shithub: libvpx

Download patch

ref: 81a67395336bdf114dc3ffbf4321d39ae26fce05
parent: 65daa4137894bd7b8bc6ec3be9fde20093ddf1c5
author: clang-format <noreply@google.com>
date: Wed Jul 13 18:26:28 EDT 2016 

vp8: apply clang-format

Change-Id: I7605b6678014a5426ceb45c27b54885e0c4e06ed


--- a/vp8/common/alloccommon.c
+++ b/vp8/common/alloccommon.c
@@ -8,7 +8,6 @@
  *  be found in the AUTHORS file in the root of the source tree.
  */
 
-
 #include "vpx_config.h"
 #include "alloccommon.h"
 #include "blockd.h"
@@ -18,176 +17,166 @@
 #include "entropymode.h"
 #include "systemdependent.h"
 
-void vp8_de_alloc_frame_buffers(VP8_COMMON *oci)
-{
-    int i;
-    for (i = 0; i < NUM_YV12_BUFFERS; i++)
-        vp8_yv12_de_alloc_frame_buffer(&oci->yv12_fb[i]);
+void vp8_de_alloc_frame_buffers(VP8_COMMON *oci) {
+  int i;
+  for (i = 0; i < NUM_YV12_BUFFERS; i++)
+    vp8_yv12_de_alloc_frame_buffer(&oci->yv12_fb[i]);
 
-    vp8_yv12_de_alloc_frame_buffer(&oci->temp_scale_frame);
+  vp8_yv12_de_alloc_frame_buffer(&oci->temp_scale_frame);
 #if CONFIG_POSTPROC
-    vp8_yv12_de_alloc_frame_buffer(&oci->post_proc_buffer);
-    if (oci->post_proc_buffer_int_used)
-        vp8_yv12_de_alloc_frame_buffer(&oci->post_proc_buffer_int);
+  vp8_yv12_de_alloc_frame_buffer(&oci->post_proc_buffer);
+  if (oci->post_proc_buffer_int_used)
+    vp8_yv12_de_alloc_frame_buffer(&oci->post_proc_buffer_int);
 
-    vpx_free(oci->pp_limits_buffer);
-    oci->pp_limits_buffer = NULL;
+  vpx_free(oci->pp_limits_buffer);
+  oci->pp_limits_buffer = NULL;
 
-    vpx_free(oci->postproc_state.generated_noise);
-    oci->postproc_state.generated_noise = NULL;
+  vpx_free(oci->postproc_state.generated_noise);
+  oci->postproc_state.generated_noise = NULL;
 #endif
 
-    vpx_free(oci->above_context);
-    vpx_free(oci->mip);
+  vpx_free(oci->above_context);
+  vpx_free(oci->mip);
 #if CONFIG_ERROR_CONCEALMENT
-    vpx_free(oci->prev_mip);
-    oci->prev_mip = NULL;
+  vpx_free(oci->prev_mip);
+  oci->prev_mip = NULL;
 #endif
 
-    oci->above_context = NULL;
-    oci->mip = NULL;
+  oci->above_context = NULL;
+  oci->mip = NULL;
 }
 
-int vp8_alloc_frame_buffers(VP8_COMMON *oci, int width, int height)
-{
-    int i;
+int vp8_alloc_frame_buffers(VP8_COMMON *oci, int width, int height) {
+  int i;
 
-    vp8_de_alloc_frame_buffers(oci);
+  vp8_de_alloc_frame_buffers(oci);
 
-    /* our internal buffers are always multiples of 16 */
-    if ((width & 0xf) != 0)
-        width += 16 - (width & 0xf);
+  /* our internal buffers are always multiples of 16 */
+  if ((width & 0xf) != 0) width += 16 - (width & 0xf);
 
-    if ((height & 0xf) != 0)
-        height += 16 - (height & 0xf);
+  if ((height & 0xf) != 0) height += 16 - (height & 0xf);
 
+  for (i = 0; i < NUM_YV12_BUFFERS; i++) {
+    oci->fb_idx_ref_cnt[i] = 0;
+    oci->yv12_fb[i].flags = 0;
+    if (vp8_yv12_alloc_frame_buffer(&oci->yv12_fb[i], width, height,
+                                    VP8BORDERINPIXELS) < 0)
+      goto allocation_fail;
+  }
 
-    for (i = 0; i < NUM_YV12_BUFFERS; i++)
-    {
-        oci->fb_idx_ref_cnt[i] = 0;
-        oci->yv12_fb[i].flags = 0;
-        if (vp8_yv12_alloc_frame_buffer(&oci->yv12_fb[i], width, height, VP8BORDERINPIXELS) < 0)
-            goto allocation_fail;
-    }
+  oci->new_fb_idx = 0;
+  oci->lst_fb_idx = 1;
+  oci->gld_fb_idx = 2;
+  oci->alt_fb_idx = 3;
 
-    oci->new_fb_idx = 0;
-    oci->lst_fb_idx = 1;
-    oci->gld_fb_idx = 2;
-    oci->alt_fb_idx = 3;
+  oci->fb_idx_ref_cnt[0] = 1;
+  oci->fb_idx_ref_cnt[1] = 1;
+  oci->fb_idx_ref_cnt[2] = 1;
+  oci->fb_idx_ref_cnt[3] = 1;
 
-    oci->fb_idx_ref_cnt[0] = 1;
-    oci->fb_idx_ref_cnt[1] = 1;
-    oci->fb_idx_ref_cnt[2] = 1;
-    oci->fb_idx_ref_cnt[3] = 1;
+  if (vp8_yv12_alloc_frame_buffer(&oci->temp_scale_frame, width, 16,
+                                  VP8BORDERINPIXELS) < 0)
+    goto allocation_fail;
 
-    if (vp8_yv12_alloc_frame_buffer(&oci->temp_scale_frame,   width, 16, VP8BORDERINPIXELS) < 0)
-        goto allocation_fail;
+  oci->mb_rows = height >> 4;
+  oci->mb_cols = width >> 4;
+  oci->MBs = oci->mb_rows * oci->mb_cols;
+  oci->mode_info_stride = oci->mb_cols + 1;
+  oci->mip =
+      vpx_calloc((oci->mb_cols + 1) * (oci->mb_rows + 1), sizeof(MODE_INFO));
 
-    oci->mb_rows = height >> 4;
-    oci->mb_cols = width >> 4;
-    oci->MBs = oci->mb_rows * oci->mb_cols;
-    oci->mode_info_stride = oci->mb_cols + 1;
-    oci->mip = vpx_calloc((oci->mb_cols + 1) * (oci->mb_rows + 1), sizeof(MODE_INFO));
+  if (!oci->mip) goto allocation_fail;
 
-    if (!oci->mip)
-        goto allocation_fail;
+  oci->mi = oci->mip + oci->mode_info_stride + 1;
 
-    oci->mi = oci->mip + oci->mode_info_stride + 1;
+  /* Allocation of previous mode info will be done in vp8_decode_frame()
+   * as it is a decoder only data */
 
-    /* Allocation of previous mode info will be done in vp8_decode_frame()
-     * as it is a decoder only data */
+  oci->above_context =
+      vpx_calloc(sizeof(ENTROPY_CONTEXT_PLANES) * oci->mb_cols, 1);
 
-    oci->above_context = vpx_calloc(sizeof(ENTROPY_CONTEXT_PLANES) * oci->mb_cols, 1);
+  if (!oci->above_context) goto allocation_fail;
 
-    if (!oci->above_context)
-        goto allocation_fail;
-
 #if CONFIG_POSTPROC
-    if (vp8_yv12_alloc_frame_buffer(&oci->post_proc_buffer, width, height, VP8BORDERINPIXELS) < 0)
-        goto allocation_fail;
+  if (vp8_yv12_alloc_frame_buffer(&oci->post_proc_buffer, width, height,
+                                  VP8BORDERINPIXELS) < 0)
+    goto allocation_fail;
 
-    oci->post_proc_buffer_int_used = 0;
-    memset(&oci->postproc_state, 0, sizeof(oci->postproc_state));
-    memset(oci->post_proc_buffer.buffer_alloc, 128,
-           oci->post_proc_buffer.frame_size);
+  oci->post_proc_buffer_int_used = 0;
+  memset(&oci->postproc_state, 0, sizeof(oci->postproc_state));
+  memset(oci->post_proc_buffer.buffer_alloc, 128,
+         oci->post_proc_buffer.frame_size);
 
-    /* Allocate buffer to store post-processing filter coefficients.
-     *
-     * Note: Round up mb_cols to support SIMD reads
-     */
-    oci->pp_limits_buffer = vpx_memalign(16, 24 * ((oci->mb_cols + 1) & ~1));
-    if (!oci->pp_limits_buffer)
-        goto allocation_fail;
+  /* Allocate buffer to store post-processing filter coefficients.
+   *
+   * Note: Round up mb_cols to support SIMD reads
+   */
+  oci->pp_limits_buffer = vpx_memalign(16, 24 * ((oci->mb_cols + 1) & ~1));
+  if (!oci->pp_limits_buffer) goto allocation_fail;
 #endif
 
-    return 0;
+  return 0;
 
 allocation_fail:
-    vp8_de_alloc_frame_buffers(oci);
-    return 1;
+  vp8_de_alloc_frame_buffers(oci);
+  return 1;
 }
 
-void vp8_setup_version(VP8_COMMON *cm)
-{
-    switch (cm->version)
-    {
+void vp8_setup_version(VP8_COMMON *cm) {
+  switch (cm->version) {
     case 0:
-        cm->no_lpf = 0;
-        cm->filter_type = NORMAL_LOOPFILTER;
-        cm->use_bilinear_mc_filter = 0;
-        cm->full_pixel = 0;
-        break;
+      cm->no_lpf = 0;
+      cm->filter_type = NORMAL_LOOPFILTER;
+      cm->use_bilinear_mc_filter = 0;
+      cm->full_pixel = 0;
+      break;
     case 1:
-        cm->no_lpf = 0;
-        cm->filter_type = SIMPLE_LOOPFILTER;
-        cm->use_bilinear_mc_filter = 1;
-        cm->full_pixel = 0;
-        break;
+      cm->no_lpf = 0;
+      cm->filter_type = SIMPLE_LOOPFILTER;
+      cm->use_bilinear_mc_filter = 1;
+      cm->full_pixel = 0;
+      break;
     case 2:
-        cm->no_lpf = 1;
-        cm->filter_type = NORMAL_LOOPFILTER;
-        cm->use_bilinear_mc_filter = 1;
-        cm->full_pixel = 0;
-        break;
+      cm->no_lpf = 1;
+      cm->filter_type = NORMAL_LOOPFILTER;
+      cm->use_bilinear_mc_filter = 1;
+      cm->full_pixel = 0;
+      break;
     case 3:
-        cm->no_lpf = 1;
-        cm->filter_type = SIMPLE_LOOPFILTER;
-        cm->use_bilinear_mc_filter = 1;
-        cm->full_pixel = 1;
-        break;
+      cm->no_lpf = 1;
+      cm->filter_type = SIMPLE_LOOPFILTER;
+      cm->use_bilinear_mc_filter = 1;
+      cm->full_pixel = 1;
+      break;
     default:
-        /*4,5,6,7 are reserved for future use*/
-        cm->no_lpf = 0;
-        cm->filter_type = NORMAL_LOOPFILTER;
-        cm->use_bilinear_mc_filter = 0;
-        cm->full_pixel = 0;
-        break;
-    }
+      /*4,5,6,7 are reserved for future use*/
+      cm->no_lpf = 0;
+      cm->filter_type = NORMAL_LOOPFILTER;
+      cm->use_bilinear_mc_filter = 0;
+      cm->full_pixel = 0;
+      break;
+  }
 }
-void vp8_create_common(VP8_COMMON *oci)
-{
-    vp8_machine_specific_config(oci);
+void vp8_create_common(VP8_COMMON *oci) {
+  vp8_machine_specific_config(oci);
 
-    vp8_init_mbmode_probs(oci);
-    vp8_default_bmode_probs(oci->fc.bmode_prob);
+  vp8_init_mbmode_probs(oci);
+  vp8_default_bmode_probs(oci->fc.bmode_prob);
 
-    oci->mb_no_coeff_skip = 1;
-    oci->no_lpf = 0;
-    oci->filter_type = NORMAL_LOOPFILTER;
-    oci->use_bilinear_mc_filter = 0;
-    oci->full_pixel = 0;
-    oci->multi_token_partition = ONE_PARTITION;
-    oci->clamp_type = RECON_CLAMP_REQUIRED;
+  oci->mb_no_coeff_skip = 1;
+  oci->no_lpf = 0;
+  oci->filter_type = NORMAL_LOOPFILTER;
+  oci->use_bilinear_mc_filter = 0;
+  oci->full_pixel = 0;
+  oci->multi_token_partition = ONE_PARTITION;
+  oci->clamp_type = RECON_CLAMP_REQUIRED;
 
-    /* Initialize reference frame sign bias structure to defaults */
-    memset(oci->ref_frame_sign_bias, 0, sizeof(oci->ref_frame_sign_bias));
+  /* Initialize reference frame sign bias structure to defaults */
+  memset(oci->ref_frame_sign_bias, 0, sizeof(oci->ref_frame_sign_bias));
 
-    /* Default disable buffer to buffer copying */
-    oci->copy_buffer_to_gf = 0;
-    oci->copy_buffer_to_arf = 0;
+  /* Default disable buffer to buffer copying */
+  oci->copy_buffer_to_gf = 0;
+  oci->copy_buffer_to_arf = 0;
 }
 
-void vp8_remove_common(VP8_COMMON *oci)
-{
-    vp8_de_alloc_frame_buffers(oci);
-}
+void vp8_remove_common(VP8_COMMON *oci) { vp8_de_alloc_frame_buffers(oci); }
--- a/vp8/common/alloccommon.h
+++ b/vp8/common/alloccommon.h
@@ -8,7 +8,6 @@
  *  be found in the AUTHORS file in the root of the source tree.
  */
 
-
 #ifndef VP8_COMMON_ALLOCCOMMON_H_
 #define VP8_COMMON_ALLOCCOMMON_H_
 
--- a/vp8/common/arm/armv6/idct_blk_v6.c
+++ b/vp8/common/arm/armv6/idct_blk_v6.c
@@ -11,105 +11,90 @@
 #include "vpx_config.h"
 #include "vp8_rtcd.h"
 
+void vp8_dequant_idct_add_y_block_v6(short *q, short *dq, unsigned char *dst,
+                                     int stride, char *eobs) {
+  int i;
 
-void vp8_dequant_idct_add_y_block_v6(short *q, short *dq,
-                                     unsigned char *dst,
-                                     int stride, char *eobs)
-{
-    int i;
+  for (i = 0; i < 4; i++) {
+    if (eobs[0] > 1)
+      vp8_dequant_idct_add_v6(q, dq, dst, stride);
+    else if (eobs[0] == 1) {
+      vp8_dc_only_idct_add_v6(q[0] * dq[0], dst, stride, dst, stride);
+      ((int *)q)[0] = 0;
+    }
 
-    for (i = 0; i < 4; i++)
-    {
-        if (eobs[0] > 1)
-            vp8_dequant_idct_add_v6 (q, dq, dst, stride);
-        else if (eobs[0] == 1)
-        {
-            vp8_dc_only_idct_add_v6 (q[0]*dq[0], dst, stride, dst, stride);
-            ((int *)q)[0] = 0;
-        }
+    if (eobs[1] > 1)
+      vp8_dequant_idct_add_v6(q + 16, dq, dst + 4, stride);
+    else if (eobs[1] == 1) {
+      vp8_dc_only_idct_add_v6(q[16] * dq[0], dst + 4, stride, dst + 4, stride);
+      ((int *)(q + 16))[0] = 0;
+    }
 
-        if (eobs[1] > 1)
-            vp8_dequant_idct_add_v6 (q+16, dq, dst+4, stride);
-        else if (eobs[1] == 1)
-        {
-            vp8_dc_only_idct_add_v6 (q[16]*dq[0], dst+4, stride, dst+4, stride);
-            ((int *)(q+16))[0] = 0;
-        }
+    if (eobs[2] > 1)
+      vp8_dequant_idct_add_v6(q + 32, dq, dst + 8, stride);
+    else if (eobs[2] == 1) {
+      vp8_dc_only_idct_add_v6(q[32] * dq[0], dst + 8, stride, dst + 8, stride);
+      ((int *)(q + 32))[0] = 0;
+    }
 
-        if (eobs[2] > 1)
-            vp8_dequant_idct_add_v6 (q+32, dq, dst+8, stride);
-        else if (eobs[2] == 1)
-        {
-            vp8_dc_only_idct_add_v6 (q[32]*dq[0], dst+8, stride, dst+8, stride);
-            ((int *)(q+32))[0] = 0;
-        }
-
-        if (eobs[3] > 1)
-            vp8_dequant_idct_add_v6 (q+48, dq, dst+12, stride);
-        else if (eobs[3] == 1)
-        {
-            vp8_dc_only_idct_add_v6 (q[48]*dq[0], dst+12, stride,dst+12,stride);
-            ((int *)(q+48))[0] = 0;
-        }
-
-        q    += 64;
-        dst  += 4*stride;
-        eobs += 4;
+    if (eobs[3] > 1)
+      vp8_dequant_idct_add_v6(q + 48, dq, dst + 12, stride);
+    else if (eobs[3] == 1) {
+      vp8_dc_only_idct_add_v6(q[48] * dq[0], dst + 12, stride, dst + 12,
+                              stride);
+      ((int *)(q + 48))[0] = 0;
     }
+
+    q += 64;
+    dst += 4 * stride;
+    eobs += 4;
+  }
 }
 
-void vp8_dequant_idct_add_uv_block_v6(short *q, short *dq,
-                                      unsigned char *dstu,
-                                      unsigned char *dstv,
-                                      int stride, char *eobs)
-{
-    int i;
+void vp8_dequant_idct_add_uv_block_v6(short *q, short *dq, unsigned char *dstu,
+                                      unsigned char *dstv, int stride,
+                                      char *eobs) {
+  int i;
 
-    for (i = 0; i < 2; i++)
-    {
-        if (eobs[0] > 1)
-            vp8_dequant_idct_add_v6 (q, dq, dstu, stride);
-        else if (eobs[0] == 1)
-        {
-            vp8_dc_only_idct_add_v6 (q[0]*dq[0], dstu, stride, dstu, stride);
-            ((int *)q)[0] = 0;
-        }
+  for (i = 0; i < 2; i++) {
+    if (eobs[0] > 1)
+      vp8_dequant_idct_add_v6(q, dq, dstu, stride);
+    else if (eobs[0] == 1) {
+      vp8_dc_only_idct_add_v6(q[0] * dq[0], dstu, stride, dstu, stride);
+      ((int *)q)[0] = 0;
+    }
 
-        if (eobs[1] > 1)
-            vp8_dequant_idct_add_v6 (q+16, dq, dstu+4, stride);
-        else if (eobs[1] == 1)
-        {
-            vp8_dc_only_idct_add_v6 (q[16]*dq[0], dstu+4, stride,
-                                                  dstu+4, stride);
-            ((int *)(q+16))[0] = 0;
-        }
-
-        q    += 32;
-        dstu += 4*stride;
-        eobs += 2;
+    if (eobs[1] > 1)
+      vp8_dequant_idct_add_v6(q + 16, dq, dstu + 4, stride);
+    else if (eobs[1] == 1) {
+      vp8_dc_only_idct_add_v6(q[16] * dq[0], dstu + 4, stride, dstu + 4,
+                              stride);
+      ((int *)(q + 16))[0] = 0;
     }
 
-    for (i = 0; i < 2; i++)
-    {
-        if (eobs[0] > 1)
-            vp8_dequant_idct_add_v6 (q, dq, dstv, stride);
-        else if (eobs[0] == 1)
-        {
-            vp8_dc_only_idct_add_v6 (q[0]*dq[0], dstv, stride, dstv, stride);
-            ((int *)q)[0] = 0;
-        }
+    q += 32;
+    dstu += 4 * stride;
+    eobs += 2;
+  }
 
-        if (eobs[1] > 1)
-            vp8_dequant_idct_add_v6 (q+16, dq, dstv+4, stride);
-        else if (eobs[1] == 1)
-        {
-            vp8_dc_only_idct_add_v6 (q[16]*dq[0], dstv+4, stride,
-                                                  dstv+4, stride);
-            ((int *)(q+16))[0] = 0;
-        }
+  for (i = 0; i < 2; i++) {
+    if (eobs[0] > 1)
+      vp8_dequant_idct_add_v6(q, dq, dstv, stride);
+    else if (eobs[0] == 1) {
+      vp8_dc_only_idct_add_v6(q[0] * dq[0], dstv, stride, dstv, stride);
+      ((int *)q)[0] = 0;
+    }
 
-        q    += 32;
-        dstv += 4*stride;
-        eobs += 2;
+    if (eobs[1] > 1)
+      vp8_dequant_idct_add_v6(q + 16, dq, dstv + 4, stride);
+    else if (eobs[1] == 1) {
+      vp8_dc_only_idct_add_v6(q[16] * dq[0], dstv + 4, stride, dstv + 4,
+                              stride);
+      ((int *)(q + 16))[0] = 0;
     }
+
+    q += 32;
+    dstv += 4 * stride;
+    eobs += 2;
+  }
 }
--- a/vp8/common/arm/bilinearfilter_arm.c
+++ b/vp8/common/arm/bilinearfilter_arm.c
@@ -14,100 +14,74 @@
 #include "vp8/common/filter.h"
 #include "bilinearfilter_arm.h"
 
-void vp8_filter_block2d_bil_armv6
-(
-    unsigned char *src_ptr,
-    unsigned char *dst_ptr,
-    unsigned int   src_pitch,
-    unsigned int   dst_pitch,
-    const short   *HFilter,
-    const short   *VFilter,
-    int            Width,
-    int            Height
-)
-{
-    unsigned short FData[36*16]; /* Temp data buffer used in filtering */
+void vp8_filter_block2d_bil_armv6(unsigned char *src_ptr,
+                                  unsigned char *dst_ptr,
+                                  unsigned int src_pitch,
+                                  unsigned int dst_pitch, const short *HFilter,
+                                  const short *VFilter, int Width, int Height) {
+  unsigned short FData[36 * 16]; /* Temp data buffer used in filtering */
 
-    /* First filter 1-D horizontally... */
-    vp8_filter_block2d_bil_first_pass_armv6(src_ptr, FData, src_pitch, Height + 1, Width, HFilter);
+  /* First filter 1-D horizontally... */
+  vp8_filter_block2d_bil_first_pass_armv6(src_ptr, FData, src_pitch, Height + 1,
+                                          Width, HFilter);
 
-    /* then 1-D vertically... */
-    vp8_filter_block2d_bil_second_pass_armv6(FData, dst_ptr, dst_pitch, Height, Width, VFilter);
+  /* then 1-D vertically... */
+  vp8_filter_block2d_bil_second_pass_armv6(FData, dst_ptr, dst_pitch, Height,
+                                           Width, VFilter);
 }
 
+void vp8_bilinear_predict4x4_armv6(unsigned char *src_ptr,
+                                   int src_pixels_per_line, int xoffset,
+                                   int yoffset, unsigned char *dst_ptr,
+                                   int dst_pitch) {
+  const short *HFilter;
+  const short *VFilter;
 
-void vp8_bilinear_predict4x4_armv6
-(
-    unsigned char  *src_ptr,
-    int   src_pixels_per_line,
-    int  xoffset,
-    int  yoffset,
-    unsigned char *dst_ptr,
-    int dst_pitch
-)
-{
-    const short  *HFilter;
-    const short  *VFilter;
+  HFilter = vp8_bilinear_filters[xoffset];
+  VFilter = vp8_bilinear_filters[yoffset];
 
-    HFilter = vp8_bilinear_filters[xoffset];
-    VFilter = vp8_bilinear_filters[yoffset];
-
-    vp8_filter_block2d_bil_armv6(src_ptr, dst_ptr, src_pixels_per_line, dst_pitch, HFilter, VFilter, 4, 4);
+  vp8_filter_block2d_bil_armv6(src_ptr, dst_ptr, src_pixels_per_line, dst_pitch,
+                               HFilter, VFilter, 4, 4);
 }
 
-void vp8_bilinear_predict8x8_armv6
-(
-    unsigned char  *src_ptr,
-    int  src_pixels_per_line,
-    int  xoffset,
-    int  yoffset,
-    unsigned char *dst_ptr,
-    int  dst_pitch
-)
-{
-    const short  *HFilter;
-    const short  *VFilter;
+void vp8_bilinear_predict8x8_armv6(unsigned char *src_ptr,
+                                   int src_pixels_per_line, int xoffset,
+                                   int yoffset, unsigned char *dst_ptr,
+                                   int dst_pitch) {
+  const short *HFilter;
+  const short *VFilter;
 
-    HFilter = vp8_bilinear_filters[xoffset];
-    VFilter = vp8_bilinear_filters[yoffset];
+  HFilter = vp8_bilinear_filters[xoffset];
+  VFilter = vp8_bilinear_filters[yoffset];
 
-    vp8_filter_block2d_bil_armv6(src_ptr, dst_ptr, src_pixels_per_line, dst_pitch, HFilter, VFilter, 8, 8);
+  vp8_filter_block2d_bil_armv6(src_ptr, dst_ptr, src_pixels_per_line, dst_pitch,
+                               HFilter, VFilter, 8, 8);
 }
 
-void vp8_bilinear_predict8x4_armv6
-(
-    unsigned char  *src_ptr,
-    int  src_pixels_per_line,
-    int  xoffset,
-    int  yoffset,
-    unsigned char *dst_ptr,
-    int  dst_pitch
-)
-{
-    const short  *HFilter;
-    const short  *VFilter;
+void vp8_bilinear_predict8x4_armv6(unsigned char *src_ptr,
+                                   int src_pixels_per_line, int xoffset,
+                                   int yoffset, unsigned char *dst_ptr,
+                                   int dst_pitch) {
+  const short *HFilter;
+  const short *VFilter;
 
-    HFilter = vp8_bilinear_filters[xoffset];
-    VFilter = vp8_bilinear_filters[yoffset];
+  HFilter = vp8_bilinear_filters[xoffset];
+  VFilter = vp8_bilinear_filters[yoffset];
 
-    vp8_filter_block2d_bil_armv6(src_ptr, dst_ptr, src_pixels_per_line, dst_pitch, HFilter, VFilter, 8, 4);
+  vp8_filter_block2d_bil_armv6(src_ptr, dst_ptr, src_pixels_per_line, dst_pitch,
+                               HFilter, VFilter, 8, 4);
 }
 
-void vp8_bilinear_predict16x16_armv6
-(
-    unsigned char  *src_ptr,
-    int  src_pixels_per_line,
-    int  xoffset,
-    int  yoffset,
-    unsigned char *dst_ptr,
-    int  dst_pitch
-)
-{
-    const short  *HFilter;
-    const short  *VFilter;
+void vp8_bilinear_predict16x16_armv6(unsigned char *src_ptr,
+                                     int src_pixels_per_line, int xoffset,
+                                     int yoffset, unsigned char *dst_ptr,
+                                     int dst_pitch) {
+  const short *HFilter;
+  const short *VFilter;
 
-    HFilter = vp8_bilinear_filters[xoffset];
-    VFilter = vp8_bilinear_filters[yoffset];
+  HFilter = vp8_bilinear_filters[xoffset];
+  VFilter = vp8_bilinear_filters[yoffset];
 
-    vp8_filter_block2d_bil_armv6(src_ptr, dst_ptr, src_pixels_per_line, dst_pitch, HFilter, VFilter, 16, 16);
+  vp8_filter_block2d_bil_armv6(src_ptr, dst_ptr, src_pixels_per_line, dst_pitch,
+                               HFilter, VFilter, 16, 16);
 }
--- a/vp8/common/arm/bilinearfilter_arm.h
+++ b/vp8/common/arm/bilinearfilter_arm.h
@@ -8,7 +8,6 @@
  *  be found in the AUTHORS file in the root of the source tree.
  */
 
-
 #ifndef VP8_COMMON_ARM_BILINEARFILTER_ARM_H_
 #define VP8_COMMON_ARM_BILINEARFILTER_ARM_H_
 
@@ -16,25 +15,14 @@
 extern "C" {
 #endif
 
-extern void vp8_filter_block2d_bil_first_pass_armv6
-(
-    const unsigned char  *src_ptr,
-    unsigned short       *dst_ptr,
-    unsigned int          src_pitch,
-    unsigned int          height,
-    unsigned int          width,
-    const short          *vp8_filter
-);
+extern void vp8_filter_block2d_bil_first_pass_armv6(
+    const unsigned char *src_ptr, unsigned short *dst_ptr,
+    unsigned int src_pitch, unsigned int height, unsigned int width,
+    const short *vp8_filter);
 
-extern void vp8_filter_block2d_bil_second_pass_armv6
-(
-    const unsigned short *src_ptr,
-    unsigned char        *dst_ptr,
-    int                   dst_pitch,
-    unsigned int          height,
-    unsigned int          width,
-    const short         *vp8_filter
-);
+extern void vp8_filter_block2d_bil_second_pass_armv6(
+    const unsigned short *src_ptr, unsigned char *dst_ptr, int dst_pitch,
+    unsigned int height, unsigned int width, const short *vp8_filter);
 
 #ifdef __cplusplus
 }  // extern "C"
--- a/vp8/common/arm/dequantize_arm.c
+++ b/vp8/common/arm/dequantize_arm.c
@@ -8,7 +8,6 @@
  *  be found in the AUTHORS file in the root of the source tree.
  */
 
-
 #include "vpx_config.h"
 #include "vp8/common/blockd.h"
 
@@ -15,11 +14,10 @@
 #if HAVE_MEDIA
 extern void vp8_dequantize_b_loop_v6(short *Q, short *DQC, short *DQ);
 
-void vp8_dequantize_b_v6(BLOCKD *d, short *DQC)
-{
-    short *DQ  = d->dqcoeff;
-    short *Q   = d->qcoeff;
+void vp8_dequantize_b_v6(BLOCKD *d, short *DQC) {
+  short *DQ = d->dqcoeff;
+  short *Q = d->qcoeff;
 
-    vp8_dequantize_b_loop_v6(Q, DQC, DQ);
+  vp8_dequantize_b_loop_v6(Q, DQC, DQ);
 }
 #endif
--- a/vp8/common/arm/filter_arm.c
+++ b/vp8/common/arm/filter_arm.c
@@ -8,7 +8,6 @@
  *  be found in the AUTHORS file in the root of the source tree.
  */
 
-
 #include "vpx_config.h"
 #include "vp8_rtcd.h"
 #include <math.h>
@@ -15,207 +14,163 @@
 #include "vp8/common/filter.h"
 #include "vpx_ports/mem.h"
 
-extern void vp8_filter_block2d_first_pass_armv6
-(
-    unsigned char *src_ptr,
-    short         *output_ptr,
-    unsigned int src_pixels_per_line,
-    unsigned int output_width,
-    unsigned int output_height,
-    const short *vp8_filter
-);
+extern void vp8_filter_block2d_first_pass_armv6(
+    unsigned char *src_ptr, short *output_ptr, unsigned int src_pixels_per_line,
+    unsigned int output_width, unsigned int output_height,
+    const short *vp8_filter);
 
 // 8x8
-extern void vp8_filter_block2d_first_pass_8x8_armv6
-(
-    unsigned char *src_ptr,
-    short         *output_ptr,
-    unsigned int src_pixels_per_line,
-    unsigned int output_width,
-    unsigned int output_height,
-    const short *vp8_filter
-);
+extern void vp8_filter_block2d_first_pass_8x8_armv6(
+    unsigned char *src_ptr, short *output_ptr, unsigned int src_pixels_per_line,
+    unsigned int output_width, unsigned int output_height,
+    const short *vp8_filter);
 
 // 16x16
-extern void vp8_filter_block2d_first_pass_16x16_armv6
-(
-    unsigned char *src_ptr,
-    short         *output_ptr,
-    unsigned int src_pixels_per_line,
-    unsigned int output_width,
-    unsigned int output_height,
-    const short *vp8_filter
-);
+extern void vp8_filter_block2d_first_pass_16x16_armv6(
+    unsigned char *src_ptr, short *output_ptr, unsigned int src_pixels_per_line,
+    unsigned int output_width, unsigned int output_height,
+    const short *vp8_filter);
 
-extern void vp8_filter_block2d_second_pass_armv6
-(
-    short         *src_ptr,
-    unsigned char *output_ptr,
-    unsigned int output_pitch,
-    unsigned int cnt,
-    const short *vp8_filter
-);
+extern void vp8_filter_block2d_second_pass_armv6(short *src_ptr,
+                                                 unsigned char *output_ptr,
+                                                 unsigned int output_pitch,
+                                                 unsigned int cnt,
+                                                 const short *vp8_filter);
 
-extern void vp8_filter4_block2d_second_pass_armv6
-(
-    short         *src_ptr,
-    unsigned char *output_ptr,
-    unsigned int output_pitch,
-    unsigned int cnt,
-    const short *vp8_filter
-);
+extern void vp8_filter4_block2d_second_pass_armv6(short *src_ptr,
+                                                  unsigned char *output_ptr,
+                                                  unsigned int output_pitch,
+                                                  unsigned int cnt,
+                                                  const short *vp8_filter);
 
-extern void vp8_filter_block2d_first_pass_only_armv6
-(
-    unsigned char *src_ptr,
-    unsigned char *output_ptr,
-    unsigned int src_pixels_per_line,
-    unsigned int cnt,
-    unsigned int output_pitch,
-    const short *vp8_filter
-);
+extern void vp8_filter_block2d_first_pass_only_armv6(
+    unsigned char *src_ptr, unsigned char *output_ptr,
+    unsigned int src_pixels_per_line, unsigned int cnt,
+    unsigned int output_pitch, const short *vp8_filter);
 
+extern void vp8_filter_block2d_second_pass_only_armv6(
+    unsigned char *src_ptr, unsigned char *output_ptr,
+    unsigned int src_pixels_per_line, unsigned int cnt,
+    unsigned int output_pitch, const short *vp8_filter);
 
-extern void vp8_filter_block2d_second_pass_only_armv6
-(
-    unsigned char *src_ptr,
-    unsigned char *output_ptr,
-    unsigned int src_pixels_per_line,
-    unsigned int cnt,
-    unsigned int output_pitch,
-    const short *vp8_filter
-);
-
 #if HAVE_MEDIA
-void vp8_sixtap_predict4x4_armv6
-(
-    unsigned char  *src_ptr,
-    int  src_pixels_per_line,
-    int  xoffset,
-    int  yoffset,
-    unsigned char *dst_ptr,
-    int  dst_pitch
-)
-{
-    const short  *HFilter;
-    const short  *VFilter;
-    DECLARE_ALIGNED(4, short, FData[12*4]); /* Temp data buffer used in filtering */
+void vp8_sixtap_predict4x4_armv6(unsigned char *src_ptr,
+                                 int src_pixels_per_line, int xoffset,
+                                 int yoffset, unsigned char *dst_ptr,
+                                 int dst_pitch) {
+  const short *HFilter;
+  const short *VFilter;
+  DECLARE_ALIGNED(4, short,
+                  FData[12 * 4]); /* Temp data buffer used in filtering */
 
+  HFilter = vp8_sub_pel_filters[xoffset]; /* 6 tap */
+  VFilter = vp8_sub_pel_filters[yoffset]; /* 6 tap */
 
-    HFilter = vp8_sub_pel_filters[xoffset];   /* 6 tap */
-    VFilter = vp8_sub_pel_filters[yoffset];   /* 6 tap */
+  /* Vfilter is null. First pass only */
+  if (xoffset && !yoffset) {
+    /*vp8_filter_block2d_first_pass_armv6 ( src_ptr, FData+2,
+    src_pixels_per_line, 4, 4, HFilter );
+    vp8_filter_block2d_second_pass_armv6 ( FData+2, dst_ptr, dst_pitch, 4,
+    VFilter );*/
 
-    /* Vfilter is null. First pass only */
-    if (xoffset && !yoffset)
-    {
-        /*vp8_filter_block2d_first_pass_armv6 ( src_ptr, FData+2, src_pixels_per_line, 4, 4, HFilter );
-        vp8_filter_block2d_second_pass_armv6 ( FData+2, dst_ptr, dst_pitch, 4, VFilter );*/
-
-        vp8_filter_block2d_first_pass_only_armv6(src_ptr, dst_ptr, src_pixels_per_line, 4, dst_pitch, HFilter);
+    vp8_filter_block2d_first_pass_only_armv6(
+        src_ptr, dst_ptr, src_pixels_per_line, 4, dst_pitch, HFilter);
+  }
+  /* Hfilter is null. Second pass only */
+  else if (!xoffset && yoffset) {
+    vp8_filter_block2d_second_pass_only_armv6(
+        src_ptr, dst_ptr, src_pixels_per_line, 4, dst_pitch, VFilter);
+  } else {
+    /* Vfilter is a 4 tap filter */
+    if (yoffset & 0x1) {
+      vp8_filter_block2d_first_pass_armv6(src_ptr - src_pixels_per_line,
+                                          FData + 1, src_pixels_per_line, 4, 7,
+                                          HFilter);
+      vp8_filter4_block2d_second_pass_armv6(FData + 2, dst_ptr, dst_pitch, 4,
+                                            VFilter);
     }
-    /* Hfilter is null. Second pass only */
-    else if (!xoffset && yoffset)
-    {
-        vp8_filter_block2d_second_pass_only_armv6(src_ptr, dst_ptr, src_pixels_per_line, 4, dst_pitch, VFilter);
+    /* Vfilter is 6 tap filter */
+    else {
+      vp8_filter_block2d_first_pass_armv6(src_ptr - (2 * src_pixels_per_line),
+                                          FData, src_pixels_per_line, 4, 9,
+                                          HFilter);
+      vp8_filter_block2d_second_pass_armv6(FData + 2, dst_ptr, dst_pitch, 4,
+                                           VFilter);
     }
-    else
-    {
-        /* Vfilter is a 4 tap filter */
-        if (yoffset & 0x1)
-        {
-            vp8_filter_block2d_first_pass_armv6(src_ptr - src_pixels_per_line, FData + 1, src_pixels_per_line, 4, 7, HFilter);
-            vp8_filter4_block2d_second_pass_armv6(FData + 2, dst_ptr, dst_pitch, 4, VFilter);
-        }
-        /* Vfilter is 6 tap filter */
-        else
-        {
-            vp8_filter_block2d_first_pass_armv6(src_ptr - (2 * src_pixels_per_line), FData, src_pixels_per_line, 4, 9, HFilter);
-            vp8_filter_block2d_second_pass_armv6(FData + 2, dst_ptr, dst_pitch, 4, VFilter);
-        }
-    }
+  }
 }
 
-void vp8_sixtap_predict8x8_armv6
-(
-    unsigned char  *src_ptr,
-    int  src_pixels_per_line,
-    int  xoffset,
-    int  yoffset,
-    unsigned char *dst_ptr,
-    int  dst_pitch
-)
-{
-    const short  *HFilter;
-    const short  *VFilter;
-    DECLARE_ALIGNED(4, short, FData[16*8]); /* Temp data buffer used in filtering */
+void vp8_sixtap_predict8x8_armv6(unsigned char *src_ptr,
+                                 int src_pixels_per_line, int xoffset,
+                                 int yoffset, unsigned char *dst_ptr,
+                                 int dst_pitch) {
+  const short *HFilter;
+  const short *VFilter;
+  DECLARE_ALIGNED(4, short,
+                  FData[16 * 8]); /* Temp data buffer used in filtering */
 
-    HFilter = vp8_sub_pel_filters[xoffset];   /* 6 tap */
-    VFilter = vp8_sub_pel_filters[yoffset];   /* 6 tap */
+  HFilter = vp8_sub_pel_filters[xoffset]; /* 6 tap */
+  VFilter = vp8_sub_pel_filters[yoffset]; /* 6 tap */
 
-    if (xoffset && !yoffset)
-    {
-        vp8_filter_block2d_first_pass_only_armv6(src_ptr, dst_ptr, src_pixels_per_line, 8, dst_pitch, HFilter);
+  if (xoffset && !yoffset) {
+    vp8_filter_block2d_first_pass_only_armv6(
+        src_ptr, dst_ptr, src_pixels_per_line, 8, dst_pitch, HFilter);
+  }
+  /* Hfilter is null. Second pass only */
+  else if (!xoffset && yoffset) {
+    vp8_filter_block2d_second_pass_only_armv6(
+        src_ptr, dst_ptr, src_pixels_per_line, 8, dst_pitch, VFilter);
+  } else {
+    if (yoffset & 0x1) {
+      vp8_filter_block2d_first_pass_8x8_armv6(src_ptr - src_pixels_per_line,
+                                              FData + 1, src_pixels_per_line, 8,
+                                              11, HFilter);
+      vp8_filter4_block2d_second_pass_armv6(FData + 2, dst_ptr, dst_pitch, 8,
+                                            VFilter);
+    } else {
+      vp8_filter_block2d_first_pass_8x8_armv6(
+          src_ptr - (2 * src_pixels_per_line), FData, src_pixels_per_line, 8,
+          13, HFilter);
+      vp8_filter_block2d_second_pass_armv6(FData + 2, dst_ptr, dst_pitch, 8,
+                                           VFilter);
     }
-    /* Hfilter is null. Second pass only */
-    else if (!xoffset && yoffset)
-    {
-        vp8_filter_block2d_second_pass_only_armv6(src_ptr, dst_ptr, src_pixels_per_line, 8, dst_pitch, VFilter);
-    }
-    else
-    {
-        if (yoffset & 0x1)
-        {
-            vp8_filter_block2d_first_pass_8x8_armv6(src_ptr - src_pixels_per_line, FData + 1, src_pixels_per_line, 8, 11, HFilter);
-            vp8_filter4_block2d_second_pass_armv6(FData + 2, dst_ptr, dst_pitch, 8, VFilter);
-        }
-        else
-        {
-            vp8_filter_block2d_first_pass_8x8_armv6(src_ptr - (2 * src_pixels_per_line), FData, src_pixels_per_line, 8, 13, HFilter);
-            vp8_filter_block2d_second_pass_armv6(FData + 2, dst_ptr, dst_pitch, 8, VFilter);
-        }
-    }
+  }
 }
 
+void vp8_sixtap_predict16x16_armv6(unsigned char *src_ptr,
+                                   int src_pixels_per_line, int xoffset,
+                                   int yoffset, unsigned char *dst_ptr,
+                                   int dst_pitch) {
+  const short *HFilter;
+  const short *VFilter;
+  DECLARE_ALIGNED(4, short,
+                  FData[24 * 16]); /* Temp data buffer used in filtering */
 
-void vp8_sixtap_predict16x16_armv6
-(
-    unsigned char  *src_ptr,
-    int  src_pixels_per_line,
-    int  xoffset,
-    int  yoffset,
-    unsigned char *dst_ptr,
-    int  dst_pitch
-)
-{
-    const short  *HFilter;
-    const short  *VFilter;
-    DECLARE_ALIGNED(4, short, FData[24*16]);    /* Temp data buffer used in filtering */
+  HFilter = vp8_sub_pel_filters[xoffset]; /* 6 tap */
+  VFilter = vp8_sub_pel_filters[yoffset]; /* 6 tap */
 
-    HFilter = vp8_sub_pel_filters[xoffset];   /* 6 tap */
-    VFilter = vp8_sub_pel_filters[yoffset];   /* 6 tap */
-
-    if (xoffset && !yoffset)
-    {
-        vp8_filter_block2d_first_pass_only_armv6(src_ptr, dst_ptr, src_pixels_per_line, 16, dst_pitch, HFilter);
+  if (xoffset && !yoffset) {
+    vp8_filter_block2d_first_pass_only_armv6(
+        src_ptr, dst_ptr, src_pixels_per_line, 16, dst_pitch, HFilter);
+  }
+  /* Hfilter is null. Second pass only */
+  else if (!xoffset && yoffset) {
+    vp8_filter_block2d_second_pass_only_armv6(
+        src_ptr, dst_ptr, src_pixels_per_line, 16, dst_pitch, VFilter);
+  } else {
+    if (yoffset & 0x1) {
+      vp8_filter_block2d_first_pass_16x16_armv6(src_ptr - src_pixels_per_line,
+                                                FData + 1, src_pixels_per_line,
+                                                16, 19, HFilter);
+      vp8_filter4_block2d_second_pass_armv6(FData + 2, dst_ptr, dst_pitch, 16,
+                                            VFilter);
+    } else {
+      vp8_filter_block2d_first_pass_16x16_armv6(
+          src_ptr - (2 * src_pixels_per_line), FData, src_pixels_per_line, 16,
+          21, HFilter);
+      vp8_filter_block2d_second_pass_armv6(FData + 2, dst_ptr, dst_pitch, 16,
+                                           VFilter);
     }
-    /* Hfilter is null. Second pass only */
-    else if (!xoffset && yoffset)
-    {
-        vp8_filter_block2d_second_pass_only_armv6(src_ptr, dst_ptr, src_pixels_per_line, 16, dst_pitch, VFilter);
-    }
-    else
-    {
-        if (yoffset & 0x1)
-        {
-            vp8_filter_block2d_first_pass_16x16_armv6(src_ptr - src_pixels_per_line, FData + 1, src_pixels_per_line, 16, 19, HFilter);
-            vp8_filter4_block2d_second_pass_armv6(FData + 2, dst_ptr, dst_pitch, 16, VFilter);
-        }
-        else
-        {
-            vp8_filter_block2d_first_pass_16x16_armv6(src_ptr - (2 * src_pixels_per_line), FData, src_pixels_per_line, 16, 21, HFilter);
-            vp8_filter_block2d_second_pass_armv6(FData + 2, dst_ptr, dst_pitch, 16, VFilter);
-        }
-    }
-
+  }
 }
 #endif
--- a/vp8/common/arm/loopfilter_arm.c
+++ b/vp8/common/arm/loopfilter_arm.c
@@ -8,15 +8,14 @@
  *  be found in the AUTHORS file in the root of the source tree.
  */
 
-
 #include "vpx_config.h"
 #include "vp8_rtcd.h"
 #include "vp8/common/loopfilter.h"
 #include "vp8/common/onyxc_int.h"
 
-#define prototype_loopfilter(sym) \
-    void sym(unsigned char *src, int pitch, const unsigned char *blimit,\
-             const unsigned char *limit, const unsigned char *thresh, int count)
+#define prototype_loopfilter(sym)                                      \
+  void sym(unsigned char *src, int pitch, const unsigned char *blimit, \
+           const unsigned char *limit, const unsigned char *thresh, int count)
 
 #if HAVE_MEDIA
 extern prototype_loopfilter(vp8_loop_filter_horizontal_edge_armv6);
@@ -27,10 +26,11 @@
 
 #if HAVE_NEON
 typedef void loopfilter_y_neon(unsigned char *src, int pitch,
-        unsigned char blimit, unsigned char limit, unsigned char thresh);
+                               unsigned char blimit, unsigned char limit,
+                               unsigned char thresh);
 typedef void loopfilter_uv_neon(unsigned char *u, int pitch,
-        unsigned char blimit, unsigned char limit, unsigned char thresh,
-        unsigned char *v);
+                                unsigned char blimit, unsigned char limit,
+                                unsigned char thresh, unsigned char *v);
 
 extern loopfilter_y_neon vp8_loop_filter_horizontal_edge_y_neon;
 extern loopfilter_y_neon vp8_loop_filter_vertical_edge_y_neon;
@@ -46,75 +46,92 @@
 #if HAVE_MEDIA
 /* ARMV6/MEDIA loopfilter functions*/
 /* Horizontal MB filtering */
-void vp8_loop_filter_mbh_armv6(unsigned char *y_ptr, unsigned char *u_ptr, unsigned char *v_ptr,
-                               int y_stride, int uv_stride, loop_filter_info *lfi)
-{
-    vp8_mbloop_filter_horizontal_edge_armv6(y_ptr, y_stride, lfi->mblim, lfi->lim, lfi->hev_thr, 2);
+void vp8_loop_filter_mbh_armv6(unsigned char *y_ptr, unsigned char *u_ptr,
+                               unsigned char *v_ptr, int y_stride,
+                               int uv_stride, loop_filter_info *lfi) {
+  vp8_mbloop_filter_horizontal_edge_armv6(y_ptr, y_stride, lfi->mblim, lfi->lim,
+                                          lfi->hev_thr, 2);
 
-    if (u_ptr)
-        vp8_mbloop_filter_horizontal_edge_armv6(u_ptr, uv_stride, lfi->mblim, lfi->lim, lfi->hev_thr, 1);
+  if (u_ptr)
+    vp8_mbloop_filter_horizontal_edge_armv6(u_ptr, uv_stride, lfi->mblim,
+                                            lfi->lim, lfi->hev_thr, 1);
 
-    if (v_ptr)
-        vp8_mbloop_filter_horizontal_edge_armv6(v_ptr, uv_stride, lfi->mblim, lfi->lim, lfi->hev_thr, 1);
+  if (v_ptr)
+    vp8_mbloop_filter_horizontal_edge_armv6(v_ptr, uv_stride, lfi->mblim,
+                                            lfi->lim, lfi->hev_thr, 1);
 }
 
 /* Vertical MB Filtering */
-void vp8_loop_filter_mbv_armv6(unsigned char *y_ptr, unsigned char *u_ptr, unsigned char *v_ptr,
-                               int y_stride, int uv_stride, loop_filter_info *lfi)
-{
-    vp8_mbloop_filter_vertical_edge_armv6(y_ptr, y_stride, lfi->mblim, lfi->lim, lfi->hev_thr, 2);
+void vp8_loop_filter_mbv_armv6(unsigned char *y_ptr, unsigned char *u_ptr,
+                               unsigned char *v_ptr, int y_stride,
+                               int uv_stride, loop_filter_info *lfi) {
+  vp8_mbloop_filter_vertical_edge_armv6(y_ptr, y_stride, lfi->mblim, lfi->lim,
+                                        lfi->hev_thr, 2);
 
-    if (u_ptr)
-        vp8_mbloop_filter_vertical_edge_armv6(u_ptr, uv_stride, lfi->mblim, lfi->lim, lfi->hev_thr, 1);
+  if (u_ptr)
+    vp8_mbloop_filter_vertical_edge_armv6(u_ptr, uv_stride, lfi->mblim,
+                                          lfi->lim, lfi->hev_thr, 1);
 
-    if (v_ptr)
-        vp8_mbloop_filter_vertical_edge_armv6(v_ptr, uv_stride, lfi->mblim, lfi->lim, lfi->hev_thr, 1);
+  if (v_ptr)
+    vp8_mbloop_filter_vertical_edge_armv6(v_ptr, uv_stride, lfi->mblim,
+                                          lfi->lim, lfi->hev_thr, 1);
 }
 
 /* Horizontal B Filtering */
-void vp8_loop_filter_bh_armv6(unsigned char *y_ptr, unsigned char *u_ptr, unsigned char *v_ptr,
-                              int y_stride, int uv_stride, loop_filter_info *lfi)
-{
-    vp8_loop_filter_horizontal_edge_armv6(y_ptr + 4 * y_stride, y_stride, lfi->blim, lfi->lim, lfi->hev_thr, 2);
-    vp8_loop_filter_horizontal_edge_armv6(y_ptr + 8 * y_stride, y_stride, lfi->blim, lfi->lim, lfi->hev_thr, 2);
-    vp8_loop_filter_horizontal_edge_armv6(y_ptr + 12 * y_stride, y_stride, lfi->blim, lfi->lim, lfi->hev_thr, 2);
+void vp8_loop_filter_bh_armv6(unsigned char *y_ptr, unsigned char *u_ptr,
+                              unsigned char *v_ptr, int y_stride, int uv_stride,
+                              loop_filter_info *lfi) {
+  vp8_loop_filter_horizontal_edge_armv6(y_ptr + 4 * y_stride, y_stride,
+                                        lfi->blim, lfi->lim, lfi->hev_thr, 2);
+  vp8_loop_filter_horizontal_edge_armv6(y_ptr + 8 * y_stride, y_stride,
+                                        lfi->blim, lfi->lim, lfi->hev_thr, 2);
+  vp8_loop_filter_horizontal_edge_armv6(y_ptr + 12 * y_stride, y_stride,
+                                        lfi->blim, lfi->lim, lfi->hev_thr, 2);
 
-    if (u_ptr)
-        vp8_loop_filter_horizontal_edge_armv6(u_ptr + 4 * uv_stride, uv_stride, lfi->blim, lfi->lim, lfi->hev_thr, 1);
+  if (u_ptr)
+    vp8_loop_filter_horizontal_edge_armv6(u_ptr + 4 * uv_stride, uv_stride,
+                                          lfi->blim, lfi->lim, lfi->hev_thr, 1);
 
-    if (v_ptr)
-        vp8_loop_filter_horizontal_edge_armv6(v_ptr + 4 * uv_stride, uv_stride, lfi->blim, lfi->lim, lfi->hev_thr, 1);
+  if (v_ptr)
+    vp8_loop_filter_horizontal_edge_armv6(v_ptr + 4 * uv_stride, uv_stride,
+                                          lfi->blim, lfi->lim, lfi->hev_thr, 1);
 }
 
 void vp8_loop_filter_bhs_armv6(unsigned char *y_ptr, int y_stride,
-                               const unsigned char *blimit)
-{
-    vp8_loop_filter_simple_horizontal_edge_armv6(y_ptr + 4 * y_stride, y_stride, blimit);
-    vp8_loop_filter_simple_horizontal_edge_armv6(y_ptr + 8 * y_stride, y_stride, blimit);
-    vp8_loop_filter_simple_horizontal_edge_armv6(y_ptr + 12 * y_stride, y_stride, blimit);
+                               const unsigned char *blimit) {
+  vp8_loop_filter_simple_horizontal_edge_armv6(y_ptr + 4 * y_stride, y_stride,
+                                               blimit);
+  vp8_loop_filter_simple_horizontal_edge_armv6(y_ptr + 8 * y_stride, y_stride,
+                                               blimit);
+  vp8_loop_filter_simple_horizontal_edge_armv6(y_ptr + 12 * y_stride, y_stride,
+                                               blimit);
 }
 
 /* Vertical B Filtering */
-void vp8_loop_filter_bv_armv6(unsigned char *y_ptr, unsigned char *u_ptr, unsigned char *v_ptr,
-                              int y_stride, int uv_stride, loop_filter_info *lfi)
-{
-    vp8_loop_filter_vertical_edge_armv6(y_ptr + 4, y_stride, lfi->blim, lfi->lim, lfi->hev_thr, 2);
-    vp8_loop_filter_vertical_edge_armv6(y_ptr + 8, y_stride, lfi->blim, lfi->lim, lfi->hev_thr, 2);
-    vp8_loop_filter_vertical_edge_armv6(y_ptr + 12, y_stride, lfi->blim, lfi->lim, lfi->hev_thr, 2);
+void vp8_loop_filter_bv_armv6(unsigned char *y_ptr, unsigned char *u_ptr,
+                              unsigned char *v_ptr, int y_stride, int uv_stride,
+                              loop_filter_info *lfi) {
+  vp8_loop_filter_vertical_edge_armv6(y_ptr + 4, y_stride, lfi->blim, lfi->lim,
+                                      lfi->hev_thr, 2);
+  vp8_loop_filter_vertical_edge_armv6(y_ptr + 8, y_stride, lfi->blim, lfi->lim,
+                                      lfi->hev_thr, 2);
+  vp8_loop_filter_vertical_edge_armv6(y_ptr + 12, y_stride, lfi->blim, lfi->lim,
+                                      lfi->hev_thr, 2);
 
-    if (u_ptr)
-        vp8_loop_filter_vertical_edge_armv6(u_ptr + 4, uv_stride, lfi->blim, lfi->lim, lfi->hev_thr, 1);
+  if (u_ptr)
+    vp8_loop_filter_vertical_edge_armv6(u_ptr + 4, uv_stride, lfi->blim,
+                                        lfi->lim, lfi->hev_thr, 1);
 
-    if (v_ptr)
-        vp8_loop_filter_vertical_edge_armv6(v_ptr + 4, uv_stride, lfi->blim, lfi->lim, lfi->hev_thr, 1);
+  if (v_ptr)
+    vp8_loop_filter_vertical_edge_armv6(v_ptr + 4, uv_stride, lfi->blim,
+                                        lfi->lim, lfi->hev_thr, 1);
 }
 
 void vp8_loop_filter_bvs_armv6(unsigned char *y_ptr, int y_stride,
-                               const unsigned char *blimit)
-{
-    vp8_loop_filter_simple_vertical_edge_armv6(y_ptr + 4, y_stride, blimit);
-    vp8_loop_filter_simple_vertical_edge_armv6(y_ptr + 8, y_stride, blimit);
-    vp8_loop_filter_simple_vertical_edge_armv6(y_ptr + 12, y_stride, blimit);
+                               const unsigned char *blimit) {
+  vp8_loop_filter_simple_vertical_edge_armv6(y_ptr + 4, y_stride, blimit);
+  vp8_loop_filter_simple_vertical_edge_armv6(y_ptr + 8, y_stride, blimit);
+  vp8_loop_filter_simple_vertical_edge_armv6(y_ptr + 12, y_stride, blimit);
 }
 #endif
 
@@ -121,61 +138,71 @@
 #if HAVE_NEON
 /* NEON loopfilter functions */
 /* Horizontal MB filtering */
-void vp8_loop_filter_mbh_neon(unsigned char *y_ptr, unsigned char *u_ptr, unsigned char *v_ptr,
-                              int y_stride, int uv_stride, loop_filter_info *lfi)
-{
-    unsigned char mblim = *lfi->mblim;
-    unsigned char lim = *lfi->lim;
-    unsigned char hev_thr = *lfi->hev_thr;
-    vp8_mbloop_filter_horizontal_edge_y_neon(y_ptr, y_stride, mblim, lim, hev_thr);
+void vp8_loop_filter_mbh_neon(unsigned char *y_ptr, unsigned char *u_ptr,
+                              unsigned char *v_ptr, int y_stride, int uv_stride,
+                              loop_filter_info *lfi) {
+  unsigned char mblim = *lfi->mblim;
+  unsigned char lim = *lfi->lim;
+  unsigned char hev_thr = *lfi->hev_thr;
+  vp8_mbloop_filter_horizontal_edge_y_neon(y_ptr, y_stride, mblim, lim,
+                                           hev_thr);
 
-    if (u_ptr)
-        vp8_mbloop_filter_horizontal_edge_uv_neon(u_ptr, uv_stride, mblim, lim, hev_thr, v_ptr);
+  if (u_ptr)
+    vp8_mbloop_filter_horizontal_edge_uv_neon(u_ptr, uv_stride, mblim, lim,
+                                              hev_thr, v_ptr);
 }
 
 /* Vertical MB Filtering */
-void vp8_loop_filter_mbv_neon(unsigned char *y_ptr, unsigned char *u_ptr, unsigned char *v_ptr,
-                              int y_stride, int uv_stride, loop_filter_info *lfi)
-{
-    unsigned char mblim = *lfi->mblim;
-    unsigned char lim = *lfi->lim;
-    unsigned char hev_thr = *lfi->hev_thr;
+void vp8_loop_filter_mbv_neon(unsigned char *y_ptr, unsigned char *u_ptr,
+                              unsigned char *v_ptr, int y_stride, int uv_stride,
+                              loop_filter_info *lfi) {
+  unsigned char mblim = *lfi->mblim;
+  unsigned char lim = *lfi->lim;
+  unsigned char hev_thr = *lfi->hev_thr;
 
-    vp8_mbloop_filter_vertical_edge_y_neon(y_ptr, y_stride, mblim, lim, hev_thr);
+  vp8_mbloop_filter_vertical_edge_y_neon(y_ptr, y_stride, mblim, lim, hev_thr);
 
-    if (u_ptr)
-        vp8_mbloop_filter_vertical_edge_uv_neon(u_ptr, uv_stride, mblim, lim, hev_thr, v_ptr);
+  if (u_ptr)
+    vp8_mbloop_filter_vertical_edge_uv_neon(u_ptr, uv_stride, mblim, lim,
+                                            hev_thr, v_ptr);
 }
 
 /* Horizontal B Filtering */
-void vp8_loop_filter_bh_neon(unsigned char *y_ptr, unsigned char *u_ptr, unsigned char *v_ptr,
-                             int y_stride, int uv_stride, loop_filter_info *lfi)
-{
-    unsigned char blim = *lfi->blim;
-    unsigned char lim = *lfi->lim;
-    unsigned char hev_thr = *lfi->hev_thr;
+void vp8_loop_filter_bh_neon(unsigned char *y_ptr, unsigned char *u_ptr,
+                             unsigned char *v_ptr, int y_stride, int uv_stride,
+                             loop_filter_info *lfi) {
+  unsigned char blim = *lfi->blim;
+  unsigned char lim = *lfi->lim;
+  unsigned char hev_thr = *lfi->hev_thr;
 
-    vp8_loop_filter_horizontal_edge_y_neon(y_ptr + 4 * y_stride, y_stride, blim, lim, hev_thr);
-    vp8_loop_filter_horizontal_edge_y_neon(y_ptr + 8 * y_stride, y_stride, blim, lim, hev_thr);
-    vp8_loop_filter_horizontal_edge_y_neon(y_ptr + 12 * y_stride, y_stride, blim, lim, hev_thr);
+  vp8_loop_filter_horizontal_edge_y_neon(y_ptr + 4 * y_stride, y_stride, blim,
+                                         lim, hev_thr);
+  vp8_loop_filter_horizontal_edge_y_neon(y_ptr + 8 * y_stride, y_stride, blim,
+                                         lim, hev_thr);
+  vp8_loop_filter_horizontal_edge_y_neon(y_ptr + 12 * y_stride, y_stride, blim,
+                                         lim, hev_thr);
 
-    if (u_ptr)
-        vp8_loop_filter_horizontal_edge_uv_neon(u_ptr + 4 * uv_stride, uv_stride, blim, lim, hev_thr, v_ptr + 4 * uv_stride);
+  if (u_ptr)
+    vp8_loop_filter_horizontal_edge_uv_neon(u_ptr + 4 * uv_stride, uv_stride,
+                                            blim, lim, hev_thr,
+                                            v_ptr + 4 * uv_stride);
 }
 
 /* Vertical B Filtering */
-void vp8_loop_filter_bv_neon(unsigned char *y_ptr, unsigned char *u_ptr, unsigned char *v_ptr,
-                             int y_stride, int uv_stride, loop_filter_info *lfi)
-{
-    unsigned char blim = *lfi->blim;
-    unsigned char lim = *lfi->lim;
-    unsigned char hev_thr = *lfi->hev_thr;
+void vp8_loop_filter_bv_neon(unsigned char *y_ptr, unsigned char *u_ptr,
+                             unsigned char *v_ptr, int y_stride, int uv_stride,
+                             loop_filter_info *lfi) {
+  unsigned char blim = *lfi->blim;
+  unsigned char lim = *lfi->lim;
+  unsigned char hev_thr = *lfi->hev_thr;
 
-    vp8_loop_filter_vertical_edge_y_neon(y_ptr + 4, y_stride, blim, lim, hev_thr);
-    vp8_loop_filter_vertical_edge_y_neon(y_ptr + 8, y_stride, blim, lim, hev_thr);
-    vp8_loop_filter_vertical_edge_y_neon(y_ptr + 12, y_stride, blim, lim, hev_thr);
+  vp8_loop_filter_vertical_edge_y_neon(y_ptr + 4, y_stride, blim, lim, hev_thr);
+  vp8_loop_filter_vertical_edge_y_neon(y_ptr + 8, y_stride, blim, lim, hev_thr);
+  vp8_loop_filter_vertical_edge_y_neon(y_ptr + 12, y_stride, blim, lim,
+                                       hev_thr);
 
-    if (u_ptr)
-        vp8_loop_filter_vertical_edge_uv_neon(u_ptr + 4, uv_stride, blim, lim, hev_thr, v_ptr + 4);
+  if (u_ptr)
+    vp8_loop_filter_vertical_edge_uv_neon(u_ptr + 4, uv_stride, blim, lim,
+                                          hev_thr, v_ptr + 4);
 }
 #endif
--- a/vp8/common/arm/neon/bilinearpredict_neon.c
+++ b/vp8/common/arm/neon/bilinearpredict_neon.c
@@ -10,582 +10,652 @@
 
 #include <arm_neon.h>
 
-static const uint8_t bifilter4_coeff[8][2] = {
-    {128,   0},
-    {112,  16},
-    { 96,  32},
-    { 80,  48},
-    { 64,  64},
-    { 48,  80},
-    { 32,  96},
-    { 16, 112}
-};
+static const uint8_t bifilter4_coeff[8][2] = { { 128, 0 }, { 112, 16 },
+                                               { 96, 32 }, { 80, 48 },
+                                               { 64, 64 }, { 48, 80 },
+                                               { 32, 96 }, { 16, 112 } };
 
-void vp8_bilinear_predict8x4_neon(
-        unsigned char *src_ptr,
-        int src_pixels_per_line,
-        int xoffset,
-        int yoffset,
-        unsigned char *dst_ptr,
-        int dst_pitch) {
-    uint8x8_t d0u8, d1u8, d2u8, d3u8, d4u8, d5u8;
-    uint8x8_t d7u8, d9u8, d11u8, d22u8, d23u8, d24u8, d25u8, d26u8;
-    uint8x16_t q1u8, q2u8, q3u8, q4u8, q5u8;
-    uint16x8_t q1u16, q2u16, q3u16, q4u16;
-    uint16x8_t q6u16, q7u16, q8u16, q9u16, q10u16;
+void vp8_bilinear_predict8x4_neon(unsigned char *src_ptr,
+                                  int src_pixels_per_line, int xoffset,
+                                  int yoffset, unsigned char *dst_ptr,
+                                  int dst_pitch) {
+  uint8x8_t d0u8, d1u8, d2u8, d3u8, d4u8, d5u8;
+  uint8x8_t d7u8, d9u8, d11u8, d22u8, d23u8, d24u8, d25u8, d26u8;
+  uint8x16_t q1u8, q2u8, q3u8, q4u8, q5u8;
+  uint16x8_t q1u16, q2u16, q3u16, q4u16;
+  uint16x8_t q6u16, q7u16, q8u16, q9u16, q10u16;
 
-    if (xoffset == 0) {  // skip_1stpass_filter
-        d22u8 = vld1_u8(src_ptr); src_ptr += src_pixels_per_line;
-        d23u8 = vld1_u8(src_ptr); src_ptr += src_pixels_per_line;
-        d24u8 = vld1_u8(src_ptr); src_ptr += src_pixels_per_line;
-        d25u8 = vld1_u8(src_ptr); src_ptr += src_pixels_per_line;
-        d26u8 = vld1_u8(src_ptr);
-    } else {
-        q1u8 = vld1q_u8(src_ptr); src_ptr += src_pixels_per_line;
-        q2u8 = vld1q_u8(src_ptr); src_ptr += src_pixels_per_line;
-        q3u8 = vld1q_u8(src_ptr); src_ptr += src_pixels_per_line;
-        q4u8 = vld1q_u8(src_ptr); src_ptr += src_pixels_per_line;
-        q5u8 = vld1q_u8(src_ptr);
+  if (xoffset == 0) {  // skip_1stpass_filter
+    d22u8 = vld1_u8(src_ptr);
+    src_ptr += src_pixels_per_line;
+    d23u8 = vld1_u8(src_ptr);
+    src_ptr += src_pixels_per_line;
+    d24u8 = vld1_u8(src_ptr);
+    src_ptr += src_pixels_per_line;
+    d25u8 = vld1_u8(src_ptr);
+    src_ptr += src_pixels_per_line;
+    d26u8 = vld1_u8(src_ptr);
+  } else {
+    q1u8 = vld1q_u8(src_ptr);
+    src_ptr += src_pixels_per_line;
+    q2u8 = vld1q_u8(src_ptr);
+    src_ptr += src_pixels_per_line;
+    q3u8 = vld1q_u8(src_ptr);
+    src_ptr += src_pixels_per_line;
+    q4u8 = vld1q_u8(src_ptr);
+    src_ptr += src_pixels_per_line;
+    q5u8 = vld1q_u8(src_ptr);
 
-        d0u8 = vdup_n_u8(bifilter4_coeff[xoffset][0]);
-        d1u8 = vdup_n_u8(bifilter4_coeff[xoffset][1]);
+    d0u8 = vdup_n_u8(bifilter4_coeff[xoffset][0]);
+    d1u8 = vdup_n_u8(bifilter4_coeff[xoffset][1]);
 
-        q6u16 = vmull_u8(vget_low_u8(q1u8), d0u8);
-        q7u16 = vmull_u8(vget_low_u8(q2u8), d0u8);
-        q8u16 = vmull_u8(vget_low_u8(q3u8), d0u8);
-        q9u16 = vmull_u8(vget_low_u8(q4u8), d0u8);
-        q10u16 = vmull_u8(vget_low_u8(q5u8), d0u8);
+    q6u16 = vmull_u8(vget_low_u8(q1u8), d0u8);
+    q7u16 = vmull_u8(vget_low_u8(q2u8), d0u8);
+    q8u16 = vmull_u8(vget_low_u8(q3u8), d0u8);
+    q9u16 = vmull_u8(vget_low_u8(q4u8), d0u8);
+    q10u16 = vmull_u8(vget_low_u8(q5u8), d0u8);
 
-        d3u8 = vext_u8(vget_low_u8(q1u8), vget_high_u8(q1u8), 1);
-        d5u8 = vext_u8(vget_low_u8(q2u8), vget_high_u8(q2u8), 1);
-        d7u8 = vext_u8(vget_low_u8(q3u8), vget_high_u8(q3u8), 1);
-        d9u8 = vext_u8(vget_low_u8(q4u8), vget_high_u8(q4u8), 1);
-        d11u8 = vext_u8(vget_low_u8(q5u8), vget_high_u8(q5u8), 1);
+    d3u8 = vext_u8(vget_low_u8(q1u8), vget_high_u8(q1u8), 1);
+    d5u8 = vext_u8(vget_low_u8(q2u8), vget_high_u8(q2u8), 1);
+    d7u8 = vext_u8(vget_low_u8(q3u8), vget_high_u8(q3u8), 1);
+    d9u8 = vext_u8(vget_low_u8(q4u8), vget_high_u8(q4u8), 1);
+    d11u8 = vext_u8(vget_low_u8(q5u8), vget_high_u8(q5u8), 1);
 
-        q6u16 = vmlal_u8(q6u16, d3u8, d1u8);
-        q7u16 = vmlal_u8(q7u16, d5u8, d1u8);
-        q8u16 = vmlal_u8(q8u16, d7u8, d1u8);
-        q9u16 = vmlal_u8(q9u16, d9u8, d1u8);
-        q10u16 = vmlal_u8(q10u16, d11u8, d1u8);
+    q6u16 = vmlal_u8(q6u16, d3u8, d1u8);
+    q7u16 = vmlal_u8(q7u16, d5u8, d1u8);
+    q8u16 = vmlal_u8(q8u16, d7u8, d1u8);
+    q9u16 = vmlal_u8(q9u16, d9u8, d1u8);
+    q10u16 = vmlal_u8(q10u16, d11u8, d1u8);
 
-        d22u8 = vqrshrn_n_u16(q6u16, 7);
-        d23u8 = vqrshrn_n_u16(q7u16, 7);
-        d24u8 = vqrshrn_n_u16(q8u16, 7);
-        d25u8 = vqrshrn_n_u16(q9u16, 7);
-        d26u8 = vqrshrn_n_u16(q10u16, 7);
-    }
+    d22u8 = vqrshrn_n_u16(q6u16, 7);
+    d23u8 = vqrshrn_n_u16(q7u16, 7);
+    d24u8 = vqrshrn_n_u16(q8u16, 7);
+    d25u8 = vqrshrn_n_u16(q9u16, 7);
+    d26u8 = vqrshrn_n_u16(q10u16, 7);
+  }
 
-    // secondpass_filter
-    if (yoffset == 0) {  // skip_2ndpass_filter
-        vst1_u8((uint8_t *)dst_ptr, d22u8); dst_ptr += dst_pitch;
-        vst1_u8((uint8_t *)dst_ptr, d23u8); dst_ptr += dst_pitch;
-        vst1_u8((uint8_t *)dst_ptr, d24u8); dst_ptr += dst_pitch;
-        vst1_u8((uint8_t *)dst_ptr, d25u8);
-    } else {
-        d0u8 = vdup_n_u8(bifilter4_coeff[yoffset][0]);
-        d1u8 = vdup_n_u8(bifilter4_coeff[yoffset][1]);
+  // secondpass_filter
+  if (yoffset == 0) {  // skip_2ndpass_filter
+    vst1_u8((uint8_t *)dst_ptr, d22u8);
+    dst_ptr += dst_pitch;
+    vst1_u8((uint8_t *)dst_ptr, d23u8);
+    dst_ptr += dst_pitch;
+    vst1_u8((uint8_t *)dst_ptr, d24u8);
+    dst_ptr += dst_pitch;
+    vst1_u8((uint8_t *)dst_ptr, d25u8);
+  } else {
+    d0u8 = vdup_n_u8(bifilter4_coeff[yoffset][0]);
+    d1u8 = vdup_n_u8(bifilter4_coeff[yoffset][1]);
 
-        q1u16 = vmull_u8(d22u8, d0u8);
-        q2u16 = vmull_u8(d23u8, d0u8);
-        q3u16 = vmull_u8(d24u8, d0u8);
-        q4u16 = vmull_u8(d25u8, d0u8);
+    q1u16 = vmull_u8(d22u8, d0u8);
+    q2u16 = vmull_u8(d23u8, d0u8);
+    q3u16 = vmull_u8(d24u8, d0u8);
+    q4u16 = vmull_u8(d25u8, d0u8);
 
-        q1u16 = vmlal_u8(q1u16, d23u8, d1u8);
-        q2u16 = vmlal_u8(q2u16, d24u8, d1u8);
-        q3u16 = vmlal_u8(q3u16, d25u8, d1u8);
-        q4u16 = vmlal_u8(q4u16, d26u8, d1u8);
+    q1u16 = vmlal_u8(q1u16, d23u8, d1u8);
+    q2u16 = vmlal_u8(q2u16, d24u8, d1u8);
+    q3u16 = vmlal_u8(q3u16, d25u8, d1u8);
+    q4u16 = vmlal_u8(q4u16, d26u8, d1u8);
 
-        d2u8 = vqrshrn_n_u16(q1u16, 7);
-        d3u8 = vqrshrn_n_u16(q2u16, 7);
-        d4u8 = vqrshrn_n_u16(q3u16, 7);
-        d5u8 = vqrshrn_n_u16(q4u16, 7);
+    d2u8 = vqrshrn_n_u16(q1u16, 7);
+    d3u8 = vqrshrn_n_u16(q2u16, 7);
+    d4u8 = vqrshrn_n_u16(q3u16, 7);
+    d5u8 = vqrshrn_n_u16(q4u16, 7);
 
-        vst1_u8((uint8_t *)dst_ptr, d2u8); dst_ptr += dst_pitch;
-        vst1_u8((uint8_t *)dst_ptr, d3u8); dst_ptr += dst_pitch;
-        vst1_u8((uint8_t *)dst_ptr, d4u8); dst_ptr += dst_pitch;
-        vst1_u8((uint8_t *)dst_ptr, d5u8);
-    }
-    return;
+    vst1_u8((uint8_t *)dst_ptr, d2u8);
+    dst_ptr += dst_pitch;
+    vst1_u8((uint8_t *)dst_ptr, d3u8);
+    dst_ptr += dst_pitch;
+    vst1_u8((uint8_t *)dst_ptr, d4u8);
+    dst_ptr += dst_pitch;
+    vst1_u8((uint8_t *)dst_ptr, d5u8);
+  }
+  return;
 }
 
-void vp8_bilinear_predict8x8_neon(
-        unsigned char *src_ptr,
-        int src_pixels_per_line,
-        int xoffset,
-        int yoffset,
-        unsigned char *dst_ptr,
-        int dst_pitch) {
-    uint8x8_t d0u8, d1u8, d2u8, d3u8, d4u8, d5u8, d6u8, d7u8, d8u8, d9u8, d11u8;
-    uint8x8_t d22u8, d23u8, d24u8, d25u8, d26u8, d27u8, d28u8, d29u8, d30u8;
-    uint8x16_t q1u8, q2u8, q3u8, q4u8, q5u8;
-    uint16x8_t q1u16, q2u16, q3u16, q4u16, q5u16;
-    uint16x8_t q6u16, q7u16, q8u16, q9u16, q10u16;
+void vp8_bilinear_predict8x8_neon(unsigned char *src_ptr,
+                                  int src_pixels_per_line, int xoffset,
+                                  int yoffset, unsigned char *dst_ptr,
+                                  int dst_pitch) {
+  uint8x8_t d0u8, d1u8, d2u8, d3u8, d4u8, d5u8, d6u8, d7u8, d8u8, d9u8, d11u8;
+  uint8x8_t d22u8, d23u8, d24u8, d25u8, d26u8, d27u8, d28u8, d29u8, d30u8;
+  uint8x16_t q1u8, q2u8, q3u8, q4u8, q5u8;
+  uint16x8_t q1u16, q2u16, q3u16, q4u16, q5u16;
+  uint16x8_t q6u16, q7u16, q8u16, q9u16, q10u16;
 
-    if (xoffset == 0) {  // skip_1stpass_filter
-        d22u8 = vld1_u8(src_ptr); src_ptr += src_pixels_per_line;
-        d23u8 = vld1_u8(src_ptr); src_ptr += src_pixels_per_line;
-        d24u8 = vld1_u8(src_ptr); src_ptr += src_pixels_per_line;
-        d25u8 = vld1_u8(src_ptr); src_ptr += src_pixels_per_line;
-        d26u8 = vld1_u8(src_ptr); src_ptr += src_pixels_per_line;
-        d27u8 = vld1_u8(src_ptr); src_ptr += src_pixels_per_line;
-        d28u8 = vld1_u8(src_ptr); src_ptr += src_pixels_per_line;
-        d29u8 = vld1_u8(src_ptr); src_ptr += src_pixels_per_line;
-        d30u8 = vld1_u8(src_ptr);
-    } else {
-        q1u8 = vld1q_u8(src_ptr); src_ptr += src_pixels_per_line;
-        q2u8 = vld1q_u8(src_ptr); src_ptr += src_pixels_per_line;
-        q3u8 = vld1q_u8(src_ptr); src_ptr += src_pixels_per_line;
-        q4u8 = vld1q_u8(src_ptr); src_ptr += src_pixels_per_line;
+  if (xoffset == 0) {  // skip_1stpass_filter
+    d22u8 = vld1_u8(src_ptr);
+    src_ptr += src_pixels_per_line;
+    d23u8 = vld1_u8(src_ptr);
+    src_ptr += src_pixels_per_line;
+    d24u8 = vld1_u8(src_ptr);
+    src_ptr += src_pixels_per_line;
+    d25u8 = vld1_u8(src_ptr);
+    src_ptr += src_pixels_per_line;
+    d26u8 = vld1_u8(src_ptr);
+    src_ptr += src_pixels_per_line;
+    d27u8 = vld1_u8(src_ptr);
+    src_ptr += src_pixels_per_line;
+    d28u8 = vld1_u8(src_ptr);
+    src_ptr += src_pixels_per_line;
+    d29u8 = vld1_u8(src_ptr);
+    src_ptr += src_pixels_per_line;
+    d30u8 = vld1_u8(src_ptr);
+  } else {
+    q1u8 = vld1q_u8(src_ptr);
+    src_ptr += src_pixels_per_line;
+    q2u8 = vld1q_u8(src_ptr);
+    src_ptr += src_pixels_per_line;
+    q3u8 = vld1q_u8(src_ptr);
+    src_ptr += src_pixels_per_line;
+    q4u8 = vld1q_u8(src_ptr);
+    src_ptr += src_pixels_per_line;
 
-        d0u8 = vdup_n_u8(bifilter4_coeff[xoffset][0]);
-        d1u8 = vdup_n_u8(bifilter4_coeff[xoffset][1]);
+    d0u8 = vdup_n_u8(bifilter4_coeff[xoffset][0]);
+    d1u8 = vdup_n_u8(bifilter4_coeff[xoffset][1]);
 
-        q6u16 = vmull_u8(vget_low_u8(q1u8), d0u8);
-        q7u16 = vmull_u8(vget_low_u8(q2u8), d0u8);
-        q8u16 = vmull_u8(vget_low_u8(q3u8), d0u8);
-        q9u16 = vmull_u8(vget_low_u8(q4u8), d0u8);
+    q6u16 = vmull_u8(vget_low_u8(q1u8), d0u8);
+    q7u16 = vmull_u8(vget_low_u8(q2u8), d0u8);
+    q8u16 = vmull_u8(vget_low_u8(q3u8), d0u8);
+    q9u16 = vmull_u8(vget_low_u8(q4u8), d0u8);
 
-        d3u8 = vext_u8(vget_low_u8(q1u8), vget_high_u8(q1u8), 1);
-        d5u8 = vext_u8(vget_low_u8(q2u8), vget_high_u8(q2u8), 1);
-        d7u8 = vext_u8(vget_low_u8(q3u8), vget_high_u8(q3u8), 1);
-        d9u8 = vext_u8(vget_low_u8(q4u8), vget_high_u8(q4u8), 1);
+    d3u8 = vext_u8(vget_low_u8(q1u8), vget_high_u8(q1u8), 1);
+    d5u8 = vext_u8(vget_low_u8(q2u8), vget_high_u8(q2u8), 1);
+    d7u8 = vext_u8(vget_low_u8(q3u8), vget_high_u8(q3u8), 1);
+    d9u8 = vext_u8(vget_low_u8(q4u8), vget_high_u8(q4u8), 1);
 
-        q6u16 = vmlal_u8(q6u16, d3u8, d1u8);
-        q7u16 = vmlal_u8(q7u16, d5u8, d1u8);
-        q8u16 = vmlal_u8(q8u16, d7u8, d1u8);
-        q9u16 = vmlal_u8(q9u16, d9u8, d1u8);
+    q6u16 = vmlal_u8(q6u16, d3u8, d1u8);
+    q7u16 = vmlal_u8(q7u16, d5u8, d1u8);
+    q8u16 = vmlal_u8(q8u16, d7u8, d1u8);
+    q9u16 = vmlal_u8(q9u16, d9u8, d1u8);
 
-        d22u8 = vqrshrn_n_u16(q6u16, 7);
-        d23u8 = vqrshrn_n_u16(q7u16, 7);
-        d24u8 = vqrshrn_n_u16(q8u16, 7);
-        d25u8 = vqrshrn_n_u16(q9u16, 7);
+    d22u8 = vqrshrn_n_u16(q6u16, 7);
+    d23u8 = vqrshrn_n_u16(q7u16, 7);
+    d24u8 = vqrshrn_n_u16(q8u16, 7);
+    d25u8 = vqrshrn_n_u16(q9u16, 7);
 
-        // first_pass filtering on the rest 5-line data
-        q1u8 = vld1q_u8(src_ptr); src_ptr += src_pixels_per_line;
-        q2u8 = vld1q_u8(src_ptr); src_ptr += src_pixels_per_line;
-        q3u8 = vld1q_u8(src_ptr); src_ptr += src_pixels_per_line;
-        q4u8 = vld1q_u8(src_ptr); src_ptr += src_pixels_per_line;
-        q5u8 = vld1q_u8(src_ptr);
+    // first_pass filtering on the rest 5-line data
+    q1u8 = vld1q_u8(src_ptr);
+    src_ptr += src_pixels_per_line;
+    q2u8 = vld1q_u8(src_ptr);
+    src_ptr += src_pixels_per_line;
+    q3u8 = vld1q_u8(src_ptr);
+    src_ptr += src_pixels_per_line;
+    q4u8 = vld1q_u8(src_ptr);
+    src_ptr += src_pixels_per_line;
+    q5u8 = vld1q_u8(src_ptr);
 
-        q6u16 = vmull_u8(vget_low_u8(q1u8), d0u8);
-        q7u16 = vmull_u8(vget_low_u8(q2u8), d0u8);
-        q8u16 = vmull_u8(vget_low_u8(q3u8), d0u8);
-        q9u16 = vmull_u8(vget_low_u8(q4u8), d0u8);
-        q10u16 = vmull_u8(vget_low_u8(q5u8), d0u8);
+    q6u16 = vmull_u8(vget_low_u8(q1u8), d0u8);
+    q7u16 = vmull_u8(vget_low_u8(q2u8), d0u8);
+    q8u16 = vmull_u8(vget_low_u8(q3u8), d0u8);
+    q9u16 = vmull_u8(vget_low_u8(q4u8), d0u8);
+    q10u16 = vmull_u8(vget_low_u8(q5u8), d0u8);
 
-        d3u8 = vext_u8(vget_low_u8(q1u8), vget_high_u8(q1u8), 1);
-        d5u8 = vext_u8(vget_low_u8(q2u8), vget_high_u8(q2u8), 1);
-        d7u8 = vext_u8(vget_low_u8(q3u8), vget_high_u8(q3u8), 1);
-        d9u8 = vext_u8(vget_low_u8(q4u8), vget_high_u8(q4u8), 1);
-        d11u8 = vext_u8(vget_low_u8(q5u8), vget_high_u8(q5u8), 1);
+    d3u8 = vext_u8(vget_low_u8(q1u8), vget_high_u8(q1u8), 1);
+    d5u8 = vext_u8(vget_low_u8(q2u8), vget_high_u8(q2u8), 1);
+    d7u8 = vext_u8(vget_low_u8(q3u8), vget_high_u8(q3u8), 1);
+    d9u8 = vext_u8(vget_low_u8(q4u8), vget_high_u8(q4u8), 1);
+    d11u8 = vext_u8(vget_low_u8(q5u8), vget_high_u8(q5u8), 1);
 
-        q6u16 = vmlal_u8(q6u16, d3u8, d1u8);
-        q7u16 = vmlal_u8(q7u16, d5u8, d1u8);
-        q8u16 = vmlal_u8(q8u16, d7u8, d1u8);
-        q9u16 = vmlal_u8(q9u16, d9u8, d1u8);
-        q10u16 = vmlal_u8(q10u16, d11u8, d1u8);
+    q6u16 = vmlal_u8(q6u16, d3u8, d1u8);
+    q7u16 = vmlal_u8(q7u16, d5u8, d1u8);
+    q8u16 = vmlal_u8(q8u16, d7u8, d1u8);
+    q9u16 = vmlal_u8(q9u16, d9u8, d1u8);
+    q10u16 = vmlal_u8(q10u16, d11u8, d1u8);
 
-        d26u8 = vqrshrn_n_u16(q6u16, 7);
-        d27u8 = vqrshrn_n_u16(q7u16, 7);
-        d28u8 = vqrshrn_n_u16(q8u16, 7);
-        d29u8 = vqrshrn_n_u16(q9u16, 7);
-        d30u8 = vqrshrn_n_u16(q10u16, 7);
-    }
+    d26u8 = vqrshrn_n_u16(q6u16, 7);
+    d27u8 = vqrshrn_n_u16(q7u16, 7);
+    d28u8 = vqrshrn_n_u16(q8u16, 7);
+    d29u8 = vqrshrn_n_u16(q9u16, 7);
+    d30u8 = vqrshrn_n_u16(q10u16, 7);
+  }
 
-    // secondpass_filter
-    if (yoffset == 0) {  // skip_2ndpass_filter
-        vst1_u8((uint8_t *)dst_ptr, d22u8); dst_ptr += dst_pitch;
-        vst1_u8((uint8_t *)dst_ptr, d23u8); dst_ptr += dst_pitch;
-        vst1_u8((uint8_t *)dst_ptr, d24u8); dst_ptr += dst_pitch;
-        vst1_u8((uint8_t *)dst_ptr, d25u8); dst_ptr += dst_pitch;
-        vst1_u8((uint8_t *)dst_ptr, d26u8); dst_ptr += dst_pitch;
-        vst1_u8((uint8_t *)dst_ptr, d27u8); dst_ptr += dst_pitch;
-        vst1_u8((uint8_t *)dst_ptr, d28u8); dst_ptr += dst_pitch;
-        vst1_u8((uint8_t *)dst_ptr, d29u8);
-    } else {
-        d0u8 = vdup_n_u8(bifilter4_coeff[yoffset][0]);
-        d1u8 = vdup_n_u8(bifilter4_coeff[yoffset][1]);
+  // secondpass_filter
+  if (yoffset == 0) {  // skip_2ndpass_filter
+    vst1_u8((uint8_t *)dst_ptr, d22u8);
+    dst_ptr += dst_pitch;
+    vst1_u8((uint8_t *)dst_ptr, d23u8);
+    dst_ptr += dst_pitch;
+    vst1_u8((uint8_t *)dst_ptr, d24u8);
+    dst_ptr += dst_pitch;
+    vst1_u8((uint8_t *)dst_ptr, d25u8);
+    dst_ptr += dst_pitch;
+    vst1_u8((uint8_t *)dst_ptr, d26u8);
+    dst_ptr += dst_pitch;
+    vst1_u8((uint8_t *)dst_ptr, d27u8);
+    dst_ptr += dst_pitch;
+    vst1_u8((uint8_t *)dst_ptr, d28u8);
+    dst_ptr += dst_pitch;
+    vst1_u8((uint8_t *)dst_ptr, d29u8);
+  } else {
+    d0u8 = vdup_n_u8(bifilter4_coeff[yoffset][0]);
+    d1u8 = vdup_n_u8(bifilter4_coeff[yoffset][1]);
 
-        q1u16 = vmull_u8(d22u8, d0u8);
-        q2u16 = vmull_u8(d23u8, d0u8);
-        q3u16 = vmull_u8(d24u8, d0u8);
-        q4u16 = vmull_u8(d25u8, d0u8);
-        q5u16 = vmull_u8(d26u8, d0u8);
-        q6u16 = vmull_u8(d27u8, d0u8);
-        q7u16 = vmull_u8(d28u8, d0u8);
-        q8u16 = vmull_u8(d29u8, d0u8);
+    q1u16 = vmull_u8(d22u8, d0u8);
+    q2u16 = vmull_u8(d23u8, d0u8);
+    q3u16 = vmull_u8(d24u8, d0u8);
+    q4u16 = vmull_u8(d25u8, d0u8);
+    q5u16 = vmull_u8(d26u8, d0u8);
+    q6u16 = vmull_u8(d27u8, d0u8);
+    q7u16 = vmull_u8(d28u8, d0u8);
+    q8u16 = vmull_u8(d29u8, d0u8);
 
-        q1u16 = vmlal_u8(q1u16, d23u8, d1u8);
-        q2u16 = vmlal_u8(q2u16, d24u8, d1u8);
-        q3u16 = vmlal_u8(q3u16, d25u8, d1u8);
-        q4u16 = vmlal_u8(q4u16, d26u8, d1u8);
-        q5u16 = vmlal_u8(q5u16, d27u8, d1u8);
-        q6u16 = vmlal_u8(q6u16, d28u8, d1u8);
-        q7u16 = vmlal_u8(q7u16, d29u8, d1u8);
-        q8u16 = vmlal_u8(q8u16, d30u8, d1u8);
+    q1u16 = vmlal_u8(q1u16, d23u8, d1u8);
+    q2u16 = vmlal_u8(q2u16, d24u8, d1u8);
+    q3u16 = vmlal_u8(q3u16, d25u8, d1u8);
+    q4u16 = vmlal_u8(q4u16, d26u8, d1u8);
+    q5u16 = vmlal_u8(q5u16, d27u8, d1u8);
+    q6u16 = vmlal_u8(q6u16, d28u8, d1u8);
+    q7u16 = vmlal_u8(q7u16, d29u8, d1u8);
+    q8u16 = vmlal_u8(q8u16, d30u8, d1u8);
 
-        d2u8 = vqrshrn_n_u16(q1u16, 7);
-        d3u8 = vqrshrn_n_u16(q2u16, 7);
-        d4u8 = vqrshrn_n_u16(q3u16, 7);
-        d5u8 = vqrshrn_n_u16(q4u16, 7);
-        d6u8 = vqrshrn_n_u16(q5u16, 7);
-        d7u8 = vqrshrn_n_u16(q6u16, 7);
-        d8u8 = vqrshrn_n_u16(q7u16, 7);
-        d9u8 = vqrshrn_n_u16(q8u16, 7);
+    d2u8 = vqrshrn_n_u16(q1u16, 7);
+    d3u8 = vqrshrn_n_u16(q2u16, 7);
+    d4u8 = vqrshrn_n_u16(q3u16, 7);
+    d5u8 = vqrshrn_n_u16(q4u16, 7);
+    d6u8 = vqrshrn_n_u16(q5u16, 7);
+    d7u8 = vqrshrn_n_u16(q6u16, 7);
+    d8u8 = vqrshrn_n_u16(q7u16, 7);
+    d9u8 = vqrshrn_n_u16(q8u16, 7);
 
-        vst1_u8((uint8_t *)dst_ptr, d2u8); dst_ptr += dst_pitch;
-        vst1_u8((uint8_t *)dst_ptr, d3u8); dst_ptr += dst_pitch;
-        vst1_u8((uint8_t *)dst_ptr, d4u8); dst_ptr += dst_pitch;
-        vst1_u8((uint8_t *)dst_ptr, d5u8); dst_ptr += dst_pitch;
-        vst1_u8((uint8_t *)dst_ptr, d6u8); dst_ptr += dst_pitch;
-        vst1_u8((uint8_t *)dst_ptr, d7u8); dst_ptr += dst_pitch;
-        vst1_u8((uint8_t *)dst_ptr, d8u8); dst_ptr += dst_pitch;
-        vst1_u8((uint8_t *)dst_ptr, d9u8);
-    }
-    return;
+    vst1_u8((uint8_t *)dst_ptr, d2u8);
+    dst_ptr += dst_pitch;
+    vst1_u8((uint8_t *)dst_ptr, d3u8);
+    dst_ptr += dst_pitch;
+    vst1_u8((uint8_t *)dst_ptr, d4u8);
+    dst_ptr += dst_pitch;
+    vst1_u8((uint8_t *)dst_ptr, d5u8);
+    dst_ptr += dst_pitch;
+    vst1_u8((uint8_t *)dst_ptr, d6u8);
+    dst_ptr += dst_pitch;
+    vst1_u8((uint8_t *)dst_ptr, d7u8);
+    dst_ptr += dst_pitch;
+    vst1_u8((uint8_t *)dst_ptr, d8u8);
+    dst_ptr += dst_pitch;
+    vst1_u8((uint8_t *)dst_ptr, d9u8);
+  }
+  return;
 }
 
-void vp8_bilinear_predict16x16_neon(
-        unsigned char *src_ptr,
-        int src_pixels_per_line,
-        int xoffset,
-        int yoffset,
-        unsigned char *dst_ptr,
-        int dst_pitch) {
-    int i;
-    unsigned char tmp[272];
-    unsigned char *tmpp;
-    uint8x8_t d0u8, d1u8, d2u8, d3u8, d4u8, d5u8, d6u8, d7u8, d8u8, d9u8;
-    uint8x8_t d10u8, d11u8, d12u8, d13u8, d14u8, d15u8, d16u8, d17u8, d18u8;
-    uint8x8_t d19u8, d20u8, d21u8;
-    uint8x16_t q1u8, q2u8, q3u8, q4u8, q5u8, q6u8, q7u8, q8u8, q9u8, q10u8;
-    uint8x16_t q11u8, q12u8, q13u8, q14u8, q15u8;
-    uint16x8_t q1u16, q2u16, q3u16, q4u16, q5u16, q6u16, q7u16, q8u16;
-    uint16x8_t q9u16, q10u16, q11u16, q12u16, q13u16, q14u16;
+void vp8_bilinear_predict16x16_neon(unsigned char *src_ptr,
+                                    int src_pixels_per_line, int xoffset,
+                                    int yoffset, unsigned char *dst_ptr,
+                                    int dst_pitch) {
+  int i;
+  unsigned char tmp[272];
+  unsigned char *tmpp;
+  uint8x8_t d0u8, d1u8, d2u8, d3u8, d4u8, d5u8, d6u8, d7u8, d8u8, d9u8;
+  uint8x8_t d10u8, d11u8, d12u8, d13u8, d14u8, d15u8, d16u8, d17u8, d18u8;
+  uint8x8_t d19u8, d20u8, d21u8;
+  uint8x16_t q1u8, q2u8, q3u8, q4u8, q5u8, q6u8, q7u8, q8u8, q9u8, q10u8;
+  uint8x16_t q11u8, q12u8, q13u8, q14u8, q15u8;
+  uint16x8_t q1u16, q2u16, q3u16, q4u16, q5u16, q6u16, q7u16, q8u16;
+  uint16x8_t q9u16, q10u16, q11u16, q12u16, q13u16, q14u16;
 
-    if (xoffset == 0) {  // secondpass_bfilter16x16_only
-        d0u8 = vdup_n_u8(bifilter4_coeff[yoffset][0]);
-        d1u8 = vdup_n_u8(bifilter4_coeff[yoffset][1]);
+  if (xoffset == 0) {  // secondpass_bfilter16x16_only
+    d0u8 = vdup_n_u8(bifilter4_coeff[yoffset][0]);
+    d1u8 = vdup_n_u8(bifilter4_coeff[yoffset][1]);
 
-        q11u8 = vld1q_u8(src_ptr);
-        src_ptr += src_pixels_per_line;
-        for (i = 4; i > 0; i--) {
-            q12u8 = vld1q_u8(src_ptr); src_ptr += src_pixels_per_line;
-            q13u8 = vld1q_u8(src_ptr); src_ptr += src_pixels_per_line;
-            q14u8 = vld1q_u8(src_ptr); src_ptr += src_pixels_per_line;
-            q15u8 = vld1q_u8(src_ptr); src_ptr += src_pixels_per_line;
+    q11u8 = vld1q_u8(src_ptr);
+    src_ptr += src_pixels_per_line;
+    for (i = 4; i > 0; i--) {
+      q12u8 = vld1q_u8(src_ptr);
+      src_ptr += src_pixels_per_line;
+      q13u8 = vld1q_u8(src_ptr);
+      src_ptr += src_pixels_per_line;
+      q14u8 = vld1q_u8(src_ptr);
+      src_ptr += src_pixels_per_line;
+      q15u8 = vld1q_u8(src_ptr);
+      src_ptr += src_pixels_per_line;
 
-            q1u16 = vmull_u8(vget_low_u8(q11u8), d0u8);
-            q2u16 = vmull_u8(vget_high_u8(q11u8), d0u8);
-            q3u16 = vmull_u8(vget_low_u8(q12u8), d0u8);
-            q4u16 = vmull_u8(vget_high_u8(q12u8), d0u8);
-            q5u16 = vmull_u8(vget_low_u8(q13u8), d0u8);
-            q6u16 = vmull_u8(vget_high_u8(q13u8), d0u8);
-            q7u16 = vmull_u8(vget_low_u8(q14u8), d0u8);
-            q8u16 = vmull_u8(vget_high_u8(q14u8), d0u8);
+      q1u16 = vmull_u8(vget_low_u8(q11u8), d0u8);
+      q2u16 = vmull_u8(vget_high_u8(q11u8), d0u8);
+      q3u16 = vmull_u8(vget_low_u8(q12u8), d0u8);
+      q4u16 = vmull_u8(vget_high_u8(q12u8), d0u8);
+      q5u16 = vmull_u8(vget_low_u8(q13u8), d0u8);
+      q6u16 = vmull_u8(vget_high_u8(q13u8), d0u8);
+      q7u16 = vmull_u8(vget_low_u8(q14u8), d0u8);
+      q8u16 = vmull_u8(vget_high_u8(q14u8), d0u8);
 
-            q1u16 = vmlal_u8(q1u16, vget_low_u8(q12u8), d1u8);
-            q2u16 = vmlal_u8(q2u16, vget_high_u8(q12u8), d1u8);
-            q3u16 = vmlal_u8(q3u16, vget_low_u8(q13u8), d1u8);
-            q4u16 = vmlal_u8(q4u16, vget_high_u8(q13u8), d1u8);
-            q5u16 = vmlal_u8(q5u16, vget_low_u8(q14u8), d1u8);
-            q6u16 = vmlal_u8(q6u16, vget_high_u8(q14u8), d1u8);
-            q7u16 = vmlal_u8(q7u16, vget_low_u8(q15u8), d1u8);
-            q8u16 = vmlal_u8(q8u16, vget_high_u8(q15u8), d1u8);
+      q1u16 = vmlal_u8(q1u16, vget_low_u8(q12u8), d1u8);
+      q2u16 = vmlal_u8(q2u16, vget_high_u8(q12u8), d1u8);
+      q3u16 = vmlal_u8(q3u16, vget_low_u8(q13u8), d1u8);
+      q4u16 = vmlal_u8(q4u16, vget_high_u8(q13u8), d1u8);
+      q5u16 = vmlal_u8(q5u16, vget_low_u8(q14u8), d1u8);
+      q6u16 = vmlal_u8(q6u16, vget_high_u8(q14u8), d1u8);
+      q7u16 = vmlal_u8(q7u16, vget_low_u8(q15u8), d1u8);
+      q8u16 = vmlal_u8(q8u16, vget_high_u8(q15u8), d1u8);
 
-            d2u8 = vqrshrn_n_u16(q1u16, 7);
-            d3u8 = vqrshrn_n_u16(q2u16, 7);
-            d4u8 = vqrshrn_n_u16(q3u16, 7);
-            d5u8 = vqrshrn_n_u16(q4u16, 7);
-            d6u8 = vqrshrn_n_u16(q5u16, 7);
-            d7u8 = vqrshrn_n_u16(q6u16, 7);
-            d8u8 = vqrshrn_n_u16(q7u16, 7);
-            d9u8 = vqrshrn_n_u16(q8u16, 7);
+      d2u8 = vqrshrn_n_u16(q1u16, 7);
+      d3u8 = vqrshrn_n_u16(q2u16, 7);
+      d4u8 = vqrshrn_n_u16(q3u16, 7);
+      d5u8 = vqrshrn_n_u16(q4u16, 7);
+      d6u8 = vqrshrn_n_u16(q5u16, 7);
+      d7u8 = vqrshrn_n_u16(q6u16, 7);
+      d8u8 = vqrshrn_n_u16(q7u16, 7);
+      d9u8 = vqrshrn_n_u16(q8u16, 7);
 
-            q1u8 = vcombine_u8(d2u8, d3u8);
-            q2u8 = vcombine_u8(d4u8, d5u8);
-            q3u8 = vcombine_u8(d6u8, d7u8);
-            q4u8 = vcombine_u8(d8u8, d9u8);
+      q1u8 = vcombine_u8(d2u8, d3u8);
+      q2u8 = vcombine_u8(d4u8, d5u8);
+      q3u8 = vcombine_u8(d6u8, d7u8);
+      q4u8 = vcombine_u8(d8u8, d9u8);
 
-            q11u8 = q15u8;
+      q11u8 = q15u8;
 
-            vst1q_u8((uint8_t *)dst_ptr, q1u8); dst_ptr += dst_pitch;
-            vst1q_u8((uint8_t *)dst_ptr, q2u8); dst_ptr += dst_pitch;
-            vst1q_u8((uint8_t *)dst_ptr, q3u8); dst_ptr += dst_pitch;
-            vst1q_u8((uint8_t *)dst_ptr, q4u8); dst_ptr += dst_pitch;
-        }
-        return;
+      vst1q_u8((uint8_t *)dst_ptr, q1u8);
+      dst_ptr += dst_pitch;
+      vst1q_u8((uint8_t *)dst_ptr, q2u8);
+      dst_ptr += dst_pitch;
+      vst1q_u8((uint8_t *)dst_ptr, q3u8);
+      dst_ptr += dst_pitch;
+      vst1q_u8((uint8_t *)dst_ptr, q4u8);
+      dst_ptr += dst_pitch;
     }
+    return;
+  }
 
-    if (yoffset == 0) {  // firstpass_bfilter16x16_only
-        d0u8 = vdup_n_u8(bifilter4_coeff[xoffset][0]);
-        d1u8 = vdup_n_u8(bifilter4_coeff[xoffset][1]);
+  if (yoffset == 0) {  // firstpass_bfilter16x16_only
+    d0u8 = vdup_n_u8(bifilter4_coeff[xoffset][0]);
+    d1u8 = vdup_n_u8(bifilter4_coeff[xoffset][1]);
 
-        for (i = 4; i > 0 ; i--) {
-            d2u8 = vld1_u8(src_ptr);
-            d3u8 = vld1_u8(src_ptr + 8);
-            d4u8 = vld1_u8(src_ptr + 16); src_ptr += src_pixels_per_line;
-            d5u8 = vld1_u8(src_ptr);
-            d6u8 = vld1_u8(src_ptr + 8);
-            d7u8 = vld1_u8(src_ptr + 16); src_ptr += src_pixels_per_line;
-            d8u8 = vld1_u8(src_ptr);
-            d9u8 = vld1_u8(src_ptr + 8);
-            d10u8 = vld1_u8(src_ptr + 16); src_ptr += src_pixels_per_line;
-            d11u8 = vld1_u8(src_ptr);
-            d12u8 = vld1_u8(src_ptr + 8);
-            d13u8 = vld1_u8(src_ptr + 16); src_ptr += src_pixels_per_line;
+    for (i = 4; i > 0; i--) {
+      d2u8 = vld1_u8(src_ptr);
+      d3u8 = vld1_u8(src_ptr + 8);
+      d4u8 = vld1_u8(src_ptr + 16);
+      src_ptr += src_pixels_per_line;
+      d5u8 = vld1_u8(src_ptr);
+      d6u8 = vld1_u8(src_ptr + 8);
+      d7u8 = vld1_u8(src_ptr + 16);
+      src_ptr += src_pixels_per_line;
+      d8u8 = vld1_u8(src_ptr);
+      d9u8 = vld1_u8(src_ptr + 8);
+      d10u8 = vld1_u8(src_ptr + 16);
+      src_ptr += src_pixels_per_line;
+      d11u8 = vld1_u8(src_ptr);
+      d12u8 = vld1_u8(src_ptr + 8);
+      d13u8 = vld1_u8(src_ptr + 16);
+      src_ptr += src_pixels_per_line;
 
-            q7u16  = vmull_u8(d2u8, d0u8);
-            q8u16  = vmull_u8(d3u8, d0u8);
-            q9u16  = vmull_u8(d5u8, d0u8);
-            q10u16 = vmull_u8(d6u8, d0u8);
-            q11u16 = vmull_u8(d8u8, d0u8);
-            q12u16 = vmull_u8(d9u8, d0u8);
-            q13u16 = vmull_u8(d11u8, d0u8);
-            q14u16 = vmull_u8(d12u8, d0u8);
+      q7u16 = vmull_u8(d2u8, d0u8);
+      q8u16 = vmull_u8(d3u8, d0u8);
+      q9u16 = vmull_u8(d5u8, d0u8);
+      q10u16 = vmull_u8(d6u8, d0u8);
+      q11u16 = vmull_u8(d8u8, d0u8);
+      q12u16 = vmull_u8(d9u8, d0u8);
+      q13u16 = vmull_u8(d11u8, d0u8);
+      q14u16 = vmull_u8(d12u8, d0u8);
 
-            d2u8  = vext_u8(d2u8, d3u8, 1);
-            d5u8  = vext_u8(d5u8, d6u8, 1);
-            d8u8  = vext_u8(d8u8, d9u8, 1);
-            d11u8 = vext_u8(d11u8, d12u8, 1);
+      d2u8 = vext_u8(d2u8, d3u8, 1);
+      d5u8 = vext_u8(d5u8, d6u8, 1);
+      d8u8 = vext_u8(d8u8, d9u8, 1);
+      d11u8 = vext_u8(d11u8, d12u8, 1);
 
-            q7u16  = vmlal_u8(q7u16, d2u8, d1u8);
-            q9u16  = vmlal_u8(q9u16, d5u8, d1u8);
-            q11u16 = vmlal_u8(q11u16, d8u8, d1u8);
-            q13u16 = vmlal_u8(q13u16, d11u8, d1u8);
+      q7u16 = vmlal_u8(q7u16, d2u8, d1u8);
+      q9u16 = vmlal_u8(q9u16, d5u8, d1u8);
+      q11u16 = vmlal_u8(q11u16, d8u8, d1u8);
+      q13u16 = vmlal_u8(q13u16, d11u8, d1u8);
 
-            d3u8  = vext_u8(d3u8, d4u8, 1);
-            d6u8  = vext_u8(d6u8, d7u8, 1);
-            d9u8  = vext_u8(d9u8, d10u8, 1);
-            d12u8 = vext_u8(d12u8, d13u8, 1);
+      d3u8 = vext_u8(d3u8, d4u8, 1);
+      d6u8 = vext_u8(d6u8, d7u8, 1);
+      d9u8 = vext_u8(d9u8, d10u8, 1);
+      d12u8 = vext_u8(d12u8, d13u8, 1);
 
-            q8u16  = vmlal_u8(q8u16,  d3u8, d1u8);
-            q10u16 = vmlal_u8(q10u16, d6u8, d1u8);
-            q12u16 = vmlal_u8(q12u16, d9u8, d1u8);
-            q14u16 = vmlal_u8(q14u16, d12u8, d1u8);
+      q8u16 = vmlal_u8(q8u16, d3u8, d1u8);
+      q10u16 = vmlal_u8(q10u16, d6u8, d1u8);
+      q12u16 = vmlal_u8(q12u16, d9u8, d1u8);
+      q14u16 = vmlal_u8(q14u16, d12u8, d1u8);
 
-            d14u8 = vqrshrn_n_u16(q7u16, 7);
-            d15u8 = vqrshrn_n_u16(q8u16, 7);
-            d16u8 = vqrshrn_n_u16(q9u16, 7);
-            d17u8 = vqrshrn_n_u16(q10u16, 7);
-            d18u8 = vqrshrn_n_u16(q11u16, 7);
-            d19u8 = vqrshrn_n_u16(q12u16, 7);
-            d20u8 = vqrshrn_n_u16(q13u16, 7);
-            d21u8 = vqrshrn_n_u16(q14u16, 7);
+      d14u8 = vqrshrn_n_u16(q7u16, 7);
+      d15u8 = vqrshrn_n_u16(q8u16, 7);
+      d16u8 = vqrshrn_n_u16(q9u16, 7);
+      d17u8 = vqrshrn_n_u16(q10u16, 7);
+      d18u8 = vqrshrn_n_u16(q11u16, 7);
+      d19u8 = vqrshrn_n_u16(q12u16, 7);
+      d20u8 = vqrshrn_n_u16(q13u16, 7);
+      d21u8 = vqrshrn_n_u16(q14u16, 7);
 
-            q7u8 = vcombine_u8(d14u8, d15u8);
-            q8u8 = vcombine_u8(d16u8, d17u8);
-            q9u8 = vcombine_u8(d18u8, d19u8);
-            q10u8 =vcombine_u8(d20u8, d21u8);
+      q7u8 = vcombine_u8(d14u8, d15u8);
+      q8u8 = vcombine_u8(d16u8, d17u8);
+      q9u8 = vcombine_u8(d18u8, d19u8);
+      q10u8 = vcombine_u8(d20u8, d21u8);
 
-            vst1q_u8((uint8_t *)dst_ptr, q7u8); dst_ptr += dst_pitch;
-            vst1q_u8((uint8_t *)dst_ptr, q8u8); dst_ptr += dst_pitch;
-            vst1q_u8((uint8_t *)dst_ptr, q9u8); dst_ptr += dst_pitch;
-            vst1q_u8((uint8_t *)dst_ptr, q10u8); dst_ptr += dst_pitch;
-        }
-        return;
+      vst1q_u8((uint8_t *)dst_ptr, q7u8);
+      dst_ptr += dst_pitch;
+      vst1q_u8((uint8_t *)dst_ptr, q8u8);
+      dst_ptr += dst_pitch;
+      vst1q_u8((uint8_t *)dst_ptr, q9u8);
+      dst_ptr += dst_pitch;
+      vst1q_u8((uint8_t *)dst_ptr, q10u8);
+      dst_ptr += dst_pitch;
     }
+    return;
+  }
 
-    d0u8 = vdup_n_u8(bifilter4_coeff[xoffset][0]);
-    d1u8 = vdup_n_u8(bifilter4_coeff[xoffset][1]);
+  d0u8 = vdup_n_u8(bifilter4_coeff[xoffset][0]);
+  d1u8 = vdup_n_u8(bifilter4_coeff[xoffset][1]);
 
+  d2u8 = vld1_u8(src_ptr);
+  d3u8 = vld1_u8(src_ptr + 8);
+  d4u8 = vld1_u8(src_ptr + 16);
+  src_ptr += src_pixels_per_line;
+  d5u8 = vld1_u8(src_ptr);
+  d6u8 = vld1_u8(src_ptr + 8);
+  d7u8 = vld1_u8(src_ptr + 16);
+  src_ptr += src_pixels_per_line;
+  d8u8 = vld1_u8(src_ptr);
+  d9u8 = vld1_u8(src_ptr + 8);
+  d10u8 = vld1_u8(src_ptr + 16);
+  src_ptr += src_pixels_per_line;
+  d11u8 = vld1_u8(src_ptr);
+  d12u8 = vld1_u8(src_ptr + 8);
+  d13u8 = vld1_u8(src_ptr + 16);
+  src_ptr += src_pixels_per_line;
+
+  // First Pass: output_height lines x output_width columns (17x16)
+  tmpp = tmp;
+  for (i = 3; i > 0; i--) {
+    q7u16 = vmull_u8(d2u8, d0u8);
+    q8u16 = vmull_u8(d3u8, d0u8);
+    q9u16 = vmull_u8(d5u8, d0u8);
+    q10u16 = vmull_u8(d6u8, d0u8);
+    q11u16 = vmull_u8(d8u8, d0u8);
+    q12u16 = vmull_u8(d9u8, d0u8);
+    q13u16 = vmull_u8(d11u8, d0u8);
+    q14u16 = vmull_u8(d12u8, d0u8);
+
+    d2u8 = vext_u8(d2u8, d3u8, 1);
+    d5u8 = vext_u8(d5u8, d6u8, 1);
+    d8u8 = vext_u8(d8u8, d9u8, 1);
+    d11u8 = vext_u8(d11u8, d12u8, 1);
+
+    q7u16 = vmlal_u8(q7u16, d2u8, d1u8);
+    q9u16 = vmlal_u8(q9u16, d5u8, d1u8);
+    q11u16 = vmlal_u8(q11u16, d8u8, d1u8);
+    q13u16 = vmlal_u8(q13u16, d11u8, d1u8);
+
+    d3u8 = vext_u8(d3u8, d4u8, 1);
+    d6u8 = vext_u8(d6u8, d7u8, 1);
+    d9u8 = vext_u8(d9u8, d10u8, 1);
+    d12u8 = vext_u8(d12u8, d13u8, 1);
+
+    q8u16 = vmlal_u8(q8u16, d3u8, d1u8);
+    q10u16 = vmlal_u8(q10u16, d6u8, d1u8);
+    q12u16 = vmlal_u8(q12u16, d9u8, d1u8);
+    q14u16 = vmlal_u8(q14u16, d12u8, d1u8);
+
+    d14u8 = vqrshrn_n_u16(q7u16, 7);
+    d15u8 = vqrshrn_n_u16(q8u16, 7);
+    d16u8 = vqrshrn_n_u16(q9u16, 7);
+    d17u8 = vqrshrn_n_u16(q10u16, 7);
+    d18u8 = vqrshrn_n_u16(q11u16, 7);
+    d19u8 = vqrshrn_n_u16(q12u16, 7);
+    d20u8 = vqrshrn_n_u16(q13u16, 7);
+    d21u8 = vqrshrn_n_u16(q14u16, 7);
+
     d2u8 = vld1_u8(src_ptr);
     d3u8 = vld1_u8(src_ptr + 8);
-    d4u8 = vld1_u8(src_ptr + 16); src_ptr += src_pixels_per_line;
+    d4u8 = vld1_u8(src_ptr + 16);
+    src_ptr += src_pixels_per_line;
     d5u8 = vld1_u8(src_ptr);
     d6u8 = vld1_u8(src_ptr + 8);
-    d7u8 = vld1_u8(src_ptr + 16); src_ptr += src_pixels_per_line;
+    d7u8 = vld1_u8(src_ptr + 16);
+    src_ptr += src_pixels_per_line;
     d8u8 = vld1_u8(src_ptr);
     d9u8 = vld1_u8(src_ptr + 8);
-    d10u8 = vld1_u8(src_ptr + 16); src_ptr += src_pixels_per_line;
+    d10u8 = vld1_u8(src_ptr + 16);
+    src_ptr += src_pixels_per_line;
     d11u8 = vld1_u8(src_ptr);
     d12u8 = vld1_u8(src_ptr + 8);
-    d13u8 = vld1_u8(src_ptr + 16); src_ptr += src_pixels_per_line;
+    d13u8 = vld1_u8(src_ptr + 16);
+    src_ptr += src_pixels_per_line;
 
-    // First Pass: output_height lines x output_width columns (17x16)
-    tmpp = tmp;
-    for (i = 3; i > 0; i--) {
-        q7u16  = vmull_u8(d2u8, d0u8);
-        q8u16  = vmull_u8(d3u8, d0u8);
-        q9u16  = vmull_u8(d5u8, d0u8);
-        q10u16 = vmull_u8(d6u8, d0u8);
-        q11u16 = vmull_u8(d8u8, d0u8);
-        q12u16 = vmull_u8(d9u8, d0u8);
-        q13u16 = vmull_u8(d11u8, d0u8);
-        q14u16 = vmull_u8(d12u8, d0u8);
+    q7u8 = vcombine_u8(d14u8, d15u8);
+    q8u8 = vcombine_u8(d16u8, d17u8);
+    q9u8 = vcombine_u8(d18u8, d19u8);
+    q10u8 = vcombine_u8(d20u8, d21u8);
 
-        d2u8  = vext_u8(d2u8, d3u8, 1);
-        d5u8  = vext_u8(d5u8, d6u8, 1);
-        d8u8  = vext_u8(d8u8, d9u8, 1);
-        d11u8 = vext_u8(d11u8, d12u8, 1);
+    vst1q_u8((uint8_t *)tmpp, q7u8);
+    tmpp += 16;
+    vst1q_u8((uint8_t *)tmpp, q8u8);
+    tmpp += 16;
+    vst1q_u8((uint8_t *)tmpp, q9u8);
+    tmpp += 16;
+    vst1q_u8((uint8_t *)tmpp, q10u8);
+    tmpp += 16;
+  }
 
-        q7u16  = vmlal_u8(q7u16, d2u8, d1u8);
-        q9u16  = vmlal_u8(q9u16, d5u8, d1u8);
-        q11u16 = vmlal_u8(q11u16, d8u8, d1u8);
-        q13u16 = vmlal_u8(q13u16, d11u8, d1u8);
+  // First-pass filtering for rest 5 lines
+  d14u8 = vld1_u8(src_ptr);
+  d15u8 = vld1_u8(src_ptr + 8);
+  d16u8 = vld1_u8(src_ptr + 16);
+  src_ptr += src_pixels_per_line;
 
-        d3u8  = vext_u8(d3u8, d4u8, 1);
-        d6u8  = vext_u8(d6u8, d7u8, 1);
-        d9u8  = vext_u8(d9u8, d10u8, 1);
-        d12u8 = vext_u8(d12u8, d13u8, 1);
+  q9u16 = vmull_u8(d2u8, d0u8);
+  q10u16 = vmull_u8(d3u8, d0u8);
+  q11u16 = vmull_u8(d5u8, d0u8);
+  q12u16 = vmull_u8(d6u8, d0u8);
+  q13u16 = vmull_u8(d8u8, d0u8);
+  q14u16 = vmull_u8(d9u8, d0u8);
 
-        q8u16  = vmlal_u8(q8u16,  d3u8, d1u8);
-        q10u16 = vmlal_u8(q10u16, d6u8, d1u8);
-        q12u16 = vmlal_u8(q12u16, d9u8, d1u8);
-        q14u16 = vmlal_u8(q14u16, d12u8, d1u8);
+  d2u8 = vext_u8(d2u8, d3u8, 1);
+  d5u8 = vext_u8(d5u8, d6u8, 1);
+  d8u8 = vext_u8(d8u8, d9u8, 1);
 
-        d14u8 = vqrshrn_n_u16(q7u16, 7);
-        d15u8 = vqrshrn_n_u16(q8u16, 7);
-        d16u8 = vqrshrn_n_u16(q9u16, 7);
-        d17u8 = vqrshrn_n_u16(q10u16, 7);
-        d18u8 = vqrshrn_n_u16(q11u16, 7);
-        d19u8 = vqrshrn_n_u16(q12u16, 7);
-        d20u8 = vqrshrn_n_u16(q13u16, 7);
-        d21u8 = vqrshrn_n_u16(q14u16, 7);
+  q9u16 = vmlal_u8(q9u16, d2u8, d1u8);
+  q11u16 = vmlal_u8(q11u16, d5u8, d1u8);
+  q13u16 = vmlal_u8(q13u16, d8u8, d1u8);
 
-        d2u8 = vld1_u8(src_ptr);
-        d3u8 = vld1_u8(src_ptr + 8);
-        d4u8 = vld1_u8(src_ptr + 16); src_ptr += src_pixels_per_line;
-        d5u8 = vld1_u8(src_ptr);
-        d6u8 = vld1_u8(src_ptr + 8);
-        d7u8 = vld1_u8(src_ptr + 16); src_ptr += src_pixels_per_line;
-        d8u8 = vld1_u8(src_ptr);
-        d9u8 = vld1_u8(src_ptr + 8);
-        d10u8 = vld1_u8(src_ptr + 16); src_ptr += src_pixels_per_line;
-        d11u8 = vld1_u8(src_ptr);
-        d12u8 = vld1_u8(src_ptr + 8);
-        d13u8 = vld1_u8(src_ptr + 16); src_ptr += src_pixels_per_line;
+  d3u8 = vext_u8(d3u8, d4u8, 1);
+  d6u8 = vext_u8(d6u8, d7u8, 1);
+  d9u8 = vext_u8(d9u8, d10u8, 1);
 
-        q7u8 = vcombine_u8(d14u8, d15u8);
-        q8u8 = vcombine_u8(d16u8, d17u8);
-        q9u8 = vcombine_u8(d18u8, d19u8);
-        q10u8 = vcombine_u8(d20u8, d21u8);
+  q10u16 = vmlal_u8(q10u16, d3u8, d1u8);
+  q12u16 = vmlal_u8(q12u16, d6u8, d1u8);
+  q14u16 = vmlal_u8(q14u16, d9u8, d1u8);
 
-        vst1q_u8((uint8_t *)tmpp, q7u8); tmpp += 16;
-        vst1q_u8((uint8_t *)tmpp, q8u8); tmpp += 16;
-        vst1q_u8((uint8_t *)tmpp, q9u8); tmpp += 16;
-        vst1q_u8((uint8_t *)tmpp, q10u8); tmpp += 16;
-    }
+  q1u16 = vmull_u8(d11u8, d0u8);
+  q2u16 = vmull_u8(d12u8, d0u8);
+  q3u16 = vmull_u8(d14u8, d0u8);
+  q4u16 = vmull_u8(d15u8, d0u8);
 
-    // First-pass filtering for rest 5 lines
-    d14u8 = vld1_u8(src_ptr);
-    d15u8 = vld1_u8(src_ptr + 8);
-    d16u8 = vld1_u8(src_ptr + 16); src_ptr += src_pixels_per_line;
+  d11u8 = vext_u8(d11u8, d12u8, 1);
+  d14u8 = vext_u8(d14u8, d15u8, 1);
 
-    q9u16  = vmull_u8(d2u8, d0u8);
-    q10u16 = vmull_u8(d3u8, d0u8);
-    q11u16 = vmull_u8(d5u8, d0u8);
-    q12u16 = vmull_u8(d6u8, d0u8);
-    q13u16 = vmull_u8(d8u8, d0u8);
-    q14u16 = vmull_u8(d9u8, d0u8);
+  q1u16 = vmlal_u8(q1u16, d11u8, d1u8);
+  q3u16 = vmlal_u8(q3u16, d14u8, d1u8);
 
-    d2u8  = vext_u8(d2u8, d3u8, 1);
-    d5u8  = vext_u8(d5u8, d6u8, 1);
-    d8u8  = vext_u8(d8u8, d9u8, 1);
+  d12u8 = vext_u8(d12u8, d13u8, 1);
+  d15u8 = vext_u8(d15u8, d16u8, 1);
 
-    q9u16  = vmlal_u8(q9u16, d2u8, d1u8);
-    q11u16 = vmlal_u8(q11u16, d5u8, d1u8);
-    q13u16 = vmlal_u8(q13u16, d8u8, d1u8);
+  q2u16 = vmlal_u8(q2u16, d12u8, d1u8);
+  q4u16 = vmlal_u8(q4u16, d15u8, d1u8);
 
-    d3u8  = vext_u8(d3u8, d4u8, 1);
-    d6u8  = vext_u8(d6u8, d7u8, 1);
-    d9u8  = vext_u8(d9u8, d10u8, 1);
+  d10u8 = vqrshrn_n_u16(q9u16, 7);
+  d11u8 = vqrshrn_n_u16(q10u16, 7);
+  d12u8 = vqrshrn_n_u16(q11u16, 7);
+  d13u8 = vqrshrn_n_u16(q12u16, 7);
+  d14u8 = vqrshrn_n_u16(q13u16, 7);
+  d15u8 = vqrshrn_n_u16(q14u16, 7);
+  d16u8 = vqrshrn_n_u16(q1u16, 7);
+  d17u8 = vqrshrn_n_u16(q2u16, 7);
+  d18u8 = vqrshrn_n_u16(q3u16, 7);
+  d19u8 = vqrshrn_n_u16(q4u16, 7);
 
-    q10u16 = vmlal_u8(q10u16, d3u8, d1u8);
-    q12u16 = vmlal_u8(q12u16, d6u8, d1u8);
-    q14u16 = vmlal_u8(q14u16, d9u8, d1u8);
+  q5u8 = vcombine_u8(d10u8, d11u8);
+  q6u8 = vcombine_u8(d12u8, d13u8);
+  q7u8 = vcombine_u8(d14u8, d15u8);
+  q8u8 = vcombine_u8(d16u8, d17u8);
+  q9u8 = vcombine_u8(d18u8, d19u8);
 
-    q1u16 = vmull_u8(d11u8, d0u8);
-    q2u16 = vmull_u8(d12u8, d0u8);
-    q3u16 = vmull_u8(d14u8, d0u8);
-    q4u16 = vmull_u8(d15u8, d0u8);
+  vst1q_u8((uint8_t *)tmpp, q5u8);
+  tmpp += 16;
+  vst1q_u8((uint8_t *)tmpp, q6u8);
+  tmpp += 16;
+  vst1q_u8((uint8_t *)tmpp, q7u8);
+  tmpp += 16;
+  vst1q_u8((uint8_t *)tmpp, q8u8);
+  tmpp += 16;
+  vst1q_u8((uint8_t *)tmpp, q9u8);
 
-    d11u8 = vext_u8(d11u8, d12u8, 1);
-    d14u8 = vext_u8(d14u8, d15u8, 1);
+  // secondpass_filter
+  d0u8 = vdup_n_u8(bifilter4_coeff[yoffset][0]);
+  d1u8 = vdup_n_u8(bifilter4_coeff[yoffset][1]);
 
-    q1u16 = vmlal_u8(q1u16, d11u8, d1u8);
-    q3u16 = vmlal_u8(q3u16, d14u8, d1u8);
-
-    d12u8 = vext_u8(d12u8, d13u8, 1);
-    d15u8 = vext_u8(d15u8, d16u8, 1);
-
-    q2u16 = vmlal_u8(q2u16, d12u8, d1u8);
-    q4u16 = vmlal_u8(q4u16, d15u8, d1u8);
-
-    d10u8 = vqrshrn_n_u16(q9u16, 7);
-    d11u8 = vqrshrn_n_u16(q10u16, 7);
-    d12u8 = vqrshrn_n_u16(q11u16, 7);
-    d13u8 = vqrshrn_n_u16(q12u16, 7);
-    d14u8 = vqrshrn_n_u16(q13u16, 7);
-    d15u8 = vqrshrn_n_u16(q14u16, 7);
-    d16u8 = vqrshrn_n_u16(q1u16, 7);
-    d17u8 = vqrshrn_n_u16(q2u16, 7);
-    d18u8 = vqrshrn_n_u16(q3u16, 7);
-    d19u8 = vqrshrn_n_u16(q4u16, 7);
-
-    q5u8 = vcombine_u8(d10u8, d11u8);
-    q6u8 = vcombine_u8(d12u8, d13u8);
-    q7u8 = vcombine_u8(d14u8, d15u8);
-    q8u8 = vcombine_u8(d16u8, d17u8);
-    q9u8 = vcombine_u8(d18u8, d19u8);
-
-    vst1q_u8((uint8_t *)tmpp, q5u8); tmpp += 16;
-    vst1q_u8((uint8_t *)tmpp, q6u8); tmpp += 16;
-    vst1q_u8((uint8_t *)tmpp, q7u8); tmpp += 16;
-    vst1q_u8((uint8_t *)tmpp, q8u8); tmpp += 16;
-    vst1q_u8((uint8_t *)tmpp, q9u8);
-
-    // secondpass_filter
-    d0u8 = vdup_n_u8(bifilter4_coeff[yoffset][0]);
-    d1u8 = vdup_n_u8(bifilter4_coeff[yoffset][1]);
-
-    tmpp = tmp;
-    q11u8 = vld1q_u8(tmpp);
+  tmpp = tmp;
+  q11u8 = vld1q_u8(tmpp);
+  tmpp += 16;
+  for (i = 4; i > 0; i--) {
+    q12u8 = vld1q_u8(tmpp);
     tmpp += 16;
-    for (i = 4; i > 0; i--) {
-        q12u8 = vld1q_u8(tmpp); tmpp += 16;
-        q13u8 = vld1q_u8(tmpp); tmpp += 16;
-        q14u8 = vld1q_u8(tmpp); tmpp += 16;
-        q15u8 = vld1q_u8(tmpp); tmpp += 16;
+    q13u8 = vld1q_u8(tmpp);
+    tmpp += 16;
+    q14u8 = vld1q_u8(tmpp);
+    tmpp += 16;
+    q15u8 = vld1q_u8(tmpp);
+    tmpp += 16;
 
-        q1u16 = vmull_u8(vget_low_u8(q11u8), d0u8);
-        q2u16 = vmull_u8(vget_high_u8(q11u8), d0u8);
-        q3u16 = vmull_u8(vget_low_u8(q12u8), d0u8);
-        q4u16 = vmull_u8(vget_high_u8(q12u8), d0u8);
-        q5u16 = vmull_u8(vget_low_u8(q13u8), d0u8);
-        q6u16 = vmull_u8(vget_high_u8(q13u8), d0u8);
-        q7u16 = vmull_u8(vget_low_u8(q14u8), d0u8);
-        q8u16 = vmull_u8(vget_high_u8(q14u8), d0u8);
+    q1u16 = vmull_u8(vget_low_u8(q11u8), d0u8);
+    q2u16 = vmull_u8(vget_high_u8(q11u8), d0u8);
+    q3u16 = vmull_u8(vget_low_u8(q12u8), d0u8);
+    q4u16 = vmull_u8(vget_high_u8(q12u8), d0u8);
+    q5u16 = vmull_u8(vget_low_u8(q13u8), d0u8);
+    q6u16 = vmull_u8(vget_high_u8(q13u8), d0u8);
+    q7u16 = vmull_u8(vget_low_u8(q14u8), d0u8);
+    q8u16 = vmull_u8(vget_high_u8(q14u8), d0u8);
 
-        q1u16 = vmlal_u8(q1u16, vget_low_u8(q12u8), d1u8);
-        q2u16 = vmlal_u8(q2u16, vget_high_u8(q12u8), d1u8);
-        q3u16 = vmlal_u8(q3u16, vget_low_u8(q13u8), d1u8);
-        q4u16 = vmlal_u8(q4u16, vget_high_u8(q13u8), d1u8);
-        q5u16 = vmlal_u8(q5u16, vget_low_u8(q14u8), d1u8);
-        q6u16 = vmlal_u8(q6u16, vget_high_u8(q14u8), d1u8);
-        q7u16 = vmlal_u8(q7u16, vget_low_u8(q15u8), d1u8);
-        q8u16 = vmlal_u8(q8u16, vget_high_u8(q15u8), d1u8);
+    q1u16 = vmlal_u8(q1u16, vget_low_u8(q12u8), d1u8);
+    q2u16 = vmlal_u8(q2u16, vget_high_u8(q12u8), d1u8);
+    q3u16 = vmlal_u8(q3u16, vget_low_u8(q13u8), d1u8);
+    q4u16 = vmlal_u8(q4u16, vget_high_u8(q13u8), d1u8);
+    q5u16 = vmlal_u8(q5u16, vget_low_u8(q14u8), d1u8);
+    q6u16 = vmlal_u8(q6u16, vget_high_u8(q14u8), d1u8);
+    q7u16 = vmlal_u8(q7u16, vget_low_u8(q15u8), d1u8);
+    q8u16 = vmlal_u8(q8u16, vget_high_u8(q15u8), d1u8);
 
-        d2u8 = vqrshrn_n_u16(q1u16, 7);
-        d3u8 = vqrshrn_n_u16(q2u16, 7);
-        d4u8 = vqrshrn_n_u16(q3u16, 7);
-        d5u8 = vqrshrn_n_u16(q4u16, 7);
-        d6u8 = vqrshrn_n_u16(q5u16, 7);
-        d7u8 = vqrshrn_n_u16(q6u16, 7);
-        d8u8 = vqrshrn_n_u16(q7u16, 7);
-        d9u8 = vqrshrn_n_u16(q8u16, 7);
+    d2u8 = vqrshrn_n_u16(q1u16, 7);
+    d3u8 = vqrshrn_n_u16(q2u16, 7);
+    d4u8 = vqrshrn_n_u16(q3u16, 7);
+    d5u8 = vqrshrn_n_u16(q4u16, 7);
+    d6u8 = vqrshrn_n_u16(q5u16, 7);
+    d7u8 = vqrshrn_n_u16(q6u16, 7);
+    d8u8 = vqrshrn_n_u16(q7u16, 7);
+    d9u8 = vqrshrn_n_u16(q8u16, 7);
 
-        q1u8 = vcombine_u8(d2u8, d3u8);
-        q2u8 = vcombine_u8(d4u8, d5u8);
-        q3u8 = vcombine_u8(d6u8, d7u8);
-        q4u8 = vcombine_u8(d8u8, d9u8);
+    q1u8 = vcombine_u8(d2u8, d3u8);
+    q2u8 = vcombine_u8(d4u8, d5u8);
+    q3u8 = vcombine_u8(d6u8, d7u8);
+    q4u8 = vcombine_u8(d8u8, d9u8);
 
-        q11u8 = q15u8;
+    q11u8 = q15u8;
 
-        vst1q_u8((uint8_t *)dst_ptr, q1u8); dst_ptr += dst_pitch;
-        vst1q_u8((uint8_t *)dst_ptr, q2u8); dst_ptr += dst_pitch;
-        vst1q_u8((uint8_t *)dst_ptr, q3u8); dst_ptr += dst_pitch;
-        vst1q_u8((uint8_t *)dst_ptr, q4u8); dst_ptr += dst_pitch;
-    }
-    return;
+    vst1q_u8((uint8_t *)dst_ptr, q1u8);
+    dst_ptr += dst_pitch;
+    vst1q_u8((uint8_t *)dst_ptr, q2u8);
+    dst_ptr += dst_pitch;
+    vst1q_u8((uint8_t *)dst_ptr, q3u8);
+    dst_ptr += dst_pitch;
+    vst1q_u8((uint8_t *)dst_ptr, q4u8);
+    dst_ptr += dst_pitch;
+  }
+  return;
 }
--- a/vp8/common/arm/neon/copymem_neon.c
+++ b/vp8/common/arm/neon/copymem_neon.c
@@ -10,50 +10,41 @@
 
 #include <arm_neon.h>
 
-void vp8_copy_mem8x4_neon(
-        unsigned char *src,
-        int src_stride,
-        unsigned char *dst,
-        int dst_stride) {
-    uint8x8_t vtmp;
-    int r;
+void vp8_copy_mem8x4_neon(unsigned char *src, int src_stride,
+                          unsigned char *dst, int dst_stride) {
+  uint8x8_t vtmp;
+  int r;
 
-    for (r = 0; r < 4; r++) {
-        vtmp = vld1_u8(src);
-        vst1_u8(dst, vtmp);
-        src += src_stride;
-        dst += dst_stride;
-    }
+  for (r = 0; r < 4; r++) {
+    vtmp = vld1_u8(src);
+    vst1_u8(dst, vtmp);
+    src += src_stride;
+    dst += dst_stride;
+  }
 }
 
-void vp8_copy_mem8x8_neon(
-        unsigned char *src,
-        int src_stride,
-        unsigned char *dst,
-        int dst_stride) {
-    uint8x8_t vtmp;
-    int r;
+void vp8_copy_mem8x8_neon(unsigned char *src, int src_stride,
+                          unsigned char *dst, int dst_stride) {
+  uint8x8_t vtmp;
+  int r;
 
-    for (r = 0; r < 8; r++) {
-        vtmp = vld1_u8(src);
-        vst1_u8(dst, vtmp);
-        src += src_stride;
-        dst += dst_stride;
-    }
+  for (r = 0; r < 8; r++) {
+    vtmp = vld1_u8(src);
+    vst1_u8(dst, vtmp);
+    src += src_stride;
+    dst += dst_stride;
+  }
 }
 
-void vp8_copy_mem16x16_neon(
-        unsigned char *src,
-        int src_stride,
-        unsigned char *dst,
-        int dst_stride) {
-    int r;
-    uint8x16_t qtmp;
+void vp8_copy_mem16x16_neon(unsigned char *src, int src_stride,
+                            unsigned char *dst, int dst_stride) {
+  int r;
+  uint8x16_t qtmp;
 
-    for (r = 0; r < 16; r++) {
-        qtmp = vld1q_u8(src);
-        vst1q_u8(dst, qtmp);
-        src += src_stride;
-        dst += dst_stride;
-    }
+  for (r = 0; r < 16; r++) {
+    qtmp = vld1q_u8(src);
+    vst1q_u8(dst, qtmp);
+    src += src_stride;
+    dst += dst_stride;
+  }
 }
--- a/vp8/common/arm/neon/dc_only_idct_add_neon.c
+++ b/vp8/common/arm/neon/dc_only_idct_add_neon.c
@@ -10,33 +10,30 @@
 
 #include <arm_neon.h>
 
-void vp8_dc_only_idct_add_neon(
-        int16_t input_dc,
-        unsigned char *pred_ptr,
-        int pred_stride,
-        unsigned char *dst_ptr,
-        int dst_stride) {
-    int i;
-    uint16_t a1 = ((input_dc + 4) >> 3);
-    uint32x2_t d2u32 = vdup_n_u32(0);
-    uint8x8_t d2u8;
-    uint16x8_t q1u16;
-    uint16x8_t qAdd;
+void vp8_dc_only_idct_add_neon(int16_t input_dc, unsigned char *pred_ptr,
+                               int pred_stride, unsigned char *dst_ptr,
+                               int dst_stride) {
+  int i;
+  uint16_t a1 = ((input_dc + 4) >> 3);
+  uint32x2_t d2u32 = vdup_n_u32(0);
+  uint8x8_t d2u8;
+  uint16x8_t q1u16;
+  uint16x8_t qAdd;
 
-    qAdd = vdupq_n_u16(a1);
+  qAdd = vdupq_n_u16(a1);
 
-    for (i = 0; i < 2; i++) {
-        d2u32 = vld1_lane_u32((const uint32_t *)pred_ptr, d2u32, 0);
-        pred_ptr += pred_stride;
-        d2u32 = vld1_lane_u32((const uint32_t *)pred_ptr, d2u32, 1);
-        pred_ptr += pred_stride;
+  for (i = 0; i < 2; i++) {
+    d2u32 = vld1_lane_u32((const uint32_t *)pred_ptr, d2u32, 0);
+    pred_ptr += pred_stride;
+    d2u32 = vld1_lane_u32((const uint32_t *)pred_ptr, d2u32, 1);
+    pred_ptr += pred_stride;
 
-        q1u16 = vaddw_u8(qAdd, vreinterpret_u8_u32(d2u32));
-        d2u8 = vqmovun_s16(vreinterpretq_s16_u16(q1u16));
+    q1u16 = vaddw_u8(qAdd, vreinterpret_u8_u32(d2u32));
+    d2u8 = vqmovun_s16(vreinterpretq_s16_u16(q1u16));
 
-        vst1_lane_u32((uint32_t *)dst_ptr, vreinterpret_u32_u8(d2u8), 0);
-        dst_ptr += dst_stride;
-        vst1_lane_u32((uint32_t *)dst_ptr, vreinterpret_u32_u8(d2u8), 1);
-        dst_ptr += dst_stride;
-    }
+    vst1_lane_u32((uint32_t *)dst_ptr, vreinterpret_u32_u8(d2u8), 0);
+    dst_ptr += dst_stride;
+    vst1_lane_u32((uint32_t *)dst_ptr, vreinterpret_u32_u8(d2u8), 1);
+    dst_ptr += dst_stride;
+  }
 }
--- a/vp8/common/arm/neon/dequant_idct_neon.c
+++ b/vp8/common/arm/neon/dequant_idct_neon.c
@@ -11,132 +11,129 @@
 #include <arm_neon.h>
 
 static const int16_t cospi8sqrt2minus1 = 20091;
-static const int16_t sinpi8sqrt2       = 35468;
+static const int16_t sinpi8sqrt2 = 35468;
 
-void vp8_dequant_idct_add_neon(
-        int16_t *input,
-        int16_t *dq,
-        unsigned char *dst,
-        int stride) {
-    unsigned char *dst0;
-    int32x2_t d14, d15;
-    int16x4_t d2, d3, d4, d5, d10, d11, d12, d13;
-    int16x8_t q1, q2, q3, q4, q5, q6;
-    int16x8_t qEmpty = vdupq_n_s16(0);
-    int32x2x2_t d2tmp0, d2tmp1;
-    int16x4x2_t d2tmp2, d2tmp3;
+void vp8_dequant_idct_add_neon(int16_t *input, int16_t *dq, unsigned char *dst,
+                               int stride) {
+  unsigned char *dst0;
+  int32x2_t d14, d15;
+  int16x4_t d2, d3, d4, d5, d10, d11, d12, d13;
+  int16x8_t q1, q2, q3, q4, q5, q6;
+  int16x8_t qEmpty = vdupq_n_s16(0);
+  int32x2x2_t d2tmp0, d2tmp1;
+  int16x4x2_t d2tmp2, d2tmp3;
 
-    d14 = d15 = vdup_n_s32(0);
+  d14 = d15 = vdup_n_s32(0);
 
-    // load input
-    q3 = vld1q_s16(input);
-    vst1q_s16(input, qEmpty);
-    input += 8;
-    q4 = vld1q_s16(input);
-    vst1q_s16(input, qEmpty);
+  // load input
+  q3 = vld1q_s16(input);
+  vst1q_s16(input, qEmpty);
+  input += 8;
+  q4 = vld1q_s16(input);
+  vst1q_s16(input, qEmpty);
 
-    // load dq
-    q5 = vld1q_s16(dq);
-    dq += 8;
-    q6 = vld1q_s16(dq);
+  // load dq
+  q5 = vld1q_s16(dq);
+  dq += 8;
+  q6 = vld1q_s16(dq);
 
-    // load src from dst
-    dst0 = dst;
-    d14 = vld1_lane_s32((const int32_t *)dst0, d14, 0);
-    dst0 += stride;
-    d14 = vld1_lane_s32((const int32_t *)dst0, d14, 1);
-    dst0 += stride;
-    d15 = vld1_lane_s32((const int32_t *)dst0, d15, 0);
-    dst0 += stride;
-    d15 = vld1_lane_s32((const int32_t *)dst0, d15, 1);
+  // load src from dst
+  dst0 = dst;
+  d14 = vld1_lane_s32((const int32_t *)dst0, d14, 0);
+  dst0 += stride;
+  d14 = vld1_lane_s32((const int32_t *)dst0, d14, 1);
+  dst0 += stride;
+  d15 = vld1_lane_s32((const int32_t *)dst0, d15, 0);
+  dst0 += stride;
+  d15 = vld1_lane_s32((const int32_t *)dst0, d15, 1);
 
-    q1 = vreinterpretq_s16_u16(vmulq_u16(vreinterpretq_u16_s16(q3),
-                                         vreinterpretq_u16_s16(q5)));
-    q2 = vreinterpretq_s16_u16(vmulq_u16(vreinterpretq_u16_s16(q4),
-                                         vreinterpretq_u16_s16(q6)));
+  q1 = vreinterpretq_s16_u16(
+      vmulq_u16(vreinterpretq_u16_s16(q3), vreinterpretq_u16_s16(q5)));
+  q2 = vreinterpretq_s16_u16(
+      vmulq_u16(vreinterpretq_u16_s16(q4), vreinterpretq_u16_s16(q6)));
 
-    d12 = vqadd_s16(vget_low_s16(q1), vget_low_s16(q2));
-    d13 = vqsub_s16(vget_low_s16(q1), vget_low_s16(q2));
+  d12 = vqadd_s16(vget_low_s16(q1), vget_low_s16(q2));
+  d13 = vqsub_s16(vget_low_s16(q1), vget_low_s16(q2));
 
-    q2 = vcombine_s16(vget_high_s16(q1), vget_high_s16(q2));
+  q2 = vcombine_s16(vget_high_s16(q1), vget_high_s16(q2));
 
-    q3 = vqdmulhq_n_s16(q2, sinpi8sqrt2);
-    q4 = vqdmulhq_n_s16(q2, cospi8sqrt2minus1);
+  q3 = vqdmulhq_n_s16(q2, sinpi8sqrt2);
+  q4 = vqdmulhq_n_s16(q2, cospi8sqrt2minus1);
 
-    q3 = vshrq_n_s16(q3, 1);
-    q4 = vshrq_n_s16(q4, 1);
+  q3 = vshrq_n_s16(q3, 1);
+  q4 = vshrq_n_s16(q4, 1);
 
-    q3 = vqaddq_s16(q3, q2);
-    q4 = vqaddq_s16(q4, q2);
+  q3 = vqaddq_s16(q3, q2);
+  q4 = vqaddq_s16(q4, q2);
 
-    d10 = vqsub_s16(vget_low_s16(q3), vget_high_s16(q4));
-    d11 = vqadd_s16(vget_high_s16(q3), vget_low_s16(q4));
+  d10 = vqsub_s16(vget_low_s16(q3), vget_high_s16(q4));
+  d11 = vqadd_s16(vget_high_s16(q3), vget_low_s16(q4));
 
-    d2 = vqadd_s16(d12, d11);
-    d3 = vqadd_s16(d13, d10);
-    d4 = vqsub_s16(d13, d10);
-    d5 = vqsub_s16(d12, d11);
+  d2 = vqadd_s16(d12, d11);
+  d3 = vqadd_s16(d13, d10);
+  d4 = vqsub_s16(d13, d10);
+  d5 = vqsub_s16(d12, d11);
 
-    d2tmp0 = vtrn_s32(vreinterpret_s32_s16(d2), vreinterpret_s32_s16(d4));
-    d2tmp1 = vtrn_s32(vreinterpret_s32_s16(d3), vreinterpret_s32_s16(d5));
-    d2tmp2 = vtrn_s16(vreinterpret_s16_s32(d2tmp0.val[0]),
-                      vreinterpret_s16_s32(d2tmp1.val[0]));
-    d2tmp3 = vtrn_s16(vreinterpret_s16_s32(d2tmp0.val[1]),
-                      vreinterpret_s16_s32(d2tmp1.val[1]));
+  d2tmp0 = vtrn_s32(vreinterpret_s32_s16(d2), vreinterpret_s32_s16(d4));
+  d2tmp1 = vtrn_s32(vreinterpret_s32_s16(d3), vreinterpret_s32_s16(d5));
+  d2tmp2 = vtrn_s16(vreinterpret_s16_s32(d2tmp0.val[0]),
+                    vreinterpret_s16_s32(d2tmp1.val[0]));
+  d2tmp3 = vtrn_s16(vreinterpret_s16_s32(d2tmp0.val[1]),
+                    vreinterpret_s16_s32(d2tmp1.val[1]));
 
-    // loop 2
-    q2 = vcombine_s16(d2tmp2.val[1], d2tmp3.val[1]);
+  // loop 2
+  q2 = vcombine_s16(d2tmp2.val[1], d2tmp3.val[1]);
 
-    q3 = vqdmulhq_n_s16(q2, sinpi8sqrt2);
-    q4 = vqdmulhq_n_s16(q2, cospi8sqrt2minus1);
+  q3 = vqdmulhq_n_s16(q2, sinpi8sqrt2);
+  q4 = vqdmulhq_n_s16(q2, cospi8sqrt2minus1);
 
-    d12 = vqadd_s16(d2tmp2.val[0], d2tmp3.val[0]);
-    d13 = vqsub_s16(d2tmp2.val[0], d2tmp3.val[0]);
+  d12 = vqadd_s16(d2tmp2.val[0], d2tmp3.val[0]);
+  d13 = vqsub_s16(d2tmp2.val[0], d2tmp3.val[0]);
 
-    q3 = vshrq_n_s16(q3, 1);
-    q4 = vshrq_n_s16(q4, 1);
+  q3 = vshrq_n_s16(q3, 1);
+  q4 = vshrq_n_s16(q4, 1);
 
-    q3 = vqaddq_s16(q3, q2);
-    q4 = vqaddq_s16(q4, q2);
+  q3 = vqaddq_s16(q3, q2);
+  q4 = vqaddq_s16(q4, q2);
 
-    d10 = vqsub_s16(vget_low_s16(q3), vget_high_s16(q4));
-    d11 = vqadd_s16(vget_high_s16(q3), vget_low_s16(q4));
+  d10 = vqsub_s16(vget_low_s16(q3), vget_high_s16(q4));
+  d11 = vqadd_s16(vget_high_s16(q3), vget_low_s16(q4));
 
-    d2 = vqadd_s16(d12, d11);
-    d3 = vqadd_s16(d13, d10);
-    d4 = vqsub_s16(d13, d10);
-    d5 = vqsub_s16(d12, d11);
+  d2 = vqadd_s16(d12, d11);
+  d3 = vqadd_s16(d13, d10);
+  d4 = vqsub_s16(d13, d10);
+  d5 = vqsub_s16(d12, d11);
 
-    d2 = vrshr_n_s16(d2, 3);
-    d3 = vrshr_n_s16(d3, 3);
-    d4 = vrshr_n_s16(d4, 3);
-    d5 = vrshr_n_s16(d5, 3);
+  d2 = vrshr_n_s16(d2, 3);
+  d3 = vrshr_n_s16(d3, 3);
+  d4 = vrshr_n_s16(d4, 3);
+  d5 = vrshr_n_s16(d5, 3);
 
-    d2tmp0 = vtrn_s32(vreinterpret_s32_s16(d2), vreinterpret_s32_s16(d4));
-    d2tmp1 = vtrn_s32(vreinterpret_s32_s16(d3), vreinterpret_s32_s16(d5));
-    d2tmp2 = vtrn_s16(vreinterpret_s16_s32(d2tmp0.val[0]),
-                      vreinterpret_s16_s32(d2tmp1.val[0]));
-    d2tmp3 = vtrn_s16(vreinterpret_s16_s32(d2tmp0.val[1]),
-                      vreinterpret_s16_s32(d2tmp1.val[1]));
+  d2tmp0 = vtrn_s32(vreinterpret_s32_s16(d2), vreinterpret_s32_s16(d4));
+  d2tmp1 = vtrn_s32(vreinterpret_s32_s16(d3), vreinterpret_s32_s16(d5));
+  d2tmp2 = vtrn_s16(vreinterpret_s16_s32(d2tmp0.val[0]),
+                    vreinterpret_s16_s32(d2tmp1.val[0]));
+  d2tmp3 = vtrn_s16(vreinterpret_s16_s32(d2tmp0.val[1]),
+                    vreinterpret_s16_s32(d2tmp1.val[1]));
 
-    q1 = vcombine_s16(d2tmp2.val[0], d2tmp2.val[1]);
-    q2 = vcombine_s16(d2tmp3.val[0], d2tmp3.val[1]);
+  q1 = vcombine_s16(d2tmp2.val[0], d2tmp2.val[1]);
+  q2 = vcombine_s16(d2tmp3.val[0], d2tmp3.val[1]);
 
-    q1 = vreinterpretq_s16_u16(vaddw_u8(vreinterpretq_u16_s16(q1),
-                                        vreinterpret_u8_s32(d14)));
-    q2 = vreinterpretq_s16_u16(vaddw_u8(vreinterpretq_u16_s16(q2),
-                                        vreinterpret_u8_s32(d15)));
+  q1 = vreinterpretq_s16_u16(
+      vaddw_u8(vreinterpretq_u16_s16(q1), vreinterpret_u8_s32(d14)));
+  q2 = vreinterpretq_s16_u16(
+      vaddw_u8(vreinterpretq_u16_s16(q2), vreinterpret_u8_s32(d15)));
 
-    d14 = vreinterpret_s32_u8(vqmovun_s16(q1));
-    d15 = vreinterpret_s32_u8(vqmovun_s16(q2));
+  d14 = vreinterpret_s32_u8(vqmovun_s16(q1));
+  d15 = vreinterpret_s32_u8(vqmovun_s16(q2));
 
-    dst0 = dst;
-    vst1_lane_s32((int32_t *)dst0, d14, 0);
-    dst0 += stride;
-    vst1_lane_s32((int32_t *)dst0, d14, 1);
-    dst0 += stride;
-    vst1_lane_s32((int32_t *)dst0, d15, 0);
-    dst0 += stride;
-    vst1_lane_s32((int32_t *)dst0, d15, 1);
-    return;
+  dst0 = dst;
+  vst1_lane_s32((int32_t *)dst0, d14, 0);
+  dst0 += stride;
+  vst1_lane_s32((int32_t *)dst0, d14, 1);
+  dst0 += stride;
+  vst1_lane_s32((int32_t *)dst0, d15, 0);
+  dst0 += stride;
+  vst1_lane_s32((int32_t *)dst0, d15, 1);
+  return;
 }
--- a/vp8/common/arm/neon/dequantizeb_neon.c
+++ b/vp8/common/arm/neon/dequantizeb_neon.c
@@ -13,13 +13,13 @@
 #include "vp8/common/blockd.h"
 
 void vp8_dequantize_b_neon(BLOCKD *d, short *DQC) {
-    int16x8x2_t qQ, qDQC, qDQ;
+  int16x8x2_t qQ, qDQC, qDQ;
 
-    qQ   = vld2q_s16(d->qcoeff);
-    qDQC = vld2q_s16(DQC);
+  qQ = vld2q_s16(d->qcoeff);
+  qDQC = vld2q_s16(DQC);
 
-    qDQ.val[0] = vmulq_s16(qQ.val[0], qDQC.val[0]);
-    qDQ.val[1] = vmulq_s16(qQ.val[1], qDQC.val[1]);
+  qDQ.val[0] = vmulq_s16(qQ.val[0], qDQC.val[0]);
+  qDQ.val[1] = vmulq_s16(qQ.val[1], qDQC.val[1]);
 
-    vst2q_s16(d->dqcoeff, qDQ);
+  vst2q_s16(d->dqcoeff, qDQ);
 }
--- a/vp8/common/arm/neon/idct_blk_neon.c
+++ b/vp8/common/arm/neon/idct_blk_neon.c
@@ -14,83 +14,71 @@
 /* place these declarations here because we don't want to maintain them
  * outside of this scope
  */
-void idct_dequant_full_2x_neon(short *q, short *dq,
-                               unsigned char *dst, int stride);
-void idct_dequant_0_2x_neon(short *q, short dq,
-                            unsigned char *dst, int stride);
+void idct_dequant_full_2x_neon(short *q, short *dq, unsigned char *dst,
+                               int stride);
+void idct_dequant_0_2x_neon(short *q, short dq, unsigned char *dst, int stride);
 
+void vp8_dequant_idct_add_y_block_neon(short *q, short *dq, unsigned char *dst,
+                                       int stride, char *eobs) {
+  int i;
 
-void vp8_dequant_idct_add_y_block_neon(short *q, short *dq,
-                                       unsigned char *dst,
-                                       int stride, char *eobs)
-{
-    int i;
+  for (i = 0; i < 4; i++) {
+    if (((short *)(eobs))[0]) {
+      if (((short *)eobs)[0] & 0xfefe)
+        idct_dequant_full_2x_neon(q, dq, dst, stride);
+      else
+        idct_dequant_0_2x_neon(q, dq[0], dst, stride);
+    }
 
-    for (i = 0; i < 4; i++)
-    {
-        if (((short *)(eobs))[0])
-        {
-            if (((short *)eobs)[0] & 0xfefe)
-                idct_dequant_full_2x_neon (q, dq, dst, stride);
-            else
-                idct_dequant_0_2x_neon (q, dq[0], dst, stride);
-        }
-
-        if (((short *)(eobs))[1])
-        {
-            if (((short *)eobs)[1] & 0xfefe)
-                idct_dequant_full_2x_neon (q+32, dq, dst+8, stride);
-            else
-                idct_dequant_0_2x_neon (q+32, dq[0], dst+8, stride);
-        }
-        q    += 64;
-        dst  += 4*stride;
-        eobs += 4;
+    if (((short *)(eobs))[1]) {
+      if (((short *)eobs)[1] & 0xfefe)
+        idct_dequant_full_2x_neon(q + 32, dq, dst + 8, stride);
+      else
+        idct_dequant_0_2x_neon(q + 32, dq[0], dst + 8, stride);
     }
+    q += 64;
+    dst += 4 * stride;
+    eobs += 4;
+  }
 }
 
 void vp8_dequant_idct_add_uv_block_neon(short *q, short *dq,
                                         unsigned char *dstu,
-                                        unsigned char *dstv,
-                                        int stride, char *eobs)
-{
-    if (((short *)(eobs))[0])
-    {
-        if (((short *)eobs)[0] & 0xfefe)
-            idct_dequant_full_2x_neon (q, dq, dstu, stride);
-        else
-            idct_dequant_0_2x_neon (q, dq[0], dstu, stride);
-    }
+                                        unsigned char *dstv, int stride,
+                                        char *eobs) {
+  if (((short *)(eobs))[0]) {
+    if (((short *)eobs)[0] & 0xfefe)
+      idct_dequant_full_2x_neon(q, dq, dstu, stride);
+    else
+      idct_dequant_0_2x_neon(q, dq[0], dstu, stride);
+  }
 
-    q    += 32;
-    dstu += 4*stride;
+  q += 32;
+  dstu += 4 * stride;
 
-    if (((short *)(eobs))[1])
-    {
-        if (((short *)eobs)[1] & 0xfefe)
-            idct_dequant_full_2x_neon (q, dq, dstu, stride);
-        else
-            idct_dequant_0_2x_neon (q, dq[0], dstu, stride);
-    }
+  if (((short *)(eobs))[1]) {
+    if (((short *)eobs)[1] & 0xfefe)
+      idct_dequant_full_2x_neon(q, dq, dstu, stride);
+    else
+      idct_dequant_0_2x_neon(q, dq[0], dstu, stride);
+  }
 
-    q += 32;
+  q += 32;
 
-    if (((short *)(eobs))[2])
-    {
-        if (((short *)eobs)[2] & 0xfefe)
-            idct_dequant_full_2x_neon (q, dq, dstv, stride);
-        else
-            idct_dequant_0_2x_neon (q, dq[0], dstv, stride);
-    }
+  if (((short *)(eobs))[2]) {
+    if (((short *)eobs)[2] & 0xfefe)
+      idct_dequant_full_2x_neon(q, dq, dstv, stride);
+    else
+      idct_dequant_0_2x_neon(q, dq[0], dstv, stride);
+  }
 
-    q    += 32;
-    dstv += 4*stride;
+  q += 32;
+  dstv += 4 * stride;
 
-    if (((short *)(eobs))[3])
-    {
-        if (((short *)eobs)[3] & 0xfefe)
-            idct_dequant_full_2x_neon (q, dq, dstv, stride);
-        else
-            idct_dequant_0_2x_neon (q, dq[0], dstv, stride);
-    }
+  if (((short *)(eobs))[3]) {
+    if (((short *)eobs)[3] & 0xfefe)
+      idct_dequant_full_2x_neon(q, dq, dstv, stride);
+    else
+      idct_dequant_0_2x_neon(q, dq[0], dstv, stride);
+  }
 }
--- a/vp8/common/arm/neon/idct_dequant_0_2x_neon.c
+++ b/vp8/common/arm/neon/idct_dequant_0_2x_neon.c
@@ -10,54 +10,50 @@
 
 #include <arm_neon.h>
 
-void idct_dequant_0_2x_neon(
-        int16_t *q,
-        int16_t dq,
-        unsigned char *dst,
-        int stride) {
-    unsigned char *dst0;
-    int i, a0, a1;
-    int16x8x2_t q2Add;
-    int32x2_t d2s32 = vdup_n_s32(0),
-              d4s32 = vdup_n_s32(0);
-    uint8x8_t d2u8, d4u8;
-    uint16x8_t q1u16, q2u16;
+void idct_dequant_0_2x_neon(int16_t *q, int16_t dq, unsigned char *dst,
+                            int stride) {
+  unsigned char *dst0;
+  int i, a0, a1;
+  int16x8x2_t q2Add;
+  int32x2_t d2s32 = vdup_n_s32(0), d4s32 = vdup_n_s32(0);
+  uint8x8_t d2u8, d4u8;
+  uint16x8_t q1u16, q2u16;
 
-    a0 = ((q[0] * dq) + 4) >> 3;
-    a1 = ((q[16] * dq) + 4) >> 3;
-    q[0] = q[16] = 0;
-    q2Add.val[0] = vdupq_n_s16((int16_t)a0);
-    q2Add.val[1] = vdupq_n_s16((int16_t)a1);
+  a0 = ((q[0] * dq) + 4) >> 3;
+  a1 = ((q[16] * dq) + 4) >> 3;
+  q[0] = q[16] = 0;
+  q2Add.val[0] = vdupq_n_s16((int16_t)a0);
+  q2Add.val[1] = vdupq_n_s16((int16_t)a1);
 
-    for (i = 0; i < 2; i++, dst += 4) {
-        dst0 = dst;
-        d2s32 = vld1_lane_s32((const int32_t *)dst0, d2s32, 0);
-        dst0 += stride;
-        d2s32 = vld1_lane_s32((const int32_t *)dst0, d2s32, 1);
-        dst0 += stride;
-        d4s32 = vld1_lane_s32((const int32_t *)dst0, d4s32, 0);
-        dst0 += stride;
-        d4s32 = vld1_lane_s32((const int32_t *)dst0, d4s32, 1);
+  for (i = 0; i < 2; i++, dst += 4) {
+    dst0 = dst;
+    d2s32 = vld1_lane_s32((const int32_t *)dst0, d2s32, 0);
+    dst0 += stride;
+    d2s32 = vld1_lane_s32((const int32_t *)dst0, d2s32, 1);
+    dst0 += stride;
+    d4s32 = vld1_lane_s32((const int32_t *)dst0, d4s32, 0);
+    dst0 += stride;
+    d4s32 = vld1_lane_s32((const int32_t *)dst0, d4s32, 1);
 
-        q1u16 = vaddw_u8(vreinterpretq_u16_s16(q2Add.val[i]),
-                         vreinterpret_u8_s32(d2s32));
-        q2u16 = vaddw_u8(vreinterpretq_u16_s16(q2Add.val[i]),
-                         vreinterpret_u8_s32(d4s32));
+    q1u16 = vaddw_u8(vreinterpretq_u16_s16(q2Add.val[i]),
+                     vreinterpret_u8_s32(d2s32));
+    q2u16 = vaddw_u8(vreinterpretq_u16_s16(q2Add.val[i]),
+                     vreinterpret_u8_s32(d4s32));
 
-        d2u8 = vqmovun_s16(vreinterpretq_s16_u16(q1u16));
-        d4u8 = vqmovun_s16(vreinterpretq_s16_u16(q2u16));
+    d2u8 = vqmovun_s16(vreinterpretq_s16_u16(q1u16));
+    d4u8 = vqmovun_s16(vreinterpretq_s16_u16(q2u16));
 
-        d2s32 = vreinterpret_s32_u8(d2u8);
-        d4s32 = vreinterpret_s32_u8(d4u8);
+    d2s32 = vreinterpret_s32_u8(d2u8);
+    d4s32 = vreinterpret_s32_u8(d4u8);
 
-        dst0 = dst;
-        vst1_lane_s32((int32_t *)dst0, d2s32, 0);
-        dst0 += stride;
-        vst1_lane_s32((int32_t *)dst0, d2s32, 1);
-        dst0 += stride;
-        vst1_lane_s32((int32_t *)dst0, d4s32, 0);
-        dst0 += stride;
-        vst1_lane_s32((int32_t *)dst0, d4s32, 1);
-    }
-    return;
+    dst0 = dst;
+    vst1_lane_s32((int32_t *)dst0, d2s32, 0);
+    dst0 += stride;
+    vst1_lane_s32((int32_t *)dst0, d2s32, 1);
+    dst0 += stride;
+    vst1_lane_s32((int32_t *)dst0, d4s32, 0);
+    dst0 += stride;
+    vst1_lane_s32((int32_t *)dst0, d4s32, 1);
+  }
+  return;
 }
--- a/vp8/common/arm/neon/idct_dequant_full_2x_neon.c
+++ b/vp8/common/arm/neon/idct_dequant_full_2x_neon.c
@@ -11,175 +11,172 @@
 #include <arm_neon.h>
 
 static const int16_t cospi8sqrt2minus1 = 20091;
-static const int16_t sinpi8sqrt2       = 17734;
+static const int16_t sinpi8sqrt2 = 17734;
 // because the lowest bit in 0x8a8c is 0, we can pre-shift this
 
-void idct_dequant_full_2x_neon(
-        int16_t *q,
-        int16_t *dq,
-        unsigned char *dst,
-        int stride) {
-    unsigned char *dst0, *dst1;
-    int32x2_t d28, d29, d30, d31;
-    int16x8_t q0, q1, q2, q3, q4, q5, q6, q7, q8, q9, q10, q11;
-    int16x8_t qEmpty = vdupq_n_s16(0);
-    int32x4x2_t q2tmp0, q2tmp1;
-    int16x8x2_t q2tmp2, q2tmp3;
-    int16x4_t dLow0, dLow1, dHigh0, dHigh1;
+void idct_dequant_full_2x_neon(int16_t *q, int16_t *dq, unsigned char *dst,
+                               int stride) {
+  unsigned char *dst0, *dst1;
+  int32x2_t d28, d29, d30, d31;
+  int16x8_t q0, q1, q2, q3, q4, q5, q6, q7, q8, q9, q10, q11;
+  int16x8_t qEmpty = vdupq_n_s16(0);
+  int32x4x2_t q2tmp0, q2tmp1;
+  int16x8x2_t q2tmp2, q2tmp3;
+  int16x4_t dLow0, dLow1, dHigh0, dHigh1;
 
-    d28 = d29 = d30 = d31 = vdup_n_s32(0);
+  d28 = d29 = d30 = d31 = vdup_n_s32(0);
 
-    // load dq
-    q0 = vld1q_s16(dq);
-    dq += 8;
-    q1 = vld1q_s16(dq);
+  // load dq
+  q0 = vld1q_s16(dq);
+  dq += 8;
+  q1 = vld1q_s16(dq);
 
-    // load q
-    q2 = vld1q_s16(q);
-    vst1q_s16(q, qEmpty);
-    q += 8;
-    q3 = vld1q_s16(q);
-    vst1q_s16(q, qEmpty);
-    q += 8;
-    q4 = vld1q_s16(q);
-    vst1q_s16(q, qEmpty);
-    q += 8;
-    q5 = vld1q_s16(q);
-    vst1q_s16(q, qEmpty);
+  // load q
+  q2 = vld1q_s16(q);
+  vst1q_s16(q, qEmpty);
+  q += 8;
+  q3 = vld1q_s16(q);
+  vst1q_s16(q, qEmpty);
+  q += 8;
+  q4 = vld1q_s16(q);
+  vst1q_s16(q, qEmpty);
+  q += 8;
+  q5 = vld1q_s16(q);
+  vst1q_s16(q, qEmpty);
 
-    // load src from dst
-    dst0 = dst;
-    dst1 = dst + 4;
-    d28 = vld1_lane_s32((const int32_t *)dst0, d28, 0);
-    dst0 += stride;
-    d28 = vld1_lane_s32((const int32_t *)dst1, d28, 1);
-    dst1 += stride;
-    d29 = vld1_lane_s32((const int32_t *)dst0, d29, 0);
-    dst0 += stride;
-    d29 = vld1_lane_s32((const int32_t *)dst1, d29, 1);
-    dst1 += stride;
+  // load src from dst
+  dst0 = dst;
+  dst1 = dst + 4;
+  d28 = vld1_lane_s32((const int32_t *)dst0, d28, 0);
+  dst0 += stride;
+  d28 = vld1_lane_s32((const int32_t *)dst1, d28, 1);
+  dst1 += stride;
+  d29 = vld1_lane_s32((const int32_t *)dst0, d29, 0);
+  dst0 += stride;
+  d29 = vld1_lane_s32((const int32_t *)dst1, d29, 1);
+  dst1 += stride;
 
-    d30 = vld1_lane_s32((const int32_t *)dst0, d30, 0);
-    dst0 += stride;
-    d30 = vld1_lane_s32((const int32_t *)dst1, d30, 1);
-    dst1 += stride;
-    d31 = vld1_lane_s32((const int32_t *)dst0, d31, 0);
-    d31 = vld1_lane_s32((const int32_t *)dst1, d31, 1);
+  d30 = vld1_lane_s32((const int32_t *)dst0, d30, 0);
+  dst0 += stride;
+  d30 = vld1_lane_s32((const int32_t *)dst1, d30, 1);
+  dst1 += stride;
+  d31 = vld1_lane_s32((const int32_t *)dst0, d31, 0);
+  d31 = vld1_lane_s32((const int32_t *)dst1, d31, 1);
 
-    q2 = vmulq_s16(q2, q0);
-    q3 = vmulq_s16(q3, q1);
-    q4 = vmulq_s16(q4, q0);
-    q5 = vmulq_s16(q5, q1);
+  q2 = vmulq_s16(q2, q0);
+  q3 = vmulq_s16(q3, q1);
+  q4 = vmulq_s16(q4, q0);
+  q5 = vmulq_s16(q5, q1);
 
-    // vswp
-    dLow0 = vget_low_s16(q2);
-    dHigh0 = vget_high_s16(q2);
-    dLow1 = vget_low_s16(q4);
-    dHigh1 = vget_high_s16(q4);
-    q2 = vcombine_s16(dLow0, dLow1);
-    q4 = vcombine_s16(dHigh0, dHigh1);
+  // vswp
+  dLow0 = vget_low_s16(q2);
+  dHigh0 = vget_high_s16(q2);
+  dLow1 = vget_low_s16(q4);
+  dHigh1 = vget_high_s16(q4);
+  q2 = vcombine_s16(dLow0, dLow1);
+  q4 = vcombine_s16(dHigh0, dHigh1);
 
-    dLow0 = vget_low_s16(q3);
-    dHigh0 = vget_high_s16(q3);
-    dLow1 = vget_low_s16(q5);
-    dHigh1 = vget_high_s16(q5);
-    q3 = vcombine_s16(dLow0, dLow1);
-    q5 = vcombine_s16(dHigh0, dHigh1);
+  dLow0 = vget_low_s16(q3);
+  dHigh0 = vget_high_s16(q3);
+  dLow1 = vget_low_s16(q5);
+  dHigh1 = vget_high_s16(q5);
+  q3 = vcombine_s16(dLow0, dLow1);
+  q5 = vcombine_s16(dHigh0, dHigh1);
 
-    q6 = vqdmulhq_n_s16(q4, sinpi8sqrt2);
-    q7 = vqdmulhq_n_s16(q5, sinpi8sqrt2);
-    q8 = vqdmulhq_n_s16(q4, cospi8sqrt2minus1);
-    q9 = vqdmulhq_n_s16(q5, cospi8sqrt2minus1);
+  q6 = vqdmulhq_n_s16(q4, sinpi8sqrt2);
+  q7 = vqdmulhq_n_s16(q5, sinpi8sqrt2);
+  q8 = vqdmulhq_n_s16(q4, cospi8sqrt2minus1);
+  q9 = vqdmulhq_n_s16(q5, cospi8sqrt2minus1);
 
-    q10 = vqaddq_s16(q2, q3);
-    q11 = vqsubq_s16(q2, q3);
+  q10 = vqaddq_s16(q2, q3);
+  q11 = vqsubq_s16(q2, q3);
 
-    q8 = vshrq_n_s16(q8, 1);
-    q9 = vshrq_n_s16(q9, 1);
+  q8 = vshrq_n_s16(q8, 1);
+  q9 = vshrq_n_s16(q9, 1);
 
-    q4 = vqaddq_s16(q4, q8);
-    q5 = vqaddq_s16(q5, q9);
+  q4 = vqaddq_s16(q4, q8);
+  q5 = vqaddq_s16(q5, q9);
 
-    q2 = vqsubq_s16(q6, q5);
-    q3 = vqaddq_s16(q7, q4);
+  q2 = vqsubq_s16(q6, q5);
+  q3 = vqaddq_s16(q7, q4);
 
-    q4 = vqaddq_s16(q10, q3);
-    q5 = vqaddq_s16(q11, q2);
-    q6 = vqsubq_s16(q11, q2);
-    q7 = vqsubq_s16(q10, q3);
+  q4 = vqaddq_s16(q10, q3);
+  q5 = vqaddq_s16(q11, q2);
+  q6 = vqsubq_s16(q11, q2);
+  q7 = vqsubq_s16(q10, q3);
 
-    q2tmp0 = vtrnq_s32(vreinterpretq_s32_s16(q4), vreinterpretq_s32_s16(q6));
-    q2tmp1 = vtrnq_s32(vreinterpretq_s32_s16(q5), vreinterpretq_s32_s16(q7));
-    q2tmp2 = vtrnq_s16(vreinterpretq_s16_s32(q2tmp0.val[0]),
-                       vreinterpretq_s16_s32(q2tmp1.val[0]));
-    q2tmp3 = vtrnq_s16(vreinterpretq_s16_s32(q2tmp0.val[1]),
-                       vreinterpretq_s16_s32(q2tmp1.val[1]));
+  q2tmp0 = vtrnq_s32(vreinterpretq_s32_s16(q4), vreinterpretq_s32_s16(q6));
+  q2tmp1 = vtrnq_s32(vreinterpretq_s32_s16(q5), vreinterpretq_s32_s16(q7));
+  q2tmp2 = vtrnq_s16(vreinterpretq_s16_s32(q2tmp0.val[0]),
+                     vreinterpretq_s16_s32(q2tmp1.val[0]));
+  q2tmp3 = vtrnq_s16(vreinterpretq_s16_s32(q2tmp0.val[1]),
+                     vreinterpretq_s16_s32(q2tmp1.val[1]));
 
-    // loop 2
-    q8  = vqdmulhq_n_s16(q2tmp2.val[1], sinpi8sqrt2);
-    q9  = vqdmulhq_n_s16(q2tmp3.val[1], sinpi8sqrt2);
-    q10 = vqdmulhq_n_s16(q2tmp2.val[1], cospi8sqrt2minus1);
-    q11 = vqdmulhq_n_s16(q2tmp3.val[1], cospi8sqrt2minus1);
+  // loop 2
+  q8 = vqdmulhq_n_s16(q2tmp2.val[1], sinpi8sqrt2);
+  q9 = vqdmulhq_n_s16(q2tmp3.val[1], sinpi8sqrt2);
+  q10 = vqdmulhq_n_s16(q2tmp2.val[1], cospi8sqrt2minus1);
+  q11 = vqdmulhq_n_s16(q2tmp3.val[1], cospi8sqrt2minus1);
 
-    q2 = vqaddq_s16(q2tmp2.val[0], q2tmp3.val[0]);
-    q3 = vqsubq_s16(q2tmp2.val[0], q2tmp3.val[0]);
+  q2 = vqaddq_s16(q2tmp2.val[0], q2tmp3.val[0]);
+  q3 = vqsubq_s16(q2tmp2.val[0], q2tmp3.val[0]);
 
-    q10 = vshrq_n_s16(q10, 1);
-    q11 = vshrq_n_s16(q11, 1);
+  q10 = vshrq_n_s16(q10, 1);
+  q11 = vshrq_n_s16(q11, 1);
 
-    q10 = vqaddq_s16(q2tmp2.val[1], q10);
-    q11 = vqaddq_s16(q2tmp3.val[1], q11);
+  q10 = vqaddq_s16(q2tmp2.val[1], q10);
+  q11 = vqaddq_s16(q2tmp3.val[1], q11);
 
-    q8 = vqsubq_s16(q8, q11);
-    q9 = vqaddq_s16(q9, q10);
+  q8 = vqsubq_s16(q8, q11);
+  q9 = vqaddq_s16(q9, q10);
 
-    q4 = vqaddq_s16(q2, q9);
-    q5 = vqaddq_s16(q3, q8);
-    q6 = vqsubq_s16(q3, q8);
-    q7 = vqsubq_s16(q2, q9);
+  q4 = vqaddq_s16(q2, q9);
+  q5 = vqaddq_s16(q3, q8);
+  q6 = vqsubq_s16(q3, q8);
+  q7 = vqsubq_s16(q2, q9);
 
-    q4 = vrshrq_n_s16(q4, 3);
-    q5 = vrshrq_n_s16(q5, 3);
-    q6 = vrshrq_n_s16(q6, 3);
-    q7 = vrshrq_n_s16(q7, 3);
+  q4 = vrshrq_n_s16(q4, 3);
+  q5 = vrshrq_n_s16(q5, 3);
+  q6 = vrshrq_n_s16(q6, 3);
+  q7 = vrshrq_n_s16(q7, 3);
 
-    q2tmp0 = vtrnq_s32(vreinterpretq_s32_s16(q4), vreinterpretq_s32_s16(q6));
-    q2tmp1 = vtrnq_s32(vreinterpretq_s32_s16(q5), vreinterpretq_s32_s16(q7));
-    q2tmp2 = vtrnq_s16(vreinterpretq_s16_s32(q2tmp0.val[0]),
-                       vreinterpretq_s16_s32(q2tmp1.val[0]));
-    q2tmp3 = vtrnq_s16(vreinterpretq_s16_s32(q2tmp0.val[1]),
-                       vreinterpretq_s16_s32(q2tmp1.val[1]));
+  q2tmp0 = vtrnq_s32(vreinterpretq_s32_s16(q4), vreinterpretq_s32_s16(q6));
+  q2tmp1 = vtrnq_s32(vreinterpretq_s32_s16(q5), vreinterpretq_s32_s16(q7));
+  q2tmp2 = vtrnq_s16(vreinterpretq_s16_s32(q2tmp0.val[0]),
+                     vreinterpretq_s16_s32(q2tmp1.val[0]));
+  q2tmp3 = vtrnq_s16(vreinterpretq_s16_s32(q2tmp0.val[1]),
+                     vreinterpretq_s16_s32(q2tmp1.val[1]));
 
-    q4 = vreinterpretq_s16_u16(vaddw_u8(vreinterpretq_u16_s16(q2tmp2.val[0]),
-                                          vreinterpret_u8_s32(d28)));
-    q5 = vreinterpretq_s16_u16(vaddw_u8(vreinterpretq_u16_s16(q2tmp2.val[1]),
-                                          vreinterpret_u8_s32(d29)));
-    q6 = vreinterpretq_s16_u16(vaddw_u8(vreinterpretq_u16_s16(q2tmp3.val[0]),
-                                          vreinterpret_u8_s32(d30)));
-    q7 = vreinterpretq_s16_u16(vaddw_u8(vreinterpretq_u16_s16(q2tmp3.val[1]),
-                                          vreinterpret_u8_s32(d31)));
+  q4 = vreinterpretq_s16_u16(
+      vaddw_u8(vreinterpretq_u16_s16(q2tmp2.val[0]), vreinterpret_u8_s32(d28)));
+  q5 = vreinterpretq_s16_u16(
+      vaddw_u8(vreinterpretq_u16_s16(q2tmp2.val[1]), vreinterpret_u8_s32(d29)));
+  q6 = vreinterpretq_s16_u16(
+      vaddw_u8(vreinterpretq_u16_s16(q2tmp3.val[0]), vreinterpret_u8_s32(d30)));
+  q7 = vreinterpretq_s16_u16(
+      vaddw_u8(vreinterpretq_u16_s16(q2tmp3.val[1]), vreinterpret_u8_s32(d31)));
 
-    d28 = vreinterpret_s32_u8(vqmovun_s16(q4));
-    d29 = vreinterpret_s32_u8(vqmovun_s16(q5));
-    d30 = vreinterpret_s32_u8(vqmovun_s16(q6));
-    d31 = vreinterpret_s32_u8(vqmovun_s16(q7));
+  d28 = vreinterpret_s32_u8(vqmovun_s16(q4));
+  d29 = vreinterpret_s32_u8(vqmovun_s16(q5));
+  d30 = vreinterpret_s32_u8(vqmovun_s16(q6));
+  d31 = vreinterpret_s32_u8(vqmovun_s16(q7));
 
-    dst0 = dst;
-    dst1 = dst + 4;
-    vst1_lane_s32((int32_t *)dst0, d28, 0);
-    dst0 += stride;
-    vst1_lane_s32((int32_t *)dst1, d28, 1);
-    dst1 += stride;
-    vst1_lane_s32((int32_t *)dst0, d29, 0);
-    dst0 += stride;
-    vst1_lane_s32((int32_t *)dst1, d29, 1);
-    dst1 += stride;
+  dst0 = dst;
+  dst1 = dst + 4;
+  vst1_lane_s32((int32_t *)dst0, d28, 0);
+  dst0 += stride;
+  vst1_lane_s32((int32_t *)dst1, d28, 1);
+  dst1 += stride;
+  vst1_lane_s32((int32_t *)dst0, d29, 0);
+  dst0 += stride;
+  vst1_lane_s32((int32_t *)dst1, d29, 1);
+  dst1 += stride;
 
-    vst1_lane_s32((int32_t *)dst0, d30, 0);
-    dst0 += stride;
-    vst1_lane_s32((int32_t *)dst1, d30, 1);
-    dst1 += stride;
-    vst1_lane_s32((int32_t *)dst0, d31, 0);
-    vst1_lane_s32((int32_t *)dst1, d31, 1);
-    return;
+  vst1_lane_s32((int32_t *)dst0, d30, 0);
+  dst0 += stride;
+  vst1_lane_s32((int32_t *)dst1, d30, 1);
+  dst1 += stride;
+  vst1_lane_s32((int32_t *)dst0, d31, 0);
+  vst1_lane_s32((int32_t *)dst1, d31, 1);
+  return;
 }
--- a/vp8/common/arm/neon/iwalsh_neon.c
+++ b/vp8/common/arm/neon/iwalsh_neon.c
@@ -10,93 +10,91 @@
 
 #include <arm_neon.h>
 
-void vp8_short_inv_walsh4x4_neon(
-        int16_t *input,
-        int16_t *mb_dqcoeff) {
-    int16x8_t q0s16, q1s16, q2s16, q3s16;
-    int16x4_t d4s16, d5s16, d6s16, d7s16;
-    int16x4x2_t v2tmp0, v2tmp1;
-    int32x2x2_t v2tmp2, v2tmp3;
-    int16x8_t qAdd3;
+void vp8_short_inv_walsh4x4_neon(int16_t *input, int16_t *mb_dqcoeff) {
+  int16x8_t q0s16, q1s16, q2s16, q3s16;
+  int16x4_t d4s16, d5s16, d6s16, d7s16;
+  int16x4x2_t v2tmp0, v2tmp1;
+  int32x2x2_t v2tmp2, v2tmp3;
+  int16x8_t qAdd3;
 
-    q0s16 = vld1q_s16(input);
-    q1s16 = vld1q_s16(input + 8);
+  q0s16 = vld1q_s16(input);
+  q1s16 = vld1q_s16(input + 8);
 
-    // 1st for loop
-    d4s16 = vadd_s16(vget_low_s16(q0s16), vget_high_s16(q1s16));
-    d6s16 = vadd_s16(vget_high_s16(q0s16), vget_low_s16(q1s16));
-    d5s16 = vsub_s16(vget_low_s16(q0s16), vget_high_s16(q1s16));
-    d7s16 = vsub_s16(vget_high_s16(q0s16), vget_low_s16(q1s16));
+  // 1st for loop
+  d4s16 = vadd_s16(vget_low_s16(q0s16), vget_high_s16(q1s16));
+  d6s16 = vadd_s16(vget_high_s16(q0s16), vget_low_s16(q1s16));
+  d5s16 = vsub_s16(vget_low_s16(q0s16), vget_high_s16(q1s16));
+  d7s16 = vsub_s16(vget_high_s16(q0s16), vget_low_s16(q1s16));
 
-    q2s16 = vcombine_s16(d4s16, d5s16);
-    q3s16 = vcombine_s16(d6s16, d7s16);
+  q2s16 = vcombine_s16(d4s16, d5s16);
+  q3s16 = vcombine_s16(d6s16, d7s16);
 
-    q0s16 = vaddq_s16(q2s16, q3s16);
-    q1s16 = vsubq_s16(q2s16, q3s16);
+  q0s16 = vaddq_s16(q2s16, q3s16);
+  q1s16 = vsubq_s16(q2s16, q3s16);
 
-    v2tmp2 = vtrn_s32(vreinterpret_s32_s16(vget_low_s16(q0s16)),
-                      vreinterpret_s32_s16(vget_low_s16(q1s16)));
-    v2tmp3 = vtrn_s32(vreinterpret_s32_s16(vget_high_s16(q0s16)),
-                      vreinterpret_s32_s16(vget_high_s16(q1s16)));
-    v2tmp0 = vtrn_s16(vreinterpret_s16_s32(v2tmp2.val[0]),
-                      vreinterpret_s16_s32(v2tmp3.val[0]));
-    v2tmp1 = vtrn_s16(vreinterpret_s16_s32(v2tmp2.val[1]),
-                      vreinterpret_s16_s32(v2tmp3.val[1]));
+  v2tmp2 = vtrn_s32(vreinterpret_s32_s16(vget_low_s16(q0s16)),
+                    vreinterpret_s32_s16(vget_low_s16(q1s16)));
+  v2tmp3 = vtrn_s32(vreinterpret_s32_s16(vget_high_s16(q0s16)),
+                    vreinterpret_s32_s16(vget_high_s16(q1s16)));
+  v2tmp0 = vtrn_s16(vreinterpret_s16_s32(v2tmp2.val[0]),
+                    vreinterpret_s16_s32(v2tmp3.val[0]));
+  v2tmp1 = vtrn_s16(vreinterpret_s16_s32(v2tmp2.val[1]),
+                    vreinterpret_s16_s32(v2tmp3.val[1]));
 
-    // 2nd for loop
-    d4s16 = vadd_s16(v2tmp0.val[0], v2tmp1.val[1]);
-    d6s16 = vadd_s16(v2tmp0.val[1], v2tmp1.val[0]);
-    d5s16 = vsub_s16(v2tmp0.val[0], v2tmp1.val[1]);
-    d7s16 = vsub_s16(v2tmp0.val[1], v2tmp1.val[0]);
-    q2s16 = vcombine_s16(d4s16, d5s16);
-    q3s16 = vcombine_s16(d6s16, d7s16);
+  // 2nd for loop
+  d4s16 = vadd_s16(v2tmp0.val[0], v2tmp1.val[1]);
+  d6s16 = vadd_s16(v2tmp0.val[1], v2tmp1.val[0]);
+  d5s16 = vsub_s16(v2tmp0.val[0], v2tmp1.val[1]);
+  d7s16 = vsub_s16(v2tmp0.val[1], v2tmp1.val[0]);
+  q2s16 = vcombine_s16(d4s16, d5s16);
+  q3s16 = vcombine_s16(d6s16, d7s16);
 
-    qAdd3 = vdupq_n_s16(3);
+  qAdd3 = vdupq_n_s16(3);
 
-    q0s16 = vaddq_s16(q2s16, q3s16);
-    q1s16 = vsubq_s16(q2s16, q3s16);
+  q0s16 = vaddq_s16(q2s16, q3s16);
+  q1s16 = vsubq_s16(q2s16, q3s16);
 
-    q0s16 = vaddq_s16(q0s16, qAdd3);
-    q1s16 = vaddq_s16(q1s16, qAdd3);
+  q0s16 = vaddq_s16(q0s16, qAdd3);
+  q1s16 = vaddq_s16(q1s16, qAdd3);
 
-    q0s16 = vshrq_n_s16(q0s16, 3);
-    q1s16 = vshrq_n_s16(q1s16, 3);
+  q0s16 = vshrq_n_s16(q0s16, 3);
+  q1s16 = vshrq_n_s16(q1s16, 3);
 
-    // store
-    vst1_lane_s16(mb_dqcoeff, vget_low_s16(q0s16),  0);
-    mb_dqcoeff += 16;
-    vst1_lane_s16(mb_dqcoeff, vget_high_s16(q0s16), 0);
-    mb_dqcoeff += 16;
-    vst1_lane_s16(mb_dqcoeff, vget_low_s16(q1s16),  0);
-    mb_dqcoeff += 16;
-    vst1_lane_s16(mb_dqcoeff, vget_high_s16(q1s16), 0);
-    mb_dqcoeff += 16;
+  // store
+  vst1_lane_s16(mb_dqcoeff, vget_low_s16(q0s16), 0);
+  mb_dqcoeff += 16;
+  vst1_lane_s16(mb_dqcoeff, vget_high_s16(q0s16), 0);
+  mb_dqcoeff += 16;
+  vst1_lane_s16(mb_dqcoeff, vget_low_s16(q1s16), 0);
+  mb_dqcoeff += 16;
+  vst1_lane_s16(mb_dqcoeff, vget_high_s16(q1s16), 0);
+  mb_dqcoeff += 16;
 
-    vst1_lane_s16(mb_dqcoeff, vget_low_s16(q0s16),  1);
-    mb_dqcoeff += 16;
-    vst1_lane_s16(mb_dqcoeff, vget_high_s16(q0s16), 1);
-    mb_dqcoeff += 16;
-    vst1_lane_s16(mb_dqcoeff, vget_low_s16(q1s16),  1);
-    mb_dqcoeff += 16;
-    vst1_lane_s16(mb_dqcoeff, vget_high_s16(q1s16), 1);
-    mb_dqcoeff += 16;
+  vst1_lane_s16(mb_dqcoeff, vget_low_s16(q0s16), 1);
+  mb_dqcoeff += 16;
+  vst1_lane_s16(mb_dqcoeff, vget_high_s16(q0s16), 1);
+  mb_dqcoeff += 16;
+  vst1_lane_s16(mb_dqcoeff, vget_low_s16(q1s16), 1);
+  mb_dqcoeff += 16;
+  vst1_lane_s16(mb_dqcoeff, vget_high_s16(q1s16), 1);
+  mb_dqcoeff += 16;
 
-    vst1_lane_s16(mb_dqcoeff, vget_low_s16(q0s16),  2);
-    mb_dqcoeff += 16;
-    vst1_lane_s16(mb_dqcoeff, vget_high_s16(q0s16), 2);
-    mb_dqcoeff += 16;
-    vst1_lane_s16(mb_dqcoeff, vget_low_s16(q1s16),  2);
-    mb_dqcoeff += 16;
-    vst1_lane_s16(mb_dqcoeff, vget_high_s16(q1s16), 2);
-    mb_dqcoeff += 16;
+  vst1_lane_s16(mb_dqcoeff, vget_low_s16(q0s16), 2);
+  mb_dqcoeff += 16;
+  vst1_lane_s16(mb_dqcoeff, vget_high_s16(q0s16), 2);
+  mb_dqcoeff += 16;
+  vst1_lane_s16(mb_dqcoeff, vget_low_s16(q1s16), 2);
+  mb_dqcoeff += 16;
+  vst1_lane_s16(mb_dqcoeff, vget_high_s16(q1s16), 2);
+  mb_dqcoeff += 16;
 
-    vst1_lane_s16(mb_dqcoeff, vget_low_s16(q0s16),  3);
-    mb_dqcoeff += 16;
-    vst1_lane_s16(mb_dqcoeff, vget_high_s16(q0s16), 3);
-    mb_dqcoeff += 16;
-    vst1_lane_s16(mb_dqcoeff, vget_low_s16(q1s16),  3);
-    mb_dqcoeff += 16;
-    vst1_lane_s16(mb_dqcoeff, vget_high_s16(q1s16), 3);
-    mb_dqcoeff += 16;
-    return;
+  vst1_lane_s16(mb_dqcoeff, vget_low_s16(q0s16), 3);
+  mb_dqcoeff += 16;
+  vst1_lane_s16(mb_dqcoeff, vget_high_s16(q0s16), 3);
+  mb_dqcoeff += 16;
+  vst1_lane_s16(mb_dqcoeff, vget_low_s16(q1s16), 3);
+  mb_dqcoeff += 16;
+  vst1_lane_s16(mb_dqcoeff, vget_high_s16(q1s16), 3);
+  mb_dqcoeff += 16;
+  return;
 }
--- a/vp8/common/arm/neon/loopfiltersimplehorizontaledge_neon.c
+++ b/vp8/common/arm/neon/loopfiltersimplehorizontaledge_neon.c
@@ -12,100 +12,93 @@
 #include "./vpx_config.h"
 
 static INLINE void vp8_loop_filter_simple_horizontal_edge_neon(
-        unsigned char *s,
-        int p,
-        const unsigned char *blimit) {
-    uint8_t *sp;
-    uint8x16_t qblimit, q0u8;
-    uint8x16_t q5u8, q6u8, q7u8, q8u8, q9u8, q10u8, q14u8, q15u8;
-    int16x8_t q2s16, q3s16, q13s16;
-    int8x8_t d8s8, d9s8;
-    int8x16_t q2s8, q3s8, q4s8, q10s8, q11s8, q14s8;
+    unsigned char *s, int p, const unsigned char *blimit) {
+  uint8_t *sp;
+  uint8x16_t qblimit, q0u8;
+  uint8x16_t q5u8, q6u8, q7u8, q8u8, q9u8, q10u8, q14u8, q15u8;
+  int16x8_t q2s16, q3s16, q13s16;
+  int8x8_t d8s8, d9s8;
+  int8x16_t q2s8, q3s8, q4s8, q10s8, q11s8, q14s8;
 
-    qblimit = vdupq_n_u8(*blimit);
+  qblimit = vdupq_n_u8(*blimit);
 
-    sp = s - (p << 1);
-    q5u8 = vld1q_u8(sp);
-    sp += p;
-    q6u8 = vld1q_u8(sp);
-    sp += p;
-    q7u8 = vld1q_u8(sp);
-    sp += p;
-    q8u8 = vld1q_u8(sp);
+  sp = s - (p << 1);
+  q5u8 = vld1q_u8(sp);
+  sp += p;
+  q6u8 = vld1q_u8(sp);
+  sp += p;
+  q7u8 = vld1q_u8(sp);
+  sp += p;
+  q8u8 = vld1q_u8(sp);
 
-    q15u8 = vabdq_u8(q6u8, q7u8);
-    q14u8 = vabdq_u8(q5u8, q8u8);
+  q15u8 = vabdq_u8(q6u8, q7u8);
+  q14u8 = vabdq_u8(q5u8, q8u8);
 
-    q15u8 = vqaddq_u8(q15u8, q15u8);
-    q14u8 = vshrq_n_u8(q14u8, 1);
-    q0u8 = vdupq_n_u8(0x80);
-    q13s16 = vdupq_n_s16(3);
-    q15u8 = vqaddq_u8(q15u8, q14u8);
+  q15u8 = vqaddq_u8(q15u8, q15u8);
+  q14u8 = vshrq_n_u8(q14u8, 1);
+  q0u8 = vdupq_n_u8(0x80);
+  q13s16 = vdupq_n_s16(3);
+  q15u8 = vqaddq_u8(q15u8, q14u8);
 
-    q5u8 = veorq_u8(q5u8, q0u8);
-    q6u8 = veorq_u8(q6u8, q0u8);
-    q7u8 = veorq_u8(q7u8, q0u8);
-    q8u8 = veorq_u8(q8u8, q0u8);
+  q5u8 = veorq_u8(q5u8, q0u8);
+  q6u8 = veorq_u8(q6u8, q0u8);
+  q7u8 = veorq_u8(q7u8, q0u8);
+  q8u8 = veorq_u8(q8u8, q0u8);
 
-    q15u8 = vcgeq_u8(qblimit, q15u8);
+  q15u8 = vcgeq_u8(qblimit, q15u8);
 
-    q2s16 = vsubl_s8(vget_low_s8(vreinterpretq_s8_u8(q7u8)),
-                     vget_low_s8(vreinterpretq_s8_u8(q6u8)));
-    q3s16 = vsubl_s8(vget_high_s8(vreinterpretq_s8_u8(q7u8)),
-                     vget_high_s8(vreinterpretq_s8_u8(q6u8)));
+  q2s16 = vsubl_s8(vget_low_s8(vreinterpretq_s8_u8(q7u8)),
+                   vget_low_s8(vreinterpretq_s8_u8(q6u8)));
+  q3s16 = vsubl_s8(vget_high_s8(vreinterpretq_s8_u8(q7u8)),
+                   vget_high_s8(vreinterpretq_s8_u8(q6u8)));
 
-    q4s8 = vqsubq_s8(vreinterpretq_s8_u8(q5u8),
-                     vreinterpretq_s8_u8(q8u8));
+  q4s8 = vqsubq_s8(vreinterpretq_s8_u8(q5u8), vreinterpretq_s8_u8(q8u8));
 
-    q2s16 = vmulq_s16(q2s16, q13s16);
-    q3s16 = vmulq_s16(q3s16, q13s16);
+  q2s16 = vmulq_s16(q2s16, q13s16);
+  q3s16 = vmulq_s16(q3s16, q13s16);
 
-    q10u8 = vdupq_n_u8(3);
-    q9u8 = vdupq_n_u8(4);
+  q10u8 = vdupq_n_u8(3);
+  q9u8 = vdupq_n_u8(4);
 
-    q2s16 = vaddw_s8(q2s16, vget_low_s8(q4s8));
-    q3s16 = vaddw_s8(q3s16, vget_high_s8(q4s8));
+  q2s16 = vaddw_s8(q2s16, vget_low_s8(q4s8));
+  q3s16 = vaddw_s8(q3s16, vget_high_s8(q4s8));
 
-    d8s8 = vqmovn_s16(q2s16);
-    d9s8 = vqmovn_s16(q3s16);
-    q4s8 = vcombine_s8(d8s8, d9s8);
+  d8s8 = vqmovn_s16(q2s16);
+  d9s8 = vqmovn_s16(q3s16);
+  q4s8 = vcombine_s8(d8s8, d9s8);
 
-    q14s8 = vandq_s8(q4s8, vreinterpretq_s8_u8(q15u8));
+  q14s8 = vandq_s8(q4s8, vreinterpretq_s8_u8(q15u8));
 
-    q2s8 = vqaddq_s8(q14s8, vreinterpretq_s8_u8(q10u8));
-    q3s8 = vqaddq_s8(q14s8, vreinterpretq_s8_u8(q9u8));
-    q2s8 = vshrq_n_s8(q2s8, 3);
-    q3s8 = vshrq_n_s8(q3s8, 3);
+  q2s8 = vqaddq_s8(q14s8, vreinterpretq_s8_u8(q10u8));
+  q3s8 = vqaddq_s8(q14s8, vreinterpretq_s8_u8(q9u8));
+  q2s8 = vshrq_n_s8(q2s8, 3);
+  q3s8 = vshrq_n_s8(q3s8, 3);
 
-    q11s8 = vqaddq_s8(vreinterpretq_s8_u8(q6u8), q2s8);
-    q10s8 = vqsubq_s8(vreinterpretq_s8_u8(q7u8), q3s8);
+  q11s8 = vqaddq_s8(vreinterpretq_s8_u8(q6u8), q2s8);
+  q10s8 = vqsubq_s8(vreinterpretq_s8_u8(q7u8), q3s8);
 
-    q6u8 = veorq_u8(vreinterpretq_u8_s8(q11s8), q0u8);
-    q7u8 = veorq_u8(vreinterpretq_u8_s8(q10s8), q0u8);
+  q6u8 = veorq_u8(vreinterpretq_u8_s8(q11s8), q0u8);
+  q7u8 = veorq_u8(vreinterpretq_u8_s8(q10s8), q0u8);
 
-    vst1q_u8(s, q7u8);
-    s -= p;
-    vst1q_u8(s, q6u8);
-    return;
+  vst1q_u8(s, q7u8);
+  s -= p;
+  vst1q_u8(s, q6u8);
+  return;
 }
 
-void vp8_loop_filter_bhs_neon(
-        unsigned char *y_ptr,
-        int y_stride,
-        const unsigned char *blimit) {
-    y_ptr += y_stride * 4;
-    vp8_loop_filter_simple_horizontal_edge_neon(y_ptr, y_stride, blimit);
-    y_ptr += y_stride * 4;
-    vp8_loop_filter_simple_horizontal_edge_neon(y_ptr, y_stride, blimit);
-    y_ptr += y_stride * 4;
-    vp8_loop_filter_simple_horizontal_edge_neon(y_ptr, y_stride, blimit);
-    return;
+void vp8_loop_filter_bhs_neon(unsigned char *y_ptr, int y_stride,
+                              const unsigned char *blimit) {
+  y_ptr += y_stride * 4;
+  vp8_loop_filter_simple_horizontal_edge_neon(y_ptr, y_stride, blimit);
+  y_ptr += y_stride * 4;
+  vp8_loop_filter_simple_horizontal_edge_neon(y_ptr, y_stride, blimit);
+  y_ptr += y_stride * 4;
+  vp8_loop_filter_simple_horizontal_edge_neon(y_ptr, y_stride, blimit);
+  return;
 }
 
-void vp8_loop_filter_mbhs_neon(
-        unsigned char *y_ptr,
-        int y_stride,
-        const unsigned char *blimit) {
-    vp8_loop_filter_simple_horizontal_edge_neon(y_ptr, y_stride, blimit);
-    return;
+void vp8_loop_filter_mbhs_neon(unsigned char *y_ptr, int y_stride,
+                               const unsigned char *blimit) {
+  vp8_loop_filter_simple_horizontal_edge_neon(y_ptr, y_stride, blimit);
+  return;
 }
--- a/vp8/common/arm/neon/loopfiltersimpleverticaledge_neon.c
+++ b/vp8/common/arm/neon/loopfiltersimpleverticaledge_neon.c
@@ -15,34 +15,33 @@
 #ifdef VPX_INCOMPATIBLE_GCC
 static INLINE void write_2x4(unsigned char *dst, int pitch,
                              const uint8x8x2_t result) {
-    /*
-     * uint8x8x2_t result
-    00 01 02 03 | 04 05 06 07
-    10 11 12 13 | 14 15 16 17
-    ---
-    * after vtrn_u8
-    00 10 02 12 | 04 14 06 16
-    01 11 03 13 | 05 15 07 17
-    */
-    const uint8x8x2_t r01_u8 = vtrn_u8(result.val[0],
-                                       result.val[1]);
-    const uint16x4_t x_0_4 = vreinterpret_u16_u8(r01_u8.val[0]);
-    const uint16x4_t x_1_5 = vreinterpret_u16_u8(r01_u8.val[1]);
-    vst1_lane_u16((uint16_t *)dst, x_0_4, 0);
-    dst += pitch;
-    vst1_lane_u16((uint16_t *)dst, x_1_5, 0);
-    dst += pitch;
-    vst1_lane_u16((uint16_t *)dst, x_0_4, 1);
-    dst += pitch;
-    vst1_lane_u16((uint16_t *)dst, x_1_5, 1);
-    dst += pitch;
-    vst1_lane_u16((uint16_t *)dst, x_0_4, 2);
-    dst += pitch;
-    vst1_lane_u16((uint16_t *)dst, x_1_5, 2);
-    dst += pitch;
-    vst1_lane_u16((uint16_t *)dst, x_0_4, 3);
-    dst += pitch;
-    vst1_lane_u16((uint16_t *)dst, x_1_5, 3);
+  /*
+   * uint8x8x2_t result
+  00 01 02 03 | 04 05 06 07
+  10 11 12 13 | 14 15 16 17
+  ---
+  * after vtrn_u8
+  00 10 02 12 | 04 14 06 16
+  01 11 03 13 | 05 15 07 17
+  */
+  const uint8x8x2_t r01_u8 = vtrn_u8(result.val[0], result.val[1]);
+  const uint16x4_t x_0_4 = vreinterpret_u16_u8(r01_u8.val[0]);
+  const uint16x4_t x_1_5 = vreinterpret_u16_u8(r01_u8.val[1]);
+  vst1_lane_u16((uint16_t *)dst, x_0_4, 0);
+  dst += pitch;
+  vst1_lane_u16((uint16_t *)dst, x_1_5, 0);
+  dst += pitch;
+  vst1_lane_u16((uint16_t *)dst, x_0_4, 1);
+  dst += pitch;
+  vst1_lane_u16((uint16_t *)dst, x_1_5, 1);
+  dst += pitch;
+  vst1_lane_u16((uint16_t *)dst, x_0_4, 2);
+  dst += pitch;
+  vst1_lane_u16((uint16_t *)dst, x_1_5, 2);
+  dst += pitch;
+  vst1_lane_u16((uint16_t *)dst, x_0_4, 3);
+  dst += pitch;
+  vst1_lane_u16((uint16_t *)dst, x_1_5, 3);
 }
 
 static INLINE void write_2x8(unsigned char *dst, int pitch,
@@ -91,193 +90,183 @@
 }
 #endif  // VPX_INCOMPATIBLE_GCC
 
-
 #ifdef VPX_INCOMPATIBLE_GCC
-static INLINE
-uint8x8x4_t read_4x8(unsigned char *src, int pitch) {
-    uint8x8x4_t x;
-    const uint8x8_t a = vld1_u8(src);
-    const uint8x8_t b = vld1_u8(src + pitch * 1);
-    const uint8x8_t c = vld1_u8(src + pitch * 2);
-    const uint8x8_t d = vld1_u8(src + pitch * 3);
-    const uint8x8_t e = vld1_u8(src + pitch * 4);
-    const uint8x8_t f = vld1_u8(src + pitch * 5);
-    const uint8x8_t g = vld1_u8(src + pitch * 6);
-    const uint8x8_t h = vld1_u8(src + pitch * 7);
-    const uint32x2x2_t r04_u32 = vtrn_u32(vreinterpret_u32_u8(a),
-                                          vreinterpret_u32_u8(e));
-    const uint32x2x2_t r15_u32 = vtrn_u32(vreinterpret_u32_u8(b),
-                                          vreinterpret_u32_u8(f));
-    const uint32x2x2_t r26_u32 = vtrn_u32(vreinterpret_u32_u8(c),
-                                          vreinterpret_u32_u8(g));
-    const uint32x2x2_t r37_u32 = vtrn_u32(vreinterpret_u32_u8(d),
-                                          vreinterpret_u32_u8(h));
-    const uint16x4x2_t r02_u16 = vtrn_u16(vreinterpret_u16_u32(r04_u32.val[0]),
-                                          vreinterpret_u16_u32(r26_u32.val[0]));
-    const uint16x4x2_t r13_u16 = vtrn_u16(vreinterpret_u16_u32(r15_u32.val[0]),
-                                          vreinterpret_u16_u32(r37_u32.val[0]));
-    const uint8x8x2_t r01_u8 = vtrn_u8(vreinterpret_u8_u16(r02_u16.val[0]),
-                                       vreinterpret_u8_u16(r13_u16.val[0]));
-    const uint8x8x2_t r23_u8 = vtrn_u8(vreinterpret_u8_u16(r02_u16.val[1]),
-                                       vreinterpret_u8_u16(r13_u16.val[1]));
-    /*
-     * after vtrn_u32
-    00 01 02 03 | 40 41 42 43
-    10 11 12 13 | 50 51 52 53
-    20 21 22 23 | 60 61 62 63
-    30 31 32 33 | 70 71 72 73
-    ---
-    * after vtrn_u16
-    00 01 20 21 | 40 41 60 61
-    02 03 22 23 | 42 43 62 63
-    10 11 30 31 | 50 51 70 71
-    12 13 32 33 | 52 52 72 73
+static INLINE uint8x8x4_t read_4x8(unsigned char *src, int pitch) {
+  uint8x8x4_t x;
+  const uint8x8_t a = vld1_u8(src);
+  const uint8x8_t b = vld1_u8(src + pitch * 1);
+  const uint8x8_t c = vld1_u8(src + pitch * 2);
+  const uint8x8_t d = vld1_u8(src + pitch * 3);
+  const uint8x8_t e = vld1_u8(src + pitch * 4);
+  const uint8x8_t f = vld1_u8(src + pitch * 5);
+  const uint8x8_t g = vld1_u8(src + pitch * 6);
+  const uint8x8_t h = vld1_u8(src + pitch * 7);
+  const uint32x2x2_t r04_u32 =
+      vtrn_u32(vreinterpret_u32_u8(a), vreinterpret_u32_u8(e));
+  const uint32x2x2_t r15_u32 =
+      vtrn_u32(vreinterpret_u32_u8(b), vreinterpret_u32_u8(f));
+  const uint32x2x2_t r26_u32 =
+      vtrn_u32(vreinterpret_u32_u8(c), vreinterpret_u32_u8(g));
+  const uint32x2x2_t r37_u32 =
+      vtrn_u32(vreinterpret_u32_u8(d), vreinterpret_u32_u8(h));
+  const uint16x4x2_t r02_u16 = vtrn_u16(vreinterpret_u16_u32(r04_u32.val[0]),
+                                        vreinterpret_u16_u32(r26_u32.val[0]));
+  const uint16x4x2_t r13_u16 = vtrn_u16(vreinterpret_u16_u32(r15_u32.val[0]),
+                                        vreinterpret_u16_u32(r37_u32.val[0]));
+  const uint8x8x2_t r01_u8 = vtrn_u8(vreinterpret_u8_u16(r02_u16.val[0]),
+                                     vreinterpret_u8_u16(r13_u16.val[0]));
+  const uint8x8x2_t r23_u8 = vtrn_u8(vreinterpret_u8_u16(r02_u16.val[1]),
+                                     vreinterpret_u8_u16(r13_u16.val[1]));
+  /*
+   * after vtrn_u32
+  00 01 02 03 | 40 41 42 43
+  10 11 12 13 | 50 51 52 53
+  20 21 22 23 | 60 61 62 63
+  30 31 32 33 | 70 71 72 73
+  ---
+  * after vtrn_u16
+  00 01 20 21 | 40 41 60 61
+  02 03 22 23 | 42 43 62 63
+  10 11 30 31 | 50 51 70 71
+  12 13 32 33 | 52 52 72 73
 
-    00 01 20 21 | 40 41 60 61
-    10 11 30 31 | 50 51 70 71
-    02 03 22 23 | 42 43 62 63
-    12 13 32 33 | 52 52 72 73
-    ---
-    * after vtrn_u8
-    00 10 20 30 | 40 50 60 70
-    01 11 21 31 | 41 51 61 71
-    02 12 22 32 | 42 52 62 72
-    03 13 23 33 | 43 53 63 73
-    */
-    x.val[0] = r01_u8.val[0];
-    x.val[1] = r01_u8.val[1];
-    x.val[2] = r23_u8.val[0];
-    x.val[3] = r23_u8.val[1];
+  00 01 20 21 | 40 41 60 61
+  10 11 30 31 | 50 51 70 71
+  02 03 22 23 | 42 43 62 63
+  12 13 32 33 | 52 52 72 73
+  ---
+  * after vtrn_u8
+  00 10 20 30 | 40 50 60 70
+  01 11 21 31 | 41 51 61 71
+  02 12 22 32 | 42 52 62 72
+  03 13 23 33 | 43 53 63 73
+  */
+  x.val[0] = r01_u8.val[0];
+  x.val[1] = r01_u8.val[1];
+  x.val[2] = r23_u8.val[0];
+  x.val[3] = r23_u8.val[1];
 
-    return x;
+  return x;
 }
 #else
-static INLINE
-uint8x8x4_t read_4x8(unsigned char *src, int pitch) {
-    uint8x8x4_t x;
-    x.val[0] = x.val[1] = x.val[2] = x.val[3] = vdup_n_u8(0);
-    x = vld4_lane_u8(src, x, 0);
-    src += pitch;
-    x = vld4_lane_u8(src, x, 1);
-    src += pitch;
-    x = vld4_lane_u8(src, x, 2);
-    src += pitch;
-    x = vld4_lane_u8(src, x, 3);
-    src += pitch;
-    x = vld4_lane_u8(src, x, 4);
-    src += pitch;
-    x = vld4_lane_u8(src, x, 5);
-    src += pitch;
-    x = vld4_lane_u8(src, x, 6);
-    src += pitch;
-    x = vld4_lane_u8(src, x, 7);
-    return x;
+static INLINE uint8x8x4_t read_4x8(unsigned char *src, int pitch) {
+  uint8x8x4_t x;
+  x.val[0] = x.val[1] = x.val[2] = x.val[3] = vdup_n_u8(0);
+  x = vld4_lane_u8(src, x, 0);
+  src += pitch;
+  x = vld4_lane_u8(src, x, 1);
+  src += pitch;
+  x = vld4_lane_u8(src, x, 2);
+  src += pitch;
+  x = vld4_lane_u8(src, x, 3);
+  src += pitch;
+  x = vld4_lane_u8(src, x, 4);
+  src += pitch;
+  x = vld4_lane_u8(src, x, 5);
+  src += pitch;
+  x = vld4_lane_u8(src, x, 6);
+  src += pitch;
+  x = vld4_lane_u8(src, x, 7);
+  return x;
 }
 #endif  // VPX_INCOMPATIBLE_GCC
 
 static INLINE void vp8_loop_filter_simple_vertical_edge_neon(
-        unsigned char *s,
-        int p,
-        const unsigned char *blimit) {
-    unsigned char *src1;
-    uint8x16_t qblimit, q0u8;
-    uint8x16_t q3u8, q4u8, q5u8, q6u8, q7u8, q11u8, q12u8, q14u8, q15u8;
-    int16x8_t q2s16, q13s16, q11s16;
-    int8x8_t d28s8, d29s8;
-    int8x16_t q2s8, q3s8, q10s8, q11s8, q14s8;
-    uint8x8x4_t d0u8x4;  // d6, d7, d8, d9
-    uint8x8x4_t d1u8x4;  // d10, d11, d12, d13
-    uint8x8x2_t d2u8x2;  // d12, d13
-    uint8x8x2_t d3u8x2;  // d14, d15
+    unsigned char *s, int p, const unsigned char *blimit) {
+  unsigned char *src1;
+  uint8x16_t qblimit, q0u8;
+  uint8x16_t q3u8, q4u8, q5u8, q6u8, q7u8, q11u8, q12u8, q14u8, q15u8;
+  int16x8_t q2s16, q13s16, q11s16;
+  int8x8_t d28s8, d29s8;
+  int8x16_t q2s8, q3s8, q10s8, q11s8, q14s8;
+  uint8x8x4_t d0u8x4;  // d6, d7, d8, d9
+  uint8x8x4_t d1u8x4;  // d10, d11, d12, d13
+  uint8x8x2_t d2u8x2;  // d12, d13
+  uint8x8x2_t d3u8x2;  // d14, d15
 
-    qblimit = vdupq_n_u8(*blimit);
+  qblimit = vdupq_n_u8(*blimit);
 
-    src1 = s - 2;
-    d0u8x4 = read_4x8(src1, p);
-    src1 += p * 8;
-    d1u8x4 = read_4x8(src1, p);
+  src1 = s - 2;
+  d0u8x4 = read_4x8(src1, p);
+  src1 += p * 8;
+  d1u8x4 = read_4x8(src1, p);
 
-    q3u8 = vcombine_u8(d0u8x4.val[0], d1u8x4.val[0]);  // d6 d10
-    q4u8 = vcombine_u8(d0u8x4.val[2], d1u8x4.val[2]);  // d8 d12
-    q5u8 = vcombine_u8(d0u8x4.val[1], d1u8x4.val[1]);  // d7 d11
-    q6u8 = vcombine_u8(d0u8x4.val[3], d1u8x4.val[3]);  // d9 d13
+  q3u8 = vcombine_u8(d0u8x4.val[0], d1u8x4.val[0]);  // d6 d10
+  q4u8 = vcombine_u8(d0u8x4.val[2], d1u8x4.val[2]);  // d8 d12
+  q5u8 = vcombine_u8(d0u8x4.val[1], d1u8x4.val[1]);  // d7 d11
+  q6u8 = vcombine_u8(d0u8x4.val[3], d1u8x4.val[3]);  // d9 d13
 
-    q15u8 = vabdq_u8(q5u8, q4u8);
-    q14u8 = vabdq_u8(q3u8, q6u8);
+  q15u8 = vabdq_u8(q5u8, q4u8);
+  q14u8 = vabdq_u8(q3u8, q6u8);
 
-    q15u8 = vqaddq_u8(q15u8, q15u8);
-    q14u8 = vshrq_n_u8(q14u8, 1);
-    q0u8 = vdupq_n_u8(0x80);
-    q11s16 = vdupq_n_s16(3);
-    q15u8 = vqaddq_u8(q15u8, q14u8);
+  q15u8 = vqaddq_u8(q15u8, q15u8);
+  q14u8 = vshrq_n_u8(q14u8, 1);
+  q0u8 = vdupq_n_u8(0x80);
+  q11s16 = vdupq_n_s16(3);
+  q15u8 = vqaddq_u8(q15u8, q14u8);
 
-    q3u8 = veorq_u8(q3u8, q0u8);
-    q4u8 = veorq_u8(q4u8, q0u8);
-    q5u8 = veorq_u8(q5u8, q0u8);
-    q6u8 = veorq_u8(q6u8, q0u8);
+  q3u8 = veorq_u8(q3u8, q0u8);
+  q4u8 = veorq_u8(q4u8, q0u8);
+  q5u8 = veorq_u8(q5u8, q0u8);
+  q6u8 = veorq_u8(q6u8, q0u8);
 
-    q15u8 = vcgeq_u8(qblimit, q15u8);
+  q15u8 = vcgeq_u8(qblimit, q15u8);
 
-    q2s16 = vsubl_s8(vget_low_s8(vreinterpretq_s8_u8(q4u8)),
-                     vget_low_s8(vreinterpretq_s8_u8(q5u8)));
-    q13s16 = vsubl_s8(vget_high_s8(vreinterpretq_s8_u8(q4u8)),
-                      vget_high_s8(vreinterpretq_s8_u8(q5u8)));
+  q2s16 = vsubl_s8(vget_low_s8(vreinterpretq_s8_u8(q4u8)),
+                   vget_low_s8(vreinterpretq_s8_u8(q5u8)));
+  q13s16 = vsubl_s8(vget_high_s8(vreinterpretq_s8_u8(q4u8)),
+                    vget_high_s8(vreinterpretq_s8_u8(q5u8)));
 
-    q14s8 = vqsubq_s8(vreinterpretq_s8_u8(q3u8),
-                      vreinterpretq_s8_u8(q6u8));
+  q14s8 = vqsubq_s8(vreinterpretq_s8_u8(q3u8), vreinterpretq_s8_u8(q6u8));
 
-    q2s16 = vmulq_s16(q2s16, q11s16);
-    q13s16 = vmulq_s16(q13s16, q11s16);
+  q2s16 = vmulq_s16(q2s16, q11s16);
+  q13s16 = vmulq_s16(q13s16, q11s16);
 
-    q11u8 = vdupq_n_u8(3);
-    q12u8 = vdupq_n_u8(4);
+  q11u8 = vdupq_n_u8(3);
+  q12u8 = vdupq_n_u8(4);
 
-    q2s16 = vaddw_s8(q2s16, vget_low_s8(q14s8));
-    q13s16 = vaddw_s8(q13s16, vget_high_s8(q14s8));
+  q2s16 = vaddw_s8(q2s16, vget_low_s8(q14s8));
+  q13s16 = vaddw_s8(q13s16, vget_high_s8(q14s8));
 
-    d28s8 = vqmovn_s16(q2s16);
-    d29s8 = vqmovn_s16(q13s16);
-    q14s8 = vcombine_s8(d28s8, d29s8);
+  d28s8 = vqmovn_s16(q2s16);
+  d29s8 = vqmovn_s16(q13s16);
+  q14s8 = vcombine_s8(d28s8, d29s8);
 
-    q14s8 = vandq_s8(q14s8, vreinterpretq_s8_u8(q15u8));
+  q14s8 = vandq_s8(q14s8, vreinterpretq_s8_u8(q15u8));
 
-    q2s8 = vqaddq_s8(q14s8, vreinterpretq_s8_u8(q11u8));
-    q3s8 = vqaddq_s8(q14s8, vreinterpretq_s8_u8(q12u8));
-    q2s8 = vshrq_n_s8(q2s8, 3);
-    q14s8 = vshrq_n_s8(q3s8, 3);
+  q2s8 = vqaddq_s8(q14s8, vreinterpretq_s8_u8(q11u8));
+  q3s8 = vqaddq_s8(q14s8, vreinterpretq_s8_u8(q12u8));
+  q2s8 = vshrq_n_s8(q2s8, 3);
+  q14s8 = vshrq_n_s8(q3s8, 3);
 
-    q11s8 = vqaddq_s8(vreinterpretq_s8_u8(q5u8), q2s8);
-    q10s8 = vqsubq_s8(vreinterpretq_s8_u8(q4u8), q14s8);
+  q11s8 = vqaddq_s8(vreinterpretq_s8_u8(q5u8), q2s8);
+  q10s8 = vqsubq_s8(vreinterpretq_s8_u8(q4u8), q14s8);
 
-    q6u8 = veorq_u8(vreinterpretq_u8_s8(q11s8), q0u8);
-    q7u8 = veorq_u8(vreinterpretq_u8_s8(q10s8), q0u8);
+  q6u8 = veorq_u8(vreinterpretq_u8_s8(q11s8), q0u8);
+  q7u8 = veorq_u8(vreinterpretq_u8_s8(q10s8), q0u8);
 
-    d2u8x2.val[0] = vget_low_u8(q6u8);   // d12
-    d2u8x2.val[1] = vget_low_u8(q7u8);   // d14
-    d3u8x2.val[0] = vget_high_u8(q6u8);  // d13
-    d3u8x2.val[1] = vget_high_u8(q7u8);  // d15
+  d2u8x2.val[0] = vget_low_u8(q6u8);   // d12
+  d2u8x2.val[1] = vget_low_u8(q7u8);   // d14
+  d3u8x2.val[0] = vget_high_u8(q6u8);  // d13
+  d3u8x2.val[1] = vget_high_u8(q7u8);  // d15
 
-    src1 = s - 1;
-    write_2x8(src1, p, d2u8x2, d3u8x2);
+  src1 = s - 1;
+  write_2x8(src1, p, d2u8x2, d3u8x2);
 }
 
-void vp8_loop_filter_bvs_neon(
-        unsigned char *y_ptr,
-        int y_stride,
-        const unsigned char *blimit) {
-    y_ptr += 4;
-    vp8_loop_filter_simple_vertical_edge_neon(y_ptr, y_stride, blimit);
-    y_ptr += 4;
-    vp8_loop_filter_simple_vertical_edge_neon(y_ptr, y_stride, blimit);
-    y_ptr += 4;
-    vp8_loop_filter_simple_vertical_edge_neon(y_ptr, y_stride, blimit);
-    return;
+void vp8_loop_filter_bvs_neon(unsigned char *y_ptr, int y_stride,
+                              const unsigned char *blimit) {
+  y_ptr += 4;
+  vp8_loop_filter_simple_vertical_edge_neon(y_ptr, y_stride, blimit);
+  y_ptr += 4;
+  vp8_loop_filter_simple_vertical_edge_neon(y_ptr, y_stride, blimit);
+  y_ptr += 4;
+  vp8_loop_filter_simple_vertical_edge_neon(y_ptr, y_stride, blimit);
+  return;
 }
 
-void vp8_loop_filter_mbvs_neon(
-        unsigned char *y_ptr,
-        int y_stride,
-        const unsigned char *blimit) {
-    vp8_loop_filter_simple_vertical_edge_neon(y_ptr, y_stride, blimit);
-    return;
+void vp8_loop_filter_mbvs_neon(unsigned char *y_ptr, int y_stride,
+                               const unsigned char *blimit) {
+  vp8_loop_filter_simple_vertical_edge_neon(y_ptr, y_stride, blimit);
+  return;
 }
--- a/vp8/common/arm/neon/mbloopfilter_neon.c
+++ b/vp8/common/arm/neon/mbloopfilter_neon.c
@@ -11,615 +11,601 @@
 #include <arm_neon.h>
 #include "./vpx_config.h"
 
-static INLINE void vp8_mbloop_filter_neon(
-        uint8x16_t qblimit,  // mblimit
-        uint8x16_t qlimit,   // limit
-        uint8x16_t qthresh,  // thresh
-        uint8x16_t q3,       // p2
-        uint8x16_t q4,       // p2
-        uint8x16_t q5,       // p1
-        uint8x16_t q6,       // p0
-        uint8x16_t q7,       // q0
-        uint8x16_t q8,       // q1
-        uint8x16_t q9,       // q2
-        uint8x16_t q10,      // q3
-        uint8x16_t *q4r,     // p1
-        uint8x16_t *q5r,     // p1
-        uint8x16_t *q6r,     // p0
-        uint8x16_t *q7r,     // q0
-        uint8x16_t *q8r,     // q1
-        uint8x16_t *q9r) {   // q1
-    uint8x16_t q0u8, q1u8, q11u8, q12u8, q13u8, q14u8, q15u8;
-    int16x8_t q0s16, q2s16, q11s16, q12s16, q13s16, q14s16, q15s16;
-    int8x16_t q1s8, q6s8, q7s8, q2s8, q11s8, q13s8;
-    uint16x8_t q0u16, q11u16, q12u16, q13u16, q14u16, q15u16;
-    int8x16_t q0s8, q12s8, q14s8, q15s8;
-    int8x8_t d0, d1, d2, d3, d4, d5, d24, d25, d28, d29;
+static INLINE void vp8_mbloop_filter_neon(uint8x16_t qblimit,  // mblimit
+                                          uint8x16_t qlimit,   // limit
+                                          uint8x16_t qthresh,  // thresh
+                                          uint8x16_t q3,       // p2
+                                          uint8x16_t q4,       // p2
+                                          uint8x16_t q5,       // p1
+                                          uint8x16_t q6,       // p0
+                                          uint8x16_t q7,       // q0
+                                          uint8x16_t q8,       // q1
+                                          uint8x16_t q9,       // q2
+                                          uint8x16_t q10,      // q3
+                                          uint8x16_t *q4r,     // p1
+                                          uint8x16_t *q5r,     // p1
+                                          uint8x16_t *q6r,     // p0
+                                          uint8x16_t *q7r,     // q0
+                                          uint8x16_t *q8r,     // q1
+                                          uint8x16_t *q9r) {   // q1
+  uint8x16_t q0u8, q1u8, q11u8, q12u8, q13u8, q14u8, q15u8;
+  int16x8_t q0s16, q2s16, q11s16, q12s16, q13s16, q14s16, q15s16;
+  int8x16_t q1s8, q6s8, q7s8, q2s8, q11s8, q13s8;
+  uint16x8_t q0u16, q11u16, q12u16, q13u16, q14u16, q15u16;
+  int8x16_t q0s8, q12s8, q14s8, q15s8;
+  int8x8_t d0, d1, d2, d3, d4, d5, d24, d25, d28, d29;
 
-    q11u8 = vabdq_u8(q3, q4);
-    q12u8 = vabdq_u8(q4, q5);
-    q13u8 = vabdq_u8(q5, q6);
-    q14u8 = vabdq_u8(q8, q7);
-    q1u8  = vabdq_u8(q9, q8);
-    q0u8  = vabdq_u8(q10, q9);
+  q11u8 = vabdq_u8(q3, q4);
+  q12u8 = vabdq_u8(q4, q5);
+  q13u8 = vabdq_u8(q5, q6);
+  q14u8 = vabdq_u8(q8, q7);
+  q1u8 = vabdq_u8(q9, q8);
+  q0u8 = vabdq_u8(q10, q9);
 
-    q11u8 = vmaxq_u8(q11u8, q12u8);
-    q12u8 = vmaxq_u8(q13u8, q14u8);
-    q1u8  = vmaxq_u8(q1u8, q0u8);
-    q15u8 = vmaxq_u8(q11u8, q12u8);
+  q11u8 = vmaxq_u8(q11u8, q12u8);
+  q12u8 = vmaxq_u8(q13u8, q14u8);
+  q1u8 = vmaxq_u8(q1u8, q0u8);
+  q15u8 = vmaxq_u8(q11u8, q12u8);
 
-    q12u8 = vabdq_u8(q6, q7);
+  q12u8 = vabdq_u8(q6, q7);
 
-    // vp8_hevmask
-    q13u8 = vcgtq_u8(q13u8, qthresh);
-    q14u8 = vcgtq_u8(q14u8, qthresh);
-    q15u8 = vmaxq_u8(q15u8, q1u8);
+  // vp8_hevmask
+  q13u8 = vcgtq_u8(q13u8, qthresh);
+  q14u8 = vcgtq_u8(q14u8, qthresh);
+  q15u8 = vmaxq_u8(q15u8, q1u8);
 
-    q15u8 = vcgeq_u8(qlimit, q15u8);
+  q15u8 = vcgeq_u8(qlimit, q15u8);
 
-    q1u8 = vabdq_u8(q5, q8);
-    q12u8 = vqaddq_u8(q12u8, q12u8);
+  q1u8 = vabdq_u8(q5, q8);
+  q12u8 = vqaddq_u8(q12u8, q12u8);
 
-    // vp8_filter() function
-    // convert to signed
-    q0u8 = vdupq_n_u8(0x80);
-    q9 = veorq_u8(q9, q0u8);
-    q8 = veorq_u8(q8, q0u8);
-    q7 = veorq_u8(q7, q0u8);
-    q6 = veorq_u8(q6, q0u8);
-    q5 = veorq_u8(q5, q0u8);
-    q4 = veorq_u8(q4, q0u8);
+  // vp8_filter() function
+  // convert to signed
+  q0u8 = vdupq_n_u8(0x80);
+  q9 = veorq_u8(q9, q0u8);
+  q8 = veorq_u8(q8, q0u8);
+  q7 = veorq_u8(q7, q0u8);
+  q6 = veorq_u8(q6, q0u8);
+  q5 = veorq_u8(q5, q0u8);
+  q4 = veorq_u8(q4, q0u8);
 
-    q1u8 = vshrq_n_u8(q1u8, 1);
-    q12u8 = vqaddq_u8(q12u8, q1u8);
+  q1u8 = vshrq_n_u8(q1u8, 1);
+  q12u8 = vqaddq_u8(q12u8, q1u8);
 
-    q14u8 = vorrq_u8(q13u8, q14u8);
-    q12u8 = vcgeq_u8(qblimit, q12u8);
+  q14u8 = vorrq_u8(q13u8, q14u8);
+  q12u8 = vcgeq_u8(qblimit, q12u8);
 
-    q2s16 = vsubl_s8(vget_low_s8(vreinterpretq_s8_u8(q7)),
-                     vget_low_s8(vreinterpretq_s8_u8(q6)));
-    q13s16 = vsubl_s8(vget_high_s8(vreinterpretq_s8_u8(q7)),
-                      vget_high_s8(vreinterpretq_s8_u8(q6)));
+  q2s16 = vsubl_s8(vget_low_s8(vreinterpretq_s8_u8(q7)),
+                   vget_low_s8(vreinterpretq_s8_u8(q6)));
+  q13s16 = vsubl_s8(vget_high_s8(vreinterpretq_s8_u8(q7)),
+                    vget_high_s8(vreinterpretq_s8_u8(q6)));
 
-    q1s8 = vqsubq_s8(vreinterpretq_s8_u8(q5),
-                     vreinterpretq_s8_u8(q8));
+  q1s8 = vqsubq_s8(vreinterpretq_s8_u8(q5), vreinterpretq_s8_u8(q8));
 
-    q11s16 = vdupq_n_s16(3);
-    q2s16  = vmulq_s16(q2s16, q11s16);
-    q13s16 = vmulq_s16(q13s16, q11s16);
+  q11s16 = vdupq_n_s16(3);
+  q2s16 = vmulq_s16(q2s16, q11s16);
+  q13s16 = vmulq_s16(q13s16, q11s16);
 
-    q15u8 = vandq_u8(q15u8, q12u8);
+  q15u8 = vandq_u8(q15u8, q12u8);
 
-    q2s16  = vaddw_s8(q2s16, vget_low_s8(q1s8));
-    q13s16 = vaddw_s8(q13s16, vget_high_s8(q1s8));
+  q2s16 = vaddw_s8(q2s16, vget_low_s8(q1s8));
+  q13s16 = vaddw_s8(q13s16, vget_high_s8(q1s8));
 
-    q12u8 = vdupq_n_u8(3);
-    q11u8 = vdupq_n_u8(4);
-    // vp8_filter = clamp(vp8_filter + 3 * ( qs0 - ps0))
-    d2 = vqmovn_s16(q2s16);
-    d3 = vqmovn_s16(q13s16);
-    q1s8 = vcombine_s8(d2, d3);
-    q1s8 = vandq_s8(q1s8, vreinterpretq_s8_u8(q15u8));
-    q13s8 = vandq_s8(q1s8, vreinterpretq_s8_u8(q14u8));
+  q12u8 = vdupq_n_u8(3);
+  q11u8 = vdupq_n_u8(4);
+  // vp8_filter = clamp(vp8_filter + 3 * ( qs0 - ps0))
+  d2 = vqmovn_s16(q2s16);
+  d3 = vqmovn_s16(q13s16);
+  q1s8 = vcombine_s8(d2, d3);
+  q1s8 = vandq_s8(q1s8, vreinterpretq_s8_u8(q15u8));
+  q13s8 = vandq_s8(q1s8, vreinterpretq_s8_u8(q14u8));
 
-    q2s8 = vqaddq_s8(q13s8, vreinterpretq_s8_u8(q11u8));
-    q13s8 = vqaddq_s8(q13s8, vreinterpretq_s8_u8(q12u8));
-    q2s8 = vshrq_n_s8(q2s8, 3);
-    q13s8 = vshrq_n_s8(q13s8, 3);
+  q2s8 = vqaddq_s8(q13s8, vreinterpretq_s8_u8(q11u8));
+  q13s8 = vqaddq_s8(q13s8, vreinterpretq_s8_u8(q12u8));
+  q2s8 = vshrq_n_s8(q2s8, 3);
+  q13s8 = vshrq_n_s8(q13s8, 3);
 
-    q7s8 = vqsubq_s8(vreinterpretq_s8_u8(q7), q2s8);
-    q6s8 = vqaddq_s8(vreinterpretq_s8_u8(q6), q13s8);
+  q7s8 = vqsubq_s8(vreinterpretq_s8_u8(q7), q2s8);
+  q6s8 = vqaddq_s8(vreinterpretq_s8_u8(q6), q13s8);
 
-    q1s8 = vbicq_s8(q1s8, vreinterpretq_s8_u8(q14u8));
+  q1s8 = vbicq_s8(q1s8, vreinterpretq_s8_u8(q14u8));
 
-    q0u16 = q11u16 = q12u16 = q13u16 = q14u16 = q15u16 = vdupq_n_u16(63);
-    d5 = vdup_n_s8(9);
-    d4 = vdup_n_s8(18);
+  q0u16 = q11u16 = q12u16 = q13u16 = q14u16 = q15u16 = vdupq_n_u16(63);
+  d5 = vdup_n_s8(9);
+  d4 = vdup_n_s8(18);
 
-    q0s16  = vmlal_s8(vreinterpretq_s16_u16(q0u16),  vget_low_s8(q1s8),  d5);
-    q11s16 = vmlal_s8(vreinterpretq_s16_u16(q11u16), vget_high_s8(q1s8), d5);
-    d5 = vdup_n_s8(27);
-    q12s16 = vmlal_s8(vreinterpretq_s16_u16(q12u16), vget_low_s8(q1s8),  d4);
-    q13s16 = vmlal_s8(vreinterpretq_s16_u16(q13u16), vget_high_s8(q1s8), d4);
-    q14s16 = vmlal_s8(vreinterpretq_s16_u16(q14u16), vget_low_s8(q1s8),  d5);
-    q15s16 = vmlal_s8(vreinterpretq_s16_u16(q15u16), vget_high_s8(q1s8), d5);
+  q0s16 = vmlal_s8(vreinterpretq_s16_u16(q0u16), vget_low_s8(q1s8), d5);
+  q11s16 = vmlal_s8(vreinterpretq_s16_u16(q11u16), vget_high_s8(q1s8), d5);
+  d5 = vdup_n_s8(27);
+  q12s16 = vmlal_s8(vreinterpretq_s16_u16(q12u16), vget_low_s8(q1s8), d4);
+  q13s16 = vmlal_s8(vreinterpretq_s16_u16(q13u16), vget_high_s8(q1s8), d4);
+  q14s16 = vmlal_s8(vreinterpretq_s16_u16(q14u16), vget_low_s8(q1s8), d5);
+  q15s16 = vmlal_s8(vreinterpretq_s16_u16(q15u16), vget_high_s8(q1s8), d5);
 
-    d0  = vqshrn_n_s16(q0s16 , 7);
-    d1  = vqshrn_n_s16(q11s16, 7);
-    d24 = vqshrn_n_s16(q12s16, 7);
-    d25 = vqshrn_n_s16(q13s16, 7);
-    d28 = vqshrn_n_s16(q14s16, 7);
-    d29 = vqshrn_n_s16(q15s16, 7);
+  d0 = vqshrn_n_s16(q0s16, 7);
+  d1 = vqshrn_n_s16(q11s16, 7);
+  d24 = vqshrn_n_s16(q12s16, 7);
+  d25 = vqshrn_n_s16(q13s16, 7);
+  d28 = vqshrn_n_s16(q14s16, 7);
+  d29 = vqshrn_n_s16(q15s16, 7);
 
-    q0s8  = vcombine_s8(d0, d1);
-    q12s8 = vcombine_s8(d24, d25);
-    q14s8 = vcombine_s8(d28, d29);
+  q0s8 = vcombine_s8(d0, d1);
+  q12s8 = vcombine_s8(d24, d25);
+  q14s8 = vcombine_s8(d28, d29);
 
-    q11s8 = vqsubq_s8(vreinterpretq_s8_u8(q9), q0s8);
-    q0s8  = vqaddq_s8(vreinterpretq_s8_u8(q4), q0s8);
-    q13s8 = vqsubq_s8(vreinterpretq_s8_u8(q8), q12s8);
-    q12s8 = vqaddq_s8(vreinterpretq_s8_u8(q5), q12s8);
-    q15s8 = vqsubq_s8((q7s8), q14s8);
-    q14s8 = vqaddq_s8((q6s8), q14s8);
+  q11s8 = vqsubq_s8(vreinterpretq_s8_u8(q9), q0s8);
+  q0s8 = vqaddq_s8(vreinterpretq_s8_u8(q4), q0s8);
+  q13s8 = vqsubq_s8(vreinterpretq_s8_u8(q8), q12s8);
+  q12s8 = vqaddq_s8(vreinterpretq_s8_u8(q5), q12s8);
+  q15s8 = vqsubq_s8((q7s8), q14s8);
+  q14s8 = vqaddq_s8((q6s8), q14s8);
 
-    q1u8 = vdupq_n_u8(0x80);
-    *q9r = veorq_u8(vreinterpretq_u8_s8(q11s8), q1u8);
-    *q8r = veorq_u8(vreinterpretq_u8_s8(q13s8), q1u8);
-    *q7r = veorq_u8(vreinterpretq_u8_s8(q15s8), q1u8);
-    *q6r = veorq_u8(vreinterpretq_u8_s8(q14s8), q1u8);
-    *q5r = veorq_u8(vreinterpretq_u8_s8(q12s8), q1u8);
-    *q4r = veorq_u8(vreinterpretq_u8_s8(q0s8), q1u8);
-    return;
+  q1u8 = vdupq_n_u8(0x80);
+  *q9r = veorq_u8(vreinterpretq_u8_s8(q11s8), q1u8);
+  *q8r = veorq_u8(vreinterpretq_u8_s8(q13s8), q1u8);
+  *q7r = veorq_u8(vreinterpretq_u8_s8(q15s8), q1u8);
+  *q6r = veorq_u8(vreinterpretq_u8_s8(q14s8), q1u8);
+  *q5r = veorq_u8(vreinterpretq_u8_s8(q12s8), q1u8);
+  *q4r = veorq_u8(vreinterpretq_u8_s8(q0s8), q1u8);
+  return;
 }
 
-void vp8_mbloop_filter_horizontal_edge_y_neon(
-        unsigned char *src,
-        int pitch,
-        unsigned char blimit,
-        unsigned char limit,
-        unsigned char thresh) {
-    uint8x16_t qblimit, qlimit, qthresh, q3, q4;
-    uint8x16_t q5, q6, q7, q8, q9, q10;
+void vp8_mbloop_filter_horizontal_edge_y_neon(unsigned char *src, int pitch,
+                                              unsigned char blimit,
+                                              unsigned char limit,
+                                              unsigned char thresh) {
+  uint8x16_t qblimit, qlimit, qthresh, q3, q4;
+  uint8x16_t q5, q6, q7, q8, q9, q10;
 
-    qblimit = vdupq_n_u8(blimit);
-    qlimit = vdupq_n_u8(limit);
-    qthresh = vdupq_n_u8(thresh);
+  qblimit = vdupq_n_u8(blimit);
+  qlimit = vdupq_n_u8(limit);
+  qthresh = vdupq_n_u8(thresh);
 
-    src -= (pitch << 2);
+  src -= (pitch << 2);
 
-    q3 = vld1q_u8(src);
-    src += pitch;
-    q4 = vld1q_u8(src);
-    src += pitch;
-    q5 = vld1q_u8(src);
-    src += pitch;
-    q6 = vld1q_u8(src);
-    src += pitch;
-    q7 = vld1q_u8(src);
-    src += pitch;
-    q8 = vld1q_u8(src);
-    src += pitch;
-    q9 = vld1q_u8(src);
-    src += pitch;
-    q10 = vld1q_u8(src);
+  q3 = vld1q_u8(src);
+  src += pitch;
+  q4 = vld1q_u8(src);
+  src += pitch;
+  q5 = vld1q_u8(src);
+  src += pitch;
+  q6 = vld1q_u8(src);
+  src += pitch;
+  q7 = vld1q_u8(src);
+  src += pitch;
+  q8 = vld1q_u8(src);
+  src += pitch;
+  q9 = vld1q_u8(src);
+  src += pitch;
+  q10 = vld1q_u8(src);
 
-    vp8_mbloop_filter_neon(qblimit, qlimit, qthresh, q3, q4,
-                         q5, q6, q7, q8, q9, q10,
-                         &q4, &q5, &q6, &q7, &q8, &q9);
+  vp8_mbloop_filter_neon(qblimit, qlimit, qthresh, q3, q4, q5, q6, q7, q8, q9,
+                         q10, &q4, &q5, &q6, &q7, &q8, &q9);
 
-    src -= (pitch * 6);
-    vst1q_u8(src, q4);
-    src += pitch;
-    vst1q_u8(src, q5);
-    src += pitch;
-    vst1q_u8(src, q6);
-    src += pitch;
-    vst1q_u8(src, q7);
-    src += pitch;
-    vst1q_u8(src, q8);
-    src += pitch;
-    vst1q_u8(src, q9);
-    return;
+  src -= (pitch * 6);
+  vst1q_u8(src, q4);
+  src += pitch;
+  vst1q_u8(src, q5);
+  src += pitch;
+  vst1q_u8(src, q6);
+  src += pitch;
+  vst1q_u8(src, q7);
+  src += pitch;
+  vst1q_u8(src, q8);
+  src += pitch;
+  vst1q_u8(src, q9);
+  return;
 }
 
-void vp8_mbloop_filter_horizontal_edge_uv_neon(
-        unsigned char *u,
-        int pitch,
-        unsigned char blimit,
-        unsigned char limit,
-        unsigned char thresh,
-        unsigned char *v) {
-    uint8x16_t qblimit, qlimit, qthresh, q3, q4;
-    uint8x16_t q5, q6, q7, q8, q9, q10;
-    uint8x8_t d6, d7, d8, d9, d10, d11, d12, d13, d14;
-    uint8x8_t d15, d16, d17, d18, d19, d20, d21;
+void vp8_mbloop_filter_horizontal_edge_uv_neon(unsigned char *u, int pitch,
+                                               unsigned char blimit,
+                                               unsigned char limit,
+                                               unsigned char thresh,
+                                               unsigned char *v) {
+  uint8x16_t qblimit, qlimit, qthresh, q3, q4;
+  uint8x16_t q5, q6, q7, q8, q9, q10;
+  uint8x8_t d6, d7, d8, d9, d10, d11, d12, d13, d14;
+  uint8x8_t d15, d16, d17, d18, d19, d20, d21;
 
-    qblimit = vdupq_n_u8(blimit);
-    qlimit = vdupq_n_u8(limit);
-    qthresh = vdupq_n_u8(thresh);
+  qblimit = vdupq_n_u8(blimit);
+  qlimit = vdupq_n_u8(limit);
+  qthresh = vdupq_n_u8(thresh);
 
-    u -= (pitch << 2);
-    v -= (pitch << 2);
+  u -= (pitch << 2);
+  v -= (pitch << 2);
 
-    d6 = vld1_u8(u);
-    u += pitch;
-    d7 = vld1_u8(v);
-    v += pitch;
-    d8 = vld1_u8(u);
-    u += pitch;
-    d9 = vld1_u8(v);
-    v += pitch;
-    d10 = vld1_u8(u);
-    u += pitch;
-    d11 = vld1_u8(v);
-    v += pitch;
-    d12 = vld1_u8(u);
-    u += pitch;
-    d13 = vld1_u8(v);
-    v += pitch;
-    d14 = vld1_u8(u);
-    u += pitch;
-    d15 = vld1_u8(v);
-    v += pitch;
-    d16 = vld1_u8(u);
-    u += pitch;
-    d17 = vld1_u8(v);
-    v += pitch;
-    d18 = vld1_u8(u);
-    u += pitch;
-    d19 = vld1_u8(v);
-    v += pitch;
-    d20 = vld1_u8(u);
-    d21 = vld1_u8(v);
+  d6 = vld1_u8(u);
+  u += pitch;
+  d7 = vld1_u8(v);
+  v += pitch;
+  d8 = vld1_u8(u);
+  u += pitch;
+  d9 = vld1_u8(v);
+  v += pitch;
+  d10 = vld1_u8(u);
+  u += pitch;
+  d11 = vld1_u8(v);
+  v += pitch;
+  d12 = vld1_u8(u);
+  u += pitch;
+  d13 = vld1_u8(v);
+  v += pitch;
+  d14 = vld1_u8(u);
+  u += pitch;
+  d15 = vld1_u8(v);
+  v += pitch;
+  d16 = vld1_u8(u);
+  u += pitch;
+  d17 = vld1_u8(v);
+  v += pitch;
+  d18 = vld1_u8(u);
+  u += pitch;
+  d19 = vld1_u8(v);
+  v += pitch;
+  d20 = vld1_u8(u);
+  d21 = vld1_u8(v);
 
-    q3 = vcombine_u8(d6, d7);
-    q4 = vcombine_u8(d8, d9);
-    q5 = vcombine_u8(d10, d11);
-    q6 = vcombine_u8(d12, d13);
-    q7 = vcombine_u8(d14, d15);
-    q8 = vcombine_u8(d16, d17);
-    q9 = vcombine_u8(d18, d19);
-    q10 = vcombine_u8(d20, d21);
+  q3 = vcombine_u8(d6, d7);
+  q4 = vcombine_u8(d8, d9);
+  q5 = vcombine_u8(d10, d11);
+  q6 = vcombine_u8(d12, d13);
+  q7 = vcombine_u8(d14, d15);
+  q8 = vcombine_u8(d16, d17);
+  q9 = vcombine_u8(d18, d19);
+  q10 = vcombine_u8(d20, d21);
 
-    vp8_mbloop_filter_neon(qblimit, qlimit, qthresh, q3, q4,
-                         q5, q6, q7, q8, q9, q10,
-                         &q4, &q5, &q6, &q7, &q8, &q9);
+  vp8_mbloop_filter_neon(qblimit, qlimit, qthresh, q3, q4, q5, q6, q7, q8, q9,
+                         q10, &q4, &q5, &q6, &q7, &q8, &q9);
 
-    u -= (pitch * 6);
-    v -= (pitch * 6);
-    vst1_u8(u, vget_low_u8(q4));
-    u += pitch;
-    vst1_u8(v, vget_high_u8(q4));
-    v += pitch;
-    vst1_u8(u, vget_low_u8(q5));
-    u += pitch;
-    vst1_u8(v, vget_high_u8(q5));
-    v += pitch;
-    vst1_u8(u, vget_low_u8(q6));
-    u += pitch;
-    vst1_u8(v, vget_high_u8(q6));
-    v += pitch;
-    vst1_u8(u, vget_low_u8(q7));
-    u += pitch;
-    vst1_u8(v, vget_high_u8(q7));
-    v += pitch;
-    vst1_u8(u, vget_low_u8(q8));
-    u += pitch;
-    vst1_u8(v, vget_high_u8(q8));
-    v += pitch;
-    vst1_u8(u, vget_low_u8(q9));
-    vst1_u8(v, vget_high_u8(q9));
-    return;
+  u -= (pitch * 6);
+  v -= (pitch * 6);
+  vst1_u8(u, vget_low_u8(q4));
+  u += pitch;
+  vst1_u8(v, vget_high_u8(q4));
+  v += pitch;
+  vst1_u8(u, vget_low_u8(q5));
+  u += pitch;
+  vst1_u8(v, vget_high_u8(q5));
+  v += pitch;
+  vst1_u8(u, vget_low_u8(q6));
+  u += pitch;
+  vst1_u8(v, vget_high_u8(q6));
+  v += pitch;
+  vst1_u8(u, vget_low_u8(q7));
+  u += pitch;
+  vst1_u8(v, vget_high_u8(q7));
+  v += pitch;
+  vst1_u8(u, vget_low_u8(q8));
+  u += pitch;
+  vst1_u8(v, vget_high_u8(q8));
+  v += pitch;
+  vst1_u8(u, vget_low_u8(q9));
+  vst1_u8(v, vget_high_u8(q9));
+  return;
 }
 
-void vp8_mbloop_filter_vertical_edge_y_neon(
-        unsigned char *src,
-        int pitch,
-        unsigned char blimit,
-        unsigned char limit,
-        unsigned char thresh) {
-    unsigned char *s1, *s2;
-    uint8x16_t qblimit, qlimit, qthresh, q3, q4;
-    uint8x16_t q5, q6, q7, q8, q9, q10;
-    uint8x8_t d6, d7, d8, d9, d10, d11, d12, d13, d14;
-    uint8x8_t d15, d16, d17, d18, d19, d20, d21;
-    uint32x4x2_t q2tmp0, q2tmp1, q2tmp2, q2tmp3;
-    uint16x8x2_t q2tmp4, q2tmp5, q2tmp6, q2tmp7;
-    uint8x16x2_t q2tmp8, q2tmp9, q2tmp10, q2tmp11;
+void vp8_mbloop_filter_vertical_edge_y_neon(unsigned char *src, int pitch,
+                                            unsigned char blimit,
+                                            unsigned char limit,
+                                            unsigned char thresh) {
+  unsigned char *s1, *s2;
+  uint8x16_t qblimit, qlimit, qthresh, q3, q4;
+  uint8x16_t q5, q6, q7, q8, q9, q10;
+  uint8x8_t d6, d7, d8, d9, d10, d11, d12, d13, d14;
+  uint8x8_t d15, d16, d17, d18, d19, d20, d21;
+  uint32x4x2_t q2tmp0, q2tmp1, q2tmp2, q2tmp3;
+  uint16x8x2_t q2tmp4, q2tmp5, q2tmp6, q2tmp7;
+  uint8x16x2_t q2tmp8, q2tmp9, q2tmp10, q2tmp11;
 
-    qblimit = vdupq_n_u8(blimit);
-    qlimit = vdupq_n_u8(limit);
-    qthresh = vdupq_n_u8(thresh);
+  qblimit = vdupq_n_u8(blimit);
+  qlimit = vdupq_n_u8(limit);
+  qthresh = vdupq_n_u8(thresh);
 
-    s1 = src - 4;
-    s2 = s1 + 8 * pitch;
-    d6  = vld1_u8(s1);
-    s1 += pitch;
-    d7  = vld1_u8(s2);
-    s2 += pitch;
-    d8  = vld1_u8(s1);
-    s1 += pitch;
-    d9  = vld1_u8(s2);
-    s2 += pitch;
-    d10 = vld1_u8(s1);
-    s1 += pitch;
-    d11 = vld1_u8(s2);
-    s2 += pitch;
-    d12 = vld1_u8(s1);
-    s1 += pitch;
-    d13 = vld1_u8(s2);
-    s2 += pitch;
-    d14 = vld1_u8(s1);
-    s1 += pitch;
-    d15 = vld1_u8(s2);
-    s2 += pitch;
-    d16 = vld1_u8(s1);
-    s1 += pitch;
-    d17 = vld1_u8(s2);
-    s2 += pitch;
-    d18 = vld1_u8(s1);
-    s1 += pitch;
-    d19 = vld1_u8(s2);
-    s2 += pitch;
-    d20 = vld1_u8(s1);
-    d21 = vld1_u8(s2);
+  s1 = src - 4;
+  s2 = s1 + 8 * pitch;
+  d6 = vld1_u8(s1);
+  s1 += pitch;
+  d7 = vld1_u8(s2);
+  s2 += pitch;
+  d8 = vld1_u8(s1);
+  s1 += pitch;
+  d9 = vld1_u8(s2);
+  s2 += pitch;
+  d10 = vld1_u8(s1);
+  s1 += pitch;
+  d11 = vld1_u8(s2);
+  s2 += pitch;
+  d12 = vld1_u8(s1);
+  s1 += pitch;
+  d13 = vld1_u8(s2);
+  s2 += pitch;
+  d14 = vld1_u8(s1);
+  s1 += pitch;
+  d15 = vld1_u8(s2);
+  s2 += pitch;
+  d16 = vld1_u8(s1);
+  s1 += pitch;
+  d17 = vld1_u8(s2);
+  s2 += pitch;
+  d18 = vld1_u8(s1);
+  s1 += pitch;
+  d19 = vld1_u8(s2);
+  s2 += pitch;
+  d20 = vld1_u8(s1);
+  d21 = vld1_u8(s2);
 
-    q3 = vcombine_u8(d6, d7);
-    q4 = vcombine_u8(d8, d9);
-    q5 = vcombine_u8(d10, d11);
-    q6 = vcombine_u8(d12, d13);
-    q7 = vcombine_u8(d14, d15);
-    q8 = vcombine_u8(d16, d17);
-    q9 = vcombine_u8(d18, d19);
-    q10 = vcombine_u8(d20, d21);
+  q3 = vcombine_u8(d6, d7);
+  q4 = vcombine_u8(d8, d9);
+  q5 = vcombine_u8(d10, d11);
+  q6 = vcombine_u8(d12, d13);
+  q7 = vcombine_u8(d14, d15);
+  q8 = vcombine_u8(d16, d17);
+  q9 = vcombine_u8(d18, d19);
+  q10 = vcombine_u8(d20, d21);
 
-    q2tmp0 = vtrnq_u32(vreinterpretq_u32_u8(q3), vreinterpretq_u32_u8(q7));
-    q2tmp1 = vtrnq_u32(vreinterpretq_u32_u8(q4), vreinterpretq_u32_u8(q8));
-    q2tmp2 = vtrnq_u32(vreinterpretq_u32_u8(q5), vreinterpretq_u32_u8(q9));
-    q2tmp3 = vtrnq_u32(vreinterpretq_u32_u8(q6), vreinterpretq_u32_u8(q10));
+  q2tmp0 = vtrnq_u32(vreinterpretq_u32_u8(q3), vreinterpretq_u32_u8(q7));
+  q2tmp1 = vtrnq_u32(vreinterpretq_u32_u8(q4), vreinterpretq_u32_u8(q8));
+  q2tmp2 = vtrnq_u32(vreinterpretq_u32_u8(q5), vreinterpretq_u32_u8(q9));
+  q2tmp3 = vtrnq_u32(vreinterpretq_u32_u8(q6), vreinterpretq_u32_u8(q10));
 
-    q2tmp4 = vtrnq_u16(vreinterpretq_u16_u32(q2tmp0.val[0]),
-                       vreinterpretq_u16_u32(q2tmp2.val[0]));
-    q2tmp5 = vtrnq_u16(vreinterpretq_u16_u32(q2tmp1.val[0]),
-                       vreinterpretq_u16_u32(q2tmp3.val[0]));
-    q2tmp6 = vtrnq_u16(vreinterpretq_u16_u32(q2tmp0.val[1]),
-                       vreinterpretq_u16_u32(q2tmp2.val[1]));
-    q2tmp7 = vtrnq_u16(vreinterpretq_u16_u32(q2tmp1.val[1]),
-                       vreinterpretq_u16_u32(q2tmp3.val[1]));
+  q2tmp4 = vtrnq_u16(vreinterpretq_u16_u32(q2tmp0.val[0]),
+                     vreinterpretq_u16_u32(q2tmp2.val[0]));
+  q2tmp5 = vtrnq_u16(vreinterpretq_u16_u32(q2tmp1.val[0]),
+                     vreinterpretq_u16_u32(q2tmp3.val[0]));
+  q2tmp6 = vtrnq_u16(vreinterpretq_u16_u32(q2tmp0.val[1]),
+                     vreinterpretq_u16_u32(q2tmp2.val[1]));
+  q2tmp7 = vtrnq_u16(vreinterpretq_u16_u32(q2tmp1.val[1]),
+                     vreinterpretq_u16_u32(q2tmp3.val[1]));
 
-    q2tmp8  = vtrnq_u8(vreinterpretq_u8_u16(q2tmp4.val[0]),
-                       vreinterpretq_u8_u16(q2tmp5.val[0]));
-    q2tmp9  = vtrnq_u8(vreinterpretq_u8_u16(q2tmp4.val[1]),
-                       vreinterpretq_u8_u16(q2tmp5.val[1]));
-    q2tmp10 = vtrnq_u8(vreinterpretq_u8_u16(q2tmp6.val[0]),
-                       vreinterpretq_u8_u16(q2tmp7.val[0]));
-    q2tmp11 = vtrnq_u8(vreinterpretq_u8_u16(q2tmp6.val[1]),
-                       vreinterpretq_u8_u16(q2tmp7.val[1]));
+  q2tmp8 = vtrnq_u8(vreinterpretq_u8_u16(q2tmp4.val[0]),
+                    vreinterpretq_u8_u16(q2tmp5.val[0]));
+  q2tmp9 = vtrnq_u8(vreinterpretq_u8_u16(q2tmp4.val[1]),
+                    vreinterpretq_u8_u16(q2tmp5.val[1]));
+  q2tmp10 = vtrnq_u8(vreinterpretq_u8_u16(q2tmp6.val[0]),
+                     vreinterpretq_u8_u16(q2tmp7.val[0]));
+  q2tmp11 = vtrnq_u8(vreinterpretq_u8_u16(q2tmp6.val[1]),
+                     vreinterpretq_u8_u16(q2tmp7.val[1]));
 
-    q3 = q2tmp8.val[0];
-    q4 = q2tmp8.val[1];
-    q5 = q2tmp9.val[0];
-    q6 = q2tmp9.val[1];
-    q7 = q2tmp10.val[0];
-    q8 = q2tmp10.val[1];
-    q9 = q2tmp11.val[0];
-    q10 = q2tmp11.val[1];
+  q3 = q2tmp8.val[0];
+  q4 = q2tmp8.val[1];
+  q5 = q2tmp9.val[0];
+  q6 = q2tmp9.val[1];
+  q7 = q2tmp10.val[0];
+  q8 = q2tmp10.val[1];
+  q9 = q2tmp11.val[0];
+  q10 = q2tmp11.val[1];
 
-    vp8_mbloop_filter_neon(qblimit, qlimit, qthresh, q3, q4,
-                         q5, q6, q7, q8, q9, q10,
-                         &q4, &q5, &q6, &q7, &q8, &q9);
+  vp8_mbloop_filter_neon(qblimit, qlimit, qthresh, q3, q4, q5, q6, q7, q8, q9,
+                         q10, &q4, &q5, &q6, &q7, &q8, &q9);
 
-    q2tmp0 = vtrnq_u32(vreinterpretq_u32_u8(q3), vreinterpretq_u32_u8(q7));
-    q2tmp1 = vtrnq_u32(vreinterpretq_u32_u8(q4), vreinterpretq_u32_u8(q8));
-    q2tmp2 = vtrnq_u32(vreinterpretq_u32_u8(q5), vreinterpretq_u32_u8(q9));
-    q2tmp3 = vtrnq_u32(vreinterpretq_u32_u8(q6), vreinterpretq_u32_u8(q10));
+  q2tmp0 = vtrnq_u32(vreinterpretq_u32_u8(q3), vreinterpretq_u32_u8(q7));
+  q2tmp1 = vtrnq_u32(vreinterpretq_u32_u8(q4), vreinterpretq_u32_u8(q8));
+  q2tmp2 = vtrnq_u32(vreinterpretq_u32_u8(q5), vreinterpretq_u32_u8(q9));
+  q2tmp3 = vtrnq_u32(vreinterpretq_u32_u8(q6), vreinterpretq_u32_u8(q10));
 
-    q2tmp4 = vtrnq_u16(vreinterpretq_u16_u32(q2tmp0.val[0]),
-                       vreinterpretq_u16_u32(q2tmp2.val[0]));
-    q2tmp5 = vtrnq_u16(vreinterpretq_u16_u32(q2tmp1.val[0]),
-                       vreinterpretq_u16_u32(q2tmp3.val[0]));
-    q2tmp6 = vtrnq_u16(vreinterpretq_u16_u32(q2tmp0.val[1]),
-                       vreinterpretq_u16_u32(q2tmp2.val[1]));
-    q2tmp7 = vtrnq_u16(vreinterpretq_u16_u32(q2tmp1.val[1]),
-                       vreinterpretq_u16_u32(q2tmp3.val[1]));
+  q2tmp4 = vtrnq_u16(vreinterpretq_u16_u32(q2tmp0.val[0]),
+                     vreinterpretq_u16_u32(q2tmp2.val[0]));
+  q2tmp5 = vtrnq_u16(vreinterpretq_u16_u32(q2tmp1.val[0]),
+                     vreinterpretq_u16_u32(q2tmp3.val[0]));
+  q2tmp6 = vtrnq_u16(vreinterpretq_u16_u32(q2tmp0.val[1]),
+                     vreinterpretq_u16_u32(q2tmp2.val[1]));
+  q2tmp7 = vtrnq_u16(vreinterpretq_u16_u32(q2tmp1.val[1]),
+                     vreinterpretq_u16_u32(q2tmp3.val[1]));
 
-    q2tmp8  = vtrnq_u8(vreinterpretq_u8_u16(q2tmp4.val[0]),
-                       vreinterpretq_u8_u16(q2tmp5.val[0]));
-    q2tmp9  = vtrnq_u8(vreinterpretq_u8_u16(q2tmp4.val[1]),
-                       vreinterpretq_u8_u16(q2tmp5.val[1]));
-    q2tmp10 = vtrnq_u8(vreinterpretq_u8_u16(q2tmp6.val[0]),
-                       vreinterpretq_u8_u16(q2tmp7.val[0]));
-    q2tmp11 = vtrnq_u8(vreinterpretq_u8_u16(q2tmp6.val[1]),
-                       vreinterpretq_u8_u16(q2tmp7.val[1]));
+  q2tmp8 = vtrnq_u8(vreinterpretq_u8_u16(q2tmp4.val[0]),
+                    vreinterpretq_u8_u16(q2tmp5.val[0]));
+  q2tmp9 = vtrnq_u8(vreinterpretq_u8_u16(q2tmp4.val[1]),
+                    vreinterpretq_u8_u16(q2tmp5.val[1]));
+  q2tmp10 = vtrnq_u8(vreinterpretq_u8_u16(q2tmp6.val[0]),
+                     vreinterpretq_u8_u16(q2tmp7.val[0]));
+  q2tmp11 = vtrnq_u8(vreinterpretq_u8_u16(q2tmp6.val[1]),
+                     vreinterpretq_u8_u16(q2tmp7.val[1]));
 
-    q3 = q2tmp8.val[0];
-    q4 = q2tmp8.val[1];
-    q5 = q2tmp9.val[0];
-    q6 = q2tmp9.val[1];
-    q7 = q2tmp10.val[0];
-    q8 = q2tmp10.val[1];
-    q9 = q2tmp11.val[0];
-    q10 = q2tmp11.val[1];
+  q3 = q2tmp8.val[0];
+  q4 = q2tmp8.val[1];
+  q5 = q2tmp9.val[0];
+  q6 = q2tmp9.val[1];
+  q7 = q2tmp10.val[0];
+  q8 = q2tmp10.val[1];
+  q9 = q2tmp11.val[0];
+  q10 = q2tmp11.val[1];
 
-    s1 -= 7 * pitch;
-    s2 -= 7 * pitch;
+  s1 -= 7 * pitch;
+  s2 -= 7 * pitch;
 
-    vst1_u8(s1, vget_low_u8(q3));
-    s1 += pitch;
-    vst1_u8(s2, vget_high_u8(q3));
-    s2 += pitch;
-    vst1_u8(s1, vget_low_u8(q4));
-    s1 += pitch;
-    vst1_u8(s2, vget_high_u8(q4));
-    s2 += pitch;
-    vst1_u8(s1, vget_low_u8(q5));
-    s1 += pitch;
-    vst1_u8(s2, vget_high_u8(q5));
-    s2 += pitch;
-    vst1_u8(s1, vget_low_u8(q6));
-    s1 += pitch;
-    vst1_u8(s2, vget_high_u8(q6));
-    s2 += pitch;
-    vst1_u8(s1, vget_low_u8(q7));
-    s1 += pitch;
-    vst1_u8(s2, vget_high_u8(q7));
-    s2 += pitch;
-    vst1_u8(s1, vget_low_u8(q8));
-    s1 += pitch;
-    vst1_u8(s2, vget_high_u8(q8));
-    s2 += pitch;
-    vst1_u8(s1, vget_low_u8(q9));
-    s1 += pitch;
-    vst1_u8(s2, vget_high_u8(q9));
-    s2 += pitch;
-    vst1_u8(s1, vget_low_u8(q10));
-    vst1_u8(s2, vget_high_u8(q10));
-    return;
+  vst1_u8(s1, vget_low_u8(q3));
+  s1 += pitch;
+  vst1_u8(s2, vget_high_u8(q3));
+  s2 += pitch;
+  vst1_u8(s1, vget_low_u8(q4));
+  s1 += pitch;
+  vst1_u8(s2, vget_high_u8(q4));
+  s2 += pitch;
+  vst1_u8(s1, vget_low_u8(q5));
+  s1 += pitch;
+  vst1_u8(s2, vget_high_u8(q5));
+  s2 += pitch;
+  vst1_u8(s1, vget_low_u8(q6));
+  s1 += pitch;
+  vst1_u8(s2, vget_high_u8(q6));
+  s2 += pitch;
+  vst1_u8(s1, vget_low_u8(q7));
+  s1 += pitch;
+  vst1_u8(s2, vget_high_u8(q7));
+  s2 += pitch;
+  vst1_u8(s1, vget_low_u8(q8));
+  s1 += pitch;
+  vst1_u8(s2, vget_high_u8(q8));
+  s2 += pitch;
+  vst1_u8(s1, vget_low_u8(q9));
+  s1 += pitch;
+  vst1_u8(s2, vget_high_u8(q9));
+  s2 += pitch;
+  vst1_u8(s1, vget_low_u8(q10));
+  vst1_u8(s2, vget_high_u8(q10));
+  return;
 }
 
-void vp8_mbloop_filter_vertical_edge_uv_neon(
-        unsigned char *u,
-        int pitch,
-        unsigned char blimit,
-        unsigned char limit,
-        unsigned char thresh,
-        unsigned char *v) {
-    unsigned char *us, *ud;
-    unsigned char *vs, *vd;
-    uint8x16_t qblimit, qlimit, qthresh, q3, q4;
-    uint8x16_t q5, q6, q7, q8, q9, q10;
-    uint8x8_t d6, d7, d8, d9, d10, d11, d12, d13, d14;
-    uint8x8_t d15, d16, d17, d18, d19, d20, d21;
-    uint32x4x2_t q2tmp0, q2tmp1, q2tmp2, q2tmp3;
-    uint16x8x2_t q2tmp4, q2tmp5, q2tmp6, q2tmp7;
-    uint8x16x2_t q2tmp8, q2tmp9, q2tmp10, q2tmp11;
+void vp8_mbloop_filter_vertical_edge_uv_neon(unsigned char *u, int pitch,
+                                             unsigned char blimit,
+                                             unsigned char limit,
+                                             unsigned char thresh,
+                                             unsigned char *v) {
+  unsigned char *us, *ud;
+  unsigned char *vs, *vd;
+  uint8x16_t qblimit, qlimit, qthresh, q3, q4;
+  uint8x16_t q5, q6, q7, q8, q9, q10;
+  uint8x8_t d6, d7, d8, d9, d10, d11, d12, d13, d14;
+  uint8x8_t d15, d16, d17, d18, d19, d20, d21;
+  uint32x4x2_t q2tmp0, q2tmp1, q2tmp2, q2tmp3;
+  uint16x8x2_t q2tmp4, q2tmp5, q2tmp6, q2tmp7;
+  uint8x16x2_t q2tmp8, q2tmp9, q2tmp10, q2tmp11;
 
-    qblimit = vdupq_n_u8(blimit);
-    qlimit = vdupq_n_u8(limit);
-    qthresh = vdupq_n_u8(thresh);
+  qblimit = vdupq_n_u8(blimit);
+  qlimit = vdupq_n_u8(limit);
+  qthresh = vdupq_n_u8(thresh);
 
-    us = u - 4;
-    vs = v - 4;
-    d6 = vld1_u8(us);
-    us += pitch;
-    d7 = vld1_u8(vs);
-    vs += pitch;
-    d8 = vld1_u8(us);
-    us += pitch;
-    d9 = vld1_u8(vs);
-    vs += pitch;
-    d10 = vld1_u8(us);
-    us += pitch;
-    d11 = vld1_u8(vs);
-    vs += pitch;
-    d12 = vld1_u8(us);
-    us += pitch;
-    d13 = vld1_u8(vs);
-    vs += pitch;
-    d14 = vld1_u8(us);
-    us += pitch;
-    d15 = vld1_u8(vs);
-    vs += pitch;
-    d16 = vld1_u8(us);
-    us += pitch;
-    d17 = vld1_u8(vs);
-    vs += pitch;
-    d18 = vld1_u8(us);
-    us += pitch;
-    d19 = vld1_u8(vs);
-    vs += pitch;
-    d20 = vld1_u8(us);
-    d21 = vld1_u8(vs);
+  us = u - 4;
+  vs = v - 4;
+  d6 = vld1_u8(us);
+  us += pitch;
+  d7 = vld1_u8(vs);
+  vs += pitch;
+  d8 = vld1_u8(us);
+  us += pitch;
+  d9 = vld1_u8(vs);
+  vs += pitch;
+  d10 = vld1_u8(us);
+  us += pitch;
+  d11 = vld1_u8(vs);
+  vs += pitch;
+  d12 = vld1_u8(us);
+  us += pitch;
+  d13 = vld1_u8(vs);
+  vs += pitch;
+  d14 = vld1_u8(us);
+  us += pitch;
+  d15 = vld1_u8(vs);
+  vs += pitch;
+  d16 = vld1_u8(us);
+  us += pitch;
+  d17 = vld1_u8(vs);
+  vs += pitch;
+  d18 = vld1_u8(us);
+  us += pitch;
+  d19 = vld1_u8(vs);
+  vs += pitch;
+  d20 = vld1_u8(us);
+  d21 = vld1_u8(vs);
 
-    q3 = vcombine_u8(d6, d7);
-    q4 = vcombine_u8(d8, d9);
-    q5 = vcombine_u8(d10, d11);
-    q6 = vcombine_u8(d12, d13);
-    q7 = vcombine_u8(d14, d15);
-    q8 = vcombine_u8(d16, d17);
-    q9 = vcombine_u8(d18, d19);
-    q10 = vcombine_u8(d20, d21);
+  q3 = vcombine_u8(d6, d7);
+  q4 = vcombine_u8(d8, d9);
+  q5 = vcombine_u8(d10, d11);
+  q6 = vcombine_u8(d12, d13);
+  q7 = vcombine_u8(d14, d15);
+  q8 = vcombine_u8(d16, d17);
+  q9 = vcombine_u8(d18, d19);
+  q10 = vcombine_u8(d20, d21);
 
-    q2tmp0 = vtrnq_u32(vreinterpretq_u32_u8(q3), vreinterpretq_u32_u8(q7));
-    q2tmp1 = vtrnq_u32(vreinterpretq_u32_u8(q4), vreinterpretq_u32_u8(q8));
-    q2tmp2 = vtrnq_u32(vreinterpretq_u32_u8(q5), vreinterpretq_u32_u8(q9));
-    q2tmp3 = vtrnq_u32(vreinterpretq_u32_u8(q6), vreinterpretq_u32_u8(q10));
+  q2tmp0 = vtrnq_u32(vreinterpretq_u32_u8(q3), vreinterpretq_u32_u8(q7));
+  q2tmp1 = vtrnq_u32(vreinterpretq_u32_u8(q4), vreinterpretq_u32_u8(q8));
+  q2tmp2 = vtrnq_u32(vreinterpretq_u32_u8(q5), vreinterpretq_u32_u8(q9));
+  q2tmp3 = vtrnq_u32(vreinterpretq_u32_u8(q6), vreinterpretq_u32_u8(q10));
 
-    q2tmp4 = vtrnq_u16(vreinterpretq_u16_u32(q2tmp0.val[0]),
-                       vreinterpretq_u16_u32(q2tmp2.val[0]));
-    q2tmp5 = vtrnq_u16(vreinterpretq_u16_u32(q2tmp1.val[0]),
-                       vreinterpretq_u16_u32(q2tmp3.val[0]));
-    q2tmp6 = vtrnq_u16(vreinterpretq_u16_u32(q2tmp0.val[1]),
-                       vreinterpretq_u16_u32(q2tmp2.val[1]));
-    q2tmp7 = vtrnq_u16(vreinterpretq_u16_u32(q2tmp1.val[1]),
-                       vreinterpretq_u16_u32(q2tmp3.val[1]));
+  q2tmp4 = vtrnq_u16(vreinterpretq_u16_u32(q2tmp0.val[0]),
+                     vreinterpretq_u16_u32(q2tmp2.val[0]));
+  q2tmp5 = vtrnq_u16(vreinterpretq_u16_u32(q2tmp1.val[0]),
+                     vreinterpretq_u16_u32(q2tmp3.val[0]));
+  q2tmp6 = vtrnq_u16(vreinterpretq_u16_u32(q2tmp0.val[1]),
+                     vreinterpretq_u16_u32(q2tmp2.val[1]));
+  q2tmp7 = vtrnq_u16(vreinterpretq_u16_u32(q2tmp1.val[1]),
+                     vreinterpretq_u16_u32(q2tmp3.val[1]));
 
-    q2tmp8  = vtrnq_u8(vreinterpretq_u8_u16(q2tmp4.val[0]),
-                       vreinterpretq_u8_u16(q2tmp5.val[0]));
-    q2tmp9  = vtrnq_u8(vreinterpretq_u8_u16(q2tmp4.val[1]),
-                       vreinterpretq_u8_u16(q2tmp5.val[1]));
-    q2tmp10 = vtrnq_u8(vreinterpretq_u8_u16(q2tmp6.val[0]),
-                       vreinterpretq_u8_u16(q2tmp7.val[0]));
-    q2tmp11 = vtrnq_u8(vreinterpretq_u8_u16(q2tmp6.val[1]),
-                       vreinterpretq_u8_u16(q2tmp7.val[1]));
+  q2tmp8 = vtrnq_u8(vreinterpretq_u8_u16(q2tmp4.val[0]),
+                    vreinterpretq_u8_u16(q2tmp5.val[0]));
+  q2tmp9 = vtrnq_u8(vreinterpretq_u8_u16(q2tmp4.val[1]),
+                    vreinterpretq_u8_u16(q2tmp5.val[1]));
+  q2tmp10 = vtrnq_u8(vreinterpretq_u8_u16(q2tmp6.val[0]),
+                     vreinterpretq_u8_u16(q2tmp7.val[0]));
+  q2tmp11 = vtrnq_u8(vreinterpretq_u8_u16(q2tmp6.val[1]),
+                     vreinterpretq_u8_u16(q2tmp7.val[1]));
 
-    q3 = q2tmp8.val[0];
-    q4 = q2tmp8.val[1];
-    q5 = q2tmp9.val[0];
-    q6 = q2tmp9.val[1];
-    q7 = q2tmp10.val[0];
-    q8 = q2tmp10.val[1];
-    q9 = q2tmp11.val[0];
-    q10 = q2tmp11.val[1];
+  q3 = q2tmp8.val[0];
+  q4 = q2tmp8.val[1];
+  q5 = q2tmp9.val[0];
+  q6 = q2tmp9.val[1];
+  q7 = q2tmp10.val[0];
+  q8 = q2tmp10.val[1];
+  q9 = q2tmp11.val[0];
+  q10 = q2tmp11.val[1];
 
-    vp8_mbloop_filter_neon(qblimit, qlimit, qthresh, q3, q4,
-                         q5, q6, q7, q8, q9, q10,
-                         &q4, &q5, &q6, &q7, &q8, &q9);
+  vp8_mbloop_filter_neon(qblimit, qlimit, qthresh, q3, q4, q5, q6, q7, q8, q9,
+                         q10, &q4, &q5, &q6, &q7, &q8, &q9);
 
-    q2tmp0 = vtrnq_u32(vreinterpretq_u32_u8(q3), vreinterpretq_u32_u8(q7));
-    q2tmp1 = vtrnq_u32(vreinterpretq_u32_u8(q4), vreinterpretq_u32_u8(q8));
-    q2tmp2 = vtrnq_u32(vreinterpretq_u32_u8(q5), vreinterpretq_u32_u8(q9));
-    q2tmp3 = vtrnq_u32(vreinterpretq_u32_u8(q6), vreinterpretq_u32_u8(q10));
+  q2tmp0 = vtrnq_u32(vreinterpretq_u32_u8(q3), vreinterpretq_u32_u8(q7));
+  q2tmp1 = vtrnq_u32(vreinterpretq_u32_u8(q4), vreinterpretq_u32_u8(q8));
+  q2tmp2 = vtrnq_u32(vreinterpretq_u32_u8(q5), vreinterpretq_u32_u8(q9));
+  q2tmp3 = vtrnq_u32(vreinterpretq_u32_u8(q6), vreinterpretq_u32_u8(q10));
 
-    q2tmp4 = vtrnq_u16(vreinterpretq_u16_u32(q2tmp0.val[0]),
-                       vreinterpretq_u16_u32(q2tmp2.val[0]));
-    q2tmp5 = vtrnq_u16(vreinterpretq_u16_u32(q2tmp1.val[0]),
-                       vreinterpretq_u16_u32(q2tmp3.val[0]));
-    q2tmp6 = vtrnq_u16(vreinterpretq_u16_u32(q2tmp0.val[1]),
-                       vreinterpretq_u16_u32(q2tmp2.val[1]));
-    q2tmp7 = vtrnq_u16(vreinterpretq_u16_u32(q2tmp1.val[1]),
-                       vreinterpretq_u16_u32(q2tmp3.val[1]));
+  q2tmp4 = vtrnq_u16(vreinterpretq_u16_u32(q2tmp0.val[0]),
+                     vreinterpretq_u16_u32(q2tmp2.val[0]));
+  q2tmp5 = vtrnq_u16(vreinterpretq_u16_u32(q2tmp1.val[0]),
+                     vreinterpretq_u16_u32(q2tmp3.val[0]));
+  q2tmp6 = vtrnq_u16(vreinterpretq_u16_u32(q2tmp0.val[1]),
+                     vreinterpretq_u16_u32(q2tmp2.val[1]));
+  q2tmp7 = vtrnq_u16(vreinterpretq_u16_u32(q2tmp1.val[1]),
+                     vreinterpretq_u16_u32(q2tmp3.val[1]));
 
-    q2tmp8  = vtrnq_u8(vreinterpretq_u8_u16(q2tmp4.val[0]),
-                       vreinterpretq_u8_u16(q2tmp5.val[0]));
-    q2tmp9  = vtrnq_u8(vreinterpretq_u8_u16(q2tmp4.val[1]),
-                       vreinterpretq_u8_u16(q2tmp5.val[1]));
-    q2tmp10 = vtrnq_u8(vreinterpretq_u8_u16(q2tmp6.val[0]),
-                       vreinterpretq_u8_u16(q2tmp7.val[0]));
-    q2tmp11 = vtrnq_u8(vreinterpretq_u8_u16(q2tmp6.val[1]),
-                       vreinterpretq_u8_u16(q2tmp7.val[1]));
+  q2tmp8 = vtrnq_u8(vreinterpretq_u8_u16(q2tmp4.val[0]),
+                    vreinterpretq_u8_u16(q2tmp5.val[0]));
+  q2tmp9 = vtrnq_u8(vreinterpretq_u8_u16(q2tmp4.val[1]),
+                    vreinterpretq_u8_u16(q2tmp5.val[1]));
+  q2tmp10 = vtrnq_u8(vreinterpretq_u8_u16(q2tmp6.val[0]),
+                     vreinterpretq_u8_u16(q2tmp7.val[0]));
+  q2tmp11 = vtrnq_u8(vreinterpretq_u8_u16(q2tmp6.val[1]),
+                     vreinterpretq_u8_u16(q2tmp7.val[1]));
 
-    q3 = q2tmp8.val[0];
-    q4 = q2tmp8.val[1];
-    q5 = q2tmp9.val[0];
-    q6 = q2tmp9.val[1];
-    q7 = q2tmp10.val[0];
-    q8 = q2tmp10.val[1];
-    q9 = q2tmp11.val[0];
-    q10 = q2tmp11.val[1];
+  q3 = q2tmp8.val[0];
+  q4 = q2tmp8.val[1];
+  q5 = q2tmp9.val[0];
+  q6 = q2tmp9.val[1];
+  q7 = q2tmp10.val[0];
+  q8 = q2tmp10.val[1];
+  q9 = q2tmp11.val[0];
+  q10 = q2tmp11.val[1];
 
-    ud = u - 4;
-    vst1_u8(ud, vget_low_u8(q3));
-    ud += pitch;
-    vst1_u8(ud, vget_low_u8(q4));
-    ud += pitch;
-    vst1_u8(ud, vget_low_u8(q5));
-    ud += pitch;
-    vst1_u8(ud, vget_low_u8(q6));
-    ud += pitch;
-    vst1_u8(ud, vget_low_u8(q7));
-    ud += pitch;
-    vst1_u8(ud, vget_low_u8(q8));
-    ud += pitch;
-    vst1_u8(ud, vget_low_u8(q9));
-    ud += pitch;
-    vst1_u8(ud, vget_low_u8(q10));
+  ud = u - 4;
+  vst1_u8(ud, vget_low_u8(q3));
+  ud += pitch;
+  vst1_u8(ud, vget_low_u8(q4));
+  ud += pitch;
+  vst1_u8(ud, vget_low_u8(q5));
+  ud += pitch;
+  vst1_u8(ud, vget_low_u8(q6));
+  ud += pitch;
+  vst1_u8(ud, vget_low_u8(q7));
+  ud += pitch;
+  vst1_u8(ud, vget_low_u8(q8));
+  ud += pitch;
+  vst1_u8(ud, vget_low_u8(q9));
+  ud += pitch;
+  vst1_u8(ud, vget_low_u8(q10));
 
-    vd = v - 4;
-    vst1_u8(vd, vget_high_u8(q3));
-    vd += pitch;
-    vst1_u8(vd, vget_high_u8(q4));
-    vd += pitch;
-    vst1_u8(vd, vget_high_u8(q5));
-    vd += pitch;
-    vst1_u8(vd, vget_high_u8(q6));
-    vd += pitch;
-    vst1_u8(vd, vget_high_u8(q7));
-    vd += pitch;
-    vst1_u8(vd, vget_high_u8(q8));
-    vd += pitch;
-    vst1_u8(vd, vget_high_u8(q9));
-    vd += pitch;
-    vst1_u8(vd, vget_high_u8(q10));
-    return;
+  vd = v - 4;
+  vst1_u8(vd, vget_high_u8(q3));
+  vd += pitch;
+  vst1_u8(vd, vget_high_u8(q4));
+  vd += pitch;
+  vst1_u8(vd, vget_high_u8(q5));
+  vd += pitch;
+  vst1_u8(vd, vget_high_u8(q6));
+  vd += pitch;
+  vst1_u8(vd, vget_high_u8(q7));
+  vd += pitch;
+  vst1_u8(vd, vget_high_u8(q8));
+  vd += pitch;
+  vst1_u8(vd, vget_high_u8(q9));
+  vd += pitch;
+  vst1_u8(vd, vget_high_u8(q10));
+  return;
 }
--- a/vp8/common/arm/neon/shortidct4x4llm_neon.c
+++ b/vp8/common/arm/neon/shortidct4x4llm_neon.c
@@ -11,113 +11,109 @@
 #include <arm_neon.h>
 
 static const int16_t cospi8sqrt2minus1 = 20091;
-static const int16_t sinpi8sqrt2       = 35468;
+static const int16_t sinpi8sqrt2 = 35468;
 
-void vp8_short_idct4x4llm_neon(
-        int16_t *input,
-        unsigned char *pred_ptr,
-        int pred_stride,
-        unsigned char *dst_ptr,
-        int dst_stride) {
-    int i;
-    uint32x2_t d6u32 = vdup_n_u32(0);
-    uint8x8_t d1u8;
-    int16x4_t d2, d3, d4, d5, d10, d11, d12, d13;
-    uint16x8_t q1u16;
-    int16x8_t q1s16, q2s16, q3s16, q4s16;
-    int32x2x2_t v2tmp0, v2tmp1;
-    int16x4x2_t v2tmp2, v2tmp3;
+void vp8_short_idct4x4llm_neon(int16_t *input, unsigned char *pred_ptr,
+                               int pred_stride, unsigned char *dst_ptr,
+                               int dst_stride) {
+  int i;
+  uint32x2_t d6u32 = vdup_n_u32(0);
+  uint8x8_t d1u8;
+  int16x4_t d2, d3, d4, d5, d10, d11, d12, d13;
+  uint16x8_t q1u16;
+  int16x8_t q1s16, q2s16, q3s16, q4s16;
+  int32x2x2_t v2tmp0, v2tmp1;
+  int16x4x2_t v2tmp2, v2tmp3;
 
-    d2 = vld1_s16(input);
-    d3 = vld1_s16(input + 4);
-    d4 = vld1_s16(input + 8);
-    d5 = vld1_s16(input + 12);
+  d2 = vld1_s16(input);
+  d3 = vld1_s16(input + 4);
+  d4 = vld1_s16(input + 8);
+  d5 = vld1_s16(input + 12);
 
-    // 1st for loop
-    q1s16 = vcombine_s16(d2, d4);  // Swap d3 d4 here
-    q2s16 = vcombine_s16(d3, d5);
+  // 1st for loop
+  q1s16 = vcombine_s16(d2, d4);  // Swap d3 d4 here
+  q2s16 = vcombine_s16(d3, d5);
 
-    q3s16 = vqdmulhq_n_s16(q2s16, sinpi8sqrt2);
-    q4s16 = vqdmulhq_n_s16(q2s16, cospi8sqrt2minus1);
+  q3s16 = vqdmulhq_n_s16(q2s16, sinpi8sqrt2);
+  q4s16 = vqdmulhq_n_s16(q2s16, cospi8sqrt2minus1);
 
-    d12 = vqadd_s16(vget_low_s16(q1s16), vget_high_s16(q1s16));  // a1
-    d13 = vqsub_s16(vget_low_s16(q1s16), vget_high_s16(q1s16));  // b1
+  d12 = vqadd_s16(vget_low_s16(q1s16), vget_high_s16(q1s16));  // a1
+  d13 = vqsub_s16(vget_low_s16(q1s16), vget_high_s16(q1s16));  // b1
 
-    q3s16 = vshrq_n_s16(q3s16, 1);
-    q4s16 = vshrq_n_s16(q4s16, 1);
+  q3s16 = vshrq_n_s16(q3s16, 1);
+  q4s16 = vshrq_n_s16(q4s16, 1);
 
-    q3s16 = vqaddq_s16(q3s16, q2s16);
-    q4s16 = vqaddq_s16(q4s16, q2s16);
+  q3s16 = vqaddq_s16(q3s16, q2s16);
+  q4s16 = vqaddq_s16(q4s16, q2s16);
 
-    d10 = vqsub_s16(vget_low_s16(q3s16), vget_high_s16(q4s16));  // c1
-    d11 = vqadd_s16(vget_high_s16(q3s16), vget_low_s16(q4s16));  // d1
+  d10 = vqsub_s16(vget_low_s16(q3s16), vget_high_s16(q4s16));  // c1
+  d11 = vqadd_s16(vget_high_s16(q3s16), vget_low_s16(q4s16));  // d1
 
-    d2 = vqadd_s16(d12, d11);
-    d3 = vqadd_s16(d13, d10);
-    d4 = vqsub_s16(d13, d10);
-    d5 = vqsub_s16(d12, d11);
+  d2 = vqadd_s16(d12, d11);
+  d3 = vqadd_s16(d13, d10);
+  d4 = vqsub_s16(d13, d10);
+  d5 = vqsub_s16(d12, d11);
 
-    v2tmp0 = vtrn_s32(vreinterpret_s32_s16(d2), vreinterpret_s32_s16(d4));
-    v2tmp1 = vtrn_s32(vreinterpret_s32_s16(d3), vreinterpret_s32_s16(d5));
-    v2tmp2 = vtrn_s16(vreinterpret_s16_s32(v2tmp0.val[0]),
-                      vreinterpret_s16_s32(v2tmp1.val[0]));
-    v2tmp3 = vtrn_s16(vreinterpret_s16_s32(v2tmp0.val[1]),
-                      vreinterpret_s16_s32(v2tmp1.val[1]));
+  v2tmp0 = vtrn_s32(vreinterpret_s32_s16(d2), vreinterpret_s32_s16(d4));
+  v2tmp1 = vtrn_s32(vreinterpret_s32_s16(d3), vreinterpret_s32_s16(d5));
+  v2tmp2 = vtrn_s16(vreinterpret_s16_s32(v2tmp0.val[0]),
+                    vreinterpret_s16_s32(v2tmp1.val[0]));
+  v2tmp3 = vtrn_s16(vreinterpret_s16_s32(v2tmp0.val[1]),
+                    vreinterpret_s16_s32(v2tmp1.val[1]));
 
-    // 2nd for loop
-    q1s16 = vcombine_s16(v2tmp2.val[0], v2tmp3.val[0]);
-    q2s16 = vcombine_s16(v2tmp2.val[1], v2tmp3.val[1]);
+  // 2nd for loop
+  q1s16 = vcombine_s16(v2tmp2.val[0], v2tmp3.val[0]);
+  q2s16 = vcombine_s16(v2tmp2.val[1], v2tmp3.val[1]);
 
-    q3s16 = vqdmulhq_n_s16(q2s16, sinpi8sqrt2);
-    q4s16 = vqdmulhq_n_s16(q2s16, cospi8sqrt2minus1);
+  q3s16 = vqdmulhq_n_s16(q2s16, sinpi8sqrt2);
+  q4s16 = vqdmulhq_n_s16(q2s16, cospi8sqrt2minus1);
 
-    d12 = vqadd_s16(vget_low_s16(q1s16), vget_high_s16(q1s16));  // a1
-    d13 = vqsub_s16(vget_low_s16(q1s16), vget_high_s16(q1s16));  // b1
+  d12 = vqadd_s16(vget_low_s16(q1s16), vget_high_s16(q1s16));  // a1
+  d13 = vqsub_s16(vget_low_s16(q1s16), vget_high_s16(q1s16));  // b1
 
-    q3s16 = vshrq_n_s16(q3s16, 1);
-    q4s16 = vshrq_n_s16(q4s16, 1);
+  q3s16 = vshrq_n_s16(q3s16, 1);
+  q4s16 = vshrq_n_s16(q4s16, 1);
 
-    q3s16 = vqaddq_s16(q3s16, q2s16);
-    q4s16 = vqaddq_s16(q4s16, q2s16);
+  q3s16 = vqaddq_s16(q3s16, q2s16);
+  q4s16 = vqaddq_s16(q4s16, q2s16);
 
-    d10 = vqsub_s16(vget_low_s16(q3s16), vget_high_s16(q4s16));  // c1
-    d11 = vqadd_s16(vget_high_s16(q3s16), vget_low_s16(q4s16));  // d1
+  d10 = vqsub_s16(vget_low_s16(q3s16), vget_high_s16(q4s16));  // c1
+  d11 = vqadd_s16(vget_high_s16(q3s16), vget_low_s16(q4s16));  // d1
 
-    d2 = vqadd_s16(d12, d11);
-    d3 = vqadd_s16(d13, d10);
-    d4 = vqsub_s16(d13, d10);
-    d5 = vqsub_s16(d12, d11);
+  d2 = vqadd_s16(d12, d11);
+  d3 = vqadd_s16(d13, d10);
+  d4 = vqsub_s16(d13, d10);
+  d5 = vqsub_s16(d12, d11);
 
-    d2 = vrshr_n_s16(d2, 3);
-    d3 = vrshr_n_s16(d3, 3);
-    d4 = vrshr_n_s16(d4, 3);
-    d5 = vrshr_n_s16(d5, 3);
+  d2 = vrshr_n_s16(d2, 3);
+  d3 = vrshr_n_s16(d3, 3);
+  d4 = vrshr_n_s16(d4, 3);
+  d5 = vrshr_n_s16(d5, 3);
 
-    v2tmp0 = vtrn_s32(vreinterpret_s32_s16(d2), vreinterpret_s32_s16(d4));
-    v2tmp1 = vtrn_s32(vreinterpret_s32_s16(d3), vreinterpret_s32_s16(d5));
-    v2tmp2 = vtrn_s16(vreinterpret_s16_s32(v2tmp0.val[0]),
-                      vreinterpret_s16_s32(v2tmp1.val[0]));
-    v2tmp3 = vtrn_s16(vreinterpret_s16_s32(v2tmp0.val[1]),
-                      vreinterpret_s16_s32(v2tmp1.val[1]));
+  v2tmp0 = vtrn_s32(vreinterpret_s32_s16(d2), vreinterpret_s32_s16(d4));
+  v2tmp1 = vtrn_s32(vreinterpret_s32_s16(d3), vreinterpret_s32_s16(d5));
+  v2tmp2 = vtrn_s16(vreinterpret_s16_s32(v2tmp0.val[0]),
+                    vreinterpret_s16_s32(v2tmp1.val[0]));
+  v2tmp3 = vtrn_s16(vreinterpret_s16_s32(v2tmp0.val[1]),
+                    vreinterpret_s16_s32(v2tmp1.val[1]));
 
-    q1s16 = vcombine_s16(v2tmp2.val[0], v2tmp2.val[1]);
-    q2s16 = vcombine_s16(v2tmp3.val[0], v2tmp3.val[1]);
+  q1s16 = vcombine_s16(v2tmp2.val[0], v2tmp2.val[1]);
+  q2s16 = vcombine_s16(v2tmp3.val[0], v2tmp3.val[1]);
 
-    // dc_only_idct_add
-    for (i = 0; i < 2; i++, q1s16 = q2s16) {
-        d6u32 = vld1_lane_u32((const uint32_t *)pred_ptr, d6u32, 0);
-        pred_ptr += pred_stride;
-        d6u32 = vld1_lane_u32((const uint32_t *)pred_ptr, d6u32, 1);
-        pred_ptr += pred_stride;
+  // dc_only_idct_add
+  for (i = 0; i < 2; i++, q1s16 = q2s16) {
+    d6u32 = vld1_lane_u32((const uint32_t *)pred_ptr, d6u32, 0);
+    pred_ptr += pred_stride;
+    d6u32 = vld1_lane_u32((const uint32_t *)pred_ptr, d6u32, 1);
+    pred_ptr += pred_stride;
 
-        q1u16 = vaddw_u8(vreinterpretq_u16_s16(q1s16),
-                         vreinterpret_u8_u32(d6u32));
-        d1u8 = vqmovun_s16(vreinterpretq_s16_u16(q1u16));
+    q1u16 = vaddw_u8(vreinterpretq_u16_s16(q1s16), vreinterpret_u8_u32(d6u32));
+    d1u8 = vqmovun_s16(vreinterpretq_s16_u16(q1u16));
 
-        vst1_lane_u32((uint32_t *)dst_ptr, vreinterpret_u32_u8(d1u8), 0);
-        dst_ptr += dst_stride;
-        vst1_lane_u32((uint32_t *)dst_ptr, vreinterpret_u32_u8(d1u8), 1);
-        dst_ptr += dst_stride;
-    }
-    return;
+    vst1_lane_u32((uint32_t *)dst_ptr, vreinterpret_u32_u8(d1u8), 0);
+    dst_ptr += dst_stride;
+    vst1_lane_u32((uint32_t *)dst_ptr, vreinterpret_u32_u8(d1u8), 1);
+    dst_ptr += dst_stride;
+  }
+  return;
 }
--- a/vp8/common/arm/neon/sixtappredict_neon.c
+++ b/vp8/common/arm/neon/sixtappredict_neon.c
@@ -12,131 +12,33 @@
 #include "vpx_ports/mem.h"
 
 static const int8_t vp8_sub_pel_filters[8][8] = {
-    {0,  0,  128,   0,   0, 0, 0, 0},  /* note that 1/8 pel positionyys are */
-    {0, -6,  123,  12,  -1, 0, 0, 0},  /*    just as per alpha -0.5 bicubic */
-    {2, -11, 108,  36,  -8, 1, 0, 0},  /* New 1/4 pel 6 tap filter */
-    {0, -9,   93,  50,  -6, 0, 0, 0},
-    {3, -16,  77,  77, -16, 3, 0, 0},  /* New 1/2 pel 6 tap filter */
-    {0, -6,   50,  93,  -9, 0, 0, 0},
-    {1, -8,   36, 108, -11, 2, 0, 0},  /* New 1/4 pel 6 tap filter */
-    {0, -1,   12, 123,  -6, 0, 0, 0},
+  { 0, 0, 128, 0, 0, 0, 0, 0 },     /* note that 1/8 pel positionyys are */
+  { 0, -6, 123, 12, -1, 0, 0, 0 },  /*    just as per alpha -0.5 bicubic */
+  { 2, -11, 108, 36, -8, 1, 0, 0 }, /* New 1/4 pel 6 tap filter */
+  { 0, -9, 93, 50, -6, 0, 0, 0 },
+  { 3, -16, 77, 77, -16, 3, 0, 0 }, /* New 1/2 pel 6 tap filter */
+  { 0, -6, 50, 93, -9, 0, 0, 0 },
+  { 1, -8, 36, 108, -11, 2, 0, 0 }, /* New 1/4 pel 6 tap filter */
+  { 0, -1, 12, 123, -6, 0, 0, 0 },
 };
 
-void vp8_sixtap_predict8x4_neon(
-        unsigned char *src_ptr,
-        int src_pixels_per_line,
-        int xoffset,
-        int yoffset,
-        unsigned char *dst_ptr,
-        int dst_pitch) {
-    unsigned char *src;
-    uint8x8_t d0u8, d1u8, d2u8, d3u8, d4u8, d5u8, d6u8, d7u8, d8u8, d9u8;
-    uint8x8_t d22u8, d23u8, d24u8, d25u8, d26u8;
-    uint8x8_t d27u8, d28u8, d29u8, d30u8, d31u8;
-    int8x8_t dtmps8, d0s8, d1s8, d2s8, d3s8, d4s8, d5s8;
-    uint16x8_t q3u16, q4u16, q5u16, q6u16, q7u16;
-    uint16x8_t q8u16, q9u16, q10u16, q11u16, q12u16;
-    int16x8_t q3s16, q4s16, q5s16, q6s16, q7s16;
-    int16x8_t q8s16, q9s16, q10s16, q11s16, q12s16;
-    uint8x16_t q3u8, q4u8, q5u8, q6u8, q7u8;
+void vp8_sixtap_predict8x4_neon(unsigned char *src_ptr, int src_pixels_per_line,
+                                int xoffset, int yoffset,
+                                unsigned char *dst_ptr, int dst_pitch) {
+  unsigned char *src;
+  uint8x8_t d0u8, d1u8, d2u8, d3u8, d4u8, d5u8, d6u8, d7u8, d8u8, d9u8;
+  uint8x8_t d22u8, d23u8, d24u8, d25u8, d26u8;
+  uint8x8_t d27u8, d28u8, d29u8, d30u8, d31u8;
+  int8x8_t dtmps8, d0s8, d1s8, d2s8, d3s8, d4s8, d5s8;
+  uint16x8_t q3u16, q4u16, q5u16, q6u16, q7u16;
+  uint16x8_t q8u16, q9u16, q10u16, q11u16, q12u16;
+  int16x8_t q3s16, q4s16, q5s16, q6s16, q7s16;
+  int16x8_t q8s16, q9s16, q10s16, q11s16, q12s16;
+  uint8x16_t q3u8, q4u8, q5u8, q6u8, q7u8;
 
-    if (xoffset == 0) {  // secondpass_filter8x4_only
-        // load second_pass filter
-        dtmps8 = vld1_s8(vp8_sub_pel_filters[yoffset]);
-        d0s8 = vdup_lane_s8(dtmps8, 0);
-        d1s8 = vdup_lane_s8(dtmps8, 1);
-        d2s8 = vdup_lane_s8(dtmps8, 2);
-        d3s8 = vdup_lane_s8(dtmps8, 3);
-        d4s8 = vdup_lane_s8(dtmps8, 4);
-        d5s8 = vdup_lane_s8(dtmps8, 5);
-        d0u8 = vreinterpret_u8_s8(vabs_s8(d0s8));
-        d1u8 = vreinterpret_u8_s8(vabs_s8(d1s8));
-        d2u8 = vreinterpret_u8_s8(vabs_s8(d2s8));
-        d3u8 = vreinterpret_u8_s8(vabs_s8(d3s8));
-        d4u8 = vreinterpret_u8_s8(vabs_s8(d4s8));
-        d5u8 = vreinterpret_u8_s8(vabs_s8(d5s8));
-
-        // load src data
-        src = src_ptr - src_pixels_per_line * 2;
-        d22u8 = vld1_u8(src);
-        src += src_pixels_per_line;
-        d23u8 = vld1_u8(src);
-        src += src_pixels_per_line;
-        d24u8 = vld1_u8(src);
-        src += src_pixels_per_line;
-        d25u8 = vld1_u8(src);
-        src += src_pixels_per_line;
-        d26u8 = vld1_u8(src);
-        src += src_pixels_per_line;
-        d27u8 = vld1_u8(src);
-        src += src_pixels_per_line;
-        d28u8 = vld1_u8(src);
-        src += src_pixels_per_line;
-        d29u8 = vld1_u8(src);
-        src += src_pixels_per_line;
-        d30u8 = vld1_u8(src);
-
-        q3u16 = vmull_u8(d22u8, d0u8);
-        q4u16 = vmull_u8(d23u8, d0u8);
-        q5u16 = vmull_u8(d24u8, d0u8);
-        q6u16 = vmull_u8(d25u8, d0u8);
-
-        q3u16 = vmlsl_u8(q3u16, d23u8, d1u8);
-        q4u16 = vmlsl_u8(q4u16, d24u8, d1u8);
-        q5u16 = vmlsl_u8(q5u16, d25u8, d1u8);
-        q6u16 = vmlsl_u8(q6u16, d26u8, d1u8);
-
-        q3u16 = vmlsl_u8(q3u16, d26u8, d4u8);
-        q4u16 = vmlsl_u8(q4u16, d27u8, d4u8);
-        q5u16 = vmlsl_u8(q5u16, d28u8, d4u8);
-        q6u16 = vmlsl_u8(q6u16, d29u8, d4u8);
-
-        q3u16 = vmlal_u8(q3u16, d24u8, d2u8);
-        q4u16 = vmlal_u8(q4u16, d25u8, d2u8);
-        q5u16 = vmlal_u8(q5u16, d26u8, d2u8);
-        q6u16 = vmlal_u8(q6u16, d27u8, d2u8);
-
-        q3u16 = vmlal_u8(q3u16, d27u8, d5u8);
-        q4u16 = vmlal_u8(q4u16, d28u8, d5u8);
-        q5u16 = vmlal_u8(q5u16, d29u8, d5u8);
-        q6u16 = vmlal_u8(q6u16, d30u8, d5u8);
-
-        q7u16 = vmull_u8(d25u8, d3u8);
-        q8u16 = vmull_u8(d26u8, d3u8);
-        q9u16 = vmull_u8(d27u8, d3u8);
-        q10u16 = vmull_u8(d28u8, d3u8);
-
-        q3s16 = vreinterpretq_s16_u16(q3u16);
-        q4s16 = vreinterpretq_s16_u16(q4u16);
-        q5s16 = vreinterpretq_s16_u16(q5u16);
-        q6s16 = vreinterpretq_s16_u16(q6u16);
-        q7s16 = vreinterpretq_s16_u16(q7u16);
-        q8s16 = vreinterpretq_s16_u16(q8u16);
-        q9s16 = vreinterpretq_s16_u16(q9u16);
-        q10s16 = vreinterpretq_s16_u16(q10u16);
-
-        q7s16 = vqaddq_s16(q7s16, q3s16);
-        q8s16 = vqaddq_s16(q8s16, q4s16);
-        q9s16 = vqaddq_s16(q9s16, q5s16);
-        q10s16 = vqaddq_s16(q10s16, q6s16);
-
-        d6u8 = vqrshrun_n_s16(q7s16, 7);
-        d7u8 = vqrshrun_n_s16(q8s16, 7);
-        d8u8 = vqrshrun_n_s16(q9s16, 7);
-        d9u8 = vqrshrun_n_s16(q10s16, 7);
-
-        vst1_u8(dst_ptr, d6u8);
-        dst_ptr += dst_pitch;
-        vst1_u8(dst_ptr, d7u8);
-        dst_ptr += dst_pitch;
-        vst1_u8(dst_ptr, d8u8);
-        dst_ptr += dst_pitch;
-        vst1_u8(dst_ptr, d9u8);
-        return;
-    }
-
-    // load first_pass filter
-    dtmps8 = vld1_s8(vp8_sub_pel_filters[xoffset]);
+  if (xoffset == 0) {  // secondpass_filter8x4_only
+    // load second_pass filter
+    dtmps8 = vld1_s8(vp8_sub_pel_filters[yoffset]);
     d0s8 = vdup_lane_s8(dtmps8, 0);
     d1s8 = vdup_lane_s8(dtmps8, 1);
     d2s8 = vdup_lane_s8(dtmps8, 2);
@@ -150,220 +52,26 @@
     d4u8 = vreinterpret_u8_s8(vabs_s8(d4s8));
     d5u8 = vreinterpret_u8_s8(vabs_s8(d5s8));
 
-    // First pass: output_height lines x output_width columns (9x4)
-    if (yoffset == 0)  // firstpass_filter4x4_only
-        src = src_ptr - 2;
-    else
-        src = src_ptr - 2 - (src_pixels_per_line * 2);
-    q3u8 = vld1q_u8(src);
+    // load src data
+    src = src_ptr - src_pixels_per_line * 2;
+    d22u8 = vld1_u8(src);
     src += src_pixels_per_line;
-    q4u8 = vld1q_u8(src);
+    d23u8 = vld1_u8(src);
     src += src_pixels_per_line;
-    q5u8 = vld1q_u8(src);
+    d24u8 = vld1_u8(src);
     src += src_pixels_per_line;
-    q6u8 = vld1q_u8(src);
-
-    q7u16  = vmull_u8(vget_low_u8(q3u8), d0u8);
-    q8u16  = vmull_u8(vget_low_u8(q4u8), d0u8);
-    q9u16  = vmull_u8(vget_low_u8(q5u8), d0u8);
-    q10u16 = vmull_u8(vget_low_u8(q6u8), d0u8);
-
-    d28u8 = vext_u8(vget_low_u8(q3u8), vget_high_u8(q3u8), 1);
-    d29u8 = vext_u8(vget_low_u8(q4u8), vget_high_u8(q4u8), 1);
-    d30u8 = vext_u8(vget_low_u8(q5u8), vget_high_u8(q5u8), 1);
-    d31u8 = vext_u8(vget_low_u8(q6u8), vget_high_u8(q6u8), 1);
-
-    q7u16  = vmlsl_u8(q7u16, d28u8, d1u8);
-    q8u16  = vmlsl_u8(q8u16, d29u8, d1u8);
-    q9u16  = vmlsl_u8(q9u16, d30u8, d1u8);
-    q10u16 = vmlsl_u8(q10u16, d31u8, d1u8);
-
-    d28u8 = vext_u8(vget_low_u8(q3u8), vget_high_u8(q3u8), 4);
-    d29u8 = vext_u8(vget_low_u8(q4u8), vget_high_u8(q4u8), 4);
-    d30u8 = vext_u8(vget_low_u8(q5u8), vget_high_u8(q5u8), 4);
-    d31u8 = vext_u8(vget_low_u8(q6u8), vget_high_u8(q6u8), 4);
-
-    q7u16  = vmlsl_u8(q7u16, d28u8, d4u8);
-    q8u16  = vmlsl_u8(q8u16, d29u8, d4u8);
-    q9u16  = vmlsl_u8(q9u16, d30u8, d4u8);
-    q10u16 = vmlsl_u8(q10u16, d31u8, d4u8);
-
-    d28u8 = vext_u8(vget_low_u8(q3u8), vget_high_u8(q3u8), 2);
-    d29u8 = vext_u8(vget_low_u8(q4u8), vget_high_u8(q4u8), 2);
-    d30u8 = vext_u8(vget_low_u8(q5u8), vget_high_u8(q5u8), 2);
-    d31u8 = vext_u8(vget_low_u8(q6u8), vget_high_u8(q6u8), 2);
-
-    q7u16  = vmlal_u8(q7u16, d28u8, d2u8);
-    q8u16  = vmlal_u8(q8u16, d29u8, d2u8);
-    q9u16  = vmlal_u8(q9u16, d30u8, d2u8);
-    q10u16 = vmlal_u8(q10u16, d31u8, d2u8);
-
-    d28u8 = vext_u8(vget_low_u8(q3u8), vget_high_u8(q3u8), 5);
-    d29u8 = vext_u8(vget_low_u8(q4u8), vget_high_u8(q4u8), 5);
-    d30u8 = vext_u8(vget_low_u8(q5u8), vget_high_u8(q5u8), 5);
-    d31u8 = vext_u8(vget_low_u8(q6u8), vget_high_u8(q6u8), 5);
-
-    q7u16 = vmlal_u8(q7u16, d28u8, d5u8);
-    q8u16 = vmlal_u8(q8u16, d29u8, d5u8);
-    q9u16 = vmlal_u8(q9u16, d30u8, d5u8);
-    q10u16 = vmlal_u8(q10u16, d31u8, d5u8);
-
-    d28u8 = vext_u8(vget_low_u8(q3u8), vget_high_u8(q3u8), 3);
-    d29u8 = vext_u8(vget_low_u8(q4u8), vget_high_u8(q4u8), 3);
-    d30u8 = vext_u8(vget_low_u8(q5u8), vget_high_u8(q5u8), 3);
-    d31u8 = vext_u8(vget_low_u8(q6u8), vget_high_u8(q6u8), 3);
-
-    q3u16 = vmull_u8(d28u8, d3u8);
-    q4u16 = vmull_u8(d29u8, d3u8);
-    q5u16 = vmull_u8(d30u8, d3u8);
-    q6u16 = vmull_u8(d31u8, d3u8);
-
-    q3s16 = vreinterpretq_s16_u16(q3u16);
-    q4s16 = vreinterpretq_s16_u16(q4u16);
-    q5s16 = vreinterpretq_s16_u16(q5u16);
-    q6s16 = vreinterpretq_s16_u16(q6u16);
-    q7s16 = vreinterpretq_s16_u16(q7u16);
-    q8s16 = vreinterpretq_s16_u16(q8u16);
-    q9s16 = vreinterpretq_s16_u16(q9u16);
-    q10s16 = vreinterpretq_s16_u16(q10u16);
-
-    q7s16 = vqaddq_s16(q7s16, q3s16);
-    q8s16 = vqaddq_s16(q8s16, q4s16);
-    q9s16 = vqaddq_s16(q9s16, q5s16);
-    q10s16 = vqaddq_s16(q10s16, q6s16);
-
-    d22u8 = vqrshrun_n_s16(q7s16, 7);
-    d23u8 = vqrshrun_n_s16(q8s16, 7);
-    d24u8 = vqrshrun_n_s16(q9s16, 7);
-    d25u8 = vqrshrun_n_s16(q10s16, 7);
-
-    if (yoffset == 0) {  // firstpass_filter8x4_only
-        vst1_u8(dst_ptr, d22u8);
-        dst_ptr += dst_pitch;
-        vst1_u8(dst_ptr, d23u8);
-        dst_ptr += dst_pitch;
-        vst1_u8(dst_ptr, d24u8);
-        dst_ptr += dst_pitch;
-        vst1_u8(dst_ptr, d25u8);
-        return;
-    }
-
-    // First Pass on rest 5-line data
+    d25u8 = vld1_u8(src);
     src += src_pixels_per_line;
-    q3u8 = vld1q_u8(src);
+    d26u8 = vld1_u8(src);
     src += src_pixels_per_line;
-    q4u8 = vld1q_u8(src);
+    d27u8 = vld1_u8(src);
     src += src_pixels_per_line;
-    q5u8 = vld1q_u8(src);
+    d28u8 = vld1_u8(src);
     src += src_pixels_per_line;
-    q6u8 = vld1q_u8(src);
+    d29u8 = vld1_u8(src);
     src += src_pixels_per_line;
-    q7u8 = vld1q_u8(src);
+    d30u8 = vld1_u8(src);
 
-    q8u16 = vmull_u8(vget_low_u8(q3u8), d0u8);
-    q9u16 = vmull_u8(vget_low_u8(q4u8), d0u8);
-    q10u16 = vmull_u8(vget_low_u8(q5u8), d0u8);
-    q11u16 = vmull_u8(vget_low_u8(q6u8), d0u8);
-    q12u16 = vmull_u8(vget_low_u8(q7u8), d0u8);
-
-    d27u8 = vext_u8(vget_low_u8(q3u8), vget_high_u8(q3u8), 1);
-    d28u8 = vext_u8(vget_low_u8(q4u8), vget_high_u8(q4u8), 1);
-    d29u8 = vext_u8(vget_low_u8(q5u8), vget_high_u8(q5u8), 1);
-    d30u8 = vext_u8(vget_low_u8(q6u8), vget_high_u8(q6u8), 1);
-    d31u8 = vext_u8(vget_low_u8(q7u8), vget_high_u8(q7u8), 1);
-
-    q8u16  = vmlsl_u8(q8u16, d27u8, d1u8);
-    q9u16  = vmlsl_u8(q9u16, d28u8, d1u8);
-    q10u16 = vmlsl_u8(q10u16, d29u8, d1u8);
-    q11u16 = vmlsl_u8(q11u16, d30u8, d1u8);
-    q12u16 = vmlsl_u8(q12u16, d31u8, d1u8);
-
-    d27u8 = vext_u8(vget_low_u8(q3u8), vget_high_u8(q3u8), 4);
-    d28u8 = vext_u8(vget_low_u8(q4u8), vget_high_u8(q4u8), 4);
-    d29u8 = vext_u8(vget_low_u8(q5u8), vget_high_u8(q5u8), 4);
-    d30u8 = vext_u8(vget_low_u8(q6u8), vget_high_u8(q6u8), 4);
-    d31u8 = vext_u8(vget_low_u8(q7u8), vget_high_u8(q7u8), 4);
-
-    q8u16  = vmlsl_u8(q8u16, d27u8, d4u8);
-    q9u16  = vmlsl_u8(q9u16, d28u8, d4u8);
-    q10u16 = vmlsl_u8(q10u16, d29u8, d4u8);
-    q11u16 = vmlsl_u8(q11u16, d30u8, d4u8);
-    q12u16 = vmlsl_u8(q12u16, d31u8, d4u8);
-
-    d27u8 = vext_u8(vget_low_u8(q3u8), vget_high_u8(q3u8), 2);
-    d28u8 = vext_u8(vget_low_u8(q4u8), vget_high_u8(q4u8), 2);
-    d29u8 = vext_u8(vget_low_u8(q5u8), vget_high_u8(q5u8), 2);
-    d30u8 = vext_u8(vget_low_u8(q6u8), vget_high_u8(q6u8), 2);
-    d31u8 = vext_u8(vget_low_u8(q7u8), vget_high_u8(q7u8), 2);
-
-    q8u16  = vmlal_u8(q8u16, d27u8, d2u8);
-    q9u16  = vmlal_u8(q9u16, d28u8, d2u8);
-    q10u16 = vmlal_u8(q10u16, d29u8, d2u8);
-    q11u16 = vmlal_u8(q11u16, d30u8, d2u8);
-    q12u16 = vmlal_u8(q12u16, d31u8, d2u8);
-
-    d27u8 = vext_u8(vget_low_u8(q3u8), vget_high_u8(q3u8), 5);
-    d28u8 = vext_u8(vget_low_u8(q4u8), vget_high_u8(q4u8), 5);
-    d29u8 = vext_u8(vget_low_u8(q5u8), vget_high_u8(q5u8), 5);
-    d30u8 = vext_u8(vget_low_u8(q6u8), vget_high_u8(q6u8), 5);
-    d31u8 = vext_u8(vget_low_u8(q7u8), vget_high_u8(q7u8), 5);
-
-    q8u16  = vmlal_u8(q8u16, d27u8, d5u8);
-    q9u16  = vmlal_u8(q9u16, d28u8, d5u8);
-    q10u16 = vmlal_u8(q10u16, d29u8, d5u8);
-    q11u16 = vmlal_u8(q11u16, d30u8, d5u8);
-    q12u16 = vmlal_u8(q12u16, d31u8, d5u8);
-
-    d27u8 = vext_u8(vget_low_u8(q3u8), vget_high_u8(q3u8), 3);
-    d28u8 = vext_u8(vget_low_u8(q4u8), vget_high_u8(q4u8), 3);
-    d29u8 = vext_u8(vget_low_u8(q5u8), vget_high_u8(q5u8), 3);
-    d30u8 = vext_u8(vget_low_u8(q6u8), vget_high_u8(q6u8), 3);
-    d31u8 = vext_u8(vget_low_u8(q7u8), vget_high_u8(q7u8), 3);
-
-    q3u16 = vmull_u8(d27u8, d3u8);
-    q4u16 = vmull_u8(d28u8, d3u8);
-    q5u16 = vmull_u8(d29u8, d3u8);
-    q6u16 = vmull_u8(d30u8, d3u8);
-    q7u16 = vmull_u8(d31u8, d3u8);
-
-    q3s16 = vreinterpretq_s16_u16(q3u16);
-    q4s16 = vreinterpretq_s16_u16(q4u16);
-    q5s16 = vreinterpretq_s16_u16(q5u16);
-    q6s16 = vreinterpretq_s16_u16(q6u16);
-    q7s16 = vreinterpretq_s16_u16(q7u16);
-    q8s16 = vreinterpretq_s16_u16(q8u16);
-    q9s16 = vreinterpretq_s16_u16(q9u16);
-    q10s16 = vreinterpretq_s16_u16(q10u16);
-    q11s16 = vreinterpretq_s16_u16(q11u16);
-    q12s16 = vreinterpretq_s16_u16(q12u16);
-
-    q8s16 = vqaddq_s16(q8s16, q3s16);
-    q9s16 = vqaddq_s16(q9s16, q4s16);
-    q10s16 = vqaddq_s16(q10s16, q5s16);
-    q11s16 = vqaddq_s16(q11s16, q6s16);
-    q12s16 = vqaddq_s16(q12s16, q7s16);
-
-    d26u8 = vqrshrun_n_s16(q8s16, 7);
-    d27u8 = vqrshrun_n_s16(q9s16, 7);
-    d28u8 = vqrshrun_n_s16(q10s16, 7);
-    d29u8 = vqrshrun_n_s16(q11s16, 7);
-    d30u8 = vqrshrun_n_s16(q12s16, 7);
-
-    // Second pass: 8x4
-    dtmps8 = vld1_s8(vp8_sub_pel_filters[yoffset]);
-    d0s8 = vdup_lane_s8(dtmps8, 0);
-    d1s8 = vdup_lane_s8(dtmps8, 1);
-    d2s8 = vdup_lane_s8(dtmps8, 2);
-    d3s8 = vdup_lane_s8(dtmps8, 3);
-    d4s8 = vdup_lane_s8(dtmps8, 4);
-    d5s8 = vdup_lane_s8(dtmps8, 5);
-    d0u8 = vreinterpret_u8_s8(vabs_s8(d0s8));
-    d1u8 = vreinterpret_u8_s8(vabs_s8(d1s8));
-    d2u8 = vreinterpret_u8_s8(vabs_s8(d2s8));
-    d3u8 = vreinterpret_u8_s8(vabs_s8(d3s8));
-    d4u8 = vreinterpret_u8_s8(vabs_s8(d4s8));
-    d5u8 = vreinterpret_u8_s8(vabs_s8(d5s8));
-
     q3u16 = vmull_u8(d22u8, d0u8);
     q4u16 = vmull_u8(d23u8, d0u8);
     q5u16 = vmull_u8(d24u8, d0u8);
@@ -421,146 +129,315 @@
     dst_ptr += dst_pitch;
     vst1_u8(dst_ptr, d9u8);
     return;
-}
+  }
 
-void vp8_sixtap_predict8x8_neon(
-        unsigned char *src_ptr,
-        int src_pixels_per_line,
-        int xoffset,
-        int yoffset,
-        unsigned char *dst_ptr,
-        int dst_pitch) {
-    unsigned char *src, *tmpp;
-    unsigned char tmp[64];
-    int i;
-    uint8x8_t d0u8, d1u8, d2u8, d3u8, d4u8, d5u8, d6u8, d7u8, d8u8, d9u8;
-    uint8x8_t d18u8, d19u8, d20u8, d21u8, d22u8, d23u8, d24u8, d25u8;
-    uint8x8_t d26u8, d27u8, d28u8, d29u8, d30u8, d31u8;
-    int8x8_t dtmps8, d0s8, d1s8, d2s8, d3s8, d4s8, d5s8;
-    uint16x8_t q3u16, q4u16, q5u16, q6u16, q7u16;
-    uint16x8_t q8u16, q9u16, q10u16, q11u16, q12u16;
-    int16x8_t q3s16, q4s16, q5s16, q6s16, q7s16;
-    int16x8_t q8s16, q9s16, q10s16, q11s16, q12s16;
-    uint8x16_t q3u8, q4u8, q5u8, q6u8, q7u8, q9u8, q10u8, q11u8, q12u8;
+  // load first_pass filter
+  dtmps8 = vld1_s8(vp8_sub_pel_filters[xoffset]);
+  d0s8 = vdup_lane_s8(dtmps8, 0);
+  d1s8 = vdup_lane_s8(dtmps8, 1);
+  d2s8 = vdup_lane_s8(dtmps8, 2);
+  d3s8 = vdup_lane_s8(dtmps8, 3);
+  d4s8 = vdup_lane_s8(dtmps8, 4);
+  d5s8 = vdup_lane_s8(dtmps8, 5);
+  d0u8 = vreinterpret_u8_s8(vabs_s8(d0s8));
+  d1u8 = vreinterpret_u8_s8(vabs_s8(d1s8));
+  d2u8 = vreinterpret_u8_s8(vabs_s8(d2s8));
+  d3u8 = vreinterpret_u8_s8(vabs_s8(d3s8));
+  d4u8 = vreinterpret_u8_s8(vabs_s8(d4s8));
+  d5u8 = vreinterpret_u8_s8(vabs_s8(d5s8));
 
-    if (xoffset == 0) {  // secondpass_filter8x8_only
-        // load second_pass filter
-        dtmps8 = vld1_s8(vp8_sub_pel_filters[yoffset]);
-        d0s8 = vdup_lane_s8(dtmps8, 0);
-        d1s8 = vdup_lane_s8(dtmps8, 1);
-        d2s8 = vdup_lane_s8(dtmps8, 2);
-        d3s8 = vdup_lane_s8(dtmps8, 3);
-        d4s8 = vdup_lane_s8(dtmps8, 4);
-        d5s8 = vdup_lane_s8(dtmps8, 5);
-        d0u8 = vreinterpret_u8_s8(vabs_s8(d0s8));
-        d1u8 = vreinterpret_u8_s8(vabs_s8(d1s8));
-        d2u8 = vreinterpret_u8_s8(vabs_s8(d2s8));
-        d3u8 = vreinterpret_u8_s8(vabs_s8(d3s8));
-        d4u8 = vreinterpret_u8_s8(vabs_s8(d4s8));
-        d5u8 = vreinterpret_u8_s8(vabs_s8(d5s8));
+  // First pass: output_height lines x output_width columns (9x4)
+  if (yoffset == 0)  // firstpass_filter4x4_only
+    src = src_ptr - 2;
+  else
+    src = src_ptr - 2 - (src_pixels_per_line * 2);
+  q3u8 = vld1q_u8(src);
+  src += src_pixels_per_line;
+  q4u8 = vld1q_u8(src);
+  src += src_pixels_per_line;
+  q5u8 = vld1q_u8(src);
+  src += src_pixels_per_line;
+  q6u8 = vld1q_u8(src);
 
-        // load src data
-        src = src_ptr - src_pixels_per_line * 2;
-        d18u8 = vld1_u8(src);
-        src += src_pixels_per_line;
-        d19u8 = vld1_u8(src);
-        src += src_pixels_per_line;
-        d20u8 = vld1_u8(src);
-        src += src_pixels_per_line;
-        d21u8 = vld1_u8(src);
-        src += src_pixels_per_line;
-        d22u8 = vld1_u8(src);
-        src += src_pixels_per_line;
-        d23u8 = vld1_u8(src);
-        src += src_pixels_per_line;
-        d24u8 = vld1_u8(src);
-        src += src_pixels_per_line;
-        d25u8 = vld1_u8(src);
-        src += src_pixels_per_line;
-        d26u8 = vld1_u8(src);
-        src += src_pixels_per_line;
-        d27u8 = vld1_u8(src);
-        src += src_pixels_per_line;
-        d28u8 = vld1_u8(src);
-        src += src_pixels_per_line;
-        d29u8 = vld1_u8(src);
-        src += src_pixels_per_line;
-        d30u8 = vld1_u8(src);
+  q7u16 = vmull_u8(vget_low_u8(q3u8), d0u8);
+  q8u16 = vmull_u8(vget_low_u8(q4u8), d0u8);
+  q9u16 = vmull_u8(vget_low_u8(q5u8), d0u8);
+  q10u16 = vmull_u8(vget_low_u8(q6u8), d0u8);
 
-        for (i = 2; i > 0; i--) {
-            q3u16 = vmull_u8(d18u8, d0u8);
-            q4u16 = vmull_u8(d19u8, d0u8);
-            q5u16 = vmull_u8(d20u8, d0u8);
-            q6u16 = vmull_u8(d21u8, d0u8);
+  d28u8 = vext_u8(vget_low_u8(q3u8), vget_high_u8(q3u8), 1);
+  d29u8 = vext_u8(vget_low_u8(q4u8), vget_high_u8(q4u8), 1);
+  d30u8 = vext_u8(vget_low_u8(q5u8), vget_high_u8(q5u8), 1);
+  d31u8 = vext_u8(vget_low_u8(q6u8), vget_high_u8(q6u8), 1);
 
-            q3u16 = vmlsl_u8(q3u16, d19u8, d1u8);
-            q4u16 = vmlsl_u8(q4u16, d20u8, d1u8);
-            q5u16 = vmlsl_u8(q5u16, d21u8, d1u8);
-            q6u16 = vmlsl_u8(q6u16, d22u8, d1u8);
+  q7u16 = vmlsl_u8(q7u16, d28u8, d1u8);
+  q8u16 = vmlsl_u8(q8u16, d29u8, d1u8);
+  q9u16 = vmlsl_u8(q9u16, d30u8, d1u8);
+  q10u16 = vmlsl_u8(q10u16, d31u8, d1u8);
 
-            q3u16 = vmlsl_u8(q3u16, d22u8, d4u8);
-            q4u16 = vmlsl_u8(q4u16, d23u8, d4u8);
-            q5u16 = vmlsl_u8(q5u16, d24u8, d4u8);
-            q6u16 = vmlsl_u8(q6u16, d25u8, d4u8);
+  d28u8 = vext_u8(vget_low_u8(q3u8), vget_high_u8(q3u8), 4);
+  d29u8 = vext_u8(vget_low_u8(q4u8), vget_high_u8(q4u8), 4);
+  d30u8 = vext_u8(vget_low_u8(q5u8), vget_high_u8(q5u8), 4);
+  d31u8 = vext_u8(vget_low_u8(q6u8), vget_high_u8(q6u8), 4);
 
-            q3u16 = vmlal_u8(q3u16, d20u8, d2u8);
-            q4u16 = vmlal_u8(q4u16, d21u8, d2u8);
-            q5u16 = vmlal_u8(q5u16, d22u8, d2u8);
-            q6u16 = vmlal_u8(q6u16, d23u8, d2u8);
+  q7u16 = vmlsl_u8(q7u16, d28u8, d4u8);
+  q8u16 = vmlsl_u8(q8u16, d29u8, d4u8);
+  q9u16 = vmlsl_u8(q9u16, d30u8, d4u8);
+  q10u16 = vmlsl_u8(q10u16, d31u8, d4u8);
 
-            q3u16 = vmlal_u8(q3u16, d23u8, d5u8);
-            q4u16 = vmlal_u8(q4u16, d24u8, d5u8);
-            q5u16 = vmlal_u8(q5u16, d25u8, d5u8);
-            q6u16 = vmlal_u8(q6u16, d26u8, d5u8);
+  d28u8 = vext_u8(vget_low_u8(q3u8), vget_high_u8(q3u8), 2);
+  d29u8 = vext_u8(vget_low_u8(q4u8), vget_high_u8(q4u8), 2);
+  d30u8 = vext_u8(vget_low_u8(q5u8), vget_high_u8(q5u8), 2);
+  d31u8 = vext_u8(vget_low_u8(q6u8), vget_high_u8(q6u8), 2);
 
-            q7u16 = vmull_u8(d21u8, d3u8);
-            q8u16 = vmull_u8(d22u8, d3u8);
-            q9u16 = vmull_u8(d23u8, d3u8);
-            q10u16 = vmull_u8(d24u8, d3u8);
+  q7u16 = vmlal_u8(q7u16, d28u8, d2u8);
+  q8u16 = vmlal_u8(q8u16, d29u8, d2u8);
+  q9u16 = vmlal_u8(q9u16, d30u8, d2u8);
+  q10u16 = vmlal_u8(q10u16, d31u8, d2u8);
 
-            q3s16 = vreinterpretq_s16_u16(q3u16);
-            q4s16 = vreinterpretq_s16_u16(q4u16);
-            q5s16 = vreinterpretq_s16_u16(q5u16);
-            q6s16 = vreinterpretq_s16_u16(q6u16);
-            q7s16 = vreinterpretq_s16_u16(q7u16);
-            q8s16 = vreinterpretq_s16_u16(q8u16);
-            q9s16 = vreinterpretq_s16_u16(q9u16);
-            q10s16 = vreinterpretq_s16_u16(q10u16);
+  d28u8 = vext_u8(vget_low_u8(q3u8), vget_high_u8(q3u8), 5);
+  d29u8 = vext_u8(vget_low_u8(q4u8), vget_high_u8(q4u8), 5);
+  d30u8 = vext_u8(vget_low_u8(q5u8), vget_high_u8(q5u8), 5);
+  d31u8 = vext_u8(vget_low_u8(q6u8), vget_high_u8(q6u8), 5);
 
-            q7s16 = vqaddq_s16(q7s16, q3s16);
-            q8s16 = vqaddq_s16(q8s16, q4s16);
-            q9s16 = vqaddq_s16(q9s16, q5s16);
-            q10s16 = vqaddq_s16(q10s16, q6s16);
+  q7u16 = vmlal_u8(q7u16, d28u8, d5u8);
+  q8u16 = vmlal_u8(q8u16, d29u8, d5u8);
+  q9u16 = vmlal_u8(q9u16, d30u8, d5u8);
+  q10u16 = vmlal_u8(q10u16, d31u8, d5u8);
 
-            d6u8 = vqrshrun_n_s16(q7s16, 7);
-            d7u8 = vqrshrun_n_s16(q8s16, 7);
-            d8u8 = vqrshrun_n_s16(q9s16, 7);
-            d9u8 = vqrshrun_n_s16(q10s16, 7);
+  d28u8 = vext_u8(vget_low_u8(q3u8), vget_high_u8(q3u8), 3);
+  d29u8 = vext_u8(vget_low_u8(q4u8), vget_high_u8(q4u8), 3);
+  d30u8 = vext_u8(vget_low_u8(q5u8), vget_high_u8(q5u8), 3);
+  d31u8 = vext_u8(vget_low_u8(q6u8), vget_high_u8(q6u8), 3);
 
-            d18u8 = d22u8;
-            d19u8 = d23u8;
-            d20u8 = d24u8;
-            d21u8 = d25u8;
-            d22u8 = d26u8;
-            d23u8 = d27u8;
-            d24u8 = d28u8;
-            d25u8 = d29u8;
-            d26u8 = d30u8;
+  q3u16 = vmull_u8(d28u8, d3u8);
+  q4u16 = vmull_u8(d29u8, d3u8);
+  q5u16 = vmull_u8(d30u8, d3u8);
+  q6u16 = vmull_u8(d31u8, d3u8);
 
-            vst1_u8(dst_ptr, d6u8);
-            dst_ptr += dst_pitch;
-            vst1_u8(dst_ptr, d7u8);
-            dst_ptr += dst_pitch;
-            vst1_u8(dst_ptr, d8u8);
-            dst_ptr += dst_pitch;
-            vst1_u8(dst_ptr, d9u8);
-            dst_ptr += dst_pitch;
-        }
-        return;
-    }
+  q3s16 = vreinterpretq_s16_u16(q3u16);
+  q4s16 = vreinterpretq_s16_u16(q4u16);
+  q5s16 = vreinterpretq_s16_u16(q5u16);
+  q6s16 = vreinterpretq_s16_u16(q6u16);
+  q7s16 = vreinterpretq_s16_u16(q7u16);
+  q8s16 = vreinterpretq_s16_u16(q8u16);
+  q9s16 = vreinterpretq_s16_u16(q9u16);
+  q10s16 = vreinterpretq_s16_u16(q10u16);
 
-    // load first_pass filter
-    dtmps8 = vld1_s8(vp8_sub_pel_filters[xoffset]);
+  q7s16 = vqaddq_s16(q7s16, q3s16);
+  q8s16 = vqaddq_s16(q8s16, q4s16);
+  q9s16 = vqaddq_s16(q9s16, q5s16);
+  q10s16 = vqaddq_s16(q10s16, q6s16);
+
+  d22u8 = vqrshrun_n_s16(q7s16, 7);
+  d23u8 = vqrshrun_n_s16(q8s16, 7);
+  d24u8 = vqrshrun_n_s16(q9s16, 7);
+  d25u8 = vqrshrun_n_s16(q10s16, 7);
+
+  if (yoffset == 0) {  // firstpass_filter8x4_only
+    vst1_u8(dst_ptr, d22u8);
+    dst_ptr += dst_pitch;
+    vst1_u8(dst_ptr, d23u8);
+    dst_ptr += dst_pitch;
+    vst1_u8(dst_ptr, d24u8);
+    dst_ptr += dst_pitch;
+    vst1_u8(dst_ptr, d25u8);
+    return;
+  }
+
+  // First Pass on rest 5-line data
+  src += src_pixels_per_line;
+  q3u8 = vld1q_u8(src);
+  src += src_pixels_per_line;
+  q4u8 = vld1q_u8(src);
+  src += src_pixels_per_line;
+  q5u8 = vld1q_u8(src);
+  src += src_pixels_per_line;
+  q6u8 = vld1q_u8(src);
+  src += src_pixels_per_line;
+  q7u8 = vld1q_u8(src);
+
+  q8u16 = vmull_u8(vget_low_u8(q3u8), d0u8);
+  q9u16 = vmull_u8(vget_low_u8(q4u8), d0u8);
+  q10u16 = vmull_u8(vget_low_u8(q5u8), d0u8);
+  q11u16 = vmull_u8(vget_low_u8(q6u8), d0u8);
+  q12u16 = vmull_u8(vget_low_u8(q7u8), d0u8);
+
+  d27u8 = vext_u8(vget_low_u8(q3u8), vget_high_u8(q3u8), 1);
+  d28u8 = vext_u8(vget_low_u8(q4u8), vget_high_u8(q4u8), 1);
+  d29u8 = vext_u8(vget_low_u8(q5u8), vget_high_u8(q5u8), 1);
+  d30u8 = vext_u8(vget_low_u8(q6u8), vget_high_u8(q6u8), 1);
+  d31u8 = vext_u8(vget_low_u8(q7u8), vget_high_u8(q7u8), 1);
+
+  q8u16 = vmlsl_u8(q8u16, d27u8, d1u8);
+  q9u16 = vmlsl_u8(q9u16, d28u8, d1u8);
+  q10u16 = vmlsl_u8(q10u16, d29u8, d1u8);
+  q11u16 = vmlsl_u8(q11u16, d30u8, d1u8);
+  q12u16 = vmlsl_u8(q12u16, d31u8, d1u8);
+
+  d27u8 = vext_u8(vget_low_u8(q3u8), vget_high_u8(q3u8), 4);
+  d28u8 = vext_u8(vget_low_u8(q4u8), vget_high_u8(q4u8), 4);
+  d29u8 = vext_u8(vget_low_u8(q5u8), vget_high_u8(q5u8), 4);
+  d30u8 = vext_u8(vget_low_u8(q6u8), vget_high_u8(q6u8), 4);
+  d31u8 = vext_u8(vget_low_u8(q7u8), vget_high_u8(q7u8), 4);
+
+  q8u16 = vmlsl_u8(q8u16, d27u8, d4u8);
+  q9u16 = vmlsl_u8(q9u16, d28u8, d4u8);
+  q10u16 = vmlsl_u8(q10u16, d29u8, d4u8);
+  q11u16 = vmlsl_u8(q11u16, d30u8, d4u8);
+  q12u16 = vmlsl_u8(q12u16, d31u8, d4u8);
+
+  d27u8 = vext_u8(vget_low_u8(q3u8), vget_high_u8(q3u8), 2);
+  d28u8 = vext_u8(vget_low_u8(q4u8), vget_high_u8(q4u8), 2);
+  d29u8 = vext_u8(vget_low_u8(q5u8), vget_high_u8(q5u8), 2);
+  d30u8 = vext_u8(vget_low_u8(q6u8), vget_high_u8(q6u8), 2);
+  d31u8 = vext_u8(vget_low_u8(q7u8), vget_high_u8(q7u8), 2);
+
+  q8u16 = vmlal_u8(q8u16, d27u8, d2u8);
+  q9u16 = vmlal_u8(q9u16, d28u8, d2u8);
+  q10u16 = vmlal_u8(q10u16, d29u8, d2u8);
+  q11u16 = vmlal_u8(q11u16, d30u8, d2u8);
+  q12u16 = vmlal_u8(q12u16, d31u8, d2u8);
+
+  d27u8 = vext_u8(vget_low_u8(q3u8), vget_high_u8(q3u8), 5);
+  d28u8 = vext_u8(vget_low_u8(q4u8), vget_high_u8(q4u8), 5);
+  d29u8 = vext_u8(vget_low_u8(q5u8), vget_high_u8(q5u8), 5);
+  d30u8 = vext_u8(vget_low_u8(q6u8), vget_high_u8(q6u8), 5);
+  d31u8 = vext_u8(vget_low_u8(q7u8), vget_high_u8(q7u8), 5);
+
+  q8u16 = vmlal_u8(q8u16, d27u8, d5u8);
+  q9u16 = vmlal_u8(q9u16, d28u8, d5u8);
+  q10u16 = vmlal_u8(q10u16, d29u8, d5u8);
+  q11u16 = vmlal_u8(q11u16, d30u8, d5u8);
+  q12u16 = vmlal_u8(q12u16, d31u8, d5u8);
+
+  d27u8 = vext_u8(vget_low_u8(q3u8), vget_high_u8(q3u8), 3);
+  d28u8 = vext_u8(vget_low_u8(q4u8), vget_high_u8(q4u8), 3);
+  d29u8 = vext_u8(vget_low_u8(q5u8), vget_high_u8(q5u8), 3);
+  d30u8 = vext_u8(vget_low_u8(q6u8), vget_high_u8(q6u8), 3);
+  d31u8 = vext_u8(vget_low_u8(q7u8), vget_high_u8(q7u8), 3);
+
+  q3u16 = vmull_u8(d27u8, d3u8);
+  q4u16 = vmull_u8(d28u8, d3u8);
+  q5u16 = vmull_u8(d29u8, d3u8);
+  q6u16 = vmull_u8(d30u8, d3u8);
+  q7u16 = vmull_u8(d31u8, d3u8);
+
+  q3s16 = vreinterpretq_s16_u16(q3u16);
+  q4s16 = vreinterpretq_s16_u16(q4u16);
+  q5s16 = vreinterpretq_s16_u16(q5u16);
+  q6s16 = vreinterpretq_s16_u16(q6u16);
+  q7s16 = vreinterpretq_s16_u16(q7u16);
+  q8s16 = vreinterpretq_s16_u16(q8u16);
+  q9s16 = vreinterpretq_s16_u16(q9u16);
+  q10s16 = vreinterpretq_s16_u16(q10u16);
+  q11s16 = vreinterpretq_s16_u16(q11u16);
+  q12s16 = vreinterpretq_s16_u16(q12u16);
+
+  q8s16 = vqaddq_s16(q8s16, q3s16);
+  q9s16 = vqaddq_s16(q9s16, q4s16);
+  q10s16 = vqaddq_s16(q10s16, q5s16);
+  q11s16 = vqaddq_s16(q11s16, q6s16);
+  q12s16 = vqaddq_s16(q12s16, q7s16);
+
+  d26u8 = vqrshrun_n_s16(q8s16, 7);
+  d27u8 = vqrshrun_n_s16(q9s16, 7);
+  d28u8 = vqrshrun_n_s16(q10s16, 7);
+  d29u8 = vqrshrun_n_s16(q11s16, 7);
+  d30u8 = vqrshrun_n_s16(q12s16, 7);
+
+  // Second pass: 8x4
+  dtmps8 = vld1_s8(vp8_sub_pel_filters[yoffset]);
+  d0s8 = vdup_lane_s8(dtmps8, 0);
+  d1s8 = vdup_lane_s8(dtmps8, 1);
+  d2s8 = vdup_lane_s8(dtmps8, 2);
+  d3s8 = vdup_lane_s8(dtmps8, 3);
+  d4s8 = vdup_lane_s8(dtmps8, 4);
+  d5s8 = vdup_lane_s8(dtmps8, 5);
+  d0u8 = vreinterpret_u8_s8(vabs_s8(d0s8));
+  d1u8 = vreinterpret_u8_s8(vabs_s8(d1s8));
+  d2u8 = vreinterpret_u8_s8(vabs_s8(d2s8));
+  d3u8 = vreinterpret_u8_s8(vabs_s8(d3s8));
+  d4u8 = vreinterpret_u8_s8(vabs_s8(d4s8));
+  d5u8 = vreinterpret_u8_s8(vabs_s8(d5s8));
+
+  q3u16 = vmull_u8(d22u8, d0u8);
+  q4u16 = vmull_u8(d23u8, d0u8);
+  q5u16 = vmull_u8(d24u8, d0u8);
+  q6u16 = vmull_u8(d25u8, d0u8);
+
+  q3u16 = vmlsl_u8(q3u16, d23u8, d1u8);
+  q4u16 = vmlsl_u8(q4u16, d24u8, d1u8);
+  q5u16 = vmlsl_u8(q5u16, d25u8, d1u8);
+  q6u16 = vmlsl_u8(q6u16, d26u8, d1u8);
+
+  q3u16 = vmlsl_u8(q3u16, d26u8, d4u8);
+  q4u16 = vmlsl_u8(q4u16, d27u8, d4u8);
+  q5u16 = vmlsl_u8(q5u16, d28u8, d4u8);
+  q6u16 = vmlsl_u8(q6u16, d29u8, d4u8);
+
+  q3u16 = vmlal_u8(q3u16, d24u8, d2u8);
+  q4u16 = vmlal_u8(q4u16, d25u8, d2u8);
+  q5u16 = vmlal_u8(q5u16, d26u8, d2u8);
+  q6u16 = vmlal_u8(q6u16, d27u8, d2u8);
+
+  q3u16 = vmlal_u8(q3u16, d27u8, d5u8);
+  q4u16 = vmlal_u8(q4u16, d28u8, d5u8);
+  q5u16 = vmlal_u8(q5u16, d29u8, d5u8);
+  q6u16 = vmlal_u8(q6u16, d30u8, d5u8);
+
+  q7u16 = vmull_u8(d25u8, d3u8);
+  q8u16 = vmull_u8(d26u8, d3u8);
+  q9u16 = vmull_u8(d27u8, d3u8);
+  q10u16 = vmull_u8(d28u8, d3u8);
+
+  q3s16 = vreinterpretq_s16_u16(q3u16);
+  q4s16 = vreinterpretq_s16_u16(q4u16);
+  q5s16 = vreinterpretq_s16_u16(q5u16);
+  q6s16 = vreinterpretq_s16_u16(q6u16);
+  q7s16 = vreinterpretq_s16_u16(q7u16);
+  q8s16 = vreinterpretq_s16_u16(q8u16);
+  q9s16 = vreinterpretq_s16_u16(q9u16);
+  q10s16 = vreinterpretq_s16_u16(q10u16);
+
+  q7s16 = vqaddq_s16(q7s16, q3s16);
+  q8s16 = vqaddq_s16(q8s16, q4s16);
+  q9s16 = vqaddq_s16(q9s16, q5s16);
+  q10s16 = vqaddq_s16(q10s16, q6s16);
+
+  d6u8 = vqrshrun_n_s16(q7s16, 7);
+  d7u8 = vqrshrun_n_s16(q8s16, 7);
+  d8u8 = vqrshrun_n_s16(q9s16, 7);
+  d9u8 = vqrshrun_n_s16(q10s16, 7);
+
+  vst1_u8(dst_ptr, d6u8);
+  dst_ptr += dst_pitch;
+  vst1_u8(dst_ptr, d7u8);
+  dst_ptr += dst_pitch;
+  vst1_u8(dst_ptr, d8u8);
+  dst_ptr += dst_pitch;
+  vst1_u8(dst_ptr, d9u8);
+  return;
+}
+
+void vp8_sixtap_predict8x8_neon(unsigned char *src_ptr, int src_pixels_per_line,
+                                int xoffset, int yoffset,
+                                unsigned char *dst_ptr, int dst_pitch) {
+  unsigned char *src, *tmpp;
+  unsigned char tmp[64];
+  int i;
+  uint8x8_t d0u8, d1u8, d2u8, d3u8, d4u8, d5u8, d6u8, d7u8, d8u8, d9u8;
+  uint8x8_t d18u8, d19u8, d20u8, d21u8, d22u8, d23u8, d24u8, d25u8;
+  uint8x8_t d26u8, d27u8, d28u8, d29u8, d30u8, d31u8;
+  int8x8_t dtmps8, d0s8, d1s8, d2s8, d3s8, d4s8, d5s8;
+  uint16x8_t q3u16, q4u16, q5u16, q6u16, q7u16;
+  uint16x8_t q8u16, q9u16, q10u16, q11u16, q12u16;
+  int16x8_t q3s16, q4s16, q5s16, q6s16, q7s16;
+  int16x8_t q8s16, q9s16, q10s16, q11s16, q12s16;
+  uint8x16_t q3u8, q4u8, q5u8, q6u8, q7u8, q9u8, q10u8, q11u8, q12u8;
+
+  if (xoffset == 0) {  // secondpass_filter8x8_only
+    // load second_pass filter
+    dtmps8 = vld1_s8(vp8_sub_pel_filters[yoffset]);
     d0s8 = vdup_lane_s8(dtmps8, 0);
     d1s8 = vdup_lane_s8(dtmps8, 1);
     d2s8 = vdup_lane_s8(dtmps8, 2);
@@ -574,125 +451,129 @@
     d4u8 = vreinterpret_u8_s8(vabs_s8(d4s8));
     d5u8 = vreinterpret_u8_s8(vabs_s8(d5s8));
 
-    // First pass: output_height lines x output_width columns (9x4)
-    if (yoffset == 0)  // firstpass_filter4x4_only
-        src = src_ptr - 2;
-    else
-        src = src_ptr - 2 - (src_pixels_per_line * 2);
+    // load src data
+    src = src_ptr - src_pixels_per_line * 2;
+    d18u8 = vld1_u8(src);
+    src += src_pixels_per_line;
+    d19u8 = vld1_u8(src);
+    src += src_pixels_per_line;
+    d20u8 = vld1_u8(src);
+    src += src_pixels_per_line;
+    d21u8 = vld1_u8(src);
+    src += src_pixels_per_line;
+    d22u8 = vld1_u8(src);
+    src += src_pixels_per_line;
+    d23u8 = vld1_u8(src);
+    src += src_pixels_per_line;
+    d24u8 = vld1_u8(src);
+    src += src_pixels_per_line;
+    d25u8 = vld1_u8(src);
+    src += src_pixels_per_line;
+    d26u8 = vld1_u8(src);
+    src += src_pixels_per_line;
+    d27u8 = vld1_u8(src);
+    src += src_pixels_per_line;
+    d28u8 = vld1_u8(src);
+    src += src_pixels_per_line;
+    d29u8 = vld1_u8(src);
+    src += src_pixels_per_line;
+    d30u8 = vld1_u8(src);
 
-    tmpp = tmp;
     for (i = 2; i > 0; i--) {
-        q3u8 = vld1q_u8(src);
-        src += src_pixels_per_line;
-        q4u8 = vld1q_u8(src);
-        src += src_pixels_per_line;
-        q5u8 = vld1q_u8(src);
-        src += src_pixels_per_line;
-        q6u8 = vld1q_u8(src);
-        src += src_pixels_per_line;
+      q3u16 = vmull_u8(d18u8, d0u8);
+      q4u16 = vmull_u8(d19u8, d0u8);
+      q5u16 = vmull_u8(d20u8, d0u8);
+      q6u16 = vmull_u8(d21u8, d0u8);
 
-        __builtin_prefetch(src);
-        __builtin_prefetch(src + src_pixels_per_line);
-        __builtin_prefetch(src + src_pixels_per_line * 2);
+      q3u16 = vmlsl_u8(q3u16, d19u8, d1u8);
+      q4u16 = vmlsl_u8(q4u16, d20u8, d1u8);
+      q5u16 = vmlsl_u8(q5u16, d21u8, d1u8);
+      q6u16 = vmlsl_u8(q6u16, d22u8, d1u8);
 
-        q7u16 = vmull_u8(vget_low_u8(q3u8), d0u8);
-        q8u16 = vmull_u8(vget_low_u8(q4u8), d0u8);
-        q9u16 = vmull_u8(vget_low_u8(q5u8), d0u8);
-        q10u16 = vmull_u8(vget_low_u8(q6u8), d0u8);
+      q3u16 = vmlsl_u8(q3u16, d22u8, d4u8);
+      q4u16 = vmlsl_u8(q4u16, d23u8, d4u8);
+      q5u16 = vmlsl_u8(q5u16, d24u8, d4u8);
+      q6u16 = vmlsl_u8(q6u16, d25u8, d4u8);
 
-        d28u8 = vext_u8(vget_low_u8(q3u8), vget_high_u8(q3u8), 1);
-        d29u8 = vext_u8(vget_low_u8(q4u8), vget_high_u8(q4u8), 1);
-        d30u8 = vext_u8(vget_low_u8(q5u8), vget_high_u8(q5u8), 1);
-        d31u8 = vext_u8(vget_low_u8(q6u8), vget_high_u8(q6u8), 1);
+      q3u16 = vmlal_u8(q3u16, d20u8, d2u8);
+      q4u16 = vmlal_u8(q4u16, d21u8, d2u8);
+      q5u16 = vmlal_u8(q5u16, d22u8, d2u8);
+      q6u16 = vmlal_u8(q6u16, d23u8, d2u8);
 
-        q7u16 = vmlsl_u8(q7u16, d28u8, d1u8);
-        q8u16 = vmlsl_u8(q8u16, d29u8, d1u8);
-        q9u16 = vmlsl_u8(q9u16, d30u8, d1u8);
-        q10u16 = vmlsl_u8(q10u16, d31u8, d1u8);
+      q3u16 = vmlal_u8(q3u16, d23u8, d5u8);
+      q4u16 = vmlal_u8(q4u16, d24u8, d5u8);
+      q5u16 = vmlal_u8(q5u16, d25u8, d5u8);
+      q6u16 = vmlal_u8(q6u16, d26u8, d5u8);
 
-        d28u8 = vext_u8(vget_low_u8(q3u8), vget_high_u8(q3u8), 4);
-        d29u8 = vext_u8(vget_low_u8(q4u8), vget_high_u8(q4u8), 4);
-        d30u8 = vext_u8(vget_low_u8(q5u8), vget_high_u8(q5u8), 4);
-        d31u8 = vext_u8(vget_low_u8(q6u8), vget_high_u8(q6u8), 4);
+      q7u16 = vmull_u8(d21u8, d3u8);
+      q8u16 = vmull_u8(d22u8, d3u8);
+      q9u16 = vmull_u8(d23u8, d3u8);
+      q10u16 = vmull_u8(d24u8, d3u8);
 
-        q7u16 = vmlsl_u8(q7u16, d28u8, d4u8);
-        q8u16 = vmlsl_u8(q8u16, d29u8, d4u8);
-        q9u16 = vmlsl_u8(q9u16, d30u8, d4u8);
-        q10u16 = vmlsl_u8(q10u16, d31u8, d4u8);
+      q3s16 = vreinterpretq_s16_u16(q3u16);
+      q4s16 = vreinterpretq_s16_u16(q4u16);
+      q5s16 = vreinterpretq_s16_u16(q5u16);
+      q6s16 = vreinterpretq_s16_u16(q6u16);
+      q7s16 = vreinterpretq_s16_u16(q7u16);
+      q8s16 = vreinterpretq_s16_u16(q8u16);
+      q9s16 = vreinterpretq_s16_u16(q9u16);
+      q10s16 = vreinterpretq_s16_u16(q10u16);
 
-        d28u8 = vext_u8(vget_low_u8(q3u8), vget_high_u8(q3u8), 2);
-        d29u8 = vext_u8(vget_low_u8(q4u8), vget_high_u8(q4u8), 2);
-        d30u8 = vext_u8(vget_low_u8(q5u8), vget_high_u8(q5u8), 2);
-        d31u8 = vext_u8(vget_low_u8(q6u8), vget_high_u8(q6u8), 2);
+      q7s16 = vqaddq_s16(q7s16, q3s16);
+      q8s16 = vqaddq_s16(q8s16, q4s16);
+      q9s16 = vqaddq_s16(q9s16, q5s16);
+      q10s16 = vqaddq_s16(q10s16, q6s16);
 
-        q7u16 = vmlal_u8(q7u16, d28u8, d2u8);
-        q8u16 = vmlal_u8(q8u16, d29u8, d2u8);
-        q9u16 = vmlal_u8(q9u16, d30u8, d2u8);
-        q10u16 = vmlal_u8(q10u16, d31u8, d2u8);
+      d6u8 = vqrshrun_n_s16(q7s16, 7);
+      d7u8 = vqrshrun_n_s16(q8s16, 7);
+      d8u8 = vqrshrun_n_s16(q9s16, 7);
+      d9u8 = vqrshrun_n_s16(q10s16, 7);
 
-        d28u8 = vext_u8(vget_low_u8(q3u8), vget_high_u8(q3u8), 5);
-        d29u8 = vext_u8(vget_low_u8(q4u8), vget_high_u8(q4u8), 5);
-        d30u8 = vext_u8(vget_low_u8(q5u8), vget_high_u8(q5u8), 5);
-        d31u8 = vext_u8(vget_low_u8(q6u8), vget_high_u8(q6u8), 5);
+      d18u8 = d22u8;
+      d19u8 = d23u8;
+      d20u8 = d24u8;
+      d21u8 = d25u8;
+      d22u8 = d26u8;
+      d23u8 = d27u8;
+      d24u8 = d28u8;
+      d25u8 = d29u8;
+      d26u8 = d30u8;
 
-        q7u16 = vmlal_u8(q7u16, d28u8, d5u8);
-        q8u16 = vmlal_u8(q8u16, d29u8, d5u8);
-        q9u16 = vmlal_u8(q9u16, d30u8, d5u8);
-        q10u16 = vmlal_u8(q10u16, d31u8, d5u8);
+      vst1_u8(dst_ptr, d6u8);
+      dst_ptr += dst_pitch;
+      vst1_u8(dst_ptr, d7u8);
+      dst_ptr += dst_pitch;
+      vst1_u8(dst_ptr, d8u8);
+      dst_ptr += dst_pitch;
+      vst1_u8(dst_ptr, d9u8);
+      dst_ptr += dst_pitch;
+    }
+    return;
+  }
 
-        d28u8 = vext_u8(vget_low_u8(q3u8), vget_high_u8(q3u8), 3);
-        d29u8 = vext_u8(vget_low_u8(q4u8), vget_high_u8(q4u8), 3);
-        d30u8 = vext_u8(vget_low_u8(q5u8), vget_high_u8(q5u8), 3);
-        d31u8 = vext_u8(vget_low_u8(q6u8), vget_high_u8(q6u8), 3);
+  // load first_pass filter
+  dtmps8 = vld1_s8(vp8_sub_pel_filters[xoffset]);
+  d0s8 = vdup_lane_s8(dtmps8, 0);
+  d1s8 = vdup_lane_s8(dtmps8, 1);
+  d2s8 = vdup_lane_s8(dtmps8, 2);
+  d3s8 = vdup_lane_s8(dtmps8, 3);
+  d4s8 = vdup_lane_s8(dtmps8, 4);
+  d5s8 = vdup_lane_s8(dtmps8, 5);
+  d0u8 = vreinterpret_u8_s8(vabs_s8(d0s8));
+  d1u8 = vreinterpret_u8_s8(vabs_s8(d1s8));
+  d2u8 = vreinterpret_u8_s8(vabs_s8(d2s8));
+  d3u8 = vreinterpret_u8_s8(vabs_s8(d3s8));
+  d4u8 = vreinterpret_u8_s8(vabs_s8(d4s8));
+  d5u8 = vreinterpret_u8_s8(vabs_s8(d5s8));
 
-        q3u16 = vmull_u8(d28u8, d3u8);
-        q4u16 = vmull_u8(d29u8, d3u8);
-        q5u16 = vmull_u8(d30u8, d3u8);
-        q6u16 = vmull_u8(d31u8, d3u8);
+  // First pass: output_height lines x output_width columns (9x4)
+  if (yoffset == 0)  // firstpass_filter4x4_only
+    src = src_ptr - 2;
+  else
+    src = src_ptr - 2 - (src_pixels_per_line * 2);
 
-        q3s16 = vreinterpretq_s16_u16(q3u16);
-        q4s16 = vreinterpretq_s16_u16(q4u16);
-        q5s16 = vreinterpretq_s16_u16(q5u16);
-        q6s16 = vreinterpretq_s16_u16(q6u16);
-        q7s16 = vreinterpretq_s16_u16(q7u16);
-        q8s16 = vreinterpretq_s16_u16(q8u16);
-        q9s16 = vreinterpretq_s16_u16(q9u16);
-        q10s16 = vreinterpretq_s16_u16(q10u16);
-
-        q7s16 = vqaddq_s16(q7s16, q3s16);
-        q8s16 = vqaddq_s16(q8s16, q4s16);
-        q9s16 = vqaddq_s16(q9s16, q5s16);
-        q10s16 = vqaddq_s16(q10s16, q6s16);
-
-        d22u8 = vqrshrun_n_s16(q7s16, 7);
-        d23u8 = vqrshrun_n_s16(q8s16, 7);
-        d24u8 = vqrshrun_n_s16(q9s16, 7);
-        d25u8 = vqrshrun_n_s16(q10s16, 7);
-
-        if (yoffset == 0) {  // firstpass_filter8x4_only
-            vst1_u8(dst_ptr, d22u8);
-            dst_ptr += dst_pitch;
-            vst1_u8(dst_ptr, d23u8);
-            dst_ptr += dst_pitch;
-            vst1_u8(dst_ptr, d24u8);
-            dst_ptr += dst_pitch;
-            vst1_u8(dst_ptr, d25u8);
-            dst_ptr += dst_pitch;
-        } else {
-            vst1_u8(tmpp, d22u8);
-            tmpp += 8;
-            vst1_u8(tmpp, d23u8);
-            tmpp += 8;
-            vst1_u8(tmpp, d24u8);
-            tmpp += 8;
-            vst1_u8(tmpp, d25u8);
-            tmpp += 8;
-        }
-    }
-    if (yoffset == 0)
-        return;
-
-    // First Pass on rest 5-line data
+  tmpp = tmp;
+  for (i = 2; i > 0; i--) {
     q3u8 = vld1q_u8(src);
     src += src_pixels_per_line;
     q4u8 = vld1q_u8(src);
@@ -701,73 +582,270 @@
     src += src_pixels_per_line;
     q6u8 = vld1q_u8(src);
     src += src_pixels_per_line;
-    q7u8 = vld1q_u8(src);
 
-    q8u16 = vmull_u8(vget_low_u8(q3u8), d0u8);
-    q9u16 = vmull_u8(vget_low_u8(q4u8), d0u8);
-    q10u16 = vmull_u8(vget_low_u8(q5u8), d0u8);
-    q11u16 = vmull_u8(vget_low_u8(q6u8), d0u8);
-    q12u16 = vmull_u8(vget_low_u8(q7u8), d0u8);
+    __builtin_prefetch(src);
+    __builtin_prefetch(src + src_pixels_per_line);
+    __builtin_prefetch(src + src_pixels_per_line * 2);
 
-    d27u8 = vext_u8(vget_low_u8(q3u8), vget_high_u8(q3u8), 1);
-    d28u8 = vext_u8(vget_low_u8(q4u8), vget_high_u8(q4u8), 1);
-    d29u8 = vext_u8(vget_low_u8(q5u8), vget_high_u8(q5u8), 1);
-    d30u8 = vext_u8(vget_low_u8(q6u8), vget_high_u8(q6u8), 1);
-    d31u8 = vext_u8(vget_low_u8(q7u8), vget_high_u8(q7u8), 1);
+    q7u16 = vmull_u8(vget_low_u8(q3u8), d0u8);
+    q8u16 = vmull_u8(vget_low_u8(q4u8), d0u8);
+    q9u16 = vmull_u8(vget_low_u8(q5u8), d0u8);
+    q10u16 = vmull_u8(vget_low_u8(q6u8), d0u8);
 
-    q8u16 = vmlsl_u8(q8u16, d27u8, d1u8);
-    q9u16 = vmlsl_u8(q9u16, d28u8, d1u8);
-    q10u16 = vmlsl_u8(q10u16, d29u8, d1u8);
-    q11u16 = vmlsl_u8(q11u16, d30u8, d1u8);
-    q12u16 = vmlsl_u8(q12u16, d31u8, d1u8);
+    d28u8 = vext_u8(vget_low_u8(q3u8), vget_high_u8(q3u8), 1);
+    d29u8 = vext_u8(vget_low_u8(q4u8), vget_high_u8(q4u8), 1);
+    d30u8 = vext_u8(vget_low_u8(q5u8), vget_high_u8(q5u8), 1);
+    d31u8 = vext_u8(vget_low_u8(q6u8), vget_high_u8(q6u8), 1);
 
-    d27u8 = vext_u8(vget_low_u8(q3u8), vget_high_u8(q3u8), 4);
-    d28u8 = vext_u8(vget_low_u8(q4u8), vget_high_u8(q4u8), 4);
-    d29u8 = vext_u8(vget_low_u8(q5u8), vget_high_u8(q5u8), 4);
-    d30u8 = vext_u8(vget_low_u8(q6u8), vget_high_u8(q6u8), 4);
-    d31u8 = vext_u8(vget_low_u8(q7u8), vget_high_u8(q7u8), 4);
+    q7u16 = vmlsl_u8(q7u16, d28u8, d1u8);
+    q8u16 = vmlsl_u8(q8u16, d29u8, d1u8);
+    q9u16 = vmlsl_u8(q9u16, d30u8, d1u8);
+    q10u16 = vmlsl_u8(q10u16, d31u8, d1u8);
 
-    q8u16 = vmlsl_u8(q8u16, d27u8, d4u8);
-    q9u16 = vmlsl_u8(q9u16, d28u8, d4u8);
-    q10u16 = vmlsl_u8(q10u16, d29u8, d4u8);
-    q11u16 = vmlsl_u8(q11u16, d30u8, d4u8);
-    q12u16 = vmlsl_u8(q12u16, d31u8, d4u8);
+    d28u8 = vext_u8(vget_low_u8(q3u8), vget_high_u8(q3u8), 4);
+    d29u8 = vext_u8(vget_low_u8(q4u8), vget_high_u8(q4u8), 4);
+    d30u8 = vext_u8(vget_low_u8(q5u8), vget_high_u8(q5u8), 4);
+    d31u8 = vext_u8(vget_low_u8(q6u8), vget_high_u8(q6u8), 4);
+
+    q7u16 = vmlsl_u8(q7u16, d28u8, d4u8);
+    q8u16 = vmlsl_u8(q8u16, d29u8, d4u8);
+    q9u16 = vmlsl_u8(q9u16, d30u8, d4u8);
+    q10u16 = vmlsl_u8(q10u16, d31u8, d4u8);
+
+    d28u8 = vext_u8(vget_low_u8(q3u8), vget_high_u8(q3u8), 2);
+    d29u8 = vext_u8(vget_low_u8(q4u8), vget_high_u8(q4u8), 2);
+    d30u8 = vext_u8(vget_low_u8(q5u8), vget_high_u8(q5u8), 2);
+    d31u8 = vext_u8(vget_low_u8(q6u8), vget_high_u8(q6u8), 2);
+
+    q7u16 = vmlal_u8(q7u16, d28u8, d2u8);
+    q8u16 = vmlal_u8(q8u16, d29u8, d2u8);
+    q9u16 = vmlal_u8(q9u16, d30u8, d2u8);
+    q10u16 = vmlal_u8(q10u16, d31u8, d2u8);
+
+    d28u8 = vext_u8(vget_low_u8(q3u8), vget_high_u8(q3u8), 5);
+    d29u8 = vext_u8(vget_low_u8(q4u8), vget_high_u8(q4u8), 5);
+    d30u8 = vext_u8(vget_low_u8(q5u8), vget_high_u8(q5u8), 5);
+    d31u8 = vext_u8(vget_low_u8(q6u8), vget_high_u8(q6u8), 5);
+
+    q7u16 = vmlal_u8(q7u16, d28u8, d5u8);
+    q8u16 = vmlal_u8(q8u16, d29u8, d5u8);
+    q9u16 = vmlal_u8(q9u16, d30u8, d5u8);
+    q10u16 = vmlal_u8(q10u16, d31u8, d5u8);
 
-    d27u8 = vext_u8(vget_low_u8(q3u8), vget_high_u8(q3u8), 2);
-    d28u8 = vext_u8(vget_low_u8(q4u8), vget_high_u8(q4u8), 2);
-    d29u8 = vext_u8(vget_low_u8(q5u8), vget_high_u8(q5u8), 2);
-    d30u8 = vext_u8(vget_low_u8(q6u8), vget_high_u8(q6u8), 2);
-    d31u8 = vext_u8(vget_low_u8(q7u8), vget_high_u8(q7u8), 2);
+    d28u8 = vext_u8(vget_low_u8(q3u8), vget_high_u8(q3u8), 3);
+    d29u8 = vext_u8(vget_low_u8(q4u8), vget_high_u8(q4u8), 3);
+    d30u8 = vext_u8(vget_low_u8(q5u8), vget_high_u8(q5u8), 3);
+    d31u8 = vext_u8(vget_low_u8(q6u8), vget_high_u8(q6u8), 3);
 
-    q8u16 = vmlal_u8(q8u16, d27u8, d2u8);
-    q9u16 = vmlal_u8(q9u16, d28u8, d2u8);
-    q10u16 = vmlal_u8(q10u16, d29u8, d2u8);
-    q11u16 = vmlal_u8(q11u16, d30u8, d2u8);
-    q12u16 = vmlal_u8(q12u16, d31u8, d2u8);
+    q3u16 = vmull_u8(d28u8, d3u8);
+    q4u16 = vmull_u8(d29u8, d3u8);
+    q5u16 = vmull_u8(d30u8, d3u8);
+    q6u16 = vmull_u8(d31u8, d3u8);
 
-    d27u8 = vext_u8(vget_low_u8(q3u8), vget_high_u8(q3u8), 5);
-    d28u8 = vext_u8(vget_low_u8(q4u8), vget_high_u8(q4u8), 5);
-    d29u8 = vext_u8(vget_low_u8(q5u8), vget_high_u8(q5u8), 5);
-    d30u8 = vext_u8(vget_low_u8(q6u8), vget_high_u8(q6u8), 5);
-    d31u8 = vext_u8(vget_low_u8(q7u8), vget_high_u8(q7u8), 5);
+    q3s16 = vreinterpretq_s16_u16(q3u16);
+    q4s16 = vreinterpretq_s16_u16(q4u16);
+    q5s16 = vreinterpretq_s16_u16(q5u16);
+    q6s16 = vreinterpretq_s16_u16(q6u16);
+    q7s16 = vreinterpretq_s16_u16(q7u16);
+    q8s16 = vreinterpretq_s16_u16(q8u16);
+    q9s16 = vreinterpretq_s16_u16(q9u16);
+    q10s16 = vreinterpretq_s16_u16(q10u16);
 
-    q8u16 = vmlal_u8(q8u16, d27u8, d5u8);
-    q9u16 = vmlal_u8(q9u16, d28u8, d5u8);
-    q10u16 = vmlal_u8(q10u16, d29u8, d5u8);
-    q11u16 = vmlal_u8(q11u16, d30u8, d5u8);
-    q12u16 = vmlal_u8(q12u16, d31u8, d5u8);
+    q7s16 = vqaddq_s16(q7s16, q3s16);
+    q8s16 = vqaddq_s16(q8s16, q4s16);
+    q9s16 = vqaddq_s16(q9s16, q5s16);
+    q10s16 = vqaddq_s16(q10s16, q6s16);
 
-    d27u8 = vext_u8(vget_low_u8(q3u8), vget_high_u8(q3u8), 3);
-    d28u8 = vext_u8(vget_low_u8(q4u8), vget_high_u8(q4u8), 3);
-    d29u8 = vext_u8(vget_low_u8(q5u8), vget_high_u8(q5u8), 3);
-    d30u8 = vext_u8(vget_low_u8(q6u8), vget_high_u8(q6u8), 3);
-    d31u8 = vext_u8(vget_low_u8(q7u8), vget_high_u8(q7u8), 3);
+    d22u8 = vqrshrun_n_s16(q7s16, 7);
+    d23u8 = vqrshrun_n_s16(q8s16, 7);
+    d24u8 = vqrshrun_n_s16(q9s16, 7);
+    d25u8 = vqrshrun_n_s16(q10s16, 7);
 
-    q3u16 = vmull_u8(d27u8, d3u8);
-    q4u16 = vmull_u8(d28u8, d3u8);
-    q5u16 = vmull_u8(d29u8, d3u8);
-    q6u16 = vmull_u8(d30u8, d3u8);
-    q7u16 = vmull_u8(d31u8, d3u8);
+    if (yoffset == 0) {  // firstpass_filter8x4_only
+      vst1_u8(dst_ptr, d22u8);
+      dst_ptr += dst_pitch;
+      vst1_u8(dst_ptr, d23u8);
+      dst_ptr += dst_pitch;
+      vst1_u8(dst_ptr, d24u8);
+      dst_ptr += dst_pitch;
+      vst1_u8(dst_ptr, d25u8);
+      dst_ptr += dst_pitch;
+    } else {
+      vst1_u8(tmpp, d22u8);
+      tmpp += 8;
+      vst1_u8(tmpp, d23u8);
+      tmpp += 8;
+      vst1_u8(tmpp, d24u8);
+      tmpp += 8;
+      vst1_u8(tmpp, d25u8);
+      tmpp += 8;
+    }
+  }
+  if (yoffset == 0) return;
+
+  // First Pass on rest 5-line data
+  q3u8 = vld1q_u8(src);
+  src += src_pixels_per_line;
+  q4u8 = vld1q_u8(src);
+  src += src_pixels_per_line;
+  q5u8 = vld1q_u8(src);
+  src += src_pixels_per_line;
+  q6u8 = vld1q_u8(src);
+  src += src_pixels_per_line;
+  q7u8 = vld1q_u8(src);
+
+  q8u16 = vmull_u8(vget_low_u8(q3u8), d0u8);
+  q9u16 = vmull_u8(vget_low_u8(q4u8), d0u8);
+  q10u16 = vmull_u8(vget_low_u8(q5u8), d0u8);
+  q11u16 = vmull_u8(vget_low_u8(q6u8), d0u8);
+  q12u16 = vmull_u8(vget_low_u8(q7u8), d0u8);
+
+  d27u8 = vext_u8(vget_low_u8(q3u8), vget_high_u8(q3u8), 1);
+  d28u8 = vext_u8(vget_low_u8(q4u8), vget_high_u8(q4u8), 1);
+  d29u8 = vext_u8(vget_low_u8(q5u8), vget_high_u8(q5u8), 1);
+  d30u8 = vext_u8(vget_low_u8(q6u8), vget_high_u8(q6u8), 1);
+  d31u8 = vext_u8(vget_low_u8(q7u8), vget_high_u8(q7u8), 1);
+
+  q8u16 = vmlsl_u8(q8u16, d27u8, d1u8);
+  q9u16 = vmlsl_u8(q9u16, d28u8, d1u8);
+  q10u16 = vmlsl_u8(q10u16, d29u8, d1u8);
+  q11u16 = vmlsl_u8(q11u16, d30u8, d1u8);
+  q12u16 = vmlsl_u8(q12u16, d31u8, d1u8);
+
+  d27u8 = vext_u8(vget_low_u8(q3u8), vget_high_u8(q3u8), 4);
+  d28u8 = vext_u8(vget_low_u8(q4u8), vget_high_u8(q4u8), 4);
+  d29u8 = vext_u8(vget_low_u8(q5u8), vget_high_u8(q5u8), 4);
+  d30u8 = vext_u8(vget_low_u8(q6u8), vget_high_u8(q6u8), 4);
+  d31u8 = vext_u8(vget_low_u8(q7u8), vget_high_u8(q7u8), 4);
+
+  q8u16 = vmlsl_u8(q8u16, d27u8, d4u8);
+  q9u16 = vmlsl_u8(q9u16, d28u8, d4u8);
+  q10u16 = vmlsl_u8(q10u16, d29u8, d4u8);
+  q11u16 = vmlsl_u8(q11u16, d30u8, d4u8);
+  q12u16 = vmlsl_u8(q12u16, d31u8, d4u8);
+
+  d27u8 = vext_u8(vget_low_u8(q3u8), vget_high_u8(q3u8), 2);
+  d28u8 = vext_u8(vget_low_u8(q4u8), vget_high_u8(q4u8), 2);
+  d29u8 = vext_u8(vget_low_u8(q5u8), vget_high_u8(q5u8), 2);
+  d30u8 = vext_u8(vget_low_u8(q6u8), vget_high_u8(q6u8), 2);
+  d31u8 = vext_u8(vget_low_u8(q7u8), vget_high_u8(q7u8), 2);
+
+  q8u16 = vmlal_u8(q8u16, d27u8, d2u8);
+  q9u16 = vmlal_u8(q9u16, d28u8, d2u8);
+  q10u16 = vmlal_u8(q10u16, d29u8, d2u8);
+  q11u16 = vmlal_u8(q11u16, d30u8, d2u8);
+  q12u16 = vmlal_u8(q12u16, d31u8, d2u8);
+
+  d27u8 = vext_u8(vget_low_u8(q3u8), vget_high_u8(q3u8), 5);
+  d28u8 = vext_u8(vget_low_u8(q4u8), vget_high_u8(q4u8), 5);
+  d29u8 = vext_u8(vget_low_u8(q5u8), vget_high_u8(q5u8), 5);
+  d30u8 = vext_u8(vget_low_u8(q6u8), vget_high_u8(q6u8), 5);
+  d31u8 = vext_u8(vget_low_u8(q7u8), vget_high_u8(q7u8), 5);
+
+  q8u16 = vmlal_u8(q8u16, d27u8, d5u8);
+  q9u16 = vmlal_u8(q9u16, d28u8, d5u8);
+  q10u16 = vmlal_u8(q10u16, d29u8, d5u8);
+  q11u16 = vmlal_u8(q11u16, d30u8, d5u8);
+  q12u16 = vmlal_u8(q12u16, d31u8, d5u8);
+
+  d27u8 = vext_u8(vget_low_u8(q3u8), vget_high_u8(q3u8), 3);
+  d28u8 = vext_u8(vget_low_u8(q4u8), vget_high_u8(q4u8), 3);
+  d29u8 = vext_u8(vget_low_u8(q5u8), vget_high_u8(q5u8), 3);
+  d30u8 = vext_u8(vget_low_u8(q6u8), vget_high_u8(q6u8), 3);
+  d31u8 = vext_u8(vget_low_u8(q7u8), vget_high_u8(q7u8), 3);
+
+  q3u16 = vmull_u8(d27u8, d3u8);
+  q4u16 = vmull_u8(d28u8, d3u8);
+  q5u16 = vmull_u8(d29u8, d3u8);
+  q6u16 = vmull_u8(d30u8, d3u8);
+  q7u16 = vmull_u8(d31u8, d3u8);
+
+  q3s16 = vreinterpretq_s16_u16(q3u16);
+  q4s16 = vreinterpretq_s16_u16(q4u16);
+  q5s16 = vreinterpretq_s16_u16(q5u16);
+  q6s16 = vreinterpretq_s16_u16(q6u16);
+  q7s16 = vreinterpretq_s16_u16(q7u16);
+  q8s16 = vreinterpretq_s16_u16(q8u16);
+  q9s16 = vreinterpretq_s16_u16(q9u16);
+  q10s16 = vreinterpretq_s16_u16(q10u16);
+  q11s16 = vreinterpretq_s16_u16(q11u16);
+  q12s16 = vreinterpretq_s16_u16(q12u16);
+
+  q8s16 = vqaddq_s16(q8s16, q3s16);
+  q9s16 = vqaddq_s16(q9s16, q4s16);
+  q10s16 = vqaddq_s16(q10s16, q5s16);
+  q11s16 = vqaddq_s16(q11s16, q6s16);
+  q12s16 = vqaddq_s16(q12s16, q7s16);
+
+  d26u8 = vqrshrun_n_s16(q8s16, 7);
+  d27u8 = vqrshrun_n_s16(q9s16, 7);
+  d28u8 = vqrshrun_n_s16(q10s16, 7);
+  d29u8 = vqrshrun_n_s16(q11s16, 7);
+  d30u8 = vqrshrun_n_s16(q12s16, 7);
+
+  // Second pass: 8x8
+  dtmps8 = vld1_s8(vp8_sub_pel_filters[yoffset]);
+  d0s8 = vdup_lane_s8(dtmps8, 0);
+  d1s8 = vdup_lane_s8(dtmps8, 1);
+  d2s8 = vdup_lane_s8(dtmps8, 2);
+  d3s8 = vdup_lane_s8(dtmps8, 3);
+  d4s8 = vdup_lane_s8(dtmps8, 4);
+  d5s8 = vdup_lane_s8(dtmps8, 5);
+  d0u8 = vreinterpret_u8_s8(vabs_s8(d0s8));
+  d1u8 = vreinterpret_u8_s8(vabs_s8(d1s8));
+  d2u8 = vreinterpret_u8_s8(vabs_s8(d2s8));
+  d3u8 = vreinterpret_u8_s8(vabs_s8(d3s8));
+  d4u8 = vreinterpret_u8_s8(vabs_s8(d4s8));
+  d5u8 = vreinterpret_u8_s8(vabs_s8(d5s8));
+
+  tmpp = tmp;
+  q9u8 = vld1q_u8(tmpp);
+  tmpp += 16;
+  q10u8 = vld1q_u8(tmpp);
+  tmpp += 16;
+  q11u8 = vld1q_u8(tmpp);
+  tmpp += 16;
+  q12u8 = vld1q_u8(tmpp);
+
+  d18u8 = vget_low_u8(q9u8);
+  d19u8 = vget_high_u8(q9u8);
+  d20u8 = vget_low_u8(q10u8);
+  d21u8 = vget_high_u8(q10u8);
+  d22u8 = vget_low_u8(q11u8);
+  d23u8 = vget_high_u8(q11u8);
+  d24u8 = vget_low_u8(q12u8);
+  d25u8 = vget_high_u8(q12u8);
+
+  for (i = 2; i > 0; i--) {
+    q3u16 = vmull_u8(d18u8, d0u8);
+    q4u16 = vmull_u8(d19u8, d0u8);
+    q5u16 = vmull_u8(d20u8, d0u8);
+    q6u16 = vmull_u8(d21u8, d0u8);
+
+    q3u16 = vmlsl_u8(q3u16, d19u8, d1u8);
+    q4u16 = vmlsl_u8(q4u16, d20u8, d1u8);
+    q5u16 = vmlsl_u8(q5u16, d21u8, d1u8);
+    q6u16 = vmlsl_u8(q6u16, d22u8, d1u8);
+
+    q3u16 = vmlsl_u8(q3u16, d22u8, d4u8);
+    q4u16 = vmlsl_u8(q4u16, d23u8, d4u8);
+    q5u16 = vmlsl_u8(q5u16, d24u8, d4u8);
+    q6u16 = vmlsl_u8(q6u16, d25u8, d4u8);
+
+    q3u16 = vmlal_u8(q3u16, d20u8, d2u8);
+    q4u16 = vmlal_u8(q4u16, d21u8, d2u8);
+    q5u16 = vmlal_u8(q5u16, d22u8, d2u8);
+    q6u16 = vmlal_u8(q6u16, d23u8, d2u8);
+
+    q3u16 = vmlal_u8(q3u16, d23u8, d5u8);
+    q4u16 = vmlal_u8(q4u16, d24u8, d5u8);
+    q5u16 = vmlal_u8(q5u16, d25u8, d5u8);
+    q6u16 = vmlal_u8(q6u16, d26u8, d5u8);
+
+    q7u16 = vmull_u8(d21u8, d3u8);
+    q8u16 = vmull_u8(d22u8, d3u8);
+    q9u16 = vmull_u8(d23u8, d3u8);
+    q10u16 = vmull_u8(d24u8, d3u8);
 
     q3s16 = vreinterpretq_s16_u16(q3u16);
     q4s16 = vreinterpretq_s16_u16(q4u16);
@@ -777,22 +855,59 @@
     q8s16 = vreinterpretq_s16_u16(q8u16);
     q9s16 = vreinterpretq_s16_u16(q9u16);
     q10s16 = vreinterpretq_s16_u16(q10u16);
-    q11s16 = vreinterpretq_s16_u16(q11u16);
-    q12s16 = vreinterpretq_s16_u16(q12u16);
 
-    q8s16 = vqaddq_s16(q8s16, q3s16);
-    q9s16 = vqaddq_s16(q9s16, q4s16);
-    q10s16 = vqaddq_s16(q10s16, q5s16);
-    q11s16 = vqaddq_s16(q11s16, q6s16);
-    q12s16 = vqaddq_s16(q12s16, q7s16);
+    q7s16 = vqaddq_s16(q7s16, q3s16);
+    q8s16 = vqaddq_s16(q8s16, q4s16);
+    q9s16 = vqaddq_s16(q9s16, q5s16);
+    q10s16 = vqaddq_s16(q10s16, q6s16);
 
-    d26u8 = vqrshrun_n_s16(q8s16, 7);
-    d27u8 = vqrshrun_n_s16(q9s16, 7);
-    d28u8 = vqrshrun_n_s16(q10s16, 7);
-    d29u8 = vqrshrun_n_s16(q11s16, 7);
-    d30u8 = vqrshrun_n_s16(q12s16, 7);
+    d6u8 = vqrshrun_n_s16(q7s16, 7);
+    d7u8 = vqrshrun_n_s16(q8s16, 7);
+    d8u8 = vqrshrun_n_s16(q9s16, 7);
+    d9u8 = vqrshrun_n_s16(q10s16, 7);
 
-    // Second pass: 8x8
+    d18u8 = d22u8;
+    d19u8 = d23u8;
+    d20u8 = d24u8;
+    d21u8 = d25u8;
+    d22u8 = d26u8;
+    d23u8 = d27u8;
+    d24u8 = d28u8;
+    d25u8 = d29u8;
+    d26u8 = d30u8;
+
+    vst1_u8(dst_ptr, d6u8);
+    dst_ptr += dst_pitch;
+    vst1_u8(dst_ptr, d7u8);
+    dst_ptr += dst_pitch;
+    vst1_u8(dst_ptr, d8u8);
+    dst_ptr += dst_pitch;
+    vst1_u8(dst_ptr, d9u8);
+    dst_ptr += dst_pitch;
+  }
+  return;
+}
+
+void vp8_sixtap_predict16x16_neon(unsigned char *src_ptr,
+                                  int src_pixels_per_line, int xoffset,
+                                  int yoffset, unsigned char *dst_ptr,
+                                  int dst_pitch) {
+  unsigned char *src, *src_tmp, *dst, *tmpp;
+  unsigned char tmp[336];
+  int i, j;
+  uint8x8_t d0u8, d1u8, d2u8, d3u8, d4u8, d5u8, d6u8, d7u8, d8u8, d9u8;
+  uint8x8_t d10u8, d11u8, d12u8, d13u8, d14u8, d15u8, d18u8, d19u8;
+  uint8x8_t d20u8, d21u8, d22u8, d23u8, d24u8, d25u8, d26u8, d27u8;
+  uint8x8_t d28u8, d29u8, d30u8, d31u8;
+  int8x8_t dtmps8, d0s8, d1s8, d2s8, d3s8, d4s8, d5s8;
+  uint8x16_t q3u8, q4u8;
+  uint16x8_t q3u16, q4u16, q5u16, q6u16, q7u16, q8u16, q9u16, q10u16;
+  uint16x8_t q11u16, q12u16, q13u16, q15u16;
+  int16x8_t q3s16, q4s16, q5s16, q6s16, q7s16, q8s16, q9s16, q10s16;
+  int16x8_t q11s16, q12s16, q13s16, q15s16;
+
+  if (xoffset == 0) {  // secondpass_filter8x8_only
+    // load second_pass filter
     dtmps8 = vld1_s8(vp8_sub_pel_filters[yoffset]);
     d0s8 = vdup_lane_s8(dtmps8, 0);
     d1s8 = vdup_lane_s8(dtmps8, 1);
@@ -807,25 +922,31 @@
     d4u8 = vreinterpret_u8_s8(vabs_s8(d4s8));
     d5u8 = vreinterpret_u8_s8(vabs_s8(d5s8));
 
-    tmpp = tmp;
-    q9u8 = vld1q_u8(tmpp);
-    tmpp += 16;
-    q10u8 = vld1q_u8(tmpp);
-    tmpp += 16;
-    q11u8 = vld1q_u8(tmpp);
-    tmpp += 16;
-    q12u8 = vld1q_u8(tmpp);
+    // load src data
+    src_tmp = src_ptr - src_pixels_per_line * 2;
+    for (i = 0; i < 2; i++) {
+      src = src_tmp + i * 8;
+      dst = dst_ptr + i * 8;
+      d18u8 = vld1_u8(src);
+      src += src_pixels_per_line;
+      d19u8 = vld1_u8(src);
+      src += src_pixels_per_line;
+      d20u8 = vld1_u8(src);
+      src += src_pixels_per_line;
+      d21u8 = vld1_u8(src);
+      src += src_pixels_per_line;
+      d22u8 = vld1_u8(src);
+      src += src_pixels_per_line;
+      for (j = 0; j < 4; j++) {
+        d23u8 = vld1_u8(src);
+        src += src_pixels_per_line;
+        d24u8 = vld1_u8(src);
+        src += src_pixels_per_line;
+        d25u8 = vld1_u8(src);
+        src += src_pixels_per_line;
+        d26u8 = vld1_u8(src);
+        src += src_pixels_per_line;
 
-    d18u8 = vget_low_u8(q9u8);
-    d19u8 = vget_high_u8(q9u8);
-    d20u8 = vget_low_u8(q10u8);
-    d21u8 = vget_high_u8(q10u8);
-    d22u8 = vget_low_u8(q11u8);
-    d23u8 = vget_high_u8(q11u8);
-    d24u8 = vget_low_u8(q12u8);
-    d25u8 = vget_high_u8(q12u8);
-
-    for (i = 2; i > 0; i--) {
         q3u16 = vmull_u8(d18u8, d0u8);
         q4u16 = vmull_u8(d19u8, d0u8);
         q5u16 = vmull_u8(d20u8, d0u8);
@@ -880,498 +1001,365 @@
         d20u8 = d24u8;
         d21u8 = d25u8;
         d22u8 = d26u8;
-        d23u8 = d27u8;
-        d24u8 = d28u8;
-        d25u8 = d29u8;
-        d26u8 = d30u8;
 
-        vst1_u8(dst_ptr, d6u8);
-        dst_ptr += dst_pitch;
-        vst1_u8(dst_ptr, d7u8);
-        dst_ptr += dst_pitch;
-        vst1_u8(dst_ptr, d8u8);
-        dst_ptr += dst_pitch;
-        vst1_u8(dst_ptr, d9u8);
-        dst_ptr += dst_pitch;
+        vst1_u8(dst, d6u8);
+        dst += dst_pitch;
+        vst1_u8(dst, d7u8);
+        dst += dst_pitch;
+        vst1_u8(dst, d8u8);
+        dst += dst_pitch;
+        vst1_u8(dst, d9u8);
+        dst += dst_pitch;
+      }
     }
     return;
-}
+  }
 
-void vp8_sixtap_predict16x16_neon(
-        unsigned char *src_ptr,
-        int src_pixels_per_line,
-        int xoffset,
-        int yoffset,
-        unsigned char *dst_ptr,
-        int dst_pitch) {
-    unsigned char *src, *src_tmp, *dst, *tmpp;
-    unsigned char tmp[336];
-    int i, j;
-    uint8x8_t d0u8, d1u8, d2u8, d3u8, d4u8, d5u8, d6u8, d7u8, d8u8, d9u8;
-    uint8x8_t d10u8, d11u8, d12u8, d13u8, d14u8, d15u8, d18u8, d19u8;
-    uint8x8_t d20u8, d21u8, d22u8, d23u8, d24u8, d25u8, d26u8, d27u8;
-    uint8x8_t d28u8, d29u8, d30u8, d31u8;
-    int8x8_t dtmps8, d0s8, d1s8, d2s8, d3s8, d4s8, d5s8;
-    uint8x16_t q3u8, q4u8;
-    uint16x8_t q3u16, q4u16, q5u16, q6u16, q7u16, q8u16, q9u16, q10u16;
-    uint16x8_t q11u16, q12u16, q13u16, q15u16;
-    int16x8_t q3s16, q4s16, q5s16, q6s16, q7s16, q8s16, q9s16, q10s16;
-    int16x8_t q11s16, q12s16, q13s16, q15s16;
+  // load first_pass filter
+  dtmps8 = vld1_s8(vp8_sub_pel_filters[xoffset]);
+  d0s8 = vdup_lane_s8(dtmps8, 0);
+  d1s8 = vdup_lane_s8(dtmps8, 1);
+  d2s8 = vdup_lane_s8(dtmps8, 2);
+  d3s8 = vdup_lane_s8(dtmps8, 3);
+  d4s8 = vdup_lane_s8(dtmps8, 4);
+  d5s8 = vdup_lane_s8(dtmps8, 5);
+  d0u8 = vreinterpret_u8_s8(vabs_s8(d0s8));
+  d1u8 = vreinterpret_u8_s8(vabs_s8(d1s8));
+  d2u8 = vreinterpret_u8_s8(vabs_s8(d2s8));
+  d3u8 = vreinterpret_u8_s8(vabs_s8(d3s8));
+  d4u8 = vreinterpret_u8_s8(vabs_s8(d4s8));
+  d5u8 = vreinterpret_u8_s8(vabs_s8(d5s8));
 
-    if (xoffset == 0) {  // secondpass_filter8x8_only
-        // load second_pass filter
-        dtmps8 = vld1_s8(vp8_sub_pel_filters[yoffset]);
-        d0s8 = vdup_lane_s8(dtmps8, 0);
-        d1s8 = vdup_lane_s8(dtmps8, 1);
-        d2s8 = vdup_lane_s8(dtmps8, 2);
-        d3s8 = vdup_lane_s8(dtmps8, 3);
-        d4s8 = vdup_lane_s8(dtmps8, 4);
-        d5s8 = vdup_lane_s8(dtmps8, 5);
-        d0u8 = vreinterpret_u8_s8(vabs_s8(d0s8));
-        d1u8 = vreinterpret_u8_s8(vabs_s8(d1s8));
-        d2u8 = vreinterpret_u8_s8(vabs_s8(d2s8));
-        d3u8 = vreinterpret_u8_s8(vabs_s8(d3s8));
-        d4u8 = vreinterpret_u8_s8(vabs_s8(d4s8));
-        d5u8 = vreinterpret_u8_s8(vabs_s8(d5s8));
+  // First pass: output_height lines x output_width columns (9x4)
+  if (yoffset == 0) {  // firstpass_filter4x4_only
+    src = src_ptr - 2;
+    dst = dst_ptr;
+    for (i = 0; i < 8; i++) {
+      d6u8 = vld1_u8(src);
+      d7u8 = vld1_u8(src + 8);
+      d8u8 = vld1_u8(src + 16);
+      src += src_pixels_per_line;
+      d9u8 = vld1_u8(src);
+      d10u8 = vld1_u8(src + 8);
+      d11u8 = vld1_u8(src + 16);
+      src += src_pixels_per_line;
 
-        // load src data
-        src_tmp = src_ptr - src_pixels_per_line * 2;
-        for (i = 0; i < 2; i++) {
-            src = src_tmp + i * 8;
-            dst = dst_ptr + i * 8;
-            d18u8 = vld1_u8(src);
-            src += src_pixels_per_line;
-            d19u8 = vld1_u8(src);
-            src += src_pixels_per_line;
-            d20u8 = vld1_u8(src);
-            src += src_pixels_per_line;
-            d21u8 = vld1_u8(src);
-            src += src_pixels_per_line;
-            d22u8 = vld1_u8(src);
-            src += src_pixels_per_line;
-            for (j = 0; j < 4; j++) {
-                d23u8 = vld1_u8(src);
-                src += src_pixels_per_line;
-                d24u8 = vld1_u8(src);
-                src += src_pixels_per_line;
-                d25u8 = vld1_u8(src);
-                src += src_pixels_per_line;
-                d26u8 = vld1_u8(src);
-                src += src_pixels_per_line;
+      __builtin_prefetch(src);
+      __builtin_prefetch(src + src_pixels_per_line);
 
-                q3u16 = vmull_u8(d18u8, d0u8);
-                q4u16 = vmull_u8(d19u8, d0u8);
-                q5u16 = vmull_u8(d20u8, d0u8);
-                q6u16 = vmull_u8(d21u8, d0u8);
+      q6u16 = vmull_u8(d6u8, d0u8);
+      q7u16 = vmull_u8(d7u8, d0u8);
+      q8u16 = vmull_u8(d9u8, d0u8);
+      q9u16 = vmull_u8(d10u8, d0u8);
 
-                q3u16 = vmlsl_u8(q3u16, d19u8, d1u8);
-                q4u16 = vmlsl_u8(q4u16, d20u8, d1u8);
-                q5u16 = vmlsl_u8(q5u16, d21u8, d1u8);
-                q6u16 = vmlsl_u8(q6u16, d22u8, d1u8);
+      d20u8 = vext_u8(d6u8, d7u8, 1);
+      d21u8 = vext_u8(d9u8, d10u8, 1);
+      d22u8 = vext_u8(d7u8, d8u8, 1);
+      d23u8 = vext_u8(d10u8, d11u8, 1);
+      d24u8 = vext_u8(d6u8, d7u8, 4);
+      d25u8 = vext_u8(d9u8, d10u8, 4);
+      d26u8 = vext_u8(d7u8, d8u8, 4);
+      d27u8 = vext_u8(d10u8, d11u8, 4);
+      d28u8 = vext_u8(d6u8, d7u8, 5);
+      d29u8 = vext_u8(d9u8, d10u8, 5);
 
-                q3u16 = vmlsl_u8(q3u16, d22u8, d4u8);
-                q4u16 = vmlsl_u8(q4u16, d23u8, d4u8);
-                q5u16 = vmlsl_u8(q5u16, d24u8, d4u8);
-                q6u16 = vmlsl_u8(q6u16, d25u8, d4u8);
+      q6u16 = vmlsl_u8(q6u16, d20u8, d1u8);
+      q8u16 = vmlsl_u8(q8u16, d21u8, d1u8);
+      q7u16 = vmlsl_u8(q7u16, d22u8, d1u8);
+      q9u16 = vmlsl_u8(q9u16, d23u8, d1u8);
+      q6u16 = vmlsl_u8(q6u16, d24u8, d4u8);
+      q8u16 = vmlsl_u8(q8u16, d25u8, d4u8);
+      q7u16 = vmlsl_u8(q7u16, d26u8, d4u8);
+      q9u16 = vmlsl_u8(q9u16, d27u8, d4u8);
+      q6u16 = vmlal_u8(q6u16, d28u8, d5u8);
+      q8u16 = vmlal_u8(q8u16, d29u8, d5u8);
 
-                q3u16 = vmlal_u8(q3u16, d20u8, d2u8);
-                q4u16 = vmlal_u8(q4u16, d21u8, d2u8);
-                q5u16 = vmlal_u8(q5u16, d22u8, d2u8);
-                q6u16 = vmlal_u8(q6u16, d23u8, d2u8);
+      d20u8 = vext_u8(d7u8, d8u8, 5);
+      d21u8 = vext_u8(d10u8, d11u8, 5);
+      d22u8 = vext_u8(d6u8, d7u8, 2);
+      d23u8 = vext_u8(d9u8, d10u8, 2);
+      d24u8 = vext_u8(d7u8, d8u8, 2);
+      d25u8 = vext_u8(d10u8, d11u8, 2);
+      d26u8 = vext_u8(d6u8, d7u8, 3);
+      d27u8 = vext_u8(d9u8, d10u8, 3);
+      d28u8 = vext_u8(d7u8, d8u8, 3);
+      d29u8 = vext_u8(d10u8, d11u8, 3);
 
-                q3u16 = vmlal_u8(q3u16, d23u8, d5u8);
-                q4u16 = vmlal_u8(q4u16, d24u8, d5u8);
-                q5u16 = vmlal_u8(q5u16, d25u8, d5u8);
-                q6u16 = vmlal_u8(q6u16, d26u8, d5u8);
+      q7u16 = vmlal_u8(q7u16, d20u8, d5u8);
+      q9u16 = vmlal_u8(q9u16, d21u8, d5u8);
+      q6u16 = vmlal_u8(q6u16, d22u8, d2u8);
+      q8u16 = vmlal_u8(q8u16, d23u8, d2u8);
+      q7u16 = vmlal_u8(q7u16, d24u8, d2u8);
+      q9u16 = vmlal_u8(q9u16, d25u8, d2u8);
 
-                q7u16 = vmull_u8(d21u8, d3u8);
-                q8u16 = vmull_u8(d22u8, d3u8);
-                q9u16 = vmull_u8(d23u8, d3u8);
-                q10u16 = vmull_u8(d24u8, d3u8);
+      q10u16 = vmull_u8(d26u8, d3u8);
+      q11u16 = vmull_u8(d27u8, d3u8);
+      q12u16 = vmull_u8(d28u8, d3u8);
+      q15u16 = vmull_u8(d29u8, d3u8);
 
-                q3s16 = vreinterpretq_s16_u16(q3u16);
-                q4s16 = vreinterpretq_s16_u16(q4u16);
-                q5s16 = vreinterpretq_s16_u16(q5u16);
-                q6s16 = vreinterpretq_s16_u16(q6u16);
-                q7s16 = vreinterpretq_s16_u16(q7u16);
-                q8s16 = vreinterpretq_s16_u16(q8u16);
-                q9s16 = vreinterpretq_s16_u16(q9u16);
-                q10s16 = vreinterpretq_s16_u16(q10u16);
+      q6s16 = vreinterpretq_s16_u16(q6u16);
+      q7s16 = vreinterpretq_s16_u16(q7u16);
+      q8s16 = vreinterpretq_s16_u16(q8u16);
+      q9s16 = vreinterpretq_s16_u16(q9u16);
+      q10s16 = vreinterpretq_s16_u16(q10u16);
+      q11s16 = vreinterpretq_s16_u16(q11u16);
+      q12s16 = vreinterpretq_s16_u16(q12u16);
+      q15s16 = vreinterpretq_s16_u16(q15u16);
 
-                q7s16 = vqaddq_s16(q7s16, q3s16);
-                q8s16 = vqaddq_s16(q8s16, q4s16);
-                q9s16 = vqaddq_s16(q9s16, q5s16);
-                q10s16 = vqaddq_s16(q10s16, q6s16);
+      q6s16 = vqaddq_s16(q6s16, q10s16);
+      q8s16 = vqaddq_s16(q8s16, q11s16);
+      q7s16 = vqaddq_s16(q7s16, q12s16);
+      q9s16 = vqaddq_s16(q9s16, q15s16);
 
-                d6u8 = vqrshrun_n_s16(q7s16, 7);
-                d7u8 = vqrshrun_n_s16(q8s16, 7);
-                d8u8 = vqrshrun_n_s16(q9s16, 7);
-                d9u8 = vqrshrun_n_s16(q10s16, 7);
+      d6u8 = vqrshrun_n_s16(q6s16, 7);
+      d7u8 = vqrshrun_n_s16(q7s16, 7);
+      d8u8 = vqrshrun_n_s16(q8s16, 7);
+      d9u8 = vqrshrun_n_s16(q9s16, 7);
 
-                d18u8 = d22u8;
-                d19u8 = d23u8;
-                d20u8 = d24u8;
-                d21u8 = d25u8;
-                d22u8 = d26u8;
-
-                vst1_u8(dst, d6u8);
-                dst += dst_pitch;
-                vst1_u8(dst, d7u8);
-                dst += dst_pitch;
-                vst1_u8(dst, d8u8);
-                dst += dst_pitch;
-                vst1_u8(dst, d9u8);
-                dst += dst_pitch;
-            }
-        }
-        return;
+      q3u8 = vcombine_u8(d6u8, d7u8);
+      q4u8 = vcombine_u8(d8u8, d9u8);
+      vst1q_u8(dst, q3u8);
+      dst += dst_pitch;
+      vst1q_u8(dst, q4u8);
+      dst += dst_pitch;
     }
+    return;
+  }
 
-    // load first_pass filter
-    dtmps8 = vld1_s8(vp8_sub_pel_filters[xoffset]);
-    d0s8 = vdup_lane_s8(dtmps8, 0);
-    d1s8 = vdup_lane_s8(dtmps8, 1);
-    d2s8 = vdup_lane_s8(dtmps8, 2);
-    d3s8 = vdup_lane_s8(dtmps8, 3);
-    d4s8 = vdup_lane_s8(dtmps8, 4);
-    d5s8 = vdup_lane_s8(dtmps8, 5);
-    d0u8 = vreinterpret_u8_s8(vabs_s8(d0s8));
-    d1u8 = vreinterpret_u8_s8(vabs_s8(d1s8));
-    d2u8 = vreinterpret_u8_s8(vabs_s8(d2s8));
-    d3u8 = vreinterpret_u8_s8(vabs_s8(d3s8));
-    d4u8 = vreinterpret_u8_s8(vabs_s8(d4s8));
-    d5u8 = vreinterpret_u8_s8(vabs_s8(d5s8));
+  src = src_ptr - 2 - src_pixels_per_line * 2;
+  tmpp = tmp;
+  for (i = 0; i < 7; i++) {
+    d6u8 = vld1_u8(src);
+    d7u8 = vld1_u8(src + 8);
+    d8u8 = vld1_u8(src + 16);
+    src += src_pixels_per_line;
+    d9u8 = vld1_u8(src);
+    d10u8 = vld1_u8(src + 8);
+    d11u8 = vld1_u8(src + 16);
+    src += src_pixels_per_line;
+    d12u8 = vld1_u8(src);
+    d13u8 = vld1_u8(src + 8);
+    d14u8 = vld1_u8(src + 16);
+    src += src_pixels_per_line;
 
-    // First pass: output_height lines x output_width columns (9x4)
-    if (yoffset == 0) {  // firstpass_filter4x4_only
-        src = src_ptr - 2;
-        dst = dst_ptr;
-        for (i = 0; i < 8; i++) {
-            d6u8 = vld1_u8(src);
-            d7u8 = vld1_u8(src + 8);
-            d8u8 = vld1_u8(src + 16);
-            src += src_pixels_per_line;
-            d9u8 = vld1_u8(src);
-            d10u8 = vld1_u8(src + 8);
-            d11u8 = vld1_u8(src + 16);
-            src += src_pixels_per_line;
+    __builtin_prefetch(src);
+    __builtin_prefetch(src + src_pixels_per_line);
+    __builtin_prefetch(src + src_pixels_per_line * 2);
 
-            __builtin_prefetch(src);
-            __builtin_prefetch(src + src_pixels_per_line);
+    q8u16 = vmull_u8(d6u8, d0u8);
+    q9u16 = vmull_u8(d7u8, d0u8);
+    q10u16 = vmull_u8(d9u8, d0u8);
+    q11u16 = vmull_u8(d10u8, d0u8);
+    q12u16 = vmull_u8(d12u8, d0u8);
+    q13u16 = vmull_u8(d13u8, d0u8);
 
-            q6u16 = vmull_u8(d6u8, d0u8);
-            q7u16 = vmull_u8(d7u8, d0u8);
-            q8u16 = vmull_u8(d9u8, d0u8);
-            q9u16 = vmull_u8(d10u8, d0u8);
+    d28u8 = vext_u8(d6u8, d7u8, 1);
+    d29u8 = vext_u8(d9u8, d10u8, 1);
+    d30u8 = vext_u8(d12u8, d13u8, 1);
+    q8u16 = vmlsl_u8(q8u16, d28u8, d1u8);
+    q10u16 = vmlsl_u8(q10u16, d29u8, d1u8);
+    q12u16 = vmlsl_u8(q12u16, d30u8, d1u8);
+    d28u8 = vext_u8(d7u8, d8u8, 1);
+    d29u8 = vext_u8(d10u8, d11u8, 1);
+    d30u8 = vext_u8(d13u8, d14u8, 1);
+    q9u16 = vmlsl_u8(q9u16, d28u8, d1u8);
+    q11u16 = vmlsl_u8(q11u16, d29u8, d1u8);
+    q13u16 = vmlsl_u8(q13u16, d30u8, d1u8);
 
-            d20u8 = vext_u8(d6u8, d7u8, 1);
-            d21u8 = vext_u8(d9u8, d10u8, 1);
-            d22u8 = vext_u8(d7u8, d8u8, 1);
-            d23u8 = vext_u8(d10u8, d11u8, 1);
-            d24u8 = vext_u8(d6u8, d7u8, 4);
-            d25u8 = vext_u8(d9u8, d10u8, 4);
-            d26u8 = vext_u8(d7u8, d8u8, 4);
-            d27u8 = vext_u8(d10u8, d11u8, 4);
-            d28u8 = vext_u8(d6u8, d7u8, 5);
-            d29u8 = vext_u8(d9u8, d10u8, 5);
+    d28u8 = vext_u8(d6u8, d7u8, 4);
+    d29u8 = vext_u8(d9u8, d10u8, 4);
+    d30u8 = vext_u8(d12u8, d13u8, 4);
+    q8u16 = vmlsl_u8(q8u16, d28u8, d4u8);
+    q10u16 = vmlsl_u8(q10u16, d29u8, d4u8);
+    q12u16 = vmlsl_u8(q12u16, d30u8, d4u8);
+    d28u8 = vext_u8(d7u8, d8u8, 4);
+    d29u8 = vext_u8(d10u8, d11u8, 4);
+    d30u8 = vext_u8(d13u8, d14u8, 4);
+    q9u16 = vmlsl_u8(q9u16, d28u8, d4u8);
+    q11u16 = vmlsl_u8(q11u16, d29u8, d4u8);
+    q13u16 = vmlsl_u8(q13u16, d30u8, d4u8);
 
-            q6u16 = vmlsl_u8(q6u16, d20u8, d1u8);
-            q8u16 = vmlsl_u8(q8u16, d21u8, d1u8);
-            q7u16 = vmlsl_u8(q7u16, d22u8, d1u8);
-            q9u16 = vmlsl_u8(q9u16, d23u8, d1u8);
-            q6u16 = vmlsl_u8(q6u16, d24u8, d4u8);
-            q8u16 = vmlsl_u8(q8u16, d25u8, d4u8);
-            q7u16 = vmlsl_u8(q7u16, d26u8, d4u8);
-            q9u16 = vmlsl_u8(q9u16, d27u8, d4u8);
-            q6u16 = vmlal_u8(q6u16, d28u8, d5u8);
-            q8u16 = vmlal_u8(q8u16, d29u8, d5u8);
+    d28u8 = vext_u8(d6u8, d7u8, 5);
+    d29u8 = vext_u8(d9u8, d10u8, 5);
+    d30u8 = vext_u8(d12u8, d13u8, 5);
+    q8u16 = vmlal_u8(q8u16, d28u8, d5u8);
+    q10u16 = vmlal_u8(q10u16, d29u8, d5u8);
+    q12u16 = vmlal_u8(q12u16, d30u8, d5u8);
+    d28u8 = vext_u8(d7u8, d8u8, 5);
+    d29u8 = vext_u8(d10u8, d11u8, 5);
+    d30u8 = vext_u8(d13u8, d14u8, 5);
+    q9u16 = vmlal_u8(q9u16, d28u8, d5u8);
+    q11u16 = vmlal_u8(q11u16, d29u8, d5u8);
+    q13u16 = vmlal_u8(q13u16, d30u8, d5u8);
 
-            d20u8 = vext_u8(d7u8, d8u8, 5);
-            d21u8 = vext_u8(d10u8, d11u8, 5);
-            d22u8 = vext_u8(d6u8, d7u8, 2);
-            d23u8 = vext_u8(d9u8, d10u8, 2);
-            d24u8 = vext_u8(d7u8, d8u8, 2);
-            d25u8 = vext_u8(d10u8, d11u8, 2);
-            d26u8 = vext_u8(d6u8, d7u8, 3);
-            d27u8 = vext_u8(d9u8, d10u8, 3);
-            d28u8 = vext_u8(d7u8, d8u8, 3);
-            d29u8 = vext_u8(d10u8, d11u8, 3);
+    d28u8 = vext_u8(d6u8, d7u8, 2);
+    d29u8 = vext_u8(d9u8, d10u8, 2);
+    d30u8 = vext_u8(d12u8, d13u8, 2);
+    q8u16 = vmlal_u8(q8u16, d28u8, d2u8);
+    q10u16 = vmlal_u8(q10u16, d29u8, d2u8);
+    q12u16 = vmlal_u8(q12u16, d30u8, d2u8);
+    d28u8 = vext_u8(d7u8, d8u8, 2);
+    d29u8 = vext_u8(d10u8, d11u8, 2);
+    d30u8 = vext_u8(d13u8, d14u8, 2);
+    q9u16 = vmlal_u8(q9u16, d28u8, d2u8);
+    q11u16 = vmlal_u8(q11u16, d29u8, d2u8);
+    q13u16 = vmlal_u8(q13u16, d30u8, d2u8);
 
-            q7u16 = vmlal_u8(q7u16, d20u8, d5u8);
-            q9u16 = vmlal_u8(q9u16, d21u8, d5u8);
-            q6u16 = vmlal_u8(q6u16, d22u8, d2u8);
-            q8u16 = vmlal_u8(q8u16, d23u8, d2u8);
-            q7u16 = vmlal_u8(q7u16, d24u8, d2u8);
-            q9u16 = vmlal_u8(q9u16, d25u8, d2u8);
+    d28u8 = vext_u8(d6u8, d7u8, 3);
+    d29u8 = vext_u8(d9u8, d10u8, 3);
+    d30u8 = vext_u8(d12u8, d13u8, 3);
+    d15u8 = vext_u8(d7u8, d8u8, 3);
+    d31u8 = vext_u8(d10u8, d11u8, 3);
+    d6u8 = vext_u8(d13u8, d14u8, 3);
+    q4u16 = vmull_u8(d28u8, d3u8);
+    q5u16 = vmull_u8(d29u8, d3u8);
+    q6u16 = vmull_u8(d30u8, d3u8);
+    q4s16 = vreinterpretq_s16_u16(q4u16);
+    q5s16 = vreinterpretq_s16_u16(q5u16);
+    q6s16 = vreinterpretq_s16_u16(q6u16);
+    q8s16 = vreinterpretq_s16_u16(q8u16);
+    q10s16 = vreinterpretq_s16_u16(q10u16);
+    q12s16 = vreinterpretq_s16_u16(q12u16);
+    q8s16 = vqaddq_s16(q8s16, q4s16);
+    q10s16 = vqaddq_s16(q10s16, q5s16);
+    q12s16 = vqaddq_s16(q12s16, q6s16);
 
-            q10u16 = vmull_u8(d26u8, d3u8);
-            q11u16 = vmull_u8(d27u8, d3u8);
-            q12u16 = vmull_u8(d28u8, d3u8);
-            q15u16 = vmull_u8(d29u8, d3u8);
+    q6u16 = vmull_u8(d15u8, d3u8);
+    q7u16 = vmull_u8(d31u8, d3u8);
+    q3u16 = vmull_u8(d6u8, d3u8);
+    q3s16 = vreinterpretq_s16_u16(q3u16);
+    q6s16 = vreinterpretq_s16_u16(q6u16);
+    q7s16 = vreinterpretq_s16_u16(q7u16);
+    q9s16 = vreinterpretq_s16_u16(q9u16);
+    q11s16 = vreinterpretq_s16_u16(q11u16);
+    q13s16 = vreinterpretq_s16_u16(q13u16);
+    q9s16 = vqaddq_s16(q9s16, q6s16);
+    q11s16 = vqaddq_s16(q11s16, q7s16);
+    q13s16 = vqaddq_s16(q13s16, q3s16);
 
-            q6s16 = vreinterpretq_s16_u16(q6u16);
-            q7s16 = vreinterpretq_s16_u16(q7u16);
-            q8s16 = vreinterpretq_s16_u16(q8u16);
-            q9s16 = vreinterpretq_s16_u16(q9u16);
-            q10s16 = vreinterpretq_s16_u16(q10u16);
-            q11s16 = vreinterpretq_s16_u16(q11u16);
-            q12s16 = vreinterpretq_s16_u16(q12u16);
-            q15s16 = vreinterpretq_s16_u16(q15u16);
+    d6u8 = vqrshrun_n_s16(q8s16, 7);
+    d7u8 = vqrshrun_n_s16(q9s16, 7);
+    d8u8 = vqrshrun_n_s16(q10s16, 7);
+    d9u8 = vqrshrun_n_s16(q11s16, 7);
+    d10u8 = vqrshrun_n_s16(q12s16, 7);
+    d11u8 = vqrshrun_n_s16(q13s16, 7);
 
-            q6s16 = vqaddq_s16(q6s16, q10s16);
-            q8s16 = vqaddq_s16(q8s16, q11s16);
-            q7s16 = vqaddq_s16(q7s16, q12s16);
-            q9s16 = vqaddq_s16(q9s16, q15s16);
+    vst1_u8(tmpp, d6u8);
+    tmpp += 8;
+    vst1_u8(tmpp, d7u8);
+    tmpp += 8;
+    vst1_u8(tmpp, d8u8);
+    tmpp += 8;
+    vst1_u8(tmpp, d9u8);
+    tmpp += 8;
+    vst1_u8(tmpp, d10u8);
+    tmpp += 8;
+    vst1_u8(tmpp, d11u8);
+    tmpp += 8;
+  }
 
-            d6u8 = vqrshrun_n_s16(q6s16, 7);
-            d7u8 = vqrshrun_n_s16(q7s16, 7);
-            d8u8 = vqrshrun_n_s16(q8s16, 7);
-            d9u8 = vqrshrun_n_s16(q9s16, 7);
+  // Second pass: 16x16
+  dtmps8 = vld1_s8(vp8_sub_pel_filters[yoffset]);
+  d0s8 = vdup_lane_s8(dtmps8, 0);
+  d1s8 = vdup_lane_s8(dtmps8, 1);
+  d2s8 = vdup_lane_s8(dtmps8, 2);
+  d3s8 = vdup_lane_s8(dtmps8, 3);
+  d4s8 = vdup_lane_s8(dtmps8, 4);
+  d5s8 = vdup_lane_s8(dtmps8, 5);
+  d0u8 = vreinterpret_u8_s8(vabs_s8(d0s8));
+  d1u8 = vreinterpret_u8_s8(vabs_s8(d1s8));
+  d2u8 = vreinterpret_u8_s8(vabs_s8(d2s8));
+  d3u8 = vreinterpret_u8_s8(vabs_s8(d3s8));
+  d4u8 = vreinterpret_u8_s8(vabs_s8(d4s8));
+  d5u8 = vreinterpret_u8_s8(vabs_s8(d5s8));
 
-            q3u8 = vcombine_u8(d6u8, d7u8);
-            q4u8 = vcombine_u8(d8u8, d9u8);
-            vst1q_u8(dst, q3u8);
-            dst += dst_pitch;
-            vst1q_u8(dst, q4u8);
-            dst += dst_pitch;
-        }
-        return;
-    }
+  for (i = 0; i < 2; i++) {
+    dst = dst_ptr + 8 * i;
+    tmpp = tmp + 8 * i;
+    d18u8 = vld1_u8(tmpp);
+    tmpp += 16;
+    d19u8 = vld1_u8(tmpp);
+    tmpp += 16;
+    d20u8 = vld1_u8(tmpp);
+    tmpp += 16;
+    d21u8 = vld1_u8(tmpp);
+    tmpp += 16;
+    d22u8 = vld1_u8(tmpp);
+    tmpp += 16;
+    for (j = 0; j < 4; j++) {
+      d23u8 = vld1_u8(tmpp);
+      tmpp += 16;
+      d24u8 = vld1_u8(tmpp);
+      tmpp += 16;
+      d25u8 = vld1_u8(tmpp);
+      tmpp += 16;
+      d26u8 = vld1_u8(tmpp);
+      tmpp += 16;
 
-    src = src_ptr - 2 - src_pixels_per_line * 2;
-    tmpp = tmp;
-    for (i = 0; i < 7; i++) {
-        d6u8 = vld1_u8(src);
-        d7u8 = vld1_u8(src + 8);
-        d8u8 = vld1_u8(src + 16);
-        src += src_pixels_per_line;
-        d9u8 = vld1_u8(src);
-        d10u8 = vld1_u8(src + 8);
-        d11u8 = vld1_u8(src + 16);
-        src += src_pixels_per_line;
-        d12u8 = vld1_u8(src);
-        d13u8 = vld1_u8(src + 8);
-        d14u8 = vld1_u8(src + 16);
-        src += src_pixels_per_line;
+      q3u16 = vmull_u8(d18u8, d0u8);
+      q4u16 = vmull_u8(d19u8, d0u8);
+      q5u16 = vmull_u8(d20u8, d0u8);
+      q6u16 = vmull_u8(d21u8, d0u8);
 
-        __builtin_prefetch(src);
-        __builtin_prefetch(src + src_pixels_per_line);
-        __builtin_prefetch(src + src_pixels_per_line * 2);
+      q3u16 = vmlsl_u8(q3u16, d19u8, d1u8);
+      q4u16 = vmlsl_u8(q4u16, d20u8, d1u8);
+      q5u16 = vmlsl_u8(q5u16, d21u8, d1u8);
+      q6u16 = vmlsl_u8(q6u16, d22u8, d1u8);
 
-        q8u16 = vmull_u8(d6u8, d0u8);
-        q9u16 = vmull_u8(d7u8, d0u8);
-        q10u16 = vmull_u8(d9u8, d0u8);
-        q11u16 = vmull_u8(d10u8, d0u8);
-        q12u16 = vmull_u8(d12u8, d0u8);
-        q13u16 = vmull_u8(d13u8, d0u8);
+      q3u16 = vmlsl_u8(q3u16, d22u8, d4u8);
+      q4u16 = vmlsl_u8(q4u16, d23u8, d4u8);
+      q5u16 = vmlsl_u8(q5u16, d24u8, d4u8);
+      q6u16 = vmlsl_u8(q6u16, d25u8, d4u8);
 
-        d28u8 = vext_u8(d6u8, d7u8, 1);
-        d29u8 = vext_u8(d9u8, d10u8, 1);
-        d30u8 = vext_u8(d12u8, d13u8, 1);
-        q8u16 = vmlsl_u8(q8u16, d28u8, d1u8);
-        q10u16 = vmlsl_u8(q10u16, d29u8, d1u8);
-        q12u16 = vmlsl_u8(q12u16, d30u8, d1u8);
-        d28u8 = vext_u8(d7u8, d8u8, 1);
-        d29u8 = vext_u8(d10u8, d11u8, 1);
-        d30u8 = vext_u8(d13u8, d14u8, 1);
-        q9u16  = vmlsl_u8(q9u16, d28u8, d1u8);
-        q11u16 = vmlsl_u8(q11u16, d29u8, d1u8);
-        q13u16 = vmlsl_u8(q13u16, d30u8, d1u8);
+      q3u16 = vmlal_u8(q3u16, d20u8, d2u8);
+      q4u16 = vmlal_u8(q4u16, d21u8, d2u8);
+      q5u16 = vmlal_u8(q5u16, d22u8, d2u8);
+      q6u16 = vmlal_u8(q6u16, d23u8, d2u8);
 
-        d28u8 = vext_u8(d6u8, d7u8, 4);
-        d29u8 = vext_u8(d9u8, d10u8, 4);
-        d30u8 = vext_u8(d12u8, d13u8, 4);
-        q8u16 = vmlsl_u8(q8u16, d28u8, d4u8);
-        q10u16 = vmlsl_u8(q10u16, d29u8, d4u8);
-        q12u16 = vmlsl_u8(q12u16, d30u8, d4u8);
-        d28u8 = vext_u8(d7u8, d8u8, 4);
-        d29u8 = vext_u8(d10u8, d11u8, 4);
-        d30u8 = vext_u8(d13u8, d14u8, 4);
-        q9u16 = vmlsl_u8(q9u16, d28u8, d4u8);
-        q11u16 = vmlsl_u8(q11u16, d29u8, d4u8);
-        q13u16 = vmlsl_u8(q13u16, d30u8, d4u8);
+      q3u16 = vmlal_u8(q3u16, d23u8, d5u8);
+      q4u16 = vmlal_u8(q4u16, d24u8, d5u8);
+      q5u16 = vmlal_u8(q5u16, d25u8, d5u8);
+      q6u16 = vmlal_u8(q6u16, d26u8, d5u8);
 
-        d28u8 = vext_u8(d6u8, d7u8, 5);
-        d29u8 = vext_u8(d9u8, d10u8, 5);
-        d30u8 = vext_u8(d12u8, d13u8, 5);
-        q8u16 = vmlal_u8(q8u16, d28u8, d5u8);
-        q10u16 = vmlal_u8(q10u16, d29u8, d5u8);
-        q12u16 = vmlal_u8(q12u16, d30u8, d5u8);
-        d28u8 = vext_u8(d7u8, d8u8, 5);
-        d29u8 = vext_u8(d10u8, d11u8, 5);
-        d30u8 = vext_u8(d13u8, d14u8, 5);
-        q9u16 = vmlal_u8(q9u16, d28u8, d5u8);
-        q11u16 = vmlal_u8(q11u16, d29u8, d5u8);
-        q13u16 = vmlal_u8(q13u16, d30u8, d5u8);
+      q7u16 = vmull_u8(d21u8, d3u8);
+      q8u16 = vmull_u8(d22u8, d3u8);
+      q9u16 = vmull_u8(d23u8, d3u8);
+      q10u16 = vmull_u8(d24u8, d3u8);
 
-        d28u8 = vext_u8(d6u8, d7u8, 2);
-        d29u8 = vext_u8(d9u8, d10u8, 2);
-        d30u8 = vext_u8(d12u8, d13u8, 2);
-        q8u16 = vmlal_u8(q8u16, d28u8, d2u8);
-        q10u16 = vmlal_u8(q10u16, d29u8, d2u8);
-        q12u16 = vmlal_u8(q12u16, d30u8, d2u8);
-        d28u8 = vext_u8(d7u8, d8u8, 2);
-        d29u8 = vext_u8(d10u8, d11u8, 2);
-        d30u8 = vext_u8(d13u8, d14u8, 2);
-        q9u16 = vmlal_u8(q9u16, d28u8, d2u8);
-        q11u16 = vmlal_u8(q11u16, d29u8, d2u8);
-        q13u16 = vmlal_u8(q13u16, d30u8, d2u8);
+      q3s16 = vreinterpretq_s16_u16(q3u16);
+      q4s16 = vreinterpretq_s16_u16(q4u16);
+      q5s16 = vreinterpretq_s16_u16(q5u16);
+      q6s16 = vreinterpretq_s16_u16(q6u16);
+      q7s16 = vreinterpretq_s16_u16(q7u16);
+      q8s16 = vreinterpretq_s16_u16(q8u16);
+      q9s16 = vreinterpretq_s16_u16(q9u16);
+      q10s16 = vreinterpretq_s16_u16(q10u16);
 
-        d28u8 = vext_u8(d6u8, d7u8, 3);
-        d29u8 = vext_u8(d9u8, d10u8, 3);
-        d30u8 = vext_u8(d12u8, d13u8, 3);
-        d15u8 = vext_u8(d7u8, d8u8, 3);
-        d31u8 = vext_u8(d10u8, d11u8, 3);
-        d6u8  = vext_u8(d13u8, d14u8, 3);
-        q4u16 = vmull_u8(d28u8, d3u8);
-        q5u16 = vmull_u8(d29u8, d3u8);
-        q6u16 = vmull_u8(d30u8, d3u8);
-        q4s16 = vreinterpretq_s16_u16(q4u16);
-        q5s16 = vreinterpretq_s16_u16(q5u16);
-        q6s16 = vreinterpretq_s16_u16(q6u16);
-        q8s16 = vreinterpretq_s16_u16(q8u16);
-        q10s16 = vreinterpretq_s16_u16(q10u16);
-        q12s16 = vreinterpretq_s16_u16(q12u16);
-        q8s16 = vqaddq_s16(q8s16, q4s16);
-        q10s16 = vqaddq_s16(q10s16, q5s16);
-        q12s16 = vqaddq_s16(q12s16, q6s16);
+      q7s16 = vqaddq_s16(q7s16, q3s16);
+      q8s16 = vqaddq_s16(q8s16, q4s16);
+      q9s16 = vqaddq_s16(q9s16, q5s16);
+      q10s16 = vqaddq_s16(q10s16, q6s16);
 
-        q6u16 = vmull_u8(d15u8, d3u8);
-        q7u16 = vmull_u8(d31u8, d3u8);
-        q3u16 = vmull_u8(d6u8, d3u8);
-        q3s16 = vreinterpretq_s16_u16(q3u16);
-        q6s16 = vreinterpretq_s16_u16(q6u16);
-        q7s16 = vreinterpretq_s16_u16(q7u16);
-        q9s16 = vreinterpretq_s16_u16(q9u16);
-        q11s16 = vreinterpretq_s16_u16(q11u16);
-        q13s16 = vreinterpretq_s16_u16(q13u16);
-        q9s16 = vqaddq_s16(q9s16, q6s16);
-        q11s16 = vqaddq_s16(q11s16, q7s16);
-        q13s16 = vqaddq_s16(q13s16, q3s16);
+      d6u8 = vqrshrun_n_s16(q7s16, 7);
+      d7u8 = vqrshrun_n_s16(q8s16, 7);
+      d8u8 = vqrshrun_n_s16(q9s16, 7);
+      d9u8 = vqrshrun_n_s16(q10s16, 7);
 
-        d6u8 = vqrshrun_n_s16(q8s16, 7);
-        d7u8 = vqrshrun_n_s16(q9s16, 7);
-        d8u8 = vqrshrun_n_s16(q10s16, 7);
-        d9u8 = vqrshrun_n_s16(q11s16, 7);
-        d10u8 = vqrshrun_n_s16(q12s16, 7);
-        d11u8 = vqrshrun_n_s16(q13s16, 7);
+      d18u8 = d22u8;
+      d19u8 = d23u8;
+      d20u8 = d24u8;
+      d21u8 = d25u8;
+      d22u8 = d26u8;
 
-        vst1_u8(tmpp, d6u8);
-        tmpp += 8;
-        vst1_u8(tmpp, d7u8);
-        tmpp += 8;
-        vst1_u8(tmpp, d8u8);
-        tmpp += 8;
-        vst1_u8(tmpp, d9u8);
-        tmpp += 8;
-        vst1_u8(tmpp, d10u8);
-        tmpp += 8;
-        vst1_u8(tmpp, d11u8);
-        tmpp += 8;
+      vst1_u8(dst, d6u8);
+      dst += dst_pitch;
+      vst1_u8(dst, d7u8);
+      dst += dst_pitch;
+      vst1_u8(dst, d8u8);
+      dst += dst_pitch;
+      vst1_u8(dst, d9u8);
+      dst += dst_pitch;
     }
-
-    // Second pass: 16x16
-    dtmps8 = vld1_s8(vp8_sub_pel_filters[yoffset]);
-    d0s8 = vdup_lane_s8(dtmps8, 0);
-    d1s8 = vdup_lane_s8(dtmps8, 1);
-    d2s8 = vdup_lane_s8(dtmps8, 2);
-    d3s8 = vdup_lane_s8(dtmps8, 3);
-    d4s8 = vdup_lane_s8(dtmps8, 4);
-    d5s8 = vdup_lane_s8(dtmps8, 5);
-    d0u8 = vreinterpret_u8_s8(vabs_s8(d0s8));
-    d1u8 = vreinterpret_u8_s8(vabs_s8(d1s8));
-    d2u8 = vreinterpret_u8_s8(vabs_s8(d2s8));
-    d3u8 = vreinterpret_u8_s8(vabs_s8(d3s8));
-    d4u8 = vreinterpret_u8_s8(vabs_s8(d4s8));
-    d5u8 = vreinterpret_u8_s8(vabs_s8(d5s8));
-
-    for (i = 0; i < 2; i++) {
-        dst = dst_ptr + 8 * i;
-        tmpp = tmp + 8 * i;
-        d18u8 = vld1_u8(tmpp);
-        tmpp += 16;
-        d19u8 = vld1_u8(tmpp);
-        tmpp += 16;
-        d20u8 = vld1_u8(tmpp);
-        tmpp += 16;
-        d21u8 = vld1_u8(tmpp);
-        tmpp += 16;
-        d22u8 = vld1_u8(tmpp);
-        tmpp += 16;
-        for (j = 0; j < 4; j++) {
-            d23u8 = vld1_u8(tmpp);
-            tmpp += 16;
-            d24u8 = vld1_u8(tmpp);
-            tmpp += 16;
-            d25u8 = vld1_u8(tmpp);
-            tmpp += 16;
-            d26u8 = vld1_u8(tmpp);
-            tmpp += 16;
-
-            q3u16 = vmull_u8(d18u8, d0u8);
-            q4u16 = vmull_u8(d19u8, d0u8);
-            q5u16 = vmull_u8(d20u8, d0u8);
-            q6u16 = vmull_u8(d21u8, d0u8);
-
-            q3u16 = vmlsl_u8(q3u16, d19u8, d1u8);
-            q4u16 = vmlsl_u8(q4u16, d20u8, d1u8);
-            q5u16 = vmlsl_u8(q5u16, d21u8, d1u8);
-            q6u16 = vmlsl_u8(q6u16, d22u8, d1u8);
-
-            q3u16 = vmlsl_u8(q3u16, d22u8, d4u8);
-            q4u16 = vmlsl_u8(q4u16, d23u8, d4u8);
-            q5u16 = vmlsl_u8(q5u16, d24u8, d4u8);
-            q6u16 = vmlsl_u8(q6u16, d25u8, d4u8);
-
-            q3u16 = vmlal_u8(q3u16, d20u8, d2u8);
-            q4u16 = vmlal_u8(q4u16, d21u8, d2u8);
-            q5u16 = vmlal_u8(q5u16, d22u8, d2u8);
-            q6u16 = vmlal_u8(q6u16, d23u8, d2u8);
-
-            q3u16 = vmlal_u8(q3u16, d23u8, d5u8);
-            q4u16 = vmlal_u8(q4u16, d24u8, d5u8);
-            q5u16 = vmlal_u8(q5u16, d25u8, d5u8);
-            q6u16 = vmlal_u8(q6u16, d26u8, d5u8);
-
-            q7u16 = vmull_u8(d21u8, d3u8);
-            q8u16 = vmull_u8(d22u8, d3u8);
-            q9u16 = vmull_u8(d23u8, d3u8);
-            q10u16 = vmull_u8(d24u8, d3u8);
-
-            q3s16 = vreinterpretq_s16_u16(q3u16);
-            q4s16 = vreinterpretq_s16_u16(q4u16);
-            q5s16 = vreinterpretq_s16_u16(q5u16);
-            q6s16 = vreinterpretq_s16_u16(q6u16);
-            q7s16 = vreinterpretq_s16_u16(q7u16);
-            q8s16 = vreinterpretq_s16_u16(q8u16);
-            q9s16 = vreinterpretq_s16_u16(q9u16);
-            q10s16 = vreinterpretq_s16_u16(q10u16);
-
-            q7s16 = vqaddq_s16(q7s16, q3s16);
-            q8s16 = vqaddq_s16(q8s16, q4s16);
-            q9s16 = vqaddq_s16(q9s16, q5s16);
-            q10s16 = vqaddq_s16(q10s16, q6s16);
-
-            d6u8 = vqrshrun_n_s16(q7s16, 7);
-            d7u8 = vqrshrun_n_s16(q8s16, 7);
-            d8u8 = vqrshrun_n_s16(q9s16, 7);
-            d9u8 = vqrshrun_n_s16(q10s16, 7);
-
-            d18u8 = d22u8;
-            d19u8 = d23u8;
-            d20u8 = d24u8;
-            d21u8 = d25u8;
-            d22u8 = d26u8;
-
-            vst1_u8(dst, d6u8);
-            dst += dst_pitch;
-            vst1_u8(dst, d7u8);
-            dst += dst_pitch;
-            vst1_u8(dst, d8u8);
-            dst += dst_pitch;
-            vst1_u8(dst, d9u8);
-            dst += dst_pitch;
-        }
-    }
-    return;
+  }
+  return;
 }
--- a/vp8/common/arm/neon/vp8_loopfilter_neon.c
+++ b/vp8/common/arm/neon/vp8_loopfilter_neon.c
@@ -12,539 +12,525 @@
 #include "./vpx_config.h"
 #include "vpx_ports/arm.h"
 
-static INLINE void vp8_loop_filter_neon(
-        uint8x16_t qblimit,  // flimit
-        uint8x16_t qlimit,   // limit
-        uint8x16_t qthresh,  // thresh
-        uint8x16_t q3,       // p3
-        uint8x16_t q4,       // p2
-        uint8x16_t q5,       // p1
-        uint8x16_t q6,       // p0
-        uint8x16_t q7,       // q0
-        uint8x16_t q8,       // q1
-        uint8x16_t q9,       // q2
-        uint8x16_t q10,      // q3
-        uint8x16_t *q5r,     // p1
-        uint8x16_t *q6r,     // p0
-        uint8x16_t *q7r,     // q0
-        uint8x16_t *q8r) {   // q1
-    uint8x16_t q0u8, q1u8, q2u8, q11u8, q12u8, q13u8, q14u8, q15u8;
-    int16x8_t q2s16, q11s16;
-    uint16x8_t q4u16;
-    int8x16_t q1s8, q2s8, q10s8, q11s8, q12s8, q13s8;
-    int8x8_t d2s8, d3s8;
+static INLINE void vp8_loop_filter_neon(uint8x16_t qblimit,  // flimit
+                                        uint8x16_t qlimit,   // limit
+                                        uint8x16_t qthresh,  // thresh
+                                        uint8x16_t q3,       // p3
+                                        uint8x16_t q4,       // p2
+                                        uint8x16_t q5,       // p1
+                                        uint8x16_t q6,       // p0
+                                        uint8x16_t q7,       // q0
+                                        uint8x16_t q8,       // q1
+                                        uint8x16_t q9,       // q2
+                                        uint8x16_t q10,      // q3
+                                        uint8x16_t *q5r,     // p1
+                                        uint8x16_t *q6r,     // p0
+                                        uint8x16_t *q7r,     // q0
+                                        uint8x16_t *q8r) {   // q1
+  uint8x16_t q0u8, q1u8, q2u8, q11u8, q12u8, q13u8, q14u8, q15u8;
+  int16x8_t q2s16, q11s16;
+  uint16x8_t q4u16;
+  int8x16_t q1s8, q2s8, q10s8, q11s8, q12s8, q13s8;
+  int8x8_t d2s8, d3s8;
 
-    q11u8 = vabdq_u8(q3, q4);
-    q12u8 = vabdq_u8(q4, q5);
-    q13u8 = vabdq_u8(q5, q6);
-    q14u8 = vabdq_u8(q8, q7);
-    q3    = vabdq_u8(q9, q8);
-    q4    = vabdq_u8(q10, q9);
+  q11u8 = vabdq_u8(q3, q4);
+  q12u8 = vabdq_u8(q4, q5);
+  q13u8 = vabdq_u8(q5, q6);
+  q14u8 = vabdq_u8(q8, q7);
+  q3 = vabdq_u8(q9, q8);
+  q4 = vabdq_u8(q10, q9);
 
-    q11u8 = vmaxq_u8(q11u8, q12u8);
-    q12u8 = vmaxq_u8(q13u8, q14u8);
-    q3    = vmaxq_u8(q3, q4);
-    q15u8 = vmaxq_u8(q11u8, q12u8);
+  q11u8 = vmaxq_u8(q11u8, q12u8);
+  q12u8 = vmaxq_u8(q13u8, q14u8);
+  q3 = vmaxq_u8(q3, q4);
+  q15u8 = vmaxq_u8(q11u8, q12u8);
 
-    q9 = vabdq_u8(q6, q7);
+  q9 = vabdq_u8(q6, q7);
 
-    // vp8_hevmask
-    q13u8 = vcgtq_u8(q13u8, qthresh);
-    q14u8 = vcgtq_u8(q14u8, qthresh);
-    q15u8 = vmaxq_u8(q15u8, q3);
+  // vp8_hevmask
+  q13u8 = vcgtq_u8(q13u8, qthresh);
+  q14u8 = vcgtq_u8(q14u8, qthresh);
+  q15u8 = vmaxq_u8(q15u8, q3);
 
-    q2u8 = vabdq_u8(q5, q8);
-    q9 = vqaddq_u8(q9, q9);
+  q2u8 = vabdq_u8(q5, q8);
+  q9 = vqaddq_u8(q9, q9);
 
-    q15u8 = vcgeq_u8(qlimit, q15u8);
+  q15u8 = vcgeq_u8(qlimit, q15u8);
 
-    // vp8_filter() function
-    // convert to signed
-    q10 = vdupq_n_u8(0x80);
-    q8 = veorq_u8(q8, q10);
-    q7 = veorq_u8(q7, q10);
-    q6 = veorq_u8(q6, q10);
-    q5 = veorq_u8(q5, q10);
+  // vp8_filter() function
+  // convert to signed
+  q10 = vdupq_n_u8(0x80);
+  q8 = veorq_u8(q8, q10);
+  q7 = veorq_u8(q7, q10);
+  q6 = veorq_u8(q6, q10);
+  q5 = veorq_u8(q5, q10);
 
-    q2u8 = vshrq_n_u8(q2u8, 1);
-    q9 = vqaddq_u8(q9, q2u8);
+  q2u8 = vshrq_n_u8(q2u8, 1);
+  q9 = vqaddq_u8(q9, q2u8);
 
-    q10 = vdupq_n_u8(3);
+  q10 = vdupq_n_u8(3);
 
-    q2s16 = vsubl_s8(vget_low_s8(vreinterpretq_s8_u8(q7)),
-                     vget_low_s8(vreinterpretq_s8_u8(q6)));
-    q11s16 = vsubl_s8(vget_high_s8(vreinterpretq_s8_u8(q7)),
-                      vget_high_s8(vreinterpretq_s8_u8(q6)));
+  q2s16 = vsubl_s8(vget_low_s8(vreinterpretq_s8_u8(q7)),
+                   vget_low_s8(vreinterpretq_s8_u8(q6)));
+  q11s16 = vsubl_s8(vget_high_s8(vreinterpretq_s8_u8(q7)),
+                    vget_high_s8(vreinterpretq_s8_u8(q6)));
 
-    q9 = vcgeq_u8(qblimit, q9);
+  q9 = vcgeq_u8(qblimit, q9);
 
-    q1s8 = vqsubq_s8(vreinterpretq_s8_u8(q5),
-                    vreinterpretq_s8_u8(q8));
+  q1s8 = vqsubq_s8(vreinterpretq_s8_u8(q5), vreinterpretq_s8_u8(q8));
 
-    q14u8 = vorrq_u8(q13u8, q14u8);
+  q14u8 = vorrq_u8(q13u8, q14u8);
 
-    q4u16 = vmovl_u8(vget_low_u8(q10));
-    q2s16 = vmulq_s16(q2s16, vreinterpretq_s16_u16(q4u16));
-    q11s16 = vmulq_s16(q11s16, vreinterpretq_s16_u16(q4u16));
+  q4u16 = vmovl_u8(vget_low_u8(q10));
+  q2s16 = vmulq_s16(q2s16, vreinterpretq_s16_u16(q4u16));
+  q11s16 = vmulq_s16(q11s16, vreinterpretq_s16_u16(q4u16));
 
-    q1u8 = vandq_u8(vreinterpretq_u8_s8(q1s8), q14u8);
-    q15u8 = vandq_u8(q15u8, q9);
+  q1u8 = vandq_u8(vreinterpretq_u8_s8(q1s8), q14u8);
+  q15u8 = vandq_u8(q15u8, q9);
 
-    q1s8 = vreinterpretq_s8_u8(q1u8);
-    q2s16 = vaddw_s8(q2s16, vget_low_s8(q1s8));
-    q11s16 = vaddw_s8(q11s16, vget_high_s8(q1s8));
+  q1s8 = vreinterpretq_s8_u8(q1u8);
+  q2s16 = vaddw_s8(q2s16, vget_low_s8(q1s8));
+  q11s16 = vaddw_s8(q11s16, vget_high_s8(q1s8));
 
-    q9 = vdupq_n_u8(4);
-    // vp8_filter = clamp(vp8_filter + 3 * ( qs0 - ps0))
-    d2s8 = vqmovn_s16(q2s16);
-    d3s8 = vqmovn_s16(q11s16);
-    q1s8 = vcombine_s8(d2s8, d3s8);
-    q1u8 = vandq_u8(vreinterpretq_u8_s8(q1s8), q15u8);
-    q1s8 = vreinterpretq_s8_u8(q1u8);
+  q9 = vdupq_n_u8(4);
+  // vp8_filter = clamp(vp8_filter + 3 * ( qs0 - ps0))
+  d2s8 = vqmovn_s16(q2s16);
+  d3s8 = vqmovn_s16(q11s16);
+  q1s8 = vcombine_s8(d2s8, d3s8);
+  q1u8 = vandq_u8(vreinterpretq_u8_s8(q1s8), q15u8);
+  q1s8 = vreinterpretq_s8_u8(q1u8);
 
-    q2s8 = vqaddq_s8(q1s8, vreinterpretq_s8_u8(q10));
-    q1s8 = vqaddq_s8(q1s8, vreinterpretq_s8_u8(q9));
-    q2s8 = vshrq_n_s8(q2s8, 3);
-    q1s8 = vshrq_n_s8(q1s8, 3);
+  q2s8 = vqaddq_s8(q1s8, vreinterpretq_s8_u8(q10));
+  q1s8 = vqaddq_s8(q1s8, vreinterpretq_s8_u8(q9));
+  q2s8 = vshrq_n_s8(q2s8, 3);
+  q1s8 = vshrq_n_s8(q1s8, 3);
 
-    q11s8 = vqaddq_s8(vreinterpretq_s8_u8(q6), q2s8);
-    q10s8 = vqsubq_s8(vreinterpretq_s8_u8(q7), q1s8);
+  q11s8 = vqaddq_s8(vreinterpretq_s8_u8(q6), q2s8);
+  q10s8 = vqsubq_s8(vreinterpretq_s8_u8(q7), q1s8);
 
-    q1s8 = vrshrq_n_s8(q1s8, 1);
-    q1s8 = vbicq_s8(q1s8, vreinterpretq_s8_u8(q14u8));
+  q1s8 = vrshrq_n_s8(q1s8, 1);
+  q1s8 = vbicq_s8(q1s8, vreinterpretq_s8_u8(q14u8));
 
-    q13s8 = vqaddq_s8(vreinterpretq_s8_u8(q5), q1s8);
-    q12s8 = vqsubq_s8(vreinterpretq_s8_u8(q8), q1s8);
+  q13s8 = vqaddq_s8(vreinterpretq_s8_u8(q5), q1s8);
+  q12s8 = vqsubq_s8(vreinterpretq_s8_u8(q8), q1s8);
 
-    q0u8 = vdupq_n_u8(0x80);
-    *q8r = veorq_u8(vreinterpretq_u8_s8(q12s8), q0u8);
-    *q7r = veorq_u8(vreinterpretq_u8_s8(q10s8), q0u8);
-    *q6r = veorq_u8(vreinterpretq_u8_s8(q11s8), q0u8);
-    *q5r = veorq_u8(vreinterpretq_u8_s8(q13s8), q0u8);
-    return;
+  q0u8 = vdupq_n_u8(0x80);
+  *q8r = veorq_u8(vreinterpretq_u8_s8(q12s8), q0u8);
+  *q7r = veorq_u8(vreinterpretq_u8_s8(q10s8), q0u8);
+  *q6r = veorq_u8(vreinterpretq_u8_s8(q11s8), q0u8);
+  *q5r = veorq_u8(vreinterpretq_u8_s8(q13s8), q0u8);
+  return;
 }
 
-void vp8_loop_filter_horizontal_edge_y_neon(
-        unsigned char *src,
-        int pitch,
-        unsigned char blimit,
-        unsigned char limit,
-        unsigned char thresh) {
-    uint8x16_t qblimit, qlimit, qthresh, q3, q4;
-    uint8x16_t q5, q6, q7, q8, q9, q10;
+void vp8_loop_filter_horizontal_edge_y_neon(unsigned char *src, int pitch,
+                                            unsigned char blimit,
+                                            unsigned char limit,
+                                            unsigned char thresh) {
+  uint8x16_t qblimit, qlimit, qthresh, q3, q4;
+  uint8x16_t q5, q6, q7, q8, q9, q10;
 
-    qblimit = vdupq_n_u8(blimit);
-    qlimit  = vdupq_n_u8(limit);
-    qthresh = vdupq_n_u8(thresh);
-    src -= (pitch << 2);
+  qblimit = vdupq_n_u8(blimit);
+  qlimit = vdupq_n_u8(limit);
+  qthresh = vdupq_n_u8(thresh);
+  src -= (pitch << 2);
 
-    q3 = vld1q_u8(src);
-    src += pitch;
-    q4 = vld1q_u8(src);
-    src += pitch;
-    q5 = vld1q_u8(src);
-    src += pitch;
-    q6 = vld1q_u8(src);
-    src += pitch;
-    q7 = vld1q_u8(src);
-    src += pitch;
-    q8 = vld1q_u8(src);
-    src += pitch;
-    q9 = vld1q_u8(src);
-    src += pitch;
-    q10 = vld1q_u8(src);
+  q3 = vld1q_u8(src);
+  src += pitch;
+  q4 = vld1q_u8(src);
+  src += pitch;
+  q5 = vld1q_u8(src);
+  src += pitch;
+  q6 = vld1q_u8(src);
+  src += pitch;
+  q7 = vld1q_u8(src);
+  src += pitch;
+  q8 = vld1q_u8(src);
+  src += pitch;
+  q9 = vld1q_u8(src);
+  src += pitch;
+  q10 = vld1q_u8(src);
 
-    vp8_loop_filter_neon(qblimit, qlimit, qthresh, q3, q4,
-                         q5, q6, q7, q8, q9, q10,
-                         &q5, &q6, &q7, &q8);
+  vp8_loop_filter_neon(qblimit, qlimit, qthresh, q3, q4, q5, q6, q7, q8, q9,
+                       q10, &q5, &q6, &q7, &q8);
 
-    src -= (pitch * 5);
-    vst1q_u8(src, q5);
-    src += pitch;
-    vst1q_u8(src, q6);
-    src += pitch;
-    vst1q_u8(src, q7);
-    src += pitch;
-    vst1q_u8(src, q8);
-    return;
+  src -= (pitch * 5);
+  vst1q_u8(src, q5);
+  src += pitch;
+  vst1q_u8(src, q6);
+  src += pitch;
+  vst1q_u8(src, q7);
+  src += pitch;
+  vst1q_u8(src, q8);
+  return;
 }
 
-void vp8_loop_filter_horizontal_edge_uv_neon(
-        unsigned char *u,
-        int pitch,
-        unsigned char blimit,
-        unsigned char limit,
-        unsigned char thresh,
-        unsigned char *v) {
-    uint8x16_t qblimit, qlimit, qthresh, q3, q4;
-    uint8x16_t q5, q6, q7, q8, q9, q10;
-    uint8x8_t d6, d7, d8, d9, d10, d11, d12, d13, d14;
-    uint8x8_t d15, d16, d17, d18, d19, d20, d21;
+void vp8_loop_filter_horizontal_edge_uv_neon(unsigned char *u, int pitch,
+                                             unsigned char blimit,
+                                             unsigned char limit,
+                                             unsigned char thresh,
+                                             unsigned char *v) {
+  uint8x16_t qblimit, qlimit, qthresh, q3, q4;
+  uint8x16_t q5, q6, q7, q8, q9, q10;
+  uint8x8_t d6, d7, d8, d9, d10, d11, d12, d13, d14;
+  uint8x8_t d15, d16, d17, d18, d19, d20, d21;
 
-    qblimit = vdupq_n_u8(blimit);
-    qlimit  = vdupq_n_u8(limit);
-    qthresh = vdupq_n_u8(thresh);
+  qblimit = vdupq_n_u8(blimit);
+  qlimit = vdupq_n_u8(limit);
+  qthresh = vdupq_n_u8(thresh);
 
-    u -= (pitch << 2);
-    v -= (pitch << 2);
+  u -= (pitch << 2);
+  v -= (pitch << 2);
 
-    d6  = vld1_u8(u);
-    u += pitch;
-    d7  = vld1_u8(v);
-    v += pitch;
-    d8  = vld1_u8(u);
-    u += pitch;
-    d9  = vld1_u8(v);
-    v += pitch;
-    d10 = vld1_u8(u);
-    u += pitch;
-    d11 = vld1_u8(v);
-    v += pitch;
-    d12 = vld1_u8(u);
-    u += pitch;
-    d13 = vld1_u8(v);
-    v += pitch;
-    d14 = vld1_u8(u);
-    u += pitch;
-    d15 = vld1_u8(v);
-    v += pitch;
-    d16 = vld1_u8(u);
-    u += pitch;
-    d17 = vld1_u8(v);
-    v += pitch;
-    d18 = vld1_u8(u);
-    u += pitch;
-    d19 = vld1_u8(v);
-    v += pitch;
-    d20 = vld1_u8(u);
-    d21 = vld1_u8(v);
+  d6 = vld1_u8(u);
+  u += pitch;
+  d7 = vld1_u8(v);
+  v += pitch;
+  d8 = vld1_u8(u);
+  u += pitch;
+  d9 = vld1_u8(v);
+  v += pitch;
+  d10 = vld1_u8(u);
+  u += pitch;
+  d11 = vld1_u8(v);
+  v += pitch;
+  d12 = vld1_u8(u);
+  u += pitch;
+  d13 = vld1_u8(v);
+  v += pitch;
+  d14 = vld1_u8(u);
+  u += pitch;
+  d15 = vld1_u8(v);
+  v += pitch;
+  d16 = vld1_u8(u);
+  u += pitch;
+  d17 = vld1_u8(v);
+  v += pitch;
+  d18 = vld1_u8(u);
+  u += pitch;
+  d19 = vld1_u8(v);
+  v += pitch;
+  d20 = vld1_u8(u);
+  d21 = vld1_u8(v);
 
-    q3 = vcombine_u8(d6, d7);
-    q4 = vcombine_u8(d8, d9);
-    q5 = vcombine_u8(d10, d11);
-    q6 = vcombine_u8(d12, d13);
-    q7 = vcombine_u8(d14, d15);
-    q8 = vcombine_u8(d16, d17);
-    q9 = vcombine_u8(d18, d19);
-    q10 = vcombine_u8(d20, d21);
+  q3 = vcombine_u8(d6, d7);
+  q4 = vcombine_u8(d8, d9);
+  q5 = vcombine_u8(d10, d11);
+  q6 = vcombine_u8(d12, d13);
+  q7 = vcombine_u8(d14, d15);
+  q8 = vcombine_u8(d16, d17);
+  q9 = vcombine_u8(d18, d19);
+  q10 = vcombine_u8(d20, d21);
 
-    vp8_loop_filter_neon(qblimit, qlimit, qthresh, q3, q4,
-                         q5, q6, q7, q8, q9, q10,
-                         &q5, &q6, &q7, &q8);
+  vp8_loop_filter_neon(qblimit, qlimit, qthresh, q3, q4, q5, q6, q7, q8, q9,
+                       q10, &q5, &q6, &q7, &q8);
 
-    u -= (pitch * 5);
-    vst1_u8(u, vget_low_u8(q5));
-    u += pitch;
-    vst1_u8(u, vget_low_u8(q6));
-    u += pitch;
-    vst1_u8(u, vget_low_u8(q7));
-    u += pitch;
-    vst1_u8(u, vget_low_u8(q8));
+  u -= (pitch * 5);
+  vst1_u8(u, vget_low_u8(q5));
+  u += pitch;
+  vst1_u8(u, vget_low_u8(q6));
+  u += pitch;
+  vst1_u8(u, vget_low_u8(q7));
+  u += pitch;
+  vst1_u8(u, vget_low_u8(q8));
 
-    v -= (pitch * 5);
-    vst1_u8(v, vget_high_u8(q5));
-    v += pitch;
-    vst1_u8(v, vget_high_u8(q6));
-    v += pitch;
-    vst1_u8(v, vget_high_u8(q7));
-    v += pitch;
-    vst1_u8(v, vget_high_u8(q8));
-    return;
+  v -= (pitch * 5);
+  vst1_u8(v, vget_high_u8(q5));
+  v += pitch;
+  vst1_u8(v, vget_high_u8(q6));
+  v += pitch;
+  vst1_u8(v, vget_high_u8(q7));
+  v += pitch;
+  vst1_u8(v, vget_high_u8(q8));
+  return;
 }
 
 static INLINE void write_4x8(unsigned char *dst, int pitch,
                              const uint8x8x4_t result) {
 #ifdef VPX_INCOMPATIBLE_GCC
-    /*
-     * uint8x8x4_t result
-    00 01 02 03 | 04 05 06 07
-    10 11 12 13 | 14 15 16 17
-    20 21 22 23 | 24 25 26 27
-    30 31 32 33 | 34 35 36 37
-    ---
-    * after vtrn_u16
-    00 01 20 21 | 04 05 24 25
-    02 03 22 23 | 06 07 26 27
-    10 11 30 31 | 14 15 34 35
-    12 13 32 33 | 16 17 36 37
-    ---
-    * after vtrn_u8
-    00 10 20 30 | 04 14 24 34
-    01 11 21 31 | 05 15 25 35
-    02 12 22 32 | 06 16 26 36
-    03 13 23 33 | 07 17 27 37
-    */
-    const uint16x4x2_t r02_u16 = vtrn_u16(vreinterpret_u16_u8(result.val[0]),
-                                          vreinterpret_u16_u8(result.val[2]));
-    const uint16x4x2_t r13_u16 = vtrn_u16(vreinterpret_u16_u8(result.val[1]),
-                                          vreinterpret_u16_u8(result.val[3]));
-    const uint8x8x2_t r01_u8 = vtrn_u8(vreinterpret_u8_u16(r02_u16.val[0]),
-                                       vreinterpret_u8_u16(r13_u16.val[0]));
-    const uint8x8x2_t r23_u8 = vtrn_u8(vreinterpret_u8_u16(r02_u16.val[1]),
-                                       vreinterpret_u8_u16(r13_u16.val[1]));
-    const uint32x2_t x_0_4 = vreinterpret_u32_u8(r01_u8.val[0]);
-    const uint32x2_t x_1_5 = vreinterpret_u32_u8(r01_u8.val[1]);
-    const uint32x2_t x_2_6 = vreinterpret_u32_u8(r23_u8.val[0]);
-    const uint32x2_t x_3_7 = vreinterpret_u32_u8(r23_u8.val[1]);
-    vst1_lane_u32((uint32_t *)dst, x_0_4, 0);
-    dst += pitch;
-    vst1_lane_u32((uint32_t *)dst, x_1_5, 0);
-    dst += pitch;
-    vst1_lane_u32((uint32_t *)dst, x_2_6, 0);
-    dst += pitch;
-    vst1_lane_u32((uint32_t *)dst, x_3_7, 0);
-    dst += pitch;
-    vst1_lane_u32((uint32_t *)dst, x_0_4, 1);
-    dst += pitch;
-    vst1_lane_u32((uint32_t *)dst, x_1_5, 1);
-    dst += pitch;
-    vst1_lane_u32((uint32_t *)dst, x_2_6, 1);
-    dst += pitch;
-    vst1_lane_u32((uint32_t *)dst, x_3_7, 1);
+  /*
+   * uint8x8x4_t result
+  00 01 02 03 | 04 05 06 07
+  10 11 12 13 | 14 15 16 17
+  20 21 22 23 | 24 25 26 27
+  30 31 32 33 | 34 35 36 37
+  ---
+  * after vtrn_u16
+  00 01 20 21 | 04 05 24 25
+  02 03 22 23 | 06 07 26 27
+  10 11 30 31 | 14 15 34 35
+  12 13 32 33 | 16 17 36 37
+  ---
+  * after vtrn_u8
+  00 10 20 30 | 04 14 24 34
+  01 11 21 31 | 05 15 25 35
+  02 12 22 32 | 06 16 26 36
+  03 13 23 33 | 07 17 27 37
+  */
+  const uint16x4x2_t r02_u16 = vtrn_u16(vreinterpret_u16_u8(result.val[0]),
+                                        vreinterpret_u16_u8(result.val[2]));
+  const uint16x4x2_t r13_u16 = vtrn_u16(vreinterpret_u16_u8(result.val[1]),
+                                        vreinterpret_u16_u8(result.val[3]));
+  const uint8x8x2_t r01_u8 = vtrn_u8(vreinterpret_u8_u16(r02_u16.val[0]),
+                                     vreinterpret_u8_u16(r13_u16.val[0]));
+  const uint8x8x2_t r23_u8 = vtrn_u8(vreinterpret_u8_u16(r02_u16.val[1]),
+                                     vreinterpret_u8_u16(r13_u16.val[1]));
+  const uint32x2_t x_0_4 = vreinterpret_u32_u8(r01_u8.val[0]);
+  const uint32x2_t x_1_5 = vreinterpret_u32_u8(r01_u8.val[1]);
+  const uint32x2_t x_2_6 = vreinterpret_u32_u8(r23_u8.val[0]);
+  const uint32x2_t x_3_7 = vreinterpret_u32_u8(r23_u8.val[1]);
+  vst1_lane_u32((uint32_t *)dst, x_0_4, 0);
+  dst += pitch;
+  vst1_lane_u32((uint32_t *)dst, x_1_5, 0);
+  dst += pitch;
+  vst1_lane_u32((uint32_t *)dst, x_2_6, 0);
+  dst += pitch;
+  vst1_lane_u32((uint32_t *)dst, x_3_7, 0);
+  dst += pitch;
+  vst1_lane_u32((uint32_t *)dst, x_0_4, 1);
+  dst += pitch;
+  vst1_lane_u32((uint32_t *)dst, x_1_5, 1);
+  dst += pitch;
+  vst1_lane_u32((uint32_t *)dst, x_2_6, 1);
+  dst += pitch;
+  vst1_lane_u32((uint32_t *)dst, x_3_7, 1);
 #else
-    vst4_lane_u8(dst, result, 0);
-    dst += pitch;
-    vst4_lane_u8(dst, result, 1);
-    dst += pitch;
-    vst4_lane_u8(dst, result, 2);
-    dst += pitch;
-    vst4_lane_u8(dst, result, 3);
-    dst += pitch;
-    vst4_lane_u8(dst, result, 4);
-    dst += pitch;
-    vst4_lane_u8(dst, result, 5);
-    dst += pitch;
-    vst4_lane_u8(dst, result, 6);
-    dst += pitch;
-    vst4_lane_u8(dst, result, 7);
+  vst4_lane_u8(dst, result, 0);
+  dst += pitch;
+  vst4_lane_u8(dst, result, 1);
+  dst += pitch;
+  vst4_lane_u8(dst, result, 2);
+  dst += pitch;
+  vst4_lane_u8(dst, result, 3);
+  dst += pitch;
+  vst4_lane_u8(dst, result, 4);
+  dst += pitch;
+  vst4_lane_u8(dst, result, 5);
+  dst += pitch;
+  vst4_lane_u8(dst, result, 6);
+  dst += pitch;
+  vst4_lane_u8(dst, result, 7);
 #endif  // VPX_INCOMPATIBLE_GCC
 }
 
-void vp8_loop_filter_vertical_edge_y_neon(
-        unsigned char *src,
-        int pitch,
-        unsigned char blimit,
-        unsigned char limit,
-        unsigned char thresh) {
-    unsigned char *s, *d;
-    uint8x16_t qblimit, qlimit, qthresh, q3, q4;
-    uint8x16_t q5, q6, q7, q8, q9, q10;
-    uint8x8_t d6, d7, d8, d9, d10, d11, d12, d13, d14;
-    uint8x8_t d15, d16, d17, d18, d19, d20, d21;
-    uint32x4x2_t q2tmp0, q2tmp1, q2tmp2, q2tmp3;
-    uint16x8x2_t q2tmp4, q2tmp5, q2tmp6, q2tmp7;
-    uint8x16x2_t q2tmp8, q2tmp9, q2tmp10, q2tmp11;
-    uint8x8x4_t q4ResultH, q4ResultL;
+void vp8_loop_filter_vertical_edge_y_neon(unsigned char *src, int pitch,
+                                          unsigned char blimit,
+                                          unsigned char limit,
+                                          unsigned char thresh) {
+  unsigned char *s, *d;
+  uint8x16_t qblimit, qlimit, qthresh, q3, q4;
+  uint8x16_t q5, q6, q7, q8, q9, q10;
+  uint8x8_t d6, d7, d8, d9, d10, d11, d12, d13, d14;
+  uint8x8_t d15, d16, d17, d18, d19, d20, d21;
+  uint32x4x2_t q2tmp0, q2tmp1, q2tmp2, q2tmp3;
+  uint16x8x2_t q2tmp4, q2tmp5, q2tmp6, q2tmp7;
+  uint8x16x2_t q2tmp8, q2tmp9, q2tmp10, q2tmp11;
+  uint8x8x4_t q4ResultH, q4ResultL;
 
-    qblimit = vdupq_n_u8(blimit);
-    qlimit  = vdupq_n_u8(limit);
-    qthresh = vdupq_n_u8(thresh);
+  qblimit = vdupq_n_u8(blimit);
+  qlimit = vdupq_n_u8(limit);
+  qthresh = vdupq_n_u8(thresh);
 
-    s = src - 4;
-    d6  = vld1_u8(s);
-    s += pitch;
-    d8  = vld1_u8(s);
-    s += pitch;
-    d10 = vld1_u8(s);
-    s += pitch;
-    d12 = vld1_u8(s);
-    s += pitch;
-    d14 = vld1_u8(s);
-    s += pitch;
-    d16 = vld1_u8(s);
-    s += pitch;
-    d18 = vld1_u8(s);
-    s += pitch;
-    d20 = vld1_u8(s);
-    s += pitch;
-    d7  = vld1_u8(s);
-    s += pitch;
-    d9  = vld1_u8(s);
-    s += pitch;
-    d11 = vld1_u8(s);
-    s += pitch;
-    d13 = vld1_u8(s);
-    s += pitch;
-    d15 = vld1_u8(s);
-    s += pitch;
-    d17 = vld1_u8(s);
-    s += pitch;
-    d19 = vld1_u8(s);
-    s += pitch;
-    d21 = vld1_u8(s);
+  s = src - 4;
+  d6 = vld1_u8(s);
+  s += pitch;
+  d8 = vld1_u8(s);
+  s += pitch;
+  d10 = vld1_u8(s);
+  s += pitch;
+  d12 = vld1_u8(s);
+  s += pitch;
+  d14 = vld1_u8(s);
+  s += pitch;
+  d16 = vld1_u8(s);
+  s += pitch;
+  d18 = vld1_u8(s);
+  s += pitch;
+  d20 = vld1_u8(s);
+  s += pitch;
+  d7 = vld1_u8(s);
+  s += pitch;
+  d9 = vld1_u8(s);
+  s += pitch;
+  d11 = vld1_u8(s);
+  s += pitch;
+  d13 = vld1_u8(s);
+  s += pitch;
+  d15 = vld1_u8(s);
+  s += pitch;
+  d17 = vld1_u8(s);
+  s += pitch;
+  d19 = vld1_u8(s);
+  s += pitch;
+  d21 = vld1_u8(s);
 
-    q3 = vcombine_u8(d6, d7);
-    q4 = vcombine_u8(d8, d9);
-    q5 = vcombine_u8(d10, d11);
-    q6 = vcombine_u8(d12, d13);
-    q7 = vcombine_u8(d14, d15);
-    q8 = vcombine_u8(d16, d17);
-    q9 = vcombine_u8(d18, d19);
-    q10 = vcombine_u8(d20, d21);
+  q3 = vcombine_u8(d6, d7);
+  q4 = vcombine_u8(d8, d9);
+  q5 = vcombine_u8(d10, d11);
+  q6 = vcombine_u8(d12, d13);
+  q7 = vcombine_u8(d14, d15);
+  q8 = vcombine_u8(d16, d17);
+  q9 = vcombine_u8(d18, d19);
+  q10 = vcombine_u8(d20, d21);
 
-    q2tmp0 = vtrnq_u32(vreinterpretq_u32_u8(q3), vreinterpretq_u32_u8(q7));
-    q2tmp1 = vtrnq_u32(vreinterpretq_u32_u8(q4), vreinterpretq_u32_u8(q8));
-    q2tmp2 = vtrnq_u32(vreinterpretq_u32_u8(q5), vreinterpretq_u32_u8(q9));
-    q2tmp3 = vtrnq_u32(vreinterpretq_u32_u8(q6), vreinterpretq_u32_u8(q10));
+  q2tmp0 = vtrnq_u32(vreinterpretq_u32_u8(q3), vreinterpretq_u32_u8(q7));
+  q2tmp1 = vtrnq_u32(vreinterpretq_u32_u8(q4), vreinterpretq_u32_u8(q8));
+  q2tmp2 = vtrnq_u32(vreinterpretq_u32_u8(q5), vreinterpretq_u32_u8(q9));
+  q2tmp3 = vtrnq_u32(vreinterpretq_u32_u8(q6), vreinterpretq_u32_u8(q10));
 
-    q2tmp4 = vtrnq_u16(vreinterpretq_u16_u32(q2tmp0.val[0]),
-                       vreinterpretq_u16_u32(q2tmp2.val[0]));
-    q2tmp5 = vtrnq_u16(vreinterpretq_u16_u32(q2tmp1.val[0]),
-                       vreinterpretq_u16_u32(q2tmp3.val[0]));
-    q2tmp6 = vtrnq_u16(vreinterpretq_u16_u32(q2tmp0.val[1]),
-                       vreinterpretq_u16_u32(q2tmp2.val[1]));
-    q2tmp7 = vtrnq_u16(vreinterpretq_u16_u32(q2tmp1.val[1]),
-                       vreinterpretq_u16_u32(q2tmp3.val[1]));
+  q2tmp4 = vtrnq_u16(vreinterpretq_u16_u32(q2tmp0.val[0]),
+                     vreinterpretq_u16_u32(q2tmp2.val[0]));
+  q2tmp5 = vtrnq_u16(vreinterpretq_u16_u32(q2tmp1.val[0]),
+                     vreinterpretq_u16_u32(q2tmp3.val[0]));
+  q2tmp6 = vtrnq_u16(vreinterpretq_u16_u32(q2tmp0.val[1]),
+                     vreinterpretq_u16_u32(q2tmp2.val[1]));
+  q2tmp7 = vtrnq_u16(vreinterpretq_u16_u32(q2tmp1.val[1]),
+                     vreinterpretq_u16_u32(q2tmp3.val[1]));
 
-    q2tmp8  = vtrnq_u8(vreinterpretq_u8_u16(q2tmp4.val[0]),
-                       vreinterpretq_u8_u16(q2tmp5.val[0]));
-    q2tmp9  = vtrnq_u8(vreinterpretq_u8_u16(q2tmp4.val[1]),
-                       vreinterpretq_u8_u16(q2tmp5.val[1]));
-    q2tmp10 = vtrnq_u8(vreinterpretq_u8_u16(q2tmp6.val[0]),
-                       vreinterpretq_u8_u16(q2tmp7.val[0]));
-    q2tmp11 = vtrnq_u8(vreinterpretq_u8_u16(q2tmp6.val[1]),
-                       vreinterpretq_u8_u16(q2tmp7.val[1]));
+  q2tmp8 = vtrnq_u8(vreinterpretq_u8_u16(q2tmp4.val[0]),
+                    vreinterpretq_u8_u16(q2tmp5.val[0]));
+  q2tmp9 = vtrnq_u8(vreinterpretq_u8_u16(q2tmp4.val[1]),
+                    vreinterpretq_u8_u16(q2tmp5.val[1]));
+  q2tmp10 = vtrnq_u8(vreinterpretq_u8_u16(q2tmp6.val[0]),
+                     vreinterpretq_u8_u16(q2tmp7.val[0]));
+  q2tmp11 = vtrnq_u8(vreinterpretq_u8_u16(q2tmp6.val[1]),
+                     vreinterpretq_u8_u16(q2tmp7.val[1]));
 
-    q3 = q2tmp8.val[0];
-    q4 = q2tmp8.val[1];
-    q5 = q2tmp9.val[0];
-    q6 = q2tmp9.val[1];
-    q7 = q2tmp10.val[0];
-    q8 = q2tmp10.val[1];
-    q9 = q2tmp11.val[0];
-    q10 = q2tmp11.val[1];
+  q3 = q2tmp8.val[0];
+  q4 = q2tmp8.val[1];
+  q5 = q2tmp9.val[0];
+  q6 = q2tmp9.val[1];
+  q7 = q2tmp10.val[0];
+  q8 = q2tmp10.val[1];
+  q9 = q2tmp11.val[0];
+  q10 = q2tmp11.val[1];
 
-    vp8_loop_filter_neon(qblimit, qlimit, qthresh, q3, q4,
-                         q5, q6, q7, q8, q9, q10,
-                         &q5, &q6, &q7, &q8);
+  vp8_loop_filter_neon(qblimit, qlimit, qthresh, q3, q4, q5, q6, q7, q8, q9,
+                       q10, &q5, &q6, &q7, &q8);
 
-    q4ResultL.val[0] = vget_low_u8(q5);   // d10
-    q4ResultL.val[1] = vget_low_u8(q6);   // d12
-    q4ResultL.val[2] = vget_low_u8(q7);   // d14
-    q4ResultL.val[3] = vget_low_u8(q8);   // d16
-    q4ResultH.val[0] = vget_high_u8(q5);  // d11
-    q4ResultH.val[1] = vget_high_u8(q6);  // d13
-    q4ResultH.val[2] = vget_high_u8(q7);  // d15
-    q4ResultH.val[3] = vget_high_u8(q8);  // d17
+  q4ResultL.val[0] = vget_low_u8(q5);   // d10
+  q4ResultL.val[1] = vget_low_u8(q6);   // d12
+  q4ResultL.val[2] = vget_low_u8(q7);   // d14
+  q4ResultL.val[3] = vget_low_u8(q8);   // d16
+  q4ResultH.val[0] = vget_high_u8(q5);  // d11
+  q4ResultH.val[1] = vget_high_u8(q6);  // d13
+  q4ResultH.val[2] = vget_high_u8(q7);  // d15
+  q4ResultH.val[3] = vget_high_u8(q8);  // d17
 
-    d = src - 2;
-    write_4x8(d, pitch, q4ResultL);
-    d += pitch * 8;
-    write_4x8(d, pitch, q4ResultH);
+  d = src - 2;
+  write_4x8(d, pitch, q4ResultL);
+  d += pitch * 8;
+  write_4x8(d, pitch, q4ResultH);
 }
 
-void vp8_loop_filter_vertical_edge_uv_neon(
-        unsigned char *u,
-        int pitch,
-        unsigned char blimit,
-        unsigned char limit,
-        unsigned char thresh,
-        unsigned char *v) {
-    unsigned char *us, *ud;
-    unsigned char *vs, *vd;
-    uint8x16_t qblimit, qlimit, qthresh, q3, q4;
-    uint8x16_t q5, q6, q7, q8, q9, q10;
-    uint8x8_t d6, d7, d8, d9, d10, d11, d12, d13, d14;
-    uint8x8_t d15, d16, d17, d18, d19, d20, d21;
-    uint32x4x2_t q2tmp0, q2tmp1, q2tmp2, q2tmp3;
-    uint16x8x2_t q2tmp4, q2tmp5, q2tmp6, q2tmp7;
-    uint8x16x2_t q2tmp8, q2tmp9, q2tmp10, q2tmp11;
-    uint8x8x4_t q4ResultH, q4ResultL;
+void vp8_loop_filter_vertical_edge_uv_neon(unsigned char *u, int pitch,
+                                           unsigned char blimit,
+                                           unsigned char limit,
+                                           unsigned char thresh,
+                                           unsigned char *v) {
+  unsigned char *us, *ud;
+  unsigned char *vs, *vd;
+  uint8x16_t qblimit, qlimit, qthresh, q3, q4;
+  uint8x16_t q5, q6, q7, q8, q9, q10;
+  uint8x8_t d6, d7, d8, d9, d10, d11, d12, d13, d14;
+  uint8x8_t d15, d16, d17, d18, d19, d20, d21;
+  uint32x4x2_t q2tmp0, q2tmp1, q2tmp2, q2tmp3;
+  uint16x8x2_t q2tmp4, q2tmp5, q2tmp6, q2tmp7;
+  uint8x16x2_t q2tmp8, q2tmp9, q2tmp10, q2tmp11;
+  uint8x8x4_t q4ResultH, q4ResultL;
 
-    qblimit = vdupq_n_u8(blimit);
-    qlimit  = vdupq_n_u8(limit);
-    qthresh = vdupq_n_u8(thresh);
+  qblimit = vdupq_n_u8(blimit);
+  qlimit = vdupq_n_u8(limit);
+  qthresh = vdupq_n_u8(thresh);
 
-    us = u - 4;
-    d6 = vld1_u8(us);
-    us += pitch;
-    d8 = vld1_u8(us);
-    us += pitch;
-    d10 = vld1_u8(us);
-    us += pitch;
-    d12 = vld1_u8(us);
-    us += pitch;
-    d14 = vld1_u8(us);
-    us += pitch;
-    d16 = vld1_u8(us);
-    us += pitch;
-    d18 = vld1_u8(us);
-    us += pitch;
-    d20 = vld1_u8(us);
+  us = u - 4;
+  d6 = vld1_u8(us);
+  us += pitch;
+  d8 = vld1_u8(us);
+  us += pitch;
+  d10 = vld1_u8(us);
+  us += pitch;
+  d12 = vld1_u8(us);
+  us += pitch;
+  d14 = vld1_u8(us);
+  us += pitch;
+  d16 = vld1_u8(us);
+  us += pitch;
+  d18 = vld1_u8(us);
+  us += pitch;
+  d20 = vld1_u8(us);
 
-    vs = v - 4;
-    d7 = vld1_u8(vs);
-    vs += pitch;
-    d9 = vld1_u8(vs);
-    vs += pitch;
-    d11 = vld1_u8(vs);
-    vs += pitch;
-    d13 = vld1_u8(vs);
-    vs += pitch;
-    d15 = vld1_u8(vs);
-    vs += pitch;
-    d17 = vld1_u8(vs);
-    vs += pitch;
-    d19 = vld1_u8(vs);
-    vs += pitch;
-    d21 = vld1_u8(vs);
+  vs = v - 4;
+  d7 = vld1_u8(vs);
+  vs += pitch;
+  d9 = vld1_u8(vs);
+  vs += pitch;
+  d11 = vld1_u8(vs);
+  vs += pitch;
+  d13 = vld1_u8(vs);
+  vs += pitch;
+  d15 = vld1_u8(vs);
+  vs += pitch;
+  d17 = vld1_u8(vs);
+  vs += pitch;
+  d19 = vld1_u8(vs);
+  vs += pitch;
+  d21 = vld1_u8(vs);
 
-    q3 = vcombine_u8(d6, d7);
-    q4 = vcombine_u8(d8, d9);
-    q5 = vcombine_u8(d10, d11);
-    q6 = vcombine_u8(d12, d13);
-    q7 = vcombine_u8(d14, d15);
-    q8 = vcombine_u8(d16, d17);
-    q9 = vcombine_u8(d18, d19);
-    q10 = vcombine_u8(d20, d21);
+  q3 = vcombine_u8(d6, d7);
+  q4 = vcombine_u8(d8, d9);
+  q5 = vcombine_u8(d10, d11);
+  q6 = vcombine_u8(d12, d13);
+  q7 = vcombine_u8(d14, d15);
+  q8 = vcombine_u8(d16, d17);
+  q9 = vcombine_u8(d18, d19);
+  q10 = vcombine_u8(d20, d21);
 
-    q2tmp0 = vtrnq_u32(vreinterpretq_u32_u8(q3), vreinterpretq_u32_u8(q7));
-    q2tmp1 = vtrnq_u32(vreinterpretq_u32_u8(q4), vreinterpretq_u32_u8(q8));
-    q2tmp2 = vtrnq_u32(vreinterpretq_u32_u8(q5), vreinterpretq_u32_u8(q9));
-    q2tmp3 = vtrnq_u32(vreinterpretq_u32_u8(q6), vreinterpretq_u32_u8(q10));
+  q2tmp0 = vtrnq_u32(vreinterpretq_u32_u8(q3), vreinterpretq_u32_u8(q7));
+  q2tmp1 = vtrnq_u32(vreinterpretq_u32_u8(q4), vreinterpretq_u32_u8(q8));
+  q2tmp2 = vtrnq_u32(vreinterpretq_u32_u8(q5), vreinterpretq_u32_u8(q9));
+  q2tmp3 = vtrnq_u32(vreinterpretq_u32_u8(q6), vreinterpretq_u32_u8(q10));
 
-    q2tmp4 = vtrnq_u16(vreinterpretq_u16_u32(q2tmp0.val[0]),
-                       vreinterpretq_u16_u32(q2tmp2.val[0]));
-    q2tmp5 = vtrnq_u16(vreinterpretq_u16_u32(q2tmp1.val[0]),
-                       vreinterpretq_u16_u32(q2tmp3.val[0]));
-    q2tmp6 = vtrnq_u16(vreinterpretq_u16_u32(q2tmp0.val[1]),
-                       vreinterpretq_u16_u32(q2tmp2.val[1]));
-    q2tmp7 = vtrnq_u16(vreinterpretq_u16_u32(q2tmp1.val[1]),
-                       vreinterpretq_u16_u32(q2tmp3.val[1]));
+  q2tmp4 = vtrnq_u16(vreinterpretq_u16_u32(q2tmp0.val[0]),
+                     vreinterpretq_u16_u32(q2tmp2.val[0]));
+  q2tmp5 = vtrnq_u16(vreinterpretq_u16_u32(q2tmp1.val[0]),
+                     vreinterpretq_u16_u32(q2tmp3.val[0]));
+  q2tmp6 = vtrnq_u16(vreinterpretq_u16_u32(q2tmp0.val[1]),
+                     vreinterpretq_u16_u32(q2tmp2.val[1]));
+  q2tmp7 = vtrnq_u16(vreinterpretq_u16_u32(q2tmp1.val[1]),
+                     vreinterpretq_u16_u32(q2tmp3.val[1]));
 
-    q2tmp8  = vtrnq_u8(vreinterpretq_u8_u16(q2tmp4.val[0]),
-                       vreinterpretq_u8_u16(q2tmp5.val[0]));
-    q2tmp9  = vtrnq_u8(vreinterpretq_u8_u16(q2tmp4.val[1]),
-                       vreinterpretq_u8_u16(q2tmp5.val[1]));
-    q2tmp10 = vtrnq_u8(vreinterpretq_u8_u16(q2tmp6.val[0]),
-                       vreinterpretq_u8_u16(q2tmp7.val[0]));
-    q2tmp11 = vtrnq_u8(vreinterpretq_u8_u16(q2tmp6.val[1]),
-                       vreinterpretq_u8_u16(q2tmp7.val[1]));
+  q2tmp8 = vtrnq_u8(vreinterpretq_u8_u16(q2tmp4.val[0]),
+                    vreinterpretq_u8_u16(q2tmp5.val[0]));
+  q2tmp9 = vtrnq_u8(vreinterpretq_u8_u16(q2tmp4.val[1]),
+                    vreinterpretq_u8_u16(q2tmp5.val[1]));
+  q2tmp10 = vtrnq_u8(vreinterpretq_u8_u16(q2tmp6.val[0]),
+                     vreinterpretq_u8_u16(q2tmp7.val[0]));
+  q2tmp11 = vtrnq_u8(vreinterpretq_u8_u16(q2tmp6.val[1]),
+                     vreinterpretq_u8_u16(q2tmp7.val[1]));
 
-    q3 = q2tmp8.val[0];
-    q4 = q2tmp8.val[1];
-    q5 = q2tmp9.val[0];
-    q6 = q2tmp9.val[1];
-    q7 = q2tmp10.val[0];
-    q8 = q2tmp10.val[1];
-    q9 = q2tmp11.val[0];
-    q10 = q2tmp11.val[1];
+  q3 = q2tmp8.val[0];
+  q4 = q2tmp8.val[1];
+  q5 = q2tmp9.val[0];
+  q6 = q2tmp9.val[1];
+  q7 = q2tmp10.val[0];
+  q8 = q2tmp10.val[1];
+  q9 = q2tmp11.val[0];
+  q10 = q2tmp11.val[1];
 
-    vp8_loop_filter_neon(qblimit, qlimit, qthresh, q3, q4,
-                         q5, q6, q7, q8, q9, q10,
-                         &q5, &q6, &q7, &q8);
+  vp8_loop_filter_neon(qblimit, qlimit, qthresh, q3, q4, q5, q6, q7, q8, q9,
+                       q10, &q5, &q6, &q7, &q8);
 
-    q4ResultL.val[0] = vget_low_u8(q5);   // d10
-    q4ResultL.val[1] = vget_low_u8(q6);   // d12
-    q4ResultL.val[2] = vget_low_u8(q7);   // d14
-    q4ResultL.val[3] = vget_low_u8(q8);   // d16
-    ud = u - 2;
-    write_4x8(ud, pitch, q4ResultL);
+  q4ResultL.val[0] = vget_low_u8(q5);  // d10
+  q4ResultL.val[1] = vget_low_u8(q6);  // d12
+  q4ResultL.val[2] = vget_low_u8(q7);  // d14
+  q4ResultL.val[3] = vget_low_u8(q8);  // d16
+  ud = u - 2;
+  write_4x8(ud, pitch, q4ResultL);
 
-    q4ResultH.val[0] = vget_high_u8(q5);  // d11
-    q4ResultH.val[1] = vget_high_u8(q6);  // d13
-    q4ResultH.val[2] = vget_high_u8(q7);  // d15
-    q4ResultH.val[3] = vget_high_u8(q8);  // d17
-    vd = v - 2;
-    write_4x8(vd, pitch, q4ResultH);
+  q4ResultH.val[0] = vget_high_u8(q5);  // d11
+  q4ResultH.val[1] = vget_high_u8(q6);  // d13
+  q4ResultH.val[2] = vget_high_u8(q7);  // d15
+  q4ResultH.val[3] = vget_high_u8(q8);  // d17
+  vd = v - 2;
+  write_4x8(vd, pitch, q4ResultH);
 }
--- a/vp8/common/blockd.c
+++ b/vp8/common/blockd.c
@@ -8,15 +8,12 @@
  *  be found in the AUTHORS file in the root of the source tree.
  */
 
-
 #include "blockd.h"
 #include "vpx_mem/vpx_mem.h"
 
-const unsigned char vp8_block2left[25] =
-{
-    0, 0, 0, 0, 1, 1, 1, 1, 2, 2, 2, 2, 3, 3, 3, 3, 4, 4, 5, 5, 6, 6, 7, 7, 8
+const unsigned char vp8_block2left[25] = {
+  0, 0, 0, 0, 1, 1, 1, 1, 2, 2, 2, 2, 3, 3, 3, 3, 4, 4, 5, 5, 6, 6, 7, 7, 8
 };
-const unsigned char vp8_block2above[25] =
-{
-    0, 1, 2, 3, 0, 1, 2, 3, 0, 1, 2, 3, 0, 1, 2, 3, 4, 5, 4, 5, 6, 7, 6, 7, 8
+const unsigned char vp8_block2above[25] = {
+  0, 1, 2, 3, 0, 1, 2, 3, 0, 1, 2, 3, 0, 1, 2, 3, 4, 5, 4, 5, 6, 7, 6, 7, 8
 };
--- a/vp8/common/blockd.h
+++ b/vp8/common/blockd.h
@@ -8,7 +8,6 @@
  *  be found in the AUTHORS file in the root of the source tree.
  */
 
-
 #ifndef VP8_COMMON_BLOCKD_H_
 #define VP8_COMMON_BLOCKD_H_
 
@@ -28,109 +27,95 @@
 #define DCPREDSIMTHRESH 0
 #define DCPREDCNTTHRESH 3
 
-#define MB_FEATURE_TREE_PROBS   3
-#define MAX_MB_SEGMENTS         4
+#define MB_FEATURE_TREE_PROBS 3
+#define MAX_MB_SEGMENTS 4
 
-#define MAX_REF_LF_DELTAS       4
-#define MAX_MODE_LF_DELTAS      4
+#define MAX_REF_LF_DELTAS 4
+#define MAX_MODE_LF_DELTAS 4
 
 /* Segment Feature Masks */
-#define SEGMENT_DELTADATA   0
-#define SEGMENT_ABSDATA     1
+#define SEGMENT_DELTADATA 0
+#define SEGMENT_ABSDATA 1
 
-typedef struct
-{
-    int r, c;
-} POS;
+typedef struct { int r, c; } POS;
 
-#define PLANE_TYPE_Y_NO_DC    0
-#define PLANE_TYPE_Y2         1
-#define PLANE_TYPE_UV         2
-#define PLANE_TYPE_Y_WITH_DC  3
+#define PLANE_TYPE_Y_NO_DC 0
+#define PLANE_TYPE_Y2 1
+#define PLANE_TYPE_UV 2
+#define PLANE_TYPE_Y_WITH_DC 3
 
-
 typedef char ENTROPY_CONTEXT;
-typedef struct
-{
-    ENTROPY_CONTEXT y1[4];
-    ENTROPY_CONTEXT u[2];
-    ENTROPY_CONTEXT v[2];
-    ENTROPY_CONTEXT y2;
+typedef struct {
+  ENTROPY_CONTEXT y1[4];
+  ENTROPY_CONTEXT u[2];
+  ENTROPY_CONTEXT v[2];
+  ENTROPY_CONTEXT y2;
 } ENTROPY_CONTEXT_PLANES;
 
 extern const unsigned char vp8_block2left[25];
 extern const unsigned char vp8_block2above[25];
 
-#define VP8_COMBINEENTROPYCONTEXTS( Dest, A, B) \
-    Dest = (A)+(B);
+#define VP8_COMBINEENTROPYCONTEXTS(Dest, A, B) Dest = (A) + (B);
 
+typedef enum { KEY_FRAME = 0, INTER_FRAME = 1 } FRAME_TYPE;
 
-typedef enum
-{
-    KEY_FRAME = 0,
-    INTER_FRAME = 1
-} FRAME_TYPE;
+typedef enum {
+  DC_PRED, /* average of above and left pixels */
+  V_PRED,  /* vertical prediction */
+  H_PRED,  /* horizontal prediction */
+  TM_PRED, /* Truemotion prediction */
+  B_PRED,  /* block based prediction, each block has its own prediction mode */
 
-typedef enum
-{
-    DC_PRED,            /* average of above and left pixels */
-    V_PRED,             /* vertical prediction */
-    H_PRED,             /* horizontal prediction */
-    TM_PRED,            /* Truemotion prediction */
-    B_PRED,             /* block based prediction, each block has its own prediction mode */
+  NEARESTMV,
+  NEARMV,
+  ZEROMV,
+  NEWMV,
+  SPLITMV,
 
-    NEARESTMV,
-    NEARMV,
-    ZEROMV,
-    NEWMV,
-    SPLITMV,
-
-    MB_MODE_COUNT
+  MB_MODE_COUNT
 } MB_PREDICTION_MODE;
 
 /* Macroblock level features */
-typedef enum
-{
-    MB_LVL_ALT_Q = 0,               /* Use alternate Quantizer .... */
-    MB_LVL_ALT_LF = 1,              /* Use alternate loop filter value... */
-    MB_LVL_MAX = 2                  /* Number of MB level features supported */
+typedef enum {
+  MB_LVL_ALT_Q = 0,  /* Use alternate Quantizer .... */
+  MB_LVL_ALT_LF = 1, /* Use alternate loop filter value... */
+  MB_LVL_MAX = 2     /* Number of MB level features supported */
 
 } MB_LVL_FEATURES;
 
 /* Segment Feature Masks */
-#define SEGMENT_ALTQ    0x01
-#define SEGMENT_ALT_LF  0x02
+#define SEGMENT_ALTQ 0x01
+#define SEGMENT_ALT_LF 0x02
 
-#define VP8_YMODES  (B_PRED + 1)
+#define VP8_YMODES (B_PRED + 1)
 #define VP8_UV_MODES (TM_PRED + 1)
 
 #define VP8_MVREFS (1 + SPLITMV - NEARESTMV)
 
-typedef enum
-{
-    B_DC_PRED,          /* average of above and left pixels */
-    B_TM_PRED,
+typedef enum {
+  B_DC_PRED, /* average of above and left pixels */
+  B_TM_PRED,
 
-    B_VE_PRED,           /* vertical prediction */
-    B_HE_PRED,           /* horizontal prediction */
+  B_VE_PRED, /* vertical prediction */
+  B_HE_PRED, /* horizontal prediction */
 
-    B_LD_PRED,
-    B_RD_PRED,
+  B_LD_PRED,
+  B_RD_PRED,
 
-    B_VR_PRED,
-    B_VL_PRED,
-    B_HD_PRED,
-    B_HU_PRED,
+  B_VR_PRED,
+  B_VL_PRED,
+  B_HD_PRED,
+  B_HU_PRED,
 
-    LEFT4X4,
-    ABOVE4X4,
-    ZERO4X4,
-    NEW4X4,
+  LEFT4X4,
+  ABOVE4X4,
+  ZERO4X4,
+  NEW4X4,
 
-    B_MODE_COUNT
+  B_MODE_COUNT
 } B_PREDICTION_MODE;
 
-#define VP8_BINTRAMODES (B_HU_PRED + 1)  /* 10 */
+#define VP8_BINTRAMODES (B_HU_PRED + 1) /* 10 */
 #define VP8_SUBMVREFS (1 + NEW4X4 - LEFT4X4)
 
 /* For keyframes, intra block modes are predicted by the (already decoded)
@@ -137,170 +122,170 @@
    modes for the Y blocks to the left and above us; for interframes, there
    is a single probability table. */
 
-union b_mode_info
-{
-    B_PREDICTION_MODE as_mode;
-    int_mv mv;
+union b_mode_info {
+  B_PREDICTION_MODE as_mode;
+  int_mv mv;
 };
 
-typedef enum
-{
-    INTRA_FRAME = 0,
-    LAST_FRAME = 1,
-    GOLDEN_FRAME = 2,
-    ALTREF_FRAME = 3,
-    MAX_REF_FRAMES = 4
+typedef enum {
+  INTRA_FRAME = 0,
+  LAST_FRAME = 1,
+  GOLDEN_FRAME = 2,
+  ALTREF_FRAME = 3,
+  MAX_REF_FRAMES = 4
 } MV_REFERENCE_FRAME;
 
-typedef struct
-{
-    uint8_t mode, uv_mode;
-    uint8_t ref_frame;
-    uint8_t is_4x4;
-    int_mv mv;
+typedef struct {
+  uint8_t mode, uv_mode;
+  uint8_t ref_frame;
+  uint8_t is_4x4;
+  int_mv mv;
 
-    uint8_t partitioning;
-    uint8_t mb_skip_coeff;                                /* does this mb has coefficients at all, 1=no coefficients, 0=need decode tokens */
-    uint8_t need_to_clamp_mvs;
-    uint8_t segment_id;                  /* Which set of segmentation parameters should be used for this MB */
+  uint8_t partitioning;
+  /* does this mb has coefficients at all, 1=no coefficients, 0=need decode
+     tokens */
+  uint8_t mb_skip_coeff;
+  uint8_t need_to_clamp_mvs;
+  /* Which set of segmentation parameters should be used for this MB */
+  uint8_t segment_id;
 } MB_MODE_INFO;
 
-typedef struct modeinfo
-{
-    MB_MODE_INFO mbmi;
-    union b_mode_info bmi[16];
+typedef struct modeinfo {
+  MB_MODE_INFO mbmi;
+  union b_mode_info bmi[16];
 } MODE_INFO;
 
 #if CONFIG_MULTI_RES_ENCODING
 /* The mb-level information needed to be stored for higher-resolution encoder */
-typedef struct
-{
-    MB_PREDICTION_MODE mode;
-    MV_REFERENCE_FRAME ref_frame;
-    int_mv mv;
-    int dissim;    /* dissimilarity level of the macroblock */
+typedef struct {
+  MB_PREDICTION_MODE mode;
+  MV_REFERENCE_FRAME ref_frame;
+  int_mv mv;
+  int dissim; /* dissimilarity level of the macroblock */
 } LOWER_RES_MB_INFO;
 
 /* The frame-level information needed to be stored for higher-resolution
  *  encoder */
-typedef struct
-{
-    FRAME_TYPE frame_type;
-    int is_frame_dropped;
-    // The frame rate for the lowest resolution.
-    double low_res_framerate;
-    /* The frame number of each reference frames */
-    unsigned int low_res_ref_frames[MAX_REF_FRAMES];
-    // The video frame counter value for the key frame, for lowest resolution.
-    unsigned int key_frame_counter_value;
-    LOWER_RES_MB_INFO *mb_info;
+typedef struct {
+  FRAME_TYPE frame_type;
+  int is_frame_dropped;
+  // The frame rate for the lowest resolution.
+  double low_res_framerate;
+  /* The frame number of each reference frames */
+  unsigned int low_res_ref_frames[MAX_REF_FRAMES];
+  // The video frame counter value for the key frame, for lowest resolution.
+  unsigned int key_frame_counter_value;
+  LOWER_RES_MB_INFO *mb_info;
 } LOWER_RES_FRAME_INFO;
 #endif
 
-typedef struct blockd
-{
-    short *qcoeff;
-    short *dqcoeff;
-    unsigned char  *predictor;
-    short *dequant;
+typedef struct blockd {
+  short *qcoeff;
+  short *dqcoeff;
+  unsigned char *predictor;
+  short *dequant;
 
-    int offset;
-    char *eob;
+  int offset;
+  char *eob;
 
-    union b_mode_info bmi;
+  union b_mode_info bmi;
 } BLOCKD;
 
-typedef void (*vp8_subpix_fn_t)(unsigned char *src, int src_pitch, int xofst, int yofst, unsigned char *dst, int dst_pitch);
+typedef void (*vp8_subpix_fn_t)(unsigned char *src, int src_pitch, int xofst,
+                                int yofst, unsigned char *dst, int dst_pitch);
 
-typedef struct macroblockd
-{
-    DECLARE_ALIGNED(16, unsigned char,  predictor[384]);
-    DECLARE_ALIGNED(16, short, qcoeff[400]);
-    DECLARE_ALIGNED(16, short, dqcoeff[400]);
-    DECLARE_ALIGNED(16, char,  eobs[25]);
+typedef struct macroblockd {
+  DECLARE_ALIGNED(16, unsigned char, predictor[384]);
+  DECLARE_ALIGNED(16, short, qcoeff[400]);
+  DECLARE_ALIGNED(16, short, dqcoeff[400]);
+  DECLARE_ALIGNED(16, char, eobs[25]);
 
-    DECLARE_ALIGNED(16, short,  dequant_y1[16]);
-    DECLARE_ALIGNED(16, short,  dequant_y1_dc[16]);
-    DECLARE_ALIGNED(16, short,  dequant_y2[16]);
-    DECLARE_ALIGNED(16, short,  dequant_uv[16]);
+  DECLARE_ALIGNED(16, short, dequant_y1[16]);
+  DECLARE_ALIGNED(16, short, dequant_y1_dc[16]);
+  DECLARE_ALIGNED(16, short, dequant_y2[16]);
+  DECLARE_ALIGNED(16, short, dequant_uv[16]);
 
-    /* 16 Y blocks, 4 U, 4 V, 1 DC 2nd order block, each with 16 entries. */
-    BLOCKD block[25];
-    int fullpixel_mask;
+  /* 16 Y blocks, 4 U, 4 V, 1 DC 2nd order block, each with 16 entries. */
+  BLOCKD block[25];
+  int fullpixel_mask;
 
-    YV12_BUFFER_CONFIG pre; /* Filtered copy of previous frame reconstruction */
-    YV12_BUFFER_CONFIG dst;
+  YV12_BUFFER_CONFIG pre; /* Filtered copy of previous frame reconstruction */
+  YV12_BUFFER_CONFIG dst;
 
-    MODE_INFO *mode_info_context;
-    int mode_info_stride;
+  MODE_INFO *mode_info_context;
+  int mode_info_stride;
 
-    FRAME_TYPE frame_type;
+  FRAME_TYPE frame_type;
 
-    int up_available;
-    int left_available;
+  int up_available;
+  int left_available;
 
-    unsigned char *recon_above[3];
-    unsigned char *recon_left[3];
-    int recon_left_stride[2];
+  unsigned char *recon_above[3];
+  unsigned char *recon_left[3];
+  int recon_left_stride[2];
 
-    /* Y,U,V,Y2 */
-    ENTROPY_CONTEXT_PLANES *above_context;
-    ENTROPY_CONTEXT_PLANES *left_context;
+  /* Y,U,V,Y2 */
+  ENTROPY_CONTEXT_PLANES *above_context;
+  ENTROPY_CONTEXT_PLANES *left_context;
 
-    /* 0 indicates segmentation at MB level is not enabled. Otherwise the individual bits indicate which features are active. */
-    unsigned char segmentation_enabled;
+  /* 0 indicates segmentation at MB level is not enabled. Otherwise the
+   * individual bits indicate which features are active. */
+  unsigned char segmentation_enabled;
 
-    /* 0 (do not update) 1 (update) the macroblock segmentation map. */
-    unsigned char update_mb_segmentation_map;
+  /* 0 (do not update) 1 (update) the macroblock segmentation map. */
+  unsigned char update_mb_segmentation_map;
 
-    /* 0 (do not update) 1 (update) the macroblock segmentation feature data. */
-    unsigned char update_mb_segmentation_data;
+  /* 0 (do not update) 1 (update) the macroblock segmentation feature data. */
+  unsigned char update_mb_segmentation_data;
 
-    /* 0 (do not update) 1 (update) the macroblock segmentation feature data. */
-    unsigned char mb_segement_abs_delta;
+  /* 0 (do not update) 1 (update) the macroblock segmentation feature data. */
+  unsigned char mb_segement_abs_delta;
 
-    /* Per frame flags that define which MB level features (such as quantizer or loop filter level) */
-    /* are enabled and when enabled the proabilities used to decode the per MB flags in MB_MODE_INFO */
-    vp8_prob mb_segment_tree_probs[MB_FEATURE_TREE_PROBS];         /* Probability Tree used to code Segment number */
+  /* Per frame flags that define which MB level features (such as quantizer or
+   * loop filter level) */
+  /* are enabled and when enabled the proabilities used to decode the per MB
+   * flags in MB_MODE_INFO */
+  /* Probability Tree used to code Segment number */
+  vp8_prob mb_segment_tree_probs[MB_FEATURE_TREE_PROBS];
+  /* Segment parameters */
+  signed char segment_feature_data[MB_LVL_MAX][MAX_MB_SEGMENTS];
 
-    signed char segment_feature_data[MB_LVL_MAX][MAX_MB_SEGMENTS];            /* Segment parameters */
+  /* mode_based Loop filter adjustment */
+  unsigned char mode_ref_lf_delta_enabled;
+  unsigned char mode_ref_lf_delta_update;
 
-    /* mode_based Loop filter adjustment */
-    unsigned char mode_ref_lf_delta_enabled;
-    unsigned char mode_ref_lf_delta_update;
+  /* Delta values have the range +/- MAX_LOOP_FILTER */
+  signed char
+      last_ref_lf_deltas[MAX_REF_LF_DELTAS];    /* 0 = Intra, Last, GF, ARF */
+  signed char ref_lf_deltas[MAX_REF_LF_DELTAS]; /* 0 = Intra, Last, GF, ARF */
+  /* 0 = BPRED, ZERO_MV, MV, SPLIT */
+  signed char last_mode_lf_deltas[MAX_MODE_LF_DELTAS];
+  signed char
+      mode_lf_deltas[MAX_MODE_LF_DELTAS]; /* 0 = BPRED, ZERO_MV, MV, SPLIT */
 
-    /* Delta values have the range +/- MAX_LOOP_FILTER */
-    signed char last_ref_lf_deltas[MAX_REF_LF_DELTAS];                /* 0 = Intra, Last, GF, ARF */
-    signed char ref_lf_deltas[MAX_REF_LF_DELTAS];                     /* 0 = Intra, Last, GF, ARF */
-    signed char last_mode_lf_deltas[MAX_MODE_LF_DELTAS];                      /* 0 = BPRED, ZERO_MV, MV, SPLIT */
-    signed char mode_lf_deltas[MAX_MODE_LF_DELTAS];                           /* 0 = BPRED, ZERO_MV, MV, SPLIT */
+  /* Distance of MB away from frame edges */
+  int mb_to_left_edge;
+  int mb_to_right_edge;
+  int mb_to_top_edge;
+  int mb_to_bottom_edge;
 
-    /* Distance of MB away from frame edges */
-    int mb_to_left_edge;
-    int mb_to_right_edge;
-    int mb_to_top_edge;
-    int mb_to_bottom_edge;
+  vp8_subpix_fn_t subpixel_predict;
+  vp8_subpix_fn_t subpixel_predict8x4;
+  vp8_subpix_fn_t subpixel_predict8x8;
+  vp8_subpix_fn_t subpixel_predict16x16;
 
+  void *current_bc;
 
+  int corrupted;
 
-    vp8_subpix_fn_t  subpixel_predict;
-    vp8_subpix_fn_t  subpixel_predict8x4;
-    vp8_subpix_fn_t  subpixel_predict8x8;
-    vp8_subpix_fn_t  subpixel_predict16x16;
-
-    void *current_bc;
-
-    int corrupted;
-
 #if ARCH_X86 || ARCH_X86_64
-    /* This is an intermediate buffer currently used in sub-pixel motion search
-     * to keep a copy of the reference area. This buffer can be used for other
-     * purpose.
-     */
-    DECLARE_ALIGNED(32, unsigned char, y_buf[22*32]);
+  /* This is an intermediate buffer currently used in sub-pixel motion search
+   * to keep a copy of the reference area. This buffer can be used for other
+   * purpose.
+   */
+  DECLARE_ALIGNED(32, unsigned char, y_buf[22 * 32]);
 #endif
 } MACROBLOCKD;
-
 
 extern void vp8_build_block_doffsets(MACROBLOCKD *x);
 extern void vp8_setup_block_dptrs(MACROBLOCKD *x);
--- a/vp8/common/coefupdateprobs.h
+++ b/vp8/common/coefupdateprobs.h
@@ -18,177 +18,177 @@
 /* Update probabilities for the nodes in the token entropy tree.
    Generated file included by entropy.c */
 
-const vp8_prob vp8_coef_update_probs [BLOCK_TYPES] [COEF_BANDS] [PREV_COEF_CONTEXTS] [ENTROPY_NODES] =
-{
-    {
-        {
-            {255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, },
-            {255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, },
-            {255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, },
-        },
-        {
-            {176, 246, 255, 255, 255, 255, 255, 255, 255, 255, 255, },
-            {223, 241, 252, 255, 255, 255, 255, 255, 255, 255, 255, },
-            {249, 253, 253, 255, 255, 255, 255, 255, 255, 255, 255, },
-        },
-        {
-            {255, 244, 252, 255, 255, 255, 255, 255, 255, 255, 255, },
-            {234, 254, 254, 255, 255, 255, 255, 255, 255, 255, 255, },
-            {253, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, },
-        },
-        {
-            {255, 246, 254, 255, 255, 255, 255, 255, 255, 255, 255, },
-            {239, 253, 254, 255, 255, 255, 255, 255, 255, 255, 255, },
-            {254, 255, 254, 255, 255, 255, 255, 255, 255, 255, 255, },
-        },
-        {
-            {255, 248, 254, 255, 255, 255, 255, 255, 255, 255, 255, },
-            {251, 255, 254, 255, 255, 255, 255, 255, 255, 255, 255, },
-            {255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, },
-        },
-        {
-            {255, 253, 254, 255, 255, 255, 255, 255, 255, 255, 255, },
-            {251, 254, 254, 255, 255, 255, 255, 255, 255, 255, 255, },
-            {254, 255, 254, 255, 255, 255, 255, 255, 255, 255, 255, },
-        },
-        {
-            {255, 254, 253, 255, 254, 255, 255, 255, 255, 255, 255, },
-            {250, 255, 254, 255, 254, 255, 255, 255, 255, 255, 255, },
-            {254, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, },
-        },
-        {
-            {255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, },
-            {255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, },
-            {255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, },
-        },
-    },
-    {
-        {
-            {217, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, },
-            {225, 252, 241, 253, 255, 255, 254, 255, 255, 255, 255, },
-            {234, 250, 241, 250, 253, 255, 253, 254, 255, 255, 255, },
-        },
-        {
-            {255, 254, 255, 255, 255, 255, 255, 255, 255, 255, 255, },
-            {223, 254, 254, 255, 255, 255, 255, 255, 255, 255, 255, },
-            {238, 253, 254, 254, 255, 255, 255, 255, 255, 255, 255, },
-        },
-        {
-            {255, 248, 254, 255, 255, 255, 255, 255, 255, 255, 255, },
-            {249, 254, 255, 255, 255, 255, 255, 255, 255, 255, 255, },
-            {255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, },
-        },
-        {
-            {255, 253, 255, 255, 255, 255, 255, 255, 255, 255, 255, },
-            {247, 254, 255, 255, 255, 255, 255, 255, 255, 255, 255, },
-            {255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, },
-        },
-        {
-            {255, 253, 254, 255, 255, 255, 255, 255, 255, 255, 255, },
-            {252, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, },
-            {255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, },
-        },
-        {
-            {255, 254, 254, 255, 255, 255, 255, 255, 255, 255, 255, },
-            {253, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, },
-            {255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, },
-        },
-        {
-            {255, 254, 253, 255, 255, 255, 255, 255, 255, 255, 255, },
-            {250, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, },
-            {254, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, },
-        },
-        {
-            {255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, },
-            {255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, },
-            {255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, },
-        },
-    },
-    {
-        {
-            {186, 251, 250, 255, 255, 255, 255, 255, 255, 255, 255, },
-            {234, 251, 244, 254, 255, 255, 255, 255, 255, 255, 255, },
-            {251, 251, 243, 253, 254, 255, 254, 255, 255, 255, 255, },
-        },
-        {
-            {255, 253, 254, 255, 255, 255, 255, 255, 255, 255, 255, },
-            {236, 253, 254, 255, 255, 255, 255, 255, 255, 255, 255, },
-            {251, 253, 253, 254, 254, 255, 255, 255, 255, 255, 255, },
-        },
-        {
-            {255, 254, 254, 255, 255, 255, 255, 255, 255, 255, 255, },
-            {254, 254, 254, 255, 255, 255, 255, 255, 255, 255, 255, },
-            {255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, },
-        },
-        {
-            {255, 254, 255, 255, 255, 255, 255, 255, 255, 255, 255, },
-            {254, 254, 255, 255, 255, 255, 255, 255, 255, 255, 255, },
-            {254, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, },
-        },
-        {
-            {255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, },
-            {254, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, },
-            {255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, },
-        },
-        {
-            {255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, },
-            {255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, },
-            {255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, },
-        },
-        {
-            {255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, },
-            {255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, },
-            {255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, },
-        },
-        {
-            {255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, },
-            {255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, },
-            {255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, },
-        },
-    },
-    {
-        {
-            {248, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, },
-            {250, 254, 252, 254, 255, 255, 255, 255, 255, 255, 255, },
-            {248, 254, 249, 253, 255, 255, 255, 255, 255, 255, 255, },
-        },
-        {
-            {255, 253, 253, 255, 255, 255, 255, 255, 255, 255, 255, },
-            {246, 253, 253, 255, 255, 255, 255, 255, 255, 255, 255, },
-            {252, 254, 251, 254, 254, 255, 255, 255, 255, 255, 255, },
-        },
-        {
-            {255, 254, 252, 255, 255, 255, 255, 255, 255, 255, 255, },
-            {248, 254, 253, 255, 255, 255, 255, 255, 255, 255, 255, },
-            {253, 255, 254, 254, 255, 255, 255, 255, 255, 255, 255, },
-        },
-        {
-            {255, 251, 254, 255, 255, 255, 255, 255, 255, 255, 255, },
-            {245, 251, 254, 255, 255, 255, 255, 255, 255, 255, 255, },
-            {253, 253, 254, 255, 255, 255, 255, 255, 255, 255, 255, },
-        },
-        {
-            {255, 251, 253, 255, 255, 255, 255, 255, 255, 255, 255, },
-            {252, 253, 254, 255, 255, 255, 255, 255, 255, 255, 255, },
-            {255, 254, 255, 255, 255, 255, 255, 255, 255, 255, 255, },
-        },
-        {
-            {255, 252, 255, 255, 255, 255, 255, 255, 255, 255, 255, },
-            {249, 255, 254, 255, 255, 255, 255, 255, 255, 255, 255, },
-            {255, 255, 254, 255, 255, 255, 255, 255, 255, 255, 255, },
-        },
-        {
-            {255, 255, 253, 255, 255, 255, 255, 255, 255, 255, 255, },
-            {250, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, },
-            {255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, },
-        },
-        {
-            {255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, },
-            {254, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, },
-            {255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, },
-        },
-    },
-};
+const vp8_prob vp8_coef_update_probs
+    [BLOCK_TYPES][COEF_BANDS][PREV_COEF_CONTEXTS][ENTROPY_NODES] = {
+      {
+          {
+              { 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255 },
+              { 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255 },
+              { 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255 },
+          },
+          {
+              { 176, 246, 255, 255, 255, 255, 255, 255, 255, 255, 255 },
+              { 223, 241, 252, 255, 255, 255, 255, 255, 255, 255, 255 },
+              { 249, 253, 253, 255, 255, 255, 255, 255, 255, 255, 255 },
+          },
+          {
+              { 255, 244, 252, 255, 255, 255, 255, 255, 255, 255, 255 },
+              { 234, 254, 254, 255, 255, 255, 255, 255, 255, 255, 255 },
+              { 253, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255 },
+          },
+          {
+              { 255, 246, 254, 255, 255, 255, 255, 255, 255, 255, 255 },
+              { 239, 253, 254, 255, 255, 255, 255, 255, 255, 255, 255 },
+              { 254, 255, 254, 255, 255, 255, 255, 255, 255, 255, 255 },
+          },
+          {
+              { 255, 248, 254, 255, 255, 255, 255, 255, 255, 255, 255 },
+              { 251, 255, 254, 255, 255, 255, 255, 255, 255, 255, 255 },
+              { 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255 },
+          },
+          {
+              { 255, 253, 254, 255, 255, 255, 255, 255, 255, 255, 255 },
+              { 251, 254, 254, 255, 255, 255, 255, 255, 255, 255, 255 },
+              { 254, 255, 254, 255, 255, 255, 255, 255, 255, 255, 255 },
+          },
+          {
+              { 255, 254, 253, 255, 254, 255, 255, 255, 255, 255, 255 },
+              { 250, 255, 254, 255, 254, 255, 255, 255, 255, 255, 255 },
+              { 254, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255 },
+          },
+          {
+              { 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255 },
+              { 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255 },
+              { 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255 },
+          },
+      },
+      {
+          {
+              { 217, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255 },
+              { 225, 252, 241, 253, 255, 255, 254, 255, 255, 255, 255 },
+              { 234, 250, 241, 250, 253, 255, 253, 254, 255, 255, 255 },
+          },
+          {
+              { 255, 254, 255, 255, 255, 255, 255, 255, 255, 255, 255 },
+              { 223, 254, 254, 255, 255, 255, 255, 255, 255, 255, 255 },
+              { 238, 253, 254, 254, 255, 255, 255, 255, 255, 255, 255 },
+          },
+          {
+              { 255, 248, 254, 255, 255, 255, 255, 255, 255, 255, 255 },
+              { 249, 254, 255, 255, 255, 255, 255, 255, 255, 255, 255 },
+              { 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255 },
+          },
+          {
+              { 255, 253, 255, 255, 255, 255, 255, 255, 255, 255, 255 },
+              { 247, 254, 255, 255, 255, 255, 255, 255, 255, 255, 255 },
+              { 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255 },
+          },
+          {
+              { 255, 253, 254, 255, 255, 255, 255, 255, 255, 255, 255 },
+              { 252, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255 },
+              { 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255 },
+          },
+          {
+              { 255, 254, 254, 255, 255, 255, 255, 255, 255, 255, 255 },
+              { 253, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255 },
+              { 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255 },
+          },
+          {
+              { 255, 254, 253, 255, 255, 255, 255, 255, 255, 255, 255 },
+              { 250, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255 },
+              { 254, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255 },
+          },
+          {
+              { 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255 },
+              { 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255 },
+              { 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255 },
+          },
+      },
+      {
+          {
+              { 186, 251, 250, 255, 255, 255, 255, 255, 255, 255, 255 },
+              { 234, 251, 244, 254, 255, 255, 255, 255, 255, 255, 255 },
+              { 251, 251, 243, 253, 254, 255, 254, 255, 255, 255, 255 },
+          },
+          {
+              { 255, 253, 254, 255, 255, 255, 255, 255, 255, 255, 255 },
+              { 236, 253, 254, 255, 255, 255, 255, 255, 255, 255, 255 },
+              { 251, 253, 253, 254, 254, 255, 255, 255, 255, 255, 255 },
+          },
+          {
+              { 255, 254, 254, 255, 255, 255, 255, 255, 255, 255, 255 },
+              { 254, 254, 254, 255, 255, 255, 255, 255, 255, 255, 255 },
+              { 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255 },
+          },
+          {
+              { 255, 254, 255, 255, 255, 255, 255, 255, 255, 255, 255 },
+              { 254, 254, 255, 255, 255, 255, 255, 255, 255, 255, 255 },
+              { 254, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255 },
+          },
+          {
+              { 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255 },
+              { 254, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255 },
+              { 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255 },
+          },
+          {
+              { 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255 },
+              { 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255 },
+              { 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255 },
+          },
+          {
+              { 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255 },
+              { 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255 },
+              { 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255 },
+          },
+          {
+              { 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255 },
+              { 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255 },
+              { 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255 },
+          },
+      },
+      {
+          {
+              { 248, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255 },
+              { 250, 254, 252, 254, 255, 255, 255, 255, 255, 255, 255 },
+              { 248, 254, 249, 253, 255, 255, 255, 255, 255, 255, 255 },
+          },
+          {
+              { 255, 253, 253, 255, 255, 255, 255, 255, 255, 255, 255 },
+              { 246, 253, 253, 255, 255, 255, 255, 255, 255, 255, 255 },
+              { 252, 254, 251, 254, 254, 255, 255, 255, 255, 255, 255 },
+          },
+          {
+              { 255, 254, 252, 255, 255, 255, 255, 255, 255, 255, 255 },
+              { 248, 254, 253, 255, 255, 255, 255, 255, 255, 255, 255 },
+              { 253, 255, 254, 254, 255, 255, 255, 255, 255, 255, 255 },
+          },
+          {
+              { 255, 251, 254, 255, 255, 255, 255, 255, 255, 255, 255 },
+              { 245, 251, 254, 255, 255, 255, 255, 255, 255, 255, 255 },
+              { 253, 253, 254, 255, 255, 255, 255, 255, 255, 255, 255 },
+          },
+          {
+              { 255, 251, 253, 255, 255, 255, 255, 255, 255, 255, 255 },
+              { 252, 253, 254, 255, 255, 255, 255, 255, 255, 255, 255 },
+              { 255, 254, 255, 255, 255, 255, 255, 255, 255, 255, 255 },
+          },
+          {
+              { 255, 252, 255, 255, 255, 255, 255, 255, 255, 255, 255 },
+              { 249, 255, 254, 255, 255, 255, 255, 255, 255, 255, 255 },
+              { 255, 255, 254, 255, 255, 255, 255, 255, 255, 255, 255 },
+          },
+          {
+              { 255, 255, 253, 255, 255, 255, 255, 255, 255, 255, 255 },
+              { 250, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255 },
+              { 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255 },
+          },
+          {
+              { 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255 },
+              { 254, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255 },
+              { 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255 },
+          },
+      },
+    };
 
 #ifdef __cplusplus
 }  // extern "C"
--- a/vp8/common/common.h
+++ b/vp8/common/common.h
@@ -8,7 +8,6 @@
  *  be found in the AUTHORS file in the root of the source tree.
  */
 
-
 #ifndef VP8_COMMON_COMMON_H_
 #define VP8_COMMON_COMMON_H_
 
@@ -24,22 +23,23 @@
 
 /* Only need this for fixed-size arrays, for structs just assign. */
 
-#define vp8_copy( Dest, Src) { \
-        assert( sizeof( Dest) == sizeof( Src)); \
-        memcpy( Dest, Src, sizeof( Src)); \
-    }
+#define vp8_copy(Dest, Src)              \
+  {                                      \
+    assert(sizeof(Dest) == sizeof(Src)); \
+    memcpy(Dest, Src, sizeof(Src));      \
+  }
 
 /* Use this for variably-sized arrays. */
 
-#define vp8_copy_array( Dest, Src, N) { \
-        assert( sizeof( *Dest) == sizeof( *Src)); \
-        memcpy( Dest, Src, N * sizeof( *Src)); \
-    }
+#define vp8_copy_array(Dest, Src, N)       \
+  {                                        \
+    assert(sizeof(*Dest) == sizeof(*Src)); \
+    memcpy(Dest, Src, N * sizeof(*Src));   \
+  }
 
-#define vp8_zero( Dest)  memset( &Dest, 0, sizeof( Dest));
+#define vp8_zero(Dest) memset(&Dest, 0, sizeof(Dest));
 
-#define vp8_zero_array( Dest, N)  memset( Dest, 0, N * sizeof( *Dest));
-
+#define vp8_zero_array(Dest, N) memset(Dest, 0, N * sizeof(*Dest));
 
 #ifdef __cplusplus
 }  // extern "C"
--- a/vp8/common/context.c
+++ b/vp8/common/context.c
@@ -8,7 +8,6 @@
  *  be found in the AUTHORS file in the root of the source tree.
  */
 
-
 #include "entropy.h"
 
 /* *** GENERATED FILE: DO NOT EDIT *** */
--- a/vp8/common/copy_c.c
+++ b/vp8/common/copy_c.c
@@ -8,7 +8,6 @@
  *  be found in the AUTHORS file in the root of the source tree.
  */
 
-
 #include <string.h>
 
 #include "./vp8_rtcd.h"
@@ -16,17 +15,13 @@
 
 /* Copy 2 macroblocks to a buffer */
 void vp8_copy32xn_c(const unsigned char *src_ptr, int src_stride,
-                    unsigned char *dst_ptr, int dst_stride,
-                    int height)
-{
-    int r;
+                    unsigned char *dst_ptr, int dst_stride, int height) {
+  int r;
 
-    for (r = 0; r < height; r++)
-    {
-        memcpy(dst_ptr, src_ptr, 32);
+  for (r = 0; r < height; r++) {
+    memcpy(dst_ptr, src_ptr, 32);
 
-        src_ptr += src_stride;
-        dst_ptr += dst_stride;
-
-    }
+    src_ptr += src_stride;
+    dst_ptr += dst_stride;
+  }
 }
--- a/vp8/common/debugmodes.c
+++ b/vp8/common/debugmodes.c
@@ -8,148 +8,128 @@
  *  be found in the AUTHORS file in the root of the source tree.
  */
 
-
 #include <stdio.h>
 #include "blockd.h"
 
+void vp8_print_modes_and_motion_vectors(MODE_INFO *mi, int rows, int cols,
+                                        int frame) {
+  int mb_row;
+  int mb_col;
+  int mb_index = 0;
+  FILE *mvs = fopen("mvs.stt", "a");
 
-void vp8_print_modes_and_motion_vectors(MODE_INFO *mi, int rows, int cols, int frame)
-{
+  /* print out the macroblock Y modes */
+  mb_index = 0;
+  fprintf(mvs, "Mb Modes for Frame %d\n", frame);
 
-    int mb_row;
-    int mb_col;
-    int mb_index = 0;
-    FILE *mvs = fopen("mvs.stt", "a");
+  for (mb_row = 0; mb_row < rows; mb_row++) {
+    for (mb_col = 0; mb_col < cols; mb_col++) {
+      fprintf(mvs, "%2d ", mi[mb_index].mbmi.mode);
 
-    /* print out the macroblock Y modes */
-    mb_index = 0;
-    fprintf(mvs, "Mb Modes for Frame %d\n", frame);
-
-    for (mb_row = 0; mb_row < rows; mb_row++)
-    {
-        for (mb_col = 0; mb_col < cols; mb_col++)
-        {
-
-            fprintf(mvs, "%2d ", mi[mb_index].mbmi.mode);
-
-            mb_index++;
-        }
-
-        fprintf(mvs, "\n");
-        mb_index++;
+      mb_index++;
     }
 
     fprintf(mvs, "\n");
+    mb_index++;
+  }
 
-    mb_index = 0;
-    fprintf(mvs, "Mb mv ref for Frame %d\n", frame);
+  fprintf(mvs, "\n");
 
-    for (mb_row = 0; mb_row < rows; mb_row++)
-    {
-        for (mb_col = 0; mb_col < cols; mb_col++)
-        {
+  mb_index = 0;
+  fprintf(mvs, "Mb mv ref for Frame %d\n", frame);
 
-            fprintf(mvs, "%2d ", mi[mb_index].mbmi.ref_frame);
+  for (mb_row = 0; mb_row < rows; mb_row++) {
+    for (mb_col = 0; mb_col < cols; mb_col++) {
+      fprintf(mvs, "%2d ", mi[mb_index].mbmi.ref_frame);
 
-            mb_index++;
-        }
-
-        fprintf(mvs, "\n");
-        mb_index++;
+      mb_index++;
     }
 
     fprintf(mvs, "\n");
+    mb_index++;
+  }
 
-    /* print out the macroblock UV modes */
-    mb_index = 0;
-    fprintf(mvs, "UV Modes for Frame %d\n", frame);
+  fprintf(mvs, "\n");
 
-    for (mb_row = 0; mb_row < rows; mb_row++)
-    {
-        for (mb_col = 0; mb_col < cols; mb_col++)
-        {
+  /* print out the macroblock UV modes */
+  mb_index = 0;
+  fprintf(mvs, "UV Modes for Frame %d\n", frame);
 
-            fprintf(mvs, "%2d ", mi[mb_index].mbmi.uv_mode);
+  for (mb_row = 0; mb_row < rows; mb_row++) {
+    for (mb_col = 0; mb_col < cols; mb_col++) {
+      fprintf(mvs, "%2d ", mi[mb_index].mbmi.uv_mode);
 
-            mb_index++;
-        }
-
-        mb_index++;
-        fprintf(mvs, "\n");
+      mb_index++;
     }
 
+    mb_index++;
     fprintf(mvs, "\n");
+  }
 
-    /* print out the block modes */
-    fprintf(mvs, "Mbs for Frame %d\n", frame);
-    {
-        int b_row;
+  fprintf(mvs, "\n");
 
-        for (b_row = 0; b_row < 4 * rows; b_row++)
-        {
-            int b_col;
-            int bindex;
+  /* print out the block modes */
+  fprintf(mvs, "Mbs for Frame %d\n", frame);
+  {
+    int b_row;
 
-            for (b_col = 0; b_col < 4 * cols; b_col++)
-            {
-                mb_index = (b_row >> 2) * (cols + 1) + (b_col >> 2);
-                bindex = (b_row & 3) * 4 + (b_col & 3);
+    for (b_row = 0; b_row < 4 * rows; b_row++) {
+      int b_col;
+      int bindex;
 
-                if (mi[mb_index].mbmi.mode == B_PRED)
-                    fprintf(mvs, "%2d ", mi[mb_index].bmi[bindex].as_mode);
-                else
-                    fprintf(mvs, "xx ");
+      for (b_col = 0; b_col < 4 * cols; b_col++) {
+        mb_index = (b_row >> 2) * (cols + 1) + (b_col >> 2);
+        bindex = (b_row & 3) * 4 + (b_col & 3);
 
-            }
+        if (mi[mb_index].mbmi.mode == B_PRED)
+          fprintf(mvs, "%2d ", mi[mb_index].bmi[bindex].as_mode);
+        else
+          fprintf(mvs, "xx ");
+      }
 
-            fprintf(mvs, "\n");
-        }
+      fprintf(mvs, "\n");
     }
-    fprintf(mvs, "\n");
+  }
+  fprintf(mvs, "\n");
 
-    /* print out the macroblock mvs */
-    mb_index = 0;
-    fprintf(mvs, "MVs for Frame %d\n", frame);
+  /* print out the macroblock mvs */
+  mb_index = 0;
+  fprintf(mvs, "MVs for Frame %d\n", frame);
 
-    for (mb_row = 0; mb_row < rows; mb_row++)
-    {
-        for (mb_col = 0; mb_col < cols; mb_col++)
-        {
-            fprintf(mvs, "%5d:%-5d", mi[mb_index].mbmi.mv.as_mv.row / 2, mi[mb_index].mbmi.mv.as_mv.col / 2);
+  for (mb_row = 0; mb_row < rows; mb_row++) {
+    for (mb_col = 0; mb_col < cols; mb_col++) {
+      fprintf(mvs, "%5d:%-5d", mi[mb_index].mbmi.mv.as_mv.row / 2,
+              mi[mb_index].mbmi.mv.as_mv.col / 2);
 
-            mb_index++;
-        }
-
-        mb_index++;
-        fprintf(mvs, "\n");
+      mb_index++;
     }
 
+    mb_index++;
     fprintf(mvs, "\n");
+  }
 
+  fprintf(mvs, "\n");
 
-    /* print out the block modes */
-    fprintf(mvs, "MVs for Frame %d\n", frame);
-    {
-        int b_row;
+  /* print out the block modes */
+  fprintf(mvs, "MVs for Frame %d\n", frame);
+  {
+    int b_row;
 
-        for (b_row = 0; b_row < 4 * rows; b_row++)
-        {
-            int b_col;
-            int bindex;
+    for (b_row = 0; b_row < 4 * rows; b_row++) {
+      int b_col;
+      int bindex;
 
-            for (b_col = 0; b_col < 4 * cols; b_col++)
-            {
-                mb_index = (b_row >> 2) * (cols + 1) + (b_col >> 2);
-                bindex = (b_row & 3) * 4 + (b_col & 3);
-                fprintf(mvs, "%3d:%-3d ", mi[mb_index].bmi[bindex].mv.as_mv.row, mi[mb_index].bmi[bindex].mv.as_mv.col);
+      for (b_col = 0; b_col < 4 * cols; b_col++) {
+        mb_index = (b_row >> 2) * (cols + 1) + (b_col >> 2);
+        bindex = (b_row & 3) * 4 + (b_col & 3);
+        fprintf(mvs, "%3d:%-3d ", mi[mb_index].bmi[bindex].mv.as_mv.row,
+                mi[mb_index].bmi[bindex].mv.as_mv.col);
+      }
 
-            }
-
-            fprintf(mvs, "\n");
-        }
+      fprintf(mvs, "\n");
     }
-    fprintf(mvs, "\n");
+  }
+  fprintf(mvs, "\n");
 
-
-    fclose(mvs);
+  fclose(mvs);
 }
--- a/vp8/common/default_coef_probs.h
+++ b/vp8/common/default_coef_probs.h
@@ -17,181 +17,141 @@
 
 /*Generated file, included by entropy.c*/
 
-
-static const vp8_prob default_coef_probs [BLOCK_TYPES]
-                                         [COEF_BANDS]
-                                         [PREV_COEF_CONTEXTS]
-                                         [ENTROPY_NODES] =
-{
-    { /* Block Type ( 0 ) */
+static const vp8_prob default_coef_probs
+    [BLOCK_TYPES][COEF_BANDS][PREV_COEF_CONTEXTS][ENTROPY_NODES] = {
+      {   /* Block Type ( 0 ) */
         { /* Coeff Band ( 0 )*/
-            { 128, 128, 128, 128, 128, 128, 128, 128, 128, 128, 128 },
-            { 128, 128, 128, 128, 128, 128, 128, 128, 128, 128, 128 },
-            { 128, 128, 128, 128, 128, 128, 128, 128, 128, 128, 128 }
-        },
+          { 128, 128, 128, 128, 128, 128, 128, 128, 128, 128, 128 },
+          { 128, 128, 128, 128, 128, 128, 128, 128, 128, 128, 128 },
+          { 128, 128, 128, 128, 128, 128, 128, 128, 128, 128, 128 } },
         { /* Coeff Band ( 1 )*/
-            { 253, 136, 254, 255, 228, 219, 128, 128, 128, 128, 128 },
-            { 189, 129, 242, 255, 227, 213, 255, 219, 128, 128, 128 },
-            { 106, 126, 227, 252, 214, 209, 255, 255, 128, 128, 128 }
-        },
+          { 253, 136, 254, 255, 228, 219, 128, 128, 128, 128, 128 },
+          { 189, 129, 242, 255, 227, 213, 255, 219, 128, 128, 128 },
+          { 106, 126, 227, 252, 214, 209, 255, 255, 128, 128, 128 } },
         { /* Coeff Band ( 2 )*/
-            {   1,  98, 248, 255, 236, 226, 255, 255, 128, 128, 128 },
-            { 181, 133, 238, 254, 221, 234, 255, 154, 128, 128, 128 },
-            {  78, 134, 202, 247, 198, 180, 255, 219, 128, 128, 128 }
-        },
+          { 1, 98, 248, 255, 236, 226, 255, 255, 128, 128, 128 },
+          { 181, 133, 238, 254, 221, 234, 255, 154, 128, 128, 128 },
+          { 78, 134, 202, 247, 198, 180, 255, 219, 128, 128, 128 } },
         { /* Coeff Band ( 3 )*/
-            {   1, 185, 249, 255, 243, 255, 128, 128, 128, 128, 128 },
-            { 184, 150, 247, 255, 236, 224, 128, 128, 128, 128, 128 },
-            {  77, 110, 216, 255, 236, 230, 128, 128, 128, 128, 128 }
-        },
+          { 1, 185, 249, 255, 243, 255, 128, 128, 128, 128, 128 },
+          { 184, 150, 247, 255, 236, 224, 128, 128, 128, 128, 128 },
+          { 77, 110, 216, 255, 236, 230, 128, 128, 128, 128, 128 } },
         { /* Coeff Band ( 4 )*/
-            {   1, 101, 251, 255, 241, 255, 128, 128, 128, 128, 128 },
-            { 170, 139, 241, 252, 236, 209, 255, 255, 128, 128, 128 },
-            {  37, 116, 196, 243, 228, 255, 255, 255, 128, 128, 128 }
-        },
+          { 1, 101, 251, 255, 241, 255, 128, 128, 128, 128, 128 },
+          { 170, 139, 241, 252, 236, 209, 255, 255, 128, 128, 128 },
+          { 37, 116, 196, 243, 228, 255, 255, 255, 128, 128, 128 } },
         { /* Coeff Band ( 5 )*/
-            {   1, 204, 254, 255, 245, 255, 128, 128, 128, 128, 128 },
-            { 207, 160, 250, 255, 238, 128, 128, 128, 128, 128, 128 },
-            { 102, 103, 231, 255, 211, 171, 128, 128, 128, 128, 128 }
-        },
+          { 1, 204, 254, 255, 245, 255, 128, 128, 128, 128, 128 },
+          { 207, 160, 250, 255, 238, 128, 128, 128, 128, 128, 128 },
+          { 102, 103, 231, 255, 211, 171, 128, 128, 128, 128, 128 } },
         { /* Coeff Band ( 6 )*/
-            {   1, 152, 252, 255, 240, 255, 128, 128, 128, 128, 128 },
-            { 177, 135, 243, 255, 234, 225, 128, 128, 128, 128, 128 },
-            {  80, 129, 211, 255, 194, 224, 128, 128, 128, 128, 128 }
-        },
+          { 1, 152, 252, 255, 240, 255, 128, 128, 128, 128, 128 },
+          { 177, 135, 243, 255, 234, 225, 128, 128, 128, 128, 128 },
+          { 80, 129, 211, 255, 194, 224, 128, 128, 128, 128, 128 } },
         { /* Coeff Band ( 7 )*/
-            {   1,   1, 255, 128, 128, 128, 128, 128, 128, 128, 128 },
-            { 246,   1, 255, 128, 128, 128, 128, 128, 128, 128, 128 },
-            { 255, 128, 128, 128, 128, 128, 128, 128, 128, 128, 128 }
-        }
-    },
-    { /* Block Type ( 1 ) */
+          { 1, 1, 255, 128, 128, 128, 128, 128, 128, 128, 128 },
+          { 246, 1, 255, 128, 128, 128, 128, 128, 128, 128, 128 },
+          { 255, 128, 128, 128, 128, 128, 128, 128, 128, 128, 128 } } },
+      {   /* Block Type ( 1 ) */
         { /* Coeff Band ( 0 )*/
-            { 198,  35, 237, 223, 193, 187, 162, 160, 145, 155,  62 },
-            { 131,  45, 198, 221, 172, 176, 220, 157, 252, 221,   1 },
-            {  68,  47, 146, 208, 149, 167, 221, 162, 255, 223, 128 }
-        },
+          { 198, 35, 237, 223, 193, 187, 162, 160, 145, 155, 62 },
+          { 131, 45, 198, 221, 172, 176, 220, 157, 252, 221, 1 },
+          { 68, 47, 146, 208, 149, 167, 221, 162, 255, 223, 128 } },
         { /* Coeff Band ( 1 )*/
-            {   1, 149, 241, 255, 221, 224, 255, 255, 128, 128, 128 },
-            { 184, 141, 234, 253, 222, 220, 255, 199, 128, 128, 128 },
-            {  81,  99, 181, 242, 176, 190, 249, 202, 255, 255, 128 }
-        },
+          { 1, 149, 241, 255, 221, 224, 255, 255, 128, 128, 128 },
+          { 184, 141, 234, 253, 222, 220, 255, 199, 128, 128, 128 },
+          { 81, 99, 181, 242, 176, 190, 249, 202, 255, 255, 128 } },
         { /* Coeff Band ( 2 )*/
-            {   1, 129, 232, 253, 214, 197, 242, 196, 255, 255, 128 },
-            {  99, 121, 210, 250, 201, 198, 255, 202, 128, 128, 128 },
-            {  23,  91, 163, 242, 170, 187, 247, 210, 255, 255, 128 }
-        },
+          { 1, 129, 232, 253, 214, 197, 242, 196, 255, 255, 128 },
+          { 99, 121, 210, 250, 201, 198, 255, 202, 128, 128, 128 },
+          { 23, 91, 163, 242, 170, 187, 247, 210, 255, 255, 128 } },
         { /* Coeff Band ( 3 )*/
-            {   1, 200, 246, 255, 234, 255, 128, 128, 128, 128, 128 },
-            { 109, 178, 241, 255, 231, 245, 255, 255, 128, 128, 128 },
-            {  44, 130, 201, 253, 205, 192, 255, 255, 128, 128, 128 }
-        },
+          { 1, 200, 246, 255, 234, 255, 128, 128, 128, 128, 128 },
+          { 109, 178, 241, 255, 231, 245, 255, 255, 128, 128, 128 },
+          { 44, 130, 201, 253, 205, 192, 255, 255, 128, 128, 128 } },
         { /* Coeff Band ( 4 )*/
-            {   1, 132, 239, 251, 219, 209, 255, 165, 128, 128, 128 },
-            {  94, 136, 225, 251, 218, 190, 255, 255, 128, 128, 128 },
-            {  22, 100, 174, 245, 186, 161, 255, 199, 128, 128, 128 }
-        },
+          { 1, 132, 239, 251, 219, 209, 255, 165, 128, 128, 128 },
+          { 94, 136, 225, 251, 218, 190, 255, 255, 128, 128, 128 },
+          { 22, 100, 174, 245, 186, 161, 255, 199, 128, 128, 128 } },
         { /* Coeff Band ( 5 )*/
-            {   1, 182, 249, 255, 232, 235, 128, 128, 128, 128, 128 },
-            { 124, 143, 241, 255, 227, 234, 128, 128, 128, 128, 128 },
-            {  35,  77, 181, 251, 193, 211, 255, 205, 128, 128, 128 }
-        },
+          { 1, 182, 249, 255, 232, 235, 128, 128, 128, 128, 128 },
+          { 124, 143, 241, 255, 227, 234, 128, 128, 128, 128, 128 },
+          { 35, 77, 181, 251, 193, 211, 255, 205, 128, 128, 128 } },
         { /* Coeff Band ( 6 )*/
-            {   1, 157, 247, 255, 236, 231, 255, 255, 128, 128, 128 },
-            { 121, 141, 235, 255, 225, 227, 255, 255, 128, 128, 128 },
-            {  45,  99, 188, 251, 195, 217, 255, 224, 128, 128, 128 }
-        },
+          { 1, 157, 247, 255, 236, 231, 255, 255, 128, 128, 128 },
+          { 121, 141, 235, 255, 225, 227, 255, 255, 128, 128, 128 },
+          { 45, 99, 188, 251, 195, 217, 255, 224, 128, 128, 128 } },
         { /* Coeff Band ( 7 )*/
-            {   1,   1, 251, 255, 213, 255, 128, 128, 128, 128, 128 },
-            { 203,   1, 248, 255, 255, 128, 128, 128, 128, 128, 128 },
-            { 137,   1, 177, 255, 224, 255, 128, 128, 128, 128, 128 }
-        }
-    },
-    { /* Block Type ( 2 ) */
+          { 1, 1, 251, 255, 213, 255, 128, 128, 128, 128, 128 },
+          { 203, 1, 248, 255, 255, 128, 128, 128, 128, 128, 128 },
+          { 137, 1, 177, 255, 224, 255, 128, 128, 128, 128, 128 } } },
+      {   /* Block Type ( 2 ) */
         { /* Coeff Band ( 0 )*/
-            { 253,   9, 248, 251, 207, 208, 255, 192, 128, 128, 128 },
-            { 175,  13, 224, 243, 193, 185, 249, 198, 255, 255, 128 },
-            {  73,  17, 171, 221, 161, 179, 236, 167, 255, 234, 128 }
-        },
+          { 253, 9, 248, 251, 207, 208, 255, 192, 128, 128, 128 },
+          { 175, 13, 224, 243, 193, 185, 249, 198, 255, 255, 128 },
+          { 73, 17, 171, 221, 161, 179, 236, 167, 255, 234, 128 } },
         { /* Coeff Band ( 1 )*/
-            {   1,  95, 247, 253, 212, 183, 255, 255, 128, 128, 128 },
-            { 239,  90, 244, 250, 211, 209, 255, 255, 128, 128, 128 },
-            { 155,  77, 195, 248, 188, 195, 255, 255, 128, 128, 128 }
-        },
+          { 1, 95, 247, 253, 212, 183, 255, 255, 128, 128, 128 },
+          { 239, 90, 244, 250, 211, 209, 255, 255, 128, 128, 128 },
+          { 155, 77, 195, 248, 188, 195, 255, 255, 128, 128, 128 } },
         { /* Coeff Band ( 2 )*/
-            {   1,  24, 239, 251, 218, 219, 255, 205, 128, 128, 128 },
-            { 201,  51, 219, 255, 196, 186, 128, 128, 128, 128, 128 },
-            {  69,  46, 190, 239, 201, 218, 255, 228, 128, 128, 128 }
-        },
+          { 1, 24, 239, 251, 218, 219, 255, 205, 128, 128, 128 },
+          { 201, 51, 219, 255, 196, 186, 128, 128, 128, 128, 128 },
+          { 69, 46, 190, 239, 201, 218, 255, 228, 128, 128, 128 } },
         { /* Coeff Band ( 3 )*/
-            {   1, 191, 251, 255, 255, 128, 128, 128, 128, 128, 128 },
-            { 223, 165, 249, 255, 213, 255, 128, 128, 128, 128, 128 },
-            { 141, 124, 248, 255, 255, 128, 128, 128, 128, 128, 128 }
-        },
+          { 1, 191, 251, 255, 255, 128, 128, 128, 128, 128, 128 },
+          { 223, 165, 249, 255, 213, 255, 128, 128, 128, 128, 128 },
+          { 141, 124, 248, 255, 255, 128, 128, 128, 128, 128, 128 } },
         { /* Coeff Band ( 4 )*/
-            {   1,  16, 248, 255, 255, 128, 128, 128, 128, 128, 128 },
-            { 190,  36, 230, 255, 236, 255, 128, 128, 128, 128, 128 },
-            { 149,   1, 255, 128, 128, 128, 128, 128, 128, 128, 128 }
-        },
+          { 1, 16, 248, 255, 255, 128, 128, 128, 128, 128, 128 },
+          { 190, 36, 230, 255, 236, 255, 128, 128, 128, 128, 128 },
+          { 149, 1, 255, 128, 128, 128, 128, 128, 128, 128, 128 } },
         { /* Coeff Band ( 5 )*/
-            {   1, 226, 255, 128, 128, 128, 128, 128, 128, 128, 128 },
-            { 247, 192, 255, 128, 128, 128, 128, 128, 128, 128, 128 },
-            { 240, 128, 255, 128, 128, 128, 128, 128, 128, 128, 128 }
-        },
+          { 1, 226, 255, 128, 128, 128, 128, 128, 128, 128, 128 },
+          { 247, 192, 255, 128, 128, 128, 128, 128, 128, 128, 128 },
+          { 240, 128, 255, 128, 128, 128, 128, 128, 128, 128, 128 } },
         { /* Coeff Band ( 6 )*/
-            {   1, 134, 252, 255, 255, 128, 128, 128, 128, 128, 128 },
-            { 213,  62, 250, 255, 255, 128, 128, 128, 128, 128, 128 },
-            {  55,  93, 255, 128, 128, 128, 128, 128, 128, 128, 128 }
-        },
+          { 1, 134, 252, 255, 255, 128, 128, 128, 128, 128, 128 },
+          { 213, 62, 250, 255, 255, 128, 128, 128, 128, 128, 128 },
+          { 55, 93, 255, 128, 128, 128, 128, 128, 128, 128, 128 } },
         { /* Coeff Band ( 7 )*/
-            { 128, 128, 128, 128, 128, 128, 128, 128, 128, 128, 128 },
-            { 128, 128, 128, 128, 128, 128, 128, 128, 128, 128, 128 },
-            { 128, 128, 128, 128, 128, 128, 128, 128, 128, 128, 128 }
-        }
-    },
-    { /* Block Type ( 3 ) */
+          { 128, 128, 128, 128, 128, 128, 128, 128, 128, 128, 128 },
+          { 128, 128, 128, 128, 128, 128, 128, 128, 128, 128, 128 },
+          { 128, 128, 128, 128, 128, 128, 128, 128, 128, 128, 128 } } },
+      {   /* Block Type ( 3 ) */
         { /* Coeff Band ( 0 )*/
-            { 202,  24, 213, 235, 186, 191, 220, 160, 240, 175, 255 },
-            { 126,  38, 182, 232, 169, 184, 228, 174, 255, 187, 128 },
-            {  61,  46, 138, 219, 151, 178, 240, 170, 255, 216, 128 }
-        },
+          { 202, 24, 213, 235, 186, 191, 220, 160, 240, 175, 255 },
+          { 126, 38, 182, 232, 169, 184, 228, 174, 255, 187, 128 },
+          { 61, 46, 138, 219, 151, 178, 240, 170, 255, 216, 128 } },
         { /* Coeff Band ( 1 )*/
-            {   1, 112, 230, 250, 199, 191, 247, 159, 255, 255, 128 },
-            { 166, 109, 228, 252, 211, 215, 255, 174, 128, 128, 128 },
-            {  39,  77, 162, 232, 172, 180, 245, 178, 255, 255, 128 }
-        },
+          { 1, 112, 230, 250, 199, 191, 247, 159, 255, 255, 128 },
+          { 166, 109, 228, 252, 211, 215, 255, 174, 128, 128, 128 },
+          { 39, 77, 162, 232, 172, 180, 245, 178, 255, 255, 128 } },
         { /* Coeff Band ( 2 )*/
-            {   1,  52, 220, 246, 198, 199, 249, 220, 255, 255, 128 },
-            { 124,  74, 191, 243, 183, 193, 250, 221, 255, 255, 128 },
-            {  24,  71, 130, 219, 154, 170, 243, 182, 255, 255, 128 }
-        },
+          { 1, 52, 220, 246, 198, 199, 249, 220, 255, 255, 128 },
+          { 124, 74, 191, 243, 183, 193, 250, 221, 255, 255, 128 },
+          { 24, 71, 130, 219, 154, 170, 243, 182, 255, 255, 128 } },
         { /* Coeff Band ( 3 )*/
-            {   1, 182, 225, 249, 219, 240, 255, 224, 128, 128, 128 },
-            { 149, 150, 226, 252, 216, 205, 255, 171, 128, 128, 128 },
-            {  28, 108, 170, 242, 183, 194, 254, 223, 255, 255, 128 }
-        },
+          { 1, 182, 225, 249, 219, 240, 255, 224, 128, 128, 128 },
+          { 149, 150, 226, 252, 216, 205, 255, 171, 128, 128, 128 },
+          { 28, 108, 170, 242, 183, 194, 254, 223, 255, 255, 128 } },
         { /* Coeff Band ( 4 )*/
-            {   1,  81, 230, 252, 204, 203, 255, 192, 128, 128, 128 },
-            { 123, 102, 209, 247, 188, 196, 255, 233, 128, 128, 128 },
-            {  20,  95, 153, 243, 164, 173, 255, 203, 128, 128, 128 }
-        },
+          { 1, 81, 230, 252, 204, 203, 255, 192, 128, 128, 128 },
+          { 123, 102, 209, 247, 188, 196, 255, 233, 128, 128, 128 },
+          { 20, 95, 153, 243, 164, 173, 255, 203, 128, 128, 128 } },
         { /* Coeff Band ( 5 )*/
-            {   1, 222, 248, 255, 216, 213, 128, 128, 128, 128, 128 },
-            { 168, 175, 246, 252, 235, 205, 255, 255, 128, 128, 128 },
-            {  47, 116, 215, 255, 211, 212, 255, 255, 128, 128, 128 }
-        },
+          { 1, 222, 248, 255, 216, 213, 128, 128, 128, 128, 128 },
+          { 168, 175, 246, 252, 235, 205, 255, 255, 128, 128, 128 },
+          { 47, 116, 215, 255, 211, 212, 255, 255, 128, 128, 128 } },
         { /* Coeff Band ( 6 )*/
-            {   1, 121, 236, 253, 212, 214, 255, 255, 128, 128, 128 },
-            { 141,  84, 213, 252, 201, 202, 255, 219, 128, 128, 128 },
-            {  42,  80, 160, 240, 162, 185, 255, 205, 128, 128, 128 }
-        },
+          { 1, 121, 236, 253, 212, 214, 255, 255, 128, 128, 128 },
+          { 141, 84, 213, 252, 201, 202, 255, 219, 128, 128, 128 },
+          { 42, 80, 160, 240, 162, 185, 255, 205, 128, 128, 128 } },
         { /* Coeff Band ( 7 )*/
-            {   1,   1, 255, 128, 128, 128, 128, 128, 128, 128, 128 },
-            { 244,   1, 255, 128, 128, 128, 128, 128, 128, 128, 128 },
-            { 238,   1, 255, 128, 128, 128, 128, 128, 128, 128, 128 }
-        }
-    }
-};
+          { 1, 1, 255, 128, 128, 128, 128, 128, 128, 128, 128 },
+          { 244, 1, 255, 128, 128, 128, 128, 128, 128, 128, 128 },
+          { 238, 1, 255, 128, 128, 128, 128, 128, 128, 128, 128 } } }
+    };
 
 #ifdef __cplusplus
 }  // extern "C"
--- a/vp8/common/dequantize.c
+++ b/vp8/common/dequantize.c
@@ -8,36 +8,30 @@
  *  be found in the AUTHORS file in the root of the source tree.
  */
 
-
 #include "vpx_config.h"
 #include "vp8_rtcd.h"
 #include "vp8/common/blockd.h"
 #include "vpx_mem/vpx_mem.h"
 
-void vp8_dequantize_b_c(BLOCKD *d, short *DQC)
-{
-    int i;
-    short *DQ  = d->dqcoeff;
-    short *Q   = d->qcoeff;
+void vp8_dequantize_b_c(BLOCKD *d, short *DQC) {
+  int i;
+  short *DQ = d->dqcoeff;
+  short *Q = d->qcoeff;
 
-    for (i = 0; i < 16; i++)
-    {
-        DQ[i] = Q[i] * DQC[i];
-    }
+  for (i = 0; i < 16; i++) {
+    DQ[i] = Q[i] * DQC[i];
+  }
 }
 
-void vp8_dequant_idct_add_c(short *input, short *dq,
-                            unsigned char *dest, int stride)
-{
-    int i;
+void vp8_dequant_idct_add_c(short *input, short *dq, unsigned char *dest,
+                            int stride) {
+  int i;
 
-    for (i = 0; i < 16; i++)
-    {
-        input[i] = dq[i] * input[i];
-    }
+  for (i = 0; i < 16; i++) {
+    input[i] = dq[i] * input[i];
+  }
 
-    vp8_short_idct4x4llm_c(input, dest, stride, dest, stride);
+  vp8_short_idct4x4llm_c(input, dest, stride, dest, stride);
 
-    memset(input, 0, 32);
-
+  memset(input, 0, 32);
 }
--- a/vp8/common/entropy.c
+++ b/vp8/common/entropy.c
@@ -15,47 +15,34 @@
 
 #include "coefupdateprobs.h"
 
-DECLARE_ALIGNED(16, const unsigned char, vp8_norm[256]) =
-{
-    0, 7, 6, 6, 5, 5, 5, 5, 4, 4, 4, 4, 4, 4, 4, 4,
-    3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3,
-    2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2,
-    2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2,
-    1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
-    1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
-    1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
-    1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
-    0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
-    0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
-    0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
-    0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
-    0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
-    0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
-    0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
-    0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0
+DECLARE_ALIGNED(16, const unsigned char, vp8_norm[256]) = {
+  0, 7, 6, 6, 5, 5, 5, 5, 4, 4, 4, 4, 4, 4, 4, 4, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3,
+  3, 3, 3, 3, 3, 3, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2,
+  2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
+  1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
+  1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0,
+  0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
+  0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
+  0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
+  0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
+  0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0
 };
 
-DECLARE_ALIGNED(16, const unsigned char, vp8_coef_bands[16]) =
-{ 0, 1, 2, 3, 6, 4, 5, 6, 6, 6, 6, 6, 6, 6, 6, 7};
+DECLARE_ALIGNED(16, const unsigned char, vp8_coef_bands[16]) = {
+  0, 1, 2, 3, 6, 4, 5, 6, 6, 6, 6, 6, 6, 6, 6, 7
+};
 
 DECLARE_ALIGNED(16, const unsigned char,
-                vp8_prev_token_class[MAX_ENTROPY_TOKENS]) =
-{ 0, 1, 2, 2, 2, 2, 2, 2, 2, 2, 2, 0};
+                vp8_prev_token_class[MAX_ENTROPY_TOKENS]) = {
+  0, 1, 2, 2, 2, 2, 2, 2, 2, 2, 2, 0
+};
 
-DECLARE_ALIGNED(16, const int, vp8_default_zig_zag1d[16]) =
-{
-    0,  1,  4,  8,
-    5,  2,  3,  6,
-    9, 12, 13, 10,
-    7, 11, 14, 15,
+DECLARE_ALIGNED(16, const int, vp8_default_zig_zag1d[16]) = {
+  0, 1, 4, 8, 5, 2, 3, 6, 9, 12, 13, 10, 7, 11, 14, 15,
 };
 
-DECLARE_ALIGNED(16, const short, vp8_default_inv_zig_zag[16]) =
-{
-    1,  2,  6,  7,
-    3,  5,  8, 13,
-    4,  9, 12, 14,
-   10, 11, 15, 16
+DECLARE_ALIGNED(16, const short, vp8_default_inv_zig_zag[16]) = {
+  1, 2, 6, 7, 3, 5, 8, 13, 4, 9, 12, 14, 10, 11, 15, 16
 };
 
 /* vp8_default_zig_zag_mask generated with:
@@ -71,87 +58,69 @@
 
     }
 */
-DECLARE_ALIGNED(16, const short, vp8_default_zig_zag_mask[16]) =
-{
-     1,    2,    32,     64,
-     4,   16,   128,   4096,
-     8,  256,  2048,   8192,
-   512, 1024, 16384, -32768
+DECLARE_ALIGNED(16, const short, vp8_default_zig_zag_mask[16]) = {
+  1, 2, 32, 64, 4, 16, 128, 4096, 8, 256, 2048, 8192, 512, 1024, 16384, -32768
 };
 
-const int vp8_mb_feature_data_bits[MB_LVL_MAX] = {7, 6};
+const int vp8_mb_feature_data_bits[MB_LVL_MAX] = { 7, 6 };
 
 /* Array indices are identical to previously-existing CONTEXT_NODE indices */
-
-const vp8_tree_index vp8_coef_tree[ 22] =     /* corresponding _CONTEXT_NODEs */
-{
-    -DCT_EOB_TOKEN, 2,                             /* 0 = EOB */
-    -ZERO_TOKEN, 4,                               /* 1 = ZERO */
-    -ONE_TOKEN, 6,                               /* 2 = ONE */
-    8, 12,                                      /* 3 = LOW_VAL */
-    -TWO_TOKEN, 10,                            /* 4 = TWO */
-    -THREE_TOKEN, -FOUR_TOKEN,                /* 5 = THREE */
-    14, 16,                                    /* 6 = HIGH_LOW */
-    -DCT_VAL_CATEGORY1, -DCT_VAL_CATEGORY2,   /* 7 = CAT_ONE */
-    18, 20,                                   /* 8 = CAT_THREEFOUR */
-    -DCT_VAL_CATEGORY3, -DCT_VAL_CATEGORY4,  /* 9 = CAT_THREE */
-    -DCT_VAL_CATEGORY5, -DCT_VAL_CATEGORY6   /* 10 = CAT_FIVE */
+/* corresponding _CONTEXT_NODEs */
+/* clang-format off */
+const vp8_tree_index vp8_coef_tree[22] = {
+  -DCT_EOB_TOKEN, 2,                       /* 0 = EOB */
+  -ZERO_TOKEN, 4,                          /* 1 = ZERO */
+  -ONE_TOKEN, 6,                           /* 2 = ONE */
+  8, 12,                                   /* 3 = LOW_VAL */
+  -TWO_TOKEN, 10,                          /* 4 = TWO */
+  -THREE_TOKEN, -FOUR_TOKEN,               /* 5 = THREE */
+  14, 16,                                  /* 6 = HIGH_LOW */
+  -DCT_VAL_CATEGORY1, -DCT_VAL_CATEGORY2,  /* 7 = CAT_ONE */
+  18, 20,                                  /* 8 = CAT_THREEFOUR */
+  -DCT_VAL_CATEGORY3, -DCT_VAL_CATEGORY4,  /* 9 = CAT_THREE */
+  -DCT_VAL_CATEGORY5, -DCT_VAL_CATEGORY6   /* 10 = CAT_FIVE */
 };
+/* clang-format on */
 
 /* vp8_coef_encodings generated with:
     vp8_tokens_from_tree(vp8_coef_encodings, vp8_coef_tree);
 */
-vp8_token vp8_coef_encodings[MAX_ENTROPY_TOKENS] =
-{
-    {2, 2},
-    {6, 3},
-    {28, 5},
-    {58, 6},
-    {59, 6},
-    {60, 6},
-    {61, 6},
-    {124, 7},
-    {125, 7},
-    {126, 7},
-    {127, 7},
-    {0, 1}
+vp8_token vp8_coef_encodings[MAX_ENTROPY_TOKENS] = {
+  { 2, 2 },  { 6, 3 },   { 28, 5 },  { 58, 6 },  { 59, 6 },  { 60, 6 },
+  { 61, 6 }, { 124, 7 }, { 125, 7 }, { 126, 7 }, { 127, 7 }, { 0, 1 }
 };
 
 /* Trees for extra bits.  Probabilities are constant and
    do not depend on previously encoded bits */
 
-static const vp8_prob Pcat1[] = { 159};
-static const vp8_prob Pcat2[] = { 165, 145};
-static const vp8_prob Pcat3[] = { 173, 148, 140};
-static const vp8_prob Pcat4[] = { 176, 155, 140, 135};
-static const vp8_prob Pcat5[] = { 180, 157, 141, 134, 130};
-static const vp8_prob Pcat6[] =
-{ 254, 254, 243, 230, 196, 177, 153, 140, 133, 130, 129};
+static const vp8_prob Pcat1[] = { 159 };
+static const vp8_prob Pcat2[] = { 165, 145 };
+static const vp8_prob Pcat3[] = { 173, 148, 140 };
+static const vp8_prob Pcat4[] = { 176, 155, 140, 135 };
+static const vp8_prob Pcat5[] = { 180, 157, 141, 134, 130 };
+static const vp8_prob Pcat6[] = { 254, 254, 243, 230, 196, 177,
+                                  153, 140, 133, 130, 129 };
 
-
 /* tree index tables generated with:
 
-    void init_bit_tree(vp8_tree_index *p, int n)
-    {
-        int i = 0;
+    void init_bit_tree(vp8_tree_index *p, int n) {
+      int i = 0;
 
-        while (++i < n)
-        {
-            p[0] = p[1] = i << 1;
-            p += 2;
-        }
+      while (++i < n) {
+          p[0] = p[1] = i << 1;
+          p += 2;
+      }
 
-        p[0] = p[1] = 0;
+      p[0] = p[1] = 0;
     }
 
-    void init_bit_trees()
-    {
-        init_bit_tree(cat1, 1);
-        init_bit_tree(cat2, 2);
-        init_bit_tree(cat3, 3);
-        init_bit_tree(cat4, 4);
-        init_bit_tree(cat5, 5);
-        init_bit_tree(cat6, 11);
+    void init_bit_trees() {
+      init_bit_tree(cat1, 1);
+      init_bit_tree(cat2, 2);
+      init_bit_tree(cat3, 3);
+      init_bit_tree(cat4, 4);
+      init_bit_tree(cat5, 5);
+      init_bit_tree(cat6, 11);
     }
 */
 
@@ -160,29 +129,19 @@
 static const vp8_tree_index cat3[6] = { 2, 2, 4, 4, 0, 0 };
 static const vp8_tree_index cat4[8] = { 2, 2, 4, 4, 6, 6, 0, 0 };
 static const vp8_tree_index cat5[10] = { 2, 2, 4, 4, 6, 6, 8, 8, 0, 0 };
-static const vp8_tree_index cat6[22] = { 2, 2, 4, 4, 6, 6, 8, 8, 10, 10, 12, 12,
-                                        14, 14, 16, 16, 18, 18, 20, 20, 0, 0 };
+static const vp8_tree_index cat6[22] = {
+  2, 2, 4, 4, 6, 6, 8, 8, 10, 10, 12, 12, 14, 14, 16, 16, 18, 18, 20, 20, 0, 0
+};
 
-const vp8_extra_bit_struct vp8_extra_bits[12] =
-{
-    { 0, 0, 0, 0},
-    { 0, 0, 0, 1},
-    { 0, 0, 0, 2},
-    { 0, 0, 0, 3},
-    { 0, 0, 0, 4},
-    { cat1, Pcat1, 1, 5},
-    { cat2, Pcat2, 2, 7},
-    { cat3, Pcat3, 3, 11},
-    { cat4, Pcat4, 4, 19},
-    { cat5, Pcat5, 5, 35},
-    { cat6, Pcat6, 11, 67},
-    { 0, 0, 0, 0}
+const vp8_extra_bit_struct vp8_extra_bits[12] = {
+  { 0, 0, 0, 0 },         { 0, 0, 0, 1 },          { 0, 0, 0, 2 },
+  { 0, 0, 0, 3 },         { 0, 0, 0, 4 },          { cat1, Pcat1, 1, 5 },
+  { cat2, Pcat2, 2, 7 },  { cat3, Pcat3, 3, 11 },  { cat4, Pcat4, 4, 19 },
+  { cat5, Pcat5, 5, 35 }, { cat6, Pcat6, 11, 67 }, { 0, 0, 0, 0 }
 };
 
 #include "default_coef_probs.h"
 
-void vp8_default_coef_probs(VP8_COMMON *pc)
-{
-    memcpy(pc->fc.coef_probs, default_coef_probs, sizeof(default_coef_probs));
+void vp8_default_coef_probs(VP8_COMMON *pc) {
+  memcpy(pc->fc.coef_probs, default_coef_probs, sizeof(default_coef_probs));
 }
-
--- a/vp8/common/entropy.h
+++ b/vp8/common/entropy.h
@@ -8,7 +8,6 @@
  *  be found in the AUTHORS file in the root of the source tree.
  */
 
-
 #ifndef VP8_COMMON_ENTROPY_H_
 #define VP8_COMMON_ENTROPY_H_
 
@@ -21,18 +20,18 @@
 
 /* Coefficient token alphabet */
 
-#define ZERO_TOKEN              0       /* 0         Extra Bits 0+0 */
-#define ONE_TOKEN               1       /* 1         Extra Bits 0+1 */
-#define TWO_TOKEN               2       /* 2         Extra Bits 0+1 */
-#define THREE_TOKEN             3       /* 3         Extra Bits 0+1 */
-#define FOUR_TOKEN              4       /* 4         Extra Bits 0+1 */
-#define DCT_VAL_CATEGORY1       5       /* 5-6       Extra Bits 1+1 */
-#define DCT_VAL_CATEGORY2       6       /* 7-10      Extra Bits 2+1 */
-#define DCT_VAL_CATEGORY3       7       /* 11-18     Extra Bits 3+1 */
-#define DCT_VAL_CATEGORY4       8       /* 19-34     Extra Bits 4+1 */
-#define DCT_VAL_CATEGORY5       9       /* 35-66     Extra Bits 5+1 */
-#define DCT_VAL_CATEGORY6       10      /* 67+       Extra Bits 11+1 */
-#define DCT_EOB_TOKEN           11      /* EOB       Extra Bits 0+0 */
+#define ZERO_TOKEN 0         /* 0         Extra Bits 0+0 */
+#define ONE_TOKEN 1          /* 1         Extra Bits 0+1 */
+#define TWO_TOKEN 2          /* 2         Extra Bits 0+1 */
+#define THREE_TOKEN 3        /* 3         Extra Bits 0+1 */
+#define FOUR_TOKEN 4         /* 4         Extra Bits 0+1 */
+#define DCT_VAL_CATEGORY1 5  /* 5-6       Extra Bits 1+1 */
+#define DCT_VAL_CATEGORY2 6  /* 7-10      Extra Bits 2+1 */
+#define DCT_VAL_CATEGORY3 7  /* 11-18     Extra Bits 3+1 */
+#define DCT_VAL_CATEGORY4 8  /* 19-34     Extra Bits 4+1 */
+#define DCT_VAL_CATEGORY5 9  /* 35-66     Extra Bits 5+1 */
+#define DCT_VAL_CATEGORY6 10 /* 67+       Extra Bits 11+1 */
+#define DCT_EOB_TOKEN 11     /* EOB       Extra Bits 0+0 */
 
 #define MAX_ENTROPY_TOKENS 12
 #define ENTROPY_NODES 11
@@ -41,22 +40,21 @@
 
 extern const struct vp8_token_struct vp8_coef_encodings[MAX_ENTROPY_TOKENS];
 
-typedef struct
-{
-    vp8_tree_p tree;
-    const vp8_prob *prob;
-    int Len;
-    int base_val;
+typedef struct {
+  vp8_tree_p tree;
+  const vp8_prob *prob;
+  int Len;
+  int base_val;
 } vp8_extra_bit_struct;
 
-extern const vp8_extra_bit_struct vp8_extra_bits[12];    /* indexed by token value */
+extern const vp8_extra_bit_struct
+    vp8_extra_bits[12]; /* indexed by token value */
 
-#define PROB_UPDATE_BASELINE_COST   7
+#define PROB_UPDATE_BASELINE_COST 7
 
-#define MAX_PROB                255
-#define DCT_MAX_VALUE           2048
+#define MAX_PROB 255
+#define DCT_MAX_VALUE 2048
 
-
 /* Coefficients are predicted via a 3-dimensional probability table. */
 
 /* Outside dimension.  0 = Y no DC, 1 = Y2, 2 = UV, 3 = Y with DC */
@@ -86,12 +84,13 @@
    distinct bands). */
 
 /*# define DC_TOKEN_CONTEXTS        3*/ /* 00, 0!0, !0!0 */
-#   define PREV_COEF_CONTEXTS       3
+#define PREV_COEF_CONTEXTS 3
 
-extern DECLARE_ALIGNED(16, const unsigned char, vp8_prev_token_class[MAX_ENTROPY_TOKENS]);
+extern DECLARE_ALIGNED(16, const unsigned char,
+                       vp8_prev_token_class[MAX_ENTROPY_TOKENS]);
 
-extern const vp8_prob vp8_coef_update_probs [BLOCK_TYPES] [COEF_BANDS] [PREV_COEF_CONTEXTS] [ENTROPY_NODES];
-
+extern const vp8_prob vp8_coef_update_probs[BLOCK_TYPES][COEF_BANDS]
+                                           [PREV_COEF_CONTEXTS][ENTROPY_NODES];
 
 struct VP8Common;
 void vp8_default_coef_probs(struct VP8Common *);
--- a/vp8/common/entropymode.c
+++ b/vp8/common/entropymode.c
@@ -16,156 +16,97 @@
 
 #include "vp8_entropymodedata.h"
 
-int vp8_mv_cont(const int_mv *l, const int_mv *a)
-{
-    int lez = (l->as_int == 0);
-    int aez = (a->as_int == 0);
-    int lea = (l->as_int == a->as_int);
+int vp8_mv_cont(const int_mv *l, const int_mv *a) {
+  int lez = (l->as_int == 0);
+  int aez = (a->as_int == 0);
+  int lea = (l->as_int == a->as_int);
 
-    if (lea && lez)
-        return SUBMVREF_LEFT_ABOVE_ZED;
+  if (lea && lez) return SUBMVREF_LEFT_ABOVE_ZED;
 
-    if (lea)
-        return SUBMVREF_LEFT_ABOVE_SAME;
+  if (lea) return SUBMVREF_LEFT_ABOVE_SAME;
 
-    if (aez)
-        return SUBMVREF_ABOVE_ZED;
+  if (aez) return SUBMVREF_ABOVE_ZED;
 
-    if (lez)
-        return SUBMVREF_LEFT_ZED;
+  if (lez) return SUBMVREF_LEFT_ZED;
 
-    return SUBMVREF_NORMAL;
+  return SUBMVREF_NORMAL;
 }
 
-static const vp8_prob sub_mv_ref_prob [VP8_SUBMVREFS-1] = { 180, 162, 25};
+static const vp8_prob sub_mv_ref_prob[VP8_SUBMVREFS - 1] = { 180, 162, 25 };
 
-const vp8_prob vp8_sub_mv_ref_prob2 [SUBMVREF_COUNT][VP8_SUBMVREFS-1] =
-{
-    { 147, 136, 18 },
-    { 106, 145, 1  },
-    { 179, 121, 1  },
-    { 223, 1  , 34 },
-    { 208, 1  , 1  }
+const vp8_prob vp8_sub_mv_ref_prob2[SUBMVREF_COUNT][VP8_SUBMVREFS - 1] = {
+  { 147, 136, 18 },
+  { 106, 145, 1 },
+  { 179, 121, 1 },
+  { 223, 1, 34 },
+  { 208, 1, 1 }
 };
 
-
-
-const vp8_mbsplit vp8_mbsplits [VP8_NUMMBSPLITS] =
-{
-    {
-        0,  0,  0,  0,
-        0,  0,  0,  0,
-        1,  1,  1,  1,
-        1,  1,  1,  1,
-    },
-    {
-        0,  0,  1,  1,
-        0,  0,  1,  1,
-        0,  0,  1,  1,
-        0,  0,  1,  1,
-    },
-    {
-        0,  0,  1,  1,
-        0,  0,  1,  1,
-        2,  2,  3,  3,
-        2,  2,  3,  3,
-    },
-    {
-        0,  1,  2,  3,
-        4,  5,  6,  7,
-        8,  9,  10, 11,
-        12, 13, 14, 15,
-    }
+const vp8_mbsplit vp8_mbsplits[VP8_NUMMBSPLITS] = {
+  {
+      0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 1, 1, 1,
+  },
+  {
+      0, 0, 1, 1, 0, 0, 1, 1, 0, 0, 1, 1, 0, 0, 1, 1,
+  },
+  {
+      0, 0, 1, 1, 0, 0, 1, 1, 2, 2, 3, 3, 2, 2, 3, 3,
+  },
+  {
+      0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15,
+  }
 };
 
-const int vp8_mbsplit_count [VP8_NUMMBSPLITS] = { 2, 2, 4, 16};
+const int vp8_mbsplit_count[VP8_NUMMBSPLITS] = { 2, 2, 4, 16 };
 
-const vp8_prob vp8_mbsplit_probs [VP8_NUMMBSPLITS-1] = { 110, 111, 150};
+const vp8_prob vp8_mbsplit_probs[VP8_NUMMBSPLITS - 1] = { 110, 111, 150 };
 
-
 /* Array indices are identical to previously-existing INTRAMODECONTEXTNODES. */
 
-const vp8_tree_index vp8_bmode_tree[18] =     /* INTRAMODECONTEXTNODE value */
-{
-    -B_DC_PRED, 2,                             /* 0 = DC_NODE */
-    -B_TM_PRED, 4,                            /* 1 = TM_NODE */
-    -B_VE_PRED, 6,                           /* 2 = VE_NODE */
-    8, 12,                                  /* 3 = COM_NODE */
-    -B_HE_PRED, 10,                        /* 4 = HE_NODE */
-    -B_RD_PRED, -B_VR_PRED,               /* 5 = RD_NODE */
-    -B_LD_PRED, 14,                        /* 6 = LD_NODE */
-    -B_VL_PRED, 16,                      /* 7 = VL_NODE */
-    -B_HD_PRED, -B_HU_PRED             /* 8 = HD_NODE */
-};
+const vp8_tree_index vp8_bmode_tree[18] = /* INTRAMODECONTEXTNODE value */
+    {
+      -B_DC_PRED, 2,          /* 0 = DC_NODE */
+      -B_TM_PRED, 4,          /* 1 = TM_NODE */
+      -B_VE_PRED, 6,          /* 2 = VE_NODE */
+      8,          12,         /* 3 = COM_NODE */
+      -B_HE_PRED, 10,         /* 4 = HE_NODE */
+      -B_RD_PRED, -B_VR_PRED, /* 5 = RD_NODE */
+      -B_LD_PRED, 14,         /* 6 = LD_NODE */
+      -B_VL_PRED, 16,         /* 7 = VL_NODE */
+      -B_HD_PRED, -B_HU_PRED  /* 8 = HD_NODE */
+    };
 
 /* Again, these trees use the same probability indices as their
    explicitly-programmed predecessors. */
 
-const vp8_tree_index vp8_ymode_tree[8] =
-{
-    -DC_PRED, 2,
-    4, 6,
-    -V_PRED, -H_PRED,
-    -TM_PRED, -B_PRED
+const vp8_tree_index vp8_ymode_tree[8] = {
+  -DC_PRED, 2, 4, 6, -V_PRED, -H_PRED, -TM_PRED, -B_PRED
 };
 
-const vp8_tree_index vp8_kf_ymode_tree[8] =
-{
-    -B_PRED, 2,
-    4, 6,
-    -DC_PRED, -V_PRED,
-    -H_PRED, -TM_PRED
+const vp8_tree_index vp8_kf_ymode_tree[8] = {
+  -B_PRED, 2, 4, 6, -DC_PRED, -V_PRED, -H_PRED, -TM_PRED
 };
 
-const vp8_tree_index vp8_uv_mode_tree[6] =
-{
-    -DC_PRED, 2,
-    -V_PRED, 4,
-    -H_PRED, -TM_PRED
-};
+const vp8_tree_index vp8_uv_mode_tree[6] = { -DC_PRED, 2,       -V_PRED,
+                                             4,        -H_PRED, -TM_PRED };
 
-const vp8_tree_index vp8_mbsplit_tree[6] =
-{
-    -3, 2,
-    -2, 4,
-    -0, -1
-};
+const vp8_tree_index vp8_mbsplit_tree[6] = { -3, 2, -2, 4, -0, -1 };
 
-const vp8_tree_index vp8_mv_ref_tree[8] =
-{
-    -ZEROMV, 2,
-    -NEARESTMV, 4,
-    -NEARMV, 6,
-    -NEWMV, -SPLITMV
-};
+const vp8_tree_index vp8_mv_ref_tree[8] = { -ZEROMV, 2, -NEARESTMV, 4,
+                                            -NEARMV, 6, -NEWMV,     -SPLITMV };
 
-const vp8_tree_index vp8_sub_mv_ref_tree[6] =
-{
-    -LEFT4X4, 2,
-    -ABOVE4X4, 4,
-    -ZERO4X4, -NEW4X4
-};
+const vp8_tree_index vp8_sub_mv_ref_tree[6] = { -LEFT4X4, 2,        -ABOVE4X4,
+                                                4,        -ZERO4X4, -NEW4X4 };
 
-const vp8_tree_index vp8_small_mvtree [14] =
-{
-    2, 8,
-    4, 6,
-    -0, -1,
-    -2, -3,
-    10, 12,
-    -4, -5,
-    -6, -7
-};
+const vp8_tree_index vp8_small_mvtree[14] = { 2,  8,  4,  6,  -0, -1, -2,
+                                              -3, 10, 12, -4, -5, -6, -7 };
 
-void vp8_init_mbmode_probs(VP8_COMMON *x)
-{
-    memcpy(x->fc.ymode_prob, vp8_ymode_prob, sizeof(vp8_ymode_prob));
-    memcpy(x->fc.uv_mode_prob, vp8_uv_mode_prob, sizeof(vp8_uv_mode_prob));
-    memcpy(x->fc.sub_mv_ref_prob, sub_mv_ref_prob, sizeof(sub_mv_ref_prob));
+void vp8_init_mbmode_probs(VP8_COMMON *x) {
+  memcpy(x->fc.ymode_prob, vp8_ymode_prob, sizeof(vp8_ymode_prob));
+  memcpy(x->fc.uv_mode_prob, vp8_uv_mode_prob, sizeof(vp8_uv_mode_prob));
+  memcpy(x->fc.sub_mv_ref_prob, sub_mv_ref_prob, sizeof(sub_mv_ref_prob));
 }
 
-void vp8_default_bmode_probs(vp8_prob p [VP8_BINTRAMODES-1])
-{
-    memcpy(p, vp8_bmode_prob, sizeof(vp8_bmode_prob));
+void vp8_default_bmode_probs(vp8_prob p[VP8_BINTRAMODES - 1]) {
+  memcpy(p, vp8_bmode_prob, sizeof(vp8_bmode_prob));
 }
-
--- a/vp8/common/entropymode.h
+++ b/vp8/common/entropymode.h
@@ -8,7 +8,6 @@
  *  be found in the AUTHORS file in the root of the source tree.
  */
 
-
 #ifndef VP8_COMMON_ENTROPYMODE_H_
 #define VP8_COMMON_ENTROPYMODE_H_
 
@@ -19,13 +18,12 @@
 extern "C" {
 #endif
 
-typedef enum
-{
-    SUBMVREF_NORMAL,
-    SUBMVREF_LEFT_ZED,
-    SUBMVREF_ABOVE_ZED,
-    SUBMVREF_LEFT_ABOVE_SAME,
-    SUBMVREF_LEFT_ABOVE_ZED
+typedef enum {
+  SUBMVREF_NORMAL,
+  SUBMVREF_LEFT_ZED,
+  SUBMVREF_ABOVE_ZED,
+  SUBMVREF_LEFT_ABOVE_SAME,
+  SUBMVREF_LEFT_ABOVE_ZED
 } sumvfref_t;
 
 typedef int vp8_mbsplit[16];
@@ -32,29 +30,28 @@
 
 #define VP8_NUMMBSPLITS 4
 
-extern const vp8_mbsplit vp8_mbsplits [VP8_NUMMBSPLITS];
+extern const vp8_mbsplit vp8_mbsplits[VP8_NUMMBSPLITS];
 
-extern const int vp8_mbsplit_count [VP8_NUMMBSPLITS];    /* # of subsets */
+extern const int vp8_mbsplit_count[VP8_NUMMBSPLITS]; /* # of subsets */
 
-extern const vp8_prob vp8_mbsplit_probs [VP8_NUMMBSPLITS-1];
+extern const vp8_prob vp8_mbsplit_probs[VP8_NUMMBSPLITS - 1];
 
 extern int vp8_mv_cont(const int_mv *l, const int_mv *a);
 #define SUBMVREF_COUNT 5
-extern const vp8_prob vp8_sub_mv_ref_prob2 [SUBMVREF_COUNT][VP8_SUBMVREFS-1];
+extern const vp8_prob vp8_sub_mv_ref_prob2[SUBMVREF_COUNT][VP8_SUBMVREFS - 1];
 
+extern const unsigned int vp8_kf_default_bmode_counts
+    [VP8_BINTRAMODES][VP8_BINTRAMODES][VP8_BINTRAMODES];
 
-extern const unsigned int vp8_kf_default_bmode_counts [VP8_BINTRAMODES] [VP8_BINTRAMODES] [VP8_BINTRAMODES];
-
-
 extern const vp8_tree_index vp8_bmode_tree[];
 
-extern const vp8_tree_index  vp8_ymode_tree[];
-extern const vp8_tree_index  vp8_kf_ymode_tree[];
-extern const vp8_tree_index  vp8_uv_mode_tree[];
+extern const vp8_tree_index vp8_ymode_tree[];
+extern const vp8_tree_index vp8_kf_ymode_tree[];
+extern const vp8_tree_index vp8_uv_mode_tree[];
 
-extern const vp8_tree_index  vp8_mbsplit_tree[];
-extern const vp8_tree_index  vp8_mv_ref_tree[];
-extern const vp8_tree_index  vp8_sub_mv_ref_tree[];
+extern const vp8_tree_index vp8_mbsplit_tree[];
+extern const vp8_tree_index vp8_mv_ref_tree[];
+extern const vp8_tree_index vp8_sub_mv_ref_tree[];
 
 extern const struct vp8_token_struct vp8_bmode_encodings[VP8_BINTRAMODES];
 extern const struct vp8_token_struct vp8_ymode_encodings[VP8_YMODES];
@@ -65,7 +62,8 @@
 /* Inter mode values do not start at zero */
 
 extern const struct vp8_token_struct vp8_mv_ref_encoding_array[VP8_MVREFS];
-extern const struct vp8_token_struct vp8_sub_mv_ref_encoding_array[VP8_SUBMVREFS];
+extern const struct vp8_token_struct
+    vp8_sub_mv_ref_encoding_array[VP8_SUBMVREFS];
 
 extern const vp8_tree_index vp8_small_mvtree[];
 
@@ -72,14 +70,15 @@
 extern const struct vp8_token_struct vp8_small_mvencodings[8];
 
 /* Key frame default mode probs */
-extern const vp8_prob vp8_kf_bmode_prob[VP8_BINTRAMODES][VP8_BINTRAMODES]
-[VP8_BINTRAMODES-1];
-extern const vp8_prob vp8_kf_uv_mode_prob[VP8_UV_MODES-1];
-extern const vp8_prob vp8_kf_ymode_prob[VP8_YMODES-1];
+extern const vp8_prob
+    vp8_kf_bmode_prob[VP8_BINTRAMODES][VP8_BINTRAMODES][VP8_BINTRAMODES - 1];
+extern const vp8_prob vp8_kf_uv_mode_prob[VP8_UV_MODES - 1];
+extern const vp8_prob vp8_kf_ymode_prob[VP8_YMODES - 1];
 
 void vp8_init_mbmode_probs(VP8_COMMON *x);
-void vp8_default_bmode_probs(vp8_prob dest [VP8_BINTRAMODES-1]);
-void vp8_kf_default_bmode_probs(vp8_prob dest [VP8_BINTRAMODES] [VP8_BINTRAMODES] [VP8_BINTRAMODES-1]);
+void vp8_default_bmode_probs(vp8_prob dest[VP8_BINTRAMODES - 1]);
+void vp8_kf_default_bmode_probs(
+    vp8_prob dest[VP8_BINTRAMODES][VP8_BINTRAMODES][VP8_BINTRAMODES - 1]);
 
 #ifdef __cplusplus
 }  // extern "C"
--- a/vp8/common/entropymv.c
+++ b/vp8/common/entropymv.c
@@ -8,42 +8,40 @@
  *  be found in the AUTHORS file in the root of the source tree.
  */
 
-
 #include "entropymv.h"
 
-const MV_CONTEXT vp8_mv_update_probs[2] =
-{
-    {{
-        237,
-        246,
-        253, 253, 254, 254, 254, 254, 254,
-        254, 254, 254, 254, 254, 250, 250, 252, 254, 254
-    }},
-    {{
-        231,
-        243,
-        245, 253, 254, 254, 254, 254, 254,
-        254, 254, 254, 254, 254, 251, 251, 254, 254, 254
-    }}
+/* clang-format off */
+const MV_CONTEXT vp8_mv_update_probs[2] = {
+  { {
+      237,
+      246,
+      253, 253, 254, 254, 254, 254, 254,
+      254, 254, 254, 254, 254, 250, 250, 252, 254, 254
+  } },
+  { {
+      231,
+      243,
+      245, 253, 254, 254, 254, 254, 254,
+      254, 254, 254, 254, 254, 251, 251, 254, 254, 254
+  } }
 };
-const MV_CONTEXT vp8_default_mv_context[2] =
-{
-    {{
-        /* row */
-        162,                                        /* is short */
-        128,                                        /* sign */
-        225, 146, 172, 147, 214,  39, 156,          /* short tree */
-        128, 129, 132,  75, 145, 178, 206, 239, 254, 254 /* long bits */
-    }},
+/* clang-format on */
 
+const MV_CONTEXT vp8_default_mv_context[2] = {
+  { {
+      /* row */
+      162,                                            /* is short */
+      128,                                            /* sign */
+      225, 146, 172, 147, 214, 39, 156,               /* short tree */
+      128, 129, 132, 75, 145, 178, 206, 239, 254, 254 /* long bits */
+  } },
 
+  { {
+      /* same for column */
+      164,                                            /* is short */
+      128,                                            /**/
+      204, 170, 119, 235, 140, 230, 228,              /**/
+      128, 130, 130, 74, 148, 180, 203, 236, 254, 254 /* long bits */
 
-    {{
-        /* same for column */
-        164,                                        /* is short */
-        128,
-        204, 170, 119, 235, 140, 230, 228,
-        128, 130, 130,  74, 148, 180, 203, 236, 254, 254 /* long bits */
-
-    }}
+  } }
 };
--- a/vp8/common/entropymv.h
+++ b/vp8/common/entropymv.h
@@ -8,7 +8,6 @@
  *  be found in the AUTHORS file in the root of the source tree.
  */
 
-
 #ifndef VP8_COMMON_ENTROPYMV_H_
 #define VP8_COMMON_ENTROPYMV_H_
 
@@ -18,29 +17,27 @@
 extern "C" {
 #endif
 
-enum
-{
-    mv_max  = 1023,              /* max absolute value of a MV component */
-    MVvals = (2 * mv_max) + 1,   /* # possible values "" */
-    mvfp_max  = 255,              /* max absolute value of a full pixel MV component */
-    MVfpvals = (2 * mvfp_max) +1, /* # possible full pixel MV values */
+enum {
+  mv_max = 1023,             /* max absolute value of a MV component */
+  MVvals = (2 * mv_max) + 1, /* # possible values "" */
+  mvfp_max = 255, /* max absolute value of a full pixel MV component */
+  MVfpvals = (2 * mvfp_max) + 1, /* # possible full pixel MV values */
 
-    mvlong_width = 10,       /* Large MVs have 9 bit magnitudes */
-    mvnum_short = 8,         /* magnitudes 0 through 7 */
+  mvlong_width = 10, /* Large MVs have 9 bit magnitudes */
+  mvnum_short = 8,   /* magnitudes 0 through 7 */
 
-    /* probability offsets for coding each MV component */
+  /* probability offsets for coding each MV component */
 
-    mvpis_short = 0,         /* short (<= 7) vs long (>= 8) */
-    MVPsign,                /* sign for non-zero */
-    MVPshort,               /* 8 short values = 7-position tree */
+  mvpis_short = 0, /* short (<= 7) vs long (>= 8) */
+  MVPsign,         /* sign for non-zero */
+  MVPshort,        /* 8 short values = 7-position tree */
 
-    MVPbits = MVPshort + mvnum_short - 1, /* mvlong_width long value bits */
-    MVPcount = MVPbits + mvlong_width    /* (with independent probabilities) */
+  MVPbits = MVPshort + mvnum_short - 1, /* mvlong_width long value bits */
+  MVPcount = MVPbits + mvlong_width     /* (with independent probabilities) */
 };
 
-typedef struct mv_context
-{
-    vp8_prob prob[MVPcount];  /* often come in row, col pairs */
+typedef struct mv_context {
+  vp8_prob prob[MVPcount]; /* often come in row, col pairs */
 } MV_CONTEXT;
 
 extern const MV_CONTEXT vp8_mv_update_probs[2], vp8_default_mv_context[2];
--- a/vp8/common/extend.c
+++ b/vp8/common/extend.c
@@ -8,181 +8,146 @@
  *  be found in the AUTHORS file in the root of the source tree.
  */
 
-
 #include "extend.h"
 #include "vpx_mem/vpx_mem.h"
 
+static void copy_and_extend_plane(unsigned char *s, /* source */
+                                  int sp,           /* source pitch */
+                                  unsigned char *d, /* destination */
+                                  int dp,           /* destination pitch */
+                                  int h,            /* height */
+                                  int w,            /* width */
+                                  int et,           /* extend top border */
+                                  int el,           /* extend left border */
+                                  int eb,           /* extend bottom border */
+                                  int er            /* extend right border */
+                                  ) {
+  int i;
+  unsigned char *src_ptr1, *src_ptr2;
+  unsigned char *dest_ptr1, *dest_ptr2;
+  int linesize;
 
-static void copy_and_extend_plane
-(
-    unsigned char *s, /* source */
-    int sp,           /* source pitch */
-    unsigned char *d, /* destination */
-    int dp,           /* destination pitch */
-    int h,            /* height */
-    int w,            /* width */
-    int et,           /* extend top border */
-    int el,           /* extend left border */
-    int eb,           /* extend bottom border */
-    int er            /* extend right border */
-)
-{
-    int i;
-    unsigned char *src_ptr1, *src_ptr2;
-    unsigned char *dest_ptr1, *dest_ptr2;
-    int linesize;
+  /* copy the left and right most columns out */
+  src_ptr1 = s;
+  src_ptr2 = s + w - 1;
+  dest_ptr1 = d - el;
+  dest_ptr2 = d + w;
 
-    /* copy the left and right most columns out */
-    src_ptr1 = s;
-    src_ptr2 = s + w - 1;
-    dest_ptr1 = d - el;
-    dest_ptr2 = d + w;
+  for (i = 0; i < h; i++) {
+    memset(dest_ptr1, src_ptr1[0], el);
+    memcpy(dest_ptr1 + el, src_ptr1, w);
+    memset(dest_ptr2, src_ptr2[0], er);
+    src_ptr1 += sp;
+    src_ptr2 += sp;
+    dest_ptr1 += dp;
+    dest_ptr2 += dp;
+  }
 
-    for (i = 0; i < h; i++)
-    {
-        memset(dest_ptr1, src_ptr1[0], el);
-        memcpy(dest_ptr1 + el, src_ptr1, w);
-        memset(dest_ptr2, src_ptr2[0], er);
-        src_ptr1  += sp;
-        src_ptr2  += sp;
-        dest_ptr1 += dp;
-        dest_ptr2 += dp;
-    }
+  /* Now copy the top and bottom lines into each line of the respective
+   * borders
+   */
+  src_ptr1 = d - el;
+  src_ptr2 = d + dp * (h - 1) - el;
+  dest_ptr1 = d + dp * (-et) - el;
+  dest_ptr2 = d + dp * (h)-el;
+  linesize = el + er + w;
 
-    /* Now copy the top and bottom lines into each line of the respective
-     * borders
-     */
-    src_ptr1 = d - el;
-    src_ptr2 = d + dp * (h - 1) - el;
-    dest_ptr1 = d + dp * (-et) - el;
-    dest_ptr2 = d + dp * (h) - el;
-    linesize = el + er + w;
+  for (i = 0; i < et; i++) {
+    memcpy(dest_ptr1, src_ptr1, linesize);
+    dest_ptr1 += dp;
+  }
 
-    for (i = 0; i < et; i++)
-    {
-        memcpy(dest_ptr1, src_ptr1, linesize);
-        dest_ptr1 += dp;
-    }
-
-    for (i = 0; i < eb; i++)
-    {
-        memcpy(dest_ptr2, src_ptr2, linesize);
-        dest_ptr2 += dp;
-    }
+  for (i = 0; i < eb; i++) {
+    memcpy(dest_ptr2, src_ptr2, linesize);
+    dest_ptr2 += dp;
+  }
 }
 
-
 void vp8_copy_and_extend_frame(YV12_BUFFER_CONFIG *src,
-                               YV12_BUFFER_CONFIG *dst)
-{
-    int et = dst->border;
-    int el = dst->border;
-    int eb = dst->border + dst->y_height - src->y_height;
-    int er = dst->border + dst->y_width - src->y_width;
+                               YV12_BUFFER_CONFIG *dst) {
+  int et = dst->border;
+  int el = dst->border;
+  int eb = dst->border + dst->y_height - src->y_height;
+  int er = dst->border + dst->y_width - src->y_width;
 
-    copy_and_extend_plane(src->y_buffer, src->y_stride,
-                          dst->y_buffer, dst->y_stride,
-                          src->y_height, src->y_width,
-                          et, el, eb, er);
+  copy_and_extend_plane(src->y_buffer, src->y_stride, dst->y_buffer,
+                        dst->y_stride, src->y_height, src->y_width, et, el, eb,
+                        er);
 
-    et = dst->border >> 1;
-    el = dst->border >> 1;
-    eb = (dst->border >> 1) + dst->uv_height - src->uv_height;
-    er = (dst->border >> 1) + dst->uv_width - src->uv_width;
+  et = dst->border >> 1;
+  el = dst->border >> 1;
+  eb = (dst->border >> 1) + dst->uv_height - src->uv_height;
+  er = (dst->border >> 1) + dst->uv_width - src->uv_width;
 
-    copy_and_extend_plane(src->u_buffer, src->uv_stride,
-                          dst->u_buffer, dst->uv_stride,
-                          src->uv_height, src->uv_width,
-                          et, el, eb, er);
+  copy_and_extend_plane(src->u_buffer, src->uv_stride, dst->u_buffer,
+                        dst->uv_stride, src->uv_height, src->uv_width, et, el,
+                        eb, er);
 
-    copy_and_extend_plane(src->v_buffer, src->uv_stride,
-                          dst->v_buffer, dst->uv_stride,
-                          src->uv_height, src->uv_width,
-                          et, el, eb, er);
+  copy_and_extend_plane(src->v_buffer, src->uv_stride, dst->v_buffer,
+                        dst->uv_stride, src->uv_height, src->uv_width, et, el,
+                        eb, er);
 }
 
-
 void vp8_copy_and_extend_frame_with_rect(YV12_BUFFER_CONFIG *src,
-                                         YV12_BUFFER_CONFIG *dst,
-                                         int srcy, int srcx,
-                                         int srch, int srcw)
-{
-    int et = dst->border;
-    int el = dst->border;
-    int eb = dst->border + dst->y_height - src->y_height;
-    int er = dst->border + dst->y_width - src->y_width;
-    int src_y_offset = srcy * src->y_stride + srcx;
-    int dst_y_offset = srcy * dst->y_stride + srcx;
-    int src_uv_offset = ((srcy * src->uv_stride) >> 1) + (srcx >> 1);
-    int dst_uv_offset = ((srcy * dst->uv_stride) >> 1) + (srcx >> 1);
+                                         YV12_BUFFER_CONFIG *dst, int srcy,
+                                         int srcx, int srch, int srcw) {
+  int et = dst->border;
+  int el = dst->border;
+  int eb = dst->border + dst->y_height - src->y_height;
+  int er = dst->border + dst->y_width - src->y_width;
+  int src_y_offset = srcy * src->y_stride + srcx;
+  int dst_y_offset = srcy * dst->y_stride + srcx;
+  int src_uv_offset = ((srcy * src->uv_stride) >> 1) + (srcx >> 1);
+  int dst_uv_offset = ((srcy * dst->uv_stride) >> 1) + (srcx >> 1);
 
-    /* If the side is not touching the bounder then don't extend. */
-    if (srcy)
-      et = 0;
-    if (srcx)
-      el = 0;
-    if (srcy + srch != src->y_height)
-      eb = 0;
-    if (srcx + srcw != src->y_width)
-      er = 0;
+  /* If the side is not touching the bounder then don't extend. */
+  if (srcy) et = 0;
+  if (srcx) el = 0;
+  if (srcy + srch != src->y_height) eb = 0;
+  if (srcx + srcw != src->y_width) er = 0;
 
-    copy_and_extend_plane(src->y_buffer + src_y_offset,
-                          src->y_stride,
-                          dst->y_buffer + dst_y_offset,
-                          dst->y_stride,
-                          srch, srcw,
-                          et, el, eb, er);
+  copy_and_extend_plane(src->y_buffer + src_y_offset, src->y_stride,
+                        dst->y_buffer + dst_y_offset, dst->y_stride, srch, srcw,
+                        et, el, eb, er);
 
-    et = (et + 1) >> 1;
-    el = (el + 1) >> 1;
-    eb = (eb + 1) >> 1;
-    er = (er + 1) >> 1;
-    srch = (srch + 1) >> 1;
-    srcw = (srcw + 1) >> 1;
+  et = (et + 1) >> 1;
+  el = (el + 1) >> 1;
+  eb = (eb + 1) >> 1;
+  er = (er + 1) >> 1;
+  srch = (srch + 1) >> 1;
+  srcw = (srcw + 1) >> 1;
 
-    copy_and_extend_plane(src->u_buffer + src_uv_offset,
-                          src->uv_stride,
-                          dst->u_buffer + dst_uv_offset,
-                          dst->uv_stride,
-                          srch, srcw,
-                          et, el, eb, er);
+  copy_and_extend_plane(src->u_buffer + src_uv_offset, src->uv_stride,
+                        dst->u_buffer + dst_uv_offset, dst->uv_stride, srch,
+                        srcw, et, el, eb, er);
 
-    copy_and_extend_plane(src->v_buffer + src_uv_offset,
-                          src->uv_stride,
-                          dst->v_buffer + dst_uv_offset,
-                          dst->uv_stride,
-                          srch, srcw,
-                          et, el, eb, er);
+  copy_and_extend_plane(src->v_buffer + src_uv_offset, src->uv_stride,
+                        dst->v_buffer + dst_uv_offset, dst->uv_stride, srch,
+                        srcw, et, el, eb, er);
 }
 
-
 /* note the extension is only for the last row, for intra prediction purpose */
-void vp8_extend_mb_row(YV12_BUFFER_CONFIG *ybf,
-                       unsigned char *YPtr,
-                       unsigned char *UPtr,
-                       unsigned char *VPtr)
-{
-    int i;
+void vp8_extend_mb_row(YV12_BUFFER_CONFIG *ybf, unsigned char *YPtr,
+                       unsigned char *UPtr, unsigned char *VPtr) {
+  int i;
 
-    YPtr += ybf->y_stride * 14;
-    UPtr += ybf->uv_stride * 6;
-    VPtr += ybf->uv_stride * 6;
+  YPtr += ybf->y_stride * 14;
+  UPtr += ybf->uv_stride * 6;
+  VPtr += ybf->uv_stride * 6;
 
-    for (i = 0; i < 4; i++)
-    {
-        YPtr[i] = YPtr[-1];
-        UPtr[i] = UPtr[-1];
-        VPtr[i] = VPtr[-1];
-    }
+  for (i = 0; i < 4; i++) {
+    YPtr[i] = YPtr[-1];
+    UPtr[i] = UPtr[-1];
+    VPtr[i] = VPtr[-1];
+  }
 
-    YPtr += ybf->y_stride;
-    UPtr += ybf->uv_stride;
-    VPtr += ybf->uv_stride;
+  YPtr += ybf->y_stride;
+  UPtr += ybf->uv_stride;
+  VPtr += ybf->uv_stride;
 
-    for (i = 0; i < 4; i++)
-    {
-        YPtr[i] = YPtr[-1];
-        UPtr[i] = UPtr[-1];
-        VPtr[i] = VPtr[-1];
-    }
+  for (i = 0; i < 4; i++) {
+    YPtr[i] = YPtr[-1];
+    UPtr[i] = UPtr[-1];
+    VPtr[i] = VPtr[-1];
+  }
 }
--- a/vp8/common/extend.h
+++ b/vp8/common/extend.h
@@ -8,7 +8,6 @@
  *  be found in the AUTHORS file in the root of the source tree.
  */
 
-
 #ifndef VP8_COMMON_EXTEND_H_
 #define VP8_COMMON_EXTEND_H_
 
@@ -18,13 +17,13 @@
 extern "C" {
 #endif
 
-void vp8_extend_mb_row(YV12_BUFFER_CONFIG *ybf, unsigned char *YPtr, unsigned char *UPtr, unsigned char *VPtr);
+void vp8_extend_mb_row(YV12_BUFFER_CONFIG *ybf, unsigned char *YPtr,
+                       unsigned char *UPtr, unsigned char *VPtr);
 void vp8_copy_and_extend_frame(YV12_BUFFER_CONFIG *src,
                                YV12_BUFFER_CONFIG *dst);
 void vp8_copy_and_extend_frame_with_rect(YV12_BUFFER_CONFIG *src,
-                                         YV12_BUFFER_CONFIG *dst,
-                                         int srcy, int srcx,
-                                         int srch, int srcw);
+                                         YV12_BUFFER_CONFIG *dst, int srcy,
+                                         int srcx, int srch, int srcw);
 
 #ifdef __cplusplus
 }  // extern "C"
--- a/vp8/common/filter.c
+++ b/vp8/common/filter.c
@@ -8,243 +8,186 @@
  *  be found in the AUTHORS file in the root of the source tree.
  */
 
-
 #include "filter.h"
 #include "./vp8_rtcd.h"
 
-DECLARE_ALIGNED(16, const short, vp8_bilinear_filters[8][2]) =
-{
-    { 128,   0 },
-    { 112,  16 },
-    {  96,  32 },
-    {  80,  48 },
-    {  64,  64 },
-    {  48,  80 },
-    {  32,  96 },
-    {  16, 112 }
+DECLARE_ALIGNED(16, const short, vp8_bilinear_filters[8][2]) = {
+  { 128, 0 }, { 112, 16 }, { 96, 32 }, { 80, 48 },
+  { 64, 64 }, { 48, 80 },  { 32, 96 }, { 16, 112 }
 };
 
-DECLARE_ALIGNED(16, const short, vp8_sub_pel_filters[8][6]) =
-{
+DECLARE_ALIGNED(16, const short, vp8_sub_pel_filters[8][6]) = {
 
-    { 0,  0,  128,    0,   0,  0 },         /* note that 1/8 pel positions are just as per alpha -0.5 bicubic */
-    { 0, -6,  123,   12,  -1,  0 },
-    { 2, -11, 108,   36,  -8,  1 },         /* New 1/4 pel 6 tap filter */
-    { 0, -9,   93,   50,  -6,  0 },
-    { 3, -16,  77,   77, -16,  3 },         /* New 1/2 pel 6 tap filter */
-    { 0, -6,   50,   93,  -9,  0 },
-    { 1, -8,   36,  108, -11,  2 },         /* New 1/4 pel 6 tap filter */
-    { 0, -1,   12,  123,  -6,  0 },
+  { 0, 0, 128, 0, 0,
+    0 }, /* note that 1/8 pel positions are just as per alpha -0.5 bicubic */
+  { 0, -6, 123, 12, -1, 0 },
+  { 2, -11, 108, 36, -8, 1 }, /* New 1/4 pel 6 tap filter */
+  { 0, -9, 93, 50, -6, 0 },
+  { 3, -16, 77, 77, -16, 3 }, /* New 1/2 pel 6 tap filter */
+  { 0, -6, 50, 93, -9, 0 },
+  { 1, -8, 36, 108, -11, 2 }, /* New 1/4 pel 6 tap filter */
+  { 0, -1, 12, 123, -6, 0 },
 };
 
-static void filter_block2d_first_pass
-(
-    unsigned char *src_ptr,
-    int *output_ptr,
-    unsigned int src_pixels_per_line,
-    unsigned int pixel_step,
-    unsigned int output_height,
-    unsigned int output_width,
-    const short *vp8_filter
-)
-{
-    unsigned int i, j;
-    int  Temp;
+static void filter_block2d_first_pass(unsigned char *src_ptr, int *output_ptr,
+                                      unsigned int src_pixels_per_line,
+                                      unsigned int pixel_step,
+                                      unsigned int output_height,
+                                      unsigned int output_width,
+                                      const short *vp8_filter) {
+  unsigned int i, j;
+  int Temp;
 
-    for (i = 0; i < output_height; i++)
-    {
-        for (j = 0; j < output_width; j++)
-        {
-            Temp = ((int)src_ptr[-2 * (int)pixel_step] * vp8_filter[0]) +
-                   ((int)src_ptr[-1 * (int)pixel_step] * vp8_filter[1]) +
-                   ((int)src_ptr[0]                 * vp8_filter[2]) +
-                   ((int)src_ptr[pixel_step]         * vp8_filter[3]) +
-                   ((int)src_ptr[2*pixel_step]       * vp8_filter[4]) +
-                   ((int)src_ptr[3*pixel_step]       * vp8_filter[5]) +
-                   (VP8_FILTER_WEIGHT >> 1);      /* Rounding */
+  for (i = 0; i < output_height; i++) {
+    for (j = 0; j < output_width; j++) {
+      Temp = ((int)src_ptr[-2 * (int)pixel_step] * vp8_filter[0]) +
+             ((int)src_ptr[-1 * (int)pixel_step] * vp8_filter[1]) +
+             ((int)src_ptr[0] * vp8_filter[2]) +
+             ((int)src_ptr[pixel_step] * vp8_filter[3]) +
+             ((int)src_ptr[2 * pixel_step] * vp8_filter[4]) +
+             ((int)src_ptr[3 * pixel_step] * vp8_filter[5]) +
+             (VP8_FILTER_WEIGHT >> 1); /* Rounding */
 
-            /* Normalize back to 0-255 */
-            Temp = Temp >> VP8_FILTER_SHIFT;
+      /* Normalize back to 0-255 */
+      Temp = Temp >> VP8_FILTER_SHIFT;
 
-            if (Temp < 0)
-                Temp = 0;
-            else if (Temp > 255)
-                Temp = 255;
+      if (Temp < 0)
+        Temp = 0;
+      else if (Temp > 255)
+        Temp = 255;
 
-            output_ptr[j] = Temp;
-            src_ptr++;
-        }
-
-        /* Next row... */
-        src_ptr    += src_pixels_per_line - output_width;
-        output_ptr += output_width;
+      output_ptr[j] = Temp;
+      src_ptr++;
     }
+
+    /* Next row... */
+    src_ptr += src_pixels_per_line - output_width;
+    output_ptr += output_width;
+  }
 }
 
-static void filter_block2d_second_pass
-(
-    int *src_ptr,
-    unsigned char *output_ptr,
-    int output_pitch,
-    unsigned int src_pixels_per_line,
-    unsigned int pixel_step,
-    unsigned int output_height,
-    unsigned int output_width,
-    const short *vp8_filter
-)
-{
-    unsigned int i, j;
-    int  Temp;
+static void filter_block2d_second_pass(int *src_ptr, unsigned char *output_ptr,
+                                       int output_pitch,
+                                       unsigned int src_pixels_per_line,
+                                       unsigned int pixel_step,
+                                       unsigned int output_height,
+                                       unsigned int output_width,
+                                       const short *vp8_filter) {
+  unsigned int i, j;
+  int Temp;
 
-    for (i = 0; i < output_height; i++)
-    {
-        for (j = 0; j < output_width; j++)
-        {
-            /* Apply filter */
-            Temp = ((int)src_ptr[-2 * (int)pixel_step] * vp8_filter[0]) +
-                   ((int)src_ptr[-1 * (int)pixel_step] * vp8_filter[1]) +
-                   ((int)src_ptr[0]                 * vp8_filter[2]) +
-                   ((int)src_ptr[pixel_step]         * vp8_filter[3]) +
-                   ((int)src_ptr[2*pixel_step]       * vp8_filter[4]) +
-                   ((int)src_ptr[3*pixel_step]       * vp8_filter[5]) +
-                   (VP8_FILTER_WEIGHT >> 1);   /* Rounding */
+  for (i = 0; i < output_height; i++) {
+    for (j = 0; j < output_width; j++) {
+      /* Apply filter */
+      Temp = ((int)src_ptr[-2 * (int)pixel_step] * vp8_filter[0]) +
+             ((int)src_ptr[-1 * (int)pixel_step] * vp8_filter[1]) +
+             ((int)src_ptr[0] * vp8_filter[2]) +
+             ((int)src_ptr[pixel_step] * vp8_filter[3]) +
+             ((int)src_ptr[2 * pixel_step] * vp8_filter[4]) +
+             ((int)src_ptr[3 * pixel_step] * vp8_filter[5]) +
+             (VP8_FILTER_WEIGHT >> 1); /* Rounding */
 
-            /* Normalize back to 0-255 */
-            Temp = Temp >> VP8_FILTER_SHIFT;
+      /* Normalize back to 0-255 */
+      Temp = Temp >> VP8_FILTER_SHIFT;
 
-            if (Temp < 0)
-                Temp = 0;
-            else if (Temp > 255)
-                Temp = 255;
+      if (Temp < 0)
+        Temp = 0;
+      else if (Temp > 255)
+        Temp = 255;
 
-            output_ptr[j] = (unsigned char)Temp;
-            src_ptr++;
-        }
-
-        /* Start next row */
-        src_ptr    += src_pixels_per_line - output_width;
-        output_ptr += output_pitch;
+      output_ptr[j] = (unsigned char)Temp;
+      src_ptr++;
     }
+
+    /* Start next row */
+    src_ptr += src_pixels_per_line - output_width;
+    output_ptr += output_pitch;
+  }
 }
 
+static void filter_block2d(unsigned char *src_ptr, unsigned char *output_ptr,
+                           unsigned int src_pixels_per_line, int output_pitch,
+                           const short *HFilter, const short *VFilter) {
+  int FData[9 * 4]; /* Temp data buffer used in filtering */
 
-static void filter_block2d
-(
-    unsigned char  *src_ptr,
-    unsigned char  *output_ptr,
-    unsigned int src_pixels_per_line,
-    int output_pitch,
-    const short  *HFilter,
-    const short  *VFilter
-)
-{
-    int FData[9*4]; /* Temp data buffer used in filtering */
+  /* First filter 1-D horizontally... */
+  filter_block2d_first_pass(src_ptr - (2 * src_pixels_per_line), FData,
+                            src_pixels_per_line, 1, 9, 4, HFilter);
 
-    /* First filter 1-D horizontally... */
-    filter_block2d_first_pass(src_ptr - (2 * src_pixels_per_line), FData, src_pixels_per_line, 1, 9, 4, HFilter);
-
-    /* then filter verticaly... */
-    filter_block2d_second_pass(FData + 8, output_ptr, output_pitch, 4, 4, 4, 4, VFilter);
+  /* then filter verticaly... */
+  filter_block2d_second_pass(FData + 8, output_ptr, output_pitch, 4, 4, 4, 4,
+                             VFilter);
 }
 
+void vp8_sixtap_predict4x4_c(unsigned char *src_ptr, int src_pixels_per_line,
+                             int xoffset, int yoffset, unsigned char *dst_ptr,
+                             int dst_pitch) {
+  const short *HFilter;
+  const short *VFilter;
 
-void vp8_sixtap_predict4x4_c
-(
-    unsigned char  *src_ptr,
-    int   src_pixels_per_line,
-    int  xoffset,
-    int  yoffset,
-    unsigned char *dst_ptr,
-    int dst_pitch
-)
-{
-    const short  *HFilter;
-    const short  *VFilter;
+  HFilter = vp8_sub_pel_filters[xoffset]; /* 6 tap */
+  VFilter = vp8_sub_pel_filters[yoffset]; /* 6 tap */
 
-    HFilter = vp8_sub_pel_filters[xoffset];   /* 6 tap */
-    VFilter = vp8_sub_pel_filters[yoffset];   /* 6 tap */
-
-    filter_block2d(src_ptr, dst_ptr, src_pixels_per_line, dst_pitch, HFilter, VFilter);
+  filter_block2d(src_ptr, dst_ptr, src_pixels_per_line, dst_pitch, HFilter,
+                 VFilter);
 }
-void vp8_sixtap_predict8x8_c
-(
-    unsigned char  *src_ptr,
-    int  src_pixels_per_line,
-    int  xoffset,
-    int  yoffset,
-    unsigned char *dst_ptr,
-    int  dst_pitch
-)
-{
-    const short  *HFilter;
-    const short  *VFilter;
-    int FData[13*16];   /* Temp data buffer used in filtering */
+void vp8_sixtap_predict8x8_c(unsigned char *src_ptr, int src_pixels_per_line,
+                             int xoffset, int yoffset, unsigned char *dst_ptr,
+                             int dst_pitch) {
+  const short *HFilter;
+  const short *VFilter;
+  int FData[13 * 16]; /* Temp data buffer used in filtering */
 
-    HFilter = vp8_sub_pel_filters[xoffset];   /* 6 tap */
-    VFilter = vp8_sub_pel_filters[yoffset];   /* 6 tap */
+  HFilter = vp8_sub_pel_filters[xoffset]; /* 6 tap */
+  VFilter = vp8_sub_pel_filters[yoffset]; /* 6 tap */
 
-    /* First filter 1-D horizontally... */
-    filter_block2d_first_pass(src_ptr - (2 * src_pixels_per_line), FData, src_pixels_per_line, 1, 13, 8, HFilter);
+  /* First filter 1-D horizontally... */
+  filter_block2d_first_pass(src_ptr - (2 * src_pixels_per_line), FData,
+                            src_pixels_per_line, 1, 13, 8, HFilter);
 
-
-    /* then filter verticaly... */
-    filter_block2d_second_pass(FData + 16, dst_ptr, dst_pitch, 8, 8, 8, 8, VFilter);
-
+  /* then filter verticaly... */
+  filter_block2d_second_pass(FData + 16, dst_ptr, dst_pitch, 8, 8, 8, 8,
+                             VFilter);
 }
 
-void vp8_sixtap_predict8x4_c
-(
-    unsigned char  *src_ptr,
-    int  src_pixels_per_line,
-    int  xoffset,
-    int  yoffset,
-    unsigned char *dst_ptr,
-    int  dst_pitch
-)
-{
-    const short  *HFilter;
-    const short  *VFilter;
-    int FData[13*16];   /* Temp data buffer used in filtering */
+void vp8_sixtap_predict8x4_c(unsigned char *src_ptr, int src_pixels_per_line,
+                             int xoffset, int yoffset, unsigned char *dst_ptr,
+                             int dst_pitch) {
+  const short *HFilter;
+  const short *VFilter;
+  int FData[13 * 16]; /* Temp data buffer used in filtering */
 
-    HFilter = vp8_sub_pel_filters[xoffset];   /* 6 tap */
-    VFilter = vp8_sub_pel_filters[yoffset];   /* 6 tap */
+  HFilter = vp8_sub_pel_filters[xoffset]; /* 6 tap */
+  VFilter = vp8_sub_pel_filters[yoffset]; /* 6 tap */
 
-    /* First filter 1-D horizontally... */
-    filter_block2d_first_pass(src_ptr - (2 * src_pixels_per_line), FData, src_pixels_per_line, 1, 9, 8, HFilter);
+  /* First filter 1-D horizontally... */
+  filter_block2d_first_pass(src_ptr - (2 * src_pixels_per_line), FData,
+                            src_pixels_per_line, 1, 9, 8, HFilter);
 
-
-    /* then filter verticaly... */
-    filter_block2d_second_pass(FData + 16, dst_ptr, dst_pitch, 8, 8, 4, 8, VFilter);
-
+  /* then filter verticaly... */
+  filter_block2d_second_pass(FData + 16, dst_ptr, dst_pitch, 8, 8, 4, 8,
+                             VFilter);
 }
 
-void vp8_sixtap_predict16x16_c
-(
-    unsigned char  *src_ptr,
-    int  src_pixels_per_line,
-    int  xoffset,
-    int  yoffset,
-    unsigned char *dst_ptr,
-    int  dst_pitch
-)
-{
-    const short  *HFilter;
-    const short  *VFilter;
-    int FData[21*24];   /* Temp data buffer used in filtering */
+void vp8_sixtap_predict16x16_c(unsigned char *src_ptr, int src_pixels_per_line,
+                               int xoffset, int yoffset, unsigned char *dst_ptr,
+                               int dst_pitch) {
+  const short *HFilter;
+  const short *VFilter;
+  int FData[21 * 24]; /* Temp data buffer used in filtering */
 
+  HFilter = vp8_sub_pel_filters[xoffset]; /* 6 tap */
+  VFilter = vp8_sub_pel_filters[yoffset]; /* 6 tap */
 
-    HFilter = vp8_sub_pel_filters[xoffset];   /* 6 tap */
-    VFilter = vp8_sub_pel_filters[yoffset];   /* 6 tap */
+  /* First filter 1-D horizontally... */
+  filter_block2d_first_pass(src_ptr - (2 * src_pixels_per_line), FData,
+                            src_pixels_per_line, 1, 21, 16, HFilter);
 
-    /* First filter 1-D horizontally... */
-    filter_block2d_first_pass(src_ptr - (2 * src_pixels_per_line), FData, src_pixels_per_line, 1, 21, 16, HFilter);
-
-    /* then filter verticaly... */
-    filter_block2d_second_pass(FData + 32, dst_ptr, dst_pitch, 16, 16, 16, 16, VFilter);
-
+  /* then filter verticaly... */
+  filter_block2d_second_pass(FData + 32, dst_ptr, dst_pitch, 16, 16, 16, 16,
+                             VFilter);
 }
 
-
 /****************************************************************************
  *
  *  ROUTINE       : filter_block2d_bil_first_pass
@@ -267,33 +210,25 @@
  *                  Two filter taps should sum to VP8_FILTER_WEIGHT.
  *
  ****************************************************************************/
-static void filter_block2d_bil_first_pass
-(
-    unsigned char  *src_ptr,
-    unsigned short *dst_ptr,
-    unsigned int    src_stride,
-    unsigned int    height,
-    unsigned int    width,
-    const short    *vp8_filter
-)
-{
-    unsigned int i, j;
+static void filter_block2d_bil_first_pass(
+    unsigned char *src_ptr, unsigned short *dst_ptr, unsigned int src_stride,
+    unsigned int height, unsigned int width, const short *vp8_filter) {
+  unsigned int i, j;
 
-    for (i = 0; i < height; i++)
-    {
-        for (j = 0; j < width; j++)
-        {
-            /* Apply bilinear filter */
-            dst_ptr[j] = (((int)src_ptr[0] * vp8_filter[0]) +
-                          ((int)src_ptr[1] * vp8_filter[1]) +
-                          (VP8_FILTER_WEIGHT / 2)) >> VP8_FILTER_SHIFT;
-            src_ptr++;
-        }
-
-        /* Next row... */
-        src_ptr += src_stride - width;
-        dst_ptr += width;
+  for (i = 0; i < height; i++) {
+    for (j = 0; j < width; j++) {
+      /* Apply bilinear filter */
+      dst_ptr[j] =
+          (((int)src_ptr[0] * vp8_filter[0]) +
+           ((int)src_ptr[1] * vp8_filter[1]) + (VP8_FILTER_WEIGHT / 2)) >>
+          VP8_FILTER_SHIFT;
+      src_ptr++;
     }
+
+    /* Next row... */
+    src_ptr += src_stride - width;
+    dst_ptr += width;
+  }
 }
 
 /****************************************************************************
@@ -312,43 +247,36 @@
  *
  *  FUNCTION      : Applies a 1-D 2-tap bi-linear filter to the source block
  *                  in the vertical direction to produce the filtered output
- *                  block. Used to implement second-pass of 2-D separable filter.
+ *                  block. Used to implement second-pass of 2-D separable
+ *                  filter.
  *
- *  SPECIAL NOTES : Requires 32-bit input as produced by filter_block2d_bil_first_pass.
+ *  SPECIAL NOTES : Requires 32-bit input as produced by
+ *                  filter_block2d_bil_first_pass.
  *                  Two filter taps should sum to VP8_FILTER_WEIGHT.
  *
  ****************************************************************************/
-static void filter_block2d_bil_second_pass
-(
-    unsigned short *src_ptr,
-    unsigned char  *dst_ptr,
-    int             dst_pitch,
-    unsigned int    height,
-    unsigned int    width,
-    const short    *vp8_filter
-)
-{
-    unsigned int  i, j;
-    int  Temp;
+static void filter_block2d_bil_second_pass(unsigned short *src_ptr,
+                                           unsigned char *dst_ptr,
+                                           int dst_pitch, unsigned int height,
+                                           unsigned int width,
+                                           const short *vp8_filter) {
+  unsigned int i, j;
+  int Temp;
 
-    for (i = 0; i < height; i++)
-    {
-        for (j = 0; j < width; j++)
-        {
-            /* Apply filter */
-            Temp = ((int)src_ptr[0]     * vp8_filter[0]) +
-                   ((int)src_ptr[width] * vp8_filter[1]) +
-                   (VP8_FILTER_WEIGHT / 2);
-            dst_ptr[j] = (unsigned int)(Temp >> VP8_FILTER_SHIFT);
-            src_ptr++;
-        }
-
-        /* Next row... */
-        dst_ptr += dst_pitch;
+  for (i = 0; i < height; i++) {
+    for (j = 0; j < width; j++) {
+      /* Apply filter */
+      Temp = ((int)src_ptr[0] * vp8_filter[0]) +
+             ((int)src_ptr[width] * vp8_filter[1]) + (VP8_FILTER_WEIGHT / 2);
+      dst_ptr[j] = (unsigned int)(Temp >> VP8_FILTER_SHIFT);
+      src_ptr++;
     }
+
+    /* Next row... */
+    dst_ptr += dst_pitch;
+  }
 }
 
-
 /****************************************************************************
  *
  *  ROUTINE       : filter_block2d_bil
@@ -356,7 +284,8 @@
  *  INPUTS        : UINT8  *src_ptr          : Pointer to source block.
  *                  UINT32  src_pitch        : Stride of source block.
  *                  UINT32  dst_pitch        : Stride of destination block.
- *                  INT32  *HFilter          : Array of 2 horizontal filter taps.
+ *                  INT32  *HFilter          : Array of 2 horizontal filter
+ *                                             taps.
  *                  INT32  *VFilter          : Array of 2 vertical filter taps.
  *                  INT32  Width             : Block width
  *                  INT32  Height            : Block height
@@ -372,44 +301,29 @@
  *  SPECIAL NOTES : The largest block size can be handled here is 16x16
  *
  ****************************************************************************/
-static void filter_block2d_bil
-(
-    unsigned char *src_ptr,
-    unsigned char *dst_ptr,
-    unsigned int   src_pitch,
-    unsigned int   dst_pitch,
-    const short   *HFilter,
-    const short   *VFilter,
-    int            Width,
-    int            Height
-)
-{
+static void filter_block2d_bil(unsigned char *src_ptr, unsigned char *dst_ptr,
+                               unsigned int src_pitch, unsigned int dst_pitch,
+                               const short *HFilter, const short *VFilter,
+                               int Width, int Height) {
+  unsigned short FData[17 * 16]; /* Temp data buffer used in filtering */
 
-    unsigned short FData[17*16];    /* Temp data buffer used in filtering */
+  /* First filter 1-D horizontally... */
+  filter_block2d_bil_first_pass(src_ptr, FData, src_pitch, Height + 1, Width,
+                                HFilter);
 
-    /* First filter 1-D horizontally... */
-    filter_block2d_bil_first_pass(src_ptr, FData, src_pitch, Height + 1, Width, HFilter);
-
-    /* then 1-D vertically... */
-    filter_block2d_bil_second_pass(FData, dst_ptr, dst_pitch, Height, Width, VFilter);
+  /* then 1-D vertically... */
+  filter_block2d_bil_second_pass(FData, dst_ptr, dst_pitch, Height, Width,
+                                 VFilter);
 }
 
+void vp8_bilinear_predict4x4_c(unsigned char *src_ptr, int src_pixels_per_line,
+                               int xoffset, int yoffset, unsigned char *dst_ptr,
+                               int dst_pitch) {
+  const short *HFilter;
+  const short *VFilter;
 
-void vp8_bilinear_predict4x4_c
-(
-    unsigned char  *src_ptr,
-    int   src_pixels_per_line,
-    int  xoffset,
-    int  yoffset,
-    unsigned char *dst_ptr,
-    int dst_pitch
-)
-{
-    const short *HFilter;
-    const short *VFilter;
-
-    HFilter = vp8_bilinear_filters[xoffset];
-    VFilter = vp8_bilinear_filters[yoffset];
+  HFilter = vp8_bilinear_filters[xoffset];
+  VFilter = vp8_bilinear_filters[yoffset];
 #if 0
     {
         int i;
@@ -429,65 +343,46 @@
         }
     }
 #endif
-    filter_block2d_bil(src_ptr, dst_ptr, src_pixels_per_line, dst_pitch, HFilter, VFilter, 4, 4);
-
+  filter_block2d_bil(src_ptr, dst_ptr, src_pixels_per_line, dst_pitch, HFilter,
+                     VFilter, 4, 4);
 }
 
-void vp8_bilinear_predict8x8_c
-(
-    unsigned char  *src_ptr,
-    int  src_pixels_per_line,
-    int  xoffset,
-    int  yoffset,
-    unsigned char *dst_ptr,
-    int  dst_pitch
-)
-{
-    const short *HFilter;
-    const short *VFilter;
+void vp8_bilinear_predict8x8_c(unsigned char *src_ptr, int src_pixels_per_line,
+                               int xoffset, int yoffset, unsigned char *dst_ptr,
+                               int dst_pitch) {
+  const short *HFilter;
+  const short *VFilter;
 
-    HFilter = vp8_bilinear_filters[xoffset];
-    VFilter = vp8_bilinear_filters[yoffset];
+  HFilter = vp8_bilinear_filters[xoffset];
+  VFilter = vp8_bilinear_filters[yoffset];
 
-    filter_block2d_bil(src_ptr, dst_ptr, src_pixels_per_line, dst_pitch, HFilter, VFilter, 8, 8);
-
+  filter_block2d_bil(src_ptr, dst_ptr, src_pixels_per_line, dst_pitch, HFilter,
+                     VFilter, 8, 8);
 }
 
-void vp8_bilinear_predict8x4_c
-(
-    unsigned char  *src_ptr,
-    int  src_pixels_per_line,
-    int  xoffset,
-    int  yoffset,
-    unsigned char *dst_ptr,
-    int  dst_pitch
-)
-{
-    const short *HFilter;
-    const short *VFilter;
+void vp8_bilinear_predict8x4_c(unsigned char *src_ptr, int src_pixels_per_line,
+                               int xoffset, int yoffset, unsigned char *dst_ptr,
+                               int dst_pitch) {
+  const short *HFilter;
+  const short *VFilter;
 
-    HFilter = vp8_bilinear_filters[xoffset];
-    VFilter = vp8_bilinear_filters[yoffset];
+  HFilter = vp8_bilinear_filters[xoffset];
+  VFilter = vp8_bilinear_filters[yoffset];
 
-    filter_block2d_bil(src_ptr, dst_ptr, src_pixels_per_line, dst_pitch, HFilter, VFilter, 8, 4);
-
+  filter_block2d_bil(src_ptr, dst_ptr, src_pixels_per_line, dst_pitch, HFilter,
+                     VFilter, 8, 4);
 }
 
-void vp8_bilinear_predict16x16_c
-(
-    unsigned char  *src_ptr,
-    int  src_pixels_per_line,
-    int  xoffset,
-    int  yoffset,
-    unsigned char *dst_ptr,
-    int  dst_pitch
-)
-{
-    const short *HFilter;
-    const short *VFilter;
+void vp8_bilinear_predict16x16_c(unsigned char *src_ptr,
+                                 int src_pixels_per_line, int xoffset,
+                                 int yoffset, unsigned char *dst_ptr,
+                                 int dst_pitch) {
+  const short *HFilter;
+  const short *VFilter;
 
-    HFilter = vp8_bilinear_filters[xoffset];
-    VFilter = vp8_bilinear_filters[yoffset];
+  HFilter = vp8_bilinear_filters[xoffset];
+  VFilter = vp8_bilinear_filters[yoffset];
 
-    filter_block2d_bil(src_ptr, dst_ptr, src_pixels_per_line, dst_pitch, HFilter, VFilter, 16, 16);
+  filter_block2d_bil(src_ptr, dst_ptr, src_pixels_per_line, dst_pitch, HFilter,
+                     VFilter, 16, 16);
 }
--- a/vp8/common/filter.h
+++ b/vp8/common/filter.h
@@ -8,7 +8,6 @@
  *  be found in the AUTHORS file in the root of the source tree.
  */
 
-
 #ifndef VP8_COMMON_FILTER_H_
 #define VP8_COMMON_FILTER_H_
 
@@ -20,7 +19,7 @@
 
 #define BLOCK_HEIGHT_WIDTH 4
 #define VP8_FILTER_WEIGHT 128
-#define VP8_FILTER_SHIFT  7
+#define VP8_FILTER_SHIFT 7
 
 extern DECLARE_ALIGNED(16, const short, vp8_bilinear_filters[8][2]);
 extern DECLARE_ALIGNED(16, const short, vp8_sub_pel_filters[8][6]);
--- a/vp8/common/findnearmv.c
+++ b/vp8/common/findnearmv.c
@@ -8,186 +8,147 @@
  *  be found in the AUTHORS file in the root of the source tree.
  */
 
-
 #include "findnearmv.h"
 
 const unsigned char vp8_mbsplit_offset[4][16] = {
-    { 0,  8,  0,  0,  0,  0,  0,  0,  0,  0,   0,  0,  0,  0,  0,  0},
-    { 0,  2,  0,  0,  0,  0,  0,  0,  0,  0,   0,  0,  0,  0,  0,  0},
-    { 0,  2,  8, 10,  0,  0,  0,  0,  0,  0,   0,  0,  0,  0,  0,  0},
-    { 0,  1,  2,  3,  4,  5,  6,  7,  8,  9,  10, 11, 12, 13, 14, 15}
+  { 0, 8, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 },
+  { 0, 2, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 },
+  { 0, 2, 8, 10, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 },
+  { 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15 }
 };
 
 /* Predict motion vectors using those from already-decoded nearby blocks.
    Note that we only consider one 4x4 subblock from each candidate 16x16
    macroblock.   */
-void vp8_find_near_mvs
-(
-    MACROBLOCKD *xd,
-    const MODE_INFO *here,
-    int_mv *nearest,
-    int_mv *nearby,
-    int_mv *best_mv,
-    int cnt[4],
-    int refframe,
-    int *ref_frame_sign_bias
-)
-{
-    const MODE_INFO *above = here - xd->mode_info_stride;
-    const MODE_INFO *left = here - 1;
-    const MODE_INFO *aboveleft = above - 1;
-    int_mv            near_mvs[4];
-    int_mv           *mv = near_mvs;
-    int             *cntx = cnt;
-    enum {CNT_INTRA, CNT_NEAREST, CNT_NEAR, CNT_SPLITMV};
+void vp8_find_near_mvs(MACROBLOCKD *xd, const MODE_INFO *here, int_mv *nearest,
+                       int_mv *nearby, int_mv *best_mv, int cnt[4],
+                       int refframe, int *ref_frame_sign_bias) {
+  const MODE_INFO *above = here - xd->mode_info_stride;
+  const MODE_INFO *left = here - 1;
+  const MODE_INFO *aboveleft = above - 1;
+  int_mv near_mvs[4];
+  int_mv *mv = near_mvs;
+  int *cntx = cnt;
+  enum { CNT_INTRA, CNT_NEAREST, CNT_NEAR, CNT_SPLITMV };
 
-    /* Zero accumulators */
-    mv[0].as_int = mv[1].as_int = mv[2].as_int = 0;
-    cnt[0] = cnt[1] = cnt[2] = cnt[3] = 0;
+  /* Zero accumulators */
+  mv[0].as_int = mv[1].as_int = mv[2].as_int = 0;
+  cnt[0] = cnt[1] = cnt[2] = cnt[3] = 0;
 
-    /* Process above */
-    if (above->mbmi.ref_frame != INTRA_FRAME)
-    {
-        if (above->mbmi.mv.as_int)
-        {
-            (++mv)->as_int = above->mbmi.mv.as_int;
-            mv_bias(ref_frame_sign_bias[above->mbmi.ref_frame], refframe, mv, ref_frame_sign_bias);
-            ++cntx;
-        }
-
-        *cntx += 2;
+  /* Process above */
+  if (above->mbmi.ref_frame != INTRA_FRAME) {
+    if (above->mbmi.mv.as_int) {
+      (++mv)->as_int = above->mbmi.mv.as_int;
+      mv_bias(ref_frame_sign_bias[above->mbmi.ref_frame], refframe, mv,
+              ref_frame_sign_bias);
+      ++cntx;
     }
 
-    /* Process left */
-    if (left->mbmi.ref_frame != INTRA_FRAME)
-    {
-        if (left->mbmi.mv.as_int)
-        {
-            int_mv this_mv;
+    *cntx += 2;
+  }
 
-            this_mv.as_int = left->mbmi.mv.as_int;
-            mv_bias(ref_frame_sign_bias[left->mbmi.ref_frame], refframe, &this_mv, ref_frame_sign_bias);
+  /* Process left */
+  if (left->mbmi.ref_frame != INTRA_FRAME) {
+    if (left->mbmi.mv.as_int) {
+      int_mv this_mv;
 
-            if (this_mv.as_int != mv->as_int)
-            {
-                (++mv)->as_int = this_mv.as_int;
-                ++cntx;
-            }
+      this_mv.as_int = left->mbmi.mv.as_int;
+      mv_bias(ref_frame_sign_bias[left->mbmi.ref_frame], refframe, &this_mv,
+              ref_frame_sign_bias);
 
-            *cntx += 2;
-        }
-        else
-            cnt[CNT_INTRA] += 2;
-    }
+      if (this_mv.as_int != mv->as_int) {
+        (++mv)->as_int = this_mv.as_int;
+        ++cntx;
+      }
 
-    /* Process above left */
-    if (aboveleft->mbmi.ref_frame != INTRA_FRAME)
-    {
-        if (aboveleft->mbmi.mv.as_int)
-        {
-            int_mv this_mv;
+      *cntx += 2;
+    } else
+      cnt[CNT_INTRA] += 2;
+  }
 
-            this_mv.as_int = aboveleft->mbmi.mv.as_int;
-            mv_bias(ref_frame_sign_bias[aboveleft->mbmi.ref_frame], refframe, &this_mv, ref_frame_sign_bias);
+  /* Process above left */
+  if (aboveleft->mbmi.ref_frame != INTRA_FRAME) {
+    if (aboveleft->mbmi.mv.as_int) {
+      int_mv this_mv;
 
-            if (this_mv.as_int != mv->as_int)
-            {
-                (++mv)->as_int = this_mv.as_int;
-                ++cntx;
-            }
+      this_mv.as_int = aboveleft->mbmi.mv.as_int;
+      mv_bias(ref_frame_sign_bias[aboveleft->mbmi.ref_frame], refframe,
+              &this_mv, ref_frame_sign_bias);
 
-            *cntx += 1;
-        }
-        else
-            cnt[CNT_INTRA] += 1;
-    }
+      if (this_mv.as_int != mv->as_int) {
+        (++mv)->as_int = this_mv.as_int;
+        ++cntx;
+      }
 
-    /* If we have three distinct MV's ... */
-    if (cnt[CNT_SPLITMV])
-    {
-        /* See if above-left MV can be merged with NEAREST */
-        if (mv->as_int == near_mvs[CNT_NEAREST].as_int)
-            cnt[CNT_NEAREST] += 1;
-    }
+      *cntx += 1;
+    } else
+      cnt[CNT_INTRA] += 1;
+  }
 
-    cnt[CNT_SPLITMV] = ((above->mbmi.mode == SPLITMV)
-                        + (left->mbmi.mode == SPLITMV)) * 2
-                       + (aboveleft->mbmi.mode == SPLITMV);
+  /* If we have three distinct MV's ... */
+  if (cnt[CNT_SPLITMV]) {
+    /* See if above-left MV can be merged with NEAREST */
+    if (mv->as_int == near_mvs[CNT_NEAREST].as_int) cnt[CNT_NEAREST] += 1;
+  }
 
-    /* Swap near and nearest if necessary */
-    if (cnt[CNT_NEAR] > cnt[CNT_NEAREST])
-    {
-        int tmp;
-        tmp = cnt[CNT_NEAREST];
-        cnt[CNT_NEAREST] = cnt[CNT_NEAR];
-        cnt[CNT_NEAR] = tmp;
-        tmp = near_mvs[CNT_NEAREST].as_int;
-        near_mvs[CNT_NEAREST].as_int = near_mvs[CNT_NEAR].as_int;
-        near_mvs[CNT_NEAR].as_int = tmp;
-    }
+  cnt[CNT_SPLITMV] =
+      ((above->mbmi.mode == SPLITMV) + (left->mbmi.mode == SPLITMV)) * 2 +
+      (aboveleft->mbmi.mode == SPLITMV);
 
-    /* Use near_mvs[0] to store the "best" MV */
-    if (cnt[CNT_NEAREST] >= cnt[CNT_INTRA])
-        near_mvs[CNT_INTRA] = near_mvs[CNT_NEAREST];
+  /* Swap near and nearest if necessary */
+  if (cnt[CNT_NEAR] > cnt[CNT_NEAREST]) {
+    int tmp;
+    tmp = cnt[CNT_NEAREST];
+    cnt[CNT_NEAREST] = cnt[CNT_NEAR];
+    cnt[CNT_NEAR] = tmp;
+    tmp = near_mvs[CNT_NEAREST].as_int;
+    near_mvs[CNT_NEAREST].as_int = near_mvs[CNT_NEAR].as_int;
+    near_mvs[CNT_NEAR].as_int = tmp;
+  }
 
-    /* Set up return values */
-    best_mv->as_int = near_mvs[0].as_int;
-    nearest->as_int = near_mvs[CNT_NEAREST].as_int;
-    nearby->as_int = near_mvs[CNT_NEAR].as_int;
+  /* Use near_mvs[0] to store the "best" MV */
+  if (cnt[CNT_NEAREST] >= cnt[CNT_INTRA])
+    near_mvs[CNT_INTRA] = near_mvs[CNT_NEAREST];
+
+  /* Set up return values */
+  best_mv->as_int = near_mvs[0].as_int;
+  nearest->as_int = near_mvs[CNT_NEAREST].as_int;
+  nearby->as_int = near_mvs[CNT_NEAR].as_int;
 }
 
-
-static void invert_and_clamp_mvs(int_mv *inv, int_mv *src, MACROBLOCKD *xd)
-{
-    inv->as_mv.row = src->as_mv.row * -1;
-    inv->as_mv.col = src->as_mv.col * -1;
-    vp8_clamp_mv2(inv, xd);
-    vp8_clamp_mv2(src, xd);
+static void invert_and_clamp_mvs(int_mv *inv, int_mv *src, MACROBLOCKD *xd) {
+  inv->as_mv.row = src->as_mv.row * -1;
+  inv->as_mv.col = src->as_mv.col * -1;
+  vp8_clamp_mv2(inv, xd);
+  vp8_clamp_mv2(src, xd);
 }
 
+int vp8_find_near_mvs_bias(MACROBLOCKD *xd, const MODE_INFO *here,
+                           int_mv mode_mv_sb[2][MB_MODE_COUNT],
+                           int_mv best_mv_sb[2], int cnt[4], int refframe,
+                           int *ref_frame_sign_bias) {
+  int sign_bias = ref_frame_sign_bias[refframe];
 
-int vp8_find_near_mvs_bias
-(
-    MACROBLOCKD *xd,
-    const MODE_INFO *here,
-    int_mv mode_mv_sb[2][MB_MODE_COUNT],
-    int_mv best_mv_sb[2],
-    int cnt[4],
-    int refframe,
-    int *ref_frame_sign_bias
-)
-{
-    int sign_bias = ref_frame_sign_bias[refframe];
+  vp8_find_near_mvs(xd, here, &mode_mv_sb[sign_bias][NEARESTMV],
+                    &mode_mv_sb[sign_bias][NEARMV], &best_mv_sb[sign_bias], cnt,
+                    refframe, ref_frame_sign_bias);
 
-    vp8_find_near_mvs(xd,
-                      here,
-                      &mode_mv_sb[sign_bias][NEARESTMV],
-                      &mode_mv_sb[sign_bias][NEARMV],
-                      &best_mv_sb[sign_bias],
-                      cnt,
-                      refframe,
-                      ref_frame_sign_bias);
+  invert_and_clamp_mvs(&mode_mv_sb[!sign_bias][NEARESTMV],
+                       &mode_mv_sb[sign_bias][NEARESTMV], xd);
+  invert_and_clamp_mvs(&mode_mv_sb[!sign_bias][NEARMV],
+                       &mode_mv_sb[sign_bias][NEARMV], xd);
+  invert_and_clamp_mvs(&best_mv_sb[!sign_bias], &best_mv_sb[sign_bias], xd);
 
-    invert_and_clamp_mvs(&mode_mv_sb[!sign_bias][NEARESTMV],
-                         &mode_mv_sb[sign_bias][NEARESTMV], xd);
-    invert_and_clamp_mvs(&mode_mv_sb[!sign_bias][NEARMV],
-                         &mode_mv_sb[sign_bias][NEARMV], xd);
-    invert_and_clamp_mvs(&best_mv_sb[!sign_bias],
-                         &best_mv_sb[sign_bias], xd);
-
-    return sign_bias;
+  return sign_bias;
 }
 
-
-vp8_prob *vp8_mv_ref_probs(
-    vp8_prob p[VP8_MVREFS-1], const int near_mv_ref_ct[4]
-)
-{
-    p[0] = vp8_mode_contexts [near_mv_ref_ct[0]] [0];
-    p[1] = vp8_mode_contexts [near_mv_ref_ct[1]] [1];
-    p[2] = vp8_mode_contexts [near_mv_ref_ct[2]] [2];
-    p[3] = vp8_mode_contexts [near_mv_ref_ct[3]] [3];
-    /*p[3] = vp8_mode_contexts [near_mv_ref_ct[1] + near_mv_ref_ct[2] + near_mv_ref_ct[3]] [3];*/
-    return p;
+vp8_prob *vp8_mv_ref_probs(vp8_prob p[VP8_MVREFS - 1],
+                           const int near_mv_ref_ct[4]) {
+  p[0] = vp8_mode_contexts[near_mv_ref_ct[0]][0];
+  p[1] = vp8_mode_contexts[near_mv_ref_ct[1]][1];
+  p[2] = vp8_mode_contexts[near_mv_ref_ct[2]][2];
+  p[3] = vp8_mode_contexts[near_mv_ref_ct[3]][3];
+  /* p[3] = vp8_mode_contexts[near_mv_ref_ct[1] + near_mv_ref_ct[2] +
+                           near_mv_ref_ct[3]][3]; */
+  return p;
 }
-
--- a/vp8/common/findnearmv.h
+++ b/vp8/common/findnearmv.h
@@ -8,7 +8,6 @@
  *  be found in the AUTHORS file in the root of the source tree.
  */
 
-
 #ifndef VP8_COMMON_FINDNEARMV_H_
 #define VP8_COMMON_FINDNEARMV_H_
 
@@ -22,170 +21,125 @@
 extern "C" {
 #endif
 
-
 static INLINE void mv_bias(int refmb_ref_frame_sign_bias, int refframe,
-                           int_mv *mvp, const int *ref_frame_sign_bias)
-{
-    if (refmb_ref_frame_sign_bias != ref_frame_sign_bias[refframe])
-    {
-        mvp->as_mv.row *= -1;
-        mvp->as_mv.col *= -1;
-    }
+                           int_mv *mvp, const int *ref_frame_sign_bias) {
+  if (refmb_ref_frame_sign_bias != ref_frame_sign_bias[refframe]) {
+    mvp->as_mv.row *= -1;
+    mvp->as_mv.col *= -1;
+  }
 }
 
 #define LEFT_TOP_MARGIN (16 << 3)
 #define RIGHT_BOTTOM_MARGIN (16 << 3)
-static INLINE void vp8_clamp_mv2(int_mv *mv, const MACROBLOCKD *xd)
-{
-    if (mv->as_mv.col < (xd->mb_to_left_edge - LEFT_TOP_MARGIN))
-        mv->as_mv.col = xd->mb_to_left_edge - LEFT_TOP_MARGIN;
-    else if (mv->as_mv.col > xd->mb_to_right_edge + RIGHT_BOTTOM_MARGIN)
-        mv->as_mv.col = xd->mb_to_right_edge + RIGHT_BOTTOM_MARGIN;
+static INLINE void vp8_clamp_mv2(int_mv *mv, const MACROBLOCKD *xd) {
+  if (mv->as_mv.col < (xd->mb_to_left_edge - LEFT_TOP_MARGIN))
+    mv->as_mv.col = xd->mb_to_left_edge - LEFT_TOP_MARGIN;
+  else if (mv->as_mv.col > xd->mb_to_right_edge + RIGHT_BOTTOM_MARGIN)
+    mv->as_mv.col = xd->mb_to_right_edge + RIGHT_BOTTOM_MARGIN;
 
-    if (mv->as_mv.row < (xd->mb_to_top_edge - LEFT_TOP_MARGIN))
-        mv->as_mv.row = xd->mb_to_top_edge - LEFT_TOP_MARGIN;
-    else if (mv->as_mv.row > xd->mb_to_bottom_edge + RIGHT_BOTTOM_MARGIN)
-        mv->as_mv.row = xd->mb_to_bottom_edge + RIGHT_BOTTOM_MARGIN;
+  if (mv->as_mv.row < (xd->mb_to_top_edge - LEFT_TOP_MARGIN))
+    mv->as_mv.row = xd->mb_to_top_edge - LEFT_TOP_MARGIN;
+  else if (mv->as_mv.row > xd->mb_to_bottom_edge + RIGHT_BOTTOM_MARGIN)
+    mv->as_mv.row = xd->mb_to_bottom_edge + RIGHT_BOTTOM_MARGIN;
 }
 
 static INLINE void vp8_clamp_mv(int_mv *mv, int mb_to_left_edge,
                                 int mb_to_right_edge, int mb_to_top_edge,
-                                int mb_to_bottom_edge)
-{
-    mv->as_mv.col = (mv->as_mv.col < mb_to_left_edge) ?
-        mb_to_left_edge : mv->as_mv.col;
-    mv->as_mv.col = (mv->as_mv.col > mb_to_right_edge) ?
-        mb_to_right_edge : mv->as_mv.col;
-    mv->as_mv.row = (mv->as_mv.row < mb_to_top_edge) ?
-        mb_to_top_edge : mv->as_mv.row;
-    mv->as_mv.row = (mv->as_mv.row > mb_to_bottom_edge) ?
-        mb_to_bottom_edge : mv->as_mv.row;
+                                int mb_to_bottom_edge) {
+  mv->as_mv.col =
+      (mv->as_mv.col < mb_to_left_edge) ? mb_to_left_edge : mv->as_mv.col;
+  mv->as_mv.col =
+      (mv->as_mv.col > mb_to_right_edge) ? mb_to_right_edge : mv->as_mv.col;
+  mv->as_mv.row =
+      (mv->as_mv.row < mb_to_top_edge) ? mb_to_top_edge : mv->as_mv.row;
+  mv->as_mv.row =
+      (mv->as_mv.row > mb_to_bottom_edge) ? mb_to_bottom_edge : mv->as_mv.row;
 }
 static INLINE unsigned int vp8_check_mv_bounds(int_mv *mv, int mb_to_left_edge,
                                                int mb_to_right_edge,
                                                int mb_to_top_edge,
-                                               int mb_to_bottom_edge)
-{
-    unsigned int need_to_clamp;
-    need_to_clamp = (mv->as_mv.col < mb_to_left_edge);
-    need_to_clamp |= (mv->as_mv.col > mb_to_right_edge);
-    need_to_clamp |= (mv->as_mv.row < mb_to_top_edge);
-    need_to_clamp |= (mv->as_mv.row > mb_to_bottom_edge);
-    return need_to_clamp;
+                                               int mb_to_bottom_edge) {
+  unsigned int need_to_clamp;
+  need_to_clamp = (mv->as_mv.col < mb_to_left_edge);
+  need_to_clamp |= (mv->as_mv.col > mb_to_right_edge);
+  need_to_clamp |= (mv->as_mv.row < mb_to_top_edge);
+  need_to_clamp |= (mv->as_mv.row > mb_to_bottom_edge);
+  return need_to_clamp;
 }
 
-void vp8_find_near_mvs
-(
-    MACROBLOCKD *xd,
-    const MODE_INFO *here,
-    int_mv *nearest, int_mv *nearby, int_mv *best,
-    int near_mv_ref_cts[4],
-    int refframe,
-    int *ref_frame_sign_bias
-);
+void vp8_find_near_mvs(MACROBLOCKD *xd, const MODE_INFO *here, int_mv *nearest,
+                       int_mv *nearby, int_mv *best, int near_mv_ref_cts[4],
+                       int refframe, int *ref_frame_sign_bias);
 
+int vp8_find_near_mvs_bias(MACROBLOCKD *xd, const MODE_INFO *here,
+                           int_mv mode_mv_sb[2][MB_MODE_COUNT],
+                           int_mv best_mv_sb[2], int cnt[4], int refframe,
+                           int *ref_frame_sign_bias);
 
-int vp8_find_near_mvs_bias
-(
-    MACROBLOCKD *xd,
-    const MODE_INFO *here,
-    int_mv mode_mv_sb[2][MB_MODE_COUNT],
-    int_mv best_mv_sb[2],
-    int cnt[4],
-    int refframe,
-    int *ref_frame_sign_bias
-);
+vp8_prob *vp8_mv_ref_probs(vp8_prob p[VP8_MVREFS - 1],
+                           const int near_mv_ref_ct[4]);
 
-
-vp8_prob *vp8_mv_ref_probs(
-    vp8_prob p[VP8_MVREFS-1], const int near_mv_ref_ct[4]
-);
-
 extern const unsigned char vp8_mbsplit_offset[4][16];
 
+static INLINE uint32_t left_block_mv(const MODE_INFO *cur_mb, int b) {
+  if (!(b & 3)) {
+    /* On L edge, get from MB to left of us */
+    --cur_mb;
 
-static INLINE uint32_t left_block_mv(const MODE_INFO *cur_mb, int b)
-{
-    if (!(b & 3))
-    {
-        /* On L edge, get from MB to left of us */
-        --cur_mb;
+    if (cur_mb->mbmi.mode != SPLITMV) return cur_mb->mbmi.mv.as_int;
+    b += 4;
+  }
 
-        if(cur_mb->mbmi.mode != SPLITMV)
-            return cur_mb->mbmi.mv.as_int;
-        b += 4;
-    }
-
-    return (cur_mb->bmi + b - 1)->mv.as_int;
+  return (cur_mb->bmi + b - 1)->mv.as_int;
 }
 
 static INLINE uint32_t above_block_mv(const MODE_INFO *cur_mb, int b,
-                                      int mi_stride)
-{
-    if (!(b >> 2))
-    {
-        /* On top edge, get from MB above us */
-        cur_mb -= mi_stride;
+                                      int mi_stride) {
+  if (!(b >> 2)) {
+    /* On top edge, get from MB above us */
+    cur_mb -= mi_stride;
 
-        if(cur_mb->mbmi.mode != SPLITMV)
-            return cur_mb->mbmi.mv.as_int;
-        b += 16;
-    }
+    if (cur_mb->mbmi.mode != SPLITMV) return cur_mb->mbmi.mv.as_int;
+    b += 16;
+  }
 
-    return (cur_mb->bmi + (b - 4))->mv.as_int;
+  return (cur_mb->bmi + (b - 4))->mv.as_int;
 }
-static INLINE B_PREDICTION_MODE left_block_mode(const MODE_INFO *cur_mb, int b)
-{
-    if (!(b & 3))
-    {
-        /* On L edge, get from MB to left of us */
-        --cur_mb;
-        switch (cur_mb->mbmi.mode)
-        {
-            case B_PRED:
-              return (cur_mb->bmi + b + 3)->as_mode;
-            case DC_PRED:
-                return B_DC_PRED;
-            case V_PRED:
-                return B_VE_PRED;
-            case H_PRED:
-                return B_HE_PRED;
-            case TM_PRED:
-                return B_TM_PRED;
-            default:
-                return B_DC_PRED;
-        }
+static INLINE B_PREDICTION_MODE left_block_mode(const MODE_INFO *cur_mb,
+                                                int b) {
+  if (!(b & 3)) {
+    /* On L edge, get from MB to left of us */
+    --cur_mb;
+    switch (cur_mb->mbmi.mode) {
+      case B_PRED: return (cur_mb->bmi + b + 3)->as_mode;
+      case DC_PRED: return B_DC_PRED;
+      case V_PRED: return B_VE_PRED;
+      case H_PRED: return B_HE_PRED;
+      case TM_PRED: return B_TM_PRED;
+      default: return B_DC_PRED;
     }
+  }
 
-    return (cur_mb->bmi + b - 1)->as_mode;
+  return (cur_mb->bmi + b - 1)->as_mode;
 }
 
 static INLINE B_PREDICTION_MODE above_block_mode(const MODE_INFO *cur_mb, int b,
-                                                 int mi_stride)
-{
-    if (!(b >> 2))
-    {
-        /* On top edge, get from MB above us */
-        cur_mb -= mi_stride;
+                                                 int mi_stride) {
+  if (!(b >> 2)) {
+    /* On top edge, get from MB above us */
+    cur_mb -= mi_stride;
 
-        switch (cur_mb->mbmi.mode)
-        {
-            case B_PRED:
-              return (cur_mb->bmi + b + 12)->as_mode;
-            case DC_PRED:
-                return B_DC_PRED;
-            case V_PRED:
-                return B_VE_PRED;
-            case H_PRED:
-                return B_HE_PRED;
-            case TM_PRED:
-                return B_TM_PRED;
-            default:
-                return B_DC_PRED;
-        }
+    switch (cur_mb->mbmi.mode) {
+      case B_PRED: return (cur_mb->bmi + b + 12)->as_mode;
+      case DC_PRED: return B_DC_PRED;
+      case V_PRED: return B_VE_PRED;
+      case H_PRED: return B_HE_PRED;
+      case TM_PRED: return B_TM_PRED;
+      default: return B_DC_PRED;
     }
+  }
 
-    return (cur_mb->bmi + b - 4)->as_mode;
+  return (cur_mb->bmi + b - 4)->as_mode;
 }
 
 #ifdef __cplusplus
--- a/vp8/common/generic/systemdependent.c
+++ b/vp8/common/generic/systemdependent.c
@@ -8,7 +8,6 @@
  *  be found in the AUTHORS file in the root of the source tree.
  */
 
-
 #include "vpx_config.h"
 #include "vp8_rtcd.h"
 #if ARCH_ARM
@@ -24,7 +23,7 @@
 #include <unistd.h>
 #elif defined(_WIN32)
 #include <windows.h>
-typedef void (WINAPI *PGNSI)(LPSYSTEM_INFO);
+typedef void(WINAPI *PGNSI)(LPSYSTEM_INFO);
 #elif defined(__OS2__)
 #define INCL_DOS
 #define INCL_DOSSPINLOCK
@@ -33,74 +32,69 @@
 #endif
 
 #if CONFIG_MULTITHREAD
-static int get_cpu_count()
-{
-    int core_count = 16;
+static int get_cpu_count() {
+  int core_count = 16;
 
 #if HAVE_UNISTD_H && !defined(__OS2__)
 #if defined(_SC_NPROCESSORS_ONLN)
-    core_count = sysconf(_SC_NPROCESSORS_ONLN);
+  core_count = sysconf(_SC_NPROCESSORS_ONLN);
 #elif defined(_SC_NPROC_ONLN)
-    core_count = sysconf(_SC_NPROC_ONLN);
+  core_count = sysconf(_SC_NPROC_ONLN);
 #endif
 #elif defined(_WIN32)
-    {
+  {
 #if _WIN32_WINNT >= 0x0501
-        SYSTEM_INFO sysinfo;
-        GetNativeSystemInfo(&sysinfo);
+    SYSTEM_INFO sysinfo;
+    GetNativeSystemInfo(&sysinfo);
 #else
-        PGNSI pGNSI;
-        SYSTEM_INFO sysinfo;
+    PGNSI pGNSI;
+    SYSTEM_INFO sysinfo;
 
-        /* Call GetNativeSystemInfo if supported or
-         * GetSystemInfo otherwise. */
+    /* Call GetNativeSystemInfo if supported or
+     * GetSystemInfo otherwise. */
 
-        pGNSI = (PGNSI) GetProcAddress(
-                GetModuleHandle(TEXT("kernel32.dll")), "GetNativeSystemInfo");
-        if (pGNSI != NULL)
-            pGNSI(&sysinfo);
-        else
-            GetSystemInfo(&sysinfo);
+    pGNSI = (PGNSI)GetProcAddress(GetModuleHandle(TEXT("kernel32.dll")),
+                                  "GetNativeSystemInfo");
+    if (pGNSI != NULL)
+      pGNSI(&sysinfo);
+    else
+      GetSystemInfo(&sysinfo);
 #endif
 
-        core_count = sysinfo.dwNumberOfProcessors;
-    }
+    core_count = sysinfo.dwNumberOfProcessors;
+  }
 #elif defined(__OS2__)
-    {
-        ULONG proc_id;
-        ULONG status;
+  {
+    ULONG proc_id;
+    ULONG status;
 
-        core_count = 0;
-        for (proc_id = 1; ; proc_id++)
-        {
-            if (DosGetProcessorStatus(proc_id, &status))
-                break;
+    core_count = 0;
+    for (proc_id = 1;; proc_id++) {
+      if (DosGetProcessorStatus(proc_id, &status)) break;
 
-            if (status == PROC_ONLINE)
-                core_count++;
-        }
+      if (status == PROC_ONLINE) core_count++;
     }
+  }
 #else
-    /* other platforms */
+/* other platforms */
 #endif
 
-    return core_count > 0 ? core_count : 1;
+  return core_count > 0 ? core_count : 1;
 }
 #endif
 
-void vp8_clear_system_state_c() {};
+void vp8_clear_system_state_c(){};
 
-void vp8_machine_specific_config(VP8_COMMON *ctx)
-{
+void vp8_machine_specific_config(VP8_COMMON *ctx) {
 #if CONFIG_MULTITHREAD
-    ctx->processor_core_count = get_cpu_count();
+  ctx->processor_core_count = get_cpu_count();
 #else
-    (void)ctx;
+  (void)ctx;
 #endif /* CONFIG_MULTITHREAD */
 
 #if ARCH_ARM
-    ctx->cpu_caps = arm_cpu_caps();
+  ctx->cpu_caps = arm_cpu_caps();
 #elif ARCH_X86 || ARCH_X86_64
-    ctx->cpu_caps = x86_simd_caps();
+  ctx->cpu_caps = x86_simd_caps();
 #endif
 }
--- a/vp8/common/header.h
+++ b/vp8/common/header.h
@@ -8,7 +8,6 @@
  *  be found in the AUTHORS file in the root of the source tree.
  */
 
-
 #ifndef VP8_COMMON_HEADER_H_
 #define VP8_COMMON_HEADER_H_
 
@@ -17,22 +16,21 @@
 #endif
 
 /* 24 bits total */
-typedef struct
-{
-    unsigned int type: 1;
-    unsigned int version: 3;
-    unsigned int show_frame: 1;
+typedef struct {
+  unsigned int type : 1;
+  unsigned int version : 3;
+  unsigned int show_frame : 1;
 
-    /* Allow 2^20 bytes = 8 megabits for first partition */
+  /* Allow 2^20 bytes = 8 megabits for first partition */
 
-    unsigned int first_partition_length_in_bytes: 19;
+  unsigned int first_partition_length_in_bytes : 19;
 
 #ifdef PACKET_TESTING
-    unsigned int frame_number;
-    unsigned int update_gold: 1;
-    unsigned int uses_gold: 1;
-    unsigned int update_last: 1;
-    unsigned int uses_last: 1;
+  unsigned int frame_number;
+  unsigned int update_gold : 1;
+  unsigned int uses_gold : 1;
+  unsigned int update_last : 1;
+  unsigned int uses_last : 1;
 #endif
 
 } VP8_HEADER;
@@ -42,7 +40,6 @@
 #else
 #define VP8_HEADER_SIZE 3
 #endif
-
 
 #ifdef __cplusplus
 }  // extern "C"
--- a/vp8/common/idct_blk.c
+++ b/vp8/common/idct_blk.c
@@ -12,79 +12,67 @@
 #include "vp8_rtcd.h"
 #include "vpx_mem/vpx_mem.h"
 
-void vp8_dequant_idct_add_c(short *input, short *dq,
-                            unsigned char *dest, int stride);
-void vp8_dc_only_idct_add_c(short input_dc, unsigned char * pred,
+void vp8_dequant_idct_add_c(short *input, short *dq, unsigned char *dest,
+                            int stride);
+void vp8_dc_only_idct_add_c(short input_dc, unsigned char *pred,
                             int pred_stride, unsigned char *dst_ptr,
                             int dst_stride);
 
-void vp8_dequant_idct_add_y_block_c
-            (short *q, short *dq,
-             unsigned char *dst, int stride, char *eobs)
-{
-    int i, j;
+void vp8_dequant_idct_add_y_block_c(short *q, short *dq, unsigned char *dst,
+                                    int stride, char *eobs) {
+  int i, j;
 
-    for (i = 0; i < 4; i++)
-    {
-        for (j = 0; j < 4; j++)
-        {
-            if (*eobs++ > 1)
-                vp8_dequant_idct_add_c (q, dq, dst, stride);
-            else
-            {
-                vp8_dc_only_idct_add_c (q[0]*dq[0], dst, stride, dst, stride);
-                memset(q, 0, 2 * sizeof(q[0]));
-            }
+  for (i = 0; i < 4; i++) {
+    for (j = 0; j < 4; j++) {
+      if (*eobs++ > 1)
+        vp8_dequant_idct_add_c(q, dq, dst, stride);
+      else {
+        vp8_dc_only_idct_add_c(q[0] * dq[0], dst, stride, dst, stride);
+        memset(q, 0, 2 * sizeof(q[0]));
+      }
 
-            q   += 16;
-            dst += 4;
-        }
-
-        dst += 4*stride - 16;
+      q += 16;
+      dst += 4;
     }
+
+    dst += 4 * stride - 16;
+  }
 }
 
-void vp8_dequant_idct_add_uv_block_c
-            (short *q, short *dq,
-             unsigned char *dstu, unsigned char *dstv, int stride, char *eobs)
-{
-    int i, j;
+void vp8_dequant_idct_add_uv_block_c(short *q, short *dq, unsigned char *dstu,
+                                     unsigned char *dstv, int stride,
+                                     char *eobs) {
+  int i, j;
 
-    for (i = 0; i < 2; i++)
-    {
-        for (j = 0; j < 2; j++)
-        {
-            if (*eobs++ > 1)
-                vp8_dequant_idct_add_c (q, dq, dstu, stride);
-            else
-            {
-                vp8_dc_only_idct_add_c (q[0]*dq[0], dstu, stride, dstu, stride);
-                memset(q, 0, 2 * sizeof(q[0]));
-            }
+  for (i = 0; i < 2; i++) {
+    for (j = 0; j < 2; j++) {
+      if (*eobs++ > 1)
+        vp8_dequant_idct_add_c(q, dq, dstu, stride);
+      else {
+        vp8_dc_only_idct_add_c(q[0] * dq[0], dstu, stride, dstu, stride);
+        memset(q, 0, 2 * sizeof(q[0]));
+      }
 
-            q    += 16;
-            dstu += 4;
-        }
-
-        dstu += 4*stride - 8;
+      q += 16;
+      dstu += 4;
     }
 
-    for (i = 0; i < 2; i++)
-    {
-        for (j = 0; j < 2; j++)
-        {
-            if (*eobs++ > 1)
-                vp8_dequant_idct_add_c (q, dq, dstv, stride);
-            else
-            {
-                vp8_dc_only_idct_add_c (q[0]*dq[0], dstv, stride, dstv, stride);
-                memset(q, 0, 2 * sizeof(q[0]));
-            }
+    dstu += 4 * stride - 8;
+  }
 
-            q    += 16;
-            dstv += 4;
-        }
+  for (i = 0; i < 2; i++) {
+    for (j = 0; j < 2; j++) {
+      if (*eobs++ > 1)
+        vp8_dequant_idct_add_c(q, dq, dstv, stride);
+      else {
+        vp8_dc_only_idct_add_c(q[0] * dq[0], dstv, stride, dstv, stride);
+        memset(q, 0, 2 * sizeof(q[0]));
+      }
 
-        dstv += 4*stride - 8;
+      q += 16;
+      dstv += 4;
     }
+
+    dstv += 4 * stride - 8;
+  }
 }
--- a/vp8/common/idctllm.c
+++ b/vp8/common/idctllm.c
@@ -24,182 +24,162 @@
  *         x * sqrt(2) * cos (pi/8) = x + x * (sqrt(2) *cos(pi/8)-1).
  **************************************************************************/
 static const int cospi8sqrt2minus1 = 20091;
-static const int sinpi8sqrt2      = 35468;
+static const int sinpi8sqrt2 = 35468;
 
 void vp8_short_idct4x4llm_c(short *input, unsigned char *pred_ptr,
                             int pred_stride, unsigned char *dst_ptr,
-                            int dst_stride)
-{
-    int i;
-    int r, c;
-    int a1, b1, c1, d1;
-    short output[16];
-    short *ip = input;
-    short *op = output;
-    int temp1, temp2;
-    int shortpitch = 4;
+                            int dst_stride) {
+  int i;
+  int r, c;
+  int a1, b1, c1, d1;
+  short output[16];
+  short *ip = input;
+  short *op = output;
+  int temp1, temp2;
+  int shortpitch = 4;
 
-    for (i = 0; i < 4; i++)
-    {
-        a1 = ip[0] + ip[8];
-        b1 = ip[0] - ip[8];
+  for (i = 0; i < 4; i++) {
+    a1 = ip[0] + ip[8];
+    b1 = ip[0] - ip[8];
 
-        temp1 = (ip[4] * sinpi8sqrt2) >> 16;
-        temp2 = ip[12] + ((ip[12] * cospi8sqrt2minus1) >> 16);
-        c1 = temp1 - temp2;
+    temp1 = (ip[4] * sinpi8sqrt2) >> 16;
+    temp2 = ip[12] + ((ip[12] * cospi8sqrt2minus1) >> 16);
+    c1 = temp1 - temp2;
 
-        temp1 = ip[4] + ((ip[4] * cospi8sqrt2minus1) >> 16);
-        temp2 = (ip[12] * sinpi8sqrt2) >> 16;
-        d1 = temp1 + temp2;
+    temp1 = ip[4] + ((ip[4] * cospi8sqrt2minus1) >> 16);
+    temp2 = (ip[12] * sinpi8sqrt2) >> 16;
+    d1 = temp1 + temp2;
 
-        op[shortpitch*0] = a1 + d1;
-        op[shortpitch*3] = a1 - d1;
+    op[shortpitch * 0] = a1 + d1;
+    op[shortpitch * 3] = a1 - d1;
 
-        op[shortpitch*1] = b1 + c1;
-        op[shortpitch*2] = b1 - c1;
+    op[shortpitch * 1] = b1 + c1;
+    op[shortpitch * 2] = b1 - c1;
 
-        ip++;
-        op++;
-    }
+    ip++;
+    op++;
+  }
 
-    ip = output;
-    op = output;
+  ip = output;
+  op = output;
 
-    for (i = 0; i < 4; i++)
-    {
-        a1 = ip[0] + ip[2];
-        b1 = ip[0] - ip[2];
+  for (i = 0; i < 4; i++) {
+    a1 = ip[0] + ip[2];
+    b1 = ip[0] - ip[2];
 
-        temp1 = (ip[1] * sinpi8sqrt2) >> 16;
-        temp2 = ip[3] + ((ip[3] * cospi8sqrt2minus1) >> 16);
-        c1 = temp1 - temp2;
+    temp1 = (ip[1] * sinpi8sqrt2) >> 16;
+    temp2 = ip[3] + ((ip[3] * cospi8sqrt2minus1) >> 16);
+    c1 = temp1 - temp2;
 
-        temp1 = ip[1] + ((ip[1] * cospi8sqrt2minus1) >> 16);
-        temp2 = (ip[3] * sinpi8sqrt2) >> 16;
-        d1 = temp1 + temp2;
+    temp1 = ip[1] + ((ip[1] * cospi8sqrt2minus1) >> 16);
+    temp2 = (ip[3] * sinpi8sqrt2) >> 16;
+    d1 = temp1 + temp2;
 
+    op[0] = (a1 + d1 + 4) >> 3;
+    op[3] = (a1 - d1 + 4) >> 3;
 
-        op[0] = (a1 + d1 + 4) >> 3;
-        op[3] = (a1 - d1 + 4) >> 3;
+    op[1] = (b1 + c1 + 4) >> 3;
+    op[2] = (b1 - c1 + 4) >> 3;
 
-        op[1] = (b1 + c1 + 4) >> 3;
-        op[2] = (b1 - c1 + 4) >> 3;
+    ip += shortpitch;
+    op += shortpitch;
+  }
 
-        ip += shortpitch;
-        op += shortpitch;
-    }
+  ip = output;
+  for (r = 0; r < 4; r++) {
+    for (c = 0; c < 4; c++) {
+      int a = ip[c] + pred_ptr[c];
 
-    ip = output;
-    for (r = 0; r < 4; r++)
-    {
-        for (c = 0; c < 4; c++)
-        {
-            int a = ip[c] + pred_ptr[c] ;
+      if (a < 0) a = 0;
 
-            if (a < 0)
-                a = 0;
+      if (a > 255) a = 255;
 
-            if (a > 255)
-                a = 255;
-
-            dst_ptr[c] = (unsigned char) a ;
-        }
-        ip += 4;
-        dst_ptr += dst_stride;
-        pred_ptr += pred_stride;
+      dst_ptr[c] = (unsigned char)a;
     }
+    ip += 4;
+    dst_ptr += dst_stride;
+    pred_ptr += pred_stride;
+  }
 }
 
 void vp8_dc_only_idct_add_c(short input_dc, unsigned char *pred_ptr,
                             int pred_stride, unsigned char *dst_ptr,
-                            int dst_stride)
-{
-    int a1 = ((input_dc + 4) >> 3);
-    int r, c;
+                            int dst_stride) {
+  int a1 = ((input_dc + 4) >> 3);
+  int r, c;
 
-    for (r = 0; r < 4; r++)
-    {
-        for (c = 0; c < 4; c++)
-        {
-            int a = a1 + pred_ptr[c] ;
+  for (r = 0; r < 4; r++) {
+    for (c = 0; c < 4; c++) {
+      int a = a1 + pred_ptr[c];
 
-            if (a < 0)
-                a = 0;
+      if (a < 0) a = 0;
 
-            if (a > 255)
-                a = 255;
+      if (a > 255) a = 255;
 
-            dst_ptr[c] = (unsigned char) a ;
-        }
-
-        dst_ptr += dst_stride;
-        pred_ptr += pred_stride;
+      dst_ptr[c] = (unsigned char)a;
     }
 
+    dst_ptr += dst_stride;
+    pred_ptr += pred_stride;
+  }
 }
 
-void vp8_short_inv_walsh4x4_c(short *input, short *mb_dqcoeff)
-{
-    short output[16];
-    int i;
-    int a1, b1, c1, d1;
-    int a2, b2, c2, d2;
-    short *ip = input;
-    short *op = output;
+void vp8_short_inv_walsh4x4_c(short *input, short *mb_dqcoeff) {
+  short output[16];
+  int i;
+  int a1, b1, c1, d1;
+  int a2, b2, c2, d2;
+  short *ip = input;
+  short *op = output;
 
-    for (i = 0; i < 4; i++)
-    {
-        a1 = ip[0] + ip[12];
-        b1 = ip[4] + ip[8];
-        c1 = ip[4] - ip[8];
-        d1 = ip[0] - ip[12];
+  for (i = 0; i < 4; i++) {
+    a1 = ip[0] + ip[12];
+    b1 = ip[4] + ip[8];
+    c1 = ip[4] - ip[8];
+    d1 = ip[0] - ip[12];
 
-        op[0] = a1 + b1;
-        op[4] = c1 + d1;
-        op[8] = a1 - b1;
-        op[12] = d1 - c1;
-        ip++;
-        op++;
-    }
+    op[0] = a1 + b1;
+    op[4] = c1 + d1;
+    op[8] = a1 - b1;
+    op[12] = d1 - c1;
+    ip++;
+    op++;
+  }
 
-    ip = output;
-    op = output;
+  ip = output;
+  op = output;
 
-    for (i = 0; i < 4; i++)
-    {
-        a1 = ip[0] + ip[3];
-        b1 = ip[1] + ip[2];
-        c1 = ip[1] - ip[2];
-        d1 = ip[0] - ip[3];
+  for (i = 0; i < 4; i++) {
+    a1 = ip[0] + ip[3];
+    b1 = ip[1] + ip[2];
+    c1 = ip[1] - ip[2];
+    d1 = ip[0] - ip[3];
 
-        a2 = a1 + b1;
-        b2 = c1 + d1;
-        c2 = a1 - b1;
-        d2 = d1 - c1;
+    a2 = a1 + b1;
+    b2 = c1 + d1;
+    c2 = a1 - b1;
+    d2 = d1 - c1;
 
-        op[0] = (a2 + 3) >> 3;
-        op[1] = (b2 + 3) >> 3;
-        op[2] = (c2 + 3) >> 3;
-        op[3] = (d2 + 3) >> 3;
+    op[0] = (a2 + 3) >> 3;
+    op[1] = (b2 + 3) >> 3;
+    op[2] = (c2 + 3) >> 3;
+    op[3] = (d2 + 3) >> 3;
 
-        ip += 4;
-        op += 4;
-    }
+    ip += 4;
+    op += 4;
+  }
 
-    for(i = 0; i < 16; i++)
-    {
-        mb_dqcoeff[i * 16] = output[i];
-    }
+  for (i = 0; i < 16; i++) {
+    mb_dqcoeff[i * 16] = output[i];
+  }
 }
 
-void vp8_short_inv_walsh4x4_1_c(short *input, short *mb_dqcoeff)
-{
-    int i;
-    int a1;
+void vp8_short_inv_walsh4x4_1_c(short *input, short *mb_dqcoeff) {
+  int i;
+  int a1;
 
-    a1 = ((input[0] + 3) >> 3);
-    for(i = 0; i < 16; i++)
-    {
-        mb_dqcoeff[i * 16] = a1;
-    }
+  a1 = ((input[0] + 3) >> 3);
+  for (i = 0; i < 16; i++) {
+    mb_dqcoeff[i * 16] = a1;
+  }
 }
--- a/vp8/common/invtrans.h
+++ b/vp8/common/invtrans.h
@@ -8,7 +8,6 @@
  *  be found in the AUTHORS file in the root of the source tree.
  */
 
-
 #ifndef VP8_COMMON_INVTRANS_H_
 #define VP8_COMMON_INVTRANS_H_
 
@@ -25,43 +24,31 @@
 extern "C" {
 #endif
 
-static void eob_adjust(char *eobs, short *diff)
-{
-    /* eob adjust.... the idct can only skip if both the dc and eob are zero */
-    int js;
-    for(js = 0; js < 16; js++)
-    {
-        if((eobs[js] == 0) && (diff[0] != 0))
-            eobs[js]++;
-        diff+=16;
-    }
+static void eob_adjust(char *eobs, short *diff) {
+  /* eob adjust.... the idct can only skip if both the dc and eob are zero */
+  int js;
+  for (js = 0; js < 16; js++) {
+    if ((eobs[js] == 0) && (diff[0] != 0)) eobs[js]++;
+    diff += 16;
+  }
 }
 
-static INLINE void vp8_inverse_transform_mby(MACROBLOCKD *xd)
-{
-    short *DQC = xd->dequant_y1;
+static INLINE void vp8_inverse_transform_mby(MACROBLOCKD *xd) {
+  short *DQC = xd->dequant_y1;
 
-    if (xd->mode_info_context->mbmi.mode != SPLITMV)
-    {
-        /* do 2nd order transform on the dc block */
-        if (xd->eobs[24] > 1)
-        {
-            vp8_short_inv_walsh4x4
-                (&xd->block[24].dqcoeff[0], xd->qcoeff);
-        }
-        else
-        {
-            vp8_short_inv_walsh4x4_1
-                (&xd->block[24].dqcoeff[0], xd->qcoeff);
-        }
-        eob_adjust(xd->eobs, xd->qcoeff);
-
-        DQC = xd->dequant_y1_dc;
+  if (xd->mode_info_context->mbmi.mode != SPLITMV) {
+    /* do 2nd order transform on the dc block */
+    if (xd->eobs[24] > 1) {
+      vp8_short_inv_walsh4x4(&xd->block[24].dqcoeff[0], xd->qcoeff);
+    } else {
+      vp8_short_inv_walsh4x4_1(&xd->block[24].dqcoeff[0], xd->qcoeff);
     }
-    vp8_dequant_idct_add_y_block
-                    (xd->qcoeff, DQC,
-                     xd->dst.y_buffer,
-                     xd->dst.y_stride, xd->eobs);
+    eob_adjust(xd->eobs, xd->qcoeff);
+
+    DQC = xd->dequant_y1_dc;
+  }
+  vp8_dequant_idct_add_y_block(xd->qcoeff, DQC, xd->dst.y_buffer,
+                               xd->dst.y_stride, xd->eobs);
 }
 #ifdef __cplusplus
 }  // extern "C"
--- a/vp8/common/loopfilter.h
+++ b/vp8/common/loopfilter.h
@@ -8,7 +8,6 @@
  *  be found in the AUTHORS file in the root of the source tree.
  */
 
-
 #ifndef VP8_COMMON_LOOPFILTER_H_
 #define VP8_COMMON_LOOPFILTER_H_
 
@@ -20,16 +19,12 @@
 extern "C" {
 #endif
 
-#define MAX_LOOP_FILTER             63
+#define MAX_LOOP_FILTER 63
 /* fraction of total macroblock rows to be used in fast filter level picking */
 /* has to be > 2 */
-#define PARTIAL_FRAME_FRACTION      8
+#define PARTIAL_FRAME_FRACTION 8
 
-typedef enum
-{
-    NORMAL_LOOPFILTER = 0,
-    SIMPLE_LOOPFILTER = 1
-} LOOPFILTERTYPE;
+typedef enum { NORMAL_LOOPFILTER = 0, SIMPLE_LOOPFILTER = 1 } LOOPFILTERTYPE;
 
 #if ARCH_ARM
 #define SIMD_WIDTH 1
@@ -40,35 +35,32 @@
 /* Need to align this structure so when it is declared and
  * passed it can be loaded into vector registers.
  */
-typedef struct
-{
-    DECLARE_ALIGNED(SIMD_WIDTH, unsigned char, mblim[MAX_LOOP_FILTER + 1][SIMD_WIDTH]);
-    DECLARE_ALIGNED(SIMD_WIDTH, unsigned char, blim[MAX_LOOP_FILTER + 1][SIMD_WIDTH]);
-    DECLARE_ALIGNED(SIMD_WIDTH, unsigned char, lim[MAX_LOOP_FILTER + 1][SIMD_WIDTH]);
-    DECLARE_ALIGNED(SIMD_WIDTH, unsigned char, hev_thr[4][SIMD_WIDTH]);
-    unsigned char lvl[4][4][4];
-    unsigned char hev_thr_lut[2][MAX_LOOP_FILTER + 1];
-    unsigned char mode_lf_lut[10];
+typedef struct {
+  DECLARE_ALIGNED(SIMD_WIDTH, unsigned char,
+                  mblim[MAX_LOOP_FILTER + 1][SIMD_WIDTH]);
+  DECLARE_ALIGNED(SIMD_WIDTH, unsigned char,
+                  blim[MAX_LOOP_FILTER + 1][SIMD_WIDTH]);
+  DECLARE_ALIGNED(SIMD_WIDTH, unsigned char,
+                  lim[MAX_LOOP_FILTER + 1][SIMD_WIDTH]);
+  DECLARE_ALIGNED(SIMD_WIDTH, unsigned char, hev_thr[4][SIMD_WIDTH]);
+  unsigned char lvl[4][4][4];
+  unsigned char hev_thr_lut[2][MAX_LOOP_FILTER + 1];
+  unsigned char mode_lf_lut[10];
 } loop_filter_info_n;
 
-typedef struct loop_filter_info
-{
-    const unsigned char * mblim;
-    const unsigned char * blim;
-    const unsigned char * lim;
-    const unsigned char * hev_thr;
+typedef struct loop_filter_info {
+  const unsigned char *mblim;
+  const unsigned char *blim;
+  const unsigned char *lim;
+  const unsigned char *hev_thr;
 } loop_filter_info;
 
-
-typedef void loop_filter_uvfunction
-(
-    unsigned char *u,   /* source pointer */
-    int p,              /* pitch */
-    const unsigned char *blimit,
-    const unsigned char *limit,
-    const unsigned char *thresh,
-    unsigned char *v
-);
+typedef void loop_filter_uvfunction(unsigned char *u, /* source pointer */
+                                    int p,            /* pitch */
+                                    const unsigned char *blimit,
+                                    const unsigned char *limit,
+                                    const unsigned char *thresh,
+                                    unsigned char *v);
 
 /* assorted loopfilter functions which get used elsewhere */
 struct VP8Common;
@@ -77,8 +69,7 @@
 
 void vp8_loop_filter_init(struct VP8Common *cm);
 
-void vp8_loop_filter_frame_init(struct VP8Common *cm,
-                                struct macroblockd *mbd,
+void vp8_loop_filter_frame_init(struct VP8Common *cm, struct macroblockd *mbd,
                                 int default_filt_lvl);
 
 void vp8_loop_filter_frame(struct VP8Common *cm, struct macroblockd *mbd,
@@ -88,8 +79,7 @@
                                    struct macroblockd *mbd,
                                    int default_filt_lvl);
 
-void vp8_loop_filter_frame_yonly(struct VP8Common *cm,
-                                 struct macroblockd *mbd,
+void vp8_loop_filter_frame_yonly(struct VP8Common *cm, struct macroblockd *mbd,
                                  int default_filt_lvl);
 
 void vp8_loop_filter_update_sharpness(loop_filter_info_n *lfi,
@@ -96,14 +86,14 @@
                                       int sharpness_lvl);
 
 void vp8_loop_filter_row_normal(struct VP8Common *cm,
-                                struct modeinfo *mode_info_context,
-                                int mb_row, int post_ystride, int post_uvstride,
+                                struct modeinfo *mode_info_context, int mb_row,
+                                int post_ystride, int post_uvstride,
                                 unsigned char *y_ptr, unsigned char *u_ptr,
                                 unsigned char *v_ptr);
 
 void vp8_loop_filter_row_simple(struct VP8Common *cm,
-                                struct modeinfo *mode_info_context,
-                                int mb_row, int post_ystride, int post_uvstride,
+                                struct modeinfo *mode_info_context, int mb_row,
+                                int post_ystride, int post_uvstride,
                                 unsigned char *y_ptr, unsigned char *u_ptr,
                                 unsigned char *v_ptr);
 #ifdef __cplusplus
--- a/vp8/common/loopfilter_filters.c
+++ b/vp8/common/loopfilter_filters.c
@@ -8,7 +8,6 @@
  *  be found in the AUTHORS file in the root of the source tree.
  */
 
-
 #include <stdlib.h>
 #include "loopfilter.h"
 #include "onyxc_int.h"
@@ -15,416 +14,368 @@
 
 typedef unsigned char uc;
 
-static signed char vp8_signed_char_clamp(int t)
-{
-    t = (t < -128 ? -128 : t);
-    t = (t > 127 ? 127 : t);
-    return (signed char) t;
+static signed char vp8_signed_char_clamp(int t) {
+  t = (t < -128 ? -128 : t);
+  t = (t > 127 ? 127 : t);
+  return (signed char)t;
 }
 
-
 /* should we apply any filter at all ( 11111111 yes, 00000000 no) */
-static signed char vp8_filter_mask(uc limit, uc blimit,
-                            uc p3, uc p2, uc p1, uc p0,
-                            uc q0, uc q1, uc q2, uc q3)
-{
-    signed char mask = 0;
-    mask |= (abs(p3 - p2) > limit);
-    mask |= (abs(p2 - p1) > limit);
-    mask |= (abs(p1 - p0) > limit);
-    mask |= (abs(q1 - q0) > limit);
-    mask |= (abs(q2 - q1) > limit);
-    mask |= (abs(q3 - q2) > limit);
-    mask |= (abs(p0 - q0) * 2 + abs(p1 - q1) / 2  > blimit);
-    return mask - 1;
+static signed char vp8_filter_mask(uc limit, uc blimit, uc p3, uc p2, uc p1,
+                                   uc p0, uc q0, uc q1, uc q2, uc q3) {
+  signed char mask = 0;
+  mask |= (abs(p3 - p2) > limit);
+  mask |= (abs(p2 - p1) > limit);
+  mask |= (abs(p1 - p0) > limit);
+  mask |= (abs(q1 - q0) > limit);
+  mask |= (abs(q2 - q1) > limit);
+  mask |= (abs(q3 - q2) > limit);
+  mask |= (abs(p0 - q0) * 2 + abs(p1 - q1) / 2 > blimit);
+  return mask - 1;
 }
 
 /* is there high variance internal edge ( 11111111 yes, 00000000 no) */
-static signed char vp8_hevmask(uc thresh, uc p1, uc p0, uc q0, uc q1)
-{
-    signed char hev = 0;
-    hev  |= (abs(p1 - p0) > thresh) * -1;
-    hev  |= (abs(q1 - q0) > thresh) * -1;
-    return hev;
+static signed char vp8_hevmask(uc thresh, uc p1, uc p0, uc q0, uc q1) {
+  signed char hev = 0;
+  hev |= (abs(p1 - p0) > thresh) * -1;
+  hev |= (abs(q1 - q0) > thresh) * -1;
+  return hev;
 }
 
-static void vp8_filter(signed char mask, uc hev, uc *op1,
-        uc *op0, uc *oq0, uc *oq1)
+static void vp8_filter(signed char mask, uc hev, uc *op1, uc *op0, uc *oq0,
+                       uc *oq1) {
+  signed char ps0, qs0;
+  signed char ps1, qs1;
+  signed char filter_value, Filter1, Filter2;
+  signed char u;
 
-{
-    signed char ps0, qs0;
-    signed char ps1, qs1;
-    signed char filter_value, Filter1, Filter2;
-    signed char u;
+  ps1 = (signed char)*op1 ^ 0x80;
+  ps0 = (signed char)*op0 ^ 0x80;
+  qs0 = (signed char)*oq0 ^ 0x80;
+  qs1 = (signed char)*oq1 ^ 0x80;
 
-    ps1 = (signed char) * op1 ^ 0x80;
-    ps0 = (signed char) * op0 ^ 0x80;
-    qs0 = (signed char) * oq0 ^ 0x80;
-    qs1 = (signed char) * oq1 ^ 0x80;
+  /* add outer taps if we have high edge variance */
+  filter_value = vp8_signed_char_clamp(ps1 - qs1);
+  filter_value &= hev;
 
-    /* add outer taps if we have high edge variance */
-    filter_value = vp8_signed_char_clamp(ps1 - qs1);
-    filter_value &= hev;
+  /* inner taps */
+  filter_value = vp8_signed_char_clamp(filter_value + 3 * (qs0 - ps0));
+  filter_value &= mask;
 
-    /* inner taps */
-    filter_value = vp8_signed_char_clamp(filter_value + 3 * (qs0 - ps0));
-    filter_value &= mask;
+  /* save bottom 3 bits so that we round one side +4 and the other +3
+   * if it equals 4 we'll set to adjust by -1 to account for the fact
+   * we'd round 3 the other way
+   */
+  Filter1 = vp8_signed_char_clamp(filter_value + 4);
+  Filter2 = vp8_signed_char_clamp(filter_value + 3);
+  Filter1 >>= 3;
+  Filter2 >>= 3;
+  u = vp8_signed_char_clamp(qs0 - Filter1);
+  *oq0 = u ^ 0x80;
+  u = vp8_signed_char_clamp(ps0 + Filter2);
+  *op0 = u ^ 0x80;
+  filter_value = Filter1;
 
-    /* save bottom 3 bits so that we round one side +4 and the other +3
-     * if it equals 4 we'll set to adjust by -1 to account for the fact
-     * we'd round 3 the other way
-     */
-    Filter1 = vp8_signed_char_clamp(filter_value + 4);
-    Filter2 = vp8_signed_char_clamp(filter_value + 3);
-    Filter1 >>= 3;
-    Filter2 >>= 3;
-    u = vp8_signed_char_clamp(qs0 - Filter1);
-    *oq0 = u ^ 0x80;
-    u = vp8_signed_char_clamp(ps0 + Filter2);
-    *op0 = u ^ 0x80;
-    filter_value = Filter1;
+  /* outer tap adjustments */
+  filter_value += 1;
+  filter_value >>= 1;
+  filter_value &= ~hev;
 
-    /* outer tap adjustments */
-    filter_value += 1;
-    filter_value >>= 1;
-    filter_value &= ~hev;
-
-    u = vp8_signed_char_clamp(qs1 - filter_value);
-    *oq1 = u ^ 0x80;
-    u = vp8_signed_char_clamp(ps1 + filter_value);
-    *op1 = u ^ 0x80;
-
+  u = vp8_signed_char_clamp(qs1 - filter_value);
+  *oq1 = u ^ 0x80;
+  u = vp8_signed_char_clamp(ps1 + filter_value);
+  *op1 = u ^ 0x80;
 }
-void vp8_loop_filter_horizontal_edge_c
-(
-    unsigned char *s,
-    int p, /* pitch */
-    const unsigned char *blimit,
-    const unsigned char *limit,
-    const unsigned char *thresh,
-    int count
-)
-{
-    int  hev = 0; /* high edge variance */
-    signed char mask = 0;
-    int i = 0;
+void vp8_loop_filter_horizontal_edge_c(unsigned char *s, int p, /* pitch */
+                                       const unsigned char *blimit,
+                                       const unsigned char *limit,
+                                       const unsigned char *thresh, int count) {
+  int hev = 0; /* high edge variance */
+  signed char mask = 0;
+  int i = 0;
 
-    /* loop filter designed to work using chars so that we can make maximum use
-     * of 8 bit simd instructions.
-     */
-    do
-    {
-        mask = vp8_filter_mask(limit[0], blimit[0],
-                               s[-4*p], s[-3*p], s[-2*p], s[-1*p],
-                               s[0*p], s[1*p], s[2*p], s[3*p]);
+  /* loop filter designed to work using chars so that we can make maximum use
+   * of 8 bit simd instructions.
+   */
+  do {
+    mask = vp8_filter_mask(limit[0], blimit[0], s[-4 * p], s[-3 * p], s[-2 * p],
+                           s[-1 * p], s[0 * p], s[1 * p], s[2 * p], s[3 * p]);
 
-        hev = vp8_hevmask(thresh[0], s[-2*p], s[-1*p], s[0*p], s[1*p]);
+    hev = vp8_hevmask(thresh[0], s[-2 * p], s[-1 * p], s[0 * p], s[1 * p]);
 
-        vp8_filter(mask, hev, s - 2 * p, s - 1 * p, s, s + 1 * p);
+    vp8_filter(mask, hev, s - 2 * p, s - 1 * p, s, s + 1 * p);
 
-        ++s;
-    }
-    while (++i < count * 8);
+    ++s;
+  } while (++i < count * 8);
 }
 
-void vp8_loop_filter_vertical_edge_c
-(
-    unsigned char *s,
-    int p,
-    const unsigned char *blimit,
-    const unsigned char *limit,
-    const unsigned char *thresh,
-    int count
-)
-{
-    int  hev = 0; /* high edge variance */
-    signed char mask = 0;
-    int i = 0;
+void vp8_loop_filter_vertical_edge_c(unsigned char *s, int p,
+                                     const unsigned char *blimit,
+                                     const unsigned char *limit,
+                                     const unsigned char *thresh, int count) {
+  int hev = 0; /* high edge variance */
+  signed char mask = 0;
+  int i = 0;
 
-    /* loop filter designed to work using chars so that we can make maximum use
-     * of 8 bit simd instructions.
-     */
-    do
-    {
-        mask = vp8_filter_mask(limit[0], blimit[0],
-                               s[-4], s[-3], s[-2], s[-1], s[0], s[1], s[2], s[3]);
+  /* loop filter designed to work using chars so that we can make maximum use
+   * of 8 bit simd instructions.
+   */
+  do {
+    mask = vp8_filter_mask(limit[0], blimit[0], s[-4], s[-3], s[-2], s[-1],
+                           s[0], s[1], s[2], s[3]);
 
-        hev = vp8_hevmask(thresh[0], s[-2], s[-1], s[0], s[1]);
+    hev = vp8_hevmask(thresh[0], s[-2], s[-1], s[0], s[1]);
 
-        vp8_filter(mask, hev, s - 2, s - 1, s, s + 1);
+    vp8_filter(mask, hev, s - 2, s - 1, s, s + 1);
 
-        s += p;
-    }
-    while (++i < count * 8);
+    s += p;
+  } while (++i < count * 8);
 }
 
-static void vp8_mbfilter(signed char mask, uc hev,
-                           uc *op2, uc *op1, uc *op0, uc *oq0, uc *oq1, uc *oq2)
-{
-    signed char s, u;
-    signed char filter_value, Filter1, Filter2;
-    signed char ps2 = (signed char) * op2 ^ 0x80;
-    signed char ps1 = (signed char) * op1 ^ 0x80;
-    signed char ps0 = (signed char) * op0 ^ 0x80;
-    signed char qs0 = (signed char) * oq0 ^ 0x80;
-    signed char qs1 = (signed char) * oq1 ^ 0x80;
-    signed char qs2 = (signed char) * oq2 ^ 0x80;
+static void vp8_mbfilter(signed char mask, uc hev, uc *op2, uc *op1, uc *op0,
+                         uc *oq0, uc *oq1, uc *oq2) {
+  signed char s, u;
+  signed char filter_value, Filter1, Filter2;
+  signed char ps2 = (signed char)*op2 ^ 0x80;
+  signed char ps1 = (signed char)*op1 ^ 0x80;
+  signed char ps0 = (signed char)*op0 ^ 0x80;
+  signed char qs0 = (signed char)*oq0 ^ 0x80;
+  signed char qs1 = (signed char)*oq1 ^ 0x80;
+  signed char qs2 = (signed char)*oq2 ^ 0x80;
 
-    /* add outer taps if we have high edge variance */
-    filter_value = vp8_signed_char_clamp(ps1 - qs1);
-    filter_value = vp8_signed_char_clamp(filter_value + 3 * (qs0 - ps0));
-    filter_value &= mask;
+  /* add outer taps if we have high edge variance */
+  filter_value = vp8_signed_char_clamp(ps1 - qs1);
+  filter_value = vp8_signed_char_clamp(filter_value + 3 * (qs0 - ps0));
+  filter_value &= mask;
 
-    Filter2 = filter_value;
-    Filter2 &= hev;
+  Filter2 = filter_value;
+  Filter2 &= hev;
 
-    /* save bottom 3 bits so that we round one side +4 and the other +3 */
-    Filter1 = vp8_signed_char_clamp(Filter2 + 4);
-    Filter2 = vp8_signed_char_clamp(Filter2 + 3);
-    Filter1 >>= 3;
-    Filter2 >>= 3;
-    qs0 = vp8_signed_char_clamp(qs0 - Filter1);
-    ps0 = vp8_signed_char_clamp(ps0 + Filter2);
+  /* save bottom 3 bits so that we round one side +4 and the other +3 */
+  Filter1 = vp8_signed_char_clamp(Filter2 + 4);
+  Filter2 = vp8_signed_char_clamp(Filter2 + 3);
+  Filter1 >>= 3;
+  Filter2 >>= 3;
+  qs0 = vp8_signed_char_clamp(qs0 - Filter1);
+  ps0 = vp8_signed_char_clamp(ps0 + Filter2);
 
+  /* only apply wider filter if not high edge variance */
+  filter_value &= ~hev;
+  Filter2 = filter_value;
 
-    /* only apply wider filter if not high edge variance */
-    filter_value &= ~hev;
-    Filter2 = filter_value;
+  /* roughly 3/7th difference across boundary */
+  u = vp8_signed_char_clamp((63 + Filter2 * 27) >> 7);
+  s = vp8_signed_char_clamp(qs0 - u);
+  *oq0 = s ^ 0x80;
+  s = vp8_signed_char_clamp(ps0 + u);
+  *op0 = s ^ 0x80;
 
-    /* roughly 3/7th difference across boundary */
-    u = vp8_signed_char_clamp((63 + Filter2 * 27) >> 7);
-    s = vp8_signed_char_clamp(qs0 - u);
-    *oq0 = s ^ 0x80;
-    s = vp8_signed_char_clamp(ps0 + u);
-    *op0 = s ^ 0x80;
+  /* roughly 2/7th difference across boundary */
+  u = vp8_signed_char_clamp((63 + Filter2 * 18) >> 7);
+  s = vp8_signed_char_clamp(qs1 - u);
+  *oq1 = s ^ 0x80;
+  s = vp8_signed_char_clamp(ps1 + u);
+  *op1 = s ^ 0x80;
 
-    /* roughly 2/7th difference across boundary */
-    u = vp8_signed_char_clamp((63 + Filter2 * 18) >> 7);
-    s = vp8_signed_char_clamp(qs1 - u);
-    *oq1 = s ^ 0x80;
-    s = vp8_signed_char_clamp(ps1 + u);
-    *op1 = s ^ 0x80;
-
-    /* roughly 1/7th difference across boundary */
-    u = vp8_signed_char_clamp((63 + Filter2 * 9) >> 7);
-    s = vp8_signed_char_clamp(qs2 - u);
-    *oq2 = s ^ 0x80;
-    s = vp8_signed_char_clamp(ps2 + u);
-    *op2 = s ^ 0x80;
+  /* roughly 1/7th difference across boundary */
+  u = vp8_signed_char_clamp((63 + Filter2 * 9) >> 7);
+  s = vp8_signed_char_clamp(qs2 - u);
+  *oq2 = s ^ 0x80;
+  s = vp8_signed_char_clamp(ps2 + u);
+  *op2 = s ^ 0x80;
 }
 
-void vp8_mbloop_filter_horizontal_edge_c
-(
-    unsigned char *s,
-    int p,
-    const unsigned char *blimit,
-    const unsigned char *limit,
-    const unsigned char *thresh,
-    int count
-)
-{
-    signed char hev = 0; /* high edge variance */
-    signed char mask = 0;
-    int i = 0;
+void vp8_mbloop_filter_horizontal_edge_c(unsigned char *s, int p,
+                                         const unsigned char *blimit,
+                                         const unsigned char *limit,
+                                         const unsigned char *thresh,
+                                         int count) {
+  signed char hev = 0; /* high edge variance */
+  signed char mask = 0;
+  int i = 0;
 
-    /* loop filter designed to work using chars so that we can make maximum use
-     * of 8 bit simd instructions.
-     */
-    do
-    {
+  /* loop filter designed to work using chars so that we can make maximum use
+   * of 8 bit simd instructions.
+   */
+  do {
+    mask = vp8_filter_mask(limit[0], blimit[0], s[-4 * p], s[-3 * p], s[-2 * p],
+                           s[-1 * p], s[0 * p], s[1 * p], s[2 * p], s[3 * p]);
 
-        mask = vp8_filter_mask(limit[0], blimit[0],
-                               s[-4*p], s[-3*p], s[-2*p], s[-1*p],
-                               s[0*p], s[1*p], s[2*p], s[3*p]);
+    hev = vp8_hevmask(thresh[0], s[-2 * p], s[-1 * p], s[0 * p], s[1 * p]);
 
-        hev = vp8_hevmask(thresh[0], s[-2*p], s[-1*p], s[0*p], s[1*p]);
+    vp8_mbfilter(mask, hev, s - 3 * p, s - 2 * p, s - 1 * p, s, s + 1 * p,
+                 s + 2 * p);
 
-        vp8_mbfilter(mask, hev, s - 3 * p, s - 2 * p, s - 1 * p, s, s + 1 * p, s + 2 * p);
-
-        ++s;
-    }
-    while (++i < count * 8);
-
+    ++s;
+  } while (++i < count * 8);
 }
 
+void vp8_mbloop_filter_vertical_edge_c(unsigned char *s, int p,
+                                       const unsigned char *blimit,
+                                       const unsigned char *limit,
+                                       const unsigned char *thresh, int count) {
+  signed char hev = 0; /* high edge variance */
+  signed char mask = 0;
+  int i = 0;
 
-void vp8_mbloop_filter_vertical_edge_c
-(
-    unsigned char *s,
-    int p,
-    const unsigned char *blimit,
-    const unsigned char *limit,
-    const unsigned char *thresh,
-    int count
-)
-{
-    signed char hev = 0; /* high edge variance */
-    signed char mask = 0;
-    int i = 0;
+  do {
+    mask = vp8_filter_mask(limit[0], blimit[0], s[-4], s[-3], s[-2], s[-1],
+                           s[0], s[1], s[2], s[3]);
 
-    do
-    {
+    hev = vp8_hevmask(thresh[0], s[-2], s[-1], s[0], s[1]);
 
-        mask = vp8_filter_mask(limit[0], blimit[0],
-                               s[-4], s[-3], s[-2], s[-1], s[0], s[1], s[2], s[3]);
+    vp8_mbfilter(mask, hev, s - 3, s - 2, s - 1, s, s + 1, s + 2);
 
-        hev = vp8_hevmask(thresh[0], s[-2], s[-1], s[0], s[1]);
-
-        vp8_mbfilter(mask, hev, s - 3, s - 2, s - 1, s, s + 1, s + 2);
-
-        s += p;
-    }
-    while (++i < count * 8);
-
+    s += p;
+  } while (++i < count * 8);
 }
 
 /* should we apply any filter at all ( 11111111 yes, 00000000 no) */
-static signed char vp8_simple_filter_mask(uc blimit, uc p1, uc p0, uc q0, uc q1)
-{
-/* Why does this cause problems for win32?
- * error C2143: syntax error : missing ';' before 'type'
- *  (void) limit;
- */
-    signed char mask = (abs(p0 - q0) * 2 + abs(p1 - q1) / 2  <= blimit) * -1;
-    return mask;
+static signed char vp8_simple_filter_mask(uc blimit, uc p1, uc p0, uc q0,
+                                          uc q1) {
+  /* Why does this cause problems for win32?
+   * error C2143: syntax error : missing ';' before 'type'
+   *  (void) limit;
+   */
+  signed char mask = (abs(p0 - q0) * 2 + abs(p1 - q1) / 2 <= blimit) * -1;
+  return mask;
 }
 
-static void vp8_simple_filter(signed char mask, uc *op1, uc *op0, uc *oq0, uc *oq1)
-{
-    signed char filter_value, Filter1, Filter2;
-    signed char p1 = (signed char) * op1 ^ 0x80;
-    signed char p0 = (signed char) * op0 ^ 0x80;
-    signed char q0 = (signed char) * oq0 ^ 0x80;
-    signed char q1 = (signed char) * oq1 ^ 0x80;
-    signed char u;
+static void vp8_simple_filter(signed char mask, uc *op1, uc *op0, uc *oq0,
+                              uc *oq1) {
+  signed char filter_value, Filter1, Filter2;
+  signed char p1 = (signed char)*op1 ^ 0x80;
+  signed char p0 = (signed char)*op0 ^ 0x80;
+  signed char q0 = (signed char)*oq0 ^ 0x80;
+  signed char q1 = (signed char)*oq1 ^ 0x80;
+  signed char u;
 
-    filter_value = vp8_signed_char_clamp(p1 - q1);
-    filter_value = vp8_signed_char_clamp(filter_value + 3 * (q0 - p0));
-    filter_value &= mask;
+  filter_value = vp8_signed_char_clamp(p1 - q1);
+  filter_value = vp8_signed_char_clamp(filter_value + 3 * (q0 - p0));
+  filter_value &= mask;
 
-    /* save bottom 3 bits so that we round one side +4 and the other +3 */
-    Filter1 = vp8_signed_char_clamp(filter_value + 4);
-    Filter1 >>= 3;
-    u = vp8_signed_char_clamp(q0 - Filter1);
-    *oq0  = u ^ 0x80;
+  /* save bottom 3 bits so that we round one side +4 and the other +3 */
+  Filter1 = vp8_signed_char_clamp(filter_value + 4);
+  Filter1 >>= 3;
+  u = vp8_signed_char_clamp(q0 - Filter1);
+  *oq0 = u ^ 0x80;
 
-    Filter2 = vp8_signed_char_clamp(filter_value + 3);
-    Filter2 >>= 3;
-    u = vp8_signed_char_clamp(p0 + Filter2);
-    *op0 = u ^ 0x80;
+  Filter2 = vp8_signed_char_clamp(filter_value + 3);
+  Filter2 >>= 3;
+  u = vp8_signed_char_clamp(p0 + Filter2);
+  *op0 = u ^ 0x80;
 }
 
-void vp8_loop_filter_simple_horizontal_edge_c
-(
-    unsigned char *s,
-    int p,
-    const unsigned char *blimit
-)
-{
-    signed char mask = 0;
-    int i = 0;
+void vp8_loop_filter_simple_horizontal_edge_c(unsigned char *s, int p,
+                                              const unsigned char *blimit) {
+  signed char mask = 0;
+  int i = 0;
 
-    do
-    {
-        mask = vp8_simple_filter_mask(blimit[0], s[-2*p], s[-1*p], s[0*p], s[1*p]);
-        vp8_simple_filter(mask, s - 2 * p, s - 1 * p, s, s + 1 * p);
-        ++s;
-    }
-    while (++i < 16);
+  do {
+    mask = vp8_simple_filter_mask(blimit[0], s[-2 * p], s[-1 * p], s[0 * p],
+                                  s[1 * p]);
+    vp8_simple_filter(mask, s - 2 * p, s - 1 * p, s, s + 1 * p);
+    ++s;
+  } while (++i < 16);
 }
 
-void vp8_loop_filter_simple_vertical_edge_c
-(
-    unsigned char *s,
-    int p,
-    const unsigned char *blimit
-)
-{
-    signed char mask = 0;
-    int i = 0;
+void vp8_loop_filter_simple_vertical_edge_c(unsigned char *s, int p,
+                                            const unsigned char *blimit) {
+  signed char mask = 0;
+  int i = 0;
 
-    do
-    {
-        mask = vp8_simple_filter_mask(blimit[0], s[-2], s[-1], s[0], s[1]);
-        vp8_simple_filter(mask, s - 2, s - 1, s, s + 1);
-        s += p;
-    }
-    while (++i < 16);
-
+  do {
+    mask = vp8_simple_filter_mask(blimit[0], s[-2], s[-1], s[0], s[1]);
+    vp8_simple_filter(mask, s - 2, s - 1, s, s + 1);
+    s += p;
+  } while (++i < 16);
 }
 
 /* Horizontal MB filtering */
 void vp8_loop_filter_mbh_c(unsigned char *y_ptr, unsigned char *u_ptr,
                            unsigned char *v_ptr, int y_stride, int uv_stride,
-                           loop_filter_info *lfi)
-{
-    vp8_mbloop_filter_horizontal_edge_c(y_ptr, y_stride, lfi->mblim, lfi->lim, lfi->hev_thr, 2);
+                           loop_filter_info *lfi) {
+  vp8_mbloop_filter_horizontal_edge_c(y_ptr, y_stride, lfi->mblim, lfi->lim,
+                                      lfi->hev_thr, 2);
 
-    if (u_ptr)
-        vp8_mbloop_filter_horizontal_edge_c(u_ptr, uv_stride, lfi->mblim, lfi->lim, lfi->hev_thr, 1);
+  if (u_ptr)
+    vp8_mbloop_filter_horizontal_edge_c(u_ptr, uv_stride, lfi->mblim, lfi->lim,
+                                        lfi->hev_thr, 1);
 
-    if (v_ptr)
-        vp8_mbloop_filter_horizontal_edge_c(v_ptr, uv_stride, lfi->mblim, lfi->lim, lfi->hev_thr, 1);
+  if (v_ptr)
+    vp8_mbloop_filter_horizontal_edge_c(v_ptr, uv_stride, lfi->mblim, lfi->lim,
+                                        lfi->hev_thr, 1);
 }
 
 /* Vertical MB Filtering */
 void vp8_loop_filter_mbv_c(unsigned char *y_ptr, unsigned char *u_ptr,
                            unsigned char *v_ptr, int y_stride, int uv_stride,
-                           loop_filter_info *lfi)
-{
-    vp8_mbloop_filter_vertical_edge_c(y_ptr, y_stride, lfi->mblim, lfi->lim, lfi->hev_thr, 2);
+                           loop_filter_info *lfi) {
+  vp8_mbloop_filter_vertical_edge_c(y_ptr, y_stride, lfi->mblim, lfi->lim,
+                                    lfi->hev_thr, 2);
 
-    if (u_ptr)
-        vp8_mbloop_filter_vertical_edge_c(u_ptr, uv_stride, lfi->mblim, lfi->lim, lfi->hev_thr, 1);
+  if (u_ptr)
+    vp8_mbloop_filter_vertical_edge_c(u_ptr, uv_stride, lfi->mblim, lfi->lim,
+                                      lfi->hev_thr, 1);
 
-    if (v_ptr)
-        vp8_mbloop_filter_vertical_edge_c(v_ptr, uv_stride, lfi->mblim, lfi->lim, lfi->hev_thr, 1);
+  if (v_ptr)
+    vp8_mbloop_filter_vertical_edge_c(v_ptr, uv_stride, lfi->mblim, lfi->lim,
+                                      lfi->hev_thr, 1);
 }
 
 /* Horizontal B Filtering */
 void vp8_loop_filter_bh_c(unsigned char *y_ptr, unsigned char *u_ptr,
                           unsigned char *v_ptr, int y_stride, int uv_stride,
-                          loop_filter_info *lfi)
-{
-    vp8_loop_filter_horizontal_edge_c(y_ptr + 4 * y_stride, y_stride, lfi->blim, lfi->lim, lfi->hev_thr, 2);
-    vp8_loop_filter_horizontal_edge_c(y_ptr + 8 * y_stride, y_stride, lfi->blim, lfi->lim, lfi->hev_thr, 2);
-    vp8_loop_filter_horizontal_edge_c(y_ptr + 12 * y_stride, y_stride, lfi->blim, lfi->lim, lfi->hev_thr, 2);
+                          loop_filter_info *lfi) {
+  vp8_loop_filter_horizontal_edge_c(y_ptr + 4 * y_stride, y_stride, lfi->blim,
+                                    lfi->lim, lfi->hev_thr, 2);
+  vp8_loop_filter_horizontal_edge_c(y_ptr + 8 * y_stride, y_stride, lfi->blim,
+                                    lfi->lim, lfi->hev_thr, 2);
+  vp8_loop_filter_horizontal_edge_c(y_ptr + 12 * y_stride, y_stride, lfi->blim,
+                                    lfi->lim, lfi->hev_thr, 2);
 
-    if (u_ptr)
-        vp8_loop_filter_horizontal_edge_c(u_ptr + 4 * uv_stride, uv_stride, lfi->blim, lfi->lim, lfi->hev_thr, 1);
+  if (u_ptr)
+    vp8_loop_filter_horizontal_edge_c(u_ptr + 4 * uv_stride, uv_stride,
+                                      lfi->blim, lfi->lim, lfi->hev_thr, 1);
 
-    if (v_ptr)
-        vp8_loop_filter_horizontal_edge_c(v_ptr + 4 * uv_stride, uv_stride, lfi->blim, lfi->lim, lfi->hev_thr, 1);
+  if (v_ptr)
+    vp8_loop_filter_horizontal_edge_c(v_ptr + 4 * uv_stride, uv_stride,
+                                      lfi->blim, lfi->lim, lfi->hev_thr, 1);
 }
 
 void vp8_loop_filter_bhs_c(unsigned char *y_ptr, int y_stride,
-                           const unsigned char *blimit)
-{
-    vp8_loop_filter_simple_horizontal_edge_c(y_ptr + 4 * y_stride, y_stride, blimit);
-    vp8_loop_filter_simple_horizontal_edge_c(y_ptr + 8 * y_stride, y_stride, blimit);
-    vp8_loop_filter_simple_horizontal_edge_c(y_ptr + 12 * y_stride, y_stride, blimit);
+                           const unsigned char *blimit) {
+  vp8_loop_filter_simple_horizontal_edge_c(y_ptr + 4 * y_stride, y_stride,
+                                           blimit);
+  vp8_loop_filter_simple_horizontal_edge_c(y_ptr + 8 * y_stride, y_stride,
+                                           blimit);
+  vp8_loop_filter_simple_horizontal_edge_c(y_ptr + 12 * y_stride, y_stride,
+                                           blimit);
 }
 
 /* Vertical B Filtering */
 void vp8_loop_filter_bv_c(unsigned char *y_ptr, unsigned char *u_ptr,
                           unsigned char *v_ptr, int y_stride, int uv_stride,
-                          loop_filter_info *lfi)
-{
-    vp8_loop_filter_vertical_edge_c(y_ptr + 4, y_stride, lfi->blim, lfi->lim, lfi->hev_thr, 2);
-    vp8_loop_filter_vertical_edge_c(y_ptr + 8, y_stride, lfi->blim, lfi->lim, lfi->hev_thr, 2);
-    vp8_loop_filter_vertical_edge_c(y_ptr + 12, y_stride, lfi->blim, lfi->lim, lfi->hev_thr, 2);
+                          loop_filter_info *lfi) {
+  vp8_loop_filter_vertical_edge_c(y_ptr + 4, y_stride, lfi->blim, lfi->lim,
+                                  lfi->hev_thr, 2);
+  vp8_loop_filter_vertical_edge_c(y_ptr + 8, y_stride, lfi->blim, lfi->lim,
+                                  lfi->hev_thr, 2);
+  vp8_loop_filter_vertical_edge_c(y_ptr + 12, y_stride, lfi->blim, lfi->lim,
+                                  lfi->hev_thr, 2);
 
-    if (u_ptr)
-        vp8_loop_filter_vertical_edge_c(u_ptr + 4, uv_stride, lfi->blim, lfi->lim, lfi->hev_thr, 1);
+  if (u_ptr)
+    vp8_loop_filter_vertical_edge_c(u_ptr + 4, uv_stride, lfi->blim, lfi->lim,
+                                    lfi->hev_thr, 1);
 
-    if (v_ptr)
-        vp8_loop_filter_vertical_edge_c(v_ptr + 4, uv_stride, lfi->blim, lfi->lim, lfi->hev_thr, 1);
+  if (v_ptr)
+    vp8_loop_filter_vertical_edge_c(v_ptr + 4, uv_stride, lfi->blim, lfi->lim,
+                                    lfi->hev_thr, 1);
 }
 
 void vp8_loop_filter_bvs_c(unsigned char *y_ptr, int y_stride,
-                           const unsigned char *blimit)
-{
-    vp8_loop_filter_simple_vertical_edge_c(y_ptr + 4, y_stride, blimit);
-    vp8_loop_filter_simple_vertical_edge_c(y_ptr + 8, y_stride, blimit);
-    vp8_loop_filter_simple_vertical_edge_c(y_ptr + 12, y_stride, blimit);
+                           const unsigned char *blimit) {
+  vp8_loop_filter_simple_vertical_edge_c(y_ptr + 4, y_stride, blimit);
+  vp8_loop_filter_simple_vertical_edge_c(y_ptr + 8, y_stride, blimit);
+  vp8_loop_filter_simple_vertical_edge_c(y_ptr + 12, y_stride, blimit);
 }
--- a/vp8/common/mbpitch.c
+++ b/vp8/common/mbpitch.c
@@ -8,61 +8,50 @@
  *  be found in the AUTHORS file in the root of the source tree.
  */
 
-
 #include "blockd.h"
 
-void vp8_setup_block_dptrs(MACROBLOCKD *x)
-{
-    int r, c;
+void vp8_setup_block_dptrs(MACROBLOCKD *x) {
+  int r, c;
 
-    for (r = 0; r < 4; r++)
-    {
-        for (c = 0; c < 4; c++)
-        {
-            x->block[r*4+c].predictor = x->predictor + r * 4 * 16 + c * 4;
-        }
+  for (r = 0; r < 4; r++) {
+    for (c = 0; c < 4; c++) {
+      x->block[r * 4 + c].predictor = x->predictor + r * 4 * 16 + c * 4;
     }
+  }
 
-    for (r = 0; r < 2; r++)
-    {
-        for (c = 0; c < 2; c++)
-        {
-            x->block[16+r*2+c].predictor = x->predictor + 256 + r * 4 * 8 + c * 4;
-
-        }
+  for (r = 0; r < 2; r++) {
+    for (c = 0; c < 2; c++) {
+      x->block[16 + r * 2 + c].predictor =
+          x->predictor + 256 + r * 4 * 8 + c * 4;
     }
+  }
 
-    for (r = 0; r < 2; r++)
-    {
-        for (c = 0; c < 2; c++)
-        {
-            x->block[20+r*2+c].predictor = x->predictor + 320 + r * 4 * 8 + c * 4;
-
-        }
+  for (r = 0; r < 2; r++) {
+    for (c = 0; c < 2; c++) {
+      x->block[20 + r * 2 + c].predictor =
+          x->predictor + 320 + r * 4 * 8 + c * 4;
     }
+  }
 
-    for (r = 0; r < 25; r++)
-    {
-        x->block[r].qcoeff  = x->qcoeff  + r * 16;
-        x->block[r].dqcoeff = x->dqcoeff + r * 16;
-        x->block[r].eob     = x->eobs + r;
-    }
+  for (r = 0; r < 25; r++) {
+    x->block[r].qcoeff = x->qcoeff + r * 16;
+    x->block[r].dqcoeff = x->dqcoeff + r * 16;
+    x->block[r].eob = x->eobs + r;
+  }
 }
 
-void vp8_build_block_doffsets(MACROBLOCKD *x)
-{
-    int block;
+void vp8_build_block_doffsets(MACROBLOCKD *x) {
+  int block;
 
-    for (block = 0; block < 16; block++) /* y blocks */
-    {
-        x->block[block].offset =
-            (block >> 2) * 4 * x->dst.y_stride + (block & 3) * 4;
-    }
+  for (block = 0; block < 16; block++) /* y blocks */
+  {
+    x->block[block].offset =
+        (block >> 2) * 4 * x->dst.y_stride + (block & 3) * 4;
+  }
 
-    for (block = 16; block < 20; block++) /* U and V blocks */
-    {
-        x->block[block+4].offset =
-        x->block[block].offset =
-            ((block - 16) >> 1) * 4 * x->dst.uv_stride + (block & 1) * 4;
-    }
+  for (block = 16; block < 20; block++) /* U and V blocks */
+  {
+    x->block[block + 4].offset = x->block[block].offset =
+        ((block - 16) >> 1) * 4 * x->dst.uv_stride + (block & 1) * 4;
+  }
 }
--- a/vp8/common/mfqe.c
+++ b/vp8/common/mfqe.c
@@ -8,7 +8,6 @@
  *  be found in the AUTHORS file in the root of the source tree.
  */
 
-
 /* MFQE: Multiframe Quality Enhancement
  * In rate limited situations keyframes may cause significant visual artifacts
  * commonly referred to as "popping." This file implements a postproccesing
@@ -28,359 +27,299 @@
 #include <stdlib.h>
 
 static void filter_by_weight(unsigned char *src, int src_stride,
-                             unsigned char *dst, int dst_stride,
-                             int block_size, int src_weight)
-{
-    int dst_weight = (1 << MFQE_PRECISION) - src_weight;
-    int rounding_bit = 1 << (MFQE_PRECISION - 1);
-    int r, c;
+                             unsigned char *dst, int dst_stride, int block_size,
+                             int src_weight) {
+  int dst_weight = (1 << MFQE_PRECISION) - src_weight;
+  int rounding_bit = 1 << (MFQE_PRECISION - 1);
+  int r, c;
 
-    for (r = 0; r < block_size; r++)
-    {
-        for (c = 0; c < block_size; c++)
-        {
-            dst[c] = (src[c] * src_weight +
-                      dst[c] * dst_weight +
-                      rounding_bit) >> MFQE_PRECISION;
-        }
-        src += src_stride;
-        dst += dst_stride;
+  for (r = 0; r < block_size; r++) {
+    for (c = 0; c < block_size; c++) {
+      dst[c] = (src[c] * src_weight + dst[c] * dst_weight + rounding_bit) >>
+               MFQE_PRECISION;
     }
+    src += src_stride;
+    dst += dst_stride;
+  }
 }
 
 void vp8_filter_by_weight16x16_c(unsigned char *src, int src_stride,
                                  unsigned char *dst, int dst_stride,
-                                 int src_weight)
-{
-    filter_by_weight(src, src_stride, dst, dst_stride, 16, src_weight);
+                                 int src_weight) {
+  filter_by_weight(src, src_stride, dst, dst_stride, 16, src_weight);
 }
 
 void vp8_filter_by_weight8x8_c(unsigned char *src, int src_stride,
                                unsigned char *dst, int dst_stride,
-                               int src_weight)
-{
-    filter_by_weight(src, src_stride, dst, dst_stride, 8, src_weight);
+                               int src_weight) {
+  filter_by_weight(src, src_stride, dst, dst_stride, 8, src_weight);
 }
 
 void vp8_filter_by_weight4x4_c(unsigned char *src, int src_stride,
                                unsigned char *dst, int dst_stride,
-                               int src_weight)
-{
-    filter_by_weight(src, src_stride, dst, dst_stride, 4, src_weight);
+                               int src_weight) {
+  filter_by_weight(src, src_stride, dst, dst_stride, 4, src_weight);
 }
 
-static void apply_ifactor(unsigned char *y_src,
-                          int y_src_stride,
-                          unsigned char *y_dst,
-                          int y_dst_stride,
-                          unsigned char *u_src,
-                          unsigned char *v_src,
-                          int uv_src_stride,
-                          unsigned char *u_dst,
-                          unsigned char *v_dst,
-                          int uv_dst_stride,
-                          int block_size,
-                          int src_weight)
-{
-    if (block_size == 16)
-    {
-        vp8_filter_by_weight16x16(y_src, y_src_stride, y_dst, y_dst_stride, src_weight);
-        vp8_filter_by_weight8x8(u_src, uv_src_stride, u_dst, uv_dst_stride, src_weight);
-        vp8_filter_by_weight8x8(v_src, uv_src_stride, v_dst, uv_dst_stride, src_weight);
-    }
-    else /* if (block_size == 8) */
-    {
-        vp8_filter_by_weight8x8(y_src, y_src_stride, y_dst, y_dst_stride, src_weight);
-        vp8_filter_by_weight4x4(u_src, uv_src_stride, u_dst, uv_dst_stride, src_weight);
-        vp8_filter_by_weight4x4(v_src, uv_src_stride, v_dst, uv_dst_stride, src_weight);
-    }
+static void apply_ifactor(unsigned char *y_src, int y_src_stride,
+                          unsigned char *y_dst, int y_dst_stride,
+                          unsigned char *u_src, unsigned char *v_src,
+                          int uv_src_stride, unsigned char *u_dst,
+                          unsigned char *v_dst, int uv_dst_stride,
+                          int block_size, int src_weight) {
+  if (block_size == 16) {
+    vp8_filter_by_weight16x16(y_src, y_src_stride, y_dst, y_dst_stride,
+                              src_weight);
+    vp8_filter_by_weight8x8(u_src, uv_src_stride, u_dst, uv_dst_stride,
+                            src_weight);
+    vp8_filter_by_weight8x8(v_src, uv_src_stride, v_dst, uv_dst_stride,
+                            src_weight);
+  } else /* if (block_size == 8) */
+  {
+    vp8_filter_by_weight8x8(y_src, y_src_stride, y_dst, y_dst_stride,
+                            src_weight);
+    vp8_filter_by_weight4x4(u_src, uv_src_stride, u_dst, uv_dst_stride,
+                            src_weight);
+    vp8_filter_by_weight4x4(v_src, uv_src_stride, v_dst, uv_dst_stride,
+                            src_weight);
+  }
 }
 
-static unsigned int int_sqrt(unsigned int x)
-{
-    unsigned int y = x;
-    unsigned int guess;
-    int p = 1;
-    while (y>>=1) p++;
-    p>>=1;
+static unsigned int int_sqrt(unsigned int x) {
+  unsigned int y = x;
+  unsigned int guess;
+  int p = 1;
+  while (y >>= 1) p++;
+  p >>= 1;
 
-    guess=0;
-    while (p>=0)
-    {
-        guess |= (1<<p);
-        if (x<guess*guess)
-            guess -= (1<<p);
-        p--;
-    }
-    /* choose between guess or guess+1 */
-    return guess+(guess*guess+guess+1<=x);
+  guess = 0;
+  while (p >= 0) {
+    guess |= (1 << p);
+    if (x < guess * guess) guess -= (1 << p);
+    p--;
+  }
+  /* choose between guess or guess+1 */
+  return guess + (guess * guess + guess + 1 <= x);
 }
 
 #define USE_SSD
-static void multiframe_quality_enhance_block
-(
+static void multiframe_quality_enhance_block(
     int blksize, /* Currently only values supported are 16, 8 */
-    int qcurr,
-    int qprev,
-    unsigned char *y,
-    unsigned char *u,
-    unsigned char *v,
-    int y_stride,
-    int uv_stride,
-    unsigned char *yd,
-    unsigned char *ud,
-    unsigned char *vd,
-    int yd_stride,
-    int uvd_stride
-)
-{
-    static const unsigned char VP8_ZEROS[16]=
-    {
-         0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0
-    };
-    int uvblksize = blksize >> 1;
-    int qdiff = qcurr - qprev;
+    int qcurr, int qprev, unsigned char *y, unsigned char *u, unsigned char *v,
+    int y_stride, int uv_stride, unsigned char *yd, unsigned char *ud,
+    unsigned char *vd, int yd_stride, int uvd_stride) {
+  static const unsigned char VP8_ZEROS[16] = { 0, 0, 0, 0, 0, 0, 0, 0,
+                                               0, 0, 0, 0, 0, 0, 0, 0 };
+  int uvblksize = blksize >> 1;
+  int qdiff = qcurr - qprev;
 
-    int i;
-    unsigned char *up;
-    unsigned char *udp;
-    unsigned char *vp;
-    unsigned char *vdp;
+  int i;
+  unsigned char *up;
+  unsigned char *udp;
+  unsigned char *vp;
+  unsigned char *vdp;
 
-    unsigned int act, actd, sad, usad, vsad, sse, thr, thrsq, actrisk;
+  unsigned int act, actd, sad, usad, vsad, sse, thr, thrsq, actrisk;
 
-    if (blksize == 16)
-    {
-        actd = (vpx_variance16x16(yd, yd_stride, VP8_ZEROS, 0, &sse)+128)>>8;
-        act = (vpx_variance16x16(y, y_stride, VP8_ZEROS, 0, &sse)+128)>>8;
+  if (blksize == 16) {
+    actd = (vpx_variance16x16(yd, yd_stride, VP8_ZEROS, 0, &sse) + 128) >> 8;
+    act = (vpx_variance16x16(y, y_stride, VP8_ZEROS, 0, &sse) + 128) >> 8;
 #ifdef USE_SSD
-        vpx_variance16x16(y, y_stride, yd, yd_stride, &sse);
-        sad = (sse + 128)>>8;
-        vpx_variance8x8(u, uv_stride, ud, uvd_stride, &sse);
-        usad = (sse + 32)>>6;
-        vpx_variance8x8(v, uv_stride, vd, uvd_stride, &sse);
-        vsad = (sse + 32)>>6;
+    vpx_variance16x16(y, y_stride, yd, yd_stride, &sse);
+    sad = (sse + 128) >> 8;
+    vpx_variance8x8(u, uv_stride, ud, uvd_stride, &sse);
+    usad = (sse + 32) >> 6;
+    vpx_variance8x8(v, uv_stride, vd, uvd_stride, &sse);
+    vsad = (sse + 32) >> 6;
 #else
-        sad = (vpx_sad16x16(y, y_stride, yd, yd_stride) + 128) >> 8;
-        usad = (vpx_sad8x8(u, uv_stride, ud, uvd_stride) + 32) >> 6;
-        vsad = (vpx_sad8x8(v, uv_stride, vd, uvd_stride)+ 32) >> 6;
+    sad = (vpx_sad16x16(y, y_stride, yd, yd_stride) + 128) >> 8;
+    usad = (vpx_sad8x8(u, uv_stride, ud, uvd_stride) + 32) >> 6;
+    vsad = (vpx_sad8x8(v, uv_stride, vd, uvd_stride) + 32) >> 6;
 #endif
-    }
-    else /* if (blksize == 8) */
-    {
-        actd = (vpx_variance8x8(yd, yd_stride, VP8_ZEROS, 0, &sse)+32)>>6;
-        act = (vpx_variance8x8(y, y_stride, VP8_ZEROS, 0, &sse)+32)>>6;
+  } else /* if (blksize == 8) */
+  {
+    actd = (vpx_variance8x8(yd, yd_stride, VP8_ZEROS, 0, &sse) + 32) >> 6;
+    act = (vpx_variance8x8(y, y_stride, VP8_ZEROS, 0, &sse) + 32) >> 6;
 #ifdef USE_SSD
-        vpx_variance8x8(y, y_stride, yd, yd_stride, &sse);
-        sad = (sse + 32)>>6;
-        vpx_variance4x4(u, uv_stride, ud, uvd_stride, &sse);
-        usad = (sse + 8)>>4;
-        vpx_variance4x4(v, uv_stride, vd, uvd_stride, &sse);
-        vsad = (sse + 8)>>4;
+    vpx_variance8x8(y, y_stride, yd, yd_stride, &sse);
+    sad = (sse + 32) >> 6;
+    vpx_variance4x4(u, uv_stride, ud, uvd_stride, &sse);
+    usad = (sse + 8) >> 4;
+    vpx_variance4x4(v, uv_stride, vd, uvd_stride, &sse);
+    vsad = (sse + 8) >> 4;
 #else
-        sad = (vpx_sad8x8(y, y_stride, yd, yd_stride) + 32) >> 6;
-        usad = (vpx_sad4x4(u, uv_stride, ud, uvd_stride) + 8) >> 4;
-        vsad = (vpx_sad4x4(v, uv_stride, vd, uvd_stride) + 8) >> 4;
+    sad = (vpx_sad8x8(y, y_stride, yd, yd_stride) + 32) >> 6;
+    usad = (vpx_sad4x4(u, uv_stride, ud, uvd_stride) + 8) >> 4;
+    vsad = (vpx_sad4x4(v, uv_stride, vd, uvd_stride) + 8) >> 4;
 #endif
-    }
+  }
 
-    actrisk = (actd > act * 5);
+  actrisk = (actd > act * 5);
 
-    /* thr = qdiff/16 + log2(act) + log4(qprev) */
-    thr = (qdiff >> 4);
-    while (actd >>= 1) thr++;
-    while (qprev >>= 2) thr++;
+  /* thr = qdiff/16 + log2(act) + log4(qprev) */
+  thr = (qdiff >> 4);
+  while (actd >>= 1) thr++;
+  while (qprev >>= 2) thr++;
 
 #ifdef USE_SSD
-    thrsq = thr * thr;
-    if (sad < thrsq &&
-        /* additional checks for color mismatch and excessive addition of
-         * high-frequencies */
-        4 * usad < thrsq && 4 * vsad < thrsq && !actrisk)
+  thrsq = thr * thr;
+  if (sad < thrsq &&
+      /* additional checks for color mismatch and excessive addition of
+       * high-frequencies */
+      4 * usad < thrsq && 4 * vsad < thrsq && !actrisk)
 #else
-    if (sad < thr &&
-        /* additional checks for color mismatch and excessive addition of
-         * high-frequencies */
-        2 * usad < thr && 2 * vsad < thr && !actrisk)
+  if (sad < thr &&
+      /* additional checks for color mismatch and excessive addition of
+       * high-frequencies */
+      2 * usad < thr && 2 * vsad < thr && !actrisk)
 #endif
-    {
-        int ifactor;
+  {
+    int ifactor;
 #ifdef USE_SSD
-        /* TODO: optimize this later to not need sqr root */
-        sad = int_sqrt(sad);
+    /* TODO: optimize this later to not need sqr root */
+    sad = int_sqrt(sad);
 #endif
-        ifactor = (sad << MFQE_PRECISION) / thr;
-        ifactor >>= (qdiff >> 5);
+    ifactor = (sad << MFQE_PRECISION) / thr;
+    ifactor >>= (qdiff >> 5);
 
-        if (ifactor)
-        {
-            apply_ifactor(y, y_stride, yd, yd_stride,
-                          u, v, uv_stride,
-                          ud, vd, uvd_stride,
-                          blksize, ifactor);
-        }
+    if (ifactor) {
+      apply_ifactor(y, y_stride, yd, yd_stride, u, v, uv_stride, ud, vd,
+                    uvd_stride, blksize, ifactor);
     }
-    else  /* else implicitly copy from previous frame */
+  } else /* else implicitly copy from previous frame */
+  {
+    if (blksize == 16) {
+      vp8_copy_mem16x16(y, y_stride, yd, yd_stride);
+      vp8_copy_mem8x8(u, uv_stride, ud, uvd_stride);
+      vp8_copy_mem8x8(v, uv_stride, vd, uvd_stride);
+    } else /* if (blksize == 8) */
     {
-        if (blksize == 16)
-        {
-            vp8_copy_mem16x16(y, y_stride, yd, yd_stride);
-            vp8_copy_mem8x8(u, uv_stride, ud, uvd_stride);
-            vp8_copy_mem8x8(v, uv_stride, vd, uvd_stride);
-        }
-        else  /* if (blksize == 8) */
-        {
-            vp8_copy_mem8x8(y, y_stride, yd, yd_stride);
-            for (up = u, udp = ud, i = 0; i < uvblksize; ++i, up += uv_stride, udp += uvd_stride)
-                memcpy(udp, up, uvblksize);
-            for (vp = v, vdp = vd, i = 0; i < uvblksize; ++i, vp += uv_stride, vdp += uvd_stride)
-                memcpy(vdp, vp, uvblksize);
-        }
+      vp8_copy_mem8x8(y, y_stride, yd, yd_stride);
+      for (up = u, udp = ud, i = 0; i < uvblksize;
+           ++i, up += uv_stride, udp += uvd_stride)
+        memcpy(udp, up, uvblksize);
+      for (vp = v, vdp = vd, i = 0; i < uvblksize;
+           ++i, vp += uv_stride, vdp += uvd_stride)
+        memcpy(vdp, vp, uvblksize);
     }
+  }
 }
 
-static int qualify_inter_mb(const MODE_INFO *mode_info_context, int *map)
-{
-    if (mode_info_context->mbmi.mb_skip_coeff)
-        map[0] = map[1] = map[2] = map[3] = 1;
-    else if (mode_info_context->mbmi.mode==SPLITMV)
-    {
-        static int ndx[4][4] =
-        {
-            {0, 1, 4, 5},
-            {2, 3, 6, 7},
-            {8, 9, 12, 13},
-            {10, 11, 14, 15}
-        };
-        int i, j;
-        for (i=0; i<4; ++i)
-        {
-            map[i] = 1;
-            for (j=0; j<4 && map[j]; ++j)
-                map[i] &= (mode_info_context->bmi[ndx[i][j]].mv.as_mv.row <= 2 &&
-                           mode_info_context->bmi[ndx[i][j]].mv.as_mv.col <= 2);
-        }
+static int qualify_inter_mb(const MODE_INFO *mode_info_context, int *map) {
+  if (mode_info_context->mbmi.mb_skip_coeff)
+    map[0] = map[1] = map[2] = map[3] = 1;
+  else if (mode_info_context->mbmi.mode == SPLITMV) {
+    static int ndx[4][4] = {
+      { 0, 1, 4, 5 }, { 2, 3, 6, 7 }, { 8, 9, 12, 13 }, { 10, 11, 14, 15 }
+    };
+    int i, j;
+    for (i = 0; i < 4; ++i) {
+      map[i] = 1;
+      for (j = 0; j < 4 && map[j]; ++j)
+        map[i] &= (mode_info_context->bmi[ndx[i][j]].mv.as_mv.row <= 2 &&
+                   mode_info_context->bmi[ndx[i][j]].mv.as_mv.col <= 2);
     }
-    else
-    {
-        map[0] = map[1] = map[2] = map[3] =
-            (mode_info_context->mbmi.mode > B_PRED &&
-             abs(mode_info_context->mbmi.mv.as_mv.row) <= 2 &&
-             abs(mode_info_context->mbmi.mv.as_mv.col) <= 2);
-    }
-    return (map[0]+map[1]+map[2]+map[3]);
+  } else {
+    map[0] = map[1] = map[2] = map[3] =
+        (mode_info_context->mbmi.mode > B_PRED &&
+         abs(mode_info_context->mbmi.mv.as_mv.row) <= 2 &&
+         abs(mode_info_context->mbmi.mv.as_mv.col) <= 2);
+  }
+  return (map[0] + map[1] + map[2] + map[3]);
 }
 
-void vp8_multiframe_quality_enhance
-(
-    VP8_COMMON *cm
-)
-{
-    YV12_BUFFER_CONFIG *show = cm->frame_to_show;
-    YV12_BUFFER_CONFIG *dest = &cm->post_proc_buffer;
+void vp8_multiframe_quality_enhance(VP8_COMMON *cm) {
+  YV12_BUFFER_CONFIG *show = cm->frame_to_show;
+  YV12_BUFFER_CONFIG *dest = &cm->post_proc_buffer;
 
-    FRAME_TYPE frame_type = cm->frame_type;
-    /* Point at base of Mb MODE_INFO list has motion vectors etc */
-    const MODE_INFO *mode_info_context = cm->show_frame_mi;
-    int mb_row;
-    int mb_col;
-    int totmap, map[4];
-    int qcurr = cm->base_qindex;
-    int qprev = cm->postproc_state.last_base_qindex;
+  FRAME_TYPE frame_type = cm->frame_type;
+  /* Point at base of Mb MODE_INFO list has motion vectors etc */
+  const MODE_INFO *mode_info_context = cm->show_frame_mi;
+  int mb_row;
+  int mb_col;
+  int totmap, map[4];
+  int qcurr = cm->base_qindex;
+  int qprev = cm->postproc_state.last_base_qindex;
 
-    unsigned char *y_ptr, *u_ptr, *v_ptr;
-    unsigned char *yd_ptr, *ud_ptr, *vd_ptr;
+  unsigned char *y_ptr, *u_ptr, *v_ptr;
+  unsigned char *yd_ptr, *ud_ptr, *vd_ptr;
 
-    /* Set up the buffer pointers */
-    y_ptr = show->y_buffer;
-    u_ptr = show->u_buffer;
-    v_ptr = show->v_buffer;
-    yd_ptr = dest->y_buffer;
-    ud_ptr = dest->u_buffer;
-    vd_ptr = dest->v_buffer;
+  /* Set up the buffer pointers */
+  y_ptr = show->y_buffer;
+  u_ptr = show->u_buffer;
+  v_ptr = show->v_buffer;
+  yd_ptr = dest->y_buffer;
+  ud_ptr = dest->u_buffer;
+  vd_ptr = dest->v_buffer;
 
-    /* postprocess each macro block */
-    for (mb_row = 0; mb_row < cm->mb_rows; mb_row++)
-    {
-        for (mb_col = 0; mb_col < cm->mb_cols; mb_col++)
-        {
-            /* if motion is high there will likely be no benefit */
-            if (frame_type == INTER_FRAME) totmap = qualify_inter_mb(mode_info_context, map);
-            else totmap = (frame_type == KEY_FRAME ? 4 : 0);
-            if (totmap)
-            {
-                if (totmap < 4)
-                {
-                    int i, j;
-                    for (i=0; i<2; ++i)
-                        for (j=0; j<2; ++j)
-                        {
-                            if (map[i*2+j])
-                            {
-                                multiframe_quality_enhance_block(8, qcurr, qprev,
-                                                                 y_ptr + 8*(i*show->y_stride+j),
-                                                                 u_ptr + 4*(i*show->uv_stride+j),
-                                                                 v_ptr + 4*(i*show->uv_stride+j),
-                                                                 show->y_stride,
-                                                                 show->uv_stride,
-                                                                 yd_ptr + 8*(i*dest->y_stride+j),
-                                                                 ud_ptr + 4*(i*dest->uv_stride+j),
-                                                                 vd_ptr + 4*(i*dest->uv_stride+j),
-                                                                 dest->y_stride,
-                                                                 dest->uv_stride);
-                            }
-                            else
-                            {
-                                /* copy a 8x8 block */
-                                int k;
-                                unsigned char *up = u_ptr + 4*(i*show->uv_stride+j);
-                                unsigned char *udp = ud_ptr + 4*(i*dest->uv_stride+j);
-                                unsigned char *vp = v_ptr + 4*(i*show->uv_stride+j);
-                                unsigned char *vdp = vd_ptr + 4*(i*dest->uv_stride+j);
-                                vp8_copy_mem8x8(y_ptr + 8*(i*show->y_stride+j), show->y_stride,
-                                                yd_ptr + 8*(i*dest->y_stride+j), dest->y_stride);
-                                for (k = 0; k < 4; ++k, up += show->uv_stride, udp += dest->uv_stride,
-                                                        vp += show->uv_stride, vdp += dest->uv_stride)
-                                {
-                                    memcpy(udp, up, 4);
-                                    memcpy(vdp, vp, 4);
-                                }
-                            }
-                        }
-                }
-                else /* totmap = 4 */
-                {
-                    multiframe_quality_enhance_block(16, qcurr, qprev, y_ptr,
-                                                     u_ptr, v_ptr,
-                                                     show->y_stride,
-                                                     show->uv_stride,
-                                                     yd_ptr, ud_ptr, vd_ptr,
-                                                     dest->y_stride,
-                                                     dest->uv_stride);
-                }
-            }
-            else
-            {
-                vp8_copy_mem16x16(y_ptr, show->y_stride, yd_ptr, dest->y_stride);
-                vp8_copy_mem8x8(u_ptr, show->uv_stride, ud_ptr, dest->uv_stride);
-                vp8_copy_mem8x8(v_ptr, show->uv_stride, vd_ptr, dest->uv_stride);
-            }
-            y_ptr += 16;
-            u_ptr += 8;
-            v_ptr += 8;
-            yd_ptr += 16;
-            ud_ptr += 8;
-            vd_ptr += 8;
-            mode_info_context++;     /* step to next MB */
-        }
+  /* postprocess each macro block */
+  for (mb_row = 0; mb_row < cm->mb_rows; mb_row++) {
+    for (mb_col = 0; mb_col < cm->mb_cols; mb_col++) {
+      /* if motion is high there will likely be no benefit */
+      if (frame_type == INTER_FRAME)
+        totmap = qualify_inter_mb(mode_info_context, map);
+      else
+        totmap = (frame_type == KEY_FRAME ? 4 : 0);
+      if (totmap) {
+        if (totmap < 4) {
+          int i, j;
+          for (i = 0; i < 2; ++i)
+            for (j = 0; j < 2; ++j) {
+              if (map[i * 2 + j]) {
+                multiframe_quality_enhance_block(
+                    8, qcurr, qprev, y_ptr + 8 * (i * show->y_stride + j),
+                    u_ptr + 4 * (i * show->uv_stride + j),
+                    v_ptr + 4 * (i * show->uv_stride + j), show->y_stride,
+                    show->uv_stride, yd_ptr + 8 * (i * dest->y_stride + j),
+                    ud_ptr + 4 * (i * dest->uv_stride + j),
+                    vd_ptr + 4 * (i * dest->uv_stride + j), dest->y_stride,
+                    dest->uv_stride);
+              } else {
+                /* copy a 8x8 block */
+                int k;
+                unsigned char *up = u_ptr + 4 * (i * show->uv_stride + j);
+                unsigned char *udp = ud_ptr + 4 * (i * dest->uv_stride + j);
+                unsigned char *vp = v_ptr + 4 * (i * show->uv_stride + j);
+                unsigned char *vdp = vd_ptr + 4 * (i * dest->uv_stride + j);
+                vp8_copy_mem8x8(
+                    y_ptr + 8 * (i * show->y_stride + j), show->y_stride,
+                    yd_ptr + 8 * (i * dest->y_stride + j), dest->y_stride);
+                for (k = 0; k < 4; ++k, up += show->uv_stride,
+                    udp += dest->uv_stride, vp += show->uv_stride,
+                    vdp += dest->uv_stride) {
+                  memcpy(udp, up, 4);
+                  memcpy(vdp, vp, 4);
+                }
+              }
+            }
+        } else /* totmap = 4 */
+        {
+          multiframe_quality_enhance_block(
+              16, qcurr, qprev, y_ptr, u_ptr, v_ptr, show->y_stride,
+              show->uv_stride, yd_ptr, ud_ptr, vd_ptr, dest->y_stride,
+              dest->uv_stride);
+        }
+      } else {
+        vp8_copy_mem16x16(y_ptr, show->y_stride, yd_ptr, dest->y_stride);
+        vp8_copy_mem8x8(u_ptr, show->uv_stride, ud_ptr, dest->uv_stride);
+        vp8_copy_mem8x8(v_ptr, show->uv_stride, vd_ptr, dest->uv_stride);
+      }
+      y_ptr += 16;
+      u_ptr += 8;
+      v_ptr += 8;
+      yd_ptr += 16;
+      ud_ptr += 8;
+      vd_ptr += 8;
+      mode_info_context++; /* step to next MB */
+    }
 
-        y_ptr += show->y_stride  * 16 - 16 * cm->mb_cols;
-        u_ptr += show->uv_stride *  8 - 8 * cm->mb_cols;
-        v_ptr += show->uv_stride *  8 - 8 * cm->mb_cols;
-        yd_ptr += dest->y_stride  * 16 - 16 * cm->mb_cols;
-        ud_ptr += dest->uv_stride *  8 - 8 * cm->mb_cols;
-        vd_ptr += dest->uv_stride *  8 - 8 * cm->mb_cols;
+    y_ptr += show->y_stride * 16 - 16 * cm->mb_cols;
+    u_ptr += show->uv_stride * 8 - 8 * cm->mb_cols;
+    v_ptr += show->uv_stride * 8 - 8 * cm->mb_cols;
+    yd_ptr += dest->y_stride * 16 - 16 * cm->mb_cols;
+    ud_ptr += dest->uv_stride * 8 - 8 * cm->mb_cols;
+    vd_ptr += dest->uv_stride * 8 - 8 * cm->mb_cols;
 
-        mode_info_context++;         /* Skip border mb */
-    }
+    mode_info_context++; /* Skip border mb */
+  }
 }
--- a/vp8/common/mips/dspr2/dequantize_dspr2.c
+++ b/vp8/common/mips/dspr2/dequantize_dspr2.c
@@ -8,26 +8,22 @@
  *  be found in the AUTHORS file in the root of the source tree.
  */
 
-
 #include "vpx_config.h"
 #include "vp8_rtcd.h"
 #include "vpx_mem/vpx_mem.h"
 
 #if HAVE_DSPR2
-void vp8_dequant_idct_add_dspr2(short *input, short *dq,
-                                unsigned char *dest, int stride)
-{
-    int i;
+void vp8_dequant_idct_add_dspr2(short *input, short *dq, unsigned char *dest,
+                                int stride) {
+  int i;
 
-    for (i = 0; i < 16; i++)
-    {
-        input[i] = dq[i] * input[i];
-    }
+  for (i = 0; i < 16; i++) {
+    input[i] = dq[i] * input[i];
+  }
 
-    vp8_short_idct4x4llm_dspr2(input, dest, stride, dest, stride);
+  vp8_short_idct4x4llm_dspr2(input, dest, stride, dest, stride);
 
-    memset(input, 0, 32);
-
+  memset(input, 0, 32);
 }
 
 #endif
--- a/vp8/common/mips/dspr2/filter_dspr2.c
+++ b/vp8/common/mips/dspr2/filter_dspr2.c
@@ -8,7 +8,6 @@
  *  be found in the AUTHORS file in the root of the source tree.
  */
 
-
 #include <stdlib.h>
 #include "vp8_rtcd.h"
 #include "vpx_ports/mem.h"
@@ -17,684 +16,483 @@
 #define CROP_WIDTH 256
 unsigned char ff_cropTbl[256 + 2 * CROP_WIDTH];
 
-static const unsigned short sub_pel_filterss[8][3] =
-{
-    {      0,      0,      0},
-    {      0, 0x0601, 0x7b0c},
-    { 0x0201, 0x0b08, 0x6c24},
-    {      0, 0x0906, 0x5d32},
-    { 0x0303, 0x1010, 0x4d4d},
-    {      0, 0x0609, 0x325d},
-    { 0x0102, 0x080b, 0x246c},
-    {      0, 0x0106, 0x0c7b},
+static const unsigned short sub_pel_filterss[8][3] = {
+  { 0, 0, 0 },
+  { 0, 0x0601, 0x7b0c },
+  { 0x0201, 0x0b08, 0x6c24 },
+  { 0, 0x0906, 0x5d32 },
+  { 0x0303, 0x1010, 0x4d4d },
+  { 0, 0x0609, 0x325d },
+  { 0x0102, 0x080b, 0x246c },
+  { 0, 0x0106, 0x0c7b },
 };
 
-
-static const int sub_pel_filters_int[8][3] =
-{
-    {          0,          0,          0},
-    { 0x0000fffa, 0x007b000c, 0xffff0000},
-    { 0x0002fff5, 0x006c0024, 0xfff80001},
-    { 0x0000fff7, 0x005d0032, 0xfffa0000},
-    { 0x0003fff0, 0x004d004d, 0xfff00003},
-    { 0x0000fffa, 0x0032005d, 0xfff70000},
-    { 0x0001fff8, 0x0024006c, 0xfff50002},
-    { 0x0000ffff, 0x000c007b, 0xfffa0000},
+static const int sub_pel_filters_int[8][3] = {
+  { 0, 0, 0 },
+  { 0x0000fffa, 0x007b000c, 0xffff0000 },
+  { 0x0002fff5, 0x006c0024, 0xfff80001 },
+  { 0x0000fff7, 0x005d0032, 0xfffa0000 },
+  { 0x0003fff0, 0x004d004d, 0xfff00003 },
+  { 0x0000fffa, 0x0032005d, 0xfff70000 },
+  { 0x0001fff8, 0x0024006c, 0xfff50002 },
+  { 0x0000ffff, 0x000c007b, 0xfffa0000 },
 };
 
-
-static const int sub_pel_filters_inv[8][3] =
-{
-    {          0,          0,          0},
-    { 0xfffa0000, 0x000c007b, 0x0000ffff},
-    { 0xfff50002, 0x0024006c, 0x0001fff8},
-    { 0xfff70000, 0x0032005d, 0x0000fffa},
-    { 0xfff00003, 0x004d004d, 0x0003fff0},
-    { 0xfffa0000, 0x005d0032, 0x0000fff7},
-    { 0xfff80001, 0x006c0024, 0x0002fff5},
-    { 0xffff0000, 0x007b000c, 0x0000fffa},
+static const int sub_pel_filters_inv[8][3] = {
+  { 0, 0, 0 },
+  { 0xfffa0000, 0x000c007b, 0x0000ffff },
+  { 0xfff50002, 0x0024006c, 0x0001fff8 },
+  { 0xfff70000, 0x0032005d, 0x0000fffa },
+  { 0xfff00003, 0x004d004d, 0x0003fff0 },
+  { 0xfffa0000, 0x005d0032, 0x0000fff7 },
+  { 0xfff80001, 0x006c0024, 0x0002fff5 },
+  { 0xffff0000, 0x007b000c, 0x0000fffa },
 };
 
-
-static const int sub_pel_filters_int_tap_4[8][2] =
-{
-    {          0,          0},
-    { 0xfffa007b, 0x000cffff},
-    {          0,          0},
-    { 0xfff7005d, 0x0032fffa},
-    {          0,          0},
-    { 0xfffa0032, 0x005dfff7},
-    {          0,          0},
-    { 0xffff000c, 0x007bfffa},
+/* clang-format off */
+static const int sub_pel_filters_int_tap_4[8][2] = {
+  {          0,          0},
+  { 0xfffa007b, 0x000cffff},
+  {          0,          0},
+  { 0xfff7005d, 0x0032fffa},
+  {          0,          0},
+  { 0xfffa0032, 0x005dfff7},
+  {          0,          0},
+  { 0xffff000c, 0x007bfffa},
 };
 
 
-static const int sub_pel_filters_inv_tap_4[8][2] =
-{
-    {          0,          0},
-    { 0x007bfffa, 0xffff000c},
-    {          0,          0},
-    { 0x005dfff7, 0xfffa0032},
-    {          0,          0},
-    { 0x0032fffa, 0xfff7005d},
-    {          0,          0},
-    { 0x000cffff, 0xfffa007b},
+static const int sub_pel_filters_inv_tap_4[8][2] = {
+  {          0,          0},
+  { 0x007bfffa, 0xffff000c},
+  {          0,          0},
+  { 0x005dfff7, 0xfffa0032},
+  {          0,          0},
+  { 0x0032fffa, 0xfff7005d},
+  {          0,          0},
+  { 0x000cffff, 0xfffa007b},
 };
+/* clang-format on */
 
-inline void prefetch_load(unsigned char *src)
-{
-    __asm__ __volatile__ (
-        "pref   0,  0(%[src])   \n\t"
-        :
-        : [src] "r" (src)
-    );
+inline void prefetch_load(unsigned char *src) {
+  __asm__ __volatile__("pref   0,  0(%[src])   \n\t" : : [src] "r"(src));
 }
 
-
-inline void prefetch_store(unsigned char *dst)
-{
-    __asm__ __volatile__ (
-        "pref   1,  0(%[dst])   \n\t"
-        :
-        : [dst] "r" (dst)
-    );
+inline void prefetch_store(unsigned char *dst) {
+  __asm__ __volatile__("pref   1,  0(%[dst])   \n\t" : : [dst] "r"(dst));
 }
 
-void dsputil_static_init(void)
-{
-    int i;
+void dsputil_static_init(void) {
+  int i;
 
-    for (i = 0; i < 256; i++) ff_cropTbl[i + CROP_WIDTH] = i;
+  for (i = 0; i < 256; i++) ff_cropTbl[i + CROP_WIDTH] = i;
 
-    for (i = 0; i < CROP_WIDTH; i++)
-    {
-        ff_cropTbl[i] = 0;
-        ff_cropTbl[i + CROP_WIDTH + 256] = 255;
-    }
+  for (i = 0; i < CROP_WIDTH; i++) {
+    ff_cropTbl[i] = 0;
+    ff_cropTbl[i + CROP_WIDTH + 256] = 255;
+  }
 }
 
-void vp8_filter_block2d_first_pass_4
-(
-    unsigned char *RESTRICT src_ptr,
-    unsigned char *RESTRICT dst_ptr,
-    unsigned int src_pixels_per_line,
-    unsigned int output_height,
-    int xoffset,
-    int pitch
-)
-{
-    unsigned int i;
-    int Temp1, Temp2, Temp3, Temp4;
+void vp8_filter_block2d_first_pass_4(unsigned char *RESTRICT src_ptr,
+                                     unsigned char *RESTRICT dst_ptr,
+                                     unsigned int src_pixels_per_line,
+                                     unsigned int output_height, int xoffset,
+                                     int pitch) {
+  unsigned int i;
+  int Temp1, Temp2, Temp3, Temp4;
 
-    unsigned int vector4a = 64;
-    int vector1b, vector2b, vector3b;
-    unsigned int tp1, tp2, tn1, tn2;
-    unsigned int p1, p2, p3;
-    unsigned int n1, n2, n3;
-    unsigned char *cm = ff_cropTbl + CROP_WIDTH;
+  unsigned int vector4a = 64;
+  int vector1b, vector2b, vector3b;
+  unsigned int tp1, tp2, tn1, tn2;
+  unsigned int p1, p2, p3;
+  unsigned int n1, n2, n3;
+  unsigned char *cm = ff_cropTbl + CROP_WIDTH;
 
-    vector3b = sub_pel_filters_inv[xoffset][2];
+  vector3b = sub_pel_filters_inv[xoffset][2];
 
-    /* if (xoffset == 0) we don't need any filtering */
-    if (vector3b == 0)
-    {
-        for (i = 0; i < output_height; i++)
-        {
-            /* prefetch src_ptr data to cache memory */
-            prefetch_load(src_ptr + src_pixels_per_line);
-            dst_ptr[0] = src_ptr[0];
-            dst_ptr[1] = src_ptr[1];
-            dst_ptr[2] = src_ptr[2];
-            dst_ptr[3] = src_ptr[3];
+  /* if (xoffset == 0) we don't need any filtering */
+  if (vector3b == 0) {
+    for (i = 0; i < output_height; i++) {
+      /* prefetch src_ptr data to cache memory */
+      prefetch_load(src_ptr + src_pixels_per_line);
+      dst_ptr[0] = src_ptr[0];
+      dst_ptr[1] = src_ptr[1];
+      dst_ptr[2] = src_ptr[2];
+      dst_ptr[3] = src_ptr[3];
 
-            /* next row... */
-            src_ptr += src_pixels_per_line;
-            dst_ptr += 4;
-        }
+      /* next row... */
+      src_ptr += src_pixels_per_line;
+      dst_ptr += 4;
     }
-    else
-    {
-        if (vector3b > 65536)
-        {
-            /* 6 tap filter */
+  } else {
+    if (vector3b > 65536) {
+      /* 6 tap filter */
 
-            vector1b = sub_pel_filters_inv[xoffset][0];
-            vector2b = sub_pel_filters_inv[xoffset][1];
+      vector1b = sub_pel_filters_inv[xoffset][0];
+      vector2b = sub_pel_filters_inv[xoffset][1];
 
-            /* prefetch src_ptr data to cache memory */
-            prefetch_load(src_ptr + src_pixels_per_line);
+      /* prefetch src_ptr data to cache memory */
+      prefetch_load(src_ptr + src_pixels_per_line);
 
-            for (i = output_height; i--;)
-            {
-                /* apply filter with vectors pairs */
-                __asm__ __volatile__ (
-                    "ulw              %[tp1],      -2(%[src_ptr])                 \n\t"
-                    "ulw              %[tp2],      2(%[src_ptr])                  \n\t"
+      for (i = output_height; i--;) {
+        /* apply filter with vectors pairs */
+        __asm__ __volatile__(
+            "ulw              %[tp1],      -2(%[src_ptr])                 \n\t"
+            "ulw              %[tp2],      2(%[src_ptr])                  \n\t"
 
-                    /* even 1. pixel */
-                    "mtlo             %[vector4a], $ac3                           \n\t"
-                    "preceu.ph.qbr    %[p1],       %[tp1]                         \n\t"
-                    "preceu.ph.qbl    %[p2],       %[tp1]                         \n\t"
-                    "preceu.ph.qbr    %[p3],       %[tp2]                         \n\t"
-                    "dpa.w.ph         $ac3,        %[p1],          %[vector1b]    \n\t"
-                    "dpa.w.ph         $ac3,        %[p2],          %[vector2b]    \n\t"
-                    "dpa.w.ph         $ac3,        %[p3],          %[vector3b]    \n\t"
+            /* even 1. pixel */
+            "mtlo             %[vector4a], $ac3                           \n\t"
+            "preceu.ph.qbr    %[p1],       %[tp1]                         \n\t"
+            "preceu.ph.qbl    %[p2],       %[tp1]                         \n\t"
+            "preceu.ph.qbr    %[p3],       %[tp2]                         \n\t"
+            "dpa.w.ph         $ac3,        %[p1],          %[vector1b]    \n\t"
+            "dpa.w.ph         $ac3,        %[p2],          %[vector2b]    \n\t"
+            "dpa.w.ph         $ac3,        %[p3],          %[vector3b]    \n\t"
 
-                    /* even 2. pixel */
-                    "mtlo             %[vector4a], $ac2                           \n\t"
-                    "preceu.ph.qbl    %[p1],       %[tp2]                         \n\t"
-                    "balign           %[tp2],      %[tp1],         3              \n\t"
-                    "extp             %[Temp1],    $ac3,           9              \n\t"
-                    "dpa.w.ph         $ac2,        %[p2],          %[vector1b]    \n\t"
-                    "dpa.w.ph         $ac2,        %[p3],          %[vector2b]    \n\t"
-                    "dpa.w.ph         $ac2,        %[p1],          %[vector3b]    \n\t"
+            /* even 2. pixel */
+            "mtlo             %[vector4a], $ac2                           \n\t"
+            "preceu.ph.qbl    %[p1],       %[tp2]                         \n\t"
+            "balign           %[tp2],      %[tp1],         3              \n\t"
+            "extp             %[Temp1],    $ac3,           9              \n\t"
+            "dpa.w.ph         $ac2,        %[p2],          %[vector1b]    \n\t"
+            "dpa.w.ph         $ac2,        %[p3],          %[vector2b]    \n\t"
+            "dpa.w.ph         $ac2,        %[p1],          %[vector3b]    \n\t"
 
-                    /* odd 1. pixel */
-                    "ulw              %[tn2],      3(%[src_ptr])                  \n\t"
-                    "mtlo             %[vector4a], $ac3                           \n\t"
-                    "preceu.ph.qbr    %[n1],       %[tp2]                         \n\t"
-                    "preceu.ph.qbl    %[n2],       %[tp2]                         \n\t"
-                    "preceu.ph.qbr    %[n3],       %[tn2]                         \n\t"
-                    "extp             %[Temp3],    $ac2,           9              \n\t"
-                    "dpa.w.ph         $ac3,        %[n1],          %[vector1b]    \n\t"
-                    "dpa.w.ph         $ac3,        %[n2],          %[vector2b]    \n\t"
-                    "dpa.w.ph         $ac3,        %[n3],          %[vector3b]    \n\t"
+            /* odd 1. pixel */
+            "ulw              %[tn2],      3(%[src_ptr])                  \n\t"
+            "mtlo             %[vector4a], $ac3                           \n\t"
+            "preceu.ph.qbr    %[n1],       %[tp2]                         \n\t"
+            "preceu.ph.qbl    %[n2],       %[tp2]                         \n\t"
+            "preceu.ph.qbr    %[n3],       %[tn2]                         \n\t"
+            "extp             %[Temp3],    $ac2,           9              \n\t"
+            "dpa.w.ph         $ac3,        %[n1],          %[vector1b]    \n\t"
+            "dpa.w.ph         $ac3,        %[n2],          %[vector2b]    \n\t"
+            "dpa.w.ph         $ac3,        %[n3],          %[vector3b]    \n\t"
 
-                    /* even 2. pixel */
-                    "mtlo             %[vector4a], $ac2                           \n\t"
-                    "preceu.ph.qbl    %[n1],       %[tn2]                         \n\t"
-                    "extp             %[Temp2],    $ac3,           9              \n\t"
-                    "dpa.w.ph         $ac2,        %[n2],          %[vector1b]    \n\t"
-                    "dpa.w.ph         $ac2,        %[n3],          %[vector2b]    \n\t"
-                    "dpa.w.ph         $ac2,        %[n1],          %[vector3b]    \n\t"
-                    "extp             %[Temp4],    $ac2,           9              \n\t"
+            /* even 2. pixel */
+            "mtlo             %[vector4a], $ac2                           \n\t"
+            "preceu.ph.qbl    %[n1],       %[tn2]                         \n\t"
+            "extp             %[Temp2],    $ac3,           9              \n\t"
+            "dpa.w.ph         $ac2,        %[n2],          %[vector1b]    \n\t"
+            "dpa.w.ph         $ac2,        %[n3],          %[vector2b]    \n\t"
+            "dpa.w.ph         $ac2,        %[n1],          %[vector3b]    \n\t"
+            "extp             %[Temp4],    $ac2,           9              \n\t"
 
-                    /* clamp */
-                    "lbux             %[tp1],      %[Temp1](%[cm])                \n\t"
-                    "lbux             %[tn1],      %[Temp2](%[cm])                \n\t"
-                    "lbux             %[tp2],      %[Temp3](%[cm])                \n\t"
-                    "lbux             %[n2],       %[Temp4](%[cm])                \n\t"
+            /* clamp */
+            "lbux             %[tp1],      %[Temp1](%[cm])                \n\t"
+            "lbux             %[tn1],      %[Temp2](%[cm])                \n\t"
+            "lbux             %[tp2],      %[Temp3](%[cm])                \n\t"
+            "lbux             %[n2],       %[Temp4](%[cm])                \n\t"
 
-                    /* store bytes */
-                    "sb               %[tp1],      0(%[dst_ptr])                  \n\t"
-                    "sb               %[tn1],      1(%[dst_ptr])                  \n\t"
-                    "sb               %[tp2],      2(%[dst_ptr])                  \n\t"
-                    "sb               %[n2],       3(%[dst_ptr])                  \n\t"
+            /* store bytes */
+            "sb               %[tp1],      0(%[dst_ptr])                  \n\t"
+            "sb               %[tn1],      1(%[dst_ptr])                  \n\t"
+            "sb               %[tp2],      2(%[dst_ptr])                  \n\t"
+            "sb               %[n2],       3(%[dst_ptr])                  \n\t"
 
-                    : [tp1] "=&r" (tp1), [tp2] "=&r" (tp2), [tn1] "=&r" (tn1),
-                      [tn2] "=&r" (tn2), [p1] "=&r" (p1), [p2] "=&r" (p2),
-                      [p3] "=&r" (p3), [n1] "=&r" (n1), [n2] "=&r" (n2),
-                      [n3] "=&r" (n3), [Temp1] "=&r" (Temp1), [Temp2] "=&r" (Temp2),
-                      [Temp3] "=&r" (Temp3), [Temp4] "=&r" (Temp4)
-                    : [vector1b] "r" (vector1b), [vector2b] "r" (vector2b),
-                      [vector4a] "r" (vector4a), [cm] "r" (cm), [dst_ptr] "r" (dst_ptr),
-                      [vector3b] "r" (vector3b), [src_ptr] "r" (src_ptr)
-                );
+            : [tp1] "=&r"(tp1), [tp2] "=&r"(tp2), [tn1] "=&r"(tn1),
+              [tn2] "=&r"(tn2), [p1] "=&r"(p1), [p2] "=&r"(p2), [p3] "=&r"(p3),
+              [n1] "=&r"(n1), [n2] "=&r"(n2), [n3] "=&r"(n3),
+              [Temp1] "=&r"(Temp1), [Temp2] "=&r"(Temp2), [Temp3] "=&r"(Temp3),
+              [Temp4] "=&r"(Temp4)
+            : [vector1b] "r"(vector1b), [vector2b] "r"(vector2b),
+              [vector4a] "r"(vector4a), [cm] "r"(cm), [dst_ptr] "r"(dst_ptr),
+              [vector3b] "r"(vector3b), [src_ptr] "r"(src_ptr));
 
-                /* Next row... */
-                src_ptr += src_pixels_per_line;
-                dst_ptr += pitch;
-            }
-        }
-        else
-        {
-            /* 4 tap filter */
+        /* Next row... */
+        src_ptr += src_pixels_per_line;
+        dst_ptr += pitch;
+      }
+    } else {
+      /* 4 tap filter */
 
-            vector1b = sub_pel_filters_inv_tap_4[xoffset][0];
-            vector2b = sub_pel_filters_inv_tap_4[xoffset][1];
+      vector1b = sub_pel_filters_inv_tap_4[xoffset][0];
+      vector2b = sub_pel_filters_inv_tap_4[xoffset][1];
 
-            for (i = output_height; i--;)
-            {
-                /* apply filter with vectors pairs */
-                __asm__ __volatile__ (
-                    "ulw              %[tp1],      -1(%[src_ptr])                 \n\t"
-                    "ulw              %[tp2],      3(%[src_ptr])                  \n\t"
+      for (i = output_height; i--;) {
+        /* apply filter with vectors pairs */
+        __asm__ __volatile__(
+            "ulw              %[tp1],      -1(%[src_ptr])                 \n\t"
+            "ulw              %[tp2],      3(%[src_ptr])                  \n\t"
 
-                    /* even 1. pixel */
-                    "mtlo             %[vector4a], $ac3                           \n\t"
-                    "preceu.ph.qbr    %[p1],       %[tp1]                         \n\t"
-                    "preceu.ph.qbl    %[p2],       %[tp1]                         \n\t"
-                    "preceu.ph.qbr    %[p3],       %[tp2]                         \n\t"
-                    "dpa.w.ph         $ac3,        %[p1],          %[vector1b]    \n\t"
-                    "dpa.w.ph         $ac3,        %[p2],          %[vector2b]    \n\t"
+            /* even 1. pixel */
+            "mtlo             %[vector4a], $ac3                           \n\t"
+            "preceu.ph.qbr    %[p1],       %[tp1]                         \n\t"
+            "preceu.ph.qbl    %[p2],       %[tp1]                         \n\t"
+            "preceu.ph.qbr    %[p3],       %[tp2]                         \n\t"
+            "dpa.w.ph         $ac3,        %[p1],          %[vector1b]    \n\t"
+            "dpa.w.ph         $ac3,        %[p2],          %[vector2b]    \n\t"
 
-                    /* even 2. pixel */
-                    "mtlo             %[vector4a], $ac2                           \n\t"
-                    "dpa.w.ph         $ac2,        %[p2],          %[vector1b]    \n\t"
-                    "dpa.w.ph         $ac2,        %[p3],          %[vector2b]    \n\t"
-                    "extp             %[Temp1],    $ac3,           9              \n\t"
+            /* even 2. pixel */
+            "mtlo             %[vector4a], $ac2                           \n\t"
+            "dpa.w.ph         $ac2,        %[p2],          %[vector1b]    \n\t"
+            "dpa.w.ph         $ac2,        %[p3],          %[vector2b]    \n\t"
+            "extp             %[Temp1],    $ac3,           9              \n\t"
 
-                    /* odd 1. pixel */
-                    "srl              %[tn1],      %[tp2],         8              \n\t"
-                    "balign           %[tp2],      %[tp1],         3              \n\t"
-                    "mtlo             %[vector4a], $ac3                           \n\t"
-                    "preceu.ph.qbr    %[n1],       %[tp2]                         \n\t"
-                    "preceu.ph.qbl    %[n2],       %[tp2]                         \n\t"
-                    "preceu.ph.qbr    %[n3],       %[tn1]                         \n\t"
-                    "extp             %[Temp3],    $ac2,           9              \n\t"
-                    "dpa.w.ph         $ac3,        %[n1],          %[vector1b]    \n\t"
-                    "dpa.w.ph         $ac3,        %[n2],          %[vector2b]    \n\t"
+            /* odd 1. pixel */
+            "srl              %[tn1],      %[tp2],         8              \n\t"
+            "balign           %[tp2],      %[tp1],         3              \n\t"
+            "mtlo             %[vector4a], $ac3                           \n\t"
+            "preceu.ph.qbr    %[n1],       %[tp2]                         \n\t"
+            "preceu.ph.qbl    %[n2],       %[tp2]                         \n\t"
+            "preceu.ph.qbr    %[n3],       %[tn1]                         \n\t"
+            "extp             %[Temp3],    $ac2,           9              \n\t"
+            "dpa.w.ph         $ac3,        %[n1],          %[vector1b]    \n\t"
+            "dpa.w.ph         $ac3,        %[n2],          %[vector2b]    \n\t"
 
-                    /* odd 2. pixel */
-                    "mtlo             %[vector4a], $ac2                           \n\t"
-                    "extp             %[Temp2],    $ac3,           9              \n\t"
-                    "dpa.w.ph         $ac2,        %[n2],          %[vector1b]    \n\t"
-                    "dpa.w.ph         $ac2,        %[n3],          %[vector2b]    \n\t"
-                    "extp             %[Temp4],    $ac2,           9              \n\t"
+            /* odd 2. pixel */
+            "mtlo             %[vector4a], $ac2                           \n\t"
+            "extp             %[Temp2],    $ac3,           9              \n\t"
+            "dpa.w.ph         $ac2,        %[n2],          %[vector1b]    \n\t"
+            "dpa.w.ph         $ac2,        %[n3],          %[vector2b]    \n\t"
+            "extp             %[Temp4],    $ac2,           9              \n\t"
 
-                    /* clamp and store results */
-                    "lbux             %[tp1],      %[Temp1](%[cm])                \n\t"
-                    "lbux             %[tn1],      %[Temp2](%[cm])                \n\t"
-                    "lbux             %[tp2],      %[Temp3](%[cm])                \n\t"
-                    "sb               %[tp1],      0(%[dst_ptr])                  \n\t"
-                    "sb               %[tn1],      1(%[dst_ptr])                  \n\t"
-                    "lbux             %[n2],       %[Temp4](%[cm])                \n\t"
-                    "sb               %[tp2],      2(%[dst_ptr])                  \n\t"
-                    "sb               %[n2],       3(%[dst_ptr])                  \n\t"
+            /* clamp and store results */
+            "lbux             %[tp1],      %[Temp1](%[cm])                \n\t"
+            "lbux             %[tn1],      %[Temp2](%[cm])                \n\t"
+            "lbux             %[tp2],      %[Temp3](%[cm])                \n\t"
+            "sb               %[tp1],      0(%[dst_ptr])                  \n\t"
+            "sb               %[tn1],      1(%[dst_ptr])                  \n\t"
+            "lbux             %[n2],       %[Temp4](%[cm])                \n\t"
+            "sb               %[tp2],      2(%[dst_ptr])                  \n\t"
+            "sb               %[n2],       3(%[dst_ptr])                  \n\t"
 
-                    : [tp1] "=&r" (tp1), [tp2] "=&r" (tp2), [tn1] "=&r" (tn1),
-                      [p1] "=&r" (p1), [p2] "=&r" (p2), [p3] "=&r" (p3),
-                      [n1] "=&r" (n1), [n2] "=&r" (n2), [n3] "=&r" (n3),
-                      [Temp1] "=&r" (Temp1), [Temp2] "=&r" (Temp2),
-                      [Temp3] "=&r" (Temp3), [Temp4] "=&r" (Temp4)
-                    : [vector1b] "r" (vector1b), [vector2b] "r" (vector2b),
-                      [vector4a] "r" (vector4a), [cm] "r" (cm), [dst_ptr] "r" (dst_ptr),
-                      [src_ptr] "r" (src_ptr)
-                );
-                /*  Next row... */
-                src_ptr += src_pixels_per_line;
-                dst_ptr += pitch;
-            }
-        }
+            : [tp1] "=&r"(tp1), [tp2] "=&r"(tp2), [tn1] "=&r"(tn1),
+              [p1] "=&r"(p1), [p2] "=&r"(p2), [p3] "=&r"(p3), [n1] "=&r"(n1),
+              [n2] "=&r"(n2), [n3] "=&r"(n3), [Temp1] "=&r"(Temp1),
+              [Temp2] "=&r"(Temp2), [Temp3] "=&r"(Temp3), [Temp4] "=&r"(Temp4)
+            : [vector1b] "r"(vector1b), [vector2b] "r"(vector2b),
+              [vector4a] "r"(vector4a), [cm] "r"(cm), [dst_ptr] "r"(dst_ptr),
+              [src_ptr] "r"(src_ptr));
+        /*  Next row... */
+        src_ptr += src_pixels_per_line;
+        dst_ptr += pitch;
+      }
     }
+  }
 }
 
-void vp8_filter_block2d_first_pass_8_all
-(
-    unsigned char *RESTRICT src_ptr,
-    unsigned char *RESTRICT dst_ptr,
-    unsigned int src_pixels_per_line,
-    unsigned int output_height,
-    int xoffset,
-    int pitch
-)
-{
-    unsigned int i;
-    int Temp1, Temp2, Temp3, Temp4;
+void vp8_filter_block2d_first_pass_8_all(unsigned char *RESTRICT src_ptr,
+                                         unsigned char *RESTRICT dst_ptr,
+                                         unsigned int src_pixels_per_line,
+                                         unsigned int output_height,
+                                         int xoffset, int pitch) {
+  unsigned int i;
+  int Temp1, Temp2, Temp3, Temp4;
 
-    unsigned int vector4a = 64;
-    unsigned int vector1b, vector2b, vector3b;
-    unsigned int tp1, tp2, tn1, tn2;
-    unsigned int p1, p2, p3, p4;
-    unsigned int n1, n2, n3, n4;
+  unsigned int vector4a = 64;
+  unsigned int vector1b, vector2b, vector3b;
+  unsigned int tp1, tp2, tn1, tn2;
+  unsigned int p1, p2, p3, p4;
+  unsigned int n1, n2, n3, n4;
 
-    unsigned char *cm = ff_cropTbl + CROP_WIDTH;
+  unsigned char *cm = ff_cropTbl + CROP_WIDTH;
 
-    /* if (xoffset == 0) we don't need any filtering */
-    if (xoffset == 0)
-    {
-        for (i = 0; i < output_height; i++)
-        {
-            /* prefetch src_ptr data to cache memory */
-            prefetch_load(src_ptr + src_pixels_per_line);
+  /* if (xoffset == 0) we don't need any filtering */
+  if (xoffset == 0) {
+    for (i = 0; i < output_height; i++) {
+      /* prefetch src_ptr data to cache memory */
+      prefetch_load(src_ptr + src_pixels_per_line);
 
-            dst_ptr[0] = src_ptr[0];
-            dst_ptr[1] = src_ptr[1];
-            dst_ptr[2] = src_ptr[2];
-            dst_ptr[3] = src_ptr[3];
-            dst_ptr[4] = src_ptr[4];
-            dst_ptr[5] = src_ptr[5];
-            dst_ptr[6] = src_ptr[6];
-            dst_ptr[7] = src_ptr[7];
+      dst_ptr[0] = src_ptr[0];
+      dst_ptr[1] = src_ptr[1];
+      dst_ptr[2] = src_ptr[2];
+      dst_ptr[3] = src_ptr[3];
+      dst_ptr[4] = src_ptr[4];
+      dst_ptr[5] = src_ptr[5];
+      dst_ptr[6] = src_ptr[6];
+      dst_ptr[7] = src_ptr[7];
 
-            /* next row... */
-            src_ptr += src_pixels_per_line;
-            dst_ptr += 8;
-        }
+      /* next row... */
+      src_ptr += src_pixels_per_line;
+      dst_ptr += 8;
     }
-    else
-    {
-        vector3b = sub_pel_filters_inv[xoffset][2];
+  } else {
+    vector3b = sub_pel_filters_inv[xoffset][2];
 
-        if (vector3b > 65536)
-        {
-            /* 6 tap filter */
+    if (vector3b > 65536) {
+      /* 6 tap filter */
 
-            vector1b = sub_pel_filters_inv[xoffset][0];
-            vector2b = sub_pel_filters_inv[xoffset][1];
+      vector1b = sub_pel_filters_inv[xoffset][0];
+      vector2b = sub_pel_filters_inv[xoffset][1];
 
-            for (i = output_height; i--;)
-            {
-                /* prefetch src_ptr data to cache memory */
-                prefetch_load(src_ptr + src_pixels_per_line);
+      for (i = output_height; i--;) {
+        /* prefetch src_ptr data to cache memory */
+        prefetch_load(src_ptr + src_pixels_per_line);
 
-                /* apply filter with vectors pairs */
-                __asm__ __volatile__ (
-                    "ulw              %[tp1],      -2(%[src_ptr])                 \n\t"
-                    "ulw              %[tp2],      2(%[src_ptr])                  \n\t"
+        /* apply filter with vectors pairs */
+        __asm__ __volatile__(
+            "ulw              %[tp1],      -2(%[src_ptr])                 \n\t"
+            "ulw              %[tp2],      2(%[src_ptr])                  \n\t"
 
-                    /* even 1. pixel */
-                    "mtlo             %[vector4a], $ac3                           \n\t"
-                    "preceu.ph.qbr    %[p1],       %[tp1]                         \n\t"
-                    "preceu.ph.qbl    %[p2],       %[tp1]                         \n\t"
-                    "preceu.ph.qbr    %[p3],       %[tp2]                         \n\t"
-                    "dpa.w.ph         $ac3,        %[p1],          %[vector1b]    \n\t"
-                    "dpa.w.ph         $ac3,        %[p2],          %[vector2b]    \n\t"
-                    "dpa.w.ph         $ac3,        %[p3],          %[vector3b]    \n\t"
+            /* even 1. pixel */
+            "mtlo             %[vector4a], $ac3                           \n\t"
+            "preceu.ph.qbr    %[p1],       %[tp1]                         \n\t"
+            "preceu.ph.qbl    %[p2],       %[tp1]                         \n\t"
+            "preceu.ph.qbr    %[p3],       %[tp2]                         \n\t"
+            "dpa.w.ph         $ac3,        %[p1],          %[vector1b]    \n\t"
+            "dpa.w.ph         $ac3,        %[p2],          %[vector2b]    \n\t"
+            "dpa.w.ph         $ac3,        %[p3],          %[vector3b]    \n\t"
 
-                    /* even 2. pixel */
-                    "mtlo             %[vector4a], $ac2                           \n\t"
-                    "preceu.ph.qbl    %[p1],       %[tp2]                         \n\t"
-                    "dpa.w.ph         $ac2,        %[p2],          %[vector1b]    \n\t"
-                    "dpa.w.ph         $ac2,        %[p3],          %[vector2b]    \n\t"
-                    "dpa.w.ph         $ac2,        %[p1],          %[vector3b]    \n\t"
+            /* even 2. pixel */
+            "mtlo             %[vector4a], $ac2                           \n\t"
+            "preceu.ph.qbl    %[p1],       %[tp2]                         \n\t"
+            "dpa.w.ph         $ac2,        %[p2],          %[vector1b]    \n\t"
+            "dpa.w.ph         $ac2,        %[p3],          %[vector2b]    \n\t"
+            "dpa.w.ph         $ac2,        %[p1],          %[vector3b]    \n\t"
 
-                    "balign           %[tp2],      %[tp1],         3              \n\t"
-                    "extp             %[Temp1],    $ac3,           9              \n\t"
-                    "ulw              %[tn2],      3(%[src_ptr])                  \n\t"
+            "balign           %[tp2],      %[tp1],         3              \n\t"
+            "extp             %[Temp1],    $ac3,           9              \n\t"
+            "ulw              %[tn2],      3(%[src_ptr])                  \n\t"
 
-                    /* odd 1. pixel */
-                    "mtlo             %[vector4a], $ac3                           \n\t"
-                    "preceu.ph.qbr    %[n1],       %[tp2]                         \n\t"
-                    "preceu.ph.qbl    %[n2],       %[tp2]                         \n\t"
-                    "preceu.ph.qbr    %[n3],       %[tn2]                         \n\t"
-                    "extp             %[Temp3],    $ac2,           9              \n\t"
-                    "dpa.w.ph         $ac3,        %[n1],          %[vector1b]    \n\t"
-                    "dpa.w.ph         $ac3,        %[n2],          %[vector2b]    \n\t"
-                    "dpa.w.ph         $ac3,        %[n3],          %[vector3b]    \n\t"
+            /* odd 1. pixel */
+            "mtlo             %[vector4a], $ac3                           \n\t"
+            "preceu.ph.qbr    %[n1],       %[tp2]                         \n\t"
+            "preceu.ph.qbl    %[n2],       %[tp2]                         \n\t"
+            "preceu.ph.qbr    %[n3],       %[tn2]                         \n\t"
+            "extp             %[Temp3],    $ac2,           9              \n\t"
+            "dpa.w.ph         $ac3,        %[n1],          %[vector1b]    \n\t"
+            "dpa.w.ph         $ac3,        %[n2],          %[vector2b]    \n\t"
+            "dpa.w.ph         $ac3,        %[n3],          %[vector3b]    \n\t"
 
-                    /* odd 2. pixel */
-                    "mtlo             %[vector4a], $ac2                           \n\t"
-                    "preceu.ph.qbl    %[n1],       %[tn2]                         \n\t"
-                    "dpa.w.ph         $ac2,        %[n2],          %[vector1b]    \n\t"
-                    "dpa.w.ph         $ac2,        %[n3],          %[vector2b]    \n\t"
-                    "dpa.w.ph         $ac2,        %[n1],          %[vector3b]    \n\t"
-                    "ulw              %[tp1],      6(%[src_ptr])                  \n\t"
-                    "extp             %[Temp2],    $ac3,           9              \n\t"
-                    "mtlo             %[vector4a], $ac3                           \n\t"
-                    "preceu.ph.qbr    %[p2],       %[tp1]                         \n\t"
-                    "extp             %[Temp4],    $ac2,           9              \n\t"
+            /* odd 2. pixel */
+            "mtlo             %[vector4a], $ac2                           \n\t"
+            "preceu.ph.qbl    %[n1],       %[tn2]                         \n\t"
+            "dpa.w.ph         $ac2,        %[n2],          %[vector1b]    \n\t"
+            "dpa.w.ph         $ac2,        %[n3],          %[vector2b]    \n\t"
+            "dpa.w.ph         $ac2,        %[n1],          %[vector3b]    \n\t"
+            "ulw              %[tp1],      6(%[src_ptr])                  \n\t"
+            "extp             %[Temp2],    $ac3,           9              \n\t"
+            "mtlo             %[vector4a], $ac3                           \n\t"
+            "preceu.ph.qbr    %[p2],       %[tp1]                         \n\t"
+            "extp             %[Temp4],    $ac2,           9              \n\t"
 
-                    : [tp1] "=&r" (tp1), [tp2] "=&r" (tp2), [tn2] "=&r" (tn2),
-                      [p1] "=&r" (p1), [p2] "=&r" (p2), [p3] "=&r" (p3),
-                      [n1] "=&r" (n1), [n2] "=&r" (n2), [n3] "=&r" (n3),
-                      [Temp1] "=&r" (Temp1), [Temp2] "=&r" (Temp2),
-                      [Temp3] "=&r" (Temp3), [Temp4] "=r" (Temp4)
-                    : [vector1b] "r" (vector1b), [vector2b] "r" (vector2b),
-                      [vector4a] "r" (vector4a), [vector3b] "r" (vector3b),
-                      [src_ptr] "r" (src_ptr)
-                );
+            : [tp1] "=&r"(tp1), [tp2] "=&r"(tp2), [tn2] "=&r"(tn2),
+              [p1] "=&r"(p1), [p2] "=&r"(p2), [p3] "=&r"(p3), [n1] "=&r"(n1),
+              [n2] "=&r"(n2), [n3] "=&r"(n3), [Temp1] "=&r"(Temp1),
+              [Temp2] "=&r"(Temp2), [Temp3] "=&r"(Temp3), [Temp4] "=r"(Temp4)
+            : [vector1b] "r"(vector1b), [vector2b] "r"(vector2b),
+              [vector4a] "r"(vector4a), [vector3b] "r"(vector3b),
+              [src_ptr] "r"(src_ptr));
 
-                /* clamp and store results */
-                dst_ptr[0] = cm[Temp1];
-                dst_ptr[1] = cm[Temp2];
-                dst_ptr[2] = cm[Temp3];
-                dst_ptr[3] = cm[Temp4];
+        /* clamp and store results */
+        dst_ptr[0] = cm[Temp1];
+        dst_ptr[1] = cm[Temp2];
+        dst_ptr[2] = cm[Temp3];
+        dst_ptr[3] = cm[Temp4];
 
-                /* next 4 pixels */
-                __asm__ __volatile__ (
-                    /* even 3. pixel */
-                    "dpa.w.ph         $ac3,        %[p3],          %[vector1b]    \n\t"
-                    "dpa.w.ph         $ac3,        %[p1],          %[vector2b]    \n\t"
-                    "dpa.w.ph         $ac3,        %[p2],          %[vector3b]    \n\t"
+        /* next 4 pixels */
+        __asm__ __volatile__(
+            /* even 3. pixel */
+            "dpa.w.ph         $ac3,        %[p3],          %[vector1b]    \n\t"
+            "dpa.w.ph         $ac3,        %[p1],          %[vector2b]    \n\t"
+            "dpa.w.ph         $ac3,        %[p2],          %[vector3b]    \n\t"
 
-                    /* even 4. pixel */
-                    "mtlo             %[vector4a], $ac2                           \n\t"
-                    "preceu.ph.qbl    %[p4],       %[tp1]                         \n\t"
-                    "dpa.w.ph         $ac2,        %[p1],          %[vector1b]    \n\t"
-                    "dpa.w.ph         $ac2,        %[p2],          %[vector2b]    \n\t"
-                    "dpa.w.ph         $ac2,        %[p4],          %[vector3b]    \n\t"
+            /* even 4. pixel */
+            "mtlo             %[vector4a], $ac2                           \n\t"
+            "preceu.ph.qbl    %[p4],       %[tp1]                         \n\t"
+            "dpa.w.ph         $ac2,        %[p1],          %[vector1b]    \n\t"
+            "dpa.w.ph         $ac2,        %[p2],          %[vector2b]    \n\t"
+            "dpa.w.ph         $ac2,        %[p4],          %[vector3b]    \n\t"
 
-                    "ulw              %[tn1],      7(%[src_ptr])                  \n\t"
-                    "extp             %[Temp1],    $ac3,           9              \n\t"
-
-                    /* odd 3. pixel */
-                    "mtlo             %[vector4a], $ac3                           \n\t"
-                    "preceu.ph.qbr    %[n2],       %[tn1]                         \n\t"
-                    "dpa.w.ph         $ac3,        %[n3],          %[vector1b]    \n\t"
-                    "dpa.w.ph         $ac3,        %[n1],          %[vector2b]    \n\t"
-                    "dpa.w.ph         $ac3,        %[n2],          %[vector3b]    \n\t"
-                    "extp             %[Temp3],    $ac2,           9              \n\t"
-
-                    /* odd 4. pixel */
-                    "mtlo             %[vector4a], $ac2                           \n\t"
-                    "preceu.ph.qbl    %[n4],       %[tn1]                         \n\t"
-                    "dpa.w.ph         $ac2,        %[n1],          %[vector1b]    \n\t"
-                    "dpa.w.ph         $ac2,        %[n2],          %[vector2b]    \n\t"
-                    "dpa.w.ph         $ac2,        %[n4],          %[vector3b]    \n\t"
-                    "extp             %[Temp2],    $ac3,           9              \n\t"
-                    "extp             %[Temp4],    $ac2,           9              \n\t"
-
-                    : [tn1] "=&r" (tn1), [n2] "=&r" (n2),
-                      [p4] "=&r" (p4), [n4] "=&r" (n4),
-                      [Temp1] "=&r" (Temp1), [Temp2] "=&r" (Temp2),
-                      [Temp3] "=&r" (Temp3), [Temp4] "=r" (Temp4)
-                    : [tp1] "r" (tp1), [vector1b] "r" (vector1b), [p2] "r" (p2),
-                      [vector2b] "r" (vector2b), [n1] "r" (n1), [p1] "r" (p1),
-                      [vector4a] "r" (vector4a), [vector3b] "r" (vector3b),
-                      [p3] "r" (p3), [n3] "r" (n3), [src_ptr] "r" (src_ptr)
-                );
-
-                /* clamp and store results */
-                dst_ptr[4] = cm[Temp1];
-                dst_ptr[5] = cm[Temp2];
-                dst_ptr[6] = cm[Temp3];
-                dst_ptr[7] = cm[Temp4];
-
-                src_ptr += src_pixels_per_line;
-                dst_ptr += pitch;
-            }
-        }
-        else
-        {
-            /* 4 tap filter */
-
-            vector1b = sub_pel_filters_inv_tap_4[xoffset][0];
-            vector2b = sub_pel_filters_inv_tap_4[xoffset][1];
-
-            for (i = output_height; i--;)
-            {
-                /* prefetch src_ptr data to cache memory */
-                prefetch_load(src_ptr + src_pixels_per_line);
+            "ulw              %[tn1],      7(%[src_ptr])                  \n\t"
+            "extp             %[Temp1],    $ac3,           9              \n\t"
 
-                /* apply filter with vectors pairs */
-                __asm__ __volatile__ (
-                    "ulw              %[tp1],      -1(%[src_ptr])                 \n\t"
+            /* odd 3. pixel */
+            "mtlo             %[vector4a], $ac3                           \n\t"
+            "preceu.ph.qbr    %[n2],       %[tn1]                         \n\t"
+            "dpa.w.ph         $ac3,        %[n3],          %[vector1b]    \n\t"
+            "dpa.w.ph         $ac3,        %[n1],          %[vector2b]    \n\t"
+            "dpa.w.ph         $ac3,        %[n2],          %[vector3b]    \n\t"
+            "extp             %[Temp3],    $ac2,           9              \n\t"
 
-                    /* even 1. pixel */
-                    "mtlo             %[vector4a], $ac3                           \n\t"
-                    "preceu.ph.qbr    %[p1],       %[tp1]                         \n\t"
-                    "preceu.ph.qbl    %[p2],       %[tp1]                         \n\t"
-                    "dpa.w.ph         $ac3,        %[p1],          %[vector1b]    \n\t"
-                    "dpa.w.ph         $ac3,        %[p2],          %[vector2b]    \n\t"
+            /* odd 4. pixel */
+            "mtlo             %[vector4a], $ac2                           \n\t"
+            "preceu.ph.qbl    %[n4],       %[tn1]                         \n\t"
+            "dpa.w.ph         $ac2,        %[n1],          %[vector1b]    \n\t"
+            "dpa.w.ph         $ac2,        %[n2],          %[vector2b]    \n\t"
+            "dpa.w.ph         $ac2,        %[n4],          %[vector3b]    \n\t"
+            "extp             %[Temp2],    $ac3,           9              \n\t"
+            "extp             %[Temp4],    $ac2,           9              \n\t"
 
-                    "ulw              %[tp2],      3(%[src_ptr])                  \n\t"
+            : [tn1] "=&r"(tn1), [n2] "=&r"(n2), [p4] "=&r"(p4), [n4] "=&r"(n4),
+              [Temp1] "=&r"(Temp1), [Temp2] "=&r"(Temp2), [Temp3] "=&r"(Temp3),
+              [Temp4] "=r"(Temp4)
+            : [tp1] "r"(tp1), [vector1b] "r"(vector1b), [p2] "r"(p2),
+              [vector2b] "r"(vector2b), [n1] "r"(n1), [p1] "r"(p1),
+              [vector4a] "r"(vector4a), [vector3b] "r"(vector3b), [p3] "r"(p3),
+              [n3] "r"(n3), [src_ptr] "r"(src_ptr));
 
-                    /* even 2. pixel  */
-                    "mtlo             %[vector4a], $ac2                           \n\t"
-                    "preceu.ph.qbr    %[p3],       %[tp2]                         \n\t"
-                    "preceu.ph.qbl    %[p4],       %[tp2]                         \n\t"
-                    "dpa.w.ph         $ac2,        %[p2],          %[vector1b]    \n\t"
-                    "dpa.w.ph         $ac2,        %[p3],          %[vector2b]    \n\t"
-                    "extp             %[Temp1],    $ac3,           9              \n\t"
+        /* clamp and store results */
+        dst_ptr[4] = cm[Temp1];
+        dst_ptr[5] = cm[Temp2];
+        dst_ptr[6] = cm[Temp3];
+        dst_ptr[7] = cm[Temp4];
 
-                    "balign           %[tp2],      %[tp1],         3              \n\t"
+        src_ptr += src_pixels_per_line;
+        dst_ptr += pitch;
+      }
+    } else {
+      /* 4 tap filter */
 
-                    /* odd 1. pixel */
-                    "mtlo             %[vector4a], $ac3                           \n\t"
-                    "preceu.ph.qbr    %[n1],       %[tp2]                         \n\t"
-                    "preceu.ph.qbl    %[n2],       %[tp2]                         \n\t"
-                    "dpa.w.ph         $ac3,        %[n1],          %[vector1b]    \n\t"
-                    "dpa.w.ph         $ac3,        %[n2],          %[vector2b]    \n\t"
-                    "extp             %[Temp3],    $ac2,           9              \n\t"
+      vector1b = sub_pel_filters_inv_tap_4[xoffset][0];
+      vector2b = sub_pel_filters_inv_tap_4[xoffset][1];
 
-                    "ulw              %[tn2],      4(%[src_ptr])                  \n\t"
-
-                    /* odd 2. pixel */
-                    "mtlo             %[vector4a], $ac2                           \n\t"
-                    "preceu.ph.qbr    %[n3],       %[tn2]                         \n\t"
-                    "preceu.ph.qbl    %[n4],       %[tn2]                         \n\t"
-                    "dpa.w.ph         $ac2,        %[n2],          %[vector1b]    \n\t"
-                    "dpa.w.ph         $ac2,        %[n3],          %[vector2b]    \n\t"
-                    "ulw              %[tp1],      7(%[src_ptr])                  \n\t"
-                    "extp             %[Temp2],    $ac3,           9              \n\t"
-                    "mtlo             %[vector4a], $ac3                           \n\t"
-                    "extp             %[Temp4],    $ac2,           9              \n\t"
-
-                    : [tp1] "=&r" (tp1), [tp2] "=&r" (tp2),
-                      [tn2] "=&r" (tn2), [p1] "=&r" (p1), [p2] "=&r" (p2),
-                      [p3] "=&r" (p3), [p4] "=&r" (p4), [n1] "=&r" (n1),
-                      [n2] "=&r" (n2), [n3] "=&r" (n3), [n4] "=&r" (n4),
-                      [Temp1] "=&r" (Temp1), [Temp2] "=&r" (Temp2),
-                      [Temp3] "=&r" (Temp3), [Temp4] "=r" (Temp4)
-                    : [vector1b] "r" (vector1b), [vector2b] "r" (vector2b),
-                      [vector4a] "r" (vector4a), [src_ptr] "r" (src_ptr)
-                );
-
-                /* clamp and store results */
-                dst_ptr[0] = cm[Temp1];
-                dst_ptr[1] = cm[Temp2];
-                dst_ptr[2] = cm[Temp3];
-                dst_ptr[3] = cm[Temp4];
-
-                /* next 4 pixels */
-                __asm__ __volatile__ (
-                    /* even 3. pixel */
-                    "dpa.w.ph         $ac3,        %[p3],          %[vector1b]    \n\t"
-                    "dpa.w.ph         $ac3,        %[p4],          %[vector2b]    \n\t"
-
-                    /* even 4. pixel */
-                    "mtlo             %[vector4a], $ac2                           \n\t"
-                    "preceu.ph.qbr    %[p2],       %[tp1]                         \n\t"
-                    "dpa.w.ph         $ac2,        %[p4],          %[vector1b]    \n\t"
-                    "dpa.w.ph         $ac2,        %[p2],          %[vector2b]    \n\t"
-                    "extp             %[Temp1],    $ac3,           9              \n\t"
-
-                    /* odd 3. pixel */
-                    "mtlo             %[vector4a], $ac3                           \n\t"
-                    "dpa.w.ph         $ac3,        %[n3],          %[vector1b]    \n\t"
-                    "dpa.w.ph         $ac3,        %[n4],          %[vector2b]    \n\t"
-                    "ulw              %[tn1],      8(%[src_ptr])                  \n\t"
-                    "extp             %[Temp3],    $ac2,           9              \n\t"
-
-                    /* odd 4. pixel */
-                    "mtlo             %[vector4a], $ac2                           \n\t"
-                    "preceu.ph.qbr    %[n2],       %[tn1]                         \n\t"
-                    "dpa.w.ph         $ac2,        %[n4],          %[vector1b]    \n\t"
-                    "dpa.w.ph         $ac2,        %[n2],          %[vector2b]    \n\t"
-                    "extp             %[Temp2],    $ac3,           9              \n\t"
-                    "extp             %[Temp4],    $ac2,           9              \n\t"
-
-                    : [tn1] "=&r" (tn1), [p2] "=&r" (p2), [n2] "=&r" (n2),
-                      [Temp1] "=&r" (Temp1), [Temp2] "=&r" (Temp2),
-                      [Temp3] "=&r" (Temp3), [Temp4] "=r" (Temp4)
-                    : [tp1] "r" (tp1), [p3] "r" (p3), [p4] "r" (p4),
-                      [vector1b] "r" (vector1b), [vector2b] "r" (vector2b),
-                      [vector4a] "r" (vector4a), [src_ptr] "r" (src_ptr),
-                      [n3] "r" (n3), [n4] "r" (n4)
-                );
-
-                /* clamp and store results */
-                dst_ptr[4] = cm[Temp1];
-                dst_ptr[5] = cm[Temp2];
-                dst_ptr[6] = cm[Temp3];
-                dst_ptr[7] = cm[Temp4];
-
-                /* next row... */
-                src_ptr += src_pixels_per_line;
-                dst_ptr += pitch;
-            }
-        }
-    }
-}
-
-
-void vp8_filter_block2d_first_pass16_6tap
-(
-    unsigned char *RESTRICT src_ptr,
-    unsigned char *RESTRICT dst_ptr,
-    unsigned int src_pixels_per_line,
-    unsigned int output_height,
-    int xoffset,
-    int pitch
-)
-{
-    unsigned int i;
-    int Temp1, Temp2, Temp3, Temp4;
-
-    unsigned int vector4a;
-    unsigned int vector1b, vector2b, vector3b;
-    unsigned int tp1, tp2, tn1, tn2;
-    unsigned int p1, p2, p3, p4;
-    unsigned int n1, n2, n3, n4;
-    unsigned char *cm = ff_cropTbl + CROP_WIDTH;
-
-    vector1b = sub_pel_filters_inv[xoffset][0];
-    vector2b = sub_pel_filters_inv[xoffset][1];
-    vector3b = sub_pel_filters_inv[xoffset][2];
-    vector4a = 64;
-
-    for (i = output_height; i--;)
-    {
+      for (i = output_height; i--;) {
         /* prefetch src_ptr data to cache memory */
         prefetch_load(src_ptr + src_pixels_per_line);
 
         /* apply filter with vectors pairs */
-        __asm__ __volatile__ (
-            "ulw                %[tp1],      -2(%[src_ptr])                 \n\t"
-            "ulw                %[tp2],      2(%[src_ptr])                  \n\t"
+        __asm__ __volatile__(
+            "ulw              %[tp1],      -1(%[src_ptr])                 \n\t"
 
             /* even 1. pixel */
-            "mtlo               %[vector4a], $ac3                           \n\t"
-            "preceu.ph.qbr      %[p1],       %[tp1]                         \n\t"
-            "preceu.ph.qbl      %[p2],       %[tp1]                         \n\t"
-            "preceu.ph.qbr      %[p3],       %[tp2]                         \n\t"
-            "dpa.w.ph           $ac3,        %[p1],           %[vector1b]   \n\t"
-            "dpa.w.ph           $ac3,        %[p2],           %[vector2b]   \n\t"
-            "dpa.w.ph           $ac3,        %[p3],           %[vector3b]   \n\t"
+            "mtlo             %[vector4a], $ac3                           \n\t"
+            "preceu.ph.qbr    %[p1],       %[tp1]                         \n\t"
+            "preceu.ph.qbl    %[p2],       %[tp1]                         \n\t"
+            "dpa.w.ph         $ac3,        %[p1],          %[vector1b]    \n\t"
+            "dpa.w.ph         $ac3,        %[p2],          %[vector2b]    \n\t"
 
-            /* even 2. pixel */
-            "mtlo               %[vector4a], $ac2                           \n\t"
-            "preceu.ph.qbl      %[p1],       %[tp2]                         \n\t"
-            "dpa.w.ph           $ac2,        %[p2],           %[vector1b]   \n\t"
-            "dpa.w.ph           $ac2,        %[p3],           %[vector2b]   \n\t"
-            "dpa.w.ph           $ac2,        %[p1],           %[vector3b]   \n\t"
+            "ulw              %[tp2],      3(%[src_ptr])                  \n\t"
 
-            "balign             %[tp2],      %[tp1],          3             \n\t"
-            "ulw                %[tn2],      3(%[src_ptr])                  \n\t"
-            "extp               %[Temp1],    $ac3,            9             \n\t"
+            /* even 2. pixel  */
+            "mtlo             %[vector4a], $ac2                           \n\t"
+            "preceu.ph.qbr    %[p3],       %[tp2]                         \n\t"
+            "preceu.ph.qbl    %[p4],       %[tp2]                         \n\t"
+            "dpa.w.ph         $ac2,        %[p2],          %[vector1b]    \n\t"
+            "dpa.w.ph         $ac2,        %[p3],          %[vector2b]    \n\t"
+            "extp             %[Temp1],    $ac3,           9              \n\t"
 
+            "balign           %[tp2],      %[tp1],         3              \n\t"
+
             /* odd 1. pixel */
-            "mtlo               %[vector4a], $ac3                           \n\t"
-            "preceu.ph.qbr      %[n1],       %[tp2]                         \n\t"
-            "preceu.ph.qbl      %[n2],       %[tp2]                         \n\t"
-            "preceu.ph.qbr      %[n3],       %[tn2]                         \n\t"
-            "extp               %[Temp3],    $ac2,            9             \n\t"
-            "dpa.w.ph           $ac3,        %[n1],           %[vector1b]   \n\t"
-            "dpa.w.ph           $ac3,        %[n2],           %[vector2b]   \n\t"
-            "dpa.w.ph           $ac3,        %[n3],           %[vector3b]   \n\t"
+            "mtlo             %[vector4a], $ac3                           \n\t"
+            "preceu.ph.qbr    %[n1],       %[tp2]                         \n\t"
+            "preceu.ph.qbl    %[n2],       %[tp2]                         \n\t"
+            "dpa.w.ph         $ac3,        %[n1],          %[vector1b]    \n\t"
+            "dpa.w.ph         $ac3,        %[n2],          %[vector2b]    \n\t"
+            "extp             %[Temp3],    $ac2,           9              \n\t"
 
+            "ulw              %[tn2],      4(%[src_ptr])                  \n\t"
+
             /* odd 2. pixel */
-            "mtlo               %[vector4a], $ac2                           \n\t"
-            "preceu.ph.qbl      %[n1],       %[tn2]                         \n\t"
-            "dpa.w.ph           $ac2,        %[n2],           %[vector1b]   \n\t"
-            "dpa.w.ph           $ac2,        %[n3],           %[vector2b]   \n\t"
-            "dpa.w.ph           $ac2,        %[n1],           %[vector3b]   \n\t"
-            "ulw                %[tp1],      6(%[src_ptr])                  \n\t"
-            "extp               %[Temp2],    $ac3,            9             \n\t"
-            "mtlo               %[vector4a], $ac3                           \n\t"
-            "preceu.ph.qbr      %[p2],       %[tp1]                         \n\t"
-            "extp               %[Temp4],    $ac2,            9             \n\t"
+            "mtlo             %[vector4a], $ac2                           \n\t"
+            "preceu.ph.qbr    %[n3],       %[tn2]                         \n\t"
+            "preceu.ph.qbl    %[n4],       %[tn2]                         \n\t"
+            "dpa.w.ph         $ac2,        %[n2],          %[vector1b]    \n\t"
+            "dpa.w.ph         $ac2,        %[n3],          %[vector2b]    \n\t"
+            "ulw              %[tp1],      7(%[src_ptr])                  \n\t"
+            "extp             %[Temp2],    $ac3,           9              \n\t"
+            "mtlo             %[vector4a], $ac3                           \n\t"
+            "extp             %[Temp4],    $ac2,           9              \n\t"
 
-            : [tp1] "=&r" (tp1), [tp2] "=&r" (tp2), [tn2] "=&r" (tn2),
-              [p1] "=&r" (p1), [p2] "=&r" (p2), [p3] "=&r" (p3),
-              [n1] "=&r" (n1), [n2] "=&r" (n2), [n3] "=&r" (n3),
-              [Temp1] "=&r" (Temp1), [Temp2] "=&r" (Temp2),
-              [Temp3] "=&r" (Temp3), [Temp4] "=r" (Temp4)
-            : [vector1b] "r" (vector1b), [vector2b] "r" (vector2b),
-              [vector4a] "r" (vector4a), [vector3b] "r" (vector3b),
-              [src_ptr] "r" (src_ptr)
-        );
+            : [tp1] "=&r"(tp1), [tp2] "=&r"(tp2), [tn2] "=&r"(tn2),
+              [p1] "=&r"(p1), [p2] "=&r"(p2), [p3] "=&r"(p3), [p4] "=&r"(p4),
+              [n1] "=&r"(n1), [n2] "=&r"(n2), [n3] "=&r"(n3), [n4] "=&r"(n4),
+              [Temp1] "=&r"(Temp1), [Temp2] "=&r"(Temp2), [Temp3] "=&r"(Temp3),
+              [Temp4] "=r"(Temp4)
+            : [vector1b] "r"(vector1b), [vector2b] "r"(vector2b),
+              [vector4a] "r"(vector4a), [src_ptr] "r"(src_ptr));
 
         /* clamp and store results */
         dst_ptr[0] = cm[Temp1];
@@ -703,50 +501,40 @@
         dst_ptr[3] = cm[Temp4];
 
         /* next 4 pixels */
-        __asm__ __volatile__ (
+        __asm__ __volatile__(
             /* even 3. pixel */
-            "dpa.w.ph           $ac3,        %[p3],           %[vector1b]   \n\t"
-            "dpa.w.ph           $ac3,        %[p1],           %[vector2b]   \n\t"
-            "dpa.w.ph           $ac3,        %[p2],           %[vector3b]   \n\t"
+            "dpa.w.ph         $ac3,        %[p3],          %[vector1b]    \n\t"
+            "dpa.w.ph         $ac3,        %[p4],          %[vector2b]    \n\t"
 
             /* even 4. pixel */
-            "mtlo               %[vector4a], $ac2                           \n\t"
-            "preceu.ph.qbl      %[p4],       %[tp1]                         \n\t"
-            "dpa.w.ph           $ac2,        %[p1],           %[vector1b]   \n\t"
-            "dpa.w.ph           $ac2,        %[p2],           %[vector2b]   \n\t"
-            "dpa.w.ph           $ac2,        %[p4],           %[vector3b]   \n\t"
-            "ulw                %[tn1],      7(%[src_ptr])                  \n\t"
-            "extp               %[Temp1],    $ac3,            9             \n\t"
+            "mtlo             %[vector4a], $ac2                           \n\t"
+            "preceu.ph.qbr    %[p2],       %[tp1]                         \n\t"
+            "dpa.w.ph         $ac2,        %[p4],          %[vector1b]    \n\t"
+            "dpa.w.ph         $ac2,        %[p2],          %[vector2b]    \n\t"
+            "extp             %[Temp1],    $ac3,           9              \n\t"
 
             /* odd 3. pixel */
-            "mtlo               %[vector4a], $ac3                           \n\t"
-            "preceu.ph.qbr      %[n2],       %[tn1]                         \n\t"
-            "dpa.w.ph           $ac3,        %[n3],           %[vector1b]   \n\t"
-            "dpa.w.ph           $ac3,        %[n1],           %[vector2b]   \n\t"
-            "dpa.w.ph           $ac3,        %[n2],           %[vector3b]   \n\t"
-            "extp               %[Temp3],    $ac2,            9             \n\t"
+            "mtlo             %[vector4a], $ac3                           \n\t"
+            "dpa.w.ph         $ac3,        %[n3],          %[vector1b]    \n\t"
+            "dpa.w.ph         $ac3,        %[n4],          %[vector2b]    \n\t"
+            "ulw              %[tn1],      8(%[src_ptr])                  \n\t"
+            "extp             %[Temp3],    $ac2,           9              \n\t"
 
             /* odd 4. pixel */
-            "mtlo               %[vector4a], $ac2                           \n\t"
-            "preceu.ph.qbl      %[n4],       %[tn1]                         \n\t"
-            "dpa.w.ph           $ac2,        %[n1],           %[vector1b]   \n\t"
-            "dpa.w.ph           $ac2,        %[n2],           %[vector2b]   \n\t"
-            "dpa.w.ph           $ac2,        %[n4],           %[vector3b]   \n\t"
-            "ulw                %[tp2],      10(%[src_ptr])                 \n\t"
-            "extp               %[Temp2],    $ac3,            9             \n\t"
-            "mtlo               %[vector4a], $ac3                           \n\t"
-            "preceu.ph.qbr      %[p1],       %[tp2]                         \n\t"
-            "extp               %[Temp4],    $ac2,            9             \n\t"
+            "mtlo             %[vector4a], $ac2                           \n\t"
+            "preceu.ph.qbr    %[n2],       %[tn1]                         \n\t"
+            "dpa.w.ph         $ac2,        %[n4],          %[vector1b]    \n\t"
+            "dpa.w.ph         $ac2,        %[n2],          %[vector2b]    \n\t"
+            "extp             %[Temp2],    $ac3,           9              \n\t"
+            "extp             %[Temp4],    $ac2,           9              \n\t"
 
-            : [tn1] "=&r" (tn1), [tp2] "=&r" (tp2), [n2] "=&r" (n2),
-              [p4] "=&r" (p4), [n4] "=&r" (n4),
-              [Temp1] "=&r" (Temp1), [Temp2] "=&r" (Temp2),
-              [Temp3] "=&r" (Temp3), [Temp4] "=r" (Temp4)
-            : [vector1b] "r" (vector1b), [vector2b] "r" (vector2b),
-              [tp1] "r" (tp1), [n1] "r" (n1), [p1] "r" (p1),
-              [vector4a] "r" (vector4a), [p2] "r" (p2), [vector3b] "r" (vector3b),
-              [p3] "r" (p3), [n3] "r" (n3), [src_ptr] "r" (src_ptr)
-        );
+            : [tn1] "=&r"(tn1), [p2] "=&r"(p2), [n2] "=&r"(n2),
+              [Temp1] "=&r"(Temp1), [Temp2] "=&r"(Temp2), [Temp3] "=&r"(Temp3),
+              [Temp4] "=r"(Temp4)
+            : [tp1] "r"(tp1), [p3] "r"(p3), [p4] "r"(p4),
+              [vector1b] "r"(vector1b), [vector2b] "r"(vector2b),
+              [vector4a] "r"(vector4a), [src_ptr] "r"(src_ptr), [n3] "r"(n3),
+              [n4] "r"(n4));
 
         /* clamp and store results */
         dst_ptr[4] = cm[Temp1];
@@ -754,2070 +542,2225 @@
         dst_ptr[6] = cm[Temp3];
         dst_ptr[7] = cm[Temp4];
 
-        /* next 4 pixels */
-        __asm__ __volatile__ (
-            /* even 5. pixel */
-            "dpa.w.ph           $ac3,        %[p2],           %[vector1b]   \n\t"
-            "dpa.w.ph           $ac3,        %[p4],           %[vector2b]   \n\t"
-            "dpa.w.ph           $ac3,        %[p1],           %[vector3b]   \n\t"
+        /* next row... */
+        src_ptr += src_pixels_per_line;
+        dst_ptr += pitch;
+      }
+    }
+  }
+}
 
-            /* even 6. pixel */
-            "mtlo               %[vector4a], $ac2                           \n\t"
-            "preceu.ph.qbl      %[p3],       %[tp2]                         \n\t"
-            "dpa.w.ph           $ac2,        %[p4],           %[vector1b]   \n\t"
-            "dpa.w.ph           $ac2,        %[p1],           %[vector2b]   \n\t"
-            "dpa.w.ph           $ac2,        %[p3],           %[vector3b]   \n\t"
+void vp8_filter_block2d_first_pass16_6tap(unsigned char *RESTRICT src_ptr,
+                                          unsigned char *RESTRICT dst_ptr,
+                                          unsigned int src_pixels_per_line,
+                                          unsigned int output_height,
+                                          int xoffset, int pitch) {
+  unsigned int i;
+  int Temp1, Temp2, Temp3, Temp4;
 
-            "ulw                %[tn1],      11(%[src_ptr])                 \n\t"
-            "extp               %[Temp1],    $ac3,            9             \n\t"
+  unsigned int vector4a;
+  unsigned int vector1b, vector2b, vector3b;
+  unsigned int tp1, tp2, tn1, tn2;
+  unsigned int p1, p2, p3, p4;
+  unsigned int n1, n2, n3, n4;
+  unsigned char *cm = ff_cropTbl + CROP_WIDTH;
 
-            /* odd 5. pixel */
-            "mtlo               %[vector4a], $ac3                           \n\t"
-            "preceu.ph.qbr      %[n1],       %[tn1]                         \n\t"
-            "dpa.w.ph           $ac3,        %[n2],           %[vector1b]   \n\t"
-            "dpa.w.ph           $ac3,        %[n4],           %[vector2b]   \n\t"
-            "dpa.w.ph           $ac3,        %[n1],           %[vector3b]   \n\t"
-            "extp               %[Temp3],    $ac2,            9             \n\t"
+  vector1b = sub_pel_filters_inv[xoffset][0];
+  vector2b = sub_pel_filters_inv[xoffset][1];
+  vector3b = sub_pel_filters_inv[xoffset][2];
+  vector4a = 64;
 
-            /* odd 6. pixel */
-            "mtlo               %[vector4a], $ac2                           \n\t"
-            "preceu.ph.qbl      %[n3],       %[tn1]                         \n\t"
-            "dpa.w.ph           $ac2,        %[n4],           %[vector1b]   \n\t"
-            "dpa.w.ph           $ac2,        %[n1],           %[vector2b]   \n\t"
-            "dpa.w.ph           $ac2,        %[n3],           %[vector3b]   \n\t"
-            "ulw                %[tp1],      14(%[src_ptr])                 \n\t"
-            "extp               %[Temp2],    $ac3,            9             \n\t"
-            "mtlo               %[vector4a], $ac3                           \n\t"
-            "preceu.ph.qbr      %[p4],       %[tp1]                         \n\t"
-            "extp               %[Temp4],    $ac2,            9             \n\t"
+  for (i = output_height; i--;) {
+    /* prefetch src_ptr data to cache memory */
+    prefetch_load(src_ptr + src_pixels_per_line);
 
-            : [tn1] "=&r" (tn1), [tp1] "=&r" (tp1),
-              [n1] "=&r" (n1), [p3] "=&r" (p3), [n3] "=&r" (n3),
-              [Temp1] "=&r" (Temp1), [Temp2] "=&r" (Temp2),
-              [Temp3] "=&r" (Temp3), [Temp4] "=r" (Temp4)
-            : [vector1b] "r" (vector1b), [vector2b] "r" (vector2b),
-              [tp2] "r" (tp2), [p2] "r" (p2), [n2] "r" (n2),
-              [p4] "r" (p4), [n4] "r" (n4), [p1] "r" (p1), [src_ptr] "r" (src_ptr),
-              [vector4a] "r" (vector4a), [vector3b] "r" (vector3b)
-        );
+    /* apply filter with vectors pairs */
+    __asm__ __volatile__(
+        "ulw                %[tp1],      -2(%[src_ptr])                 \n\t"
+        "ulw                %[tp2],      2(%[src_ptr])                  \n\t"
 
-        /* clamp and store results */
-        dst_ptr[8] = cm[Temp1];
-        dst_ptr[9] = cm[Temp2];
-        dst_ptr[10] = cm[Temp3];
-        dst_ptr[11] = cm[Temp4];
+        /* even 1. pixel */
+        "mtlo               %[vector4a], $ac3                           \n\t"
+        "preceu.ph.qbr      %[p1],       %[tp1]                         \n\t"
+        "preceu.ph.qbl      %[p2],       %[tp1]                         \n\t"
+        "preceu.ph.qbr      %[p3],       %[tp2]                         \n\t"
+        "dpa.w.ph           $ac3,        %[p1],           %[vector1b]   \n\t"
+        "dpa.w.ph           $ac3,        %[p2],           %[vector2b]   \n\t"
+        "dpa.w.ph           $ac3,        %[p3],           %[vector3b]   \n\t"
 
-        /* next 4 pixels */
-        __asm__ __volatile__ (
-            /* even 7. pixel */
-            "dpa.w.ph           $ac3,        %[p1],           %[vector1b]   \n\t"
-            "dpa.w.ph           $ac3,        %[p3],           %[vector2b]   \n\t"
-            "dpa.w.ph           $ac3,        %[p4],           %[vector3b]   \n\t"
+        /* even 2. pixel */
+        "mtlo               %[vector4a], $ac2                           \n\t"
+        "preceu.ph.qbl      %[p1],       %[tp2]                         \n\t"
+        "dpa.w.ph           $ac2,        %[p2],           %[vector1b]   \n\t"
+        "dpa.w.ph           $ac2,        %[p3],           %[vector2b]   \n\t"
+        "dpa.w.ph           $ac2,        %[p1],           %[vector3b]   \n\t"
 
-            /* even 8. pixel */
-            "mtlo               %[vector4a], $ac2                           \n\t"
-            "preceu.ph.qbl      %[p2],       %[tp1]                         \n\t"
-            "dpa.w.ph           $ac2,        %[p3],           %[vector1b]   \n\t"
-            "dpa.w.ph           $ac2,        %[p4],           %[vector2b]   \n\t"
-            "dpa.w.ph           $ac2,        %[p2],           %[vector3b]   \n\t"
-            "ulw                %[tn1],      15(%[src_ptr])                 \n\t"
-            "extp               %[Temp1],    $ac3,            9             \n\t"
+        "balign             %[tp2],      %[tp1],          3             \n\t"
+        "ulw                %[tn2],      3(%[src_ptr])                  \n\t"
+        "extp               %[Temp1],    $ac3,            9             \n\t"
 
-            /* odd 7. pixel */
-            "mtlo               %[vector4a], $ac3                           \n\t"
-            "preceu.ph.qbr      %[n4],       %[tn1]                         \n\t"
-            "dpa.w.ph           $ac3,        %[n1],           %[vector1b]   \n\t"
-            "dpa.w.ph           $ac3,        %[n3],           %[vector2b]   \n\t"
-            "dpa.w.ph           $ac3,        %[n4],           %[vector3b]   \n\t"
-            "extp               %[Temp3],    $ac2,            9             \n\t"
+        /* odd 1. pixel */
+        "mtlo               %[vector4a], $ac3                           \n\t"
+        "preceu.ph.qbr      %[n1],       %[tp2]                         \n\t"
+        "preceu.ph.qbl      %[n2],       %[tp2]                         \n\t"
+        "preceu.ph.qbr      %[n3],       %[tn2]                         \n\t"
+        "extp               %[Temp3],    $ac2,            9             \n\t"
+        "dpa.w.ph           $ac3,        %[n1],           %[vector1b]   \n\t"
+        "dpa.w.ph           $ac3,        %[n2],           %[vector2b]   \n\t"
+        "dpa.w.ph           $ac3,        %[n3],           %[vector3b]   \n\t"
 
-            /* odd 8. pixel */
-            "mtlo               %[vector4a], $ac2                           \n\t"
-            "preceu.ph.qbl      %[n2],       %[tn1]                         \n\t"
-            "dpa.w.ph           $ac2,        %[n3],           %[vector1b]   \n\t"
-            "dpa.w.ph           $ac2,        %[n4],           %[vector2b]   \n\t"
-            "dpa.w.ph           $ac2,        %[n2],           %[vector3b]   \n\t"
-            "extp               %[Temp2],    $ac3,            9             \n\t"
-            "extp               %[Temp4],    $ac2,            9             \n\t"
+        /* odd 2. pixel */
+        "mtlo               %[vector4a], $ac2                           \n\t"
+        "preceu.ph.qbl      %[n1],       %[tn2]                         \n\t"
+        "dpa.w.ph           $ac2,        %[n2],           %[vector1b]   \n\t"
+        "dpa.w.ph           $ac2,        %[n3],           %[vector2b]   \n\t"
+        "dpa.w.ph           $ac2,        %[n1],           %[vector3b]   \n\t"
+        "ulw                %[tp1],      6(%[src_ptr])                  \n\t"
+        "extp               %[Temp2],    $ac3,            9             \n\t"
+        "mtlo               %[vector4a], $ac3                           \n\t"
+        "preceu.ph.qbr      %[p2],       %[tp1]                         \n\t"
+        "extp               %[Temp4],    $ac2,            9             \n\t"
 
-            /* clamp and store results */
-            "lbux               %[tp1],      %[Temp1](%[cm])                \n\t"
-            "lbux               %[tn1],      %[Temp2](%[cm])                \n\t"
-            "lbux               %[p2],       %[Temp3](%[cm])                \n\t"
-            "sb                 %[tp1],      12(%[dst_ptr])                 \n\t"
-            "sb                 %[tn1],      13(%[dst_ptr])                 \n\t"
-            "lbux               %[n2],       %[Temp4](%[cm])                \n\t"
-            "sb                 %[p2],       14(%[dst_ptr])                 \n\t"
-            "sb                 %[n2],       15(%[dst_ptr])                 \n\t"
+        : [tp1] "=&r"(tp1), [tp2] "=&r"(tp2), [tn2] "=&r"(tn2), [p1] "=&r"(p1),
+          [p2] "=&r"(p2), [p3] "=&r"(p3), [n1] "=&r"(n1), [n2] "=&r"(n2),
+          [n3] "=&r"(n3), [Temp1] "=&r"(Temp1), [Temp2] "=&r"(Temp2),
+          [Temp3] "=&r"(Temp3), [Temp4] "=r"(Temp4)
+        : [vector1b] "r"(vector1b), [vector2b] "r"(vector2b),
+          [vector4a] "r"(vector4a), [vector3b] "r"(vector3b),
+          [src_ptr] "r"(src_ptr));
 
-            : [tn1] "=&r" (tn1), [p2] "=&r" (p2), [n2] "=&r" (n2), [n4] "=&r" (n4),
-              [Temp1] "=&r" (Temp1), [Temp2] "=&r" (Temp2),
-              [Temp3] "=&r" (Temp3), [Temp4] "=r" (Temp4)
-            : [vector1b] "r" (vector1b), [vector2b] "r" (vector2b),
-              [tp1] "r" (tp1), [p4] "r" (p4), [n1] "r" (n1), [p1] "r" (p1),
-              [vector4a] "r" (vector4a), [vector3b] "r" (vector3b), [p3] "r" (p3),
-              [n3] "r" (n3), [src_ptr] "r" (src_ptr),
-              [cm] "r" (cm), [dst_ptr] "r" (dst_ptr)
-        );
+    /* clamp and store results */
+    dst_ptr[0] = cm[Temp1];
+    dst_ptr[1] = cm[Temp2];
+    dst_ptr[2] = cm[Temp3];
+    dst_ptr[3] = cm[Temp4];
 
-        src_ptr += src_pixels_per_line;
-        dst_ptr += pitch;
-    }
-}
+    /* next 4 pixels */
+    __asm__ __volatile__(
+        /* even 3. pixel */
+        "dpa.w.ph           $ac3,        %[p3],           %[vector1b]   \n\t"
+        "dpa.w.ph           $ac3,        %[p1],           %[vector2b]   \n\t"
+        "dpa.w.ph           $ac3,        %[p2],           %[vector3b]   \n\t"
 
+        /* even 4. pixel */
+        "mtlo               %[vector4a], $ac2                           \n\t"
+        "preceu.ph.qbl      %[p4],       %[tp1]                         \n\t"
+        "dpa.w.ph           $ac2,        %[p1],           %[vector1b]   \n\t"
+        "dpa.w.ph           $ac2,        %[p2],           %[vector2b]   \n\t"
+        "dpa.w.ph           $ac2,        %[p4],           %[vector3b]   \n\t"
+        "ulw                %[tn1],      7(%[src_ptr])                  \n\t"
+        "extp               %[Temp1],    $ac3,            9             \n\t"
 
-void vp8_filter_block2d_first_pass16_0
-(
-    unsigned char *RESTRICT src_ptr,
-    unsigned char *RESTRICT output_ptr,
-    unsigned int src_pixels_per_line
-)
-{
-    int Temp1, Temp2, Temp3, Temp4;
-    int i;
+        /* odd 3. pixel */
+        "mtlo               %[vector4a], $ac3                           \n\t"
+        "preceu.ph.qbr      %[n2],       %[tn1]                         \n\t"
+        "dpa.w.ph           $ac3,        %[n3],           %[vector1b]   \n\t"
+        "dpa.w.ph           $ac3,        %[n1],           %[vector2b]   \n\t"
+        "dpa.w.ph           $ac3,        %[n2],           %[vector3b]   \n\t"
+        "extp               %[Temp3],    $ac2,            9             \n\t"
 
-    /* prefetch src_ptr data to cache memory */
-    prefetch_store(output_ptr + 32);
+        /* odd 4. pixel */
+        "mtlo               %[vector4a], $ac2                           \n\t"
+        "preceu.ph.qbl      %[n4],       %[tn1]                         \n\t"
+        "dpa.w.ph           $ac2,        %[n1],           %[vector1b]   \n\t"
+        "dpa.w.ph           $ac2,        %[n2],           %[vector2b]   \n\t"
+        "dpa.w.ph           $ac2,        %[n4],           %[vector3b]   \n\t"
+        "ulw                %[tp2],      10(%[src_ptr])                 \n\t"
+        "extp               %[Temp2],    $ac3,            9             \n\t"
+        "mtlo               %[vector4a], $ac3                           \n\t"
+        "preceu.ph.qbr      %[p1],       %[tp2]                         \n\t"
+        "extp               %[Temp4],    $ac2,            9             \n\t"
 
-    /* copy memory from src buffer to dst buffer */
-    for (i = 0; i < 7; i++)
-    {
-        __asm__ __volatile__ (
-            "ulw    %[Temp1],   0(%[src_ptr])                               \n\t"
-            "ulw    %[Temp2],   4(%[src_ptr])                               \n\t"
-            "ulw    %[Temp3],   8(%[src_ptr])                               \n\t"
-            "ulw    %[Temp4],   12(%[src_ptr])                              \n\t"
-            "sw     %[Temp1],   0(%[output_ptr])                            \n\t"
-            "sw     %[Temp2],   4(%[output_ptr])                            \n\t"
-            "sw     %[Temp3],   8(%[output_ptr])                            \n\t"
-            "sw     %[Temp4],   12(%[output_ptr])                           \n\t"
-            "addu   %[src_ptr], %[src_ptr],        %[src_pixels_per_line]   \n\t"
+        : [tn1] "=&r"(tn1), [tp2] "=&r"(tp2), [n2] "=&r"(n2), [p4] "=&r"(p4),
+          [n4] "=&r"(n4), [Temp1] "=&r"(Temp1), [Temp2] "=&r"(Temp2),
+          [Temp3] "=&r"(Temp3), [Temp4] "=r"(Temp4)
+        : [vector1b] "r"(vector1b), [vector2b] "r"(vector2b), [tp1] "r"(tp1),
+          [n1] "r"(n1), [p1] "r"(p1), [vector4a] "r"(vector4a), [p2] "r"(p2),
+          [vector3b] "r"(vector3b), [p3] "r"(p3), [n3] "r"(n3),
+          [src_ptr] "r"(src_ptr));
 
-            : [Temp1] "=&r" (Temp1), [Temp2] "=&r" (Temp2), [Temp3] "=&r" (Temp3),
-              [Temp4] "=&r" (Temp4), [src_ptr] "+r" (src_ptr)
-            : [src_pixels_per_line] "r" (src_pixels_per_line),
-              [output_ptr] "r" (output_ptr)
-        );
+    /* clamp and store results */
+    dst_ptr[4] = cm[Temp1];
+    dst_ptr[5] = cm[Temp2];
+    dst_ptr[6] = cm[Temp3];
+    dst_ptr[7] = cm[Temp4];
 
-        __asm__ __volatile__ (
-            "ulw    %[Temp1],   0(%[src_ptr])                               \n\t"
-            "ulw    %[Temp2],   4(%[src_ptr])                               \n\t"
-            "ulw    %[Temp3],   8(%[src_ptr])                               \n\t"
-            "ulw    %[Temp4],   12(%[src_ptr])                              \n\t"
-            "sw     %[Temp1],   16(%[output_ptr])                           \n\t"
-            "sw     %[Temp2],   20(%[output_ptr])                           \n\t"
-            "sw     %[Temp3],   24(%[output_ptr])                           \n\t"
-            "sw     %[Temp4],   28(%[output_ptr])                           \n\t"
-            "addu   %[src_ptr], %[src_ptr],        %[src_pixels_per_line]   \n\t"
+    /* next 4 pixels */
+    __asm__ __volatile__(
+        /* even 5. pixel */
+        "dpa.w.ph           $ac3,        %[p2],           %[vector1b]   \n\t"
+        "dpa.w.ph           $ac3,        %[p4],           %[vector2b]   \n\t"
+        "dpa.w.ph           $ac3,        %[p1],           %[vector3b]   \n\t"
 
-            : [Temp1] "=&r" (Temp1), [Temp2] "=&r" (Temp2), [Temp3] "=&r" (Temp3),
-              [Temp4] "=&r" (Temp4), [src_ptr] "+r" (src_ptr)
-            : [src_pixels_per_line] "r" (src_pixels_per_line),
-              [output_ptr] "r" (output_ptr)
-        );
+        /* even 6. pixel */
+        "mtlo               %[vector4a], $ac2                           \n\t"
+        "preceu.ph.qbl      %[p3],       %[tp2]                         \n\t"
+        "dpa.w.ph           $ac2,        %[p4],           %[vector1b]   \n\t"
+        "dpa.w.ph           $ac2,        %[p1],           %[vector2b]   \n\t"
+        "dpa.w.ph           $ac2,        %[p3],           %[vector3b]   \n\t"
 
-        __asm__ __volatile__ (
-            "ulw    %[Temp1],   0(%[src_ptr])                               \n\t"
-            "ulw    %[Temp2],   4(%[src_ptr])                               \n\t"
-            "ulw    %[Temp3],   8(%[src_ptr])                               \n\t"
-            "ulw    %[Temp4],   12(%[src_ptr])                              \n\t"
-            "sw     %[Temp1],   32(%[output_ptr])                           \n\t"
-            "sw     %[Temp2],   36(%[output_ptr])                           \n\t"
-            "sw     %[Temp3],   40(%[output_ptr])                           \n\t"
-            "sw     %[Temp4],   44(%[output_ptr])                           \n\t"
-            "addu   %[src_ptr], %[src_ptr],        %[src_pixels_per_line]   \n\t"
+        "ulw                %[tn1],      11(%[src_ptr])                 \n\t"
+        "extp               %[Temp1],    $ac3,            9             \n\t"
 
-            : [Temp1] "=&r" (Temp1), [Temp2] "=&r" (Temp2), [Temp3] "=&r" (Temp3),
-              [Temp4] "=&r" (Temp4), [src_ptr] "+r" (src_ptr)
-            : [src_pixels_per_line] "r" (src_pixels_per_line),
-              [output_ptr] "r" (output_ptr)
-        );
+        /* odd 5. pixel */
+        "mtlo               %[vector4a], $ac3                           \n\t"
+        "preceu.ph.qbr      %[n1],       %[tn1]                         \n\t"
+        "dpa.w.ph           $ac3,        %[n2],           %[vector1b]   \n\t"
+        "dpa.w.ph           $ac3,        %[n4],           %[vector2b]   \n\t"
+        "dpa.w.ph           $ac3,        %[n1],           %[vector3b]   \n\t"
+        "extp               %[Temp3],    $ac2,            9             \n\t"
 
-        output_ptr += 48;
-    }
+        /* odd 6. pixel */
+        "mtlo               %[vector4a], $ac2                           \n\t"
+        "preceu.ph.qbl      %[n3],       %[tn1]                         \n\t"
+        "dpa.w.ph           $ac2,        %[n4],           %[vector1b]   \n\t"
+        "dpa.w.ph           $ac2,        %[n1],           %[vector2b]   \n\t"
+        "dpa.w.ph           $ac2,        %[n3],           %[vector3b]   \n\t"
+        "ulw                %[tp1],      14(%[src_ptr])                 \n\t"
+        "extp               %[Temp2],    $ac3,            9             \n\t"
+        "mtlo               %[vector4a], $ac3                           \n\t"
+        "preceu.ph.qbr      %[p4],       %[tp1]                         \n\t"
+        "extp               %[Temp4],    $ac2,            9             \n\t"
+
+        : [tn1] "=&r"(tn1), [tp1] "=&r"(tp1), [n1] "=&r"(n1), [p3] "=&r"(p3),
+          [n3] "=&r"(n3), [Temp1] "=&r"(Temp1), [Temp2] "=&r"(Temp2),
+          [Temp3] "=&r"(Temp3), [Temp4] "=r"(Temp4)
+        : [vector1b] "r"(vector1b), [vector2b] "r"(vector2b), [tp2] "r"(tp2),
+          [p2] "r"(p2), [n2] "r"(n2), [p4] "r"(p4), [n4] "r"(n4), [p1] "r"(p1),
+          [src_ptr] "r"(src_ptr), [vector4a] "r"(vector4a),
+          [vector3b] "r"(vector3b));
+
+    /* clamp and store results */
+    dst_ptr[8] = cm[Temp1];
+    dst_ptr[9] = cm[Temp2];
+    dst_ptr[10] = cm[Temp3];
+    dst_ptr[11] = cm[Temp4];
+
+    /* next 4 pixels */
+    __asm__ __volatile__(
+        /* even 7. pixel */
+        "dpa.w.ph           $ac3,        %[p1],           %[vector1b]   \n\t"
+        "dpa.w.ph           $ac3,        %[p3],           %[vector2b]   \n\t"
+        "dpa.w.ph           $ac3,        %[p4],           %[vector3b]   \n\t"
+
+        /* even 8. pixel */
+        "mtlo               %[vector4a], $ac2                           \n\t"
+        "preceu.ph.qbl      %[p2],       %[tp1]                         \n\t"
+        "dpa.w.ph           $ac2,        %[p3],           %[vector1b]   \n\t"
+        "dpa.w.ph           $ac2,        %[p4],           %[vector2b]   \n\t"
+        "dpa.w.ph           $ac2,        %[p2],           %[vector3b]   \n\t"
+        "ulw                %[tn1],      15(%[src_ptr])                 \n\t"
+        "extp               %[Temp1],    $ac3,            9             \n\t"
+
+        /* odd 7. pixel */
+        "mtlo               %[vector4a], $ac3                           \n\t"
+        "preceu.ph.qbr      %[n4],       %[tn1]                         \n\t"
+        "dpa.w.ph           $ac3,        %[n1],           %[vector1b]   \n\t"
+        "dpa.w.ph           $ac3,        %[n3],           %[vector2b]   \n\t"
+        "dpa.w.ph           $ac3,        %[n4],           %[vector3b]   \n\t"
+        "extp               %[Temp3],    $ac2,            9             \n\t"
+
+        /* odd 8. pixel */
+        "mtlo               %[vector4a], $ac2                           \n\t"
+        "preceu.ph.qbl      %[n2],       %[tn1]                         \n\t"
+        "dpa.w.ph           $ac2,        %[n3],           %[vector1b]   \n\t"
+        "dpa.w.ph           $ac2,        %[n4],           %[vector2b]   \n\t"
+        "dpa.w.ph           $ac2,        %[n2],           %[vector3b]   \n\t"
+        "extp               %[Temp2],    $ac3,            9             \n\t"
+        "extp               %[Temp4],    $ac2,            9             \n\t"
+
+        /* clamp and store results */
+        "lbux               %[tp1],      %[Temp1](%[cm])                \n\t"
+        "lbux               %[tn1],      %[Temp2](%[cm])                \n\t"
+        "lbux               %[p2],       %[Temp3](%[cm])                \n\t"
+        "sb                 %[tp1],      12(%[dst_ptr])                 \n\t"
+        "sb                 %[tn1],      13(%[dst_ptr])                 \n\t"
+        "lbux               %[n2],       %[Temp4](%[cm])                \n\t"
+        "sb                 %[p2],       14(%[dst_ptr])                 \n\t"
+        "sb                 %[n2],       15(%[dst_ptr])                 \n\t"
+
+        : [tn1] "=&r"(tn1), [p2] "=&r"(p2), [n2] "=&r"(n2), [n4] "=&r"(n4),
+          [Temp1] "=&r"(Temp1), [Temp2] "=&r"(Temp2), [Temp3] "=&r"(Temp3),
+          [Temp4] "=r"(Temp4)
+        : [vector1b] "r"(vector1b), [vector2b] "r"(vector2b), [tp1] "r"(tp1),
+          [p4] "r"(p4), [n1] "r"(n1), [p1] "r"(p1), [vector4a] "r"(vector4a),
+          [vector3b] "r"(vector3b), [p3] "r"(p3), [n3] "r"(n3),
+          [src_ptr] "r"(src_ptr), [cm] "r"(cm), [dst_ptr] "r"(dst_ptr));
+
+    src_ptr += src_pixels_per_line;
+    dst_ptr += pitch;
+  }
 }
 
+void vp8_filter_block2d_first_pass16_0(unsigned char *RESTRICT src_ptr,
+                                       unsigned char *RESTRICT output_ptr,
+                                       unsigned int src_pixels_per_line) {
+  int Temp1, Temp2, Temp3, Temp4;
+  int i;
 
-void vp8_filter_block2d_first_pass16_4tap
-(
-    unsigned char *RESTRICT src_ptr,
-    unsigned char *RESTRICT output_ptr,
-    unsigned int src_pixels_per_line,
-    unsigned int output_width,
-    unsigned int output_height,
-    int xoffset,
-    int yoffset,
-    unsigned char *RESTRICT dst_ptr,
-    int pitch
-)
-{
-    unsigned int i, j;
-    int Temp1, Temp2, Temp3, Temp4;
+  /* prefetch src_ptr data to cache memory */
+  prefetch_store(output_ptr + 32);
 
-    unsigned int vector4a;
-    int vector1b, vector2b;
-    unsigned int tp1, tp2, tp3, tn1;
-    unsigned int p1, p2, p3;
-    unsigned int n1, n2, n3;
-    unsigned char *cm = ff_cropTbl + CROP_WIDTH;
+  /* copy memory from src buffer to dst buffer */
+  for (i = 0; i < 7; i++) {
+    __asm__ __volatile__(
+        "ulw    %[Temp1],   0(%[src_ptr])                               \n\t"
+        "ulw    %[Temp2],   4(%[src_ptr])                               \n\t"
+        "ulw    %[Temp3],   8(%[src_ptr])                               \n\t"
+        "ulw    %[Temp4],   12(%[src_ptr])                              \n\t"
+        "sw     %[Temp1],   0(%[output_ptr])                            \n\t"
+        "sw     %[Temp2],   4(%[output_ptr])                            \n\t"
+        "sw     %[Temp3],   8(%[output_ptr])                            \n\t"
+        "sw     %[Temp4],   12(%[output_ptr])                           \n\t"
+        "addu   %[src_ptr], %[src_ptr],        %[src_pixels_per_line]   \n\t"
 
-    vector4a = 64;
+        : [Temp1] "=&r"(Temp1), [Temp2] "=&r"(Temp2), [Temp3] "=&r"(Temp3),
+          [Temp4] "=&r"(Temp4), [src_ptr] "+r"(src_ptr)
+        : [src_pixels_per_line] "r"(src_pixels_per_line),
+          [output_ptr] "r"(output_ptr));
 
-    vector1b = sub_pel_filters_inv_tap_4[xoffset][0];
-    vector2b = sub_pel_filters_inv_tap_4[xoffset][1];
+    __asm__ __volatile__(
+        "ulw    %[Temp1],   0(%[src_ptr])                               \n\t"
+        "ulw    %[Temp2],   4(%[src_ptr])                               \n\t"
+        "ulw    %[Temp3],   8(%[src_ptr])                               \n\t"
+        "ulw    %[Temp4],   12(%[src_ptr])                              \n\t"
+        "sw     %[Temp1],   16(%[output_ptr])                           \n\t"
+        "sw     %[Temp2],   20(%[output_ptr])                           \n\t"
+        "sw     %[Temp3],   24(%[output_ptr])                           \n\t"
+        "sw     %[Temp4],   28(%[output_ptr])                           \n\t"
+        "addu   %[src_ptr], %[src_ptr],        %[src_pixels_per_line]   \n\t"
 
-    /* if (yoffset == 0) don't need temp buffer, data will be stored in dst_ptr */
-    if (yoffset == 0)
-    {
-        output_height -= 5;
-        src_ptr += (src_pixels_per_line + src_pixels_per_line);
+        : [Temp1] "=&r"(Temp1), [Temp2] "=&r"(Temp2), [Temp3] "=&r"(Temp3),
+          [Temp4] "=&r"(Temp4), [src_ptr] "+r"(src_ptr)
+        : [src_pixels_per_line] "r"(src_pixels_per_line),
+          [output_ptr] "r"(output_ptr));
 
-        for (i = output_height; i--;)
-        {
-            __asm__ __volatile__ (
-                "ulw     %[tp3],   -1(%[src_ptr])               \n\t"
-                : [tp3] "=&r" (tp3)
-                : [src_ptr] "r" (src_ptr)
-            );
+    __asm__ __volatile__(
+        "ulw    %[Temp1],   0(%[src_ptr])                               \n\t"
+        "ulw    %[Temp2],   4(%[src_ptr])                               \n\t"
+        "ulw    %[Temp3],   8(%[src_ptr])                               \n\t"
+        "ulw    %[Temp4],   12(%[src_ptr])                              \n\t"
+        "sw     %[Temp1],   32(%[output_ptr])                           \n\t"
+        "sw     %[Temp2],   36(%[output_ptr])                           \n\t"
+        "sw     %[Temp3],   40(%[output_ptr])                           \n\t"
+        "sw     %[Temp4],   44(%[output_ptr])                           \n\t"
+        "addu   %[src_ptr], %[src_ptr],        %[src_pixels_per_line]   \n\t"
 
-            /* processing 4 adjacent pixels */
-            for (j = 0; j < 16; j += 4)
-            {
-                /* apply filter with vectors pairs */
-                __asm__ __volatile__ (
-                    "ulw              %[tp2],      3(%[src_ptr])                    \n\t"
-                    "move             %[tp1],      %[tp3]                           \n\t"
+        : [Temp1] "=&r"(Temp1), [Temp2] "=&r"(Temp2), [Temp3] "=&r"(Temp3),
+          [Temp4] "=&r"(Temp4), [src_ptr] "+r"(src_ptr)
+        : [src_pixels_per_line] "r"(src_pixels_per_line),
+          [output_ptr] "r"(output_ptr));
 
-                    /* even 1. pixel */
-                    "mtlo             %[vector4a], $ac3                             \n\t"
-                    "mthi             $0,          $ac3                             \n\t"
-                    "move             %[tp3],      %[tp2]                           \n\t"
-                    "preceu.ph.qbr    %[p1],       %[tp1]                           \n\t"
-                    "preceu.ph.qbl    %[p2],       %[tp1]                           \n\t"
-                    "preceu.ph.qbr    %[p3],       %[tp2]                           \n\t"
-                    "dpa.w.ph         $ac3,        %[p1],           %[vector1b]     \n\t"
-                    "dpa.w.ph         $ac3,        %[p2],           %[vector2b]     \n\t"
+    output_ptr += 48;
+  }
+}
 
-                    /* even 2. pixel */
-                    "mtlo             %[vector4a], $ac2                             \n\t"
-                    "mthi             $0,          $ac2                             \n\t"
-                    "dpa.w.ph         $ac2,        %[p2],           %[vector1b]     \n\t"
-                    "dpa.w.ph         $ac2,        %[p3],           %[vector2b]     \n\t"
-                    "extr.w           %[Temp1],    $ac3,            7               \n\t"
+void vp8_filter_block2d_first_pass16_4tap(
+    unsigned char *RESTRICT src_ptr, unsigned char *RESTRICT output_ptr,
+    unsigned int src_pixels_per_line, unsigned int output_width,
+    unsigned int output_height, int xoffset, int yoffset,
+    unsigned char *RESTRICT dst_ptr, int pitch) {
+  unsigned int i, j;
+  int Temp1, Temp2, Temp3, Temp4;
 
-                    /* odd 1. pixel */
-                    "ulw              %[tn1],      4(%[src_ptr])                    \n\t"
-                    "balign           %[tp2],      %[tp1],          3               \n\t"
-                    "mtlo             %[vector4a], $ac3                             \n\t"
-                    "mthi             $0,          $ac3                             \n\t"
-                    "preceu.ph.qbr    %[n1],       %[tp2]                           \n\t"
-                    "preceu.ph.qbl    %[n2],       %[tp2]                           \n\t"
-                    "preceu.ph.qbr    %[n3],       %[tn1]                           \n\t"
-                    "extr.w           %[Temp3],    $ac2,            7               \n\t"
-                    "dpa.w.ph         $ac3,        %[n1],           %[vector1b]     \n\t"
-                    "dpa.w.ph         $ac3,        %[n2],           %[vector2b]     \n\t"
+  unsigned int vector4a;
+  int vector1b, vector2b;
+  unsigned int tp1, tp2, tp3, tn1;
+  unsigned int p1, p2, p3;
+  unsigned int n1, n2, n3;
+  unsigned char *cm = ff_cropTbl + CROP_WIDTH;
 
-                    /* odd 2. pixel */
-                    "mtlo             %[vector4a], $ac2                             \n\t"
-                    "mthi             $0,          $ac2                             \n\t"
-                    "extr.w           %[Temp2],    $ac3,            7               \n\t"
-                    "dpa.w.ph         $ac2,        %[n2],           %[vector1b]     \n\t"
-                    "dpa.w.ph         $ac2,        %[n3],           %[vector2b]     \n\t"
-                    "extr.w           %[Temp4],    $ac2,            7               \n\t"
+  vector4a = 64;
 
-                    /* clamp and store results */
-                    "lbux             %[tp1],      %[Temp1](%[cm])                  \n\t"
-                    "lbux             %[tn1],      %[Temp2](%[cm])                  \n\t"
-                    "lbux             %[tp2],      %[Temp3](%[cm])                  \n\t"
-                    "sb               %[tp1],      0(%[dst_ptr])                    \n\t"
-                    "sb               %[tn1],      1(%[dst_ptr])                    \n\t"
-                    "lbux             %[n2],       %[Temp4](%[cm])                  \n\t"
-                    "sb               %[tp2],      2(%[dst_ptr])                    \n\t"
-                    "sb               %[n2],       3(%[dst_ptr])                    \n\t"
+  vector1b = sub_pel_filters_inv_tap_4[xoffset][0];
+  vector2b = sub_pel_filters_inv_tap_4[xoffset][1];
 
-                    : [tp1] "=&r" (tp1), [tp2] "=&r" (tp2), [tp3] "=&r" (tp3),
-                      [tn1] "=&r" (tn1), [p1] "=&r" (p1), [p2] "=&r" (p2),
-                      [n1] "=&r" (n1), [n2] "=&r" (n2), [n3] "=&r" (n3),
-                      [Temp1] "=&r" (Temp1), [Temp2] "=&r" (Temp2), [p3] "=&r" (p3),
-                      [Temp3] "=&r" (Temp3), [Temp4] "=&r" (Temp4)
-                    : [vector1b] "r" (vector1b), [vector2b] "r" (vector2b),
-                      [vector4a] "r" (vector4a), [cm] "r" (cm), [dst_ptr] "r" (dst_ptr),
-                      [src_ptr] "r" (src_ptr)
-                );
+  /* if (yoffset == 0) don't need temp buffer, data will be stored in dst_ptr */
+  if (yoffset == 0) {
+    output_height -= 5;
+    src_ptr += (src_pixels_per_line + src_pixels_per_line);
 
-                src_ptr += 4;
-            }
+    for (i = output_height; i--;) {
+      __asm__ __volatile__("ulw     %[tp3],   -1(%[src_ptr])               \n\t"
+                           : [tp3] "=&r"(tp3)
+                           : [src_ptr] "r"(src_ptr));
 
-            /* Next row... */
-            src_ptr += src_pixels_per_line - 16;
-            dst_ptr += pitch;
-        }
-    }
-    else
-    {
-        for (i = output_height; i--;)
-        {
-            /* processing 4 adjacent pixels */
-            for (j = 0; j < 16; j += 4)
-            {
-                /* apply filter with vectors pairs */
-                __asm__ __volatile__ (
-                    "ulw              %[tp1],      -1(%[src_ptr])                   \n\t"
-                    "ulw              %[tp2],      3(%[src_ptr])                    \n\t"
+      /* processing 4 adjacent pixels */
+      for (j = 0; j < 16; j += 4) {
+        /* apply filter with vectors pairs */
+        __asm__ __volatile__(
+            "ulw              %[tp2],      3(%[src_ptr])                    "
+            "\n\t"
+            "move             %[tp1],      %[tp3]                           "
+            "\n\t"
 
-                    /* even 1. pixel */
-                    "mtlo             %[vector4a], $ac3                             \n\t"
-                    "mthi             $0,          $ac3                             \n\t"
-                    "preceu.ph.qbr    %[p1],       %[tp1]                           \n\t"
-                    "preceu.ph.qbl    %[p2],       %[tp1]                           \n\t"
-                    "preceu.ph.qbr    %[p3],       %[tp2]                           \n\t"
-                    "dpa.w.ph         $ac3,        %[p1],           %[vector1b]     \n\t"
-                    "dpa.w.ph         $ac3,        %[p2],           %[vector2b]     \n\t"
+            /* even 1. pixel */
+            "mtlo             %[vector4a], $ac3                             "
+            "\n\t"
+            "mthi             $0,          $ac3                             "
+            "\n\t"
+            "move             %[tp3],      %[tp2]                           "
+            "\n\t"
+            "preceu.ph.qbr    %[p1],       %[tp1]                           "
+            "\n\t"
+            "preceu.ph.qbl    %[p2],       %[tp1]                           "
+            "\n\t"
+            "preceu.ph.qbr    %[p3],       %[tp2]                           "
+            "\n\t"
+            "dpa.w.ph         $ac3,        %[p1],           %[vector1b]     "
+            "\n\t"
+            "dpa.w.ph         $ac3,        %[p2],           %[vector2b]     "
+            "\n\t"
 
-                    /* even 2. pixel */
-                    "mtlo             %[vector4a], $ac2                             \n\t"
-                    "mthi             $0,          $ac2                             \n\t"
-                    "dpa.w.ph         $ac2,        %[p2],           %[vector1b]     \n\t"
-                    "dpa.w.ph         $ac2,        %[p3],           %[vector2b]     \n\t"
-                    "extr.w           %[Temp1],    $ac3,            7               \n\t"
+            /* even 2. pixel */
+            "mtlo             %[vector4a], $ac2                             "
+            "\n\t"
+            "mthi             $0,          $ac2                             "
+            "\n\t"
+            "dpa.w.ph         $ac2,        %[p2],           %[vector1b]     "
+            "\n\t"
+            "dpa.w.ph         $ac2,        %[p3],           %[vector2b]     "
+            "\n\t"
+            "extr.w           %[Temp1],    $ac3,            7               "
+            "\n\t"
 
-                    /* odd 1. pixel */
-                    "ulw              %[tn1],      4(%[src_ptr])                    \n\t"
-                    "balign           %[tp2],      %[tp1],          3               \n\t"
-                    "mtlo             %[vector4a], $ac3                             \n\t"
-                    "mthi             $0,          $ac3                             \n\t"
-                    "preceu.ph.qbr    %[n1],       %[tp2]                           \n\t"
-                    "preceu.ph.qbl    %[n2],       %[tp2]                           \n\t"
-                    "preceu.ph.qbr    %[n3],       %[tn1]                           \n\t"
-                    "extr.w           %[Temp3],    $ac2,            7               \n\t"
-                    "dpa.w.ph         $ac3,        %[n1],           %[vector1b]     \n\t"
-                    "dpa.w.ph         $ac3,        %[n2],           %[vector2b]     \n\t"
+            /* odd 1. pixel */
+            "ulw              %[tn1],      4(%[src_ptr])                    "
+            "\n\t"
+            "balign           %[tp2],      %[tp1],          3               "
+            "\n\t"
+            "mtlo             %[vector4a], $ac3                             "
+            "\n\t"
+            "mthi             $0,          $ac3                             "
+            "\n\t"
+            "preceu.ph.qbr    %[n1],       %[tp2]                           "
+            "\n\t"
+            "preceu.ph.qbl    %[n2],       %[tp2]                           "
+            "\n\t"
+            "preceu.ph.qbr    %[n3],       %[tn1]                           "
+            "\n\t"
+            "extr.w           %[Temp3],    $ac2,            7               "
+            "\n\t"
+            "dpa.w.ph         $ac3,        %[n1],           %[vector1b]     "
+            "\n\t"
+            "dpa.w.ph         $ac3,        %[n2],           %[vector2b]     "
+            "\n\t"
 
-                    /* odd 2. pixel */
-                    "mtlo             %[vector4a], $ac2                             \n\t"
-                    "mthi             $0,          $ac2                             \n\t"
-                    "extr.w           %[Temp2],    $ac3,            7               \n\t"
-                    "dpa.w.ph         $ac2,        %[n2],           %[vector1b]     \n\t"
-                    "dpa.w.ph         $ac2,        %[n3],           %[vector2b]     \n\t"
-                    "extr.w           %[Temp4],    $ac2,            7               \n\t"
+            /* odd 2. pixel */
+            "mtlo             %[vector4a], $ac2                             "
+            "\n\t"
+            "mthi             $0,          $ac2                             "
+            "\n\t"
+            "extr.w           %[Temp2],    $ac3,            7               "
+            "\n\t"
+            "dpa.w.ph         $ac2,        %[n2],           %[vector1b]     "
+            "\n\t"
+            "dpa.w.ph         $ac2,        %[n3],           %[vector2b]     "
+            "\n\t"
+            "extr.w           %[Temp4],    $ac2,            7               "
+            "\n\t"
 
-                    /* clamp and store results */
-                    "lbux             %[tp1],      %[Temp1](%[cm])                  \n\t"
-                    "lbux             %[tn1],      %[Temp2](%[cm])                  \n\t"
-                    "lbux             %[tp2],      %[Temp3](%[cm])                  \n\t"
-                    "sb               %[tp1],      0(%[output_ptr])                 \n\t"
-                    "sb               %[tn1],      1(%[output_ptr])                 \n\t"
-                    "lbux             %[n2],       %[Temp4](%[cm])                  \n\t"
-                    "sb               %[tp2],      2(%[output_ptr])                 \n\t"
-                    "sb               %[n2],       3(%[output_ptr])                 \n\t"
+            /* clamp and store results */
+            "lbux             %[tp1],      %[Temp1](%[cm])                  "
+            "\n\t"
+            "lbux             %[tn1],      %[Temp2](%[cm])                  "
+            "\n\t"
+            "lbux             %[tp2],      %[Temp3](%[cm])                  "
+            "\n\t"
+            "sb               %[tp1],      0(%[dst_ptr])                    "
+            "\n\t"
+            "sb               %[tn1],      1(%[dst_ptr])                    "
+            "\n\t"
+            "lbux             %[n2],       %[Temp4](%[cm])                  "
+            "\n\t"
+            "sb               %[tp2],      2(%[dst_ptr])                    "
+            "\n\t"
+            "sb               %[n2],       3(%[dst_ptr])                    "
+            "\n\t"
 
-                    : [tp1] "=&r" (tp1), [tp2] "=&r" (tp2), [tn1] "=&r" (tn1),
-                      [p1] "=&r" (p1), [p2] "=&r" (p2), [p3] "=&r" (p3),
-                      [n1] "=&r" (n1), [n2] "=&r" (n2), [n3] "=&r" (n3),
-                      [Temp1] "=&r" (Temp1), [Temp2] "=&r" (Temp2),
-                      [Temp3] "=&r" (Temp3), [Temp4] "=&r" (Temp4)
-                    : [vector1b] "r" (vector1b), [vector2b] "r" (vector2b),
-                      [vector4a] "r" (vector4a), [cm] "r" (cm),
-                      [output_ptr] "r" (output_ptr), [src_ptr] "r" (src_ptr)
-                );
+            : [tp1] "=&r"(tp1), [tp2] "=&r"(tp2), [tp3] "=&r"(tp3),
+              [tn1] "=&r"(tn1), [p1] "=&r"(p1), [p2] "=&r"(p2), [n1] "=&r"(n1),
+              [n2] "=&r"(n2), [n3] "=&r"(n3), [Temp1] "=&r"(Temp1),
+              [Temp2] "=&r"(Temp2), [p3] "=&r"(p3), [Temp3] "=&r"(Temp3),
+              [Temp4] "=&r"(Temp4)
+            : [vector1b] "r"(vector1b), [vector2b] "r"(vector2b),
+              [vector4a] "r"(vector4a), [cm] "r"(cm), [dst_ptr] "r"(dst_ptr),
+              [src_ptr] "r"(src_ptr));
+
+        src_ptr += 4;
+      }
+
+      /* Next row... */
+      src_ptr += src_pixels_per_line - 16;
+      dst_ptr += pitch;
+    }
+  } else {
+    for (i = output_height; i--;) {
+      /* processing 4 adjacent pixels */
+      for (j = 0; j < 16; j += 4) {
+        /* apply filter with vectors pairs */
+        __asm__ __volatile__(
+            "ulw              %[tp1],      -1(%[src_ptr])                   "
+            "\n\t"
+            "ulw              %[tp2],      3(%[src_ptr])                    "
+            "\n\t"
+
+            /* even 1. pixel */
+            "mtlo             %[vector4a], $ac3                             "
+            "\n\t"
+            "mthi             $0,          $ac3                             "
+            "\n\t"
+            "preceu.ph.qbr    %[p1],       %[tp1]                           "
+            "\n\t"
+            "preceu.ph.qbl    %[p2],       %[tp1]                           "
+            "\n\t"
+            "preceu.ph.qbr    %[p3],       %[tp2]                           "
+            "\n\t"
+            "dpa.w.ph         $ac3,        %[p1],           %[vector1b]     "
+            "\n\t"
+            "dpa.w.ph         $ac3,        %[p2],           %[vector2b]     "
+            "\n\t"
+
+            /* even 2. pixel */
+            "mtlo             %[vector4a], $ac2                             "
+            "\n\t"
+            "mthi             $0,          $ac2                             "
+            "\n\t"
+            "dpa.w.ph         $ac2,        %[p2],           %[vector1b]     "
+            "\n\t"
+            "dpa.w.ph         $ac2,        %[p3],           %[vector2b]     "
+            "\n\t"
+            "extr.w           %[Temp1],    $ac3,            7               "
+            "\n\t"
+
+            /* odd 1. pixel */
+            "ulw              %[tn1],      4(%[src_ptr])                    "
+            "\n\t"
+            "balign           %[tp2],      %[tp1],          3               "
+            "\n\t"
+            "mtlo             %[vector4a], $ac3                             "
+            "\n\t"
+            "mthi             $0,          $ac3                             "
+            "\n\t"
+            "preceu.ph.qbr    %[n1],       %[tp2]                           "
+            "\n\t"
+            "preceu.ph.qbl    %[n2],       %[tp2]                           "
+            "\n\t"
+            "preceu.ph.qbr    %[n3],       %[tn1]                           "
+            "\n\t"
+            "extr.w           %[Temp3],    $ac2,            7               "
+            "\n\t"
+            "dpa.w.ph         $ac3,        %[n1],           %[vector1b]     "
+            "\n\t"
+            "dpa.w.ph         $ac3,        %[n2],           %[vector2b]     "
+            "\n\t"
+
+            /* odd 2. pixel */
+            "mtlo             %[vector4a], $ac2                             "
+            "\n\t"
+            "mthi             $0,          $ac2                             "
+            "\n\t"
+            "extr.w           %[Temp2],    $ac3,            7               "
+            "\n\t"
+            "dpa.w.ph         $ac2,        %[n2],           %[vector1b]     "
+            "\n\t"
+            "dpa.w.ph         $ac2,        %[n3],           %[vector2b]     "
+            "\n\t"
+            "extr.w           %[Temp4],    $ac2,            7               "
+            "\n\t"
 
-                src_ptr += 4;
-            }
+            /* clamp and store results */
+            "lbux             %[tp1],      %[Temp1](%[cm])                  "
+            "\n\t"
+            "lbux             %[tn1],      %[Temp2](%[cm])                  "
+            "\n\t"
+            "lbux             %[tp2],      %[Temp3](%[cm])                  "
+            "\n\t"
+            "sb               %[tp1],      0(%[output_ptr])                 "
+            "\n\t"
+            "sb               %[tn1],      1(%[output_ptr])                 "
+            "\n\t"
+            "lbux             %[n2],       %[Temp4](%[cm])                  "
+            "\n\t"
+            "sb               %[tp2],      2(%[output_ptr])                 "
+            "\n\t"
+            "sb               %[n2],       3(%[output_ptr])                 "
+            "\n\t"
 
-            /* next row... */
-            src_ptr += src_pixels_per_line;
-            output_ptr += output_width;
-        }
+            : [tp1] "=&r"(tp1), [tp2] "=&r"(tp2), [tn1] "=&r"(tn1),
+              [p1] "=&r"(p1), [p2] "=&r"(p2), [p3] "=&r"(p3), [n1] "=&r"(n1),
+              [n2] "=&r"(n2), [n3] "=&r"(n3), [Temp1] "=&r"(Temp1),
+              [Temp2] "=&r"(Temp2), [Temp3] "=&r"(Temp3), [Temp4] "=&r"(Temp4)
+            : [vector1b] "r"(vector1b), [vector2b] "r"(vector2b),
+              [vector4a] "r"(vector4a), [cm] "r"(cm),
+              [output_ptr] "r"(output_ptr), [src_ptr] "r"(src_ptr));
+
+        src_ptr += 4;
+      }
+
+      /* next row... */
+      src_ptr += src_pixels_per_line;
+      output_ptr += output_width;
     }
+  }
 }
 
+void vp8_filter_block2d_second_pass4(unsigned char *RESTRICT src_ptr,
+                                     unsigned char *RESTRICT output_ptr,
+                                     int output_pitch, int yoffset) {
+  unsigned int i;
 
-void vp8_filter_block2d_second_pass4
-(
-    unsigned char *RESTRICT src_ptr,
-    unsigned char *RESTRICT output_ptr,
-    int output_pitch,
-    int yoffset
-)
-{
-    unsigned int i;
+  int Temp1, Temp2, Temp3, Temp4;
+  unsigned int vector1b, vector2b, vector3b, vector4a;
 
-    int Temp1, Temp2, Temp3, Temp4;
-    unsigned int vector1b, vector2b, vector3b, vector4a;
+  unsigned char src_ptr_l2;
+  unsigned char src_ptr_l1;
+  unsigned char src_ptr_0;
+  unsigned char src_ptr_r1;
+  unsigned char src_ptr_r2;
+  unsigned char src_ptr_r3;
 
-    unsigned char src_ptr_l2;
-    unsigned char src_ptr_l1;
-    unsigned char src_ptr_0;
-    unsigned char src_ptr_r1;
-    unsigned char src_ptr_r2;
-    unsigned char src_ptr_r3;
+  unsigned char *cm = ff_cropTbl + CROP_WIDTH;
 
-    unsigned char *cm = ff_cropTbl + CROP_WIDTH;
+  vector4a = 64;
 
-    vector4a = 64;
+  /* load filter coefficients */
+  vector1b = sub_pel_filterss[yoffset][0];
+  vector2b = sub_pel_filterss[yoffset][2];
+  vector3b = sub_pel_filterss[yoffset][1];
 
-    /* load filter coefficients */
-    vector1b = sub_pel_filterss[yoffset][0];
-    vector2b = sub_pel_filterss[yoffset][2];
-    vector3b = sub_pel_filterss[yoffset][1];
+  if (vector1b) {
+    /* 6 tap filter */
 
-    if (vector1b)
-    {
-        /* 6 tap filter */
+    for (i = 2; i--;) {
+      /* prefetch src_ptr data to cache memory */
+      prefetch_load(src_ptr);
 
-        for (i = 2; i--;)
-        {
-            /* prefetch src_ptr data to cache memory */
-            prefetch_load(src_ptr);
+      /* do not allow compiler to reorder instructions */
+      __asm__ __volatile__(
+          ".set noreorder                                                 \n\t"
+          :
+          :);
 
-            /* do not allow compiler to reorder instructions */
-            __asm__ __volatile__ (
-                ".set noreorder                                                 \n\t"
-                :
-                :
-            );
+      /* apply filter with vectors pairs */
+      __asm__ __volatile__(
+          "lbu            %[src_ptr_l2],  -8(%[src_ptr])                  \n\t"
+          "lbu            %[src_ptr_l1],  -4(%[src_ptr])                  \n\t"
+          "lbu            %[src_ptr_0],   0(%[src_ptr])                   \n\t"
+          "lbu            %[src_ptr_r1],  4(%[src_ptr])                   \n\t"
+          "lbu            %[src_ptr_r2],  8(%[src_ptr])                   \n\t"
+          "lbu            %[src_ptr_r3],  12(%[src_ptr])                  \n\t"
+          "mtlo           %[vector4a],    $ac2                            \n\t"
 
-            /* apply filter with vectors pairs */
-            __asm__ __volatile__ (
-                "lbu            %[src_ptr_l2],  -8(%[src_ptr])                  \n\t"
-                "lbu            %[src_ptr_l1],  -4(%[src_ptr])                  \n\t"
-                "lbu            %[src_ptr_0],   0(%[src_ptr])                   \n\t"
-                "lbu            %[src_ptr_r1],  4(%[src_ptr])                   \n\t"
-                "lbu            %[src_ptr_r2],  8(%[src_ptr])                   \n\t"
-                "lbu            %[src_ptr_r3],  12(%[src_ptr])                  \n\t"
-                "mtlo           %[vector4a],    $ac2                            \n\t"
+          "append         %[src_ptr_l2],  %[src_ptr_r3],  8               \n\t"
+          "append         %[src_ptr_0],   %[src_ptr_r1],  8               \n\t"
+          "append         %[src_ptr_l1],  %[src_ptr_r2],  8               \n\t"
+          "dpau.h.qbr     $ac2,           %[src_ptr_l2],  %[vector1b]     \n\t"
+          "dpau.h.qbr     $ac2,           %[src_ptr_0],   %[vector2b]     \n\t"
+          "dpsu.h.qbr     $ac2,           %[src_ptr_l1],  %[vector3b]     \n\t"
 
-                "append         %[src_ptr_l2],  %[src_ptr_r3],  8               \n\t"
-                "append         %[src_ptr_0],   %[src_ptr_r1],  8               \n\t"
-                "append         %[src_ptr_l1],  %[src_ptr_r2],  8               \n\t"
-                "dpau.h.qbr     $ac2,           %[src_ptr_l2],  %[vector1b]     \n\t"
-                "dpau.h.qbr     $ac2,           %[src_ptr_0],   %[vector2b]     \n\t"
-                "dpsu.h.qbr     $ac2,           %[src_ptr_l1],  %[vector3b]     \n\t"
+          "lbu            %[src_ptr_l2],  -7(%[src_ptr])                  \n\t"
+          "lbu            %[src_ptr_l1],  -3(%[src_ptr])                  \n\t"
+          "lbu            %[src_ptr_0],   1(%[src_ptr])                   \n\t"
+          "lbu            %[src_ptr_r1],  5(%[src_ptr])                   \n\t"
+          "lbu            %[src_ptr_r2],  9(%[src_ptr])                   \n\t"
+          "lbu            %[src_ptr_r3],  13(%[src_ptr])                  \n\t"
+          "mtlo           %[vector4a],    $ac3                            \n\t"
+          "extp           %[Temp1],       $ac2,           9               \n\t"
 
-                "lbu            %[src_ptr_l2],  -7(%[src_ptr])                  \n\t"
-                "lbu            %[src_ptr_l1],  -3(%[src_ptr])                  \n\t"
-                "lbu            %[src_ptr_0],   1(%[src_ptr])                   \n\t"
-                "lbu            %[src_ptr_r1],  5(%[src_ptr])                   \n\t"
-                "lbu            %[src_ptr_r2],  9(%[src_ptr])                   \n\t"
-                "lbu            %[src_ptr_r3],  13(%[src_ptr])                  \n\t"
-                "mtlo           %[vector4a],    $ac3                            \n\t"
-                "extp           %[Temp1],       $ac2,           9               \n\t"
+          "append         %[src_ptr_l2],  %[src_ptr_r3],  8               \n\t"
+          "append         %[src_ptr_0],   %[src_ptr_r1],  8               \n\t"
+          "append         %[src_ptr_l1],  %[src_ptr_r2],  8               \n\t"
+          "dpau.h.qbr     $ac3,           %[src_ptr_l2],  %[vector1b]     \n\t"
+          "dpau.h.qbr     $ac3,           %[src_ptr_0],   %[vector2b]     \n\t"
+          "dpsu.h.qbr     $ac3,           %[src_ptr_l1],  %[vector3b]     \n\t"
 
-                "append         %[src_ptr_l2],  %[src_ptr_r3],  8               \n\t"
-                "append         %[src_ptr_0],   %[src_ptr_r1],  8               \n\t"
-                "append         %[src_ptr_l1],  %[src_ptr_r2],  8               \n\t"
-                "dpau.h.qbr     $ac3,           %[src_ptr_l2],  %[vector1b]     \n\t"
-                "dpau.h.qbr     $ac3,           %[src_ptr_0],   %[vector2b]     \n\t"
-                "dpsu.h.qbr     $ac3,           %[src_ptr_l1],  %[vector3b]     \n\t"
+          "lbu            %[src_ptr_l2],  -6(%[src_ptr])                  \n\t"
+          "lbu            %[src_ptr_l1],  -2(%[src_ptr])                  \n\t"
+          "lbu            %[src_ptr_0],   2(%[src_ptr])                   \n\t"
+          "lbu            %[src_ptr_r1],  6(%[src_ptr])                   \n\t"
+          "lbu            %[src_ptr_r2],  10(%[src_ptr])                  \n\t"
+          "lbu            %[src_ptr_r3],  14(%[src_ptr])                  \n\t"
+          "mtlo           %[vector4a],    $ac0                            \n\t"
+          "extp           %[Temp2],       $ac3,           9               \n\t"
 
-                "lbu            %[src_ptr_l2],  -6(%[src_ptr])                  \n\t"
-                "lbu            %[src_ptr_l1],  -2(%[src_ptr])                  \n\t"
-                "lbu            %[src_ptr_0],   2(%[src_ptr])                   \n\t"
-                "lbu            %[src_ptr_r1],  6(%[src_ptr])                   \n\t"
-                "lbu            %[src_ptr_r2],  10(%[src_ptr])                  \n\t"
-                "lbu            %[src_ptr_r3],  14(%[src_ptr])                  \n\t"
-                "mtlo           %[vector4a],    $ac0                            \n\t"
-                "extp           %[Temp2],       $ac3,           9               \n\t"
+          "append         %[src_ptr_l2],  %[src_ptr_r3],  8               \n\t"
+          "append         %[src_ptr_0],   %[src_ptr_r1],  8               \n\t"
+          "append         %[src_ptr_l1],  %[src_ptr_r2],  8               \n\t"
+          "dpau.h.qbr     $ac0,           %[src_ptr_l2],  %[vector1b]     \n\t"
+          "dpau.h.qbr     $ac0,           %[src_ptr_0],   %[vector2b]     \n\t"
+          "dpsu.h.qbr     $ac0,           %[src_ptr_l1],  %[vector3b]     \n\t"
 
-                "append         %[src_ptr_l2],  %[src_ptr_r3],  8               \n\t"
-                "append         %[src_ptr_0],   %[src_ptr_r1],  8               \n\t"
-                "append         %[src_ptr_l1],  %[src_ptr_r2],  8               \n\t"
-                "dpau.h.qbr     $ac0,           %[src_ptr_l2],  %[vector1b]     \n\t"
-                "dpau.h.qbr     $ac0,           %[src_ptr_0],   %[vector2b]     \n\t"
-                "dpsu.h.qbr     $ac0,           %[src_ptr_l1],  %[vector3b]     \n\t"
+          "lbu            %[src_ptr_l2],  -5(%[src_ptr])                  \n\t"
+          "lbu            %[src_ptr_l1],  -1(%[src_ptr])                  \n\t"
+          "lbu            %[src_ptr_0],   3(%[src_ptr])                   \n\t"
+          "lbu            %[src_ptr_r1],  7(%[src_ptr])                   \n\t"
+          "lbu            %[src_ptr_r2],  11(%[src_ptr])                  \n\t"
+          "lbu            %[src_ptr_r3],  15(%[src_ptr])                  \n\t"
+          "mtlo           %[vector4a],    $ac1                            \n\t"
+          "extp           %[Temp3],       $ac0,           9               \n\t"
 
-                "lbu            %[src_ptr_l2],  -5(%[src_ptr])                  \n\t"
-                "lbu            %[src_ptr_l1],  -1(%[src_ptr])                  \n\t"
-                "lbu            %[src_ptr_0],   3(%[src_ptr])                   \n\t"
-                "lbu            %[src_ptr_r1],  7(%[src_ptr])                   \n\t"
-                "lbu            %[src_ptr_r2],  11(%[src_ptr])                  \n\t"
-                "lbu            %[src_ptr_r3],  15(%[src_ptr])                  \n\t"
-                "mtlo           %[vector4a],    $ac1                            \n\t"
-                "extp           %[Temp3],       $ac0,           9               \n\t"
+          "append         %[src_ptr_l2],  %[src_ptr_r3],  8               \n\t"
+          "append         %[src_ptr_0],   %[src_ptr_r1],  8               \n\t"
+          "append         %[src_ptr_l1],  %[src_ptr_r2],  8               \n\t"
+          "dpau.h.qbr     $ac1,           %[src_ptr_l2],  %[vector1b]     \n\t"
+          "dpau.h.qbr     $ac1,           %[src_ptr_0],   %[vector2b]     \n\t"
+          "dpsu.h.qbr     $ac1,           %[src_ptr_l1],  %[vector3b]     \n\t"
+          "extp           %[Temp4],       $ac1,           9               \n\t"
 
-                "append         %[src_ptr_l2],  %[src_ptr_r3],  8               \n\t"
-                "append         %[src_ptr_0],   %[src_ptr_r1],  8               \n\t"
-                "append         %[src_ptr_l1],  %[src_ptr_r2],  8               \n\t"
-                "dpau.h.qbr     $ac1,           %[src_ptr_l2],  %[vector1b]     \n\t"
-                "dpau.h.qbr     $ac1,           %[src_ptr_0],   %[vector2b]     \n\t"
-                "dpsu.h.qbr     $ac1,           %[src_ptr_l1],  %[vector3b]     \n\t"
-                "extp           %[Temp4],       $ac1,           9               \n\t"
+          : [Temp1] "=&r"(Temp1), [Temp2] "=&r"(Temp2), [Temp3] "=&r"(Temp3),
+            [Temp4] "=r"(Temp4), [src_ptr_l1] "=&r"(src_ptr_l1),
+            [src_ptr_0] "=&r"(src_ptr_0), [src_ptr_r1] "=&r"(src_ptr_r1),
+            [src_ptr_r2] "=&r"(src_ptr_r2), [src_ptr_l2] "=&r"(src_ptr_l2),
+            [src_ptr_r3] "=&r"(src_ptr_r3)
+          : [vector1b] "r"(vector1b), [vector2b] "r"(vector2b),
+            [vector3b] "r"(vector3b), [vector4a] "r"(vector4a),
+            [src_ptr] "r"(src_ptr));
 
-                : [Temp1] "=&r" (Temp1), [Temp2] "=&r" (Temp2),
-                  [Temp3] "=&r" (Temp3), [Temp4] "=r" (Temp4),
-                  [src_ptr_l1] "=&r" (src_ptr_l1), [src_ptr_0] "=&r" (src_ptr_0),
-                  [src_ptr_r1] "=&r" (src_ptr_r1), [src_ptr_r2] "=&r" (src_ptr_r2),
-                  [src_ptr_l2] "=&r" (src_ptr_l2), [src_ptr_r3] "=&r" (src_ptr_r3)
-                : [vector1b] "r" (vector1b), [vector2b] "r" (vector2b),
-                  [vector3b] "r" (vector3b), [vector4a] "r" (vector4a),
-                  [src_ptr] "r" (src_ptr)
-            );
+      /* clamp and store results */
+      output_ptr[0] = cm[Temp1];
+      output_ptr[1] = cm[Temp2];
+      output_ptr[2] = cm[Temp3];
+      output_ptr[3] = cm[Temp4];
 
-            /* clamp and store results */
-            output_ptr[0] = cm[Temp1];
-            output_ptr[1] = cm[Temp2];
-            output_ptr[2] = cm[Temp3];
-            output_ptr[3] = cm[Temp4];
+      output_ptr += output_pitch;
 
-            output_ptr += output_pitch;
+      /* apply filter with vectors pairs */
+      __asm__ __volatile__(
+          "lbu            %[src_ptr_l2],  -4(%[src_ptr])                  \n\t"
+          "lbu            %[src_ptr_l1],  0(%[src_ptr])                   \n\t"
+          "lbu            %[src_ptr_0],   4(%[src_ptr])                   \n\t"
+          "lbu            %[src_ptr_r1],  8(%[src_ptr])                   \n\t"
+          "lbu            %[src_ptr_r2],  12(%[src_ptr])                  \n\t"
+          "lbu            %[src_ptr_r3],  16(%[src_ptr])                  \n\t"
+          "mtlo           %[vector4a],    $ac2                            \n\t"
+          "append         %[src_ptr_l2],  %[src_ptr_r3],  8               \n\t"
+          "append         %[src_ptr_0],   %[src_ptr_r1],  8               \n\t"
+          "append         %[src_ptr_l1],  %[src_ptr_r2],  8               \n\t"
+          "dpau.h.qbr     $ac2,           %[src_ptr_l2],  %[vector1b]     \n\t"
+          "dpau.h.qbr     $ac2,           %[src_ptr_0],   %[vector2b]     \n\t"
+          "dpsu.h.qbr     $ac2,           %[src_ptr_l1],  %[vector3b]     \n\t"
 
-            /* apply filter with vectors pairs */
-            __asm__ __volatile__ (
-                "lbu            %[src_ptr_l2],  -4(%[src_ptr])                  \n\t"
-                "lbu            %[src_ptr_l1],  0(%[src_ptr])                   \n\t"
-                "lbu            %[src_ptr_0],   4(%[src_ptr])                   \n\t"
-                "lbu            %[src_ptr_r1],  8(%[src_ptr])                   \n\t"
-                "lbu            %[src_ptr_r2],  12(%[src_ptr])                  \n\t"
-                "lbu            %[src_ptr_r3],  16(%[src_ptr])                  \n\t"
-                "mtlo           %[vector4a],    $ac2                            \n\t"
-                "append         %[src_ptr_l2],  %[src_ptr_r3],  8               \n\t"
-                "append         %[src_ptr_0],   %[src_ptr_r1],  8               \n\t"
-                "append         %[src_ptr_l1],  %[src_ptr_r2],  8               \n\t"
-                "dpau.h.qbr     $ac2,           %[src_ptr_l2],  %[vector1b]     \n\t"
-                "dpau.h.qbr     $ac2,           %[src_ptr_0],   %[vector2b]     \n\t"
-                "dpsu.h.qbr     $ac2,           %[src_ptr_l1],  %[vector3b]     \n\t"
+          "lbu            %[src_ptr_l2],  -3(%[src_ptr])                  \n\t"
+          "lbu            %[src_ptr_l1],  1(%[src_ptr])                   \n\t"
+          "lbu            %[src_ptr_0],   5(%[src_ptr])                   \n\t"
+          "lbu            %[src_ptr_r1],  9(%[src_ptr])                   \n\t"
+          "lbu            %[src_ptr_r2],  13(%[src_ptr])                  \n\t"
+          "lbu            %[src_ptr_r3],  17(%[src_ptr])                  \n\t"
+          "mtlo           %[vector4a],    $ac3                            \n\t"
+          "extp           %[Temp1],       $ac2,           9               \n\t"
 
-                "lbu            %[src_ptr_l2],  -3(%[src_ptr])                  \n\t"
-                "lbu            %[src_ptr_l1],  1(%[src_ptr])                   \n\t"
-                "lbu            %[src_ptr_0],   5(%[src_ptr])                   \n\t"
-                "lbu            %[src_ptr_r1],  9(%[src_ptr])                   \n\t"
-                "lbu            %[src_ptr_r2],  13(%[src_ptr])                  \n\t"
-                "lbu            %[src_ptr_r3],  17(%[src_ptr])                  \n\t"
-                "mtlo           %[vector4a],    $ac3                            \n\t"
-                "extp           %[Temp1],       $ac2,           9               \n\t"
+          "append         %[src_ptr_l2],  %[src_ptr_r3],  8               \n\t"
+          "append         %[src_ptr_0],   %[src_ptr_r1],  8               \n\t"
+          "append         %[src_ptr_l1],  %[src_ptr_r2],  8               \n\t"
+          "dpau.h.qbr     $ac3,           %[src_ptr_l2],  %[vector1b]     \n\t"
+          "dpau.h.qbr     $ac3,           %[src_ptr_0],   %[vector2b]     \n\t"
+          "dpsu.h.qbr     $ac3,           %[src_ptr_l1],  %[vector3b]     \n\t"
 
-                "append         %[src_ptr_l2],  %[src_ptr_r3],  8               \n\t"
-                "append         %[src_ptr_0],   %[src_ptr_r1],  8               \n\t"
-                "append         %[src_ptr_l1],  %[src_ptr_r2],  8               \n\t"
-                "dpau.h.qbr     $ac3,           %[src_ptr_l2],  %[vector1b]     \n\t"
-                "dpau.h.qbr     $ac3,           %[src_ptr_0],   %[vector2b]     \n\t"
-                "dpsu.h.qbr     $ac3,           %[src_ptr_l1],  %[vector3b]     \n\t"
+          "lbu            %[src_ptr_l2],  -2(%[src_ptr])                  \n\t"
+          "lbu            %[src_ptr_l1],  2(%[src_ptr])                   \n\t"
+          "lbu            %[src_ptr_0],   6(%[src_ptr])                   \n\t"
+          "lbu            %[src_ptr_r1],  10(%[src_ptr])                  \n\t"
+          "lbu            %[src_ptr_r2],  14(%[src_ptr])                  \n\t"
+          "lbu            %[src_ptr_r3],  18(%[src_ptr])                  \n\t"
+          "mtlo           %[vector4a],    $ac0                            \n\t"
+          "extp           %[Temp2],       $ac3,           9               \n\t"
 
-                "lbu            %[src_ptr_l2],  -2(%[src_ptr])                  \n\t"
-                "lbu            %[src_ptr_l1],  2(%[src_ptr])                   \n\t"
-                "lbu            %[src_ptr_0],   6(%[src_ptr])                   \n\t"
-                "lbu            %[src_ptr_r1],  10(%[src_ptr])                  \n\t"
-                "lbu            %[src_ptr_r2],  14(%[src_ptr])                  \n\t"
-                "lbu            %[src_ptr_r3],  18(%[src_ptr])                  \n\t"
-                "mtlo           %[vector4a],    $ac0                            \n\t"
-                "extp           %[Temp2],       $ac3,           9               \n\t"
+          "append         %[src_ptr_l2],  %[src_ptr_r3],  8               \n\t"
+          "append         %[src_ptr_0],   %[src_ptr_r1],  8               \n\t"
+          "append         %[src_ptr_l1],  %[src_ptr_r2],  8               \n\t"
+          "dpau.h.qbr     $ac0,           %[src_ptr_l2],  %[vector1b]     \n\t"
+          "dpau.h.qbr     $ac0,           %[src_ptr_0],   %[vector2b]     \n\t"
+          "dpsu.h.qbr     $ac0,           %[src_ptr_l1],  %[vector3b]     \n\t"
 
-                "append         %[src_ptr_l2],  %[src_ptr_r3],  8               \n\t"
-                "append         %[src_ptr_0],   %[src_ptr_r1],  8               \n\t"
-                "append         %[src_ptr_l1],  %[src_ptr_r2],  8               \n\t"
-                "dpau.h.qbr     $ac0,           %[src_ptr_l2],  %[vector1b]     \n\t"
-                "dpau.h.qbr     $ac0,           %[src_ptr_0],   %[vector2b]     \n\t"
-                "dpsu.h.qbr     $ac0,           %[src_ptr_l1],  %[vector3b]     \n\t"
+          "lbu            %[src_ptr_l2],  -1(%[src_ptr])                  \n\t"
+          "lbu            %[src_ptr_l1],  3(%[src_ptr])                   \n\t"
+          "lbu            %[src_ptr_0],   7(%[src_ptr])                   \n\t"
+          "lbu            %[src_ptr_r1],  11(%[src_ptr])                  \n\t"
+          "lbu            %[src_ptr_r2],  15(%[src_ptr])                  \n\t"
+          "lbu            %[src_ptr_r3],  19(%[src_ptr])                  \n\t"
+          "mtlo           %[vector4a],    $ac1                            \n\t"
+          "extp           %[Temp3],       $ac0,           9               \n\t"
 
-                "lbu            %[src_ptr_l2],  -1(%[src_ptr])                  \n\t"
-                "lbu            %[src_ptr_l1],  3(%[src_ptr])                   \n\t"
-                "lbu            %[src_ptr_0],   7(%[src_ptr])                   \n\t"
-                "lbu            %[src_ptr_r1],  11(%[src_ptr])                  \n\t"
-                "lbu            %[src_ptr_r2],  15(%[src_ptr])                  \n\t"
-                "lbu            %[src_ptr_r3],  19(%[src_ptr])                  \n\t"
-                "mtlo           %[vector4a],    $ac1                            \n\t"
-                "extp           %[Temp3],       $ac0,           9               \n\t"
+          "append         %[src_ptr_l2],  %[src_ptr_r3],  8               \n\t"
+          "append         %[src_ptr_0],   %[src_ptr_r1],  8               \n\t"
+          "append         %[src_ptr_l1],  %[src_ptr_r2],  8               \n\t"
+          "dpau.h.qbr     $ac1,           %[src_ptr_l2],  %[vector1b]     \n\t"
+          "dpau.h.qbr     $ac1,           %[src_ptr_0],   %[vector2b]     \n\t"
+          "dpsu.h.qbr     $ac1,           %[src_ptr_l1],  %[vector3b]     \n\t"
+          "extp           %[Temp4],       $ac1,           9               \n\t"
 
-                "append         %[src_ptr_l2],  %[src_ptr_r3],  8               \n\t"
-                "append         %[src_ptr_0],   %[src_ptr_r1],  8               \n\t"
-                "append         %[src_ptr_l1],  %[src_ptr_r2],  8               \n\t"
-                "dpau.h.qbr     $ac1,           %[src_ptr_l2],  %[vector1b]     \n\t"
-                "dpau.h.qbr     $ac1,           %[src_ptr_0],   %[vector2b]     \n\t"
-                "dpsu.h.qbr     $ac1,           %[src_ptr_l1],  %[vector3b]     \n\t"
-                "extp           %[Temp4],       $ac1,           9               \n\t"
+          : [Temp1] "=&r"(Temp1), [Temp2] "=&r"(Temp2), [Temp3] "=&r"(Temp3),
+            [Temp4] "=r"(Temp4), [src_ptr_l1] "=&r"(src_ptr_l1),
+            [src_ptr_0] "=&r"(src_ptr_0), [src_ptr_r1] "=&r"(src_ptr_r1),
+            [src_ptr_r2] "=&r"(src_ptr_r2), [src_ptr_l2] "=&r"(src_ptr_l2),
+            [src_ptr_r3] "=&r"(src_ptr_r3)
+          : [vector1b] "r"(vector1b), [vector2b] "r"(vector2b),
+            [vector3b] "r"(vector3b), [vector4a] "r"(vector4a),
+            [src_ptr] "r"(src_ptr));
 
-                : [Temp1] "=&r" (Temp1), [Temp2] "=&r" (Temp2),
-                  [Temp3] "=&r" (Temp3), [Temp4] "=r" (Temp4),
-                  [src_ptr_l1] "=&r" (src_ptr_l1), [src_ptr_0] "=&r" (src_ptr_0),
-                  [src_ptr_r1] "=&r" (src_ptr_r1), [src_ptr_r2] "=&r" (src_ptr_r2),
-                  [src_ptr_l2] "=&r" (src_ptr_l2), [src_ptr_r3] "=&r" (src_ptr_r3)
-                : [vector1b] "r" (vector1b), [vector2b] "r" (vector2b),
-                  [vector3b] "r" (vector3b), [vector4a] "r" (vector4a),
-                  [src_ptr] "r" (src_ptr)
-            );
+      /* clamp and store results */
+      output_ptr[0] = cm[Temp1];
+      output_ptr[1] = cm[Temp2];
+      output_ptr[2] = cm[Temp3];
+      output_ptr[3] = cm[Temp4];
 
-            /* clamp and store results */
-            output_ptr[0] = cm[Temp1];
-            output_ptr[1] = cm[Temp2];
-            output_ptr[2] = cm[Temp3];
-            output_ptr[3] = cm[Temp4];
-
-            src_ptr += 8;
-            output_ptr += output_pitch;
-        }
+      src_ptr += 8;
+      output_ptr += output_pitch;
     }
-    else
-    {
-        /* 4 tap filter */
+  } else {
+    /* 4 tap filter */
 
-        /* prefetch src_ptr data to cache memory */
-        prefetch_load(src_ptr);
+    /* prefetch src_ptr data to cache memory */
+    prefetch_load(src_ptr);
 
-        for (i = 2; i--;)
-        {
-            /* do not allow compiler to reorder instructions */
-            __asm__ __volatile__ (
-                ".set noreorder                                                 \n\t"
-                :
-                :
-            );
+    for (i = 2; i--;) {
+      /* do not allow compiler to reorder instructions */
+      __asm__ __volatile__(
+          ".set noreorder                                                 \n\t"
+          :
+          :);
 
-            /* apply filter with vectors pairs */
-            __asm__ __volatile__ (
-                "lbu            %[src_ptr_l1],  -4(%[src_ptr])                  \n\t"
-                "lbu            %[src_ptr_0],   0(%[src_ptr])                   \n\t"
-                "lbu            %[src_ptr_r1],  4(%[src_ptr])                   \n\t"
-                "lbu            %[src_ptr_r2],  8(%[src_ptr])                   \n\t"
-                "mtlo           %[vector4a],    $ac2                            \n\t"
-                "append         %[src_ptr_0],   %[src_ptr_r1],  8               \n\t"
-                "append         %[src_ptr_l1],  %[src_ptr_r2],  8               \n\t"
-                "dpau.h.qbr     $ac2,           %[src_ptr_0],   %[vector2b]     \n\t"
-                "dpsu.h.qbr     $ac2,           %[src_ptr_l1],  %[vector3b]     \n\t"
+      /* apply filter with vectors pairs */
+      __asm__ __volatile__(
+          "lbu            %[src_ptr_l1],  -4(%[src_ptr])                  \n\t"
+          "lbu            %[src_ptr_0],   0(%[src_ptr])                   \n\t"
+          "lbu            %[src_ptr_r1],  4(%[src_ptr])                   \n\t"
+          "lbu            %[src_ptr_r2],  8(%[src_ptr])                   \n\t"
+          "mtlo           %[vector4a],    $ac2                            \n\t"
+          "append         %[src_ptr_0],   %[src_ptr_r1],  8               \n\t"
+          "append         %[src_ptr_l1],  %[src_ptr_r2],  8               \n\t"
+          "dpau.h.qbr     $ac2,           %[src_ptr_0],   %[vector2b]     \n\t"
+          "dpsu.h.qbr     $ac2,           %[src_ptr_l1],  %[vector3b]     \n\t"
 
-                "lbu            %[src_ptr_l1],  -3(%[src_ptr])                  \n\t"
-                "lbu            %[src_ptr_0],   1(%[src_ptr])                   \n\t"
-                "lbu            %[src_ptr_r1],  5(%[src_ptr])                   \n\t"
-                "lbu            %[src_ptr_r2],  9(%[src_ptr])                   \n\t"
-                "mtlo           %[vector4a],    $ac3                            \n\t"
-                "extp           %[Temp1],       $ac2,           9               \n\t"
+          "lbu            %[src_ptr_l1],  -3(%[src_ptr])                  \n\t"
+          "lbu            %[src_ptr_0],   1(%[src_ptr])                   \n\t"
+          "lbu            %[src_ptr_r1],  5(%[src_ptr])                   \n\t"
+          "lbu            %[src_ptr_r2],  9(%[src_ptr])                   \n\t"
+          "mtlo           %[vector4a],    $ac3                            \n\t"
+          "extp           %[Temp1],       $ac2,           9               \n\t"
 
-                "append         %[src_ptr_0],   %[src_ptr_r1],  8               \n\t"
-                "append         %[src_ptr_l1],  %[src_ptr_r2],  8               \n\t"
-                "dpau.h.qbr     $ac3,           %[src_ptr_0],   %[vector2b]     \n\t"
-                "dpsu.h.qbr     $ac3,           %[src_ptr_l1],  %[vector3b]     \n\t"
+          "append         %[src_ptr_0],   %[src_ptr_r1],  8               \n\t"
+          "append         %[src_ptr_l1],  %[src_ptr_r2],  8               \n\t"
+          "dpau.h.qbr     $ac3,           %[src_ptr_0],   %[vector2b]     \n\t"
+          "dpsu.h.qbr     $ac3,           %[src_ptr_l1],  %[vector3b]     \n\t"
 
-                "lbu            %[src_ptr_l1],  -2(%[src_ptr])                  \n\t"
-                "lbu            %[src_ptr_0],   2(%[src_ptr])                   \n\t"
-                "lbu            %[src_ptr_r1],  6(%[src_ptr])                   \n\t"
-                "lbu            %[src_ptr_r2],  10(%[src_ptr])                  \n\t"
-                "mtlo           %[vector4a],    $ac0                            \n\t"
-                "extp           %[Temp2],       $ac3,           9               \n\t"
+          "lbu            %[src_ptr_l1],  -2(%[src_ptr])                  \n\t"
+          "lbu            %[src_ptr_0],   2(%[src_ptr])                   \n\t"
+          "lbu            %[src_ptr_r1],  6(%[src_ptr])                   \n\t"
+          "lbu            %[src_ptr_r2],  10(%[src_ptr])                  \n\t"
+          "mtlo           %[vector4a],    $ac0                            \n\t"
+          "extp           %[Temp2],       $ac3,           9               \n\t"
 
-                "append         %[src_ptr_0],   %[src_ptr_r1],  8               \n\t"
-                "append         %[src_ptr_l1],  %[src_ptr_r2],  8               \n\t"
-                "dpau.h.qbr     $ac0,           %[src_ptr_0],   %[vector2b]     \n\t"
-                "dpsu.h.qbr     $ac0,           %[src_ptr_l1],  %[vector3b]     \n\t"
+          "append         %[src_ptr_0],   %[src_ptr_r1],  8               \n\t"
+          "append         %[src_ptr_l1],  %[src_ptr_r2],  8               \n\t"
+          "dpau.h.qbr     $ac0,           %[src_ptr_0],   %[vector2b]     \n\t"
+          "dpsu.h.qbr     $ac0,           %[src_ptr_l1],  %[vector3b]     \n\t"
 
-                "lbu            %[src_ptr_l1],  -1(%[src_ptr])                  \n\t"
-                "lbu            %[src_ptr_0],   3(%[src_ptr])                   \n\t"
-                "lbu            %[src_ptr_r1],  7(%[src_ptr])                   \n\t"
-                "lbu            %[src_ptr_r2],  11(%[src_ptr])                  \n\t"
-                "mtlo           %[vector4a],    $ac1                            \n\t"
-                "extp           %[Temp3],       $ac0,           9               \n\t"
-                "append         %[src_ptr_0],   %[src_ptr_r1],  8               \n\t"
-                "append         %[src_ptr_l1],  %[src_ptr_r2],  8               \n\t"
-                "dpau.h.qbr     $ac1,           %[src_ptr_0],   %[vector2b]     \n\t"
-                "dpsu.h.qbr     $ac1,           %[src_ptr_l1],  %[vector3b]     \n\t"
-                "extp           %[Temp4],       $ac1,           9               \n\t"
+          "lbu            %[src_ptr_l1],  -1(%[src_ptr])                  \n\t"
+          "lbu            %[src_ptr_0],   3(%[src_ptr])                   \n\t"
+          "lbu            %[src_ptr_r1],  7(%[src_ptr])                   \n\t"
+          "lbu            %[src_ptr_r2],  11(%[src_ptr])                  \n\t"
+          "mtlo           %[vector4a],    $ac1                            \n\t"
+          "extp           %[Temp3],       $ac0,           9               \n\t"
+          "append         %[src_ptr_0],   %[src_ptr_r1],  8               \n\t"
+          "append         %[src_ptr_l1],  %[src_ptr_r2],  8               \n\t"
+          "dpau.h.qbr     $ac1,           %[src_ptr_0],   %[vector2b]     \n\t"
+          "dpsu.h.qbr     $ac1,           %[src_ptr_l1],  %[vector3b]     \n\t"
+          "extp           %[Temp4],       $ac1,           9               \n\t"
 
-                : [Temp1] "=&r" (Temp1), [Temp2] "=&r" (Temp2),
-                  [Temp3] "=&r" (Temp3), [Temp4] "=r" (Temp4),
-                  [src_ptr_l1] "=&r" (src_ptr_l1), [src_ptr_0] "=&r" (src_ptr_0),
-                  [src_ptr_r1] "=&r" (src_ptr_r1), [src_ptr_r2] "=&r" (src_ptr_r2)
-                : [vector2b] "r" (vector2b), [vector3b] "r" (vector3b),
-                  [vector4a] "r" (vector4a), [src_ptr] "r" (src_ptr)
-            );
+          : [Temp1] "=&r"(Temp1), [Temp2] "=&r"(Temp2), [Temp3] "=&r"(Temp3),
+            [Temp4] "=r"(Temp4), [src_ptr_l1] "=&r"(src_ptr_l1),
+            [src_ptr_0] "=&r"(src_ptr_0), [src_ptr_r1] "=&r"(src_ptr_r1),
+            [src_ptr_r2] "=&r"(src_ptr_r2)
+          : [vector2b] "r"(vector2b), [vector3b] "r"(vector3b),
+            [vector4a] "r"(vector4a), [src_ptr] "r"(src_ptr));
 
-            /* clamp and store results */
-            output_ptr[0] = cm[Temp1];
-            output_ptr[1] = cm[Temp2];
-            output_ptr[2] = cm[Temp3];
-            output_ptr[3] = cm[Temp4];
+      /* clamp and store results */
+      output_ptr[0] = cm[Temp1];
+      output_ptr[1] = cm[Temp2];
+      output_ptr[2] = cm[Temp3];
+      output_ptr[3] = cm[Temp4];
 
-            output_ptr += output_pitch;
+      output_ptr += output_pitch;
 
-            /* apply filter with vectors pairs */
-            __asm__ __volatile__ (
-                "lbu            %[src_ptr_l1],  0(%[src_ptr])                   \n\t"
-                "lbu            %[src_ptr_0],   4(%[src_ptr])                   \n\t"
-                "lbu            %[src_ptr_r1],  8(%[src_ptr])                   \n\t"
-                "lbu            %[src_ptr_r2],  12(%[src_ptr])                  \n\t"
-                "mtlo           %[vector4a],    $ac2                            \n\t"
-                "append         %[src_ptr_0],   %[src_ptr_r1],  8               \n\t"
-                "append         %[src_ptr_l1],  %[src_ptr_r2],  8               \n\t"
-                "dpau.h.qbr     $ac2,           %[src_ptr_0],   %[vector2b]     \n\t"
-                "dpsu.h.qbr     $ac2,           %[src_ptr_l1],  %[vector3b]     \n\t"
+      /* apply filter with vectors pairs */
+      __asm__ __volatile__(
+          "lbu            %[src_ptr_l1],  0(%[src_ptr])                   \n\t"
+          "lbu            %[src_ptr_0],   4(%[src_ptr])                   \n\t"
+          "lbu            %[src_ptr_r1],  8(%[src_ptr])                   \n\t"
+          "lbu            %[src_ptr_r2],  12(%[src_ptr])                  \n\t"
+          "mtlo           %[vector4a],    $ac2                            \n\t"
+          "append         %[src_ptr_0],   %[src_ptr_r1],  8               \n\t"
+          "append         %[src_ptr_l1],  %[src_ptr_r2],  8               \n\t"
+          "dpau.h.qbr     $ac2,           %[src_ptr_0],   %[vector2b]     \n\t"
+          "dpsu.h.qbr     $ac2,           %[src_ptr_l1],  %[vector3b]     \n\t"
 
-                "lbu            %[src_ptr_l1],  1(%[src_ptr])                   \n\t"
-                "lbu            %[src_ptr_0],   5(%[src_ptr])                   \n\t"
-                "lbu            %[src_ptr_r1],  9(%[src_ptr])                   \n\t"
-                "lbu            %[src_ptr_r2],  13(%[src_ptr])                  \n\t"
-                "mtlo           %[vector4a],    $ac3                            \n\t"
-                "extp           %[Temp1],       $ac2,           9               \n\t"
+          "lbu            %[src_ptr_l1],  1(%[src_ptr])                   \n\t"
+          "lbu            %[src_ptr_0],   5(%[src_ptr])                   \n\t"
+          "lbu            %[src_ptr_r1],  9(%[src_ptr])                   \n\t"
+          "lbu            %[src_ptr_r2],  13(%[src_ptr])                  \n\t"
+          "mtlo           %[vector4a],    $ac3                            \n\t"
+          "extp           %[Temp1],       $ac2,           9               \n\t"
 
-                "append         %[src_ptr_0],   %[src_ptr_r1],  8               \n\t"
-                "append         %[src_ptr_l1],  %[src_ptr_r2],  8               \n\t"
-                "dpau.h.qbr     $ac3,           %[src_ptr_0],   %[vector2b]     \n\t"
-                "dpsu.h.qbr     $ac3,           %[src_ptr_l1],  %[vector3b]     \n\t"
+          "append         %[src_ptr_0],   %[src_ptr_r1],  8               \n\t"
+          "append         %[src_ptr_l1],  %[src_ptr_r2],  8               \n\t"
+          "dpau.h.qbr     $ac3,           %[src_ptr_0],   %[vector2b]     \n\t"
+          "dpsu.h.qbr     $ac3,           %[src_ptr_l1],  %[vector3b]     \n\t"
 
-                "lbu            %[src_ptr_l1],  2(%[src_ptr])                   \n\t"
-                "lbu            %[src_ptr_0],   6(%[src_ptr])                   \n\t"
-                "lbu            %[src_ptr_r1],  10(%[src_ptr])                  \n\t"
-                "lbu            %[src_ptr_r2],  14(%[src_ptr])                  \n\t"
-                "mtlo           %[vector4a],    $ac0                            \n\t"
-                "extp           %[Temp2],       $ac3,           9               \n\t"
+          "lbu            %[src_ptr_l1],  2(%[src_ptr])                   \n\t"
+          "lbu            %[src_ptr_0],   6(%[src_ptr])                   \n\t"
+          "lbu            %[src_ptr_r1],  10(%[src_ptr])                  \n\t"
+          "lbu            %[src_ptr_r2],  14(%[src_ptr])                  \n\t"
+          "mtlo           %[vector4a],    $ac0                            \n\t"
+          "extp           %[Temp2],       $ac3,           9               \n\t"
 
-                "append         %[src_ptr_0],   %[src_ptr_r1],  8               \n\t"
-                "append         %[src_ptr_l1],  %[src_ptr_r2],  8               \n\t"
-                "dpau.h.qbr     $ac0,           %[src_ptr_0],   %[vector2b]     \n\t"
-                "dpsu.h.qbr     $ac0,           %[src_ptr_l1],  %[vector3b]     \n\t"
+          "append         %[src_ptr_0],   %[src_ptr_r1],  8               \n\t"
+          "append         %[src_ptr_l1],  %[src_ptr_r2],  8               \n\t"
+          "dpau.h.qbr     $ac0,           %[src_ptr_0],   %[vector2b]     \n\t"
+          "dpsu.h.qbr     $ac0,           %[src_ptr_l1],  %[vector3b]     \n\t"
 
-                "lbu            %[src_ptr_l1],  3(%[src_ptr])                   \n\t"
-                "lbu            %[src_ptr_0],   7(%[src_ptr])                   \n\t"
-                "lbu            %[src_ptr_r1],  11(%[src_ptr])                  \n\t"
-                "lbu            %[src_ptr_r2],  15(%[src_ptr])                  \n\t"
-                "mtlo           %[vector4a],    $ac1                            \n\t"
-                "extp           %[Temp3],       $ac0,           9               \n\t"
-                "append         %[src_ptr_0],   %[src_ptr_r1],  8               \n\t"
-                "append         %[src_ptr_l1],  %[src_ptr_r2],  8               \n\t"
-                "dpau.h.qbr     $ac1,           %[src_ptr_0],   %[vector2b]     \n\t"
-                "dpsu.h.qbr     $ac1,           %[src_ptr_l1],  %[vector3b]     \n\t"
-                "extp           %[Temp4],       $ac1,           9               \n\t"
+          "lbu            %[src_ptr_l1],  3(%[src_ptr])                   \n\t"
+          "lbu            %[src_ptr_0],   7(%[src_ptr])                   \n\t"
+          "lbu            %[src_ptr_r1],  11(%[src_ptr])                  \n\t"
+          "lbu            %[src_ptr_r2],  15(%[src_ptr])                  \n\t"
+          "mtlo           %[vector4a],    $ac1                            \n\t"
+          "extp           %[Temp3],       $ac0,           9               \n\t"
+          "append         %[src_ptr_0],   %[src_ptr_r1],  8               \n\t"
+          "append         %[src_ptr_l1],  %[src_ptr_r2],  8               \n\t"
+          "dpau.h.qbr     $ac1,           %[src_ptr_0],   %[vector2b]     \n\t"
+          "dpsu.h.qbr     $ac1,           %[src_ptr_l1],  %[vector3b]     \n\t"
+          "extp           %[Temp4],       $ac1,           9               \n\t"
 
-                : [Temp1] "=&r" (Temp1), [Temp2] "=&r" (Temp2),
-                  [Temp3] "=&r" (Temp3), [Temp4] "=r" (Temp4),
-                  [src_ptr_l1] "=&r" (src_ptr_l1), [src_ptr_0] "=&r" (src_ptr_0),
-                  [src_ptr_r1] "=&r" (src_ptr_r1), [src_ptr_r2] "=&r" (src_ptr_r2)
-                : [vector2b] "r" (vector2b), [vector3b] "r" (vector3b),
-                  [vector4a] "r" (vector4a), [src_ptr] "r" (src_ptr)
-            );
+          : [Temp1] "=&r"(Temp1), [Temp2] "=&r"(Temp2), [Temp3] "=&r"(Temp3),
+            [Temp4] "=r"(Temp4), [src_ptr_l1] "=&r"(src_ptr_l1),
+            [src_ptr_0] "=&r"(src_ptr_0), [src_ptr_r1] "=&r"(src_ptr_r1),
+            [src_ptr_r2] "=&r"(src_ptr_r2)
+          : [vector2b] "r"(vector2b), [vector3b] "r"(vector3b),
+            [vector4a] "r"(vector4a), [src_ptr] "r"(src_ptr));
 
-            /* clamp and store results */
-            output_ptr[0] = cm[Temp1];
-            output_ptr[1] = cm[Temp2];
-            output_ptr[2] = cm[Temp3];
-            output_ptr[3] = cm[Temp4];
+      /* clamp and store results */
+      output_ptr[0] = cm[Temp1];
+      output_ptr[1] = cm[Temp2];
+      output_ptr[2] = cm[Temp3];
+      output_ptr[3] = cm[Temp4];
 
-            src_ptr += 8;
-            output_ptr += output_pitch;
-        }
+      src_ptr += 8;
+      output_ptr += output_pitch;
     }
+  }
 }
 
+void vp8_filter_block2d_second_pass_8(unsigned char *RESTRICT src_ptr,
+                                      unsigned char *RESTRICT output_ptr,
+                                      int output_pitch,
+                                      unsigned int output_height,
+                                      unsigned int output_width,
+                                      unsigned int yoffset) {
+  unsigned int i;
 
-void vp8_filter_block2d_second_pass_8
-(
-    unsigned char *RESTRICT src_ptr,
-    unsigned char *RESTRICT output_ptr,
-    int output_pitch,
-    unsigned int output_height,
-    unsigned int output_width,
-    unsigned int yoffset
-)
-{
-    unsigned int i;
+  int Temp1, Temp2, Temp3, Temp4, Temp5, Temp6, Temp7, Temp8;
+  unsigned int vector1b, vector2b, vector3b, vector4a;
 
-    int Temp1, Temp2, Temp3, Temp4, Temp5, Temp6, Temp7, Temp8;
-    unsigned int vector1b, vector2b, vector3b, vector4a;
+  unsigned char src_ptr_l2;
+  unsigned char src_ptr_l1;
+  unsigned char src_ptr_0;
+  unsigned char src_ptr_r1;
+  unsigned char src_ptr_r2;
+  unsigned char src_ptr_r3;
+  unsigned char *cm = ff_cropTbl + CROP_WIDTH;
 
-    unsigned char src_ptr_l2;
-    unsigned char src_ptr_l1;
-    unsigned char src_ptr_0;
-    unsigned char src_ptr_r1;
-    unsigned char src_ptr_r2;
-    unsigned char src_ptr_r3;
-    unsigned char *cm = ff_cropTbl + CROP_WIDTH;
+  vector4a = 64;
 
-    vector4a = 64;
+  vector1b = sub_pel_filterss[yoffset][0];
+  vector2b = sub_pel_filterss[yoffset][2];
+  vector3b = sub_pel_filterss[yoffset][1];
 
-    vector1b = sub_pel_filterss[yoffset][0];
-    vector2b = sub_pel_filterss[yoffset][2];
-    vector3b = sub_pel_filterss[yoffset][1];
+  if (vector1b) {
+    /* 6 tap filter */
 
-    if (vector1b)
-    {
-        /* 6 tap filter */
+    /* prefetch src_ptr data to cache memory */
+    prefetch_load(src_ptr);
 
-        /* prefetch src_ptr data to cache memory */
-        prefetch_load(src_ptr);
+    for (i = output_height; i--;) {
+      /* apply filter with vectors pairs */
+      __asm__ __volatile__(
+          "lbu            %[src_ptr_l2],  -16(%[src_ptr])                 \n\t"
+          "lbu            %[src_ptr_l1],  -8(%[src_ptr])                  \n\t"
+          "lbu            %[src_ptr_0],   0(%[src_ptr])                   \n\t"
+          "lbu            %[src_ptr_r1],  8(%[src_ptr])                   \n\t"
+          "lbu            %[src_ptr_r2],  16(%[src_ptr])                  \n\t"
+          "lbu            %[src_ptr_r3],  24(%[src_ptr])                  \n\t"
+          "mtlo           %[vector4a],    $ac2                            \n\t"
 
-        for (i = output_height; i--;)
-        {
-            /* apply filter with vectors pairs */
-            __asm__ __volatile__ (
-                "lbu            %[src_ptr_l2],  -16(%[src_ptr])                 \n\t"
-                "lbu            %[src_ptr_l1],  -8(%[src_ptr])                  \n\t"
-                "lbu            %[src_ptr_0],   0(%[src_ptr])                   \n\t"
-                "lbu            %[src_ptr_r1],  8(%[src_ptr])                   \n\t"
-                "lbu            %[src_ptr_r2],  16(%[src_ptr])                  \n\t"
-                "lbu            %[src_ptr_r3],  24(%[src_ptr])                  \n\t"
-                "mtlo           %[vector4a],    $ac2                            \n\t"
+          "append         %[src_ptr_0],   %[src_ptr_r1],  8               \n\t"
+          "append         %[src_ptr_l1],  %[src_ptr_r2],  8               \n\t"
+          "append         %[src_ptr_l2],  %[src_ptr_r3],  8               \n\t"
+          "dpau.h.qbr     $ac2,           %[src_ptr_l2],  %[vector1b]     \n\t"
+          "dpau.h.qbr     $ac2,           %[src_ptr_0],   %[vector2b]     \n\t"
+          "dpsu.h.qbr     $ac2,           %[src_ptr_l1],  %[vector3b]     \n\t"
 
-                "append         %[src_ptr_0],   %[src_ptr_r1],  8               \n\t"
-                "append         %[src_ptr_l1],  %[src_ptr_r2],  8               \n\t"
-                "append         %[src_ptr_l2],  %[src_ptr_r3],  8               \n\t"
-                "dpau.h.qbr     $ac2,           %[src_ptr_l2],  %[vector1b]     \n\t"
-                "dpau.h.qbr     $ac2,           %[src_ptr_0],   %[vector2b]     \n\t"
-                "dpsu.h.qbr     $ac2,           %[src_ptr_l1],  %[vector3b]     \n\t"
+          "lbu            %[src_ptr_l2],  -15(%[src_ptr])                 \n\t"
+          "lbu            %[src_ptr_l1],  -7(%[src_ptr])                  \n\t"
+          "lbu            %[src_ptr_0],   1(%[src_ptr])                   \n\t"
+          "lbu            %[src_ptr_r1],  9(%[src_ptr])                   \n\t"
+          "lbu            %[src_ptr_r2],  17(%[src_ptr])                  \n\t"
+          "lbu            %[src_ptr_r3],  25(%[src_ptr])                  \n\t"
+          "mtlo           %[vector4a],    $ac3                            \n\t"
+          "extp           %[Temp1],       $ac2,           9               \n\t"
 
-                "lbu            %[src_ptr_l2],  -15(%[src_ptr])                 \n\t"
-                "lbu            %[src_ptr_l1],  -7(%[src_ptr])                  \n\t"
-                "lbu            %[src_ptr_0],   1(%[src_ptr])                   \n\t"
-                "lbu            %[src_ptr_r1],  9(%[src_ptr])                   \n\t"
-                "lbu            %[src_ptr_r2],  17(%[src_ptr])                  \n\t"
-                "lbu            %[src_ptr_r3],  25(%[src_ptr])                  \n\t"
-                "mtlo           %[vector4a],    $ac3                            \n\t"
-                "extp           %[Temp1],       $ac2,           9               \n\t"
+          "append         %[src_ptr_l2],  %[src_ptr_r3],  8               \n\t"
+          "append         %[src_ptr_0],   %[src_ptr_r1],  8               \n\t"
+          "append         %[src_ptr_l1],  %[src_ptr_r2],  8               \n\t"
+          "dpau.h.qbr     $ac3,           %[src_ptr_l2],  %[vector1b]     \n\t"
+          "dpau.h.qbr     $ac3,           %[src_ptr_0],   %[vector2b]     \n\t"
+          "dpsu.h.qbr     $ac3,           %[src_ptr_l1],  %[vector3b]     \n\t"
 
-                "append         %[src_ptr_l2],  %[src_ptr_r3],  8               \n\t"
-                "append         %[src_ptr_0],   %[src_ptr_r1],  8               \n\t"
-                "append         %[src_ptr_l1],  %[src_ptr_r2],  8               \n\t"
-                "dpau.h.qbr     $ac3,           %[src_ptr_l2],  %[vector1b]     \n\t"
-                "dpau.h.qbr     $ac3,           %[src_ptr_0],   %[vector2b]     \n\t"
-                "dpsu.h.qbr     $ac3,           %[src_ptr_l1],  %[vector3b]     \n\t"
+          "lbu            %[src_ptr_l2],  -14(%[src_ptr])                 \n\t"
+          "lbu            %[src_ptr_l1],  -6(%[src_ptr])                  \n\t"
+          "lbu            %[src_ptr_0],   2(%[src_ptr])                   \n\t"
+          "lbu            %[src_ptr_r1],  10(%[src_ptr])                  \n\t"
+          "lbu            %[src_ptr_r2],  18(%[src_ptr])                  \n\t"
+          "lbu            %[src_ptr_r3],  26(%[src_ptr])                  \n\t"
+          "mtlo           %[vector4a],    $ac0                            \n\t"
+          "extp           %[Temp2],       $ac3,           9               \n\t"
 
-                "lbu            %[src_ptr_l2],  -14(%[src_ptr])                 \n\t"
-                "lbu            %[src_ptr_l1],  -6(%[src_ptr])                  \n\t"
-                "lbu            %[src_ptr_0],   2(%[src_ptr])                   \n\t"
-                "lbu            %[src_ptr_r1],  10(%[src_ptr])                  \n\t"
-                "lbu            %[src_ptr_r2],  18(%[src_ptr])                  \n\t"
-                "lbu            %[src_ptr_r3],  26(%[src_ptr])                  \n\t"
-                "mtlo           %[vector4a],    $ac0                            \n\t"
-                "extp           %[Temp2],       $ac3,           9               \n\t"
+          "append         %[src_ptr_l2],  %[src_ptr_r3],  8               \n\t"
+          "append         %[src_ptr_0],   %[src_ptr_r1],  8               \n\t"
+          "append         %[src_ptr_l1],  %[src_ptr_r2],  8               \n\t"
+          "dpau.h.qbr     $ac0,           %[src_ptr_l2],  %[vector1b]     \n\t"
+          "dpau.h.qbr     $ac0,           %[src_ptr_0],   %[vector2b]     \n\t"
+          "dpsu.h.qbr     $ac0,           %[src_ptr_l1],  %[vector3b]     \n\t"
 
-                "append         %[src_ptr_l2],  %[src_ptr_r3],  8               \n\t"
-                "append         %[src_ptr_0],   %[src_ptr_r1],  8               \n\t"
-                "append         %[src_ptr_l1],  %[src_ptr_r2],  8               \n\t"
-                "dpau.h.qbr     $ac0,           %[src_ptr_l2],  %[vector1b]     \n\t"
-                "dpau.h.qbr     $ac0,           %[src_ptr_0],   %[vector2b]     \n\t"
-                "dpsu.h.qbr     $ac0,           %[src_ptr_l1],  %[vector3b]     \n\t"
+          "lbu            %[src_ptr_l2],  -13(%[src_ptr])                 \n\t"
+          "lbu            %[src_ptr_l1],  -5(%[src_ptr])                  \n\t"
+          "lbu            %[src_ptr_0],   3(%[src_ptr])                   \n\t"
+          "lbu            %[src_ptr_r1],  11(%[src_ptr])                  \n\t"
+          "lbu            %[src_ptr_r2],  19(%[src_ptr])                  \n\t"
+          "lbu            %[src_ptr_r3],  27(%[src_ptr])                  \n\t"
+          "mtlo           %[vector4a],    $ac1                            \n\t"
+          "extp           %[Temp3],       $ac0,           9               \n\t"
 
-                "lbu            %[src_ptr_l2],  -13(%[src_ptr])                 \n\t"
-                "lbu            %[src_ptr_l1],  -5(%[src_ptr])                  \n\t"
-                "lbu            %[src_ptr_0],   3(%[src_ptr])                   \n\t"
-                "lbu            %[src_ptr_r1],  11(%[src_ptr])                  \n\t"
-                "lbu            %[src_ptr_r2],  19(%[src_ptr])                  \n\t"
-                "lbu            %[src_ptr_r3],  27(%[src_ptr])                  \n\t"
-                "mtlo           %[vector4a],    $ac1                            \n\t"
-                "extp           %[Temp3],       $ac0,           9               \n\t"
+          "append         %[src_ptr_l2],  %[src_ptr_r3],  8               \n\t"
+          "append         %[src_ptr_0],   %[src_ptr_r1],  8               \n\t"
+          "append         %[src_ptr_l1],  %[src_ptr_r2],  8               \n\t"
+          "dpau.h.qbr     $ac1,           %[src_ptr_l2],  %[vector1b]     \n\t"
+          "dpau.h.qbr     $ac1,           %[src_ptr_0],   %[vector2b]     \n\t"
+          "dpsu.h.qbr     $ac1,           %[src_ptr_l1],  %[vector3b]     \n\t"
 
-                "append         %[src_ptr_l2],  %[src_ptr_r3],  8               \n\t"
-                "append         %[src_ptr_0],   %[src_ptr_r1],  8               \n\t"
-                "append         %[src_ptr_l1],  %[src_ptr_r2],  8               \n\t"
-                "dpau.h.qbr     $ac1,           %[src_ptr_l2],  %[vector1b]     \n\t"
-                "dpau.h.qbr     $ac1,           %[src_ptr_0],   %[vector2b]     \n\t"
-                "dpsu.h.qbr     $ac1,           %[src_ptr_l1],  %[vector3b]     \n\t"
+          : [Temp1] "=&r"(Temp1), [Temp2] "=&r"(Temp2), [Temp3] "=&r"(Temp3),
+            [src_ptr_l1] "=&r"(src_ptr_l1), [src_ptr_0] "=&r"(src_ptr_0),
+            [src_ptr_r1] "=&r"(src_ptr_r1), [src_ptr_r2] "=&r"(src_ptr_r2),
+            [src_ptr_l2] "=&r"(src_ptr_l2), [src_ptr_r3] "=&r"(src_ptr_r3)
+          : [vector1b] "r"(vector1b), [vector2b] "r"(vector2b),
+            [vector3b] "r"(vector3b), [vector4a] "r"(vector4a),
+            [src_ptr] "r"(src_ptr));
 
-                : [Temp1] "=&r" (Temp1), [Temp2] "=&r" (Temp2), [Temp3] "=&r" (Temp3),
-                  [src_ptr_l1] "=&r" (src_ptr_l1), [src_ptr_0] "=&r" (src_ptr_0),
-                  [src_ptr_r1] "=&r" (src_ptr_r1), [src_ptr_r2] "=&r" (src_ptr_r2),
-                  [src_ptr_l2] "=&r" (src_ptr_l2), [src_ptr_r3] "=&r" (src_ptr_r3)
-                : [vector1b] "r" (vector1b), [vector2b] "r" (vector2b),
-                  [vector3b] "r" (vector3b), [vector4a] "r" (vector4a),
-                  [src_ptr] "r" (src_ptr)
-            );
+      /* apply filter with vectors pairs */
+      __asm__ __volatile__(
+          "lbu            %[src_ptr_l2],  -12(%[src_ptr])                 \n\t"
+          "lbu            %[src_ptr_l1],  -4(%[src_ptr])                  \n\t"
+          "lbu            %[src_ptr_0],   4(%[src_ptr])                   \n\t"
+          "lbu            %[src_ptr_r1],  12(%[src_ptr])                  \n\t"
+          "lbu            %[src_ptr_r2],  20(%[src_ptr])                  \n\t"
+          "lbu            %[src_ptr_r3],  28(%[src_ptr])                  \n\t"
+          "mtlo           %[vector4a],    $ac2                            \n\t"
 
-            /* apply filter with vectors pairs */
-            __asm__ __volatile__ (
-                "lbu            %[src_ptr_l2],  -12(%[src_ptr])                 \n\t"
-                "lbu            %[src_ptr_l1],  -4(%[src_ptr])                  \n\t"
-                "lbu            %[src_ptr_0],   4(%[src_ptr])                   \n\t"
-                "lbu            %[src_ptr_r1],  12(%[src_ptr])                  \n\t"
-                "lbu            %[src_ptr_r2],  20(%[src_ptr])                  \n\t"
-                "lbu            %[src_ptr_r3],  28(%[src_ptr])                  \n\t"
-                "mtlo           %[vector4a],    $ac2                            \n\t"
+          "append         %[src_ptr_l2],  %[src_ptr_r3],  8               \n\t"
+          "append         %[src_ptr_0],   %[src_ptr_r1],  8               \n\t"
+          "append         %[src_ptr_l1],  %[src_ptr_r2],  8               \n\t"
+          "dpau.h.qbr     $ac2,           %[src_ptr_l2],  %[vector1b]     \n\t"
+          "dpau.h.qbr     $ac2,           %[src_ptr_0],   %[vector2b]     \n\t"
+          "dpsu.h.qbr     $ac2,           %[src_ptr_l1],  %[vector3b]     \n\t"
+          "extp           %[Temp4],       $ac1,           9               \n\t"
 
-                "append         %[src_ptr_l2],  %[src_ptr_r3],  8               \n\t"
-                "append         %[src_ptr_0],   %[src_ptr_r1],  8               \n\t"
-                "append         %[src_ptr_l1],  %[src_ptr_r2],  8               \n\t"
-                "dpau.h.qbr     $ac2,           %[src_ptr_l2],  %[vector1b]     \n\t"
-                "dpau.h.qbr     $ac2,           %[src_ptr_0],   %[vector2b]     \n\t"
-                "dpsu.h.qbr     $ac2,           %[src_ptr_l1],  %[vector3b]     \n\t"
-                "extp           %[Temp4],       $ac1,           9               \n\t"
+          "lbu            %[src_ptr_l2],  -11(%[src_ptr])                 \n\t"
+          "lbu            %[src_ptr_l1],  -3(%[src_ptr])                  \n\t"
+          "lbu            %[src_ptr_0],   5(%[src_ptr])                   \n\t"
+          "lbu            %[src_ptr_r1],  13(%[src_ptr])                  \n\t"
+          "lbu            %[src_ptr_r2],  21(%[src_ptr])                  \n\t"
+          "lbu            %[src_ptr_r3],  29(%[src_ptr])                  \n\t"
+          "mtlo           %[vector4a],    $ac3                            \n\t"
+          "extp           %[Temp5],       $ac2,           9               \n\t"
 
-                "lbu            %[src_ptr_l2],  -11(%[src_ptr])                 \n\t"
-                "lbu            %[src_ptr_l1],  -3(%[src_ptr])                  \n\t"
-                "lbu            %[src_ptr_0],   5(%[src_ptr])                   \n\t"
-                "lbu            %[src_ptr_r1],  13(%[src_ptr])                  \n\t"
-                "lbu            %[src_ptr_r2],  21(%[src_ptr])                  \n\t"
-                "lbu            %[src_ptr_r3],  29(%[src_ptr])                  \n\t"
-                "mtlo           %[vector4a],    $ac3                            \n\t"
-                "extp           %[Temp5],       $ac2,           9               \n\t"
+          "append         %[src_ptr_l2],  %[src_ptr_r3],  8               \n\t"
+          "append         %[src_ptr_0],   %[src_ptr_r1],  8               \n\t"
+          "append         %[src_ptr_l1],  %[src_ptr_r2],  8               \n\t"
+          "dpau.h.qbr     $ac3,           %[src_ptr_l2],  %[vector1b]     \n\t"
+          "dpau.h.qbr     $ac3,           %[src_ptr_0],   %[vector2b]     \n\t"
+          "dpsu.h.qbr     $ac3,           %[src_ptr_l1],  %[vector3b]     \n\t"
 
-                "append         %[src_ptr_l2],  %[src_ptr_r3],  8               \n\t"
-                "append         %[src_ptr_0],   %[src_ptr_r1],  8               \n\t"
-                "append         %[src_ptr_l1],  %[src_ptr_r2],  8               \n\t"
-                "dpau.h.qbr     $ac3,           %[src_ptr_l2],  %[vector1b]     \n\t"
-                "dpau.h.qbr     $ac3,           %[src_ptr_0],   %[vector2b]     \n\t"
-                "dpsu.h.qbr     $ac3,           %[src_ptr_l1],  %[vector3b]     \n\t"
+          "lbu            %[src_ptr_l2],  -10(%[src_ptr])                 \n\t"
+          "lbu            %[src_ptr_l1],  -2(%[src_ptr])                  \n\t"
+          "lbu            %[src_ptr_0],   6(%[src_ptr])                   \n\t"
+          "lbu            %[src_ptr_r1],  14(%[src_ptr])                  \n\t"
+          "lbu            %[src_ptr_r2],  22(%[src_ptr])                  \n\t"
+          "lbu            %[src_ptr_r3],  30(%[src_ptr])                  \n\t"
+          "mtlo           %[vector4a],    $ac0                            \n\t"
+          "extp           %[Temp6],       $ac3,           9               \n\t"
 
-                "lbu            %[src_ptr_l2],  -10(%[src_ptr])                 \n\t"
-                "lbu            %[src_ptr_l1],  -2(%[src_ptr])                  \n\t"
-                "lbu            %[src_ptr_0],   6(%[src_ptr])                   \n\t"
-                "lbu            %[src_ptr_r1],  14(%[src_ptr])                  \n\t"
-                "lbu            %[src_ptr_r2],  22(%[src_ptr])                  \n\t"
-                "lbu            %[src_ptr_r3],  30(%[src_ptr])                  \n\t"
-                "mtlo           %[vector4a],    $ac0                            \n\t"
-                "extp           %[Temp6],       $ac3,           9               \n\t"
+          "append         %[src_ptr_l2],  %[src_ptr_r3],  8               \n\t"
+          "append         %[src_ptr_0],   %[src_ptr_r1],  8               \n\t"
+          "append         %[src_ptr_l1],  %[src_ptr_r2],  8               \n\t"
+          "dpau.h.qbr     $ac0,           %[src_ptr_l2],  %[vector1b]     \n\t"
+          "dpau.h.qbr     $ac0,           %[src_ptr_0],   %[vector2b]     \n\t"
+          "dpsu.h.qbr     $ac0,           %[src_ptr_l1],  %[vector3b]     \n\t"
 
-                "append         %[src_ptr_l2],  %[src_ptr_r3],  8               \n\t"
-                "append         %[src_ptr_0],   %[src_ptr_r1],  8               \n\t"
-                "append         %[src_ptr_l1],  %[src_ptr_r2],  8               \n\t"
-                "dpau.h.qbr     $ac0,           %[src_ptr_l2],  %[vector1b]     \n\t"
-                "dpau.h.qbr     $ac0,           %[src_ptr_0],   %[vector2b]     \n\t"
-                "dpsu.h.qbr     $ac0,           %[src_ptr_l1],  %[vector3b]     \n\t"
+          "lbu            %[src_ptr_l2],  -9(%[src_ptr])                  \n\t"
+          "lbu            %[src_ptr_l1],  -1(%[src_ptr])                  \n\t"
+          "lbu            %[src_ptr_0],   7(%[src_ptr])                   \n\t"
+          "lbu            %[src_ptr_r1],  15(%[src_ptr])                  \n\t"
+          "lbu            %[src_ptr_r2],  23(%[src_ptr])                  \n\t"
+          "lbu            %[src_ptr_r3],  31(%[src_ptr])                  \n\t"
+          "mtlo           %[vector4a],    $ac1                            \n\t"
+          "extp           %[Temp7],       $ac0,           9               \n\t"
 
-                "lbu            %[src_ptr_l2],  -9(%[src_ptr])                  \n\t"
-                "lbu            %[src_ptr_l1],  -1(%[src_ptr])                  \n\t"
-                "lbu            %[src_ptr_0],   7(%[src_ptr])                   \n\t"
-                "lbu            %[src_ptr_r1],  15(%[src_ptr])                  \n\t"
-                "lbu            %[src_ptr_r2],  23(%[src_ptr])                  \n\t"
-                "lbu            %[src_ptr_r3],  31(%[src_ptr])                  \n\t"
-                "mtlo           %[vector4a],    $ac1                            \n\t"
-                "extp           %[Temp7],       $ac0,           9               \n\t"
+          "append         %[src_ptr_l2],  %[src_ptr_r3],  8               \n\t"
+          "append         %[src_ptr_0],   %[src_ptr_r1],  8               \n\t"
+          "append         %[src_ptr_l1],  %[src_ptr_r2],  8               \n\t"
+          "dpau.h.qbr     $ac1,           %[src_ptr_l2],  %[vector1b]     \n\t"
+          "dpau.h.qbr     $ac1,           %[src_ptr_0],   %[vector2b]     \n\t"
+          "dpsu.h.qbr     $ac1,           %[src_ptr_l1],  %[vector3b]     \n\t"
+          "extp           %[Temp8],       $ac1,           9               \n\t"
 
-                "append         %[src_ptr_l2],  %[src_ptr_r3],  8               \n\t"
-                "append         %[src_ptr_0],   %[src_ptr_r1],  8               \n\t"
-                "append         %[src_ptr_l1],  %[src_ptr_r2],  8               \n\t"
-                "dpau.h.qbr     $ac1,           %[src_ptr_l2],  %[vector1b]     \n\t"
-                "dpau.h.qbr     $ac1,           %[src_ptr_0],   %[vector2b]     \n\t"
-                "dpsu.h.qbr     $ac1,           %[src_ptr_l1],  %[vector3b]     \n\t"
-                "extp           %[Temp8],       $ac1,           9               \n\t"
+          : [Temp4] "=&r"(Temp4), [Temp5] "=&r"(Temp5), [Temp6] "=&r"(Temp6),
+            [Temp7] "=&r"(Temp7), [Temp8] "=r"(Temp8),
+            [src_ptr_l1] "=&r"(src_ptr_l1), [src_ptr_0] "=&r"(src_ptr_0),
+            [src_ptr_r1] "=&r"(src_ptr_r1), [src_ptr_r2] "=&r"(src_ptr_r2),
+            [src_ptr_l2] "=&r"(src_ptr_l2), [src_ptr_r3] "=&r"(src_ptr_r3)
+          : [vector1b] "r"(vector1b), [vector2b] "r"(vector2b),
+            [vector3b] "r"(vector3b), [vector4a] "r"(vector4a),
+            [src_ptr] "r"(src_ptr));
 
-                : [Temp4] "=&r" (Temp4), [Temp5] "=&r" (Temp5),
-                  [Temp6] "=&r" (Temp6), [Temp7] "=&r" (Temp7), [Temp8] "=r" (Temp8),
-                  [src_ptr_l1] "=&r" (src_ptr_l1), [src_ptr_0] "=&r" (src_ptr_0),
-                  [src_ptr_r1] "=&r" (src_ptr_r1), [src_ptr_r2] "=&r" (src_ptr_r2),
-                  [src_ptr_l2] "=&r" (src_ptr_l2),[src_ptr_r3] "=&r" (src_ptr_r3)
-                : [vector1b] "r" (vector1b), [vector2b] "r" (vector2b),
-                  [vector3b] "r" (vector3b), [vector4a] "r" (vector4a),
-                  [src_ptr] "r" (src_ptr)
-            );
+      /* clamp and store results */
+      output_ptr[0] = cm[Temp1];
+      output_ptr[1] = cm[Temp2];
+      output_ptr[2] = cm[Temp3];
+      output_ptr[3] = cm[Temp4];
+      output_ptr[4] = cm[Temp5];
+      output_ptr[5] = cm[Temp6];
+      output_ptr[6] = cm[Temp7];
+      output_ptr[7] = cm[Temp8];
 
-            /* clamp and store results */
-            output_ptr[0] = cm[Temp1];
-            output_ptr[1] = cm[Temp2];
-            output_ptr[2] = cm[Temp3];
-            output_ptr[3] = cm[Temp4];
-            output_ptr[4] = cm[Temp5];
-            output_ptr[5] = cm[Temp6];
-            output_ptr[6] = cm[Temp7];
-            output_ptr[7] = cm[Temp8];
-
-            src_ptr += 8;
-            output_ptr += output_pitch;
-        }
+      src_ptr += 8;
+      output_ptr += output_pitch;
     }
-    else
-    {
-        /* 4 tap filter */
+  } else {
+    /* 4 tap filter */
 
-        /* prefetch src_ptr data to cache memory */
-        prefetch_load(src_ptr);
+    /* prefetch src_ptr data to cache memory */
+    prefetch_load(src_ptr);
 
-        for (i = output_height; i--;)
-        {
-            __asm__ __volatile__ (
-                "lbu            %[src_ptr_l1],  -8(%[src_ptr])                  \n\t"
-                "lbu            %[src_ptr_0],   0(%[src_ptr])                   \n\t"
-                "lbu            %[src_ptr_r1],  8(%[src_ptr])                   \n\t"
-                "lbu            %[src_ptr_r2],  16(%[src_ptr])                  \n\t"
-                "mtlo           %[vector4a],    $ac2                            \n\t"
-                "append         %[src_ptr_0],   %[src_ptr_r1],  8               \n\t"
-                "append         %[src_ptr_l1],  %[src_ptr_r2],  8               \n\t"
-                "dpau.h.qbr     $ac2,           %[src_ptr_0],   %[vector2b]     \n\t"
-                "dpsu.h.qbr     $ac2,           %[src_ptr_l1],  %[vector3b]     \n\t"
+    for (i = output_height; i--;) {
+      __asm__ __volatile__(
+          "lbu            %[src_ptr_l1],  -8(%[src_ptr])                  \n\t"
+          "lbu            %[src_ptr_0],   0(%[src_ptr])                   \n\t"
+          "lbu            %[src_ptr_r1],  8(%[src_ptr])                   \n\t"
+          "lbu            %[src_ptr_r2],  16(%[src_ptr])                  \n\t"
+          "mtlo           %[vector4a],    $ac2                            \n\t"
+          "append         %[src_ptr_0],   %[src_ptr_r1],  8               \n\t"
+          "append         %[src_ptr_l1],  %[src_ptr_r2],  8               \n\t"
+          "dpau.h.qbr     $ac2,           %[src_ptr_0],   %[vector2b]     \n\t"
+          "dpsu.h.qbr     $ac2,           %[src_ptr_l1],  %[vector3b]     \n\t"
 
-                : [src_ptr_l1] "=&r" (src_ptr_l1), [src_ptr_0] "=&r" (src_ptr_0),
-                  [src_ptr_r1] "=&r" (src_ptr_r1), [src_ptr_r2] "=&r" (src_ptr_r2)
-                : [vector2b] "r" (vector2b), [vector3b] "r" (vector3b),
-                  [vector4a] "r" (vector4a), [src_ptr] "r" (src_ptr)
-            );
+          : [src_ptr_l1] "=&r"(src_ptr_l1), [src_ptr_0] "=&r"(src_ptr_0),
+            [src_ptr_r1] "=&r"(src_ptr_r1), [src_ptr_r2] "=&r"(src_ptr_r2)
+          : [vector2b] "r"(vector2b), [vector3b] "r"(vector3b),
+            [vector4a] "r"(vector4a), [src_ptr] "r"(src_ptr));
 
-            __asm__ __volatile__ (
-                "lbu            %[src_ptr_l1],  -7(%[src_ptr])                  \n\t"
-                "lbu            %[src_ptr_0],   1(%[src_ptr])                   \n\t"
-                "lbu            %[src_ptr_r1],  9(%[src_ptr])                   \n\t"
-                "lbu            %[src_ptr_r2],  17(%[src_ptr])                  \n\t"
-                "mtlo           %[vector4a],    $ac3                            \n\t"
-                "append         %[src_ptr_0],   %[src_ptr_r1],  8               \n\t"
-                "append         %[src_ptr_l1],  %[src_ptr_r2],  8               \n\t"
-                "dpau.h.qbr     $ac3,           %[src_ptr_0],   %[vector2b]     \n\t"
-                "dpsu.h.qbr     $ac3,           %[src_ptr_l1],  %[vector3b]     \n\t"
-                "extp           %[Temp1],       $ac2,           9               \n\t"
+      __asm__ __volatile__(
+          "lbu            %[src_ptr_l1],  -7(%[src_ptr])                  \n\t"
+          "lbu            %[src_ptr_0],   1(%[src_ptr])                   \n\t"
+          "lbu            %[src_ptr_r1],  9(%[src_ptr])                   \n\t"
+          "lbu            %[src_ptr_r2],  17(%[src_ptr])                  \n\t"
+          "mtlo           %[vector4a],    $ac3                            \n\t"
+          "append         %[src_ptr_0],   %[src_ptr_r1],  8               \n\t"
+          "append         %[src_ptr_l1],  %[src_ptr_r2],  8               \n\t"
+          "dpau.h.qbr     $ac3,           %[src_ptr_0],   %[vector2b]     \n\t"
+          "dpsu.h.qbr     $ac3,           %[src_ptr_l1],  %[vector3b]     \n\t"
+          "extp           %[Temp1],       $ac2,           9               \n\t"
 
-                : [Temp1] "=r" (Temp1),
-                  [src_ptr_l1] "=&r" (src_ptr_l1), [src_ptr_0] "=&r" (src_ptr_0),
-                  [src_ptr_r1] "=&r" (src_ptr_r1), [src_ptr_r2] "=&r" (src_ptr_r2)
-                : [vector2b] "r" (vector2b), [vector3b] "r" (vector3b),
-                  [vector4a] "r" (vector4a), [src_ptr] "r" (src_ptr)
-            );
+          : [Temp1] "=r"(Temp1), [src_ptr_l1] "=&r"(src_ptr_l1),
+            [src_ptr_0] "=&r"(src_ptr_0), [src_ptr_r1] "=&r"(src_ptr_r1),
+            [src_ptr_r2] "=&r"(src_ptr_r2)
+          : [vector2b] "r"(vector2b), [vector3b] "r"(vector3b),
+            [vector4a] "r"(vector4a), [src_ptr] "r"(src_ptr));
 
-            src_ptr_l1 = src_ptr[-6];
-            src_ptr_0  = src_ptr[2];
-            src_ptr_r1 = src_ptr[10];
-            src_ptr_r2 = src_ptr[18];
+      src_ptr_l1 = src_ptr[-6];
+      src_ptr_0 = src_ptr[2];
+      src_ptr_r1 = src_ptr[10];
+      src_ptr_r2 = src_ptr[18];
 
-            __asm__ __volatile__ (
-                "mtlo           %[vector4a],    $ac0                            \n\t"
-                "append         %[src_ptr_0],   %[src_ptr_r1],  8               \n\t"
-                "append         %[src_ptr_l1],  %[src_ptr_r2],  8               \n\t"
-                "dpau.h.qbr     $ac0,           %[src_ptr_0],   %[vector2b]     \n\t"
-                "dpsu.h.qbr     $ac0,           %[src_ptr_l1],  %[vector3b]     \n\t"
-                "extp           %[Temp2],       $ac3,           9               \n\t"
+      __asm__ __volatile__(
+          "mtlo           %[vector4a],    $ac0                            \n\t"
+          "append         %[src_ptr_0],   %[src_ptr_r1],  8               \n\t"
+          "append         %[src_ptr_l1],  %[src_ptr_r2],  8               \n\t"
+          "dpau.h.qbr     $ac0,           %[src_ptr_0],   %[vector2b]     \n\t"
+          "dpsu.h.qbr     $ac0,           %[src_ptr_l1],  %[vector3b]     \n\t"
+          "extp           %[Temp2],       $ac3,           9               \n\t"
 
-                : [Temp2] "=r" (Temp2)
-                : [vector2b] "r" (vector2b), [vector3b] "r" (vector3b),
-                  [src_ptr_l1] "r" (src_ptr_l1), [src_ptr_0] "r" (src_ptr_0),
-                  [src_ptr_r1] "r" (src_ptr_r1), [src_ptr_r2] "r" (src_ptr_r2),
-                  [vector4a] "r" (vector4a)
-            );
-
-            src_ptr_l1 = src_ptr[-5];
-            src_ptr_0  = src_ptr[3];
-            src_ptr_r1 = src_ptr[11];
-            src_ptr_r2 = src_ptr[19];
+          : [Temp2] "=r"(Temp2)
+          : [vector2b] "r"(vector2b), [vector3b] "r"(vector3b),
+            [src_ptr_l1] "r"(src_ptr_l1), [src_ptr_0] "r"(src_ptr_0),
+            [src_ptr_r1] "r"(src_ptr_r1), [src_ptr_r2] "r"(src_ptr_r2),
+            [vector4a] "r"(vector4a));
 
-            __asm__ __volatile__ (
-                "mtlo           %[vector4a],    $ac1                            \n\t"
-                "append         %[src_ptr_0],   %[src_ptr_r1],  8               \n\t"
-                "append         %[src_ptr_l1],  %[src_ptr_r2],  8               \n\t"
-                "dpau.h.qbr     $ac1,           %[src_ptr_0],   %[vector2b]     \n\t"
-                "dpsu.h.qbr     $ac1,           %[src_ptr_l1],  %[vector3b]     \n\t"
-                "extp           %[Temp3],       $ac0,           9               \n\t"
+      src_ptr_l1 = src_ptr[-5];
+      src_ptr_0 = src_ptr[3];
+      src_ptr_r1 = src_ptr[11];
+      src_ptr_r2 = src_ptr[19];
 
-                : [Temp3] "=r" (Temp3)
-                : [vector2b] "r" (vector2b), [vector3b] "r" (vector3b),
-                  [src_ptr_l1] "r" (src_ptr_l1), [src_ptr_0] "r" (src_ptr_0),
-                  [src_ptr_r1] "r" (src_ptr_r1), [src_ptr_r2] "r" (src_ptr_r2),
-                  [vector4a] "r" (vector4a)
-            );
+      __asm__ __volatile__(
+          "mtlo           %[vector4a],    $ac1                            \n\t"
+          "append         %[src_ptr_0],   %[src_ptr_r1],  8               \n\t"
+          "append         %[src_ptr_l1],  %[src_ptr_r2],  8               \n\t"
+          "dpau.h.qbr     $ac1,           %[src_ptr_0],   %[vector2b]     \n\t"
+          "dpsu.h.qbr     $ac1,           %[src_ptr_l1],  %[vector3b]     \n\t"
+          "extp           %[Temp3],       $ac0,           9               \n\t"
 
-            src_ptr_l1 = src_ptr[-4];
-            src_ptr_0  = src_ptr[4];
-            src_ptr_r1 = src_ptr[12];
-            src_ptr_r2 = src_ptr[20];
+          : [Temp3] "=r"(Temp3)
+          : [vector2b] "r"(vector2b), [vector3b] "r"(vector3b),
+            [src_ptr_l1] "r"(src_ptr_l1), [src_ptr_0] "r"(src_ptr_0),
+            [src_ptr_r1] "r"(src_ptr_r1), [src_ptr_r2] "r"(src_ptr_r2),
+            [vector4a] "r"(vector4a));
 
-            __asm__ __volatile__ (
-                "mtlo           %[vector4a],    $ac2                            \n\t"
-                "append         %[src_ptr_0],   %[src_ptr_r1],  8               \n\t"
-                "append         %[src_ptr_l1],  %[src_ptr_r2],  8               \n\t"
-                "dpau.h.qbr     $ac2,           %[src_ptr_0],   %[vector2b]     \n\t"
-                "dpsu.h.qbr     $ac2,           %[src_ptr_l1],  %[vector3b]     \n\t"
-                "extp           %[Temp4],       $ac1,           9               \n\t"
+      src_ptr_l1 = src_ptr[-4];
+      src_ptr_0 = src_ptr[4];
+      src_ptr_r1 = src_ptr[12];
+      src_ptr_r2 = src_ptr[20];
 
-                : [Temp4] "=r" (Temp4)
-                : [vector2b] "r" (vector2b), [vector3b] "r" (vector3b),
-                  [src_ptr_l1] "r" (src_ptr_l1), [src_ptr_0] "r" (src_ptr_0),
-                  [src_ptr_r1] "r" (src_ptr_r1), [src_ptr_r2] "r" (src_ptr_r2),
-                  [vector4a] "r" (vector4a)
-            );
+      __asm__ __volatile__(
+          "mtlo           %[vector4a],    $ac2                            \n\t"
+          "append         %[src_ptr_0],   %[src_ptr_r1],  8               \n\t"
+          "append         %[src_ptr_l1],  %[src_ptr_r2],  8               \n\t"
+          "dpau.h.qbr     $ac2,           %[src_ptr_0],   %[vector2b]     \n\t"
+          "dpsu.h.qbr     $ac2,           %[src_ptr_l1],  %[vector3b]     \n\t"
+          "extp           %[Temp4],       $ac1,           9               \n\t"
 
-            src_ptr_l1 = src_ptr[-3];
-            src_ptr_0  = src_ptr[5];
-            src_ptr_r1 = src_ptr[13];
-            src_ptr_r2 = src_ptr[21];
+          : [Temp4] "=r"(Temp4)
+          : [vector2b] "r"(vector2b), [vector3b] "r"(vector3b),
+            [src_ptr_l1] "r"(src_ptr_l1), [src_ptr_0] "r"(src_ptr_0),
+            [src_ptr_r1] "r"(src_ptr_r1), [src_ptr_r2] "r"(src_ptr_r2),
+            [vector4a] "r"(vector4a));
 
-            __asm__ __volatile__ (
-                "mtlo           %[vector4a],    $ac3                            \n\t"
-                "append         %[src_ptr_0],   %[src_ptr_r1],  8               \n\t"
-                "append         %[src_ptr_l1],  %[src_ptr_r2],  8               \n\t"
-                "dpau.h.qbr     $ac3,           %[src_ptr_0],   %[vector2b]     \n\t"
-                "dpsu.h.qbr     $ac3,           %[src_ptr_l1],  %[vector3b]     \n\t"
-                "extp           %[Temp5],       $ac2,           9               \n\t"
+      src_ptr_l1 = src_ptr[-3];
+      src_ptr_0 = src_ptr[5];
+      src_ptr_r1 = src_ptr[13];
+      src_ptr_r2 = src_ptr[21];
 
-                : [Temp5] "=&r" (Temp5)
-                : [vector2b] "r" (vector2b), [vector3b] "r" (vector3b),
-                  [src_ptr_l1] "r" (src_ptr_l1), [src_ptr_0] "r" (src_ptr_0),
-                  [src_ptr_r1] "r" (src_ptr_r1), [src_ptr_r2] "r" (src_ptr_r2),
-                  [vector4a] "r" (vector4a)
-            );
+      __asm__ __volatile__(
+          "mtlo           %[vector4a],    $ac3                            \n\t"
+          "append         %[src_ptr_0],   %[src_ptr_r1],  8               \n\t"
+          "append         %[src_ptr_l1],  %[src_ptr_r2],  8               \n\t"
+          "dpau.h.qbr     $ac3,           %[src_ptr_0],   %[vector2b]     \n\t"
+          "dpsu.h.qbr     $ac3,           %[src_ptr_l1],  %[vector3b]     \n\t"
+          "extp           %[Temp5],       $ac2,           9               \n\t"
 
-            src_ptr_l1 = src_ptr[-2];
-            src_ptr_0  = src_ptr[6];
-            src_ptr_r1 = src_ptr[14];
-            src_ptr_r2 = src_ptr[22];
+          : [Temp5] "=&r"(Temp5)
+          : [vector2b] "r"(vector2b), [vector3b] "r"(vector3b),
+            [src_ptr_l1] "r"(src_ptr_l1), [src_ptr_0] "r"(src_ptr_0),
+            [src_ptr_r1] "r"(src_ptr_r1), [src_ptr_r2] "r"(src_ptr_r2),
+            [vector4a] "r"(vector4a));
 
-            __asm__ __volatile__ (
-                "mtlo           %[vector4a],    $ac0                            \n\t"
-                "append         %[src_ptr_0],   %[src_ptr_r1],  8               \n\t"
-                "append         %[src_ptr_l1],  %[src_ptr_r2],  8               \n\t"
-                "dpau.h.qbr     $ac0,           %[src_ptr_0],   %[vector2b]     \n\t"
-                "dpsu.h.qbr     $ac0,           %[src_ptr_l1],  %[vector3b]     \n\t"
-                "extp           %[Temp6],       $ac3,           9               \n\t"
+      src_ptr_l1 = src_ptr[-2];
+      src_ptr_0 = src_ptr[6];
+      src_ptr_r1 = src_ptr[14];
+      src_ptr_r2 = src_ptr[22];
 
-                : [Temp6] "=r" (Temp6)
-                : [vector2b] "r" (vector2b), [vector3b] "r" (vector3b),
-                  [src_ptr_l1] "r" (src_ptr_l1), [src_ptr_0] "r" (src_ptr_0),
-                  [src_ptr_r1] "r" (src_ptr_r1), [src_ptr_r2] "r" (src_ptr_r2),
-                  [vector4a] "r" (vector4a)
-            );
+      __asm__ __volatile__(
+          "mtlo           %[vector4a],    $ac0                            \n\t"
+          "append         %[src_ptr_0],   %[src_ptr_r1],  8               \n\t"
+          "append         %[src_ptr_l1],  %[src_ptr_r2],  8               \n\t"
+          "dpau.h.qbr     $ac0,           %[src_ptr_0],   %[vector2b]     \n\t"
+          "dpsu.h.qbr     $ac0,           %[src_ptr_l1],  %[vector3b]     \n\t"
+          "extp           %[Temp6],       $ac3,           9               \n\t"
 
-            src_ptr_l1 = src_ptr[-1];
-            src_ptr_0  = src_ptr[7];
-            src_ptr_r1 = src_ptr[15];
-            src_ptr_r2 = src_ptr[23];
+          : [Temp6] "=r"(Temp6)
+          : [vector2b] "r"(vector2b), [vector3b] "r"(vector3b),
+            [src_ptr_l1] "r"(src_ptr_l1), [src_ptr_0] "r"(src_ptr_0),
+            [src_ptr_r1] "r"(src_ptr_r1), [src_ptr_r2] "r"(src_ptr_r2),
+            [vector4a] "r"(vector4a));
 
-            __asm__ __volatile__ (
-                "mtlo           %[vector4a],    $ac1                            \n\t"
-                "append         %[src_ptr_0],   %[src_ptr_r1],  8               \n\t"
-                "append         %[src_ptr_l1],  %[src_ptr_r2],  8               \n\t"
-                "dpau.h.qbr     $ac1,           %[src_ptr_0],   %[vector2b]     \n\t"
-                "dpsu.h.qbr     $ac1,           %[src_ptr_l1],  %[vector3b]     \n\t"
-                "extp           %[Temp7],       $ac0,           9               \n\t"
-                "extp           %[Temp8],       $ac1,           9               \n\t"
+      src_ptr_l1 = src_ptr[-1];
+      src_ptr_0 = src_ptr[7];
+      src_ptr_r1 = src_ptr[15];
+      src_ptr_r2 = src_ptr[23];
 
-                : [Temp7] "=&r" (Temp7), [Temp8] "=r" (Temp8)
-                : [vector2b] "r" (vector2b), [vector3b] "r" (vector3b),
-                  [src_ptr_l1] "r" (src_ptr_l1), [src_ptr_0] "r" (src_ptr_0),
-                  [src_ptr_r1] "r" (src_ptr_r1), [src_ptr_r2] "r" (src_ptr_r2),
-                  [vector4a] "r" (vector4a)
-            );
+      __asm__ __volatile__(
+          "mtlo           %[vector4a],    $ac1                            \n\t"
+          "append         %[src_ptr_0],   %[src_ptr_r1],  8               \n\t"
+          "append         %[src_ptr_l1],  %[src_ptr_r2],  8               \n\t"
+          "dpau.h.qbr     $ac1,           %[src_ptr_0],   %[vector2b]     \n\t"
+          "dpsu.h.qbr     $ac1,           %[src_ptr_l1],  %[vector3b]     \n\t"
+          "extp           %[Temp7],       $ac0,           9               \n\t"
+          "extp           %[Temp8],       $ac1,           9               \n\t"
 
-            /* clamp and store results */
-            output_ptr[0] = cm[Temp1];
-            output_ptr[1] = cm[Temp2];
-            output_ptr[2] = cm[Temp3];
-            output_ptr[3] = cm[Temp4];
-            output_ptr[4] = cm[Temp5];
-            output_ptr[5] = cm[Temp6];
-            output_ptr[6] = cm[Temp7];
-            output_ptr[7] = cm[Temp8];
+          : [Temp7] "=&r"(Temp7), [Temp8] "=r"(Temp8)
+          : [vector2b] "r"(vector2b), [vector3b] "r"(vector3b),
+            [src_ptr_l1] "r"(src_ptr_l1), [src_ptr_0] "r"(src_ptr_0),
+            [src_ptr_r1] "r"(src_ptr_r1), [src_ptr_r2] "r"(src_ptr_r2),
+            [vector4a] "r"(vector4a));
 
-            src_ptr += 8;
-            output_ptr += output_pitch;
-        }
+      /* clamp and store results */
+      output_ptr[0] = cm[Temp1];
+      output_ptr[1] = cm[Temp2];
+      output_ptr[2] = cm[Temp3];
+      output_ptr[3] = cm[Temp4];
+      output_ptr[4] = cm[Temp5];
+      output_ptr[5] = cm[Temp6];
+      output_ptr[6] = cm[Temp7];
+      output_ptr[7] = cm[Temp8];
+
+      src_ptr += 8;
+      output_ptr += output_pitch;
     }
+  }
 }
 
+void vp8_filter_block2d_second_pass161(unsigned char *RESTRICT src_ptr,
+                                       unsigned char *RESTRICT output_ptr,
+                                       int output_pitch,
+                                       const unsigned short *vp8_filter) {
+  unsigned int i, j;
 
-void vp8_filter_block2d_second_pass161
-(
-    unsigned char *RESTRICT src_ptr,
-    unsigned char *RESTRICT output_ptr,
-    int output_pitch,
-    const unsigned short *vp8_filter
-)
-{
-    unsigned int i, j;
+  int Temp1, Temp2, Temp3, Temp4, Temp5, Temp6, Temp7, Temp8;
+  unsigned int vector4a;
+  unsigned int vector1b, vector2b, vector3b;
 
-    int Temp1, Temp2, Temp3, Temp4, Temp5, Temp6, Temp7, Temp8;
-    unsigned int vector4a;
-    unsigned int vector1b, vector2b, vector3b;
+  unsigned char src_ptr_l2;
+  unsigned char src_ptr_l1;
+  unsigned char src_ptr_0;
+  unsigned char src_ptr_r1;
+  unsigned char src_ptr_r2;
+  unsigned char src_ptr_r3;
+  unsigned char *cm = ff_cropTbl + CROP_WIDTH;
 
-    unsigned char src_ptr_l2;
-    unsigned char src_ptr_l1;
-    unsigned char src_ptr_0;
-    unsigned char src_ptr_r1;
-    unsigned char src_ptr_r2;
-    unsigned char src_ptr_r3;
-    unsigned char *cm = ff_cropTbl + CROP_WIDTH;
+  vector4a = 64;
 
-    vector4a = 64;
+  vector1b = vp8_filter[0];
+  vector2b = vp8_filter[2];
+  vector3b = vp8_filter[1];
 
-    vector1b = vp8_filter[0];
-    vector2b = vp8_filter[2];
-    vector3b = vp8_filter[1];
+  if (vector1b == 0) {
+    /* 4 tap filter */
 
-    if (vector1b == 0)
-    {
-        /* 4 tap filter */
+    /* prefetch src_ptr data to cache memory */
+    prefetch_load(src_ptr + 16);
 
-        /* prefetch src_ptr data to cache memory */
-        prefetch_load(src_ptr + 16);
+    for (i = 16; i--;) {
+      /* unrolling for loop */
+      for (j = 0; j < 16; j += 8) {
+        /* apply filter with vectors pairs */
+        __asm__ __volatile__(
+            "lbu            %[src_ptr_l1],  -16(%[src_ptr])                 "
+            "\n\t"
+            "lbu            %[src_ptr_0],   0(%[src_ptr])                   "
+            "\n\t"
+            "lbu            %[src_ptr_r1],  16(%[src_ptr])                  "
+            "\n\t"
+            "lbu            %[src_ptr_r2],  32(%[src_ptr])                  "
+            "\n\t"
+            "mtlo           %[vector4a],    $ac2                            "
+            "\n\t"
+            "append         %[src_ptr_0],   %[src_ptr_r1],  8               "
+            "\n\t"
+            "append         %[src_ptr_l1],  %[src_ptr_r2],  8               "
+            "\n\t"
+            "dpau.h.qbr     $ac2,           %[src_ptr_0],   %[vector2b]     "
+            "\n\t"
+            "dpsu.h.qbr     $ac2,           %[src_ptr_l1],  %[vector3b]     "
+            "\n\t"
 
-        for (i = 16; i--;)
-        {
-            /* unrolling for loop */
-            for (j = 0; j < 16; j += 8)
-            {
-                /* apply filter with vectors pairs */
-                __asm__ __volatile__ (
-                    "lbu            %[src_ptr_l1],  -16(%[src_ptr])                 \n\t"
-                    "lbu            %[src_ptr_0],   0(%[src_ptr])                   \n\t"
-                    "lbu            %[src_ptr_r1],  16(%[src_ptr])                  \n\t"
-                    "lbu            %[src_ptr_r2],  32(%[src_ptr])                  \n\t"
-                    "mtlo           %[vector4a],    $ac2                            \n\t"
-                    "append         %[src_ptr_0],   %[src_ptr_r1],  8               \n\t"
-                    "append         %[src_ptr_l1],  %[src_ptr_r2],  8               \n\t"
-                    "dpau.h.qbr     $ac2,           %[src_ptr_0],   %[vector2b]     \n\t"
-                    "dpsu.h.qbr     $ac2,           %[src_ptr_l1],  %[vector3b]     \n\t"
+            "lbu            %[src_ptr_l1],  -15(%[src_ptr])                 "
+            "\n\t"
+            "lbu            %[src_ptr_0],   1(%[src_ptr])                   "
+            "\n\t"
+            "lbu            %[src_ptr_r1],  17(%[src_ptr])                  "
+            "\n\t"
+            "lbu            %[src_ptr_r2],  33(%[src_ptr])                  "
+            "\n\t"
+            "mtlo           %[vector4a],    $ac3                            "
+            "\n\t"
+            "extp           %[Temp1],       $ac2,           9               "
+            "\n\t"
 
-                    "lbu            %[src_ptr_l1],  -15(%[src_ptr])                 \n\t"
-                    "lbu            %[src_ptr_0],   1(%[src_ptr])                   \n\t"
-                    "lbu            %[src_ptr_r1],  17(%[src_ptr])                  \n\t"
-                    "lbu            %[src_ptr_r2],  33(%[src_ptr])                  \n\t"
-                    "mtlo           %[vector4a],    $ac3                            \n\t"
-                    "extp           %[Temp1],       $ac2,           9               \n\t"
+            "append         %[src_ptr_0],   %[src_ptr_r1],  8               "
+            "\n\t"
+            "append         %[src_ptr_l1],  %[src_ptr_r2],  8               "
+            "\n\t"
+            "dpau.h.qbr     $ac3,           %[src_ptr_0],   %[vector2b]     "
+            "\n\t"
+            "dpsu.h.qbr     $ac3,           %[src_ptr_l1],  %[vector3b]     "
+            "\n\t"
 
-                    "append         %[src_ptr_0],   %[src_ptr_r1],  8               \n\t"
-                    "append         %[src_ptr_l1],  %[src_ptr_r2],  8               \n\t"
-                    "dpau.h.qbr     $ac3,           %[src_ptr_0],   %[vector2b]     \n\t"
-                    "dpsu.h.qbr     $ac3,           %[src_ptr_l1],  %[vector3b]     \n\t"
+            "lbu            %[src_ptr_l1],  -14(%[src_ptr])                 "
+            "\n\t"
+            "lbu            %[src_ptr_0],   2(%[src_ptr])                   "
+            "\n\t"
+            "lbu            %[src_ptr_r1],  18(%[src_ptr])                  "
+            "\n\t"
+            "lbu            %[src_ptr_r2],  34(%[src_ptr])                  "
+            "\n\t"
+            "mtlo           %[vector4a],    $ac1                            "
+            "\n\t"
+            "extp           %[Temp2],       $ac3,           9               "
+            "\n\t"
 
-                    "lbu            %[src_ptr_l1],  -14(%[src_ptr])                 \n\t"
-                    "lbu            %[src_ptr_0],   2(%[src_ptr])                   \n\t"
-                    "lbu            %[src_ptr_r1],  18(%[src_ptr])                  \n\t"
-                    "lbu            %[src_ptr_r2],  34(%[src_ptr])                  \n\t"
-                    "mtlo           %[vector4a],    $ac1                            \n\t"
-                    "extp           %[Temp2],       $ac3,           9               \n\t"
+            "append         %[src_ptr_0],   %[src_ptr_r1],  8               "
+            "\n\t"
+            "append         %[src_ptr_l1],  %[src_ptr_r2],  8               "
+            "\n\t"
+            "dpau.h.qbr     $ac1,           %[src_ptr_0],   %[vector2b]     "
+            "\n\t"
+            "dpsu.h.qbr     $ac1,           %[src_ptr_l1],  %[vector3b]     "
+            "\n\t"
 
-                    "append         %[src_ptr_0],   %[src_ptr_r1],  8               \n\t"
-                    "append         %[src_ptr_l1],  %[src_ptr_r2],  8               \n\t"
-                    "dpau.h.qbr     $ac1,           %[src_ptr_0],   %[vector2b]     \n\t"
-                    "dpsu.h.qbr     $ac1,           %[src_ptr_l1],  %[vector3b]     \n\t"
+            "lbu            %[src_ptr_l1],  -13(%[src_ptr])                 "
+            "\n\t"
+            "lbu            %[src_ptr_0],   3(%[src_ptr])                   "
+            "\n\t"
+            "lbu            %[src_ptr_r1],  19(%[src_ptr])                  "
+            "\n\t"
+            "lbu            %[src_ptr_r2],  35(%[src_ptr])                  "
+            "\n\t"
+            "mtlo           %[vector4a],    $ac3                            "
+            "\n\t"
+            "extp           %[Temp3],       $ac1,           9               "
+            "\n\t"
 
-                    "lbu            %[src_ptr_l1],  -13(%[src_ptr])                 \n\t"
-                    "lbu            %[src_ptr_0],   3(%[src_ptr])                   \n\t"
-                    "lbu            %[src_ptr_r1],  19(%[src_ptr])                  \n\t"
-                    "lbu            %[src_ptr_r2],  35(%[src_ptr])                  \n\t"
-                    "mtlo           %[vector4a],    $ac3                            \n\t"
-                    "extp           %[Temp3],       $ac1,           9               \n\t"
+            "append         %[src_ptr_0],   %[src_ptr_r1],  8               "
+            "\n\t"
+            "append         %[src_ptr_l1],  %[src_ptr_r2],  8               "
+            "\n\t"
+            "dpau.h.qbr     $ac3,           %[src_ptr_0],   %[vector2b]     "
+            "\n\t"
+            "dpsu.h.qbr     $ac3,           %[src_ptr_l1],  %[vector3b]     "
+            "\n\t"
 
-                    "append         %[src_ptr_0],   %[src_ptr_r1],  8               \n\t"
-                    "append         %[src_ptr_l1],  %[src_ptr_r2],  8               \n\t"
-                    "dpau.h.qbr     $ac3,           %[src_ptr_0],   %[vector2b]     \n\t"
-                    "dpsu.h.qbr     $ac3,           %[src_ptr_l1],  %[vector3b]     \n\t"
+            "lbu            %[src_ptr_l1],  -12(%[src_ptr])                 "
+            "\n\t"
+            "lbu            %[src_ptr_0],   4(%[src_ptr])                   "
+            "\n\t"
+            "lbu            %[src_ptr_r1],  20(%[src_ptr])                  "
+            "\n\t"
+            "lbu            %[src_ptr_r2],  36(%[src_ptr])                  "
+            "\n\t"
+            "mtlo           %[vector4a],    $ac2                            "
+            "\n\t"
+            "extp           %[Temp4],       $ac3,           9               "
+            "\n\t"
 
-                    "lbu            %[src_ptr_l1],  -12(%[src_ptr])                 \n\t"
-                    "lbu            %[src_ptr_0],   4(%[src_ptr])                   \n\t"
-                    "lbu            %[src_ptr_r1],  20(%[src_ptr])                  \n\t"
-                    "lbu            %[src_ptr_r2],  36(%[src_ptr])                  \n\t"
-                    "mtlo           %[vector4a],    $ac2                            \n\t"
-                    "extp           %[Temp4],       $ac3,           9               \n\t"
+            "append         %[src_ptr_0],   %[src_ptr_r1],  8               "
+            "\n\t"
+            "append         %[src_ptr_l1],  %[src_ptr_r2],  8               "
+            "\n\t"
+            "dpau.h.qbr     $ac2,           %[src_ptr_0],   %[vector2b]     "
+            "\n\t"
+            "dpsu.h.qbr     $ac2,           %[src_ptr_l1],  %[vector3b]     "
+            "\n\t"
 
-                    "append         %[src_ptr_0],   %[src_ptr_r1],  8               \n\t"
-                    "append         %[src_ptr_l1],  %[src_ptr_r2],  8               \n\t"
-                    "dpau.h.qbr     $ac2,           %[src_ptr_0],   %[vector2b]     \n\t"
-                    "dpsu.h.qbr     $ac2,           %[src_ptr_l1],  %[vector3b]     \n\t"
+            "lbu            %[src_ptr_l1],  -11(%[src_ptr])                 "
+            "\n\t"
+            "lbu            %[src_ptr_0],   5(%[src_ptr])                   "
+            "\n\t"
+            "lbu            %[src_ptr_r1],  21(%[src_ptr])                  "
+            "\n\t"
+            "lbu            %[src_ptr_r2],  37(%[src_ptr])                  "
+            "\n\t"
+            "mtlo           %[vector4a],    $ac3                            "
+            "\n\t"
+            "extp           %[Temp5],       $ac2,           9               "
+            "\n\t"
 
-                    "lbu            %[src_ptr_l1],  -11(%[src_ptr])                 \n\t"
-                    "lbu            %[src_ptr_0],   5(%[src_ptr])                   \n\t"
-                    "lbu            %[src_ptr_r1],  21(%[src_ptr])                  \n\t"
-                    "lbu            %[src_ptr_r2],  37(%[src_ptr])                  \n\t"
-                    "mtlo           %[vector4a],    $ac3                            \n\t"
-                    "extp           %[Temp5],       $ac2,           9               \n\t"
+            "append         %[src_ptr_0],   %[src_ptr_r1],  8               "
+            "\n\t"
+            "append         %[src_ptr_l1],  %[src_ptr_r2],  8               "
+            "\n\t"
+            "dpau.h.qbr     $ac3,           %[src_ptr_0],   %[vector2b]     "
+            "\n\t"
+            "dpsu.h.qbr     $ac3,           %[src_ptr_l1],  %[vector3b]     "
+            "\n\t"
 
-                    "append         %[src_ptr_0],   %[src_ptr_r1],  8               \n\t"
-                    "append         %[src_ptr_l1],  %[src_ptr_r2],  8               \n\t"
-                    "dpau.h.qbr     $ac3,           %[src_ptr_0],   %[vector2b]     \n\t"
-                    "dpsu.h.qbr     $ac3,           %[src_ptr_l1],  %[vector3b]     \n\t"
+            "lbu            %[src_ptr_l1],  -10(%[src_ptr])                 "
+            "\n\t"
+            "lbu            %[src_ptr_0],   6(%[src_ptr])                   "
+            "\n\t"
+            "lbu            %[src_ptr_r1],  22(%[src_ptr])                  "
+            "\n\t"
+            "lbu            %[src_ptr_r2],  38(%[src_ptr])                  "
+            "\n\t"
+            "mtlo           %[vector4a],    $ac1                            "
+            "\n\t"
+            "extp           %[Temp6],       $ac3,           9               "
+            "\n\t"
 
-                    "lbu            %[src_ptr_l1],  -10(%[src_ptr])                 \n\t"
-                    "lbu            %[src_ptr_0],   6(%[src_ptr])                   \n\t"
-                    "lbu            %[src_ptr_r1],  22(%[src_ptr])                  \n\t"
-                    "lbu            %[src_ptr_r2],  38(%[src_ptr])                  \n\t"
-                    "mtlo           %[vector4a],    $ac1                            \n\t"
-                    "extp           %[Temp6],       $ac3,           9               \n\t"
+            "append         %[src_ptr_0],   %[src_ptr_r1],  8               "
+            "\n\t"
+            "append         %[src_ptr_l1],  %[src_ptr_r2],  8               "
+            "\n\t"
+            "dpau.h.qbr     $ac1,           %[src_ptr_0],   %[vector2b]     "
+            "\n\t"
+            "dpsu.h.qbr     $ac1,           %[src_ptr_l1],  %[vector3b]     "
+            "\n\t"
 
-                    "append         %[src_ptr_0],   %[src_ptr_r1],  8               \n\t"
-                    "append         %[src_ptr_l1],  %[src_ptr_r2],  8               \n\t"
-                    "dpau.h.qbr     $ac1,           %[src_ptr_0],   %[vector2b]     \n\t"
-                    "dpsu.h.qbr     $ac1,           %[src_ptr_l1],  %[vector3b]     \n\t"
+            "lbu            %[src_ptr_l1],  -9(%[src_ptr])                  "
+            "\n\t"
+            "lbu            %[src_ptr_0],   7(%[src_ptr])                   "
+            "\n\t"
+            "lbu            %[src_ptr_r1],  23(%[src_ptr])                  "
+            "\n\t"
+            "lbu            %[src_ptr_r2],  39(%[src_ptr])                  "
+            "\n\t"
+            "mtlo           %[vector4a],    $ac3                            "
+            "\n\t"
+            "extp           %[Temp7],       $ac1,           9               "
+            "\n\t"
 
-                    "lbu            %[src_ptr_l1],  -9(%[src_ptr])                  \n\t"
-                    "lbu            %[src_ptr_0],   7(%[src_ptr])                   \n\t"
-                    "lbu            %[src_ptr_r1],  23(%[src_ptr])                  \n\t"
-                    "lbu            %[src_ptr_r2],  39(%[src_ptr])                  \n\t"
-                    "mtlo           %[vector4a],    $ac3                            \n\t"
-                    "extp           %[Temp7],       $ac1,           9               \n\t"
+            "append         %[src_ptr_0],   %[src_ptr_r1],  8               "
+            "\n\t"
+            "append         %[src_ptr_l1],  %[src_ptr_r2],  8               "
+            "\n\t"
+            "dpau.h.qbr     $ac3,           %[src_ptr_0],   %[vector2b]     "
+            "\n\t"
+            "dpsu.h.qbr     $ac3,           %[src_ptr_l1],  %[vector3b]     "
+            "\n\t"
+            "extp           %[Temp8],       $ac3,           9               "
+            "\n\t"
 
-                    "append         %[src_ptr_0],   %[src_ptr_r1],  8               \n\t"
-                    "append         %[src_ptr_l1],  %[src_ptr_r2],  8               \n\t"
-                    "dpau.h.qbr     $ac3,           %[src_ptr_0],   %[vector2b]     \n\t"
-                    "dpsu.h.qbr     $ac3,           %[src_ptr_l1],  %[vector3b]     \n\t"
-                    "extp           %[Temp8],       $ac3,           9               \n\t"
+            : [Temp1] "=&r"(Temp1), [Temp2] "=&r"(Temp2), [Temp3] "=&r"(Temp3),
+              [Temp4] "=&r"(Temp4), [Temp5] "=&r"(Temp5), [Temp6] "=&r"(Temp6),
+              [Temp7] "=&r"(Temp7), [Temp8] "=r"(Temp8),
+              [src_ptr_l1] "=&r"(src_ptr_l1), [src_ptr_0] "=&r"(src_ptr_0),
+              [src_ptr_r1] "=&r"(src_ptr_r1), [src_ptr_r2] "=&r"(src_ptr_r2)
+            : [vector2b] "r"(vector2b), [vector3b] "r"(vector3b),
+              [vector4a] "r"(vector4a), [src_ptr] "r"(src_ptr));
 
-                    : [Temp1] "=&r" (Temp1), [Temp2] "=&r" (Temp2),
-                      [Temp3] "=&r" (Temp3), [Temp4] "=&r" (Temp4),
-                      [Temp5] "=&r" (Temp5), [Temp6] "=&r" (Temp6),
-                      [Temp7] "=&r" (Temp7), [Temp8] "=r" (Temp8),
-                      [src_ptr_l1] "=&r" (src_ptr_l1), [src_ptr_0] "=&r" (src_ptr_0),
-                      [src_ptr_r1] "=&r" (src_ptr_r1), [src_ptr_r2] "=&r" (src_ptr_r2)
-                    : [vector2b] "r" (vector2b), [vector3b] "r" (vector3b),
-                      [vector4a] "r" (vector4a), [src_ptr] "r" (src_ptr)
-                );
+        /* clamp and store results */
+        output_ptr[j] = cm[Temp1];
+        output_ptr[j + 1] = cm[Temp2];
+        output_ptr[j + 2] = cm[Temp3];
+        output_ptr[j + 3] = cm[Temp4];
+        output_ptr[j + 4] = cm[Temp5];
+        output_ptr[j + 5] = cm[Temp6];
+        output_ptr[j + 6] = cm[Temp7];
+        output_ptr[j + 7] = cm[Temp8];
 
-                /* clamp and store results */
-                output_ptr[j] = cm[Temp1];
-                output_ptr[j + 1] = cm[Temp2];
-                output_ptr[j + 2] = cm[Temp3];
-                output_ptr[j + 3] = cm[Temp4];
-                output_ptr[j + 4] = cm[Temp5];
-                output_ptr[j + 5] = cm[Temp6];
-                output_ptr[j + 6] = cm[Temp7];
-                output_ptr[j + 7] = cm[Temp8];
+        src_ptr += 8;
+      }
 
-                src_ptr += 8;
-            }
-
-            output_ptr += output_pitch;
-        }
+      output_ptr += output_pitch;
     }
-    else
-    {
-        /* 4 tap filter */
+  } else {
+    /* 4 tap filter */
 
-        /* prefetch src_ptr data to cache memory */
-        prefetch_load(src_ptr + 16);
+    /* prefetch src_ptr data to cache memory */
+    prefetch_load(src_ptr + 16);
 
-        /* unroll for loop */
-        for (i = 16; i--;)
-        {
-            /* apply filter with vectors pairs */
-            __asm__ __volatile__ (
-                "lbu            %[src_ptr_l2],  -32(%[src_ptr])                 \n\t"
-                "lbu            %[src_ptr_l1],  -16(%[src_ptr])                 \n\t"
-                "lbu            %[src_ptr_0],   0(%[src_ptr])                   \n\t"
-                "lbu            %[src_ptr_r1],  16(%[src_ptr])                  \n\t"
-                "lbu            %[src_ptr_r2],  32(%[src_ptr])                  \n\t"
-                "lbu            %[src_ptr_r3],  48(%[src_ptr])                  \n\t"
-                "mtlo           %[vector4a],    $ac2                            \n\t"
+    /* unroll for loop */
+    for (i = 16; i--;) {
+      /* apply filter with vectors pairs */
+      __asm__ __volatile__(
+          "lbu            %[src_ptr_l2],  -32(%[src_ptr])                 \n\t"
+          "lbu            %[src_ptr_l1],  -16(%[src_ptr])                 \n\t"
+          "lbu            %[src_ptr_0],   0(%[src_ptr])                   \n\t"
+          "lbu            %[src_ptr_r1],  16(%[src_ptr])                  \n\t"
+          "lbu            %[src_ptr_r2],  32(%[src_ptr])                  \n\t"
+          "lbu            %[src_ptr_r3],  48(%[src_ptr])                  \n\t"
+          "mtlo           %[vector4a],    $ac2                            \n\t"
 
-                "append         %[src_ptr_l2],  %[src_ptr_r3],  8               \n\t"
-                "append         %[src_ptr_0],   %[src_ptr_r1],  8               \n\t"
-                "append         %[src_ptr_l1],  %[src_ptr_r2],  8               \n\t"
-                "dpau.h.qbr     $ac2,           %[src_ptr_l2],  %[vector1b]     \n\t"
-                "dpau.h.qbr     $ac2,           %[src_ptr_0],   %[vector2b]     \n\t"
-                "dpsu.h.qbr     $ac2,           %[src_ptr_l1],  %[vector3b]     \n\t"
+          "append         %[src_ptr_l2],  %[src_ptr_r3],  8               \n\t"
+          "append         %[src_ptr_0],   %[src_ptr_r1],  8               \n\t"
+          "append         %[src_ptr_l1],  %[src_ptr_r2],  8               \n\t"
+          "dpau.h.qbr     $ac2,           %[src_ptr_l2],  %[vector1b]     \n\t"
+          "dpau.h.qbr     $ac2,           %[src_ptr_0],   %[vector2b]     \n\t"
+          "dpsu.h.qbr     $ac2,           %[src_ptr_l1],  %[vector3b]     \n\t"
 
-                "lbu            %[src_ptr_l2],  -31(%[src_ptr])                 \n\t"
-                "lbu            %[src_ptr_l1],  -15(%[src_ptr])                 \n\t"
-                "lbu            %[src_ptr_0],   1(%[src_ptr])                   \n\t"
-                "lbu            %[src_ptr_r1],  17(%[src_ptr])                  \n\t"
-                "lbu            %[src_ptr_r2],  33(%[src_ptr])                  \n\t"
-                "lbu            %[src_ptr_r3],  49(%[src_ptr])                  \n\t"
-                "mtlo           %[vector4a],    $ac0                            \n\t"
-                "extp           %[Temp1],       $ac2,           9               \n\t"
+          "lbu            %[src_ptr_l2],  -31(%[src_ptr])                 \n\t"
+          "lbu            %[src_ptr_l1],  -15(%[src_ptr])                 \n\t"
+          "lbu            %[src_ptr_0],   1(%[src_ptr])                   \n\t"
+          "lbu            %[src_ptr_r1],  17(%[src_ptr])                  \n\t"
+          "lbu            %[src_ptr_r2],  33(%[src_ptr])                  \n\t"
+          "lbu            %[src_ptr_r3],  49(%[src_ptr])                  \n\t"
+          "mtlo           %[vector4a],    $ac0                            \n\t"
+          "extp           %[Temp1],       $ac2,           9               \n\t"
 
-                "append         %[src_ptr_l2],  %[src_ptr_r3],  8               \n\t"
-                "append         %[src_ptr_0],   %[src_ptr_r1],  8               \n\t"
-                "append         %[src_ptr_l1],  %[src_ptr_r2],  8               \n\t"
-                "dpau.h.qbr     $ac0,           %[src_ptr_l2],  %[vector1b]     \n\t"
-                "dpau.h.qbr     $ac0,           %[src_ptr_0],   %[vector2b]     \n\t"
-                "dpsu.h.qbr     $ac0,           %[src_ptr_l1],  %[vector3b]     \n\t"
+          "append         %[src_ptr_l2],  %[src_ptr_r3],  8               \n\t"
+          "append         %[src_ptr_0],   %[src_ptr_r1],  8               \n\t"
+          "append         %[src_ptr_l1],  %[src_ptr_r2],  8               \n\t"
+          "dpau.h.qbr     $ac0,           %[src_ptr_l2],  %[vector1b]     \n\t"
+          "dpau.h.qbr     $ac0,           %[src_ptr_0],   %[vector2b]     \n\t"
+          "dpsu.h.qbr     $ac0,           %[src_ptr_l1],  %[vector3b]     \n\t"
 
-                "lbu            %[src_ptr_l2],  -30(%[src_ptr])                 \n\t"
-                "lbu            %[src_ptr_l1],  -14(%[src_ptr])                 \n\t"
-                "lbu            %[src_ptr_0],   2(%[src_ptr])                   \n\t"
-                "lbu            %[src_ptr_r1],  18(%[src_ptr])                  \n\t"
-                "lbu            %[src_ptr_r2],  34(%[src_ptr])                  \n\t"
-                "lbu            %[src_ptr_r3],  50(%[src_ptr])                  \n\t"
-                "mtlo           %[vector4a],    $ac1                            \n\t"
-                "extp           %[Temp2],       $ac0,           9               \n\t"
+          "lbu            %[src_ptr_l2],  -30(%[src_ptr])                 \n\t"
+          "lbu            %[src_ptr_l1],  -14(%[src_ptr])                 \n\t"
+          "lbu            %[src_ptr_0],   2(%[src_ptr])                   \n\t"
+          "lbu            %[src_ptr_r1],  18(%[src_ptr])                  \n\t"
+          "lbu            %[src_ptr_r2],  34(%[src_ptr])                  \n\t"
+          "lbu            %[src_ptr_r3],  50(%[src_ptr])                  \n\t"
+          "mtlo           %[vector4a],    $ac1                            \n\t"
+          "extp           %[Temp2],       $ac0,           9               \n\t"
 
-                "append         %[src_ptr_l2],  %[src_ptr_r3],  8               \n\t"
-                "append         %[src_ptr_0],   %[src_ptr_r1],  8               \n\t"
-                "append         %[src_ptr_l1],  %[src_ptr_r2],  8               \n\t"
-                "dpau.h.qbr     $ac1,           %[src_ptr_l2],  %[vector1b]     \n\t"
-                "dpau.h.qbr     $ac1,           %[src_ptr_0],   %[vector2b]     \n\t"
-                "dpsu.h.qbr     $ac1,           %[src_ptr_l1],  %[vector3b]     \n\t"
+          "append         %[src_ptr_l2],  %[src_ptr_r3],  8               \n\t"
+          "append         %[src_ptr_0],   %[src_ptr_r1],  8               \n\t"
+          "append         %[src_ptr_l1],  %[src_ptr_r2],  8               \n\t"
+          "dpau.h.qbr     $ac1,           %[src_ptr_l2],  %[vector1b]     \n\t"
+          "dpau.h.qbr     $ac1,           %[src_ptr_0],   %[vector2b]     \n\t"
+          "dpsu.h.qbr     $ac1,           %[src_ptr_l1],  %[vector3b]     \n\t"
 
-                "lbu            %[src_ptr_l2],  -29(%[src_ptr])                 \n\t"
-                "lbu            %[src_ptr_l1],  -13(%[src_ptr])                 \n\t"
-                "lbu            %[src_ptr_0],   3(%[src_ptr])                   \n\t"
-                "lbu            %[src_ptr_r1],  19(%[src_ptr])                  \n\t"
-                "lbu            %[src_ptr_r2],  35(%[src_ptr])                  \n\t"
-                "lbu            %[src_ptr_r3],  51(%[src_ptr])                  \n\t"
-                "mtlo           %[vector4a],    $ac3                            \n\t"
-                "extp           %[Temp3],       $ac1,           9               \n\t"
+          "lbu            %[src_ptr_l2],  -29(%[src_ptr])                 \n\t"
+          "lbu            %[src_ptr_l1],  -13(%[src_ptr])                 \n\t"
+          "lbu            %[src_ptr_0],   3(%[src_ptr])                   \n\t"
+          "lbu            %[src_ptr_r1],  19(%[src_ptr])                  \n\t"
+          "lbu            %[src_ptr_r2],  35(%[src_ptr])                  \n\t"
+          "lbu            %[src_ptr_r3],  51(%[src_ptr])                  \n\t"
+          "mtlo           %[vector4a],    $ac3                            \n\t"
+          "extp           %[Temp3],       $ac1,           9               \n\t"
 
-                "append         %[src_ptr_l2],  %[src_ptr_r3],  8               \n\t"
-                "append         %[src_ptr_0],   %[src_ptr_r1],  8               \n\t"
-                "append         %[src_ptr_l1],  %[src_ptr_r2],  8               \n\t"
-                "dpau.h.qbr     $ac3,           %[src_ptr_l2],  %[vector1b]     \n\t"
-                "dpau.h.qbr     $ac3,           %[src_ptr_0],   %[vector2b]     \n\t"
-                "dpsu.h.qbr     $ac3,           %[src_ptr_l1],  %[vector3b]     \n\t"
+          "append         %[src_ptr_l2],  %[src_ptr_r3],  8               \n\t"
+          "append         %[src_ptr_0],   %[src_ptr_r1],  8               \n\t"
+          "append         %[src_ptr_l1],  %[src_ptr_r2],  8               \n\t"
+          "dpau.h.qbr     $ac3,           %[src_ptr_l2],  %[vector1b]     \n\t"
+          "dpau.h.qbr     $ac3,           %[src_ptr_0],   %[vector2b]     \n\t"
+          "dpsu.h.qbr     $ac3,           %[src_ptr_l1],  %[vector3b]     \n\t"
 
-                "lbu            %[src_ptr_l2],  -28(%[src_ptr])                 \n\t"
-                "lbu            %[src_ptr_l1],  -12(%[src_ptr])                 \n\t"
-                "lbu            %[src_ptr_0],   4(%[src_ptr])                   \n\t"
-                "lbu            %[src_ptr_r1],  20(%[src_ptr])                  \n\t"
-                "lbu            %[src_ptr_r2],  36(%[src_ptr])                  \n\t"
-                "lbu            %[src_ptr_r3],  52(%[src_ptr])                  \n\t"
-                "mtlo           %[vector4a],    $ac2                            \n\t"
-                "extp           %[Temp4],       $ac3,           9               \n\t"
+          "lbu            %[src_ptr_l2],  -28(%[src_ptr])                 \n\t"
+          "lbu            %[src_ptr_l1],  -12(%[src_ptr])                 \n\t"
+          "lbu            %[src_ptr_0],   4(%[src_ptr])                   \n\t"
+          "lbu            %[src_ptr_r1],  20(%[src_ptr])                  \n\t"
+          "lbu            %[src_ptr_r2],  36(%[src_ptr])                  \n\t"
+          "lbu            %[src_ptr_r3],  52(%[src_ptr])                  \n\t"
+          "mtlo           %[vector4a],    $ac2                            \n\t"
+          "extp           %[Temp4],       $ac3,           9               \n\t"
 
-                "append         %[src_ptr_l2],  %[src_ptr_r3],  8               \n\t"
-                "append         %[src_ptr_0],   %[src_ptr_r1],  8               \n\t"
-                "append         %[src_ptr_l1],  %[src_ptr_r2],  8               \n\t"
-                "dpau.h.qbr     $ac2,           %[src_ptr_l2],  %[vector1b]     \n\t"
-                "dpau.h.qbr     $ac2,           %[src_ptr_0],   %[vector2b]     \n\t"
-                "dpsu.h.qbr     $ac2,           %[src_ptr_l1],  %[vector3b]     \n\t"
+          "append         %[src_ptr_l2],  %[src_ptr_r3],  8               \n\t"
+          "append         %[src_ptr_0],   %[src_ptr_r1],  8               \n\t"
+          "append         %[src_ptr_l1],  %[src_ptr_r2],  8               \n\t"
+          "dpau.h.qbr     $ac2,           %[src_ptr_l2],  %[vector1b]     \n\t"
+          "dpau.h.qbr     $ac2,           %[src_ptr_0],   %[vector2b]     \n\t"
+          "dpsu.h.qbr     $ac2,           %[src_ptr_l1],  %[vector3b]     \n\t"
 
-                "lbu            %[src_ptr_l2],  -27(%[src_ptr])                 \n\t"
-                "lbu            %[src_ptr_l1],  -11(%[src_ptr])                 \n\t"
-                "lbu            %[src_ptr_0],   5(%[src_ptr])                   \n\t"
-                "lbu            %[src_ptr_r1],  21(%[src_ptr])                  \n\t"
-                "lbu            %[src_ptr_r2],  37(%[src_ptr])                  \n\t"
-                "lbu            %[src_ptr_r3],  53(%[src_ptr])                  \n\t"
-                "mtlo           %[vector4a],    $ac0                            \n\t"
-                "extp           %[Temp5],       $ac2,           9               \n\t"
+          "lbu            %[src_ptr_l2],  -27(%[src_ptr])                 \n\t"
+          "lbu            %[src_ptr_l1],  -11(%[src_ptr])                 \n\t"
+          "lbu            %[src_ptr_0],   5(%[src_ptr])                   \n\t"
+          "lbu            %[src_ptr_r1],  21(%[src_ptr])                  \n\t"
+          "lbu            %[src_ptr_r2],  37(%[src_ptr])                  \n\t"
+          "lbu            %[src_ptr_r3],  53(%[src_ptr])                  \n\t"
+          "mtlo           %[vector4a],    $ac0                            \n\t"
+          "extp           %[Temp5],       $ac2,           9               \n\t"
 
-                "append         %[src_ptr_l2],  %[src_ptr_r3],  8               \n\t"
-                "append         %[src_ptr_0],   %[src_ptr_r1],  8               \n\t"
-                "append         %[src_ptr_l1],  %[src_ptr_r2],  8               \n\t"
-                "dpau.h.qbr     $ac0,           %[src_ptr_l2],  %[vector1b]     \n\t"
-                "dpau.h.qbr     $ac0,           %[src_ptr_0],   %[vector2b]     \n\t"
-                "dpsu.h.qbr     $ac0,           %[src_ptr_l1],  %[vector3b]     \n\t"
+          "append         %[src_ptr_l2],  %[src_ptr_r3],  8               \n\t"
+          "append         %[src_ptr_0],   %[src_ptr_r1],  8               \n\t"
+          "append         %[src_ptr_l1],  %[src_ptr_r2],  8               \n\t"
+          "dpau.h.qbr     $ac0,           %[src_ptr_l2],  %[vector1b]     \n\t"
+          "dpau.h.qbr     $ac0,           %[src_ptr_0],   %[vector2b]     \n\t"
+          "dpsu.h.qbr     $ac0,           %[src_ptr_l1],  %[vector3b]     \n\t"
 
-                "lbu            %[src_ptr_l2],  -26(%[src_ptr])                 \n\t"
-                "lbu            %[src_ptr_l1],  -10(%[src_ptr])                 \n\t"
-                "lbu            %[src_ptr_0],   6(%[src_ptr])                   \n\t"
-                "lbu            %[src_ptr_r1],  22(%[src_ptr])                  \n\t"
-                "lbu            %[src_ptr_r2],  38(%[src_ptr])                  \n\t"
-                "lbu            %[src_ptr_r3],  54(%[src_ptr])                  \n\t"
-                "mtlo           %[vector4a],    $ac1                            \n\t"
-                "extp           %[Temp6],       $ac0,           9               \n\t"
+          "lbu            %[src_ptr_l2],  -26(%[src_ptr])                 \n\t"
+          "lbu            %[src_ptr_l1],  -10(%[src_ptr])                 \n\t"
+          "lbu            %[src_ptr_0],   6(%[src_ptr])                   \n\t"
+          "lbu            %[src_ptr_r1],  22(%[src_ptr])                  \n\t"
+          "lbu            %[src_ptr_r2],  38(%[src_ptr])                  \n\t"
+          "lbu            %[src_ptr_r3],  54(%[src_ptr])                  \n\t"
+          "mtlo           %[vector4a],    $ac1                            \n\t"
+          "extp           %[Temp6],       $ac0,           9               \n\t"
 
-                "append         %[src_ptr_l2],  %[src_ptr_r3],  8               \n\t"
-                "append         %[src_ptr_0],   %[src_ptr_r1],  8               \n\t"
-                "append         %[src_ptr_l1],  %[src_ptr_r2],  8               \n\t"
-                "dpau.h.qbr     $ac1,           %[src_ptr_l2],  %[vector1b]     \n\t"
-                "dpau.h.qbr     $ac1,           %[src_ptr_0],   %[vector2b]     \n\t"
-                "dpsu.h.qbr     $ac1,           %[src_ptr_l1],  %[vector3b]     \n\t"
+          "append         %[src_ptr_l2],  %[src_ptr_r3],  8               \n\t"
+          "append         %[src_ptr_0],   %[src_ptr_r1],  8               \n\t"
+          "append         %[src_ptr_l1],  %[src_ptr_r2],  8               \n\t"
+          "dpau.h.qbr     $ac1,           %[src_ptr_l2],  %[vector1b]     \n\t"
+          "dpau.h.qbr     $ac1,           %[src_ptr_0],   %[vector2b]     \n\t"
+          "dpsu.h.qbr     $ac1,           %[src_ptr_l1],  %[vector3b]     \n\t"
 
-                "lbu            %[src_ptr_l2],  -25(%[src_ptr])                 \n\t"
-                "lbu            %[src_ptr_l1],  -9(%[src_ptr])                  \n\t"
-                "lbu            %[src_ptr_0],   7(%[src_ptr])                   \n\t"
-                "lbu            %[src_ptr_r1],  23(%[src_ptr])                  \n\t"
-                "lbu            %[src_ptr_r2],  39(%[src_ptr])                  \n\t"
-                "lbu            %[src_ptr_r3],  55(%[src_ptr])                  \n\t"
-                "mtlo           %[vector4a],    $ac3                            \n\t"
-                "extp           %[Temp7],       $ac1,           9               \n\t"
+          "lbu            %[src_ptr_l2],  -25(%[src_ptr])                 \n\t"
+          "lbu            %[src_ptr_l1],  -9(%[src_ptr])                  \n\t"
+          "lbu            %[src_ptr_0],   7(%[src_ptr])                   \n\t"
+          "lbu            %[src_ptr_r1],  23(%[src_ptr])                  \n\t"
+          "lbu            %[src_ptr_r2],  39(%[src_ptr])                  \n\t"
+          "lbu            %[src_ptr_r3],  55(%[src_ptr])                  \n\t"
+          "mtlo           %[vector4a],    $ac3                            \n\t"
+          "extp           %[Temp7],       $ac1,           9               \n\t"
 
-                "append         %[src_ptr_l2],  %[src_ptr_r3],  8               \n\t"
-                "append         %[src_ptr_0],   %[src_ptr_r1],  8               \n\t"
-                "append         %[src_ptr_l1],  %[src_ptr_r2],  8               \n\t"
-                "dpau.h.qbr     $ac3,           %[src_ptr_l2],  %[vector1b]     \n\t"
-                "dpau.h.qbr     $ac3,           %[src_ptr_0],   %[vector2b]     \n\t"
-                "dpsu.h.qbr     $ac3,           %[src_ptr_l1],  %[vector3b]     \n\t"
-                "extp           %[Temp8],       $ac3,           9               \n\t"
+          "append         %[src_ptr_l2],  %[src_ptr_r3],  8               \n\t"
+          "append         %[src_ptr_0],   %[src_ptr_r1],  8               \n\t"
+          "append         %[src_ptr_l1],  %[src_ptr_r2],  8               \n\t"
+          "dpau.h.qbr     $ac3,           %[src_ptr_l2],  %[vector1b]     \n\t"
+          "dpau.h.qbr     $ac3,           %[src_ptr_0],   %[vector2b]     \n\t"
+          "dpsu.h.qbr     $ac3,           %[src_ptr_l1],  %[vector3b]     \n\t"
+          "extp           %[Temp8],       $ac3,           9               \n\t"
 
-                : [Temp1] "=&r" (Temp1), [Temp2] "=&r" (Temp2),
-                  [Temp3] "=&r" (Temp3), [Temp4] "=&r" (Temp4),
-                  [Temp5] "=&r" (Temp5), [Temp6] "=&r" (Temp6),
-                  [Temp7] "=&r" (Temp7), [Temp8] "=r" (Temp8),
-                  [src_ptr_l1] "=&r" (src_ptr_l1), [src_ptr_0] "=&r" (src_ptr_0),
-                  [src_ptr_r1] "=&r" (src_ptr_r1), [src_ptr_r2] "=&r" (src_ptr_r2),
-                  [src_ptr_l2] "=&r" (src_ptr_l2),[src_ptr_r3] "=&r" (src_ptr_r3)
-                : [vector1b] "r" (vector1b), [vector2b] "r" (vector2b),
-                  [vector3b] "r" (vector3b), [vector4a] "r" (vector4a),
-                  [src_ptr] "r" (src_ptr)
-            );
+          : [Temp1] "=&r"(Temp1), [Temp2] "=&r"(Temp2), [Temp3] "=&r"(Temp3),
+            [Temp4] "=&r"(Temp4), [Temp5] "=&r"(Temp5), [Temp6] "=&r"(Temp6),
+            [Temp7] "=&r"(Temp7), [Temp8] "=r"(Temp8),
+            [src_ptr_l1] "=&r"(src_ptr_l1), [src_ptr_0] "=&r"(src_ptr_0),
+            [src_ptr_r1] "=&r"(src_ptr_r1), [src_ptr_r2] "=&r"(src_ptr_r2),
+            [src_ptr_l2] "=&r"(src_ptr_l2), [src_ptr_r3] "=&r"(src_ptr_r3)
+          : [vector1b] "r"(vector1b), [vector2b] "r"(vector2b),
+            [vector3b] "r"(vector3b), [vector4a] "r"(vector4a),
+            [src_ptr] "r"(src_ptr));
 
-            /* clamp and store results */
-            output_ptr[0] = cm[Temp1];
-            output_ptr[1] = cm[Temp2];
-            output_ptr[2] = cm[Temp3];
-            output_ptr[3] = cm[Temp4];
-            output_ptr[4] = cm[Temp5];
-            output_ptr[5] = cm[Temp6];
-            output_ptr[6] = cm[Temp7];
-            output_ptr[7] = cm[Temp8];
+      /* clamp and store results */
+      output_ptr[0] = cm[Temp1];
+      output_ptr[1] = cm[Temp2];
+      output_ptr[2] = cm[Temp3];
+      output_ptr[3] = cm[Temp4];
+      output_ptr[4] = cm[Temp5];
+      output_ptr[5] = cm[Temp6];
+      output_ptr[6] = cm[Temp7];
+      output_ptr[7] = cm[Temp8];
 
-            /* apply filter with vectors pairs */
-            __asm__ __volatile__ (
-                "lbu            %[src_ptr_l2],  -24(%[src_ptr])                 \n\t"
-                "lbu            %[src_ptr_l1],  -8(%[src_ptr])                  \n\t"
-                "lbu            %[src_ptr_0],   8(%[src_ptr])                   \n\t"
-                "lbu            %[src_ptr_r1],  24(%[src_ptr])                  \n\t"
-                "lbu            %[src_ptr_r2],  40(%[src_ptr])                  \n\t"
-                "lbu            %[src_ptr_r3],  56(%[src_ptr])                  \n\t"
-                "mtlo           %[vector4a],    $ac2                            \n\t"
+      /* apply filter with vectors pairs */
+      __asm__ __volatile__(
+          "lbu            %[src_ptr_l2],  -24(%[src_ptr])                 \n\t"
+          "lbu            %[src_ptr_l1],  -8(%[src_ptr])                  \n\t"
+          "lbu            %[src_ptr_0],   8(%[src_ptr])                   \n\t"
+          "lbu            %[src_ptr_r1],  24(%[src_ptr])                  \n\t"
+          "lbu            %[src_ptr_r2],  40(%[src_ptr])                  \n\t"
+          "lbu            %[src_ptr_r3],  56(%[src_ptr])                  \n\t"
+          "mtlo           %[vector4a],    $ac2                            \n\t"
 
-                "append         %[src_ptr_l2],  %[src_ptr_r3],  8               \n\t"
-                "append         %[src_ptr_0],   %[src_ptr_r1],  8               \n\t"
-                "append         %[src_ptr_l1],  %[src_ptr_r2],  8               \n\t"
-                "dpau.h.qbr     $ac2,           %[src_ptr_l2],  %[vector1b]     \n\t"
-                "dpau.h.qbr     $ac2,           %[src_ptr_0],   %[vector2b]     \n\t"
-                "dpsu.h.qbr     $ac2,           %[src_ptr_l1],  %[vector3b]     \n\t"
+          "append         %[src_ptr_l2],  %[src_ptr_r3],  8               \n\t"
+          "append         %[src_ptr_0],   %[src_ptr_r1],  8               \n\t"
+          "append         %[src_ptr_l1],  %[src_ptr_r2],  8               \n\t"
+          "dpau.h.qbr     $ac2,           %[src_ptr_l2],  %[vector1b]     \n\t"
+          "dpau.h.qbr     $ac2,           %[src_ptr_0],   %[vector2b]     \n\t"
+          "dpsu.h.qbr     $ac2,           %[src_ptr_l1],  %[vector3b]     \n\t"
 
-                "lbu            %[src_ptr_l2],  -23(%[src_ptr])                 \n\t"
-                "lbu            %[src_ptr_l1],  -7(%[src_ptr])                  \n\t"
-                "lbu            %[src_ptr_0],   9(%[src_ptr])                   \n\t"
-                "lbu            %[src_ptr_r1],  25(%[src_ptr])                  \n\t"
-                "lbu            %[src_ptr_r2],  41(%[src_ptr])                  \n\t"
-                "lbu            %[src_ptr_r3],  57(%[src_ptr])                  \n\t"
-                "mtlo           %[vector4a],    $ac0                            \n\t"
-                "extp           %[Temp1],       $ac2,           9               \n\t"
+          "lbu            %[src_ptr_l2],  -23(%[src_ptr])                 \n\t"
+          "lbu            %[src_ptr_l1],  -7(%[src_ptr])                  \n\t"
+          "lbu            %[src_ptr_0],   9(%[src_ptr])                   \n\t"
+          "lbu            %[src_ptr_r1],  25(%[src_ptr])                  \n\t"
+          "lbu            %[src_ptr_r2],  41(%[src_ptr])                  \n\t"
+          "lbu            %[src_ptr_r3],  57(%[src_ptr])                  \n\t"
+          "mtlo           %[vector4a],    $ac0                            \n\t"
+          "extp           %[Temp1],       $ac2,           9               \n\t"
 
-                "append         %[src_ptr_l2],  %[src_ptr_r3],  8               \n\t"
-                "append         %[src_ptr_0],   %[src_ptr_r1],  8               \n\t"
-                "append         %[src_ptr_l1],  %[src_ptr_r2],  8               \n\t"
-                "dpau.h.qbr     $ac0,           %[src_ptr_l2],  %[vector1b]     \n\t"
-                "dpau.h.qbr     $ac0,           %[src_ptr_0],   %[vector2b]     \n\t"
-                "dpsu.h.qbr     $ac0,           %[src_ptr_l1],  %[vector3b]     \n\t"
+          "append         %[src_ptr_l2],  %[src_ptr_r3],  8               \n\t"
+          "append         %[src_ptr_0],   %[src_ptr_r1],  8               \n\t"
+          "append         %[src_ptr_l1],  %[src_ptr_r2],  8               \n\t"
+          "dpau.h.qbr     $ac0,           %[src_ptr_l2],  %[vector1b]     \n\t"
+          "dpau.h.qbr     $ac0,           %[src_ptr_0],   %[vector2b]     \n\t"
+          "dpsu.h.qbr     $ac0,           %[src_ptr_l1],  %[vector3b]     \n\t"
 
-                "lbu            %[src_ptr_l2],  -22(%[src_ptr])                 \n\t"
-                "lbu            %[src_ptr_l1],  -6(%[src_ptr])                  \n\t"
-                "lbu            %[src_ptr_0],   10(%[src_ptr])                  \n\t"
-                "lbu            %[src_ptr_r1],  26(%[src_ptr])                  \n\t"
-                "lbu            %[src_ptr_r2],  42(%[src_ptr])                  \n\t"
-                "lbu            %[src_ptr_r3],  58(%[src_ptr])                  \n\t"
-                "mtlo           %[vector4a],    $ac1                            \n\t"
-                "extp           %[Temp2],       $ac0,           9               \n\t"
+          "lbu            %[src_ptr_l2],  -22(%[src_ptr])                 \n\t"
+          "lbu            %[src_ptr_l1],  -6(%[src_ptr])                  \n\t"
+          "lbu            %[src_ptr_0],   10(%[src_ptr])                  \n\t"
+          "lbu            %[src_ptr_r1],  26(%[src_ptr])                  \n\t"
+          "lbu            %[src_ptr_r2],  42(%[src_ptr])                  \n\t"
+          "lbu            %[src_ptr_r3],  58(%[src_ptr])                  \n\t"
+          "mtlo           %[vector4a],    $ac1                            \n\t"
+          "extp           %[Temp2],       $ac0,           9               \n\t"
 
-                "append         %[src_ptr_l2],  %[src_ptr_r3],  8               \n\t"
-                "append         %[src_ptr_0],   %[src_ptr_r1],  8               \n\t"
-                "append         %[src_ptr_l1],  %[src_ptr_r2],  8               \n\t"
-                "dpau.h.qbr     $ac1,           %[src_ptr_l2],  %[vector1b]     \n\t"
-                "dpau.h.qbr     $ac1,           %[src_ptr_0],   %[vector2b]     \n\t"
-                "dpsu.h.qbr     $ac1,           %[src_ptr_l1],  %[vector3b]     \n\t"
+          "append         %[src_ptr_l2],  %[src_ptr_r3],  8               \n\t"
+          "append         %[src_ptr_0],   %[src_ptr_r1],  8               \n\t"
+          "append         %[src_ptr_l1],  %[src_ptr_r2],  8               \n\t"
+          "dpau.h.qbr     $ac1,           %[src_ptr_l2],  %[vector1b]     \n\t"
+          "dpau.h.qbr     $ac1,           %[src_ptr_0],   %[vector2b]     \n\t"
+          "dpsu.h.qbr     $ac1,           %[src_ptr_l1],  %[vector3b]     \n\t"
 
-                "lbu            %[src_ptr_l2],  -21(%[src_ptr])                 \n\t"
-                "lbu            %[src_ptr_l1],  -5(%[src_ptr])                  \n\t"
-                "lbu            %[src_ptr_0],   11(%[src_ptr])                  \n\t"
-                "lbu            %[src_ptr_r1],  27(%[src_ptr])                  \n\t"
-                "lbu            %[src_ptr_r2],  43(%[src_ptr])                  \n\t"
-                "lbu            %[src_ptr_r3],  59(%[src_ptr])                  \n\t"
-                "mtlo           %[vector4a],    $ac3                            \n\t"
-                "extp           %[Temp3],       $ac1,           9               \n\t"
+          "lbu            %[src_ptr_l2],  -21(%[src_ptr])                 \n\t"
+          "lbu            %[src_ptr_l1],  -5(%[src_ptr])                  \n\t"
+          "lbu            %[src_ptr_0],   11(%[src_ptr])                  \n\t"
+          "lbu            %[src_ptr_r1],  27(%[src_ptr])                  \n\t"
+          "lbu            %[src_ptr_r2],  43(%[src_ptr])                  \n\t"
+          "lbu            %[src_ptr_r3],  59(%[src_ptr])                  \n\t"
+          "mtlo           %[vector4a],    $ac3                            \n\t"
+          "extp           %[Temp3],       $ac1,           9               \n\t"
 
-                "append         %[src_ptr_l2],  %[src_ptr_r3],  8               \n\t"
-                "append         %[src_ptr_0],   %[src_ptr_r1],  8               \n\t"
-                "append         %[src_ptr_l1],  %[src_ptr_r2],  8               \n\t"
-                "dpau.h.qbr     $ac3,           %[src_ptr_l2],  %[vector1b]     \n\t"
-                "dpau.h.qbr     $ac3,           %[src_ptr_0],   %[vector2b]     \n\t"
-                "dpsu.h.qbr     $ac3,           %[src_ptr_l1],  %[vector3b]     \n\t"
+          "append         %[src_ptr_l2],  %[src_ptr_r3],  8               \n\t"
+          "append         %[src_ptr_0],   %[src_ptr_r1],  8               \n\t"
+          "append         %[src_ptr_l1],  %[src_ptr_r2],  8               \n\t"
+          "dpau.h.qbr     $ac3,           %[src_ptr_l2],  %[vector1b]     \n\t"
+          "dpau.h.qbr     $ac3,           %[src_ptr_0],   %[vector2b]     \n\t"
+          "dpsu.h.qbr     $ac3,           %[src_ptr_l1],  %[vector3b]     \n\t"
 
-                "lbu            %[src_ptr_l2],  -20(%[src_ptr])                 \n\t"
-                "lbu            %[src_ptr_l1],  -4(%[src_ptr])                  \n\t"
-                "lbu            %[src_ptr_0],   12(%[src_ptr])                  \n\t"
-                "lbu            %[src_ptr_r1],  28(%[src_ptr])                  \n\t"
-                "lbu            %[src_ptr_r2],  44(%[src_ptr])                  \n\t"
-                "lbu            %[src_ptr_r3],  60(%[src_ptr])                  \n\t"
-                "mtlo           %[vector4a],    $ac2                            \n\t"
-                "extp           %[Temp4],       $ac3,           9               \n\t"
+          "lbu            %[src_ptr_l2],  -20(%[src_ptr])                 \n\t"
+          "lbu            %[src_ptr_l1],  -4(%[src_ptr])                  \n\t"
+          "lbu            %[src_ptr_0],   12(%[src_ptr])                  \n\t"
+          "lbu            %[src_ptr_r1],  28(%[src_ptr])                  \n\t"
+          "lbu            %[src_ptr_r2],  44(%[src_ptr])                  \n\t"
+          "lbu            %[src_ptr_r3],  60(%[src_ptr])                  \n\t"
+          "mtlo           %[vector4a],    $ac2                            \n\t"
+          "extp           %[Temp4],       $ac3,           9               \n\t"
 
-                "append         %[src_ptr_l2],  %[src_ptr_r3],  8               \n\t"
-                "append         %[src_ptr_0],   %[src_ptr_r1],  8               \n\t"
-                "append         %[src_ptr_l1],  %[src_ptr_r2],  8               \n\t"
-                "dpau.h.qbr     $ac2,           %[src_ptr_l2],  %[vector1b]     \n\t"
-                "dpau.h.qbr     $ac2,           %[src_ptr_0],   %[vector2b]     \n\t"
-                "dpsu.h.qbr     $ac2,           %[src_ptr_l1],  %[vector3b]     \n\t"
+          "append         %[src_ptr_l2],  %[src_ptr_r3],  8               \n\t"
+          "append         %[src_ptr_0],   %[src_ptr_r1],  8               \n\t"
+          "append         %[src_ptr_l1],  %[src_ptr_r2],  8               \n\t"
+          "dpau.h.qbr     $ac2,           %[src_ptr_l2],  %[vector1b]     \n\t"
+          "dpau.h.qbr     $ac2,           %[src_ptr_0],   %[vector2b]     \n\t"
+          "dpsu.h.qbr     $ac2,           %[src_ptr_l1],  %[vector3b]     \n\t"
 
-                "lbu            %[src_ptr_l2],  -19(%[src_ptr])                 \n\t"
-                "lbu            %[src_ptr_l1],  -3(%[src_ptr])                  \n\t"
-                "lbu            %[src_ptr_0],   13(%[src_ptr])                  \n\t"
-                "lbu            %[src_ptr_r1],  29(%[src_ptr])                  \n\t"
-                "lbu            %[src_ptr_r2],  45(%[src_ptr])                  \n\t"
-                "lbu            %[src_ptr_r3],  61(%[src_ptr])                  \n\t"
-                "mtlo           %[vector4a],    $ac0                            \n\t"
-                "extp           %[Temp5],       $ac2,           9               \n\t"
+          "lbu            %[src_ptr_l2],  -19(%[src_ptr])                 \n\t"
+          "lbu            %[src_ptr_l1],  -3(%[src_ptr])                  \n\t"
+          "lbu            %[src_ptr_0],   13(%[src_ptr])                  \n\t"
+          "lbu            %[src_ptr_r1],  29(%[src_ptr])                  \n\t"
+          "lbu            %[src_ptr_r2],  45(%[src_ptr])                  \n\t"
+          "lbu            %[src_ptr_r3],  61(%[src_ptr])                  \n\t"
+          "mtlo           %[vector4a],    $ac0                            \n\t"
+          "extp           %[Temp5],       $ac2,           9               \n\t"
 
-                "append         %[src_ptr_l2],  %[src_ptr_r3],  8               \n\t"
-                "append         %[src_ptr_0],   %[src_ptr_r1],  8               \n\t"
-                "append         %[src_ptr_l1],  %[src_ptr_r2],  8               \n\t"
-                "dpau.h.qbr     $ac0,           %[src_ptr_l2],  %[vector1b]     \n\t"
-                "dpau.h.qbr     $ac0,           %[src_ptr_0],   %[vector2b]     \n\t"
-                "dpsu.h.qbr     $ac0,           %[src_ptr_l1],  %[vector3b]     \n\t"
+          "append         %[src_ptr_l2],  %[src_ptr_r3],  8               \n\t"
+          "append         %[src_ptr_0],   %[src_ptr_r1],  8               \n\t"
+          "append         %[src_ptr_l1],  %[src_ptr_r2],  8               \n\t"
+          "dpau.h.qbr     $ac0,           %[src_ptr_l2],  %[vector1b]     \n\t"
+          "dpau.h.qbr     $ac0,           %[src_ptr_0],   %[vector2b]     \n\t"
+          "dpsu.h.qbr     $ac0,           %[src_ptr_l1],  %[vector3b]     \n\t"
 
-                "lbu            %[src_ptr_l2],  -18(%[src_ptr])                 \n\t"
-                "lbu            %[src_ptr_l1],  -2(%[src_ptr])                  \n\t"
-                "lbu            %[src_ptr_0],   14(%[src_ptr])                  \n\t"
-                "lbu            %[src_ptr_r1],  30(%[src_ptr])                  \n\t"
-                "lbu            %[src_ptr_r2],  46(%[src_ptr])                  \n\t"
-                "lbu            %[src_ptr_r3],  62(%[src_ptr])                  \n\t"
-                "mtlo           %[vector4a],    $ac1                            \n\t"
-                "extp           %[Temp6],       $ac0,           9               \n\t"
+          "lbu            %[src_ptr_l2],  -18(%[src_ptr])                 \n\t"
+          "lbu            %[src_ptr_l1],  -2(%[src_ptr])                  \n\t"
+          "lbu            %[src_ptr_0],   14(%[src_ptr])                  \n\t"
+          "lbu            %[src_ptr_r1],  30(%[src_ptr])                  \n\t"
+          "lbu            %[src_ptr_r2],  46(%[src_ptr])                  \n\t"
+          "lbu            %[src_ptr_r3],  62(%[src_ptr])                  \n\t"
+          "mtlo           %[vector4a],    $ac1                            \n\t"
+          "extp           %[Temp6],       $ac0,           9               \n\t"
 
-                "append         %[src_ptr_l2],  %[src_ptr_r3],  8               \n\t"
-                "append         %[src_ptr_0],   %[src_ptr_r1],  8               \n\t"
-                "append         %[src_ptr_l1],  %[src_ptr_r2],  8               \n\t"
-                "dpau.h.qbr     $ac1,           %[src_ptr_l2],  %[vector1b]     \n\t"
-                "dpau.h.qbr     $ac1,           %[src_ptr_0],   %[vector2b]     \n\t"
-                "dpsu.h.qbr     $ac1,           %[src_ptr_l1],  %[vector3b]     \n\t"
+          "append         %[src_ptr_l2],  %[src_ptr_r3],  8               \n\t"
+          "append         %[src_ptr_0],   %[src_ptr_r1],  8               \n\t"
+          "append         %[src_ptr_l1],  %[src_ptr_r2],  8               \n\t"
+          "dpau.h.qbr     $ac1,           %[src_ptr_l2],  %[vector1b]     \n\t"
+          "dpau.h.qbr     $ac1,           %[src_ptr_0],   %[vector2b]     \n\t"
+          "dpsu.h.qbr     $ac1,           %[src_ptr_l1],  %[vector3b]     \n\t"
 
-                "lbu            %[src_ptr_l2],  -17(%[src_ptr])                 \n\t"
-                "lbu            %[src_ptr_l1],  -1(%[src_ptr])                  \n\t"
-                "lbu            %[src_ptr_0],   15(%[src_ptr])                  \n\t"
-                "lbu            %[src_ptr_r1],  31(%[src_ptr])                  \n\t"
-                "lbu            %[src_ptr_r2],  47(%[src_ptr])                  \n\t"
-                "lbu            %[src_ptr_r3],  63(%[src_ptr])                  \n\t"
-                "mtlo           %[vector4a],    $ac3                            \n\t"
-                "extp           %[Temp7],       $ac1,           9               \n\t"
+          "lbu            %[src_ptr_l2],  -17(%[src_ptr])                 \n\t"
+          "lbu            %[src_ptr_l1],  -1(%[src_ptr])                  \n\t"
+          "lbu            %[src_ptr_0],   15(%[src_ptr])                  \n\t"
+          "lbu            %[src_ptr_r1],  31(%[src_ptr])                  \n\t"
+          "lbu            %[src_ptr_r2],  47(%[src_ptr])                  \n\t"
+          "lbu            %[src_ptr_r3],  63(%[src_ptr])                  \n\t"
+          "mtlo           %[vector4a],    $ac3                            \n\t"
+          "extp           %[Temp7],       $ac1,           9               \n\t"
 
-                "append         %[src_ptr_l2],  %[src_ptr_r3],  8               \n\t"
-                "append         %[src_ptr_0],   %[src_ptr_r1],  8               \n\t"
-                "append         %[src_ptr_l1],  %[src_ptr_r2],  8               \n\t"
-                "dpau.h.qbr     $ac3,           %[src_ptr_l2],  %[vector1b]     \n\t"
-                "dpau.h.qbr     $ac3,           %[src_ptr_0],   %[vector2b]     \n\t"
-                "dpsu.h.qbr     $ac3,           %[src_ptr_l1],  %[vector3b]     \n\t"
-                "extp           %[Temp8],       $ac3,           9               \n\t"
+          "append         %[src_ptr_l2],  %[src_ptr_r3],  8               \n\t"
+          "append         %[src_ptr_0],   %[src_ptr_r1],  8               \n\t"
+          "append         %[src_ptr_l1],  %[src_ptr_r2],  8               \n\t"
+          "dpau.h.qbr     $ac3,           %[src_ptr_l2],  %[vector1b]     \n\t"
+          "dpau.h.qbr     $ac3,           %[src_ptr_0],   %[vector2b]     \n\t"
+          "dpsu.h.qbr     $ac3,           %[src_ptr_l1],  %[vector3b]     \n\t"
+          "extp           %[Temp8],       $ac3,           9               \n\t"
 
-                : [Temp1] "=&r" (Temp1), [Temp2] "=&r" (Temp2),
-                  [Temp3] "=&r" (Temp3), [Temp4] "=&r" (Temp4),
-                  [Temp5] "=&r" (Temp5), [Temp6] "=&r" (Temp6),
-                  [Temp7] "=&r" (Temp7), [Temp8] "=r" (Temp8),
-                  [src_ptr_l1] "=&r" (src_ptr_l1), [src_ptr_0] "=&r" (src_ptr_0),
-                  [src_ptr_r1] "=&r" (src_ptr_r1), [src_ptr_r2] "=&r" (src_ptr_r2),
-                  [src_ptr_l2] "=&r" (src_ptr_l2), [src_ptr_r3] "=&r" (src_ptr_r3)
-                : [vector1b] "r" (vector1b), [vector2b] "r" (vector2b),
-                  [vector3b] "r" (vector3b), [vector4a] "r" (vector4a),
-                  [src_ptr] "r" (src_ptr)
-            );
+          : [Temp1] "=&r"(Temp1), [Temp2] "=&r"(Temp2), [Temp3] "=&r"(Temp3),
+            [Temp4] "=&r"(Temp4), [Temp5] "=&r"(Temp5), [Temp6] "=&r"(Temp6),
+            [Temp7] "=&r"(Temp7), [Temp8] "=r"(Temp8),
+            [src_ptr_l1] "=&r"(src_ptr_l1), [src_ptr_0] "=&r"(src_ptr_0),
+            [src_ptr_r1] "=&r"(src_ptr_r1), [src_ptr_r2] "=&r"(src_ptr_r2),
+            [src_ptr_l2] "=&r"(src_ptr_l2), [src_ptr_r3] "=&r"(src_ptr_r3)
+          : [vector1b] "r"(vector1b), [vector2b] "r"(vector2b),
+            [vector3b] "r"(vector3b), [vector4a] "r"(vector4a),
+            [src_ptr] "r"(src_ptr));
 
-            src_ptr += 16;
-            output_ptr[8] = cm[Temp1];
-            output_ptr[9] = cm[Temp2];
-            output_ptr[10] = cm[Temp3];
-            output_ptr[11] = cm[Temp4];
-            output_ptr[12] = cm[Temp5];
-            output_ptr[13] = cm[Temp6];
-            output_ptr[14] = cm[Temp7];
-            output_ptr[15] = cm[Temp8];
+      src_ptr += 16;
+      output_ptr[8] = cm[Temp1];
+      output_ptr[9] = cm[Temp2];
+      output_ptr[10] = cm[Temp3];
+      output_ptr[11] = cm[Temp4];
+      output_ptr[12] = cm[Temp5];
+      output_ptr[13] = cm[Temp6];
+      output_ptr[14] = cm[Temp7];
+      output_ptr[15] = cm[Temp8];
 
-            output_ptr += output_pitch;
-        }
+      output_ptr += output_pitch;
     }
+  }
 }
 
+void vp8_sixtap_predict4x4_dspr2(unsigned char *RESTRICT src_ptr,
+                                 int src_pixels_per_line, int xoffset,
+                                 int yoffset, unsigned char *RESTRICT dst_ptr,
+                                 int dst_pitch) {
+  unsigned char FData[9 * 4]; /* Temp data bufffer used in filtering */
+  unsigned int pos = 16;
 
-void vp8_sixtap_predict4x4_dspr2
-(
-    unsigned char *RESTRICT src_ptr,
-    int   src_pixels_per_line,
-    int  xoffset,
-    int  yoffset,
-    unsigned char *RESTRICT dst_ptr,
-    int dst_pitch
-)
-{
-    unsigned char FData[9 * 4]; /* Temp data bufffer used in filtering */
-    unsigned int pos = 16;
+  /* bit positon for extract from acc */
+  __asm__ __volatile__("wrdsp      %[pos],     1           \n\t"
+                       :
+                       : [pos] "r"(pos));
 
-    /* bit positon for extract from acc */
-    __asm__ __volatile__ (
-        "wrdsp      %[pos],     1           \n\t"
-        :
-        : [pos] "r" (pos)
-    );
-
-    if (yoffset)
-    {
-        /* First filter 1-D horizontally... */
-        vp8_filter_block2d_first_pass_4(src_ptr - (2 * src_pixels_per_line), FData,
-                                        src_pixels_per_line, 9, xoffset, 4);
-        /* then filter verticaly... */
-        vp8_filter_block2d_second_pass4(FData + 8, dst_ptr, dst_pitch, yoffset);
-    }
-    else
-        /* if (yoffsset == 0) vp8_filter_block2d_first_pass save data to dst_ptr */
-        vp8_filter_block2d_first_pass_4(src_ptr, dst_ptr, src_pixels_per_line,
-                                        4, xoffset, dst_pitch);
+  if (yoffset) {
+    /* First filter 1-D horizontally... */
+    vp8_filter_block2d_first_pass_4(src_ptr - (2 * src_pixels_per_line), FData,
+                                    src_pixels_per_line, 9, xoffset, 4);
+    /* then filter verticaly... */
+    vp8_filter_block2d_second_pass4(FData + 8, dst_ptr, dst_pitch, yoffset);
+  } else
+    /* if (yoffsset == 0) vp8_filter_block2d_first_pass save data to dst_ptr */
+    vp8_filter_block2d_first_pass_4(src_ptr, dst_ptr, src_pixels_per_line, 4,
+                                    xoffset, dst_pitch);
 }
 
+void vp8_sixtap_predict8x8_dspr2(unsigned char *RESTRICT src_ptr,
+                                 int src_pixels_per_line, int xoffset,
+                                 int yoffset, unsigned char *RESTRICT dst_ptr,
+                                 int dst_pitch) {
+  unsigned char FData[13 * 8]; /* Temp data bufffer used in filtering */
+  unsigned int pos, Temp1, Temp2;
 
-void vp8_sixtap_predict8x8_dspr2
-(
-    unsigned char   *RESTRICT src_ptr,
-    int  src_pixels_per_line,
-    int  xoffset,
-    int  yoffset,
-    unsigned char *RESTRICT dst_ptr,
-    int  dst_pitch
-)
-{
+  pos = 16;
 
-    unsigned char FData[13 * 8]; /* Temp data bufffer used in filtering */
-    unsigned int pos, Temp1, Temp2;
+  /* bit positon for extract from acc */
+  __asm__ __volatile__("wrdsp      %[pos],     1               \n\t"
+                       :
+                       : [pos] "r"(pos));
 
-    pos = 16;
+  if (yoffset) {
+    src_ptr = src_ptr - (2 * src_pixels_per_line);
 
-    /* bit positon for extract from acc */
-    __asm__ __volatile__ (
-        "wrdsp      %[pos],     1               \n\t"
-        :
-        : [pos] "r" (pos)
-    );
+    if (xoffset) /* filter 1-D horizontally... */
+      vp8_filter_block2d_first_pass_8_all(src_ptr, FData, src_pixels_per_line,
+                                          13, xoffset, 8);
 
-    if (yoffset)
-    {
+    else {
+      /* prefetch src_ptr data to cache memory */
+      prefetch_load(src_ptr + 2 * src_pixels_per_line);
 
-        src_ptr = src_ptr - (2 * src_pixels_per_line);
+      __asm__ __volatile__(
+          "ulw    %[Temp1],   0(%[src_ptr])                           \n\t"
+          "ulw    %[Temp2],   4(%[src_ptr])                           \n\t"
+          "sw     %[Temp1],   0(%[FData])                             \n\t"
+          "sw     %[Temp2],   4(%[FData])                             \n\t"
+          "addu   %[src_ptr], %[src_ptr],    %[src_pixels_per_line]   \n\t"
 
-        if (xoffset)
-            /* filter 1-D horizontally... */
-            vp8_filter_block2d_first_pass_8_all(src_ptr, FData, src_pixels_per_line,
-                                                13, xoffset, 8);
+          "ulw    %[Temp1],   0(%[src_ptr])                           \n\t"
+          "ulw    %[Temp2],   4(%[src_ptr])                           \n\t"
+          "sw     %[Temp1],   8(%[FData])                             \n\t"
+          "sw     %[Temp2],   12(%[FData])                            \n\t"
+          "addu   %[src_ptr], %[src_ptr],    %[src_pixels_per_line]   \n\t"
 
-        else
-        {
-            /* prefetch src_ptr data to cache memory */
-            prefetch_load(src_ptr + 2 * src_pixels_per_line);
+          "ulw    %[Temp1],   0(%[src_ptr])                           \n\t"
+          "ulw    %[Temp2],   4(%[src_ptr])                           \n\t"
+          "sw     %[Temp1],   16(%[FData])                            \n\t"
+          "sw     %[Temp2],   20(%[FData])                            \n\t"
+          "addu   %[src_ptr], %[src_ptr],    %[src_pixels_per_line]   \n\t"
 
-            __asm__ __volatile__ (
-                "ulw    %[Temp1],   0(%[src_ptr])                           \n\t"
-                "ulw    %[Temp2],   4(%[src_ptr])                           \n\t"
-                "sw     %[Temp1],   0(%[FData])                             \n\t"
-                "sw     %[Temp2],   4(%[FData])                             \n\t"
-                "addu   %[src_ptr], %[src_ptr],    %[src_pixels_per_line]   \n\t"
+          "ulw    %[Temp1],   0(%[src_ptr])                           \n\t"
+          "ulw    %[Temp2],   4(%[src_ptr])                           \n\t"
+          "sw     %[Temp1],   24(%[FData])                            \n\t"
+          "sw     %[Temp2],   28(%[FData])                            \n\t"
+          "addu   %[src_ptr], %[src_ptr],    %[src_pixels_per_line]   \n\t"
 
-                "ulw    %[Temp1],   0(%[src_ptr])                           \n\t"
-                "ulw    %[Temp2],   4(%[src_ptr])                           \n\t"
-                "sw     %[Temp1],   8(%[FData])                             \n\t"
-                "sw     %[Temp2],   12(%[FData])                            \n\t"
-                "addu   %[src_ptr], %[src_ptr],    %[src_pixels_per_line]   \n\t"
+          "ulw    %[Temp1],   0(%[src_ptr])                           \n\t"
+          "ulw    %[Temp2],   4(%[src_ptr])                           \n\t"
+          "sw     %[Temp1],   32(%[FData])                            \n\t"
+          "sw     %[Temp2],   36(%[FData])                            \n\t"
+          "addu   %[src_ptr], %[src_ptr],    %[src_pixels_per_line]   \n\t"
 
-                "ulw    %[Temp1],   0(%[src_ptr])                           \n\t"
-                "ulw    %[Temp2],   4(%[src_ptr])                           \n\t"
-                "sw     %[Temp1],   16(%[FData])                            \n\t"
-                "sw     %[Temp2],   20(%[FData])                            \n\t"
-                "addu   %[src_ptr], %[src_ptr],    %[src_pixels_per_line]   \n\t"
+          "ulw    %[Temp1],   0(%[src_ptr])                           \n\t"
+          "ulw    %[Temp2],   4(%[src_ptr])                           \n\t"
+          "sw     %[Temp1],   40(%[FData])                            \n\t"
+          "sw     %[Temp2],   44(%[FData])                            \n\t"
+          "addu   %[src_ptr], %[src_ptr],    %[src_pixels_per_line]   \n\t"
 
-                "ulw    %[Temp1],   0(%[src_ptr])                           \n\t"
-                "ulw    %[Temp2],   4(%[src_ptr])                           \n\t"
-                "sw     %[Temp1],   24(%[FData])                            \n\t"
-                "sw     %[Temp2],   28(%[FData])                            \n\t"
-                "addu   %[src_ptr], %[src_ptr],    %[src_pixels_per_line]   \n\t"
+          "ulw    %[Temp1],   0(%[src_ptr])                           \n\t"
+          "ulw    %[Temp2],   4(%[src_ptr])                           \n\t"
+          "sw     %[Temp1],   48(%[FData])                            \n\t"
+          "sw     %[Temp2],   52(%[FData])                            \n\t"
+          "addu   %[src_ptr], %[src_ptr],    %[src_pixels_per_line]   \n\t"
 
-                "ulw    %[Temp1],   0(%[src_ptr])                           \n\t"
-                "ulw    %[Temp2],   4(%[src_ptr])                           \n\t"
-                "sw     %[Temp1],   32(%[FData])                            \n\t"
-                "sw     %[Temp2],   36(%[FData])                            \n\t"
-                "addu   %[src_ptr], %[src_ptr],    %[src_pixels_per_line]   \n\t"
+          "ulw    %[Temp1],   0(%[src_ptr])                           \n\t"
+          "ulw    %[Temp2],   4(%[src_ptr])                           \n\t"
+          "sw     %[Temp1],   56(%[FData])                            \n\t"
+          "sw     %[Temp2],   60(%[FData])                            \n\t"
+          "addu   %[src_ptr], %[src_ptr],    %[src_pixels_per_line]   \n\t"
 
-                "ulw    %[Temp1],   0(%[src_ptr])                           \n\t"
-                "ulw    %[Temp2],   4(%[src_ptr])                           \n\t"
-                "sw     %[Temp1],   40(%[FData])                            \n\t"
-                "sw     %[Temp2],   44(%[FData])                            \n\t"
-                "addu   %[src_ptr], %[src_ptr],    %[src_pixels_per_line]   \n\t"
+          "ulw    %[Temp1],   0(%[src_ptr])                           \n\t"
+          "ulw    %[Temp2],   4(%[src_ptr])                           \n\t"
+          "sw     %[Temp1],   64(%[FData])                            \n\t"
+          "sw     %[Temp2],   68(%[FData])                            \n\t"
+          "addu   %[src_ptr], %[src_ptr],    %[src_pixels_per_line]   \n\t"
 
-                "ulw    %[Temp1],   0(%[src_ptr])                           \n\t"
-                "ulw    %[Temp2],   4(%[src_ptr])                           \n\t"
-                "sw     %[Temp1],   48(%[FData])                            \n\t"
-                "sw     %[Temp2],   52(%[FData])                            \n\t"
-                "addu   %[src_ptr], %[src_ptr],    %[src_pixels_per_line]   \n\t"
+          "ulw    %[Temp1],   0(%[src_ptr])                           \n\t"
+          "ulw    %[Temp2],   4(%[src_ptr])                           \n\t"
+          "sw     %[Temp1],   72(%[FData])                            \n\t"
+          "sw     %[Temp2],   76(%[FData])                            \n\t"
+          "addu   %[src_ptr], %[src_ptr],    %[src_pixels_per_line]   \n\t"
 
-                "ulw    %[Temp1],   0(%[src_ptr])                           \n\t"
-                "ulw    %[Temp2],   4(%[src_ptr])                           \n\t"
-                "sw     %[Temp1],   56(%[FData])                            \n\t"
-                "sw     %[Temp2],   60(%[FData])                            \n\t"
-                "addu   %[src_ptr], %[src_ptr],    %[src_pixels_per_line]   \n\t"
+          "ulw    %[Temp1],   0(%[src_ptr])                           \n\t"
+          "ulw    %[Temp2],   4(%[src_ptr])                           \n\t"
+          "sw     %[Temp1],   80(%[FData])                            \n\t"
+          "sw     %[Temp2],   84(%[FData])                            \n\t"
+          "addu   %[src_ptr], %[src_ptr],    %[src_pixels_per_line]   \n\t"
 
-                "ulw    %[Temp1],   0(%[src_ptr])                           \n\t"
-                "ulw    %[Temp2],   4(%[src_ptr])                           \n\t"
-                "sw     %[Temp1],   64(%[FData])                            \n\t"
-                "sw     %[Temp2],   68(%[FData])                            \n\t"
-                "addu   %[src_ptr], %[src_ptr],    %[src_pixels_per_line]   \n\t"
+          "ulw    %[Temp1],   0(%[src_ptr])                           \n\t"
+          "ulw    %[Temp2],   4(%[src_ptr])                           \n\t"
+          "sw     %[Temp1],   88(%[FData])                            \n\t"
+          "sw     %[Temp2],   92(%[FData])                            \n\t"
+          "addu   %[src_ptr], %[src_ptr],    %[src_pixels_per_line]   \n\t"
 
-                "ulw    %[Temp1],   0(%[src_ptr])                           \n\t"
-                "ulw    %[Temp2],   4(%[src_ptr])                           \n\t"
-                "sw     %[Temp1],   72(%[FData])                            \n\t"
-                "sw     %[Temp2],   76(%[FData])                            \n\t"
-                "addu   %[src_ptr], %[src_ptr],    %[src_pixels_per_line]   \n\t"
+          "ulw    %[Temp1],   0(%[src_ptr])                           \n\t"
+          "ulw    %[Temp2],   4(%[src_ptr])                           \n\t"
+          "sw     %[Temp1],   96(%[FData])                            \n\t"
+          "sw     %[Temp2],   100(%[FData])                           \n\t"
 
-                "ulw    %[Temp1],   0(%[src_ptr])                           \n\t"
-                "ulw    %[Temp2],   4(%[src_ptr])                           \n\t"
-                "sw     %[Temp1],   80(%[FData])                            \n\t"
-                "sw     %[Temp2],   84(%[FData])                            \n\t"
-                "addu   %[src_ptr], %[src_ptr],    %[src_pixels_per_line]   \n\t"
-
-                "ulw    %[Temp1],   0(%[src_ptr])                           \n\t"
-                "ulw    %[Temp2],   4(%[src_ptr])                           \n\t"
-                "sw     %[Temp1],   88(%[FData])                            \n\t"
-                "sw     %[Temp2],   92(%[FData])                            \n\t"
-                "addu   %[src_ptr], %[src_ptr],    %[src_pixels_per_line]   \n\t"
-
-                "ulw    %[Temp1],   0(%[src_ptr])                           \n\t"
-                "ulw    %[Temp2],   4(%[src_ptr])                           \n\t"
-                "sw     %[Temp1],   96(%[FData])                            \n\t"
-                "sw     %[Temp2],   100(%[FData])                           \n\t"
-
-                : [Temp1] "=&r" (Temp1), [Temp2] "=&r" (Temp2)
-                : [FData] "r" (FData), [src_ptr] "r" (src_ptr),
-                  [src_pixels_per_line] "r" (src_pixels_per_line)
-            );
-        }
-
-        /* filter verticaly... */
-        vp8_filter_block2d_second_pass_8(FData + 16, dst_ptr, dst_pitch, 8, 8, yoffset);
+          : [Temp1] "=&r"(Temp1), [Temp2] "=&r"(Temp2)
+          : [FData] "r"(FData), [src_ptr] "r"(src_ptr),
+            [src_pixels_per_line] "r"(src_pixels_per_line));
     }
 
-    /* if (yoffsset == 0) vp8_filter_block2d_first_pass save data to dst_ptr */
-    else
-    {
-        if (xoffset)
-            vp8_filter_block2d_first_pass_8_all(src_ptr, dst_ptr, src_pixels_per_line,
-                                                8, xoffset, dst_pitch);
+    /* filter verticaly... */
+    vp8_filter_block2d_second_pass_8(FData + 16, dst_ptr, dst_pitch, 8, 8,
+                                     yoffset);
+  }
 
-        else
-        {
-            /* copy from src buffer to dst buffer */
-            __asm__ __volatile__ (
-                "ulw    %[Temp1],   0(%[src_ptr])                           \n\t"
-                "ulw    %[Temp2],   4(%[src_ptr])                           \n\t"
-                "sw     %[Temp1],   0(%[dst_ptr])                           \n\t"
-                "sw     %[Temp2],   4(%[dst_ptr])                           \n\t"
-                "addu   %[src_ptr], %[src_ptr],    %[src_pixels_per_line]   \n\t"
+  /* if (yoffsset == 0) vp8_filter_block2d_first_pass save data to dst_ptr */
+  else {
+    if (xoffset)
+      vp8_filter_block2d_first_pass_8_all(src_ptr, dst_ptr, src_pixels_per_line,
+                                          8, xoffset, dst_pitch);
 
-                "ulw    %[Temp1],   0(%[src_ptr])                           \n\t"
-                "ulw    %[Temp2],   4(%[src_ptr])                           \n\t"
-                "sw     %[Temp1],   8(%[dst_ptr])                           \n\t"
-                "sw     %[Temp2],   12(%[dst_ptr])                          \n\t"
-                "addu   %[src_ptr], %[src_ptr],    %[src_pixels_per_line]   \n\t"
+    else {
+      /* copy from src buffer to dst buffer */
+      __asm__ __volatile__(
+          "ulw    %[Temp1],   0(%[src_ptr])                           \n\t"
+          "ulw    %[Temp2],   4(%[src_ptr])                           \n\t"
+          "sw     %[Temp1],   0(%[dst_ptr])                           \n\t"
+          "sw     %[Temp2],   4(%[dst_ptr])                           \n\t"
+          "addu   %[src_ptr], %[src_ptr],    %[src_pixels_per_line]   \n\t"
 
-                "ulw    %[Temp1],   0(%[src_ptr])                           \n\t"
-                "ulw    %[Temp2],   4(%[src_ptr])                           \n\t"
-                "sw     %[Temp1],   16(%[dst_ptr])                          \n\t"
-                "sw     %[Temp2],   20(%[dst_ptr])                          \n\t"
-                "addu   %[src_ptr], %[src_ptr],    %[src_pixels_per_line]   \n\t"
+          "ulw    %[Temp1],   0(%[src_ptr])                           \n\t"
+          "ulw    %[Temp2],   4(%[src_ptr])                           \n\t"
+          "sw     %[Temp1],   8(%[dst_ptr])                           \n\t"
+          "sw     %[Temp2],   12(%[dst_ptr])                          \n\t"
+          "addu   %[src_ptr], %[src_ptr],    %[src_pixels_per_line]   \n\t"
 
-                "ulw    %[Temp1],   0(%[src_ptr])                           \n\t"
-                "ulw    %[Temp2],   4(%[src_ptr])                           \n\t"
-                "sw     %[Temp1],   24(%[dst_ptr])                          \n\t"
-                "sw     %[Temp2],   28(%[dst_ptr])                          \n\t"
-                "addu   %[src_ptr], %[src_ptr],   %[src_pixels_per_line]    \n\t"
+          "ulw    %[Temp1],   0(%[src_ptr])                           \n\t"
+          "ulw    %[Temp2],   4(%[src_ptr])                           \n\t"
+          "sw     %[Temp1],   16(%[dst_ptr])                          \n\t"
+          "sw     %[Temp2],   20(%[dst_ptr])                          \n\t"
+          "addu   %[src_ptr], %[src_ptr],    %[src_pixels_per_line]   \n\t"
 
-                "ulw    %[Temp1],   0(%[src_ptr])                           \n\t"
-                "ulw    %[Temp2],   4(%[src_ptr])                           \n\t"
-                "sw     %[Temp1],   32(%[dst_ptr])                          \n\t"
-                "sw     %[Temp2],   36(%[dst_ptr])                          \n\t"
-                "addu   %[src_ptr], %[src_ptr],    %[src_pixels_per_line]   \n\t"
+          "ulw    %[Temp1],   0(%[src_ptr])                           \n\t"
+          "ulw    %[Temp2],   4(%[src_ptr])                           \n\t"
+          "sw     %[Temp1],   24(%[dst_ptr])                          \n\t"
+          "sw     %[Temp2],   28(%[dst_ptr])                          \n\t"
+          "addu   %[src_ptr], %[src_ptr],   %[src_pixels_per_line]    \n\t"
 
-                "ulw    %[Temp1],   0(%[src_ptr])                           \n\t"
-                "ulw    %[Temp2],   4(%[src_ptr])                           \n\t"
-                "sw     %[Temp1],   40(%[dst_ptr])                          \n\t"
-                "sw     %[Temp2],   44(%[dst_ptr])                          \n\t"
-                "addu   %[src_ptr], %[src_ptr],    %[src_pixels_per_line]   \n\t"
+          "ulw    %[Temp1],   0(%[src_ptr])                           \n\t"
+          "ulw    %[Temp2],   4(%[src_ptr])                           \n\t"
+          "sw     %[Temp1],   32(%[dst_ptr])                          \n\t"
+          "sw     %[Temp2],   36(%[dst_ptr])                          \n\t"
+          "addu   %[src_ptr], %[src_ptr],    %[src_pixels_per_line]   \n\t"
 
-                "ulw    %[Temp1],   0(%[src_ptr])                           \n\t"
-                "ulw    %[Temp2],   4(%[src_ptr])                           \n\t"
-                "sw     %[Temp1],   48(%[dst_ptr])                          \n\t"
-                "sw     %[Temp2],   52(%[dst_ptr])                          \n\t"
-                "addu   %[src_ptr], %[src_ptr],    %[src_pixels_per_line]   \n\t"
+          "ulw    %[Temp1],   0(%[src_ptr])                           \n\t"
+          "ulw    %[Temp2],   4(%[src_ptr])                           \n\t"
+          "sw     %[Temp1],   40(%[dst_ptr])                          \n\t"
+          "sw     %[Temp2],   44(%[dst_ptr])                          \n\t"
+          "addu   %[src_ptr], %[src_ptr],    %[src_pixels_per_line]   \n\t"
 
-                "ulw    %[Temp1],   0(%[src_ptr])                           \n\t"
-                "ulw    %[Temp2],   4(%[src_ptr])                           \n\t"
-                "sw     %[Temp1],   56(%[dst_ptr])                          \n\t"
-                "sw     %[Temp2],   60(%[dst_ptr])                          \n\t"
+          "ulw    %[Temp1],   0(%[src_ptr])                           \n\t"
+          "ulw    %[Temp2],   4(%[src_ptr])                           \n\t"
+          "sw     %[Temp1],   48(%[dst_ptr])                          \n\t"
+          "sw     %[Temp2],   52(%[dst_ptr])                          \n\t"
+          "addu   %[src_ptr], %[src_ptr],    %[src_pixels_per_line]   \n\t"
 
-                : [Temp1] "=&r" (Temp1), [Temp2] "=&r" (Temp2)
-                : [dst_ptr] "r" (dst_ptr), [src_ptr] "r" (src_ptr),
-                  [src_pixels_per_line] "r" (src_pixels_per_line)
-            );
-        }
+          "ulw    %[Temp1],   0(%[src_ptr])                           \n\t"
+          "ulw    %[Temp2],   4(%[src_ptr])                           \n\t"
+          "sw     %[Temp1],   56(%[dst_ptr])                          \n\t"
+          "sw     %[Temp2],   60(%[dst_ptr])                          \n\t"
+
+          : [Temp1] "=&r"(Temp1), [Temp2] "=&r"(Temp2)
+          : [dst_ptr] "r"(dst_ptr), [src_ptr] "r"(src_ptr),
+            [src_pixels_per_line] "r"(src_pixels_per_line));
     }
+  }
 }
 
+void vp8_sixtap_predict8x4_dspr2(unsigned char *RESTRICT src_ptr,
+                                 int src_pixels_per_line, int xoffset,
+                                 int yoffset, unsigned char *RESTRICT dst_ptr,
+                                 int dst_pitch) {
+  unsigned char FData[9 * 8]; /* Temp data bufffer used in filtering */
+  unsigned int pos, Temp1, Temp2;
 
-void vp8_sixtap_predict8x4_dspr2
-(
-    unsigned char   *RESTRICT src_ptr,
-    int  src_pixels_per_line,
-    int  xoffset,
-    int  yoffset,
-    unsigned char *RESTRICT dst_ptr,
-    int  dst_pitch
-)
-{
-    unsigned char FData[9 * 8]; /* Temp data bufffer used in filtering */
-    unsigned int pos, Temp1, Temp2;
+  pos = 16;
 
-    pos = 16;
+  /* bit positon for extract from acc */
+  __asm__ __volatile__("wrdsp      %[pos],     1           \n\t"
+                       :
+                       : [pos] "r"(pos));
 
-    /* bit positon for extract from acc */
-    __asm__ __volatile__ (
-        "wrdsp      %[pos],     1           \n\t"
-        :
-        : [pos] "r" (pos)
-    );
+  if (yoffset) {
+    src_ptr = src_ptr - (2 * src_pixels_per_line);
 
-    if (yoffset)
-    {
+    if (xoffset) /* filter 1-D horizontally... */
+      vp8_filter_block2d_first_pass_8_all(src_ptr, FData, src_pixels_per_line,
+                                          9, xoffset, 8);
 
-        src_ptr = src_ptr - (2 * src_pixels_per_line);
+    else {
+      /* prefetch src_ptr data to cache memory */
+      prefetch_load(src_ptr + 2 * src_pixels_per_line);
 
-        if (xoffset)
-            /* filter 1-D horizontally... */
-            vp8_filter_block2d_first_pass_8_all(src_ptr, FData, src_pixels_per_line,
-                                                9, xoffset, 8);
+      __asm__ __volatile__(
+          "ulw    %[Temp1],   0(%[src_ptr])                           \n\t"
+          "ulw    %[Temp2],   4(%[src_ptr])                           \n\t"
+          "sw     %[Temp1],   0(%[FData])                             \n\t"
+          "sw     %[Temp2],   4(%[FData])                             \n\t"
+          "addu   %[src_ptr], %[src_ptr],    %[src_pixels_per_line]   \n\t"
 
-        else
-        {
-            /* prefetch src_ptr data to cache memory */
-            prefetch_load(src_ptr + 2 * src_pixels_per_line);
+          "ulw    %[Temp1],   0(%[src_ptr])                           \n\t"
+          "ulw    %[Temp2],   4(%[src_ptr])                           \n\t"
+          "sw     %[Temp1],   8(%[FData])                             \n\t"
+          "sw     %[Temp2],   12(%[FData])                            \n\t"
+          "addu   %[src_ptr], %[src_ptr],    %[src_pixels_per_line]   \n\t"
 
-            __asm__ __volatile__ (
-                "ulw    %[Temp1],   0(%[src_ptr])                           \n\t"
-                "ulw    %[Temp2],   4(%[src_ptr])                           \n\t"
-                "sw     %[Temp1],   0(%[FData])                             \n\t"
-                "sw     %[Temp2],   4(%[FData])                             \n\t"
-                "addu   %[src_ptr], %[src_ptr],    %[src_pixels_per_line]   \n\t"
+          "ulw    %[Temp1],   0(%[src_ptr])                           \n\t"
+          "ulw    %[Temp2],   4(%[src_ptr])                           \n\t"
+          "sw     %[Temp1],   16(%[FData])                            \n\t"
+          "sw     %[Temp2],   20(%[FData])                            \n\t"
+          "addu   %[src_ptr], %[src_ptr],    %[src_pixels_per_line]   \n\t"
 
-                "ulw    %[Temp1],   0(%[src_ptr])                           \n\t"
-                "ulw    %[Temp2],   4(%[src_ptr])                           \n\t"
-                "sw     %[Temp1],   8(%[FData])                             \n\t"
-                "sw     %[Temp2],   12(%[FData])                            \n\t"
-                "addu   %[src_ptr], %[src_ptr],    %[src_pixels_per_line]   \n\t"
+          "ulw    %[Temp1],   0(%[src_ptr])                           \n\t"
+          "ulw    %[Temp2],   4(%[src_ptr])                           \n\t"
+          "sw     %[Temp1],   24(%[FData])                            \n\t"
+          "sw     %[Temp2],   28(%[FData])                            \n\t"
+          "addu   %[src_ptr], %[src_ptr],    %[src_pixels_per_line]   \n\t"
 
-                "ulw    %[Temp1],   0(%[src_ptr])                           \n\t"
-                "ulw    %[Temp2],   4(%[src_ptr])                           \n\t"
-                "sw     %[Temp1],   16(%[FData])                            \n\t"
-                "sw     %[Temp2],   20(%[FData])                            \n\t"
-                "addu   %[src_ptr], %[src_ptr],    %[src_pixels_per_line]   \n\t"
+          "ulw    %[Temp1],   0(%[src_ptr])                           \n\t"
+          "ulw    %[Temp2],   4(%[src_ptr])                           \n\t"
+          "sw     %[Temp1],   32(%[FData])                            \n\t"
+          "sw     %[Temp2],   36(%[FData])                            \n\t"
+          "addu   %[src_ptr], %[src_ptr],    %[src_pixels_per_line]   \n\t"
 
-                "ulw    %[Temp1],   0(%[src_ptr])                           \n\t"
-                "ulw    %[Temp2],   4(%[src_ptr])                           \n\t"
-                "sw     %[Temp1],   24(%[FData])                            \n\t"
-                "sw     %[Temp2],   28(%[FData])                            \n\t"
-                "addu   %[src_ptr], %[src_ptr],    %[src_pixels_per_line]   \n\t"
+          "ulw    %[Temp1],   0(%[src_ptr])                           \n\t"
+          "ulw    %[Temp2],   4(%[src_ptr])                           \n\t"
+          "sw     %[Temp1],   40(%[FData])                            \n\t"
+          "sw     %[Temp2],   44(%[FData])                            \n\t"
+          "addu   %[src_ptr], %[src_ptr],    %[src_pixels_per_line]   \n\t"
 
-                "ulw    %[Temp1],   0(%[src_ptr])                           \n\t"
-                "ulw    %[Temp2],   4(%[src_ptr])                           \n\t"
-                "sw     %[Temp1],   32(%[FData])                            \n\t"
-                "sw     %[Temp2],   36(%[FData])                            \n\t"
-                "addu   %[src_ptr], %[src_ptr],    %[src_pixels_per_line]   \n\t"
+          "ulw    %[Temp1],   0(%[src_ptr])                           \n\t"
+          "ulw    %[Temp2],   4(%[src_ptr])                           \n\t"
+          "sw     %[Temp1],   48(%[FData])                            \n\t"
+          "sw     %[Temp2],   52(%[FData])                            \n\t"
+          "addu   %[src_ptr], %[src_ptr],    %[src_pixels_per_line]   \n\t"
 
-                "ulw    %[Temp1],   0(%[src_ptr])                           \n\t"
-                "ulw    %[Temp2],   4(%[src_ptr])                           \n\t"
-                "sw     %[Temp1],   40(%[FData])                            \n\t"
-                "sw     %[Temp2],   44(%[FData])                            \n\t"
-                "addu   %[src_ptr], %[src_ptr],    %[src_pixels_per_line]   \n\t"
+          "ulw    %[Temp1],   0(%[src_ptr])                           \n\t"
+          "ulw    %[Temp2],   4(%[src_ptr])                           \n\t"
+          "sw     %[Temp1],   56(%[FData])                            \n\t"
+          "sw     %[Temp2],   60(%[FData])                            \n\t"
+          "addu   %[src_ptr], %[src_ptr],    %[src_pixels_per_line]   \n\t"
 
-                "ulw    %[Temp1],   0(%[src_ptr])                           \n\t"
-                "ulw    %[Temp2],   4(%[src_ptr])                           \n\t"
-                "sw     %[Temp1],   48(%[FData])                            \n\t"
-                "sw     %[Temp2],   52(%[FData])                            \n\t"
-                "addu   %[src_ptr], %[src_ptr],    %[src_pixels_per_line]   \n\t"
+          "ulw    %[Temp1],   0(%[src_ptr])                           \n\t"
+          "ulw    %[Temp2],   4(%[src_ptr])                           \n\t"
+          "sw     %[Temp1],   64(%[FData])                            \n\t"
+          "sw     %[Temp2],   68(%[FData])                            \n\t"
 
-                "ulw    %[Temp1],   0(%[src_ptr])                           \n\t"
-                "ulw    %[Temp2],   4(%[src_ptr])                           \n\t"
-                "sw     %[Temp1],   56(%[FData])                            \n\t"
-                "sw     %[Temp2],   60(%[FData])                            \n\t"
-                "addu   %[src_ptr], %[src_ptr],    %[src_pixels_per_line]   \n\t"
-
-                "ulw    %[Temp1],   0(%[src_ptr])                           \n\t"
-                "ulw    %[Temp2],   4(%[src_ptr])                           \n\t"
-                "sw     %[Temp1],   64(%[FData])                            \n\t"
-                "sw     %[Temp2],   68(%[FData])                            \n\t"
-
-                : [Temp1] "=&r" (Temp1), [Temp2] "=&r" (Temp2)
-                : [FData] "r" (FData), [src_ptr] "r" (src_ptr),
-                  [src_pixels_per_line] "r" (src_pixels_per_line)
-            );
-        }
-
-        /* filter verticaly... */
-        vp8_filter_block2d_second_pass_8(FData + 16, dst_ptr, dst_pitch, 4, 8, yoffset);
+          : [Temp1] "=&r"(Temp1), [Temp2] "=&r"(Temp2)
+          : [FData] "r"(FData), [src_ptr] "r"(src_ptr),
+            [src_pixels_per_line] "r"(src_pixels_per_line));
     }
 
-    /* if (yoffsset == 0) vp8_filter_block2d_first_pass save data to dst_ptr */
-    else
-    {
-        if (xoffset)
-            vp8_filter_block2d_first_pass_8_all(src_ptr, dst_ptr, src_pixels_per_line,
-                                                4, xoffset, dst_pitch);
+    /* filter verticaly... */
+    vp8_filter_block2d_second_pass_8(FData + 16, dst_ptr, dst_pitch, 4, 8,
+                                     yoffset);
+  }
 
-        else
-        {
-            /* copy from src buffer to dst buffer */
-            __asm__ __volatile__ (
-                "ulw    %[Temp1],   0(%[src_ptr])                           \n\t"
-                "ulw    %[Temp2],   4(%[src_ptr])                           \n\t"
-                "sw     %[Temp1],   0(%[dst_ptr])                           \n\t"
-                "sw     %[Temp2],   4(%[dst_ptr])                           \n\t"
-                "addu   %[src_ptr], %[src_ptr],    %[src_pixels_per_line]   \n\t"
+  /* if (yoffsset == 0) vp8_filter_block2d_first_pass save data to dst_ptr */
+  else {
+    if (xoffset)
+      vp8_filter_block2d_first_pass_8_all(src_ptr, dst_ptr, src_pixels_per_line,
+                                          4, xoffset, dst_pitch);
 
-                "ulw    %[Temp1],   0(%[src_ptr])                           \n\t"
-                "ulw    %[Temp2],   4(%[src_ptr])                           \n\t"
-                "sw     %[Temp1],   8(%[dst_ptr])                           \n\t"
-                "sw     %[Temp2],   12(%[dst_ptr])                          \n\t"
-                "addu   %[src_ptr], %[src_ptr],    %[src_pixels_per_line]   \n\t"
+    else {
+      /* copy from src buffer to dst buffer */
+      __asm__ __volatile__(
+          "ulw    %[Temp1],   0(%[src_ptr])                           \n\t"
+          "ulw    %[Temp2],   4(%[src_ptr])                           \n\t"
+          "sw     %[Temp1],   0(%[dst_ptr])                           \n\t"
+          "sw     %[Temp2],   4(%[dst_ptr])                           \n\t"
+          "addu   %[src_ptr], %[src_ptr],    %[src_pixels_per_line]   \n\t"
 
-                "ulw    %[Temp1],   0(%[src_ptr])                           \n\t"
-                "ulw    %[Temp2],   4(%[src_ptr])                           \n\t"
-                "sw     %[Temp1],   16(%[dst_ptr])                          \n\t"
-                "sw     %[Temp2],   20(%[dst_ptr])                          \n\t"
-                "addu   %[src_ptr], %[src_ptr],    %[src_pixels_per_line]   \n\t"
+          "ulw    %[Temp1],   0(%[src_ptr])                           \n\t"
+          "ulw    %[Temp2],   4(%[src_ptr])                           \n\t"
+          "sw     %[Temp1],   8(%[dst_ptr])                           \n\t"
+          "sw     %[Temp2],   12(%[dst_ptr])                          \n\t"
+          "addu   %[src_ptr], %[src_ptr],    %[src_pixels_per_line]   \n\t"
 
-                "ulw    %[Temp1],   0(%[src_ptr])                           \n\t"
-                "ulw    %[Temp2],   4(%[src_ptr])                           \n\t"
-                "sw     %[Temp1],   24(%[dst_ptr])                          \n\t"
-                "sw     %[Temp2],   28(%[dst_ptr])                          \n\t"
+          "ulw    %[Temp1],   0(%[src_ptr])                           \n\t"
+          "ulw    %[Temp2],   4(%[src_ptr])                           \n\t"
+          "sw     %[Temp1],   16(%[dst_ptr])                          \n\t"
+          "sw     %[Temp2],   20(%[dst_ptr])                          \n\t"
+          "addu   %[src_ptr], %[src_ptr],    %[src_pixels_per_line]   \n\t"
 
-                : [Temp1] "=&r" (Temp1), [Temp2] "=&r" (Temp2)
-                : [dst_ptr] "r" (dst_ptr), [src_ptr] "r" (src_ptr),
-                  [src_pixels_per_line] "r" (src_pixels_per_line)
-            );
-        }
+          "ulw    %[Temp1],   0(%[src_ptr])                           \n\t"
+          "ulw    %[Temp2],   4(%[src_ptr])                           \n\t"
+          "sw     %[Temp1],   24(%[dst_ptr])                          \n\t"
+          "sw     %[Temp2],   28(%[dst_ptr])                          \n\t"
+
+          : [Temp1] "=&r"(Temp1), [Temp2] "=&r"(Temp2)
+          : [dst_ptr] "r"(dst_ptr), [src_ptr] "r"(src_ptr),
+            [src_pixels_per_line] "r"(src_pixels_per_line));
     }
+  }
 }
 
+void vp8_sixtap_predict16x16_dspr2(unsigned char *RESTRICT src_ptr,
+                                   int src_pixels_per_line, int xoffset,
+                                   int yoffset, unsigned char *RESTRICT dst_ptr,
+                                   int dst_pitch) {
+  const unsigned short *VFilter;
+  unsigned char FData[21 * 16]; /* Temp data bufffer used in filtering */
+  unsigned int pos;
 
-void vp8_sixtap_predict16x16_dspr2
-(
-    unsigned char   *RESTRICT src_ptr,
-    int  src_pixels_per_line,
-    int  xoffset,
-    int  yoffset,
-    unsigned char *RESTRICT dst_ptr,
-    int  dst_pitch
-)
-{
-    const unsigned short *VFilter;
-    unsigned char FData[21 * 16]; /* Temp data bufffer used in filtering */
-    unsigned int pos;
+  VFilter = sub_pel_filterss[yoffset];
 
-    VFilter = sub_pel_filterss[yoffset];
+  pos = 16;
 
-    pos = 16;
+  /* bit positon for extract from acc */
+  __asm__ __volatile__("wrdsp      %[pos],     1           \n\t"
+                       :
+                       : [pos] "r"(pos));
 
-    /* bit positon for extract from acc */
-    __asm__ __volatile__ (
-        "wrdsp      %[pos],     1           \n\t"
-        :
-        : [pos] "r" (pos)
-    );
+  if (yoffset) {
+    src_ptr = src_ptr - (2 * src_pixels_per_line);
 
-    if (yoffset)
-    {
+    switch (xoffset) {
+      /* filter 1-D horizontally... */
+      case 2:
+      case 4:
+      case 6:
+        /* 6 tap filter */
+        vp8_filter_block2d_first_pass16_6tap(
+            src_ptr, FData, src_pixels_per_line, 21, xoffset, 16);
+        break;
 
-        src_ptr = src_ptr - (2 * src_pixels_per_line);
+      case 0:
+        /* only copy buffer */
+        vp8_filter_block2d_first_pass16_0(src_ptr, FData, src_pixels_per_line);
+        break;
 
-        switch (xoffset)
-        {
-            /* filter 1-D horizontally... */
-        case 2:
-        case 4:
-        case 6:
-            /* 6 tap filter */
-            vp8_filter_block2d_first_pass16_6tap(src_ptr, FData, src_pixels_per_line,
-                                                 21, xoffset, 16);
-            break;
+      case 1:
+      case 3:
+      case 5:
+      case 7:
+        /* 4 tap filter */
+        vp8_filter_block2d_first_pass16_4tap(
+            src_ptr, FData, src_pixels_per_line, 16, 21, xoffset, yoffset,
+            dst_ptr, dst_pitch);
+        break;
+    }
 
-        case 0:
-            /* only copy buffer */
-            vp8_filter_block2d_first_pass16_0(src_ptr, FData, src_pixels_per_line);
-            break;
+    /* filter verticaly... */
+    vp8_filter_block2d_second_pass161(FData + 32, dst_ptr, dst_pitch, VFilter);
+  } else {
+    /* if (yoffsset == 0) vp8_filter_block2d_first_pass save data to dst_ptr */
+    switch (xoffset) {
+      case 2:
+      case 4:
+      case 6:
+        /* 6 tap filter */
+        vp8_filter_block2d_first_pass16_6tap(
+            src_ptr, dst_ptr, src_pixels_per_line, 16, xoffset, dst_pitch);
+        break;
 
-        case 1:
-        case 3:
-        case 5:
-        case 7:
-            /* 4 tap filter */
-            vp8_filter_block2d_first_pass16_4tap(src_ptr, FData, src_pixels_per_line, 16,
-                                                 21, xoffset, yoffset, dst_ptr, dst_pitch);
-            break;
-        }
-
-        /* filter verticaly... */
-        vp8_filter_block2d_second_pass161(FData + 32, dst_ptr, dst_pitch, VFilter);
+      case 1:
+      case 3:
+      case 5:
+      case 7:
+        /* 4 tap filter */
+        vp8_filter_block2d_first_pass16_4tap(
+            src_ptr, dst_ptr, src_pixels_per_line, 16, 21, xoffset, yoffset,
+            dst_ptr, dst_pitch);
+        break;
     }
-    else
-    {
-        /* if (yoffsset == 0) vp8_filter_block2d_first_pass save data to dst_ptr */
-        switch (xoffset)
-        {
-        case 2:
-        case 4:
-        case 6:
-            /* 6 tap filter */
-            vp8_filter_block2d_first_pass16_6tap(src_ptr, dst_ptr, src_pixels_per_line,
-                                                 16, xoffset, dst_pitch);
-            break;
-
-        case 1:
-        case 3:
-        case 5:
-        case 7:
-            /* 4 tap filter */
-            vp8_filter_block2d_first_pass16_4tap(src_ptr, dst_ptr, src_pixels_per_line, 16,
-                                                 21, xoffset, yoffset, dst_ptr, dst_pitch);
-            break;
-        }
-    }
+  }
 }
 
 #endif
--- a/vp8/common/mips/dspr2/idct_blk_dspr2.c
+++ b/vp8/common/mips/dspr2/idct_blk_dspr2.c
@@ -13,76 +13,64 @@
 
 #if HAVE_DSPR2
 
-void vp8_dequant_idct_add_y_block_dspr2
-(short *q, short *dq,
- unsigned char *dst, int stride, char *eobs)
-{
-    int i, j;
+void vp8_dequant_idct_add_y_block_dspr2(short *q, short *dq, unsigned char *dst,
+                                        int stride, char *eobs) {
+  int i, j;
 
-    for (i = 0; i < 4; i++)
-    {
-        for (j = 0; j < 4; j++)
-        {
-            if (*eobs++ > 1)
-                vp8_dequant_idct_add_dspr2(q, dq, dst, stride);
-            else
-            {
-                vp8_dc_only_idct_add_dspr2(q[0]*dq[0], dst, stride, dst, stride);
-                ((int *)q)[0] = 0;
-            }
+  for (i = 0; i < 4; i++) {
+    for (j = 0; j < 4; j++) {
+      if (*eobs++ > 1)
+        vp8_dequant_idct_add_dspr2(q, dq, dst, stride);
+      else {
+        vp8_dc_only_idct_add_dspr2(q[0] * dq[0], dst, stride, dst, stride);
+        ((int *)q)[0] = 0;
+      }
 
-            q   += 16;
-            dst += 4;
-        }
-
-        dst += 4 * stride - 16;
+      q += 16;
+      dst += 4;
     }
+
+    dst += 4 * stride - 16;
+  }
 }
 
-void vp8_dequant_idct_add_uv_block_dspr2
-(short *q, short *dq,
- unsigned char *dstu, unsigned char *dstv, int stride, char *eobs)
-{
-    int i, j;
+void vp8_dequant_idct_add_uv_block_dspr2(short *q, short *dq,
+                                         unsigned char *dstu,
+                                         unsigned char *dstv, int stride,
+                                         char *eobs) {
+  int i, j;
 
-    for (i = 0; i < 2; i++)
-    {
-        for (j = 0; j < 2; j++)
-        {
-            if (*eobs++ > 1)
-                vp8_dequant_idct_add_dspr2(q, dq, dstu, stride);
-            else
-            {
-                vp8_dc_only_idct_add_dspr2(q[0]*dq[0], dstu, stride, dstu, stride);
-                ((int *)q)[0] = 0;
-            }
+  for (i = 0; i < 2; i++) {
+    for (j = 0; j < 2; j++) {
+      if (*eobs++ > 1)
+        vp8_dequant_idct_add_dspr2(q, dq, dstu, stride);
+      else {
+        vp8_dc_only_idct_add_dspr2(q[0] * dq[0], dstu, stride, dstu, stride);
+        ((int *)q)[0] = 0;
+      }
 
-            q    += 16;
-            dstu += 4;
-        }
-
-        dstu += 4 * stride - 8;
+      q += 16;
+      dstu += 4;
     }
 
-    for (i = 0; i < 2; i++)
-    {
-        for (j = 0; j < 2; j++)
-        {
-            if (*eobs++ > 1)
-                vp8_dequant_idct_add_dspr2(q, dq, dstv, stride);
-            else
-            {
-                vp8_dc_only_idct_add_dspr2(q[0]*dq[0], dstv, stride, dstv, stride);
-                ((int *)q)[0] = 0;
-            }
+    dstu += 4 * stride - 8;
+  }
 
-            q    += 16;
-            dstv += 4;
-        }
+  for (i = 0; i < 2; i++) {
+    for (j = 0; j < 2; j++) {
+      if (*eobs++ > 1)
+        vp8_dequant_idct_add_dspr2(q, dq, dstv, stride);
+      else {
+        vp8_dc_only_idct_add_dspr2(q[0] * dq[0], dstv, stride, dstv, stride);
+        ((int *)q)[0] = 0;
+      }
 
-        dstv += 4 * stride - 8;
+      q += 16;
+      dstv += 4;
     }
+
+    dstv += 4 * stride - 8;
+  }
 }
 
 #endif
-
--- a/vp8/common/mips/dspr2/idctllm_dspr2.c
+++ b/vp8/common/mips/dspr2/idctllm_dspr2.c
@@ -28,342 +28,319 @@
  ****************************************************************************/
 extern unsigned char ff_cropTbl[256 + 2 * CROP_WIDTH];
 static const int cospi8sqrt2minus1 = 20091;
-static const int sinpi8sqrt2      = 35468;
+static const int sinpi8sqrt2 = 35468;
 
-inline void prefetch_load_short(short *src)
-{
-    __asm__ __volatile__ (
-        "pref   0,  0(%[src])   \n\t"
-        :
-        : [src] "r" (src)
-    );
+inline void prefetch_load_short(short *src) {
+  __asm__ __volatile__("pref   0,  0(%[src])   \n\t" : : [src] "r"(src));
 }
 
 void vp8_short_idct4x4llm_dspr2(short *input, unsigned char *pred_ptr,
                                 int pred_stride, unsigned char *dst_ptr,
-                                int dst_stride)
-{
-    int r, c;
-    int a1, b1, c1, d1;
-    short output[16];
-    short *ip = input;
-    short *op = output;
-    int temp1, temp2;
-    int shortpitch = 4;
+                                int dst_stride) {
+  int r, c;
+  int a1, b1, c1, d1;
+  short output[16];
+  short *ip = input;
+  short *op = output;
+  int temp1, temp2;
+  int shortpitch = 4;
 
-    int c2, d2;
-    int temp3, temp4;
-    unsigned char *cm = ff_cropTbl + CROP_WIDTH;
+  int c2, d2;
+  int temp3, temp4;
+  unsigned char *cm = ff_cropTbl + CROP_WIDTH;
 
-    /* prepare data for load */
-    prefetch_load_short(ip + 8);
+  /* prepare data for load */
+  prefetch_load_short(ip + 8);
 
-    /* first loop is unrolled */
-    a1 = ip[0] + ip[8];
-    b1 = ip[0] - ip[8];
+  /* first loop is unrolled */
+  a1 = ip[0] + ip[8];
+  b1 = ip[0] - ip[8];
 
-    temp1 = (ip[4] * sinpi8sqrt2) >> 16;
-    temp2 = ip[12] + ((ip[12] * cospi8sqrt2minus1) >> 16);
-    c1 = temp1 - temp2;
+  temp1 = (ip[4] * sinpi8sqrt2) >> 16;
+  temp2 = ip[12] + ((ip[12] * cospi8sqrt2minus1) >> 16);
+  c1 = temp1 - temp2;
 
-    temp1 = ip[4] + ((ip[4] * cospi8sqrt2minus1) >> 16);
-    temp2 = (ip[12] * sinpi8sqrt2) >> 16;
-    d1 = temp1 + temp2;
+  temp1 = ip[4] + ((ip[4] * cospi8sqrt2minus1) >> 16);
+  temp2 = (ip[12] * sinpi8sqrt2) >> 16;
+  d1 = temp1 + temp2;
 
-    temp3 = (ip[5] * sinpi8sqrt2) >> 16;
-    temp4 = ip[13] + ((ip[13] * cospi8sqrt2minus1) >> 16);
-    c2 = temp3 - temp4;
+  temp3 = (ip[5] * sinpi8sqrt2) >> 16;
+  temp4 = ip[13] + ((ip[13] * cospi8sqrt2minus1) >> 16);
+  c2 = temp3 - temp4;
 
-    temp3 = ip[5] + ((ip[5] * cospi8sqrt2minus1) >> 16);
-    temp4 = (ip[13] * sinpi8sqrt2) >> 16;
-    d2 = temp3 + temp4;
+  temp3 = ip[5] + ((ip[5] * cospi8sqrt2minus1) >> 16);
+  temp4 = (ip[13] * sinpi8sqrt2) >> 16;
+  d2 = temp3 + temp4;
 
-    op[0] = a1 + d1;
-    op[12] = a1 - d1;
-    op[4] = b1 + c1;
-    op[8] = b1 - c1;
+  op[0] = a1 + d1;
+  op[12] = a1 - d1;
+  op[4] = b1 + c1;
+  op[8] = b1 - c1;
 
-    a1 = ip[1] + ip[9];
-    b1 = ip[1] - ip[9];
+  a1 = ip[1] + ip[9];
+  b1 = ip[1] - ip[9];
 
-    op[1] = a1 + d2;
-    op[13] = a1 - d2;
-    op[5] = b1 + c2;
-    op[9] = b1 - c2;
+  op[1] = a1 + d2;
+  op[13] = a1 - d2;
+  op[5] = b1 + c2;
+  op[9] = b1 - c2;
 
-    a1 = ip[2] + ip[10];
-    b1 = ip[2] - ip[10];
+  a1 = ip[2] + ip[10];
+  b1 = ip[2] - ip[10];
 
-    temp1 = (ip[6] * sinpi8sqrt2) >> 16;
-    temp2 = ip[14] + ((ip[14] * cospi8sqrt2minus1) >> 16);
-    c1 = temp1 - temp2;
+  temp1 = (ip[6] * sinpi8sqrt2) >> 16;
+  temp2 = ip[14] + ((ip[14] * cospi8sqrt2minus1) >> 16);
+  c1 = temp1 - temp2;
 
-    temp1 = ip[6] + ((ip[6] * cospi8sqrt2minus1) >> 16);
-    temp2 = (ip[14] * sinpi8sqrt2) >> 16;
-    d1 = temp1 + temp2;
+  temp1 = ip[6] + ((ip[6] * cospi8sqrt2minus1) >> 16);
+  temp2 = (ip[14] * sinpi8sqrt2) >> 16;
+  d1 = temp1 + temp2;
 
-    temp3 = (ip[7] * sinpi8sqrt2) >> 16;
-    temp4 = ip[15] + ((ip[15] * cospi8sqrt2minus1) >> 16);
-    c2 = temp3 - temp4;
+  temp3 = (ip[7] * sinpi8sqrt2) >> 16;
+  temp4 = ip[15] + ((ip[15] * cospi8sqrt2minus1) >> 16);
+  c2 = temp3 - temp4;
 
-    temp3 = ip[7] + ((ip[7] * cospi8sqrt2minus1) >> 16);
-    temp4 = (ip[15] * sinpi8sqrt2) >> 16;
-    d2 = temp3 + temp4;
+  temp3 = ip[7] + ((ip[7] * cospi8sqrt2minus1) >> 16);
+  temp4 = (ip[15] * sinpi8sqrt2) >> 16;
+  d2 = temp3 + temp4;
 
-    op[2] = a1 + d1;
-    op[14] = a1 - d1;
-    op[6] = b1 + c1;
-    op[10] = b1 - c1;
+  op[2] = a1 + d1;
+  op[14] = a1 - d1;
+  op[6] = b1 + c1;
+  op[10] = b1 - c1;
 
-    a1 = ip[3] + ip[11];
-    b1 = ip[3] - ip[11];
+  a1 = ip[3] + ip[11];
+  b1 = ip[3] - ip[11];
 
-    op[3] = a1 + d2;
-    op[15] = a1 - d2;
-    op[7] = b1 + c2;
-    op[11] = b1 - c2;
+  op[3] = a1 + d2;
+  op[15] = a1 - d2;
+  op[7] = b1 + c2;
+  op[11] = b1 - c2;
 
-    ip = output;
+  ip = output;
 
-    /* prepare data for load */
-    prefetch_load_short(ip + shortpitch);
+  /* prepare data for load */
+  prefetch_load_short(ip + shortpitch);
 
-    /* second loop is unrolled */
-    a1 = ip[0] + ip[2];
-    b1 = ip[0] - ip[2];
+  /* second loop is unrolled */
+  a1 = ip[0] + ip[2];
+  b1 = ip[0] - ip[2];
 
-    temp1 = (ip[1] * sinpi8sqrt2) >> 16;
-    temp2 = ip[3] + ((ip[3] * cospi8sqrt2minus1) >> 16);
-    c1 = temp1 - temp2;
+  temp1 = (ip[1] * sinpi8sqrt2) >> 16;
+  temp2 = ip[3] + ((ip[3] * cospi8sqrt2minus1) >> 16);
+  c1 = temp1 - temp2;
 
-    temp1 = ip[1] + ((ip[1] * cospi8sqrt2minus1) >> 16);
-    temp2 = (ip[3] * sinpi8sqrt2) >> 16;
-    d1 = temp1 + temp2;
+  temp1 = ip[1] + ((ip[1] * cospi8sqrt2minus1) >> 16);
+  temp2 = (ip[3] * sinpi8sqrt2) >> 16;
+  d1 = temp1 + temp2;
 
-    temp3 = (ip[5] * sinpi8sqrt2) >> 16;
-    temp4 = ip[7] + ((ip[7] * cospi8sqrt2minus1) >> 16);
-    c2 = temp3 - temp4;
+  temp3 = (ip[5] * sinpi8sqrt2) >> 16;
+  temp4 = ip[7] + ((ip[7] * cospi8sqrt2minus1) >> 16);
+  c2 = temp3 - temp4;
 
-    temp3 = ip[5] + ((ip[5] * cospi8sqrt2minus1) >> 16);
-    temp4 = (ip[7] * sinpi8sqrt2) >> 16;
-    d2 = temp3 + temp4;
+  temp3 = ip[5] + ((ip[5] * cospi8sqrt2minus1) >> 16);
+  temp4 = (ip[7] * sinpi8sqrt2) >> 16;
+  d2 = temp3 + temp4;
 
-    op[0] = (a1 + d1 + 4) >> 3;
-    op[3] = (a1 - d1 + 4) >> 3;
-    op[1] = (b1 + c1 + 4) >> 3;
-    op[2] = (b1 - c1 + 4) >> 3;
+  op[0] = (a1 + d1 + 4) >> 3;
+  op[3] = (a1 - d1 + 4) >> 3;
+  op[1] = (b1 + c1 + 4) >> 3;
+  op[2] = (b1 - c1 + 4) >> 3;
 
-    a1 = ip[4] + ip[6];
-    b1 = ip[4] - ip[6];
+  a1 = ip[4] + ip[6];
+  b1 = ip[4] - ip[6];
 
-    op[4] = (a1 + d2 + 4) >> 3;
-    op[7] = (a1 - d2 + 4) >> 3;
-    op[5] = (b1 + c2 + 4) >> 3;
-    op[6] = (b1 - c2 + 4) >> 3;
+  op[4] = (a1 + d2 + 4) >> 3;
+  op[7] = (a1 - d2 + 4) >> 3;
+  op[5] = (b1 + c2 + 4) >> 3;
+  op[6] = (b1 - c2 + 4) >> 3;
 
-    a1 = ip[8] + ip[10];
-    b1 = ip[8] - ip[10];
+  a1 = ip[8] + ip[10];
+  b1 = ip[8] - ip[10];
 
-    temp1 = (ip[9] * sinpi8sqrt2) >> 16;
-    temp2 = ip[11] + ((ip[11] * cospi8sqrt2minus1) >> 16);
-    c1 = temp1 - temp2;
+  temp1 = (ip[9] * sinpi8sqrt2) >> 16;
+  temp2 = ip[11] + ((ip[11] * cospi8sqrt2minus1) >> 16);
+  c1 = temp1 - temp2;
 
-    temp1 = ip[9] + ((ip[9] * cospi8sqrt2minus1) >> 16);
-    temp2 = (ip[11] * sinpi8sqrt2) >> 16;
-    d1 = temp1 + temp2;
+  temp1 = ip[9] + ((ip[9] * cospi8sqrt2minus1) >> 16);
+  temp2 = (ip[11] * sinpi8sqrt2) >> 16;
+  d1 = temp1 + temp2;
 
-    temp3 = (ip[13] * sinpi8sqrt2) >> 16;
-    temp4 = ip[15] + ((ip[15] * cospi8sqrt2minus1) >> 16);
-    c2 = temp3 - temp4;
+  temp3 = (ip[13] * sinpi8sqrt2) >> 16;
+  temp4 = ip[15] + ((ip[15] * cospi8sqrt2minus1) >> 16);
+  c2 = temp3 - temp4;
 
-    temp3 = ip[13] + ((ip[13] * cospi8sqrt2minus1) >> 16);
-    temp4 = (ip[15] * sinpi8sqrt2) >> 16;
-    d2 = temp3 + temp4;
+  temp3 = ip[13] + ((ip[13] * cospi8sqrt2minus1) >> 16);
+  temp4 = (ip[15] * sinpi8sqrt2) >> 16;
+  d2 = temp3 + temp4;
 
-    op[8] = (a1 + d1 + 4) >> 3;
-    op[11] = (a1 - d1 + 4) >> 3;
-    op[9] = (b1 + c1 + 4) >> 3;
-    op[10] = (b1 - c1 + 4) >> 3;
+  op[8] = (a1 + d1 + 4) >> 3;
+  op[11] = (a1 - d1 + 4) >> 3;
+  op[9] = (b1 + c1 + 4) >> 3;
+  op[10] = (b1 - c1 + 4) >> 3;
 
-    a1 = ip[12] + ip[14];
-    b1 = ip[12] - ip[14];
+  a1 = ip[12] + ip[14];
+  b1 = ip[12] - ip[14];
 
-    op[12] = (a1 + d2 + 4) >> 3;
-    op[15] = (a1 - d2 + 4) >> 3;
-    op[13] = (b1 + c2 + 4) >> 3;
-    op[14] = (b1 - c2 + 4) >> 3;
+  op[12] = (a1 + d2 + 4) >> 3;
+  op[15] = (a1 - d2 + 4) >> 3;
+  op[13] = (b1 + c2 + 4) >> 3;
+  op[14] = (b1 - c2 + 4) >> 3;
 
-    ip = output;
+  ip = output;
 
-    for (r = 0; r < 4; r++)
-    {
-        for (c = 0; c < 4; c++)
-        {
-            short a = ip[c] + pred_ptr[c] ;
-            dst_ptr[c] = cm[a] ;
-        }
-
-        ip += 4;
-        dst_ptr += dst_stride;
-        pred_ptr += pred_stride;
+  for (r = 0; r < 4; r++) {
+    for (c = 0; c < 4; c++) {
+      short a = ip[c] + pred_ptr[c];
+      dst_ptr[c] = cm[a];
     }
+
+    ip += 4;
+    dst_ptr += dst_stride;
+    pred_ptr += pred_stride;
+  }
 }
 
-void vp8_dc_only_idct_add_dspr2(short input_dc, unsigned char *pred_ptr, int pred_stride, unsigned char *dst_ptr, int dst_stride)
-{
-    int a1;
-    int i, absa1;
-    int t2, vector_a1, vector_a;
+void vp8_dc_only_idct_add_dspr2(short input_dc, unsigned char *pred_ptr,
+                                int pred_stride, unsigned char *dst_ptr,
+                                int dst_stride) {
+  int a1;
+  int i, absa1;
+  int t2, vector_a1, vector_a;
 
-    /* a1 = ((input_dc + 4) >> 3); */
-    __asm__ __volatile__ (
-        "addi  %[a1], %[input_dc], 4   \n\t"
-        "sra   %[a1], %[a1],       3   \n\t"
-        : [a1] "=r" (a1)
-        : [input_dc] "r" (input_dc)
-    );
+  /* a1 = ((input_dc + 4) >> 3); */
+  __asm__ __volatile__(
+      "addi  %[a1], %[input_dc], 4   \n\t"
+      "sra   %[a1], %[a1],       3   \n\t"
+      : [a1] "=r"(a1)
+      : [input_dc] "r"(input_dc));
 
-    if (a1 < 0)
-    {
-        /* use quad-byte
-         * input and output memory are four byte aligned
-         */
-        __asm__ __volatile__ (
-            "abs        %[absa1],     %[a1]         \n\t"
-            "replv.qb   %[vector_a1], %[absa1]      \n\t"
-            : [absa1] "=r" (absa1), [vector_a1] "=r" (vector_a1)
-            : [a1] "r" (a1)
-        );
+  if (a1 < 0) {
+    /* use quad-byte
+     * input and output memory are four byte aligned
+     */
+    __asm__ __volatile__(
+        "abs        %[absa1],     %[a1]         \n\t"
+        "replv.qb   %[vector_a1], %[absa1]      \n\t"
+        : [absa1] "=r"(absa1), [vector_a1] "=r"(vector_a1)
+        : [a1] "r"(a1));
 
-        /* use (a1 - predptr[c]) instead a1 + predptr[c] */
-        for (i = 4; i--;)
-        {
-            __asm__ __volatile__ (
-                "lw             %[t2],       0(%[pred_ptr])                     \n\t"
-                "add            %[pred_ptr], %[pred_ptr],    %[pred_stride]     \n\t"
-                "subu_s.qb      %[vector_a], %[t2],          %[vector_a1]       \n\t"
-                "sw             %[vector_a], 0(%[dst_ptr])                      \n\t"
-                "add            %[dst_ptr],  %[dst_ptr],     %[dst_stride]      \n\t"
-                : [t2] "=&r" (t2), [vector_a] "=&r" (vector_a),
-                  [dst_ptr] "+&r" (dst_ptr), [pred_ptr] "+&r" (pred_ptr)
-                : [dst_stride] "r" (dst_stride), [pred_stride] "r" (pred_stride), [vector_a1] "r" (vector_a1)
-            );
-        }
+    /* use (a1 - predptr[c]) instead a1 + predptr[c] */
+    for (i = 4; i--;) {
+      __asm__ __volatile__(
+          "lw             %[t2],       0(%[pred_ptr])                     \n\t"
+          "add            %[pred_ptr], %[pred_ptr],    %[pred_stride]     \n\t"
+          "subu_s.qb      %[vector_a], %[t2],          %[vector_a1]       \n\t"
+          "sw             %[vector_a], 0(%[dst_ptr])                      \n\t"
+          "add            %[dst_ptr],  %[dst_ptr],     %[dst_stride]      \n\t"
+          : [t2] "=&r"(t2), [vector_a] "=&r"(vector_a),
+            [dst_ptr] "+&r"(dst_ptr), [pred_ptr] "+&r"(pred_ptr)
+          : [dst_stride] "r"(dst_stride), [pred_stride] "r"(pred_stride),
+            [vector_a1] "r"(vector_a1));
     }
-    else
-    {
-        /* use quad-byte
-         * input and output memory are four byte aligned
-         */
-        __asm__ __volatile__ (
-            "replv.qb       %[vector_a1], %[a1]     \n\t"
-            : [vector_a1] "=r" (vector_a1)
-            : [a1] "r" (a1)
-        );
+  } else {
+    /* use quad-byte
+     * input and output memory are four byte aligned
+     */
+    __asm__ __volatile__("replv.qb       %[vector_a1], %[a1]     \n\t"
+                         : [vector_a1] "=r"(vector_a1)
+                         : [a1] "r"(a1));
 
-        for (i = 4; i--;)
-        {
-            __asm__ __volatile__ (
-                "lw             %[t2],       0(%[pred_ptr])                 \n\t"
-                "add            %[pred_ptr], %[pred_ptr],    %[pred_stride] \n\t"
-                "addu_s.qb      %[vector_a], %[vector_a1],   %[t2]          \n\t"
-                "sw             %[vector_a], 0(%[dst_ptr])                  \n\t"
-                "add            %[dst_ptr],  %[dst_ptr],     %[dst_stride]  \n\t"
-                : [t2] "=&r" (t2), [vector_a] "=&r" (vector_a),
-                  [dst_ptr] "+&r" (dst_ptr), [pred_ptr] "+&r" (pred_ptr)
-                : [dst_stride] "r" (dst_stride), [pred_stride] "r" (pred_stride), [vector_a1] "r" (vector_a1)
-            );
-        }
+    for (i = 4; i--;) {
+      __asm__ __volatile__(
+          "lw             %[t2],       0(%[pred_ptr])                 \n\t"
+          "add            %[pred_ptr], %[pred_ptr],    %[pred_stride] \n\t"
+          "addu_s.qb      %[vector_a], %[vector_a1],   %[t2]          \n\t"
+          "sw             %[vector_a], 0(%[dst_ptr])                  \n\t"
+          "add            %[dst_ptr],  %[dst_ptr],     %[dst_stride]  \n\t"
+          : [t2] "=&r"(t2), [vector_a] "=&r"(vector_a),
+            [dst_ptr] "+&r"(dst_ptr), [pred_ptr] "+&r"(pred_ptr)
+          : [dst_stride] "r"(dst_stride), [pred_stride] "r"(pred_stride),
+            [vector_a1] "r"(vector_a1));
     }
-
+  }
 }
 
-void vp8_short_inv_walsh4x4_dspr2(short *input, short *mb_dqcoeff)
-{
-    short output[16];
-    int i;
-    int a1, b1, c1, d1;
-    int a2, b2, c2, d2;
-    short *ip = input;
-    short *op = output;
+void vp8_short_inv_walsh4x4_dspr2(short *input, short *mb_dqcoeff) {
+  short output[16];
+  int i;
+  int a1, b1, c1, d1;
+  int a2, b2, c2, d2;
+  short *ip = input;
+  short *op = output;
 
-    prefetch_load_short(ip);
+  prefetch_load_short(ip);
 
-    for (i = 4; i--;)
-    {
-        a1 = ip[0] + ip[12];
-        b1 = ip[4] + ip[8];
-        c1 = ip[4] - ip[8];
-        d1 = ip[0] - ip[12];
+  for (i = 4; i--;) {
+    a1 = ip[0] + ip[12];
+    b1 = ip[4] + ip[8];
+    c1 = ip[4] - ip[8];
+    d1 = ip[0] - ip[12];
 
-        op[0] = a1 + b1;
-        op[4] = c1 + d1;
-        op[8] = a1 - b1;
-        op[12] = d1 - c1;
+    op[0] = a1 + b1;
+    op[4] = c1 + d1;
+    op[8] = a1 - b1;
+    op[12] = d1 - c1;
 
-        ip++;
-        op++;
-    }
+    ip++;
+    op++;
+  }
 
-    ip = output;
-    op = output;
+  ip = output;
+  op = output;
 
-    prefetch_load_short(ip);
+  prefetch_load_short(ip);
 
-    for (i = 4; i--;)
-    {
-        a1 = ip[0] + ip[3] + 3;
-        b1 = ip[1] + ip[2];
-        c1 = ip[1] - ip[2];
-        d1 = ip[0] - ip[3] + 3;
+  for (i = 4; i--;) {
+    a1 = ip[0] + ip[3] + 3;
+    b1 = ip[1] + ip[2];
+    c1 = ip[1] - ip[2];
+    d1 = ip[0] - ip[3] + 3;
 
-        a2 = a1 + b1;
-        b2 = d1 + c1;
-        c2 = a1 - b1;
-        d2 = d1 - c1;
+    a2 = a1 + b1;
+    b2 = d1 + c1;
+    c2 = a1 - b1;
+    d2 = d1 - c1;
 
-        op[0] = a2 >> 3;
-        op[1] = b2 >> 3;
-        op[2] = c2 >> 3;
-        op[3] = d2 >> 3;
+    op[0] = a2 >> 3;
+    op[1] = b2 >> 3;
+    op[2] = c2 >> 3;
+    op[3] = d2 >> 3;
 
-        ip += 4;
-        op += 4;
-    }
+    ip += 4;
+    op += 4;
+  }
 
-    for (i = 0; i < 16; i++)
-    {
-        mb_dqcoeff[i * 16] = output[i];
-    }
+  for (i = 0; i < 16; i++) {
+    mb_dqcoeff[i * 16] = output[i];
+  }
 }
 
-void vp8_short_inv_walsh4x4_1_dspr2(short *input, short *mb_dqcoeff)
-{
-    int a1;
+void vp8_short_inv_walsh4x4_1_dspr2(short *input, short *mb_dqcoeff) {
+  int a1;
 
-    a1 = ((input[0] + 3) >> 3);
+  a1 = ((input[0] + 3) >> 3);
 
-    __asm__ __volatile__ (
-        "sh             %[a1], 0(%[mb_dqcoeff])                    \n\t"
-        "sh             %[a1], 32(%[mb_dqcoeff])                   \n\t"
-        "sh             %[a1], 64(%[mb_dqcoeff])                   \n\t"
-        "sh             %[a1], 96(%[mb_dqcoeff])                   \n\t"
-        "sh             %[a1], 128(%[mb_dqcoeff])                  \n\t"
-        "sh             %[a1], 160(%[mb_dqcoeff])                  \n\t"
-        "sh             %[a1], 192(%[mb_dqcoeff])                  \n\t"
-        "sh             %[a1], 224(%[mb_dqcoeff])                  \n\t"
-        "sh             %[a1], 256(%[mb_dqcoeff])                  \n\t"
-        "sh             %[a1], 288(%[mb_dqcoeff])                  \n\t"
-        "sh             %[a1], 320(%[mb_dqcoeff])                  \n\t"
-        "sh             %[a1], 352(%[mb_dqcoeff])                  \n\t"
-        "sh             %[a1], 384(%[mb_dqcoeff])                  \n\t"
-        "sh             %[a1], 416(%[mb_dqcoeff])                  \n\t"
-        "sh             %[a1], 448(%[mb_dqcoeff])                  \n\t"
-        "sh             %[a1], 480(%[mb_dqcoeff])                  \n\t"
+  __asm__ __volatile__(
+      "sh             %[a1], 0(%[mb_dqcoeff])                    \n\t"
+      "sh             %[a1], 32(%[mb_dqcoeff])                   \n\t"
+      "sh             %[a1], 64(%[mb_dqcoeff])                   \n\t"
+      "sh             %[a1], 96(%[mb_dqcoeff])                   \n\t"
+      "sh             %[a1], 128(%[mb_dqcoeff])                  \n\t"
+      "sh             %[a1], 160(%[mb_dqcoeff])                  \n\t"
+      "sh             %[a1], 192(%[mb_dqcoeff])                  \n\t"
+      "sh             %[a1], 224(%[mb_dqcoeff])                  \n\t"
+      "sh             %[a1], 256(%[mb_dqcoeff])                  \n\t"
+      "sh             %[a1], 288(%[mb_dqcoeff])                  \n\t"
+      "sh             %[a1], 320(%[mb_dqcoeff])                  \n\t"
+      "sh             %[a1], 352(%[mb_dqcoeff])                  \n\t"
+      "sh             %[a1], 384(%[mb_dqcoeff])                  \n\t"
+      "sh             %[a1], 416(%[mb_dqcoeff])                  \n\t"
+      "sh             %[a1], 448(%[mb_dqcoeff])                  \n\t"
+      "sh             %[a1], 480(%[mb_dqcoeff])                  \n\t"
 
-        :
-        : [a1] "r" (a1), [mb_dqcoeff] "r" (mb_dqcoeff)
-    );
+      :
+      : [a1] "r"(a1), [mb_dqcoeff] "r"(mb_dqcoeff));
 }
 
 #endif
--- a/vp8/common/mips/dspr2/reconinter_dspr2.c
+++ b/vp8/common/mips/dspr2/reconinter_dspr2.c
@@ -8,114 +8,90 @@
  *  be found in the AUTHORS file in the root of the source tree.
  */
 
-
 #include "vpx_config.h"
 #include "vp8_rtcd.h"
 #include "vpx/vpx_integer.h"
 
 #if HAVE_DSPR2
-inline void prefetch_load_int(unsigned char *src)
-{
-    __asm__ __volatile__ (
-        "pref   0,  0(%[src])   \n\t"
-        :
-        : [src] "r" (src)
-    );
+inline void prefetch_load_int(unsigned char *src) {
+  __asm__ __volatile__("pref   0,  0(%[src])   \n\t" : : [src] "r"(src));
 }
 
+__inline void vp8_copy_mem16x16_dspr2(unsigned char *RESTRICT src,
+                                      int src_stride,
+                                      unsigned char *RESTRICT dst,
+                                      int dst_stride) {
+  int r;
+  unsigned int a0, a1, a2, a3;
 
-__inline void vp8_copy_mem16x16_dspr2(
-    unsigned char *RESTRICT src,
-    int src_stride,
-    unsigned char *RESTRICT dst,
-    int dst_stride)
-{
-    int r;
-    unsigned int a0, a1, a2, a3;
+  for (r = 16; r--;) {
+    /* load src data in cache memory */
+    prefetch_load_int(src + src_stride);
 
-    for (r = 16; r--;)
-    {
-        /* load src data in cache memory */
-        prefetch_load_int(src + src_stride);
+    /* use unaligned memory load and store */
+    __asm__ __volatile__(
+        "ulw    %[a0], 0(%[src])            \n\t"
+        "ulw    %[a1], 4(%[src])            \n\t"
+        "ulw    %[a2], 8(%[src])            \n\t"
+        "ulw    %[a3], 12(%[src])           \n\t"
+        "sw     %[a0], 0(%[dst])            \n\t"
+        "sw     %[a1], 4(%[dst])            \n\t"
+        "sw     %[a2], 8(%[dst])            \n\t"
+        "sw     %[a3], 12(%[dst])           \n\t"
+        : [a0] "=&r"(a0), [a1] "=&r"(a1), [a2] "=&r"(a2), [a3] "=&r"(a3)
+        : [src] "r"(src), [dst] "r"(dst));
 
-        /* use unaligned memory load and store */
-        __asm__ __volatile__ (
-            "ulw    %[a0], 0(%[src])            \n\t"
-            "ulw    %[a1], 4(%[src])            \n\t"
-            "ulw    %[a2], 8(%[src])            \n\t"
-            "ulw    %[a3], 12(%[src])           \n\t"
-            "sw     %[a0], 0(%[dst])            \n\t"
-            "sw     %[a1], 4(%[dst])            \n\t"
-            "sw     %[a2], 8(%[dst])            \n\t"
-            "sw     %[a3], 12(%[dst])           \n\t"
-            : [a0] "=&r" (a0), [a1] "=&r" (a1),
-              [a2] "=&r" (a2), [a3] "=&r" (a3)
-            : [src] "r" (src), [dst] "r" (dst)
-        );
-
-        src += src_stride;
-        dst += dst_stride;
-    }
+    src += src_stride;
+    dst += dst_stride;
+  }
 }
 
-
-__inline void vp8_copy_mem8x8_dspr2(
-    unsigned char *RESTRICT src,
-    int src_stride,
-    unsigned char *RESTRICT dst,
-    int dst_stride)
-{
-    int r;
-    unsigned int a0, a1;
+__inline void vp8_copy_mem8x8_dspr2(unsigned char *RESTRICT src, int src_stride,
+                                    unsigned char *RESTRICT dst,
+                                    int dst_stride) {
+  int r;
+  unsigned int a0, a1;
 
-    /* load src data in cache memory */
-    prefetch_load_int(src + src_stride);
+  /* load src data in cache memory */
+  prefetch_load_int(src + src_stride);
 
-    for (r = 8; r--;)
-    {
-        /* use unaligned memory load and store */
-        __asm__ __volatile__ (
-            "ulw    %[a0], 0(%[src])            \n\t"
-            "ulw    %[a1], 4(%[src])            \n\t"
-            "sw     %[a0], 0(%[dst])            \n\t"
-            "sw     %[a1], 4(%[dst])            \n\t"
-            : [a0] "=&r" (a0), [a1] "=&r" (a1)
-            : [src] "r" (src), [dst] "r" (dst)
-        );
+  for (r = 8; r--;) {
+    /* use unaligned memory load and store */
+    __asm__ __volatile__(
+        "ulw    %[a0], 0(%[src])            \n\t"
+        "ulw    %[a1], 4(%[src])            \n\t"
+        "sw     %[a0], 0(%[dst])            \n\t"
+        "sw     %[a1], 4(%[dst])            \n\t"
+        : [a0] "=&r"(a0), [a1] "=&r"(a1)
+        : [src] "r"(src), [dst] "r"(dst));
 
-        src += src_stride;
-        dst += dst_stride;
-    }
+    src += src_stride;
+    dst += dst_stride;
+  }
 }
 
-
-__inline void vp8_copy_mem8x4_dspr2(
-    unsigned char *RESTRICT src,
-    int src_stride,
-    unsigned char *RESTRICT dst,
-    int dst_stride)
-{
-    int r;
-    unsigned int a0, a1;
+__inline void vp8_copy_mem8x4_dspr2(unsigned char *RESTRICT src, int src_stride,
+                                    unsigned char *RESTRICT dst,
+                                    int dst_stride) {
+  int r;
+  unsigned int a0, a1;
 
-    /* load src data in cache memory */
-    prefetch_load_int(src + src_stride);
+  /* load src data in cache memory */
+  prefetch_load_int(src + src_stride);
 
-    for (r = 4; r--;)
-    {
-        /* use unaligned memory load and store */
-        __asm__ __volatile__ (
-            "ulw    %[a0], 0(%[src])            \n\t"
-            "ulw    %[a1], 4(%[src])            \n\t"
-            "sw     %[a0], 0(%[dst])            \n\t"
-            "sw     %[a1], 4(%[dst])            \n\t"
-           : [a0] "=&r" (a0), [a1] "=&r" (a1)
-           : [src] "r" (src), [dst] "r" (dst)
-        );
+  for (r = 4; r--;) {
+    /* use unaligned memory load and store */
+    __asm__ __volatile__(
+        "ulw    %[a0], 0(%[src])            \n\t"
+        "ulw    %[a1], 4(%[src])            \n\t"
+        "sw     %[a0], 0(%[dst])            \n\t"
+        "sw     %[a1], 4(%[dst])            \n\t"
+        : [a0] "=&r"(a0), [a1] "=&r"(a1)
+        : [src] "r"(src), [dst] "r"(dst));
 
-        src += src_stride;
-        dst += dst_stride;
-    }
+    src += src_stride;
+    dst += dst_stride;
+  }
 }
 
 #endif
--- a/vp8/common/mips/dspr2/vp8_loopfilter_filters_dspr2.c
+++ b/vp8/common/mips/dspr2/vp8_loopfilter_filters_dspr2.c
@@ -8,7 +8,6 @@
  *  be found in the AUTHORS file in the root of the source tree.
  */
 
-
 #include <stdlib.h>
 #include "vp8_rtcd.h"
 #include "vp8/common/onyxc_int.h"
@@ -17,1047 +16,636 @@
 typedef unsigned char uc;
 
 /* prefetch data for load */
-inline void prefetch_load_lf(unsigned char *src)
-{
-    __asm__ __volatile__ (
-        "pref   0,  0(%[src])   \n\t"
-        :
-        : [src] "r" (src)
-    );
+inline void prefetch_load_lf(unsigned char *src) {
+  __asm__ __volatile__("pref   0,  0(%[src])   \n\t" : : [src] "r"(src));
 }
 
-
 /* prefetch data for store */
-inline void prefetch_store_lf(unsigned char *dst)
-{
-    __asm__ __volatile__ (
-        "pref   1,  0(%[dst])   \n\t"
-        :
-        : [dst] "r" (dst)
-    );
+inline void prefetch_store_lf(unsigned char *dst) {
+  __asm__ __volatile__("pref   1,  0(%[dst])   \n\t" : : [dst] "r"(dst));
 }
 
 /* processing 4 pixels at the same time
  * compute hev and mask in the same function
  */
-static __inline void vp8_filter_mask_vec_mips
-(
-    uint32_t limit,
-    uint32_t flimit,
-    uint32_t p1,
-    uint32_t p0,
-    uint32_t p3,
-    uint32_t p2,
-    uint32_t q0,
-    uint32_t q1,
-    uint32_t q2,
-    uint32_t q3,
-    uint32_t thresh,
-    uint32_t *hev,
-    uint32_t *mask
-)
-{
-    uint32_t c, r, r3, r_k;
-    uint32_t s1, s2, s3;
-    uint32_t ones = 0xFFFFFFFF;
-    uint32_t hev1;
+static __inline void vp8_filter_mask_vec_mips(
+    uint32_t limit, uint32_t flimit, uint32_t p1, uint32_t p0, uint32_t p3,
+    uint32_t p2, uint32_t q0, uint32_t q1, uint32_t q2, uint32_t q3,
+    uint32_t thresh, uint32_t *hev, uint32_t *mask) {
+  uint32_t c, r, r3, r_k;
+  uint32_t s1, s2, s3;
+  uint32_t ones = 0xFFFFFFFF;
+  uint32_t hev1;
 
-    __asm__ __volatile__ (
-        /* mask |= (abs(p3 - p2) > limit) */
-        "subu_s.qb      %[c],   %[p3],     %[p2]        \n\t"
-        "subu_s.qb      %[r_k], %[p2],     %[p3]        \n\t"
-        "or             %[r_k], %[r_k],    %[c]         \n\t"
-        "cmpgu.lt.qb    %[c],   %[limit],  %[r_k]       \n\t"
-        "or             %[r],   $0,        %[c]         \n\t"
+  __asm__ __volatile__(
+      /* mask |= (abs(p3 - p2) > limit) */
+      "subu_s.qb      %[c],   %[p3],     %[p2]        \n\t"
+      "subu_s.qb      %[r_k], %[p2],     %[p3]        \n\t"
+      "or             %[r_k], %[r_k],    %[c]         \n\t"
+      "cmpgu.lt.qb    %[c],   %[limit],  %[r_k]       \n\t"
+      "or             %[r],   $0,        %[c]         \n\t"
 
-        /* mask |= (abs(p2 - p1) > limit) */
-        "subu_s.qb      %[c],   %[p2],     %[p1]        \n\t"
-        "subu_s.qb      %[r_k], %[p1],     %[p2]        \n\t"
-        "or             %[r_k], %[r_k],    %[c]         \n\t"
-        "cmpgu.lt.qb    %[c],   %[limit],  %[r_k]       \n\t"
-        "or             %[r],   %[r],      %[c]         \n\t"
+      /* mask |= (abs(p2 - p1) > limit) */
+      "subu_s.qb      %[c],   %[p2],     %[p1]        \n\t"
+      "subu_s.qb      %[r_k], %[p1],     %[p2]        \n\t"
+      "or             %[r_k], %[r_k],    %[c]         \n\t"
+      "cmpgu.lt.qb    %[c],   %[limit],  %[r_k]       \n\t"
+      "or             %[r],   %[r],      %[c]         \n\t"
 
-        /* mask |= (abs(p1 - p0) > limit)
-         * hev  |= (abs(p1 - p0) > thresh)
-         */
-        "subu_s.qb      %[c],   %[p1],     %[p0]        \n\t"
-        "subu_s.qb      %[r_k], %[p0],     %[p1]        \n\t"
-        "or             %[r_k], %[r_k],    %[c]         \n\t"
-        "cmpgu.lt.qb    %[c],   %[thresh], %[r_k]       \n\t"
-        "or             %[r3],  $0,        %[c]         \n\t"
-        "cmpgu.lt.qb    %[c],   %[limit],  %[r_k]       \n\t"
-        "or             %[r],   %[r],      %[c]         \n\t"
+      /* mask |= (abs(p1 - p0) > limit)
+       * hev  |= (abs(p1 - p0) > thresh)
+       */
+      "subu_s.qb      %[c],   %[p1],     %[p0]        \n\t"
+      "subu_s.qb      %[r_k], %[p0],     %[p1]        \n\t"
+      "or             %[r_k], %[r_k],    %[c]         \n\t"
+      "cmpgu.lt.qb    %[c],   %[thresh], %[r_k]       \n\t"
+      "or             %[r3],  $0,        %[c]         \n\t"
+      "cmpgu.lt.qb    %[c],   %[limit],  %[r_k]       \n\t"
+      "or             %[r],   %[r],      %[c]         \n\t"
 
-        /* mask |= (abs(q1 - q0) > limit)
-         * hev  |= (abs(q1 - q0) > thresh)
-         */
-        "subu_s.qb      %[c],   %[q1],     %[q0]        \n\t"
-        "subu_s.qb      %[r_k], %[q0],     %[q1]        \n\t"
-        "or             %[r_k], %[r_k],    %[c]         \n\t"
-        "cmpgu.lt.qb    %[c],   %[thresh], %[r_k]       \n\t"
-        "or             %[r3],  %[r3],     %[c]         \n\t"
-        "cmpgu.lt.qb    %[c],   %[limit],  %[r_k]       \n\t"
-        "or             %[r],   %[r],      %[c]         \n\t"
+      /* mask |= (abs(q1 - q0) > limit)
+       * hev  |= (abs(q1 - q0) > thresh)
+       */
+      "subu_s.qb      %[c],   %[q1],     %[q0]        \n\t"
+      "subu_s.qb      %[r_k], %[q0],     %[q1]        \n\t"
+      "or             %[r_k], %[r_k],    %[c]         \n\t"
+      "cmpgu.lt.qb    %[c],   %[thresh], %[r_k]       \n\t"
+      "or             %[r3],  %[r3],     %[c]         \n\t"
+      "cmpgu.lt.qb    %[c],   %[limit],  %[r_k]       \n\t"
+      "or             %[r],   %[r],      %[c]         \n\t"
 
-        /* mask |= (abs(q2 - q1) > limit) */
-        "subu_s.qb      %[c],   %[q2],     %[q1]        \n\t"
-        "subu_s.qb      %[r_k], %[q1],     %[q2]        \n\t"
-        "or             %[r_k], %[r_k],    %[c]         \n\t"
-        "cmpgu.lt.qb    %[c],   %[limit],  %[r_k]       \n\t"
-        "or             %[r],   %[r],      %[c]         \n\t"
-        "sll            %[r3],    %[r3],    24          \n\t"
+      /* mask |= (abs(q2 - q1) > limit) */
+      "subu_s.qb      %[c],   %[q2],     %[q1]        \n\t"
+      "subu_s.qb      %[r_k], %[q1],     %[q2]        \n\t"
+      "or             %[r_k], %[r_k],    %[c]         \n\t"
+      "cmpgu.lt.qb    %[c],   %[limit],  %[r_k]       \n\t"
+      "or             %[r],   %[r],      %[c]         \n\t"
+      "sll            %[r3],    %[r3],    24          \n\t"
 
-        /* mask |= (abs(q3 - q2) > limit) */
-        "subu_s.qb      %[c],   %[q3],     %[q2]        \n\t"
-        "subu_s.qb      %[r_k], %[q2],     %[q3]        \n\t"
-        "or             %[r_k], %[r_k],    %[c]         \n\t"
-        "cmpgu.lt.qb    %[c],   %[limit],  %[r_k]       \n\t"
-        "or             %[r],   %[r],      %[c]         \n\t"
+      /* mask |= (abs(q3 - q2) > limit) */
+      "subu_s.qb      %[c],   %[q3],     %[q2]        \n\t"
+      "subu_s.qb      %[r_k], %[q2],     %[q3]        \n\t"
+      "or             %[r_k], %[r_k],    %[c]         \n\t"
+      "cmpgu.lt.qb    %[c],   %[limit],  %[r_k]       \n\t"
+      "or             %[r],   %[r],      %[c]         \n\t"
 
-        : [c] "=&r" (c), [r_k] "=&r" (r_k),
-          [r] "=&r" (r), [r3] "=&r" (r3)
-        : [limit] "r" (limit), [p3] "r" (p3), [p2] "r" (p2),
-          [p1] "r" (p1), [p0] "r" (p0), [q1] "r" (q1), [q0] "r" (q0),
-          [q2] "r" (q2), [q3] "r" (q3), [thresh] "r" (thresh)
-    );
+      : [c] "=&r"(c), [r_k] "=&r"(r_k), [r] "=&r"(r), [r3] "=&r"(r3)
+      : [limit] "r"(limit), [p3] "r"(p3), [p2] "r"(p2), [p1] "r"(p1),
+        [p0] "r"(p0), [q1] "r"(q1), [q0] "r"(q0), [q2] "r"(q2), [q3] "r"(q3),
+        [thresh] "r"(thresh));
 
-    __asm__ __volatile__ (
-        /* abs(p0 - q0) */
-        "subu_s.qb      %[c],   %[p0],     %[q0]        \n\t"
-        "subu_s.qb      %[r_k], %[q0],     %[p0]        \n\t"
-        "wrdsp          %[r3]                           \n\t"
-        "or             %[s1],  %[r_k],    %[c]         \n\t"
+  __asm__ __volatile__(
+      /* abs(p0 - q0) */
+      "subu_s.qb      %[c],   %[p0],     %[q0]        \n\t"
+      "subu_s.qb      %[r_k], %[q0],     %[p0]        \n\t"
+      "wrdsp          %[r3]                           \n\t"
+      "or             %[s1],  %[r_k],    %[c]         \n\t"
 
-        /* abs(p1 - q1) */
-        "subu_s.qb      %[c],    %[p1],    %[q1]        \n\t"
-        "addu_s.qb      %[s3],   %[s1],    %[s1]        \n\t"
-        "pick.qb        %[hev1], %[ones],  $0           \n\t"
-        "subu_s.qb      %[r_k],  %[q1],    %[p1]        \n\t"
-        "or             %[s2],   %[r_k],   %[c]         \n\t"
+      /* abs(p1 - q1) */
+      "subu_s.qb      %[c],    %[p1],    %[q1]        \n\t"
+      "addu_s.qb      %[s3],   %[s1],    %[s1]        \n\t"
+      "pick.qb        %[hev1], %[ones],  $0           \n\t"
+      "subu_s.qb      %[r_k],  %[q1],    %[p1]        \n\t"
+      "or             %[s2],   %[r_k],   %[c]         \n\t"
 
-        /* abs(p0 - q0) * 2 + abs(p1 - q1) / 2  > flimit * 2 + limit */
-        "shrl.qb        %[s2],   %[s2],     1           \n\t"
-        "addu_s.qb      %[s1],   %[s2],     %[s3]       \n\t"
-        "cmpgu.lt.qb    %[c],    %[flimit], %[s1]       \n\t"
-        "or             %[r],    %[r],      %[c]        \n\t"
-        "sll            %[r],    %[r],      24          \n\t"
+      /* abs(p0 - q0) * 2 + abs(p1 - q1) / 2  > flimit * 2 + limit */
+      "shrl.qb        %[s2],   %[s2],     1           \n\t"
+      "addu_s.qb      %[s1],   %[s2],     %[s3]       \n\t"
+      "cmpgu.lt.qb    %[c],    %[flimit], %[s1]       \n\t"
+      "or             %[r],    %[r],      %[c]        \n\t"
+      "sll            %[r],    %[r],      24          \n\t"
 
-        "wrdsp          %[r]                            \n\t"
-        "pick.qb        %[s2],  $0,         %[ones]     \n\t"
+      "wrdsp          %[r]                            \n\t"
+      "pick.qb        %[s2],  $0,         %[ones]     \n\t"
 
-        : [c] "=&r" (c), [r_k] "=&r" (r_k), [s1] "=&r" (s1), [hev1] "=&r" (hev1),
-          [s2] "=&r" (s2), [r] "+r" (r), [s3] "=&r" (s3)
-        : [p0] "r" (p0), [q0] "r" (q0), [p1] "r" (p1), [r3] "r" (r3),
-          [q1] "r" (q1), [ones] "r" (ones), [flimit] "r" (flimit)
-    );
+      : [c] "=&r"(c), [r_k] "=&r"(r_k), [s1] "=&r"(s1), [hev1] "=&r"(hev1),
+        [s2] "=&r"(s2), [r] "+r"(r), [s3] "=&r"(s3)
+      : [p0] "r"(p0), [q0] "r"(q0), [p1] "r"(p1), [r3] "r"(r3), [q1] "r"(q1),
+        [ones] "r"(ones), [flimit] "r"(flimit));
 
-    *hev = hev1;
-    *mask = s2;
+  *hev = hev1;
+  *mask = s2;
 }
 
-
 /* inputs & outputs are quad-byte vectors */
-static __inline void vp8_filter_mips
-(
-    uint32_t mask,
-    uint32_t hev,
-    uint32_t *ps1,
-    uint32_t *ps0,
-    uint32_t *qs0,
-    uint32_t *qs1
-)
-{
-    int32_t vp8_filter_l, vp8_filter_r;
-    int32_t Filter1_l, Filter1_r, Filter2_l, Filter2_r;
-    int32_t subr_r, subr_l;
-    uint32_t t1, t2, HWM, t3;
-    uint32_t hev_l, hev_r, mask_l, mask_r, invhev_l, invhev_r;
+static __inline void vp8_filter_mips(uint32_t mask, uint32_t hev, uint32_t *ps1,
+                                     uint32_t *ps0, uint32_t *qs0,
+                                     uint32_t *qs1) {
+  int32_t vp8_filter_l, vp8_filter_r;
+  int32_t Filter1_l, Filter1_r, Filter2_l, Filter2_r;
+  int32_t subr_r, subr_l;
+  uint32_t t1, t2, HWM, t3;
+  uint32_t hev_l, hev_r, mask_l, mask_r, invhev_l, invhev_r;
 
-    int32_t vps1, vps0, vqs0, vqs1;
-    int32_t vps1_l, vps1_r, vps0_l, vps0_r, vqs0_l, vqs0_r, vqs1_l, vqs1_r;
-    uint32_t N128;
+  int32_t vps1, vps0, vqs0, vqs1;
+  int32_t vps1_l, vps1_r, vps0_l, vps0_r, vqs0_l, vqs0_r, vqs1_l, vqs1_r;
+  uint32_t N128;
 
-    N128 = 0x80808080;
-    t1  = 0x03000300;
-    t2  = 0x04000400;
-    t3  = 0x01000100;
-    HWM = 0xFF00FF00;
+  N128 = 0x80808080;
+  t1 = 0x03000300;
+  t2 = 0x04000400;
+  t3 = 0x01000100;
+  HWM = 0xFF00FF00;
 
-    vps0 = (*ps0) ^ N128;
-    vps1 = (*ps1) ^ N128;
-    vqs0 = (*qs0) ^ N128;
-    vqs1 = (*qs1) ^ N128;
+  vps0 = (*ps0) ^ N128;
+  vps1 = (*ps1) ^ N128;
+  vqs0 = (*qs0) ^ N128;
+  vqs1 = (*qs1) ^ N128;
 
-    /* use halfword pairs instead quad-bytes because of accuracy */
-    vps0_l = vps0 & HWM;
-    vps0_r = vps0 << 8;
-    vps0_r = vps0_r & HWM;
+  /* use halfword pairs instead quad-bytes because of accuracy */
+  vps0_l = vps0 & HWM;
+  vps0_r = vps0 << 8;
+  vps0_r = vps0_r & HWM;
 
-    vps1_l = vps1 & HWM;
-    vps1_r = vps1 << 8;
-    vps1_r = vps1_r & HWM;
+  vps1_l = vps1 & HWM;
+  vps1_r = vps1 << 8;
+  vps1_r = vps1_r & HWM;
 
-    vqs0_l = vqs0 & HWM;
-    vqs0_r = vqs0 << 8;
-    vqs0_r = vqs0_r & HWM;
+  vqs0_l = vqs0 & HWM;
+  vqs0_r = vqs0 << 8;
+  vqs0_r = vqs0_r & HWM;
 
-    vqs1_l = vqs1 & HWM;
-    vqs1_r = vqs1 << 8;
-    vqs1_r = vqs1_r & HWM;
+  vqs1_l = vqs1 & HWM;
+  vqs1_r = vqs1 << 8;
+  vqs1_r = vqs1_r & HWM;
 
-    mask_l = mask & HWM;
-    mask_r = mask << 8;
-    mask_r = mask_r & HWM;
+  mask_l = mask & HWM;
+  mask_r = mask << 8;
+  mask_r = mask_r & HWM;
 
-    hev_l = hev & HWM;
-    hev_r = hev << 8;
-    hev_r = hev_r & HWM;
+  hev_l = hev & HWM;
+  hev_r = hev << 8;
+  hev_r = hev_r & HWM;
 
-    __asm__ __volatile__ (
-        /* vp8_filter = vp8_signed_char_clamp(ps1 - qs1); */
-        "subq_s.ph    %[vp8_filter_l], %[vps1_l],       %[vqs1_l]       \n\t"
-        "subq_s.ph    %[vp8_filter_r], %[vps1_r],       %[vqs1_r]       \n\t"
+  __asm__ __volatile__(
+      /* vp8_filter = vp8_signed_char_clamp(ps1 - qs1); */
+      "subq_s.ph    %[vp8_filter_l], %[vps1_l],       %[vqs1_l]       \n\t"
+      "subq_s.ph    %[vp8_filter_r], %[vps1_r],       %[vqs1_r]       \n\t"
 
-        /* qs0 - ps0 */
-        "subq_s.ph    %[subr_l],       %[vqs0_l],       %[vps0_l]       \n\t"
-        "subq_s.ph    %[subr_r],       %[vqs0_r],       %[vps0_r]       \n\t"
+      /* qs0 - ps0 */
+      "subq_s.ph    %[subr_l],       %[vqs0_l],       %[vps0_l]       \n\t"
+      "subq_s.ph    %[subr_r],       %[vqs0_r],       %[vps0_r]       \n\t"
 
-        /* vp8_filter &= hev; */
-        "and          %[vp8_filter_l], %[vp8_filter_l], %[hev_l]        \n\t"
-        "and          %[vp8_filter_r], %[vp8_filter_r], %[hev_r]        \n\t"
+      /* vp8_filter &= hev; */
+      "and          %[vp8_filter_l], %[vp8_filter_l], %[hev_l]        \n\t"
+      "and          %[vp8_filter_r], %[vp8_filter_r], %[hev_r]        \n\t"
 
-        /* vp8_filter = vp8_signed_char_clamp(vp8_filter + 3 * (qs0 - ps0)); */
-        "addq_s.ph    %[vp8_filter_l], %[vp8_filter_l], %[subr_l]       \n\t"
-        "addq_s.ph    %[vp8_filter_r], %[vp8_filter_r], %[subr_r]       \n\t"
-        "xor          %[invhev_l],     %[hev_l],        %[HWM]          \n\t"
-        "addq_s.ph    %[vp8_filter_l], %[vp8_filter_l], %[subr_l]       \n\t"
-        "addq_s.ph    %[vp8_filter_r], %[vp8_filter_r], %[subr_r]       \n\t"
-        "xor          %[invhev_r],     %[hev_r],        %[HWM]          \n\t"
-        "addq_s.ph    %[vp8_filter_l], %[vp8_filter_l], %[subr_l]       \n\t"
-        "addq_s.ph    %[vp8_filter_r], %[vp8_filter_r], %[subr_r]       \n\t"
+      /* vp8_filter = vp8_signed_char_clamp(vp8_filter + 3 * (qs0 - ps0)); */
+      "addq_s.ph    %[vp8_filter_l], %[vp8_filter_l], %[subr_l]       \n\t"
+      "addq_s.ph    %[vp8_filter_r], %[vp8_filter_r], %[subr_r]       \n\t"
+      "xor          %[invhev_l],     %[hev_l],        %[HWM]          \n\t"
+      "addq_s.ph    %[vp8_filter_l], %[vp8_filter_l], %[subr_l]       \n\t"
+      "addq_s.ph    %[vp8_filter_r], %[vp8_filter_r], %[subr_r]       \n\t"
+      "xor          %[invhev_r],     %[hev_r],        %[HWM]          \n\t"
+      "addq_s.ph    %[vp8_filter_l], %[vp8_filter_l], %[subr_l]       \n\t"
+      "addq_s.ph    %[vp8_filter_r], %[vp8_filter_r], %[subr_r]       \n\t"
 
-        /* vp8_filter &= mask; */
-        "and          %[vp8_filter_l], %[vp8_filter_l], %[mask_l]       \n\t"
-        "and          %[vp8_filter_r], %[vp8_filter_r], %[mask_r]       \n\t"
+      /* vp8_filter &= mask; */
+      "and          %[vp8_filter_l], %[vp8_filter_l], %[mask_l]       \n\t"
+      "and          %[vp8_filter_r], %[vp8_filter_r], %[mask_r]       \n\t"
 
-        : [vp8_filter_l] "=&r" (vp8_filter_l), [vp8_filter_r] "=&r" (vp8_filter_r),
-          [subr_l] "=&r" (subr_l), [subr_r] "=&r" (subr_r),
-          [invhev_l] "=&r" (invhev_l), [invhev_r] "=&r" (invhev_r)
+      : [vp8_filter_l] "=&r"(vp8_filter_l), [vp8_filter_r] "=&r"(vp8_filter_r),
+        [subr_l] "=&r"(subr_l), [subr_r] "=&r"(subr_r),
+        [invhev_l] "=&r"(invhev_l), [invhev_r] "=&r"(invhev_r)
 
-        : [vps0_l] "r" (vps0_l), [vps0_r] "r" (vps0_r), [vps1_l] "r" (vps1_l),
-          [vps1_r] "r" (vps1_r), [vqs0_l] "r" (vqs0_l), [vqs0_r] "r" (vqs0_r),
-          [vqs1_l] "r" (vqs1_l), [vqs1_r] "r" (vqs1_r),
-          [mask_l] "r" (mask_l), [mask_r] "r" (mask_r),
-          [hev_l] "r" (hev_l), [hev_r] "r" (hev_r),
-          [HWM] "r" (HWM)
-    );
+      : [vps0_l] "r"(vps0_l), [vps0_r] "r"(vps0_r), [vps1_l] "r"(vps1_l),
+        [vps1_r] "r"(vps1_r), [vqs0_l] "r"(vqs0_l), [vqs0_r] "r"(vqs0_r),
+        [vqs1_l] "r"(vqs1_l), [vqs1_r] "r"(vqs1_r), [mask_l] "r"(mask_l),
+        [mask_r] "r"(mask_r), [hev_l] "r"(hev_l), [hev_r] "r"(hev_r),
+        [HWM] "r"(HWM));
 
-    /* save bottom 3 bits so that we round one side +4 and the other +3 */
-    __asm__ __volatile__ (
-        /* Filter2 = vp8_signed_char_clamp(vp8_filter + 3) >>= 3; */
-        "addq_s.ph    %[Filter1_l],    %[vp8_filter_l], %[t2]           \n\t"
-        "addq_s.ph    %[Filter1_r],    %[vp8_filter_r], %[t2]           \n\t"
+  /* save bottom 3 bits so that we round one side +4 and the other +3 */
+  __asm__ __volatile__(
+      /* Filter2 = vp8_signed_char_clamp(vp8_filter + 3) >>= 3; */
+      "addq_s.ph    %[Filter1_l],    %[vp8_filter_l], %[t2]           \n\t"
+      "addq_s.ph    %[Filter1_r],    %[vp8_filter_r], %[t2]           \n\t"
 
-        /* Filter1 = vp8_signed_char_clamp(vp8_filter + 4) >>= 3; */
-        "addq_s.ph    %[Filter2_l],    %[vp8_filter_l], %[t1]           \n\t"
-        "addq_s.ph    %[Filter2_r],    %[vp8_filter_r], %[t1]           \n\t"
-        "shra.ph      %[Filter1_r],    %[Filter1_r],    3               \n\t"
-        "shra.ph      %[Filter1_l],    %[Filter1_l],    3               \n\t"
+      /* Filter1 = vp8_signed_char_clamp(vp8_filter + 4) >>= 3; */
+      "addq_s.ph    %[Filter2_l],    %[vp8_filter_l], %[t1]           \n\t"
+      "addq_s.ph    %[Filter2_r],    %[vp8_filter_r], %[t1]           \n\t"
+      "shra.ph      %[Filter1_r],    %[Filter1_r],    3               \n\t"
+      "shra.ph      %[Filter1_l],    %[Filter1_l],    3               \n\t"
 
-        "shra.ph      %[Filter2_l],    %[Filter2_l],    3               \n\t"
-        "shra.ph      %[Filter2_r],    %[Filter2_r],    3               \n\t"
+      "shra.ph      %[Filter2_l],    %[Filter2_l],    3               \n\t"
+      "shra.ph      %[Filter2_r],    %[Filter2_r],    3               \n\t"
 
-        "and          %[Filter1_l],    %[Filter1_l],    %[HWM]          \n\t"
-        "and          %[Filter1_r],    %[Filter1_r],    %[HWM]          \n\t"
+      "and          %[Filter1_l],    %[Filter1_l],    %[HWM]          \n\t"
+      "and          %[Filter1_r],    %[Filter1_r],    %[HWM]          \n\t"
 
-        /* vps0 = vp8_signed_char_clamp(ps0 + Filter2); */
-        "addq_s.ph    %[vps0_l],       %[vps0_l],       %[Filter2_l]    \n\t"
-        "addq_s.ph    %[vps0_r],       %[vps0_r],       %[Filter2_r]    \n\t"
+      /* vps0 = vp8_signed_char_clamp(ps0 + Filter2); */
+      "addq_s.ph    %[vps0_l],       %[vps0_l],       %[Filter2_l]    \n\t"
+      "addq_s.ph    %[vps0_r],       %[vps0_r],       %[Filter2_r]    \n\t"
 
-        /* vqs0 = vp8_signed_char_clamp(qs0 - Filter1); */
-        "subq_s.ph    %[vqs0_l],       %[vqs0_l],       %[Filter1_l]    \n\t"
-        "subq_s.ph    %[vqs0_r],       %[vqs0_r],       %[Filter1_r]    \n\t"
+      /* vqs0 = vp8_signed_char_clamp(qs0 - Filter1); */
+      "subq_s.ph    %[vqs0_l],       %[vqs0_l],       %[Filter1_l]    \n\t"
+      "subq_s.ph    %[vqs0_r],       %[vqs0_r],       %[Filter1_r]    \n\t"
 
-        : [Filter1_l] "=&r" (Filter1_l), [Filter1_r] "=&r" (Filter1_r),
-          [Filter2_l] "=&r" (Filter2_l), [Filter2_r] "=&r" (Filter2_r),
-          [vps0_l] "+r" (vps0_l), [vps0_r] "+r" (vps0_r),
-          [vqs0_l] "+r" (vqs0_l), [vqs0_r] "+r" (vqs0_r)
+      : [Filter1_l] "=&r"(Filter1_l), [Filter1_r] "=&r"(Filter1_r),
+        [Filter2_l] "=&r"(Filter2_l), [Filter2_r] "=&r"(Filter2_r),
+        [vps0_l] "+r"(vps0_l), [vps0_r] "+r"(vps0_r), [vqs0_l] "+r"(vqs0_l),
+        [vqs0_r] "+r"(vqs0_r)
 
-        : [t1] "r" (t1), [t2] "r" (t2),
-          [vp8_filter_l] "r" (vp8_filter_l), [vp8_filter_r] "r" (vp8_filter_r),
-          [HWM] "r" (HWM)
-    );
+      : [t1] "r"(t1), [t2] "r"(t2), [vp8_filter_l] "r"(vp8_filter_l),
+        [vp8_filter_r] "r"(vp8_filter_r), [HWM] "r"(HWM));
 
-    __asm__ __volatile__ (
-        /* (vp8_filter += 1) >>= 1 */
-        "addqh.ph    %[Filter1_l],    %[Filter1_l],     %[t3]           \n\t"
-        "addqh.ph    %[Filter1_r],    %[Filter1_r],     %[t3]           \n\t"
+  __asm__ __volatile__(
+      /* (vp8_filter += 1) >>= 1 */
+      "addqh.ph    %[Filter1_l],    %[Filter1_l],     %[t3]           \n\t"
+      "addqh.ph    %[Filter1_r],    %[Filter1_r],     %[t3]           \n\t"
 
-        /* vp8_filter &= ~hev; */
-        "and          %[Filter1_l],    %[Filter1_l],    %[invhev_l]     \n\t"
-        "and          %[Filter1_r],    %[Filter1_r],    %[invhev_r]     \n\t"
+      /* vp8_filter &= ~hev; */
+      "and          %[Filter1_l],    %[Filter1_l],    %[invhev_l]     \n\t"
+      "and          %[Filter1_r],    %[Filter1_r],    %[invhev_r]     \n\t"
 
-        /* vps1 = vp8_signed_char_clamp(ps1 + vp8_filter); */
-        "addq_s.ph    %[vps1_l],       %[vps1_l],       %[Filter1_l]    \n\t"
-        "addq_s.ph    %[vps1_r],       %[vps1_r],       %[Filter1_r]    \n\t"
+      /* vps1 = vp8_signed_char_clamp(ps1 + vp8_filter); */
+      "addq_s.ph    %[vps1_l],       %[vps1_l],       %[Filter1_l]    \n\t"
+      "addq_s.ph    %[vps1_r],       %[vps1_r],       %[Filter1_r]    \n\t"
 
-        /* vqs1 = vp8_signed_char_clamp(qs1 - vp8_filter); */
-        "subq_s.ph    %[vqs1_l],       %[vqs1_l],       %[Filter1_l]    \n\t"
-        "subq_s.ph    %[vqs1_r],       %[vqs1_r],       %[Filter1_r]    \n\t"
+      /* vqs1 = vp8_signed_char_clamp(qs1 - vp8_filter); */
+      "subq_s.ph    %[vqs1_l],       %[vqs1_l],       %[Filter1_l]    \n\t"
+      "subq_s.ph    %[vqs1_r],       %[vqs1_r],       %[Filter1_r]    \n\t"
 
-        : [Filter1_l] "+r" (Filter1_l), [Filter1_r] "+r" (Filter1_r),
-          [vps1_l] "+r" (vps1_l), [vps1_r] "+r" (vps1_r),
-          [vqs1_l] "+r" (vqs1_l), [vqs1_r] "+r" (vqs1_r)
+      : [Filter1_l] "+r"(Filter1_l), [Filter1_r] "+r"(Filter1_r),
+        [vps1_l] "+r"(vps1_l), [vps1_r] "+r"(vps1_r), [vqs1_l] "+r"(vqs1_l),
+        [vqs1_r] "+r"(vqs1_r)
 
-        : [t3] "r" (t3), [invhev_l] "r" (invhev_l), [invhev_r] "r" (invhev_r)
-    );
+      : [t3] "r"(t3), [invhev_l] "r"(invhev_l), [invhev_r] "r"(invhev_r));
 
-    /* Create quad-bytes from halfword pairs */
-    vqs0_l = vqs0_l & HWM;
-    vqs1_l = vqs1_l & HWM;
-    vps0_l = vps0_l & HWM;
-    vps1_l = vps1_l & HWM;
+  /* Create quad-bytes from halfword pairs */
+  vqs0_l = vqs0_l & HWM;
+  vqs1_l = vqs1_l & HWM;
+  vps0_l = vps0_l & HWM;
+  vps1_l = vps1_l & HWM;
 
-    __asm__ __volatile__ (
-        "shrl.ph      %[vqs0_r],       %[vqs0_r],       8               \n\t"
-        "shrl.ph      %[vps0_r],       %[vps0_r],       8               \n\t"
-        "shrl.ph      %[vqs1_r],       %[vqs1_r],       8               \n\t"
-        "shrl.ph      %[vps1_r],       %[vps1_r],       8               \n\t"
+  __asm__ __volatile__(
+      "shrl.ph      %[vqs0_r],       %[vqs0_r],       8               \n\t"
+      "shrl.ph      %[vps0_r],       %[vps0_r],       8               \n\t"
+      "shrl.ph      %[vqs1_r],       %[vqs1_r],       8               \n\t"
+      "shrl.ph      %[vps1_r],       %[vps1_r],       8               \n\t"
 
-        : [vps1_r] "+r" (vps1_r), [vqs1_r] "+r" (vqs1_r),
-          [vps0_r] "+r" (vps0_r), [vqs0_r] "+r" (vqs0_r)
-        :
-    );
+      : [vps1_r] "+r"(vps1_r), [vqs1_r] "+r"(vqs1_r), [vps0_r] "+r"(vps0_r),
+        [vqs0_r] "+r"(vqs0_r)
+      :);
 
-    vqs0 = vqs0_l | vqs0_r;
-    vqs1 = vqs1_l | vqs1_r;
-    vps0 = vps0_l | vps0_r;
-    vps1 = vps1_l | vps1_r;
+  vqs0 = vqs0_l | vqs0_r;
+  vqs1 = vqs1_l | vqs1_r;
+  vps0 = vps0_l | vps0_r;
+  vps1 = vps1_l | vps1_r;
 
-    *ps0 = vps0 ^ N128;
-    *ps1 = vps1 ^ N128;
-    *qs0 = vqs0 ^ N128;
-    *qs1 = vqs1 ^ N128;
+  *ps0 = vps0 ^ N128;
+  *ps1 = vps1 ^ N128;
+  *qs0 = vqs0 ^ N128;
+  *qs1 = vqs1 ^ N128;
 }
 
-void vp8_loop_filter_horizontal_edge_mips
-(
-    unsigned char *s,
-    int p,
-    unsigned int flimit,
-    unsigned int limit,
-    unsigned int thresh,
-    int count
-)
-{
-    uint32_t mask;
-    uint32_t hev;
-    uint32_t pm1, p0, p1, p2, p3, p4, p5, p6;
-    unsigned char *sm1, *s0, *s1, *s2, *s3, *s4, *s5, *s6;
+void vp8_loop_filter_horizontal_edge_mips(unsigned char *s, int p,
+                                          unsigned int flimit,
+                                          unsigned int limit,
+                                          unsigned int thresh, int count) {
+  uint32_t mask;
+  uint32_t hev;
+  uint32_t pm1, p0, p1, p2, p3, p4, p5, p6;
+  unsigned char *sm1, *s0, *s1, *s2, *s3, *s4, *s5, *s6;
 
-    mask = 0;
-    hev = 0;
-    p1 = 0;
-    p2 = 0;
-    p3 = 0;
-    p4 = 0;
+  mask = 0;
+  hev = 0;
+  p1 = 0;
+  p2 = 0;
+  p3 = 0;
+  p4 = 0;
 
-    /* prefetch data for store */
-    prefetch_store_lf(s);
+  /* prefetch data for store */
+  prefetch_store_lf(s);
 
-    /* loop filter designed to work using chars so that we can make maximum use
-     * of 8 bit simd instructions.
-     */
+  /* loop filter designed to work using chars so that we can make maximum use
+   * of 8 bit simd instructions.
+   */
 
-    sm1 = s - (p << 2);
-    s0 = s - p - p - p;
-    s1 = s - p - p ;
-    s2 = s - p;
-    s3 = s;
-    s4 = s + p;
-    s5 = s + p + p;
-    s6 = s + p + p + p;
+  sm1 = s - (p << 2);
+  s0 = s - p - p - p;
+  s1 = s - p - p;
+  s2 = s - p;
+  s3 = s;
+  s4 = s + p;
+  s5 = s + p + p;
+  s6 = s + p + p + p;
 
-    /* load quad-byte vectors
-     * memory is 4 byte aligned
-     */
-    p1 = *((uint32_t *)(s1));
-    p2 = *((uint32_t *)(s2));
-    p3 = *((uint32_t *)(s3));
-    p4 = *((uint32_t *)(s4));
+  /* load quad-byte vectors
+   * memory is 4 byte aligned
+   */
+  p1 = *((uint32_t *)(s1));
+  p2 = *((uint32_t *)(s2));
+  p3 = *((uint32_t *)(s3));
+  p4 = *((uint32_t *)(s4));
 
-    /* if (p1 - p4 == 0) and (p2 - p3 == 0)
-     * mask will be zero and filtering is not needed
-     */
-    if (!(((p1 - p4) == 0) && ((p2 - p3) == 0)))
-    {
+  /* if (p1 - p4 == 0) and (p2 - p3 == 0)
+   * mask will be zero and filtering is not needed
+   */
+  if (!(((p1 - p4) == 0) && ((p2 - p3) == 0))) {
+    pm1 = *((uint32_t *)(sm1));
+    p0 = *((uint32_t *)(s0));
+    p5 = *((uint32_t *)(s5));
+    p6 = *((uint32_t *)(s6));
 
-        pm1 = *((uint32_t *)(sm1));
-        p0  = *((uint32_t *)(s0));
-        p5  = *((uint32_t *)(s5));
-        p6  = *((uint32_t *)(s6));
+    vp8_filter_mask_vec_mips(limit, flimit, p1, p2, pm1, p0, p3, p4, p5, p6,
+                             thresh, &hev, &mask);
 
-        vp8_filter_mask_vec_mips(limit, flimit, p1, p2, pm1, p0, p3, p4, p5, p6,
-                                 thresh, &hev, &mask);
+    /* if mask == 0 do filtering is not needed */
+    if (mask) {
+      /* filtering */
+      vp8_filter_mips(mask, hev, &p1, &p2, &p3, &p4);
 
-        /* if mask == 0 do filtering is not needed */
-        if (mask)
-        {
-            /* filtering */
-            vp8_filter_mips(mask, hev, &p1, &p2, &p3, &p4);
-
-            /* unpack processed 4x4 neighborhood */
-            *((uint32_t *)s1) = p1;
-            *((uint32_t *)s2) = p2;
-            *((uint32_t *)s3) = p3;
-            *((uint32_t *)s4) = p4;
-        }
+      /* unpack processed 4x4 neighborhood */
+      *((uint32_t *)s1) = p1;
+      *((uint32_t *)s2) = p2;
+      *((uint32_t *)s3) = p3;
+      *((uint32_t *)s4) = p4;
     }
+  }
 
-    sm1 += 4;
-    s0  += 4;
-    s1  += 4;
-    s2  += 4;
-    s3  += 4;
-    s4  += 4;
-    s5  += 4;
-    s6  += 4;
+  sm1 += 4;
+  s0 += 4;
+  s1 += 4;
+  s2 += 4;
+  s3 += 4;
+  s4 += 4;
+  s5 += 4;
+  s6 += 4;
 
-    /* load quad-byte vectors
-     * memory is 4 byte aligned
-     */
-    p1 = *((uint32_t *)(s1));
-    p2 = *((uint32_t *)(s2));
-    p3 = *((uint32_t *)(s3));
-    p4 = *((uint32_t *)(s4));
+  /* load quad-byte vectors
+   * memory is 4 byte aligned
+   */
+  p1 = *((uint32_t *)(s1));
+  p2 = *((uint32_t *)(s2));
+  p3 = *((uint32_t *)(s3));
+  p4 = *((uint32_t *)(s4));
 
-    /* if (p1 - p4 == 0) and (p2 - p3 == 0)
-     * mask will be zero and filtering is not needed
-     */
-    if (!(((p1 - p4) == 0) && ((p2 - p3) == 0)))
-    {
+  /* if (p1 - p4 == 0) and (p2 - p3 == 0)
+   * mask will be zero and filtering is not needed
+   */
+  if (!(((p1 - p4) == 0) && ((p2 - p3) == 0))) {
+    pm1 = *((uint32_t *)(sm1));
+    p0 = *((uint32_t *)(s0));
+    p5 = *((uint32_t *)(s5));
+    p6 = *((uint32_t *)(s6));
 
-        pm1 = *((uint32_t *)(sm1));
-        p0  = *((uint32_t *)(s0));
-        p5  = *((uint32_t *)(s5));
-        p6  = *((uint32_t *)(s6));
+    vp8_filter_mask_vec_mips(limit, flimit, p1, p2, pm1, p0, p3, p4, p5, p6,
+                             thresh, &hev, &mask);
 
-        vp8_filter_mask_vec_mips(limit, flimit, p1, p2, pm1, p0, p3, p4, p5, p6,
-                                 thresh, &hev, &mask);
+    /* if mask == 0 do filtering is not needed */
+    if (mask) {
+      /* filtering */
+      vp8_filter_mips(mask, hev, &p1, &p2, &p3, &p4);
 
-        /* if mask == 0 do filtering is not needed */
-        if (mask)
-        {
-            /* filtering */
-            vp8_filter_mips(mask, hev, &p1, &p2, &p3, &p4);
-
-            /* unpack processed 4x4 neighborhood */
-            *((uint32_t *)s1) = p1;
-            *((uint32_t *)s2) = p2;
-            *((uint32_t *)s3) = p3;
-            *((uint32_t *)s4) = p4;
-        }
+      /* unpack processed 4x4 neighborhood */
+      *((uint32_t *)s1) = p1;
+      *((uint32_t *)s2) = p2;
+      *((uint32_t *)s3) = p3;
+      *((uint32_t *)s4) = p4;
     }
+  }
 
-    sm1 += 4;
-    s0  += 4;
-    s1  += 4;
-    s2  += 4;
-    s3  += 4;
-    s4  += 4;
-    s5  += 4;
-    s6  += 4;
+  sm1 += 4;
+  s0 += 4;
+  s1 += 4;
+  s2 += 4;
+  s3 += 4;
+  s4 += 4;
+  s5 += 4;
+  s6 += 4;
 
-    /* load quad-byte vectors
-     * memory is 4 byte aligned
-     */
-    p1 = *((uint32_t *)(s1));
-    p2 = *((uint32_t *)(s2));
-    p3 = *((uint32_t *)(s3));
-    p4 = *((uint32_t *)(s4));
+  /* load quad-byte vectors
+   * memory is 4 byte aligned
+   */
+  p1 = *((uint32_t *)(s1));
+  p2 = *((uint32_t *)(s2));
+  p3 = *((uint32_t *)(s3));
+  p4 = *((uint32_t *)(s4));
 
-    /* if (p1 - p4 == 0) and (p2 - p3 == 0)
-     * mask will be zero and filtering is not needed
-     */
-    if (!(((p1 - p4) == 0) && ((p2 - p3) == 0)))
-    {
+  /* if (p1 - p4 == 0) and (p2 - p3 == 0)
+   * mask will be zero and filtering is not needed
+   */
+  if (!(((p1 - p4) == 0) && ((p2 - p3) == 0))) {
+    pm1 = *((uint32_t *)(sm1));
+    p0 = *((uint32_t *)(s0));
+    p5 = *((uint32_t *)(s5));
+    p6 = *((uint32_t *)(s6));
 
-        pm1 = *((uint32_t *)(sm1));
-        p0  = *((uint32_t *)(s0));
-        p5  = *((uint32_t *)(s5));
-        p6  = *((uint32_t *)(s6));
+    vp8_filter_mask_vec_mips(limit, flimit, p1, p2, pm1, p0, p3, p4, p5, p6,
+                             thresh, &hev, &mask);
 
-        vp8_filter_mask_vec_mips(limit, flimit, p1, p2, pm1, p0, p3, p4, p5, p6,
-                                 thresh, &hev, &mask);
+    /* if mask == 0 do filtering is not needed */
+    if (mask) {
+      /* filtering */
+      vp8_filter_mips(mask, hev, &p1, &p2, &p3, &p4);
 
-        /* if mask == 0 do filtering is not needed */
-        if (mask)
-        {
-            /* filtering */
-            vp8_filter_mips(mask, hev, &p1, &p2, &p3, &p4);
-
-            /* unpack processed 4x4 neighborhood */
-            *((uint32_t *)s1) = p1;
-            *((uint32_t *)s2) = p2;
-            *((uint32_t *)s3) = p3;
-            *((uint32_t *)s4) = p4;
-        }
+      /* unpack processed 4x4 neighborhood */
+      *((uint32_t *)s1) = p1;
+      *((uint32_t *)s2) = p2;
+      *((uint32_t *)s3) = p3;
+      *((uint32_t *)s4) = p4;
     }
+  }
 
-    sm1 += 4;
-    s0  += 4;
-    s1  += 4;
-    s2  += 4;
-    s3  += 4;
-    s4  += 4;
-    s5  += 4;
-    s6  += 4;
+  sm1 += 4;
+  s0 += 4;
+  s1 += 4;
+  s2 += 4;
+  s3 += 4;
+  s4 += 4;
+  s5 += 4;
+  s6 += 4;
 
-    /* load quad-byte vectors
-     * memory is 4 byte aligned
-     */
-    p1 = *((uint32_t *)(s1));
-    p2 = *((uint32_t *)(s2));
-    p3 = *((uint32_t *)(s3));
-    p4 = *((uint32_t *)(s4));
+  /* load quad-byte vectors
+   * memory is 4 byte aligned
+   */
+  p1 = *((uint32_t *)(s1));
+  p2 = *((uint32_t *)(s2));
+  p3 = *((uint32_t *)(s3));
+  p4 = *((uint32_t *)(s4));
 
-    /* if (p1 - p4 == 0) and (p2 - p3 == 0)
-     * mask will be zero and filtering is not needed
-     */
-    if (!(((p1 - p4) == 0) && ((p2 - p3) == 0)))
-    {
+  /* if (p1 - p4 == 0) and (p2 - p3 == 0)
+   * mask will be zero and filtering is not needed
+   */
+  if (!(((p1 - p4) == 0) && ((p2 - p3) == 0))) {
+    pm1 = *((uint32_t *)(sm1));
+    p0 = *((uint32_t *)(s0));
+    p5 = *((uint32_t *)(s5));
+    p6 = *((uint32_t *)(s6));
 
-        pm1 = *((uint32_t *)(sm1));
-        p0  = *((uint32_t *)(s0));
-        p5  = *((uint32_t *)(s5));
-        p6  = *((uint32_t *)(s6));
+    vp8_filter_mask_vec_mips(limit, flimit, p1, p2, pm1, p0, p3, p4, p5, p6,
+                             thresh, &hev, &mask);
 
-        vp8_filter_mask_vec_mips(limit, flimit, p1, p2, pm1, p0, p3, p4, p5, p6,
-                                 thresh, &hev, &mask);
+    /* if mask == 0 do filtering is not needed */
+    if (mask) {
+      /* filtering */
+      vp8_filter_mips(mask, hev, &p1, &p2, &p3, &p4);
 
-        /* if mask == 0 do filtering is not needed */
-        if (mask)
-        {
-            /* filtering */
-            vp8_filter_mips(mask, hev, &p1, &p2, &p3, &p4);
-
-            /* unpack processed 4x4 neighborhood */
-            *((uint32_t *)s1) = p1;
-            *((uint32_t *)s2) = p2;
-            *((uint32_t *)s3) = p3;
-            *((uint32_t *)s4) = p4;
-        }
+      /* unpack processed 4x4 neighborhood */
+      *((uint32_t *)s1) = p1;
+      *((uint32_t *)s2) = p2;
+      *((uint32_t *)s3) = p3;
+      *((uint32_t *)s4) = p4;
     }
+  }
 }
 
-void vp8_loop_filter_uvhorizontal_edge_mips
-(
-    unsigned char *s,
-    int p,
-    unsigned int flimit,
-    unsigned int limit,
-    unsigned int thresh,
-    int count
-)
-{
-    uint32_t mask;
-    uint32_t hev;
-    uint32_t pm1, p0, p1, p2, p3, p4, p5, p6;
-    unsigned char *sm1, *s0, *s1, *s2, *s3, *s4, *s5, *s6;
+void vp8_loop_filter_uvhorizontal_edge_mips(unsigned char *s, int p,
+                                            unsigned int flimit,
+                                            unsigned int limit,
+                                            unsigned int thresh, int count) {
+  uint32_t mask;
+  uint32_t hev;
+  uint32_t pm1, p0, p1, p2, p3, p4, p5, p6;
+  unsigned char *sm1, *s0, *s1, *s2, *s3, *s4, *s5, *s6;
 
-    mask = 0;
-    hev = 0;
-    p1 = 0;
-    p2 = 0;
-    p3 = 0;
-    p4 = 0;
+  mask = 0;
+  hev = 0;
+  p1 = 0;
+  p2 = 0;
+  p3 = 0;
+  p4 = 0;
 
-    /* loop filter designed to work using chars so that we can make maximum use
-     * of 8 bit simd instructions.
-     */
+  /* loop filter designed to work using chars so that we can make maximum use
+   * of 8 bit simd instructions.
+   */
 
-    sm1 = s - (p << 2);
-    s0  = s - p - p - p;
-    s1  = s - p - p ;
-    s2  = s - p;
-    s3  = s;
-    s4  = s + p;
-    s5  = s + p + p;
-    s6  = s + p + p + p;
+  sm1 = s - (p << 2);
+  s0 = s - p - p - p;
+  s1 = s - p - p;
+  s2 = s - p;
+  s3 = s;
+  s4 = s + p;
+  s5 = s + p + p;
+  s6 = s + p + p + p;
 
-    /* load quad-byte vectors
-     * memory is 4 byte aligned
-     */
-    p1 = *((uint32_t *)(s1));
-    p2 = *((uint32_t *)(s2));
-    p3 = *((uint32_t *)(s3));
-    p4 = *((uint32_t *)(s4));
+  /* load quad-byte vectors
+   * memory is 4 byte aligned
+   */
+  p1 = *((uint32_t *)(s1));
+  p2 = *((uint32_t *)(s2));
+  p3 = *((uint32_t *)(s3));
+  p4 = *((uint32_t *)(s4));
 
-    /* if (p1 - p4 == 0) and (p2 - p3 == 0)
-     * mask will be zero and filtering is not needed
-     */
-    if (!(((p1 - p4) == 0) && ((p2 - p3) == 0)))
-    {
+  /* if (p1 - p4 == 0) and (p2 - p3 == 0)
+   * mask will be zero and filtering is not needed
+   */
+  if (!(((p1 - p4) == 0) && ((p2 - p3) == 0))) {
+    pm1 = *((uint32_t *)(sm1));
+    p0 = *((uint32_t *)(s0));
+    p5 = *((uint32_t *)(s5));
+    p6 = *((uint32_t *)(s6));
 
-        pm1 = *((uint32_t *)(sm1));
-        p0  = *((uint32_t *)(s0));
-        p5  = *((uint32_t *)(s5));
-        p6  = *((uint32_t *)(s6));
+    vp8_filter_mask_vec_mips(limit, flimit, p1, p2, pm1, p0, p3, p4, p5, p6,
+                             thresh, &hev, &mask);
 
-        vp8_filter_mask_vec_mips(limit, flimit, p1, p2, pm1, p0, p3, p4, p5, p6,
-                                 thresh, &hev, &mask);
+    /* if mask == 0 do filtering is not needed */
+    if (mask) {
+      /* filtering */
+      vp8_filter_mips(mask, hev, &p1, &p2, &p3, &p4);
 
-        /* if mask == 0 do filtering is not needed */
-        if (mask)
-        {
-            /* filtering */
-            vp8_filter_mips(mask, hev, &p1, &p2, &p3, &p4);
-
-            /* unpack processed 4x4 neighborhood */
-            *((uint32_t *)s1) = p1;
-            *((uint32_t *)s2) = p2;
-            *((uint32_t *)s3) = p3;
-            *((uint32_t *)s4) = p4;
-        }
+      /* unpack processed 4x4 neighborhood */
+      *((uint32_t *)s1) = p1;
+      *((uint32_t *)s2) = p2;
+      *((uint32_t *)s3) = p3;
+      *((uint32_t *)s4) = p4;
     }
+  }
 
-    sm1 += 4;
-    s0  += 4;
-    s1  += 4;
-    s2  += 4;
-    s3  += 4;
-    s4  += 4;
-    s5  += 4;
-    s6  += 4;
+  sm1 += 4;
+  s0 += 4;
+  s1 += 4;
+  s2 += 4;
+  s3 += 4;
+  s4 += 4;
+  s5 += 4;
+  s6 += 4;
 
-    /* load quad-byte vectors
-     * memory is 4 byte aligned
-     */
-    p1 = *((uint32_t *)(s1));
-    p2 = *((uint32_t *)(s2));
-    p3 = *((uint32_t *)(s3));
-    p4 = *((uint32_t *)(s4));
+  /* load quad-byte vectors
+   * memory is 4 byte aligned
+   */
+  p1 = *((uint32_t *)(s1));
+  p2 = *((uint32_t *)(s2));
+  p3 = *((uint32_t *)(s3));
+  p4 = *((uint32_t *)(s4));
 
-    /* if (p1 - p4 == 0) and (p2 - p3 == 0)
-     * mask will be zero and filtering is not needed
-     */
-    if (!(((p1 - p4) == 0) && ((p2 - p3) == 0)))
-    {
+  /* if (p1 - p4 == 0) and (p2 - p3 == 0)
+   * mask will be zero and filtering is not needed
+   */
+  if (!(((p1 - p4) == 0) && ((p2 - p3) == 0))) {
+    pm1 = *((uint32_t *)(sm1));
+    p0 = *((uint32_t *)(s0));
+    p5 = *((uint32_t *)(s5));
+    p6 = *((uint32_t *)(s6));
 
-        pm1 = *((uint32_t *)(sm1));
-        p0  = *((uint32_t *)(s0));
-        p5  = *((uint32_t *)(s5));
-        p6  = *((uint32_t *)(s6));
+    vp8_filter_mask_vec_mips(limit, flimit, p1, p2, pm1, p0, p3, p4, p5, p6,
+                             thresh, &hev, &mask);
 
-        vp8_filter_mask_vec_mips(limit, flimit, p1, p2, pm1, p0, p3, p4, p5, p6,
-                                 thresh, &hev, &mask);
+    /* if mask == 0 do filtering is not needed */
+    if (mask) {
+      /* filtering */
+      vp8_filter_mips(mask, hev, &p1, &p2, &p3, &p4);
 
-        /* if mask == 0 do filtering is not needed */
-        if (mask)
-        {
-            /* filtering */
-            vp8_filter_mips(mask, hev, &p1, &p2, &p3, &p4);
-
-            /* unpack processed 4x4 neighborhood */
-            *((uint32_t *)s1) = p1;
-            *((uint32_t *)s2) = p2;
-            *((uint32_t *)s3) = p3;
-            *((uint32_t *)s4) = p4;
-        }
+      /* unpack processed 4x4 neighborhood */
+      *((uint32_t *)s1) = p1;
+      *((uint32_t *)s2) = p2;
+      *((uint32_t *)s3) = p3;
+      *((uint32_t *)s4) = p4;
     }
+  }
 }
 
-void vp8_loop_filter_vertical_edge_mips
-(
-    unsigned char *s,
-    int p,
-    const unsigned int flimit,
-    const unsigned int limit,
-    const unsigned int thresh,
-    int count
-)
-{
-    int i;
-    uint32_t mask, hev;
-    uint32_t pm1, p0, p1, p2, p3, p4, p5, p6;
-    unsigned char *s1, *s2, *s3, *s4;
-    uint32_t prim1, prim2, sec3, sec4, prim3, prim4;
+void vp8_loop_filter_vertical_edge_mips(unsigned char *s, int p,
+                                        const unsigned int flimit,
+                                        const unsigned int limit,
+                                        const unsigned int thresh, int count) {
+  int i;
+  uint32_t mask, hev;
+  uint32_t pm1, p0, p1, p2, p3, p4, p5, p6;
+  unsigned char *s1, *s2, *s3, *s4;
+  uint32_t prim1, prim2, sec3, sec4, prim3, prim4;
 
-    hev = 0;
-    mask = 0;
-    i = 0;
-    pm1 = 0;
-    p0 = 0;
-    p1 = 0;
-    p2 = 0;
-    p3 = 0;
-    p4 = 0;
-    p5 = 0;
-    p6 = 0;
+  hev = 0;
+  mask = 0;
+  i = 0;
+  pm1 = 0;
+  p0 = 0;
+  p1 = 0;
+  p2 = 0;
+  p3 = 0;
+  p4 = 0;
+  p5 = 0;
+  p6 = 0;
 
-    /* loop filter designed to work using chars so that we can make maximum use
-     * of 8 bit simd instructions.
-     */
+  /* loop filter designed to work using chars so that we can make maximum use
+   * of 8 bit simd instructions.
+   */
 
-    /* apply filter on 4 pixesl at the same time */
-    do
-    {
+  /* apply filter on 4 pixesl at the same time */
+  do {
+    /* prefetch data for store */
+    prefetch_store_lf(s + p);
 
-        /* prefetch data for store */
-        prefetch_store_lf(s + p);
-
-        s1 = s;
-        s2 = s + p;
-        s3 = s2 + p;
-        s4 = s3 + p;
-        s  = s4 + p;
-
-        /* load quad-byte vectors
-         * memory is 4 byte aligned
-         */
-        p2  = *((uint32_t *)(s1 - 4));
-        p6  = *((uint32_t *)(s1));
-        p1  = *((uint32_t *)(s2 - 4));
-        p5  = *((uint32_t *)(s2));
-        p0  = *((uint32_t *)(s3 - 4));
-        p4  = *((uint32_t *)(s3));
-        pm1 = *((uint32_t *)(s4 - 4));
-        p3  = *((uint32_t *)(s4));
-
-        /* transpose pm1, p0, p1, p2 */
-        __asm__ __volatile__ (
-            "precrq.qb.ph   %[prim1],   %[p2],      %[p1]       \n\t"
-            "precr.qb.ph    %[prim2],   %[p2],      %[p1]       \n\t"
-            "precrq.qb.ph   %[prim3],   %[p0],      %[pm1]      \n\t"
-            "precr.qb.ph    %[prim4],   %[p0],      %[pm1]      \n\t"
-
-            "precrq.qb.ph   %[p1],      %[prim1],   %[prim2]    \n\t"
-            "precr.qb.ph    %[pm1],     %[prim1],   %[prim2]    \n\t"
-            "precrq.qb.ph   %[sec3],    %[prim3],   %[prim4]    \n\t"
-            "precr.qb.ph    %[sec4],    %[prim3],   %[prim4]    \n\t"
-
-            "precrq.ph.w    %[p2],      %[p1],      %[sec3]     \n\t"
-            "precrq.ph.w    %[p0],      %[pm1],     %[sec4]     \n\t"
-            "append         %[p1],      %[sec3],    16          \n\t"
-            "append         %[pm1],     %[sec4],    16          \n\t"
-
-            : [prim1] "=&r" (prim1), [prim2] "=&r" (prim2),
-              [prim3] "=&r" (prim3), [prim4] "=&r" (prim4),
-              [p2] "+r" (p2), [p1] "+r" (p1), [p0] "+r" (p0), [pm1] "+r" (pm1),
-              [sec3] "=&r" (sec3), [sec4] "=&r" (sec4)
-            :
-        );
-
-        /* transpose p3, p4, p5, p6 */
-        __asm__ __volatile__ (
-            "precrq.qb.ph   %[prim1],   %[p6],      %[p5]       \n\t"
-            "precr.qb.ph    %[prim2],   %[p6],      %[p5]       \n\t"
-            "precrq.qb.ph   %[prim3],   %[p4],      %[p3]       \n\t"
-            "precr.qb.ph    %[prim4],   %[p4],      %[p3]       \n\t"
-
-            "precrq.qb.ph   %[p5],      %[prim1],   %[prim2]    \n\t"
-            "precr.qb.ph    %[p3],      %[prim1],   %[prim2]    \n\t"
-            "precrq.qb.ph   %[sec3],    %[prim3],   %[prim4]    \n\t"
-            "precr.qb.ph    %[sec4],    %[prim3],   %[prim4]    \n\t"
-
-            "precrq.ph.w    %[p6],      %[p5],      %[sec3]     \n\t"
-            "precrq.ph.w    %[p4],      %[p3],      %[sec4]     \n\t"
-            "append         %[p5],      %[sec3],    16          \n\t"
-            "append         %[p3],      %[sec4],    16          \n\t"
-
-            : [prim1] "=&r" (prim1), [prim2] "=&r" (prim2),
-              [prim3] "=&r" (prim3), [prim4] "=&r" (prim4),
-              [p6] "+r" (p6), [p5] "+r" (p5), [p4] "+r" (p4), [p3] "+r" (p3),
-              [sec3] "=&r" (sec3), [sec4] "=&r" (sec4)
-            :
-        );
-
-        /* if (p1 - p4 == 0) and (p2 - p3 == 0)
-         * mask will be zero and filtering is not needed
-         */
-        if (!(((p1 - p4) == 0) && ((p2 - p3) == 0)))
-        {
-
-            vp8_filter_mask_vec_mips(limit, flimit, p1, p2, pm1, p0, p3, p4, p5, p6,
-                                     thresh, &hev, &mask);
-
-            /* if mask == 0 do filtering is not needed */
-            if (mask)
-            {
-                /* filtering */
-                vp8_filter_mips(mask, hev, &p1, &p2, &p3, &p4);
-
-                /* unpack processed 4x4 neighborhood
-                 * don't use transpose on output data
-                 * because memory isn't aligned
-                 */
-                __asm__ __volatile__ (
-                    "sb         %[p4],  1(%[s4])    \n\t"
-                    "sb         %[p3],  0(%[s4])    \n\t"
-                    "sb         %[p2], -1(%[s4])    \n\t"
-                    "sb         %[p1], -2(%[s4])    \n\t"
-                    :
-                    : [p4] "r" (p4), [p3] "r" (p3), [s4] "r" (s4),
-                      [p2] "r" (p2), [p1] "r" (p1)
-                );
-
-                __asm__ __volatile__ (
-                    "srl        %[p4], %[p4], 8     \n\t"
-                    "srl        %[p3], %[p3], 8     \n\t"
-                    "srl        %[p2], %[p2], 8     \n\t"
-                    "srl        %[p1], %[p1], 8     \n\t"
-                    : [p4] "+r" (p4), [p3] "+r" (p3), [p2] "+r" (p2), [p1] "+r" (p1)
-                    :
-                );
-
-                __asm__ __volatile__ (
-                    "sb         %[p4],  1(%[s3])    \n\t"
-                    "sb         %[p3],  0(%[s3])    \n\t"
-                    "sb         %[p2], -1(%[s3])    \n\t"
-                    "sb         %[p1], -2(%[s3])    \n\t"
-                    : [p1] "+r" (p1)
-                    : [p4] "r" (p4), [p3] "r" (p3), [s3] "r" (s3), [p2] "r" (p2)
-                );
-
-                __asm__ __volatile__ (
-                    "srl        %[p4], %[p4], 8     \n\t"
-                    "srl        %[p3], %[p3], 8     \n\t"
-                    "srl        %[p2], %[p2], 8     \n\t"
-                    "srl        %[p1], %[p1], 8     \n\t"
-                    : [p4] "+r" (p4), [p3] "+r" (p3), [p2] "+r" (p2), [p1] "+r" (p1)
-                    :
-                );
-
-                __asm__ __volatile__ (
-                    "sb         %[p4],  1(%[s2])    \n\t"
-                    "sb         %[p3],  0(%[s2])    \n\t"
-                    "sb         %[p2], -1(%[s2])    \n\t"
-                    "sb         %[p1], -2(%[s2])    \n\t"
-                    :
-                    : [p4] "r" (p4), [p3] "r" (p3), [s2] "r" (s2),
-                      [p2] "r" (p2), [p1] "r" (p1)
-                );
-
-                __asm__ __volatile__ (
-                    "srl        %[p4], %[p4], 8     \n\t"
-                    "srl        %[p3], %[p3], 8     \n\t"
-                    "srl        %[p2], %[p2], 8     \n\t"
-                    "srl        %[p1], %[p1], 8     \n\t"
-                    : [p4] "+r" (p4), [p3] "+r" (p3), [p2] "+r" (p2), [p1] "+r" (p1)
-                    :
-                );
-
-                __asm__ __volatile__ (
-                    "sb         %[p4],  1(%[s1])    \n\t"
-                    "sb         %[p3],  0(%[s1])    \n\t"
-                    "sb         %[p2], -1(%[s1])    \n\t"
-                    "sb         %[p1], -2(%[s1])    \n\t"
-                    :
-                    : [p4] "r" (p4), [p3] "r" (p3), [s1] "r" (s1),
-                      [p2] "r" (p2), [p1] "r" (p1)
-                );
-            }
-        }
-
-        s1 = s;
-        s2 = s + p;
-        s3 = s2 + p;
-        s4 = s3 + p;
-        s  = s4 + p;
-
-        /* load quad-byte vectors
-         * memory is 4 byte aligned
-         */
-        p2  = *((uint32_t *)(s1 - 4));
-        p6  = *((uint32_t *)(s1));
-        p1  = *((uint32_t *)(s2 - 4));
-        p5  = *((uint32_t *)(s2));
-        p0  = *((uint32_t *)(s3 - 4));
-        p4  = *((uint32_t *)(s3));
-        pm1 = *((uint32_t *)(s4 - 4));
-        p3  = *((uint32_t *)(s4));
-
-        /* transpose pm1, p0, p1, p2 */
-        __asm__ __volatile__ (
-            "precrq.qb.ph   %[prim1],   %[p2],      %[p1]       \n\t"
-            "precr.qb.ph    %[prim2],   %[p2],      %[p1]       \n\t"
-            "precrq.qb.ph   %[prim3],   %[p0],      %[pm1]      \n\t"
-            "precr.qb.ph    %[prim4],   %[p0],      %[pm1]      \n\t"
-
-            "precrq.qb.ph   %[p1],      %[prim1],   %[prim2]    \n\t"
-            "precr.qb.ph    %[pm1],     %[prim1],   %[prim2]    \n\t"
-            "precrq.qb.ph   %[sec3],    %[prim3],   %[prim4]    \n\t"
-            "precr.qb.ph    %[sec4],    %[prim3],   %[prim4]    \n\t"
-
-            "precrq.ph.w    %[p2],      %[p1],      %[sec3]     \n\t"
-            "precrq.ph.w    %[p0],      %[pm1],     %[sec4]     \n\t"
-            "append         %[p1],      %[sec3],    16          \n\t"
-            "append         %[pm1],     %[sec4],    16          \n\t"
-
-            : [prim1] "=&r" (prim1), [prim2] "=&r" (prim2),
-              [prim3] "=&r" (prim3), [prim4] "=&r" (prim4),
-              [p2] "+r" (p2), [p1] "+r" (p1), [p0] "+r" (p0), [pm1] "+r" (pm1),
-              [sec3] "=&r" (sec3), [sec4] "=&r" (sec4)
-            :
-        );
-
-        /* transpose p3, p4, p5, p6 */
-        __asm__ __volatile__ (
-            "precrq.qb.ph   %[prim1],   %[p6],      %[p5]       \n\t"
-            "precr.qb.ph    %[prim2],   %[p6],      %[p5]       \n\t"
-            "precrq.qb.ph   %[prim3],   %[p4],      %[p3]       \n\t"
-            "precr.qb.ph    %[prim4],   %[p4],      %[p3]       \n\t"
-
-            "precrq.qb.ph   %[p5],      %[prim1],   %[prim2]    \n\t"
-            "precr.qb.ph    %[p3],      %[prim1],   %[prim2]    \n\t"
-            "precrq.qb.ph   %[sec3],    %[prim3],   %[prim4]    \n\t"
-            "precr.qb.ph    %[sec4],    %[prim3],   %[prim4]    \n\t"
-
-            "precrq.ph.w    %[p6],      %[p5],      %[sec3]     \n\t"
-            "precrq.ph.w    %[p4],      %[p3],      %[sec4]     \n\t"
-            "append         %[p5],      %[sec3],    16          \n\t"
-            "append         %[p3],      %[sec4],    16          \n\t"
-
-            : [prim1] "=&r" (prim1), [prim2] "=&r" (prim2),
-              [prim3] "=&r" (prim3), [prim4] "=&r" (prim4),
-              [p6] "+r" (p6), [p5] "+r" (p5), [p4] "+r" (p4), [p3] "+r" (p3),
-              [sec3] "=&r" (sec3), [sec4] "=&r" (sec4)
-            :
-        );
-
-        /* if (p1 - p4 == 0) and (p2 - p3 == 0)
-         * mask will be zero and filtering is not needed
-         */
-        if (!(((p1 - p4) == 0) && ((p2 - p3) == 0)))
-        {
-
-            vp8_filter_mask_vec_mips(limit, flimit, p1, p2, pm1, p0, p3, p4, p5, p6,
-                                     thresh, &hev, &mask);
-
-            /* if mask == 0 do filtering is not needed */
-            if (mask)
-            {
-                /* filtering */
-                vp8_filter_mips(mask, hev, &p1, &p2, &p3, &p4);
-
-                /* unpack processed 4x4 neighborhood
-                 * don't use transpose on output data
-                 * because memory isn't aligned
-                 */
-                __asm__ __volatile__ (
-                    "sb         %[p4],  1(%[s4])    \n\t"
-                    "sb         %[p3],  0(%[s4])    \n\t"
-                    "sb         %[p2], -1(%[s4])    \n\t"
-                    "sb         %[p1], -2(%[s4])    \n\t"
-                    :
-                    : [p4] "r" (p4), [p3] "r" (p3), [s4] "r" (s4),
-                      [p2] "r" (p2), [p1] "r" (p1)
-                );
-
-                __asm__ __volatile__ (
-                    "srl        %[p4], %[p4], 8     \n\t"
-                    "srl        %[p3], %[p3], 8     \n\t"
-                    "srl        %[p2], %[p2], 8     \n\t"
-                    "srl        %[p1], %[p1], 8     \n\t"
-                    : [p4] "+r" (p4), [p3] "+r" (p3), [p2] "+r" (p2), [p1] "+r" (p1)
-                    :
-                );
-
-                __asm__ __volatile__ (
-                    "sb         %[p4],  1(%[s3])    \n\t"
-                    "sb         %[p3],  0(%[s3])    \n\t"
-                    "sb         %[p2], -1(%[s3])    \n\t"
-                    "sb         %[p1], -2(%[s3])    \n\t"
-                    : [p1] "+r" (p1)
-                    : [p4] "r" (p4), [p3] "r" (p3), [s3] "r" (s3), [p2] "r" (p2)
-                );
-
-                __asm__ __volatile__ (
-                    "srl        %[p4], %[p4], 8     \n\t"
-                    "srl        %[p3], %[p3], 8     \n\t"
-                    "srl        %[p2], %[p2], 8     \n\t"
-                    "srl        %[p1], %[p1], 8     \n\t"
-                    : [p4] "+r" (p4), [p3] "+r" (p3), [p2] "+r" (p2), [p1] "+r" (p1)
-                    :
-                );
-
-                __asm__ __volatile__ (
-                    "sb         %[p4],  1(%[s2])    \n\t"
-                    "sb         %[p3],  0(%[s2])    \n\t"
-                    "sb         %[p2], -1(%[s2])    \n\t"
-                    "sb         %[p1], -2(%[s2])    \n\t"
-                    :
-                    : [p4] "r" (p4), [p3] "r" (p3), [s2] "r" (s2),
-                      [p2] "r" (p2), [p1] "r" (p1)
-                );
-
-                __asm__ __volatile__ (
-                    "srl        %[p4], %[p4], 8     \n\t"
-                    "srl        %[p3], %[p3], 8     \n\t"
-                    "srl        %[p2], %[p2], 8     \n\t"
-                    "srl        %[p1], %[p1], 8     \n\t"
-                    : [p4] "+r" (p4), [p3] "+r" (p3), [p2] "+r" (p2), [p1] "+r" (p1)
-                    :
-                );
-
-                __asm__ __volatile__ (
-                    "sb         %[p4],  1(%[s1])    \n\t"
-                    "sb         %[p3],  0(%[s1])    \n\t"
-                    "sb         %[p2], -1(%[s1])    \n\t"
-                    "sb         %[p1], -2(%[s1])    \n\t"
-                    :
-                    : [p4] "r" (p4), [p3] "r" (p3), [s1] "r" (s1),
-                      [p2] "r" (p2), [p1] "r" (p1)
-                );
-            }
-        }
-
-        i += 8;
-    }
-
-    while (i < count);
-}
-
-void vp8_loop_filter_uvvertical_edge_mips
-(
-    unsigned char *s,
-    int p,
-    unsigned int flimit,
-    unsigned int limit,
-    unsigned int thresh,
-    int count
-)
-{
-    uint32_t mask, hev;
-    uint32_t pm1, p0, p1, p2, p3, p4, p5, p6;
-    unsigned char *s1, *s2, *s3, *s4;
-    uint32_t prim1, prim2, sec3, sec4, prim3, prim4;
-
-    /* loop filter designed to work using chars so that we can make maximum use
-     * of 8 bit simd instructions.
-     */
-
-    /* apply filter on 4 pixesl at the same time */
-
     s1 = s;
     s2 = s + p;
     s3 = s2 + p;
     s4 = s3 + p;
+    s = s4 + p;
 
     /* load quad-byte vectors
-    * memory is 4 byte aligned
-    */
-    p2  = *((uint32_t *)(s1 - 4));
-    p6  = *((uint32_t *)(s1));
-    p1  = *((uint32_t *)(s2 - 4));
-    p5  = *((uint32_t *)(s2));
-    p0  = *((uint32_t *)(s3 - 4));
-    p4  = *((uint32_t *)(s3));
+     * memory is 4 byte aligned
+     */
+    p2 = *((uint32_t *)(s1 - 4));
+    p6 = *((uint32_t *)(s1));
+    p1 = *((uint32_t *)(s2 - 4));
+    p5 = *((uint32_t *)(s2));
+    p0 = *((uint32_t *)(s3 - 4));
+    p4 = *((uint32_t *)(s3));
     pm1 = *((uint32_t *)(s4 - 4));
-    p3  = *((uint32_t *)(s4));
+    p3 = *((uint32_t *)(s4));
 
     /* transpose pm1, p0, p1, p2 */
-    __asm__ __volatile__ (
+    __asm__ __volatile__(
         "precrq.qb.ph   %[prim1],   %[p2],      %[p1]       \n\t"
         "precr.qb.ph    %[prim2],   %[p2],      %[p1]       \n\t"
         "precrq.qb.ph   %[prim3],   %[p0],      %[pm1]      \n\t"
@@ -1073,15 +661,13 @@
         "append         %[p1],      %[sec3],    16          \n\t"
         "append         %[pm1],     %[sec4],    16          \n\t"
 
-        : [prim1] "=&r" (prim1), [prim2] "=&r" (prim2),
-          [prim3] "=&r" (prim3), [prim4] "=&r" (prim4),
-          [p2] "+r" (p2), [p1] "+r" (p1), [p0] "+r" (p0), [pm1] "+r" (pm1),
-          [sec3] "=&r" (sec3), [sec4] "=&r" (sec4)
-        :
-    );
+        : [prim1] "=&r"(prim1), [prim2] "=&r"(prim2), [prim3] "=&r"(prim3),
+          [prim4] "=&r"(prim4), [p2] "+r"(p2), [p1] "+r"(p1), [p0] "+r"(p0),
+          [pm1] "+r"(pm1), [sec3] "=&r"(sec3), [sec4] "=&r"(sec4)
+        :);
 
     /* transpose p3, p4, p5, p6 */
-    __asm__ __volatile__ (
+    __asm__ __volatile__(
         "precrq.qb.ph   %[prim1],   %[p6],      %[p5]       \n\t"
         "precr.qb.ph    %[prim2],   %[p6],      %[p5]       \n\t"
         "precrq.qb.ph   %[prim3],   %[p4],      %[p3]       \n\t"
@@ -1097,118 +683,108 @@
         "append         %[p5],      %[sec3],    16          \n\t"
         "append         %[p3],      %[sec4],    16          \n\t"
 
-        : [prim1] "=&r" (prim1), [prim2] "=&r" (prim2),
-          [prim3] "=&r" (prim3), [prim4] "=&r" (prim4),
-          [p6] "+r" (p6), [p5] "+r" (p5), [p4] "+r" (p4), [p3] "+r" (p3),
-          [sec3] "=&r" (sec3), [sec4] "=&r" (sec4)
-        :
-    );
+        : [prim1] "=&r"(prim1), [prim2] "=&r"(prim2), [prim3] "=&r"(prim3),
+          [prim4] "=&r"(prim4), [p6] "+r"(p6), [p5] "+r"(p5), [p4] "+r"(p4),
+          [p3] "+r"(p3), [sec3] "=&r"(sec3), [sec4] "=&r"(sec4)
+        :);
 
     /* if (p1 - p4 == 0) and (p2 - p3 == 0)
-    * mask will be zero and filtering is not needed
-    */
-    if (!(((p1 - p4) == 0) && ((p2 - p3) == 0)))
-    {
+     * mask will be zero and filtering is not needed
+     */
+    if (!(((p1 - p4) == 0) && ((p2 - p3) == 0))) {
+      vp8_filter_mask_vec_mips(limit, flimit, p1, p2, pm1, p0, p3, p4, p5, p6,
+                               thresh, &hev, &mask);
 
-        vp8_filter_mask_vec_mips(limit, flimit, p1, p2, pm1, p0, p3, p4, p5, p6,
-                                 thresh, &hev, &mask);
+      /* if mask == 0 do filtering is not needed */
+      if (mask) {
+        /* filtering */
+        vp8_filter_mips(mask, hev, &p1, &p2, &p3, &p4);
 
-        /* if mask == 0 do filtering is not needed */
-        if (mask)
-        {
-            /* filtering */
-            vp8_filter_mips(mask, hev, &p1, &p2, &p3, &p4);
+        /* unpack processed 4x4 neighborhood
+         * don't use transpose on output data
+         * because memory isn't aligned
+         */
+        __asm__ __volatile__(
+            "sb         %[p4],  1(%[s4])    \n\t"
+            "sb         %[p3],  0(%[s4])    \n\t"
+            "sb         %[p2], -1(%[s4])    \n\t"
+            "sb         %[p1], -2(%[s4])    \n\t"
+            :
+            : [p4] "r"(p4), [p3] "r"(p3), [s4] "r"(s4), [p2] "r"(p2),
+              [p1] "r"(p1));
 
-            /* unpack processed 4x4 neighborhood
-             * don't use transpose on output data
-             * because memory isn't aligned
-             */
-            __asm__ __volatile__ (
-                "sb         %[p4],  1(%[s4])    \n\t"
-                "sb         %[p3],  0(%[s4])    \n\t"
-                "sb         %[p2], -1(%[s4])    \n\t"
-                "sb         %[p1], -2(%[s4])    \n\t"
-                :
-                : [p4] "r" (p4), [p3] "r" (p3), [s4] "r" (s4),
-                  [p2] "r" (p2), [p1] "r" (p1)
-            );
+        __asm__ __volatile__(
+            "srl        %[p4], %[p4], 8     \n\t"
+            "srl        %[p3], %[p3], 8     \n\t"
+            "srl        %[p2], %[p2], 8     \n\t"
+            "srl        %[p1], %[p1], 8     \n\t"
+            : [p4] "+r"(p4), [p3] "+r"(p3), [p2] "+r"(p2), [p1] "+r"(p1)
+            :);
 
-            __asm__ __volatile__ (
-                "srl        %[p4], %[p4], 8     \n\t"
-                "srl        %[p3], %[p3], 8     \n\t"
-                "srl        %[p2], %[p2], 8     \n\t"
-                "srl        %[p1], %[p1], 8     \n\t"
-                : [p4] "+r" (p4), [p3] "+r" (p3), [p2] "+r" (p2), [p1] "+r" (p1)
-                :
-            );
+        __asm__ __volatile__(
+            "sb         %[p4],  1(%[s3])    \n\t"
+            "sb         %[p3],  0(%[s3])    \n\t"
+            "sb         %[p2], -1(%[s3])    \n\t"
+            "sb         %[p1], -2(%[s3])    \n\t"
+            : [p1] "+r"(p1)
+            : [p4] "r"(p4), [p3] "r"(p3), [s3] "r"(s3), [p2] "r"(p2));
 
-            __asm__ __volatile__ (
-                "sb         %[p4],  1(%[s3])    \n\t"
-                "sb         %[p3],  0(%[s3])    \n\t"
-                "sb         %[p2], -1(%[s3])    \n\t"
-                "sb         %[p1], -2(%[s3])    \n\t"
-                : [p1] "+r" (p1)
-                : [p4] "r" (p4), [p3] "r" (p3), [s3] "r" (s3), [p2] "r" (p2)
-            );
+        __asm__ __volatile__(
+            "srl        %[p4], %[p4], 8     \n\t"
+            "srl        %[p3], %[p3], 8     \n\t"
+            "srl        %[p2], %[p2], 8     \n\t"
+            "srl        %[p1], %[p1], 8     \n\t"
+            : [p4] "+r"(p4), [p3] "+r"(p3), [p2] "+r"(p2), [p1] "+r"(p1)
+            :);
 
-            __asm__ __volatile__ (
-                "srl        %[p4], %[p4], 8     \n\t"
-                "srl        %[p3], %[p3], 8     \n\t"
-                "srl        %[p2], %[p2], 8     \n\t"
-                "srl        %[p1], %[p1], 8     \n\t"
-                : [p4] "+r" (p4), [p3] "+r" (p3), [p2] "+r" (p2), [p1] "+r" (p1)
-                :
-            );
+        __asm__ __volatile__(
+            "sb         %[p4],  1(%[s2])    \n\t"
+            "sb         %[p3],  0(%[s2])    \n\t"
+            "sb         %[p2], -1(%[s2])    \n\t"
+            "sb         %[p1], -2(%[s2])    \n\t"
+            :
+            : [p4] "r"(p4), [p3] "r"(p3), [s2] "r"(s2), [p2] "r"(p2),
+              [p1] "r"(p1));
 
-            __asm__ __volatile__ (
-                "sb         %[p4],  1(%[s2])    \n\t"
-                "sb         %[p3],  0(%[s2])    \n\t"
-                "sb         %[p2], -1(%[s2])    \n\t"
-                "sb         %[p1], -2(%[s2])    \n\t"
-                :
-                : [p4] "r" (p4), [p3] "r" (p3), [s2] "r" (s2),
-                  [p2] "r" (p2), [p1] "r" (p1)
-            );
+        __asm__ __volatile__(
+            "srl        %[p4], %[p4], 8     \n\t"
+            "srl        %[p3], %[p3], 8     \n\t"
+            "srl        %[p2], %[p2], 8     \n\t"
+            "srl        %[p1], %[p1], 8     \n\t"
+            : [p4] "+r"(p4), [p3] "+r"(p3), [p2] "+r"(p2), [p1] "+r"(p1)
+            :);
 
-            __asm__ __volatile__ (
-                "srl        %[p4], %[p4], 8     \n\t"
-                "srl        %[p3], %[p3], 8     \n\t"
-                "srl        %[p2], %[p2], 8     \n\t"
-                "srl        %[p1], %[p1], 8     \n\t"
-                : [p4] "+r" (p4), [p3] "+r" (p3), [p2] "+r" (p2), [p1] "+r" (p1)
-                :
-            );
-
-            __asm__ __volatile__ (
-                "sb         %[p4],  1(%[s1])    \n\t"
-                "sb         %[p3],  0(%[s1])    \n\t"
-                "sb         %[p2], -1(%[s1])    \n\t"
-                "sb         %[p1], -2(%[s1])    \n\t"
-                :
-                : [p4] "r" (p4), [p3] "r" (p3), [s1] "r" (s1), [p2] "r" (p2), [p1] "r" (p1)
-            );
-        }
+        __asm__ __volatile__(
+            "sb         %[p4],  1(%[s1])    \n\t"
+            "sb         %[p3],  0(%[s1])    \n\t"
+            "sb         %[p2], -1(%[s1])    \n\t"
+            "sb         %[p1], -2(%[s1])    \n\t"
+            :
+            : [p4] "r"(p4), [p3] "r"(p3), [s1] "r"(s1), [p2] "r"(p2),
+              [p1] "r"(p1));
+      }
     }
 
-    s1 = s4 + p;
-    s2 = s1 + p;
+    s1 = s;
+    s2 = s + p;
     s3 = s2 + p;
     s4 = s3 + p;
+    s = s4 + p;
 
     /* load quad-byte vectors
      * memory is 4 byte aligned
      */
-    p2  = *((uint32_t *)(s1 - 4));
-    p6  = *((uint32_t *)(s1));
-    p1  = *((uint32_t *)(s2 - 4));
-    p5  = *((uint32_t *)(s2));
-    p0  = *((uint32_t *)(s3 - 4));
-    p4  = *((uint32_t *)(s3));
+    p2 = *((uint32_t *)(s1 - 4));
+    p6 = *((uint32_t *)(s1));
+    p1 = *((uint32_t *)(s2 - 4));
+    p5 = *((uint32_t *)(s2));
+    p0 = *((uint32_t *)(s3 - 4));
+    p4 = *((uint32_t *)(s3));
     pm1 = *((uint32_t *)(s4 - 4));
-    p3  = *((uint32_t *)(s4));
+    p3 = *((uint32_t *)(s4));
 
     /* transpose pm1, p0, p1, p2 */
-    __asm__ __volatile__ (
+    __asm__ __volatile__(
         "precrq.qb.ph   %[prim1],   %[p2],      %[p1]       \n\t"
         "precr.qb.ph    %[prim2],   %[p2],      %[p1]       \n\t"
         "precrq.qb.ph   %[prim3],   %[p0],      %[pm1]      \n\t"
@@ -1224,15 +800,13 @@
         "append         %[p1],      %[sec3],    16          \n\t"
         "append         %[pm1],     %[sec4],    16          \n\t"
 
-        : [prim1] "=&r" (prim1), [prim2] "=&r" (prim2),
-          [prim3] "=&r" (prim3), [prim4] "=&r" (prim4),
-          [p2] "+r" (p2), [p1] "+r" (p1), [p0] "+r" (p0), [pm1] "+r" (pm1),
-          [sec3] "=&r" (sec3), [sec4] "=&r" (sec4)
-        :
-    );
+        : [prim1] "=&r"(prim1), [prim2] "=&r"(prim2), [prim3] "=&r"(prim3),
+          [prim4] "=&r"(prim4), [p2] "+r"(p2), [p1] "+r"(p1), [p0] "+r"(p0),
+          [pm1] "+r"(pm1), [sec3] "=&r"(sec3), [sec4] "=&r"(sec4)
+        :);
 
     /* transpose p3, p4, p5, p6 */
-    __asm__ __volatile__ (
+    __asm__ __volatile__(
         "precrq.qb.ph   %[prim1],   %[p6],      %[p5]       \n\t"
         "precr.qb.ph    %[prim2],   %[p6],      %[p5]       \n\t"
         "precrq.qb.ph   %[prim3],   %[p4],      %[p3]       \n\t"
@@ -1248,566 +822,691 @@
         "append         %[p5],      %[sec3],    16          \n\t"
         "append         %[p3],      %[sec4],    16          \n\t"
 
-        : [prim1] "=&r" (prim1), [prim2] "=&r" (prim2),
-          [prim3] "=&r" (prim3), [prim4] "=&r" (prim4),
-          [p6] "+r" (p6), [p5] "+r" (p5), [p4] "+r" (p4), [p3] "+r" (p3),
-          [sec3] "=&r" (sec3), [sec4] "=&r" (sec4)
-        :
-    );
+        : [prim1] "=&r"(prim1), [prim2] "=&r"(prim2), [prim3] "=&r"(prim3),
+          [prim4] "=&r"(prim4), [p6] "+r"(p6), [p5] "+r"(p5), [p4] "+r"(p4),
+          [p3] "+r"(p3), [sec3] "=&r"(sec3), [sec4] "=&r"(sec4)
+        :);
 
     /* if (p1 - p4 == 0) and (p2 - p3 == 0)
      * mask will be zero and filtering is not needed
      */
-    if (!(((p1 - p4) == 0) && ((p2 - p3) == 0)))
-    {
+    if (!(((p1 - p4) == 0) && ((p2 - p3) == 0))) {
+      vp8_filter_mask_vec_mips(limit, flimit, p1, p2, pm1, p0, p3, p4, p5, p6,
+                               thresh, &hev, &mask);
 
-        vp8_filter_mask_vec_mips(limit, flimit, p1, p2, pm1, p0, p3, p4, p5, p6,
-                                 thresh, &hev, &mask);
+      /* if mask == 0 do filtering is not needed */
+      if (mask) {
+        /* filtering */
+        vp8_filter_mips(mask, hev, &p1, &p2, &p3, &p4);
 
-        /* if mask == 0 do filtering is not needed */
-        if (mask)
-        {
-            /* filtering */
-            vp8_filter_mips(mask, hev, &p1, &p2, &p3, &p4);
+        /* unpack processed 4x4 neighborhood
+         * don't use transpose on output data
+         * because memory isn't aligned
+         */
+        __asm__ __volatile__(
+            "sb         %[p4],  1(%[s4])    \n\t"
+            "sb         %[p3],  0(%[s4])    \n\t"
+            "sb         %[p2], -1(%[s4])    \n\t"
+            "sb         %[p1], -2(%[s4])    \n\t"
+            :
+            : [p4] "r"(p4), [p3] "r"(p3), [s4] "r"(s4), [p2] "r"(p2),
+              [p1] "r"(p1));
 
-            /* unpack processed 4x4 neighborhood
-             * don't use transpose on output data
-             * because memory isn't aligned
-             */
-            __asm__ __volatile__ (
-                "sb         %[p4],  1(%[s4])    \n\t"
-                "sb         %[p3],  0(%[s4])    \n\t"
-                "sb         %[p2], -1(%[s4])    \n\t"
-                "sb         %[p1], -2(%[s4])    \n\t"
-                :
-                : [p4] "r" (p4), [p3] "r" (p3), [s4] "r" (s4),
-                  [p2] "r" (p2), [p1] "r" (p1)
-            );
+        __asm__ __volatile__(
+            "srl        %[p4], %[p4], 8     \n\t"
+            "srl        %[p3], %[p3], 8     \n\t"
+            "srl        %[p2], %[p2], 8     \n\t"
+            "srl        %[p1], %[p1], 8     \n\t"
+            : [p4] "+r"(p4), [p3] "+r"(p3), [p2] "+r"(p2), [p1] "+r"(p1)
+            :);
 
-            __asm__ __volatile__ (
-                "srl        %[p4], %[p4], 8     \n\t"
-                "srl        %[p3], %[p3], 8     \n\t"
-                "srl        %[p2], %[p2], 8     \n\t"
-                "srl        %[p1], %[p1], 8     \n\t"
-                : [p4] "+r" (p4), [p3] "+r" (p3), [p2] "+r" (p2), [p1] "+r" (p1)
-                :
-            );
+        __asm__ __volatile__(
+            "sb         %[p4],  1(%[s3])    \n\t"
+            "sb         %[p3],  0(%[s3])    \n\t"
+            "sb         %[p2], -1(%[s3])    \n\t"
+            "sb         %[p1], -2(%[s3])    \n\t"
+            : [p1] "+r"(p1)
+            : [p4] "r"(p4), [p3] "r"(p3), [s3] "r"(s3), [p2] "r"(p2));
 
-            __asm__ __volatile__ (
-                "sb         %[p4],  1(%[s3])    \n\t"
-                "sb         %[p3],  0(%[s3])    \n\t"
-                "sb         %[p2], -1(%[s3])    \n\t"
-                "sb         %[p1], -2(%[s3])    \n\t"
-                : [p1] "+r" (p1)
-                : [p4] "r" (p4), [p3] "r" (p3), [s3] "r" (s3), [p2] "r" (p2)
-            );
+        __asm__ __volatile__(
+            "srl        %[p4], %[p4], 8     \n\t"
+            "srl        %[p3], %[p3], 8     \n\t"
+            "srl        %[p2], %[p2], 8     \n\t"
+            "srl        %[p1], %[p1], 8     \n\t"
+            : [p4] "+r"(p4), [p3] "+r"(p3), [p2] "+r"(p2), [p1] "+r"(p1)
+            :);
 
-            __asm__ __volatile__ (
-                "srl        %[p4], %[p4], 8     \n\t"
-                "srl        %[p3], %[p3], 8     \n\t"
-                "srl        %[p2], %[p2], 8     \n\t"
-                "srl        %[p1], %[p1], 8     \n\t"
-                : [p4] "+r" (p4), [p3] "+r" (p3), [p2] "+r" (p2), [p1] "+r" (p1)
-                :
-            );
+        __asm__ __volatile__(
+            "sb         %[p4],  1(%[s2])    \n\t"
+            "sb         %[p3],  0(%[s2])    \n\t"
+            "sb         %[p2], -1(%[s2])    \n\t"
+            "sb         %[p1], -2(%[s2])    \n\t"
+            :
+            : [p4] "r"(p4), [p3] "r"(p3), [s2] "r"(s2), [p2] "r"(p2),
+              [p1] "r"(p1));
 
-            __asm__ __volatile__ (
-                "sb         %[p4],  1(%[s2])    \n\t"
-                "sb         %[p3],  0(%[s2])    \n\t"
-                "sb         %[p2], -1(%[s2])    \n\t"
-                "sb         %[p1], -2(%[s2])    \n\t"
-                :
-                : [p4] "r" (p4), [p3] "r" (p3), [s2] "r" (s2),
-                  [p2] "r" (p2), [p1] "r" (p1)
-            );
+        __asm__ __volatile__(
+            "srl        %[p4], %[p4], 8     \n\t"
+            "srl        %[p3], %[p3], 8     \n\t"
+            "srl        %[p2], %[p2], 8     \n\t"
+            "srl        %[p1], %[p1], 8     \n\t"
+            : [p4] "+r"(p4), [p3] "+r"(p3), [p2] "+r"(p2), [p1] "+r"(p1)
+            :);
 
-            __asm__ __volatile__ (
-                "srl        %[p4], %[p4], 8     \n\t"
-                "srl        %[p3], %[p3], 8     \n\t"
-                "srl        %[p2], %[p2], 8     \n\t"
-                "srl        %[p1], %[p1], 8     \n\t"
-                : [p4] "+r" (p4), [p3] "+r" (p3), [p2] "+r" (p2), [p1] "+r" (p1)
-                :
-            );
-
-            __asm__ __volatile__ (
-                "sb         %[p4],  1(%[s1])    \n\t"
-                "sb         %[p3],  0(%[s1])    \n\t"
-                "sb         %[p2], -1(%[s1])    \n\t"
-                "sb         %[p1], -2(%[s1])    \n\t"
-                :
-                : [p4] "r" (p4), [p3] "r" (p3), [s1] "r" (s1),
-                  [p2] "r" (p2), [p1] "r" (p1)
-            );
-        }
+        __asm__ __volatile__(
+            "sb         %[p4],  1(%[s1])    \n\t"
+            "sb         %[p3],  0(%[s1])    \n\t"
+            "sb         %[p2], -1(%[s1])    \n\t"
+            "sb         %[p1], -2(%[s1])    \n\t"
+            :
+            : [p4] "r"(p4), [p3] "r"(p3), [s1] "r"(s1), [p2] "r"(p2),
+              [p1] "r"(p1));
+      }
     }
+
+    i += 8;
+  }
+
+  while (i < count);
 }
 
-/* inputs & outputs are quad-byte vectors */
-static __inline void vp8_mbfilter_mips
-(
-    uint32_t mask,
-    uint32_t hev,
-    uint32_t *ps2,
-    uint32_t *ps1,
-    uint32_t *ps0,
-    uint32_t *qs0,
-    uint32_t *qs1,
-    uint32_t *qs2
-)
-{
-    int32_t vps2, vps1, vps0, vqs0, vqs1, vqs2;
-    int32_t vps2_l, vps1_l, vps0_l, vqs0_l, vqs1_l, vqs2_l;
-    int32_t vps2_r, vps1_r, vps0_r, vqs0_r, vqs1_r, vqs2_r;
-    uint32_t HWM, vp8_filter_l, vp8_filter_r, mask_l, mask_r, hev_l, hev_r, subr_r, subr_l;
-    uint32_t Filter2_l, Filter2_r, t1, t2, Filter1_l, Filter1_r, invhev_l, invhev_r;
-    uint32_t N128, R63;
-    uint32_t u1_l, u1_r, u2_l, u2_r, u3_l, u3_r;
+void vp8_loop_filter_uvvertical_edge_mips(unsigned char *s, int p,
+                                          unsigned int flimit,
+                                          unsigned int limit,
+                                          unsigned int thresh, int count) {
+  uint32_t mask, hev;
+  uint32_t pm1, p0, p1, p2, p3, p4, p5, p6;
+  unsigned char *s1, *s2, *s3, *s4;
+  uint32_t prim1, prim2, sec3, sec4, prim3, prim4;
 
-    R63  = 0x003F003F;
-    HWM  = 0xFF00FF00;
-    N128 = 0x80808080;
-    t1   = 0x03000300;
-    t2   = 0x04000400;
+  /* loop filter designed to work using chars so that we can make maximum use
+   * of 8 bit simd instructions.
+   */
 
-    vps0 = (*ps0) ^ N128;
-    vps1 = (*ps1) ^ N128;
-    vps2 = (*ps2) ^ N128;
-    vqs0 = (*qs0) ^ N128;
-    vqs1 = (*qs1) ^ N128;
-    vqs2 = (*qs2) ^ N128;
+  /* apply filter on 4 pixesl at the same time */
 
-    /* use halfword pairs instead quad-bytes because of accuracy */
-    vps0_l = vps0 & HWM;
-    vps0_r = vps0 << 8;
-    vps0_r = vps0_r & HWM;
+  s1 = s;
+  s2 = s + p;
+  s3 = s2 + p;
+  s4 = s3 + p;
 
-    vqs0_l = vqs0 & HWM;
-    vqs0_r = vqs0 << 8;
-    vqs0_r = vqs0_r & HWM;
+  /* load quad-byte vectors
+  * memory is 4 byte aligned
+  */
+  p2 = *((uint32_t *)(s1 - 4));
+  p6 = *((uint32_t *)(s1));
+  p1 = *((uint32_t *)(s2 - 4));
+  p5 = *((uint32_t *)(s2));
+  p0 = *((uint32_t *)(s3 - 4));
+  p4 = *((uint32_t *)(s3));
+  pm1 = *((uint32_t *)(s4 - 4));
+  p3 = *((uint32_t *)(s4));
 
-    vps1_l = vps1 & HWM;
-    vps1_r = vps1 << 8;
-    vps1_r = vps1_r & HWM;
+  /* transpose pm1, p0, p1, p2 */
+  __asm__ __volatile__(
+      "precrq.qb.ph   %[prim1],   %[p2],      %[p1]       \n\t"
+      "precr.qb.ph    %[prim2],   %[p2],      %[p1]       \n\t"
+      "precrq.qb.ph   %[prim3],   %[p0],      %[pm1]      \n\t"
+      "precr.qb.ph    %[prim4],   %[p0],      %[pm1]      \n\t"
 
-    vqs1_l = vqs1 & HWM;
-    vqs1_r = vqs1 << 8;
-    vqs1_r = vqs1_r & HWM;
+      "precrq.qb.ph   %[p1],      %[prim1],   %[prim2]    \n\t"
+      "precr.qb.ph    %[pm1],     %[prim1],   %[prim2]    \n\t"
+      "precrq.qb.ph   %[sec3],    %[prim3],   %[prim4]    \n\t"
+      "precr.qb.ph    %[sec4],    %[prim3],   %[prim4]    \n\t"
 
-    vqs2_l = vqs2 & HWM;
-    vqs2_r = vqs2 << 8;
-    vqs2_r = vqs2_r & HWM;
+      "precrq.ph.w    %[p2],      %[p1],      %[sec3]     \n\t"
+      "precrq.ph.w    %[p0],      %[pm1],     %[sec4]     \n\t"
+      "append         %[p1],      %[sec3],    16          \n\t"
+      "append         %[pm1],     %[sec4],    16          \n\t"
 
-    __asm__ __volatile__ (
-        /* qs0 - ps0 */
-        "subq_s.ph    %[subr_l],       %[vqs0_l],       %[vps0_l]       \n\t"
-        "subq_s.ph    %[subr_r],       %[vqs0_r],       %[vps0_r]       \n\t"
+      : [prim1] "=&r"(prim1), [prim2] "=&r"(prim2), [prim3] "=&r"(prim3),
+        [prim4] "=&r"(prim4), [p2] "+r"(p2), [p1] "+r"(p1), [p0] "+r"(p0),
+        [pm1] "+r"(pm1), [sec3] "=&r"(sec3), [sec4] "=&r"(sec4)
+      :);
 
-        /* vp8_filter = vp8_signed_char_clamp(ps1 - qs1); */
-        "subq_s.ph    %[vp8_filter_l], %[vps1_l],       %[vqs1_l]       \n\t"
-        "subq_s.ph    %[vp8_filter_r], %[vps1_r],       %[vqs1_r]       \n\t"
+  /* transpose p3, p4, p5, p6 */
+  __asm__ __volatile__(
+      "precrq.qb.ph   %[prim1],   %[p6],      %[p5]       \n\t"
+      "precr.qb.ph    %[prim2],   %[p6],      %[p5]       \n\t"
+      "precrq.qb.ph   %[prim3],   %[p4],      %[p3]       \n\t"
+      "precr.qb.ph    %[prim4],   %[p4],      %[p3]       \n\t"
 
-        : [vp8_filter_l] "=&r" (vp8_filter_l), [vp8_filter_r] "=r" (vp8_filter_r),
-          [subr_l] "=&r" (subr_l), [subr_r] "=&r" (subr_r)
-        : [vps0_l] "r" (vps0_l), [vps0_r] "r" (vps0_r), [vps1_l] "r" (vps1_l),
-          [vps1_r] "r" (vps1_r), [vqs0_l] "r" (vqs0_l), [vqs0_r] "r" (vqs0_r),
-          [vqs1_l] "r" (vqs1_l), [vqs1_r] "r" (vqs1_r)
-    );
+      "precrq.qb.ph   %[p5],      %[prim1],   %[prim2]    \n\t"
+      "precr.qb.ph    %[p3],      %[prim1],   %[prim2]    \n\t"
+      "precrq.qb.ph   %[sec3],    %[prim3],   %[prim4]    \n\t"
+      "precr.qb.ph    %[sec4],    %[prim3],   %[prim4]    \n\t"
 
-    vps2_l = vps2 & HWM;
-    vps2_r = vps2 << 8;
-    vps2_r = vps2_r & HWM;
+      "precrq.ph.w    %[p6],      %[p5],      %[sec3]     \n\t"
+      "precrq.ph.w    %[p4],      %[p3],      %[sec4]     \n\t"
+      "append         %[p5],      %[sec3],    16          \n\t"
+      "append         %[p3],      %[sec4],    16          \n\t"
 
-    /* add outer taps if we have high edge variance */
-    __asm__ __volatile__ (
-        /* vp8_filter = vp8_signed_char_clamp(vp8_filter + 3 * (qs0 - ps0)); */
-        "addq_s.ph    %[vp8_filter_l], %[vp8_filter_l], %[subr_l]       \n\t"
-        "addq_s.ph    %[vp8_filter_r], %[vp8_filter_r], %[subr_r]       \n\t"
-        "and          %[mask_l],       %[HWM],          %[mask]         \n\t"
-        "sll          %[mask_r],       %[mask],         8               \n\t"
-        "and          %[mask_r],       %[HWM],          %[mask_r]       \n\t"
-        "addq_s.ph    %[vp8_filter_l], %[vp8_filter_l], %[subr_l]       \n\t"
-        "addq_s.ph    %[vp8_filter_r], %[vp8_filter_r], %[subr_r]       \n\t"
-        "and          %[hev_l],        %[HWM],          %[hev]          \n\t"
-        "sll          %[hev_r],        %[hev],          8               \n\t"
-        "and          %[hev_r],        %[HWM],          %[hev_r]        \n\t"
-        "addq_s.ph    %[vp8_filter_l], %[vp8_filter_l], %[subr_l]       \n\t"
-        "addq_s.ph    %[vp8_filter_r], %[vp8_filter_r], %[subr_r]       \n\t"
+      : [prim1] "=&r"(prim1), [prim2] "=&r"(prim2), [prim3] "=&r"(prim3),
+        [prim4] "=&r"(prim4), [p6] "+r"(p6), [p5] "+r"(p5), [p4] "+r"(p4),
+        [p3] "+r"(p3), [sec3] "=&r"(sec3), [sec4] "=&r"(sec4)
+      :);
 
-        /* vp8_filter &= mask; */
-        "and          %[vp8_filter_l], %[vp8_filter_l], %[mask_l]       \n\t"
-        "and          %[vp8_filter_r], %[vp8_filter_r], %[mask_r]       \n\t"
+  /* if (p1 - p4 == 0) and (p2 - p3 == 0)
+  * mask will be zero and filtering is not needed
+  */
+  if (!(((p1 - p4) == 0) && ((p2 - p3) == 0))) {
+    vp8_filter_mask_vec_mips(limit, flimit, p1, p2, pm1, p0, p3, p4, p5, p6,
+                             thresh, &hev, &mask);
 
-        /* Filter2 = vp8_filter & hev; */
-        "and          %[Filter2_l],    %[vp8_filter_l], %[hev_l]        \n\t"
-        "and          %[Filter2_r],    %[vp8_filter_r], %[hev_r]        \n\t"
+    /* if mask == 0 do filtering is not needed */
+    if (mask) {
+      /* filtering */
+      vp8_filter_mips(mask, hev, &p1, &p2, &p3, &p4);
 
-        : [vp8_filter_l] "+r" (vp8_filter_l), [vp8_filter_r] "+r" (vp8_filter_r),
-          [hev_l] "=&r" (hev_l), [hev_r] "=&r" (hev_r),
-          [mask_l] "=&r" (mask_l), [mask_r] "=&r" (mask_r),
-          [Filter2_l] "=&r" (Filter2_l), [Filter2_r] "=&r" (Filter2_r)
-        : [subr_l] "r" (subr_l), [subr_r] "r" (subr_r),
-          [HWM] "r" (HWM), [hev]  "r" (hev), [mask] "r" (mask)
-    );
+      /* unpack processed 4x4 neighborhood
+       * don't use transpose on output data
+       * because memory isn't aligned
+       */
+      __asm__ __volatile__(
+          "sb         %[p4],  1(%[s4])    \n\t"
+          "sb         %[p3],  0(%[s4])    \n\t"
+          "sb         %[p2], -1(%[s4])    \n\t"
+          "sb         %[p1], -2(%[s4])    \n\t"
+          :
+          :
+          [p4] "r"(p4), [p3] "r"(p3), [s4] "r"(s4), [p2] "r"(p2), [p1] "r"(p1));
 
-    /* save bottom 3 bits so that we round one side +4 and the other +3 */
-    __asm__ __volatile__ (
-        /* Filter1 = vp8_signed_char_clamp(Filter2 + 4) >>= 3; */
-        "addq_s.ph    %[Filter1_l],    %[Filter2_l],    %[t2]           \n\t"
-        "xor          %[invhev_l],     %[hev_l],        %[HWM]          \n\t"
-        "addq_s.ph    %[Filter1_r],    %[Filter2_r],    %[t2]           \n\t"
+      __asm__ __volatile__(
+          "srl        %[p4], %[p4], 8     \n\t"
+          "srl        %[p3], %[p3], 8     \n\t"
+          "srl        %[p2], %[p2], 8     \n\t"
+          "srl        %[p1], %[p1], 8     \n\t"
+          : [p4] "+r"(p4), [p3] "+r"(p3), [p2] "+r"(p2), [p1] "+r"(p1)
+          :);
 
-        /* Filter2 = vp8_signed_char_clamp(Filter2 + 3) >>= 3; */
-        "addq_s.ph    %[Filter2_l],    %[Filter2_l],    %[t1]           \n\t"
-        "addq_s.ph    %[Filter2_r],    %[Filter2_r],    %[t1]           \n\t"
+      __asm__ __volatile__(
+          "sb         %[p4],  1(%[s3])    \n\t"
+          "sb         %[p3],  0(%[s3])    \n\t"
+          "sb         %[p2], -1(%[s3])    \n\t"
+          "sb         %[p1], -2(%[s3])    \n\t"
+          : [p1] "+r"(p1)
+          : [p4] "r"(p4), [p3] "r"(p3), [s3] "r"(s3), [p2] "r"(p2));
 
-        "shra.ph      %[Filter1_l],    %[Filter1_l],    3               \n\t"
-        "shra.ph      %[Filter1_r],    %[Filter1_r],    3               \n\t"
+      __asm__ __volatile__(
+          "srl        %[p4], %[p4], 8     \n\t"
+          "srl        %[p3], %[p3], 8     \n\t"
+          "srl        %[p2], %[p2], 8     \n\t"
+          "srl        %[p1], %[p1], 8     \n\t"
+          : [p4] "+r"(p4), [p3] "+r"(p3), [p2] "+r"(p2), [p1] "+r"(p1)
+          :);
 
-        "shra.ph      %[Filter2_l],    %[Filter2_l],    3               \n\t"
-        "shra.ph      %[Filter2_r],    %[Filter2_r],    3               \n\t"
-        "and          %[Filter1_l],    %[Filter1_l],    %[HWM]          \n\t"
-        "and          %[Filter1_r],    %[Filter1_r],    %[HWM]          \n\t"
-        "xor          %[invhev_r],     %[hev_r],        %[HWM]          \n\t"
+      __asm__ __volatile__(
+          "sb         %[p4],  1(%[s2])    \n\t"
+          "sb         %[p3],  0(%[s2])    \n\t"
+          "sb         %[p2], -1(%[s2])    \n\t"
+          "sb         %[p1], -2(%[s2])    \n\t"
+          :
+          :
+          [p4] "r"(p4), [p3] "r"(p3), [s2] "r"(s2), [p2] "r"(p2), [p1] "r"(p1));
 
-        /* qs0 = vp8_signed_char_clamp(qs0 - Filter1); */
-        "subq_s.ph    %[vqs0_l],       %[vqs0_l],       %[Filter1_l]    \n\t"
-        "subq_s.ph    %[vqs0_r],       %[vqs0_r],       %[Filter1_r]    \n\t"
+      __asm__ __volatile__(
+          "srl        %[p4], %[p4], 8     \n\t"
+          "srl        %[p3], %[p3], 8     \n\t"
+          "srl        %[p2], %[p2], 8     \n\t"
+          "srl        %[p1], %[p1], 8     \n\t"
+          : [p4] "+r"(p4), [p3] "+r"(p3), [p2] "+r"(p2), [p1] "+r"(p1)
+          :);
 
-        /* ps0 = vp8_signed_char_clamp(ps0 + Filter2); */
-        "addq_s.ph    %[vps0_l],       %[vps0_l],       %[Filter2_l]    \n\t"
-        "addq_s.ph    %[vps0_r],       %[vps0_r],       %[Filter2_r]    \n\t"
+      __asm__ __volatile__(
+          "sb         %[p4],  1(%[s1])    \n\t"
+          "sb         %[p3],  0(%[s1])    \n\t"
+          "sb         %[p2], -1(%[s1])    \n\t"
+          "sb         %[p1], -2(%[s1])    \n\t"
+          :
+          :
+          [p4] "r"(p4), [p3] "r"(p3), [s1] "r"(s1), [p2] "r"(p2), [p1] "r"(p1));
+    }
+  }
 
-        : [invhev_l] "=&r" (invhev_l), [invhev_r] "=&r" (invhev_r),
-          [Filter1_l] "=&r" (Filter1_l), [Filter1_r] "=&r" (Filter1_r),
-          [Filter2_l] "+r" (Filter2_l), [Filter2_r] "+r" (Filter2_r),
-          [vps0_l] "+r" (vps0_l), [vps0_r] "+r" (vps0_r),
-          [vqs0_l] "+r" (vqs0_l), [vqs0_r] "+r" (vqs0_r)
-        : [t1] "r" (t1), [t2] "r" (t2), [HWM] "r" (HWM),
-          [hev_l] "r" (hev_l), [hev_r] "r" (hev_r)
-    );
+  s1 = s4 + p;
+  s2 = s1 + p;
+  s3 = s2 + p;
+  s4 = s3 + p;
 
-    /* only apply wider filter if not high edge variance */
-    __asm__ __volatile__ (
-        /* vp8_filter &= ~hev; */
-        "and          %[Filter2_l],    %[vp8_filter_l], %[invhev_l]     \n\t"
-        "and          %[Filter2_r],    %[vp8_filter_r], %[invhev_r]     \n\t"
+  /* load quad-byte vectors
+   * memory is 4 byte aligned
+   */
+  p2 = *((uint32_t *)(s1 - 4));
+  p6 = *((uint32_t *)(s1));
+  p1 = *((uint32_t *)(s2 - 4));
+  p5 = *((uint32_t *)(s2));
+  p0 = *((uint32_t *)(s3 - 4));
+  p4 = *((uint32_t *)(s3));
+  pm1 = *((uint32_t *)(s4 - 4));
+  p3 = *((uint32_t *)(s4));
 
-        "shra.ph      %[Filter2_l],    %[Filter2_l],    8               \n\t"
-        "shra.ph      %[Filter2_r],    %[Filter2_r],    8               \n\t"
+  /* transpose pm1, p0, p1, p2 */
+  __asm__ __volatile__(
+      "precrq.qb.ph   %[prim1],   %[p2],      %[p1]       \n\t"
+      "precr.qb.ph    %[prim2],   %[p2],      %[p1]       \n\t"
+      "precrq.qb.ph   %[prim3],   %[p0],      %[pm1]      \n\t"
+      "precr.qb.ph    %[prim4],   %[p0],      %[pm1]      \n\t"
 
-        : [Filter2_l] "=&r" (Filter2_l), [Filter2_r] "=&r" (Filter2_r)
-        : [vp8_filter_l] "r" (vp8_filter_l), [vp8_filter_r] "r" (vp8_filter_r),
-          [invhev_l] "r" (invhev_l), [invhev_r] "r" (invhev_r)
-    );
+      "precrq.qb.ph   %[p1],      %[prim1],   %[prim2]    \n\t"
+      "precr.qb.ph    %[pm1],     %[prim1],   %[prim2]    \n\t"
+      "precrq.qb.ph   %[sec3],    %[prim3],   %[prim4]    \n\t"
+      "precr.qb.ph    %[sec4],    %[prim3],   %[prim4]    \n\t"
 
-    /* roughly 3/7th difference across boundary */
-    __asm__ __volatile__ (
-        "shll.ph      %[u3_l],         %[Filter2_l],    3               \n\t"
-        "shll.ph      %[u3_r],         %[Filter2_r],    3               \n\t"
+      "precrq.ph.w    %[p2],      %[p1],      %[sec3]     \n\t"
+      "precrq.ph.w    %[p0],      %[pm1],     %[sec4]     \n\t"
+      "append         %[p1],      %[sec3],    16          \n\t"
+      "append         %[pm1],     %[sec4],    16          \n\t"
 
-        "addq.ph      %[u3_l],         %[u3_l],         %[Filter2_l]    \n\t"
-        "addq.ph      %[u3_r],         %[u3_r],         %[Filter2_r]    \n\t"
+      : [prim1] "=&r"(prim1), [prim2] "=&r"(prim2), [prim3] "=&r"(prim3),
+        [prim4] "=&r"(prim4), [p2] "+r"(p2), [p1] "+r"(p1), [p0] "+r"(p0),
+        [pm1] "+r"(pm1), [sec3] "=&r"(sec3), [sec4] "=&r"(sec4)
+      :);
 
-        "shll.ph      %[u2_l],         %[u3_l],         1               \n\t"
-        "shll.ph      %[u2_r],         %[u3_r],         1               \n\t"
+  /* transpose p3, p4, p5, p6 */
+  __asm__ __volatile__(
+      "precrq.qb.ph   %[prim1],   %[p6],      %[p5]       \n\t"
+      "precr.qb.ph    %[prim2],   %[p6],      %[p5]       \n\t"
+      "precrq.qb.ph   %[prim3],   %[p4],      %[p3]       \n\t"
+      "precr.qb.ph    %[prim4],   %[p4],      %[p3]       \n\t"
 
-        "addq.ph      %[u1_l],         %[u3_l],         %[u2_l]         \n\t"
-        "addq.ph      %[u1_r],         %[u3_r],         %[u2_r]         \n\t"
+      "precrq.qb.ph   %[p5],      %[prim1],   %[prim2]    \n\t"
+      "precr.qb.ph    %[p3],      %[prim1],   %[prim2]    \n\t"
+      "precrq.qb.ph   %[sec3],    %[prim3],   %[prim4]    \n\t"
+      "precr.qb.ph    %[sec4],    %[prim3],   %[prim4]    \n\t"
 
-        "addq.ph      %[u2_l],         %[u2_l],         %[R63]          \n\t"
-        "addq.ph      %[u2_r],         %[u2_r],         %[R63]          \n\t"
+      "precrq.ph.w    %[p6],      %[p5],      %[sec3]     \n\t"
+      "precrq.ph.w    %[p4],      %[p3],      %[sec4]     \n\t"
+      "append         %[p5],      %[sec3],    16          \n\t"
+      "append         %[p3],      %[sec4],    16          \n\t"
 
-        "addq.ph      %[u3_l],         %[u3_l],         %[R63]          \n\t"
-        "addq.ph      %[u3_r],         %[u3_r],         %[R63]          \n\t"
+      : [prim1] "=&r"(prim1), [prim2] "=&r"(prim2), [prim3] "=&r"(prim3),
+        [prim4] "=&r"(prim4), [p6] "+r"(p6), [p5] "+r"(p5), [p4] "+r"(p4),
+        [p3] "+r"(p3), [sec3] "=&r"(sec3), [sec4] "=&r"(sec4)
+      :);
 
-        /* vp8_signed_char_clamp((63 + Filter2 * 27) >> 7)
-         * vp8_signed_char_clamp((63 + Filter2 * 18) >> 7)
-         */
-        "addq.ph      %[u1_l],         %[u1_l],         %[R63]          \n\t"
-        "addq.ph      %[u1_r],         %[u1_r],         %[R63]          \n\t"
-        "shra.ph      %[u1_l],         %[u1_l],         7               \n\t"
-        "shra.ph      %[u1_r],         %[u1_r],         7               \n\t"
-        "shra.ph      %[u2_l],         %[u2_l],         7               \n\t"
-        "shra.ph      %[u2_r],         %[u2_r],         7               \n\t"
-        "shll.ph      %[u1_l],         %[u1_l],         8               \n\t"
-        "shll.ph      %[u1_r],         %[u1_r],         8               \n\t"
-        "shll.ph      %[u2_l],         %[u2_l],         8               \n\t"
-        "shll.ph      %[u2_r],         %[u2_r],         8               \n\t"
+  /* if (p1 - p4 == 0) and (p2 - p3 == 0)
+   * mask will be zero and filtering is not needed
+   */
+  if (!(((p1 - p4) == 0) && ((p2 - p3) == 0))) {
+    vp8_filter_mask_vec_mips(limit, flimit, p1, p2, pm1, p0, p3, p4, p5, p6,
+                             thresh, &hev, &mask);
 
-        /* vqs0 = vp8_signed_char_clamp(qs0 - u); */
-        "subq_s.ph    %[vqs0_l],       %[vqs0_l],       %[u1_l]         \n\t"
-        "subq_s.ph    %[vqs0_r],       %[vqs0_r],       %[u1_r]         \n\t"
+    /* if mask == 0 do filtering is not needed */
+    if (mask) {
+      /* filtering */
+      vp8_filter_mips(mask, hev, &p1, &p2, &p3, &p4);
 
-        /* vps0 = vp8_signed_char_clamp(ps0 + u); */
-        "addq_s.ph    %[vps0_l],       %[vps0_l],       %[u1_l]         \n\t"
-        "addq_s.ph    %[vps0_r],       %[vps0_r],       %[u1_r]         \n\t"
+      /* unpack processed 4x4 neighborhood
+       * don't use transpose on output data
+       * because memory isn't aligned
+       */
+      __asm__ __volatile__(
+          "sb         %[p4],  1(%[s4])    \n\t"
+          "sb         %[p3],  0(%[s4])    \n\t"
+          "sb         %[p2], -1(%[s4])    \n\t"
+          "sb         %[p1], -2(%[s4])    \n\t"
+          :
+          :
+          [p4] "r"(p4), [p3] "r"(p3), [s4] "r"(s4), [p2] "r"(p2), [p1] "r"(p1));
 
-        : [u1_l] "=&r" (u1_l), [u1_r] "=&r" (u1_r), [u2_l] "=&r" (u2_l),
-          [u2_r] "=&r" (u2_r), [u3_l] "=&r" (u3_l), [u3_r] "=&r" (u3_r),
-          [vps0_l] "+r" (vps0_l), [vps0_r] "+r" (vps0_r),
-          [vqs0_l] "+r" (vqs0_l), [vqs0_r] "+r" (vqs0_r)
-        : [R63]  "r" (R63),
-          [Filter2_l] "r" (Filter2_l), [Filter2_r] "r" (Filter2_r)
-    );
+      __asm__ __volatile__(
+          "srl        %[p4], %[p4], 8     \n\t"
+          "srl        %[p3], %[p3], 8     \n\t"
+          "srl        %[p2], %[p2], 8     \n\t"
+          "srl        %[p1], %[p1], 8     \n\t"
+          : [p4] "+r"(p4), [p3] "+r"(p3), [p2] "+r"(p2), [p1] "+r"(p1)
+          :);
 
-    __asm__ __volatile__ (
-        /* vqs1 = vp8_signed_char_clamp(qs1 - u); */
-        "subq_s.ph    %[vqs1_l],       %[vqs1_l],       %[u2_l]         \n\t"
-        "addq_s.ph    %[vps1_l],       %[vps1_l],       %[u2_l]         \n\t"
+      __asm__ __volatile__(
+          "sb         %[p4],  1(%[s3])    \n\t"
+          "sb         %[p3],  0(%[s3])    \n\t"
+          "sb         %[p2], -1(%[s3])    \n\t"
+          "sb         %[p1], -2(%[s3])    \n\t"
+          : [p1] "+r"(p1)
+          : [p4] "r"(p4), [p3] "r"(p3), [s3] "r"(s3), [p2] "r"(p2));
 
-        /* vps1 = vp8_signed_char_clamp(ps1 + u); */
-        "addq_s.ph    %[vps1_r],       %[vps1_r],       %[u2_r]         \n\t"
-        "subq_s.ph    %[vqs1_r],       %[vqs1_r],       %[u2_r]         \n\t"
+      __asm__ __volatile__(
+          "srl        %[p4], %[p4], 8     \n\t"
+          "srl        %[p3], %[p3], 8     \n\t"
+          "srl        %[p2], %[p2], 8     \n\t"
+          "srl        %[p1], %[p1], 8     \n\t"
+          : [p4] "+r"(p4), [p3] "+r"(p3), [p2] "+r"(p2), [p1] "+r"(p1)
+          :);
 
-        : [vps1_l] "+r" (vps1_l), [vps1_r] "+r" (vps1_r),
-          [vqs1_l] "+r" (vqs1_l), [vqs1_r] "+r" (vqs1_r)
-        : [u2_l] "r" (u2_l), [u2_r] "r" (u2_r)
-    );
+      __asm__ __volatile__(
+          "sb         %[p4],  1(%[s2])    \n\t"
+          "sb         %[p3],  0(%[s2])    \n\t"
+          "sb         %[p2], -1(%[s2])    \n\t"
+          "sb         %[p1], -2(%[s2])    \n\t"
+          :
+          :
+          [p4] "r"(p4), [p3] "r"(p3), [s2] "r"(s2), [p2] "r"(p2), [p1] "r"(p1));
 
-    /* roughly 1/7th difference across boundary */
-    __asm__ __volatile__ (
-        /* u = vp8_signed_char_clamp((63 + Filter2 * 9) >> 7); */
-        "shra.ph      %[u3_l],         %[u3_l],         7               \n\t"
-        "shra.ph      %[u3_r],         %[u3_r],         7               \n\t"
-        "shll.ph      %[u3_l],         %[u3_l],         8               \n\t"
-        "shll.ph      %[u3_r],         %[u3_r],         8               \n\t"
+      __asm__ __volatile__(
+          "srl        %[p4], %[p4], 8     \n\t"
+          "srl        %[p3], %[p3], 8     \n\t"
+          "srl        %[p2], %[p2], 8     \n\t"
+          "srl        %[p1], %[p1], 8     \n\t"
+          : [p4] "+r"(p4), [p3] "+r"(p3), [p2] "+r"(p2), [p1] "+r"(p1)
+          :);
 
-        /* vqs2 = vp8_signed_char_clamp(qs2 - u); */
-        "subq_s.ph    %[vqs2_l],       %[vqs2_l],       %[u3_l]         \n\t"
-        "subq_s.ph    %[vqs2_r],       %[vqs2_r],       %[u3_r]         \n\t"
+      __asm__ __volatile__(
+          "sb         %[p4],  1(%[s1])    \n\t"
+          "sb         %[p3],  0(%[s1])    \n\t"
+          "sb         %[p2], -1(%[s1])    \n\t"
+          "sb         %[p1], -2(%[s1])    \n\t"
+          :
+          :
+          [p4] "r"(p4), [p3] "r"(p3), [s1] "r"(s1), [p2] "r"(p2), [p1] "r"(p1));
+    }
+  }
+}
 
-        /* vps2 = vp8_signed_char_clamp(ps2 + u); */
-        "addq_s.ph    %[vps2_l],       %[vps2_l],       %[u3_l]         \n\t"
-        "addq_s.ph    %[vps2_r],       %[vps2_r],       %[u3_r]         \n\t"
+/* inputs & outputs are quad-byte vectors */
+static __inline void vp8_mbfilter_mips(uint32_t mask, uint32_t hev,
+                                       uint32_t *ps2, uint32_t *ps1,
+                                       uint32_t *ps0, uint32_t *qs0,
+                                       uint32_t *qs1, uint32_t *qs2) {
+  int32_t vps2, vps1, vps0, vqs0, vqs1, vqs2;
+  int32_t vps2_l, vps1_l, vps0_l, vqs0_l, vqs1_l, vqs2_l;
+  int32_t vps2_r, vps1_r, vps0_r, vqs0_r, vqs1_r, vqs2_r;
+  uint32_t HWM, vp8_filter_l, vp8_filter_r, mask_l, mask_r, hev_l, hev_r,
+      subr_r, subr_l;
+  uint32_t Filter2_l, Filter2_r, t1, t2, Filter1_l, Filter1_r, invhev_l,
+      invhev_r;
+  uint32_t N128, R63;
+  uint32_t u1_l, u1_r, u2_l, u2_r, u3_l, u3_r;
 
-        : [u3_l] "+r" (u3_l), [u3_r] "+r" (u3_r), [vps2_l] "+r" (vps2_l),
-          [vps2_r] "+r" (vps2_r), [vqs2_l] "+r" (vqs2_l), [vqs2_r] "+r" (vqs2_r)
-        :
-    );
+  R63 = 0x003F003F;
+  HWM = 0xFF00FF00;
+  N128 = 0x80808080;
+  t1 = 0x03000300;
+  t2 = 0x04000400;
 
-    /* Create quad-bytes from halfword pairs */
-    __asm__ __volatile__ (
-        "and          %[vqs0_l],       %[vqs0_l],       %[HWM]          \n\t"
-        "shrl.ph      %[vqs0_r],       %[vqs0_r],       8               \n\t"
+  vps0 = (*ps0) ^ N128;
+  vps1 = (*ps1) ^ N128;
+  vps2 = (*ps2) ^ N128;
+  vqs0 = (*qs0) ^ N128;
+  vqs1 = (*qs1) ^ N128;
+  vqs2 = (*qs2) ^ N128;
 
-        "and          %[vps0_l],       %[vps0_l],       %[HWM]          \n\t"
-        "shrl.ph      %[vps0_r],       %[vps0_r],       8               \n\t"
+  /* use halfword pairs instead quad-bytes because of accuracy */
+  vps0_l = vps0 & HWM;
+  vps0_r = vps0 << 8;
+  vps0_r = vps0_r & HWM;
 
-        "and          %[vqs1_l],       %[vqs1_l],       %[HWM]          \n\t"
-        "shrl.ph      %[vqs1_r],       %[vqs1_r],       8               \n\t"
+  vqs0_l = vqs0 & HWM;
+  vqs0_r = vqs0 << 8;
+  vqs0_r = vqs0_r & HWM;
 
-        "and          %[vps1_l],       %[vps1_l],       %[HWM]          \n\t"
-        "shrl.ph      %[vps1_r],       %[vps1_r],       8               \n\t"
+  vps1_l = vps1 & HWM;
+  vps1_r = vps1 << 8;
+  vps1_r = vps1_r & HWM;
 
-        "and          %[vqs2_l],       %[vqs2_l],       %[HWM]          \n\t"
-        "shrl.ph      %[vqs2_r],       %[vqs2_r],       8               \n\t"
+  vqs1_l = vqs1 & HWM;
+  vqs1_r = vqs1 << 8;
+  vqs1_r = vqs1_r & HWM;
 
-        "and          %[vps2_l],       %[vps2_l],       %[HWM]          \n\t"
-        "shrl.ph      %[vps2_r],       %[vps2_r],       8               \n\t"
+  vqs2_l = vqs2 & HWM;
+  vqs2_r = vqs2 << 8;
+  vqs2_r = vqs2_r & HWM;
 
-        "or           %[vqs0_r],       %[vqs0_l],       %[vqs0_r]       \n\t"
-        "or           %[vps0_r],       %[vps0_l],       %[vps0_r]       \n\t"
-        "or           %[vqs1_r],       %[vqs1_l],       %[vqs1_r]       \n\t"
-        "or           %[vps1_r],       %[vps1_l],       %[vps1_r]       \n\t"
-        "or           %[vqs2_r],       %[vqs2_l],       %[vqs2_r]       \n\t"
-        "or           %[vps2_r],       %[vps2_l],       %[vps2_r]       \n\t"
+  __asm__ __volatile__(
+      /* qs0 - ps0 */
+      "subq_s.ph    %[subr_l],       %[vqs0_l],       %[vps0_l]       \n\t"
+      "subq_s.ph    %[subr_r],       %[vqs0_r],       %[vps0_r]       \n\t"
 
-        : [vps1_l] "+r" (vps1_l), [vps1_r] "+r" (vps1_r), [vqs1_l] "+r" (vqs1_l),
-          [vqs1_r] "+r" (vqs1_r), [vps0_l] "+r" (vps0_l), [vps0_r] "+r" (vps0_r),
-          [vqs0_l] "+r" (vqs0_l), [vqs0_r] "+r" (vqs0_r), [vqs2_l] "+r" (vqs2_l),
-          [vqs2_r] "+r" (vqs2_r), [vps2_r] "+r" (vps2_r), [vps2_l] "+r" (vps2_l)
-        : [HWM] "r" (HWM)
-    );
+      /* vp8_filter = vp8_signed_char_clamp(ps1 - qs1); */
+      "subq_s.ph    %[vp8_filter_l], %[vps1_l],       %[vqs1_l]       \n\t"
+      "subq_s.ph    %[vp8_filter_r], %[vps1_r],       %[vqs1_r]       \n\t"
 
-    *ps0 = vps0_r ^ N128;
-    *ps1 = vps1_r ^ N128;
-    *ps2 = vps2_r ^ N128;
-    *qs0 = vqs0_r ^ N128;
-    *qs1 = vqs1_r ^ N128;
-    *qs2 = vqs2_r ^ N128;
-}
+      : [vp8_filter_l] "=&r"(vp8_filter_l), [vp8_filter_r] "=r"(vp8_filter_r),
+        [subr_l] "=&r"(subr_l), [subr_r] "=&r"(subr_r)
+      : [vps0_l] "r"(vps0_l), [vps0_r] "r"(vps0_r), [vps1_l] "r"(vps1_l),
+        [vps1_r] "r"(vps1_r), [vqs0_l] "r"(vqs0_l), [vqs0_r] "r"(vqs0_r),
+        [vqs1_l] "r"(vqs1_l), [vqs1_r] "r"(vqs1_r));
 
-void vp8_mbloop_filter_horizontal_edge_mips
-(
-    unsigned char *s,
-    int p,
-    unsigned int flimit,
-    unsigned int limit,
-    unsigned int thresh,
-    int count
-)
-{
-    int i;
-    uint32_t mask, hev;
-    uint32_t pm1, p0, p1, p2, p3, p4, p5, p6;
-    unsigned char *sm1, *s0, *s1, *s2, *s3, *s4, *s5, *s6;
+  vps2_l = vps2 & HWM;
+  vps2_r = vps2 << 8;
+  vps2_r = vps2_r & HWM;
 
-    mask = 0;
-    hev = 0;
-    i = 0;
-    p1 = 0;
-    p2 = 0;
-    p3 = 0;
-    p4 = 0;
+  /* add outer taps if we have high edge variance */
+  __asm__ __volatile__(
+      /* vp8_filter = vp8_signed_char_clamp(vp8_filter + 3 * (qs0 - ps0)); */
+      "addq_s.ph    %[vp8_filter_l], %[vp8_filter_l], %[subr_l]       \n\t"
+      "addq_s.ph    %[vp8_filter_r], %[vp8_filter_r], %[subr_r]       \n\t"
+      "and          %[mask_l],       %[HWM],          %[mask]         \n\t"
+      "sll          %[mask_r],       %[mask],         8               \n\t"
+      "and          %[mask_r],       %[HWM],          %[mask_r]       \n\t"
+      "addq_s.ph    %[vp8_filter_l], %[vp8_filter_l], %[subr_l]       \n\t"
+      "addq_s.ph    %[vp8_filter_r], %[vp8_filter_r], %[subr_r]       \n\t"
+      "and          %[hev_l],        %[HWM],          %[hev]          \n\t"
+      "sll          %[hev_r],        %[hev],          8               \n\t"
+      "and          %[hev_r],        %[HWM],          %[hev_r]        \n\t"
+      "addq_s.ph    %[vp8_filter_l], %[vp8_filter_l], %[subr_l]       \n\t"
+      "addq_s.ph    %[vp8_filter_r], %[vp8_filter_r], %[subr_r]       \n\t"
 
-    /* loop filter designed to work using chars so that we can make maximum use
-     * of 8 bit simd instructions.
-     */
+      /* vp8_filter &= mask; */
+      "and          %[vp8_filter_l], %[vp8_filter_l], %[mask_l]       \n\t"
+      "and          %[vp8_filter_r], %[vp8_filter_r], %[mask_r]       \n\t"
 
-    sm1 = s - (p << 2);
-    s0  = s - p - p - p;
-    s1  = s - p - p;
-    s2  = s - p;
-    s3  = s;
-    s4  = s + p;
-    s5  = s + p + p;
-    s6  = s + p + p + p;
+      /* Filter2 = vp8_filter & hev; */
+      "and          %[Filter2_l],    %[vp8_filter_l], %[hev_l]        \n\t"
+      "and          %[Filter2_r],    %[vp8_filter_r], %[hev_r]        \n\t"
 
-    /* prefetch data for load */
-    prefetch_load_lf(s + p);
+      : [vp8_filter_l] "+r"(vp8_filter_l), [vp8_filter_r] "+r"(vp8_filter_r),
+        [hev_l] "=&r"(hev_l), [hev_r] "=&r"(hev_r), [mask_l] "=&r"(mask_l),
+        [mask_r] "=&r"(mask_r), [Filter2_l] "=&r"(Filter2_l),
+        [Filter2_r] "=&r"(Filter2_r)
+      : [subr_l] "r"(subr_l), [subr_r] "r"(subr_r), [HWM] "r"(HWM),
+        [hev] "r"(hev), [mask] "r"(mask));
 
-    /* apply filter on 4 pixesl at the same time */
-    do
-    {
-        /* load quad-byte vectors
-         * memory is 4 byte aligned
-         */
-        p1 = *((uint32_t *)(s1));
-        p2 = *((uint32_t *)(s2));
-        p3 = *((uint32_t *)(s3));
-        p4 = *((uint32_t *)(s4));
+  /* save bottom 3 bits so that we round one side +4 and the other +3 */
+  __asm__ __volatile__(
+      /* Filter1 = vp8_signed_char_clamp(Filter2 + 4) >>= 3; */
+      "addq_s.ph    %[Filter1_l],    %[Filter2_l],    %[t2]           \n\t"
+      "xor          %[invhev_l],     %[hev_l],        %[HWM]          \n\t"
+      "addq_s.ph    %[Filter1_r],    %[Filter2_r],    %[t2]           \n\t"
 
-        /* if (p1 - p4 == 0) and (p2 - p3 == 0)
-         * mask will be zero and filtering is not needed
-         */
-        if (!(((p1 - p4) == 0) && ((p2 - p3) == 0)))
-        {
+      /* Filter2 = vp8_signed_char_clamp(Filter2 + 3) >>= 3; */
+      "addq_s.ph    %[Filter2_l],    %[Filter2_l],    %[t1]           \n\t"
+      "addq_s.ph    %[Filter2_r],    %[Filter2_r],    %[t1]           \n\t"
 
-            pm1 = *((uint32_t *)(sm1));
-            p0  = *((uint32_t *)(s0));
-            p5  = *((uint32_t *)(s5));
-            p6  = *((uint32_t *)(s6));
+      "shra.ph      %[Filter1_l],    %[Filter1_l],    3               \n\t"
+      "shra.ph      %[Filter1_r],    %[Filter1_r],    3               \n\t"
 
-            vp8_filter_mask_vec_mips(limit, flimit, p1, p2, pm1, p0, p3, p4, p5, p6,
-                                     thresh, &hev, &mask);
+      "shra.ph      %[Filter2_l],    %[Filter2_l],    3               \n\t"
+      "shra.ph      %[Filter2_r],    %[Filter2_r],    3               \n\t"
+      "and          %[Filter1_l],    %[Filter1_l],    %[HWM]          \n\t"
+      "and          %[Filter1_r],    %[Filter1_r],    %[HWM]          \n\t"
+      "xor          %[invhev_r],     %[hev_r],        %[HWM]          \n\t"
 
-            /* if mask == 0 do filtering is not needed */
-            if (mask)
-            {
-                /* filtering */
-                vp8_mbfilter_mips(mask, hev, &p0, &p1, &p2, &p3, &p4, &p5);
+      /* qs0 = vp8_signed_char_clamp(qs0 - Filter1); */
+      "subq_s.ph    %[vqs0_l],       %[vqs0_l],       %[Filter1_l]    \n\t"
+      "subq_s.ph    %[vqs0_r],       %[vqs0_r],       %[Filter1_r]    \n\t"
 
-                /* unpack processed 4x4 neighborhood
-                 * memory is 4 byte aligned
-                 */
-                *((uint32_t *)s0) = p0;
-                *((uint32_t *)s1) = p1;
-                *((uint32_t *)s2) = p2;
-                *((uint32_t *)s3) = p3;
-                *((uint32_t *)s4) = p4;
-                *((uint32_t *)s5) = p5;
-            }
-        }
+      /* ps0 = vp8_signed_char_clamp(ps0 + Filter2); */
+      "addq_s.ph    %[vps0_l],       %[vps0_l],       %[Filter2_l]    \n\t"
+      "addq_s.ph    %[vps0_r],       %[vps0_r],       %[Filter2_r]    \n\t"
 
-        sm1 += 4;
-        s0  += 4;
-        s1  += 4;
-        s2  += 4;
-        s3  += 4;
-        s4  += 4;
-        s5  += 4;
-        s6  += 4;
+      : [invhev_l] "=&r"(invhev_l), [invhev_r] "=&r"(invhev_r),
+        [Filter1_l] "=&r"(Filter1_l), [Filter1_r] "=&r"(Filter1_r),
+        [Filter2_l] "+r"(Filter2_l), [Filter2_r] "+r"(Filter2_r),
+        [vps0_l] "+r"(vps0_l), [vps0_r] "+r"(vps0_r), [vqs0_l] "+r"(vqs0_l),
+        [vqs0_r] "+r"(vqs0_r)
+      : [t1] "r"(t1), [t2] "r"(t2), [HWM] "r"(HWM), [hev_l] "r"(hev_l),
+        [hev_r] "r"(hev_r));
 
-        /* load quad-byte vectors
-         * memory is 4 byte aligned
-         */
-        p1 = *((uint32_t *)(s1));
-        p2 = *((uint32_t *)(s2));
-        p3 = *((uint32_t *)(s3));
-        p4 = *((uint32_t *)(s4));
+  /* only apply wider filter if not high edge variance */
+  __asm__ __volatile__(
+      /* vp8_filter &= ~hev; */
+      "and          %[Filter2_l],    %[vp8_filter_l], %[invhev_l]     \n\t"
+      "and          %[Filter2_r],    %[vp8_filter_r], %[invhev_r]     \n\t"
 
-        /* if (p1 - p4 == 0) and (p2 - p3 == 0)
-         * mask will be zero and filtering is not needed
-         */
-        if (!(((p1 - p4) == 0) && ((p2 - p3) == 0)))
-        {
+      "shra.ph      %[Filter2_l],    %[Filter2_l],    8               \n\t"
+      "shra.ph      %[Filter2_r],    %[Filter2_r],    8               \n\t"
 
-            pm1 = *((uint32_t *)(sm1));
-            p0  = *((uint32_t *)(s0));
-            p5  = *((uint32_t *)(s5));
-            p6  = *((uint32_t *)(s6));
+      : [Filter2_l] "=&r"(Filter2_l), [Filter2_r] "=&r"(Filter2_r)
+      : [vp8_filter_l] "r"(vp8_filter_l), [vp8_filter_r] "r"(vp8_filter_r),
+        [invhev_l] "r"(invhev_l), [invhev_r] "r"(invhev_r));
 
-            vp8_filter_mask_vec_mips(limit, flimit, p1, p2, pm1, p0, p3, p4, p5, p6,
-                                     thresh, &hev, &mask);
+  /* roughly 3/7th difference across boundary */
+  __asm__ __volatile__(
+      "shll.ph      %[u3_l],         %[Filter2_l],    3               \n\t"
+      "shll.ph      %[u3_r],         %[Filter2_r],    3               \n\t"
 
-            /* if mask == 0 do filtering is not needed */
-            if (mask)
-            {
-                /* filtering */
-                vp8_mbfilter_mips(mask, hev, &p0, &p1, &p2, &p3, &p4, &p5);
+      "addq.ph      %[u3_l],         %[u3_l],         %[Filter2_l]    \n\t"
+      "addq.ph      %[u3_r],         %[u3_r],         %[Filter2_r]    \n\t"
 
-                /* unpack processed 4x4 neighborhood
-                 * memory is 4 byte aligned
-                 */
-                *((uint32_t *)s0) = p0;
-                *((uint32_t *)s1) = p1;
-                *((uint32_t *)s2) = p2;
-                *((uint32_t *)s3) = p3;
-                *((uint32_t *)s4) = p4;
-                *((uint32_t *)s5) = p5;
-            }
-        }
+      "shll.ph      %[u2_l],         %[u3_l],         1               \n\t"
+      "shll.ph      %[u2_r],         %[u3_r],         1               \n\t"
 
-        sm1 += 4;
-        s0  += 4;
-        s1  += 4;
-        s2  += 4;
-        s3  += 4;
-        s4  += 4;
-        s5  += 4;
-        s6  += 4;
+      "addq.ph      %[u1_l],         %[u3_l],         %[u2_l]         \n\t"
+      "addq.ph      %[u1_r],         %[u3_r],         %[u2_r]         \n\t"
 
-        i += 8;
-    }
+      "addq.ph      %[u2_l],         %[u2_l],         %[R63]          \n\t"
+      "addq.ph      %[u2_r],         %[u2_r],         %[R63]          \n\t"
 
-    while (i < count);
+      "addq.ph      %[u3_l],         %[u3_l],         %[R63]          \n\t"
+      "addq.ph      %[u3_r],         %[u3_r],         %[R63]          \n\t"
+
+      /* vp8_signed_char_clamp((63 + Filter2 * 27) >> 7)
+       * vp8_signed_char_clamp((63 + Filter2 * 18) >> 7)
+       */
+      "addq.ph      %[u1_l],         %[u1_l],         %[R63]          \n\t"
+      "addq.ph      %[u1_r],         %[u1_r],         %[R63]          \n\t"
+      "shra.ph      %[u1_l],         %[u1_l],         7               \n\t"
+      "shra.ph      %[u1_r],         %[u1_r],         7               \n\t"
+      "shra.ph      %[u2_l],         %[u2_l],         7               \n\t"
+      "shra.ph      %[u2_r],         %[u2_r],         7               \n\t"
+      "shll.ph      %[u1_l],         %[u1_l],         8               \n\t"
+      "shll.ph      %[u1_r],         %[u1_r],         8               \n\t"
+      "shll.ph      %[u2_l],         %[u2_l],         8               \n\t"
+      "shll.ph      %[u2_r],         %[u2_r],         8               \n\t"
+
+      /* vqs0 = vp8_signed_char_clamp(qs0 - u); */
+      "subq_s.ph    %[vqs0_l],       %[vqs0_l],       %[u1_l]         \n\t"
+      "subq_s.ph    %[vqs0_r],       %[vqs0_r],       %[u1_r]         \n\t"
+
+      /* vps0 = vp8_signed_char_clamp(ps0 + u); */
+      "addq_s.ph    %[vps0_l],       %[vps0_l],       %[u1_l]         \n\t"
+      "addq_s.ph    %[vps0_r],       %[vps0_r],       %[u1_r]         \n\t"
+
+      : [u1_l] "=&r"(u1_l), [u1_r] "=&r"(u1_r), [u2_l] "=&r"(u2_l),
+        [u2_r] "=&r"(u2_r), [u3_l] "=&r"(u3_l), [u3_r] "=&r"(u3_r),
+        [vps0_l] "+r"(vps0_l), [vps0_r] "+r"(vps0_r), [vqs0_l] "+r"(vqs0_l),
+        [vqs0_r] "+r"(vqs0_r)
+      : [R63] "r"(R63), [Filter2_l] "r"(Filter2_l), [Filter2_r] "r"(Filter2_r));
+
+  __asm__ __volatile__(
+      /* vqs1 = vp8_signed_char_clamp(qs1 - u); */
+      "subq_s.ph    %[vqs1_l],       %[vqs1_l],       %[u2_l]         \n\t"
+      "addq_s.ph    %[vps1_l],       %[vps1_l],       %[u2_l]         \n\t"
+
+      /* vps1 = vp8_signed_char_clamp(ps1 + u); */
+      "addq_s.ph    %[vps1_r],       %[vps1_r],       %[u2_r]         \n\t"
+      "subq_s.ph    %[vqs1_r],       %[vqs1_r],       %[u2_r]         \n\t"
+
+      : [vps1_l] "+r"(vps1_l), [vps1_r] "+r"(vps1_r), [vqs1_l] "+r"(vqs1_l),
+        [vqs1_r] "+r"(vqs1_r)
+      : [u2_l] "r"(u2_l), [u2_r] "r"(u2_r));
+
+  /* roughly 1/7th difference across boundary */
+  __asm__ __volatile__(
+      /* u = vp8_signed_char_clamp((63 + Filter2 * 9) >> 7); */
+      "shra.ph      %[u3_l],         %[u3_l],         7               \n\t"
+      "shra.ph      %[u3_r],         %[u3_r],         7               \n\t"
+      "shll.ph      %[u3_l],         %[u3_l],         8               \n\t"
+      "shll.ph      %[u3_r],         %[u3_r],         8               \n\t"
+
+      /* vqs2 = vp8_signed_char_clamp(qs2 - u); */
+      "subq_s.ph    %[vqs2_l],       %[vqs2_l],       %[u3_l]         \n\t"
+      "subq_s.ph    %[vqs2_r],       %[vqs2_r],       %[u3_r]         \n\t"
+
+      /* vps2 = vp8_signed_char_clamp(ps2 + u); */
+      "addq_s.ph    %[vps2_l],       %[vps2_l],       %[u3_l]         \n\t"
+      "addq_s.ph    %[vps2_r],       %[vps2_r],       %[u3_r]         \n\t"
+
+      : [u3_l] "+r"(u3_l), [u3_r] "+r"(u3_r), [vps2_l] "+r"(vps2_l),
+        [vps2_r] "+r"(vps2_r), [vqs2_l] "+r"(vqs2_l), [vqs2_r] "+r"(vqs2_r)
+      :);
+
+  /* Create quad-bytes from halfword pairs */
+  __asm__ __volatile__(
+      "and          %[vqs0_l],       %[vqs0_l],       %[HWM]          \n\t"
+      "shrl.ph      %[vqs0_r],       %[vqs0_r],       8               \n\t"
+
+      "and          %[vps0_l],       %[vps0_l],       %[HWM]          \n\t"
+      "shrl.ph      %[vps0_r],       %[vps0_r],       8               \n\t"
+
+      "and          %[vqs1_l],       %[vqs1_l],       %[HWM]          \n\t"
+      "shrl.ph      %[vqs1_r],       %[vqs1_r],       8               \n\t"
+
+      "and          %[vps1_l],       %[vps1_l],       %[HWM]          \n\t"
+      "shrl.ph      %[vps1_r],       %[vps1_r],       8               \n\t"
+
+      "and          %[vqs2_l],       %[vqs2_l],       %[HWM]          \n\t"
+      "shrl.ph      %[vqs2_r],       %[vqs2_r],       8               \n\t"
+
+      "and          %[vps2_l],       %[vps2_l],       %[HWM]          \n\t"
+      "shrl.ph      %[vps2_r],       %[vps2_r],       8               \n\t"
+
+      "or           %[vqs0_r],       %[vqs0_l],       %[vqs0_r]       \n\t"
+      "or           %[vps0_r],       %[vps0_l],       %[vps0_r]       \n\t"
+      "or           %[vqs1_r],       %[vqs1_l],       %[vqs1_r]       \n\t"
+      "or           %[vps1_r],       %[vps1_l],       %[vps1_r]       \n\t"
+      "or           %[vqs2_r],       %[vqs2_l],       %[vqs2_r]       \n\t"
+      "or           %[vps2_r],       %[vps2_l],       %[vps2_r]       \n\t"
+
+      : [vps1_l] "+r"(vps1_l), [vps1_r] "+r"(vps1_r), [vqs1_l] "+r"(vqs1_l),
+        [vqs1_r] "+r"(vqs1_r), [vps0_l] "+r"(vps0_l), [vps0_r] "+r"(vps0_r),
+        [vqs0_l] "+r"(vqs0_l), [vqs0_r] "+r"(vqs0_r), [vqs2_l] "+r"(vqs2_l),
+        [vqs2_r] "+r"(vqs2_r), [vps2_r] "+r"(vps2_r), [vps2_l] "+r"(vps2_l)
+      : [HWM] "r"(HWM));
+
+  *ps0 = vps0_r ^ N128;
+  *ps1 = vps1_r ^ N128;
+  *ps2 = vps2_r ^ N128;
+  *qs0 = vqs0_r ^ N128;
+  *qs1 = vqs1_r ^ N128;
+  *qs2 = vqs2_r ^ N128;
 }
 
-void vp8_mbloop_filter_uvhorizontal_edge_mips
-(
-    unsigned char *s,
-    int p,
-    unsigned int flimit,
-    unsigned int limit,
-    unsigned int thresh,
-    int count
-)
-{
-    uint32_t mask, hev;
-    uint32_t pm1, p0, p1, p2, p3, p4, p5, p6;
-    unsigned char *sm1, *s0, *s1, *s2, *s3, *s4, *s5, *s6;
+void vp8_mbloop_filter_horizontal_edge_mips(unsigned char *s, int p,
+                                            unsigned int flimit,
+                                            unsigned int limit,
+                                            unsigned int thresh, int count) {
+  int i;
+  uint32_t mask, hev;
+  uint32_t pm1, p0, p1, p2, p3, p4, p5, p6;
+  unsigned char *sm1, *s0, *s1, *s2, *s3, *s4, *s5, *s6;
 
-    mask = 0;
-    hev = 0;
-    p1 = 0;
-    p2 = 0;
-    p3 = 0;
-    p4 = 0;
+  mask = 0;
+  hev = 0;
+  i = 0;
+  p1 = 0;
+  p2 = 0;
+  p3 = 0;
+  p4 = 0;
 
-    /* loop filter designed to work using chars so that we can make maximum use
-     * of 8 bit simd instructions.
-     */
+  /* loop filter designed to work using chars so that we can make maximum use
+   * of 8 bit simd instructions.
+   */
 
-    sm1 = s - (p << 2);
-    s0  = s - p - p - p;
-    s1  = s - p - p;
-    s2  = s - p;
-    s3  = s;
-    s4  = s + p;
-    s5  = s + p + p;
-    s6  = s + p + p + p;
+  sm1 = s - (p << 2);
+  s0 = s - p - p - p;
+  s1 = s - p - p;
+  s2 = s - p;
+  s3 = s;
+  s4 = s + p;
+  s5 = s + p + p;
+  s6 = s + p + p + p;
 
+  /* prefetch data for load */
+  prefetch_load_lf(s + p);
+
+  /* apply filter on 4 pixesl at the same time */
+  do {
     /* load quad-byte vectors
      * memory is 4 byte aligned
      */
@@ -1819,43 +1518,40 @@
     /* if (p1 - p4 == 0) and (p2 - p3 == 0)
      * mask will be zero and filtering is not needed
      */
-    if (!(((p1 - p4) == 0) && ((p2 - p3) == 0)))
-    {
+    if (!(((p1 - p4) == 0) && ((p2 - p3) == 0))) {
+      pm1 = *((uint32_t *)(sm1));
+      p0 = *((uint32_t *)(s0));
+      p5 = *((uint32_t *)(s5));
+      p6 = *((uint32_t *)(s6));
 
-        pm1 = *((uint32_t *)(sm1));
-        p0  = *((uint32_t *)(s0));
-        p5  = *((uint32_t *)(s5));
-        p6  = *((uint32_t *)(s6));
+      vp8_filter_mask_vec_mips(limit, flimit, p1, p2, pm1, p0, p3, p4, p5, p6,
+                               thresh, &hev, &mask);
 
-        /* if mask == 0 do filtering is not needed */
-        vp8_filter_mask_vec_mips(limit, flimit, p1, p2, pm1, p0, p3, p4, p5, p6,
-                                 thresh, &hev, &mask);
+      /* if mask == 0 do filtering is not needed */
+      if (mask) {
+        /* filtering */
+        vp8_mbfilter_mips(mask, hev, &p0, &p1, &p2, &p3, &p4, &p5);
 
-        if (mask)
-        {
-            /* filtering */
-            vp8_mbfilter_mips(mask, hev, &p0, &p1, &p2, &p3, &p4, &p5);
-
-            /* unpack processed 4x4 neighborhood
-             * memory is 4 byte aligned
-             */
-            *((uint32_t *)s0) = p0;
-            *((uint32_t *)s1) = p1;
-            *((uint32_t *)s2) = p2;
-            *((uint32_t *)s3) = p3;
-            *((uint32_t *)s4) = p4;
-            *((uint32_t *)s5) = p5;
-        }
+        /* unpack processed 4x4 neighborhood
+         * memory is 4 byte aligned
+         */
+        *((uint32_t *)s0) = p0;
+        *((uint32_t *)s1) = p1;
+        *((uint32_t *)s2) = p2;
+        *((uint32_t *)s3) = p3;
+        *((uint32_t *)s4) = p4;
+        *((uint32_t *)s5) = p5;
+      }
     }
 
     sm1 += 4;
-    s0  += 4;
-    s1  += 4;
-    s2  += 4;
-    s3  += 4;
-    s4  += 4;
-    s5  += 4;
-    s6  += 4;
+    s0 += 4;
+    s1 += 4;
+    s2 += 4;
+    s3 += 4;
+    s4 += 4;
+    s5 += 4;
+    s6 += 4;
 
     /* load quad-byte vectors
      * memory is 4 byte aligned
@@ -1868,303 +1564,207 @@
     /* if (p1 - p4 == 0) and (p2 - p3 == 0)
      * mask will be zero and filtering is not needed
      */
-    if (!(((p1 - p4) == 0) && ((p2 - p3) == 0)))
-    {
+    if (!(((p1 - p4) == 0) && ((p2 - p3) == 0))) {
+      pm1 = *((uint32_t *)(sm1));
+      p0 = *((uint32_t *)(s0));
+      p5 = *((uint32_t *)(s5));
+      p6 = *((uint32_t *)(s6));
 
-        pm1 = *((uint32_t *)(sm1));
-        p0  = *((uint32_t *)(s0));
-        p5  = *((uint32_t *)(s5));
-        p6  = *((uint32_t *)(s6));
+      vp8_filter_mask_vec_mips(limit, flimit, p1, p2, pm1, p0, p3, p4, p5, p6,
+                               thresh, &hev, &mask);
 
-        vp8_filter_mask_vec_mips(limit, flimit, p1, p2, pm1, p0, p3, p4, p5, p6,
-                                 thresh, &hev, &mask);
+      /* if mask == 0 do filtering is not needed */
+      if (mask) {
+        /* filtering */
+        vp8_mbfilter_mips(mask, hev, &p0, &p1, &p2, &p3, &p4, &p5);
 
-        /* if mask == 0 do filtering is not needed */
-        if (mask)
-        {
-            /* filtering */
-            vp8_mbfilter_mips(mask, hev, &p0, &p1, &p2, &p3, &p4, &p5);
-
-            /* unpack processed 4x4 neighborhood
-             * memory is 4 byte aligned
-             */
-            *((uint32_t *)s0) = p0;
-            *((uint32_t *)s1) = p1;
-            *((uint32_t *)s2) = p2;
-            *((uint32_t *)s3) = p3;
-            *((uint32_t *)s4) = p4;
-            *((uint32_t *)s5) = p5;
-        }
-    }
-}
-
-
-void vp8_mbloop_filter_vertical_edge_mips
-(
-    unsigned char *s,
-    int p,
-    unsigned int flimit,
-    unsigned int limit,
-    unsigned int thresh,
-    int count
-)
-{
-
-    int i;
-    uint32_t mask, hev;
-    uint32_t pm1, p0, p1, p2, p3, p4, p5, p6;
-    unsigned char *s1, *s2, *s3, *s4;
-    uint32_t prim1, prim2, sec3, sec4, prim3, prim4;
-
-    mask = 0;
-    hev = 0;
-    i = 0;
-    pm1 = 0;
-    p0 = 0;
-    p1 = 0;
-    p2 = 0;
-    p3 = 0;
-    p4 = 0;
-    p5 = 0;
-    p6 = 0;
-
-    /* loop filter designed to work using chars so that we can make maximum use
-     * of 8 bit simd instructions.
-     */
-
-    /* apply filter on 4 pixesl at the same time */
-    do
-    {
-        s1 = s;
-        s2 = s + p;
-        s3 = s2 + p;
-        s4 = s3 + p;
-        s  = s4 + p;
-
-        /* load quad-byte vectors
+        /* unpack processed 4x4 neighborhood
          * memory is 4 byte aligned
          */
-        p2  = *((uint32_t *)(s1 - 4));
-        p6  = *((uint32_t *)(s1));
-        p1  = *((uint32_t *)(s2 - 4));
-        p5  = *((uint32_t *)(s2));
-        p0  = *((uint32_t *)(s3 - 4));
-        p4  = *((uint32_t *)(s3));
-        pm1 = *((uint32_t *)(s4 - 4));
-        p3  = *((uint32_t *)(s4));
+        *((uint32_t *)s0) = p0;
+        *((uint32_t *)s1) = p1;
+        *((uint32_t *)s2) = p2;
+        *((uint32_t *)s3) = p3;
+        *((uint32_t *)s4) = p4;
+        *((uint32_t *)s5) = p5;
+      }
+    }
 
-        /* transpose pm1, p0, p1, p2 */
-        __asm__ __volatile__ (
-            "precrq.qb.ph   %[prim1],   %[p2],      %[p1]       \n\t"
-            "precr.qb.ph    %[prim2],   %[p2],      %[p1]       \n\t"
-            "precrq.qb.ph   %[prim3],   %[p0],      %[pm1]      \n\t"
-            "precr.qb.ph    %[prim4],   %[p0],      %[pm1]      \n\t"
+    sm1 += 4;
+    s0 += 4;
+    s1 += 4;
+    s2 += 4;
+    s3 += 4;
+    s4 += 4;
+    s5 += 4;
+    s6 += 4;
 
-            "precrq.qb.ph   %[p1],      %[prim1],   %[prim2]    \n\t"
-            "precr.qb.ph    %[pm1],     %[prim1],   %[prim2]    \n\t"
-            "precrq.qb.ph   %[sec3],    %[prim3],   %[prim4]    \n\t"
-            "precr.qb.ph    %[sec4],    %[prim3],   %[prim4]    \n\t"
+    i += 8;
+  }
 
-            "precrq.ph.w    %[p2],      %[p1],      %[sec3]     \n\t"
-            "precrq.ph.w    %[p0],      %[pm1],     %[sec4]     \n\t"
-            "append         %[p1],      %[sec3],    16          \n\t"
-            "append         %[pm1],     %[sec4],    16          \n\t"
+  while (i < count);
+}
 
-            : [prim1] "=&r" (prim1), [prim2] "=&r" (prim2),
-              [prim3] "=&r" (prim3), [prim4] "=&r" (prim4),
-              [p2] "+r" (p2), [p1] "+r" (p1), [p0] "+r" (p0), [pm1] "+r" (pm1),
-              [sec3] "=&r" (sec3), [sec4] "=&r" (sec4)
-            :
-        );
+void vp8_mbloop_filter_uvhorizontal_edge_mips(unsigned char *s, int p,
+                                              unsigned int flimit,
+                                              unsigned int limit,
+                                              unsigned int thresh, int count) {
+  uint32_t mask, hev;
+  uint32_t pm1, p0, p1, p2, p3, p4, p5, p6;
+  unsigned char *sm1, *s0, *s1, *s2, *s3, *s4, *s5, *s6;
 
-        /* transpose p3, p4, p5, p6 */
-        __asm__ __volatile__ (
-            "precrq.qb.ph   %[prim1],   %[p6],      %[p5]       \n\t"
-            "precr.qb.ph    %[prim2],   %[p6],      %[p5]       \n\t"
-            "precrq.qb.ph   %[prim3],   %[p4],      %[p3]       \n\t"
-            "precr.qb.ph    %[prim4],   %[p4],      %[p3]       \n\t"
+  mask = 0;
+  hev = 0;
+  p1 = 0;
+  p2 = 0;
+  p3 = 0;
+  p4 = 0;
 
-            "precrq.qb.ph   %[p5],      %[prim1],   %[prim2]    \n\t"
-            "precr.qb.ph    %[p3],      %[prim1],   %[prim2]    \n\t"
-            "precrq.qb.ph   %[sec3],    %[prim3],   %[prim4]    \n\t"
-            "precr.qb.ph    %[sec4],    %[prim3],   %[prim4]    \n\t"
+  /* loop filter designed to work using chars so that we can make maximum use
+   * of 8 bit simd instructions.
+   */
 
-            "precrq.ph.w    %[p6],      %[p5],      %[sec3]     \n\t"
-            "precrq.ph.w    %[p4],      %[p3],      %[sec4]     \n\t"
-            "append         %[p5],      %[sec3],    16          \n\t"
-            "append         %[p3],      %[sec4],    16          \n\t"
+  sm1 = s - (p << 2);
+  s0 = s - p - p - p;
+  s1 = s - p - p;
+  s2 = s - p;
+  s3 = s;
+  s4 = s + p;
+  s5 = s + p + p;
+  s6 = s + p + p + p;
 
-            : [prim1] "=&r" (prim1), [prim2] "=&r" (prim2),
-              [prim3] "=&r" (prim3), [prim4] "=&r" (prim4),
-              [p6] "+r" (p6), [p5] "+r" (p5), [p4] "+r" (p4), [p3] "+r" (p3),
-              [sec3] "=&r" (sec3), [sec4] "=&r" (sec4)
-            :
-        );
+  /* load quad-byte vectors
+   * memory is 4 byte aligned
+   */
+  p1 = *((uint32_t *)(s1));
+  p2 = *((uint32_t *)(s2));
+  p3 = *((uint32_t *)(s3));
+  p4 = *((uint32_t *)(s4));
 
-        /* if (p1 - p4 == 0) and (p2 - p3 == 0)
-         * mask will be zero and filtering is not needed
-         */
-        if (!(((p1 - p4) == 0) && ((p2 - p3) == 0)))
-        {
+  /* if (p1 - p4 == 0) and (p2 - p3 == 0)
+   * mask will be zero and filtering is not needed
+   */
+  if (!(((p1 - p4) == 0) && ((p2 - p3) == 0))) {
+    pm1 = *((uint32_t *)(sm1));
+    p0 = *((uint32_t *)(s0));
+    p5 = *((uint32_t *)(s5));
+    p6 = *((uint32_t *)(s6));
 
-            vp8_filter_mask_vec_mips(limit, flimit, p1, p2, pm1, p0, p3, p4, p5, p6,
-                                     thresh, &hev, &mask);
+    /* if mask == 0 do filtering is not needed */
+    vp8_filter_mask_vec_mips(limit, flimit, p1, p2, pm1, p0, p3, p4, p5, p6,
+                             thresh, &hev, &mask);
 
-            /* if mask == 0 do filtering is not needed */
-            if (mask)
-            {
-                /* filtering */
-                vp8_mbfilter_mips(mask, hev, &p0, &p1, &p2, &p3, &p4, &p5);
+    if (mask) {
+      /* filtering */
+      vp8_mbfilter_mips(mask, hev, &p0, &p1, &p2, &p3, &p4, &p5);
 
-                /* don't use transpose on output data
-                 * because memory isn't aligned
-                 */
-                __asm__ __volatile__ (
-                    "sb         %[p5],  2(%[s4])        \n\t"
-                    "sb         %[p4],  1(%[s4])        \n\t"
-                    "sb         %[p3],  0(%[s4])        \n\t"
-                    "sb         %[p2], -1(%[s4])        \n\t"
-                    "sb         %[p1], -2(%[s4])        \n\t"
-                    "sb         %[p0], -3(%[s4])        \n\t"
-                    :
-                    : [p5] "r" (p5), [p4] "r" (p4), [p3] "r" (p3), [s4] "r" (s4),
-                      [p2] "r" (p2), [p1] "r" (p1), [p0] "r" (p0)
-                );
+      /* unpack processed 4x4 neighborhood
+       * memory is 4 byte aligned
+       */
+      *((uint32_t *)s0) = p0;
+      *((uint32_t *)s1) = p1;
+      *((uint32_t *)s2) = p2;
+      *((uint32_t *)s3) = p3;
+      *((uint32_t *)s4) = p4;
+      *((uint32_t *)s5) = p5;
+    }
+  }
 
-                __asm__ __volatile__ (
-                    "srl        %[p5], %[p5], 8         \n\t"
-                    "srl        %[p4], %[p4], 8         \n\t"
-                    "srl        %[p3], %[p3], 8         \n\t"
-                    "srl        %[p2], %[p2], 8         \n\t"
-                    "srl        %[p1], %[p1], 8         \n\t"
-                    "srl        %[p0], %[p0], 8         \n\t"
-                    : [p5] "+r" (p5), [p4] "+r" (p4), [p3] "+r" (p3),
-                      [p2] "+r" (p2), [p1] "+r" (p1), [p0] "+r" (p0)
-                    :
-                );
+  sm1 += 4;
+  s0 += 4;
+  s1 += 4;
+  s2 += 4;
+  s3 += 4;
+  s4 += 4;
+  s5 += 4;
+  s6 += 4;
 
-                __asm__ __volatile__ (
-                    "sb         %[p5],  2(%[s3])        \n\t"
-                    "sb         %[p4],  1(%[s3])        \n\t"
-                    "sb         %[p3],  0(%[s3])        \n\t"
-                    "sb         %[p2], -1(%[s3])        \n\t"
-                    "sb         %[p1], -2(%[s3])        \n\t"
-                    "sb         %[p0], -3(%[s3])        \n\t"
-                    :
-                    : [p5] "r" (p5), [p4] "r" (p4), [p3] "r" (p3), [s3] "r" (s3),
-                      [p2] "r" (p2), [p1] "r" (p1), [p0] "r" (p0)
-                );
+  /* load quad-byte vectors
+   * memory is 4 byte aligned
+   */
+  p1 = *((uint32_t *)(s1));
+  p2 = *((uint32_t *)(s2));
+  p3 = *((uint32_t *)(s3));
+  p4 = *((uint32_t *)(s4));
 
-                __asm__ __volatile__ (
-                    "srl        %[p5], %[p5], 8         \n\t"
-                    "srl        %[p4], %[p4], 8         \n\t"
-                    "srl        %[p3], %[p3], 8         \n\t"
-                    "srl        %[p2], %[p2], 8         \n\t"
-                    "srl        %[p1], %[p1], 8         \n\t"
-                    "srl        %[p0], %[p0], 8         \n\t"
-                    : [p5] "+r" (p5), [p4] "+r" (p4), [p3] "+r" (p3),
-                      [p2] "+r" (p2), [p1] "+r" (p1), [p0] "+r" (p0)
-                    :
-                );
+  /* if (p1 - p4 == 0) and (p2 - p3 == 0)
+   * mask will be zero and filtering is not needed
+   */
+  if (!(((p1 - p4) == 0) && ((p2 - p3) == 0))) {
+    pm1 = *((uint32_t *)(sm1));
+    p0 = *((uint32_t *)(s0));
+    p5 = *((uint32_t *)(s5));
+    p6 = *((uint32_t *)(s6));
 
-                __asm__ __volatile__ (
-                    "sb         %[p5],  2(%[s2])        \n\t"
-                    "sb         %[p4],  1(%[s2])        \n\t"
-                    "sb         %[p3],  0(%[s2])        \n\t"
-                    "sb         %[p2], -1(%[s2])        \n\t"
-                    "sb         %[p1], -2(%[s2])        \n\t"
-                    "sb         %[p0], -3(%[s2])        \n\t"
-                    :
-                    : [p5] "r" (p5), [p4] "r" (p4), [p3] "r" (p3), [s2] "r" (s2),
-                      [p2] "r" (p2), [p1] "r" (p1), [p0] "r" (p0)
-                );
+    vp8_filter_mask_vec_mips(limit, flimit, p1, p2, pm1, p0, p3, p4, p5, p6,
+                             thresh, &hev, &mask);
 
-                __asm__ __volatile__ (
-                    "srl        %[p5], %[p5], 8         \n\t"
-                    "srl        %[p4], %[p4], 8         \n\t"
-                    "srl        %[p3], %[p3], 8         \n\t"
-                    "srl        %[p2], %[p2], 8         \n\t"
-                    "srl        %[p1], %[p1], 8         \n\t"
-                    "srl        %[p0], %[p0], 8         \n\t"
-                    : [p5] "+r" (p5), [p4] "+r" (p4), [p3] "+r" (p3),
-                      [p2] "+r" (p2), [p1] "+r" (p1), [p0] "+r" (p0)
-                    :
-                );
+    /* if mask == 0 do filtering is not needed */
+    if (mask) {
+      /* filtering */
+      vp8_mbfilter_mips(mask, hev, &p0, &p1, &p2, &p3, &p4, &p5);
 
-                __asm__ __volatile__ (
-                    "sb         %[p5],  2(%[s1])        \n\t"
-                    "sb         %[p4],  1(%[s1])        \n\t"
-                    "sb         %[p3],  0(%[s1])        \n\t"
-                    "sb         %[p2], -1(%[s1])        \n\t"
-                    "sb         %[p1], -2(%[s1])        \n\t"
-                    "sb         %[p0], -3(%[s1])        \n\t"
-                    :
-                    : [p5] "r" (p5), [p4] "r" (p4), [p3] "r" (p3), [s1] "r" (s1),
-                      [p2] "r" (p2), [p1] "r" (p1), [p0] "r" (p0)
-                );
-            }
-        }
-
-        i += 4;
+      /* unpack processed 4x4 neighborhood
+       * memory is 4 byte aligned
+       */
+      *((uint32_t *)s0) = p0;
+      *((uint32_t *)s1) = p1;
+      *((uint32_t *)s2) = p2;
+      *((uint32_t *)s3) = p3;
+      *((uint32_t *)s4) = p4;
+      *((uint32_t *)s5) = p5;
     }
-
-    while (i < count);
+  }
 }
 
-void vp8_mbloop_filter_uvvertical_edge_mips
-(
-    unsigned char *s,
-    int p,
-    unsigned int flimit,
-    unsigned int limit,
-    unsigned int thresh,
-    int count
-)
-{
-    uint32_t mask, hev;
-    uint32_t pm1, p0, p1, p2, p3, p4, p5, p6;
-    unsigned char *s1, *s2, *s3, *s4;
-    uint32_t prim1, prim2, sec3, sec4, prim3, prim4;
+void vp8_mbloop_filter_vertical_edge_mips(unsigned char *s, int p,
+                                          unsigned int flimit,
+                                          unsigned int limit,
+                                          unsigned int thresh, int count) {
+  int i;
+  uint32_t mask, hev;
+  uint32_t pm1, p0, p1, p2, p3, p4, p5, p6;
+  unsigned char *s1, *s2, *s3, *s4;
+  uint32_t prim1, prim2, sec3, sec4, prim3, prim4;
 
-    mask = 0;
-    hev = 0;
-    pm1 = 0;
-    p0 = 0;
-    p1 = 0;
-    p2 = 0;
-    p3 = 0;
-    p4 = 0;
-    p5 = 0;
-    p6 = 0;
+  mask = 0;
+  hev = 0;
+  i = 0;
+  pm1 = 0;
+  p0 = 0;
+  p1 = 0;
+  p2 = 0;
+  p3 = 0;
+  p4 = 0;
+  p5 = 0;
+  p6 = 0;
 
-    /* loop filter designed to work using chars so that we can make maximum use
-     * of 8 bit simd instructions.
-     */
+  /* loop filter designed to work using chars so that we can make maximum use
+   * of 8 bit simd instructions.
+   */
 
-    /* apply filter on 4 pixesl at the same time */
-
+  /* apply filter on 4 pixesl at the same time */
+  do {
     s1 = s;
     s2 = s + p;
     s3 = s2 + p;
     s4 = s3 + p;
+    s = s4 + p;
 
-    /* prefetch data for load */
-    prefetch_load_lf(s + 2 * p);
-
     /* load quad-byte vectors
      * memory is 4 byte aligned
      */
-    p2  = *((uint32_t *)(s1 - 4));
-    p6  = *((uint32_t *)(s1));
-    p1  = *((uint32_t *)(s2 - 4));
-    p5  = *((uint32_t *)(s2));
-    p0  = *((uint32_t *)(s3 - 4));
-    p4  = *((uint32_t *)(s3));
+    p2 = *((uint32_t *)(s1 - 4));
+    p6 = *((uint32_t *)(s1));
+    p1 = *((uint32_t *)(s2 - 4));
+    p5 = *((uint32_t *)(s2));
+    p0 = *((uint32_t *)(s3 - 4));
+    p4 = *((uint32_t *)(s3));
     pm1 = *((uint32_t *)(s4 - 4));
-    p3  = *((uint32_t *)(s4));
+    p3 = *((uint32_t *)(s4));
 
     /* transpose pm1, p0, p1, p2 */
-    __asm__ __volatile__ (
+    __asm__ __volatile__(
         "precrq.qb.ph   %[prim1],   %[p2],      %[p1]       \n\t"
         "precr.qb.ph    %[prim2],   %[p2],      %[p1]       \n\t"
         "precrq.qb.ph   %[prim3],   %[p0],      %[pm1]      \n\t"
@@ -2180,15 +1780,13 @@
         "append         %[p1],      %[sec3],    16          \n\t"
         "append         %[pm1],     %[sec4],    16          \n\t"
 
-        : [prim1] "=&r" (prim1), [prim2] "=&r" (prim2),
-          [prim3] "=&r" (prim3), [prim4] "=&r" (prim4),
-          [p2] "+r" (p2), [p1] "+r" (p1), [p0] "+r" (p0), [pm1] "+r" (pm1),
-          [sec3] "=&r" (sec3), [sec4] "=&r" (sec4)
-        :
-    );
+        : [prim1] "=&r"(prim1), [prim2] "=&r"(prim2), [prim3] "=&r"(prim3),
+          [prim4] "=&r"(prim4), [p2] "+r"(p2), [p1] "+r"(p1), [p0] "+r"(p0),
+          [pm1] "+r"(pm1), [sec3] "=&r"(sec3), [sec4] "=&r"(sec4)
+        :);
 
     /* transpose p3, p4, p5, p6 */
-    __asm__ __volatile__ (
+    __asm__ __volatile__(
         "precrq.qb.ph   %[prim1],   %[p6],      %[p5]       \n\t"
         "precr.qb.ph    %[prim2],   %[p6],      %[p5]       \n\t"
         "precrq.qb.ph   %[prim3],   %[p4],      %[p3]       \n\t"
@@ -2204,419 +1802,595 @@
         "append         %[p5],      %[sec3],    16          \n\t"
         "append         %[p3],      %[sec4],    16          \n\t"
 
-        : [prim1] "=&r" (prim1), [prim2] "=&r" (prim2),
-          [prim3] "=&r" (prim3), [prim4] "=&r" (prim4),
-          [p6] "+r" (p6), [p5] "+r" (p5), [p4] "+r" (p4), [p3] "+r" (p3),
-          [sec3] "=&r" (sec3), [sec4] "=&r" (sec4)
-        :
-    );
+        : [prim1] "=&r"(prim1), [prim2] "=&r"(prim2), [prim3] "=&r"(prim3),
+          [prim4] "=&r"(prim4), [p6] "+r"(p6), [p5] "+r"(p5), [p4] "+r"(p4),
+          [p3] "+r"(p3), [sec3] "=&r"(sec3), [sec4] "=&r"(sec4)
+        :);
 
     /* if (p1 - p4 == 0) and (p2 - p3 == 0)
      * mask will be zero and filtering is not needed
      */
-    if (!(((p1 - p4) == 0) && ((p2 - p3) == 0)))
-    {
+    if (!(((p1 - p4) == 0) && ((p2 - p3) == 0))) {
+      vp8_filter_mask_vec_mips(limit, flimit, p1, p2, pm1, p0, p3, p4, p5, p6,
+                               thresh, &hev, &mask);
 
-        vp8_filter_mask_vec_mips(limit, flimit, p1, p2, pm1, p0, p3, p4, p5, p6,
-                                 thresh, &hev, &mask);
+      /* if mask == 0 do filtering is not needed */
+      if (mask) {
+        /* filtering */
+        vp8_mbfilter_mips(mask, hev, &p0, &p1, &p2, &p3, &p4, &p5);
 
-        /* if mask == 0 do filtering is not needed */
-        if (mask)
-        {
-            /* filtering */
-            vp8_mbfilter_mips(mask, hev, &p0, &p1, &p2, &p3, &p4, &p5);
+        /* don't use transpose on output data
+         * because memory isn't aligned
+         */
+        __asm__ __volatile__(
+            "sb         %[p5],  2(%[s4])        \n\t"
+            "sb         %[p4],  1(%[s4])        \n\t"
+            "sb         %[p3],  0(%[s4])        \n\t"
+            "sb         %[p2], -1(%[s4])        \n\t"
+            "sb         %[p1], -2(%[s4])        \n\t"
+            "sb         %[p0], -3(%[s4])        \n\t"
+            :
+            : [p5] "r"(p5), [p4] "r"(p4), [p3] "r"(p3), [s4] "r"(s4),
+              [p2] "r"(p2), [p1] "r"(p1), [p0] "r"(p0));
 
-            /* don't use transpose on output data
-             * because memory isn't aligned
-             */
-            __asm__ __volatile__ (
-                "sb         %[p5],  2(%[s4])        \n\t"
-                "sb         %[p4],  1(%[s4])        \n\t"
-                "sb         %[p3],  0(%[s4])        \n\t"
-                "sb         %[p2], -1(%[s4])        \n\t"
-                "sb         %[p1], -2(%[s4])        \n\t"
-                "sb         %[p0], -3(%[s4])        \n\t"
-                :
-                : [p5] "r" (p5), [p4] "r" (p4), [p3] "r" (p3), [s4] "r" (s4),
-                  [p2] "r" (p2), [p1] "r" (p1), [p0] "r" (p0)
-            );
+        __asm__ __volatile__(
+            "srl        %[p5], %[p5], 8         \n\t"
+            "srl        %[p4], %[p4], 8         \n\t"
+            "srl        %[p3], %[p3], 8         \n\t"
+            "srl        %[p2], %[p2], 8         \n\t"
+            "srl        %[p1], %[p1], 8         \n\t"
+            "srl        %[p0], %[p0], 8         \n\t"
+            : [p5] "+r"(p5), [p4] "+r"(p4), [p3] "+r"(p3), [p2] "+r"(p2),
+              [p1] "+r"(p1), [p0] "+r"(p0)
+            :);
 
-            __asm__ __volatile__ (
-                "srl        %[p5], %[p5], 8         \n\t"
-                "srl        %[p4], %[p4], 8         \n\t"
-                "srl        %[p3], %[p3], 8         \n\t"
-                "srl        %[p2], %[p2], 8         \n\t"
-                "srl        %[p1], %[p1], 8         \n\t"
-                "srl        %[p0], %[p0], 8         \n\t"
-                : [p5] "+r" (p5), [p4] "+r" (p4), [p3] "+r" (p3),
-                  [p2] "+r" (p2), [p1] "+r" (p1), [p0] "+r" (p0)
-                :
-            );
+        __asm__ __volatile__(
+            "sb         %[p5],  2(%[s3])        \n\t"
+            "sb         %[p4],  1(%[s3])        \n\t"
+            "sb         %[p3],  0(%[s3])        \n\t"
+            "sb         %[p2], -1(%[s3])        \n\t"
+            "sb         %[p1], -2(%[s3])        \n\t"
+            "sb         %[p0], -3(%[s3])        \n\t"
+            :
+            : [p5] "r"(p5), [p4] "r"(p4), [p3] "r"(p3), [s3] "r"(s3),
+              [p2] "r"(p2), [p1] "r"(p1), [p0] "r"(p0));
 
-            __asm__ __volatile__ (
-                "sb         %[p5],  2(%[s3])        \n\t"
-                "sb         %[p4],  1(%[s3])        \n\t"
-                "sb         %[p3],  0(%[s3])        \n\t"
-                "sb         %[p2], -1(%[s3])        \n\t"
-                "sb         %[p1], -2(%[s3])        \n\t"
-                "sb         %[p0], -3(%[s3])        \n\t"
-                :
-                : [p5] "r" (p5), [p4] "r" (p4), [p3] "r" (p3), [s3] "r" (s3),
-                  [p2] "r" (p2), [p1] "r" (p1), [p0] "r" (p0)
-            );
+        __asm__ __volatile__(
+            "srl        %[p5], %[p5], 8         \n\t"
+            "srl        %[p4], %[p4], 8         \n\t"
+            "srl        %[p3], %[p3], 8         \n\t"
+            "srl        %[p2], %[p2], 8         \n\t"
+            "srl        %[p1], %[p1], 8         \n\t"
+            "srl        %[p0], %[p0], 8         \n\t"
+            : [p5] "+r"(p5), [p4] "+r"(p4), [p3] "+r"(p3), [p2] "+r"(p2),
+              [p1] "+r"(p1), [p0] "+r"(p0)
+            :);
 
-            __asm__ __volatile__ (
-                "srl        %[p5], %[p5], 8         \n\t"
-                "srl        %[p4], %[p4], 8         \n\t"
-                "srl        %[p3], %[p3], 8         \n\t"
-                "srl        %[p2], %[p2], 8         \n\t"
-                "srl        %[p1], %[p1], 8         \n\t"
-                "srl        %[p0], %[p0], 8         \n\t"
-                : [p5] "+r" (p5), [p4] "+r" (p4), [p3] "+r" (p3),
-                  [p2] "+r" (p2), [p1] "+r" (p1), [p0] "+r" (p0)
-                :
-            );
+        __asm__ __volatile__(
+            "sb         %[p5],  2(%[s2])        \n\t"
+            "sb         %[p4],  1(%[s2])        \n\t"
+            "sb         %[p3],  0(%[s2])        \n\t"
+            "sb         %[p2], -1(%[s2])        \n\t"
+            "sb         %[p1], -2(%[s2])        \n\t"
+            "sb         %[p0], -3(%[s2])        \n\t"
+            :
+            : [p5] "r"(p5), [p4] "r"(p4), [p3] "r"(p3), [s2] "r"(s2),
+              [p2] "r"(p2), [p1] "r"(p1), [p0] "r"(p0));
 
-            __asm__ __volatile__ (
-                "sb         %[p5],  2(%[s2])        \n\t"
-                "sb         %[p4],  1(%[s2])        \n\t"
-                "sb         %[p3],  0(%[s2])        \n\t"
-                "sb         %[p2], -1(%[s2])        \n\t"
-                "sb         %[p1], -2(%[s2])        \n\t"
-                "sb         %[p0], -3(%[s2])        \n\t"
-                :
-                : [p5] "r" (p5), [p4] "r" (p4), [p3] "r" (p3), [s2] "r" (s2),
-                  [p2] "r" (p2), [p1] "r" (p1), [p0] "r" (p0)
-            );
+        __asm__ __volatile__(
+            "srl        %[p5], %[p5], 8         \n\t"
+            "srl        %[p4], %[p4], 8         \n\t"
+            "srl        %[p3], %[p3], 8         \n\t"
+            "srl        %[p2], %[p2], 8         \n\t"
+            "srl        %[p1], %[p1], 8         \n\t"
+            "srl        %[p0], %[p0], 8         \n\t"
+            : [p5] "+r"(p5), [p4] "+r"(p4), [p3] "+r"(p3), [p2] "+r"(p2),
+              [p1] "+r"(p1), [p0] "+r"(p0)
+            :);
 
-            __asm__ __volatile__ (
-                "srl        %[p5], %[p5], 8         \n\t"
-                "srl        %[p4], %[p4], 8         \n\t"
-                "srl        %[p3], %[p3], 8         \n\t"
-                "srl        %[p2], %[p2], 8         \n\t"
-                "srl        %[p1], %[p1], 8         \n\t"
-                "srl        %[p0], %[p0], 8         \n\t"
-                : [p5] "+r" (p5), [p4] "+r" (p4), [p3] "+r" (p3),
-                  [p2] "+r" (p2), [p1] "+r" (p1), [p0] "+r" (p0)
-                :
-            );
-
-            __asm__ __volatile__ (
-                "sb         %[p5],  2(%[s1])        \n\t"
-                "sb         %[p4],  1(%[s1])        \n\t"
-                "sb         %[p3],  0(%[s1])        \n\t"
-                "sb         %[p2], -1(%[s1])        \n\t"
-                "sb         %[p1], -2(%[s1])        \n\t"
-                "sb         %[p0], -3(%[s1])        \n\t"
-                :
-                : [p5] "r" (p5), [p4] "r" (p4), [p3] "r" (p3), [s1] "r" (s1),
-                  [p2] "r" (p2), [p1] "r" (p1), [p0] "r" (p0)
-            );
-        }
+        __asm__ __volatile__(
+            "sb         %[p5],  2(%[s1])        \n\t"
+            "sb         %[p4],  1(%[s1])        \n\t"
+            "sb         %[p3],  0(%[s1])        \n\t"
+            "sb         %[p2], -1(%[s1])        \n\t"
+            "sb         %[p1], -2(%[s1])        \n\t"
+            "sb         %[p0], -3(%[s1])        \n\t"
+            :
+            : [p5] "r"(p5), [p4] "r"(p4), [p3] "r"(p3), [s1] "r"(s1),
+              [p2] "r"(p2), [p1] "r"(p1), [p0] "r"(p0));
+      }
     }
 
-    s1 = s4 + p;
-    s2 = s1 + p;
-    s3 = s2 + p;
-    s4 = s3 + p;
+    i += 4;
+  }
 
-    /* load quad-byte vectors
-    * memory is 4 byte aligned
-    */
-    p2  = *((uint32_t *)(s1 - 4));
-    p6  = *((uint32_t *)(s1));
-    p1  = *((uint32_t *)(s2 - 4));
-    p5  = *((uint32_t *)(s2));
-    p0  = *((uint32_t *)(s3 - 4));
-    p4  = *((uint32_t *)(s3));
-    pm1 = *((uint32_t *)(s4 - 4));
-    p3  = *((uint32_t *)(s4));
+  while (i < count);
+}
 
-    /* transpose pm1, p0, p1, p2 */
-    __asm__ __volatile__ (
-        "precrq.qb.ph   %[prim1],   %[p2],      %[p1]       \n\t"
-        "precr.qb.ph    %[prim2],   %[p2],      %[p1]       \n\t"
-        "precrq.qb.ph   %[prim3],   %[p0],      %[pm1]      \n\t"
-        "precr.qb.ph    %[prim4],   %[p0],      %[pm1]      \n\t"
+void vp8_mbloop_filter_uvvertical_edge_mips(unsigned char *s, int p,
+                                            unsigned int flimit,
+                                            unsigned int limit,
+                                            unsigned int thresh, int count) {
+  uint32_t mask, hev;
+  uint32_t pm1, p0, p1, p2, p3, p4, p5, p6;
+  unsigned char *s1, *s2, *s3, *s4;
+  uint32_t prim1, prim2, sec3, sec4, prim3, prim4;
 
-        "precrq.qb.ph   %[p1],      %[prim1],   %[prim2]    \n\t"
-        "precr.qb.ph    %[pm1],     %[prim1],   %[prim2]    \n\t"
-        "precrq.qb.ph   %[sec3],    %[prim3],   %[prim4]    \n\t"
-        "precr.qb.ph    %[sec4],    %[prim3],   %[prim4]    \n\t"
+  mask = 0;
+  hev = 0;
+  pm1 = 0;
+  p0 = 0;
+  p1 = 0;
+  p2 = 0;
+  p3 = 0;
+  p4 = 0;
+  p5 = 0;
+  p6 = 0;
 
-        "precrq.ph.w    %[p2],      %[p1],      %[sec3]     \n\t"
-        "precrq.ph.w    %[p0],      %[pm1],     %[sec4]     \n\t"
-        "append         %[p1],      %[sec3],    16          \n\t"
-        "append         %[pm1],     %[sec4],    16          \n\t"
+  /* loop filter designed to work using chars so that we can make maximum use
+   * of 8 bit simd instructions.
+   */
 
-        : [prim1] "=&r" (prim1), [prim2] "=&r" (prim2),
-          [prim3] "=&r" (prim3), [prim4] "=&r" (prim4),
-          [p2] "+r" (p2), [p1] "+r" (p1), [p0] "+r" (p0), [pm1] "+r" (pm1),
-          [sec3] "=&r" (sec3), [sec4] "=&r" (sec4)
-        :
-    );
+  /* apply filter on 4 pixesl at the same time */
 
-    /* transpose p3, p4, p5, p6 */
-    __asm__ __volatile__ (
-        "precrq.qb.ph   %[prim1],   %[p6],      %[p5]       \n\t"
-        "precr.qb.ph    %[prim2],   %[p6],      %[p5]       \n\t"
-        "precrq.qb.ph   %[prim3],   %[p4],      %[p3]       \n\t"
-        "precr.qb.ph    %[prim4],   %[p4],      %[p3]       \n\t"
+  s1 = s;
+  s2 = s + p;
+  s3 = s2 + p;
+  s4 = s3 + p;
 
-        "precrq.qb.ph   %[p5],      %[prim1],   %[prim2]    \n\t"
-        "precr.qb.ph    %[p3],      %[prim1],   %[prim2]    \n\t"
-        "precrq.qb.ph   %[sec3],    %[prim3],   %[prim4]    \n\t"
-        "precr.qb.ph    %[sec4],    %[prim3],   %[prim4]    \n\t"
+  /* prefetch data for load */
+  prefetch_load_lf(s + 2 * p);
 
-        "precrq.ph.w    %[p6],      %[p5],      %[sec3]     \n\t"
-        "precrq.ph.w    %[p4],      %[p3],      %[sec4]     \n\t"
-        "append         %[p5],      %[sec3],    16          \n\t"
-        "append         %[p3],      %[sec4],    16          \n\t"
+  /* load quad-byte vectors
+   * memory is 4 byte aligned
+   */
+  p2 = *((uint32_t *)(s1 - 4));
+  p6 = *((uint32_t *)(s1));
+  p1 = *((uint32_t *)(s2 - 4));
+  p5 = *((uint32_t *)(s2));
+  p0 = *((uint32_t *)(s3 - 4));
+  p4 = *((uint32_t *)(s3));
+  pm1 = *((uint32_t *)(s4 - 4));
+  p3 = *((uint32_t *)(s4));
 
-        : [prim1] "=&r" (prim1), [prim2] "=&r" (prim2),
-          [prim3] "=&r" (prim3), [prim4] "=&r" (prim4),
-          [p6] "+r" (p6), [p5] "+r" (p5), [p4] "+r" (p4), [p3] "+r" (p3),
-          [sec3] "=&r" (sec3), [sec4] "=&r" (sec4)
-        :
-    );
+  /* transpose pm1, p0, p1, p2 */
+  __asm__ __volatile__(
+      "precrq.qb.ph   %[prim1],   %[p2],      %[p1]       \n\t"
+      "precr.qb.ph    %[prim2],   %[p2],      %[p1]       \n\t"
+      "precrq.qb.ph   %[prim3],   %[p0],      %[pm1]      \n\t"
+      "precr.qb.ph    %[prim4],   %[p0],      %[pm1]      \n\t"
 
-    /* if (p1 - p4 == 0) and (p2 - p3 == 0)
-     * mask will be zero and filtering is not needed
-     */
-    if (!(((p1 - p4) == 0) && ((p2 - p3) == 0)))
-    {
+      "precrq.qb.ph   %[p1],      %[prim1],   %[prim2]    \n\t"
+      "precr.qb.ph    %[pm1],     %[prim1],   %[prim2]    \n\t"
+      "precrq.qb.ph   %[sec3],    %[prim3],   %[prim4]    \n\t"
+      "precr.qb.ph    %[sec4],    %[prim3],   %[prim4]    \n\t"
 
-        vp8_filter_mask_vec_mips(limit, flimit, p1, p2, pm1, p0, p3, p4, p5, p6, thresh, &hev, &mask);
+      "precrq.ph.w    %[p2],      %[p1],      %[sec3]     \n\t"
+      "precrq.ph.w    %[p0],      %[pm1],     %[sec4]     \n\t"
+      "append         %[p1],      %[sec3],    16          \n\t"
+      "append         %[pm1],     %[sec4],    16          \n\t"
 
-        /* if mask == 0 do filtering is not needed */
-        if (mask)
-        {
-            /* filtering */
-            vp8_mbfilter_mips(mask, hev, &p0, &p1, &p2, &p3, &p4, &p5);
+      : [prim1] "=&r"(prim1), [prim2] "=&r"(prim2), [prim3] "=&r"(prim3),
+        [prim4] "=&r"(prim4), [p2] "+r"(p2), [p1] "+r"(p1), [p0] "+r"(p0),
+        [pm1] "+r"(pm1), [sec3] "=&r"(sec3), [sec4] "=&r"(sec4)
+      :);
 
-            /* don't use transpose on output data
-             * because memory isn't aligned
-             */
-            __asm__ __volatile__ (
-                "sb         %[p5],  2(%[s4])        \n\t"
-                "sb         %[p4],  1(%[s4])        \n\t"
-                "sb         %[p3],  0(%[s4])        \n\t"
-                "sb         %[p2], -1(%[s4])        \n\t"
-                "sb         %[p1], -2(%[s4])        \n\t"
-                "sb         %[p0], -3(%[s4])        \n\t"
-                :
-                : [p5] "r" (p5), [p4] "r" (p4), [p3] "r" (p3), [s4] "r" (s4),
-                  [p2] "r" (p2), [p1] "r" (p1), [p0] "r" (p0)
-            );
+  /* transpose p3, p4, p5, p6 */
+  __asm__ __volatile__(
+      "precrq.qb.ph   %[prim1],   %[p6],      %[p5]       \n\t"
+      "precr.qb.ph    %[prim2],   %[p6],      %[p5]       \n\t"
+      "precrq.qb.ph   %[prim3],   %[p4],      %[p3]       \n\t"
+      "precr.qb.ph    %[prim4],   %[p4],      %[p3]       \n\t"
 
-            __asm__ __volatile__ (
-                "srl        %[p5], %[p5], 8         \n\t"
-                "srl        %[p4], %[p4], 8         \n\t"
-                "srl        %[p3], %[p3], 8         \n\t"
-                "srl        %[p2], %[p2], 8         \n\t"
-                "srl        %[p1], %[p1], 8         \n\t"
-                "srl        %[p0], %[p0], 8         \n\t"
-                : [p5] "+r" (p5), [p4] "+r" (p4), [p3] "+r" (p3),
-                  [p2] "+r" (p2), [p1] "+r" (p1), [p0] "+r" (p0)
-                :
-            );
+      "precrq.qb.ph   %[p5],      %[prim1],   %[prim2]    \n\t"
+      "precr.qb.ph    %[p3],      %[prim1],   %[prim2]    \n\t"
+      "precrq.qb.ph   %[sec3],    %[prim3],   %[prim4]    \n\t"
+      "precr.qb.ph    %[sec4],    %[prim3],   %[prim4]    \n\t"
 
-            __asm__ __volatile__ (
-                "sb         %[p5],  2(%[s3])        \n\t"
-                "sb         %[p4],  1(%[s3])        \n\t"
-                "sb         %[p3],  0(%[s3])        \n\t"
-                "sb         %[p2], -1(%[s3])        \n\t"
-                "sb         %[p1], -2(%[s3])        \n\t"
-                "sb         %[p0], -3(%[s3])        \n\t"
-                :
-                : [p5] "r" (p5), [p4] "r" (p4), [p3] "r" (p3), [s3] "r" (s3),
-                  [p2] "r" (p2), [p1] "r" (p1), [p0] "r" (p0)
-            );
+      "precrq.ph.w    %[p6],      %[p5],      %[sec3]     \n\t"
+      "precrq.ph.w    %[p4],      %[p3],      %[sec4]     \n\t"
+      "append         %[p5],      %[sec3],    16          \n\t"
+      "append         %[p3],      %[sec4],    16          \n\t"
 
-            __asm__ __volatile__ (
-                "srl        %[p5], %[p5], 8         \n\t"
-                "srl        %[p4], %[p4], 8         \n\t"
-                "srl        %[p3], %[p3], 8         \n\t"
-                "srl        %[p2], %[p2], 8         \n\t"
-                "srl        %[p1], %[p1], 8         \n\t"
-                "srl        %[p0], %[p0], 8         \n\t"
-                : [p5] "+r" (p5), [p4] "+r" (p4), [p3] "+r" (p3),
-                  [p2] "+r" (p2), [p1] "+r" (p1), [p0] "+r" (p0)
-                :
-            );
+      : [prim1] "=&r"(prim1), [prim2] "=&r"(prim2), [prim3] "=&r"(prim3),
+        [prim4] "=&r"(prim4), [p6] "+r"(p6), [p5] "+r"(p5), [p4] "+r"(p4),
+        [p3] "+r"(p3), [sec3] "=&r"(sec3), [sec4] "=&r"(sec4)
+      :);
 
-            __asm__ __volatile__ (
-                "sb         %[p5],  2(%[s2])        \n\t"
-                "sb         %[p4],  1(%[s2])        \n\t"
-                "sb         %[p3],  0(%[s2])        \n\t"
-                "sb         %[p2], -1(%[s2])        \n\t"
-                "sb         %[p1], -2(%[s2])        \n\t"
-                "sb         %[p0], -3(%[s2])        \n\t"
-                :
-                : [p5] "r" (p5), [p4] "r" (p4), [p3] "r" (p3), [s2] "r" (s2),
-                  [p2] "r" (p2), [p1] "r" (p1), [p0] "r" (p0)
-            );
+  /* if (p1 - p4 == 0) and (p2 - p3 == 0)
+   * mask will be zero and filtering is not needed
+   */
+  if (!(((p1 - p4) == 0) && ((p2 - p3) == 0))) {
+    vp8_filter_mask_vec_mips(limit, flimit, p1, p2, pm1, p0, p3, p4, p5, p6,
+                             thresh, &hev, &mask);
 
-            __asm__ __volatile__ (
-                "srl        %[p5], %[p5], 8         \n\t"
-                "srl        %[p4], %[p4], 8         \n\t"
-                "srl        %[p3], %[p3], 8         \n\t"
-                "srl        %[p2], %[p2], 8         \n\t"
-                "srl        %[p1], %[p1], 8         \n\t"
-                "srl        %[p0], %[p0], 8         \n\t"
-                : [p5] "+r" (p5), [p4] "+r" (p4), [p3] "+r" (p3),
-                  [p2] "+r" (p2), [p1] "+r" (p1), [p0] "+r" (p0)
-                :
-            );
+    /* if mask == 0 do filtering is not needed */
+    if (mask) {
+      /* filtering */
+      vp8_mbfilter_mips(mask, hev, &p0, &p1, &p2, &p3, &p4, &p5);
 
-            __asm__ __volatile__ (
-                "sb         %[p5],  2(%[s1])        \n\t"
-                "sb         %[p4],  1(%[s1])        \n\t"
-                "sb         %[p3],  0(%[s1])        \n\t"
-                "sb         %[p2], -1(%[s1])        \n\t"
-                "sb         %[p1], -2(%[s1])        \n\t"
-                "sb         %[p0], -3(%[s1])        \n\t"
-                :
-                : [p5] "r" (p5), [p4] "r" (p4), [p3] "r" (p3), [s1] "r" (s1),
-                  [p2] "r" (p2), [p1] "r" (p1), [p0] "r" (p0)
-            );
-        }
-    }
-}
+      /* don't use transpose on output data
+       * because memory isn't aligned
+       */
+      __asm__ __volatile__(
+          "sb         %[p5],  2(%[s4])        \n\t"
+          "sb         %[p4],  1(%[s4])        \n\t"
+          "sb         %[p3],  0(%[s4])        \n\t"
+          "sb         %[p2], -1(%[s4])        \n\t"
+          "sb         %[p1], -2(%[s4])        \n\t"
+          "sb         %[p0], -3(%[s4])        \n\t"
+          :
+          : [p5] "r"(p5), [p4] "r"(p4), [p3] "r"(p3), [s4] "r"(s4),
+            [p2] "r"(p2), [p1] "r"(p1), [p0] "r"(p0));
 
-/* Horizontal MB filtering */
-void vp8_loop_filter_mbh_dspr2(unsigned char *y_ptr, unsigned char *u_ptr, unsigned char *v_ptr,
-                               int y_stride, int uv_stride, loop_filter_info *lfi)
-{
-    unsigned int thresh_vec, flimit_vec, limit_vec;
-    unsigned char thresh, flimit, limit, flimit_temp;
+      __asm__ __volatile__(
+          "srl        %[p5], %[p5], 8         \n\t"
+          "srl        %[p4], %[p4], 8         \n\t"
+          "srl        %[p3], %[p3], 8         \n\t"
+          "srl        %[p2], %[p2], 8         \n\t"
+          "srl        %[p1], %[p1], 8         \n\t"
+          "srl        %[p0], %[p0], 8         \n\t"
+          : [p5] "+r"(p5), [p4] "+r"(p4), [p3] "+r"(p3), [p2] "+r"(p2),
+            [p1] "+r"(p1), [p0] "+r"(p0)
+          :);
 
-    /* use direct value instead pointers */
-    limit = *(lfi->lim);
-    flimit_temp = *(lfi->mblim);
-    thresh = *(lfi->hev_thr);
-    flimit = flimit_temp;
+      __asm__ __volatile__(
+          "sb         %[p5],  2(%[s3])        \n\t"
+          "sb         %[p4],  1(%[s3])        \n\t"
+          "sb         %[p3],  0(%[s3])        \n\t"
+          "sb         %[p2], -1(%[s3])        \n\t"
+          "sb         %[p1], -2(%[s3])        \n\t"
+          "sb         %[p0], -3(%[s3])        \n\t"
+          :
+          : [p5] "r"(p5), [p4] "r"(p4), [p3] "r"(p3), [s3] "r"(s3),
+            [p2] "r"(p2), [p1] "r"(p1), [p0] "r"(p0));
 
-    /* create quad-byte */
-    __asm__ __volatile__ (
-        "replv.qb       %[thresh_vec], %[thresh]    \n\t"
-        "replv.qb       %[flimit_vec], %[flimit]    \n\t"
-        "replv.qb       %[limit_vec],  %[limit]     \n\t"
-        : [thresh_vec] "=&r" (thresh_vec), [flimit_vec] "=&r" (flimit_vec), [limit_vec] "=r" (limit_vec)
-        : [thresh] "r" (thresh), [flimit] "r" (flimit), [limit] "r" (limit)
-    );
+      __asm__ __volatile__(
+          "srl        %[p5], %[p5], 8         \n\t"
+          "srl        %[p4], %[p4], 8         \n\t"
+          "srl        %[p3], %[p3], 8         \n\t"
+          "srl        %[p2], %[p2], 8         \n\t"
+          "srl        %[p1], %[p1], 8         \n\t"
+          "srl        %[p0], %[p0], 8         \n\t"
+          : [p5] "+r"(p5), [p4] "+r"(p4), [p3] "+r"(p3), [p2] "+r"(p2),
+            [p1] "+r"(p1), [p0] "+r"(p0)
+          :);
 
-    vp8_mbloop_filter_horizontal_edge_mips(y_ptr, y_stride, flimit_vec, limit_vec, thresh_vec, 16);
+      __asm__ __volatile__(
+          "sb         %[p5],  2(%[s2])        \n\t"
+          "sb         %[p4],  1(%[s2])        \n\t"
+          "sb         %[p3],  0(%[s2])        \n\t"
+          "sb         %[p2], -1(%[s2])        \n\t"
+          "sb         %[p1], -2(%[s2])        \n\t"
+          "sb         %[p0], -3(%[s2])        \n\t"
+          :
+          : [p5] "r"(p5), [p4] "r"(p4), [p3] "r"(p3), [s2] "r"(s2),
+            [p2] "r"(p2), [p1] "r"(p1), [p0] "r"(p0));
 
-    if (u_ptr)
-    {
-        vp8_mbloop_filter_uvhorizontal_edge_mips(u_ptr, uv_stride, flimit_vec, limit_vec, thresh_vec, 0);
+      __asm__ __volatile__(
+          "srl        %[p5], %[p5], 8         \n\t"
+          "srl        %[p4], %[p4], 8         \n\t"
+          "srl        %[p3], %[p3], 8         \n\t"
+          "srl        %[p2], %[p2], 8         \n\t"
+          "srl        %[p1], %[p1], 8         \n\t"
+          "srl        %[p0], %[p0], 8         \n\t"
+          : [p5] "+r"(p5), [p4] "+r"(p4), [p3] "+r"(p3), [p2] "+r"(p2),
+            [p1] "+r"(p1), [p0] "+r"(p0)
+          :);
+
+      __asm__ __volatile__(
+          "sb         %[p5],  2(%[s1])        \n\t"
+          "sb         %[p4],  1(%[s1])        \n\t"
+          "sb         %[p3],  0(%[s1])        \n\t"
+          "sb         %[p2], -1(%[s1])        \n\t"
+          "sb         %[p1], -2(%[s1])        \n\t"
+          "sb         %[p0], -3(%[s1])        \n\t"
+          :
+          : [p5] "r"(p5), [p4] "r"(p4), [p3] "r"(p3), [s1] "r"(s1),
+            [p2] "r"(p2), [p1] "r"(p1), [p0] "r"(p0));
     }
+  }
 
-    if (v_ptr)
-    {
-        vp8_mbloop_filter_uvhorizontal_edge_mips(v_ptr, uv_stride, flimit_vec, limit_vec, thresh_vec, 0);
+  s1 = s4 + p;
+  s2 = s1 + p;
+  s3 = s2 + p;
+  s4 = s3 + p;
+
+  /* load quad-byte vectors
+  * memory is 4 byte aligned
+  */
+  p2 = *((uint32_t *)(s1 - 4));
+  p6 = *((uint32_t *)(s1));
+  p1 = *((uint32_t *)(s2 - 4));
+  p5 = *((uint32_t *)(s2));
+  p0 = *((uint32_t *)(s3 - 4));
+  p4 = *((uint32_t *)(s3));
+  pm1 = *((uint32_t *)(s4 - 4));
+  p3 = *((uint32_t *)(s4));
+
+  /* transpose pm1, p0, p1, p2 */
+  __asm__ __volatile__(
+      "precrq.qb.ph   %[prim1],   %[p2],      %[p1]       \n\t"
+      "precr.qb.ph    %[prim2],   %[p2],      %[p1]       \n\t"
+      "precrq.qb.ph   %[prim3],   %[p0],      %[pm1]      \n\t"
+      "precr.qb.ph    %[prim4],   %[p0],      %[pm1]      \n\t"
+
+      "precrq.qb.ph   %[p1],      %[prim1],   %[prim2]    \n\t"
+      "precr.qb.ph    %[pm1],     %[prim1],   %[prim2]    \n\t"
+      "precrq.qb.ph   %[sec3],    %[prim3],   %[prim4]    \n\t"
+      "precr.qb.ph    %[sec4],    %[prim3],   %[prim4]    \n\t"
+
+      "precrq.ph.w    %[p2],      %[p1],      %[sec3]     \n\t"
+      "precrq.ph.w    %[p0],      %[pm1],     %[sec4]     \n\t"
+      "append         %[p1],      %[sec3],    16          \n\t"
+      "append         %[pm1],     %[sec4],    16          \n\t"
+
+      : [prim1] "=&r"(prim1), [prim2] "=&r"(prim2), [prim3] "=&r"(prim3),
+        [prim4] "=&r"(prim4), [p2] "+r"(p2), [p1] "+r"(p1), [p0] "+r"(p0),
+        [pm1] "+r"(pm1), [sec3] "=&r"(sec3), [sec4] "=&r"(sec4)
+      :);
+
+  /* transpose p3, p4, p5, p6 */
+  __asm__ __volatile__(
+      "precrq.qb.ph   %[prim1],   %[p6],      %[p5]       \n\t"
+      "precr.qb.ph    %[prim2],   %[p6],      %[p5]       \n\t"
+      "precrq.qb.ph   %[prim3],   %[p4],      %[p3]       \n\t"
+      "precr.qb.ph    %[prim4],   %[p4],      %[p3]       \n\t"
+
+      "precrq.qb.ph   %[p5],      %[prim1],   %[prim2]    \n\t"
+      "precr.qb.ph    %[p3],      %[prim1],   %[prim2]    \n\t"
+      "precrq.qb.ph   %[sec3],    %[prim3],   %[prim4]    \n\t"
+      "precr.qb.ph    %[sec4],    %[prim3],   %[prim4]    \n\t"
+
+      "precrq.ph.w    %[p6],      %[p5],      %[sec3]     \n\t"
+      "precrq.ph.w    %[p4],      %[p3],      %[sec4]     \n\t"
+      "append         %[p5],      %[sec3],    16          \n\t"
+      "append         %[p3],      %[sec4],    16          \n\t"
+
+      : [prim1] "=&r"(prim1), [prim2] "=&r"(prim2), [prim3] "=&r"(prim3),
+        [prim4] "=&r"(prim4), [p6] "+r"(p6), [p5] "+r"(p5), [p4] "+r"(p4),
+        [p3] "+r"(p3), [sec3] "=&r"(sec3), [sec4] "=&r"(sec4)
+      :);
+
+  /* if (p1 - p4 == 0) and (p2 - p3 == 0)
+   * mask will be zero and filtering is not needed
+   */
+  if (!(((p1 - p4) == 0) && ((p2 - p3) == 0))) {
+    vp8_filter_mask_vec_mips(limit, flimit, p1, p2, pm1, p0, p3, p4, p5, p6,
+                             thresh, &hev, &mask);
+
+    /* if mask == 0 do filtering is not needed */
+    if (mask) {
+      /* filtering */
+      vp8_mbfilter_mips(mask, hev, &p0, &p1, &p2, &p3, &p4, &p5);
+
+      /* don't use transpose on output data
+       * because memory isn't aligned
+       */
+      __asm__ __volatile__(
+          "sb         %[p5],  2(%[s4])        \n\t"
+          "sb         %[p4],  1(%[s4])        \n\t"
+          "sb         %[p3],  0(%[s4])        \n\t"
+          "sb         %[p2], -1(%[s4])        \n\t"
+          "sb         %[p1], -2(%[s4])        \n\t"
+          "sb         %[p0], -3(%[s4])        \n\t"
+          :
+          : [p5] "r"(p5), [p4] "r"(p4), [p3] "r"(p3), [s4] "r"(s4),
+            [p2] "r"(p2), [p1] "r"(p1), [p0] "r"(p0));
+
+      __asm__ __volatile__(
+          "srl        %[p5], %[p5], 8         \n\t"
+          "srl        %[p4], %[p4], 8         \n\t"
+          "srl        %[p3], %[p3], 8         \n\t"
+          "srl        %[p2], %[p2], 8         \n\t"
+          "srl        %[p1], %[p1], 8         \n\t"
+          "srl        %[p0], %[p0], 8         \n\t"
+          : [p5] "+r"(p5), [p4] "+r"(p4), [p3] "+r"(p3), [p2] "+r"(p2),
+            [p1] "+r"(p1), [p0] "+r"(p0)
+          :);
+
+      __asm__ __volatile__(
+          "sb         %[p5],  2(%[s3])        \n\t"
+          "sb         %[p4],  1(%[s3])        \n\t"
+          "sb         %[p3],  0(%[s3])        \n\t"
+          "sb         %[p2], -1(%[s3])        \n\t"
+          "sb         %[p1], -2(%[s3])        \n\t"
+          "sb         %[p0], -3(%[s3])        \n\t"
+          :
+          : [p5] "r"(p5), [p4] "r"(p4), [p3] "r"(p3), [s3] "r"(s3),
+            [p2] "r"(p2), [p1] "r"(p1), [p0] "r"(p0));
+
+      __asm__ __volatile__(
+          "srl        %[p5], %[p5], 8         \n\t"
+          "srl        %[p4], %[p4], 8         \n\t"
+          "srl        %[p3], %[p3], 8         \n\t"
+          "srl        %[p2], %[p2], 8         \n\t"
+          "srl        %[p1], %[p1], 8         \n\t"
+          "srl        %[p0], %[p0], 8         \n\t"
+          : [p5] "+r"(p5), [p4] "+r"(p4), [p3] "+r"(p3), [p2] "+r"(p2),
+            [p1] "+r"(p1), [p0] "+r"(p0)
+          :);
+
+      __asm__ __volatile__(
+          "sb         %[p5],  2(%[s2])        \n\t"
+          "sb         %[p4],  1(%[s2])        \n\t"
+          "sb         %[p3],  0(%[s2])        \n\t"
+          "sb         %[p2], -1(%[s2])        \n\t"
+          "sb         %[p1], -2(%[s2])        \n\t"
+          "sb         %[p0], -3(%[s2])        \n\t"
+          :
+          : [p5] "r"(p5), [p4] "r"(p4), [p3] "r"(p3), [s2] "r"(s2),
+            [p2] "r"(p2), [p1] "r"(p1), [p0] "r"(p0));
+
+      __asm__ __volatile__(
+          "srl        %[p5], %[p5], 8         \n\t"
+          "srl        %[p4], %[p4], 8         \n\t"
+          "srl        %[p3], %[p3], 8         \n\t"
+          "srl        %[p2], %[p2], 8         \n\t"
+          "srl        %[p1], %[p1], 8         \n\t"
+          "srl        %[p0], %[p0], 8         \n\t"
+          : [p5] "+r"(p5), [p4] "+r"(p4), [p3] "+r"(p3), [p2] "+r"(p2),
+            [p1] "+r"(p1), [p0] "+r"(p0)
+          :);
+
+      __asm__ __volatile__(
+          "sb         %[p5],  2(%[s1])        \n\t"
+          "sb         %[p4],  1(%[s1])        \n\t"
+          "sb         %[p3],  0(%[s1])        \n\t"
+          "sb         %[p2], -1(%[s1])        \n\t"
+          "sb         %[p1], -2(%[s1])        \n\t"
+          "sb         %[p0], -3(%[s1])        \n\t"
+          :
+          : [p5] "r"(p5), [p4] "r"(p4), [p3] "r"(p3), [s1] "r"(s1),
+            [p2] "r"(p2), [p1] "r"(p1), [p0] "r"(p0));
     }
+  }
 }
 
+/* Horizontal MB filtering */
+void vp8_loop_filter_mbh_dspr2(unsigned char *y_ptr, unsigned char *u_ptr,
+                               unsigned char *v_ptr, int y_stride,
+                               int uv_stride, loop_filter_info *lfi) {
+  unsigned int thresh_vec, flimit_vec, limit_vec;
+  unsigned char thresh, flimit, limit, flimit_temp;
 
+  /* use direct value instead pointers */
+  limit = *(lfi->lim);
+  flimit_temp = *(lfi->mblim);
+  thresh = *(lfi->hev_thr);
+  flimit = flimit_temp;
+
+  /* create quad-byte */
+  __asm__ __volatile__(
+      "replv.qb       %[thresh_vec], %[thresh]    \n\t"
+      "replv.qb       %[flimit_vec], %[flimit]    \n\t"
+      "replv.qb       %[limit_vec],  %[limit]     \n\t"
+      : [thresh_vec] "=&r"(thresh_vec), [flimit_vec] "=&r"(flimit_vec),
+        [limit_vec] "=r"(limit_vec)
+      : [thresh] "r"(thresh), [flimit] "r"(flimit), [limit] "r"(limit));
+
+  vp8_mbloop_filter_horizontal_edge_mips(y_ptr, y_stride, flimit_vec, limit_vec,
+                                         thresh_vec, 16);
+
+  if (u_ptr) {
+    vp8_mbloop_filter_uvhorizontal_edge_mips(u_ptr, uv_stride, flimit_vec,
+                                             limit_vec, thresh_vec, 0);
+  }
+
+  if (v_ptr) {
+    vp8_mbloop_filter_uvhorizontal_edge_mips(v_ptr, uv_stride, flimit_vec,
+                                             limit_vec, thresh_vec, 0);
+  }
+}
+
 /* Vertical MB Filtering */
-void vp8_loop_filter_mbv_dspr2(unsigned char *y_ptr, unsigned char *u_ptr, unsigned char *v_ptr,
-                               int y_stride, int uv_stride, loop_filter_info *lfi)
-{
-    unsigned int thresh_vec, flimit_vec, limit_vec;
-    unsigned char thresh, flimit, limit, flimit_temp;
+void vp8_loop_filter_mbv_dspr2(unsigned char *y_ptr, unsigned char *u_ptr,
+                               unsigned char *v_ptr, int y_stride,
+                               int uv_stride, loop_filter_info *lfi) {
+  unsigned int thresh_vec, flimit_vec, limit_vec;
+  unsigned char thresh, flimit, limit, flimit_temp;
 
-    /* use direct value instead pointers */
-    limit = *(lfi->lim);
-    flimit_temp = *(lfi->mblim);
-    thresh = *(lfi->hev_thr);
-    flimit = flimit_temp;
+  /* use direct value instead pointers */
+  limit = *(lfi->lim);
+  flimit_temp = *(lfi->mblim);
+  thresh = *(lfi->hev_thr);
+  flimit = flimit_temp;
 
-    /* create quad-byte */
-    __asm__ __volatile__ (
-        "replv.qb       %[thresh_vec], %[thresh]    \n\t"
-        "replv.qb       %[flimit_vec], %[flimit]    \n\t"
-        "replv.qb       %[limit_vec],  %[limit]     \n\t"
-        : [thresh_vec] "=&r" (thresh_vec), [flimit_vec] "=&r" (flimit_vec), [limit_vec] "=r" (limit_vec)
-        : [thresh] "r" (thresh), [flimit] "r" (flimit), [limit] "r" (limit)
-    );
+  /* create quad-byte */
+  __asm__ __volatile__(
+      "replv.qb       %[thresh_vec], %[thresh]    \n\t"
+      "replv.qb       %[flimit_vec], %[flimit]    \n\t"
+      "replv.qb       %[limit_vec],  %[limit]     \n\t"
+      : [thresh_vec] "=&r"(thresh_vec), [flimit_vec] "=&r"(flimit_vec),
+        [limit_vec] "=r"(limit_vec)
+      : [thresh] "r"(thresh), [flimit] "r"(flimit), [limit] "r"(limit));
 
-    vp8_mbloop_filter_vertical_edge_mips(y_ptr, y_stride, flimit_vec, limit_vec, thresh_vec, 16);
+  vp8_mbloop_filter_vertical_edge_mips(y_ptr, y_stride, flimit_vec, limit_vec,
+                                       thresh_vec, 16);
 
-    if (u_ptr)
-        vp8_mbloop_filter_uvvertical_edge_mips(u_ptr, uv_stride, flimit_vec, limit_vec, thresh_vec, 0);
+  if (u_ptr)
+    vp8_mbloop_filter_uvvertical_edge_mips(u_ptr, uv_stride, flimit_vec,
+                                           limit_vec, thresh_vec, 0);
 
-    if (v_ptr)
-        vp8_mbloop_filter_uvvertical_edge_mips(v_ptr, uv_stride, flimit_vec, limit_vec, thresh_vec, 0);
+  if (v_ptr)
+    vp8_mbloop_filter_uvvertical_edge_mips(v_ptr, uv_stride, flimit_vec,
+                                           limit_vec, thresh_vec, 0);
 }
 
-
 /* Horizontal B Filtering */
-void vp8_loop_filter_bh_dspr2(unsigned char *y_ptr, unsigned char *u_ptr, unsigned char *v_ptr,
-                              int y_stride, int uv_stride, loop_filter_info *lfi)
-{
-    unsigned int thresh_vec, flimit_vec, limit_vec;
-    unsigned char thresh, flimit, limit, flimit_temp;
+void vp8_loop_filter_bh_dspr2(unsigned char *y_ptr, unsigned char *u_ptr,
+                              unsigned char *v_ptr, int y_stride, int uv_stride,
+                              loop_filter_info *lfi) {
+  unsigned int thresh_vec, flimit_vec, limit_vec;
+  unsigned char thresh, flimit, limit, flimit_temp;
 
-    /* use direct value instead pointers */
-    limit = *(lfi->lim);
-    flimit_temp = *(lfi->blim);
-    thresh = *(lfi->hev_thr);
-    flimit = flimit_temp;
+  /* use direct value instead pointers */
+  limit = *(lfi->lim);
+  flimit_temp = *(lfi->blim);
+  thresh = *(lfi->hev_thr);
+  flimit = flimit_temp;
 
-    /* create quad-byte */
-    __asm__ __volatile__ (
-        "replv.qb       %[thresh_vec], %[thresh]    \n\t"
-        "replv.qb       %[flimit_vec], %[flimit]    \n\t"
-        "replv.qb       %[limit_vec],  %[limit]     \n\t"
-        : [thresh_vec] "=&r" (thresh_vec), [flimit_vec] "=&r" (flimit_vec), [limit_vec] "=r" (limit_vec)
-        : [thresh] "r" (thresh), [flimit] "r" (flimit), [limit] "r" (limit)
-    );
+  /* create quad-byte */
+  __asm__ __volatile__(
+      "replv.qb       %[thresh_vec], %[thresh]    \n\t"
+      "replv.qb       %[flimit_vec], %[flimit]    \n\t"
+      "replv.qb       %[limit_vec],  %[limit]     \n\t"
+      : [thresh_vec] "=&r"(thresh_vec), [flimit_vec] "=&r"(flimit_vec),
+        [limit_vec] "=r"(limit_vec)
+      : [thresh] "r"(thresh), [flimit] "r"(flimit), [limit] "r"(limit));
 
-    vp8_loop_filter_horizontal_edge_mips(y_ptr + 4 * y_stride, y_stride, flimit_vec, limit_vec, thresh_vec, 16);
-    vp8_loop_filter_horizontal_edge_mips(y_ptr + 8 * y_stride, y_stride, flimit_vec, limit_vec, thresh_vec, 16);
-    vp8_loop_filter_horizontal_edge_mips(y_ptr + 12 * y_stride, y_stride, flimit_vec, limit_vec, thresh_vec, 16);
+  vp8_loop_filter_horizontal_edge_mips(y_ptr + 4 * y_stride, y_stride,
+                                       flimit_vec, limit_vec, thresh_vec, 16);
+  vp8_loop_filter_horizontal_edge_mips(y_ptr + 8 * y_stride, y_stride,
+                                       flimit_vec, limit_vec, thresh_vec, 16);
+  vp8_loop_filter_horizontal_edge_mips(y_ptr + 12 * y_stride, y_stride,
+                                       flimit_vec, limit_vec, thresh_vec, 16);
 
-    if (u_ptr)
-        vp8_loop_filter_uvhorizontal_edge_mips(u_ptr + 4 * uv_stride, uv_stride, flimit_vec, limit_vec, thresh_vec, 0);
+  if (u_ptr)
+    vp8_loop_filter_uvhorizontal_edge_mips(
+        u_ptr + 4 * uv_stride, uv_stride, flimit_vec, limit_vec, thresh_vec, 0);
 
-    if (v_ptr)
-        vp8_loop_filter_uvhorizontal_edge_mips(v_ptr + 4 * uv_stride, uv_stride, flimit_vec, limit_vec, thresh_vec, 0);
+  if (v_ptr)
+    vp8_loop_filter_uvhorizontal_edge_mips(
+        v_ptr + 4 * uv_stride, uv_stride, flimit_vec, limit_vec, thresh_vec, 0);
 }
 
-
 /* Vertical B Filtering */
-void vp8_loop_filter_bv_dspr2(unsigned char *y_ptr, unsigned char *u_ptr, unsigned char *v_ptr,
-                              int y_stride, int uv_stride, loop_filter_info *lfi)
-{
-    unsigned int thresh_vec, flimit_vec, limit_vec;
-    unsigned char thresh, flimit, limit, flimit_temp;
+void vp8_loop_filter_bv_dspr2(unsigned char *y_ptr, unsigned char *u_ptr,
+                              unsigned char *v_ptr, int y_stride, int uv_stride,
+                              loop_filter_info *lfi) {
+  unsigned int thresh_vec, flimit_vec, limit_vec;
+  unsigned char thresh, flimit, limit, flimit_temp;
 
-    /* use direct value instead pointers */
-    limit = *(lfi->lim);
-    flimit_temp = *(lfi->blim);
-    thresh = *(lfi->hev_thr);
-    flimit = flimit_temp;
+  /* use direct value instead pointers */
+  limit = *(lfi->lim);
+  flimit_temp = *(lfi->blim);
+  thresh = *(lfi->hev_thr);
+  flimit = flimit_temp;
 
-    /* create quad-byte */
-    __asm__ __volatile__ (
-        "replv.qb       %[thresh_vec], %[thresh]    \n\t"
-        "replv.qb       %[flimit_vec], %[flimit]    \n\t"
-        "replv.qb       %[limit_vec],  %[limit]     \n\t"
-        : [thresh_vec] "=&r" (thresh_vec), [flimit_vec] "=&r" (flimit_vec), [limit_vec] "=r" (limit_vec)
-        : [thresh] "r" (thresh), [flimit] "r" (flimit), [limit] "r" (limit)
-    );
+  /* create quad-byte */
+  __asm__ __volatile__(
+      "replv.qb       %[thresh_vec], %[thresh]    \n\t"
+      "replv.qb       %[flimit_vec], %[flimit]    \n\t"
+      "replv.qb       %[limit_vec],  %[limit]     \n\t"
+      : [thresh_vec] "=&r"(thresh_vec), [flimit_vec] "=&r"(flimit_vec),
+        [limit_vec] "=r"(limit_vec)
+      : [thresh] "r"(thresh), [flimit] "r"(flimit), [limit] "r"(limit));
 
-    vp8_loop_filter_vertical_edge_mips(y_ptr + 4, y_stride, flimit_vec, limit_vec, thresh_vec, 16);
-    vp8_loop_filter_vertical_edge_mips(y_ptr + 8, y_stride, flimit_vec, limit_vec, thresh_vec, 16);
-    vp8_loop_filter_vertical_edge_mips(y_ptr + 12, y_stride, flimit_vec, limit_vec, thresh_vec, 16);
+  vp8_loop_filter_vertical_edge_mips(y_ptr + 4, y_stride, flimit_vec, limit_vec,
+                                     thresh_vec, 16);
+  vp8_loop_filter_vertical_edge_mips(y_ptr + 8, y_stride, flimit_vec, limit_vec,
+                                     thresh_vec, 16);
+  vp8_loop_filter_vertical_edge_mips(y_ptr + 12, y_stride, flimit_vec,
+                                     limit_vec, thresh_vec, 16);
 
-    if (u_ptr)
-        vp8_loop_filter_uvvertical_edge_mips(u_ptr + 4, uv_stride, flimit_vec, limit_vec, thresh_vec, 0);
+  if (u_ptr)
+    vp8_loop_filter_uvvertical_edge_mips(u_ptr + 4, uv_stride, flimit_vec,
+                                         limit_vec, thresh_vec, 0);
 
-    if (v_ptr)
-        vp8_loop_filter_uvvertical_edge_mips(v_ptr + 4, uv_stride, flimit_vec, limit_vec, thresh_vec, 0);
+  if (v_ptr)
+    vp8_loop_filter_uvvertical_edge_mips(v_ptr + 4, uv_stride, flimit_vec,
+                                         limit_vec, thresh_vec, 0);
 }
 
 #endif
--- a/vp8/common/mips/msa/bilinear_filter_msa.c
+++ b/vp8/common/mips/msa/bilinear_filter_msa.c
@@ -13,206 +13,209 @@
 #include "vp8/common/filter.h"
 #include "vp8/common/mips/msa/vp8_macros_msa.h"
 
-DECLARE_ALIGNED(16, static const int8_t, vp8_bilinear_filters_msa[7][2]) =
-{
-    { 112, 16 },
-    { 96, 32 },
-    { 80, 48 },
-    { 64, 64 },
-    { 48, 80 },
-    { 32, 96 },
-    { 16, 112 }
+DECLARE_ALIGNED(16, static const int8_t, vp8_bilinear_filters_msa[7][2]) = {
+  { 112, 16 }, { 96, 32 }, { 80, 48 }, { 64, 64 },
+  { 48, 80 },  { 32, 96 }, { 16, 112 }
 };
 
-static const uint8_t vp8_mc_filt_mask_arr[16 * 3] =
-{
-    /* 8 width cases */
-    0, 1, 1, 2, 2, 3, 3, 4, 4, 5, 5, 6, 6, 7, 7, 8,
-    /* 4 width cases */
-    0, 1, 1, 2, 2, 3, 3, 4, 16, 17, 17, 18, 18, 19, 19, 20,
-    /* 4 width cases */
-    8, 9, 9, 10, 10, 11, 11, 12, 24, 25, 25, 26, 26, 27, 27, 28
+static const uint8_t vp8_mc_filt_mask_arr[16 * 3] = {
+  /* 8 width cases */
+  0, 1, 1, 2, 2, 3, 3, 4, 4, 5, 5, 6, 6, 7, 7, 8,
+  /* 4 width cases */
+  0, 1, 1, 2, 2, 3, 3, 4, 16, 17, 17, 18, 18, 19, 19, 20,
+  /* 4 width cases */
+  8, 9, 9, 10, 10, 11, 11, 12, 24, 25, 25, 26, 26, 27, 27, 28
 };
 
 static void common_hz_2t_4x4_msa(uint8_t *RESTRICT src, int32_t src_stride,
                                  uint8_t *RESTRICT dst, int32_t dst_stride,
-                                 const int8_t *filter)
-{
-    v16i8 src0, src1, src2, src3, mask;
-    v16u8 filt0, vec0, vec1, res0, res1;
-    v8u16 vec2, vec3, filt;
+                                 const int8_t *filter) {
+  v16i8 src0, src1, src2, src3, mask;
+  v16u8 filt0, vec0, vec1, res0, res1;
+  v8u16 vec2, vec3, filt;
 
-    mask = LD_SB(&vp8_mc_filt_mask_arr[16]);
+  mask = LD_SB(&vp8_mc_filt_mask_arr[16]);
 
-    filt = LD_UH(filter);
-    filt0 = (v16u8)__msa_splati_h((v8i16)filt, 0);
+  filt = LD_UH(filter);
+  filt0 = (v16u8)__msa_splati_h((v8i16)filt, 0);
 
-    LD_SB4(src, src_stride, src0, src1, src2, src3);
-    VSHF_B2_UB(src0, src1, src2, src3, mask, mask, vec0, vec1);
-    DOTP_UB2_UH(vec0, vec1, filt0, filt0, vec2, vec3);
-    SRARI_H2_UH(vec2, vec3, VP8_FILTER_SHIFT);
-    PCKEV_B2_UB(vec2, vec2, vec3, vec3, res0, res1);
-    ST4x4_UB(res0, res1, 0, 1, 0, 1, dst, dst_stride);
+  LD_SB4(src, src_stride, src0, src1, src2, src3);
+  VSHF_B2_UB(src0, src1, src2, src3, mask, mask, vec0, vec1);
+  DOTP_UB2_UH(vec0, vec1, filt0, filt0, vec2, vec3);
+  SRARI_H2_UH(vec2, vec3, VP8_FILTER_SHIFT);
+  PCKEV_B2_UB(vec2, vec2, vec3, vec3, res0, res1);
+  ST4x4_UB(res0, res1, 0, 1, 0, 1, dst, dst_stride);
 }
 
 static void common_hz_2t_4x8_msa(uint8_t *RESTRICT src, int32_t src_stride,
                                  uint8_t *RESTRICT dst, int32_t dst_stride,
-                                 const int8_t *filter)
-{
-    v16u8 vec0, vec1, vec2, vec3, filt0;
-    v16i8 src0, src1, src2, src3, src4, src5, src6, src7, mask;
-    v16i8 res0, res1, res2, res3;
-    v8u16 vec4, vec5, vec6, vec7, filt;
+                                 const int8_t *filter) {
+  v16u8 vec0, vec1, vec2, vec3, filt0;
+  v16i8 src0, src1, src2, src3, src4, src5, src6, src7, mask;
+  v16i8 res0, res1, res2, res3;
+  v8u16 vec4, vec5, vec6, vec7, filt;
 
-    mask = LD_SB(&vp8_mc_filt_mask_arr[16]);
+  mask = LD_SB(&vp8_mc_filt_mask_arr[16]);
 
-    filt = LD_UH(filter);
-    filt0 = (v16u8)__msa_splati_h((v8i16)filt, 0);
+  filt = LD_UH(filter);
+  filt0 = (v16u8)__msa_splati_h((v8i16)filt, 0);
 
-    LD_SB8(src, src_stride, src0, src1, src2, src3, src4, src5, src6, src7);
-    VSHF_B2_UB(src0, src1, src2, src3, mask, mask, vec0, vec1);
-    VSHF_B2_UB(src4, src5, src6, src7, mask, mask, vec2, vec3);
-    DOTP_UB4_UH(vec0, vec1, vec2, vec3, filt0, filt0, filt0, filt0,
-                vec4, vec5, vec6, vec7);
-    SRARI_H4_UH(vec4, vec5, vec6, vec7, VP8_FILTER_SHIFT);
-    PCKEV_B4_SB(vec4, vec4, vec5, vec5, vec6, vec6, vec7, vec7,
-                res0, res1, res2, res3);
-    ST4x4_UB(res0, res1, 0, 1, 0, 1, dst, dst_stride);
-    dst += (4 * dst_stride);
-    ST4x4_UB(res2, res3, 0, 1, 0, 1, dst, dst_stride);
+  LD_SB8(src, src_stride, src0, src1, src2, src3, src4, src5, src6, src7);
+  VSHF_B2_UB(src0, src1, src2, src3, mask, mask, vec0, vec1);
+  VSHF_B2_UB(src4, src5, src6, src7, mask, mask, vec2, vec3);
+  DOTP_UB4_UH(vec0, vec1, vec2, vec3, filt0, filt0, filt0, filt0, vec4, vec5,
+              vec6, vec7);
+  SRARI_H4_UH(vec4, vec5, vec6, vec7, VP8_FILTER_SHIFT);
+  PCKEV_B4_SB(vec4, vec4, vec5, vec5, vec6, vec6, vec7, vec7, res0, res1, res2,
+              res3);
+  ST4x4_UB(res0, res1, 0, 1, 0, 1, dst, dst_stride);
+  dst += (4 * dst_stride);
+  ST4x4_UB(res2, res3, 0, 1, 0, 1, dst, dst_stride);
 }
 
 static void common_hz_2t_4w_msa(uint8_t *RESTRICT src, int32_t src_stride,
                                 uint8_t *RESTRICT dst, int32_t dst_stride,
-                                const int8_t *filter, int32_t height)
-{
-    if (4 == height)
-    {
-        common_hz_2t_4x4_msa(src, src_stride, dst, dst_stride, filter);
-    }
-    else if (8 == height)
-    {
-        common_hz_2t_4x8_msa(src, src_stride, dst, dst_stride, filter);
-    }
+                                const int8_t *filter, int32_t height) {
+  if (4 == height) {
+    common_hz_2t_4x4_msa(src, src_stride, dst, dst_stride, filter);
+  } else if (8 == height) {
+    common_hz_2t_4x8_msa(src, src_stride, dst, dst_stride, filter);
+  }
 }
 
 static void common_hz_2t_8x4_msa(uint8_t *RESTRICT src, int32_t src_stride,
                                  uint8_t *RESTRICT dst, int32_t dst_stride,
-                                 const int8_t *filter)
-{
-    v16u8 filt0;
-    v16i8 src0, src1, src2, src3, mask;
-    v8u16 vec0, vec1, vec2, vec3, filt;
+                                 const int8_t *filter) {
+  v16u8 filt0;
+  v16i8 src0, src1, src2, src3, mask;
+  v8u16 vec0, vec1, vec2, vec3, filt;
 
-    mask = LD_SB(&vp8_mc_filt_mask_arr[0]);
+  mask = LD_SB(&vp8_mc_filt_mask_arr[0]);
 
-    filt = LD_UH(filter);
-    filt0 = (v16u8)__msa_splati_h((v8i16)filt, 0);
+  filt = LD_UH(filter);
+  filt0 = (v16u8)__msa_splati_h((v8i16)filt, 0);
 
-    LD_SB4(src, src_stride, src0, src1, src2, src3);
-    VSHF_B2_UH(src0, src0, src1, src1, mask, mask, vec0, vec1);
-    VSHF_B2_UH(src2, src2, src3, src3, mask, mask, vec2, vec3);
-    DOTP_UB4_UH(vec0, vec1, vec2, vec3, filt0, filt0, filt0, filt0,
-                vec0, vec1, vec2, vec3);
-    SRARI_H4_UH(vec0, vec1, vec2, vec3, VP8_FILTER_SHIFT);
-    PCKEV_B2_SB(vec1, vec0, vec3, vec2, src0, src1);
-    ST8x4_UB(src0, src1, dst, dst_stride);
+  LD_SB4(src, src_stride, src0, src1, src2, src3);
+  VSHF_B2_UH(src0, src0, src1, src1, mask, mask, vec0, vec1);
+  VSHF_B2_UH(src2, src2, src3, src3, mask, mask, vec2, vec3);
+  DOTP_UB4_UH(vec0, vec1, vec2, vec3, filt0, filt0, filt0, filt0, vec0, vec1,
+              vec2, vec3);
+  SRARI_H4_UH(vec0, vec1, vec2, vec3, VP8_FILTER_SHIFT);
+  PCKEV_B2_SB(vec1, vec0, vec3, vec2, src0, src1);
+  ST8x4_UB(src0, src1, dst, dst_stride);
 }
 
 static void common_hz_2t_8x8mult_msa(uint8_t *RESTRICT src, int32_t src_stride,
                                      uint8_t *RESTRICT dst, int32_t dst_stride,
-                                     const int8_t *filter, int32_t height)
-{
-    v16u8 filt0;
-    v16i8 src0, src1, src2, src3, mask, out0, out1;
-    v8u16 vec0, vec1, vec2, vec3, filt;
+                                     const int8_t *filter, int32_t height) {
+  v16u8 filt0;
+  v16i8 src0, src1, src2, src3, mask, out0, out1;
+  v8u16 vec0, vec1, vec2, vec3, filt;
 
-    mask = LD_SB(&vp8_mc_filt_mask_arr[0]);
+  mask = LD_SB(&vp8_mc_filt_mask_arr[0]);
 
-    filt = LD_UH(filter);
-    filt0 = (v16u8)__msa_splati_h((v8i16)filt, 0);
+  filt = LD_UH(filter);
+  filt0 = (v16u8)__msa_splati_h((v8i16)filt, 0);
 
+  LD_SB4(src, src_stride, src0, src1, src2, src3);
+  src += (4 * src_stride);
+
+  VSHF_B2_UH(src0, src0, src1, src1, mask, mask, vec0, vec1);
+  VSHF_B2_UH(src2, src2, src3, src3, mask, mask, vec2, vec3);
+  DOTP_UB4_UH(vec0, vec1, vec2, vec3, filt0, filt0, filt0, filt0, vec0, vec1,
+              vec2, vec3);
+  SRARI_H4_UH(vec0, vec1, vec2, vec3, VP8_FILTER_SHIFT);
+
+  LD_SB4(src, src_stride, src0, src1, src2, src3);
+  src += (4 * src_stride);
+
+  PCKEV_B2_SB(vec1, vec0, vec3, vec2, out0, out1);
+  ST8x4_UB(out0, out1, dst, dst_stride);
+  dst += (4 * dst_stride);
+
+  VSHF_B2_UH(src0, src0, src1, src1, mask, mask, vec0, vec1);
+  VSHF_B2_UH(src2, src2, src3, src3, mask, mask, vec2, vec3);
+  DOTP_UB4_UH(vec0, vec1, vec2, vec3, filt0, filt0, filt0, filt0, vec0, vec1,
+              vec2, vec3);
+  SRARI_H4_UH(vec0, vec1, vec2, vec3, VP8_FILTER_SHIFT);
+  PCKEV_B2_SB(vec1, vec0, vec3, vec2, out0, out1);
+  ST8x4_UB(out0, out1, dst, dst_stride);
+  dst += (4 * dst_stride);
+
+  if (16 == height) {
     LD_SB4(src, src_stride, src0, src1, src2, src3);
     src += (4 * src_stride);
 
     VSHF_B2_UH(src0, src0, src1, src1, mask, mask, vec0, vec1);
     VSHF_B2_UH(src2, src2, src3, src3, mask, mask, vec2, vec3);
-    DOTP_UB4_UH(vec0, vec1, vec2, vec3, filt0, filt0, filt0, filt0,
-                vec0, vec1, vec2, vec3);
+    DOTP_UB4_UH(vec0, vec1, vec2, vec3, filt0, filt0, filt0, filt0, vec0, vec1,
+                vec2, vec3);
     SRARI_H4_UH(vec0, vec1, vec2, vec3, VP8_FILTER_SHIFT);
-
     LD_SB4(src, src_stride, src0, src1, src2, src3);
     src += (4 * src_stride);
 
     PCKEV_B2_SB(vec1, vec0, vec3, vec2, out0, out1);
     ST8x4_UB(out0, out1, dst, dst_stride);
-    dst += (4 * dst_stride);
 
     VSHF_B2_UH(src0, src0, src1, src1, mask, mask, vec0, vec1);
     VSHF_B2_UH(src2, src2, src3, src3, mask, mask, vec2, vec3);
-    DOTP_UB4_UH(vec0, vec1, vec2, vec3, filt0, filt0, filt0, filt0,
-                vec0, vec1, vec2, vec3);
+    DOTP_UB4_UH(vec0, vec1, vec2, vec3, filt0, filt0, filt0, filt0, vec0, vec1,
+                vec2, vec3);
     SRARI_H4_UH(vec0, vec1, vec2, vec3, VP8_FILTER_SHIFT);
     PCKEV_B2_SB(vec1, vec0, vec3, vec2, out0, out1);
-    ST8x4_UB(out0, out1, dst, dst_stride);
-    dst += (4 * dst_stride);
-
-    if (16 == height)
-    {
-        LD_SB4(src, src_stride, src0, src1, src2, src3);
-        src += (4 * src_stride);
-
-        VSHF_B2_UH(src0, src0, src1, src1, mask, mask, vec0, vec1);
-        VSHF_B2_UH(src2, src2, src3, src3, mask, mask, vec2, vec3);
-        DOTP_UB4_UH(vec0, vec1, vec2, vec3, filt0, filt0, filt0, filt0,
-                    vec0, vec1, vec2, vec3);
-        SRARI_H4_UH(vec0, vec1, vec2, vec3, VP8_FILTER_SHIFT);
-        LD_SB4(src, src_stride, src0, src1, src2, src3);
-        src += (4 * src_stride);
-
-        PCKEV_B2_SB(vec1, vec0, vec3, vec2, out0, out1);
-        ST8x4_UB(out0, out1, dst, dst_stride);
-
-        VSHF_B2_UH(src0, src0, src1, src1, mask, mask, vec0, vec1);
-        VSHF_B2_UH(src2, src2, src3, src3, mask, mask, vec2, vec3);
-        DOTP_UB4_UH(vec0, vec1, vec2, vec3, filt0, filt0, filt0, filt0,
-                    vec0, vec1, vec2, vec3);
-        SRARI_H4_UH(vec0, vec1, vec2, vec3, VP8_FILTER_SHIFT);
-        PCKEV_B2_SB(vec1, vec0, vec3, vec2, out0, out1);
-        ST8x4_UB(out0, out1, dst + 4 * dst_stride, dst_stride);
-    }
+    ST8x4_UB(out0, out1, dst + 4 * dst_stride, dst_stride);
+  }
 }
 
 static void common_hz_2t_8w_msa(uint8_t *RESTRICT src, int32_t src_stride,
                                 uint8_t *RESTRICT dst, int32_t dst_stride,
-                                const int8_t *filter, int32_t height)
-{
-    if (4 == height)
-    {
-        common_hz_2t_8x4_msa(src, src_stride, dst, dst_stride, filter);
-    }
-    else
-    {
-        common_hz_2t_8x8mult_msa(src, src_stride, dst, dst_stride, filter, height);
-    }
+                                const int8_t *filter, int32_t height) {
+  if (4 == height) {
+    common_hz_2t_8x4_msa(src, src_stride, dst, dst_stride, filter);
+  } else {
+    common_hz_2t_8x8mult_msa(src, src_stride, dst, dst_stride, filter, height);
+  }
 }
 
 static void common_hz_2t_16w_msa(uint8_t *RESTRICT src, int32_t src_stride,
                                  uint8_t *RESTRICT dst, int32_t dst_stride,
-                                 const int8_t *filter, int32_t height)
-{
-    uint32_t loop_cnt;
-    v16i8 src0, src1, src2, src3, src4, src5, src6, src7, mask;
-    v16u8 filt0, vec0, vec1, vec2, vec3, vec4, vec5, vec6, vec7;
-    v8u16 out0, out1, out2, out3, out4, out5, out6, out7, filt;
+                                 const int8_t *filter, int32_t height) {
+  uint32_t loop_cnt;
+  v16i8 src0, src1, src2, src3, src4, src5, src6, src7, mask;
+  v16u8 filt0, vec0, vec1, vec2, vec3, vec4, vec5, vec6, vec7;
+  v8u16 out0, out1, out2, out3, out4, out5, out6, out7, filt;
 
-    mask = LD_SB(&vp8_mc_filt_mask_arr[0]);
+  mask = LD_SB(&vp8_mc_filt_mask_arr[0]);
 
-    loop_cnt = (height >> 2) - 1;
+  loop_cnt = (height >> 2) - 1;
 
-    filt = LD_UH(filter);
-    filt0 = (v16u8)__msa_splati_h((v8i16)filt, 0);
+  filt = LD_UH(filter);
+  filt0 = (v16u8)__msa_splati_h((v8i16)filt, 0);
 
+  LD_SB4(src, src_stride, src0, src2, src4, src6);
+  LD_SB4(src + 8, src_stride, src1, src3, src5, src7);
+  src += (4 * src_stride);
+
+  VSHF_B2_UB(src0, src0, src1, src1, mask, mask, vec0, vec1);
+  VSHF_B2_UB(src2, src2, src3, src3, mask, mask, vec2, vec3);
+  VSHF_B2_UB(src4, src4, src5, src5, mask, mask, vec4, vec5);
+  VSHF_B2_UB(src6, src6, src7, src7, mask, mask, vec6, vec7);
+  DOTP_UB4_UH(vec0, vec1, vec2, vec3, filt0, filt0, filt0, filt0, out0, out1,
+              out2, out3);
+  DOTP_UB4_UH(vec4, vec5, vec6, vec7, filt0, filt0, filt0, filt0, out4, out5,
+              out6, out7);
+  SRARI_H4_UH(out0, out1, out2, out3, VP8_FILTER_SHIFT);
+  SRARI_H4_UH(out4, out5, out6, out7, VP8_FILTER_SHIFT);
+  PCKEV_ST_SB(out0, out1, dst);
+  dst += dst_stride;
+  PCKEV_ST_SB(out2, out3, dst);
+  dst += dst_stride;
+  PCKEV_ST_SB(out4, out5, dst);
+  dst += dst_stride;
+  PCKEV_ST_SB(out6, out7, dst);
+  dst += dst_stride;
+
+  for (; loop_cnt--;) {
     LD_SB4(src, src_stride, src0, src2, src4, src6);
     LD_SB4(src + 8, src_stride, src1, src3, src5, src7);
     src += (4 * src_stride);
@@ -221,10 +224,10 @@
     VSHF_B2_UB(src2, src2, src3, src3, mask, mask, vec2, vec3);
     VSHF_B2_UB(src4, src4, src5, src5, mask, mask, vec4, vec5);
     VSHF_B2_UB(src6, src6, src7, src7, mask, mask, vec6, vec7);
-    DOTP_UB4_UH(vec0, vec1, vec2, vec3, filt0, filt0, filt0, filt0,
-                out0, out1, out2, out3);
-    DOTP_UB4_UH(vec4, vec5, vec6, vec7, filt0, filt0, filt0, filt0,
-                out4, out5, out6, out7);
+    DOTP_UB4_UH(vec0, vec1, vec2, vec3, filt0, filt0, filt0, filt0, out0, out1,
+                out2, out3);
+    DOTP_UB4_UH(vec4, vec5, vec6, vec7, filt0, filt0, filt0, filt0, out4, out5,
+                out6, out7);
     SRARI_H4_UH(out0, out1, out2, out3, VP8_FILTER_SHIFT);
     SRARI_H4_UH(out4, out5, out6, out7, VP8_FILTER_SHIFT);
     PCKEV_ST_SB(out0, out1, dst);
@@ -235,308 +238,265 @@
     dst += dst_stride;
     PCKEV_ST_SB(out6, out7, dst);
     dst += dst_stride;
-
-    for (; loop_cnt--;)
-    {
-        LD_SB4(src, src_stride, src0, src2, src4, src6);
-        LD_SB4(src + 8, src_stride, src1, src3, src5, src7);
-        src += (4 * src_stride);
-
-        VSHF_B2_UB(src0, src0, src1, src1, mask, mask, vec0, vec1);
-        VSHF_B2_UB(src2, src2, src3, src3, mask, mask, vec2, vec3);
-        VSHF_B2_UB(src4, src4, src5, src5, mask, mask, vec4, vec5);
-        VSHF_B2_UB(src6, src6, src7, src7, mask, mask, vec6, vec7);
-        DOTP_UB4_UH(vec0, vec1, vec2, vec3, filt0, filt0, filt0, filt0,
-                    out0, out1, out2, out3);
-        DOTP_UB4_UH(vec4, vec5, vec6, vec7, filt0, filt0, filt0, filt0,
-                    out4, out5, out6, out7);
-        SRARI_H4_UH(out0, out1, out2, out3, VP8_FILTER_SHIFT);
-        SRARI_H4_UH(out4, out5, out6, out7, VP8_FILTER_SHIFT);
-        PCKEV_ST_SB(out0, out1, dst);
-        dst += dst_stride;
-        PCKEV_ST_SB(out2, out3, dst);
-        dst += dst_stride;
-        PCKEV_ST_SB(out4, out5, dst);
-        dst += dst_stride;
-        PCKEV_ST_SB(out6, out7, dst);
-        dst += dst_stride;
-    }
+  }
 }
 
 static void common_vt_2t_4x4_msa(uint8_t *RESTRICT src, int32_t src_stride,
                                  uint8_t *RESTRICT dst, int32_t dst_stride,
-                                 const int8_t *filter)
-{
-    v16i8 src0, src1, src2, src3, src4;
-    v16i8 src10_r, src32_r, src21_r, src43_r, src2110, src4332;
-    v16u8 filt0;
-    v8i16 filt;
-    v8u16 tmp0, tmp1;
+                                 const int8_t *filter) {
+  v16i8 src0, src1, src2, src3, src4;
+  v16i8 src10_r, src32_r, src21_r, src43_r, src2110, src4332;
+  v16u8 filt0;
+  v8i16 filt;
+  v8u16 tmp0, tmp1;
 
-    filt = LD_SH(filter);
-    filt0 = (v16u8)__msa_splati_h(filt, 0);
+  filt = LD_SH(filter);
+  filt0 = (v16u8)__msa_splati_h(filt, 0);
 
-    LD_SB5(src, src_stride, src0, src1, src2, src3, src4);
-    src += (5 * src_stride);
+  LD_SB5(src, src_stride, src0, src1, src2, src3, src4);
+  src += (5 * src_stride);
 
-    ILVR_B4_SB(src1, src0, src2, src1, src3, src2, src4, src3,
-               src10_r, src21_r, src32_r, src43_r);
-    ILVR_D2_SB(src21_r, src10_r, src43_r, src32_r, src2110, src4332);
-    DOTP_UB2_UH(src2110, src4332, filt0, filt0, tmp0, tmp1);
-    SRARI_H2_UH(tmp0, tmp1, VP8_FILTER_SHIFT);
-    src2110 = __msa_pckev_b((v16i8)tmp1, (v16i8)tmp0);
-    ST4x4_UB(src2110, src2110, 0, 1, 2, 3, dst, dst_stride);
+  ILVR_B4_SB(src1, src0, src2, src1, src3, src2, src4, src3, src10_r, src21_r,
+             src32_r, src43_r);
+  ILVR_D2_SB(src21_r, src10_r, src43_r, src32_r, src2110, src4332);
+  DOTP_UB2_UH(src2110, src4332, filt0, filt0, tmp0, tmp1);
+  SRARI_H2_UH(tmp0, tmp1, VP8_FILTER_SHIFT);
+  src2110 = __msa_pckev_b((v16i8)tmp1, (v16i8)tmp0);
+  ST4x4_UB(src2110, src2110, 0, 1, 2, 3, dst, dst_stride);
 }
 
 static void common_vt_2t_4x8_msa(uint8_t *RESTRICT src, int32_t src_stride,
                                  uint8_t *RESTRICT dst, int32_t dst_stride,
-                                 const int8_t *filter)
-{
-    v16i8 src0, src1, src2, src3, src4, src5, src6, src7, src8;
-    v16i8 src10_r, src32_r, src54_r, src76_r, src21_r, src43_r;
-    v16i8 src65_r, src87_r, src2110, src4332, src6554, src8776;
-    v8u16 tmp0, tmp1, tmp2, tmp3;
-    v16u8 filt0;
-    v8i16 filt;
+                                 const int8_t *filter) {
+  v16i8 src0, src1, src2, src3, src4, src5, src6, src7, src8;
+  v16i8 src10_r, src32_r, src54_r, src76_r, src21_r, src43_r;
+  v16i8 src65_r, src87_r, src2110, src4332, src6554, src8776;
+  v8u16 tmp0, tmp1, tmp2, tmp3;
+  v16u8 filt0;
+  v8i16 filt;
 
-    filt = LD_SH(filter);
-    filt0 = (v16u8)__msa_splati_h(filt, 0);
+  filt = LD_SH(filter);
+  filt0 = (v16u8)__msa_splati_h(filt, 0);
 
-    LD_SB8(src, src_stride, src0, src1, src2, src3, src4, src5, src6, src7);
-    src += (8 * src_stride);
+  LD_SB8(src, src_stride, src0, src1, src2, src3, src4, src5, src6, src7);
+  src += (8 * src_stride);
 
-    src8 = LD_SB(src);
-    src += src_stride;
+  src8 = LD_SB(src);
+  src += src_stride;
 
-    ILVR_B4_SB(src1, src0, src2, src1, src3, src2, src4, src3, src10_r, src21_r,
-               src32_r, src43_r);
-    ILVR_B4_SB(src5, src4, src6, src5, src7, src6, src8, src7, src54_r, src65_r,
-               src76_r, src87_r);
-    ILVR_D4_SB(src21_r, src10_r, src43_r, src32_r, src65_r, src54_r,
-               src87_r, src76_r, src2110, src4332, src6554, src8776);
-    DOTP_UB4_UH(src2110, src4332, src6554, src8776, filt0, filt0, filt0, filt0,
-                tmp0, tmp1, tmp2, tmp3);
-    SRARI_H4_UH(tmp0, tmp1, tmp2, tmp3, VP8_FILTER_SHIFT);
-    PCKEV_B2_SB(tmp1, tmp0, tmp3, tmp2, src2110, src4332);
-    ST4x4_UB(src2110, src2110, 0, 1, 2, 3, dst, dst_stride);
-    ST4x4_UB(src4332, src4332, 0, 1, 2, 3, dst + 4 * dst_stride, dst_stride);
+  ILVR_B4_SB(src1, src0, src2, src1, src3, src2, src4, src3, src10_r, src21_r,
+             src32_r, src43_r);
+  ILVR_B4_SB(src5, src4, src6, src5, src7, src6, src8, src7, src54_r, src65_r,
+             src76_r, src87_r);
+  ILVR_D4_SB(src21_r, src10_r, src43_r, src32_r, src65_r, src54_r, src87_r,
+             src76_r, src2110, src4332, src6554, src8776);
+  DOTP_UB4_UH(src2110, src4332, src6554, src8776, filt0, filt0, filt0, filt0,
+              tmp0, tmp1, tmp2, tmp3);
+  SRARI_H4_UH(tmp0, tmp1, tmp2, tmp3, VP8_FILTER_SHIFT);
+  PCKEV_B2_SB(tmp1, tmp0, tmp3, tmp2, src2110, src4332);
+  ST4x4_UB(src2110, src2110, 0, 1, 2, 3, dst, dst_stride);
+  ST4x4_UB(src4332, src4332, 0, 1, 2, 3, dst + 4 * dst_stride, dst_stride);
 }
 
 static void common_vt_2t_4w_msa(uint8_t *RESTRICT src, int32_t src_stride,
                                 uint8_t *RESTRICT dst, int32_t dst_stride,
-                                const int8_t *filter, int32_t height)
-{
-    if (4 == height)
-    {
-        common_vt_2t_4x4_msa(src, src_stride, dst, dst_stride, filter);
-    }
-    else if (8 == height)
-    {
-        common_vt_2t_4x8_msa(src, src_stride, dst, dst_stride, filter);
-    }
+                                const int8_t *filter, int32_t height) {
+  if (4 == height) {
+    common_vt_2t_4x4_msa(src, src_stride, dst, dst_stride, filter);
+  } else if (8 == height) {
+    common_vt_2t_4x8_msa(src, src_stride, dst, dst_stride, filter);
+  }
 }
 
 static void common_vt_2t_8x4_msa(uint8_t *RESTRICT src, int32_t src_stride,
                                  uint8_t *RESTRICT dst, int32_t dst_stride,
-                                 const int8_t *filter)
-{
-    v16u8 src0, src1, src2, src3, src4, vec0, vec1, vec2, vec3, filt0;
-    v16i8 out0, out1;
-    v8u16 tmp0, tmp1, tmp2, tmp3;
-    v8i16 filt;
+                                 const int8_t *filter) {
+  v16u8 src0, src1, src2, src3, src4, vec0, vec1, vec2, vec3, filt0;
+  v16i8 out0, out1;
+  v8u16 tmp0, tmp1, tmp2, tmp3;
+  v8i16 filt;
 
-    filt = LD_SH(filter);
-    filt0 = (v16u8)__msa_splati_h(filt, 0);
+  filt = LD_SH(filter);
+  filt0 = (v16u8)__msa_splati_h(filt, 0);
 
-    LD_UB5(src, src_stride, src0, src1, src2, src3, src4);
-    ILVR_B2_UB(src1, src0, src2, src1, vec0, vec1);
-    ILVR_B2_UB(src3, src2, src4, src3, vec2, vec3);
-    DOTP_UB4_UH(vec0, vec1, vec2, vec3, filt0, filt0, filt0, filt0,
-                tmp0, tmp1, tmp2, tmp3);
-    SRARI_H4_UH(tmp0, tmp1, tmp2, tmp3, VP8_FILTER_SHIFT);
-    PCKEV_B2_SB(tmp1, tmp0, tmp3, tmp2, out0, out1);
-    ST8x4_UB(out0, out1, dst, dst_stride);
+  LD_UB5(src, src_stride, src0, src1, src2, src3, src4);
+  ILVR_B2_UB(src1, src0, src2, src1, vec0, vec1);
+  ILVR_B2_UB(src3, src2, src4, src3, vec2, vec3);
+  DOTP_UB4_UH(vec0, vec1, vec2, vec3, filt0, filt0, filt0, filt0, tmp0, tmp1,
+              tmp2, tmp3);
+  SRARI_H4_UH(tmp0, tmp1, tmp2, tmp3, VP8_FILTER_SHIFT);
+  PCKEV_B2_SB(tmp1, tmp0, tmp3, tmp2, out0, out1);
+  ST8x4_UB(out0, out1, dst, dst_stride);
 }
 
 static void common_vt_2t_8x8mult_msa(uint8_t *RESTRICT src, int32_t src_stride,
                                      uint8_t *RESTRICT dst, int32_t dst_stride,
-                                     const int8_t *filter, int32_t height)
-{
-    uint32_t loop_cnt;
-    v16u8 src0, src1, src2, src3, src4, src5, src6, src7, src8;
-    v16u8 vec0, vec1, vec2, vec3, vec4, vec5, vec6, vec7, filt0;
-    v16i8 out0, out1;
-    v8u16 tmp0, tmp1, tmp2, tmp3;
-    v8i16 filt;
+                                     const int8_t *filter, int32_t height) {
+  uint32_t loop_cnt;
+  v16u8 src0, src1, src2, src3, src4, src5, src6, src7, src8;
+  v16u8 vec0, vec1, vec2, vec3, vec4, vec5, vec6, vec7, filt0;
+  v16i8 out0, out1;
+  v8u16 tmp0, tmp1, tmp2, tmp3;
+  v8i16 filt;
 
-    filt = LD_SH(filter);
-    filt0 = (v16u8)__msa_splati_h(filt, 0);
+  filt = LD_SH(filter);
+  filt0 = (v16u8)__msa_splati_h(filt, 0);
 
-    src0 = LD_UB(src);
-    src += src_stride;
+  src0 = LD_UB(src);
+  src += src_stride;
 
-    for (loop_cnt = (height >> 3); loop_cnt--;)
-    {
-        LD_UB8(src, src_stride, src1, src2, src3, src4, src5, src6, src7, src8);
-        src += (8 * src_stride);
+  for (loop_cnt = (height >> 3); loop_cnt--;) {
+    LD_UB8(src, src_stride, src1, src2, src3, src4, src5, src6, src7, src8);
+    src += (8 * src_stride);
 
-        ILVR_B4_UB(src1, src0, src2, src1, src3, src2, src4, src3,
-                   vec0, vec1, vec2, vec3);
-        ILVR_B4_UB(src5, src4, src6, src5, src7, src6, src8, src7,
-                   vec4, vec5, vec6, vec7);
-        DOTP_UB4_UH(vec0, vec1, vec2, vec3, filt0, filt0, filt0, filt0,
-                    tmp0, tmp1, tmp2, tmp3);
-        SRARI_H4_UH(tmp0, tmp1, tmp2, tmp3, VP8_FILTER_SHIFT);
-        PCKEV_B2_SB(tmp1, tmp0, tmp3, tmp2, out0, out1);
-        ST8x4_UB(out0, out1, dst, dst_stride);
-        dst += (4 * dst_stride);
+    ILVR_B4_UB(src1, src0, src2, src1, src3, src2, src4, src3, vec0, vec1, vec2,
+               vec3);
+    ILVR_B4_UB(src5, src4, src6, src5, src7, src6, src8, src7, vec4, vec5, vec6,
+               vec7);
+    DOTP_UB4_UH(vec0, vec1, vec2, vec3, filt0, filt0, filt0, filt0, tmp0, tmp1,
+                tmp2, tmp3);
+    SRARI_H4_UH(tmp0, tmp1, tmp2, tmp3, VP8_FILTER_SHIFT);
+    PCKEV_B2_SB(tmp1, tmp0, tmp3, tmp2, out0, out1);
+    ST8x4_UB(out0, out1, dst, dst_stride);
+    dst += (4 * dst_stride);
 
-        DOTP_UB4_UH(vec4, vec5, vec6, vec7, filt0, filt0, filt0, filt0,
-                    tmp0, tmp1, tmp2, tmp3);
-        SRARI_H4_UH(tmp0, tmp1, tmp2, tmp3, VP8_FILTER_SHIFT);
-        PCKEV_B2_SB(tmp1, tmp0, tmp3, tmp2, out0, out1);
-        ST8x4_UB(out0, out1, dst, dst_stride);
-        dst += (4 * dst_stride);
+    DOTP_UB4_UH(vec4, vec5, vec6, vec7, filt0, filt0, filt0, filt0, tmp0, tmp1,
+                tmp2, tmp3);
+    SRARI_H4_UH(tmp0, tmp1, tmp2, tmp3, VP8_FILTER_SHIFT);
+    PCKEV_B2_SB(tmp1, tmp0, tmp3, tmp2, out0, out1);
+    ST8x4_UB(out0, out1, dst, dst_stride);
+    dst += (4 * dst_stride);
 
-        src0 = src8;
-    }
+    src0 = src8;
+  }
 }
 
 static void common_vt_2t_8w_msa(uint8_t *RESTRICT src, int32_t src_stride,
                                 uint8_t *RESTRICT dst, int32_t dst_stride,
-                                const int8_t *filter, int32_t height)
-{
-    if (4 == height)
-    {
-        common_vt_2t_8x4_msa(src, src_stride, dst, dst_stride, filter);
-    }
-    else
-    {
-        common_vt_2t_8x8mult_msa(src, src_stride, dst, dst_stride, filter,
-                                 height);
-    }
+                                const int8_t *filter, int32_t height) {
+  if (4 == height) {
+    common_vt_2t_8x4_msa(src, src_stride, dst, dst_stride, filter);
+  } else {
+    common_vt_2t_8x8mult_msa(src, src_stride, dst, dst_stride, filter, height);
+  }
 }
 
 static void common_vt_2t_16w_msa(uint8_t *RESTRICT src, int32_t src_stride,
                                  uint8_t *RESTRICT dst, int32_t dst_stride,
-                                 const int8_t *filter, int32_t height)
-{
-    uint32_t loop_cnt;
-    v16u8 src0, src1, src2, src3, src4;
-    v16u8 vec0, vec1, vec2, vec3, vec4, vec5, vec6, vec7, filt0;
-    v8u16 tmp0, tmp1, tmp2, tmp3;
-    v8i16 filt;
+                                 const int8_t *filter, int32_t height) {
+  uint32_t loop_cnt;
+  v16u8 src0, src1, src2, src3, src4;
+  v16u8 vec0, vec1, vec2, vec3, vec4, vec5, vec6, vec7, filt0;
+  v8u16 tmp0, tmp1, tmp2, tmp3;
+  v8i16 filt;
 
-    filt = LD_SH(filter);
-    filt0 = (v16u8)__msa_splati_h(filt, 0);
+  filt = LD_SH(filter);
+  filt0 = (v16u8)__msa_splati_h(filt, 0);
 
-    src0 = LD_UB(src);
-    src += src_stride;
+  src0 = LD_UB(src);
+  src += src_stride;
 
-    for (loop_cnt = (height >> 2); loop_cnt--;)
-    {
-        LD_UB4(src, src_stride, src1, src2, src3, src4);
-        src += (4 * src_stride);
+  for (loop_cnt = (height >> 2); loop_cnt--;) {
+    LD_UB4(src, src_stride, src1, src2, src3, src4);
+    src += (4 * src_stride);
 
-        ILVR_B2_UB(src1, src0, src2, src1, vec0, vec2);
-        ILVL_B2_UB(src1, src0, src2, src1, vec1, vec3);
-        DOTP_UB2_UH(vec0, vec1, filt0, filt0, tmp0, tmp1);
-        SRARI_H2_UH(tmp0, tmp1, VP8_FILTER_SHIFT);
-        PCKEV_ST_SB(tmp0, tmp1, dst);
-        dst += dst_stride;
+    ILVR_B2_UB(src1, src0, src2, src1, vec0, vec2);
+    ILVL_B2_UB(src1, src0, src2, src1, vec1, vec3);
+    DOTP_UB2_UH(vec0, vec1, filt0, filt0, tmp0, tmp1);
+    SRARI_H2_UH(tmp0, tmp1, VP8_FILTER_SHIFT);
+    PCKEV_ST_SB(tmp0, tmp1, dst);
+    dst += dst_stride;
 
-        ILVR_B2_UB(src3, src2, src4, src3, vec4, vec6);
-        ILVL_B2_UB(src3, src2, src4, src3, vec5, vec7);
-        DOTP_UB2_UH(vec2, vec3, filt0, filt0, tmp2, tmp3);
-        SRARI_H2_UH(tmp2, tmp3, VP8_FILTER_SHIFT);
-        PCKEV_ST_SB(tmp2, tmp3, dst);
-        dst += dst_stride;
+    ILVR_B2_UB(src3, src2, src4, src3, vec4, vec6);
+    ILVL_B2_UB(src3, src2, src4, src3, vec5, vec7);
+    DOTP_UB2_UH(vec2, vec3, filt0, filt0, tmp2, tmp3);
+    SRARI_H2_UH(tmp2, tmp3, VP8_FILTER_SHIFT);
+    PCKEV_ST_SB(tmp2, tmp3, dst);
+    dst += dst_stride;
 
-        DOTP_UB2_UH(vec4, vec5, filt0, filt0, tmp0, tmp1);
-        SRARI_H2_UH(tmp0, tmp1, VP8_FILTER_SHIFT);
-        PCKEV_ST_SB(tmp0, tmp1, dst);
-        dst += dst_stride;
+    DOTP_UB2_UH(vec4, vec5, filt0, filt0, tmp0, tmp1);
+    SRARI_H2_UH(tmp0, tmp1, VP8_FILTER_SHIFT);
+    PCKEV_ST_SB(tmp0, tmp1, dst);
+    dst += dst_stride;
 
-        DOTP_UB2_UH(vec6, vec7, filt0, filt0, tmp2, tmp3);
-        SRARI_H2_UH(tmp2, tmp3, VP8_FILTER_SHIFT);
-        PCKEV_ST_SB(tmp2, tmp3, dst);
-        dst += dst_stride;
+    DOTP_UB2_UH(vec6, vec7, filt0, filt0, tmp2, tmp3);
+    SRARI_H2_UH(tmp2, tmp3, VP8_FILTER_SHIFT);
+    PCKEV_ST_SB(tmp2, tmp3, dst);
+    dst += dst_stride;
 
-        src0 = src4;
-    }
+    src0 = src4;
+  }
 }
 
 static void common_hv_2ht_2vt_4x4_msa(uint8_t *RESTRICT src, int32_t src_stride,
                                       uint8_t *RESTRICT dst, int32_t dst_stride,
                                       const int8_t *filter_horiz,
-                                      const int8_t *filter_vert)
-{
-    v16i8 src0, src1, src2, src3, src4, mask;
-    v16u8 filt_vt, filt_hz, vec0, vec1, res0, res1;
-    v8u16 hz_out0, hz_out1, hz_out2, hz_out3, hz_out4, filt, tmp0, tmp1;
+                                      const int8_t *filter_vert) {
+  v16i8 src0, src1, src2, src3, src4, mask;
+  v16u8 filt_vt, filt_hz, vec0, vec1, res0, res1;
+  v8u16 hz_out0, hz_out1, hz_out2, hz_out3, hz_out4, filt, tmp0, tmp1;
 
-    mask = LD_SB(&vp8_mc_filt_mask_arr[16]);
+  mask = LD_SB(&vp8_mc_filt_mask_arr[16]);
 
-    filt = LD_UH(filter_horiz);
-    filt_hz = (v16u8)__msa_splati_h((v8i16)filt, 0);
-    filt = LD_UH(filter_vert);
-    filt_vt = (v16u8)__msa_splati_h((v8i16)filt, 0);
+  filt = LD_UH(filter_horiz);
+  filt_hz = (v16u8)__msa_splati_h((v8i16)filt, 0);
+  filt = LD_UH(filter_vert);
+  filt_vt = (v16u8)__msa_splati_h((v8i16)filt, 0);
 
-    LD_SB5(src, src_stride, src0, src1, src2, src3, src4);
-    hz_out0 = HORIZ_2TAP_FILT_UH(src0, src1, mask, filt_hz, VP8_FILTER_SHIFT);
-    hz_out2 = HORIZ_2TAP_FILT_UH(src2, src3, mask, filt_hz, VP8_FILTER_SHIFT);
-    hz_out4 = HORIZ_2TAP_FILT_UH(src4, src4, mask, filt_hz, VP8_FILTER_SHIFT);
-    hz_out1 = (v8u16)__msa_sldi_b((v16i8)hz_out2, (v16i8)hz_out0, 8);
-    hz_out3 = (v8u16)__msa_pckod_d((v2i64)hz_out4, (v2i64)hz_out2);
+  LD_SB5(src, src_stride, src0, src1, src2, src3, src4);
+  hz_out0 = HORIZ_2TAP_FILT_UH(src0, src1, mask, filt_hz, VP8_FILTER_SHIFT);
+  hz_out2 = HORIZ_2TAP_FILT_UH(src2, src3, mask, filt_hz, VP8_FILTER_SHIFT);
+  hz_out4 = HORIZ_2TAP_FILT_UH(src4, src4, mask, filt_hz, VP8_FILTER_SHIFT);
+  hz_out1 = (v8u16)__msa_sldi_b((v16i8)hz_out2, (v16i8)hz_out0, 8);
+  hz_out3 = (v8u16)__msa_pckod_d((v2i64)hz_out4, (v2i64)hz_out2);
 
-    ILVEV_B2_UB(hz_out0, hz_out1, hz_out2, hz_out3, vec0, vec1);
-    DOTP_UB2_UH(vec0, vec1, filt_vt, filt_vt, tmp0, tmp1);
-    SRARI_H2_UH(tmp0, tmp1, VP8_FILTER_SHIFT);
-    PCKEV_B2_UB(tmp0, tmp0, tmp1, tmp1, res0, res1);
-    ST4x4_UB(res0, res1, 0, 1, 0, 1, dst, dst_stride);
+  ILVEV_B2_UB(hz_out0, hz_out1, hz_out2, hz_out3, vec0, vec1);
+  DOTP_UB2_UH(vec0, vec1, filt_vt, filt_vt, tmp0, tmp1);
+  SRARI_H2_UH(tmp0, tmp1, VP8_FILTER_SHIFT);
+  PCKEV_B2_UB(tmp0, tmp0, tmp1, tmp1, res0, res1);
+  ST4x4_UB(res0, res1, 0, 1, 0, 1, dst, dst_stride);
 }
 
 static void common_hv_2ht_2vt_4x8_msa(uint8_t *RESTRICT src, int32_t src_stride,
                                       uint8_t *RESTRICT dst, int32_t dst_stride,
                                       const int8_t *filter_horiz,
-                                      const int8_t *filter_vert)
-{
-    v16i8 src0, src1, src2, src3, src4, src5, src6, src7, src8, mask;
-    v16i8 res0, res1, res2, res3;
-    v16u8 filt_hz, filt_vt, vec0, vec1, vec2, vec3;
-    v8u16 hz_out0, hz_out1, hz_out2, hz_out3, hz_out4, hz_out5, hz_out6;
-    v8u16 hz_out7, hz_out8, vec4, vec5, vec6, vec7, filt;
+                                      const int8_t *filter_vert) {
+  v16i8 src0, src1, src2, src3, src4, src5, src6, src7, src8, mask;
+  v16i8 res0, res1, res2, res3;
+  v16u8 filt_hz, filt_vt, vec0, vec1, vec2, vec3;
+  v8u16 hz_out0, hz_out1, hz_out2, hz_out3, hz_out4, hz_out5, hz_out6;
+  v8u16 hz_out7, hz_out8, vec4, vec5, vec6, vec7, filt;
 
-    mask = LD_SB(&vp8_mc_filt_mask_arr[16]);
+  mask = LD_SB(&vp8_mc_filt_mask_arr[16]);
 
-    filt = LD_UH(filter_horiz);
-    filt_hz = (v16u8)__msa_splati_h((v8i16)filt, 0);
-    filt = LD_UH(filter_vert);
-    filt_vt = (v16u8)__msa_splati_h((v8i16)filt, 0);
+  filt = LD_UH(filter_horiz);
+  filt_hz = (v16u8)__msa_splati_h((v8i16)filt, 0);
+  filt = LD_UH(filter_vert);
+  filt_vt = (v16u8)__msa_splati_h((v8i16)filt, 0);
 
-    LD_SB8(src, src_stride, src0, src1, src2, src3, src4, src5, src6, src7);
-    src += (8 * src_stride);
-    src8 = LD_SB(src);
+  LD_SB8(src, src_stride, src0, src1, src2, src3, src4, src5, src6, src7);
+  src += (8 * src_stride);
+  src8 = LD_SB(src);
 
-    hz_out0 = HORIZ_2TAP_FILT_UH(src0, src1, mask, filt_hz, VP8_FILTER_SHIFT);
-    hz_out2 = HORIZ_2TAP_FILT_UH(src2, src3, mask, filt_hz, VP8_FILTER_SHIFT);
-    hz_out4 = HORIZ_2TAP_FILT_UH(src4, src5, mask, filt_hz, VP8_FILTER_SHIFT);
-    hz_out6 = HORIZ_2TAP_FILT_UH(src6, src7, mask, filt_hz, VP8_FILTER_SHIFT);
-    hz_out8 = HORIZ_2TAP_FILT_UH(src8, src8, mask, filt_hz, VP8_FILTER_SHIFT);
-    SLDI_B3_UH(hz_out2, hz_out4, hz_out6, hz_out0, hz_out2, hz_out4, hz_out1,
-               hz_out3, hz_out5, 8);
-    hz_out7 = (v8u16)__msa_pckod_d((v2i64)hz_out8, (v2i64)hz_out6);
+  hz_out0 = HORIZ_2TAP_FILT_UH(src0, src1, mask, filt_hz, VP8_FILTER_SHIFT);
+  hz_out2 = HORIZ_2TAP_FILT_UH(src2, src3, mask, filt_hz, VP8_FILTER_SHIFT);
+  hz_out4 = HORIZ_2TAP_FILT_UH(src4, src5, mask, filt_hz, VP8_FILTER_SHIFT);
+  hz_out6 = HORIZ_2TAP_FILT_UH(src6, src7, mask, filt_hz, VP8_FILTER_SHIFT);
+  hz_out8 = HORIZ_2TAP_FILT_UH(src8, src8, mask, filt_hz, VP8_FILTER_SHIFT);
+  SLDI_B3_UH(hz_out2, hz_out4, hz_out6, hz_out0, hz_out2, hz_out4, hz_out1,
+             hz_out3, hz_out5, 8);
+  hz_out7 = (v8u16)__msa_pckod_d((v2i64)hz_out8, (v2i64)hz_out6);
 
-    ILVEV_B2_UB(hz_out0, hz_out1, hz_out2, hz_out3, vec0, vec1);
-    ILVEV_B2_UB(hz_out4, hz_out5, hz_out6, hz_out7, vec2, vec3);
-    DOTP_UB4_UH(vec0, vec1, vec2, vec3, filt_vt, filt_vt, filt_vt, filt_vt,
-                vec4, vec5, vec6, vec7);
-    SRARI_H4_UH(vec4, vec5, vec6, vec7, VP8_FILTER_SHIFT);
-    PCKEV_B4_SB(vec4, vec4, vec5, vec5, vec6, vec6, vec7, vec7,
-                res0, res1, res2, res3);
-    ST4x4_UB(res0, res1, 0, 1, 0, 1, dst, dst_stride);
-    dst += (4 * dst_stride);
-    ST4x4_UB(res2, res3, 0, 1, 0, 1, dst, dst_stride);
+  ILVEV_B2_UB(hz_out0, hz_out1, hz_out2, hz_out3, vec0, vec1);
+  ILVEV_B2_UB(hz_out4, hz_out5, hz_out6, hz_out7, vec2, vec3);
+  DOTP_UB4_UH(vec0, vec1, vec2, vec3, filt_vt, filt_vt, filt_vt, filt_vt, vec4,
+              vec5, vec6, vec7);
+  SRARI_H4_UH(vec4, vec5, vec6, vec7, VP8_FILTER_SHIFT);
+  PCKEV_B4_SB(vec4, vec4, vec5, vec5, vec6, vec6, vec7, vec7, res0, res1, res2,
+              res3);
+  ST4x4_UB(res0, res1, 0, 1, 0, 1, dst, dst_stride);
+  dst += (4 * dst_stride);
+  ST4x4_UB(res2, res3, 0, 1, 0, 1, dst, dst_stride);
 }
 
 static void common_hv_2ht_2vt_4w_msa(uint8_t *RESTRICT src, int32_t src_stride,
@@ -543,146 +503,128 @@
                                      uint8_t *RESTRICT dst, int32_t dst_stride,
                                      const int8_t *filter_horiz,
                                      const int8_t *filter_vert,
-                                     int32_t height)
-{
-    if (4 == height)
-    {
-        common_hv_2ht_2vt_4x4_msa(src, src_stride, dst, dst_stride,
-                                  filter_horiz, filter_vert);
-    }
-    else if (8 == height)
-    {
-        common_hv_2ht_2vt_4x8_msa(src, src_stride, dst, dst_stride,
-                                  filter_horiz, filter_vert);
-    }
+                                     int32_t height) {
+  if (4 == height) {
+    common_hv_2ht_2vt_4x4_msa(src, src_stride, dst, dst_stride, filter_horiz,
+                              filter_vert);
+  } else if (8 == height) {
+    common_hv_2ht_2vt_4x8_msa(src, src_stride, dst, dst_stride, filter_horiz,
+                              filter_vert);
+  }
 }
 
 static void common_hv_2ht_2vt_8x4_msa(uint8_t *RESTRICT src, int32_t src_stride,
                                       uint8_t *RESTRICT dst, int32_t dst_stride,
                                       const int8_t *filter_horiz,
-                                      const int8_t *filter_vert)
-{
-    v16i8 src0, src1, src2, src3, src4, mask, out0, out1;
-    v16u8 filt_hz, filt_vt, vec0, vec1, vec2, vec3;
-    v8u16 hz_out0, hz_out1, tmp0, tmp1, tmp2, tmp3;
-    v8i16 filt;
+                                      const int8_t *filter_vert) {
+  v16i8 src0, src1, src2, src3, src4, mask, out0, out1;
+  v16u8 filt_hz, filt_vt, vec0, vec1, vec2, vec3;
+  v8u16 hz_out0, hz_out1, tmp0, tmp1, tmp2, tmp3;
+  v8i16 filt;
 
-    mask = LD_SB(&vp8_mc_filt_mask_arr[0]);
+  mask = LD_SB(&vp8_mc_filt_mask_arr[0]);
 
-    filt = LD_SH(filter_horiz);
-    filt_hz = (v16u8)__msa_splati_h(filt, 0);
-    filt = LD_SH(filter_vert);
-    filt_vt = (v16u8)__msa_splati_h(filt, 0);
+  filt = LD_SH(filter_horiz);
+  filt_hz = (v16u8)__msa_splati_h(filt, 0);
+  filt = LD_SH(filter_vert);
+  filt_vt = (v16u8)__msa_splati_h(filt, 0);
 
-    LD_SB5(src, src_stride, src0, src1, src2, src3, src4);
+  LD_SB5(src, src_stride, src0, src1, src2, src3, src4);
 
-    hz_out0 = HORIZ_2TAP_FILT_UH(src0, src0, mask, filt_hz, VP8_FILTER_SHIFT);
-    hz_out1 = HORIZ_2TAP_FILT_UH(src1, src1, mask, filt_hz, VP8_FILTER_SHIFT);
-    vec0 = (v16u8)__msa_ilvev_b((v16i8)hz_out1, (v16i8)hz_out0);
-    tmp0 = __msa_dotp_u_h(vec0, filt_vt);
+  hz_out0 = HORIZ_2TAP_FILT_UH(src0, src0, mask, filt_hz, VP8_FILTER_SHIFT);
+  hz_out1 = HORIZ_2TAP_FILT_UH(src1, src1, mask, filt_hz, VP8_FILTER_SHIFT);
+  vec0 = (v16u8)__msa_ilvev_b((v16i8)hz_out1, (v16i8)hz_out0);
+  tmp0 = __msa_dotp_u_h(vec0, filt_vt);
 
-    hz_out0 = HORIZ_2TAP_FILT_UH(src2, src2, mask, filt_hz, VP8_FILTER_SHIFT);
-    vec1 = (v16u8)__msa_ilvev_b((v16i8)hz_out0, (v16i8)hz_out1);
-    tmp1 = __msa_dotp_u_h(vec1, filt_vt);
+  hz_out0 = HORIZ_2TAP_FILT_UH(src2, src2, mask, filt_hz, VP8_FILTER_SHIFT);
+  vec1 = (v16u8)__msa_ilvev_b((v16i8)hz_out0, (v16i8)hz_out1);
+  tmp1 = __msa_dotp_u_h(vec1, filt_vt);
 
-    hz_out1 = HORIZ_2TAP_FILT_UH(src3, src3, mask, filt_hz, VP8_FILTER_SHIFT);
-    vec2 = (v16u8)__msa_ilvev_b((v16i8)hz_out1, (v16i8)hz_out0);
-    tmp2 = __msa_dotp_u_h(vec2, filt_vt);
+  hz_out1 = HORIZ_2TAP_FILT_UH(src3, src3, mask, filt_hz, VP8_FILTER_SHIFT);
+  vec2 = (v16u8)__msa_ilvev_b((v16i8)hz_out1, (v16i8)hz_out0);
+  tmp2 = __msa_dotp_u_h(vec2, filt_vt);
 
-    hz_out0 = HORIZ_2TAP_FILT_UH(src4, src4, mask, filt_hz, VP8_FILTER_SHIFT);
-    vec3 = (v16u8)__msa_ilvev_b((v16i8)hz_out0, (v16i8)hz_out1);
-    tmp3 = __msa_dotp_u_h(vec3, filt_vt);
+  hz_out0 = HORIZ_2TAP_FILT_UH(src4, src4, mask, filt_hz, VP8_FILTER_SHIFT);
+  vec3 = (v16u8)__msa_ilvev_b((v16i8)hz_out0, (v16i8)hz_out1);
+  tmp3 = __msa_dotp_u_h(vec3, filt_vt);
 
-    SRARI_H4_UH(tmp0, tmp1, tmp2, tmp3, VP8_FILTER_SHIFT);
-    PCKEV_B2_SB(tmp1, tmp0, tmp3, tmp2, out0, out1);
-    ST8x4_UB(out0, out1, dst, dst_stride);
+  SRARI_H4_UH(tmp0, tmp1, tmp2, tmp3, VP8_FILTER_SHIFT);
+  PCKEV_B2_SB(tmp1, tmp0, tmp3, tmp2, out0, out1);
+  ST8x4_UB(out0, out1, dst, dst_stride);
 }
 
-static void common_hv_2ht_2vt_8x8mult_msa(uint8_t *RESTRICT src,
-                                          int32_t src_stride,
-                                          uint8_t *RESTRICT dst,
-                                          int32_t dst_stride,
-                                          const int8_t *filter_horiz,
-                                          const int8_t *filter_vert,
-                                          int32_t height)
-{
-    uint32_t loop_cnt;
-    v16i8 src0, src1, src2, src3, src4, mask, out0, out1;
-    v16u8 filt_hz, filt_vt, vec0;
-    v8u16 hz_out0, hz_out1, tmp1, tmp2, tmp3, tmp4, tmp5, tmp6, tmp7, tmp8;
-    v8i16 filt;
+static void common_hv_2ht_2vt_8x8mult_msa(
+    uint8_t *RESTRICT src, int32_t src_stride, uint8_t *RESTRICT dst,
+    int32_t dst_stride, const int8_t *filter_horiz, const int8_t *filter_vert,
+    int32_t height) {
+  uint32_t loop_cnt;
+  v16i8 src0, src1, src2, src3, src4, mask, out0, out1;
+  v16u8 filt_hz, filt_vt, vec0;
+  v8u16 hz_out0, hz_out1, tmp1, tmp2, tmp3, tmp4, tmp5, tmp6, tmp7, tmp8;
+  v8i16 filt;
 
-    mask = LD_SB(&vp8_mc_filt_mask_arr[0]);
+  mask = LD_SB(&vp8_mc_filt_mask_arr[0]);
 
-    filt = LD_SH(filter_horiz);
-    filt_hz = (v16u8)__msa_splati_h(filt, 0);
-    filt = LD_SH(filter_vert);
-    filt_vt = (v16u8)__msa_splati_h(filt, 0);
+  filt = LD_SH(filter_horiz);
+  filt_hz = (v16u8)__msa_splati_h(filt, 0);
+  filt = LD_SH(filter_vert);
+  filt_vt = (v16u8)__msa_splati_h(filt, 0);
 
-    src0 = LD_SB(src);
-    src += src_stride;
+  src0 = LD_SB(src);
+  src += src_stride;
 
-    hz_out0 = HORIZ_2TAP_FILT_UH(src0, src0, mask, filt_hz, VP8_FILTER_SHIFT);
+  hz_out0 = HORIZ_2TAP_FILT_UH(src0, src0, mask, filt_hz, VP8_FILTER_SHIFT);
 
-    for (loop_cnt = (height >> 3); loop_cnt--;)
-    {
-        LD_SB4(src, src_stride, src1, src2, src3, src4);
-        src += (4 * src_stride);
+  for (loop_cnt = (height >> 3); loop_cnt--;) {
+    LD_SB4(src, src_stride, src1, src2, src3, src4);
+    src += (4 * src_stride);
 
-        hz_out1 = HORIZ_2TAP_FILT_UH(src1, src1, mask, filt_hz,
-                                     VP8_FILTER_SHIFT);
-        vec0 = (v16u8)__msa_ilvev_b((v16i8)hz_out1, (v16i8)hz_out0);
-        tmp1 = __msa_dotp_u_h(vec0, filt_vt);
+    hz_out1 = HORIZ_2TAP_FILT_UH(src1, src1, mask, filt_hz, VP8_FILTER_SHIFT);
+    vec0 = (v16u8)__msa_ilvev_b((v16i8)hz_out1, (v16i8)hz_out0);
+    tmp1 = __msa_dotp_u_h(vec0, filt_vt);
 
-        hz_out0 = HORIZ_2TAP_FILT_UH(src2, src2, mask, filt_hz,
-                                     VP8_FILTER_SHIFT);
-        vec0 = (v16u8)__msa_ilvev_b((v16i8)hz_out0, (v16i8)hz_out1);
-        tmp2 = __msa_dotp_u_h(vec0, filt_vt);
+    hz_out0 = HORIZ_2TAP_FILT_UH(src2, src2, mask, filt_hz, VP8_FILTER_SHIFT);
+    vec0 = (v16u8)__msa_ilvev_b((v16i8)hz_out0, (v16i8)hz_out1);
+    tmp2 = __msa_dotp_u_h(vec0, filt_vt);
 
-        SRARI_H2_UH(tmp1, tmp2, VP8_FILTER_SHIFT);
+    SRARI_H2_UH(tmp1, tmp2, VP8_FILTER_SHIFT);
 
-        hz_out1 = HORIZ_2TAP_FILT_UH(src3, src3, mask, filt_hz,
-                                     VP8_FILTER_SHIFT);
-        vec0 = (v16u8)__msa_ilvev_b((v16i8)hz_out1, (v16i8)hz_out0);
-        tmp3 = __msa_dotp_u_h(vec0, filt_vt);
+    hz_out1 = HORIZ_2TAP_FILT_UH(src3, src3, mask, filt_hz, VP8_FILTER_SHIFT);
+    vec0 = (v16u8)__msa_ilvev_b((v16i8)hz_out1, (v16i8)hz_out0);
+    tmp3 = __msa_dotp_u_h(vec0, filt_vt);
 
-        hz_out0 = HORIZ_2TAP_FILT_UH(src4, src4, mask, filt_hz,
-                                     VP8_FILTER_SHIFT);
-        LD_SB4(src, src_stride, src1, src2, src3, src4);
-        src += (4 * src_stride);
-        vec0 = (v16u8)__msa_ilvev_b((v16i8)hz_out0, (v16i8)hz_out1);
-        tmp4 = __msa_dotp_u_h(vec0, filt_vt);
+    hz_out0 = HORIZ_2TAP_FILT_UH(src4, src4, mask, filt_hz, VP8_FILTER_SHIFT);
+    LD_SB4(src, src_stride, src1, src2, src3, src4);
+    src += (4 * src_stride);
+    vec0 = (v16u8)__msa_ilvev_b((v16i8)hz_out0, (v16i8)hz_out1);
+    tmp4 = __msa_dotp_u_h(vec0, filt_vt);
 
-        SRARI_H2_UH(tmp3, tmp4, VP8_FILTER_SHIFT);
-        PCKEV_B2_SB(tmp2, tmp1, tmp4, tmp3, out0, out1);
-        ST8x4_UB(out0, out1, dst, dst_stride);
-        dst += (4 * dst_stride);
+    SRARI_H2_UH(tmp3, tmp4, VP8_FILTER_SHIFT);
+    PCKEV_B2_SB(tmp2, tmp1, tmp4, tmp3, out0, out1);
+    ST8x4_UB(out0, out1, dst, dst_stride);
+    dst += (4 * dst_stride);
 
-        hz_out1 = HORIZ_2TAP_FILT_UH(src1, src1, mask, filt_hz,
-                                     VP8_FILTER_SHIFT);
-        vec0 = (v16u8)__msa_ilvev_b((v16i8)hz_out1, (v16i8)hz_out0);
-        tmp5 = __msa_dotp_u_h(vec0, filt_vt);
+    hz_out1 = HORIZ_2TAP_FILT_UH(src1, src1, mask, filt_hz, VP8_FILTER_SHIFT);
+    vec0 = (v16u8)__msa_ilvev_b((v16i8)hz_out1, (v16i8)hz_out0);
+    tmp5 = __msa_dotp_u_h(vec0, filt_vt);
 
-        hz_out0 = HORIZ_2TAP_FILT_UH(src2, src2, mask, filt_hz,
-                                     VP8_FILTER_SHIFT);
-        vec0 = (v16u8)__msa_ilvev_b((v16i8)hz_out0, (v16i8)hz_out1);
-        tmp6 = __msa_dotp_u_h(vec0, filt_vt);
+    hz_out0 = HORIZ_2TAP_FILT_UH(src2, src2, mask, filt_hz, VP8_FILTER_SHIFT);
+    vec0 = (v16u8)__msa_ilvev_b((v16i8)hz_out0, (v16i8)hz_out1);
+    tmp6 = __msa_dotp_u_h(vec0, filt_vt);
 
-        hz_out1 = HORIZ_2TAP_FILT_UH(src3, src3, mask, filt_hz,
-                                     VP8_FILTER_SHIFT);
-        vec0 = (v16u8)__msa_ilvev_b((v16i8)hz_out1, (v16i8)hz_out0);
-        tmp7 = __msa_dotp_u_h(vec0, filt_vt);
+    hz_out1 = HORIZ_2TAP_FILT_UH(src3, src3, mask, filt_hz, VP8_FILTER_SHIFT);
+    vec0 = (v16u8)__msa_ilvev_b((v16i8)hz_out1, (v16i8)hz_out0);
+    tmp7 = __msa_dotp_u_h(vec0, filt_vt);
 
-        hz_out0 = HORIZ_2TAP_FILT_UH(src4, src4, mask, filt_hz,
-                                     VP8_FILTER_SHIFT);
-        vec0 = (v16u8)__msa_ilvev_b((v16i8)hz_out0, (v16i8)hz_out1);
-        tmp8 = __msa_dotp_u_h(vec0, filt_vt);
+    hz_out0 = HORIZ_2TAP_FILT_UH(src4, src4, mask, filt_hz, VP8_FILTER_SHIFT);
+    vec0 = (v16u8)__msa_ilvev_b((v16i8)hz_out0, (v16i8)hz_out1);
+    tmp8 = __msa_dotp_u_h(vec0, filt_vt);
 
-        SRARI_H4_UH(tmp5, tmp6, tmp7, tmp8, VP8_FILTER_SHIFT);
-        PCKEV_B2_SB(tmp6, tmp5, tmp8, tmp7, out0, out1);
-        ST8x4_UB(out0, out1, dst, dst_stride);
-        dst += (4 * dst_stride);
-    }
+    SRARI_H4_UH(tmp5, tmp6, tmp7, tmp8, VP8_FILTER_SHIFT);
+    PCKEV_B2_SB(tmp6, tmp5, tmp8, tmp7, out0, out1);
+    ST8x4_UB(out0, out1, dst, dst_stride);
+    dst += (4 * dst_stride);
+  }
 }
 
 static void common_hv_2ht_2vt_8w_msa(uint8_t *RESTRICT src, int32_t src_stride,
@@ -689,18 +631,14 @@
                                      uint8_t *RESTRICT dst, int32_t dst_stride,
                                      const int8_t *filter_horiz,
                                      const int8_t *filter_vert,
-                                     int32_t height)
-{
-    if (4 == height)
-    {
-        common_hv_2ht_2vt_8x4_msa(src, src_stride, dst, dst_stride,
-                                  filter_horiz, filter_vert);
-    }
-    else
-    {
-        common_hv_2ht_2vt_8x8mult_msa(src, src_stride, dst, dst_stride,
-                                      filter_horiz, filter_vert, height);
-    }
+                                     int32_t height) {
+  if (4 == height) {
+    common_hv_2ht_2vt_8x4_msa(src, src_stride, dst, dst_stride, filter_horiz,
+                              filter_vert);
+  } else {
+    common_hv_2ht_2vt_8x8mult_msa(src, src_stride, dst, dst_stride,
+                                  filter_horiz, filter_vert, height);
+  }
 }
 
 static void common_hv_2ht_2vt_16w_msa(uint8_t *RESTRICT src, int32_t src_stride,
@@ -707,205 +645,153 @@
                                       uint8_t *RESTRICT dst, int32_t dst_stride,
                                       const int8_t *filter_horiz,
                                       const int8_t *filter_vert,
-                                      int32_t height)
-{
-    uint32_t loop_cnt;
-    v16i8 src0, src1, src2, src3, src4, src5, src6, src7, mask;
-    v16u8 filt_hz, filt_vt, vec0, vec1;
-    v8u16 tmp1, tmp2, hz_out0, hz_out1, hz_out2, hz_out3;
-    v8i16 filt;
+                                      int32_t height) {
+  uint32_t loop_cnt;
+  v16i8 src0, src1, src2, src3, src4, src5, src6, src7, mask;
+  v16u8 filt_hz, filt_vt, vec0, vec1;
+  v8u16 tmp1, tmp2, hz_out0, hz_out1, hz_out2, hz_out3;
+  v8i16 filt;
 
-    mask = LD_SB(&vp8_mc_filt_mask_arr[0]);
+  mask = LD_SB(&vp8_mc_filt_mask_arr[0]);
 
-    /* rearranging filter */
-    filt = LD_SH(filter_horiz);
-    filt_hz = (v16u8)__msa_splati_h(filt, 0);
-    filt = LD_SH(filter_vert);
-    filt_vt = (v16u8)__msa_splati_h(filt, 0);
+  /* rearranging filter */
+  filt = LD_SH(filter_horiz);
+  filt_hz = (v16u8)__msa_splati_h(filt, 0);
+  filt = LD_SH(filter_vert);
+  filt_vt = (v16u8)__msa_splati_h(filt, 0);
 
-    LD_SB2(src, 8, src0, src1);
-    src += src_stride;
+  LD_SB2(src, 8, src0, src1);
+  src += src_stride;
 
-    hz_out0 = HORIZ_2TAP_FILT_UH(src0, src0, mask, filt_hz, VP8_FILTER_SHIFT);
-    hz_out2 = HORIZ_2TAP_FILT_UH(src1, src1, mask, filt_hz, VP8_FILTER_SHIFT);
+  hz_out0 = HORIZ_2TAP_FILT_UH(src0, src0, mask, filt_hz, VP8_FILTER_SHIFT);
+  hz_out2 = HORIZ_2TAP_FILT_UH(src1, src1, mask, filt_hz, VP8_FILTER_SHIFT);
 
-    for (loop_cnt = (height >> 2); loop_cnt--;)
-    {
-        LD_SB4(src, src_stride, src0, src2, src4, src6);
-        LD_SB4(src + 8, src_stride, src1, src3, src5, src7);
-        src += (4 * src_stride);
+  for (loop_cnt = (height >> 2); loop_cnt--;) {
+    LD_SB4(src, src_stride, src0, src2, src4, src6);
+    LD_SB4(src + 8, src_stride, src1, src3, src5, src7);
+    src += (4 * src_stride);
 
-        hz_out1 = HORIZ_2TAP_FILT_UH(src0, src0, mask, filt_hz,
-                                     VP8_FILTER_SHIFT);
-        hz_out3 = HORIZ_2TAP_FILT_UH(src1, src1, mask, filt_hz,
-                                     VP8_FILTER_SHIFT);
-        ILVEV_B2_UB(hz_out0, hz_out1, hz_out2, hz_out3, vec0, vec1);
-        DOTP_UB2_UH(vec0, vec1, filt_vt, filt_vt, tmp1, tmp2);
-        SRARI_H2_UH(tmp1, tmp2, VP8_FILTER_SHIFT);
-        PCKEV_ST_SB(tmp1, tmp2, dst);
-        dst += dst_stride;
+    hz_out1 = HORIZ_2TAP_FILT_UH(src0, src0, mask, filt_hz, VP8_FILTER_SHIFT);
+    hz_out3 = HORIZ_2TAP_FILT_UH(src1, src1, mask, filt_hz, VP8_FILTER_SHIFT);
+    ILVEV_B2_UB(hz_out0, hz_out1, hz_out2, hz_out3, vec0, vec1);
+    DOTP_UB2_UH(vec0, vec1, filt_vt, filt_vt, tmp1, tmp2);
+    SRARI_H2_UH(tmp1, tmp2, VP8_FILTER_SHIFT);
+    PCKEV_ST_SB(tmp1, tmp2, dst);
+    dst += dst_stride;
 
-        hz_out0 = HORIZ_2TAP_FILT_UH(src2, src2, mask, filt_hz,
-                                     VP8_FILTER_SHIFT);
-        hz_out2 = HORIZ_2TAP_FILT_UH(src3, src3, mask, filt_hz,
-                                     VP8_FILTER_SHIFT);
-        ILVEV_B2_UB(hz_out1, hz_out0, hz_out3, hz_out2, vec0, vec1);
-        DOTP_UB2_UH(vec0, vec1, filt_vt, filt_vt, tmp1, tmp2);
-        SRARI_H2_UH(tmp1, tmp2, VP8_FILTER_SHIFT);
-        PCKEV_ST_SB(tmp1, tmp2, dst);
-        dst += dst_stride;
+    hz_out0 = HORIZ_2TAP_FILT_UH(src2, src2, mask, filt_hz, VP8_FILTER_SHIFT);
+    hz_out2 = HORIZ_2TAP_FILT_UH(src3, src3, mask, filt_hz, VP8_FILTER_SHIFT);
+    ILVEV_B2_UB(hz_out1, hz_out0, hz_out3, hz_out2, vec0, vec1);
+    DOTP_UB2_UH(vec0, vec1, filt_vt, filt_vt, tmp1, tmp2);
+    SRARI_H2_UH(tmp1, tmp2, VP8_FILTER_SHIFT);
+    PCKEV_ST_SB(tmp1, tmp2, dst);
+    dst += dst_stride;
 
-        hz_out1 = HORIZ_2TAP_FILT_UH(src4, src4, mask, filt_hz,
-                                     VP8_FILTER_SHIFT);
-        hz_out3 = HORIZ_2TAP_FILT_UH(src5, src5, mask, filt_hz,
-                                     VP8_FILTER_SHIFT);
-        ILVEV_B2_UB(hz_out0, hz_out1, hz_out2, hz_out3, vec0, vec1);
-        DOTP_UB2_UH(vec0, vec1, filt_vt, filt_vt, tmp1, tmp2);
-        SRARI_H2_UH(tmp1, tmp2, VP8_FILTER_SHIFT);
-        PCKEV_ST_SB(tmp1, tmp2, dst);
-        dst += dst_stride;
+    hz_out1 = HORIZ_2TAP_FILT_UH(src4, src4, mask, filt_hz, VP8_FILTER_SHIFT);
+    hz_out3 = HORIZ_2TAP_FILT_UH(src5, src5, mask, filt_hz, VP8_FILTER_SHIFT);
+    ILVEV_B2_UB(hz_out0, hz_out1, hz_out2, hz_out3, vec0, vec1);
+    DOTP_UB2_UH(vec0, vec1, filt_vt, filt_vt, tmp1, tmp2);
+    SRARI_H2_UH(tmp1, tmp2, VP8_FILTER_SHIFT);
+    PCKEV_ST_SB(tmp1, tmp2, dst);
+    dst += dst_stride;
 
-        hz_out0 = HORIZ_2TAP_FILT_UH(src6, src6, mask, filt_hz,
-                                     VP8_FILTER_SHIFT);
-        hz_out2 = HORIZ_2TAP_FILT_UH(src7, src7, mask, filt_hz,
-                                     VP8_FILTER_SHIFT);
-        ILVEV_B2_UB(hz_out1, hz_out0, hz_out3, hz_out2, vec0, vec1);
-        DOTP_UB2_UH(vec0, vec1, filt_vt, filt_vt, tmp1, tmp2);
-        SRARI_H2_UH(tmp1, tmp2, VP8_FILTER_SHIFT);
-        PCKEV_ST_SB(tmp1, tmp2, dst);
-        dst += dst_stride;
-    }
+    hz_out0 = HORIZ_2TAP_FILT_UH(src6, src6, mask, filt_hz, VP8_FILTER_SHIFT);
+    hz_out2 = HORIZ_2TAP_FILT_UH(src7, src7, mask, filt_hz, VP8_FILTER_SHIFT);
+    ILVEV_B2_UB(hz_out1, hz_out0, hz_out3, hz_out2, vec0, vec1);
+    DOTP_UB2_UH(vec0, vec1, filt_vt, filt_vt, tmp1, tmp2);
+    SRARI_H2_UH(tmp1, tmp2, VP8_FILTER_SHIFT);
+    PCKEV_ST_SB(tmp1, tmp2, dst);
+    dst += dst_stride;
+  }
 }
 
 void vp8_bilinear_predict4x4_msa(uint8_t *RESTRICT src, int32_t src_stride,
                                  int32_t xoffset, int32_t yoffset,
-                                 uint8_t *RESTRICT dst, int32_t dst_stride)
-{
-    const int8_t *h_filter = vp8_bilinear_filters_msa[xoffset - 1];
-    const int8_t *v_filter = vp8_bilinear_filters_msa[yoffset - 1];
+                                 uint8_t *RESTRICT dst, int32_t dst_stride) {
+  const int8_t *h_filter = vp8_bilinear_filters_msa[xoffset - 1];
+  const int8_t *v_filter = vp8_bilinear_filters_msa[yoffset - 1];
 
-    if (yoffset)
-    {
-        if (xoffset)
-        {
-            common_hv_2ht_2vt_4w_msa(src, src_stride, dst, dst_stride,
-                                     h_filter, v_filter, 4);
-        }
-        else
-        {
-            common_vt_2t_4w_msa(src, src_stride, dst, dst_stride, v_filter, 4);
-        }
+  if (yoffset) {
+    if (xoffset) {
+      common_hv_2ht_2vt_4w_msa(src, src_stride, dst, dst_stride, h_filter,
+                               v_filter, 4);
+    } else {
+      common_vt_2t_4w_msa(src, src_stride, dst, dst_stride, v_filter, 4);
     }
-    else
-    {
-        if (xoffset)
-        {
-            common_hz_2t_4w_msa(src, src_stride, dst, dst_stride, h_filter, 4);
-        }
-        else
-        {
-            uint32_t tp0, tp1, tp2, tp3;
+  } else {
+    if (xoffset) {
+      common_hz_2t_4w_msa(src, src_stride, dst, dst_stride, h_filter, 4);
+    } else {
+      uint32_t tp0, tp1, tp2, tp3;
 
-            LW4(src, src_stride, tp0, tp1, tp2, tp3);
-            SW4(tp0, tp1, tp2, tp3, dst, dst_stride);
-        }
+      LW4(src, src_stride, tp0, tp1, tp2, tp3);
+      SW4(tp0, tp1, tp2, tp3, dst, dst_stride);
     }
+  }
 }
 
 void vp8_bilinear_predict8x4_msa(uint8_t *RESTRICT src, int32_t src_stride,
                                  int32_t xoffset, int32_t yoffset,
-                                 uint8_t *RESTRICT dst, int32_t dst_stride)
-{
-    const int8_t *h_filter = vp8_bilinear_filters_msa[xoffset - 1];
-    const int8_t *v_filter = vp8_bilinear_filters_msa[yoffset - 1];
+                                 uint8_t *RESTRICT dst, int32_t dst_stride) {
+  const int8_t *h_filter = vp8_bilinear_filters_msa[xoffset - 1];
+  const int8_t *v_filter = vp8_bilinear_filters_msa[yoffset - 1];
 
-    if (yoffset)
-    {
-        if (xoffset)
-        {
-            common_hv_2ht_2vt_8w_msa(src, src_stride, dst, dst_stride,
-                                     h_filter, v_filter, 4);
-        }
-        else
-        {
-            common_vt_2t_8w_msa(src, src_stride, dst, dst_stride, v_filter, 4);
-        }
+  if (yoffset) {
+    if (xoffset) {
+      common_hv_2ht_2vt_8w_msa(src, src_stride, dst, dst_stride, h_filter,
+                               v_filter, 4);
+    } else {
+      common_vt_2t_8w_msa(src, src_stride, dst, dst_stride, v_filter, 4);
     }
-    else
-    {
-        if (xoffset)
-        {
-            common_hz_2t_8w_msa(src, src_stride, dst, dst_stride, h_filter, 4);
-        }
-        else
-        {
-            vp8_copy_mem8x4(src, src_stride, dst, dst_stride);
-        }
+  } else {
+    if (xoffset) {
+      common_hz_2t_8w_msa(src, src_stride, dst, dst_stride, h_filter, 4);
+    } else {
+      vp8_copy_mem8x4(src, src_stride, dst, dst_stride);
     }
+  }
 }
 
 void vp8_bilinear_predict8x8_msa(uint8_t *RESTRICT src, int32_t src_stride,
                                  int32_t xoffset, int32_t yoffset,
-                                 uint8_t *RESTRICT dst, int32_t dst_stride)
-{
-    const int8_t *h_filter = vp8_bilinear_filters_msa[xoffset - 1];
-    const int8_t *v_filter = vp8_bilinear_filters_msa[yoffset - 1];
+                                 uint8_t *RESTRICT dst, int32_t dst_stride) {
+  const int8_t *h_filter = vp8_bilinear_filters_msa[xoffset - 1];
+  const int8_t *v_filter = vp8_bilinear_filters_msa[yoffset - 1];
 
-    if (yoffset)
-    {
-        if (xoffset)
-        {
-            common_hv_2ht_2vt_8w_msa(src, src_stride, dst, dst_stride,
-                                     h_filter, v_filter, 8);
-        }
-        else
-        {
-            common_vt_2t_8w_msa(src, src_stride, dst, dst_stride, v_filter, 8);
-        }
+  if (yoffset) {
+    if (xoffset) {
+      common_hv_2ht_2vt_8w_msa(src, src_stride, dst, dst_stride, h_filter,
+                               v_filter, 8);
+    } else {
+      common_vt_2t_8w_msa(src, src_stride, dst, dst_stride, v_filter, 8);
     }
-    else
-    {
-        if (xoffset)
-        {
-            common_hz_2t_8w_msa(src, src_stride, dst, dst_stride, h_filter, 8);
-        }
-        else
-        {
-            vp8_copy_mem8x8(src, src_stride, dst, dst_stride);
-        }
+  } else {
+    if (xoffset) {
+      common_hz_2t_8w_msa(src, src_stride, dst, dst_stride, h_filter, 8);
+    } else {
+      vp8_copy_mem8x8(src, src_stride, dst, dst_stride);
     }
+  }
 }
 
 void vp8_bilinear_predict16x16_msa(uint8_t *RESTRICT src, int32_t src_stride,
                                    int32_t xoffset, int32_t yoffset,
-                                   uint8_t *RESTRICT dst, int32_t dst_stride)
-{
-    const int8_t *h_filter = vp8_bilinear_filters_msa[xoffset - 1];
-    const int8_t *v_filter = vp8_bilinear_filters_msa[yoffset - 1];
+                                   uint8_t *RESTRICT dst, int32_t dst_stride) {
+  const int8_t *h_filter = vp8_bilinear_filters_msa[xoffset - 1];
+  const int8_t *v_filter = vp8_bilinear_filters_msa[yoffset - 1];
 
-    if (yoffset)
-    {
-        if (xoffset)
-        {
-            common_hv_2ht_2vt_16w_msa(src, src_stride, dst, dst_stride,
-                                      h_filter, v_filter, 16);
-        }
-        else
-        {
-            common_vt_2t_16w_msa(src, src_stride, dst, dst_stride, v_filter,
-                                 16);
-        }
+  if (yoffset) {
+    if (xoffset) {
+      common_hv_2ht_2vt_16w_msa(src, src_stride, dst, dst_stride, h_filter,
+                                v_filter, 16);
+    } else {
+      common_vt_2t_16w_msa(src, src_stride, dst, dst_stride, v_filter, 16);
     }
-    else
-    {
-        if (xoffset)
-        {
-            common_hz_2t_16w_msa(src, src_stride, dst, dst_stride, h_filter,
-                                 16);
-        }
-        else
-        {
-            vp8_copy_mem16x16(src, src_stride, dst, dst_stride);
-        }
+  } else {
+    if (xoffset) {
+      common_hz_2t_16w_msa(src, src_stride, dst, dst_stride, h_filter, 16);
+    } else {
+      vp8_copy_mem16x16(src, src_stride, dst, dst_stride);
     }
+  }
 }
--- a/vp8/common/mips/msa/copymem_msa.c
+++ b/vp8/common/mips/msa/copymem_msa.c
@@ -11,60 +11,52 @@
 #include "./vp8_rtcd.h"
 #include "vp8/common/mips/msa/vp8_macros_msa.h"
 
-static void copy_8x4_msa(uint8_t *src, int32_t src_stride,
-                         uint8_t *dst, int32_t dst_stride)
-{
-    uint64_t src0, src1, src2, src3;
+static void copy_8x4_msa(uint8_t *src, int32_t src_stride, uint8_t *dst,
+                         int32_t dst_stride) {
+  uint64_t src0, src1, src2, src3;
 
-    LD4(src, src_stride, src0, src1, src2, src3);
-    SD4(src0, src1, src2, src3, dst, dst_stride);
+  LD4(src, src_stride, src0, src1, src2, src3);
+  SD4(src0, src1, src2, src3, dst, dst_stride);
 }
 
-static void copy_8x8_msa(uint8_t *src, int32_t src_stride,
-                         uint8_t *dst, int32_t dst_stride)
-{
-    uint64_t src0, src1, src2, src3;
+static void copy_8x8_msa(uint8_t *src, int32_t src_stride, uint8_t *dst,
+                         int32_t dst_stride) {
+  uint64_t src0, src1, src2, src3;
 
-    LD4(src, src_stride, src0, src1, src2, src3);
-    src += (4 * src_stride);
-    SD4(src0, src1, src2, src3, dst, dst_stride);
-    dst += (4 * dst_stride);
+  LD4(src, src_stride, src0, src1, src2, src3);
+  src += (4 * src_stride);
+  SD4(src0, src1, src2, src3, dst, dst_stride);
+  dst += (4 * dst_stride);
 
-    LD4(src, src_stride, src0, src1, src2, src3);
-    SD4(src0, src1, src2, src3, dst, dst_stride);
+  LD4(src, src_stride, src0, src1, src2, src3);
+  SD4(src0, src1, src2, src3, dst, dst_stride);
 }
 
-static void copy_16x16_msa(uint8_t *src, int32_t src_stride,
-                           uint8_t *dst, int32_t dst_stride)
-{
-    v16u8 src0, src1, src2, src3, src4, src5, src6, src7;
-    v16u8 src8, src9, src10, src11, src12, src13, src14, src15;
+static void copy_16x16_msa(uint8_t *src, int32_t src_stride, uint8_t *dst,
+                           int32_t dst_stride) {
+  v16u8 src0, src1, src2, src3, src4, src5, src6, src7;
+  v16u8 src8, src9, src10, src11, src12, src13, src14, src15;
 
-    LD_UB8(src, src_stride, src0, src1, src2, src3, src4, src5, src6, src7);
-    src += (8 * src_stride);
-    LD_UB8(src, src_stride, src8, src9, src10, src11, src12, src13, src14,
-           src15);
+  LD_UB8(src, src_stride, src0, src1, src2, src3, src4, src5, src6, src7);
+  src += (8 * src_stride);
+  LD_UB8(src, src_stride, src8, src9, src10, src11, src12, src13, src14, src15);
 
-    ST_UB8(src0, src1, src2, src3, src4, src5, src6, src7, dst, dst_stride);
-    dst += (8 * dst_stride);
-    ST_UB8(src8, src9, src10, src11, src12, src13, src14, src15, dst,
-           dst_stride);
+  ST_UB8(src0, src1, src2, src3, src4, src5, src6, src7, dst, dst_stride);
+  dst += (8 * dst_stride);
+  ST_UB8(src8, src9, src10, src11, src12, src13, src14, src15, dst, dst_stride);
 }
 
-void vp8_copy_mem16x16_msa(uint8_t *src, int32_t src_stride,
-                           uint8_t *dst, int32_t dst_stride)
-{
-    copy_16x16_msa(src, src_stride, dst, dst_stride);
+void vp8_copy_mem16x16_msa(uint8_t *src, int32_t src_stride, uint8_t *dst,
+                           int32_t dst_stride) {
+  copy_16x16_msa(src, src_stride, dst, dst_stride);
 }
 
-void vp8_copy_mem8x8_msa(uint8_t *src, int32_t src_stride,
-                         uint8_t *dst, int32_t dst_stride)
-{
-    copy_8x8_msa(src, src_stride, dst, dst_stride);
+void vp8_copy_mem8x8_msa(uint8_t *src, int32_t src_stride, uint8_t *dst,
+                         int32_t dst_stride) {
+  copy_8x8_msa(src, src_stride, dst, dst_stride);
 }
 
-void vp8_copy_mem8x4_msa(uint8_t *src, int32_t src_stride,
-                         uint8_t *dst, int32_t dst_stride)
-{
-    copy_8x4_msa(src, src_stride, dst, dst_stride);
+void vp8_copy_mem8x4_msa(uint8_t *src, int32_t src_stride, uint8_t *dst,
+                         int32_t dst_stride) {
+  copy_8x4_msa(src, src_stride, dst, dst_stride);
 }
--- a/vp8/common/mips/msa/idct_msa.c
+++ b/vp8/common/mips/msa/idct_msa.c
@@ -15,443 +15,401 @@
 static const int32_t cospi8sqrt2minus1 = 20091;
 static const int32_t sinpi8sqrt2 = 35468;
 
-#define TRANSPOSE_TWO_4x4_H(in0, in1, in2, in3, out0, out1, out2, out3)  \
-{                                                                        \
-    v8i16 s4_m, s5_m, s6_m, s7_m;                                        \
-                                                                         \
-    TRANSPOSE8X4_SH_SH(in0, in1, in2, in3, s4_m, s5_m, s6_m, s7_m);      \
-    ILVR_D2_SH(s6_m, s4_m, s7_m, s5_m, out0, out2);                      \
-    out1 = (v8i16)__msa_ilvl_d((v2i64)s6_m, (v2i64)s4_m);                \
-    out3 = (v8i16)__msa_ilvl_d((v2i64)s7_m, (v2i64)s5_m);                \
-}
+#define TRANSPOSE_TWO_4x4_H(in0, in1, in2, in3, out0, out1, out2, out3) \
+  {                                                                     \
+    v8i16 s4_m, s5_m, s6_m, s7_m;                                       \
+                                                                        \
+    TRANSPOSE8X4_SH_SH(in0, in1, in2, in3, s4_m, s5_m, s6_m, s7_m);     \
+    ILVR_D2_SH(s6_m, s4_m, s7_m, s5_m, out0, out2);                     \
+    out1 = (v8i16)__msa_ilvl_d((v2i64)s6_m, (v2i64)s4_m);               \
+    out3 = (v8i16)__msa_ilvl_d((v2i64)s7_m, (v2i64)s5_m);               \
+  }
+
+#define EXPAND_TO_H_MULTIPLY_SINPI8SQRT2_PCK_TO_W(in)    \
+  ({                                                     \
+    v8i16 out_m;                                         \
+    v8i16 zero_m = { 0 };                                \
+    v4i32 tmp1_m, tmp2_m;                                \
+    v4i32 sinpi8_sqrt2_m = __msa_fill_w(sinpi8sqrt2);    \
+                                                         \
+    ILVRL_H2_SW(in, zero_m, tmp1_m, tmp2_m);             \
+    tmp1_m >>= 16;                                       \
+    tmp2_m >>= 16;                                       \
+    tmp1_m = (tmp1_m * sinpi8_sqrt2_m) >> 16;            \
+    tmp2_m = (tmp2_m * sinpi8_sqrt2_m) >> 16;            \
+    out_m = __msa_pckev_h((v8i16)tmp2_m, (v8i16)tmp1_m); \
+                                                         \
+    out_m;                                               \
+  })
 
-#define EXPAND_TO_H_MULTIPLY_SINPI8SQRT2_PCK_TO_W(in)     \
-({                                                        \
-    v8i16 out_m;                                          \
-    v8i16 zero_m = { 0 };                                 \
-    v4i32 tmp1_m, tmp2_m;                                 \
-    v4i32 sinpi8_sqrt2_m = __msa_fill_w(sinpi8sqrt2);     \
-                                                          \
-    ILVRL_H2_SW(in, zero_m, tmp1_m, tmp2_m);              \
-    tmp1_m >>= 16;                                        \
-    tmp2_m >>= 16;                                        \
-    tmp1_m = (tmp1_m * sinpi8_sqrt2_m) >> 16;             \
-    tmp2_m = (tmp2_m * sinpi8_sqrt2_m) >> 16;             \
-    out_m = __msa_pckev_h((v8i16)tmp2_m, (v8i16)tmp1_m);  \
-                                                          \
-    out_m;                                                \
-})
+#define VP8_IDCT_1D_H(in0, in1, in2, in3, out0, out1, out2, out3) \
+  {                                                               \
+    v8i16 a1_m, b1_m, c1_m, d1_m;                                 \
+    v8i16 c_tmp1_m, c_tmp2_m, d_tmp1_m, d_tmp2_m;                 \
+    v8i16 const_cospi8sqrt2minus1_m;                              \
+                                                                  \
+    const_cospi8sqrt2minus1_m = __msa_fill_h(cospi8sqrt2minus1);  \
+    a1_m = in0 + in2;                                             \
+    b1_m = in0 - in2;                                             \
+    c_tmp1_m = EXPAND_TO_H_MULTIPLY_SINPI8SQRT2_PCK_TO_W(in1);    \
+    c_tmp2_m = __msa_mul_q_h(in3, const_cospi8sqrt2minus1_m);     \
+    c_tmp2_m = c_tmp2_m >> 1;                                     \
+    c_tmp2_m = in3 + c_tmp2_m;                                    \
+    c1_m = c_tmp1_m - c_tmp2_m;                                   \
+    d_tmp1_m = __msa_mul_q_h(in1, const_cospi8sqrt2minus1_m);     \
+    d_tmp1_m = d_tmp1_m >> 1;                                     \
+    d_tmp1_m = in1 + d_tmp1_m;                                    \
+    d_tmp2_m = EXPAND_TO_H_MULTIPLY_SINPI8SQRT2_PCK_TO_W(in3);    \
+    d1_m = d_tmp1_m + d_tmp2_m;                                   \
+    BUTTERFLY_4(a1_m, b1_m, c1_m, d1_m, out0, out1, out2, out3);  \
+  }
 
-#define VP8_IDCT_1D_H(in0, in1, in2, in3, out0, out1, out2, out3)  \
-{                                                                  \
-    v8i16 a1_m, b1_m, c1_m, d1_m;                                  \
-    v8i16 c_tmp1_m, c_tmp2_m, d_tmp1_m, d_tmp2_m;                  \
-    v8i16 const_cospi8sqrt2minus1_m;                               \
-                                                                   \
-    const_cospi8sqrt2minus1_m = __msa_fill_h(cospi8sqrt2minus1);   \
-    a1_m = in0 + in2;                                              \
-    b1_m = in0 - in2;                                              \
-    c_tmp1_m = EXPAND_TO_H_MULTIPLY_SINPI8SQRT2_PCK_TO_W(in1);     \
-    c_tmp2_m = __msa_mul_q_h(in3, const_cospi8sqrt2minus1_m);      \
-    c_tmp2_m = c_tmp2_m >> 1;                                      \
-    c_tmp2_m = in3 + c_tmp2_m;                                     \
-    c1_m = c_tmp1_m - c_tmp2_m;                                    \
-    d_tmp1_m = __msa_mul_q_h(in1, const_cospi8sqrt2minus1_m);      \
-    d_tmp1_m = d_tmp1_m >> 1;                                      \
-    d_tmp1_m = in1 + d_tmp1_m;                                     \
-    d_tmp2_m = EXPAND_TO_H_MULTIPLY_SINPI8SQRT2_PCK_TO_W(in3);     \
-    d1_m = d_tmp1_m + d_tmp2_m;                                    \
-    BUTTERFLY_4(a1_m, b1_m, c1_m, d1_m, out0, out1, out2, out3);   \
-}
+#define VP8_IDCT_1D_W(in0, in1, in2, in3, out0, out1, out2, out3) \
+  {                                                               \
+    v4i32 a1_m, b1_m, c1_m, d1_m;                                 \
+    v4i32 c_tmp1_m, c_tmp2_m, d_tmp1_m, d_tmp2_m;                 \
+    v4i32 const_cospi8sqrt2minus1_m, sinpi8_sqrt2_m;              \
+                                                                  \
+    const_cospi8sqrt2minus1_m = __msa_fill_w(cospi8sqrt2minus1);  \
+    sinpi8_sqrt2_m = __msa_fill_w(sinpi8sqrt2);                   \
+    a1_m = in0 + in2;                                             \
+    b1_m = in0 - in2;                                             \
+    c_tmp1_m = (in1 * sinpi8_sqrt2_m) >> 16;                      \
+    c_tmp2_m = in3 + ((in3 * const_cospi8sqrt2minus1_m) >> 16);   \
+    c1_m = c_tmp1_m - c_tmp2_m;                                   \
+    d_tmp1_m = in1 + ((in1 * const_cospi8sqrt2minus1_m) >> 16);   \
+    d_tmp2_m = (in3 * sinpi8_sqrt2_m) >> 16;                      \
+    d1_m = d_tmp1_m + d_tmp2_m;                                   \
+    BUTTERFLY_4(a1_m, b1_m, c1_m, d1_m, out0, out1, out2, out3);  \
+  }
 
-#define VP8_IDCT_1D_W(in0, in1, in2, in3, out0, out1, out2, out3)  \
-{                                                                  \
-    v4i32 a1_m, b1_m, c1_m, d1_m;                                  \
-    v4i32 c_tmp1_m, c_tmp2_m, d_tmp1_m, d_tmp2_m;                  \
-    v4i32 const_cospi8sqrt2minus1_m, sinpi8_sqrt2_m;               \
-                                                                   \
-    const_cospi8sqrt2minus1_m = __msa_fill_w(cospi8sqrt2minus1);   \
-    sinpi8_sqrt2_m = __msa_fill_w(sinpi8sqrt2);                    \
-    a1_m = in0 + in2;                                              \
-    b1_m = in0 - in2;                                              \
-    c_tmp1_m = (in1 * sinpi8_sqrt2_m) >> 16;                       \
-    c_tmp2_m = in3 + ((in3 * const_cospi8sqrt2minus1_m) >> 16);    \
-    c1_m = c_tmp1_m - c_tmp2_m;                                    \
-    d_tmp1_m = in1 + ((in1 * const_cospi8sqrt2minus1_m) >> 16);    \
-    d_tmp2_m = (in3 * sinpi8_sqrt2_m) >> 16;                       \
-    d1_m = d_tmp1_m + d_tmp2_m;                                    \
-    BUTTERFLY_4(a1_m, b1_m, c1_m, d1_m, out0, out1, out2, out3);   \
-}
-
 static void idct4x4_addblk_msa(int16_t *input, uint8_t *pred,
-                               int32_t pred_stride,
-                               uint8_t *dest, int32_t dest_stride)
-{
-    v8i16 input0, input1;
-    v4i32 in0, in1, in2, in3, hz0, hz1, hz2, hz3, vt0, vt1, vt2, vt3;
-    v4i32 res0, res1, res2, res3;
-    v16i8 zero = { 0 };
-    v16i8 pred0, pred1, pred2, pred3, dest0, dest1, dest2, dest3;
-    v16i8 mask = { 0, 4, 8, 12, 20, 21, 22, 23, 24,
-                   25, 26, 27, 28, 29, 30, 31 };
+                               int32_t pred_stride, uint8_t *dest,
+                               int32_t dest_stride) {
+  v8i16 input0, input1;
+  v4i32 in0, in1, in2, in3, hz0, hz1, hz2, hz3, vt0, vt1, vt2, vt3;
+  v4i32 res0, res1, res2, res3;
+  v16i8 zero = { 0 };
+  v16i8 pred0, pred1, pred2, pred3, dest0, dest1, dest2, dest3;
+  v16i8 mask = { 0, 4, 8, 12, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31 };
 
-    LD_SH2(input, 8, input0, input1);
-    UNPCK_SH_SW(input0, in0, in1);
-    UNPCK_SH_SW(input1, in2, in3);
-    VP8_IDCT_1D_W(in0, in1, in2, in3, hz0, hz1, hz2, hz3);
-    TRANSPOSE4x4_SW_SW(hz0, hz1, hz2, hz3, hz0, hz1, hz2, hz3);
-    VP8_IDCT_1D_W(hz0, hz1, hz2, hz3, vt0, vt1, vt2, vt3);
-    SRARI_W4_SW(vt0, vt1, vt2, vt3, 3);
-    TRANSPOSE4x4_SW_SW(vt0, vt1, vt2, vt3, vt0, vt1, vt2, vt3);
-    LD_SB4(pred, pred_stride, pred0, pred1, pred2, pred3);
-    ILVR_B4_SW(zero, pred0, zero, pred1, zero, pred2, zero, pred3, res0, res1,
-               res2, res3);
-    ILVR_H4_SW(zero, res0, zero, res1, zero, res2, zero, res3, res0, res1,
-               res2, res3);
-    ADD4(res0, vt0, res1, vt1, res2, vt2, res3, vt3, res0, res1, res2, res3);
-    res0 = CLIP_SW_0_255(res0);
-    res1 = CLIP_SW_0_255(res1);
-    res2 = CLIP_SW_0_255(res2);
-    res3 = CLIP_SW_0_255(res3);
-    LD_SB4(dest, dest_stride, dest0, dest1, dest2, dest3);
-    VSHF_B2_SB(res0, dest0, res1, dest1, mask, mask, dest0, dest1);
-    VSHF_B2_SB(res2, dest2, res3, dest3, mask, mask, dest2, dest3);
-    ST_SB4(dest0, dest1, dest2, dest3, dest, dest_stride);
+  LD_SH2(input, 8, input0, input1);
+  UNPCK_SH_SW(input0, in0, in1);
+  UNPCK_SH_SW(input1, in2, in3);
+  VP8_IDCT_1D_W(in0, in1, in2, in3, hz0, hz1, hz2, hz3);
+  TRANSPOSE4x4_SW_SW(hz0, hz1, hz2, hz3, hz0, hz1, hz2, hz3);
+  VP8_IDCT_1D_W(hz0, hz1, hz2, hz3, vt0, vt1, vt2, vt3);
+  SRARI_W4_SW(vt0, vt1, vt2, vt3, 3);
+  TRANSPOSE4x4_SW_SW(vt0, vt1, vt2, vt3, vt0, vt1, vt2, vt3);
+  LD_SB4(pred, pred_stride, pred0, pred1, pred2, pred3);
+  ILVR_B4_SW(zero, pred0, zero, pred1, zero, pred2, zero, pred3, res0, res1,
+             res2, res3);
+  ILVR_H4_SW(zero, res0, zero, res1, zero, res2, zero, res3, res0, res1, res2,
+             res3);
+  ADD4(res0, vt0, res1, vt1, res2, vt2, res3, vt3, res0, res1, res2, res3);
+  res0 = CLIP_SW_0_255(res0);
+  res1 = CLIP_SW_0_255(res1);
+  res2 = CLIP_SW_0_255(res2);
+  res3 = CLIP_SW_0_255(res3);
+  LD_SB4(dest, dest_stride, dest0, dest1, dest2, dest3);
+  VSHF_B2_SB(res0, dest0, res1, dest1, mask, mask, dest0, dest1);
+  VSHF_B2_SB(res2, dest2, res3, dest3, mask, mask, dest2, dest3);
+  ST_SB4(dest0, dest1, dest2, dest3, dest, dest_stride);
 }
 
 static void idct4x4_addconst_msa(int16_t in_dc, uint8_t *pred,
-                                 int32_t pred_stride,
-                                 uint8_t *dest, int32_t dest_stride)
-{
-    v8i16 vec;
-    v8i16 res0, res1, res2, res3;
-    v16i8 zero = { 0 };
-    v16i8 pred0, pred1, pred2, pred3, dest0, dest1, dest2, dest3;
-    v16i8 mask = { 0, 2, 4, 6, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31 };
+                                 int32_t pred_stride, uint8_t *dest,
+                                 int32_t dest_stride) {
+  v8i16 vec;
+  v8i16 res0, res1, res2, res3;
+  v16i8 zero = { 0 };
+  v16i8 pred0, pred1, pred2, pred3, dest0, dest1, dest2, dest3;
+  v16i8 mask = { 0, 2, 4, 6, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31 };
 
-    vec = __msa_fill_h(in_dc);
-    vec = __msa_srari_h(vec, 3);
-    LD_SB4(pred, pred_stride, pred0, pred1, pred2, pred3);
-    ILVR_B4_SH(zero, pred0, zero, pred1, zero, pred2, zero, pred3, res0, res1,
-               res2, res3);
-    ADD4(res0, vec, res1, vec, res2, vec, res3, vec, res0, res1, res2, res3);
-    CLIP_SH4_0_255(res0, res1, res2, res3);
-    LD_SB4(dest, dest_stride, dest0, dest1, dest2, dest3);
-    VSHF_B2_SB(res0, dest0, res1, dest1, mask, mask, dest0, dest1);
-    VSHF_B2_SB(res2, dest2, res3, dest3, mask, mask, dest2, dest3);
-    ST_SB4(dest0, dest1, dest2, dest3, dest, dest_stride);
+  vec = __msa_fill_h(in_dc);
+  vec = __msa_srari_h(vec, 3);
+  LD_SB4(pred, pred_stride, pred0, pred1, pred2, pred3);
+  ILVR_B4_SH(zero, pred0, zero, pred1, zero, pred2, zero, pred3, res0, res1,
+             res2, res3);
+  ADD4(res0, vec, res1, vec, res2, vec, res3, vec, res0, res1, res2, res3);
+  CLIP_SH4_0_255(res0, res1, res2, res3);
+  LD_SB4(dest, dest_stride, dest0, dest1, dest2, dest3);
+  VSHF_B2_SB(res0, dest0, res1, dest1, mask, mask, dest0, dest1);
+  VSHF_B2_SB(res2, dest2, res3, dest3, mask, mask, dest2, dest3);
+  ST_SB4(dest0, dest1, dest2, dest3, dest, dest_stride);
 }
 
-void vp8_short_inv_walsh4x4_msa(int16_t *input, int16_t *mb_dq_coeff)
-{
-    v8i16 input0, input1;
-    v4i32 in0, in1, in2, in3, a1, b1, c1, d1;
-    v4i32 hz0, hz1, hz2, hz3, vt0, vt1, vt2, vt3;
+void vp8_short_inv_walsh4x4_msa(int16_t *input, int16_t *mb_dq_coeff) {
+  v8i16 input0, input1;
+  v4i32 in0, in1, in2, in3, a1, b1, c1, d1;
+  v4i32 hz0, hz1, hz2, hz3, vt0, vt1, vt2, vt3;
 
-    LD_SH2(input, 8, input0, input1);
-    UNPCK_SH_SW(input0, in0, in1);
-    UNPCK_SH_SW(input1, in2, in3);
-    BUTTERFLY_4(in0, in1, in2, in3, a1, b1, c1, d1);
-    BUTTERFLY_4(a1, d1, c1, b1, hz0, hz1, hz3, hz2);
-    TRANSPOSE4x4_SW_SW(hz0, hz1, hz2, hz3, hz0, hz1, hz2, hz3);
-    BUTTERFLY_4(hz0, hz1, hz2, hz3, a1, b1, c1, d1);
-    BUTTERFLY_4(a1, d1, c1, b1, vt0, vt1, vt3, vt2);
-    ADD4(vt0, 3, vt1, 3, vt2, 3, vt3, 3, vt0, vt1, vt2, vt3);
-    SRA_4V(vt0, vt1, vt2, vt3, 3);
-    mb_dq_coeff[0] = __msa_copy_s_h((v8i16)vt0, 0);
-    mb_dq_coeff[16] = __msa_copy_s_h((v8i16)vt1, 0);
-    mb_dq_coeff[32] = __msa_copy_s_h((v8i16)vt2, 0);
-    mb_dq_coeff[48] = __msa_copy_s_h((v8i16)vt3, 0);
-    mb_dq_coeff[64] = __msa_copy_s_h((v8i16)vt0, 2);
-    mb_dq_coeff[80] = __msa_copy_s_h((v8i16)vt1, 2);
-    mb_dq_coeff[96] = __msa_copy_s_h((v8i16)vt2, 2);
-    mb_dq_coeff[112] = __msa_copy_s_h((v8i16)vt3, 2);
-    mb_dq_coeff[128] = __msa_copy_s_h((v8i16)vt0, 4);
-    mb_dq_coeff[144] = __msa_copy_s_h((v8i16)vt1, 4);
-    mb_dq_coeff[160] = __msa_copy_s_h((v8i16)vt2, 4);
-    mb_dq_coeff[176] = __msa_copy_s_h((v8i16)vt3, 4);
-    mb_dq_coeff[192] = __msa_copy_s_h((v8i16)vt0, 6);
-    mb_dq_coeff[208] = __msa_copy_s_h((v8i16)vt1, 6);
-    mb_dq_coeff[224] = __msa_copy_s_h((v8i16)vt2, 6);
-    mb_dq_coeff[240] = __msa_copy_s_h((v8i16)vt3, 6);
+  LD_SH2(input, 8, input0, input1);
+  UNPCK_SH_SW(input0, in0, in1);
+  UNPCK_SH_SW(input1, in2, in3);
+  BUTTERFLY_4(in0, in1, in2, in3, a1, b1, c1, d1);
+  BUTTERFLY_4(a1, d1, c1, b1, hz0, hz1, hz3, hz2);
+  TRANSPOSE4x4_SW_SW(hz0, hz1, hz2, hz3, hz0, hz1, hz2, hz3);
+  BUTTERFLY_4(hz0, hz1, hz2, hz3, a1, b1, c1, d1);
+  BUTTERFLY_4(a1, d1, c1, b1, vt0, vt1, vt3, vt2);
+  ADD4(vt0, 3, vt1, 3, vt2, 3, vt3, 3, vt0, vt1, vt2, vt3);
+  SRA_4V(vt0, vt1, vt2, vt3, 3);
+  mb_dq_coeff[0] = __msa_copy_s_h((v8i16)vt0, 0);
+  mb_dq_coeff[16] = __msa_copy_s_h((v8i16)vt1, 0);
+  mb_dq_coeff[32] = __msa_copy_s_h((v8i16)vt2, 0);
+  mb_dq_coeff[48] = __msa_copy_s_h((v8i16)vt3, 0);
+  mb_dq_coeff[64] = __msa_copy_s_h((v8i16)vt0, 2);
+  mb_dq_coeff[80] = __msa_copy_s_h((v8i16)vt1, 2);
+  mb_dq_coeff[96] = __msa_copy_s_h((v8i16)vt2, 2);
+  mb_dq_coeff[112] = __msa_copy_s_h((v8i16)vt3, 2);
+  mb_dq_coeff[128] = __msa_copy_s_h((v8i16)vt0, 4);
+  mb_dq_coeff[144] = __msa_copy_s_h((v8i16)vt1, 4);
+  mb_dq_coeff[160] = __msa_copy_s_h((v8i16)vt2, 4);
+  mb_dq_coeff[176] = __msa_copy_s_h((v8i16)vt3, 4);
+  mb_dq_coeff[192] = __msa_copy_s_h((v8i16)vt0, 6);
+  mb_dq_coeff[208] = __msa_copy_s_h((v8i16)vt1, 6);
+  mb_dq_coeff[224] = __msa_copy_s_h((v8i16)vt2, 6);
+  mb_dq_coeff[240] = __msa_copy_s_h((v8i16)vt3, 6);
 }
 
 static void dequant_idct4x4_addblk_msa(int16_t *input, int16_t *dequant_input,
-                                       uint8_t *dest, int32_t dest_stride)
-{
-    v8i16 input0, input1, dequant_in0, dequant_in1, mul0, mul1;
-    v8i16 in0, in1, in2, in3;
-    v8i16 hz0_h, hz1_h, hz2_h, hz3_h;
-    v16i8 dest0, dest1, dest2, dest3;
-    v4i32 hz0_w, hz1_w, hz2_w, hz3_w;
-    v4i32 vt0, vt1, vt2, vt3, res0, res1, res2, res3;
-    v2i64 zero = { 0 };
-    v16i8 mask = { 0, 4, 8, 12, 20, 21, 22, 23, 24,
-                   25, 26, 27, 28, 29, 30, 31 };
+                                       uint8_t *dest, int32_t dest_stride) {
+  v8i16 input0, input1, dequant_in0, dequant_in1, mul0, mul1;
+  v8i16 in0, in1, in2, in3;
+  v8i16 hz0_h, hz1_h, hz2_h, hz3_h;
+  v16i8 dest0, dest1, dest2, dest3;
+  v4i32 hz0_w, hz1_w, hz2_w, hz3_w;
+  v4i32 vt0, vt1, vt2, vt3, res0, res1, res2, res3;
+  v2i64 zero = { 0 };
+  v16i8 mask = { 0, 4, 8, 12, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31 };
 
-    LD_SH2(input, 8, input0, input1);
-    LD_SH2(dequant_input, 8, dequant_in0, dequant_in1);
-    MUL2(input0, dequant_in0, input1, dequant_in1, mul0, mul1);
-    PCKEV_D2_SH(zero, mul0, zero, mul1, in0, in2);
-    PCKOD_D2_SH(zero, mul0, zero, mul1, in1, in3);
-    VP8_IDCT_1D_H(in0, in1, in2, in3, hz0_h, hz1_h, hz2_h, hz3_h);
-    PCKEV_D2_SH(hz1_h, hz0_h, hz3_h, hz2_h, mul0, mul1);
-    UNPCK_SH_SW(mul0, hz0_w, hz1_w);
-    UNPCK_SH_SW(mul1, hz2_w, hz3_w);
-    TRANSPOSE4x4_SW_SW(hz0_w, hz1_w, hz2_w, hz3_w, hz0_w, hz1_w, hz2_w, hz3_w);
-    VP8_IDCT_1D_W(hz0_w, hz1_w, hz2_w, hz3_w, vt0, vt1, vt2, vt3);
-    SRARI_W4_SW(vt0, vt1, vt2, vt3, 3);
-    TRANSPOSE4x4_SW_SW(vt0, vt1, vt2, vt3, vt0, vt1, vt2, vt3);
-    LD_SB4(dest, dest_stride, dest0, dest1, dest2, dest3);
-    ILVR_B4_SW(zero, dest0, zero, dest1, zero, dest2, zero, dest3, res0, res1,
-               res2, res3);
-    ILVR_H4_SW(zero, res0, zero, res1, zero, res2, zero, res3, res0, res1,
-               res2, res3);
-    ADD4(res0, vt0, res1, vt1, res2, vt2, res3, vt3, res0, res1, res2, res3);
-    res0 = CLIP_SW_0_255(res0);
-    res1 = CLIP_SW_0_255(res1);
-    res2 = CLIP_SW_0_255(res2);
-    res3 = CLIP_SW_0_255(res3);
-    VSHF_B2_SB(res0, dest0, res1, dest1, mask, mask, dest0, dest1);
-    VSHF_B2_SB(res2, dest2, res3, dest3, mask, mask, dest2, dest3);
-    ST_SB4(dest0, dest1, dest2, dest3, dest, dest_stride);
+  LD_SH2(input, 8, input0, input1);
+  LD_SH2(dequant_input, 8, dequant_in0, dequant_in1);
+  MUL2(input0, dequant_in0, input1, dequant_in1, mul0, mul1);
+  PCKEV_D2_SH(zero, mul0, zero, mul1, in0, in2);
+  PCKOD_D2_SH(zero, mul0, zero, mul1, in1, in3);
+  VP8_IDCT_1D_H(in0, in1, in2, in3, hz0_h, hz1_h, hz2_h, hz3_h);
+  PCKEV_D2_SH(hz1_h, hz0_h, hz3_h, hz2_h, mul0, mul1);
+  UNPCK_SH_SW(mul0, hz0_w, hz1_w);
+  UNPCK_SH_SW(mul1, hz2_w, hz3_w);
+  TRANSPOSE4x4_SW_SW(hz0_w, hz1_w, hz2_w, hz3_w, hz0_w, hz1_w, hz2_w, hz3_w);
+  VP8_IDCT_1D_W(hz0_w, hz1_w, hz2_w, hz3_w, vt0, vt1, vt2, vt3);
+  SRARI_W4_SW(vt0, vt1, vt2, vt3, 3);
+  TRANSPOSE4x4_SW_SW(vt0, vt1, vt2, vt3, vt0, vt1, vt2, vt3);
+  LD_SB4(dest, dest_stride, dest0, dest1, dest2, dest3);
+  ILVR_B4_SW(zero, dest0, zero, dest1, zero, dest2, zero, dest3, res0, res1,
+             res2, res3);
+  ILVR_H4_SW(zero, res0, zero, res1, zero, res2, zero, res3, res0, res1, res2,
+             res3);
+  ADD4(res0, vt0, res1, vt1, res2, vt2, res3, vt3, res0, res1, res2, res3);
+  res0 = CLIP_SW_0_255(res0);
+  res1 = CLIP_SW_0_255(res1);
+  res2 = CLIP_SW_0_255(res2);
+  res3 = CLIP_SW_0_255(res3);
+  VSHF_B2_SB(res0, dest0, res1, dest1, mask, mask, dest0, dest1);
+  VSHF_B2_SB(res2, dest2, res3, dest3, mask, mask, dest2, dest3);
+  ST_SB4(dest0, dest1, dest2, dest3, dest, dest_stride);
 }
 
 static void dequant_idct4x4_addblk_2x_msa(int16_t *input,
-                                          int16_t *dequant_input,
-                                          uint8_t *dest, int32_t dest_stride)
-{
-    v16u8 dest0, dest1, dest2, dest3;
-    v8i16 in0, in1, in2, in3;
-    v8i16 mul0, mul1, mul2, mul3, dequant_in0, dequant_in1;
-    v8i16 hz0, hz1, hz2, hz3, vt0, vt1, vt2, vt3;
-    v8i16 res0, res1, res2, res3;
-    v4i32 hz0l, hz1l, hz2l, hz3l, hz0r, hz1r, hz2r, hz3r;
-    v4i32 vt0l, vt1l, vt2l, vt3l, vt0r, vt1r, vt2r, vt3r;
-    v16i8 zero = { 0 };
+                                          int16_t *dequant_input, uint8_t *dest,
+                                          int32_t dest_stride) {
+  v16u8 dest0, dest1, dest2, dest3;
+  v8i16 in0, in1, in2, in3;
+  v8i16 mul0, mul1, mul2, mul3, dequant_in0, dequant_in1;
+  v8i16 hz0, hz1, hz2, hz3, vt0, vt1, vt2, vt3;
+  v8i16 res0, res1, res2, res3;
+  v4i32 hz0l, hz1l, hz2l, hz3l, hz0r, hz1r, hz2r, hz3r;
+  v4i32 vt0l, vt1l, vt2l, vt3l, vt0r, vt1r, vt2r, vt3r;
+  v16i8 zero = { 0 };
 
-    LD_SH4(input, 8, in0, in1, in2, in3);
-    LD_SH2(dequant_input, 8, dequant_in0, dequant_in1);
-    MUL4(in0, dequant_in0, in1, dequant_in1, in2, dequant_in0, in3, dequant_in1,
-         mul0, mul1, mul2, mul3);
-    PCKEV_D2_SH(mul2, mul0, mul3, mul1, in0, in2);
-    PCKOD_D2_SH(mul2, mul0, mul3, mul1, in1, in3);
-    VP8_IDCT_1D_H(in0, in1, in2, in3, hz0, hz1, hz2, hz3);
-    TRANSPOSE_TWO_4x4_H(hz0, hz1, hz2, hz3, hz0, hz1, hz2, hz3);
-    UNPCK_SH_SW(hz0, hz0r, hz0l);
-    UNPCK_SH_SW(hz1, hz1r, hz1l);
-    UNPCK_SH_SW(hz2, hz2r, hz2l);
-    UNPCK_SH_SW(hz3, hz3r, hz3l);
-    VP8_IDCT_1D_W(hz0l, hz1l, hz2l, hz3l, vt0l, vt1l, vt2l, vt3l);
-    SRARI_W4_SW(vt0l, vt1l, vt2l, vt3l, 3);
-    VP8_IDCT_1D_W(hz0r, hz1r, hz2r, hz3r, vt0r, vt1r, vt2r, vt3r);
-    SRARI_W4_SW(vt0r, vt1r, vt2r, vt3r, 3);
-    PCKEV_H4_SH(vt0l, vt0r, vt1l, vt1r, vt2l, vt2r, vt3l, vt3r, vt0, vt1, vt2,
-                vt3);
-    TRANSPOSE_TWO_4x4_H(vt0, vt1, vt2, vt3, vt0, vt1, vt2, vt3);
-    LD_UB4(dest, dest_stride, dest0, dest1, dest2, dest3);
-    ILVR_B4_SH(zero, dest0, zero, dest1, zero, dest2, zero, dest3, res0, res1,
-               res2, res3);
-    ADD4(res0, vt0, res1, vt1, res2, vt2, res3, vt3, res0, res1, res2, res3);
-    CLIP_SH4_0_255(res0, res1, res2, res3);
-    PCKEV_B4_SH(res0, res0, res1, res1, res2, res2, res3, res3, res0, res1,
-                res2, res3);
-    PCKOD_D2_UB(dest0, res0, dest1, res1, dest0, dest1);
-    PCKOD_D2_UB(dest2, res2, dest3, res3, dest2, dest3);
-    ST_UB4(dest0, dest1, dest2, dest3, dest, dest_stride);
+  LD_SH4(input, 8, in0, in1, in2, in3);
+  LD_SH2(dequant_input, 8, dequant_in0, dequant_in1);
+  MUL4(in0, dequant_in0, in1, dequant_in1, in2, dequant_in0, in3, dequant_in1,
+       mul0, mul1, mul2, mul3);
+  PCKEV_D2_SH(mul2, mul0, mul3, mul1, in0, in2);
+  PCKOD_D2_SH(mul2, mul0, mul3, mul1, in1, in3);
+  VP8_IDCT_1D_H(in0, in1, in2, in3, hz0, hz1, hz2, hz3);
+  TRANSPOSE_TWO_4x4_H(hz0, hz1, hz2, hz3, hz0, hz1, hz2, hz3);
+  UNPCK_SH_SW(hz0, hz0r, hz0l);
+  UNPCK_SH_SW(hz1, hz1r, hz1l);
+  UNPCK_SH_SW(hz2, hz2r, hz2l);
+  UNPCK_SH_SW(hz3, hz3r, hz3l);
+  VP8_IDCT_1D_W(hz0l, hz1l, hz2l, hz3l, vt0l, vt1l, vt2l, vt3l);
+  SRARI_W4_SW(vt0l, vt1l, vt2l, vt3l, 3);
+  VP8_IDCT_1D_W(hz0r, hz1r, hz2r, hz3r, vt0r, vt1r, vt2r, vt3r);
+  SRARI_W4_SW(vt0r, vt1r, vt2r, vt3r, 3);
+  PCKEV_H4_SH(vt0l, vt0r, vt1l, vt1r, vt2l, vt2r, vt3l, vt3r, vt0, vt1, vt2,
+              vt3);
+  TRANSPOSE_TWO_4x4_H(vt0, vt1, vt2, vt3, vt0, vt1, vt2, vt3);
+  LD_UB4(dest, dest_stride, dest0, dest1, dest2, dest3);
+  ILVR_B4_SH(zero, dest0, zero, dest1, zero, dest2, zero, dest3, res0, res1,
+             res2, res3);
+  ADD4(res0, vt0, res1, vt1, res2, vt2, res3, vt3, res0, res1, res2, res3);
+  CLIP_SH4_0_255(res0, res1, res2, res3);
+  PCKEV_B4_SH(res0, res0, res1, res1, res2, res2, res3, res3, res0, res1, res2,
+              res3);
+  PCKOD_D2_UB(dest0, res0, dest1, res1, dest0, dest1);
+  PCKOD_D2_UB(dest2, res2, dest3, res3, dest2, dest3);
+  ST_UB4(dest0, dest1, dest2, dest3, dest, dest_stride);
 
-    __asm__ __volatile__(
-        "sw   $zero,    0(%[input])  \n\t"
-        "sw   $zero,    4(%[input])  \n\t"
-        "sw   $zero,    8(%[input])  \n\t"
-        "sw   $zero,   12(%[input])  \n\t"
-        "sw   $zero,   16(%[input])  \n\t"
-        "sw   $zero,   20(%[input])  \n\t"
-        "sw   $zero,   24(%[input])  \n\t"
-        "sw   $zero,   28(%[input])  \n\t"
-        "sw   $zero,   32(%[input])  \n\t"
-        "sw   $zero,   36(%[input])  \n\t"
-        "sw   $zero,   40(%[input])  \n\t"
-        "sw   $zero,   44(%[input])  \n\t"
-        "sw   $zero,   48(%[input])  \n\t"
-        "sw   $zero,   52(%[input])  \n\t"
-        "sw   $zero,   56(%[input])  \n\t"
-        "sw   $zero,   60(%[input])  \n\t"::
+  __asm__ __volatile__(
+      "sw   $zero,    0(%[input])  \n\t"
+      "sw   $zero,    4(%[input])  \n\t"
+      "sw   $zero,    8(%[input])  \n\t"
+      "sw   $zero,   12(%[input])  \n\t"
+      "sw   $zero,   16(%[input])  \n\t"
+      "sw   $zero,   20(%[input])  \n\t"
+      "sw   $zero,   24(%[input])  \n\t"
+      "sw   $zero,   28(%[input])  \n\t"
+      "sw   $zero,   32(%[input])  \n\t"
+      "sw   $zero,   36(%[input])  \n\t"
+      "sw   $zero,   40(%[input])  \n\t"
+      "sw   $zero,   44(%[input])  \n\t"
+      "sw   $zero,   48(%[input])  \n\t"
+      "sw   $zero,   52(%[input])  \n\t"
+      "sw   $zero,   56(%[input])  \n\t"
+      "sw   $zero,   60(%[input])  \n\t" ::
 
-        [input] "r"(input)
-    );
+          [input] "r"(input));
 }
 
 static void dequant_idct_addconst_2x_msa(int16_t *input, int16_t *dequant_input,
-                                         uint8_t *dest, int32_t dest_stride)
-{
-    v8i16 input_dc0, input_dc1, vec;
-    v16u8 dest0, dest1, dest2, dest3;
-    v16i8 zero = { 0 };
-    v8i16 res0, res1, res2, res3;
+                                         uint8_t *dest, int32_t dest_stride) {
+  v8i16 input_dc0, input_dc1, vec;
+  v16u8 dest0, dest1, dest2, dest3;
+  v16i8 zero = { 0 };
+  v8i16 res0, res1, res2, res3;
 
-    input_dc0 = __msa_fill_h(input[0] * dequant_input[0]);
-    input_dc1 = __msa_fill_h(input[16] * dequant_input[0]);
-    SRARI_H2_SH(input_dc0, input_dc1, 3);
-    vec = (v8i16)__msa_pckev_d((v2i64)input_dc1, (v2i64)input_dc0);
-    input[0] = 0;
-    input[16] = 0;
-    LD_UB4(dest, dest_stride, dest0, dest1, dest2, dest3);
-    ILVR_B4_SH(zero, dest0, zero, dest1, zero, dest2, zero, dest3, res0,
-               res1, res2, res3);
-    ADD4(res0, vec, res1, vec, res2, vec, res3, vec, res0, res1, res2, res3);
-    CLIP_SH4_0_255(res0, res1, res2, res3);
-    PCKEV_B4_SH(res0, res0, res1, res1, res2, res2, res3, res3, res0, res1,
-                res2, res3);
-    PCKOD_D2_UB(dest0, res0, dest1, res1, dest0, dest1);
-    PCKOD_D2_UB(dest2, res2, dest3, res3, dest2, dest3);
-    ST_UB4(dest0, dest1, dest2, dest3, dest, dest_stride);
+  input_dc0 = __msa_fill_h(input[0] * dequant_input[0]);
+  input_dc1 = __msa_fill_h(input[16] * dequant_input[0]);
+  SRARI_H2_SH(input_dc0, input_dc1, 3);
+  vec = (v8i16)__msa_pckev_d((v2i64)input_dc1, (v2i64)input_dc0);
+  input[0] = 0;
+  input[16] = 0;
+  LD_UB4(dest, dest_stride, dest0, dest1, dest2, dest3);
+  ILVR_B4_SH(zero, dest0, zero, dest1, zero, dest2, zero, dest3, res0, res1,
+             res2, res3);
+  ADD4(res0, vec, res1, vec, res2, vec, res3, vec, res0, res1, res2, res3);
+  CLIP_SH4_0_255(res0, res1, res2, res3);
+  PCKEV_B4_SH(res0, res0, res1, res1, res2, res2, res3, res3, res0, res1, res2,
+              res3);
+  PCKOD_D2_UB(dest0, res0, dest1, res1, dest0, dest1);
+  PCKOD_D2_UB(dest2, res2, dest3, res3, dest2, dest3);
+  ST_UB4(dest0, dest1, dest2, dest3, dest, dest_stride);
 }
 
 void vp8_short_idct4x4llm_msa(int16_t *input, uint8_t *pred_ptr,
                               int32_t pred_stride, uint8_t *dst_ptr,
-                              int32_t dst_stride)
-{
-    idct4x4_addblk_msa(input, pred_ptr, pred_stride, dst_ptr, dst_stride);
+                              int32_t dst_stride) {
+  idct4x4_addblk_msa(input, pred_ptr, pred_stride, dst_ptr, dst_stride);
 }
 
 void vp8_dc_only_idct_add_msa(int16_t input_dc, uint8_t *pred_ptr,
                               int32_t pred_stride, uint8_t *dst_ptr,
-                              int32_t dst_stride)
-{
-    idct4x4_addconst_msa(input_dc, pred_ptr, pred_stride, dst_ptr, dst_stride);
+                              int32_t dst_stride) {
+  idct4x4_addconst_msa(input_dc, pred_ptr, pred_stride, dst_ptr, dst_stride);
 }
 
-void vp8_dequantize_b_msa(BLOCKD *d, int16_t *DQC)
-{
-    v8i16 dqc0, dqc1, q0, q1, dq0, dq1;
+void vp8_dequantize_b_msa(BLOCKD *d, int16_t *DQC) {
+  v8i16 dqc0, dqc1, q0, q1, dq0, dq1;
 
-    LD_SH2(DQC, 8, dqc0, dqc1);
-    LD_SH2(d->qcoeff, 8, q0, q1);
-    MUL2(dqc0, q0, dqc1, q1, dq0, dq1);
-    ST_SH2(dq0, dq1, d->dqcoeff, 8);
+  LD_SH2(DQC, 8, dqc0, dqc1);
+  LD_SH2(d->qcoeff, 8, q0, q1);
+  MUL2(dqc0, q0, dqc1, q1, dq0, dq1);
+  ST_SH2(dq0, dq1, d->dqcoeff, 8);
 }
 
-void vp8_dequant_idct_add_msa(int16_t *input, int16_t *dq,
-                              uint8_t *dest, int32_t stride)
-{
-    dequant_idct4x4_addblk_msa(input, dq, dest, stride);
+void vp8_dequant_idct_add_msa(int16_t *input, int16_t *dq, uint8_t *dest,
+                              int32_t stride) {
+  dequant_idct4x4_addblk_msa(input, dq, dest, stride);
 
-    __asm__ __volatile__ (
-        "sw     $zero,    0(%[input])     \n\t"
-        "sw     $zero,    4(%[input])     \n\t"
-        "sw     $zero,    8(%[input])     \n\t"
-        "sw     $zero,   12(%[input])     \n\t"
-        "sw     $zero,   16(%[input])     \n\t"
-        "sw     $zero,   20(%[input])     \n\t"
-        "sw     $zero,   24(%[input])     \n\t"
-        "sw     $zero,   28(%[input])     \n\t"
+  __asm__ __volatile__(
+      "sw     $zero,    0(%[input])     \n\t"
+      "sw     $zero,    4(%[input])     \n\t"
+      "sw     $zero,    8(%[input])     \n\t"
+      "sw     $zero,   12(%[input])     \n\t"
+      "sw     $zero,   16(%[input])     \n\t"
+      "sw     $zero,   20(%[input])     \n\t"
+      "sw     $zero,   24(%[input])     \n\t"
+      "sw     $zero,   28(%[input])     \n\t"
 
-        :
-        : [input] "r" (input)
-    );
+      :
+      : [input] "r"(input));
 }
 
-void vp8_dequant_idct_add_y_block_msa(int16_t *q, int16_t *dq,
-                                      uint8_t *dst, int32_t stride,
-                                      char *eobs)
-{
-    int16_t *eobs_h = (int16_t *)eobs;
-    uint8_t i;
+void vp8_dequant_idct_add_y_block_msa(int16_t *q, int16_t *dq, uint8_t *dst,
+                                      int32_t stride, char *eobs) {
+  int16_t *eobs_h = (int16_t *)eobs;
+  uint8_t i;
 
-    for (i = 4; i--;)
-    {
-        if (eobs_h[0])
-        {
-            if (eobs_h[0] & 0xfefe)
-            {
-                dequant_idct4x4_addblk_2x_msa(q, dq, dst, stride);
-            }
-            else
-            {
-                dequant_idct_addconst_2x_msa(q, dq, dst, stride);
-            }
-        }
+  for (i = 4; i--;) {
+    if (eobs_h[0]) {
+      if (eobs_h[0] & 0xfefe) {
+        dequant_idct4x4_addblk_2x_msa(q, dq, dst, stride);
+      } else {
+        dequant_idct_addconst_2x_msa(q, dq, dst, stride);
+      }
+    }
 
-        q += 32;
+    q += 32;
 
-        if (eobs_h[1])
-        {
-            if (eobs_h[1] & 0xfefe)
-            {
-                dequant_idct4x4_addblk_2x_msa(q, dq, dst + 8, stride);
-            }
-            else
-            {
-                dequant_idct_addconst_2x_msa(q, dq, dst + 8, stride);
-            }
-        }
-
-        q += 32;
-        dst += (4 * stride);
-        eobs_h += 2;
+    if (eobs_h[1]) {
+      if (eobs_h[1] & 0xfefe) {
+        dequant_idct4x4_addblk_2x_msa(q, dq, dst + 8, stride);
+      } else {
+        dequant_idct_addconst_2x_msa(q, dq, dst + 8, stride);
+      }
     }
+
+    q += 32;
+    dst += (4 * stride);
+    eobs_h += 2;
+  }
 }
 
-void vp8_dequant_idct_add_uv_block_msa(int16_t *q, int16_t *dq,
-                                       uint8_t *dstu, uint8_t *dstv,
-                                       int32_t stride, char *eobs)
-{
-    int16_t *eobs_h = (int16_t *)eobs;
+void vp8_dequant_idct_add_uv_block_msa(int16_t *q, int16_t *dq, uint8_t *dstu,
+                                       uint8_t *dstv, int32_t stride,
+                                       char *eobs) {
+  int16_t *eobs_h = (int16_t *)eobs;
 
-    if (eobs_h[0])
-    {
-        if (eobs_h[0] & 0xfefe)
-        {
-            dequant_idct4x4_addblk_2x_msa(q, dq, dstu, stride);
-        }
-        else
-        {
-            dequant_idct_addconst_2x_msa(q, dq, dstu, stride);
-        }
+  if (eobs_h[0]) {
+    if (eobs_h[0] & 0xfefe) {
+      dequant_idct4x4_addblk_2x_msa(q, dq, dstu, stride);
+    } else {
+      dequant_idct_addconst_2x_msa(q, dq, dstu, stride);
     }
+  }
 
-    q += 32;
-    dstu += (stride * 4);
+  q += 32;
+  dstu += (stride * 4);
 
-    if (eobs_h[1])
-    {
-        if (eobs_h[1] & 0xfefe)
-        {
-            dequant_idct4x4_addblk_2x_msa(q, dq, dstu, stride);
-        }
-        else
-        {
-            dequant_idct_addconst_2x_msa(q, dq, dstu, stride);
-        }
+  if (eobs_h[1]) {
+    if (eobs_h[1] & 0xfefe) {
+      dequant_idct4x4_addblk_2x_msa(q, dq, dstu, stride);
+    } else {
+      dequant_idct_addconst_2x_msa(q, dq, dstu, stride);
     }
+  }
 
-    q += 32;
+  q += 32;
 
-    if (eobs_h[2])
-    {
-        if (eobs_h[2] & 0xfefe)
-        {
-            dequant_idct4x4_addblk_2x_msa(q, dq, dstv, stride);
-        }
-        else
-        {
-            dequant_idct_addconst_2x_msa(q, dq, dstv, stride);
-        }
+  if (eobs_h[2]) {
+    if (eobs_h[2] & 0xfefe) {
+      dequant_idct4x4_addblk_2x_msa(q, dq, dstv, stride);
+    } else {
+      dequant_idct_addconst_2x_msa(q, dq, dstv, stride);
     }
+  }
 
-    q += 32;
-    dstv += (stride * 4);
+  q += 32;
+  dstv += (stride * 4);
 
-    if (eobs_h[3])
-    {
-        if (eobs_h[3] & 0xfefe)
-        {
-            dequant_idct4x4_addblk_2x_msa(q, dq, dstv, stride);
-        }
-        else
-        {
-            dequant_idct_addconst_2x_msa(q, dq, dstv, stride);
-        }
+  if (eobs_h[3]) {
+    if (eobs_h[3] & 0xfefe) {
+      dequant_idct4x4_addblk_2x_msa(q, dq, dstv, stride);
+    } else {
+      dequant_idct_addconst_2x_msa(q, dq, dstv, stride);
     }
+  }
 }
--- a/vp8/common/mips/msa/loopfilter_filters_msa.c
+++ b/vp8/common/mips/msa/loopfilter_filters_msa.c
@@ -12,265 +12,263 @@
 #include "vp8/common/loopfilter.h"
 #include "vp8/common/mips/msa/vp8_macros_msa.h"
 
-#define VP8_SIMPLE_MASK(p1, p0, q0, q1, b_limit, mask)         \
-{                                                              \
-    v16u8 p1_a_sub_q1, p0_a_sub_q0;                            \
-                                                               \
-    p0_a_sub_q0 = __msa_asub_u_b(p0, q0);                      \
-    p1_a_sub_q1 = __msa_asub_u_b(p1, q1);                      \
-    p1_a_sub_q1 = (v16u8)__msa_srli_b((v16i8)p1_a_sub_q1, 1);  \
-    p0_a_sub_q0 = __msa_adds_u_b(p0_a_sub_q0, p0_a_sub_q0);    \
-    mask = __msa_adds_u_b(p0_a_sub_q0, p1_a_sub_q1);           \
-    mask = ((v16u8)mask <= b_limit);                           \
-}
+#define VP8_SIMPLE_MASK(p1, p0, q0, q1, b_limit, mask)        \
+  {                                                           \
+    v16u8 p1_a_sub_q1, p0_a_sub_q0;                           \
+                                                              \
+    p0_a_sub_q0 = __msa_asub_u_b(p0, q0);                     \
+    p1_a_sub_q1 = __msa_asub_u_b(p1, q1);                     \
+    p1_a_sub_q1 = (v16u8)__msa_srli_b((v16i8)p1_a_sub_q1, 1); \
+    p0_a_sub_q0 = __msa_adds_u_b(p0_a_sub_q0, p0_a_sub_q0);   \
+    mask = __msa_adds_u_b(p0_a_sub_q0, p1_a_sub_q1);          \
+    mask = ((v16u8)mask <= b_limit);                          \
+  }
 
-#define VP8_LPF_FILTER4_4W(p1_in_out, p0_in_out, q0_in_out, q1_in_out,  \
-                           mask_in, hev_in)                             \
-{                                                                       \
-    v16i8 p1_m, p0_m, q0_m, q1_m, q0_sub_p0, filt_sign;                 \
-    v16i8 filt, filt1, filt2, cnst4b, cnst3b;                           \
-    v8i16 q0_sub_p0_r, q0_sub_p0_l, filt_l, filt_r, cnst3h;             \
-                                                                        \
-    p1_m = (v16i8)__msa_xori_b(p1_in_out, 0x80);                        \
-    p0_m = (v16i8)__msa_xori_b(p0_in_out, 0x80);                        \
-    q0_m = (v16i8)__msa_xori_b(q0_in_out, 0x80);                        \
-    q1_m = (v16i8)__msa_xori_b(q1_in_out, 0x80);                        \
-                                                                        \
-    filt = __msa_subs_s_b(p1_m, q1_m);                                  \
-                                                                        \
-    filt = filt & (v16i8)hev_in;                                        \
-                                                                        \
-    q0_sub_p0 = q0_m - p0_m;                                            \
-    filt_sign = __msa_clti_s_b(filt, 0);                                \
-                                                                        \
-    cnst3h = __msa_ldi_h(3);                                            \
-    q0_sub_p0_r = (v8i16)__msa_ilvr_b(q0_sub_p0, q0_sub_p0);            \
-    q0_sub_p0_r = __msa_dotp_s_h((v16i8)q0_sub_p0_r, (v16i8)cnst3h);    \
-    filt_r = (v8i16)__msa_ilvr_b(filt_sign, filt);                      \
-    filt_r += q0_sub_p0_r;                                              \
-    filt_r = __msa_sat_s_h(filt_r, 7);                                  \
-                                                                        \
-    q0_sub_p0_l = (v8i16)__msa_ilvl_b(q0_sub_p0, q0_sub_p0);            \
-    q0_sub_p0_l = __msa_dotp_s_h((v16i8)q0_sub_p0_l, (v16i8)cnst3h);    \
-    filt_l = (v8i16)__msa_ilvl_b(filt_sign, filt);                      \
-    filt_l += q0_sub_p0_l;                                              \
-    filt_l = __msa_sat_s_h(filt_l, 7);                                  \
-                                                                        \
-    filt = __msa_pckev_b((v16i8)filt_l, (v16i8)filt_r);                 \
-    filt = filt & (v16i8)mask_in;                                       \
-                                                                        \
-    cnst4b = __msa_ldi_b(4);                                            \
-    filt1 = __msa_adds_s_b(filt, cnst4b);                               \
-    filt1 >>= 3;                                                        \
-                                                                        \
-    cnst3b = __msa_ldi_b(3);                                            \
-    filt2 = __msa_adds_s_b(filt, cnst3b);                               \
-    filt2 >>= 3;                                                        \
-                                                                        \
-    q0_m = __msa_subs_s_b(q0_m, filt1);                                 \
-    q0_in_out = __msa_xori_b((v16u8)q0_m, 0x80);                        \
-    p0_m = __msa_adds_s_b(p0_m, filt2);                                 \
-    p0_in_out = __msa_xori_b((v16u8)p0_m, 0x80);                        \
-                                                                        \
-    filt = __msa_srari_b(filt1, 1);                                     \
-    hev_in = __msa_xori_b((v16u8)hev_in, 0xff);                         \
-    filt = filt & (v16i8)hev_in;                                        \
-                                                                        \
-    q1_m = __msa_subs_s_b(q1_m, filt);                                  \
-    q1_in_out = __msa_xori_b((v16u8)q1_m, 0x80);                        \
-    p1_m = __msa_adds_s_b(p1_m, filt);                                  \
-    p1_in_out = __msa_xori_b((v16u8)p1_m, 0x80);                        \
-}
+#define VP8_LPF_FILTER4_4W(p1_in_out, p0_in_out, q0_in_out, q1_in_out, \
+                           mask_in, hev_in)                            \
+  {                                                                    \
+    v16i8 p1_m, p0_m, q0_m, q1_m, q0_sub_p0, filt_sign;                \
+    v16i8 filt, filt1, filt2, cnst4b, cnst3b;                          \
+    v8i16 q0_sub_p0_r, q0_sub_p0_l, filt_l, filt_r, cnst3h;            \
+                                                                       \
+    p1_m = (v16i8)__msa_xori_b(p1_in_out, 0x80);                       \
+    p0_m = (v16i8)__msa_xori_b(p0_in_out, 0x80);                       \
+    q0_m = (v16i8)__msa_xori_b(q0_in_out, 0x80);                       \
+    q1_m = (v16i8)__msa_xori_b(q1_in_out, 0x80);                       \
+                                                                       \
+    filt = __msa_subs_s_b(p1_m, q1_m);                                 \
+                                                                       \
+    filt = filt & (v16i8)hev_in;                                       \
+                                                                       \
+    q0_sub_p0 = q0_m - p0_m;                                           \
+    filt_sign = __msa_clti_s_b(filt, 0);                               \
+                                                                       \
+    cnst3h = __msa_ldi_h(3);                                           \
+    q0_sub_p0_r = (v8i16)__msa_ilvr_b(q0_sub_p0, q0_sub_p0);           \
+    q0_sub_p0_r = __msa_dotp_s_h((v16i8)q0_sub_p0_r, (v16i8)cnst3h);   \
+    filt_r = (v8i16)__msa_ilvr_b(filt_sign, filt);                     \
+    filt_r += q0_sub_p0_r;                                             \
+    filt_r = __msa_sat_s_h(filt_r, 7);                                 \
+                                                                       \
+    q0_sub_p0_l = (v8i16)__msa_ilvl_b(q0_sub_p0, q0_sub_p0);           \
+    q0_sub_p0_l = __msa_dotp_s_h((v16i8)q0_sub_p0_l, (v16i8)cnst3h);   \
+    filt_l = (v8i16)__msa_ilvl_b(filt_sign, filt);                     \
+    filt_l += q0_sub_p0_l;                                             \
+    filt_l = __msa_sat_s_h(filt_l, 7);                                 \
+                                                                       \
+    filt = __msa_pckev_b((v16i8)filt_l, (v16i8)filt_r);                \
+    filt = filt & (v16i8)mask_in;                                      \
+                                                                       \
+    cnst4b = __msa_ldi_b(4);                                           \
+    filt1 = __msa_adds_s_b(filt, cnst4b);                              \
+    filt1 >>= 3;                                                       \
+                                                                       \
+    cnst3b = __msa_ldi_b(3);                                           \
+    filt2 = __msa_adds_s_b(filt, cnst3b);                              \
+    filt2 >>= 3;                                                       \
+                                                                       \
+    q0_m = __msa_subs_s_b(q0_m, filt1);                                \
+    q0_in_out = __msa_xori_b((v16u8)q0_m, 0x80);                       \
+    p0_m = __msa_adds_s_b(p0_m, filt2);                                \
+    p0_in_out = __msa_xori_b((v16u8)p0_m, 0x80);                       \
+                                                                       \
+    filt = __msa_srari_b(filt1, 1);                                    \
+    hev_in = __msa_xori_b((v16u8)hev_in, 0xff);                        \
+    filt = filt & (v16i8)hev_in;                                       \
+                                                                       \
+    q1_m = __msa_subs_s_b(q1_m, filt);                                 \
+    q1_in_out = __msa_xori_b((v16u8)q1_m, 0x80);                       \
+    p1_m = __msa_adds_s_b(p1_m, filt);                                 \
+    p1_in_out = __msa_xori_b((v16u8)p1_m, 0x80);                       \
+  }
 
-#define VP8_SIMPLE_FILT(p1_in, p0_in, q0_in, q1_in, mask)          \
-{                                                                  \
-    v16i8 p1_m, p0_m, q0_m, q1_m, q0_sub_p0, q0_sub_p0_sign;       \
-    v16i8 filt, filt1, filt2, cnst4b, cnst3b, filt_sign;           \
-    v8i16 q0_sub_p0_r, q0_sub_p0_l, filt_l, filt_r, cnst3h;        \
-                                                                   \
-    p1_m = (v16i8)__msa_xori_b(p1_in, 0x80);                       \
-    p0_m = (v16i8)__msa_xori_b(p0_in, 0x80);                       \
-    q0_m = (v16i8)__msa_xori_b(q0_in, 0x80);                       \
-    q1_m = (v16i8)__msa_xori_b(q1_in, 0x80);                       \
-                                                                   \
-    filt = __msa_subs_s_b(p1_m, q1_m);                             \
-                                                                   \
-    q0_sub_p0 = q0_m - p0_m;                                       \
-    filt_sign = __msa_clti_s_b(filt, 0);                           \
-                                                                   \
-    cnst3h = __msa_ldi_h(3);                                       \
-    q0_sub_p0_sign = __msa_clti_s_b(q0_sub_p0, 0);                 \
-    q0_sub_p0_r = (v8i16)__msa_ilvr_b(q0_sub_p0_sign, q0_sub_p0);  \
-    q0_sub_p0_r *= cnst3h;                                         \
-    filt_r = (v8i16)__msa_ilvr_b(filt_sign, filt);                 \
-    filt_r += q0_sub_p0_r;                                         \
-    filt_r = __msa_sat_s_h(filt_r, 7);                             \
-                                                                   \
-    q0_sub_p0_l = (v8i16)__msa_ilvl_b(q0_sub_p0_sign, q0_sub_p0);  \
-    q0_sub_p0_l *= cnst3h;                                         \
-    filt_l = (v8i16)__msa_ilvl_b(filt_sign, filt);                 \
-    filt_l += q0_sub_p0_l;                                         \
-    filt_l = __msa_sat_s_h(filt_l, 7);                             \
-                                                                   \
-    filt = __msa_pckev_b((v16i8)filt_l, (v16i8)filt_r);            \
-    filt = filt & (v16i8)(mask);                                   \
-                                                                   \
-    cnst4b = __msa_ldi_b(4);                                       \
-    filt1 = __msa_adds_s_b(filt, cnst4b);                          \
-    filt1 >>= 3;                                                   \
-                                                                   \
-    cnst3b = __msa_ldi_b(3);                                       \
-    filt2 = __msa_adds_s_b(filt, cnst3b);                          \
-    filt2 >>= 3;                                                   \
-                                                                   \
-    q0_m = __msa_subs_s_b(q0_m, filt1);                            \
-    p0_m = __msa_adds_s_b(p0_m, filt2);                            \
-    q0_in = __msa_xori_b((v16u8)q0_m, 0x80);                       \
-    p0_in = __msa_xori_b((v16u8)p0_m, 0x80);                       \
-}
+#define VP8_SIMPLE_FILT(p1_in, p0_in, q0_in, q1_in, mask)         \
+  {                                                               \
+    v16i8 p1_m, p0_m, q0_m, q1_m, q0_sub_p0, q0_sub_p0_sign;      \
+    v16i8 filt, filt1, filt2, cnst4b, cnst3b, filt_sign;          \
+    v8i16 q0_sub_p0_r, q0_sub_p0_l, filt_l, filt_r, cnst3h;       \
+                                                                  \
+    p1_m = (v16i8)__msa_xori_b(p1_in, 0x80);                      \
+    p0_m = (v16i8)__msa_xori_b(p0_in, 0x80);                      \
+    q0_m = (v16i8)__msa_xori_b(q0_in, 0x80);                      \
+    q1_m = (v16i8)__msa_xori_b(q1_in, 0x80);                      \
+                                                                  \
+    filt = __msa_subs_s_b(p1_m, q1_m);                            \
+                                                                  \
+    q0_sub_p0 = q0_m - p0_m;                                      \
+    filt_sign = __msa_clti_s_b(filt, 0);                          \
+                                                                  \
+    cnst3h = __msa_ldi_h(3);                                      \
+    q0_sub_p0_sign = __msa_clti_s_b(q0_sub_p0, 0);                \
+    q0_sub_p0_r = (v8i16)__msa_ilvr_b(q0_sub_p0_sign, q0_sub_p0); \
+    q0_sub_p0_r *= cnst3h;                                        \
+    filt_r = (v8i16)__msa_ilvr_b(filt_sign, filt);                \
+    filt_r += q0_sub_p0_r;                                        \
+    filt_r = __msa_sat_s_h(filt_r, 7);                            \
+                                                                  \
+    q0_sub_p0_l = (v8i16)__msa_ilvl_b(q0_sub_p0_sign, q0_sub_p0); \
+    q0_sub_p0_l *= cnst3h;                                        \
+    filt_l = (v8i16)__msa_ilvl_b(filt_sign, filt);                \
+    filt_l += q0_sub_p0_l;                                        \
+    filt_l = __msa_sat_s_h(filt_l, 7);                            \
+                                                                  \
+    filt = __msa_pckev_b((v16i8)filt_l, (v16i8)filt_r);           \
+    filt = filt & (v16i8)(mask);                                  \
+                                                                  \
+    cnst4b = __msa_ldi_b(4);                                      \
+    filt1 = __msa_adds_s_b(filt, cnst4b);                         \
+    filt1 >>= 3;                                                  \
+                                                                  \
+    cnst3b = __msa_ldi_b(3);                                      \
+    filt2 = __msa_adds_s_b(filt, cnst3b);                         \
+    filt2 >>= 3;                                                  \
+                                                                  \
+    q0_m = __msa_subs_s_b(q0_m, filt1);                           \
+    p0_m = __msa_adds_s_b(p0_m, filt2);                           \
+    q0_in = __msa_xori_b((v16u8)q0_m, 0x80);                      \
+    p0_in = __msa_xori_b((v16u8)p0_m, 0x80);                      \
+  }
 
-#define VP8_MBFILTER(p2, p1, p0, q0, q1, q2, mask, hev)            \
-{                                                                  \
-    v16i8 p2_m, p1_m, p0_m, q2_m, q1_m, q0_m;                      \
-    v16i8 filt, q0_sub_p0, cnst4b, cnst3b;                         \
-    v16i8 u, filt1, filt2, filt_sign, q0_sub_p0_sign;              \
-    v8i16 q0_sub_p0_r, q0_sub_p0_l, filt_r, u_r, u_l, filt_l;      \
-    v8i16 cnst3h, cnst27h, cnst18h, cnst63h;                       \
-                                                                   \
-    cnst3h = __msa_ldi_h(3);                                       \
-                                                                   \
-    p2_m = (v16i8)__msa_xori_b(p2, 0x80);                          \
-    p1_m = (v16i8)__msa_xori_b(p1, 0x80);                          \
-    p0_m = (v16i8)__msa_xori_b(p0, 0x80);                          \
-    q0_m = (v16i8)__msa_xori_b(q0, 0x80);                          \
-    q1_m = (v16i8)__msa_xori_b(q1, 0x80);                          \
-    q2_m = (v16i8)__msa_xori_b(q2, 0x80);                          \
-                                                                   \
-    filt = __msa_subs_s_b(p1_m, q1_m);                             \
-    q0_sub_p0 = q0_m - p0_m;                                       \
-    q0_sub_p0_sign = __msa_clti_s_b(q0_sub_p0, 0);                 \
-    filt_sign = __msa_clti_s_b(filt, 0);                           \
-                                                                   \
-    q0_sub_p0_r = (v8i16)__msa_ilvr_b(q0_sub_p0_sign, q0_sub_p0);  \
-    q0_sub_p0_r *= cnst3h;                                         \
-    filt_r = (v8i16)__msa_ilvr_b(filt_sign, filt);                 \
-    filt_r = filt_r + q0_sub_p0_r;                                 \
-    filt_r = __msa_sat_s_h(filt_r, 7);                             \
-                                                                   \
-    q0_sub_p0_l = (v8i16)__msa_ilvl_b(q0_sub_p0_sign, q0_sub_p0);  \
-    q0_sub_p0_l *= cnst3h;                                         \
-    filt_l = (v8i16)__msa_ilvl_b(filt_sign, filt);                 \
-    filt_l = filt_l + q0_sub_p0_l;                                 \
-    filt_l = __msa_sat_s_h(filt_l, 7);                             \
-                                                                   \
-    filt = __msa_pckev_b((v16i8)filt_l, (v16i8)filt_r);            \
-    filt = filt & (v16i8)mask;                                     \
-    filt2 = filt & (v16i8)hev;                                     \
-                                                                   \
-    hev = __msa_xori_b(hev, 0xff);                                 \
-    filt = filt & (v16i8)hev;                                      \
-    cnst4b = __msa_ldi_b(4);                                       \
-    filt1 = __msa_adds_s_b(filt2, cnst4b);                         \
-    filt1 >>= 3;                                                   \
-    cnst3b = __msa_ldi_b(3);                                       \
-    filt2 = __msa_adds_s_b(filt2, cnst3b);                         \
-    filt2 >>= 3;                                                   \
-    q0_m = __msa_subs_s_b(q0_m, filt1);                            \
-    p0_m = __msa_adds_s_b(p0_m, filt2);                            \
-                                                                   \
-    filt_sign = __msa_clti_s_b(filt, 0);                           \
-    ILVRL_B2_SH(filt_sign, filt, filt_r, filt_l);                  \
-                                                                   \
-    cnst27h = __msa_ldi_h(27);                                     \
-    cnst63h = __msa_ldi_h(63);                                     \
-                                                                   \
-    u_r = filt_r * cnst27h;                                        \
-    u_r += cnst63h;                                                \
-    u_r >>= 7;                                                     \
-    u_r = __msa_sat_s_h(u_r, 7);                                   \
-    u_l = filt_l * cnst27h;                                        \
-    u_l += cnst63h;                                                \
-    u_l >>= 7;                                                     \
-    u_l = __msa_sat_s_h(u_l, 7);                                   \
-    u = __msa_pckev_b((v16i8)u_l, (v16i8)u_r);                     \
-    q0_m = __msa_subs_s_b(q0_m, u);                                \
-    q0 = __msa_xori_b((v16u8)q0_m, 0x80);                          \
-    p0_m = __msa_adds_s_b(p0_m, u);                                \
-    p0 = __msa_xori_b((v16u8)p0_m, 0x80);                          \
-    cnst18h = __msa_ldi_h(18);                                     \
-    u_r = filt_r * cnst18h;                                        \
-    u_r += cnst63h;                                                \
-    u_r >>= 7;                                                     \
-    u_r = __msa_sat_s_h(u_r, 7);                                   \
-                                                                   \
-    u_l = filt_l * cnst18h;                                        \
-    u_l += cnst63h;                                                \
-    u_l >>= 7;                                                     \
-    u_l = __msa_sat_s_h(u_l, 7);                                   \
-    u = __msa_pckev_b((v16i8)u_l, (v16i8)u_r);                     \
-    q1_m = __msa_subs_s_b(q1_m, u);                                \
-    q1 = __msa_xori_b((v16u8)q1_m, 0x80);                          \
-    p1_m = __msa_adds_s_b(p1_m, u);                                \
-    p1 = __msa_xori_b((v16u8)p1_m, 0x80);                          \
-    u_r = filt_r << 3;                                             \
-    u_r += filt_r + cnst63h;                                       \
-    u_r >>= 7;                                                     \
-    u_r = __msa_sat_s_h(u_r, 7);                                   \
-                                                                   \
-    u_l = filt_l << 3;                                             \
-    u_l += filt_l + cnst63h;                                       \
-    u_l >>= 7;                                                     \
-    u_l = __msa_sat_s_h(u_l, 7);                                   \
-    u = __msa_pckev_b((v16i8)u_l, (v16i8)u_r);                     \
-    q2_m = __msa_subs_s_b(q2_m, u);                                \
-    q2 = __msa_xori_b((v16u8)q2_m, 0x80);                          \
-    p2_m = __msa_adds_s_b(p2_m, u);                                \
-    p2 = __msa_xori_b((v16u8)p2_m, 0x80);                          \
-}
+#define VP8_MBFILTER(p2, p1, p0, q0, q1, q2, mask, hev)           \
+  {                                                               \
+    v16i8 p2_m, p1_m, p0_m, q2_m, q1_m, q0_m;                     \
+    v16i8 filt, q0_sub_p0, cnst4b, cnst3b;                        \
+    v16i8 u, filt1, filt2, filt_sign, q0_sub_p0_sign;             \
+    v8i16 q0_sub_p0_r, q0_sub_p0_l, filt_r, u_r, u_l, filt_l;     \
+    v8i16 cnst3h, cnst27h, cnst18h, cnst63h;                      \
+                                                                  \
+    cnst3h = __msa_ldi_h(3);                                      \
+                                                                  \
+    p2_m = (v16i8)__msa_xori_b(p2, 0x80);                         \
+    p1_m = (v16i8)__msa_xori_b(p1, 0x80);                         \
+    p0_m = (v16i8)__msa_xori_b(p0, 0x80);                         \
+    q0_m = (v16i8)__msa_xori_b(q0, 0x80);                         \
+    q1_m = (v16i8)__msa_xori_b(q1, 0x80);                         \
+    q2_m = (v16i8)__msa_xori_b(q2, 0x80);                         \
+                                                                  \
+    filt = __msa_subs_s_b(p1_m, q1_m);                            \
+    q0_sub_p0 = q0_m - p0_m;                                      \
+    q0_sub_p0_sign = __msa_clti_s_b(q0_sub_p0, 0);                \
+    filt_sign = __msa_clti_s_b(filt, 0);                          \
+                                                                  \
+    q0_sub_p0_r = (v8i16)__msa_ilvr_b(q0_sub_p0_sign, q0_sub_p0); \
+    q0_sub_p0_r *= cnst3h;                                        \
+    filt_r = (v8i16)__msa_ilvr_b(filt_sign, filt);                \
+    filt_r = filt_r + q0_sub_p0_r;                                \
+    filt_r = __msa_sat_s_h(filt_r, 7);                            \
+                                                                  \
+    q0_sub_p0_l = (v8i16)__msa_ilvl_b(q0_sub_p0_sign, q0_sub_p0); \
+    q0_sub_p0_l *= cnst3h;                                        \
+    filt_l = (v8i16)__msa_ilvl_b(filt_sign, filt);                \
+    filt_l = filt_l + q0_sub_p0_l;                                \
+    filt_l = __msa_sat_s_h(filt_l, 7);                            \
+                                                                  \
+    filt = __msa_pckev_b((v16i8)filt_l, (v16i8)filt_r);           \
+    filt = filt & (v16i8)mask;                                    \
+    filt2 = filt & (v16i8)hev;                                    \
+                                                                  \
+    hev = __msa_xori_b(hev, 0xff);                                \
+    filt = filt & (v16i8)hev;                                     \
+    cnst4b = __msa_ldi_b(4);                                      \
+    filt1 = __msa_adds_s_b(filt2, cnst4b);                        \
+    filt1 >>= 3;                                                  \
+    cnst3b = __msa_ldi_b(3);                                      \
+    filt2 = __msa_adds_s_b(filt2, cnst3b);                        \
+    filt2 >>= 3;                                                  \
+    q0_m = __msa_subs_s_b(q0_m, filt1);                           \
+    p0_m = __msa_adds_s_b(p0_m, filt2);                           \
+                                                                  \
+    filt_sign = __msa_clti_s_b(filt, 0);                          \
+    ILVRL_B2_SH(filt_sign, filt, filt_r, filt_l);                 \
+                                                                  \
+    cnst27h = __msa_ldi_h(27);                                    \
+    cnst63h = __msa_ldi_h(63);                                    \
+                                                                  \
+    u_r = filt_r * cnst27h;                                       \
+    u_r += cnst63h;                                               \
+    u_r >>= 7;                                                    \
+    u_r = __msa_sat_s_h(u_r, 7);                                  \
+    u_l = filt_l * cnst27h;                                       \
+    u_l += cnst63h;                                               \
+    u_l >>= 7;                                                    \
+    u_l = __msa_sat_s_h(u_l, 7);                                  \
+    u = __msa_pckev_b((v16i8)u_l, (v16i8)u_r);                    \
+    q0_m = __msa_subs_s_b(q0_m, u);                               \
+    q0 = __msa_xori_b((v16u8)q0_m, 0x80);                         \
+    p0_m = __msa_adds_s_b(p0_m, u);                               \
+    p0 = __msa_xori_b((v16u8)p0_m, 0x80);                         \
+    cnst18h = __msa_ldi_h(18);                                    \
+    u_r = filt_r * cnst18h;                                       \
+    u_r += cnst63h;                                               \
+    u_r >>= 7;                                                    \
+    u_r = __msa_sat_s_h(u_r, 7);                                  \
+                                                                  \
+    u_l = filt_l * cnst18h;                                       \
+    u_l += cnst63h;                                               \
+    u_l >>= 7;                                                    \
+    u_l = __msa_sat_s_h(u_l, 7);                                  \
+    u = __msa_pckev_b((v16i8)u_l, (v16i8)u_r);                    \
+    q1_m = __msa_subs_s_b(q1_m, u);                               \
+    q1 = __msa_xori_b((v16u8)q1_m, 0x80);                         \
+    p1_m = __msa_adds_s_b(p1_m, u);                               \
+    p1 = __msa_xori_b((v16u8)p1_m, 0x80);                         \
+    u_r = filt_r << 3;                                            \
+    u_r += filt_r + cnst63h;                                      \
+    u_r >>= 7;                                                    \
+    u_r = __msa_sat_s_h(u_r, 7);                                  \
+                                                                  \
+    u_l = filt_l << 3;                                            \
+    u_l += filt_l + cnst63h;                                      \
+    u_l >>= 7;                                                    \
+    u_l = __msa_sat_s_h(u_l, 7);                                  \
+    u = __msa_pckev_b((v16i8)u_l, (v16i8)u_r);                    \
+    q2_m = __msa_subs_s_b(q2_m, u);                               \
+    q2 = __msa_xori_b((v16u8)q2_m, 0x80);                         \
+    p2_m = __msa_adds_s_b(p2_m, u);                               \
+    p2 = __msa_xori_b((v16u8)p2_m, 0x80);                         \
+  }
 
-#define LPF_MASK_HEV(p3_in, p2_in, p1_in, p0_in,                   \
-                     q0_in, q1_in, q2_in, q3_in,                   \
-                     limit_in, b_limit_in, thresh_in,              \
-                     hev_out, mask_out, flat_out)                  \
-{                                                                  \
-    v16u8 p3_asub_p2_m, p2_asub_p1_m, p1_asub_p0_m, q1_asub_q0_m;  \
-    v16u8 p1_asub_q1_m, p0_asub_q0_m, q3_asub_q2_m, q2_asub_q1_m;  \
-                                                                   \
-    p3_asub_p2_m = __msa_asub_u_b((p3_in), (p2_in));               \
-    p2_asub_p1_m = __msa_asub_u_b((p2_in), (p1_in));               \
-    p1_asub_p0_m = __msa_asub_u_b((p1_in), (p0_in));               \
-    q1_asub_q0_m = __msa_asub_u_b((q1_in), (q0_in));               \
-    q2_asub_q1_m = __msa_asub_u_b((q2_in), (q1_in));               \
-    q3_asub_q2_m = __msa_asub_u_b((q3_in), (q2_in));               \
-    p0_asub_q0_m = __msa_asub_u_b((p0_in), (q0_in));               \
-    p1_asub_q1_m = __msa_asub_u_b((p1_in), (q1_in));               \
-    flat_out = __msa_max_u_b(p1_asub_p0_m, q1_asub_q0_m);          \
-    hev_out = (thresh_in) < (v16u8)flat_out;                       \
-    p0_asub_q0_m = __msa_adds_u_b(p0_asub_q0_m, p0_asub_q0_m);     \
-    p1_asub_q1_m >>= 1;                                            \
-    p0_asub_q0_m = __msa_adds_u_b(p0_asub_q0_m, p1_asub_q1_m);     \
-    mask_out = (b_limit_in) < p0_asub_q0_m;                        \
-    mask_out = __msa_max_u_b(flat_out, mask_out);                  \
-    p3_asub_p2_m = __msa_max_u_b(p3_asub_p2_m, p2_asub_p1_m);      \
-    mask_out = __msa_max_u_b(p3_asub_p2_m, mask_out);              \
-    q2_asub_q1_m = __msa_max_u_b(q2_asub_q1_m, q3_asub_q2_m);      \
-    mask_out = __msa_max_u_b(q2_asub_q1_m, mask_out);              \
-    mask_out = (limit_in) < (v16u8)mask_out;                       \
-    mask_out = __msa_xori_b(mask_out, 0xff);                       \
-}
+#define LPF_MASK_HEV(p3_in, p2_in, p1_in, p0_in, q0_in, q1_in, q2_in, q3_in, \
+                     limit_in, b_limit_in, thresh_in, hev_out, mask_out,     \
+                     flat_out)                                               \
+  {                                                                          \
+    v16u8 p3_asub_p2_m, p2_asub_p1_m, p1_asub_p0_m, q1_asub_q0_m;            \
+    v16u8 p1_asub_q1_m, p0_asub_q0_m, q3_asub_q2_m, q2_asub_q1_m;            \
+                                                                             \
+    p3_asub_p2_m = __msa_asub_u_b((p3_in), (p2_in));                         \
+    p2_asub_p1_m = __msa_asub_u_b((p2_in), (p1_in));                         \
+    p1_asub_p0_m = __msa_asub_u_b((p1_in), (p0_in));                         \
+    q1_asub_q0_m = __msa_asub_u_b((q1_in), (q0_in));                         \
+    q2_asub_q1_m = __msa_asub_u_b((q2_in), (q1_in));                         \
+    q3_asub_q2_m = __msa_asub_u_b((q3_in), (q2_in));                         \
+    p0_asub_q0_m = __msa_asub_u_b((p0_in), (q0_in));                         \
+    p1_asub_q1_m = __msa_asub_u_b((p1_in), (q1_in));                         \
+    flat_out = __msa_max_u_b(p1_asub_p0_m, q1_asub_q0_m);                    \
+    hev_out = (thresh_in) < (v16u8)flat_out;                                 \
+    p0_asub_q0_m = __msa_adds_u_b(p0_asub_q0_m, p0_asub_q0_m);               \
+    p1_asub_q1_m >>= 1;                                                      \
+    p0_asub_q0_m = __msa_adds_u_b(p0_asub_q0_m, p1_asub_q1_m);               \
+    mask_out = (b_limit_in) < p0_asub_q0_m;                                  \
+    mask_out = __msa_max_u_b(flat_out, mask_out);                            \
+    p3_asub_p2_m = __msa_max_u_b(p3_asub_p2_m, p2_asub_p1_m);                \
+    mask_out = __msa_max_u_b(p3_asub_p2_m, mask_out);                        \
+    q2_asub_q1_m = __msa_max_u_b(q2_asub_q1_m, q3_asub_q2_m);                \
+    mask_out = __msa_max_u_b(q2_asub_q1_m, mask_out);                        \
+    mask_out = (limit_in) < (v16u8)mask_out;                                 \
+    mask_out = __msa_xori_b(mask_out, 0xff);                                 \
+  }
 
-#define VP8_ST6x1_UB(in0, in0_idx, in1, in1_idx, pdst, stride)  \
-{                                                               \
-    uint16_t tmp0_h;                                            \
-    uint32_t tmp0_w;                                            \
-                                                                \
-    tmp0_w = __msa_copy_u_w((v4i32)in0, in0_idx);               \
-    tmp0_h = __msa_copy_u_h((v8i16)in1, in1_idx);               \
-    SW(tmp0_w, pdst);                                           \
-    SH(tmp0_h, pdst + stride);                                  \
-}
+#define VP8_ST6x1_UB(in0, in0_idx, in1, in1_idx, pdst, stride) \
+  {                                                            \
+    uint16_t tmp0_h;                                           \
+    uint32_t tmp0_w;                                           \
+                                                               \
+    tmp0_w = __msa_copy_u_w((v4i32)in0, in0_idx);              \
+    tmp0_h = __msa_copy_u_h((v8i16)in1, in1_idx);              \
+    SW(tmp0_w, pdst);                                          \
+    SH(tmp0_h, pdst + stride);                                 \
+  }
 
-
 static void loop_filter_horizontal_4_dual_msa(uint8_t *src, int32_t pitch,
                                               const uint8_t *b_limit0_ptr,
                                               const uint8_t *limit0_ptr,
@@ -277,30 +275,29 @@
                                               const uint8_t *thresh0_ptr,
                                               const uint8_t *b_limit1_ptr,
                                               const uint8_t *limit1_ptr,
-                                              const uint8_t *thresh1_ptr)
-{
-    v16u8 mask, hev, flat;
-    v16u8 thresh0, b_limit0, limit0, thresh1, b_limit1, limit1;
-    v16u8 p3, p2, p1, p0, q3, q2, q1, q0;
+                                              const uint8_t *thresh1_ptr) {
+  v16u8 mask, hev, flat;
+  v16u8 thresh0, b_limit0, limit0, thresh1, b_limit1, limit1;
+  v16u8 p3, p2, p1, p0, q3, q2, q1, q0;
 
-    LD_UB8((src - 4 * pitch), pitch, p3, p2, p1, p0, q0, q1, q2, q3);
-    thresh0 = (v16u8)__msa_fill_b(*thresh0_ptr);
-    thresh1 = (v16u8)__msa_fill_b(*thresh1_ptr);
-    thresh0 = (v16u8)__msa_ilvr_d((v2i64)thresh1, (v2i64)thresh0);
+  LD_UB8((src - 4 * pitch), pitch, p3, p2, p1, p0, q0, q1, q2, q3);
+  thresh0 = (v16u8)__msa_fill_b(*thresh0_ptr);
+  thresh1 = (v16u8)__msa_fill_b(*thresh1_ptr);
+  thresh0 = (v16u8)__msa_ilvr_d((v2i64)thresh1, (v2i64)thresh0);
 
-    b_limit0 = (v16u8)__msa_fill_b(*b_limit0_ptr);
-    b_limit1 = (v16u8)__msa_fill_b(*b_limit1_ptr);
-    b_limit0 = (v16u8)__msa_ilvr_d((v2i64)b_limit1, (v2i64)b_limit0);
+  b_limit0 = (v16u8)__msa_fill_b(*b_limit0_ptr);
+  b_limit1 = (v16u8)__msa_fill_b(*b_limit1_ptr);
+  b_limit0 = (v16u8)__msa_ilvr_d((v2i64)b_limit1, (v2i64)b_limit0);
 
-    limit0 = (v16u8)__msa_fill_b(*limit0_ptr);
-    limit1 = (v16u8)__msa_fill_b(*limit1_ptr);
-    limit0 = (v16u8)__msa_ilvr_d((v2i64)limit1, (v2i64)limit0);
+  limit0 = (v16u8)__msa_fill_b(*limit0_ptr);
+  limit1 = (v16u8)__msa_fill_b(*limit1_ptr);
+  limit0 = (v16u8)__msa_ilvr_d((v2i64)limit1, (v2i64)limit0);
 
-    LPF_MASK_HEV(p3, p2, p1, p0, q0, q1, q2, q3, limit0, b_limit0, thresh0,
-                 hev, mask, flat);
-    VP8_LPF_FILTER4_4W(p1, p0, q0, q1, mask, hev);
+  LPF_MASK_HEV(p3, p2, p1, p0, q0, q1, q2, q3, limit0, b_limit0, thresh0, hev,
+               mask, flat);
+  VP8_LPF_FILTER4_4W(p1, p0, q0, q1, mask, hev);
 
-    ST_UB4(p1, p0, q0, q1, (src - 2 * pitch), pitch);
+  ST_UB4(p1, p0, q0, q1, (src - 2 * pitch), pitch);
 }
 
 static void loop_filter_vertical_4_dual_msa(uint8_t *src, int32_t pitch,
@@ -309,69 +306,67 @@
                                             const uint8_t *thresh0_ptr,
                                             const uint8_t *b_limit1_ptr,
                                             const uint8_t *limit1_ptr,
-                                            const uint8_t *thresh1_ptr)
-{
-    v16u8 mask, hev, flat;
-    v16u8 thresh0, b_limit0, limit0, thresh1, b_limit1, limit1;
-    v16u8 p3, p2, p1, p0, q3, q2, q1, q0;
-    v16u8 row0, row1, row2, row3, row4, row5, row6, row7;
-    v16u8 row8, row9, row10, row11, row12, row13, row14, row15;
-    v8i16 tmp0, tmp1, tmp2, tmp3, tmp4, tmp5;
+                                            const uint8_t *thresh1_ptr) {
+  v16u8 mask, hev, flat;
+  v16u8 thresh0, b_limit0, limit0, thresh1, b_limit1, limit1;
+  v16u8 p3, p2, p1, p0, q3, q2, q1, q0;
+  v16u8 row0, row1, row2, row3, row4, row5, row6, row7;
+  v16u8 row8, row9, row10, row11, row12, row13, row14, row15;
+  v8i16 tmp0, tmp1, tmp2, tmp3, tmp4, tmp5;
 
-    LD_UB8(src - 4, pitch, row0, row1, row2, row3, row4, row5, row6, row7);
-    LD_UB8(src - 4 + (8 * pitch), pitch,
-           row8, row9, row10, row11, row12, row13, row14, row15);
-    TRANSPOSE16x8_UB_UB(row0, row1, row2, row3, row4, row5, row6, row7,
-                        row8, row9, row10, row11, row12, row13, row14, row15,
-                        p3, p2, p1, p0, q0, q1, q2, q3);
+  LD_UB8(src - 4, pitch, row0, row1, row2, row3, row4, row5, row6, row7);
+  LD_UB8(src - 4 + (8 * pitch), pitch, row8, row9, row10, row11, row12, row13,
+         row14, row15);
+  TRANSPOSE16x8_UB_UB(row0, row1, row2, row3, row4, row5, row6, row7, row8,
+                      row9, row10, row11, row12, row13, row14, row15, p3, p2,
+                      p1, p0, q0, q1, q2, q3);
 
-    thresh0 = (v16u8)__msa_fill_b(*thresh0_ptr);
-    thresh1 = (v16u8)__msa_fill_b(*thresh1_ptr);
-    thresh0 = (v16u8)__msa_ilvr_d((v2i64)thresh1, (v2i64)thresh0);
+  thresh0 = (v16u8)__msa_fill_b(*thresh0_ptr);
+  thresh1 = (v16u8)__msa_fill_b(*thresh1_ptr);
+  thresh0 = (v16u8)__msa_ilvr_d((v2i64)thresh1, (v2i64)thresh0);
 
-    b_limit0 = (v16u8)__msa_fill_b(*b_limit0_ptr);
-    b_limit1 = (v16u8)__msa_fill_b(*b_limit1_ptr);
-    b_limit0 = (v16u8)__msa_ilvr_d((v2i64)b_limit1, (v2i64)b_limit0);
+  b_limit0 = (v16u8)__msa_fill_b(*b_limit0_ptr);
+  b_limit1 = (v16u8)__msa_fill_b(*b_limit1_ptr);
+  b_limit0 = (v16u8)__msa_ilvr_d((v2i64)b_limit1, (v2i64)b_limit0);
 
-    limit0 = (v16u8)__msa_fill_b(*limit0_ptr);
-    limit1 = (v16u8)__msa_fill_b(*limit1_ptr);
-    limit0 = (v16u8)__msa_ilvr_d((v2i64)limit1, (v2i64)limit0);
+  limit0 = (v16u8)__msa_fill_b(*limit0_ptr);
+  limit1 = (v16u8)__msa_fill_b(*limit1_ptr);
+  limit0 = (v16u8)__msa_ilvr_d((v2i64)limit1, (v2i64)limit0);
 
-    LPF_MASK_HEV(p3, p2, p1, p0, q0, q1, q2, q3, limit0, b_limit0, thresh0,
-                 hev, mask, flat);
-    VP8_LPF_FILTER4_4W(p1, p0, q0, q1, mask, hev);
-    ILVR_B2_SH(p0, p1, q1, q0, tmp0, tmp1);
-    ILVRL_H2_SH(tmp1, tmp0, tmp2, tmp3);
-    ILVL_B2_SH(p0, p1, q1, q0, tmp0, tmp1);
-    ILVRL_H2_SH(tmp1, tmp0, tmp4, tmp5);
+  LPF_MASK_HEV(p3, p2, p1, p0, q0, q1, q2, q3, limit0, b_limit0, thresh0, hev,
+               mask, flat);
+  VP8_LPF_FILTER4_4W(p1, p0, q0, q1, mask, hev);
+  ILVR_B2_SH(p0, p1, q1, q0, tmp0, tmp1);
+  ILVRL_H2_SH(tmp1, tmp0, tmp2, tmp3);
+  ILVL_B2_SH(p0, p1, q1, q0, tmp0, tmp1);
+  ILVRL_H2_SH(tmp1, tmp0, tmp4, tmp5);
 
-    src -= 2;
-    ST4x8_UB(tmp2, tmp3, src, pitch);
-    src += (8 * pitch);
-    ST4x8_UB(tmp4, tmp5, src, pitch);
+  src -= 2;
+  ST4x8_UB(tmp2, tmp3, src, pitch);
+  src += (8 * pitch);
+  ST4x8_UB(tmp4, tmp5, src, pitch);
 }
 
 static void mbloop_filter_horizontal_edge_y_msa(uint8_t *src, int32_t pitch,
                                                 const uint8_t b_limit_in,
                                                 const uint8_t limit_in,
-                                                const uint8_t thresh_in)
-{
-    uint8_t *temp_src;
-    v16u8 p3, p2, p1, p0, q3, q2, q1, q0;
-    v16u8 mask, hev, flat, thresh, limit, b_limit;
+                                                const uint8_t thresh_in) {
+  uint8_t *temp_src;
+  v16u8 p3, p2, p1, p0, q3, q2, q1, q0;
+  v16u8 mask, hev, flat, thresh, limit, b_limit;
 
-    b_limit = (v16u8)__msa_fill_b(b_limit_in);
-    limit = (v16u8)__msa_fill_b(limit_in);
-    thresh = (v16u8)__msa_fill_b(thresh_in);
-    temp_src = src - (pitch << 2);
-    LD_UB8(temp_src, pitch, p3, p2, p1, p0, q0, q1, q2, q3);
-    LPF_MASK_HEV(p3, p2, p1, p0, q0, q1, q2, q3, limit, b_limit, thresh,
-                 hev, mask, flat);
-    VP8_MBFILTER(p2, p1, p0, q0, q1, q2, mask, hev);
-    temp_src = src - 3 * pitch;
-    ST_UB4(p2, p1, p0, q0, temp_src, pitch);
-    temp_src += (4 * pitch);
-    ST_UB2(q1, q2, temp_src, pitch);
+  b_limit = (v16u8)__msa_fill_b(b_limit_in);
+  limit = (v16u8)__msa_fill_b(limit_in);
+  thresh = (v16u8)__msa_fill_b(thresh_in);
+  temp_src = src - (pitch << 2);
+  LD_UB8(temp_src, pitch, p3, p2, p1, p0, q0, q1, q2, q3);
+  LPF_MASK_HEV(p3, p2, p1, p0, q0, q1, q2, q3, limit, b_limit, thresh, hev,
+               mask, flat);
+  VP8_MBFILTER(p2, p1, p0, q0, q1, q2, mask, hev);
+  temp_src = src - 3 * pitch;
+  ST_UB4(p2, p1, p0, q0, temp_src, pitch);
+  temp_src += (4 * pitch);
+  ST_UB2(q1, q2, temp_src, pitch);
 }
 
 static void mbloop_filter_horizontal_edge_uv_msa(uint8_t *src_u, uint8_t *src_v,
@@ -378,122 +373,119 @@
                                                  int32_t pitch,
                                                  const uint8_t b_limit_in,
                                                  const uint8_t limit_in,
-                                                 const uint8_t thresh_in)
-{
-    uint8_t *temp_src;
-    uint64_t p2_d, p1_d, p0_d, q0_d, q1_d, q2_d;
-    v16u8 p3, p2, p1, p0, q3, q2, q1, q0;
-    v16u8 mask, hev, flat, thresh, limit, b_limit;
-    v16u8 p3_u, p2_u, p1_u, p0_u, q3_u, q2_u, q1_u, q0_u;
-    v16u8 p3_v, p2_v, p1_v, p0_v, q3_v, q2_v, q1_v, q0_v;
+                                                 const uint8_t thresh_in) {
+  uint8_t *temp_src;
+  uint64_t p2_d, p1_d, p0_d, q0_d, q1_d, q2_d;
+  v16u8 p3, p2, p1, p0, q3, q2, q1, q0;
+  v16u8 mask, hev, flat, thresh, limit, b_limit;
+  v16u8 p3_u, p2_u, p1_u, p0_u, q3_u, q2_u, q1_u, q0_u;
+  v16u8 p3_v, p2_v, p1_v, p0_v, q3_v, q2_v, q1_v, q0_v;
 
-    b_limit = (v16u8)__msa_fill_b(b_limit_in);
-    limit = (v16u8)__msa_fill_b(limit_in);
-    thresh = (v16u8)__msa_fill_b(thresh_in);
+  b_limit = (v16u8)__msa_fill_b(b_limit_in);
+  limit = (v16u8)__msa_fill_b(limit_in);
+  thresh = (v16u8)__msa_fill_b(thresh_in);
 
-    temp_src = src_u - (pitch << 2);
-    LD_UB8(temp_src, pitch, p3_u, p2_u, p1_u, p0_u, q0_u, q1_u, q2_u, q3_u);
-    temp_src = src_v - (pitch << 2);
-    LD_UB8(temp_src, pitch, p3_v, p2_v, p1_v, p0_v, q0_v, q1_v, q2_v, q3_v);
+  temp_src = src_u - (pitch << 2);
+  LD_UB8(temp_src, pitch, p3_u, p2_u, p1_u, p0_u, q0_u, q1_u, q2_u, q3_u);
+  temp_src = src_v - (pitch << 2);
+  LD_UB8(temp_src, pitch, p3_v, p2_v, p1_v, p0_v, q0_v, q1_v, q2_v, q3_v);
 
-    ILVR_D4_UB(p3_v, p3_u, p2_v, p2_u, p1_v, p1_u, p0_v, p0_u, p3, p2, p1, p0);
-    ILVR_D4_UB(q0_v, q0_u, q1_v, q1_u, q2_v, q2_u, q3_v, q3_u, q0, q1, q2, q3);
-    LPF_MASK_HEV(p3, p2, p1, p0, q0, q1, q2, q3, limit, b_limit, thresh,
-                 hev, mask, flat);
-    VP8_MBFILTER(p2, p1, p0, q0, q1, q2, mask, hev);
+  ILVR_D4_UB(p3_v, p3_u, p2_v, p2_u, p1_v, p1_u, p0_v, p0_u, p3, p2, p1, p0);
+  ILVR_D4_UB(q0_v, q0_u, q1_v, q1_u, q2_v, q2_u, q3_v, q3_u, q0, q1, q2, q3);
+  LPF_MASK_HEV(p3, p2, p1, p0, q0, q1, q2, q3, limit, b_limit, thresh, hev,
+               mask, flat);
+  VP8_MBFILTER(p2, p1, p0, q0, q1, q2, mask, hev);
 
-    p2_d = __msa_copy_u_d((v2i64)p2, 0);
-    p1_d = __msa_copy_u_d((v2i64)p1, 0);
-    p0_d = __msa_copy_u_d((v2i64)p0, 0);
-    q0_d = __msa_copy_u_d((v2i64)q0, 0);
-    q1_d = __msa_copy_u_d((v2i64)q1, 0);
-    q2_d = __msa_copy_u_d((v2i64)q2, 0);
-    src_u -= (pitch * 3);
-    SD4(p2_d, p1_d, p0_d, q0_d, src_u, pitch);
-    src_u += 4 * pitch;
-    SD(q1_d, src_u);
-    src_u += pitch;
-    SD(q2_d, src_u);
+  p2_d = __msa_copy_u_d((v2i64)p2, 0);
+  p1_d = __msa_copy_u_d((v2i64)p1, 0);
+  p0_d = __msa_copy_u_d((v2i64)p0, 0);
+  q0_d = __msa_copy_u_d((v2i64)q0, 0);
+  q1_d = __msa_copy_u_d((v2i64)q1, 0);
+  q2_d = __msa_copy_u_d((v2i64)q2, 0);
+  src_u -= (pitch * 3);
+  SD4(p2_d, p1_d, p0_d, q0_d, src_u, pitch);
+  src_u += 4 * pitch;
+  SD(q1_d, src_u);
+  src_u += pitch;
+  SD(q2_d, src_u);
 
-    p2_d = __msa_copy_u_d((v2i64)p2, 1);
-    p1_d = __msa_copy_u_d((v2i64)p1, 1);
-    p0_d = __msa_copy_u_d((v2i64)p0, 1);
-    q0_d = __msa_copy_u_d((v2i64)q0, 1);
-    q1_d = __msa_copy_u_d((v2i64)q1, 1);
-    q2_d = __msa_copy_u_d((v2i64)q2, 1);
-    src_v -= (pitch * 3);
-    SD4(p2_d, p1_d, p0_d, q0_d, src_v, pitch);
-    src_v += 4 * pitch;
-    SD(q1_d, src_v);
-    src_v += pitch;
-    SD(q2_d, src_v);
+  p2_d = __msa_copy_u_d((v2i64)p2, 1);
+  p1_d = __msa_copy_u_d((v2i64)p1, 1);
+  p0_d = __msa_copy_u_d((v2i64)p0, 1);
+  q0_d = __msa_copy_u_d((v2i64)q0, 1);
+  q1_d = __msa_copy_u_d((v2i64)q1, 1);
+  q2_d = __msa_copy_u_d((v2i64)q2, 1);
+  src_v -= (pitch * 3);
+  SD4(p2_d, p1_d, p0_d, q0_d, src_v, pitch);
+  src_v += 4 * pitch;
+  SD(q1_d, src_v);
+  src_v += pitch;
+  SD(q2_d, src_v);
 }
 
 static void mbloop_filter_vertical_edge_y_msa(uint8_t *src, int32_t pitch,
                                               const uint8_t b_limit_in,
                                               const uint8_t limit_in,
-                                              const uint8_t thresh_in)
-{
-    uint8_t *temp_src;
-    v16u8 p3, p2, p1, p0, q3, q2, q1, q0;
-    v16u8 mask, hev, flat, thresh, limit, b_limit;
-    v16u8 row0, row1, row2, row3, row4, row5, row6, row7, row8;
-    v16u8 row9, row10, row11, row12, row13, row14, row15;
-    v8i16 tmp0, tmp1, tmp2, tmp3, tmp4, tmp5, tmp6, tmp7;
+                                              const uint8_t thresh_in) {
+  uint8_t *temp_src;
+  v16u8 p3, p2, p1, p0, q3, q2, q1, q0;
+  v16u8 mask, hev, flat, thresh, limit, b_limit;
+  v16u8 row0, row1, row2, row3, row4, row5, row6, row7, row8;
+  v16u8 row9, row10, row11, row12, row13, row14, row15;
+  v8i16 tmp0, tmp1, tmp2, tmp3, tmp4, tmp5, tmp6, tmp7;
 
-    b_limit = (v16u8)__msa_fill_b(b_limit_in);
-    limit = (v16u8)__msa_fill_b(limit_in);
-    thresh = (v16u8)__msa_fill_b(thresh_in);
-    temp_src = src - 4;
-    LD_UB8(temp_src, pitch, row0, row1, row2, row3, row4, row5, row6, row7);
-    temp_src += (8 * pitch);
-    LD_UB8(temp_src, pitch,
-           row8, row9, row10, row11, row12, row13, row14, row15);
-    TRANSPOSE16x8_UB_UB(row0, row1, row2, row3, row4, row5, row6, row7,
-                        row8, row9, row10, row11, row12, row13, row14, row15,
-                        p3, p2, p1, p0, q0, q1, q2, q3);
+  b_limit = (v16u8)__msa_fill_b(b_limit_in);
+  limit = (v16u8)__msa_fill_b(limit_in);
+  thresh = (v16u8)__msa_fill_b(thresh_in);
+  temp_src = src - 4;
+  LD_UB8(temp_src, pitch, row0, row1, row2, row3, row4, row5, row6, row7);
+  temp_src += (8 * pitch);
+  LD_UB8(temp_src, pitch, row8, row9, row10, row11, row12, row13, row14, row15);
+  TRANSPOSE16x8_UB_UB(row0, row1, row2, row3, row4, row5, row6, row7, row8,
+                      row9, row10, row11, row12, row13, row14, row15, p3, p2,
+                      p1, p0, q0, q1, q2, q3);
 
-    LPF_MASK_HEV(p3, p2, p1, p0, q0, q1, q2, q3, limit, b_limit, thresh,
-                 hev, mask, flat);
-    VP8_MBFILTER(p2, p1, p0, q0, q1, q2, mask, hev);
-    ILVR_B2_SH(p1, p2, q0, p0, tmp0, tmp1);
-    ILVRL_H2_SH(tmp1, tmp0, tmp3, tmp4);
-    ILVL_B2_SH(p1, p2, q0, p0, tmp0, tmp1);
-    ILVRL_H2_SH(tmp1, tmp0, tmp6, tmp7);
-    ILVRL_B2_SH(q2, q1, tmp2, tmp5);
+  LPF_MASK_HEV(p3, p2, p1, p0, q0, q1, q2, q3, limit, b_limit, thresh, hev,
+               mask, flat);
+  VP8_MBFILTER(p2, p1, p0, q0, q1, q2, mask, hev);
+  ILVR_B2_SH(p1, p2, q0, p0, tmp0, tmp1);
+  ILVRL_H2_SH(tmp1, tmp0, tmp3, tmp4);
+  ILVL_B2_SH(p1, p2, q0, p0, tmp0, tmp1);
+  ILVRL_H2_SH(tmp1, tmp0, tmp6, tmp7);
+  ILVRL_B2_SH(q2, q1, tmp2, tmp5);
 
-    temp_src = src - 3;
-    VP8_ST6x1_UB(tmp3, 0, tmp2, 0, temp_src, 4);
-    temp_src += pitch;
-    VP8_ST6x1_UB(tmp3, 1, tmp2, 1, temp_src, 4);
-    temp_src += pitch;
-    VP8_ST6x1_UB(tmp3, 2, tmp2, 2, temp_src, 4);
-    temp_src += pitch;
-    VP8_ST6x1_UB(tmp3, 3, tmp2, 3, temp_src, 4);
-    temp_src += pitch;
-    VP8_ST6x1_UB(tmp4, 0, tmp2, 4, temp_src, 4);
-    temp_src += pitch;
-    VP8_ST6x1_UB(tmp4, 1, tmp2, 5, temp_src, 4);
-    temp_src += pitch;
-    VP8_ST6x1_UB(tmp4, 2, tmp2, 6, temp_src, 4);
-    temp_src += pitch;
-    VP8_ST6x1_UB(tmp4, 3, tmp2, 7, temp_src, 4);
-    temp_src += pitch;
-    VP8_ST6x1_UB(tmp6, 0, tmp5, 0, temp_src, 4);
-    temp_src += pitch;
-    VP8_ST6x1_UB(tmp6, 1, tmp5, 1, temp_src, 4);
-    temp_src += pitch;
-    VP8_ST6x1_UB(tmp6, 2, tmp5, 2, temp_src, 4);
-    temp_src += pitch;
-    VP8_ST6x1_UB(tmp6, 3, tmp5, 3, temp_src, 4);
-    temp_src += pitch;
-    VP8_ST6x1_UB(tmp7, 0, tmp5, 4, temp_src, 4);
-    temp_src += pitch;
-    VP8_ST6x1_UB(tmp7, 1, tmp5, 5, temp_src, 4);
-    temp_src += pitch;
-    VP8_ST6x1_UB(tmp7, 2, tmp5, 6, temp_src, 4);
-    temp_src += pitch;
-    VP8_ST6x1_UB(tmp7, 3, tmp5, 7, temp_src, 4);
+  temp_src = src - 3;
+  VP8_ST6x1_UB(tmp3, 0, tmp2, 0, temp_src, 4);
+  temp_src += pitch;
+  VP8_ST6x1_UB(tmp3, 1, tmp2, 1, temp_src, 4);
+  temp_src += pitch;
+  VP8_ST6x1_UB(tmp3, 2, tmp2, 2, temp_src, 4);
+  temp_src += pitch;
+  VP8_ST6x1_UB(tmp3, 3, tmp2, 3, temp_src, 4);
+  temp_src += pitch;
+  VP8_ST6x1_UB(tmp4, 0, tmp2, 4, temp_src, 4);
+  temp_src += pitch;
+  VP8_ST6x1_UB(tmp4, 1, tmp2, 5, temp_src, 4);
+  temp_src += pitch;
+  VP8_ST6x1_UB(tmp4, 2, tmp2, 6, temp_src, 4);
+  temp_src += pitch;
+  VP8_ST6x1_UB(tmp4, 3, tmp2, 7, temp_src, 4);
+  temp_src += pitch;
+  VP8_ST6x1_UB(tmp6, 0, tmp5, 0, temp_src, 4);
+  temp_src += pitch;
+  VP8_ST6x1_UB(tmp6, 1, tmp5, 1, temp_src, 4);
+  temp_src += pitch;
+  VP8_ST6x1_UB(tmp6, 2, tmp5, 2, temp_src, 4);
+  temp_src += pitch;
+  VP8_ST6x1_UB(tmp6, 3, tmp5, 3, temp_src, 4);
+  temp_src += pitch;
+  VP8_ST6x1_UB(tmp7, 0, tmp5, 4, temp_src, 4);
+  temp_src += pitch;
+  VP8_ST6x1_UB(tmp7, 1, tmp5, 5, temp_src, 4);
+  temp_src += pitch;
+  VP8_ST6x1_UB(tmp7, 2, tmp5, 6, temp_src, 4);
+  temp_src += pitch;
+  VP8_ST6x1_UB(tmp7, 3, tmp5, 7, temp_src, 4);
 }
 
 static void mbloop_filter_vertical_edge_uv_msa(uint8_t *src_u, uint8_t *src_v,
@@ -500,115 +492,111 @@
                                                int32_t pitch,
                                                const uint8_t b_limit_in,
                                                const uint8_t limit_in,
-                                               const uint8_t thresh_in)
-{
-    v16u8 p3, p2, p1, p0, q3, q2, q1, q0;
-    v16u8 mask, hev, flat, thresh, limit, b_limit;
-    v16u8 row0, row1, row2, row3, row4, row5, row6, row7, row8;
-    v16u8 row9, row10, row11, row12, row13, row14, row15;
-    v8i16 tmp0, tmp1, tmp2, tmp3, tmp4, tmp5, tmp6, tmp7;
+                                               const uint8_t thresh_in) {
+  v16u8 p3, p2, p1, p0, q3, q2, q1, q0;
+  v16u8 mask, hev, flat, thresh, limit, b_limit;
+  v16u8 row0, row1, row2, row3, row4, row5, row6, row7, row8;
+  v16u8 row9, row10, row11, row12, row13, row14, row15;
+  v8i16 tmp0, tmp1, tmp2, tmp3, tmp4, tmp5, tmp6, tmp7;
 
-    b_limit = (v16u8)__msa_fill_b(b_limit_in);
-    limit = (v16u8)__msa_fill_b(limit_in);
-    thresh = (v16u8)__msa_fill_b(thresh_in);
+  b_limit = (v16u8)__msa_fill_b(b_limit_in);
+  limit = (v16u8)__msa_fill_b(limit_in);
+  thresh = (v16u8)__msa_fill_b(thresh_in);
 
-    LD_UB8(src_u - 4, pitch, row0, row1, row2, row3, row4, row5, row6, row7);
-    LD_UB8(src_v - 4, pitch,
-           row8, row9, row10, row11, row12, row13, row14, row15);
-    TRANSPOSE16x8_UB_UB(row0, row1, row2, row3, row4, row5, row6, row7,
-                        row8, row9, row10, row11, row12, row13, row14, row15,
-                        p3, p2, p1, p0, q0, q1, q2, q3);
+  LD_UB8(src_u - 4, pitch, row0, row1, row2, row3, row4, row5, row6, row7);
+  LD_UB8(src_v - 4, pitch, row8, row9, row10, row11, row12, row13, row14,
+         row15);
+  TRANSPOSE16x8_UB_UB(row0, row1, row2, row3, row4, row5, row6, row7, row8,
+                      row9, row10, row11, row12, row13, row14, row15, p3, p2,
+                      p1, p0, q0, q1, q2, q3);
 
-    LPF_MASK_HEV(p3, p2, p1, p0, q0, q1, q2, q3, limit, b_limit, thresh,
-                 hev, mask, flat);
-    VP8_MBFILTER(p2, p1, p0, q0, q1, q2, mask, hev);
+  LPF_MASK_HEV(p3, p2, p1, p0, q0, q1, q2, q3, limit, b_limit, thresh, hev,
+               mask, flat);
+  VP8_MBFILTER(p2, p1, p0, q0, q1, q2, mask, hev);
 
-    ILVR_B2_SH(p1, p2, q0, p0, tmp0, tmp1);
-    ILVRL_H2_SH(tmp1, tmp0, tmp3, tmp4);
-    ILVL_B2_SH(p1, p2, q0, p0, tmp0, tmp1);
-    ILVRL_H2_SH(tmp1, tmp0, tmp6, tmp7);
-    ILVRL_B2_SH(q2, q1, tmp2, tmp5);
+  ILVR_B2_SH(p1, p2, q0, p0, tmp0, tmp1);
+  ILVRL_H2_SH(tmp1, tmp0, tmp3, tmp4);
+  ILVL_B2_SH(p1, p2, q0, p0, tmp0, tmp1);
+  ILVRL_H2_SH(tmp1, tmp0, tmp6, tmp7);
+  ILVRL_B2_SH(q2, q1, tmp2, tmp5);
 
-    src_u -= 3;
-    VP8_ST6x1_UB(tmp3, 0, tmp2, 0, src_u, 4);
-    src_u += pitch;
-    VP8_ST6x1_UB(tmp3, 1, tmp2, 1, src_u, 4);
-    src_u += pitch;
-    VP8_ST6x1_UB(tmp3, 2, tmp2, 2, src_u, 4);
-    src_u += pitch;
-    VP8_ST6x1_UB(tmp3, 3, tmp2, 3, src_u, 4);
-    src_u += pitch;
-    VP8_ST6x1_UB(tmp4, 0, tmp2, 4, src_u, 4);
-    src_u += pitch;
-    VP8_ST6x1_UB(tmp4, 1, tmp2, 5, src_u, 4);
-    src_u += pitch;
-    VP8_ST6x1_UB(tmp4, 2, tmp2, 6, src_u, 4);
-    src_u += pitch;
-    VP8_ST6x1_UB(tmp4, 3, tmp2, 7, src_u, 4);
+  src_u -= 3;
+  VP8_ST6x1_UB(tmp3, 0, tmp2, 0, src_u, 4);
+  src_u += pitch;
+  VP8_ST6x1_UB(tmp3, 1, tmp2, 1, src_u, 4);
+  src_u += pitch;
+  VP8_ST6x1_UB(tmp3, 2, tmp2, 2, src_u, 4);
+  src_u += pitch;
+  VP8_ST6x1_UB(tmp3, 3, tmp2, 3, src_u, 4);
+  src_u += pitch;
+  VP8_ST6x1_UB(tmp4, 0, tmp2, 4, src_u, 4);
+  src_u += pitch;
+  VP8_ST6x1_UB(tmp4, 1, tmp2, 5, src_u, 4);
+  src_u += pitch;
+  VP8_ST6x1_UB(tmp4, 2, tmp2, 6, src_u, 4);
+  src_u += pitch;
+  VP8_ST6x1_UB(tmp4, 3, tmp2, 7, src_u, 4);
 
-    src_v -= 3;
-    VP8_ST6x1_UB(tmp6, 0, tmp5, 0, src_v, 4);
-    src_v += pitch;
-    VP8_ST6x1_UB(tmp6, 1, tmp5, 1, src_v, 4);
-    src_v += pitch;
-    VP8_ST6x1_UB(tmp6, 2, tmp5, 2, src_v, 4);
-    src_v += pitch;
-    VP8_ST6x1_UB(tmp6, 3, tmp5, 3, src_v, 4);
-    src_v += pitch;
-    VP8_ST6x1_UB(tmp7, 0, tmp5, 4, src_v, 4);
-    src_v += pitch;
-    VP8_ST6x1_UB(tmp7, 1, tmp5, 5, src_v, 4);
-    src_v += pitch;
-    VP8_ST6x1_UB(tmp7, 2, tmp5, 6, src_v, 4);
-    src_v += pitch;
-    VP8_ST6x1_UB(tmp7, 3, tmp5, 7, src_v, 4);
+  src_v -= 3;
+  VP8_ST6x1_UB(tmp6, 0, tmp5, 0, src_v, 4);
+  src_v += pitch;
+  VP8_ST6x1_UB(tmp6, 1, tmp5, 1, src_v, 4);
+  src_v += pitch;
+  VP8_ST6x1_UB(tmp6, 2, tmp5, 2, src_v, 4);
+  src_v += pitch;
+  VP8_ST6x1_UB(tmp6, 3, tmp5, 3, src_v, 4);
+  src_v += pitch;
+  VP8_ST6x1_UB(tmp7, 0, tmp5, 4, src_v, 4);
+  src_v += pitch;
+  VP8_ST6x1_UB(tmp7, 1, tmp5, 5, src_v, 4);
+  src_v += pitch;
+  VP8_ST6x1_UB(tmp7, 2, tmp5, 6, src_v, 4);
+  src_v += pitch;
+  VP8_ST6x1_UB(tmp7, 3, tmp5, 7, src_v, 4);
 }
 
 void vp8_loop_filter_simple_horizontal_edge_msa(uint8_t *src, int32_t pitch,
-                                                const uint8_t *b_limit_ptr)
-{
-    v16u8 p1, p0, q1, q0;
-    v16u8 mask, b_limit;
+                                                const uint8_t *b_limit_ptr) {
+  v16u8 p1, p0, q1, q0;
+  v16u8 mask, b_limit;
 
-    b_limit = (v16u8)__msa_fill_b(*b_limit_ptr);
-    LD_UB4(src - (pitch << 1), pitch, p1, p0, q0, q1);
-    VP8_SIMPLE_MASK(p1, p0, q0, q1, b_limit, mask);
-    VP8_SIMPLE_FILT(p1, p0, q0, q1, mask);
-    ST_UB2(p0, q0, (src - pitch), pitch);
+  b_limit = (v16u8)__msa_fill_b(*b_limit_ptr);
+  LD_UB4(src - (pitch << 1), pitch, p1, p0, q0, q1);
+  VP8_SIMPLE_MASK(p1, p0, q0, q1, b_limit, mask);
+  VP8_SIMPLE_FILT(p1, p0, q0, q1, mask);
+  ST_UB2(p0, q0, (src - pitch), pitch);
 }
 
 void vp8_loop_filter_simple_vertical_edge_msa(uint8_t *src, int32_t pitch,
-                                              const uint8_t *b_limit_ptr)
-{
-    uint8_t *temp_src;
-    v16u8 p1, p0, q1, q0;
-    v16u8 mask, b_limit;
-    v16u8 row0, row1, row2, row3, row4, row5, row6, row7, row8;
-    v16u8 row9, row10, row11, row12, row13, row14, row15;
-    v8i16 tmp0, tmp1;
+                                              const uint8_t *b_limit_ptr) {
+  uint8_t *temp_src;
+  v16u8 p1, p0, q1, q0;
+  v16u8 mask, b_limit;
+  v16u8 row0, row1, row2, row3, row4, row5, row6, row7, row8;
+  v16u8 row9, row10, row11, row12, row13, row14, row15;
+  v8i16 tmp0, tmp1;
 
-    b_limit = (v16u8)__msa_fill_b(*b_limit_ptr);
-    temp_src = src - 2;
-    LD_UB8(temp_src, pitch, row0, row1, row2, row3, row4, row5, row6, row7);
-    temp_src += (8 * pitch);
-    LD_UB8(temp_src, pitch,
-           row8, row9, row10, row11, row12, row13, row14, row15);
-    TRANSPOSE16x4_UB_UB(row0, row1, row2, row3, row4, row5, row6, row7,
-                        row8, row9, row10, row11, row12, row13, row14, row15,
-                        p1, p0, q0, q1);
-    VP8_SIMPLE_MASK(p1, p0, q0, q1, b_limit, mask);
-    VP8_SIMPLE_FILT(p1, p0, q0, q1, mask);
-    ILVRL_B2_SH(q0, p0, tmp1, tmp0);
+  b_limit = (v16u8)__msa_fill_b(*b_limit_ptr);
+  temp_src = src - 2;
+  LD_UB8(temp_src, pitch, row0, row1, row2, row3, row4, row5, row6, row7);
+  temp_src += (8 * pitch);
+  LD_UB8(temp_src, pitch, row8, row9, row10, row11, row12, row13, row14, row15);
+  TRANSPOSE16x4_UB_UB(row0, row1, row2, row3, row4, row5, row6, row7, row8,
+                      row9, row10, row11, row12, row13, row14, row15, p1, p0,
+                      q0, q1);
+  VP8_SIMPLE_MASK(p1, p0, q0, q1, b_limit, mask);
+  VP8_SIMPLE_FILT(p1, p0, q0, q1, mask);
+  ILVRL_B2_SH(q0, p0, tmp1, tmp0);
 
-    src -= 1;
-    ST2x4_UB(tmp1, 0, src, pitch);
-    src += 4 * pitch;
-    ST2x4_UB(tmp1, 4, src, pitch);
-    src += 4 * pitch;
-    ST2x4_UB(tmp0, 0, src, pitch);
-    src += 4 * pitch;
-    ST2x4_UB(tmp0, 4, src, pitch);
-    src += 4 * pitch;
+  src -= 1;
+  ST2x4_UB(tmp1, 0, src, pitch);
+  src += 4 * pitch;
+  ST2x4_UB(tmp1, 4, src, pitch);
+  src += 4 * pitch;
+  ST2x4_UB(tmp0, 0, src, pitch);
+  src += 4 * pitch;
+  ST2x4_UB(tmp0, 4, src, pitch);
+  src += 4 * pitch;
 }
 
 static void loop_filter_horizontal_edge_uv_msa(uint8_t *src_u, uint8_t *src_v,
@@ -615,44 +603,43 @@
                                                int32_t pitch,
                                                const uint8_t b_limit_in,
                                                const uint8_t limit_in,
-                                               const uint8_t thresh_in)
-{
-    uint64_t p1_d, p0_d, q0_d, q1_d;
-    v16u8 p3, p2, p1, p0, q3, q2, q1, q0;
-    v16u8 mask, hev, flat, thresh, limit, b_limit;
-    v16u8 p3_u, p2_u, p1_u, p0_u, q3_u, q2_u, q1_u, q0_u;
-    v16u8 p3_v, p2_v, p1_v, p0_v, q3_v, q2_v, q1_v, q0_v;
+                                               const uint8_t thresh_in) {
+  uint64_t p1_d, p0_d, q0_d, q1_d;
+  v16u8 p3, p2, p1, p0, q3, q2, q1, q0;
+  v16u8 mask, hev, flat, thresh, limit, b_limit;
+  v16u8 p3_u, p2_u, p1_u, p0_u, q3_u, q2_u, q1_u, q0_u;
+  v16u8 p3_v, p2_v, p1_v, p0_v, q3_v, q2_v, q1_v, q0_v;
 
-    thresh = (v16u8)__msa_fill_b(thresh_in);
-    limit = (v16u8)__msa_fill_b(limit_in);
-    b_limit = (v16u8)__msa_fill_b(b_limit_in);
+  thresh = (v16u8)__msa_fill_b(thresh_in);
+  limit = (v16u8)__msa_fill_b(limit_in);
+  b_limit = (v16u8)__msa_fill_b(b_limit_in);
 
-    src_u = src_u - (pitch << 2);
-    LD_UB8(src_u, pitch, p3_u, p2_u, p1_u, p0_u, q0_u, q1_u, q2_u, q3_u);
-    src_u += (5 * pitch);
-    src_v = src_v - (pitch << 2);
-    LD_UB8(src_v, pitch, p3_v, p2_v, p1_v, p0_v, q0_v, q1_v, q2_v, q3_v);
-    src_v += (5 * pitch);
+  src_u = src_u - (pitch << 2);
+  LD_UB8(src_u, pitch, p3_u, p2_u, p1_u, p0_u, q0_u, q1_u, q2_u, q3_u);
+  src_u += (5 * pitch);
+  src_v = src_v - (pitch << 2);
+  LD_UB8(src_v, pitch, p3_v, p2_v, p1_v, p0_v, q0_v, q1_v, q2_v, q3_v);
+  src_v += (5 * pitch);
 
-    /* right 8 element of p3 are u pixel and
-       left 8 element of p3 are v pixel */
-    ILVR_D4_UB(p3_v, p3_u, p2_v, p2_u, p1_v, p1_u, p0_v, p0_u, p3, p2, p1, p0);
-    ILVR_D4_UB(q0_v, q0_u, q1_v, q1_u, q2_v, q2_u, q3_v, q3_u, q0, q1, q2, q3);
-    LPF_MASK_HEV(p3, p2, p1, p0, q0, q1, q2, q3, limit, b_limit, thresh,
-                 hev, mask, flat);
-    VP8_LPF_FILTER4_4W(p1, p0, q0, q1, mask, hev);
+  /* right 8 element of p3 are u pixel and
+     left 8 element of p3 are v pixel */
+  ILVR_D4_UB(p3_v, p3_u, p2_v, p2_u, p1_v, p1_u, p0_v, p0_u, p3, p2, p1, p0);
+  ILVR_D4_UB(q0_v, q0_u, q1_v, q1_u, q2_v, q2_u, q3_v, q3_u, q0, q1, q2, q3);
+  LPF_MASK_HEV(p3, p2, p1, p0, q0, q1, q2, q3, limit, b_limit, thresh, hev,
+               mask, flat);
+  VP8_LPF_FILTER4_4W(p1, p0, q0, q1, mask, hev);
 
-    p1_d = __msa_copy_u_d((v2i64)p1, 0);
-    p0_d = __msa_copy_u_d((v2i64)p0, 0);
-    q0_d = __msa_copy_u_d((v2i64)q0, 0);
-    q1_d = __msa_copy_u_d((v2i64)q1, 0);
-    SD4(q1_d, q0_d, p0_d, p1_d, src_u, (- pitch));
+  p1_d = __msa_copy_u_d((v2i64)p1, 0);
+  p0_d = __msa_copy_u_d((v2i64)p0, 0);
+  q0_d = __msa_copy_u_d((v2i64)q0, 0);
+  q1_d = __msa_copy_u_d((v2i64)q1, 0);
+  SD4(q1_d, q0_d, p0_d, p1_d, src_u, (-pitch));
 
-    p1_d = __msa_copy_u_d((v2i64)p1, 1);
-    p0_d = __msa_copy_u_d((v2i64)p0, 1);
-    q0_d = __msa_copy_u_d((v2i64)q0, 1);
-    q1_d = __msa_copy_u_d((v2i64)q1, 1);
-    SD4(q1_d, q0_d, p0_d, p1_d, src_v, (- pitch));
+  p1_d = __msa_copy_u_d((v2i64)p1, 1);
+  p0_d = __msa_copy_u_d((v2i64)p0, 1);
+  q0_d = __msa_copy_u_d((v2i64)q0, 1);
+  q1_d = __msa_copy_u_d((v2i64)q1, 1);
+  SD4(q1_d, q0_d, p0_d, p1_d, src_v, (-pitch));
 }
 
 static void loop_filter_vertical_edge_uv_msa(uint8_t *src_u, uint8_t *src_v,
@@ -659,168 +646,127 @@
                                              int32_t pitch,
                                              const uint8_t b_limit_in,
                                              const uint8_t limit_in,
-                                             const uint8_t thresh_in)
-{
-    uint8_t *temp_src_u, *temp_src_v;
-    v16u8 p3, p2, p1, p0, q3, q2, q1, q0;
-    v16u8 mask, hev, flat, thresh, limit, b_limit;
-    v16u8 row0, row1, row2, row3, row4, row5, row6, row7, row8;
-    v16u8 row9, row10, row11, row12, row13, row14, row15;
-    v4i32 tmp0, tmp1, tmp2, tmp3, tmp4, tmp5;
+                                             const uint8_t thresh_in) {
+  uint8_t *temp_src_u, *temp_src_v;
+  v16u8 p3, p2, p1, p0, q3, q2, q1, q0;
+  v16u8 mask, hev, flat, thresh, limit, b_limit;
+  v16u8 row0, row1, row2, row3, row4, row5, row6, row7, row8;
+  v16u8 row9, row10, row11, row12, row13, row14, row15;
+  v4i32 tmp0, tmp1, tmp2, tmp3, tmp4, tmp5;
 
-    thresh = (v16u8)__msa_fill_b(thresh_in);
-    limit = (v16u8)__msa_fill_b(limit_in);
-    b_limit = (v16u8)__msa_fill_b(b_limit_in);
+  thresh = (v16u8)__msa_fill_b(thresh_in);
+  limit = (v16u8)__msa_fill_b(limit_in);
+  b_limit = (v16u8)__msa_fill_b(b_limit_in);
 
-    LD_UB8(src_u - 4, pitch, row0, row1, row2, row3, row4, row5, row6, row7);
-    LD_UB8(src_v - 4, pitch,
-           row8, row9, row10, row11, row12, row13, row14, row15);
-    TRANSPOSE16x8_UB_UB(row0, row1, row2, row3, row4, row5, row6, row7,
-                        row8, row9, row10, row11, row12, row13, row14, row15,
-                        p3, p2, p1, p0, q0, q1, q2, q3);
+  LD_UB8(src_u - 4, pitch, row0, row1, row2, row3, row4, row5, row6, row7);
+  LD_UB8(src_v - 4, pitch, row8, row9, row10, row11, row12, row13, row14,
+         row15);
+  TRANSPOSE16x8_UB_UB(row0, row1, row2, row3, row4, row5, row6, row7, row8,
+                      row9, row10, row11, row12, row13, row14, row15, p3, p2,
+                      p1, p0, q0, q1, q2, q3);
 
-    LPF_MASK_HEV(p3, p2, p1, p0, q0, q1, q2, q3, limit, b_limit, thresh,
-                 hev, mask, flat);
-    VP8_LPF_FILTER4_4W(p1, p0, q0, q1, mask, hev);
-    ILVR_B2_SW(p0, p1, q1, q0, tmp0, tmp1);
-    ILVRL_H2_SW(tmp1, tmp0, tmp2, tmp3);
-    tmp0 = (v4i32)__msa_ilvl_b((v16i8)p0, (v16i8)p1);
-    tmp1 = (v4i32)__msa_ilvl_b((v16i8)q1, (v16i8)q0);
-    ILVRL_H2_SW(tmp1, tmp0, tmp4, tmp5);
+  LPF_MASK_HEV(p3, p2, p1, p0, q0, q1, q2, q3, limit, b_limit, thresh, hev,
+               mask, flat);
+  VP8_LPF_FILTER4_4W(p1, p0, q0, q1, mask, hev);
+  ILVR_B2_SW(p0, p1, q1, q0, tmp0, tmp1);
+  ILVRL_H2_SW(tmp1, tmp0, tmp2, tmp3);
+  tmp0 = (v4i32)__msa_ilvl_b((v16i8)p0, (v16i8)p1);
+  tmp1 = (v4i32)__msa_ilvl_b((v16i8)q1, (v16i8)q0);
+  ILVRL_H2_SW(tmp1, tmp0, tmp4, tmp5);
 
-    temp_src_u = src_u - 2;
-    ST4x4_UB(tmp2, tmp2, 0, 1, 2, 3, temp_src_u, pitch);
-    temp_src_u += 4 * pitch;
-    ST4x4_UB(tmp3, tmp3, 0, 1, 2, 3, temp_src_u, pitch);
+  temp_src_u = src_u - 2;
+  ST4x4_UB(tmp2, tmp2, 0, 1, 2, 3, temp_src_u, pitch);
+  temp_src_u += 4 * pitch;
+  ST4x4_UB(tmp3, tmp3, 0, 1, 2, 3, temp_src_u, pitch);
 
-    temp_src_v = src_v - 2;
-    ST4x4_UB(tmp4, tmp4, 0, 1, 2, 3, temp_src_v, pitch);
-    temp_src_v += 4 * pitch;
-    ST4x4_UB(tmp5, tmp5, 0, 1, 2, 3, temp_src_v, pitch);
+  temp_src_v = src_v - 2;
+  ST4x4_UB(tmp4, tmp4, 0, 1, 2, 3, temp_src_v, pitch);
+  temp_src_v += 4 * pitch;
+  ST4x4_UB(tmp5, tmp5, 0, 1, 2, 3, temp_src_v, pitch);
 }
 
-void vp8_loop_filter_mbh_msa(uint8_t *src_y, uint8_t *src_u,
-                             uint8_t *src_v, int32_t pitch_y,
-                             int32_t pitch_u_v,
-                             loop_filter_info *lpf_info_ptr)
-{
-    mbloop_filter_horizontal_edge_y_msa(src_y, pitch_y,
-                                        *lpf_info_ptr->mblim,
-                                        *lpf_info_ptr->lim,
-                                        *lpf_info_ptr->hev_thr);
-    if (src_u)
-    {
-        mbloop_filter_horizontal_edge_uv_msa(src_u, src_v, pitch_u_v,
-                                             *lpf_info_ptr->mblim,
-                                             *lpf_info_ptr->lim,
-                                             *lpf_info_ptr->hev_thr);
-    }
+void vp8_loop_filter_mbh_msa(uint8_t *src_y, uint8_t *src_u, uint8_t *src_v,
+                             int32_t pitch_y, int32_t pitch_u_v,
+                             loop_filter_info *lpf_info_ptr) {
+  mbloop_filter_horizontal_edge_y_msa(src_y, pitch_y, *lpf_info_ptr->mblim,
+                                      *lpf_info_ptr->lim,
+                                      *lpf_info_ptr->hev_thr);
+  if (src_u) {
+    mbloop_filter_horizontal_edge_uv_msa(
+        src_u, src_v, pitch_u_v, *lpf_info_ptr->mblim, *lpf_info_ptr->lim,
+        *lpf_info_ptr->hev_thr);
+  }
 }
 
-void vp8_loop_filter_mbv_msa(uint8_t *src_y, uint8_t *src_u,
-                             uint8_t *src_v, int32_t pitch_y,
-                             int32_t pitch_u_v,
-                             loop_filter_info *lpf_info_ptr)
-{
-    mbloop_filter_vertical_edge_y_msa(src_y, pitch_y,
-                                      *lpf_info_ptr->mblim,
-                                      *lpf_info_ptr->lim,
-                                      *lpf_info_ptr->hev_thr);
-    if (src_u)
-    {
-        mbloop_filter_vertical_edge_uv_msa(src_u, src_v, pitch_u_v,
-                                           *lpf_info_ptr->mblim,
-                                           *lpf_info_ptr->lim,
-                                           *lpf_info_ptr->hev_thr);
-    }
+void vp8_loop_filter_mbv_msa(uint8_t *src_y, uint8_t *src_u, uint8_t *src_v,
+                             int32_t pitch_y, int32_t pitch_u_v,
+                             loop_filter_info *lpf_info_ptr) {
+  mbloop_filter_vertical_edge_y_msa(src_y, pitch_y, *lpf_info_ptr->mblim,
+                                    *lpf_info_ptr->lim, *lpf_info_ptr->hev_thr);
+  if (src_u) {
+    mbloop_filter_vertical_edge_uv_msa(src_u, src_v, pitch_u_v,
+                                       *lpf_info_ptr->mblim, *lpf_info_ptr->lim,
+                                       *lpf_info_ptr->hev_thr);
+  }
 }
 
-void vp8_loop_filter_bh_msa(uint8_t *src_y, uint8_t *src_u,
-                            uint8_t *src_v, int32_t pitch_y,
-                            int32_t pitch_u_v,
-                            loop_filter_info *lpf_info_ptr)
-{
-    loop_filter_horizontal_4_dual_msa(src_y + 4 * pitch_y, pitch_y,
-                                      lpf_info_ptr->blim,
-                                      lpf_info_ptr->lim,
-                                      lpf_info_ptr->hev_thr,
-                                      lpf_info_ptr->blim,
-                                      lpf_info_ptr->lim,
-                                      lpf_info_ptr->hev_thr);
-    loop_filter_horizontal_4_dual_msa(src_y + 8 * pitch_y, pitch_y,
-                                      lpf_info_ptr->blim,
-                                      lpf_info_ptr->lim,
-                                      lpf_info_ptr->hev_thr,
-                                      lpf_info_ptr->blim,
-                                      lpf_info_ptr->lim,
-                                      lpf_info_ptr->hev_thr);
-    loop_filter_horizontal_4_dual_msa(src_y + 12 * pitch_y, pitch_y,
-                                      lpf_info_ptr->blim,
-                                      lpf_info_ptr->lim,
-                                      lpf_info_ptr->hev_thr,
-                                      lpf_info_ptr->blim,
-                                      lpf_info_ptr->lim,
-                                      lpf_info_ptr->hev_thr);
-    if (src_u)
-    {
-        loop_filter_horizontal_edge_uv_msa(src_u + (4 * pitch_u_v),
-                                           src_v + (4 * pitch_u_v),
-                                           pitch_u_v,
-                                           *lpf_info_ptr->blim,
-                                           *lpf_info_ptr->lim,
-                                           *lpf_info_ptr->hev_thr);
-    }
+void vp8_loop_filter_bh_msa(uint8_t *src_y, uint8_t *src_u, uint8_t *src_v,
+                            int32_t pitch_y, int32_t pitch_u_v,
+                            loop_filter_info *lpf_info_ptr) {
+  loop_filter_horizontal_4_dual_msa(src_y + 4 * pitch_y, pitch_y,
+                                    lpf_info_ptr->blim, lpf_info_ptr->lim,
+                                    lpf_info_ptr->hev_thr, lpf_info_ptr->blim,
+                                    lpf_info_ptr->lim, lpf_info_ptr->hev_thr);
+  loop_filter_horizontal_4_dual_msa(src_y + 8 * pitch_y, pitch_y,
+                                    lpf_info_ptr->blim, lpf_info_ptr->lim,
+                                    lpf_info_ptr->hev_thr, lpf_info_ptr->blim,
+                                    lpf_info_ptr->lim, lpf_info_ptr->hev_thr);
+  loop_filter_horizontal_4_dual_msa(src_y + 12 * pitch_y, pitch_y,
+                                    lpf_info_ptr->blim, lpf_info_ptr->lim,
+                                    lpf_info_ptr->hev_thr, lpf_info_ptr->blim,
+                                    lpf_info_ptr->lim, lpf_info_ptr->hev_thr);
+  if (src_u) {
+    loop_filter_horizontal_edge_uv_msa(
+        src_u + (4 * pitch_u_v), src_v + (4 * pitch_u_v), pitch_u_v,
+        *lpf_info_ptr->blim, *lpf_info_ptr->lim, *lpf_info_ptr->hev_thr);
+  }
 }
 
-void vp8_loop_filter_bv_msa(uint8_t *src_y, uint8_t *src_u,
-                            uint8_t *src_v, int32_t pitch_y,
-                            int32_t pitch_u_v,
-                            loop_filter_info *lpf_info_ptr)
-{
-    loop_filter_vertical_4_dual_msa(src_y + 4, pitch_y, lpf_info_ptr->blim,
-                                    lpf_info_ptr->lim,
-                                    lpf_info_ptr->hev_thr,
-                                    lpf_info_ptr->blim,
-                                    lpf_info_ptr->lim,
-                                    lpf_info_ptr->hev_thr);
-    loop_filter_vertical_4_dual_msa(src_y + 8, pitch_y,
-                                    lpf_info_ptr->blim,
-                                    lpf_info_ptr->lim,
-                                    lpf_info_ptr->hev_thr,
-                                    lpf_info_ptr->blim,
-                                    lpf_info_ptr->lim,
-                                    lpf_info_ptr->hev_thr);
-    loop_filter_vertical_4_dual_msa(src_y + 12, pitch_y,
-                                    lpf_info_ptr->blim,
-                                    lpf_info_ptr->lim,
-                                    lpf_info_ptr->hev_thr,
-                                    lpf_info_ptr->blim,
-                                    lpf_info_ptr->lim,
-                                    lpf_info_ptr->hev_thr);
-    if (src_u)
-    {
-        loop_filter_vertical_edge_uv_msa(src_u + 4, src_v + 4, pitch_u_v,
-                                         *lpf_info_ptr->blim,
-                                         *lpf_info_ptr->lim,
-                                         *lpf_info_ptr->hev_thr);
-    }
+void vp8_loop_filter_bv_msa(uint8_t *src_y, uint8_t *src_u, uint8_t *src_v,
+                            int32_t pitch_y, int32_t pitch_u_v,
+                            loop_filter_info *lpf_info_ptr) {
+  loop_filter_vertical_4_dual_msa(src_y + 4, pitch_y, lpf_info_ptr->blim,
+                                  lpf_info_ptr->lim, lpf_info_ptr->hev_thr,
+                                  lpf_info_ptr->blim, lpf_info_ptr->lim,
+                                  lpf_info_ptr->hev_thr);
+  loop_filter_vertical_4_dual_msa(src_y + 8, pitch_y, lpf_info_ptr->blim,
+                                  lpf_info_ptr->lim, lpf_info_ptr->hev_thr,
+                                  lpf_info_ptr->blim, lpf_info_ptr->lim,
+                                  lpf_info_ptr->hev_thr);
+  loop_filter_vertical_4_dual_msa(src_y + 12, pitch_y, lpf_info_ptr->blim,
+                                  lpf_info_ptr->lim, lpf_info_ptr->hev_thr,
+                                  lpf_info_ptr->blim, lpf_info_ptr->lim,
+                                  lpf_info_ptr->hev_thr);
+  if (src_u) {
+    loop_filter_vertical_edge_uv_msa(src_u + 4, src_v + 4, pitch_u_v,
+                                     *lpf_info_ptr->blim, *lpf_info_ptr->lim,
+                                     *lpf_info_ptr->hev_thr);
+  }
 }
 
 void vp8_loop_filter_bhs_msa(uint8_t *src_y, int32_t pitch_y,
-                             const uint8_t *b_limit_ptr)
-{
-    vp8_loop_filter_simple_horizontal_edge_msa(src_y + (4 * pitch_y),
-                                               pitch_y, b_limit_ptr);
-    vp8_loop_filter_simple_horizontal_edge_msa(src_y + (8 * pitch_y),
-                                               pitch_y, b_limit_ptr);
-    vp8_loop_filter_simple_horizontal_edge_msa(src_y + (12 * pitch_y),
-                                               pitch_y, b_limit_ptr);
+                             const uint8_t *b_limit_ptr) {
+  vp8_loop_filter_simple_horizontal_edge_msa(src_y + (4 * pitch_y), pitch_y,
+                                             b_limit_ptr);
+  vp8_loop_filter_simple_horizontal_edge_msa(src_y + (8 * pitch_y), pitch_y,
+                                             b_limit_ptr);
+  vp8_loop_filter_simple_horizontal_edge_msa(src_y + (12 * pitch_y), pitch_y,
+                                             b_limit_ptr);
 }
 
 void vp8_loop_filter_bvs_msa(uint8_t *src_y, int32_t pitch_y,
-                             const uint8_t *b_limit_ptr)
-{
-    vp8_loop_filter_simple_vertical_edge_msa(src_y + 4, pitch_y, b_limit_ptr);
-    vp8_loop_filter_simple_vertical_edge_msa(src_y + 8, pitch_y, b_limit_ptr);
-    vp8_loop_filter_simple_vertical_edge_msa(src_y + 12, pitch_y, b_limit_ptr);
+                             const uint8_t *b_limit_ptr) {
+  vp8_loop_filter_simple_vertical_edge_msa(src_y + 4, pitch_y, b_limit_ptr);
+  vp8_loop_filter_simple_vertical_edge_msa(src_y + 8, pitch_y, b_limit_ptr);
+  vp8_loop_filter_simple_vertical_edge_msa(src_y + 12, pitch_y, b_limit_ptr);
 }
--- a/vp8/common/mips/msa/mfqe_msa.c
+++ b/vp8/common/mips/msa/mfqe_msa.c
@@ -14,133 +14,126 @@
 
 static void filter_by_weight8x8_msa(uint8_t *src_ptr, int32_t src_stride,
                                     uint8_t *dst_ptr, int32_t dst_stride,
-                                    int32_t src_weight)
-{
-    int32_t dst_weight = (1 << MFQE_PRECISION) - src_weight;
-    int32_t row;
-    uint64_t src0_d, src1_d, dst0_d, dst1_d;
-    v16i8 src0 = { 0 };
-    v16i8 src1 = { 0 };
-    v16i8 dst0 = { 0 };
-    v16i8 dst1 = { 0 };
-    v8i16 src_wt, dst_wt, res_h_r, res_h_l, src_r, src_l, dst_r, dst_l;
+                                    int32_t src_weight) {
+  int32_t dst_weight = (1 << MFQE_PRECISION) - src_weight;
+  int32_t row;
+  uint64_t src0_d, src1_d, dst0_d, dst1_d;
+  v16i8 src0 = { 0 };
+  v16i8 src1 = { 0 };
+  v16i8 dst0 = { 0 };
+  v16i8 dst1 = { 0 };
+  v8i16 src_wt, dst_wt, res_h_r, res_h_l, src_r, src_l, dst_r, dst_l;
 
-    src_wt = __msa_fill_h(src_weight);
-    dst_wt = __msa_fill_h(dst_weight);
+  src_wt = __msa_fill_h(src_weight);
+  dst_wt = __msa_fill_h(dst_weight);
 
-    for (row = 2; row--;)
-    {
-        LD2(src_ptr, src_stride, src0_d, src1_d);
-        src_ptr += (2 * src_stride);
-        LD2(dst_ptr, dst_stride, dst0_d, dst1_d);
-        INSERT_D2_SB(src0_d, src1_d, src0);
-        INSERT_D2_SB(dst0_d, dst1_d, dst0);
+  for (row = 2; row--;) {
+    LD2(src_ptr, src_stride, src0_d, src1_d);
+    src_ptr += (2 * src_stride);
+    LD2(dst_ptr, dst_stride, dst0_d, dst1_d);
+    INSERT_D2_SB(src0_d, src1_d, src0);
+    INSERT_D2_SB(dst0_d, dst1_d, dst0);
 
-        LD2(src_ptr, src_stride, src0_d, src1_d);
-        src_ptr += (2 * src_stride);
-        LD2((dst_ptr + 2 * dst_stride), dst_stride, dst0_d, dst1_d);
-        INSERT_D2_SB(src0_d, src1_d, src1);
-        INSERT_D2_SB(dst0_d, dst1_d, dst1);
+    LD2(src_ptr, src_stride, src0_d, src1_d);
+    src_ptr += (2 * src_stride);
+    LD2((dst_ptr + 2 * dst_stride), dst_stride, dst0_d, dst1_d);
+    INSERT_D2_SB(src0_d, src1_d, src1);
+    INSERT_D2_SB(dst0_d, dst1_d, dst1);
 
-        UNPCK_UB_SH(src0, src_r, src_l);
-        UNPCK_UB_SH(dst0, dst_r, dst_l);
-        res_h_r = (src_r * src_wt);
-        res_h_r += (dst_r * dst_wt);
-        res_h_l = (src_l * src_wt);
-        res_h_l += (dst_l * dst_wt);
-        SRARI_H2_SH(res_h_r, res_h_l, MFQE_PRECISION);
-        dst0 = (v16i8)__msa_pckev_b((v16i8)res_h_l, (v16i8)res_h_r);
-        ST8x2_UB(dst0, dst_ptr, dst_stride);
-        dst_ptr += (2 * dst_stride);
+    UNPCK_UB_SH(src0, src_r, src_l);
+    UNPCK_UB_SH(dst0, dst_r, dst_l);
+    res_h_r = (src_r * src_wt);
+    res_h_r += (dst_r * dst_wt);
+    res_h_l = (src_l * src_wt);
+    res_h_l += (dst_l * dst_wt);
+    SRARI_H2_SH(res_h_r, res_h_l, MFQE_PRECISION);
+    dst0 = (v16i8)__msa_pckev_b((v16i8)res_h_l, (v16i8)res_h_r);
+    ST8x2_UB(dst0, dst_ptr, dst_stride);
+    dst_ptr += (2 * dst_stride);
 
-        UNPCK_UB_SH(src1, src_r, src_l);
-        UNPCK_UB_SH(dst1, dst_r, dst_l);
-        res_h_r = (src_r * src_wt);
-        res_h_r += (dst_r * dst_wt);
-        res_h_l = (src_l * src_wt);
-        res_h_l += (dst_l * dst_wt);
-        SRARI_H2_SH(res_h_r, res_h_l, MFQE_PRECISION);
-        dst1 = (v16i8)__msa_pckev_b((v16i8)res_h_l, (v16i8)res_h_r);
-        ST8x2_UB(dst1, dst_ptr, dst_stride);
-        dst_ptr += (2 * dst_stride);
-    }
+    UNPCK_UB_SH(src1, src_r, src_l);
+    UNPCK_UB_SH(dst1, dst_r, dst_l);
+    res_h_r = (src_r * src_wt);
+    res_h_r += (dst_r * dst_wt);
+    res_h_l = (src_l * src_wt);
+    res_h_l += (dst_l * dst_wt);
+    SRARI_H2_SH(res_h_r, res_h_l, MFQE_PRECISION);
+    dst1 = (v16i8)__msa_pckev_b((v16i8)res_h_l, (v16i8)res_h_r);
+    ST8x2_UB(dst1, dst_ptr, dst_stride);
+    dst_ptr += (2 * dst_stride);
+  }
 }
 
 static void filter_by_weight16x16_msa(uint8_t *src_ptr, int32_t src_stride,
                                       uint8_t *dst_ptr, int32_t dst_stride,
-                                      int32_t src_weight)
-{
-    int32_t dst_weight = (1 << MFQE_PRECISION) - src_weight;
-    int32_t row;
-    v16i8 src0, src1, src2, src3;
-    v16i8 dst0, dst1, dst2, dst3;
-    v8i16 src_wt, dst_wt;
-    v8i16 res_h_r, res_h_l;
-    v8i16 src_r, src_l, dst_r, dst_l;
+                                      int32_t src_weight) {
+  int32_t dst_weight = (1 << MFQE_PRECISION) - src_weight;
+  int32_t row;
+  v16i8 src0, src1, src2, src3;
+  v16i8 dst0, dst1, dst2, dst3;
+  v8i16 src_wt, dst_wt;
+  v8i16 res_h_r, res_h_l;
+  v8i16 src_r, src_l, dst_r, dst_l;
 
-    src_wt = __msa_fill_h(src_weight);
-    dst_wt = __msa_fill_h(dst_weight);
+  src_wt = __msa_fill_h(src_weight);
+  dst_wt = __msa_fill_h(dst_weight);
 
-    for (row = 4; row--;)
-    {
-        LD_SB4(src_ptr, src_stride, src0, src1, src2, src3);
-        src_ptr += (4 * src_stride);
-        LD_SB4(dst_ptr, dst_stride, dst0, dst1, dst2, dst3);
+  for (row = 4; row--;) {
+    LD_SB4(src_ptr, src_stride, src0, src1, src2, src3);
+    src_ptr += (4 * src_stride);
+    LD_SB4(dst_ptr, dst_stride, dst0, dst1, dst2, dst3);
 
-        UNPCK_UB_SH(src0, src_r, src_l);
-        UNPCK_UB_SH(dst0, dst_r, dst_l);
-        res_h_r = (src_r * src_wt);
-        res_h_r += (dst_r * dst_wt);
-        res_h_l = (src_l * src_wt);
-        res_h_l += (dst_l * dst_wt);
-        SRARI_H2_SH(res_h_r, res_h_l, MFQE_PRECISION);
-        PCKEV_ST_SB(res_h_r, res_h_l, dst_ptr);
-        dst_ptr += dst_stride;
+    UNPCK_UB_SH(src0, src_r, src_l);
+    UNPCK_UB_SH(dst0, dst_r, dst_l);
+    res_h_r = (src_r * src_wt);
+    res_h_r += (dst_r * dst_wt);
+    res_h_l = (src_l * src_wt);
+    res_h_l += (dst_l * dst_wt);
+    SRARI_H2_SH(res_h_r, res_h_l, MFQE_PRECISION);
+    PCKEV_ST_SB(res_h_r, res_h_l, dst_ptr);
+    dst_ptr += dst_stride;
 
-        UNPCK_UB_SH(src1, src_r, src_l);
-        UNPCK_UB_SH(dst1, dst_r, dst_l);
-        res_h_r = (src_r * src_wt);
-        res_h_r += (dst_r * dst_wt);
-        res_h_l = (src_l * src_wt);
-        res_h_l += (dst_l * dst_wt);
-        SRARI_H2_SH(res_h_r, res_h_l, MFQE_PRECISION);
-        PCKEV_ST_SB(res_h_r, res_h_l, dst_ptr);
-        dst_ptr += dst_stride;
+    UNPCK_UB_SH(src1, src_r, src_l);
+    UNPCK_UB_SH(dst1, dst_r, dst_l);
+    res_h_r = (src_r * src_wt);
+    res_h_r += (dst_r * dst_wt);
+    res_h_l = (src_l * src_wt);
+    res_h_l += (dst_l * dst_wt);
+    SRARI_H2_SH(res_h_r, res_h_l, MFQE_PRECISION);
+    PCKEV_ST_SB(res_h_r, res_h_l, dst_ptr);
+    dst_ptr += dst_stride;
 
-        UNPCK_UB_SH(src2, src_r, src_l);
-        UNPCK_UB_SH(dst2, dst_r, dst_l);
-        res_h_r = (src_r * src_wt);
-        res_h_r += (dst_r * dst_wt);
-        res_h_l = (src_l * src_wt);
-        res_h_l += (dst_l * dst_wt);
-        SRARI_H2_SH(res_h_r, res_h_l, MFQE_PRECISION);
-        PCKEV_ST_SB(res_h_r, res_h_l, dst_ptr);
-        dst_ptr += dst_stride;
+    UNPCK_UB_SH(src2, src_r, src_l);
+    UNPCK_UB_SH(dst2, dst_r, dst_l);
+    res_h_r = (src_r * src_wt);
+    res_h_r += (dst_r * dst_wt);
+    res_h_l = (src_l * src_wt);
+    res_h_l += (dst_l * dst_wt);
+    SRARI_H2_SH(res_h_r, res_h_l, MFQE_PRECISION);
+    PCKEV_ST_SB(res_h_r, res_h_l, dst_ptr);
+    dst_ptr += dst_stride;
 
-        UNPCK_UB_SH(src3, src_r, src_l);
-        UNPCK_UB_SH(dst3, dst_r, dst_l);
-        res_h_r = (src_r * src_wt);
-        res_h_r += (dst_r * dst_wt);
-        res_h_l = (src_l * src_wt);
-        res_h_l += (dst_l * dst_wt);
-        SRARI_H2_SH(res_h_r, res_h_l, MFQE_PRECISION);
-        PCKEV_ST_SB(res_h_r, res_h_l, dst_ptr);
-        dst_ptr += dst_stride;
-    }
+    UNPCK_UB_SH(src3, src_r, src_l);
+    UNPCK_UB_SH(dst3, dst_r, dst_l);
+    res_h_r = (src_r * src_wt);
+    res_h_r += (dst_r * dst_wt);
+    res_h_l = (src_l * src_wt);
+    res_h_l += (dst_l * dst_wt);
+    SRARI_H2_SH(res_h_r, res_h_l, MFQE_PRECISION);
+    PCKEV_ST_SB(res_h_r, res_h_l, dst_ptr);
+    dst_ptr += dst_stride;
+  }
 }
 
 void vp8_filter_by_weight16x16_msa(uint8_t *src_ptr, int32_t src_stride,
                                    uint8_t *dst_ptr, int32_t dst_stride,
-                                   int32_t src_weight)
-{
-    filter_by_weight16x16_msa(src_ptr, src_stride, dst_ptr, dst_stride,
-                              src_weight);
+                                   int32_t src_weight) {
+  filter_by_weight16x16_msa(src_ptr, src_stride, dst_ptr, dst_stride,
+                            src_weight);
 }
 
 void vp8_filter_by_weight8x8_msa(uint8_t *src_ptr, int32_t src_stride,
                                  uint8_t *dst_ptr, int32_t dst_stride,
-                                 int32_t src_weight)
-{
-    filter_by_weight8x8_msa(src_ptr, src_stride, dst_ptr, dst_stride,
-                            src_weight);
+                                 int32_t src_weight) {
+  filter_by_weight8x8_msa(src_ptr, src_stride, dst_ptr, dst_stride, src_weight);
 }
--- a/vp8/common/mips/msa/sixtap_filter_msa.c
+++ b/vp8/common/mips/msa/sixtap_filter_msa.c
@@ -13,226 +13,225 @@
 #include "vp8/common/filter.h"
 #include "vp8/common/mips/msa/vp8_macros_msa.h"
 
-DECLARE_ALIGNED(16, static const int8_t, vp8_subpel_filters_msa[7][8]) =
-{
-    { 0, -6, 123, 12, -1, 0, 0, 0 },
-    { 2, -11, 108, 36, -8, 1, 0, 0 },  /* New 1/4 pel 6 tap filter */
-    { 0, -9, 93, 50, -6, 0, 0, 0 },
-    { 3, -16, 77, 77, -16, 3, 0, 0 },  /* New 1/2 pel 6 tap filter */
-    { 0, -6, 50, 93, -9, 0, 0, 0 },
-    { 1, -8, 36, 108, -11, 2, 0, 0 },  /* New 1/4 pel 6 tap filter */
-    { 0, -1, 12, 123, -6, 0, 0, 0 },
+DECLARE_ALIGNED(16, static const int8_t, vp8_subpel_filters_msa[7][8]) = {
+  { 0, -6, 123, 12, -1, 0, 0, 0 },
+  { 2, -11, 108, 36, -8, 1, 0, 0 }, /* New 1/4 pel 6 tap filter */
+  { 0, -9, 93, 50, -6, 0, 0, 0 },
+  { 3, -16, 77, 77, -16, 3, 0, 0 }, /* New 1/2 pel 6 tap filter */
+  { 0, -6, 50, 93, -9, 0, 0, 0 },
+  { 1, -8, 36, 108, -11, 2, 0, 0 }, /* New 1/4 pel 6 tap filter */
+  { 0, -1, 12, 123, -6, 0, 0, 0 },
 };
 
-static const uint8_t vp8_mc_filt_mask_arr[16 * 3] =
-{
-    /* 8 width cases */
-    0, 1, 1, 2, 2, 3, 3, 4, 4, 5, 5, 6, 6, 7, 7, 8,
-    /* 4 width cases */
-    0, 1, 1, 2, 2, 3, 3, 4, 16, 17, 17, 18, 18, 19, 19, 20,
-    /* 4 width cases */
-    8, 9, 9, 10, 10, 11, 11, 12, 24, 25, 25, 26, 26, 27, 27, 28
+static const uint8_t vp8_mc_filt_mask_arr[16 * 3] = {
+  /* 8 width cases */
+  0, 1, 1, 2, 2, 3, 3, 4, 4, 5, 5, 6, 6, 7, 7, 8,
+  /* 4 width cases */
+  0, 1, 1, 2, 2, 3, 3, 4, 16, 17, 17, 18, 18, 19, 19, 20,
+  /* 4 width cases */
+  8, 9, 9, 10, 10, 11, 11, 12, 24, 25, 25, 26, 26, 27, 27, 28
 };
 
-#define HORIZ_6TAP_FILT(src0, src1, mask0, mask1, mask2,                 \
-                        filt_h0, filt_h1, filt_h2)                       \
-({                                                                       \
-    v16i8 vec0_m, vec1_m, vec2_m;                                        \
-    v8i16 hz_out_m;                                                      \
-                                                                         \
-    VSHF_B3_SB(src0, src1, src0, src1, src0, src1, mask0, mask1, mask2,  \
-               vec0_m, vec1_m, vec2_m);                                  \
-    hz_out_m = DPADD_SH3_SH(vec0_m, vec1_m, vec2_m,                      \
-                            filt_h0, filt_h1, filt_h2);                  \
-                                                                         \
-    hz_out_m = __msa_srari_h(hz_out_m, VP8_FILTER_SHIFT);                \
-    hz_out_m = __msa_sat_s_h(hz_out_m, 7);                               \
-                                                                         \
-    hz_out_m;                                                            \
-})
+#define HORIZ_6TAP_FILT(src0, src1, mask0, mask1, mask2, filt_h0, filt_h1, \
+                        filt_h2)                                           \
+  ({                                                                       \
+    v16i8 vec0_m, vec1_m, vec2_m;                                          \
+    v8i16 hz_out_m;                                                        \
+                                                                           \
+    VSHF_B3_SB(src0, src1, src0, src1, src0, src1, mask0, mask1, mask2,    \
+               vec0_m, vec1_m, vec2_m);                                    \
+    hz_out_m =                                                             \
+        DPADD_SH3_SH(vec0_m, vec1_m, vec2_m, filt_h0, filt_h1, filt_h2);   \
+                                                                           \
+    hz_out_m = __msa_srari_h(hz_out_m, VP8_FILTER_SHIFT);                  \
+    hz_out_m = __msa_sat_s_h(hz_out_m, 7);                                 \
+                                                                           \
+    hz_out_m;                                                              \
+  })
 
-#define HORIZ_6TAP_4WID_4VECS_FILT(src0, src1, src2, src3,             \
-                                   mask0, mask1, mask2,                \
-                                   filt0, filt1, filt2,                \
-                                   out0, out1)                         \
-{                                                                      \
-    v16i8 vec0_m, vec1_m, vec2_m, vec3_m, vec4_m, vec5_m;              \
-                                                                       \
-    VSHF_B2_SB(src0, src1, src2, src3, mask0, mask0, vec0_m, vec1_m);  \
-    DOTP_SB2_SH(vec0_m, vec1_m, filt0, filt0, out0, out1);             \
-    VSHF_B2_SB(src0, src1, src2, src3, mask1, mask1, vec2_m, vec3_m);  \
-    DPADD_SB2_SH(vec2_m, vec3_m, filt1, filt1, out0, out1);            \
-    VSHF_B2_SB(src0, src1, src2, src3, mask2, mask2, vec4_m, vec5_m);  \
-    DPADD_SB2_SH(vec4_m, vec5_m, filt2, filt2, out0, out1);            \
-}
+#define HORIZ_6TAP_4WID_4VECS_FILT(src0, src1, src2, src3, mask0, mask1,   \
+                                   mask2, filt0, filt1, filt2, out0, out1) \
+  {                                                                        \
+    v16i8 vec0_m, vec1_m, vec2_m, vec3_m, vec4_m, vec5_m;                  \
+                                                                           \
+    VSHF_B2_SB(src0, src1, src2, src3, mask0, mask0, vec0_m, vec1_m);      \
+    DOTP_SB2_SH(vec0_m, vec1_m, filt0, filt0, out0, out1);                 \
+    VSHF_B2_SB(src0, src1, src2, src3, mask1, mask1, vec2_m, vec3_m);      \
+    DPADD_SB2_SH(vec2_m, vec3_m, filt1, filt1, out0, out1);                \
+    VSHF_B2_SB(src0, src1, src2, src3, mask2, mask2, vec4_m, vec5_m);      \
+    DPADD_SB2_SH(vec4_m, vec5_m, filt2, filt2, out0, out1);                \
+  }
 
-#define HORIZ_6TAP_8WID_4VECS_FILT(src0, src1, src2, src3,                    \
-                                   mask0, mask1, mask2,                       \
-                                   filt0, filt1, filt2,                       \
-                                   out0, out1, out2, out3)                    \
-{                                                                             \
-    v16i8 vec0_m, vec1_m, vec2_m, vec3_m, vec4_m, vec5_m, vec6_m, vec7_m;     \
-                                                                              \
-    VSHF_B2_SB(src0, src0, src1, src1, mask0, mask0, vec0_m, vec1_m);         \
-    VSHF_B2_SB(src2, src2, src3, src3, mask0, mask0, vec2_m, vec3_m);         \
-    DOTP_SB4_SH(vec0_m, vec1_m, vec2_m, vec3_m, filt0, filt0, filt0, filt0,   \
-                out0, out1, out2, out3);                                      \
-    VSHF_B2_SB(src0, src0, src1, src1, mask1, mask1, vec0_m, vec1_m);         \
-    VSHF_B2_SB(src2, src2, src3, src3, mask1, mask1, vec2_m, vec3_m);         \
-    VSHF_B2_SB(src0, src0, src1, src1, mask2, mask2, vec4_m, vec5_m);         \
-    VSHF_B2_SB(src2, src2, src3, src3, mask2, mask2, vec6_m, vec7_m);         \
-    DPADD_SB4_SH(vec0_m, vec1_m, vec2_m, vec3_m, filt1, filt1, filt1, filt1,  \
-                 out0, out1, out2, out3);                                     \
-    DPADD_SB4_SH(vec4_m, vec5_m, vec6_m, vec7_m, filt2, filt2, filt2, filt2,  \
-                 out0, out1, out2, out3);                                     \
-}
+#define HORIZ_6TAP_8WID_4VECS_FILT(src0, src1, src2, src3, mask0, mask1,     \
+                                   mask2, filt0, filt1, filt2, out0, out1,   \
+                                   out2, out3)                               \
+  {                                                                          \
+    v16i8 vec0_m, vec1_m, vec2_m, vec3_m, vec4_m, vec5_m, vec6_m, vec7_m;    \
+                                                                             \
+    VSHF_B2_SB(src0, src0, src1, src1, mask0, mask0, vec0_m, vec1_m);        \
+    VSHF_B2_SB(src2, src2, src3, src3, mask0, mask0, vec2_m, vec3_m);        \
+    DOTP_SB4_SH(vec0_m, vec1_m, vec2_m, vec3_m, filt0, filt0, filt0, filt0,  \
+                out0, out1, out2, out3);                                     \
+    VSHF_B2_SB(src0, src0, src1, src1, mask1, mask1, vec0_m, vec1_m);        \
+    VSHF_B2_SB(src2, src2, src3, src3, mask1, mask1, vec2_m, vec3_m);        \
+    VSHF_B2_SB(src0, src0, src1, src1, mask2, mask2, vec4_m, vec5_m);        \
+    VSHF_B2_SB(src2, src2, src3, src3, mask2, mask2, vec6_m, vec7_m);        \
+    DPADD_SB4_SH(vec0_m, vec1_m, vec2_m, vec3_m, filt1, filt1, filt1, filt1, \
+                 out0, out1, out2, out3);                                    \
+    DPADD_SB4_SH(vec4_m, vec5_m, vec6_m, vec7_m, filt2, filt2, filt2, filt2, \
+                 out0, out1, out2, out3);                                    \
+  }
 
-#define FILT_4TAP_DPADD_S_H(vec0, vec1, filt0, filt1)         \
-({                                                            \
-    v8i16 tmp0;                                               \
-                                                              \
-    tmp0 = __msa_dotp_s_h((v16i8)vec0, (v16i8)filt0);         \
-    tmp0 = __msa_dpadd_s_h(tmp0, (v16i8)vec1, (v16i8)filt1);  \
-                                                              \
-    tmp0;                                                     \
-})
+#define FILT_4TAP_DPADD_S_H(vec0, vec1, filt0, filt1)        \
+  ({                                                         \
+    v8i16 tmp0;                                              \
+                                                             \
+    tmp0 = __msa_dotp_s_h((v16i8)vec0, (v16i8)filt0);        \
+    tmp0 = __msa_dpadd_s_h(tmp0, (v16i8)vec1, (v16i8)filt1); \
+                                                             \
+    tmp0;                                                    \
+  })
 
-#define HORIZ_4TAP_FILT(src0, src1, mask0, mask1, filt_h0, filt_h1)    \
-({                                                                     \
-    v16i8 vec0_m, vec1_m;                                              \
-    v8i16 hz_out_m;                                                    \
-                                                                       \
-    VSHF_B2_SB(src0, src1, src0, src1, mask0, mask1, vec0_m, vec1_m);  \
-    hz_out_m = FILT_4TAP_DPADD_S_H(vec0_m, vec1_m, filt_h0, filt_h1);  \
-                                                                       \
-    hz_out_m = __msa_srari_h(hz_out_m, VP8_FILTER_SHIFT);              \
-    hz_out_m = __msa_sat_s_h(hz_out_m, 7);                             \
-                                                                       \
-    hz_out_m;                                                          \
-})
+#define HORIZ_4TAP_FILT(src0, src1, mask0, mask1, filt_h0, filt_h1)   \
+  ({                                                                  \
+    v16i8 vec0_m, vec1_m;                                             \
+    v8i16 hz_out_m;                                                   \
+                                                                      \
+    VSHF_B2_SB(src0, src1, src0, src1, mask0, mask1, vec0_m, vec1_m); \
+    hz_out_m = FILT_4TAP_DPADD_S_H(vec0_m, vec1_m, filt_h0, filt_h1); \
+                                                                      \
+    hz_out_m = __msa_srari_h(hz_out_m, VP8_FILTER_SHIFT);             \
+    hz_out_m = __msa_sat_s_h(hz_out_m, 7);                            \
+                                                                      \
+    hz_out_m;                                                         \
+  })
 
-#define HORIZ_4TAP_4WID_4VECS_FILT(src0, src1, src2, src3,             \
-                                   mask0, mask1, filt0, filt1,         \
-                                   out0, out1)                         \
-{                                                                      \
-    v16i8 vec0_m, vec1_m, vec2_m, vec3_m;                              \
-                                                                       \
-    VSHF_B2_SB(src0, src1, src2, src3, mask0, mask0, vec0_m, vec1_m);  \
-    DOTP_SB2_SH(vec0_m, vec1_m, filt0, filt0, out0, out1);             \
-    VSHF_B2_SB(src0, src1, src2, src3, mask1, mask1, vec2_m, vec3_m);  \
-    DPADD_SB2_SH(vec2_m, vec3_m, filt1, filt1, out0, out1);            \
-}
+#define HORIZ_4TAP_4WID_4VECS_FILT(src0, src1, src2, src3, mask0, mask1, \
+                                   filt0, filt1, out0, out1)             \
+  {                                                                      \
+    v16i8 vec0_m, vec1_m, vec2_m, vec3_m;                                \
+                                                                         \
+    VSHF_B2_SB(src0, src1, src2, src3, mask0, mask0, vec0_m, vec1_m);    \
+    DOTP_SB2_SH(vec0_m, vec1_m, filt0, filt0, out0, out1);               \
+    VSHF_B2_SB(src0, src1, src2, src3, mask1, mask1, vec2_m, vec3_m);    \
+    DPADD_SB2_SH(vec2_m, vec3_m, filt1, filt1, out0, out1);              \
+  }
 
-#define HORIZ_4TAP_8WID_4VECS_FILT(src0, src1, src2, src3,                    \
-                                   mask0, mask1, filt0, filt1,                \
-                                   out0, out1, out2, out3)                    \
-{                                                                             \
-    v16i8 vec0_m, vec1_m, vec2_m, vec3_m;                                     \
-                                                                              \
-    VSHF_B2_SB(src0, src0, src1, src1, mask0, mask0, vec0_m, vec1_m);         \
-    VSHF_B2_SB(src2, src2, src3, src3, mask0, mask0, vec2_m, vec3_m);         \
-    DOTP_SB4_SH(vec0_m, vec1_m, vec2_m, vec3_m, filt0, filt0, filt0, filt0,   \
-                out0, out1, out2, out3);                                      \
-    VSHF_B2_SB(src0, src0, src1, src1, mask1, mask1, vec0_m, vec1_m);         \
-    VSHF_B2_SB(src2, src2, src3, src3, mask1, mask1, vec2_m, vec3_m);         \
-    DPADD_SB4_SH(vec0_m, vec1_m, vec2_m, vec3_m, filt1, filt1, filt1, filt1,  \
-                 out0, out1, out2, out3);                                     \
-}
+#define HORIZ_4TAP_8WID_4VECS_FILT(src0, src1, src2, src3, mask0, mask1,     \
+                                   filt0, filt1, out0, out1, out2, out3)     \
+  {                                                                          \
+    v16i8 vec0_m, vec1_m, vec2_m, vec3_m;                                    \
+                                                                             \
+    VSHF_B2_SB(src0, src0, src1, src1, mask0, mask0, vec0_m, vec1_m);        \
+    VSHF_B2_SB(src2, src2, src3, src3, mask0, mask0, vec2_m, vec3_m);        \
+    DOTP_SB4_SH(vec0_m, vec1_m, vec2_m, vec3_m, filt0, filt0, filt0, filt0,  \
+                out0, out1, out2, out3);                                     \
+    VSHF_B2_SB(src0, src0, src1, src1, mask1, mask1, vec0_m, vec1_m);        \
+    VSHF_B2_SB(src2, src2, src3, src3, mask1, mask1, vec2_m, vec3_m);        \
+    DPADD_SB4_SH(vec0_m, vec1_m, vec2_m, vec3_m, filt1, filt1, filt1, filt1, \
+                 out0, out1, out2, out3);                                    \
+  }
 
 static void common_hz_6t_4x4_msa(uint8_t *RESTRICT src, int32_t src_stride,
                                  uint8_t *RESTRICT dst, int32_t dst_stride,
-                                 const int8_t *filter)
-{
-    v16i8 src0, src1, src2, src3, filt0, filt1, filt2;
-    v16u8 mask0, mask1, mask2, out;
-    v8i16 filt, out0, out1;
+                                 const int8_t *filter) {
+  v16i8 src0, src1, src2, src3, filt0, filt1, filt2;
+  v16u8 mask0, mask1, mask2, out;
+  v8i16 filt, out0, out1;
 
-    mask0 = LD_UB(&vp8_mc_filt_mask_arr[16]);
-    src -= 2;
+  mask0 = LD_UB(&vp8_mc_filt_mask_arr[16]);
+  src -= 2;
 
-    filt = LD_SH(filter);
-    SPLATI_H3_SB(filt, 0, 1, 2, filt0, filt1, filt2);
+  filt = LD_SH(filter);
+  SPLATI_H3_SB(filt, 0, 1, 2, filt0, filt1, filt2);
 
-    mask1 = mask0 + 2;
-    mask2 = mask0 + 4;
+  mask1 = mask0 + 2;
+  mask2 = mask0 + 4;
 
-    LD_SB4(src, src_stride, src0, src1, src2, src3);
-    XORI_B4_128_SB(src0, src1, src2, src3);
-    HORIZ_6TAP_4WID_4VECS_FILT(src0, src1, src2, src3, mask0, mask1, mask2,
-                               filt0, filt1, filt2, out0, out1);
-    SRARI_H2_SH(out0, out1, VP8_FILTER_SHIFT);
-    SAT_SH2_SH(out0, out1, 7);
-    out = PCKEV_XORI128_UB(out0, out1);
-    ST4x4_UB(out, out, 0, 1, 2, 3, dst, dst_stride);
+  LD_SB4(src, src_stride, src0, src1, src2, src3);
+  XORI_B4_128_SB(src0, src1, src2, src3);
+  HORIZ_6TAP_4WID_4VECS_FILT(src0, src1, src2, src3, mask0, mask1, mask2, filt0,
+                             filt1, filt2, out0, out1);
+  SRARI_H2_SH(out0, out1, VP8_FILTER_SHIFT);
+  SAT_SH2_SH(out0, out1, 7);
+  out = PCKEV_XORI128_UB(out0, out1);
+  ST4x4_UB(out, out, 0, 1, 2, 3, dst, dst_stride);
 }
 
 static void common_hz_6t_4x8_msa(uint8_t *RESTRICT src, int32_t src_stride,
                                  uint8_t *RESTRICT dst, int32_t dst_stride,
-                                 const int8_t *filter)
-{
-    v16i8 src0, src1, src2, src3, filt0, filt1, filt2;
-    v16u8 mask0, mask1, mask2, out;
-    v8i16 filt, out0, out1, out2, out3;
+                                 const int8_t *filter) {
+  v16i8 src0, src1, src2, src3, filt0, filt1, filt2;
+  v16u8 mask0, mask1, mask2, out;
+  v8i16 filt, out0, out1, out2, out3;
 
-    mask0 = LD_UB(&vp8_mc_filt_mask_arr[16]);
-    src -= 2;
+  mask0 = LD_UB(&vp8_mc_filt_mask_arr[16]);
+  src -= 2;
 
-    filt = LD_SH(filter);
-    SPLATI_H3_SB(filt, 0, 1, 2, filt0, filt1, filt2);
+  filt = LD_SH(filter);
+  SPLATI_H3_SB(filt, 0, 1, 2, filt0, filt1, filt2);
 
-    mask1 = mask0 + 2;
-    mask2 = mask0 + 4;
+  mask1 = mask0 + 2;
+  mask2 = mask0 + 4;
 
-    LD_SB4(src, src_stride, src0, src1, src2, src3);
-    XORI_B4_128_SB(src0, src1, src2, src3);
-    src += (4 * src_stride);
-    HORIZ_6TAP_4WID_4VECS_FILT(src0, src1, src2, src3, mask0, mask1, mask2,
-                               filt0, filt1, filt2, out0, out1);
-    LD_SB4(src, src_stride, src0, src1, src2, src3);
-    XORI_B4_128_SB(src0, src1, src2, src3);
-    HORIZ_6TAP_4WID_4VECS_FILT(src0, src1, src2, src3, mask0, mask1, mask2,
-                               filt0, filt1, filt2, out2, out3);
-    SRARI_H4_SH(out0, out1, out2, out3, VP8_FILTER_SHIFT);
-    SAT_SH4_SH(out0, out1, out2, out3, 7);
-    out = PCKEV_XORI128_UB(out0, out1);
-    ST4x4_UB(out, out, 0, 1, 2, 3, dst, dst_stride);
-    dst += (4 * dst_stride);
-    out = PCKEV_XORI128_UB(out2, out3);
-    ST4x4_UB(out, out, 0, 1, 2, 3, dst, dst_stride);
+  LD_SB4(src, src_stride, src0, src1, src2, src3);
+  XORI_B4_128_SB(src0, src1, src2, src3);
+  src += (4 * src_stride);
+  HORIZ_6TAP_4WID_4VECS_FILT(src0, src1, src2, src3, mask0, mask1, mask2, filt0,
+                             filt1, filt2, out0, out1);
+  LD_SB4(src, src_stride, src0, src1, src2, src3);
+  XORI_B4_128_SB(src0, src1, src2, src3);
+  HORIZ_6TAP_4WID_4VECS_FILT(src0, src1, src2, src3, mask0, mask1, mask2, filt0,
+                             filt1, filt2, out2, out3);
+  SRARI_H4_SH(out0, out1, out2, out3, VP8_FILTER_SHIFT);
+  SAT_SH4_SH(out0, out1, out2, out3, 7);
+  out = PCKEV_XORI128_UB(out0, out1);
+  ST4x4_UB(out, out, 0, 1, 2, 3, dst, dst_stride);
+  dst += (4 * dst_stride);
+  out = PCKEV_XORI128_UB(out2, out3);
+  ST4x4_UB(out, out, 0, 1, 2, 3, dst, dst_stride);
 }
 
 static void common_hz_6t_4w_msa(uint8_t *RESTRICT src, int32_t src_stride,
                                 uint8_t *RESTRICT dst, int32_t dst_stride,
-                                const int8_t *filter, int32_t height)
-{
-    if (4 == height)
-    {
-        common_hz_6t_4x4_msa(src, src_stride, dst, dst_stride, filter);
-    }
-    else if (8 == height)
-    {
-        common_hz_6t_4x8_msa(src, src_stride, dst, dst_stride, filter);
-    }
+                                const int8_t *filter, int32_t height) {
+  if (4 == height) {
+    common_hz_6t_4x4_msa(src, src_stride, dst, dst_stride, filter);
+  } else if (8 == height) {
+    common_hz_6t_4x8_msa(src, src_stride, dst, dst_stride, filter);
+  }
 }
 
 static void common_hz_6t_8w_msa(uint8_t *RESTRICT src, int32_t src_stride,
                                 uint8_t *RESTRICT dst, int32_t dst_stride,
-                                const int8_t *filter, int32_t height)
-{
-    uint32_t loop_cnt;
-    v16i8 src0, src1, src2, src3, filt0, filt1, filt2;
-    v16u8 mask0, mask1, mask2, tmp0, tmp1;
-    v8i16 filt, out0, out1, out2, out3;
+                                const int8_t *filter, int32_t height) {
+  uint32_t loop_cnt;
+  v16i8 src0, src1, src2, src3, filt0, filt1, filt2;
+  v16u8 mask0, mask1, mask2, tmp0, tmp1;
+  v8i16 filt, out0, out1, out2, out3;
 
-    mask0 = LD_UB(&vp8_mc_filt_mask_arr[0]);
-    src -= 2;
+  mask0 = LD_UB(&vp8_mc_filt_mask_arr[0]);
+  src -= 2;
 
-    filt = LD_SH(filter);
-    SPLATI_H3_SB(filt, 0, 1, 2, filt0, filt1, filt2);
+  filt = LD_SH(filter);
+  SPLATI_H3_SB(filt, 0, 1, 2, filt0, filt1, filt2);
 
-    mask1 = mask0 + 2;
-    mask2 = mask0 + 4;
+  mask1 = mask0 + 2;
+  mask2 = mask0 + 4;
 
+  LD_SB4(src, src_stride, src0, src1, src2, src3);
+  XORI_B4_128_SB(src0, src1, src2, src3);
+  src += (4 * src_stride);
+  HORIZ_6TAP_8WID_4VECS_FILT(src0, src1, src2, src3, mask0, mask1, mask2, filt0,
+                             filt1, filt2, out0, out1, out2, out3);
+  SRARI_H4_SH(out0, out1, out2, out3, VP8_FILTER_SHIFT);
+  SAT_SH4_SH(out0, out1, out2, out3, 7);
+  tmp0 = PCKEV_XORI128_UB(out0, out1);
+  tmp1 = PCKEV_XORI128_UB(out2, out3);
+  ST8x4_UB(tmp0, tmp1, dst, dst_stride);
+  dst += (4 * dst_stride);
+
+  for (loop_cnt = (height >> 2) - 1; loop_cnt--;) {
     LD_SB4(src, src_stride, src0, src1, src2, src3);
     XORI_B4_128_SB(src0, src1, src2, src3);
     src += (4 * src_stride);
@@ -244,232 +243,210 @@
     tmp1 = PCKEV_XORI128_UB(out2, out3);
     ST8x4_UB(tmp0, tmp1, dst, dst_stride);
     dst += (4 * dst_stride);
-
-    for (loop_cnt = (height >> 2) - 1; loop_cnt--;)
-    {
-        LD_SB4(src, src_stride, src0, src1, src2, src3);
-        XORI_B4_128_SB(src0, src1, src2, src3);
-        src += (4 * src_stride);
-        HORIZ_6TAP_8WID_4VECS_FILT(src0, src1, src2, src3, mask0, mask1, mask2,
-                                   filt0, filt1, filt2, out0, out1, out2, out3);
-        SRARI_H4_SH(out0, out1, out2, out3, VP8_FILTER_SHIFT);
-        SAT_SH4_SH(out0, out1, out2, out3, 7);
-        tmp0 = PCKEV_XORI128_UB(out0, out1);
-        tmp1 = PCKEV_XORI128_UB(out2, out3);
-        ST8x4_UB(tmp0, tmp1, dst, dst_stride);
-        dst += (4 * dst_stride);
-    }
+  }
 }
 
 static void common_hz_6t_16w_msa(uint8_t *RESTRICT src, int32_t src_stride,
                                  uint8_t *RESTRICT dst, int32_t dst_stride,
-                                 const int8_t *filter, int32_t height)
-{
-    uint32_t loop_cnt;
-    v16i8 src0, src1, src2, src3, src4, src5, src6, src7, filt0, filt1, filt2;
-    v16u8 mask0, mask1, mask2, out;
-    v8i16 filt, out0, out1, out2, out3, out4, out5, out6, out7;
+                                 const int8_t *filter, int32_t height) {
+  uint32_t loop_cnt;
+  v16i8 src0, src1, src2, src3, src4, src5, src6, src7, filt0, filt1, filt2;
+  v16u8 mask0, mask1, mask2, out;
+  v8i16 filt, out0, out1, out2, out3, out4, out5, out6, out7;
 
-    mask0 = LD_UB(&vp8_mc_filt_mask_arr[0]);
-    src -= 2;
+  mask0 = LD_UB(&vp8_mc_filt_mask_arr[0]);
+  src -= 2;
 
-    filt = LD_SH(filter);
-    SPLATI_H3_SB(filt, 0, 1, 2, filt0, filt1, filt2);
+  filt = LD_SH(filter);
+  SPLATI_H3_SB(filt, 0, 1, 2, filt0, filt1, filt2);
 
-    mask1 = mask0 + 2;
-    mask2 = mask0 + 4;
+  mask1 = mask0 + 2;
+  mask2 = mask0 + 4;
 
-    for (loop_cnt = (height >> 2); loop_cnt--;)
-    {
-        LD_SB4(src, src_stride, src0, src2, src4, src6);
-        LD_SB4(src + 8, src_stride, src1, src3, src5, src7);
-        XORI_B8_128_SB(src0, src1, src2, src3, src4, src5, src6, src7);
-        src += (4 * src_stride);
+  for (loop_cnt = (height >> 2); loop_cnt--;) {
+    LD_SB4(src, src_stride, src0, src2, src4, src6);
+    LD_SB4(src + 8, src_stride, src1, src3, src5, src7);
+    XORI_B8_128_SB(src0, src1, src2, src3, src4, src5, src6, src7);
+    src += (4 * src_stride);
 
-        HORIZ_6TAP_8WID_4VECS_FILT(src0, src1, src2, src3, mask0, mask1, mask2,
-                                   filt0, filt1, filt2, out0, out1, out2, out3);
-        HORIZ_6TAP_8WID_4VECS_FILT(src4, src5, src6, src7, mask0, mask1, mask2,
-                                   filt0, filt1, filt2, out4, out5, out6, out7);
-        SRARI_H4_SH(out0, out1, out2, out3, VP8_FILTER_SHIFT);
-        SRARI_H4_SH(out4, out5, out6, out7, VP8_FILTER_SHIFT);
-        SAT_SH4_SH(out0, out1, out2, out3, 7);
-        SAT_SH4_SH(out4, out5, out6, out7, 7);
-        out = PCKEV_XORI128_UB(out0, out1);
-        ST_UB(out, dst);
-        dst += dst_stride;
-        out = PCKEV_XORI128_UB(out2, out3);
-        ST_UB(out, dst);
-        dst += dst_stride;
-        out = PCKEV_XORI128_UB(out4, out5);
-        ST_UB(out, dst);
-        dst += dst_stride;
-        out = PCKEV_XORI128_UB(out6, out7);
-        ST_UB(out, dst);
-        dst += dst_stride;
-    }
+    HORIZ_6TAP_8WID_4VECS_FILT(src0, src1, src2, src3, mask0, mask1, mask2,
+                               filt0, filt1, filt2, out0, out1, out2, out3);
+    HORIZ_6TAP_8WID_4VECS_FILT(src4, src5, src6, src7, mask0, mask1, mask2,
+                               filt0, filt1, filt2, out4, out5, out6, out7);
+    SRARI_H4_SH(out0, out1, out2, out3, VP8_FILTER_SHIFT);
+    SRARI_H4_SH(out4, out5, out6, out7, VP8_FILTER_SHIFT);
+    SAT_SH4_SH(out0, out1, out2, out3, 7);
+    SAT_SH4_SH(out4, out5, out6, out7, 7);
+    out = PCKEV_XORI128_UB(out0, out1);
+    ST_UB(out, dst);
+    dst += dst_stride;
+    out = PCKEV_XORI128_UB(out2, out3);
+    ST_UB(out, dst);
+    dst += dst_stride;
+    out = PCKEV_XORI128_UB(out4, out5);
+    ST_UB(out, dst);
+    dst += dst_stride;
+    out = PCKEV_XORI128_UB(out6, out7);
+    ST_UB(out, dst);
+    dst += dst_stride;
+  }
 }
 
 static void common_vt_6t_4w_msa(uint8_t *RESTRICT src, int32_t src_stride,
                                 uint8_t *RESTRICT dst, int32_t dst_stride,
-                                const int8_t *filter, int32_t height)
-{
-    uint32_t loop_cnt;
-    v16i8 src0, src1, src2, src3, src4, src5, src6, src7, src8;
-    v16i8 src10_r, src32_r, src54_r, src76_r, src21_r, src43_r, src65_r;
-    v16i8 src87_r, src2110, src4332, src6554, src8776, filt0, filt1, filt2;
-    v16u8 out;
-    v8i16 filt, out10, out32;
+                                const int8_t *filter, int32_t height) {
+  uint32_t loop_cnt;
+  v16i8 src0, src1, src2, src3, src4, src5, src6, src7, src8;
+  v16i8 src10_r, src32_r, src54_r, src76_r, src21_r, src43_r, src65_r;
+  v16i8 src87_r, src2110, src4332, src6554, src8776, filt0, filt1, filt2;
+  v16u8 out;
+  v8i16 filt, out10, out32;
 
-    src -= (2 * src_stride);
+  src -= (2 * src_stride);
 
-    filt = LD_SH(filter);
-    SPLATI_H3_SB(filt, 0, 1, 2, filt0, filt1, filt2);
+  filt = LD_SH(filter);
+  SPLATI_H3_SB(filt, 0, 1, 2, filt0, filt1, filt2);
 
-    LD_SB5(src, src_stride, src0, src1, src2, src3, src4);
-    src += (5 * src_stride);
+  LD_SB5(src, src_stride, src0, src1, src2, src3, src4);
+  src += (5 * src_stride);
 
-    ILVR_B4_SB(src1, src0, src2, src1, src3, src2, src4, src3, src10_r, src21_r,
-               src32_r, src43_r);
-    ILVR_D2_SB(src21_r, src10_r, src43_r, src32_r, src2110, src4332);
-    XORI_B2_128_SB(src2110, src4332);
+  ILVR_B4_SB(src1, src0, src2, src1, src3, src2, src4, src3, src10_r, src21_r,
+             src32_r, src43_r);
+  ILVR_D2_SB(src21_r, src10_r, src43_r, src32_r, src2110, src4332);
+  XORI_B2_128_SB(src2110, src4332);
 
-    for (loop_cnt = (height >> 2); loop_cnt--;)
-    {
-        LD_SB4(src, src_stride, src5, src6, src7, src8);
-        src += (4 * src_stride);
+  for (loop_cnt = (height >> 2); loop_cnt--;) {
+    LD_SB4(src, src_stride, src5, src6, src7, src8);
+    src += (4 * src_stride);
 
-        ILVR_B4_SB(src5, src4, src6, src5, src7, src6, src8, src7, src54_r,
-                   src65_r, src76_r, src87_r);
-        ILVR_D2_SB(src65_r, src54_r, src87_r, src76_r, src6554, src8776);
-        XORI_B2_128_SB(src6554, src8776);
-        out10 = DPADD_SH3_SH(src2110, src4332, src6554, filt0, filt1, filt2);
-        out32 = DPADD_SH3_SH(src4332, src6554, src8776, filt0, filt1, filt2);
-        SRARI_H2_SH(out10, out32, VP8_FILTER_SHIFT);
-        SAT_SH2_SH(out10, out32, 7);
-        out = PCKEV_XORI128_UB(out10, out32);
-        ST4x4_UB(out, out, 0, 1, 2, 3, dst, dst_stride);
-        dst += (4 * dst_stride);
+    ILVR_B4_SB(src5, src4, src6, src5, src7, src6, src8, src7, src54_r, src65_r,
+               src76_r, src87_r);
+    ILVR_D2_SB(src65_r, src54_r, src87_r, src76_r, src6554, src8776);
+    XORI_B2_128_SB(src6554, src8776);
+    out10 = DPADD_SH3_SH(src2110, src4332, src6554, filt0, filt1, filt2);
+    out32 = DPADD_SH3_SH(src4332, src6554, src8776, filt0, filt1, filt2);
+    SRARI_H2_SH(out10, out32, VP8_FILTER_SHIFT);
+    SAT_SH2_SH(out10, out32, 7);
+    out = PCKEV_XORI128_UB(out10, out32);
+    ST4x4_UB(out, out, 0, 1, 2, 3, dst, dst_stride);
+    dst += (4 * dst_stride);
 
-        src2110 = src6554;
-        src4332 = src8776;
-        src4 = src8;
-    }
+    src2110 = src6554;
+    src4332 = src8776;
+    src4 = src8;
+  }
 }
 
 static void common_vt_6t_8w_msa(uint8_t *RESTRICT src, int32_t src_stride,
                                 uint8_t *RESTRICT dst, int32_t dst_stride,
-                                const int8_t *filter, int32_t height)
-{
-    uint32_t loop_cnt;
-    v16i8 src0, src1, src2, src3, src4, src7, src8, src9, src10;
-    v16i8 src10_r, src32_r, src76_r, src98_r, src21_r, src43_r, src87_r;
-    v16i8 src109_r, filt0, filt1, filt2;
-    v16u8 tmp0, tmp1;
-    v8i16 filt, out0_r, out1_r, out2_r, out3_r;
+                                const int8_t *filter, int32_t height) {
+  uint32_t loop_cnt;
+  v16i8 src0, src1, src2, src3, src4, src7, src8, src9, src10;
+  v16i8 src10_r, src32_r, src76_r, src98_r, src21_r, src43_r, src87_r;
+  v16i8 src109_r, filt0, filt1, filt2;
+  v16u8 tmp0, tmp1;
+  v8i16 filt, out0_r, out1_r, out2_r, out3_r;
 
-    src -= (2 * src_stride);
+  src -= (2 * src_stride);
 
-    filt = LD_SH(filter);
-    SPLATI_H3_SB(filt, 0, 1, 2, filt0, filt1, filt2);
+  filt = LD_SH(filter);
+  SPLATI_H3_SB(filt, 0, 1, 2, filt0, filt1, filt2);
 
-    LD_SB5(src, src_stride, src0, src1, src2, src3, src4);
-    src += (5 * src_stride);
+  LD_SB5(src, src_stride, src0, src1, src2, src3, src4);
+  src += (5 * src_stride);
 
-    XORI_B5_128_SB(src0, src1, src2, src3, src4);
-    ILVR_B4_SB(src1, src0, src3, src2, src2, src1, src4, src3, src10_r, src32_r,
-               src21_r, src43_r);
+  XORI_B5_128_SB(src0, src1, src2, src3, src4);
+  ILVR_B4_SB(src1, src0, src3, src2, src2, src1, src4, src3, src10_r, src32_r,
+             src21_r, src43_r);
 
-    for (loop_cnt = (height >> 2); loop_cnt--;)
-    {
-        LD_SB4(src, src_stride, src7, src8, src9, src10);
-        XORI_B4_128_SB(src7, src8, src9, src10);
-        src += (4 * src_stride);
+  for (loop_cnt = (height >> 2); loop_cnt--;) {
+    LD_SB4(src, src_stride, src7, src8, src9, src10);
+    XORI_B4_128_SB(src7, src8, src9, src10);
+    src += (4 * src_stride);
 
-        ILVR_B4_SB(src7, src4, src8, src7, src9, src8, src10, src9, src76_r,
-                   src87_r, src98_r, src109_r);
-        out0_r = DPADD_SH3_SH(src10_r, src32_r, src76_r, filt0, filt1, filt2);
-        out1_r = DPADD_SH3_SH(src21_r, src43_r, src87_r, filt0, filt1, filt2);
-        out2_r = DPADD_SH3_SH(src32_r, src76_r, src98_r, filt0, filt1, filt2);
-        out3_r = DPADD_SH3_SH(src43_r, src87_r, src109_r, filt0, filt1, filt2);
-        SRARI_H4_SH(out0_r, out1_r, out2_r, out3_r, VP8_FILTER_SHIFT);
-        SAT_SH4_SH(out0_r, out1_r, out2_r, out3_r, 7);
-        tmp0 = PCKEV_XORI128_UB(out0_r, out1_r);
-        tmp1 = PCKEV_XORI128_UB(out2_r, out3_r);
-        ST8x4_UB(tmp0, tmp1, dst, dst_stride);
-        dst += (4 * dst_stride);
+    ILVR_B4_SB(src7, src4, src8, src7, src9, src8, src10, src9, src76_r,
+               src87_r, src98_r, src109_r);
+    out0_r = DPADD_SH3_SH(src10_r, src32_r, src76_r, filt0, filt1, filt2);
+    out1_r = DPADD_SH3_SH(src21_r, src43_r, src87_r, filt0, filt1, filt2);
+    out2_r = DPADD_SH3_SH(src32_r, src76_r, src98_r, filt0, filt1, filt2);
+    out3_r = DPADD_SH3_SH(src43_r, src87_r, src109_r, filt0, filt1, filt2);
+    SRARI_H4_SH(out0_r, out1_r, out2_r, out3_r, VP8_FILTER_SHIFT);
+    SAT_SH4_SH(out0_r, out1_r, out2_r, out3_r, 7);
+    tmp0 = PCKEV_XORI128_UB(out0_r, out1_r);
+    tmp1 = PCKEV_XORI128_UB(out2_r, out3_r);
+    ST8x4_UB(tmp0, tmp1, dst, dst_stride);
+    dst += (4 * dst_stride);
 
-        src10_r = src76_r;
-        src32_r = src98_r;
-        src21_r = src87_r;
-        src43_r = src109_r;
-        src4 = src10;
-    }
+    src10_r = src76_r;
+    src32_r = src98_r;
+    src21_r = src87_r;
+    src43_r = src109_r;
+    src4 = src10;
+  }
 }
 
 static void common_vt_6t_16w_msa(uint8_t *RESTRICT src, int32_t src_stride,
                                  uint8_t *RESTRICT dst, int32_t dst_stride,
-                                 const int8_t *filter, int32_t height)
-{
-    uint32_t loop_cnt;
-    v16i8 src0, src1, src2, src3, src4, src5, src6, src7, src8;
-    v16i8 src10_r, src32_r, src54_r, src76_r, src21_r, src43_r, src65_r;
-    v16i8 src87_r, src10_l, src32_l, src54_l, src76_l, src21_l, src43_l;
-    v16i8 src65_l, src87_l, filt0, filt1, filt2;
-    v16u8 tmp0, tmp1, tmp2, tmp3;
-    v8i16 out0_r, out1_r, out2_r, out3_r, out0_l, out1_l, out2_l, out3_l, filt;
+                                 const int8_t *filter, int32_t height) {
+  uint32_t loop_cnt;
+  v16i8 src0, src1, src2, src3, src4, src5, src6, src7, src8;
+  v16i8 src10_r, src32_r, src54_r, src76_r, src21_r, src43_r, src65_r;
+  v16i8 src87_r, src10_l, src32_l, src54_l, src76_l, src21_l, src43_l;
+  v16i8 src65_l, src87_l, filt0, filt1, filt2;
+  v16u8 tmp0, tmp1, tmp2, tmp3;
+  v8i16 out0_r, out1_r, out2_r, out3_r, out0_l, out1_l, out2_l, out3_l, filt;
 
-    src -= (2 * src_stride);
+  src -= (2 * src_stride);
 
-    filt = LD_SH(filter);
-    SPLATI_H3_SB(filt, 0, 1, 2, filt0, filt1, filt2);
+  filt = LD_SH(filter);
+  SPLATI_H3_SB(filt, 0, 1, 2, filt0, filt1, filt2);
 
-    LD_SB5(src, src_stride, src0, src1, src2, src3, src4);
-    src += (5 * src_stride);
+  LD_SB5(src, src_stride, src0, src1, src2, src3, src4);
+  src += (5 * src_stride);
 
-    XORI_B5_128_SB(src0, src1, src2, src3, src4);
-    ILVR_B4_SB(src1, src0, src3, src2, src4, src3, src2, src1, src10_r,
-               src32_r, src43_r, src21_r);
-    ILVL_B4_SB(src1, src0, src3, src2, src4, src3, src2, src1, src10_l,
-               src32_l, src43_l, src21_l);
+  XORI_B5_128_SB(src0, src1, src2, src3, src4);
+  ILVR_B4_SB(src1, src0, src3, src2, src4, src3, src2, src1, src10_r, src32_r,
+             src43_r, src21_r);
+  ILVL_B4_SB(src1, src0, src3, src2, src4, src3, src2, src1, src10_l, src32_l,
+             src43_l, src21_l);
 
-    for (loop_cnt = (height >> 2); loop_cnt--;)
-    {
-        LD_SB4(src, src_stride, src5, src6, src7, src8);
-        src += (4 * src_stride);
+  for (loop_cnt = (height >> 2); loop_cnt--;) {
+    LD_SB4(src, src_stride, src5, src6, src7, src8);
+    src += (4 * src_stride);
 
-        XORI_B4_128_SB(src5, src6, src7, src8);
-        ILVR_B4_SB(src5, src4, src6, src5, src7, src6, src8, src7, src54_r,
-                   src65_r, src76_r, src87_r);
-        ILVL_B4_SB(src5, src4, src6, src5, src7, src6, src8, src7, src54_l,
-                   src65_l, src76_l, src87_l);
-        out0_r = DPADD_SH3_SH(src10_r, src32_r, src54_r, filt0, filt1, filt2);
-        out1_r = DPADD_SH3_SH(src21_r, src43_r, src65_r, filt0, filt1, filt2);
-        out2_r = DPADD_SH3_SH(src32_r, src54_r, src76_r, filt0, filt1, filt2);
-        out3_r = DPADD_SH3_SH(src43_r, src65_r, src87_r, filt0, filt1, filt2);
-        out0_l = DPADD_SH3_SH(src10_l, src32_l, src54_l, filt0, filt1, filt2);
-        out1_l = DPADD_SH3_SH(src21_l, src43_l, src65_l, filt0, filt1, filt2);
-        out2_l = DPADD_SH3_SH(src32_l, src54_l, src76_l, filt0, filt1, filt2);
-        out3_l = DPADD_SH3_SH(src43_l, src65_l, src87_l, filt0, filt1, filt2);
-        SRARI_H4_SH(out0_r, out1_r, out2_r, out3_r, VP8_FILTER_SHIFT);
-        SRARI_H4_SH(out0_l, out1_l, out2_l, out3_l, VP8_FILTER_SHIFT);
-        SAT_SH4_SH(out0_r, out1_r, out2_r, out3_r, 7);
-        SAT_SH4_SH(out0_l, out1_l, out2_l, out3_l, 7);
-        PCKEV_B4_UB(out0_l, out0_r, out1_l, out1_r, out2_l, out2_r, out3_l,
-                    out3_r, tmp0, tmp1, tmp2, tmp3);
-        XORI_B4_128_UB(tmp0, tmp1, tmp2, tmp3);
-        ST_UB4(tmp0, tmp1, tmp2, tmp3, dst, dst_stride);
-        dst += (4 * dst_stride);
+    XORI_B4_128_SB(src5, src6, src7, src8);
+    ILVR_B4_SB(src5, src4, src6, src5, src7, src6, src8, src7, src54_r, src65_r,
+               src76_r, src87_r);
+    ILVL_B4_SB(src5, src4, src6, src5, src7, src6, src8, src7, src54_l, src65_l,
+               src76_l, src87_l);
+    out0_r = DPADD_SH3_SH(src10_r, src32_r, src54_r, filt0, filt1, filt2);
+    out1_r = DPADD_SH3_SH(src21_r, src43_r, src65_r, filt0, filt1, filt2);
+    out2_r = DPADD_SH3_SH(src32_r, src54_r, src76_r, filt0, filt1, filt2);
+    out3_r = DPADD_SH3_SH(src43_r, src65_r, src87_r, filt0, filt1, filt2);
+    out0_l = DPADD_SH3_SH(src10_l, src32_l, src54_l, filt0, filt1, filt2);
+    out1_l = DPADD_SH3_SH(src21_l, src43_l, src65_l, filt0, filt1, filt2);
+    out2_l = DPADD_SH3_SH(src32_l, src54_l, src76_l, filt0, filt1, filt2);
+    out3_l = DPADD_SH3_SH(src43_l, src65_l, src87_l, filt0, filt1, filt2);
+    SRARI_H4_SH(out0_r, out1_r, out2_r, out3_r, VP8_FILTER_SHIFT);
+    SRARI_H4_SH(out0_l, out1_l, out2_l, out3_l, VP8_FILTER_SHIFT);
+    SAT_SH4_SH(out0_r, out1_r, out2_r, out3_r, 7);
+    SAT_SH4_SH(out0_l, out1_l, out2_l, out3_l, 7);
+    PCKEV_B4_UB(out0_l, out0_r, out1_l, out1_r, out2_l, out2_r, out3_l, out3_r,
+                tmp0, tmp1, tmp2, tmp3);
+    XORI_B4_128_UB(tmp0, tmp1, tmp2, tmp3);
+    ST_UB4(tmp0, tmp1, tmp2, tmp3, dst, dst_stride);
+    dst += (4 * dst_stride);
 
-        src10_r = src54_r;
-        src32_r = src76_r;
-        src21_r = src65_r;
-        src43_r = src87_r;
-        src10_l = src54_l;
-        src32_l = src76_l;
-        src21_l = src65_l;
-        src43_l = src87_l;
-        src4 = src8;
-    }
+    src10_r = src54_r;
+    src32_r = src76_r;
+    src21_r = src65_r;
+    src43_r = src87_r;
+    src10_l = src54_l;
+    src32_l = src76_l;
+    src21_l = src65_l;
+    src43_l = src87_l;
+    src4 = src8;
+  }
 }
 
 static void common_hv_6ht_6vt_4w_msa(uint8_t *RESTRICT src, int32_t src_stride,
@@ -476,74 +453,72 @@
                                      uint8_t *RESTRICT dst, int32_t dst_stride,
                                      const int8_t *filter_horiz,
                                      const int8_t *filter_vert,
-                                     int32_t height)
-{
-    uint32_t loop_cnt;
-    v16i8 src0, src1, src2, src3, src4, src5, src6, src7, src8;
-    v16i8 filt_hz0, filt_hz1, filt_hz2;
-    v16u8 mask0, mask1, mask2, out;
-    v8i16 tmp0, tmp1;
-    v8i16 hz_out0, hz_out1, hz_out2, hz_out3, hz_out4, hz_out5, hz_out6;
-    v8i16 hz_out7, filt, filt_vt0, filt_vt1, filt_vt2, out0, out1, out2, out3;
+                                     int32_t height) {
+  uint32_t loop_cnt;
+  v16i8 src0, src1, src2, src3, src4, src5, src6, src7, src8;
+  v16i8 filt_hz0, filt_hz1, filt_hz2;
+  v16u8 mask0, mask1, mask2, out;
+  v8i16 tmp0, tmp1;
+  v8i16 hz_out0, hz_out1, hz_out2, hz_out3, hz_out4, hz_out5, hz_out6;
+  v8i16 hz_out7, filt, filt_vt0, filt_vt1, filt_vt2, out0, out1, out2, out3;
 
-    mask0 = LD_UB(&vp8_mc_filt_mask_arr[16]);
-    src -= (2 + 2 * src_stride);
+  mask0 = LD_UB(&vp8_mc_filt_mask_arr[16]);
+  src -= (2 + 2 * src_stride);
 
-    filt = LD_SH(filter_horiz);
-    SPLATI_H3_SB(filt, 0, 1, 2, filt_hz0, filt_hz1, filt_hz2);
-    filt = LD_SH(filter_vert);
-    SPLATI_H3_SH(filt, 0, 1, 2, filt_vt0, filt_vt1, filt_vt2);
+  filt = LD_SH(filter_horiz);
+  SPLATI_H3_SB(filt, 0, 1, 2, filt_hz0, filt_hz1, filt_hz2);
+  filt = LD_SH(filter_vert);
+  SPLATI_H3_SH(filt, 0, 1, 2, filt_vt0, filt_vt1, filt_vt2);
 
-    mask1 = mask0 + 2;
-    mask2 = mask0 + 4;
+  mask1 = mask0 + 2;
+  mask2 = mask0 + 4;
 
-    LD_SB5(src, src_stride, src0, src1, src2, src3, src4);
-    src += (5 * src_stride);
+  LD_SB5(src, src_stride, src0, src1, src2, src3, src4);
+  src += (5 * src_stride);
 
-    XORI_B5_128_SB(src0, src1, src2, src3, src4);
-    hz_out0 = HORIZ_6TAP_FILT(src0, src1, mask0, mask1, mask2, filt_hz0,
-                              filt_hz1, filt_hz2);
-    hz_out2 = HORIZ_6TAP_FILT(src2, src3, mask0, mask1, mask2, filt_hz0,
-                              filt_hz1, filt_hz2);
-    hz_out1 = (v8i16)__msa_sldi_b((v16i8)hz_out2, (v16i8)hz_out0, 8);
-    hz_out3 = HORIZ_6TAP_FILT(src3, src4, mask0, mask1, mask2, filt_hz0,
-                              filt_hz1, filt_hz2);
-    ILVEV_B2_SH(hz_out0, hz_out1, hz_out2, hz_out3, out0, out1);
+  XORI_B5_128_SB(src0, src1, src2, src3, src4);
+  hz_out0 = HORIZ_6TAP_FILT(src0, src1, mask0, mask1, mask2, filt_hz0, filt_hz1,
+                            filt_hz2);
+  hz_out2 = HORIZ_6TAP_FILT(src2, src3, mask0, mask1, mask2, filt_hz0, filt_hz1,
+                            filt_hz2);
+  hz_out1 = (v8i16)__msa_sldi_b((v16i8)hz_out2, (v16i8)hz_out0, 8);
+  hz_out3 = HORIZ_6TAP_FILT(src3, src4, mask0, mask1, mask2, filt_hz0, filt_hz1,
+                            filt_hz2);
+  ILVEV_B2_SH(hz_out0, hz_out1, hz_out2, hz_out3, out0, out1);
 
-    for (loop_cnt = (height >> 2); loop_cnt--;)
-    {
-        LD_SB2(src, src_stride, src5, src6);
-        src += (2 * src_stride);
+  for (loop_cnt = (height >> 2); loop_cnt--;) {
+    LD_SB2(src, src_stride, src5, src6);
+    src += (2 * src_stride);
 
-        XORI_B2_128_SB(src5, src6);
-        hz_out5 = HORIZ_6TAP_FILT(src5, src6, mask0, mask1, mask2, filt_hz0,
-                                  filt_hz1, filt_hz2);
-        hz_out4 = (v8i16)__msa_sldi_b((v16i8)hz_out5, (v16i8)hz_out3, 8);
+    XORI_B2_128_SB(src5, src6);
+    hz_out5 = HORIZ_6TAP_FILT(src5, src6, mask0, mask1, mask2, filt_hz0,
+                              filt_hz1, filt_hz2);
+    hz_out4 = (v8i16)__msa_sldi_b((v16i8)hz_out5, (v16i8)hz_out3, 8);
 
-        LD_SB2(src, src_stride, src7, src8);
-        src += (2 * src_stride);
+    LD_SB2(src, src_stride, src7, src8);
+    src += (2 * src_stride);
 
-        XORI_B2_128_SB(src7, src8);
-        hz_out7 = HORIZ_6TAP_FILT(src7, src8, mask0, mask1, mask2, filt_hz0,
-                                  filt_hz1, filt_hz2);
-        hz_out6 = (v8i16)__msa_sldi_b((v16i8)hz_out7, (v16i8)hz_out5, 8);
+    XORI_B2_128_SB(src7, src8);
+    hz_out7 = HORIZ_6TAP_FILT(src7, src8, mask0, mask1, mask2, filt_hz0,
+                              filt_hz1, filt_hz2);
+    hz_out6 = (v8i16)__msa_sldi_b((v16i8)hz_out7, (v16i8)hz_out5, 8);
 
-        out2 = (v8i16)__msa_ilvev_b((v16i8)hz_out5, (v16i8)hz_out4);
-        tmp0 = DPADD_SH3_SH(out0, out1, out2, filt_vt0, filt_vt1, filt_vt2);
+    out2 = (v8i16)__msa_ilvev_b((v16i8)hz_out5, (v16i8)hz_out4);
+    tmp0 = DPADD_SH3_SH(out0, out1, out2, filt_vt0, filt_vt1, filt_vt2);
 
-        out3 = (v8i16)__msa_ilvev_b((v16i8)hz_out7, (v16i8)hz_out6);
-        tmp1 = DPADD_SH3_SH(out1, out2, out3, filt_vt0, filt_vt1, filt_vt2);
+    out3 = (v8i16)__msa_ilvev_b((v16i8)hz_out7, (v16i8)hz_out6);
+    tmp1 = DPADD_SH3_SH(out1, out2, out3, filt_vt0, filt_vt1, filt_vt2);
 
-        SRARI_H2_SH(tmp0, tmp1, 7);
-        SAT_SH2_SH(tmp0, tmp1, 7);
-        out = PCKEV_XORI128_UB(tmp0, tmp1);
-        ST4x4_UB(out, out, 0, 1, 2, 3, dst, dst_stride);
-        dst += (4 * dst_stride);
+    SRARI_H2_SH(tmp0, tmp1, 7);
+    SAT_SH2_SH(tmp0, tmp1, 7);
+    out = PCKEV_XORI128_UB(tmp0, tmp1);
+    ST4x4_UB(out, out, 0, 1, 2, 3, dst, dst_stride);
+    dst += (4 * dst_stride);
 
-        hz_out3 = hz_out7;
-        out0 = out2;
-        out1 = out3;
-    }
+    hz_out3 = hz_out7;
+    out0 = out2;
+    out1 = out3;
+  }
 }
 
 static void common_hv_6ht_6vt_8w_msa(uint8_t *RESTRICT src, int32_t src_stride,
@@ -550,86 +525,84 @@
                                      uint8_t *RESTRICT dst, int32_t dst_stride,
                                      const int8_t *filter_horiz,
                                      const int8_t *filter_vert,
-                                     int32_t height)
-{
-    uint32_t loop_cnt;
-    v16i8 src0, src1, src2, src3, src4, src5, src6, src7, src8;
-    v16i8 filt_hz0, filt_hz1, filt_hz2;
-    v16u8 mask0, mask1, mask2, vec0, vec1;
-    v8i16 filt, filt_vt0, filt_vt1, filt_vt2;
-    v8i16 hz_out0, hz_out1, hz_out2, hz_out3, hz_out4,  hz_out5, hz_out6;
-    v8i16 hz_out7, hz_out8, out0, out1, out2, out3, out4, out5, out6, out7;
-    v8i16 tmp0, tmp1, tmp2, tmp3;
+                                     int32_t height) {
+  uint32_t loop_cnt;
+  v16i8 src0, src1, src2, src3, src4, src5, src6, src7, src8;
+  v16i8 filt_hz0, filt_hz1, filt_hz2;
+  v16u8 mask0, mask1, mask2, vec0, vec1;
+  v8i16 filt, filt_vt0, filt_vt1, filt_vt2;
+  v8i16 hz_out0, hz_out1, hz_out2, hz_out3, hz_out4, hz_out5, hz_out6;
+  v8i16 hz_out7, hz_out8, out0, out1, out2, out3, out4, out5, out6, out7;
+  v8i16 tmp0, tmp1, tmp2, tmp3;
 
-    mask0 = LD_UB(&vp8_mc_filt_mask_arr[0]);
-    src -= (2 + 2 * src_stride);
+  mask0 = LD_UB(&vp8_mc_filt_mask_arr[0]);
+  src -= (2 + 2 * src_stride);
 
-    filt = LD_SH(filter_horiz);
-    SPLATI_H3_SB(filt, 0, 1, 2, filt_hz0, filt_hz1, filt_hz2);
+  filt = LD_SH(filter_horiz);
+  SPLATI_H3_SB(filt, 0, 1, 2, filt_hz0, filt_hz1, filt_hz2);
 
-    mask1 = mask0 + 2;
-    mask2 = mask0 + 4;
+  mask1 = mask0 + 2;
+  mask2 = mask0 + 4;
 
-    LD_SB5(src, src_stride, src0, src1, src2, src3, src4);
-    src += (5 * src_stride);
+  LD_SB5(src, src_stride, src0, src1, src2, src3, src4);
+  src += (5 * src_stride);
 
-    XORI_B5_128_SB(src0, src1, src2, src3, src4);
-    hz_out0 = HORIZ_6TAP_FILT(src0, src0, mask0, mask1, mask2, filt_hz0,
-                              filt_hz1, filt_hz2);
-    hz_out1 = HORIZ_6TAP_FILT(src1, src1, mask0, mask1, mask2, filt_hz0,
-                              filt_hz1, filt_hz2);
-    hz_out2 = HORIZ_6TAP_FILT(src2, src2, mask0, mask1, mask2, filt_hz0,
-                              filt_hz1, filt_hz2);
-    hz_out3 = HORIZ_6TAP_FILT(src3, src3, mask0, mask1, mask2, filt_hz0,
-                              filt_hz1, filt_hz2);
-    hz_out4 = HORIZ_6TAP_FILT(src4, src4, mask0, mask1, mask2, filt_hz0,
-                              filt_hz1, filt_hz2);
+  XORI_B5_128_SB(src0, src1, src2, src3, src4);
+  hz_out0 = HORIZ_6TAP_FILT(src0, src0, mask0, mask1, mask2, filt_hz0, filt_hz1,
+                            filt_hz2);
+  hz_out1 = HORIZ_6TAP_FILT(src1, src1, mask0, mask1, mask2, filt_hz0, filt_hz1,
+                            filt_hz2);
+  hz_out2 = HORIZ_6TAP_FILT(src2, src2, mask0, mask1, mask2, filt_hz0, filt_hz1,
+                            filt_hz2);
+  hz_out3 = HORIZ_6TAP_FILT(src3, src3, mask0, mask1, mask2, filt_hz0, filt_hz1,
+                            filt_hz2);
+  hz_out4 = HORIZ_6TAP_FILT(src4, src4, mask0, mask1, mask2, filt_hz0, filt_hz1,
+                            filt_hz2);
 
-    filt = LD_SH(filter_vert);
-    SPLATI_H3_SH(filt, 0, 1, 2, filt_vt0, filt_vt1, filt_vt2);
+  filt = LD_SH(filter_vert);
+  SPLATI_H3_SH(filt, 0, 1, 2, filt_vt0, filt_vt1, filt_vt2);
 
-    ILVEV_B2_SH(hz_out0, hz_out1, hz_out2, hz_out3, out0, out1);
-    ILVEV_B2_SH(hz_out1, hz_out2, hz_out3, hz_out4, out3, out4);
+  ILVEV_B2_SH(hz_out0, hz_out1, hz_out2, hz_out3, out0, out1);
+  ILVEV_B2_SH(hz_out1, hz_out2, hz_out3, hz_out4, out3, out4);
 
-    for (loop_cnt = (height >> 2); loop_cnt--;)
-    {
-        LD_SB4(src, src_stride, src5, src6, src7, src8);
-        src += (4 * src_stride);
+  for (loop_cnt = (height >> 2); loop_cnt--;) {
+    LD_SB4(src, src_stride, src5, src6, src7, src8);
+    src += (4 * src_stride);
 
-        XORI_B4_128_SB(src5, src6, src7, src8);
-        hz_out5 = HORIZ_6TAP_FILT(src5, src5, mask0, mask1, mask2, filt_hz0,
-                                  filt_hz1, filt_hz2);
-        out2 = (v8i16)__msa_ilvev_b((v16i8)hz_out5, (v16i8)hz_out4);
-        tmp0 = DPADD_SH3_SH(out0, out1, out2, filt_vt0, filt_vt1, filt_vt2);
+    XORI_B4_128_SB(src5, src6, src7, src8);
+    hz_out5 = HORIZ_6TAP_FILT(src5, src5, mask0, mask1, mask2, filt_hz0,
+                              filt_hz1, filt_hz2);
+    out2 = (v8i16)__msa_ilvev_b((v16i8)hz_out5, (v16i8)hz_out4);
+    tmp0 = DPADD_SH3_SH(out0, out1, out2, filt_vt0, filt_vt1, filt_vt2);
 
-        hz_out6 = HORIZ_6TAP_FILT(src6, src6, mask0, mask1, mask2, filt_hz0,
-                                  filt_hz1, filt_hz2);
-        out5 = (v8i16)__msa_ilvev_b((v16i8)hz_out6, (v16i8)hz_out5);
-        tmp1 = DPADD_SH3_SH(out3, out4, out5, filt_vt0, filt_vt1, filt_vt2);
+    hz_out6 = HORIZ_6TAP_FILT(src6, src6, mask0, mask1, mask2, filt_hz0,
+                              filt_hz1, filt_hz2);
+    out5 = (v8i16)__msa_ilvev_b((v16i8)hz_out6, (v16i8)hz_out5);
+    tmp1 = DPADD_SH3_SH(out3, out4, out5, filt_vt0, filt_vt1, filt_vt2);
 
-        hz_out7 = HORIZ_6TAP_FILT(src7, src7, mask0, mask1, mask2, filt_hz0,
-                                  filt_hz1, filt_hz2);
-        out7 = (v8i16)__msa_ilvev_b((v16i8)hz_out7, (v16i8)hz_out6);
-        tmp2 = DPADD_SH3_SH(out1, out2, out7, filt_vt0, filt_vt1, filt_vt2);
+    hz_out7 = HORIZ_6TAP_FILT(src7, src7, mask0, mask1, mask2, filt_hz0,
+                              filt_hz1, filt_hz2);
+    out7 = (v8i16)__msa_ilvev_b((v16i8)hz_out7, (v16i8)hz_out6);
+    tmp2 = DPADD_SH3_SH(out1, out2, out7, filt_vt0, filt_vt1, filt_vt2);
 
-        hz_out8 = HORIZ_6TAP_FILT(src8, src8, mask0, mask1, mask2, filt_hz0,
-                                  filt_hz1, filt_hz2);
-        out6 = (v8i16)__msa_ilvev_b((v16i8)hz_out8, (v16i8)hz_out7);
-        tmp3 = DPADD_SH3_SH(out4, out5, out6, filt_vt0, filt_vt1, filt_vt2);
+    hz_out8 = HORIZ_6TAP_FILT(src8, src8, mask0, mask1, mask2, filt_hz0,
+                              filt_hz1, filt_hz2);
+    out6 = (v8i16)__msa_ilvev_b((v16i8)hz_out8, (v16i8)hz_out7);
+    tmp3 = DPADD_SH3_SH(out4, out5, out6, filt_vt0, filt_vt1, filt_vt2);
 
-        SRARI_H4_SH(tmp0, tmp1, tmp2, tmp3, 7);
-        SAT_SH4_SH(tmp0, tmp1, tmp2, tmp3, 7);
-        vec0 = PCKEV_XORI128_UB(tmp0, tmp1);
-        vec1 = PCKEV_XORI128_UB(tmp2, tmp3);
-        ST8x4_UB(vec0, vec1, dst, dst_stride);
-        dst += (4 * dst_stride);
+    SRARI_H4_SH(tmp0, tmp1, tmp2, tmp3, 7);
+    SAT_SH4_SH(tmp0, tmp1, tmp2, tmp3, 7);
+    vec0 = PCKEV_XORI128_UB(tmp0, tmp1);
+    vec1 = PCKEV_XORI128_UB(tmp2, tmp3);
+    ST8x4_UB(vec0, vec1, dst, dst_stride);
+    dst += (4 * dst_stride);
 
-        hz_out4 = hz_out8;
-        out0 = out2;
-        out1 = out7;
-        out3 = out5;
-        out4 = out6;
-    }
+    hz_out4 = hz_out8;
+    out0 = out2;
+    out1 = out7;
+    out3 = out5;
+    out4 = out6;
+  }
 }
 
 static void common_hv_6ht_6vt_16w_msa(uint8_t *RESTRICT src, int32_t src_stride,
@@ -636,325 +609,307 @@
                                       uint8_t *RESTRICT dst, int32_t dst_stride,
                                       const int8_t *filter_horiz,
                                       const int8_t *filter_vert,
-                                      int32_t height)
-{
-    int32_t multiple8_cnt;
-    for (multiple8_cnt = 2; multiple8_cnt--;)
-    {
-        common_hv_6ht_6vt_8w_msa(src, src_stride, dst, dst_stride, filter_horiz,
-                                 filter_vert, height);
-        src += 8;
-        dst += 8;
-    }
+                                      int32_t height) {
+  int32_t multiple8_cnt;
+  for (multiple8_cnt = 2; multiple8_cnt--;) {
+    common_hv_6ht_6vt_8w_msa(src, src_stride, dst, dst_stride, filter_horiz,
+                             filter_vert, height);
+    src += 8;
+    dst += 8;
+  }
 }
 
 static void common_hz_4t_4x4_msa(uint8_t *RESTRICT src, int32_t src_stride,
                                  uint8_t *RESTRICT dst, int32_t dst_stride,
-                                 const int8_t *filter)
-{
-    v16i8 src0, src1, src2, src3, filt0, filt1, mask0, mask1;
-    v8i16 filt, out0, out1;
-    v16u8 out;
+                                 const int8_t *filter) {
+  v16i8 src0, src1, src2, src3, filt0, filt1, mask0, mask1;
+  v8i16 filt, out0, out1;
+  v16u8 out;
 
-    mask0 = LD_SB(&vp8_mc_filt_mask_arr[16]);
-    src -= 1;
+  mask0 = LD_SB(&vp8_mc_filt_mask_arr[16]);
+  src -= 1;
 
-    filt = LD_SH(filter);
-    SPLATI_H2_SB(filt, 0, 1, filt0, filt1);
+  filt = LD_SH(filter);
+  SPLATI_H2_SB(filt, 0, 1, filt0, filt1);
 
-    mask1 = mask0 + 2;
+  mask1 = mask0 + 2;
 
-    LD_SB4(src, src_stride, src0, src1, src2, src3);
-    XORI_B4_128_SB(src0, src1, src2, src3);
-    HORIZ_4TAP_4WID_4VECS_FILT(src0, src1, src2, src3, mask0, mask1,
-                               filt0, filt1, out0, out1);
-    SRARI_H2_SH(out0, out1, VP8_FILTER_SHIFT);
-    SAT_SH2_SH(out0, out1, 7);
-    out = PCKEV_XORI128_UB(out0, out1);
-    ST4x4_UB(out, out, 0, 1, 2, 3, dst, dst_stride);
+  LD_SB4(src, src_stride, src0, src1, src2, src3);
+  XORI_B4_128_SB(src0, src1, src2, src3);
+  HORIZ_4TAP_4WID_4VECS_FILT(src0, src1, src2, src3, mask0, mask1, filt0, filt1,
+                             out0, out1);
+  SRARI_H2_SH(out0, out1, VP8_FILTER_SHIFT);
+  SAT_SH2_SH(out0, out1, 7);
+  out = PCKEV_XORI128_UB(out0, out1);
+  ST4x4_UB(out, out, 0, 1, 2, 3, dst, dst_stride);
 }
 
 static void common_hz_4t_4x8_msa(uint8_t *RESTRICT src, int32_t src_stride,
                                  uint8_t *RESTRICT dst, int32_t dst_stride,
-                                 const int8_t *filter)
-{
-    v16i8 src0, src1, src2, src3, filt0, filt1, mask0, mask1;
-    v16u8 out;
-    v8i16 filt, out0, out1, out2, out3;
+                                 const int8_t *filter) {
+  v16i8 src0, src1, src2, src3, filt0, filt1, mask0, mask1;
+  v16u8 out;
+  v8i16 filt, out0, out1, out2, out3;
 
-    mask0 = LD_SB(&vp8_mc_filt_mask_arr[16]);
-    src -= 1;
+  mask0 = LD_SB(&vp8_mc_filt_mask_arr[16]);
+  src -= 1;
 
-    filt = LD_SH(filter);
-    SPLATI_H2_SB(filt, 0, 1, filt0, filt1);
+  filt = LD_SH(filter);
+  SPLATI_H2_SB(filt, 0, 1, filt0, filt1);
 
-    mask1 = mask0 + 2;
+  mask1 = mask0 + 2;
 
-    LD_SB4(src, src_stride, src0, src1, src2, src3);
-    src += (4 * src_stride);
+  LD_SB4(src, src_stride, src0, src1, src2, src3);
+  src += (4 * src_stride);
 
-    XORI_B4_128_SB(src0, src1, src2, src3);
-    HORIZ_4TAP_4WID_4VECS_FILT(src0, src1, src2, src3, mask0, mask1,
-                               filt0, filt1, out0, out1);
-    LD_SB4(src, src_stride, src0, src1, src2, src3);
-    XORI_B4_128_SB(src0, src1, src2, src3);
-    HORIZ_4TAP_4WID_4VECS_FILT(src0, src1, src2, src3, mask0, mask1,
-                               filt0, filt1, out2, out3);
-    SRARI_H4_SH(out0, out1, out2, out3, VP8_FILTER_SHIFT);
-    SAT_SH4_SH(out0, out1, out2, out3, 7);
-    out = PCKEV_XORI128_UB(out0, out1);
-    ST4x4_UB(out, out, 0, 1, 2, 3, dst, dst_stride);
-    dst += (4 * dst_stride);
-    out = PCKEV_XORI128_UB(out2, out3);
-    ST4x4_UB(out, out, 0, 1, 2, 3, dst, dst_stride);
+  XORI_B4_128_SB(src0, src1, src2, src3);
+  HORIZ_4TAP_4WID_4VECS_FILT(src0, src1, src2, src3, mask0, mask1, filt0, filt1,
+                             out0, out1);
+  LD_SB4(src, src_stride, src0, src1, src2, src3);
+  XORI_B4_128_SB(src0, src1, src2, src3);
+  HORIZ_4TAP_4WID_4VECS_FILT(src0, src1, src2, src3, mask0, mask1, filt0, filt1,
+                             out2, out3);
+  SRARI_H4_SH(out0, out1, out2, out3, VP8_FILTER_SHIFT);
+  SAT_SH4_SH(out0, out1, out2, out3, 7);
+  out = PCKEV_XORI128_UB(out0, out1);
+  ST4x4_UB(out, out, 0, 1, 2, 3, dst, dst_stride);
+  dst += (4 * dst_stride);
+  out = PCKEV_XORI128_UB(out2, out3);
+  ST4x4_UB(out, out, 0, 1, 2, 3, dst, dst_stride);
 }
 
 static void common_hz_4t_4w_msa(uint8_t *RESTRICT src, int32_t src_stride,
                                 uint8_t *RESTRICT dst, int32_t dst_stride,
-                                const int8_t *filter, int32_t height)
-{
-    if (4 == height)
-    {
-        common_hz_4t_4x4_msa(src, src_stride, dst, dst_stride, filter);
-    }
-    else if (8 == height)
-    {
-        common_hz_4t_4x8_msa(src, src_stride, dst, dst_stride, filter);
-    }
+                                const int8_t *filter, int32_t height) {
+  if (4 == height) {
+    common_hz_4t_4x4_msa(src, src_stride, dst, dst_stride, filter);
+  } else if (8 == height) {
+    common_hz_4t_4x8_msa(src, src_stride, dst, dst_stride, filter);
+  }
 }
 
 static void common_hz_4t_8w_msa(uint8_t *RESTRICT src, int32_t src_stride,
                                 uint8_t *RESTRICT dst, int32_t dst_stride,
-                                const int8_t *filter, int32_t height)
-{
-    uint32_t loop_cnt;
-    v16i8 src0, src1, src2, src3, filt0, filt1, mask0, mask1;
-    v16u8 tmp0, tmp1;
-    v8i16 filt, out0, out1, out2, out3;
+                                const int8_t *filter, int32_t height) {
+  uint32_t loop_cnt;
+  v16i8 src0, src1, src2, src3, filt0, filt1, mask0, mask1;
+  v16u8 tmp0, tmp1;
+  v8i16 filt, out0, out1, out2, out3;
 
-    mask0 = LD_SB(&vp8_mc_filt_mask_arr[0]);
-    src -= 1;
+  mask0 = LD_SB(&vp8_mc_filt_mask_arr[0]);
+  src -= 1;
 
-    filt = LD_SH(filter);
-    SPLATI_H2_SB(filt, 0, 1, filt0, filt1);
+  filt = LD_SH(filter);
+  SPLATI_H2_SB(filt, 0, 1, filt0, filt1);
 
-    mask1 = mask0 + 2;
+  mask1 = mask0 + 2;
 
-    for (loop_cnt = (height >> 2); loop_cnt--;)
-    {
-        LD_SB4(src, src_stride, src0, src1, src2, src3);
-        src += (4 * src_stride);
+  for (loop_cnt = (height >> 2); loop_cnt--;) {
+    LD_SB4(src, src_stride, src0, src1, src2, src3);
+    src += (4 * src_stride);
 
-        XORI_B4_128_SB(src0, src1, src2, src3);
-        HORIZ_4TAP_8WID_4VECS_FILT(src0, src1, src2, src3, mask0, mask1, filt0,
-                                   filt1, out0, out1, out2, out3);
-        SRARI_H4_SH(out0, out1, out2, out3, VP8_FILTER_SHIFT);
-        SAT_SH4_SH(out0, out1, out2, out3, 7);
-        tmp0 = PCKEV_XORI128_UB(out0, out1);
-        tmp1 = PCKEV_XORI128_UB(out2, out3);
-        ST8x4_UB(tmp0, tmp1, dst, dst_stride);
-        dst += (4 * dst_stride);
-    }
+    XORI_B4_128_SB(src0, src1, src2, src3);
+    HORIZ_4TAP_8WID_4VECS_FILT(src0, src1, src2, src3, mask0, mask1, filt0,
+                               filt1, out0, out1, out2, out3);
+    SRARI_H4_SH(out0, out1, out2, out3, VP8_FILTER_SHIFT);
+    SAT_SH4_SH(out0, out1, out2, out3, 7);
+    tmp0 = PCKEV_XORI128_UB(out0, out1);
+    tmp1 = PCKEV_XORI128_UB(out2, out3);
+    ST8x4_UB(tmp0, tmp1, dst, dst_stride);
+    dst += (4 * dst_stride);
+  }
 }
 
 static void common_hz_4t_16w_msa(uint8_t *RESTRICT src, int32_t src_stride,
                                  uint8_t *RESTRICT dst, int32_t dst_stride,
-                                 const int8_t *filter, int32_t height)
-{
-    uint32_t loop_cnt;
-    v16i8 src0, src1, src2, src3, src4, src5, src6, src7;
-    v16i8 filt0, filt1, mask0, mask1;
-    v8i16 filt, out0, out1, out2, out3, out4, out5, out6, out7;
-    v16u8 out;
+                                 const int8_t *filter, int32_t height) {
+  uint32_t loop_cnt;
+  v16i8 src0, src1, src2, src3, src4, src5, src6, src7;
+  v16i8 filt0, filt1, mask0, mask1;
+  v8i16 filt, out0, out1, out2, out3, out4, out5, out6, out7;
+  v16u8 out;
 
-    mask0 = LD_SB(&vp8_mc_filt_mask_arr[0]);
-    src -= 1;
+  mask0 = LD_SB(&vp8_mc_filt_mask_arr[0]);
+  src -= 1;
 
-    filt = LD_SH(filter);
-    SPLATI_H2_SB(filt, 0, 1, filt0, filt1);
+  filt = LD_SH(filter);
+  SPLATI_H2_SB(filt, 0, 1, filt0, filt1);
 
-    mask1 = mask0 + 2;
+  mask1 = mask0 + 2;
 
-    for (loop_cnt = (height >> 2); loop_cnt--;)
-    {
-        LD_SB4(src, src_stride, src0, src2, src4, src6);
-        LD_SB4(src + 8, src_stride, src1, src3, src5, src7);
-        src += (4 * src_stride);
+  for (loop_cnt = (height >> 2); loop_cnt--;) {
+    LD_SB4(src, src_stride, src0, src2, src4, src6);
+    LD_SB4(src + 8, src_stride, src1, src3, src5, src7);
+    src += (4 * src_stride);
 
-        XORI_B8_128_SB(src0, src1, src2, src3, src4, src5, src6, src7);
-        HORIZ_4TAP_8WID_4VECS_FILT(src0, src1, src2, src3, mask0, mask1, filt0,
-                                   filt1, out0, out1, out2, out3);
-        HORIZ_4TAP_8WID_4VECS_FILT(src4, src5, src6, src7, mask0, mask1, filt0,
-                                   filt1, out4, out5, out6, out7);
-        SRARI_H4_SH(out0, out1, out2, out3, VP8_FILTER_SHIFT);
-        SRARI_H4_SH(out4, out5, out6, out7, VP8_FILTER_SHIFT);
-        SAT_SH4_SH(out0, out1, out2, out3, 7);
-        SAT_SH4_SH(out4, out5, out6, out7, 7);
-        out = PCKEV_XORI128_UB(out0, out1);
-        ST_UB(out, dst);
-        dst += dst_stride;
-        out = PCKEV_XORI128_UB(out2, out3);
-        ST_UB(out, dst);
-        dst += dst_stride;
-        out = PCKEV_XORI128_UB(out4, out5);
-        ST_UB(out, dst);
-        dst += dst_stride;
-        out = PCKEV_XORI128_UB(out6, out7);
-        ST_UB(out, dst);
-        dst += dst_stride;
-    }
+    XORI_B8_128_SB(src0, src1, src2, src3, src4, src5, src6, src7);
+    HORIZ_4TAP_8WID_4VECS_FILT(src0, src1, src2, src3, mask0, mask1, filt0,
+                               filt1, out0, out1, out2, out3);
+    HORIZ_4TAP_8WID_4VECS_FILT(src4, src5, src6, src7, mask0, mask1, filt0,
+                               filt1, out4, out5, out6, out7);
+    SRARI_H4_SH(out0, out1, out2, out3, VP8_FILTER_SHIFT);
+    SRARI_H4_SH(out4, out5, out6, out7, VP8_FILTER_SHIFT);
+    SAT_SH4_SH(out0, out1, out2, out3, 7);
+    SAT_SH4_SH(out4, out5, out6, out7, 7);
+    out = PCKEV_XORI128_UB(out0, out1);
+    ST_UB(out, dst);
+    dst += dst_stride;
+    out = PCKEV_XORI128_UB(out2, out3);
+    ST_UB(out, dst);
+    dst += dst_stride;
+    out = PCKEV_XORI128_UB(out4, out5);
+    ST_UB(out, dst);
+    dst += dst_stride;
+    out = PCKEV_XORI128_UB(out6, out7);
+    ST_UB(out, dst);
+    dst += dst_stride;
+  }
 }
 
 static void common_vt_4t_4w_msa(uint8_t *RESTRICT src, int32_t src_stride,
                                 uint8_t *RESTRICT dst, int32_t dst_stride,
-                                const int8_t *filter, int32_t height)
-{
-    uint32_t loop_cnt;
-    v16i8 src0, src1, src2, src3, src4, src5;
-    v16i8 src10_r, src32_r, src54_r, src21_r, src43_r, src65_r;
-    v16i8 src2110, src4332, filt0, filt1;
-    v8i16 filt, out10, out32;
-    v16u8 out;
+                                const int8_t *filter, int32_t height) {
+  uint32_t loop_cnt;
+  v16i8 src0, src1, src2, src3, src4, src5;
+  v16i8 src10_r, src32_r, src54_r, src21_r, src43_r, src65_r;
+  v16i8 src2110, src4332, filt0, filt1;
+  v8i16 filt, out10, out32;
+  v16u8 out;
 
-    src -= src_stride;
+  src -= src_stride;
 
-    filt = LD_SH(filter);
-    SPLATI_H2_SB(filt, 0, 1, filt0, filt1);
+  filt = LD_SH(filter);
+  SPLATI_H2_SB(filt, 0, 1, filt0, filt1);
 
-    LD_SB3(src, src_stride, src0, src1, src2);
-    src += (3 * src_stride);
+  LD_SB3(src, src_stride, src0, src1, src2);
+  src += (3 * src_stride);
 
-    ILVR_B2_SB(src1, src0, src2, src1, src10_r, src21_r);
+  ILVR_B2_SB(src1, src0, src2, src1, src10_r, src21_r);
 
-    src2110 = (v16i8)__msa_ilvr_d((v2i64)src21_r, (v2i64)src10_r);
-    src2110 = (v16i8)__msa_xori_b((v16u8)src2110, 128);
+  src2110 = (v16i8)__msa_ilvr_d((v2i64)src21_r, (v2i64)src10_r);
+  src2110 = (v16i8)__msa_xori_b((v16u8)src2110, 128);
 
-    for (loop_cnt = (height >> 2); loop_cnt--;)
-    {
-        LD_SB3(src, src_stride, src3, src4, src5);
-        src += (3 * src_stride);
-        ILVR_B2_SB(src3, src2, src4, src3, src32_r, src43_r);
-        src4332 = (v16i8)__msa_ilvr_d((v2i64)src43_r, (v2i64)src32_r);
-        src4332 = (v16i8)__msa_xori_b((v16u8)src4332, 128);
-        out10 = FILT_4TAP_DPADD_S_H(src2110, src4332, filt0, filt1);
+  for (loop_cnt = (height >> 2); loop_cnt--;) {
+    LD_SB3(src, src_stride, src3, src4, src5);
+    src += (3 * src_stride);
+    ILVR_B2_SB(src3, src2, src4, src3, src32_r, src43_r);
+    src4332 = (v16i8)__msa_ilvr_d((v2i64)src43_r, (v2i64)src32_r);
+    src4332 = (v16i8)__msa_xori_b((v16u8)src4332, 128);
+    out10 = FILT_4TAP_DPADD_S_H(src2110, src4332, filt0, filt1);
 
-        src2 = LD_SB(src);
-        src += (src_stride);
-        ILVR_B2_SB(src5, src4, src2, src5, src54_r, src65_r);
-        src2110 = (v16i8)__msa_ilvr_d((v2i64)src65_r, (v2i64)src54_r);
-        src2110 = (v16i8)__msa_xori_b((v16u8)src2110, 128);
-        out32 = FILT_4TAP_DPADD_S_H(src4332, src2110, filt0, filt1);
-        SRARI_H2_SH(out10, out32, VP8_FILTER_SHIFT);
-        SAT_SH2_SH(out10, out32, 7);
-        out = PCKEV_XORI128_UB(out10, out32);
-        ST4x4_UB(out, out, 0, 1, 2, 3, dst, dst_stride);
-        dst += (4 * dst_stride);
-    }
+    src2 = LD_SB(src);
+    src += (src_stride);
+    ILVR_B2_SB(src5, src4, src2, src5, src54_r, src65_r);
+    src2110 = (v16i8)__msa_ilvr_d((v2i64)src65_r, (v2i64)src54_r);
+    src2110 = (v16i8)__msa_xori_b((v16u8)src2110, 128);
+    out32 = FILT_4TAP_DPADD_S_H(src4332, src2110, filt0, filt1);
+    SRARI_H2_SH(out10, out32, VP8_FILTER_SHIFT);
+    SAT_SH2_SH(out10, out32, 7);
+    out = PCKEV_XORI128_UB(out10, out32);
+    ST4x4_UB(out, out, 0, 1, 2, 3, dst, dst_stride);
+    dst += (4 * dst_stride);
+  }
 }
 
 static void common_vt_4t_8w_msa(uint8_t *RESTRICT src, int32_t src_stride,
                                 uint8_t *RESTRICT dst, int32_t dst_stride,
-                                const int8_t *filter, int32_t height)
-{
-    uint32_t loop_cnt;
-    v16i8 src0, src1, src2, src7, src8, src9, src10;
-    v16i8 src10_r, src72_r, src98_r, src21_r, src87_r, src109_r, filt0, filt1;
-    v16u8 tmp0, tmp1;
-    v8i16 filt, out0_r, out1_r, out2_r, out3_r;
+                                const int8_t *filter, int32_t height) {
+  uint32_t loop_cnt;
+  v16i8 src0, src1, src2, src7, src8, src9, src10;
+  v16i8 src10_r, src72_r, src98_r, src21_r, src87_r, src109_r, filt0, filt1;
+  v16u8 tmp0, tmp1;
+  v8i16 filt, out0_r, out1_r, out2_r, out3_r;
 
-    src -= src_stride;
+  src -= src_stride;
 
-    filt = LD_SH(filter);
-    SPLATI_H2_SB(filt, 0, 1, filt0, filt1);
+  filt = LD_SH(filter);
+  SPLATI_H2_SB(filt, 0, 1, filt0, filt1);
 
-    LD_SB3(src, src_stride, src0, src1, src2);
-    src += (3 * src_stride);
+  LD_SB3(src, src_stride, src0, src1, src2);
+  src += (3 * src_stride);
 
-    XORI_B3_128_SB(src0, src1, src2);
-    ILVR_B2_SB(src1, src0, src2, src1, src10_r, src21_r);
+  XORI_B3_128_SB(src0, src1, src2);
+  ILVR_B2_SB(src1, src0, src2, src1, src10_r, src21_r);
 
-    for (loop_cnt = (height >> 2); loop_cnt--;)
-    {
-        LD_SB4(src, src_stride, src7, src8, src9, src10);
-        src += (4 * src_stride);
+  for (loop_cnt = (height >> 2); loop_cnt--;) {
+    LD_SB4(src, src_stride, src7, src8, src9, src10);
+    src += (4 * src_stride);
 
-        XORI_B4_128_SB(src7, src8, src9, src10);
-        ILVR_B4_SB(src7, src2, src8, src7, src9, src8, src10, src9,
-                   src72_r, src87_r, src98_r, src109_r);
-        out0_r = FILT_4TAP_DPADD_S_H(src10_r, src72_r, filt0, filt1);
-        out1_r = FILT_4TAP_DPADD_S_H(src21_r, src87_r, filt0, filt1);
-        out2_r = FILT_4TAP_DPADD_S_H(src72_r, src98_r, filt0, filt1);
-        out3_r = FILT_4TAP_DPADD_S_H(src87_r, src109_r, filt0, filt1);
-        SRARI_H4_SH(out0_r, out1_r, out2_r, out3_r, VP8_FILTER_SHIFT);
-        SAT_SH4_SH(out0_r, out1_r, out2_r, out3_r, 7);
-        tmp0 = PCKEV_XORI128_UB(out0_r, out1_r);
-        tmp1 = PCKEV_XORI128_UB(out2_r, out3_r);
-        ST8x4_UB(tmp0, tmp1, dst, dst_stride);
-        dst += (4 * dst_stride);
+    XORI_B4_128_SB(src7, src8, src9, src10);
+    ILVR_B4_SB(src7, src2, src8, src7, src9, src8, src10, src9, src72_r,
+               src87_r, src98_r, src109_r);
+    out0_r = FILT_4TAP_DPADD_S_H(src10_r, src72_r, filt0, filt1);
+    out1_r = FILT_4TAP_DPADD_S_H(src21_r, src87_r, filt0, filt1);
+    out2_r = FILT_4TAP_DPADD_S_H(src72_r, src98_r, filt0, filt1);
+    out3_r = FILT_4TAP_DPADD_S_H(src87_r, src109_r, filt0, filt1);
+    SRARI_H4_SH(out0_r, out1_r, out2_r, out3_r, VP8_FILTER_SHIFT);
+    SAT_SH4_SH(out0_r, out1_r, out2_r, out3_r, 7);
+    tmp0 = PCKEV_XORI128_UB(out0_r, out1_r);
+    tmp1 = PCKEV_XORI128_UB(out2_r, out3_r);
+    ST8x4_UB(tmp0, tmp1, dst, dst_stride);
+    dst += (4 * dst_stride);
 
-        src10_r = src98_r;
-        src21_r = src109_r;
-        src2 = src10;
-    }
+    src10_r = src98_r;
+    src21_r = src109_r;
+    src2 = src10;
+  }
 }
 
 static void common_vt_4t_16w_msa(uint8_t *RESTRICT src, int32_t src_stride,
                                  uint8_t *RESTRICT dst, int32_t dst_stride,
-                                 const int8_t *filter, int32_t height)
-{
-    uint32_t loop_cnt;
-    v16i8 src0, src1, src2, src3, src4, src5, src6;
-    v16i8 src10_r, src32_r, src54_r, src21_r, src43_r, src65_r, src10_l;
-    v16i8 src32_l, src54_l, src21_l, src43_l, src65_l, filt0, filt1;
-    v16u8 tmp0, tmp1, tmp2, tmp3;
-    v8i16 filt, out0_r, out1_r, out2_r, out3_r, out0_l, out1_l, out2_l, out3_l;
+                                 const int8_t *filter, int32_t height) {
+  uint32_t loop_cnt;
+  v16i8 src0, src1, src2, src3, src4, src5, src6;
+  v16i8 src10_r, src32_r, src54_r, src21_r, src43_r, src65_r, src10_l;
+  v16i8 src32_l, src54_l, src21_l, src43_l, src65_l, filt0, filt1;
+  v16u8 tmp0, tmp1, tmp2, tmp3;
+  v8i16 filt, out0_r, out1_r, out2_r, out3_r, out0_l, out1_l, out2_l, out3_l;
 
-    src -= src_stride;
+  src -= src_stride;
 
-    filt = LD_SH(filter);
-    SPLATI_H2_SB(filt, 0, 1, filt0, filt1);
+  filt = LD_SH(filter);
+  SPLATI_H2_SB(filt, 0, 1, filt0, filt1);
 
-    LD_SB3(src, src_stride, src0, src1, src2);
-    src += (3 * src_stride);
+  LD_SB3(src, src_stride, src0, src1, src2);
+  src += (3 * src_stride);
 
-    XORI_B3_128_SB(src0, src1, src2);
-    ILVR_B2_SB(src1, src0, src2, src1, src10_r, src21_r);
-    ILVL_B2_SB(src1, src0, src2, src1, src10_l, src21_l);
+  XORI_B3_128_SB(src0, src1, src2);
+  ILVR_B2_SB(src1, src0, src2, src1, src10_r, src21_r);
+  ILVL_B2_SB(src1, src0, src2, src1, src10_l, src21_l);
 
-    for (loop_cnt = (height >> 2); loop_cnt--;)
-    {
-        LD_SB4(src, src_stride, src3, src4, src5, src6);
-        src += (4 * src_stride);
+  for (loop_cnt = (height >> 2); loop_cnt--;) {
+    LD_SB4(src, src_stride, src3, src4, src5, src6);
+    src += (4 * src_stride);
 
-        XORI_B4_128_SB(src3, src4, src5, src6);
-        ILVR_B4_SB(src3, src2, src4, src3, src5, src4, src6, src5,
-                   src32_r, src43_r, src54_r, src65_r);
-        ILVL_B4_SB(src3, src2, src4, src3, src5, src4, src6, src5,
-                   src32_l, src43_l, src54_l, src65_l);
-        out0_r = FILT_4TAP_DPADD_S_H(src10_r, src32_r, filt0, filt1);
-        out1_r = FILT_4TAP_DPADD_S_H(src21_r, src43_r, filt0, filt1);
-        out2_r = FILT_4TAP_DPADD_S_H(src32_r, src54_r, filt0, filt1);
-        out3_r = FILT_4TAP_DPADD_S_H(src43_r, src65_r, filt0, filt1);
-        out0_l = FILT_4TAP_DPADD_S_H(src10_l, src32_l, filt0, filt1);
-        out1_l = FILT_4TAP_DPADD_S_H(src21_l, src43_l, filt0, filt1);
-        out2_l = FILT_4TAP_DPADD_S_H(src32_l, src54_l, filt0, filt1);
-        out3_l = FILT_4TAP_DPADD_S_H(src43_l, src65_l, filt0, filt1);
-        SRARI_H4_SH(out0_r, out1_r, out2_r, out3_r, VP8_FILTER_SHIFT);
-        SRARI_H4_SH(out0_l, out1_l, out2_l, out3_l, VP8_FILTER_SHIFT);
-        SAT_SH4_SH(out0_r, out1_r, out2_r, out3_r, 7);
-        SAT_SH4_SH(out0_l, out1_l, out2_l, out3_l, 7);
-        PCKEV_B4_UB(out0_l, out0_r, out1_l, out1_r, out2_l, out2_r, out3_l,
-                    out3_r, tmp0, tmp1, tmp2, tmp3);
-        XORI_B4_128_UB(tmp0, tmp1, tmp2, tmp3);
-        ST_UB4(tmp0, tmp1, tmp2, tmp3, dst, dst_stride);
-        dst += (4 * dst_stride);
+    XORI_B4_128_SB(src3, src4, src5, src6);
+    ILVR_B4_SB(src3, src2, src4, src3, src5, src4, src6, src5, src32_r, src43_r,
+               src54_r, src65_r);
+    ILVL_B4_SB(src3, src2, src4, src3, src5, src4, src6, src5, src32_l, src43_l,
+               src54_l, src65_l);
+    out0_r = FILT_4TAP_DPADD_S_H(src10_r, src32_r, filt0, filt1);
+    out1_r = FILT_4TAP_DPADD_S_H(src21_r, src43_r, filt0, filt1);
+    out2_r = FILT_4TAP_DPADD_S_H(src32_r, src54_r, filt0, filt1);
+    out3_r = FILT_4TAP_DPADD_S_H(src43_r, src65_r, filt0, filt1);
+    out0_l = FILT_4TAP_DPADD_S_H(src10_l, src32_l, filt0, filt1);
+    out1_l = FILT_4TAP_DPADD_S_H(src21_l, src43_l, filt0, filt1);
+    out2_l = FILT_4TAP_DPADD_S_H(src32_l, src54_l, filt0, filt1);
+    out3_l = FILT_4TAP_DPADD_S_H(src43_l, src65_l, filt0, filt1);
+    SRARI_H4_SH(out0_r, out1_r, out2_r, out3_r, VP8_FILTER_SHIFT);
+    SRARI_H4_SH(out0_l, out1_l, out2_l, out3_l, VP8_FILTER_SHIFT);
+    SAT_SH4_SH(out0_r, out1_r, out2_r, out3_r, 7);
+    SAT_SH4_SH(out0_l, out1_l, out2_l, out3_l, 7);
+    PCKEV_B4_UB(out0_l, out0_r, out1_l, out1_r, out2_l, out2_r, out3_l, out3_r,
+                tmp0, tmp1, tmp2, tmp3);
+    XORI_B4_128_UB(tmp0, tmp1, tmp2, tmp3);
+    ST_UB4(tmp0, tmp1, tmp2, tmp3, dst, dst_stride);
+    dst += (4 * dst_stride);
 
-        src10_r = src54_r;
-        src21_r = src65_r;
-        src10_l = src54_l;
-        src21_l = src65_l;
-        src2 = src6;
-    }
+    src10_r = src54_r;
+    src21_r = src65_r;
+    src10_l = src54_l;
+    src21_l = src65_l;
+    src2 = src6;
+  }
 }
 
 static void common_hv_4ht_4vt_4w_msa(uint8_t *RESTRICT src, int32_t src_stride,
@@ -961,59 +916,57 @@
                                      uint8_t *RESTRICT dst, int32_t dst_stride,
                                      const int8_t *filter_horiz,
                                      const int8_t *filter_vert,
-                                     int32_t height)
-{
-    uint32_t loop_cnt;
-    v16i8 src0, src1, src2, src3, src4, src5, src6, filt_hz0, filt_hz1;
-    v16u8 mask0, mask1, out;
-    v8i16 filt, filt_vt0, filt_vt1, tmp0, tmp1, vec0, vec1, vec2;
-    v8i16 hz_out0, hz_out1, hz_out2, hz_out3, hz_out4, hz_out5;
+                                     int32_t height) {
+  uint32_t loop_cnt;
+  v16i8 src0, src1, src2, src3, src4, src5, src6, filt_hz0, filt_hz1;
+  v16u8 mask0, mask1, out;
+  v8i16 filt, filt_vt0, filt_vt1, tmp0, tmp1, vec0, vec1, vec2;
+  v8i16 hz_out0, hz_out1, hz_out2, hz_out3, hz_out4, hz_out5;
 
-    mask0 = LD_UB(&vp8_mc_filt_mask_arr[16]);
-    src -= (1 + 1 * src_stride);
+  mask0 = LD_UB(&vp8_mc_filt_mask_arr[16]);
+  src -= (1 + 1 * src_stride);
 
-    filt = LD_SH(filter_horiz);
-    SPLATI_H2_SB(filt, 0, 1, filt_hz0, filt_hz1);
+  filt = LD_SH(filter_horiz);
+  SPLATI_H2_SB(filt, 0, 1, filt_hz0, filt_hz1);
 
-    mask1 = mask0 + 2;
+  mask1 = mask0 + 2;
 
-    LD_SB3(src, src_stride, src0, src1, src2);
-    src += (3 * src_stride);
+  LD_SB3(src, src_stride, src0, src1, src2);
+  src += (3 * src_stride);
 
-    XORI_B3_128_SB(src0, src1, src2);
-    hz_out0 = HORIZ_4TAP_FILT(src0, src1, mask0, mask1, filt_hz0, filt_hz1);
-    hz_out1 = HORIZ_4TAP_FILT(src1, src2, mask0, mask1, filt_hz0, filt_hz1);
-    vec0 = (v8i16)__msa_ilvev_b((v16i8)hz_out1, (v16i8)hz_out0);
+  XORI_B3_128_SB(src0, src1, src2);
+  hz_out0 = HORIZ_4TAP_FILT(src0, src1, mask0, mask1, filt_hz0, filt_hz1);
+  hz_out1 = HORIZ_4TAP_FILT(src1, src2, mask0, mask1, filt_hz0, filt_hz1);
+  vec0 = (v8i16)__msa_ilvev_b((v16i8)hz_out1, (v16i8)hz_out0);
 
-    filt = LD_SH(filter_vert);
-    SPLATI_H2_SH(filt, 0, 1, filt_vt0, filt_vt1);
+  filt = LD_SH(filter_vert);
+  SPLATI_H2_SH(filt, 0, 1, filt_vt0, filt_vt1);
 
-    for (loop_cnt = (height >> 2); loop_cnt--;)
-    {
-        LD_SB4(src, src_stride, src3, src4, src5, src6);
-        src += (4 * src_stride);
+  for (loop_cnt = (height >> 2); loop_cnt--;) {
+    LD_SB4(src, src_stride, src3, src4, src5, src6);
+    src += (4 * src_stride);
 
-        XORI_B2_128_SB(src3, src4);
-        hz_out3 = HORIZ_4TAP_FILT(src3, src4, mask0, mask1, filt_hz0, filt_hz1);
-        hz_out2 = (v8i16)__msa_sldi_b((v16i8)hz_out3, (v16i8)hz_out1, 8);
-        vec1 = (v8i16)__msa_ilvev_b((v16i8)hz_out3, (v16i8)hz_out2);
-        tmp0 = FILT_4TAP_DPADD_S_H(vec0, vec1, filt_vt0, filt_vt1);
+    XORI_B2_128_SB(src3, src4);
+    hz_out3 = HORIZ_4TAP_FILT(src3, src4, mask0, mask1, filt_hz0, filt_hz1);
+    hz_out2 = (v8i16)__msa_sldi_b((v16i8)hz_out3, (v16i8)hz_out1, 8);
+    vec1 = (v8i16)__msa_ilvev_b((v16i8)hz_out3, (v16i8)hz_out2);
+    tmp0 = FILT_4TAP_DPADD_S_H(vec0, vec1, filt_vt0, filt_vt1);
 
-        XORI_B2_128_SB(src5, src6);
-        hz_out5 = HORIZ_4TAP_FILT(src5, src6, mask0, mask1, filt_hz0, filt_hz1);
-        hz_out4 = (v8i16)__msa_sldi_b((v16i8)hz_out5, (v16i8)hz_out3, 8);
-        vec2 = (v8i16)__msa_ilvev_b((v16i8)hz_out5, (v16i8)hz_out4);
-        tmp1 = FILT_4TAP_DPADD_S_H(vec1, vec2, filt_vt0, filt_vt1);
+    XORI_B2_128_SB(src5, src6);
+    hz_out5 = HORIZ_4TAP_FILT(src5, src6, mask0, mask1, filt_hz0, filt_hz1);
+    hz_out4 = (v8i16)__msa_sldi_b((v16i8)hz_out5, (v16i8)hz_out3, 8);
+    vec2 = (v8i16)__msa_ilvev_b((v16i8)hz_out5, (v16i8)hz_out4);
+    tmp1 = FILT_4TAP_DPADD_S_H(vec1, vec2, filt_vt0, filt_vt1);
 
-        SRARI_H2_SH(tmp0, tmp1, 7);
-        SAT_SH2_SH(tmp0, tmp1, 7);
-        out = PCKEV_XORI128_UB(tmp0, tmp1);
-        ST4x4_UB(out, out, 0, 1, 2, 3, dst, dst_stride);
-        dst += (4 * dst_stride);
+    SRARI_H2_SH(tmp0, tmp1, 7);
+    SAT_SH2_SH(tmp0, tmp1, 7);
+    out = PCKEV_XORI128_UB(tmp0, tmp1);
+    ST4x4_UB(out, out, 0, 1, 2, 3, dst, dst_stride);
+    dst += (4 * dst_stride);
 
-        hz_out1 = hz_out5;
-        vec0 = vec2;
-    }
+    hz_out1 = hz_out5;
+    vec0 = vec2;
+  }
 }
 
 static void common_hv_4ht_4vt_8w_msa(uint8_t *RESTRICT src, int32_t src_stride,
@@ -1020,67 +973,65 @@
                                      uint8_t *RESTRICT dst, int32_t dst_stride,
                                      const int8_t *filter_horiz,
                                      const int8_t *filter_vert,
-                                     int32_t height)
-{
-    uint32_t loop_cnt;
-    v16i8 src0, src1, src2, src3, src4, src5, src6, filt_hz0, filt_hz1;
-    v16u8 mask0, mask1, out0, out1;
-    v8i16 filt, filt_vt0, filt_vt1, tmp0, tmp1, tmp2, tmp3;
-    v8i16 hz_out0, hz_out1, hz_out2, hz_out3;
-    v8i16 vec0, vec1, vec2, vec3, vec4;
+                                     int32_t height) {
+  uint32_t loop_cnt;
+  v16i8 src0, src1, src2, src3, src4, src5, src6, filt_hz0, filt_hz1;
+  v16u8 mask0, mask1, out0, out1;
+  v8i16 filt, filt_vt0, filt_vt1, tmp0, tmp1, tmp2, tmp3;
+  v8i16 hz_out0, hz_out1, hz_out2, hz_out3;
+  v8i16 vec0, vec1, vec2, vec3, vec4;
 
-    mask0 = LD_UB(&vp8_mc_filt_mask_arr[0]);
-    src -= (1 + 1 * src_stride);
+  mask0 = LD_UB(&vp8_mc_filt_mask_arr[0]);
+  src -= (1 + 1 * src_stride);
 
-    filt = LD_SH(filter_horiz);
-    SPLATI_H2_SB(filt, 0, 1, filt_hz0, filt_hz1);
+  filt = LD_SH(filter_horiz);
+  SPLATI_H2_SB(filt, 0, 1, filt_hz0, filt_hz1);
 
-    mask1 = mask0 + 2;
+  mask1 = mask0 + 2;
 
-    LD_SB3(src, src_stride, src0, src1, src2);
-    src += (3 * src_stride);
+  LD_SB3(src, src_stride, src0, src1, src2);
+  src += (3 * src_stride);
 
-    XORI_B3_128_SB(src0, src1, src2);
-    hz_out0 = HORIZ_4TAP_FILT(src0, src0, mask0, mask1, filt_hz0, filt_hz1);
-    hz_out1 = HORIZ_4TAP_FILT(src1, src1, mask0, mask1, filt_hz0, filt_hz1);
-    hz_out2 = HORIZ_4TAP_FILT(src2, src2, mask0, mask1, filt_hz0, filt_hz1);
-    ILVEV_B2_SH(hz_out0, hz_out1, hz_out1, hz_out2, vec0, vec2);
+  XORI_B3_128_SB(src0, src1, src2);
+  hz_out0 = HORIZ_4TAP_FILT(src0, src0, mask0, mask1, filt_hz0, filt_hz1);
+  hz_out1 = HORIZ_4TAP_FILT(src1, src1, mask0, mask1, filt_hz0, filt_hz1);
+  hz_out2 = HORIZ_4TAP_FILT(src2, src2, mask0, mask1, filt_hz0, filt_hz1);
+  ILVEV_B2_SH(hz_out0, hz_out1, hz_out1, hz_out2, vec0, vec2);
 
-    filt = LD_SH(filter_vert);
-    SPLATI_H2_SH(filt, 0, 1, filt_vt0, filt_vt1);
+  filt = LD_SH(filter_vert);
+  SPLATI_H2_SH(filt, 0, 1, filt_vt0, filt_vt1);
 
-    for (loop_cnt = (height >> 2); loop_cnt--;)
-    {
-        LD_SB4(src, src_stride, src3, src4, src5, src6);
-        src += (4 * src_stride);
+  for (loop_cnt = (height >> 2); loop_cnt--;) {
+    LD_SB4(src, src_stride, src3, src4, src5, src6);
+    src += (4 * src_stride);
 
-        XORI_B4_128_SB(src3, src4, src5, src6);
-        hz_out3 = HORIZ_4TAP_FILT(src3, src3, mask0, mask1, filt_hz0, filt_hz1);
-        vec1 = (v8i16)__msa_ilvev_b((v16i8)hz_out3, (v16i8)hz_out2);
-        tmp0 = FILT_4TAP_DPADD_S_H(vec0, vec1, filt_vt0, filt_vt1);
+    XORI_B4_128_SB(src3, src4, src5, src6);
+    hz_out3 = HORIZ_4TAP_FILT(src3, src3, mask0, mask1, filt_hz0, filt_hz1);
+    vec1 = (v8i16)__msa_ilvev_b((v16i8)hz_out3, (v16i8)hz_out2);
+    tmp0 = FILT_4TAP_DPADD_S_H(vec0, vec1, filt_vt0, filt_vt1);
 
-        hz_out0 = HORIZ_4TAP_FILT(src4, src4, mask0, mask1, filt_hz0, filt_hz1);
-        vec3 = (v8i16)__msa_ilvev_b((v16i8)hz_out0, (v16i8)hz_out3);
-        tmp1 = FILT_4TAP_DPADD_S_H(vec2, vec3, filt_vt0, filt_vt1);
+    hz_out0 = HORIZ_4TAP_FILT(src4, src4, mask0, mask1, filt_hz0, filt_hz1);
+    vec3 = (v8i16)__msa_ilvev_b((v16i8)hz_out0, (v16i8)hz_out3);
+    tmp1 = FILT_4TAP_DPADD_S_H(vec2, vec3, filt_vt0, filt_vt1);
 
-        hz_out1 = HORIZ_4TAP_FILT(src5, src5, mask0, mask1, filt_hz0, filt_hz1);
-        vec4 = (v8i16)__msa_ilvev_b((v16i8)hz_out1, (v16i8)hz_out0);
-        tmp2 = FILT_4TAP_DPADD_S_H(vec1, vec4, filt_vt0, filt_vt1);
+    hz_out1 = HORIZ_4TAP_FILT(src5, src5, mask0, mask1, filt_hz0, filt_hz1);
+    vec4 = (v8i16)__msa_ilvev_b((v16i8)hz_out1, (v16i8)hz_out0);
+    tmp2 = FILT_4TAP_DPADD_S_H(vec1, vec4, filt_vt0, filt_vt1);
 
-        hz_out2 = HORIZ_4TAP_FILT(src6, src6, mask0, mask1, filt_hz0, filt_hz1);
-        ILVEV_B2_SH(hz_out3, hz_out0, hz_out1, hz_out2, vec0, vec1);
-        tmp3 = FILT_4TAP_DPADD_S_H(vec0, vec1, filt_vt0, filt_vt1);
+    hz_out2 = HORIZ_4TAP_FILT(src6, src6, mask0, mask1, filt_hz0, filt_hz1);
+    ILVEV_B2_SH(hz_out3, hz_out0, hz_out1, hz_out2, vec0, vec1);
+    tmp3 = FILT_4TAP_DPADD_S_H(vec0, vec1, filt_vt0, filt_vt1);
 
-        SRARI_H4_SH(tmp0, tmp1, tmp2, tmp3, 7);
-        SAT_SH4_SH(tmp0, tmp1, tmp2, tmp3, 7);
-        out0 = PCKEV_XORI128_UB(tmp0, tmp1);
-        out1 = PCKEV_XORI128_UB(tmp2, tmp3);
-        ST8x4_UB(out0, out1, dst, dst_stride);
-        dst += (4 * dst_stride);
+    SRARI_H4_SH(tmp0, tmp1, tmp2, tmp3, 7);
+    SAT_SH4_SH(tmp0, tmp1, tmp2, tmp3, 7);
+    out0 = PCKEV_XORI128_UB(tmp0, tmp1);
+    out1 = PCKEV_XORI128_UB(tmp2, tmp3);
+    ST8x4_UB(out0, out1, dst, dst_stride);
+    dst += (4 * dst_stride);
 
-        vec0 = vec4;
-        vec2 = vec1;
-    }
+    vec0 = vec4;
+    vec2 = vec1;
+  }
 }
 
 static void common_hv_4ht_4vt_16w_msa(uint8_t *RESTRICT src, int32_t src_stride,
@@ -1087,16 +1038,14 @@
                                       uint8_t *RESTRICT dst, int32_t dst_stride,
                                       const int8_t *filter_horiz,
                                       const int8_t *filter_vert,
-                                      int32_t height)
-{
-    int32_t multiple8_cnt;
-    for (multiple8_cnt = 2; multiple8_cnt--;)
-    {
-        common_hv_4ht_4vt_8w_msa(src, src_stride, dst, dst_stride, filter_horiz,
-                                 filter_vert, height);
-        src += 8;
-        dst += 8;
-    }
+                                      int32_t height) {
+  int32_t multiple8_cnt;
+  for (multiple8_cnt = 2; multiple8_cnt--;) {
+    common_hv_4ht_4vt_8w_msa(src, src_stride, dst, dst_stride, filter_horiz,
+                             filter_vert, height);
+    src += 8;
+    dst += 8;
+  }
 }
 
 static void common_hv_6ht_4vt_4w_msa(uint8_t *RESTRICT src, int32_t src_stride,
@@ -1103,65 +1052,63 @@
                                      uint8_t *RESTRICT dst, int32_t dst_stride,
                                      const int8_t *filter_horiz,
                                      const int8_t *filter_vert,
-                                     int32_t height)
-{
-    uint32_t loop_cnt;
-    v16i8 src0, src1, src2, src3, src4, src5, src6;
-    v16i8 filt_hz0, filt_hz1, filt_hz2;
-    v16u8 res0, res1, mask0, mask1, mask2;
-    v8i16 filt, filt_vt0, filt_vt1, tmp0, tmp1, vec0, vec1, vec2;
-    v8i16 hz_out0, hz_out1, hz_out2, hz_out3, hz_out4, hz_out5;
+                                     int32_t height) {
+  uint32_t loop_cnt;
+  v16i8 src0, src1, src2, src3, src4, src5, src6;
+  v16i8 filt_hz0, filt_hz1, filt_hz2;
+  v16u8 res0, res1, mask0, mask1, mask2;
+  v8i16 filt, filt_vt0, filt_vt1, tmp0, tmp1, vec0, vec1, vec2;
+  v8i16 hz_out0, hz_out1, hz_out2, hz_out3, hz_out4, hz_out5;
 
-    mask0 = LD_UB(&vp8_mc_filt_mask_arr[16]);
-    src -= (2 + 1 * src_stride);
+  mask0 = LD_UB(&vp8_mc_filt_mask_arr[16]);
+  src -= (2 + 1 * src_stride);
 
-    filt = LD_SH(filter_horiz);
-    SPLATI_H3_SB(filt, 0, 1, 2, filt_hz0, filt_hz1, filt_hz2);
+  filt = LD_SH(filter_horiz);
+  SPLATI_H3_SB(filt, 0, 1, 2, filt_hz0, filt_hz1, filt_hz2);
 
-    mask1 = mask0 + 2;
-    mask2 = mask0 + 4;
+  mask1 = mask0 + 2;
+  mask2 = mask0 + 4;
 
-    LD_SB3(src, src_stride, src0, src1, src2);
-    src += (3 * src_stride);
+  LD_SB3(src, src_stride, src0, src1, src2);
+  src += (3 * src_stride);
 
-    XORI_B3_128_SB(src0, src1, src2);
-    hz_out0 = HORIZ_6TAP_FILT(src0, src1, mask0, mask1, mask2, filt_hz0,
-                              filt_hz1, filt_hz2);
-    hz_out1 = HORIZ_6TAP_FILT(src1, src2, mask0, mask1, mask2, filt_hz0,
-                              filt_hz1, filt_hz2);
-    vec0 = (v8i16)__msa_ilvev_b((v16i8)hz_out1, (v16i8)hz_out0);
+  XORI_B3_128_SB(src0, src1, src2);
+  hz_out0 = HORIZ_6TAP_FILT(src0, src1, mask0, mask1, mask2, filt_hz0, filt_hz1,
+                            filt_hz2);
+  hz_out1 = HORIZ_6TAP_FILT(src1, src2, mask0, mask1, mask2, filt_hz0, filt_hz1,
+                            filt_hz2);
+  vec0 = (v8i16)__msa_ilvev_b((v16i8)hz_out1, (v16i8)hz_out0);
 
-    filt = LD_SH(filter_vert);
-    SPLATI_H2_SH(filt, 0, 1, filt_vt0, filt_vt1);
+  filt = LD_SH(filter_vert);
+  SPLATI_H2_SH(filt, 0, 1, filt_vt0, filt_vt1);
 
-    for (loop_cnt = (height >> 2); loop_cnt--;)
-    {
-        LD_SB4(src, src_stride, src3, src4, src5, src6);
-        src += (4 * src_stride);
+  for (loop_cnt = (height >> 2); loop_cnt--;) {
+    LD_SB4(src, src_stride, src3, src4, src5, src6);
+    src += (4 * src_stride);
 
-        XORI_B4_128_SB(src3, src4, src5, src6);
-        hz_out3 = HORIZ_6TAP_FILT(src3, src4, mask0, mask1, mask2, filt_hz0,
-                                  filt_hz1, filt_hz2);
-        hz_out2 = (v8i16)__msa_sldi_b((v16i8)hz_out3, (v16i8)hz_out1, 8);
-        vec1 = (v8i16)__msa_ilvev_b((v16i8)hz_out3, (v16i8)hz_out2);
-        tmp0 = FILT_4TAP_DPADD_S_H(vec0, vec1, filt_vt0, filt_vt1);
+    XORI_B4_128_SB(src3, src4, src5, src6);
+    hz_out3 = HORIZ_6TAP_FILT(src3, src4, mask0, mask1, mask2, filt_hz0,
+                              filt_hz1, filt_hz2);
+    hz_out2 = (v8i16)__msa_sldi_b((v16i8)hz_out3, (v16i8)hz_out1, 8);
+    vec1 = (v8i16)__msa_ilvev_b((v16i8)hz_out3, (v16i8)hz_out2);
+    tmp0 = FILT_4TAP_DPADD_S_H(vec0, vec1, filt_vt0, filt_vt1);
 
-        hz_out5 = HORIZ_6TAP_FILT(src5, src6, mask0, mask1, mask2, filt_hz0,
-                                  filt_hz1, filt_hz2);
-        hz_out4 = (v8i16)__msa_sldi_b((v16i8)hz_out5, (v16i8)hz_out3, 8);
-        vec2 = (v8i16)__msa_ilvev_b((v16i8)hz_out5, (v16i8)hz_out4);
-        tmp1 = FILT_4TAP_DPADD_S_H(vec1, vec2, filt_vt0, filt_vt1);
+    hz_out5 = HORIZ_6TAP_FILT(src5, src6, mask0, mask1, mask2, filt_hz0,
+                              filt_hz1, filt_hz2);
+    hz_out4 = (v8i16)__msa_sldi_b((v16i8)hz_out5, (v16i8)hz_out3, 8);
+    vec2 = (v8i16)__msa_ilvev_b((v16i8)hz_out5, (v16i8)hz_out4);
+    tmp1 = FILT_4TAP_DPADD_S_H(vec1, vec2, filt_vt0, filt_vt1);
 
-        SRARI_H2_SH(tmp0, tmp1, 7);
-        SAT_SH2_SH(tmp0, tmp1, 7);
-        PCKEV_B2_UB(tmp0, tmp0, tmp1, tmp1, res0, res1);
-        XORI_B2_128_UB(res0, res1);
-        ST4x4_UB(res0, res1, 0, 1, 0, 1, dst, dst_stride);
-        dst += (4 * dst_stride);
+    SRARI_H2_SH(tmp0, tmp1, 7);
+    SAT_SH2_SH(tmp0, tmp1, 7);
+    PCKEV_B2_UB(tmp0, tmp0, tmp1, tmp1, res0, res1);
+    XORI_B2_128_UB(res0, res1);
+    ST4x4_UB(res0, res1, 0, 1, 0, 1, dst, dst_stride);
+    dst += (4 * dst_stride);
 
-        hz_out1 = hz_out5;
-        vec0 = vec2;
-    }
+    hz_out1 = hz_out5;
+    vec0 = vec2;
+  }
 }
 
 static void common_hv_6ht_4vt_8w_msa(uint8_t *RESTRICT src, int32_t src_stride,
@@ -1168,73 +1115,71 @@
                                      uint8_t *RESTRICT dst, int32_t dst_stride,
                                      const int8_t *filter_horiz,
                                      const int8_t *filter_vert,
-                                     int32_t height)
-{
-    uint32_t loop_cnt;
-    v16i8 src0, src1, src2, src3, src4, src5, src6;
-    v16i8 filt_hz0, filt_hz1, filt_hz2, mask0, mask1, mask2;
-    v8i16 filt, filt_vt0, filt_vt1, hz_out0, hz_out1, hz_out2, hz_out3;
-    v8i16 tmp0, tmp1, tmp2, tmp3, vec0, vec1, vec2, vec3;
-    v16u8 out0, out1;
+                                     int32_t height) {
+  uint32_t loop_cnt;
+  v16i8 src0, src1, src2, src3, src4, src5, src6;
+  v16i8 filt_hz0, filt_hz1, filt_hz2, mask0, mask1, mask2;
+  v8i16 filt, filt_vt0, filt_vt1, hz_out0, hz_out1, hz_out2, hz_out3;
+  v8i16 tmp0, tmp1, tmp2, tmp3, vec0, vec1, vec2, vec3;
+  v16u8 out0, out1;
 
-    mask0 = LD_SB(&vp8_mc_filt_mask_arr[0]);
-    src -= (2 + src_stride);
+  mask0 = LD_SB(&vp8_mc_filt_mask_arr[0]);
+  src -= (2 + src_stride);
 
-    filt = LD_SH(filter_horiz);
-    SPLATI_H3_SB(filt, 0, 1, 2, filt_hz0, filt_hz1, filt_hz2);
+  filt = LD_SH(filter_horiz);
+  SPLATI_H3_SB(filt, 0, 1, 2, filt_hz0, filt_hz1, filt_hz2);
 
-    mask1 = mask0 + 2;
-    mask2 = mask0 + 4;
+  mask1 = mask0 + 2;
+  mask2 = mask0 + 4;
 
-    LD_SB3(src, src_stride, src0, src1, src2);
-    src += (3 * src_stride);
+  LD_SB3(src, src_stride, src0, src1, src2);
+  src += (3 * src_stride);
 
-    XORI_B3_128_SB(src0, src1, src2);
-    hz_out0 = HORIZ_6TAP_FILT(src0, src0, mask0, mask1, mask2, filt_hz0,
-                              filt_hz1, filt_hz2);
-    hz_out1 = HORIZ_6TAP_FILT(src1, src1, mask0, mask1, mask2, filt_hz0,
-                              filt_hz1, filt_hz2);
-    hz_out2 = HORIZ_6TAP_FILT(src2, src2, mask0, mask1, mask2, filt_hz0,
-                              filt_hz1, filt_hz2);
-    ILVEV_B2_SH(hz_out0, hz_out1, hz_out1, hz_out2, vec0, vec2);
+  XORI_B3_128_SB(src0, src1, src2);
+  hz_out0 = HORIZ_6TAP_FILT(src0, src0, mask0, mask1, mask2, filt_hz0, filt_hz1,
+                            filt_hz2);
+  hz_out1 = HORIZ_6TAP_FILT(src1, src1, mask0, mask1, mask2, filt_hz0, filt_hz1,
+                            filt_hz2);
+  hz_out2 = HORIZ_6TAP_FILT(src2, src2, mask0, mask1, mask2, filt_hz0, filt_hz1,
+                            filt_hz2);
+  ILVEV_B2_SH(hz_out0, hz_out1, hz_out1, hz_out2, vec0, vec2);
 
-    filt = LD_SH(filter_vert);
-    SPLATI_H2_SH(filt, 0, 1, filt_vt0, filt_vt1);
+  filt = LD_SH(filter_vert);
+  SPLATI_H2_SH(filt, 0, 1, filt_vt0, filt_vt1);
 
-    for (loop_cnt = (height >> 2); loop_cnt--;)
-    {
-        LD_SB4(src, src_stride, src3, src4, src5, src6);
-        src += (4 * src_stride);
+  for (loop_cnt = (height >> 2); loop_cnt--;) {
+    LD_SB4(src, src_stride, src3, src4, src5, src6);
+    src += (4 * src_stride);
 
-        XORI_B4_128_SB(src3, src4, src5, src6);
+    XORI_B4_128_SB(src3, src4, src5, src6);
 
-        hz_out3 = HORIZ_6TAP_FILT(src3, src3, mask0, mask1, mask2, filt_hz0,
-                                  filt_hz1, filt_hz2);
-        vec1 = (v8i16)__msa_ilvev_b((v16i8)hz_out3, (v16i8)hz_out2);
-        tmp0 = FILT_4TAP_DPADD_S_H(vec0, vec1, filt_vt0, filt_vt1);
+    hz_out3 = HORIZ_6TAP_FILT(src3, src3, mask0, mask1, mask2, filt_hz0,
+                              filt_hz1, filt_hz2);
+    vec1 = (v8i16)__msa_ilvev_b((v16i8)hz_out3, (v16i8)hz_out2);
+    tmp0 = FILT_4TAP_DPADD_S_H(vec0, vec1, filt_vt0, filt_vt1);
 
-        hz_out0 = HORIZ_6TAP_FILT(src4, src4, mask0, mask1, mask2, filt_hz0,
-                                  filt_hz1, filt_hz2);
-        vec3 = (v8i16)__msa_ilvev_b((v16i8)hz_out0, (v16i8)hz_out3);
-        tmp1 = FILT_4TAP_DPADD_S_H(vec2, vec3, filt_vt0, filt_vt1);
+    hz_out0 = HORIZ_6TAP_FILT(src4, src4, mask0, mask1, mask2, filt_hz0,
+                              filt_hz1, filt_hz2);
+    vec3 = (v8i16)__msa_ilvev_b((v16i8)hz_out0, (v16i8)hz_out3);
+    tmp1 = FILT_4TAP_DPADD_S_H(vec2, vec3, filt_vt0, filt_vt1);
 
-        hz_out1 = HORIZ_6TAP_FILT(src5, src5, mask0, mask1, mask2, filt_hz0,
-                                  filt_hz1, filt_hz2);
-        vec0 = (v8i16)__msa_ilvev_b((v16i8)hz_out1, (v16i8)hz_out0);
-        tmp2 = FILT_4TAP_DPADD_S_H(vec1, vec0, filt_vt0, filt_vt1);
+    hz_out1 = HORIZ_6TAP_FILT(src5, src5, mask0, mask1, mask2, filt_hz0,
+                              filt_hz1, filt_hz2);
+    vec0 = (v8i16)__msa_ilvev_b((v16i8)hz_out1, (v16i8)hz_out0);
+    tmp2 = FILT_4TAP_DPADD_S_H(vec1, vec0, filt_vt0, filt_vt1);
 
-        hz_out2 = HORIZ_6TAP_FILT(src6, src6, mask0, mask1, mask2, filt_hz0,
-                                  filt_hz1, filt_hz2);
-        ILVEV_B2_SH(hz_out3, hz_out0, hz_out1, hz_out2, vec1, vec2);
-        tmp3 = FILT_4TAP_DPADD_S_H(vec1, vec2, filt_vt0, filt_vt1);
+    hz_out2 = HORIZ_6TAP_FILT(src6, src6, mask0, mask1, mask2, filt_hz0,
+                              filt_hz1, filt_hz2);
+    ILVEV_B2_SH(hz_out3, hz_out0, hz_out1, hz_out2, vec1, vec2);
+    tmp3 = FILT_4TAP_DPADD_S_H(vec1, vec2, filt_vt0, filt_vt1);
 
-        SRARI_H4_SH(tmp0, tmp1, tmp2, tmp3, 7);
-        SAT_SH4_SH(tmp0, tmp1, tmp2, tmp3, 7);
-        out0 = PCKEV_XORI128_UB(tmp0, tmp1);
-        out1 = PCKEV_XORI128_UB(tmp2, tmp3);
-        ST8x4_UB(out0, out1, dst, dst_stride);
-        dst += (4 * dst_stride);
-    }
+    SRARI_H4_SH(tmp0, tmp1, tmp2, tmp3, 7);
+    SAT_SH4_SH(tmp0, tmp1, tmp2, tmp3, 7);
+    out0 = PCKEV_XORI128_UB(tmp0, tmp1);
+    out1 = PCKEV_XORI128_UB(tmp2, tmp3);
+    ST8x4_UB(out0, out1, dst, dst_stride);
+    dst += (4 * dst_stride);
+  }
 }
 
 static void common_hv_6ht_4vt_16w_msa(uint8_t *RESTRICT src, int32_t src_stride,
@@ -1241,16 +1186,14 @@
                                       uint8_t *RESTRICT dst, int32_t dst_stride,
                                       const int8_t *filter_horiz,
                                       const int8_t *filter_vert,
-                                      int32_t height)
-{
-    int32_t multiple8_cnt;
-    for (multiple8_cnt = 2; multiple8_cnt--;)
-    {
-        common_hv_6ht_4vt_8w_msa(src, src_stride, dst, dst_stride, filter_horiz,
-                                 filter_vert, height);
-        src += 8;
-        dst += 8;
-    }
+                                      int32_t height) {
+  int32_t multiple8_cnt;
+  for (multiple8_cnt = 2; multiple8_cnt--;) {
+    common_hv_6ht_4vt_8w_msa(src, src_stride, dst, dst_stride, filter_horiz,
+                             filter_vert, height);
+    src += 8;
+    dst += 8;
+  }
 }
 
 static void common_hv_4ht_6vt_4w_msa(uint8_t *RESTRICT src, int32_t src_stride,
@@ -1257,64 +1200,62 @@
                                      uint8_t *RESTRICT dst, int32_t dst_stride,
                                      const int8_t *filter_horiz,
                                      const int8_t *filter_vert,
-                                     int32_t height)
-{
-    uint32_t loop_cnt;
-    v16i8 src0, src1, src2, src3, src4, src5, src6, src7, src8;
-    v16i8 filt_hz0, filt_hz1, mask0, mask1;
-    v16u8 out;
-    v8i16 hz_out0, hz_out1, hz_out2, hz_out3, hz_out4, hz_out5, hz_out6;
-    v8i16 hz_out7, tmp0, tmp1, out0, out1, out2, out3;
-    v8i16 filt, filt_vt0, filt_vt1, filt_vt2;
+                                     int32_t height) {
+  uint32_t loop_cnt;
+  v16i8 src0, src1, src2, src3, src4, src5, src6, src7, src8;
+  v16i8 filt_hz0, filt_hz1, mask0, mask1;
+  v16u8 out;
+  v8i16 hz_out0, hz_out1, hz_out2, hz_out3, hz_out4, hz_out5, hz_out6;
+  v8i16 hz_out7, tmp0, tmp1, out0, out1, out2, out3;
+  v8i16 filt, filt_vt0, filt_vt1, filt_vt2;
 
-    mask0 = LD_SB(&vp8_mc_filt_mask_arr[16]);
+  mask0 = LD_SB(&vp8_mc_filt_mask_arr[16]);
 
-    src -= (1 + 2 * src_stride);
+  src -= (1 + 2 * src_stride);
 
-    filt = LD_SH(filter_horiz);
-    SPLATI_H2_SB(filt, 0, 1, filt_hz0, filt_hz1);
+  filt = LD_SH(filter_horiz);
+  SPLATI_H2_SB(filt, 0, 1, filt_hz0, filt_hz1);
 
-    mask1 = mask0 + 2;
+  mask1 = mask0 + 2;
 
-    LD_SB5(src, src_stride, src0, src1, src2, src3, src4);
-    src += (5 * src_stride);
+  LD_SB5(src, src_stride, src0, src1, src2, src3, src4);
+  src += (5 * src_stride);
 
-    XORI_B5_128_SB(src0, src1, src2, src3, src4);
-    hz_out0 = HORIZ_4TAP_FILT(src0, src1, mask0, mask1, filt_hz0, filt_hz1);
-    hz_out2 = HORIZ_4TAP_FILT(src2, src3, mask0, mask1, filt_hz0, filt_hz1);
-    hz_out3 = HORIZ_4TAP_FILT(src3, src4, mask0, mask1, filt_hz0, filt_hz1);
-    hz_out1 = (v8i16)__msa_sldi_b((v16i8)hz_out2, (v16i8)hz_out0, 8);
-    ILVEV_B2_SH(hz_out0, hz_out1, hz_out2, hz_out3, out0, out1);
+  XORI_B5_128_SB(src0, src1, src2, src3, src4);
+  hz_out0 = HORIZ_4TAP_FILT(src0, src1, mask0, mask1, filt_hz0, filt_hz1);
+  hz_out2 = HORIZ_4TAP_FILT(src2, src3, mask0, mask1, filt_hz0, filt_hz1);
+  hz_out3 = HORIZ_4TAP_FILT(src3, src4, mask0, mask1, filt_hz0, filt_hz1);
+  hz_out1 = (v8i16)__msa_sldi_b((v16i8)hz_out2, (v16i8)hz_out0, 8);
+  ILVEV_B2_SH(hz_out0, hz_out1, hz_out2, hz_out3, out0, out1);
 
-    filt = LD_SH(filter_vert);
-    SPLATI_H3_SH(filt, 0, 1, 2, filt_vt0, filt_vt1, filt_vt2);
+  filt = LD_SH(filter_vert);
+  SPLATI_H3_SH(filt, 0, 1, 2, filt_vt0, filt_vt1, filt_vt2);
 
-    for (loop_cnt = (height >> 2); loop_cnt--;)
-    {
-        LD_SB4(src, src_stride, src5, src6, src7, src8);
-        XORI_B4_128_SB(src5, src6, src7, src8);
-        src += (4 * src_stride);
+  for (loop_cnt = (height >> 2); loop_cnt--;) {
+    LD_SB4(src, src_stride, src5, src6, src7, src8);
+    XORI_B4_128_SB(src5, src6, src7, src8);
+    src += (4 * src_stride);
 
-        hz_out5 = HORIZ_4TAP_FILT(src5, src6, mask0, mask1, filt_hz0, filt_hz1);
-        hz_out4 = (v8i16)__msa_sldi_b((v16i8)hz_out5, (v16i8)hz_out3, 8);
-        out2 = (v8i16)__msa_ilvev_b((v16i8)hz_out5, (v16i8)hz_out4);
-        tmp0 = DPADD_SH3_SH(out0, out1, out2, filt_vt0, filt_vt1, filt_vt2);
+    hz_out5 = HORIZ_4TAP_FILT(src5, src6, mask0, mask1, filt_hz0, filt_hz1);
+    hz_out4 = (v8i16)__msa_sldi_b((v16i8)hz_out5, (v16i8)hz_out3, 8);
+    out2 = (v8i16)__msa_ilvev_b((v16i8)hz_out5, (v16i8)hz_out4);
+    tmp0 = DPADD_SH3_SH(out0, out1, out2, filt_vt0, filt_vt1, filt_vt2);
 
-        hz_out7 = HORIZ_4TAP_FILT(src7, src8, mask0, mask1, filt_hz0, filt_hz1);
-        hz_out6 = (v8i16)__msa_sldi_b((v16i8)hz_out7, (v16i8)hz_out5, 8);
-        out3 = (v8i16)__msa_ilvev_b((v16i8)hz_out7, (v16i8)hz_out6);
-        tmp1 = DPADD_SH3_SH(out1, out2, out3, filt_vt0, filt_vt1, filt_vt2);
+    hz_out7 = HORIZ_4TAP_FILT(src7, src8, mask0, mask1, filt_hz0, filt_hz1);
+    hz_out6 = (v8i16)__msa_sldi_b((v16i8)hz_out7, (v16i8)hz_out5, 8);
+    out3 = (v8i16)__msa_ilvev_b((v16i8)hz_out7, (v16i8)hz_out6);
+    tmp1 = DPADD_SH3_SH(out1, out2, out3, filt_vt0, filt_vt1, filt_vt2);
 
-        SRARI_H2_SH(tmp0, tmp1, 7);
-        SAT_SH2_SH(tmp0, tmp1, 7);
-        out = PCKEV_XORI128_UB(tmp0, tmp1);
-        ST4x4_UB(out, out, 0, 1, 2, 3, dst, dst_stride);
-        dst += (4 * dst_stride);
+    SRARI_H2_SH(tmp0, tmp1, 7);
+    SAT_SH2_SH(tmp0, tmp1, 7);
+    out = PCKEV_XORI128_UB(tmp0, tmp1);
+    ST4x4_UB(out, out, 0, 1, 2, 3, dst, dst_stride);
+    dst += (4 * dst_stride);
 
-        hz_out3 = hz_out7;
-        out0 = out2;
-        out1 = out3;
-    }
+    hz_out3 = hz_out7;
+    out0 = out2;
+    out1 = out3;
+  }
 }
 
 static void common_hv_4ht_6vt_8w_msa(uint8_t *RESTRICT src, int32_t src_stride,
@@ -1321,75 +1262,73 @@
                                      uint8_t *RESTRICT dst, int32_t dst_stride,
                                      const int8_t *filter_horiz,
                                      const int8_t *filter_vert,
-                                     int32_t height)
-{
-    uint32_t loop_cnt;
-    v16i8 src0, src1, src2, src3, src4, src5, src6, src7, src8;
-    v16i8 filt_hz0, filt_hz1, mask0, mask1;
-    v8i16 filt, filt_vt0, filt_vt1, filt_vt2, tmp0, tmp1, tmp2, tmp3;
-    v8i16 hz_out0, hz_out1, hz_out2, hz_out3, hz_out4, hz_out5, hz_out6;
-    v8i16 hz_out7, hz_out8, out0, out1, out2, out3, out4, out5, out6, out7;
-    v16u8 vec0, vec1;
+                                     int32_t height) {
+  uint32_t loop_cnt;
+  v16i8 src0, src1, src2, src3, src4, src5, src6, src7, src8;
+  v16i8 filt_hz0, filt_hz1, mask0, mask1;
+  v8i16 filt, filt_vt0, filt_vt1, filt_vt2, tmp0, tmp1, tmp2, tmp3;
+  v8i16 hz_out0, hz_out1, hz_out2, hz_out3, hz_out4, hz_out5, hz_out6;
+  v8i16 hz_out7, hz_out8, out0, out1, out2, out3, out4, out5, out6, out7;
+  v16u8 vec0, vec1;
 
-    mask0 = LD_SB(&vp8_mc_filt_mask_arr[0]);
-    src -= (1 + 2 * src_stride);
+  mask0 = LD_SB(&vp8_mc_filt_mask_arr[0]);
+  src -= (1 + 2 * src_stride);
 
-    filt = LD_SH(filter_horiz);
-    SPLATI_H2_SB(filt, 0, 1, filt_hz0, filt_hz1);
+  filt = LD_SH(filter_horiz);
+  SPLATI_H2_SB(filt, 0, 1, filt_hz0, filt_hz1);
 
-    mask1 = mask0 + 2;
+  mask1 = mask0 + 2;
 
-    LD_SB5(src, src_stride, src0, src1, src2, src3, src4);
-    src += (5 * src_stride);
+  LD_SB5(src, src_stride, src0, src1, src2, src3, src4);
+  src += (5 * src_stride);
 
-    XORI_B5_128_SB(src0, src1, src2, src3, src4);
-    hz_out0 = HORIZ_4TAP_FILT(src0, src0, mask0, mask1, filt_hz0, filt_hz1);
-    hz_out1 = HORIZ_4TAP_FILT(src1, src1, mask0, mask1, filt_hz0, filt_hz1);
-    hz_out2 = HORIZ_4TAP_FILT(src2, src2, mask0, mask1, filt_hz0, filt_hz1);
-    hz_out3 = HORIZ_4TAP_FILT(src3, src3, mask0, mask1, filt_hz0, filt_hz1);
-    hz_out4 = HORIZ_4TAP_FILT(src4, src4, mask0, mask1, filt_hz0, filt_hz1);
-    ILVEV_B2_SH(hz_out0, hz_out1, hz_out2, hz_out3, out0, out1);
-    ILVEV_B2_SH(hz_out1, hz_out2, hz_out3, hz_out4, out3, out4);
+  XORI_B5_128_SB(src0, src1, src2, src3, src4);
+  hz_out0 = HORIZ_4TAP_FILT(src0, src0, mask0, mask1, filt_hz0, filt_hz1);
+  hz_out1 = HORIZ_4TAP_FILT(src1, src1, mask0, mask1, filt_hz0, filt_hz1);
+  hz_out2 = HORIZ_4TAP_FILT(src2, src2, mask0, mask1, filt_hz0, filt_hz1);
+  hz_out3 = HORIZ_4TAP_FILT(src3, src3, mask0, mask1, filt_hz0, filt_hz1);
+  hz_out4 = HORIZ_4TAP_FILT(src4, src4, mask0, mask1, filt_hz0, filt_hz1);
+  ILVEV_B2_SH(hz_out0, hz_out1, hz_out2, hz_out3, out0, out1);
+  ILVEV_B2_SH(hz_out1, hz_out2, hz_out3, hz_out4, out3, out4);
 
-    filt = LD_SH(filter_vert);
-    SPLATI_H3_SH(filt, 0, 1, 2, filt_vt0, filt_vt1, filt_vt2);
+  filt = LD_SH(filter_vert);
+  SPLATI_H3_SH(filt, 0, 1, 2, filt_vt0, filt_vt1, filt_vt2);
 
-    for (loop_cnt = (height >> 2); loop_cnt--;)
-    {
-        LD_SB4(src, src_stride, src5, src6, src7, src8);
-        src += (4 * src_stride);
+  for (loop_cnt = (height >> 2); loop_cnt--;) {
+    LD_SB4(src, src_stride, src5, src6, src7, src8);
+    src += (4 * src_stride);
 
-        XORI_B4_128_SB(src5, src6, src7, src8);
+    XORI_B4_128_SB(src5, src6, src7, src8);
 
-        hz_out5 = HORIZ_4TAP_FILT(src5, src5, mask0, mask1, filt_hz0, filt_hz1);
-        out2 = (v8i16)__msa_ilvev_b((v16i8)hz_out5, (v16i8)hz_out4);
-        tmp0 = DPADD_SH3_SH(out0, out1, out2, filt_vt0, filt_vt1, filt_vt2);
+    hz_out5 = HORIZ_4TAP_FILT(src5, src5, mask0, mask1, filt_hz0, filt_hz1);
+    out2 = (v8i16)__msa_ilvev_b((v16i8)hz_out5, (v16i8)hz_out4);
+    tmp0 = DPADD_SH3_SH(out0, out1, out2, filt_vt0, filt_vt1, filt_vt2);
 
-        hz_out6 = HORIZ_4TAP_FILT(src6, src6, mask0, mask1, filt_hz0, filt_hz1);
-        out5 = (v8i16)__msa_ilvev_b((v16i8)hz_out6, (v16i8)hz_out5);
-        tmp1 = DPADD_SH3_SH(out3, out4, out5, filt_vt0, filt_vt1, filt_vt2);
+    hz_out6 = HORIZ_4TAP_FILT(src6, src6, mask0, mask1, filt_hz0, filt_hz1);
+    out5 = (v8i16)__msa_ilvev_b((v16i8)hz_out6, (v16i8)hz_out5);
+    tmp1 = DPADD_SH3_SH(out3, out4, out5, filt_vt0, filt_vt1, filt_vt2);
 
-        hz_out7 = HORIZ_4TAP_FILT(src7, src7, mask0, mask1, filt_hz0, filt_hz1);
-        out6 = (v8i16)__msa_ilvev_b((v16i8)hz_out7, (v16i8)hz_out6);
-        tmp2 = DPADD_SH3_SH(out1, out2, out6, filt_vt0, filt_vt1, filt_vt2);
+    hz_out7 = HORIZ_4TAP_FILT(src7, src7, mask0, mask1, filt_hz0, filt_hz1);
+    out6 = (v8i16)__msa_ilvev_b((v16i8)hz_out7, (v16i8)hz_out6);
+    tmp2 = DPADD_SH3_SH(out1, out2, out6, filt_vt0, filt_vt1, filt_vt2);
 
-        hz_out8 = HORIZ_4TAP_FILT(src8, src8, mask0, mask1, filt_hz0, filt_hz1);
-        out7 = (v8i16)__msa_ilvev_b((v16i8)hz_out8, (v16i8)hz_out7);
-        tmp3 = DPADD_SH3_SH(out4, out5, out7, filt_vt0, filt_vt1, filt_vt2);
+    hz_out8 = HORIZ_4TAP_FILT(src8, src8, mask0, mask1, filt_hz0, filt_hz1);
+    out7 = (v8i16)__msa_ilvev_b((v16i8)hz_out8, (v16i8)hz_out7);
+    tmp3 = DPADD_SH3_SH(out4, out5, out7, filt_vt0, filt_vt1, filt_vt2);
 
-        SRARI_H4_SH(tmp0, tmp1, tmp2, tmp3, 7);
-        SAT_SH4_SH(tmp0, tmp1, tmp2, tmp3, 7);
-        vec0 = PCKEV_XORI128_UB(tmp0, tmp1);
-        vec1 = PCKEV_XORI128_UB(tmp2, tmp3);
-        ST8x4_UB(vec0, vec1, dst, dst_stride);
-        dst += (4 * dst_stride);
+    SRARI_H4_SH(tmp0, tmp1, tmp2, tmp3, 7);
+    SAT_SH4_SH(tmp0, tmp1, tmp2, tmp3, 7);
+    vec0 = PCKEV_XORI128_UB(tmp0, tmp1);
+    vec1 = PCKEV_XORI128_UB(tmp2, tmp3);
+    ST8x4_UB(vec0, vec1, dst, dst_stride);
+    dst += (4 * dst_stride);
 
-        hz_out4 = hz_out8;
-        out0 = out2;
-        out1 = out6;
-        out3 = out5;
-        out4 = out7;
-    }
+    hz_out4 = hz_out8;
+    out0 = out2;
+    out1 = out6;
+    out3 = out5;
+    out4 = out7;
+  }
 }
 
 static void common_hv_4ht_6vt_16w_msa(uint8_t *RESTRICT src, int32_t src_stride,
@@ -1396,455 +1335,371 @@
                                       uint8_t *RESTRICT dst, int32_t dst_stride,
                                       const int8_t *filter_horiz,
                                       const int8_t *filter_vert,
-                                      int32_t height)
-{
-    int32_t multiple8_cnt;
-    for (multiple8_cnt = 2; multiple8_cnt--;)
-    {
-        common_hv_4ht_6vt_8w_msa(src, src_stride, dst, dst_stride, filter_horiz,
-                                 filter_vert, height);
-        src += 8;
-        dst += 8;
-    }
+                                      int32_t height) {
+  int32_t multiple8_cnt;
+  for (multiple8_cnt = 2; multiple8_cnt--;) {
+    common_hv_4ht_6vt_8w_msa(src, src_stride, dst, dst_stride, filter_horiz,
+                             filter_vert, height);
+    src += 8;
+    dst += 8;
+  }
 }
 
 void vp8_sixtap_predict4x4_msa(uint8_t *RESTRICT src, int32_t src_stride,
                                int32_t xoffset, int32_t yoffset,
-                               uint8_t *RESTRICT dst, int32_t dst_stride)
-{
-    const int8_t *h_filter = vp8_subpel_filters_msa[xoffset - 1];
-    const int8_t *v_filter = vp8_subpel_filters_msa[yoffset - 1];
+                               uint8_t *RESTRICT dst, int32_t dst_stride) {
+  const int8_t *h_filter = vp8_subpel_filters_msa[xoffset - 1];
+  const int8_t *v_filter = vp8_subpel_filters_msa[yoffset - 1];
 
-    if (yoffset)
-    {
-        if (xoffset)
-        {
-            switch (xoffset)
-            {
-                case 2:
-                case 4:
-                case 6:
-                    switch (yoffset)
-                    {
-                        case 2:
-                        case 4:
-                        case 6:
-                            common_hv_6ht_6vt_4w_msa(src, src_stride, dst,
-                                                     dst_stride, h_filter,
-                                                     v_filter, 4);
-                            break;
+  if (yoffset) {
+    if (xoffset) {
+      switch (xoffset) {
+        case 2:
+        case 4:
+        case 6:
+          switch (yoffset) {
+            case 2:
+            case 4:
+            case 6:
+              common_hv_6ht_6vt_4w_msa(src, src_stride, dst, dst_stride,
+                                       h_filter, v_filter, 4);
+              break;
 
-                        case 1:
-                        case 3:
-                        case 5:
-                        case 7:
-                            common_hv_6ht_4vt_4w_msa(src, src_stride, dst,
-                                                     dst_stride, h_filter,
-                                                     v_filter + 1, 4);
-                            break;
-                    }
-                    break;
+            case 1:
+            case 3:
+            case 5:
+            case 7:
+              common_hv_6ht_4vt_4w_msa(src, src_stride, dst, dst_stride,
+                                       h_filter, v_filter + 1, 4);
+              break;
+          }
+          break;
 
-                case 1:
-                case 3:
-                case 5:
-                case 7:
-                    switch (yoffset)
-                    {
-                        case 2:
-                        case 4:
-                        case 6:
-                            common_hv_4ht_6vt_4w_msa(src, src_stride, dst,
-                                                     dst_stride, h_filter + 1,
-                                                     v_filter, 4);
-                            break;
-
-                        case 1:
-                        case 3:
-                        case 5:
-                        case 7:
-                            common_hv_4ht_4vt_4w_msa(src, src_stride, dst,
-                                                     dst_stride, h_filter + 1,
-                                                     v_filter + 1, 4);
-                            break;
-                    }
-                    break;
-            }
-        }
-        else
-        {
-            switch (yoffset)
-            {
-                case 2:
-                case 4:
-                case 6:
-                    common_vt_6t_4w_msa(src, src_stride, dst, dst_stride,
-                                        v_filter, 4);
-                    break;
-
-                case 1:
-                case 3:
-                case 5:
-                case 7:
-                    common_vt_4t_4w_msa(src, src_stride, dst, dst_stride,
-                                        v_filter + 1, 4);
-                    break;
-            }
-        }
-    }
-    else
-    {
-        switch (xoffset)
-        {
-            case 0:
-                {
-                uint32_t tp0, tp1, tp2, tp3;
-
-                LW4(src, src_stride, tp0, tp1, tp2, tp3);
-                SW4(tp0, tp1, tp2, tp3, dst, dst_stride);
-                break;
-                }
+        case 1:
+        case 3:
+        case 5:
+        case 7:
+          switch (yoffset) {
             case 2:
             case 4:
             case 6:
-                common_hz_6t_4w_msa(src, src_stride, dst, dst_stride,
-                                    h_filter, 4);
-                break;
+              common_hv_4ht_6vt_4w_msa(src, src_stride, dst, dst_stride,
+                                       h_filter + 1, v_filter, 4);
+              break;
 
             case 1:
             case 3:
             case 5:
             case 7:
-                common_hz_4t_4w_msa(src, src_stride, dst, dst_stride,
-                                    h_filter + 1, 4);
-                break;
-        }
-    }
-}
-
-void vp8_sixtap_predict8x4_msa(uint8_t *RESTRICT src, int32_t src_stride,
-                               int32_t xoffset, int32_t yoffset,
-                               uint8_t *RESTRICT dst, int32_t dst_stride)
-{
-    const int8_t *h_filter = vp8_subpel_filters_msa[xoffset - 1];
-    const int8_t *v_filter = vp8_subpel_filters_msa[yoffset - 1];
-
-    if (yoffset)
-    {
-        if (xoffset)
-        {
-            switch (xoffset)
-            {
-                case 2:
-                case 4:
-                case 6:
-                    switch (yoffset)
-                    {
-                        case 2:
-                        case 4:
-                        case 6:
-                            common_hv_6ht_6vt_8w_msa(src, src_stride, dst,
-                                                     dst_stride, h_filter,
-                                                     v_filter, 4);
-                            break;
+              common_hv_4ht_4vt_4w_msa(src, src_stride, dst, dst_stride,
+                                       h_filter + 1, v_filter + 1, 4);
+              break;
+          }
+          break;
+      }
+    } else {
+      switch (yoffset) {
+        case 2:
+        case 4:
+        case 6:
+          common_vt_6t_4w_msa(src, src_stride, dst, dst_stride, v_filter, 4);
+          break;
 
-                        case 1:
-                        case 3:
-                        case 5:
-                        case 7:
-                            common_hv_6ht_4vt_8w_msa(src, src_stride, dst,
-                                                     dst_stride, h_filter,
-                                                     v_filter + 1, 4);
-                            break;
-                    }
-                    break;
+        case 1:
+        case 3:
+        case 5:
+        case 7:
+          common_vt_4t_4w_msa(src, src_stride, dst, dst_stride, v_filter + 1,
+                              4);
+          break;
+      }
+    }
+  } else {
+    switch (xoffset) {
+      case 0: {
+        uint32_t tp0, tp1, tp2, tp3;
 
-                case 1:
-                case 3:
-                case 5:
-                case 7:
-                    switch (yoffset)
-                    {
-                        case 2:
-                        case 4:
-                        case 6:
-                            common_hv_4ht_6vt_8w_msa(src, src_stride, dst,
-                                                     dst_stride, h_filter + 1,
-                                                     v_filter, 4);
-                            break;
+        LW4(src, src_stride, tp0, tp1, tp2, tp3);
+        SW4(tp0, tp1, tp2, tp3, dst, dst_stride);
+        break;
+      }
+      case 2:
+      case 4:
+      case 6:
+        common_hz_6t_4w_msa(src, src_stride, dst, dst_stride, h_filter, 4);
+        break;
+
+      case 1:
+      case 3:
+      case 5:
+      case 7:
+        common_hz_4t_4w_msa(src, src_stride, dst, dst_stride, h_filter + 1, 4);
+        break;
+    }
+  }
+}
+
+void vp8_sixtap_predict8x4_msa(uint8_t *RESTRICT src, int32_t src_stride,
+                               int32_t xoffset, int32_t yoffset,
+                               uint8_t *RESTRICT dst, int32_t dst_stride) {
+  const int8_t *h_filter = vp8_subpel_filters_msa[xoffset - 1];
+  const int8_t *v_filter = vp8_subpel_filters_msa[yoffset - 1];
+
+  if (yoffset) {
+    if (xoffset) {
+      switch (xoffset) {
+        case 2:
+        case 4:
+        case 6:
+          switch (yoffset) {
+            case 2:
+            case 4:
+            case 6:
+              common_hv_6ht_6vt_8w_msa(src, src_stride, dst, dst_stride,
+                                       h_filter, v_filter, 4);
+              break;
 
-                        case 1:
-                        case 3:
-                        case 5:
-                        case 7:
-                            common_hv_4ht_4vt_8w_msa(src, src_stride, dst,
-                                                     dst_stride, h_filter + 1,
-                                                     v_filter + 1, 4);
-                            break;
-                    }
-                    break;
-            }
-        }
-        else
-        {
-            switch (yoffset)
-            {
-                case 2:
-                case 4:
-                case 6:
-                    common_vt_6t_8w_msa(src, src_stride, dst, dst_stride,
-                                        v_filter, 4);
-                    break;
-
-                case 1:
-                case 3:
-                case 5:
-                case 7:
-                    common_vt_4t_8w_msa(src, src_stride, dst, dst_stride,
-                                        v_filter + 1, 4);
-                    break;
-            }
-        }
-    }
-    else
-    {
-        switch (xoffset)
-        {
-            case 0:
-                vp8_copy_mem8x4(src, src_stride, dst, dst_stride);
-                break;
+            case 1:
+            case 3:
+            case 5:
+            case 7:
+              common_hv_6ht_4vt_8w_msa(src, src_stride, dst, dst_stride,
+                                       h_filter, v_filter + 1, 4);
+              break;
+          }
+          break;
+
+        case 1:
+        case 3:
+        case 5:
+        case 7:
+          switch (yoffset) {
             case 2:
             case 4:
             case 6:
-                common_hz_6t_8w_msa(src, src_stride, dst, dst_stride,
-                                    h_filter, 4);
-                break;
+              common_hv_4ht_6vt_8w_msa(src, src_stride, dst, dst_stride,
+                                       h_filter + 1, v_filter, 4);
+              break;
 
             case 1:
             case 3:
             case 5:
             case 7:
-                common_hz_4t_8w_msa(src, src_stride, dst, dst_stride,
-                                    h_filter + 1, 4);
-                break;
-        }
-    }
-}
-
-void vp8_sixtap_predict8x8_msa(uint8_t *RESTRICT src, int32_t src_stride,
-                               int32_t xoffset, int32_t yoffset,
-                               uint8_t *RESTRICT dst, int32_t dst_stride)
-{
-    const int8_t *h_filter = vp8_subpel_filters_msa[xoffset - 1];
-    const int8_t *v_filter = vp8_subpel_filters_msa[yoffset - 1];
-
-    if (yoffset)
-    {
-        if (xoffset)
-        {
-            switch (xoffset)
-            {
-                case 2:
-                case 4:
-                case 6:
-                    switch (yoffset)
-                    {
-                        case 2:
-                        case 4:
-                        case 6:
-                            common_hv_6ht_6vt_8w_msa(src, src_stride, dst,
-                                                     dst_stride, h_filter,
-                                                     v_filter, 8);
-                            break;
+              common_hv_4ht_4vt_8w_msa(src, src_stride, dst, dst_stride,
+                                       h_filter + 1, v_filter + 1, 4);
+              break;
+          }
+          break;
+      }
+    } else {
+      switch (yoffset) {
+        case 2:
+        case 4:
+        case 6:
+          common_vt_6t_8w_msa(src, src_stride, dst, dst_stride, v_filter, 4);
+          break;
 
-                        case 1:
-                        case 3:
-                        case 5:
-                        case 7:
-                            common_hv_6ht_4vt_8w_msa(src, src_stride, dst,
-                                                     dst_stride, h_filter,
-                                                     v_filter + 1, 8);
-                            break;
-                    }
-                    break;
+        case 1:
+        case 3:
+        case 5:
+        case 7:
+          common_vt_4t_8w_msa(src, src_stride, dst, dst_stride, v_filter + 1,
+                              4);
+          break;
+      }
+    }
+  } else {
+    switch (xoffset) {
+      case 0: vp8_copy_mem8x4(src, src_stride, dst, dst_stride); break;
+      case 2:
+      case 4:
+      case 6:
+        common_hz_6t_8w_msa(src, src_stride, dst, dst_stride, h_filter, 4);
+        break;
 
-                case 1:
-                case 3:
-                case 5:
-                case 7:
-                    switch (yoffset)
-                    {
-                        case 2:
-                        case 4:
-                        case 6:
-                            common_hv_4ht_6vt_8w_msa(src, src_stride, dst,
-                                                     dst_stride, h_filter + 1,
-                                                     v_filter, 8);
-                            break;
+      case 1:
+      case 3:
+      case 5:
+      case 7:
+        common_hz_4t_8w_msa(src, src_stride, dst, dst_stride, h_filter + 1, 4);
+        break;
+    }
+  }
+}
+
+void vp8_sixtap_predict8x8_msa(uint8_t *RESTRICT src, int32_t src_stride,
+                               int32_t xoffset, int32_t yoffset,
+                               uint8_t *RESTRICT dst, int32_t dst_stride) {
+  const int8_t *h_filter = vp8_subpel_filters_msa[xoffset - 1];
+  const int8_t *v_filter = vp8_subpel_filters_msa[yoffset - 1];
+
+  if (yoffset) {
+    if (xoffset) {
+      switch (xoffset) {
+        case 2:
+        case 4:
+        case 6:
+          switch (yoffset) {
+            case 2:
+            case 4:
+            case 6:
+              common_hv_6ht_6vt_8w_msa(src, src_stride, dst, dst_stride,
+                                       h_filter, v_filter, 8);
+              break;
 
-                        case 1:
-                        case 3:
-                        case 5:
-                        case 7:
-                            common_hv_4ht_4vt_8w_msa(src, src_stride, dst,
-                                                     dst_stride, h_filter + 1,
-                                                     v_filter + 1, 8);
-                            break;
-                    }
-                    break;
-                }
-        }
-        else
-        {
-            switch (yoffset)
-            {
-                case 2:
-                case 4:
-                case 6:
-                    common_vt_6t_8w_msa(src, src_stride, dst, dst_stride,
-                                        v_filter, 8);
-                    break;
-
-                case 1:
-                case 3:
-                case 5:
-                case 7:
-                    common_vt_4t_8w_msa(src, src_stride, dst, dst_stride,
-                                        v_filter + 1, 8);
-                    break;
-            }
-        }
-    }
-    else
-    {
-        switch (xoffset)
-        {
-            case 0:
-                vp8_copy_mem8x8(src, src_stride, dst, dst_stride);
-                break;
+            case 1:
+            case 3:
+            case 5:
+            case 7:
+              common_hv_6ht_4vt_8w_msa(src, src_stride, dst, dst_stride,
+                                       h_filter, v_filter + 1, 8);
+              break;
+          }
+          break;
+
+        case 1:
+        case 3:
+        case 5:
+        case 7:
+          switch (yoffset) {
             case 2:
             case 4:
             case 6:
-                common_hz_6t_8w_msa(src, src_stride, dst, dst_stride, h_filter,
-                                    8);
-                break;
+              common_hv_4ht_6vt_8w_msa(src, src_stride, dst, dst_stride,
+                                       h_filter + 1, v_filter, 8);
+              break;
 
             case 1:
             case 3:
             case 5:
             case 7:
-                common_hz_4t_8w_msa(src, src_stride, dst, dst_stride,
-                                    h_filter + 1, 8);
-                break;
-        }
-    }
-}
-
-void vp8_sixtap_predict16x16_msa(uint8_t *RESTRICT src, int32_t src_stride,
-                                 int32_t xoffset, int32_t yoffset,
-                                 uint8_t *RESTRICT dst, int32_t dst_stride)
-{
-    const int8_t *h_filter = vp8_subpel_filters_msa[xoffset - 1];
-    const int8_t *v_filter = vp8_subpel_filters_msa[yoffset - 1];
-
-    if (yoffset)
-    {
-        if (xoffset)
-        {
-            switch (xoffset)
-            {
-                case 2:
-                case 4:
-                case 6:
-                    switch (yoffset)
-                    {
-                        case 2:
-                        case 4:
-                        case 6:
-                            common_hv_6ht_6vt_16w_msa(src, src_stride, dst,
-                                                      dst_stride, h_filter,
-                                                      v_filter, 16);
-                            break;
+              common_hv_4ht_4vt_8w_msa(src, src_stride, dst, dst_stride,
+                                       h_filter + 1, v_filter + 1, 8);
+              break;
+          }
+          break;
+      }
+    } else {
+      switch (yoffset) {
+        case 2:
+        case 4:
+        case 6:
+          common_vt_6t_8w_msa(src, src_stride, dst, dst_stride, v_filter, 8);
+          break;
 
-                        case 1:
-                        case 3:
-                        case 5:
-                        case 7:
-                            common_hv_6ht_4vt_16w_msa(src, src_stride, dst,
-                                                      dst_stride, h_filter,
-                                                      v_filter + 1, 16);
-                            break;
-                    }
-                    break;
+        case 1:
+        case 3:
+        case 5:
+        case 7:
+          common_vt_4t_8w_msa(src, src_stride, dst, dst_stride, v_filter + 1,
+                              8);
+          break;
+      }
+    }
+  } else {
+    switch (xoffset) {
+      case 0: vp8_copy_mem8x8(src, src_stride, dst, dst_stride); break;
+      case 2:
+      case 4:
+      case 6:
+        common_hz_6t_8w_msa(src, src_stride, dst, dst_stride, h_filter, 8);
+        break;
 
-                case 1:
-                case 3:
-                case 5:
-                case 7:
-                    switch (yoffset)
-                    {
-                        case 2:
-                        case 4:
-                        case 6:
-                            common_hv_4ht_6vt_16w_msa(src, src_stride, dst,
-                                                      dst_stride, h_filter + 1,
-                                                      v_filter, 16);
-                            break;
+      case 1:
+      case 3:
+      case 5:
+      case 7:
+        common_hz_4t_8w_msa(src, src_stride, dst, dst_stride, h_filter + 1, 8);
+        break;
+    }
+  }
+}
+
+void vp8_sixtap_predict16x16_msa(uint8_t *RESTRICT src, int32_t src_stride,
+                                 int32_t xoffset, int32_t yoffset,
+                                 uint8_t *RESTRICT dst, int32_t dst_stride) {
+  const int8_t *h_filter = vp8_subpel_filters_msa[xoffset - 1];
+  const int8_t *v_filter = vp8_subpel_filters_msa[yoffset - 1];
+
+  if (yoffset) {
+    if (xoffset) {
+      switch (xoffset) {
+        case 2:
+        case 4:
+        case 6:
+          switch (yoffset) {
+            case 2:
+            case 4:
+            case 6:
+              common_hv_6ht_6vt_16w_msa(src, src_stride, dst, dst_stride,
+                                        h_filter, v_filter, 16);
+              break;
 
-                        case 1:
-                        case 3:
-                        case 5:
-                        case 7:
-                            common_hv_4ht_4vt_16w_msa(src, src_stride, dst,
-                                                      dst_stride, h_filter + 1,
-                                                      v_filter + 1, 16);
-                            break;
-                    }
-                    break;
-            }
-        }
-        else
-        {
-            switch (yoffset)
-            {
-                case 2:
-                case 4:
-                case 6:
-                    common_vt_6t_16w_msa(src, src_stride, dst, dst_stride,
-                                         v_filter, 16);
-                    break;
-
-                case 1:
-                case 3:
-                case 5:
-                case 7:
-                    common_vt_4t_16w_msa(src, src_stride, dst, dst_stride,
-                                         v_filter + 1, 16);
-                    break;
-            }
-        }
-    }
-    else
-    {
-        switch (xoffset)
-        {
-            case 0:
-                vp8_copy_mem16x16(src, src_stride, dst, dst_stride);
-                break;
+            case 1:
+            case 3:
+            case 5:
+            case 7:
+              common_hv_6ht_4vt_16w_msa(src, src_stride, dst, dst_stride,
+                                        h_filter, v_filter + 1, 16);
+              break;
+          }
+          break;
+
+        case 1:
+        case 3:
+        case 5:
+        case 7:
+          switch (yoffset) {
             case 2:
             case 4:
             case 6:
-                common_hz_6t_16w_msa(src, src_stride, dst, dst_stride,
-                                     h_filter, 16);
-                break;
+              common_hv_4ht_6vt_16w_msa(src, src_stride, dst, dst_stride,
+                                        h_filter + 1, v_filter, 16);
+              break;
 
             case 1:
             case 3:
             case 5:
             case 7:
-                common_hz_4t_16w_msa(src, src_stride, dst, dst_stride,
-                                     h_filter + 1, 16);
-                break;
-        }
+              common_hv_4ht_4vt_16w_msa(src, src_stride, dst, dst_stride,
+                                        h_filter + 1, v_filter + 1, 16);
+              break;
+          }
+          break;
+      }
+    } else {
+      switch (yoffset) {
+        case 2:
+        case 4:
+        case 6:
+          common_vt_6t_16w_msa(src, src_stride, dst, dst_stride, v_filter, 16);
+          break;
+
+        case 1:
+        case 3:
+        case 5:
+        case 7:
+          common_vt_4t_16w_msa(src, src_stride, dst, dst_stride, v_filter + 1,
+                               16);
+          break;
+      }
     }
+  } else {
+    switch (xoffset) {
+      case 0: vp8_copy_mem16x16(src, src_stride, dst, dst_stride); break;
+      case 2:
+      case 4:
+      case 6:
+        common_hz_6t_16w_msa(src, src_stride, dst, dst_stride, h_filter, 16);
+        break;
+
+      case 1:
+      case 3:
+      case 5:
+      case 7:
+        common_hz_4t_16w_msa(src, src_stride, dst, dst_stride, h_filter + 1,
+                             16);
+        break;
+    }
+  }
 }
--- a/vp8/common/mips/msa/vp8_macros_msa.h
+++ b/vp8/common/mips/msa/vp8_macros_msa.h
@@ -40,177 +40,159 @@
 #define ST_SW(...) ST_W(v4i32, __VA_ARGS__)
 
 #if (__mips_isa_rev >= 6)
-#define LW(psrc)                                      \
-({                                                    \
-    const uint8_t *psrc_m = (const uint8_t *)(psrc);  \
-    uint32_t val_m;                                   \
-                                                      \
-    asm volatile (                                    \
-        "lw  %[val_m],  %[psrc_m]  \n\t"              \
-                                                      \
-        : [val_m] "=r" (val_m)                        \
-        : [psrc_m] "m" (*psrc_m)                      \
-    );                                                \
-                                                      \
-    val_m;                                            \
-})
+#define LW(psrc)                                     \
+  ({                                                 \
+    const uint8_t *psrc_m = (const uint8_t *)(psrc); \
+    uint32_t val_m;                                  \
+                                                     \
+    asm volatile("lw  %[val_m],  %[psrc_m]  \n\t"    \
+                                                     \
+                 : [val_m] "=r"(val_m)               \
+                 : [psrc_m] "m"(*psrc_m));           \
+                                                     \
+    val_m;                                           \
+  })
 
 #if (__mips == 64)
-#define LD(psrc)                                      \
-({                                                    \
-    const uint8_t *psrc_m = (const uint8_t *)(psrc);  \
-    uint64_t val_m = 0;                               \
-                                                      \
-    asm volatile (                                    \
-        "ld  %[val_m],  %[psrc_m]  \n\t"              \
-                                                      \
-        : [val_m] "=r" (val_m)                        \
-        : [psrc_m] "m" (*psrc_m)                      \
-    );                                                \
-                                                      \
-    val_m;                                            \
-})
+#define LD(psrc)                                     \
+  ({                                                 \
+    const uint8_t *psrc_m = (const uint8_t *)(psrc); \
+    uint64_t val_m = 0;                              \
+                                                     \
+    asm volatile("ld  %[val_m],  %[psrc_m]  \n\t"    \
+                                                     \
+                 : [val_m] "=r"(val_m)               \
+                 : [psrc_m] "m"(*psrc_m));           \
+                                                     \
+    val_m;                                           \
+  })
 #else  // !(__mips == 64)
-#define LD(psrc)                                             \
-({                                                           \
-    const uint8_t *psrc_m = (const uint8_t *)(psrc);         \
-    uint32_t val0_m, val1_m;                                 \
-    uint64_t val_m = 0;                                      \
-                                                             \
-    val0_m = LW(psrc_m);                                     \
-    val1_m = LW(psrc_m + 4);                                 \
-                                                             \
-    val_m = (uint64_t)(val1_m);                              \
-    val_m = (uint64_t)((val_m << 32) & 0xFFFFFFFF00000000);  \
-    val_m = (uint64_t)(val_m | (uint64_t)val0_m);            \
-                                                             \
-    val_m;                                                   \
-})
+#define LD(psrc)                                            \
+  ({                                                        \
+    const uint8_t *psrc_m = (const uint8_t *)(psrc);        \
+    uint32_t val0_m, val1_m;                                \
+    uint64_t val_m = 0;                                     \
+                                                            \
+    val0_m = LW(psrc_m);                                    \
+    val1_m = LW(psrc_m + 4);                                \
+                                                            \
+    val_m = (uint64_t)(val1_m);                             \
+    val_m = (uint64_t)((val_m << 32) & 0xFFFFFFFF00000000); \
+    val_m = (uint64_t)(val_m | (uint64_t)val0_m);           \
+                                                            \
+    val_m;                                                  \
+  })
 #endif  // (__mips == 64)
 
-#define SH(val, pdst)                     \
-{                                         \
-    uint8_t *pdst_m = (uint8_t *)(pdst);  \
-    const uint16_t val_m = (val);         \
-                                          \
-    asm volatile (                        \
-        "sh  %[val_m],  %[pdst_m]  \n\t"  \
-                                          \
-        : [pdst_m] "=m" (*pdst_m)         \
-        : [val_m] "r" (val_m)             \
-    );                                    \
-}
+#define SH(val, pdst)                             \
+  {                                               \
+    uint8_t *pdst_m = (uint8_t *)(pdst);          \
+    const uint16_t val_m = (val);                 \
+                                                  \
+    asm volatile("sh  %[val_m],  %[pdst_m]  \n\t" \
+                                                  \
+                 : [pdst_m] "=m"(*pdst_m)         \
+                 : [val_m] "r"(val_m));           \
+  }
 
-#define SW(val, pdst)                     \
-{                                         \
-    uint8_t *pdst_m = (uint8_t *)(pdst);  \
-    const uint32_t val_m = (val);         \
-                                          \
-    asm volatile (                        \
-        "sw  %[val_m],  %[pdst_m]  \n\t"  \
-                                          \
-        : [pdst_m] "=m" (*pdst_m)         \
-        : [val_m] "r" (val_m)             \
-    );                                    \
-}
+#define SW(val, pdst)                             \
+  {                                               \
+    uint8_t *pdst_m = (uint8_t *)(pdst);          \
+    const uint32_t val_m = (val);                 \
+                                                  \
+    asm volatile("sw  %[val_m],  %[pdst_m]  \n\t" \
+                                                  \
+                 : [pdst_m] "=m"(*pdst_m)         \
+                 : [val_m] "r"(val_m));           \
+  }
 
-#define SD(val, pdst)                     \
-{                                         \
-    uint8_t *pdst_m = (uint8_t *)(pdst);  \
-    const uint64_t val_m = (val);         \
-                                          \
-    asm volatile (                        \
-        "sd  %[val_m],  %[pdst_m]  \n\t"  \
-                                          \
-        : [pdst_m] "=m" (*pdst_m)         \
-        : [val_m] "r" (val_m)             \
-    );                                    \
-}
+#define SD(val, pdst)                             \
+  {                                               \
+    uint8_t *pdst_m = (uint8_t *)(pdst);          \
+    const uint64_t val_m = (val);                 \
+                                                  \
+    asm volatile("sd  %[val_m],  %[pdst_m]  \n\t" \
+                                                  \
+                 : [pdst_m] "=m"(*pdst_m)         \
+                 : [val_m] "r"(val_m));           \
+  }
 #else  // !(__mips_isa_rev >= 6)
-#define LW(psrc)                                      \
-({                                                    \
-    const uint8_t *psrc_m = (const uint8_t *)(psrc);  \
-    uint32_t val_m;                                   \
-                                                      \
-    asm volatile (                                    \
-        "ulw  %[val_m],  %[psrc_m]  \n\t"             \
-                                                      \
-        : [val_m] "=r" (val_m)                        \
-        : [psrc_m] "m" (*psrc_m)                      \
-    );                                                \
-                                                      \
-    val_m;                                            \
-})
+#define LW(psrc)                                     \
+  ({                                                 \
+    const uint8_t *psrc_m = (const uint8_t *)(psrc); \
+    uint32_t val_m;                                  \
+                                                     \
+    asm volatile("ulw  %[val_m],  %[psrc_m]  \n\t"   \
+                                                     \
+                 : [val_m] "=r"(val_m)               \
+                 : [psrc_m] "m"(*psrc_m));           \
+                                                     \
+    val_m;                                           \
+  })
 
 #if (__mips == 64)
-#define LD(psrc)                                      \
-({                                                    \
-    const uint8_t *psrc_m = (const uint8_t *)(psrc);  \
-    uint64_t val_m = 0;                               \
-                                                      \
-    asm volatile (                                    \
-        "uld  %[val_m],  %[psrc_m]  \n\t"             \
-                                                      \
-        : [val_m] "=r" (val_m)                        \
-        : [psrc_m] "m" (*psrc_m)                      \
-    );                                                \
-                                                      \
-    val_m;                                            \
-})
+#define LD(psrc)                                     \
+  ({                                                 \
+    const uint8_t *psrc_m = (const uint8_t *)(psrc); \
+    uint64_t val_m = 0;                              \
+                                                     \
+    asm volatile("uld  %[val_m],  %[psrc_m]  \n\t"   \
+                                                     \
+                 : [val_m] "=r"(val_m)               \
+                 : [psrc_m] "m"(*psrc_m));           \
+                                                     \
+    val_m;                                           \
+  })
 #else  // !(__mips == 64)
-#define LD(psrc)                                             \
-({                                                           \
-    const uint8_t *psrc_m1 = (const uint8_t *)(psrc);        \
-    uint32_t val0_m, val1_m;                                 \
-    uint64_t val_m = 0;                                      \
-                                                             \
-    val0_m = LW(psrc_m1);                                    \
-    val1_m = LW(psrc_m1 + 4);                                \
-                                                             \
-    val_m = (uint64_t)(val1_m);                              \
-    val_m = (uint64_t)((val_m << 32) & 0xFFFFFFFF00000000);  \
-    val_m = (uint64_t)(val_m | (uint64_t)val0_m);            \
-                                                             \
-    val_m;                                                   \
-})
+#define LD(psrc)                                            \
+  ({                                                        \
+    const uint8_t *psrc_m1 = (const uint8_t *)(psrc);       \
+    uint32_t val0_m, val1_m;                                \
+    uint64_t val_m = 0;                                     \
+                                                            \
+    val0_m = LW(psrc_m1);                                   \
+    val1_m = LW(psrc_m1 + 4);                               \
+                                                            \
+    val_m = (uint64_t)(val1_m);                             \
+    val_m = (uint64_t)((val_m << 32) & 0xFFFFFFFF00000000); \
+    val_m = (uint64_t)(val_m | (uint64_t)val0_m);           \
+                                                            \
+    val_m;                                                  \
+  })
 #endif  // (__mips == 64)
-#define SH(val, pdst)                      \
-{                                          \
-    uint8_t *pdst_m = (uint8_t *)(pdst);   \
-    const uint16_t val_m = (val);          \
-                                           \
-    asm volatile (                         \
-        "ush  %[val_m],  %[pdst_m]  \n\t"  \
-                                           \
-        : [pdst_m] "=m" (*pdst_m)          \
-        : [val_m] "r" (val_m)              \
-    );                                     \
-}
+#define SH(val, pdst)                              \
+  {                                                \
+    uint8_t *pdst_m = (uint8_t *)(pdst);           \
+    const uint16_t val_m = (val);                  \
+                                                   \
+    asm volatile("ush  %[val_m],  %[pdst_m]  \n\t" \
+                                                   \
+                 : [pdst_m] "=m"(*pdst_m)          \
+                 : [val_m] "r"(val_m));            \
+  }
 
-#define SW(val, pdst)                      \
-{                                          \
-    uint8_t *pdst_m = (uint8_t *)(pdst);   \
-    const uint32_t val_m = (val);          \
-                                           \
-    asm volatile (                         \
-        "usw  %[val_m],  %[pdst_m]  \n\t"  \
-                                           \
-        : [pdst_m] "=m" (*pdst_m)          \
-        : [val_m] "r" (val_m)              \
-    );                                     \
-}
+#define SW(val, pdst)                              \
+  {                                                \
+    uint8_t *pdst_m = (uint8_t *)(pdst);           \
+    const uint32_t val_m = (val);                  \
+                                                   \
+    asm volatile("usw  %[val_m],  %[pdst_m]  \n\t" \
+                                                   \
+                 : [pdst_m] "=m"(*pdst_m)          \
+                 : [val_m] "r"(val_m));            \
+  }
 
-#define SD(val, pdst)                                         \
-{                                                             \
-    uint8_t *pdst_m1 = (uint8_t *)(pdst);                     \
-    uint32_t val0_m, val1_m;                                  \
-                                                              \
-    val0_m = (uint32_t)((val) & 0x00000000FFFFFFFF);          \
-    val1_m = (uint32_t)(((val) >> 32) & 0x00000000FFFFFFFF);  \
-                                                              \
-    SW(val0_m, pdst_m1);                                      \
-    SW(val1_m, pdst_m1 + 4);                                  \
-}
+#define SD(val, pdst)                                        \
+  {                                                          \
+    uint8_t *pdst_m1 = (uint8_t *)(pdst);                    \
+    uint32_t val0_m, val1_m;                                 \
+                                                             \
+    val0_m = (uint32_t)((val)&0x00000000FFFFFFFF);           \
+    val1_m = (uint32_t)(((val) >> 32) & 0x00000000FFFFFFFF); \
+                                                             \
+    SW(val0_m, pdst_m1);                                     \
+    SW(val1_m, pdst_m1 + 4);                                 \
+  }
 #endif  // (__mips_isa_rev >= 6)
 
 /* Description : Load 4 words with stride
@@ -221,13 +203,13 @@
                  Load word in 'out2' from (psrc + 2 * stride)
                  Load word in 'out3' from (psrc + 3 * stride)
 */
-#define LW4(psrc, stride, out0, out1, out2, out3)  \
-{                                                  \
-    out0 = LW((psrc));                             \
-    out1 = LW((psrc) + stride);                    \
-    out2 = LW((psrc) + 2 * stride);                \
-    out3 = LW((psrc) + 3 * stride);                \
-}
+#define LW4(psrc, stride, out0, out1, out2, out3) \
+  {                                               \
+    out0 = LW((psrc));                            \
+    out1 = LW((psrc) + stride);                   \
+    out2 = LW((psrc) + 2 * stride);               \
+    out3 = LW((psrc) + 3 * stride);               \
+  }
 
 /* Description : Load double words with stride
    Arguments   : Inputs  - psrc, stride
@@ -235,16 +217,16 @@
    Details     : Load double word in 'out0' from (psrc)
                  Load double word in 'out1' from (psrc + stride)
 */
-#define LD2(psrc, stride, out0, out1)  \
-{                                      \
-    out0 = LD((psrc));                 \
-    out1 = LD((psrc) + stride);        \
-}
-#define LD4(psrc, stride, out0, out1, out2, out3)  \
-{                                                  \
-    LD2((psrc), stride, out0, out1);               \
-    LD2((psrc) + 2 * stride, stride, out2, out3);  \
-}
+#define LD2(psrc, stride, out0, out1) \
+  {                                   \
+    out0 = LD((psrc));                \
+    out1 = LD((psrc) + stride);       \
+  }
+#define LD4(psrc, stride, out0, out1, out2, out3) \
+  {                                               \
+    LD2((psrc), stride, out0, out1);              \
+    LD2((psrc) + 2 * stride, stride, out2, out3); \
+  }
 
 /* Description : Store 4 words with stride
    Arguments   : Inputs - in0, in1, in2, in3, pdst, stride
@@ -253,13 +235,13 @@
                  Store word from 'in2' to (pdst + 2 * stride)
                  Store word from 'in3' to (pdst + 3 * stride)
 */
-#define SW4(in0, in1, in2, in3, pdst, stride)  \
-{                                              \
-    SW(in0, (pdst));                           \
-    SW(in1, (pdst) + stride);                  \
-    SW(in2, (pdst) + 2 * stride);              \
-    SW(in3, (pdst) + 3 * stride);              \
-}
+#define SW4(in0, in1, in2, in3, pdst, stride) \
+  {                                           \
+    SW(in0, (pdst));                          \
+    SW(in1, (pdst) + stride);                 \
+    SW(in2, (pdst) + 2 * stride);             \
+    SW(in3, (pdst) + 3 * stride);             \
+  }
 
 /* Description : Store 4 double words with stride
    Arguments   : Inputs - in0, in1, in2, in3, pdst, stride
@@ -268,13 +250,13 @@
                  Store double word from 'in2' to (pdst + 2 * stride)
                  Store double word from 'in3' to (pdst + 3 * stride)
 */
-#define SD4(in0, in1, in2, in3, pdst, stride)  \
-{                                              \
-    SD(in0, (pdst));                           \
-    SD(in1, (pdst) + stride);                  \
-    SD(in2, (pdst) + 2 * stride);              \
-    SD(in3, (pdst) + 3 * stride);              \
-}
+#define SD4(in0, in1, in2, in3, pdst, stride) \
+  {                                           \
+    SD(in0, (pdst));                          \
+    SD(in1, (pdst) + stride);                 \
+    SD(in2, (pdst) + 2 * stride);             \
+    SD(in3, (pdst) + 3 * stride);             \
+  }
 
 /* Description : Load vectors with 16 byte elements with stride
    Arguments   : Inputs  - psrc, stride
@@ -283,44 +265,44 @@
    Details     : Load 16 byte elements in 'out0' from (psrc)
                  Load 16 byte elements in 'out1' from (psrc + stride)
 */
-#define LD_B2(RTYPE, psrc, stride, out0, out1)  \
-{                                               \
-    out0 = LD_B(RTYPE, (psrc));                 \
-    out1 = LD_B(RTYPE, (psrc) + stride);        \
-}
+#define LD_B2(RTYPE, psrc, stride, out0, out1) \
+  {                                            \
+    out0 = LD_B(RTYPE, (psrc));                \
+    out1 = LD_B(RTYPE, (psrc) + stride);       \
+  }
 #define LD_UB2(...) LD_B2(v16u8, __VA_ARGS__)
 #define LD_SB2(...) LD_B2(v16i8, __VA_ARGS__)
 
-#define LD_B3(RTYPE, psrc, stride, out0, out1, out2)  \
-{                                                     \
-    LD_B2(RTYPE, (psrc), stride, out0, out1);         \
-    out2 = LD_B(RTYPE, (psrc) + 2 * stride);          \
-}
+#define LD_B3(RTYPE, psrc, stride, out0, out1, out2) \
+  {                                                  \
+    LD_B2(RTYPE, (psrc), stride, out0, out1);        \
+    out2 = LD_B(RTYPE, (psrc) + 2 * stride);         \
+  }
 #define LD_UB3(...) LD_B3(v16u8, __VA_ARGS__)
 #define LD_SB3(...) LD_B3(v16i8, __VA_ARGS__)
 
-#define LD_B4(RTYPE, psrc, stride, out0, out1, out2, out3)   \
-{                                                            \
-    LD_B2(RTYPE, (psrc), stride, out0, out1);                \
-    LD_B2(RTYPE, (psrc) + 2 * stride , stride, out2, out3);  \
-}
+#define LD_B4(RTYPE, psrc, stride, out0, out1, out2, out3) \
+  {                                                        \
+    LD_B2(RTYPE, (psrc), stride, out0, out1);              \
+    LD_B2(RTYPE, (psrc) + 2 * stride, stride, out2, out3); \
+  }
 #define LD_UB4(...) LD_B4(v16u8, __VA_ARGS__)
 #define LD_SB4(...) LD_B4(v16i8, __VA_ARGS__)
 
-#define LD_B5(RTYPE, psrc, stride, out0, out1, out2, out3, out4)  \
-{                                                                 \
-    LD_B4(RTYPE, (psrc), stride, out0, out1, out2, out3);         \
-    out4 = LD_B(RTYPE, (psrc) + 4 * stride);                      \
-}
+#define LD_B5(RTYPE, psrc, stride, out0, out1, out2, out3, out4) \
+  {                                                              \
+    LD_B4(RTYPE, (psrc), stride, out0, out1, out2, out3);        \
+    out4 = LD_B(RTYPE, (psrc) + 4 * stride);                     \
+  }
 #define LD_UB5(...) LD_B5(v16u8, __VA_ARGS__)
 #define LD_SB5(...) LD_B5(v16i8, __VA_ARGS__)
 
-#define LD_B8(RTYPE, psrc, stride,                                      \
-              out0, out1, out2, out3, out4, out5, out6, out7)           \
-{                                                                       \
-    LD_B4(RTYPE, (psrc), stride, out0, out1, out2, out3);               \
-    LD_B4(RTYPE, (psrc) + 4 * stride, stride, out4, out5, out6, out7);  \
-}
+#define LD_B8(RTYPE, psrc, stride, out0, out1, out2, out3, out4, out5, out6, \
+              out7)                                                          \
+  {                                                                          \
+    LD_B4(RTYPE, (psrc), stride, out0, out1, out2, out3);                    \
+    LD_B4(RTYPE, (psrc) + 4 * stride, stride, out4, out5, out6, out7);       \
+  }
 #define LD_UB8(...) LD_B8(v16u8, __VA_ARGS__)
 #define LD_SB8(...) LD_B8(v16i8, __VA_ARGS__)
 
@@ -330,18 +312,18 @@
    Details     : Load 8 halfword elements in 'out0' from (psrc)
                  Load 8 halfword elements in 'out1' from (psrc + stride)
 */
-#define LD_H2(RTYPE, psrc, stride, out0, out1)  \
-{                                               \
-    out0 = LD_H(RTYPE, (psrc));                 \
-    out1 = LD_H(RTYPE, (psrc) + (stride));      \
-}
+#define LD_H2(RTYPE, psrc, stride, out0, out1) \
+  {                                            \
+    out0 = LD_H(RTYPE, (psrc));                \
+    out1 = LD_H(RTYPE, (psrc) + (stride));     \
+  }
 #define LD_SH2(...) LD_H2(v8i16, __VA_ARGS__)
 
-#define LD_H4(RTYPE, psrc, stride, out0, out1, out2, out3)  \
-{                                                           \
-    LD_H2(RTYPE, (psrc), stride, out0, out1);               \
-    LD_H2(RTYPE, (psrc) + 2 * stride, stride, out2, out3);  \
-}
+#define LD_H4(RTYPE, psrc, stride, out0, out1, out2, out3) \
+  {                                                        \
+    LD_H2(RTYPE, (psrc), stride, out0, out1);              \
+    LD_H2(RTYPE, (psrc) + 2 * stride, stride, out2, out3); \
+  }
 #define LD_SH4(...) LD_H4(v8i16, __VA_ARGS__)
 
 /* Description : Load 2 vectors of signed word elements with stride
@@ -349,11 +331,11 @@
                  Outputs - out0, out1
                  Return Type - signed word
 */
-#define LD_SW2(psrc, stride, out0, out1)  \
-{                                         \
-    out0 = LD_SW((psrc));                 \
-    out1 = LD_SW((psrc) + stride);        \
-}
+#define LD_SW2(psrc, stride, out0, out1) \
+  {                                      \
+    out0 = LD_SW((psrc));                \
+    out1 = LD_SW((psrc) + stride);       \
+  }
 
 /* Description : Store vectors of 16 byte elements with stride
    Arguments   : Inputs - in0, in1, pdst, stride
@@ -360,27 +342,26 @@
    Details     : Store 16 byte elements from 'in0' to (pdst)
                  Store 16 byte elements from 'in1' to (pdst + stride)
 */
-#define ST_B2(RTYPE, in0, in1, pdst, stride)  \
-{                                             \
-    ST_B(RTYPE, in0, (pdst));                 \
-    ST_B(RTYPE, in1, (pdst) + stride);        \
-}
+#define ST_B2(RTYPE, in0, in1, pdst, stride) \
+  {                                          \
+    ST_B(RTYPE, in0, (pdst));                \
+    ST_B(RTYPE, in1, (pdst) + stride);       \
+  }
 #define ST_UB2(...) ST_B2(v16u8, __VA_ARGS__)
 
-#define ST_B4(RTYPE, in0, in1, in2, in3, pdst, stride)    \
-{                                                         \
-    ST_B2(RTYPE, in0, in1, (pdst), stride);               \
-    ST_B2(RTYPE, in2, in3, (pdst) + 2 * stride, stride);  \
-}
+#define ST_B4(RTYPE, in0, in1, in2, in3, pdst, stride)   \
+  {                                                      \
+    ST_B2(RTYPE, in0, in1, (pdst), stride);              \
+    ST_B2(RTYPE, in2, in3, (pdst) + 2 * stride, stride); \
+  }
 #define ST_UB4(...) ST_B4(v16u8, __VA_ARGS__)
 #define ST_SB4(...) ST_B4(v16i8, __VA_ARGS__)
 
-#define ST_B8(RTYPE, in0, in1, in2, in3, in4, in5, in6, in7,        \
-              pdst, stride)                                         \
-{                                                                   \
-    ST_B4(RTYPE, in0, in1, in2, in3, pdst, stride);                 \
-    ST_B4(RTYPE, in4, in5, in6, in7, (pdst) + 4 * stride, stride);  \
-}
+#define ST_B8(RTYPE, in0, in1, in2, in3, in4, in5, in6, in7, pdst, stride) \
+  {                                                                        \
+    ST_B4(RTYPE, in0, in1, in2, in3, pdst, stride);                        \
+    ST_B4(RTYPE, in4, in5, in6, in7, (pdst) + 4 * stride, stride);         \
+  }
 #define ST_UB8(...) ST_B8(v16u8, __VA_ARGS__)
 
 /* Description : Store vectors of 8 halfword elements with stride
@@ -388,11 +369,11 @@
    Details     : Store 8 halfword elements from 'in0' to (pdst)
                  Store 8 halfword elements from 'in1' to (pdst + stride)
 */
-#define ST_H2(RTYPE, in0, in1, pdst, stride)  \
-{                                             \
-    ST_H(RTYPE, in0, (pdst));                 \
-    ST_H(RTYPE, in1, (pdst) + stride);        \
-}
+#define ST_H2(RTYPE, in0, in1, pdst, stride) \
+  {                                          \
+    ST_H(RTYPE, in0, (pdst));                \
+    ST_H(RTYPE, in1, (pdst) + stride);       \
+  }
 #define ST_SH2(...) ST_H2(v8i16, __VA_ARGS__)
 
 /* Description : Store vectors of word elements with stride
@@ -400,11 +381,11 @@
    Details     : Store 4 word elements from 'in0' to (pdst)
                  Store 4 word elements from 'in1' to (pdst + stride)
 */
-#define ST_SW2(in0, in1, pdst, stride)  \
-{                                       \
-    ST_SW(in0, (pdst));                 \
-    ST_SW(in1, (pdst) + stride);        \
-}
+#define ST_SW2(in0, in1, pdst, stride) \
+  {                                    \
+    ST_SW(in0, (pdst));                \
+    ST_SW(in1, (pdst) + stride);       \
+  }
 
 /* Description : Store 2x4 byte block to destination memory from input vector
    Arguments   : Inputs - in, stidx, pdst, stride
@@ -417,21 +398,21 @@
                  Index 'stidx+3' halfword element from 'in' vector is copied to
                  the GP register and stored to (pdst + 3 * stride)
 */
-#define ST2x4_UB(in, stidx, pdst, stride)             \
-{                                                     \
-    uint16_t out0_m, out1_m, out2_m, out3_m;          \
-    uint8_t *pblk_2x4_m = (uint8_t *)(pdst);          \
-                                                      \
-    out0_m = __msa_copy_u_h((v8i16)in, (stidx));      \
-    out1_m = __msa_copy_u_h((v8i16)in, (stidx + 1));  \
-    out2_m = __msa_copy_u_h((v8i16)in, (stidx + 2));  \
-    out3_m = __msa_copy_u_h((v8i16)in, (stidx + 3));  \
-                                                      \
-    SH(out0_m, pblk_2x4_m);                           \
-    SH(out1_m, pblk_2x4_m + stride);                  \
-    SH(out2_m, pblk_2x4_m + 2 * stride);              \
-    SH(out3_m, pblk_2x4_m + 3 * stride);              \
-}
+#define ST2x4_UB(in, stidx, pdst, stride)            \
+  {                                                  \
+    uint16_t out0_m, out1_m, out2_m, out3_m;         \
+    uint8_t *pblk_2x4_m = (uint8_t *)(pdst);         \
+                                                     \
+    out0_m = __msa_copy_u_h((v8i16)in, (stidx));     \
+    out1_m = __msa_copy_u_h((v8i16)in, (stidx + 1)); \
+    out2_m = __msa_copy_u_h((v8i16)in, (stidx + 2)); \
+    out3_m = __msa_copy_u_h((v8i16)in, (stidx + 3)); \
+                                                     \
+    SH(out0_m, pblk_2x4_m);                          \
+    SH(out1_m, pblk_2x4_m + stride);                 \
+    SH(out2_m, pblk_2x4_m + 2 * stride);             \
+    SH(out3_m, pblk_2x4_m + 3 * stride);             \
+  }
 
 /* Description : Store 4x4 byte block to destination memory from input vector
    Arguments   : Inputs - in0, in1, pdst, stride
@@ -444,25 +425,25 @@
                  'Idx3' word element from input vector 'in0' is copied to the
                  GP register and stored to (pdst + 3 * stride)
 */
-#define ST4x4_UB(in0, in1, idx0, idx1, idx2, idx3, pdst, stride)  \
-{                                                                 \
-    uint32_t out0_m, out1_m, out2_m, out3_m;                      \
-    uint8_t *pblk_4x4_m = (uint8_t *)(pdst);                      \
-                                                                  \
-    out0_m = __msa_copy_u_w((v4i32)in0, idx0);                    \
-    out1_m = __msa_copy_u_w((v4i32)in0, idx1);                    \
-    out2_m = __msa_copy_u_w((v4i32)in1, idx2);                    \
-    out3_m = __msa_copy_u_w((v4i32)in1, idx3);                    \
-                                                                  \
-    SW4(out0_m, out1_m, out2_m, out3_m, pblk_4x4_m, stride);      \
-}
-#define ST4x8_UB(in0, in1, pdst, stride)                            \
-{                                                                   \
-    uint8_t *pblk_4x8 = (uint8_t *)(pdst);                          \
-                                                                    \
-    ST4x4_UB(in0, in0, 0, 1, 2, 3, pblk_4x8, stride);               \
-    ST4x4_UB(in1, in1, 0, 1, 2, 3, pblk_4x8 + 4 * stride, stride);  \
-}
+#define ST4x4_UB(in0, in1, idx0, idx1, idx2, idx3, pdst, stride) \
+  {                                                              \
+    uint32_t out0_m, out1_m, out2_m, out3_m;                     \
+    uint8_t *pblk_4x4_m = (uint8_t *)(pdst);                     \
+                                                                 \
+    out0_m = __msa_copy_u_w((v4i32)in0, idx0);                   \
+    out1_m = __msa_copy_u_w((v4i32)in0, idx1);                   \
+    out2_m = __msa_copy_u_w((v4i32)in1, idx2);                   \
+    out3_m = __msa_copy_u_w((v4i32)in1, idx3);                   \
+                                                                 \
+    SW4(out0_m, out1_m, out2_m, out3_m, pblk_4x4_m, stride);     \
+  }
+#define ST4x8_UB(in0, in1, pdst, stride)                           \
+  {                                                                \
+    uint8_t *pblk_4x8 = (uint8_t *)(pdst);                         \
+                                                                   \
+    ST4x4_UB(in0, in0, 0, 1, 2, 3, pblk_4x8, stride);              \
+    ST4x4_UB(in1, in1, 0, 1, 2, 3, pblk_4x8 + 4 * stride, stride); \
+  }
 
 /* Description : Store 8x1 byte block to destination memory from input vector
    Arguments   : Inputs - in, pdst
@@ -469,13 +450,13 @@
    Details     : Index 0 double word element from 'in' vector is copied to the
                  GP register and stored to (pdst)
 */
-#define ST8x1_UB(in, pdst)                  \
-{                                           \
-    uint64_t out0_m;                        \
-                                            \
-    out0_m = __msa_copy_u_d((v2i64)in, 0);  \
-    SD(out0_m, pdst);                       \
-}
+#define ST8x1_UB(in, pdst)                 \
+  {                                        \
+    uint64_t out0_m;                       \
+                                           \
+    out0_m = __msa_copy_u_d((v2i64)in, 0); \
+    SD(out0_m, pdst);                      \
+  }
 
 /* Description : Store 8x2 byte block to destination memory from input vector
    Arguments   : Inputs - in, pdst, stride
@@ -484,17 +465,17 @@
                  Index 1 double word element from 'in' vector is copied to the
                  GP register and stored to (pdst + stride)
 */
-#define ST8x2_UB(in, pdst, stride)            \
-{                                             \
-    uint64_t out0_m, out1_m;                  \
-    uint8_t *pblk_8x2_m = (uint8_t *)(pdst);  \
-                                              \
-    out0_m = __msa_copy_u_d((v2i64)in, 0);    \
-    out1_m = __msa_copy_u_d((v2i64)in, 1);    \
-                                              \
-    SD(out0_m, pblk_8x2_m);                   \
-    SD(out1_m, pblk_8x2_m + stride);          \
-}
+#define ST8x2_UB(in, pdst, stride)           \
+  {                                          \
+    uint64_t out0_m, out1_m;                 \
+    uint8_t *pblk_8x2_m = (uint8_t *)(pdst); \
+                                             \
+    out0_m = __msa_copy_u_d((v2i64)in, 0);   \
+    out1_m = __msa_copy_u_d((v2i64)in, 1);   \
+                                             \
+    SD(out0_m, pblk_8x2_m);                  \
+    SD(out1_m, pblk_8x2_m + stride);         \
+  }
 
 /* Description : Store 8x4 byte block to destination memory from input
                  vectors
@@ -508,18 +489,18 @@
                  Index 1 double word element from 'in1' vector is copied to the
                  GP register and stored to (pdst + 3 * stride)
 */
-#define ST8x4_UB(in0, in1, pdst, stride)                      \
-{                                                             \
-    uint64_t out0_m, out1_m, out2_m, out3_m;                  \
-    uint8_t *pblk_8x4_m = (uint8_t *)(pdst);                  \
-                                                              \
-    out0_m = __msa_copy_u_d((v2i64)in0, 0);                   \
-    out1_m = __msa_copy_u_d((v2i64)in0, 1);                   \
-    out2_m = __msa_copy_u_d((v2i64)in1, 0);                   \
-    out3_m = __msa_copy_u_d((v2i64)in1, 1);                   \
-                                                              \
-    SD4(out0_m, out1_m, out2_m, out3_m, pblk_8x4_m, stride);  \
-}
+#define ST8x4_UB(in0, in1, pdst, stride)                     \
+  {                                                          \
+    uint64_t out0_m, out1_m, out2_m, out3_m;                 \
+    uint8_t *pblk_8x4_m = (uint8_t *)(pdst);                 \
+                                                             \
+    out0_m = __msa_copy_u_d((v2i64)in0, 0);                  \
+    out1_m = __msa_copy_u_d((v2i64)in0, 1);                  \
+    out2_m = __msa_copy_u_d((v2i64)in1, 0);                  \
+    out3_m = __msa_copy_u_d((v2i64)in1, 1);                  \
+                                                             \
+    SD4(out0_m, out1_m, out2_m, out3_m, pblk_8x4_m, stride); \
+  }
 
 /* Description : Immediate number of elements to slide with zero
    Arguments   : Inputs  - in0, in1, slide_val
@@ -528,13 +509,13 @@
    Details     : Byte elements from 'zero_m' vector are slid into 'in0' by
                  value specified in the 'slide_val'
 */
-#define SLDI_B2_0(RTYPE, in0, in1, out0, out1, slide_val)              \
-{                                                                      \
-    v16i8 zero_m = { 0 };                                              \
-                                                                       \
-    out0 = (RTYPE)__msa_sldi_b((v16i8)zero_m, (v16i8)in0, slide_val);  \
-    out1 = (RTYPE)__msa_sldi_b((v16i8)zero_m, (v16i8)in1, slide_val);  \
-}
+#define SLDI_B2_0(RTYPE, in0, in1, out0, out1, slide_val)             \
+  {                                                                   \
+    v16i8 zero_m = { 0 };                                             \
+                                                                      \
+    out0 = (RTYPE)__msa_sldi_b((v16i8)zero_m, (v16i8)in0, slide_val); \
+    out1 = (RTYPE)__msa_sldi_b((v16i8)zero_m, (v16i8)in1, slide_val); \
+  }
 #define SLDI_B2_0_UB(...) SLDI_B2_0(v16u8, __VA_ARGS__)
 
 /* Description : Immediate number of elements to slide
@@ -544,18 +525,18 @@
    Details     : Byte elements from 'in0_0' vector are slid into 'in1_0' by
                  value specified in the 'slide_val'
 */
-#define SLDI_B2(RTYPE, in0_0, in0_1, in1_0, in1_1, out0, out1, slide_val)  \
-{                                                                          \
-    out0 = (RTYPE)__msa_sldi_b((v16i8)in0_0, (v16i8)in1_0, slide_val);     \
-    out1 = (RTYPE)__msa_sldi_b((v16i8)in0_1, (v16i8)in1_1, slide_val);     \
-}
+#define SLDI_B2(RTYPE, in0_0, in0_1, in1_0, in1_1, out0, out1, slide_val) \
+  {                                                                       \
+    out0 = (RTYPE)__msa_sldi_b((v16i8)in0_0, (v16i8)in1_0, slide_val);    \
+    out1 = (RTYPE)__msa_sldi_b((v16i8)in0_1, (v16i8)in1_1, slide_val);    \
+  }
 
-#define SLDI_B3(RTYPE, in0_0, in0_1, in0_2, in1_0, in1_1, in1_2,        \
-                out0, out1, out2, slide_val)                            \
-{                                                                       \
-    SLDI_B2(RTYPE, in0_0, in0_1, in1_0, in1_1, out0, out1, slide_val);  \
-    out2 = (RTYPE)__msa_sldi_b((v16i8)in0_2, (v16i8)in1_2, slide_val);  \
-}
+#define SLDI_B3(RTYPE, in0_0, in0_1, in0_2, in1_0, in1_1, in1_2, out0, out1, \
+                out2, slide_val)                                             \
+  {                                                                          \
+    SLDI_B2(RTYPE, in0_0, in0_1, in1_0, in1_1, out0, out1, slide_val);       \
+    out2 = (RTYPE)__msa_sldi_b((v16i8)in0_2, (v16i8)in1_2, slide_val);       \
+  }
 #define SLDI_B3_UH(...) SLDI_B3(v8u16, __VA_ARGS__)
 
 /* Description : Shuffle byte vector elements as per mask vector
@@ -565,21 +546,21 @@
    Details     : Byte elements from 'in0' & 'in1' are copied selectively to
                  'out0' as per control vector 'mask0'
 */
-#define VSHF_B2(RTYPE, in0, in1, in2, in3, mask0, mask1, out0, out1)   \
-{                                                                      \
-    out0 = (RTYPE)__msa_vshf_b((v16i8)mask0, (v16i8)in1, (v16i8)in0);  \
-    out1 = (RTYPE)__msa_vshf_b((v16i8)mask1, (v16i8)in3, (v16i8)in2);  \
-}
+#define VSHF_B2(RTYPE, in0, in1, in2, in3, mask0, mask1, out0, out1)  \
+  {                                                                   \
+    out0 = (RTYPE)__msa_vshf_b((v16i8)mask0, (v16i8)in1, (v16i8)in0); \
+    out1 = (RTYPE)__msa_vshf_b((v16i8)mask1, (v16i8)in3, (v16i8)in2); \
+  }
 #define VSHF_B2_UB(...) VSHF_B2(v16u8, __VA_ARGS__)
 #define VSHF_B2_SB(...) VSHF_B2(v16i8, __VA_ARGS__)
 #define VSHF_B2_UH(...) VSHF_B2(v8u16, __VA_ARGS__)
 
-#define VSHF_B3(RTYPE, in0, in1, in2, in3, in4, in5, mask0, mask1, mask2,  \
-                out0, out1, out2)                                          \
-{                                                                          \
-    VSHF_B2(RTYPE, in0, in1, in2, in3, mask0, mask1, out0, out1);          \
-    out2 = (RTYPE)__msa_vshf_b((v16i8)mask2, (v16i8)in5, (v16i8)in4);      \
-}
+#define VSHF_B3(RTYPE, in0, in1, in2, in3, in4, in5, mask0, mask1, mask2, \
+                out0, out1, out2)                                         \
+  {                                                                       \
+    VSHF_B2(RTYPE, in0, in1, in2, in3, mask0, mask1, out0, out1);         \
+    out2 = (RTYPE)__msa_vshf_b((v16i8)mask2, (v16i8)in5, (v16i8)in4);     \
+  }
 #define VSHF_B3_SB(...) VSHF_B3(v16i8, __VA_ARGS__)
 
 /* Description : Shuffle halfword vector elements as per mask vector
@@ -589,11 +570,11 @@
    Details     : halfword elements from 'in0' & 'in1' are copied selectively to
                  'out0' as per control vector 'mask0'
 */
-#define VSHF_H2(RTYPE, in0, in1, in2, in3, mask0, mask1, out0, out1)   \
-{                                                                      \
-    out0 = (RTYPE)__msa_vshf_h((v8i16)mask0, (v8i16)in1, (v8i16)in0);  \
-    out1 = (RTYPE)__msa_vshf_h((v8i16)mask1, (v8i16)in3, (v8i16)in2);  \
-}
+#define VSHF_H2(RTYPE, in0, in1, in2, in3, mask0, mask1, out0, out1)  \
+  {                                                                   \
+    out0 = (RTYPE)__msa_vshf_h((v8i16)mask0, (v8i16)in1, (v8i16)in0); \
+    out1 = (RTYPE)__msa_vshf_h((v8i16)mask1, (v8i16)in3, (v8i16)in2); \
+  }
 #define VSHF_H2_SH(...) VSHF_H2(v8i16, __VA_ARGS__)
 
 /* Description : Dot product of byte vector elements
@@ -606,20 +587,19 @@
                  The multiplication result of adjacent odd-even elements
                  are added together and written to the 'out0' vector
 */
-#define DOTP_UB2(RTYPE, mult0, mult1, cnst0, cnst1, out0, out1)  \
-{                                                                \
-    out0 = (RTYPE)__msa_dotp_u_h((v16u8)mult0, (v16u8)cnst0);    \
-    out1 = (RTYPE)__msa_dotp_u_h((v16u8)mult1, (v16u8)cnst1);    \
-}
+#define DOTP_UB2(RTYPE, mult0, mult1, cnst0, cnst1, out0, out1) \
+  {                                                             \
+    out0 = (RTYPE)__msa_dotp_u_h((v16u8)mult0, (v16u8)cnst0);   \
+    out1 = (RTYPE)__msa_dotp_u_h((v16u8)mult1, (v16u8)cnst1);   \
+  }
 #define DOTP_UB2_UH(...) DOTP_UB2(v8u16, __VA_ARGS__)
 
-#define DOTP_UB4(RTYPE, mult0, mult1, mult2, mult3,           \
-                 cnst0, cnst1, cnst2, cnst3,                  \
-                 out0, out1, out2, out3)                      \
-{                                                             \
-    DOTP_UB2(RTYPE, mult0, mult1, cnst0, cnst1, out0, out1);  \
-    DOTP_UB2(RTYPE, mult2, mult3, cnst2, cnst3, out2, out3);  \
-}
+#define DOTP_UB4(RTYPE, mult0, mult1, mult2, mult3, cnst0, cnst1, cnst2, \
+                 cnst3, out0, out1, out2, out3)                          \
+  {                                                                      \
+    DOTP_UB2(RTYPE, mult0, mult1, cnst0, cnst1, out0, out1);             \
+    DOTP_UB2(RTYPE, mult2, mult3, cnst2, cnst3, out2, out3);             \
+  }
 #define DOTP_UB4_UH(...) DOTP_UB4(v8u16, __VA_ARGS__)
 
 /* Description : Dot product of byte vector elements
@@ -632,19 +612,19 @@
                  The multiplication result of adjacent odd-even elements
                  are added together and written to the 'out0' vector
 */
-#define DOTP_SB2(RTYPE, mult0, mult1, cnst0, cnst1, out0, out1)  \
-{                                                                \
-    out0 = (RTYPE)__msa_dotp_s_h((v16i8)mult0, (v16i8)cnst0);    \
-    out1 = (RTYPE)__msa_dotp_s_h((v16i8)mult1, (v16i8)cnst1);    \
-}
+#define DOTP_SB2(RTYPE, mult0, mult1, cnst0, cnst1, out0, out1) \
+  {                                                             \
+    out0 = (RTYPE)__msa_dotp_s_h((v16i8)mult0, (v16i8)cnst0);   \
+    out1 = (RTYPE)__msa_dotp_s_h((v16i8)mult1, (v16i8)cnst1);   \
+  }
 #define DOTP_SB2_SH(...) DOTP_SB2(v8i16, __VA_ARGS__)
 
-#define DOTP_SB4(RTYPE, mult0, mult1, mult2, mult3,                   \
-                 cnst0, cnst1, cnst2, cnst3, out0, out1, out2, out3)  \
-{                                                                     \
-    DOTP_SB2(RTYPE, mult0, mult1, cnst0, cnst1, out0, out1);          \
-    DOTP_SB2(RTYPE, mult2, mult3, cnst2, cnst3, out2, out3);          \
-}
+#define DOTP_SB4(RTYPE, mult0, mult1, mult2, mult3, cnst0, cnst1, cnst2, \
+                 cnst3, out0, out1, out2, out3)                          \
+  {                                                                      \
+    DOTP_SB2(RTYPE, mult0, mult1, cnst0, cnst1, out0, out1);             \
+    DOTP_SB2(RTYPE, mult2, mult3, cnst2, cnst3, out2, out3);             \
+  }
 #define DOTP_SB4_SH(...) DOTP_SB4(v8i16, __VA_ARGS__)
 
 /* Description : Dot product of halfword vector elements
@@ -657,19 +637,18 @@
                  The multiplication result of adjacent odd-even elements
                  are added together and written to the 'out0' vector
 */
-#define DOTP_SH2(RTYPE, mult0, mult1, cnst0, cnst1, out0, out1)  \
-{                                                                \
-    out0 = (RTYPE)__msa_dotp_s_w((v8i16)mult0, (v8i16)cnst0);    \
-    out1 = (RTYPE)__msa_dotp_s_w((v8i16)mult1, (v8i16)cnst1);    \
-}
+#define DOTP_SH2(RTYPE, mult0, mult1, cnst0, cnst1, out0, out1) \
+  {                                                             \
+    out0 = (RTYPE)__msa_dotp_s_w((v8i16)mult0, (v8i16)cnst0);   \
+    out1 = (RTYPE)__msa_dotp_s_w((v8i16)mult1, (v8i16)cnst1);   \
+  }
 
-#define DOTP_SH4(RTYPE, mult0, mult1, mult2, mult3,           \
-                 cnst0, cnst1, cnst2, cnst3,                  \
-                 out0, out1, out2, out3)                      \
-{                                                             \
-    DOTP_SH2(RTYPE, mult0, mult1, cnst0, cnst1, out0, out1);  \
-    DOTP_SH2(RTYPE, mult2, mult3, cnst2, cnst3, out2, out3);  \
-}
+#define DOTP_SH4(RTYPE, mult0, mult1, mult2, mult3, cnst0, cnst1, cnst2, \
+                 cnst3, out0, out1, out2, out3)                          \
+  {                                                                      \
+    DOTP_SH2(RTYPE, mult0, mult1, cnst0, cnst1, out0, out1);             \
+    DOTP_SH2(RTYPE, mult2, mult3, cnst2, cnst3, out2, out3);             \
+  }
 #define DOTP_SH4_SW(...) DOTP_SH4(v4i32, __VA_ARGS__)
 
 /* Description : Dot product of word vector elements
@@ -682,11 +661,11 @@
                  The multiplication result of adjacent odd-even elements
                  are added together and written to the 'out0' vector
 */
-#define DOTP_SW2(RTYPE, mult0, mult1, cnst0, cnst1, out0, out1)  \
-{                                                                \
-    out0 = (RTYPE)__msa_dotp_s_d((v4i32)mult0, (v4i32)cnst0);    \
-    out1 = (RTYPE)__msa_dotp_s_d((v4i32)mult1, (v4i32)cnst1);    \
-}
+#define DOTP_SW2(RTYPE, mult0, mult1, cnst0, cnst1, out0, out1) \
+  {                                                             \
+    out0 = (RTYPE)__msa_dotp_s_d((v4i32)mult0, (v4i32)cnst0);   \
+    out1 = (RTYPE)__msa_dotp_s_d((v4i32)mult1, (v4i32)cnst1);   \
+  }
 #define DOTP_SW2_SD(...) DOTP_SW2(v2i64, __VA_ARGS__)
 
 /* Description : Dot product & addition of byte vector elements
@@ -699,19 +678,19 @@
                  The multiplication result of adjacent odd-even elements
                  are added to the 'out0' vector
 */
-#define DPADD_SB2(RTYPE, mult0, mult1, cnst0, cnst1, out0, out1)             \
-{                                                                            \
-    out0 = (RTYPE)__msa_dpadd_s_h((v8i16)out0, (v16i8)mult0, (v16i8)cnst0);  \
-    out1 = (RTYPE)__msa_dpadd_s_h((v8i16)out1, (v16i8)mult1, (v16i8)cnst1);  \
-}
+#define DPADD_SB2(RTYPE, mult0, mult1, cnst0, cnst1, out0, out1)            \
+  {                                                                         \
+    out0 = (RTYPE)__msa_dpadd_s_h((v8i16)out0, (v16i8)mult0, (v16i8)cnst0); \
+    out1 = (RTYPE)__msa_dpadd_s_h((v8i16)out1, (v16i8)mult1, (v16i8)cnst1); \
+  }
 #define DPADD_SB2_SH(...) DPADD_SB2(v8i16, __VA_ARGS__)
 
-#define DPADD_SB4(RTYPE, mult0, mult1, mult2, mult3,                   \
-                  cnst0, cnst1, cnst2, cnst3, out0, out1, out2, out3)  \
-{                                                                      \
-    DPADD_SB2(RTYPE, mult0, mult1, cnst0, cnst1, out0, out1);          \
-    DPADD_SB2(RTYPE, mult2, mult3, cnst2, cnst3, out2, out3);          \
-}
+#define DPADD_SB4(RTYPE, mult0, mult1, mult2, mult3, cnst0, cnst1, cnst2, \
+                  cnst3, out0, out1, out2, out3)                          \
+  {                                                                       \
+    DPADD_SB2(RTYPE, mult0, mult1, cnst0, cnst1, out0, out1);             \
+    DPADD_SB2(RTYPE, mult2, mult3, cnst2, cnst3, out2, out3);             \
+  }
 #define DPADD_SB4_SH(...) DPADD_SB4(v8i16, __VA_ARGS__)
 
 /* Description : Dot product & addition of halfword vector elements
@@ -724,19 +703,19 @@
                  The multiplication result of adjacent odd-even elements
                  are added to the 'out0' vector
 */
-#define DPADD_SH2(RTYPE, mult0, mult1, cnst0, cnst1, out0, out1)             \
-{                                                                            \
-    out0 = (RTYPE)__msa_dpadd_s_w((v4i32)out0, (v8i16)mult0, (v8i16)cnst0);  \
-    out1 = (RTYPE)__msa_dpadd_s_w((v4i32)out1, (v8i16)mult1, (v8i16)cnst1);  \
-}
+#define DPADD_SH2(RTYPE, mult0, mult1, cnst0, cnst1, out0, out1)            \
+  {                                                                         \
+    out0 = (RTYPE)__msa_dpadd_s_w((v4i32)out0, (v8i16)mult0, (v8i16)cnst0); \
+    out1 = (RTYPE)__msa_dpadd_s_w((v4i32)out1, (v8i16)mult1, (v8i16)cnst1); \
+  }
 #define DPADD_SH2_SW(...) DPADD_SH2(v4i32, __VA_ARGS__)
 
-#define DPADD_SH4(RTYPE, mult0, mult1, mult2, mult3,                   \
-                  cnst0, cnst1, cnst2, cnst3, out0, out1, out2, out3)  \
-{                                                                      \
-    DPADD_SH2(RTYPE, mult0, mult1, cnst0, cnst1, out0, out1);          \
-    DPADD_SH2(RTYPE, mult2, mult3, cnst2, cnst3, out2, out3);          \
-}
+#define DPADD_SH4(RTYPE, mult0, mult1, mult2, mult3, cnst0, cnst1, cnst2, \
+                  cnst3, out0, out1, out2, out3)                          \
+  {                                                                       \
+    DPADD_SH2(RTYPE, mult0, mult1, cnst0, cnst1, out0, out1);             \
+    DPADD_SH2(RTYPE, mult2, mult3, cnst2, cnst3, out2, out3);             \
+  }
 #define DPADD_SH4_SW(...) DPADD_SH4(v4i32, __VA_ARGS__)
 
 /* Description : Dot product & addition of double word vector elements
@@ -749,11 +728,11 @@
                  The multiplication result of adjacent odd-even elements
                  are added to the 'out0' vector
 */
-#define DPADD_SD2(RTYPE, mult0, mult1, out0, out1)                           \
-{                                                                            \
-    out0 = (RTYPE)__msa_dpadd_s_d((v2i64)out0, (v4i32)mult0, (v4i32)mult0);  \
-    out1 = (RTYPE)__msa_dpadd_s_d((v2i64)out1, (v4i32)mult1, (v4i32)mult1);  \
-}
+#define DPADD_SD2(RTYPE, mult0, mult1, out0, out1)                          \
+  {                                                                         \
+    out0 = (RTYPE)__msa_dpadd_s_d((v2i64)out0, (v4i32)mult0, (v4i32)mult0); \
+    out1 = (RTYPE)__msa_dpadd_s_d((v2i64)out1, (v4i32)mult1, (v4i32)mult1); \
+  }
 #define DPADD_SD2_SD(...) DPADD_SD2(v2i64, __VA_ARGS__)
 
 /* Description : Clips all signed halfword elements of input vector
@@ -762,25 +741,25 @@
                  Output - out_m
                  Return Type - signed halfword
 */
-#define CLIP_SH_0_255(in)                               \
-({                                                      \
-    v8i16 max_m = __msa_ldi_h(255);                     \
-    v8i16 out_m;                                        \
-                                                        \
-    out_m = __msa_maxi_s_h((v8i16)in, 0);               \
-    out_m = __msa_min_s_h((v8i16)max_m, (v8i16)out_m);  \
-    out_m;                                              \
-})
-#define CLIP_SH2_0_255(in0, in1)  \
-{                                 \
-    in0 = CLIP_SH_0_255(in0);     \
-    in1 = CLIP_SH_0_255(in1);     \
-}
-#define CLIP_SH4_0_255(in0, in1, in2, in3)  \
-{                                           \
-    CLIP_SH2_0_255(in0, in1);               \
-    CLIP_SH2_0_255(in2, in3);               \
-}
+#define CLIP_SH_0_255(in)                              \
+  ({                                                   \
+    v8i16 max_m = __msa_ldi_h(255);                    \
+    v8i16 out_m;                                       \
+                                                       \
+    out_m = __msa_maxi_s_h((v8i16)in, 0);              \
+    out_m = __msa_min_s_h((v8i16)max_m, (v8i16)out_m); \
+    out_m;                                             \
+  })
+#define CLIP_SH2_0_255(in0, in1) \
+  {                              \
+    in0 = CLIP_SH_0_255(in0);    \
+    in1 = CLIP_SH_0_255(in1);    \
+  }
+#define CLIP_SH4_0_255(in0, in1, in2, in3) \
+  {                                        \
+    CLIP_SH2_0_255(in0, in1);              \
+    CLIP_SH2_0_255(in2, in3);              \
+  }
 
 /* Description : Clips all signed word elements of input vector
                  between 0 & 255
@@ -788,15 +767,15 @@
                  Output - out_m
                  Return Type - signed word
 */
-#define CLIP_SW_0_255(in)                               \
-({                                                      \
-    v4i32 max_m = __msa_ldi_w(255);                     \
-    v4i32 out_m;                                        \
-                                                        \
-    out_m = __msa_maxi_s_w((v4i32)in, 0);               \
-    out_m = __msa_min_s_w((v4i32)max_m, (v4i32)out_m);  \
-    out_m;                                              \
-})
+#define CLIP_SW_0_255(in)                              \
+  ({                                                   \
+    v4i32 max_m = __msa_ldi_w(255);                    \
+    v4i32 out_m;                                       \
+                                                       \
+    out_m = __msa_maxi_s_w((v4i32)in, 0);              \
+    out_m = __msa_min_s_w((v4i32)max_m, (v4i32)out_m); \
+    out_m;                                             \
+  })
 
 /* Description : Horizontal addition of 4 signed word elements of input vector
    Arguments   : Input  - in       (signed word vector)
@@ -805,17 +784,17 @@
    Details     : 4 signed word elements of 'in' vector are added together and
                  the resulting integer sum is returned
 */
-#define HADD_SW_S32(in)                             \
-({                                                  \
-    v2i64 res0_m, res1_m;                           \
-    int32_t sum_m;                                  \
-                                                    \
-    res0_m = __msa_hadd_s_d((v4i32)in, (v4i32)in);  \
-    res1_m = __msa_splati_d(res0_m, 1);             \
-    res0_m = res0_m + res1_m;                       \
-    sum_m = __msa_copy_s_w((v4i32)res0_m, 0);       \
-    sum_m;                                          \
-})
+#define HADD_SW_S32(in)                            \
+  ({                                               \
+    v2i64 res0_m, res1_m;                          \
+    int32_t sum_m;                                 \
+                                                   \
+    res0_m = __msa_hadd_s_d((v4i32)in, (v4i32)in); \
+    res1_m = __msa_splati_d(res0_m, 1);            \
+    res0_m = res0_m + res1_m;                      \
+    sum_m = __msa_copy_s_w((v4i32)res0_m, 0);      \
+    sum_m;                                         \
+  })
 
 /* Description : Horizontal addition of 8 unsigned halfword elements
    Arguments   : Inputs  - in       (unsigned halfword vector)
@@ -824,19 +803,19 @@
    Details     : 8 unsigned halfword elements of input vector are added
                  together and the resulting integer sum is returned
 */
-#define HADD_UH_U32(in)                                \
-({                                                     \
-    v4u32 res_m;                                       \
-    v2u64 res0_m, res1_m;                              \
-    uint32_t sum_m;                                    \
-                                                       \
-    res_m = __msa_hadd_u_w((v8u16)in, (v8u16)in);      \
-    res0_m = __msa_hadd_u_d(res_m, res_m);             \
-    res1_m = (v2u64)__msa_splati_d((v2i64)res0_m, 1);  \
-    res0_m = res0_m + res1_m;                          \
-    sum_m = __msa_copy_u_w((v4i32)res0_m, 0);          \
-    sum_m;                                             \
-})
+#define HADD_UH_U32(in)                               \
+  ({                                                  \
+    v4u32 res_m;                                      \
+    v2u64 res0_m, res1_m;                             \
+    uint32_t sum_m;                                   \
+                                                      \
+    res_m = __msa_hadd_u_w((v8u16)in, (v8u16)in);     \
+    res0_m = __msa_hadd_u_d(res_m, res_m);            \
+    res1_m = (v2u64)__msa_splati_d((v2i64)res0_m, 1); \
+    res0_m = res0_m + res1_m;                         \
+    sum_m = __msa_copy_u_w((v4i32)res0_m, 0);         \
+    sum_m;                                            \
+  })
 
 /* Description : Horizontal addition of unsigned byte vector elements
    Arguments   : Inputs  - in0, in1
@@ -846,11 +825,11 @@
                  even unsigned byte element from 'in0' (pairwise) and the
                  halfword result is written to 'out0'
 */
-#define HADD_UB2(RTYPE, in0, in1, out0, out1)              \
-{                                                          \
-    out0 = (RTYPE)__msa_hadd_u_h((v16u8)in0, (v16u8)in0);  \
-    out1 = (RTYPE)__msa_hadd_u_h((v16u8)in1, (v16u8)in1);  \
-}
+#define HADD_UB2(RTYPE, in0, in1, out0, out1)             \
+  {                                                       \
+    out0 = (RTYPE)__msa_hadd_u_h((v16u8)in0, (v16u8)in0); \
+    out1 = (RTYPE)__msa_hadd_u_h((v16u8)in1, (v16u8)in1); \
+  }
 #define HADD_UB2_UH(...) HADD_UB2(v8u16, __VA_ARGS__)
 
 /* Description : Horizontal subtraction of unsigned byte vector elements
@@ -861,11 +840,11 @@
                  even unsigned byte element from 'in0' (pairwise) and the
                  halfword result is written to 'out0'
 */
-#define HSUB_UB2(RTYPE, in0, in1, out0, out1)              \
-{                                                          \
-    out0 = (RTYPE)__msa_hsub_u_h((v16u8)in0, (v16u8)in0);  \
-    out1 = (RTYPE)__msa_hsub_u_h((v16u8)in1, (v16u8)in1);  \
-}
+#define HSUB_UB2(RTYPE, in0, in1, out0, out1)             \
+  {                                                       \
+    out0 = (RTYPE)__msa_hsub_u_h((v16u8)in0, (v16u8)in0); \
+    out1 = (RTYPE)__msa_hsub_u_h((v16u8)in1, (v16u8)in1); \
+  }
 #define HSUB_UB2_SH(...) HSUB_UB2(v8i16, __VA_ARGS__)
 
 /* Description : Horizontal subtraction of signed halfword vector elements
@@ -876,11 +855,11 @@
                  even signed halfword element from 'in0' (pairwise) and the
                  word result is written to 'out0'
 */
-#define HSUB_UH2(RTYPE, in0, in1, out0, out1)              \
-{                                                          \
-    out0 = (RTYPE)__msa_hsub_s_w((v8i16)in0, (v8i16)in0);  \
-    out1 = (RTYPE)__msa_hsub_s_w((v8i16)in1, (v8i16)in1);  \
-}
+#define HSUB_UH2(RTYPE, in0, in1, out0, out1)             \
+  {                                                       \
+    out0 = (RTYPE)__msa_hsub_s_w((v8i16)in0, (v8i16)in0); \
+    out1 = (RTYPE)__msa_hsub_s_w((v8i16)in1, (v8i16)in1); \
+  }
 #define HSUB_UH2_SW(...) HSUB_UH2(v4i32, __VA_ARGS__)
 
 /* Description : Set element n input vector to GPR value
@@ -889,11 +868,11 @@
                  Return Type - as per RTYPE
    Details     : Set element 0 in vector 'out' to value specified in 'in0'
 */
-#define INSERT_D2(RTYPE, in0, in1, out)               \
-{                                                     \
-    out = (RTYPE)__msa_insert_d((v2i64)out, 0, in0);  \
-    out = (RTYPE)__msa_insert_d((v2i64)out, 1, in1);  \
-}
+#define INSERT_D2(RTYPE, in0, in1, out)              \
+  {                                                  \
+    out = (RTYPE)__msa_insert_d((v2i64)out, 0, in0); \
+    out = (RTYPE)__msa_insert_d((v2i64)out, 1, in1); \
+  }
 #define INSERT_D2_SB(...) INSERT_D2(v16i8, __VA_ARGS__)
 
 /* Description : Interleave even byte elements from vectors
@@ -903,11 +882,11 @@
    Details     : Even byte elements of 'in0' and 'in1' are interleaved
                  and written to 'out0'
 */
-#define ILVEV_B2(RTYPE, in0, in1, in2, in3, out0, out1)   \
-{                                                         \
-    out0 = (RTYPE)__msa_ilvev_b((v16i8)in1, (v16i8)in0);  \
-    out1 = (RTYPE)__msa_ilvev_b((v16i8)in3, (v16i8)in2);  \
-}
+#define ILVEV_B2(RTYPE, in0, in1, in2, in3, out0, out1)  \
+  {                                                      \
+    out0 = (RTYPE)__msa_ilvev_b((v16i8)in1, (v16i8)in0); \
+    out1 = (RTYPE)__msa_ilvev_b((v16i8)in3, (v16i8)in2); \
+  }
 #define ILVEV_B2_UB(...) ILVEV_B2(v16u8, __VA_ARGS__)
 #define ILVEV_B2_SH(...) ILVEV_B2(v8i16, __VA_ARGS__)
 #define ILVEV_B2_SD(...) ILVEV_B2(v2i64, __VA_ARGS__)
@@ -919,11 +898,11 @@
    Details     : Even halfword elements of 'in0' and 'in1' are interleaved
                  and written to 'out0'
 */
-#define ILVEV_H2(RTYPE, in0, in1, in2, in3, out0, out1)   \
-{                                                         \
-    out0 = (RTYPE)__msa_ilvev_h((v8i16)in1, (v8i16)in0);  \
-    out1 = (RTYPE)__msa_ilvev_h((v8i16)in3, (v8i16)in2);  \
-}
+#define ILVEV_H2(RTYPE, in0, in1, in2, in3, out0, out1)  \
+  {                                                      \
+    out0 = (RTYPE)__msa_ilvev_h((v8i16)in1, (v8i16)in0); \
+    out1 = (RTYPE)__msa_ilvev_h((v8i16)in3, (v8i16)in2); \
+  }
 #define ILVEV_H2_UB(...) ILVEV_H2(v16u8, __VA_ARGS__)
 #define ILVEV_H2_SH(...) ILVEV_H2(v8i16, __VA_ARGS__)
 
@@ -934,11 +913,11 @@
    Details     : Even word elements of 'in0' and 'in1' are interleaved
                  and written to 'out0'
 */
-#define ILVEV_W2(RTYPE, in0, in1, in2, in3, out0, out1)   \
-{                                                         \
-    out0 = (RTYPE)__msa_ilvev_w((v4i32)in1, (v4i32)in0);  \
-    out1 = (RTYPE)__msa_ilvev_w((v4i32)in3, (v4i32)in2);  \
-}
+#define ILVEV_W2(RTYPE, in0, in1, in2, in3, out0, out1)  \
+  {                                                      \
+    out0 = (RTYPE)__msa_ilvev_w((v4i32)in1, (v4i32)in0); \
+    out1 = (RTYPE)__msa_ilvev_w((v4i32)in3, (v4i32)in2); \
+  }
 #define ILVEV_W2_SD(...) ILVEV_W2(v2i64, __VA_ARGS__)
 
 /* Description : Interleave even double word elements from vectors
@@ -948,11 +927,11 @@
    Details     : Even double word elements of 'in0' and 'in1' are interleaved
                  and written to 'out0'
 */
-#define ILVEV_D2(RTYPE, in0, in1, in2, in3, out0, out1)   \
-{                                                         \
-    out0 = (RTYPE)__msa_ilvev_d((v2i64)in1, (v2i64)in0);  \
-    out1 = (RTYPE)__msa_ilvev_d((v2i64)in3, (v2i64)in2);  \
-}
+#define ILVEV_D2(RTYPE, in0, in1, in2, in3, out0, out1)  \
+  {                                                      \
+    out0 = (RTYPE)__msa_ilvev_d((v2i64)in1, (v2i64)in0); \
+    out1 = (RTYPE)__msa_ilvev_d((v2i64)in3, (v2i64)in2); \
+  }
 #define ILVEV_D2_UB(...) ILVEV_D2(v16u8, __VA_ARGS__)
 
 /* Description : Interleave left half of byte elements from vectors
@@ -962,21 +941,21 @@
    Details     : Left half of byte elements of 'in0' and 'in1' are interleaved
                  and written to 'out0'.
 */
-#define ILVL_B2(RTYPE, in0, in1, in2, in3, out0, out1)   \
-{                                                        \
-    out0 = (RTYPE)__msa_ilvl_b((v16i8)in0, (v16i8)in1);  \
-    out1 = (RTYPE)__msa_ilvl_b((v16i8)in2, (v16i8)in3);  \
-}
+#define ILVL_B2(RTYPE, in0, in1, in2, in3, out0, out1)  \
+  {                                                     \
+    out0 = (RTYPE)__msa_ilvl_b((v16i8)in0, (v16i8)in1); \
+    out1 = (RTYPE)__msa_ilvl_b((v16i8)in2, (v16i8)in3); \
+  }
 #define ILVL_B2_UB(...) ILVL_B2(v16u8, __VA_ARGS__)
 #define ILVL_B2_SB(...) ILVL_B2(v16i8, __VA_ARGS__)
 #define ILVL_B2_SH(...) ILVL_B2(v8i16, __VA_ARGS__)
 
-#define ILVL_B4(RTYPE, in0, in1, in2, in3, in4, in5, in6, in7,  \
-                out0, out1, out2, out3)                         \
-{                                                               \
-    ILVL_B2(RTYPE, in0, in1, in2, in3, out0, out1);             \
-    ILVL_B2(RTYPE, in4, in5, in6, in7, out2, out3);             \
-}
+#define ILVL_B4(RTYPE, in0, in1, in2, in3, in4, in5, in6, in7, out0, out1, \
+                out2, out3)                                                \
+  {                                                                        \
+    ILVL_B2(RTYPE, in0, in1, in2, in3, out0, out1);                        \
+    ILVL_B2(RTYPE, in4, in5, in6, in7, out2, out3);                        \
+  }
 #define ILVL_B4_SB(...) ILVL_B4(v16i8, __VA_ARGS__)
 #define ILVL_B4_SH(...) ILVL_B4(v8i16, __VA_ARGS__)
 
@@ -987,11 +966,11 @@
    Details     : Left half of halfword elements of 'in0' and 'in1' are
                  interleaved and written to 'out0'.
 */
-#define ILVL_H2(RTYPE, in0, in1, in2, in3, out0, out1)   \
-{                                                        \
-    out0 = (RTYPE)__msa_ilvl_h((v8i16)in0, (v8i16)in1);  \
-    out1 = (RTYPE)__msa_ilvl_h((v8i16)in2, (v8i16)in3);  \
-}
+#define ILVL_H2(RTYPE, in0, in1, in2, in3, out0, out1)  \
+  {                                                     \
+    out0 = (RTYPE)__msa_ilvl_h((v8i16)in0, (v8i16)in1); \
+    out1 = (RTYPE)__msa_ilvl_h((v8i16)in2, (v8i16)in3); \
+  }
 #define ILVL_H2_SH(...) ILVL_H2(v8i16, __VA_ARGS__)
 #define ILVL_H2_SW(...) ILVL_H2(v4i32, __VA_ARGS__)
 
@@ -1002,11 +981,11 @@
    Details     : Left half of word elements of 'in0' and 'in1' are interleaved
                  and written to 'out0'.
 */
-#define ILVL_W2(RTYPE, in0, in1, in2, in3, out0, out1)   \
-{                                                        \
-    out0 = (RTYPE)__msa_ilvl_w((v4i32)in0, (v4i32)in1);  \
-    out1 = (RTYPE)__msa_ilvl_w((v4i32)in2, (v4i32)in3);  \
-}
+#define ILVL_W2(RTYPE, in0, in1, in2, in3, out0, out1)  \
+  {                                                     \
+    out0 = (RTYPE)__msa_ilvl_w((v4i32)in0, (v4i32)in1); \
+    out1 = (RTYPE)__msa_ilvl_w((v4i32)in2, (v4i32)in3); \
+  }
 #define ILVL_W2_SH(...) ILVL_W2(v8i16, __VA_ARGS__)
 
 /* Description : Interleave right half of byte elements from vectors
@@ -1016,22 +995,22 @@
    Details     : Right half of byte elements of 'in0' and 'in1' are interleaved
                  and written to out0.
 */
-#define ILVR_B2(RTYPE, in0, in1, in2, in3, out0, out1)   \
-{                                                        \
-    out0 = (RTYPE)__msa_ilvr_b((v16i8)in0, (v16i8)in1);  \
-    out1 = (RTYPE)__msa_ilvr_b((v16i8)in2, (v16i8)in3);  \
-}
+#define ILVR_B2(RTYPE, in0, in1, in2, in3, out0, out1)  \
+  {                                                     \
+    out0 = (RTYPE)__msa_ilvr_b((v16i8)in0, (v16i8)in1); \
+    out1 = (RTYPE)__msa_ilvr_b((v16i8)in2, (v16i8)in3); \
+  }
 #define ILVR_B2_UB(...) ILVR_B2(v16u8, __VA_ARGS__)
 #define ILVR_B2_SB(...) ILVR_B2(v16i8, __VA_ARGS__)
 #define ILVR_B2_SH(...) ILVR_B2(v8i16, __VA_ARGS__)
 #define ILVR_B2_SW(...) ILVR_B2(v4i32, __VA_ARGS__)
 
-#define ILVR_B4(RTYPE, in0, in1, in2, in3, in4, in5, in6, in7,  \
-                out0, out1, out2, out3)                         \
-{                                                               \
-    ILVR_B2(RTYPE, in0, in1, in2, in3, out0, out1);             \
-    ILVR_B2(RTYPE, in4, in5, in6, in7, out2, out3);             \
-}
+#define ILVR_B4(RTYPE, in0, in1, in2, in3, in4, in5, in6, in7, out0, out1, \
+                out2, out3)                                                \
+  {                                                                        \
+    ILVR_B2(RTYPE, in0, in1, in2, in3, out0, out1);                        \
+    ILVR_B2(RTYPE, in4, in5, in6, in7, out2, out3);                        \
+  }
 #define ILVR_B4_UB(...) ILVR_B4(v16u8, __VA_ARGS__)
 #define ILVR_B4_SB(...) ILVR_B4(v16i8, __VA_ARGS__)
 #define ILVR_B4_UH(...) ILVR_B4(v8u16, __VA_ARGS__)
@@ -1045,28 +1024,28 @@
    Details     : Right half of halfword elements of 'in0' and 'in1' are
                  interleaved and written to 'out0'.
 */
-#define ILVR_H2(RTYPE, in0, in1, in2, in3, out0, out1)   \
-{                                                        \
-    out0 = (RTYPE)__msa_ilvr_h((v8i16)in0, (v8i16)in1);  \
-    out1 = (RTYPE)__msa_ilvr_h((v8i16)in2, (v8i16)in3);  \
-}
+#define ILVR_H2(RTYPE, in0, in1, in2, in3, out0, out1)  \
+  {                                                     \
+    out0 = (RTYPE)__msa_ilvr_h((v8i16)in0, (v8i16)in1); \
+    out1 = (RTYPE)__msa_ilvr_h((v8i16)in2, (v8i16)in3); \
+  }
 #define ILVR_H2_SH(...) ILVR_H2(v8i16, __VA_ARGS__)
 #define ILVR_H2_SW(...) ILVR_H2(v4i32, __VA_ARGS__)
 
-#define ILVR_H4(RTYPE, in0, in1, in2, in3, in4, in5, in6, in7,  \
-                out0, out1, out2, out3)                         \
-{                                                               \
-    ILVR_H2(RTYPE, in0, in1, in2, in3, out0, out1);             \
-    ILVR_H2(RTYPE, in4, in5, in6, in7, out2, out3);             \
-}
+#define ILVR_H4(RTYPE, in0, in1, in2, in3, in4, in5, in6, in7, out0, out1, \
+                out2, out3)                                                \
+  {                                                                        \
+    ILVR_H2(RTYPE, in0, in1, in2, in3, out0, out1);                        \
+    ILVR_H2(RTYPE, in4, in5, in6, in7, out2, out3);                        \
+  }
 #define ILVR_H4_SH(...) ILVR_H4(v8i16, __VA_ARGS__)
 #define ILVR_H4_SW(...) ILVR_H4(v4i32, __VA_ARGS__)
 
-#define ILVR_W2(RTYPE, in0, in1, in2, in3, out0, out1)   \
-{                                                        \
-    out0 = (RTYPE)__msa_ilvr_w((v4i32)in0, (v4i32)in1);  \
-    out1 = (RTYPE)__msa_ilvr_w((v4i32)in2, (v4i32)in3);  \
-}
+#define ILVR_W2(RTYPE, in0, in1, in2, in3, out0, out1)  \
+  {                                                     \
+    out0 = (RTYPE)__msa_ilvr_w((v4i32)in0, (v4i32)in1); \
+    out1 = (RTYPE)__msa_ilvr_w((v4i32)in2, (v4i32)in3); \
+  }
 #define ILVR_W2_SH(...) ILVR_W2(v8i16, __VA_ARGS__)
 
 /* Description : Interleave right half of double word elements from vectors
@@ -1076,21 +1055,21 @@
    Details     : Right half of double word elements of 'in0' and 'in1' are
                  interleaved and written to 'out0'.
 */
-#define ILVR_D2(RTYPE, in0, in1, in2, in3, out0, out1)       \
-{                                                            \
-    out0 = (RTYPE)__msa_ilvr_d((v2i64)(in0), (v2i64)(in1));  \
-    out1 = (RTYPE)__msa_ilvr_d((v2i64)(in2), (v2i64)(in3));  \
-}
+#define ILVR_D2(RTYPE, in0, in1, in2, in3, out0, out1)      \
+  {                                                         \
+    out0 = (RTYPE)__msa_ilvr_d((v2i64)(in0), (v2i64)(in1)); \
+    out1 = (RTYPE)__msa_ilvr_d((v2i64)(in2), (v2i64)(in3)); \
+  }
 #define ILVR_D2_UB(...) ILVR_D2(v16u8, __VA_ARGS__)
 #define ILVR_D2_SB(...) ILVR_D2(v16i8, __VA_ARGS__)
 #define ILVR_D2_SH(...) ILVR_D2(v8i16, __VA_ARGS__)
 
-#define ILVR_D4(RTYPE, in0, in1, in2, in3, in4, in5, in6, in7,  \
-                out0, out1, out2, out3)                         \
-{                                                               \
-    ILVR_D2(RTYPE, in0, in1, in2, in3, out0, out1);             \
-    ILVR_D2(RTYPE, in4, in5, in6, in7, out2, out3);             \
-}
+#define ILVR_D4(RTYPE, in0, in1, in2, in3, in4, in5, in6, in7, out0, out1, \
+                out2, out3)                                                \
+  {                                                                        \
+    ILVR_D2(RTYPE, in0, in1, in2, in3, out0, out1);                        \
+    ILVR_D2(RTYPE, in4, in5, in6, in7, out2, out3);                        \
+  }
 #define ILVR_D4_SB(...) ILVR_D4(v16i8, __VA_ARGS__)
 #define ILVR_D4_UB(...) ILVR_D4(v16u8, __VA_ARGS__)
 
@@ -1101,29 +1080,29 @@
    Details     : Right half of byte elements from 'in0' and 'in1' are
                  interleaved and written to 'out0'
 */
-#define ILVRL_B2(RTYPE, in0, in1, out0, out1)            \
-{                                                        \
-    out0 = (RTYPE)__msa_ilvr_b((v16i8)in0, (v16i8)in1);  \
-    out1 = (RTYPE)__msa_ilvl_b((v16i8)in0, (v16i8)in1);  \
-}
+#define ILVRL_B2(RTYPE, in0, in1, out0, out1)           \
+  {                                                     \
+    out0 = (RTYPE)__msa_ilvr_b((v16i8)in0, (v16i8)in1); \
+    out1 = (RTYPE)__msa_ilvl_b((v16i8)in0, (v16i8)in1); \
+  }
 #define ILVRL_B2_UB(...) ILVRL_B2(v16u8, __VA_ARGS__)
 #define ILVRL_B2_SB(...) ILVRL_B2(v16i8, __VA_ARGS__)
 #define ILVRL_B2_UH(...) ILVRL_B2(v8u16, __VA_ARGS__)
 #define ILVRL_B2_SH(...) ILVRL_B2(v8i16, __VA_ARGS__)
 
-#define ILVRL_H2(RTYPE, in0, in1, out0, out1)            \
-{                                                        \
-    out0 = (RTYPE)__msa_ilvr_h((v8i16)in0, (v8i16)in1);  \
-    out1 = (RTYPE)__msa_ilvl_h((v8i16)in0, (v8i16)in1);  \
-}
+#define ILVRL_H2(RTYPE, in0, in1, out0, out1)           \
+  {                                                     \
+    out0 = (RTYPE)__msa_ilvr_h((v8i16)in0, (v8i16)in1); \
+    out1 = (RTYPE)__msa_ilvl_h((v8i16)in0, (v8i16)in1); \
+  }
 #define ILVRL_H2_SH(...) ILVRL_H2(v8i16, __VA_ARGS__)
 #define ILVRL_H2_SW(...) ILVRL_H2(v4i32, __VA_ARGS__)
 
-#define ILVRL_W2(RTYPE, in0, in1, out0, out1)            \
-{                                                        \
-    out0 = (RTYPE)__msa_ilvr_w((v4i32)in0, (v4i32)in1);  \
-    out1 = (RTYPE)__msa_ilvl_w((v4i32)in0, (v4i32)in1);  \
-}
+#define ILVRL_W2(RTYPE, in0, in1, out0, out1)           \
+  {                                                     \
+    out0 = (RTYPE)__msa_ilvr_w((v4i32)in0, (v4i32)in1); \
+    out1 = (RTYPE)__msa_ilvl_w((v4i32)in0, (v4i32)in1); \
+  }
 #define ILVRL_W2_UB(...) ILVRL_W2(v16u8, __VA_ARGS__)
 #define ILVRL_W2_SH(...) ILVRL_W2(v8i16, __VA_ARGS__)
 #define ILVRL_W2_SW(...) ILVRL_W2(v4i32, __VA_ARGS__)
@@ -1136,11 +1115,11 @@
    Details     : Maximum of signed halfword element values from 'in0' and
                  'max_val' are written in place
 */
-#define MAXI_SH2(RTYPE, in0, in1, max_val)               \
-{                                                        \
-    in0 = (RTYPE)__msa_maxi_s_h((v8i16)in0, (max_val));  \
-    in1 = (RTYPE)__msa_maxi_s_h((v8i16)in1, (max_val));  \
-}
+#define MAXI_SH2(RTYPE, in0, in1, max_val)              \
+  {                                                     \
+    in0 = (RTYPE)__msa_maxi_s_h((v8i16)in0, (max_val)); \
+    in1 = (RTYPE)__msa_maxi_s_h((v8i16)in1, (max_val)); \
+  }
 #define MAXI_SH2_SH(...) MAXI_SH2(v8i16, __VA_ARGS__)
 
 /* Description : Saturate the halfword element values to the max
@@ -1153,11 +1132,11 @@
                  value generated with (sat_val + 1) bit range.
                  The results are written in place
 */
-#define SAT_UH2(RTYPE, in0, in1, sat_val)             \
-{                                                     \
-    in0 = (RTYPE)__msa_sat_u_h((v8u16)in0, sat_val);  \
-    in1 = (RTYPE)__msa_sat_u_h((v8u16)in1, sat_val);  \
-}
+#define SAT_UH2(RTYPE, in0, in1, sat_val)            \
+  {                                                  \
+    in0 = (RTYPE)__msa_sat_u_h((v8u16)in0, sat_val); \
+    in1 = (RTYPE)__msa_sat_u_h((v8u16)in1, sat_val); \
+  }
 #define SAT_UH2_SH(...) SAT_UH2(v8i16, __VA_ARGS__)
 
 /* Description : Saturate the halfword element values to the max
@@ -1170,18 +1149,18 @@
                  value generated with (sat_val + 1) bit range
                  The results are written in place
 */
-#define SAT_SH2(RTYPE, in0, in1, sat_val)             \
-{                                                     \
-    in0 = (RTYPE)__msa_sat_s_h((v8i16)in0, sat_val);  \
-    in1 = (RTYPE)__msa_sat_s_h((v8i16)in1, sat_val);  \
-}
+#define SAT_SH2(RTYPE, in0, in1, sat_val)            \
+  {                                                  \
+    in0 = (RTYPE)__msa_sat_s_h((v8i16)in0, sat_val); \
+    in1 = (RTYPE)__msa_sat_s_h((v8i16)in1, sat_val); \
+  }
 #define SAT_SH2_SH(...) SAT_SH2(v8i16, __VA_ARGS__)
 
-#define SAT_SH4(RTYPE, in0, in1, in2, in3, sat_val)  \
-{                                                    \
-    SAT_SH2(RTYPE, in0, in1, sat_val);               \
-    SAT_SH2(RTYPE, in2, in3, sat_val);               \
-}
+#define SAT_SH4(RTYPE, in0, in1, in2, in3, sat_val) \
+  {                                                 \
+    SAT_SH2(RTYPE, in0, in1, sat_val);              \
+    SAT_SH2(RTYPE, in2, in3, sat_val);              \
+  }
 #define SAT_SH4_SH(...) SAT_SH4(v8i16, __VA_ARGS__)
 
 /* Description : Indexed halfword element values are replicated to all
@@ -1193,20 +1172,19 @@
                   elements in 'out0' vector
                   Valid index range for halfword operation is 0-7
 */
-#define SPLATI_H2(RTYPE, in, idx0, idx1, out0, out1)  \
-{                                                     \
-    out0 = (RTYPE)__msa_splati_h((v8i16)in, idx0);    \
-    out1 = (RTYPE)__msa_splati_h((v8i16)in, idx1);    \
-}
+#define SPLATI_H2(RTYPE, in, idx0, idx1, out0, out1) \
+  {                                                  \
+    out0 = (RTYPE)__msa_splati_h((v8i16)in, idx0);   \
+    out1 = (RTYPE)__msa_splati_h((v8i16)in, idx1);   \
+  }
 #define SPLATI_H2_SB(...) SPLATI_H2(v16i8, __VA_ARGS__)
 #define SPLATI_H2_SH(...) SPLATI_H2(v8i16, __VA_ARGS__)
 
-#define SPLATI_H3(RTYPE, in, idx0, idx1, idx2,      \
-                  out0, out1, out2)                 \
-{                                                   \
-    SPLATI_H2(RTYPE, in, idx0, idx1, out0, out1);   \
-    out2 = (RTYPE)__msa_splati_h((v8i16)in, idx2);  \
-}
+#define SPLATI_H3(RTYPE, in, idx0, idx1, idx2, out0, out1, out2) \
+  {                                                              \
+    SPLATI_H2(RTYPE, in, idx0, idx1, out0, out1);                \
+    out2 = (RTYPE)__msa_splati_h((v8i16)in, idx2);               \
+  }
 #define SPLATI_H3_SB(...) SPLATI_H3(v16i8, __VA_ARGS__)
 #define SPLATI_H3_SH(...) SPLATI_H3(v8i16, __VA_ARGS__)
 
@@ -1221,11 +1199,11 @@
                  elements in 'out1' vector
                  Valid index range for word operation is 0-3
 */
-#define SPLATI_W2(RTYPE, in, stidx, out0, out1)          \
-{                                                        \
-    out0 = (RTYPE)__msa_splati_w((v4i32)in, stidx);      \
-    out1 = (RTYPE)__msa_splati_w((v4i32)in, (stidx+1));  \
-}
+#define SPLATI_W2(RTYPE, in, stidx, out0, out1)           \
+  {                                                       \
+    out0 = (RTYPE)__msa_splati_w((v4i32)in, stidx);       \
+    out1 = (RTYPE)__msa_splati_w((v4i32)in, (stidx + 1)); \
+  }
 #define SPLATI_W2_SW(...) SPLATI_W2(v4i32, __VA_ARGS__)
 
 /* Description : Pack even byte elements of vector pairs
@@ -1236,20 +1214,20 @@
                  'out0' & even byte elements of 'in1' are copied to the right
                  half of 'out0'.
 */
-#define PCKEV_B2(RTYPE, in0, in1, in2, in3, out0, out1)   \
-{                                                         \
-    out0 = (RTYPE)__msa_pckev_b((v16i8)in0, (v16i8)in1);  \
-    out1 = (RTYPE)__msa_pckev_b((v16i8)in2, (v16i8)in3);  \
-}
+#define PCKEV_B2(RTYPE, in0, in1, in2, in3, out0, out1)  \
+  {                                                      \
+    out0 = (RTYPE)__msa_pckev_b((v16i8)in0, (v16i8)in1); \
+    out1 = (RTYPE)__msa_pckev_b((v16i8)in2, (v16i8)in3); \
+  }
 #define PCKEV_B2_SB(...) PCKEV_B2(v16i8, __VA_ARGS__)
 #define PCKEV_B2_UB(...) PCKEV_B2(v16u8, __VA_ARGS__)
 
-#define PCKEV_B4(RTYPE, in0, in1, in2, in3, in4, in5, in6, in7,  \
-                 out0, out1, out2, out3)                         \
-{                                                                \
-    PCKEV_B2(RTYPE, in0, in1, in2, in3, out0, out1);             \
-    PCKEV_B2(RTYPE, in4, in5, in6, in7, out2, out3);             \
-}
+#define PCKEV_B4(RTYPE, in0, in1, in2, in3, in4, in5, in6, in7, out0, out1, \
+                 out2, out3)                                                \
+  {                                                                         \
+    PCKEV_B2(RTYPE, in0, in1, in2, in3, out0, out1);                        \
+    PCKEV_B2(RTYPE, in4, in5, in6, in7, out2, out3);                        \
+  }
 #define PCKEV_B4_SB(...) PCKEV_B4(v16i8, __VA_ARGS__)
 #define PCKEV_B4_UB(...) PCKEV_B4(v16u8, __VA_ARGS__)
 #define PCKEV_B4_SH(...) PCKEV_B4(v8i16, __VA_ARGS__)
@@ -1262,19 +1240,19 @@
                  'out0' & even halfword elements of 'in1' are copied to the
                  right half of 'out0'.
 */
-#define PCKEV_H2(RTYPE, in0, in1, in2, in3, out0, out1)   \
-{                                                         \
-    out0 = (RTYPE)__msa_pckev_h((v8i16)in0, (v8i16)in1);  \
-    out1 = (RTYPE)__msa_pckev_h((v8i16)in2, (v8i16)in3);  \
-}
+#define PCKEV_H2(RTYPE, in0, in1, in2, in3, out0, out1)  \
+  {                                                      \
+    out0 = (RTYPE)__msa_pckev_h((v8i16)in0, (v8i16)in1); \
+    out1 = (RTYPE)__msa_pckev_h((v8i16)in2, (v8i16)in3); \
+  }
 #define PCKEV_H2_SH(...) PCKEV_H2(v8i16, __VA_ARGS__)
 
-#define PCKEV_H4(RTYPE, in0, in1, in2, in3, in4, in5, in6, in7,  \
-                 out0, out1, out2, out3)                         \
-{                                                                \
-    PCKEV_H2(RTYPE, in0, in1, in2, in3, out0, out1);             \
-    PCKEV_H2(RTYPE, in4, in5, in6, in7, out2, out3);             \
-}
+#define PCKEV_H4(RTYPE, in0, in1, in2, in3, in4, in5, in6, in7, out0, out1, \
+                 out2, out3)                                                \
+  {                                                                         \
+    PCKEV_H2(RTYPE, in0, in1, in2, in3, out0, out1);                        \
+    PCKEV_H2(RTYPE, in4, in5, in6, in7, out2, out3);                        \
+  }
 #define PCKEV_H4_SH(...) PCKEV_H4(v8i16, __VA_ARGS__)
 
 /* Description : Pack even double word elements of vector pairs
@@ -1285,11 +1263,11 @@
                  'out0' & even double elements of 'in1' are copied to the right
                  half of 'out0'.
 */
-#define PCKEV_D2(RTYPE, in0, in1, in2, in3, out0, out1)   \
-{                                                         \
-    out0 = (RTYPE)__msa_pckev_d((v2i64)in0, (v2i64)in1);  \
-    out1 = (RTYPE)__msa_pckev_d((v2i64)in2, (v2i64)in3);  \
-}
+#define PCKEV_D2(RTYPE, in0, in1, in2, in3, out0, out1)  \
+  {                                                      \
+    out0 = (RTYPE)__msa_pckev_d((v2i64)in0, (v2i64)in1); \
+    out1 = (RTYPE)__msa_pckev_d((v2i64)in2, (v2i64)in3); \
+  }
 #define PCKEV_D2_UB(...) PCKEV_D2(v16u8, __VA_ARGS__)
 #define PCKEV_D2_SH(...) PCKEV_D2(v8i16, __VA_ARGS__)
 
@@ -1301,11 +1279,11 @@
                  of 'out0' & odd double word elements of 'in1' are copied to
                  the right half of 'out0'.
 */
-#define PCKOD_D2(RTYPE, in0, in1, in2, in3, out0, out1)   \
-{                                                         \
-    out0 = (RTYPE)__msa_pckod_d((v2i64)in0, (v2i64)in1);  \
-    out1 = (RTYPE)__msa_pckod_d((v2i64)in2, (v2i64)in3);  \
-}
+#define PCKOD_D2(RTYPE, in0, in1, in2, in3, out0, out1)  \
+  {                                                      \
+    out0 = (RTYPE)__msa_pckod_d((v2i64)in0, (v2i64)in1); \
+    out1 = (RTYPE)__msa_pckod_d((v2i64)in2, (v2i64)in3); \
+  }
 #define PCKOD_D2_UB(...) PCKOD_D2(v16u8, __VA_ARGS__)
 #define PCKOD_D2_SH(...) PCKOD_D2(v8i16, __VA_ARGS__)
 
@@ -1316,41 +1294,41 @@
    Details     : Each unsigned byte element from input vector 'in0' is
                  logically xor'ed with 128 and the result is stored in-place.
 */
-#define XORI_B2_128(RTYPE, in0, in1)             \
-{                                                \
-    in0 = (RTYPE)__msa_xori_b((v16u8)in0, 128);  \
-    in1 = (RTYPE)__msa_xori_b((v16u8)in1, 128);  \
-}
+#define XORI_B2_128(RTYPE, in0, in1)            \
+  {                                             \
+    in0 = (RTYPE)__msa_xori_b((v16u8)in0, 128); \
+    in1 = (RTYPE)__msa_xori_b((v16u8)in1, 128); \
+  }
 #define XORI_B2_128_UB(...) XORI_B2_128(v16u8, __VA_ARGS__)
 #define XORI_B2_128_SB(...) XORI_B2_128(v16i8, __VA_ARGS__)
 
-#define XORI_B3_128(RTYPE, in0, in1, in2)        \
-{                                                \
-    XORI_B2_128(RTYPE, in0, in1);                \
-    in2 = (RTYPE)__msa_xori_b((v16u8)in2, 128);  \
-}
+#define XORI_B3_128(RTYPE, in0, in1, in2)       \
+  {                                             \
+    XORI_B2_128(RTYPE, in0, in1);               \
+    in2 = (RTYPE)__msa_xori_b((v16u8)in2, 128); \
+  }
 #define XORI_B3_128_SB(...) XORI_B3_128(v16i8, __VA_ARGS__)
 
-#define XORI_B4_128(RTYPE, in0, in1, in2, in3)  \
-{                                               \
-    XORI_B2_128(RTYPE, in0, in1);               \
-    XORI_B2_128(RTYPE, in2, in3);               \
-}
+#define XORI_B4_128(RTYPE, in0, in1, in2, in3) \
+  {                                            \
+    XORI_B2_128(RTYPE, in0, in1);              \
+    XORI_B2_128(RTYPE, in2, in3);              \
+  }
 #define XORI_B4_128_UB(...) XORI_B4_128(v16u8, __VA_ARGS__)
 #define XORI_B4_128_SB(...) XORI_B4_128(v16i8, __VA_ARGS__)
 
-#define XORI_B5_128(RTYPE, in0, in1, in2, in3, in4)  \
-{                                                    \
-    XORI_B3_128(RTYPE, in0, in1, in2);               \
-    XORI_B2_128(RTYPE, in3, in4);                    \
-}
+#define XORI_B5_128(RTYPE, in0, in1, in2, in3, in4) \
+  {                                                 \
+    XORI_B3_128(RTYPE, in0, in1, in2);              \
+    XORI_B2_128(RTYPE, in3, in4);                   \
+  }
 #define XORI_B5_128_SB(...) XORI_B5_128(v16i8, __VA_ARGS__)
 
-#define XORI_B8_128(RTYPE, in0, in1, in2, in3, in4, in5, in6, in7)  \
-{                                                                   \
-    XORI_B4_128(RTYPE, in0, in1, in2, in3);                         \
-    XORI_B4_128(RTYPE, in4, in5, in6, in7);                         \
-}
+#define XORI_B8_128(RTYPE, in0, in1, in2, in3, in4, in5, in6, in7) \
+  {                                                                \
+    XORI_B4_128(RTYPE, in0, in1, in2, in3);                        \
+    XORI_B4_128(RTYPE, in4, in5, in6, in7);                        \
+  }
 #define XORI_B8_128_SB(...) XORI_B8_128(v16i8, __VA_ARGS__)
 
 /* Description : Shift left all elements of vector (generic for all data types)
@@ -1360,13 +1338,13 @@
    Details     : Each element of vector 'in0' is left shifted by 'shift' and
                  the result is written in-place.
 */
-#define SLLI_4V(in0, in1, in2, in3, shift)  \
-{                                           \
-    in0 = in0 << shift;                     \
-    in1 = in1 << shift;                     \
-    in2 = in2 << shift;                     \
-    in3 = in3 << shift;                     \
-}
+#define SLLI_4V(in0, in1, in2, in3, shift) \
+  {                                        \
+    in0 = in0 << shift;                    \
+    in1 = in1 << shift;                    \
+    in2 = in2 << shift;                    \
+    in3 = in3 << shift;                    \
+  }
 
 /* Description : Arithmetic shift right all elements of vector
                  (generic for all data types)
@@ -1376,13 +1354,13 @@
    Details     : Each element of vector 'in0' is right shifted by 'shift' and
                  the result is written in-place. 'shift' is a GP variable.
 */
-#define SRA_4V(in0, in1, in2, in3, shift)  \
-{                                          \
-    in0 = in0 >> shift;                    \
-    in1 = in1 >> shift;                    \
-    in2 = in2 >> shift;                    \
-    in3 = in3 >> shift;                    \
-}
+#define SRA_4V(in0, in1, in2, in3, shift) \
+  {                                       \
+    in0 = in0 >> shift;                   \
+    in1 = in1 >> shift;                   \
+    in2 = in2 >> shift;                   \
+    in3 = in3 >> shift;                   \
+  }
 
 /* Description : Shift right arithmetic rounded words
    Arguments   : Inputs  - in0, in1, shift
@@ -1394,17 +1372,17 @@
                  rounding and the result is written in-place.
                  'shift' is a vector.
 */
-#define SRAR_W2(RTYPE, in0, in1, shift)                   \
-{                                                         \
-    in0 = (RTYPE)__msa_srar_w((v4i32)in0, (v4i32)shift);  \
-    in1 = (RTYPE)__msa_srar_w((v4i32)in1, (v4i32)shift);  \
-}
+#define SRAR_W2(RTYPE, in0, in1, shift)                  \
+  {                                                      \
+    in0 = (RTYPE)__msa_srar_w((v4i32)in0, (v4i32)shift); \
+    in1 = (RTYPE)__msa_srar_w((v4i32)in1, (v4i32)shift); \
+  }
 
-#define SRAR_W4(RTYPE, in0, in1, in2, in3, shift)  \
-{                                                  \
-    SRAR_W2(RTYPE, in0, in1, shift);               \
-    SRAR_W2(RTYPE, in2, in3, shift);               \
-}
+#define SRAR_W4(RTYPE, in0, in1, in2, in3, shift) \
+  {                                               \
+    SRAR_W2(RTYPE, in0, in1, shift);              \
+    SRAR_W2(RTYPE, in2, in3, shift);              \
+  }
 #define SRAR_W4_SW(...) SRAR_W4(v4i32, __VA_ARGS__)
 
 /* Description : Shift right arithmetic rounded (immediate)
@@ -1416,33 +1394,33 @@
                  shifted value for rounding and the result is written in-place.
                  'shift' is an immediate value.
 */
-#define SRARI_H2(RTYPE, in0, in1, shift)            \
-{                                                   \
-    in0 = (RTYPE)__msa_srari_h((v8i16)in0, shift);  \
-    in1 = (RTYPE)__msa_srari_h((v8i16)in1, shift);  \
-}
+#define SRARI_H2(RTYPE, in0, in1, shift)           \
+  {                                                \
+    in0 = (RTYPE)__msa_srari_h((v8i16)in0, shift); \
+    in1 = (RTYPE)__msa_srari_h((v8i16)in1, shift); \
+  }
 #define SRARI_H2_UH(...) SRARI_H2(v8u16, __VA_ARGS__)
 #define SRARI_H2_SH(...) SRARI_H2(v8i16, __VA_ARGS__)
 
-#define SRARI_H4(RTYPE, in0, in1, in2, in3, shift)  \
-{                                                   \
-    SRARI_H2(RTYPE, in0, in1, shift);               \
-    SRARI_H2(RTYPE, in2, in3, shift);               \
-}
+#define SRARI_H4(RTYPE, in0, in1, in2, in3, shift) \
+  {                                                \
+    SRARI_H2(RTYPE, in0, in1, shift);              \
+    SRARI_H2(RTYPE, in2, in3, shift);              \
+  }
 #define SRARI_H4_UH(...) SRARI_H4(v8u16, __VA_ARGS__)
 #define SRARI_H4_SH(...) SRARI_H4(v8i16, __VA_ARGS__)
 
-#define SRARI_W2(RTYPE, in0, in1, shift)            \
-{                                                   \
-    in0 = (RTYPE)__msa_srari_w((v4i32)in0, shift);  \
-    in1 = (RTYPE)__msa_srari_w((v4i32)in1, shift);  \
-}
+#define SRARI_W2(RTYPE, in0, in1, shift)           \
+  {                                                \
+    in0 = (RTYPE)__msa_srari_w((v4i32)in0, shift); \
+    in1 = (RTYPE)__msa_srari_w((v4i32)in1, shift); \
+  }
 
-#define SRARI_W4(RTYPE, in0, in1, in2, in3, shift)  \
-{                                                   \
-    SRARI_W2(RTYPE, in0, in1, shift);               \
-    SRARI_W2(RTYPE, in2, in3, shift);               \
-}
+#define SRARI_W4(RTYPE, in0, in1, in2, in3, shift) \
+  {                                                \
+    SRARI_W2(RTYPE, in0, in1, shift);              \
+    SRARI_W2(RTYPE, in2, in3, shift);              \
+  }
 #define SRARI_W4_SW(...) SRARI_W4(v4i32, __VA_ARGS__)
 
 /* Description : Multiplication of pairs of vectors
@@ -1451,17 +1429,16 @@
    Details     : Each element from 'in0' is multiplied with elements from 'in1'
                  and the result is written to 'out0'
 */
-#define MUL2(in0, in1, in2, in3, out0, out1)  \
-{                                             \
-    out0 = in0 * in1;                         \
-    out1 = in2 * in3;                         \
-}
-#define MUL4(in0, in1, in2, in3, in4, in5, in6, in7,  \
-             out0, out1, out2, out3)                  \
-{                                                     \
-    MUL2(in0, in1, in2, in3, out0, out1);             \
-    MUL2(in4, in5, in6, in7, out2, out3);             \
-}
+#define MUL2(in0, in1, in2, in3, out0, out1) \
+  {                                          \
+    out0 = in0 * in1;                        \
+    out1 = in2 * in3;                        \
+  }
+#define MUL4(in0, in1, in2, in3, in4, in5, in6, in7, out0, out1, out2, out3) \
+  {                                                                          \
+    MUL2(in0, in1, in2, in3, out0, out1);                                    \
+    MUL2(in4, in5, in6, in7, out2, out3);                                    \
+  }
 
 /* Description : Addition of 2 pairs of vectors
    Arguments   : Inputs  - in0, in1, in2, in3
@@ -1469,17 +1446,16 @@
    Details     : Each element in 'in0' is added to 'in1' and result is written
                  to 'out0'.
 */
-#define ADD2(in0, in1, in2, in3, out0, out1)  \
-{                                             \
-    out0 = in0 + in1;                         \
-    out1 = in2 + in3;                         \
-}
-#define ADD4(in0, in1, in2, in3, in4, in5, in6, in7,  \
-             out0, out1, out2, out3)                  \
-{                                                     \
-    ADD2(in0, in1, in2, in3, out0, out1);             \
-    ADD2(in4, in5, in6, in7, out2, out3);             \
-}
+#define ADD2(in0, in1, in2, in3, out0, out1) \
+  {                                          \
+    out0 = in0 + in1;                        \
+    out1 = in2 + in3;                        \
+  }
+#define ADD4(in0, in1, in2, in3, in4, in5, in6, in7, out0, out1, out2, out3) \
+  {                                                                          \
+    ADD2(in0, in1, in2, in3, out0, out1);                                    \
+    ADD2(in4, in5, in6, in7, out2, out3);                                    \
+  }
 
 /* Description : Subtraction of 2 pairs of vectors
    Arguments   : Inputs  - in0, in1, in2, in3
@@ -1487,19 +1463,18 @@
    Details     : Each element in 'in1' is subtracted from 'in0' and result is
                  written to 'out0'.
 */
-#define SUB2(in0, in1, in2, in3, out0, out1)  \
-{                                             \
-    out0 = in0 - in1;                         \
-    out1 = in2 - in3;                         \
-}
-#define SUB4(in0, in1, in2, in3, in4, in5, in6, in7,  \
-             out0, out1, out2, out3)                  \
-{                                                     \
-    out0 = in0 - in1;                                 \
-    out1 = in2 - in3;                                 \
-    out2 = in4 - in5;                                 \
-    out3 = in6 - in7;                                 \
-}
+#define SUB2(in0, in1, in2, in3, out0, out1) \
+  {                                          \
+    out0 = in0 - in1;                        \
+    out1 = in2 - in3;                        \
+  }
+#define SUB4(in0, in1, in2, in3, in4, in5, in6, in7, out0, out1, out2, out3) \
+  {                                                                          \
+    out0 = in0 - in1;                                                        \
+    out1 = in2 - in3;                                                        \
+    out2 = in4 - in5;                                                        \
+    out3 = in6 - in7;                                                        \
+  }
 
 /* Description : Sign extend halfword elements from right half of the vector
    Arguments   : Input  - in    (halfword vector)
@@ -1509,13 +1484,13 @@
                  extracted and interleaved with same vector 'in0' to generate
                  4 word elements keeping sign intact
 */
-#define UNPCK_R_SH_SW(in, out)                     \
-{                                                  \
-    v8i16 sign_m;                                  \
-                                                   \
-    sign_m = __msa_clti_s_h((v8i16)in, 0);         \
-    out = (v4i32)__msa_ilvr_h(sign_m, (v8i16)in);  \
-}
+#define UNPCK_R_SH_SW(in, out)                    \
+  {                                               \
+    v8i16 sign_m;                                 \
+                                                  \
+    sign_m = __msa_clti_s_h((v8i16)in, 0);        \
+    out = (v4i32)__msa_ilvr_h(sign_m, (v8i16)in); \
+  }
 
 /* Description : Zero extend unsigned byte elements to halfword elements
    Arguments   : Input   - in          (unsigned byte vector)
@@ -1524,12 +1499,12 @@
    Details     : Zero extended right half of vector is returned in 'out0'
                  Zero extended left half of vector is returned in 'out1'
 */
-#define UNPCK_UB_SH(in, out0, out1)       \
-{                                         \
-    v16i8 zero_m = { 0 };                 \
-                                          \
-    ILVRL_B2_SH(zero_m, in, out0, out1);  \
-}
+#define UNPCK_UB_SH(in, out0, out1)      \
+  {                                      \
+    v16i8 zero_m = { 0 };                \
+                                         \
+    ILVRL_B2_SH(zero_m, in, out0, out1); \
+  }
 
 /* Description : Sign extend halfword elements from input vector and return
                  the result in pair of vectors
@@ -1542,13 +1517,13 @@
                  Then interleaved left with same vector 'in0' to
                  generate 4 signed word elements in 'out1'
 */
-#define UNPCK_SH_SW(in, out0, out1)        \
-{                                          \
-    v8i16 tmp_m;                           \
-                                           \
-    tmp_m = __msa_clti_s_h((v8i16)in, 0);  \
-    ILVRL_H2_SW(tmp_m, in, out0, out1);    \
-}
+#define UNPCK_SH_SW(in, out0, out1)       \
+  {                                       \
+    v8i16 tmp_m;                          \
+                                          \
+    tmp_m = __msa_clti_s_h((v8i16)in, 0); \
+    ILVRL_H2_SW(tmp_m, in, out0, out1);   \
+  }
 
 /* Description : Butterfly of 4 input vectors
    Arguments   : Inputs  - in0, in1, in2, in3
@@ -1555,14 +1530,14 @@
                  Outputs - out0, out1, out2, out3
    Details     : Butterfly operation
 */
-#define BUTTERFLY_4(in0, in1, in2, in3, out0, out1, out2, out3)  \
-{                                                                \
-    out0 = in0 + in3;                                            \
-    out1 = in1 + in2;                                            \
-                                                                 \
-    out2 = in1 - in2;                                            \
-    out3 = in0 - in3;                                            \
-}
+#define BUTTERFLY_4(in0, in1, in2, in3, out0, out1, out2, out3) \
+  {                                                             \
+    out0 = in0 + in3;                                           \
+    out1 = in1 + in2;                                           \
+                                                                \
+    out2 = in1 - in2;                                           \
+    out3 = in0 - in3;                                           \
+  }
 
 /* Description : Transpose input 8x8 byte block
    Arguments   : Inputs  - in0, in1, in2, in3, in4, in5, in6, in7
@@ -1569,21 +1544,21 @@
                  Outputs - out0, out1, out2, out3, out4, out5, out6, out7
                  Return Type - as per RTYPE
 */
-#define TRANSPOSE8x8_UB(RTYPE, in0, in1, in2, in3, in4, in5, in6, in7,   \
-                        out0, out1, out2, out3, out4, out5, out6, out7)  \
-{                                                                        \
-    v16i8 tmp0_m, tmp1_m, tmp2_m, tmp3_m;                                \
-    v16i8 tmp4_m, tmp5_m, tmp6_m, tmp7_m;                                \
-                                                                         \
-    ILVR_B4_SB(in2, in0, in3, in1, in6, in4, in7, in5,                   \
-               tmp0_m, tmp1_m, tmp2_m, tmp3_m);                          \
-    ILVRL_B2_SB(tmp1_m, tmp0_m, tmp4_m, tmp5_m);                         \
-    ILVRL_B2_SB(tmp3_m, tmp2_m, tmp6_m, tmp7_m);                         \
-    ILVRL_W2(RTYPE, tmp6_m, tmp4_m, out0, out2);                         \
-    ILVRL_W2(RTYPE, tmp7_m, tmp5_m, out4, out6);                         \
-    SLDI_B2_0(RTYPE, out0, out2, out1, out3, 8);                         \
-    SLDI_B2_0(RTYPE, out4, out6, out5, out7, 8);                         \
-}
+#define TRANSPOSE8x8_UB(RTYPE, in0, in1, in2, in3, in4, in5, in6, in7, out0,   \
+                        out1, out2, out3, out4, out5, out6, out7)              \
+  {                                                                            \
+    v16i8 tmp0_m, tmp1_m, tmp2_m, tmp3_m;                                      \
+    v16i8 tmp4_m, tmp5_m, tmp6_m, tmp7_m;                                      \
+                                                                               \
+    ILVR_B4_SB(in2, in0, in3, in1, in6, in4, in7, in5, tmp0_m, tmp1_m, tmp2_m, \
+               tmp3_m);                                                        \
+    ILVRL_B2_SB(tmp1_m, tmp0_m, tmp4_m, tmp5_m);                               \
+    ILVRL_B2_SB(tmp3_m, tmp2_m, tmp6_m, tmp7_m);                               \
+    ILVRL_W2(RTYPE, tmp6_m, tmp4_m, out0, out2);                               \
+    ILVRL_W2(RTYPE, tmp7_m, tmp5_m, out4, out6);                               \
+    SLDI_B2_0(RTYPE, out0, out2, out1, out3, 8);                               \
+    SLDI_B2_0(RTYPE, out4, out6, out5, out7, 8);                               \
+  }
 #define TRANSPOSE8x8_UB_UB(...) TRANSPOSE8x8_UB(v16u8, __VA_ARGS__)
 
 /* Description : Transpose 16x4 block into 4x16 with byte elements in vectors
@@ -1592,33 +1567,33 @@
                  Outputs - out0, out1, out2, out3
                  Return Type - unsigned byte
 */
-#define TRANSPOSE16x4_UB_UB(in0, in1, in2, in3, in4, in5, in6, in7,        \
-                            in8, in9, in10, in11, in12, in13, in14, in15,  \
-                            out0, out1, out2, out3)                        \
-{                                                                          \
-    v2i64 tmp0_m, tmp1_m, tmp2_m, tmp3_m;                                  \
-                                                                           \
-    ILVEV_W2_SD(in0, in4, in8, in12, tmp0_m, tmp1_m);                      \
-    out1 = (v16u8)__msa_ilvev_d(tmp1_m, tmp0_m);                           \
-                                                                           \
-    ILVEV_W2_SD(in1, in5, in9, in13, tmp0_m, tmp1_m);                      \
-    out3 = (v16u8)__msa_ilvev_d(tmp1_m, tmp0_m);                           \
-                                                                           \
-    ILVEV_W2_SD(in2, in6, in10, in14, tmp0_m, tmp1_m);                     \
-                                                                           \
-    tmp2_m = __msa_ilvev_d(tmp1_m, tmp0_m);                                \
-    ILVEV_W2_SD(in3, in7, in11, in15, tmp0_m, tmp1_m);                     \
-                                                                           \
-    tmp3_m = __msa_ilvev_d(tmp1_m, tmp0_m);                                \
-    ILVEV_B2_SD(out1, out3, tmp2_m, tmp3_m, tmp0_m, tmp1_m);               \
-    out0 = (v16u8)__msa_ilvev_h((v8i16)tmp1_m, (v8i16)tmp0_m);             \
-    out2 = (v16u8)__msa_ilvod_h((v8i16)tmp1_m, (v8i16)tmp0_m);             \
-                                                                           \
-    tmp0_m = (v2i64)__msa_ilvod_b((v16i8)out3, (v16i8)out1);               \
-    tmp1_m = (v2i64)__msa_ilvod_b((v16i8)tmp3_m, (v16i8)tmp2_m);           \
-    out1 = (v16u8)__msa_ilvev_h((v8i16)tmp1_m, (v8i16)tmp0_m);             \
-    out3 = (v16u8)__msa_ilvod_h((v8i16)tmp1_m, (v8i16)tmp0_m);             \
-}
+#define TRANSPOSE16x4_UB_UB(in0, in1, in2, in3, in4, in5, in6, in7, in8, in9, \
+                            in10, in11, in12, in13, in14, in15, out0, out1,   \
+                            out2, out3)                                       \
+  {                                                                           \
+    v2i64 tmp0_m, tmp1_m, tmp2_m, tmp3_m;                                     \
+                                                                              \
+    ILVEV_W2_SD(in0, in4, in8, in12, tmp0_m, tmp1_m);                         \
+    out1 = (v16u8)__msa_ilvev_d(tmp1_m, tmp0_m);                              \
+                                                                              \
+    ILVEV_W2_SD(in1, in5, in9, in13, tmp0_m, tmp1_m);                         \
+    out3 = (v16u8)__msa_ilvev_d(tmp1_m, tmp0_m);                              \
+                                                                              \
+    ILVEV_W2_SD(in2, in6, in10, in14, tmp0_m, tmp1_m);                        \
+                                                                              \
+    tmp2_m = __msa_ilvev_d(tmp1_m, tmp0_m);                                   \
+    ILVEV_W2_SD(in3, in7, in11, in15, tmp0_m, tmp1_m);                        \
+                                                                              \
+    tmp3_m = __msa_ilvev_d(tmp1_m, tmp0_m);                                   \
+    ILVEV_B2_SD(out1, out3, tmp2_m, tmp3_m, tmp0_m, tmp1_m);                  \
+    out0 = (v16u8)__msa_ilvev_h((v8i16)tmp1_m, (v8i16)tmp0_m);                \
+    out2 = (v16u8)__msa_ilvod_h((v8i16)tmp1_m, (v8i16)tmp0_m);                \
+                                                                              \
+    tmp0_m = (v2i64)__msa_ilvod_b((v16i8)out3, (v16i8)out1);                  \
+    tmp1_m = (v2i64)__msa_ilvod_b((v16i8)tmp3_m, (v16i8)tmp2_m);              \
+    out1 = (v16u8)__msa_ilvev_h((v8i16)tmp1_m, (v8i16)tmp0_m);                \
+    out3 = (v16u8)__msa_ilvod_h((v8i16)tmp1_m, (v8i16)tmp0_m);                \
+  }
 
 /* Description : Transpose 16x8 block into 8x16 with byte elements in vectors
    Arguments   : Inputs  - in0, in1, in2, in3, in4, in5, in6, in7,
@@ -1626,47 +1601,47 @@
                  Outputs - out0, out1, out2, out3, out4, out5, out6, out7
                  Return Type - unsigned byte
 */
-#define TRANSPOSE16x8_UB_UB(in0, in1, in2, in3, in4, in5, in6, in7,          \
-                            in8, in9, in10, in11, in12, in13, in14, in15,    \
-                            out0, out1, out2, out3, out4, out5, out6, out7)  \
-{                                                                            \
-    v16u8 tmp0_m, tmp1_m, tmp2_m, tmp3_m;                                    \
-    v16u8 tmp4_m, tmp5_m, tmp6_m, tmp7_m;                                    \
-                                                                             \
-    ILVEV_D2_UB(in0, in8, in1, in9, out7, out6);                             \
-    ILVEV_D2_UB(in2, in10, in3, in11, out5, out4);                           \
-    ILVEV_D2_UB(in4, in12, in5, in13, out3, out2);                           \
-    ILVEV_D2_UB(in6, in14, in7, in15, out1, out0);                           \
-                                                                             \
-    tmp0_m = (v16u8)__msa_ilvev_b((v16i8)out6, (v16i8)out7);                 \
-    tmp4_m = (v16u8)__msa_ilvod_b((v16i8)out6, (v16i8)out7);                 \
-    tmp1_m = (v16u8)__msa_ilvev_b((v16i8)out4, (v16i8)out5);                 \
-    tmp5_m = (v16u8)__msa_ilvod_b((v16i8)out4, (v16i8)out5);                 \
-    out5 = (v16u8)__msa_ilvev_b((v16i8)out2, (v16i8)out3);                   \
-    tmp6_m = (v16u8)__msa_ilvod_b((v16i8)out2, (v16i8)out3);                 \
-    out7 = (v16u8)__msa_ilvev_b((v16i8)out0, (v16i8)out1);                   \
-    tmp7_m = (v16u8)__msa_ilvod_b((v16i8)out0, (v16i8)out1);                 \
-                                                                             \
-    ILVEV_H2_UB(tmp0_m, tmp1_m, out5, out7, tmp2_m, tmp3_m);                 \
-    out0 = (v16u8)__msa_ilvev_w((v4i32)tmp3_m, (v4i32)tmp2_m);               \
-    out4 = (v16u8)__msa_ilvod_w((v4i32)tmp3_m, (v4i32)tmp2_m);               \
-                                                                             \
-    tmp2_m = (v16u8)__msa_ilvod_h((v8i16)tmp1_m, (v8i16)tmp0_m);             \
-    tmp3_m = (v16u8)__msa_ilvod_h((v8i16)out7, (v8i16)out5);                 \
-    out2 = (v16u8)__msa_ilvev_w((v4i32)tmp3_m, (v4i32)tmp2_m);               \
-    out6 = (v16u8)__msa_ilvod_w((v4i32)tmp3_m, (v4i32)tmp2_m);               \
-                                                                             \
-    ILVEV_H2_UB(tmp4_m, tmp5_m, tmp6_m, tmp7_m, tmp2_m, tmp3_m);             \
-    out1 = (v16u8)__msa_ilvev_w((v4i32)tmp3_m, (v4i32)tmp2_m);               \
-    out5 = (v16u8)__msa_ilvod_w((v4i32)tmp3_m, (v4i32)tmp2_m);               \
-                                                                             \
-    tmp2_m = (v16u8)__msa_ilvod_h((v8i16)tmp5_m, (v8i16)tmp4_m);             \
-    tmp2_m = (v16u8)__msa_ilvod_h((v8i16)tmp5_m, (v8i16)tmp4_m);             \
-    tmp3_m = (v16u8)__msa_ilvod_h((v8i16)tmp7_m, (v8i16)tmp6_m);             \
-    tmp3_m = (v16u8)__msa_ilvod_h((v8i16)tmp7_m, (v8i16)tmp6_m);             \
-    out3 = (v16u8)__msa_ilvev_w((v4i32)tmp3_m, (v4i32)tmp2_m);               \
-    out7 = (v16u8)__msa_ilvod_w((v4i32)tmp3_m, (v4i32)tmp2_m);               \
-}
+#define TRANSPOSE16x8_UB_UB(in0, in1, in2, in3, in4, in5, in6, in7, in8, in9, \
+                            in10, in11, in12, in13, in14, in15, out0, out1,   \
+                            out2, out3, out4, out5, out6, out7)               \
+  {                                                                           \
+    v16u8 tmp0_m, tmp1_m, tmp2_m, tmp3_m;                                     \
+    v16u8 tmp4_m, tmp5_m, tmp6_m, tmp7_m;                                     \
+                                                                              \
+    ILVEV_D2_UB(in0, in8, in1, in9, out7, out6);                              \
+    ILVEV_D2_UB(in2, in10, in3, in11, out5, out4);                            \
+    ILVEV_D2_UB(in4, in12, in5, in13, out3, out2);                            \
+    ILVEV_D2_UB(in6, in14, in7, in15, out1, out0);                            \
+                                                                              \
+    tmp0_m = (v16u8)__msa_ilvev_b((v16i8)out6, (v16i8)out7);                  \
+    tmp4_m = (v16u8)__msa_ilvod_b((v16i8)out6, (v16i8)out7);                  \
+    tmp1_m = (v16u8)__msa_ilvev_b((v16i8)out4, (v16i8)out5);                  \
+    tmp5_m = (v16u8)__msa_ilvod_b((v16i8)out4, (v16i8)out5);                  \
+    out5 = (v16u8)__msa_ilvev_b((v16i8)out2, (v16i8)out3);                    \
+    tmp6_m = (v16u8)__msa_ilvod_b((v16i8)out2, (v16i8)out3);                  \
+    out7 = (v16u8)__msa_ilvev_b((v16i8)out0, (v16i8)out1);                    \
+    tmp7_m = (v16u8)__msa_ilvod_b((v16i8)out0, (v16i8)out1);                  \
+                                                                              \
+    ILVEV_H2_UB(tmp0_m, tmp1_m, out5, out7, tmp2_m, tmp3_m);                  \
+    out0 = (v16u8)__msa_ilvev_w((v4i32)tmp3_m, (v4i32)tmp2_m);                \
+    out4 = (v16u8)__msa_ilvod_w((v4i32)tmp3_m, (v4i32)tmp2_m);                \
+                                                                              \
+    tmp2_m = (v16u8)__msa_ilvod_h((v8i16)tmp1_m, (v8i16)tmp0_m);              \
+    tmp3_m = (v16u8)__msa_ilvod_h((v8i16)out7, (v8i16)out5);                  \
+    out2 = (v16u8)__msa_ilvev_w((v4i32)tmp3_m, (v4i32)tmp2_m);                \
+    out6 = (v16u8)__msa_ilvod_w((v4i32)tmp3_m, (v4i32)tmp2_m);                \
+                                                                              \
+    ILVEV_H2_UB(tmp4_m, tmp5_m, tmp6_m, tmp7_m, tmp2_m, tmp3_m);              \
+    out1 = (v16u8)__msa_ilvev_w((v4i32)tmp3_m, (v4i32)tmp2_m);                \
+    out5 = (v16u8)__msa_ilvod_w((v4i32)tmp3_m, (v4i32)tmp2_m);                \
+                                                                              \
+    tmp2_m = (v16u8)__msa_ilvod_h((v8i16)tmp5_m, (v8i16)tmp4_m);              \
+    tmp2_m = (v16u8)__msa_ilvod_h((v8i16)tmp5_m, (v8i16)tmp4_m);              \
+    tmp3_m = (v16u8)__msa_ilvod_h((v8i16)tmp7_m, (v8i16)tmp6_m);              \
+    tmp3_m = (v16u8)__msa_ilvod_h((v8i16)tmp7_m, (v8i16)tmp6_m);              \
+    out3 = (v16u8)__msa_ilvev_w((v4i32)tmp3_m, (v4i32)tmp2_m);                \
+    out7 = (v16u8)__msa_ilvod_w((v4i32)tmp3_m, (v4i32)tmp2_m);                \
+  }
 
 /* Description : Transpose 4x4 block with half word elements in vectors
    Arguments   : Inputs  - in0, in1, in2, in3
@@ -1673,15 +1648,15 @@
                  Outputs - out0, out1, out2, out3
                  Return Type - signed halfword
 */
-#define TRANSPOSE4x4_SH_SH(in0, in1, in2, in3, out0, out1, out2, out3)  \
-{                                                                       \
-    v8i16 s0_m, s1_m;                                                   \
-                                                                        \
-    ILVR_H2_SH(in1, in0, in3, in2, s0_m, s1_m);                         \
-    ILVRL_W2_SH(s1_m, s0_m, out0, out2);                                \
-    out1 = (v8i16)__msa_ilvl_d((v2i64)out0, (v2i64)out0);               \
-    out3 = (v8i16)__msa_ilvl_d((v2i64)out0, (v2i64)out2);               \
-}
+#define TRANSPOSE4x4_SH_SH(in0, in1, in2, in3, out0, out1, out2, out3) \
+  {                                                                    \
+    v8i16 s0_m, s1_m;                                                  \
+                                                                       \
+    ILVR_H2_SH(in1, in0, in3, in2, s0_m, s1_m);                        \
+    ILVRL_W2_SH(s1_m, s0_m, out0, out2);                               \
+    out1 = (v8i16)__msa_ilvl_d((v2i64)out0, (v2i64)out0);              \
+    out3 = (v8i16)__msa_ilvl_d((v2i64)out0, (v2i64)out2);              \
+  }
 
 /* Description : Transpose 8x4 block with half word elements in vectors
    Arguments   : Inputs  - in0, in1, in2, in3, in4, in5, in6, in7
@@ -1688,15 +1663,15 @@
                  Outputs - out0, out1, out2, out3, out4, out5, out6, out7
                  Return Type - signed halfword
 */
-#define TRANSPOSE8X4_SH_SH(in0, in1, in2, in3, out0, out1, out2, out3)  \
-{                                                                       \
-    v8i16 tmp0_m, tmp1_m, tmp2_m, tmp3_m;                               \
-                                                                        \
-    ILVR_H2_SH(in1, in0, in3, in2, tmp0_m, tmp1_m);                     \
-    ILVL_H2_SH(in1, in0, in3, in2, tmp2_m, tmp3_m);                     \
-    ILVR_W2_SH(tmp1_m, tmp0_m, tmp3_m, tmp2_m, out0, out2);             \
-    ILVL_W2_SH(tmp1_m, tmp0_m, tmp3_m, tmp2_m, out1, out3);             \
-}
+#define TRANSPOSE8X4_SH_SH(in0, in1, in2, in3, out0, out1, out2, out3) \
+  {                                                                    \
+    v8i16 tmp0_m, tmp1_m, tmp2_m, tmp3_m;                              \
+                                                                       \
+    ILVR_H2_SH(in1, in0, in3, in2, tmp0_m, tmp1_m);                    \
+    ILVL_H2_SH(in1, in0, in3, in2, tmp2_m, tmp3_m);                    \
+    ILVR_W2_SH(tmp1_m, tmp0_m, tmp3_m, tmp2_m, out0, out2);            \
+    ILVL_W2_SH(tmp1_m, tmp0_m, tmp3_m, tmp2_m, out1, out3);            \
+  }
 
 /* Description : Transpose 4x4 block with word elements in vectors
    Arguments   : Inputs  - in0, in1, in2, in3
@@ -1703,18 +1678,18 @@
                  Outputs - out0, out1, out2, out3
                  Return Type - signed word
 */
-#define TRANSPOSE4x4_SW_SW(in0, in1, in2, in3, out0, out1, out2, out3)  \
-{                                                                       \
-    v4i32 s0_m, s1_m, s2_m, s3_m;                                       \
-                                                                        \
-    ILVRL_W2_SW(in1, in0, s0_m, s1_m);                                  \
-    ILVRL_W2_SW(in3, in2, s2_m, s3_m);                                  \
-                                                                        \
-    out0 = (v4i32)__msa_ilvr_d((v2i64)s2_m, (v2i64)s0_m);               \
-    out1 = (v4i32)__msa_ilvl_d((v2i64)s2_m, (v2i64)s0_m);               \
-    out2 = (v4i32)__msa_ilvr_d((v2i64)s3_m, (v2i64)s1_m);               \
-    out3 = (v4i32)__msa_ilvl_d((v2i64)s3_m, (v2i64)s1_m);               \
-}
+#define TRANSPOSE4x4_SW_SW(in0, in1, in2, in3, out0, out1, out2, out3) \
+  {                                                                    \
+    v4i32 s0_m, s1_m, s2_m, s3_m;                                      \
+                                                                       \
+    ILVRL_W2_SW(in1, in0, s0_m, s1_m);                                 \
+    ILVRL_W2_SW(in3, in2, s2_m, s3_m);                                 \
+                                                                       \
+    out0 = (v4i32)__msa_ilvr_d((v2i64)s2_m, (v2i64)s0_m);              \
+    out1 = (v4i32)__msa_ilvl_d((v2i64)s2_m, (v2i64)s0_m);              \
+    out2 = (v4i32)__msa_ilvr_d((v2i64)s3_m, (v2i64)s1_m);              \
+    out3 = (v4i32)__msa_ilvl_d((v2i64)s3_m, (v2i64)s1_m);              \
+  }
 
 /* Description : Dot product and addition of 3 signed halfword input vectors
    Arguments   : Inputs - in0, in1, in2, coeff0, coeff1, coeff2
@@ -1726,18 +1701,18 @@
                  Addition of all the 3 vector results
                  out0_m = (in0 * coeff0) + (in1 * coeff1) + (in2 * coeff2)
 */
-#define DPADD_SH3_SH(in0, in1, in2, coeff0, coeff1, coeff2)       \
-({                                                                \
-    v8i16 tmp1_m;                                                 \
-    v8i16 out0_m;                                                 \
-                                                                  \
-    out0_m = __msa_dotp_s_h((v16i8)in0, (v16i8)coeff0);           \
-    out0_m = __msa_dpadd_s_h(out0_m, (v16i8)in1, (v16i8)coeff1);  \
-    tmp1_m = __msa_dotp_s_h((v16i8)in2, (v16i8)coeff2);           \
-    out0_m = __msa_adds_s_h(out0_m, tmp1_m);                      \
-                                                                  \
-    out0_m;                                                       \
-})
+#define DPADD_SH3_SH(in0, in1, in2, coeff0, coeff1, coeff2)      \
+  ({                                                             \
+    v8i16 tmp1_m;                                                \
+    v8i16 out0_m;                                                \
+                                                                 \
+    out0_m = __msa_dotp_s_h((v16i8)in0, (v16i8)coeff0);          \
+    out0_m = __msa_dpadd_s_h(out0_m, (v16i8)in1, (v16i8)coeff1); \
+    tmp1_m = __msa_dotp_s_h((v16i8)in2, (v16i8)coeff2);          \
+    out0_m = __msa_adds_s_h(out0_m, tmp1_m);                     \
+                                                                 \
+    out0_m;                                                      \
+  })
 
 /* Description : Pack even elements of input vectors & xor with 128
    Arguments   : Inputs - in0, in1
@@ -1747,37 +1722,37 @@
                  together in one vector and the resulting vector is xor'ed with
                  128 to shift the range from signed to unsigned byte
 */
-#define PCKEV_XORI128_UB(in0, in1)                         \
-({                                                         \
-    v16u8 out_m;                                           \
-    out_m = (v16u8)__msa_pckev_b((v16i8)in1, (v16i8)in0);  \
-    out_m = (v16u8)__msa_xori_b((v16u8)out_m, 128);        \
-    out_m;                                                 \
-})
+#define PCKEV_XORI128_UB(in0, in1)                        \
+  ({                                                      \
+    v16u8 out_m;                                          \
+    out_m = (v16u8)__msa_pckev_b((v16i8)in1, (v16i8)in0); \
+    out_m = (v16u8)__msa_xori_b((v16u8)out_m, 128);       \
+    out_m;                                                \
+  })
 
 /* Description : Pack even byte elements and store byte vector in destination
                  memory
    Arguments   : Inputs - in0, in1, pdst
 */
-#define PCKEV_ST_SB(in0, in1, pdst)                 \
-{                                                   \
-    v16i8 tmp_m;                                    \
-    tmp_m = __msa_pckev_b((v16i8)in1, (v16i8)in0);  \
-    ST_SB(tmp_m, (pdst));                           \
-}
+#define PCKEV_ST_SB(in0, in1, pdst)                \
+  {                                                \
+    v16i8 tmp_m;                                   \
+    tmp_m = __msa_pckev_b((v16i8)in1, (v16i8)in0); \
+    ST_SB(tmp_m, (pdst));                          \
+  }
 
 /* Description : Horizontal 2 tap filter kernel code
    Arguments   : Inputs - in0, in1, mask, coeff, shift
 */
-#define HORIZ_2TAP_FILT_UH(in0, in1, mask, coeff, shift)         \
-({                                                               \
-    v16i8 tmp0_m;                                                \
-    v8u16 tmp1_m;                                                \
-                                                                 \
-    tmp0_m = __msa_vshf_b((v16i8)mask, (v16i8)in1, (v16i8)in0);  \
-    tmp1_m = __msa_dotp_u_h((v16u8)tmp0_m, (v16u8)coeff);        \
-    tmp1_m = (v8u16)__msa_srari_h((v8i16)tmp1_m, shift);         \
-                                                                 \
-    tmp1_m;                                                      \
-})
-#endif  /* VP8_COMMON_MIPS_MSA_VP8_MACROS_MSA_H_ */
+#define HORIZ_2TAP_FILT_UH(in0, in1, mask, coeff, shift)        \
+  ({                                                            \
+    v16i8 tmp0_m;                                               \
+    v8u16 tmp1_m;                                               \
+                                                                \
+    tmp0_m = __msa_vshf_b((v16i8)mask, (v16i8)in1, (v16i8)in0); \
+    tmp1_m = __msa_dotp_u_h((v16u8)tmp0_m, (v16u8)coeff);       \
+    tmp1_m = (v8u16)__msa_srari_h((v8i16)tmp1_m, shift);        \
+                                                                \
+    tmp1_m;                                                     \
+  })
+#endif /* VP8_COMMON_MIPS_MSA_VP8_MACROS_MSA_H_ */
--- a/vp8/common/modecont.c
+++ b/vp8/common/modecont.c
@@ -8,33 +8,31 @@
  *  be found in the AUTHORS file in the root of the source tree.
  */
 
-
 #include "entropy.h"
 
-const int vp8_mode_contexts[6][4] =
-{
-    {
-        /* 0 */
-        7,     1,     1,   143,
-    },
-    {
-        /* 1 */
-        14,    18,    14,   107,
-    },
-    {
-        /* 2 */
-        135,    64,    57,    68,
-    },
-    {
-        /* 3 */
-        60,    56,   128,    65,
-    },
-    {
-        /* 4 */
-        159,   134,   128,    34,
-    },
-    {
-        /* 5 */
-        234,   188,   128,    28,
-    },
+const int vp8_mode_contexts[6][4] = {
+  {
+      /* 0 */
+      7, 1, 1, 143,
+  },
+  {
+      /* 1 */
+      14, 18, 14, 107,
+  },
+  {
+      /* 2 */
+      135, 64, 57, 68,
+  },
+  {
+      /* 3 */
+      60, 56, 128, 65,
+  },
+  {
+      /* 4 */
+      159, 134, 128, 34,
+  },
+  {
+      /* 5 */
+      234, 188, 128, 28,
+  },
 };
--- a/vp8/common/modecont.h
+++ b/vp8/common/modecont.h
@@ -8,7 +8,6 @@
  *  be found in the AUTHORS file in the root of the source tree.
  */
 
-
 #ifndef VP8_COMMON_MODECONT_H_
 #define VP8_COMMON_MODECONT_H_
 
--- a/vp8/common/mv.h
+++ b/vp8/common/mv.h
@@ -8,7 +8,6 @@
  *  be found in the AUTHORS file in the root of the source tree.
  */
 
-
 #ifndef VP8_COMMON_MV_H_
 #define VP8_COMMON_MV_H_
 #include "vpx/vpx_integer.h"
@@ -17,17 +16,15 @@
 extern "C" {
 #endif
 
-typedef struct
-{
-    short row;
-    short col;
+typedef struct {
+  short row;
+  short col;
 } MV;
 
-typedef union int_mv
-{
-    uint32_t  as_int;
-    MV        as_mv;
-} int_mv;        /* facilitates faster equality tests and copies */
+typedef union int_mv {
+  uint32_t as_int;
+  MV as_mv;
+} int_mv; /* facilitates faster equality tests and copies */
 
 #ifdef __cplusplus
 }  // extern "C"
--- a/vp8/common/onyx.h
+++ b/vp8/common/onyx.h
@@ -8,13 +8,11 @@
  *  be found in the AUTHORS file in the root of the source tree.
  */
 
-
 #ifndef VP8_COMMON_ONYX_H_
 #define VP8_COMMON_ONYX_H_
 
 #ifdef __cplusplus
-extern "C"
-{
+extern "C" {
 #endif
 
 #include "vpx_config.h"
@@ -24,256 +22,258 @@
 #include "vpx_scale/yv12config.h"
 #include "ppflags.h"
 
-    struct VP8_COMP;
+struct VP8_COMP;
 
-    /* Create/destroy static data structures. */
+/* Create/destroy static data structures. */
 
-    typedef enum
-    {
-        NORMAL      = 0,
-        FOURFIVE    = 1,
-        THREEFIVE   = 2,
-        ONETWO      = 3
+typedef enum {
+  NORMAL = 0,
+  FOURFIVE = 1,
+  THREEFIVE = 2,
+  ONETWO = 3
+} VPX_SCALING;
 
-    } VPX_SCALING;
+typedef enum {
+  USAGE_LOCAL_FILE_PLAYBACK = 0x0,
+  USAGE_STREAM_FROM_SERVER = 0x1,
+  USAGE_CONSTRAINED_QUALITY = 0x2,
+  USAGE_CONSTANT_QUALITY = 0x3
+} END_USAGE;
 
-    typedef enum
-    {
-        USAGE_LOCAL_FILE_PLAYBACK   = 0x0,
-        USAGE_STREAM_FROM_SERVER    = 0x1,
-        USAGE_CONSTRAINED_QUALITY   = 0x2,
-        USAGE_CONSTANT_QUALITY      = 0x3
-    } END_USAGE;
+typedef enum {
+  MODE_REALTIME = 0x0,
+  MODE_GOODQUALITY = 0x1,
+  MODE_BESTQUALITY = 0x2,
+  MODE_FIRSTPASS = 0x3,
+  MODE_SECONDPASS = 0x4,
+  MODE_SECONDPASS_BEST = 0x5
+} MODE;
 
+typedef enum {
+  FRAMEFLAGS_KEY = 1,
+  FRAMEFLAGS_GOLDEN = 2,
+  FRAMEFLAGS_ALTREF = 4
+} FRAMETYPE_FLAGS;
 
-    typedef enum
-    {
-        MODE_REALTIME       = 0x0,
-        MODE_GOODQUALITY    = 0x1,
-        MODE_BESTQUALITY    = 0x2,
-        MODE_FIRSTPASS      = 0x3,
-        MODE_SECONDPASS     = 0x4,
-        MODE_SECONDPASS_BEST = 0x5
-    } MODE;
-
-    typedef enum
-    {
-        FRAMEFLAGS_KEY    = 1,
-        FRAMEFLAGS_GOLDEN = 2,
-        FRAMEFLAGS_ALTREF = 4
-    } FRAMETYPE_FLAGS;
-
-
 #include <assert.h>
-    static INLINE void Scale2Ratio(int mode, int *hr, int *hs)
-    {
-        switch (mode)
-        {
-        case    NORMAL:
-            *hr = 1;
-            *hs = 1;
-            break;
-        case    FOURFIVE:
-            *hr = 4;
-            *hs = 5;
-            break;
-        case    THREEFIVE:
-            *hr = 3;
-            *hs = 5;
-            break;
-        case    ONETWO:
-            *hr = 1;
-            *hs = 2;
-            break;
-        default:
-            *hr = 1;
-            *hs = 1;
-            assert(0);
-            break;
-        }
-    }
+static INLINE void Scale2Ratio(int mode, int *hr, int *hs) {
+  switch (mode) {
+    case NORMAL:
+      *hr = 1;
+      *hs = 1;
+      break;
+    case FOURFIVE:
+      *hr = 4;
+      *hs = 5;
+      break;
+    case THREEFIVE:
+      *hr = 3;
+      *hs = 5;
+      break;
+    case ONETWO:
+      *hr = 1;
+      *hs = 2;
+      break;
+    default:
+      *hr = 1;
+      *hs = 1;
+      assert(0);
+      break;
+  }
+}
 
-    typedef struct
-    {
-        /* 4 versions of bitstream defined:
-         *   0 best quality/slowest decode, 3 lowest quality/fastest decode
-         */
-        int Version;
-        int Width;
-        int Height;
-        struct vpx_rational  timebase;
-        unsigned int target_bandwidth;    /* kilobits per second */
+typedef struct {
+  /* 4 versions of bitstream defined:
+   *   0 best quality/slowest decode, 3 lowest quality/fastest decode
+   */
+  int Version;
+  int Width;
+  int Height;
+  struct vpx_rational timebase;
+  unsigned int target_bandwidth; /* kilobits per second */
 
-        /* Parameter used for applying denoiser.
-         * For temporal denoiser: noise_sensitivity = 0 means off,
-         * noise_sensitivity = 1 means temporal denoiser on for Y channel only,
-         * noise_sensitivity = 2 means temporal denoiser on for all channels.
-         * noise_sensitivity = 3 means aggressive denoising mode.
-         * noise_sensitivity >= 4 means adaptive denoising mode.
-         * Temporal denoiser is enabled via the configuration option:
-         * CONFIG_TEMPORAL_DENOISING.
-         * For spatial denoiser: noise_sensitivity controls the amount of
-         * pre-processing blur: noise_sensitivity = 0 means off.
-         * Spatial denoiser invoked under !CONFIG_TEMPORAL_DENOISING.
-         */
-        int noise_sensitivity;
+  /* Parameter used for applying denoiser.
+   * For temporal denoiser: noise_sensitivity = 0 means off,
+   * noise_sensitivity = 1 means temporal denoiser on for Y channel only,
+   * noise_sensitivity = 2 means temporal denoiser on for all channels.
+   * noise_sensitivity = 3 means aggressive denoising mode.
+   * noise_sensitivity >= 4 means adaptive denoising mode.
+   * Temporal denoiser is enabled via the configuration option:
+   * CONFIG_TEMPORAL_DENOISING.
+   * For spatial denoiser: noise_sensitivity controls the amount of
+   * pre-processing blur: noise_sensitivity = 0 means off.
+   * Spatial denoiser invoked under !CONFIG_TEMPORAL_DENOISING.
+   */
+  int noise_sensitivity;
 
-        /* parameter used for sharpening output: recommendation 0: */
-        int Sharpness;
-        int cpu_used;
-        unsigned int rc_max_intra_bitrate_pct;
-        unsigned int screen_content_mode;
+  /* parameter used for sharpening output: recommendation 0: */
+  int Sharpness;
+  int cpu_used;
+  unsigned int rc_max_intra_bitrate_pct;
+  unsigned int screen_content_mode;
 
-        /* mode ->
-         *(0)=Realtime/Live Encoding. This mode is optimized for realtim
-         *    encoding (for example, capturing a television signal or feed
-         *    from a live camera). ( speed setting controls how fast )
-         *(1)=Good Quality Fast Encoding. The encoder balances quality with
-         *    the amount of time it takes to encode the output. ( speed
-         *    setting controls how fast )
-         *(2)=One Pass - Best Quality. The encoder places priority on the
-         *    quality of the output over encoding speed. The output is
-         *    compressed at the highest possible quality. This option takes
-         *    the longest amount of time to encode. ( speed setting ignored
-         *    )
-         *(3)=Two Pass - First Pass. The encoder generates a file of
-         *    statistics for use in the second encoding pass. ( speed
-         *    setting controls how fast )
-         *(4)=Two Pass - Second Pass. The encoder uses the statistics that
-         *    were generated in the first encoding pass to create the
-         *    compressed output. ( speed setting controls how fast )
-         *(5)=Two Pass - Second Pass Best.  The encoder uses the statistics
-         *    that were generated in the first encoding pass to create the
-         *    compressed output using the highest possible quality, and
-         *    taking a longer amount of time to encode.. ( speed setting
-         *    ignored )
-         */
-        int Mode;
+  /* mode ->
+   *(0)=Realtime/Live Encoding. This mode is optimized for realtim
+   *    encoding (for example, capturing a television signal or feed
+   *    from a live camera). ( speed setting controls how fast )
+   *(1)=Good Quality Fast Encoding. The encoder balances quality with
+   *    the amount of time it takes to encode the output. ( speed
+   *    setting controls how fast )
+   *(2)=One Pass - Best Quality. The encoder places priority on the
+   *    quality of the output over encoding speed. The output is
+   *    compressed at the highest possible quality. This option takes
+   *    the longest amount of time to encode. ( speed setting ignored
+   *    )
+   *(3)=Two Pass - First Pass. The encoder generates a file of
+   *    statistics for use in the second encoding pass. ( speed
+   *    setting controls how fast )
+   *(4)=Two Pass - Second Pass. The encoder uses the statistics that
+   *    were generated in the first encoding pass to create the
+   *    compressed output. ( speed setting controls how fast )
+   *(5)=Two Pass - Second Pass Best.  The encoder uses the statistics
+   *    that were generated in the first encoding pass to create the
+   *    compressed output using the highest possible quality, and
+   *    taking a longer amount of time to encode.. ( speed setting
+   *    ignored )
+   */
+  int Mode;
 
-        /* Key Framing Operations */
-        int auto_key;       /* automatically detect cut scenes */
-        int key_freq;       /* maximum distance to key frame. */
+  /* Key Framing Operations */
+  int auto_key; /* automatically detect cut scenes */
+  int key_freq; /* maximum distance to key frame. */
 
-        /* lagged compression (if allow_lag == 0 lag_in_frames is ignored) */
-        int allow_lag;
-        int lag_in_frames; /* how many frames lag before we start encoding */
+  /* lagged compression (if allow_lag == 0 lag_in_frames is ignored) */
+  int allow_lag;
+  int lag_in_frames; /* how many frames lag before we start encoding */
 
-        /*
-         * DATARATE CONTROL OPTIONS
-         */
+  /*
+   * DATARATE CONTROL OPTIONS
+   */
 
-        int end_usage; /* vbr or cbr */
+  int end_usage; /* vbr or cbr */
 
-        /* buffer targeting aggressiveness */
-        int under_shoot_pct;
-        int over_shoot_pct;
+  /* buffer targeting aggressiveness */
+  int under_shoot_pct;
+  int over_shoot_pct;
 
-        /* buffering parameters */
-        int64_t starting_buffer_level;
-        int64_t optimal_buffer_level;
-        int64_t maximum_buffer_size;
+  /* buffering parameters */
+  int64_t starting_buffer_level;
+  int64_t optimal_buffer_level;
+  int64_t maximum_buffer_size;
 
-        int64_t starting_buffer_level_in_ms;
-        int64_t optimal_buffer_level_in_ms;
-        int64_t maximum_buffer_size_in_ms;
+  int64_t starting_buffer_level_in_ms;
+  int64_t optimal_buffer_level_in_ms;
+  int64_t maximum_buffer_size_in_ms;
 
-        /* controlling quality */
-        int fixed_q;
-        int worst_allowed_q;
-        int best_allowed_q;
-        int cq_level;
+  /* controlling quality */
+  int fixed_q;
+  int worst_allowed_q;
+  int best_allowed_q;
+  int cq_level;
 
-        /* allow internal resizing */
-        int allow_spatial_resampling;
-        int resample_down_water_mark;
-        int resample_up_water_mark;
+  /* allow internal resizing */
+  int allow_spatial_resampling;
+  int resample_down_water_mark;
+  int resample_up_water_mark;
 
-        /* allow internal frame rate alterations */
-        int allow_df;
-        int drop_frames_water_mark;
+  /* allow internal frame rate alterations */
+  int allow_df;
+  int drop_frames_water_mark;
 
-        /* two pass datarate control */
-        int two_pass_vbrbias;
-        int two_pass_vbrmin_section;
-        int two_pass_vbrmax_section;
+  /* two pass datarate control */
+  int two_pass_vbrbias;
+  int two_pass_vbrmin_section;
+  int two_pass_vbrmax_section;
 
-        /*
-         * END DATARATE CONTROL OPTIONS
-         */
+  /*
+   * END DATARATE CONTROL OPTIONS
+   */
 
-        /* these parameters aren't to be used in final build don't use!!! */
-        int play_alternate;
-        int alt_freq;
-        int alt_q;
-        int key_q;
-        int gold_q;
+  /* these parameters aren't to be used in final build don't use!!! */
+  int play_alternate;
+  int alt_freq;
+  int alt_q;
+  int key_q;
+  int gold_q;
 
+  int multi_threaded;   /* how many threads to run the encoder on */
+  int token_partitions; /* how many token partitions to create */
 
-        int multi_threaded;   /* how many threads to run the encoder on */
-        int token_partitions; /* how many token partitions to create */
+  /* early breakout threshold: for video conf recommend 800 */
+  int encode_breakout;
 
-        /* early breakout threshold: for video conf recommend 800 */
-        int encode_breakout;
+  /* Bitfield defining the error resiliency features to enable.
+   * Can provide decodable frames after losses in previous
+   * frames and decodable partitions after losses in the same frame.
+   */
+  unsigned int error_resilient_mode;
 
-        /* Bitfield defining the error resiliency features to enable.
-         * Can provide decodable frames after losses in previous
-         * frames and decodable partitions after losses in the same frame.
-         */
-        unsigned int error_resilient_mode;
+  int arnr_max_frames;
+  int arnr_strength;
+  int arnr_type;
 
-        int arnr_max_frames;
-        int arnr_strength;
-        int arnr_type;
+  vpx_fixed_buf_t two_pass_stats_in;
+  struct vpx_codec_pkt_list *output_pkt_list;
 
-        vpx_fixed_buf_t        two_pass_stats_in;
-        struct vpx_codec_pkt_list  *output_pkt_list;
+  vp8e_tuning tuning;
 
-        vp8e_tuning tuning;
+  /* Temporal scaling parameters */
+  unsigned int number_of_layers;
+  unsigned int target_bitrate[VPX_TS_MAX_PERIODICITY];
+  unsigned int rate_decimator[VPX_TS_MAX_PERIODICITY];
+  unsigned int periodicity;
+  unsigned int layer_id[VPX_TS_MAX_PERIODICITY];
 
-        /* Temporal scaling parameters */
-        unsigned int number_of_layers;
-        unsigned int target_bitrate[VPX_TS_MAX_PERIODICITY];
-        unsigned int rate_decimator[VPX_TS_MAX_PERIODICITY];
-        unsigned int periodicity;
-        unsigned int layer_id[VPX_TS_MAX_PERIODICITY];
-
 #if CONFIG_MULTI_RES_ENCODING
-        /* Number of total resolutions encoded */
-        unsigned int mr_total_resolutions;
+  /* Number of total resolutions encoded */
+  unsigned int mr_total_resolutions;
 
-        /* Current encoder ID */
-        unsigned int mr_encoder_id;
+  /* Current encoder ID */
+  unsigned int mr_encoder_id;
 
-        /* Down-sampling factor */
-        vpx_rational_t mr_down_sampling_factor;
+  /* Down-sampling factor */
+  vpx_rational_t mr_down_sampling_factor;
 
-        /* Memory location to store low-resolution encoder's mode info */
-        void* mr_low_res_mode_info;
+  /* Memory location to store low-resolution encoder's mode info */
+  void *mr_low_res_mode_info;
 #endif
-    } VP8_CONFIG;
+} VP8_CONFIG;
 
+void vp8_initialize();
 
-    void vp8_initialize();
+struct VP8_COMP *vp8_create_compressor(VP8_CONFIG *oxcf);
+void vp8_remove_compressor(struct VP8_COMP **comp);
 
-    struct VP8_COMP* vp8_create_compressor(VP8_CONFIG *oxcf);
-    void vp8_remove_compressor(struct VP8_COMP* *comp);
+void vp8_init_config(struct VP8_COMP *onyx, VP8_CONFIG *oxcf);
+void vp8_change_config(struct VP8_COMP *onyx, VP8_CONFIG *oxcf);
 
-    void vp8_init_config(struct VP8_COMP* onyx, VP8_CONFIG *oxcf);
-    void vp8_change_config(struct VP8_COMP* onyx, VP8_CONFIG *oxcf);
+int vp8_receive_raw_frame(struct VP8_COMP *comp, unsigned int frame_flags,
+                          YV12_BUFFER_CONFIG *sd, int64_t time_stamp,
+                          int64_t end_time_stamp);
+int vp8_get_compressed_data(struct VP8_COMP *comp, unsigned int *frame_flags,
+                            unsigned long *size, unsigned char *dest,
+                            unsigned char *dest_end, int64_t *time_stamp,
+                            int64_t *time_end, int flush);
+int vp8_get_preview_raw_frame(struct VP8_COMP *comp, YV12_BUFFER_CONFIG *dest,
+                              vp8_ppflags_t *flags);
 
-    int vp8_receive_raw_frame(struct VP8_COMP* comp, unsigned int frame_flags, YV12_BUFFER_CONFIG *sd, int64_t time_stamp, int64_t end_time_stamp);
-    int vp8_get_compressed_data(struct VP8_COMP* comp, unsigned int *frame_flags, unsigned long *size, unsigned char *dest, unsigned char *dest_end, int64_t *time_stamp, int64_t *time_end, int flush);
-    int vp8_get_preview_raw_frame(struct VP8_COMP* comp, YV12_BUFFER_CONFIG *dest, vp8_ppflags_t *flags);
-
-    int vp8_use_as_reference(struct VP8_COMP* comp, int ref_frame_flags);
-    int vp8_update_reference(struct VP8_COMP* comp, int ref_frame_flags);
-    int vp8_get_reference(struct VP8_COMP* comp, enum vpx_ref_frame_type ref_frame_flag, YV12_BUFFER_CONFIG *sd);
-    int vp8_set_reference(struct VP8_COMP* comp, enum vpx_ref_frame_type ref_frame_flag, YV12_BUFFER_CONFIG *sd);
-    int vp8_update_entropy(struct VP8_COMP* comp, int update);
-    int vp8_set_roimap(struct VP8_COMP* comp, unsigned char *map, unsigned int rows, unsigned int cols, int delta_q[4], int delta_lf[4], unsigned int threshold[4]);
-    int vp8_set_active_map(struct VP8_COMP* comp, unsigned char *map, unsigned int rows, unsigned int cols);
-    int vp8_set_internal_size(struct VP8_COMP* comp, VPX_SCALING horiz_mode, VPX_SCALING vert_mode);
-    int vp8_get_quantizer(struct VP8_COMP* c);
+int vp8_use_as_reference(struct VP8_COMP *comp, int ref_frame_flags);
+int vp8_update_reference(struct VP8_COMP *comp, int ref_frame_flags);
+int vp8_get_reference(struct VP8_COMP *comp,
+                      enum vpx_ref_frame_type ref_frame_flag,
+                      YV12_BUFFER_CONFIG *sd);
+int vp8_set_reference(struct VP8_COMP *comp,
+                      enum vpx_ref_frame_type ref_frame_flag,
+                      YV12_BUFFER_CONFIG *sd);
+int vp8_update_entropy(struct VP8_COMP *comp, int update);
+int vp8_set_roimap(struct VP8_COMP *comp, unsigned char *map, unsigned int rows,
+                   unsigned int cols, int delta_q[4], int delta_lf[4],
+                   unsigned int threshold[4]);
+int vp8_set_active_map(struct VP8_COMP *comp, unsigned char *map,
+                       unsigned int rows, unsigned int cols);
+int vp8_set_internal_size(struct VP8_COMP *comp, VPX_SCALING horiz_mode,
+                          VPX_SCALING vert_mode);
+int vp8_get_quantizer(struct VP8_COMP *c);
 
 #ifdef __cplusplus
 }
--- a/vp8/common/onyxc_int.h
+++ b/vp8/common/onyxc_int.h
@@ -8,7 +8,6 @@
  *  be found in the AUTHORS file in the root of the source tree.
  */
 
-
 #ifndef VP8_COMMON_ONYXC_INT_H_
 #define VP8_COMMON_ONYXC_INT_H_
 
@@ -38,144 +37,143 @@
 
 #define MAX_PARTITIONS 9
 
-typedef struct frame_contexts
-{
-    vp8_prob bmode_prob [VP8_BINTRAMODES-1];
-    vp8_prob ymode_prob [VP8_YMODES-1];   /* interframe intra mode probs */
-    vp8_prob uv_mode_prob [VP8_UV_MODES-1];
-    vp8_prob sub_mv_ref_prob [VP8_SUBMVREFS-1];
-    vp8_prob coef_probs [BLOCK_TYPES] [COEF_BANDS] [PREV_COEF_CONTEXTS] [ENTROPY_NODES];
-    MV_CONTEXT mvc[2];
+typedef struct frame_contexts {
+  vp8_prob bmode_prob[VP8_BINTRAMODES - 1];
+  vp8_prob ymode_prob[VP8_YMODES - 1]; /* interframe intra mode probs */
+  vp8_prob uv_mode_prob[VP8_UV_MODES - 1];
+  vp8_prob sub_mv_ref_prob[VP8_SUBMVREFS - 1];
+  vp8_prob
+      coef_probs[BLOCK_TYPES][COEF_BANDS][PREV_COEF_CONTEXTS][ENTROPY_NODES];
+  MV_CONTEXT mvc[2];
 } FRAME_CONTEXT;
 
-typedef enum
-{
-    ONE_PARTITION  = 0,
-    TWO_PARTITION  = 1,
-    FOUR_PARTITION = 2,
-    EIGHT_PARTITION = 3
+typedef enum {
+  ONE_PARTITION = 0,
+  TWO_PARTITION = 1,
+  FOUR_PARTITION = 2,
+  EIGHT_PARTITION = 3
 } TOKEN_PARTITION;
 
-typedef enum
-{
-    RECON_CLAMP_REQUIRED        = 0,
-    RECON_CLAMP_NOTREQUIRED     = 1
+typedef enum {
+  RECON_CLAMP_REQUIRED = 0,
+  RECON_CLAMP_NOTREQUIRED = 1
 } CLAMP_TYPE;
 
 typedef struct VP8Common
 
-{
-    struct vpx_internal_error_info  error;
+    {
+  struct vpx_internal_error_info error;
 
-    DECLARE_ALIGNED(16, short, Y1dequant[QINDEX_RANGE][2]);
-    DECLARE_ALIGNED(16, short, Y2dequant[QINDEX_RANGE][2]);
-    DECLARE_ALIGNED(16, short, UVdequant[QINDEX_RANGE][2]);
+  DECLARE_ALIGNED(16, short, Y1dequant[QINDEX_RANGE][2]);
+  DECLARE_ALIGNED(16, short, Y2dequant[QINDEX_RANGE][2]);
+  DECLARE_ALIGNED(16, short, UVdequant[QINDEX_RANGE][2]);
 
-    int Width;
-    int Height;
-    int horiz_scale;
-    int vert_scale;
+  int Width;
+  int Height;
+  int horiz_scale;
+  int vert_scale;
 
-    CLAMP_TYPE  clamp_type;
+  CLAMP_TYPE clamp_type;
 
-    YV12_BUFFER_CONFIG *frame_to_show;
+  YV12_BUFFER_CONFIG *frame_to_show;
 
-    YV12_BUFFER_CONFIG yv12_fb[NUM_YV12_BUFFERS];
-    int fb_idx_ref_cnt[NUM_YV12_BUFFERS];
-    int new_fb_idx, lst_fb_idx, gld_fb_idx, alt_fb_idx;
+  YV12_BUFFER_CONFIG yv12_fb[NUM_YV12_BUFFERS];
+  int fb_idx_ref_cnt[NUM_YV12_BUFFERS];
+  int new_fb_idx, lst_fb_idx, gld_fb_idx, alt_fb_idx;
 
-    YV12_BUFFER_CONFIG temp_scale_frame;
+  YV12_BUFFER_CONFIG temp_scale_frame;
 
 #if CONFIG_POSTPROC
-    YV12_BUFFER_CONFIG post_proc_buffer;
-    YV12_BUFFER_CONFIG post_proc_buffer_int;
-    int post_proc_buffer_int_used;
-    unsigned char *pp_limits_buffer;   /* post-processing filter coefficients */
+  YV12_BUFFER_CONFIG post_proc_buffer;
+  YV12_BUFFER_CONFIG post_proc_buffer_int;
+  int post_proc_buffer_int_used;
+  unsigned char *pp_limits_buffer; /* post-processing filter coefficients */
 #endif
 
-    FRAME_TYPE last_frame_type;  /* Save last frame's frame type for motion search. */
-    FRAME_TYPE frame_type;
+  FRAME_TYPE
+  last_frame_type; /* Save last frame's frame type for motion search. */
+  FRAME_TYPE frame_type;
 
-    int show_frame;
+  int show_frame;
 
-    int frame_flags;
-    int MBs;
-    int mb_rows;
-    int mb_cols;
-    int mode_info_stride;
+  int frame_flags;
+  int MBs;
+  int mb_rows;
+  int mb_cols;
+  int mode_info_stride;
 
-    /* profile settings */
-    int mb_no_coeff_skip;
-    int no_lpf;
-    int use_bilinear_mc_filter;
-    int full_pixel;
+  /* profile settings */
+  int mb_no_coeff_skip;
+  int no_lpf;
+  int use_bilinear_mc_filter;
+  int full_pixel;
 
-    int base_qindex;
+  int base_qindex;
 
-    int y1dc_delta_q;
-    int y2dc_delta_q;
-    int y2ac_delta_q;
-    int uvdc_delta_q;
-    int uvac_delta_q;
+  int y1dc_delta_q;
+  int y2dc_delta_q;
+  int y2ac_delta_q;
+  int uvdc_delta_q;
+  int uvac_delta_q;
 
-    /* We allocate a MODE_INFO struct for each macroblock, together with
-       an extra row on top and column on the left to simplify prediction. */
+  /* We allocate a MODE_INFO struct for each macroblock, together with
+     an extra row on top and column on the left to simplify prediction. */
 
-    MODE_INFO *mip; /* Base of allocated array */
-    MODE_INFO *mi;  /* Corresponds to upper left visible macroblock */
+  MODE_INFO *mip; /* Base of allocated array */
+  MODE_INFO *mi;  /* Corresponds to upper left visible macroblock */
 #if CONFIG_ERROR_CONCEALMENT
-    MODE_INFO *prev_mip; /* MODE_INFO array 'mip' from last decoded frame */
-    MODE_INFO *prev_mi;  /* 'mi' from last frame (points into prev_mip) */
+  MODE_INFO *prev_mip; /* MODE_INFO array 'mip' from last decoded frame */
+  MODE_INFO *prev_mi;  /* 'mi' from last frame (points into prev_mip) */
 #endif
-    MODE_INFO *show_frame_mi;  /* MODE_INFO for the last decoded frame
-                                  to show */
-    LOOPFILTERTYPE filter_type;
+  /* MODE_INFO for the last decoded frame to show */
+  MODE_INFO *show_frame_mi;
+  LOOPFILTERTYPE filter_type;
 
-    loop_filter_info_n lf_info;
+  loop_filter_info_n lf_info;
 
-    int filter_level;
-    int last_sharpness_level;
-    int sharpness_level;
+  int filter_level;
+  int last_sharpness_level;
+  int sharpness_level;
 
-    int refresh_last_frame;       /* Two state 0 = NO, 1 = YES */
-    int refresh_golden_frame;     /* Two state 0 = NO, 1 = YES */
-    int refresh_alt_ref_frame;     /* Two state 0 = NO, 1 = YES */
+  int refresh_last_frame;    /* Two state 0 = NO, 1 = YES */
+  int refresh_golden_frame;  /* Two state 0 = NO, 1 = YES */
+  int refresh_alt_ref_frame; /* Two state 0 = NO, 1 = YES */
 
-    int copy_buffer_to_gf;         /* 0 none, 1 Last to GF, 2 ARF to GF */
-    int copy_buffer_to_arf;        /* 0 none, 1 Last to ARF, 2 GF to ARF */
+  int copy_buffer_to_gf;  /* 0 none, 1 Last to GF, 2 ARF to GF */
+  int copy_buffer_to_arf; /* 0 none, 1 Last to ARF, 2 GF to ARF */
 
-    int refresh_entropy_probs;    /* Two state 0 = NO, 1 = YES */
+  int refresh_entropy_probs; /* Two state 0 = NO, 1 = YES */
 
-    int ref_frame_sign_bias[MAX_REF_FRAMES];    /* Two state 0, 1 */
+  int ref_frame_sign_bias[MAX_REF_FRAMES]; /* Two state 0, 1 */
 
-    /* Y,U,V,Y2 */
-    ENTROPY_CONTEXT_PLANES *above_context;   /* row of context for each plane */
-    ENTROPY_CONTEXT_PLANES left_context;  /* (up to) 4 contexts "" */
+  /* Y,U,V,Y2 */
+  ENTROPY_CONTEXT_PLANES *above_context; /* row of context for each plane */
+  ENTROPY_CONTEXT_PLANES left_context;   /* (up to) 4 contexts "" */
 
-    FRAME_CONTEXT lfc; /* last frame entropy */
-    FRAME_CONTEXT fc;  /* this frame entropy */
+  FRAME_CONTEXT lfc; /* last frame entropy */
+  FRAME_CONTEXT fc;  /* this frame entropy */
 
-    unsigned int current_video_frame;
+  unsigned int current_video_frame;
 
-    int version;
+  int version;
 
-    TOKEN_PARTITION multi_token_partition;
+  TOKEN_PARTITION multi_token_partition;
 
 #ifdef PACKET_TESTING
-    VP8_HEADER oh;
+  VP8_HEADER oh;
 #endif
 #if CONFIG_POSTPROC_VISUALIZER
-    double bitrate;
-    double framerate;
+  double bitrate;
+  double framerate;
 #endif
 
 #if CONFIG_MULTITHREAD
-    int processor_core_count;
+  int processor_core_count;
 #endif
 #if CONFIG_POSTPROC
-    struct postproc_state  postproc_state;
+  struct postproc_state postproc_state;
 #endif
-    int cpu_caps;
+  int cpu_caps;
 } VP8_COMMON;
 
 #ifdef __cplusplus
--- a/vp8/common/onyxd.h
+++ b/vp8/common/onyxd.h
@@ -8,15 +8,12 @@
  *  be found in the AUTHORS file in the root of the source tree.
  */
 
-
 #ifndef VP8_COMMON_ONYXD_H_
 #define VP8_COMMON_ONYXD_H_
 
-
 /* Create/destroy static data structures. */
 #ifdef __cplusplus
-extern "C"
-{
+extern "C" {
 #endif
 #include "vpx_scale/yv12config.h"
 #include "ppflags.h"
@@ -24,40 +21,40 @@
 #include "vpx/vpx_codec.h"
 #include "vpx/vp8.h"
 
-    struct VP8D_COMP;
+struct VP8D_COMP;
 
-    typedef struct
-    {
-        int     Width;
-        int     Height;
-        int     Version;
-        int     postprocess;
-        int     max_threads;
-        int     error_concealment;
-    } VP8D_CONFIG;
+typedef struct {
+  int Width;
+  int Height;
+  int Version;
+  int postprocess;
+  int max_threads;
+  int error_concealment;
+} VP8D_CONFIG;
 
-    typedef enum
-    {
-        VP8D_OK = 0
-    } VP8D_SETTING;
+typedef enum { VP8D_OK = 0 } VP8D_SETTING;
 
-    void vp8dx_initialize(void);
+void vp8dx_initialize(void);
 
-    void vp8dx_set_setting(struct VP8D_COMP* comp, VP8D_SETTING oxst, int x);
+void vp8dx_set_setting(struct VP8D_COMP *comp, VP8D_SETTING oxst, int x);
 
-    int vp8dx_get_setting(struct VP8D_COMP* comp, VP8D_SETTING oxst);
+int vp8dx_get_setting(struct VP8D_COMP *comp, VP8D_SETTING oxst);
 
-    int vp8dx_receive_compressed_data(struct VP8D_COMP* comp,
-                                      size_t size, const uint8_t *dest,
-                                      int64_t time_stamp);
-    int vp8dx_get_raw_frame(struct VP8D_COMP* comp, YV12_BUFFER_CONFIG *sd, int64_t *time_stamp, int64_t *time_end_stamp, vp8_ppflags_t *flags);
+int vp8dx_receive_compressed_data(struct VP8D_COMP *comp, size_t size,
+                                  const uint8_t *dest, int64_t time_stamp);
+int vp8dx_get_raw_frame(struct VP8D_COMP *comp, YV12_BUFFER_CONFIG *sd,
+                        int64_t *time_stamp, int64_t *time_end_stamp,
+                        vp8_ppflags_t *flags);
 
-    vpx_codec_err_t vp8dx_get_reference(struct VP8D_COMP* comp, enum vpx_ref_frame_type ref_frame_flag, YV12_BUFFER_CONFIG *sd);
-    vpx_codec_err_t vp8dx_set_reference(struct VP8D_COMP* comp, enum vpx_ref_frame_type ref_frame_flag, YV12_BUFFER_CONFIG *sd);
+vpx_codec_err_t vp8dx_get_reference(struct VP8D_COMP *comp,
+                                    enum vpx_ref_frame_type ref_frame_flag,
+                                    YV12_BUFFER_CONFIG *sd);
+vpx_codec_err_t vp8dx_set_reference(struct VP8D_COMP *comp,
+                                    enum vpx_ref_frame_type ref_frame_flag,
+                                    YV12_BUFFER_CONFIG *sd);
 
 #ifdef __cplusplus
 }
 #endif
-
 
 #endif  // VP8_COMMON_ONYXD_H_
--- a/vp8/common/postproc.c
+++ b/vp8/common/postproc.c
@@ -8,7 +8,6 @@
  *  be found in the AUTHORS file in the root of the source tree.
  */
 
-
 #include "vpx_config.h"
 #include "vpx_dsp_rtcd.h"
 #include "vp8_rtcd.h"
@@ -25,182 +24,165 @@
 #include <stdlib.h>
 #include <stdio.h>
 
-#define RGB_TO_YUV(t)                                                                       \
-    ( (0.257*(float)(t>>16)) + (0.504*(float)(t>>8&0xff)) + (0.098*(float)(t&0xff)) + 16),  \
-    (-(0.148*(float)(t>>16)) - (0.291*(float)(t>>8&0xff)) + (0.439*(float)(t&0xff)) + 128), \
-    ( (0.439*(float)(t>>16)) - (0.368*(float)(t>>8&0xff)) - (0.071*(float)(t&0xff)) + 128)
+#define RGB_TO_YUV(t)                                                   \
+  ((0.257 * (float)(t >> 16)) + (0.504 * (float)(t >> 8 & 0xff)) +      \
+   (0.098 * (float)(t & 0xff)) + 16),                                   \
+      (-(0.148 * (float)(t >> 16)) - (0.291 * (float)(t >> 8 & 0xff)) + \
+       (0.439 * (float)(t & 0xff)) + 128),                              \
+      ((0.439 * (float)(t >> 16)) - (0.368 * (float)(t >> 8 & 0xff)) -  \
+       (0.071 * (float)(t & 0xff)) + 128)
 
 /* global constants */
 #if CONFIG_POSTPROC_VISUALIZER
-static const unsigned char MB_PREDICTION_MODE_colors[MB_MODE_COUNT][3] =
-{
-    { RGB_TO_YUV(0x98FB98) },   /* PaleGreen */
-    { RGB_TO_YUV(0x00FF00) },   /* Green */
-    { RGB_TO_YUV(0xADFF2F) },   /* GreenYellow */
-    { RGB_TO_YUV(0x228B22) },   /* ForestGreen */
-    { RGB_TO_YUV(0x006400) },   /* DarkGreen */
-    { RGB_TO_YUV(0x98F5FF) },   /* Cadet Blue */
-    { RGB_TO_YUV(0x6CA6CD) },   /* Sky Blue */
-    { RGB_TO_YUV(0x00008B) },   /* Dark blue */
-    { RGB_TO_YUV(0x551A8B) },   /* Purple */
-    { RGB_TO_YUV(0xFF0000) }    /* Red */
+static const unsigned char MB_PREDICTION_MODE_colors[MB_MODE_COUNT][3] = {
+  { RGB_TO_YUV(0x98FB98) }, /* PaleGreen */
+  { RGB_TO_YUV(0x00FF00) }, /* Green */
+  { RGB_TO_YUV(0xADFF2F) }, /* GreenYellow */
+  { RGB_TO_YUV(0x228B22) }, /* ForestGreen */
+  { RGB_TO_YUV(0x006400) }, /* DarkGreen */
+  { RGB_TO_YUV(0x98F5FF) }, /* Cadet Blue */
+  { RGB_TO_YUV(0x6CA6CD) }, /* Sky Blue */
+  { RGB_TO_YUV(0x00008B) }, /* Dark blue */
+  { RGB_TO_YUV(0x551A8B) }, /* Purple */
+  { RGB_TO_YUV(0xFF0000) }  /* Red */
 };
 
-static const unsigned char B_PREDICTION_MODE_colors[B_MODE_COUNT][3] =
-{
-    { RGB_TO_YUV(0x6633ff) },   /* Purple */
-    { RGB_TO_YUV(0xcc33ff) },   /* Magenta */
-    { RGB_TO_YUV(0xff33cc) },   /* Pink */
-    { RGB_TO_YUV(0xff3366) },   /* Coral */
-    { RGB_TO_YUV(0x3366ff) },   /* Blue */
-    { RGB_TO_YUV(0xed00f5) },   /* Dark Blue */
-    { RGB_TO_YUV(0x2e00b8) },   /* Dark Purple */
-    { RGB_TO_YUV(0xff6633) },   /* Orange */
-    { RGB_TO_YUV(0x33ccff) },   /* Light Blue */
-    { RGB_TO_YUV(0x8ab800) },   /* Green */
-    { RGB_TO_YUV(0xffcc33) },   /* Light Orange */
-    { RGB_TO_YUV(0x33ffcc) },   /* Aqua */
-    { RGB_TO_YUV(0x66ff33) },   /* Light Green */
-    { RGB_TO_YUV(0xccff33) },   /* Yellow */
+static const unsigned char B_PREDICTION_MODE_colors[B_MODE_COUNT][3] = {
+  { RGB_TO_YUV(0x6633ff) }, /* Purple */
+  { RGB_TO_YUV(0xcc33ff) }, /* Magenta */
+  { RGB_TO_YUV(0xff33cc) }, /* Pink */
+  { RGB_TO_YUV(0xff3366) }, /* Coral */
+  { RGB_TO_YUV(0x3366ff) }, /* Blue */
+  { RGB_TO_YUV(0xed00f5) }, /* Dark Blue */
+  { RGB_TO_YUV(0x2e00b8) }, /* Dark Purple */
+  { RGB_TO_YUV(0xff6633) }, /* Orange */
+  { RGB_TO_YUV(0x33ccff) }, /* Light Blue */
+  { RGB_TO_YUV(0x8ab800) }, /* Green */
+  { RGB_TO_YUV(0xffcc33) }, /* Light Orange */
+  { RGB_TO_YUV(0x33ffcc) }, /* Aqua */
+  { RGB_TO_YUV(0x66ff33) }, /* Light Green */
+  { RGB_TO_YUV(0xccff33) }, /* Yellow */
 };
 
-static const unsigned char MV_REFERENCE_FRAME_colors[MAX_REF_FRAMES][3] =
-{
-    { RGB_TO_YUV(0x00ff00) },   /* Blue */
-    { RGB_TO_YUV(0x0000ff) },   /* Green */
-    { RGB_TO_YUV(0xffff00) },   /* Yellow */
-    { RGB_TO_YUV(0xff0000) },   /* Red */
+static const unsigned char MV_REFERENCE_FRAME_colors[MAX_REF_FRAMES][3] = {
+  { RGB_TO_YUV(0x00ff00) }, /* Blue */
+  { RGB_TO_YUV(0x0000ff) }, /* Green */
+  { RGB_TO_YUV(0xffff00) }, /* Yellow */
+  { RGB_TO_YUV(0xff0000) }, /* Red */
 };
 #endif
 
-
-extern void vp8_blit_text(const char *msg, unsigned char *address, const int pitch);
-extern void vp8_blit_line(int x0, int x1, int y0, int y1, unsigned char *image, const int pitch);
+extern void vp8_blit_text(const char *msg, unsigned char *address,
+                          const int pitch);
+extern void vp8_blit_line(int x0, int x1, int y0, int y1, unsigned char *image,
+                          const int pitch);
 /***********************************************************************************************************
  */
-static int q2mbl(int x)
-{
-    if (x < 20) x = 20;
+static int q2mbl(int x) {
+  if (x < 20) x = 20;
 
-    x = 50 + (x - 50) * 10 / 8;
-    return x * x / 3;
+  x = 50 + (x - 50) * 10 / 8;
+  return x * x / 3;
 }
 
 #if CONFIG_POSTPROC
-static void vp8_de_mblock(YV12_BUFFER_CONFIG         *post,
-                          int                         q)
-{
-    vpx_mbpost_proc_across_ip(post->y_buffer, post->y_stride, post->y_height,
-                              post->y_width, q2mbl(q));
-    vpx_mbpost_proc_down(post->y_buffer, post->y_stride, post->y_height,
-                         post->y_width, q2mbl(q));
+static void vp8_de_mblock(YV12_BUFFER_CONFIG *post, int q) {
+  vpx_mbpost_proc_across_ip(post->y_buffer, post->y_stride, post->y_height,
+                            post->y_width, q2mbl(q));
+  vpx_mbpost_proc_down(post->y_buffer, post->y_stride, post->y_height,
+                       post->y_width, q2mbl(q));
 }
 
-void vp8_deblock(VP8_COMMON                 *cm,
-                 YV12_BUFFER_CONFIG         *source,
-                 YV12_BUFFER_CONFIG         *post,
-                 int                         q,
-                 int                         low_var_thresh,
-                 int                         flag)
-{
-    double level = 6.0e-05 * q * q * q - .0067 * q * q + .306 * q + .0065;
-    int ppl = (int)(level + .5);
+void vp8_deblock(VP8_COMMON *cm, YV12_BUFFER_CONFIG *source,
+                 YV12_BUFFER_CONFIG *post, int q, int low_var_thresh,
+                 int flag) {
+  double level = 6.0e-05 * q * q * q - .0067 * q * q + .306 * q + .0065;
+  int ppl = (int)(level + .5);
 
-    const MODE_INFO *mode_info_context = cm->show_frame_mi;
-    int mbr, mbc;
+  const MODE_INFO *mode_info_context = cm->show_frame_mi;
+  int mbr, mbc;
 
-    /* The pixel thresholds are adjusted according to if or not the macroblock
-     * is a skipped block.  */
-    unsigned char *ylimits = cm->pp_limits_buffer;
-    unsigned char *uvlimits = cm->pp_limits_buffer + 16 * cm->mb_cols;
-    (void) low_var_thresh;
-    (void) flag;
+  /* The pixel thresholds are adjusted according to if or not the macroblock
+   * is a skipped block.  */
+  unsigned char *ylimits = cm->pp_limits_buffer;
+  unsigned char *uvlimits = cm->pp_limits_buffer + 16 * cm->mb_cols;
+  (void)low_var_thresh;
+  (void)flag;
 
-    if (ppl > 0)
-    {
-        for (mbr = 0; mbr < cm->mb_rows; mbr++)
-        {
-            unsigned char *ylptr = ylimits;
-            unsigned char *uvlptr = uvlimits;
-            for (mbc = 0; mbc < cm->mb_cols; mbc++)
-            {
-                unsigned char mb_ppl;
+  if (ppl > 0) {
+    for (mbr = 0; mbr < cm->mb_rows; mbr++) {
+      unsigned char *ylptr = ylimits;
+      unsigned char *uvlptr = uvlimits;
+      for (mbc = 0; mbc < cm->mb_cols; mbc++) {
+        unsigned char mb_ppl;
 
-                if (mode_info_context->mbmi.mb_skip_coeff)
-                    mb_ppl = (unsigned char)ppl >> 1;
-                else
-                    mb_ppl = (unsigned char)ppl;
+        if (mode_info_context->mbmi.mb_skip_coeff)
+          mb_ppl = (unsigned char)ppl >> 1;
+        else
+          mb_ppl = (unsigned char)ppl;
 
-                memset(ylptr, mb_ppl, 16);
-                memset(uvlptr, mb_ppl, 8);
+        memset(ylptr, mb_ppl, 16);
+        memset(uvlptr, mb_ppl, 8);
 
-                ylptr += 16;
-                uvlptr += 8;
-                mode_info_context++;
-            }
-            mode_info_context++;
+        ylptr += 16;
+        uvlptr += 8;
+        mode_info_context++;
+      }
+      mode_info_context++;
 
-            vpx_post_proc_down_and_across_mb_row(
-                source->y_buffer + 16 * mbr * source->y_stride,
-                post->y_buffer + 16 * mbr * post->y_stride, source->y_stride,
-                post->y_stride, source->y_width, ylimits, 16);
+      vpx_post_proc_down_and_across_mb_row(
+          source->y_buffer + 16 * mbr * source->y_stride,
+          post->y_buffer + 16 * mbr * post->y_stride, source->y_stride,
+          post->y_stride, source->y_width, ylimits, 16);
 
-            vpx_post_proc_down_and_across_mb_row(
-                source->u_buffer + 8 * mbr * source->uv_stride,
-                post->u_buffer + 8 * mbr * post->uv_stride, source->uv_stride,
-                post->uv_stride, source->uv_width, uvlimits, 8);
-            vpx_post_proc_down_and_across_mb_row(
-                source->v_buffer + 8 * mbr * source->uv_stride,
-                post->v_buffer + 8 * mbr * post->uv_stride, source->uv_stride,
-                post->uv_stride, source->uv_width, uvlimits, 8);
-        }
-    } else
-    {
-        vp8_yv12_copy_frame(source, post);
+      vpx_post_proc_down_and_across_mb_row(
+          source->u_buffer + 8 * mbr * source->uv_stride,
+          post->u_buffer + 8 * mbr * post->uv_stride, source->uv_stride,
+          post->uv_stride, source->uv_width, uvlimits, 8);
+      vpx_post_proc_down_and_across_mb_row(
+          source->v_buffer + 8 * mbr * source->uv_stride,
+          post->v_buffer + 8 * mbr * post->uv_stride, source->uv_stride,
+          post->uv_stride, source->uv_width, uvlimits, 8);
     }
+  } else {
+    vp8_yv12_copy_frame(source, post);
+  }
 }
 #endif
 
-void vp8_de_noise(VP8_COMMON                 *cm,
-                  YV12_BUFFER_CONFIG         *source,
-                  YV12_BUFFER_CONFIG         *post,
-                  int                         q,
-                  int                         low_var_thresh,
-                  int                         flag,
-                  int                         uvfilter)
-{
-    int mbr;
-    double level = 6.0e-05 * q * q * q - .0067 * q * q + .306 * q + .0065;
-    int ppl = (int)(level + .5);
-    int mb_rows = cm->mb_rows;
-    int mb_cols = cm->mb_cols;
-    unsigned char *limits = cm->pp_limits_buffer;;
-    (void) post;
-    (void) low_var_thresh;
-    (void) flag;
+void vp8_de_noise(VP8_COMMON *cm, YV12_BUFFER_CONFIG *source,
+                  YV12_BUFFER_CONFIG *post, int q, int low_var_thresh, int flag,
+                  int uvfilter) {
+  int mbr;
+  double level = 6.0e-05 * q * q * q - .0067 * q * q + .306 * q + .0065;
+  int ppl = (int)(level + .5);
+  int mb_rows = cm->mb_rows;
+  int mb_cols = cm->mb_cols;
+  unsigned char *limits = cm->pp_limits_buffer;
+  ;
+  (void)post;
+  (void)low_var_thresh;
+  (void)flag;
 
-    memset(limits, (unsigned char)ppl, 16 * mb_cols);
+  memset(limits, (unsigned char)ppl, 16 * mb_cols);
 
-    /* TODO: The original code don't filter the 2 outer rows and columns. */
-    for (mbr = 0; mbr < mb_rows; mbr++)
-    {
-        vpx_post_proc_down_and_across_mb_row(
-            source->y_buffer + 16 * mbr * source->y_stride,
-            source->y_buffer + 16 * mbr * source->y_stride,
-            source->y_stride, source->y_stride, source->y_width, limits, 16);
-        if (uvfilter == 1) {
-          vpx_post_proc_down_and_across_mb_row(
-              source->u_buffer + 8 * mbr * source->uv_stride,
-              source->u_buffer + 8 * mbr * source->uv_stride,
-              source->uv_stride, source->uv_stride, source->uv_width, limits,
-              8);
-          vpx_post_proc_down_and_across_mb_row(
-              source->v_buffer + 8 * mbr * source->uv_stride,
-              source->v_buffer + 8 * mbr * source->uv_stride,
-              source->uv_stride, source->uv_stride, source->uv_width, limits,
-              8);
-        }
+  /* TODO: The original code don't filter the 2 outer rows and columns. */
+  for (mbr = 0; mbr < mb_rows; mbr++) {
+    vpx_post_proc_down_and_across_mb_row(
+        source->y_buffer + 16 * mbr * source->y_stride,
+        source->y_buffer + 16 * mbr * source->y_stride, source->y_stride,
+        source->y_stride, source->y_width, limits, 16);
+    if (uvfilter == 1) {
+      vpx_post_proc_down_and_across_mb_row(
+          source->u_buffer + 8 * mbr * source->uv_stride,
+          source->u_buffer + 8 * mbr * source->uv_stride, source->uv_stride,
+          source->uv_stride, source->uv_width, limits, 8);
+      vpx_post_proc_down_and_across_mb_row(
+          source->v_buffer + 8 * mbr * source->uv_stride,
+          source->v_buffer + 8 * mbr * source->uv_stride, source->uv_stride,
+          source->uv_stride, source->uv_width, limits, 8);
     }
+  }
 }
 
 /* Blend the macro block with a solid colored square.  Leave the
@@ -207,683 +189,597 @@
  * edges unblended to give distinction to macro blocks in areas
  * filled with the same color block.
  */
-void vp8_blend_mb_inner_c (unsigned char *y, unsigned char *u, unsigned char *v,
-                        int y_1, int u_1, int v_1, int alpha, int stride)
-{
-    int i, j;
-    int y1_const = y_1*((1<<16)-alpha);
-    int u1_const = u_1*((1<<16)-alpha);
-    int v1_const = v_1*((1<<16)-alpha);
+void vp8_blend_mb_inner_c(unsigned char *y, unsigned char *u, unsigned char *v,
+                          int y_1, int u_1, int v_1, int alpha, int stride) {
+  int i, j;
+  int y1_const = y_1 * ((1 << 16) - alpha);
+  int u1_const = u_1 * ((1 << 16) - alpha);
+  int v1_const = v_1 * ((1 << 16) - alpha);
 
-    y += 2*stride + 2;
-    for (i = 0; i < 12; i++)
-    {
-        for (j = 0; j < 12; j++)
-        {
-            y[j] = (y[j]*alpha + y1_const)>>16;
-        }
-        y += stride;
+  y += 2 * stride + 2;
+  for (i = 0; i < 12; i++) {
+    for (j = 0; j < 12; j++) {
+      y[j] = (y[j] * alpha + y1_const) >> 16;
     }
+    y += stride;
+  }
 
-    stride >>= 1;
+  stride >>= 1;
 
-    u += stride + 1;
-    v += stride + 1;
+  u += stride + 1;
+  v += stride + 1;
 
-    for (i = 0; i < 6; i++)
-    {
-        for (j = 0; j < 6; j++)
-        {
-            u[j] = (u[j]*alpha + u1_const)>>16;
-            v[j] = (v[j]*alpha + v1_const)>>16;
-        }
-        u += stride;
-        v += stride;
+  for (i = 0; i < 6; i++) {
+    for (j = 0; j < 6; j++) {
+      u[j] = (u[j] * alpha + u1_const) >> 16;
+      v[j] = (v[j] * alpha + v1_const) >> 16;
     }
+    u += stride;
+    v += stride;
+  }
 }
 
 /* Blend only the edge of the macro block.  Leave center
  * unblended to allow for other visualizations to be layered.
  */
-void vp8_blend_mb_outer_c (unsigned char *y, unsigned char *u, unsigned char *v,
-                        int y_1, int u_1, int v_1, int alpha, int stride)
-{
-    int i, j;
-    int y1_const = y_1*((1<<16)-alpha);
-    int u1_const = u_1*((1<<16)-alpha);
-    int v1_const = v_1*((1<<16)-alpha);
+void vp8_blend_mb_outer_c(unsigned char *y, unsigned char *u, unsigned char *v,
+                          int y_1, int u_1, int v_1, int alpha, int stride) {
+  int i, j;
+  int y1_const = y_1 * ((1 << 16) - alpha);
+  int u1_const = u_1 * ((1 << 16) - alpha);
+  int v1_const = v_1 * ((1 << 16) - alpha);
 
-    for (i = 0; i < 2; i++)
-    {
-        for (j = 0; j < 16; j++)
-        {
-            y[j] = (y[j]*alpha + y1_const)>>16;
-        }
-        y += stride;
+  for (i = 0; i < 2; i++) {
+    for (j = 0; j < 16; j++) {
+      y[j] = (y[j] * alpha + y1_const) >> 16;
     }
+    y += stride;
+  }
 
-    for (i = 0; i < 12; i++)
-    {
-        y[0]  = (y[0]*alpha  + y1_const)>>16;
-        y[1]  = (y[1]*alpha  + y1_const)>>16;
-        y[14] = (y[14]*alpha + y1_const)>>16;
-        y[15] = (y[15]*alpha + y1_const)>>16;
-        y += stride;
-    }
+  for (i = 0; i < 12; i++) {
+    y[0] = (y[0] * alpha + y1_const) >> 16;
+    y[1] = (y[1] * alpha + y1_const) >> 16;
+    y[14] = (y[14] * alpha + y1_const) >> 16;
+    y[15] = (y[15] * alpha + y1_const) >> 16;
+    y += stride;
+  }
 
-    for (i = 0; i < 2; i++)
-    {
-        for (j = 0; j < 16; j++)
-        {
-            y[j] = (y[j]*alpha + y1_const)>>16;
-        }
-        y += stride;
+  for (i = 0; i < 2; i++) {
+    for (j = 0; j < 16; j++) {
+      y[j] = (y[j] * alpha + y1_const) >> 16;
     }
+    y += stride;
+  }
 
-    stride >>= 1;
+  stride >>= 1;
 
-    for (j = 0; j < 8; j++)
-    {
-        u[j] = (u[j]*alpha + u1_const)>>16;
-        v[j] = (v[j]*alpha + v1_const)>>16;
-    }
-    u += stride;
-    v += stride;
+  for (j = 0; j < 8; j++) {
+    u[j] = (u[j] * alpha + u1_const) >> 16;
+    v[j] = (v[j] * alpha + v1_const) >> 16;
+  }
+  u += stride;
+  v += stride;
 
-    for (i = 0; i < 6; i++)
-    {
-        u[0] = (u[0]*alpha + u1_const)>>16;
-        v[0] = (v[0]*alpha + v1_const)>>16;
+  for (i = 0; i < 6; i++) {
+    u[0] = (u[0] * alpha + u1_const) >> 16;
+    v[0] = (v[0] * alpha + v1_const) >> 16;
 
-        u[7] = (u[7]*alpha + u1_const)>>16;
-        v[7] = (v[7]*alpha + v1_const)>>16;
+    u[7] = (u[7] * alpha + u1_const) >> 16;
+    v[7] = (v[7] * alpha + v1_const) >> 16;
 
-        u += stride;
-        v += stride;
-    }
+    u += stride;
+    v += stride;
+  }
 
-    for (j = 0; j < 8; j++)
-    {
-        u[j] = (u[j]*alpha + u1_const)>>16;
-        v[j] = (v[j]*alpha + v1_const)>>16;
-    }
+  for (j = 0; j < 8; j++) {
+    u[j] = (u[j] * alpha + u1_const) >> 16;
+    v[j] = (v[j] * alpha + v1_const) >> 16;
+  }
 }
 
-void vp8_blend_b_c (unsigned char *y, unsigned char *u, unsigned char *v,
-                        int y_1, int u_1, int v_1, int alpha, int stride)
-{
-    int i, j;
-    int y1_const = y_1*((1<<16)-alpha);
-    int u1_const = u_1*((1<<16)-alpha);
-    int v1_const = v_1*((1<<16)-alpha);
+void vp8_blend_b_c(unsigned char *y, unsigned char *u, unsigned char *v,
+                   int y_1, int u_1, int v_1, int alpha, int stride) {
+  int i, j;
+  int y1_const = y_1 * ((1 << 16) - alpha);
+  int u1_const = u_1 * ((1 << 16) - alpha);
+  int v1_const = v_1 * ((1 << 16) - alpha);
 
-    for (i = 0; i < 4; i++)
-    {
-        for (j = 0; j < 4; j++)
-        {
-            y[j] = (y[j]*alpha + y1_const)>>16;
-        }
-        y += stride;
+  for (i = 0; i < 4; i++) {
+    for (j = 0; j < 4; j++) {
+      y[j] = (y[j] * alpha + y1_const) >> 16;
     }
+    y += stride;
+  }
 
-    stride >>= 1;
+  stride >>= 1;
 
-    for (i = 0; i < 2; i++)
-    {
-        for (j = 0; j < 2; j++)
-        {
-            u[j] = (u[j]*alpha + u1_const)>>16;
-            v[j] = (v[j]*alpha + v1_const)>>16;
-        }
-        u += stride;
-        v += stride;
+  for (i = 0; i < 2; i++) {
+    for (j = 0; j < 2; j++) {
+      u[j] = (u[j] * alpha + u1_const) >> 16;
+      v[j] = (v[j] * alpha + v1_const) >> 16;
     }
+    u += stride;
+    v += stride;
+  }
 }
 
 #if CONFIG_POSTPROC_VISUALIZER
-static void constrain_line (int x_0, int *x_1, int y_0, int *y_1, int width, int height)
-{
-    int dx;
-    int dy;
+static void constrain_line(int x_0, int *x_1, int y_0, int *y_1, int width,
+                           int height) {
+  int dx;
+  int dy;
 
-    if (*x_1 > width)
-    {
-        dx = *x_1 - x_0;
-        dy = *y_1 - y_0;
+  if (*x_1 > width) {
+    dx = *x_1 - x_0;
+    dy = *y_1 - y_0;
 
-        *x_1 = width;
-        if (dx)
-            *y_1 = ((width-x_0)*dy)/dx + y_0;
-    }
-    if (*x_1 < 0)
-    {
-        dx = *x_1 - x_0;
-        dy = *y_1 - y_0;
+    *x_1 = width;
+    if (dx) *y_1 = ((width - x_0) * dy) / dx + y_0;
+  }
+  if (*x_1 < 0) {
+    dx = *x_1 - x_0;
+    dy = *y_1 - y_0;
 
-        *x_1 = 0;
-        if (dx)
-            *y_1 = ((0-x_0)*dy)/dx + y_0;
-    }
-    if (*y_1 > height)
-    {
-        dx = *x_1 - x_0;
-        dy = *y_1 - y_0;
+    *x_1 = 0;
+    if (dx) *y_1 = ((0 - x_0) * dy) / dx + y_0;
+  }
+  if (*y_1 > height) {
+    dx = *x_1 - x_0;
+    dy = *y_1 - y_0;
 
-        *y_1 = height;
-        if (dy)
-            *x_1 = ((height-y_0)*dx)/dy + x_0;
-    }
-    if (*y_1 < 0)
-    {
-        dx = *x_1 - x_0;
-        dy = *y_1 - y_0;
+    *y_1 = height;
+    if (dy) *x_1 = ((height - y_0) * dx) / dy + x_0;
+  }
+  if (*y_1 < 0) {
+    dx = *x_1 - x_0;
+    dy = *y_1 - y_0;
 
-        *y_1 = 0;
-        if (dy)
-            *x_1 = ((0-y_0)*dx)/dy + x_0;
-    }
+    *y_1 = 0;
+    if (dy) *x_1 = ((0 - y_0) * dx) / dy + x_0;
+  }
 }
 #endif  // CONFIG_POSTPROC_VISUALIZER
 
 #if CONFIG_POSTPROC
-int vp8_post_proc_frame(VP8_COMMON *oci, YV12_BUFFER_CONFIG *dest, vp8_ppflags_t *ppflags)
-{
-    int q = oci->filter_level * 10 / 6;
-    int flags = ppflags->post_proc_flag;
-    int deblock_level = ppflags->deblocking_level;
-    int noise_level = ppflags->noise_level;
+int vp8_post_proc_frame(VP8_COMMON *oci, YV12_BUFFER_CONFIG *dest,
+                        vp8_ppflags_t *ppflags) {
+  int q = oci->filter_level * 10 / 6;
+  int flags = ppflags->post_proc_flag;
+  int deblock_level = ppflags->deblocking_level;
+  int noise_level = ppflags->noise_level;
 
-    if (!oci->frame_to_show)
-        return -1;
+  if (!oci->frame_to_show) return -1;
 
-    if (q > 63)
-        q = 63;
+  if (q > 63) q = 63;
 
-    if (!flags)
-    {
-        *dest = *oci->frame_to_show;
+  if (!flags) {
+    *dest = *oci->frame_to_show;
 
-        /* handle problem with extending borders */
-        dest->y_width = oci->Width;
-        dest->y_height = oci->Height;
-        dest->uv_height = dest->y_height / 2;
-        oci->postproc_state.last_base_qindex = oci->base_qindex;
-        oci->postproc_state.last_frame_valid = 1;
-        return 0;
+    /* handle problem with extending borders */
+    dest->y_width = oci->Width;
+    dest->y_height = oci->Height;
+    dest->uv_height = dest->y_height / 2;
+    oci->postproc_state.last_base_qindex = oci->base_qindex;
+    oci->postproc_state.last_frame_valid = 1;
+    return 0;
+  }
+  if (flags & VP8D_ADDNOISE) {
+    if (!oci->postproc_state.generated_noise) {
+      oci->postproc_state.generated_noise = vpx_calloc(
+          oci->Width + 256, sizeof(*oci->postproc_state.generated_noise));
+      if (!oci->postproc_state.generated_noise) return 1;
     }
-    if (flags & VP8D_ADDNOISE)
-    {
-      if (!oci->postproc_state.generated_noise)
-      {
-          oci->postproc_state.generated_noise = vpx_calloc(
-              oci->Width + 256, sizeof(*oci->postproc_state.generated_noise));
-          if (!oci->postproc_state.generated_noise)
-            return 1;
-      }
-    }
+  }
 
-    /* Allocate post_proc_buffer_int if needed */
-    if ((flags & VP8D_MFQE) && !oci->post_proc_buffer_int_used)
-    {
-        if ((flags & VP8D_DEBLOCK) || (flags & VP8D_DEMACROBLOCK))
-        {
-            int width = (oci->Width + 15) & ~15;
-            int height = (oci->Height + 15) & ~15;
+  /* Allocate post_proc_buffer_int if needed */
+  if ((flags & VP8D_MFQE) && !oci->post_proc_buffer_int_used) {
+    if ((flags & VP8D_DEBLOCK) || (flags & VP8D_DEMACROBLOCK)) {
+      int width = (oci->Width + 15) & ~15;
+      int height = (oci->Height + 15) & ~15;
 
-            if (vp8_yv12_alloc_frame_buffer(&oci->post_proc_buffer_int,
-                                            width, height, VP8BORDERINPIXELS))
-                vpx_internal_error(&oci->error, VPX_CODEC_MEM_ERROR,
-                                   "Failed to allocate MFQE framebuffer");
+      if (vp8_yv12_alloc_frame_buffer(&oci->post_proc_buffer_int, width, height,
+                                      VP8BORDERINPIXELS))
+        vpx_internal_error(&oci->error, VPX_CODEC_MEM_ERROR,
+                           "Failed to allocate MFQE framebuffer");
 
-            oci->post_proc_buffer_int_used = 1;
+      oci->post_proc_buffer_int_used = 1;
 
-            /* insure that postproc is set to all 0's so that post proc
-             * doesn't pull random data in from edge
-             */
-            memset((&oci->post_proc_buffer_int)->buffer_alloc,128,(&oci->post_proc_buffer)->frame_size);
-
-        }
+      /* insure that postproc is set to all 0's so that post proc
+       * doesn't pull random data in from edge
+       */
+      memset((&oci->post_proc_buffer_int)->buffer_alloc, 128,
+             (&oci->post_proc_buffer)->frame_size);
     }
+  }
 
-    vp8_clear_system_state();
+  vp8_clear_system_state();
 
-    if ((flags & VP8D_MFQE) &&
-         oci->postproc_state.last_frame_valid &&
-         oci->current_video_frame >= 2 &&
-         oci->postproc_state.last_base_qindex < 60 &&
-         oci->base_qindex - oci->postproc_state.last_base_qindex >= 20)
-    {
-        vp8_multiframe_quality_enhance(oci);
-        if (((flags & VP8D_DEBLOCK) || (flags & VP8D_DEMACROBLOCK)) &&
-            oci->post_proc_buffer_int_used)
-        {
-            vp8_yv12_copy_frame(&oci->post_proc_buffer, &oci->post_proc_buffer_int);
-            if (flags & VP8D_DEMACROBLOCK)
-            {
-                vp8_deblock(oci, &oci->post_proc_buffer_int, &oci->post_proc_buffer,
-                                               q + (deblock_level - 5) * 10, 1, 0);
-                vp8_de_mblock(&oci->post_proc_buffer,
-                              q + (deblock_level - 5) * 10);
-            }
-            else if (flags & VP8D_DEBLOCK)
-            {
-                vp8_deblock(oci, &oci->post_proc_buffer_int, &oci->post_proc_buffer,
-                            q, 1, 0);
-            }
-        }
-        /* Move partially towards the base q of the previous frame */
-        oci->postproc_state.last_base_qindex = (3*oci->postproc_state.last_base_qindex + oci->base_qindex)>>2;
-    }
-    else if (flags & VP8D_DEMACROBLOCK)
-    {
-        vp8_deblock(oci, oci->frame_to_show, &oci->post_proc_buffer,
-                                     q + (deblock_level - 5) * 10, 1, 0);
+  if ((flags & VP8D_MFQE) && oci->postproc_state.last_frame_valid &&
+      oci->current_video_frame >= 2 &&
+      oci->postproc_state.last_base_qindex < 60 &&
+      oci->base_qindex - oci->postproc_state.last_base_qindex >= 20) {
+    vp8_multiframe_quality_enhance(oci);
+    if (((flags & VP8D_DEBLOCK) || (flags & VP8D_DEMACROBLOCK)) &&
+        oci->post_proc_buffer_int_used) {
+      vp8_yv12_copy_frame(&oci->post_proc_buffer, &oci->post_proc_buffer_int);
+      if (flags & VP8D_DEMACROBLOCK) {
+        vp8_deblock(oci, &oci->post_proc_buffer_int, &oci->post_proc_buffer,
+                    q + (deblock_level - 5) * 10, 1, 0);
         vp8_de_mblock(&oci->post_proc_buffer, q + (deblock_level - 5) * 10);
-
-        oci->postproc_state.last_base_qindex = oci->base_qindex;
+      } else if (flags & VP8D_DEBLOCK) {
+        vp8_deblock(oci, &oci->post_proc_buffer_int, &oci->post_proc_buffer, q,
+                    1, 0);
+      }
     }
-    else if (flags & VP8D_DEBLOCK)
-    {
-        vp8_deblock(oci, oci->frame_to_show, &oci->post_proc_buffer,
-                    q, 1, 0);
-        oci->postproc_state.last_base_qindex = oci->base_qindex;
-    }
-    else
-    {
-        vp8_yv12_copy_frame(oci->frame_to_show, &oci->post_proc_buffer);
-        oci->postproc_state.last_base_qindex = oci->base_qindex;
-    }
-    oci->postproc_state.last_frame_valid = 1;
+    /* Move partially towards the base q of the previous frame */
+    oci->postproc_state.last_base_qindex =
+        (3 * oci->postproc_state.last_base_qindex + oci->base_qindex) >> 2;
+  } else if (flags & VP8D_DEMACROBLOCK) {
+    vp8_deblock(oci, oci->frame_to_show, &oci->post_proc_buffer,
+                q + (deblock_level - 5) * 10, 1, 0);
+    vp8_de_mblock(&oci->post_proc_buffer, q + (deblock_level - 5) * 10);
 
-    if (flags & VP8D_ADDNOISE)
-    {
-        if (oci->postproc_state.last_q != q
-            || oci->postproc_state.last_noise != noise_level)
-        {
-            double sigma;
-            struct postproc_state *ppstate = &oci->postproc_state;
-            vp8_clear_system_state();
-            sigma = noise_level + .5 + .6 * q / 63.0;
-            ppstate->clamp = vpx_setup_noise(sigma, ppstate->generated_noise,
-                                             oci->Width + 256);
-            ppstate->last_q = q;
-            ppstate->last_noise = noise_level;
-        }
+    oci->postproc_state.last_base_qindex = oci->base_qindex;
+  } else if (flags & VP8D_DEBLOCK) {
+    vp8_deblock(oci, oci->frame_to_show, &oci->post_proc_buffer, q, 1, 0);
+    oci->postproc_state.last_base_qindex = oci->base_qindex;
+  } else {
+    vp8_yv12_copy_frame(oci->frame_to_show, &oci->post_proc_buffer);
+    oci->postproc_state.last_base_qindex = oci->base_qindex;
+  }
+  oci->postproc_state.last_frame_valid = 1;
 
-        vpx_plane_add_noise
-        (oci->post_proc_buffer.y_buffer,
-         oci->postproc_state.generated_noise,
-         oci->postproc_state.clamp,
-         oci->postproc_state.clamp,
-         oci->post_proc_buffer.y_width, oci->post_proc_buffer.y_height,
-         oci->post_proc_buffer.y_stride);
+  if (flags & VP8D_ADDNOISE) {
+    if (oci->postproc_state.last_q != q ||
+        oci->postproc_state.last_noise != noise_level) {
+      double sigma;
+      struct postproc_state *ppstate = &oci->postproc_state;
+      vp8_clear_system_state();
+      sigma = noise_level + .5 + .6 * q / 63.0;
+      ppstate->clamp =
+          vpx_setup_noise(sigma, ppstate->generated_noise, oci->Width + 256);
+      ppstate->last_q = q;
+      ppstate->last_noise = noise_level;
     }
 
+    vpx_plane_add_noise(
+        oci->post_proc_buffer.y_buffer, oci->postproc_state.generated_noise,
+        oci->postproc_state.clamp, oci->postproc_state.clamp,
+        oci->post_proc_buffer.y_width, oci->post_proc_buffer.y_height,
+        oci->post_proc_buffer.y_stride);
+  }
+
 #if CONFIG_POSTPROC_VISUALIZER
-    if (flags & VP8D_DEBUG_TXT_FRAME_INFO)
-    {
-        char message[512];
-        sprintf(message, "F%1dG%1dQ%3dF%3dP%d_s%dx%d",
-                (oci->frame_type == KEY_FRAME),
-                oci->refresh_golden_frame,
-                oci->base_qindex,
-                oci->filter_level,
-                flags,
-                oci->mb_cols, oci->mb_rows);
-        vp8_blit_text(message, oci->post_proc_buffer.y_buffer, oci->post_proc_buffer.y_stride);
-    }
+  if (flags & VP8D_DEBUG_TXT_FRAME_INFO) {
+    char message[512];
+    sprintf(message, "F%1dG%1dQ%3dF%3dP%d_s%dx%d",
+            (oci->frame_type == KEY_FRAME), oci->refresh_golden_frame,
+            oci->base_qindex, oci->filter_level, flags, oci->mb_cols,
+            oci->mb_rows);
+    vp8_blit_text(message, oci->post_proc_buffer.y_buffer,
+                  oci->post_proc_buffer.y_stride);
+  }
 
-    if (flags & VP8D_DEBUG_TXT_MBLK_MODES)
-    {
-        int i, j;
-        unsigned char *y_ptr;
-        YV12_BUFFER_CONFIG *post = &oci->post_proc_buffer;
-        int mb_rows = post->y_height >> 4;
-        int mb_cols = post->y_width  >> 4;
-        int mb_index = 0;
-        MODE_INFO *mi = oci->mi;
+  if (flags & VP8D_DEBUG_TXT_MBLK_MODES) {
+    int i, j;
+    unsigned char *y_ptr;
+    YV12_BUFFER_CONFIG *post = &oci->post_proc_buffer;
+    int mb_rows = post->y_height >> 4;
+    int mb_cols = post->y_width >> 4;
+    int mb_index = 0;
+    MODE_INFO *mi = oci->mi;
 
-        y_ptr = post->y_buffer + 4 * post->y_stride + 4;
+    y_ptr = post->y_buffer + 4 * post->y_stride + 4;
 
-        /* vp8_filter each macro block */
-        for (i = 0; i < mb_rows; i++)
-        {
-            for (j = 0; j < mb_cols; j++)
-            {
-                char zz[4];
+    /* vp8_filter each macro block */
+    for (i = 0; i < mb_rows; i++) {
+      for (j = 0; j < mb_cols; j++) {
+        char zz[4];
 
-                sprintf(zz, "%c", mi[mb_index].mbmi.mode + 'a');
+        sprintf(zz, "%c", mi[mb_index].mbmi.mode + 'a');
 
-                vp8_blit_text(zz, y_ptr, post->y_stride);
-                mb_index ++;
-                y_ptr += 16;
-            }
+        vp8_blit_text(zz, y_ptr, post->y_stride);
+        mb_index++;
+        y_ptr += 16;
+      }
 
-            mb_index ++; /* border */
-            y_ptr += post->y_stride  * 16 - post->y_width;
-
-        }
+      mb_index++; /* border */
+      y_ptr += post->y_stride * 16 - post->y_width;
     }
+  }
 
-    if (flags & VP8D_DEBUG_TXT_DC_DIFF)
-    {
-        int i, j;
-        unsigned char *y_ptr;
-        YV12_BUFFER_CONFIG *post = &oci->post_proc_buffer;
-        int mb_rows = post->y_height >> 4;
-        int mb_cols = post->y_width  >> 4;
-        int mb_index = 0;
-        MODE_INFO *mi = oci->mi;
+  if (flags & VP8D_DEBUG_TXT_DC_DIFF) {
+    int i, j;
+    unsigned char *y_ptr;
+    YV12_BUFFER_CONFIG *post = &oci->post_proc_buffer;
+    int mb_rows = post->y_height >> 4;
+    int mb_cols = post->y_width >> 4;
+    int mb_index = 0;
+    MODE_INFO *mi = oci->mi;
 
-        y_ptr = post->y_buffer + 4 * post->y_stride + 4;
+    y_ptr = post->y_buffer + 4 * post->y_stride + 4;
 
-        /* vp8_filter each macro block */
-        for (i = 0; i < mb_rows; i++)
-        {
-            for (j = 0; j < mb_cols; j++)
-            {
-                char zz[4];
-                int dc_diff = !(mi[mb_index].mbmi.mode != B_PRED &&
-                              mi[mb_index].mbmi.mode != SPLITMV &&
-                              mi[mb_index].mbmi.mb_skip_coeff);
+    /* vp8_filter each macro block */
+    for (i = 0; i < mb_rows; i++) {
+      for (j = 0; j < mb_cols; j++) {
+        char zz[4];
+        int dc_diff = !(mi[mb_index].mbmi.mode != B_PRED &&
+                        mi[mb_index].mbmi.mode != SPLITMV &&
+                        mi[mb_index].mbmi.mb_skip_coeff);
 
-                if (oci->frame_type == KEY_FRAME)
-                    sprintf(zz, "a");
-                else
-                    sprintf(zz, "%c", dc_diff + '0');
+        if (oci->frame_type == KEY_FRAME)
+          sprintf(zz, "a");
+        else
+          sprintf(zz, "%c", dc_diff + '0');
 
-                vp8_blit_text(zz, y_ptr, post->y_stride);
-                mb_index ++;
-                y_ptr += 16;
-            }
+        vp8_blit_text(zz, y_ptr, post->y_stride);
+        mb_index++;
+        y_ptr += 16;
+      }
 
-            mb_index ++; /* border */
-            y_ptr += post->y_stride  * 16 - post->y_width;
-
-        }
+      mb_index++; /* border */
+      y_ptr += post->y_stride * 16 - post->y_width;
     }
+  }
 
-    if (flags & VP8D_DEBUG_TXT_RATE_INFO)
-    {
-        char message[512];
-        sprintf(message, "Bitrate: %10.2f framerate: %10.2f ", oci->bitrate, oci->framerate);
-        vp8_blit_text(message, oci->post_proc_buffer.y_buffer, oci->post_proc_buffer.y_stride);
-    }
+  if (flags & VP8D_DEBUG_TXT_RATE_INFO) {
+    char message[512];
+    sprintf(message, "Bitrate: %10.2f framerate: %10.2f ", oci->bitrate,
+            oci->framerate);
+    vp8_blit_text(message, oci->post_proc_buffer.y_buffer,
+                  oci->post_proc_buffer.y_stride);
+  }
 
-    /* Draw motion vectors */
-    if ((flags & VP8D_DEBUG_DRAW_MV) && ppflags->display_mv_flag)
-    {
-        YV12_BUFFER_CONFIG *post = &oci->post_proc_buffer;
-        int width  = post->y_width;
-        int height = post->y_height;
-        unsigned char *y_buffer = oci->post_proc_buffer.y_buffer;
-        int y_stride = oci->post_proc_buffer.y_stride;
-        MODE_INFO *mi = oci->mi;
-        int x0, y0;
+  /* Draw motion vectors */
+  if ((flags & VP8D_DEBUG_DRAW_MV) && ppflags->display_mv_flag) {
+    YV12_BUFFER_CONFIG *post = &oci->post_proc_buffer;
+    int width = post->y_width;
+    int height = post->y_height;
+    unsigned char *y_buffer = oci->post_proc_buffer.y_buffer;
+    int y_stride = oci->post_proc_buffer.y_stride;
+    MODE_INFO *mi = oci->mi;
+    int x0, y0;
 
-        for (y0 = 0; y0 < height; y0 += 16)
-        {
-            for (x0 = 0; x0 < width; x0 += 16)
-            {
-                int x1, y1;
+    for (y0 = 0; y0 < height; y0 += 16) {
+      for (x0 = 0; x0 < width; x0 += 16) {
+        int x1, y1;
 
-                if (!(ppflags->display_mv_flag & (1<<mi->mbmi.mode)))
-                {
-                    mi++;
-                    continue;
-                }
+        if (!(ppflags->display_mv_flag & (1 << mi->mbmi.mode))) {
+          mi++;
+          continue;
+        }
 
-                if (mi->mbmi.mode == SPLITMV)
-                {
-                    switch (mi->mbmi.partitioning)
-                    {
-                        case 0 :    /* mv_top_bottom */
-                        {
-                            union b_mode_info *bmi = &mi->bmi[0];
-                            MV *mv = &bmi->mv.as_mv;
+        if (mi->mbmi.mode == SPLITMV) {
+          switch (mi->mbmi.partitioning) {
+            case 0: /* mv_top_bottom */
+            {
+              union b_mode_info *bmi = &mi->bmi[0];
+              MV *mv = &bmi->mv.as_mv;
 
-                            x1 = x0 + 8 + (mv->col >> 3);
-                            y1 = y0 + 4 + (mv->row >> 3);
+              x1 = x0 + 8 + (mv->col >> 3);
+              y1 = y0 + 4 + (mv->row >> 3);
 
-                            constrain_line (x0+8, &x1, y0+4, &y1, width, height);
-                            vp8_blit_line  (x0+8,  x1, y0+4,  y1, y_buffer, y_stride);
+              constrain_line(x0 + 8, &x1, y0 + 4, &y1, width, height);
+              vp8_blit_line(x0 + 8, x1, y0 + 4, y1, y_buffer, y_stride);
 
-                            bmi = &mi->bmi[8];
+              bmi = &mi->bmi[8];
 
-                            x1 = x0 + 8 + (mv->col >> 3);
-                            y1 = y0 +12 + (mv->row >> 3);
+              x1 = x0 + 8 + (mv->col >> 3);
+              y1 = y0 + 12 + (mv->row >> 3);
 
-                            constrain_line (x0+8, &x1, y0+12, &y1, width, height);
-                            vp8_blit_line  (x0+8,  x1, y0+12,  y1, y_buffer, y_stride);
+              constrain_line(x0 + 8, &x1, y0 + 12, &y1, width, height);
+              vp8_blit_line(x0 + 8, x1, y0 + 12, y1, y_buffer, y_stride);
 
-                            break;
-                        }
-                        case 1 :    /* mv_left_right */
-                        {
-                            union b_mode_info *bmi = &mi->bmi[0];
-                            MV *mv = &bmi->mv.as_mv;
+              break;
+            }
+            case 1: /* mv_left_right */
+            {
+              union b_mode_info *bmi = &mi->bmi[0];
+              MV *mv = &bmi->mv.as_mv;
 
-                            x1 = x0 + 4 + (mv->col >> 3);
-                            y1 = y0 + 8 + (mv->row >> 3);
+              x1 = x0 + 4 + (mv->col >> 3);
+              y1 = y0 + 8 + (mv->row >> 3);
 
-                            constrain_line (x0+4, &x1, y0+8, &y1, width, height);
-                            vp8_blit_line  (x0+4,  x1, y0+8,  y1, y_buffer, y_stride);
+              constrain_line(x0 + 4, &x1, y0 + 8, &y1, width, height);
+              vp8_blit_line(x0 + 4, x1, y0 + 8, y1, y_buffer, y_stride);
 
-                            bmi = &mi->bmi[2];
+              bmi = &mi->bmi[2];
 
-                            x1 = x0 +12 + (mv->col >> 3);
-                            y1 = y0 + 8 + (mv->row >> 3);
+              x1 = x0 + 12 + (mv->col >> 3);
+              y1 = y0 + 8 + (mv->row >> 3);
 
-                            constrain_line (x0+12, &x1, y0+8, &y1, width, height);
-                            vp8_blit_line  (x0+12,  x1, y0+8,  y1, y_buffer, y_stride);
+              constrain_line(x0 + 12, &x1, y0 + 8, &y1, width, height);
+              vp8_blit_line(x0 + 12, x1, y0 + 8, y1, y_buffer, y_stride);
 
-                            break;
-                        }
-                        case 2 :    /* mv_quarters   */
-                        {
-                            union b_mode_info *bmi = &mi->bmi[0];
-                            MV *mv = &bmi->mv.as_mv;
+              break;
+            }
+            case 2: /* mv_quarters   */
+            {
+              union b_mode_info *bmi = &mi->bmi[0];
+              MV *mv = &bmi->mv.as_mv;
 
-                            x1 = x0 + 4 + (mv->col >> 3);
-                            y1 = y0 + 4 + (mv->row >> 3);
+              x1 = x0 + 4 + (mv->col >> 3);
+              y1 = y0 + 4 + (mv->row >> 3);
 
-                            constrain_line (x0+4, &x1, y0+4, &y1, width, height);
-                            vp8_blit_line  (x0+4,  x1, y0+4,  y1, y_buffer, y_stride);
+              constrain_line(x0 + 4, &x1, y0 + 4, &y1, width, height);
+              vp8_blit_line(x0 + 4, x1, y0 + 4, y1, y_buffer, y_stride);
 
-                            bmi = &mi->bmi[2];
+              bmi = &mi->bmi[2];
 
-                            x1 = x0 +12 + (mv->col >> 3);
-                            y1 = y0 + 4 + (mv->row >> 3);
+              x1 = x0 + 12 + (mv->col >> 3);
+              y1 = y0 + 4 + (mv->row >> 3);
 
-                            constrain_line (x0+12, &x1, y0+4, &y1, width, height);
-                            vp8_blit_line  (x0+12,  x1, y0+4,  y1, y_buffer, y_stride);
+              constrain_line(x0 + 12, &x1, y0 + 4, &y1, width, height);
+              vp8_blit_line(x0 + 12, x1, y0 + 4, y1, y_buffer, y_stride);
 
-                            bmi = &mi->bmi[8];
+              bmi = &mi->bmi[8];
 
-                            x1 = x0 + 4 + (mv->col >> 3);
-                            y1 = y0 +12 + (mv->row >> 3);
+              x1 = x0 + 4 + (mv->col >> 3);
+              y1 = y0 + 12 + (mv->row >> 3);
 
-                            constrain_line (x0+4, &x1, y0+12, &y1, width, height);
-                            vp8_blit_line  (x0+4,  x1, y0+12,  y1, y_buffer, y_stride);
+              constrain_line(x0 + 4, &x1, y0 + 12, &y1, width, height);
+              vp8_blit_line(x0 + 4, x1, y0 + 12, y1, y_buffer, y_stride);
 
-                            bmi = &mi->bmi[10];
+              bmi = &mi->bmi[10];
 
-                            x1 = x0 +12 + (mv->col >> 3);
-                            y1 = y0 +12 + (mv->row >> 3);
+              x1 = x0 + 12 + (mv->col >> 3);
+              y1 = y0 + 12 + (mv->row >> 3);
 
-                            constrain_line (x0+12, &x1, y0+12, &y1, width, height);
-                            vp8_blit_line  (x0+12,  x1, y0+12,  y1, y_buffer, y_stride);
-                            break;
-                        }
-                        default :
-                        {
-                            union b_mode_info *bmi = mi->bmi;
-                            int bx0, by0;
+              constrain_line(x0 + 12, &x1, y0 + 12, &y1, width, height);
+              vp8_blit_line(x0 + 12, x1, y0 + 12, y1, y_buffer, y_stride);
+              break;
+            }
+            default: {
+              union b_mode_info *bmi = mi->bmi;
+              int bx0, by0;
 
-                            for (by0 = y0; by0 < (y0+16); by0 += 4)
-                            {
-                                for (bx0 = x0; bx0 < (x0+16); bx0 += 4)
-                                {
-                                    MV *mv = &bmi->mv.as_mv;
+              for (by0 = y0; by0 < (y0 + 16); by0 += 4) {
+                for (bx0 = x0; bx0 < (x0 + 16); bx0 += 4) {
+                  MV *mv = &bmi->mv.as_mv;
 
-                                    x1 = bx0 + 2 + (mv->col >> 3);
-                                    y1 = by0 + 2 + (mv->row >> 3);
+                  x1 = bx0 + 2 + (mv->col >> 3);
+                  y1 = by0 + 2 + (mv->row >> 3);
 
-                                    constrain_line (bx0+2, &x1, by0+2, &y1, width, height);
-                                    vp8_blit_line  (bx0+2,  x1, by0+2,  y1, y_buffer, y_stride);
+                  constrain_line(bx0 + 2, &x1, by0 + 2, &y1, width, height);
+                  vp8_blit_line(bx0 + 2, x1, by0 + 2, y1, y_buffer, y_stride);
 
-                                    bmi++;
-                                }
-                            }
-                        }
-                    }
-                }
-                else if (mi->mbmi.mode >= NEARESTMV)
-                {
-                    MV *mv = &mi->mbmi.mv.as_mv;
-                    const int lx0 = x0 + 8;
-                    const int ly0 = y0 + 8;
+                  bmi++;
+                }
+              }
+            }
+          }
+        } else if (mi->mbmi.mode >= NEARESTMV) {
+          MV *mv = &mi->mbmi.mv.as_mv;
+          const int lx0 = x0 + 8;
+          const int ly0 = y0 + 8;
 
-                    x1 = lx0 + (mv->col >> 3);
-                    y1 = ly0 + (mv->row >> 3);
+          x1 = lx0 + (mv->col >> 3);
+          y1 = ly0 + (mv->row >> 3);
 
-                    if (x1 != lx0 && y1 != ly0)
-                    {
-                        constrain_line (lx0, &x1, ly0-1, &y1, width, height);
-                        vp8_blit_line  (lx0,  x1, ly0-1,  y1, y_buffer, y_stride);
+          if (x1 != lx0 && y1 != ly0) {
+            constrain_line(lx0, &x1, ly0 - 1, &y1, width, height);
+            vp8_blit_line(lx0, x1, ly0 - 1, y1, y_buffer, y_stride);
 
-                        constrain_line (lx0, &x1, ly0+1, &y1, width, height);
-                        vp8_blit_line  (lx0,  x1, ly0+1,  y1, y_buffer, y_stride);
-                    }
-                    else
-                        vp8_blit_line  (lx0,  x1, ly0,  y1, y_buffer, y_stride);
-                }
-
-                mi++;
-            }
-            mi++;
+            constrain_line(lx0, &x1, ly0 + 1, &y1, width, height);
+            vp8_blit_line(lx0, x1, ly0 + 1, y1, y_buffer, y_stride);
+          } else
+            vp8_blit_line(lx0, x1, ly0, y1, y_buffer, y_stride);
         }
+
+        mi++;
+      }
+      mi++;
     }
+  }
 
-    /* Color in block modes */
-    if ((flags & VP8D_DEBUG_CLR_BLK_MODES)
-        && (ppflags->display_mb_modes_flag || ppflags->display_b_modes_flag))
-    {
-        int y, x;
-        YV12_BUFFER_CONFIG *post = &oci->post_proc_buffer;
-        int width  = post->y_width;
-        int height = post->y_height;
-        unsigned char *y_ptr = oci->post_proc_buffer.y_buffer;
-        unsigned char *u_ptr = oci->post_proc_buffer.u_buffer;
-        unsigned char *v_ptr = oci->post_proc_buffer.v_buffer;
-        int y_stride = oci->post_proc_buffer.y_stride;
-        MODE_INFO *mi = oci->mi;
+  /* Color in block modes */
+  if ((flags & VP8D_DEBUG_CLR_BLK_MODES) &&
+      (ppflags->display_mb_modes_flag || ppflags->display_b_modes_flag)) {
+    int y, x;
+    YV12_BUFFER_CONFIG *post = &oci->post_proc_buffer;
+    int width = post->y_width;
+    int height = post->y_height;
+    unsigned char *y_ptr = oci->post_proc_buffer.y_buffer;
+    unsigned char *u_ptr = oci->post_proc_buffer.u_buffer;
+    unsigned char *v_ptr = oci->post_proc_buffer.v_buffer;
+    int y_stride = oci->post_proc_buffer.y_stride;
+    MODE_INFO *mi = oci->mi;
 
-        for (y = 0; y < height; y += 16)
-        {
-            for (x = 0; x < width; x += 16)
-            {
-                int Y = 0, U = 0, V = 0;
+    for (y = 0; y < height; y += 16) {
+      for (x = 0; x < width; x += 16) {
+        int Y = 0, U = 0, V = 0;
 
-                if (mi->mbmi.mode == B_PRED &&
-                    ((ppflags->display_mb_modes_flag & B_PRED) || ppflags->display_b_modes_flag))
-                {
-                    int by, bx;
-                    unsigned char *yl, *ul, *vl;
-                    union b_mode_info *bmi = mi->bmi;
+        if (mi->mbmi.mode == B_PRED &&
+            ((ppflags->display_mb_modes_flag & B_PRED) ||
+             ppflags->display_b_modes_flag)) {
+          int by, bx;
+          unsigned char *yl, *ul, *vl;
+          union b_mode_info *bmi = mi->bmi;
 
-                    yl = y_ptr + x;
-                    ul = u_ptr + (x>>1);
-                    vl = v_ptr + (x>>1);
+          yl = y_ptr + x;
+          ul = u_ptr + (x >> 1);
+          vl = v_ptr + (x >> 1);
 
-                    for (by = 0; by < 16; by += 4)
-                    {
-                        for (bx = 0; bx < 16; bx += 4)
-                        {
-                            if ((ppflags->display_b_modes_flag & (1<<mi->mbmi.mode))
-                                || (ppflags->display_mb_modes_flag & B_PRED))
-                            {
-                                Y = B_PREDICTION_MODE_colors[bmi->as_mode][0];
-                                U = B_PREDICTION_MODE_colors[bmi->as_mode][1];
-                                V = B_PREDICTION_MODE_colors[bmi->as_mode][2];
+          for (by = 0; by < 16; by += 4) {
+            for (bx = 0; bx < 16; bx += 4) {
+              if ((ppflags->display_b_modes_flag & (1 << mi->mbmi.mode)) ||
+                  (ppflags->display_mb_modes_flag & B_PRED)) {
+                Y = B_PREDICTION_MODE_colors[bmi->as_mode][0];
+                U = B_PREDICTION_MODE_colors[bmi->as_mode][1];
+                V = B_PREDICTION_MODE_colors[bmi->as_mode][2];
 
-                                vp8_blend_b
-                                    (yl+bx, ul+(bx>>1), vl+(bx>>1), Y, U, V, 0xc000, y_stride);
-                            }
-                            bmi++;
-                        }
+                vp8_blend_b(yl + bx, ul + (bx >> 1), vl + (bx >> 1), Y, U, V,
+                            0xc000, y_stride);
+              }
+              bmi++;
+            }
 
-                        yl += y_stride*4;
-                        ul += y_stride*1;
-                        vl += y_stride*1;
-                    }
-                }
-                else if (ppflags->display_mb_modes_flag & (1<<mi->mbmi.mode))
-                {
-                    Y = MB_PREDICTION_MODE_colors[mi->mbmi.mode][0];
-                    U = MB_PREDICTION_MODE_colors[mi->mbmi.mode][1];
-                    V = MB_PREDICTION_MODE_colors[mi->mbmi.mode][2];
+            yl += y_stride * 4;
+            ul += y_stride * 1;
+            vl += y_stride * 1;
+          }
+        } else if (ppflags->display_mb_modes_flag & (1 << mi->mbmi.mode)) {
+          Y = MB_PREDICTION_MODE_colors[mi->mbmi.mode][0];
+          U = MB_PREDICTION_MODE_colors[mi->mbmi.mode][1];
+          V = MB_PREDICTION_MODE_colors[mi->mbmi.mode][2];
 
-                    vp8_blend_mb_inner
-                        (y_ptr+x, u_ptr+(x>>1), v_ptr+(x>>1), Y, U, V, 0xc000, y_stride);
-                }
+          vp8_blend_mb_inner(y_ptr + x, u_ptr + (x >> 1), v_ptr + (x >> 1), Y,
+                             U, V, 0xc000, y_stride);
+        }
 
-                mi++;
-            }
-            y_ptr += y_stride*16;
-            u_ptr += y_stride*4;
-            v_ptr += y_stride*4;
+        mi++;
+      }
+      y_ptr += y_stride * 16;
+      u_ptr += y_stride * 4;
+      v_ptr += y_stride * 4;
 
-            mi++;
-        }
+      mi++;
     }
+  }
 
-    /* Color in frame reference blocks */
-    if ((flags & VP8D_DEBUG_CLR_FRM_REF_BLKS) && ppflags->display_ref_frame_flag)
-    {
-        int y, x;
-        YV12_BUFFER_CONFIG *post = &oci->post_proc_buffer;
-        int width  = post->y_width;
-        int height = post->y_height;
-        unsigned char *y_ptr = oci->post_proc_buffer.y_buffer;
-        unsigned char *u_ptr = oci->post_proc_buffer.u_buffer;
-        unsigned char *v_ptr = oci->post_proc_buffer.v_buffer;
-        int y_stride = oci->post_proc_buffer.y_stride;
-        MODE_INFO *mi = oci->mi;
+  /* Color in frame reference blocks */
+  if ((flags & VP8D_DEBUG_CLR_FRM_REF_BLKS) &&
+      ppflags->display_ref_frame_flag) {
+    int y, x;
+    YV12_BUFFER_CONFIG *post = &oci->post_proc_buffer;
+    int width = post->y_width;
+    int height = post->y_height;
+    unsigned char *y_ptr = oci->post_proc_buffer.y_buffer;
+    unsigned char *u_ptr = oci->post_proc_buffer.u_buffer;
+    unsigned char *v_ptr = oci->post_proc_buffer.v_buffer;
+    int y_stride = oci->post_proc_buffer.y_stride;
+    MODE_INFO *mi = oci->mi;
 
-        for (y = 0; y < height; y += 16)
-        {
-            for (x = 0; x < width; x +=16)
-            {
-                int Y = 0, U = 0, V = 0;
+    for (y = 0; y < height; y += 16) {
+      for (x = 0; x < width; x += 16) {
+        int Y = 0, U = 0, V = 0;
 
-                if (ppflags->display_ref_frame_flag & (1<<mi->mbmi.ref_frame))
-                {
-                    Y = MV_REFERENCE_FRAME_colors[mi->mbmi.ref_frame][0];
-                    U = MV_REFERENCE_FRAME_colors[mi->mbmi.ref_frame][1];
-                    V = MV_REFERENCE_FRAME_colors[mi->mbmi.ref_frame][2];
+        if (ppflags->display_ref_frame_flag & (1 << mi->mbmi.ref_frame)) {
+          Y = MV_REFERENCE_FRAME_colors[mi->mbmi.ref_frame][0];
+          U = MV_REFERENCE_FRAME_colors[mi->mbmi.ref_frame][1];
+          V = MV_REFERENCE_FRAME_colors[mi->mbmi.ref_frame][2];
 
-                    vp8_blend_mb_outer
-                        (y_ptr+x, u_ptr+(x>>1), v_ptr+(x>>1), Y, U, V, 0xc000, y_stride);
-                }
+          vp8_blend_mb_outer(y_ptr + x, u_ptr + (x >> 1), v_ptr + (x >> 1), Y,
+                             U, V, 0xc000, y_stride);
+        }
 
-                mi++;
-            }
-            y_ptr += y_stride*16;
-            u_ptr += y_stride*4;
-            v_ptr += y_stride*4;
+        mi++;
+      }
+      y_ptr += y_stride * 16;
+      u_ptr += y_stride * 4;
+      v_ptr += y_stride * 4;
 
-            mi++;
-        }
+      mi++;
     }
+  }
 #endif
 
-    *dest = oci->post_proc_buffer;
+  *dest = oci->post_proc_buffer;
 
-    /* handle problem with extending borders */
-    dest->y_width = oci->Width;
-    dest->y_height = oci->Height;
-    dest->uv_height = dest->y_height / 2;
-    return 0;
+  /* handle problem with extending borders */
+  dest->y_width = oci->Width;
+  dest->y_height = oci->Height;
+  dest->uv_height = dest->y_height / 2;
+  return 0;
 }
 #endif
--- a/vp8/common/postproc.h
+++ b/vp8/common/postproc.h
@@ -8,19 +8,17 @@
  *  be found in the AUTHORS file in the root of the source tree.
  */
 
-
 #ifndef VP8_COMMON_POSTPROC_H_
 #define VP8_COMMON_POSTPROC_H_
 
 #include "vpx_ports/mem.h"
-struct postproc_state
-{
-    int           last_q;
-    int           last_noise;
-    int           last_base_qindex;
-    int           last_frame_valid;
-    int           clamp;
-    int8_t       *generated_noise;
+struct postproc_state {
+  int last_q;
+  int last_noise;
+  int last_base_qindex;
+  int last_frame_valid;
+  int clamp;
+  int8_t *generated_noise;
 };
 #include "onyxc_int.h"
 #include "ppflags.h"
@@ -31,21 +29,12 @@
 int vp8_post_proc_frame(struct VP8Common *oci, YV12_BUFFER_CONFIG *dest,
                         vp8_ppflags_t *flags);
 
+void vp8_de_noise(struct VP8Common *oci, YV12_BUFFER_CONFIG *source,
+                  YV12_BUFFER_CONFIG *post, int q, int low_var_thresh, int flag,
+                  int uvfilter);
 
-void vp8_de_noise(struct VP8Common           *oci,
-                  YV12_BUFFER_CONFIG         *source,
-                  YV12_BUFFER_CONFIG         *post,
-                  int                         q,
-                  int                         low_var_thresh,
-                  int                         flag,
-                  int                         uvfilter);
-
-void vp8_deblock(struct VP8Common           *oci,
-                 YV12_BUFFER_CONFIG         *source,
-                 YV12_BUFFER_CONFIG         *post,
-                 int                         q,
-                 int                         low_var_thresh,
-                 int                         flag);
+void vp8_deblock(struct VP8Common *oci, YV12_BUFFER_CONFIG *source,
+                 YV12_BUFFER_CONFIG *post, int q, int low_var_thresh, int flag);
 
 #define MFQE_PRECISION 4
 
--- a/vp8/common/ppflags.h
+++ b/vp8/common/ppflags.h
@@ -8,7 +8,6 @@
  *  be found in the AUTHORS file in the root of the source tree.
  */
 
-
 #ifndef VP8_COMMON_PPFLAGS_H_
 #define VP8_COMMON_PPFLAGS_H_
 
@@ -15,31 +14,29 @@
 #ifdef __cplusplus
 extern "C" {
 #endif
-enum
-{
-    VP8D_NOFILTERING            = 0,
-    VP8D_DEBLOCK                = 1<<0,
-    VP8D_DEMACROBLOCK           = 1<<1,
-    VP8D_ADDNOISE               = 1<<2,
-    VP8D_DEBUG_TXT_FRAME_INFO   = 1<<3,
-    VP8D_DEBUG_TXT_MBLK_MODES   = 1<<4,
-    VP8D_DEBUG_TXT_DC_DIFF      = 1<<5,
-    VP8D_DEBUG_TXT_RATE_INFO    = 1<<6,
-    VP8D_DEBUG_DRAW_MV          = 1<<7,
-    VP8D_DEBUG_CLR_BLK_MODES    = 1<<8,
-    VP8D_DEBUG_CLR_FRM_REF_BLKS = 1<<9,
-    VP8D_MFQE                   = 1<<10
+enum {
+  VP8D_NOFILTERING = 0,
+  VP8D_DEBLOCK = 1 << 0,
+  VP8D_DEMACROBLOCK = 1 << 1,
+  VP8D_ADDNOISE = 1 << 2,
+  VP8D_DEBUG_TXT_FRAME_INFO = 1 << 3,
+  VP8D_DEBUG_TXT_MBLK_MODES = 1 << 4,
+  VP8D_DEBUG_TXT_DC_DIFF = 1 << 5,
+  VP8D_DEBUG_TXT_RATE_INFO = 1 << 6,
+  VP8D_DEBUG_DRAW_MV = 1 << 7,
+  VP8D_DEBUG_CLR_BLK_MODES = 1 << 8,
+  VP8D_DEBUG_CLR_FRM_REF_BLKS = 1 << 9,
+  VP8D_MFQE = 1 << 10
 };
 
-typedef struct
-{
-    int post_proc_flag;
-    int deblocking_level;
-    int noise_level;
-    int display_ref_frame_flag;
-    int display_mb_modes_flag;
-    int display_b_modes_flag;
-    int display_mv_flag;
+typedef struct {
+  int post_proc_flag;
+  int deblocking_level;
+  int noise_level;
+  int display_ref_frame_flag;
+  int display_mb_modes_flag;
+  int display_b_modes_flag;
+  int display_mv_flag;
 } vp8_ppflags_t;
 
 #ifdef __cplusplus
--- a/vp8/common/quant_common.c
+++ b/vp8/common/quant_common.c
@@ -8,128 +8,117 @@
  *  be found in the AUTHORS file in the root of the source tree.
  */
 
-
 #include "quant_common.h"
 
-static const int dc_qlookup[QINDEX_RANGE] =
-{
-    4,    5,    6,    7,    8,    9,   10,   10,   11,   12,   13,   14,   15,   16,   17,   17,
-    18,   19,   20,   20,   21,   21,   22,   22,   23,   23,   24,   25,   25,   26,   27,   28,
-    29,   30,   31,   32,   33,   34,   35,   36,   37,   37,   38,   39,   40,   41,   42,   43,
-    44,   45,   46,   46,   47,   48,   49,   50,   51,   52,   53,   54,   55,   56,   57,   58,
-    59,   60,   61,   62,   63,   64,   65,   66,   67,   68,   69,   70,   71,   72,   73,   74,
-    75,   76,   76,   77,   78,   79,   80,   81,   82,   83,   84,   85,   86,   87,   88,   89,
-    91,   93,   95,   96,   98,  100,  101,  102,  104,  106,  108,  110,  112,  114,  116,  118,
-    122,  124,  126,  128,  130,  132,  134,  136,  138,  140,  143,  145,  148,  151,  154,  157,
+static const int dc_qlookup[QINDEX_RANGE] = {
+  4,   5,   6,   7,   8,   9,   10,  10,  11,  12,  13,  14,  15,  16,  17,
+  17,  18,  19,  20,  20,  21,  21,  22,  22,  23,  23,  24,  25,  25,  26,
+  27,  28,  29,  30,  31,  32,  33,  34,  35,  36,  37,  37,  38,  39,  40,
+  41,  42,  43,  44,  45,  46,  46,  47,  48,  49,  50,  51,  52,  53,  54,
+  55,  56,  57,  58,  59,  60,  61,  62,  63,  64,  65,  66,  67,  68,  69,
+  70,  71,  72,  73,  74,  75,  76,  76,  77,  78,  79,  80,  81,  82,  83,
+  84,  85,  86,  87,  88,  89,  91,  93,  95,  96,  98,  100, 101, 102, 104,
+  106, 108, 110, 112, 114, 116, 118, 122, 124, 126, 128, 130, 132, 134, 136,
+  138, 140, 143, 145, 148, 151, 154, 157,
 };
 
-static const int ac_qlookup[QINDEX_RANGE] =
-{
-    4,    5,    6,    7,    8,    9,   10,   11,   12,   13,   14,   15,   16,   17,   18,   19,
-    20,   21,   22,   23,   24,   25,   26,   27,   28,   29,   30,   31,   32,   33,   34,   35,
-    36,   37,   38,   39,   40,   41,   42,   43,   44,   45,   46,   47,   48,   49,   50,   51,
-    52,   53,   54,   55,   56,   57,   58,   60,   62,   64,   66,   68,   70,   72,   74,   76,
-    78,   80,   82,   84,   86,   88,   90,   92,   94,   96,   98,  100,  102,  104,  106,  108,
-    110,  112,  114,  116,  119,  122,  125,  128,  131,  134,  137,  140,  143,  146,  149,  152,
-    155,  158,  161,  164,  167,  170,  173,  177,  181,  185,  189,  193,  197,  201,  205,  209,
-    213,  217,  221,  225,  229,  234,  239,  245,  249,  254,  259,  264,  269,  274,  279,  284,
+static const int ac_qlookup[QINDEX_RANGE] = {
+  4,   5,   6,   7,   8,   9,   10,  11,  12,  13,  14,  15,  16,  17,  18,
+  19,  20,  21,  22,  23,  24,  25,  26,  27,  28,  29,  30,  31,  32,  33,
+  34,  35,  36,  37,  38,  39,  40,  41,  42,  43,  44,  45,  46,  47,  48,
+  49,  50,  51,  52,  53,  54,  55,  56,  57,  58,  60,  62,  64,  66,  68,
+  70,  72,  74,  76,  78,  80,  82,  84,  86,  88,  90,  92,  94,  96,  98,
+  100, 102, 104, 106, 108, 110, 112, 114, 116, 119, 122, 125, 128, 131, 134,
+  137, 140, 143, 146, 149, 152, 155, 158, 161, 164, 167, 170, 173, 177, 181,
+  185, 189, 193, 197, 201, 205, 209, 213, 217, 221, 225, 229, 234, 239, 245,
+  249, 254, 259, 264, 269, 274, 279, 284,
 };
 
+int vp8_dc_quant(int QIndex, int Delta) {
+  int retval;
 
-int vp8_dc_quant(int QIndex, int Delta)
-{
-    int retval;
+  QIndex = QIndex + Delta;
 
-    QIndex = QIndex + Delta;
+  if (QIndex > 127)
+    QIndex = 127;
+  else if (QIndex < 0)
+    QIndex = 0;
 
-    if (QIndex > 127)
-        QIndex = 127;
-    else if (QIndex < 0)
-        QIndex = 0;
-
-    retval = dc_qlookup[ QIndex ];
-    return retval;
+  retval = dc_qlookup[QIndex];
+  return retval;
 }
 
-int vp8_dc2quant(int QIndex, int Delta)
-{
-    int retval;
+int vp8_dc2quant(int QIndex, int Delta) {
+  int retval;
 
-    QIndex = QIndex + Delta;
+  QIndex = QIndex + Delta;
 
-    if (QIndex > 127)
-        QIndex = 127;
-    else if (QIndex < 0)
-        QIndex = 0;
+  if (QIndex > 127)
+    QIndex = 127;
+  else if (QIndex < 0)
+    QIndex = 0;
 
-    retval = dc_qlookup[ QIndex ] * 2;
-    return retval;
-
+  retval = dc_qlookup[QIndex] * 2;
+  return retval;
 }
-int vp8_dc_uv_quant(int QIndex, int Delta)
-{
-    int retval;
+int vp8_dc_uv_quant(int QIndex, int Delta) {
+  int retval;
 
-    QIndex = QIndex + Delta;
+  QIndex = QIndex + Delta;
 
-    if (QIndex > 127)
-        QIndex = 127;
-    else if (QIndex < 0)
-        QIndex = 0;
+  if (QIndex > 127)
+    QIndex = 127;
+  else if (QIndex < 0)
+    QIndex = 0;
 
-    retval = dc_qlookup[ QIndex ];
+  retval = dc_qlookup[QIndex];
 
-    if (retval > 132)
-        retval = 132;
+  if (retval > 132) retval = 132;
 
-    return retval;
+  return retval;
 }
 
-int vp8_ac_yquant(int QIndex)
-{
-    int retval;
+int vp8_ac_yquant(int QIndex) {
+  int retval;
 
-    if (QIndex > 127)
-        QIndex = 127;
-    else if (QIndex < 0)
-        QIndex = 0;
+  if (QIndex > 127)
+    QIndex = 127;
+  else if (QIndex < 0)
+    QIndex = 0;
 
-    retval = ac_qlookup[ QIndex ];
-    return retval;
+  retval = ac_qlookup[QIndex];
+  return retval;
 }
 
-int vp8_ac2quant(int QIndex, int Delta)
-{
-    int retval;
+int vp8_ac2quant(int QIndex, int Delta) {
+  int retval;
 
-    QIndex = QIndex + Delta;
+  QIndex = QIndex + Delta;
 
-    if (QIndex > 127)
-        QIndex = 127;
-    else if (QIndex < 0)
-        QIndex = 0;
+  if (QIndex > 127)
+    QIndex = 127;
+  else if (QIndex < 0)
+    QIndex = 0;
 
-    /* For all x in [0..284], x*155/100 is bitwise equal to (x*101581) >> 16.
-     * The smallest precision for that is '(x*6349) >> 12' but 16 is a good
-     * word size. */
-    retval = (ac_qlookup[ QIndex ] * 101581) >> 16;
+  /* For all x in [0..284], x*155/100 is bitwise equal to (x*101581) >> 16.
+   * The smallest precision for that is '(x*6349) >> 12' but 16 is a good
+   * word size. */
+  retval = (ac_qlookup[QIndex] * 101581) >> 16;
 
-    if (retval < 8)
-        retval = 8;
+  if (retval < 8) retval = 8;
 
-    return retval;
+  return retval;
 }
-int vp8_ac_uv_quant(int QIndex, int Delta)
-{
-    int retval;
+int vp8_ac_uv_quant(int QIndex, int Delta) {
+  int retval;
 
-    QIndex = QIndex + Delta;
+  QIndex = QIndex + Delta;
 
-    if (QIndex > 127)
-        QIndex = 127;
-    else if (QIndex < 0)
-        QIndex = 0;
+  if (QIndex > 127)
+    QIndex = 127;
+  else if (QIndex < 0)
+    QIndex = 0;
 
-    retval = ac_qlookup[ QIndex ];
-    return retval;
+  retval = ac_qlookup[QIndex];
+  return retval;
 }
--- a/vp8/common/quant_common.h
+++ b/vp8/common/quant_common.h
@@ -11,7 +11,6 @@
 #ifndef VP8_COMMON_QUANT_COMMON_H_
 #define VP8_COMMON_QUANT_COMMON_H_
 
-
 #include "string.h"
 #include "blockd.h"
 #include "onyxc_int.h"
--- a/vp8/common/reconinter.c
+++ b/vp8/common/reconinter.c
@@ -8,7 +8,6 @@
  *  be found in the AUTHORS file in the root of the source tree.
  */
 
-
 #include <limits.h>
 #include <string.h>
 
@@ -21,524 +20,474 @@
 #include "onyxc_int.h"
 #endif
 
-void vp8_copy_mem16x16_c(
-    unsigned char *src,
-    int src_stride,
-    unsigned char *dst,
-    int dst_stride)
-{
+void vp8_copy_mem16x16_c(unsigned char *src, int src_stride, unsigned char *dst,
+                         int dst_stride) {
+  int r;
 
-    int r;
+  for (r = 0; r < 16; r++) {
+    memcpy(dst, src, 16);
 
-    for (r = 0; r < 16; r++)
-    {
-        memcpy(dst, src, 16);
-
-        src += src_stride;
-        dst += dst_stride;
-
-    }
-
+    src += src_stride;
+    dst += dst_stride;
+  }
 }
 
-void vp8_copy_mem8x8_c(
-    unsigned char *src,
-    int src_stride,
-    unsigned char *dst,
-    int dst_stride)
-{
-    int r;
+void vp8_copy_mem8x8_c(unsigned char *src, int src_stride, unsigned char *dst,
+                       int dst_stride) {
+  int r;
 
-    for (r = 0; r < 8; r++)
-    {
-        memcpy(dst, src, 8);
+  for (r = 0; r < 8; r++) {
+    memcpy(dst, src, 8);
 
-        src += src_stride;
-        dst += dst_stride;
-
-    }
-
+    src += src_stride;
+    dst += dst_stride;
+  }
 }
 
-void vp8_copy_mem8x4_c(
-    unsigned char *src,
-    int src_stride,
-    unsigned char *dst,
-    int dst_stride)
-{
-    int r;
+void vp8_copy_mem8x4_c(unsigned char *src, int src_stride, unsigned char *dst,
+                       int dst_stride) {
+  int r;
 
-    for (r = 0; r < 4; r++)
-    {
-        memcpy(dst, src, 8);
+  for (r = 0; r < 4; r++) {
+    memcpy(dst, src, 8);
 
-        src += src_stride;
-        dst += dst_stride;
-
-    }
-
+    src += src_stride;
+    dst += dst_stride;
+  }
 }
 
+void vp8_build_inter_predictors_b(BLOCKD *d, int pitch, unsigned char *base_pre,
+                                  int pre_stride, vp8_subpix_fn_t sppf) {
+  int r;
+  unsigned char *pred_ptr = d->predictor;
+  unsigned char *ptr;
+  ptr = base_pre + d->offset + (d->bmi.mv.as_mv.row >> 3) * pre_stride +
+        (d->bmi.mv.as_mv.col >> 3);
 
-void vp8_build_inter_predictors_b(BLOCKD *d, int pitch, unsigned char *base_pre, int pre_stride, vp8_subpix_fn_t sppf)
-{
-    int r;
-    unsigned char *pred_ptr = d->predictor;
-    unsigned char *ptr;
-    ptr = base_pre + d->offset + (d->bmi.mv.as_mv.row >> 3) * pre_stride + (d->bmi.mv.as_mv.col >> 3);
-
-    if (d->bmi.mv.as_mv.row & 7 || d->bmi.mv.as_mv.col & 7)
-    {
-        sppf(ptr, pre_stride, d->bmi.mv.as_mv.col & 7, d->bmi.mv.as_mv.row & 7, pred_ptr, pitch);
+  if (d->bmi.mv.as_mv.row & 7 || d->bmi.mv.as_mv.col & 7) {
+    sppf(ptr, pre_stride, d->bmi.mv.as_mv.col & 7, d->bmi.mv.as_mv.row & 7,
+         pred_ptr, pitch);
+  } else {
+    for (r = 0; r < 4; r++) {
+      pred_ptr[0] = ptr[0];
+      pred_ptr[1] = ptr[1];
+      pred_ptr[2] = ptr[2];
+      pred_ptr[3] = ptr[3];
+      pred_ptr += pitch;
+      ptr += pre_stride;
     }
-    else
-    {
-        for (r = 0; r < 4; r++)
-        {
-            pred_ptr[0]  = ptr[0];
-            pred_ptr[1]  = ptr[1];
-            pred_ptr[2]  = ptr[2];
-            pred_ptr[3]  = ptr[3];
-            pred_ptr     += pitch;
-            ptr         += pre_stride;
-        }
-    }
+  }
 }
 
-static void build_inter_predictors4b(MACROBLOCKD *x, BLOCKD *d, unsigned char *dst, int dst_stride, unsigned char *base_pre, int pre_stride)
-{
-    unsigned char *ptr;
-    ptr = base_pre + d->offset + (d->bmi.mv.as_mv.row >> 3) * pre_stride + (d->bmi.mv.as_mv.col >> 3);
+static void build_inter_predictors4b(MACROBLOCKD *x, BLOCKD *d,
+                                     unsigned char *dst, int dst_stride,
+                                     unsigned char *base_pre, int pre_stride) {
+  unsigned char *ptr;
+  ptr = base_pre + d->offset + (d->bmi.mv.as_mv.row >> 3) * pre_stride +
+        (d->bmi.mv.as_mv.col >> 3);
 
-    if (d->bmi.mv.as_mv.row & 7 || d->bmi.mv.as_mv.col & 7)
-    {
-        x->subpixel_predict8x8(ptr, pre_stride, d->bmi.mv.as_mv.col & 7, d->bmi.mv.as_mv.row & 7, dst, dst_stride);
-    }
-    else
-    {
-        vp8_copy_mem8x8(ptr, pre_stride, dst, dst_stride);
-    }
+  if (d->bmi.mv.as_mv.row & 7 || d->bmi.mv.as_mv.col & 7) {
+    x->subpixel_predict8x8(ptr, pre_stride, d->bmi.mv.as_mv.col & 7,
+                           d->bmi.mv.as_mv.row & 7, dst, dst_stride);
+  } else {
+    vp8_copy_mem8x8(ptr, pre_stride, dst, dst_stride);
+  }
 }
 
-static void build_inter_predictors2b(MACROBLOCKD *x, BLOCKD *d, unsigned char *dst, int dst_stride, unsigned char *base_pre, int pre_stride)
-{
-    unsigned char *ptr;
-    ptr = base_pre + d->offset + (d->bmi.mv.as_mv.row >> 3) * pre_stride + (d->bmi.mv.as_mv.col >> 3);
+static void build_inter_predictors2b(MACROBLOCKD *x, BLOCKD *d,
+                                     unsigned char *dst, int dst_stride,
+                                     unsigned char *base_pre, int pre_stride) {
+  unsigned char *ptr;
+  ptr = base_pre + d->offset + (d->bmi.mv.as_mv.row >> 3) * pre_stride +
+        (d->bmi.mv.as_mv.col >> 3);
 
-    if (d->bmi.mv.as_mv.row & 7 || d->bmi.mv.as_mv.col & 7)
-    {
-        x->subpixel_predict8x4(ptr, pre_stride, d->bmi.mv.as_mv.col & 7, d->bmi.mv.as_mv.row & 7, dst, dst_stride);
-    }
-    else
-    {
-        vp8_copy_mem8x4(ptr, pre_stride, dst, dst_stride);
-    }
+  if (d->bmi.mv.as_mv.row & 7 || d->bmi.mv.as_mv.col & 7) {
+    x->subpixel_predict8x4(ptr, pre_stride, d->bmi.mv.as_mv.col & 7,
+                           d->bmi.mv.as_mv.row & 7, dst, dst_stride);
+  } else {
+    vp8_copy_mem8x4(ptr, pre_stride, dst, dst_stride);
+  }
 }
 
-static void build_inter_predictors_b(BLOCKD *d, unsigned char *dst, int dst_stride, unsigned char *base_pre, int pre_stride, vp8_subpix_fn_t sppf)
-{
-    int r;
-    unsigned char *ptr;
-    ptr = base_pre + d->offset + (d->bmi.mv.as_mv.row >> 3) * pre_stride + (d->bmi.mv.as_mv.col >> 3);
+static void build_inter_predictors_b(BLOCKD *d, unsigned char *dst,
+                                     int dst_stride, unsigned char *base_pre,
+                                     int pre_stride, vp8_subpix_fn_t sppf) {
+  int r;
+  unsigned char *ptr;
+  ptr = base_pre + d->offset + (d->bmi.mv.as_mv.row >> 3) * pre_stride +
+        (d->bmi.mv.as_mv.col >> 3);
 
-    if (d->bmi.mv.as_mv.row & 7 || d->bmi.mv.as_mv.col & 7)
-    {
-        sppf(ptr, pre_stride, d->bmi.mv.as_mv.col & 7, d->bmi.mv.as_mv.row & 7, dst, dst_stride);
+  if (d->bmi.mv.as_mv.row & 7 || d->bmi.mv.as_mv.col & 7) {
+    sppf(ptr, pre_stride, d->bmi.mv.as_mv.col & 7, d->bmi.mv.as_mv.row & 7, dst,
+         dst_stride);
+  } else {
+    for (r = 0; r < 4; r++) {
+      dst[0] = ptr[0];
+      dst[1] = ptr[1];
+      dst[2] = ptr[2];
+      dst[3] = ptr[3];
+      dst += dst_stride;
+      ptr += pre_stride;
     }
-    else
-    {
-        for (r = 0; r < 4; r++)
-        {
-          dst[0]  = ptr[0];
-          dst[1]  = ptr[1];
-          dst[2]  = ptr[2];
-          dst[3]  = ptr[3];
-          dst     += dst_stride;
-          ptr     += pre_stride;
-        }
-    }
+  }
 }
 
-
 /*encoder only*/
-void vp8_build_inter16x16_predictors_mbuv(MACROBLOCKD *x)
-{
-    unsigned char *uptr, *vptr;
-    unsigned char *upred_ptr = &x->predictor[256];
-    unsigned char *vpred_ptr = &x->predictor[320];
+void vp8_build_inter16x16_predictors_mbuv(MACROBLOCKD *x) {
+  unsigned char *uptr, *vptr;
+  unsigned char *upred_ptr = &x->predictor[256];
+  unsigned char *vpred_ptr = &x->predictor[320];
 
-    int mv_row = x->mode_info_context->mbmi.mv.as_mv.row;
-    int mv_col = x->mode_info_context->mbmi.mv.as_mv.col;
-    int offset;
-    int pre_stride = x->pre.uv_stride;
+  int mv_row = x->mode_info_context->mbmi.mv.as_mv.row;
+  int mv_col = x->mode_info_context->mbmi.mv.as_mv.col;
+  int offset;
+  int pre_stride = x->pre.uv_stride;
 
-    /* calc uv motion vectors */
-    mv_row += 1 | (mv_row >> (sizeof(int) * CHAR_BIT - 1));
-    mv_col += 1 | (mv_col >> (sizeof(int) * CHAR_BIT - 1));
-    mv_row /= 2;
-    mv_col /= 2;
-    mv_row &= x->fullpixel_mask;
-    mv_col &= x->fullpixel_mask;
+  /* calc uv motion vectors */
+  mv_row += 1 | (mv_row >> (sizeof(int) * CHAR_BIT - 1));
+  mv_col += 1 | (mv_col >> (sizeof(int) * CHAR_BIT - 1));
+  mv_row /= 2;
+  mv_col /= 2;
+  mv_row &= x->fullpixel_mask;
+  mv_col &= x->fullpixel_mask;
 
-    offset = (mv_row >> 3) * pre_stride + (mv_col >> 3);
-    uptr = x->pre.u_buffer + offset;
-    vptr = x->pre.v_buffer + offset;
+  offset = (mv_row >> 3) * pre_stride + (mv_col >> 3);
+  uptr = x->pre.u_buffer + offset;
+  vptr = x->pre.v_buffer + offset;
 
-    if ((mv_row | mv_col) & 7)
-    {
-        x->subpixel_predict8x8(uptr, pre_stride, mv_col & 7, mv_row & 7, upred_ptr, 8);
-        x->subpixel_predict8x8(vptr, pre_stride, mv_col & 7, mv_row & 7, vpred_ptr, 8);
-    }
-    else
-    {
-        vp8_copy_mem8x8(uptr, pre_stride, upred_ptr, 8);
-        vp8_copy_mem8x8(vptr, pre_stride, vpred_ptr, 8);
-    }
+  if ((mv_row | mv_col) & 7) {
+    x->subpixel_predict8x8(uptr, pre_stride, mv_col & 7, mv_row & 7, upred_ptr,
+                           8);
+    x->subpixel_predict8x8(vptr, pre_stride, mv_col & 7, mv_row & 7, vpred_ptr,
+                           8);
+  } else {
+    vp8_copy_mem8x8(uptr, pre_stride, upred_ptr, 8);
+    vp8_copy_mem8x8(vptr, pre_stride, vpred_ptr, 8);
+  }
 }
 
 /*encoder only*/
-void vp8_build_inter4x4_predictors_mbuv(MACROBLOCKD *x)
-{
-    int i, j;
-    int pre_stride = x->pre.uv_stride;
-    unsigned char *base_pre;
+void vp8_build_inter4x4_predictors_mbuv(MACROBLOCKD *x) {
+  int i, j;
+  int pre_stride = x->pre.uv_stride;
+  unsigned char *base_pre;
 
-    /* build uv mvs */
-    for (i = 0; i < 2; i++)
-    {
-        for (j = 0; j < 2; j++)
-        {
-            int yoffset = i * 8 + j * 2;
-            int uoffset = 16 + i * 2 + j;
-            int voffset = 20 + i * 2 + j;
+  /* build uv mvs */
+  for (i = 0; i < 2; i++) {
+    for (j = 0; j < 2; j++) {
+      int yoffset = i * 8 + j * 2;
+      int uoffset = 16 + i * 2 + j;
+      int voffset = 20 + i * 2 + j;
 
-            int temp;
+      int temp;
 
-            temp = x->block[yoffset  ].bmi.mv.as_mv.row
-                   + x->block[yoffset+1].bmi.mv.as_mv.row
-                   + x->block[yoffset+4].bmi.mv.as_mv.row
-                   + x->block[yoffset+5].bmi.mv.as_mv.row;
+      temp = x->block[yoffset].bmi.mv.as_mv.row +
+             x->block[yoffset + 1].bmi.mv.as_mv.row +
+             x->block[yoffset + 4].bmi.mv.as_mv.row +
+             x->block[yoffset + 5].bmi.mv.as_mv.row;
 
-            temp += 4 + ((temp >> (sizeof(temp) * CHAR_BIT - 1)) * 8);
+      temp += 4 + ((temp >> (sizeof(temp) * CHAR_BIT - 1)) * 8);
 
-            x->block[uoffset].bmi.mv.as_mv.row = (temp / 8) & x->fullpixel_mask;
+      x->block[uoffset].bmi.mv.as_mv.row = (temp / 8) & x->fullpixel_mask;
 
-            temp = x->block[yoffset  ].bmi.mv.as_mv.col
-                   + x->block[yoffset+1].bmi.mv.as_mv.col
-                   + x->block[yoffset+4].bmi.mv.as_mv.col
-                   + x->block[yoffset+5].bmi.mv.as_mv.col;
+      temp = x->block[yoffset].bmi.mv.as_mv.col +
+             x->block[yoffset + 1].bmi.mv.as_mv.col +
+             x->block[yoffset + 4].bmi.mv.as_mv.col +
+             x->block[yoffset + 5].bmi.mv.as_mv.col;
 
-            temp += 4 + ((temp >> (sizeof(temp) * CHAR_BIT - 1)) * 8);
+      temp += 4 + ((temp >> (sizeof(temp) * CHAR_BIT - 1)) * 8);
 
-            x->block[uoffset].bmi.mv.as_mv.col = (temp / 8) & x->fullpixel_mask;
+      x->block[uoffset].bmi.mv.as_mv.col = (temp / 8) & x->fullpixel_mask;
 
-            x->block[voffset].bmi.mv.as_int = x->block[uoffset].bmi.mv.as_int;
-        }
+      x->block[voffset].bmi.mv.as_int = x->block[uoffset].bmi.mv.as_int;
     }
+  }
 
-    base_pre = x->pre.u_buffer;
-    for (i = 16; i < 20; i += 2)
-    {
-        BLOCKD *d0 = &x->block[i];
-        BLOCKD *d1 = &x->block[i+1];
+  base_pre = x->pre.u_buffer;
+  for (i = 16; i < 20; i += 2) {
+    BLOCKD *d0 = &x->block[i];
+    BLOCKD *d1 = &x->block[i + 1];
 
-        if (d0->bmi.mv.as_int == d1->bmi.mv.as_int)
-            build_inter_predictors2b(x, d0, d0->predictor, 8, base_pre, pre_stride);
-        else
-        {
-            vp8_build_inter_predictors_b(d0, 8, base_pre, pre_stride, x->subpixel_predict);
-            vp8_build_inter_predictors_b(d1, 8, base_pre, pre_stride, x->subpixel_predict);
-        }
+    if (d0->bmi.mv.as_int == d1->bmi.mv.as_int)
+      build_inter_predictors2b(x, d0, d0->predictor, 8, base_pre, pre_stride);
+    else {
+      vp8_build_inter_predictors_b(d0, 8, base_pre, pre_stride,
+                                   x->subpixel_predict);
+      vp8_build_inter_predictors_b(d1, 8, base_pre, pre_stride,
+                                   x->subpixel_predict);
     }
+  }
 
-    base_pre = x->pre.v_buffer;
-    for (i = 20; i < 24; i += 2)
-    {
-        BLOCKD *d0 = &x->block[i];
-        BLOCKD *d1 = &x->block[i+1];
+  base_pre = x->pre.v_buffer;
+  for (i = 20; i < 24; i += 2) {
+    BLOCKD *d0 = &x->block[i];
+    BLOCKD *d1 = &x->block[i + 1];
 
-        if (d0->bmi.mv.as_int == d1->bmi.mv.as_int)
-            build_inter_predictors2b(x, d0, d0->predictor, 8, base_pre, pre_stride);
-        else
-        {
-            vp8_build_inter_predictors_b(d0, 8, base_pre, pre_stride, x->subpixel_predict);
-            vp8_build_inter_predictors_b(d1, 8, base_pre, pre_stride, x->subpixel_predict);
-        }
+    if (d0->bmi.mv.as_int == d1->bmi.mv.as_int)
+      build_inter_predictors2b(x, d0, d0->predictor, 8, base_pre, pre_stride);
+    else {
+      vp8_build_inter_predictors_b(d0, 8, base_pre, pre_stride,
+                                   x->subpixel_predict);
+      vp8_build_inter_predictors_b(d1, 8, base_pre, pre_stride,
+                                   x->subpixel_predict);
     }
+  }
 }
 
-
 /*encoder only*/
-void vp8_build_inter16x16_predictors_mby(MACROBLOCKD *x,
-                                         unsigned char *dst_y,
-                                         int dst_ystride)
-{
-    unsigned char *ptr_base;
-    unsigned char *ptr;
-    int mv_row = x->mode_info_context->mbmi.mv.as_mv.row;
-    int mv_col = x->mode_info_context->mbmi.mv.as_mv.col;
-    int pre_stride = x->pre.y_stride;
+void vp8_build_inter16x16_predictors_mby(MACROBLOCKD *x, unsigned char *dst_y,
+                                         int dst_ystride) {
+  unsigned char *ptr_base;
+  unsigned char *ptr;
+  int mv_row = x->mode_info_context->mbmi.mv.as_mv.row;
+  int mv_col = x->mode_info_context->mbmi.mv.as_mv.col;
+  int pre_stride = x->pre.y_stride;
 
-    ptr_base = x->pre.y_buffer;
-    ptr = ptr_base + (mv_row >> 3) * pre_stride + (mv_col >> 3);
+  ptr_base = x->pre.y_buffer;
+  ptr = ptr_base + (mv_row >> 3) * pre_stride + (mv_col >> 3);
 
-    if ((mv_row | mv_col) & 7)
-    {
-        x->subpixel_predict16x16(ptr, pre_stride, mv_col & 7, mv_row & 7,
-                                 dst_y, dst_ystride);
-    }
-    else
-    {
-        vp8_copy_mem16x16(ptr, pre_stride, dst_y,
-            dst_ystride);
-    }
+  if ((mv_row | mv_col) & 7) {
+    x->subpixel_predict16x16(ptr, pre_stride, mv_col & 7, mv_row & 7, dst_y,
+                             dst_ystride);
+  } else {
+    vp8_copy_mem16x16(ptr, pre_stride, dst_y, dst_ystride);
+  }
 }
 
-static void clamp_mv_to_umv_border(MV *mv, const MACROBLOCKD *xd)
-{
-    /* If the MV points so far into the UMV border that no visible pixels
-     * are used for reconstruction, the subpel part of the MV can be
-     * discarded and the MV limited to 16 pixels with equivalent results.
-     *
-     * This limit kicks in at 19 pixels for the top and left edges, for
-     * the 16 pixels plus 3 taps right of the central pixel when subpel
-     * filtering. The bottom and right edges use 16 pixels plus 2 pixels
-     * left of the central pixel when filtering.
-     */
-    if (mv->col < (xd->mb_to_left_edge - (19 << 3)))
-        mv->col = xd->mb_to_left_edge - (16 << 3);
-    else if (mv->col > xd->mb_to_right_edge + (18 << 3))
-        mv->col = xd->mb_to_right_edge + (16 << 3);
+static void clamp_mv_to_umv_border(MV *mv, const MACROBLOCKD *xd) {
+  /* If the MV points so far into the UMV border that no visible pixels
+   * are used for reconstruction, the subpel part of the MV can be
+   * discarded and the MV limited to 16 pixels with equivalent results.
+   *
+   * This limit kicks in at 19 pixels for the top and left edges, for
+   * the 16 pixels plus 3 taps right of the central pixel when subpel
+   * filtering. The bottom and right edges use 16 pixels plus 2 pixels
+   * left of the central pixel when filtering.
+   */
+  if (mv->col < (xd->mb_to_left_edge - (19 << 3)))
+    mv->col = xd->mb_to_left_edge - (16 << 3);
+  else if (mv->col > xd->mb_to_right_edge + (18 << 3))
+    mv->col = xd->mb_to_right_edge + (16 << 3);
 
-    if (mv->row < (xd->mb_to_top_edge - (19 << 3)))
-        mv->row = xd->mb_to_top_edge - (16 << 3);
-    else if (mv->row > xd->mb_to_bottom_edge + (18 << 3))
-        mv->row = xd->mb_to_bottom_edge + (16 << 3);
+  if (mv->row < (xd->mb_to_top_edge - (19 << 3)))
+    mv->row = xd->mb_to_top_edge - (16 << 3);
+  else if (mv->row > xd->mb_to_bottom_edge + (18 << 3))
+    mv->row = xd->mb_to_bottom_edge + (16 << 3);
 }
 
 /* A version of the above function for chroma block MVs.*/
-static void clamp_uvmv_to_umv_border(MV *mv, const MACROBLOCKD *xd)
-{
-    mv->col = (2*mv->col < (xd->mb_to_left_edge - (19 << 3))) ?
-        (xd->mb_to_left_edge - (16 << 3)) >> 1 : mv->col;
-    mv->col = (2*mv->col > xd->mb_to_right_edge + (18 << 3)) ?
-        (xd->mb_to_right_edge + (16 << 3)) >> 1 : mv->col;
+static void clamp_uvmv_to_umv_border(MV *mv, const MACROBLOCKD *xd) {
+  mv->col = (2 * mv->col < (xd->mb_to_left_edge - (19 << 3)))
+                ? (xd->mb_to_left_edge - (16 << 3)) >> 1
+                : mv->col;
+  mv->col = (2 * mv->col > xd->mb_to_right_edge + (18 << 3))
+                ? (xd->mb_to_right_edge + (16 << 3)) >> 1
+                : mv->col;
 
-    mv->row = (2*mv->row < (xd->mb_to_top_edge - (19 << 3))) ?
-        (xd->mb_to_top_edge - (16 << 3)) >> 1 : mv->row;
-    mv->row = (2*mv->row > xd->mb_to_bottom_edge + (18 << 3)) ?
-        (xd->mb_to_bottom_edge + (16 << 3)) >> 1 : mv->row;
+  mv->row = (2 * mv->row < (xd->mb_to_top_edge - (19 << 3)))
+                ? (xd->mb_to_top_edge - (16 << 3)) >> 1
+                : mv->row;
+  mv->row = (2 * mv->row > xd->mb_to_bottom_edge + (18 << 3))
+                ? (xd->mb_to_bottom_edge + (16 << 3)) >> 1
+                : mv->row;
 }
 
-void vp8_build_inter16x16_predictors_mb(MACROBLOCKD *x,
-                                        unsigned char *dst_y,
+void vp8_build_inter16x16_predictors_mb(MACROBLOCKD *x, unsigned char *dst_y,
                                         unsigned char *dst_u,
-                                        unsigned char *dst_v,
-                                        int dst_ystride,
-                                        int dst_uvstride)
-{
-    int offset;
-    unsigned char *ptr;
-    unsigned char *uptr, *vptr;
+                                        unsigned char *dst_v, int dst_ystride,
+                                        int dst_uvstride) {
+  int offset;
+  unsigned char *ptr;
+  unsigned char *uptr, *vptr;
 
-    int_mv _16x16mv;
+  int_mv _16x16mv;
 
-    unsigned char *ptr_base = x->pre.y_buffer;
-    int pre_stride = x->pre.y_stride;
+  unsigned char *ptr_base = x->pre.y_buffer;
+  int pre_stride = x->pre.y_stride;
 
-    _16x16mv.as_int = x->mode_info_context->mbmi.mv.as_int;
+  _16x16mv.as_int = x->mode_info_context->mbmi.mv.as_int;
 
-    if (x->mode_info_context->mbmi.need_to_clamp_mvs)
-    {
-        clamp_mv_to_umv_border(&_16x16mv.as_mv, x);
-    }
+  if (x->mode_info_context->mbmi.need_to_clamp_mvs) {
+    clamp_mv_to_umv_border(&_16x16mv.as_mv, x);
+  }
 
-    ptr = ptr_base + ( _16x16mv.as_mv.row >> 3) * pre_stride + (_16x16mv.as_mv.col >> 3);
+  ptr = ptr_base + (_16x16mv.as_mv.row >> 3) * pre_stride +
+        (_16x16mv.as_mv.col >> 3);
 
-    if ( _16x16mv.as_int & 0x00070007)
-    {
-        x->subpixel_predict16x16(ptr, pre_stride, _16x16mv.as_mv.col & 7,  _16x16mv.as_mv.row & 7, dst_y, dst_ystride);
-    }
-    else
-    {
-        vp8_copy_mem16x16(ptr, pre_stride, dst_y, dst_ystride);
-    }
+  if (_16x16mv.as_int & 0x00070007) {
+    x->subpixel_predict16x16(ptr, pre_stride, _16x16mv.as_mv.col & 7,
+                             _16x16mv.as_mv.row & 7, dst_y, dst_ystride);
+  } else {
+    vp8_copy_mem16x16(ptr, pre_stride, dst_y, dst_ystride);
+  }
 
-    /* calc uv motion vectors */
-    _16x16mv.as_mv.row += 1 | (_16x16mv.as_mv.row >> (sizeof(int) * CHAR_BIT - 1));
-    _16x16mv.as_mv.col += 1 | (_16x16mv.as_mv.col >> (sizeof(int) * CHAR_BIT - 1));
-    _16x16mv.as_mv.row /= 2;
-    _16x16mv.as_mv.col /= 2;
-    _16x16mv.as_mv.row &= x->fullpixel_mask;
-    _16x16mv.as_mv.col &= x->fullpixel_mask;
+  /* calc uv motion vectors */
+  _16x16mv.as_mv.row +=
+      1 | (_16x16mv.as_mv.row >> (sizeof(int) * CHAR_BIT - 1));
+  _16x16mv.as_mv.col +=
+      1 | (_16x16mv.as_mv.col >> (sizeof(int) * CHAR_BIT - 1));
+  _16x16mv.as_mv.row /= 2;
+  _16x16mv.as_mv.col /= 2;
+  _16x16mv.as_mv.row &= x->fullpixel_mask;
+  _16x16mv.as_mv.col &= x->fullpixel_mask;
 
-    pre_stride >>= 1;
-    offset = ( _16x16mv.as_mv.row >> 3) * pre_stride + (_16x16mv.as_mv.col >> 3);
-    uptr = x->pre.u_buffer + offset;
-    vptr = x->pre.v_buffer + offset;
+  pre_stride >>= 1;
+  offset = (_16x16mv.as_mv.row >> 3) * pre_stride + (_16x16mv.as_mv.col >> 3);
+  uptr = x->pre.u_buffer + offset;
+  vptr = x->pre.v_buffer + offset;
 
-    if ( _16x16mv.as_int & 0x00070007)
-    {
-        x->subpixel_predict8x8(uptr, pre_stride, _16x16mv.as_mv.col & 7,  _16x16mv.as_mv.row & 7, dst_u, dst_uvstride);
-        x->subpixel_predict8x8(vptr, pre_stride, _16x16mv.as_mv.col & 7,  _16x16mv.as_mv.row & 7, dst_v, dst_uvstride);
-    }
-    else
-    {
-        vp8_copy_mem8x8(uptr, pre_stride, dst_u, dst_uvstride);
-        vp8_copy_mem8x8(vptr, pre_stride, dst_v, dst_uvstride);
-    }
+  if (_16x16mv.as_int & 0x00070007) {
+    x->subpixel_predict8x8(uptr, pre_stride, _16x16mv.as_mv.col & 7,
+                           _16x16mv.as_mv.row & 7, dst_u, dst_uvstride);
+    x->subpixel_predict8x8(vptr, pre_stride, _16x16mv.as_mv.col & 7,
+                           _16x16mv.as_mv.row & 7, dst_v, dst_uvstride);
+  } else {
+    vp8_copy_mem8x8(uptr, pre_stride, dst_u, dst_uvstride);
+    vp8_copy_mem8x8(vptr, pre_stride, dst_v, dst_uvstride);
+  }
 }
 
-static void build_inter4x4_predictors_mb(MACROBLOCKD *x)
-{
-    int i;
-    unsigned char *base_dst = x->dst.y_buffer;
-    unsigned char *base_pre = x->pre.y_buffer;
+static void build_inter4x4_predictors_mb(MACROBLOCKD *x) {
+  int i;
+  unsigned char *base_dst = x->dst.y_buffer;
+  unsigned char *base_pre = x->pre.y_buffer;
 
-    if (x->mode_info_context->mbmi.partitioning < 3)
-    {
-        BLOCKD *b;
-        int dst_stride = x->dst.y_stride;
+  if (x->mode_info_context->mbmi.partitioning < 3) {
+    BLOCKD *b;
+    int dst_stride = x->dst.y_stride;
 
-        x->block[ 0].bmi = x->mode_info_context->bmi[ 0];
-        x->block[ 2].bmi = x->mode_info_context->bmi[ 2];
-        x->block[ 8].bmi = x->mode_info_context->bmi[ 8];
-        x->block[10].bmi = x->mode_info_context->bmi[10];
-        if (x->mode_info_context->mbmi.need_to_clamp_mvs)
-        {
-            clamp_mv_to_umv_border(&x->block[ 0].bmi.mv.as_mv, x);
-            clamp_mv_to_umv_border(&x->block[ 2].bmi.mv.as_mv, x);
-            clamp_mv_to_umv_border(&x->block[ 8].bmi.mv.as_mv, x);
-            clamp_mv_to_umv_border(&x->block[10].bmi.mv.as_mv, x);
-        }
-
-        b = &x->block[ 0];
-        build_inter_predictors4b(x, b, base_dst + b->offset, dst_stride, base_pre, dst_stride);
-        b = &x->block[ 2];
-        build_inter_predictors4b(x, b, base_dst + b->offset, dst_stride, base_pre, dst_stride);
-        b = &x->block[ 8];
-        build_inter_predictors4b(x, b, base_dst + b->offset, dst_stride, base_pre, dst_stride);
-        b = &x->block[10];
-        build_inter_predictors4b(x, b, base_dst + b->offset, dst_stride, base_pre, dst_stride);
+    x->block[0].bmi = x->mode_info_context->bmi[0];
+    x->block[2].bmi = x->mode_info_context->bmi[2];
+    x->block[8].bmi = x->mode_info_context->bmi[8];
+    x->block[10].bmi = x->mode_info_context->bmi[10];
+    if (x->mode_info_context->mbmi.need_to_clamp_mvs) {
+      clamp_mv_to_umv_border(&x->block[0].bmi.mv.as_mv, x);
+      clamp_mv_to_umv_border(&x->block[2].bmi.mv.as_mv, x);
+      clamp_mv_to_umv_border(&x->block[8].bmi.mv.as_mv, x);
+      clamp_mv_to_umv_border(&x->block[10].bmi.mv.as_mv, x);
     }
-    else
-    {
-        for (i = 0; i < 16; i += 2)
-        {
-            BLOCKD *d0 = &x->block[i];
-            BLOCKD *d1 = &x->block[i+1];
-            int dst_stride = x->dst.y_stride;
 
-            x->block[i+0].bmi = x->mode_info_context->bmi[i+0];
-            x->block[i+1].bmi = x->mode_info_context->bmi[i+1];
-            if (x->mode_info_context->mbmi.need_to_clamp_mvs)
-            {
-                clamp_mv_to_umv_border(&x->block[i+0].bmi.mv.as_mv, x);
-                clamp_mv_to_umv_border(&x->block[i+1].bmi.mv.as_mv, x);
-            }
+    b = &x->block[0];
+    build_inter_predictors4b(x, b, base_dst + b->offset, dst_stride, base_pre,
+                             dst_stride);
+    b = &x->block[2];
+    build_inter_predictors4b(x, b, base_dst + b->offset, dst_stride, base_pre,
+                             dst_stride);
+    b = &x->block[8];
+    build_inter_predictors4b(x, b, base_dst + b->offset, dst_stride, base_pre,
+                             dst_stride);
+    b = &x->block[10];
+    build_inter_predictors4b(x, b, base_dst + b->offset, dst_stride, base_pre,
+                             dst_stride);
+  } else {
+    for (i = 0; i < 16; i += 2) {
+      BLOCKD *d0 = &x->block[i];
+      BLOCKD *d1 = &x->block[i + 1];
+      int dst_stride = x->dst.y_stride;
 
-            if (d0->bmi.mv.as_int == d1->bmi.mv.as_int)
-                build_inter_predictors2b(x, d0, base_dst + d0->offset, dst_stride, base_pre, dst_stride);
-            else
-            {
-                build_inter_predictors_b(d0, base_dst + d0->offset, dst_stride, base_pre, dst_stride, x->subpixel_predict);
-                build_inter_predictors_b(d1, base_dst + d1->offset, dst_stride, base_pre, dst_stride, x->subpixel_predict);
-            }
+      x->block[i + 0].bmi = x->mode_info_context->bmi[i + 0];
+      x->block[i + 1].bmi = x->mode_info_context->bmi[i + 1];
+      if (x->mode_info_context->mbmi.need_to_clamp_mvs) {
+        clamp_mv_to_umv_border(&x->block[i + 0].bmi.mv.as_mv, x);
+        clamp_mv_to_umv_border(&x->block[i + 1].bmi.mv.as_mv, x);
+      }
 
-        }
-
+      if (d0->bmi.mv.as_int == d1->bmi.mv.as_int)
+        build_inter_predictors2b(x, d0, base_dst + d0->offset, dst_stride,
+                                 base_pre, dst_stride);
+      else {
+        build_inter_predictors_b(d0, base_dst + d0->offset, dst_stride,
+                                 base_pre, dst_stride, x->subpixel_predict);
+        build_inter_predictors_b(d1, base_dst + d1->offset, dst_stride,
+                                 base_pre, dst_stride, x->subpixel_predict);
+      }
     }
-    base_dst = x->dst.u_buffer;
-    base_pre = x->pre.u_buffer;
-    for (i = 16; i < 20; i += 2)
-    {
-        BLOCKD *d0 = &x->block[i];
-        BLOCKD *d1 = &x->block[i+1];
-        int dst_stride = x->dst.uv_stride;
+  }
+  base_dst = x->dst.u_buffer;
+  base_pre = x->pre.u_buffer;
+  for (i = 16; i < 20; i += 2) {
+    BLOCKD *d0 = &x->block[i];
+    BLOCKD *d1 = &x->block[i + 1];
+    int dst_stride = x->dst.uv_stride;
 
-        /* Note: uv mvs already clamped in build_4x4uvmvs() */
+    /* Note: uv mvs already clamped in build_4x4uvmvs() */
 
-        if (d0->bmi.mv.as_int == d1->bmi.mv.as_int)
-            build_inter_predictors2b(x, d0, base_dst + d0->offset, dst_stride, base_pre, dst_stride);
-        else
-        {
-            build_inter_predictors_b(d0, base_dst + d0->offset, dst_stride, base_pre, dst_stride, x->subpixel_predict);
-            build_inter_predictors_b(d1, base_dst + d1->offset, dst_stride, base_pre, dst_stride, x->subpixel_predict);
-        }
+    if (d0->bmi.mv.as_int == d1->bmi.mv.as_int)
+      build_inter_predictors2b(x, d0, base_dst + d0->offset, dst_stride,
+                               base_pre, dst_stride);
+    else {
+      build_inter_predictors_b(d0, base_dst + d0->offset, dst_stride, base_pre,
+                               dst_stride, x->subpixel_predict);
+      build_inter_predictors_b(d1, base_dst + d1->offset, dst_stride, base_pre,
+                               dst_stride, x->subpixel_predict);
     }
+  }
 
-    base_dst = x->dst.v_buffer;
-    base_pre = x->pre.v_buffer;
-    for (i = 20; i < 24; i += 2)
-    {
-        BLOCKD *d0 = &x->block[i];
-        BLOCKD *d1 = &x->block[i+1];
-        int dst_stride = x->dst.uv_stride;
+  base_dst = x->dst.v_buffer;
+  base_pre = x->pre.v_buffer;
+  for (i = 20; i < 24; i += 2) {
+    BLOCKD *d0 = &x->block[i];
+    BLOCKD *d1 = &x->block[i + 1];
+    int dst_stride = x->dst.uv_stride;
 
-        /* Note: uv mvs already clamped in build_4x4uvmvs() */
+    /* Note: uv mvs already clamped in build_4x4uvmvs() */
 
-        if (d0->bmi.mv.as_int == d1->bmi.mv.as_int)
-            build_inter_predictors2b(x, d0, base_dst + d0->offset, dst_stride, base_pre, dst_stride);
-        else
-        {
-            build_inter_predictors_b(d0, base_dst + d0->offset, dst_stride, base_pre, dst_stride, x->subpixel_predict);
-            build_inter_predictors_b(d1, base_dst + d1->offset, dst_stride, base_pre, dst_stride, x->subpixel_predict);
-        }
+    if (d0->bmi.mv.as_int == d1->bmi.mv.as_int)
+      build_inter_predictors2b(x, d0, base_dst + d0->offset, dst_stride,
+                               base_pre, dst_stride);
+    else {
+      build_inter_predictors_b(d0, base_dst + d0->offset, dst_stride, base_pre,
+                               dst_stride, x->subpixel_predict);
+      build_inter_predictors_b(d1, base_dst + d1->offset, dst_stride, base_pre,
+                               dst_stride, x->subpixel_predict);
     }
+  }
 }
 
-static
-void build_4x4uvmvs(MACROBLOCKD *x)
-{
-    int i, j;
+static void build_4x4uvmvs(MACROBLOCKD *x) {
+  int i, j;
 
-    for (i = 0; i < 2; i++)
-    {
-        for (j = 0; j < 2; j++)
-        {
-            int yoffset = i * 8 + j * 2;
-            int uoffset = 16 + i * 2 + j;
-            int voffset = 20 + i * 2 + j;
+  for (i = 0; i < 2; i++) {
+    for (j = 0; j < 2; j++) {
+      int yoffset = i * 8 + j * 2;
+      int uoffset = 16 + i * 2 + j;
+      int voffset = 20 + i * 2 + j;
 
-            int temp;
+      int temp;
 
-            temp = x->mode_info_context->bmi[yoffset + 0].mv.as_mv.row
-                 + x->mode_info_context->bmi[yoffset + 1].mv.as_mv.row
-                 + x->mode_info_context->bmi[yoffset + 4].mv.as_mv.row
-                 + x->mode_info_context->bmi[yoffset + 5].mv.as_mv.row;
+      temp = x->mode_info_context->bmi[yoffset + 0].mv.as_mv.row +
+             x->mode_info_context->bmi[yoffset + 1].mv.as_mv.row +
+             x->mode_info_context->bmi[yoffset + 4].mv.as_mv.row +
+             x->mode_info_context->bmi[yoffset + 5].mv.as_mv.row;
 
-            temp += 4 + ((temp >> (sizeof(temp) * CHAR_BIT - 1)) * 8);
+      temp += 4 + ((temp >> (sizeof(temp) * CHAR_BIT - 1)) * 8);
 
-            x->block[uoffset].bmi.mv.as_mv.row = (temp / 8) & x->fullpixel_mask;
+      x->block[uoffset].bmi.mv.as_mv.row = (temp / 8) & x->fullpixel_mask;
 
-            temp = x->mode_info_context->bmi[yoffset + 0].mv.as_mv.col
-                 + x->mode_info_context->bmi[yoffset + 1].mv.as_mv.col
-                 + x->mode_info_context->bmi[yoffset + 4].mv.as_mv.col
-                 + x->mode_info_context->bmi[yoffset + 5].mv.as_mv.col;
+      temp = x->mode_info_context->bmi[yoffset + 0].mv.as_mv.col +
+             x->mode_info_context->bmi[yoffset + 1].mv.as_mv.col +
+             x->mode_info_context->bmi[yoffset + 4].mv.as_mv.col +
+             x->mode_info_context->bmi[yoffset + 5].mv.as_mv.col;
 
-            temp += 4 + ((temp >> (sizeof(temp) * CHAR_BIT - 1)) * 8);
+      temp += 4 + ((temp >> (sizeof(temp) * CHAR_BIT - 1)) * 8);
 
-            x->block[uoffset].bmi.mv.as_mv.col = (temp / 8) & x->fullpixel_mask;
+      x->block[uoffset].bmi.mv.as_mv.col = (temp / 8) & x->fullpixel_mask;
 
-            if (x->mode_info_context->mbmi.need_to_clamp_mvs)
-                clamp_uvmv_to_umv_border(&x->block[uoffset].bmi.mv.as_mv, x);
+      if (x->mode_info_context->mbmi.need_to_clamp_mvs)
+        clamp_uvmv_to_umv_border(&x->block[uoffset].bmi.mv.as_mv, x);
 
-            x->block[voffset].bmi.mv.as_int = x->block[uoffset].bmi.mv.as_int;
-        }
+      x->block[voffset].bmi.mv.as_int = x->block[uoffset].bmi.mv.as_int;
     }
+  }
 }
 
-void vp8_build_inter_predictors_mb(MACROBLOCKD *xd)
-{
-    if (xd->mode_info_context->mbmi.mode != SPLITMV)
-    {
-        vp8_build_inter16x16_predictors_mb(xd, xd->dst.y_buffer,
-                                           xd->dst.u_buffer, xd->dst.v_buffer,
-                                           xd->dst.y_stride, xd->dst.uv_stride);
-    }
-    else
-    {
-        build_4x4uvmvs(xd);
-        build_inter4x4_predictors_mb(xd);
-    }
+void vp8_build_inter_predictors_mb(MACROBLOCKD *xd) {
+  if (xd->mode_info_context->mbmi.mode != SPLITMV) {
+    vp8_build_inter16x16_predictors_mb(xd, xd->dst.y_buffer, xd->dst.u_buffer,
+                                       xd->dst.v_buffer, xd->dst.y_stride,
+                                       xd->dst.uv_stride);
+  } else {
+    build_4x4uvmvs(xd);
+    build_inter4x4_predictors_mb(xd);
+  }
 }
--- a/vp8/common/reconinter.h
+++ b/vp8/common/reconinter.h
@@ -8,7 +8,6 @@
  *  be found in the AUTHORS file in the root of the source tree.
  */
 
-
 #ifndef VP8_COMMON_RECONINTER_H_
 #define VP8_COMMON_RECONINTER_H_
 
@@ -17,21 +16,16 @@
 #endif
 
 extern void vp8_build_inter_predictors_mb(MACROBLOCKD *x);
-extern void vp8_build_inter16x16_predictors_mb(MACROBLOCKD *x,
-                                               unsigned char *dst_y,
-                                               unsigned char *dst_u,
-                                               unsigned char *dst_v,
-                                               int dst_ystride,
-                                               int dst_uvstride);
+extern void vp8_build_inter16x16_predictors_mb(
+    MACROBLOCKD *x, unsigned char *dst_y, unsigned char *dst_u,
+    unsigned char *dst_v, int dst_ystride, int dst_uvstride);
 
-
 extern void vp8_build_inter16x16_predictors_mby(MACROBLOCKD *x,
                                                 unsigned char *dst_y,
                                                 int dst_ystride);
 extern void vp8_build_inter_predictors_b(BLOCKD *d, int pitch,
                                          unsigned char *base_pre,
-                                         int pre_stride,
-                                         vp8_subpix_fn_t sppf);
+                                         int pre_stride, vp8_subpix_fn_t sppf);
 
 extern void vp8_build_inter16x16_predictors_mbuv(MACROBLOCKD *x);
 extern void vp8_build_inter4x4_predictors_mbuv(MACROBLOCKD *x);
--- a/vp8/common/reconintra.c
+++ b/vp8/common/reconintra.c
@@ -8,7 +8,6 @@
  *  be found in the AUTHORS file in the root of the source tree.
  */
 
-
 #include "./vpx_config.h"
 #include "./vpx_dsp_rtcd.h"
 #include "./vp8_rtcd.h"
@@ -19,9 +18,9 @@
 #include "vp8/common/reconintra4x4.h"
 
 enum {
-    SIZE_16,
-    SIZE_8,
-    NUM_SIZES,
+  SIZE_16,
+  SIZE_8,
+  NUM_SIZES,
 };
 
 typedef void (*intra_pred_fn)(uint8_t *dst, ptrdiff_t stride,
@@ -30,88 +29,68 @@
 static intra_pred_fn pred[4][NUM_SIZES];
 static intra_pred_fn dc_pred[2][2][NUM_SIZES];
 
-static void vp8_init_intra_predictors_internal(void)
-{
-#define INIT_SIZE(sz) \
-    pred[V_PRED][SIZE_##sz] = vpx_v_predictor_##sz##x##sz; \
-    pred[H_PRED][SIZE_##sz] = vpx_h_predictor_##sz##x##sz; \
-    pred[TM_PRED][SIZE_##sz] = vpx_tm_predictor_##sz##x##sz; \
- \
-    dc_pred[0][0][SIZE_##sz] = vpx_dc_128_predictor_##sz##x##sz; \
-    dc_pred[0][1][SIZE_##sz] = vpx_dc_top_predictor_##sz##x##sz; \
-    dc_pred[1][0][SIZE_##sz] = vpx_dc_left_predictor_##sz##x##sz; \
-    dc_pred[1][1][SIZE_##sz] = vpx_dc_predictor_##sz##x##sz
+static void vp8_init_intra_predictors_internal(void) {
+#define INIT_SIZE(sz)                                           \
+  pred[V_PRED][SIZE_##sz] = vpx_v_predictor_##sz##x##sz;        \
+  pred[H_PRED][SIZE_##sz] = vpx_h_predictor_##sz##x##sz;        \
+  pred[TM_PRED][SIZE_##sz] = vpx_tm_predictor_##sz##x##sz;      \
+                                                                \
+  dc_pred[0][0][SIZE_##sz] = vpx_dc_128_predictor_##sz##x##sz;  \
+  dc_pred[0][1][SIZE_##sz] = vpx_dc_top_predictor_##sz##x##sz;  \
+  dc_pred[1][0][SIZE_##sz] = vpx_dc_left_predictor_##sz##x##sz; \
+  dc_pred[1][1][SIZE_##sz] = vpx_dc_predictor_##sz##x##sz
 
-    INIT_SIZE(16);
-    INIT_SIZE(8);
-    vp8_init_intra4x4_predictors_internal();
+  INIT_SIZE(16);
+  INIT_SIZE(8);
+  vp8_init_intra4x4_predictors_internal();
 }
 
-void vp8_build_intra_predictors_mby_s(MACROBLOCKD *x,
-                                      unsigned char * yabove_row,
-                                      unsigned char * yleft,
-                                      int left_stride,
-                                      unsigned char * ypred_ptr,
-                                      int y_stride)
-{
-    MB_PREDICTION_MODE mode = x->mode_info_context->mbmi.mode;
-    DECLARE_ALIGNED(16, uint8_t, yleft_col[16]);
-    int i;
-    intra_pred_fn fn;
+void vp8_build_intra_predictors_mby_s(MACROBLOCKD *x, unsigned char *yabove_row,
+                                      unsigned char *yleft, int left_stride,
+                                      unsigned char *ypred_ptr, int y_stride) {
+  MB_PREDICTION_MODE mode = x->mode_info_context->mbmi.mode;
+  DECLARE_ALIGNED(16, uint8_t, yleft_col[16]);
+  int i;
+  intra_pred_fn fn;
 
-    for (i = 0; i < 16; i++)
-    {
-        yleft_col[i] = yleft[i* left_stride];
-    }
+  for (i = 0; i < 16; i++) {
+    yleft_col[i] = yleft[i * left_stride];
+  }
 
-    if (mode == DC_PRED)
-    {
-        fn = dc_pred[x->left_available][x->up_available][SIZE_16];
-    }
-    else
-    {
-        fn = pred[mode][SIZE_16];
-    }
+  if (mode == DC_PRED) {
+    fn = dc_pred[x->left_available][x->up_available][SIZE_16];
+  } else {
+    fn = pred[mode][SIZE_16];
+  }
 
-    fn(ypred_ptr, y_stride, yabove_row, yleft_col);
+  fn(ypred_ptr, y_stride, yabove_row, yleft_col);
 }
 
-void vp8_build_intra_predictors_mbuv_s(MACROBLOCKD *x,
-                                       unsigned char * uabove_row,
-                                       unsigned char * vabove_row,
-                                       unsigned char * uleft,
-                                       unsigned char * vleft,
-                                       int left_stride,
-                                       unsigned char * upred_ptr,
-                                       unsigned char * vpred_ptr,
-                                       int pred_stride)
-{
-    MB_PREDICTION_MODE uvmode = x->mode_info_context->mbmi.uv_mode;
-    unsigned char uleft_col[8];
-    unsigned char vleft_col[8];
-    int i;
-    intra_pred_fn fn;
+void vp8_build_intra_predictors_mbuv_s(
+    MACROBLOCKD *x, unsigned char *uabove_row, unsigned char *vabove_row,
+    unsigned char *uleft, unsigned char *vleft, int left_stride,
+    unsigned char *upred_ptr, unsigned char *vpred_ptr, int pred_stride) {
+  MB_PREDICTION_MODE uvmode = x->mode_info_context->mbmi.uv_mode;
+  unsigned char uleft_col[8];
+  unsigned char vleft_col[8];
+  int i;
+  intra_pred_fn fn;
 
-    for (i = 0; i < 8; i++)
-    {
-        uleft_col[i] = uleft[i * left_stride];
-        vleft_col[i] = vleft[i * left_stride];
-    }
+  for (i = 0; i < 8; i++) {
+    uleft_col[i] = uleft[i * left_stride];
+    vleft_col[i] = vleft[i * left_stride];
+  }
 
-    if (uvmode == DC_PRED)
-    {
-        fn = dc_pred[x->left_available][x->up_available][SIZE_8];
-    }
-    else
-    {
-        fn = pred[uvmode][SIZE_8];
-    }
+  if (uvmode == DC_PRED) {
+    fn = dc_pred[x->left_available][x->up_available][SIZE_8];
+  } else {
+    fn = pred[uvmode][SIZE_8];
+  }
 
-    fn(upred_ptr, pred_stride, uabove_row, uleft_col);
-    fn(vpred_ptr, pred_stride, vabove_row, vleft_col);
+  fn(upred_ptr, pred_stride, uabove_row, uleft_col);
+  fn(vpred_ptr, pred_stride, vabove_row, vleft_col);
 }
 
-void vp8_init_intra_predictors(void)
-{
-    once(vp8_init_intra_predictors_internal);
+void vp8_init_intra_predictors(void) {
+  once(vp8_init_intra_predictors_internal);
 }
--- a/vp8/common/reconintra.h
+++ b/vp8/common/reconintra.h
@@ -8,7 +8,6 @@
  *  be found in the AUTHORS file in the root of the source tree.
  */
 
-
 #ifndef VP8_COMMON_RECONINTRA_H_
 #define VP8_COMMON_RECONINTRA_H_
 
@@ -18,22 +17,14 @@
 extern "C" {
 #endif
 
-void vp8_build_intra_predictors_mby_s(MACROBLOCKD *x,
-                                      unsigned char *yabove_row,
-                                      unsigned char *yleft,
-                                      int left_stride,
-                                      unsigned char *ypred_ptr,
-                                      int y_stride);
+void vp8_build_intra_predictors_mby_s(MACROBLOCKD *x, unsigned char *yabove_row,
+                                      unsigned char *yleft, int left_stride,
+                                      unsigned char *ypred_ptr, int y_stride);
 
-void vp8_build_intra_predictors_mbuv_s(MACROBLOCKD *x,
-                                       unsigned char * uabove_row,
-                                       unsigned char * vabove_row,
-                                       unsigned char * uleft,
-                                       unsigned char * vleft,
-                                       int left_stride,
-                                       unsigned char * upred_ptr,
-                                       unsigned char * vpred_ptr,
-                                       int pred_stride);
+void vp8_build_intra_predictors_mbuv_s(
+    MACROBLOCKD *x, unsigned char *uabove_row, unsigned char *vabove_row,
+    unsigned char *uleft, unsigned char *vleft, int left_stride,
+    unsigned char *upred_ptr, unsigned char *vpred_ptr, int pred_stride);
 
 void vp8_init_intra_predictors(void);
 
--- a/vp8/common/reconintra4x4.c
+++ b/vp8/common/reconintra4x4.c
@@ -21,35 +21,32 @@
 
 static intra_pred_fn pred[10];
 
-void vp8_init_intra4x4_predictors_internal(void)
-{
-    pred[B_DC_PRED] = vpx_dc_predictor_4x4;
-    pred[B_TM_PRED] = vpx_tm_predictor_4x4;
-    pred[B_VE_PRED] = vpx_ve_predictor_4x4;
-    pred[B_HE_PRED] = vpx_he_predictor_4x4;
-    pred[B_LD_PRED] = vpx_d45e_predictor_4x4;
-    pred[B_RD_PRED] = vpx_d135_predictor_4x4;
-    pred[B_VR_PRED] = vpx_d117_predictor_4x4;
-    pred[B_VL_PRED] = vpx_d63f_predictor_4x4;
-    pred[B_HD_PRED] = vpx_d153_predictor_4x4;
-    pred[B_HU_PRED] = vpx_d207_predictor_4x4;
+void vp8_init_intra4x4_predictors_internal(void) {
+  pred[B_DC_PRED] = vpx_dc_predictor_4x4;
+  pred[B_TM_PRED] = vpx_tm_predictor_4x4;
+  pred[B_VE_PRED] = vpx_ve_predictor_4x4;
+  pred[B_HE_PRED] = vpx_he_predictor_4x4;
+  pred[B_LD_PRED] = vpx_d45e_predictor_4x4;
+  pred[B_RD_PRED] = vpx_d135_predictor_4x4;
+  pred[B_VR_PRED] = vpx_d117_predictor_4x4;
+  pred[B_VL_PRED] = vpx_d63f_predictor_4x4;
+  pred[B_HD_PRED] = vpx_d153_predictor_4x4;
+  pred[B_HU_PRED] = vpx_d207_predictor_4x4;
 }
 
-void vp8_intra4x4_predict(unsigned char *above,
-                          unsigned char *yleft, int left_stride,
-                          B_PREDICTION_MODE b_mode,
+void vp8_intra4x4_predict(unsigned char *above, unsigned char *yleft,
+                          int left_stride, B_PREDICTION_MODE b_mode,
                           unsigned char *dst, int dst_stride,
-                          unsigned char top_left)
-{
-    unsigned char Left[4];
-    unsigned char Aboveb[12], *Above = Aboveb + 4;
+                          unsigned char top_left) {
+  unsigned char Left[4];
+  unsigned char Aboveb[12], *Above = Aboveb + 4;
 
-    Left[0] = yleft[0];
-    Left[1] = yleft[left_stride];
-    Left[2] = yleft[2 * left_stride];
-    Left[3] = yleft[3 * left_stride];
-    memcpy(Above, above, 8);
-    Above[-1] = top_left;
+  Left[0] = yleft[0];
+  Left[1] = yleft[left_stride];
+  Left[2] = yleft[2 * left_stride];
+  Left[3] = yleft[3 * left_stride];
+  memcpy(Above, above, 8);
+  Above[-1] = top_left;
 
-    pred[b_mode](dst, dst_stride, Above, Left);
+  pred[b_mode](dst, dst_stride, Above, Left);
 }
--- a/vp8/common/reconintra4x4.h
+++ b/vp8/common/reconintra4x4.h
@@ -8,7 +8,6 @@
  *  be found in the AUTHORS file in the root of the source tree.
  */
 
-
 #ifndef VP8_COMMON_RECONINTRA4X4_H_
 #define VP8_COMMON_RECONINTRA4X4_H_
 #include "vp8/common/blockd.h"
@@ -18,24 +17,22 @@
 #endif
 
 static INLINE void intra_prediction_down_copy(MACROBLOCKD *xd,
-                                              unsigned char *above_right_src)
-{
-    int dst_stride = xd->dst.y_stride;
-    unsigned char *above_right_dst = xd->dst.y_buffer - dst_stride + 16;
+                                              unsigned char *above_right_src) {
+  int dst_stride = xd->dst.y_stride;
+  unsigned char *above_right_dst = xd->dst.y_buffer - dst_stride + 16;
 
-    unsigned int *src_ptr = (unsigned int *)above_right_src;
-    unsigned int *dst_ptr0 = (unsigned int *)(above_right_dst + 4 * dst_stride);
-    unsigned int *dst_ptr1 = (unsigned int *)(above_right_dst + 8 * dst_stride);
-    unsigned int *dst_ptr2 = (unsigned int *)(above_right_dst + 12 * dst_stride);
+  unsigned int *src_ptr = (unsigned int *)above_right_src;
+  unsigned int *dst_ptr0 = (unsigned int *)(above_right_dst + 4 * dst_stride);
+  unsigned int *dst_ptr1 = (unsigned int *)(above_right_dst + 8 * dst_stride);
+  unsigned int *dst_ptr2 = (unsigned int *)(above_right_dst + 12 * dst_stride);
 
-    *dst_ptr0 = *src_ptr;
-    *dst_ptr1 = *src_ptr;
-    *dst_ptr2 = *src_ptr;
+  *dst_ptr0 = *src_ptr;
+  *dst_ptr1 = *src_ptr;
+  *dst_ptr2 = *src_ptr;
 }
 
-void vp8_intra4x4_predict(unsigned char *Above,
-                          unsigned char *yleft, int left_stride,
-                          B_PREDICTION_MODE b_mode,
+void vp8_intra4x4_predict(unsigned char *Above, unsigned char *yleft,
+                          int left_stride, B_PREDICTION_MODE b_mode,
                           unsigned char *dst, int dst_stride,
                           unsigned char top_left);
 
--- a/vp8/common/rtcd.c
+++ b/vp8/common/rtcd.c
@@ -12,8 +12,4 @@
 #include "./vp8_rtcd.h"
 #include "vpx_ports/vpx_once.h"
 
-
-void vp8_rtcd()
-{
-    once(setup_rtcd_internal);
-}
+void vp8_rtcd() { once(setup_rtcd_internal); }
--- a/vp8/common/setupintrarecon.c
+++ b/vp8/common/setupintrarecon.c
@@ -8,32 +8,28 @@
  *  be found in the AUTHORS file in the root of the source tree.
  */
 
-
 #include "setupintrarecon.h"
 #include "vpx_mem/vpx_mem.h"
 
-void vp8_setup_intra_recon(YV12_BUFFER_CONFIG *ybf)
-{
-    int i;
+void vp8_setup_intra_recon(YV12_BUFFER_CONFIG *ybf) {
+  int i;
 
-    /* set up frame new frame for intra coded blocks */
-    memset(ybf->y_buffer - 1 - ybf->y_stride, 127, ybf->y_width + 5);
-    for (i = 0; i < ybf->y_height; i++)
-        ybf->y_buffer[ybf->y_stride *i - 1] = (unsigned char) 129;
+  /* set up frame new frame for intra coded blocks */
+  memset(ybf->y_buffer - 1 - ybf->y_stride, 127, ybf->y_width + 5);
+  for (i = 0; i < ybf->y_height; i++)
+    ybf->y_buffer[ybf->y_stride * i - 1] = (unsigned char)129;
 
-    memset(ybf->u_buffer - 1 - ybf->uv_stride, 127, ybf->uv_width + 5);
-    for (i = 0; i < ybf->uv_height; i++)
-        ybf->u_buffer[ybf->uv_stride *i - 1] = (unsigned char) 129;
+  memset(ybf->u_buffer - 1 - ybf->uv_stride, 127, ybf->uv_width + 5);
+  for (i = 0; i < ybf->uv_height; i++)
+    ybf->u_buffer[ybf->uv_stride * i - 1] = (unsigned char)129;
 
-    memset(ybf->v_buffer - 1 - ybf->uv_stride, 127, ybf->uv_width + 5);
-    for (i = 0; i < ybf->uv_height; i++)
-        ybf->v_buffer[ybf->uv_stride *i - 1] = (unsigned char) 129;
-
+  memset(ybf->v_buffer - 1 - ybf->uv_stride, 127, ybf->uv_width + 5);
+  for (i = 0; i < ybf->uv_height; i++)
+    ybf->v_buffer[ybf->uv_stride * i - 1] = (unsigned char)129;
 }
 
-void vp8_setup_intra_recon_top_line(YV12_BUFFER_CONFIG *ybf)
-{
-    memset(ybf->y_buffer - 1 - ybf->y_stride, 127, ybf->y_width + 5);
-    memset(ybf->u_buffer - 1 - ybf->uv_stride, 127, ybf->uv_width + 5);
-    memset(ybf->v_buffer - 1 - ybf->uv_stride, 127, ybf->uv_width + 5);
+void vp8_setup_intra_recon_top_line(YV12_BUFFER_CONFIG *ybf) {
+  memset(ybf->y_buffer - 1 - ybf->y_stride, 127, ybf->y_width + 5);
+  memset(ybf->u_buffer - 1 - ybf->uv_stride, 127, ybf->uv_width + 5);
+  memset(ybf->v_buffer - 1 - ybf->uv_stride, 127, ybf->uv_width + 5);
 }
--- a/vp8/common/setupintrarecon.h
+++ b/vp8/common/setupintrarecon.h
@@ -22,20 +22,15 @@
 
 static INLINE void setup_intra_recon_left(unsigned char *y_buffer,
                                           unsigned char *u_buffer,
-                                          unsigned char *v_buffer,
-                                          int y_stride,
-                                          int uv_stride)
-{
-    int i;
+                                          unsigned char *v_buffer, int y_stride,
+                                          int uv_stride) {
+  int i;
 
-    for (i = 0; i < 16; i++)
-        y_buffer[y_stride *i] = (unsigned char) 129;
+  for (i = 0; i < 16; i++) y_buffer[y_stride * i] = (unsigned char)129;
 
-    for (i = 0; i < 8; i++)
-        u_buffer[uv_stride *i] = (unsigned char) 129;
+  for (i = 0; i < 8; i++) u_buffer[uv_stride * i] = (unsigned char)129;
 
-    for (i = 0; i < 8; i++)
-        v_buffer[uv_stride *i] = (unsigned char) 129;
+  for (i = 0; i < 8; i++) v_buffer[uv_stride * i] = (unsigned char)129;
 }
 
 #ifdef __cplusplus
--- a/vp8/common/swapyv12buffer.c
+++ b/vp8/common/swapyv12buffer.c
@@ -8,27 +8,25 @@
  *  be found in the AUTHORS file in the root of the source tree.
  */
 
-
 #include "swapyv12buffer.h"
 
-void vp8_swap_yv12_buffer(YV12_BUFFER_CONFIG *new_frame, YV12_BUFFER_CONFIG *last_frame)
-{
-    unsigned char *temp;
+void vp8_swap_yv12_buffer(YV12_BUFFER_CONFIG *new_frame,
+                          YV12_BUFFER_CONFIG *last_frame) {
+  unsigned char *temp;
 
-    temp = last_frame->buffer_alloc;
-    last_frame->buffer_alloc = new_frame->buffer_alloc;
-    new_frame->buffer_alloc = temp;
+  temp = last_frame->buffer_alloc;
+  last_frame->buffer_alloc = new_frame->buffer_alloc;
+  new_frame->buffer_alloc = temp;
 
-    temp = last_frame->y_buffer;
-    last_frame->y_buffer = new_frame->y_buffer;
-    new_frame->y_buffer = temp;
+  temp = last_frame->y_buffer;
+  last_frame->y_buffer = new_frame->y_buffer;
+  new_frame->y_buffer = temp;
 
-    temp = last_frame->u_buffer;
-    last_frame->u_buffer = new_frame->u_buffer;
-    new_frame->u_buffer = temp;
+  temp = last_frame->u_buffer;
+  last_frame->u_buffer = new_frame->u_buffer;
+  new_frame->u_buffer = temp;
 
-    temp = last_frame->v_buffer;
-    last_frame->v_buffer = new_frame->v_buffer;
-    new_frame->v_buffer = temp;
-
+  temp = last_frame->v_buffer;
+  last_frame->v_buffer = new_frame->v_buffer;
+  new_frame->v_buffer = temp;
 }
--- a/vp8/common/swapyv12buffer.h
+++ b/vp8/common/swapyv12buffer.h
@@ -8,7 +8,6 @@
  *  be found in the AUTHORS file in the root of the source tree.
  */
 
-
 #ifndef VP8_COMMON_SWAPYV12BUFFER_H_
 #define VP8_COMMON_SWAPYV12BUFFER_H_
 
@@ -18,7 +17,8 @@
 extern "C" {
 #endif
 
-void vp8_swap_yv12_buffer(YV12_BUFFER_CONFIG *new_frame, YV12_BUFFER_CONFIG *last_frame);
+void vp8_swap_yv12_buffer(YV12_BUFFER_CONFIG *new_frame,
+                          YV12_BUFFER_CONFIG *last_frame);
 
 #ifdef __cplusplus
 }  // extern "C"
--- a/vp8/common/textblit.c
+++ b/vp8/common/textblit.c
@@ -10,121 +10,109 @@
 
 #include <stdlib.h>
 
+void vp8_blit_text(const char *msg, unsigned char *address, const int pitch) {
+  int letter_bitmap;
+  unsigned char *output_pos = address;
+  int colpos;
+  const int font[] = {
+    0x0,       0x5C00,    0x8020,    0xAFABEA,  0xD7EC0,   0x1111111, 0x1855740,
+    0x18000,   0x45C0,    0x74400,   0x51140,   0x23880,   0xC4000,   0x21080,
+    0x80000,   0x111110,  0xE9D72E,  0x87E40,   0x12AD732, 0xAAD62A,  0x4F94C4,
+    0x4D6B7,   0x456AA,   0x3E8423,  0xAAD6AA,  0xAAD6A2,  0x2800,    0x2A00,
+    0x8A880,   0x52940,   0x22A20,   0x15422,   0x6AD62E,  0x1E4A53E, 0xAAD6BF,
+    0x8C62E,   0xE8C63F,  0x118D6BF, 0x1094BF,  0xCAC62E,  0x1F2109F, 0x118FE31,
+    0xF8C628,  0x8A89F,   0x108421F, 0x1F1105F, 0x1F4105F, 0xE8C62E,  0x2294BF,
+    0x164C62E, 0x12694BF, 0x8AD6A2,  0x10FC21,  0x1F8421F, 0x744107,  0xF8220F,
+    0x1151151, 0x117041,  0x119D731, 0x47E0,    0x1041041, 0xFC400,   0x10440,
+    0x1084210, 0x820
+  };
+  colpos = 0;
 
-void vp8_blit_text(const char *msg, unsigned char *address, const int pitch)
-{
-    int letter_bitmap;
-    unsigned char *output_pos = address;
-    int colpos;
-    const int font[] =
-    {
-        0x0, 0x5C00, 0x8020, 0xAFABEA, 0xD7EC0, 0x1111111, 0x1855740, 0x18000,
-        0x45C0, 0x74400, 0x51140, 0x23880, 0xC4000, 0x21080, 0x80000, 0x111110,
-        0xE9D72E, 0x87E40, 0x12AD732, 0xAAD62A, 0x4F94C4, 0x4D6B7, 0x456AA,
-        0x3E8423, 0xAAD6AA, 0xAAD6A2, 0x2800, 0x2A00, 0x8A880, 0x52940, 0x22A20,
-        0x15422, 0x6AD62E, 0x1E4A53E, 0xAAD6BF, 0x8C62E, 0xE8C63F, 0x118D6BF,
-        0x1094BF, 0xCAC62E, 0x1F2109F, 0x118FE31, 0xF8C628, 0x8A89F, 0x108421F,
-        0x1F1105F, 0x1F4105F, 0xE8C62E, 0x2294BF, 0x164C62E, 0x12694BF, 0x8AD6A2,
-        0x10FC21, 0x1F8421F, 0x744107, 0xF8220F, 0x1151151, 0x117041, 0x119D731,
-        0x47E0, 0x1041041, 0xFC400, 0x10440, 0x1084210, 0x820
-    };
-    colpos = 0;
+  while (msg[colpos] != 0) {
+    char letter = msg[colpos];
+    int fontcol, fontrow;
 
-    while (msg[colpos] != 0)
-    {
-        char letter = msg[colpos];
-        int fontcol, fontrow;
+    if (letter <= 'Z' && letter >= ' ')
+      letter_bitmap = font[letter - ' '];
+    else if (letter <= 'z' && letter >= 'a')
+      letter_bitmap = font[letter - 'a' + 'A' - ' '];
+    else
+      letter_bitmap = font[0];
 
-        if (letter <= 'Z' && letter >= ' ')
-            letter_bitmap = font[letter-' '];
-        else if (letter <= 'z' && letter >= 'a')
-            letter_bitmap = font[letter-'a'+'A' - ' '];
-        else
-            letter_bitmap = font[0];
+    for (fontcol = 6; fontcol >= 0; fontcol--)
+      for (fontrow = 0; fontrow < 5; fontrow++)
+        output_pos[fontrow * pitch + fontcol] =
+            ((letter_bitmap >> (fontcol * 5)) & (1 << fontrow) ? 255 : 0);
 
-        for (fontcol = 6; fontcol >= 0 ; fontcol--)
-            for (fontrow = 0; fontrow < 5; fontrow++)
-                output_pos[fontrow *pitch + fontcol] =
-                    ((letter_bitmap >> (fontcol * 5)) & (1 << fontrow) ? 255 : 0);
-
-        output_pos += 7;
-        colpos++;
-    }
+    output_pos += 7;
+    colpos++;
+  }
 }
 
-static void plot (const int x, const int y, unsigned char *image, const int pitch)
-{
-    image [x+y*pitch] ^= 255;
+static void plot(const int x, const int y, unsigned char *image,
+                 const int pitch) {
+  image[x + y * pitch] ^= 255;
 }
 
 /* Bresenham line algorithm */
-void vp8_blit_line(int x0, int x1, int y0, int y1, unsigned char *image, const int pitch)
-{
-    int steep = abs(y1 - y0) > abs(x1 - x0);
-    int deltax, deltay;
-    int error, ystep, y, x;
+void vp8_blit_line(int x0, int x1, int y0, int y1, unsigned char *image,
+                   const int pitch) {
+  int steep = abs(y1 - y0) > abs(x1 - x0);
+  int deltax, deltay;
+  int error, ystep, y, x;
 
-    if (steep)
-    {
-        int t;
-        t = x0;
-        x0 = y0;
-        y0 = t;
+  if (steep) {
+    int t;
+    t = x0;
+    x0 = y0;
+    y0 = t;
 
-        t = x1;
-        x1 = y1;
-        y1 = t;
-    }
+    t = x1;
+    x1 = y1;
+    y1 = t;
+  }
 
-    if (x0 > x1)
-    {
-        int t;
-        t = x0;
-        x0 = x1;
-        x1 = t;
+  if (x0 > x1) {
+    int t;
+    t = x0;
+    x0 = x1;
+    x1 = t;
 
-        t = y0;
-        y0 = y1;
-        y1 = t;
-    }
+    t = y0;
+    y0 = y1;
+    y1 = t;
+  }
 
-    deltax = x1 - x0;
-    deltay = abs(y1 - y0);
-    error  = deltax / 2;
+  deltax = x1 - x0;
+  deltay = abs(y1 - y0);
+  error = deltax / 2;
 
-    y = y0;
+  y = y0;
 
-    if (y0 < y1)
-        ystep = 1;
-    else
-        ystep = -1;
+  if (y0 < y1)
+    ystep = 1;
+  else
+    ystep = -1;
 
-    if (steep)
-    {
-        for (x = x0; x <= x1; x++)
-        {
-            plot(y,x, image, pitch);
+  if (steep) {
+    for (x = x0; x <= x1; x++) {
+      plot(y, x, image, pitch);
 
-            error = error - deltay;
-            if (error < 0)
-            {
-                y = y + ystep;
-                error = error + deltax;
-            }
-        }
+      error = error - deltay;
+      if (error < 0) {
+        y = y + ystep;
+        error = error + deltax;
+      }
     }
-    else
-    {
-        for (x = x0; x <= x1; x++)
-        {
-            plot(x,y, image, pitch);
+  } else {
+    for (x = x0; x <= x1; x++) {
+      plot(x, y, image, pitch);
 
-            error = error - deltay;
-            if (error < 0)
-            {
-                y = y + ystep;
-                error = error + deltax;
-            }
-        }
+      error = error - deltay;
+      if (error < 0) {
+        y = y + ystep;
+        error = error + deltax;
+      }
     }
+  }
 }
--- a/vp8/common/threading.h
+++ b/vp8/common/threading.h
@@ -8,7 +8,6 @@
  *  be found in the AUTHORS file in the root of the source tree.
  */
 
-
 #ifndef VP8_COMMON_THREADING_H_
 #define VP8_COMMON_THREADING_H_
 
@@ -30,10 +29,12 @@
 #define THREAD_SPECIFIC_INDEX DWORD
 #define pthread_t HANDLE
 #define pthread_attr_t DWORD
-#define pthread_detach(thread) if(thread!=NULL)CloseHandle(thread)
+#define pthread_detach(thread) \
+  if (thread != NULL) CloseHandle(thread)
 #define thread_sleep(nms) Sleep(nms)
-#define pthread_cancel(thread) terminate_thread(thread,0)
-#define ts_key_create(ts_key, destructor) {ts_key = TlsAlloc();};
+#define pthread_cancel(thread) terminate_thread(thread, 0)
+#define ts_key_create(ts_key, destructor) \
+  { ts_key = TlsAlloc(); };
 #define pthread_getspecific(ts_key) TlsGetValue(ts_key)
 #define pthread_setspecific(ts_key, value) TlsSetValue(ts_key, (void *)value)
 #define pthread_self() GetCurrentThreadId()
@@ -53,9 +54,9 @@
 #define thread_sleep(nms) DosSleep(nms)
 #define pthread_cancel(thread) DosKillThread(thread)
 #define ts_key_create(ts_key, destructor) \
-    DosAllocThreadLocalMemory(1, &(ts_key));
+  DosAllocThreadLocalMemory(1, &(ts_key));
 #define pthread_getspecific(ts_key) ((void *)(*(ts_key)))
-#define pthread_setspecific(ts_key, value) (*(ts_key)=(ULONG)(value))
+#define pthread_setspecific(ts_key, value) (*(ts_key) = (ULONG)(value))
 #define pthread_self() _gettid()
 #else
 #ifdef __APPLE__
@@ -75,7 +76,8 @@
 #define THREAD_FUNCTION void *
 #define THREAD_FUNCTION_RETURN void *
 #define THREAD_SPECIFIC_INDEX pthread_key_t
-#define ts_key_create(ts_key, destructor) pthread_key_create (&(ts_key), destructor);
+#define ts_key_create(ts_key, destructor) \
+  pthread_key_create(&(ts_key), destructor);
 #endif
 
 /* Synchronization macros: Win32 and Pthreads */
@@ -82,78 +84,74 @@
 #if defined(_WIN32) && !HAVE_PTHREAD_H
 #define sem_t HANDLE
 #define pause(voidpara) __asm PAUSE
-#define sem_init(sem, sem_attr1, sem_init_value) (int)((*sem = CreateSemaphore(NULL,0,32768,NULL))==NULL)
-#define sem_wait(sem) (int)(WAIT_OBJECT_0 != WaitForSingleObject(*sem,INFINITE))
-#define sem_post(sem) ReleaseSemaphore(*sem,1,NULL)
-#define sem_destroy(sem) if(*sem)((int)(CloseHandle(*sem))==TRUE)
+#define sem_init(sem, sem_attr1, sem_init_value) \
+  (int)((*sem = CreateSemaphore(NULL, 0, 32768, NULL)) == NULL)
+#define sem_wait(sem) \
+  (int)(WAIT_OBJECT_0 != WaitForSingleObject(*sem, INFINITE))
+#define sem_post(sem) ReleaseSemaphore(*sem, 1, NULL)
+#define sem_destroy(sem) \
+  if (*sem) ((int)(CloseHandle(*sem)) == TRUE)
 #define thread_sleep(nms) Sleep(nms)
 
 #elif defined(__OS2__)
-typedef struct
-{
-    HEV  event;
-    HMTX wait_mutex;
-    HMTX count_mutex;
-    int  count;
+typedef struct {
+  HEV event;
+  HMTX wait_mutex;
+  HMTX count_mutex;
+  int count;
 } sem_t;
 
-static inline int sem_init(sem_t *sem, int pshared, unsigned int value)
-{
-    DosCreateEventSem(NULL, &sem->event, pshared ? DC_SEM_SHARED : 0,
-                      value > 0 ? TRUE : FALSE);
-    DosCreateMutexSem(NULL, &sem->wait_mutex, 0, FALSE);
-    DosCreateMutexSem(NULL, &sem->count_mutex, 0, FALSE);
+static inline int sem_init(sem_t *sem, int pshared, unsigned int value) {
+  DosCreateEventSem(NULL, &sem->event, pshared ? DC_SEM_SHARED : 0,
+                    value > 0 ? TRUE : FALSE);
+  DosCreateMutexSem(NULL, &sem->wait_mutex, 0, FALSE);
+  DosCreateMutexSem(NULL, &sem->count_mutex, 0, FALSE);
 
-    sem->count = value;
+  sem->count = value;
 
-    return 0;
+  return 0;
 }
 
-static inline int sem_wait(sem_t * sem)
-{
-    DosRequestMutexSem(sem->wait_mutex, -1);
+static inline int sem_wait(sem_t *sem) {
+  DosRequestMutexSem(sem->wait_mutex, -1);
 
-    DosWaitEventSem(sem->event, -1);
+  DosWaitEventSem(sem->event, -1);
 
-    DosRequestMutexSem(sem->count_mutex, -1);
+  DosRequestMutexSem(sem->count_mutex, -1);
 
-    sem->count--;
-    if (sem->count == 0)
-    {
-        ULONG post_count;
+  sem->count--;
+  if (sem->count == 0) {
+    ULONG post_count;
 
-        DosResetEventSem(sem->event, &post_count);
-    }
+    DosResetEventSem(sem->event, &post_count);
+  }
 
-    DosReleaseMutexSem(sem->count_mutex);
+  DosReleaseMutexSem(sem->count_mutex);
 
-    DosReleaseMutexSem(sem->wait_mutex);
+  DosReleaseMutexSem(sem->wait_mutex);
 
-    return 0;
+  return 0;
 }
 
-static inline int sem_post(sem_t * sem)
-{
-    DosRequestMutexSem(sem->count_mutex, -1);
+static inline int sem_post(sem_t *sem) {
+  DosRequestMutexSem(sem->count_mutex, -1);
 
-    if (sem->count < 32768)
-    {
-        sem->count++;
-        DosPostEventSem(sem->event);
-    }
+  if (sem->count < 32768) {
+    sem->count++;
+    DosPostEventSem(sem->event);
+  }
 
-    DosReleaseMutexSem(sem->count_mutex);
+  DosReleaseMutexSem(sem->count_mutex);
 
-    return 0;
+  return 0;
 }
 
-static inline int sem_destroy(sem_t * sem)
-{
-    DosCloseEventSem(sem->event);
-    DosCloseMutexSem(sem->wait_mutex);
-    DosCloseMutexSem(sem->count_mutex);
+static inline int sem_destroy(sem_t *sem) {
+  DosCloseEventSem(sem->event);
+  DosCloseMutexSem(sem->wait_mutex);
+  DosCloseMutexSem(sem->count_mutex);
 
-    return 0;
+  return 0;
 }
 
 #define thread_sleep(nms) DosSleep(nms)
@@ -162,15 +160,20 @@
 
 #ifdef __APPLE__
 #define sem_t semaphore_t
-#define sem_init(X,Y,Z) semaphore_create(mach_task_self(), X, SYNC_POLICY_FIFO, Z)
-#define sem_wait(sem) (semaphore_wait(*sem) )
+#define sem_init(X, Y, Z) \
+  semaphore_create(mach_task_self(), X, SYNC_POLICY_FIFO, Z)
+#define sem_wait(sem) (semaphore_wait(*sem))
 #define sem_post(sem) semaphore_signal(*sem)
-#define sem_destroy(sem) semaphore_destroy(mach_task_self(),*sem)
-#define thread_sleep(nms) /* { struct timespec ts;ts.tv_sec=0; ts.tv_nsec = 1000*nms;nanosleep(&ts, NULL);} */
+#define sem_destroy(sem) semaphore_destroy(mach_task_self(), *sem)
+#define thread_sleep(nms)
+/* { struct timespec ts;ts.tv_sec=0; ts.tv_nsec =
+   1000*nms;nanosleep(&ts, NULL);} */
 #else
 #include <unistd.h>
 #include <sched.h>
-#define thread_sleep(nms) sched_yield();/* {struct timespec ts;ts.tv_sec=0; ts.tv_nsec = 1000*nms;nanosleep(&ts, NULL);} */
+#define thread_sleep(nms) sched_yield();
+/* {struct timespec ts;ts.tv_sec=0;
+    ts.tv_nsec = 1000*nms;nanosleep(&ts, NULL);} */
 #endif
 /* Not Windows. Assume pthreads */
 
@@ -185,42 +188,41 @@
 #include "vpx_util/vpx_thread.h"
 
 static INLINE void mutex_lock(pthread_mutex_t *const mutex) {
-    const int kMaxTryLocks = 4000;
-    int locked = 0;
-    int i;
+  const int kMaxTryLocks = 4000;
+  int locked = 0;
+  int i;
 
-    for (i = 0; i < kMaxTryLocks; ++i) {
-        if (!pthread_mutex_trylock(mutex)) {
-            locked = 1;
-            break;
-        }
+  for (i = 0; i < kMaxTryLocks; ++i) {
+    if (!pthread_mutex_trylock(mutex)) {
+      locked = 1;
+      break;
     }
+  }
 
-    if (!locked)
-        pthread_mutex_lock(mutex);
+  if (!locked) pthread_mutex_lock(mutex);
 }
 
 static INLINE int protected_read(pthread_mutex_t *const mutex, const int *p) {
-    int ret;
-    mutex_lock(mutex);
-    ret = *p;
-    pthread_mutex_unlock(mutex);
-    return ret;
+  int ret;
+  mutex_lock(mutex);
+  ret = *p;
+  pthread_mutex_unlock(mutex);
+  return ret;
 }
 
 static INLINE void sync_read(pthread_mutex_t *const mutex, int mb_col,
                              const int *last_row_current_mb_col,
                              const int nsync) {
-    while (mb_col > (protected_read(mutex, last_row_current_mb_col) - nsync)) {
-        x86_pause_hint();
-        thread_sleep(0);
-    }
+  while (mb_col > (protected_read(mutex, last_row_current_mb_col) - nsync)) {
+    x86_pause_hint();
+    thread_sleep(0);
+  }
 }
 
 static INLINE void protected_write(pthread_mutex_t *mutex, int *p, int v) {
-    mutex_lock(mutex);
-    *p = v;
-    pthread_mutex_unlock(mutex);
+  mutex_lock(mutex);
+  *p = v;
+  pthread_mutex_unlock(mutex);
 }
 
 #endif /* CONFIG_OS_SUPPORT && CONFIG_MULTITHREAD */
--- a/vp8/common/treecoder.c
+++ b/vp8/common/treecoder.c
@@ -8,7 +8,6 @@
  *  be found in the AUTHORS file in the root of the source tree.
  */
 
-
 #if CONFIG_DEBUG
 #include <assert.h>
 #endif
@@ -16,128 +15,95 @@
 
 #include "treecoder.h"
 
-static void tree2tok(
-    struct vp8_token_struct *const p,
-    vp8_tree t,
-    int i,
-    int v,
-    int L
-)
-{
-    v += v;
-    ++L;
+static void tree2tok(struct vp8_token_struct *const p, vp8_tree t, int i, int v,
+                     int L) {
+  v += v;
+  ++L;
 
-    do
-    {
-        const vp8_tree_index j = t[i++];
+  do {
+    const vp8_tree_index j = t[i++];
 
-        if (j <= 0)
-        {
-            p[-j].value = v;
-            p[-j].Len = L;
-        }
-        else
-            tree2tok(p, t, j, v, L);
-    }
-    while (++v & 1);
+    if (j <= 0) {
+      p[-j].value = v;
+      p[-j].Len = L;
+    } else
+      tree2tok(p, t, j, v, L);
+  } while (++v & 1);
 }
 
-void vp8_tokens_from_tree(struct vp8_token_struct *p, vp8_tree t)
-{
-    tree2tok(p, t, 0, 0, 0);
+void vp8_tokens_from_tree(struct vp8_token_struct *p, vp8_tree t) {
+  tree2tok(p, t, 0, 0, 0);
 }
 
 void vp8_tokens_from_tree_offset(struct vp8_token_struct *p, vp8_tree t,
-                                 int offset)
-{
-    tree2tok(p - offset, t, 0, 0, 0);
+                                 int offset) {
+  tree2tok(p - offset, t, 0, 0, 0);
 }
 
-static void branch_counts(
-    int n,                      /* n = size of alphabet */
-    vp8_token tok               [ /* n */ ],
-    vp8_tree tree,
-    unsigned int branch_ct       [ /* n-1 */ ] [2],
-    const unsigned int num_events[ /* n */ ]
-)
-{
-    const int tree_len = n - 1;
-    int t = 0;
+static void branch_counts(int n, /* n = size of alphabet */
+                          vp8_token tok[/* n */], vp8_tree tree,
+                          unsigned int branch_ct[/* n-1 */][2],
+                          const unsigned int num_events[/* n */]) {
+  const int tree_len = n - 1;
+  int t = 0;
 
 #if CONFIG_DEBUG
-    assert(tree_len);
+  assert(tree_len);
 #endif
 
-    do
-    {
-        branch_ct[t][0] = branch_ct[t][1] = 0;
-    }
-    while (++t < tree_len);
+  do {
+    branch_ct[t][0] = branch_ct[t][1] = 0;
+  } while (++t < tree_len);
 
-    t = 0;
+  t = 0;
 
-    do
-    {
-        int L = tok[t].Len;
-        const int enc = tok[t].value;
-        const unsigned int ct = num_events[t];
+  do {
+    int L = tok[t].Len;
+    const int enc = tok[t].value;
+    const unsigned int ct = num_events[t];
 
-        vp8_tree_index i = 0;
+    vp8_tree_index i = 0;
 
-        do
-        {
-            const int b = (enc >> --L) & 1;
-            const int j = i >> 1;
+    do {
+      const int b = (enc >> --L) & 1;
+      const int j = i >> 1;
 #if CONFIG_DEBUG
-            assert(j < tree_len  &&  0 <= L);
+      assert(j < tree_len && 0 <= L);
 #endif
 
-            branch_ct [j] [b] += ct;
-            i = tree[ i + b];
-        }
-        while (i > 0);
+      branch_ct[j][b] += ct;
+      i = tree[i + b];
+    } while (i > 0);
 
 #if CONFIG_DEBUG
-        assert(!L);
+    assert(!L);
 #endif
-    }
-    while (++t < n);
-
+  } while (++t < n);
 }
 
+void vp8_tree_probs_from_distribution(int n, /* n = size of alphabet */
+                                      vp8_token tok[/* n */], vp8_tree tree,
+                                      vp8_prob probs[/* n-1 */],
+                                      unsigned int branch_ct[/* n-1 */][2],
+                                      const unsigned int num_events[/* n */],
+                                      unsigned int Pfac, int rd) {
+  const int tree_len = n - 1;
+  int t = 0;
 
-void vp8_tree_probs_from_distribution(
-    int n,                      /* n = size of alphabet */
-    vp8_token tok               [ /* n */ ],
-    vp8_tree tree,
-    vp8_prob probs          [ /* n-1 */ ],
-    unsigned int branch_ct       [ /* n-1 */ ] [2],
-    const unsigned int num_events[ /* n */ ],
-    unsigned int Pfac,
-    int rd
-)
-{
-    const int tree_len = n - 1;
-    int t = 0;
+  branch_counts(n, tok, tree, branch_ct, num_events);
 
-    branch_counts(n, tok, tree, branch_ct, num_events);
+  do {
+    const unsigned int *const c = branch_ct[t];
+    const unsigned int tot = c[0] + c[1];
 
-    do
-    {
-        const unsigned int *const c = branch_ct[t];
-        const unsigned int tot = c[0] + c[1];
-
 #if CONFIG_DEBUG
-        assert(tot < (1 << 24));        /* no overflow below */
+    assert(tot < (1 << 24)); /* no overflow below */
 #endif
 
-        if (tot)
-        {
-            const unsigned int p = ((c[0] * Pfac) + (rd ? tot >> 1 : 0)) / tot;
-            probs[t] = p < 256 ? (p ? p : 1) : 255; /* agree w/old version for now */
-        }
-        else
-            probs[t] = vp8_prob_half;
-    }
-    while (++t < tree_len);
+    if (tot) {
+      const unsigned int p = ((c[0] * Pfac) + (rd ? tot >> 1 : 0)) / tot;
+      probs[t] = p < 256 ? (p ? p : 1) : 255; /* agree w/old version for now */
+    } else
+      probs[t] = vp8_prob_half;
+  } while (++t < tree_len);
 }
--- a/vp8/common/treecoder.h
+++ b/vp8/common/treecoder.h
@@ -8,7 +8,6 @@
  *  be found in the AUTHORS file in the root of the source tree.
  */
 
-
 #ifndef VP8_COMMON_TREECODER_H_
 #define VP8_COMMON_TREECODER_H_
 
@@ -18,10 +17,9 @@
 
 typedef unsigned char vp8bc_index_t; /* probability index */
 
-
 typedef unsigned char vp8_prob;
 
-#define vp8_prob_half ( (vp8_prob) 128)
+#define vp8_prob_half ((vp8_prob)128)
 
 typedef signed char vp8_tree_index;
 struct bool_coder_spec;
@@ -34,11 +32,8 @@
 typedef const bool_writer c_bool_writer;
 typedef const bool_reader c_bool_reader;
 
+#define vp8_complement(x) (255 - x)
 
-
-# define vp8_complement( x) (255 - x)
-
-
 /* We build coding trees compactly in arrays.
    Each node of the tree is a pair of vp8_tree_indices.
    Array index often references a corresponding probability table.
@@ -48,11 +43,9 @@
 
 typedef const vp8_tree_index vp8_tree[], *vp8_tree_p;
 
-
-typedef const struct vp8_token_struct
-{
-    int value;
-    int Len;
+typedef const struct vp8_token_struct {
+  int value;
+  int Len;
 } vp8_token;
 
 /* Construct encoding array from tree. */
@@ -61,35 +54,26 @@
 void vp8_tokens_from_tree_offset(struct vp8_token_struct *, vp8_tree,
                                  int offset);
 
-
 /* Convert array of token occurrence counts into a table of probabilities
    for the associated binary encoding tree.  Also writes count of branches
    taken for each node on the tree; this facilitiates decisions as to
    probability updates. */
 
-void vp8_tree_probs_from_distribution(
-    int n,                      /* n = size of alphabet */
-    vp8_token tok               [ /* n */ ],
-    vp8_tree tree,
-    vp8_prob probs          [ /* n-1 */ ],
-    unsigned int branch_ct       [ /* n-1 */ ] [2],
-    const unsigned int num_events[ /* n */ ],
-    unsigned int Pfactor,
-    int Round
-);
+void vp8_tree_probs_from_distribution(int n, /* n = size of alphabet */
+                                      vp8_token tok[/* n */], vp8_tree tree,
+                                      vp8_prob probs[/* n-1 */],
+                                      unsigned int branch_ct[/* n-1 */][2],
+                                      const unsigned int num_events[/* n */],
+                                      unsigned int Pfactor, int Round);
 
 /* Variant of above using coder spec rather than hardwired 8-bit probs. */
 
-void vp8bc_tree_probs_from_distribution(
-    int n,                      /* n = size of alphabet */
-    vp8_token tok               [ /* n */ ],
-    vp8_tree tree,
-    vp8_prob probs          [ /* n-1 */ ],
-    unsigned int branch_ct       [ /* n-1 */ ] [2],
-    const unsigned int num_events[ /* n */ ],
-    c_bool_coder_spec *s
-);
-
+void vp8bc_tree_probs_from_distribution(int n, /* n = size of alphabet */
+                                        vp8_token tok[/* n */], vp8_tree tree,
+                                        vp8_prob probs[/* n-1 */],
+                                        unsigned int branch_ct[/* n-1 */][2],
+                                        const unsigned int num_events[/* n */],
+                                        c_bool_coder_spec *s);
 
 #ifdef __cplusplus
 }  // extern "C"
--- a/vp8/common/vp8_entropymodedata.h
+++ b/vp8/common/vp8_entropymodedata.h
@@ -17,235 +17,153 @@
 
 /*Generated file, included by entropymode.c*/
 
-
-const struct vp8_token_struct vp8_bmode_encodings[VP8_BINTRAMODES] =
-{
-    { 0, 1 },
-    { 2, 2 },
-    { 6, 3 },
-    { 28, 5 },
-    { 30, 5 },
-    { 58, 6 },
-    { 59, 6 },
-    { 62, 6 },
-    { 126, 7 },
-    { 127, 7 }
+const struct vp8_token_struct vp8_bmode_encodings[VP8_BINTRAMODES] = {
+  { 0, 1 },  { 2, 2 },  { 6, 3 },  { 28, 5 },  { 30, 5 },
+  { 58, 6 }, { 59, 6 }, { 62, 6 }, { 126, 7 }, { 127, 7 }
 };
 
-const struct vp8_token_struct vp8_ymode_encodings[VP8_YMODES] =
-{
-    { 0, 1 },
-    { 4, 3 },
-    { 5, 3 },
-    { 6, 3 },
-    { 7, 3 }
+const struct vp8_token_struct vp8_ymode_encodings[VP8_YMODES] = {
+  { 0, 1 }, { 4, 3 }, { 5, 3 }, { 6, 3 }, { 7, 3 }
 };
 
-const struct vp8_token_struct vp8_kf_ymode_encodings[VP8_YMODES] =
-{
-    { 4, 3 },
-    { 5, 3 },
-    { 6, 3 },
-    { 7, 3 },
-    { 0, 1 }
+const struct vp8_token_struct vp8_kf_ymode_encodings[VP8_YMODES] = {
+  { 4, 3 }, { 5, 3 }, { 6, 3 }, { 7, 3 }, { 0, 1 }
 };
 
-const struct vp8_token_struct vp8_uv_mode_encodings[VP8_UV_MODES] =
-{
-    { 0, 1 },
-    { 2, 2 },
-    { 6, 3 },
-    { 7, 3 }
+const struct vp8_token_struct vp8_uv_mode_encodings[VP8_UV_MODES] = {
+  { 0, 1 }, { 2, 2 }, { 6, 3 }, { 7, 3 }
 };
 
-const struct vp8_token_struct vp8_mbsplit_encodings[VP8_NUMMBSPLITS] =
-{
-    { 6, 3 },
-    { 7, 3 },
-    { 2, 2 },
-    { 0, 1 }
+const struct vp8_token_struct vp8_mbsplit_encodings[VP8_NUMMBSPLITS] = {
+  { 6, 3 }, { 7, 3 }, { 2, 2 }, { 0, 1 }
 };
 
-const struct vp8_token_struct vp8_mv_ref_encoding_array[VP8_MVREFS] =
-{
-    { 2, 2 },
-    { 6, 3 },
-    { 0, 1 },
-    { 14, 4 },
-    { 15, 4 }
+const struct vp8_token_struct vp8_mv_ref_encoding_array[VP8_MVREFS] = {
+  { 2, 2 }, { 6, 3 }, { 0, 1 }, { 14, 4 }, { 15, 4 }
 };
 
-const struct vp8_token_struct vp8_sub_mv_ref_encoding_array[VP8_SUBMVREFS] =
-{
-    { 0, 1 },
-    { 2, 2 },
-    { 6, 3 },
-    { 7, 3 }
+const struct vp8_token_struct vp8_sub_mv_ref_encoding_array[VP8_SUBMVREFS] = {
+  { 0, 1 }, { 2, 2 }, { 6, 3 }, { 7, 3 }
 };
 
-const struct vp8_token_struct vp8_small_mvencodings[8] =
-{
-    { 0, 3 },
-    { 1, 3 },
-    { 2, 3 },
-    { 3, 3 },
-    { 4, 3 },
-    { 5, 3 },
-    { 6, 3 },
-    { 7, 3 }
+const struct vp8_token_struct vp8_small_mvencodings[8] = {
+  { 0, 3 }, { 1, 3 }, { 2, 3 }, { 3, 3 }, { 4, 3 }, { 5, 3 }, { 6, 3 }, { 7, 3 }
 };
 
-const vp8_prob vp8_ymode_prob[VP8_YMODES-1] =
-{
-    112, 86, 140, 37
-};
+const vp8_prob vp8_ymode_prob[VP8_YMODES - 1] = { 112, 86, 140, 37 };
 
-const vp8_prob vp8_kf_ymode_prob[VP8_YMODES-1] =
-{
-    145, 156, 163, 128
-};
+const vp8_prob vp8_kf_ymode_prob[VP8_YMODES - 1] = { 145, 156, 163, 128 };
 
-const vp8_prob vp8_uv_mode_prob[VP8_UV_MODES-1] =
-{
-    162, 101, 204
-};
+const vp8_prob vp8_uv_mode_prob[VP8_UV_MODES - 1] = { 162, 101, 204 };
 
-const vp8_prob vp8_kf_uv_mode_prob[VP8_UV_MODES-1] =
-{
-    142, 114, 183
-};
+const vp8_prob vp8_kf_uv_mode_prob[VP8_UV_MODES - 1] = { 142, 114, 183 };
 
-const vp8_prob vp8_bmode_prob[VP8_BINTRAMODES-1] =
-{
-    120, 90, 79, 133, 87, 85, 80, 111, 151
-};
+const vp8_prob vp8_bmode_prob[VP8_BINTRAMODES - 1] = { 120, 90, 79,  133, 87,
+                                                       85,  80, 111, 151 };
 
-
-
-const vp8_prob vp8_kf_bmode_prob
-[VP8_BINTRAMODES] [VP8_BINTRAMODES] [VP8_BINTRAMODES-1] =
-{
-    {
-        { 231, 120,  48,  89, 115, 113, 120, 152, 112 },
-        { 152, 179,  64, 126, 170, 118,  46,  70,  95 },
-        { 175,  69, 143,  80,  85,  82,  72, 155, 103 },
-        {  56,  58,  10, 171, 218, 189,  17,  13, 152 },
-        { 144,  71,  10,  38, 171, 213, 144,  34,  26 },
-        { 114,  26,  17, 163,  44, 195,  21,  10, 173 },
-        { 121,  24,  80, 195,  26,  62,  44,  64,  85 },
-        { 170,  46,  55,  19, 136, 160,  33, 206,  71 },
-        {  63,  20,   8, 114, 114, 208,  12,   9, 226 },
-        {  81,  40,  11,  96, 182,  84,  29,  16,  36 }
-    },
-    {
-        { 134, 183,  89, 137,  98, 101, 106, 165, 148 },
-        {  72, 187, 100, 130, 157, 111,  32,  75,  80 },
-        {  66, 102, 167,  99,  74,  62,  40, 234, 128 },
-        {  41,  53,   9, 178, 241, 141,  26,   8, 107 },
-        { 104,  79,  12,  27, 217, 255,  87,  17,   7 },
-        {  74,  43,  26, 146,  73, 166,  49,  23, 157 },
-        {  65,  38, 105, 160,  51,  52,  31, 115, 128 },
-        {  87,  68,  71,  44, 114,  51,  15, 186,  23 },
-        {  47,  41,  14, 110, 182, 183,  21,  17, 194 },
-        {  66,  45,  25, 102, 197, 189,  23,  18,  22 }
-    },
-    {
-        {  88,  88, 147, 150,  42,  46,  45, 196, 205 },
-        {  43,  97, 183, 117,  85,  38,  35, 179,  61 },
-        {  39,  53, 200,  87,  26,  21,  43, 232, 171 },
-        {  56,  34,  51, 104, 114, 102,  29,  93,  77 },
-        { 107,  54,  32,  26,  51,   1,  81,  43,  31 },
-        {  39,  28,  85, 171,  58, 165,  90,  98,  64 },
-        {  34,  22, 116, 206,  23,  34,  43, 166,  73 },
-        {  68,  25, 106,  22,  64, 171,  36, 225, 114 },
-        {  34,  19,  21, 102, 132, 188,  16,  76, 124 },
-        {  62,  18,  78,  95,  85,  57,  50,  48,  51 }
-    },
-    {
-        { 193, 101,  35, 159, 215, 111,  89,  46, 111 },
-        {  60, 148,  31, 172, 219, 228,  21,  18, 111 },
-        { 112, 113,  77,  85, 179, 255,  38, 120, 114 },
-        {  40,  42,   1, 196, 245, 209,  10,  25, 109 },
-        { 100,  80,   8,  43, 154,   1,  51,  26,  71 },
-        {  88,  43,  29, 140, 166, 213,  37,  43, 154 },
-        {  61,  63,  30, 155,  67,  45,  68,   1, 209 },
-        { 142,  78,  78,  16, 255, 128,  34, 197, 171 },
-        {  41,  40,   5, 102, 211, 183,   4,   1, 221 },
-        {  51,  50,  17, 168, 209, 192,  23,  25,  82 }
-    },
-    {
-        { 125,  98,  42,  88, 104,  85, 117, 175,  82 },
-        {  95,  84,  53,  89, 128, 100, 113, 101,  45 },
-        {  75,  79, 123,  47,  51, 128,  81, 171,   1 },
-        {  57,  17,   5,  71, 102,  57,  53,  41,  49 },
-        { 115,  21,   2,  10, 102, 255, 166,  23,   6 },
-        {  38,  33,  13, 121,  57,  73,  26,   1,  85 },
-        {  41,  10,  67, 138,  77, 110,  90,  47, 114 },
-        { 101,  29,  16,  10,  85, 128, 101, 196,  26 },
-        {  57,  18,  10, 102, 102, 213,  34,  20,  43 },
-        { 117,  20,  15,  36, 163, 128,  68,   1,  26 }
-    },
-    {
-        { 138,  31,  36, 171,  27, 166,  38,  44, 229 },
-        {  67,  87,  58, 169,  82, 115,  26,  59, 179 },
-        {  63,  59,  90, 180,  59, 166,  93,  73, 154 },
-        {  40,  40,  21, 116, 143, 209,  34,  39, 175 },
-        {  57,  46,  22,  24, 128,   1,  54,  17,  37 },
-        {  47,  15,  16, 183,  34, 223,  49,  45, 183 },
-        {  46,  17,  33, 183,   6,  98,  15,  32, 183 },
-        {  65,  32,  73, 115,  28, 128,  23, 128, 205 },
-        {  40,   3,   9, 115,  51, 192,  18,   6, 223 },
-        {  87,  37,   9, 115,  59,  77,  64,  21,  47 }
-    },
-    {
-        { 104,  55,  44, 218,   9,  54,  53, 130, 226 },
-        {  64,  90,  70, 205,  40,  41,  23,  26,  57 },
-        {  54,  57, 112, 184,   5,  41,  38, 166, 213 },
-        {  30,  34,  26, 133, 152, 116,  10,  32, 134 },
-        {  75,  32,  12,  51, 192, 255, 160,  43,  51 },
-        {  39,  19,  53, 221,  26, 114,  32,  73, 255 },
-        {  31,   9,  65, 234,   2,  15,   1, 118,  73 },
-        {  88,  31,  35,  67, 102,  85,  55, 186,  85 },
-        {  56,  21,  23, 111,  59, 205,  45,  37, 192 },
-        {  55,  38,  70, 124,  73, 102,   1,  34,  98 }
-    },
-    {
-        { 102,  61,  71,  37,  34,  53,  31, 243, 192 },
-        {  69,  60,  71,  38,  73, 119,  28, 222,  37 },
-        {  68,  45, 128,  34,   1,  47,  11, 245, 171 },
-        {  62,  17,  19,  70, 146,  85,  55,  62,  70 },
-        {  75,  15,   9,   9,  64, 255, 184, 119,  16 },
-        {  37,  43,  37, 154, 100, 163,  85, 160,   1 },
-        {  63,   9,  92, 136,  28,  64,  32, 201,  85 },
-        {  86,   6,  28,   5,  64, 255,  25, 248,   1 },
-        {  56,   8,  17, 132, 137, 255,  55, 116, 128 },
-        {  58,  15,  20,  82, 135,  57,  26, 121,  40 }
-    },
-    {
-        { 164,  50,  31, 137, 154, 133,  25,  35, 218 },
-        {  51, 103,  44, 131, 131, 123,  31,   6, 158 },
-        {  86,  40,  64, 135, 148, 224,  45, 183, 128 },
-        {  22,  26,  17, 131, 240, 154,  14,   1, 209 },
-        {  83,  12,  13,  54, 192, 255,  68,  47,  28 },
-        {  45,  16,  21,  91,  64, 222,   7,   1, 197 },
-        {  56,  21,  39, 155,  60, 138,  23, 102, 213 },
-        {  85,  26,  85,  85, 128, 128,  32, 146, 171 },
-        {  18,  11,   7,  63, 144, 171,   4,   4, 246 },
-        {  35,  27,  10, 146, 174, 171,  12,  26, 128 }
-    },
-    {
-        { 190,  80,  35,  99, 180,  80, 126,  54,  45 },
-        {  85, 126,  47,  87, 176,  51,  41,  20,  32 },
-        { 101,  75, 128, 139, 118, 146, 116, 128,  85 },
-        {  56,  41,  15, 176, 236,  85,  37,   9,  62 },
-        { 146,  36,  19,  30, 171, 255,  97,  27,  20 },
-        {  71,  30,  17, 119, 118, 255,  17,  18, 138 },
-        { 101,  38,  60, 138,  55,  70,  43,  26, 142 },
-        { 138,  45,  61,  62, 219,   1,  81, 188,  64 },
-        {  32,  41,  20, 117, 151, 142,  20,  21, 163 },
-        { 112,  19,  12,  61, 195, 128,  48,   4,  24 }
-    }
-};
+const vp8_prob
+    vp8_kf_bmode_prob[VP8_BINTRAMODES][VP8_BINTRAMODES][VP8_BINTRAMODES - 1] = {
+      { { 231, 120, 48, 89, 115, 113, 120, 152, 112 },
+        { 152, 179, 64, 126, 170, 118, 46, 70, 95 },
+        { 175, 69, 143, 80, 85, 82, 72, 155, 103 },
+        { 56, 58, 10, 171, 218, 189, 17, 13, 152 },
+        { 144, 71, 10, 38, 171, 213, 144, 34, 26 },
+        { 114, 26, 17, 163, 44, 195, 21, 10, 173 },
+        { 121, 24, 80, 195, 26, 62, 44, 64, 85 },
+        { 170, 46, 55, 19, 136, 160, 33, 206, 71 },
+        { 63, 20, 8, 114, 114, 208, 12, 9, 226 },
+        { 81, 40, 11, 96, 182, 84, 29, 16, 36 } },
+      { { 134, 183, 89, 137, 98, 101, 106, 165, 148 },
+        { 72, 187, 100, 130, 157, 111, 32, 75, 80 },
+        { 66, 102, 167, 99, 74, 62, 40, 234, 128 },
+        { 41, 53, 9, 178, 241, 141, 26, 8, 107 },
+        { 104, 79, 12, 27, 217, 255, 87, 17, 7 },
+        { 74, 43, 26, 146, 73, 166, 49, 23, 157 },
+        { 65, 38, 105, 160, 51, 52, 31, 115, 128 },
+        { 87, 68, 71, 44, 114, 51, 15, 186, 23 },
+        { 47, 41, 14, 110, 182, 183, 21, 17, 194 },
+        { 66, 45, 25, 102, 197, 189, 23, 18, 22 } },
+      { { 88, 88, 147, 150, 42, 46, 45, 196, 205 },
+        { 43, 97, 183, 117, 85, 38, 35, 179, 61 },
+        { 39, 53, 200, 87, 26, 21, 43, 232, 171 },
+        { 56, 34, 51, 104, 114, 102, 29, 93, 77 },
+        { 107, 54, 32, 26, 51, 1, 81, 43, 31 },
+        { 39, 28, 85, 171, 58, 165, 90, 98, 64 },
+        { 34, 22, 116, 206, 23, 34, 43, 166, 73 },
+        { 68, 25, 106, 22, 64, 171, 36, 225, 114 },
+        { 34, 19, 21, 102, 132, 188, 16, 76, 124 },
+        { 62, 18, 78, 95, 85, 57, 50, 48, 51 } },
+      { { 193, 101, 35, 159, 215, 111, 89, 46, 111 },
+        { 60, 148, 31, 172, 219, 228, 21, 18, 111 },
+        { 112, 113, 77, 85, 179, 255, 38, 120, 114 },
+        { 40, 42, 1, 196, 245, 209, 10, 25, 109 },
+        { 100, 80, 8, 43, 154, 1, 51, 26, 71 },
+        { 88, 43, 29, 140, 166, 213, 37, 43, 154 },
+        { 61, 63, 30, 155, 67, 45, 68, 1, 209 },
+        { 142, 78, 78, 16, 255, 128, 34, 197, 171 },
+        { 41, 40, 5, 102, 211, 183, 4, 1, 221 },
+        { 51, 50, 17, 168, 209, 192, 23, 25, 82 } },
+      { { 125, 98, 42, 88, 104, 85, 117, 175, 82 },
+        { 95, 84, 53, 89, 128, 100, 113, 101, 45 },
+        { 75, 79, 123, 47, 51, 128, 81, 171, 1 },
+        { 57, 17, 5, 71, 102, 57, 53, 41, 49 },
+        { 115, 21, 2, 10, 102, 255, 166, 23, 6 },
+        { 38, 33, 13, 121, 57, 73, 26, 1, 85 },
+        { 41, 10, 67, 138, 77, 110, 90, 47, 114 },
+        { 101, 29, 16, 10, 85, 128, 101, 196, 26 },
+        { 57, 18, 10, 102, 102, 213, 34, 20, 43 },
+        { 117, 20, 15, 36, 163, 128, 68, 1, 26 } },
+      { { 138, 31, 36, 171, 27, 166, 38, 44, 229 },
+        { 67, 87, 58, 169, 82, 115, 26, 59, 179 },
+        { 63, 59, 90, 180, 59, 166, 93, 73, 154 },
+        { 40, 40, 21, 116, 143, 209, 34, 39, 175 },
+        { 57, 46, 22, 24, 128, 1, 54, 17, 37 },
+        { 47, 15, 16, 183, 34, 223, 49, 45, 183 },
+        { 46, 17, 33, 183, 6, 98, 15, 32, 183 },
+        { 65, 32, 73, 115, 28, 128, 23, 128, 205 },
+        { 40, 3, 9, 115, 51, 192, 18, 6, 223 },
+        { 87, 37, 9, 115, 59, 77, 64, 21, 47 } },
+      { { 104, 55, 44, 218, 9, 54, 53, 130, 226 },
+        { 64, 90, 70, 205, 40, 41, 23, 26, 57 },
+        { 54, 57, 112, 184, 5, 41, 38, 166, 213 },
+        { 30, 34, 26, 133, 152, 116, 10, 32, 134 },
+        { 75, 32, 12, 51, 192, 255, 160, 43, 51 },
+        { 39, 19, 53, 221, 26, 114, 32, 73, 255 },
+        { 31, 9, 65, 234, 2, 15, 1, 118, 73 },
+        { 88, 31, 35, 67, 102, 85, 55, 186, 85 },
+        { 56, 21, 23, 111, 59, 205, 45, 37, 192 },
+        { 55, 38, 70, 124, 73, 102, 1, 34, 98 } },
+      { { 102, 61, 71, 37, 34, 53, 31, 243, 192 },
+        { 69, 60, 71, 38, 73, 119, 28, 222, 37 },
+        { 68, 45, 128, 34, 1, 47, 11, 245, 171 },
+        { 62, 17, 19, 70, 146, 85, 55, 62, 70 },
+        { 75, 15, 9, 9, 64, 255, 184, 119, 16 },
+        { 37, 43, 37, 154, 100, 163, 85, 160, 1 },
+        { 63, 9, 92, 136, 28, 64, 32, 201, 85 },
+        { 86, 6, 28, 5, 64, 255, 25, 248, 1 },
+        { 56, 8, 17, 132, 137, 255, 55, 116, 128 },
+        { 58, 15, 20, 82, 135, 57, 26, 121, 40 } },
+      { { 164, 50, 31, 137, 154, 133, 25, 35, 218 },
+        { 51, 103, 44, 131, 131, 123, 31, 6, 158 },
+        { 86, 40, 64, 135, 148, 224, 45, 183, 128 },
+        { 22, 26, 17, 131, 240, 154, 14, 1, 209 },
+        { 83, 12, 13, 54, 192, 255, 68, 47, 28 },
+        { 45, 16, 21, 91, 64, 222, 7, 1, 197 },
+        { 56, 21, 39, 155, 60, 138, 23, 102, 213 },
+        { 85, 26, 85, 85, 128, 128, 32, 146, 171 },
+        { 18, 11, 7, 63, 144, 171, 4, 4, 246 },
+        { 35, 27, 10, 146, 174, 171, 12, 26, 128 } },
+      { { 190, 80, 35, 99, 180, 80, 126, 54, 45 },
+        { 85, 126, 47, 87, 176, 51, 41, 20, 32 },
+        { 101, 75, 128, 139, 118, 146, 116, 128, 85 },
+        { 56, 41, 15, 176, 236, 85, 37, 9, 62 },
+        { 146, 36, 19, 30, 171, 255, 97, 27, 20 },
+        { 71, 30, 17, 119, 118, 255, 17, 18, 138 },
+        { 101, 38, 60, 138, 55, 70, 43, 26, 142 },
+        { 138, 45, 61, 62, 219, 1, 81, 188, 64 },
+        { 32, 41, 20, 117, 151, 142, 20, 21, 163 },
+        { 112, 19, 12, 61, 195, 128, 48, 4, 24 } }
+    };
 
 #ifdef __cplusplus
 }  // extern "C"
--- a/vp8/common/vp8_loopfilter.c
+++ b/vp8/common/vp8_loopfilter.c
@@ -8,7 +8,6 @@
  *  be found in the AUTHORS file in the root of the source tree.
  */
 
-
 #include "vpx_config.h"
 #include "vp8_rtcd.h"
 #include "loopfilter.h"
@@ -15,455 +14,395 @@
 #include "onyxc_int.h"
 #include "vpx_mem/vpx_mem.h"
 
+static void lf_init_lut(loop_filter_info_n *lfi) {
+  int filt_lvl;
 
-static void lf_init_lut(loop_filter_info_n *lfi)
-{
-    int filt_lvl;
-
-    for (filt_lvl = 0; filt_lvl <= MAX_LOOP_FILTER; filt_lvl++)
-    {
-        if (filt_lvl >= 40)
-        {
-            lfi->hev_thr_lut[KEY_FRAME][filt_lvl] = 2;
-            lfi->hev_thr_lut[INTER_FRAME][filt_lvl] = 3;
-        }
-        else if (filt_lvl >= 20)
-        {
-            lfi->hev_thr_lut[KEY_FRAME][filt_lvl] = 1;
-            lfi->hev_thr_lut[INTER_FRAME][filt_lvl] = 2;
-        }
-        else if (filt_lvl >= 15)
-        {
-            lfi->hev_thr_lut[KEY_FRAME][filt_lvl] = 1;
-            lfi->hev_thr_lut[INTER_FRAME][filt_lvl] = 1;
-        }
-        else
-        {
-            lfi->hev_thr_lut[KEY_FRAME][filt_lvl] = 0;
-            lfi->hev_thr_lut[INTER_FRAME][filt_lvl] = 0;
-        }
+  for (filt_lvl = 0; filt_lvl <= MAX_LOOP_FILTER; filt_lvl++) {
+    if (filt_lvl >= 40) {
+      lfi->hev_thr_lut[KEY_FRAME][filt_lvl] = 2;
+      lfi->hev_thr_lut[INTER_FRAME][filt_lvl] = 3;
+    } else if (filt_lvl >= 20) {
+      lfi->hev_thr_lut[KEY_FRAME][filt_lvl] = 1;
+      lfi->hev_thr_lut[INTER_FRAME][filt_lvl] = 2;
+    } else if (filt_lvl >= 15) {
+      lfi->hev_thr_lut[KEY_FRAME][filt_lvl] = 1;
+      lfi->hev_thr_lut[INTER_FRAME][filt_lvl] = 1;
+    } else {
+      lfi->hev_thr_lut[KEY_FRAME][filt_lvl] = 0;
+      lfi->hev_thr_lut[INTER_FRAME][filt_lvl] = 0;
     }
+  }
 
-    lfi->mode_lf_lut[DC_PRED] = 1;
-    lfi->mode_lf_lut[V_PRED] = 1;
-    lfi->mode_lf_lut[H_PRED] = 1;
-    lfi->mode_lf_lut[TM_PRED] = 1;
-    lfi->mode_lf_lut[B_PRED]  = 0;
+  lfi->mode_lf_lut[DC_PRED] = 1;
+  lfi->mode_lf_lut[V_PRED] = 1;
+  lfi->mode_lf_lut[H_PRED] = 1;
+  lfi->mode_lf_lut[TM_PRED] = 1;
+  lfi->mode_lf_lut[B_PRED] = 0;
 
-    lfi->mode_lf_lut[ZEROMV]  = 1;
-    lfi->mode_lf_lut[NEARESTMV] = 2;
-    lfi->mode_lf_lut[NEARMV] = 2;
-    lfi->mode_lf_lut[NEWMV] = 2;
-    lfi->mode_lf_lut[SPLITMV] = 3;
-
+  lfi->mode_lf_lut[ZEROMV] = 1;
+  lfi->mode_lf_lut[NEARESTMV] = 2;
+  lfi->mode_lf_lut[NEARMV] = 2;
+  lfi->mode_lf_lut[NEWMV] = 2;
+  lfi->mode_lf_lut[SPLITMV] = 3;
 }
 
 void vp8_loop_filter_update_sharpness(loop_filter_info_n *lfi,
-                                      int sharpness_lvl)
-{
-    int i;
+                                      int sharpness_lvl) {
+  int i;
 
-    /* For each possible value for the loop filter fill out limits */
-    for (i = 0; i <= MAX_LOOP_FILTER; i++)
-    {
-        int filt_lvl = i;
-        int block_inside_limit = 0;
+  /* For each possible value for the loop filter fill out limits */
+  for (i = 0; i <= MAX_LOOP_FILTER; i++) {
+    int filt_lvl = i;
+    int block_inside_limit = 0;
 
-        /* Set loop filter paramaeters that control sharpness. */
-        block_inside_limit = filt_lvl >> (sharpness_lvl > 0);
-        block_inside_limit = block_inside_limit >> (sharpness_lvl > 4);
+    /* Set loop filter paramaeters that control sharpness. */
+    block_inside_limit = filt_lvl >> (sharpness_lvl > 0);
+    block_inside_limit = block_inside_limit >> (sharpness_lvl > 4);
 
-        if (sharpness_lvl > 0)
-        {
-            if (block_inside_limit > (9 - sharpness_lvl))
-                block_inside_limit = (9 - sharpness_lvl);
-        }
+    if (sharpness_lvl > 0) {
+      if (block_inside_limit > (9 - sharpness_lvl))
+        block_inside_limit = (9 - sharpness_lvl);
+    }
 
-        if (block_inside_limit < 1)
-            block_inside_limit = 1;
+    if (block_inside_limit < 1) block_inside_limit = 1;
 
-        memset(lfi->lim[i], block_inside_limit, SIMD_WIDTH);
-        memset(lfi->blim[i], (2 * filt_lvl + block_inside_limit), SIMD_WIDTH);
-        memset(lfi->mblim[i], (2 * (filt_lvl + 2) + block_inside_limit),
-               SIMD_WIDTH);
-    }
+    memset(lfi->lim[i], block_inside_limit, SIMD_WIDTH);
+    memset(lfi->blim[i], (2 * filt_lvl + block_inside_limit), SIMD_WIDTH);
+    memset(lfi->mblim[i], (2 * (filt_lvl + 2) + block_inside_limit),
+           SIMD_WIDTH);
+  }
 }
 
-void vp8_loop_filter_init(VP8_COMMON *cm)
-{
-    loop_filter_info_n *lfi = &cm->lf_info;
-    int i;
+void vp8_loop_filter_init(VP8_COMMON *cm) {
+  loop_filter_info_n *lfi = &cm->lf_info;
+  int i;
 
-    /* init limits for given sharpness*/
-    vp8_loop_filter_update_sharpness(lfi, cm->sharpness_level);
-    cm->last_sharpness_level = cm->sharpness_level;
+  /* init limits for given sharpness*/
+  vp8_loop_filter_update_sharpness(lfi, cm->sharpness_level);
+  cm->last_sharpness_level = cm->sharpness_level;
 
-    /* init LUT for lvl  and hev thr picking */
-    lf_init_lut(lfi);
+  /* init LUT for lvl  and hev thr picking */
+  lf_init_lut(lfi);
 
-    /* init hev threshold const vectors */
-    for(i = 0; i < 4 ; i++)
-    {
-        memset(lfi->hev_thr[i], i, SIMD_WIDTH);
-    }
+  /* init hev threshold const vectors */
+  for (i = 0; i < 4; i++) {
+    memset(lfi->hev_thr[i], i, SIMD_WIDTH);
+  }
 }
 
-void vp8_loop_filter_frame_init(VP8_COMMON *cm,
-                                MACROBLOCKD *mbd,
-                                int default_filt_lvl)
-{
-    int seg,  /* segment number */
-        ref,  /* index in ref_lf_deltas */
-        mode; /* index in mode_lf_deltas */
+void vp8_loop_filter_frame_init(VP8_COMMON *cm, MACROBLOCKD *mbd,
+                                int default_filt_lvl) {
+  int seg,  /* segment number */
+      ref,  /* index in ref_lf_deltas */
+      mode; /* index in mode_lf_deltas */
 
-    loop_filter_info_n *lfi = &cm->lf_info;
+  loop_filter_info_n *lfi = &cm->lf_info;
 
-    /* update limits if sharpness has changed */
-    if(cm->last_sharpness_level != cm->sharpness_level)
-    {
-        vp8_loop_filter_update_sharpness(lfi, cm->sharpness_level);
-        cm->last_sharpness_level = cm->sharpness_level;
+  /* update limits if sharpness has changed */
+  if (cm->last_sharpness_level != cm->sharpness_level) {
+    vp8_loop_filter_update_sharpness(lfi, cm->sharpness_level);
+    cm->last_sharpness_level = cm->sharpness_level;
+  }
+
+  for (seg = 0; seg < MAX_MB_SEGMENTS; seg++) {
+    int lvl_seg = default_filt_lvl;
+    int lvl_ref, lvl_mode;
+
+    /* Note the baseline filter values for each segment */
+    if (mbd->segmentation_enabled) {
+      /* Abs value */
+      if (mbd->mb_segement_abs_delta == SEGMENT_ABSDATA) {
+        lvl_seg = mbd->segment_feature_data[MB_LVL_ALT_LF][seg];
+      } else /* Delta Value */
+      {
+        lvl_seg += mbd->segment_feature_data[MB_LVL_ALT_LF][seg];
+      }
+      lvl_seg = (lvl_seg > 0) ? ((lvl_seg > 63) ? 63 : lvl_seg) : 0;
     }
 
-    for(seg = 0; seg < MAX_MB_SEGMENTS; seg++)
-    {
-        int lvl_seg = default_filt_lvl;
-        int lvl_ref, lvl_mode;
+    if (!mbd->mode_ref_lf_delta_enabled) {
+      /* we could get rid of this if we assume that deltas are set to
+       * zero when not in use; encoder always uses deltas
+       */
+      memset(lfi->lvl[seg][0], lvl_seg, 4 * 4);
+      continue;
+    }
 
-        /* Note the baseline filter values for each segment */
-        if (mbd->segmentation_enabled)
-        {
-            /* Abs value */
-            if (mbd->mb_segement_abs_delta == SEGMENT_ABSDATA)
-            {
-                lvl_seg = mbd->segment_feature_data[MB_LVL_ALT_LF][seg];
-            }
-            else  /* Delta Value */
-            {
-                lvl_seg += mbd->segment_feature_data[MB_LVL_ALT_LF][seg];
-            }
-            lvl_seg = (lvl_seg > 0) ? ((lvl_seg > 63) ? 63: lvl_seg) : 0;
-        }
+    /* INTRA_FRAME */
+    ref = INTRA_FRAME;
 
-        if (!mbd->mode_ref_lf_delta_enabled)
-        {
-            /* we could get rid of this if we assume that deltas are set to
-             * zero when not in use; encoder always uses deltas
-             */
-            memset(lfi->lvl[seg][0], lvl_seg, 4 * 4 );
-            continue;
-        }
+    /* Apply delta for reference frame */
+    lvl_ref = lvl_seg + mbd->ref_lf_deltas[ref];
 
-        /* INTRA_FRAME */
-        ref = INTRA_FRAME;
+    /* Apply delta for Intra modes */
+    mode = 0; /* B_PRED */
+    /* Only the split mode BPRED has a further special case */
+    lvl_mode = lvl_ref + mbd->mode_lf_deltas[mode];
+    /* clamp */
+    lvl_mode = (lvl_mode > 0) ? (lvl_mode > 63 ? 63 : lvl_mode) : 0;
 
-        /* Apply delta for reference frame */
-        lvl_ref = lvl_seg + mbd->ref_lf_deltas[ref];
+    lfi->lvl[seg][ref][mode] = lvl_mode;
 
-        /* Apply delta for Intra modes */
-        mode = 0; /* B_PRED */
-        /* Only the split mode BPRED has a further special case */
+    mode = 1; /* all the rest of Intra modes */
+    /* clamp */
+    lvl_mode = (lvl_ref > 0) ? (lvl_ref > 63 ? 63 : lvl_ref) : 0;
+    lfi->lvl[seg][ref][mode] = lvl_mode;
+
+    /* LAST, GOLDEN, ALT */
+    for (ref = 1; ref < MAX_REF_FRAMES; ref++) {
+      /* Apply delta for reference frame */
+      lvl_ref = lvl_seg + mbd->ref_lf_deltas[ref];
+
+      /* Apply delta for Inter modes */
+      for (mode = 1; mode < 4; mode++) {
         lvl_mode = lvl_ref + mbd->mode_lf_deltas[mode];
         /* clamp */
         lvl_mode = (lvl_mode > 0) ? (lvl_mode > 63 ? 63 : lvl_mode) : 0;
 
         lfi->lvl[seg][ref][mode] = lvl_mode;
-
-        mode = 1; /* all the rest of Intra modes */
-        /* clamp */
-        lvl_mode = (lvl_ref > 0) ? (lvl_ref > 63 ? 63 : lvl_ref) : 0;
-        lfi->lvl[seg][ref][mode] = lvl_mode;
-
-        /* LAST, GOLDEN, ALT */
-        for(ref = 1; ref < MAX_REF_FRAMES; ref++)
-        {
-            /* Apply delta for reference frame */
-            lvl_ref = lvl_seg + mbd->ref_lf_deltas[ref];
-
-            /* Apply delta for Inter modes */
-            for (mode = 1; mode < 4; mode++)
-            {
-                lvl_mode = lvl_ref + mbd->mode_lf_deltas[mode];
-                /* clamp */
-                lvl_mode = (lvl_mode > 0) ? (lvl_mode > 63 ? 63 : lvl_mode) : 0;
-
-                lfi->lvl[seg][ref][mode] = lvl_mode;
-            }
-        }
+      }
     }
+  }
 }
 
-
 void vp8_loop_filter_row_normal(VP8_COMMON *cm, MODE_INFO *mode_info_context,
-                         int mb_row, int post_ystride, int post_uvstride,
-                         unsigned char *y_ptr, unsigned char *u_ptr,
-                         unsigned char *v_ptr)
-{
-    int mb_col;
-    int filter_level;
-    loop_filter_info_n *lfi_n = &cm->lf_info;
-    loop_filter_info lfi;
-    FRAME_TYPE frame_type = cm->frame_type;
+                                int mb_row, int post_ystride, int post_uvstride,
+                                unsigned char *y_ptr, unsigned char *u_ptr,
+                                unsigned char *v_ptr) {
+  int mb_col;
+  int filter_level;
+  loop_filter_info_n *lfi_n = &cm->lf_info;
+  loop_filter_info lfi;
+  FRAME_TYPE frame_type = cm->frame_type;
 
-    for (mb_col = 0; mb_col < cm->mb_cols; mb_col++)
-    {
-        int skip_lf = (mode_info_context->mbmi.mode != B_PRED &&
-                        mode_info_context->mbmi.mode != SPLITMV &&
-                        mode_info_context->mbmi.mb_skip_coeff);
+  for (mb_col = 0; mb_col < cm->mb_cols; mb_col++) {
+    int skip_lf = (mode_info_context->mbmi.mode != B_PRED &&
+                   mode_info_context->mbmi.mode != SPLITMV &&
+                   mode_info_context->mbmi.mb_skip_coeff);
 
-        const int mode_index = lfi_n->mode_lf_lut[mode_info_context->mbmi.mode];
-        const int seg = mode_info_context->mbmi.segment_id;
-        const int ref_frame = mode_info_context->mbmi.ref_frame;
+    const int mode_index = lfi_n->mode_lf_lut[mode_info_context->mbmi.mode];
+    const int seg = mode_info_context->mbmi.segment_id;
+    const int ref_frame = mode_info_context->mbmi.ref_frame;
 
-        filter_level = lfi_n->lvl[seg][ref_frame][mode_index];
+    filter_level = lfi_n->lvl[seg][ref_frame][mode_index];
 
-        if (filter_level)
-        {
-            const int hev_index = lfi_n->hev_thr_lut[frame_type][filter_level];
-            lfi.mblim = lfi_n->mblim[filter_level];
-            lfi.blim = lfi_n->blim[filter_level];
-            lfi.lim = lfi_n->lim[filter_level];
-            lfi.hev_thr = lfi_n->hev_thr[hev_index];
+    if (filter_level) {
+      const int hev_index = lfi_n->hev_thr_lut[frame_type][filter_level];
+      lfi.mblim = lfi_n->mblim[filter_level];
+      lfi.blim = lfi_n->blim[filter_level];
+      lfi.lim = lfi_n->lim[filter_level];
+      lfi.hev_thr = lfi_n->hev_thr[hev_index];
 
-            if (mb_col > 0)
-                vp8_loop_filter_mbv
-                (y_ptr, u_ptr, v_ptr, post_ystride, post_uvstride, &lfi);
+      if (mb_col > 0)
+        vp8_loop_filter_mbv(y_ptr, u_ptr, v_ptr, post_ystride, post_uvstride,
+                            &lfi);
 
-            if (!skip_lf)
-                vp8_loop_filter_bv
-                (y_ptr, u_ptr, v_ptr, post_ystride, post_uvstride, &lfi);
+      if (!skip_lf)
+        vp8_loop_filter_bv(y_ptr, u_ptr, v_ptr, post_ystride, post_uvstride,
+                           &lfi);
 
-            /* don't apply across umv border */
-            if (mb_row > 0)
-                vp8_loop_filter_mbh
-                (y_ptr, u_ptr, v_ptr, post_ystride, post_uvstride, &lfi);
+      /* don't apply across umv border */
+      if (mb_row > 0)
+        vp8_loop_filter_mbh(y_ptr, u_ptr, v_ptr, post_ystride, post_uvstride,
+                            &lfi);
 
-            if (!skip_lf)
-                vp8_loop_filter_bh
-                (y_ptr, u_ptr, v_ptr, post_ystride, post_uvstride, &lfi);
-        }
-
-        y_ptr += 16;
-        u_ptr += 8;
-        v_ptr += 8;
-
-        mode_info_context++;     /* step to next MB */
+      if (!skip_lf)
+        vp8_loop_filter_bh(y_ptr, u_ptr, v_ptr, post_ystride, post_uvstride,
+                           &lfi);
     }
 
+    y_ptr += 16;
+    u_ptr += 8;
+    v_ptr += 8;
+
+    mode_info_context++; /* step to next MB */
+  }
 }
 
 void vp8_loop_filter_row_simple(VP8_COMMON *cm, MODE_INFO *mode_info_context,
-                         int mb_row, int post_ystride, int post_uvstride,
-                         unsigned char *y_ptr, unsigned char *u_ptr,
-                         unsigned char *v_ptr)
-{
-    int mb_col;
-    int filter_level;
-    loop_filter_info_n *lfi_n = &cm->lf_info;
-    (void)post_uvstride;
+                                int mb_row, int post_ystride, int post_uvstride,
+                                unsigned char *y_ptr, unsigned char *u_ptr,
+                                unsigned char *v_ptr) {
+  int mb_col;
+  int filter_level;
+  loop_filter_info_n *lfi_n = &cm->lf_info;
+  (void)post_uvstride;
 
-    for (mb_col = 0; mb_col < cm->mb_cols; mb_col++)
-    {
-        int skip_lf = (mode_info_context->mbmi.mode != B_PRED &&
-                        mode_info_context->mbmi.mode != SPLITMV &&
-                        mode_info_context->mbmi.mb_skip_coeff);
+  for (mb_col = 0; mb_col < cm->mb_cols; mb_col++) {
+    int skip_lf = (mode_info_context->mbmi.mode != B_PRED &&
+                   mode_info_context->mbmi.mode != SPLITMV &&
+                   mode_info_context->mbmi.mb_skip_coeff);
 
-        const int mode_index = lfi_n->mode_lf_lut[mode_info_context->mbmi.mode];
-        const int seg = mode_info_context->mbmi.segment_id;
-        const int ref_frame = mode_info_context->mbmi.ref_frame;
+    const int mode_index = lfi_n->mode_lf_lut[mode_info_context->mbmi.mode];
+    const int seg = mode_info_context->mbmi.segment_id;
+    const int ref_frame = mode_info_context->mbmi.ref_frame;
 
-        filter_level = lfi_n->lvl[seg][ref_frame][mode_index];
+    filter_level = lfi_n->lvl[seg][ref_frame][mode_index];
 
-        if (filter_level)
-        {
-            if (mb_col > 0)
-                vp8_loop_filter_simple_mbv
-                (y_ptr, post_ystride, lfi_n->mblim[filter_level]);
+    if (filter_level) {
+      if (mb_col > 0)
+        vp8_loop_filter_simple_mbv(y_ptr, post_ystride,
+                                   lfi_n->mblim[filter_level]);
 
-            if (!skip_lf)
-                vp8_loop_filter_simple_bv
-                (y_ptr, post_ystride, lfi_n->blim[filter_level]);
+      if (!skip_lf)
+        vp8_loop_filter_simple_bv(y_ptr, post_ystride,
+                                  lfi_n->blim[filter_level]);
 
-            /* don't apply across umv border */
-            if (mb_row > 0)
-                vp8_loop_filter_simple_mbh
-                (y_ptr, post_ystride, lfi_n->mblim[filter_level]);
+      /* don't apply across umv border */
+      if (mb_row > 0)
+        vp8_loop_filter_simple_mbh(y_ptr, post_ystride,
+                                   lfi_n->mblim[filter_level]);
 
-            if (!skip_lf)
-                vp8_loop_filter_simple_bh
-                (y_ptr, post_ystride, lfi_n->blim[filter_level]);
-        }
-
-        y_ptr += 16;
-        u_ptr += 8;
-        v_ptr += 8;
-
-        mode_info_context++;     /* step to next MB */
+      if (!skip_lf)
+        vp8_loop_filter_simple_bh(y_ptr, post_ystride,
+                                  lfi_n->blim[filter_level]);
     }
 
+    y_ptr += 16;
+    u_ptr += 8;
+    v_ptr += 8;
+
+    mode_info_context++; /* step to next MB */
+  }
 }
-void vp8_loop_filter_frame(VP8_COMMON *cm,
-                           MACROBLOCKD *mbd,
-                           int frame_type)
-{
-    YV12_BUFFER_CONFIG *post = cm->frame_to_show;
-    loop_filter_info_n *lfi_n = &cm->lf_info;
-    loop_filter_info lfi;
+void vp8_loop_filter_frame(VP8_COMMON *cm, MACROBLOCKD *mbd, int frame_type) {
+  YV12_BUFFER_CONFIG *post = cm->frame_to_show;
+  loop_filter_info_n *lfi_n = &cm->lf_info;
+  loop_filter_info lfi;
 
-    int mb_row;
-    int mb_col;
-    int mb_rows = cm->mb_rows;
-    int mb_cols = cm->mb_cols;
+  int mb_row;
+  int mb_col;
+  int mb_rows = cm->mb_rows;
+  int mb_cols = cm->mb_cols;
 
-    int filter_level;
+  int filter_level;
 
-    unsigned char *y_ptr, *u_ptr, *v_ptr;
+  unsigned char *y_ptr, *u_ptr, *v_ptr;
 
-    /* Point at base of Mb MODE_INFO list */
-    const MODE_INFO *mode_info_context = cm->mi;
-    int post_y_stride = post->y_stride;
-    int post_uv_stride = post->uv_stride;
+  /* Point at base of Mb MODE_INFO list */
+  const MODE_INFO *mode_info_context = cm->mi;
+  int post_y_stride = post->y_stride;
+  int post_uv_stride = post->uv_stride;
 
-    /* Initialize the loop filter for this frame. */
-    vp8_loop_filter_frame_init(cm, mbd, cm->filter_level);
+  /* Initialize the loop filter for this frame. */
+  vp8_loop_filter_frame_init(cm, mbd, cm->filter_level);
 
-    /* Set up the buffer pointers */
-    y_ptr = post->y_buffer;
-    u_ptr = post->u_buffer;
-    v_ptr = post->v_buffer;
+  /* Set up the buffer pointers */
+  y_ptr = post->y_buffer;
+  u_ptr = post->u_buffer;
+  v_ptr = post->v_buffer;
 
-    /* vp8_filter each macro block */
-    if (cm->filter_type == NORMAL_LOOPFILTER)
-    {
-        for (mb_row = 0; mb_row < mb_rows; mb_row++)
-        {
-            for (mb_col = 0; mb_col < mb_cols; mb_col++)
-            {
-                int skip_lf = (mode_info_context->mbmi.mode != B_PRED &&
-                                mode_info_context->mbmi.mode != SPLITMV &&
-                                mode_info_context->mbmi.mb_skip_coeff);
+  /* vp8_filter each macro block */
+  if (cm->filter_type == NORMAL_LOOPFILTER) {
+    for (mb_row = 0; mb_row < mb_rows; mb_row++) {
+      for (mb_col = 0; mb_col < mb_cols; mb_col++) {
+        int skip_lf = (mode_info_context->mbmi.mode != B_PRED &&
+                       mode_info_context->mbmi.mode != SPLITMV &&
+                       mode_info_context->mbmi.mb_skip_coeff);
 
-                const int mode_index = lfi_n->mode_lf_lut[mode_info_context->mbmi.mode];
-                const int seg = mode_info_context->mbmi.segment_id;
-                const int ref_frame = mode_info_context->mbmi.ref_frame;
+        const int mode_index = lfi_n->mode_lf_lut[mode_info_context->mbmi.mode];
+        const int seg = mode_info_context->mbmi.segment_id;
+        const int ref_frame = mode_info_context->mbmi.ref_frame;
 
-                filter_level = lfi_n->lvl[seg][ref_frame][mode_index];
+        filter_level = lfi_n->lvl[seg][ref_frame][mode_index];
 
-                if (filter_level)
-                {
-                    const int hev_index = lfi_n->hev_thr_lut[frame_type][filter_level];
-                    lfi.mblim = lfi_n->mblim[filter_level];
-                    lfi.blim = lfi_n->blim[filter_level];
-                    lfi.lim = lfi_n->lim[filter_level];
-                    lfi.hev_thr = lfi_n->hev_thr[hev_index];
+        if (filter_level) {
+          const int hev_index = lfi_n->hev_thr_lut[frame_type][filter_level];
+          lfi.mblim = lfi_n->mblim[filter_level];
+          lfi.blim = lfi_n->blim[filter_level];
+          lfi.lim = lfi_n->lim[filter_level];
+          lfi.hev_thr = lfi_n->hev_thr[hev_index];
 
-                    if (mb_col > 0)
-                        vp8_loop_filter_mbv
-                        (y_ptr, u_ptr, v_ptr, post_y_stride, post_uv_stride, &lfi);
+          if (mb_col > 0)
+            vp8_loop_filter_mbv(y_ptr, u_ptr, v_ptr, post_y_stride,
+                                post_uv_stride, &lfi);
 
-                    if (!skip_lf)
-                        vp8_loop_filter_bv
-                        (y_ptr, u_ptr, v_ptr, post_y_stride, post_uv_stride, &lfi);
+          if (!skip_lf)
+            vp8_loop_filter_bv(y_ptr, u_ptr, v_ptr, post_y_stride,
+                               post_uv_stride, &lfi);
 
-                    /* don't apply across umv border */
-                    if (mb_row > 0)
-                        vp8_loop_filter_mbh
-                        (y_ptr, u_ptr, v_ptr, post_y_stride, post_uv_stride, &lfi);
+          /* don't apply across umv border */
+          if (mb_row > 0)
+            vp8_loop_filter_mbh(y_ptr, u_ptr, v_ptr, post_y_stride,
+                                post_uv_stride, &lfi);
 
-                    if (!skip_lf)
-                        vp8_loop_filter_bh
-                        (y_ptr, u_ptr, v_ptr, post_y_stride, post_uv_stride, &lfi);
-                }
+          if (!skip_lf)
+            vp8_loop_filter_bh(y_ptr, u_ptr, v_ptr, post_y_stride,
+                               post_uv_stride, &lfi);
+        }
 
-                y_ptr += 16;
-                u_ptr += 8;
-                v_ptr += 8;
+        y_ptr += 16;
+        u_ptr += 8;
+        v_ptr += 8;
 
-                mode_info_context++;     /* step to next MB */
-            }
-            y_ptr += post_y_stride  * 16 - post->y_width;
-            u_ptr += post_uv_stride *  8 - post->uv_width;
-            v_ptr += post_uv_stride *  8 - post->uv_width;
+        mode_info_context++; /* step to next MB */
+      }
+      y_ptr += post_y_stride * 16 - post->y_width;
+      u_ptr += post_uv_stride * 8 - post->uv_width;
+      v_ptr += post_uv_stride * 8 - post->uv_width;
 
-            mode_info_context++;         /* Skip border mb */
-
-        }
+      mode_info_context++; /* Skip border mb */
     }
-    else /* SIMPLE_LOOPFILTER */
-    {
-        for (mb_row = 0; mb_row < mb_rows; mb_row++)
-        {
-            for (mb_col = 0; mb_col < mb_cols; mb_col++)
-            {
-                int skip_lf = (mode_info_context->mbmi.mode != B_PRED &&
-                                mode_info_context->mbmi.mode != SPLITMV &&
-                                mode_info_context->mbmi.mb_skip_coeff);
+  } else /* SIMPLE_LOOPFILTER */
+  {
+    for (mb_row = 0; mb_row < mb_rows; mb_row++) {
+      for (mb_col = 0; mb_col < mb_cols; mb_col++) {
+        int skip_lf = (mode_info_context->mbmi.mode != B_PRED &&
+                       mode_info_context->mbmi.mode != SPLITMV &&
+                       mode_info_context->mbmi.mb_skip_coeff);
 
-                const int mode_index = lfi_n->mode_lf_lut[mode_info_context->mbmi.mode];
-                const int seg = mode_info_context->mbmi.segment_id;
-                const int ref_frame = mode_info_context->mbmi.ref_frame;
+        const int mode_index = lfi_n->mode_lf_lut[mode_info_context->mbmi.mode];
+        const int seg = mode_info_context->mbmi.segment_id;
+        const int ref_frame = mode_info_context->mbmi.ref_frame;
 
-                filter_level = lfi_n->lvl[seg][ref_frame][mode_index];
-                if (filter_level)
-                {
-                    const unsigned char * mblim = lfi_n->mblim[filter_level];
-                    const unsigned char * blim = lfi_n->blim[filter_level];
+        filter_level = lfi_n->lvl[seg][ref_frame][mode_index];
+        if (filter_level) {
+          const unsigned char *mblim = lfi_n->mblim[filter_level];
+          const unsigned char *blim = lfi_n->blim[filter_level];
 
-                    if (mb_col > 0)
-                        vp8_loop_filter_simple_mbv
-                        (y_ptr, post_y_stride, mblim);
+          if (mb_col > 0)
+            vp8_loop_filter_simple_mbv(y_ptr, post_y_stride, mblim);
 
-                    if (!skip_lf)
-                        vp8_loop_filter_simple_bv
-                        (y_ptr, post_y_stride, blim);
+          if (!skip_lf) vp8_loop_filter_simple_bv(y_ptr, post_y_stride, blim);
 
-                    /* don't apply across umv border */
-                    if (mb_row > 0)
-                        vp8_loop_filter_simple_mbh
-                        (y_ptr, post_y_stride, mblim);
+          /* don't apply across umv border */
+          if (mb_row > 0)
+            vp8_loop_filter_simple_mbh(y_ptr, post_y_stride, mblim);
 
-                    if (!skip_lf)
-                        vp8_loop_filter_simple_bh
-                        (y_ptr, post_y_stride, blim);
-                }
+          if (!skip_lf) vp8_loop_filter_simple_bh(y_ptr, post_y_stride, blim);
+        }
 
-                y_ptr += 16;
-                u_ptr += 8;
-                v_ptr += 8;
+        y_ptr += 16;
+        u_ptr += 8;
+        v_ptr += 8;
 
-                mode_info_context++;     /* step to next MB */
-            }
-            y_ptr += post_y_stride  * 16 - post->y_width;
-            u_ptr += post_uv_stride *  8 - post->uv_width;
-            v_ptr += post_uv_stride *  8 - post->uv_width;
+        mode_info_context++; /* step to next MB */
+      }
+      y_ptr += post_y_stride * 16 - post->y_width;
+      u_ptr += post_uv_stride * 8 - post->uv_width;
+      v_ptr += post_uv_stride * 8 - post->uv_width;
 
-            mode_info_context++;         /* Skip border mb */
-
-        }
+      mode_info_context++; /* Skip border mb */
     }
+  }
 }
 
-void vp8_loop_filter_frame_yonly
-(
-    VP8_COMMON *cm,
-    MACROBLOCKD *mbd,
-    int default_filt_lvl
-)
-{
-    YV12_BUFFER_CONFIG *post = cm->frame_to_show;
+void vp8_loop_filter_frame_yonly(VP8_COMMON *cm, MACROBLOCKD *mbd,
+                                 int default_filt_lvl) {
+  YV12_BUFFER_CONFIG *post = cm->frame_to_show;
 
-    unsigned char *y_ptr;
-    int mb_row;
-    int mb_col;
+  unsigned char *y_ptr;
+  int mb_row;
+  int mb_col;
 
-    loop_filter_info_n *lfi_n = &cm->lf_info;
-    loop_filter_info lfi;
+  loop_filter_info_n *lfi_n = &cm->lf_info;
+  loop_filter_info lfi;
 
-    int filter_level;
-    FRAME_TYPE frame_type = cm->frame_type;
+  int filter_level;
+  FRAME_TYPE frame_type = cm->frame_type;
 
-    /* Point at base of Mb MODE_INFO list */
-    const MODE_INFO *mode_info_context = cm->mi;
+  /* Point at base of Mb MODE_INFO list */
+  const MODE_INFO *mode_info_context = cm->mi;
 
 #if 0
     if(default_filt_lvl == 0) /* no filter applied */
@@ -470,110 +409,93 @@
         return;
 #endif
 
-    /* Initialize the loop filter for this frame. */
-    vp8_loop_filter_frame_init( cm, mbd, default_filt_lvl);
+  /* Initialize the loop filter for this frame. */
+  vp8_loop_filter_frame_init(cm, mbd, default_filt_lvl);
 
-    /* Set up the buffer pointers */
-    y_ptr = post->y_buffer;
+  /* Set up the buffer pointers */
+  y_ptr = post->y_buffer;
 
-    /* vp8_filter each macro block */
-    for (mb_row = 0; mb_row < cm->mb_rows; mb_row++)
-    {
-        for (mb_col = 0; mb_col < cm->mb_cols; mb_col++)
-        {
-            int skip_lf = (mode_info_context->mbmi.mode != B_PRED &&
-                            mode_info_context->mbmi.mode != SPLITMV &&
-                            mode_info_context->mbmi.mb_skip_coeff);
+  /* vp8_filter each macro block */
+  for (mb_row = 0; mb_row < cm->mb_rows; mb_row++) {
+    for (mb_col = 0; mb_col < cm->mb_cols; mb_col++) {
+      int skip_lf = (mode_info_context->mbmi.mode != B_PRED &&
+                     mode_info_context->mbmi.mode != SPLITMV &&
+                     mode_info_context->mbmi.mb_skip_coeff);
 
-            const int mode_index = lfi_n->mode_lf_lut[mode_info_context->mbmi.mode];
-            const int seg = mode_info_context->mbmi.segment_id;
-            const int ref_frame = mode_info_context->mbmi.ref_frame;
+      const int mode_index = lfi_n->mode_lf_lut[mode_info_context->mbmi.mode];
+      const int seg = mode_info_context->mbmi.segment_id;
+      const int ref_frame = mode_info_context->mbmi.ref_frame;
 
-            filter_level = lfi_n->lvl[seg][ref_frame][mode_index];
+      filter_level = lfi_n->lvl[seg][ref_frame][mode_index];
 
-            if (filter_level)
-            {
-                if (cm->filter_type == NORMAL_LOOPFILTER)
-                {
-                    const int hev_index = lfi_n->hev_thr_lut[frame_type][filter_level];
-                    lfi.mblim = lfi_n->mblim[filter_level];
-                    lfi.blim = lfi_n->blim[filter_level];
-                    lfi.lim = lfi_n->lim[filter_level];
-                    lfi.hev_thr = lfi_n->hev_thr[hev_index];
+      if (filter_level) {
+        if (cm->filter_type == NORMAL_LOOPFILTER) {
+          const int hev_index = lfi_n->hev_thr_lut[frame_type][filter_level];
+          lfi.mblim = lfi_n->mblim[filter_level];
+          lfi.blim = lfi_n->blim[filter_level];
+          lfi.lim = lfi_n->lim[filter_level];
+          lfi.hev_thr = lfi_n->hev_thr[hev_index];
 
-                    if (mb_col > 0)
-                        vp8_loop_filter_mbv
-                        (y_ptr, 0, 0, post->y_stride, 0, &lfi);
+          if (mb_col > 0)
+            vp8_loop_filter_mbv(y_ptr, 0, 0, post->y_stride, 0, &lfi);
 
-                    if (!skip_lf)
-                        vp8_loop_filter_bv
-                        (y_ptr, 0, 0, post->y_stride, 0, &lfi);
+          if (!skip_lf)
+            vp8_loop_filter_bv(y_ptr, 0, 0, post->y_stride, 0, &lfi);
 
-                    /* don't apply across umv border */
-                    if (mb_row > 0)
-                        vp8_loop_filter_mbh
-                        (y_ptr, 0, 0, post->y_stride, 0, &lfi);
+          /* don't apply across umv border */
+          if (mb_row > 0)
+            vp8_loop_filter_mbh(y_ptr, 0, 0, post->y_stride, 0, &lfi);
 
-                    if (!skip_lf)
-                        vp8_loop_filter_bh
-                        (y_ptr, 0, 0, post->y_stride, 0, &lfi);
-                }
-                else
-                {
-                    if (mb_col > 0)
-                        vp8_loop_filter_simple_mbv
-                        (y_ptr, post->y_stride, lfi_n->mblim[filter_level]);
+          if (!skip_lf)
+            vp8_loop_filter_bh(y_ptr, 0, 0, post->y_stride, 0, &lfi);
+        } else {
+          if (mb_col > 0)
+            vp8_loop_filter_simple_mbv(y_ptr, post->y_stride,
+                                       lfi_n->mblim[filter_level]);
 
-                    if (!skip_lf)
-                        vp8_loop_filter_simple_bv
-                        (y_ptr, post->y_stride, lfi_n->blim[filter_level]);
+          if (!skip_lf)
+            vp8_loop_filter_simple_bv(y_ptr, post->y_stride,
+                                      lfi_n->blim[filter_level]);
 
-                    /* don't apply across umv border */
-                    if (mb_row > 0)
-                        vp8_loop_filter_simple_mbh
-                        (y_ptr, post->y_stride, lfi_n->mblim[filter_level]);
+          /* don't apply across umv border */
+          if (mb_row > 0)
+            vp8_loop_filter_simple_mbh(y_ptr, post->y_stride,
+                                       lfi_n->mblim[filter_level]);
 
-                    if (!skip_lf)
-                        vp8_loop_filter_simple_bh
-                        (y_ptr, post->y_stride, lfi_n->blim[filter_level]);
-                }
-            }
-
-            y_ptr += 16;
-            mode_info_context ++;        /* step to next MB */
-
+          if (!skip_lf)
+            vp8_loop_filter_simple_bh(y_ptr, post->y_stride,
+                                      lfi_n->blim[filter_level]);
         }
+      }
 
-        y_ptr += post->y_stride  * 16 - post->y_width;
-        mode_info_context ++;            /* Skip border mb */
+      y_ptr += 16;
+      mode_info_context++; /* step to next MB */
     }
 
+    y_ptr += post->y_stride * 16 - post->y_width;
+    mode_info_context++; /* Skip border mb */
+  }
 }
 
-void vp8_loop_filter_partial_frame
-(
-    VP8_COMMON *cm,
-    MACROBLOCKD *mbd,
-    int default_filt_lvl
-)
-{
-    YV12_BUFFER_CONFIG *post = cm->frame_to_show;
+void vp8_loop_filter_partial_frame(VP8_COMMON *cm, MACROBLOCKD *mbd,
+                                   int default_filt_lvl) {
+  YV12_BUFFER_CONFIG *post = cm->frame_to_show;
 
-    unsigned char *y_ptr;
-    int mb_row;
-    int mb_col;
-    int mb_cols = post->y_width >> 4;
-    int mb_rows = post->y_height >> 4;
+  unsigned char *y_ptr;
+  int mb_row;
+  int mb_col;
+  int mb_cols = post->y_width >> 4;
+  int mb_rows = post->y_height >> 4;
 
-    int linestocopy;
+  int linestocopy;
 
-    loop_filter_info_n *lfi_n = &cm->lf_info;
-    loop_filter_info lfi;
+  loop_filter_info_n *lfi_n = &cm->lf_info;
+  loop_filter_info lfi;
 
-    int filter_level;
-    FRAME_TYPE frame_type = cm->frame_type;
+  int filter_level;
+  FRAME_TYPE frame_type = cm->frame_type;
 
-    const MODE_INFO *mode_info_context;
+  const MODE_INFO *mode_info_context;
 
 #if 0
     if(default_filt_lvl == 0) /* no filter applied */
@@ -580,82 +502,71 @@
         return;
 #endif
 
-    /* Initialize the loop filter for this frame. */
-    vp8_loop_filter_frame_init( cm, mbd, default_filt_lvl);
+  /* Initialize the loop filter for this frame. */
+  vp8_loop_filter_frame_init(cm, mbd, default_filt_lvl);
 
-    /* number of MB rows to use in partial filtering */
-    linestocopy = mb_rows / PARTIAL_FRAME_FRACTION;
-    linestocopy = linestocopy ? linestocopy << 4 : 16;     /* 16 lines per MB */
+  /* number of MB rows to use in partial filtering */
+  linestocopy = mb_rows / PARTIAL_FRAME_FRACTION;
+  linestocopy = linestocopy ? linestocopy << 4 : 16; /* 16 lines per MB */
 
-    /* Set up the buffer pointers; partial image starts at ~middle of frame */
-    y_ptr = post->y_buffer + ((post->y_height >> 5) * 16) * post->y_stride;
-    mode_info_context = cm->mi + (post->y_height >> 5) * (mb_cols + 1);
+  /* Set up the buffer pointers; partial image starts at ~middle of frame */
+  y_ptr = post->y_buffer + ((post->y_height >> 5) * 16) * post->y_stride;
+  mode_info_context = cm->mi + (post->y_height >> 5) * (mb_cols + 1);
 
-    /* vp8_filter each macro block */
-    for (mb_row = 0; mb_row<(linestocopy >> 4); mb_row++)
-    {
-        for (mb_col = 0; mb_col < mb_cols; mb_col++)
-        {
-            int skip_lf = (mode_info_context->mbmi.mode != B_PRED &&
-                           mode_info_context->mbmi.mode != SPLITMV &&
-                           mode_info_context->mbmi.mb_skip_coeff);
+  /* vp8_filter each macro block */
+  for (mb_row = 0; mb_row < (linestocopy >> 4); mb_row++) {
+    for (mb_col = 0; mb_col < mb_cols; mb_col++) {
+      int skip_lf = (mode_info_context->mbmi.mode != B_PRED &&
+                     mode_info_context->mbmi.mode != SPLITMV &&
+                     mode_info_context->mbmi.mb_skip_coeff);
 
-            const int mode_index =
-                lfi_n->mode_lf_lut[mode_info_context->mbmi.mode];
-            const int seg = mode_info_context->mbmi.segment_id;
-            const int ref_frame = mode_info_context->mbmi.ref_frame;
+      const int mode_index = lfi_n->mode_lf_lut[mode_info_context->mbmi.mode];
+      const int seg = mode_info_context->mbmi.segment_id;
+      const int ref_frame = mode_info_context->mbmi.ref_frame;
 
-            filter_level = lfi_n->lvl[seg][ref_frame][mode_index];
+      filter_level = lfi_n->lvl[seg][ref_frame][mode_index];
 
-            if (filter_level)
-            {
-                if (cm->filter_type == NORMAL_LOOPFILTER)
-                {
-                    const int hev_index = lfi_n->hev_thr_lut[frame_type][filter_level];
-                    lfi.mblim = lfi_n->mblim[filter_level];
-                    lfi.blim = lfi_n->blim[filter_level];
-                    lfi.lim = lfi_n->lim[filter_level];
-                    lfi.hev_thr = lfi_n->hev_thr[hev_index];
+      if (filter_level) {
+        if (cm->filter_type == NORMAL_LOOPFILTER) {
+          const int hev_index = lfi_n->hev_thr_lut[frame_type][filter_level];
+          lfi.mblim = lfi_n->mblim[filter_level];
+          lfi.blim = lfi_n->blim[filter_level];
+          lfi.lim = lfi_n->lim[filter_level];
+          lfi.hev_thr = lfi_n->hev_thr[hev_index];
 
-                    if (mb_col > 0)
-                        vp8_loop_filter_mbv
-                        (y_ptr, 0, 0, post->y_stride, 0, &lfi);
+          if (mb_col > 0)
+            vp8_loop_filter_mbv(y_ptr, 0, 0, post->y_stride, 0, &lfi);
 
-                    if (!skip_lf)
-                        vp8_loop_filter_bv
-                        (y_ptr, 0, 0, post->y_stride, 0, &lfi);
+          if (!skip_lf)
+            vp8_loop_filter_bv(y_ptr, 0, 0, post->y_stride, 0, &lfi);
 
-                    vp8_loop_filter_mbh
-                        (y_ptr, 0, 0, post->y_stride, 0, &lfi);
+          vp8_loop_filter_mbh(y_ptr, 0, 0, post->y_stride, 0, &lfi);
 
-                    if (!skip_lf)
-                        vp8_loop_filter_bh
-                        (y_ptr, 0, 0, post->y_stride, 0, &lfi);
-                }
-                else
-                {
-                    if (mb_col > 0)
-                        vp8_loop_filter_simple_mbv
-                        (y_ptr, post->y_stride, lfi_n->mblim[filter_level]);
+          if (!skip_lf)
+            vp8_loop_filter_bh(y_ptr, 0, 0, post->y_stride, 0, &lfi);
+        } else {
+          if (mb_col > 0)
+            vp8_loop_filter_simple_mbv(y_ptr, post->y_stride,
+                                       lfi_n->mblim[filter_level]);
 
-                    if (!skip_lf)
-                        vp8_loop_filter_simple_bv
-                        (y_ptr, post->y_stride, lfi_n->blim[filter_level]);
+          if (!skip_lf)
+            vp8_loop_filter_simple_bv(y_ptr, post->y_stride,
+                                      lfi_n->blim[filter_level]);
 
-                    vp8_loop_filter_simple_mbh
-                        (y_ptr, post->y_stride, lfi_n->mblim[filter_level]);
+          vp8_loop_filter_simple_mbh(y_ptr, post->y_stride,
+                                     lfi_n->mblim[filter_level]);
 
-                    if (!skip_lf)
-                        vp8_loop_filter_simple_bh
-                        (y_ptr, post->y_stride, lfi_n->blim[filter_level]);
-                }
-            }
-
-            y_ptr += 16;
-            mode_info_context += 1;      /* step to next MB */
+          if (!skip_lf)
+            vp8_loop_filter_simple_bh(y_ptr, post->y_stride,
+                                      lfi_n->blim[filter_level]);
         }
+      }
 
-        y_ptr += post->y_stride  * 16 - post->y_width;
-        mode_info_context += 1;          /* Skip border mb */
+      y_ptr += 16;
+      mode_info_context += 1; /* step to next MB */
     }
+
+    y_ptr += post->y_stride * 16 - post->y_width;
+    mode_info_context += 1; /* Skip border mb */
+  }
 }
--- a/vp8/common/x86/filter_x86.c
+++ b/vp8/common/x86/filter_x86.c
@@ -10,26 +10,20 @@
 
 #include "vp8/common/x86/filter_x86.h"
 
-DECLARE_ALIGNED(16, const short, vp8_bilinear_filters_x86_4[8][8]) =
-{
-    { 128, 128, 128, 128,   0,   0,   0,   0 },
-    { 112, 112, 112, 112,  16,  16,  16,  16 },
-    {  96,  96,  96,  96,  32,  32,  32,  32 },
-    {  80,  80,  80,  80,  48,  48,  48,  48 },
-    {  64,  64,  64,  64,  64,  64,  64,  64 },
-    {  48,  48,  48,  48,  80,  80,  80,  80 },
-    {  32,  32,  32,  32,  96,  96,  96,  96 },
-    {  16,  16,  16,  16, 112, 112, 112, 112 }
+DECLARE_ALIGNED(16, const short, vp8_bilinear_filters_x86_4[8][8]) = {
+  { 128, 128, 128, 128, 0, 0, 0, 0 }, { 112, 112, 112, 112, 16, 16, 16, 16 },
+  { 96, 96, 96, 96, 32, 32, 32, 32 }, { 80, 80, 80, 80, 48, 48, 48, 48 },
+  { 64, 64, 64, 64, 64, 64, 64, 64 }, { 48, 48, 48, 48, 80, 80, 80, 80 },
+  { 32, 32, 32, 32, 96, 96, 96, 96 }, { 16, 16, 16, 16, 112, 112, 112, 112 }
 };
 
-DECLARE_ALIGNED(16, const short, vp8_bilinear_filters_x86_8[8][16]) =
-{
-    { 128, 128, 128, 128, 128, 128, 128, 128,   0,   0,   0,   0,   0,   0,   0,   0 },
-    { 112, 112, 112, 112, 112, 112, 112, 112,  16,  16,  16,  16,  16,  16,  16,  16 },
-    {  96,  96,  96,  96,  96,  96,  96,  96,  32,  32,  32,  32,  32,  32,  32,  32 },
-    {  80,  80,  80,  80,  80,  80,  80,  80,  48,  48,  48,  48,  48,  48,  48,  48 },
-    {  64,  64,  64,  64,  64,  64,  64,  64,  64,  64,  64,  64,  64,  64,  64,  64 },
-    {  48,  48,  48,  48,  48,  48,  48,  48,  80,  80,  80,  80,  80,  80,  80,  80 },
-    {  32,  32,  32,  32,  32,  32,  32,  32,  96,  96,  96,  96,  96,  96,  96,  96 },
-    {  16,  16,  16,  16,  16,  16,  16,  16, 112, 112, 112, 112, 112, 112, 112, 112 }
+DECLARE_ALIGNED(16, const short, vp8_bilinear_filters_x86_8[8][16]) = {
+  { 128, 128, 128, 128, 128, 128, 128, 128, 0, 0, 0, 0, 0, 0, 0, 0 },
+  { 112, 112, 112, 112, 112, 112, 112, 112, 16, 16, 16, 16, 16, 16, 16, 16 },
+  { 96, 96, 96, 96, 96, 96, 96, 96, 32, 32, 32, 32, 32, 32, 32, 32 },
+  { 80, 80, 80, 80, 80, 80, 80, 80, 48, 48, 48, 48, 48, 48, 48, 48 },
+  { 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64 },
+  { 48, 48, 48, 48, 48, 48, 48, 48, 80, 80, 80, 80, 80, 80, 80, 80 },
+  { 32, 32, 32, 32, 32, 32, 32, 32, 96, 96, 96, 96, 96, 96, 96, 96 },
+  { 16, 16, 16, 16, 16, 16, 16, 16, 112, 112, 112, 112, 112, 112, 112, 112 }
 };
--- a/vp8/common/x86/idct_blk_mmx.c
+++ b/vp8/common/x86/idct_blk_mmx.c
@@ -15,114 +15,97 @@
 
 extern void vp8_dequantize_b_impl_mmx(short *sq, short *dq, short *q);
 
-void vp8_dequantize_b_mmx(BLOCKD *d, short *DQC)
-{
-    short *sq = (short *) d->qcoeff;
-    short *dq = (short *) d->dqcoeff;
+void vp8_dequantize_b_mmx(BLOCKD *d, short *DQC) {
+  short *sq = (short *)d->qcoeff;
+  short *dq = (short *)d->dqcoeff;
 
-    vp8_dequantize_b_impl_mmx(sq, dq, DQC);
+  vp8_dequantize_b_impl_mmx(sq, dq, DQC);
 }
 
-void vp8_dequant_idct_add_y_block_mmx
-            (short *q, short *dq,
-             unsigned char *dst, int stride, char *eobs)
-{
-    int i;
+void vp8_dequant_idct_add_y_block_mmx(short *q, short *dq, unsigned char *dst,
+                                      int stride, char *eobs) {
+  int i;
 
-    for (i = 0; i < 4; i++)
-    {
-        if (eobs[0] > 1)
-            vp8_dequant_idct_add_mmx (q, dq, dst, stride);
-        else if (eobs[0] == 1)
-        {
-            vp8_dc_only_idct_add_mmx (q[0]*dq[0], dst, stride, dst, stride);
-            memset(q, 0, 2 * sizeof(q[0]));
-        }
+  for (i = 0; i < 4; i++) {
+    if (eobs[0] > 1)
+      vp8_dequant_idct_add_mmx(q, dq, dst, stride);
+    else if (eobs[0] == 1) {
+      vp8_dc_only_idct_add_mmx(q[0] * dq[0], dst, stride, dst, stride);
+      memset(q, 0, 2 * sizeof(q[0]));
+    }
 
-        if (eobs[1] > 1)
-            vp8_dequant_idct_add_mmx (q+16, dq, dst+4, stride);
-        else if (eobs[1] == 1)
-        {
-            vp8_dc_only_idct_add_mmx (q[16]*dq[0], dst+4, stride,
-                                      dst+4, stride);
-            memset(q + 16, 0, 2 * sizeof(q[0]));
-        }
+    if (eobs[1] > 1)
+      vp8_dequant_idct_add_mmx(q + 16, dq, dst + 4, stride);
+    else if (eobs[1] == 1) {
+      vp8_dc_only_idct_add_mmx(q[16] * dq[0], dst + 4, stride, dst + 4, stride);
+      memset(q + 16, 0, 2 * sizeof(q[0]));
+    }
 
-        if (eobs[2] > 1)
-            vp8_dequant_idct_add_mmx (q+32, dq, dst+8, stride);
-        else if (eobs[2] == 1)
-        {
-            vp8_dc_only_idct_add_mmx (q[32]*dq[0], dst+8, stride,
-                                      dst+8, stride);
-            memset(q + 32, 0, 2 * sizeof(q[0]));
-        }
+    if (eobs[2] > 1)
+      vp8_dequant_idct_add_mmx(q + 32, dq, dst + 8, stride);
+    else if (eobs[2] == 1) {
+      vp8_dc_only_idct_add_mmx(q[32] * dq[0], dst + 8, stride, dst + 8, stride);
+      memset(q + 32, 0, 2 * sizeof(q[0]));
+    }
 
-        if (eobs[3] > 1)
-            vp8_dequant_idct_add_mmx (q+48, dq, dst+12, stride);
-        else if (eobs[3] == 1)
-        {
-            vp8_dc_only_idct_add_mmx (q[48]*dq[0], dst+12, stride,
-                                      dst+12, stride);
-            memset(q + 48, 0, 2 * sizeof(q[0]));
-        }
-
-        q    += 64;
-        dst  += 4*stride;
-        eobs += 4;
+    if (eobs[3] > 1)
+      vp8_dequant_idct_add_mmx(q + 48, dq, dst + 12, stride);
+    else if (eobs[3] == 1) {
+      vp8_dc_only_idct_add_mmx(q[48] * dq[0], dst + 12, stride, dst + 12,
+                               stride);
+      memset(q + 48, 0, 2 * sizeof(q[0]));
     }
+
+    q += 64;
+    dst += 4 * stride;
+    eobs += 4;
+  }
 }
 
-void vp8_dequant_idct_add_uv_block_mmx
-            (short *q, short *dq,
-             unsigned char *dstu, unsigned char *dstv, int stride, char *eobs)
-{
-    int i;
+void vp8_dequant_idct_add_uv_block_mmx(short *q, short *dq, unsigned char *dstu,
+                                       unsigned char *dstv, int stride,
+                                       char *eobs) {
+  int i;
 
-    for (i = 0; i < 2; i++)
-    {
-        if (eobs[0] > 1)
-            vp8_dequant_idct_add_mmx (q, dq, dstu, stride);
-        else if (eobs[0] == 1)
-        {
-            vp8_dc_only_idct_add_mmx (q[0]*dq[0], dstu, stride, dstu, stride);
-            memset(q, 0, 2 * sizeof(q[0]));
-        }
+  for (i = 0; i < 2; i++) {
+    if (eobs[0] > 1)
+      vp8_dequant_idct_add_mmx(q, dq, dstu, stride);
+    else if (eobs[0] == 1) {
+      vp8_dc_only_idct_add_mmx(q[0] * dq[0], dstu, stride, dstu, stride);
+      memset(q, 0, 2 * sizeof(q[0]));
+    }
 
-        if (eobs[1] > 1)
-            vp8_dequant_idct_add_mmx (q+16, dq, dstu+4, stride);
-        else if (eobs[1] == 1)
-        {
-            vp8_dc_only_idct_add_mmx (q[16]*dq[0], dstu+4, stride,
-                                      dstu+4, stride);
-            memset(q + 16, 0, 2 * sizeof(q[0]));
-        }
-
-        q    += 32;
-        dstu += 4*stride;
-        eobs += 2;
+    if (eobs[1] > 1)
+      vp8_dequant_idct_add_mmx(q + 16, dq, dstu + 4, stride);
+    else if (eobs[1] == 1) {
+      vp8_dc_only_idct_add_mmx(q[16] * dq[0], dstu + 4, stride, dstu + 4,
+                               stride);
+      memset(q + 16, 0, 2 * sizeof(q[0]));
     }
 
-    for (i = 0; i < 2; i++)
-    {
-        if (eobs[0] > 1)
-            vp8_dequant_idct_add_mmx (q, dq, dstv, stride);
-        else if (eobs[0] == 1)
-        {
-            vp8_dc_only_idct_add_mmx (q[0]*dq[0], dstv, stride, dstv, stride);
-            memset(q, 0, 2 * sizeof(q[0]));
-        }
+    q += 32;
+    dstu += 4 * stride;
+    eobs += 2;
+  }
 
-        if (eobs[1] > 1)
-            vp8_dequant_idct_add_mmx (q+16, dq, dstv+4, stride);
-        else if (eobs[1] == 1)
-        {
-            vp8_dc_only_idct_add_mmx (q[16]*dq[0], dstv+4, stride,
-                                      dstv+4, stride);
-            memset(q + 16, 0, 2 * sizeof(q[0]));
-        }
+  for (i = 0; i < 2; i++) {
+    if (eobs[0] > 1)
+      vp8_dequant_idct_add_mmx(q, dq, dstv, stride);
+    else if (eobs[0] == 1) {
+      vp8_dc_only_idct_add_mmx(q[0] * dq[0], dstv, stride, dstv, stride);
+      memset(q, 0, 2 * sizeof(q[0]));
+    }
 
-        q    += 32;
-        dstv += 4*stride;
-        eobs += 2;
+    if (eobs[1] > 1)
+      vp8_dequant_idct_add_mmx(q + 16, dq, dstv + 4, stride);
+    else if (eobs[1] == 1) {
+      vp8_dc_only_idct_add_mmx(q[16] * dq[0], dstv + 4, stride, dstv + 4,
+                               stride);
+      memset(q + 16, 0, 2 * sizeof(q[0]));
     }
+
+    q += 32;
+    dstv += 4 * stride;
+    eobs += 2;
+  }
 }
--- a/vp8/common/x86/idct_blk_sse2.c
+++ b/vp8/common/x86/idct_blk_sse2.c
@@ -11,79 +11,68 @@
 #include "vpx_config.h"
 #include "vp8_rtcd.h"
 
-void vp8_idct_dequant_0_2x_sse2
-            (short *q, short *dq ,
-             unsigned char *dst, int dst_stride);
-void vp8_idct_dequant_full_2x_sse2
-            (short *q, short *dq ,
-             unsigned char *dst, int dst_stride);
+void vp8_idct_dequant_0_2x_sse2(short *q, short *dq, unsigned char *dst,
+                                int dst_stride);
+void vp8_idct_dequant_full_2x_sse2(short *q, short *dq, unsigned char *dst,
+                                   int dst_stride);
 
-void vp8_dequant_idct_add_y_block_sse2
-            (short *q, short *dq,
-             unsigned char *dst, int stride, char *eobs)
-{
-    int i;
+void vp8_dequant_idct_add_y_block_sse2(short *q, short *dq, unsigned char *dst,
+                                       int stride, char *eobs) {
+  int i;
 
-    for (i = 0; i < 4; i++)
-    {
-        if (((short *)(eobs))[0])
-        {
-            if (((short *)(eobs))[0] & 0xfefe)
-                vp8_idct_dequant_full_2x_sse2 (q, dq, dst, stride);
-            else
-                vp8_idct_dequant_0_2x_sse2 (q, dq, dst, stride);
-        }
-        if (((short *)(eobs))[1])
-        {
-            if (((short *)(eobs))[1] & 0xfefe)
-                vp8_idct_dequant_full_2x_sse2 (q+32, dq, dst+8, stride);
-            else
-                vp8_idct_dequant_0_2x_sse2 (q+32, dq, dst+8, stride);
-        }
-        q    += 64;
-        dst  += stride*4;
-        eobs += 4;
+  for (i = 0; i < 4; i++) {
+    if (((short *)(eobs))[0]) {
+      if (((short *)(eobs))[0] & 0xfefe)
+        vp8_idct_dequant_full_2x_sse2(q, dq, dst, stride);
+      else
+        vp8_idct_dequant_0_2x_sse2(q, dq, dst, stride);
     }
+    if (((short *)(eobs))[1]) {
+      if (((short *)(eobs))[1] & 0xfefe)
+        vp8_idct_dequant_full_2x_sse2(q + 32, dq, dst + 8, stride);
+      else
+        vp8_idct_dequant_0_2x_sse2(q + 32, dq, dst + 8, stride);
+    }
+    q += 64;
+    dst += stride * 4;
+    eobs += 4;
+  }
 }
 
-void vp8_dequant_idct_add_uv_block_sse2
-            (short *q, short *dq,
-             unsigned char *dstu, unsigned char *dstv, int stride, char *eobs)
-{
-    if (((short *)(eobs))[0])
-    {
-        if (((short *)(eobs))[0] & 0xfefe)
-            vp8_idct_dequant_full_2x_sse2 (q, dq, dstu, stride);
-        else
-            vp8_idct_dequant_0_2x_sse2 (q, dq, dstu, stride);
-    }
-    q    += 32;
-    dstu += stride*4;
+void vp8_dequant_idct_add_uv_block_sse2(short *q, short *dq,
+                                        unsigned char *dstu,
+                                        unsigned char *dstv, int stride,
+                                        char *eobs) {
+  if (((short *)(eobs))[0]) {
+    if (((short *)(eobs))[0] & 0xfefe)
+      vp8_idct_dequant_full_2x_sse2(q, dq, dstu, stride);
+    else
+      vp8_idct_dequant_0_2x_sse2(q, dq, dstu, stride);
+  }
+  q += 32;
+  dstu += stride * 4;
 
-    if (((short *)(eobs))[1])
-    {
-        if (((short *)(eobs))[1] & 0xfefe)
-            vp8_idct_dequant_full_2x_sse2 (q, dq, dstu, stride);
-        else
-            vp8_idct_dequant_0_2x_sse2 (q, dq, dstu, stride);
-    }
-    q    += 32;
+  if (((short *)(eobs))[1]) {
+    if (((short *)(eobs))[1] & 0xfefe)
+      vp8_idct_dequant_full_2x_sse2(q, dq, dstu, stride);
+    else
+      vp8_idct_dequant_0_2x_sse2(q, dq, dstu, stride);
+  }
+  q += 32;
 
-    if (((short *)(eobs))[2])
-    {
-        if (((short *)(eobs))[2] & 0xfefe)
-            vp8_idct_dequant_full_2x_sse2 (q, dq, dstv, stride);
-        else
-            vp8_idct_dequant_0_2x_sse2 (q, dq, dstv, stride);
-    }
-    q    += 32;
-    dstv += stride*4;
+  if (((short *)(eobs))[2]) {
+    if (((short *)(eobs))[2] & 0xfefe)
+      vp8_idct_dequant_full_2x_sse2(q, dq, dstv, stride);
+    else
+      vp8_idct_dequant_0_2x_sse2(q, dq, dstv, stride);
+  }
+  q += 32;
+  dstv += stride * 4;
 
-    if (((short *)(eobs))[3])
-    {
-      if (((short *)(eobs))[3] & 0xfefe)
-          vp8_idct_dequant_full_2x_sse2 (q, dq, dstv, stride);
-      else
-          vp8_idct_dequant_0_2x_sse2 (q, dq, dstv, stride);
-    }
+  if (((short *)(eobs))[3]) {
+    if (((short *)(eobs))[3] & 0xfefe)
+      vp8_idct_dequant_full_2x_sse2(q, dq, dstv, stride);
+    else
+      vp8_idct_dequant_0_2x_sse2(q, dq, dstv, stride);
+  }
 }
--- a/vp8/common/x86/loopfilter_x86.c
+++ b/vp8/common/x86/loopfilter_x86.c
@@ -8,20 +8,19 @@
  *  be found in the AUTHORS file in the root of the source tree.
  */
 
-
 #include "vpx_config.h"
 #include "vp8/common/loopfilter.h"
 
-#define prototype_loopfilter(sym) \
-    void sym(unsigned char *src, int pitch, const unsigned char *blimit,\
-             const unsigned char *limit, const unsigned char *thresh, int count)
+#define prototype_loopfilter(sym)                                      \
+  void sym(unsigned char *src, int pitch, const unsigned char *blimit, \
+           const unsigned char *limit, const unsigned char *thresh, int count)
 
-#define prototype_loopfilter_nc(sym) \
-    void sym(unsigned char *src, int pitch, const unsigned char *blimit,\
-             const unsigned char *limit, const unsigned char *thresh)
+#define prototype_loopfilter_nc(sym)                                   \
+  void sym(unsigned char *src, int pitch, const unsigned char *blimit, \
+           const unsigned char *limit, const unsigned char *thresh)
 
 #define prototype_simple_loopfilter(sym) \
-    void sym(unsigned char *y, int ystride, const unsigned char *blimit)
+  void sym(unsigned char *y, int ystride, const unsigned char *blimit)
 
 prototype_loopfilter(vp8_mbloop_filter_vertical_edge_mmx);
 prototype_loopfilter(vp8_mbloop_filter_horizontal_edge_mmx);
@@ -47,152 +46,178 @@
 
 #if HAVE_MMX
 /* Horizontal MB filtering */
-void vp8_loop_filter_mbh_mmx(unsigned char *y_ptr, unsigned char *u_ptr, unsigned char *v_ptr,
-                             int y_stride, int uv_stride, loop_filter_info *lfi)
-{
-    vp8_mbloop_filter_horizontal_edge_mmx(y_ptr, y_stride, lfi->mblim, lfi->lim, lfi->hev_thr, 2);
+void vp8_loop_filter_mbh_mmx(unsigned char *y_ptr, unsigned char *u_ptr,
+                             unsigned char *v_ptr, int y_stride, int uv_stride,
+                             loop_filter_info *lfi) {
+  vp8_mbloop_filter_horizontal_edge_mmx(y_ptr, y_stride, lfi->mblim, lfi->lim,
+                                        lfi->hev_thr, 2);
 
-    if (u_ptr)
-        vp8_mbloop_filter_horizontal_edge_mmx(u_ptr, uv_stride, lfi->mblim, lfi->lim, lfi->hev_thr, 1);
+  if (u_ptr)
+    vp8_mbloop_filter_horizontal_edge_mmx(u_ptr, uv_stride, lfi->mblim,
+                                          lfi->lim, lfi->hev_thr, 1);
 
-    if (v_ptr)
-        vp8_mbloop_filter_horizontal_edge_mmx(v_ptr, uv_stride, lfi->mblim, lfi->lim, lfi->hev_thr, 1);
+  if (v_ptr)
+    vp8_mbloop_filter_horizontal_edge_mmx(v_ptr, uv_stride, lfi->mblim,
+                                          lfi->lim, lfi->hev_thr, 1);
 }
 
-
 /* Vertical MB Filtering */
-void vp8_loop_filter_mbv_mmx(unsigned char *y_ptr, unsigned char *u_ptr, unsigned char *v_ptr,
-                             int y_stride, int uv_stride, loop_filter_info *lfi)
-{
-    vp8_mbloop_filter_vertical_edge_mmx(y_ptr, y_stride, lfi->mblim, lfi->lim, lfi->hev_thr, 2);
+void vp8_loop_filter_mbv_mmx(unsigned char *y_ptr, unsigned char *u_ptr,
+                             unsigned char *v_ptr, int y_stride, int uv_stride,
+                             loop_filter_info *lfi) {
+  vp8_mbloop_filter_vertical_edge_mmx(y_ptr, y_stride, lfi->mblim, lfi->lim,
+                                      lfi->hev_thr, 2);
 
-    if (u_ptr)
-        vp8_mbloop_filter_vertical_edge_mmx(u_ptr, uv_stride, lfi->mblim, lfi->lim, lfi->hev_thr, 1);
+  if (u_ptr)
+    vp8_mbloop_filter_vertical_edge_mmx(u_ptr, uv_stride, lfi->mblim, lfi->lim,
+                                        lfi->hev_thr, 1);
 
-    if (v_ptr)
-        vp8_mbloop_filter_vertical_edge_mmx(v_ptr, uv_stride, lfi->mblim, lfi->lim, lfi->hev_thr, 1);
+  if (v_ptr)
+    vp8_mbloop_filter_vertical_edge_mmx(v_ptr, uv_stride, lfi->mblim, lfi->lim,
+                                        lfi->hev_thr, 1);
 }
 
-
 /* Horizontal B Filtering */
-void vp8_loop_filter_bh_mmx(unsigned char *y_ptr, unsigned char *u_ptr, unsigned char *v_ptr,
-                            int y_stride, int uv_stride, loop_filter_info *lfi)
-{
-    vp8_loop_filter_horizontal_edge_mmx(y_ptr + 4 * y_stride, y_stride, lfi->blim, lfi->lim, lfi->hev_thr, 2);
-    vp8_loop_filter_horizontal_edge_mmx(y_ptr + 8 * y_stride, y_stride, lfi->blim, lfi->lim, lfi->hev_thr, 2);
-    vp8_loop_filter_horizontal_edge_mmx(y_ptr + 12 * y_stride, y_stride, lfi->blim, lfi->lim, lfi->hev_thr, 2);
+void vp8_loop_filter_bh_mmx(unsigned char *y_ptr, unsigned char *u_ptr,
+                            unsigned char *v_ptr, int y_stride, int uv_stride,
+                            loop_filter_info *lfi) {
+  vp8_loop_filter_horizontal_edge_mmx(y_ptr + 4 * y_stride, y_stride, lfi->blim,
+                                      lfi->lim, lfi->hev_thr, 2);
+  vp8_loop_filter_horizontal_edge_mmx(y_ptr + 8 * y_stride, y_stride, lfi->blim,
+                                      lfi->lim, lfi->hev_thr, 2);
+  vp8_loop_filter_horizontal_edge_mmx(y_ptr + 12 * y_stride, y_stride,
+                                      lfi->blim, lfi->lim, lfi->hev_thr, 2);
 
-    if (u_ptr)
-        vp8_loop_filter_horizontal_edge_mmx(u_ptr + 4 * uv_stride, uv_stride, lfi->blim, lfi->lim, lfi->hev_thr, 1);
+  if (u_ptr)
+    vp8_loop_filter_horizontal_edge_mmx(u_ptr + 4 * uv_stride, uv_stride,
+                                        lfi->blim, lfi->lim, lfi->hev_thr, 1);
 
-    if (v_ptr)
-        vp8_loop_filter_horizontal_edge_mmx(v_ptr + 4 * uv_stride, uv_stride, lfi->blim, lfi->lim, lfi->hev_thr, 1);
+  if (v_ptr)
+    vp8_loop_filter_horizontal_edge_mmx(v_ptr + 4 * uv_stride, uv_stride,
+                                        lfi->blim, lfi->lim, lfi->hev_thr, 1);
 }
 
-
-void vp8_loop_filter_bhs_mmx(unsigned char *y_ptr, int y_stride, const unsigned char *blimit)
-{
-    vp8_loop_filter_simple_horizontal_edge_mmx(y_ptr + 4 * y_stride, y_stride, blimit);
-    vp8_loop_filter_simple_horizontal_edge_mmx(y_ptr + 8 * y_stride, y_stride, blimit);
-    vp8_loop_filter_simple_horizontal_edge_mmx(y_ptr + 12 * y_stride, y_stride, blimit);
+void vp8_loop_filter_bhs_mmx(unsigned char *y_ptr, int y_stride,
+                             const unsigned char *blimit) {
+  vp8_loop_filter_simple_horizontal_edge_mmx(y_ptr + 4 * y_stride, y_stride,
+                                             blimit);
+  vp8_loop_filter_simple_horizontal_edge_mmx(y_ptr + 8 * y_stride, y_stride,
+                                             blimit);
+  vp8_loop_filter_simple_horizontal_edge_mmx(y_ptr + 12 * y_stride, y_stride,
+                                             blimit);
 }
 
-
 /* Vertical B Filtering */
-void vp8_loop_filter_bv_mmx(unsigned char *y_ptr, unsigned char *u_ptr, unsigned char *v_ptr,
-                            int y_stride, int uv_stride, loop_filter_info *lfi)
-{
-    vp8_loop_filter_vertical_edge_mmx(y_ptr + 4, y_stride, lfi->blim, lfi->lim, lfi->hev_thr, 2);
-    vp8_loop_filter_vertical_edge_mmx(y_ptr + 8, y_stride, lfi->blim, lfi->lim, lfi->hev_thr, 2);
-    vp8_loop_filter_vertical_edge_mmx(y_ptr + 12, y_stride, lfi->blim, lfi->lim, lfi->hev_thr, 2);
+void vp8_loop_filter_bv_mmx(unsigned char *y_ptr, unsigned char *u_ptr,
+                            unsigned char *v_ptr, int y_stride, int uv_stride,
+                            loop_filter_info *lfi) {
+  vp8_loop_filter_vertical_edge_mmx(y_ptr + 4, y_stride, lfi->blim, lfi->lim,
+                                    lfi->hev_thr, 2);
+  vp8_loop_filter_vertical_edge_mmx(y_ptr + 8, y_stride, lfi->blim, lfi->lim,
+                                    lfi->hev_thr, 2);
+  vp8_loop_filter_vertical_edge_mmx(y_ptr + 12, y_stride, lfi->blim, lfi->lim,
+                                    lfi->hev_thr, 2);
 
-    if (u_ptr)
-        vp8_loop_filter_vertical_edge_mmx(u_ptr + 4, uv_stride, lfi->blim, lfi->lim, lfi->hev_thr, 1);
+  if (u_ptr)
+    vp8_loop_filter_vertical_edge_mmx(u_ptr + 4, uv_stride, lfi->blim, lfi->lim,
+                                      lfi->hev_thr, 1);
 
-    if (v_ptr)
-        vp8_loop_filter_vertical_edge_mmx(v_ptr + 4, uv_stride, lfi->blim, lfi->lim, lfi->hev_thr, 1);
+  if (v_ptr)
+    vp8_loop_filter_vertical_edge_mmx(v_ptr + 4, uv_stride, lfi->blim, lfi->lim,
+                                      lfi->hev_thr, 1);
 }
 
-
-void vp8_loop_filter_bvs_mmx(unsigned char *y_ptr, int y_stride, const unsigned char *blimit)
-{
-    vp8_loop_filter_simple_vertical_edge_mmx(y_ptr + 4, y_stride, blimit);
-    vp8_loop_filter_simple_vertical_edge_mmx(y_ptr + 8, y_stride, blimit);
-    vp8_loop_filter_simple_vertical_edge_mmx(y_ptr + 12, y_stride, blimit);
+void vp8_loop_filter_bvs_mmx(unsigned char *y_ptr, int y_stride,
+                             const unsigned char *blimit) {
+  vp8_loop_filter_simple_vertical_edge_mmx(y_ptr + 4, y_stride, blimit);
+  vp8_loop_filter_simple_vertical_edge_mmx(y_ptr + 8, y_stride, blimit);
+  vp8_loop_filter_simple_vertical_edge_mmx(y_ptr + 12, y_stride, blimit);
 }
 #endif
 
-
 /* Horizontal MB filtering */
 #if HAVE_SSE2
-void vp8_loop_filter_mbh_sse2(unsigned char *y_ptr, unsigned char *u_ptr, unsigned char *v_ptr,
-                              int y_stride, int uv_stride, loop_filter_info *lfi)
-{
-    vp8_mbloop_filter_horizontal_edge_sse2(y_ptr, y_stride, lfi->mblim, lfi->lim, lfi->hev_thr);
+void vp8_loop_filter_mbh_sse2(unsigned char *y_ptr, unsigned char *u_ptr,
+                              unsigned char *v_ptr, int y_stride, int uv_stride,
+                              loop_filter_info *lfi) {
+  vp8_mbloop_filter_horizontal_edge_sse2(y_ptr, y_stride, lfi->mblim, lfi->lim,
+                                         lfi->hev_thr);
 
-    if (u_ptr)
-        vp8_mbloop_filter_horizontal_edge_uv_sse2(u_ptr, uv_stride, lfi->mblim, lfi->lim, lfi->hev_thr, v_ptr);
+  if (u_ptr)
+    vp8_mbloop_filter_horizontal_edge_uv_sse2(u_ptr, uv_stride, lfi->mblim,
+                                              lfi->lim, lfi->hev_thr, v_ptr);
 }
 
-
 /* Vertical MB Filtering */
-void vp8_loop_filter_mbv_sse2(unsigned char *y_ptr, unsigned char *u_ptr, unsigned char *v_ptr,
-                              int y_stride, int uv_stride, loop_filter_info *lfi)
-{
-    vp8_mbloop_filter_vertical_edge_sse2(y_ptr, y_stride, lfi->mblim, lfi->lim, lfi->hev_thr);
+void vp8_loop_filter_mbv_sse2(unsigned char *y_ptr, unsigned char *u_ptr,
+                              unsigned char *v_ptr, int y_stride, int uv_stride,
+                              loop_filter_info *lfi) {
+  vp8_mbloop_filter_vertical_edge_sse2(y_ptr, y_stride, lfi->mblim, lfi->lim,
+                                       lfi->hev_thr);
 
-    if (u_ptr)
-        vp8_mbloop_filter_vertical_edge_uv_sse2(u_ptr, uv_stride, lfi->mblim, lfi->lim, lfi->hev_thr, v_ptr);
+  if (u_ptr)
+    vp8_mbloop_filter_vertical_edge_uv_sse2(u_ptr, uv_stride, lfi->mblim,
+                                            lfi->lim, lfi->hev_thr, v_ptr);
 }
 
-
 /* Horizontal B Filtering */
-void vp8_loop_filter_bh_sse2(unsigned char *y_ptr, unsigned char *u_ptr, unsigned char *v_ptr,
-                             int y_stride, int uv_stride, loop_filter_info *lfi)
-{
+void vp8_loop_filter_bh_sse2(unsigned char *y_ptr, unsigned char *u_ptr,
+                             unsigned char *v_ptr, int y_stride, int uv_stride,
+                             loop_filter_info *lfi) {
 #if ARCH_X86_64
-    vp8_loop_filter_bh_y_sse2(y_ptr, y_stride, lfi->blim, lfi->lim, lfi->hev_thr, 2);
+  vp8_loop_filter_bh_y_sse2(y_ptr, y_stride, lfi->blim, lfi->lim, lfi->hev_thr,
+                            2);
 #else
-    vp8_loop_filter_horizontal_edge_sse2(y_ptr + 4 * y_stride, y_stride, lfi->blim, lfi->lim, lfi->hev_thr);
-    vp8_loop_filter_horizontal_edge_sse2(y_ptr + 8 * y_stride, y_stride, lfi->blim, lfi->lim, lfi->hev_thr);
-    vp8_loop_filter_horizontal_edge_sse2(y_ptr + 12 * y_stride, y_stride, lfi->blim, lfi->lim, lfi->hev_thr);
+  vp8_loop_filter_horizontal_edge_sse2(y_ptr + 4 * y_stride, y_stride,
+                                       lfi->blim, lfi->lim, lfi->hev_thr);
+  vp8_loop_filter_horizontal_edge_sse2(y_ptr + 8 * y_stride, y_stride,
+                                       lfi->blim, lfi->lim, lfi->hev_thr);
+  vp8_loop_filter_horizontal_edge_sse2(y_ptr + 12 * y_stride, y_stride,
+                                       lfi->blim, lfi->lim, lfi->hev_thr);
 #endif
 
-    if (u_ptr)
-        vp8_loop_filter_horizontal_edge_uv_sse2(u_ptr + 4 * uv_stride, uv_stride, lfi->blim, lfi->lim, lfi->hev_thr, v_ptr + 4 * uv_stride);
+  if (u_ptr)
+    vp8_loop_filter_horizontal_edge_uv_sse2(u_ptr + 4 * uv_stride, uv_stride,
+                                            lfi->blim, lfi->lim, lfi->hev_thr,
+                                            v_ptr + 4 * uv_stride);
 }
 
-
-void vp8_loop_filter_bhs_sse2(unsigned char *y_ptr, int y_stride, const unsigned char *blimit)
-{
-    vp8_loop_filter_simple_horizontal_edge_sse2(y_ptr + 4 * y_stride, y_stride, blimit);
-    vp8_loop_filter_simple_horizontal_edge_sse2(y_ptr + 8 * y_stride, y_stride, blimit);
-    vp8_loop_filter_simple_horizontal_edge_sse2(y_ptr + 12 * y_stride, y_stride, blimit);
+void vp8_loop_filter_bhs_sse2(unsigned char *y_ptr, int y_stride,
+                              const unsigned char *blimit) {
+  vp8_loop_filter_simple_horizontal_edge_sse2(y_ptr + 4 * y_stride, y_stride,
+                                              blimit);
+  vp8_loop_filter_simple_horizontal_edge_sse2(y_ptr + 8 * y_stride, y_stride,
+                                              blimit);
+  vp8_loop_filter_simple_horizontal_edge_sse2(y_ptr + 12 * y_stride, y_stride,
+                                              blimit);
 }
 
-
 /* Vertical B Filtering */
-void vp8_loop_filter_bv_sse2(unsigned char *y_ptr, unsigned char *u_ptr, unsigned char *v_ptr,
-                             int y_stride, int uv_stride, loop_filter_info *lfi)
-{
+void vp8_loop_filter_bv_sse2(unsigned char *y_ptr, unsigned char *u_ptr,
+                             unsigned char *v_ptr, int y_stride, int uv_stride,
+                             loop_filter_info *lfi) {
 #if ARCH_X86_64
-    vp8_loop_filter_bv_y_sse2(y_ptr, y_stride, lfi->blim, lfi->lim, lfi->hev_thr, 2);
+  vp8_loop_filter_bv_y_sse2(y_ptr, y_stride, lfi->blim, lfi->lim, lfi->hev_thr,
+                            2);
 #else
-    vp8_loop_filter_vertical_edge_sse2(y_ptr + 4, y_stride, lfi->blim, lfi->lim, lfi->hev_thr);
-    vp8_loop_filter_vertical_edge_sse2(y_ptr + 8, y_stride, lfi->blim, lfi->lim, lfi->hev_thr);
-    vp8_loop_filter_vertical_edge_sse2(y_ptr + 12, y_stride, lfi->blim, lfi->lim, lfi->hev_thr);
+  vp8_loop_filter_vertical_edge_sse2(y_ptr + 4, y_stride, lfi->blim, lfi->lim,
+                                     lfi->hev_thr);
+  vp8_loop_filter_vertical_edge_sse2(y_ptr + 8, y_stride, lfi->blim, lfi->lim,
+                                     lfi->hev_thr);
+  vp8_loop_filter_vertical_edge_sse2(y_ptr + 12, y_stride, lfi->blim, lfi->lim,
+                                     lfi->hev_thr);
 #endif
 
-    if (u_ptr)
-        vp8_loop_filter_vertical_edge_uv_sse2(u_ptr + 4, uv_stride, lfi->blim, lfi->lim, lfi->hev_thr, v_ptr + 4);
+  if (u_ptr)
+    vp8_loop_filter_vertical_edge_uv_sse2(u_ptr + 4, uv_stride, lfi->blim,
+                                          lfi->lim, lfi->hev_thr, v_ptr + 4);
 }
 
-
-void vp8_loop_filter_bvs_sse2(unsigned char *y_ptr, int y_stride, const unsigned char *blimit)
-{
-    vp8_loop_filter_simple_vertical_edge_sse2(y_ptr + 4, y_stride, blimit);
-    vp8_loop_filter_simple_vertical_edge_sse2(y_ptr + 8, y_stride, blimit);
-    vp8_loop_filter_simple_vertical_edge_sse2(y_ptr + 12, y_stride, blimit);
+void vp8_loop_filter_bvs_sse2(unsigned char *y_ptr, int y_stride,
+                              const unsigned char *blimit) {
+  vp8_loop_filter_simple_vertical_edge_sse2(y_ptr + 4, y_stride, blimit);
+  vp8_loop_filter_simple_vertical_edge_sse2(y_ptr + 8, y_stride, blimit);
+  vp8_loop_filter_simple_vertical_edge_sse2(y_ptr + 12, y_stride, blimit);
 }
 
 #endif
--- a/vp8/common/x86/vp8_asm_stubs.c
+++ b/vp8/common/x86/vp8_asm_stubs.c
@@ -8,353 +8,273 @@
  *  be found in the AUTHORS file in the root of the source tree.
  */
 
-
 #include "vpx_config.h"
 #include "vp8_rtcd.h"
 #include "vpx_ports/mem.h"
 #include "filter_x86.h"
 
-extern const short vp8_six_tap_mmx[8][6*8];
+extern const short vp8_six_tap_mmx[8][6 * 8];
 
-extern void vp8_filter_block1d_h6_mmx
-(
-    unsigned char   *src_ptr,
-    unsigned short  *output_ptr,
-    unsigned int    src_pixels_per_line,
-    unsigned int    pixel_step,
-    unsigned int    output_height,
-    unsigned int    output_width,
-    const short      *vp8_filter
-);
-extern void vp8_filter_block1dc_v6_mmx
-(
-    unsigned short *src_ptr,
-    unsigned char  *output_ptr,
-    int             output_pitch,
-    unsigned int    pixels_per_line,
-    unsigned int    pixel_step,
-    unsigned int    output_height,
-    unsigned int    output_width,
-    const short    *vp8_filter
-);
-extern void vp8_filter_block1d8_h6_sse2
-(
-    unsigned char  *src_ptr,
-    unsigned short *output_ptr,
-    unsigned int    src_pixels_per_line,
-    unsigned int    pixel_step,
-    unsigned int    output_height,
-    unsigned int    output_width,
-    const short    *vp8_filter
-);
-extern void vp8_filter_block1d16_h6_sse2
-(
-    unsigned char  *src_ptr,
-    unsigned short *output_ptr,
-    unsigned int    src_pixels_per_line,
-    unsigned int    pixel_step,
-    unsigned int    output_height,
-    unsigned int    output_width,
-    const short    *vp8_filter
-);
-extern void vp8_filter_block1d8_v6_sse2
-(
-    unsigned short *src_ptr,
-    unsigned char *output_ptr,
-    int dst_ptich,
-    unsigned int pixels_per_line,
-    unsigned int pixel_step,
-    unsigned int output_height,
-    unsigned int output_width,
-    const short    *vp8_filter
-);
-extern void vp8_filter_block1d16_v6_sse2
-(
-    unsigned short *src_ptr,
-    unsigned char *output_ptr,
-    int dst_ptich,
-    unsigned int pixels_per_line,
-    unsigned int pixel_step,
-    unsigned int output_height,
-    unsigned int output_width,
-    const short    *vp8_filter
-);
-extern void vp8_unpack_block1d16_h6_sse2
-(
-    unsigned char  *src_ptr,
-    unsigned short *output_ptr,
-    unsigned int    src_pixels_per_line,
-    unsigned int    output_height,
-    unsigned int    output_width
-);
-extern void vp8_filter_block1d8_h6_only_sse2
-(
-    unsigned char  *src_ptr,
-    unsigned int    src_pixels_per_line,
-    unsigned char  *output_ptr,
-    int dst_ptich,
-    unsigned int    output_height,
-    const short    *vp8_filter
-);
-extern void vp8_filter_block1d16_h6_only_sse2
-(
-    unsigned char  *src_ptr,
-    unsigned int    src_pixels_per_line,
-    unsigned char  *output_ptr,
-    int dst_ptich,
-    unsigned int    output_height,
-    const short    *vp8_filter
-);
-extern void vp8_filter_block1d8_v6_only_sse2
-(
-    unsigned char *src_ptr,
-    unsigned int   src_pixels_per_line,
-    unsigned char *output_ptr,
-    int dst_ptich,
-    unsigned int   output_height,
-    const short   *vp8_filter
-);
+extern void vp8_filter_block1d_h6_mmx(unsigned char *src_ptr,
+                                      unsigned short *output_ptr,
+                                      unsigned int src_pixels_per_line,
+                                      unsigned int pixel_step,
+                                      unsigned int output_height,
+                                      unsigned int output_width,
+                                      const short *vp8_filter);
+extern void vp8_filter_block1dc_v6_mmx(
+    unsigned short *src_ptr, unsigned char *output_ptr, int output_pitch,
+    unsigned int pixels_per_line, unsigned int pixel_step,
+    unsigned int output_height, unsigned int output_width,
+    const short *vp8_filter);
+extern void vp8_filter_block1d8_h6_sse2(unsigned char *src_ptr,
+                                        unsigned short *output_ptr,
+                                        unsigned int src_pixels_per_line,
+                                        unsigned int pixel_step,
+                                        unsigned int output_height,
+                                        unsigned int output_width,
+                                        const short *vp8_filter);
+extern void vp8_filter_block1d16_h6_sse2(unsigned char *src_ptr,
+                                         unsigned short *output_ptr,
+                                         unsigned int src_pixels_per_line,
+                                         unsigned int pixel_step,
+                                         unsigned int output_height,
+                                         unsigned int output_width,
+                                         const short *vp8_filter);
+extern void vp8_filter_block1d8_v6_sse2(
+    unsigned short *src_ptr, unsigned char *output_ptr, int dst_ptich,
+    unsigned int pixels_per_line, unsigned int pixel_step,
+    unsigned int output_height, unsigned int output_width,
+    const short *vp8_filter);
+extern void vp8_filter_block1d16_v6_sse2(
+    unsigned short *src_ptr, unsigned char *output_ptr, int dst_ptich,
+    unsigned int pixels_per_line, unsigned int pixel_step,
+    unsigned int output_height, unsigned int output_width,
+    const short *vp8_filter);
+extern void vp8_unpack_block1d16_h6_sse2(unsigned char *src_ptr,
+                                         unsigned short *output_ptr,
+                                         unsigned int src_pixels_per_line,
+                                         unsigned int output_height,
+                                         unsigned int output_width);
+extern void vp8_filter_block1d8_h6_only_sse2(unsigned char *src_ptr,
+                                             unsigned int src_pixels_per_line,
+                                             unsigned char *output_ptr,
+                                             int dst_ptich,
+                                             unsigned int output_height,
+                                             const short *vp8_filter);
+extern void vp8_filter_block1d16_h6_only_sse2(unsigned char *src_ptr,
+                                              unsigned int src_pixels_per_line,
+                                              unsigned char *output_ptr,
+                                              int dst_ptich,
+                                              unsigned int output_height,
+                                              const short *vp8_filter);
+extern void vp8_filter_block1d8_v6_only_sse2(unsigned char *src_ptr,
+                                             unsigned int src_pixels_per_line,
+                                             unsigned char *output_ptr,
+                                             int dst_ptich,
+                                             unsigned int output_height,
+                                             const short *vp8_filter);
 
-
 #if HAVE_MMX
-void vp8_sixtap_predict4x4_mmx
-(
-    unsigned char  *src_ptr,
-    int   src_pixels_per_line,
-    int  xoffset,
-    int  yoffset,
-    unsigned char *dst_ptr,
-    int dst_pitch
-)
-{
-    DECLARE_ALIGNED(16, unsigned short, FData2[16*16]);  /* Temp data bufffer used in filtering */
-    const short *HFilter, *VFilter;
-    HFilter = vp8_six_tap_mmx[xoffset];
-    vp8_filter_block1d_h6_mmx(src_ptr - (2 * src_pixels_per_line), FData2, src_pixels_per_line, 1, 9, 8, HFilter);
-    VFilter = vp8_six_tap_mmx[yoffset];
-    vp8_filter_block1dc_v6_mmx(FData2 + 8, dst_ptr, dst_pitch, 8, 4 , 4, 4, VFilter);
-
+void vp8_sixtap_predict4x4_mmx(unsigned char *src_ptr, int src_pixels_per_line,
+                               int xoffset, int yoffset, unsigned char *dst_ptr,
+                               int dst_pitch) {
+  DECLARE_ALIGNED(16, unsigned short,
+                  FData2[16 * 16]); /* Temp data bufffer used in filtering */
+  const short *HFilter, *VFilter;
+  HFilter = vp8_six_tap_mmx[xoffset];
+  vp8_filter_block1d_h6_mmx(src_ptr - (2 * src_pixels_per_line), FData2,
+                            src_pixels_per_line, 1, 9, 8, HFilter);
+  VFilter = vp8_six_tap_mmx[yoffset];
+  vp8_filter_block1dc_v6_mmx(FData2 + 8, dst_ptr, dst_pitch, 8, 4, 4, 4,
+                             VFilter);
 }
 
+void vp8_sixtap_predict16x16_mmx(unsigned char *src_ptr,
+                                 int src_pixels_per_line, int xoffset,
+                                 int yoffset, unsigned char *dst_ptr,
+                                 int dst_pitch) {
+  DECLARE_ALIGNED(16, unsigned short,
+                  FData2[24 * 24]); /* Temp data bufffer used in filtering */
 
-void vp8_sixtap_predict16x16_mmx
-(
-    unsigned char  *src_ptr,
-    int   src_pixels_per_line,
-    int  xoffset,
-    int  yoffset,
-    unsigned char *dst_ptr,
-    int dst_pitch
-)
-{
+  const short *HFilter, *VFilter;
 
-    DECLARE_ALIGNED(16, unsigned short, FData2[24*24]);  /* Temp data bufffer used in filtering */
+  HFilter = vp8_six_tap_mmx[xoffset];
 
-    const short *HFilter, *VFilter;
+  vp8_filter_block1d_h6_mmx(src_ptr - (2 * src_pixels_per_line), FData2,
+                            src_pixels_per_line, 1, 21, 32, HFilter);
+  vp8_filter_block1d_h6_mmx(src_ptr - (2 * src_pixels_per_line) + 4, FData2 + 4,
+                            src_pixels_per_line, 1, 21, 32, HFilter);
+  vp8_filter_block1d_h6_mmx(src_ptr - (2 * src_pixels_per_line) + 8, FData2 + 8,
+                            src_pixels_per_line, 1, 21, 32, HFilter);
+  vp8_filter_block1d_h6_mmx(src_ptr - (2 * src_pixels_per_line) + 12,
+                            FData2 + 12, src_pixels_per_line, 1, 21, 32,
+                            HFilter);
 
-
-    HFilter = vp8_six_tap_mmx[xoffset];
-
-    vp8_filter_block1d_h6_mmx(src_ptr - (2 * src_pixels_per_line),    FData2,   src_pixels_per_line, 1, 21, 32, HFilter);
-    vp8_filter_block1d_h6_mmx(src_ptr - (2 * src_pixels_per_line) + 4,  FData2 + 4, src_pixels_per_line, 1, 21, 32, HFilter);
-    vp8_filter_block1d_h6_mmx(src_ptr - (2 * src_pixels_per_line) + 8,  FData2 + 8, src_pixels_per_line, 1, 21, 32, HFilter);
-    vp8_filter_block1d_h6_mmx(src_ptr - (2 * src_pixels_per_line) + 12, FData2 + 12, src_pixels_per_line, 1, 21, 32, HFilter);
-
-    VFilter = vp8_six_tap_mmx[yoffset];
-    vp8_filter_block1dc_v6_mmx(FData2 + 32, dst_ptr,   dst_pitch, 32, 16 , 16, 16, VFilter);
-    vp8_filter_block1dc_v6_mmx(FData2 + 36, dst_ptr + 4, dst_pitch, 32, 16 , 16, 16, VFilter);
-    vp8_filter_block1dc_v6_mmx(FData2 + 40, dst_ptr + 8, dst_pitch, 32, 16 , 16, 16, VFilter);
-    vp8_filter_block1dc_v6_mmx(FData2 + 44, dst_ptr + 12, dst_pitch, 32, 16 , 16, 16, VFilter);
-
+  VFilter = vp8_six_tap_mmx[yoffset];
+  vp8_filter_block1dc_v6_mmx(FData2 + 32, dst_ptr, dst_pitch, 32, 16, 16, 16,
+                             VFilter);
+  vp8_filter_block1dc_v6_mmx(FData2 + 36, dst_ptr + 4, dst_pitch, 32, 16, 16,
+                             16, VFilter);
+  vp8_filter_block1dc_v6_mmx(FData2 + 40, dst_ptr + 8, dst_pitch, 32, 16, 16,
+                             16, VFilter);
+  vp8_filter_block1dc_v6_mmx(FData2 + 44, dst_ptr + 12, dst_pitch, 32, 16, 16,
+                             16, VFilter);
 }
 
+void vp8_sixtap_predict8x8_mmx(unsigned char *src_ptr, int src_pixels_per_line,
+                               int xoffset, int yoffset, unsigned char *dst_ptr,
+                               int dst_pitch) {
+  DECLARE_ALIGNED(16, unsigned short,
+                  FData2[256]); /* Temp data bufffer used in filtering */
 
-void vp8_sixtap_predict8x8_mmx
-(
-    unsigned char  *src_ptr,
-    int   src_pixels_per_line,
-    int  xoffset,
-    int  yoffset,
-    unsigned char *dst_ptr,
-    int dst_pitch
-)
-{
+  const short *HFilter, *VFilter;
 
-    DECLARE_ALIGNED(16, unsigned short, FData2[256]);    /* Temp data bufffer used in filtering */
+  HFilter = vp8_six_tap_mmx[xoffset];
+  vp8_filter_block1d_h6_mmx(src_ptr - (2 * src_pixels_per_line), FData2,
+                            src_pixels_per_line, 1, 13, 16, HFilter);
+  vp8_filter_block1d_h6_mmx(src_ptr - (2 * src_pixels_per_line) + 4, FData2 + 4,
+                            src_pixels_per_line, 1, 13, 16, HFilter);
 
-    const short *HFilter, *VFilter;
-
-    HFilter = vp8_six_tap_mmx[xoffset];
-    vp8_filter_block1d_h6_mmx(src_ptr - (2 * src_pixels_per_line),    FData2,   src_pixels_per_line, 1, 13, 16, HFilter);
-    vp8_filter_block1d_h6_mmx(src_ptr - (2 * src_pixels_per_line) + 4,  FData2 + 4, src_pixels_per_line, 1, 13, 16, HFilter);
-
-    VFilter = vp8_six_tap_mmx[yoffset];
-    vp8_filter_block1dc_v6_mmx(FData2 + 16, dst_ptr,   dst_pitch, 16, 8 , 8, 8, VFilter);
-    vp8_filter_block1dc_v6_mmx(FData2 + 20, dst_ptr + 4, dst_pitch, 16, 8 , 8, 8, VFilter);
-
+  VFilter = vp8_six_tap_mmx[yoffset];
+  vp8_filter_block1dc_v6_mmx(FData2 + 16, dst_ptr, dst_pitch, 16, 8, 8, 8,
+                             VFilter);
+  vp8_filter_block1dc_v6_mmx(FData2 + 20, dst_ptr + 4, dst_pitch, 16, 8, 8, 8,
+                             VFilter);
 }
 
+void vp8_sixtap_predict8x4_mmx(unsigned char *src_ptr, int src_pixels_per_line,
+                               int xoffset, int yoffset, unsigned char *dst_ptr,
+                               int dst_pitch) {
+  DECLARE_ALIGNED(16, unsigned short,
+                  FData2[256]); /* Temp data bufffer used in filtering */
 
-void vp8_sixtap_predict8x4_mmx
-(
-    unsigned char  *src_ptr,
-    int   src_pixels_per_line,
-    int  xoffset,
-    int  yoffset,
-    unsigned char *dst_ptr,
-    int dst_pitch
-)
-{
+  const short *HFilter, *VFilter;
 
-    DECLARE_ALIGNED(16, unsigned short, FData2[256]);    /* Temp data bufffer used in filtering */
+  HFilter = vp8_six_tap_mmx[xoffset];
+  vp8_filter_block1d_h6_mmx(src_ptr - (2 * src_pixels_per_line), FData2,
+                            src_pixels_per_line, 1, 9, 16, HFilter);
+  vp8_filter_block1d_h6_mmx(src_ptr - (2 * src_pixels_per_line) + 4, FData2 + 4,
+                            src_pixels_per_line, 1, 9, 16, HFilter);
 
-    const short *HFilter, *VFilter;
-
-    HFilter = vp8_six_tap_mmx[xoffset];
-    vp8_filter_block1d_h6_mmx(src_ptr - (2 * src_pixels_per_line),    FData2,   src_pixels_per_line, 1, 9, 16, HFilter);
-    vp8_filter_block1d_h6_mmx(src_ptr - (2 * src_pixels_per_line) + 4,  FData2 + 4, src_pixels_per_line, 1, 9, 16, HFilter);
-
-    VFilter = vp8_six_tap_mmx[yoffset];
-    vp8_filter_block1dc_v6_mmx(FData2 + 16, dst_ptr,   dst_pitch, 16, 8 , 4, 8, VFilter);
-    vp8_filter_block1dc_v6_mmx(FData2 + 20, dst_ptr + 4, dst_pitch, 16, 8 , 4, 8, VFilter);
-
+  VFilter = vp8_six_tap_mmx[yoffset];
+  vp8_filter_block1dc_v6_mmx(FData2 + 16, dst_ptr, dst_pitch, 16, 8, 4, 8,
+                             VFilter);
+  vp8_filter_block1dc_v6_mmx(FData2 + 20, dst_ptr + 4, dst_pitch, 16, 8, 4, 8,
+                             VFilter);
 }
 
-
-
-void vp8_bilinear_predict16x16_mmx
-(
-    unsigned char  *src_ptr,
-    int   src_pixels_per_line,
-    int  xoffset,
-    int  yoffset,
-    unsigned char *dst_ptr,
-    int dst_pitch
-)
-{
-    vp8_bilinear_predict8x8_mmx(src_ptr,   src_pixels_per_line, xoffset, yoffset, dst_ptr,   dst_pitch);
-    vp8_bilinear_predict8x8_mmx(src_ptr + 8, src_pixels_per_line, xoffset, yoffset, dst_ptr + 8, dst_pitch);
-    vp8_bilinear_predict8x8_mmx(src_ptr + 8 * src_pixels_per_line,   src_pixels_per_line, xoffset, yoffset, dst_ptr + dst_pitch * 8,   dst_pitch);
-    vp8_bilinear_predict8x8_mmx(src_ptr + 8 * src_pixels_per_line + 8, src_pixels_per_line, xoffset, yoffset, dst_ptr + dst_pitch * 8 + 8, dst_pitch);
+void vp8_bilinear_predict16x16_mmx(unsigned char *src_ptr,
+                                   int src_pixels_per_line, int xoffset,
+                                   int yoffset, unsigned char *dst_ptr,
+                                   int dst_pitch) {
+  vp8_bilinear_predict8x8_mmx(src_ptr, src_pixels_per_line, xoffset, yoffset,
+                              dst_ptr, dst_pitch);
+  vp8_bilinear_predict8x8_mmx(src_ptr + 8, src_pixels_per_line, xoffset,
+                              yoffset, dst_ptr + 8, dst_pitch);
+  vp8_bilinear_predict8x8_mmx(src_ptr + 8 * src_pixels_per_line,
+                              src_pixels_per_line, xoffset, yoffset,
+                              dst_ptr + dst_pitch * 8, dst_pitch);
+  vp8_bilinear_predict8x8_mmx(src_ptr + 8 * src_pixels_per_line + 8,
+                              src_pixels_per_line, xoffset, yoffset,
+                              dst_ptr + dst_pitch * 8 + 8, dst_pitch);
 }
 #endif
 
-
 #if HAVE_SSE2
-void vp8_sixtap_predict16x16_sse2
-(
-    unsigned char  *src_ptr,
-    int   src_pixels_per_line,
-    int  xoffset,
-    int  yoffset,
-    unsigned char *dst_ptr,
-    int dst_pitch
+void vp8_sixtap_predict16x16_sse2(unsigned char *src_ptr,
+                                  int src_pixels_per_line, int xoffset,
+                                  int yoffset, unsigned char *dst_ptr,
+                                  int dst_pitch
 
-)
-{
-    DECLARE_ALIGNED(16, unsigned short, FData2[24*24]);    /* Temp data bufffer used in filtering */
+                                  ) {
+  DECLARE_ALIGNED(16, unsigned short,
+                  FData2[24 * 24]); /* Temp data bufffer used in filtering */
 
-    const short *HFilter, *VFilter;
+  const short *HFilter, *VFilter;
 
-    if (xoffset)
-    {
-        if (yoffset)
-        {
-            HFilter = vp8_six_tap_mmx[xoffset];
-            vp8_filter_block1d16_h6_sse2(src_ptr - (2 * src_pixels_per_line), FData2,   src_pixels_per_line, 1, 21, 32, HFilter);
-            VFilter = vp8_six_tap_mmx[yoffset];
-            vp8_filter_block1d16_v6_sse2(FData2 + 32, dst_ptr,   dst_pitch, 32, 16 , 16, dst_pitch, VFilter);
-        }
-        else
-        {
-            /* First-pass only */
-            HFilter = vp8_six_tap_mmx[xoffset];
-            vp8_filter_block1d16_h6_only_sse2(src_ptr, src_pixels_per_line, dst_ptr, dst_pitch, 16, HFilter);
-        }
+  if (xoffset) {
+    if (yoffset) {
+      HFilter = vp8_six_tap_mmx[xoffset];
+      vp8_filter_block1d16_h6_sse2(src_ptr - (2 * src_pixels_per_line), FData2,
+                                   src_pixels_per_line, 1, 21, 32, HFilter);
+      VFilter = vp8_six_tap_mmx[yoffset];
+      vp8_filter_block1d16_v6_sse2(FData2 + 32, dst_ptr, dst_pitch, 32, 16, 16,
+                                   dst_pitch, VFilter);
+    } else {
+      /* First-pass only */
+      HFilter = vp8_six_tap_mmx[xoffset];
+      vp8_filter_block1d16_h6_only_sse2(src_ptr, src_pixels_per_line, dst_ptr,
+                                        dst_pitch, 16, HFilter);
     }
-    else
-    {
-        /* Second-pass only */
-        VFilter = vp8_six_tap_mmx[yoffset];
-        vp8_unpack_block1d16_h6_sse2(src_ptr - (2 * src_pixels_per_line), FData2,   src_pixels_per_line, 21, 32);
-        vp8_filter_block1d16_v6_sse2(FData2 + 32, dst_ptr,   dst_pitch, 32, 16 , 16, dst_pitch, VFilter);
-    }
+  } else {
+    /* Second-pass only */
+    VFilter = vp8_six_tap_mmx[yoffset];
+    vp8_unpack_block1d16_h6_sse2(src_ptr - (2 * src_pixels_per_line), FData2,
+                                 src_pixels_per_line, 21, 32);
+    vp8_filter_block1d16_v6_sse2(FData2 + 32, dst_ptr, dst_pitch, 32, 16, 16,
+                                 dst_pitch, VFilter);
+  }
 }
 
+void vp8_sixtap_predict8x8_sse2(unsigned char *src_ptr, int src_pixels_per_line,
+                                int xoffset, int yoffset,
+                                unsigned char *dst_ptr, int dst_pitch) {
+  DECLARE_ALIGNED(16, unsigned short,
+                  FData2[256]); /* Temp data bufffer used in filtering */
+  const short *HFilter, *VFilter;
 
-void vp8_sixtap_predict8x8_sse2
-(
-    unsigned char  *src_ptr,
-    int   src_pixels_per_line,
-    int  xoffset,
-    int  yoffset,
-    unsigned char *dst_ptr,
-    int dst_pitch
-)
-{
-    DECLARE_ALIGNED(16, unsigned short, FData2[256]);  /* Temp data bufffer used in filtering */
-    const short *HFilter, *VFilter;
-
-    if (xoffset)
-    {
-        if (yoffset)
-        {
-            HFilter = vp8_six_tap_mmx[xoffset];
-            vp8_filter_block1d8_h6_sse2(src_ptr - (2 * src_pixels_per_line), FData2,   src_pixels_per_line, 1, 13, 16, HFilter);
-            VFilter = vp8_six_tap_mmx[yoffset];
-            vp8_filter_block1d8_v6_sse2(FData2 + 16, dst_ptr,   dst_pitch, 16, 8 , 8, dst_pitch, VFilter);
-        }
-        else
-        {
-            /* First-pass only */
-            HFilter = vp8_six_tap_mmx[xoffset];
-            vp8_filter_block1d8_h6_only_sse2(src_ptr, src_pixels_per_line, dst_ptr, dst_pitch, 8, HFilter);
-        }
+  if (xoffset) {
+    if (yoffset) {
+      HFilter = vp8_six_tap_mmx[xoffset];
+      vp8_filter_block1d8_h6_sse2(src_ptr - (2 * src_pixels_per_line), FData2,
+                                  src_pixels_per_line, 1, 13, 16, HFilter);
+      VFilter = vp8_six_tap_mmx[yoffset];
+      vp8_filter_block1d8_v6_sse2(FData2 + 16, dst_ptr, dst_pitch, 16, 8, 8,
+                                  dst_pitch, VFilter);
+    } else {
+      /* First-pass only */
+      HFilter = vp8_six_tap_mmx[xoffset];
+      vp8_filter_block1d8_h6_only_sse2(src_ptr, src_pixels_per_line, dst_ptr,
+                                       dst_pitch, 8, HFilter);
     }
-    else
-    {
-        /* Second-pass only */
-        VFilter = vp8_six_tap_mmx[yoffset];
-        vp8_filter_block1d8_v6_only_sse2(src_ptr - (2 * src_pixels_per_line), src_pixels_per_line, dst_ptr, dst_pitch, 8, VFilter);
-    }
+  } else {
+    /* Second-pass only */
+    VFilter = vp8_six_tap_mmx[yoffset];
+    vp8_filter_block1d8_v6_only_sse2(src_ptr - (2 * src_pixels_per_line),
+                                     src_pixels_per_line, dst_ptr, dst_pitch, 8,
+                                     VFilter);
+  }
 }
 
+void vp8_sixtap_predict8x4_sse2(unsigned char *src_ptr, int src_pixels_per_line,
+                                int xoffset, int yoffset,
+                                unsigned char *dst_ptr, int dst_pitch) {
+  DECLARE_ALIGNED(16, unsigned short,
+                  FData2[256]); /* Temp data bufffer used in filtering */
+  const short *HFilter, *VFilter;
 
-void vp8_sixtap_predict8x4_sse2
-(
-    unsigned char  *src_ptr,
-    int   src_pixels_per_line,
-    int  xoffset,
-    int  yoffset,
-    unsigned char *dst_ptr,
-    int dst_pitch
-)
-{
-    DECLARE_ALIGNED(16, unsigned short, FData2[256]);  /* Temp data bufffer used in filtering */
-    const short *HFilter, *VFilter;
-
-    if (xoffset)
-    {
-        if (yoffset)
-        {
-            HFilter = vp8_six_tap_mmx[xoffset];
-            vp8_filter_block1d8_h6_sse2(src_ptr - (2 * src_pixels_per_line), FData2,   src_pixels_per_line, 1, 9, 16, HFilter);
-            VFilter = vp8_six_tap_mmx[yoffset];
-            vp8_filter_block1d8_v6_sse2(FData2 + 16, dst_ptr,   dst_pitch, 16, 8 , 4, dst_pitch, VFilter);
-        }
-        else
-        {
-            /* First-pass only */
-            HFilter = vp8_six_tap_mmx[xoffset];
-            vp8_filter_block1d8_h6_only_sse2(src_ptr, src_pixels_per_line, dst_ptr, dst_pitch, 4, HFilter);
-        }
+  if (xoffset) {
+    if (yoffset) {
+      HFilter = vp8_six_tap_mmx[xoffset];
+      vp8_filter_block1d8_h6_sse2(src_ptr - (2 * src_pixels_per_line), FData2,
+                                  src_pixels_per_line, 1, 9, 16, HFilter);
+      VFilter = vp8_six_tap_mmx[yoffset];
+      vp8_filter_block1d8_v6_sse2(FData2 + 16, dst_ptr, dst_pitch, 16, 8, 4,
+                                  dst_pitch, VFilter);
+    } else {
+      /* First-pass only */
+      HFilter = vp8_six_tap_mmx[xoffset];
+      vp8_filter_block1d8_h6_only_sse2(src_ptr, src_pixels_per_line, dst_ptr,
+                                       dst_pitch, 4, HFilter);
     }
-    else
-    {
-        /* Second-pass only */
-        VFilter = vp8_six_tap_mmx[yoffset];
-        vp8_filter_block1d8_v6_only_sse2(src_ptr - (2 * src_pixels_per_line), src_pixels_per_line, dst_ptr, dst_pitch, 4, VFilter);
-    }
+  } else {
+    /* Second-pass only */
+    VFilter = vp8_six_tap_mmx[yoffset];
+    vp8_filter_block1d8_v6_only_sse2(src_ptr - (2 * src_pixels_per_line),
+                                     src_pixels_per_line, dst_ptr, dst_pitch, 4,
+                                     VFilter);
+  }
 }
 
 #endif
@@ -361,264 +281,179 @@
 
 #if HAVE_SSSE3
 
-extern void vp8_filter_block1d8_h6_ssse3
-(
-    unsigned char  *src_ptr,
-    unsigned int    src_pixels_per_line,
-    unsigned char  *output_ptr,
-    unsigned int    output_pitch,
-    unsigned int    output_height,
-    unsigned int    vp8_filter_index
-);
+extern void vp8_filter_block1d8_h6_ssse3(unsigned char *src_ptr,
+                                         unsigned int src_pixels_per_line,
+                                         unsigned char *output_ptr,
+                                         unsigned int output_pitch,
+                                         unsigned int output_height,
+                                         unsigned int vp8_filter_index);
 
-extern void vp8_filter_block1d16_h6_ssse3
-(
-    unsigned char  *src_ptr,
-    unsigned int    src_pixels_per_line,
-    unsigned char  *output_ptr,
-    unsigned int    output_pitch,
-    unsigned int    output_height,
-    unsigned int    vp8_filter_index
-);
+extern void vp8_filter_block1d16_h6_ssse3(unsigned char *src_ptr,
+                                          unsigned int src_pixels_per_line,
+                                          unsigned char *output_ptr,
+                                          unsigned int output_pitch,
+                                          unsigned int output_height,
+                                          unsigned int vp8_filter_index);
 
-extern void vp8_filter_block1d16_v6_ssse3
-(
-    unsigned char *src_ptr,
-    unsigned int   src_pitch,
-    unsigned char *output_ptr,
-    unsigned int   out_pitch,
-    unsigned int   output_height,
-    unsigned int   vp8_filter_index
-);
+extern void vp8_filter_block1d16_v6_ssse3(unsigned char *src_ptr,
+                                          unsigned int src_pitch,
+                                          unsigned char *output_ptr,
+                                          unsigned int out_pitch,
+                                          unsigned int output_height,
+                                          unsigned int vp8_filter_index);
 
-extern void vp8_filter_block1d8_v6_ssse3
-(
-    unsigned char *src_ptr,
-    unsigned int   src_pitch,
-    unsigned char *output_ptr,
-    unsigned int   out_pitch,
-    unsigned int   output_height,
-    unsigned int   vp8_filter_index
-);
+extern void vp8_filter_block1d8_v6_ssse3(unsigned char *src_ptr,
+                                         unsigned int src_pitch,
+                                         unsigned char *output_ptr,
+                                         unsigned int out_pitch,
+                                         unsigned int output_height,
+                                         unsigned int vp8_filter_index);
 
-extern void vp8_filter_block1d4_h6_ssse3
-(
-    unsigned char  *src_ptr,
-    unsigned int    src_pixels_per_line,
-    unsigned char  *output_ptr,
-    unsigned int    output_pitch,
-    unsigned int    output_height,
-    unsigned int    vp8_filter_index
-);
+extern void vp8_filter_block1d4_h6_ssse3(unsigned char *src_ptr,
+                                         unsigned int src_pixels_per_line,
+                                         unsigned char *output_ptr,
+                                         unsigned int output_pitch,
+                                         unsigned int output_height,
+                                         unsigned int vp8_filter_index);
 
-extern void vp8_filter_block1d4_v6_ssse3
-(
-    unsigned char *src_ptr,
-    unsigned int   src_pitch,
-    unsigned char *output_ptr,
-    unsigned int   out_pitch,
-    unsigned int   output_height,
-    unsigned int   vp8_filter_index
-);
+extern void vp8_filter_block1d4_v6_ssse3(unsigned char *src_ptr,
+                                         unsigned int src_pitch,
+                                         unsigned char *output_ptr,
+                                         unsigned int out_pitch,
+                                         unsigned int output_height,
+                                         unsigned int vp8_filter_index);
 
-void vp8_sixtap_predict16x16_ssse3
-(
-    unsigned char  *src_ptr,
-    int   src_pixels_per_line,
-    int  xoffset,
-    int  yoffset,
-    unsigned char *dst_ptr,
-    int dst_pitch
+void vp8_sixtap_predict16x16_ssse3(unsigned char *src_ptr,
+                                   int src_pixels_per_line, int xoffset,
+                                   int yoffset, unsigned char *dst_ptr,
+                                   int dst_pitch
 
-)
-{
-    DECLARE_ALIGNED(16, unsigned char, FData2[24*24]);
+                                   ) {
+  DECLARE_ALIGNED(16, unsigned char, FData2[24 * 24]);
 
-    if (xoffset)
-    {
-        if (yoffset)
-        {
-            vp8_filter_block1d16_h6_ssse3(src_ptr - (2 * src_pixels_per_line),
-                                          src_pixels_per_line, FData2,
-                                          16, 21, xoffset);
-            vp8_filter_block1d16_v6_ssse3(FData2 , 16, dst_ptr, dst_pitch,
-                                          16, yoffset);
-        }
-        else
-        {
-            /* First-pass only */
-            vp8_filter_block1d16_h6_ssse3(src_ptr, src_pixels_per_line,
-                                          dst_ptr, dst_pitch, 16, xoffset);
-        }
+  if (xoffset) {
+    if (yoffset) {
+      vp8_filter_block1d16_h6_ssse3(src_ptr - (2 * src_pixels_per_line),
+                                    src_pixels_per_line, FData2, 16, 21,
+                                    xoffset);
+      vp8_filter_block1d16_v6_ssse3(FData2, 16, dst_ptr, dst_pitch, 16,
+                                    yoffset);
+    } else {
+      /* First-pass only */
+      vp8_filter_block1d16_h6_ssse3(src_ptr, src_pixels_per_line, dst_ptr,
+                                    dst_pitch, 16, xoffset);
     }
-    else
-    {
-        if (yoffset)
-        {
-            /* Second-pass only */
-            vp8_filter_block1d16_v6_ssse3(src_ptr - (2 * src_pixels_per_line),
-                                          src_pixels_per_line,
-                                          dst_ptr, dst_pitch, 16, yoffset);
-        }
-        else
-        {
-            /* ssse3 second-pass only function couldn't handle (xoffset==0 &&
-             * yoffset==0) case correctly. Add copy function here to guarantee
-             * six-tap function handles all possible offsets. */
-            vp8_copy_mem16x16(src_ptr, src_pixels_per_line, dst_ptr, dst_pitch);
-        }
+  } else {
+    if (yoffset) {
+      /* Second-pass only */
+      vp8_filter_block1d16_v6_ssse3(src_ptr - (2 * src_pixels_per_line),
+                                    src_pixels_per_line, dst_ptr, dst_pitch, 16,
+                                    yoffset);
+    } else {
+      /* ssse3 second-pass only function couldn't handle (xoffset==0 &&
+       * yoffset==0) case correctly. Add copy function here to guarantee
+       * six-tap function handles all possible offsets. */
+      vp8_copy_mem16x16(src_ptr, src_pixels_per_line, dst_ptr, dst_pitch);
     }
+  }
 }
 
-void vp8_sixtap_predict8x8_ssse3
-(
-    unsigned char  *src_ptr,
-    int   src_pixels_per_line,
-    int  xoffset,
-    int  yoffset,
-    unsigned char *dst_ptr,
-    int dst_pitch
-)
-{
-    DECLARE_ALIGNED(16, unsigned char, FData2[256]);
+void vp8_sixtap_predict8x8_ssse3(unsigned char *src_ptr,
+                                 int src_pixels_per_line, int xoffset,
+                                 int yoffset, unsigned char *dst_ptr,
+                                 int dst_pitch) {
+  DECLARE_ALIGNED(16, unsigned char, FData2[256]);
 
-    if (xoffset)
-    {
-        if (yoffset)
-        {
-            vp8_filter_block1d8_h6_ssse3(src_ptr - (2 * src_pixels_per_line),
-                                         src_pixels_per_line, FData2,
-                                         8, 13, xoffset);
-            vp8_filter_block1d8_v6_ssse3(FData2, 8, dst_ptr, dst_pitch,
-                                         8, yoffset);
-        }
-        else
-        {
-            vp8_filter_block1d8_h6_ssse3(src_ptr, src_pixels_per_line,
-                                         dst_ptr, dst_pitch, 8, xoffset);
-        }
+  if (xoffset) {
+    if (yoffset) {
+      vp8_filter_block1d8_h6_ssse3(src_ptr - (2 * src_pixels_per_line),
+                                   src_pixels_per_line, FData2, 8, 13, xoffset);
+      vp8_filter_block1d8_v6_ssse3(FData2, 8, dst_ptr, dst_pitch, 8, yoffset);
+    } else {
+      vp8_filter_block1d8_h6_ssse3(src_ptr, src_pixels_per_line, dst_ptr,
+                                   dst_pitch, 8, xoffset);
     }
-    else
-    {
-        if (yoffset)
-        {
-            /* Second-pass only */
-            vp8_filter_block1d8_v6_ssse3(src_ptr - (2 * src_pixels_per_line),
-                                         src_pixels_per_line,
-                                         dst_ptr, dst_pitch, 8, yoffset);
-        }
-        else
-        {
-            /* ssse3 second-pass only function couldn't handle (xoffset==0 &&
-             * yoffset==0) case correctly. Add copy function here to guarantee
-             * six-tap function handles all possible offsets. */
-            vp8_copy_mem8x8(src_ptr, src_pixels_per_line, dst_ptr, dst_pitch);
-        }
+  } else {
+    if (yoffset) {
+      /* Second-pass only */
+      vp8_filter_block1d8_v6_ssse3(src_ptr - (2 * src_pixels_per_line),
+                                   src_pixels_per_line, dst_ptr, dst_pitch, 8,
+                                   yoffset);
+    } else {
+      /* ssse3 second-pass only function couldn't handle (xoffset==0 &&
+       * yoffset==0) case correctly. Add copy function here to guarantee
+       * six-tap function handles all possible offsets. */
+      vp8_copy_mem8x8(src_ptr, src_pixels_per_line, dst_ptr, dst_pitch);
     }
+  }
 }
 
+void vp8_sixtap_predict8x4_ssse3(unsigned char *src_ptr,
+                                 int src_pixels_per_line, int xoffset,
+                                 int yoffset, unsigned char *dst_ptr,
+                                 int dst_pitch) {
+  DECLARE_ALIGNED(16, unsigned char, FData2[256]);
 
-void vp8_sixtap_predict8x4_ssse3
-(
-    unsigned char  *src_ptr,
-    int   src_pixels_per_line,
-    int  xoffset,
-    int  yoffset,
-    unsigned char *dst_ptr,
-    int dst_pitch
-)
-{
-    DECLARE_ALIGNED(16, unsigned char, FData2[256]);
-
-    if (xoffset)
-    {
-        if (yoffset)
-        {
-            vp8_filter_block1d8_h6_ssse3(src_ptr - (2 * src_pixels_per_line),
-                                         src_pixels_per_line, FData2,
-                                         8, 9, xoffset);
-            vp8_filter_block1d8_v6_ssse3(FData2, 8, dst_ptr, dst_pitch,
-                                         4, yoffset);
-        }
-        else
-        {
-            /* First-pass only */
-            vp8_filter_block1d8_h6_ssse3(src_ptr, src_pixels_per_line,
-                                         dst_ptr, dst_pitch, 4, xoffset);
-        }
+  if (xoffset) {
+    if (yoffset) {
+      vp8_filter_block1d8_h6_ssse3(src_ptr - (2 * src_pixels_per_line),
+                                   src_pixels_per_line, FData2, 8, 9, xoffset);
+      vp8_filter_block1d8_v6_ssse3(FData2, 8, dst_ptr, dst_pitch, 4, yoffset);
+    } else {
+      /* First-pass only */
+      vp8_filter_block1d8_h6_ssse3(src_ptr, src_pixels_per_line, dst_ptr,
+                                   dst_pitch, 4, xoffset);
     }
-    else
-    {
-        if (yoffset)
-        {
-            /* Second-pass only */
-            vp8_filter_block1d8_v6_ssse3(src_ptr - (2 * src_pixels_per_line),
-                                         src_pixels_per_line,
-                                         dst_ptr, dst_pitch, 4, yoffset);
-        }
-        else
-        {
-            /* ssse3 second-pass only function couldn't handle (xoffset==0 &&
-             * yoffset==0) case correctly. Add copy function here to guarantee
-             * six-tap function handles all possible offsets. */
-            vp8_copy_mem8x4(src_ptr, src_pixels_per_line, dst_ptr, dst_pitch);
-        }
+  } else {
+    if (yoffset) {
+      /* Second-pass only */
+      vp8_filter_block1d8_v6_ssse3(src_ptr - (2 * src_pixels_per_line),
+                                   src_pixels_per_line, dst_ptr, dst_pitch, 4,
+                                   yoffset);
+    } else {
+      /* ssse3 second-pass only function couldn't handle (xoffset==0 &&
+       * yoffset==0) case correctly. Add copy function here to guarantee
+       * six-tap function handles all possible offsets. */
+      vp8_copy_mem8x4(src_ptr, src_pixels_per_line, dst_ptr, dst_pitch);
     }
+  }
 }
 
-void vp8_sixtap_predict4x4_ssse3
-(
-    unsigned char  *src_ptr,
-    int   src_pixels_per_line,
-    int  xoffset,
-    int  yoffset,
-    unsigned char *dst_ptr,
-    int dst_pitch
-)
-{
-  DECLARE_ALIGNED(16, unsigned char, FData2[4*9]);
+void vp8_sixtap_predict4x4_ssse3(unsigned char *src_ptr,
+                                 int src_pixels_per_line, int xoffset,
+                                 int yoffset, unsigned char *dst_ptr,
+                                 int dst_pitch) {
+  DECLARE_ALIGNED(16, unsigned char, FData2[4 * 9]);
 
-  if (xoffset)
-  {
-      if (yoffset)
-      {
-          vp8_filter_block1d4_h6_ssse3(src_ptr - (2 * src_pixels_per_line),
-                                       src_pixels_per_line,
-                                       FData2, 4, 9, xoffset);
-          vp8_filter_block1d4_v6_ssse3(FData2, 4, dst_ptr, dst_pitch,
-                                       4, yoffset);
-      }
-      else
-      {
-          vp8_filter_block1d4_h6_ssse3(src_ptr, src_pixels_per_line,
-                                       dst_ptr, dst_pitch, 4, xoffset);
-      }
-  }
-  else
-  {
-      if (yoffset)
-      {
-          vp8_filter_block1d4_v6_ssse3(src_ptr - (2 * src_pixels_per_line),
-                                       src_pixels_per_line,
-                                       dst_ptr, dst_pitch, 4, yoffset);
-      }
-      else
-      {
-        /* ssse3 second-pass only function couldn't handle (xoffset==0 &&
-          * yoffset==0) case correctly. Add copy function here to guarantee
-          * six-tap function handles all possible offsets. */
-          int r;
+  if (xoffset) {
+    if (yoffset) {
+      vp8_filter_block1d4_h6_ssse3(src_ptr - (2 * src_pixels_per_line),
+                                   src_pixels_per_line, FData2, 4, 9, xoffset);
+      vp8_filter_block1d4_v6_ssse3(FData2, 4, dst_ptr, dst_pitch, 4, yoffset);
+    } else {
+      vp8_filter_block1d4_h6_ssse3(src_ptr, src_pixels_per_line, dst_ptr,
+                                   dst_pitch, 4, xoffset);
+    }
+  } else {
+    if (yoffset) {
+      vp8_filter_block1d4_v6_ssse3(src_ptr - (2 * src_pixels_per_line),
+                                   src_pixels_per_line, dst_ptr, dst_pitch, 4,
+                                   yoffset);
+    } else {
+      /* ssse3 second-pass only function couldn't handle (xoffset==0 &&
+        * yoffset==0) case correctly. Add copy function here to guarantee
+        * six-tap function handles all possible offsets. */
+      int r;
 
-          for (r = 0; r < 4; r++)
-          {
-            dst_ptr[0]  = src_ptr[0];
-            dst_ptr[1]  = src_ptr[1];
-            dst_ptr[2]  = src_ptr[2];
-            dst_ptr[3]  = src_ptr[3];
-            dst_ptr     += dst_pitch;
-            src_ptr     += src_pixels_per_line;
-          }
+      for (r = 0; r < 4; r++) {
+        dst_ptr[0] = src_ptr[0];
+        dst_ptr[1] = src_ptr[1];
+        dst_ptr[2] = src_ptr[2];
+        dst_ptr[3] = src_ptr[3];
+        dst_ptr += dst_pitch;
+        src_ptr += src_pixels_per_line;
       }
+    }
   }
 }
 
--- a/vp8/decoder/dboolhuff.c
+++ b/vp8/decoder/dboolhuff.c
@@ -8,70 +8,61 @@
  *  be found in the AUTHORS file in the root of the source tree.
  */
 
-
 #include "dboolhuff.h"
 #include "vp8/common/common.h"
 #include "vpx_dsp/vpx_dsp_common.h"
 
-int vp8dx_start_decode(BOOL_DECODER *br,
-                       const unsigned char *source,
-                       unsigned int source_sz,
-                       vpx_decrypt_cb decrypt_cb,
-                       void *decrypt_state)
-{
-    br->user_buffer_end = source+source_sz;
-    br->user_buffer     = source;
-    br->value    = 0;
-    br->count    = -8;
-    br->range    = 255;
-    br->decrypt_cb = decrypt_cb;
-    br->decrypt_state = decrypt_state;
+int vp8dx_start_decode(BOOL_DECODER *br, const unsigned char *source,
+                       unsigned int source_sz, vpx_decrypt_cb decrypt_cb,
+                       void *decrypt_state) {
+  br->user_buffer_end = source + source_sz;
+  br->user_buffer = source;
+  br->value = 0;
+  br->count = -8;
+  br->range = 255;
+  br->decrypt_cb = decrypt_cb;
+  br->decrypt_state = decrypt_state;
 
-    if (source_sz && !source)
-        return 1;
+  if (source_sz && !source) return 1;
 
-    /* Populate the buffer */
-    vp8dx_bool_decoder_fill(br);
+  /* Populate the buffer */
+  vp8dx_bool_decoder_fill(br);
 
-    return 0;
+  return 0;
 }
 
-void vp8dx_bool_decoder_fill(BOOL_DECODER *br)
-{
-    const unsigned char *bufptr = br->user_buffer;
-    VP8_BD_VALUE value = br->value;
-    int count = br->count;
-    int shift = VP8_BD_VALUE_SIZE - CHAR_BIT - (count + CHAR_BIT);
-    size_t bytes_left = br->user_buffer_end - bufptr;
-    size_t bits_left = bytes_left * CHAR_BIT;
-    int x = shift + CHAR_BIT - (int)bits_left;
-    int loop_end = 0;
-    unsigned char decrypted[sizeof(VP8_BD_VALUE) + 1];
+void vp8dx_bool_decoder_fill(BOOL_DECODER *br) {
+  const unsigned char *bufptr = br->user_buffer;
+  VP8_BD_VALUE value = br->value;
+  int count = br->count;
+  int shift = VP8_BD_VALUE_SIZE - CHAR_BIT - (count + CHAR_BIT);
+  size_t bytes_left = br->user_buffer_end - bufptr;
+  size_t bits_left = bytes_left * CHAR_BIT;
+  int x = shift + CHAR_BIT - (int)bits_left;
+  int loop_end = 0;
+  unsigned char decrypted[sizeof(VP8_BD_VALUE) + 1];
 
-    if (br->decrypt_cb) {
-        size_t n = VPXMIN(sizeof(decrypted), bytes_left);
-        br->decrypt_cb(br->decrypt_state, bufptr, decrypted, (int)n);
-        bufptr = decrypted;
-    }
+  if (br->decrypt_cb) {
+    size_t n = VPXMIN(sizeof(decrypted), bytes_left);
+    br->decrypt_cb(br->decrypt_state, bufptr, decrypted, (int)n);
+    bufptr = decrypted;
+  }
 
-    if(x >= 0)
-    {
-        count += VP8_LOTS_OF_BITS;
-        loop_end = x;
-    }
+  if (x >= 0) {
+    count += VP8_LOTS_OF_BITS;
+    loop_end = x;
+  }
 
-    if (x < 0 || bits_left)
-    {
-        while(shift >= loop_end)
-        {
-            count += CHAR_BIT;
-            value |= (VP8_BD_VALUE)*bufptr << shift;
-            ++bufptr;
-            ++br->user_buffer;
-            shift -= CHAR_BIT;
-        }
+  if (x < 0 || bits_left) {
+    while (shift >= loop_end) {
+      count += CHAR_BIT;
+      value |= (VP8_BD_VALUE)*bufptr << shift;
+      ++bufptr;
+      ++br->user_buffer;
+      shift -= CHAR_BIT;
     }
+  }
 
-    br->value = value;
-    br->count = count;
+  br->value = value;
+  br->count = count;
 }
--- a/vp8/decoder/dboolhuff.h
+++ b/vp8/decoder/dboolhuff.h
@@ -8,7 +8,6 @@
  *  be found in the AUTHORS file in the root of the source tree.
  */
 
-
 #ifndef VP8_DECODER_DBOOLHUFF_H_
 #define VP8_DECODER_DBOOLHUFF_H_
 
@@ -26,7 +25,7 @@
 
 typedef size_t VP8_BD_VALUE;
 
-#define VP8_BD_VALUE_SIZE ((int)sizeof(VP8_BD_VALUE)*CHAR_BIT)
+#define VP8_BD_VALUE_SIZE ((int)sizeof(VP8_BD_VALUE) * CHAR_BIT)
 
 /*This is meant to be a large, positive constant that can still be efficiently
    loaded as an immediate (on platforms like ARM, for example).
@@ -33,105 +32,95 @@
   Even relatively modest values like 100 would work fine.*/
 #define VP8_LOTS_OF_BITS (0x40000000)
 
-typedef struct
-{
-    const unsigned char *user_buffer_end;
-    const unsigned char *user_buffer;
-    VP8_BD_VALUE         value;
-    int                  count;
-    unsigned int         range;
-    vpx_decrypt_cb       decrypt_cb;
-    void                *decrypt_state;
+typedef struct {
+  const unsigned char *user_buffer_end;
+  const unsigned char *user_buffer;
+  VP8_BD_VALUE value;
+  int count;
+  unsigned int range;
+  vpx_decrypt_cb decrypt_cb;
+  void *decrypt_state;
 } BOOL_DECODER;
 
 DECLARE_ALIGNED(16, extern const unsigned char, vp8_norm[256]);
 
-int vp8dx_start_decode(BOOL_DECODER *br,
-                       const unsigned char *source,
-                       unsigned int source_sz,
-                       vpx_decrypt_cb decrypt_cb,
+int vp8dx_start_decode(BOOL_DECODER *br, const unsigned char *source,
+                       unsigned int source_sz, vpx_decrypt_cb decrypt_cb,
                        void *decrypt_state);
 
 void vp8dx_bool_decoder_fill(BOOL_DECODER *br);
 
-
 static int vp8dx_decode_bool(BOOL_DECODER *br, int probability) {
-    unsigned int bit = 0;
-    VP8_BD_VALUE value;
-    unsigned int split;
-    VP8_BD_VALUE bigsplit;
-    int count;
-    unsigned int range;
+  unsigned int bit = 0;
+  VP8_BD_VALUE value;
+  unsigned int split;
+  VP8_BD_VALUE bigsplit;
+  int count;
+  unsigned int range;
 
-    split = 1 + (((br->range - 1) * probability) >> 8);
+  split = 1 + (((br->range - 1) * probability) >> 8);
 
-    if(br->count < 0)
-        vp8dx_bool_decoder_fill(br);
+  if (br->count < 0) vp8dx_bool_decoder_fill(br);
 
-    value = br->value;
-    count = br->count;
+  value = br->value;
+  count = br->count;
 
-    bigsplit = (VP8_BD_VALUE)split << (VP8_BD_VALUE_SIZE - 8);
+  bigsplit = (VP8_BD_VALUE)split << (VP8_BD_VALUE_SIZE - 8);
 
-    range = split;
+  range = split;
 
-    if (value >= bigsplit)
-    {
-        range = br->range - split;
-        value = value - bigsplit;
-        bit = 1;
-    }
+  if (value >= bigsplit) {
+    range = br->range - split;
+    value = value - bigsplit;
+    bit = 1;
+  }
 
-    {
-        register int shift = vp8_norm[range];
-        range <<= shift;
-        value <<= shift;
-        count -= shift;
-    }
-    br->value = value;
-    br->count = count;
-    br->range = range;
+  {
+    register int shift = vp8_norm[range];
+    range <<= shift;
+    value <<= shift;
+    count -= shift;
+  }
+  br->value = value;
+  br->count = count;
+  br->range = range;
 
-    return bit;
+  return bit;
 }
 
-static INLINE int vp8_decode_value(BOOL_DECODER *br, int bits)
-{
-    int z = 0;
-    int bit;
+static INLINE int vp8_decode_value(BOOL_DECODER *br, int bits) {
+  int z = 0;
+  int bit;
 
-    for (bit = bits - 1; bit >= 0; bit--)
-    {
-        z |= (vp8dx_decode_bool(br, 0x80) << bit);
-    }
+  for (bit = bits - 1; bit >= 0; bit--) {
+    z |= (vp8dx_decode_bool(br, 0x80) << bit);
+  }
 
-    return z;
+  return z;
 }
 
-static INLINE int vp8dx_bool_error(BOOL_DECODER *br)
-{
-    /* Check if we have reached the end of the buffer.
-     *
-     * Variable 'count' stores the number of bits in the 'value' buffer, minus
-     * 8. The top byte is part of the algorithm, and the remainder is buffered
-     * to be shifted into it. So if count == 8, the top 16 bits of 'value' are
-     * occupied, 8 for the algorithm and 8 in the buffer.
-     *
-     * When reading a byte from the user's buffer, count is filled with 8 and
-     * one byte is filled into the value buffer. When we reach the end of the
-     * data, count is additionally filled with VP8_LOTS_OF_BITS. So when
-     * count == VP8_LOTS_OF_BITS - 1, the user's data has been exhausted.
+static INLINE int vp8dx_bool_error(BOOL_DECODER *br) {
+  /* Check if we have reached the end of the buffer.
+   *
+   * Variable 'count' stores the number of bits in the 'value' buffer, minus
+   * 8. The top byte is part of the algorithm, and the remainder is buffered
+   * to be shifted into it. So if count == 8, the top 16 bits of 'value' are
+   * occupied, 8 for the algorithm and 8 in the buffer.
+   *
+   * When reading a byte from the user's buffer, count is filled with 8 and
+   * one byte is filled into the value buffer. When we reach the end of the
+   * data, count is additionally filled with VP8_LOTS_OF_BITS. So when
+   * count == VP8_LOTS_OF_BITS - 1, the user's data has been exhausted.
+   */
+  if ((br->count > VP8_BD_VALUE_SIZE) && (br->count < VP8_LOTS_OF_BITS)) {
+    /* We have tried to decode bits after the end of
+     * stream was encountered.
      */
-    if ((br->count > VP8_BD_VALUE_SIZE) && (br->count < VP8_LOTS_OF_BITS))
-    {
-       /* We have tried to decode bits after the end of
-        * stream was encountered.
-        */
-        return 1;
-    }
+    return 1;
+  }
 
-    /* No error. */
-    return 0;
+  /* No error. */
+  return 0;
 }
 
 #ifdef __cplusplus
--- a/vp8/decoder/decodeframe.c
+++ b/vp8/decoder/decodeframe.c
@@ -8,7 +8,6 @@
  *  be found in the AUTHORS file in the root of the source tree.
  */
 
-
 #include "vpx_config.h"
 #include "vp8_rtcd.h"
 #include "./vpx_scale_rtcd.h"
@@ -40,262 +39,210 @@
 #include <assert.h>
 #include <stdio.h>
 
-void vp8cx_init_de_quantizer(VP8D_COMP *pbi)
-{
-    int Q;
-    VP8_COMMON *const pc = & pbi->common;
+void vp8cx_init_de_quantizer(VP8D_COMP *pbi) {
+  int Q;
+  VP8_COMMON *const pc = &pbi->common;
 
-    for (Q = 0; Q < QINDEX_RANGE; Q++)
-    {
-        pc->Y1dequant[Q][0] = (short)vp8_dc_quant(Q, pc->y1dc_delta_q);
-        pc->Y2dequant[Q][0] = (short)vp8_dc2quant(Q, pc->y2dc_delta_q);
-        pc->UVdequant[Q][0] = (short)vp8_dc_uv_quant(Q, pc->uvdc_delta_q);
+  for (Q = 0; Q < QINDEX_RANGE; Q++) {
+    pc->Y1dequant[Q][0] = (short)vp8_dc_quant(Q, pc->y1dc_delta_q);
+    pc->Y2dequant[Q][0] = (short)vp8_dc2quant(Q, pc->y2dc_delta_q);
+    pc->UVdequant[Q][0] = (short)vp8_dc_uv_quant(Q, pc->uvdc_delta_q);
 
-        pc->Y1dequant[Q][1] = (short)vp8_ac_yquant(Q);
-        pc->Y2dequant[Q][1] = (short)vp8_ac2quant(Q, pc->y2ac_delta_q);
-        pc->UVdequant[Q][1] = (short)vp8_ac_uv_quant(Q, pc->uvac_delta_q);
-    }
+    pc->Y1dequant[Q][1] = (short)vp8_ac_yquant(Q);
+    pc->Y2dequant[Q][1] = (short)vp8_ac2quant(Q, pc->y2ac_delta_q);
+    pc->UVdequant[Q][1] = (short)vp8_ac_uv_quant(Q, pc->uvac_delta_q);
+  }
 }
 
-void vp8_mb_init_dequantizer(VP8D_COMP *pbi, MACROBLOCKD *xd)
-{
-    int i;
-    int QIndex;
-    MB_MODE_INFO *mbmi = &xd->mode_info_context->mbmi;
-    VP8_COMMON *const pc = & pbi->common;
+void vp8_mb_init_dequantizer(VP8D_COMP *pbi, MACROBLOCKD *xd) {
+  int i;
+  int QIndex;
+  MB_MODE_INFO *mbmi = &xd->mode_info_context->mbmi;
+  VP8_COMMON *const pc = &pbi->common;
 
-    /* Decide whether to use the default or alternate baseline Q value. */
-    if (xd->segmentation_enabled)
-    {
-        /* Abs Value */
-        if (xd->mb_segement_abs_delta == SEGMENT_ABSDATA)
-            QIndex = xd->segment_feature_data[MB_LVL_ALT_Q][mbmi->segment_id];
+  /* Decide whether to use the default or alternate baseline Q value. */
+  if (xd->segmentation_enabled) {
+    /* Abs Value */
+    if (xd->mb_segement_abs_delta == SEGMENT_ABSDATA)
+      QIndex = xd->segment_feature_data[MB_LVL_ALT_Q][mbmi->segment_id];
 
-        /* Delta Value */
-        else
-            QIndex = pc->base_qindex + xd->segment_feature_data[MB_LVL_ALT_Q][mbmi->segment_id];
-
-        QIndex = (QIndex >= 0) ? ((QIndex <= MAXQ) ? QIndex : MAXQ) : 0;    /* Clamp to valid range */
-    }
+    /* Delta Value */
     else
-        QIndex = pc->base_qindex;
+      QIndex = pc->base_qindex +
+               xd->segment_feature_data[MB_LVL_ALT_Q][mbmi->segment_id];
 
-    /* Set up the macroblock dequant constants */
-    xd->dequant_y1_dc[0] = 1;
-    xd->dequant_y1[0] = pc->Y1dequant[QIndex][0];
-    xd->dequant_y2[0] = pc->Y2dequant[QIndex][0];
-    xd->dequant_uv[0] = pc->UVdequant[QIndex][0];
+    QIndex = (QIndex >= 0) ? ((QIndex <= MAXQ) ? QIndex : MAXQ)
+                           : 0; /* Clamp to valid range */
+  } else
+    QIndex = pc->base_qindex;
 
-    for (i = 1; i < 16; i++)
-    {
-        xd->dequant_y1_dc[i] =
-        xd->dequant_y1[i] = pc->Y1dequant[QIndex][1];
-        xd->dequant_y2[i] = pc->Y2dequant[QIndex][1];
-        xd->dequant_uv[i] = pc->UVdequant[QIndex][1];
-    }
+  /* Set up the macroblock dequant constants */
+  xd->dequant_y1_dc[0] = 1;
+  xd->dequant_y1[0] = pc->Y1dequant[QIndex][0];
+  xd->dequant_y2[0] = pc->Y2dequant[QIndex][0];
+  xd->dequant_uv[0] = pc->UVdequant[QIndex][0];
+
+  for (i = 1; i < 16; i++) {
+    xd->dequant_y1_dc[i] = xd->dequant_y1[i] = pc->Y1dequant[QIndex][1];
+    xd->dequant_y2[i] = pc->Y2dequant[QIndex][1];
+    xd->dequant_uv[i] = pc->UVdequant[QIndex][1];
+  }
 }
 
 static void decode_macroblock(VP8D_COMP *pbi, MACROBLOCKD *xd,
-                              unsigned int mb_idx)
-{
-    MB_PREDICTION_MODE mode;
-    int i;
+                              unsigned int mb_idx) {
+  MB_PREDICTION_MODE mode;
+  int i;
 #if CONFIG_ERROR_CONCEALMENT
-    int corruption_detected = 0;
+  int corruption_detected = 0;
 #else
-    (void)mb_idx;
+  (void)mb_idx;
 #endif
 
-    if (xd->mode_info_context->mbmi.mb_skip_coeff)
-    {
-        vp8_reset_mb_tokens_context(xd);
-    }
-    else if (!vp8dx_bool_error(xd->current_bc))
-    {
-        int eobtotal;
-        eobtotal = vp8_decode_mb_tokens(pbi, xd);
+  if (xd->mode_info_context->mbmi.mb_skip_coeff) {
+    vp8_reset_mb_tokens_context(xd);
+  } else if (!vp8dx_bool_error(xd->current_bc)) {
+    int eobtotal;
+    eobtotal = vp8_decode_mb_tokens(pbi, xd);
 
-        /* Special case:  Force the loopfilter to skip when eobtotal is zero */
-        xd->mode_info_context->mbmi.mb_skip_coeff = (eobtotal==0);
-    }
+    /* Special case:  Force the loopfilter to skip when eobtotal is zero */
+    xd->mode_info_context->mbmi.mb_skip_coeff = (eobtotal == 0);
+  }
 
-    mode = xd->mode_info_context->mbmi.mode;
+  mode = xd->mode_info_context->mbmi.mode;
 
-    if (xd->segmentation_enabled)
-        vp8_mb_init_dequantizer(pbi, xd);
+  if (xd->segmentation_enabled) vp8_mb_init_dequantizer(pbi, xd);
 
-
 #if CONFIG_ERROR_CONCEALMENT
 
-    if(pbi->ec_active)
-    {
-        int throw_residual;
-        /* When we have independent partitions we can apply residual even
-         * though other partitions within the frame are corrupt.
-         */
-        throw_residual = (!pbi->independent_partitions &&
-                          pbi->frame_corrupt_residual);
-        throw_residual = (throw_residual || vp8dx_bool_error(xd->current_bc));
+  if (pbi->ec_active) {
+    int throw_residual;
+    /* When we have independent partitions we can apply residual even
+     * though other partitions within the frame are corrupt.
+     */
+    throw_residual =
+        (!pbi->independent_partitions && pbi->frame_corrupt_residual);
+    throw_residual = (throw_residual || vp8dx_bool_error(xd->current_bc));
 
-        if ((mb_idx >= pbi->mvs_corrupt_from_mb || throw_residual))
-        {
-            /* MB with corrupt residuals or corrupt mode/motion vectors.
-             * Better to use the predictor as reconstruction.
-             */
-            pbi->frame_corrupt_residual = 1;
-            memset(xd->qcoeff, 0, sizeof(xd->qcoeff));
+    if ((mb_idx >= pbi->mvs_corrupt_from_mb || throw_residual)) {
+      /* MB with corrupt residuals or corrupt mode/motion vectors.
+       * Better to use the predictor as reconstruction.
+       */
+      pbi->frame_corrupt_residual = 1;
+      memset(xd->qcoeff, 0, sizeof(xd->qcoeff));
 
-            corruption_detected = 1;
+      corruption_detected = 1;
 
-            /* force idct to be skipped for B_PRED and use the
-             * prediction only for reconstruction
-             * */
-            memset(xd->eobs, 0, 25);
-        }
+      /* force idct to be skipped for B_PRED and use the
+       * prediction only for reconstruction
+       * */
+      memset(xd->eobs, 0, 25);
     }
+  }
 #endif
 
-    /* do prediction */
-    if (xd->mode_info_context->mbmi.ref_frame == INTRA_FRAME)
-    {
-        vp8_build_intra_predictors_mbuv_s(xd,
-                                          xd->recon_above[1],
-                                          xd->recon_above[2],
-                                          xd->recon_left[1],
-                                          xd->recon_left[2],
-                                          xd->recon_left_stride[1],
-                                          xd->dst.u_buffer, xd->dst.v_buffer,
-                                          xd->dst.uv_stride);
+  /* do prediction */
+  if (xd->mode_info_context->mbmi.ref_frame == INTRA_FRAME) {
+    vp8_build_intra_predictors_mbuv_s(
+        xd, xd->recon_above[1], xd->recon_above[2], xd->recon_left[1],
+        xd->recon_left[2], xd->recon_left_stride[1], xd->dst.u_buffer,
+        xd->dst.v_buffer, xd->dst.uv_stride);
 
-        if (mode != B_PRED)
-        {
-            vp8_build_intra_predictors_mby_s(xd,
-                                                 xd->recon_above[0],
-                                                 xd->recon_left[0],
-                                                 xd->recon_left_stride[0],
-                                                 xd->dst.y_buffer,
-                                                 xd->dst.y_stride);
-        }
-        else
-        {
-            short *DQC = xd->dequant_y1;
-            int dst_stride = xd->dst.y_stride;
+    if (mode != B_PRED) {
+      vp8_build_intra_predictors_mby_s(
+          xd, xd->recon_above[0], xd->recon_left[0], xd->recon_left_stride[0],
+          xd->dst.y_buffer, xd->dst.y_stride);
+    } else {
+      short *DQC = xd->dequant_y1;
+      int dst_stride = xd->dst.y_stride;
 
-            /* clear out residual eob info */
-            if(xd->mode_info_context->mbmi.mb_skip_coeff)
-                memset(xd->eobs, 0, 25);
+      /* clear out residual eob info */
+      if (xd->mode_info_context->mbmi.mb_skip_coeff) memset(xd->eobs, 0, 25);
 
-            intra_prediction_down_copy(xd, xd->recon_above[0] + 16);
+      intra_prediction_down_copy(xd, xd->recon_above[0] + 16);
 
-            for (i = 0; i < 16; i++)
-            {
-                BLOCKD *b = &xd->block[i];
-                unsigned char *dst = xd->dst.y_buffer + b->offset;
-                B_PREDICTION_MODE b_mode =
-                    xd->mode_info_context->bmi[i].as_mode;
-                unsigned char *Above = dst - dst_stride;
-                unsigned char *yleft = dst - 1;
-                int left_stride = dst_stride;
-                unsigned char top_left = Above[-1];
+      for (i = 0; i < 16; i++) {
+        BLOCKD *b = &xd->block[i];
+        unsigned char *dst = xd->dst.y_buffer + b->offset;
+        B_PREDICTION_MODE b_mode = xd->mode_info_context->bmi[i].as_mode;
+        unsigned char *Above = dst - dst_stride;
+        unsigned char *yleft = dst - 1;
+        int left_stride = dst_stride;
+        unsigned char top_left = Above[-1];
 
-                vp8_intra4x4_predict(Above, yleft, left_stride, b_mode,
-                                     dst, dst_stride, top_left);
+        vp8_intra4x4_predict(Above, yleft, left_stride, b_mode, dst, dst_stride,
+                             top_left);
 
-                if (xd->eobs[i])
-                {
-                    if (xd->eobs[i] > 1)
-                    {
-                    vp8_dequant_idct_add(b->qcoeff, DQC, dst, dst_stride);
-                    }
-                    else
-                    {
-                        vp8_dc_only_idct_add
-                            (b->qcoeff[0] * DQC[0],
-                                dst, dst_stride,
-                                dst, dst_stride);
-                        memset(b->qcoeff, 0, 2 * sizeof(b->qcoeff[0]));
-                    }
-                }
-            }
+        if (xd->eobs[i]) {
+          if (xd->eobs[i] > 1) {
+            vp8_dequant_idct_add(b->qcoeff, DQC, dst, dst_stride);
+          } else {
+            vp8_dc_only_idct_add(b->qcoeff[0] * DQC[0], dst, dst_stride, dst,
+                                 dst_stride);
+            memset(b->qcoeff, 0, 2 * sizeof(b->qcoeff[0]));
+          }
         }
+      }
     }
-    else
-    {
-        vp8_build_inter_predictors_mb(xd);
-    }
+  } else {
+    vp8_build_inter_predictors_mb(xd);
+  }
 
-
 #if CONFIG_ERROR_CONCEALMENT
-    if (corruption_detected)
-    {
-        return;
-    }
+  if (corruption_detected) {
+    return;
+  }
 #endif
 
-    if(!xd->mode_info_context->mbmi.mb_skip_coeff)
-    {
-        /* dequantization and idct */
-        if (mode != B_PRED)
-        {
-            short *DQC = xd->dequant_y1;
+  if (!xd->mode_info_context->mbmi.mb_skip_coeff) {
+    /* dequantization and idct */
+    if (mode != B_PRED) {
+      short *DQC = xd->dequant_y1;
 
-            if (mode != SPLITMV)
-            {
-                BLOCKD *b = &xd->block[24];
+      if (mode != SPLITMV) {
+        BLOCKD *b = &xd->block[24];
 
-                /* do 2nd order transform on the dc block */
-                if (xd->eobs[24] > 1)
-                {
-                    vp8_dequantize_b(b, xd->dequant_y2);
+        /* do 2nd order transform on the dc block */
+        if (xd->eobs[24] > 1) {
+          vp8_dequantize_b(b, xd->dequant_y2);
 
-                    vp8_short_inv_walsh4x4(&b->dqcoeff[0],
-                        xd->qcoeff);
-                    memset(b->qcoeff, 0, 16 * sizeof(b->qcoeff[0]));
-                }
-                else
-                {
-                    b->dqcoeff[0] = b->qcoeff[0] * xd->dequant_y2[0];
-                    vp8_short_inv_walsh4x4_1(&b->dqcoeff[0],
-                        xd->qcoeff);
-                    memset(b->qcoeff, 0, 2 * sizeof(b->qcoeff[0]));
-                }
-
-                /* override the dc dequant constant in order to preserve the
-                 * dc components
-                 */
-                DQC = xd->dequant_y1_dc;
-            }
-
-            vp8_dequant_idct_add_y_block
-                            (xd->qcoeff, DQC,
-                             xd->dst.y_buffer,
-                             xd->dst.y_stride, xd->eobs);
+          vp8_short_inv_walsh4x4(&b->dqcoeff[0], xd->qcoeff);
+          memset(b->qcoeff, 0, 16 * sizeof(b->qcoeff[0]));
+        } else {
+          b->dqcoeff[0] = b->qcoeff[0] * xd->dequant_y2[0];
+          vp8_short_inv_walsh4x4_1(&b->dqcoeff[0], xd->qcoeff);
+          memset(b->qcoeff, 0, 2 * sizeof(b->qcoeff[0]));
         }
 
-        vp8_dequant_idct_add_uv_block
-                        (xd->qcoeff+16*16, xd->dequant_uv,
-                         xd->dst.u_buffer, xd->dst.v_buffer,
-                         xd->dst.uv_stride, xd->eobs+16);
+        /* override the dc dequant constant in order to preserve the
+         * dc components
+         */
+        DQC = xd->dequant_y1_dc;
+      }
+
+      vp8_dequant_idct_add_y_block(xd->qcoeff, DQC, xd->dst.y_buffer,
+                                   xd->dst.y_stride, xd->eobs);
     }
+
+    vp8_dequant_idct_add_uv_block(xd->qcoeff + 16 * 16, xd->dequant_uv,
+                                  xd->dst.u_buffer, xd->dst.v_buffer,
+                                  xd->dst.uv_stride, xd->eobs + 16);
+  }
 }
 
-static int get_delta_q(vp8_reader *bc, int prev, int *q_update)
-{
-    int ret_val = 0;
+static int get_delta_q(vp8_reader *bc, int prev, int *q_update) {
+  int ret_val = 0;
 
-    if (vp8_read_bit(bc))
-    {
-        ret_val = vp8_read_literal(bc, 4);
+  if (vp8_read_bit(bc)) {
+    ret_val = vp8_read_literal(bc, 4);
 
-        if (vp8_read_bit(bc))
-            ret_val = -ret_val;
-    }
+    if (vp8_read_bit(bc)) ret_val = -ret_val;
+  }
 
-    /* Trigger a quantizer update if the delta-q value has changed */
-    if (ret_val != prev)
-        *q_update = 1;
+  /* Trigger a quantizer update if the delta-q value has changed */
+  if (ret_val != prev) *q_update = 1;
 
-    return ret_val;
+  return ret_val;
 }
 
 #ifdef PACKET_TESTING
@@ -303,1095 +250,970 @@
 FILE *vpxlog = 0;
 #endif
 
-static void yv12_extend_frame_top_c(YV12_BUFFER_CONFIG *ybf)
-{
-    int i;
-    unsigned char *src_ptr1;
-    unsigned char *dest_ptr1;
+static void yv12_extend_frame_top_c(YV12_BUFFER_CONFIG *ybf) {
+  int i;
+  unsigned char *src_ptr1;
+  unsigned char *dest_ptr1;
 
-    unsigned int Border;
-    int plane_stride;
+  unsigned int Border;
+  int plane_stride;
 
-    /***********/
-    /* Y Plane */
-    /***********/
-    Border = ybf->border;
-    plane_stride = ybf->y_stride;
-    src_ptr1 = ybf->y_buffer - Border;
-    dest_ptr1 = src_ptr1 - (Border * plane_stride);
+  /***********/
+  /* Y Plane */
+  /***********/
+  Border = ybf->border;
+  plane_stride = ybf->y_stride;
+  src_ptr1 = ybf->y_buffer - Border;
+  dest_ptr1 = src_ptr1 - (Border * plane_stride);
 
-    for (i = 0; i < (int)Border; i++)
-    {
-        memcpy(dest_ptr1, src_ptr1, plane_stride);
-        dest_ptr1 += plane_stride;
-    }
+  for (i = 0; i < (int)Border; i++) {
+    memcpy(dest_ptr1, src_ptr1, plane_stride);
+    dest_ptr1 += plane_stride;
+  }
 
+  /***********/
+  /* U Plane */
+  /***********/
+  plane_stride = ybf->uv_stride;
+  Border /= 2;
+  src_ptr1 = ybf->u_buffer - Border;
+  dest_ptr1 = src_ptr1 - (Border * plane_stride);
 
-    /***********/
-    /* U Plane */
-    /***********/
-    plane_stride = ybf->uv_stride;
-    Border /= 2;
-    src_ptr1 = ybf->u_buffer - Border;
-    dest_ptr1 = src_ptr1 - (Border * plane_stride);
+  for (i = 0; i < (int)(Border); i++) {
+    memcpy(dest_ptr1, src_ptr1, plane_stride);
+    dest_ptr1 += plane_stride;
+  }
 
-    for (i = 0; i < (int)(Border); i++)
-    {
-        memcpy(dest_ptr1, src_ptr1, plane_stride);
-        dest_ptr1 += plane_stride;
-    }
+  /***********/
+  /* V Plane */
+  /***********/
 
-    /***********/
-    /* V Plane */
-    /***********/
+  src_ptr1 = ybf->v_buffer - Border;
+  dest_ptr1 = src_ptr1 - (Border * plane_stride);
 
-    src_ptr1 = ybf->v_buffer - Border;
-    dest_ptr1 = src_ptr1 - (Border * plane_stride);
-
-    for (i = 0; i < (int)(Border); i++)
-    {
-        memcpy(dest_ptr1, src_ptr1, plane_stride);
-        dest_ptr1 += plane_stride;
-    }
+  for (i = 0; i < (int)(Border); i++) {
+    memcpy(dest_ptr1, src_ptr1, plane_stride);
+    dest_ptr1 += plane_stride;
+  }
 }
 
-static void yv12_extend_frame_bottom_c(YV12_BUFFER_CONFIG *ybf)
-{
-    int i;
-    unsigned char *src_ptr1, *src_ptr2;
-    unsigned char *dest_ptr2;
+static void yv12_extend_frame_bottom_c(YV12_BUFFER_CONFIG *ybf) {
+  int i;
+  unsigned char *src_ptr1, *src_ptr2;
+  unsigned char *dest_ptr2;
 
-    unsigned int Border;
-    int plane_stride;
-    int plane_height;
+  unsigned int Border;
+  int plane_stride;
+  int plane_height;
 
-    /***********/
-    /* Y Plane */
-    /***********/
-    Border = ybf->border;
-    plane_stride = ybf->y_stride;
-    plane_height = ybf->y_height;
+  /***********/
+  /* Y Plane */
+  /***********/
+  Border = ybf->border;
+  plane_stride = ybf->y_stride;
+  plane_height = ybf->y_height;
 
-    src_ptr1 = ybf->y_buffer - Border;
-    src_ptr2 = src_ptr1 + (plane_height * plane_stride) - plane_stride;
-    dest_ptr2 = src_ptr2 + plane_stride;
+  src_ptr1 = ybf->y_buffer - Border;
+  src_ptr2 = src_ptr1 + (plane_height * plane_stride) - plane_stride;
+  dest_ptr2 = src_ptr2 + plane_stride;
 
-    for (i = 0; i < (int)Border; i++)
-    {
-        memcpy(dest_ptr2, src_ptr2, plane_stride);
-        dest_ptr2 += plane_stride;
-    }
+  for (i = 0; i < (int)Border; i++) {
+    memcpy(dest_ptr2, src_ptr2, plane_stride);
+    dest_ptr2 += plane_stride;
+  }
 
+  /***********/
+  /* U Plane */
+  /***********/
+  plane_stride = ybf->uv_stride;
+  plane_height = ybf->uv_height;
+  Border /= 2;
 
-    /***********/
-    /* U Plane */
-    /***********/
-    plane_stride = ybf->uv_stride;
-    plane_height = ybf->uv_height;
-    Border /= 2;
+  src_ptr1 = ybf->u_buffer - Border;
+  src_ptr2 = src_ptr1 + (plane_height * plane_stride) - plane_stride;
+  dest_ptr2 = src_ptr2 + plane_stride;
 
-    src_ptr1 = ybf->u_buffer - Border;
-    src_ptr2 = src_ptr1 + (plane_height * plane_stride) - plane_stride;
-    dest_ptr2 = src_ptr2 + plane_stride;
+  for (i = 0; i < (int)(Border); i++) {
+    memcpy(dest_ptr2, src_ptr2, plane_stride);
+    dest_ptr2 += plane_stride;
+  }
 
-    for (i = 0; i < (int)(Border); i++)
-    {
-        memcpy(dest_ptr2, src_ptr2, plane_stride);
-        dest_ptr2 += plane_stride;
-    }
+  /***********/
+  /* V Plane */
+  /***********/
 
-    /***********/
-    /* V Plane */
-    /***********/
+  src_ptr1 = ybf->v_buffer - Border;
+  src_ptr2 = src_ptr1 + (plane_height * plane_stride) - plane_stride;
+  dest_ptr2 = src_ptr2 + plane_stride;
 
-    src_ptr1 = ybf->v_buffer - Border;
-    src_ptr2 = src_ptr1 + (plane_height * plane_stride) - plane_stride;
-    dest_ptr2 = src_ptr2 + plane_stride;
-
-    for (i = 0; i < (int)(Border); i++)
-    {
-        memcpy(dest_ptr2, src_ptr2, plane_stride);
-        dest_ptr2 += plane_stride;
-    }
+  for (i = 0; i < (int)(Border); i++) {
+    memcpy(dest_ptr2, src_ptr2, plane_stride);
+    dest_ptr2 += plane_stride;
+  }
 }
 
 static void yv12_extend_frame_left_right_c(YV12_BUFFER_CONFIG *ybf,
                                            unsigned char *y_src,
                                            unsigned char *u_src,
-                                           unsigned char *v_src)
-{
-    int i;
-    unsigned char *src_ptr1, *src_ptr2;
-    unsigned char *dest_ptr1, *dest_ptr2;
+                                           unsigned char *v_src) {
+  int i;
+  unsigned char *src_ptr1, *src_ptr2;
+  unsigned char *dest_ptr1, *dest_ptr2;
 
-    unsigned int Border;
-    int plane_stride;
-    int plane_height;
-    int plane_width;
+  unsigned int Border;
+  int plane_stride;
+  int plane_height;
+  int plane_width;
 
-    /***********/
-    /* Y Plane */
-    /***********/
-    Border = ybf->border;
-    plane_stride = ybf->y_stride;
-    plane_height = 16;
-    plane_width = ybf->y_width;
+  /***********/
+  /* Y Plane */
+  /***********/
+  Border = ybf->border;
+  plane_stride = ybf->y_stride;
+  plane_height = 16;
+  plane_width = ybf->y_width;
 
-    /* copy the left and right most columns out */
-    src_ptr1 = y_src;
-    src_ptr2 = src_ptr1 + plane_width - 1;
-    dest_ptr1 = src_ptr1 - Border;
-    dest_ptr2 = src_ptr2 + 1;
+  /* copy the left and right most columns out */
+  src_ptr1 = y_src;
+  src_ptr2 = src_ptr1 + plane_width - 1;
+  dest_ptr1 = src_ptr1 - Border;
+  dest_ptr2 = src_ptr2 + 1;
 
-    for (i = 0; i < plane_height; i++)
-    {
-        memset(dest_ptr1, src_ptr1[0], Border);
-        memset(dest_ptr2, src_ptr2[0], Border);
-        src_ptr1  += plane_stride;
-        src_ptr2  += plane_stride;
-        dest_ptr1 += plane_stride;
-        dest_ptr2 += plane_stride;
-    }
+  for (i = 0; i < plane_height; i++) {
+    memset(dest_ptr1, src_ptr1[0], Border);
+    memset(dest_ptr2, src_ptr2[0], Border);
+    src_ptr1 += plane_stride;
+    src_ptr2 += plane_stride;
+    dest_ptr1 += plane_stride;
+    dest_ptr2 += plane_stride;
+  }
 
-    /***********/
-    /* U Plane */
-    /***********/
-    plane_stride = ybf->uv_stride;
-    plane_height = 8;
-    plane_width = ybf->uv_width;
-    Border /= 2;
+  /***********/
+  /* U Plane */
+  /***********/
+  plane_stride = ybf->uv_stride;
+  plane_height = 8;
+  plane_width = ybf->uv_width;
+  Border /= 2;
 
-    /* copy the left and right most columns out */
-    src_ptr1 = u_src;
-    src_ptr2 = src_ptr1 + plane_width - 1;
-    dest_ptr1 = src_ptr1 - Border;
-    dest_ptr2 = src_ptr2 + 1;
+  /* copy the left and right most columns out */
+  src_ptr1 = u_src;
+  src_ptr2 = src_ptr1 + plane_width - 1;
+  dest_ptr1 = src_ptr1 - Border;
+  dest_ptr2 = src_ptr2 + 1;
 
-    for (i = 0; i < plane_height; i++)
-    {
-        memset(dest_ptr1, src_ptr1[0], Border);
-        memset(dest_ptr2, src_ptr2[0], Border);
-        src_ptr1  += plane_stride;
-        src_ptr2  += plane_stride;
-        dest_ptr1 += plane_stride;
-        dest_ptr2 += plane_stride;
-    }
+  for (i = 0; i < plane_height; i++) {
+    memset(dest_ptr1, src_ptr1[0], Border);
+    memset(dest_ptr2, src_ptr2[0], Border);
+    src_ptr1 += plane_stride;
+    src_ptr2 += plane_stride;
+    dest_ptr1 += plane_stride;
+    dest_ptr2 += plane_stride;
+  }
 
-    /***********/
-    /* V Plane */
-    /***********/
+  /***********/
+  /* V Plane */
+  /***********/
 
-    /* copy the left and right most columns out */
-    src_ptr1 = v_src;
-    src_ptr2 = src_ptr1 + plane_width - 1;
-    dest_ptr1 = src_ptr1 - Border;
-    dest_ptr2 = src_ptr2 + 1;
+  /* copy the left and right most columns out */
+  src_ptr1 = v_src;
+  src_ptr2 = src_ptr1 + plane_width - 1;
+  dest_ptr1 = src_ptr1 - Border;
+  dest_ptr2 = src_ptr2 + 1;
 
-    for (i = 0; i < plane_height; i++)
-    {
-        memset(dest_ptr1, src_ptr1[0], Border);
-        memset(dest_ptr2, src_ptr2[0], Border);
-        src_ptr1  += plane_stride;
-        src_ptr2  += plane_stride;
-        dest_ptr1 += plane_stride;
-        dest_ptr2 += plane_stride;
-    }
+  for (i = 0; i < plane_height; i++) {
+    memset(dest_ptr1, src_ptr1[0], Border);
+    memset(dest_ptr2, src_ptr2[0], Border);
+    src_ptr1 += plane_stride;
+    src_ptr2 += plane_stride;
+    dest_ptr1 += plane_stride;
+    dest_ptr2 += plane_stride;
+  }
 }
 
-static void decode_mb_rows(VP8D_COMP *pbi)
-{
-    VP8_COMMON *const pc = & pbi->common;
-    MACROBLOCKD *const xd  = & pbi->mb;
+static void decode_mb_rows(VP8D_COMP *pbi) {
+  VP8_COMMON *const pc = &pbi->common;
+  MACROBLOCKD *const xd = &pbi->mb;
 
-    MODE_INFO *lf_mic = xd->mode_info_context;
+  MODE_INFO *lf_mic = xd->mode_info_context;
 
-    int ibc = 0;
-    int num_part = 1 << pc->multi_token_partition;
+  int ibc = 0;
+  int num_part = 1 << pc->multi_token_partition;
 
-    int recon_yoffset, recon_uvoffset;
-    int mb_row, mb_col;
-    int mb_idx = 0;
+  int recon_yoffset, recon_uvoffset;
+  int mb_row, mb_col;
+  int mb_idx = 0;
 
-    YV12_BUFFER_CONFIG *yv12_fb_new = pbi->dec_fb_ref[INTRA_FRAME];
+  YV12_BUFFER_CONFIG *yv12_fb_new = pbi->dec_fb_ref[INTRA_FRAME];
 
-    int recon_y_stride = yv12_fb_new->y_stride;
-    int recon_uv_stride = yv12_fb_new->uv_stride;
+  int recon_y_stride = yv12_fb_new->y_stride;
+  int recon_uv_stride = yv12_fb_new->uv_stride;
 
-    unsigned char *ref_buffer[MAX_REF_FRAMES][3];
-    unsigned char *dst_buffer[3];
-    unsigned char *lf_dst[3];
-    unsigned char *eb_dst[3];
-    int i;
-    int ref_fb_corrupted[MAX_REF_FRAMES];
+  unsigned char *ref_buffer[MAX_REF_FRAMES][3];
+  unsigned char *dst_buffer[3];
+  unsigned char *lf_dst[3];
+  unsigned char *eb_dst[3];
+  int i;
+  int ref_fb_corrupted[MAX_REF_FRAMES];
 
-    ref_fb_corrupted[INTRA_FRAME] = 0;
+  ref_fb_corrupted[INTRA_FRAME] = 0;
 
-    for(i = 1; i < MAX_REF_FRAMES; i++)
-    {
-        YV12_BUFFER_CONFIG *this_fb = pbi->dec_fb_ref[i];
+  for (i = 1; i < MAX_REF_FRAMES; i++) {
+    YV12_BUFFER_CONFIG *this_fb = pbi->dec_fb_ref[i];
 
-        ref_buffer[i][0] = this_fb->y_buffer;
-        ref_buffer[i][1] = this_fb->u_buffer;
-        ref_buffer[i][2] = this_fb->v_buffer;
+    ref_buffer[i][0] = this_fb->y_buffer;
+    ref_buffer[i][1] = this_fb->u_buffer;
+    ref_buffer[i][2] = this_fb->v_buffer;
 
-        ref_fb_corrupted[i] = this_fb->corrupted;
-    }
+    ref_fb_corrupted[i] = this_fb->corrupted;
+  }
 
-    /* Set up the buffer pointers */
-    eb_dst[0] = lf_dst[0] = dst_buffer[0] = yv12_fb_new->y_buffer;
-    eb_dst[1] = lf_dst[1] = dst_buffer[1] = yv12_fb_new->u_buffer;
-    eb_dst[2] = lf_dst[2] = dst_buffer[2] = yv12_fb_new->v_buffer;
+  /* Set up the buffer pointers */
+  eb_dst[0] = lf_dst[0] = dst_buffer[0] = yv12_fb_new->y_buffer;
+  eb_dst[1] = lf_dst[1] = dst_buffer[1] = yv12_fb_new->u_buffer;
+  eb_dst[2] = lf_dst[2] = dst_buffer[2] = yv12_fb_new->v_buffer;
 
-    xd->up_available = 0;
+  xd->up_available = 0;
 
-    /* Initialize the loop filter for this frame. */
-    if(pc->filter_level)
-        vp8_loop_filter_frame_init(pc, xd, pc->filter_level);
+  /* Initialize the loop filter for this frame. */
+  if (pc->filter_level) vp8_loop_filter_frame_init(pc, xd, pc->filter_level);
 
-    vp8_setup_intra_recon_top_line(yv12_fb_new);
+  vp8_setup_intra_recon_top_line(yv12_fb_new);
 
-    /* Decode the individual macro block */
-    for (mb_row = 0; mb_row < pc->mb_rows; mb_row++)
-    {
-        if (num_part > 1)
-        {
-            xd->current_bc = & pbi->mbc[ibc];
-            ibc++;
+  /* Decode the individual macro block */
+  for (mb_row = 0; mb_row < pc->mb_rows; mb_row++) {
+    if (num_part > 1) {
+      xd->current_bc = &pbi->mbc[ibc];
+      ibc++;
 
-            if (ibc == num_part)
-                ibc = 0;
-        }
+      if (ibc == num_part) ibc = 0;
+    }
 
-        recon_yoffset = mb_row * recon_y_stride * 16;
-        recon_uvoffset = mb_row * recon_uv_stride * 8;
+    recon_yoffset = mb_row * recon_y_stride * 16;
+    recon_uvoffset = mb_row * recon_uv_stride * 8;
 
-        /* reset contexts */
-        xd->above_context = pc->above_context;
-        memset(xd->left_context, 0, sizeof(ENTROPY_CONTEXT_PLANES));
+    /* reset contexts */
+    xd->above_context = pc->above_context;
+    memset(xd->left_context, 0, sizeof(ENTROPY_CONTEXT_PLANES));
 
-        xd->left_available = 0;
+    xd->left_available = 0;
 
-        xd->mb_to_top_edge = -((mb_row * 16) << 3);
-        xd->mb_to_bottom_edge = ((pc->mb_rows - 1 - mb_row) * 16) << 3;
+    xd->mb_to_top_edge = -((mb_row * 16) << 3);
+    xd->mb_to_bottom_edge = ((pc->mb_rows - 1 - mb_row) * 16) << 3;
 
-        xd->recon_above[0] = dst_buffer[0] + recon_yoffset;
-        xd->recon_above[1] = dst_buffer[1] + recon_uvoffset;
-        xd->recon_above[2] = dst_buffer[2] + recon_uvoffset;
+    xd->recon_above[0] = dst_buffer[0] + recon_yoffset;
+    xd->recon_above[1] = dst_buffer[1] + recon_uvoffset;
+    xd->recon_above[2] = dst_buffer[2] + recon_uvoffset;
 
-        xd->recon_left[0] = xd->recon_above[0] - 1;
-        xd->recon_left[1] = xd->recon_above[1] - 1;
-        xd->recon_left[2] = xd->recon_above[2] - 1;
+    xd->recon_left[0] = xd->recon_above[0] - 1;
+    xd->recon_left[1] = xd->recon_above[1] - 1;
+    xd->recon_left[2] = xd->recon_above[2] - 1;
 
-        xd->recon_above[0] -= xd->dst.y_stride;
-        xd->recon_above[1] -= xd->dst.uv_stride;
-        xd->recon_above[2] -= xd->dst.uv_stride;
+    xd->recon_above[0] -= xd->dst.y_stride;
+    xd->recon_above[1] -= xd->dst.uv_stride;
+    xd->recon_above[2] -= xd->dst.uv_stride;
 
-        /* TODO: move to outside row loop */
-        xd->recon_left_stride[0] = xd->dst.y_stride;
-        xd->recon_left_stride[1] = xd->dst.uv_stride;
+    /* TODO: move to outside row loop */
+    xd->recon_left_stride[0] = xd->dst.y_stride;
+    xd->recon_left_stride[1] = xd->dst.uv_stride;
 
-        setup_intra_recon_left(xd->recon_left[0], xd->recon_left[1],
-                               xd->recon_left[2], xd->dst.y_stride,
-                               xd->dst.uv_stride);
+    setup_intra_recon_left(xd->recon_left[0], xd->recon_left[1],
+                           xd->recon_left[2], xd->dst.y_stride,
+                           xd->dst.uv_stride);
 
-        for (mb_col = 0; mb_col < pc->mb_cols; mb_col++)
-        {
-            /* Distance of Mb to the various image edges.
-             * These are specified to 8th pel as they are always compared to values
-             * that are in 1/8th pel units
-             */
-            xd->mb_to_left_edge = -((mb_col * 16) << 3);
-            xd->mb_to_right_edge = ((pc->mb_cols - 1 - mb_col) * 16) << 3;
+    for (mb_col = 0; mb_col < pc->mb_cols; mb_col++) {
+      /* Distance of Mb to the various image edges.
+       * These are specified to 8th pel as they are always compared to values
+       * that are in 1/8th pel units
+       */
+      xd->mb_to_left_edge = -((mb_col * 16) << 3);
+      xd->mb_to_right_edge = ((pc->mb_cols - 1 - mb_col) * 16) << 3;
 
 #if CONFIG_ERROR_CONCEALMENT
-            {
-                int corrupt_residual = (!pbi->independent_partitions &&
-                                       pbi->frame_corrupt_residual) ||
-                                       vp8dx_bool_error(xd->current_bc);
-                if (pbi->ec_active &&
-                    xd->mode_info_context->mbmi.ref_frame == INTRA_FRAME &&
-                    corrupt_residual)
-                {
-                    /* We have an intra block with corrupt coefficients, better to
-                     * conceal with an inter block. Interpolate MVs from neighboring
-                     * MBs.
-                     *
-                     * Note that for the first mb with corrupt residual in a frame,
-                     * we might not discover that before decoding the residual. That
-                     * happens after this check, and therefore no inter concealment
-                     * will be done.
-                     */
-                    vp8_interpolate_motion(xd,
-                                           mb_row, mb_col,
-                                           pc->mb_rows, pc->mb_cols);
-                }
-            }
+      {
+        int corrupt_residual =
+            (!pbi->independent_partitions && pbi->frame_corrupt_residual) ||
+            vp8dx_bool_error(xd->current_bc);
+        if (pbi->ec_active &&
+            xd->mode_info_context->mbmi.ref_frame == INTRA_FRAME &&
+            corrupt_residual) {
+          /* We have an intra block with corrupt coefficients, better to
+           * conceal with an inter block. Interpolate MVs from neighboring
+           * MBs.
+           *
+           * Note that for the first mb with corrupt residual in a frame,
+           * we might not discover that before decoding the residual. That
+           * happens after this check, and therefore no inter concealment
+           * will be done.
+           */
+          vp8_interpolate_motion(xd, mb_row, mb_col, pc->mb_rows, pc->mb_cols);
+        }
+      }
 #endif
 
-            xd->dst.y_buffer = dst_buffer[0] + recon_yoffset;
-            xd->dst.u_buffer = dst_buffer[1] + recon_uvoffset;
-            xd->dst.v_buffer = dst_buffer[2] + recon_uvoffset;
+      xd->dst.y_buffer = dst_buffer[0] + recon_yoffset;
+      xd->dst.u_buffer = dst_buffer[1] + recon_uvoffset;
+      xd->dst.v_buffer = dst_buffer[2] + recon_uvoffset;
 
-            if (xd->mode_info_context->mbmi.ref_frame >= LAST_FRAME) {
-              const MV_REFERENCE_FRAME ref = xd->mode_info_context->mbmi.ref_frame;
-              xd->pre.y_buffer = ref_buffer[ref][0] + recon_yoffset;
-              xd->pre.u_buffer = ref_buffer[ref][1] + recon_uvoffset;
-              xd->pre.v_buffer = ref_buffer[ref][2] + recon_uvoffset;
-            } else {
-              // ref_frame is INTRA_FRAME, pre buffer should not be used.
-              xd->pre.y_buffer = 0;
-              xd->pre.u_buffer = 0;
-              xd->pre.v_buffer = 0;
-            }
+      if (xd->mode_info_context->mbmi.ref_frame >= LAST_FRAME) {
+        const MV_REFERENCE_FRAME ref = xd->mode_info_context->mbmi.ref_frame;
+        xd->pre.y_buffer = ref_buffer[ref][0] + recon_yoffset;
+        xd->pre.u_buffer = ref_buffer[ref][1] + recon_uvoffset;
+        xd->pre.v_buffer = ref_buffer[ref][2] + recon_uvoffset;
+      } else {
+        // ref_frame is INTRA_FRAME, pre buffer should not be used.
+        xd->pre.y_buffer = 0;
+        xd->pre.u_buffer = 0;
+        xd->pre.v_buffer = 0;
+      }
 
-            /* propagate errors from reference frames */
-            xd->corrupted |= ref_fb_corrupted[xd->mode_info_context->mbmi.ref_frame];
+      /* propagate errors from reference frames */
+      xd->corrupted |= ref_fb_corrupted[xd->mode_info_context->mbmi.ref_frame];
 
-            decode_macroblock(pbi, xd, mb_idx);
+      decode_macroblock(pbi, xd, mb_idx);
 
-            mb_idx++;
-            xd->left_available = 1;
+      mb_idx++;
+      xd->left_available = 1;
 
-            /* check if the boolean decoder has suffered an error */
-            xd->corrupted |= vp8dx_bool_error(xd->current_bc);
+      /* check if the boolean decoder has suffered an error */
+      xd->corrupted |= vp8dx_bool_error(xd->current_bc);
 
-            xd->recon_above[0] += 16;
-            xd->recon_above[1] += 8;
-            xd->recon_above[2] += 8;
-            xd->recon_left[0] += 16;
-            xd->recon_left[1] += 8;
-            xd->recon_left[2] += 8;
+      xd->recon_above[0] += 16;
+      xd->recon_above[1] += 8;
+      xd->recon_above[2] += 8;
+      xd->recon_left[0] += 16;
+      xd->recon_left[1] += 8;
+      xd->recon_left[2] += 8;
 
-            recon_yoffset += 16;
-            recon_uvoffset += 8;
+      recon_yoffset += 16;
+      recon_uvoffset += 8;
 
-            ++xd->mode_info_context;  /* next mb */
+      ++xd->mode_info_context; /* next mb */
 
-            xd->above_context++;
-        }
+      xd->above_context++;
+    }
 
-        /* adjust to the next row of mbs */
-        vp8_extend_mb_row(yv12_fb_new, xd->dst.y_buffer + 16,
-                          xd->dst.u_buffer + 8, xd->dst.v_buffer + 8);
+    /* adjust to the next row of mbs */
+    vp8_extend_mb_row(yv12_fb_new, xd->dst.y_buffer + 16, xd->dst.u_buffer + 8,
+                      xd->dst.v_buffer + 8);
 
-        ++xd->mode_info_context;      /* skip prediction column */
-        xd->up_available = 1;
+    ++xd->mode_info_context; /* skip prediction column */
+    xd->up_available = 1;
 
-        if(pc->filter_level)
-        {
-            if(mb_row > 0)
-            {
-                if (pc->filter_type == NORMAL_LOOPFILTER)
-                    vp8_loop_filter_row_normal(pc, lf_mic, mb_row-1,
-                                               recon_y_stride, recon_uv_stride,
-                                               lf_dst[0], lf_dst[1], lf_dst[2]);
-                else
-                    vp8_loop_filter_row_simple(pc, lf_mic, mb_row-1,
-                                               recon_y_stride, recon_uv_stride,
-                                               lf_dst[0], lf_dst[1], lf_dst[2]);
-                if(mb_row > 1)
-                {
-                    yv12_extend_frame_left_right_c(yv12_fb_new,
-                                                   eb_dst[0],
-                                                   eb_dst[1],
-                                                   eb_dst[2]);
-
-                    eb_dst[0] += recon_y_stride  * 16;
-                    eb_dst[1] += recon_uv_stride *  8;
-                    eb_dst[2] += recon_uv_stride *  8;
-                }
-
-                lf_dst[0] += recon_y_stride  * 16;
-                lf_dst[1] += recon_uv_stride *  8;
-                lf_dst[2] += recon_uv_stride *  8;
-                lf_mic += pc->mb_cols;
-                lf_mic++;         /* Skip border mb */
-            }
-        }
-        else
-        {
-            if(mb_row > 0)
-            {
-                /**/
-                yv12_extend_frame_left_right_c(yv12_fb_new,
-                                               eb_dst[0],
-                                               eb_dst[1],
-                                               eb_dst[2]);
-                eb_dst[0] += recon_y_stride  * 16;
-                eb_dst[1] += recon_uv_stride *  8;
-                eb_dst[2] += recon_uv_stride *  8;
-            }
-        }
-    }
-
-    if(pc->filter_level)
-    {
+    if (pc->filter_level) {
+      if (mb_row > 0) {
         if (pc->filter_type == NORMAL_LOOPFILTER)
-            vp8_loop_filter_row_normal(pc, lf_mic, mb_row-1, recon_y_stride,
-                                       recon_uv_stride, lf_dst[0], lf_dst[1],
-                                       lf_dst[2]);
+          vp8_loop_filter_row_normal(pc, lf_mic, mb_row - 1, recon_y_stride,
+                                     recon_uv_stride, lf_dst[0], lf_dst[1],
+                                     lf_dst[2]);
         else
-            vp8_loop_filter_row_simple(pc, lf_mic, mb_row-1, recon_y_stride,
-                                       recon_uv_stride, lf_dst[0], lf_dst[1],
-                                       lf_dst[2]);
+          vp8_loop_filter_row_simple(pc, lf_mic, mb_row - 1, recon_y_stride,
+                                     recon_uv_stride, lf_dst[0], lf_dst[1],
+                                     lf_dst[2]);
+        if (mb_row > 1) {
+          yv12_extend_frame_left_right_c(yv12_fb_new, eb_dst[0], eb_dst[1],
+                                         eb_dst[2]);
 
-        yv12_extend_frame_left_right_c(yv12_fb_new,
-                                       eb_dst[0],
-                                       eb_dst[1],
+          eb_dst[0] += recon_y_stride * 16;
+          eb_dst[1] += recon_uv_stride * 8;
+          eb_dst[2] += recon_uv_stride * 8;
+        }
+
+        lf_dst[0] += recon_y_stride * 16;
+        lf_dst[1] += recon_uv_stride * 8;
+        lf_dst[2] += recon_uv_stride * 8;
+        lf_mic += pc->mb_cols;
+        lf_mic++; /* Skip border mb */
+      }
+    } else {
+      if (mb_row > 0) {
+        /**/
+        yv12_extend_frame_left_right_c(yv12_fb_new, eb_dst[0], eb_dst[1],
                                        eb_dst[2]);
-        eb_dst[0] += recon_y_stride  * 16;
-        eb_dst[1] += recon_uv_stride *  8;
-        eb_dst[2] += recon_uv_stride *  8;
+        eb_dst[0] += recon_y_stride * 16;
+        eb_dst[1] += recon_uv_stride * 8;
+        eb_dst[2] += recon_uv_stride * 8;
+      }
     }
-    yv12_extend_frame_left_right_c(yv12_fb_new,
-                                   eb_dst[0],
-                                   eb_dst[1],
-                                   eb_dst[2]);
-    yv12_extend_frame_top_c(yv12_fb_new);
-    yv12_extend_frame_bottom_c(yv12_fb_new);
+  }
 
+  if (pc->filter_level) {
+    if (pc->filter_type == NORMAL_LOOPFILTER)
+      vp8_loop_filter_row_normal(pc, lf_mic, mb_row - 1, recon_y_stride,
+                                 recon_uv_stride, lf_dst[0], lf_dst[1],
+                                 lf_dst[2]);
+    else
+      vp8_loop_filter_row_simple(pc, lf_mic, mb_row - 1, recon_y_stride,
+                                 recon_uv_stride, lf_dst[0], lf_dst[1],
+                                 lf_dst[2]);
+
+    yv12_extend_frame_left_right_c(yv12_fb_new, eb_dst[0], eb_dst[1],
+                                   eb_dst[2]);
+    eb_dst[0] += recon_y_stride * 16;
+    eb_dst[1] += recon_uv_stride * 8;
+    eb_dst[2] += recon_uv_stride * 8;
+  }
+  yv12_extend_frame_left_right_c(yv12_fb_new, eb_dst[0], eb_dst[1], eb_dst[2]);
+  yv12_extend_frame_top_c(yv12_fb_new);
+  yv12_extend_frame_bottom_c(yv12_fb_new);
 }
 
 static unsigned int read_partition_size(VP8D_COMP *pbi,
-                                        const unsigned char *cx_size)
-{
-    unsigned char temp[3];
-    if (pbi->decrypt_cb)
-    {
-        pbi->decrypt_cb(pbi->decrypt_state, cx_size, temp, 3);
-        cx_size = temp;
-    }
-    return cx_size[0] + (cx_size[1] << 8) + (cx_size[2] << 16);
+                                        const unsigned char *cx_size) {
+  unsigned char temp[3];
+  if (pbi->decrypt_cb) {
+    pbi->decrypt_cb(pbi->decrypt_state, cx_size, temp, 3);
+    cx_size = temp;
+  }
+  return cx_size[0] + (cx_size[1] << 8) + (cx_size[2] << 16);
 }
 
-static int read_is_valid(const unsigned char *start,
-                         size_t               len,
-                         const unsigned char *end)
-{
-    return (start + len > start && start + len <= end);
+static int read_is_valid(const unsigned char *start, size_t len,
+                         const unsigned char *end) {
+  return (start + len > start && start + len <= end);
 }
 
 static unsigned int read_available_partition_size(
-                                       VP8D_COMP *pbi,
-                                       const unsigned char *token_part_sizes,
-                                       const unsigned char *fragment_start,
-                                       const unsigned char *first_fragment_end,
-                                       const unsigned char *fragment_end,
-                                       int i,
-                                       int num_part)
-{
-    VP8_COMMON* pc = &pbi->common;
-    const unsigned char *partition_size_ptr = token_part_sizes + i * 3;
-    unsigned int partition_size = 0;
-    ptrdiff_t bytes_left = fragment_end - fragment_start;
-    /* Calculate the length of this partition. The last partition
-     * size is implicit. If the partition size can't be read, then
-     * either use the remaining data in the buffer (for EC mode)
-     * or throw an error.
-     */
-    if (i < num_part - 1)
-    {
-        if (read_is_valid(partition_size_ptr, 3, first_fragment_end))
-            partition_size = read_partition_size(pbi, partition_size_ptr);
-        else if (pbi->ec_active)
-            partition_size = (unsigned int)bytes_left;
-        else
-            vpx_internal_error(&pc->error, VPX_CODEC_CORRUPT_FRAME,
-                               "Truncated partition size data");
-    }
+    VP8D_COMP *pbi, const unsigned char *token_part_sizes,
+    const unsigned char *fragment_start,
+    const unsigned char *first_fragment_end, const unsigned char *fragment_end,
+    int i, int num_part) {
+  VP8_COMMON *pc = &pbi->common;
+  const unsigned char *partition_size_ptr = token_part_sizes + i * 3;
+  unsigned int partition_size = 0;
+  ptrdiff_t bytes_left = fragment_end - fragment_start;
+  /* Calculate the length of this partition. The last partition
+   * size is implicit. If the partition size can't be read, then
+   * either use the remaining data in the buffer (for EC mode)
+   * or throw an error.
+   */
+  if (i < num_part - 1) {
+    if (read_is_valid(partition_size_ptr, 3, first_fragment_end))
+      partition_size = read_partition_size(pbi, partition_size_ptr);
+    else if (pbi->ec_active)
+      partition_size = (unsigned int)bytes_left;
     else
-        partition_size = (unsigned int)bytes_left;
+      vpx_internal_error(&pc->error, VPX_CODEC_CORRUPT_FRAME,
+                         "Truncated partition size data");
+  } else
+    partition_size = (unsigned int)bytes_left;
 
-    /* Validate the calculated partition length. If the buffer
-     * described by the partition can't be fully read, then restrict
-     * it to the portion that can be (for EC mode) or throw an error.
-     */
-    if (!read_is_valid(fragment_start, partition_size, fragment_end))
-    {
-        if (pbi->ec_active)
-            partition_size = (unsigned int)bytes_left;
-        else
-            vpx_internal_error(&pc->error, VPX_CODEC_CORRUPT_FRAME,
-                               "Truncated packet or corrupt partition "
-                               "%d length", i + 1);
-    }
-    return partition_size;
+  /* Validate the calculated partition length. If the buffer
+   * described by the partition can't be fully read, then restrict
+   * it to the portion that can be (for EC mode) or throw an error.
+   */
+  if (!read_is_valid(fragment_start, partition_size, fragment_end)) {
+    if (pbi->ec_active)
+      partition_size = (unsigned int)bytes_left;
+    else
+      vpx_internal_error(&pc->error, VPX_CODEC_CORRUPT_FRAME,
+                         "Truncated packet or corrupt partition "
+                         "%d length",
+                         i + 1);
+  }
+  return partition_size;
 }
 
-
 static void setup_token_decoder(VP8D_COMP *pbi,
-                                const unsigned char* token_part_sizes)
-{
-    vp8_reader *bool_decoder = &pbi->mbc[0];
-    unsigned int partition_idx;
-    unsigned int fragment_idx;
-    unsigned int num_token_partitions;
-    const unsigned char *first_fragment_end = pbi->fragments.ptrs[0] +
-                                          pbi->fragments.sizes[0];
+                                const unsigned char *token_part_sizes) {
+  vp8_reader *bool_decoder = &pbi->mbc[0];
+  unsigned int partition_idx;
+  unsigned int fragment_idx;
+  unsigned int num_token_partitions;
+  const unsigned char *first_fragment_end =
+      pbi->fragments.ptrs[0] + pbi->fragments.sizes[0];
 
-    TOKEN_PARTITION multi_token_partition =
-            (TOKEN_PARTITION)vp8_read_literal(&pbi->mbc[8], 2);
-    if (!vp8dx_bool_error(&pbi->mbc[8]))
-        pbi->common.multi_token_partition = multi_token_partition;
-    num_token_partitions = 1 << pbi->common.multi_token_partition;
+  TOKEN_PARTITION multi_token_partition =
+      (TOKEN_PARTITION)vp8_read_literal(&pbi->mbc[8], 2);
+  if (!vp8dx_bool_error(&pbi->mbc[8]))
+    pbi->common.multi_token_partition = multi_token_partition;
+  num_token_partitions = 1 << pbi->common.multi_token_partition;
 
-    /* Check for partitions within the fragments and unpack the fragments
-     * so that each fragment pointer points to its corresponding partition. */
-    for (fragment_idx = 0; fragment_idx < pbi->fragments.count; ++fragment_idx)
-    {
-        unsigned int fragment_size = pbi->fragments.sizes[fragment_idx];
-        const unsigned char *fragment_end = pbi->fragments.ptrs[fragment_idx] +
-                                            fragment_size;
-        /* Special case for handling the first partition since we have already
-         * read its size. */
-        if (fragment_idx == 0)
-        {
-            /* Size of first partition + token partition sizes element */
-            ptrdiff_t ext_first_part_size = token_part_sizes -
-                pbi->fragments.ptrs[0] + 3 * (num_token_partitions - 1);
-            fragment_size -= (unsigned int)ext_first_part_size;
-            if (fragment_size > 0)
-            {
-                pbi->fragments.sizes[0] = (unsigned int)ext_first_part_size;
-                /* The fragment contains an additional partition. Move to
-                 * next. */
-                fragment_idx++;
-                pbi->fragments.ptrs[fragment_idx] = pbi->fragments.ptrs[0] +
-                  pbi->fragments.sizes[0];
-            }
-        }
-        /* Split the chunk into partitions read from the bitstream */
-        while (fragment_size > 0)
-        {
-            ptrdiff_t partition_size = read_available_partition_size(
-                                                 pbi,
-                                                 token_part_sizes,
-                                                 pbi->fragments.ptrs[fragment_idx],
-                                                 first_fragment_end,
-                                                 fragment_end,
-                                                 fragment_idx - 1,
-                                                 num_token_partitions);
-            pbi->fragments.sizes[fragment_idx] = (unsigned int)partition_size;
-            fragment_size -= (unsigned int)partition_size;
-            assert(fragment_idx <= num_token_partitions);
-            if (fragment_size > 0)
-            {
-                /* The fragment contains an additional partition.
-                 * Move to next. */
-                fragment_idx++;
-                pbi->fragments.ptrs[fragment_idx] =
-                    pbi->fragments.ptrs[fragment_idx - 1] + partition_size;
-            }
-        }
+  /* Check for partitions within the fragments and unpack the fragments
+   * so that each fragment pointer points to its corresponding partition. */
+  for (fragment_idx = 0; fragment_idx < pbi->fragments.count; ++fragment_idx) {
+    unsigned int fragment_size = pbi->fragments.sizes[fragment_idx];
+    const unsigned char *fragment_end =
+        pbi->fragments.ptrs[fragment_idx] + fragment_size;
+    /* Special case for handling the first partition since we have already
+     * read its size. */
+    if (fragment_idx == 0) {
+      /* Size of first partition + token partition sizes element */
+      ptrdiff_t ext_first_part_size = token_part_sizes -
+                                      pbi->fragments.ptrs[0] +
+                                      3 * (num_token_partitions - 1);
+      fragment_size -= (unsigned int)ext_first_part_size;
+      if (fragment_size > 0) {
+        pbi->fragments.sizes[0] = (unsigned int)ext_first_part_size;
+        /* The fragment contains an additional partition. Move to
+         * next. */
+        fragment_idx++;
+        pbi->fragments.ptrs[fragment_idx] =
+            pbi->fragments.ptrs[0] + pbi->fragments.sizes[0];
+      }
     }
+    /* Split the chunk into partitions read from the bitstream */
+    while (fragment_size > 0) {
+      ptrdiff_t partition_size = read_available_partition_size(
+          pbi, token_part_sizes, pbi->fragments.ptrs[fragment_idx],
+          first_fragment_end, fragment_end, fragment_idx - 1,
+          num_token_partitions);
+      pbi->fragments.sizes[fragment_idx] = (unsigned int)partition_size;
+      fragment_size -= (unsigned int)partition_size;
+      assert(fragment_idx <= num_token_partitions);
+      if (fragment_size > 0) {
+        /* The fragment contains an additional partition.
+         * Move to next. */
+        fragment_idx++;
+        pbi->fragments.ptrs[fragment_idx] =
+            pbi->fragments.ptrs[fragment_idx - 1] + partition_size;
+      }
+    }
+  }
 
-    pbi->fragments.count = num_token_partitions + 1;
+  pbi->fragments.count = num_token_partitions + 1;
 
-    for (partition_idx = 1; partition_idx < pbi->fragments.count; ++partition_idx)
-    {
-        if (vp8dx_start_decode(bool_decoder,
-                               pbi->fragments.ptrs[partition_idx],
-                               pbi->fragments.sizes[partition_idx],
-                               pbi->decrypt_cb, pbi->decrypt_state))
-            vpx_internal_error(&pbi->common.error, VPX_CODEC_MEM_ERROR,
-                               "Failed to allocate bool decoder %d",
-                               partition_idx);
+  for (partition_idx = 1; partition_idx < pbi->fragments.count;
+       ++partition_idx) {
+    if (vp8dx_start_decode(bool_decoder, pbi->fragments.ptrs[partition_idx],
+                           pbi->fragments.sizes[partition_idx], pbi->decrypt_cb,
+                           pbi->decrypt_state))
+      vpx_internal_error(&pbi->common.error, VPX_CODEC_MEM_ERROR,
+                         "Failed to allocate bool decoder %d", partition_idx);
 
-        bool_decoder++;
-    }
+    bool_decoder++;
+  }
 
 #if CONFIG_MULTITHREAD
-    /* Clamp number of decoder threads */
-    if (pbi->decoding_thread_count > num_token_partitions - 1)
-        pbi->decoding_thread_count = num_token_partitions - 1;
+  /* Clamp number of decoder threads */
+  if (pbi->decoding_thread_count > num_token_partitions - 1)
+    pbi->decoding_thread_count = num_token_partitions - 1;
 #endif
 }
 
+static void init_frame(VP8D_COMP *pbi) {
+  VP8_COMMON *const pc = &pbi->common;
+  MACROBLOCKD *const xd = &pbi->mb;
 
-static void init_frame(VP8D_COMP *pbi)
-{
-    VP8_COMMON *const pc = & pbi->common;
-    MACROBLOCKD *const xd  = & pbi->mb;
+  if (pc->frame_type == KEY_FRAME) {
+    /* Various keyframe initializations */
+    memcpy(pc->fc.mvc, vp8_default_mv_context, sizeof(vp8_default_mv_context));
 
-    if (pc->frame_type == KEY_FRAME)
-    {
-        /* Various keyframe initializations */
-        memcpy(pc->fc.mvc, vp8_default_mv_context, sizeof(vp8_default_mv_context));
+    vp8_init_mbmode_probs(pc);
 
-        vp8_init_mbmode_probs(pc);
+    vp8_default_coef_probs(pc);
 
-        vp8_default_coef_probs(pc);
+    /* reset the segment feature data to 0 with delta coding (Default state). */
+    memset(xd->segment_feature_data, 0, sizeof(xd->segment_feature_data));
+    xd->mb_segement_abs_delta = SEGMENT_DELTADATA;
 
-        /* reset the segment feature data to 0 with delta coding (Default state). */
-        memset(xd->segment_feature_data, 0, sizeof(xd->segment_feature_data));
-        xd->mb_segement_abs_delta = SEGMENT_DELTADATA;
+    /* reset the mode ref deltasa for loop filter */
+    memset(xd->ref_lf_deltas, 0, sizeof(xd->ref_lf_deltas));
+    memset(xd->mode_lf_deltas, 0, sizeof(xd->mode_lf_deltas));
 
-        /* reset the mode ref deltasa for loop filter */
-        memset(xd->ref_lf_deltas, 0, sizeof(xd->ref_lf_deltas));
-        memset(xd->mode_lf_deltas, 0, sizeof(xd->mode_lf_deltas));
+    /* All buffers are implicitly updated on key frames. */
+    pc->refresh_golden_frame = 1;
+    pc->refresh_alt_ref_frame = 1;
+    pc->copy_buffer_to_gf = 0;
+    pc->copy_buffer_to_arf = 0;
 
-        /* All buffers are implicitly updated on key frames. */
-        pc->refresh_golden_frame = 1;
-        pc->refresh_alt_ref_frame = 1;
-        pc->copy_buffer_to_gf = 0;
-        pc->copy_buffer_to_arf = 0;
-
-        /* Note that Golden and Altref modes cannot be used on a key frame so
-         * ref_frame_sign_bias[] is undefined and meaningless
-         */
-        pc->ref_frame_sign_bias[GOLDEN_FRAME] = 0;
-        pc->ref_frame_sign_bias[ALTREF_FRAME] = 0;
+    /* Note that Golden and Altref modes cannot be used on a key frame so
+     * ref_frame_sign_bias[] is undefined and meaningless
+     */
+    pc->ref_frame_sign_bias[GOLDEN_FRAME] = 0;
+    pc->ref_frame_sign_bias[ALTREF_FRAME] = 0;
+  } else {
+    /* To enable choice of different interploation filters */
+    if (!pc->use_bilinear_mc_filter) {
+      xd->subpixel_predict = vp8_sixtap_predict4x4;
+      xd->subpixel_predict8x4 = vp8_sixtap_predict8x4;
+      xd->subpixel_predict8x8 = vp8_sixtap_predict8x8;
+      xd->subpixel_predict16x16 = vp8_sixtap_predict16x16;
+    } else {
+      xd->subpixel_predict = vp8_bilinear_predict4x4;
+      xd->subpixel_predict8x4 = vp8_bilinear_predict8x4;
+      xd->subpixel_predict8x8 = vp8_bilinear_predict8x8;
+      xd->subpixel_predict16x16 = vp8_bilinear_predict16x16;
     }
-    else
-    {
-        /* To enable choice of different interploation filters */
-        if (!pc->use_bilinear_mc_filter)
-        {
-            xd->subpixel_predict        = vp8_sixtap_predict4x4;
-            xd->subpixel_predict8x4     = vp8_sixtap_predict8x4;
-            xd->subpixel_predict8x8     = vp8_sixtap_predict8x8;
-            xd->subpixel_predict16x16   = vp8_sixtap_predict16x16;
-        }
-        else
-        {
-            xd->subpixel_predict        = vp8_bilinear_predict4x4;
-            xd->subpixel_predict8x4     = vp8_bilinear_predict8x4;
-            xd->subpixel_predict8x8     = vp8_bilinear_predict8x8;
-            xd->subpixel_predict16x16   = vp8_bilinear_predict16x16;
-        }
 
-        if (pbi->decoded_key_frame && pbi->ec_enabled && !pbi->ec_active)
-            pbi->ec_active = 1;
-    }
+    if (pbi->decoded_key_frame && pbi->ec_enabled && !pbi->ec_active)
+      pbi->ec_active = 1;
+  }
 
-    xd->left_context = &pc->left_context;
-    xd->mode_info_context = pc->mi;
-    xd->frame_type = pc->frame_type;
-    xd->mode_info_context->mbmi.mode = DC_PRED;
-    xd->mode_info_stride = pc->mode_info_stride;
-    xd->corrupted = 0; /* init without corruption */
+  xd->left_context = &pc->left_context;
+  xd->mode_info_context = pc->mi;
+  xd->frame_type = pc->frame_type;
+  xd->mode_info_context->mbmi.mode = DC_PRED;
+  xd->mode_info_stride = pc->mode_info_stride;
+  xd->corrupted = 0; /* init without corruption */
 
-    xd->fullpixel_mask = 0xffffffff;
-    if(pc->full_pixel)
-        xd->fullpixel_mask = 0xfffffff8;
-
+  xd->fullpixel_mask = 0xffffffff;
+  if (pc->full_pixel) xd->fullpixel_mask = 0xfffffff8;
 }
 
-int vp8_decode_frame(VP8D_COMP *pbi)
-{
-    vp8_reader *const bc = &pbi->mbc[8];
-    VP8_COMMON *const pc = &pbi->common;
-    MACROBLOCKD *const xd  = &pbi->mb;
-    const unsigned char *data = pbi->fragments.ptrs[0];
-    const unsigned int data_sz = pbi->fragments.sizes[0];
-    const unsigned char *data_end = data + data_sz;
-    ptrdiff_t first_partition_length_in_bytes;
+int vp8_decode_frame(VP8D_COMP *pbi) {
+  vp8_reader *const bc = &pbi->mbc[8];
+  VP8_COMMON *const pc = &pbi->common;
+  MACROBLOCKD *const xd = &pbi->mb;
+  const unsigned char *data = pbi->fragments.ptrs[0];
+  const unsigned int data_sz = pbi->fragments.sizes[0];
+  const unsigned char *data_end = data + data_sz;
+  ptrdiff_t first_partition_length_in_bytes;
 
-    int i, j, k, l;
-    const int *const mb_feature_data_bits = vp8_mb_feature_data_bits;
-    int corrupt_tokens = 0;
-    int prev_independent_partitions = pbi->independent_partitions;
+  int i, j, k, l;
+  const int *const mb_feature_data_bits = vp8_mb_feature_data_bits;
+  int corrupt_tokens = 0;
+  int prev_independent_partitions = pbi->independent_partitions;
 
-    YV12_BUFFER_CONFIG *yv12_fb_new = pbi->dec_fb_ref[INTRA_FRAME];
+  YV12_BUFFER_CONFIG *yv12_fb_new = pbi->dec_fb_ref[INTRA_FRAME];
 
-    /* start with no corruption of current frame */
-    xd->corrupted = 0;
-    yv12_fb_new->corrupted = 0;
+  /* start with no corruption of current frame */
+  xd->corrupted = 0;
+  yv12_fb_new->corrupted = 0;
 
-    if (data_end - data < 3)
-    {
-        if (!pbi->ec_active)
-        {
-            vpx_internal_error(&pc->error, VPX_CODEC_CORRUPT_FRAME,
-                               "Truncated packet");
-        }
+  if (data_end - data < 3) {
+    if (!pbi->ec_active) {
+      vpx_internal_error(&pc->error, VPX_CODEC_CORRUPT_FRAME,
+                         "Truncated packet");
+    }
 
-        /* Declare the missing frame as an inter frame since it will
-           be handled as an inter frame when we have estimated its
-           motion vectors. */
-        pc->frame_type = INTER_FRAME;
-        pc->version = 0;
-        pc->show_frame = 1;
-        first_partition_length_in_bytes = 0;
+    /* Declare the missing frame as an inter frame since it will
+       be handled as an inter frame when we have estimated its
+       motion vectors. */
+    pc->frame_type = INTER_FRAME;
+    pc->version = 0;
+    pc->show_frame = 1;
+    first_partition_length_in_bytes = 0;
+  } else {
+    unsigned char clear_buffer[10];
+    const unsigned char *clear = data;
+    if (pbi->decrypt_cb) {
+      int n = (int)VPXMIN(sizeof(clear_buffer), data_sz);
+      pbi->decrypt_cb(pbi->decrypt_state, data, clear_buffer, n);
+      clear = clear_buffer;
     }
-    else
-    {
-        unsigned char clear_buffer[10];
-        const unsigned char *clear = data;
-        if (pbi->decrypt_cb)
-        {
-            int n = (int)VPXMIN(sizeof(clear_buffer), data_sz);
-            pbi->decrypt_cb(pbi->decrypt_state, data, clear_buffer, n);
-            clear = clear_buffer;
-        }
 
-        pc->frame_type = (FRAME_TYPE)(clear[0] & 1);
-        pc->version = (clear[0] >> 1) & 7;
-        pc->show_frame = (clear[0] >> 4) & 1;
-        first_partition_length_in_bytes =
-            (clear[0] | (clear[1] << 8) | (clear[2] << 16)) >> 5;
+    pc->frame_type = (FRAME_TYPE)(clear[0] & 1);
+    pc->version = (clear[0] >> 1) & 7;
+    pc->show_frame = (clear[0] >> 4) & 1;
+    first_partition_length_in_bytes =
+        (clear[0] | (clear[1] << 8) | (clear[2] << 16)) >> 5;
 
-        if (!pbi->ec_active &&
-            (data + first_partition_length_in_bytes > data_end
-            || data + first_partition_length_in_bytes < data))
-            vpx_internal_error(&pc->error, VPX_CODEC_CORRUPT_FRAME,
-                               "Truncated packet or corrupt partition 0 length");
+    if (!pbi->ec_active && (data + first_partition_length_in_bytes > data_end ||
+                            data + first_partition_length_in_bytes < data))
+      vpx_internal_error(&pc->error, VPX_CODEC_CORRUPT_FRAME,
+                         "Truncated packet or corrupt partition 0 length");
 
-        data += 3;
-        clear += 3;
+    data += 3;
+    clear += 3;
 
-        vp8_setup_version(pc);
+    vp8_setup_version(pc);
 
+    if (pc->frame_type == KEY_FRAME) {
+      /* vet via sync code */
+      /* When error concealment is enabled we should only check the sync
+       * code if we have enough bits available
+       */
+      if (!pbi->ec_active || data + 3 < data_end) {
+        if (clear[0] != 0x9d || clear[1] != 0x01 || clear[2] != 0x2a)
+          vpx_internal_error(&pc->error, VPX_CODEC_UNSUP_BITSTREAM,
+                             "Invalid frame sync code");
+      }
 
-        if (pc->frame_type == KEY_FRAME)
-        {
-            /* vet via sync code */
-            /* When error concealment is enabled we should only check the sync
-             * code if we have enough bits available
-             */
-            if (!pbi->ec_active || data + 3 < data_end)
-            {
-                if (clear[0] != 0x9d || clear[1] != 0x01 || clear[2] != 0x2a)
-                    vpx_internal_error(&pc->error, VPX_CODEC_UNSUP_BITSTREAM,
-                                   "Invalid frame sync code");
-            }
-
-            /* If error concealment is enabled we should only parse the new size
-             * if we have enough data. Otherwise we will end up with the wrong
-             * size.
-             */
-            if (!pbi->ec_active || data + 6 < data_end)
-            {
-                pc->Width = (clear[3] | (clear[4] << 8)) & 0x3fff;
-                pc->horiz_scale = clear[4] >> 6;
-                pc->Height = (clear[5] | (clear[6] << 8)) & 0x3fff;
-                pc->vert_scale = clear[6] >> 6;
-            }
-            data += 7;
-        }
-        else
-        {
-          memcpy(&xd->pre, yv12_fb_new, sizeof(YV12_BUFFER_CONFIG));
-          memcpy(&xd->dst, yv12_fb_new, sizeof(YV12_BUFFER_CONFIG));
-        }
+      /* If error concealment is enabled we should only parse the new size
+       * if we have enough data. Otherwise we will end up with the wrong
+       * size.
+       */
+      if (!pbi->ec_active || data + 6 < data_end) {
+        pc->Width = (clear[3] | (clear[4] << 8)) & 0x3fff;
+        pc->horiz_scale = clear[4] >> 6;
+        pc->Height = (clear[5] | (clear[6] << 8)) & 0x3fff;
+        pc->vert_scale = clear[6] >> 6;
+      }
+      data += 7;
+    } else {
+      memcpy(&xd->pre, yv12_fb_new, sizeof(YV12_BUFFER_CONFIG));
+      memcpy(&xd->dst, yv12_fb_new, sizeof(YV12_BUFFER_CONFIG));
     }
-    if ((!pbi->decoded_key_frame && pc->frame_type != KEY_FRAME))
-    {
-        return -1;
-    }
+  }
+  if ((!pbi->decoded_key_frame && pc->frame_type != KEY_FRAME)) {
+    return -1;
+  }
 
-    init_frame(pbi);
+  init_frame(pbi);
 
-    if (vp8dx_start_decode(bc, data, (unsigned int)(data_end - data),
-                           pbi->decrypt_cb, pbi->decrypt_state))
-        vpx_internal_error(&pc->error, VPX_CODEC_MEM_ERROR,
-                           "Failed to allocate bool decoder 0");
-    if (pc->frame_type == KEY_FRAME) {
-        (void)vp8_read_bit(bc);  // colorspace
-        pc->clamp_type  = (CLAMP_TYPE)vp8_read_bit(bc);
-    }
+  if (vp8dx_start_decode(bc, data, (unsigned int)(data_end - data),
+                         pbi->decrypt_cb, pbi->decrypt_state))
+    vpx_internal_error(&pc->error, VPX_CODEC_MEM_ERROR,
+                       "Failed to allocate bool decoder 0");
+  if (pc->frame_type == KEY_FRAME) {
+    (void)vp8_read_bit(bc);  // colorspace
+    pc->clamp_type = (CLAMP_TYPE)vp8_read_bit(bc);
+  }
 
-    /* Is segmentation enabled */
-    xd->segmentation_enabled = (unsigned char)vp8_read_bit(bc);
+  /* Is segmentation enabled */
+  xd->segmentation_enabled = (unsigned char)vp8_read_bit(bc);
 
-    if (xd->segmentation_enabled)
-    {
-        /* Signal whether or not the segmentation map is being explicitly updated this frame. */
-        xd->update_mb_segmentation_map = (unsigned char)vp8_read_bit(bc);
-        xd->update_mb_segmentation_data = (unsigned char)vp8_read_bit(bc);
+  if (xd->segmentation_enabled) {
+    /* Signal whether or not the segmentation map is being explicitly updated
+     * this frame. */
+    xd->update_mb_segmentation_map = (unsigned char)vp8_read_bit(bc);
+    xd->update_mb_segmentation_data = (unsigned char)vp8_read_bit(bc);
 
-        if (xd->update_mb_segmentation_data)
-        {
-            xd->mb_segement_abs_delta = (unsigned char)vp8_read_bit(bc);
+    if (xd->update_mb_segmentation_data) {
+      xd->mb_segement_abs_delta = (unsigned char)vp8_read_bit(bc);
 
-            memset(xd->segment_feature_data, 0, sizeof(xd->segment_feature_data));
+      memset(xd->segment_feature_data, 0, sizeof(xd->segment_feature_data));
 
-            /* For each segmentation feature (Quant and loop filter level) */
-            for (i = 0; i < MB_LVL_MAX; i++)
-            {
-                for (j = 0; j < MAX_MB_SEGMENTS; j++)
-                {
-                    /* Frame level data */
-                    if (vp8_read_bit(bc))
-                    {
-                        xd->segment_feature_data[i][j] = (signed char)vp8_read_literal(bc, mb_feature_data_bits[i]);
+      /* For each segmentation feature (Quant and loop filter level) */
+      for (i = 0; i < MB_LVL_MAX; i++) {
+        for (j = 0; j < MAX_MB_SEGMENTS; j++) {
+          /* Frame level data */
+          if (vp8_read_bit(bc)) {
+            xd->segment_feature_data[i][j] =
+                (signed char)vp8_read_literal(bc, mb_feature_data_bits[i]);
 
-                        if (vp8_read_bit(bc))
-                            xd->segment_feature_data[i][j] = -xd->segment_feature_data[i][j];
-                    }
-                    else
-                        xd->segment_feature_data[i][j] = 0;
-                }
-            }
+            if (vp8_read_bit(bc))
+              xd->segment_feature_data[i][j] = -xd->segment_feature_data[i][j];
+          } else
+            xd->segment_feature_data[i][j] = 0;
         }
+      }
+    }
 
-        if (xd->update_mb_segmentation_map)
-        {
-            /* Which macro block level features are enabled */
-            memset(xd->mb_segment_tree_probs, 255, sizeof(xd->mb_segment_tree_probs));
+    if (xd->update_mb_segmentation_map) {
+      /* Which macro block level features are enabled */
+      memset(xd->mb_segment_tree_probs, 255, sizeof(xd->mb_segment_tree_probs));
 
-            /* Read the probs used to decode the segment id for each macro block. */
-            for (i = 0; i < MB_FEATURE_TREE_PROBS; i++)
-            {
-                /* If not explicitly set value is defaulted to 255 by memset above */
-                if (vp8_read_bit(bc))
-                    xd->mb_segment_tree_probs[i] = (vp8_prob)vp8_read_literal(bc, 8);
-            }
-        }
+      /* Read the probs used to decode the segment id for each macro block. */
+      for (i = 0; i < MB_FEATURE_TREE_PROBS; i++) {
+        /* If not explicitly set value is defaulted to 255 by memset above */
+        if (vp8_read_bit(bc))
+          xd->mb_segment_tree_probs[i] = (vp8_prob)vp8_read_literal(bc, 8);
+      }
     }
-    else
-    {
-        /* No segmentation updates on this frame */
-        xd->update_mb_segmentation_map = 0;
-        xd->update_mb_segmentation_data = 0;
-    }
+  } else {
+    /* No segmentation updates on this frame */
+    xd->update_mb_segmentation_map = 0;
+    xd->update_mb_segmentation_data = 0;
+  }
 
-    /* Read the loop filter level and type */
-    pc->filter_type = (LOOPFILTERTYPE) vp8_read_bit(bc);
-    pc->filter_level = vp8_read_literal(bc, 6);
-    pc->sharpness_level = vp8_read_literal(bc, 3);
+  /* Read the loop filter level and type */
+  pc->filter_type = (LOOPFILTERTYPE)vp8_read_bit(bc);
+  pc->filter_level = vp8_read_literal(bc, 6);
+  pc->sharpness_level = vp8_read_literal(bc, 3);
 
-    /* Read in loop filter deltas applied at the MB level based on mode or ref frame. */
-    xd->mode_ref_lf_delta_update = 0;
-    xd->mode_ref_lf_delta_enabled = (unsigned char)vp8_read_bit(bc);
+  /* Read in loop filter deltas applied at the MB level based on mode or ref
+   * frame. */
+  xd->mode_ref_lf_delta_update = 0;
+  xd->mode_ref_lf_delta_enabled = (unsigned char)vp8_read_bit(bc);
 
-    if (xd->mode_ref_lf_delta_enabled)
-    {
-        /* Do the deltas need to be updated */
-        xd->mode_ref_lf_delta_update = (unsigned char)vp8_read_bit(bc);
+  if (xd->mode_ref_lf_delta_enabled) {
+    /* Do the deltas need to be updated */
+    xd->mode_ref_lf_delta_update = (unsigned char)vp8_read_bit(bc);
 
-        if (xd->mode_ref_lf_delta_update)
-        {
-            /* Send update */
-            for (i = 0; i < MAX_REF_LF_DELTAS; i++)
-            {
-                if (vp8_read_bit(bc))
-                {
-                    /*sign = vp8_read_bit( bc );*/
-                    xd->ref_lf_deltas[i] = (signed char)vp8_read_literal(bc, 6);
+    if (xd->mode_ref_lf_delta_update) {
+      /* Send update */
+      for (i = 0; i < MAX_REF_LF_DELTAS; i++) {
+        if (vp8_read_bit(bc)) {
+          /*sign = vp8_read_bit( bc );*/
+          xd->ref_lf_deltas[i] = (signed char)vp8_read_literal(bc, 6);
 
-                    if (vp8_read_bit(bc))        /* Apply sign */
-                        xd->ref_lf_deltas[i] = xd->ref_lf_deltas[i] * -1;
-                }
-            }
+          if (vp8_read_bit(bc)) /* Apply sign */
+            xd->ref_lf_deltas[i] = xd->ref_lf_deltas[i] * -1;
+        }
+      }
 
-            /* Send update */
-            for (i = 0; i < MAX_MODE_LF_DELTAS; i++)
-            {
-                if (vp8_read_bit(bc))
-                {
-                    /*sign = vp8_read_bit( bc );*/
-                    xd->mode_lf_deltas[i] = (signed char)vp8_read_literal(bc, 6);
+      /* Send update */
+      for (i = 0; i < MAX_MODE_LF_DELTAS; i++) {
+        if (vp8_read_bit(bc)) {
+          /*sign = vp8_read_bit( bc );*/
+          xd->mode_lf_deltas[i] = (signed char)vp8_read_literal(bc, 6);
 
-                    if (vp8_read_bit(bc))        /* Apply sign */
-                        xd->mode_lf_deltas[i] = xd->mode_lf_deltas[i] * -1;
-                }
-            }
+          if (vp8_read_bit(bc)) /* Apply sign */
+            xd->mode_lf_deltas[i] = xd->mode_lf_deltas[i] * -1;
         }
+      }
     }
+  }
 
-    setup_token_decoder(pbi, data + first_partition_length_in_bytes);
+  setup_token_decoder(pbi, data + first_partition_length_in_bytes);
 
-    xd->current_bc = &pbi->mbc[0];
+  xd->current_bc = &pbi->mbc[0];
 
-    /* Read the default quantizers. */
-    {
-        int Q, q_update;
+  /* Read the default quantizers. */
+  {
+    int Q, q_update;
 
-        Q = vp8_read_literal(bc, 7);  /* AC 1st order Q = default */
-        pc->base_qindex = Q;
-        q_update = 0;
-        pc->y1dc_delta_q = get_delta_q(bc, pc->y1dc_delta_q, &q_update);
-        pc->y2dc_delta_q = get_delta_q(bc, pc->y2dc_delta_q, &q_update);
-        pc->y2ac_delta_q = get_delta_q(bc, pc->y2ac_delta_q, &q_update);
-        pc->uvdc_delta_q = get_delta_q(bc, pc->uvdc_delta_q, &q_update);
-        pc->uvac_delta_q = get_delta_q(bc, pc->uvac_delta_q, &q_update);
+    Q = vp8_read_literal(bc, 7); /* AC 1st order Q = default */
+    pc->base_qindex = Q;
+    q_update = 0;
+    pc->y1dc_delta_q = get_delta_q(bc, pc->y1dc_delta_q, &q_update);
+    pc->y2dc_delta_q = get_delta_q(bc, pc->y2dc_delta_q, &q_update);
+    pc->y2ac_delta_q = get_delta_q(bc, pc->y2ac_delta_q, &q_update);
+    pc->uvdc_delta_q = get_delta_q(bc, pc->uvdc_delta_q, &q_update);
+    pc->uvac_delta_q = get_delta_q(bc, pc->uvac_delta_q, &q_update);
 
-        if (q_update)
-            vp8cx_init_de_quantizer(pbi);
+    if (q_update) vp8cx_init_de_quantizer(pbi);
 
-        /* MB level dequantizer setup */
-        vp8_mb_init_dequantizer(pbi, &pbi->mb);
-    }
+    /* MB level dequantizer setup */
+    vp8_mb_init_dequantizer(pbi, &pbi->mb);
+  }
 
-    /* Determine if the golden frame or ARF buffer should be updated and how.
-     * For all non key frames the GF and ARF refresh flags and sign bias
-     * flags must be set explicitly.
-     */
-    if (pc->frame_type != KEY_FRAME)
-    {
-        /* Should the GF or ARF be updated from the current frame */
-        pc->refresh_golden_frame = vp8_read_bit(bc);
+  /* Determine if the golden frame or ARF buffer should be updated and how.
+   * For all non key frames the GF and ARF refresh flags and sign bias
+   * flags must be set explicitly.
+   */
+  if (pc->frame_type != KEY_FRAME) {
+    /* Should the GF or ARF be updated from the current frame */
+    pc->refresh_golden_frame = vp8_read_bit(bc);
 #if CONFIG_ERROR_CONCEALMENT
-        /* Assume we shouldn't refresh golden if the bit is missing */
-        xd->corrupted |= vp8dx_bool_error(bc);
-        if (pbi->ec_active && xd->corrupted)
-            pc->refresh_golden_frame = 0;
+    /* Assume we shouldn't refresh golden if the bit is missing */
+    xd->corrupted |= vp8dx_bool_error(bc);
+    if (pbi->ec_active && xd->corrupted) pc->refresh_golden_frame = 0;
 #endif
 
-        pc->refresh_alt_ref_frame = vp8_read_bit(bc);
+    pc->refresh_alt_ref_frame = vp8_read_bit(bc);
 #if CONFIG_ERROR_CONCEALMENT
-        /* Assume we shouldn't refresh altref if the bit is missing */
-        xd->corrupted |= vp8dx_bool_error(bc);
-        if (pbi->ec_active && xd->corrupted)
-            pc->refresh_alt_ref_frame = 0;
+    /* Assume we shouldn't refresh altref if the bit is missing */
+    xd->corrupted |= vp8dx_bool_error(bc);
+    if (pbi->ec_active && xd->corrupted) pc->refresh_alt_ref_frame = 0;
 #endif
 
-        /* Buffer to buffer copy flags. */
-        pc->copy_buffer_to_gf = 0;
+    /* Buffer to buffer copy flags. */
+    pc->copy_buffer_to_gf = 0;
 
-        if (!pc->refresh_golden_frame)
-            pc->copy_buffer_to_gf = vp8_read_literal(bc, 2);
+    if (!pc->refresh_golden_frame)
+      pc->copy_buffer_to_gf = vp8_read_literal(bc, 2);
 
 #if CONFIG_ERROR_CONCEALMENT
-        /* Assume we shouldn't copy to the golden if the bit is missing */
-        xd->corrupted |= vp8dx_bool_error(bc);
-        if (pbi->ec_active && xd->corrupted)
-            pc->copy_buffer_to_gf = 0;
+    /* Assume we shouldn't copy to the golden if the bit is missing */
+    xd->corrupted |= vp8dx_bool_error(bc);
+    if (pbi->ec_active && xd->corrupted) pc->copy_buffer_to_gf = 0;
 #endif
 
-        pc->copy_buffer_to_arf = 0;
+    pc->copy_buffer_to_arf = 0;
 
-        if (!pc->refresh_alt_ref_frame)
-            pc->copy_buffer_to_arf = vp8_read_literal(bc, 2);
+    if (!pc->refresh_alt_ref_frame)
+      pc->copy_buffer_to_arf = vp8_read_literal(bc, 2);
 
 #if CONFIG_ERROR_CONCEALMENT
-        /* Assume we shouldn't copy to the alt-ref if the bit is missing */
-        xd->corrupted |= vp8dx_bool_error(bc);
-        if (pbi->ec_active && xd->corrupted)
-            pc->copy_buffer_to_arf = 0;
+    /* Assume we shouldn't copy to the alt-ref if the bit is missing */
+    xd->corrupted |= vp8dx_bool_error(bc);
+    if (pbi->ec_active && xd->corrupted) pc->copy_buffer_to_arf = 0;
 #endif
 
+    pc->ref_frame_sign_bias[GOLDEN_FRAME] = vp8_read_bit(bc);
+    pc->ref_frame_sign_bias[ALTREF_FRAME] = vp8_read_bit(bc);
+  }
 
-        pc->ref_frame_sign_bias[GOLDEN_FRAME] = vp8_read_bit(bc);
-        pc->ref_frame_sign_bias[ALTREF_FRAME] = vp8_read_bit(bc);
-    }
-
-    pc->refresh_entropy_probs = vp8_read_bit(bc);
+  pc->refresh_entropy_probs = vp8_read_bit(bc);
 #if CONFIG_ERROR_CONCEALMENT
-    /* Assume we shouldn't refresh the probabilities if the bit is
-     * missing */
-    xd->corrupted |= vp8dx_bool_error(bc);
-    if (pbi->ec_active && xd->corrupted)
-        pc->refresh_entropy_probs = 0;
+  /* Assume we shouldn't refresh the probabilities if the bit is
+   * missing */
+  xd->corrupted |= vp8dx_bool_error(bc);
+  if (pbi->ec_active && xd->corrupted) pc->refresh_entropy_probs = 0;
 #endif
-    if (pc->refresh_entropy_probs == 0)
-    {
-        memcpy(&pc->lfc, &pc->fc, sizeof(pc->fc));
-    }
+  if (pc->refresh_entropy_probs == 0) {
+    memcpy(&pc->lfc, &pc->fc, sizeof(pc->fc));
+  }
 
-    pc->refresh_last_frame = pc->frame_type == KEY_FRAME  ||  vp8_read_bit(bc);
+  pc->refresh_last_frame = pc->frame_type == KEY_FRAME || vp8_read_bit(bc);
 
 #if CONFIG_ERROR_CONCEALMENT
-    /* Assume we should refresh the last frame if the bit is missing */
-    xd->corrupted |= vp8dx_bool_error(bc);
-    if (pbi->ec_active && xd->corrupted)
-        pc->refresh_last_frame = 1;
+  /* Assume we should refresh the last frame if the bit is missing */
+  xd->corrupted |= vp8dx_bool_error(bc);
+  if (pbi->ec_active && xd->corrupted) pc->refresh_last_frame = 1;
 #endif
 
-    if (0)
-    {
-        FILE *z = fopen("decodestats.stt", "a");
-        fprintf(z, "%6d F:%d,G:%d,A:%d,L:%d,Q:%d\n",
-                pc->current_video_frame,
-                pc->frame_type,
-                pc->refresh_golden_frame,
-                pc->refresh_alt_ref_frame,
-                pc->refresh_last_frame,
-                pc->base_qindex);
-        fclose(z);
-    }
+  if (0) {
+    FILE *z = fopen("decodestats.stt", "a");
+    fprintf(z, "%6d F:%d,G:%d,A:%d,L:%d,Q:%d\n", pc->current_video_frame,
+            pc->frame_type, pc->refresh_golden_frame, pc->refresh_alt_ref_frame,
+            pc->refresh_last_frame, pc->base_qindex);
+    fclose(z);
+  }
 
-    {
-        pbi->independent_partitions = 1;
+  {
+    pbi->independent_partitions = 1;
 
-        /* read coef probability tree */
-        for (i = 0; i < BLOCK_TYPES; i++)
-            for (j = 0; j < COEF_BANDS; j++)
-                for (k = 0; k < PREV_COEF_CONTEXTS; k++)
-                    for (l = 0; l < ENTROPY_NODES; l++)
-                    {
+    /* read coef probability tree */
+    for (i = 0; i < BLOCK_TYPES; i++)
+      for (j = 0; j < COEF_BANDS; j++)
+        for (k = 0; k < PREV_COEF_CONTEXTS; k++)
+          for (l = 0; l < ENTROPY_NODES; l++) {
+            vp8_prob *const p = pc->fc.coef_probs[i][j][k] + l;
 
-                        vp8_prob *const p = pc->fc.coef_probs [i][j][k] + l;
+            if (vp8_read(bc, vp8_coef_update_probs[i][j][k][l])) {
+              *p = (vp8_prob)vp8_read_literal(bc, 8);
+            }
+            if (k > 0 && *p != pc->fc.coef_probs[i][j][k - 1][l])
+              pbi->independent_partitions = 0;
+          }
+  }
 
-                        if (vp8_read(bc, vp8_coef_update_probs [i][j][k][l]))
-                        {
-                            *p = (vp8_prob)vp8_read_literal(bc, 8);
+  /* clear out the coeff buffer */
+  memset(xd->qcoeff, 0, sizeof(xd->qcoeff));
 
-                        }
-                        if (k > 0 && *p != pc->fc.coef_probs[i][j][k-1][l])
-                            pbi->independent_partitions = 0;
+  vp8_decode_mode_mvs(pbi);
 
-                    }
-    }
-
-    /* clear out the coeff buffer */
-    memset(xd->qcoeff, 0, sizeof(xd->qcoeff));
-
-    vp8_decode_mode_mvs(pbi);
-
 #if CONFIG_ERROR_CONCEALMENT
-    if (pbi->ec_active &&
-            pbi->mvs_corrupt_from_mb < (unsigned int)pc->mb_cols * pc->mb_rows)
-    {
-        /* Motion vectors are missing in this frame. We will try to estimate
-         * them and then continue decoding the frame as usual */
-        vp8_estimate_missing_mvs(pbi);
-    }
+  if (pbi->ec_active &&
+      pbi->mvs_corrupt_from_mb < (unsigned int)pc->mb_cols * pc->mb_rows) {
+    /* Motion vectors are missing in this frame. We will try to estimate
+     * them and then continue decoding the frame as usual */
+    vp8_estimate_missing_mvs(pbi);
+  }
 #endif
 
-    memset(pc->above_context, 0, sizeof(ENTROPY_CONTEXT_PLANES) * pc->mb_cols);
-    pbi->frame_corrupt_residual = 0;
+  memset(pc->above_context, 0, sizeof(ENTROPY_CONTEXT_PLANES) * pc->mb_cols);
+  pbi->frame_corrupt_residual = 0;
 
 #if CONFIG_MULTITHREAD
-    if (pbi->b_multithreaded_rd && pc->multi_token_partition != ONE_PARTITION)
-    {
-        unsigned int thread;
-        vp8mt_decode_mb_rows(pbi, xd);
-        vp8_yv12_extend_frame_borders(yv12_fb_new);
-        for (thread = 0; thread < pbi->decoding_thread_count; ++thread)
-            corrupt_tokens |= pbi->mb_row_di[thread].mbd.corrupted;
-    }
-    else
+  if (pbi->b_multithreaded_rd && pc->multi_token_partition != ONE_PARTITION) {
+    unsigned int thread;
+    vp8mt_decode_mb_rows(pbi, xd);
+    vp8_yv12_extend_frame_borders(yv12_fb_new);
+    for (thread = 0; thread < pbi->decoding_thread_count; ++thread)
+      corrupt_tokens |= pbi->mb_row_di[thread].mbd.corrupted;
+  } else
 #endif
-    {
-        decode_mb_rows(pbi);
-        corrupt_tokens |= xd->corrupted;
-    }
+  {
+    decode_mb_rows(pbi);
+    corrupt_tokens |= xd->corrupted;
+  }
 
-    /* Collect information about decoder corruption. */
-    /* 1. Check first boolean decoder for errors. */
-    yv12_fb_new->corrupted = vp8dx_bool_error(bc);
-    /* 2. Check the macroblock information */
-    yv12_fb_new->corrupted |= corrupt_tokens;
+  /* Collect information about decoder corruption. */
+  /* 1. Check first boolean decoder for errors. */
+  yv12_fb_new->corrupted = vp8dx_bool_error(bc);
+  /* 2. Check the macroblock information */
+  yv12_fb_new->corrupted |= corrupt_tokens;
 
-    if (!pbi->decoded_key_frame)
-    {
-        if (pc->frame_type == KEY_FRAME &&
-            !yv12_fb_new->corrupted)
-            pbi->decoded_key_frame = 1;
-        else
-            vpx_internal_error(&pbi->common.error, VPX_CODEC_CORRUPT_FRAME,
-                               "A stream must start with a complete key frame");
-    }
+  if (!pbi->decoded_key_frame) {
+    if (pc->frame_type == KEY_FRAME && !yv12_fb_new->corrupted)
+      pbi->decoded_key_frame = 1;
+    else
+      vpx_internal_error(&pbi->common.error, VPX_CODEC_CORRUPT_FRAME,
+                         "A stream must start with a complete key frame");
+  }
 
-    /* vpx_log("Decoder: Frame Decoded, Size Roughly:%d bytes  \n",bc->pos+pbi->bc2.pos); */
+  /* vpx_log("Decoder: Frame Decoded, Size Roughly:%d bytes
+   * \n",bc->pos+pbi->bc2.pos); */
 
-    if (pc->refresh_entropy_probs == 0)
-    {
-        memcpy(&pc->fc, &pc->lfc, sizeof(pc->fc));
-        pbi->independent_partitions = prev_independent_partitions;
-    }
+  if (pc->refresh_entropy_probs == 0) {
+    memcpy(&pc->fc, &pc->lfc, sizeof(pc->fc));
+    pbi->independent_partitions = prev_independent_partitions;
+  }
 
 #ifdef PACKET_TESTING
-    {
-        FILE *f = fopen("decompressor.VP8", "ab");
-        unsigned int size = pbi->bc2.pos + pbi->bc.pos + 8;
-        fwrite((void *) &size, 4, 1, f);
-        fwrite((void *) pbi->Source, size, 1, f);
-        fclose(f);
-    }
+  {
+    FILE *f = fopen("decompressor.VP8", "ab");
+    unsigned int size = pbi->bc2.pos + pbi->bc.pos + 8;
+    fwrite((void *)&size, 4, 1, f);
+    fwrite((void *)pbi->Source, size, 1, f);
+    fclose(f);
+  }
 #endif
 
-    return 0;
+  return 0;
 }
--- a/vp8/decoder/decodemv.c
+++ b/vp8/decoder/decodemv.c
@@ -8,7 +8,6 @@
  *  be found in the AUTHORS file in the root of the source tree.
  */
 
-
 #include "treereader.h"
 #include "vp8/common/entropymv.h"
 #include "vp8/common/entropymode.h"
@@ -18,653 +17,540 @@
 #if CONFIG_DEBUG
 #include <assert.h>
 #endif
-static B_PREDICTION_MODE read_bmode(vp8_reader *bc, const vp8_prob *p)
-{
-    const int i = vp8_treed_read(bc, vp8_bmode_tree, p);
+static B_PREDICTION_MODE read_bmode(vp8_reader *bc, const vp8_prob *p) {
+  const int i = vp8_treed_read(bc, vp8_bmode_tree, p);
 
-    return (B_PREDICTION_MODE)i;
+  return (B_PREDICTION_MODE)i;
 }
 
-static MB_PREDICTION_MODE read_ymode(vp8_reader *bc, const vp8_prob *p)
-{
-    const int i = vp8_treed_read(bc, vp8_ymode_tree, p);
+static MB_PREDICTION_MODE read_ymode(vp8_reader *bc, const vp8_prob *p) {
+  const int i = vp8_treed_read(bc, vp8_ymode_tree, p);
 
-    return (MB_PREDICTION_MODE)i;
+  return (MB_PREDICTION_MODE)i;
 }
 
-static MB_PREDICTION_MODE read_kf_ymode(vp8_reader *bc, const vp8_prob *p)
-{
-    const int i = vp8_treed_read(bc, vp8_kf_ymode_tree, p);
+static MB_PREDICTION_MODE read_kf_ymode(vp8_reader *bc, const vp8_prob *p) {
+  const int i = vp8_treed_read(bc, vp8_kf_ymode_tree, p);
 
-    return (MB_PREDICTION_MODE)i;
+  return (MB_PREDICTION_MODE)i;
 }
 
-static MB_PREDICTION_MODE read_uv_mode(vp8_reader *bc, const vp8_prob *p)
-{
-    const int i = vp8_treed_read(bc, vp8_uv_mode_tree, p);
+static MB_PREDICTION_MODE read_uv_mode(vp8_reader *bc, const vp8_prob *p) {
+  const int i = vp8_treed_read(bc, vp8_uv_mode_tree, p);
 
-    return (MB_PREDICTION_MODE)i;
+  return (MB_PREDICTION_MODE)i;
 }
 
-static void read_kf_modes(VP8D_COMP *pbi, MODE_INFO *mi)
-{
-    vp8_reader *const bc = & pbi->mbc[8];
-    const int mis = pbi->common.mode_info_stride;
+static void read_kf_modes(VP8D_COMP *pbi, MODE_INFO *mi) {
+  vp8_reader *const bc = &pbi->mbc[8];
+  const int mis = pbi->common.mode_info_stride;
 
-    mi->mbmi.ref_frame = INTRA_FRAME;
-    mi->mbmi.mode = read_kf_ymode(bc, vp8_kf_ymode_prob);
+  mi->mbmi.ref_frame = INTRA_FRAME;
+  mi->mbmi.mode = read_kf_ymode(bc, vp8_kf_ymode_prob);
 
-    if (mi->mbmi.mode == B_PRED)
-    {
-        int i = 0;
-        mi->mbmi.is_4x4 = 1;
+  if (mi->mbmi.mode == B_PRED) {
+    int i = 0;
+    mi->mbmi.is_4x4 = 1;
 
-        do
-        {
-            const B_PREDICTION_MODE A = above_block_mode(mi, i, mis);
-            const B_PREDICTION_MODE L = left_block_mode(mi, i);
+    do {
+      const B_PREDICTION_MODE A = above_block_mode(mi, i, mis);
+      const B_PREDICTION_MODE L = left_block_mode(mi, i);
 
-            mi->bmi[i].as_mode =
-                read_bmode(bc, vp8_kf_bmode_prob [A] [L]);
-        }
-        while (++i < 16);
-    }
+      mi->bmi[i].as_mode = read_bmode(bc, vp8_kf_bmode_prob[A][L]);
+    } while (++i < 16);
+  }
 
-    mi->mbmi.uv_mode = read_uv_mode(bc, vp8_kf_uv_mode_prob);
+  mi->mbmi.uv_mode = read_uv_mode(bc, vp8_kf_uv_mode_prob);
 }
 
-static int read_mvcomponent(vp8_reader *r, const MV_CONTEXT *mvc)
-{
-    const vp8_prob *const p = (const vp8_prob *) mvc;
-    int x = 0;
+static int read_mvcomponent(vp8_reader *r, const MV_CONTEXT *mvc) {
+  const vp8_prob *const p = (const vp8_prob *)mvc;
+  int x = 0;
 
-    if (vp8_read(r, p [mvpis_short]))  /* Large */
-    {
-        int i = 0;
+  if (vp8_read(r, p[mvpis_short])) /* Large */
+  {
+    int i = 0;
 
-        do
-        {
-            x += vp8_read(r, p [MVPbits + i]) << i;
-        }
-        while (++i < 3);
+    do {
+      x += vp8_read(r, p[MVPbits + i]) << i;
+    } while (++i < 3);
 
-        i = mvlong_width - 1;  /* Skip bit 3, which is sometimes implicit */
+    i = mvlong_width - 1; /* Skip bit 3, which is sometimes implicit */
 
-        do
-        {
-            x += vp8_read(r, p [MVPbits + i]) << i;
-        }
-        while (--i > 3);
+    do {
+      x += vp8_read(r, p[MVPbits + i]) << i;
+    } while (--i > 3);
 
-        if (!(x & 0xFFF0)  ||  vp8_read(r, p [MVPbits + 3]))
-            x += 8;
-    }
-    else   /* small */
-        x = vp8_treed_read(r, vp8_small_mvtree, p + MVPshort);
+    if (!(x & 0xFFF0) || vp8_read(r, p[MVPbits + 3])) x += 8;
+  } else /* small */
+    x = vp8_treed_read(r, vp8_small_mvtree, p + MVPshort);
 
-    if (x  &&  vp8_read(r, p [MVPsign]))
-        x = -x;
+  if (x && vp8_read(r, p[MVPsign])) x = -x;
 
-    return x;
+  return x;
 }
 
-static void read_mv(vp8_reader *r, MV *mv, const MV_CONTEXT *mvc)
-{
-    mv->row = (short)(read_mvcomponent(r,   mvc) * 2);
-    mv->col = (short)(read_mvcomponent(r, ++mvc) * 2);
+static void read_mv(vp8_reader *r, MV *mv, const MV_CONTEXT *mvc) {
+  mv->row = (short)(read_mvcomponent(r, mvc) * 2);
+  mv->col = (short)(read_mvcomponent(r, ++mvc) * 2);
 }
 
+static void read_mvcontexts(vp8_reader *bc, MV_CONTEXT *mvc) {
+  int i = 0;
 
-static void read_mvcontexts(vp8_reader *bc, MV_CONTEXT *mvc)
-{
-    int i = 0;
+  do {
+    const vp8_prob *up = vp8_mv_update_probs[i].prob;
+    vp8_prob *p = (vp8_prob *)(mvc + i);
+    vp8_prob *const pstop = p + MVPcount;
 
-    do
-    {
-        const vp8_prob *up = vp8_mv_update_probs[i].prob;
-        vp8_prob *p = (vp8_prob *)(mvc + i);
-        vp8_prob *const pstop = p + MVPcount;
+    do {
+      if (vp8_read(bc, *up++)) {
+        const vp8_prob x = (vp8_prob)vp8_read_literal(bc, 7);
 
-        do
-        {
-            if (vp8_read(bc, *up++))
-            {
-                const vp8_prob x = (vp8_prob)vp8_read_literal(bc, 7);
-
-                *p = x ? x << 1 : 1;
-            }
-        }
-        while (++p < pstop);
-    }
-    while (++i < 2);
+        *p = x ? x << 1 : 1;
+      }
+    } while (++p < pstop);
+  } while (++i < 2);
 }
 
-static const unsigned char mbsplit_fill_count[4] = {8, 8, 4, 1};
+static const unsigned char mbsplit_fill_count[4] = { 8, 8, 4, 1 };
 static const unsigned char mbsplit_fill_offset[4][16] = {
-    { 0,  1,  2,  3,  4,  5,  6,  7,  8,  9,  10, 11, 12, 13, 14, 15},
-    { 0,  1,  4,  5,  8,  9, 12, 13,  2,  3,   6,  7, 10, 11, 14, 15},
-    { 0,  1,  4,  5,  2,  3,  6,  7,  8,  9,  12, 13, 10, 11, 14, 15},
-    { 0,  1,  2,  3,  4,  5,  6,  7,  8,  9,  10, 11, 12, 13, 14, 15}
+  { 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15 },
+  { 0, 1, 4, 5, 8, 9, 12, 13, 2, 3, 6, 7, 10, 11, 14, 15 },
+  { 0, 1, 4, 5, 2, 3, 6, 7, 8, 9, 12, 13, 10, 11, 14, 15 },
+  { 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15 }
 };
 
+static void mb_mode_mv_init(VP8D_COMP *pbi) {
+  vp8_reader *const bc = &pbi->mbc[8];
+  MV_CONTEXT *const mvc = pbi->common.fc.mvc;
 
-static void mb_mode_mv_init(VP8D_COMP *pbi)
-{
-    vp8_reader *const bc = & pbi->mbc[8];
-    MV_CONTEXT *const mvc = pbi->common.fc.mvc;
-
 #if CONFIG_ERROR_CONCEALMENT
-    /* Default is that no macroblock is corrupt, therefore we initialize
-     * mvs_corrupt_from_mb to something very big, which we can be sure is
-     * outside the frame. */
-    pbi->mvs_corrupt_from_mb = UINT_MAX;
+  /* Default is that no macroblock is corrupt, therefore we initialize
+   * mvs_corrupt_from_mb to something very big, which we can be sure is
+   * outside the frame. */
+  pbi->mvs_corrupt_from_mb = UINT_MAX;
 #endif
-    /* Read the mb_no_coeff_skip flag */
-    pbi->common.mb_no_coeff_skip = (int)vp8_read_bit(bc);
+  /* Read the mb_no_coeff_skip flag */
+  pbi->common.mb_no_coeff_skip = (int)vp8_read_bit(bc);
 
-    pbi->prob_skip_false = 0;
-    if (pbi->common.mb_no_coeff_skip)
-        pbi->prob_skip_false = (vp8_prob)vp8_read_literal(bc, 8);
+  pbi->prob_skip_false = 0;
+  if (pbi->common.mb_no_coeff_skip)
+    pbi->prob_skip_false = (vp8_prob)vp8_read_literal(bc, 8);
 
-    if(pbi->common.frame_type != KEY_FRAME)
-    {
-        pbi->prob_intra = (vp8_prob)vp8_read_literal(bc, 8);
-        pbi->prob_last  = (vp8_prob)vp8_read_literal(bc, 8);
-        pbi->prob_gf    = (vp8_prob)vp8_read_literal(bc, 8);
+  if (pbi->common.frame_type != KEY_FRAME) {
+    pbi->prob_intra = (vp8_prob)vp8_read_literal(bc, 8);
+    pbi->prob_last = (vp8_prob)vp8_read_literal(bc, 8);
+    pbi->prob_gf = (vp8_prob)vp8_read_literal(bc, 8);
 
-        if (vp8_read_bit(bc))
-        {
-            int i = 0;
+    if (vp8_read_bit(bc)) {
+      int i = 0;
 
-            do
-            {
-                pbi->common.fc.ymode_prob[i] =
-                    (vp8_prob) vp8_read_literal(bc, 8);
-            }
-            while (++i < 4);
-        }
+      do {
+        pbi->common.fc.ymode_prob[i] = (vp8_prob)vp8_read_literal(bc, 8);
+      } while (++i < 4);
+    }
 
-        if (vp8_read_bit(bc))
-        {
-            int i = 0;
+    if (vp8_read_bit(bc)) {
+      int i = 0;
 
-            do
-            {
-                pbi->common.fc.uv_mode_prob[i] =
-                    (vp8_prob) vp8_read_literal(bc, 8);
-            }
-            while (++i < 3);
-        }
-
-        read_mvcontexts(bc, mvc);
+      do {
+        pbi->common.fc.uv_mode_prob[i] = (vp8_prob)vp8_read_literal(bc, 8);
+      } while (++i < 3);
     }
+
+    read_mvcontexts(bc, mvc);
+  }
 }
 
-const vp8_prob vp8_sub_mv_ref_prob3 [8][VP8_SUBMVREFS-1] =
-{
-    { 147, 136, 18 },   /* SUBMVREF_NORMAL          */
-    { 223, 1  , 34 },   /* SUBMVREF_LEFT_ABOVE_SAME */
-    { 106, 145, 1  },   /* SUBMVREF_LEFT_ZED        */
-    { 208, 1  , 1  },   /* SUBMVREF_LEFT_ABOVE_ZED  */
-    { 179, 121, 1  },   /* SUBMVREF_ABOVE_ZED       */
-    { 223, 1  , 34 },   /* SUBMVREF_LEFT_ABOVE_SAME */
-    { 179, 121, 1  },   /* SUBMVREF_ABOVE_ZED       */
-    { 208, 1  , 1  }    /* SUBMVREF_LEFT_ABOVE_ZED  */
+const vp8_prob vp8_sub_mv_ref_prob3[8][VP8_SUBMVREFS - 1] = {
+  { 147, 136, 18 }, /* SUBMVREF_NORMAL          */
+  { 223, 1, 34 },   /* SUBMVREF_LEFT_ABOVE_SAME */
+  { 106, 145, 1 },  /* SUBMVREF_LEFT_ZED        */
+  { 208, 1, 1 },    /* SUBMVREF_LEFT_ABOVE_ZED  */
+  { 179, 121, 1 },  /* SUBMVREF_ABOVE_ZED       */
+  { 223, 1, 34 },   /* SUBMVREF_LEFT_ABOVE_SAME */
+  { 179, 121, 1 },  /* SUBMVREF_ABOVE_ZED       */
+  { 208, 1, 1 }     /* SUBMVREF_LEFT_ABOVE_ZED  */
 };
 
-static
-const vp8_prob * get_sub_mv_ref_prob(const int left, const int above)
-{
-    int lez = (left == 0);
-    int aez = (above == 0);
-    int lea = (left == above);
-    const vp8_prob * prob;
+static const vp8_prob *get_sub_mv_ref_prob(const int left, const int above) {
+  int lez = (left == 0);
+  int aez = (above == 0);
+  int lea = (left == above);
+  const vp8_prob *prob;
 
-    prob = vp8_sub_mv_ref_prob3[(aez << 2) |
-                                (lez << 1) |
-                                (lea)];
+  prob = vp8_sub_mv_ref_prob3[(aez << 2) | (lez << 1) | (lea)];
 
-    return prob;
+  return prob;
 }
 
 static void decode_split_mv(vp8_reader *const bc, MODE_INFO *mi,
-                        const MODE_INFO *left_mb, const MODE_INFO *above_mb,
-                        MB_MODE_INFO *mbmi, int_mv best_mv,
-                        MV_CONTEXT *const mvc, int mb_to_left_edge,
-                        int mb_to_right_edge, int mb_to_top_edge,
-                        int mb_to_bottom_edge)
-{
-    int s;      /* split configuration (16x8, 8x16, 8x8, 4x4) */
-    int num_p;  /* number of partitions in the split configuration
-                  (see vp8_mbsplit_count) */
-    int j = 0;
+                            const MODE_INFO *left_mb, const MODE_INFO *above_mb,
+                            MB_MODE_INFO *mbmi, int_mv best_mv,
+                            MV_CONTEXT *const mvc, int mb_to_left_edge,
+                            int mb_to_right_edge, int mb_to_top_edge,
+                            int mb_to_bottom_edge) {
+  int s; /* split configuration (16x8, 8x16, 8x8, 4x4) */
+  /* number of partitions in the split configuration (see vp8_mbsplit_count) */
+  int num_p;
+  int j = 0;
 
-    s = 3;
-    num_p = 16;
-    if( vp8_read(bc, 110) )
-    {
-        s = 2;
-        num_p = 4;
-        if( vp8_read(bc, 111) )
-        {
-            s = vp8_read(bc, 150);
-            num_p = 2;
-        }
+  s = 3;
+  num_p = 16;
+  if (vp8_read(bc, 110)) {
+    s = 2;
+    num_p = 4;
+    if (vp8_read(bc, 111)) {
+      s = vp8_read(bc, 150);
+      num_p = 2;
     }
+  }
 
-    do  /* for each subset j */
-    {
-        int_mv leftmv, abovemv;
-        int_mv blockmv;
-        int k;  /* first block in subset j */
+  do /* for each subset j */
+  {
+    int_mv leftmv, abovemv;
+    int_mv blockmv;
+    int k; /* first block in subset j */
 
-        const vp8_prob *prob;
-        k = vp8_mbsplit_offset[s][j];
+    const vp8_prob *prob;
+    k = vp8_mbsplit_offset[s][j];
 
-        if (!(k & 3))
-        {
-            /* On L edge, get from MB to left of us */
-            if(left_mb->mbmi.mode != SPLITMV)
-                leftmv.as_int =  left_mb->mbmi.mv.as_int;
-            else
-                leftmv.as_int =  (left_mb->bmi + k + 4 - 1)->mv.as_int;
-        }
-        else
-            leftmv.as_int =  (mi->bmi + k - 1)->mv.as_int;
+    if (!(k & 3)) {
+      /* On L edge, get from MB to left of us */
+      if (left_mb->mbmi.mode != SPLITMV)
+        leftmv.as_int = left_mb->mbmi.mv.as_int;
+      else
+        leftmv.as_int = (left_mb->bmi + k + 4 - 1)->mv.as_int;
+    } else
+      leftmv.as_int = (mi->bmi + k - 1)->mv.as_int;
 
-        if (!(k >> 2))
-        {
-            /* On top edge, get from MB above us */
-            if(above_mb->mbmi.mode != SPLITMV)
-                abovemv.as_int =  above_mb->mbmi.mv.as_int;
-            else
-                abovemv.as_int =  (above_mb->bmi + k + 16 - 4)->mv.as_int;
-        }
-        else
-            abovemv.as_int = (mi->bmi + k - 4)->mv.as_int;
+    if (!(k >> 2)) {
+      /* On top edge, get from MB above us */
+      if (above_mb->mbmi.mode != SPLITMV)
+        abovemv.as_int = above_mb->mbmi.mv.as_int;
+      else
+        abovemv.as_int = (above_mb->bmi + k + 16 - 4)->mv.as_int;
+    } else
+      abovemv.as_int = (mi->bmi + k - 4)->mv.as_int;
 
-        prob = get_sub_mv_ref_prob(leftmv.as_int, abovemv.as_int);
+    prob = get_sub_mv_ref_prob(leftmv.as_int, abovemv.as_int);
 
-        if( vp8_read(bc, prob[0]) )
-        {
-            if( vp8_read(bc, prob[1]) )
-            {
-                blockmv.as_int = 0;
-                if( vp8_read(bc, prob[2]) )
-                {
-                    blockmv.as_mv.row = read_mvcomponent(bc, &mvc[0]) * 2;
-                    blockmv.as_mv.row += best_mv.as_mv.row;
-                    blockmv.as_mv.col = read_mvcomponent(bc, &mvc[1]) * 2;
-                    blockmv.as_mv.col += best_mv.as_mv.col;
-                }
-            }
-            else
-            {
-                blockmv.as_int = abovemv.as_int;
-            }
+    if (vp8_read(bc, prob[0])) {
+      if (vp8_read(bc, prob[1])) {
+        blockmv.as_int = 0;
+        if (vp8_read(bc, prob[2])) {
+          blockmv.as_mv.row = read_mvcomponent(bc, &mvc[0]) * 2;
+          blockmv.as_mv.row += best_mv.as_mv.row;
+          blockmv.as_mv.col = read_mvcomponent(bc, &mvc[1]) * 2;
+          blockmv.as_mv.col += best_mv.as_mv.col;
         }
-        else
-        {
-            blockmv.as_int = leftmv.as_int;
-        }
+      } else {
+        blockmv.as_int = abovemv.as_int;
+      }
+    } else {
+      blockmv.as_int = leftmv.as_int;
+    }
 
-        mbmi->need_to_clamp_mvs |= vp8_check_mv_bounds(&blockmv,
-                                                  mb_to_left_edge,
-                                                  mb_to_right_edge,
-                                                  mb_to_top_edge,
-                                                  mb_to_bottom_edge);
+    mbmi->need_to_clamp_mvs |=
+        vp8_check_mv_bounds(&blockmv, mb_to_left_edge, mb_to_right_edge,
+                            mb_to_top_edge, mb_to_bottom_edge);
 
-        {
-            /* Fill (uniform) modes, mvs of jth subset.
-             Must do it here because ensuing subsets can
-             refer back to us via "left" or "above". */
-            const unsigned char *fill_offset;
-            unsigned int fill_count = mbsplit_fill_count[s];
+    {
+      /* Fill (uniform) modes, mvs of jth subset.
+       Must do it here because ensuing subsets can
+       refer back to us via "left" or "above". */
+      const unsigned char *fill_offset;
+      unsigned int fill_count = mbsplit_fill_count[s];
 
-            fill_offset = &mbsplit_fill_offset[s]
-                             [(unsigned char)j * mbsplit_fill_count[s]];
+      fill_offset =
+          &mbsplit_fill_offset[s][(unsigned char)j * mbsplit_fill_count[s]];
 
-            do {
-                mi->bmi[ *fill_offset].mv.as_int = blockmv.as_int;
-                fill_offset++;
-            }while (--fill_count);
-        }
-
+      do {
+        mi->bmi[*fill_offset].mv.as_int = blockmv.as_int;
+        fill_offset++;
+      } while (--fill_count);
     }
-    while (++j < num_p);
 
-    mbmi->partitioning = s;
+  } while (++j < num_p);
+
+  mbmi->partitioning = s;
 }
 
-static void read_mb_modes_mv(VP8D_COMP *pbi, MODE_INFO *mi, MB_MODE_INFO *mbmi)
-{
-    vp8_reader *const bc = & pbi->mbc[8];
-    mbmi->ref_frame = (MV_REFERENCE_FRAME) vp8_read(bc, pbi->prob_intra);
-    if (mbmi->ref_frame)    /* inter MB */
-    {
-        enum {CNT_INTRA, CNT_NEAREST, CNT_NEAR, CNT_SPLITMV};
-        int cnt[4];
-        int *cntx = cnt;
-        int_mv near_mvs[4];
-        int_mv *nmv = near_mvs;
-        const int mis = pbi->mb.mode_info_stride;
-        const MODE_INFO *above = mi - mis;
-        const MODE_INFO *left = mi - 1;
-        const MODE_INFO *aboveleft = above - 1;
-        int *ref_frame_sign_bias = pbi->common.ref_frame_sign_bias;
+static void read_mb_modes_mv(VP8D_COMP *pbi, MODE_INFO *mi,
+                             MB_MODE_INFO *mbmi) {
+  vp8_reader *const bc = &pbi->mbc[8];
+  mbmi->ref_frame = (MV_REFERENCE_FRAME)vp8_read(bc, pbi->prob_intra);
+  if (mbmi->ref_frame) /* inter MB */
+  {
+    enum { CNT_INTRA, CNT_NEAREST, CNT_NEAR, CNT_SPLITMV };
+    int cnt[4];
+    int *cntx = cnt;
+    int_mv near_mvs[4];
+    int_mv *nmv = near_mvs;
+    const int mis = pbi->mb.mode_info_stride;
+    const MODE_INFO *above = mi - mis;
+    const MODE_INFO *left = mi - 1;
+    const MODE_INFO *aboveleft = above - 1;
+    int *ref_frame_sign_bias = pbi->common.ref_frame_sign_bias;
 
-        mbmi->need_to_clamp_mvs = 0;
+    mbmi->need_to_clamp_mvs = 0;
 
-        if (vp8_read(bc, pbi->prob_last))
-        {
-            mbmi->ref_frame =
-                (MV_REFERENCE_FRAME)((int)(2 + vp8_read(bc, pbi->prob_gf)));
-        }
+    if (vp8_read(bc, pbi->prob_last)) {
+      mbmi->ref_frame =
+          (MV_REFERENCE_FRAME)((int)(2 + vp8_read(bc, pbi->prob_gf)));
+    }
 
-        /* Zero accumulators */
-        nmv[0].as_int = nmv[1].as_int = nmv[2].as_int = 0;
-        cnt[0] = cnt[1] = cnt[2] = cnt[3] = 0;
+    /* Zero accumulators */
+    nmv[0].as_int = nmv[1].as_int = nmv[2].as_int = 0;
+    cnt[0] = cnt[1] = cnt[2] = cnt[3] = 0;
 
-        /* Process above */
-        if (above->mbmi.ref_frame != INTRA_FRAME)
-        {
-            if (above->mbmi.mv.as_int)
-            {
-                (++nmv)->as_int = above->mbmi.mv.as_int;
-                mv_bias(ref_frame_sign_bias[above->mbmi.ref_frame],
-                        mbmi->ref_frame, nmv, ref_frame_sign_bias);
-                ++cntx;
-            }
+    /* Process above */
+    if (above->mbmi.ref_frame != INTRA_FRAME) {
+      if (above->mbmi.mv.as_int) {
+        (++nmv)->as_int = above->mbmi.mv.as_int;
+        mv_bias(ref_frame_sign_bias[above->mbmi.ref_frame], mbmi->ref_frame,
+                nmv, ref_frame_sign_bias);
+        ++cntx;
+      }
 
-            *cntx += 2;
-        }
+      *cntx += 2;
+    }
 
-        /* Process left */
-        if (left->mbmi.ref_frame != INTRA_FRAME)
-        {
-            if (left->mbmi.mv.as_int)
-            {
-                int_mv this_mv;
+    /* Process left */
+    if (left->mbmi.ref_frame != INTRA_FRAME) {
+      if (left->mbmi.mv.as_int) {
+        int_mv this_mv;
 
-                this_mv.as_int = left->mbmi.mv.as_int;
-                mv_bias(ref_frame_sign_bias[left->mbmi.ref_frame],
-                        mbmi->ref_frame, &this_mv, ref_frame_sign_bias);
+        this_mv.as_int = left->mbmi.mv.as_int;
+        mv_bias(ref_frame_sign_bias[left->mbmi.ref_frame], mbmi->ref_frame,
+                &this_mv, ref_frame_sign_bias);
 
-                if (this_mv.as_int != nmv->as_int)
-                {
-                    (++nmv)->as_int = this_mv.as_int;
-                    ++cntx;
-                }
-
-                *cntx += 2;
-            }
-            else
-                cnt[CNT_INTRA] += 2;
+        if (this_mv.as_int != nmv->as_int) {
+          (++nmv)->as_int = this_mv.as_int;
+          ++cntx;
         }
 
-        /* Process above left */
-        if (aboveleft->mbmi.ref_frame != INTRA_FRAME)
-        {
-            if (aboveleft->mbmi.mv.as_int)
-            {
-                int_mv this_mv;
+        *cntx += 2;
+      } else
+        cnt[CNT_INTRA] += 2;
+    }
 
-                this_mv.as_int = aboveleft->mbmi.mv.as_int;
-                mv_bias(ref_frame_sign_bias[aboveleft->mbmi.ref_frame],
-                        mbmi->ref_frame, &this_mv, ref_frame_sign_bias);
+    /* Process above left */
+    if (aboveleft->mbmi.ref_frame != INTRA_FRAME) {
+      if (aboveleft->mbmi.mv.as_int) {
+        int_mv this_mv;
 
-                if (this_mv.as_int != nmv->as_int)
-                {
-                    (++nmv)->as_int = this_mv.as_int;
-                    ++cntx;
-                }
+        this_mv.as_int = aboveleft->mbmi.mv.as_int;
+        mv_bias(ref_frame_sign_bias[aboveleft->mbmi.ref_frame], mbmi->ref_frame,
+                &this_mv, ref_frame_sign_bias);
 
-                *cntx += 1;
-            }
-            else
-                cnt[CNT_INTRA] += 1;
+        if (this_mv.as_int != nmv->as_int) {
+          (++nmv)->as_int = this_mv.as_int;
+          ++cntx;
         }
 
-        if( vp8_read(bc, vp8_mode_contexts [cnt[CNT_INTRA]] [0]) )
-        {
+        *cntx += 1;
+      } else
+        cnt[CNT_INTRA] += 1;
+    }
 
-            /* If we have three distinct MV's ... */
-            /* See if above-left MV can be merged with NEAREST */
-            cnt[CNT_NEAREST] += ( (cnt[CNT_SPLITMV] > 0) &
-                (nmv->as_int == near_mvs[CNT_NEAREST].as_int));
+    if (vp8_read(bc, vp8_mode_contexts[cnt[CNT_INTRA]][0])) {
+      /* If we have three distinct MV's ... */
+      /* See if above-left MV can be merged with NEAREST */
+      cnt[CNT_NEAREST] += ((cnt[CNT_SPLITMV] > 0) &
+                           (nmv->as_int == near_mvs[CNT_NEAREST].as_int));
 
-            /* Swap near and nearest if necessary */
-            if (cnt[CNT_NEAR] > cnt[CNT_NEAREST])
-            {
-                int tmp;
-                tmp = cnt[CNT_NEAREST];
-                cnt[CNT_NEAREST] = cnt[CNT_NEAR];
-                cnt[CNT_NEAR] = tmp;
-                tmp = near_mvs[CNT_NEAREST].as_int;
-                near_mvs[CNT_NEAREST].as_int = near_mvs[CNT_NEAR].as_int;
-                near_mvs[CNT_NEAR].as_int = tmp;
-            }
+      /* Swap near and nearest if necessary */
+      if (cnt[CNT_NEAR] > cnt[CNT_NEAREST]) {
+        int tmp;
+        tmp = cnt[CNT_NEAREST];
+        cnt[CNT_NEAREST] = cnt[CNT_NEAR];
+        cnt[CNT_NEAR] = tmp;
+        tmp = near_mvs[CNT_NEAREST].as_int;
+        near_mvs[CNT_NEAREST].as_int = near_mvs[CNT_NEAR].as_int;
+        near_mvs[CNT_NEAR].as_int = tmp;
+      }
 
-            if( vp8_read(bc, vp8_mode_contexts [cnt[CNT_NEAREST]] [1]) )
-            {
+      if (vp8_read(bc, vp8_mode_contexts[cnt[CNT_NEAREST]][1])) {
+        if (vp8_read(bc, vp8_mode_contexts[cnt[CNT_NEAR]][2])) {
+          int mb_to_top_edge;
+          int mb_to_bottom_edge;
+          int mb_to_left_edge;
+          int mb_to_right_edge;
+          MV_CONTEXT *const mvc = pbi->common.fc.mvc;
+          int near_index;
 
-                if( vp8_read(bc, vp8_mode_contexts [cnt[CNT_NEAR]] [2]) )
-                {
-                    int mb_to_top_edge;
-                    int mb_to_bottom_edge;
-                    int mb_to_left_edge;
-                    int mb_to_right_edge;
-                    MV_CONTEXT *const mvc = pbi->common.fc.mvc;
-                    int near_index;
+          mb_to_top_edge = pbi->mb.mb_to_top_edge;
+          mb_to_bottom_edge = pbi->mb.mb_to_bottom_edge;
+          mb_to_top_edge -= LEFT_TOP_MARGIN;
+          mb_to_bottom_edge += RIGHT_BOTTOM_MARGIN;
+          mb_to_right_edge = pbi->mb.mb_to_right_edge;
+          mb_to_right_edge += RIGHT_BOTTOM_MARGIN;
+          mb_to_left_edge = pbi->mb.mb_to_left_edge;
+          mb_to_left_edge -= LEFT_TOP_MARGIN;
 
-                    mb_to_top_edge = pbi->mb.mb_to_top_edge;
-                    mb_to_bottom_edge = pbi->mb.mb_to_bottom_edge;
-                    mb_to_top_edge -= LEFT_TOP_MARGIN;
-                    mb_to_bottom_edge += RIGHT_BOTTOM_MARGIN;
-                    mb_to_right_edge = pbi->mb.mb_to_right_edge;
-                    mb_to_right_edge += RIGHT_BOTTOM_MARGIN;
-                    mb_to_left_edge = pbi->mb.mb_to_left_edge;
-                    mb_to_left_edge -= LEFT_TOP_MARGIN;
+          /* Use near_mvs[0] to store the "best" MV */
+          near_index = CNT_INTRA + (cnt[CNT_NEAREST] >= cnt[CNT_INTRA]);
 
-                    /* Use near_mvs[0] to store the "best" MV */
-                    near_index = CNT_INTRA +
-                        (cnt[CNT_NEAREST] >= cnt[CNT_INTRA]);
+          vp8_clamp_mv2(&near_mvs[near_index], &pbi->mb);
 
-                    vp8_clamp_mv2(&near_mvs[near_index], &pbi->mb);
+          cnt[CNT_SPLITMV] =
+              ((above->mbmi.mode == SPLITMV) + (left->mbmi.mode == SPLITMV)) *
+                  2 +
+              (aboveleft->mbmi.mode == SPLITMV);
 
-                    cnt[CNT_SPLITMV] = ((above->mbmi.mode == SPLITMV)
-                                        + (left->mbmi.mode == SPLITMV)) * 2
-                                       + (aboveleft->mbmi.mode == SPLITMV);
+          if (vp8_read(bc, vp8_mode_contexts[cnt[CNT_SPLITMV]][3])) {
+            decode_split_mv(bc, mi, left, above, mbmi, near_mvs[near_index],
+                            mvc, mb_to_left_edge, mb_to_right_edge,
+                            mb_to_top_edge, mb_to_bottom_edge);
+            mbmi->mv.as_int = mi->bmi[15].mv.as_int;
+            mbmi->mode = SPLITMV;
+            mbmi->is_4x4 = 1;
+          } else {
+            int_mv *const mbmi_mv = &mbmi->mv;
+            read_mv(bc, &mbmi_mv->as_mv, (const MV_CONTEXT *)mvc);
+            mbmi_mv->as_mv.row += near_mvs[near_index].as_mv.row;
+            mbmi_mv->as_mv.col += near_mvs[near_index].as_mv.col;
 
-                    if( vp8_read(bc, vp8_mode_contexts [cnt[CNT_SPLITMV]] [3]) )
-                    {
-                        decode_split_mv(bc, mi, left, above,
-                                                    mbmi,
-                                                    near_mvs[near_index],
-                                                    mvc, mb_to_left_edge,
-                                                    mb_to_right_edge,
-                                                    mb_to_top_edge,
-                                                    mb_to_bottom_edge);
-                        mbmi->mv.as_int = mi->bmi[15].mv.as_int;
-                        mbmi->mode =  SPLITMV;
-                        mbmi->is_4x4 = 1;
-                    }
-                    else
-                    {
-                        int_mv *const mbmi_mv = & mbmi->mv;
-                        read_mv(bc, &mbmi_mv->as_mv, (const MV_CONTEXT *) mvc);
-                        mbmi_mv->as_mv.row += near_mvs[near_index].as_mv.row;
-                        mbmi_mv->as_mv.col += near_mvs[near_index].as_mv.col;
+            /* Don't need to check this on NEARMV and NEARESTMV
+             * modes since those modes clamp the MV. The NEWMV mode
+             * does not, so signal to the prediction stage whether
+             * special handling may be required.
+             */
+            mbmi->need_to_clamp_mvs =
+                vp8_check_mv_bounds(mbmi_mv, mb_to_left_edge, mb_to_right_edge,
+                                    mb_to_top_edge, mb_to_bottom_edge);
+            mbmi->mode = NEWMV;
+          }
+        } else {
+          mbmi->mode = NEARMV;
+          mbmi->mv.as_int = near_mvs[CNT_NEAR].as_int;
+          vp8_clamp_mv2(&mbmi->mv, &pbi->mb);
+        }
+      } else {
+        mbmi->mode = NEARESTMV;
+        mbmi->mv.as_int = near_mvs[CNT_NEAREST].as_int;
+        vp8_clamp_mv2(&mbmi->mv, &pbi->mb);
+      }
+    } else {
+      mbmi->mode = ZEROMV;
+      mbmi->mv.as_int = 0;
+    }
 
-                        /* Don't need to check this on NEARMV and NEARESTMV
-                         * modes since those modes clamp the MV. The NEWMV mode
-                         * does not, so signal to the prediction stage whether
-                         * special handling may be required.
-                         */
-                        mbmi->need_to_clamp_mvs =
-                            vp8_check_mv_bounds(mbmi_mv, mb_to_left_edge,
-                                                mb_to_right_edge,
-                                                mb_to_top_edge,
-                                                mb_to_bottom_edge);
-                        mbmi->mode =  NEWMV;
-                    }
-                }
-                else
-                {
-                    mbmi->mode =  NEARMV;
-                    mbmi->mv.as_int = near_mvs[CNT_NEAR].as_int;
-                    vp8_clamp_mv2(&mbmi->mv, &pbi->mb);
-                }
-            }
-            else
-            {
-                mbmi->mode =  NEARESTMV;
-                mbmi->mv.as_int = near_mvs[CNT_NEAREST].as_int;
-                vp8_clamp_mv2(&mbmi->mv, &pbi->mb);
-            }
-        }
-        else
-        {
-            mbmi->mode =  ZEROMV;
-            mbmi->mv.as_int = 0;
-        }
-
 #if CONFIG_ERROR_CONCEALMENT
-        if(pbi->ec_enabled && (mbmi->mode != SPLITMV))
-        {
-            mi->bmi[ 0].mv.as_int =
-            mi->bmi[ 1].mv.as_int =
-            mi->bmi[ 2].mv.as_int =
-            mi->bmi[ 3].mv.as_int =
-            mi->bmi[ 4].mv.as_int =
-            mi->bmi[ 5].mv.as_int =
-            mi->bmi[ 6].mv.as_int =
-            mi->bmi[ 7].mv.as_int =
-            mi->bmi[ 8].mv.as_int =
-            mi->bmi[ 9].mv.as_int =
-            mi->bmi[10].mv.as_int =
-            mi->bmi[11].mv.as_int =
-            mi->bmi[12].mv.as_int =
-            mi->bmi[13].mv.as_int =
-            mi->bmi[14].mv.as_int =
-            mi->bmi[15].mv.as_int = mbmi->mv.as_int;
-        }
-#endif
+    if (pbi->ec_enabled && (mbmi->mode != SPLITMV)) {
+      mi->bmi[0].mv.as_int = mi->bmi[1].mv.as_int = mi->bmi[2].mv.as_int =
+          mi->bmi[3].mv.as_int = mi->bmi[4].mv.as_int = mi->bmi[5].mv.as_int =
+              mi->bmi[6].mv.as_int = mi->bmi[7].mv.as_int =
+                  mi->bmi[8].mv.as_int = mi->bmi[9].mv.as_int =
+                      mi->bmi[10].mv.as_int = mi->bmi[11].mv.as_int =
+                          mi->bmi[12].mv.as_int = mi->bmi[13].mv.as_int =
+                              mi->bmi[14].mv.as_int = mi->bmi[15].mv.as_int =
+                                  mbmi->mv.as_int;
     }
-    else
-    {
-        /* required for left and above block mv */
-        mbmi->mv.as_int = 0;
+#endif
+  } else {
+    /* required for left and above block mv */
+    mbmi->mv.as_int = 0;
 
-        /* MB is intra coded */
-        if ((mbmi->mode = read_ymode(bc, pbi->common.fc.ymode_prob)) == B_PRED)
-        {
-            int j = 0;
-            mbmi->is_4x4 = 1;
-            do
-            {
-                mi->bmi[j].as_mode = read_bmode(bc, pbi->common.fc.bmode_prob);
-            }
-            while (++j < 16);
-        }
-
-        mbmi->uv_mode = read_uv_mode(bc, pbi->common.fc.uv_mode_prob);
+    /* MB is intra coded */
+    if ((mbmi->mode = read_ymode(bc, pbi->common.fc.ymode_prob)) == B_PRED) {
+      int j = 0;
+      mbmi->is_4x4 = 1;
+      do {
+        mi->bmi[j].as_mode = read_bmode(bc, pbi->common.fc.bmode_prob);
+      } while (++j < 16);
     }
 
+    mbmi->uv_mode = read_uv_mode(bc, pbi->common.fc.uv_mode_prob);
+  }
 }
 
-static void read_mb_features(vp8_reader *r, MB_MODE_INFO *mi, MACROBLOCKD *x)
-{
-    /* Is segmentation enabled */
-    if (x->segmentation_enabled && x->update_mb_segmentation_map)
-    {
-        /* If so then read the segment id. */
-        if (vp8_read(r, x->mb_segment_tree_probs[0]))
-            mi->segment_id =
-                (unsigned char)(2 + vp8_read(r, x->mb_segment_tree_probs[2]));
-        else
-            mi->segment_id =
-                (unsigned char)(vp8_read(r, x->mb_segment_tree_probs[1]));
-    }
+static void read_mb_features(vp8_reader *r, MB_MODE_INFO *mi, MACROBLOCKD *x) {
+  /* Is segmentation enabled */
+  if (x->segmentation_enabled && x->update_mb_segmentation_map) {
+    /* If so then read the segment id. */
+    if (vp8_read(r, x->mb_segment_tree_probs[0]))
+      mi->segment_id =
+          (unsigned char)(2 + vp8_read(r, x->mb_segment_tree_probs[2]));
+    else
+      mi->segment_id =
+          (unsigned char)(vp8_read(r, x->mb_segment_tree_probs[1]));
+  }
 }
 
 static void decode_mb_mode_mvs(VP8D_COMP *pbi, MODE_INFO *mi,
-                               MB_MODE_INFO *mbmi)
-{
-    (void)mbmi;
+                               MB_MODE_INFO *mbmi) {
+  (void)mbmi;
 
-    /* Read the Macroblock segmentation map if it is being updated explicitly
-     * this frame (reset to 0 above by default)
-     * By default on a key frame reset all MBs to segment 0
-     */
-    if (pbi->mb.update_mb_segmentation_map)
-        read_mb_features(&pbi->mbc[8], &mi->mbmi, &pbi->mb);
-    else if(pbi->common.frame_type == KEY_FRAME)
-        mi->mbmi.segment_id = 0;
+  /* Read the Macroblock segmentation map if it is being updated explicitly
+   * this frame (reset to 0 above by default)
+   * By default on a key frame reset all MBs to segment 0
+   */
+  if (pbi->mb.update_mb_segmentation_map)
+    read_mb_features(&pbi->mbc[8], &mi->mbmi, &pbi->mb);
+  else if (pbi->common.frame_type == KEY_FRAME)
+    mi->mbmi.segment_id = 0;
 
-    /* Read the macroblock coeff skip flag if this feature is in use,
-     * else default to 0 */
-    if (pbi->common.mb_no_coeff_skip)
-        mi->mbmi.mb_skip_coeff = vp8_read(&pbi->mbc[8], pbi->prob_skip_false);
-    else
-        mi->mbmi.mb_skip_coeff = 0;
+  /* Read the macroblock coeff skip flag if this feature is in use,
+   * else default to 0 */
+  if (pbi->common.mb_no_coeff_skip)
+    mi->mbmi.mb_skip_coeff = vp8_read(&pbi->mbc[8], pbi->prob_skip_false);
+  else
+    mi->mbmi.mb_skip_coeff = 0;
 
-    mi->mbmi.is_4x4 = 0;
-    if(pbi->common.frame_type == KEY_FRAME)
-        read_kf_modes(pbi, mi);
-    else
-        read_mb_modes_mv(pbi, mi, &mi->mbmi);
-
+  mi->mbmi.is_4x4 = 0;
+  if (pbi->common.frame_type == KEY_FRAME)
+    read_kf_modes(pbi, mi);
+  else
+    read_mb_modes_mv(pbi, mi, &mi->mbmi);
 }
 
-void vp8_decode_mode_mvs(VP8D_COMP *pbi)
-{
-    MODE_INFO *mi = pbi->common.mi;
-    int mb_row = -1;
-    int mb_to_right_edge_start;
+void vp8_decode_mode_mvs(VP8D_COMP *pbi) {
+  MODE_INFO *mi = pbi->common.mi;
+  int mb_row = -1;
+  int mb_to_right_edge_start;
 
-    mb_mode_mv_init(pbi);
+  mb_mode_mv_init(pbi);
 
-    pbi->mb.mb_to_top_edge = 0;
-    pbi->mb.mb_to_bottom_edge = ((pbi->common.mb_rows - 1) * 16) << 3;
-    mb_to_right_edge_start = ((pbi->common.mb_cols - 1) * 16) << 3;
+  pbi->mb.mb_to_top_edge = 0;
+  pbi->mb.mb_to_bottom_edge = ((pbi->common.mb_rows - 1) * 16) << 3;
+  mb_to_right_edge_start = ((pbi->common.mb_cols - 1) * 16) << 3;
 
-    while (++mb_row < pbi->common.mb_rows)
-    {
-        int mb_col = -1;
+  while (++mb_row < pbi->common.mb_rows) {
+    int mb_col = -1;
 
-        pbi->mb.mb_to_left_edge =  0;
-        pbi->mb.mb_to_right_edge = mb_to_right_edge_start;
+    pbi->mb.mb_to_left_edge = 0;
+    pbi->mb.mb_to_right_edge = mb_to_right_edge_start;
 
-        while (++mb_col < pbi->common.mb_cols)
-        {
+    while (++mb_col < pbi->common.mb_cols) {
 #if CONFIG_ERROR_CONCEALMENT
-            int mb_num = mb_row * pbi->common.mb_cols + mb_col;
+      int mb_num = mb_row * pbi->common.mb_cols + mb_col;
 #endif
 
-            decode_mb_mode_mvs(pbi, mi, &mi->mbmi);
+      decode_mb_mode_mvs(pbi, mi, &mi->mbmi);
 
 #if CONFIG_ERROR_CONCEALMENT
-            /* look for corruption. set mvs_corrupt_from_mb to the current
-             * mb_num if the frame is corrupt from this macroblock. */
-            if (vp8dx_bool_error(&pbi->mbc[8]) && mb_num <
-                (int)pbi->mvs_corrupt_from_mb)
-            {
-                pbi->mvs_corrupt_from_mb = mb_num;
-                /* no need to continue since the partition is corrupt from
-                 * here on.
-                 */
-                return;
-            }
+      /* look for corruption. set mvs_corrupt_from_mb to the current
+       * mb_num if the frame is corrupt from this macroblock. */
+      if (vp8dx_bool_error(&pbi->mbc[8]) &&
+          mb_num < (int)pbi->mvs_corrupt_from_mb) {
+        pbi->mvs_corrupt_from_mb = mb_num;
+        /* no need to continue since the partition is corrupt from
+         * here on.
+         */
+        return;
+      }
 #endif
 
-            pbi->mb.mb_to_left_edge -= (16 << 3);
-            pbi->mb.mb_to_right_edge -= (16 << 3);
-            mi++;       /* next macroblock */
-        }
-        pbi->mb.mb_to_top_edge -= (16 << 3);
-        pbi->mb.mb_to_bottom_edge -= (16 << 3);
-
-        mi++;           /* skip left predictor each row */
+      pbi->mb.mb_to_left_edge -= (16 << 3);
+      pbi->mb.mb_to_right_edge -= (16 << 3);
+      mi++; /* next macroblock */
     }
+    pbi->mb.mb_to_top_edge -= (16 << 3);
+    pbi->mb.mb_to_bottom_edge -= (16 << 3);
+
+    mi++; /* skip left predictor each row */
+  }
 }
--- a/vp8/decoder/detokenize.c
+++ b/vp8/decoder/detokenize.c
@@ -8,7 +8,6 @@
  *  be found in the AUTHORS file in the root of the source tree.
  */
 
-
 #include "vp8/common/blockd.h"
 #include "onyxd_int.h"
 #include "vpx_mem/vpx_mem.h"
@@ -15,19 +14,17 @@
 #include "vpx_ports/mem.h"
 #include "detokenize.h"
 
-void vp8_reset_mb_tokens_context(MACROBLOCKD *x)
-{
-    ENTROPY_CONTEXT *a_ctx = ((ENTROPY_CONTEXT *)x->above_context);
-    ENTROPY_CONTEXT *l_ctx = ((ENTROPY_CONTEXT *)x->left_context);
+void vp8_reset_mb_tokens_context(MACROBLOCKD *x) {
+  ENTROPY_CONTEXT *a_ctx = ((ENTROPY_CONTEXT *)x->above_context);
+  ENTROPY_CONTEXT *l_ctx = ((ENTROPY_CONTEXT *)x->left_context);
 
-    memset(a_ctx, 0, sizeof(ENTROPY_CONTEXT_PLANES)-1);
-    memset(l_ctx, 0, sizeof(ENTROPY_CONTEXT_PLANES)-1);
+  memset(a_ctx, 0, sizeof(ENTROPY_CONTEXT_PLANES) - 1);
+  memset(l_ctx, 0, sizeof(ENTROPY_CONTEXT_PLANES) - 1);
 
-    /* Clear entropy contexts for Y2 blocks */
-    if (!x->mode_info_context->mbmi.is_4x4)
-    {
-        a_ctx[8] = l_ctx[8] = 0;
-    }
+  /* Clear entropy contexts for Y2 blocks */
+  if (!x->mode_info_context->mbmi.is_4x4) {
+    a_ctx[8] = l_ctx[8] = 0;
+  }
 }
 
 /*
@@ -35,211 +32,175 @@
     Residual decoding (Paragraph 13.2 / 13.3)
 */
 static const uint8_t kBands[16 + 1] = {
-  0, 1, 2, 3, 6, 4, 5, 6, 6, 6, 6, 6, 6, 6, 6, 7,
-  0  /* extra entry as sentinel */
+  0, 1, 2, 3, 6, 4, 5, 6, 6,
+  6, 6, 6, 6, 6, 6, 7, 0 /* extra entry as sentinel */
 };
 
 static const uint8_t kCat3[] = { 173, 148, 140, 0 };
 static const uint8_t kCat4[] = { 176, 155, 140, 135, 0 };
 static const uint8_t kCat5[] = { 180, 157, 141, 134, 130, 0 };
-static const uint8_t kCat6[] =
-  { 254, 254, 243, 230, 196, 177, 153, 140, 133, 130, 129, 0 };
-static const uint8_t* const kCat3456[] = { kCat3, kCat4, kCat5, kCat6 };
-static const uint8_t kZigzag[16] = {
-  0, 1, 4, 8,  5, 2, 3, 6,  9, 12, 13, 10,  7, 11, 14, 15
-};
+static const uint8_t kCat6[] = { 254, 254, 243, 230, 196, 177,
+                                 153, 140, 133, 130, 129, 0 };
+static const uint8_t *const kCat3456[] = { kCat3, kCat4, kCat5, kCat6 };
+static const uint8_t kZigzag[16] = { 0, 1,  4,  8,  5, 2,  3,  6,
+                                     9, 12, 13, 10, 7, 11, 14, 15 };
 
 #define VP8GetBit vp8dx_decode_bool
-#define NUM_PROBAS  11
-#define NUM_CTX  3
+#define NUM_PROBAS 11
+#define NUM_CTX 3
 
 /* for const-casting */
 typedef const uint8_t (*ProbaArray)[NUM_CTX][NUM_PROBAS];
 
-static int GetSigned(BOOL_DECODER *br, int value_to_sign)
-{
-    int split = (br->range + 1) >> 1;
-    VP8_BD_VALUE bigsplit = (VP8_BD_VALUE)split << (VP8_BD_VALUE_SIZE - 8);
-    int v;
+static int GetSigned(BOOL_DECODER *br, int value_to_sign) {
+  int split = (br->range + 1) >> 1;
+  VP8_BD_VALUE bigsplit = (VP8_BD_VALUE)split << (VP8_BD_VALUE_SIZE - 8);
+  int v;
 
-    if(br->count < 0)
-        vp8dx_bool_decoder_fill(br);
+  if (br->count < 0) vp8dx_bool_decoder_fill(br);
 
-    if ( br->value < bigsplit )
-    {
-        br->range = split;
-        v= value_to_sign;
-    }
-    else
-    {
-        br->range = br->range-split;
-        br->value = br->value-bigsplit;
-        v = -value_to_sign;
-    }
-    br->range +=br->range;
-    br->value +=br->value;
-    br->count--;
+  if (br->value < bigsplit) {
+    br->range = split;
+    v = value_to_sign;
+  } else {
+    br->range = br->range - split;
+    br->value = br->value - bigsplit;
+    v = -value_to_sign;
+  }
+  br->range += br->range;
+  br->value += br->value;
+  br->count--;
 
-    return v;
+  return v;
 }
 /*
    Returns the position of the last non-zero coeff plus one
    (and 0 if there's no coeff at all)
 */
-static int GetCoeffs(BOOL_DECODER *br, ProbaArray prob,
-                     int ctx, int n, int16_t* out)
-{
-    const uint8_t* p = prob[n][ctx];
-    if (!VP8GetBit(br, p[0]))
-    {   /* first EOB is more a 'CBP' bit. */
-        return 0;
-    }
-    while (1)
-    {
-        ++n;
-        if (!VP8GetBit(br, p[1]))
-        {
-            p = prob[kBands[n]][0];
-        }
-        else
-        {  /* non zero coeff */
-            int v, j;
-            if (!VP8GetBit(br, p[2]))
-            {
-                p = prob[kBands[n]][1];
-                v = 1;
+static int GetCoeffs(BOOL_DECODER *br, ProbaArray prob, int ctx, int n,
+                     int16_t *out) {
+  const uint8_t *p = prob[n][ctx];
+  if (!VP8GetBit(br, p[0])) { /* first EOB is more a 'CBP' bit. */
+    return 0;
+  }
+  while (1) {
+    ++n;
+    if (!VP8GetBit(br, p[1])) {
+      p = prob[kBands[n]][0];
+    } else { /* non zero coeff */
+      int v, j;
+      if (!VP8GetBit(br, p[2])) {
+        p = prob[kBands[n]][1];
+        v = 1;
+      } else {
+        if (!VP8GetBit(br, p[3])) {
+          if (!VP8GetBit(br, p[4])) {
+            v = 2;
+          } else {
+            v = 3 + VP8GetBit(br, p[5]);
+          }
+        } else {
+          if (!VP8GetBit(br, p[6])) {
+            if (!VP8GetBit(br, p[7])) {
+              v = 5 + VP8GetBit(br, 159);
+            } else {
+              v = 7 + 2 * VP8GetBit(br, 165);
+              v += VP8GetBit(br, 145);
             }
-            else
-            {
-                if (!VP8GetBit(br, p[3]))
-                {
-                    if (!VP8GetBit(br, p[4]))
-                    {
-                        v = 2;
-                    }
-                    else
-                    {
-                        v = 3 + VP8GetBit(br, p[5]);
-                    }
-                }
-                else
-                {
-                    if (!VP8GetBit(br, p[6]))
-                    {
-                        if (!VP8GetBit(br, p[7]))
-                        {
-                            v = 5 + VP8GetBit(br, 159);
-                        } else
-                        {
-                            v = 7 + 2 * VP8GetBit(br, 165);
-                            v += VP8GetBit(br, 145);
-                        }
-                    }
-                    else
-                    {
-                        const uint8_t* tab;
-                        const int bit1 = VP8GetBit(br, p[8]);
-                        const int bit0 = VP8GetBit(br, p[9 + bit1]);
-                        const int cat = 2 * bit1 + bit0;
-                        v = 0;
-                        for (tab = kCat3456[cat]; *tab; ++tab)
-                        {
-                            v += v + VP8GetBit(br, *tab);
-                        }
-                        v += 3 + (8 << cat);
-                    }
-                }
-                p = prob[kBands[n]][2];
+          } else {
+            const uint8_t *tab;
+            const int bit1 = VP8GetBit(br, p[8]);
+            const int bit0 = VP8GetBit(br, p[9 + bit1]);
+            const int cat = 2 * bit1 + bit0;
+            v = 0;
+            for (tab = kCat3456[cat]; *tab; ++tab) {
+              v += v + VP8GetBit(br, *tab);
             }
-            j = kZigzag[n - 1];
+            v += 3 + (8 << cat);
+          }
+        }
+        p = prob[kBands[n]][2];
+      }
+      j = kZigzag[n - 1];
 
-            out[j] = GetSigned(br, v);
+      out[j] = GetSigned(br, v);
 
-            if (n == 16 || !VP8GetBit(br, p[0]))
-            {   /* EOB */
-                return n;
-            }
-        }
-        if (n == 16)
-        {
-            return 16;
-        }
+      if (n == 16 || !VP8GetBit(br, p[0])) { /* EOB */
+        return n;
+      }
     }
+    if (n == 16) {
+      return 16;
+    }
+  }
 }
 
-int vp8_decode_mb_tokens(VP8D_COMP *dx, MACROBLOCKD *x)
-{
-    BOOL_DECODER *bc = x->current_bc;
-    const FRAME_CONTEXT * const fc = &dx->common.fc;
-    char *eobs = x->eobs;
+int vp8_decode_mb_tokens(VP8D_COMP *dx, MACROBLOCKD *x) {
+  BOOL_DECODER *bc = x->current_bc;
+  const FRAME_CONTEXT *const fc = &dx->common.fc;
+  char *eobs = x->eobs;
 
-    int i;
-    int nonzeros;
-    int eobtotal = 0;
+  int i;
+  int nonzeros;
+  int eobtotal = 0;
 
-    short *qcoeff_ptr;
-    ProbaArray coef_probs;
-    ENTROPY_CONTEXT *a_ctx = ((ENTROPY_CONTEXT *)x->above_context);
-    ENTROPY_CONTEXT *l_ctx = ((ENTROPY_CONTEXT *)x->left_context);
-    ENTROPY_CONTEXT *a;
-    ENTROPY_CONTEXT *l;
-    int skip_dc = 0;
+  short *qcoeff_ptr;
+  ProbaArray coef_probs;
+  ENTROPY_CONTEXT *a_ctx = ((ENTROPY_CONTEXT *)x->above_context);
+  ENTROPY_CONTEXT *l_ctx = ((ENTROPY_CONTEXT *)x->left_context);
+  ENTROPY_CONTEXT *a;
+  ENTROPY_CONTEXT *l;
+  int skip_dc = 0;
 
-    qcoeff_ptr = &x->qcoeff[0];
+  qcoeff_ptr = &x->qcoeff[0];
 
-    if (!x->mode_info_context->mbmi.is_4x4)
-    {
-        a = a_ctx + 8;
-        l = l_ctx + 8;
+  if (!x->mode_info_context->mbmi.is_4x4) {
+    a = a_ctx + 8;
+    l = l_ctx + 8;
 
-        coef_probs = fc->coef_probs [1];
+    coef_probs = fc->coef_probs[1];
 
-        nonzeros = GetCoeffs(bc, coef_probs, (*a + *l), 0, qcoeff_ptr + 24 * 16);
-        *a = *l = (nonzeros > 0);
+    nonzeros = GetCoeffs(bc, coef_probs, (*a + *l), 0, qcoeff_ptr + 24 * 16);
+    *a = *l = (nonzeros > 0);
 
-        eobs[24] = nonzeros;
-        eobtotal += nonzeros - 16;
+    eobs[24] = nonzeros;
+    eobtotal += nonzeros - 16;
 
-        coef_probs = fc->coef_probs [0];
-        skip_dc = 1;
-    }
-    else
-    {
-        coef_probs = fc->coef_probs [3];
-        skip_dc = 0;
-    }
+    coef_probs = fc->coef_probs[0];
+    skip_dc = 1;
+  } else {
+    coef_probs = fc->coef_probs[3];
+    skip_dc = 0;
+  }
 
-    for (i = 0; i < 16; ++i)
-    {
-        a = a_ctx + (i&3);
-        l = l_ctx + ((i&0xc)>>2);
+  for (i = 0; i < 16; ++i) {
+    a = a_ctx + (i & 3);
+    l = l_ctx + ((i & 0xc) >> 2);
 
-        nonzeros = GetCoeffs(bc, coef_probs, (*a + *l), skip_dc, qcoeff_ptr);
-        *a = *l = (nonzeros > 0);
+    nonzeros = GetCoeffs(bc, coef_probs, (*a + *l), skip_dc, qcoeff_ptr);
+    *a = *l = (nonzeros > 0);
 
-        nonzeros += skip_dc;
-        eobs[i] = nonzeros;
-        eobtotal += nonzeros;
-        qcoeff_ptr += 16;
-    }
+    nonzeros += skip_dc;
+    eobs[i] = nonzeros;
+    eobtotal += nonzeros;
+    qcoeff_ptr += 16;
+  }
 
-    coef_probs = fc->coef_probs [2];
+  coef_probs = fc->coef_probs[2];
 
-    a_ctx += 4;
-    l_ctx += 4;
-    for (i = 16; i < 24; ++i)
-    {
-        a = a_ctx + ((i > 19)<<1) + (i&1);
-        l = l_ctx + ((i > 19)<<1) + ((i&3)>1);
+  a_ctx += 4;
+  l_ctx += 4;
+  for (i = 16; i < 24; ++i) {
+    a = a_ctx + ((i > 19) << 1) + (i & 1);
+    l = l_ctx + ((i > 19) << 1) + ((i & 3) > 1);
 
-        nonzeros = GetCoeffs(bc, coef_probs, (*a + *l), 0, qcoeff_ptr);
-        *a = *l = (nonzeros > 0);
+    nonzeros = GetCoeffs(bc, coef_probs, (*a + *l), 0, qcoeff_ptr);
+    *a = *l = (nonzeros > 0);
 
-        eobs[i] = nonzeros;
-        eobtotal += nonzeros;
-        qcoeff_ptr += 16;
-    }
+    eobs[i] = nonzeros;
+    eobtotal += nonzeros;
+    qcoeff_ptr += 16;
+  }
 
-    return eobtotal;
+  return eobtotal;
 }
-
--- a/vp8/decoder/ec_types.h
+++ b/vp8/decoder/ec_types.h
@@ -17,38 +17,31 @@
 
 #define MAX_OVERLAPS 16
 
-
 /* The area (pixel area in Q6) the block pointed to by bmi overlaps
  * another block with.
  */
-typedef struct
-{
-    int overlap;
-    union b_mode_info *bmi;
+typedef struct {
+  int overlap;
+  union b_mode_info *bmi;
 } OVERLAP_NODE;
 
 /* Structure to keep track of overlapping blocks on a block level. */
-typedef struct
-{
-    /* TODO(holmer): This array should be exchanged for a linked list */
-    OVERLAP_NODE overlaps[MAX_OVERLAPS];
+typedef struct {
+  /* TODO(holmer): This array should be exchanged for a linked list */
+  OVERLAP_NODE overlaps[MAX_OVERLAPS];
 } B_OVERLAP;
 
 /* Structure used to hold all the overlaps of a macroblock. The overlaps of a
  * macroblock is further divided into block overlaps.
  */
-typedef struct
-{
-    B_OVERLAP overlaps[16];
-} MB_OVERLAP;
+typedef struct { B_OVERLAP overlaps[16]; } MB_OVERLAP;
 
 /* Structure for keeping track of motion vectors and which reference frame they
  * refer to. Used for motion vector interpolation.
  */
-typedef struct
-{
-    MV mv;
-    MV_REFERENCE_FRAME ref_frame;
+typedef struct {
+  MV mv;
+  MV_REFERENCE_FRAME ref_frame;
 } EC_BLOCK;
 
 #ifdef __cplusplus
--- a/vp8/decoder/error_concealment.c
+++ b/vp8/decoder/error_concealment.c
@@ -18,14 +18,13 @@
 #include "vp8/common/common.h"
 #include "vpx_dsp/vpx_dsp_common.h"
 
-#define FLOOR(x,q) ((x) & -(1 << (q)))
+#define FLOOR(x, q) ((x) & -(1 << (q)))
 
 #define NUM_NEIGHBORS 20
 
-typedef struct ec_position
-{
-    int row;
-    int col;
+typedef struct ec_position {
+  int row;
+  int col;
 } EC_POS;
 
 /*
@@ -35,56 +34,45 @@
  * W = round((1./(sqrt(x.^2 + y.^2))*2^7));
  * W(1,1) = 0;
  */
-static const int weights_q7[5][5] = {
-       {  0,   128,    64,    43,    32 },
-       {128,    91,    57,    40,    31 },
-       { 64,    57,    45,    36,    29 },
-       { 43,    40,    36,    30,    26 },
-       { 32,    31,    29,    26,    23 }
-};
+static const int weights_q7[5][5] = { { 0, 128, 64, 43, 32 },
+                                      { 128, 91, 57, 40, 31 },
+                                      { 64, 57, 45, 36, 29 },
+                                      { 43, 40, 36, 30, 26 },
+                                      { 32, 31, 29, 26, 23 } };
 
-int vp8_alloc_overlap_lists(VP8D_COMP *pbi)
-{
-    if (pbi->overlaps != NULL)
-    {
-        vpx_free(pbi->overlaps);
-        pbi->overlaps = NULL;
-    }
+int vp8_alloc_overlap_lists(VP8D_COMP *pbi) {
+  if (pbi->overlaps != NULL) {
+    vpx_free(pbi->overlaps);
+    pbi->overlaps = NULL;
+  }
 
-    pbi->overlaps = vpx_calloc(pbi->common.mb_rows * pbi->common.mb_cols,
-                               sizeof(MB_OVERLAP));
+  pbi->overlaps =
+      vpx_calloc(pbi->common.mb_rows * pbi->common.mb_cols, sizeof(MB_OVERLAP));
 
-    if (pbi->overlaps == NULL)
-        return -1;
+  if (pbi->overlaps == NULL) return -1;
 
-    return 0;
+  return 0;
 }
 
-void vp8_de_alloc_overlap_lists(VP8D_COMP *pbi)
-{
-    vpx_free(pbi->overlaps);
-    pbi->overlaps = NULL;
+void vp8_de_alloc_overlap_lists(VP8D_COMP *pbi) {
+  vpx_free(pbi->overlaps);
+  pbi->overlaps = NULL;
 }
 
 /* Inserts a new overlap area value to the list of overlaps of a block */
-static void assign_overlap(OVERLAP_NODE* overlaps,
-                           union b_mode_info *bmi,
-                           int overlap)
-{
-    int i;
-    if (overlap <= 0)
-        return;
-    /* Find and assign to the next empty overlap node in the list of overlaps.
-     * Empty is defined as bmi == NULL */
-    for (i = 0; i < MAX_OVERLAPS; i++)
-    {
-        if (overlaps[i].bmi == NULL)
-        {
-            overlaps[i].bmi = bmi;
-            overlaps[i].overlap = overlap;
-            break;
-        }
+static void assign_overlap(OVERLAP_NODE *overlaps, union b_mode_info *bmi,
+                           int overlap) {
+  int i;
+  if (overlap <= 0) return;
+  /* Find and assign to the next empty overlap node in the list of overlaps.
+   * Empty is defined as bmi == NULL */
+  for (i = 0; i < MAX_OVERLAPS; i++) {
+    if (overlaps[i].bmi == NULL) {
+      overlaps[i].bmi = bmi;
+      overlaps[i].overlap = overlap;
+      break;
     }
+  }
 }
 
 /* Calculates the overlap area between two 4x4 squares, where the first
@@ -92,16 +80,16 @@
  * square has its upper-left corner at (b2_row, b2_col). Doesn't
  * properly handle squares which do not overlap.
  */
-static int block_overlap(int b1_row, int b1_col, int b2_row, int b2_col)
-{
-    const int int_top = VPXMAX(b1_row, b2_row); // top
-    const int int_left = VPXMAX(b1_col, b2_col); // left
-    /* Since each block is 4x4 pixels, adding 4 (Q3) to the left/top edge
-     * gives us the right/bottom edge.
-     */
-    const int int_right = VPXMIN(b1_col + (4<<3), b2_col + (4<<3)); // right
-    const int int_bottom = VPXMIN(b1_row + (4<<3), b2_row + (4<<3)); // bottom
-    return (int_bottom - int_top) * (int_right - int_left);
+static int block_overlap(int b1_row, int b1_col, int b2_row, int b2_col) {
+  const int int_top = VPXMAX(b1_row, b2_row);   // top
+  const int int_left = VPXMAX(b1_col, b2_col);  // left
+  /* Since each block is 4x4 pixels, adding 4 (Q3) to the left/top edge
+   * gives us the right/bottom edge.
+   */
+  const int int_right = VPXMIN(b1_col + (4 << 3), b2_col + (4 << 3));  // right
+  const int int_bottom =
+      VPXMIN(b1_row + (4 << 3), b2_row + (4 << 3));  // bottom
+  return (int_bottom - int_top) * (int_right - int_left);
 }
 
 /* Calculates the overlap area for all blocks in a macroblock at position
@@ -111,134 +99,112 @@
  * first_blk_col) in blocks relative the upper-left corner of the image.
  */
 static void calculate_overlaps_mb(B_OVERLAP *b_overlaps, union b_mode_info *bmi,
-                                  int new_row, int new_col,
-                                  int mb_row, int mb_col,
-                                  int first_blk_row, int first_blk_col)
-{
-    /* Find the blocks within this MB (defined by mb_row, mb_col) which are
-     * overlapped by bmi and calculate and assign overlap for each of those
-     * blocks. */
+                                  int new_row, int new_col, int mb_row,
+                                  int mb_col, int first_blk_row,
+                                  int first_blk_col) {
+  /* Find the blocks within this MB (defined by mb_row, mb_col) which are
+   * overlapped by bmi and calculate and assign overlap for each of those
+   * blocks. */
 
-    /* Block coordinates relative the upper-left block */
-    const int rel_ol_blk_row = first_blk_row - mb_row * 4;
-    const int rel_ol_blk_col = first_blk_col - mb_col * 4;
-    /* If the block partly overlaps any previous MB, these coordinates
-     * can be < 0. We don't want to access blocks in previous MBs.
-     */
-    const int blk_idx = VPXMAX(rel_ol_blk_row,0) * 4 + VPXMAX(rel_ol_blk_col,0);
-    /* Upper left overlapping block */
-    B_OVERLAP *b_ol_ul = &(b_overlaps[blk_idx]);
+  /* Block coordinates relative the upper-left block */
+  const int rel_ol_blk_row = first_blk_row - mb_row * 4;
+  const int rel_ol_blk_col = first_blk_col - mb_col * 4;
+  /* If the block partly overlaps any previous MB, these coordinates
+   * can be < 0. We don't want to access blocks in previous MBs.
+   */
+  const int blk_idx = VPXMAX(rel_ol_blk_row, 0) * 4 + VPXMAX(rel_ol_blk_col, 0);
+  /* Upper left overlapping block */
+  B_OVERLAP *b_ol_ul = &(b_overlaps[blk_idx]);
 
-    /* Calculate and assign overlaps for all blocks in this MB
-     * which the motion compensated block overlaps
-     */
-    /* Avoid calculating overlaps for blocks in later MBs */
-    int end_row = VPXMIN(4 + mb_row * 4 - first_blk_row, 2);
-    int end_col = VPXMIN(4 + mb_col * 4 - first_blk_col, 2);
-    int row, col;
+  /* Calculate and assign overlaps for all blocks in this MB
+   * which the motion compensated block overlaps
+   */
+  /* Avoid calculating overlaps for blocks in later MBs */
+  int end_row = VPXMIN(4 + mb_row * 4 - first_blk_row, 2);
+  int end_col = VPXMIN(4 + mb_col * 4 - first_blk_col, 2);
+  int row, col;
 
-    /* Check if new_row and new_col are evenly divisible by 4 (Q3),
-     * and if so we shouldn't check neighboring blocks
-     */
-    if (new_row >= 0 && (new_row & 0x1F) == 0)
-        end_row = 1;
-    if (new_col >= 0 && (new_col & 0x1F) == 0)
-        end_col = 1;
+  /* Check if new_row and new_col are evenly divisible by 4 (Q3),
+   * and if so we shouldn't check neighboring blocks
+   */
+  if (new_row >= 0 && (new_row & 0x1F) == 0) end_row = 1;
+  if (new_col >= 0 && (new_col & 0x1F) == 0) end_col = 1;
 
-    /* Check if the overlapping block partly overlaps a previous MB
-     * and if so, we're overlapping fewer blocks in this MB.
-     */
-    if (new_row < (mb_row*16)<<3)
-        end_row = 1;
-    if (new_col < (mb_col*16)<<3)
-        end_col = 1;
+  /* Check if the overlapping block partly overlaps a previous MB
+   * and if so, we're overlapping fewer blocks in this MB.
+   */
+  if (new_row < (mb_row * 16) << 3) end_row = 1;
+  if (new_col < (mb_col * 16) << 3) end_col = 1;
 
-    for (row = 0; row < end_row; ++row)
-    {
-        for (col = 0; col < end_col; ++col)
-        {
-            /* input in Q3, result in Q6 */
-            const int overlap = block_overlap(new_row, new_col,
-                                                  (((first_blk_row + row) *
-                                                      4) << 3),
-                                                  (((first_blk_col + col) *
-                                                      4) << 3));
-            assign_overlap(b_ol_ul[row * 4 + col].overlaps, bmi, overlap);
-        }
+  for (row = 0; row < end_row; ++row) {
+    for (col = 0; col < end_col; ++col) {
+      /* input in Q3, result in Q6 */
+      const int overlap =
+          block_overlap(new_row, new_col, (((first_blk_row + row) * 4) << 3),
+                        (((first_blk_col + col) * 4) << 3));
+      assign_overlap(b_ol_ul[row * 4 + col].overlaps, bmi, overlap);
     }
+  }
 }
 
-void vp8_calculate_overlaps(MB_OVERLAP *overlap_ul,
-                            int mb_rows, int mb_cols,
-                            union b_mode_info *bmi,
-                            int b_row, int b_col)
-{
-    MB_OVERLAP *mb_overlap;
-    int row, col, rel_row, rel_col;
-    int new_row, new_col;
-    int end_row, end_col;
-    int overlap_b_row, overlap_b_col;
-    int overlap_mb_row, overlap_mb_col;
+void vp8_calculate_overlaps(MB_OVERLAP *overlap_ul, int mb_rows, int mb_cols,
+                            union b_mode_info *bmi, int b_row, int b_col) {
+  MB_OVERLAP *mb_overlap;
+  int row, col, rel_row, rel_col;
+  int new_row, new_col;
+  int end_row, end_col;
+  int overlap_b_row, overlap_b_col;
+  int overlap_mb_row, overlap_mb_col;
 
-    /* mb subpixel position */
-    row = (4 * b_row) << 3; /* Q3 */
-    col = (4 * b_col) << 3; /* Q3 */
+  /* mb subpixel position */
+  row = (4 * b_row) << 3; /* Q3 */
+  col = (4 * b_col) << 3; /* Q3 */
 
-    /* reverse compensate for motion */
-    new_row = row - bmi->mv.as_mv.row;
-    new_col = col - bmi->mv.as_mv.col;
+  /* reverse compensate for motion */
+  new_row = row - bmi->mv.as_mv.row;
+  new_col = col - bmi->mv.as_mv.col;
 
-    if (new_row >= ((16*mb_rows) << 3) || new_col >= ((16*mb_cols) << 3))
-    {
-        /* the new block ended up outside the frame */
-        return;
-    }
+  if (new_row >= ((16 * mb_rows) << 3) || new_col >= ((16 * mb_cols) << 3)) {
+    /* the new block ended up outside the frame */
+    return;
+  }
 
-    if (new_row <= -32 || new_col <= -32)
-    {
-        /* outside the frame */
-        return;
-    }
-    /* overlapping block's position in blocks */
-    overlap_b_row = FLOOR(new_row / 4, 3) >> 3;
-    overlap_b_col = FLOOR(new_col / 4, 3) >> 3;
+  if (new_row <= -32 || new_col <= -32) {
+    /* outside the frame */
+    return;
+  }
+  /* overlapping block's position in blocks */
+  overlap_b_row = FLOOR(new_row / 4, 3) >> 3;
+  overlap_b_col = FLOOR(new_col / 4, 3) >> 3;
 
-    /* overlapping block's MB position in MBs
-     * operations are done in Q3
-     */
-    overlap_mb_row = FLOOR((overlap_b_row << 3) / 4, 3) >> 3;
-    overlap_mb_col = FLOOR((overlap_b_col << 3) / 4, 3) >> 3;
+  /* overlapping block's MB position in MBs
+   * operations are done in Q3
+   */
+  overlap_mb_row = FLOOR((overlap_b_row << 3) / 4, 3) >> 3;
+  overlap_mb_col = FLOOR((overlap_b_col << 3) / 4, 3) >> 3;
 
-    end_row = VPXMIN(mb_rows - overlap_mb_row, 2);
-    end_col = VPXMIN(mb_cols - overlap_mb_col, 2);
+  end_row = VPXMIN(mb_rows - overlap_mb_row, 2);
+  end_col = VPXMIN(mb_cols - overlap_mb_col, 2);
 
-    /* Don't calculate overlap for MBs we don't overlap */
-    /* Check if the new block row starts at the last block row of the MB */
-    if (abs(new_row - ((16*overlap_mb_row) << 3)) < ((3*4) << 3))
-        end_row = 1;
-    /* Check if the new block col starts at the last block col of the MB */
-    if (abs(new_col - ((16*overlap_mb_col) << 3)) < ((3*4) << 3))
-        end_col = 1;
+  /* Don't calculate overlap for MBs we don't overlap */
+  /* Check if the new block row starts at the last block row of the MB */
+  if (abs(new_row - ((16 * overlap_mb_row) << 3)) < ((3 * 4) << 3)) end_row = 1;
+  /* Check if the new block col starts at the last block col of the MB */
+  if (abs(new_col - ((16 * overlap_mb_col) << 3)) < ((3 * 4) << 3)) end_col = 1;
 
-    /* find the MB(s) this block is overlapping */
-    for (rel_row = 0; rel_row < end_row; ++rel_row)
-    {
-        for (rel_col = 0; rel_col < end_col; ++rel_col)
-        {
-            if (overlap_mb_row + rel_row < 0 ||
-                overlap_mb_col + rel_col < 0)
-                continue;
-            mb_overlap = overlap_ul + (overlap_mb_row + rel_row) * mb_cols +
-                 overlap_mb_col + rel_col;
+  /* find the MB(s) this block is overlapping */
+  for (rel_row = 0; rel_row < end_row; ++rel_row) {
+    for (rel_col = 0; rel_col < end_col; ++rel_col) {
+      if (overlap_mb_row + rel_row < 0 || overlap_mb_col + rel_col < 0)
+        continue;
+      mb_overlap = overlap_ul + (overlap_mb_row + rel_row) * mb_cols +
+                   overlap_mb_col + rel_col;
 
-            calculate_overlaps_mb(mb_overlap->overlaps, bmi,
-                                  new_row, new_col,
-                                  overlap_mb_row + rel_row,
-                                  overlap_mb_col + rel_col,
-                                  overlap_b_row + rel_row,
-                                  overlap_b_col + rel_col);
-        }
+      calculate_overlaps_mb(mb_overlap->overlaps, bmi, new_row, new_col,
+                            overlap_mb_row + rel_row, overlap_mb_col + rel_col,
+                            overlap_b_row + rel_row, overlap_b_col + rel_col);
     }
+  }
 }
 
 /* Estimates a motion vector given the overlapping blocks' motion vectors.
@@ -245,180 +211,142 @@
  * Filters out all overlapping blocks which do not refer to the correct
  * reference frame type.
  */
-static void estimate_mv(const OVERLAP_NODE *overlaps, union b_mode_info *bmi)
-{
-    int i;
-    int overlap_sum = 0;
-    int row_acc = 0;
-    int col_acc = 0;
+static void estimate_mv(const OVERLAP_NODE *overlaps, union b_mode_info *bmi) {
+  int i;
+  int overlap_sum = 0;
+  int row_acc = 0;
+  int col_acc = 0;
 
-    bmi->mv.as_int = 0;
-    for (i=0; i < MAX_OVERLAPS; ++i)
-    {
-        if (overlaps[i].bmi == NULL)
-            break;
-        col_acc += overlaps[i].overlap * overlaps[i].bmi->mv.as_mv.col;
-        row_acc += overlaps[i].overlap * overlaps[i].bmi->mv.as_mv.row;
-        overlap_sum += overlaps[i].overlap;
-    }
-    if (overlap_sum > 0)
-    {
-        /* Q9 / Q6 = Q3 */
-        bmi->mv.as_mv.col = col_acc / overlap_sum;
-        bmi->mv.as_mv.row = row_acc / overlap_sum;
-    }
-    else
-    {
-        bmi->mv.as_mv.col = 0;
-        bmi->mv.as_mv.row = 0;
-    }
+  bmi->mv.as_int = 0;
+  for (i = 0; i < MAX_OVERLAPS; ++i) {
+    if (overlaps[i].bmi == NULL) break;
+    col_acc += overlaps[i].overlap * overlaps[i].bmi->mv.as_mv.col;
+    row_acc += overlaps[i].overlap * overlaps[i].bmi->mv.as_mv.row;
+    overlap_sum += overlaps[i].overlap;
+  }
+  if (overlap_sum > 0) {
+    /* Q9 / Q6 = Q3 */
+    bmi->mv.as_mv.col = col_acc / overlap_sum;
+    bmi->mv.as_mv.row = row_acc / overlap_sum;
+  } else {
+    bmi->mv.as_mv.col = 0;
+    bmi->mv.as_mv.row = 0;
+  }
 }
 
 /* Estimates all motion vectors for a macroblock given the lists of
  * overlaps for each block. Decides whether or not the MVs must be clamped.
  */
-static void estimate_mb_mvs(const B_OVERLAP *block_overlaps,
-                            MODE_INFO *mi,
-                            int mb_to_left_edge,
-                            int mb_to_right_edge,
-                            int mb_to_top_edge,
-                            int mb_to_bottom_edge)
-{
-    int row, col;
-    int non_zero_count = 0;
-    MV * const filtered_mv = &(mi->mbmi.mv.as_mv);
-    union b_mode_info * const bmi = mi->bmi;
-    filtered_mv->col = 0;
-    filtered_mv->row = 0;
-    mi->mbmi.need_to_clamp_mvs = 0;
-    for (row = 0; row < 4; ++row)
-    {
-        int this_b_to_top_edge = mb_to_top_edge + ((row*4)<<3);
-        int this_b_to_bottom_edge = mb_to_bottom_edge - ((row*4)<<3);
-        for (col = 0; col < 4; ++col)
-        {
-            int i = row * 4 + col;
-            int this_b_to_left_edge = mb_to_left_edge + ((col*4)<<3);
-            int this_b_to_right_edge = mb_to_right_edge - ((col*4)<<3);
-            /* Estimate vectors for all blocks which are overlapped by this */
-            /* type. Interpolate/extrapolate the rest of the block's MVs */
-            estimate_mv(block_overlaps[i].overlaps, &(bmi[i]));
-            mi->mbmi.need_to_clamp_mvs |= vp8_check_mv_bounds(
-                                                         &bmi[i].mv,
-                                                         this_b_to_left_edge,
-                                                         this_b_to_right_edge,
-                                                         this_b_to_top_edge,
-                                                         this_b_to_bottom_edge);
-            if (bmi[i].mv.as_int != 0)
-            {
-                ++non_zero_count;
-                filtered_mv->col += bmi[i].mv.as_mv.col;
-                filtered_mv->row += bmi[i].mv.as_mv.row;
-            }
-        }
+static void estimate_mb_mvs(const B_OVERLAP *block_overlaps, MODE_INFO *mi,
+                            int mb_to_left_edge, int mb_to_right_edge,
+                            int mb_to_top_edge, int mb_to_bottom_edge) {
+  int row, col;
+  int non_zero_count = 0;
+  MV *const filtered_mv = &(mi->mbmi.mv.as_mv);
+  union b_mode_info *const bmi = mi->bmi;
+  filtered_mv->col = 0;
+  filtered_mv->row = 0;
+  mi->mbmi.need_to_clamp_mvs = 0;
+  for (row = 0; row < 4; ++row) {
+    int this_b_to_top_edge = mb_to_top_edge + ((row * 4) << 3);
+    int this_b_to_bottom_edge = mb_to_bottom_edge - ((row * 4) << 3);
+    for (col = 0; col < 4; ++col) {
+      int i = row * 4 + col;
+      int this_b_to_left_edge = mb_to_left_edge + ((col * 4) << 3);
+      int this_b_to_right_edge = mb_to_right_edge - ((col * 4) << 3);
+      /* Estimate vectors for all blocks which are overlapped by this */
+      /* type. Interpolate/extrapolate the rest of the block's MVs */
+      estimate_mv(block_overlaps[i].overlaps, &(bmi[i]));
+      mi->mbmi.need_to_clamp_mvs |= vp8_check_mv_bounds(
+          &bmi[i].mv, this_b_to_left_edge, this_b_to_right_edge,
+          this_b_to_top_edge, this_b_to_bottom_edge);
+      if (bmi[i].mv.as_int != 0) {
+        ++non_zero_count;
+        filtered_mv->col += bmi[i].mv.as_mv.col;
+        filtered_mv->row += bmi[i].mv.as_mv.row;
+      }
     }
-    if (non_zero_count > 0)
-    {
-        filtered_mv->col /= non_zero_count;
-        filtered_mv->row /= non_zero_count;
-    }
+  }
+  if (non_zero_count > 0) {
+    filtered_mv->col /= non_zero_count;
+    filtered_mv->row /= non_zero_count;
+  }
 }
 
 static void calc_prev_mb_overlaps(MB_OVERLAP *overlaps, MODE_INFO *prev_mi,
-                                    int mb_row, int mb_col,
-                                    int mb_rows, int mb_cols)
-{
-    int sub_row;
-    int sub_col;
-    for (sub_row = 0; sub_row < 4; ++sub_row)
-    {
-        for (sub_col = 0; sub_col < 4; ++sub_col)
-        {
-            vp8_calculate_overlaps(
-                                overlaps, mb_rows, mb_cols,
-                                &(prev_mi->bmi[sub_row * 4 + sub_col]),
-                                4 * mb_row + sub_row,
-                                4 * mb_col + sub_col);
-        }
+                                  int mb_row, int mb_col, int mb_rows,
+                                  int mb_cols) {
+  int sub_row;
+  int sub_col;
+  for (sub_row = 0; sub_row < 4; ++sub_row) {
+    for (sub_col = 0; sub_col < 4; ++sub_col) {
+      vp8_calculate_overlaps(overlaps, mb_rows, mb_cols,
+                             &(prev_mi->bmi[sub_row * 4 + sub_col]),
+                             4 * mb_row + sub_row, 4 * mb_col + sub_col);
     }
+  }
 }
 
 /* Estimate all missing motion vectors. This function does the same as the one
  * above, but has different input arguments. */
-static void estimate_missing_mvs(MB_OVERLAP *overlaps,
-                                 MODE_INFO *mi, MODE_INFO *prev_mi,
-                                 int mb_rows, int mb_cols,
-                                 unsigned int first_corrupt)
-{
-    int mb_row, mb_col;
-    memset(overlaps, 0, sizeof(MB_OVERLAP) * mb_rows * mb_cols);
-    /* First calculate the overlaps for all blocks */
-    for (mb_row = 0; mb_row < mb_rows; ++mb_row)
-    {
-        for (mb_col = 0; mb_col < mb_cols; ++mb_col)
-        {
-            /* We're only able to use blocks referring to the last frame
-             * when extrapolating new vectors.
-             */
-            if (prev_mi->mbmi.ref_frame == LAST_FRAME)
-            {
-                calc_prev_mb_overlaps(overlaps, prev_mi,
-                                      mb_row, mb_col,
-                                      mb_rows, mb_cols);
-            }
-            ++prev_mi;
-        }
-        ++prev_mi;
+static void estimate_missing_mvs(MB_OVERLAP *overlaps, MODE_INFO *mi,
+                                 MODE_INFO *prev_mi, int mb_rows, int mb_cols,
+                                 unsigned int first_corrupt) {
+  int mb_row, mb_col;
+  memset(overlaps, 0, sizeof(MB_OVERLAP) * mb_rows * mb_cols);
+  /* First calculate the overlaps for all blocks */
+  for (mb_row = 0; mb_row < mb_rows; ++mb_row) {
+    for (mb_col = 0; mb_col < mb_cols; ++mb_col) {
+      /* We're only able to use blocks referring to the last frame
+       * when extrapolating new vectors.
+       */
+      if (prev_mi->mbmi.ref_frame == LAST_FRAME) {
+        calc_prev_mb_overlaps(overlaps, prev_mi, mb_row, mb_col, mb_rows,
+                              mb_cols);
+      }
+      ++prev_mi;
     }
+    ++prev_mi;
+  }
 
-    mb_row = first_corrupt / mb_cols;
-    mb_col = first_corrupt - mb_row * mb_cols;
-    mi += mb_row*(mb_cols + 1) + mb_col;
-    /* Go through all macroblocks in the current image with missing MVs
-     * and calculate new MVs using the overlaps.
-     */
-    for (; mb_row < mb_rows; ++mb_row)
-    {
-        int mb_to_top_edge = -((mb_row * 16)) << 3;
-        int mb_to_bottom_edge = ((mb_rows - 1 - mb_row) * 16) << 3;
-        for (; mb_col < mb_cols; ++mb_col)
-        {
-            int mb_to_left_edge = -((mb_col * 16) << 3);
-            int mb_to_right_edge = ((mb_cols - 1 - mb_col) * 16) << 3;
-            const B_OVERLAP *block_overlaps =
-                    overlaps[mb_row*mb_cols + mb_col].overlaps;
-            mi->mbmi.ref_frame = LAST_FRAME;
-            mi->mbmi.mode = SPLITMV;
-            mi->mbmi.uv_mode = DC_PRED;
-            mi->mbmi.partitioning = 3;
-            mi->mbmi.segment_id = 0;
-            estimate_mb_mvs(block_overlaps,
-                            mi,
-                            mb_to_left_edge,
-                            mb_to_right_edge,
-                            mb_to_top_edge,
-                            mb_to_bottom_edge);
-            ++mi;
-        }
-        mb_col = 0;
-        ++mi;
+  mb_row = first_corrupt / mb_cols;
+  mb_col = first_corrupt - mb_row * mb_cols;
+  mi += mb_row * (mb_cols + 1) + mb_col;
+  /* Go through all macroblocks in the current image with missing MVs
+   * and calculate new MVs using the overlaps.
+   */
+  for (; mb_row < mb_rows; ++mb_row) {
+    int mb_to_top_edge = -((mb_row * 16)) << 3;
+    int mb_to_bottom_edge = ((mb_rows - 1 - mb_row) * 16) << 3;
+    for (; mb_col < mb_cols; ++mb_col) {
+      int mb_to_left_edge = -((mb_col * 16) << 3);
+      int mb_to_right_edge = ((mb_cols - 1 - mb_col) * 16) << 3;
+      const B_OVERLAP *block_overlaps =
+          overlaps[mb_row * mb_cols + mb_col].overlaps;
+      mi->mbmi.ref_frame = LAST_FRAME;
+      mi->mbmi.mode = SPLITMV;
+      mi->mbmi.uv_mode = DC_PRED;
+      mi->mbmi.partitioning = 3;
+      mi->mbmi.segment_id = 0;
+      estimate_mb_mvs(block_overlaps, mi, mb_to_left_edge, mb_to_right_edge,
+                      mb_to_top_edge, mb_to_bottom_edge);
+      ++mi;
     }
+    mb_col = 0;
+    ++mi;
+  }
 }
 
-void vp8_estimate_missing_mvs(VP8D_COMP *pbi)
-{
-    VP8_COMMON * const pc = &pbi->common;
-    estimate_missing_mvs(pbi->overlaps,
-                         pc->mi, pc->prev_mi,
-                         pc->mb_rows, pc->mb_cols,
-                         pbi->mvs_corrupt_from_mb);
+void vp8_estimate_missing_mvs(VP8D_COMP *pbi) {
+  VP8_COMMON *const pc = &pbi->common;
+  estimate_missing_mvs(pbi->overlaps, pc->mi, pc->prev_mi, pc->mb_rows,
+                       pc->mb_cols, pbi->mvs_corrupt_from_mb);
 }
 
-static void assign_neighbor(EC_BLOCK *neighbor, MODE_INFO *mi, int block_idx)
-{
-    assert(mi->mbmi.ref_frame < MAX_REF_FRAMES);
-    neighbor->ref_frame = mi->mbmi.ref_frame;
-    neighbor->mv = mi->bmi[block_idx].mv.as_mv;
+static void assign_neighbor(EC_BLOCK *neighbor, MODE_INFO *mi, int block_idx) {
+  assert(mi->mbmi.ref_frame < MAX_REF_FRAMES);
+  neighbor->ref_frame = mi->mbmi.ref_frame;
+  neighbor->mv = mi->bmi[block_idx].mv.as_mv;
 }
 
 /* Finds the neighboring blocks of a macroblocks. In the general case
@@ -429,158 +357,126 @@
  * neighbor, and so on. The last element refers to the neighbor below the first
  * neighbor.
  */
-static void find_neighboring_blocks(MODE_INFO *mi,
-                                    EC_BLOCK *neighbors,
-                                    int mb_row, int mb_col,
-                                    int mb_rows, int mb_cols,
-                                    int mi_stride)
-{
-    int i = 0;
-    int j;
-    if (mb_row > 0)
-    {
-        /* upper left */
-        if (mb_col > 0)
-            assign_neighbor(&neighbors[i], mi - mi_stride - 1, 15);
-        ++i;
-        /* above */
-        for (j = 12; j < 16; ++j, ++i)
-            assign_neighbor(&neighbors[i], mi - mi_stride, j);
-    }
-    else
-        i += 5;
+static void find_neighboring_blocks(MODE_INFO *mi, EC_BLOCK *neighbors,
+                                    int mb_row, int mb_col, int mb_rows,
+                                    int mb_cols, int mi_stride) {
+  int i = 0;
+  int j;
+  if (mb_row > 0) {
+    /* upper left */
+    if (mb_col > 0) assign_neighbor(&neighbors[i], mi - mi_stride - 1, 15);
+    ++i;
+    /* above */
+    for (j = 12; j < 16; ++j, ++i)
+      assign_neighbor(&neighbors[i], mi - mi_stride, j);
+  } else
+    i += 5;
+  if (mb_col < mb_cols - 1) {
+    /* upper right */
+    if (mb_row > 0) assign_neighbor(&neighbors[i], mi - mi_stride + 1, 12);
+    ++i;
+    /* right */
+    for (j = 0; j <= 12; j += 4, ++i) assign_neighbor(&neighbors[i], mi + 1, j);
+  } else
+    i += 5;
+  if (mb_row < mb_rows - 1) {
+    /* lower right */
     if (mb_col < mb_cols - 1)
-    {
-        /* upper right */
-        if (mb_row > 0)
-            assign_neighbor(&neighbors[i], mi - mi_stride + 1, 12);
-        ++i;
-        /* right */
-        for (j = 0; j <= 12; j += 4, ++i)
-            assign_neighbor(&neighbors[i], mi + 1, j);
-    }
-    else
-        i += 5;
+      assign_neighbor(&neighbors[i], mi + mi_stride + 1, 0);
+    ++i;
+    /* below */
+    for (j = 0; j < 4; ++j, ++i)
+      assign_neighbor(&neighbors[i], mi + mi_stride, j);
+  } else
+    i += 5;
+  if (mb_col > 0) {
+    /* lower left */
     if (mb_row < mb_rows - 1)
-    {
-        /* lower right */
-        if (mb_col < mb_cols - 1)
-            assign_neighbor(&neighbors[i], mi + mi_stride + 1, 0);
-        ++i;
-        /* below */
-        for (j = 0; j < 4; ++j, ++i)
-            assign_neighbor(&neighbors[i], mi + mi_stride, j);
+      assign_neighbor(&neighbors[i], mi + mi_stride - 1, 4);
+    ++i;
+    /* left */
+    for (j = 3; j < 16; j += 4, ++i) {
+      assign_neighbor(&neighbors[i], mi - 1, j);
     }
-    else
-        i += 5;
-    if (mb_col > 0)
-    {
-        /* lower left */
-        if (mb_row < mb_rows - 1)
-            assign_neighbor(&neighbors[i], mi + mi_stride - 1, 4);
-        ++i;
-        /* left */
-        for (j = 3; j < 16; j += 4, ++i)
-        {
-            assign_neighbor(&neighbors[i], mi - 1, j);
-        }
-    }
-    else
-        i += 5;
-    assert(i == 20);
+  } else
+    i += 5;
+  assert(i == 20);
 }
 
 /* Interpolates all motion vectors for a macroblock from the neighboring blocks'
  * motion vectors.
  */
-static void interpolate_mvs(MACROBLOCKD *mb,
-                         EC_BLOCK *neighbors,
-                         MV_REFERENCE_FRAME dom_ref_frame)
-{
-    int row, col, i;
-    MODE_INFO * const mi = mb->mode_info_context;
-    /* Table with the position of the neighboring blocks relative the position
-     * of the upper left block of the current MB. Starting with the upper left
-     * neighbor and going to the right.
-     */
-    const EC_POS neigh_pos[NUM_NEIGHBORS] = {
-                                        {-1,-1}, {-1,0}, {-1,1}, {-1,2}, {-1,3},
-                                        {-1,4}, {0,4}, {1,4}, {2,4}, {3,4},
-                                        {4,4}, {4,3}, {4,2}, {4,1}, {4,0},
-                                        {4,-1}, {3,-1}, {2,-1}, {1,-1}, {0,-1}
-                                      };
-    mi->mbmi.need_to_clamp_mvs = 0;
-    for (row = 0; row < 4; ++row)
-    {
-        int mb_to_top_edge = mb->mb_to_top_edge + ((row*4)<<3);
-        int mb_to_bottom_edge = mb->mb_to_bottom_edge - ((row*4)<<3);
-        for (col = 0; col < 4; ++col)
-        {
-            int mb_to_left_edge = mb->mb_to_left_edge + ((col*4)<<3);
-            int mb_to_right_edge = mb->mb_to_right_edge - ((col*4)<<3);
-            int w_sum = 0;
-            int mv_row_sum = 0;
-            int mv_col_sum = 0;
-            int_mv * const mv = &(mi->bmi[row*4 + col].mv);
-            mv->as_int = 0;
-            for (i = 0; i < NUM_NEIGHBORS; ++i)
-            {
-                /* Calculate the weighted sum of neighboring MVs referring
-                 * to the dominant frame type.
-                 */
-                const int w = weights_q7[abs(row - neigh_pos[i].row)]
-                                        [abs(col - neigh_pos[i].col)];
-                if (neighbors[i].ref_frame != dom_ref_frame)
-                    continue;
-                w_sum += w;
-                /* Q7 * Q3 = Q10 */
-                mv_row_sum += w*neighbors[i].mv.row;
-                mv_col_sum += w*neighbors[i].mv.col;
-            }
-            if (w_sum > 0)
-            {
-                /* Avoid division by zero.
-                 * Normalize with the sum of the coefficients
-                 * Q3 = Q10 / Q7
-                 */
-                mv->as_mv.row = mv_row_sum / w_sum;
-                mv->as_mv.col = mv_col_sum / w_sum;
-                mi->mbmi.need_to_clamp_mvs |= vp8_check_mv_bounds(
-                                                            mv,
-                                                            mb_to_left_edge,
-                                                            mb_to_right_edge,
-                                                            mb_to_top_edge,
-                                                            mb_to_bottom_edge);
-            }
-        }
+static void interpolate_mvs(MACROBLOCKD *mb, EC_BLOCK *neighbors,
+                            MV_REFERENCE_FRAME dom_ref_frame) {
+  int row, col, i;
+  MODE_INFO *const mi = mb->mode_info_context;
+  /* Table with the position of the neighboring blocks relative the position
+   * of the upper left block of the current MB. Starting with the upper left
+   * neighbor and going to the right.
+   */
+  const EC_POS neigh_pos[NUM_NEIGHBORS] = {
+    { -1, -1 }, { -1, 0 }, { -1, 1 }, { -1, 2 }, { -1, 3 }, { -1, 4 }, { 0, 4 },
+    { 1, 4 },   { 2, 4 },  { 3, 4 },  { 4, 4 },  { 4, 3 },  { 4, 2 },  { 4, 1 },
+    { 4, 0 },   { 4, -1 }, { 3, -1 }, { 2, -1 }, { 1, -1 }, { 0, -1 }
+  };
+  mi->mbmi.need_to_clamp_mvs = 0;
+  for (row = 0; row < 4; ++row) {
+    int mb_to_top_edge = mb->mb_to_top_edge + ((row * 4) << 3);
+    int mb_to_bottom_edge = mb->mb_to_bottom_edge - ((row * 4) << 3);
+    for (col = 0; col < 4; ++col) {
+      int mb_to_left_edge = mb->mb_to_left_edge + ((col * 4) << 3);
+      int mb_to_right_edge = mb->mb_to_right_edge - ((col * 4) << 3);
+      int w_sum = 0;
+      int mv_row_sum = 0;
+      int mv_col_sum = 0;
+      int_mv *const mv = &(mi->bmi[row * 4 + col].mv);
+      mv->as_int = 0;
+      for (i = 0; i < NUM_NEIGHBORS; ++i) {
+        /* Calculate the weighted sum of neighboring MVs referring
+         * to the dominant frame type.
+         */
+        const int w = weights_q7[abs(row - neigh_pos[i].row)][abs(
+            col - neigh_pos[i].col)];
+        if (neighbors[i].ref_frame != dom_ref_frame) continue;
+        w_sum += w;
+        /* Q7 * Q3 = Q10 */
+        mv_row_sum += w * neighbors[i].mv.row;
+        mv_col_sum += w * neighbors[i].mv.col;
+      }
+      if (w_sum > 0) {
+        /* Avoid division by zero.
+         * Normalize with the sum of the coefficients
+         * Q3 = Q10 / Q7
+         */
+        mv->as_mv.row = mv_row_sum / w_sum;
+        mv->as_mv.col = mv_col_sum / w_sum;
+        mi->mbmi.need_to_clamp_mvs |=
+            vp8_check_mv_bounds(mv, mb_to_left_edge, mb_to_right_edge,
+                                mb_to_top_edge, mb_to_bottom_edge);
+      }
     }
+  }
 }
 
-void vp8_interpolate_motion(MACROBLOCKD *mb,
-                        int mb_row, int mb_col,
-                        int mb_rows, int mb_cols)
-{
-    /* Find relevant neighboring blocks */
-    EC_BLOCK neighbors[NUM_NEIGHBORS];
-    int i;
-    /* Initialize the array. MAX_REF_FRAMES is interpreted as "doesn't exist" */
-    for (i = 0; i < NUM_NEIGHBORS; ++i)
-    {
-        neighbors[i].ref_frame = MAX_REF_FRAMES;
-        neighbors[i].mv.row = neighbors[i].mv.col = 0;
-    }
-    find_neighboring_blocks(mb->mode_info_context,
-                                neighbors,
-                                mb_row, mb_col,
-                                mb_rows, mb_cols,
-                                mb->mode_info_stride);
-    /* Interpolate MVs for the missing blocks from the surrounding
-     * blocks which refer to the last frame. */
-    interpolate_mvs(mb, neighbors, LAST_FRAME);
+void vp8_interpolate_motion(MACROBLOCKD *mb, int mb_row, int mb_col,
+                            int mb_rows, int mb_cols) {
+  /* Find relevant neighboring blocks */
+  EC_BLOCK neighbors[NUM_NEIGHBORS];
+  int i;
+  /* Initialize the array. MAX_REF_FRAMES is interpreted as "doesn't exist" */
+  for (i = 0; i < NUM_NEIGHBORS; ++i) {
+    neighbors[i].ref_frame = MAX_REF_FRAMES;
+    neighbors[i].mv.row = neighbors[i].mv.col = 0;
+  }
+  find_neighboring_blocks(mb->mode_info_context, neighbors, mb_row, mb_col,
+                          mb_rows, mb_cols, mb->mode_info_stride);
+  /* Interpolate MVs for the missing blocks from the surrounding
+   * blocks which refer to the last frame. */
+  interpolate_mvs(mb, neighbors, LAST_FRAME);
 
-    mb->mode_info_context->mbmi.ref_frame = LAST_FRAME;
-    mb->mode_info_context->mbmi.mode = SPLITMV;
-    mb->mode_info_context->mbmi.uv_mode = DC_PRED;
-    mb->mode_info_context->mbmi.partitioning = 3;
-    mb->mode_info_context->mbmi.segment_id = 0;
+  mb->mode_info_context->mbmi.ref_frame = LAST_FRAME;
+  mb->mode_info_context->mbmi.mode = SPLITMV;
+  mb->mode_info_context->mbmi.uv_mode = DC_PRED;
+  mb->mode_info_context->mbmi.partitioning = 3;
+  mb->mode_info_context->mbmi.segment_id = 0;
 }
--- a/vp8/decoder/error_concealment.h
+++ b/vp8/decoder/error_concealment.h
@@ -8,7 +8,6 @@
  *  be found in the AUTHORS file in the root of the source tree.
  */
 
-
 #ifndef VP8_DECODER_ERROR_CONCEALMENT_H_
 #define VP8_DECODER_ERROR_CONCEALMENT_H_
 
@@ -32,8 +31,7 @@
 
 /* Interpolates all motion vectors for a macroblock mb at position
  * (mb_row, mb_col). */
-void vp8_interpolate_motion(MACROBLOCKD *mb,
-                            int mb_row, int mb_col,
+void vp8_interpolate_motion(MACROBLOCKD *mb, int mb_row, int mb_col,
                             int mb_rows, int mb_cols);
 
 #ifdef __cplusplus
--- a/vp8/decoder/onyxd_if.c
+++ b/vp8/decoder/onyxd_if.c
@@ -8,7 +8,6 @@
  *  be found in the AUTHORS file in the root of the source tree.
  */
 
-
 #include "vp8/common/onyxc_int.h"
 #if CONFIG_POSTPROC
 #include "vp8/common/postproc.h"
@@ -42,480 +41,432 @@
 
 extern void vp8_init_loop_filter(VP8_COMMON *cm);
 extern void vp8cx_init_de_quantizer(VP8D_COMP *pbi);
-static int get_free_fb (VP8_COMMON *cm);
-static void ref_cnt_fb (int *buf, int *idx, int new_idx);
+static int get_free_fb(VP8_COMMON *cm);
+static void ref_cnt_fb(int *buf, int *idx, int new_idx);
 
 static void initialize_dec(void) {
-    static volatile int init_done = 0;
+  static volatile int init_done = 0;
 
-    if (!init_done)
-    {
-        vpx_dsp_rtcd();
-        vp8_init_intra_predictors();
-        init_done = 1;
-    }
+  if (!init_done) {
+    vpx_dsp_rtcd();
+    vp8_init_intra_predictors();
+    init_done = 1;
+  }
 }
 
-static void remove_decompressor(VP8D_COMP *pbi)
-{
+static void remove_decompressor(VP8D_COMP *pbi) {
 #if CONFIG_ERROR_CONCEALMENT
-    vp8_de_alloc_overlap_lists(pbi);
+  vp8_de_alloc_overlap_lists(pbi);
 #endif
-    vp8_remove_common(&pbi->common);
-    vpx_free(pbi);
+  vp8_remove_common(&pbi->common);
+  vpx_free(pbi);
 }
 
-static struct VP8D_COMP * create_decompressor(VP8D_CONFIG *oxcf)
-{
-    VP8D_COMP *pbi = vpx_memalign(32, sizeof(VP8D_COMP));
+static struct VP8D_COMP *create_decompressor(VP8D_CONFIG *oxcf) {
+  VP8D_COMP *pbi = vpx_memalign(32, sizeof(VP8D_COMP));
 
-    if (!pbi)
-        return NULL;
+  if (!pbi) return NULL;
 
-    memset(pbi, 0, sizeof(VP8D_COMP));
+  memset(pbi, 0, sizeof(VP8D_COMP));
 
-    if (setjmp(pbi->common.error.jmp))
-    {
-        pbi->common.error.setjmp = 0;
-        remove_decompressor(pbi);
-        return 0;
-    }
+  if (setjmp(pbi->common.error.jmp)) {
+    pbi->common.error.setjmp = 0;
+    remove_decompressor(pbi);
+    return 0;
+  }
 
-    pbi->common.error.setjmp = 1;
+  pbi->common.error.setjmp = 1;
 
-    vp8_create_common(&pbi->common);
+  vp8_create_common(&pbi->common);
 
-    pbi->common.current_video_frame = 0;
-    pbi->ready_for_new_data = 1;
+  pbi->common.current_video_frame = 0;
+  pbi->ready_for_new_data = 1;
 
-    /* vp8cx_init_de_quantizer() is first called here. Add check in frame_init_dequantizer() to avoid
-     *  unnecessary calling of vp8cx_init_de_quantizer() for every frame.
-     */
-    vp8cx_init_de_quantizer(pbi);
+  /* vp8cx_init_de_quantizer() is first called here. Add check in
+   * frame_init_dequantizer() to avoid
+   *  unnecessary calling of vp8cx_init_de_quantizer() for every frame.
+   */
+  vp8cx_init_de_quantizer(pbi);
 
-    vp8_loop_filter_init(&pbi->common);
+  vp8_loop_filter_init(&pbi->common);
 
-    pbi->common.error.setjmp = 0;
+  pbi->common.error.setjmp = 0;
 
 #if CONFIG_ERROR_CONCEALMENT
-    pbi->ec_enabled = oxcf->error_concealment;
-    pbi->overlaps = NULL;
+  pbi->ec_enabled = oxcf->error_concealment;
+  pbi->overlaps = NULL;
 #else
-    (void)oxcf;
-    pbi->ec_enabled = 0;
+  (void)oxcf;
+  pbi->ec_enabled = 0;
 #endif
-    /* Error concealment is activated after a key frame has been
-     * decoded without errors when error concealment is enabled.
-     */
-    pbi->ec_active = 0;
+  /* Error concealment is activated after a key frame has been
+   * decoded without errors when error concealment is enabled.
+   */
+  pbi->ec_active = 0;
 
-    pbi->decoded_key_frame = 0;
+  pbi->decoded_key_frame = 0;
 
-    /* Independent partitions is activated when a frame updates the
-     * token probability table to have equal probabilities over the
-     * PREV_COEF context.
-     */
-    pbi->independent_partitions = 0;
+  /* Independent partitions is activated when a frame updates the
+   * token probability table to have equal probabilities over the
+   * PREV_COEF context.
+   */
+  pbi->independent_partitions = 0;
 
-    vp8_setup_block_dptrs(&pbi->mb);
+  vp8_setup_block_dptrs(&pbi->mb);
 
-    once(initialize_dec);
+  once(initialize_dec);
 
-    return pbi;
+  return pbi;
 }
 
-vpx_codec_err_t vp8dx_get_reference(VP8D_COMP *pbi, enum vpx_ref_frame_type ref_frame_flag, YV12_BUFFER_CONFIG *sd)
-{
-    VP8_COMMON *cm = &pbi->common;
-    int ref_fb_idx;
+vpx_codec_err_t vp8dx_get_reference(VP8D_COMP *pbi,
+                                    enum vpx_ref_frame_type ref_frame_flag,
+                                    YV12_BUFFER_CONFIG *sd) {
+  VP8_COMMON *cm = &pbi->common;
+  int ref_fb_idx;
 
-    if (ref_frame_flag == VP8_LAST_FRAME)
-        ref_fb_idx = cm->lst_fb_idx;
-    else if (ref_frame_flag == VP8_GOLD_FRAME)
-        ref_fb_idx = cm->gld_fb_idx;
-    else if (ref_frame_flag == VP8_ALTR_FRAME)
-        ref_fb_idx = cm->alt_fb_idx;
-    else{
-        vpx_internal_error(&pbi->common.error, VPX_CODEC_ERROR,
-            "Invalid reference frame");
-        return pbi->common.error.error_code;
-    }
+  if (ref_frame_flag == VP8_LAST_FRAME)
+    ref_fb_idx = cm->lst_fb_idx;
+  else if (ref_frame_flag == VP8_GOLD_FRAME)
+    ref_fb_idx = cm->gld_fb_idx;
+  else if (ref_frame_flag == VP8_ALTR_FRAME)
+    ref_fb_idx = cm->alt_fb_idx;
+  else {
+    vpx_internal_error(&pbi->common.error, VPX_CODEC_ERROR,
+                       "Invalid reference frame");
+    return pbi->common.error.error_code;
+  }
 
-    if(cm->yv12_fb[ref_fb_idx].y_height != sd->y_height ||
-        cm->yv12_fb[ref_fb_idx].y_width != sd->y_width ||
-        cm->yv12_fb[ref_fb_idx].uv_height != sd->uv_height ||
-        cm->yv12_fb[ref_fb_idx].uv_width != sd->uv_width){
-        vpx_internal_error(&pbi->common.error, VPX_CODEC_ERROR,
-            "Incorrect buffer dimensions");
-    }
-    else
-        vp8_yv12_copy_frame(&cm->yv12_fb[ref_fb_idx], sd);
+  if (cm->yv12_fb[ref_fb_idx].y_height != sd->y_height ||
+      cm->yv12_fb[ref_fb_idx].y_width != sd->y_width ||
+      cm->yv12_fb[ref_fb_idx].uv_height != sd->uv_height ||
+      cm->yv12_fb[ref_fb_idx].uv_width != sd->uv_width) {
+    vpx_internal_error(&pbi->common.error, VPX_CODEC_ERROR,
+                       "Incorrect buffer dimensions");
+  } else
+    vp8_yv12_copy_frame(&cm->yv12_fb[ref_fb_idx], sd);
 
-    return pbi->common.error.error_code;
+  return pbi->common.error.error_code;
 }
 
+vpx_codec_err_t vp8dx_set_reference(VP8D_COMP *pbi,
+                                    enum vpx_ref_frame_type ref_frame_flag,
+                                    YV12_BUFFER_CONFIG *sd) {
+  VP8_COMMON *cm = &pbi->common;
+  int *ref_fb_ptr = NULL;
+  int free_fb;
 
-vpx_codec_err_t vp8dx_set_reference(VP8D_COMP *pbi, enum vpx_ref_frame_type ref_frame_flag, YV12_BUFFER_CONFIG *sd)
-{
-    VP8_COMMON *cm = &pbi->common;
-    int *ref_fb_ptr = NULL;
-    int free_fb;
+  if (ref_frame_flag == VP8_LAST_FRAME)
+    ref_fb_ptr = &cm->lst_fb_idx;
+  else if (ref_frame_flag == VP8_GOLD_FRAME)
+    ref_fb_ptr = &cm->gld_fb_idx;
+  else if (ref_frame_flag == VP8_ALTR_FRAME)
+    ref_fb_ptr = &cm->alt_fb_idx;
+  else {
+    vpx_internal_error(&pbi->common.error, VPX_CODEC_ERROR,
+                       "Invalid reference frame");
+    return pbi->common.error.error_code;
+  }
 
-    if (ref_frame_flag == VP8_LAST_FRAME)
-        ref_fb_ptr = &cm->lst_fb_idx;
-    else if (ref_frame_flag == VP8_GOLD_FRAME)
-        ref_fb_ptr = &cm->gld_fb_idx;
-    else if (ref_frame_flag == VP8_ALTR_FRAME)
-        ref_fb_ptr = &cm->alt_fb_idx;
-    else{
-        vpx_internal_error(&pbi->common.error, VPX_CODEC_ERROR,
-            "Invalid reference frame");
-        return pbi->common.error.error_code;
-    }
+  if (cm->yv12_fb[*ref_fb_ptr].y_height != sd->y_height ||
+      cm->yv12_fb[*ref_fb_ptr].y_width != sd->y_width ||
+      cm->yv12_fb[*ref_fb_ptr].uv_height != sd->uv_height ||
+      cm->yv12_fb[*ref_fb_ptr].uv_width != sd->uv_width) {
+    vpx_internal_error(&pbi->common.error, VPX_CODEC_ERROR,
+                       "Incorrect buffer dimensions");
+  } else {
+    /* Find an empty frame buffer. */
+    free_fb = get_free_fb(cm);
+    /* Decrease fb_idx_ref_cnt since it will be increased again in
+     * ref_cnt_fb() below. */
+    cm->fb_idx_ref_cnt[free_fb]--;
 
-    if(cm->yv12_fb[*ref_fb_ptr].y_height != sd->y_height ||
-        cm->yv12_fb[*ref_fb_ptr].y_width != sd->y_width ||
-        cm->yv12_fb[*ref_fb_ptr].uv_height != sd->uv_height ||
-        cm->yv12_fb[*ref_fb_ptr].uv_width != sd->uv_width){
-        vpx_internal_error(&pbi->common.error, VPX_CODEC_ERROR,
-            "Incorrect buffer dimensions");
-    }
-    else{
-        /* Find an empty frame buffer. */
-        free_fb = get_free_fb(cm);
-        /* Decrease fb_idx_ref_cnt since it will be increased again in
-         * ref_cnt_fb() below. */
-        cm->fb_idx_ref_cnt[free_fb]--;
+    /* Manage the reference counters and copy image. */
+    ref_cnt_fb(cm->fb_idx_ref_cnt, ref_fb_ptr, free_fb);
+    vp8_yv12_copy_frame(sd, &cm->yv12_fb[*ref_fb_ptr]);
+  }
 
-        /* Manage the reference counters and copy image. */
-        ref_cnt_fb (cm->fb_idx_ref_cnt, ref_fb_ptr, free_fb);
-        vp8_yv12_copy_frame(sd, &cm->yv12_fb[*ref_fb_ptr]);
-    }
-
-   return pbi->common.error.error_code;
+  return pbi->common.error.error_code;
 }
 
-static int get_free_fb (VP8_COMMON *cm)
-{
-    int i;
-    for (i = 0; i < NUM_YV12_BUFFERS; i++)
-        if (cm->fb_idx_ref_cnt[i] == 0)
-            break;
+static int get_free_fb(VP8_COMMON *cm) {
+  int i;
+  for (i = 0; i < NUM_YV12_BUFFERS; i++)
+    if (cm->fb_idx_ref_cnt[i] == 0) break;
 
-    assert(i < NUM_YV12_BUFFERS);
-    cm->fb_idx_ref_cnt[i] = 1;
-    return i;
+  assert(i < NUM_YV12_BUFFERS);
+  cm->fb_idx_ref_cnt[i] = 1;
+  return i;
 }
 
-static void ref_cnt_fb (int *buf, int *idx, int new_idx)
-{
-    if (buf[*idx] > 0)
-        buf[*idx]--;
+static void ref_cnt_fb(int *buf, int *idx, int new_idx) {
+  if (buf[*idx] > 0) buf[*idx]--;
 
-    *idx = new_idx;
+  *idx = new_idx;
 
-    buf[new_idx]++;
+  buf[new_idx]++;
 }
 
 /* If any buffer copy / swapping is signalled it should be done here. */
-static int swap_frame_buffers (VP8_COMMON *cm)
-{
-    int err = 0;
+static int swap_frame_buffers(VP8_COMMON *cm) {
+  int err = 0;
 
-    /* The alternate reference frame or golden frame can be updated
-     *  using the new, last, or golden/alt ref frame.  If it
-     *  is updated using the newly decoded frame it is a refresh.
-     *  An update using the last or golden/alt ref frame is a copy.
-     */
-    if (cm->copy_buffer_to_arf)
-    {
-        int new_fb = 0;
+  /* The alternate reference frame or golden frame can be updated
+   *  using the new, last, or golden/alt ref frame.  If it
+   *  is updated using the newly decoded frame it is a refresh.
+   *  An update using the last or golden/alt ref frame is a copy.
+   */
+  if (cm->copy_buffer_to_arf) {
+    int new_fb = 0;
 
-        if (cm->copy_buffer_to_arf == 1)
-            new_fb = cm->lst_fb_idx;
-        else if (cm->copy_buffer_to_arf == 2)
-            new_fb = cm->gld_fb_idx;
-        else
-            err = -1;
+    if (cm->copy_buffer_to_arf == 1)
+      new_fb = cm->lst_fb_idx;
+    else if (cm->copy_buffer_to_arf == 2)
+      new_fb = cm->gld_fb_idx;
+    else
+      err = -1;
 
-        ref_cnt_fb (cm->fb_idx_ref_cnt, &cm->alt_fb_idx, new_fb);
-    }
+    ref_cnt_fb(cm->fb_idx_ref_cnt, &cm->alt_fb_idx, new_fb);
+  }
 
-    if (cm->copy_buffer_to_gf)
-    {
-        int new_fb = 0;
+  if (cm->copy_buffer_to_gf) {
+    int new_fb = 0;
 
-        if (cm->copy_buffer_to_gf == 1)
-            new_fb = cm->lst_fb_idx;
-        else if (cm->copy_buffer_to_gf == 2)
-            new_fb = cm->alt_fb_idx;
-        else
-            err = -1;
+    if (cm->copy_buffer_to_gf == 1)
+      new_fb = cm->lst_fb_idx;
+    else if (cm->copy_buffer_to_gf == 2)
+      new_fb = cm->alt_fb_idx;
+    else
+      err = -1;
 
-        ref_cnt_fb (cm->fb_idx_ref_cnt, &cm->gld_fb_idx, new_fb);
-    }
+    ref_cnt_fb(cm->fb_idx_ref_cnt, &cm->gld_fb_idx, new_fb);
+  }
 
-    if (cm->refresh_golden_frame)
-        ref_cnt_fb (cm->fb_idx_ref_cnt, &cm->gld_fb_idx, cm->new_fb_idx);
+  if (cm->refresh_golden_frame)
+    ref_cnt_fb(cm->fb_idx_ref_cnt, &cm->gld_fb_idx, cm->new_fb_idx);
 
-    if (cm->refresh_alt_ref_frame)
-        ref_cnt_fb (cm->fb_idx_ref_cnt, &cm->alt_fb_idx, cm->new_fb_idx);
+  if (cm->refresh_alt_ref_frame)
+    ref_cnt_fb(cm->fb_idx_ref_cnt, &cm->alt_fb_idx, cm->new_fb_idx);
 
-    if (cm->refresh_last_frame)
-    {
-        ref_cnt_fb (cm->fb_idx_ref_cnt, &cm->lst_fb_idx, cm->new_fb_idx);
+  if (cm->refresh_last_frame) {
+    ref_cnt_fb(cm->fb_idx_ref_cnt, &cm->lst_fb_idx, cm->new_fb_idx);
 
-        cm->frame_to_show = &cm->yv12_fb[cm->lst_fb_idx];
-    }
-    else
-        cm->frame_to_show = &cm->yv12_fb[cm->new_fb_idx];
+    cm->frame_to_show = &cm->yv12_fb[cm->lst_fb_idx];
+  } else
+    cm->frame_to_show = &cm->yv12_fb[cm->new_fb_idx];
 
-    cm->fb_idx_ref_cnt[cm->new_fb_idx]--;
+  cm->fb_idx_ref_cnt[cm->new_fb_idx]--;
 
-    return err;
+  return err;
 }
 
-static int check_fragments_for_errors(VP8D_COMP *pbi)
-{
-    if (!pbi->ec_active &&
-        pbi->fragments.count <= 1 && pbi->fragments.sizes[0] == 0)
-    {
-        VP8_COMMON *cm = &pbi->common;
+static int check_fragments_for_errors(VP8D_COMP *pbi) {
+  if (!pbi->ec_active && pbi->fragments.count <= 1 &&
+      pbi->fragments.sizes[0] == 0) {
+    VP8_COMMON *cm = &pbi->common;
 
-        /* If error concealment is disabled we won't signal missing frames
-         * to the decoder.
-         */
-        if (cm->fb_idx_ref_cnt[cm->lst_fb_idx] > 1)
-        {
-            /* The last reference shares buffer with another reference
-             * buffer. Move it to its own buffer before setting it as
-             * corrupt, otherwise we will make multiple buffers corrupt.
-             */
-            const int prev_idx = cm->lst_fb_idx;
-            cm->fb_idx_ref_cnt[prev_idx]--;
-            cm->lst_fb_idx = get_free_fb(cm);
-            vp8_yv12_copy_frame(&cm->yv12_fb[prev_idx],
-                                    &cm->yv12_fb[cm->lst_fb_idx]);
-        }
-        /* This is used to signal that we are missing frames.
-         * We do not know if the missing frame(s) was supposed to update
-         * any of the reference buffers, but we act conservative and
-         * mark only the last buffer as corrupted.
-         */
-        cm->yv12_fb[cm->lst_fb_idx].corrupted = 1;
+    /* If error concealment is disabled we won't signal missing frames
+     * to the decoder.
+     */
+    if (cm->fb_idx_ref_cnt[cm->lst_fb_idx] > 1) {
+      /* The last reference shares buffer with another reference
+       * buffer. Move it to its own buffer before setting it as
+       * corrupt, otherwise we will make multiple buffers corrupt.
+       */
+      const int prev_idx = cm->lst_fb_idx;
+      cm->fb_idx_ref_cnt[prev_idx]--;
+      cm->lst_fb_idx = get_free_fb(cm);
+      vp8_yv12_copy_frame(&cm->yv12_fb[prev_idx], &cm->yv12_fb[cm->lst_fb_idx]);
+    }
+    /* This is used to signal that we are missing frames.
+     * We do not know if the missing frame(s) was supposed to update
+     * any of the reference buffers, but we act conservative and
+     * mark only the last buffer as corrupted.
+     */
+    cm->yv12_fb[cm->lst_fb_idx].corrupted = 1;
 
-        /* Signal that we have no frame to show. */
-        cm->show_frame = 0;
+    /* Signal that we have no frame to show. */
+    cm->show_frame = 0;
 
-        /* Nothing more to do. */
-        return 0;
-    }
+    /* Nothing more to do. */
+    return 0;
+  }
 
-    return 1;
+  return 1;
 }
 
 int vp8dx_receive_compressed_data(VP8D_COMP *pbi, size_t size,
-                                  const uint8_t *source,
-                                  int64_t time_stamp)
-{
-    VP8_COMMON *cm = &pbi->common;
-    int retcode = -1;
-    (void)size;
-    (void)source;
+                                  const uint8_t *source, int64_t time_stamp) {
+  VP8_COMMON *cm = &pbi->common;
+  int retcode = -1;
+  (void)size;
+  (void)source;
 
-    pbi->common.error.error_code = VPX_CODEC_OK;
+  pbi->common.error.error_code = VPX_CODEC_OK;
 
-    retcode = check_fragments_for_errors(pbi);
-    if(retcode <= 0)
-        return retcode;
+  retcode = check_fragments_for_errors(pbi);
+  if (retcode <= 0) return retcode;
 
-    cm->new_fb_idx = get_free_fb (cm);
+  cm->new_fb_idx = get_free_fb(cm);
 
-    /* setup reference frames for vp8_decode_frame */
-    pbi->dec_fb_ref[INTRA_FRAME]  = &cm->yv12_fb[cm->new_fb_idx];
-    pbi->dec_fb_ref[LAST_FRAME]   = &cm->yv12_fb[cm->lst_fb_idx];
-    pbi->dec_fb_ref[GOLDEN_FRAME] = &cm->yv12_fb[cm->gld_fb_idx];
-    pbi->dec_fb_ref[ALTREF_FRAME] = &cm->yv12_fb[cm->alt_fb_idx];
+  /* setup reference frames for vp8_decode_frame */
+  pbi->dec_fb_ref[INTRA_FRAME] = &cm->yv12_fb[cm->new_fb_idx];
+  pbi->dec_fb_ref[LAST_FRAME] = &cm->yv12_fb[cm->lst_fb_idx];
+  pbi->dec_fb_ref[GOLDEN_FRAME] = &cm->yv12_fb[cm->gld_fb_idx];
+  pbi->dec_fb_ref[ALTREF_FRAME] = &cm->yv12_fb[cm->alt_fb_idx];
 
-    if (setjmp(pbi->common.error.jmp))
-    {
-       /* We do not know if the missing frame(s) was supposed to update
-        * any of the reference buffers, but we act conservative and
-        * mark only the last buffer as corrupted.
-        */
-        cm->yv12_fb[cm->lst_fb_idx].corrupted = 1;
+  if (setjmp(pbi->common.error.jmp)) {
+    /* We do not know if the missing frame(s) was supposed to update
+     * any of the reference buffers, but we act conservative and
+     * mark only the last buffer as corrupted.
+     */
+    cm->yv12_fb[cm->lst_fb_idx].corrupted = 1;
 
-        if (cm->fb_idx_ref_cnt[cm->new_fb_idx] > 0)
-          cm->fb_idx_ref_cnt[cm->new_fb_idx]--;
+    if (cm->fb_idx_ref_cnt[cm->new_fb_idx] > 0)
+      cm->fb_idx_ref_cnt[cm->new_fb_idx]--;
 
-        goto decode_exit;
-    }
+    goto decode_exit;
+  }
 
-    pbi->common.error.setjmp = 1;
+  pbi->common.error.setjmp = 1;
 
-    retcode = vp8_decode_frame(pbi);
+  retcode = vp8_decode_frame(pbi);
 
-    if (retcode < 0)
-    {
-        if (cm->fb_idx_ref_cnt[cm->new_fb_idx] > 0)
-          cm->fb_idx_ref_cnt[cm->new_fb_idx]--;
+  if (retcode < 0) {
+    if (cm->fb_idx_ref_cnt[cm->new_fb_idx] > 0)
+      cm->fb_idx_ref_cnt[cm->new_fb_idx]--;
 
-        pbi->common.error.error_code = VPX_CODEC_ERROR;
-        goto decode_exit;
-    }
+    pbi->common.error.error_code = VPX_CODEC_ERROR;
+    goto decode_exit;
+  }
 
-    if (swap_frame_buffers (cm))
-    {
-        pbi->common.error.error_code = VPX_CODEC_ERROR;
-        goto decode_exit;
-    }
+  if (swap_frame_buffers(cm)) {
+    pbi->common.error.error_code = VPX_CODEC_ERROR;
+    goto decode_exit;
+  }
 
-    vp8_clear_system_state();
+  vp8_clear_system_state();
 
-    if (cm->show_frame)
-    {
-        cm->current_video_frame++;
-        cm->show_frame_mi = cm->mi;
-    }
+  if (cm->show_frame) {
+    cm->current_video_frame++;
+    cm->show_frame_mi = cm->mi;
+  }
 
-    #if CONFIG_ERROR_CONCEALMENT
-    /* swap the mode infos to storage for future error concealment */
-    if (pbi->ec_enabled && pbi->common.prev_mi)
-    {
-        MODE_INFO* tmp = pbi->common.prev_mi;
-        int row, col;
-        pbi->common.prev_mi = pbi->common.mi;
-        pbi->common.mi = tmp;
+#if CONFIG_ERROR_CONCEALMENT
+  /* swap the mode infos to storage for future error concealment */
+  if (pbi->ec_enabled && pbi->common.prev_mi) {
+    MODE_INFO *tmp = pbi->common.prev_mi;
+    int row, col;
+    pbi->common.prev_mi = pbi->common.mi;
+    pbi->common.mi = tmp;
 
-        /* Propagate the segment_ids to the next frame */
-        for (row = 0; row < pbi->common.mb_rows; ++row)
-        {
-            for (col = 0; col < pbi->common.mb_cols; ++col)
-            {
-                const int i = row*pbi->common.mode_info_stride + col;
-                pbi->common.mi[i].mbmi.segment_id =
-                        pbi->common.prev_mi[i].mbmi.segment_id;
-            }
-        }
+    /* Propagate the segment_ids to the next frame */
+    for (row = 0; row < pbi->common.mb_rows; ++row) {
+      for (col = 0; col < pbi->common.mb_cols; ++col) {
+        const int i = row * pbi->common.mode_info_stride + col;
+        pbi->common.mi[i].mbmi.segment_id =
+            pbi->common.prev_mi[i].mbmi.segment_id;
+      }
     }
+  }
 #endif
 
-    pbi->ready_for_new_data = 0;
-    pbi->last_time_stamp = time_stamp;
+  pbi->ready_for_new_data = 0;
+  pbi->last_time_stamp = time_stamp;
 
 decode_exit:
-    pbi->common.error.setjmp = 0;
-    vp8_clear_system_state();
-    return retcode;
+  pbi->common.error.setjmp = 0;
+  vp8_clear_system_state();
+  return retcode;
 }
-int vp8dx_get_raw_frame(VP8D_COMP *pbi, YV12_BUFFER_CONFIG *sd, int64_t *time_stamp, int64_t *time_end_stamp, vp8_ppflags_t *flags)
-{
-    int ret = -1;
+int vp8dx_get_raw_frame(VP8D_COMP *pbi, YV12_BUFFER_CONFIG *sd,
+                        int64_t *time_stamp, int64_t *time_end_stamp,
+                        vp8_ppflags_t *flags) {
+  int ret = -1;
 
-    if (pbi->ready_for_new_data == 1)
-        return ret;
+  if (pbi->ready_for_new_data == 1) return ret;
 
-    /* ie no raw frame to show!!! */
-    if (pbi->common.show_frame == 0)
-        return ret;
+  /* ie no raw frame to show!!! */
+  if (pbi->common.show_frame == 0) return ret;
 
-    pbi->ready_for_new_data = 1;
-    *time_stamp = pbi->last_time_stamp;
-    *time_end_stamp = 0;
+  pbi->ready_for_new_data = 1;
+  *time_stamp = pbi->last_time_stamp;
+  *time_end_stamp = 0;
 
 #if CONFIG_POSTPROC
-    ret = vp8_post_proc_frame(&pbi->common, sd, flags);
+  ret = vp8_post_proc_frame(&pbi->common, sd, flags);
 #else
-    (void)flags;
+  (void)flags;
 
-    if (pbi->common.frame_to_show)
-    {
-        *sd = *pbi->common.frame_to_show;
-        sd->y_width = pbi->common.Width;
-        sd->y_height = pbi->common.Height;
-        sd->uv_height = pbi->common.Height / 2;
-        ret = 0;
-    }
-    else
-    {
-        ret = -1;
-    }
+  if (pbi->common.frame_to_show) {
+    *sd = *pbi->common.frame_to_show;
+    sd->y_width = pbi->common.Width;
+    sd->y_height = pbi->common.Height;
+    sd->uv_height = pbi->common.Height / 2;
+    ret = 0;
+  } else {
+    ret = -1;
+  }
 
 #endif /*!CONFIG_POSTPROC*/
-    vp8_clear_system_state();
-    return ret;
+  vp8_clear_system_state();
+  return ret;
 }
 
-
 /* This function as written isn't decoder specific, but the encoder has
  * much faster ways of computing this, so it's ok for it to live in a
  * decode specific file.
  */
-int vp8dx_references_buffer( VP8_COMMON *oci, int ref_frame )
-{
-    const MODE_INFO *mi = oci->mi;
-    int mb_row, mb_col;
+int vp8dx_references_buffer(VP8_COMMON *oci, int ref_frame) {
+  const MODE_INFO *mi = oci->mi;
+  int mb_row, mb_col;
 
-    for (mb_row = 0; mb_row < oci->mb_rows; mb_row++)
-    {
-        for (mb_col = 0; mb_col < oci->mb_cols; mb_col++,mi++)
-        {
-            if( mi->mbmi.ref_frame == ref_frame)
-              return 1;
-        }
-        mi++;
+  for (mb_row = 0; mb_row < oci->mb_rows; mb_row++) {
+    for (mb_col = 0; mb_col < oci->mb_cols; mb_col++, mi++) {
+      if (mi->mbmi.ref_frame == ref_frame) return 1;
     }
-    return 0;
-
+    mi++;
+  }
+  return 0;
 }
 
-int vp8_create_decoder_instances(struct frame_buffers *fb, VP8D_CONFIG *oxcf)
-{
-    if(!fb->use_frame_threads)
-    {
-        /* decoder instance for single thread mode */
-        fb->pbi[0] = create_decompressor(oxcf);
-        if(!fb->pbi[0])
-            return VPX_CODEC_ERROR;
+int vp8_create_decoder_instances(struct frame_buffers *fb, VP8D_CONFIG *oxcf) {
+  if (!fb->use_frame_threads) {
+    /* decoder instance for single thread mode */
+    fb->pbi[0] = create_decompressor(oxcf);
+    if (!fb->pbi[0]) return VPX_CODEC_ERROR;
 
 #if CONFIG_MULTITHREAD
-        /* enable row-based threading only when use_frame_threads
-         * is disabled */
-        fb->pbi[0]->max_threads = oxcf->max_threads;
-        vp8_decoder_create_threads(fb->pbi[0]);
+    /* enable row-based threading only when use_frame_threads
+     * is disabled */
+    fb->pbi[0]->max_threads = oxcf->max_threads;
+    vp8_decoder_create_threads(fb->pbi[0]);
 #endif
-    }
-    else
-    {
-        /* TODO : create frame threads and decoder instances for each
-         * thread here */
-    }
+  } else {
+    /* TODO : create frame threads and decoder instances for each
+     * thread here */
+  }
 
-    return VPX_CODEC_OK;
+  return VPX_CODEC_OK;
 }
 
-int vp8_remove_decoder_instances(struct frame_buffers *fb)
-{
-    if(!fb->use_frame_threads)
-    {
-        VP8D_COMP *pbi = fb->pbi[0];
+int vp8_remove_decoder_instances(struct frame_buffers *fb) {
+  if (!fb->use_frame_threads) {
+    VP8D_COMP *pbi = fb->pbi[0];
 
-        if (!pbi)
-            return VPX_CODEC_ERROR;
+    if (!pbi) return VPX_CODEC_ERROR;
 #if CONFIG_MULTITHREAD
-        if (pbi->b_multithreaded_rd)
-            vp8mt_de_alloc_temp_buffers(pbi, pbi->common.mb_rows);
-        vp8_decoder_remove_threads(pbi);
+    if (pbi->b_multithreaded_rd)
+      vp8mt_de_alloc_temp_buffers(pbi, pbi->common.mb_rows);
+    vp8_decoder_remove_threads(pbi);
 #endif
 
-        /* decoder instance for single thread mode */
-        remove_decompressor(pbi);
-    }
-    else
-    {
-        /* TODO : remove frame threads and decoder instances for each
-         * thread here */
-    }
+    /* decoder instance for single thread mode */
+    remove_decompressor(pbi);
+  } else {
+    /* TODO : remove frame threads and decoder instances for each
+     * thread here */
+  }
 
-    return VPX_CODEC_OK;
+  return VPX_CODEC_OK;
 }
--- a/vp8/decoder/onyxd_int.h
+++ b/vp8/decoder/onyxd_int.h
@@ -8,7 +8,6 @@
  *  be found in the AUTHORS file in the root of the source tree.
  */
 
-
 #ifndef VP8_DECODER_ONYXD_INT_H_
 #define VP8_DECODER_ONYXD_INT_H_
 
@@ -26,110 +25,101 @@
 extern "C" {
 #endif
 
-typedef struct
-{
-    int ithread;
-    void *ptr1;
-    void *ptr2;
+typedef struct {
+  int ithread;
+  void *ptr1;
+  void *ptr2;
 } DECODETHREAD_DATA;
 
-typedef struct
-{
-    MACROBLOCKD  mbd;
-} MB_ROW_DEC;
+typedef struct { MACROBLOCKD mbd; } MB_ROW_DEC;
 
-
-typedef struct
-{
-    int enabled;
-    unsigned int count;
-    const unsigned char *ptrs[MAX_PARTITIONS];
-    unsigned int sizes[MAX_PARTITIONS];
+typedef struct {
+  int enabled;
+  unsigned int count;
+  const unsigned char *ptrs[MAX_PARTITIONS];
+  unsigned int sizes[MAX_PARTITIONS];
 } FRAGMENT_DATA;
 
 #define MAX_FB_MT_DEC 32
 
-struct frame_buffers
-{
-    /*
-     * this struct will be populated with frame buffer management
-     * info in future commits. */
+struct frame_buffers {
+  /*
+   * this struct will be populated with frame buffer management
+   * info in future commits. */
 
-    /* enable/disable frame-based threading */
-    int     use_frame_threads;
+  /* enable/disable frame-based threading */
+  int use_frame_threads;
 
-    /* decoder instances */
-    struct VP8D_COMP *pbi[MAX_FB_MT_DEC];
-
+  /* decoder instances */
+  struct VP8D_COMP *pbi[MAX_FB_MT_DEC];
 };
 
-typedef struct VP8D_COMP
-{
-    DECLARE_ALIGNED(16, MACROBLOCKD, mb);
+typedef struct VP8D_COMP {
+  DECLARE_ALIGNED(16, MACROBLOCKD, mb);
 
-    YV12_BUFFER_CONFIG *dec_fb_ref[NUM_YV12_BUFFERS];
+  YV12_BUFFER_CONFIG *dec_fb_ref[NUM_YV12_BUFFERS];
 
-    DECLARE_ALIGNED(16, VP8_COMMON, common);
+  DECLARE_ALIGNED(16, VP8_COMMON, common);
 
-    /* the last partition will be used for the modes/mvs */
-    vp8_reader mbc[MAX_PARTITIONS];
+  /* the last partition will be used for the modes/mvs */
+  vp8_reader mbc[MAX_PARTITIONS];
 
-    VP8D_CONFIG oxcf;
+  VP8D_CONFIG oxcf;
 
-    FRAGMENT_DATA fragments;
+  FRAGMENT_DATA fragments;
 
 #if CONFIG_MULTITHREAD
-    /* variable for threading */
+  /* variable for threading */
 
-    int b_multithreaded_rd;
-    int max_threads;
-    int current_mb_col_main;
-    unsigned int decoding_thread_count;
-    int allocated_decoding_thread_count;
+  int b_multithreaded_rd;
+  int max_threads;
+  int current_mb_col_main;
+  unsigned int decoding_thread_count;
+  int allocated_decoding_thread_count;
 
-    int mt_baseline_filter_level[MAX_MB_SEGMENTS];
-    int sync_range;
-    int *mt_current_mb_col;                  /* Each row remembers its already decoded column. */
-    pthread_mutex_t *pmutex;
-    pthread_mutex_t mt_mutex;                /* mutex for b_multithreaded_rd */
+  int mt_baseline_filter_level[MAX_MB_SEGMENTS];
+  int sync_range;
+  int *mt_current_mb_col; /* Each row remembers its already decoded column. */
+  pthread_mutex_t *pmutex;
+  pthread_mutex_t mt_mutex; /* mutex for b_multithreaded_rd */
 
-    unsigned char **mt_yabove_row;           /* mb_rows x width */
-    unsigned char **mt_uabove_row;
-    unsigned char **mt_vabove_row;
-    unsigned char **mt_yleft_col;            /* mb_rows x 16 */
-    unsigned char **mt_uleft_col;            /* mb_rows x 8 */
-    unsigned char **mt_vleft_col;            /* mb_rows x 8 */
+  unsigned char **mt_yabove_row; /* mb_rows x width */
+  unsigned char **mt_uabove_row;
+  unsigned char **mt_vabove_row;
+  unsigned char **mt_yleft_col; /* mb_rows x 16 */
+  unsigned char **mt_uleft_col; /* mb_rows x 8 */
+  unsigned char **mt_vleft_col; /* mb_rows x 8 */
 
-    MB_ROW_DEC           *mb_row_di;
-    DECODETHREAD_DATA    *de_thread_data;
+  MB_ROW_DEC *mb_row_di;
+  DECODETHREAD_DATA *de_thread_data;
 
-    pthread_t           *h_decoding_thread;
-    sem_t               *h_event_start_decoding;
-    sem_t                h_event_end_decoding;
-    /* end of threading data */
+  pthread_t *h_decoding_thread;
+  sem_t *h_event_start_decoding;
+  sem_t h_event_end_decoding;
+/* end of threading data */
 #endif
 
-    int64_t last_time_stamp;
-    int   ready_for_new_data;
+  int64_t last_time_stamp;
+  int ready_for_new_data;
 
-    vp8_prob prob_intra;
-    vp8_prob prob_last;
-    vp8_prob prob_gf;
-    vp8_prob prob_skip_false;
+  vp8_prob prob_intra;
+  vp8_prob prob_last;
+  vp8_prob prob_gf;
+  vp8_prob prob_skip_false;
 
 #if CONFIG_ERROR_CONCEALMENT
-    MB_OVERLAP *overlaps;
-    /* the mb num from which modes and mvs (first partition) are corrupt */
-    unsigned int mvs_corrupt_from_mb;
+  MB_OVERLAP *overlaps;
+  /* the mb num from which modes and mvs (first partition) are corrupt */
+  unsigned int mvs_corrupt_from_mb;
 #endif
-    int ec_enabled;
-    int ec_active;
-    int decoded_key_frame;
-    int independent_partitions;
-    int frame_corrupt_residual;
+  int ec_enabled;
+  int ec_active;
+  int decoded_key_frame;
+  int independent_partitions;
+  int frame_corrupt_residual;
 
-    vpx_decrypt_cb decrypt_cb;
-    void *decrypt_state;
+  vpx_decrypt_cb decrypt_cb;
+  void *decrypt_state;
 } VP8D_COMP;
 
 int vp8_decode_frame(VP8D_COMP *cpi);
@@ -138,20 +128,22 @@
 int vp8_remove_decoder_instances(struct frame_buffers *fb);
 
 #if CONFIG_DEBUG
-#define CHECK_MEM_ERROR(lval,expr) do {\
-        lval = (expr); \
-        if(!lval) \
-            vpx_internal_error(&pbi->common.error, VPX_CODEC_MEM_ERROR,\
-                               "Failed to allocate "#lval" at %s:%d", \
-                               __FILE__,__LINE__);\
-    } while(0)
+#define CHECK_MEM_ERROR(lval, expr)                                         \
+  do {                                                                      \
+    lval = (expr);                                                          \
+    if (!lval)                                                              \
+      vpx_internal_error(&pbi->common.error, VPX_CODEC_MEM_ERROR,           \
+                         "Failed to allocate " #lval " at %s:%d", __FILE__, \
+                         __LINE__);                                         \
+  } while (0)
 #else
-#define CHECK_MEM_ERROR(lval,expr) do {\
-        lval = (expr); \
-        if(!lval) \
-            vpx_internal_error(&pbi->common.error, VPX_CODEC_MEM_ERROR,\
-                               "Failed to allocate "#lval);\
-    } while(0)
+#define CHECK_MEM_ERROR(lval, expr)                               \
+  do {                                                            \
+    lval = (expr);                                                \
+    if (!lval)                                                    \
+      vpx_internal_error(&pbi->common.error, VPX_CODEC_MEM_ERROR, \
+                         "Failed to allocate " #lval);            \
+  } while (0)
 #endif
 
 #ifdef __cplusplus
--- a/vp8/decoder/threading.c
+++ b/vp8/decoder/threading.c
@@ -8,11 +8,10 @@
  *  be found in the AUTHORS file in the root of the source tree.
  */
 
-
 #include "vpx_config.h"
 #include "vp8_rtcd.h"
 #if !defined(WIN32) && CONFIG_OS_SUPPORT == 1
-# include <unistd.h>
+#include <unistd.h>
 #endif
 #include "onyxd_int.h"
 #include "vpx_mem/vpx_mem.h"
@@ -31,898 +30,824 @@
 #endif
 
 #define CALLOC_ARRAY(p, n) CHECK_MEM_ERROR((p), vpx_calloc(sizeof(*(p)), (n)))
-#define CALLOC_ARRAY_ALIGNED(p, n, algn) do {                      \
-  CHECK_MEM_ERROR((p), vpx_memalign((algn), sizeof(*(p)) * (n)));  \
-  memset((p), 0, (n) * sizeof(*(p)));                              \
-} while (0)
+#define CALLOC_ARRAY_ALIGNED(p, n, algn)                            \
+  do {                                                              \
+    CHECK_MEM_ERROR((p), vpx_memalign((algn), sizeof(*(p)) * (n))); \
+    memset((p), 0, (n) * sizeof(*(p)));                             \
+  } while (0)
 
-
 void vp8_mb_init_dequantizer(VP8D_COMP *pbi, MACROBLOCKD *xd);
 
-static void setup_decoding_thread_data(VP8D_COMP *pbi, MACROBLOCKD *xd, MB_ROW_DEC *mbrd, int count)
-{
-    VP8_COMMON *const pc = & pbi->common;
-    int i;
+static void setup_decoding_thread_data(VP8D_COMP *pbi, MACROBLOCKD *xd,
+                                       MB_ROW_DEC *mbrd, int count) {
+  VP8_COMMON *const pc = &pbi->common;
+  int i;
 
-    for (i = 0; i < count; i++)
-    {
-        MACROBLOCKD *mbd = &mbrd[i].mbd;
-        mbd->subpixel_predict        = xd->subpixel_predict;
-        mbd->subpixel_predict8x4     = xd->subpixel_predict8x4;
-        mbd->subpixel_predict8x8     = xd->subpixel_predict8x8;
-        mbd->subpixel_predict16x16   = xd->subpixel_predict16x16;
+  for (i = 0; i < count; i++) {
+    MACROBLOCKD *mbd = &mbrd[i].mbd;
+    mbd->subpixel_predict = xd->subpixel_predict;
+    mbd->subpixel_predict8x4 = xd->subpixel_predict8x4;
+    mbd->subpixel_predict8x8 = xd->subpixel_predict8x8;
+    mbd->subpixel_predict16x16 = xd->subpixel_predict16x16;
 
-        mbd->frame_type = pc->frame_type;
-        mbd->pre = xd->pre;
-        mbd->dst = xd->dst;
+    mbd->frame_type = pc->frame_type;
+    mbd->pre = xd->pre;
+    mbd->dst = xd->dst;
 
-        mbd->segmentation_enabled    = xd->segmentation_enabled;
-        mbd->mb_segement_abs_delta     = xd->mb_segement_abs_delta;
-        memcpy(mbd->segment_feature_data, xd->segment_feature_data, sizeof(xd->segment_feature_data));
+    mbd->segmentation_enabled = xd->segmentation_enabled;
+    mbd->mb_segement_abs_delta = xd->mb_segement_abs_delta;
+    memcpy(mbd->segment_feature_data, xd->segment_feature_data,
+           sizeof(xd->segment_feature_data));
 
-        /*signed char ref_lf_deltas[MAX_REF_LF_DELTAS];*/
-        memcpy(mbd->ref_lf_deltas, xd->ref_lf_deltas, sizeof(xd->ref_lf_deltas));
-        /*signed char mode_lf_deltas[MAX_MODE_LF_DELTAS];*/
-        memcpy(mbd->mode_lf_deltas, xd->mode_lf_deltas, sizeof(xd->mode_lf_deltas));
-        /*unsigned char mode_ref_lf_delta_enabled;
-        unsigned char mode_ref_lf_delta_update;*/
-        mbd->mode_ref_lf_delta_enabled    = xd->mode_ref_lf_delta_enabled;
-        mbd->mode_ref_lf_delta_update    = xd->mode_ref_lf_delta_update;
+    /*signed char ref_lf_deltas[MAX_REF_LF_DELTAS];*/
+    memcpy(mbd->ref_lf_deltas, xd->ref_lf_deltas, sizeof(xd->ref_lf_deltas));
+    /*signed char mode_lf_deltas[MAX_MODE_LF_DELTAS];*/
+    memcpy(mbd->mode_lf_deltas, xd->mode_lf_deltas, sizeof(xd->mode_lf_deltas));
+    /*unsigned char mode_ref_lf_delta_enabled;
+    unsigned char mode_ref_lf_delta_update;*/
+    mbd->mode_ref_lf_delta_enabled = xd->mode_ref_lf_delta_enabled;
+    mbd->mode_ref_lf_delta_update = xd->mode_ref_lf_delta_update;
 
-        mbd->current_bc = &pbi->mbc[0];
+    mbd->current_bc = &pbi->mbc[0];
 
-        memcpy(mbd->dequant_y1_dc, xd->dequant_y1_dc, sizeof(xd->dequant_y1_dc));
-        memcpy(mbd->dequant_y1, xd->dequant_y1, sizeof(xd->dequant_y1));
-        memcpy(mbd->dequant_y2, xd->dequant_y2, sizeof(xd->dequant_y2));
-        memcpy(mbd->dequant_uv, xd->dequant_uv, sizeof(xd->dequant_uv));
+    memcpy(mbd->dequant_y1_dc, xd->dequant_y1_dc, sizeof(xd->dequant_y1_dc));
+    memcpy(mbd->dequant_y1, xd->dequant_y1, sizeof(xd->dequant_y1));
+    memcpy(mbd->dequant_y2, xd->dequant_y2, sizeof(xd->dequant_y2));
+    memcpy(mbd->dequant_uv, xd->dequant_uv, sizeof(xd->dequant_uv));
 
-        mbd->fullpixel_mask = 0xffffffff;
+    mbd->fullpixel_mask = 0xffffffff;
 
-        if (pc->full_pixel)
-            mbd->fullpixel_mask = 0xfffffff8;
+    if (pc->full_pixel) mbd->fullpixel_mask = 0xfffffff8;
+  }
 
-    }
-
-    for (i = 0; i < pc->mb_rows; i++)
-        pbi->mt_current_mb_col[i] = -1;
+  for (i = 0; i < pc->mb_rows; i++) pbi->mt_current_mb_col[i] = -1;
 }
 
 static void mt_decode_macroblock(VP8D_COMP *pbi, MACROBLOCKD *xd,
-                                 unsigned int mb_idx)
-{
-    MB_PREDICTION_MODE mode;
-    int i;
+                                 unsigned int mb_idx) {
+  MB_PREDICTION_MODE mode;
+  int i;
 #if CONFIG_ERROR_CONCEALMENT
-    int corruption_detected = 0;
+  int corruption_detected = 0;
 #else
-    (void)mb_idx;
+  (void)mb_idx;
 #endif
 
-    if (xd->mode_info_context->mbmi.mb_skip_coeff)
-    {
-        vp8_reset_mb_tokens_context(xd);
-    }
-    else if (!vp8dx_bool_error(xd->current_bc))
-    {
-        int eobtotal;
-        eobtotal = vp8_decode_mb_tokens(pbi, xd);
+  if (xd->mode_info_context->mbmi.mb_skip_coeff) {
+    vp8_reset_mb_tokens_context(xd);
+  } else if (!vp8dx_bool_error(xd->current_bc)) {
+    int eobtotal;
+    eobtotal = vp8_decode_mb_tokens(pbi, xd);
 
-        /* Special case:  Force the loopfilter to skip when eobtotal is zero */
-        xd->mode_info_context->mbmi.mb_skip_coeff = (eobtotal==0);
-    }
+    /* Special case:  Force the loopfilter to skip when eobtotal is zero */
+    xd->mode_info_context->mbmi.mb_skip_coeff = (eobtotal == 0);
+  }
 
-    mode = xd->mode_info_context->mbmi.mode;
+  mode = xd->mode_info_context->mbmi.mode;
 
-    if (xd->segmentation_enabled)
-        vp8_mb_init_dequantizer(pbi, xd);
+  if (xd->segmentation_enabled) vp8_mb_init_dequantizer(pbi, xd);
 
-
 #if CONFIG_ERROR_CONCEALMENT
 
-    if(pbi->ec_active)
-    {
-        int throw_residual;
-        /* When we have independent partitions we can apply residual even
-         * though other partitions within the frame are corrupt.
-         */
-        throw_residual = (!pbi->independent_partitions &&
-                          pbi->frame_corrupt_residual);
-        throw_residual = (throw_residual || vp8dx_bool_error(xd->current_bc));
+  if (pbi->ec_active) {
+    int throw_residual;
+    /* When we have independent partitions we can apply residual even
+     * though other partitions within the frame are corrupt.
+     */
+    throw_residual =
+        (!pbi->independent_partitions && pbi->frame_corrupt_residual);
+    throw_residual = (throw_residual || vp8dx_bool_error(xd->current_bc));
 
-        if ((mb_idx >= pbi->mvs_corrupt_from_mb || throw_residual))
-        {
-            /* MB with corrupt residuals or corrupt mode/motion vectors.
-             * Better to use the predictor as reconstruction.
-             */
-            pbi->frame_corrupt_residual = 1;
-            memset(xd->qcoeff, 0, sizeof(xd->qcoeff));
+    if ((mb_idx >= pbi->mvs_corrupt_from_mb || throw_residual)) {
+      /* MB with corrupt residuals or corrupt mode/motion vectors.
+       * Better to use the predictor as reconstruction.
+       */
+      pbi->frame_corrupt_residual = 1;
+      memset(xd->qcoeff, 0, sizeof(xd->qcoeff));
 
-            corruption_detected = 1;
+      corruption_detected = 1;
 
-            /* force idct to be skipped for B_PRED and use the
-             * prediction only for reconstruction
-             * */
-            memset(xd->eobs, 0, 25);
-        }
+      /* force idct to be skipped for B_PRED and use the
+       * prediction only for reconstruction
+       * */
+      memset(xd->eobs, 0, 25);
     }
+  }
 #endif
 
-    /* do prediction */
-    if (xd->mode_info_context->mbmi.ref_frame == INTRA_FRAME)
-    {
-        vp8_build_intra_predictors_mbuv_s(xd,
-                                          xd->recon_above[1],
-                                          xd->recon_above[2],
-                                          xd->recon_left[1],
-                                          xd->recon_left[2],
-                                          xd->recon_left_stride[1],
-                                          xd->dst.u_buffer, xd->dst.v_buffer,
-                                          xd->dst.uv_stride);
+  /* do prediction */
+  if (xd->mode_info_context->mbmi.ref_frame == INTRA_FRAME) {
+    vp8_build_intra_predictors_mbuv_s(
+        xd, xd->recon_above[1], xd->recon_above[2], xd->recon_left[1],
+        xd->recon_left[2], xd->recon_left_stride[1], xd->dst.u_buffer,
+        xd->dst.v_buffer, xd->dst.uv_stride);
 
-        if (mode != B_PRED)
-        {
-            vp8_build_intra_predictors_mby_s(xd,
-                                                 xd->recon_above[0],
-                                                 xd->recon_left[0],
-                                                 xd->recon_left_stride[0],
-                                                 xd->dst.y_buffer,
-                                                 xd->dst.y_stride);
-        }
-        else
-        {
-            short *DQC = xd->dequant_y1;
-            int dst_stride = xd->dst.y_stride;
+    if (mode != B_PRED) {
+      vp8_build_intra_predictors_mby_s(
+          xd, xd->recon_above[0], xd->recon_left[0], xd->recon_left_stride[0],
+          xd->dst.y_buffer, xd->dst.y_stride);
+    } else {
+      short *DQC = xd->dequant_y1;
+      int dst_stride = xd->dst.y_stride;
 
-            /* clear out residual eob info */
-            if(xd->mode_info_context->mbmi.mb_skip_coeff)
-                memset(xd->eobs, 0, 25);
+      /* clear out residual eob info */
+      if (xd->mode_info_context->mbmi.mb_skip_coeff) memset(xd->eobs, 0, 25);
 
-            intra_prediction_down_copy(xd, xd->recon_above[0] + 16);
+      intra_prediction_down_copy(xd, xd->recon_above[0] + 16);
 
-            for (i = 0; i < 16; i++)
-            {
-                BLOCKD *b = &xd->block[i];
-                unsigned char *dst = xd->dst.y_buffer + b->offset;
-                B_PREDICTION_MODE b_mode =
-                    xd->mode_info_context->bmi[i].as_mode;
-                unsigned char *Above;
-                unsigned char *yleft;
-                int left_stride;
-                unsigned char top_left;
+      for (i = 0; i < 16; i++) {
+        BLOCKD *b = &xd->block[i];
+        unsigned char *dst = xd->dst.y_buffer + b->offset;
+        B_PREDICTION_MODE b_mode = xd->mode_info_context->bmi[i].as_mode;
+        unsigned char *Above;
+        unsigned char *yleft;
+        int left_stride;
+        unsigned char top_left;
 
-                /*Caution: For some b_mode, it needs 8 pixels (4 above + 4 above-right).*/
-                if (i < 4 && pbi->common.filter_level)
-                    Above = xd->recon_above[0] + b->offset;
-                else
-                    Above = dst - dst_stride;
+        /*Caution: For some b_mode, it needs 8 pixels (4 above + 4
+         * above-right).*/
+        if (i < 4 && pbi->common.filter_level)
+          Above = xd->recon_above[0] + b->offset;
+        else
+          Above = dst - dst_stride;
 
-                if (i%4==0 && pbi->common.filter_level)
-                {
-                    yleft = xd->recon_left[0] + i;
-                    left_stride = 1;
-                }
-                else
-                {
-                    yleft = dst - 1;
-                    left_stride = dst_stride;
-                }
+        if (i % 4 == 0 && pbi->common.filter_level) {
+          yleft = xd->recon_left[0] + i;
+          left_stride = 1;
+        } else {
+          yleft = dst - 1;
+          left_stride = dst_stride;
+        }
 
-                if ((i==4 || i==8 || i==12) && pbi->common.filter_level)
-                    top_left = *(xd->recon_left[0] + i - 1);
-                else
-                    top_left = Above[-1];
+        if ((i == 4 || i == 8 || i == 12) && pbi->common.filter_level)
+          top_left = *(xd->recon_left[0] + i - 1);
+        else
+          top_left = Above[-1];
 
-                vp8_intra4x4_predict(Above, yleft, left_stride,
-                                     b_mode, dst, dst_stride, top_left);
+        vp8_intra4x4_predict(Above, yleft, left_stride, b_mode, dst, dst_stride,
+                             top_left);
 
-                if (xd->eobs[i] )
-                {
-                    if (xd->eobs[i] > 1)
-                    {
-                        vp8_dequant_idct_add(b->qcoeff, DQC, dst, dst_stride);
-                    }
-                    else
-                    {
-                        vp8_dc_only_idct_add(b->qcoeff[0] * DQC[0],
-                                             dst, dst_stride, dst, dst_stride);
-                        memset(b->qcoeff, 0, 2 * sizeof(b->qcoeff[0]));
-                    }
-                }
-            }
+        if (xd->eobs[i]) {
+          if (xd->eobs[i] > 1) {
+            vp8_dequant_idct_add(b->qcoeff, DQC, dst, dst_stride);
+          } else {
+            vp8_dc_only_idct_add(b->qcoeff[0] * DQC[0], dst, dst_stride, dst,
+                                 dst_stride);
+            memset(b->qcoeff, 0, 2 * sizeof(b->qcoeff[0]));
+          }
         }
+      }
     }
-    else
-    {
-        vp8_build_inter_predictors_mb(xd);
-    }
+  } else {
+    vp8_build_inter_predictors_mb(xd);
+  }
 
-
 #if CONFIG_ERROR_CONCEALMENT
-    if (corruption_detected)
-    {
-        return;
-    }
+  if (corruption_detected) {
+    return;
+  }
 #endif
 
-    if(!xd->mode_info_context->mbmi.mb_skip_coeff)
-    {
-        /* dequantization and idct */
-        if (mode != B_PRED)
-        {
-            short *DQC = xd->dequant_y1;
+  if (!xd->mode_info_context->mbmi.mb_skip_coeff) {
+    /* dequantization and idct */
+    if (mode != B_PRED) {
+      short *DQC = xd->dequant_y1;
 
-            if (mode != SPLITMV)
-            {
-                BLOCKD *b = &xd->block[24];
+      if (mode != SPLITMV) {
+        BLOCKD *b = &xd->block[24];
 
-                /* do 2nd order transform on the dc block */
-                if (xd->eobs[24] > 1)
-                {
-                    vp8_dequantize_b(b, xd->dequant_y2);
+        /* do 2nd order transform on the dc block */
+        if (xd->eobs[24] > 1) {
+          vp8_dequantize_b(b, xd->dequant_y2);
 
-                    vp8_short_inv_walsh4x4(&b->dqcoeff[0],
-                        xd->qcoeff);
-                    memset(b->qcoeff, 0, 16 * sizeof(b->qcoeff[0]));
-                }
-                else
-                {
-                    b->dqcoeff[0] = b->qcoeff[0] * xd->dequant_y2[0];
-                    vp8_short_inv_walsh4x4_1(&b->dqcoeff[0],
-                        xd->qcoeff);
-                    memset(b->qcoeff, 0, 2 * sizeof(b->qcoeff[0]));
-                }
-
-                /* override the dc dequant constant in order to preserve the
-                 * dc components
-                 */
-                DQC = xd->dequant_y1_dc;
-            }
-
-            vp8_dequant_idct_add_y_block
-                            (xd->qcoeff, DQC,
-                             xd->dst.y_buffer,
-                             xd->dst.y_stride, xd->eobs);
+          vp8_short_inv_walsh4x4(&b->dqcoeff[0], xd->qcoeff);
+          memset(b->qcoeff, 0, 16 * sizeof(b->qcoeff[0]));
+        } else {
+          b->dqcoeff[0] = b->qcoeff[0] * xd->dequant_y2[0];
+          vp8_short_inv_walsh4x4_1(&b->dqcoeff[0], xd->qcoeff);
+          memset(b->qcoeff, 0, 2 * sizeof(b->qcoeff[0]));
         }
 
-        vp8_dequant_idct_add_uv_block
-                        (xd->qcoeff+16*16, xd->dequant_uv,
-                         xd->dst.u_buffer, xd->dst.v_buffer,
-                         xd->dst.uv_stride, xd->eobs+16);
+        /* override the dc dequant constant in order to preserve the
+         * dc components
+         */
+        DQC = xd->dequant_y1_dc;
+      }
+
+      vp8_dequant_idct_add_y_block(xd->qcoeff, DQC, xd->dst.y_buffer,
+                                   xd->dst.y_stride, xd->eobs);
     }
+
+    vp8_dequant_idct_add_uv_block(xd->qcoeff + 16 * 16, xd->dequant_uv,
+                                  xd->dst.u_buffer, xd->dst.v_buffer,
+                                  xd->dst.uv_stride, xd->eobs + 16);
+  }
 }
 
-static void mt_decode_mb_rows(VP8D_COMP *pbi, MACROBLOCKD *xd, int start_mb_row)
-{
-    const int *last_row_current_mb_col;
-    int *current_mb_col;
-    int mb_row;
-    VP8_COMMON *pc = &pbi->common;
-    const int nsync = pbi->sync_range;
-    const int first_row_no_sync_above = pc->mb_cols + nsync;
-    int num_part = 1 << pbi->common.multi_token_partition;
-    int last_mb_row = start_mb_row;
+static void mt_decode_mb_rows(VP8D_COMP *pbi, MACROBLOCKD *xd,
+                              int start_mb_row) {
+  const int *last_row_current_mb_col;
+  int *current_mb_col;
+  int mb_row;
+  VP8_COMMON *pc = &pbi->common;
+  const int nsync = pbi->sync_range;
+  const int first_row_no_sync_above = pc->mb_cols + nsync;
+  int num_part = 1 << pbi->common.multi_token_partition;
+  int last_mb_row = start_mb_row;
 
-    YV12_BUFFER_CONFIG *yv12_fb_new = pbi->dec_fb_ref[INTRA_FRAME];
-    YV12_BUFFER_CONFIG *yv12_fb_lst = pbi->dec_fb_ref[LAST_FRAME];
+  YV12_BUFFER_CONFIG *yv12_fb_new = pbi->dec_fb_ref[INTRA_FRAME];
+  YV12_BUFFER_CONFIG *yv12_fb_lst = pbi->dec_fb_ref[LAST_FRAME];
 
-    int recon_y_stride = yv12_fb_new->y_stride;
-    int recon_uv_stride = yv12_fb_new->uv_stride;
+  int recon_y_stride = yv12_fb_new->y_stride;
+  int recon_uv_stride = yv12_fb_new->uv_stride;
 
-    unsigned char *ref_buffer[MAX_REF_FRAMES][3];
-    unsigned char *dst_buffer[3];
-    int i;
-    int ref_fb_corrupted[MAX_REF_FRAMES];
+  unsigned char *ref_buffer[MAX_REF_FRAMES][3];
+  unsigned char *dst_buffer[3];
+  int i;
+  int ref_fb_corrupted[MAX_REF_FRAMES];
 
-    ref_fb_corrupted[INTRA_FRAME] = 0;
+  ref_fb_corrupted[INTRA_FRAME] = 0;
 
-    for(i = 1; i < MAX_REF_FRAMES; i++)
-    {
-        YV12_BUFFER_CONFIG *this_fb = pbi->dec_fb_ref[i];
+  for (i = 1; i < MAX_REF_FRAMES; i++) {
+    YV12_BUFFER_CONFIG *this_fb = pbi->dec_fb_ref[i];
 
-        ref_buffer[i][0] = this_fb->y_buffer;
-        ref_buffer[i][1] = this_fb->u_buffer;
-        ref_buffer[i][2] = this_fb->v_buffer;
+    ref_buffer[i][0] = this_fb->y_buffer;
+    ref_buffer[i][1] = this_fb->u_buffer;
+    ref_buffer[i][2] = this_fb->v_buffer;
 
-        ref_fb_corrupted[i] = this_fb->corrupted;
-    }
+    ref_fb_corrupted[i] = this_fb->corrupted;
+  }
 
-    dst_buffer[0] = yv12_fb_new->y_buffer;
-    dst_buffer[1] = yv12_fb_new->u_buffer;
-    dst_buffer[2] = yv12_fb_new->v_buffer;
+  dst_buffer[0] = yv12_fb_new->y_buffer;
+  dst_buffer[1] = yv12_fb_new->u_buffer;
+  dst_buffer[2] = yv12_fb_new->v_buffer;
 
-    xd->up_available = (start_mb_row != 0);
+  xd->up_available = (start_mb_row != 0);
 
-    xd->mode_info_context = pc->mi + pc->mode_info_stride * start_mb_row;
-    xd->mode_info_stride = pc->mode_info_stride;
+  xd->mode_info_context = pc->mi + pc->mode_info_stride * start_mb_row;
+  xd->mode_info_stride = pc->mode_info_stride;
 
-    for (mb_row = start_mb_row; mb_row < pc->mb_rows; mb_row += (pbi->decoding_thread_count + 1))
-    {
-       int recon_yoffset, recon_uvoffset;
-       int mb_col;
-       int filter_level;
-       loop_filter_info_n *lfi_n = &pc->lf_info;
+  for (mb_row = start_mb_row; mb_row < pc->mb_rows;
+       mb_row += (pbi->decoding_thread_count + 1)) {
+    int recon_yoffset, recon_uvoffset;
+    int mb_col;
+    int filter_level;
+    loop_filter_info_n *lfi_n = &pc->lf_info;
 
-       /* save last row processed by this thread */
-       last_mb_row = mb_row;
-       /* select bool coder for current partition */
-       xd->current_bc =  &pbi->mbc[mb_row%num_part];
+    /* save last row processed by this thread */
+    last_mb_row = mb_row;
+    /* select bool coder for current partition */
+    xd->current_bc = &pbi->mbc[mb_row % num_part];
 
-       if (mb_row > 0)
-           last_row_current_mb_col = &pbi->mt_current_mb_col[mb_row -1];
-       else
-           last_row_current_mb_col = &first_row_no_sync_above;
+    if (mb_row > 0)
+      last_row_current_mb_col = &pbi->mt_current_mb_col[mb_row - 1];
+    else
+      last_row_current_mb_col = &first_row_no_sync_above;
 
-       current_mb_col = &pbi->mt_current_mb_col[mb_row];
+    current_mb_col = &pbi->mt_current_mb_col[mb_row];
 
-       recon_yoffset = mb_row * recon_y_stride * 16;
-       recon_uvoffset = mb_row * recon_uv_stride * 8;
+    recon_yoffset = mb_row * recon_y_stride * 16;
+    recon_uvoffset = mb_row * recon_uv_stride * 8;
 
-       /* reset contexts */
-       xd->above_context = pc->above_context;
-       memset(xd->left_context, 0, sizeof(ENTROPY_CONTEXT_PLANES));
+    /* reset contexts */
+    xd->above_context = pc->above_context;
+    memset(xd->left_context, 0, sizeof(ENTROPY_CONTEXT_PLANES));
 
-       xd->left_available = 0;
+    xd->left_available = 0;
 
-       xd->mb_to_top_edge = -((mb_row * 16)) << 3;
-       xd->mb_to_bottom_edge = ((pc->mb_rows - 1 - mb_row) * 16) << 3;
+    xd->mb_to_top_edge = -((mb_row * 16)) << 3;
+    xd->mb_to_bottom_edge = ((pc->mb_rows - 1 - mb_row) * 16) << 3;
 
-       if (pbi->common.filter_level)
-       {
-          xd->recon_above[0] = pbi->mt_yabove_row[mb_row] + 0*16 +32;
-          xd->recon_above[1] = pbi->mt_uabove_row[mb_row] + 0*8 +16;
-          xd->recon_above[2] = pbi->mt_vabove_row[mb_row] + 0*8 +16;
+    if (pbi->common.filter_level) {
+      xd->recon_above[0] = pbi->mt_yabove_row[mb_row] + 0 * 16 + 32;
+      xd->recon_above[1] = pbi->mt_uabove_row[mb_row] + 0 * 8 + 16;
+      xd->recon_above[2] = pbi->mt_vabove_row[mb_row] + 0 * 8 + 16;
 
-          xd->recon_left[0] = pbi->mt_yleft_col[mb_row];
-          xd->recon_left[1] = pbi->mt_uleft_col[mb_row];
-          xd->recon_left[2] = pbi->mt_vleft_col[mb_row];
+      xd->recon_left[0] = pbi->mt_yleft_col[mb_row];
+      xd->recon_left[1] = pbi->mt_uleft_col[mb_row];
+      xd->recon_left[2] = pbi->mt_vleft_col[mb_row];
 
-          /* TODO: move to outside row loop */
-          xd->recon_left_stride[0] = 1;
-          xd->recon_left_stride[1] = 1;
-       }
-       else
-       {
-          xd->recon_above[0] = dst_buffer[0] + recon_yoffset;
-          xd->recon_above[1] = dst_buffer[1] + recon_uvoffset;
-          xd->recon_above[2] = dst_buffer[2] + recon_uvoffset;
+      /* TODO: move to outside row loop */
+      xd->recon_left_stride[0] = 1;
+      xd->recon_left_stride[1] = 1;
+    } else {
+      xd->recon_above[0] = dst_buffer[0] + recon_yoffset;
+      xd->recon_above[1] = dst_buffer[1] + recon_uvoffset;
+      xd->recon_above[2] = dst_buffer[2] + recon_uvoffset;
 
-          xd->recon_left[0] = xd->recon_above[0] - 1;
-          xd->recon_left[1] = xd->recon_above[1] - 1;
-          xd->recon_left[2] = xd->recon_above[2] - 1;
+      xd->recon_left[0] = xd->recon_above[0] - 1;
+      xd->recon_left[1] = xd->recon_above[1] - 1;
+      xd->recon_left[2] = xd->recon_above[2] - 1;
 
-          xd->recon_above[0] -= xd->dst.y_stride;
-          xd->recon_above[1] -= xd->dst.uv_stride;
-          xd->recon_above[2] -= xd->dst.uv_stride;
+      xd->recon_above[0] -= xd->dst.y_stride;
+      xd->recon_above[1] -= xd->dst.uv_stride;
+      xd->recon_above[2] -= xd->dst.uv_stride;
 
-          /* TODO: move to outside row loop */
-          xd->recon_left_stride[0] = xd->dst.y_stride;
-          xd->recon_left_stride[1] = xd->dst.uv_stride;
+      /* TODO: move to outside row loop */
+      xd->recon_left_stride[0] = xd->dst.y_stride;
+      xd->recon_left_stride[1] = xd->dst.uv_stride;
 
-          setup_intra_recon_left(xd->recon_left[0], xd->recon_left[1],
-                                 xd->recon_left[2], xd->dst.y_stride,
-                                 xd->dst.uv_stride);
-       }
+      setup_intra_recon_left(xd->recon_left[0], xd->recon_left[1],
+                             xd->recon_left[2], xd->dst.y_stride,
+                             xd->dst.uv_stride);
+    }
 
-       for (mb_col = 0; mb_col < pc->mb_cols; mb_col++) {
-           if (((mb_col - 1) % nsync) == 0) {
-               pthread_mutex_t *mutex = &pbi->pmutex[mb_row];
-               protected_write(mutex, current_mb_col, mb_col - 1);
-           }
+    for (mb_col = 0; mb_col < pc->mb_cols; mb_col++) {
+      if (((mb_col - 1) % nsync) == 0) {
+        pthread_mutex_t *mutex = &pbi->pmutex[mb_row];
+        protected_write(mutex, current_mb_col, mb_col - 1);
+      }
 
-           if (mb_row && !(mb_col & (nsync - 1))) {
-               pthread_mutex_t *mutex = &pbi->pmutex[mb_row-1];
-               sync_read(mutex, mb_col, last_row_current_mb_col, nsync);
-           }
+      if (mb_row && !(mb_col & (nsync - 1))) {
+        pthread_mutex_t *mutex = &pbi->pmutex[mb_row - 1];
+        sync_read(mutex, mb_col, last_row_current_mb_col, nsync);
+      }
 
-           /* Distance of MB to the various image edges.
-            * These are specified to 8th pel as they are always
-            * compared to values that are in 1/8th pel units.
-            */
-           xd->mb_to_left_edge = -((mb_col * 16) << 3);
-           xd->mb_to_right_edge = ((pc->mb_cols - 1 - mb_col) * 16) << 3;
+      /* Distance of MB to the various image edges.
+       * These are specified to 8th pel as they are always
+       * compared to values that are in 1/8th pel units.
+       */
+      xd->mb_to_left_edge = -((mb_col * 16) << 3);
+      xd->mb_to_right_edge = ((pc->mb_cols - 1 - mb_col) * 16) << 3;
 
-    #if CONFIG_ERROR_CONCEALMENT
-           {
-               int corrupt_residual =
-                           (!pbi->independent_partitions &&
-                           pbi->frame_corrupt_residual) ||
-                           vp8dx_bool_error(xd->current_bc);
-               if (pbi->ec_active &&
-                   (xd->mode_info_context->mbmi.ref_frame ==
-                                                    INTRA_FRAME) &&
-                   corrupt_residual)
-               {
-                   /* We have an intra block with corrupt
-                    * coefficients, better to conceal with an inter
-                    * block.
-                    * Interpolate MVs from neighboring MBs
-                    *
-                    * Note that for the first mb with corrupt
-                    * residual in a frame, we might not discover
-                    * that before decoding the residual. That
-                    * happens after this check, and therefore no
-                    * inter concealment will be done.
-                    */
-                   vp8_interpolate_motion(xd,
-                                          mb_row, mb_col,
-                                          pc->mb_rows, pc->mb_cols);
-               }
-           }
-    #endif
+#if CONFIG_ERROR_CONCEALMENT
+      {
+        int corrupt_residual =
+            (!pbi->independent_partitions && pbi->frame_corrupt_residual) ||
+            vp8dx_bool_error(xd->current_bc);
+        if (pbi->ec_active &&
+            (xd->mode_info_context->mbmi.ref_frame == INTRA_FRAME) &&
+            corrupt_residual) {
+          /* We have an intra block with corrupt
+           * coefficients, better to conceal with an inter
+           * block.
+           * Interpolate MVs from neighboring MBs
+           *
+           * Note that for the first mb with corrupt
+           * residual in a frame, we might not discover
+           * that before decoding the residual. That
+           * happens after this check, and therefore no
+           * inter concealment will be done.
+           */
+          vp8_interpolate_motion(xd, mb_row, mb_col, pc->mb_rows, pc->mb_cols);
+        }
+      }
+#endif
 
+      xd->dst.y_buffer = dst_buffer[0] + recon_yoffset;
+      xd->dst.u_buffer = dst_buffer[1] + recon_uvoffset;
+      xd->dst.v_buffer = dst_buffer[2] + recon_uvoffset;
 
-           xd->dst.y_buffer = dst_buffer[0] + recon_yoffset;
-           xd->dst.u_buffer = dst_buffer[1] + recon_uvoffset;
-           xd->dst.v_buffer = dst_buffer[2] + recon_uvoffset;
+      xd->pre.y_buffer =
+          ref_buffer[xd->mode_info_context->mbmi.ref_frame][0] + recon_yoffset;
+      xd->pre.u_buffer =
+          ref_buffer[xd->mode_info_context->mbmi.ref_frame][1] + recon_uvoffset;
+      xd->pre.v_buffer =
+          ref_buffer[xd->mode_info_context->mbmi.ref_frame][2] + recon_uvoffset;
 
-           xd->pre.y_buffer = ref_buffer[xd->mode_info_context->mbmi.ref_frame][0] + recon_yoffset;
-           xd->pre.u_buffer = ref_buffer[xd->mode_info_context->mbmi.ref_frame][1] + recon_uvoffset;
-           xd->pre.v_buffer = ref_buffer[xd->mode_info_context->mbmi.ref_frame][2] + recon_uvoffset;
+      /* propagate errors from reference frames */
+      xd->corrupted |= ref_fb_corrupted[xd->mode_info_context->mbmi.ref_frame];
 
-           /* propagate errors from reference frames */
-           xd->corrupted |= ref_fb_corrupted[xd->mode_info_context->mbmi.ref_frame];
+      mt_decode_macroblock(pbi, xd, 0);
 
-           mt_decode_macroblock(pbi, xd, 0);
+      xd->left_available = 1;
 
-           xd->left_available = 1;
+      /* check if the boolean decoder has suffered an error */
+      xd->corrupted |= vp8dx_bool_error(xd->current_bc);
 
-           /* check if the boolean decoder has suffered an error */
-           xd->corrupted |= vp8dx_bool_error(xd->current_bc);
+      xd->recon_above[0] += 16;
+      xd->recon_above[1] += 8;
+      xd->recon_above[2] += 8;
 
-           xd->recon_above[0] += 16;
-           xd->recon_above[1] += 8;
-           xd->recon_above[2] += 8;
+      if (!pbi->common.filter_level) {
+        xd->recon_left[0] += 16;
+        xd->recon_left[1] += 8;
+        xd->recon_left[2] += 8;
+      }
 
-           if (!pbi->common.filter_level)
-           {
-              xd->recon_left[0] += 16;
-              xd->recon_left[1] += 8;
-              xd->recon_left[2] += 8;
-           }
+      if (pbi->common.filter_level) {
+        int skip_lf = (xd->mode_info_context->mbmi.mode != B_PRED &&
+                       xd->mode_info_context->mbmi.mode != SPLITMV &&
+                       xd->mode_info_context->mbmi.mb_skip_coeff);
 
-           if (pbi->common.filter_level)
-           {
-               int skip_lf = (xd->mode_info_context->mbmi.mode != B_PRED &&
-                               xd->mode_info_context->mbmi.mode != SPLITMV &&
-                               xd->mode_info_context->mbmi.mb_skip_coeff);
+        const int mode_index =
+            lfi_n->mode_lf_lut[xd->mode_info_context->mbmi.mode];
+        const int seg = xd->mode_info_context->mbmi.segment_id;
+        const int ref_frame = xd->mode_info_context->mbmi.ref_frame;
 
-               const int mode_index = lfi_n->mode_lf_lut[xd->mode_info_context->mbmi.mode];
-               const int seg = xd->mode_info_context->mbmi.segment_id;
-               const int ref_frame = xd->mode_info_context->mbmi.ref_frame;
+        filter_level = lfi_n->lvl[seg][ref_frame][mode_index];
 
-               filter_level = lfi_n->lvl[seg][ref_frame][mode_index];
+        if (mb_row != pc->mb_rows - 1) {
+          /* Save decoded MB last row data for next-row decoding */
+          memcpy((pbi->mt_yabove_row[mb_row + 1] + 32 + mb_col * 16),
+                 (xd->dst.y_buffer + 15 * recon_y_stride), 16);
+          memcpy((pbi->mt_uabove_row[mb_row + 1] + 16 + mb_col * 8),
+                 (xd->dst.u_buffer + 7 * recon_uv_stride), 8);
+          memcpy((pbi->mt_vabove_row[mb_row + 1] + 16 + mb_col * 8),
+                 (xd->dst.v_buffer + 7 * recon_uv_stride), 8);
+        }
 
-               if( mb_row != pc->mb_rows-1 )
-               {
-                   /* Save decoded MB last row data for next-row decoding */
-                   memcpy((pbi->mt_yabove_row[mb_row + 1] + 32 + mb_col*16), (xd->dst.y_buffer + 15 * recon_y_stride), 16);
-                   memcpy((pbi->mt_uabove_row[mb_row + 1] + 16 + mb_col*8), (xd->dst.u_buffer + 7 * recon_uv_stride), 8);
-                   memcpy((pbi->mt_vabove_row[mb_row + 1] + 16 + mb_col*8), (xd->dst.v_buffer + 7 * recon_uv_stride), 8);
-               }
+        /* save left_col for next MB decoding */
+        if (mb_col != pc->mb_cols - 1) {
+          MODE_INFO *next = xd->mode_info_context + 1;
 
-               /* save left_col for next MB decoding */
-               if(mb_col != pc->mb_cols-1)
-               {
-                   MODE_INFO *next = xd->mode_info_context +1;
+          if (next->mbmi.ref_frame == INTRA_FRAME) {
+            for (i = 0; i < 16; i++)
+              pbi->mt_yleft_col[mb_row][i] =
+                  xd->dst.y_buffer[i * recon_y_stride + 15];
+            for (i = 0; i < 8; i++) {
+              pbi->mt_uleft_col[mb_row][i] =
+                  xd->dst.u_buffer[i * recon_uv_stride + 7];
+              pbi->mt_vleft_col[mb_row][i] =
+                  xd->dst.v_buffer[i * recon_uv_stride + 7];
+            }
+          }
+        }
 
-                   if (next->mbmi.ref_frame == INTRA_FRAME)
-                   {
-                       for (i = 0; i < 16; i++)
-                           pbi->mt_yleft_col[mb_row][i] = xd->dst.y_buffer [i* recon_y_stride + 15];
-                       for (i = 0; i < 8; i++)
-                       {
-                           pbi->mt_uleft_col[mb_row][i] = xd->dst.u_buffer [i* recon_uv_stride + 7];
-                           pbi->mt_vleft_col[mb_row][i] = xd->dst.v_buffer [i* recon_uv_stride + 7];
-                       }
-                   }
-               }
+        /* loopfilter on this macroblock. */
+        if (filter_level) {
+          if (pc->filter_type == NORMAL_LOOPFILTER) {
+            loop_filter_info lfi;
+            FRAME_TYPE frame_type = pc->frame_type;
+            const int hev_index = lfi_n->hev_thr_lut[frame_type][filter_level];
+            lfi.mblim = lfi_n->mblim[filter_level];
+            lfi.blim = lfi_n->blim[filter_level];
+            lfi.lim = lfi_n->lim[filter_level];
+            lfi.hev_thr = lfi_n->hev_thr[hev_index];
 
-               /* loopfilter on this macroblock. */
-               if (filter_level)
-               {
-                   if(pc->filter_type == NORMAL_LOOPFILTER)
-                   {
-                       loop_filter_info lfi;
-                       FRAME_TYPE frame_type = pc->frame_type;
-                       const int hev_index = lfi_n->hev_thr_lut[frame_type][filter_level];
-                       lfi.mblim = lfi_n->mblim[filter_level];
-                       lfi.blim = lfi_n->blim[filter_level];
-                       lfi.lim = lfi_n->lim[filter_level];
-                       lfi.hev_thr = lfi_n->hev_thr[hev_index];
+            if (mb_col > 0)
+              vp8_loop_filter_mbv(xd->dst.y_buffer, xd->dst.u_buffer,
+                                  xd->dst.v_buffer, recon_y_stride,
+                                  recon_uv_stride, &lfi);
 
-                       if (mb_col > 0)
-                           vp8_loop_filter_mbv
-                           (xd->dst.y_buffer, xd->dst.u_buffer, xd->dst.v_buffer, recon_y_stride, recon_uv_stride, &lfi);
+            if (!skip_lf)
+              vp8_loop_filter_bv(xd->dst.y_buffer, xd->dst.u_buffer,
+                                 xd->dst.v_buffer, recon_y_stride,
+                                 recon_uv_stride, &lfi);
 
-                       if (!skip_lf)
-                           vp8_loop_filter_bv
-                           (xd->dst.y_buffer, xd->dst.u_buffer, xd->dst.v_buffer, recon_y_stride, recon_uv_stride, &lfi);
+            /* don't apply across umv border */
+            if (mb_row > 0)
+              vp8_loop_filter_mbh(xd->dst.y_buffer, xd->dst.u_buffer,
+                                  xd->dst.v_buffer, recon_y_stride,
+                                  recon_uv_stride, &lfi);
 
-                       /* don't apply across umv border */
-                       if (mb_row > 0)
-                           vp8_loop_filter_mbh
-                           (xd->dst.y_buffer, xd->dst.u_buffer, xd->dst.v_buffer, recon_y_stride, recon_uv_stride, &lfi);
+            if (!skip_lf)
+              vp8_loop_filter_bh(xd->dst.y_buffer, xd->dst.u_buffer,
+                                 xd->dst.v_buffer, recon_y_stride,
+                                 recon_uv_stride, &lfi);
+          } else {
+            if (mb_col > 0)
+              vp8_loop_filter_simple_mbv(xd->dst.y_buffer, recon_y_stride,
+                                         lfi_n->mblim[filter_level]);
 
-                       if (!skip_lf)
-                           vp8_loop_filter_bh
-                           (xd->dst.y_buffer, xd->dst.u_buffer, xd->dst.v_buffer,  recon_y_stride, recon_uv_stride, &lfi);
-                   }
-                   else
-                   {
-                       if (mb_col > 0)
-                           vp8_loop_filter_simple_mbv
-                           (xd->dst.y_buffer, recon_y_stride, lfi_n->mblim[filter_level]);
+            if (!skip_lf)
+              vp8_loop_filter_simple_bv(xd->dst.y_buffer, recon_y_stride,
+                                        lfi_n->blim[filter_level]);
 
-                       if (!skip_lf)
-                           vp8_loop_filter_simple_bv
-                           (xd->dst.y_buffer, recon_y_stride, lfi_n->blim[filter_level]);
+            /* don't apply across umv border */
+            if (mb_row > 0)
+              vp8_loop_filter_simple_mbh(xd->dst.y_buffer, recon_y_stride,
+                                         lfi_n->mblim[filter_level]);
 
-                       /* don't apply across umv border */
-                       if (mb_row > 0)
-                           vp8_loop_filter_simple_mbh
-                           (xd->dst.y_buffer, recon_y_stride, lfi_n->mblim[filter_level]);
+            if (!skip_lf)
+              vp8_loop_filter_simple_bh(xd->dst.y_buffer, recon_y_stride,
+                                        lfi_n->blim[filter_level]);
+          }
+        }
+      }
 
-                       if (!skip_lf)
-                           vp8_loop_filter_simple_bh
-                           (xd->dst.y_buffer, recon_y_stride, lfi_n->blim[filter_level]);
-                   }
-               }
+      recon_yoffset += 16;
+      recon_uvoffset += 8;
 
-           }
+      ++xd->mode_info_context; /* next mb */
 
-           recon_yoffset += 16;
-           recon_uvoffset += 8;
+      xd->above_context++;
+    }
 
-           ++xd->mode_info_context;  /* next mb */
+    /* adjust to the next row of mbs */
+    if (pbi->common.filter_level) {
+      if (mb_row != pc->mb_rows - 1) {
+        int lasty = yv12_fb_lst->y_width + VP8BORDERINPIXELS;
+        int lastuv = (yv12_fb_lst->y_width >> 1) + (VP8BORDERINPIXELS >> 1);
 
-           xd->above_context++;
-       }
+        for (i = 0; i < 4; i++) {
+          pbi->mt_yabove_row[mb_row + 1][lasty + i] =
+              pbi->mt_yabove_row[mb_row + 1][lasty - 1];
+          pbi->mt_uabove_row[mb_row + 1][lastuv + i] =
+              pbi->mt_uabove_row[mb_row + 1][lastuv - 1];
+          pbi->mt_vabove_row[mb_row + 1][lastuv + i] =
+              pbi->mt_vabove_row[mb_row + 1][lastuv - 1];
+        }
+      }
+    } else
+      vp8_extend_mb_row(yv12_fb_new, xd->dst.y_buffer + 16,
+                        xd->dst.u_buffer + 8, xd->dst.v_buffer + 8);
 
-       /* adjust to the next row of mbs */
-       if (pbi->common.filter_level)
-       {
-           if(mb_row != pc->mb_rows-1)
-           {
-               int lasty = yv12_fb_lst->y_width + VP8BORDERINPIXELS;
-               int lastuv = (yv12_fb_lst->y_width>>1) + (VP8BORDERINPIXELS>>1);
+    /* last MB of row is ready just after extension is done */
+    protected_write(&pbi->pmutex[mb_row], current_mb_col, mb_col + nsync);
 
-               for (i = 0; i < 4; i++)
-               {
-                   pbi->mt_yabove_row[mb_row +1][lasty + i] = pbi->mt_yabove_row[mb_row +1][lasty -1];
-                   pbi->mt_uabove_row[mb_row +1][lastuv + i] = pbi->mt_uabove_row[mb_row +1][lastuv -1];
-                   pbi->mt_vabove_row[mb_row +1][lastuv + i] = pbi->mt_vabove_row[mb_row +1][lastuv -1];
-               }
-           }
-       }
-       else
-           vp8_extend_mb_row(yv12_fb_new, xd->dst.y_buffer + 16,
-                             xd->dst.u_buffer + 8, xd->dst.v_buffer + 8);
+    ++xd->mode_info_context; /* skip prediction column */
+    xd->up_available = 1;
 
-       /* last MB of row is ready just after extension is done */
-       protected_write(&pbi->pmutex[mb_row], current_mb_col, mb_col + nsync);
+    /* since we have multithread */
+    xd->mode_info_context += xd->mode_info_stride * pbi->decoding_thread_count;
+  }
 
-       ++xd->mode_info_context;      /* skip prediction column */
-       xd->up_available = 1;
-
-       /* since we have multithread */
-       xd->mode_info_context += xd->mode_info_stride * pbi->decoding_thread_count;
-    }
-
-    /* signal end of frame decoding if this thread processed the last mb_row */
-    if (last_mb_row == (pc->mb_rows - 1))
-        sem_post(&pbi->h_event_end_decoding);
-
+  /* signal end of frame decoding if this thread processed the last mb_row */
+  if (last_mb_row == (pc->mb_rows - 1)) sem_post(&pbi->h_event_end_decoding);
 }
 
+static THREAD_FUNCTION thread_decoding_proc(void *p_data) {
+  int ithread = ((DECODETHREAD_DATA *)p_data)->ithread;
+  VP8D_COMP *pbi = (VP8D_COMP *)(((DECODETHREAD_DATA *)p_data)->ptr1);
+  MB_ROW_DEC *mbrd = (MB_ROW_DEC *)(((DECODETHREAD_DATA *)p_data)->ptr2);
+  ENTROPY_CONTEXT_PLANES mb_row_left_context;
 
-static THREAD_FUNCTION thread_decoding_proc(void *p_data)
-{
-    int ithread = ((DECODETHREAD_DATA *)p_data)->ithread;
-    VP8D_COMP *pbi = (VP8D_COMP *)(((DECODETHREAD_DATA *)p_data)->ptr1);
-    MB_ROW_DEC *mbrd = (MB_ROW_DEC *)(((DECODETHREAD_DATA *)p_data)->ptr2);
-    ENTROPY_CONTEXT_PLANES mb_row_left_context;
+  while (1) {
+    if (protected_read(&pbi->mt_mutex, &pbi->b_multithreaded_rd) == 0) break;
 
-    while (1)
-    {
-        if (protected_read(&pbi->mt_mutex, &pbi->b_multithreaded_rd) == 0)
-            break;
+    if (sem_wait(&pbi->h_event_start_decoding[ithread]) == 0) {
+      if (protected_read(&pbi->mt_mutex, &pbi->b_multithreaded_rd) == 0)
+        break;
+      else {
+        MACROBLOCKD *xd = &mbrd->mbd;
+        xd->left_context = &mb_row_left_context;
 
-        if (sem_wait(&pbi->h_event_start_decoding[ithread]) == 0)
-        {
-            if (protected_read(&pbi->mt_mutex, &pbi->b_multithreaded_rd) == 0)
-                break;
-            else
-            {
-                MACROBLOCKD *xd = &mbrd->mbd;
-                xd->left_context = &mb_row_left_context;
-
-                mt_decode_mb_rows(pbi, xd, ithread+1);
-            }
-        }
+        mt_decode_mb_rows(pbi, xd, ithread + 1);
+      }
     }
+  }
 
-    return 0 ;
+  return 0;
 }
 
+void vp8_decoder_create_threads(VP8D_COMP *pbi) {
+  int core_count = 0;
+  unsigned int ithread;
 
-void vp8_decoder_create_threads(VP8D_COMP *pbi)
-{
-    int core_count = 0;
-    unsigned int ithread;
+  pbi->b_multithreaded_rd = 0;
+  pbi->allocated_decoding_thread_count = 0;
+  pthread_mutex_init(&pbi->mt_mutex, NULL);
 
-    pbi->b_multithreaded_rd = 0;
-    pbi->allocated_decoding_thread_count = 0;
-    pthread_mutex_init(&pbi->mt_mutex, NULL);
+  /* limit decoding threads to the max number of token partitions */
+  core_count = (pbi->max_threads > 8) ? 8 : pbi->max_threads;
 
-    /* limit decoding threads to the max number of token partitions */
-    core_count = (pbi->max_threads > 8) ? 8 : pbi->max_threads;
+  /* limit decoding threads to the available cores */
+  if (core_count > pbi->common.processor_core_count)
+    core_count = pbi->common.processor_core_count;
 
-    /* limit decoding threads to the available cores */
-    if (core_count > pbi->common.processor_core_count)
-        core_count = pbi->common.processor_core_count;
+  if (core_count > 1) {
+    pbi->b_multithreaded_rd = 1;
+    pbi->decoding_thread_count = core_count - 1;
 
-    if (core_count > 1)
-    {
-        pbi->b_multithreaded_rd = 1;
-        pbi->decoding_thread_count = core_count - 1;
+    CALLOC_ARRAY(pbi->h_decoding_thread, pbi->decoding_thread_count);
+    CALLOC_ARRAY(pbi->h_event_start_decoding, pbi->decoding_thread_count);
+    CALLOC_ARRAY_ALIGNED(pbi->mb_row_di, pbi->decoding_thread_count, 32);
+    CALLOC_ARRAY(pbi->de_thread_data, pbi->decoding_thread_count);
 
-        CALLOC_ARRAY(pbi->h_decoding_thread, pbi->decoding_thread_count);
-        CALLOC_ARRAY(pbi->h_event_start_decoding, pbi->decoding_thread_count);
-        CALLOC_ARRAY_ALIGNED(pbi->mb_row_di, pbi->decoding_thread_count, 32);
-        CALLOC_ARRAY(pbi->de_thread_data, pbi->decoding_thread_count);
+    for (ithread = 0; ithread < pbi->decoding_thread_count; ithread++) {
+      sem_init(&pbi->h_event_start_decoding[ithread], 0, 0);
 
-        for (ithread = 0; ithread < pbi->decoding_thread_count; ithread++)
-        {
-            sem_init(&pbi->h_event_start_decoding[ithread], 0, 0);
+      vp8_setup_block_dptrs(&pbi->mb_row_di[ithread].mbd);
 
-            vp8_setup_block_dptrs(&pbi->mb_row_di[ithread].mbd);
+      pbi->de_thread_data[ithread].ithread = ithread;
+      pbi->de_thread_data[ithread].ptr1 = (void *)pbi;
+      pbi->de_thread_data[ithread].ptr2 = (void *)&pbi->mb_row_di[ithread];
 
-            pbi->de_thread_data[ithread].ithread  = ithread;
-            pbi->de_thread_data[ithread].ptr1     = (void *)pbi;
-            pbi->de_thread_data[ithread].ptr2     = (void *) &pbi->mb_row_di[ithread];
+      pthread_create(&pbi->h_decoding_thread[ithread], 0, thread_decoding_proc,
+                     (&pbi->de_thread_data[ithread]));
+    }
 
-            pthread_create(&pbi->h_decoding_thread[ithread], 0, thread_decoding_proc, (&pbi->de_thread_data[ithread]));
-        }
+    sem_init(&pbi->h_event_end_decoding, 0, 0);
 
-        sem_init(&pbi->h_event_end_decoding, 0, 0);
-
-        pbi->allocated_decoding_thread_count = pbi->decoding_thread_count;
-    }
+    pbi->allocated_decoding_thread_count = pbi->decoding_thread_count;
+  }
 }
 
+void vp8mt_de_alloc_temp_buffers(VP8D_COMP *pbi, int mb_rows) {
+  int i;
 
-void vp8mt_de_alloc_temp_buffers(VP8D_COMP *pbi, int mb_rows)
-{
-    int i;
+  if (protected_read(&pbi->mt_mutex, &pbi->b_multithreaded_rd)) {
+    /* De-allocate mutex */
+    if (pbi->pmutex != NULL) {
+      for (i = 0; i < mb_rows; i++) {
+        pthread_mutex_destroy(&pbi->pmutex[i]);
+      }
+      vpx_free(pbi->pmutex);
+      pbi->pmutex = NULL;
+    }
 
-    if (protected_read(&pbi->mt_mutex, &pbi->b_multithreaded_rd))
-    {
-        /* De-allocate mutex */
-        if (pbi->pmutex != NULL) {
-            for (i = 0; i < mb_rows; i++) {
-                pthread_mutex_destroy(&pbi->pmutex[i]);
-            }
-            vpx_free(pbi->pmutex);
-            pbi->pmutex = NULL;
-        }
+    vpx_free(pbi->mt_current_mb_col);
+    pbi->mt_current_mb_col = NULL;
 
-            vpx_free(pbi->mt_current_mb_col);
-            pbi->mt_current_mb_col = NULL ;
+    /* Free above_row buffers. */
+    if (pbi->mt_yabove_row) {
+      for (i = 0; i < mb_rows; i++) {
+        vpx_free(pbi->mt_yabove_row[i]);
+        pbi->mt_yabove_row[i] = NULL;
+      }
+      vpx_free(pbi->mt_yabove_row);
+      pbi->mt_yabove_row = NULL;
+    }
 
-        /* Free above_row buffers. */
-        if (pbi->mt_yabove_row)
-        {
-            for (i=0; i< mb_rows; i++)
-            {
-                    vpx_free(pbi->mt_yabove_row[i]);
-                    pbi->mt_yabove_row[i] = NULL ;
-            }
-            vpx_free(pbi->mt_yabove_row);
-            pbi->mt_yabove_row = NULL ;
-        }
+    if (pbi->mt_uabove_row) {
+      for (i = 0; i < mb_rows; i++) {
+        vpx_free(pbi->mt_uabove_row[i]);
+        pbi->mt_uabove_row[i] = NULL;
+      }
+      vpx_free(pbi->mt_uabove_row);
+      pbi->mt_uabove_row = NULL;
+    }
 
-        if (pbi->mt_uabove_row)
-        {
-            for (i=0; i< mb_rows; i++)
-            {
-                    vpx_free(pbi->mt_uabove_row[i]);
-                    pbi->mt_uabove_row[i] = NULL ;
-            }
-            vpx_free(pbi->mt_uabove_row);
-            pbi->mt_uabove_row = NULL ;
-        }
+    if (pbi->mt_vabove_row) {
+      for (i = 0; i < mb_rows; i++) {
+        vpx_free(pbi->mt_vabove_row[i]);
+        pbi->mt_vabove_row[i] = NULL;
+      }
+      vpx_free(pbi->mt_vabove_row);
+      pbi->mt_vabove_row = NULL;
+    }
 
-        if (pbi->mt_vabove_row)
-        {
-            for (i=0; i< mb_rows; i++)
-            {
-                    vpx_free(pbi->mt_vabove_row[i]);
-                    pbi->mt_vabove_row[i] = NULL ;
-            }
-            vpx_free(pbi->mt_vabove_row);
-            pbi->mt_vabove_row = NULL ;
-        }
+    /* Free left_col buffers. */
+    if (pbi->mt_yleft_col) {
+      for (i = 0; i < mb_rows; i++) {
+        vpx_free(pbi->mt_yleft_col[i]);
+        pbi->mt_yleft_col[i] = NULL;
+      }
+      vpx_free(pbi->mt_yleft_col);
+      pbi->mt_yleft_col = NULL;
+    }
 
-        /* Free left_col buffers. */
-        if (pbi->mt_yleft_col)
-        {
-            for (i=0; i< mb_rows; i++)
-            {
-                    vpx_free(pbi->mt_yleft_col[i]);
-                    pbi->mt_yleft_col[i] = NULL ;
-            }
-            vpx_free(pbi->mt_yleft_col);
-            pbi->mt_yleft_col = NULL ;
-        }
+    if (pbi->mt_uleft_col) {
+      for (i = 0; i < mb_rows; i++) {
+        vpx_free(pbi->mt_uleft_col[i]);
+        pbi->mt_uleft_col[i] = NULL;
+      }
+      vpx_free(pbi->mt_uleft_col);
+      pbi->mt_uleft_col = NULL;
+    }
 
-        if (pbi->mt_uleft_col)
-        {
-            for (i=0; i< mb_rows; i++)
-            {
-                    vpx_free(pbi->mt_uleft_col[i]);
-                    pbi->mt_uleft_col[i] = NULL ;
-            }
-            vpx_free(pbi->mt_uleft_col);
-            pbi->mt_uleft_col = NULL ;
-        }
-
-        if (pbi->mt_vleft_col)
-        {
-            for (i=0; i< mb_rows; i++)
-            {
-                    vpx_free(pbi->mt_vleft_col[i]);
-                    pbi->mt_vleft_col[i] = NULL ;
-            }
-            vpx_free(pbi->mt_vleft_col);
-            pbi->mt_vleft_col = NULL ;
-        }
+    if (pbi->mt_vleft_col) {
+      for (i = 0; i < mb_rows; i++) {
+        vpx_free(pbi->mt_vleft_col[i]);
+        pbi->mt_vleft_col[i] = NULL;
+      }
+      vpx_free(pbi->mt_vleft_col);
+      pbi->mt_vleft_col = NULL;
     }
+  }
 }
 
+void vp8mt_alloc_temp_buffers(VP8D_COMP *pbi, int width, int prev_mb_rows) {
+  VP8_COMMON *const pc = &pbi->common;
+  int i;
+  int uv_width;
 
-void vp8mt_alloc_temp_buffers(VP8D_COMP *pbi, int width, int prev_mb_rows)
-{
-    VP8_COMMON *const pc = & pbi->common;
-    int i;
-    int uv_width;
+  if (protected_read(&pbi->mt_mutex, &pbi->b_multithreaded_rd)) {
+    vp8mt_de_alloc_temp_buffers(pbi, prev_mb_rows);
 
-    if (protected_read(&pbi->mt_mutex, &pbi->b_multithreaded_rd))
-    {
-        vp8mt_de_alloc_temp_buffers(pbi, prev_mb_rows);
+    /* our internal buffers are always multiples of 16 */
+    if ((width & 0xf) != 0) width += 16 - (width & 0xf);
 
-        /* our internal buffers are always multiples of 16 */
-        if ((width & 0xf) != 0)
-            width += 16 - (width & 0xf);
+    if (width < 640)
+      pbi->sync_range = 1;
+    else if (width <= 1280)
+      pbi->sync_range = 8;
+    else if (width <= 2560)
+      pbi->sync_range = 16;
+    else
+      pbi->sync_range = 32;
 
-        if (width < 640) pbi->sync_range = 1;
-        else if (width <= 1280) pbi->sync_range = 8;
-        else if (width <= 2560) pbi->sync_range =16;
-        else pbi->sync_range = 32;
+    uv_width = width >> 1;
 
-        uv_width = width >>1;
+    /* Allocate mutex */
+    CHECK_MEM_ERROR(pbi->pmutex,
+                    vpx_malloc(sizeof(*pbi->pmutex) * pc->mb_rows));
+    if (pbi->pmutex) {
+      for (i = 0; i < pc->mb_rows; i++) {
+        pthread_mutex_init(&pbi->pmutex[i], NULL);
+      }
+    }
 
-        /* Allocate mutex */
-        CHECK_MEM_ERROR(pbi->pmutex, vpx_malloc(sizeof(*pbi->pmutex) *
-                                                pc->mb_rows));
-        if (pbi->pmutex) {
-            for (i = 0; i < pc->mb_rows; i++) {
-                pthread_mutex_init(&pbi->pmutex[i], NULL);
-            }
-        }
+    /* Allocate an int for each mb row. */
+    CALLOC_ARRAY(pbi->mt_current_mb_col, pc->mb_rows);
 
-        /* Allocate an int for each mb row. */
-        CALLOC_ARRAY(pbi->mt_current_mb_col, pc->mb_rows);
+    /* Allocate memory for above_row buffers. */
+    CALLOC_ARRAY(pbi->mt_yabove_row, pc->mb_rows);
+    for (i = 0; i < pc->mb_rows; i++)
+      CHECK_MEM_ERROR(pbi->mt_yabove_row[i],
+                      vpx_memalign(16, sizeof(unsigned char) *
+                                           (width + (VP8BORDERINPIXELS << 1))));
 
-        /* Allocate memory for above_row buffers. */
-        CALLOC_ARRAY(pbi->mt_yabove_row, pc->mb_rows);
-        for (i = 0; i < pc->mb_rows; i++)
-            CHECK_MEM_ERROR(pbi->mt_yabove_row[i], vpx_memalign(16,sizeof(unsigned char) * (width + (VP8BORDERINPIXELS<<1))));
+    CALLOC_ARRAY(pbi->mt_uabove_row, pc->mb_rows);
+    for (i = 0; i < pc->mb_rows; i++)
+      CHECK_MEM_ERROR(pbi->mt_uabove_row[i],
+                      vpx_memalign(16, sizeof(unsigned char) *
+                                           (uv_width + VP8BORDERINPIXELS)));
 
-        CALLOC_ARRAY(pbi->mt_uabove_row, pc->mb_rows);
-        for (i = 0; i < pc->mb_rows; i++)
-            CHECK_MEM_ERROR(pbi->mt_uabove_row[i], vpx_memalign(16,sizeof(unsigned char) * (uv_width + VP8BORDERINPIXELS)));
+    CALLOC_ARRAY(pbi->mt_vabove_row, pc->mb_rows);
+    for (i = 0; i < pc->mb_rows; i++)
+      CHECK_MEM_ERROR(pbi->mt_vabove_row[i],
+                      vpx_memalign(16, sizeof(unsigned char) *
+                                           (uv_width + VP8BORDERINPIXELS)));
 
-        CALLOC_ARRAY(pbi->mt_vabove_row, pc->mb_rows);
-        for (i = 0; i < pc->mb_rows; i++)
-            CHECK_MEM_ERROR(pbi->mt_vabove_row[i], vpx_memalign(16,sizeof(unsigned char) * (uv_width + VP8BORDERINPIXELS)));
+    /* Allocate memory for left_col buffers. */
+    CALLOC_ARRAY(pbi->mt_yleft_col, pc->mb_rows);
+    for (i = 0; i < pc->mb_rows; i++)
+      CHECK_MEM_ERROR(pbi->mt_yleft_col[i],
+                      vpx_calloc(sizeof(unsigned char) * 16, 1));
 
-        /* Allocate memory for left_col buffers. */
-        CALLOC_ARRAY(pbi->mt_yleft_col, pc->mb_rows);
-        for (i = 0; i < pc->mb_rows; i++)
-            CHECK_MEM_ERROR(pbi->mt_yleft_col[i], vpx_calloc(sizeof(unsigned char) * 16, 1));
+    CALLOC_ARRAY(pbi->mt_uleft_col, pc->mb_rows);
+    for (i = 0; i < pc->mb_rows; i++)
+      CHECK_MEM_ERROR(pbi->mt_uleft_col[i],
+                      vpx_calloc(sizeof(unsigned char) * 8, 1));
 
-        CALLOC_ARRAY(pbi->mt_uleft_col, pc->mb_rows);
-        for (i = 0; i < pc->mb_rows; i++)
-            CHECK_MEM_ERROR(pbi->mt_uleft_col[i], vpx_calloc(sizeof(unsigned char) * 8, 1));
-
-        CALLOC_ARRAY(pbi->mt_vleft_col, pc->mb_rows);
-        for (i = 0; i < pc->mb_rows; i++)
-            CHECK_MEM_ERROR(pbi->mt_vleft_col[i], vpx_calloc(sizeof(unsigned char) * 8, 1));
-    }
+    CALLOC_ARRAY(pbi->mt_vleft_col, pc->mb_rows);
+    for (i = 0; i < pc->mb_rows; i++)
+      CHECK_MEM_ERROR(pbi->mt_vleft_col[i],
+                      vpx_calloc(sizeof(unsigned char) * 8, 1));
+  }
 }
 
+void vp8_decoder_remove_threads(VP8D_COMP *pbi) {
+  /* shutdown MB Decoding thread; */
+  if (protected_read(&pbi->mt_mutex, &pbi->b_multithreaded_rd)) {
+    int i;
 
-void vp8_decoder_remove_threads(VP8D_COMP *pbi)
-{
-    /* shutdown MB Decoding thread; */
-    if (protected_read(&pbi->mt_mutex, &pbi->b_multithreaded_rd))
-    {
-        int i;
+    protected_write(&pbi->mt_mutex, &pbi->b_multithreaded_rd, 0);
 
-        protected_write(&pbi->mt_mutex, &pbi->b_multithreaded_rd, 0);
+    /* allow all threads to exit */
+    for (i = 0; i < pbi->allocated_decoding_thread_count; i++) {
+      sem_post(&pbi->h_event_start_decoding[i]);
+      pthread_join(pbi->h_decoding_thread[i], NULL);
+    }
 
-        /* allow all threads to exit */
-        for (i = 0; i < pbi->allocated_decoding_thread_count; i++)
-        {
-            sem_post(&pbi->h_event_start_decoding[i]);
-            pthread_join(pbi->h_decoding_thread[i], NULL);
-        }
+    for (i = 0; i < pbi->allocated_decoding_thread_count; i++) {
+      sem_destroy(&pbi->h_event_start_decoding[i]);
+    }
 
-        for (i = 0; i < pbi->allocated_decoding_thread_count; i++)
-        {
-            sem_destroy(&pbi->h_event_start_decoding[i]);
-        }
+    sem_destroy(&pbi->h_event_end_decoding);
 
-        sem_destroy(&pbi->h_event_end_decoding);
+    vpx_free(pbi->h_decoding_thread);
+    pbi->h_decoding_thread = NULL;
 
-            vpx_free(pbi->h_decoding_thread);
-            pbi->h_decoding_thread = NULL;
+    vpx_free(pbi->h_event_start_decoding);
+    pbi->h_event_start_decoding = NULL;
 
-            vpx_free(pbi->h_event_start_decoding);
-            pbi->h_event_start_decoding = NULL;
+    vpx_free(pbi->mb_row_di);
+    pbi->mb_row_di = NULL;
 
-            vpx_free(pbi->mb_row_di);
-            pbi->mb_row_di = NULL ;
-
-            vpx_free(pbi->de_thread_data);
-            pbi->de_thread_data = NULL;
-    }
-    pthread_mutex_destroy(&pbi->mt_mutex);
+    vpx_free(pbi->de_thread_data);
+    pbi->de_thread_data = NULL;
+  }
+  pthread_mutex_destroy(&pbi->mt_mutex);
 }
 
-void vp8mt_decode_mb_rows( VP8D_COMP *pbi, MACROBLOCKD *xd)
-{
-    VP8_COMMON *pc = &pbi->common;
-    unsigned int i;
-    int j;
+void vp8mt_decode_mb_rows(VP8D_COMP *pbi, MACROBLOCKD *xd) {
+  VP8_COMMON *pc = &pbi->common;
+  unsigned int i;
+  int j;
 
-    int filter_level = pc->filter_level;
-    YV12_BUFFER_CONFIG *yv12_fb_new = pbi->dec_fb_ref[INTRA_FRAME];
+  int filter_level = pc->filter_level;
+  YV12_BUFFER_CONFIG *yv12_fb_new = pbi->dec_fb_ref[INTRA_FRAME];
 
-    if (filter_level)
-    {
-        /* Set above_row buffer to 127 for decoding first MB row */
-        memset(pbi->mt_yabove_row[0] + VP8BORDERINPIXELS-1, 127, yv12_fb_new->y_width + 5);
-        memset(pbi->mt_uabove_row[0] + (VP8BORDERINPIXELS>>1)-1, 127, (yv12_fb_new->y_width>>1) +5);
-        memset(pbi->mt_vabove_row[0] + (VP8BORDERINPIXELS>>1)-1, 127, (yv12_fb_new->y_width>>1) +5);
+  if (filter_level) {
+    /* Set above_row buffer to 127 for decoding first MB row */
+    memset(pbi->mt_yabove_row[0] + VP8BORDERINPIXELS - 1, 127,
+           yv12_fb_new->y_width + 5);
+    memset(pbi->mt_uabove_row[0] + (VP8BORDERINPIXELS >> 1) - 1, 127,
+           (yv12_fb_new->y_width >> 1) + 5);
+    memset(pbi->mt_vabove_row[0] + (VP8BORDERINPIXELS >> 1) - 1, 127,
+           (yv12_fb_new->y_width >> 1) + 5);
 
-        for (j=1; j<pc->mb_rows; j++)
-        {
-            memset(pbi->mt_yabove_row[j] + VP8BORDERINPIXELS-1, (unsigned char)129, 1);
-            memset(pbi->mt_uabove_row[j] + (VP8BORDERINPIXELS>>1)-1, (unsigned char)129, 1);
-            memset(pbi->mt_vabove_row[j] + (VP8BORDERINPIXELS>>1)-1, (unsigned char)129, 1);
-        }
+    for (j = 1; j < pc->mb_rows; j++) {
+      memset(pbi->mt_yabove_row[j] + VP8BORDERINPIXELS - 1, (unsigned char)129,
+             1);
+      memset(pbi->mt_uabove_row[j] + (VP8BORDERINPIXELS >> 1) - 1,
+             (unsigned char)129, 1);
+      memset(pbi->mt_vabove_row[j] + (VP8BORDERINPIXELS >> 1) - 1,
+             (unsigned char)129, 1);
+    }
 
-        /* Set left_col to 129 initially */
-        for (j=0; j<pc->mb_rows; j++)
-        {
-            memset(pbi->mt_yleft_col[j], (unsigned char)129, 16);
-            memset(pbi->mt_uleft_col[j], (unsigned char)129, 8);
-            memset(pbi->mt_vleft_col[j], (unsigned char)129, 8);
-        }
-
-        /* Initialize the loop filter for this frame. */
-        vp8_loop_filter_frame_init(pc, &pbi->mb, filter_level);
+    /* Set left_col to 129 initially */
+    for (j = 0; j < pc->mb_rows; j++) {
+      memset(pbi->mt_yleft_col[j], (unsigned char)129, 16);
+      memset(pbi->mt_uleft_col[j], (unsigned char)129, 8);
+      memset(pbi->mt_vleft_col[j], (unsigned char)129, 8);
     }
-    else
-        vp8_setup_intra_recon_top_line(yv12_fb_new);
 
-    setup_decoding_thread_data(pbi, xd, pbi->mb_row_di, pbi->decoding_thread_count);
+    /* Initialize the loop filter for this frame. */
+    vp8_loop_filter_frame_init(pc, &pbi->mb, filter_level);
+  } else
+    vp8_setup_intra_recon_top_line(yv12_fb_new);
 
-    for (i = 0; i < pbi->decoding_thread_count; i++)
-        sem_post(&pbi->h_event_start_decoding[i]);
+  setup_decoding_thread_data(pbi, xd, pbi->mb_row_di,
+                             pbi->decoding_thread_count);
 
-    mt_decode_mb_rows(pbi, xd, 0);
+  for (i = 0; i < pbi->decoding_thread_count; i++)
+    sem_post(&pbi->h_event_start_decoding[i]);
 
-    sem_wait(&pbi->h_event_end_decoding);   /* add back for each frame */
+  mt_decode_mb_rows(pbi, xd, 0);
+
+  sem_wait(&pbi->h_event_end_decoding); /* add back for each frame */
 }
--- a/vp8/decoder/treereader.h
+++ b/vp8/decoder/treereader.h
@@ -8,7 +8,6 @@
  *  be found in the AUTHORS file in the root of the source tree.
  */
 
-
 #ifndef VP8_DECODER_TREEREADER_H_
 #define VP8_DECODER_TREEREADER_H_
 
@@ -26,20 +25,17 @@
 #define vp8_read_literal vp8_decode_value
 #define vp8_read_bit(R) vp8_read(R, vp8_prob_half)
 
-
 /* Intent of tree data structure is to make decoding trivial. */
 
 static INLINE int vp8_treed_read(
-    vp8_reader *const r,        /* !!! must return a 0 or 1 !!! */
-    vp8_tree t,
-    const vp8_prob *const p
-)
-{
-    register vp8_tree_index i = 0;
+    vp8_reader *const r, /* !!! must return a 0 or 1 !!! */
+    vp8_tree t, const vp8_prob *const p) {
+  register vp8_tree_index i = 0;
 
-    while ((i = t[ i + vp8_read(r, p[i>>1])]) > 0) ;
+  while ((i = t[i + vp8_read(r, p[i >> 1])]) > 0) {
+  }
 
-    return -i;
+  return -i;
 }
 
 #ifdef __cplusplus
--- a/vp8/encoder/arm/dct_arm.c
+++ b/vp8/encoder/arm/dct_arm.c
@@ -13,10 +13,9 @@
 
 #if HAVE_MEDIA
 
-void vp8_short_fdct8x4_armv6(short *input, short *output, int pitch)
-{
-    vp8_short_fdct4x4_armv6(input,   output,    pitch);
-    vp8_short_fdct4x4_armv6(input + 4, output + 16, pitch);
+void vp8_short_fdct8x4_armv6(short *input, short *output, int pitch) {
+  vp8_short_fdct4x4_armv6(input, output, pitch);
+  vp8_short_fdct4x4_armv6(input + 4, output + 16, pitch);
 }
 
 #endif /* HAVE_MEDIA */
--- a/vp8/encoder/arm/neon/denoising_neon.c
+++ b/vp8/encoder/arm/neon/denoising_neon.c
@@ -48,431 +48,413 @@
 int vp8_denoiser_filter_neon(unsigned char *mc_running_avg_y,
                              int mc_running_avg_y_stride,
                              unsigned char *running_avg_y,
-                             int running_avg_y_stride,
-                             unsigned char *sig, int sig_stride,
-                             unsigned int motion_magnitude,
+                             int running_avg_y_stride, unsigned char *sig,
+                             int sig_stride, unsigned int motion_magnitude,
                              int increase_denoising) {
-    /* If motion_magnitude is small, making the denoiser more aggressive by
-     * increasing the adjustment for each level, level1 adjustment is
-     * increased, the deltas stay the same.
-     */
-    int shift_inc  = (increase_denoising &&
-        motion_magnitude <= MOTION_MAGNITUDE_THRESHOLD) ? 1 : 0;
-    const uint8x16_t v_level1_adjustment = vmovq_n_u8(
-        (motion_magnitude <= MOTION_MAGNITUDE_THRESHOLD) ? 4 + shift_inc : 3);
-    const uint8x16_t v_delta_level_1_and_2 = vdupq_n_u8(1);
-    const uint8x16_t v_delta_level_2_and_3 = vdupq_n_u8(2);
-    const uint8x16_t v_level1_threshold = vmovq_n_u8(4 + shift_inc);
-    const uint8x16_t v_level2_threshold = vdupq_n_u8(8);
-    const uint8x16_t v_level3_threshold = vdupq_n_u8(16);
-    int64x2_t v_sum_diff_total = vdupq_n_s64(0);
+  /* If motion_magnitude is small, making the denoiser more aggressive by
+   * increasing the adjustment for each level, level1 adjustment is
+   * increased, the deltas stay the same.
+   */
+  int shift_inc =
+      (increase_denoising && motion_magnitude <= MOTION_MAGNITUDE_THRESHOLD)
+          ? 1
+          : 0;
+  const uint8x16_t v_level1_adjustment = vmovq_n_u8(
+      (motion_magnitude <= MOTION_MAGNITUDE_THRESHOLD) ? 4 + shift_inc : 3);
+  const uint8x16_t v_delta_level_1_and_2 = vdupq_n_u8(1);
+  const uint8x16_t v_delta_level_2_and_3 = vdupq_n_u8(2);
+  const uint8x16_t v_level1_threshold = vmovq_n_u8(4 + shift_inc);
+  const uint8x16_t v_level2_threshold = vdupq_n_u8(8);
+  const uint8x16_t v_level3_threshold = vdupq_n_u8(16);
+  int64x2_t v_sum_diff_total = vdupq_n_s64(0);
 
-    /* Go over lines. */
-    int r;
-    for (r = 0; r < 16; ++r) {
-        /* Load inputs. */
-        const uint8x16_t v_sig = vld1q_u8(sig);
-        const uint8x16_t v_mc_running_avg_y = vld1q_u8(mc_running_avg_y);
+  /* Go over lines. */
+  int r;
+  for (r = 0; r < 16; ++r) {
+    /* Load inputs. */
+    const uint8x16_t v_sig = vld1q_u8(sig);
+    const uint8x16_t v_mc_running_avg_y = vld1q_u8(mc_running_avg_y);
 
-        /* Calculate absolute difference and sign masks. */
-        const uint8x16_t v_abs_diff      = vabdq_u8(v_sig, v_mc_running_avg_y);
-        const uint8x16_t v_diff_pos_mask = vcltq_u8(v_sig, v_mc_running_avg_y);
-        const uint8x16_t v_diff_neg_mask = vcgtq_u8(v_sig, v_mc_running_avg_y);
+    /* Calculate absolute difference and sign masks. */
+    const uint8x16_t v_abs_diff = vabdq_u8(v_sig, v_mc_running_avg_y);
+    const uint8x16_t v_diff_pos_mask = vcltq_u8(v_sig, v_mc_running_avg_y);
+    const uint8x16_t v_diff_neg_mask = vcgtq_u8(v_sig, v_mc_running_avg_y);
 
-        /* Figure out which level that put us in. */
-        const uint8x16_t v_level1_mask = vcleq_u8(v_level1_threshold,
-                                                  v_abs_diff);
-        const uint8x16_t v_level2_mask = vcleq_u8(v_level2_threshold,
-                                                  v_abs_diff);
-        const uint8x16_t v_level3_mask = vcleq_u8(v_level3_threshold,
-                                                  v_abs_diff);
+    /* Figure out which level that put us in. */
+    const uint8x16_t v_level1_mask = vcleq_u8(v_level1_threshold, v_abs_diff);
+    const uint8x16_t v_level2_mask = vcleq_u8(v_level2_threshold, v_abs_diff);
+    const uint8x16_t v_level3_mask = vcleq_u8(v_level3_threshold, v_abs_diff);
 
-        /* Calculate absolute adjustments for level 1, 2 and 3. */
-        const uint8x16_t v_level2_adjustment = vandq_u8(v_level2_mask,
-                                                        v_delta_level_1_and_2);
-        const uint8x16_t v_level3_adjustment = vandq_u8(v_level3_mask,
-                                                        v_delta_level_2_and_3);
-        const uint8x16_t v_level1and2_adjustment = vaddq_u8(v_level1_adjustment,
-            v_level2_adjustment);
-        const uint8x16_t v_level1and2and3_adjustment = vaddq_u8(
-            v_level1and2_adjustment, v_level3_adjustment);
+    /* Calculate absolute adjustments for level 1, 2 and 3. */
+    const uint8x16_t v_level2_adjustment =
+        vandq_u8(v_level2_mask, v_delta_level_1_and_2);
+    const uint8x16_t v_level3_adjustment =
+        vandq_u8(v_level3_mask, v_delta_level_2_and_3);
+    const uint8x16_t v_level1and2_adjustment =
+        vaddq_u8(v_level1_adjustment, v_level2_adjustment);
+    const uint8x16_t v_level1and2and3_adjustment =
+        vaddq_u8(v_level1and2_adjustment, v_level3_adjustment);
 
-        /* Figure adjustment absolute value by selecting between the absolute
-         * difference if in level0 or the value for level 1, 2 and 3.
-         */
-        const uint8x16_t v_abs_adjustment = vbslq_u8(v_level1_mask,
-            v_level1and2and3_adjustment, v_abs_diff);
+    /* Figure adjustment absolute value by selecting between the absolute
+     * difference if in level0 or the value for level 1, 2 and 3.
+     */
+    const uint8x16_t v_abs_adjustment =
+        vbslq_u8(v_level1_mask, v_level1and2and3_adjustment, v_abs_diff);
 
-        /* Calculate positive and negative adjustments. Apply them to the signal
-         * and accumulate them. Adjustments are less than eight and the maximum
-         * sum of them (7 * 16) can fit in a signed char.
-         */
-        const uint8x16_t v_pos_adjustment = vandq_u8(v_diff_pos_mask,
-                                                     v_abs_adjustment);
-        const uint8x16_t v_neg_adjustment = vandq_u8(v_diff_neg_mask,
-                                                     v_abs_adjustment);
+    /* Calculate positive and negative adjustments. Apply them to the signal
+     * and accumulate them. Adjustments are less than eight and the maximum
+     * sum of them (7 * 16) can fit in a signed char.
+     */
+    const uint8x16_t v_pos_adjustment =
+        vandq_u8(v_diff_pos_mask, v_abs_adjustment);
+    const uint8x16_t v_neg_adjustment =
+        vandq_u8(v_diff_neg_mask, v_abs_adjustment);
 
-        uint8x16_t v_running_avg_y = vqaddq_u8(v_sig, v_pos_adjustment);
-        v_running_avg_y = vqsubq_u8(v_running_avg_y, v_neg_adjustment);
+    uint8x16_t v_running_avg_y = vqaddq_u8(v_sig, v_pos_adjustment);
+    v_running_avg_y = vqsubq_u8(v_running_avg_y, v_neg_adjustment);
 
-        /* Store results. */
-        vst1q_u8(running_avg_y, v_running_avg_y);
+    /* Store results. */
+    vst1q_u8(running_avg_y, v_running_avg_y);
 
-        /* Sum all the accumulators to have the sum of all pixel differences
-         * for this macroblock.
-         */
-        {
-            const int8x16_t v_sum_diff =
-                vqsubq_s8(vreinterpretq_s8_u8(v_pos_adjustment),
-                          vreinterpretq_s8_u8(v_neg_adjustment));
+    /* Sum all the accumulators to have the sum of all pixel differences
+     * for this macroblock.
+     */
+    {
+      const int8x16_t v_sum_diff =
+          vqsubq_s8(vreinterpretq_s8_u8(v_pos_adjustment),
+                    vreinterpretq_s8_u8(v_neg_adjustment));
 
-            const int16x8_t fe_dc_ba_98_76_54_32_10 = vpaddlq_s8(v_sum_diff);
+      const int16x8_t fe_dc_ba_98_76_54_32_10 = vpaddlq_s8(v_sum_diff);
 
-            const int32x4_t fedc_ba98_7654_3210 =
-                vpaddlq_s16(fe_dc_ba_98_76_54_32_10);
+      const int32x4_t fedc_ba98_7654_3210 =
+          vpaddlq_s16(fe_dc_ba_98_76_54_32_10);
 
-            const int64x2_t fedcba98_76543210 =
-                vpaddlq_s32(fedc_ba98_7654_3210);
+      const int64x2_t fedcba98_76543210 = vpaddlq_s32(fedc_ba98_7654_3210);
 
-            v_sum_diff_total = vqaddq_s64(v_sum_diff_total, fedcba98_76543210);
-        }
-
-        /* Update pointers for next iteration. */
-        sig += sig_stride;
-        mc_running_avg_y += mc_running_avg_y_stride;
-        running_avg_y += running_avg_y_stride;
+      v_sum_diff_total = vqaddq_s64(v_sum_diff_total, fedcba98_76543210);
     }
 
-    /* Too much adjustments => copy block. */
-    {
-        int64x1_t x = vqadd_s64(vget_high_s64(v_sum_diff_total),
-                                      vget_low_s64(v_sum_diff_total));
-        int sum_diff = vget_lane_s32(vabs_s32(vreinterpret_s32_s64(x)), 0);
-        int sum_diff_thresh = SUM_DIFF_THRESHOLD;
+    /* Update pointers for next iteration. */
+    sig += sig_stride;
+    mc_running_avg_y += mc_running_avg_y_stride;
+    running_avg_y += running_avg_y_stride;
+  }
 
-        if (increase_denoising) sum_diff_thresh = SUM_DIFF_THRESHOLD_HIGH;
-        if (sum_diff > sum_diff_thresh) {
-          // Before returning to copy the block (i.e., apply no denoising),
-          // checK if we can still apply some (weaker) temporal filtering to
-          // this block, that would otherwise not be denoised at all. Simplest
-          // is to apply an additional adjustment to running_avg_y to bring it
-          // closer to sig. The adjustment is capped by a maximum delta, and
-          // chosen such that in most cases the resulting sum_diff will be
-          // within the accceptable range given by sum_diff_thresh.
+  /* Too much adjustments => copy block. */
+  {
+    int64x1_t x = vqadd_s64(vget_high_s64(v_sum_diff_total),
+                            vget_low_s64(v_sum_diff_total));
+    int sum_diff = vget_lane_s32(vabs_s32(vreinterpret_s32_s64(x)), 0);
+    int sum_diff_thresh = SUM_DIFF_THRESHOLD;
 
-          // The delta is set by the excess of absolute pixel diff over the
-          // threshold.
-          int delta = ((sum_diff - sum_diff_thresh) >> 8) + 1;
-          // Only apply the adjustment for max delta up to 3.
-          if (delta < 4) {
-            const uint8x16_t k_delta = vmovq_n_u8(delta);
-            sig -= sig_stride * 16;
-            mc_running_avg_y -= mc_running_avg_y_stride * 16;
-            running_avg_y -= running_avg_y_stride * 16;
-            for (r = 0; r < 16; ++r) {
-              uint8x16_t v_running_avg_y = vld1q_u8(running_avg_y);
-              const uint8x16_t v_sig = vld1q_u8(sig);
-              const uint8x16_t v_mc_running_avg_y = vld1q_u8(mc_running_avg_y);
+    if (increase_denoising) sum_diff_thresh = SUM_DIFF_THRESHOLD_HIGH;
+    if (sum_diff > sum_diff_thresh) {
+      // Before returning to copy the block (i.e., apply no denoising),
+      // checK if we can still apply some (weaker) temporal filtering to
+      // this block, that would otherwise not be denoised at all. Simplest
+      // is to apply an additional adjustment to running_avg_y to bring it
+      // closer to sig. The adjustment is capped by a maximum delta, and
+      // chosen such that in most cases the resulting sum_diff will be
+      // within the accceptable range given by sum_diff_thresh.
 
-              /* Calculate absolute difference and sign masks. */
-              const uint8x16_t v_abs_diff      = vabdq_u8(v_sig,
-                                                          v_mc_running_avg_y);
-              const uint8x16_t v_diff_pos_mask = vcltq_u8(v_sig,
-                                                          v_mc_running_avg_y);
-              const uint8x16_t v_diff_neg_mask = vcgtq_u8(v_sig,
-                                                          v_mc_running_avg_y);
-              // Clamp absolute difference to delta to get the adjustment.
-              const uint8x16_t v_abs_adjustment =
-                  vminq_u8(v_abs_diff, (k_delta));
+      // The delta is set by the excess of absolute pixel diff over the
+      // threshold.
+      int delta = ((sum_diff - sum_diff_thresh) >> 8) + 1;
+      // Only apply the adjustment for max delta up to 3.
+      if (delta < 4) {
+        const uint8x16_t k_delta = vmovq_n_u8(delta);
+        sig -= sig_stride * 16;
+        mc_running_avg_y -= mc_running_avg_y_stride * 16;
+        running_avg_y -= running_avg_y_stride * 16;
+        for (r = 0; r < 16; ++r) {
+          uint8x16_t v_running_avg_y = vld1q_u8(running_avg_y);
+          const uint8x16_t v_sig = vld1q_u8(sig);
+          const uint8x16_t v_mc_running_avg_y = vld1q_u8(mc_running_avg_y);
 
-              const uint8x16_t v_pos_adjustment = vandq_u8(v_diff_pos_mask,
-                                                           v_abs_adjustment);
-              const uint8x16_t v_neg_adjustment = vandq_u8(v_diff_neg_mask,
-                                                           v_abs_adjustment);
+          /* Calculate absolute difference and sign masks. */
+          const uint8x16_t v_abs_diff = vabdq_u8(v_sig, v_mc_running_avg_y);
+          const uint8x16_t v_diff_pos_mask =
+              vcltq_u8(v_sig, v_mc_running_avg_y);
+          const uint8x16_t v_diff_neg_mask =
+              vcgtq_u8(v_sig, v_mc_running_avg_y);
+          // Clamp absolute difference to delta to get the adjustment.
+          const uint8x16_t v_abs_adjustment = vminq_u8(v_abs_diff, (k_delta));
 
-              v_running_avg_y = vqsubq_u8(v_running_avg_y, v_pos_adjustment);
-              v_running_avg_y = vqaddq_u8(v_running_avg_y, v_neg_adjustment);
+          const uint8x16_t v_pos_adjustment =
+              vandq_u8(v_diff_pos_mask, v_abs_adjustment);
+          const uint8x16_t v_neg_adjustment =
+              vandq_u8(v_diff_neg_mask, v_abs_adjustment);
 
-              /* Store results. */
-              vst1q_u8(running_avg_y, v_running_avg_y);
+          v_running_avg_y = vqsubq_u8(v_running_avg_y, v_pos_adjustment);
+          v_running_avg_y = vqaddq_u8(v_running_avg_y, v_neg_adjustment);
 
-              {
-                  const int8x16_t v_sum_diff =
-                      vqsubq_s8(vreinterpretq_s8_u8(v_neg_adjustment),
-                                vreinterpretq_s8_u8(v_pos_adjustment));
+          /* Store results. */
+          vst1q_u8(running_avg_y, v_running_avg_y);
 
-                  const int16x8_t fe_dc_ba_98_76_54_32_10 =
-                      vpaddlq_s8(v_sum_diff);
-                  const int32x4_t fedc_ba98_7654_3210 =
-                      vpaddlq_s16(fe_dc_ba_98_76_54_32_10);
-                  const int64x2_t fedcba98_76543210 =
-                      vpaddlq_s32(fedc_ba98_7654_3210);
+          {
+            const int8x16_t v_sum_diff =
+                vqsubq_s8(vreinterpretq_s8_u8(v_neg_adjustment),
+                          vreinterpretq_s8_u8(v_pos_adjustment));
 
-                  v_sum_diff_total = vqaddq_s64(v_sum_diff_total,
-                                                fedcba98_76543210);
-              }
-              /* Update pointers for next iteration. */
-              sig += sig_stride;
-              mc_running_avg_y += mc_running_avg_y_stride;
-              running_avg_y += running_avg_y_stride;
-            }
-            {
-              // Update the sum of all pixel differences of this MB.
-              x = vqadd_s64(vget_high_s64(v_sum_diff_total),
-                            vget_low_s64(v_sum_diff_total));
-              sum_diff = vget_lane_s32(vabs_s32(vreinterpret_s32_s64(x)), 0);
+            const int16x8_t fe_dc_ba_98_76_54_32_10 = vpaddlq_s8(v_sum_diff);
+            const int32x4_t fedc_ba98_7654_3210 =
+                vpaddlq_s16(fe_dc_ba_98_76_54_32_10);
+            const int64x2_t fedcba98_76543210 =
+                vpaddlq_s32(fedc_ba98_7654_3210);
 
-              if (sum_diff > sum_diff_thresh) {
-                return COPY_BLOCK;
-              }
-            }
-          } else {
+            v_sum_diff_total = vqaddq_s64(v_sum_diff_total, fedcba98_76543210);
+          }
+          /* Update pointers for next iteration. */
+          sig += sig_stride;
+          mc_running_avg_y += mc_running_avg_y_stride;
+          running_avg_y += running_avg_y_stride;
+        }
+        {
+          // Update the sum of all pixel differences of this MB.
+          x = vqadd_s64(vget_high_s64(v_sum_diff_total),
+                        vget_low_s64(v_sum_diff_total));
+          sum_diff = vget_lane_s32(vabs_s32(vreinterpret_s32_s64(x)), 0);
+
+          if (sum_diff > sum_diff_thresh) {
             return COPY_BLOCK;
           }
         }
+      } else {
+        return COPY_BLOCK;
+      }
     }
+  }
 
-    /* Tell above level that block was filtered. */
-    running_avg_y -= running_avg_y_stride * 16;
-    sig -= sig_stride * 16;
+  /* Tell above level that block was filtered. */
+  running_avg_y -= running_avg_y_stride * 16;
+  sig -= sig_stride * 16;
 
-    vp8_copy_mem16x16(running_avg_y, running_avg_y_stride, sig, sig_stride);
+  vp8_copy_mem16x16(running_avg_y, running_avg_y_stride, sig, sig_stride);
 
-    return FILTER_BLOCK;
+  return FILTER_BLOCK;
 }
 
 int vp8_denoiser_filter_uv_neon(unsigned char *mc_running_avg,
-                             int mc_running_avg_stride,
-                             unsigned char *running_avg,
-                             int running_avg_stride,
-                             unsigned char *sig, int sig_stride,
-                             unsigned int motion_magnitude,
-                             int increase_denoising) {
-    /* If motion_magnitude is small, making the denoiser more aggressive by
-     * increasing the adjustment for each level, level1 adjustment is
-     * increased, the deltas stay the same.
-     */
-    int shift_inc  = (increase_denoising &&
-        motion_magnitude <= MOTION_MAGNITUDE_THRESHOLD_UV) ? 1 : 0;
-    const uint8x16_t v_level1_adjustment = vmovq_n_u8(
-        (motion_magnitude <= MOTION_MAGNITUDE_THRESHOLD_UV) ? 4 + shift_inc : 3);
+                                int mc_running_avg_stride,
+                                unsigned char *running_avg,
+                                int running_avg_stride, unsigned char *sig,
+                                int sig_stride, unsigned int motion_magnitude,
+                                int increase_denoising) {
+  /* If motion_magnitude is small, making the denoiser more aggressive by
+   * increasing the adjustment for each level, level1 adjustment is
+   * increased, the deltas stay the same.
+   */
+  int shift_inc =
+      (increase_denoising && motion_magnitude <= MOTION_MAGNITUDE_THRESHOLD_UV)
+          ? 1
+          : 0;
+  const uint8x16_t v_level1_adjustment = vmovq_n_u8(
+      (motion_magnitude <= MOTION_MAGNITUDE_THRESHOLD_UV) ? 4 + shift_inc : 3);
 
-    const uint8x16_t v_delta_level_1_and_2 = vdupq_n_u8(1);
-    const uint8x16_t v_delta_level_2_and_3 = vdupq_n_u8(2);
-    const uint8x16_t v_level1_threshold = vmovq_n_u8(4 + shift_inc);
-    const uint8x16_t v_level2_threshold = vdupq_n_u8(8);
-    const uint8x16_t v_level3_threshold = vdupq_n_u8(16);
-    int64x2_t v_sum_diff_total = vdupq_n_s64(0);
-    int r;
+  const uint8x16_t v_delta_level_1_and_2 = vdupq_n_u8(1);
+  const uint8x16_t v_delta_level_2_and_3 = vdupq_n_u8(2);
+  const uint8x16_t v_level1_threshold = vmovq_n_u8(4 + shift_inc);
+  const uint8x16_t v_level2_threshold = vdupq_n_u8(8);
+  const uint8x16_t v_level3_threshold = vdupq_n_u8(16);
+  int64x2_t v_sum_diff_total = vdupq_n_s64(0);
+  int r;
 
-    {
-      uint16x4_t v_sum_block = vdup_n_u16(0);
+  {
+    uint16x4_t v_sum_block = vdup_n_u16(0);
 
-      // Avoid denoising color signal if its close to average level.
-      for (r = 0; r < 8; ++r) {
-        const uint8x8_t v_sig = vld1_u8(sig);
-        const uint16x4_t _76_54_32_10 = vpaddl_u8(v_sig);
-        v_sum_block = vqadd_u16(v_sum_block, _76_54_32_10);
-        sig += sig_stride;
+    // Avoid denoising color signal if its close to average level.
+    for (r = 0; r < 8; ++r) {
+      const uint8x8_t v_sig = vld1_u8(sig);
+      const uint16x4_t _76_54_32_10 = vpaddl_u8(v_sig);
+      v_sum_block = vqadd_u16(v_sum_block, _76_54_32_10);
+      sig += sig_stride;
+    }
+    sig -= sig_stride * 8;
+    {
+      const uint32x2_t _7654_3210 = vpaddl_u16(v_sum_block);
+      const uint64x1_t _76543210 = vpaddl_u32(_7654_3210);
+      const int sum_block = vget_lane_s32(vreinterpret_s32_u64(_76543210), 0);
+      if (abs(sum_block - (128 * 8 * 8)) < SUM_DIFF_FROM_AVG_THRESH_UV) {
+        return COPY_BLOCK;
       }
-      sig -= sig_stride * 8;
-      {
-        const uint32x2_t _7654_3210 = vpaddl_u16(v_sum_block);
-        const uint64x1_t _76543210 = vpaddl_u32(_7654_3210);
-        const int sum_block =
-            vget_lane_s32(vreinterpret_s32_u64(_76543210), 0);
-        if (abs(sum_block - (128 * 8 * 8)) < SUM_DIFF_FROM_AVG_THRESH_UV) {
-          return COPY_BLOCK;
-        }
-      }
     }
+  }
 
-    /* Go over lines. */
-    for (r = 0; r < 4; ++r) {
-        /* Load inputs. */
-        const uint8x8_t v_sig_lo = vld1_u8(sig);
-        const uint8x8_t v_sig_hi = vld1_u8(&sig[sig_stride]);
-        const uint8x16_t v_sig = vcombine_u8(v_sig_lo, v_sig_hi);
-        const uint8x8_t v_mc_running_avg_lo = vld1_u8(mc_running_avg);
-        const uint8x8_t v_mc_running_avg_hi =
-            vld1_u8(&mc_running_avg[mc_running_avg_stride]);
-        const uint8x16_t v_mc_running_avg =
-            vcombine_u8(v_mc_running_avg_lo, v_mc_running_avg_hi);
-        /* Calculate absolute difference and sign masks. */
-        const uint8x16_t v_abs_diff      = vabdq_u8(v_sig, v_mc_running_avg);
-        const uint8x16_t v_diff_pos_mask = vcltq_u8(v_sig, v_mc_running_avg);
-        const uint8x16_t v_diff_neg_mask = vcgtq_u8(v_sig, v_mc_running_avg);
+  /* Go over lines. */
+  for (r = 0; r < 4; ++r) {
+    /* Load inputs. */
+    const uint8x8_t v_sig_lo = vld1_u8(sig);
+    const uint8x8_t v_sig_hi = vld1_u8(&sig[sig_stride]);
+    const uint8x16_t v_sig = vcombine_u8(v_sig_lo, v_sig_hi);
+    const uint8x8_t v_mc_running_avg_lo = vld1_u8(mc_running_avg);
+    const uint8x8_t v_mc_running_avg_hi =
+        vld1_u8(&mc_running_avg[mc_running_avg_stride]);
+    const uint8x16_t v_mc_running_avg =
+        vcombine_u8(v_mc_running_avg_lo, v_mc_running_avg_hi);
+    /* Calculate absolute difference and sign masks. */
+    const uint8x16_t v_abs_diff = vabdq_u8(v_sig, v_mc_running_avg);
+    const uint8x16_t v_diff_pos_mask = vcltq_u8(v_sig, v_mc_running_avg);
+    const uint8x16_t v_diff_neg_mask = vcgtq_u8(v_sig, v_mc_running_avg);
 
-        /* Figure out which level that put us in. */
-        const uint8x16_t v_level1_mask = vcleq_u8(v_level1_threshold,
-                                                  v_abs_diff);
-        const uint8x16_t v_level2_mask = vcleq_u8(v_level2_threshold,
-                                                  v_abs_diff);
-        const uint8x16_t v_level3_mask = vcleq_u8(v_level3_threshold,
-                                                  v_abs_diff);
+    /* Figure out which level that put us in. */
+    const uint8x16_t v_level1_mask = vcleq_u8(v_level1_threshold, v_abs_diff);
+    const uint8x16_t v_level2_mask = vcleq_u8(v_level2_threshold, v_abs_diff);
+    const uint8x16_t v_level3_mask = vcleq_u8(v_level3_threshold, v_abs_diff);
 
-        /* Calculate absolute adjustments for level 1, 2 and 3. */
-        const uint8x16_t v_level2_adjustment = vandq_u8(v_level2_mask,
-                                                        v_delta_level_1_and_2);
-        const uint8x16_t v_level3_adjustment = vandq_u8(v_level3_mask,
-                                                        v_delta_level_2_and_3);
-        const uint8x16_t v_level1and2_adjustment = vaddq_u8(v_level1_adjustment,
-            v_level2_adjustment);
-        const uint8x16_t v_level1and2and3_adjustment = vaddq_u8(
-            v_level1and2_adjustment, v_level3_adjustment);
+    /* Calculate absolute adjustments for level 1, 2 and 3. */
+    const uint8x16_t v_level2_adjustment =
+        vandq_u8(v_level2_mask, v_delta_level_1_and_2);
+    const uint8x16_t v_level3_adjustment =
+        vandq_u8(v_level3_mask, v_delta_level_2_and_3);
+    const uint8x16_t v_level1and2_adjustment =
+        vaddq_u8(v_level1_adjustment, v_level2_adjustment);
+    const uint8x16_t v_level1and2and3_adjustment =
+        vaddq_u8(v_level1and2_adjustment, v_level3_adjustment);
 
-        /* Figure adjustment absolute value by selecting between the absolute
-         * difference if in level0 or the value for level 1, 2 and 3.
-         */
-        const uint8x16_t v_abs_adjustment = vbslq_u8(v_level1_mask,
-            v_level1and2and3_adjustment, v_abs_diff);
+    /* Figure adjustment absolute value by selecting between the absolute
+     * difference if in level0 or the value for level 1, 2 and 3.
+     */
+    const uint8x16_t v_abs_adjustment =
+        vbslq_u8(v_level1_mask, v_level1and2and3_adjustment, v_abs_diff);
 
-        /* Calculate positive and negative adjustments. Apply them to the signal
-         * and accumulate them. Adjustments are less than eight and the maximum
-         * sum of them (7 * 16) can fit in a signed char.
-         */
-        const uint8x16_t v_pos_adjustment = vandq_u8(v_diff_pos_mask,
-                                                     v_abs_adjustment);
-        const uint8x16_t v_neg_adjustment = vandq_u8(v_diff_neg_mask,
-                                                     v_abs_adjustment);
+    /* Calculate positive and negative adjustments. Apply them to the signal
+     * and accumulate them. Adjustments are less than eight and the maximum
+     * sum of them (7 * 16) can fit in a signed char.
+     */
+    const uint8x16_t v_pos_adjustment =
+        vandq_u8(v_diff_pos_mask, v_abs_adjustment);
+    const uint8x16_t v_neg_adjustment =
+        vandq_u8(v_diff_neg_mask, v_abs_adjustment);
 
-        uint8x16_t v_running_avg = vqaddq_u8(v_sig, v_pos_adjustment);
-        v_running_avg = vqsubq_u8(v_running_avg, v_neg_adjustment);
+    uint8x16_t v_running_avg = vqaddq_u8(v_sig, v_pos_adjustment);
+    v_running_avg = vqsubq_u8(v_running_avg, v_neg_adjustment);
 
-        /* Store results. */
-        vst1_u8(running_avg, vget_low_u8(v_running_avg));
-        vst1_u8(&running_avg[running_avg_stride], vget_high_u8(v_running_avg));
+    /* Store results. */
+    vst1_u8(running_avg, vget_low_u8(v_running_avg));
+    vst1_u8(&running_avg[running_avg_stride], vget_high_u8(v_running_avg));
 
-        /* Sum all the accumulators to have the sum of all pixel differences
-         * for this macroblock.
-         */
-        {
-            const int8x16_t v_sum_diff =
-                vqsubq_s8(vreinterpretq_s8_u8(v_pos_adjustment),
-                          vreinterpretq_s8_u8(v_neg_adjustment));
+    /* Sum all the accumulators to have the sum of all pixel differences
+     * for this macroblock.
+     */
+    {
+      const int8x16_t v_sum_diff =
+          vqsubq_s8(vreinterpretq_s8_u8(v_pos_adjustment),
+                    vreinterpretq_s8_u8(v_neg_adjustment));
 
-            const int16x8_t fe_dc_ba_98_76_54_32_10 = vpaddlq_s8(v_sum_diff);
+      const int16x8_t fe_dc_ba_98_76_54_32_10 = vpaddlq_s8(v_sum_diff);
 
-            const int32x4_t fedc_ba98_7654_3210 =
-                vpaddlq_s16(fe_dc_ba_98_76_54_32_10);
+      const int32x4_t fedc_ba98_7654_3210 =
+          vpaddlq_s16(fe_dc_ba_98_76_54_32_10);
 
-            const int64x2_t fedcba98_76543210 =
-                vpaddlq_s32(fedc_ba98_7654_3210);
+      const int64x2_t fedcba98_76543210 = vpaddlq_s32(fedc_ba98_7654_3210);
 
-            v_sum_diff_total = vqaddq_s64(v_sum_diff_total, fedcba98_76543210);
-        }
-
-        /* Update pointers for next iteration. */
-        sig += sig_stride * 2;
-        mc_running_avg += mc_running_avg_stride * 2;
-        running_avg += running_avg_stride * 2;
+      v_sum_diff_total = vqaddq_s64(v_sum_diff_total, fedcba98_76543210);
     }
 
+    /* Update pointers for next iteration. */
+    sig += sig_stride * 2;
+    mc_running_avg += mc_running_avg_stride * 2;
+    running_avg += running_avg_stride * 2;
+  }
 
-    /* Too much adjustments => copy block. */
-    {
-        int64x1_t x = vqadd_s64(vget_high_s64(v_sum_diff_total),
-                                      vget_low_s64(v_sum_diff_total));
-        int sum_diff = vget_lane_s32(vabs_s32(vreinterpret_s32_s64(x)), 0);
-        int sum_diff_thresh = SUM_DIFF_THRESHOLD_UV;
-        if (increase_denoising) sum_diff_thresh = SUM_DIFF_THRESHOLD_HIGH_UV;
-        if (sum_diff > sum_diff_thresh) {
-          // Before returning to copy the block (i.e., apply no denoising),
-          // checK if we can still apply some (weaker) temporal filtering to
-          // this block, that would otherwise not be denoised at all. Simplest
-          // is to apply an additional adjustment to running_avg_y to bring it
-          // closer to sig. The adjustment is capped by a maximum delta, and
-          // chosen such that in most cases the resulting sum_diff will be
-          // within the accceptable range given by sum_diff_thresh.
+  /* Too much adjustments => copy block. */
+  {
+    int64x1_t x = vqadd_s64(vget_high_s64(v_sum_diff_total),
+                            vget_low_s64(v_sum_diff_total));
+    int sum_diff = vget_lane_s32(vabs_s32(vreinterpret_s32_s64(x)), 0);
+    int sum_diff_thresh = SUM_DIFF_THRESHOLD_UV;
+    if (increase_denoising) sum_diff_thresh = SUM_DIFF_THRESHOLD_HIGH_UV;
+    if (sum_diff > sum_diff_thresh) {
+      // Before returning to copy the block (i.e., apply no denoising),
+      // checK if we can still apply some (weaker) temporal filtering to
+      // this block, that would otherwise not be denoised at all. Simplest
+      // is to apply an additional adjustment to running_avg_y to bring it
+      // closer to sig. The adjustment is capped by a maximum delta, and
+      // chosen such that in most cases the resulting sum_diff will be
+      // within the accceptable range given by sum_diff_thresh.
 
-          // The delta is set by the excess of absolute pixel diff over the
-          // threshold.
-          int delta = ((sum_diff - sum_diff_thresh) >> 8) + 1;
-          // Only apply the adjustment for max delta up to 3.
-          if (delta < 4) {
-            const uint8x16_t k_delta = vmovq_n_u8(delta);
-            sig -= sig_stride * 8;
-            mc_running_avg -= mc_running_avg_stride * 8;
-            running_avg -= running_avg_stride * 8;
-            for (r = 0; r < 4; ++r) {
-              const uint8x8_t v_sig_lo = vld1_u8(sig);
-              const uint8x8_t v_sig_hi = vld1_u8(&sig[sig_stride]);
-              const uint8x16_t v_sig = vcombine_u8(v_sig_lo, v_sig_hi);
-              const uint8x8_t v_mc_running_avg_lo = vld1_u8(mc_running_avg);
-              const uint8x8_t v_mc_running_avg_hi =
-                  vld1_u8(&mc_running_avg[mc_running_avg_stride]);
-              const uint8x16_t v_mc_running_avg =
-                  vcombine_u8(v_mc_running_avg_lo, v_mc_running_avg_hi);
-              /* Calculate absolute difference and sign masks. */
-              const uint8x16_t v_abs_diff      = vabdq_u8(v_sig,
-                                                          v_mc_running_avg);
-              const uint8x16_t v_diff_pos_mask = vcltq_u8(v_sig,
-                                                          v_mc_running_avg);
-              const uint8x16_t v_diff_neg_mask = vcgtq_u8(v_sig,
-                                                          v_mc_running_avg);
-              // Clamp absolute difference to delta to get the adjustment.
-              const uint8x16_t v_abs_adjustment =
-                  vminq_u8(v_abs_diff, (k_delta));
+      // The delta is set by the excess of absolute pixel diff over the
+      // threshold.
+      int delta = ((sum_diff - sum_diff_thresh) >> 8) + 1;
+      // Only apply the adjustment for max delta up to 3.
+      if (delta < 4) {
+        const uint8x16_t k_delta = vmovq_n_u8(delta);
+        sig -= sig_stride * 8;
+        mc_running_avg -= mc_running_avg_stride * 8;
+        running_avg -= running_avg_stride * 8;
+        for (r = 0; r < 4; ++r) {
+          const uint8x8_t v_sig_lo = vld1_u8(sig);
+          const uint8x8_t v_sig_hi = vld1_u8(&sig[sig_stride]);
+          const uint8x16_t v_sig = vcombine_u8(v_sig_lo, v_sig_hi);
+          const uint8x8_t v_mc_running_avg_lo = vld1_u8(mc_running_avg);
+          const uint8x8_t v_mc_running_avg_hi =
+              vld1_u8(&mc_running_avg[mc_running_avg_stride]);
+          const uint8x16_t v_mc_running_avg =
+              vcombine_u8(v_mc_running_avg_lo, v_mc_running_avg_hi);
+          /* Calculate absolute difference and sign masks. */
+          const uint8x16_t v_abs_diff = vabdq_u8(v_sig, v_mc_running_avg);
+          const uint8x16_t v_diff_pos_mask = vcltq_u8(v_sig, v_mc_running_avg);
+          const uint8x16_t v_diff_neg_mask = vcgtq_u8(v_sig, v_mc_running_avg);
+          // Clamp absolute difference to delta to get the adjustment.
+          const uint8x16_t v_abs_adjustment = vminq_u8(v_abs_diff, (k_delta));
 
-              const uint8x16_t v_pos_adjustment = vandq_u8(v_diff_pos_mask,
-                                                           v_abs_adjustment);
-              const uint8x16_t v_neg_adjustment = vandq_u8(v_diff_neg_mask,
-                                                           v_abs_adjustment);
-              const uint8x8_t v_running_avg_lo = vld1_u8(running_avg);
-              const uint8x8_t v_running_avg_hi =
-                  vld1_u8(&running_avg[running_avg_stride]);
-              uint8x16_t v_running_avg =
-                  vcombine_u8(v_running_avg_lo, v_running_avg_hi);
+          const uint8x16_t v_pos_adjustment =
+              vandq_u8(v_diff_pos_mask, v_abs_adjustment);
+          const uint8x16_t v_neg_adjustment =
+              vandq_u8(v_diff_neg_mask, v_abs_adjustment);
+          const uint8x8_t v_running_avg_lo = vld1_u8(running_avg);
+          const uint8x8_t v_running_avg_hi =
+              vld1_u8(&running_avg[running_avg_stride]);
+          uint8x16_t v_running_avg =
+              vcombine_u8(v_running_avg_lo, v_running_avg_hi);
 
-              v_running_avg = vqsubq_u8(v_running_avg, v_pos_adjustment);
-              v_running_avg = vqaddq_u8(v_running_avg, v_neg_adjustment);
+          v_running_avg = vqsubq_u8(v_running_avg, v_pos_adjustment);
+          v_running_avg = vqaddq_u8(v_running_avg, v_neg_adjustment);
 
-              /* Store results. */
-              vst1_u8(running_avg, vget_low_u8(v_running_avg));
-              vst1_u8(&running_avg[running_avg_stride],
-                      vget_high_u8(v_running_avg));
+          /* Store results. */
+          vst1_u8(running_avg, vget_low_u8(v_running_avg));
+          vst1_u8(&running_avg[running_avg_stride],
+                  vget_high_u8(v_running_avg));
 
-              {
-                  const int8x16_t v_sum_diff =
-                      vqsubq_s8(vreinterpretq_s8_u8(v_neg_adjustment),
-                                vreinterpretq_s8_u8(v_pos_adjustment));
+          {
+            const int8x16_t v_sum_diff =
+                vqsubq_s8(vreinterpretq_s8_u8(v_neg_adjustment),
+                          vreinterpretq_s8_u8(v_pos_adjustment));
 
-                  const int16x8_t fe_dc_ba_98_76_54_32_10 =
-                      vpaddlq_s8(v_sum_diff);
-                  const int32x4_t fedc_ba98_7654_3210 =
-                      vpaddlq_s16(fe_dc_ba_98_76_54_32_10);
-                  const int64x2_t fedcba98_76543210 =
-                      vpaddlq_s32(fedc_ba98_7654_3210);
+            const int16x8_t fe_dc_ba_98_76_54_32_10 = vpaddlq_s8(v_sum_diff);
+            const int32x4_t fedc_ba98_7654_3210 =
+                vpaddlq_s16(fe_dc_ba_98_76_54_32_10);
+            const int64x2_t fedcba98_76543210 =
+                vpaddlq_s32(fedc_ba98_7654_3210);
 
-                  v_sum_diff_total = vqaddq_s64(v_sum_diff_total,
-                                                fedcba98_76543210);
-              }
-              /* Update pointers for next iteration. */
-              sig += sig_stride * 2;
-              mc_running_avg += mc_running_avg_stride * 2;
-              running_avg += running_avg_stride * 2;
-            }
-            {
-              // Update the sum of all pixel differences of this MB.
-              x = vqadd_s64(vget_high_s64(v_sum_diff_total),
-                            vget_low_s64(v_sum_diff_total));
-              sum_diff = vget_lane_s32(vabs_s32(vreinterpret_s32_s64(x)), 0);
+            v_sum_diff_total = vqaddq_s64(v_sum_diff_total, fedcba98_76543210);
+          }
+          /* Update pointers for next iteration. */
+          sig += sig_stride * 2;
+          mc_running_avg += mc_running_avg_stride * 2;
+          running_avg += running_avg_stride * 2;
+        }
+        {
+          // Update the sum of all pixel differences of this MB.
+          x = vqadd_s64(vget_high_s64(v_sum_diff_total),
+                        vget_low_s64(v_sum_diff_total));
+          sum_diff = vget_lane_s32(vabs_s32(vreinterpret_s32_s64(x)), 0);
 
-              if (sum_diff > sum_diff_thresh) {
-                return COPY_BLOCK;
-              }
-            }
-          } else {
+          if (sum_diff > sum_diff_thresh) {
             return COPY_BLOCK;
           }
         }
+      } else {
+        return COPY_BLOCK;
+      }
     }
+  }
 
-    /* Tell above level that block was filtered. */
-    running_avg -= running_avg_stride * 8;
-    sig -= sig_stride * 8;
+  /* Tell above level that block was filtered. */
+  running_avg -= running_avg_stride * 8;
+  sig -= sig_stride * 8;
 
-    vp8_copy_mem8x8(running_avg, running_avg_stride, sig, sig_stride);
+  vp8_copy_mem8x8(running_avg, running_avg_stride, sig, sig_stride);
 
-    return FILTER_BLOCK;
+  return FILTER_BLOCK;
 }
--- a/vp8/encoder/arm/neon/fastquantizeb_neon.c
+++ b/vp8/encoder/arm/neon/fastquantizeb_neon.c
@@ -11,79 +11,73 @@
 #include <arm_neon.h>
 #include "vp8/encoder/block.h"
 
-static const uint16_t inv_zig_zag[16] = {
-    1,  2,  6,   7,
-    3,  5,  8,  13,
-    4,  9,  12, 14,
-    10, 11, 15, 16
-};
+static const uint16_t inv_zig_zag[16] = { 1, 2, 6,  7,  3,  5,  8,  13,
+                                          4, 9, 12, 14, 10, 11, 15, 16 };
 
 void vp8_fast_quantize_b_neon(BLOCK *b, BLOCKD *d) {
-    const int16x8_t one_q = vdupq_n_s16(-1),
-                    z0 = vld1q_s16(b->coeff),
-                    z1 = vld1q_s16(b->coeff + 8),
-                    round0 = vld1q_s16(b->round),
-                    round1 = vld1q_s16(b->round + 8),
-                    quant0 = vld1q_s16(b->quant_fast),
-                    quant1 = vld1q_s16(b->quant_fast + 8),
-                    dequant0 = vld1q_s16(d->dequant),
-                    dequant1 = vld1q_s16(d->dequant + 8);
-    const uint16x8_t zig_zag0 = vld1q_u16(inv_zig_zag),
-                     zig_zag1 = vld1q_u16(inv_zig_zag + 8);
-    int16x8_t x0, x1, sz0, sz1, y0, y1;
-    uint16x8_t eob0, eob1;
-    uint16x4_t eob_d16;
-    uint32x2_t eob_d32;
-    uint32x4_t eob_q32;
+  const int16x8_t one_q = vdupq_n_s16(-1), z0 = vld1q_s16(b->coeff),
+                  z1 = vld1q_s16(b->coeff + 8), round0 = vld1q_s16(b->round),
+                  round1 = vld1q_s16(b->round + 8),
+                  quant0 = vld1q_s16(b->quant_fast),
+                  quant1 = vld1q_s16(b->quant_fast + 8),
+                  dequant0 = vld1q_s16(d->dequant),
+                  dequant1 = vld1q_s16(d->dequant + 8);
+  const uint16x8_t zig_zag0 = vld1q_u16(inv_zig_zag),
+                   zig_zag1 = vld1q_u16(inv_zig_zag + 8);
+  int16x8_t x0, x1, sz0, sz1, y0, y1;
+  uint16x8_t eob0, eob1;
+  uint16x4_t eob_d16;
+  uint32x2_t eob_d32;
+  uint32x4_t eob_q32;
 
-    /* sign of z: z >> 15 */
-    sz0 = vshrq_n_s16(z0, 15);
-    sz1 = vshrq_n_s16(z1, 15);
+  /* sign of z: z >> 15 */
+  sz0 = vshrq_n_s16(z0, 15);
+  sz1 = vshrq_n_s16(z1, 15);
 
-    /* x = abs(z) */
-    x0 = vabsq_s16(z0);
-    x1 = vabsq_s16(z1);
+  /* x = abs(z) */
+  x0 = vabsq_s16(z0);
+  x1 = vabsq_s16(z1);
 
-    /* x += round */
-    x0 = vaddq_s16(x0, round0);
-    x1 = vaddq_s16(x1, round1);
+  /* x += round */
+  x0 = vaddq_s16(x0, round0);
+  x1 = vaddq_s16(x1, round1);
 
-    /* y = 2 * (x * quant) >> 16 */
-    y0 = vqdmulhq_s16(x0, quant0);
-    y1 = vqdmulhq_s16(x1, quant1);
+  /* y = 2 * (x * quant) >> 16 */
+  y0 = vqdmulhq_s16(x0, quant0);
+  y1 = vqdmulhq_s16(x1, quant1);
 
-    /* Compensate for doubling in vqdmulhq */
-    y0 = vshrq_n_s16(y0, 1);
-    y1 = vshrq_n_s16(y1, 1);
+  /* Compensate for doubling in vqdmulhq */
+  y0 = vshrq_n_s16(y0, 1);
+  y1 = vshrq_n_s16(y1, 1);
 
-    /* Restore sign bit */
-    y0 = veorq_s16(y0, sz0);
-    y1 = veorq_s16(y1, sz1);
-    x0 = vsubq_s16(y0, sz0);
-    x1 = vsubq_s16(y1, sz1);
+  /* Restore sign bit */
+  y0 = veorq_s16(y0, sz0);
+  y1 = veorq_s16(y1, sz1);
+  x0 = vsubq_s16(y0, sz0);
+  x1 = vsubq_s16(y1, sz1);
 
-    /* find non-zero elements */
-    eob0 = vtstq_s16(x0, one_q);
-    eob1 = vtstq_s16(x1, one_q);
+  /* find non-zero elements */
+  eob0 = vtstq_s16(x0, one_q);
+  eob1 = vtstq_s16(x1, one_q);
 
-    /* mask zig zag */
-    eob0 = vandq_u16(eob0, zig_zag0);
-    eob1 = vandq_u16(eob1, zig_zag1);
+  /* mask zig zag */
+  eob0 = vandq_u16(eob0, zig_zag0);
+  eob1 = vandq_u16(eob1, zig_zag1);
 
-    /* select the largest value */
-    eob0 = vmaxq_u16(eob0, eob1);
-    eob_d16 = vmax_u16(vget_low_u16(eob0), vget_high_u16(eob0));
-    eob_q32 = vmovl_u16(eob_d16);
-    eob_d32 = vmax_u32(vget_low_u32(eob_q32), vget_high_u32(eob_q32));
-    eob_d32 = vpmax_u32(eob_d32, eob_d32);
+  /* select the largest value */
+  eob0 = vmaxq_u16(eob0, eob1);
+  eob_d16 = vmax_u16(vget_low_u16(eob0), vget_high_u16(eob0));
+  eob_q32 = vmovl_u16(eob_d16);
+  eob_d32 = vmax_u32(vget_low_u32(eob_q32), vget_high_u32(eob_q32));
+  eob_d32 = vpmax_u32(eob_d32, eob_d32);
 
-    /* qcoeff = x */
-    vst1q_s16(d->qcoeff, x0);
-    vst1q_s16(d->qcoeff + 8, x1);
+  /* qcoeff = x */
+  vst1q_s16(d->qcoeff, x0);
+  vst1q_s16(d->qcoeff + 8, x1);
 
-    /* dqcoeff = x * dequant */
-    vst1q_s16(d->dqcoeff, vmulq_s16(dequant0, x0));
-    vst1q_s16(d->dqcoeff + 8, vmulq_s16(dequant1, x1));
+  /* dqcoeff = x * dequant */
+  vst1q_s16(d->dqcoeff, vmulq_s16(dequant0, x0));
+  vst1q_s16(d->dqcoeff + 8, vmulq_s16(dequant1, x1));
 
-    vst1_lane_s8((int8_t *)d->eob, vreinterpret_s8_u32(eob_d32), 0);
+  vst1_lane_s8((int8_t *)d->eob, vreinterpret_s8_u32(eob_d32), 0);
 }
--- a/vp8/encoder/arm/neon/shortfdct_neon.c
+++ b/vp8/encoder/arm/neon/shortfdct_neon.c
@@ -10,260 +10,250 @@
 
 #include <arm_neon.h>
 
-void vp8_short_fdct4x4_neon(
-        int16_t *input,
-        int16_t *output,
-        int pitch) {
-    int16x4_t d0s16, d1s16, d2s16, d3s16, d4s16, d5s16, d6s16, d7s16;
-    int16x4_t d16s16, d17s16, d26s16, dEmptys16;
-    uint16x4_t d4u16;
-    int16x8_t q0s16, q1s16;
-    int32x4_t q9s32, q10s32, q11s32, q12s32;
-    int16x4x2_t v2tmp0, v2tmp1;
-    int32x2x2_t v2tmp2, v2tmp3;
+void vp8_short_fdct4x4_neon(int16_t *input, int16_t *output, int pitch) {
+  int16x4_t d0s16, d1s16, d2s16, d3s16, d4s16, d5s16, d6s16, d7s16;
+  int16x4_t d16s16, d17s16, d26s16, dEmptys16;
+  uint16x4_t d4u16;
+  int16x8_t q0s16, q1s16;
+  int32x4_t q9s32, q10s32, q11s32, q12s32;
+  int16x4x2_t v2tmp0, v2tmp1;
+  int32x2x2_t v2tmp2, v2tmp3;
 
-    d16s16 = vdup_n_s16(5352);
-    d17s16 = vdup_n_s16(2217);
-    q9s32 = vdupq_n_s32(14500);
-    q10s32 = vdupq_n_s32(7500);
-    q11s32 = vdupq_n_s32(12000);
-    q12s32 = vdupq_n_s32(51000);
+  d16s16 = vdup_n_s16(5352);
+  d17s16 = vdup_n_s16(2217);
+  q9s32 = vdupq_n_s32(14500);
+  q10s32 = vdupq_n_s32(7500);
+  q11s32 = vdupq_n_s32(12000);
+  q12s32 = vdupq_n_s32(51000);
 
-    // Part one
-    pitch >>= 1;
-    d0s16 = vld1_s16(input);
-    input += pitch;
-    d1s16 = vld1_s16(input);
-    input += pitch;
-    d2s16 = vld1_s16(input);
-    input += pitch;
-    d3s16 = vld1_s16(input);
+  // Part one
+  pitch >>= 1;
+  d0s16 = vld1_s16(input);
+  input += pitch;
+  d1s16 = vld1_s16(input);
+  input += pitch;
+  d2s16 = vld1_s16(input);
+  input += pitch;
+  d3s16 = vld1_s16(input);
 
-    v2tmp2 = vtrn_s32(vreinterpret_s32_s16(d0s16),
-                      vreinterpret_s32_s16(d2s16));
-    v2tmp3 = vtrn_s32(vreinterpret_s32_s16(d1s16),
-                      vreinterpret_s32_s16(d3s16));
-    v2tmp0 = vtrn_s16(vreinterpret_s16_s32(v2tmp2.val[0]),   // d0
-                      vreinterpret_s16_s32(v2tmp3.val[0]));  // d1
-    v2tmp1 = vtrn_s16(vreinterpret_s16_s32(v2tmp2.val[1]),   // d2
-                      vreinterpret_s16_s32(v2tmp3.val[1]));  // d3
+  v2tmp2 = vtrn_s32(vreinterpret_s32_s16(d0s16), vreinterpret_s32_s16(d2s16));
+  v2tmp3 = vtrn_s32(vreinterpret_s32_s16(d1s16), vreinterpret_s32_s16(d3s16));
+  v2tmp0 = vtrn_s16(vreinterpret_s16_s32(v2tmp2.val[0]),   // d0
+                    vreinterpret_s16_s32(v2tmp3.val[0]));  // d1
+  v2tmp1 = vtrn_s16(vreinterpret_s16_s32(v2tmp2.val[1]),   // d2
+                    vreinterpret_s16_s32(v2tmp3.val[1]));  // d3
 
-    d4s16 = vadd_s16(v2tmp0.val[0], v2tmp1.val[1]);
-    d5s16 = vadd_s16(v2tmp0.val[1], v2tmp1.val[0]);
-    d6s16 = vsub_s16(v2tmp0.val[1], v2tmp1.val[0]);
-    d7s16 = vsub_s16(v2tmp0.val[0], v2tmp1.val[1]);
+  d4s16 = vadd_s16(v2tmp0.val[0], v2tmp1.val[1]);
+  d5s16 = vadd_s16(v2tmp0.val[1], v2tmp1.val[0]);
+  d6s16 = vsub_s16(v2tmp0.val[1], v2tmp1.val[0]);
+  d7s16 = vsub_s16(v2tmp0.val[0], v2tmp1.val[1]);
 
-    d4s16 = vshl_n_s16(d4s16, 3);
-    d5s16 = vshl_n_s16(d5s16, 3);
-    d6s16 = vshl_n_s16(d6s16, 3);
-    d7s16 = vshl_n_s16(d7s16, 3);
+  d4s16 = vshl_n_s16(d4s16, 3);
+  d5s16 = vshl_n_s16(d5s16, 3);
+  d6s16 = vshl_n_s16(d6s16, 3);
+  d7s16 = vshl_n_s16(d7s16, 3);
 
-    d0s16 = vadd_s16(d4s16, d5s16);
-    d2s16 = vsub_s16(d4s16, d5s16);
+  d0s16 = vadd_s16(d4s16, d5s16);
+  d2s16 = vsub_s16(d4s16, d5s16);
 
-    q9s32 = vmlal_s16(q9s32, d7s16, d16s16);
-    q10s32 = vmlal_s16(q10s32, d7s16, d17s16);
-    q9s32 = vmlal_s16(q9s32, d6s16, d17s16);
-    q10s32 = vmlsl_s16(q10s32, d6s16, d16s16);
+  q9s32 = vmlal_s16(q9s32, d7s16, d16s16);
+  q10s32 = vmlal_s16(q10s32, d7s16, d17s16);
+  q9s32 = vmlal_s16(q9s32, d6s16, d17s16);
+  q10s32 = vmlsl_s16(q10s32, d6s16, d16s16);
 
-    d1s16 = vshrn_n_s32(q9s32, 12);
-    d3s16 = vshrn_n_s32(q10s32, 12);
+  d1s16 = vshrn_n_s32(q9s32, 12);
+  d3s16 = vshrn_n_s32(q10s32, 12);
 
-    // Part two
-    v2tmp2 = vtrn_s32(vreinterpret_s32_s16(d0s16),
-                      vreinterpret_s32_s16(d2s16));
-    v2tmp3 = vtrn_s32(vreinterpret_s32_s16(d1s16),
-                      vreinterpret_s32_s16(d3s16));
-    v2tmp0 = vtrn_s16(vreinterpret_s16_s32(v2tmp2.val[0]),   // d0
-                      vreinterpret_s16_s32(v2tmp3.val[0]));  // d1
-    v2tmp1 = vtrn_s16(vreinterpret_s16_s32(v2tmp2.val[1]),   // d2
-                      vreinterpret_s16_s32(v2tmp3.val[1]));  // d3
+  // Part two
+  v2tmp2 = vtrn_s32(vreinterpret_s32_s16(d0s16), vreinterpret_s32_s16(d2s16));
+  v2tmp3 = vtrn_s32(vreinterpret_s32_s16(d1s16), vreinterpret_s32_s16(d3s16));
+  v2tmp0 = vtrn_s16(vreinterpret_s16_s32(v2tmp2.val[0]),   // d0
+                    vreinterpret_s16_s32(v2tmp3.val[0]));  // d1
+  v2tmp1 = vtrn_s16(vreinterpret_s16_s32(v2tmp2.val[1]),   // d2
+                    vreinterpret_s16_s32(v2tmp3.val[1]));  // d3
 
-    d4s16 = vadd_s16(v2tmp0.val[0], v2tmp1.val[1]);
-    d5s16 = vadd_s16(v2tmp0.val[1], v2tmp1.val[0]);
-    d6s16 = vsub_s16(v2tmp0.val[1], v2tmp1.val[0]);
-    d7s16 = vsub_s16(v2tmp0.val[0], v2tmp1.val[1]);
+  d4s16 = vadd_s16(v2tmp0.val[0], v2tmp1.val[1]);
+  d5s16 = vadd_s16(v2tmp0.val[1], v2tmp1.val[0]);
+  d6s16 = vsub_s16(v2tmp0.val[1], v2tmp1.val[0]);
+  d7s16 = vsub_s16(v2tmp0.val[0], v2tmp1.val[1]);
 
-    d26s16 = vdup_n_s16(7);
-    d4s16 = vadd_s16(d4s16, d26s16);
+  d26s16 = vdup_n_s16(7);
+  d4s16 = vadd_s16(d4s16, d26s16);
 
-    d0s16 = vadd_s16(d4s16, d5s16);
-    d2s16 = vsub_s16(d4s16, d5s16);
+  d0s16 = vadd_s16(d4s16, d5s16);
+  d2s16 = vsub_s16(d4s16, d5s16);
 
-    q11s32 = vmlal_s16(q11s32, d7s16, d16s16);
-    q12s32 = vmlal_s16(q12s32, d7s16, d17s16);
+  q11s32 = vmlal_s16(q11s32, d7s16, d16s16);
+  q12s32 = vmlal_s16(q12s32, d7s16, d17s16);
 
-    dEmptys16 = vdup_n_s16(0);
-    d4u16 = vceq_s16(d7s16, dEmptys16);
+  dEmptys16 = vdup_n_s16(0);
+  d4u16 = vceq_s16(d7s16, dEmptys16);
 
-    d0s16 = vshr_n_s16(d0s16, 4);
-    d2s16 = vshr_n_s16(d2s16, 4);
+  d0s16 = vshr_n_s16(d0s16, 4);
+  d2s16 = vshr_n_s16(d2s16, 4);
 
-    q11s32 = vmlal_s16(q11s32, d6s16, d17s16);
-    q12s32 = vmlsl_s16(q12s32, d6s16, d16s16);
+  q11s32 = vmlal_s16(q11s32, d6s16, d17s16);
+  q12s32 = vmlsl_s16(q12s32, d6s16, d16s16);
 
-    d4u16 = vmvn_u16(d4u16);
-    d1s16 = vshrn_n_s32(q11s32, 16);
-    d1s16 = vsub_s16(d1s16, vreinterpret_s16_u16(d4u16));
-    d3s16 = vshrn_n_s32(q12s32, 16);
+  d4u16 = vmvn_u16(d4u16);
+  d1s16 = vshrn_n_s32(q11s32, 16);
+  d1s16 = vsub_s16(d1s16, vreinterpret_s16_u16(d4u16));
+  d3s16 = vshrn_n_s32(q12s32, 16);
 
-    q0s16 = vcombine_s16(d0s16, d1s16);
-    q1s16 = vcombine_s16(d2s16, d3s16);
+  q0s16 = vcombine_s16(d0s16, d1s16);
+  q1s16 = vcombine_s16(d2s16, d3s16);
 
-    vst1q_s16(output, q0s16);
-    vst1q_s16(output + 8, q1s16);
-    return;
+  vst1q_s16(output, q0s16);
+  vst1q_s16(output + 8, q1s16);
+  return;
 }
 
-void vp8_short_fdct8x4_neon(
-        int16_t *input,
-        int16_t *output,
-        int pitch) {
-    int16x4_t d0s16, d1s16, d2s16, d3s16, d4s16, d5s16, d6s16, d7s16;
-    int16x4_t d16s16, d17s16, d26s16, d27s16, d28s16, d29s16;
-    uint16x4_t d28u16, d29u16;
-    uint16x8_t q14u16;
-    int16x8_t q0s16, q1s16, q2s16, q3s16;
-    int16x8_t q11s16, q12s16, q13s16, q14s16, q15s16, qEmptys16;
-    int32x4_t q9s32, q10s32, q11s32, q12s32;
-    int16x8x2_t v2tmp0, v2tmp1;
-    int32x4x2_t v2tmp2, v2tmp3;
+void vp8_short_fdct8x4_neon(int16_t *input, int16_t *output, int pitch) {
+  int16x4_t d0s16, d1s16, d2s16, d3s16, d4s16, d5s16, d6s16, d7s16;
+  int16x4_t d16s16, d17s16, d26s16, d27s16, d28s16, d29s16;
+  uint16x4_t d28u16, d29u16;
+  uint16x8_t q14u16;
+  int16x8_t q0s16, q1s16, q2s16, q3s16;
+  int16x8_t q11s16, q12s16, q13s16, q14s16, q15s16, qEmptys16;
+  int32x4_t q9s32, q10s32, q11s32, q12s32;
+  int16x8x2_t v2tmp0, v2tmp1;
+  int32x4x2_t v2tmp2, v2tmp3;
 
-    d16s16 = vdup_n_s16(5352);
-    d17s16 = vdup_n_s16(2217);
-    q9s32 = vdupq_n_s32(14500);
-    q10s32 = vdupq_n_s32(7500);
+  d16s16 = vdup_n_s16(5352);
+  d17s16 = vdup_n_s16(2217);
+  q9s32 = vdupq_n_s32(14500);
+  q10s32 = vdupq_n_s32(7500);
 
-    // Part one
-    pitch >>= 1;
-    q0s16 = vld1q_s16(input);
-    input += pitch;
-    q1s16 = vld1q_s16(input);
-    input += pitch;
-    q2s16 = vld1q_s16(input);
-    input += pitch;
-    q3s16 = vld1q_s16(input);
+  // Part one
+  pitch >>= 1;
+  q0s16 = vld1q_s16(input);
+  input += pitch;
+  q1s16 = vld1q_s16(input);
+  input += pitch;
+  q2s16 = vld1q_s16(input);
+  input += pitch;
+  q3s16 = vld1q_s16(input);
 
-    v2tmp2 = vtrnq_s32(vreinterpretq_s32_s16(q0s16),
-                       vreinterpretq_s32_s16(q2s16));
-    v2tmp3 = vtrnq_s32(vreinterpretq_s32_s16(q1s16),
-                       vreinterpretq_s32_s16(q3s16));
-    v2tmp0 = vtrnq_s16(vreinterpretq_s16_s32(v2tmp2.val[0]),   // q0
-                       vreinterpretq_s16_s32(v2tmp3.val[0]));  // q1
-    v2tmp1 = vtrnq_s16(vreinterpretq_s16_s32(v2tmp2.val[1]),   // q2
-                       vreinterpretq_s16_s32(v2tmp3.val[1]));  // q3
+  v2tmp2 =
+      vtrnq_s32(vreinterpretq_s32_s16(q0s16), vreinterpretq_s32_s16(q2s16));
+  v2tmp3 =
+      vtrnq_s32(vreinterpretq_s32_s16(q1s16), vreinterpretq_s32_s16(q3s16));
+  v2tmp0 = vtrnq_s16(vreinterpretq_s16_s32(v2tmp2.val[0]),   // q0
+                     vreinterpretq_s16_s32(v2tmp3.val[0]));  // q1
+  v2tmp1 = vtrnq_s16(vreinterpretq_s16_s32(v2tmp2.val[1]),   // q2
+                     vreinterpretq_s16_s32(v2tmp3.val[1]));  // q3
 
-    q11s16 = vaddq_s16(v2tmp0.val[0], v2tmp1.val[1]);
-    q12s16 = vaddq_s16(v2tmp0.val[1], v2tmp1.val[0]);
-    q13s16 = vsubq_s16(v2tmp0.val[1], v2tmp1.val[0]);
-    q14s16 = vsubq_s16(v2tmp0.val[0], v2tmp1.val[1]);
+  q11s16 = vaddq_s16(v2tmp0.val[0], v2tmp1.val[1]);
+  q12s16 = vaddq_s16(v2tmp0.val[1], v2tmp1.val[0]);
+  q13s16 = vsubq_s16(v2tmp0.val[1], v2tmp1.val[0]);
+  q14s16 = vsubq_s16(v2tmp0.val[0], v2tmp1.val[1]);
 
-    q11s16 = vshlq_n_s16(q11s16, 3);
-    q12s16 = vshlq_n_s16(q12s16, 3);
-    q13s16 = vshlq_n_s16(q13s16, 3);
-    q14s16 = vshlq_n_s16(q14s16, 3);
+  q11s16 = vshlq_n_s16(q11s16, 3);
+  q12s16 = vshlq_n_s16(q12s16, 3);
+  q13s16 = vshlq_n_s16(q13s16, 3);
+  q14s16 = vshlq_n_s16(q14s16, 3);
 
-    q0s16 = vaddq_s16(q11s16, q12s16);
-    q2s16 = vsubq_s16(q11s16, q12s16);
+  q0s16 = vaddq_s16(q11s16, q12s16);
+  q2s16 = vsubq_s16(q11s16, q12s16);
 
-    q11s32 = q9s32;
-    q12s32 = q10s32;
+  q11s32 = q9s32;
+  q12s32 = q10s32;
 
-    d26s16 = vget_low_s16(q13s16);
-    d27s16 = vget_high_s16(q13s16);
-    d28s16 = vget_low_s16(q14s16);
-    d29s16 = vget_high_s16(q14s16);
+  d26s16 = vget_low_s16(q13s16);
+  d27s16 = vget_high_s16(q13s16);
+  d28s16 = vget_low_s16(q14s16);
+  d29s16 = vget_high_s16(q14s16);
 
-    q9s32 = vmlal_s16(q9s32, d28s16, d16s16);
-    q10s32 = vmlal_s16(q10s32, d28s16, d17s16);
-    q11s32 = vmlal_s16(q11s32, d29s16, d16s16);
-    q12s32 = vmlal_s16(q12s32, d29s16, d17s16);
+  q9s32 = vmlal_s16(q9s32, d28s16, d16s16);
+  q10s32 = vmlal_s16(q10s32, d28s16, d17s16);
+  q11s32 = vmlal_s16(q11s32, d29s16, d16s16);
+  q12s32 = vmlal_s16(q12s32, d29s16, d17s16);
 
-    q9s32 = vmlal_s16(q9s32, d26s16, d17s16);
-    q10s32 = vmlsl_s16(q10s32, d26s16, d16s16);
-    q11s32 = vmlal_s16(q11s32, d27s16, d17s16);
-    q12s32 = vmlsl_s16(q12s32, d27s16, d16s16);
+  q9s32 = vmlal_s16(q9s32, d26s16, d17s16);
+  q10s32 = vmlsl_s16(q10s32, d26s16, d16s16);
+  q11s32 = vmlal_s16(q11s32, d27s16, d17s16);
+  q12s32 = vmlsl_s16(q12s32, d27s16, d16s16);
 
-    d2s16 = vshrn_n_s32(q9s32, 12);
-    d6s16 = vshrn_n_s32(q10s32, 12);
-    d3s16 = vshrn_n_s32(q11s32, 12);
-    d7s16 = vshrn_n_s32(q12s32, 12);
-    q1s16 = vcombine_s16(d2s16, d3s16);
-    q3s16 = vcombine_s16(d6s16, d7s16);
+  d2s16 = vshrn_n_s32(q9s32, 12);
+  d6s16 = vshrn_n_s32(q10s32, 12);
+  d3s16 = vshrn_n_s32(q11s32, 12);
+  d7s16 = vshrn_n_s32(q12s32, 12);
+  q1s16 = vcombine_s16(d2s16, d3s16);
+  q3s16 = vcombine_s16(d6s16, d7s16);
 
-    // Part two
-    q9s32 = vdupq_n_s32(12000);
-    q10s32 = vdupq_n_s32(51000);
+  // Part two
+  q9s32 = vdupq_n_s32(12000);
+  q10s32 = vdupq_n_s32(51000);
 
-    v2tmp2 = vtrnq_s32(vreinterpretq_s32_s16(q0s16),
-                       vreinterpretq_s32_s16(q2s16));
-    v2tmp3 = vtrnq_s32(vreinterpretq_s32_s16(q1s16),
-                       vreinterpretq_s32_s16(q3s16));
-    v2tmp0 = vtrnq_s16(vreinterpretq_s16_s32(v2tmp2.val[0]),   // q0
-                       vreinterpretq_s16_s32(v2tmp3.val[0]));  // q1
-    v2tmp1 = vtrnq_s16(vreinterpretq_s16_s32(v2tmp2.val[1]),   // q2
-                       vreinterpretq_s16_s32(v2tmp3.val[1]));  // q3
+  v2tmp2 =
+      vtrnq_s32(vreinterpretq_s32_s16(q0s16), vreinterpretq_s32_s16(q2s16));
+  v2tmp3 =
+      vtrnq_s32(vreinterpretq_s32_s16(q1s16), vreinterpretq_s32_s16(q3s16));
+  v2tmp0 = vtrnq_s16(vreinterpretq_s16_s32(v2tmp2.val[0]),   // q0
+                     vreinterpretq_s16_s32(v2tmp3.val[0]));  // q1
+  v2tmp1 = vtrnq_s16(vreinterpretq_s16_s32(v2tmp2.val[1]),   // q2
+                     vreinterpretq_s16_s32(v2tmp3.val[1]));  // q3
 
-    q11s16 = vaddq_s16(v2tmp0.val[0], v2tmp1.val[1]);
-    q12s16 = vaddq_s16(v2tmp0.val[1], v2tmp1.val[0]);
-    q13s16 = vsubq_s16(v2tmp0.val[1], v2tmp1.val[0]);
-    q14s16 = vsubq_s16(v2tmp0.val[0], v2tmp1.val[1]);
+  q11s16 = vaddq_s16(v2tmp0.val[0], v2tmp1.val[1]);
+  q12s16 = vaddq_s16(v2tmp0.val[1], v2tmp1.val[0]);
+  q13s16 = vsubq_s16(v2tmp0.val[1], v2tmp1.val[0]);
+  q14s16 = vsubq_s16(v2tmp0.val[0], v2tmp1.val[1]);
 
-    q15s16 = vdupq_n_s16(7);
-    q11s16 = vaddq_s16(q11s16, q15s16);
-    q0s16 = vaddq_s16(q11s16, q12s16);
-    q1s16 = vsubq_s16(q11s16, q12s16);
+  q15s16 = vdupq_n_s16(7);
+  q11s16 = vaddq_s16(q11s16, q15s16);
+  q0s16 = vaddq_s16(q11s16, q12s16);
+  q1s16 = vsubq_s16(q11s16, q12s16);
 
-    q11s32 = q9s32;
-    q12s32 = q10s32;
+  q11s32 = q9s32;
+  q12s32 = q10s32;
 
-    d0s16 = vget_low_s16(q0s16);
-    d1s16 = vget_high_s16(q0s16);
-    d2s16 = vget_low_s16(q1s16);
-    d3s16 = vget_high_s16(q1s16);
+  d0s16 = vget_low_s16(q0s16);
+  d1s16 = vget_high_s16(q0s16);
+  d2s16 = vget_low_s16(q1s16);
+  d3s16 = vget_high_s16(q1s16);
 
-    d0s16 = vshr_n_s16(d0s16, 4);
-    d4s16 = vshr_n_s16(d1s16, 4);
-    d2s16 = vshr_n_s16(d2s16, 4);
-    d6s16 = vshr_n_s16(d3s16, 4);
+  d0s16 = vshr_n_s16(d0s16, 4);
+  d4s16 = vshr_n_s16(d1s16, 4);
+  d2s16 = vshr_n_s16(d2s16, 4);
+  d6s16 = vshr_n_s16(d3s16, 4);
 
-    d26s16 = vget_low_s16(q13s16);
-    d27s16 = vget_high_s16(q13s16);
-    d28s16 = vget_low_s16(q14s16);
-    d29s16 = vget_high_s16(q14s16);
+  d26s16 = vget_low_s16(q13s16);
+  d27s16 = vget_high_s16(q13s16);
+  d28s16 = vget_low_s16(q14s16);
+  d29s16 = vget_high_s16(q14s16);
 
-    q9s32 = vmlal_s16(q9s32, d28s16, d16s16);
-    q10s32 = vmlal_s16(q10s32, d28s16, d17s16);
-    q11s32 = vmlal_s16(q11s32, d29s16, d16s16);
-    q12s32 = vmlal_s16(q12s32, d29s16, d17s16);
+  q9s32 = vmlal_s16(q9s32, d28s16, d16s16);
+  q10s32 = vmlal_s16(q10s32, d28s16, d17s16);
+  q11s32 = vmlal_s16(q11s32, d29s16, d16s16);
+  q12s32 = vmlal_s16(q12s32, d29s16, d17s16);
 
-    q9s32 = vmlal_s16(q9s32, d26s16, d17s16);
-    q10s32 = vmlsl_s16(q10s32, d26s16, d16s16);
-    q11s32 = vmlal_s16(q11s32, d27s16, d17s16);
-    q12s32 = vmlsl_s16(q12s32, d27s16, d16s16);
+  q9s32 = vmlal_s16(q9s32, d26s16, d17s16);
+  q10s32 = vmlsl_s16(q10s32, d26s16, d16s16);
+  q11s32 = vmlal_s16(q11s32, d27s16, d17s16);
+  q12s32 = vmlsl_s16(q12s32, d27s16, d16s16);
 
-    d1s16 = vshrn_n_s32(q9s32, 16);
-    d3s16 = vshrn_n_s32(q10s32, 16);
-    d5s16 = vshrn_n_s32(q11s32, 16);
-    d7s16 = vshrn_n_s32(q12s32, 16);
+  d1s16 = vshrn_n_s32(q9s32, 16);
+  d3s16 = vshrn_n_s32(q10s32, 16);
+  d5s16 = vshrn_n_s32(q11s32, 16);
+  d7s16 = vshrn_n_s32(q12s32, 16);
 
-    qEmptys16 = vdupq_n_s16(0);
-    q14u16 = vceqq_s16(q14s16, qEmptys16);
-    q14u16 = vmvnq_u16(q14u16);
+  qEmptys16 = vdupq_n_s16(0);
+  q14u16 = vceqq_s16(q14s16, qEmptys16);
+  q14u16 = vmvnq_u16(q14u16);
 
-    d28u16 = vget_low_u16(q14u16);
-    d29u16 = vget_high_u16(q14u16);
-    d1s16 = vsub_s16(d1s16, vreinterpret_s16_u16(d28u16));
-    d5s16 = vsub_s16(d5s16, vreinterpret_s16_u16(d29u16));
+  d28u16 = vget_low_u16(q14u16);
+  d29u16 = vget_high_u16(q14u16);
+  d1s16 = vsub_s16(d1s16, vreinterpret_s16_u16(d28u16));
+  d5s16 = vsub_s16(d5s16, vreinterpret_s16_u16(d29u16));
 
-    q0s16 = vcombine_s16(d0s16, d1s16);
-    q1s16 = vcombine_s16(d2s16, d3s16);
-    q2s16 = vcombine_s16(d4s16, d5s16);
-    q3s16 = vcombine_s16(d6s16, d7s16);
+  q0s16 = vcombine_s16(d0s16, d1s16);
+  q1s16 = vcombine_s16(d2s16, d3s16);
+  q2s16 = vcombine_s16(d4s16, d5s16);
+  q3s16 = vcombine_s16(d6s16, d7s16);
 
-    vst1q_s16(output, q0s16);
-    vst1q_s16(output + 8, q1s16);
-    vst1q_s16(output + 16, q2s16);
-    vst1q_s16(output + 24, q3s16);
-    return;
+  vst1q_s16(output, q0s16);
+  vst1q_s16(output + 8, q1s16);
+  vst1q_s16(output + 16, q2s16);
+  vst1q_s16(output + 24, q3s16);
+  return;
 }
--- a/vp8/encoder/arm/neon/vp8_shortwalsh4x4_neon.c
+++ b/vp8/encoder/arm/neon/vp8_shortwalsh4x4_neon.c
@@ -13,117 +13,107 @@
 
 #ifdef VPX_INCOMPATIBLE_GCC
 #include "./vp8_rtcd.h"
-void vp8_short_walsh4x4_neon(
-        int16_t *input,
-        int16_t *output,
-        int pitch) {
+void vp8_short_walsh4x4_neon(int16_t *input, int16_t *output, int pitch) {
   vp8_short_walsh4x4_c(input, output, pitch);
 }
 #else
-void vp8_short_walsh4x4_neon(
-        int16_t *input,
-        int16_t *output,
-        int pitch) {
-    uint16x4_t d16u16;
-    int16x8_t q0s16, q1s16;
-    int16x4_t dEmptys16, d0s16, d1s16, d2s16, d3s16, d4s16, d5s16, d6s16, d7s16;
-    int32x4_t qEmptys32, q0s32, q1s32, q2s32, q3s32, q8s32;
-    int32x4_t q9s32, q10s32, q11s32, q15s32;
-    uint32x4_t q8u32, q9u32, q10u32, q11u32;
-    int16x4x2_t v2tmp0, v2tmp1;
-    int32x2x2_t v2tmp2, v2tmp3;
+void vp8_short_walsh4x4_neon(int16_t *input, int16_t *output, int pitch) {
+  uint16x4_t d16u16;
+  int16x8_t q0s16, q1s16;
+  int16x4_t dEmptys16, d0s16, d1s16, d2s16, d3s16, d4s16, d5s16, d6s16, d7s16;
+  int32x4_t qEmptys32, q0s32, q1s32, q2s32, q3s32, q8s32;
+  int32x4_t q9s32, q10s32, q11s32, q15s32;
+  uint32x4_t q8u32, q9u32, q10u32, q11u32;
+  int16x4x2_t v2tmp0, v2tmp1;
+  int32x2x2_t v2tmp2, v2tmp3;
 
-    dEmptys16 = vdup_n_s16(0);
-    qEmptys32 = vdupq_n_s32(0);
-    q15s32 = vdupq_n_s32(3);
+  dEmptys16 = vdup_n_s16(0);
+  qEmptys32 = vdupq_n_s32(0);
+  q15s32 = vdupq_n_s32(3);
 
-    d0s16 = vld1_s16(input);
-    input += pitch/2;
-    d1s16 = vld1_s16(input);
-    input += pitch/2;
-    d2s16 = vld1_s16(input);
-    input += pitch/2;
-    d3s16 = vld1_s16(input);
+  d0s16 = vld1_s16(input);
+  input += pitch / 2;
+  d1s16 = vld1_s16(input);
+  input += pitch / 2;
+  d2s16 = vld1_s16(input);
+  input += pitch / 2;
+  d3s16 = vld1_s16(input);
 
-    v2tmp2 = vtrn_s32(vreinterpret_s32_s16(d0s16),
-                      vreinterpret_s32_s16(d2s16));
-    v2tmp3 = vtrn_s32(vreinterpret_s32_s16(d1s16),
-                      vreinterpret_s32_s16(d3s16));
-    v2tmp0 = vtrn_s16(vreinterpret_s16_s32(v2tmp2.val[0]),   // d0
-                      vreinterpret_s16_s32(v2tmp3.val[0]));  // d1
-    v2tmp1 = vtrn_s16(vreinterpret_s16_s32(v2tmp2.val[1]),   // d2
-                      vreinterpret_s16_s32(v2tmp3.val[1]));  // d3
+  v2tmp2 = vtrn_s32(vreinterpret_s32_s16(d0s16), vreinterpret_s32_s16(d2s16));
+  v2tmp3 = vtrn_s32(vreinterpret_s32_s16(d1s16), vreinterpret_s32_s16(d3s16));
+  v2tmp0 = vtrn_s16(vreinterpret_s16_s32(v2tmp2.val[0]),   // d0
+                    vreinterpret_s16_s32(v2tmp3.val[0]));  // d1
+  v2tmp1 = vtrn_s16(vreinterpret_s16_s32(v2tmp2.val[1]),   // d2
+                    vreinterpret_s16_s32(v2tmp3.val[1]));  // d3
 
-    d4s16 = vadd_s16(v2tmp0.val[0], v2tmp1.val[0]);
-    d5s16 = vadd_s16(v2tmp0.val[1], v2tmp1.val[1]);
-    d6s16 = vsub_s16(v2tmp0.val[1], v2tmp1.val[1]);
-    d7s16 = vsub_s16(v2tmp0.val[0], v2tmp1.val[0]);
+  d4s16 = vadd_s16(v2tmp0.val[0], v2tmp1.val[0]);
+  d5s16 = vadd_s16(v2tmp0.val[1], v2tmp1.val[1]);
+  d6s16 = vsub_s16(v2tmp0.val[1], v2tmp1.val[1]);
+  d7s16 = vsub_s16(v2tmp0.val[0], v2tmp1.val[0]);
 
-    d4s16 = vshl_n_s16(d4s16, 2);
-    d5s16 = vshl_n_s16(d5s16, 2);
-    d6s16 = vshl_n_s16(d6s16, 2);
-    d7s16 = vshl_n_s16(d7s16, 2);
+  d4s16 = vshl_n_s16(d4s16, 2);
+  d5s16 = vshl_n_s16(d5s16, 2);
+  d6s16 = vshl_n_s16(d6s16, 2);
+  d7s16 = vshl_n_s16(d7s16, 2);
 
-    d16u16 = vceq_s16(d4s16, dEmptys16);
-    d16u16 = vmvn_u16(d16u16);
+  d16u16 = vceq_s16(d4s16, dEmptys16);
+  d16u16 = vmvn_u16(d16u16);
 
-    d0s16 = vadd_s16(d4s16, d5s16);
-    d3s16 = vsub_s16(d4s16, d5s16);
-    d1s16 = vadd_s16(d7s16, d6s16);
-    d2s16 = vsub_s16(d7s16, d6s16);
+  d0s16 = vadd_s16(d4s16, d5s16);
+  d3s16 = vsub_s16(d4s16, d5s16);
+  d1s16 = vadd_s16(d7s16, d6s16);
+  d2s16 = vsub_s16(d7s16, d6s16);
 
-    d0s16 = vsub_s16(d0s16, vreinterpret_s16_u16(d16u16));
+  d0s16 = vsub_s16(d0s16, vreinterpret_s16_u16(d16u16));
 
-    // Second for-loop
-    v2tmp2 = vtrn_s32(vreinterpret_s32_s16(d1s16),
-                      vreinterpret_s32_s16(d3s16));
-    v2tmp3 = vtrn_s32(vreinterpret_s32_s16(d0s16),
-                      vreinterpret_s32_s16(d2s16));
-    v2tmp0 = vtrn_s16(vreinterpret_s16_s32(v2tmp3.val[1]),   // d2
-                      vreinterpret_s16_s32(v2tmp2.val[1]));  // d3
-    v2tmp1 = vtrn_s16(vreinterpret_s16_s32(v2tmp3.val[0]),   // d0
-                      vreinterpret_s16_s32(v2tmp2.val[0]));  // d1
+  // Second for-loop
+  v2tmp2 = vtrn_s32(vreinterpret_s32_s16(d1s16), vreinterpret_s32_s16(d3s16));
+  v2tmp3 = vtrn_s32(vreinterpret_s32_s16(d0s16), vreinterpret_s32_s16(d2s16));
+  v2tmp0 = vtrn_s16(vreinterpret_s16_s32(v2tmp3.val[1]),   // d2
+                    vreinterpret_s16_s32(v2tmp2.val[1]));  // d3
+  v2tmp1 = vtrn_s16(vreinterpret_s16_s32(v2tmp3.val[0]),   // d0
+                    vreinterpret_s16_s32(v2tmp2.val[0]));  // d1
 
-    q8s32  = vaddl_s16(v2tmp1.val[0], v2tmp0.val[0]);
-    q9s32  = vaddl_s16(v2tmp1.val[1], v2tmp0.val[1]);
-    q10s32 = vsubl_s16(v2tmp1.val[1], v2tmp0.val[1]);
-    q11s32 = vsubl_s16(v2tmp1.val[0], v2tmp0.val[0]);
+  q8s32 = vaddl_s16(v2tmp1.val[0], v2tmp0.val[0]);
+  q9s32 = vaddl_s16(v2tmp1.val[1], v2tmp0.val[1]);
+  q10s32 = vsubl_s16(v2tmp1.val[1], v2tmp0.val[1]);
+  q11s32 = vsubl_s16(v2tmp1.val[0], v2tmp0.val[0]);
 
-    q0s32 = vaddq_s32(q8s32, q9s32);
-    q1s32 = vaddq_s32(q11s32, q10s32);
-    q2s32 = vsubq_s32(q11s32, q10s32);
-    q3s32 = vsubq_s32(q8s32, q9s32);
+  q0s32 = vaddq_s32(q8s32, q9s32);
+  q1s32 = vaddq_s32(q11s32, q10s32);
+  q2s32 = vsubq_s32(q11s32, q10s32);
+  q3s32 = vsubq_s32(q8s32, q9s32);
 
-    q8u32  = vcltq_s32(q0s32, qEmptys32);
-    q9u32  = vcltq_s32(q1s32, qEmptys32);
-    q10u32 = vcltq_s32(q2s32, qEmptys32);
-    q11u32 = vcltq_s32(q3s32, qEmptys32);
+  q8u32 = vcltq_s32(q0s32, qEmptys32);
+  q9u32 = vcltq_s32(q1s32, qEmptys32);
+  q10u32 = vcltq_s32(q2s32, qEmptys32);
+  q11u32 = vcltq_s32(q3s32, qEmptys32);
 
-    q8s32  = vreinterpretq_s32_u32(q8u32);
-    q9s32  = vreinterpretq_s32_u32(q9u32);
-    q10s32 = vreinterpretq_s32_u32(q10u32);
-    q11s32 = vreinterpretq_s32_u32(q11u32);
+  q8s32 = vreinterpretq_s32_u32(q8u32);
+  q9s32 = vreinterpretq_s32_u32(q9u32);
+  q10s32 = vreinterpretq_s32_u32(q10u32);
+  q11s32 = vreinterpretq_s32_u32(q11u32);
 
-    q0s32 = vsubq_s32(q0s32, q8s32);
-    q1s32 = vsubq_s32(q1s32, q9s32);
-    q2s32 = vsubq_s32(q2s32, q10s32);
-    q3s32 = vsubq_s32(q3s32, q11s32);
+  q0s32 = vsubq_s32(q0s32, q8s32);
+  q1s32 = vsubq_s32(q1s32, q9s32);
+  q2s32 = vsubq_s32(q2s32, q10s32);
+  q3s32 = vsubq_s32(q3s32, q11s32);
 
-    q8s32  = vaddq_s32(q0s32, q15s32);
-    q9s32  = vaddq_s32(q1s32, q15s32);
-    q10s32 = vaddq_s32(q2s32, q15s32);
-    q11s32 = vaddq_s32(q3s32, q15s32);
+  q8s32 = vaddq_s32(q0s32, q15s32);
+  q9s32 = vaddq_s32(q1s32, q15s32);
+  q10s32 = vaddq_s32(q2s32, q15s32);
+  q11s32 = vaddq_s32(q3s32, q15s32);
 
-    d0s16 = vshrn_n_s32(q8s32, 3);
-    d1s16 = vshrn_n_s32(q9s32, 3);
-    d2s16 = vshrn_n_s32(q10s32, 3);
-    d3s16 = vshrn_n_s32(q11s32, 3);
+  d0s16 = vshrn_n_s32(q8s32, 3);
+  d1s16 = vshrn_n_s32(q9s32, 3);
+  d2s16 = vshrn_n_s32(q10s32, 3);
+  d3s16 = vshrn_n_s32(q11s32, 3);
 
-    q0s16 = vcombine_s16(d0s16, d1s16);
-    q1s16 = vcombine_s16(d2s16, d3s16);
+  q0s16 = vcombine_s16(d0s16, d1s16);
+  q1s16 = vcombine_s16(d2s16, d3s16);
 
-    vst1q_s16(output, q0s16);
-    vst1q_s16(output + 8, q1s16);
-    return;
+  vst1q_s16(output, q0s16);
+  vst1q_s16(output + 8, q1s16);
+  return;
 }
 #endif  // VPX_INCOMPATIBLE_GCC
--- a/vp8/encoder/bitstream.c
+++ b/vp8/encoder/bitstream.c
@@ -8,7 +8,6 @@
  *  be found in the AUTHORS file in the root of the source tree.
  */
 
-
 #include "vp8/common/header.h"
 #include "encodemv.h"
 #include "vp8/common/entropymode.h"
@@ -25,24 +24,16 @@
 #include "defaultcoefcounts.h"
 #include "vp8/common/common.h"
 
-const int vp8cx_base_skip_false_prob[128] =
-{
-    255, 255, 255, 255, 255, 255, 255, 255,
-    255, 255, 255, 255, 255, 255, 255, 255,
-    255, 255, 255, 255, 255, 255, 255, 255,
-    255, 255, 255, 255, 255, 255, 255, 255,
-    255, 255, 255, 255, 255, 255, 255, 255,
-    255, 255, 255, 255, 255, 255, 255, 255,
-    255, 255, 255, 255, 255, 255, 255, 255,
-    251, 248, 244, 240, 236, 232, 229, 225,
-    221, 217, 213, 208, 204, 199, 194, 190,
-    187, 183, 179, 175, 172, 168, 164, 160,
-    157, 153, 149, 145, 142, 138, 134, 130,
-    127, 124, 120, 117, 114, 110, 107, 104,
-    101, 98,  95,  92,  89,  86,  83, 80,
-    77,  74,  71,  68,  65,  62,  59, 56,
-    53,  50,  47,  44,  41,  38,  35, 32,
-    30,  28,  26,  24,  22,  20,  18, 16,
+const int vp8cx_base_skip_false_prob[128] = {
+  255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255,
+  255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255,
+  255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255,
+  255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 251, 248, 244, 240,
+  236, 232, 229, 225, 221, 217, 213, 208, 204, 199, 194, 190, 187, 183, 179,
+  175, 172, 168, 164, 160, 157, 153, 149, 145, 142, 138, 134, 130, 127, 124,
+  120, 117, 114, 110, 107, 104, 101, 98,  95,  92,  89,  86,  83,  80,  77,
+  74,  71,  68,  65,  62,  59,  56,  53,  50,  47,  44,  41,  38,  35,  32,
+  30,  28,  26,  24,  22,  20,  18,  16,
 };
 
 #if defined(SECTIONBITS_OUTPUT)
@@ -51,7 +42,8 @@
 
 #ifdef VP8_ENTROPY_STATS
 int intra_mode_stats[10][10][10];
-static unsigned int tree_update_hist [BLOCK_TYPES] [COEF_BANDS] [PREV_COEF_CONTEXTS] [ENTROPY_NODES] [2];
+static unsigned int tree_update_hist[BLOCK_TYPES][COEF_BANDS]
+                                    [PREV_COEF_CONTEXTS][ENTROPY_NODES][2];
 extern unsigned int active_section;
 #endif
 
@@ -59,731 +51,612 @@
 int count_mb_seg[4] = { 0, 0, 0, 0 };
 #endif
 
+static void update_mode(vp8_writer *const w, int n, vp8_token tok[/* n */],
+                        vp8_tree tree, vp8_prob Pnew[/* n-1 */],
+                        vp8_prob Pcur[/* n-1 */],
+                        unsigned int bct[/* n-1 */][2],
+                        const unsigned int num_events[/* n */]) {
+  unsigned int new_b = 0, old_b = 0;
+  int i = 0;
 
-static void update_mode(
-    vp8_writer *const w,
-    int n,
-    vp8_token tok               [/* n */],
-    vp8_tree tree,
-    vp8_prob Pnew               [/* n-1 */],
-    vp8_prob Pcur               [/* n-1 */],
-    unsigned int bct            [/* n-1 */] [2],
-    const unsigned int num_events[/* n */]
-)
-{
-    unsigned int new_b = 0, old_b = 0;
-    int i = 0;
+  vp8_tree_probs_from_distribution(n--, tok, tree, Pnew, bct, num_events, 256,
+                                   1);
 
-    vp8_tree_probs_from_distribution(
-        n--, tok, tree,
-        Pnew, bct, num_events,
-        256, 1
-    );
+  do {
+    new_b += vp8_cost_branch(bct[i], Pnew[i]);
+    old_b += vp8_cost_branch(bct[i], Pcur[i]);
+  } while (++i < n);
 
-    do
-    {
-        new_b += vp8_cost_branch(bct[i], Pnew[i]);
-        old_b += vp8_cost_branch(bct[i], Pcur[i]);
-    }
-    while (++i < n);
+  if (new_b + (n << 8) < old_b) {
+    int j = 0;
 
-    if (new_b + (n << 8) < old_b)
-    {
-        int j = 0;
+    vp8_write_bit(w, 1);
 
-        vp8_write_bit(w, 1);
+    do {
+      const vp8_prob p = Pnew[j];
 
-        do
-        {
-            const vp8_prob p = Pnew[j];
-
-            vp8_write_literal(w, Pcur[j] = p ? p : 1, 8);
-        }
-        while (++j < n);
-    }
-    else
-        vp8_write_bit(w, 0);
+      vp8_write_literal(w, Pcur[j] = p ? p : 1, 8);
+    } while (++j < n);
+  } else
+    vp8_write_bit(w, 0);
 }
 
-static void update_mbintra_mode_probs(VP8_COMP *cpi)
-{
-    VP8_COMMON *const x = & cpi->common;
+static void update_mbintra_mode_probs(VP8_COMP *cpi) {
+  VP8_COMMON *const x = &cpi->common;
 
-    vp8_writer *const w = cpi->bc;
+  vp8_writer *const w = cpi->bc;
 
-    {
-        vp8_prob Pnew   [VP8_YMODES-1];
-        unsigned int bct [VP8_YMODES-1] [2];
+  {
+    vp8_prob Pnew[VP8_YMODES - 1];
+    unsigned int bct[VP8_YMODES - 1][2];
 
-        update_mode(
-            w, VP8_YMODES, vp8_ymode_encodings, vp8_ymode_tree,
-            Pnew, x->fc.ymode_prob, bct, (unsigned int *)cpi->mb.ymode_count
-        );
-    }
-    {
-        vp8_prob Pnew   [VP8_UV_MODES-1];
-        unsigned int bct [VP8_UV_MODES-1] [2];
+    update_mode(w, VP8_YMODES, vp8_ymode_encodings, vp8_ymode_tree, Pnew,
+                x->fc.ymode_prob, bct, (unsigned int *)cpi->mb.ymode_count);
+  }
+  {
+    vp8_prob Pnew[VP8_UV_MODES - 1];
+    unsigned int bct[VP8_UV_MODES - 1][2];
 
-        update_mode(
-            w, VP8_UV_MODES, vp8_uv_mode_encodings, vp8_uv_mode_tree,
-            Pnew, x->fc.uv_mode_prob, bct, (unsigned int *)cpi->mb.uv_mode_count
-        );
-    }
+    update_mode(w, VP8_UV_MODES, vp8_uv_mode_encodings, vp8_uv_mode_tree, Pnew,
+                x->fc.uv_mode_prob, bct, (unsigned int *)cpi->mb.uv_mode_count);
+  }
 }
 
-static void write_ymode(vp8_writer *bc, int m, const vp8_prob *p)
-{
-    vp8_write_token(bc, vp8_ymode_tree, p, vp8_ymode_encodings + m);
+static void write_ymode(vp8_writer *bc, int m, const vp8_prob *p) {
+  vp8_write_token(bc, vp8_ymode_tree, p, vp8_ymode_encodings + m);
 }
 
-static void kfwrite_ymode(vp8_writer *bc, int m, const vp8_prob *p)
-{
-    vp8_write_token(bc, vp8_kf_ymode_tree, p, vp8_kf_ymode_encodings + m);
+static void kfwrite_ymode(vp8_writer *bc, int m, const vp8_prob *p) {
+  vp8_write_token(bc, vp8_kf_ymode_tree, p, vp8_kf_ymode_encodings + m);
 }
 
-static void write_uv_mode(vp8_writer *bc, int m, const vp8_prob *p)
-{
-    vp8_write_token(bc, vp8_uv_mode_tree, p, vp8_uv_mode_encodings + m);
+static void write_uv_mode(vp8_writer *bc, int m, const vp8_prob *p) {
+  vp8_write_token(bc, vp8_uv_mode_tree, p, vp8_uv_mode_encodings + m);
 }
 
-
-static void write_bmode(vp8_writer *bc, int m, const vp8_prob *p)
-{
-    vp8_write_token(bc, vp8_bmode_tree, p, vp8_bmode_encodings + m);
+static void write_bmode(vp8_writer *bc, int m, const vp8_prob *p) {
+  vp8_write_token(bc, vp8_bmode_tree, p, vp8_bmode_encodings + m);
 }
 
-static void write_split(vp8_writer *bc, int x)
-{
-    vp8_write_token(
-        bc, vp8_mbsplit_tree, vp8_mbsplit_probs, vp8_mbsplit_encodings + x
-    );
+static void write_split(vp8_writer *bc, int x) {
+  vp8_write_token(bc, vp8_mbsplit_tree, vp8_mbsplit_probs,
+                  vp8_mbsplit_encodings + x);
 }
 
-void vp8_pack_tokens(vp8_writer *w, const TOKENEXTRA *p, int xcount)
-{
-    const TOKENEXTRA *stop = p + xcount;
-    unsigned int split;
-    int shift;
-    int count = w->count;
-    unsigned int range = w->range;
-    unsigned int lowvalue = w->lowvalue;
+void vp8_pack_tokens(vp8_writer *w, const TOKENEXTRA *p, int xcount) {
+  const TOKENEXTRA *stop = p + xcount;
+  unsigned int split;
+  int shift;
+  int count = w->count;
+  unsigned int range = w->range;
+  unsigned int lowvalue = w->lowvalue;
 
-    while (p < stop)
-    {
-        const int t = p->Token;
-        vp8_token *a = vp8_coef_encodings + t;
-        const vp8_extra_bit_struct *b = vp8_extra_bits + t;
-        int i = 0;
-        const unsigned char *pp = p->context_tree;
-        int v = a->value;
-        int n = a->Len;
+  while (p < stop) {
+    const int t = p->Token;
+    vp8_token *a = vp8_coef_encodings + t;
+    const vp8_extra_bit_struct *b = vp8_extra_bits + t;
+    int i = 0;
+    const unsigned char *pp = p->context_tree;
+    int v = a->value;
+    int n = a->Len;
 
-        if (p->skip_eob_node)
-        {
-            n--;
-            i = 2;
-        }
-
-        do
-        {
-            const int bb = (v >> --n) & 1;
-            split = 1 + (((range - 1) * pp[i>>1]) >> 8);
-            i = vp8_coef_tree[i+bb];
-
-            if (bb)
-            {
-                lowvalue += split;
-                range = range - split;
-            }
-            else
-            {
-                range = split;
-            }
-
-            shift = vp8_norm[range];
-            range <<= shift;
-            count += shift;
-
-            if (count >= 0)
-            {
-                int offset = shift - count;
-
-                if ((lowvalue << (offset - 1)) & 0x80000000)
-                {
-                    int x = w->pos - 1;
-
-                    while (x >= 0 && w->buffer[x] == 0xff)
-                    {
-                        w->buffer[x] = (unsigned char)0;
-                        x--;
-                    }
+    if (p->skip_eob_node) {
+      n--;
+      i = 2;
+    }
 
-                    w->buffer[x] += 1;
-                }
+    do {
+      const int bb = (v >> --n) & 1;
+      split = 1 + (((range - 1) * pp[i >> 1]) >> 8);
+      i = vp8_coef_tree[i + bb];
 
-                validate_buffer(w->buffer + w->pos,
-                                1,
-                                w->buffer_end,
-                                w->error);
-
-                w->buffer[w->pos++] = (lowvalue >> (24 - offset));
-                lowvalue <<= offset;
-                shift = count;
-                lowvalue &= 0xffffff;
-                count -= 8 ;
-            }
-
-            lowvalue <<= shift;
-        }
-        while (n);
-
-
-        if (b->base_val)
-        {
-            const int e = p->Extra, L = b->Len;
-
-            if (L)
-            {
-                const unsigned char *proba = b->prob;
-                const int v2 = e >> 1;
-                int n2 = L;              /* number of bits in v2, assumed nonzero */
-                i = 0;
-
-                do
-                {
-                    const int bb = (v2 >> --n2) & 1;
-                    split = 1 + (((range - 1) * proba[i>>1]) >> 8);
-                    i = b->tree[i+bb];
-
-                    if (bb)
-                    {
-                        lowvalue += split;
-                        range = range - split;
-                    }
-                    else
-                    {
-                        range = split;
-                    }
+      if (bb) {
+        lowvalue += split;
+        range = range - split;
+      } else {
+        range = split;
+      }
 
-                    shift = vp8_norm[range];
-                    range <<= shift;
-                    count += shift;
+      shift = vp8_norm[range];
+      range <<= shift;
+      count += shift;
 
-                    if (count >= 0)
-                    {
-                        int offset = shift - count;
+      if (count >= 0) {
+        int offset = shift - count;
 
-                        if ((lowvalue << (offset - 1)) & 0x80000000)
-                        {
-                            int x = w->pos - 1;
+        if ((lowvalue << (offset - 1)) & 0x80000000) {
+          int x = w->pos - 1;
 
-                            while (x >= 0 && w->buffer[x] == 0xff)
-                            {
-                                w->buffer[x] = (unsigned char)0;
-                                x--;
-                            }
+          while (x >= 0 && w->buffer[x] == 0xff) {
+            w->buffer[x] = (unsigned char)0;
+            x--;
+          }
 
-                            w->buffer[x] += 1;
-                        }
+          w->buffer[x] += 1;
+        }
 
-                        validate_buffer(w->buffer + w->pos,
-                                        1,
-                                        w->buffer_end,
-                                        w->error);
+        validate_buffer(w->buffer + w->pos, 1, w->buffer_end, w->error);
 
-                        w->buffer[w->pos++] = (lowvalue >> (24 - offset));
-                        lowvalue <<= offset;
-                        shift = count;
-                        lowvalue &= 0xffffff;
-                        count -= 8 ;
-                    }
+        w->buffer[w->pos++] = (lowvalue >> (24 - offset));
+        lowvalue <<= offset;
+        shift = count;
+        lowvalue &= 0xffffff;
+        count -= 8;
+      }
 
-                    lowvalue <<= shift;
-                }
-                while (n2);
-            }
+      lowvalue <<= shift;
+    } while (n);
 
+    if (b->base_val) {
+      const int e = p->Extra, L = b->Len;
 
-            {
+      if (L) {
+        const unsigned char *proba = b->prob;
+        const int v2 = e >> 1;
+        int n2 = L; /* number of bits in v2, assumed nonzero */
+        i = 0;
 
-                split = (range + 1) >> 1;
+        do {
+          const int bb = (v2 >> --n2) & 1;
+          split = 1 + (((range - 1) * proba[i >> 1]) >> 8);
+          i = b->tree[i + bb];
 
-                if (e & 1)
-                {
-                    lowvalue += split;
-                    range = range - split;
-                }
-                else
-                {
-                    range = split;
-                }
+          if (bb) {
+            lowvalue += split;
+            range = range - split;
+          } else {
+            range = split;
+          }
 
-                range <<= 1;
+          shift = vp8_norm[range];
+          range <<= shift;
+          count += shift;
 
-                if ((lowvalue & 0x80000000))
-                {
-                    int x = w->pos - 1;
+          if (count >= 0) {
+            int offset = shift - count;
 
-                    while (x >= 0 && w->buffer[x] == 0xff)
-                    {
-                        w->buffer[x] = (unsigned char)0;
-                        x--;
-                    }
+            if ((lowvalue << (offset - 1)) & 0x80000000) {
+              int x = w->pos - 1;
 
-                    w->buffer[x] += 1;
+              while (x >= 0 && w->buffer[x] == 0xff) {
+                w->buffer[x] = (unsigned char)0;
+                x--;
+              }
 
-                }
+              w->buffer[x] += 1;
+            }
 
-                lowvalue  <<= 1;
+            validate_buffer(w->buffer + w->pos, 1, w->buffer_end, w->error);
 
-                if (!++count)
-                {
-                    count = -8;
+            w->buffer[w->pos++] = (lowvalue >> (24 - offset));
+            lowvalue <<= offset;
+            shift = count;
+            lowvalue &= 0xffffff;
+            count -= 8;
+          }
 
-                    validate_buffer(w->buffer + w->pos,
-                                    1,
-                                    w->buffer_end,
-                                    w->error);
+          lowvalue <<= shift;
+        } while (n2);
+      }
 
-                    w->buffer[w->pos++] = (lowvalue >> 24);
-                    lowvalue &= 0xffffff;
-                }
-            }
+      {
+        split = (range + 1) >> 1;
 
+        if (e & 1) {
+          lowvalue += split;
+          range = range - split;
+        } else {
+          range = split;
         }
 
-        ++p;
+        range <<= 1;
+
+        if ((lowvalue & 0x80000000)) {
+          int x = w->pos - 1;
+
+          while (x >= 0 && w->buffer[x] == 0xff) {
+            w->buffer[x] = (unsigned char)0;
+            x--;
+          }
+
+          w->buffer[x] += 1;
+        }
+
+        lowvalue <<= 1;
+
+        if (!++count) {
+          count = -8;
+
+          validate_buffer(w->buffer + w->pos, 1, w->buffer_end, w->error);
+
+          w->buffer[w->pos++] = (lowvalue >> 24);
+          lowvalue &= 0xffffff;
+        }
+      }
     }
 
-    w->count = count;
-    w->lowvalue = lowvalue;
-    w->range = range;
+    ++p;
+  }
 
+  w->count = count;
+  w->lowvalue = lowvalue;
+  w->range = range;
 }
 
-static void write_partition_size(unsigned char *cx_data, int size)
-{
-    signed char csize;
+static void write_partition_size(unsigned char *cx_data, int size) {
+  signed char csize;
 
-    csize = size & 0xff;
-    *cx_data = csize;
-    csize = (size >> 8) & 0xff;
-    *(cx_data + 1) = csize;
-    csize = (size >> 16) & 0xff;
-    *(cx_data + 2) = csize;
-
+  csize = size & 0xff;
+  *cx_data = csize;
+  csize = (size >> 8) & 0xff;
+  *(cx_data + 1) = csize;
+  csize = (size >> 16) & 0xff;
+  *(cx_data + 2) = csize;
 }
 
 static void pack_tokens_into_partitions(VP8_COMP *cpi, unsigned char *cx_data,
-                                          unsigned char * cx_data_end,
-                                          int num_part)
-{
+                                        unsigned char *cx_data_end,
+                                        int num_part) {
+  int i;
+  unsigned char *ptr = cx_data;
+  unsigned char *ptr_end = cx_data_end;
+  vp8_writer *w;
 
-    int i;
-    unsigned char *ptr = cx_data;
-    unsigned char *ptr_end = cx_data_end;
-    vp8_writer * w;
+  for (i = 0; i < num_part; i++) {
+    int mb_row;
 
-    for (i = 0; i < num_part; i++)
-    {
-        int mb_row;
+    w = cpi->bc + i + 1;
 
-        w = cpi->bc + i + 1;
+    vp8_start_encode(w, ptr, ptr_end);
 
-        vp8_start_encode(w, ptr, ptr_end);
+    for (mb_row = i; mb_row < cpi->common.mb_rows; mb_row += num_part) {
+      const TOKENEXTRA *p = cpi->tplist[mb_row].start;
+      const TOKENEXTRA *stop = cpi->tplist[mb_row].stop;
+      int tokens = (int)(stop - p);
 
-        for (mb_row = i; mb_row < cpi->common.mb_rows; mb_row += num_part)
-        {
-            const TOKENEXTRA *p    = cpi->tplist[mb_row].start;
-            const TOKENEXTRA *stop = cpi->tplist[mb_row].stop;
-            int tokens = (int)(stop - p);
-
-            vp8_pack_tokens(w, p, tokens);
-        }
-
-        vp8_stop_encode(w);
-        ptr += w->pos;
+      vp8_pack_tokens(w, p, tokens);
     }
+
+    vp8_stop_encode(w);
+    ptr += w->pos;
+  }
 }
 
-
 #if CONFIG_MULTITHREAD
-static void pack_mb_row_tokens(VP8_COMP *cpi, vp8_writer *w)
-{
-    int mb_row;
+static void pack_mb_row_tokens(VP8_COMP *cpi, vp8_writer *w) {
+  int mb_row;
 
-    for (mb_row = 0; mb_row < cpi->common.mb_rows; mb_row++)
-    {
-        const TOKENEXTRA *p    = cpi->tplist[mb_row].start;
-        const TOKENEXTRA *stop = cpi->tplist[mb_row].stop;
-        int tokens = (int)(stop - p);
+  for (mb_row = 0; mb_row < cpi->common.mb_rows; mb_row++) {
+    const TOKENEXTRA *p = cpi->tplist[mb_row].start;
+    const TOKENEXTRA *stop = cpi->tplist[mb_row].stop;
+    int tokens = (int)(stop - p);
 
-        vp8_pack_tokens(w, p, tokens);
-    }
-
+    vp8_pack_tokens(w, p, tokens);
+  }
 }
 #endif  // CONFIG_MULTITHREAD
 
-static void write_mv_ref
-(
-    vp8_writer *w, MB_PREDICTION_MODE m, const vp8_prob *p
-)
-{
+static void write_mv_ref(vp8_writer *w, MB_PREDICTION_MODE m,
+                         const vp8_prob *p) {
 #if CONFIG_DEBUG
-    assert(NEARESTMV <= m  &&  m <= SPLITMV);
+  assert(NEARESTMV <= m && m <= SPLITMV);
 #endif
-    vp8_write_token(w, vp8_mv_ref_tree, p,
-                    vp8_mv_ref_encoding_array + (m - NEARESTMV));
+  vp8_write_token(w, vp8_mv_ref_tree, p,
+                  vp8_mv_ref_encoding_array + (m - NEARESTMV));
 }
 
-static void write_sub_mv_ref
-(
-    vp8_writer *w, B_PREDICTION_MODE m, const vp8_prob *p
-)
-{
+static void write_sub_mv_ref(vp8_writer *w, B_PREDICTION_MODE m,
+                             const vp8_prob *p) {
 #if CONFIG_DEBUG
-    assert(LEFT4X4 <= m  &&  m <= NEW4X4);
+  assert(LEFT4X4 <= m && m <= NEW4X4);
 #endif
-    vp8_write_token(w, vp8_sub_mv_ref_tree, p,
-                    vp8_sub_mv_ref_encoding_array + (m - LEFT4X4));
+  vp8_write_token(w, vp8_sub_mv_ref_tree, p,
+                  vp8_sub_mv_ref_encoding_array + (m - LEFT4X4));
 }
 
-static void write_mv
-(
-    vp8_writer *w, const MV *mv, const int_mv *ref, const MV_CONTEXT *mvc
-)
-{
-    MV e;
-    e.row = mv->row - ref->as_mv.row;
-    e.col = mv->col - ref->as_mv.col;
+static void write_mv(vp8_writer *w, const MV *mv, const int_mv *ref,
+                     const MV_CONTEXT *mvc) {
+  MV e;
+  e.row = mv->row - ref->as_mv.row;
+  e.col = mv->col - ref->as_mv.col;
 
-    vp8_encode_motion_vector(w, &e, mvc);
+  vp8_encode_motion_vector(w, &e, mvc);
 }
 
-static void write_mb_features(vp8_writer *w, const MB_MODE_INFO *mi, const MACROBLOCKD *x)
-{
-    /* Encode the MB segment id. */
-    if (x->segmentation_enabled && x->update_mb_segmentation_map)
-    {
-        switch (mi->segment_id)
-        {
-        case 0:
-            vp8_write(w, 0, x->mb_segment_tree_probs[0]);
-            vp8_write(w, 0, x->mb_segment_tree_probs[1]);
-            break;
-        case 1:
-            vp8_write(w, 0, x->mb_segment_tree_probs[0]);
-            vp8_write(w, 1, x->mb_segment_tree_probs[1]);
-            break;
-        case 2:
-            vp8_write(w, 1, x->mb_segment_tree_probs[0]);
-            vp8_write(w, 0, x->mb_segment_tree_probs[2]);
-            break;
-        case 3:
-            vp8_write(w, 1, x->mb_segment_tree_probs[0]);
-            vp8_write(w, 1, x->mb_segment_tree_probs[2]);
-            break;
+static void write_mb_features(vp8_writer *w, const MB_MODE_INFO *mi,
+                              const MACROBLOCKD *x) {
+  /* Encode the MB segment id. */
+  if (x->segmentation_enabled && x->update_mb_segmentation_map) {
+    switch (mi->segment_id) {
+      case 0:
+        vp8_write(w, 0, x->mb_segment_tree_probs[0]);
+        vp8_write(w, 0, x->mb_segment_tree_probs[1]);
+        break;
+      case 1:
+        vp8_write(w, 0, x->mb_segment_tree_probs[0]);
+        vp8_write(w, 1, x->mb_segment_tree_probs[1]);
+        break;
+      case 2:
+        vp8_write(w, 1, x->mb_segment_tree_probs[0]);
+        vp8_write(w, 0, x->mb_segment_tree_probs[2]);
+        break;
+      case 3:
+        vp8_write(w, 1, x->mb_segment_tree_probs[0]);
+        vp8_write(w, 1, x->mb_segment_tree_probs[2]);
+        break;
 
-            /* TRAP.. This should not happen */
-        default:
-            vp8_write(w, 0, x->mb_segment_tree_probs[0]);
-            vp8_write(w, 0, x->mb_segment_tree_probs[1]);
-            break;
-        }
+      /* TRAP.. This should not happen */
+      default:
+        vp8_write(w, 0, x->mb_segment_tree_probs[0]);
+        vp8_write(w, 0, x->mb_segment_tree_probs[1]);
+        break;
     }
+  }
 }
-void vp8_convert_rfct_to_prob(VP8_COMP *const cpi)
-{
-    const int *const rfct = cpi->mb.count_mb_ref_frame_usage;
-    const int rf_intra = rfct[INTRA_FRAME];
-    const int rf_inter = rfct[LAST_FRAME] + rfct[GOLDEN_FRAME] + rfct[ALTREF_FRAME];
+void vp8_convert_rfct_to_prob(VP8_COMP *const cpi) {
+  const int *const rfct = cpi->mb.count_mb_ref_frame_usage;
+  const int rf_intra = rfct[INTRA_FRAME];
+  const int rf_inter =
+      rfct[LAST_FRAME] + rfct[GOLDEN_FRAME] + rfct[ALTREF_FRAME];
 
-    /* Calculate the probabilities used to code the ref frame based on usage */
-    if (!(cpi->prob_intra_coded = rf_intra * 255 / (rf_intra + rf_inter)))
-        cpi->prob_intra_coded = 1;
+  /* Calculate the probabilities used to code the ref frame based on usage */
+  if (!(cpi->prob_intra_coded = rf_intra * 255 / (rf_intra + rf_inter)))
+    cpi->prob_intra_coded = 1;
 
-    cpi->prob_last_coded = rf_inter ? (rfct[LAST_FRAME] * 255) / rf_inter : 128;
+  cpi->prob_last_coded = rf_inter ? (rfct[LAST_FRAME] * 255) / rf_inter : 128;
 
-    if (!cpi->prob_last_coded)
-        cpi->prob_last_coded = 1;
+  if (!cpi->prob_last_coded) cpi->prob_last_coded = 1;
 
-    cpi->prob_gf_coded = (rfct[GOLDEN_FRAME] + rfct[ALTREF_FRAME])
-                  ? (rfct[GOLDEN_FRAME] * 255) / (rfct[GOLDEN_FRAME] + rfct[ALTREF_FRAME]) : 128;
+  cpi->prob_gf_coded = (rfct[GOLDEN_FRAME] + rfct[ALTREF_FRAME])
+                           ? (rfct[GOLDEN_FRAME] * 255) /
+                                 (rfct[GOLDEN_FRAME] + rfct[ALTREF_FRAME])
+                           : 128;
 
-    if (!cpi->prob_gf_coded)
-        cpi->prob_gf_coded = 1;
-
+  if (!cpi->prob_gf_coded) cpi->prob_gf_coded = 1;
 }
 
-static void pack_inter_mode_mvs(VP8_COMP *const cpi)
-{
-    VP8_COMMON *const pc = & cpi->common;
-    vp8_writer *const w = cpi->bc;
-    const MV_CONTEXT *mvc = pc->fc.mvc;
+static void pack_inter_mode_mvs(VP8_COMP *const cpi) {
+  VP8_COMMON *const pc = &cpi->common;
+  vp8_writer *const w = cpi->bc;
+  const MV_CONTEXT *mvc = pc->fc.mvc;
 
+  MODE_INFO *m = pc->mi;
+  const int mis = pc->mode_info_stride;
+  int mb_row = -1;
 
-    MODE_INFO *m = pc->mi;
-    const int mis = pc->mode_info_stride;
-    int mb_row = -1;
+  int prob_skip_false = 0;
 
-    int prob_skip_false = 0;
+  cpi->mb.partition_info = cpi->mb.pi;
 
-    cpi->mb.partition_info = cpi->mb.pi;
+  vp8_convert_rfct_to_prob(cpi);
 
-    vp8_convert_rfct_to_prob(cpi);
-
 #ifdef VP8_ENTROPY_STATS
-    active_section = 1;
+  active_section = 1;
 #endif
 
-    if (pc->mb_no_coeff_skip)
-    {
-        int total_mbs = pc->mb_rows * pc->mb_cols;
+  if (pc->mb_no_coeff_skip) {
+    int total_mbs = pc->mb_rows * pc->mb_cols;
 
-        prob_skip_false = (total_mbs - cpi->mb.skip_true_count ) * 256 / total_mbs;
+    prob_skip_false = (total_mbs - cpi->mb.skip_true_count) * 256 / total_mbs;
 
-        if (prob_skip_false <= 1)
-            prob_skip_false = 1;
+    if (prob_skip_false <= 1) prob_skip_false = 1;
 
-        if (prob_skip_false > 255)
-            prob_skip_false = 255;
+    if (prob_skip_false > 255) prob_skip_false = 255;
 
-        cpi->prob_skip_false = prob_skip_false;
-        vp8_write_literal(w, prob_skip_false, 8);
-    }
+    cpi->prob_skip_false = prob_skip_false;
+    vp8_write_literal(w, prob_skip_false, 8);
+  }
 
-    vp8_write_literal(w, cpi->prob_intra_coded, 8);
-    vp8_write_literal(w, cpi->prob_last_coded, 8);
-    vp8_write_literal(w, cpi->prob_gf_coded, 8);
+  vp8_write_literal(w, cpi->prob_intra_coded, 8);
+  vp8_write_literal(w, cpi->prob_last_coded, 8);
+  vp8_write_literal(w, cpi->prob_gf_coded, 8);
 
-    update_mbintra_mode_probs(cpi);
+  update_mbintra_mode_probs(cpi);
 
-    vp8_write_mvprobs(cpi);
+  vp8_write_mvprobs(cpi);
 
-    while (++mb_row < pc->mb_rows)
-    {
-        int mb_col = -1;
+  while (++mb_row < pc->mb_rows) {
+    int mb_col = -1;
 
-        while (++mb_col < pc->mb_cols)
-        {
-            const MB_MODE_INFO *const mi = & m->mbmi;
-            const MV_REFERENCE_FRAME rf = mi->ref_frame;
-            const MB_PREDICTION_MODE mode = mi->mode;
+    while (++mb_col < pc->mb_cols) {
+      const MB_MODE_INFO *const mi = &m->mbmi;
+      const MV_REFERENCE_FRAME rf = mi->ref_frame;
+      const MB_PREDICTION_MODE mode = mi->mode;
 
-            MACROBLOCKD *xd = &cpi->mb.e_mbd;
+      MACROBLOCKD *xd = &cpi->mb.e_mbd;
 
-            /* Distance of Mb to the various image edges.
-             * These specified to 8th pel as they are always compared to MV
-             * values that are in 1/8th pel units
-             */
-            xd->mb_to_left_edge = -((mb_col * 16) << 3);
-            xd->mb_to_right_edge = ((pc->mb_cols - 1 - mb_col) * 16) << 3;
-            xd->mb_to_top_edge = -((mb_row * 16) << 3);
-            xd->mb_to_bottom_edge = ((pc->mb_rows - 1 - mb_row) * 16) << 3;
+      /* Distance of Mb to the various image edges.
+       * These specified to 8th pel as they are always compared to MV
+       * values that are in 1/8th pel units
+       */
+      xd->mb_to_left_edge = -((mb_col * 16) << 3);
+      xd->mb_to_right_edge = ((pc->mb_cols - 1 - mb_col) * 16) << 3;
+      xd->mb_to_top_edge = -((mb_row * 16) << 3);
+      xd->mb_to_bottom_edge = ((pc->mb_rows - 1 - mb_row) * 16) << 3;
 
 #ifdef VP8_ENTROPY_STATS
-            active_section = 9;
+      active_section = 9;
 #endif
 
-            if (cpi->mb.e_mbd.update_mb_segmentation_map)
-                write_mb_features(w, mi, &cpi->mb.e_mbd);
+      if (cpi->mb.e_mbd.update_mb_segmentation_map)
+        write_mb_features(w, mi, &cpi->mb.e_mbd);
 
-            if (pc->mb_no_coeff_skip)
-                vp8_encode_bool(w, m->mbmi.mb_skip_coeff, prob_skip_false);
+      if (pc->mb_no_coeff_skip)
+        vp8_encode_bool(w, m->mbmi.mb_skip_coeff, prob_skip_false);
 
-            if (rf == INTRA_FRAME)
-            {
-                vp8_write(w, 0, cpi->prob_intra_coded);
+      if (rf == INTRA_FRAME) {
+        vp8_write(w, 0, cpi->prob_intra_coded);
 #ifdef VP8_ENTROPY_STATS
-                active_section = 6;
+        active_section = 6;
 #endif
-                write_ymode(w, mode, pc->fc.ymode_prob);
+        write_ymode(w, mode, pc->fc.ymode_prob);
 
-                if (mode == B_PRED)
-                {
-                    int j = 0;
+        if (mode == B_PRED) {
+          int j = 0;
 
-                    do
-                        write_bmode(w, m->bmi[j].as_mode, pc->fc.bmode_prob);
-                    while (++j < 16);
-                }
+          do
+            write_bmode(w, m->bmi[j].as_mode, pc->fc.bmode_prob);
+          while (++j < 16);
+        }
 
-                write_uv_mode(w, mi->uv_mode, pc->fc.uv_mode_prob);
-            }
-            else    /* inter coded */
-            {
-                int_mv best_mv;
-                vp8_prob mv_ref_p [VP8_MVREFS-1];
+        write_uv_mode(w, mi->uv_mode, pc->fc.uv_mode_prob);
+      } else /* inter coded */
+      {
+        int_mv best_mv;
+        vp8_prob mv_ref_p[VP8_MVREFS - 1];
 
-                vp8_write(w, 1, cpi->prob_intra_coded);
+        vp8_write(w, 1, cpi->prob_intra_coded);
 
-                if (rf == LAST_FRAME)
-                    vp8_write(w, 0, cpi->prob_last_coded);
-                else
-                {
-                    vp8_write(w, 1, cpi->prob_last_coded);
-                    vp8_write(w, (rf == GOLDEN_FRAME) ? 0 : 1, cpi->prob_gf_coded);
-                }
+        if (rf == LAST_FRAME)
+          vp8_write(w, 0, cpi->prob_last_coded);
+        else {
+          vp8_write(w, 1, cpi->prob_last_coded);
+          vp8_write(w, (rf == GOLDEN_FRAME) ? 0 : 1, cpi->prob_gf_coded);
+        }
 
-                {
-                    int_mv n1, n2;
-                    int ct[4];
+        {
+          int_mv n1, n2;
+          int ct[4];
 
-                    vp8_find_near_mvs(xd, m, &n1, &n2, &best_mv, ct, rf, cpi->common.ref_frame_sign_bias);
-                    vp8_clamp_mv2(&best_mv, xd);
+          vp8_find_near_mvs(xd, m, &n1, &n2, &best_mv, ct, rf,
+                            cpi->common.ref_frame_sign_bias);
+          vp8_clamp_mv2(&best_mv, xd);
 
-                    vp8_mv_ref_probs(mv_ref_p, ct);
+          vp8_mv_ref_probs(mv_ref_p, ct);
 
 #ifdef VP8_ENTROPY_STATS
-                    accum_mv_refs(mode, ct);
+          accum_mv_refs(mode, ct);
 #endif
+        }
 
-                }
-
 #ifdef VP8_ENTROPY_STATS
-                active_section = 3;
+        active_section = 3;
 #endif
 
-                write_mv_ref(w, mode, mv_ref_p);
+        write_mv_ref(w, mode, mv_ref_p);
 
-                switch (mode)   /* new, split require MVs */
-                {
-                case NEWMV:
+        switch (mode) /* new, split require MVs */
+        {
+          case NEWMV:
 
 #ifdef VP8_ENTROPY_STATS
-                    active_section = 5;
+            active_section = 5;
 #endif
 
-                    write_mv(w, &mi->mv.as_mv, &best_mv, mvc);
-                    break;
+            write_mv(w, &mi->mv.as_mv, &best_mv, mvc);
+            break;
 
-                case SPLITMV:
-                {
-                    int j = 0;
+          case SPLITMV: {
+            int j = 0;
 
 #ifdef MODE_STATS
-                    ++count_mb_seg [mi->partitioning];
+            ++count_mb_seg[mi->partitioning];
 #endif
 
-                    write_split(w, mi->partitioning);
+            write_split(w, mi->partitioning);
 
-                    do
-                    {
-                        B_PREDICTION_MODE blockmode;
-                        int_mv blockmv;
-                        const int *const  L = vp8_mbsplits [mi->partitioning];
-                        int k = -1;  /* first block in subset j */
-                        int mv_contz;
-                        int_mv leftmv, abovemv;
+            do {
+              B_PREDICTION_MODE blockmode;
+              int_mv blockmv;
+              const int *const L = vp8_mbsplits[mi->partitioning];
+              int k = -1; /* first block in subset j */
+              int mv_contz;
+              int_mv leftmv, abovemv;
 
-                        blockmode =  cpi->mb.partition_info->bmi[j].mode;
-                        blockmv =  cpi->mb.partition_info->bmi[j].mv;
+              blockmode = cpi->mb.partition_info->bmi[j].mode;
+              blockmv = cpi->mb.partition_info->bmi[j].mv;
 #if CONFIG_DEBUG
-                        while (j != L[++k])
-                            if (k >= 16)
-                                assert(0);
+              while (j != L[++k])
+                if (k >= 16) assert(0);
 #else
-                        while (j != L[++k]);
+              while (j != L[++k])
+                ;
 #endif
-                        leftmv.as_int = left_block_mv(m, k);
-                        abovemv.as_int = above_block_mv(m, k, mis);
-                        mv_contz = vp8_mv_cont(&leftmv, &abovemv);
+              leftmv.as_int = left_block_mv(m, k);
+              abovemv.as_int = above_block_mv(m, k, mis);
+              mv_contz = vp8_mv_cont(&leftmv, &abovemv);
 
-                        write_sub_mv_ref(w, blockmode, vp8_sub_mv_ref_prob2 [mv_contz]);
+              write_sub_mv_ref(w, blockmode, vp8_sub_mv_ref_prob2[mv_contz]);
 
-                        if (blockmode == NEW4X4)
-                        {
+              if (blockmode == NEW4X4) {
 #ifdef VP8_ENTROPY_STATS
-                            active_section = 11;
+                active_section = 11;
 #endif
-                            write_mv(w, &blockmv.as_mv, &best_mv, (const MV_CONTEXT *) mvc);
-                        }
-                    }
-                    while (++j < cpi->mb.partition_info->count);
-                }
-                break;
-                default:
-                    break;
-                }
-            }
+                write_mv(w, &blockmv.as_mv, &best_mv, (const MV_CONTEXT *)mvc);
+              }
+            } while (++j < cpi->mb.partition_info->count);
+          } break;
+          default: break;
+        }
+      }
 
-            ++m;
-            cpi->mb.partition_info++;
-        }
-
-        ++m;  /* skip L prediction border */
-        cpi->mb.partition_info++;
+      ++m;
+      cpi->mb.partition_info++;
     }
+
+    ++m; /* skip L prediction border */
+    cpi->mb.partition_info++;
+  }
 }
 
+static void write_kfmodes(VP8_COMP *cpi) {
+  vp8_writer *const bc = cpi->bc;
+  const VP8_COMMON *const c = &cpi->common;
+  /* const */
+  MODE_INFO *m = c->mi;
 
-static void write_kfmodes(VP8_COMP *cpi)
-{
-    vp8_writer *const bc = cpi->bc;
-    const VP8_COMMON *const c = & cpi->common;
-    /* const */
-    MODE_INFO *m = c->mi;
+  int mb_row = -1;
+  int prob_skip_false = 0;
 
-    int mb_row = -1;
-    int prob_skip_false = 0;
+  if (c->mb_no_coeff_skip) {
+    int total_mbs = c->mb_rows * c->mb_cols;
 
-    if (c->mb_no_coeff_skip)
-    {
-        int total_mbs = c->mb_rows * c->mb_cols;
+    prob_skip_false = (total_mbs - cpi->mb.skip_true_count) * 256 / total_mbs;
 
-        prob_skip_false = (total_mbs - cpi->mb.skip_true_count ) * 256 / total_mbs;
+    if (prob_skip_false <= 1) prob_skip_false = 1;
 
-        if (prob_skip_false <= 1)
-            prob_skip_false = 1;
+    if (prob_skip_false >= 255) prob_skip_false = 255;
 
-        if (prob_skip_false >= 255)
-            prob_skip_false = 255;
+    cpi->prob_skip_false = prob_skip_false;
+    vp8_write_literal(bc, prob_skip_false, 8);
+  }
 
-        cpi->prob_skip_false = prob_skip_false;
-        vp8_write_literal(bc, prob_skip_false, 8);
-    }
+  while (++mb_row < c->mb_rows) {
+    int mb_col = -1;
 
-    while (++mb_row < c->mb_rows)
-    {
-        int mb_col = -1;
+    while (++mb_col < c->mb_cols) {
+      const int ym = m->mbmi.mode;
 
-        while (++mb_col < c->mb_cols)
-        {
-            const int ym = m->mbmi.mode;
+      if (cpi->mb.e_mbd.update_mb_segmentation_map)
+        write_mb_features(bc, &m->mbmi, &cpi->mb.e_mbd);
 
-            if (cpi->mb.e_mbd.update_mb_segmentation_map)
-                write_mb_features(bc, &m->mbmi, &cpi->mb.e_mbd);
+      if (c->mb_no_coeff_skip)
+        vp8_encode_bool(bc, m->mbmi.mb_skip_coeff, prob_skip_false);
 
-            if (c->mb_no_coeff_skip)
-                vp8_encode_bool(bc, m->mbmi.mb_skip_coeff, prob_skip_false);
+      kfwrite_ymode(bc, ym, vp8_kf_ymode_prob);
 
-            kfwrite_ymode(bc, ym, vp8_kf_ymode_prob);
+      if (ym == B_PRED) {
+        const int mis = c->mode_info_stride;
+        int i = 0;
 
-            if (ym == B_PRED)
-            {
-                const int mis = c->mode_info_stride;
-                int i = 0;
+        do {
+          const B_PREDICTION_MODE A = above_block_mode(m, i, mis);
+          const B_PREDICTION_MODE L = left_block_mode(m, i);
+          const int bm = m->bmi[i].as_mode;
 
-                do
-                {
-                    const B_PREDICTION_MODE A = above_block_mode(m, i, mis);
-                    const B_PREDICTION_MODE L = left_block_mode(m, i);
-                    const int bm = m->bmi[i].as_mode;
-
 #ifdef VP8_ENTROPY_STATS
-                    ++intra_mode_stats [A] [L] [bm];
+          ++intra_mode_stats[A][L][bm];
 #endif
 
-                    write_bmode(bc, bm, vp8_kf_bmode_prob [A] [L]);
-                }
-                while (++i < 16);
-            }
+          write_bmode(bc, bm, vp8_kf_bmode_prob[A][L]);
+        } while (++i < 16);
+      }
 
-            write_uv_mode(bc, (m++)->mbmi.uv_mode, vp8_kf_uv_mode_prob);
-        }
-
-        m++;    /* skip L prediction border */
+      write_uv_mode(bc, (m++)->mbmi.uv_mode, vp8_kf_uv_mode_prob);
     }
+
+    m++; /* skip L prediction border */
+  }
 }
 
 #if 0
@@ -821,428 +694,345 @@
 #endif
 
 static void sum_probs_over_prev_coef_context(
-        const unsigned int probs[PREV_COEF_CONTEXTS][MAX_ENTROPY_TOKENS],
-        unsigned int* out)
-{
-    int i, j;
-    for (i=0; i < MAX_ENTROPY_TOKENS; ++i)
-    {
-        for (j=0; j < PREV_COEF_CONTEXTS; ++j)
-        {
-            const unsigned int tmp = out[i];
-            out[i] += probs[j][i];
-            /* check for wrap */
-            if (out[i] < tmp)
-                out[i] = UINT_MAX;
-        }
+    const unsigned int probs[PREV_COEF_CONTEXTS][MAX_ENTROPY_TOKENS],
+    unsigned int *out) {
+  int i, j;
+  for (i = 0; i < MAX_ENTROPY_TOKENS; ++i) {
+    for (j = 0; j < PREV_COEF_CONTEXTS; ++j) {
+      const unsigned int tmp = out[i];
+      out[i] += probs[j][i];
+      /* check for wrap */
+      if (out[i] < tmp) out[i] = UINT_MAX;
     }
+  }
 }
 
-static int prob_update_savings(const unsigned int *ct,
-                                   const vp8_prob oldp, const vp8_prob newp,
-                                   const vp8_prob upd)
-{
-    const int old_b = vp8_cost_branch(ct, oldp);
-    const int new_b = vp8_cost_branch(ct, newp);
-    const int update_b = 8 +
-                         ((vp8_cost_one(upd) - vp8_cost_zero(upd)) >> 8);
+static int prob_update_savings(const unsigned int *ct, const vp8_prob oldp,
+                               const vp8_prob newp, const vp8_prob upd) {
+  const int old_b = vp8_cost_branch(ct, oldp);
+  const int new_b = vp8_cost_branch(ct, newp);
+  const int update_b = 8 + ((vp8_cost_one(upd) - vp8_cost_zero(upd)) >> 8);
 
-    return old_b - new_b - update_b;
+  return old_b - new_b - update_b;
 }
 
-static int independent_coef_context_savings(VP8_COMP *cpi)
-{
-    MACROBLOCK *const x = & cpi->mb;
-    int savings = 0;
-    int i = 0;
-    do
-    {
-        int j = 0;
-        do
-        {
-            int k = 0;
-            unsigned int prev_coef_count_sum[MAX_ENTROPY_TOKENS] = {0};
-            int prev_coef_savings[MAX_ENTROPY_TOKENS] = {0};
-            const unsigned int (*probs)[MAX_ENTROPY_TOKENS];
-            /* Calculate new probabilities given the constraint that
-             * they must be equal over the prev coef contexts
-             */
+static int independent_coef_context_savings(VP8_COMP *cpi) {
+  MACROBLOCK *const x = &cpi->mb;
+  int savings = 0;
+  int i = 0;
+  do {
+    int j = 0;
+    do {
+      int k = 0;
+      unsigned int prev_coef_count_sum[MAX_ENTROPY_TOKENS] = { 0 };
+      int prev_coef_savings[MAX_ENTROPY_TOKENS] = { 0 };
+      const unsigned int(*probs)[MAX_ENTROPY_TOKENS];
+      /* Calculate new probabilities given the constraint that
+       * they must be equal over the prev coef contexts
+       */
 
-            probs = (const unsigned int (*)[MAX_ENTROPY_TOKENS])
-                x->coef_counts[i][j];
+      probs = (const unsigned int(*)[MAX_ENTROPY_TOKENS])x->coef_counts[i][j];
 
-            /* Reset to default probabilities at key frames */
-            if (cpi->common.frame_type == KEY_FRAME)
-                probs = default_coef_counts[i][j];
+      /* Reset to default probabilities at key frames */
+      if (cpi->common.frame_type == KEY_FRAME)
+        probs = default_coef_counts[i][j];
 
-            sum_probs_over_prev_coef_context(probs, prev_coef_count_sum);
+      sum_probs_over_prev_coef_context(probs, prev_coef_count_sum);
 
-            do
-            {
-                /* at every context */
+      do {
+        /* at every context */
 
-                /* calc probs and branch cts for this frame only */
-                int t = 0;      /* token/prob index */
+        /* calc probs and branch cts for this frame only */
+        int t = 0; /* token/prob index */
 
-                vp8_tree_probs_from_distribution(
-                    MAX_ENTROPY_TOKENS, vp8_coef_encodings, vp8_coef_tree,
-                    cpi->frame_coef_probs[i][j][k],
-                    cpi->frame_branch_ct [i][j][k],
-                    prev_coef_count_sum,
-                    256, 1);
+        vp8_tree_probs_from_distribution(
+            MAX_ENTROPY_TOKENS, vp8_coef_encodings, vp8_coef_tree,
+            cpi->frame_coef_probs[i][j][k], cpi->frame_branch_ct[i][j][k],
+            prev_coef_count_sum, 256, 1);
 
-                do
-                {
-                    const unsigned int *ct  = cpi->frame_branch_ct [i][j][k][t];
-                    const vp8_prob newp = cpi->frame_coef_probs [i][j][k][t];
-                    const vp8_prob oldp = cpi->common.fc.coef_probs [i][j][k][t];
-                    const vp8_prob upd = vp8_coef_update_probs [i][j][k][t];
-                    const int s = prob_update_savings(ct, oldp, newp, upd);
+        do {
+          const unsigned int *ct = cpi->frame_branch_ct[i][j][k][t];
+          const vp8_prob newp = cpi->frame_coef_probs[i][j][k][t];
+          const vp8_prob oldp = cpi->common.fc.coef_probs[i][j][k][t];
+          const vp8_prob upd = vp8_coef_update_probs[i][j][k][t];
+          const int s = prob_update_savings(ct, oldp, newp, upd);
 
-                    if (cpi->common.frame_type != KEY_FRAME ||
-                        (cpi->common.frame_type == KEY_FRAME && newp != oldp))
-                        prev_coef_savings[t] += s;
-                }
-                while (++t < ENTROPY_NODES);
-            }
-            while (++k < PREV_COEF_CONTEXTS);
-            k = 0;
-            do
-            {
-                /* We only update probabilities if we can save bits, except
-                 * for key frames where we have to update all probabilities
-                 * to get the equal probabilities across the prev coef
-                 * contexts.
-                 */
-                if (prev_coef_savings[k] > 0 ||
-                    cpi->common.frame_type == KEY_FRAME)
-                    savings += prev_coef_savings[k];
-            }
-            while (++k < ENTROPY_NODES);
-        }
-        while (++j < COEF_BANDS);
-    }
-    while (++i < BLOCK_TYPES);
-    return savings;
-}
+          if (cpi->common.frame_type != KEY_FRAME ||
+              (cpi->common.frame_type == KEY_FRAME && newp != oldp))
+            prev_coef_savings[t] += s;
+        } while (++t < ENTROPY_NODES);
+      } while (++k < PREV_COEF_CONTEXTS);
+      k = 0;
+      do {
+        /* We only update probabilities if we can save bits, except
+         * for key frames where we have to update all probabilities
+         * to get the equal probabilities across the prev coef
+         * contexts.
+         */
+        if (prev_coef_savings[k] > 0 || cpi->common.frame_type == KEY_FRAME)
+          savings += prev_coef_savings[k];
+      } while (++k < ENTROPY_NODES);
+    } while (++j < COEF_BANDS);
+  } while (++i < BLOCK_TYPES);
+  return savings;
+}
 
-static int default_coef_context_savings(VP8_COMP *cpi)
-{
-    MACROBLOCK *const x = & cpi->mb;
-    int savings = 0;
-    int i = 0;
-    do
-    {
-        int j = 0;
-        do
-        {
-            int k = 0;
-            do
-            {
-                /* at every context */
+static int default_coef_context_savings(VP8_COMP *cpi) {
+  MACROBLOCK *const x = &cpi->mb;
+  int savings = 0;
+  int i = 0;
+  do {
+    int j = 0;
+    do {
+      int k = 0;
+      do {
+        /* at every context */
 
-                /* calc probs and branch cts for this frame only */
-                int t = 0;      /* token/prob index */
+        /* calc probs and branch cts for this frame only */
+        int t = 0; /* token/prob index */
 
-                vp8_tree_probs_from_distribution(
-                    MAX_ENTROPY_TOKENS, vp8_coef_encodings, vp8_coef_tree,
-                    cpi->frame_coef_probs [i][j][k],
-                    cpi->frame_branch_ct [i][j][k],
-                    x->coef_counts [i][j][k],
-                    256, 1
-                );
+        vp8_tree_probs_from_distribution(
+            MAX_ENTROPY_TOKENS, vp8_coef_encodings, vp8_coef_tree,
+            cpi->frame_coef_probs[i][j][k], cpi->frame_branch_ct[i][j][k],
+            x->coef_counts[i][j][k], 256, 1);
 
-                do
-                {
-                    const unsigned int *ct  = cpi->frame_branch_ct [i][j][k][t];
-                    const vp8_prob newp = cpi->frame_coef_probs [i][j][k][t];
-                    const vp8_prob oldp = cpi->common.fc.coef_probs [i][j][k][t];
-                    const vp8_prob upd = vp8_coef_update_probs [i][j][k][t];
-                    const int s = prob_update_savings(ct, oldp, newp, upd);
+        do {
+          const unsigned int *ct = cpi->frame_branch_ct[i][j][k][t];
+          const vp8_prob newp = cpi->frame_coef_probs[i][j][k][t];
+          const vp8_prob oldp = cpi->common.fc.coef_probs[i][j][k][t];
+          const vp8_prob upd = vp8_coef_update_probs[i][j][k][t];
+          const int s = prob_update_savings(ct, oldp, newp, upd);
 
-                    if (s > 0)
-                    {
-                        savings += s;
-                    }
-                }
-                while (++t < ENTROPY_NODES);
-            }
-            while (++k < PREV_COEF_CONTEXTS);
-        }
-        while (++j < COEF_BANDS);
-    }
-    while (++i < BLOCK_TYPES);
-    return savings;
-}
+          if (s > 0) {
+            savings += s;
+          }
+        } while (++t < ENTROPY_NODES);
+      } while (++k < PREV_COEF_CONTEXTS);
+    } while (++j < COEF_BANDS);
+  } while (++i < BLOCK_TYPES);
+  return savings;
+}
 
-void vp8_calc_ref_frame_costs(int *ref_frame_cost,
-                              int prob_intra,
-                              int prob_last,
-                              int prob_garf
-                             )
-{
-    assert(prob_intra >= 0);
-    assert(prob_intra <= 255);
-    assert(prob_last >= 0);
-    assert(prob_last <= 255);
-    assert(prob_garf >= 0);
-    assert(prob_garf <= 255);
-    ref_frame_cost[INTRA_FRAME]   = vp8_cost_zero(prob_intra);
-    ref_frame_cost[LAST_FRAME]    = vp8_cost_one(prob_intra)
-                                    + vp8_cost_zero(prob_last);
-    ref_frame_cost[GOLDEN_FRAME]  = vp8_cost_one(prob_intra)
-                                    + vp8_cost_one(prob_last)
-                                    + vp8_cost_zero(prob_garf);
-    ref_frame_cost[ALTREF_FRAME]  = vp8_cost_one(prob_intra)
-                                    + vp8_cost_one(prob_last)
-                                    + vp8_cost_one(prob_garf);
-
+void vp8_calc_ref_frame_costs(int *ref_frame_cost, int prob_intra,
+                              int prob_last, int prob_garf) {
+  assert(prob_intra >= 0);
+  assert(prob_intra <= 255);
+  assert(prob_last >= 0);
+  assert(prob_last <= 255);
+  assert(prob_garf >= 0);
+  assert(prob_garf <= 255);
+  ref_frame_cost[INTRA_FRAME] = vp8_cost_zero(prob_intra);
+  ref_frame_cost[LAST_FRAME] =
+      vp8_cost_one(prob_intra) + vp8_cost_zero(prob_last);
+  ref_frame_cost[GOLDEN_FRAME] = vp8_cost_one(prob_intra) +
+                                 vp8_cost_one(prob_last) +
+                                 vp8_cost_zero(prob_garf);
+  ref_frame_cost[ALTREF_FRAME] = vp8_cost_one(prob_intra) +
+                                 vp8_cost_one(prob_last) +
+                                 vp8_cost_one(prob_garf);
 }
 
-int vp8_estimate_entropy_savings(VP8_COMP *cpi)
-{
-    int savings = 0;
+int vp8_estimate_entropy_savings(VP8_COMP *cpi) {
+  int savings = 0;
 
-    const int *const rfct = cpi->mb.count_mb_ref_frame_usage;
-    const int rf_intra = rfct[INTRA_FRAME];
-    const int rf_inter = rfct[LAST_FRAME] + rfct[GOLDEN_FRAME] + rfct[ALTREF_FRAME];
-    int new_intra, new_last, new_garf, oldtotal, newtotal;
-    int ref_frame_cost[MAX_REF_FRAMES];
+  const int *const rfct = cpi->mb.count_mb_ref_frame_usage;
+  const int rf_intra = rfct[INTRA_FRAME];
+  const int rf_inter =
+      rfct[LAST_FRAME] + rfct[GOLDEN_FRAME] + rfct[ALTREF_FRAME];
+  int new_intra, new_last, new_garf, oldtotal, newtotal;
+  int ref_frame_cost[MAX_REF_FRAMES];
 
-    vp8_clear_system_state();
+  vp8_clear_system_state();
 
-    if (cpi->common.frame_type != KEY_FRAME)
-    {
-        if (!(new_intra = rf_intra * 255 / (rf_intra + rf_inter)))
-            new_intra = 1;
+  if (cpi->common.frame_type != KEY_FRAME) {
+    if (!(new_intra = rf_intra * 255 / (rf_intra + rf_inter))) new_intra = 1;
 
-        new_last = rf_inter ? (rfct[LAST_FRAME] * 255) / rf_inter : 128;
+    new_last = rf_inter ? (rfct[LAST_FRAME] * 255) / rf_inter : 128;
 
-        new_garf = (rfct[GOLDEN_FRAME] + rfct[ALTREF_FRAME])
-                  ? (rfct[GOLDEN_FRAME] * 255) / (rfct[GOLDEN_FRAME] + rfct[ALTREF_FRAME]) : 128;
+    new_garf = (rfct[GOLDEN_FRAME] + rfct[ALTREF_FRAME])
+                   ? (rfct[GOLDEN_FRAME] * 255) /
+                         (rfct[GOLDEN_FRAME] + rfct[ALTREF_FRAME])
+                   : 128;
 
+    vp8_calc_ref_frame_costs(ref_frame_cost, new_intra, new_last, new_garf);
 
-        vp8_calc_ref_frame_costs(ref_frame_cost,new_intra,new_last,new_garf);
+    newtotal = rfct[INTRA_FRAME] * ref_frame_cost[INTRA_FRAME] +
+               rfct[LAST_FRAME] * ref_frame_cost[LAST_FRAME] +
+               rfct[GOLDEN_FRAME] * ref_frame_cost[GOLDEN_FRAME] +
+               rfct[ALTREF_FRAME] * ref_frame_cost[ALTREF_FRAME];
 
-        newtotal =
-            rfct[INTRA_FRAME] * ref_frame_cost[INTRA_FRAME] +
-            rfct[LAST_FRAME] * ref_frame_cost[LAST_FRAME] +
-            rfct[GOLDEN_FRAME] * ref_frame_cost[GOLDEN_FRAME] +
-            rfct[ALTREF_FRAME] * ref_frame_cost[ALTREF_FRAME];
+    /* old costs */
+    vp8_calc_ref_frame_costs(ref_frame_cost, cpi->prob_intra_coded,
+                             cpi->prob_last_coded, cpi->prob_gf_coded);
 
+    oldtotal = rfct[INTRA_FRAME] * ref_frame_cost[INTRA_FRAME] +
+               rfct[LAST_FRAME] * ref_frame_cost[LAST_FRAME] +
+               rfct[GOLDEN_FRAME] * ref_frame_cost[GOLDEN_FRAME] +
+               rfct[ALTREF_FRAME] * ref_frame_cost[ALTREF_FRAME];
 
-        /* old costs */
-        vp8_calc_ref_frame_costs(ref_frame_cost,cpi->prob_intra_coded,
-                                 cpi->prob_last_coded,cpi->prob_gf_coded);
+    savings += (oldtotal - newtotal) / 256;
+  }
 
-        oldtotal =
-            rfct[INTRA_FRAME] * ref_frame_cost[INTRA_FRAME] +
-            rfct[LAST_FRAME] * ref_frame_cost[LAST_FRAME] +
-            rfct[GOLDEN_FRAME] * ref_frame_cost[GOLDEN_FRAME] +
-            rfct[ALTREF_FRAME] * ref_frame_cost[ALTREF_FRAME];
+  if (cpi->oxcf.error_resilient_mode & VPX_ERROR_RESILIENT_PARTITIONS)
+    savings += independent_coef_context_savings(cpi);
+  else
+    savings += default_coef_context_savings(cpi);
 
-        savings += (oldtotal - newtotal) / 256;
-    }
-
-
-    if (cpi->oxcf.error_resilient_mode & VPX_ERROR_RESILIENT_PARTITIONS)
-        savings += independent_coef_context_savings(cpi);
-    else
-        savings += default_coef_context_savings(cpi);
-
-
-    return savings;
+  return savings;
 }
 
 #if CONFIG_REALTIME_ONLY & CONFIG_ONTHEFLY_BITPACKING
-int vp8_update_coef_context(VP8_COMP *cpi)
-{
-    int savings = 0;
+int vp8_update_coef_context(VP8_COMP *cpi) {
+  int savings = 0;
 
+  if (cpi->common.frame_type == KEY_FRAME) {
+    /* Reset to default counts/probabilities at key frames */
+    vp8_copy(cpi->mb.coef_counts, default_coef_counts);
+  }
 
-    if (cpi->common.frame_type == KEY_FRAME)
-    {
-        /* Reset to default counts/probabilities at key frames */
-        vp8_copy(cpi->mb.coef_counts, default_coef_counts);
-    }
+  if (cpi->oxcf.error_resilient_mode & VPX_ERROR_RESILIENT_PARTITIONS)
+    savings += independent_coef_context_savings(cpi);
+  else
+    savings += default_coef_context_savings(cpi);
 
-    if (cpi->oxcf.error_resilient_mode & VPX_ERROR_RESILIENT_PARTITIONS)
-        savings += independent_coef_context_savings(cpi);
-    else
-        savings += default_coef_context_savings(cpi);
-
-    return savings;
+  return savings;
 }
 #endif
 
-void vp8_update_coef_probs(VP8_COMP *cpi)
-{
-    int i = 0;
+void vp8_update_coef_probs(VP8_COMP *cpi) {
+  int i = 0;
 #if !(CONFIG_REALTIME_ONLY & CONFIG_ONTHEFLY_BITPACKING)
-    vp8_writer *const w = cpi->bc;
+  vp8_writer *const w = cpi->bc;
 #endif
-    int savings = 0;
+  int savings = 0;
 
-    vp8_clear_system_state();
+  vp8_clear_system_state();
 
-    do
-    {
-        int j = 0;
+  do {
+    int j = 0;
 
-        do
-        {
-            int k = 0;
-            int prev_coef_savings[ENTROPY_NODES] = {0};
-            if (cpi->oxcf.error_resilient_mode & VPX_ERROR_RESILIENT_PARTITIONS)
-            {
-                for (k = 0; k < PREV_COEF_CONTEXTS; ++k)
-                {
-                    int t;      /* token/prob index */
-                    for (t = 0; t < ENTROPY_NODES; ++t)
-                    {
-                        const unsigned int *ct = cpi->frame_branch_ct [i][j]
-                                                                      [k][t];
-                        const vp8_prob newp = cpi->frame_coef_probs[i][j][k][t];
-                        const vp8_prob oldp = cpi->common.fc.coef_probs[i][j]
-                                                                       [k][t];
-                        const vp8_prob upd = vp8_coef_update_probs[i][j][k][t];
+    do {
+      int k = 0;
+      int prev_coef_savings[ENTROPY_NODES] = { 0 };
+      if (cpi->oxcf.error_resilient_mode & VPX_ERROR_RESILIENT_PARTITIONS) {
+        for (k = 0; k < PREV_COEF_CONTEXTS; ++k) {
+          int t; /* token/prob index */
+          for (t = 0; t < ENTROPY_NODES; ++t) {
+            const unsigned int *ct = cpi->frame_branch_ct[i][j][k][t];
+            const vp8_prob newp = cpi->frame_coef_probs[i][j][k][t];
+            const vp8_prob oldp = cpi->common.fc.coef_probs[i][j][k][t];
+            const vp8_prob upd = vp8_coef_update_probs[i][j][k][t];
 
-                        prev_coef_savings[t] +=
-                                prob_update_savings(ct, oldp, newp, upd);
-                    }
-                }
-                k = 0;
-            }
-            do
-            {
-                /* note: use result from vp8_estimate_entropy_savings, so no
-                 * need to call vp8_tree_probs_from_distribution here.
-                 */
+            prev_coef_savings[t] += prob_update_savings(ct, oldp, newp, upd);
+          }
+        }
+        k = 0;
+      }
+      do {
+        /* note: use result from vp8_estimate_entropy_savings, so no
+         * need to call vp8_tree_probs_from_distribution here.
+         */
 
-                /* at every context */
+        /* at every context */
 
-                /* calc probs and branch cts for this frame only */
-                int t = 0;      /* token/prob index */
+        /* calc probs and branch cts for this frame only */
+        int t = 0; /* token/prob index */
 
-                do
-                {
-                    const vp8_prob newp = cpi->frame_coef_probs [i][j][k][t];
+        do {
+          const vp8_prob newp = cpi->frame_coef_probs[i][j][k][t];
 
-                    vp8_prob *Pold = cpi->common.fc.coef_probs [i][j][k] + t;
-                    const vp8_prob upd = vp8_coef_update_probs [i][j][k][t];
+          vp8_prob *Pold = cpi->common.fc.coef_probs[i][j][k] + t;
+          const vp8_prob upd = vp8_coef_update_probs[i][j][k][t];
 
-                    int s = prev_coef_savings[t];
-                    int u = 0;
+          int s = prev_coef_savings[t];
+          int u = 0;
 
-                    if (!(cpi->oxcf.error_resilient_mode &
-                            VPX_ERROR_RESILIENT_PARTITIONS))
-                    {
-                        s = prob_update_savings(
-                                cpi->frame_branch_ct [i][j][k][t],
-                                *Pold, newp, upd);
-                    }
+          if (!(cpi->oxcf.error_resilient_mode &
+                VPX_ERROR_RESILIENT_PARTITIONS)) {
+            s = prob_update_savings(cpi->frame_branch_ct[i][j][k][t], *Pold,
+                                    newp, upd);
+          }
 
-                    if (s > 0)
-                        u = 1;
+          if (s > 0) u = 1;
 
-                    /* Force updates on key frames if the new is different,
-                     * so that we can be sure we end up with equal probabilities
-                     * over the prev coef contexts.
-                     */
-                    if ((cpi->oxcf.error_resilient_mode &
-                            VPX_ERROR_RESILIENT_PARTITIONS) &&
-                        cpi->common.frame_type == KEY_FRAME && newp != *Pold)
-                        u = 1;
+          /* Force updates on key frames if the new is different,
+           * so that we can be sure we end up with equal probabilities
+           * over the prev coef contexts.
+           */
+          if ((cpi->oxcf.error_resilient_mode &
+               VPX_ERROR_RESILIENT_PARTITIONS) &&
+              cpi->common.frame_type == KEY_FRAME && newp != *Pold)
+            u = 1;
 
 #if CONFIG_REALTIME_ONLY & CONFIG_ONTHEFLY_BITPACKING
-                    cpi->update_probs[i][j][k][t] = u;
+          cpi->update_probs[i][j][k][t] = u;
 #else
-                    vp8_write(w, u, upd);
+          vp8_write(w, u, upd);
 #endif
 
-
 #ifdef VP8_ENTROPY_STATS
-                    ++ tree_update_hist [i][j][k][t] [u];
+          ++tree_update_hist[i][j][k][t][u];
 #endif
 
-                    if (u)
-                    {
-                        /* send/use new probability */
+          if (u) {
+            /* send/use new probability */
 
-                        *Pold = newp;
+            *Pold = newp;
 #if !(CONFIG_REALTIME_ONLY & CONFIG_ONTHEFLY_BITPACKING)
-                        vp8_write_literal(w, newp, 8);
+            vp8_write_literal(w, newp, 8);
 #endif
 
-                        savings += s;
+            savings += s;
+          }
 
-                    }
+        } while (++t < ENTROPY_NODES);
 
-                }
-                while (++t < ENTROPY_NODES);
-
-                /* Accum token counts for generation of default statistics */
+/* Accum token counts for generation of default statistics */
 #ifdef VP8_ENTROPY_STATS
-                t = 0;
+        t = 0;
 
-                do
-                {
-                    context_counters [i][j][k][t] += cpi->coef_counts [i][j][k][t];
-                }
-                while (++t < MAX_ENTROPY_TOKENS);
+        do {
+          context_counters[i][j][k][t] += cpi->coef_counts[i][j][k][t];
+        } while (++t < MAX_ENTROPY_TOKENS);
 
 #endif
 
-            }
-            while (++k < PREV_COEF_CONTEXTS);
-        }
-        while (++j < COEF_BANDS);
-    }
-    while (++i < BLOCK_TYPES);
-
+      } while (++k < PREV_COEF_CONTEXTS);
+    } while (++j < COEF_BANDS);
+  } while (++i < BLOCK_TYPES);
 }
 
 #if CONFIG_REALTIME_ONLY & CONFIG_ONTHEFLY_BITPACKING
-static void pack_coef_probs(VP8_COMP *cpi)
-{
-    int i = 0;
-    vp8_writer *const w = cpi->bc;
+static void pack_coef_probs(VP8_COMP *cpi) {
+  int i = 0;
+  vp8_writer *const w = cpi->bc;
 
-    do
-    {
-        int j = 0;
+  do {
+    int j = 0;
 
-        do
-        {
-            int k = 0;
+    do {
+      int k = 0;
 
-            do
-            {
-                int t = 0;      /* token/prob index */
+      do {
+        int t = 0; /* token/prob index */
 
-                do
-                {
-                    const vp8_prob newp = cpi->common.fc.coef_probs [i][j][k][t];
-                    const vp8_prob upd = vp8_coef_update_probs [i][j][k][t];
+        do {
+          const vp8_prob newp = cpi->common.fc.coef_probs[i][j][k][t];
+          const vp8_prob upd = vp8_coef_update_probs[i][j][k][t];
 
-                    const char u = cpi->update_probs[i][j][k][t] ;
+          const char u = cpi->update_probs[i][j][k][t];
 
-                    vp8_write(w, u, upd);
+          vp8_write(w, u, upd);
 
-                    if (u)
-                    {
-                        /* send/use new probability */
-                        vp8_write_literal(w, newp, 8);
-                    }
-                }
-                while (++t < ENTROPY_NODES);
-            }
-            while (++k < PREV_COEF_CONTEXTS);
-        }
-        while (++j < COEF_BANDS);
-    }
-    while (++i < BLOCK_TYPES);
-}
+          if (u) {
+            /* send/use new probability */
+            vp8_write_literal(w, newp, 8);
+          }
+        } while (++t < ENTROPY_NODES);
+      } while (++k < PREV_COEF_CONTEXTS);
+    } while (++j < COEF_BANDS);
+  } while (++i < BLOCK_TYPES);
+}
 #endif
 
 #ifdef PACKET_TESTING
@@ -1249,491 +1039,434 @@
 FILE *vpxlogc = 0;
 #endif
 
-static void put_delta_q(vp8_writer *bc, int delta_q)
-{
-    if (delta_q != 0)
-    {
-        vp8_write_bit(bc, 1);
-        vp8_write_literal(bc, abs(delta_q), 4);
+static void put_delta_q(vp8_writer *bc, int delta_q) {
+  if (delta_q != 0) {
+    vp8_write_bit(bc, 1);
+    vp8_write_literal(bc, abs(delta_q), 4);
 
-        if (delta_q < 0)
-            vp8_write_bit(bc, 1);
-        else
-            vp8_write_bit(bc, 0);
-    }
+    if (delta_q < 0)
+      vp8_write_bit(bc, 1);
     else
-        vp8_write_bit(bc, 0);
+      vp8_write_bit(bc, 0);
+  } else
+    vp8_write_bit(bc, 0);
 }
 
-void vp8_pack_bitstream(VP8_COMP *cpi, unsigned char *dest, unsigned char * dest_end, unsigned long *size)
-{
-    int i, j;
-    VP8_HEADER oh;
-    VP8_COMMON *const pc = & cpi->common;
-    vp8_writer *const bc = cpi->bc;
-    MACROBLOCKD *const xd = & cpi->mb.e_mbd;
-    int extra_bytes_packed = 0;
+void vp8_pack_bitstream(VP8_COMP *cpi, unsigned char *dest,
+                        unsigned char *dest_end, unsigned long *size) {
+  int i, j;
+  VP8_HEADER oh;
+  VP8_COMMON *const pc = &cpi->common;
+  vp8_writer *const bc = cpi->bc;
+  MACROBLOCKD *const xd = &cpi->mb.e_mbd;
+  int extra_bytes_packed = 0;
 
-    unsigned char *cx_data = dest;
-    unsigned char *cx_data_end = dest_end;
-    const int *mb_feature_data_bits;
+  unsigned char *cx_data = dest;
+  unsigned char *cx_data_end = dest_end;
+  const int *mb_feature_data_bits;
 
-    oh.show_frame = (int) pc->show_frame;
-    oh.type = (int)pc->frame_type;
-    oh.version = pc->version;
-    oh.first_partition_length_in_bytes = 0;
+  oh.show_frame = (int)pc->show_frame;
+  oh.type = (int)pc->frame_type;
+  oh.version = pc->version;
+  oh.first_partition_length_in_bytes = 0;
 
-    mb_feature_data_bits = vp8_mb_feature_data_bits;
+  mb_feature_data_bits = vp8_mb_feature_data_bits;
 
-    bc[0].error = &pc->error;
+  bc[0].error = &pc->error;
 
-    validate_buffer(cx_data, 3, cx_data_end, &cpi->common.error);
-    cx_data += 3;
+  validate_buffer(cx_data, 3, cx_data_end, &cpi->common.error);
+  cx_data += 3;
 
 #if defined(SECTIONBITS_OUTPUT)
-    Sectionbits[active_section = 1] += sizeof(VP8_HEADER) * 8 * 256;
+  Sectionbits[active_section = 1] += sizeof(VP8_HEADER) * 8 * 256;
 #endif
 
-    /* every keyframe send startcode, width, height, scale factor, clamp
-     * and color type
-     */
-    if (oh.type == KEY_FRAME)
-    {
-        int v;
+  /* every keyframe send startcode, width, height, scale factor, clamp
+   * and color type
+   */
+  if (oh.type == KEY_FRAME) {
+    int v;
 
-        validate_buffer(cx_data, 7, cx_data_end, &cpi->common.error);
+    validate_buffer(cx_data, 7, cx_data_end, &cpi->common.error);
 
-        /* Start / synch code */
-        cx_data[0] = 0x9D;
-        cx_data[1] = 0x01;
-        cx_data[2] = 0x2a;
+    /* Start / synch code */
+    cx_data[0] = 0x9D;
+    cx_data[1] = 0x01;
+    cx_data[2] = 0x2a;
 
-        v = (pc->horiz_scale << 14) | pc->Width;
-        cx_data[3] = v;
-        cx_data[4] = v >> 8;
+    v = (pc->horiz_scale << 14) | pc->Width;
+    cx_data[3] = v;
+    cx_data[4] = v >> 8;
 
-        v = (pc->vert_scale << 14) | pc->Height;
-        cx_data[5] = v;
-        cx_data[6] = v >> 8;
+    v = (pc->vert_scale << 14) | pc->Height;
+    cx_data[5] = v;
+    cx_data[6] = v >> 8;
 
+    extra_bytes_packed = 7;
+    cx_data += extra_bytes_packed;
 
-        extra_bytes_packed = 7;
-        cx_data += extra_bytes_packed ;
+    vp8_start_encode(bc, cx_data, cx_data_end);
 
-        vp8_start_encode(bc, cx_data, cx_data_end);
+    /* signal clr type */
+    vp8_write_bit(bc, 0);
+    vp8_write_bit(bc, pc->clamp_type);
 
-        /* signal clr type */
-        vp8_write_bit(bc, 0);
-        vp8_write_bit(bc, pc->clamp_type);
+  } else
+    vp8_start_encode(bc, cx_data, cx_data_end);
 
-    }
-    else
-        vp8_start_encode(bc, cx_data, cx_data_end);
+  /* Signal whether or not Segmentation is enabled */
+  vp8_write_bit(bc, xd->segmentation_enabled);
 
+  /*  Indicate which features are enabled */
+  if (xd->segmentation_enabled) {
+    /* Signal whether or not the segmentation map is being updated. */
+    vp8_write_bit(bc, xd->update_mb_segmentation_map);
+    vp8_write_bit(bc, xd->update_mb_segmentation_data);
 
-    /* Signal whether or not Segmentation is enabled */
-    vp8_write_bit(bc, xd->segmentation_enabled);
+    if (xd->update_mb_segmentation_data) {
+      signed char Data;
 
-    /*  Indicate which features are enabled */
-    if (xd->segmentation_enabled)
-    {
-        /* Signal whether or not the segmentation map is being updated. */
-        vp8_write_bit(bc, xd->update_mb_segmentation_map);
-        vp8_write_bit(bc, xd->update_mb_segmentation_data);
+      vp8_write_bit(bc, xd->mb_segement_abs_delta);
 
-        if (xd->update_mb_segmentation_data)
-        {
-            signed char Data;
+      /* For each segmentation feature (Quant and loop filter level) */
+      for (i = 0; i < MB_LVL_MAX; i++) {
+        /* For each of the segments */
+        for (j = 0; j < MAX_MB_SEGMENTS; j++) {
+          Data = xd->segment_feature_data[i][j];
 
-            vp8_write_bit(bc, xd->mb_segement_abs_delta);
-
-            /* For each segmentation feature (Quant and loop filter level) */
-            for (i = 0; i < MB_LVL_MAX; i++)
-            {
-                /* For each of the segments */
-                for (j = 0; j < MAX_MB_SEGMENTS; j++)
-                {
-                    Data = xd->segment_feature_data[i][j];
-
-                    /* Frame level data */
-                    if (Data)
-                    {
-                        vp8_write_bit(bc, 1);
-
-                        if (Data < 0)
-                        {
-                            Data = - Data;
-                            vp8_write_literal(bc, Data, mb_feature_data_bits[i]);
-                            vp8_write_bit(bc, 1);
-                        }
-                        else
-                        {
-                            vp8_write_literal(bc, Data, mb_feature_data_bits[i]);
-                            vp8_write_bit(bc, 0);
-                        }
-                    }
-                    else
-                        vp8_write_bit(bc, 0);
-                }
-            }
-        }
+          /* Frame level data */
+          if (Data) {
+            vp8_write_bit(bc, 1);
+
+            if (Data < 0) {
+              Data = -Data;
+              vp8_write_literal(bc, Data, mb_feature_data_bits[i]);
+              vp8_write_bit(bc, 1);
+            } else {
+              vp8_write_literal(bc, Data, mb_feature_data_bits[i]);
+              vp8_write_bit(bc, 0);
+            }
+          } else
+            vp8_write_bit(bc, 0);
+        }
+      }
+    }
 
-        if (xd->update_mb_segmentation_map)
-        {
-            /* Write the probs used to decode the segment id for each mb */
-            for (i = 0; i < MB_FEATURE_TREE_PROBS; i++)
-            {
-                int Data = xd->mb_segment_tree_probs[i];
+    if (xd->update_mb_segmentation_map) {
+      /* Write the probs used to decode the segment id for each mb */
+      for (i = 0; i < MB_FEATURE_TREE_PROBS; i++) {
+        int Data = xd->mb_segment_tree_probs[i];
 
-                if (Data != 255)
-                {
-                    vp8_write_bit(bc, 1);
-                    vp8_write_literal(bc, Data, 8);
-                }
-                else
-                    vp8_write_bit(bc, 0);
-            }
-        }
+        if (Data != 255) {
+          vp8_write_bit(bc, 1);
+          vp8_write_literal(bc, Data, 8);
+        } else
+          vp8_write_bit(bc, 0);
+      }
     }
+  }
 
-    vp8_write_bit(bc, pc->filter_type);
-    vp8_write_literal(bc, pc->filter_level, 6);
-    vp8_write_literal(bc, pc->sharpness_level, 3);
+  vp8_write_bit(bc, pc->filter_type);
+  vp8_write_literal(bc, pc->filter_level, 6);
+  vp8_write_literal(bc, pc->sharpness_level, 3);
 
-    /* Write out loop filter deltas applied at the MB level based on mode
-     * or ref frame (if they are enabled).
-     */
-    vp8_write_bit(bc, xd->mode_ref_lf_delta_enabled);
+  /* Write out loop filter deltas applied at the MB level based on mode
+   * or ref frame (if they are enabled).
+   */
+  vp8_write_bit(bc, xd->mode_ref_lf_delta_enabled);
 
-    if (xd->mode_ref_lf_delta_enabled)
-    {
-        /* Do the deltas need to be updated */
-        int send_update = xd->mode_ref_lf_delta_update
-                          || cpi->oxcf.error_resilient_mode;
+  if (xd->mode_ref_lf_delta_enabled) {
+    /* Do the deltas need to be updated */
+    int send_update =
+        xd->mode_ref_lf_delta_update || cpi->oxcf.error_resilient_mode;
 
-        vp8_write_bit(bc, send_update);
-        if (send_update)
-        {
-            int Data;
+    vp8_write_bit(bc, send_update);
+    if (send_update) {
+      int Data;
 
-            /* Send update */
-            for (i = 0; i < MAX_REF_LF_DELTAS; i++)
-            {
-                Data = xd->ref_lf_deltas[i];
+      /* Send update */
+      for (i = 0; i < MAX_REF_LF_DELTAS; i++) {
+        Data = xd->ref_lf_deltas[i];
 
-                /* Frame level data */
-                if (xd->ref_lf_deltas[i] != xd->last_ref_lf_deltas[i]
-                    || cpi->oxcf.error_resilient_mode)
-                {
-                    xd->last_ref_lf_deltas[i] = xd->ref_lf_deltas[i];
-                    vp8_write_bit(bc, 1);
+        /* Frame level data */
+        if (xd->ref_lf_deltas[i] != xd->last_ref_lf_deltas[i] ||
+            cpi->oxcf.error_resilient_mode) {
+          xd->last_ref_lf_deltas[i] = xd->ref_lf_deltas[i];
+          vp8_write_bit(bc, 1);
 
-                    if (Data > 0)
-                    {
-                        vp8_write_literal(bc, (Data & 0x3F), 6);
-                        vp8_write_bit(bc, 0);    /* sign */
-                    }
-                    else
-                    {
-                        Data = -Data;
-                        vp8_write_literal(bc, (Data & 0x3F), 6);
-                        vp8_write_bit(bc, 1);    /* sign */
-                    }
-                }
-                else
-                    vp8_write_bit(bc, 0);
-            }
+          if (Data > 0) {
+            vp8_write_literal(bc, (Data & 0x3F), 6);
+            vp8_write_bit(bc, 0); /* sign */
+          } else {
+            Data = -Data;
+            vp8_write_literal(bc, (Data & 0x3F), 6);
+            vp8_write_bit(bc, 1); /* sign */
+          }
+        } else
+          vp8_write_bit(bc, 0);
+      }
 
-            /* Send update */
-            for (i = 0; i < MAX_MODE_LF_DELTAS; i++)
-            {
-                Data = xd->mode_lf_deltas[i];
+      /* Send update */
+      for (i = 0; i < MAX_MODE_LF_DELTAS; i++) {
+        Data = xd->mode_lf_deltas[i];
 
-                if (xd->mode_lf_deltas[i] != xd->last_mode_lf_deltas[i]
-                    || cpi->oxcf.error_resilient_mode)
-                {
-                    xd->last_mode_lf_deltas[i] = xd->mode_lf_deltas[i];
-                    vp8_write_bit(bc, 1);
+        if (xd->mode_lf_deltas[i] != xd->last_mode_lf_deltas[i] ||
+            cpi->oxcf.error_resilient_mode) {
+          xd->last_mode_lf_deltas[i] = xd->mode_lf_deltas[i];
+          vp8_write_bit(bc, 1);
 
-                    if (Data > 0)
-                    {
-                        vp8_write_literal(bc, (Data & 0x3F), 6);
-                        vp8_write_bit(bc, 0);    /* sign */
-                    }
-                    else
-                    {
-                        Data = -Data;
-                        vp8_write_literal(bc, (Data & 0x3F), 6);
-                        vp8_write_bit(bc, 1);    /* sign */
-                    }
-                }
-                else
-                    vp8_write_bit(bc, 0);
-            }
-        }
-    }
+          if (Data > 0) {
+            vp8_write_literal(bc, (Data & 0x3F), 6);
+            vp8_write_bit(bc, 0); /* sign */
+          } else {
+            Data = -Data;
+            vp8_write_literal(bc, (Data & 0x3F), 6);
+            vp8_write_bit(bc, 1); /* sign */
+          }
+        } else
+          vp8_write_bit(bc, 0);
+      }
+    }
+  }
 
-    /* signal here is multi token partition is enabled */
-    vp8_write_literal(bc, pc->multi_token_partition, 2);
+  /* signal here is multi token partition is enabled */
+  vp8_write_literal(bc, pc->multi_token_partition, 2);
 
-    /* Frame Qbaseline quantizer index */
-    vp8_write_literal(bc, pc->base_qindex, 7);
+  /* Frame Qbaseline quantizer index */
+  vp8_write_literal(bc, pc->base_qindex, 7);
 
-    /* Transmit Dc, Second order and Uv quantizer delta information */
-    put_delta_q(bc, pc->y1dc_delta_q);
-    put_delta_q(bc, pc->y2dc_delta_q);
-    put_delta_q(bc, pc->y2ac_delta_q);
-    put_delta_q(bc, pc->uvdc_delta_q);
-    put_delta_q(bc, pc->uvac_delta_q);
+  /* Transmit Dc, Second order and Uv quantizer delta information */
+  put_delta_q(bc, pc->y1dc_delta_q);
+  put_delta_q(bc, pc->y2dc_delta_q);
+  put_delta_q(bc, pc->y2ac_delta_q);
+  put_delta_q(bc, pc->uvdc_delta_q);
+  put_delta_q(bc, pc->uvac_delta_q);
 
-    /* When there is a key frame all reference buffers are updated using
-     * the new key frame
+  /* When there is a key frame all reference buffers are updated using
+   * the new key frame
+   */
+  if (pc->frame_type != KEY_FRAME) {
+    /* Should the GF or ARF be updated using the transmitted frame
+     * or buffer
      */
-    if (pc->frame_type != KEY_FRAME)
-    {
-        /* Should the GF or ARF be updated using the transmitted frame
-         * or buffer
-         */
-        vp8_write_bit(bc, pc->refresh_golden_frame);
-        vp8_write_bit(bc, pc->refresh_alt_ref_frame);
+    vp8_write_bit(bc, pc->refresh_golden_frame);
+    vp8_write_bit(bc, pc->refresh_alt_ref_frame);
 
-        /* If not being updated from current frame should either GF or ARF
-         * be updated from another buffer
-         */
-        if (!pc->refresh_golden_frame)
-            vp8_write_literal(bc, pc->copy_buffer_to_gf, 2);
+    /* If not being updated from current frame should either GF or ARF
+     * be updated from another buffer
+     */
+    if (!pc->refresh_golden_frame)
+      vp8_write_literal(bc, pc->copy_buffer_to_gf, 2);
 
-        if (!pc->refresh_alt_ref_frame)
-            vp8_write_literal(bc, pc->copy_buffer_to_arf, 2);
+    if (!pc->refresh_alt_ref_frame)
+      vp8_write_literal(bc, pc->copy_buffer_to_arf, 2);
 
-        /* Indicate reference frame sign bias for Golden and ARF frames
-         * (always 0 for last frame buffer)
-         */
-        vp8_write_bit(bc, pc->ref_frame_sign_bias[GOLDEN_FRAME]);
-        vp8_write_bit(bc, pc->ref_frame_sign_bias[ALTREF_FRAME]);
-    }
+    /* Indicate reference frame sign bias for Golden and ARF frames
+     * (always 0 for last frame buffer)
+     */
+    vp8_write_bit(bc, pc->ref_frame_sign_bias[GOLDEN_FRAME]);
+    vp8_write_bit(bc, pc->ref_frame_sign_bias[ALTREF_FRAME]);
+  }
 
 #if !(CONFIG_REALTIME_ONLY & CONFIG_ONTHEFLY_BITPACKING)
-    if (cpi->oxcf.error_resilient_mode & VPX_ERROR_RESILIENT_PARTITIONS)
-    {
-        if (pc->frame_type == KEY_FRAME)
-            pc->refresh_entropy_probs = 1;
-        else
-            pc->refresh_entropy_probs = 0;
-    }
+  if (cpi->oxcf.error_resilient_mode & VPX_ERROR_RESILIENT_PARTITIONS) {
+    if (pc->frame_type == KEY_FRAME)
+      pc->refresh_entropy_probs = 1;
+    else
+      pc->refresh_entropy_probs = 0;
+  }
 #endif
 
-    vp8_write_bit(bc, pc->refresh_entropy_probs);
+  vp8_write_bit(bc, pc->refresh_entropy_probs);
 
-    if (pc->frame_type != KEY_FRAME)
-        vp8_write_bit(bc, pc->refresh_last_frame);
+  if (pc->frame_type != KEY_FRAME) vp8_write_bit(bc, pc->refresh_last_frame);
 
 #ifdef VP8_ENTROPY_STATS
 
-    if (pc->frame_type == INTER_FRAME)
-        active_section = 0;
-    else
-        active_section = 7;
+  if (pc->frame_type == INTER_FRAME)
+    active_section = 0;
+  else
+    active_section = 7;
 
 #endif
 
-    vp8_clear_system_state();
+  vp8_clear_system_state();
 
 #if CONFIG_REALTIME_ONLY & CONFIG_ONTHEFLY_BITPACKING
-    pack_coef_probs(cpi);
+  pack_coef_probs(cpi);
 #else
-    if (pc->refresh_entropy_probs == 0)
-    {
-        /* save a copy for later refresh */
-        memcpy(&cpi->common.lfc, &cpi->common.fc, sizeof(cpi->common.fc));
-    }
+  if (pc->refresh_entropy_probs == 0) {
+    /* save a copy for later refresh */
+    memcpy(&cpi->common.lfc, &cpi->common.fc, sizeof(cpi->common.fc));
+  }
 
-    vp8_update_coef_probs(cpi);
+  vp8_update_coef_probs(cpi);
 #endif
 
 #ifdef VP8_ENTROPY_STATS
-    active_section = 2;
+  active_section = 2;
 #endif
 
-    /* Write out the mb_no_coeff_skip flag */
-    vp8_write_bit(bc, pc->mb_no_coeff_skip);
+  /* Write out the mb_no_coeff_skip flag */
+  vp8_write_bit(bc, pc->mb_no_coeff_skip);
 
-    if (pc->frame_type == KEY_FRAME)
-    {
-        write_kfmodes(cpi);
+  if (pc->frame_type == KEY_FRAME) {
+    write_kfmodes(cpi);
 
 #ifdef VP8_ENTROPY_STATS
-        active_section = 8;
+    active_section = 8;
 #endif
-    }
-    else
-    {
-        pack_inter_mode_mvs(cpi);
+  } else {
+    pack_inter_mode_mvs(cpi);
 
 #ifdef VP8_ENTROPY_STATS
-        active_section = 1;
+    active_section = 1;
 #endif
-    }
+  }
 
-    vp8_stop_encode(bc);
+  vp8_stop_encode(bc);
 
-    cx_data += bc->pos;
+  cx_data += bc->pos;
 
-    oh.first_partition_length_in_bytes = cpi->bc->pos;
+  oh.first_partition_length_in_bytes = cpi->bc->pos;
 
-    /* update frame tag */
-    {
-        int v = (oh.first_partition_length_in_bytes << 5) |
-                (oh.show_frame << 4) |
-                (oh.version << 1) |
-                oh.type;
+  /* update frame tag */
+  {
+    int v = (oh.first_partition_length_in_bytes << 5) | (oh.show_frame << 4) |
+            (oh.version << 1) | oh.type;
 
-        dest[0] = v;
-        dest[1] = v >> 8;
-        dest[2] = v >> 16;
-    }
+    dest[0] = v;
+    dest[1] = v >> 8;
+    dest[2] = v >> 16;
+  }
 
-    *size = VP8_HEADER_SIZE + extra_bytes_packed + cpi->bc->pos;
+  *size = VP8_HEADER_SIZE + extra_bytes_packed + cpi->bc->pos;
 
-    cpi->partition_sz[0] = *size;
+  cpi->partition_sz[0] = *size;
 
 #if CONFIG_REALTIME_ONLY & CONFIG_ONTHEFLY_BITPACKING
-    {
-        const int num_part = (1 << pc->multi_token_partition);
-        unsigned char * dp = cpi->partition_d[0] + cpi->partition_sz[0];
+  {
+    const int num_part = (1 << pc->multi_token_partition);
+    unsigned char *dp = cpi->partition_d[0] + cpi->partition_sz[0];
 
-        if (num_part > 1)
-        {
-            /* write token part sizes (all but last) if more than 1 */
-            validate_buffer(dp, 3 * (num_part - 1), cpi->partition_d_end[0],
-                            &pc->error);
+    if (num_part > 1) {
+      /* write token part sizes (all but last) if more than 1 */
+      validate_buffer(dp, 3 * (num_part - 1), cpi->partition_d_end[0],
+                      &pc->error);
 
-            cpi->partition_sz[0] += 3*(num_part-1);
+      cpi->partition_sz[0] += 3 * (num_part - 1);
 
-            for(i = 1; i < num_part; i++)
-            {
-                write_partition_size(dp, cpi->partition_sz[i]);
-                dp += 3;
-            }
-        }
+      for (i = 1; i < num_part; i++) {
+        write_partition_size(dp, cpi->partition_sz[i]);
+        dp += 3;
+      }
+    }
 
-        if (!cpi->output_partition)
-        {
-            /* concatenate partition buffers */
-            for(i = 0; i < num_part; i++)
-            {
-                memmove(dp, cpi->partition_d[i+1], cpi->partition_sz[i+1]);
-                cpi->partition_d[i+1] = dp;
-                dp += cpi->partition_sz[i+1];
-            }
-        }
+    if (!cpi->output_partition) {
+      /* concatenate partition buffers */
+      for (i = 0; i < num_part; i++) {
+        memmove(dp, cpi->partition_d[i + 1], cpi->partition_sz[i + 1]);
+        cpi->partition_d[i + 1] = dp;
+        dp += cpi->partition_sz[i + 1];
+      }
+    }
 
-        /* update total size */
-        *size = 0;
-        for(i = 0; i < num_part+1; i++)
-        {
-            *size += cpi->partition_sz[i];
-        }
+    /* update total size */
+    *size = 0;
+    for (i = 0; i < num_part + 1; i++) {
+      *size += cpi->partition_sz[i];
     }
+  }
 #else
-    if (pc->multi_token_partition != ONE_PARTITION)
-    {
-        int num_part = 1 << pc->multi_token_partition;
+  if (pc->multi_token_partition != ONE_PARTITION) {
+    int num_part = 1 << pc->multi_token_partition;
 
-        /* partition size table at the end of first partition */
-        cpi->partition_sz[0] += 3 * (num_part - 1);
-        *size += 3 * (num_part - 1);
+    /* partition size table at the end of first partition */
+    cpi->partition_sz[0] += 3 * (num_part - 1);
+    *size += 3 * (num_part - 1);
 
-        validate_buffer(cx_data, 3 * (num_part - 1), cx_data_end,
-                        &pc->error);
+    validate_buffer(cx_data, 3 * (num_part - 1), cx_data_end, &pc->error);
 
-        for(i = 1; i < num_part + 1; i++)
-        {
-            cpi->bc[i].error = &pc->error;
-        }
+    for (i = 1; i < num_part + 1; i++) {
+      cpi->bc[i].error = &pc->error;
+    }
 
-        pack_tokens_into_partitions(cpi, cx_data + 3 * (num_part - 1),
-                                    cx_data_end, num_part);
+    pack_tokens_into_partitions(cpi, cx_data + 3 * (num_part - 1), cx_data_end,
+                                num_part);
 
-        for(i = 1; i < num_part; i++)
-        {
-            cpi->partition_sz[i] = cpi->bc[i].pos;
-            write_partition_size(cx_data, cpi->partition_sz[i]);
-            cx_data += 3;
-            *size += cpi->partition_sz[i]; /* add to total */
-        }
-
-        /* add last partition to total size */
-        cpi->partition_sz[i] = cpi->bc[i].pos;
-        *size += cpi->partition_sz[i];
+    for (i = 1; i < num_part; i++) {
+      cpi->partition_sz[i] = cpi->bc[i].pos;
+      write_partition_size(cx_data, cpi->partition_sz[i]);
+      cx_data += 3;
+      *size += cpi->partition_sz[i]; /* add to total */
     }
-    else
-    {
-        bc[1].error = &pc->error;
 
-        vp8_start_encode(&cpi->bc[1], cx_data, cx_data_end);
+    /* add last partition to total size */
+    cpi->partition_sz[i] = cpi->bc[i].pos;
+    *size += cpi->partition_sz[i];
+  } else {
+    bc[1].error = &pc->error;
 
+    vp8_start_encode(&cpi->bc[1], cx_data, cx_data_end);
+
 #if CONFIG_MULTITHREAD
-        if (cpi->b_multi_threaded)
-            pack_mb_row_tokens(cpi, &cpi->bc[1]);
-        else
+    if (cpi->b_multi_threaded)
+      pack_mb_row_tokens(cpi, &cpi->bc[1]);
+    else
 #endif  // CONFIG_MULTITHREAD
-            vp8_pack_tokens(&cpi->bc[1], cpi->tok, cpi->tok_count);
+      vp8_pack_tokens(&cpi->bc[1], cpi->tok, cpi->tok_count);
 
-        vp8_stop_encode(&cpi->bc[1]);
+    vp8_stop_encode(&cpi->bc[1]);
 
-        *size += cpi->bc[1].pos;
-        cpi->partition_sz[1] = cpi->bc[1].pos;
-    }
+    *size += cpi->bc[1].pos;
+    cpi->partition_sz[1] = cpi->bc[1].pos;
+  }
 #endif
 }
 
 #ifdef VP8_ENTROPY_STATS
-void print_tree_update_probs()
-{
-    int i, j, k, l;
-    FILE *f = fopen("context.c", "a");
-    int Sum;
-    fprintf(f, "\n/* Update probabilities for token entropy tree. */\n\n");
-    fprintf(f, "const vp8_prob tree_update_probs[BLOCK_TYPES] [COEF_BANDS] [PREV_COEF_CONTEXTS] [ENTROPY_NODES] = {\n");
+void print_tree_update_probs() {
+  int i, j, k, l;
+  FILE *f = fopen("context.c", "a");
+  int Sum;
+  fprintf(f, "\n/* Update probabilities for token entropy tree. */\n\n");
+  fprintf(f,
+          "const vp8_prob tree_update_probs[BLOCK_TYPES] [COEF_BANDS] "
+          "[PREV_COEF_CONTEXTS] [ENTROPY_NODES] = {\n");
 
-    for (i = 0; i < BLOCK_TYPES; i++)
-    {
-        fprintf(f, "  { \n");
+  for (i = 0; i < BLOCK_TYPES; i++) {
+    fprintf(f, "  { \n");
 
-        for (j = 0; j < COEF_BANDS; j++)
-        {
-            fprintf(f, "    {\n");
+    for (j = 0; j < COEF_BANDS; j++) {
+      fprintf(f, "    {\n");
 
-            for (k = 0; k < PREV_COEF_CONTEXTS; k++)
-            {
-                fprintf(f, "      {");
+      for (k = 0; k < PREV_COEF_CONTEXTS; k++) {
+        fprintf(f, "      {");
 
-                for (l = 0; l < ENTROPY_NODES; l++)
-                {
-                    Sum = tree_update_hist[i][j][k][l][0] + tree_update_hist[i][j][k][l][1];
+        for (l = 0; l < ENTROPY_NODES; l++) {
+          Sum =
+              tree_update_hist[i][j][k][l][0] + tree_update_hist[i][j][k][l][1];
 
-                    if (Sum > 0)
-                    {
-                        if (((tree_update_hist[i][j][k][l][0] * 255) / Sum) > 0)
-                            fprintf(f, "%3ld, ", (tree_update_hist[i][j][k][l][0] * 255) / Sum);
-                        else
-                            fprintf(f, "%3ld, ", 1);
-                    }
-                    else
-                        fprintf(f, "%3ld, ", 128);
-                }
-
-                fprintf(f, "},\n");
-            }
-
-            fprintf(f, "    },\n");
+          if (Sum > 0) {
+            if (((tree_update_hist[i][j][k][l][0] * 255) / Sum) > 0)
+              fprintf(f, "%3ld, ",
+                      (tree_update_hist[i][j][k][l][0] * 255) / Sum);
+            else
+              fprintf(f, "%3ld, ", 1);
+          } else
+            fprintf(f, "%3ld, ", 128);
         }
 
-        fprintf(f, "  },\n");
+        fprintf(f, "},\n");
+      }
+
+      fprintf(f, "    },\n");
     }
 
-    fprintf(f, "};\n");
-    fclose(f);
+    fprintf(f, "  },\n");
+  }
+
+  fprintf(f, "};\n");
+  fclose(f);
 }
 #endif
--- a/vp8/encoder/bitstream.h
+++ b/vp8/encoder/bitstream.h
@@ -8,7 +8,6 @@
  *  be found in the AUTHORS file in the root of the source tree.
  */
 
-
 #ifndef VP8_ENCODER_BITSTREAM_H_
 #define VP8_ENCODER_BITSTREAM_H_
 
--- a/vp8/encoder/block.h
+++ b/vp8/encoder/block.h
@@ -8,7 +8,6 @@
  *  be found in the AUTHORS file in the root of the source tree.
  */
 
-
 #ifndef VP8_ENCODER_BLOCK_H_
 #define VP8_ENCODER_BLOCK_H_
 
@@ -26,147 +25,141 @@
 #define MAX_ERROR_BINS 1024
 
 /* motion search site */
-typedef struct
-{
-    MV mv;
-    int offset;
+typedef struct {
+  MV mv;
+  int offset;
 } search_site;
 
-typedef struct block
-{
-    /* 16 Y blocks, 4 U blocks, 4 V blocks each with 16 entries */
-    short *src_diff;
-    short *coeff;
+typedef struct block {
+  /* 16 Y blocks, 4 U blocks, 4 V blocks each with 16 entries */
+  short *src_diff;
+  short *coeff;
 
-    /* 16 Y blocks, 4 U blocks, 4 V blocks each with 16 entries */
-    short *quant;
-    short *quant_fast;
-    short *quant_shift;
-    short *zbin;
-    short *zrun_zbin_boost;
-    short *round;
+  /* 16 Y blocks, 4 U blocks, 4 V blocks each with 16 entries */
+  short *quant;
+  short *quant_fast;
+  short *quant_shift;
+  short *zbin;
+  short *zrun_zbin_boost;
+  short *round;
 
-    /* Zbin Over Quant value */
-    short zbin_extra;
+  /* Zbin Over Quant value */
+  short zbin_extra;
 
-    unsigned char **base_src;
-    int src;
-    int src_stride;
+  unsigned char **base_src;
+  int src;
+  int src_stride;
 } BLOCK;
 
-typedef struct
-{
-    int count;
-    struct
-    {
-        B_PREDICTION_MODE mode;
-        int_mv mv;
-    } bmi[16];
+typedef struct {
+  int count;
+  struct {
+    B_PREDICTION_MODE mode;
+    int_mv mv;
+  } bmi[16];
 } PARTITION_INFO;
 
-typedef struct macroblock
-{
-    DECLARE_ALIGNED(16, short, src_diff[400]); /* 25 blocks Y,U,V,Y2 */
-    DECLARE_ALIGNED(16, short, coeff[400]); /* 25 blocks Y,U,V,Y2 */
-    DECLARE_ALIGNED(16, unsigned char, thismb[256]);
+typedef struct macroblock {
+  DECLARE_ALIGNED(16, short, src_diff[400]); /* 25 blocks Y,U,V,Y2 */
+  DECLARE_ALIGNED(16, short, coeff[400]);    /* 25 blocks Y,U,V,Y2 */
+  DECLARE_ALIGNED(16, unsigned char, thismb[256]);
 
-    unsigned char *thismb_ptr;
-    /* 16 Y, 4 U, 4 V, 1 DC 2nd order block */
-    BLOCK block[25];
+  unsigned char *thismb_ptr;
+  /* 16 Y, 4 U, 4 V, 1 DC 2nd order block */
+  BLOCK block[25];
 
-    YV12_BUFFER_CONFIG src;
+  YV12_BUFFER_CONFIG src;
 
-    MACROBLOCKD e_mbd;
-    PARTITION_INFO *partition_info; /* work pointer */
-    PARTITION_INFO *pi;   /* Corresponds to upper left visible macroblock */
-    PARTITION_INFO *pip;  /* Base of allocated array */
+  MACROBLOCKD e_mbd;
+  PARTITION_INFO *partition_info; /* work pointer */
+  PARTITION_INFO *pi;  /* Corresponds to upper left visible macroblock */
+  PARTITION_INFO *pip; /* Base of allocated array */
 
-    int ref_frame_cost[MAX_REF_FRAMES];
+  int ref_frame_cost[MAX_REF_FRAMES];
 
-    search_site *ss;
-    int ss_count;
-    int searches_per_step;
+  search_site *ss;
+  int ss_count;
+  int searches_per_step;
 
-    int errorperbit;
-    int sadperbit16;
-    int sadperbit4;
-    int rddiv;
-    int rdmult;
-    unsigned int * mb_activity_ptr;
-    int * mb_norm_activity_ptr;
-    signed int act_zbin_adj;
-    signed int last_act_zbin_adj;
+  int errorperbit;
+  int sadperbit16;
+  int sadperbit4;
+  int rddiv;
+  int rdmult;
+  unsigned int *mb_activity_ptr;
+  int *mb_norm_activity_ptr;
+  signed int act_zbin_adj;
+  signed int last_act_zbin_adj;
 
-    int *mvcost[2];
-    int *mvsadcost[2];
-    int (*mbmode_cost)[MB_MODE_COUNT];
-    int (*intra_uv_mode_cost)[MB_MODE_COUNT];
-    int (*bmode_costs)[10][10];
-    int *inter_bmode_costs;
-    int (*token_costs)[COEF_BANDS][PREV_COEF_CONTEXTS]
-    [MAX_ENTROPY_TOKENS];
+  int *mvcost[2];
+  int *mvsadcost[2];
+  int (*mbmode_cost)[MB_MODE_COUNT];
+  int (*intra_uv_mode_cost)[MB_MODE_COUNT];
+  int (*bmode_costs)[10][10];
+  int *inter_bmode_costs;
+  int (*token_costs)[COEF_BANDS][PREV_COEF_CONTEXTS][MAX_ENTROPY_TOKENS];
 
-    /* These define limits to motion vector components to prevent
-     * them from extending outside the UMV borders.
-     */
-    int mv_col_min;
-    int mv_col_max;
-    int mv_row_min;
-    int mv_row_max;
+  /* These define limits to motion vector components to prevent
+   * them from extending outside the UMV borders.
+   */
+  int mv_col_min;
+  int mv_col_max;
+  int mv_row_min;
+  int mv_row_max;
 
-    int skip;
+  int skip;
 
-    unsigned int encode_breakout;
+  unsigned int encode_breakout;
 
-    signed char *gf_active_ptr;
+  signed char *gf_active_ptr;
 
-    unsigned char *active_ptr;
-    MV_CONTEXT *mvc;
+  unsigned char *active_ptr;
+  MV_CONTEXT *mvc;
 
-    int optimize;
-    int q_index;
-    int is_skin;
-    int denoise_zeromv;
+  int optimize;
+  int q_index;
+  int is_skin;
+  int denoise_zeromv;
 
 #if CONFIG_TEMPORAL_DENOISING
-    int increase_denoising;
-    MB_PREDICTION_MODE best_sse_inter_mode;
-    int_mv best_sse_mv;
-    MV_REFERENCE_FRAME best_reference_frame;
-    MV_REFERENCE_FRAME best_zeromv_reference_frame;
-    unsigned char need_to_clamp_best_mvs;
+  int increase_denoising;
+  MB_PREDICTION_MODE best_sse_inter_mode;
+  int_mv best_sse_mv;
+  MV_REFERENCE_FRAME best_reference_frame;
+  MV_REFERENCE_FRAME best_zeromv_reference_frame;
+  unsigned char need_to_clamp_best_mvs;
 #endif
 
-    int skip_true_count;
-    unsigned int coef_counts [BLOCK_TYPES] [COEF_BANDS] [PREV_COEF_CONTEXTS] [MAX_ENTROPY_TOKENS];
-    unsigned int MVcount [2] [MVvals];  /* (row,col) MV cts this frame */
-    int ymode_count [VP8_YMODES];        /* intra MB type cts this frame */
-    int uv_mode_count[VP8_UV_MODES];     /* intra MB type cts this frame */
-    int64_t prediction_error;
-    int64_t intra_error;
-    int count_mb_ref_frame_usage[MAX_REF_FRAMES];
+  int skip_true_count;
+  unsigned int coef_counts[BLOCK_TYPES][COEF_BANDS][PREV_COEF_CONTEXTS]
+                          [MAX_ENTROPY_TOKENS];
+  unsigned int MVcount[2][MVvals]; /* (row,col) MV cts this frame */
+  int ymode_count[VP8_YMODES];     /* intra MB type cts this frame */
+  int uv_mode_count[VP8_UV_MODES]; /* intra MB type cts this frame */
+  int64_t prediction_error;
+  int64_t intra_error;
+  int count_mb_ref_frame_usage[MAX_REF_FRAMES];
 
-    int rd_thresh_mult[MAX_MODES];
-    int rd_threshes[MAX_MODES];
-    unsigned int mbs_tested_so_far;
-    unsigned int mode_test_hit_counts[MAX_MODES];
-    int zbin_mode_boost_enabled;
-    int zbin_mode_boost;
-    int last_zbin_mode_boost;
+  int rd_thresh_mult[MAX_MODES];
+  int rd_threshes[MAX_MODES];
+  unsigned int mbs_tested_so_far;
+  unsigned int mode_test_hit_counts[MAX_MODES];
+  int zbin_mode_boost_enabled;
+  int zbin_mode_boost;
+  int last_zbin_mode_boost;
 
-    int last_zbin_over_quant;
-    int zbin_over_quant;
-    int error_bins[MAX_ERROR_BINS];
+  int last_zbin_over_quant;
+  int zbin_over_quant;
+  int error_bins[MAX_ERROR_BINS];
 
-    void (*short_fdct4x4)(short *input, short *output, int pitch);
-    void (*short_fdct8x4)(short *input, short *output, int pitch);
-    void (*short_walsh4x4)(short *input, short *output, int pitch);
-    void (*quantize_b)(BLOCK *b, BLOCKD *d);
+  void (*short_fdct4x4)(short *input, short *output, int pitch);
+  void (*short_fdct8x4)(short *input, short *output, int pitch);
+  void (*short_walsh4x4)(short *input, short *output, int pitch);
+  void (*quantize_b)(BLOCK *b, BLOCKD *d);
 
-    unsigned int mbs_zero_last_dot_suppress;
-    int zero_last_dot_suppress;
+  unsigned int mbs_zero_last_dot_suppress;
+  int zero_last_dot_suppress;
 } MACROBLOCK;
-
 
 #ifdef __cplusplus
 }  // extern "C"
--- a/vp8/encoder/boolhuff.c
+++ b/vp8/encoder/boolhuff.c
@@ -8,7 +8,6 @@
  *  be found in the AUTHORS file in the root of the source tree.
  */
 
-
 #include "boolhuff.h"
 
 #if defined(SECTIONBITS_OUTPUT)
@@ -20,51 +19,48 @@
 unsigned int active_section = 0;
 #endif
 
-const unsigned int vp8_prob_cost[256] =
-{
-    2047, 2047, 1791, 1641, 1535, 1452, 1385, 1328, 1279, 1235, 1196, 1161, 1129, 1099, 1072, 1046,
-    1023, 1000,  979,  959,  940,  922,  905,  889,  873,  858,  843,  829,  816,  803,  790,  778,
-    767,  755,  744,  733,  723,  713,  703,  693,  684,  675,  666,  657,  649,  641,  633,  625,
-    617,  609,  602,  594,  587,  580,  573,  567,  560,  553,  547,  541,  534,  528,  522,  516,
-    511,  505,  499,  494,  488,  483,  477,  472,  467,  462,  457,  452,  447,  442,  437,  433,
-    428,  424,  419,  415,  410,  406,  401,  397,  393,  389,  385,  381,  377,  373,  369,  365,
-    361,  357,  353,  349,  346,  342,  338,  335,  331,  328,  324,  321,  317,  314,  311,  307,
-    304,  301,  297,  294,  291,  288,  285,  281,  278,  275,  272,  269,  266,  263,  260,  257,
-    255,  252,  249,  246,  243,  240,  238,  235,  232,  229,  227,  224,  221,  219,  216,  214,
-    211,  208,  206,  203,  201,  198,  196,  194,  191,  189,  186,  184,  181,  179,  177,  174,
-    172,  170,  168,  165,  163,  161,  159,  156,  154,  152,  150,  148,  145,  143,  141,  139,
-    137,  135,  133,  131,  129,  127,  125,  123,  121,  119,  117,  115,  113,  111,  109,  107,
-    105,  103,  101,   99,   97,   95,   93,   92,   90,   88,   86,   84,   82,   81,   79,   77,
-    75,   73,   72,   70,   68,   66,   65,   63,   61,   60,   58,   56,   55,   53,   51,   50,
-    48,   46,   45,   43,   41,   40,   38,   37,   35,   33,   32,   30,   29,   27,   25,   24,
-    22,   21,   19,   18,   16,   15,   13,   12,   10,    9,    7,    6,    4,    3,    1,   1
+const unsigned int vp8_prob_cost[256] = {
+  2047, 2047, 1791, 1641, 1535, 1452, 1385, 1328, 1279, 1235, 1196, 1161, 1129,
+  1099, 1072, 1046, 1023, 1000, 979,  959,  940,  922,  905,  889,  873,  858,
+  843,  829,  816,  803,  790,  778,  767,  755,  744,  733,  723,  713,  703,
+  693,  684,  675,  666,  657,  649,  641,  633,  625,  617,  609,  602,  594,
+  587,  580,  573,  567,  560,  553,  547,  541,  534,  528,  522,  516,  511,
+  505,  499,  494,  488,  483,  477,  472,  467,  462,  457,  452,  447,  442,
+  437,  433,  428,  424,  419,  415,  410,  406,  401,  397,  393,  389,  385,
+  381,  377,  373,  369,  365,  361,  357,  353,  349,  346,  342,  338,  335,
+  331,  328,  324,  321,  317,  314,  311,  307,  304,  301,  297,  294,  291,
+  288,  285,  281,  278,  275,  272,  269,  266,  263,  260,  257,  255,  252,
+  249,  246,  243,  240,  238,  235,  232,  229,  227,  224,  221,  219,  216,
+  214,  211,  208,  206,  203,  201,  198,  196,  194,  191,  189,  186,  184,
+  181,  179,  177,  174,  172,  170,  168,  165,  163,  161,  159,  156,  154,
+  152,  150,  148,  145,  143,  141,  139,  137,  135,  133,  131,  129,  127,
+  125,  123,  121,  119,  117,  115,  113,  111,  109,  107,  105,  103,  101,
+  99,   97,   95,   93,   92,   90,   88,   86,   84,   82,   81,   79,   77,
+  75,   73,   72,   70,   68,   66,   65,   63,   61,   60,   58,   56,   55,
+  53,   51,   50,   48,   46,   45,   43,   41,   40,   38,   37,   35,   33,
+  32,   30,   29,   27,   25,   24,   22,   21,   19,   18,   16,   15,   13,
+  12,   10,   9,    7,    6,    4,    3,    1,    1
 };
 
-void vp8_start_encode(BOOL_CODER *br, unsigned char *source, unsigned char *source_end)
-{
-
-    br->lowvalue   = 0;
-    br->range      = 255;
-    br->count      = -24;
-    br->buffer     = source;
-    br->buffer_end = source_end;
-    br->pos        = 0;
+void vp8_start_encode(BOOL_CODER *br, unsigned char *source,
+                      unsigned char *source_end) {
+  br->lowvalue = 0;
+  br->range = 255;
+  br->count = -24;
+  br->buffer = source;
+  br->buffer_end = source_end;
+  br->pos = 0;
 }
 
-void vp8_stop_encode(BOOL_CODER *br)
-{
-    int i;
+void vp8_stop_encode(BOOL_CODER *br) {
+  int i;
 
-    for (i = 0; i < 32; i++)
-        vp8_encode_bool(br, 0, 128);
+  for (i = 0; i < 32; i++) vp8_encode_bool(br, 0, 128);
 }
 
+void vp8_encode_value(BOOL_CODER *br, int data, int bits) {
+  int bit;
 
-void vp8_encode_value(BOOL_CODER *br, int data, int bits)
-{
-    int bit;
-
-    for (bit = bits - 1; bit >= 0; bit--)
-        vp8_encode_bool(br, (1 & (data >> bit)), 0x80);
-
+  for (bit = bits - 1; bit >= 0; bit--)
+    vp8_encode_bool(br, (1 & (data >> bit)), 0x80);
 }
--- a/vp8/encoder/boolhuff.h
+++ b/vp8/encoder/boolhuff.h
@@ -8,7 +8,6 @@
  *  be found in the AUTHORS file in the root of the source tree.
  */
 
-
 /****************************************************************************
 *
 *   Module Title :     boolhuff.h
@@ -26,103 +25,95 @@
 extern "C" {
 #endif
 
-typedef struct
-{
-    unsigned int lowvalue;
-    unsigned int range;
-    int count;
-    unsigned int pos;
-    unsigned char *buffer;
-    unsigned char *buffer_end;
-    struct vpx_internal_error_info *error;
+typedef struct {
+  unsigned int lowvalue;
+  unsigned int range;
+  int count;
+  unsigned int pos;
+  unsigned char *buffer;
+  unsigned char *buffer_end;
+  struct vpx_internal_error_info *error;
 } BOOL_CODER;
 
-extern void vp8_start_encode(BOOL_CODER *bc, unsigned char *buffer, unsigned char *buffer_end);
+extern void vp8_start_encode(BOOL_CODER *bc, unsigned char *buffer,
+                             unsigned char *buffer_end);
 
 extern void vp8_encode_value(BOOL_CODER *br, int data, int bits);
 extern void vp8_stop_encode(BOOL_CODER *bc);
 extern const unsigned int vp8_prob_cost[256];
 
-
 DECLARE_ALIGNED(16, extern const unsigned char, vp8_norm[256]);
 
-static int validate_buffer(const unsigned char *start,
-                           size_t               len,
+static int validate_buffer(const unsigned char *start, size_t len,
                            const unsigned char *end,
-                           struct vpx_internal_error_info *error)
-{
-    if (start + len > start && start + len < end)
-        return 1;
-    else
-        vpx_internal_error(error, VPX_CODEC_CORRUPT_FRAME,
-            "Truncated packet or corrupt partition ");
+                           struct vpx_internal_error_info *error) {
+  if (start + len > start && start + len < end)
+    return 1;
+  else
+    vpx_internal_error(error, VPX_CODEC_CORRUPT_FRAME,
+                       "Truncated packet or corrupt partition ");
 
-    return 0;
+  return 0;
 }
-static void vp8_encode_bool(BOOL_CODER *br, int bit, int probability)
-{
-    unsigned int split;
-    int count = br->count;
-    unsigned int range = br->range;
-    unsigned int lowvalue = br->lowvalue;
-    register int shift;
+static void vp8_encode_bool(BOOL_CODER *br, int bit, int probability) {
+  unsigned int split;
+  int count = br->count;
+  unsigned int range = br->range;
+  unsigned int lowvalue = br->lowvalue;
+  register int shift;
 
 #ifdef VP8_ENTROPY_STATS
 #if defined(SECTIONBITS_OUTPUT)
 
-    if (bit)
-        Sectionbits[active_section] += vp8_prob_cost[255-probability];
-    else
-        Sectionbits[active_section] += vp8_prob_cost[probability];
+  if (bit)
+    Sectionbits[active_section] += vp8_prob_cost[255 - probability];
+  else
+    Sectionbits[active_section] += vp8_prob_cost[probability];
 
 #endif
 #endif
 
-    split = 1 + (((range - 1) * probability) >> 8);
+  split = 1 + (((range - 1) * probability) >> 8);
 
-    range = split;
+  range = split;
 
-    if (bit)
-    {
-        lowvalue += split;
-        range = br->range - split;
-    }
+  if (bit) {
+    lowvalue += split;
+    range = br->range - split;
+  }
 
-    shift = vp8_norm[range];
+  shift = vp8_norm[range];
 
-    range <<= shift;
-    count += shift;
+  range <<= shift;
+  count += shift;
 
-    if (count >= 0)
-    {
-        int offset = shift - count;
+  if (count >= 0) {
+    int offset = shift - count;
 
-        if ((lowvalue << (offset - 1)) & 0x80000000)
-        {
-            int x = br->pos - 1;
+    if ((lowvalue << (offset - 1)) & 0x80000000) {
+      int x = br->pos - 1;
 
-            while (x >= 0 && br->buffer[x] == 0xff)
-            {
-                br->buffer[x] = (unsigned char)0;
-                x--;
-            }
+      while (x >= 0 && br->buffer[x] == 0xff) {
+        br->buffer[x] = (unsigned char)0;
+        x--;
+      }
 
-            br->buffer[x] += 1;
-        }
+      br->buffer[x] += 1;
+    }
 
-        validate_buffer(br->buffer + br->pos, 1, br->buffer_end, br->error);
-        br->buffer[br->pos++] = (lowvalue >> (24 - offset));
+    validate_buffer(br->buffer + br->pos, 1, br->buffer_end, br->error);
+    br->buffer[br->pos++] = (lowvalue >> (24 - offset));
 
-        lowvalue <<= offset;
-        shift = count;
-        lowvalue &= 0xffffff;
-        count -= 8 ;
-    }
+    lowvalue <<= offset;
+    shift = count;
+    lowvalue &= 0xffffff;
+    count -= 8;
+  }
 
-    lowvalue <<= shift;
-    br->count = count;
-    br->lowvalue = lowvalue;
-    br->range = range;
+  lowvalue <<= shift;
+  br->count = count;
+  br->lowvalue = lowvalue;
+  br->range = range;
 }
 
 #ifdef __cplusplus
--- a/vp8/encoder/dct.c
+++ b/vp8/encoder/dct.c
@@ -8,111 +8,101 @@
  *  be found in the AUTHORS file in the root of the source tree.
  */
 
-
 #include <math.h>
 
 #include "./vp8_rtcd.h"
 
-void vp8_short_fdct4x4_c(short *input, short *output, int pitch)
-{
-    int i;
-    int a1, b1, c1, d1;
-    short *ip = input;
-    short *op = output;
+void vp8_short_fdct4x4_c(short *input, short *output, int pitch) {
+  int i;
+  int a1, b1, c1, d1;
+  short *ip = input;
+  short *op = output;
 
-    for (i = 0; i < 4; i++)
-    {
-        a1 = ((ip[0] + ip[3]) * 8);
-        b1 = ((ip[1] + ip[2]) * 8);
-        c1 = ((ip[1] - ip[2]) * 8);
-        d1 = ((ip[0] - ip[3]) * 8);
+  for (i = 0; i < 4; i++) {
+    a1 = ((ip[0] + ip[3]) * 8);
+    b1 = ((ip[1] + ip[2]) * 8);
+    c1 = ((ip[1] - ip[2]) * 8);
+    d1 = ((ip[0] - ip[3]) * 8);
 
-        op[0] = a1 + b1;
-        op[2] = a1 - b1;
+    op[0] = a1 + b1;
+    op[2] = a1 - b1;
 
-        op[1] = (c1 * 2217 + d1 * 5352 +  14500)>>12;
-        op[3] = (d1 * 2217 - c1 * 5352 +   7500)>>12;
+    op[1] = (c1 * 2217 + d1 * 5352 + 14500) >> 12;
+    op[3] = (d1 * 2217 - c1 * 5352 + 7500) >> 12;
 
-        ip += pitch / 2;
-        op += 4;
+    ip += pitch / 2;
+    op += 4;
+  }
+  ip = output;
+  op = output;
+  for (i = 0; i < 4; i++) {
+    a1 = ip[0] + ip[12];
+    b1 = ip[4] + ip[8];
+    c1 = ip[4] - ip[8];
+    d1 = ip[0] - ip[12];
 
-    }
-    ip = output;
-    op = output;
-    for (i = 0; i < 4; i++)
-    {
-        a1 = ip[0] + ip[12];
-        b1 = ip[4] + ip[8];
-        c1 = ip[4] - ip[8];
-        d1 = ip[0] - ip[12];
+    op[0] = (a1 + b1 + 7) >> 4;
+    op[8] = (a1 - b1 + 7) >> 4;
 
-        op[0]  = ( a1 + b1 + 7)>>4;
-        op[8]  = ( a1 - b1 + 7)>>4;
+    op[4] = ((c1 * 2217 + d1 * 5352 + 12000) >> 16) + (d1 != 0);
+    op[12] = (d1 * 2217 - c1 * 5352 + 51000) >> 16;
 
-        op[4]  =((c1 * 2217 + d1 * 5352 +  12000)>>16) + (d1!=0);
-        op[12] = (d1 * 2217 - c1 * 5352 +  51000)>>16;
-
-        ip++;
-        op++;
-    }
+    ip++;
+    op++;
+  }
 }
 
-void vp8_short_fdct8x4_c(short *input, short *output, int pitch)
-{
-    vp8_short_fdct4x4_c(input,   output,    pitch);
-    vp8_short_fdct4x4_c(input + 4, output + 16, pitch);
+void vp8_short_fdct8x4_c(short *input, short *output, int pitch) {
+  vp8_short_fdct4x4_c(input, output, pitch);
+  vp8_short_fdct4x4_c(input + 4, output + 16, pitch);
 }
 
-void vp8_short_walsh4x4_c(short *input, short *output, int pitch)
-{
-    int i;
-    int a1, b1, c1, d1;
-    int a2, b2, c2, d2;
-    short *ip = input;
-    short *op = output;
+void vp8_short_walsh4x4_c(short *input, short *output, int pitch) {
+  int i;
+  int a1, b1, c1, d1;
+  int a2, b2, c2, d2;
+  short *ip = input;
+  short *op = output;
 
+  for (i = 0; i < 4; i++) {
+    a1 = ((ip[0] + ip[2]) * 4);
+    d1 = ((ip[1] + ip[3]) * 4);
+    c1 = ((ip[1] - ip[3]) * 4);
+    b1 = ((ip[0] - ip[2]) * 4);
 
-    for (i = 0; i < 4; i++)
-    {
-        a1 = ((ip[0] + ip[2]) * 4);
-        d1 = ((ip[1] + ip[3]) * 4);
-        c1 = ((ip[1] - ip[3]) * 4);
-        b1 = ((ip[0] - ip[2]) * 4);
+    op[0] = a1 + d1 + (a1 != 0);
+    op[1] = b1 + c1;
+    op[2] = b1 - c1;
+    op[3] = a1 - d1;
+    ip += pitch / 2;
+    op += 4;
+  }
 
-        op[0] = a1 + d1 + (a1!=0);
-        op[1] = b1 + c1;
-        op[2] = b1 - c1;
-        op[3] = a1 - d1;
-        ip += pitch / 2;
-        op += 4;
-    }
+  ip = output;
+  op = output;
 
-    ip = output;
-    op = output;
+  for (i = 0; i < 4; i++) {
+    a1 = ip[0] + ip[8];
+    d1 = ip[4] + ip[12];
+    c1 = ip[4] - ip[12];
+    b1 = ip[0] - ip[8];
 
-    for (i = 0; i < 4; i++)
-    {
-        a1 = ip[0] + ip[8];
-        d1 = ip[4] + ip[12];
-        c1 = ip[4] - ip[12];
-        b1 = ip[0] - ip[8];
+    a2 = a1 + d1;
+    b2 = b1 + c1;
+    c2 = b1 - c1;
+    d2 = a1 - d1;
 
-        a2 = a1 + d1;
-        b2 = b1 + c1;
-        c2 = b1 - c1;
-        d2 = a1 - d1;
+    a2 += a2 < 0;
+    b2 += b2 < 0;
+    c2 += c2 < 0;
+    d2 += d2 < 0;
 
-        a2 += a2<0;
-        b2 += b2<0;
-        c2 += c2<0;
-        d2 += d2<0;
+    op[0] = (a2 + 3) >> 3;
+    op[4] = (b2 + 3) >> 3;
+    op[8] = (c2 + 3) >> 3;
+    op[12] = (d2 + 3) >> 3;
 
-        op[0] = (a2+3) >> 3;
-        op[4] = (b2+3) >> 3;
-        op[8] = (c2+3) >> 3;
-        op[12]= (d2+3) >> 3;
-
-        ip++;
-        op++;
-    }
+    ip++;
+    op++;
+  }
 }
--- a/vp8/encoder/dct_value_cost.h
+++ b/vp8/encoder/dct_value_cost.h
@@ -18,350 +18,323 @@
 /* Generated file, included by tokenize.c  */
 /* Values generated by fill_value_tokens() */
 
-static const short dct_value_cost[2048*2] =
-{
-    8285, 8277, 8267, 8259, 8253, 8245, 8226, 8218, 8212, 8204, 8194, 8186,
-    8180, 8172, 8150, 8142, 8136, 8128, 8118, 8110, 8104, 8096, 8077, 8069,
-    8063, 8055, 8045, 8037, 8031, 8023, 7997, 7989, 7983, 7975, 7965, 7957,
-    7951, 7943, 7924, 7916, 7910, 7902, 7892, 7884, 7878, 7870, 7848, 7840,
-    7834, 7826, 7816, 7808, 7802, 7794, 7775, 7767, 7761, 7753, 7743, 7735,
-    7729, 7721, 7923, 7915, 7909, 7901, 7891, 7883, 7877, 7869, 7850, 7842,
-    7836, 7828, 7818, 7810, 7804, 7796, 7774, 7766, 7760, 7752, 7742, 7734,
-    7728, 7720, 7701, 7693, 7687, 7679, 7669, 7661, 7655, 7647, 7621, 7613,
-    7607, 7599, 7589, 7581, 7575, 7567, 7548, 7540, 7534, 7526, 7516, 7508,
-    7502, 7494, 7472, 7464, 7458, 7450, 7440, 7432, 7426, 7418, 7399, 7391,
-    7385, 7377, 7367, 7359, 7353, 7345, 7479, 7471, 7465, 7457, 7447, 7439,
-    7433, 7425, 7406, 7398, 7392, 7384, 7374, 7366, 7360, 7352, 7330, 7322,
-    7316, 7308, 7298, 7290, 7284, 7276, 7257, 7249, 7243, 7235, 7225, 7217,
-    7211, 7203, 7177, 7169, 7163, 7155, 7145, 7137, 7131, 7123, 7104, 7096,
-    7090, 7082, 7072, 7064, 7058, 7050, 7028, 7020, 7014, 7006, 6996, 6988,
-    6982, 6974, 6955, 6947, 6941, 6933, 6923, 6915, 6909, 6901, 7632, 7624,
-    7618, 7610, 7600, 7592, 7586, 7578, 7559, 7551, 7545, 7537, 7527, 7519,
-    7513, 7505, 7483, 7475, 7469, 7461, 7451, 7443, 7437, 7429, 7410, 7402,
-    7396, 7388, 7378, 7370, 7364, 7356, 7330, 7322, 7316, 7308, 7298, 7290,
-    7284, 7276, 7257, 7249, 7243, 7235, 7225, 7217, 7211, 7203, 7181, 7173,
-    7167, 7159, 7149, 7141, 7135, 7127, 7108, 7100, 7094, 7086, 7076, 7068,
-    7062, 7054, 7188, 7180, 7174, 7166, 7156, 7148, 7142, 7134, 7115, 7107,
-    7101, 7093, 7083, 7075, 7069, 7061, 7039, 7031, 7025, 7017, 7007, 6999,
-    6993, 6985, 6966, 6958, 6952, 6944, 6934, 6926, 6920, 6912, 6886, 6878,
-    6872, 6864, 6854, 6846, 6840, 6832, 6813, 6805, 6799, 6791, 6781, 6773,
-    6767, 6759, 6737, 6729, 6723, 6715, 6705, 6697, 6691, 6683, 6664, 6656,
-    6650, 6642, 6632, 6624, 6618, 6610, 6812, 6804, 6798, 6790, 6780, 6772,
-    6766, 6758, 6739, 6731, 6725, 6717, 6707, 6699, 6693, 6685, 6663, 6655,
-    6649, 6641, 6631, 6623, 6617, 6609, 6590, 6582, 6576, 6568, 6558, 6550,
-    6544, 6536, 6510, 6502, 6496, 6488, 6478, 6470, 6464, 6456, 6437, 6429,
-    6423, 6415, 6405, 6397, 6391, 6383, 6361, 6353, 6347, 6339, 6329, 6321,
-    6315, 6307, 6288, 6280, 6274, 6266, 6256, 6248, 6242, 6234, 6368, 6360,
-    6354, 6346, 6336, 6328, 6322, 6314, 6295, 6287, 6281, 6273, 6263, 6255,
-    6249, 6241, 6219, 6211, 6205, 6197, 6187, 6179, 6173, 6165, 6146, 6138,
-    6132, 6124, 6114, 6106, 6100, 6092, 6066, 6058, 6052, 6044, 6034, 6026,
-    6020, 6012, 5993, 5985, 5979, 5971, 5961, 5953, 5947, 5939, 5917, 5909,
-    5903, 5895, 5885, 5877, 5871, 5863, 5844, 5836, 5830, 5822, 5812, 5804,
-    5798, 5790, 6697, 6689, 6683, 6675, 6665, 6657, 6651, 6643, 6624, 6616,
-    6610, 6602, 6592, 6584, 6578, 6570, 6548, 6540, 6534, 6526, 6516, 6508,
-    6502, 6494, 6475, 6467, 6461, 6453, 6443, 6435, 6429, 6421, 6395, 6387,
-    6381, 6373, 6363, 6355, 6349, 6341, 6322, 6314, 6308, 6300, 6290, 6282,
-    6276, 6268, 6246, 6238, 6232, 6224, 6214, 6206, 6200, 6192, 6173, 6165,
-    6159, 6151, 6141, 6133, 6127, 6119, 6253, 6245, 6239, 6231, 6221, 6213,
-    6207, 6199, 6180, 6172, 6166, 6158, 6148, 6140, 6134, 6126, 6104, 6096,
-    6090, 6082, 6072, 6064, 6058, 6050, 6031, 6023, 6017, 6009, 5999, 5991,
-    5985, 5977, 5951, 5943, 5937, 5929, 5919, 5911, 5905, 5897, 5878, 5870,
-    5864, 5856, 5846, 5838, 5832, 5824, 5802, 5794, 5788, 5780, 5770, 5762,
-    5756, 5748, 5729, 5721, 5715, 5707, 5697, 5689, 5683, 5675, 5877, 5869,
-    5863, 5855, 5845, 5837, 5831, 5823, 5804, 5796, 5790, 5782, 5772, 5764,
-    5758, 5750, 5728, 5720, 5714, 5706, 5696, 5688, 5682, 5674, 5655, 5647,
-    5641, 5633, 5623, 5615, 5609, 5601, 5575, 5567, 5561, 5553, 5543, 5535,
-    5529, 5521, 5502, 5494, 5488, 5480, 5470, 5462, 5456, 5448, 5426, 5418,
-    5412, 5404, 5394, 5386, 5380, 5372, 5353, 5345, 5339, 5331, 5321, 5313,
-    5307, 5299, 5433, 5425, 5419, 5411, 5401, 5393, 5387, 5379, 5360, 5352,
-    5346, 5338, 5328, 5320, 5314, 5306, 5284, 5276, 5270, 5262, 5252, 5244,
-    5238, 5230, 5211, 5203, 5197, 5189, 5179, 5171, 5165, 5157, 5131, 5123,
-    5117, 5109, 5099, 5091, 5085, 5077, 5058, 5050, 5044, 5036, 5026, 5018,
-    5012, 5004, 4982, 4974, 4968, 4960, 4950, 4942, 4936, 4928, 4909, 4901,
-    4895, 4887, 4877, 4869, 4863, 4855, 5586, 5578, 5572, 5564, 5554, 5546,
-    5540, 5532, 5513, 5505, 5499, 5491, 5481, 5473, 5467, 5459, 5437, 5429,
-    5423, 5415, 5405, 5397, 5391, 5383, 5364, 5356, 5350, 5342, 5332, 5324,
-    5318, 5310, 5284, 5276, 5270, 5262, 5252, 5244, 5238, 5230, 5211, 5203,
-    5197, 5189, 5179, 5171, 5165, 5157, 5135, 5127, 5121, 5113, 5103, 5095,
-    5089, 5081, 5062, 5054, 5048, 5040, 5030, 5022, 5016, 5008, 5142, 5134,
-    5128, 5120, 5110, 5102, 5096, 5088, 5069, 5061, 5055, 5047, 5037, 5029,
-    5023, 5015, 4993, 4985, 4979, 4971, 4961, 4953, 4947, 4939, 4920, 4912,
-    4906, 4898, 4888, 4880, 4874, 4866, 4840, 4832, 4826, 4818, 4808, 4800,
-    4794, 4786, 4767, 4759, 4753, 4745, 4735, 4727, 4721, 4713, 4691, 4683,
-    4677, 4669, 4659, 4651, 4645, 4637, 4618, 4610, 4604, 4596, 4586, 4578,
-    4572, 4564, 4766, 4758, 4752, 4744, 4734, 4726, 4720, 4712, 4693, 4685,
-    4679, 4671, 4661, 4653, 4647, 4639, 4617, 4609, 4603, 4595, 4585, 4577,
-    4571, 4563, 4544, 4536, 4530, 4522, 4512, 4504, 4498, 4490, 4464, 4456,
-    4450, 4442, 4432, 4424, 4418, 4410, 4391, 4383, 4377, 4369, 4359, 4351,
-    4345, 4337, 4315, 4307, 4301, 4293, 4283, 4275, 4269, 4261, 4242, 4234,
-    4228, 4220, 4210, 4202, 4196, 4188, 4322, 4314, 4308, 4300, 4290, 4282,
-    4276, 4268, 4249, 4241, 4235, 4227, 4217, 4209, 4203, 4195, 4173, 4165,
-    4159, 4151, 4141, 4133, 4127, 4119, 4100, 4092, 4086, 4078, 4068, 4060,
-    4054, 4046, 4020, 4012, 4006, 3998, 3988, 3980, 3974, 3966, 3947, 3939,
-    3933, 3925, 3915, 3907, 3901, 3893, 3871, 3863, 3857, 3849, 3839, 3831,
-    3825, 3817, 3798, 3790, 3784, 3776, 3766, 3758, 3752, 3744, 6697, 6689,
-    6683, 6675, 6665, 6657, 6651, 6643, 6624, 6616, 6610, 6602, 6592, 6584,
-    6578, 6570, 6548, 6540, 6534, 6526, 6516, 6508, 6502, 6494, 6475, 6467,
-    6461, 6453, 6443, 6435, 6429, 6421, 6395, 6387, 6381, 6373, 6363, 6355,
-    6349, 6341, 6322, 6314, 6308, 6300, 6290, 6282, 6276, 6268, 6246, 6238,
-    6232, 6224, 6214, 6206, 6200, 6192, 6173, 6165, 6159, 6151, 6141, 6133,
-    6127, 6119, 6253, 6245, 6239, 6231, 6221, 6213, 6207, 6199, 6180, 6172,
-    6166, 6158, 6148, 6140, 6134, 6126, 6104, 6096, 6090, 6082, 6072, 6064,
-    6058, 6050, 6031, 6023, 6017, 6009, 5999, 5991, 5985, 5977, 5951, 5943,
-    5937, 5929, 5919, 5911, 5905, 5897, 5878, 5870, 5864, 5856, 5846, 5838,
-    5832, 5824, 5802, 5794, 5788, 5780, 5770, 5762, 5756, 5748, 5729, 5721,
-    5715, 5707, 5697, 5689, 5683, 5675, 5877, 5869, 5863, 5855, 5845, 5837,
-    5831, 5823, 5804, 5796, 5790, 5782, 5772, 5764, 5758, 5750, 5728, 5720,
-    5714, 5706, 5696, 5688, 5682, 5674, 5655, 5647, 5641, 5633, 5623, 5615,
-    5609, 5601, 5575, 5567, 5561, 5553, 5543, 5535, 5529, 5521, 5502, 5494,
-    5488, 5480, 5470, 5462, 5456, 5448, 5426, 5418, 5412, 5404, 5394, 5386,
-    5380, 5372, 5353, 5345, 5339, 5331, 5321, 5313, 5307, 5299, 5433, 5425,
-    5419, 5411, 5401, 5393, 5387, 5379, 5360, 5352, 5346, 5338, 5328, 5320,
-    5314, 5306, 5284, 5276, 5270, 5262, 5252, 5244, 5238, 5230, 5211, 5203,
-    5197, 5189, 5179, 5171, 5165, 5157, 5131, 5123, 5117, 5109, 5099, 5091,
-    5085, 5077, 5058, 5050, 5044, 5036, 5026, 5018, 5012, 5004, 4982, 4974,
-    4968, 4960, 4950, 4942, 4936, 4928, 4909, 4901, 4895, 4887, 4877, 4869,
-    4863, 4855, 5586, 5578, 5572, 5564, 5554, 5546, 5540, 5532, 5513, 5505,
-    5499, 5491, 5481, 5473, 5467, 5459, 5437, 5429, 5423, 5415, 5405, 5397,
-    5391, 5383, 5364, 5356, 5350, 5342, 5332, 5324, 5318, 5310, 5284, 5276,
-    5270, 5262, 5252, 5244, 5238, 5230, 5211, 5203, 5197, 5189, 5179, 5171,
-    5165, 5157, 5135, 5127, 5121, 5113, 5103, 5095, 5089, 5081, 5062, 5054,
-    5048, 5040, 5030, 5022, 5016, 5008, 5142, 5134, 5128, 5120, 5110, 5102,
-    5096, 5088, 5069, 5061, 5055, 5047, 5037, 5029, 5023, 5015, 4993, 4985,
-    4979, 4971, 4961, 4953, 4947, 4939, 4920, 4912, 4906, 4898, 4888, 4880,
-    4874, 4866, 4840, 4832, 4826, 4818, 4808, 4800, 4794, 4786, 4767, 4759,
-    4753, 4745, 4735, 4727, 4721, 4713, 4691, 4683, 4677, 4669, 4659, 4651,
-    4645, 4637, 4618, 4610, 4604, 4596, 4586, 4578, 4572, 4564, 4766, 4758,
-    4752, 4744, 4734, 4726, 4720, 4712, 4693, 4685, 4679, 4671, 4661, 4653,
-    4647, 4639, 4617, 4609, 4603, 4595, 4585, 4577, 4571, 4563, 4544, 4536,
-    4530, 4522, 4512, 4504, 4498, 4490, 4464, 4456, 4450, 4442, 4432, 4424,
-    4418, 4410, 4391, 4383, 4377, 4369, 4359, 4351, 4345, 4337, 4315, 4307,
-    4301, 4293, 4283, 4275, 4269, 4261, 4242, 4234, 4228, 4220, 4210, 4202,
-    4196, 4188, 4322, 4314, 4308, 4300, 4290, 4282, 4276, 4268, 4249, 4241,
-    4235, 4227, 4217, 4209, 4203, 4195, 4173, 4165, 4159, 4151, 4141, 4133,
-    4127, 4119, 4100, 4092, 4086, 4078, 4068, 4060, 4054, 4046, 4020, 4012,
-    4006, 3998, 3988, 3980, 3974, 3966, 3947, 3939, 3933, 3925, 3915, 3907,
-    3901, 3893, 3871, 3863, 3857, 3849, 3839, 3831, 3825, 3817, 3798, 3790,
-    3784, 3776, 3766, 3758, 3752, 3744, 4651, 4643, 4637, 4629, 4619, 4611,
-    4605, 4597, 4578, 4570, 4564, 4556, 4546, 4538, 4532, 4524, 4502, 4494,
-    4488, 4480, 4470, 4462, 4456, 4448, 4429, 4421, 4415, 4407, 4397, 4389,
-    4383, 4375, 4349, 4341, 4335, 4327, 4317, 4309, 4303, 4295, 4276, 4268,
-    4262, 4254, 4244, 4236, 4230, 4222, 4200, 4192, 4186, 4178, 4168, 4160,
-    4154, 4146, 4127, 4119, 4113, 4105, 4095, 4087, 4081, 4073, 4207, 4199,
-    4193, 4185, 4175, 4167, 4161, 4153, 4134, 4126, 4120, 4112, 4102, 4094,
-    4088, 4080, 4058, 4050, 4044, 4036, 4026, 4018, 4012, 4004, 3985, 3977,
-    3971, 3963, 3953, 3945, 3939, 3931, 3905, 3897, 3891, 3883, 3873, 3865,
-    3859, 3851, 3832, 3824, 3818, 3810, 3800, 3792, 3786, 3778, 3756, 3748,
-    3742, 3734, 3724, 3716, 3710, 3702, 3683, 3675, 3669, 3661, 3651, 3643,
-    3637, 3629, 3831, 3823, 3817, 3809, 3799, 3791, 3785, 3777, 3758, 3750,
-    3744, 3736, 3726, 3718, 3712, 3704, 3682, 3674, 3668, 3660, 3650, 3642,
-    3636, 3628, 3609, 3601, 3595, 3587, 3577, 3569, 3563, 3555, 3529, 3521,
-    3515, 3507, 3497, 3489, 3483, 3475, 3456, 3448, 3442, 3434, 3424, 3416,
-    3410, 3402, 3380, 3372, 3366, 3358, 3348, 3340, 3334, 3326, 3307, 3299,
-    3293, 3285, 3275, 3267, 3261, 3253, 3387, 3379, 3373, 3365, 3355, 3347,
-    3341, 3333, 3314, 3306, 3300, 3292, 3282, 3274, 3268, 3260, 3238, 3230,
-    3224, 3216, 3206, 3198, 3192, 3184, 3165, 3157, 3151, 3143, 3133, 3125,
-    3119, 3111, 3085, 3077, 3071, 3063, 3053, 3045, 3039, 3031, 3012, 3004,
-    2998, 2990, 2980, 2972, 2966, 2958, 2936, 2928, 2922, 2914, 2904, 2896,
-    2890, 2882, 2863, 2855, 2849, 2841, 2831, 2823, 2817, 2809, 3540, 3532,
-    3526, 3518, 3508, 3500, 3494, 3486, 3467, 3459, 3453, 3445, 3435, 3427,
-    3421, 3413, 3391, 3383, 3377, 3369, 3359, 3351, 3345, 3337, 3318, 3310,
-    3304, 3296, 3286, 3278, 3272, 3264, 3238, 3230, 3224, 3216, 3206, 3198,
-    3192, 3184, 3165, 3157, 3151, 3143, 3133, 3125, 3119, 3111, 3089, 3081,
-    3075, 3067, 3057, 3049, 3043, 3035, 3016, 3008, 3002, 2994, 2984, 2976,
-    2970, 2962, 3096, 3088, 3082, 3074, 3064, 3056, 3050, 3042, 3023, 3015,
-    3009, 3001, 2991, 2983, 2977, 2969, 2947, 2939, 2933, 2925, 2915, 2907,
-    2901, 2893, 2874, 2866, 2860, 2852, 2842, 2834, 2828, 2820, 2794, 2786,
-    2780, 2772, 2762, 2754, 2748, 2740, 2721, 2713, 2707, 2699, 2689, 2681,
-    2675, 2667, 2645, 2637, 2631, 2623, 2613, 2605, 2599, 2591, 2572, 2564,
-    2558, 2550, 2540, 2532, 2526, 2518, 2720, 2712, 2706, 2698, 2688, 2680,
-    2674, 2666, 2647, 2639, 2633, 2625, 2615, 2607, 2601, 2593, 2571, 2563,
-    2557, 2549, 2539, 2531, 2525, 2517, 2498, 2490, 2484, 2476, 2466, 2458,
-    2452, 2444, 2418, 2410, 2404, 2396, 2386, 2378, 2372, 2364, 2345, 2337,
-    2331, 2323, 2313, 2305, 2299, 2291, 2269, 2261, 2255, 2247, 2237, 2229,
-    2223, 2215, 2196, 2188, 2182, 2174, 2164, 2156, 2150, 2142, 2276, 2268,
-    2262, 2254, 2244, 2236, 2230, 2222, 2203, 2195, 2189, 2181, 2171, 2163,
-    2157, 2149, 2127, 2119, 2113, 2105, 2095, 2087, 2081, 2073, 2054, 2046,
-    2040, 2032, 2022, 2014, 2008, 2000, 1974, 1966, 1960, 1952, 1942, 1934,
-    1928, 1920, 1901, 1893, 1887, 1879, 1869, 1861, 1855, 1847, 1825, 1817,
-    1811, 1803, 1793, 1785, 1779, 1771, 1752, 1744, 1738, 1730, 1720, 1712,
-    1706, 1698, 1897, 1883, 1860, 1846, 1819, 1805, 1782, 1768, 1723, 1709,
-    1686, 1672, 1645, 1631, 1608, 1594, 1574, 1560, 1537, 1523, 1496, 1482,
-    1459, 1445, 1400, 1386, 1363, 1349, 1322, 1308, 1285, 1271, 1608, 1565,
-    1535, 1492, 1446, 1403, 1373, 1330, 1312, 1269, 1239, 1196, 1150, 1107,
-    1077, 1034, 1291, 1218, 1171, 1098, 1015, 942, 895, 822, 953, 850,
-    729, 626, 618, 431, 257, 257, 257, 257, 0, 255, 255, 255,
-    255, 429, 616, 624, 727, 848, 951, 820, 893, 940, 1013, 1096,
-    1169, 1216, 1289, 1032, 1075, 1105, 1148, 1194, 1237, 1267, 1310, 1328,
-    1371, 1401, 1444, 1490, 1533, 1563, 1606, 1269, 1283, 1306, 1320, 1347,
-    1361, 1384, 1398, 1443, 1457, 1480, 1494, 1521, 1535, 1558, 1572, 1592,
-    1606, 1629, 1643, 1670, 1684, 1707, 1721, 1766, 1780, 1803, 1817, 1844,
-    1858, 1881, 1895, 1696, 1704, 1710, 1718, 1728, 1736, 1742, 1750, 1769,
-    1777, 1783, 1791, 1801, 1809, 1815, 1823, 1845, 1853, 1859, 1867, 1877,
-    1885, 1891, 1899, 1918, 1926, 1932, 1940, 1950, 1958, 1964, 1972, 1998,
-    2006, 2012, 2020, 2030, 2038, 2044, 2052, 2071, 2079, 2085, 2093, 2103,
-    2111, 2117, 2125, 2147, 2155, 2161, 2169, 2179, 2187, 2193, 2201, 2220,
-    2228, 2234, 2242, 2252, 2260, 2266, 2274, 2140, 2148, 2154, 2162, 2172,
-    2180, 2186, 2194, 2213, 2221, 2227, 2235, 2245, 2253, 2259, 2267, 2289,
-    2297, 2303, 2311, 2321, 2329, 2335, 2343, 2362, 2370, 2376, 2384, 2394,
-    2402, 2408, 2416, 2442, 2450, 2456, 2464, 2474, 2482, 2488, 2496, 2515,
-    2523, 2529, 2537, 2547, 2555, 2561, 2569, 2591, 2599, 2605, 2613, 2623,
-    2631, 2637, 2645, 2664, 2672, 2678, 2686, 2696, 2704, 2710, 2718, 2516,
-    2524, 2530, 2538, 2548, 2556, 2562, 2570, 2589, 2597, 2603, 2611, 2621,
-    2629, 2635, 2643, 2665, 2673, 2679, 2687, 2697, 2705, 2711, 2719, 2738,
-    2746, 2752, 2760, 2770, 2778, 2784, 2792, 2818, 2826, 2832, 2840, 2850,
-    2858, 2864, 2872, 2891, 2899, 2905, 2913, 2923, 2931, 2937, 2945, 2967,
-    2975, 2981, 2989, 2999, 3007, 3013, 3021, 3040, 3048, 3054, 3062, 3072,
-    3080, 3086, 3094, 2960, 2968, 2974, 2982, 2992, 3000, 3006, 3014, 3033,
-    3041, 3047, 3055, 3065, 3073, 3079, 3087, 3109, 3117, 3123, 3131, 3141,
-    3149, 3155, 3163, 3182, 3190, 3196, 3204, 3214, 3222, 3228, 3236, 3262,
-    3270, 3276, 3284, 3294, 3302, 3308, 3316, 3335, 3343, 3349, 3357, 3367,
-    3375, 3381, 3389, 3411, 3419, 3425, 3433, 3443, 3451, 3457, 3465, 3484,
-    3492, 3498, 3506, 3516, 3524, 3530, 3538, 2807, 2815, 2821, 2829, 2839,
-    2847, 2853, 2861, 2880, 2888, 2894, 2902, 2912, 2920, 2926, 2934, 2956,
-    2964, 2970, 2978, 2988, 2996, 3002, 3010, 3029, 3037, 3043, 3051, 3061,
-    3069, 3075, 3083, 3109, 3117, 3123, 3131, 3141, 3149, 3155, 3163, 3182,
-    3190, 3196, 3204, 3214, 3222, 3228, 3236, 3258, 3266, 3272, 3280, 3290,
-    3298, 3304, 3312, 3331, 3339, 3345, 3353, 3363, 3371, 3377, 3385, 3251,
-    3259, 3265, 3273, 3283, 3291, 3297, 3305, 3324, 3332, 3338, 3346, 3356,
-    3364, 3370, 3378, 3400, 3408, 3414, 3422, 3432, 3440, 3446, 3454, 3473,
-    3481, 3487, 3495, 3505, 3513, 3519, 3527, 3553, 3561, 3567, 3575, 3585,
-    3593, 3599, 3607, 3626, 3634, 3640, 3648, 3658, 3666, 3672, 3680, 3702,
-    3710, 3716, 3724, 3734, 3742, 3748, 3756, 3775, 3783, 3789, 3797, 3807,
-    3815, 3821, 3829, 3627, 3635, 3641, 3649, 3659, 3667, 3673, 3681, 3700,
-    3708, 3714, 3722, 3732, 3740, 3746, 3754, 3776, 3784, 3790, 3798, 3808,
-    3816, 3822, 3830, 3849, 3857, 3863, 3871, 3881, 3889, 3895, 3903, 3929,
-    3937, 3943, 3951, 3961, 3969, 3975, 3983, 4002, 4010, 4016, 4024, 4034,
-    4042, 4048, 4056, 4078, 4086, 4092, 4100, 4110, 4118, 4124, 4132, 4151,
-    4159, 4165, 4173, 4183, 4191, 4197, 4205, 4071, 4079, 4085, 4093, 4103,
-    4111, 4117, 4125, 4144, 4152, 4158, 4166, 4176, 4184, 4190, 4198, 4220,
-    4228, 4234, 4242, 4252, 4260, 4266, 4274, 4293, 4301, 4307, 4315, 4325,
-    4333, 4339, 4347, 4373, 4381, 4387, 4395, 4405, 4413, 4419, 4427, 4446,
-    4454, 4460, 4468, 4478, 4486, 4492, 4500, 4522, 4530, 4536, 4544, 4554,
-    4562, 4568, 4576, 4595, 4603, 4609, 4617, 4627, 4635, 4641, 4649, 3742,
-    3750, 3756, 3764, 3774, 3782, 3788, 3796, 3815, 3823, 3829, 3837, 3847,
-    3855, 3861, 3869, 3891, 3899, 3905, 3913, 3923, 3931, 3937, 3945, 3964,
-    3972, 3978, 3986, 3996, 4004, 4010, 4018, 4044, 4052, 4058, 4066, 4076,
-    4084, 4090, 4098, 4117, 4125, 4131, 4139, 4149, 4157, 4163, 4171, 4193,
-    4201, 4207, 4215, 4225, 4233, 4239, 4247, 4266, 4274, 4280, 4288, 4298,
-    4306, 4312, 4320, 4186, 4194, 4200, 4208, 4218, 4226, 4232, 4240, 4259,
-    4267, 4273, 4281, 4291, 4299, 4305, 4313, 4335, 4343, 4349, 4357, 4367,
-    4375, 4381, 4389, 4408, 4416, 4422, 4430, 4440, 4448, 4454, 4462, 4488,
-    4496, 4502, 4510, 4520, 4528, 4534, 4542, 4561, 4569, 4575, 4583, 4593,
-    4601, 4607, 4615, 4637, 4645, 4651, 4659, 4669, 4677, 4683, 4691, 4710,
-    4718, 4724, 4732, 4742, 4750, 4756, 4764, 4562, 4570, 4576, 4584, 4594,
-    4602, 4608, 4616, 4635, 4643, 4649, 4657, 4667, 4675, 4681, 4689, 4711,
-    4719, 4725, 4733, 4743, 4751, 4757, 4765, 4784, 4792, 4798, 4806, 4816,
-    4824, 4830, 4838, 4864, 4872, 4878, 4886, 4896, 4904, 4910, 4918, 4937,
-    4945, 4951, 4959, 4969, 4977, 4983, 4991, 5013, 5021, 5027, 5035, 5045,
-    5053, 5059, 5067, 5086, 5094, 5100, 5108, 5118, 5126, 5132, 5140, 5006,
-    5014, 5020, 5028, 5038, 5046, 5052, 5060, 5079, 5087, 5093, 5101, 5111,
-    5119, 5125, 5133, 5155, 5163, 5169, 5177, 5187, 5195, 5201, 5209, 5228,
-    5236, 5242, 5250, 5260, 5268, 5274, 5282, 5308, 5316, 5322, 5330, 5340,
-    5348, 5354, 5362, 5381, 5389, 5395, 5403, 5413, 5421, 5427, 5435, 5457,
-    5465, 5471, 5479, 5489, 5497, 5503, 5511, 5530, 5538, 5544, 5552, 5562,
-    5570, 5576, 5584, 4853, 4861, 4867, 4875, 4885, 4893, 4899, 4907, 4926,
-    4934, 4940, 4948, 4958, 4966, 4972, 4980, 5002, 5010, 5016, 5024, 5034,
-    5042, 5048, 5056, 5075, 5083, 5089, 5097, 5107, 5115, 5121, 5129, 5155,
-    5163, 5169, 5177, 5187, 5195, 5201, 5209, 5228, 5236, 5242, 5250, 5260,
-    5268, 5274, 5282, 5304, 5312, 5318, 5326, 5336, 5344, 5350, 5358, 5377,
-    5385, 5391, 5399, 5409, 5417, 5423, 5431, 5297, 5305, 5311, 5319, 5329,
-    5337, 5343, 5351, 5370, 5378, 5384, 5392, 5402, 5410, 5416, 5424, 5446,
-    5454, 5460, 5468, 5478, 5486, 5492, 5500, 5519, 5527, 5533, 5541, 5551,
-    5559, 5565, 5573, 5599, 5607, 5613, 5621, 5631, 5639, 5645, 5653, 5672,
-    5680, 5686, 5694, 5704, 5712, 5718, 5726, 5748, 5756, 5762, 5770, 5780,
-    5788, 5794, 5802, 5821, 5829, 5835, 5843, 5853, 5861, 5867, 5875, 5673,
-    5681, 5687, 5695, 5705, 5713, 5719, 5727, 5746, 5754, 5760, 5768, 5778,
-    5786, 5792, 5800, 5822, 5830, 5836, 5844, 5854, 5862, 5868, 5876, 5895,
-    5903, 5909, 5917, 5927, 5935, 5941, 5949, 5975, 5983, 5989, 5997, 6007,
-    6015, 6021, 6029, 6048, 6056, 6062, 6070, 6080, 6088, 6094, 6102, 6124,
-    6132, 6138, 6146, 6156, 6164, 6170, 6178, 6197, 6205, 6211, 6219, 6229,
-    6237, 6243, 6251, 6117, 6125, 6131, 6139, 6149, 6157, 6163, 6171, 6190,
-    6198, 6204, 6212, 6222, 6230, 6236, 6244, 6266, 6274, 6280, 6288, 6298,
-    6306, 6312, 6320, 6339, 6347, 6353, 6361, 6371, 6379, 6385, 6393, 6419,
-    6427, 6433, 6441, 6451, 6459, 6465, 6473, 6492, 6500, 6506, 6514, 6524,
-    6532, 6538, 6546, 6568, 6576, 6582, 6590, 6600, 6608, 6614, 6622, 6641,
-    6649, 6655, 6663, 6673, 6681, 6687, 6695, 3742, 3750, 3756, 3764, 3774,
-    3782, 3788, 3796, 3815, 3823, 3829, 3837, 3847, 3855, 3861, 3869, 3891,
-    3899, 3905, 3913, 3923, 3931, 3937, 3945, 3964, 3972, 3978, 3986, 3996,
-    4004, 4010, 4018, 4044, 4052, 4058, 4066, 4076, 4084, 4090, 4098, 4117,
-    4125, 4131, 4139, 4149, 4157, 4163, 4171, 4193, 4201, 4207, 4215, 4225,
-    4233, 4239, 4247, 4266, 4274, 4280, 4288, 4298, 4306, 4312, 4320, 4186,
-    4194, 4200, 4208, 4218, 4226, 4232, 4240, 4259, 4267, 4273, 4281, 4291,
-    4299, 4305, 4313, 4335, 4343, 4349, 4357, 4367, 4375, 4381, 4389, 4408,
-    4416, 4422, 4430, 4440, 4448, 4454, 4462, 4488, 4496, 4502, 4510, 4520,
-    4528, 4534, 4542, 4561, 4569, 4575, 4583, 4593, 4601, 4607, 4615, 4637,
-    4645, 4651, 4659, 4669, 4677, 4683, 4691, 4710, 4718, 4724, 4732, 4742,
-    4750, 4756, 4764, 4562, 4570, 4576, 4584, 4594, 4602, 4608, 4616, 4635,
-    4643, 4649, 4657, 4667, 4675, 4681, 4689, 4711, 4719, 4725, 4733, 4743,
-    4751, 4757, 4765, 4784, 4792, 4798, 4806, 4816, 4824, 4830, 4838, 4864,
-    4872, 4878, 4886, 4896, 4904, 4910, 4918, 4937, 4945, 4951, 4959, 4969,
-    4977, 4983, 4991, 5013, 5021, 5027, 5035, 5045, 5053, 5059, 5067, 5086,
-    5094, 5100, 5108, 5118, 5126, 5132, 5140, 5006, 5014, 5020, 5028, 5038,
-    5046, 5052, 5060, 5079, 5087, 5093, 5101, 5111, 5119, 5125, 5133, 5155,
-    5163, 5169, 5177, 5187, 5195, 5201, 5209, 5228, 5236, 5242, 5250, 5260,
-    5268, 5274, 5282, 5308, 5316, 5322, 5330, 5340, 5348, 5354, 5362, 5381,
-    5389, 5395, 5403, 5413, 5421, 5427, 5435, 5457, 5465, 5471, 5479, 5489,
-    5497, 5503, 5511, 5530, 5538, 5544, 5552, 5562, 5570, 5576, 5584, 4853,
-    4861, 4867, 4875, 4885, 4893, 4899, 4907, 4926, 4934, 4940, 4948, 4958,
-    4966, 4972, 4980, 5002, 5010, 5016, 5024, 5034, 5042, 5048, 5056, 5075,
-    5083, 5089, 5097, 5107, 5115, 5121, 5129, 5155, 5163, 5169, 5177, 5187,
-    5195, 5201, 5209, 5228, 5236, 5242, 5250, 5260, 5268, 5274, 5282, 5304,
-    5312, 5318, 5326, 5336, 5344, 5350, 5358, 5377, 5385, 5391, 5399, 5409,
-    5417, 5423, 5431, 5297, 5305, 5311, 5319, 5329, 5337, 5343, 5351, 5370,
-    5378, 5384, 5392, 5402, 5410, 5416, 5424, 5446, 5454, 5460, 5468, 5478,
-    5486, 5492, 5500, 5519, 5527, 5533, 5541, 5551, 5559, 5565, 5573, 5599,
-    5607, 5613, 5621, 5631, 5639, 5645, 5653, 5672, 5680, 5686, 5694, 5704,
-    5712, 5718, 5726, 5748, 5756, 5762, 5770, 5780, 5788, 5794, 5802, 5821,
-    5829, 5835, 5843, 5853, 5861, 5867, 5875, 5673, 5681, 5687, 5695, 5705,
-    5713, 5719, 5727, 5746, 5754, 5760, 5768, 5778, 5786, 5792, 5800, 5822,
-    5830, 5836, 5844, 5854, 5862, 5868, 5876, 5895, 5903, 5909, 5917, 5927,
-    5935, 5941, 5949, 5975, 5983, 5989, 5997, 6007, 6015, 6021, 6029, 6048,
-    6056, 6062, 6070, 6080, 6088, 6094, 6102, 6124, 6132, 6138, 6146, 6156,
-    6164, 6170, 6178, 6197, 6205, 6211, 6219, 6229, 6237, 6243, 6251, 6117,
-    6125, 6131, 6139, 6149, 6157, 6163, 6171, 6190, 6198, 6204, 6212, 6222,
-    6230, 6236, 6244, 6266, 6274, 6280, 6288, 6298, 6306, 6312, 6320, 6339,
-    6347, 6353, 6361, 6371, 6379, 6385, 6393, 6419, 6427, 6433, 6441, 6451,
-    6459, 6465, 6473, 6492, 6500, 6506, 6514, 6524, 6532, 6538, 6546, 6568,
-    6576, 6582, 6590, 6600, 6608, 6614, 6622, 6641, 6649, 6655, 6663, 6673,
-    6681, 6687, 6695, 5788, 5796, 5802, 5810, 5820, 5828, 5834, 5842, 5861,
-    5869, 5875, 5883, 5893, 5901, 5907, 5915, 5937, 5945, 5951, 5959, 5969,
-    5977, 5983, 5991, 6010, 6018, 6024, 6032, 6042, 6050, 6056, 6064, 6090,
-    6098, 6104, 6112, 6122, 6130, 6136, 6144, 6163, 6171, 6177, 6185, 6195,
-    6203, 6209, 6217, 6239, 6247, 6253, 6261, 6271, 6279, 6285, 6293, 6312,
-    6320, 6326, 6334, 6344, 6352, 6358, 6366, 6232, 6240, 6246, 6254, 6264,
-    6272, 6278, 6286, 6305, 6313, 6319, 6327, 6337, 6345, 6351, 6359, 6381,
-    6389, 6395, 6403, 6413, 6421, 6427, 6435, 6454, 6462, 6468, 6476, 6486,
-    6494, 6500, 6508, 6534, 6542, 6548, 6556, 6566, 6574, 6580, 6588, 6607,
-    6615, 6621, 6629, 6639, 6647, 6653, 6661, 6683, 6691, 6697, 6705, 6715,
-    6723, 6729, 6737, 6756, 6764, 6770, 6778, 6788, 6796, 6802, 6810, 6608,
-    6616, 6622, 6630, 6640, 6648, 6654, 6662, 6681, 6689, 6695, 6703, 6713,
-    6721, 6727, 6735, 6757, 6765, 6771, 6779, 6789, 6797, 6803, 6811, 6830,
-    6838, 6844, 6852, 6862, 6870, 6876, 6884, 6910, 6918, 6924, 6932, 6942,
-    6950, 6956, 6964, 6983, 6991, 6997, 7005, 7015, 7023, 7029, 7037, 7059,
-    7067, 7073, 7081, 7091, 7099, 7105, 7113, 7132, 7140, 7146, 7154, 7164,
-    7172, 7178, 7186, 7052, 7060, 7066, 7074, 7084, 7092, 7098, 7106, 7125,
-    7133, 7139, 7147, 7157, 7165, 7171, 7179, 7201, 7209, 7215, 7223, 7233,
-    7241, 7247, 7255, 7274, 7282, 7288, 7296, 7306, 7314, 7320, 7328, 7354,
-    7362, 7368, 7376, 7386, 7394, 7400, 7408, 7427, 7435, 7441, 7449, 7459,
-    7467, 7473, 7481, 7503, 7511, 7517, 7525, 7535, 7543, 7549, 7557, 7576,
-    7584, 7590, 7598, 7608, 7616, 7622, 7630, 6899, 6907, 6913, 6921, 6931,
-    6939, 6945, 6953, 6972, 6980, 6986, 6994, 7004, 7012, 7018, 7026, 7048,
-    7056, 7062, 7070, 7080, 7088, 7094, 7102, 7121, 7129, 7135, 7143, 7153,
-    7161, 7167, 7175, 7201, 7209, 7215, 7223, 7233, 7241, 7247, 7255, 7274,
-    7282, 7288, 7296, 7306, 7314, 7320, 7328, 7350, 7358, 7364, 7372, 7382,
-    7390, 7396, 7404, 7423, 7431, 7437, 7445, 7455, 7463, 7469, 7477, 7343,
-    7351, 7357, 7365, 7375, 7383, 7389, 7397, 7416, 7424, 7430, 7438, 7448,
-    7456, 7462, 7470, 7492, 7500, 7506, 7514, 7524, 7532, 7538, 7546, 7565,
-    7573, 7579, 7587, 7597, 7605, 7611, 7619, 7645, 7653, 7659, 7667, 7677,
-    7685, 7691, 7699, 7718, 7726, 7732, 7740, 7750, 7758, 7764, 7772, 7794,
-    7802, 7808, 7816, 7826, 7834, 7840, 7848, 7867, 7875, 7881, 7889, 7899,
-    7907, 7913, 7921, 7719, 7727, 7733, 7741, 7751, 7759, 7765, 7773, 7792,
-    7800, 7806, 7814, 7824, 7832, 7838, 7846, 7868, 7876, 7882, 7890, 7900,
-    7908, 7914, 7922, 7941, 7949, 7955, 7963, 7973, 7981, 7987, 7995, 8021,
-    8029, 8035, 8043, 8053, 8061, 8067, 8075, 8094, 8102, 8108, 8116, 8126,
-    8134, 8140, 8148, 8170, 8178, 8184, 8192, 8202, 8210, 8216, 8224, 8243,
-    8251, 8257, 8265, 8275
+static const short dct_value_cost[2048 * 2] = {
+  8285, 8277, 8267, 8259, 8253, 8245, 8226, 8218, 8212, 8204, 8194, 8186, 8180,
+  8172, 8150, 8142, 8136, 8128, 8118, 8110, 8104, 8096, 8077, 8069, 8063, 8055,
+  8045, 8037, 8031, 8023, 7997, 7989, 7983, 7975, 7965, 7957, 7951, 7943, 7924,
+  7916, 7910, 7902, 7892, 7884, 7878, 7870, 7848, 7840, 7834, 7826, 7816, 7808,
+  7802, 7794, 7775, 7767, 7761, 7753, 7743, 7735, 7729, 7721, 7923, 7915, 7909,
+  7901, 7891, 7883, 7877, 7869, 7850, 7842, 7836, 7828, 7818, 7810, 7804, 7796,
+  7774, 7766, 7760, 7752, 7742, 7734, 7728, 7720, 7701, 7693, 7687, 7679, 7669,
+  7661, 7655, 7647, 7621, 7613, 7607, 7599, 7589, 7581, 7575, 7567, 7548, 7540,
+  7534, 7526, 7516, 7508, 7502, 7494, 7472, 7464, 7458, 7450, 7440, 7432, 7426,
+  7418, 7399, 7391, 7385, 7377, 7367, 7359, 7353, 7345, 7479, 7471, 7465, 7457,
+  7447, 7439, 7433, 7425, 7406, 7398, 7392, 7384, 7374, 7366, 7360, 7352, 7330,
+  7322, 7316, 7308, 7298, 7290, 7284, 7276, 7257, 7249, 7243, 7235, 7225, 7217,
+  7211, 7203, 7177, 7169, 7163, 7155, 7145, 7137, 7131, 7123, 7104, 7096, 7090,
+  7082, 7072, 7064, 7058, 7050, 7028, 7020, 7014, 7006, 6996, 6988, 6982, 6974,
+  6955, 6947, 6941, 6933, 6923, 6915, 6909, 6901, 7632, 7624, 7618, 7610, 7600,
+  7592, 7586, 7578, 7559, 7551, 7545, 7537, 7527, 7519, 7513, 7505, 7483, 7475,
+  7469, 7461, 7451, 7443, 7437, 7429, 7410, 7402, 7396, 7388, 7378, 7370, 7364,
+  7356, 7330, 7322, 7316, 7308, 7298, 7290, 7284, 7276, 7257, 7249, 7243, 7235,
+  7225, 7217, 7211, 7203, 7181, 7173, 7167, 7159, 7149, 7141, 7135, 7127, 7108,
+  7100, 7094, 7086, 7076, 7068, 7062, 7054, 7188, 7180, 7174, 7166, 7156, 7148,
+  7142, 7134, 7115, 7107, 7101, 7093, 7083, 7075, 7069, 7061, 7039, 7031, 7025,
+  7017, 7007, 6999, 6993, 6985, 6966, 6958, 6952, 6944, 6934, 6926, 6920, 6912,
+  6886, 6878, 6872, 6864, 6854, 6846, 6840, 6832, 6813, 6805, 6799, 6791, 6781,
+  6773, 6767, 6759, 6737, 6729, 6723, 6715, 6705, 6697, 6691, 6683, 6664, 6656,
+  6650, 6642, 6632, 6624, 6618, 6610, 6812, 6804, 6798, 6790, 6780, 6772, 6766,
+  6758, 6739, 6731, 6725, 6717, 6707, 6699, 6693, 6685, 6663, 6655, 6649, 6641,
+  6631, 6623, 6617, 6609, 6590, 6582, 6576, 6568, 6558, 6550, 6544, 6536, 6510,
+  6502, 6496, 6488, 6478, 6470, 6464, 6456, 6437, 6429, 6423, 6415, 6405, 6397,
+  6391, 6383, 6361, 6353, 6347, 6339, 6329, 6321, 6315, 6307, 6288, 6280, 6274,
+  6266, 6256, 6248, 6242, 6234, 6368, 6360, 6354, 6346, 6336, 6328, 6322, 6314,
+  6295, 6287, 6281, 6273, 6263, 6255, 6249, 6241, 6219, 6211, 6205, 6197, 6187,
+  6179, 6173, 6165, 6146, 6138, 6132, 6124, 6114, 6106, 6100, 6092, 6066, 6058,
+  6052, 6044, 6034, 6026, 6020, 6012, 5993, 5985, 5979, 5971, 5961, 5953, 5947,
+  5939, 5917, 5909, 5903, 5895, 5885, 5877, 5871, 5863, 5844, 5836, 5830, 5822,
+  5812, 5804, 5798, 5790, 6697, 6689, 6683, 6675, 6665, 6657, 6651, 6643, 6624,
+  6616, 6610, 6602, 6592, 6584, 6578, 6570, 6548, 6540, 6534, 6526, 6516, 6508,
+  6502, 6494, 6475, 6467, 6461, 6453, 6443, 6435, 6429, 6421, 6395, 6387, 6381,
+  6373, 6363, 6355, 6349, 6341, 6322, 6314, 6308, 6300, 6290, 6282, 6276, 6268,
+  6246, 6238, 6232, 6224, 6214, 6206, 6200, 6192, 6173, 6165, 6159, 6151, 6141,
+  6133, 6127, 6119, 6253, 6245, 6239, 6231, 6221, 6213, 6207, 6199, 6180, 6172,
+  6166, 6158, 6148, 6140, 6134, 6126, 6104, 6096, 6090, 6082, 6072, 6064, 6058,
+  6050, 6031, 6023, 6017, 6009, 5999, 5991, 5985, 5977, 5951, 5943, 5937, 5929,
+  5919, 5911, 5905, 5897, 5878, 5870, 5864, 5856, 5846, 5838, 5832, 5824, 5802,
+  5794, 5788, 5780, 5770, 5762, 5756, 5748, 5729, 5721, 5715, 5707, 5697, 5689,
+  5683, 5675, 5877, 5869, 5863, 5855, 5845, 5837, 5831, 5823, 5804, 5796, 5790,
+  5782, 5772, 5764, 5758, 5750, 5728, 5720, 5714, 5706, 5696, 5688, 5682, 5674,
+  5655, 5647, 5641, 5633, 5623, 5615, 5609, 5601, 5575, 5567, 5561, 5553, 5543,
+  5535, 5529, 5521, 5502, 5494, 5488, 5480, 5470, 5462, 5456, 5448, 5426, 5418,
+  5412, 5404, 5394, 5386, 5380, 5372, 5353, 5345, 5339, 5331, 5321, 5313, 5307,
+  5299, 5433, 5425, 5419, 5411, 5401, 5393, 5387, 5379, 5360, 5352, 5346, 5338,
+  5328, 5320, 5314, 5306, 5284, 5276, 5270, 5262, 5252, 5244, 5238, 5230, 5211,
+  5203, 5197, 5189, 5179, 5171, 5165, 5157, 5131, 5123, 5117, 5109, 5099, 5091,
+  5085, 5077, 5058, 5050, 5044, 5036, 5026, 5018, 5012, 5004, 4982, 4974, 4968,
+  4960, 4950, 4942, 4936, 4928, 4909, 4901, 4895, 4887, 4877, 4869, 4863, 4855,
+  5586, 5578, 5572, 5564, 5554, 5546, 5540, 5532, 5513, 5505, 5499, 5491, 5481,
+  5473, 5467, 5459, 5437, 5429, 5423, 5415, 5405, 5397, 5391, 5383, 5364, 5356,
+  5350, 5342, 5332, 5324, 5318, 5310, 5284, 5276, 5270, 5262, 5252, 5244, 5238,
+  5230, 5211, 5203, 5197, 5189, 5179, 5171, 5165, 5157, 5135, 5127, 5121, 5113,
+  5103, 5095, 5089, 5081, 5062, 5054, 5048, 5040, 5030, 5022, 5016, 5008, 5142,
+  5134, 5128, 5120, 5110, 5102, 5096, 5088, 5069, 5061, 5055, 5047, 5037, 5029,
+  5023, 5015, 4993, 4985, 4979, 4971, 4961, 4953, 4947, 4939, 4920, 4912, 4906,
+  4898, 4888, 4880, 4874, 4866, 4840, 4832, 4826, 4818, 4808, 4800, 4794, 4786,
+  4767, 4759, 4753, 4745, 4735, 4727, 4721, 4713, 4691, 4683, 4677, 4669, 4659,
+  4651, 4645, 4637, 4618, 4610, 4604, 4596, 4586, 4578, 4572, 4564, 4766, 4758,
+  4752, 4744, 4734, 4726, 4720, 4712, 4693, 4685, 4679, 4671, 4661, 4653, 4647,
+  4639, 4617, 4609, 4603, 4595, 4585, 4577, 4571, 4563, 4544, 4536, 4530, 4522,
+  4512, 4504, 4498, 4490, 4464, 4456, 4450, 4442, 4432, 4424, 4418, 4410, 4391,
+  4383, 4377, 4369, 4359, 4351, 4345, 4337, 4315, 4307, 4301, 4293, 4283, 4275,
+  4269, 4261, 4242, 4234, 4228, 4220, 4210, 4202, 4196, 4188, 4322, 4314, 4308,
+  4300, 4290, 4282, 4276, 4268, 4249, 4241, 4235, 4227, 4217, 4209, 4203, 4195,
+  4173, 4165, 4159, 4151, 4141, 4133, 4127, 4119, 4100, 4092, 4086, 4078, 4068,
+  4060, 4054, 4046, 4020, 4012, 4006, 3998, 3988, 3980, 3974, 3966, 3947, 3939,
+  3933, 3925, 3915, 3907, 3901, 3893, 3871, 3863, 3857, 3849, 3839, 3831, 3825,
+  3817, 3798, 3790, 3784, 3776, 3766, 3758, 3752, 3744, 6697, 6689, 6683, 6675,
+  6665, 6657, 6651, 6643, 6624, 6616, 6610, 6602, 6592, 6584, 6578, 6570, 6548,
+  6540, 6534, 6526, 6516, 6508, 6502, 6494, 6475, 6467, 6461, 6453, 6443, 6435,
+  6429, 6421, 6395, 6387, 6381, 6373, 6363, 6355, 6349, 6341, 6322, 6314, 6308,
+  6300, 6290, 6282, 6276, 6268, 6246, 6238, 6232, 6224, 6214, 6206, 6200, 6192,
+  6173, 6165, 6159, 6151, 6141, 6133, 6127, 6119, 6253, 6245, 6239, 6231, 6221,
+  6213, 6207, 6199, 6180, 6172, 6166, 6158, 6148, 6140, 6134, 6126, 6104, 6096,
+  6090, 6082, 6072, 6064, 6058, 6050, 6031, 6023, 6017, 6009, 5999, 5991, 5985,
+  5977, 5951, 5943, 5937, 5929, 5919, 5911, 5905, 5897, 5878, 5870, 5864, 5856,
+  5846, 5838, 5832, 5824, 5802, 5794, 5788, 5780, 5770, 5762, 5756, 5748, 5729,
+  5721, 5715, 5707, 5697, 5689, 5683, 5675, 5877, 5869, 5863, 5855, 5845, 5837,
+  5831, 5823, 5804, 5796, 5790, 5782, 5772, 5764, 5758, 5750, 5728, 5720, 5714,
+  5706, 5696, 5688, 5682, 5674, 5655, 5647, 5641, 5633, 5623, 5615, 5609, 5601,
+  5575, 5567, 5561, 5553, 5543, 5535, 5529, 5521, 5502, 5494, 5488, 5480, 5470,
+  5462, 5456, 5448, 5426, 5418, 5412, 5404, 5394, 5386, 5380, 5372, 5353, 5345,
+  5339, 5331, 5321, 5313, 5307, 5299, 5433, 5425, 5419, 5411, 5401, 5393, 5387,
+  5379, 5360, 5352, 5346, 5338, 5328, 5320, 5314, 5306, 5284, 5276, 5270, 5262,
+  5252, 5244, 5238, 5230, 5211, 5203, 5197, 5189, 5179, 5171, 5165, 5157, 5131,
+  5123, 5117, 5109, 5099, 5091, 5085, 5077, 5058, 5050, 5044, 5036, 5026, 5018,
+  5012, 5004, 4982, 4974, 4968, 4960, 4950, 4942, 4936, 4928, 4909, 4901, 4895,
+  4887, 4877, 4869, 4863, 4855, 5586, 5578, 5572, 5564, 5554, 5546, 5540, 5532,
+  5513, 5505, 5499, 5491, 5481, 5473, 5467, 5459, 5437, 5429, 5423, 5415, 5405,
+  5397, 5391, 5383, 5364, 5356, 5350, 5342, 5332, 5324, 5318, 5310, 5284, 5276,
+  5270, 5262, 5252, 5244, 5238, 5230, 5211, 5203, 5197, 5189, 5179, 5171, 5165,
+  5157, 5135, 5127, 5121, 5113, 5103, 5095, 5089, 5081, 5062, 5054, 5048, 5040,
+  5030, 5022, 5016, 5008, 5142, 5134, 5128, 5120, 5110, 5102, 5096, 5088, 5069,
+  5061, 5055, 5047, 5037, 5029, 5023, 5015, 4993, 4985, 4979, 4971, 4961, 4953,
+  4947, 4939, 4920, 4912, 4906, 4898, 4888, 4880, 4874, 4866, 4840, 4832, 4826,
+  4818, 4808, 4800, 4794, 4786, 4767, 4759, 4753, 4745, 4735, 4727, 4721, 4713,
+  4691, 4683, 4677, 4669, 4659, 4651, 4645, 4637, 4618, 4610, 4604, 4596, 4586,
+  4578, 4572, 4564, 4766, 4758, 4752, 4744, 4734, 4726, 4720, 4712, 4693, 4685,
+  4679, 4671, 4661, 4653, 4647, 4639, 4617, 4609, 4603, 4595, 4585, 4577, 4571,
+  4563, 4544, 4536, 4530, 4522, 4512, 4504, 4498, 4490, 4464, 4456, 4450, 4442,
+  4432, 4424, 4418, 4410, 4391, 4383, 4377, 4369, 4359, 4351, 4345, 4337, 4315,
+  4307, 4301, 4293, 4283, 4275, 4269, 4261, 4242, 4234, 4228, 4220, 4210, 4202,
+  4196, 4188, 4322, 4314, 4308, 4300, 4290, 4282, 4276, 4268, 4249, 4241, 4235,
+  4227, 4217, 4209, 4203, 4195, 4173, 4165, 4159, 4151, 4141, 4133, 4127, 4119,
+  4100, 4092, 4086, 4078, 4068, 4060, 4054, 4046, 4020, 4012, 4006, 3998, 3988,
+  3980, 3974, 3966, 3947, 3939, 3933, 3925, 3915, 3907, 3901, 3893, 3871, 3863,
+  3857, 3849, 3839, 3831, 3825, 3817, 3798, 3790, 3784, 3776, 3766, 3758, 3752,
+  3744, 4651, 4643, 4637, 4629, 4619, 4611, 4605, 4597, 4578, 4570, 4564, 4556,
+  4546, 4538, 4532, 4524, 4502, 4494, 4488, 4480, 4470, 4462, 4456, 4448, 4429,
+  4421, 4415, 4407, 4397, 4389, 4383, 4375, 4349, 4341, 4335, 4327, 4317, 4309,
+  4303, 4295, 4276, 4268, 4262, 4254, 4244, 4236, 4230, 4222, 4200, 4192, 4186,
+  4178, 4168, 4160, 4154, 4146, 4127, 4119, 4113, 4105, 4095, 4087, 4081, 4073,
+  4207, 4199, 4193, 4185, 4175, 4167, 4161, 4153, 4134, 4126, 4120, 4112, 4102,
+  4094, 4088, 4080, 4058, 4050, 4044, 4036, 4026, 4018, 4012, 4004, 3985, 3977,
+  3971, 3963, 3953, 3945, 3939, 3931, 3905, 3897, 3891, 3883, 3873, 3865, 3859,
+  3851, 3832, 3824, 3818, 3810, 3800, 3792, 3786, 3778, 3756, 3748, 3742, 3734,
+  3724, 3716, 3710, 3702, 3683, 3675, 3669, 3661, 3651, 3643, 3637, 3629, 3831,
+  3823, 3817, 3809, 3799, 3791, 3785, 3777, 3758, 3750, 3744, 3736, 3726, 3718,
+  3712, 3704, 3682, 3674, 3668, 3660, 3650, 3642, 3636, 3628, 3609, 3601, 3595,
+  3587, 3577, 3569, 3563, 3555, 3529, 3521, 3515, 3507, 3497, 3489, 3483, 3475,
+  3456, 3448, 3442, 3434, 3424, 3416, 3410, 3402, 3380, 3372, 3366, 3358, 3348,
+  3340, 3334, 3326, 3307, 3299, 3293, 3285, 3275, 3267, 3261, 3253, 3387, 3379,
+  3373, 3365, 3355, 3347, 3341, 3333, 3314, 3306, 3300, 3292, 3282, 3274, 3268,
+  3260, 3238, 3230, 3224, 3216, 3206, 3198, 3192, 3184, 3165, 3157, 3151, 3143,
+  3133, 3125, 3119, 3111, 3085, 3077, 3071, 3063, 3053, 3045, 3039, 3031, 3012,
+  3004, 2998, 2990, 2980, 2972, 2966, 2958, 2936, 2928, 2922, 2914, 2904, 2896,
+  2890, 2882, 2863, 2855, 2849, 2841, 2831, 2823, 2817, 2809, 3540, 3532, 3526,
+  3518, 3508, 3500, 3494, 3486, 3467, 3459, 3453, 3445, 3435, 3427, 3421, 3413,
+  3391, 3383, 3377, 3369, 3359, 3351, 3345, 3337, 3318, 3310, 3304, 3296, 3286,
+  3278, 3272, 3264, 3238, 3230, 3224, 3216, 3206, 3198, 3192, 3184, 3165, 3157,
+  3151, 3143, 3133, 3125, 3119, 3111, 3089, 3081, 3075, 3067, 3057, 3049, 3043,
+  3035, 3016, 3008, 3002, 2994, 2984, 2976, 2970, 2962, 3096, 3088, 3082, 3074,
+  3064, 3056, 3050, 3042, 3023, 3015, 3009, 3001, 2991, 2983, 2977, 2969, 2947,
+  2939, 2933, 2925, 2915, 2907, 2901, 2893, 2874, 2866, 2860, 2852, 2842, 2834,
+  2828, 2820, 2794, 2786, 2780, 2772, 2762, 2754, 2748, 2740, 2721, 2713, 2707,
+  2699, 2689, 2681, 2675, 2667, 2645, 2637, 2631, 2623, 2613, 2605, 2599, 2591,
+  2572, 2564, 2558, 2550, 2540, 2532, 2526, 2518, 2720, 2712, 2706, 2698, 2688,
+  2680, 2674, 2666, 2647, 2639, 2633, 2625, 2615, 2607, 2601, 2593, 2571, 2563,
+  2557, 2549, 2539, 2531, 2525, 2517, 2498, 2490, 2484, 2476, 2466, 2458, 2452,
+  2444, 2418, 2410, 2404, 2396, 2386, 2378, 2372, 2364, 2345, 2337, 2331, 2323,
+  2313, 2305, 2299, 2291, 2269, 2261, 2255, 2247, 2237, 2229, 2223, 2215, 2196,
+  2188, 2182, 2174, 2164, 2156, 2150, 2142, 2276, 2268, 2262, 2254, 2244, 2236,
+  2230, 2222, 2203, 2195, 2189, 2181, 2171, 2163, 2157, 2149, 2127, 2119, 2113,
+  2105, 2095, 2087, 2081, 2073, 2054, 2046, 2040, 2032, 2022, 2014, 2008, 2000,
+  1974, 1966, 1960, 1952, 1942, 1934, 1928, 1920, 1901, 1893, 1887, 1879, 1869,
+  1861, 1855, 1847, 1825, 1817, 1811, 1803, 1793, 1785, 1779, 1771, 1752, 1744,
+  1738, 1730, 1720, 1712, 1706, 1698, 1897, 1883, 1860, 1846, 1819, 1805, 1782,
+  1768, 1723, 1709, 1686, 1672, 1645, 1631, 1608, 1594, 1574, 1560, 1537, 1523,
+  1496, 1482, 1459, 1445, 1400, 1386, 1363, 1349, 1322, 1308, 1285, 1271, 1608,
+  1565, 1535, 1492, 1446, 1403, 1373, 1330, 1312, 1269, 1239, 1196, 1150, 1107,
+  1077, 1034, 1291, 1218, 1171, 1098, 1015, 942,  895,  822,  953,  850,  729,
+  626,  618,  431,  257,  257,  257,  257,  0,    255,  255,  255,  255,  429,
+  616,  624,  727,  848,  951,  820,  893,  940,  1013, 1096, 1169, 1216, 1289,
+  1032, 1075, 1105, 1148, 1194, 1237, 1267, 1310, 1328, 1371, 1401, 1444, 1490,
+  1533, 1563, 1606, 1269, 1283, 1306, 1320, 1347, 1361, 1384, 1398, 1443, 1457,
+  1480, 1494, 1521, 1535, 1558, 1572, 1592, 1606, 1629, 1643, 1670, 1684, 1707,
+  1721, 1766, 1780, 1803, 1817, 1844, 1858, 1881, 1895, 1696, 1704, 1710, 1718,
+  1728, 1736, 1742, 1750, 1769, 1777, 1783, 1791, 1801, 1809, 1815, 1823, 1845,
+  1853, 1859, 1867, 1877, 1885, 1891, 1899, 1918, 1926, 1932, 1940, 1950, 1958,
+  1964, 1972, 1998, 2006, 2012, 2020, 2030, 2038, 2044, 2052, 2071, 2079, 2085,
+  2093, 2103, 2111, 2117, 2125, 2147, 2155, 2161, 2169, 2179, 2187, 2193, 2201,
+  2220, 2228, 2234, 2242, 2252, 2260, 2266, 2274, 2140, 2148, 2154, 2162, 2172,
+  2180, 2186, 2194, 2213, 2221, 2227, 2235, 2245, 2253, 2259, 2267, 2289, 2297,
+  2303, 2311, 2321, 2329, 2335, 2343, 2362, 2370, 2376, 2384, 2394, 2402, 2408,
+  2416, 2442, 2450, 2456, 2464, 2474, 2482, 2488, 2496, 2515, 2523, 2529, 2537,
+  2547, 2555, 2561, 2569, 2591, 2599, 2605, 2613, 2623, 2631, 2637, 2645, 2664,
+  2672, 2678, 2686, 2696, 2704, 2710, 2718, 2516, 2524, 2530, 2538, 2548, 2556,
+  2562, 2570, 2589, 2597, 2603, 2611, 2621, 2629, 2635, 2643, 2665, 2673, 2679,
+  2687, 2697, 2705, 2711, 2719, 2738, 2746, 2752, 2760, 2770, 2778, 2784, 2792,
+  2818, 2826, 2832, 2840, 2850, 2858, 2864, 2872, 2891, 2899, 2905, 2913, 2923,
+  2931, 2937, 2945, 2967, 2975, 2981, 2989, 2999, 3007, 3013, 3021, 3040, 3048,
+  3054, 3062, 3072, 3080, 3086, 3094, 2960, 2968, 2974, 2982, 2992, 3000, 3006,
+  3014, 3033, 3041, 3047, 3055, 3065, 3073, 3079, 3087, 3109, 3117, 3123, 3131,
+  3141, 3149, 3155, 3163, 3182, 3190, 3196, 3204, 3214, 3222, 3228, 3236, 3262,
+  3270, 3276, 3284, 3294, 3302, 3308, 3316, 3335, 3343, 3349, 3357, 3367, 3375,
+  3381, 3389, 3411, 3419, 3425, 3433, 3443, 3451, 3457, 3465, 3484, 3492, 3498,
+  3506, 3516, 3524, 3530, 3538, 2807, 2815, 2821, 2829, 2839, 2847, 2853, 2861,
+  2880, 2888, 2894, 2902, 2912, 2920, 2926, 2934, 2956, 2964, 2970, 2978, 2988,
+  2996, 3002, 3010, 3029, 3037, 3043, 3051, 3061, 3069, 3075, 3083, 3109, 3117,
+  3123, 3131, 3141, 3149, 3155, 3163, 3182, 3190, 3196, 3204, 3214, 3222, 3228,
+  3236, 3258, 3266, 3272, 3280, 3290, 3298, 3304, 3312, 3331, 3339, 3345, 3353,
+  3363, 3371, 3377, 3385, 3251, 3259, 3265, 3273, 3283, 3291, 3297, 3305, 3324,
+  3332, 3338, 3346, 3356, 3364, 3370, 3378, 3400, 3408, 3414, 3422, 3432, 3440,
+  3446, 3454, 3473, 3481, 3487, 3495, 3505, 3513, 3519, 3527, 3553, 3561, 3567,
+  3575, 3585, 3593, 3599, 3607, 3626, 3634, 3640, 3648, 3658, 3666, 3672, 3680,
+  3702, 3710, 3716, 3724, 3734, 3742, 3748, 3756, 3775, 3783, 3789, 3797, 3807,
+  3815, 3821, 3829, 3627, 3635, 3641, 3649, 3659, 3667, 3673, 3681, 3700, 3708,
+  3714, 3722, 3732, 3740, 3746, 3754, 3776, 3784, 3790, 3798, 3808, 3816, 3822,
+  3830, 3849, 3857, 3863, 3871, 3881, 3889, 3895, 3903, 3929, 3937, 3943, 3951,
+  3961, 3969, 3975, 3983, 4002, 4010, 4016, 4024, 4034, 4042, 4048, 4056, 4078,
+  4086, 4092, 4100, 4110, 4118, 4124, 4132, 4151, 4159, 4165, 4173, 4183, 4191,
+  4197, 4205, 4071, 4079, 4085, 4093, 4103, 4111, 4117, 4125, 4144, 4152, 4158,
+  4166, 4176, 4184, 4190, 4198, 4220, 4228, 4234, 4242, 4252, 4260, 4266, 4274,
+  4293, 4301, 4307, 4315, 4325, 4333, 4339, 4347, 4373, 4381, 4387, 4395, 4405,
+  4413, 4419, 4427, 4446, 4454, 4460, 4468, 4478, 4486, 4492, 4500, 4522, 4530,
+  4536, 4544, 4554, 4562, 4568, 4576, 4595, 4603, 4609, 4617, 4627, 4635, 4641,
+  4649, 3742, 3750, 3756, 3764, 3774, 3782, 3788, 3796, 3815, 3823, 3829, 3837,
+  3847, 3855, 3861, 3869, 3891, 3899, 3905, 3913, 3923, 3931, 3937, 3945, 3964,
+  3972, 3978, 3986, 3996, 4004, 4010, 4018, 4044, 4052, 4058, 4066, 4076, 4084,
+  4090, 4098, 4117, 4125, 4131, 4139, 4149, 4157, 4163, 4171, 4193, 4201, 4207,
+  4215, 4225, 4233, 4239, 4247, 4266, 4274, 4280, 4288, 4298, 4306, 4312, 4320,
+  4186, 4194, 4200, 4208, 4218, 4226, 4232, 4240, 4259, 4267, 4273, 4281, 4291,
+  4299, 4305, 4313, 4335, 4343, 4349, 4357, 4367, 4375, 4381, 4389, 4408, 4416,
+  4422, 4430, 4440, 4448, 4454, 4462, 4488, 4496, 4502, 4510, 4520, 4528, 4534,
+  4542, 4561, 4569, 4575, 4583, 4593, 4601, 4607, 4615, 4637, 4645, 4651, 4659,
+  4669, 4677, 4683, 4691, 4710, 4718, 4724, 4732, 4742, 4750, 4756, 4764, 4562,
+  4570, 4576, 4584, 4594, 4602, 4608, 4616, 4635, 4643, 4649, 4657, 4667, 4675,
+  4681, 4689, 4711, 4719, 4725, 4733, 4743, 4751, 4757, 4765, 4784, 4792, 4798,
+  4806, 4816, 4824, 4830, 4838, 4864, 4872, 4878, 4886, 4896, 4904, 4910, 4918,
+  4937, 4945, 4951, 4959, 4969, 4977, 4983, 4991, 5013, 5021, 5027, 5035, 5045,
+  5053, 5059, 5067, 5086, 5094, 5100, 5108, 5118, 5126, 5132, 5140, 5006, 5014,
+  5020, 5028, 5038, 5046, 5052, 5060, 5079, 5087, 5093, 5101, 5111, 5119, 5125,
+  5133, 5155, 5163, 5169, 5177, 5187, 5195, 5201, 5209, 5228, 5236, 5242, 5250,
+  5260, 5268, 5274, 5282, 5308, 5316, 5322, 5330, 5340, 5348, 5354, 5362, 5381,
+  5389, 5395, 5403, 5413, 5421, 5427, 5435, 5457, 5465, 5471, 5479, 5489, 5497,
+  5503, 5511, 5530, 5538, 5544, 5552, 5562, 5570, 5576, 5584, 4853, 4861, 4867,
+  4875, 4885, 4893, 4899, 4907, 4926, 4934, 4940, 4948, 4958, 4966, 4972, 4980,
+  5002, 5010, 5016, 5024, 5034, 5042, 5048, 5056, 5075, 5083, 5089, 5097, 5107,
+  5115, 5121, 5129, 5155, 5163, 5169, 5177, 5187, 5195, 5201, 5209, 5228, 5236,
+  5242, 5250, 5260, 5268, 5274, 5282, 5304, 5312, 5318, 5326, 5336, 5344, 5350,
+  5358, 5377, 5385, 5391, 5399, 5409, 5417, 5423, 5431, 5297, 5305, 5311, 5319,
+  5329, 5337, 5343, 5351, 5370, 5378, 5384, 5392, 5402, 5410, 5416, 5424, 5446,
+  5454, 5460, 5468, 5478, 5486, 5492, 5500, 5519, 5527, 5533, 5541, 5551, 5559,
+  5565, 5573, 5599, 5607, 5613, 5621, 5631, 5639, 5645, 5653, 5672, 5680, 5686,
+  5694, 5704, 5712, 5718, 5726, 5748, 5756, 5762, 5770, 5780, 5788, 5794, 5802,
+  5821, 5829, 5835, 5843, 5853, 5861, 5867, 5875, 5673, 5681, 5687, 5695, 5705,
+  5713, 5719, 5727, 5746, 5754, 5760, 5768, 5778, 5786, 5792, 5800, 5822, 5830,
+  5836, 5844, 5854, 5862, 5868, 5876, 5895, 5903, 5909, 5917, 5927, 5935, 5941,
+  5949, 5975, 5983, 5989, 5997, 6007, 6015, 6021, 6029, 6048, 6056, 6062, 6070,
+  6080, 6088, 6094, 6102, 6124, 6132, 6138, 6146, 6156, 6164, 6170, 6178, 6197,
+  6205, 6211, 6219, 6229, 6237, 6243, 6251, 6117, 6125, 6131, 6139, 6149, 6157,
+  6163, 6171, 6190, 6198, 6204, 6212, 6222, 6230, 6236, 6244, 6266, 6274, 6280,
+  6288, 6298, 6306, 6312, 6320, 6339, 6347, 6353, 6361, 6371, 6379, 6385, 6393,
+  6419, 6427, 6433, 6441, 6451, 6459, 6465, 6473, 6492, 6500, 6506, 6514, 6524,
+  6532, 6538, 6546, 6568, 6576, 6582, 6590, 6600, 6608, 6614, 6622, 6641, 6649,
+  6655, 6663, 6673, 6681, 6687, 6695, 3742, 3750, 3756, 3764, 3774, 3782, 3788,
+  3796, 3815, 3823, 3829, 3837, 3847, 3855, 3861, 3869, 3891, 3899, 3905, 3913,
+  3923, 3931, 3937, 3945, 3964, 3972, 3978, 3986, 3996, 4004, 4010, 4018, 4044,
+  4052, 4058, 4066, 4076, 4084, 4090, 4098, 4117, 4125, 4131, 4139, 4149, 4157,
+  4163, 4171, 4193, 4201, 4207, 4215, 4225, 4233, 4239, 4247, 4266, 4274, 4280,
+  4288, 4298, 4306, 4312, 4320, 4186, 4194, 4200, 4208, 4218, 4226, 4232, 4240,
+  4259, 4267, 4273, 4281, 4291, 4299, 4305, 4313, 4335, 4343, 4349, 4357, 4367,
+  4375, 4381, 4389, 4408, 4416, 4422, 4430, 4440, 4448, 4454, 4462, 4488, 4496,
+  4502, 4510, 4520, 4528, 4534, 4542, 4561, 4569, 4575, 4583, 4593, 4601, 4607,
+  4615, 4637, 4645, 4651, 4659, 4669, 4677, 4683, 4691, 4710, 4718, 4724, 4732,
+  4742, 4750, 4756, 4764, 4562, 4570, 4576, 4584, 4594, 4602, 4608, 4616, 4635,
+  4643, 4649, 4657, 4667, 4675, 4681, 4689, 4711, 4719, 4725, 4733, 4743, 4751,
+  4757, 4765, 4784, 4792, 4798, 4806, 4816, 4824, 4830, 4838, 4864, 4872, 4878,
+  4886, 4896, 4904, 4910, 4918, 4937, 4945, 4951, 4959, 4969, 4977, 4983, 4991,
+  5013, 5021, 5027, 5035, 5045, 5053, 5059, 5067, 5086, 5094, 5100, 5108, 5118,
+  5126, 5132, 5140, 5006, 5014, 5020, 5028, 5038, 5046, 5052, 5060, 5079, 5087,
+  5093, 5101, 5111, 5119, 5125, 5133, 5155, 5163, 5169, 5177, 5187, 5195, 5201,
+  5209, 5228, 5236, 5242, 5250, 5260, 5268, 5274, 5282, 5308, 5316, 5322, 5330,
+  5340, 5348, 5354, 5362, 5381, 5389, 5395, 5403, 5413, 5421, 5427, 5435, 5457,
+  5465, 5471, 5479, 5489, 5497, 5503, 5511, 5530, 5538, 5544, 5552, 5562, 5570,
+  5576, 5584, 4853, 4861, 4867, 4875, 4885, 4893, 4899, 4907, 4926, 4934, 4940,
+  4948, 4958, 4966, 4972, 4980, 5002, 5010, 5016, 5024, 5034, 5042, 5048, 5056,
+  5075, 5083, 5089, 5097, 5107, 5115, 5121, 5129, 5155, 5163, 5169, 5177, 5187,
+  5195, 5201, 5209, 5228, 5236, 5242, 5250, 5260, 5268, 5274, 5282, 5304, 5312,
+  5318, 5326, 5336, 5344, 5350, 5358, 5377, 5385, 5391, 5399, 5409, 5417, 5423,
+  5431, 5297, 5305, 5311, 5319, 5329, 5337, 5343, 5351, 5370, 5378, 5384, 5392,
+  5402, 5410, 5416, 5424, 5446, 5454, 5460, 5468, 5478, 5486, 5492, 5500, 5519,
+  5527, 5533, 5541, 5551, 5559, 5565, 5573, 5599, 5607, 5613, 5621, 5631, 5639,
+  5645, 5653, 5672, 5680, 5686, 5694, 5704, 5712, 5718, 5726, 5748, 5756, 5762,
+  5770, 5780, 5788, 5794, 5802, 5821, 5829, 5835, 5843, 5853, 5861, 5867, 5875,
+  5673, 5681, 5687, 5695, 5705, 5713, 5719, 5727, 5746, 5754, 5760, 5768, 5778,
+  5786, 5792, 5800, 5822, 5830, 5836, 5844, 5854, 5862, 5868, 5876, 5895, 5903,
+  5909, 5917, 5927, 5935, 5941, 5949, 5975, 5983, 5989, 5997, 6007, 6015, 6021,
+  6029, 6048, 6056, 6062, 6070, 6080, 6088, 6094, 6102, 6124, 6132, 6138, 6146,
+  6156, 6164, 6170, 6178, 6197, 6205, 6211, 6219, 6229, 6237, 6243, 6251, 6117,
+  6125, 6131, 6139, 6149, 6157, 6163, 6171, 6190, 6198, 6204, 6212, 6222, 6230,
+  6236, 6244, 6266, 6274, 6280, 6288, 6298, 6306, 6312, 6320, 6339, 6347, 6353,
+  6361, 6371, 6379, 6385, 6393, 6419, 6427, 6433, 6441, 6451, 6459, 6465, 6473,
+  6492, 6500, 6506, 6514, 6524, 6532, 6538, 6546, 6568, 6576, 6582, 6590, 6600,
+  6608, 6614, 6622, 6641, 6649, 6655, 6663, 6673, 6681, 6687, 6695, 5788, 5796,
+  5802, 5810, 5820, 5828, 5834, 5842, 5861, 5869, 5875, 5883, 5893, 5901, 5907,
+  5915, 5937, 5945, 5951, 5959, 5969, 5977, 5983, 5991, 6010, 6018, 6024, 6032,
+  6042, 6050, 6056, 6064, 6090, 6098, 6104, 6112, 6122, 6130, 6136, 6144, 6163,
+  6171, 6177, 6185, 6195, 6203, 6209, 6217, 6239, 6247, 6253, 6261, 6271, 6279,
+  6285, 6293, 6312, 6320, 6326, 6334, 6344, 6352, 6358, 6366, 6232, 6240, 6246,
+  6254, 6264, 6272, 6278, 6286, 6305, 6313, 6319, 6327, 6337, 6345, 6351, 6359,
+  6381, 6389, 6395, 6403, 6413, 6421, 6427, 6435, 6454, 6462, 6468, 6476, 6486,
+  6494, 6500, 6508, 6534, 6542, 6548, 6556, 6566, 6574, 6580, 6588, 6607, 6615,
+  6621, 6629, 6639, 6647, 6653, 6661, 6683, 6691, 6697, 6705, 6715, 6723, 6729,
+  6737, 6756, 6764, 6770, 6778, 6788, 6796, 6802, 6810, 6608, 6616, 6622, 6630,
+  6640, 6648, 6654, 6662, 6681, 6689, 6695, 6703, 6713, 6721, 6727, 6735, 6757,
+  6765, 6771, 6779, 6789, 6797, 6803, 6811, 6830, 6838, 6844, 6852, 6862, 6870,
+  6876, 6884, 6910, 6918, 6924, 6932, 6942, 6950, 6956, 6964, 6983, 6991, 6997,
+  7005, 7015, 7023, 7029, 7037, 7059, 7067, 7073, 7081, 7091, 7099, 7105, 7113,
+  7132, 7140, 7146, 7154, 7164, 7172, 7178, 7186, 7052, 7060, 7066, 7074, 7084,
+  7092, 7098, 7106, 7125, 7133, 7139, 7147, 7157, 7165, 7171, 7179, 7201, 7209,
+  7215, 7223, 7233, 7241, 7247, 7255, 7274, 7282, 7288, 7296, 7306, 7314, 7320,
+  7328, 7354, 7362, 7368, 7376, 7386, 7394, 7400, 7408, 7427, 7435, 7441, 7449,
+  7459, 7467, 7473, 7481, 7503, 7511, 7517, 7525, 7535, 7543, 7549, 7557, 7576,
+  7584, 7590, 7598, 7608, 7616, 7622, 7630, 6899, 6907, 6913, 6921, 6931, 6939,
+  6945, 6953, 6972, 6980, 6986, 6994, 7004, 7012, 7018, 7026, 7048, 7056, 7062,
+  7070, 7080, 7088, 7094, 7102, 7121, 7129, 7135, 7143, 7153, 7161, 7167, 7175,
+  7201, 7209, 7215, 7223, 7233, 7241, 7247, 7255, 7274, 7282, 7288, 7296, 7306,
+  7314, 7320, 7328, 7350, 7358, 7364, 7372, 7382, 7390, 7396, 7404, 7423, 7431,
+  7437, 7445, 7455, 7463, 7469, 7477, 7343, 7351, 7357, 7365, 7375, 7383, 7389,
+  7397, 7416, 7424, 7430, 7438, 7448, 7456, 7462, 7470, 7492, 7500, 7506, 7514,
+  7524, 7532, 7538, 7546, 7565, 7573, 7579, 7587, 7597, 7605, 7611, 7619, 7645,
+  7653, 7659, 7667, 7677, 7685, 7691, 7699, 7718, 7726, 7732, 7740, 7750, 7758,
+  7764, 7772, 7794, 7802, 7808, 7816, 7826, 7834, 7840, 7848, 7867, 7875, 7881,
+  7889, 7899, 7907, 7913, 7921, 7719, 7727, 7733, 7741, 7751, 7759, 7765, 7773,
+  7792, 7800, 7806, 7814, 7824, 7832, 7838, 7846, 7868, 7876, 7882, 7890, 7900,
+  7908, 7914, 7922, 7941, 7949, 7955, 7963, 7973, 7981, 7987, 7995, 8021, 8029,
+  8035, 8043, 8053, 8061, 8067, 8075, 8094, 8102, 8108, 8116, 8126, 8134, 8140,
+  8148, 8170, 8178, 8184, 8192, 8202, 8210, 8216, 8224, 8243, 8251, 8257, 8265,
+  8275
 };
 
 #ifdef __cplusplus
--- a/vp8/encoder/dct_value_tokens.h
+++ b/vp8/encoder/dct_value_tokens.h
@@ -18,691 +18,827 @@
 /* Generated file, included by tokenize.c  */
 /* Values generated by fill_value_tokens() */
 
-static const TOKENVALUE dct_value_tokens[2048*2] =
-{
-    {10, 3963}, {10, 3961}, {10, 3959}, {10, 3957}, {10, 3955}, {10, 3953},
-    {10, 3951}, {10, 3949}, {10, 3947}, {10, 3945}, {10, 3943}, {10, 3941},
-    {10, 3939}, {10, 3937}, {10, 3935}, {10, 3933}, {10, 3931}, {10, 3929},
-    {10, 3927}, {10, 3925}, {10, 3923}, {10, 3921}, {10, 3919}, {10, 3917},
-    {10, 3915}, {10, 3913}, {10, 3911}, {10, 3909}, {10, 3907}, {10, 3905},
-    {10, 3903}, {10, 3901}, {10, 3899}, {10, 3897}, {10, 3895}, {10, 3893},
-    {10, 3891}, {10, 3889}, {10, 3887}, {10, 3885}, {10, 3883}, {10, 3881},
-    {10, 3879}, {10, 3877}, {10, 3875}, {10, 3873}, {10, 3871}, {10, 3869},
-    {10, 3867}, {10, 3865}, {10, 3863}, {10, 3861}, {10, 3859}, {10, 3857},
-    {10, 3855}, {10, 3853}, {10, 3851}, {10, 3849}, {10, 3847}, {10, 3845},
-    {10, 3843}, {10, 3841}, {10, 3839}, {10, 3837}, {10, 3835}, {10, 3833},
-    {10, 3831}, {10, 3829}, {10, 3827}, {10, 3825}, {10, 3823}, {10, 3821},
-    {10, 3819}, {10, 3817}, {10, 3815}, {10, 3813}, {10, 3811}, {10, 3809},
-    {10, 3807}, {10, 3805}, {10, 3803}, {10, 3801}, {10, 3799}, {10, 3797},
-    {10, 3795}, {10, 3793}, {10, 3791}, {10, 3789}, {10, 3787}, {10, 3785},
-    {10, 3783}, {10, 3781}, {10, 3779}, {10, 3777}, {10, 3775}, {10, 3773},
-    {10, 3771}, {10, 3769}, {10, 3767}, {10, 3765}, {10, 3763}, {10, 3761},
-    {10, 3759}, {10, 3757}, {10, 3755}, {10, 3753}, {10, 3751}, {10, 3749},
-    {10, 3747}, {10, 3745}, {10, 3743}, {10, 3741}, {10, 3739}, {10, 3737},
-    {10, 3735}, {10, 3733}, {10, 3731}, {10, 3729}, {10, 3727}, {10, 3725},
-    {10, 3723}, {10, 3721}, {10, 3719}, {10, 3717}, {10, 3715}, {10, 3713},
-    {10, 3711}, {10, 3709}, {10, 3707}, {10, 3705}, {10, 3703}, {10, 3701},
-    {10, 3699}, {10, 3697}, {10, 3695}, {10, 3693}, {10, 3691}, {10, 3689},
-    {10, 3687}, {10, 3685}, {10, 3683}, {10, 3681}, {10, 3679}, {10, 3677},
-    {10, 3675}, {10, 3673}, {10, 3671}, {10, 3669}, {10, 3667}, {10, 3665},
-    {10, 3663}, {10, 3661}, {10, 3659}, {10, 3657}, {10, 3655}, {10, 3653},
-    {10, 3651}, {10, 3649}, {10, 3647}, {10, 3645}, {10, 3643}, {10, 3641},
-    {10, 3639}, {10, 3637}, {10, 3635}, {10, 3633}, {10, 3631}, {10, 3629},
-    {10, 3627}, {10, 3625}, {10, 3623}, {10, 3621}, {10, 3619}, {10, 3617},
-    {10, 3615}, {10, 3613}, {10, 3611}, {10, 3609}, {10, 3607}, {10, 3605},
-    {10, 3603}, {10, 3601}, {10, 3599}, {10, 3597}, {10, 3595}, {10, 3593},
-    {10, 3591}, {10, 3589}, {10, 3587}, {10, 3585}, {10, 3583}, {10, 3581},
-    {10, 3579}, {10, 3577}, {10, 3575}, {10, 3573}, {10, 3571}, {10, 3569},
-    {10, 3567}, {10, 3565}, {10, 3563}, {10, 3561}, {10, 3559}, {10, 3557},
-    {10, 3555}, {10, 3553}, {10, 3551}, {10, 3549}, {10, 3547}, {10, 3545},
-    {10, 3543}, {10, 3541}, {10, 3539}, {10, 3537}, {10, 3535}, {10, 3533},
-    {10, 3531}, {10, 3529}, {10, 3527}, {10, 3525}, {10, 3523}, {10, 3521},
-    {10, 3519}, {10, 3517}, {10, 3515}, {10, 3513}, {10, 3511}, {10, 3509},
-    {10, 3507}, {10, 3505}, {10, 3503}, {10, 3501}, {10, 3499}, {10, 3497},
-    {10, 3495}, {10, 3493}, {10, 3491}, {10, 3489}, {10, 3487}, {10, 3485},
-    {10, 3483}, {10, 3481}, {10, 3479}, {10, 3477}, {10, 3475}, {10, 3473},
-    {10, 3471}, {10, 3469}, {10, 3467}, {10, 3465}, {10, 3463}, {10, 3461},
-    {10, 3459}, {10, 3457}, {10, 3455}, {10, 3453}, {10, 3451}, {10, 3449},
-    {10, 3447}, {10, 3445}, {10, 3443}, {10, 3441}, {10, 3439}, {10, 3437},
-    {10, 3435}, {10, 3433}, {10, 3431}, {10, 3429}, {10, 3427}, {10, 3425},
-    {10, 3423}, {10, 3421}, {10, 3419}, {10, 3417}, {10, 3415}, {10, 3413},
-    {10, 3411}, {10, 3409}, {10, 3407}, {10, 3405}, {10, 3403}, {10, 3401},
-    {10, 3399}, {10, 3397}, {10, 3395}, {10, 3393}, {10, 3391}, {10, 3389},
-    {10, 3387}, {10, 3385}, {10, 3383}, {10, 3381}, {10, 3379}, {10, 3377},
-    {10, 3375}, {10, 3373}, {10, 3371}, {10, 3369}, {10, 3367}, {10, 3365},
-    {10, 3363}, {10, 3361}, {10, 3359}, {10, 3357}, {10, 3355}, {10, 3353},
-    {10, 3351}, {10, 3349}, {10, 3347}, {10, 3345}, {10, 3343}, {10, 3341},
-    {10, 3339}, {10, 3337}, {10, 3335}, {10, 3333}, {10, 3331}, {10, 3329},
-    {10, 3327}, {10, 3325}, {10, 3323}, {10, 3321}, {10, 3319}, {10, 3317},
-    {10, 3315}, {10, 3313}, {10, 3311}, {10, 3309}, {10, 3307}, {10, 3305},
-    {10, 3303}, {10, 3301}, {10, 3299}, {10, 3297}, {10, 3295}, {10, 3293},
-    {10, 3291}, {10, 3289}, {10, 3287}, {10, 3285}, {10, 3283}, {10, 3281},
-    {10, 3279}, {10, 3277}, {10, 3275}, {10, 3273}, {10, 3271}, {10, 3269},
-    {10, 3267}, {10, 3265}, {10, 3263}, {10, 3261}, {10, 3259}, {10, 3257},
-    {10, 3255}, {10, 3253}, {10, 3251}, {10, 3249}, {10, 3247}, {10, 3245},
-    {10, 3243}, {10, 3241}, {10, 3239}, {10, 3237}, {10, 3235}, {10, 3233},
-    {10, 3231}, {10, 3229}, {10, 3227}, {10, 3225}, {10, 3223}, {10, 3221},
-    {10, 3219}, {10, 3217}, {10, 3215}, {10, 3213}, {10, 3211}, {10, 3209},
-    {10, 3207}, {10, 3205}, {10, 3203}, {10, 3201}, {10, 3199}, {10, 3197},
-    {10, 3195}, {10, 3193}, {10, 3191}, {10, 3189}, {10, 3187}, {10, 3185},
-    {10, 3183}, {10, 3181}, {10, 3179}, {10, 3177}, {10, 3175}, {10, 3173},
-    {10, 3171}, {10, 3169}, {10, 3167}, {10, 3165}, {10, 3163}, {10, 3161},
-    {10, 3159}, {10, 3157}, {10, 3155}, {10, 3153}, {10, 3151}, {10, 3149},
-    {10, 3147}, {10, 3145}, {10, 3143}, {10, 3141}, {10, 3139}, {10, 3137},
-    {10, 3135}, {10, 3133}, {10, 3131}, {10, 3129}, {10, 3127}, {10, 3125},
-    {10, 3123}, {10, 3121}, {10, 3119}, {10, 3117}, {10, 3115}, {10, 3113},
-    {10, 3111}, {10, 3109}, {10, 3107}, {10, 3105}, {10, 3103}, {10, 3101},
-    {10, 3099}, {10, 3097}, {10, 3095}, {10, 3093}, {10, 3091}, {10, 3089},
-    {10, 3087}, {10, 3085}, {10, 3083}, {10, 3081}, {10, 3079}, {10, 3077},
-    {10, 3075}, {10, 3073}, {10, 3071}, {10, 3069}, {10, 3067}, {10, 3065},
-    {10, 3063}, {10, 3061}, {10, 3059}, {10, 3057}, {10, 3055}, {10, 3053},
-    {10, 3051}, {10, 3049}, {10, 3047}, {10, 3045}, {10, 3043}, {10, 3041},
-    {10, 3039}, {10, 3037}, {10, 3035}, {10, 3033}, {10, 3031}, {10, 3029},
-    {10, 3027}, {10, 3025}, {10, 3023}, {10, 3021}, {10, 3019}, {10, 3017},
-    {10, 3015}, {10, 3013}, {10, 3011}, {10, 3009}, {10, 3007}, {10, 3005},
-    {10, 3003}, {10, 3001}, {10, 2999}, {10, 2997}, {10, 2995}, {10, 2993},
-    {10, 2991}, {10, 2989}, {10, 2987}, {10, 2985}, {10, 2983}, {10, 2981},
-    {10, 2979}, {10, 2977}, {10, 2975}, {10, 2973}, {10, 2971}, {10, 2969},
-    {10, 2967}, {10, 2965}, {10, 2963}, {10, 2961}, {10, 2959}, {10, 2957},
-    {10, 2955}, {10, 2953}, {10, 2951}, {10, 2949}, {10, 2947}, {10, 2945},
-    {10, 2943}, {10, 2941}, {10, 2939}, {10, 2937}, {10, 2935}, {10, 2933},
-    {10, 2931}, {10, 2929}, {10, 2927}, {10, 2925}, {10, 2923}, {10, 2921},
-    {10, 2919}, {10, 2917}, {10, 2915}, {10, 2913}, {10, 2911}, {10, 2909},
-    {10, 2907}, {10, 2905}, {10, 2903}, {10, 2901}, {10, 2899}, {10, 2897},
-    {10, 2895}, {10, 2893}, {10, 2891}, {10, 2889}, {10, 2887}, {10, 2885},
-    {10, 2883}, {10, 2881}, {10, 2879}, {10, 2877}, {10, 2875}, {10, 2873},
-    {10, 2871}, {10, 2869}, {10, 2867}, {10, 2865}, {10, 2863}, {10, 2861},
-    {10, 2859}, {10, 2857}, {10, 2855}, {10, 2853}, {10, 2851}, {10, 2849},
-    {10, 2847}, {10, 2845}, {10, 2843}, {10, 2841}, {10, 2839}, {10, 2837},
-    {10, 2835}, {10, 2833}, {10, 2831}, {10, 2829}, {10, 2827}, {10, 2825},
-    {10, 2823}, {10, 2821}, {10, 2819}, {10, 2817}, {10, 2815}, {10, 2813},
-    {10, 2811}, {10, 2809}, {10, 2807}, {10, 2805}, {10, 2803}, {10, 2801},
-    {10, 2799}, {10, 2797}, {10, 2795}, {10, 2793}, {10, 2791}, {10, 2789},
-    {10, 2787}, {10, 2785}, {10, 2783}, {10, 2781}, {10, 2779}, {10, 2777},
-    {10, 2775}, {10, 2773}, {10, 2771}, {10, 2769}, {10, 2767}, {10, 2765},
-    {10, 2763}, {10, 2761}, {10, 2759}, {10, 2757}, {10, 2755}, {10, 2753},
-    {10, 2751}, {10, 2749}, {10, 2747}, {10, 2745}, {10, 2743}, {10, 2741},
-    {10, 2739}, {10, 2737}, {10, 2735}, {10, 2733}, {10, 2731}, {10, 2729},
-    {10, 2727}, {10, 2725}, {10, 2723}, {10, 2721}, {10, 2719}, {10, 2717},
-    {10, 2715}, {10, 2713}, {10, 2711}, {10, 2709}, {10, 2707}, {10, 2705},
-    {10, 2703}, {10, 2701}, {10, 2699}, {10, 2697}, {10, 2695}, {10, 2693},
-    {10, 2691}, {10, 2689}, {10, 2687}, {10, 2685}, {10, 2683}, {10, 2681},
-    {10, 2679}, {10, 2677}, {10, 2675}, {10, 2673}, {10, 2671}, {10, 2669},
-    {10, 2667}, {10, 2665}, {10, 2663}, {10, 2661}, {10, 2659}, {10, 2657},
-    {10, 2655}, {10, 2653}, {10, 2651}, {10, 2649}, {10, 2647}, {10, 2645},
-    {10, 2643}, {10, 2641}, {10, 2639}, {10, 2637}, {10, 2635}, {10, 2633},
-    {10, 2631}, {10, 2629}, {10, 2627}, {10, 2625}, {10, 2623}, {10, 2621},
-    {10, 2619}, {10, 2617}, {10, 2615}, {10, 2613}, {10, 2611}, {10, 2609},
-    {10, 2607}, {10, 2605}, {10, 2603}, {10, 2601}, {10, 2599}, {10, 2597},
-    {10, 2595}, {10, 2593}, {10, 2591}, {10, 2589}, {10, 2587}, {10, 2585},
-    {10, 2583}, {10, 2581}, {10, 2579}, {10, 2577}, {10, 2575}, {10, 2573},
-    {10, 2571}, {10, 2569}, {10, 2567}, {10, 2565}, {10, 2563}, {10, 2561},
-    {10, 2559}, {10, 2557}, {10, 2555}, {10, 2553}, {10, 2551}, {10, 2549},
-    {10, 2547}, {10, 2545}, {10, 2543}, {10, 2541}, {10, 2539}, {10, 2537},
-    {10, 2535}, {10, 2533}, {10, 2531}, {10, 2529}, {10, 2527}, {10, 2525},
-    {10, 2523}, {10, 2521}, {10, 2519}, {10, 2517}, {10, 2515}, {10, 2513},
-    {10, 2511}, {10, 2509}, {10, 2507}, {10, 2505}, {10, 2503}, {10, 2501},
-    {10, 2499}, {10, 2497}, {10, 2495}, {10, 2493}, {10, 2491}, {10, 2489},
-    {10, 2487}, {10, 2485}, {10, 2483}, {10, 2481}, {10, 2479}, {10, 2477},
-    {10, 2475}, {10, 2473}, {10, 2471}, {10, 2469}, {10, 2467}, {10, 2465},
-    {10, 2463}, {10, 2461}, {10, 2459}, {10, 2457}, {10, 2455}, {10, 2453},
-    {10, 2451}, {10, 2449}, {10, 2447}, {10, 2445}, {10, 2443}, {10, 2441},
-    {10, 2439}, {10, 2437}, {10, 2435}, {10, 2433}, {10, 2431}, {10, 2429},
-    {10, 2427}, {10, 2425}, {10, 2423}, {10, 2421}, {10, 2419}, {10, 2417},
-    {10, 2415}, {10, 2413}, {10, 2411}, {10, 2409}, {10, 2407}, {10, 2405},
-    {10, 2403}, {10, 2401}, {10, 2399}, {10, 2397}, {10, 2395}, {10, 2393},
-    {10, 2391}, {10, 2389}, {10, 2387}, {10, 2385}, {10, 2383}, {10, 2381},
-    {10, 2379}, {10, 2377}, {10, 2375}, {10, 2373}, {10, 2371}, {10, 2369},
-    {10, 2367}, {10, 2365}, {10, 2363}, {10, 2361}, {10, 2359}, {10, 2357},
-    {10, 2355}, {10, 2353}, {10, 2351}, {10, 2349}, {10, 2347}, {10, 2345},
-    {10, 2343}, {10, 2341}, {10, 2339}, {10, 2337}, {10, 2335}, {10, 2333},
-    {10, 2331}, {10, 2329}, {10, 2327}, {10, 2325}, {10, 2323}, {10, 2321},
-    {10, 2319}, {10, 2317}, {10, 2315}, {10, 2313}, {10, 2311}, {10, 2309},
-    {10, 2307}, {10, 2305}, {10, 2303}, {10, 2301}, {10, 2299}, {10, 2297},
-    {10, 2295}, {10, 2293}, {10, 2291}, {10, 2289}, {10, 2287}, {10, 2285},
-    {10, 2283}, {10, 2281}, {10, 2279}, {10, 2277}, {10, 2275}, {10, 2273},
-    {10, 2271}, {10, 2269}, {10, 2267}, {10, 2265}, {10, 2263}, {10, 2261},
-    {10, 2259}, {10, 2257}, {10, 2255}, {10, 2253}, {10, 2251}, {10, 2249},
-    {10, 2247}, {10, 2245}, {10, 2243}, {10, 2241}, {10, 2239}, {10, 2237},
-    {10, 2235}, {10, 2233}, {10, 2231}, {10, 2229}, {10, 2227}, {10, 2225},
-    {10, 2223}, {10, 2221}, {10, 2219}, {10, 2217}, {10, 2215}, {10, 2213},
-    {10, 2211}, {10, 2209}, {10, 2207}, {10, 2205}, {10, 2203}, {10, 2201},
-    {10, 2199}, {10, 2197}, {10, 2195}, {10, 2193}, {10, 2191}, {10, 2189},
-    {10, 2187}, {10, 2185}, {10, 2183}, {10, 2181}, {10, 2179}, {10, 2177},
-    {10, 2175}, {10, 2173}, {10, 2171}, {10, 2169}, {10, 2167}, {10, 2165},
-    {10, 2163}, {10, 2161}, {10, 2159}, {10, 2157}, {10, 2155}, {10, 2153},
-    {10, 2151}, {10, 2149}, {10, 2147}, {10, 2145}, {10, 2143}, {10, 2141},
-    {10, 2139}, {10, 2137}, {10, 2135}, {10, 2133}, {10, 2131}, {10, 2129},
-    {10, 2127}, {10, 2125}, {10, 2123}, {10, 2121}, {10, 2119}, {10, 2117},
-    {10, 2115}, {10, 2113}, {10, 2111}, {10, 2109}, {10, 2107}, {10, 2105},
-    {10, 2103}, {10, 2101}, {10, 2099}, {10, 2097}, {10, 2095}, {10, 2093},
-    {10, 2091}, {10, 2089}, {10, 2087}, {10, 2085}, {10, 2083}, {10, 2081},
-    {10, 2079}, {10, 2077}, {10, 2075}, {10, 2073}, {10, 2071}, {10, 2069},
-    {10, 2067}, {10, 2065}, {10, 2063}, {10, 2061}, {10, 2059}, {10, 2057},
-    {10, 2055}, {10, 2053}, {10, 2051}, {10, 2049}, {10, 2047}, {10, 2045},
-    {10, 2043}, {10, 2041}, {10, 2039}, {10, 2037}, {10, 2035}, {10, 2033},
-    {10, 2031}, {10, 2029}, {10, 2027}, {10, 2025}, {10, 2023}, {10, 2021},
-    {10, 2019}, {10, 2017}, {10, 2015}, {10, 2013}, {10, 2011}, {10, 2009},
-    {10, 2007}, {10, 2005}, {10, 2003}, {10, 2001}, {10, 1999}, {10, 1997},
-    {10, 1995}, {10, 1993}, {10, 1991}, {10, 1989}, {10, 1987}, {10, 1985},
-    {10, 1983}, {10, 1981}, {10, 1979}, {10, 1977}, {10, 1975}, {10, 1973},
-    {10, 1971}, {10, 1969}, {10, 1967}, {10, 1965}, {10, 1963}, {10, 1961},
-    {10, 1959}, {10, 1957}, {10, 1955}, {10, 1953}, {10, 1951}, {10, 1949},
-    {10, 1947}, {10, 1945}, {10, 1943}, {10, 1941}, {10, 1939}, {10, 1937},
-    {10, 1935}, {10, 1933}, {10, 1931}, {10, 1929}, {10, 1927}, {10, 1925},
-    {10, 1923}, {10, 1921}, {10, 1919}, {10, 1917}, {10, 1915}, {10, 1913},
-    {10, 1911}, {10, 1909}, {10, 1907}, {10, 1905}, {10, 1903}, {10, 1901},
-    {10, 1899}, {10, 1897}, {10, 1895}, {10, 1893}, {10, 1891}, {10, 1889},
-    {10, 1887}, {10, 1885}, {10, 1883}, {10, 1881}, {10, 1879}, {10, 1877},
-    {10, 1875}, {10, 1873}, {10, 1871}, {10, 1869}, {10, 1867}, {10, 1865},
-    {10, 1863}, {10, 1861}, {10, 1859}, {10, 1857}, {10, 1855}, {10, 1853},
-    {10, 1851}, {10, 1849}, {10, 1847}, {10, 1845}, {10, 1843}, {10, 1841},
-    {10, 1839}, {10, 1837}, {10, 1835}, {10, 1833}, {10, 1831}, {10, 1829},
-    {10, 1827}, {10, 1825}, {10, 1823}, {10, 1821}, {10, 1819}, {10, 1817},
-    {10, 1815}, {10, 1813}, {10, 1811}, {10, 1809}, {10, 1807}, {10, 1805},
-    {10, 1803}, {10, 1801}, {10, 1799}, {10, 1797}, {10, 1795}, {10, 1793},
-    {10, 1791}, {10, 1789}, {10, 1787}, {10, 1785}, {10, 1783}, {10, 1781},
-    {10, 1779}, {10, 1777}, {10, 1775}, {10, 1773}, {10, 1771}, {10, 1769},
-    {10, 1767}, {10, 1765}, {10, 1763}, {10, 1761}, {10, 1759}, {10, 1757},
-    {10, 1755}, {10, 1753}, {10, 1751}, {10, 1749}, {10, 1747}, {10, 1745},
-    {10, 1743}, {10, 1741}, {10, 1739}, {10, 1737}, {10, 1735}, {10, 1733},
-    {10, 1731}, {10, 1729}, {10, 1727}, {10, 1725}, {10, 1723}, {10, 1721},
-    {10, 1719}, {10, 1717}, {10, 1715}, {10, 1713}, {10, 1711}, {10, 1709},
-    {10, 1707}, {10, 1705}, {10, 1703}, {10, 1701}, {10, 1699}, {10, 1697},
-    {10, 1695}, {10, 1693}, {10, 1691}, {10, 1689}, {10, 1687}, {10, 1685},
-    {10, 1683}, {10, 1681}, {10, 1679}, {10, 1677}, {10, 1675}, {10, 1673},
-    {10, 1671}, {10, 1669}, {10, 1667}, {10, 1665}, {10, 1663}, {10, 1661},
-    {10, 1659}, {10, 1657}, {10, 1655}, {10, 1653}, {10, 1651}, {10, 1649},
-    {10, 1647}, {10, 1645}, {10, 1643}, {10, 1641}, {10, 1639}, {10, 1637},
-    {10, 1635}, {10, 1633}, {10, 1631}, {10, 1629}, {10, 1627}, {10, 1625},
-    {10, 1623}, {10, 1621}, {10, 1619}, {10, 1617}, {10, 1615}, {10, 1613},
-    {10, 1611}, {10, 1609}, {10, 1607}, {10, 1605}, {10, 1603}, {10, 1601},
-    {10, 1599}, {10, 1597}, {10, 1595}, {10, 1593}, {10, 1591}, {10, 1589},
-    {10, 1587}, {10, 1585}, {10, 1583}, {10, 1581}, {10, 1579}, {10, 1577},
-    {10, 1575}, {10, 1573}, {10, 1571}, {10, 1569}, {10, 1567}, {10, 1565},
-    {10, 1563}, {10, 1561}, {10, 1559}, {10, 1557}, {10, 1555}, {10, 1553},
-    {10, 1551}, {10, 1549}, {10, 1547}, {10, 1545}, {10, 1543}, {10, 1541},
-    {10, 1539}, {10, 1537}, {10, 1535}, {10, 1533}, {10, 1531}, {10, 1529},
-    {10, 1527}, {10, 1525}, {10, 1523}, {10, 1521}, {10, 1519}, {10, 1517},
-    {10, 1515}, {10, 1513}, {10, 1511}, {10, 1509}, {10, 1507}, {10, 1505},
-    {10, 1503}, {10, 1501}, {10, 1499}, {10, 1497}, {10, 1495}, {10, 1493},
-    {10, 1491}, {10, 1489}, {10, 1487}, {10, 1485}, {10, 1483}, {10, 1481},
-    {10, 1479}, {10, 1477}, {10, 1475}, {10, 1473}, {10, 1471}, {10, 1469},
-    {10, 1467}, {10, 1465}, {10, 1463}, {10, 1461}, {10, 1459}, {10, 1457},
-    {10, 1455}, {10, 1453}, {10, 1451}, {10, 1449}, {10, 1447}, {10, 1445},
-    {10, 1443}, {10, 1441}, {10, 1439}, {10, 1437}, {10, 1435}, {10, 1433},
-    {10, 1431}, {10, 1429}, {10, 1427}, {10, 1425}, {10, 1423}, {10, 1421},
-    {10, 1419}, {10, 1417}, {10, 1415}, {10, 1413}, {10, 1411}, {10, 1409},
-    {10, 1407}, {10, 1405}, {10, 1403}, {10, 1401}, {10, 1399}, {10, 1397},
-    {10, 1395}, {10, 1393}, {10, 1391}, {10, 1389}, {10, 1387}, {10, 1385},
-    {10, 1383}, {10, 1381}, {10, 1379}, {10, 1377}, {10, 1375}, {10, 1373},
-    {10, 1371}, {10, 1369}, {10, 1367}, {10, 1365}, {10, 1363}, {10, 1361},
-    {10, 1359}, {10, 1357}, {10, 1355}, {10, 1353}, {10, 1351}, {10, 1349},
-    {10, 1347}, {10, 1345}, {10, 1343}, {10, 1341}, {10, 1339}, {10, 1337},
-    {10, 1335}, {10, 1333}, {10, 1331}, {10, 1329}, {10, 1327}, {10, 1325},
-    {10, 1323}, {10, 1321}, {10, 1319}, {10, 1317}, {10, 1315}, {10, 1313},
-    {10, 1311}, {10, 1309}, {10, 1307}, {10, 1305}, {10, 1303}, {10, 1301},
-    {10, 1299}, {10, 1297}, {10, 1295}, {10, 1293}, {10, 1291}, {10, 1289},
-    {10, 1287}, {10, 1285}, {10, 1283}, {10, 1281}, {10, 1279}, {10, 1277},
-    {10, 1275}, {10, 1273}, {10, 1271}, {10, 1269}, {10, 1267}, {10, 1265},
-    {10, 1263}, {10, 1261}, {10, 1259}, {10, 1257}, {10, 1255}, {10, 1253},
-    {10, 1251}, {10, 1249}, {10, 1247}, {10, 1245}, {10, 1243}, {10, 1241},
-    {10, 1239}, {10, 1237}, {10, 1235}, {10, 1233}, {10, 1231}, {10, 1229},
-    {10, 1227}, {10, 1225}, {10, 1223}, {10, 1221}, {10, 1219}, {10, 1217},
-    {10, 1215}, {10, 1213}, {10, 1211}, {10, 1209}, {10, 1207}, {10, 1205},
-    {10, 1203}, {10, 1201}, {10, 1199}, {10, 1197}, {10, 1195}, {10, 1193},
-    {10, 1191}, {10, 1189}, {10, 1187}, {10, 1185}, {10, 1183}, {10, 1181},
-    {10, 1179}, {10, 1177}, {10, 1175}, {10, 1173}, {10, 1171}, {10, 1169},
-    {10, 1167}, {10, 1165}, {10, 1163}, {10, 1161}, {10, 1159}, {10, 1157},
-    {10, 1155}, {10, 1153}, {10, 1151}, {10, 1149}, {10, 1147}, {10, 1145},
-    {10, 1143}, {10, 1141}, {10, 1139}, {10, 1137}, {10, 1135}, {10, 1133},
-    {10, 1131}, {10, 1129}, {10, 1127}, {10, 1125}, {10, 1123}, {10, 1121},
-    {10, 1119}, {10, 1117}, {10, 1115}, {10, 1113}, {10, 1111}, {10, 1109},
-    {10, 1107}, {10, 1105}, {10, 1103}, {10, 1101}, {10, 1099}, {10, 1097},
-    {10, 1095}, {10, 1093}, {10, 1091}, {10, 1089}, {10, 1087}, {10, 1085},
-    {10, 1083}, {10, 1081}, {10, 1079}, {10, 1077}, {10, 1075}, {10, 1073},
-    {10, 1071}, {10, 1069}, {10, 1067}, {10, 1065}, {10, 1063}, {10, 1061},
-    {10, 1059}, {10, 1057}, {10, 1055}, {10, 1053}, {10, 1051}, {10, 1049},
-    {10, 1047}, {10, 1045}, {10, 1043}, {10, 1041}, {10, 1039}, {10, 1037},
-    {10, 1035}, {10, 1033}, {10, 1031}, {10, 1029}, {10, 1027}, {10, 1025},
-    {10, 1023}, {10, 1021}, {10, 1019}, {10, 1017}, {10, 1015}, {10, 1013},
-    {10, 1011}, {10, 1009}, {10, 1007}, {10, 1005}, {10, 1003}, {10, 1001},
-    {10, 999}, {10, 997}, {10, 995}, {10, 993}, {10, 991}, {10, 989},
-    {10, 987}, {10, 985}, {10, 983}, {10, 981}, {10, 979}, {10, 977},
-    {10, 975}, {10, 973}, {10, 971}, {10, 969}, {10, 967}, {10, 965},
-    {10, 963}, {10, 961}, {10, 959}, {10, 957}, {10, 955}, {10, 953},
-    {10, 951}, {10, 949}, {10, 947}, {10, 945}, {10, 943}, {10, 941},
-    {10, 939}, {10, 937}, {10, 935}, {10, 933}, {10, 931}, {10, 929},
-    {10, 927}, {10, 925}, {10, 923}, {10, 921}, {10, 919}, {10, 917},
-    {10, 915}, {10, 913}, {10, 911}, {10, 909}, {10, 907}, {10, 905},
-    {10, 903}, {10, 901}, {10, 899}, {10, 897}, {10, 895}, {10, 893},
-    {10, 891}, {10, 889}, {10, 887}, {10, 885}, {10, 883}, {10, 881},
-    {10, 879}, {10, 877}, {10, 875}, {10, 873}, {10, 871}, {10, 869},
-    {10, 867}, {10, 865}, {10, 863}, {10, 861}, {10, 859}, {10, 857},
-    {10, 855}, {10, 853}, {10, 851}, {10, 849}, {10, 847}, {10, 845},
-    {10, 843}, {10, 841}, {10, 839}, {10, 837}, {10, 835}, {10, 833},
-    {10, 831}, {10, 829}, {10, 827}, {10, 825}, {10, 823}, {10, 821},
-    {10, 819}, {10, 817}, {10, 815}, {10, 813}, {10, 811}, {10, 809},
-    {10, 807}, {10, 805}, {10, 803}, {10, 801}, {10, 799}, {10, 797},
-    {10, 795}, {10, 793}, {10, 791}, {10, 789}, {10, 787}, {10, 785},
-    {10, 783}, {10, 781}, {10, 779}, {10, 777}, {10, 775}, {10, 773},
-    {10, 771}, {10, 769}, {10, 767}, {10, 765}, {10, 763}, {10, 761},
-    {10, 759}, {10, 757}, {10, 755}, {10, 753}, {10, 751}, {10, 749},
-    {10, 747}, {10, 745}, {10, 743}, {10, 741}, {10, 739}, {10, 737},
-    {10, 735}, {10, 733}, {10, 731}, {10, 729}, {10, 727}, {10, 725},
-    {10, 723}, {10, 721}, {10, 719}, {10, 717}, {10, 715}, {10, 713},
-    {10, 711}, {10, 709}, {10, 707}, {10, 705}, {10, 703}, {10, 701},
-    {10, 699}, {10, 697}, {10, 695}, {10, 693}, {10, 691}, {10, 689},
-    {10, 687}, {10, 685}, {10, 683}, {10, 681}, {10, 679}, {10, 677},
-    {10, 675}, {10, 673}, {10, 671}, {10, 669}, {10, 667}, {10, 665},
-    {10, 663}, {10, 661}, {10, 659}, {10, 657}, {10, 655}, {10, 653},
-    {10, 651}, {10, 649}, {10, 647}, {10, 645}, {10, 643}, {10, 641},
-    {10, 639}, {10, 637}, {10, 635}, {10, 633}, {10, 631}, {10, 629},
-    {10, 627}, {10, 625}, {10, 623}, {10, 621}, {10, 619}, {10, 617},
-    {10, 615}, {10, 613}, {10, 611}, {10, 609}, {10, 607}, {10, 605},
-    {10, 603}, {10, 601}, {10, 599}, {10, 597}, {10, 595}, {10, 593},
-    {10, 591}, {10, 589}, {10, 587}, {10, 585}, {10, 583}, {10, 581},
-    {10, 579}, {10, 577}, {10, 575}, {10, 573}, {10, 571}, {10, 569},
-    {10, 567}, {10, 565}, {10, 563}, {10, 561}, {10, 559}, {10, 557},
-    {10, 555}, {10, 553}, {10, 551}, {10, 549}, {10, 547}, {10, 545},
-    {10, 543}, {10, 541}, {10, 539}, {10, 537}, {10, 535}, {10, 533},
-    {10, 531}, {10, 529}, {10, 527}, {10, 525}, {10, 523}, {10, 521},
-    {10, 519}, {10, 517}, {10, 515}, {10, 513}, {10, 511}, {10, 509},
-    {10, 507}, {10, 505}, {10, 503}, {10, 501}, {10, 499}, {10, 497},
-    {10, 495}, {10, 493}, {10, 491}, {10, 489}, {10, 487}, {10, 485},
-    {10, 483}, {10, 481}, {10, 479}, {10, 477}, {10, 475}, {10, 473},
-    {10, 471}, {10, 469}, {10, 467}, {10, 465}, {10, 463}, {10, 461},
-    {10, 459}, {10, 457}, {10, 455}, {10, 453}, {10, 451}, {10, 449},
-    {10, 447}, {10, 445}, {10, 443}, {10, 441}, {10, 439}, {10, 437},
-    {10, 435}, {10, 433}, {10, 431}, {10, 429}, {10, 427}, {10, 425},
-    {10, 423}, {10, 421}, {10, 419}, {10, 417}, {10, 415}, {10, 413},
-    {10, 411}, {10, 409}, {10, 407}, {10, 405}, {10, 403}, {10, 401},
-    {10, 399}, {10, 397}, {10, 395}, {10, 393}, {10, 391}, {10, 389},
-    {10, 387}, {10, 385}, {10, 383}, {10, 381}, {10, 379}, {10, 377},
-    {10, 375}, {10, 373}, {10, 371}, {10, 369}, {10, 367}, {10, 365},
-    {10, 363}, {10, 361}, {10, 359}, {10, 357}, {10, 355}, {10, 353},
-    {10, 351}, {10, 349}, {10, 347}, {10, 345}, {10, 343}, {10, 341},
-    {10, 339}, {10, 337}, {10, 335}, {10, 333}, {10, 331}, {10, 329},
-    {10, 327}, {10, 325}, {10, 323}, {10, 321}, {10, 319}, {10, 317},
-    {10, 315}, {10, 313}, {10, 311}, {10, 309}, {10, 307}, {10, 305},
-    {10, 303}, {10, 301}, {10, 299}, {10, 297}, {10, 295}, {10, 293},
-    {10, 291}, {10, 289}, {10, 287}, {10, 285}, {10, 283}, {10, 281},
-    {10, 279}, {10, 277}, {10, 275}, {10, 273}, {10, 271}, {10, 269},
-    {10, 267}, {10, 265}, {10, 263}, {10, 261}, {10, 259}, {10, 257},
-    {10, 255}, {10, 253}, {10, 251}, {10, 249}, {10, 247}, {10, 245},
-    {10, 243}, {10, 241}, {10, 239}, {10, 237}, {10, 235}, {10, 233},
-    {10, 231}, {10, 229}, {10, 227}, {10, 225}, {10, 223}, {10, 221},
-    {10, 219}, {10, 217}, {10, 215}, {10, 213}, {10, 211}, {10, 209},
-    {10, 207}, {10, 205}, {10, 203}, {10, 201}, {10, 199}, {10, 197},
-    {10, 195}, {10, 193}, {10, 191}, {10, 189}, {10, 187}, {10, 185},
-    {10, 183}, {10, 181}, {10, 179}, {10, 177}, {10, 175}, {10, 173},
-    {10, 171}, {10, 169}, {10, 167}, {10, 165}, {10, 163}, {10, 161},
-    {10, 159}, {10, 157}, {10, 155}, {10, 153}, {10, 151}, {10, 149},
-    {10, 147}, {10, 145}, {10, 143}, {10, 141}, {10, 139}, {10, 137},
-    {10, 135}, {10, 133}, {10, 131}, {10, 129}, {10, 127}, {10, 125},
-    {10, 123}, {10, 121}, {10, 119}, {10, 117}, {10, 115}, {10, 113},
-    {10, 111}, {10, 109}, {10, 107}, {10, 105}, {10, 103}, {10, 101},
-    {10, 99}, {10, 97}, {10, 95}, {10, 93}, {10, 91}, {10, 89},
-    {10, 87}, {10, 85}, {10, 83}, {10, 81}, {10, 79}, {10, 77},
-    {10, 75}, {10, 73}, {10, 71}, {10, 69}, {10, 67}, {10, 65},
-    {10, 63}, {10, 61}, {10, 59}, {10, 57}, {10, 55}, {10, 53},
-    {10, 51}, {10, 49}, {10, 47}, {10, 45}, {10, 43}, {10, 41},
-    {10, 39}, {10, 37}, {10, 35}, {10, 33}, {10, 31}, {10, 29},
-    {10, 27}, {10, 25}, {10, 23}, {10, 21}, {10, 19}, {10, 17},
-    {10, 15}, {10, 13}, {10, 11}, {10, 9}, {10, 7}, {10, 5},
-    {10, 3}, {10, 1}, {9, 63}, {9, 61}, {9, 59}, {9, 57},
-    {9, 55}, {9, 53}, {9, 51}, {9, 49}, {9, 47}, {9, 45},
-    {9, 43}, {9, 41}, {9, 39}, {9, 37}, {9, 35}, {9, 33},
-    {9, 31}, {9, 29}, {9, 27}, {9, 25}, {9, 23}, {9, 21},
-    {9, 19}, {9, 17}, {9, 15}, {9, 13}, {9, 11}, {9, 9},
-    {9, 7}, {9, 5}, {9, 3}, {9, 1}, {8, 31}, {8, 29},
-    {8, 27}, {8, 25}, {8, 23}, {8, 21}, {8, 19}, {8, 17},
-    {8, 15}, {8, 13}, {8, 11}, {8, 9}, {8, 7}, {8, 5},
-    {8, 3}, {8, 1}, {7, 15}, {7, 13}, {7, 11}, {7, 9},
-    {7, 7}, {7, 5}, {7, 3}, {7, 1}, {6, 7}, {6, 5},
-    {6, 3}, {6, 1}, {5, 3}, {5, 1}, {4, 1}, {3, 1},
-    {2, 1}, {1, 1}, {0, 0}, {1, 0}, {2, 0}, {3, 0},
-    {4, 0}, {5, 0}, {5, 2}, {6, 0}, {6, 2}, {6, 4},
-    {6, 6}, {7, 0}, {7, 2}, {7, 4}, {7, 6}, {7, 8},
-    {7, 10}, {7, 12}, {7, 14}, {8, 0}, {8, 2}, {8, 4},
-    {8, 6}, {8, 8}, {8, 10}, {8, 12}, {8, 14}, {8, 16},
-    {8, 18}, {8, 20}, {8, 22}, {8, 24}, {8, 26}, {8, 28},
-    {8, 30}, {9, 0}, {9, 2}, {9, 4}, {9, 6}, {9, 8},
-    {9, 10}, {9, 12}, {9, 14}, {9, 16}, {9, 18}, {9, 20},
-    {9, 22}, {9, 24}, {9, 26}, {9, 28}, {9, 30}, {9, 32},
-    {9, 34}, {9, 36}, {9, 38}, {9, 40}, {9, 42}, {9, 44},
-    {9, 46}, {9, 48}, {9, 50}, {9, 52}, {9, 54}, {9, 56},
-    {9, 58}, {9, 60}, {9, 62}, {10, 0}, {10, 2}, {10, 4},
-    {10, 6}, {10, 8}, {10, 10}, {10, 12}, {10, 14}, {10, 16},
-    {10, 18}, {10, 20}, {10, 22}, {10, 24}, {10, 26}, {10, 28},
-    {10, 30}, {10, 32}, {10, 34}, {10, 36}, {10, 38}, {10, 40},
-    {10, 42}, {10, 44}, {10, 46}, {10, 48}, {10, 50}, {10, 52},
-    {10, 54}, {10, 56}, {10, 58}, {10, 60}, {10, 62}, {10, 64},
-    {10, 66}, {10, 68}, {10, 70}, {10, 72}, {10, 74}, {10, 76},
-    {10, 78}, {10, 80}, {10, 82}, {10, 84}, {10, 86}, {10, 88},
-    {10, 90}, {10, 92}, {10, 94}, {10, 96}, {10, 98}, {10, 100},
-    {10, 102}, {10, 104}, {10, 106}, {10, 108}, {10, 110}, {10, 112},
-    {10, 114}, {10, 116}, {10, 118}, {10, 120}, {10, 122}, {10, 124},
-    {10, 126}, {10, 128}, {10, 130}, {10, 132}, {10, 134}, {10, 136},
-    {10, 138}, {10, 140}, {10, 142}, {10, 144}, {10, 146}, {10, 148},
-    {10, 150}, {10, 152}, {10, 154}, {10, 156}, {10, 158}, {10, 160},
-    {10, 162}, {10, 164}, {10, 166}, {10, 168}, {10, 170}, {10, 172},
-    {10, 174}, {10, 176}, {10, 178}, {10, 180}, {10, 182}, {10, 184},
-    {10, 186}, {10, 188}, {10, 190}, {10, 192}, {10, 194}, {10, 196},
-    {10, 198}, {10, 200}, {10, 202}, {10, 204}, {10, 206}, {10, 208},
-    {10, 210}, {10, 212}, {10, 214}, {10, 216}, {10, 218}, {10, 220},
-    {10, 222}, {10, 224}, {10, 226}, {10, 228}, {10, 230}, {10, 232},
-    {10, 234}, {10, 236}, {10, 238}, {10, 240}, {10, 242}, {10, 244},
-    {10, 246}, {10, 248}, {10, 250}, {10, 252}, {10, 254}, {10, 256},
-    {10, 258}, {10, 260}, {10, 262}, {10, 264}, {10, 266}, {10, 268},
-    {10, 270}, {10, 272}, {10, 274}, {10, 276}, {10, 278}, {10, 280},
-    {10, 282}, {10, 284}, {10, 286}, {10, 288}, {10, 290}, {10, 292},
-    {10, 294}, {10, 296}, {10, 298}, {10, 300}, {10, 302}, {10, 304},
-    {10, 306}, {10, 308}, {10, 310}, {10, 312}, {10, 314}, {10, 316},
-    {10, 318}, {10, 320}, {10, 322}, {10, 324}, {10, 326}, {10, 328},
-    {10, 330}, {10, 332}, {10, 334}, {10, 336}, {10, 338}, {10, 340},
-    {10, 342}, {10, 344}, {10, 346}, {10, 348}, {10, 350}, {10, 352},
-    {10, 354}, {10, 356}, {10, 358}, {10, 360}, {10, 362}, {10, 364},
-    {10, 366}, {10, 368}, {10, 370}, {10, 372}, {10, 374}, {10, 376},
-    {10, 378}, {10, 380}, {10, 382}, {10, 384}, {10, 386}, {10, 388},
-    {10, 390}, {10, 392}, {10, 394}, {10, 396}, {10, 398}, {10, 400},
-    {10, 402}, {10, 404}, {10, 406}, {10, 408}, {10, 410}, {10, 412},
-    {10, 414}, {10, 416}, {10, 418}, {10, 420}, {10, 422}, {10, 424},
-    {10, 426}, {10, 428}, {10, 430}, {10, 432}, {10, 434}, {10, 436},
-    {10, 438}, {10, 440}, {10, 442}, {10, 444}, {10, 446}, {10, 448},
-    {10, 450}, {10, 452}, {10, 454}, {10, 456}, {10, 458}, {10, 460},
-    {10, 462}, {10, 464}, {10, 466}, {10, 468}, {10, 470}, {10, 472},
-    {10, 474}, {10, 476}, {10, 478}, {10, 480}, {10, 482}, {10, 484},
-    {10, 486}, {10, 488}, {10, 490}, {10, 492}, {10, 494}, {10, 496},
-    {10, 498}, {10, 500}, {10, 502}, {10, 504}, {10, 506}, {10, 508},
-    {10, 510}, {10, 512}, {10, 514}, {10, 516}, {10, 518}, {10, 520},
-    {10, 522}, {10, 524}, {10, 526}, {10, 528}, {10, 530}, {10, 532},
-    {10, 534}, {10, 536}, {10, 538}, {10, 540}, {10, 542}, {10, 544},
-    {10, 546}, {10, 548}, {10, 550}, {10, 552}, {10, 554}, {10, 556},
-    {10, 558}, {10, 560}, {10, 562}, {10, 564}, {10, 566}, {10, 568},
-    {10, 570}, {10, 572}, {10, 574}, {10, 576}, {10, 578}, {10, 580},
-    {10, 582}, {10, 584}, {10, 586}, {10, 588}, {10, 590}, {10, 592},
-    {10, 594}, {10, 596}, {10, 598}, {10, 600}, {10, 602}, {10, 604},
-    {10, 606}, {10, 608}, {10, 610}, {10, 612}, {10, 614}, {10, 616},
-    {10, 618}, {10, 620}, {10, 622}, {10, 624}, {10, 626}, {10, 628},
-    {10, 630}, {10, 632}, {10, 634}, {10, 636}, {10, 638}, {10, 640},
-    {10, 642}, {10, 644}, {10, 646}, {10, 648}, {10, 650}, {10, 652},
-    {10, 654}, {10, 656}, {10, 658}, {10, 660}, {10, 662}, {10, 664},
-    {10, 666}, {10, 668}, {10, 670}, {10, 672}, {10, 674}, {10, 676},
-    {10, 678}, {10, 680}, {10, 682}, {10, 684}, {10, 686}, {10, 688},
-    {10, 690}, {10, 692}, {10, 694}, {10, 696}, {10, 698}, {10, 700},
-    {10, 702}, {10, 704}, {10, 706}, {10, 708}, {10, 710}, {10, 712},
-    {10, 714}, {10, 716}, {10, 718}, {10, 720}, {10, 722}, {10, 724},
-    {10, 726}, {10, 728}, {10, 730}, {10, 732}, {10, 734}, {10, 736},
-    {10, 738}, {10, 740}, {10, 742}, {10, 744}, {10, 746}, {10, 748},
-    {10, 750}, {10, 752}, {10, 754}, {10, 756}, {10, 758}, {10, 760},
-    {10, 762}, {10, 764}, {10, 766}, {10, 768}, {10, 770}, {10, 772},
-    {10, 774}, {10, 776}, {10, 778}, {10, 780}, {10, 782}, {10, 784},
-    {10, 786}, {10, 788}, {10, 790}, {10, 792}, {10, 794}, {10, 796},
-    {10, 798}, {10, 800}, {10, 802}, {10, 804}, {10, 806}, {10, 808},
-    {10, 810}, {10, 812}, {10, 814}, {10, 816}, {10, 818}, {10, 820},
-    {10, 822}, {10, 824}, {10, 826}, {10, 828}, {10, 830}, {10, 832},
-    {10, 834}, {10, 836}, {10, 838}, {10, 840}, {10, 842}, {10, 844},
-    {10, 846}, {10, 848}, {10, 850}, {10, 852}, {10, 854}, {10, 856},
-    {10, 858}, {10, 860}, {10, 862}, {10, 864}, {10, 866}, {10, 868},
-    {10, 870}, {10, 872}, {10, 874}, {10, 876}, {10, 878}, {10, 880},
-    {10, 882}, {10, 884}, {10, 886}, {10, 888}, {10, 890}, {10, 892},
-    {10, 894}, {10, 896}, {10, 898}, {10, 900}, {10, 902}, {10, 904},
-    {10, 906}, {10, 908}, {10, 910}, {10, 912}, {10, 914}, {10, 916},
-    {10, 918}, {10, 920}, {10, 922}, {10, 924}, {10, 926}, {10, 928},
-    {10, 930}, {10, 932}, {10, 934}, {10, 936}, {10, 938}, {10, 940},
-    {10, 942}, {10, 944}, {10, 946}, {10, 948}, {10, 950}, {10, 952},
-    {10, 954}, {10, 956}, {10, 958}, {10, 960}, {10, 962}, {10, 964},
-    {10, 966}, {10, 968}, {10, 970}, {10, 972}, {10, 974}, {10, 976},
-    {10, 978}, {10, 980}, {10, 982}, {10, 984}, {10, 986}, {10, 988},
-    {10, 990}, {10, 992}, {10, 994}, {10, 996}, {10, 998}, {10, 1000},
-    {10, 1002}, {10, 1004}, {10, 1006}, {10, 1008}, {10, 1010}, {10, 1012},
-    {10, 1014}, {10, 1016}, {10, 1018}, {10, 1020}, {10, 1022}, {10, 1024},
-    {10, 1026}, {10, 1028}, {10, 1030}, {10, 1032}, {10, 1034}, {10, 1036},
-    {10, 1038}, {10, 1040}, {10, 1042}, {10, 1044}, {10, 1046}, {10, 1048},
-    {10, 1050}, {10, 1052}, {10, 1054}, {10, 1056}, {10, 1058}, {10, 1060},
-    {10, 1062}, {10, 1064}, {10, 1066}, {10, 1068}, {10, 1070}, {10, 1072},
-    {10, 1074}, {10, 1076}, {10, 1078}, {10, 1080}, {10, 1082}, {10, 1084},
-    {10, 1086}, {10, 1088}, {10, 1090}, {10, 1092}, {10, 1094}, {10, 1096},
-    {10, 1098}, {10, 1100}, {10, 1102}, {10, 1104}, {10, 1106}, {10, 1108},
-    {10, 1110}, {10, 1112}, {10, 1114}, {10, 1116}, {10, 1118}, {10, 1120},
-    {10, 1122}, {10, 1124}, {10, 1126}, {10, 1128}, {10, 1130}, {10, 1132},
-    {10, 1134}, {10, 1136}, {10, 1138}, {10, 1140}, {10, 1142}, {10, 1144},
-    {10, 1146}, {10, 1148}, {10, 1150}, {10, 1152}, {10, 1154}, {10, 1156},
-    {10, 1158}, {10, 1160}, {10, 1162}, {10, 1164}, {10, 1166}, {10, 1168},
-    {10, 1170}, {10, 1172}, {10, 1174}, {10, 1176}, {10, 1178}, {10, 1180},
-    {10, 1182}, {10, 1184}, {10, 1186}, {10, 1188}, {10, 1190}, {10, 1192},
-    {10, 1194}, {10, 1196}, {10, 1198}, {10, 1200}, {10, 1202}, {10, 1204},
-    {10, 1206}, {10, 1208}, {10, 1210}, {10, 1212}, {10, 1214}, {10, 1216},
-    {10, 1218}, {10, 1220}, {10, 1222}, {10, 1224}, {10, 1226}, {10, 1228},
-    {10, 1230}, {10, 1232}, {10, 1234}, {10, 1236}, {10, 1238}, {10, 1240},
-    {10, 1242}, {10, 1244}, {10, 1246}, {10, 1248}, {10, 1250}, {10, 1252},
-    {10, 1254}, {10, 1256}, {10, 1258}, {10, 1260}, {10, 1262}, {10, 1264},
-    {10, 1266}, {10, 1268}, {10, 1270}, {10, 1272}, {10, 1274}, {10, 1276},
-    {10, 1278}, {10, 1280}, {10, 1282}, {10, 1284}, {10, 1286}, {10, 1288},
-    {10, 1290}, {10, 1292}, {10, 1294}, {10, 1296}, {10, 1298}, {10, 1300},
-    {10, 1302}, {10, 1304}, {10, 1306}, {10, 1308}, {10, 1310}, {10, 1312},
-    {10, 1314}, {10, 1316}, {10, 1318}, {10, 1320}, {10, 1322}, {10, 1324},
-    {10, 1326}, {10, 1328}, {10, 1330}, {10, 1332}, {10, 1334}, {10, 1336},
-    {10, 1338}, {10, 1340}, {10, 1342}, {10, 1344}, {10, 1346}, {10, 1348},
-    {10, 1350}, {10, 1352}, {10, 1354}, {10, 1356}, {10, 1358}, {10, 1360},
-    {10, 1362}, {10, 1364}, {10, 1366}, {10, 1368}, {10, 1370}, {10, 1372},
-    {10, 1374}, {10, 1376}, {10, 1378}, {10, 1380}, {10, 1382}, {10, 1384},
-    {10, 1386}, {10, 1388}, {10, 1390}, {10, 1392}, {10, 1394}, {10, 1396},
-    {10, 1398}, {10, 1400}, {10, 1402}, {10, 1404}, {10, 1406}, {10, 1408},
-    {10, 1410}, {10, 1412}, {10, 1414}, {10, 1416}, {10, 1418}, {10, 1420},
-    {10, 1422}, {10, 1424}, {10, 1426}, {10, 1428}, {10, 1430}, {10, 1432},
-    {10, 1434}, {10, 1436}, {10, 1438}, {10, 1440}, {10, 1442}, {10, 1444},
-    {10, 1446}, {10, 1448}, {10, 1450}, {10, 1452}, {10, 1454}, {10, 1456},
-    {10, 1458}, {10, 1460}, {10, 1462}, {10, 1464}, {10, 1466}, {10, 1468},
-    {10, 1470}, {10, 1472}, {10, 1474}, {10, 1476}, {10, 1478}, {10, 1480},
-    {10, 1482}, {10, 1484}, {10, 1486}, {10, 1488}, {10, 1490}, {10, 1492},
-    {10, 1494}, {10, 1496}, {10, 1498}, {10, 1500}, {10, 1502}, {10, 1504},
-    {10, 1506}, {10, 1508}, {10, 1510}, {10, 1512}, {10, 1514}, {10, 1516},
-    {10, 1518}, {10, 1520}, {10, 1522}, {10, 1524}, {10, 1526}, {10, 1528},
-    {10, 1530}, {10, 1532}, {10, 1534}, {10, 1536}, {10, 1538}, {10, 1540},
-    {10, 1542}, {10, 1544}, {10, 1546}, {10, 1548}, {10, 1550}, {10, 1552},
-    {10, 1554}, {10, 1556}, {10, 1558}, {10, 1560}, {10, 1562}, {10, 1564},
-    {10, 1566}, {10, 1568}, {10, 1570}, {10, 1572}, {10, 1574}, {10, 1576},
-    {10, 1578}, {10, 1580}, {10, 1582}, {10, 1584}, {10, 1586}, {10, 1588},
-    {10, 1590}, {10, 1592}, {10, 1594}, {10, 1596}, {10, 1598}, {10, 1600},
-    {10, 1602}, {10, 1604}, {10, 1606}, {10, 1608}, {10, 1610}, {10, 1612},
-    {10, 1614}, {10, 1616}, {10, 1618}, {10, 1620}, {10, 1622}, {10, 1624},
-    {10, 1626}, {10, 1628}, {10, 1630}, {10, 1632}, {10, 1634}, {10, 1636},
-    {10, 1638}, {10, 1640}, {10, 1642}, {10, 1644}, {10, 1646}, {10, 1648},
-    {10, 1650}, {10, 1652}, {10, 1654}, {10, 1656}, {10, 1658}, {10, 1660},
-    {10, 1662}, {10, 1664}, {10, 1666}, {10, 1668}, {10, 1670}, {10, 1672},
-    {10, 1674}, {10, 1676}, {10, 1678}, {10, 1680}, {10, 1682}, {10, 1684},
-    {10, 1686}, {10, 1688}, {10, 1690}, {10, 1692}, {10, 1694}, {10, 1696},
-    {10, 1698}, {10, 1700}, {10, 1702}, {10, 1704}, {10, 1706}, {10, 1708},
-    {10, 1710}, {10, 1712}, {10, 1714}, {10, 1716}, {10, 1718}, {10, 1720},
-    {10, 1722}, {10, 1724}, {10, 1726}, {10, 1728}, {10, 1730}, {10, 1732},
-    {10, 1734}, {10, 1736}, {10, 1738}, {10, 1740}, {10, 1742}, {10, 1744},
-    {10, 1746}, {10, 1748}, {10, 1750}, {10, 1752}, {10, 1754}, {10, 1756},
-    {10, 1758}, {10, 1760}, {10, 1762}, {10, 1764}, {10, 1766}, {10, 1768},
-    {10, 1770}, {10, 1772}, {10, 1774}, {10, 1776}, {10, 1778}, {10, 1780},
-    {10, 1782}, {10, 1784}, {10, 1786}, {10, 1788}, {10, 1790}, {10, 1792},
-    {10, 1794}, {10, 1796}, {10, 1798}, {10, 1800}, {10, 1802}, {10, 1804},
-    {10, 1806}, {10, 1808}, {10, 1810}, {10, 1812}, {10, 1814}, {10, 1816},
-    {10, 1818}, {10, 1820}, {10, 1822}, {10, 1824}, {10, 1826}, {10, 1828},
-    {10, 1830}, {10, 1832}, {10, 1834}, {10, 1836}, {10, 1838}, {10, 1840},
-    {10, 1842}, {10, 1844}, {10, 1846}, {10, 1848}, {10, 1850}, {10, 1852},
-    {10, 1854}, {10, 1856}, {10, 1858}, {10, 1860}, {10, 1862}, {10, 1864},
-    {10, 1866}, {10, 1868}, {10, 1870}, {10, 1872}, {10, 1874}, {10, 1876},
-    {10, 1878}, {10, 1880}, {10, 1882}, {10, 1884}, {10, 1886}, {10, 1888},
-    {10, 1890}, {10, 1892}, {10, 1894}, {10, 1896}, {10, 1898}, {10, 1900},
-    {10, 1902}, {10, 1904}, {10, 1906}, {10, 1908}, {10, 1910}, {10, 1912},
-    {10, 1914}, {10, 1916}, {10, 1918}, {10, 1920}, {10, 1922}, {10, 1924},
-    {10, 1926}, {10, 1928}, {10, 1930}, {10, 1932}, {10, 1934}, {10, 1936},
-    {10, 1938}, {10, 1940}, {10, 1942}, {10, 1944}, {10, 1946}, {10, 1948},
-    {10, 1950}, {10, 1952}, {10, 1954}, {10, 1956}, {10, 1958}, {10, 1960},
-    {10, 1962}, {10, 1964}, {10, 1966}, {10, 1968}, {10, 1970}, {10, 1972},
-    {10, 1974}, {10, 1976}, {10, 1978}, {10, 1980}, {10, 1982}, {10, 1984},
-    {10, 1986}, {10, 1988}, {10, 1990}, {10, 1992}, {10, 1994}, {10, 1996},
-    {10, 1998}, {10, 2000}, {10, 2002}, {10, 2004}, {10, 2006}, {10, 2008},
-    {10, 2010}, {10, 2012}, {10, 2014}, {10, 2016}, {10, 2018}, {10, 2020},
-    {10, 2022}, {10, 2024}, {10, 2026}, {10, 2028}, {10, 2030}, {10, 2032},
-    {10, 2034}, {10, 2036}, {10, 2038}, {10, 2040}, {10, 2042}, {10, 2044},
-    {10, 2046}, {10, 2048}, {10, 2050}, {10, 2052}, {10, 2054}, {10, 2056},
-    {10, 2058}, {10, 2060}, {10, 2062}, {10, 2064}, {10, 2066}, {10, 2068},
-    {10, 2070}, {10, 2072}, {10, 2074}, {10, 2076}, {10, 2078}, {10, 2080},
-    {10, 2082}, {10, 2084}, {10, 2086}, {10, 2088}, {10, 2090}, {10, 2092},
-    {10, 2094}, {10, 2096}, {10, 2098}, {10, 2100}, {10, 2102}, {10, 2104},
-    {10, 2106}, {10, 2108}, {10, 2110}, {10, 2112}, {10, 2114}, {10, 2116},
-    {10, 2118}, {10, 2120}, {10, 2122}, {10, 2124}, {10, 2126}, {10, 2128},
-    {10, 2130}, {10, 2132}, {10, 2134}, {10, 2136}, {10, 2138}, {10, 2140},
-    {10, 2142}, {10, 2144}, {10, 2146}, {10, 2148}, {10, 2150}, {10, 2152},
-    {10, 2154}, {10, 2156}, {10, 2158}, {10, 2160}, {10, 2162}, {10, 2164},
-    {10, 2166}, {10, 2168}, {10, 2170}, {10, 2172}, {10, 2174}, {10, 2176},
-    {10, 2178}, {10, 2180}, {10, 2182}, {10, 2184}, {10, 2186}, {10, 2188},
-    {10, 2190}, {10, 2192}, {10, 2194}, {10, 2196}, {10, 2198}, {10, 2200},
-    {10, 2202}, {10, 2204}, {10, 2206}, {10, 2208}, {10, 2210}, {10, 2212},
-    {10, 2214}, {10, 2216}, {10, 2218}, {10, 2220}, {10, 2222}, {10, 2224},
-    {10, 2226}, {10, 2228}, {10, 2230}, {10, 2232}, {10, 2234}, {10, 2236},
-    {10, 2238}, {10, 2240}, {10, 2242}, {10, 2244}, {10, 2246}, {10, 2248},
-    {10, 2250}, {10, 2252}, {10, 2254}, {10, 2256}, {10, 2258}, {10, 2260},
-    {10, 2262}, {10, 2264}, {10, 2266}, {10, 2268}, {10, 2270}, {10, 2272},
-    {10, 2274}, {10, 2276}, {10, 2278}, {10, 2280}, {10, 2282}, {10, 2284},
-    {10, 2286}, {10, 2288}, {10, 2290}, {10, 2292}, {10, 2294}, {10, 2296},
-    {10, 2298}, {10, 2300}, {10, 2302}, {10, 2304}, {10, 2306}, {10, 2308},
-    {10, 2310}, {10, 2312}, {10, 2314}, {10, 2316}, {10, 2318}, {10, 2320},
-    {10, 2322}, {10, 2324}, {10, 2326}, {10, 2328}, {10, 2330}, {10, 2332},
-    {10, 2334}, {10, 2336}, {10, 2338}, {10, 2340}, {10, 2342}, {10, 2344},
-    {10, 2346}, {10, 2348}, {10, 2350}, {10, 2352}, {10, 2354}, {10, 2356},
-    {10, 2358}, {10, 2360}, {10, 2362}, {10, 2364}, {10, 2366}, {10, 2368},
-    {10, 2370}, {10, 2372}, {10, 2374}, {10, 2376}, {10, 2378}, {10, 2380},
-    {10, 2382}, {10, 2384}, {10, 2386}, {10, 2388}, {10, 2390}, {10, 2392},
-    {10, 2394}, {10, 2396}, {10, 2398}, {10, 2400}, {10, 2402}, {10, 2404},
-    {10, 2406}, {10, 2408}, {10, 2410}, {10, 2412}, {10, 2414}, {10, 2416},
-    {10, 2418}, {10, 2420}, {10, 2422}, {10, 2424}, {10, 2426}, {10, 2428},
-    {10, 2430}, {10, 2432}, {10, 2434}, {10, 2436}, {10, 2438}, {10, 2440},
-    {10, 2442}, {10, 2444}, {10, 2446}, {10, 2448}, {10, 2450}, {10, 2452},
-    {10, 2454}, {10, 2456}, {10, 2458}, {10, 2460}, {10, 2462}, {10, 2464},
-    {10, 2466}, {10, 2468}, {10, 2470}, {10, 2472}, {10, 2474}, {10, 2476},
-    {10, 2478}, {10, 2480}, {10, 2482}, {10, 2484}, {10, 2486}, {10, 2488},
-    {10, 2490}, {10, 2492}, {10, 2494}, {10, 2496}, {10, 2498}, {10, 2500},
-    {10, 2502}, {10, 2504}, {10, 2506}, {10, 2508}, {10, 2510}, {10, 2512},
-    {10, 2514}, {10, 2516}, {10, 2518}, {10, 2520}, {10, 2522}, {10, 2524},
-    {10, 2526}, {10, 2528}, {10, 2530}, {10, 2532}, {10, 2534}, {10, 2536},
-    {10, 2538}, {10, 2540}, {10, 2542}, {10, 2544}, {10, 2546}, {10, 2548},
-    {10, 2550}, {10, 2552}, {10, 2554}, {10, 2556}, {10, 2558}, {10, 2560},
-    {10, 2562}, {10, 2564}, {10, 2566}, {10, 2568}, {10, 2570}, {10, 2572},
-    {10, 2574}, {10, 2576}, {10, 2578}, {10, 2580}, {10, 2582}, {10, 2584},
-    {10, 2586}, {10, 2588}, {10, 2590}, {10, 2592}, {10, 2594}, {10, 2596},
-    {10, 2598}, {10, 2600}, {10, 2602}, {10, 2604}, {10, 2606}, {10, 2608},
-    {10, 2610}, {10, 2612}, {10, 2614}, {10, 2616}, {10, 2618}, {10, 2620},
-    {10, 2622}, {10, 2624}, {10, 2626}, {10, 2628}, {10, 2630}, {10, 2632},
-    {10, 2634}, {10, 2636}, {10, 2638}, {10, 2640}, {10, 2642}, {10, 2644},
-    {10, 2646}, {10, 2648}, {10, 2650}, {10, 2652}, {10, 2654}, {10, 2656},
-    {10, 2658}, {10, 2660}, {10, 2662}, {10, 2664}, {10, 2666}, {10, 2668},
-    {10, 2670}, {10, 2672}, {10, 2674}, {10, 2676}, {10, 2678}, {10, 2680},
-    {10, 2682}, {10, 2684}, {10, 2686}, {10, 2688}, {10, 2690}, {10, 2692},
-    {10, 2694}, {10, 2696}, {10, 2698}, {10, 2700}, {10, 2702}, {10, 2704},
-    {10, 2706}, {10, 2708}, {10, 2710}, {10, 2712}, {10, 2714}, {10, 2716},
-    {10, 2718}, {10, 2720}, {10, 2722}, {10, 2724}, {10, 2726}, {10, 2728},
-    {10, 2730}, {10, 2732}, {10, 2734}, {10, 2736}, {10, 2738}, {10, 2740},
-    {10, 2742}, {10, 2744}, {10, 2746}, {10, 2748}, {10, 2750}, {10, 2752},
-    {10, 2754}, {10, 2756}, {10, 2758}, {10, 2760}, {10, 2762}, {10, 2764},
-    {10, 2766}, {10, 2768}, {10, 2770}, {10, 2772}, {10, 2774}, {10, 2776},
-    {10, 2778}, {10, 2780}, {10, 2782}, {10, 2784}, {10, 2786}, {10, 2788},
-    {10, 2790}, {10, 2792}, {10, 2794}, {10, 2796}, {10, 2798}, {10, 2800},
-    {10, 2802}, {10, 2804}, {10, 2806}, {10, 2808}, {10, 2810}, {10, 2812},
-    {10, 2814}, {10, 2816}, {10, 2818}, {10, 2820}, {10, 2822}, {10, 2824},
-    {10, 2826}, {10, 2828}, {10, 2830}, {10, 2832}, {10, 2834}, {10, 2836},
-    {10, 2838}, {10, 2840}, {10, 2842}, {10, 2844}, {10, 2846}, {10, 2848},
-    {10, 2850}, {10, 2852}, {10, 2854}, {10, 2856}, {10, 2858}, {10, 2860},
-    {10, 2862}, {10, 2864}, {10, 2866}, {10, 2868}, {10, 2870}, {10, 2872},
-    {10, 2874}, {10, 2876}, {10, 2878}, {10, 2880}, {10, 2882}, {10, 2884},
-    {10, 2886}, {10, 2888}, {10, 2890}, {10, 2892}, {10, 2894}, {10, 2896},
-    {10, 2898}, {10, 2900}, {10, 2902}, {10, 2904}, {10, 2906}, {10, 2908},
-    {10, 2910}, {10, 2912}, {10, 2914}, {10, 2916}, {10, 2918}, {10, 2920},
-    {10, 2922}, {10, 2924}, {10, 2926}, {10, 2928}, {10, 2930}, {10, 2932},
-    {10, 2934}, {10, 2936}, {10, 2938}, {10, 2940}, {10, 2942}, {10, 2944},
-    {10, 2946}, {10, 2948}, {10, 2950}, {10, 2952}, {10, 2954}, {10, 2956},
-    {10, 2958}, {10, 2960}, {10, 2962}, {10, 2964}, {10, 2966}, {10, 2968},
-    {10, 2970}, {10, 2972}, {10, 2974}, {10, 2976}, {10, 2978}, {10, 2980},
-    {10, 2982}, {10, 2984}, {10, 2986}, {10, 2988}, {10, 2990}, {10, 2992},
-    {10, 2994}, {10, 2996}, {10, 2998}, {10, 3000}, {10, 3002}, {10, 3004},
-    {10, 3006}, {10, 3008}, {10, 3010}, {10, 3012}, {10, 3014}, {10, 3016},
-    {10, 3018}, {10, 3020}, {10, 3022}, {10, 3024}, {10, 3026}, {10, 3028},
-    {10, 3030}, {10, 3032}, {10, 3034}, {10, 3036}, {10, 3038}, {10, 3040},
-    {10, 3042}, {10, 3044}, {10, 3046}, {10, 3048}, {10, 3050}, {10, 3052},
-    {10, 3054}, {10, 3056}, {10, 3058}, {10, 3060}, {10, 3062}, {10, 3064},
-    {10, 3066}, {10, 3068}, {10, 3070}, {10, 3072}, {10, 3074}, {10, 3076},
-    {10, 3078}, {10, 3080}, {10, 3082}, {10, 3084}, {10, 3086}, {10, 3088},
-    {10, 3090}, {10, 3092}, {10, 3094}, {10, 3096}, {10, 3098}, {10, 3100},
-    {10, 3102}, {10, 3104}, {10, 3106}, {10, 3108}, {10, 3110}, {10, 3112},
-    {10, 3114}, {10, 3116}, {10, 3118}, {10, 3120}, {10, 3122}, {10, 3124},
-    {10, 3126}, {10, 3128}, {10, 3130}, {10, 3132}, {10, 3134}, {10, 3136},
-    {10, 3138}, {10, 3140}, {10, 3142}, {10, 3144}, {10, 3146}, {10, 3148},
-    {10, 3150}, {10, 3152}, {10, 3154}, {10, 3156}, {10, 3158}, {10, 3160},
-    {10, 3162}, {10, 3164}, {10, 3166}, {10, 3168}, {10, 3170}, {10, 3172},
-    {10, 3174}, {10, 3176}, {10, 3178}, {10, 3180}, {10, 3182}, {10, 3184},
-    {10, 3186}, {10, 3188}, {10, 3190}, {10, 3192}, {10, 3194}, {10, 3196},
-    {10, 3198}, {10, 3200}, {10, 3202}, {10, 3204}, {10, 3206}, {10, 3208},
-    {10, 3210}, {10, 3212}, {10, 3214}, {10, 3216}, {10, 3218}, {10, 3220},
-    {10, 3222}, {10, 3224}, {10, 3226}, {10, 3228}, {10, 3230}, {10, 3232},
-    {10, 3234}, {10, 3236}, {10, 3238}, {10, 3240}, {10, 3242}, {10, 3244},
-    {10, 3246}, {10, 3248}, {10, 3250}, {10, 3252}, {10, 3254}, {10, 3256},
-    {10, 3258}, {10, 3260}, {10, 3262}, {10, 3264}, {10, 3266}, {10, 3268},
-    {10, 3270}, {10, 3272}, {10, 3274}, {10, 3276}, {10, 3278}, {10, 3280},
-    {10, 3282}, {10, 3284}, {10, 3286}, {10, 3288}, {10, 3290}, {10, 3292},
-    {10, 3294}, {10, 3296}, {10, 3298}, {10, 3300}, {10, 3302}, {10, 3304},
-    {10, 3306}, {10, 3308}, {10, 3310}, {10, 3312}, {10, 3314}, {10, 3316},
-    {10, 3318}, {10, 3320}, {10, 3322}, {10, 3324}, {10, 3326}, {10, 3328},
-    {10, 3330}, {10, 3332}, {10, 3334}, {10, 3336}, {10, 3338}, {10, 3340},
-    {10, 3342}, {10, 3344}, {10, 3346}, {10, 3348}, {10, 3350}, {10, 3352},
-    {10, 3354}, {10, 3356}, {10, 3358}, {10, 3360}, {10, 3362}, {10, 3364},
-    {10, 3366}, {10, 3368}, {10, 3370}, {10, 3372}, {10, 3374}, {10, 3376},
-    {10, 3378}, {10, 3380}, {10, 3382}, {10, 3384}, {10, 3386}, {10, 3388},
-    {10, 3390}, {10, 3392}, {10, 3394}, {10, 3396}, {10, 3398}, {10, 3400},
-    {10, 3402}, {10, 3404}, {10, 3406}, {10, 3408}, {10, 3410}, {10, 3412},
-    {10, 3414}, {10, 3416}, {10, 3418}, {10, 3420}, {10, 3422}, {10, 3424},
-    {10, 3426}, {10, 3428}, {10, 3430}, {10, 3432}, {10, 3434}, {10, 3436},
-    {10, 3438}, {10, 3440}, {10, 3442}, {10, 3444}, {10, 3446}, {10, 3448},
-    {10, 3450}, {10, 3452}, {10, 3454}, {10, 3456}, {10, 3458}, {10, 3460},
-    {10, 3462}, {10, 3464}, {10, 3466}, {10, 3468}, {10, 3470}, {10, 3472},
-    {10, 3474}, {10, 3476}, {10, 3478}, {10, 3480}, {10, 3482}, {10, 3484},
-    {10, 3486}, {10, 3488}, {10, 3490}, {10, 3492}, {10, 3494}, {10, 3496},
-    {10, 3498}, {10, 3500}, {10, 3502}, {10, 3504}, {10, 3506}, {10, 3508},
-    {10, 3510}, {10, 3512}, {10, 3514}, {10, 3516}, {10, 3518}, {10, 3520},
-    {10, 3522}, {10, 3524}, {10, 3526}, {10, 3528}, {10, 3530}, {10, 3532},
-    {10, 3534}, {10, 3536}, {10, 3538}, {10, 3540}, {10, 3542}, {10, 3544},
-    {10, 3546}, {10, 3548}, {10, 3550}, {10, 3552}, {10, 3554}, {10, 3556},
-    {10, 3558}, {10, 3560}, {10, 3562}, {10, 3564}, {10, 3566}, {10, 3568},
-    {10, 3570}, {10, 3572}, {10, 3574}, {10, 3576}, {10, 3578}, {10, 3580},
-    {10, 3582}, {10, 3584}, {10, 3586}, {10, 3588}, {10, 3590}, {10, 3592},
-    {10, 3594}, {10, 3596}, {10, 3598}, {10, 3600}, {10, 3602}, {10, 3604},
-    {10, 3606}, {10, 3608}, {10, 3610}, {10, 3612}, {10, 3614}, {10, 3616},
-    {10, 3618}, {10, 3620}, {10, 3622}, {10, 3624}, {10, 3626}, {10, 3628},
-    {10, 3630}, {10, 3632}, {10, 3634}, {10, 3636}, {10, 3638}, {10, 3640},
-    {10, 3642}, {10, 3644}, {10, 3646}, {10, 3648}, {10, 3650}, {10, 3652},
-    {10, 3654}, {10, 3656}, {10, 3658}, {10, 3660}, {10, 3662}, {10, 3664},
-    {10, 3666}, {10, 3668}, {10, 3670}, {10, 3672}, {10, 3674}, {10, 3676},
-    {10, 3678}, {10, 3680}, {10, 3682}, {10, 3684}, {10, 3686}, {10, 3688},
-    {10, 3690}, {10, 3692}, {10, 3694}, {10, 3696}, {10, 3698}, {10, 3700},
-    {10, 3702}, {10, 3704}, {10, 3706}, {10, 3708}, {10, 3710}, {10, 3712},
-    {10, 3714}, {10, 3716}, {10, 3718}, {10, 3720}, {10, 3722}, {10, 3724},
-    {10, 3726}, {10, 3728}, {10, 3730}, {10, 3732}, {10, 3734}, {10, 3736},
-    {10, 3738}, {10, 3740}, {10, 3742}, {10, 3744}, {10, 3746}, {10, 3748},
-    {10, 3750}, {10, 3752}, {10, 3754}, {10, 3756}, {10, 3758}, {10, 3760},
-    {10, 3762}, {10, 3764}, {10, 3766}, {10, 3768}, {10, 3770}, {10, 3772},
-    {10, 3774}, {10, 3776}, {10, 3778}, {10, 3780}, {10, 3782}, {10, 3784},
-    {10, 3786}, {10, 3788}, {10, 3790}, {10, 3792}, {10, 3794}, {10, 3796},
-    {10, 3798}, {10, 3800}, {10, 3802}, {10, 3804}, {10, 3806}, {10, 3808},
-    {10, 3810}, {10, 3812}, {10, 3814}, {10, 3816}, {10, 3818}, {10, 3820},
-    {10, 3822}, {10, 3824}, {10, 3826}, {10, 3828}, {10, 3830}, {10, 3832},
-    {10, 3834}, {10, 3836}, {10, 3838}, {10, 3840}, {10, 3842}, {10, 3844},
-    {10, 3846}, {10, 3848}, {10, 3850}, {10, 3852}, {10, 3854}, {10, 3856},
-    {10, 3858}, {10, 3860}, {10, 3862}, {10, 3864}, {10, 3866}, {10, 3868},
-    {10, 3870}, {10, 3872}, {10, 3874}, {10, 3876}, {10, 3878}, {10, 3880},
-    {10, 3882}, {10, 3884}, {10, 3886}, {10, 3888}, {10, 3890}, {10, 3892},
-    {10, 3894}, {10, 3896}, {10, 3898}, {10, 3900}, {10, 3902}, {10, 3904},
-    {10, 3906}, {10, 3908}, {10, 3910}, {10, 3912}, {10, 3914}, {10, 3916},
-    {10, 3918}, {10, 3920}, {10, 3922}, {10, 3924}, {10, 3926}, {10, 3928},
-    {10, 3930}, {10, 3932}, {10, 3934}, {10, 3936}, {10, 3938}, {10, 3940},
-    {10, 3942}, {10, 3944}, {10, 3946}, {10, 3948}, {10, 3950}, {10, 3952},
-    {10, 3954}, {10, 3956}, {10, 3958}, {10, 3960}
+static const TOKENVALUE dct_value_tokens[2048 * 2] = {
+  { 10, 3963 }, { 10, 3961 }, { 10, 3959 }, { 10, 3957 }, { 10, 3955 },
+  { 10, 3953 }, { 10, 3951 }, { 10, 3949 }, { 10, 3947 }, { 10, 3945 },
+  { 10, 3943 }, { 10, 3941 }, { 10, 3939 }, { 10, 3937 }, { 10, 3935 },
+  { 10, 3933 }, { 10, 3931 }, { 10, 3929 }, { 10, 3927 }, { 10, 3925 },
+  { 10, 3923 }, { 10, 3921 }, { 10, 3919 }, { 10, 3917 }, { 10, 3915 },
+  { 10, 3913 }, { 10, 3911 }, { 10, 3909 }, { 10, 3907 }, { 10, 3905 },
+  { 10, 3903 }, { 10, 3901 }, { 10, 3899 }, { 10, 3897 }, { 10, 3895 },
+  { 10, 3893 }, { 10, 3891 }, { 10, 3889 }, { 10, 3887 }, { 10, 3885 },
+  { 10, 3883 }, { 10, 3881 }, { 10, 3879 }, { 10, 3877 }, { 10, 3875 },
+  { 10, 3873 }, { 10, 3871 }, { 10, 3869 }, { 10, 3867 }, { 10, 3865 },
+  { 10, 3863 }, { 10, 3861 }, { 10, 3859 }, { 10, 3857 }, { 10, 3855 },
+  { 10, 3853 }, { 10, 3851 }, { 10, 3849 }, { 10, 3847 }, { 10, 3845 },
+  { 10, 3843 }, { 10, 3841 }, { 10, 3839 }, { 10, 3837 }, { 10, 3835 },
+  { 10, 3833 }, { 10, 3831 }, { 10, 3829 }, { 10, 3827 }, { 10, 3825 },
+  { 10, 3823 }, { 10, 3821 }, { 10, 3819 }, { 10, 3817 }, { 10, 3815 },
+  { 10, 3813 }, { 10, 3811 }, { 10, 3809 }, { 10, 3807 }, { 10, 3805 },
+  { 10, 3803 }, { 10, 3801 }, { 10, 3799 }, { 10, 3797 }, { 10, 3795 },
+  { 10, 3793 }, { 10, 3791 }, { 10, 3789 }, { 10, 3787 }, { 10, 3785 },
+  { 10, 3783 }, { 10, 3781 }, { 10, 3779 }, { 10, 3777 }, { 10, 3775 },
+  { 10, 3773 }, { 10, 3771 }, { 10, 3769 }, { 10, 3767 }, { 10, 3765 },
+  { 10, 3763 }, { 10, 3761 }, { 10, 3759 }, { 10, 3757 }, { 10, 3755 },
+  { 10, 3753 }, { 10, 3751 }, { 10, 3749 }, { 10, 3747 }, { 10, 3745 },
+  { 10, 3743 }, { 10, 3741 }, { 10, 3739 }, { 10, 3737 }, { 10, 3735 },
+  { 10, 3733 }, { 10, 3731 }, { 10, 3729 }, { 10, 3727 }, { 10, 3725 },
+  { 10, 3723 }, { 10, 3721 }, { 10, 3719 }, { 10, 3717 }, { 10, 3715 },
+  { 10, 3713 }, { 10, 3711 }, { 10, 3709 }, { 10, 3707 }, { 10, 3705 },
+  { 10, 3703 }, { 10, 3701 }, { 10, 3699 }, { 10, 3697 }, { 10, 3695 },
+  { 10, 3693 }, { 10, 3691 }, { 10, 3689 }, { 10, 3687 }, { 10, 3685 },
+  { 10, 3683 }, { 10, 3681 }, { 10, 3679 }, { 10, 3677 }, { 10, 3675 },
+  { 10, 3673 }, { 10, 3671 }, { 10, 3669 }, { 10, 3667 }, { 10, 3665 },
+  { 10, 3663 }, { 10, 3661 }, { 10, 3659 }, { 10, 3657 }, { 10, 3655 },
+  { 10, 3653 }, { 10, 3651 }, { 10, 3649 }, { 10, 3647 }, { 10, 3645 },
+  { 10, 3643 }, { 10, 3641 }, { 10, 3639 }, { 10, 3637 }, { 10, 3635 },
+  { 10, 3633 }, { 10, 3631 }, { 10, 3629 }, { 10, 3627 }, { 10, 3625 },
+  { 10, 3623 }, { 10, 3621 }, { 10, 3619 }, { 10, 3617 }, { 10, 3615 },
+  { 10, 3613 }, { 10, 3611 }, { 10, 3609 }, { 10, 3607 }, { 10, 3605 },
+  { 10, 3603 }, { 10, 3601 }, { 10, 3599 }, { 10, 3597 }, { 10, 3595 },
+  { 10, 3593 }, { 10, 3591 }, { 10, 3589 }, { 10, 3587 }, { 10, 3585 },
+  { 10, 3583 }, { 10, 3581 }, { 10, 3579 }, { 10, 3577 }, { 10, 3575 },
+  { 10, 3573 }, { 10, 3571 }, { 10, 3569 }, { 10, 3567 }, { 10, 3565 },
+  { 10, 3563 }, { 10, 3561 }, { 10, 3559 }, { 10, 3557 }, { 10, 3555 },
+  { 10, 3553 }, { 10, 3551 }, { 10, 3549 }, { 10, 3547 }, { 10, 3545 },
+  { 10, 3543 }, { 10, 3541 }, { 10, 3539 }, { 10, 3537 }, { 10, 3535 },
+  { 10, 3533 }, { 10, 3531 }, { 10, 3529 }, { 10, 3527 }, { 10, 3525 },
+  { 10, 3523 }, { 10, 3521 }, { 10, 3519 }, { 10, 3517 }, { 10, 3515 },
+  { 10, 3513 }, { 10, 3511 }, { 10, 3509 }, { 10, 3507 }, { 10, 3505 },
+  { 10, 3503 }, { 10, 3501 }, { 10, 3499 }, { 10, 3497 }, { 10, 3495 },
+  { 10, 3493 }, { 10, 3491 }, { 10, 3489 }, { 10, 3487 }, { 10, 3485 },
+  { 10, 3483 }, { 10, 3481 }, { 10, 3479 }, { 10, 3477 }, { 10, 3475 },
+  { 10, 3473 }, { 10, 3471 }, { 10, 3469 }, { 10, 3467 }, { 10, 3465 },
+  { 10, 3463 }, { 10, 3461 }, { 10, 3459 }, { 10, 3457 }, { 10, 3455 },
+  { 10, 3453 }, { 10, 3451 }, { 10, 3449 }, { 10, 3447 }, { 10, 3445 },
+  { 10, 3443 }, { 10, 3441 }, { 10, 3439 }, { 10, 3437 }, { 10, 3435 },
+  { 10, 3433 }, { 10, 3431 }, { 10, 3429 }, { 10, 3427 }, { 10, 3425 },
+  { 10, 3423 }, { 10, 3421 }, { 10, 3419 }, { 10, 3417 }, { 10, 3415 },
+  { 10, 3413 }, { 10, 3411 }, { 10, 3409 }, { 10, 3407 }, { 10, 3405 },
+  { 10, 3403 }, { 10, 3401 }, { 10, 3399 }, { 10, 3397 }, { 10, 3395 },
+  { 10, 3393 }, { 10, 3391 }, { 10, 3389 }, { 10, 3387 }, { 10, 3385 },
+  { 10, 3383 }, { 10, 3381 }, { 10, 3379 }, { 10, 3377 }, { 10, 3375 },
+  { 10, 3373 }, { 10, 3371 }, { 10, 3369 }, { 10, 3367 }, { 10, 3365 },
+  { 10, 3363 }, { 10, 3361 }, { 10, 3359 }, { 10, 3357 }, { 10, 3355 },
+  { 10, 3353 }, { 10, 3351 }, { 10, 3349 }, { 10, 3347 }, { 10, 3345 },
+  { 10, 3343 }, { 10, 3341 }, { 10, 3339 }, { 10, 3337 }, { 10, 3335 },
+  { 10, 3333 }, { 10, 3331 }, { 10, 3329 }, { 10, 3327 }, { 10, 3325 },
+  { 10, 3323 }, { 10, 3321 }, { 10, 3319 }, { 10, 3317 }, { 10, 3315 },
+  { 10, 3313 }, { 10, 3311 }, { 10, 3309 }, { 10, 3307 }, { 10, 3305 },
+  { 10, 3303 }, { 10, 3301 }, { 10, 3299 }, { 10, 3297 }, { 10, 3295 },
+  { 10, 3293 }, { 10, 3291 }, { 10, 3289 }, { 10, 3287 }, { 10, 3285 },
+  { 10, 3283 }, { 10, 3281 }, { 10, 3279 }, { 10, 3277 }, { 10, 3275 },
+  { 10, 3273 }, { 10, 3271 }, { 10, 3269 }, { 10, 3267 }, { 10, 3265 },
+  { 10, 3263 }, { 10, 3261 }, { 10, 3259 }, { 10, 3257 }, { 10, 3255 },
+  { 10, 3253 }, { 10, 3251 }, { 10, 3249 }, { 10, 3247 }, { 10, 3245 },
+  { 10, 3243 }, { 10, 3241 }, { 10, 3239 }, { 10, 3237 }, { 10, 3235 },
+  { 10, 3233 }, { 10, 3231 }, { 10, 3229 }, { 10, 3227 }, { 10, 3225 },
+  { 10, 3223 }, { 10, 3221 }, { 10, 3219 }, { 10, 3217 }, { 10, 3215 },
+  { 10, 3213 }, { 10, 3211 }, { 10, 3209 }, { 10, 3207 }, { 10, 3205 },
+  { 10, 3203 }, { 10, 3201 }, { 10, 3199 }, { 10, 3197 }, { 10, 3195 },
+  { 10, 3193 }, { 10, 3191 }, { 10, 3189 }, { 10, 3187 }, { 10, 3185 },
+  { 10, 3183 }, { 10, 3181 }, { 10, 3179 }, { 10, 3177 }, { 10, 3175 },
+  { 10, 3173 }, { 10, 3171 }, { 10, 3169 }, { 10, 3167 }, { 10, 3165 },
+  { 10, 3163 }, { 10, 3161 }, { 10, 3159 }, { 10, 3157 }, { 10, 3155 },
+  { 10, 3153 }, { 10, 3151 }, { 10, 3149 }, { 10, 3147 }, { 10, 3145 },
+  { 10, 3143 }, { 10, 3141 }, { 10, 3139 }, { 10, 3137 }, { 10, 3135 },
+  { 10, 3133 }, { 10, 3131 }, { 10, 3129 }, { 10, 3127 }, { 10, 3125 },
+  { 10, 3123 }, { 10, 3121 }, { 10, 3119 }, { 10, 3117 }, { 10, 3115 },
+  { 10, 3113 }, { 10, 3111 }, { 10, 3109 }, { 10, 3107 }, { 10, 3105 },
+  { 10, 3103 }, { 10, 3101 }, { 10, 3099 }, { 10, 3097 }, { 10, 3095 },
+  { 10, 3093 }, { 10, 3091 }, { 10, 3089 }, { 10, 3087 }, { 10, 3085 },
+  { 10, 3083 }, { 10, 3081 }, { 10, 3079 }, { 10, 3077 }, { 10, 3075 },
+  { 10, 3073 }, { 10, 3071 }, { 10, 3069 }, { 10, 3067 }, { 10, 3065 },
+  { 10, 3063 }, { 10, 3061 }, { 10, 3059 }, { 10, 3057 }, { 10, 3055 },
+  { 10, 3053 }, { 10, 3051 }, { 10, 3049 }, { 10, 3047 }, { 10, 3045 },
+  { 10, 3043 }, { 10, 3041 }, { 10, 3039 }, { 10, 3037 }, { 10, 3035 },
+  { 10, 3033 }, { 10, 3031 }, { 10, 3029 }, { 10, 3027 }, { 10, 3025 },
+  { 10, 3023 }, { 10, 3021 }, { 10, 3019 }, { 10, 3017 }, { 10, 3015 },
+  { 10, 3013 }, { 10, 3011 }, { 10, 3009 }, { 10, 3007 }, { 10, 3005 },
+  { 10, 3003 }, { 10, 3001 }, { 10, 2999 }, { 10, 2997 }, { 10, 2995 },
+  { 10, 2993 }, { 10, 2991 }, { 10, 2989 }, { 10, 2987 }, { 10, 2985 },
+  { 10, 2983 }, { 10, 2981 }, { 10, 2979 }, { 10, 2977 }, { 10, 2975 },
+  { 10, 2973 }, { 10, 2971 }, { 10, 2969 }, { 10, 2967 }, { 10, 2965 },
+  { 10, 2963 }, { 10, 2961 }, { 10, 2959 }, { 10, 2957 }, { 10, 2955 },
+  { 10, 2953 }, { 10, 2951 }, { 10, 2949 }, { 10, 2947 }, { 10, 2945 },
+  { 10, 2943 }, { 10, 2941 }, { 10, 2939 }, { 10, 2937 }, { 10, 2935 },
+  { 10, 2933 }, { 10, 2931 }, { 10, 2929 }, { 10, 2927 }, { 10, 2925 },
+  { 10, 2923 }, { 10, 2921 }, { 10, 2919 }, { 10, 2917 }, { 10, 2915 },
+  { 10, 2913 }, { 10, 2911 }, { 10, 2909 }, { 10, 2907 }, { 10, 2905 },
+  { 10, 2903 }, { 10, 2901 }, { 10, 2899 }, { 10, 2897 }, { 10, 2895 },
+  { 10, 2893 }, { 10, 2891 }, { 10, 2889 }, { 10, 2887 }, { 10, 2885 },
+  { 10, 2883 }, { 10, 2881 }, { 10, 2879 }, { 10, 2877 }, { 10, 2875 },
+  { 10, 2873 }, { 10, 2871 }, { 10, 2869 }, { 10, 2867 }, { 10, 2865 },
+  { 10, 2863 }, { 10, 2861 }, { 10, 2859 }, { 10, 2857 }, { 10, 2855 },
+  { 10, 2853 }, { 10, 2851 }, { 10, 2849 }, { 10, 2847 }, { 10, 2845 },
+  { 10, 2843 }, { 10, 2841 }, { 10, 2839 }, { 10, 2837 }, { 10, 2835 },
+  { 10, 2833 }, { 10, 2831 }, { 10, 2829 }, { 10, 2827 }, { 10, 2825 },
+  { 10, 2823 }, { 10, 2821 }, { 10, 2819 }, { 10, 2817 }, { 10, 2815 },
+  { 10, 2813 }, { 10, 2811 }, { 10, 2809 }, { 10, 2807 }, { 10, 2805 },
+  { 10, 2803 }, { 10, 2801 }, { 10, 2799 }, { 10, 2797 }, { 10, 2795 },
+  { 10, 2793 }, { 10, 2791 }, { 10, 2789 }, { 10, 2787 }, { 10, 2785 },
+  { 10, 2783 }, { 10, 2781 }, { 10, 2779 }, { 10, 2777 }, { 10, 2775 },
+  { 10, 2773 }, { 10, 2771 }, { 10, 2769 }, { 10, 2767 }, { 10, 2765 },
+  { 10, 2763 }, { 10, 2761 }, { 10, 2759 }, { 10, 2757 }, { 10, 2755 },
+  { 10, 2753 }, { 10, 2751 }, { 10, 2749 }, { 10, 2747 }, { 10, 2745 },
+  { 10, 2743 }, { 10, 2741 }, { 10, 2739 }, { 10, 2737 }, { 10, 2735 },
+  { 10, 2733 }, { 10, 2731 }, { 10, 2729 }, { 10, 2727 }, { 10, 2725 },
+  { 10, 2723 }, { 10, 2721 }, { 10, 2719 }, { 10, 2717 }, { 10, 2715 },
+  { 10, 2713 }, { 10, 2711 }, { 10, 2709 }, { 10, 2707 }, { 10, 2705 },
+  { 10, 2703 }, { 10, 2701 }, { 10, 2699 }, { 10, 2697 }, { 10, 2695 },
+  { 10, 2693 }, { 10, 2691 }, { 10, 2689 }, { 10, 2687 }, { 10, 2685 },
+  { 10, 2683 }, { 10, 2681 }, { 10, 2679 }, { 10, 2677 }, { 10, 2675 },
+  { 10, 2673 }, { 10, 2671 }, { 10, 2669 }, { 10, 2667 }, { 10, 2665 },
+  { 10, 2663 }, { 10, 2661 }, { 10, 2659 }, { 10, 2657 }, { 10, 2655 },
+  { 10, 2653 }, { 10, 2651 }, { 10, 2649 }, { 10, 2647 }, { 10, 2645 },
+  { 10, 2643 }, { 10, 2641 }, { 10, 2639 }, { 10, 2637 }, { 10, 2635 },
+  { 10, 2633 }, { 10, 2631 }, { 10, 2629 }, { 10, 2627 }, { 10, 2625 },
+  { 10, 2623 }, { 10, 2621 }, { 10, 2619 }, { 10, 2617 }, { 10, 2615 },
+  { 10, 2613 }, { 10, 2611 }, { 10, 2609 }, { 10, 2607 }, { 10, 2605 },
+  { 10, 2603 }, { 10, 2601 }, { 10, 2599 }, { 10, 2597 }, { 10, 2595 },
+  { 10, 2593 }, { 10, 2591 }, { 10, 2589 }, { 10, 2587 }, { 10, 2585 },
+  { 10, 2583 }, { 10, 2581 }, { 10, 2579 }, { 10, 2577 }, { 10, 2575 },
+  { 10, 2573 }, { 10, 2571 }, { 10, 2569 }, { 10, 2567 }, { 10, 2565 },
+  { 10, 2563 }, { 10, 2561 }, { 10, 2559 }, { 10, 2557 }, { 10, 2555 },
+  { 10, 2553 }, { 10, 2551 }, { 10, 2549 }, { 10, 2547 }, { 10, 2545 },
+  { 10, 2543 }, { 10, 2541 }, { 10, 2539 }, { 10, 2537 }, { 10, 2535 },
+  { 10, 2533 }, { 10, 2531 }, { 10, 2529 }, { 10, 2527 }, { 10, 2525 },
+  { 10, 2523 }, { 10, 2521 }, { 10, 2519 }, { 10, 2517 }, { 10, 2515 },
+  { 10, 2513 }, { 10, 2511 }, { 10, 2509 }, { 10, 2507 }, { 10, 2505 },
+  { 10, 2503 }, { 10, 2501 }, { 10, 2499 }, { 10, 2497 }, { 10, 2495 },
+  { 10, 2493 }, { 10, 2491 }, { 10, 2489 }, { 10, 2487 }, { 10, 2485 },
+  { 10, 2483 }, { 10, 2481 }, { 10, 2479 }, { 10, 2477 }, { 10, 2475 },
+  { 10, 2473 }, { 10, 2471 }, { 10, 2469 }, { 10, 2467 }, { 10, 2465 },
+  { 10, 2463 }, { 10, 2461 }, { 10, 2459 }, { 10, 2457 }, { 10, 2455 },
+  { 10, 2453 }, { 10, 2451 }, { 10, 2449 }, { 10, 2447 }, { 10, 2445 },
+  { 10, 2443 }, { 10, 2441 }, { 10, 2439 }, { 10, 2437 }, { 10, 2435 },
+  { 10, 2433 }, { 10, 2431 }, { 10, 2429 }, { 10, 2427 }, { 10, 2425 },
+  { 10, 2423 }, { 10, 2421 }, { 10, 2419 }, { 10, 2417 }, { 10, 2415 },
+  { 10, 2413 }, { 10, 2411 }, { 10, 2409 }, { 10, 2407 }, { 10, 2405 },
+  { 10, 2403 }, { 10, 2401 }, { 10, 2399 }, { 10, 2397 }, { 10, 2395 },
+  { 10, 2393 }, { 10, 2391 }, { 10, 2389 }, { 10, 2387 }, { 10, 2385 },
+  { 10, 2383 }, { 10, 2381 }, { 10, 2379 }, { 10, 2377 }, { 10, 2375 },
+  { 10, 2373 }, { 10, 2371 }, { 10, 2369 }, { 10, 2367 }, { 10, 2365 },
+  { 10, 2363 }, { 10, 2361 }, { 10, 2359 }, { 10, 2357 }, { 10, 2355 },
+  { 10, 2353 }, { 10, 2351 }, { 10, 2349 }, { 10, 2347 }, { 10, 2345 },
+  { 10, 2343 }, { 10, 2341 }, { 10, 2339 }, { 10, 2337 }, { 10, 2335 },
+  { 10, 2333 }, { 10, 2331 }, { 10, 2329 }, { 10, 2327 }, { 10, 2325 },
+  { 10, 2323 }, { 10, 2321 }, { 10, 2319 }, { 10, 2317 }, { 10, 2315 },
+  { 10, 2313 }, { 10, 2311 }, { 10, 2309 }, { 10, 2307 }, { 10, 2305 },
+  { 10, 2303 }, { 10, 2301 }, { 10, 2299 }, { 10, 2297 }, { 10, 2295 },
+  { 10, 2293 }, { 10, 2291 }, { 10, 2289 }, { 10, 2287 }, { 10, 2285 },
+  { 10, 2283 }, { 10, 2281 }, { 10, 2279 }, { 10, 2277 }, { 10, 2275 },
+  { 10, 2273 }, { 10, 2271 }, { 10, 2269 }, { 10, 2267 }, { 10, 2265 },
+  { 10, 2263 }, { 10, 2261 }, { 10, 2259 }, { 10, 2257 }, { 10, 2255 },
+  { 10, 2253 }, { 10, 2251 }, { 10, 2249 }, { 10, 2247 }, { 10, 2245 },
+  { 10, 2243 }, { 10, 2241 }, { 10, 2239 }, { 10, 2237 }, { 10, 2235 },
+  { 10, 2233 }, { 10, 2231 }, { 10, 2229 }, { 10, 2227 }, { 10, 2225 },
+  { 10, 2223 }, { 10, 2221 }, { 10, 2219 }, { 10, 2217 }, { 10, 2215 },
+  { 10, 2213 }, { 10, 2211 }, { 10, 2209 }, { 10, 2207 }, { 10, 2205 },
+  { 10, 2203 }, { 10, 2201 }, { 10, 2199 }, { 10, 2197 }, { 10, 2195 },
+  { 10, 2193 }, { 10, 2191 }, { 10, 2189 }, { 10, 2187 }, { 10, 2185 },
+  { 10, 2183 }, { 10, 2181 }, { 10, 2179 }, { 10, 2177 }, { 10, 2175 },
+  { 10, 2173 }, { 10, 2171 }, { 10, 2169 }, { 10, 2167 }, { 10, 2165 },
+  { 10, 2163 }, { 10, 2161 }, { 10, 2159 }, { 10, 2157 }, { 10, 2155 },
+  { 10, 2153 }, { 10, 2151 }, { 10, 2149 }, { 10, 2147 }, { 10, 2145 },
+  { 10, 2143 }, { 10, 2141 }, { 10, 2139 }, { 10, 2137 }, { 10, 2135 },
+  { 10, 2133 }, { 10, 2131 }, { 10, 2129 }, { 10, 2127 }, { 10, 2125 },
+  { 10, 2123 }, { 10, 2121 }, { 10, 2119 }, { 10, 2117 }, { 10, 2115 },
+  { 10, 2113 }, { 10, 2111 }, { 10, 2109 }, { 10, 2107 }, { 10, 2105 },
+  { 10, 2103 }, { 10, 2101 }, { 10, 2099 }, { 10, 2097 }, { 10, 2095 },
+  { 10, 2093 }, { 10, 2091 }, { 10, 2089 }, { 10, 2087 }, { 10, 2085 },
+  { 10, 2083 }, { 10, 2081 }, { 10, 2079 }, { 10, 2077 }, { 10, 2075 },
+  { 10, 2073 }, { 10, 2071 }, { 10, 2069 }, { 10, 2067 }, { 10, 2065 },
+  { 10, 2063 }, { 10, 2061 }, { 10, 2059 }, { 10, 2057 }, { 10, 2055 },
+  { 10, 2053 }, { 10, 2051 }, { 10, 2049 }, { 10, 2047 }, { 10, 2045 },
+  { 10, 2043 }, { 10, 2041 }, { 10, 2039 }, { 10, 2037 }, { 10, 2035 },
+  { 10, 2033 }, { 10, 2031 }, { 10, 2029 }, { 10, 2027 }, { 10, 2025 },
+  { 10, 2023 }, { 10, 2021 }, { 10, 2019 }, { 10, 2017 }, { 10, 2015 },
+  { 10, 2013 }, { 10, 2011 }, { 10, 2009 }, { 10, 2007 }, { 10, 2005 },
+  { 10, 2003 }, { 10, 2001 }, { 10, 1999 }, { 10, 1997 }, { 10, 1995 },
+  { 10, 1993 }, { 10, 1991 }, { 10, 1989 }, { 10, 1987 }, { 10, 1985 },
+  { 10, 1983 }, { 10, 1981 }, { 10, 1979 }, { 10, 1977 }, { 10, 1975 },
+  { 10, 1973 }, { 10, 1971 }, { 10, 1969 }, { 10, 1967 }, { 10, 1965 },
+  { 10, 1963 }, { 10, 1961 }, { 10, 1959 }, { 10, 1957 }, { 10, 1955 },
+  { 10, 1953 }, { 10, 1951 }, { 10, 1949 }, { 10, 1947 }, { 10, 1945 },
+  { 10, 1943 }, { 10, 1941 }, { 10, 1939 }, { 10, 1937 }, { 10, 1935 },
+  { 10, 1933 }, { 10, 1931 }, { 10, 1929 }, { 10, 1927 }, { 10, 1925 },
+  { 10, 1923 }, { 10, 1921 }, { 10, 1919 }, { 10, 1917 }, { 10, 1915 },
+  { 10, 1913 }, { 10, 1911 }, { 10, 1909 }, { 10, 1907 }, { 10, 1905 },
+  { 10, 1903 }, { 10, 1901 }, { 10, 1899 }, { 10, 1897 }, { 10, 1895 },
+  { 10, 1893 }, { 10, 1891 }, { 10, 1889 }, { 10, 1887 }, { 10, 1885 },
+  { 10, 1883 }, { 10, 1881 }, { 10, 1879 }, { 10, 1877 }, { 10, 1875 },
+  { 10, 1873 }, { 10, 1871 }, { 10, 1869 }, { 10, 1867 }, { 10, 1865 },
+  { 10, 1863 }, { 10, 1861 }, { 10, 1859 }, { 10, 1857 }, { 10, 1855 },
+  { 10, 1853 }, { 10, 1851 }, { 10, 1849 }, { 10, 1847 }, { 10, 1845 },
+  { 10, 1843 }, { 10, 1841 }, { 10, 1839 }, { 10, 1837 }, { 10, 1835 },
+  { 10, 1833 }, { 10, 1831 }, { 10, 1829 }, { 10, 1827 }, { 10, 1825 },
+  { 10, 1823 }, { 10, 1821 }, { 10, 1819 }, { 10, 1817 }, { 10, 1815 },
+  { 10, 1813 }, { 10, 1811 }, { 10, 1809 }, { 10, 1807 }, { 10, 1805 },
+  { 10, 1803 }, { 10, 1801 }, { 10, 1799 }, { 10, 1797 }, { 10, 1795 },
+  { 10, 1793 }, { 10, 1791 }, { 10, 1789 }, { 10, 1787 }, { 10, 1785 },
+  { 10, 1783 }, { 10, 1781 }, { 10, 1779 }, { 10, 1777 }, { 10, 1775 },
+  { 10, 1773 }, { 10, 1771 }, { 10, 1769 }, { 10, 1767 }, { 10, 1765 },
+  { 10, 1763 }, { 10, 1761 }, { 10, 1759 }, { 10, 1757 }, { 10, 1755 },
+  { 10, 1753 }, { 10, 1751 }, { 10, 1749 }, { 10, 1747 }, { 10, 1745 },
+  { 10, 1743 }, { 10, 1741 }, { 10, 1739 }, { 10, 1737 }, { 10, 1735 },
+  { 10, 1733 }, { 10, 1731 }, { 10, 1729 }, { 10, 1727 }, { 10, 1725 },
+  { 10, 1723 }, { 10, 1721 }, { 10, 1719 }, { 10, 1717 }, { 10, 1715 },
+  { 10, 1713 }, { 10, 1711 }, { 10, 1709 }, { 10, 1707 }, { 10, 1705 },
+  { 10, 1703 }, { 10, 1701 }, { 10, 1699 }, { 10, 1697 }, { 10, 1695 },
+  { 10, 1693 }, { 10, 1691 }, { 10, 1689 }, { 10, 1687 }, { 10, 1685 },
+  { 10, 1683 }, { 10, 1681 }, { 10, 1679 }, { 10, 1677 }, { 10, 1675 },
+  { 10, 1673 }, { 10, 1671 }, { 10, 1669 }, { 10, 1667 }, { 10, 1665 },
+  { 10, 1663 }, { 10, 1661 }, { 10, 1659 }, { 10, 1657 }, { 10, 1655 },
+  { 10, 1653 }, { 10, 1651 }, { 10, 1649 }, { 10, 1647 }, { 10, 1645 },
+  { 10, 1643 }, { 10, 1641 }, { 10, 1639 }, { 10, 1637 }, { 10, 1635 },
+  { 10, 1633 }, { 10, 1631 }, { 10, 1629 }, { 10, 1627 }, { 10, 1625 },
+  { 10, 1623 }, { 10, 1621 }, { 10, 1619 }, { 10, 1617 }, { 10, 1615 },
+  { 10, 1613 }, { 10, 1611 }, { 10, 1609 }, { 10, 1607 }, { 10, 1605 },
+  { 10, 1603 }, { 10, 1601 }, { 10, 1599 }, { 10, 1597 }, { 10, 1595 },
+  { 10, 1593 }, { 10, 1591 }, { 10, 1589 }, { 10, 1587 }, { 10, 1585 },
+  { 10, 1583 }, { 10, 1581 }, { 10, 1579 }, { 10, 1577 }, { 10, 1575 },
+  { 10, 1573 }, { 10, 1571 }, { 10, 1569 }, { 10, 1567 }, { 10, 1565 },
+  { 10, 1563 }, { 10, 1561 }, { 10, 1559 }, { 10, 1557 }, { 10, 1555 },
+  { 10, 1553 }, { 10, 1551 }, { 10, 1549 }, { 10, 1547 }, { 10, 1545 },
+  { 10, 1543 }, { 10, 1541 }, { 10, 1539 }, { 10, 1537 }, { 10, 1535 },
+  { 10, 1533 }, { 10, 1531 }, { 10, 1529 }, { 10, 1527 }, { 10, 1525 },
+  { 10, 1523 }, { 10, 1521 }, { 10, 1519 }, { 10, 1517 }, { 10, 1515 },
+  { 10, 1513 }, { 10, 1511 }, { 10, 1509 }, { 10, 1507 }, { 10, 1505 },
+  { 10, 1503 }, { 10, 1501 }, { 10, 1499 }, { 10, 1497 }, { 10, 1495 },
+  { 10, 1493 }, { 10, 1491 }, { 10, 1489 }, { 10, 1487 }, { 10, 1485 },
+  { 10, 1483 }, { 10, 1481 }, { 10, 1479 }, { 10, 1477 }, { 10, 1475 },
+  { 10, 1473 }, { 10, 1471 }, { 10, 1469 }, { 10, 1467 }, { 10, 1465 },
+  { 10, 1463 }, { 10, 1461 }, { 10, 1459 }, { 10, 1457 }, { 10, 1455 },
+  { 10, 1453 }, { 10, 1451 }, { 10, 1449 }, { 10, 1447 }, { 10, 1445 },
+  { 10, 1443 }, { 10, 1441 }, { 10, 1439 }, { 10, 1437 }, { 10, 1435 },
+  { 10, 1433 }, { 10, 1431 }, { 10, 1429 }, { 10, 1427 }, { 10, 1425 },
+  { 10, 1423 }, { 10, 1421 }, { 10, 1419 }, { 10, 1417 }, { 10, 1415 },
+  { 10, 1413 }, { 10, 1411 }, { 10, 1409 }, { 10, 1407 }, { 10, 1405 },
+  { 10, 1403 }, { 10, 1401 }, { 10, 1399 }, { 10, 1397 }, { 10, 1395 },
+  { 10, 1393 }, { 10, 1391 }, { 10, 1389 }, { 10, 1387 }, { 10, 1385 },
+  { 10, 1383 }, { 10, 1381 }, { 10, 1379 }, { 10, 1377 }, { 10, 1375 },
+  { 10, 1373 }, { 10, 1371 }, { 10, 1369 }, { 10, 1367 }, { 10, 1365 },
+  { 10, 1363 }, { 10, 1361 }, { 10, 1359 }, { 10, 1357 }, { 10, 1355 },
+  { 10, 1353 }, { 10, 1351 }, { 10, 1349 }, { 10, 1347 }, { 10, 1345 },
+  { 10, 1343 }, { 10, 1341 }, { 10, 1339 }, { 10, 1337 }, { 10, 1335 },
+  { 10, 1333 }, { 10, 1331 }, { 10, 1329 }, { 10, 1327 }, { 10, 1325 },
+  { 10, 1323 }, { 10, 1321 }, { 10, 1319 }, { 10, 1317 }, { 10, 1315 },
+  { 10, 1313 }, { 10, 1311 }, { 10, 1309 }, { 10, 1307 }, { 10, 1305 },
+  { 10, 1303 }, { 10, 1301 }, { 10, 1299 }, { 10, 1297 }, { 10, 1295 },
+  { 10, 1293 }, { 10, 1291 }, { 10, 1289 }, { 10, 1287 }, { 10, 1285 },
+  { 10, 1283 }, { 10, 1281 }, { 10, 1279 }, { 10, 1277 }, { 10, 1275 },
+  { 10, 1273 }, { 10, 1271 }, { 10, 1269 }, { 10, 1267 }, { 10, 1265 },
+  { 10, 1263 }, { 10, 1261 }, { 10, 1259 }, { 10, 1257 }, { 10, 1255 },
+  { 10, 1253 }, { 10, 1251 }, { 10, 1249 }, { 10, 1247 }, { 10, 1245 },
+  { 10, 1243 }, { 10, 1241 }, { 10, 1239 }, { 10, 1237 }, { 10, 1235 },
+  { 10, 1233 }, { 10, 1231 }, { 10, 1229 }, { 10, 1227 }, { 10, 1225 },
+  { 10, 1223 }, { 10, 1221 }, { 10, 1219 }, { 10, 1217 }, { 10, 1215 },
+  { 10, 1213 }, { 10, 1211 }, { 10, 1209 }, { 10, 1207 }, { 10, 1205 },
+  { 10, 1203 }, { 10, 1201 }, { 10, 1199 }, { 10, 1197 }, { 10, 1195 },
+  { 10, 1193 }, { 10, 1191 }, { 10, 1189 }, { 10, 1187 }, { 10, 1185 },
+  { 10, 1183 }, { 10, 1181 }, { 10, 1179 }, { 10, 1177 }, { 10, 1175 },
+  { 10, 1173 }, { 10, 1171 }, { 10, 1169 }, { 10, 1167 }, { 10, 1165 },
+  { 10, 1163 }, { 10, 1161 }, { 10, 1159 }, { 10, 1157 }, { 10, 1155 },
+  { 10, 1153 }, { 10, 1151 }, { 10, 1149 }, { 10, 1147 }, { 10, 1145 },
+  { 10, 1143 }, { 10, 1141 }, { 10, 1139 }, { 10, 1137 }, { 10, 1135 },
+  { 10, 1133 }, { 10, 1131 }, { 10, 1129 }, { 10, 1127 }, { 10, 1125 },
+  { 10, 1123 }, { 10, 1121 }, { 10, 1119 }, { 10, 1117 }, { 10, 1115 },
+  { 10, 1113 }, { 10, 1111 }, { 10, 1109 }, { 10, 1107 }, { 10, 1105 },
+  { 10, 1103 }, { 10, 1101 }, { 10, 1099 }, { 10, 1097 }, { 10, 1095 },
+  { 10, 1093 }, { 10, 1091 }, { 10, 1089 }, { 10, 1087 }, { 10, 1085 },
+  { 10, 1083 }, { 10, 1081 }, { 10, 1079 }, { 10, 1077 }, { 10, 1075 },
+  { 10, 1073 }, { 10, 1071 }, { 10, 1069 }, { 10, 1067 }, { 10, 1065 },
+  { 10, 1063 }, { 10, 1061 }, { 10, 1059 }, { 10, 1057 }, { 10, 1055 },
+  { 10, 1053 }, { 10, 1051 }, { 10, 1049 }, { 10, 1047 }, { 10, 1045 },
+  { 10, 1043 }, { 10, 1041 }, { 10, 1039 }, { 10, 1037 }, { 10, 1035 },
+  { 10, 1033 }, { 10, 1031 }, { 10, 1029 }, { 10, 1027 }, { 10, 1025 },
+  { 10, 1023 }, { 10, 1021 }, { 10, 1019 }, { 10, 1017 }, { 10, 1015 },
+  { 10, 1013 }, { 10, 1011 }, { 10, 1009 }, { 10, 1007 }, { 10, 1005 },
+  { 10, 1003 }, { 10, 1001 }, { 10, 999 },  { 10, 997 },  { 10, 995 },
+  { 10, 993 },  { 10, 991 },  { 10, 989 },  { 10, 987 },  { 10, 985 },
+  { 10, 983 },  { 10, 981 },  { 10, 979 },  { 10, 977 },  { 10, 975 },
+  { 10, 973 },  { 10, 971 },  { 10, 969 },  { 10, 967 },  { 10, 965 },
+  { 10, 963 },  { 10, 961 },  { 10, 959 },  { 10, 957 },  { 10, 955 },
+  { 10, 953 },  { 10, 951 },  { 10, 949 },  { 10, 947 },  { 10, 945 },
+  { 10, 943 },  { 10, 941 },  { 10, 939 },  { 10, 937 },  { 10, 935 },
+  { 10, 933 },  { 10, 931 },  { 10, 929 },  { 10, 927 },  { 10, 925 },
+  { 10, 923 },  { 10, 921 },  { 10, 919 },  { 10, 917 },  { 10, 915 },
+  { 10, 913 },  { 10, 911 },  { 10, 909 },  { 10, 907 },  { 10, 905 },
+  { 10, 903 },  { 10, 901 },  { 10, 899 },  { 10, 897 },  { 10, 895 },
+  { 10, 893 },  { 10, 891 },  { 10, 889 },  { 10, 887 },  { 10, 885 },
+  { 10, 883 },  { 10, 881 },  { 10, 879 },  { 10, 877 },  { 10, 875 },
+  { 10, 873 },  { 10, 871 },  { 10, 869 },  { 10, 867 },  { 10, 865 },
+  { 10, 863 },  { 10, 861 },  { 10, 859 },  { 10, 857 },  { 10, 855 },
+  { 10, 853 },  { 10, 851 },  { 10, 849 },  { 10, 847 },  { 10, 845 },
+  { 10, 843 },  { 10, 841 },  { 10, 839 },  { 10, 837 },  { 10, 835 },
+  { 10, 833 },  { 10, 831 },  { 10, 829 },  { 10, 827 },  { 10, 825 },
+  { 10, 823 },  { 10, 821 },  { 10, 819 },  { 10, 817 },  { 10, 815 },
+  { 10, 813 },  { 10, 811 },  { 10, 809 },  { 10, 807 },  { 10, 805 },
+  { 10, 803 },  { 10, 801 },  { 10, 799 },  { 10, 797 },  { 10, 795 },
+  { 10, 793 },  { 10, 791 },  { 10, 789 },  { 10, 787 },  { 10, 785 },
+  { 10, 783 },  { 10, 781 },  { 10, 779 },  { 10, 777 },  { 10, 775 },
+  { 10, 773 },  { 10, 771 },  { 10, 769 },  { 10, 767 },  { 10, 765 },
+  { 10, 763 },  { 10, 761 },  { 10, 759 },  { 10, 757 },  { 10, 755 },
+  { 10, 753 },  { 10, 751 },  { 10, 749 },  { 10, 747 },  { 10, 745 },
+  { 10, 743 },  { 10, 741 },  { 10, 739 },  { 10, 737 },  { 10, 735 },
+  { 10, 733 },  { 10, 731 },  { 10, 729 },  { 10, 727 },  { 10, 725 },
+  { 10, 723 },  { 10, 721 },  { 10, 719 },  { 10, 717 },  { 10, 715 },
+  { 10, 713 },  { 10, 711 },  { 10, 709 },  { 10, 707 },  { 10, 705 },
+  { 10, 703 },  { 10, 701 },  { 10, 699 },  { 10, 697 },  { 10, 695 },
+  { 10, 693 },  { 10, 691 },  { 10, 689 },  { 10, 687 },  { 10, 685 },
+  { 10, 683 },  { 10, 681 },  { 10, 679 },  { 10, 677 },  { 10, 675 },
+  { 10, 673 },  { 10, 671 },  { 10, 669 },  { 10, 667 },  { 10, 665 },
+  { 10, 663 },  { 10, 661 },  { 10, 659 },  { 10, 657 },  { 10, 655 },
+  { 10, 653 },  { 10, 651 },  { 10, 649 },  { 10, 647 },  { 10, 645 },
+  { 10, 643 },  { 10, 641 },  { 10, 639 },  { 10, 637 },  { 10, 635 },
+  { 10, 633 },  { 10, 631 },  { 10, 629 },  { 10, 627 },  { 10, 625 },
+  { 10, 623 },  { 10, 621 },  { 10, 619 },  { 10, 617 },  { 10, 615 },
+  { 10, 613 },  { 10, 611 },  { 10, 609 },  { 10, 607 },  { 10, 605 },
+  { 10, 603 },  { 10, 601 },  { 10, 599 },  { 10, 597 },  { 10, 595 },
+  { 10, 593 },  { 10, 591 },  { 10, 589 },  { 10, 587 },  { 10, 585 },
+  { 10, 583 },  { 10, 581 },  { 10, 579 },  { 10, 577 },  { 10, 575 },
+  { 10, 573 },  { 10, 571 },  { 10, 569 },  { 10, 567 },  { 10, 565 },
+  { 10, 563 },  { 10, 561 },  { 10, 559 },  { 10, 557 },  { 10, 555 },
+  { 10, 553 },  { 10, 551 },  { 10, 549 },  { 10, 547 },  { 10, 545 },
+  { 10, 543 },  { 10, 541 },  { 10, 539 },  { 10, 537 },  { 10, 535 },
+  { 10, 533 },  { 10, 531 },  { 10, 529 },  { 10, 527 },  { 10, 525 },
+  { 10, 523 },  { 10, 521 },  { 10, 519 },  { 10, 517 },  { 10, 515 },
+  { 10, 513 },  { 10, 511 },  { 10, 509 },  { 10, 507 },  { 10, 505 },
+  { 10, 503 },  { 10, 501 },  { 10, 499 },  { 10, 497 },  { 10, 495 },
+  { 10, 493 },  { 10, 491 },  { 10, 489 },  { 10, 487 },  { 10, 485 },
+  { 10, 483 },  { 10, 481 },  { 10, 479 },  { 10, 477 },  { 10, 475 },
+  { 10, 473 },  { 10, 471 },  { 10, 469 },  { 10, 467 },  { 10, 465 },
+  { 10, 463 },  { 10, 461 },  { 10, 459 },  { 10, 457 },  { 10, 455 },
+  { 10, 453 },  { 10, 451 },  { 10, 449 },  { 10, 447 },  { 10, 445 },
+  { 10, 443 },  { 10, 441 },  { 10, 439 },  { 10, 437 },  { 10, 435 },
+  { 10, 433 },  { 10, 431 },  { 10, 429 },  { 10, 427 },  { 10, 425 },
+  { 10, 423 },  { 10, 421 },  { 10, 419 },  { 10, 417 },  { 10, 415 },
+  { 10, 413 },  { 10, 411 },  { 10, 409 },  { 10, 407 },  { 10, 405 },
+  { 10, 403 },  { 10, 401 },  { 10, 399 },  { 10, 397 },  { 10, 395 },
+  { 10, 393 },  { 10, 391 },  { 10, 389 },  { 10, 387 },  { 10, 385 },
+  { 10, 383 },  { 10, 381 },  { 10, 379 },  { 10, 377 },  { 10, 375 },
+  { 10, 373 },  { 10, 371 },  { 10, 369 },  { 10, 367 },  { 10, 365 },
+  { 10, 363 },  { 10, 361 },  { 10, 359 },  { 10, 357 },  { 10, 355 },
+  { 10, 353 },  { 10, 351 },  { 10, 349 },  { 10, 347 },  { 10, 345 },
+  { 10, 343 },  { 10, 341 },  { 10, 339 },  { 10, 337 },  { 10, 335 },
+  { 10, 333 },  { 10, 331 },  { 10, 329 },  { 10, 327 },  { 10, 325 },
+  { 10, 323 },  { 10, 321 },  { 10, 319 },  { 10, 317 },  { 10, 315 },
+  { 10, 313 },  { 10, 311 },  { 10, 309 },  { 10, 307 },  { 10, 305 },
+  { 10, 303 },  { 10, 301 },  { 10, 299 },  { 10, 297 },  { 10, 295 },
+  { 10, 293 },  { 10, 291 },  { 10, 289 },  { 10, 287 },  { 10, 285 },
+  { 10, 283 },  { 10, 281 },  { 10, 279 },  { 10, 277 },  { 10, 275 },
+  { 10, 273 },  { 10, 271 },  { 10, 269 },  { 10, 267 },  { 10, 265 },
+  { 10, 263 },  { 10, 261 },  { 10, 259 },  { 10, 257 },  { 10, 255 },
+  { 10, 253 },  { 10, 251 },  { 10, 249 },  { 10, 247 },  { 10, 245 },
+  { 10, 243 },  { 10, 241 },  { 10, 239 },  { 10, 237 },  { 10, 235 },
+  { 10, 233 },  { 10, 231 },  { 10, 229 },  { 10, 227 },  { 10, 225 },
+  { 10, 223 },  { 10, 221 },  { 10, 219 },  { 10, 217 },  { 10, 215 },
+  { 10, 213 },  { 10, 211 },  { 10, 209 },  { 10, 207 },  { 10, 205 },
+  { 10, 203 },  { 10, 201 },  { 10, 199 },  { 10, 197 },  { 10, 195 },
+  { 10, 193 },  { 10, 191 },  { 10, 189 },  { 10, 187 },  { 10, 185 },
+  { 10, 183 },  { 10, 181 },  { 10, 179 },  { 10, 177 },  { 10, 175 },
+  { 10, 173 },  { 10, 171 },  { 10, 169 },  { 10, 167 },  { 10, 165 },
+  { 10, 163 },  { 10, 161 },  { 10, 159 },  { 10, 157 },  { 10, 155 },
+  { 10, 153 },  { 10, 151 },  { 10, 149 },  { 10, 147 },  { 10, 145 },
+  { 10, 143 },  { 10, 141 },  { 10, 139 },  { 10, 137 },  { 10, 135 },
+  { 10, 133 },  { 10, 131 },  { 10, 129 },  { 10, 127 },  { 10, 125 },
+  { 10, 123 },  { 10, 121 },  { 10, 119 },  { 10, 117 },  { 10, 115 },
+  { 10, 113 },  { 10, 111 },  { 10, 109 },  { 10, 107 },  { 10, 105 },
+  { 10, 103 },  { 10, 101 },  { 10, 99 },   { 10, 97 },   { 10, 95 },
+  { 10, 93 },   { 10, 91 },   { 10, 89 },   { 10, 87 },   { 10, 85 },
+  { 10, 83 },   { 10, 81 },   { 10, 79 },   { 10, 77 },   { 10, 75 },
+  { 10, 73 },   { 10, 71 },   { 10, 69 },   { 10, 67 },   { 10, 65 },
+  { 10, 63 },   { 10, 61 },   { 10, 59 },   { 10, 57 },   { 10, 55 },
+  { 10, 53 },   { 10, 51 },   { 10, 49 },   { 10, 47 },   { 10, 45 },
+  { 10, 43 },   { 10, 41 },   { 10, 39 },   { 10, 37 },   { 10, 35 },
+  { 10, 33 },   { 10, 31 },   { 10, 29 },   { 10, 27 },   { 10, 25 },
+  { 10, 23 },   { 10, 21 },   { 10, 19 },   { 10, 17 },   { 10, 15 },
+  { 10, 13 },   { 10, 11 },   { 10, 9 },    { 10, 7 },    { 10, 5 },
+  { 10, 3 },    { 10, 1 },    { 9, 63 },    { 9, 61 },    { 9, 59 },
+  { 9, 57 },    { 9, 55 },    { 9, 53 },    { 9, 51 },    { 9, 49 },
+  { 9, 47 },    { 9, 45 },    { 9, 43 },    { 9, 41 },    { 9, 39 },
+  { 9, 37 },    { 9, 35 },    { 9, 33 },    { 9, 31 },    { 9, 29 },
+  { 9, 27 },    { 9, 25 },    { 9, 23 },    { 9, 21 },    { 9, 19 },
+  { 9, 17 },    { 9, 15 },    { 9, 13 },    { 9, 11 },    { 9, 9 },
+  { 9, 7 },     { 9, 5 },     { 9, 3 },     { 9, 1 },     { 8, 31 },
+  { 8, 29 },    { 8, 27 },    { 8, 25 },    { 8, 23 },    { 8, 21 },
+  { 8, 19 },    { 8, 17 },    { 8, 15 },    { 8, 13 },    { 8, 11 },
+  { 8, 9 },     { 8, 7 },     { 8, 5 },     { 8, 3 },     { 8, 1 },
+  { 7, 15 },    { 7, 13 },    { 7, 11 },    { 7, 9 },     { 7, 7 },
+  { 7, 5 },     { 7, 3 },     { 7, 1 },     { 6, 7 },     { 6, 5 },
+  { 6, 3 },     { 6, 1 },     { 5, 3 },     { 5, 1 },     { 4, 1 },
+  { 3, 1 },     { 2, 1 },     { 1, 1 },     { 0, 0 },     { 1, 0 },
+  { 2, 0 },     { 3, 0 },     { 4, 0 },     { 5, 0 },     { 5, 2 },
+  { 6, 0 },     { 6, 2 },     { 6, 4 },     { 6, 6 },     { 7, 0 },
+  { 7, 2 },     { 7, 4 },     { 7, 6 },     { 7, 8 },     { 7, 10 },
+  { 7, 12 },    { 7, 14 },    { 8, 0 },     { 8, 2 },     { 8, 4 },
+  { 8, 6 },     { 8, 8 },     { 8, 10 },    { 8, 12 },    { 8, 14 },
+  { 8, 16 },    { 8, 18 },    { 8, 20 },    { 8, 22 },    { 8, 24 },
+  { 8, 26 },    { 8, 28 },    { 8, 30 },    { 9, 0 },     { 9, 2 },
+  { 9, 4 },     { 9, 6 },     { 9, 8 },     { 9, 10 },    { 9, 12 },
+  { 9, 14 },    { 9, 16 },    { 9, 18 },    { 9, 20 },    { 9, 22 },
+  { 9, 24 },    { 9, 26 },    { 9, 28 },    { 9, 30 },    { 9, 32 },
+  { 9, 34 },    { 9, 36 },    { 9, 38 },    { 9, 40 },    { 9, 42 },
+  { 9, 44 },    { 9, 46 },    { 9, 48 },    { 9, 50 },    { 9, 52 },
+  { 9, 54 },    { 9, 56 },    { 9, 58 },    { 9, 60 },    { 9, 62 },
+  { 10, 0 },    { 10, 2 },    { 10, 4 },    { 10, 6 },    { 10, 8 },
+  { 10, 10 },   { 10, 12 },   { 10, 14 },   { 10, 16 },   { 10, 18 },
+  { 10, 20 },   { 10, 22 },   { 10, 24 },   { 10, 26 },   { 10, 28 },
+  { 10, 30 },   { 10, 32 },   { 10, 34 },   { 10, 36 },   { 10, 38 },
+  { 10, 40 },   { 10, 42 },   { 10, 44 },   { 10, 46 },   { 10, 48 },
+  { 10, 50 },   { 10, 52 },   { 10, 54 },   { 10, 56 },   { 10, 58 },
+  { 10, 60 },   { 10, 62 },   { 10, 64 },   { 10, 66 },   { 10, 68 },
+  { 10, 70 },   { 10, 72 },   { 10, 74 },   { 10, 76 },   { 10, 78 },
+  { 10, 80 },   { 10, 82 },   { 10, 84 },   { 10, 86 },   { 10, 88 },
+  { 10, 90 },   { 10, 92 },   { 10, 94 },   { 10, 96 },   { 10, 98 },
+  { 10, 100 },  { 10, 102 },  { 10, 104 },  { 10, 106 },  { 10, 108 },
+  { 10, 110 },  { 10, 112 },  { 10, 114 },  { 10, 116 },  { 10, 118 },
+  { 10, 120 },  { 10, 122 },  { 10, 124 },  { 10, 126 },  { 10, 128 },
+  { 10, 130 },  { 10, 132 },  { 10, 134 },  { 10, 136 },  { 10, 138 },
+  { 10, 140 },  { 10, 142 },  { 10, 144 },  { 10, 146 },  { 10, 148 },
+  { 10, 150 },  { 10, 152 },  { 10, 154 },  { 10, 156 },  { 10, 158 },
+  { 10, 160 },  { 10, 162 },  { 10, 164 },  { 10, 166 },  { 10, 168 },
+  { 10, 170 },  { 10, 172 },  { 10, 174 },  { 10, 176 },  { 10, 178 },
+  { 10, 180 },  { 10, 182 },  { 10, 184 },  { 10, 186 },  { 10, 188 },
+  { 10, 190 },  { 10, 192 },  { 10, 194 },  { 10, 196 },  { 10, 198 },
+  { 10, 200 },  { 10, 202 },  { 10, 204 },  { 10, 206 },  { 10, 208 },
+  { 10, 210 },  { 10, 212 },  { 10, 214 },  { 10, 216 },  { 10, 218 },
+  { 10, 220 },  { 10, 222 },  { 10, 224 },  { 10, 226 },  { 10, 228 },
+  { 10, 230 },  { 10, 232 },  { 10, 234 },  { 10, 236 },  { 10, 238 },
+  { 10, 240 },  { 10, 242 },  { 10, 244 },  { 10, 246 },  { 10, 248 },
+  { 10, 250 },  { 10, 252 },  { 10, 254 },  { 10, 256 },  { 10, 258 },
+  { 10, 260 },  { 10, 262 },  { 10, 264 },  { 10, 266 },  { 10, 268 },
+  { 10, 270 },  { 10, 272 },  { 10, 274 },  { 10, 276 },  { 10, 278 },
+  { 10, 280 },  { 10, 282 },  { 10, 284 },  { 10, 286 },  { 10, 288 },
+  { 10, 290 },  { 10, 292 },  { 10, 294 },  { 10, 296 },  { 10, 298 },
+  { 10, 300 },  { 10, 302 },  { 10, 304 },  { 10, 306 },  { 10, 308 },
+  { 10, 310 },  { 10, 312 },  { 10, 314 },  { 10, 316 },  { 10, 318 },
+  { 10, 320 },  { 10, 322 },  { 10, 324 },  { 10, 326 },  { 10, 328 },
+  { 10, 330 },  { 10, 332 },  { 10, 334 },  { 10, 336 },  { 10, 338 },
+  { 10, 340 },  { 10, 342 },  { 10, 344 },  { 10, 346 },  { 10, 348 },
+  { 10, 350 },  { 10, 352 },  { 10, 354 },  { 10, 356 },  { 10, 358 },
+  { 10, 360 },  { 10, 362 },  { 10, 364 },  { 10, 366 },  { 10, 368 },
+  { 10, 370 },  { 10, 372 },  { 10, 374 },  { 10, 376 },  { 10, 378 },
+  { 10, 380 },  { 10, 382 },  { 10, 384 },  { 10, 386 },  { 10, 388 },
+  { 10, 390 },  { 10, 392 },  { 10, 394 },  { 10, 396 },  { 10, 398 },
+  { 10, 400 },  { 10, 402 },  { 10, 404 },  { 10, 406 },  { 10, 408 },
+  { 10, 410 },  { 10, 412 },  { 10, 414 },  { 10, 416 },  { 10, 418 },
+  { 10, 420 },  { 10, 422 },  { 10, 424 },  { 10, 426 },  { 10, 428 },
+  { 10, 430 },  { 10, 432 },  { 10, 434 },  { 10, 436 },  { 10, 438 },
+  { 10, 440 },  { 10, 442 },  { 10, 444 },  { 10, 446 },  { 10, 448 },
+  { 10, 450 },  { 10, 452 },  { 10, 454 },  { 10, 456 },  { 10, 458 },
+  { 10, 460 },  { 10, 462 },  { 10, 464 },  { 10, 466 },  { 10, 468 },
+  { 10, 470 },  { 10, 472 },  { 10, 474 },  { 10, 476 },  { 10, 478 },
+  { 10, 480 },  { 10, 482 },  { 10, 484 },  { 10, 486 },  { 10, 488 },
+  { 10, 490 },  { 10, 492 },  { 10, 494 },  { 10, 496 },  { 10, 498 },
+  { 10, 500 },  { 10, 502 },  { 10, 504 },  { 10, 506 },  { 10, 508 },
+  { 10, 510 },  { 10, 512 },  { 10, 514 },  { 10, 516 },  { 10, 518 },
+  { 10, 520 },  { 10, 522 },  { 10, 524 },  { 10, 526 },  { 10, 528 },
+  { 10, 530 },  { 10, 532 },  { 10, 534 },  { 10, 536 },  { 10, 538 },
+  { 10, 540 },  { 10, 542 },  { 10, 544 },  { 10, 546 },  { 10, 548 },
+  { 10, 550 },  { 10, 552 },  { 10, 554 },  { 10, 556 },  { 10, 558 },
+  { 10, 560 },  { 10, 562 },  { 10, 564 },  { 10, 566 },  { 10, 568 },
+  { 10, 570 },  { 10, 572 },  { 10, 574 },  { 10, 576 },  { 10, 578 },
+  { 10, 580 },  { 10, 582 },  { 10, 584 },  { 10, 586 },  { 10, 588 },
+  { 10, 590 },  { 10, 592 },  { 10, 594 },  { 10, 596 },  { 10, 598 },
+  { 10, 600 },  { 10, 602 },  { 10, 604 },  { 10, 606 },  { 10, 608 },
+  { 10, 610 },  { 10, 612 },  { 10, 614 },  { 10, 616 },  { 10, 618 },
+  { 10, 620 },  { 10, 622 },  { 10, 624 },  { 10, 626 },  { 10, 628 },
+  { 10, 630 },  { 10, 632 },  { 10, 634 },  { 10, 636 },  { 10, 638 },
+  { 10, 640 },  { 10, 642 },  { 10, 644 },  { 10, 646 },  { 10, 648 },
+  { 10, 650 },  { 10, 652 },  { 10, 654 },  { 10, 656 },  { 10, 658 },
+  { 10, 660 },  { 10, 662 },  { 10, 664 },  { 10, 666 },  { 10, 668 },
+  { 10, 670 },  { 10, 672 },  { 10, 674 },  { 10, 676 },  { 10, 678 },
+  { 10, 680 },  { 10, 682 },  { 10, 684 },  { 10, 686 },  { 10, 688 },
+  { 10, 690 },  { 10, 692 },  { 10, 694 },  { 10, 696 },  { 10, 698 },
+  { 10, 700 },  { 10, 702 },  { 10, 704 },  { 10, 706 },  { 10, 708 },
+  { 10, 710 },  { 10, 712 },  { 10, 714 },  { 10, 716 },  { 10, 718 },
+  { 10, 720 },  { 10, 722 },  { 10, 724 },  { 10, 726 },  { 10, 728 },
+  { 10, 730 },  { 10, 732 },  { 10, 734 },  { 10, 736 },  { 10, 738 },
+  { 10, 740 },  { 10, 742 },  { 10, 744 },  { 10, 746 },  { 10, 748 },
+  { 10, 750 },  { 10, 752 },  { 10, 754 },  { 10, 756 },  { 10, 758 },
+  { 10, 760 },  { 10, 762 },  { 10, 764 },  { 10, 766 },  { 10, 768 },
+  { 10, 770 },  { 10, 772 },  { 10, 774 },  { 10, 776 },  { 10, 778 },
+  { 10, 780 },  { 10, 782 },  { 10, 784 },  { 10, 786 },  { 10, 788 },
+  { 10, 790 },  { 10, 792 },  { 10, 794 },  { 10, 796 },  { 10, 798 },
+  { 10, 800 },  { 10, 802 },  { 10, 804 },  { 10, 806 },  { 10, 808 },
+  { 10, 810 },  { 10, 812 },  { 10, 814 },  { 10, 816 },  { 10, 818 },
+  { 10, 820 },  { 10, 822 },  { 10, 824 },  { 10, 826 },  { 10, 828 },
+  { 10, 830 },  { 10, 832 },  { 10, 834 },  { 10, 836 },  { 10, 838 },
+  { 10, 840 },  { 10, 842 },  { 10, 844 },  { 10, 846 },  { 10, 848 },
+  { 10, 850 },  { 10, 852 },  { 10, 854 },  { 10, 856 },  { 10, 858 },
+  { 10, 860 },  { 10, 862 },  { 10, 864 },  { 10, 866 },  { 10, 868 },
+  { 10, 870 },  { 10, 872 },  { 10, 874 },  { 10, 876 },  { 10, 878 },
+  { 10, 880 },  { 10, 882 },  { 10, 884 },  { 10, 886 },  { 10, 888 },
+  { 10, 890 },  { 10, 892 },  { 10, 894 },  { 10, 896 },  { 10, 898 },
+  { 10, 900 },  { 10, 902 },  { 10, 904 },  { 10, 906 },  { 10, 908 },
+  { 10, 910 },  { 10, 912 },  { 10, 914 },  { 10, 916 },  { 10, 918 },
+  { 10, 920 },  { 10, 922 },  { 10, 924 },  { 10, 926 },  { 10, 928 },
+  { 10, 930 },  { 10, 932 },  { 10, 934 },  { 10, 936 },  { 10, 938 },
+  { 10, 940 },  { 10, 942 },  { 10, 944 },  { 10, 946 },  { 10, 948 },
+  { 10, 950 },  { 10, 952 },  { 10, 954 },  { 10, 956 },  { 10, 958 },
+  { 10, 960 },  { 10, 962 },  { 10, 964 },  { 10, 966 },  { 10, 968 },
+  { 10, 970 },  { 10, 972 },  { 10, 974 },  { 10, 976 },  { 10, 978 },
+  { 10, 980 },  { 10, 982 },  { 10, 984 },  { 10, 986 },  { 10, 988 },
+  { 10, 990 },  { 10, 992 },  { 10, 994 },  { 10, 996 },  { 10, 998 },
+  { 10, 1000 }, { 10, 1002 }, { 10, 1004 }, { 10, 1006 }, { 10, 1008 },
+  { 10, 1010 }, { 10, 1012 }, { 10, 1014 }, { 10, 1016 }, { 10, 1018 },
+  { 10, 1020 }, { 10, 1022 }, { 10, 1024 }, { 10, 1026 }, { 10, 1028 },
+  { 10, 1030 }, { 10, 1032 }, { 10, 1034 }, { 10, 1036 }, { 10, 1038 },
+  { 10, 1040 }, { 10, 1042 }, { 10, 1044 }, { 10, 1046 }, { 10, 1048 },
+  { 10, 1050 }, { 10, 1052 }, { 10, 1054 }, { 10, 1056 }, { 10, 1058 },
+  { 10, 1060 }, { 10, 1062 }, { 10, 1064 }, { 10, 1066 }, { 10, 1068 },
+  { 10, 1070 }, { 10, 1072 }, { 10, 1074 }, { 10, 1076 }, { 10, 1078 },
+  { 10, 1080 }, { 10, 1082 }, { 10, 1084 }, { 10, 1086 }, { 10, 1088 },
+  { 10, 1090 }, { 10, 1092 }, { 10, 1094 }, { 10, 1096 }, { 10, 1098 },
+  { 10, 1100 }, { 10, 1102 }, { 10, 1104 }, { 10, 1106 }, { 10, 1108 },
+  { 10, 1110 }, { 10, 1112 }, { 10, 1114 }, { 10, 1116 }, { 10, 1118 },
+  { 10, 1120 }, { 10, 1122 }, { 10, 1124 }, { 10, 1126 }, { 10, 1128 },
+  { 10, 1130 }, { 10, 1132 }, { 10, 1134 }, { 10, 1136 }, { 10, 1138 },
+  { 10, 1140 }, { 10, 1142 }, { 10, 1144 }, { 10, 1146 }, { 10, 1148 },
+  { 10, 1150 }, { 10, 1152 }, { 10, 1154 }, { 10, 1156 }, { 10, 1158 },
+  { 10, 1160 }, { 10, 1162 }, { 10, 1164 }, { 10, 1166 }, { 10, 1168 },
+  { 10, 1170 }, { 10, 1172 }, { 10, 1174 }, { 10, 1176 }, { 10, 1178 },
+  { 10, 1180 }, { 10, 1182 }, { 10, 1184 }, { 10, 1186 }, { 10, 1188 },
+  { 10, 1190 }, { 10, 1192 }, { 10, 1194 }, { 10, 1196 }, { 10, 1198 },
+  { 10, 1200 }, { 10, 1202 }, { 10, 1204 }, { 10, 1206 }, { 10, 1208 },
+  { 10, 1210 }, { 10, 1212 }, { 10, 1214 }, { 10, 1216 }, { 10, 1218 },
+  { 10, 1220 }, { 10, 1222 }, { 10, 1224 }, { 10, 1226 }, { 10, 1228 },
+  { 10, 1230 }, { 10, 1232 }, { 10, 1234 }, { 10, 1236 }, { 10, 1238 },
+  { 10, 1240 }, { 10, 1242 }, { 10, 1244 }, { 10, 1246 }, { 10, 1248 },
+  { 10, 1250 }, { 10, 1252 }, { 10, 1254 }, { 10, 1256 }, { 10, 1258 },
+  { 10, 1260 }, { 10, 1262 }, { 10, 1264 }, { 10, 1266 }, { 10, 1268 },
+  { 10, 1270 }, { 10, 1272 }, { 10, 1274 }, { 10, 1276 }, { 10, 1278 },
+  { 10, 1280 }, { 10, 1282 }, { 10, 1284 }, { 10, 1286 }, { 10, 1288 },
+  { 10, 1290 }, { 10, 1292 }, { 10, 1294 }, { 10, 1296 }, { 10, 1298 },
+  { 10, 1300 }, { 10, 1302 }, { 10, 1304 }, { 10, 1306 }, { 10, 1308 },
+  { 10, 1310 }, { 10, 1312 }, { 10, 1314 }, { 10, 1316 }, { 10, 1318 },
+  { 10, 1320 }, { 10, 1322 }, { 10, 1324 }, { 10, 1326 }, { 10, 1328 },
+  { 10, 1330 }, { 10, 1332 }, { 10, 1334 }, { 10, 1336 }, { 10, 1338 },
+  { 10, 1340 }, { 10, 1342 }, { 10, 1344 }, { 10, 1346 }, { 10, 1348 },
+  { 10, 1350 }, { 10, 1352 }, { 10, 1354 }, { 10, 1356 }, { 10, 1358 },
+  { 10, 1360 }, { 10, 1362 }, { 10, 1364 }, { 10, 1366 }, { 10, 1368 },
+  { 10, 1370 }, { 10, 1372 }, { 10, 1374 }, { 10, 1376 }, { 10, 1378 },
+  { 10, 1380 }, { 10, 1382 }, { 10, 1384 }, { 10, 1386 }, { 10, 1388 },
+  { 10, 1390 }, { 10, 1392 }, { 10, 1394 }, { 10, 1396 }, { 10, 1398 },
+  { 10, 1400 }, { 10, 1402 }, { 10, 1404 }, { 10, 1406 }, { 10, 1408 },
+  { 10, 1410 }, { 10, 1412 }, { 10, 1414 }, { 10, 1416 }, { 10, 1418 },
+  { 10, 1420 }, { 10, 1422 }, { 10, 1424 }, { 10, 1426 }, { 10, 1428 },
+  { 10, 1430 }, { 10, 1432 }, { 10, 1434 }, { 10, 1436 }, { 10, 1438 },
+  { 10, 1440 }, { 10, 1442 }, { 10, 1444 }, { 10, 1446 }, { 10, 1448 },
+  { 10, 1450 }, { 10, 1452 }, { 10, 1454 }, { 10, 1456 }, { 10, 1458 },
+  { 10, 1460 }, { 10, 1462 }, { 10, 1464 }, { 10, 1466 }, { 10, 1468 },
+  { 10, 1470 }, { 10, 1472 }, { 10, 1474 }, { 10, 1476 }, { 10, 1478 },
+  { 10, 1480 }, { 10, 1482 }, { 10, 1484 }, { 10, 1486 }, { 10, 1488 },
+  { 10, 1490 }, { 10, 1492 }, { 10, 1494 }, { 10, 1496 }, { 10, 1498 },
+  { 10, 1500 }, { 10, 1502 }, { 10, 1504 }, { 10, 1506 }, { 10, 1508 },
+  { 10, 1510 }, { 10, 1512 }, { 10, 1514 }, { 10, 1516 }, { 10, 1518 },
+  { 10, 1520 }, { 10, 1522 }, { 10, 1524 }, { 10, 1526 }, { 10, 1528 },
+  { 10, 1530 }, { 10, 1532 }, { 10, 1534 }, { 10, 1536 }, { 10, 1538 },
+  { 10, 1540 }, { 10, 1542 }, { 10, 1544 }, { 10, 1546 }, { 10, 1548 },
+  { 10, 1550 }, { 10, 1552 }, { 10, 1554 }, { 10, 1556 }, { 10, 1558 },
+  { 10, 1560 }, { 10, 1562 }, { 10, 1564 }, { 10, 1566 }, { 10, 1568 },
+  { 10, 1570 }, { 10, 1572 }, { 10, 1574 }, { 10, 1576 }, { 10, 1578 },
+  { 10, 1580 }, { 10, 1582 }, { 10, 1584 }, { 10, 1586 }, { 10, 1588 },
+  { 10, 1590 }, { 10, 1592 }, { 10, 1594 }, { 10, 1596 }, { 10, 1598 },
+  { 10, 1600 }, { 10, 1602 }, { 10, 1604 }, { 10, 1606 }, { 10, 1608 },
+  { 10, 1610 }, { 10, 1612 }, { 10, 1614 }, { 10, 1616 }, { 10, 1618 },
+  { 10, 1620 }, { 10, 1622 }, { 10, 1624 }, { 10, 1626 }, { 10, 1628 },
+  { 10, 1630 }, { 10, 1632 }, { 10, 1634 }, { 10, 1636 }, { 10, 1638 },
+  { 10, 1640 }, { 10, 1642 }, { 10, 1644 }, { 10, 1646 }, { 10, 1648 },
+  { 10, 1650 }, { 10, 1652 }, { 10, 1654 }, { 10, 1656 }, { 10, 1658 },
+  { 10, 1660 }, { 10, 1662 }, { 10, 1664 }, { 10, 1666 }, { 10, 1668 },
+  { 10, 1670 }, { 10, 1672 }, { 10, 1674 }, { 10, 1676 }, { 10, 1678 },
+  { 10, 1680 }, { 10, 1682 }, { 10, 1684 }, { 10, 1686 }, { 10, 1688 },
+  { 10, 1690 }, { 10, 1692 }, { 10, 1694 }, { 10, 1696 }, { 10, 1698 },
+  { 10, 1700 }, { 10, 1702 }, { 10, 1704 }, { 10, 1706 }, { 10, 1708 },
+  { 10, 1710 }, { 10, 1712 }, { 10, 1714 }, { 10, 1716 }, { 10, 1718 },
+  { 10, 1720 }, { 10, 1722 }, { 10, 1724 }, { 10, 1726 }, { 10, 1728 },
+  { 10, 1730 }, { 10, 1732 }, { 10, 1734 }, { 10, 1736 }, { 10, 1738 },
+  { 10, 1740 }, { 10, 1742 }, { 10, 1744 }, { 10, 1746 }, { 10, 1748 },
+  { 10, 1750 }, { 10, 1752 }, { 10, 1754 }, { 10, 1756 }, { 10, 1758 },
+  { 10, 1760 }, { 10, 1762 }, { 10, 1764 }, { 10, 1766 }, { 10, 1768 },
+  { 10, 1770 }, { 10, 1772 }, { 10, 1774 }, { 10, 1776 }, { 10, 1778 },
+  { 10, 1780 }, { 10, 1782 }, { 10, 1784 }, { 10, 1786 }, { 10, 1788 },
+  { 10, 1790 }, { 10, 1792 }, { 10, 1794 }, { 10, 1796 }, { 10, 1798 },
+  { 10, 1800 }, { 10, 1802 }, { 10, 1804 }, { 10, 1806 }, { 10, 1808 },
+  { 10, 1810 }, { 10, 1812 }, { 10, 1814 }, { 10, 1816 }, { 10, 1818 },
+  { 10, 1820 }, { 10, 1822 }, { 10, 1824 }, { 10, 1826 }, { 10, 1828 },
+  { 10, 1830 }, { 10, 1832 }, { 10, 1834 }, { 10, 1836 }, { 10, 1838 },
+  { 10, 1840 }, { 10, 1842 }, { 10, 1844 }, { 10, 1846 }, { 10, 1848 },
+  { 10, 1850 }, { 10, 1852 }, { 10, 1854 }, { 10, 1856 }, { 10, 1858 },
+  { 10, 1860 }, { 10, 1862 }, { 10, 1864 }, { 10, 1866 }, { 10, 1868 },
+  { 10, 1870 }, { 10, 1872 }, { 10, 1874 }, { 10, 1876 }, { 10, 1878 },
+  { 10, 1880 }, { 10, 1882 }, { 10, 1884 }, { 10, 1886 }, { 10, 1888 },
+  { 10, 1890 }, { 10, 1892 }, { 10, 1894 }, { 10, 1896 }, { 10, 1898 },
+  { 10, 1900 }, { 10, 1902 }, { 10, 1904 }, { 10, 1906 }, { 10, 1908 },
+  { 10, 1910 }, { 10, 1912 }, { 10, 1914 }, { 10, 1916 }, { 10, 1918 },
+  { 10, 1920 }, { 10, 1922 }, { 10, 1924 }, { 10, 1926 }, { 10, 1928 },
+  { 10, 1930 }, { 10, 1932 }, { 10, 1934 }, { 10, 1936 }, { 10, 1938 },
+  { 10, 1940 }, { 10, 1942 }, { 10, 1944 }, { 10, 1946 }, { 10, 1948 },
+  { 10, 1950 }, { 10, 1952 }, { 10, 1954 }, { 10, 1956 }, { 10, 1958 },
+  { 10, 1960 }, { 10, 1962 }, { 10, 1964 }, { 10, 1966 }, { 10, 1968 },
+  { 10, 1970 }, { 10, 1972 }, { 10, 1974 }, { 10, 1976 }, { 10, 1978 },
+  { 10, 1980 }, { 10, 1982 }, { 10, 1984 }, { 10, 1986 }, { 10, 1988 },
+  { 10, 1990 }, { 10, 1992 }, { 10, 1994 }, { 10, 1996 }, { 10, 1998 },
+  { 10, 2000 }, { 10, 2002 }, { 10, 2004 }, { 10, 2006 }, { 10, 2008 },
+  { 10, 2010 }, { 10, 2012 }, { 10, 2014 }, { 10, 2016 }, { 10, 2018 },
+  { 10, 2020 }, { 10, 2022 }, { 10, 2024 }, { 10, 2026 }, { 10, 2028 },
+  { 10, 2030 }, { 10, 2032 }, { 10, 2034 }, { 10, 2036 }, { 10, 2038 },
+  { 10, 2040 }, { 10, 2042 }, { 10, 2044 }, { 10, 2046 }, { 10, 2048 },
+  { 10, 2050 }, { 10, 2052 }, { 10, 2054 }, { 10, 2056 }, { 10, 2058 },
+  { 10, 2060 }, { 10, 2062 }, { 10, 2064 }, { 10, 2066 }, { 10, 2068 },
+  { 10, 2070 }, { 10, 2072 }, { 10, 2074 }, { 10, 2076 }, { 10, 2078 },
+  { 10, 2080 }, { 10, 2082 }, { 10, 2084 }, { 10, 2086 }, { 10, 2088 },
+  { 10, 2090 }, { 10, 2092 }, { 10, 2094 }, { 10, 2096 }, { 10, 2098 },
+  { 10, 2100 }, { 10, 2102 }, { 10, 2104 }, { 10, 2106 }, { 10, 2108 },
+  { 10, 2110 }, { 10, 2112 }, { 10, 2114 }, { 10, 2116 }, { 10, 2118 },
+  { 10, 2120 }, { 10, 2122 }, { 10, 2124 }, { 10, 2126 }, { 10, 2128 },
+  { 10, 2130 }, { 10, 2132 }, { 10, 2134 }, { 10, 2136 }, { 10, 2138 },
+  { 10, 2140 }, { 10, 2142 }, { 10, 2144 }, { 10, 2146 }, { 10, 2148 },
+  { 10, 2150 }, { 10, 2152 }, { 10, 2154 }, { 10, 2156 }, { 10, 2158 },
+  { 10, 2160 }, { 10, 2162 }, { 10, 2164 }, { 10, 2166 }, { 10, 2168 },
+  { 10, 2170 }, { 10, 2172 }, { 10, 2174 }, { 10, 2176 }, { 10, 2178 },
+  { 10, 2180 }, { 10, 2182 }, { 10, 2184 }, { 10, 2186 }, { 10, 2188 },
+  { 10, 2190 }, { 10, 2192 }, { 10, 2194 }, { 10, 2196 }, { 10, 2198 },
+  { 10, 2200 }, { 10, 2202 }, { 10, 2204 }, { 10, 2206 }, { 10, 2208 },
+  { 10, 2210 }, { 10, 2212 }, { 10, 2214 }, { 10, 2216 }, { 10, 2218 },
+  { 10, 2220 }, { 10, 2222 }, { 10, 2224 }, { 10, 2226 }, { 10, 2228 },
+  { 10, 2230 }, { 10, 2232 }, { 10, 2234 }, { 10, 2236 }, { 10, 2238 },
+  { 10, 2240 }, { 10, 2242 }, { 10, 2244 }, { 10, 2246 }, { 10, 2248 },
+  { 10, 2250 }, { 10, 2252 }, { 10, 2254 }, { 10, 2256 }, { 10, 2258 },
+  { 10, 2260 }, { 10, 2262 }, { 10, 2264 }, { 10, 2266 }, { 10, 2268 },
+  { 10, 2270 }, { 10, 2272 }, { 10, 2274 }, { 10, 2276 }, { 10, 2278 },
+  { 10, 2280 }, { 10, 2282 }, { 10, 2284 }, { 10, 2286 }, { 10, 2288 },
+  { 10, 2290 }, { 10, 2292 }, { 10, 2294 }, { 10, 2296 }, { 10, 2298 },
+  { 10, 2300 }, { 10, 2302 }, { 10, 2304 }, { 10, 2306 }, { 10, 2308 },
+  { 10, 2310 }, { 10, 2312 }, { 10, 2314 }, { 10, 2316 }, { 10, 2318 },
+  { 10, 2320 }, { 10, 2322 }, { 10, 2324 }, { 10, 2326 }, { 10, 2328 },
+  { 10, 2330 }, { 10, 2332 }, { 10, 2334 }, { 10, 2336 }, { 10, 2338 },
+  { 10, 2340 }, { 10, 2342 }, { 10, 2344 }, { 10, 2346 }, { 10, 2348 },
+  { 10, 2350 }, { 10, 2352 }, { 10, 2354 }, { 10, 2356 }, { 10, 2358 },
+  { 10, 2360 }, { 10, 2362 }, { 10, 2364 }, { 10, 2366 }, { 10, 2368 },
+  { 10, 2370 }, { 10, 2372 }, { 10, 2374 }, { 10, 2376 }, { 10, 2378 },
+  { 10, 2380 }, { 10, 2382 }, { 10, 2384 }, { 10, 2386 }, { 10, 2388 },
+  { 10, 2390 }, { 10, 2392 }, { 10, 2394 }, { 10, 2396 }, { 10, 2398 },
+  { 10, 2400 }, { 10, 2402 }, { 10, 2404 }, { 10, 2406 }, { 10, 2408 },
+  { 10, 2410 }, { 10, 2412 }, { 10, 2414 }, { 10, 2416 }, { 10, 2418 },
+  { 10, 2420 }, { 10, 2422 }, { 10, 2424 }, { 10, 2426 }, { 10, 2428 },
+  { 10, 2430 }, { 10, 2432 }, { 10, 2434 }, { 10, 2436 }, { 10, 2438 },
+  { 10, 2440 }, { 10, 2442 }, { 10, 2444 }, { 10, 2446 }, { 10, 2448 },
+  { 10, 2450 }, { 10, 2452 }, { 10, 2454 }, { 10, 2456 }, { 10, 2458 },
+  { 10, 2460 }, { 10, 2462 }, { 10, 2464 }, { 10, 2466 }, { 10, 2468 },
+  { 10, 2470 }, { 10, 2472 }, { 10, 2474 }, { 10, 2476 }, { 10, 2478 },
+  { 10, 2480 }, { 10, 2482 }, { 10, 2484 }, { 10, 2486 }, { 10, 2488 },
+  { 10, 2490 }, { 10, 2492 }, { 10, 2494 }, { 10, 2496 }, { 10, 2498 },
+  { 10, 2500 }, { 10, 2502 }, { 10, 2504 }, { 10, 2506 }, { 10, 2508 },
+  { 10, 2510 }, { 10, 2512 }, { 10, 2514 }, { 10, 2516 }, { 10, 2518 },
+  { 10, 2520 }, { 10, 2522 }, { 10, 2524 }, { 10, 2526 }, { 10, 2528 },
+  { 10, 2530 }, { 10, 2532 }, { 10, 2534 }, { 10, 2536 }, { 10, 2538 },
+  { 10, 2540 }, { 10, 2542 }, { 10, 2544 }, { 10, 2546 }, { 10, 2548 },
+  { 10, 2550 }, { 10, 2552 }, { 10, 2554 }, { 10, 2556 }, { 10, 2558 },
+  { 10, 2560 }, { 10, 2562 }, { 10, 2564 }, { 10, 2566 }, { 10, 2568 },
+  { 10, 2570 }, { 10, 2572 }, { 10, 2574 }, { 10, 2576 }, { 10, 2578 },
+  { 10, 2580 }, { 10, 2582 }, { 10, 2584 }, { 10, 2586 }, { 10, 2588 },
+  { 10, 2590 }, { 10, 2592 }, { 10, 2594 }, { 10, 2596 }, { 10, 2598 },
+  { 10, 2600 }, { 10, 2602 }, { 10, 2604 }, { 10, 2606 }, { 10, 2608 },
+  { 10, 2610 }, { 10, 2612 }, { 10, 2614 }, { 10, 2616 }, { 10, 2618 },
+  { 10, 2620 }, { 10, 2622 }, { 10, 2624 }, { 10, 2626 }, { 10, 2628 },
+  { 10, 2630 }, { 10, 2632 }, { 10, 2634 }, { 10, 2636 }, { 10, 2638 },
+  { 10, 2640 }, { 10, 2642 }, { 10, 2644 }, { 10, 2646 }, { 10, 2648 },
+  { 10, 2650 }, { 10, 2652 }, { 10, 2654 }, { 10, 2656 }, { 10, 2658 },
+  { 10, 2660 }, { 10, 2662 }, { 10, 2664 }, { 10, 2666 }, { 10, 2668 },
+  { 10, 2670 }, { 10, 2672 }, { 10, 2674 }, { 10, 2676 }, { 10, 2678 },
+  { 10, 2680 }, { 10, 2682 }, { 10, 2684 }, { 10, 2686 }, { 10, 2688 },
+  { 10, 2690 }, { 10, 2692 }, { 10, 2694 }, { 10, 2696 }, { 10, 2698 },
+  { 10, 2700 }, { 10, 2702 }, { 10, 2704 }, { 10, 2706 }, { 10, 2708 },
+  { 10, 2710 }, { 10, 2712 }, { 10, 2714 }, { 10, 2716 }, { 10, 2718 },
+  { 10, 2720 }, { 10, 2722 }, { 10, 2724 }, { 10, 2726 }, { 10, 2728 },
+  { 10, 2730 }, { 10, 2732 }, { 10, 2734 }, { 10, 2736 }, { 10, 2738 },
+  { 10, 2740 }, { 10, 2742 }, { 10, 2744 }, { 10, 2746 }, { 10, 2748 },
+  { 10, 2750 }, { 10, 2752 }, { 10, 2754 }, { 10, 2756 }, { 10, 2758 },
+  { 10, 2760 }, { 10, 2762 }, { 10, 2764 }, { 10, 2766 }, { 10, 2768 },
+  { 10, 2770 }, { 10, 2772 }, { 10, 2774 }, { 10, 2776 }, { 10, 2778 },
+  { 10, 2780 }, { 10, 2782 }, { 10, 2784 }, { 10, 2786 }, { 10, 2788 },
+  { 10, 2790 }, { 10, 2792 }, { 10, 2794 }, { 10, 2796 }, { 10, 2798 },
+  { 10, 2800 }, { 10, 2802 }, { 10, 2804 }, { 10, 2806 }, { 10, 2808 },
+  { 10, 2810 }, { 10, 2812 }, { 10, 2814 }, { 10, 2816 }, { 10, 2818 },
+  { 10, 2820 }, { 10, 2822 }, { 10, 2824 }, { 10, 2826 }, { 10, 2828 },
+  { 10, 2830 }, { 10, 2832 }, { 10, 2834 }, { 10, 2836 }, { 10, 2838 },
+  { 10, 2840 }, { 10, 2842 }, { 10, 2844 }, { 10, 2846 }, { 10, 2848 },
+  { 10, 2850 }, { 10, 2852 }, { 10, 2854 }, { 10, 2856 }, { 10, 2858 },
+  { 10, 2860 }, { 10, 2862 }, { 10, 2864 }, { 10, 2866 }, { 10, 2868 },
+  { 10, 2870 }, { 10, 2872 }, { 10, 2874 }, { 10, 2876 }, { 10, 2878 },
+  { 10, 2880 }, { 10, 2882 }, { 10, 2884 }, { 10, 2886 }, { 10, 2888 },
+  { 10, 2890 }, { 10, 2892 }, { 10, 2894 }, { 10, 2896 }, { 10, 2898 },
+  { 10, 2900 }, { 10, 2902 }, { 10, 2904 }, { 10, 2906 }, { 10, 2908 },
+  { 10, 2910 }, { 10, 2912 }, { 10, 2914 }, { 10, 2916 }, { 10, 2918 },
+  { 10, 2920 }, { 10, 2922 }, { 10, 2924 }, { 10, 2926 }, { 10, 2928 },
+  { 10, 2930 }, { 10, 2932 }, { 10, 2934 }, { 10, 2936 }, { 10, 2938 },
+  { 10, 2940 }, { 10, 2942 }, { 10, 2944 }, { 10, 2946 }, { 10, 2948 },
+  { 10, 2950 }, { 10, 2952 }, { 10, 2954 }, { 10, 2956 }, { 10, 2958 },
+  { 10, 2960 }, { 10, 2962 }, { 10, 2964 }, { 10, 2966 }, { 10, 2968 },
+  { 10, 2970 }, { 10, 2972 }, { 10, 2974 }, { 10, 2976 }, { 10, 2978 },
+  { 10, 2980 }, { 10, 2982 }, { 10, 2984 }, { 10, 2986 }, { 10, 2988 },
+  { 10, 2990 }, { 10, 2992 }, { 10, 2994 }, { 10, 2996 }, { 10, 2998 },
+  { 10, 3000 }, { 10, 3002 }, { 10, 3004 }, { 10, 3006 }, { 10, 3008 },
+  { 10, 3010 }, { 10, 3012 }, { 10, 3014 }, { 10, 3016 }, { 10, 3018 },
+  { 10, 3020 }, { 10, 3022 }, { 10, 3024 }, { 10, 3026 }, { 10, 3028 },
+  { 10, 3030 }, { 10, 3032 }, { 10, 3034 }, { 10, 3036 }, { 10, 3038 },
+  { 10, 3040 }, { 10, 3042 }, { 10, 3044 }, { 10, 3046 }, { 10, 3048 },
+  { 10, 3050 }, { 10, 3052 }, { 10, 3054 }, { 10, 3056 }, { 10, 3058 },
+  { 10, 3060 }, { 10, 3062 }, { 10, 3064 }, { 10, 3066 }, { 10, 3068 },
+  { 10, 3070 }, { 10, 3072 }, { 10, 3074 }, { 10, 3076 }, { 10, 3078 },
+  { 10, 3080 }, { 10, 3082 }, { 10, 3084 }, { 10, 3086 }, { 10, 3088 },
+  { 10, 3090 }, { 10, 3092 }, { 10, 3094 }, { 10, 3096 }, { 10, 3098 },
+  { 10, 3100 }, { 10, 3102 }, { 10, 3104 }, { 10, 3106 }, { 10, 3108 },
+  { 10, 3110 }, { 10, 3112 }, { 10, 3114 }, { 10, 3116 }, { 10, 3118 },
+  { 10, 3120 }, { 10, 3122 }, { 10, 3124 }, { 10, 3126 }, { 10, 3128 },
+  { 10, 3130 }, { 10, 3132 }, { 10, 3134 }, { 10, 3136 }, { 10, 3138 },
+  { 10, 3140 }, { 10, 3142 }, { 10, 3144 }, { 10, 3146 }, { 10, 3148 },
+  { 10, 3150 }, { 10, 3152 }, { 10, 3154 }, { 10, 3156 }, { 10, 3158 },
+  { 10, 3160 }, { 10, 3162 }, { 10, 3164 }, { 10, 3166 }, { 10, 3168 },
+  { 10, 3170 }, { 10, 3172 }, { 10, 3174 }, { 10, 3176 }, { 10, 3178 },
+  { 10, 3180 }, { 10, 3182 }, { 10, 3184 }, { 10, 3186 }, { 10, 3188 },
+  { 10, 3190 }, { 10, 3192 }, { 10, 3194 }, { 10, 3196 }, { 10, 3198 },
+  { 10, 3200 }, { 10, 3202 }, { 10, 3204 }, { 10, 3206 }, { 10, 3208 },
+  { 10, 3210 }, { 10, 3212 }, { 10, 3214 }, { 10, 3216 }, { 10, 3218 },
+  { 10, 3220 }, { 10, 3222 }, { 10, 3224 }, { 10, 3226 }, { 10, 3228 },
+  { 10, 3230 }, { 10, 3232 }, { 10, 3234 }, { 10, 3236 }, { 10, 3238 },
+  { 10, 3240 }, { 10, 3242 }, { 10, 3244 }, { 10, 3246 }, { 10, 3248 },
+  { 10, 3250 }, { 10, 3252 }, { 10, 3254 }, { 10, 3256 }, { 10, 3258 },
+  { 10, 3260 }, { 10, 3262 }, { 10, 3264 }, { 10, 3266 }, { 10, 3268 },
+  { 10, 3270 }, { 10, 3272 }, { 10, 3274 }, { 10, 3276 }, { 10, 3278 },
+  { 10, 3280 }, { 10, 3282 }, { 10, 3284 }, { 10, 3286 }, { 10, 3288 },
+  { 10, 3290 }, { 10, 3292 }, { 10, 3294 }, { 10, 3296 }, { 10, 3298 },
+  { 10, 3300 }, { 10, 3302 }, { 10, 3304 }, { 10, 3306 }, { 10, 3308 },
+  { 10, 3310 }, { 10, 3312 }, { 10, 3314 }, { 10, 3316 }, { 10, 3318 },
+  { 10, 3320 }, { 10, 3322 }, { 10, 3324 }, { 10, 3326 }, { 10, 3328 },
+  { 10, 3330 }, { 10, 3332 }, { 10, 3334 }, { 10, 3336 }, { 10, 3338 },
+  { 10, 3340 }, { 10, 3342 }, { 10, 3344 }, { 10, 3346 }, { 10, 3348 },
+  { 10, 3350 }, { 10, 3352 }, { 10, 3354 }, { 10, 3356 }, { 10, 3358 },
+  { 10, 3360 }, { 10, 3362 }, { 10, 3364 }, { 10, 3366 }, { 10, 3368 },
+  { 10, 3370 }, { 10, 3372 }, { 10, 3374 }, { 10, 3376 }, { 10, 3378 },
+  { 10, 3380 }, { 10, 3382 }, { 10, 3384 }, { 10, 3386 }, { 10, 3388 },
+  { 10, 3390 }, { 10, 3392 }, { 10, 3394 }, { 10, 3396 }, { 10, 3398 },
+  { 10, 3400 }, { 10, 3402 }, { 10, 3404 }, { 10, 3406 }, { 10, 3408 },
+  { 10, 3410 }, { 10, 3412 }, { 10, 3414 }, { 10, 3416 }, { 10, 3418 },
+  { 10, 3420 }, { 10, 3422 }, { 10, 3424 }, { 10, 3426 }, { 10, 3428 },
+  { 10, 3430 }, { 10, 3432 }, { 10, 3434 }, { 10, 3436 }, { 10, 3438 },
+  { 10, 3440 }, { 10, 3442 }, { 10, 3444 }, { 10, 3446 }, { 10, 3448 },
+  { 10, 3450 }, { 10, 3452 }, { 10, 3454 }, { 10, 3456 }, { 10, 3458 },
+  { 10, 3460 }, { 10, 3462 }, { 10, 3464 }, { 10, 3466 }, { 10, 3468 },
+  { 10, 3470 }, { 10, 3472 }, { 10, 3474 }, { 10, 3476 }, { 10, 3478 },
+  { 10, 3480 }, { 10, 3482 }, { 10, 3484 }, { 10, 3486 }, { 10, 3488 },
+  { 10, 3490 }, { 10, 3492 }, { 10, 3494 }, { 10, 3496 }, { 10, 3498 },
+  { 10, 3500 }, { 10, 3502 }, { 10, 3504 }, { 10, 3506 }, { 10, 3508 },
+  { 10, 3510 }, { 10, 3512 }, { 10, 3514 }, { 10, 3516 }, { 10, 3518 },
+  { 10, 3520 }, { 10, 3522 }, { 10, 3524 }, { 10, 3526 }, { 10, 3528 },
+  { 10, 3530 }, { 10, 3532 }, { 10, 3534 }, { 10, 3536 }, { 10, 3538 },
+  { 10, 3540 }, { 10, 3542 }, { 10, 3544 }, { 10, 3546 }, { 10, 3548 },
+  { 10, 3550 }, { 10, 3552 }, { 10, 3554 }, { 10, 3556 }, { 10, 3558 },
+  { 10, 3560 }, { 10, 3562 }, { 10, 3564 }, { 10, 3566 }, { 10, 3568 },
+  { 10, 3570 }, { 10, 3572 }, { 10, 3574 }, { 10, 3576 }, { 10, 3578 },
+  { 10, 3580 }, { 10, 3582 }, { 10, 3584 }, { 10, 3586 }, { 10, 3588 },
+  { 10, 3590 }, { 10, 3592 }, { 10, 3594 }, { 10, 3596 }, { 10, 3598 },
+  { 10, 3600 }, { 10, 3602 }, { 10, 3604 }, { 10, 3606 }, { 10, 3608 },
+  { 10, 3610 }, { 10, 3612 }, { 10, 3614 }, { 10, 3616 }, { 10, 3618 },
+  { 10, 3620 }, { 10, 3622 }, { 10, 3624 }, { 10, 3626 }, { 10, 3628 },
+  { 10, 3630 }, { 10, 3632 }, { 10, 3634 }, { 10, 3636 }, { 10, 3638 },
+  { 10, 3640 }, { 10, 3642 }, { 10, 3644 }, { 10, 3646 }, { 10, 3648 },
+  { 10, 3650 }, { 10, 3652 }, { 10, 3654 }, { 10, 3656 }, { 10, 3658 },
+  { 10, 3660 }, { 10, 3662 }, { 10, 3664 }, { 10, 3666 }, { 10, 3668 },
+  { 10, 3670 }, { 10, 3672 }, { 10, 3674 }, { 10, 3676 }, { 10, 3678 },
+  { 10, 3680 }, { 10, 3682 }, { 10, 3684 }, { 10, 3686 }, { 10, 3688 },
+  { 10, 3690 }, { 10, 3692 }, { 10, 3694 }, { 10, 3696 }, { 10, 3698 },
+  { 10, 3700 }, { 10, 3702 }, { 10, 3704 }, { 10, 3706 }, { 10, 3708 },
+  { 10, 3710 }, { 10, 3712 }, { 10, 3714 }, { 10, 3716 }, { 10, 3718 },
+  { 10, 3720 }, { 10, 3722 }, { 10, 3724 }, { 10, 3726 }, { 10, 3728 },
+  { 10, 3730 }, { 10, 3732 }, { 10, 3734 }, { 10, 3736 }, { 10, 3738 },
+  { 10, 3740 }, { 10, 3742 }, { 10, 3744 }, { 10, 3746 }, { 10, 3748 },
+  { 10, 3750 }, { 10, 3752 }, { 10, 3754 }, { 10, 3756 }, { 10, 3758 },
+  { 10, 3760 }, { 10, 3762 }, { 10, 3764 }, { 10, 3766 }, { 10, 3768 },
+  { 10, 3770 }, { 10, 3772 }, { 10, 3774 }, { 10, 3776 }, { 10, 3778 },
+  { 10, 3780 }, { 10, 3782 }, { 10, 3784 }, { 10, 3786 }, { 10, 3788 },
+  { 10, 3790 }, { 10, 3792 }, { 10, 3794 }, { 10, 3796 }, { 10, 3798 },
+  { 10, 3800 }, { 10, 3802 }, { 10, 3804 }, { 10, 3806 }, { 10, 3808 },
+  { 10, 3810 }, { 10, 3812 }, { 10, 3814 }, { 10, 3816 }, { 10, 3818 },
+  { 10, 3820 }, { 10, 3822 }, { 10, 3824 }, { 10, 3826 }, { 10, 3828 },
+  { 10, 3830 }, { 10, 3832 }, { 10, 3834 }, { 10, 3836 }, { 10, 3838 },
+  { 10, 3840 }, { 10, 3842 }, { 10, 3844 }, { 10, 3846 }, { 10, 3848 },
+  { 10, 3850 }, { 10, 3852 }, { 10, 3854 }, { 10, 3856 }, { 10, 3858 },
+  { 10, 3860 }, { 10, 3862 }, { 10, 3864 }, { 10, 3866 }, { 10, 3868 },
+  { 10, 3870 }, { 10, 3872 }, { 10, 3874 }, { 10, 3876 }, { 10, 3878 },
+  { 10, 3880 }, { 10, 3882 }, { 10, 3884 }, { 10, 3886 }, { 10, 3888 },
+  { 10, 3890 }, { 10, 3892 }, { 10, 3894 }, { 10, 3896 }, { 10, 3898 },
+  { 10, 3900 }, { 10, 3902 }, { 10, 3904 }, { 10, 3906 }, { 10, 3908 },
+  { 10, 3910 }, { 10, 3912 }, { 10, 3914 }, { 10, 3916 }, { 10, 3918 },
+  { 10, 3920 }, { 10, 3922 }, { 10, 3924 }, { 10, 3926 }, { 10, 3928 },
+  { 10, 3930 }, { 10, 3932 }, { 10, 3934 }, { 10, 3936 }, { 10, 3938 },
+  { 10, 3940 }, { 10, 3942 }, { 10, 3944 }, { 10, 3946 }, { 10, 3948 },
+  { 10, 3950 }, { 10, 3952 }, { 10, 3954 }, { 10, 3956 }, { 10, 3958 },
+  { 10, 3960 }
 };
 
 #ifdef __cplusplus
--- a/vp8/encoder/defaultcoefcounts.h
+++ b/vp8/encoder/defaultcoefcounts.h
@@ -17,217 +17,216 @@
 
 /* Generated file, included by entropy.c */
 
-static const unsigned int default_coef_counts[BLOCK_TYPES]
-                                             [COEF_BANDS]
-                                             [PREV_COEF_CONTEXTS]
-                                             [MAX_ENTROPY_TOKENS] =
-{
+static const unsigned int default_coef_counts
+    [BLOCK_TYPES][COEF_BANDS][PREV_COEF_CONTEXTS][MAX_ENTROPY_TOKENS] = {
 
-    {
-        /* Block Type ( 0 ) */
-        {
-            /* Coeff Band ( 0 ) */
-            {   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,},
-            {   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,},
-            {   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,},
-        },
-        {
-            /* Coeff Band ( 1 ) */
-            {30190, 26544, 225,  24,   4,   0,   0,   0,   0,   0,   0, 4171593,},
-            {26846, 25157, 1241, 130,  26,   6,   1,   0,   0,   0,   0, 149987,},
-            {10484, 9538, 1006, 160,  36,  18,   0,   0,   0,   0,   0, 15104,},
-        },
-        {
-            /* Coeff Band ( 2 ) */
-            {25842, 40456, 1126,  83,  11,   2,   0,   0,   0,   0,   0,   0,},
-            {9338, 8010, 512,  73,   7,   3,   2,   0,   0,   0,   0, 43294,},
-            {1047, 751, 149,  31,  13,   6,   1,   0,   0,   0,   0, 879,},
-        },
-        {
-            /* Coeff Band ( 3 ) */
-            {26136, 9826, 252,  13,   0,   0,   0,   0,   0,   0,   0,   0,},
-            {8134, 5574, 191,  14,   2,   0,   0,   0,   0,   0,   0, 35302,},
-            { 605, 677, 116,   9,   1,   0,   0,   0,   0,   0,   0, 611,},
-        },
-        {
-            /* Coeff Band ( 4 ) */
-            {10263, 15463, 283,  17,   0,   0,   0,   0,   0,   0,   0,   0,},
-            {2773, 2191, 128,   9,   2,   2,   0,   0,   0,   0,   0, 10073,},
-            { 134, 125,  32,   4,   0,   2,   0,   0,   0,   0,   0,  50,},
-        },
-        {
-            /* Coeff Band ( 5 ) */
-            {10483, 2663,  23,   1,   0,   0,   0,   0,   0,   0,   0,   0,},
-            {2137, 1251,  27,   1,   1,   0,   0,   0,   0,   0,   0, 14362,},
-            { 116, 156,  14,   2,   1,   0,   0,   0,   0,   0,   0, 190,},
-        },
-        {
-            /* Coeff Band ( 6 ) */
-            {40977, 27614, 412,  28,   0,   0,   0,   0,   0,   0,   0,   0,},
-            {6113, 5213, 261,  22,   3,   0,   0,   0,   0,   0,   0, 26164,},
-            { 382, 312,  50,  14,   2,   0,   0,   0,   0,   0,   0, 345,},
-        },
-        {
-            /* Coeff Band ( 7 ) */
-            {   0,  26,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,},
-            {   0,  13,   0,   0,   0,   0,   0,   0,   0,   0,   0, 319,},
-            {   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   8,},
-        },
-    },
-    {
-        /* Block Type ( 1 ) */
-        {
-            /* Coeff Band ( 0 ) */
-            {3268, 19382, 1043, 250,  93,  82,  49,  26,  17,   8,  25, 82289,},
-            {8758, 32110, 5436, 1832, 827, 668, 420, 153,  24,   0,   3, 52914,},
-            {9337, 23725, 8487, 3954, 2107, 1836, 1069, 399,  59,   0,   0, 18620,},
-        },
-        {
-            /* Coeff Band ( 1 ) */
-            {12419, 8420, 452,  62,   9,   1,   0,   0,   0,   0,   0,   0,},
-            {11715, 8705, 693,  92,  15,   7,   2,   0,   0,   0,   0, 53988,},
-            {7603, 8585, 2306, 778, 270, 145,  39,   5,   0,   0,   0, 9136,},
-        },
-        {
-            /* Coeff Band ( 2 ) */
-            {15938, 14335, 1207, 184,  55,  13,   4,   1,   0,   0,   0,   0,},
-            {7415, 6829, 1138, 244,  71,  26,   7,   0,   0,   0,   0, 9980,},
-            {1580, 1824, 655, 241,  89,  46,  10,   2,   0,   0,   0, 429,},
-        },
-        {
-            /* Coeff Band ( 3 ) */
-            {19453, 5260, 201,  19,   0,   0,   0,   0,   0,   0,   0,   0,},
-            {9173, 3758, 213,  22,   1,   1,   0,   0,   0,   0,   0, 9820,},
-            {1689, 1277, 276,  51,  17,   4,   0,   0,   0,   0,   0, 679,},
-        },
-        {
-            /* Coeff Band ( 4 ) */
-            {12076, 10667, 620,  85,  19,   9,   5,   0,   0,   0,   0,   0,},
-            {4665, 3625, 423,  55,  19,   9,   0,   0,   0,   0,   0, 5127,},
-            { 415, 440, 143,  34,  20,   7,   2,   0,   0,   0,   0, 101,},
-        },
-        {
-            /* Coeff Band ( 5 ) */
-            {12183, 4846, 115,  11,   1,   0,   0,   0,   0,   0,   0,   0,},
-            {4226, 3149, 177,  21,   2,   0,   0,   0,   0,   0,   0, 7157,},
-            { 375, 621, 189,  51,  11,   4,   1,   0,   0,   0,   0, 198,},
-        },
-        {
-            /* Coeff Band ( 6 ) */
-            {61658, 37743, 1203,  94,  10,   3,   0,   0,   0,   0,   0,   0,},
-            {15514, 11563, 903, 111,  14,   5,   0,   0,   0,   0,   0, 25195,},
-            { 929, 1077, 291,  78,  14,   7,   1,   0,   0,   0,   0, 507,},
-        },
-        {
-            /* Coeff Band ( 7 ) */
-            {   0, 990,  15,   3,   0,   0,   0,   0,   0,   0,   0,   0,},
-            {   0, 412,  13,   0,   0,   0,   0,   0,   0,   0,   0, 1641,},
-            {   0,  18,   7,   1,   0,   0,   0,   0,   0,   0,   0,  30,},
-        },
-    },
-    {
-        /* Block Type ( 2 ) */
-        {
-            /* Coeff Band ( 0 ) */
-            { 953, 24519, 628, 120,  28,  12,   4,   0,   0,   0,   0, 2248798,},
-            {1525, 25654, 2647, 617, 239, 143,  42,   5,   0,   0,   0, 66837,},
-            {1180, 11011, 3001, 1237, 532, 448, 239,  54,   5,   0,   0, 7122,},
-        },
-        {
-            /* Coeff Band ( 1 ) */
-            {1356, 2220,  67,  10,   4,   1,   0,   0,   0,   0,   0,   0,},
-            {1450, 2544, 102,  18,   4,   3,   0,   0,   0,   0,   0, 57063,},
-            {1182, 2110, 470, 130,  41,  21,   0,   0,   0,   0,   0, 6047,},
-        },
-        {
-            /* Coeff Band ( 2 ) */
-            { 370, 3378, 200,  30,   5,   4,   1,   0,   0,   0,   0,   0,},
-            { 293, 1006, 131,  29,  11,   0,   0,   0,   0,   0,   0, 5404,},
-            { 114, 387,  98,  23,   4,   8,   1,   0,   0,   0,   0, 236,},
-        },
-        {
-            /* Coeff Band ( 3 ) */
-            { 579, 194,   4,   0,   0,   0,   0,   0,   0,   0,   0,   0,},
-            { 395, 213,   5,   1,   0,   0,   0,   0,   0,   0,   0, 4157,},
-            { 119, 122,   4,   0,   0,   0,   0,   0,   0,   0,   0, 300,},
-        },
-        {
-            /* Coeff Band ( 4 ) */
-            {  38, 557,  19,   0,   0,   0,   0,   0,   0,   0,   0,   0,},
-            {  21, 114,  12,   1,   0,   0,   0,   0,   0,   0,   0, 427,},
-            {   0,   5,   0,   0,   0,   0,   0,   0,   0,   0,   0,   7,},
-        },
-        {
-            /* Coeff Band ( 5 ) */
-            {  52,   7,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,},
-            {  18,   6,   0,   0,   0,   0,   0,   0,   0,   0,   0, 652,},
-            {   1,   1,   0,   0,   0,   0,   0,   0,   0,   0,   0,  30,},
-        },
-        {
-            /* Coeff Band ( 6 ) */
-            { 640, 569,  10,   0,   0,   0,   0,   0,   0,   0,   0,   0,},
-            {  25,  77,   2,   0,   0,   0,   0,   0,   0,   0,   0, 517,},
-            {   4,   7,   0,   0,   0,   0,   0,   0,   0,   0,   0,   3,},
-        },
-        {
-            /* Coeff Band ( 7 ) */
-            {   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,},
-            {   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,},
-            {   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,},
-        },
-    },
-    {
-        /* Block Type ( 3 ) */
-        {
-            /* Coeff Band ( 0 ) */
-            {2506, 20161, 2707, 767, 261, 178, 107,  30,  14,   3,   0, 100694,},
-            {8806, 36478, 8817, 3268, 1280, 850, 401, 114,  42,   0,   0, 58572,},
-            {11003, 27214, 11798, 5716, 2482, 2072, 1048, 175,  32,   0,   0, 19284,},
-        },
-        {
-            /* Coeff Band ( 1 ) */
-            {9738, 11313, 959, 205,  70,  18,  11,   1,   0,   0,   0,   0,},
-            {12628, 15085, 1507, 273,  52,  19,   9,   0,   0,   0,   0, 54280,},
-            {10701, 15846, 5561, 1926, 813, 570, 249,  36,   0,   0,   0, 6460,},
-        },
-        {
-            /* Coeff Band ( 2 ) */
-            {6781, 22539, 2784, 634, 182, 123,  20,   4,   0,   0,   0,   0,},
-            {6263, 11544, 2649, 790, 259, 168,  27,   5,   0,   0,   0, 20539,},
-            {3109, 4075, 2031, 896, 457, 386, 158,  29,   0,   0,   0, 1138,},
-        },
-        {
-            /* Coeff Band ( 3 ) */
-            {11515, 4079, 465,  73,   5,  14,   2,   0,   0,   0,   0,   0,},
-            {9361, 5834, 650,  96,  24,   8,   4,   0,   0,   0,   0, 22181,},
-            {4343, 3974, 1360, 415, 132,  96,  14,   1,   0,   0,   0, 1267,},
-        },
-        {
-            /* Coeff Band ( 4 ) */
-            {4787, 9297, 823, 168,  44,  12,   4,   0,   0,   0,   0,   0,},
-            {3619, 4472, 719, 198,  60,  31,   3,   0,   0,   0,   0, 8401,},
-            {1157, 1175, 483, 182,  88,  31,   8,   0,   0,   0,   0, 268,},
-        },
-        {
-            /* Coeff Band ( 5 ) */
-            {8299, 1226,  32,   5,   1,   0,   0,   0,   0,   0,   0,   0,},
-            {3502, 1568,  57,   4,   1,   1,   0,   0,   0,   0,   0, 9811,},
-            {1055, 1070, 166,  29,   6,   1,   0,   0,   0,   0,   0, 527,},
-        },
-        {
-            /* Coeff Band ( 6 ) */
-            {27414, 27927, 1989, 347,  69,  26,   0,   0,   0,   0,   0,   0,},
-            {5876, 10074, 1574, 341,  91,  24,   4,   0,   0,   0,   0, 21954,},
-            {1571, 2171, 778, 324, 124,  65,  16,   0,   0,   0,   0, 979,},
-        },
-        {
-            /* Coeff Band ( 7 ) */
-            {   0,  29,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,},
-            {   0,  23,   0,   0,   0,   0,   0,   0,   0,   0,   0, 459,},
-            {   0,   1,   0,   0,   0,   0,   0,   0,   0,   0,   0,  13,},
-        },
-    },
-};
+      {
+          /* Block Type ( 0 ) */
+          {
+              /* Coeff Band ( 0 ) */
+              { 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 },
+              { 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 },
+              { 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 },
+          },
+          {
+              /* Coeff Band ( 1 ) */
+              { 30190, 26544, 225, 24, 4, 0, 0, 0, 0, 0, 0, 4171593 },
+              { 26846, 25157, 1241, 130, 26, 6, 1, 0, 0, 0, 0, 149987 },
+              { 10484, 9538, 1006, 160, 36, 18, 0, 0, 0, 0, 0, 15104 },
+          },
+          {
+              /* Coeff Band ( 2 ) */
+              { 25842, 40456, 1126, 83, 11, 2, 0, 0, 0, 0, 0, 0 },
+              { 9338, 8010, 512, 73, 7, 3, 2, 0, 0, 0, 0, 43294 },
+              { 1047, 751, 149, 31, 13, 6, 1, 0, 0, 0, 0, 879 },
+          },
+          {
+              /* Coeff Band ( 3 ) */
+              { 26136, 9826, 252, 13, 0, 0, 0, 0, 0, 0, 0, 0 },
+              { 8134, 5574, 191, 14, 2, 0, 0, 0, 0, 0, 0, 35302 },
+              { 605, 677, 116, 9, 1, 0, 0, 0, 0, 0, 0, 611 },
+          },
+          {
+              /* Coeff Band ( 4 ) */
+              { 10263, 15463, 283, 17, 0, 0, 0, 0, 0, 0, 0, 0 },
+              { 2773, 2191, 128, 9, 2, 2, 0, 0, 0, 0, 0, 10073 },
+              { 134, 125, 32, 4, 0, 2, 0, 0, 0, 0, 0, 50 },
+          },
+          {
+              /* Coeff Band ( 5 ) */
+              { 10483, 2663, 23, 1, 0, 0, 0, 0, 0, 0, 0, 0 },
+              { 2137, 1251, 27, 1, 1, 0, 0, 0, 0, 0, 0, 14362 },
+              { 116, 156, 14, 2, 1, 0, 0, 0, 0, 0, 0, 190 },
+          },
+          {
+              /* Coeff Band ( 6 ) */
+              { 40977, 27614, 412, 28, 0, 0, 0, 0, 0, 0, 0, 0 },
+              { 6113, 5213, 261, 22, 3, 0, 0, 0, 0, 0, 0, 26164 },
+              { 382, 312, 50, 14, 2, 0, 0, 0, 0, 0, 0, 345 },
+          },
+          {
+              /* Coeff Band ( 7 ) */
+              { 0, 26, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 },
+              { 0, 13, 0, 0, 0, 0, 0, 0, 0, 0, 0, 319 },
+              { 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 8 },
+          },
+      },
+      {
+          /* Block Type ( 1 ) */
+          {
+              /* Coeff Band ( 0 ) */
+              { 3268, 19382, 1043, 250, 93, 82, 49, 26, 17, 8, 25, 82289 },
+              { 8758, 32110, 5436, 1832, 827, 668, 420, 153, 24, 0, 3, 52914 },
+              { 9337, 23725, 8487, 3954, 2107, 1836, 1069, 399, 59, 0, 0,
+                18620 },
+          },
+          {
+              /* Coeff Band ( 1 ) */
+              { 12419, 8420, 452, 62, 9, 1, 0, 0, 0, 0, 0, 0 },
+              { 11715, 8705, 693, 92, 15, 7, 2, 0, 0, 0, 0, 53988 },
+              { 7603, 8585, 2306, 778, 270, 145, 39, 5, 0, 0, 0, 9136 },
+          },
+          {
+              /* Coeff Band ( 2 ) */
+              { 15938, 14335, 1207, 184, 55, 13, 4, 1, 0, 0, 0, 0 },
+              { 7415, 6829, 1138, 244, 71, 26, 7, 0, 0, 0, 0, 9980 },
+              { 1580, 1824, 655, 241, 89, 46, 10, 2, 0, 0, 0, 429 },
+          },
+          {
+              /* Coeff Band ( 3 ) */
+              { 19453, 5260, 201, 19, 0, 0, 0, 0, 0, 0, 0, 0 },
+              { 9173, 3758, 213, 22, 1, 1, 0, 0, 0, 0, 0, 9820 },
+              { 1689, 1277, 276, 51, 17, 4, 0, 0, 0, 0, 0, 679 },
+          },
+          {
+              /* Coeff Band ( 4 ) */
+              { 12076, 10667, 620, 85, 19, 9, 5, 0, 0, 0, 0, 0 },
+              { 4665, 3625, 423, 55, 19, 9, 0, 0, 0, 0, 0, 5127 },
+              { 415, 440, 143, 34, 20, 7, 2, 0, 0, 0, 0, 101 },
+          },
+          {
+              /* Coeff Band ( 5 ) */
+              { 12183, 4846, 115, 11, 1, 0, 0, 0, 0, 0, 0, 0 },
+              { 4226, 3149, 177, 21, 2, 0, 0, 0, 0, 0, 0, 7157 },
+              { 375, 621, 189, 51, 11, 4, 1, 0, 0, 0, 0, 198 },
+          },
+          {
+              /* Coeff Band ( 6 ) */
+              { 61658, 37743, 1203, 94, 10, 3, 0, 0, 0, 0, 0, 0 },
+              { 15514, 11563, 903, 111, 14, 5, 0, 0, 0, 0, 0, 25195 },
+              { 929, 1077, 291, 78, 14, 7, 1, 0, 0, 0, 0, 507 },
+          },
+          {
+              /* Coeff Band ( 7 ) */
+              { 0, 990, 15, 3, 0, 0, 0, 0, 0, 0, 0, 0 },
+              { 0, 412, 13, 0, 0, 0, 0, 0, 0, 0, 0, 1641 },
+              { 0, 18, 7, 1, 0, 0, 0, 0, 0, 0, 0, 30 },
+          },
+      },
+      {
+          /* Block Type ( 2 ) */
+          {
+              /* Coeff Band ( 0 ) */
+              { 953, 24519, 628, 120, 28, 12, 4, 0, 0, 0, 0, 2248798 },
+              { 1525, 25654, 2647, 617, 239, 143, 42, 5, 0, 0, 0, 66837 },
+              { 1180, 11011, 3001, 1237, 532, 448, 239, 54, 5, 0, 0, 7122 },
+          },
+          {
+              /* Coeff Band ( 1 ) */
+              { 1356, 2220, 67, 10, 4, 1, 0, 0, 0, 0, 0, 0 },
+              { 1450, 2544, 102, 18, 4, 3, 0, 0, 0, 0, 0, 57063 },
+              { 1182, 2110, 470, 130, 41, 21, 0, 0, 0, 0, 0, 6047 },
+          },
+          {
+              /* Coeff Band ( 2 ) */
+              { 370, 3378, 200, 30, 5, 4, 1, 0, 0, 0, 0, 0 },
+              { 293, 1006, 131, 29, 11, 0, 0, 0, 0, 0, 0, 5404 },
+              { 114, 387, 98, 23, 4, 8, 1, 0, 0, 0, 0, 236 },
+          },
+          {
+              /* Coeff Band ( 3 ) */
+              { 579, 194, 4, 0, 0, 0, 0, 0, 0, 0, 0, 0 },
+              { 395, 213, 5, 1, 0, 0, 0, 0, 0, 0, 0, 4157 },
+              { 119, 122, 4, 0, 0, 0, 0, 0, 0, 0, 0, 300 },
+          },
+          {
+              /* Coeff Band ( 4 ) */
+              { 38, 557, 19, 0, 0, 0, 0, 0, 0, 0, 0, 0 },
+              { 21, 114, 12, 1, 0, 0, 0, 0, 0, 0, 0, 427 },
+              { 0, 5, 0, 0, 0, 0, 0, 0, 0, 0, 0, 7 },
+          },
+          {
+              /* Coeff Band ( 5 ) */
+              { 52, 7, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 },
+              { 18, 6, 0, 0, 0, 0, 0, 0, 0, 0, 0, 652 },
+              { 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 30 },
+          },
+          {
+              /* Coeff Band ( 6 ) */
+              { 640, 569, 10, 0, 0, 0, 0, 0, 0, 0, 0, 0 },
+              { 25, 77, 2, 0, 0, 0, 0, 0, 0, 0, 0, 517 },
+              { 4, 7, 0, 0, 0, 0, 0, 0, 0, 0, 0, 3 },
+          },
+          {
+              /* Coeff Band ( 7 ) */
+              { 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 },
+              { 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 },
+              { 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 },
+          },
+      },
+      {
+          /* Block Type ( 3 ) */
+          {
+              /* Coeff Band ( 0 ) */
+              { 2506, 20161, 2707, 767, 261, 178, 107, 30, 14, 3, 0, 100694 },
+              { 8806, 36478, 8817, 3268, 1280, 850, 401, 114, 42, 0, 0, 58572 },
+              { 11003, 27214, 11798, 5716, 2482, 2072, 1048, 175, 32, 0, 0,
+                19284 },
+          },
+          {
+              /* Coeff Band ( 1 ) */
+              { 9738, 11313, 959, 205, 70, 18, 11, 1, 0, 0, 0, 0 },
+              { 12628, 15085, 1507, 273, 52, 19, 9, 0, 0, 0, 0, 54280 },
+              { 10701, 15846, 5561, 1926, 813, 570, 249, 36, 0, 0, 0, 6460 },
+          },
+          {
+              /* Coeff Band ( 2 ) */
+              { 6781, 22539, 2784, 634, 182, 123, 20, 4, 0, 0, 0, 0 },
+              { 6263, 11544, 2649, 790, 259, 168, 27, 5, 0, 0, 0, 20539 },
+              { 3109, 4075, 2031, 896, 457, 386, 158, 29, 0, 0, 0, 1138 },
+          },
+          {
+              /* Coeff Band ( 3 ) */
+              { 11515, 4079, 465, 73, 5, 14, 2, 0, 0, 0, 0, 0 },
+              { 9361, 5834, 650, 96, 24, 8, 4, 0, 0, 0, 0, 22181 },
+              { 4343, 3974, 1360, 415, 132, 96, 14, 1, 0, 0, 0, 1267 },
+          },
+          {
+              /* Coeff Band ( 4 ) */
+              { 4787, 9297, 823, 168, 44, 12, 4, 0, 0, 0, 0, 0 },
+              { 3619, 4472, 719, 198, 60, 31, 3, 0, 0, 0, 0, 8401 },
+              { 1157, 1175, 483, 182, 88, 31, 8, 0, 0, 0, 0, 268 },
+          },
+          {
+              /* Coeff Band ( 5 ) */
+              { 8299, 1226, 32, 5, 1, 0, 0, 0, 0, 0, 0, 0 },
+              { 3502, 1568, 57, 4, 1, 1, 0, 0, 0, 0, 0, 9811 },
+              { 1055, 1070, 166, 29, 6, 1, 0, 0, 0, 0, 0, 527 },
+          },
+          {
+              /* Coeff Band ( 6 ) */
+              { 27414, 27927, 1989, 347, 69, 26, 0, 0, 0, 0, 0, 0 },
+              { 5876, 10074, 1574, 341, 91, 24, 4, 0, 0, 0, 0, 21954 },
+              { 1571, 2171, 778, 324, 124, 65, 16, 0, 0, 0, 0, 979 },
+          },
+          {
+              /* Coeff Band ( 7 ) */
+              { 0, 29, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 },
+              { 0, 23, 0, 0, 0, 0, 0, 0, 0, 0, 0, 459 },
+              { 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 13 },
+          },
+      },
+    };
 
 #ifdef __cplusplus
 }  // extern "C"
--- a/vp8/encoder/denoising.c
+++ b/vp8/encoder/denoising.c
@@ -58,311 +58,291 @@
                           unsigned char *running_avg_y, int avg_y_stride,
                           unsigned char *sig, int sig_stride,
                           unsigned int motion_magnitude,
-                          int increase_denoising)
-{
-    unsigned char *running_avg_y_start = running_avg_y;
-    unsigned char *sig_start = sig;
-    int sum_diff_thresh;
-    int r, c;
-    int sum_diff = 0;
-    int adj_val[3] = {3, 4, 6};
-    int shift_inc1 = 0;
-    int shift_inc2 = 1;
-    int col_sum[16] = {0, 0, 0, 0,
-                       0, 0, 0, 0,
-                       0, 0, 0, 0,
-                       0, 0, 0, 0};
-    /* If motion_magnitude is small, making the denoiser more aggressive by
-     * increasing the adjustment for each level. Add another increment for
-     * blocks that are labeled for increase denoising. */
-    if (motion_magnitude <= MOTION_MAGNITUDE_THRESHOLD)
-    {
-      if (increase_denoising) {
-        shift_inc1 = 1;
-        shift_inc2 = 2;
-      }
-      adj_val[0] += shift_inc2;
-      adj_val[1] += shift_inc2;
-      adj_val[2] += shift_inc2;
+                          int increase_denoising) {
+  unsigned char *running_avg_y_start = running_avg_y;
+  unsigned char *sig_start = sig;
+  int sum_diff_thresh;
+  int r, c;
+  int sum_diff = 0;
+  int adj_val[3] = { 3, 4, 6 };
+  int shift_inc1 = 0;
+  int shift_inc2 = 1;
+  int col_sum[16] = { 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 };
+  /* If motion_magnitude is small, making the denoiser more aggressive by
+   * increasing the adjustment for each level. Add another increment for
+   * blocks that are labeled for increase denoising. */
+  if (motion_magnitude <= MOTION_MAGNITUDE_THRESHOLD) {
+    if (increase_denoising) {
+      shift_inc1 = 1;
+      shift_inc2 = 2;
     }
+    adj_val[0] += shift_inc2;
+    adj_val[1] += shift_inc2;
+    adj_val[2] += shift_inc2;
+  }
 
-    for (r = 0; r < 16; ++r)
-    {
-        for (c = 0; c < 16; ++c)
-        {
-            int diff = 0;
-            int adjustment = 0;
-            int absdiff = 0;
+  for (r = 0; r < 16; ++r) {
+    for (c = 0; c < 16; ++c) {
+      int diff = 0;
+      int adjustment = 0;
+      int absdiff = 0;
 
-            diff = mc_running_avg_y[c] - sig[c];
-            absdiff = abs(diff);
+      diff = mc_running_avg_y[c] - sig[c];
+      absdiff = abs(diff);
 
-            // When |diff| <= |3 + shift_inc1|, use pixel value from
-            // last denoised raw.
-            if (absdiff <= 3 + shift_inc1)
-            {
-                running_avg_y[c] = mc_running_avg_y[c];
-                col_sum[c] += diff;
-            }
-            else
-            {
-                if (absdiff >= 4 + shift_inc1 && absdiff <= 7)
-                    adjustment = adj_val[0];
-                else if (absdiff >= 8 && absdiff <= 15)
-                    adjustment = adj_val[1];
-                else
-                    adjustment = adj_val[2];
+      // When |diff| <= |3 + shift_inc1|, use pixel value from
+      // last denoised raw.
+      if (absdiff <= 3 + shift_inc1) {
+        running_avg_y[c] = mc_running_avg_y[c];
+        col_sum[c] += diff;
+      } else {
+        if (absdiff >= 4 + shift_inc1 && absdiff <= 7)
+          adjustment = adj_val[0];
+        else if (absdiff >= 8 && absdiff <= 15)
+          adjustment = adj_val[1];
+        else
+          adjustment = adj_val[2];
 
-                if (diff > 0)
-                {
-                    if ((sig[c] + adjustment) > 255)
-                        running_avg_y[c] = 255;
-                    else
-                        running_avg_y[c] = sig[c] + adjustment;
+        if (diff > 0) {
+          if ((sig[c] + adjustment) > 255)
+            running_avg_y[c] = 255;
+          else
+            running_avg_y[c] = sig[c] + adjustment;
 
-                    col_sum[c] += adjustment;
-                }
-                else
-                {
-                    if ((sig[c] - adjustment) < 0)
-                        running_avg_y[c] = 0;
-                    else
-                        running_avg_y[c] = sig[c] - adjustment;
+          col_sum[c] += adjustment;
+        } else {
+          if ((sig[c] - adjustment) < 0)
+            running_avg_y[c] = 0;
+          else
+            running_avg_y[c] = sig[c] - adjustment;
 
-                    col_sum[c] -= adjustment;
-                }
-            }
+          col_sum[c] -= adjustment;
         }
-
-        /* Update pointers for next iteration. */
-        sig += sig_stride;
-        mc_running_avg_y += mc_avg_y_stride;
-        running_avg_y += avg_y_stride;
-    }
-
-    for (c = 0; c < 16; ++c) {
-      // Below we clip the value in the same way which SSE code use.
-      // When adopting aggressive denoiser, the adj_val for each pixel
-      // could be at most 8 (this is current max adjustment of the map).
-      // In SSE code, we calculate the sum of adj_val for
-      // the columns, so the sum could be upto 128(16 rows). However,
-      // the range of the value is -128 ~ 127 in SSE code, that's why
-      // we do this change in C code.
-      // We don't do this for UV denoiser, since there are only 8 rows,
-      // and max adjustments <= 8, so the sum of the columns will not
-      // exceed 64.
-      if (col_sum[c] >= 128) {
-        col_sum[c] = 127;
       }
-      sum_diff += col_sum[c];
     }
 
-    sum_diff_thresh= SUM_DIFF_THRESHOLD;
-    if (increase_denoising) sum_diff_thresh = SUM_DIFF_THRESHOLD_HIGH;
-    if (abs(sum_diff) > sum_diff_thresh) {
-      // Before returning to copy the block (i.e., apply no denoising), check
-      // if we can still apply some (weaker) temporal filtering to this block,
-      // that would otherwise not be denoised at all. Simplest is to apply
-      // an additional adjustment to running_avg_y to bring it closer to sig.
-      // The adjustment is capped by a maximum delta, and chosen such that
-      // in most cases the resulting sum_diff will be within the
-      // accceptable range given by sum_diff_thresh.
+    /* Update pointers for next iteration. */
+    sig += sig_stride;
+    mc_running_avg_y += mc_avg_y_stride;
+    running_avg_y += avg_y_stride;
+  }
 
-      // The delta is set by the excess of absolute pixel diff over threshold.
-      int delta = ((abs(sum_diff) - sum_diff_thresh) >> 8) + 1;
-      // Only apply the adjustment for max delta up to 3.
-      if (delta < 4) {
-        sig -= sig_stride * 16;
-        mc_running_avg_y -= mc_avg_y_stride * 16;
-        running_avg_y -= avg_y_stride * 16;
-        for (r = 0; r < 16; ++r) {
-          for (c = 0; c < 16; ++c) {
-            int diff = mc_running_avg_y[c] - sig[c];
-            int adjustment = abs(diff);
-            if (adjustment > delta)
-              adjustment = delta;
-            if (diff > 0) {
-              // Bring denoised signal down.
-              if (running_avg_y[c] - adjustment < 0)
-                running_avg_y[c] = 0;
-              else
-                running_avg_y[c] = running_avg_y[c] - adjustment;
-              col_sum[c] -= adjustment;
-            } else if (diff < 0) {
-              // Bring denoised signal up.
-              if (running_avg_y[c] + adjustment > 255)
-                running_avg_y[c] = 255;
-              else
-                running_avg_y[c] = running_avg_y[c] + adjustment;
-              col_sum[c] += adjustment;
-            }
-          }
-          // TODO(marpan): Check here if abs(sum_diff) has gone below the
-          // threshold sum_diff_thresh, and if so, we can exit the row loop.
-          sig += sig_stride;
-          mc_running_avg_y += mc_avg_y_stride;
-          running_avg_y += avg_y_stride;
-        }
+  for (c = 0; c < 16; ++c) {
+    // Below we clip the value in the same way which SSE code use.
+    // When adopting aggressive denoiser, the adj_val for each pixel
+    // could be at most 8 (this is current max adjustment of the map).
+    // In SSE code, we calculate the sum of adj_val for
+    // the columns, so the sum could be upto 128(16 rows). However,
+    // the range of the value is -128 ~ 127 in SSE code, that's why
+    // we do this change in C code.
+    // We don't do this for UV denoiser, since there are only 8 rows,
+    // and max adjustments <= 8, so the sum of the columns will not
+    // exceed 64.
+    if (col_sum[c] >= 128) {
+      col_sum[c] = 127;
+    }
+    sum_diff += col_sum[c];
+  }
 
-        sum_diff = 0;
+  sum_diff_thresh = SUM_DIFF_THRESHOLD;
+  if (increase_denoising) sum_diff_thresh = SUM_DIFF_THRESHOLD_HIGH;
+  if (abs(sum_diff) > sum_diff_thresh) {
+    // Before returning to copy the block (i.e., apply no denoising), check
+    // if we can still apply some (weaker) temporal filtering to this block,
+    // that would otherwise not be denoised at all. Simplest is to apply
+    // an additional adjustment to running_avg_y to bring it closer to sig.
+    // The adjustment is capped by a maximum delta, and chosen such that
+    // in most cases the resulting sum_diff will be within the
+    // accceptable range given by sum_diff_thresh.
+
+    // The delta is set by the excess of absolute pixel diff over threshold.
+    int delta = ((abs(sum_diff) - sum_diff_thresh) >> 8) + 1;
+    // Only apply the adjustment for max delta up to 3.
+    if (delta < 4) {
+      sig -= sig_stride * 16;
+      mc_running_avg_y -= mc_avg_y_stride * 16;
+      running_avg_y -= avg_y_stride * 16;
+      for (r = 0; r < 16; ++r) {
         for (c = 0; c < 16; ++c) {
-          if (col_sum[c] >= 128) {
-            col_sum[c] = 127;
+          int diff = mc_running_avg_y[c] - sig[c];
+          int adjustment = abs(diff);
+          if (adjustment > delta) adjustment = delta;
+          if (diff > 0) {
+            // Bring denoised signal down.
+            if (running_avg_y[c] - adjustment < 0)
+              running_avg_y[c] = 0;
+            else
+              running_avg_y[c] = running_avg_y[c] - adjustment;
+            col_sum[c] -= adjustment;
+          } else if (diff < 0) {
+            // Bring denoised signal up.
+            if (running_avg_y[c] + adjustment > 255)
+              running_avg_y[c] = 255;
+            else
+              running_avg_y[c] = running_avg_y[c] + adjustment;
+            col_sum[c] += adjustment;
           }
-          sum_diff += col_sum[c];
         }
+        // TODO(marpan): Check here if abs(sum_diff) has gone below the
+        // threshold sum_diff_thresh, and if so, we can exit the row loop.
+        sig += sig_stride;
+        mc_running_avg_y += mc_avg_y_stride;
+        running_avg_y += avg_y_stride;
+      }
 
-        if (abs(sum_diff) > sum_diff_thresh)
-          return COPY_BLOCK;
-      } else {
-        return COPY_BLOCK;
+      sum_diff = 0;
+      for (c = 0; c < 16; ++c) {
+        if (col_sum[c] >= 128) {
+          col_sum[c] = 127;
+        }
+        sum_diff += col_sum[c];
       }
+
+      if (abs(sum_diff) > sum_diff_thresh) return COPY_BLOCK;
+    } else {
+      return COPY_BLOCK;
     }
+  }
 
-    vp8_copy_mem16x16(running_avg_y_start, avg_y_stride, sig_start, sig_stride);
-    return FILTER_BLOCK;
+  vp8_copy_mem16x16(running_avg_y_start, avg_y_stride, sig_start, sig_stride);
+  return FILTER_BLOCK;
 }
 
 int vp8_denoiser_filter_uv_c(unsigned char *mc_running_avg_uv,
                              int mc_avg_uv_stride,
-                             unsigned char *running_avg_uv,
-                             int avg_uv_stride,
-                             unsigned char *sig,
-                             int sig_stride,
+                             unsigned char *running_avg_uv, int avg_uv_stride,
+                             unsigned char *sig, int sig_stride,
                              unsigned int motion_magnitude,
                              int increase_denoising) {
-    unsigned char *running_avg_uv_start = running_avg_uv;
-    unsigned char *sig_start = sig;
-    int sum_diff_thresh;
-    int r, c;
-    int sum_diff = 0;
-    int sum_block = 0;
-    int adj_val[3] = {3, 4, 6};
-    int shift_inc1 = 0;
-    int shift_inc2 = 1;
-    /* If motion_magnitude is small, making the denoiser more aggressive by
-     * increasing the adjustment for each level. Add another increment for
-     * blocks that are labeled for increase denoising. */
-    if (motion_magnitude <= MOTION_MAGNITUDE_THRESHOLD_UV) {
-      if (increase_denoising) {
-        shift_inc1 = 1;
-        shift_inc2 = 2;
-      }
-      adj_val[0] += shift_inc2;
-      adj_val[1] += shift_inc2;
-      adj_val[2] += shift_inc2;
+  unsigned char *running_avg_uv_start = running_avg_uv;
+  unsigned char *sig_start = sig;
+  int sum_diff_thresh;
+  int r, c;
+  int sum_diff = 0;
+  int sum_block = 0;
+  int adj_val[3] = { 3, 4, 6 };
+  int shift_inc1 = 0;
+  int shift_inc2 = 1;
+  /* If motion_magnitude is small, making the denoiser more aggressive by
+   * increasing the adjustment for each level. Add another increment for
+   * blocks that are labeled for increase denoising. */
+  if (motion_magnitude <= MOTION_MAGNITUDE_THRESHOLD_UV) {
+    if (increase_denoising) {
+      shift_inc1 = 1;
+      shift_inc2 = 2;
     }
+    adj_val[0] += shift_inc2;
+    adj_val[1] += shift_inc2;
+    adj_val[2] += shift_inc2;
+  }
 
-    // Avoid denoising color signal if its close to average level.
-    for (r = 0; r < 8; ++r) {
-      for (c = 0; c < 8; ++c) {
-        sum_block += sig[c];
-      }
-      sig += sig_stride;
+  // Avoid denoising color signal if its close to average level.
+  for (r = 0; r < 8; ++r) {
+    for (c = 0; c < 8; ++c) {
+      sum_block += sig[c];
     }
-    if (abs(sum_block - (128 * 8 * 8)) < SUM_DIFF_FROM_AVG_THRESH_UV) {
-      return COPY_BLOCK;
-    }
+    sig += sig_stride;
+  }
+  if (abs(sum_block - (128 * 8 * 8)) < SUM_DIFF_FROM_AVG_THRESH_UV) {
+    return COPY_BLOCK;
+  }
 
-    sig -= sig_stride * 8;
-    for (r = 0; r < 8; ++r) {
-      for (c = 0; c < 8; ++c) {
-        int diff = 0;
-        int adjustment = 0;
-        int absdiff = 0;
+  sig -= sig_stride * 8;
+  for (r = 0; r < 8; ++r) {
+    for (c = 0; c < 8; ++c) {
+      int diff = 0;
+      int adjustment = 0;
+      int absdiff = 0;
 
-        diff = mc_running_avg_uv[c] - sig[c];
-        absdiff = abs(diff);
+      diff = mc_running_avg_uv[c] - sig[c];
+      absdiff = abs(diff);
 
-        // When |diff| <= |3 + shift_inc1|, use pixel value from
-        // last denoised raw.
-        if (absdiff <= 3 + shift_inc1) {
-          running_avg_uv[c] = mc_running_avg_uv[c];
-          sum_diff += diff;
+      // When |diff| <= |3 + shift_inc1|, use pixel value from
+      // last denoised raw.
+      if (absdiff <= 3 + shift_inc1) {
+        running_avg_uv[c] = mc_running_avg_uv[c];
+        sum_diff += diff;
+      } else {
+        if (absdiff >= 4 && absdiff <= 7)
+          adjustment = adj_val[0];
+        else if (absdiff >= 8 && absdiff <= 15)
+          adjustment = adj_val[1];
+        else
+          adjustment = adj_val[2];
+        if (diff > 0) {
+          if ((sig[c] + adjustment) > 255)
+            running_avg_uv[c] = 255;
+          else
+            running_avg_uv[c] = sig[c] + adjustment;
+          sum_diff += adjustment;
         } else {
-          if (absdiff >= 4 && absdiff <= 7)
-            adjustment = adj_val[0];
-          else if (absdiff >= 8 && absdiff <= 15)
-            adjustment = adj_val[1];
+          if ((sig[c] - adjustment) < 0)
+            running_avg_uv[c] = 0;
           else
-            adjustment = adj_val[2];
-          if (diff > 0) {
-            if ((sig[c] + adjustment) > 255)
-              running_avg_uv[c] = 255;
-            else
-              running_avg_uv[c] = sig[c] + adjustment;
-            sum_diff += adjustment;
-          } else {
-            if ((sig[c] - adjustment) < 0)
-              running_avg_uv[c] = 0;
-            else
-              running_avg_uv[c] = sig[c] - adjustment;
-            sum_diff -= adjustment;
-          }
+            running_avg_uv[c] = sig[c] - adjustment;
+          sum_diff -= adjustment;
         }
       }
-      /* Update pointers for next iteration. */
-      sig += sig_stride;
-      mc_running_avg_uv += mc_avg_uv_stride;
-      running_avg_uv += avg_uv_stride;
     }
+    /* Update pointers for next iteration. */
+    sig += sig_stride;
+    mc_running_avg_uv += mc_avg_uv_stride;
+    running_avg_uv += avg_uv_stride;
+  }
 
-    sum_diff_thresh= SUM_DIFF_THRESHOLD_UV;
-    if (increase_denoising) sum_diff_thresh = SUM_DIFF_THRESHOLD_HIGH_UV;
-    if (abs(sum_diff) > sum_diff_thresh) {
-      // Before returning to copy the block (i.e., apply no denoising), check
-      // if we can still apply some (weaker) temporal filtering to this block,
-      // that would otherwise not be denoised at all. Simplest is to apply
-      // an additional adjustment to running_avg_y to bring it closer to sig.
-      // The adjustment is capped by a maximum delta, and chosen such that
-      // in most cases the resulting sum_diff will be within the
-      // accceptable range given by sum_diff_thresh.
+  sum_diff_thresh = SUM_DIFF_THRESHOLD_UV;
+  if (increase_denoising) sum_diff_thresh = SUM_DIFF_THRESHOLD_HIGH_UV;
+  if (abs(sum_diff) > sum_diff_thresh) {
+    // Before returning to copy the block (i.e., apply no denoising), check
+    // if we can still apply some (weaker) temporal filtering to this block,
+    // that would otherwise not be denoised at all. Simplest is to apply
+    // an additional adjustment to running_avg_y to bring it closer to sig.
+    // The adjustment is capped by a maximum delta, and chosen such that
+    // in most cases the resulting sum_diff will be within the
+    // accceptable range given by sum_diff_thresh.
 
-      // The delta is set by the excess of absolute pixel diff over threshold.
-      int delta = ((abs(sum_diff) - sum_diff_thresh) >> 8) + 1;
-      // Only apply the adjustment for max delta up to 3.
-      if (delta < 4) {
-        sig -= sig_stride * 8;
-        mc_running_avg_uv -= mc_avg_uv_stride * 8;
-        running_avg_uv -= avg_uv_stride * 8;
-        for (r = 0; r < 8; ++r) {
-          for (c = 0; c < 8; ++c) {
-            int diff = mc_running_avg_uv[c] - sig[c];
-            int adjustment = abs(diff);
-            if (adjustment > delta)
-              adjustment = delta;
-            if (diff > 0) {
-              // Bring denoised signal down.
-              if (running_avg_uv[c] - adjustment < 0)
-                running_avg_uv[c] = 0;
-              else
-                running_avg_uv[c] = running_avg_uv[c] - adjustment;
-              sum_diff -= adjustment;
-            } else if (diff < 0) {
-              // Bring denoised signal up.
-              if (running_avg_uv[c] + adjustment > 255)
-                running_avg_uv[c] = 255;
-              else
-                running_avg_uv[c] = running_avg_uv[c] + adjustment;
-              sum_diff += adjustment;
-            }
+    // The delta is set by the excess of absolute pixel diff over threshold.
+    int delta = ((abs(sum_diff) - sum_diff_thresh) >> 8) + 1;
+    // Only apply the adjustment for max delta up to 3.
+    if (delta < 4) {
+      sig -= sig_stride * 8;
+      mc_running_avg_uv -= mc_avg_uv_stride * 8;
+      running_avg_uv -= avg_uv_stride * 8;
+      for (r = 0; r < 8; ++r) {
+        for (c = 0; c < 8; ++c) {
+          int diff = mc_running_avg_uv[c] - sig[c];
+          int adjustment = abs(diff);
+          if (adjustment > delta) adjustment = delta;
+          if (diff > 0) {
+            // Bring denoised signal down.
+            if (running_avg_uv[c] - adjustment < 0)
+              running_avg_uv[c] = 0;
+            else
+              running_avg_uv[c] = running_avg_uv[c] - adjustment;
+            sum_diff -= adjustment;
+          } else if (diff < 0) {
+            // Bring denoised signal up.
+            if (running_avg_uv[c] + adjustment > 255)
+              running_avg_uv[c] = 255;
+            else
+              running_avg_uv[c] = running_avg_uv[c] + adjustment;
+            sum_diff += adjustment;
           }
-          // TODO(marpan): Check here if abs(sum_diff) has gone below the
-          // threshold sum_diff_thresh, and if so, we can exit the row loop.
-          sig += sig_stride;
-          mc_running_avg_uv += mc_avg_uv_stride;
-          running_avg_uv += avg_uv_stride;
         }
-        if (abs(sum_diff) > sum_diff_thresh)
-          return COPY_BLOCK;
-      } else {
-        return COPY_BLOCK;
+        // TODO(marpan): Check here if abs(sum_diff) has gone below the
+        // threshold sum_diff_thresh, and if so, we can exit the row loop.
+        sig += sig_stride;
+        mc_running_avg_uv += mc_avg_uv_stride;
+        running_avg_uv += avg_uv_stride;
       }
+      if (abs(sum_diff) > sum_diff_thresh) return COPY_BLOCK;
+    } else {
+      return COPY_BLOCK;
     }
+  }
 
-    vp8_copy_mem8x8(running_avg_uv_start, avg_uv_stride, sig_start,
-                    sig_stride);
-    return FILTER_BLOCK;
+  vp8_copy_mem8x8(running_avg_uv_start, avg_uv_stride, sig_start, sig_stride);
+  return FILTER_BLOCK;
 }
 
 void vp8_denoiser_set_parameters(VP8_DENOISER *denoiser, int mode) {
@@ -398,364 +378,334 @@
 }
 
 int vp8_denoiser_allocate(VP8_DENOISER *denoiser, int width, int height,
-                          int num_mb_rows, int num_mb_cols, int mode)
-{
-    int i;
-    assert(denoiser);
-    denoiser->num_mb_cols = num_mb_cols;
+                          int num_mb_rows, int num_mb_cols, int mode) {
+  int i;
+  assert(denoiser);
+  denoiser->num_mb_cols = num_mb_cols;
 
-    for (i = 0; i < MAX_REF_FRAMES; i++)
-    {
-        denoiser->yv12_running_avg[i].flags = 0;
+  for (i = 0; i < MAX_REF_FRAMES; i++) {
+    denoiser->yv12_running_avg[i].flags = 0;
 
-        if (vp8_yv12_alloc_frame_buffer(&(denoiser->yv12_running_avg[i]), width,
-                                        height, VP8BORDERINPIXELS)
-            < 0)
-        {
-            vp8_denoiser_free(denoiser);
-            return 1;
-        }
-        memset(denoiser->yv12_running_avg[i].buffer_alloc, 0,
-               denoiser->yv12_running_avg[i].frame_size);
-
-    }
-    denoiser->yv12_mc_running_avg.flags = 0;
-
-    if (vp8_yv12_alloc_frame_buffer(&(denoiser->yv12_mc_running_avg), width,
-                                   height, VP8BORDERINPIXELS) < 0)
-    {
-        vp8_denoiser_free(denoiser);
-        return 1;
-    }
-
-    memset(denoiser->yv12_mc_running_avg.buffer_alloc, 0,
-           denoiser->yv12_mc_running_avg.frame_size);
-
-    if (vp8_yv12_alloc_frame_buffer(&denoiser->yv12_last_source, width,
+    if (vp8_yv12_alloc_frame_buffer(&(denoiser->yv12_running_avg[i]), width,
                                     height, VP8BORDERINPIXELS) < 0) {
       vp8_denoiser_free(denoiser);
       return 1;
     }
-    memset(denoiser->yv12_last_source.buffer_alloc, 0,
-           denoiser->yv12_last_source.frame_size);
+    memset(denoiser->yv12_running_avg[i].buffer_alloc, 0,
+           denoiser->yv12_running_avg[i].frame_size);
+  }
+  denoiser->yv12_mc_running_avg.flags = 0;
 
-    denoiser->denoise_state = vpx_calloc((num_mb_rows * num_mb_cols), 1);
-    if (!denoiser->denoise_state)
-    {
-        vp8_denoiser_free(denoiser);
-        return 1;
-    }
-    memset(denoiser->denoise_state, 0, (num_mb_rows * num_mb_cols));
-    vp8_denoiser_set_parameters(denoiser, mode);
-    denoiser->nmse_source_diff = 0;
-    denoiser->nmse_source_diff_count = 0;
-    denoiser->qp_avg = 0;
-    // QP threshold below which we can go up to aggressive mode.
-    denoiser->qp_threshold_up = 80;
-    // QP threshold above which we can go back down to normal mode.
-    // For now keep this second threshold high, so not used currently.
-    denoiser->qp_threshold_down = 128;
-    // Bitrate thresholds and noise metric (nmse) thresholds for switching to
-    // aggressive mode.
-    // TODO(marpan): Adjust thresholds, including effect on resolution.
-    denoiser->bitrate_threshold = 400000;  // (bits/sec).
-    denoiser->threshold_aggressive_mode = 80;
-    if (width * height > 1280 * 720) {
-      denoiser->bitrate_threshold = 3000000;
-      denoiser->threshold_aggressive_mode = 200;
-    } else if (width * height > 960 * 540) {
-      denoiser->bitrate_threshold = 1200000;
-      denoiser->threshold_aggressive_mode = 120;
-    } else if (width * height > 640 * 480) {
-      denoiser->bitrate_threshold = 600000;
-      denoiser->threshold_aggressive_mode = 100;
-    }
-    return 0;
-}
+  if (vp8_yv12_alloc_frame_buffer(&(denoiser->yv12_mc_running_avg), width,
+                                  height, VP8BORDERINPIXELS) < 0) {
+    vp8_denoiser_free(denoiser);
+    return 1;
+  }
 
+  memset(denoiser->yv12_mc_running_avg.buffer_alloc, 0,
+         denoiser->yv12_mc_running_avg.frame_size);
 
-void vp8_denoiser_free(VP8_DENOISER *denoiser)
-{
-    int i;
-    assert(denoiser);
+  if (vp8_yv12_alloc_frame_buffer(&denoiser->yv12_last_source, width, height,
+                                  VP8BORDERINPIXELS) < 0) {
+    vp8_denoiser_free(denoiser);
+    return 1;
+  }
+  memset(denoiser->yv12_last_source.buffer_alloc, 0,
+         denoiser->yv12_last_source.frame_size);
 
-    for (i = 0; i < MAX_REF_FRAMES ; i++)
-    {
-        vp8_yv12_de_alloc_frame_buffer(&denoiser->yv12_running_avg[i]);
-    }
-    vp8_yv12_de_alloc_frame_buffer(&denoiser->yv12_mc_running_avg);
-    vp8_yv12_de_alloc_frame_buffer(&denoiser->yv12_last_source);
-    vpx_free(denoiser->denoise_state);
+  denoiser->denoise_state = vpx_calloc((num_mb_rows * num_mb_cols), 1);
+  if (!denoiser->denoise_state) {
+    vp8_denoiser_free(denoiser);
+    return 1;
+  }
+  memset(denoiser->denoise_state, 0, (num_mb_rows * num_mb_cols));
+  vp8_denoiser_set_parameters(denoiser, mode);
+  denoiser->nmse_source_diff = 0;
+  denoiser->nmse_source_diff_count = 0;
+  denoiser->qp_avg = 0;
+  // QP threshold below which we can go up to aggressive mode.
+  denoiser->qp_threshold_up = 80;
+  // QP threshold above which we can go back down to normal mode.
+  // For now keep this second threshold high, so not used currently.
+  denoiser->qp_threshold_down = 128;
+  // Bitrate thresholds and noise metric (nmse) thresholds for switching to
+  // aggressive mode.
+  // TODO(marpan): Adjust thresholds, including effect on resolution.
+  denoiser->bitrate_threshold = 400000;  // (bits/sec).
+  denoiser->threshold_aggressive_mode = 80;
+  if (width * height > 1280 * 720) {
+    denoiser->bitrate_threshold = 3000000;
+    denoiser->threshold_aggressive_mode = 200;
+  } else if (width * height > 960 * 540) {
+    denoiser->bitrate_threshold = 1200000;
+    denoiser->threshold_aggressive_mode = 120;
+  } else if (width * height > 640 * 480) {
+    denoiser->bitrate_threshold = 600000;
+    denoiser->threshold_aggressive_mode = 100;
+  }
+  return 0;
 }
 
-void vp8_denoiser_denoise_mb(VP8_DENOISER *denoiser,
-                             MACROBLOCK *x,
-                             unsigned int best_sse,
-                             unsigned int zero_mv_sse,
-                             int recon_yoffset,
-                             int recon_uvoffset,
-                             loop_filter_info_n *lfi_n,
-                             int mb_row,
-                             int mb_col,
-                             int block_index,
-                             int consec_zero_last)
+void vp8_denoiser_free(VP8_DENOISER *denoiser) {
+  int i;
+  assert(denoiser);
 
-{
-    int mv_row;
-    int mv_col;
-    unsigned int motion_threshold;
-    unsigned int motion_magnitude2;
-    unsigned int sse_thresh;
-    int sse_diff_thresh = 0;
-    // Spatial loop filter: only applied selectively based on
-    // temporal filter state of block relative to top/left neighbors.
-    int apply_spatial_loop_filter = 1;
-    MV_REFERENCE_FRAME frame = x->best_reference_frame;
-    MV_REFERENCE_FRAME zero_frame = x->best_zeromv_reference_frame;
+  for (i = 0; i < MAX_REF_FRAMES; i++) {
+    vp8_yv12_de_alloc_frame_buffer(&denoiser->yv12_running_avg[i]);
+  }
+  vp8_yv12_de_alloc_frame_buffer(&denoiser->yv12_mc_running_avg);
+  vp8_yv12_de_alloc_frame_buffer(&denoiser->yv12_last_source);
+  vpx_free(denoiser->denoise_state);
+}
 
-    enum vp8_denoiser_decision decision = FILTER_BLOCK;
-    enum vp8_denoiser_decision decision_u = COPY_BLOCK;
-    enum vp8_denoiser_decision decision_v = COPY_BLOCK;
+void vp8_denoiser_denoise_mb(VP8_DENOISER *denoiser, MACROBLOCK *x,
+                             unsigned int best_sse, unsigned int zero_mv_sse,
+                             int recon_yoffset, int recon_uvoffset,
+                             loop_filter_info_n *lfi_n, int mb_row, int mb_col,
+                             int block_index, int consec_zero_last)
 
-    if (zero_frame)
-    {
-        YV12_BUFFER_CONFIG *src = &denoiser->yv12_running_avg[frame];
-        YV12_BUFFER_CONFIG *dst = &denoiser->yv12_mc_running_avg;
-        YV12_BUFFER_CONFIG saved_pre,saved_dst;
-        MB_MODE_INFO saved_mbmi;
-        MACROBLOCKD *filter_xd = &x->e_mbd;
-        MB_MODE_INFO *mbmi = &filter_xd->mode_info_context->mbmi;
-        int sse_diff = 0;
-        // Bias on zero motion vector sse.
-        const int zero_bias = denoiser->denoise_pars.denoise_mv_bias;
-        zero_mv_sse = (unsigned int)((int64_t)zero_mv_sse * zero_bias / 100);
-        sse_diff = (int)zero_mv_sse - (int)best_sse;
+{
+  int mv_row;
+  int mv_col;
+  unsigned int motion_threshold;
+  unsigned int motion_magnitude2;
+  unsigned int sse_thresh;
+  int sse_diff_thresh = 0;
+  // Spatial loop filter: only applied selectively based on
+  // temporal filter state of block relative to top/left neighbors.
+  int apply_spatial_loop_filter = 1;
+  MV_REFERENCE_FRAME frame = x->best_reference_frame;
+  MV_REFERENCE_FRAME zero_frame = x->best_zeromv_reference_frame;
 
-        saved_mbmi = *mbmi;
+  enum vp8_denoiser_decision decision = FILTER_BLOCK;
+  enum vp8_denoiser_decision decision_u = COPY_BLOCK;
+  enum vp8_denoiser_decision decision_v = COPY_BLOCK;
 
-        /* Use the best MV for the compensation. */
-        mbmi->ref_frame = x->best_reference_frame;
-        mbmi->mode = x->best_sse_inter_mode;
-        mbmi->mv = x->best_sse_mv;
-        mbmi->need_to_clamp_mvs = x->need_to_clamp_best_mvs;
-        mv_col = x->best_sse_mv.as_mv.col;
-        mv_row = x->best_sse_mv.as_mv.row;
-        // Bias to zero_mv if small amount of motion.
-        // Note sse_diff_thresh is intialized to zero, so this ensures
-        // we will always choose zero_mv for denoising if
-        // zero_mv_see <= best_sse (i.e., sse_diff <= 0).
-        if ((unsigned int)(mv_row * mv_row + mv_col * mv_col)
-            <= NOISE_MOTION_THRESHOLD)
-            sse_diff_thresh = (int)SSE_DIFF_THRESHOLD;
+  if (zero_frame) {
+    YV12_BUFFER_CONFIG *src = &denoiser->yv12_running_avg[frame];
+    YV12_BUFFER_CONFIG *dst = &denoiser->yv12_mc_running_avg;
+    YV12_BUFFER_CONFIG saved_pre, saved_dst;
+    MB_MODE_INFO saved_mbmi;
+    MACROBLOCKD *filter_xd = &x->e_mbd;
+    MB_MODE_INFO *mbmi = &filter_xd->mode_info_context->mbmi;
+    int sse_diff = 0;
+    // Bias on zero motion vector sse.
+    const int zero_bias = denoiser->denoise_pars.denoise_mv_bias;
+    zero_mv_sse = (unsigned int)((int64_t)zero_mv_sse * zero_bias / 100);
+    sse_diff = (int)zero_mv_sse - (int)best_sse;
 
-        if (frame == INTRA_FRAME ||
-            sse_diff <= sse_diff_thresh)
-        {
-            /*
-             * Handle intra blocks as referring to last frame with zero motion
-             * and let the absolute pixel difference affect the filter factor.
-             * Also consider small amount of motion as being random walk due
-             * to noise, if it doesn't mean that we get a much bigger error.
-             * Note that any changes to the mode info only affects the
-             * denoising.
-             */
-            x->denoise_zeromv = 1;
-            mbmi->ref_frame =
-                    x->best_zeromv_reference_frame;
+    saved_mbmi = *mbmi;
 
-            src = &denoiser->yv12_running_avg[zero_frame];
+    /* Use the best MV for the compensation. */
+    mbmi->ref_frame = x->best_reference_frame;
+    mbmi->mode = x->best_sse_inter_mode;
+    mbmi->mv = x->best_sse_mv;
+    mbmi->need_to_clamp_mvs = x->need_to_clamp_best_mvs;
+    mv_col = x->best_sse_mv.as_mv.col;
+    mv_row = x->best_sse_mv.as_mv.row;
+    // Bias to zero_mv if small amount of motion.
+    // Note sse_diff_thresh is intialized to zero, so this ensures
+    // we will always choose zero_mv for denoising if
+    // zero_mv_see <= best_sse (i.e., sse_diff <= 0).
+    if ((unsigned int)(mv_row * mv_row + mv_col * mv_col) <=
+        NOISE_MOTION_THRESHOLD)
+      sse_diff_thresh = (int)SSE_DIFF_THRESHOLD;
 
-            mbmi->mode = ZEROMV;
-            mbmi->mv.as_int = 0;
-            x->best_sse_inter_mode = ZEROMV;
-            x->best_sse_mv.as_int = 0;
-            best_sse = zero_mv_sse;
-        }
+    if (frame == INTRA_FRAME || sse_diff <= sse_diff_thresh) {
+      /*
+       * Handle intra blocks as referring to last frame with zero motion
+       * and let the absolute pixel difference affect the filter factor.
+       * Also consider small amount of motion as being random walk due
+       * to noise, if it doesn't mean that we get a much bigger error.
+       * Note that any changes to the mode info only affects the
+       * denoising.
+       */
+      x->denoise_zeromv = 1;
+      mbmi->ref_frame = x->best_zeromv_reference_frame;
 
-        mv_row = x->best_sse_mv.as_mv.row;
-        mv_col = x->best_sse_mv.as_mv.col;
-        motion_magnitude2 = mv_row * mv_row + mv_col * mv_col;
-        motion_threshold = denoiser->denoise_pars.scale_motion_thresh *
-            NOISE_MOTION_THRESHOLD;
+      src = &denoiser->yv12_running_avg[zero_frame];
 
-        if (motion_magnitude2 <
-          denoiser->denoise_pars.scale_increase_filter * NOISE_MOTION_THRESHOLD)
-          x->increase_denoising = 1;
+      mbmi->mode = ZEROMV;
+      mbmi->mv.as_int = 0;
+      x->best_sse_inter_mode = ZEROMV;
+      x->best_sse_mv.as_int = 0;
+      best_sse = zero_mv_sse;
+    }
 
-        sse_thresh = denoiser->denoise_pars.scale_sse_thresh * SSE_THRESHOLD;
-        if (x->increase_denoising)
-          sse_thresh =
-              denoiser->denoise_pars.scale_sse_thresh * SSE_THRESHOLD_HIGH;
+    mv_row = x->best_sse_mv.as_mv.row;
+    mv_col = x->best_sse_mv.as_mv.col;
+    motion_magnitude2 = mv_row * mv_row + mv_col * mv_col;
+    motion_threshold =
+        denoiser->denoise_pars.scale_motion_thresh * NOISE_MOTION_THRESHOLD;
 
-        if (best_sse > sse_thresh || motion_magnitude2 > motion_threshold)
-          decision = COPY_BLOCK;
+    if (motion_magnitude2 <
+        denoiser->denoise_pars.scale_increase_filter * NOISE_MOTION_THRESHOLD)
+      x->increase_denoising = 1;
 
-        // If block is considered skin, don't denoise if the block
-        // (1) is selected as non-zero motion for current frame, or
-        // (2) has not been selected as ZERO_LAST mode at least x past frames
-        // in a row.
-        // TODO(marpan): Parameter "x" should be varied with framerate.
-        // In particualar, should be reduced for layers (base layer/LAST).
-        if (x->is_skin && (consec_zero_last < 2 || motion_magnitude2 > 0))
-          decision = COPY_BLOCK;
+    sse_thresh = denoiser->denoise_pars.scale_sse_thresh * SSE_THRESHOLD;
+    if (x->increase_denoising)
+      sse_thresh = denoiser->denoise_pars.scale_sse_thresh * SSE_THRESHOLD_HIGH;
 
-        if (decision == FILTER_BLOCK) {
-          saved_pre = filter_xd->pre;
-          saved_dst = filter_xd->dst;
+    if (best_sse > sse_thresh || motion_magnitude2 > motion_threshold)
+      decision = COPY_BLOCK;
 
-          /* Compensate the running average. */
-          filter_xd->pre.y_buffer = src->y_buffer + recon_yoffset;
-          filter_xd->pre.u_buffer = src->u_buffer + recon_uvoffset;
-          filter_xd->pre.v_buffer = src->v_buffer + recon_uvoffset;
-          /* Write the compensated running average to the destination buffer. */
-          filter_xd->dst.y_buffer = dst->y_buffer + recon_yoffset;
-          filter_xd->dst.u_buffer = dst->u_buffer + recon_uvoffset;
-          filter_xd->dst.v_buffer = dst->v_buffer + recon_uvoffset;
-
-          if (!x->skip)
-          {
-              vp8_build_inter_predictors_mb(filter_xd);
-          }
-          else
-          {
-              vp8_build_inter16x16_predictors_mb(filter_xd,
-                                                 filter_xd->dst.y_buffer,
-                                                 filter_xd->dst.u_buffer,
-                                                 filter_xd->dst.v_buffer,
-                                                 filter_xd->dst.y_stride,
-                                                 filter_xd->dst.uv_stride);
-          }
-          filter_xd->pre = saved_pre;
-          filter_xd->dst = saved_dst;
-          *mbmi = saved_mbmi;
-        }
-    } else {
-      // zero_frame should always be 1 for real-time mode, as the
-      // ZEROMV mode is always checked, so we should never go into this branch.
-      // If case ZEROMV is not checked, then we will force no denoise (COPY).
+    // If block is considered skin, don't denoise if the block
+    // (1) is selected as non-zero motion for current frame, or
+    // (2) has not been selected as ZERO_LAST mode at least x past frames
+    // in a row.
+    // TODO(marpan): Parameter "x" should be varied with framerate.
+    // In particualar, should be reduced for layers (base layer/LAST).
+    if (x->is_skin && (consec_zero_last < 2 || motion_magnitude2 > 0))
       decision = COPY_BLOCK;
+
+    if (decision == FILTER_BLOCK) {
+      saved_pre = filter_xd->pre;
+      saved_dst = filter_xd->dst;
+
+      /* Compensate the running average. */
+      filter_xd->pre.y_buffer = src->y_buffer + recon_yoffset;
+      filter_xd->pre.u_buffer = src->u_buffer + recon_uvoffset;
+      filter_xd->pre.v_buffer = src->v_buffer + recon_uvoffset;
+      /* Write the compensated running average to the destination buffer. */
+      filter_xd->dst.y_buffer = dst->y_buffer + recon_yoffset;
+      filter_xd->dst.u_buffer = dst->u_buffer + recon_uvoffset;
+      filter_xd->dst.v_buffer = dst->v_buffer + recon_uvoffset;
+
+      if (!x->skip) {
+        vp8_build_inter_predictors_mb(filter_xd);
+      } else {
+        vp8_build_inter16x16_predictors_mb(
+            filter_xd, filter_xd->dst.y_buffer, filter_xd->dst.u_buffer,
+            filter_xd->dst.v_buffer, filter_xd->dst.y_stride,
+            filter_xd->dst.uv_stride);
+      }
+      filter_xd->pre = saved_pre;
+      filter_xd->dst = saved_dst;
+      *mbmi = saved_mbmi;
     }
+  } else {
+    // zero_frame should always be 1 for real-time mode, as the
+    // ZEROMV mode is always checked, so we should never go into this branch.
+    // If case ZEROMV is not checked, then we will force no denoise (COPY).
+    decision = COPY_BLOCK;
+  }
 
-    if (decision == FILTER_BLOCK)
-    {
-        unsigned char *mc_running_avg_y =
-            denoiser->yv12_mc_running_avg.y_buffer + recon_yoffset;
-        int mc_avg_y_stride = denoiser->yv12_mc_running_avg.y_stride;
-        unsigned char *running_avg_y =
-            denoiser->yv12_running_avg[INTRA_FRAME].y_buffer + recon_yoffset;
-        int avg_y_stride = denoiser->yv12_running_avg[INTRA_FRAME].y_stride;
+  if (decision == FILTER_BLOCK) {
+    unsigned char *mc_running_avg_y =
+        denoiser->yv12_mc_running_avg.y_buffer + recon_yoffset;
+    int mc_avg_y_stride = denoiser->yv12_mc_running_avg.y_stride;
+    unsigned char *running_avg_y =
+        denoiser->yv12_running_avg[INTRA_FRAME].y_buffer + recon_yoffset;
+    int avg_y_stride = denoiser->yv12_running_avg[INTRA_FRAME].y_stride;
 
-        /* Filter. */
-        decision = vp8_denoiser_filter(mc_running_avg_y, mc_avg_y_stride,
-                                       running_avg_y, avg_y_stride,
-                                       x->thismb, 16, motion_magnitude2,
-                                       x->increase_denoising);
-        denoiser->denoise_state[block_index] = motion_magnitude2 > 0 ?
-            kFilterNonZeroMV : kFilterZeroMV;
-        // Only denoise UV for zero motion, and if y channel was denoised.
-        if (denoiser->denoiser_mode != kDenoiserOnYOnly &&
-            motion_magnitude2 == 0 &&
-            decision == FILTER_BLOCK) {
-          unsigned char *mc_running_avg_u =
-              denoiser->yv12_mc_running_avg.u_buffer + recon_uvoffset;
-          unsigned char *running_avg_u =
-              denoiser->yv12_running_avg[INTRA_FRAME].u_buffer + recon_uvoffset;
-          unsigned char *mc_running_avg_v =
-              denoiser->yv12_mc_running_avg.v_buffer + recon_uvoffset;
-          unsigned char *running_avg_v =
-              denoiser->yv12_running_avg[INTRA_FRAME].v_buffer + recon_uvoffset;
-          int mc_avg_uv_stride = denoiser->yv12_mc_running_avg.uv_stride;
-          int avg_uv_stride = denoiser->yv12_running_avg[INTRA_FRAME].uv_stride;
-          int signal_stride = x->block[16].src_stride;
-          decision_u =
-              vp8_denoiser_filter_uv(mc_running_avg_u, mc_avg_uv_stride,
-                                      running_avg_u, avg_uv_stride,
-                                      x->block[16].src + *x->block[16].base_src,
-                                      signal_stride, motion_magnitude2, 0);
-          decision_v =
-              vp8_denoiser_filter_uv(mc_running_avg_v, mc_avg_uv_stride,
-                                      running_avg_v, avg_uv_stride,
-                                      x->block[20].src + *x->block[20].base_src,
-                                      signal_stride, motion_magnitude2, 0);
-        }
+    /* Filter. */
+    decision = vp8_denoiser_filter(mc_running_avg_y, mc_avg_y_stride,
+                                   running_avg_y, avg_y_stride, x->thismb, 16,
+                                   motion_magnitude2, x->increase_denoising);
+    denoiser->denoise_state[block_index] =
+        motion_magnitude2 > 0 ? kFilterNonZeroMV : kFilterZeroMV;
+    // Only denoise UV for zero motion, and if y channel was denoised.
+    if (denoiser->denoiser_mode != kDenoiserOnYOnly && motion_magnitude2 == 0 &&
+        decision == FILTER_BLOCK) {
+      unsigned char *mc_running_avg_u =
+          denoiser->yv12_mc_running_avg.u_buffer + recon_uvoffset;
+      unsigned char *running_avg_u =
+          denoiser->yv12_running_avg[INTRA_FRAME].u_buffer + recon_uvoffset;
+      unsigned char *mc_running_avg_v =
+          denoiser->yv12_mc_running_avg.v_buffer + recon_uvoffset;
+      unsigned char *running_avg_v =
+          denoiser->yv12_running_avg[INTRA_FRAME].v_buffer + recon_uvoffset;
+      int mc_avg_uv_stride = denoiser->yv12_mc_running_avg.uv_stride;
+      int avg_uv_stride = denoiser->yv12_running_avg[INTRA_FRAME].uv_stride;
+      int signal_stride = x->block[16].src_stride;
+      decision_u = vp8_denoiser_filter_uv(
+          mc_running_avg_u, mc_avg_uv_stride, running_avg_u, avg_uv_stride,
+          x->block[16].src + *x->block[16].base_src, signal_stride,
+          motion_magnitude2, 0);
+      decision_v = vp8_denoiser_filter_uv(
+          mc_running_avg_v, mc_avg_uv_stride, running_avg_v, avg_uv_stride,
+          x->block[20].src + *x->block[20].base_src, signal_stride,
+          motion_magnitude2, 0);
     }
-    if (decision == COPY_BLOCK)
-    {
-        /* No filtering of this block; it differs too much from the predictor,
-         * or the motion vector magnitude is considered too big.
-         */
-        x->denoise_zeromv = 0;
-        vp8_copy_mem16x16(
-                x->thismb, 16,
-                denoiser->yv12_running_avg[INTRA_FRAME].y_buffer + recon_yoffset,
-                denoiser->yv12_running_avg[INTRA_FRAME].y_stride);
-        denoiser->denoise_state[block_index] = kNoFilter;
+  }
+  if (decision == COPY_BLOCK) {
+    /* No filtering of this block; it differs too much from the predictor,
+     * or the motion vector magnitude is considered too big.
+     */
+    x->denoise_zeromv = 0;
+    vp8_copy_mem16x16(
+        x->thismb, 16,
+        denoiser->yv12_running_avg[INTRA_FRAME].y_buffer + recon_yoffset,
+        denoiser->yv12_running_avg[INTRA_FRAME].y_stride);
+    denoiser->denoise_state[block_index] = kNoFilter;
+  }
+  if (denoiser->denoiser_mode != kDenoiserOnYOnly) {
+    if (decision_u == COPY_BLOCK) {
+      vp8_copy_mem8x8(
+          x->block[16].src + *x->block[16].base_src, x->block[16].src_stride,
+          denoiser->yv12_running_avg[INTRA_FRAME].u_buffer + recon_uvoffset,
+          denoiser->yv12_running_avg[INTRA_FRAME].uv_stride);
     }
-    if (denoiser->denoiser_mode != kDenoiserOnYOnly) {
-      if (decision_u == COPY_BLOCK) {
-        vp8_copy_mem8x8(
-            x->block[16].src + *x->block[16].base_src, x->block[16].src_stride,
-            denoiser->yv12_running_avg[INTRA_FRAME].u_buffer + recon_uvoffset,
-            denoiser->yv12_running_avg[INTRA_FRAME].uv_stride);
-      }
-      if (decision_v == COPY_BLOCK) {
-        vp8_copy_mem8x8(
-            x->block[20].src + *x->block[20].base_src, x->block[16].src_stride,
-            denoiser->yv12_running_avg[INTRA_FRAME].v_buffer + recon_uvoffset,
-            denoiser->yv12_running_avg[INTRA_FRAME].uv_stride);
-      }
+    if (decision_v == COPY_BLOCK) {
+      vp8_copy_mem8x8(
+          x->block[20].src + *x->block[20].base_src, x->block[16].src_stride,
+          denoiser->yv12_running_avg[INTRA_FRAME].v_buffer + recon_uvoffset,
+          denoiser->yv12_running_avg[INTRA_FRAME].uv_stride);
     }
-    // Option to selectively deblock the denoised signal, for y channel only.
-    if (apply_spatial_loop_filter) {
-      loop_filter_info lfi;
-      int apply_filter_col = 0;
-      int apply_filter_row = 0;
-      int apply_filter = 0;
-      int y_stride = denoiser->yv12_running_avg[INTRA_FRAME].y_stride;
-      int uv_stride =denoiser->yv12_running_avg[INTRA_FRAME].uv_stride;
+  }
+  // Option to selectively deblock the denoised signal, for y channel only.
+  if (apply_spatial_loop_filter) {
+    loop_filter_info lfi;
+    int apply_filter_col = 0;
+    int apply_filter_row = 0;
+    int apply_filter = 0;
+    int y_stride = denoiser->yv12_running_avg[INTRA_FRAME].y_stride;
+    int uv_stride = denoiser->yv12_running_avg[INTRA_FRAME].uv_stride;
 
-      // Fix filter level to some nominal value for now.
-      int filter_level = 48;
+    // Fix filter level to some nominal value for now.
+    int filter_level = 48;
 
-      int hev_index = lfi_n->hev_thr_lut[INTER_FRAME][filter_level];
-      lfi.mblim = lfi_n->mblim[filter_level];
-      lfi.blim = lfi_n->blim[filter_level];
-      lfi.lim = lfi_n->lim[filter_level];
-      lfi.hev_thr = lfi_n->hev_thr[hev_index];
+    int hev_index = lfi_n->hev_thr_lut[INTER_FRAME][filter_level];
+    lfi.mblim = lfi_n->mblim[filter_level];
+    lfi.blim = lfi_n->blim[filter_level];
+    lfi.lim = lfi_n->lim[filter_level];
+    lfi.hev_thr = lfi_n->hev_thr[hev_index];
 
-      // Apply filter if there is a difference in the denoiser filter state
-      // between the current and left/top block, or if non-zero motion vector
-      // is used for the motion-compensated filtering.
-      if (mb_col > 0) {
-        apply_filter_col = !((denoiser->denoise_state[block_index] ==
-            denoiser->denoise_state[block_index - 1]) &&
+    // Apply filter if there is a difference in the denoiser filter state
+    // between the current and left/top block, or if non-zero motion vector
+    // is used for the motion-compensated filtering.
+    if (mb_col > 0) {
+      apply_filter_col =
+          !((denoiser->denoise_state[block_index] ==
+             denoiser->denoise_state[block_index - 1]) &&
             denoiser->denoise_state[block_index] != kFilterNonZeroMV);
-        if (apply_filter_col) {
-          // Filter left vertical edge.
-          apply_filter = 1;
-          vp8_loop_filter_mbv(
-              denoiser->yv12_running_avg[INTRA_FRAME].y_buffer + recon_yoffset,
-              NULL, NULL, y_stride, uv_stride, &lfi);
-        }
+      if (apply_filter_col) {
+        // Filter left vertical edge.
+        apply_filter = 1;
+        vp8_loop_filter_mbv(
+            denoiser->yv12_running_avg[INTRA_FRAME].y_buffer + recon_yoffset,
+            NULL, NULL, y_stride, uv_stride, &lfi);
       }
-      if (mb_row > 0) {
-        apply_filter_row = !((denoiser->denoise_state[block_index] ==
-            denoiser->denoise_state[block_index - denoiser->num_mb_cols]) &&
+    }
+    if (mb_row > 0) {
+      apply_filter_row =
+          !((denoiser->denoise_state[block_index] ==
+             denoiser->denoise_state[block_index - denoiser->num_mb_cols]) &&
             denoiser->denoise_state[block_index] != kFilterNonZeroMV);
-        if (apply_filter_row) {
-          // Filter top horizontal edge.
-          apply_filter = 1;
-          vp8_loop_filter_mbh(
-              denoiser->yv12_running_avg[INTRA_FRAME].y_buffer + recon_yoffset,
-              NULL, NULL, y_stride, uv_stride, &lfi);
-        }
-      }
-      if (apply_filter) {
-        // Update the signal block |x|. Pixel changes are only to top and/or
-        // left boundary pixels: can we avoid full block copy here.
-        vp8_copy_mem16x16(
+      if (apply_filter_row) {
+        // Filter top horizontal edge.
+        apply_filter = 1;
+        vp8_loop_filter_mbh(
             denoiser->yv12_running_avg[INTRA_FRAME].y_buffer + recon_yoffset,
-            y_stride, x->thismb, 16);
+            NULL, NULL, y_stride, uv_stride, &lfi);
       }
     }
+    if (apply_filter) {
+      // Update the signal block |x|. Pixel changes are only to top and/or
+      // left boundary pixels: can we avoid full block copy here.
+      vp8_copy_mem16x16(
+          denoiser->yv12_running_avg[INTRA_FRAME].y_buffer + recon_yoffset,
+          y_stride, x->thismb, 16);
+    }
+  }
 }
--- a/vp8/encoder/denoising.h
+++ b/vp8/encoder/denoising.h
@@ -20,26 +20,18 @@
 
 #define SUM_DIFF_THRESHOLD 512
 #define SUM_DIFF_THRESHOLD_HIGH 600
-#define MOTION_MAGNITUDE_THRESHOLD (8*3)
+#define MOTION_MAGNITUDE_THRESHOLD (8 * 3)
 
-#define SUM_DIFF_THRESHOLD_UV (96)   // (8 * 8 * 1.5)
+#define SUM_DIFF_THRESHOLD_UV (96)  // (8 * 8 * 1.5)
 #define SUM_DIFF_THRESHOLD_HIGH_UV (8 * 8 * 2)
 #define SUM_DIFF_FROM_AVG_THRESH_UV (8 * 8 * 8)
-#define MOTION_MAGNITUDE_THRESHOLD_UV (8*3)
+#define MOTION_MAGNITUDE_THRESHOLD_UV (8 * 3)
 
 #define MAX_GF_ARF_DENOISE_RANGE (8)
 
-enum vp8_denoiser_decision
-{
-  COPY_BLOCK,
-  FILTER_BLOCK
-};
+enum vp8_denoiser_decision { COPY_BLOCK, FILTER_BLOCK };
 
-enum vp8_denoiser_filter_state {
-  kNoFilter,
-  kFilterZeroMV,
-  kFilterNonZeroMV
-};
+enum vp8_denoiser_filter_state { kNoFilter, kFilterZeroMV, kFilterNonZeroMV };
 
 enum vp8_denoiser_mode {
   kDenoiserOff,
@@ -73,23 +65,22 @@
   unsigned int spatial_blur;
 } denoise_params;
 
-typedef struct vp8_denoiser
-{
-    YV12_BUFFER_CONFIG yv12_running_avg[MAX_REF_FRAMES];
-    YV12_BUFFER_CONFIG yv12_mc_running_avg;
-    // TODO(marpan): Should remove yv12_last_source and use vp8_lookahead_peak.
-    YV12_BUFFER_CONFIG yv12_last_source;
-    unsigned char* denoise_state;
-    int num_mb_cols;
-    int denoiser_mode;
-    int threshold_aggressive_mode;
-    int nmse_source_diff;
-    int nmse_source_diff_count;
-    int qp_avg;
-    int qp_threshold_up;
-    int qp_threshold_down;
-    int bitrate_threshold;
-    denoise_params denoise_pars;
+typedef struct vp8_denoiser {
+  YV12_BUFFER_CONFIG yv12_running_avg[MAX_REF_FRAMES];
+  YV12_BUFFER_CONFIG yv12_mc_running_avg;
+  // TODO(marpan): Should remove yv12_last_source and use vp8_lookahead_peak.
+  YV12_BUFFER_CONFIG yv12_last_source;
+  unsigned char *denoise_state;
+  int num_mb_cols;
+  int denoiser_mode;
+  int threshold_aggressive_mode;
+  int nmse_source_diff;
+  int nmse_source_diff_count;
+  int qp_avg;
+  int qp_threshold_up;
+  int qp_threshold_down;
+  int bitrate_threshold;
+  denoise_params denoise_pars;
 } VP8_DENOISER;
 
 int vp8_denoiser_allocate(VP8_DENOISER *denoiser, int width, int height,
@@ -99,17 +90,11 @@
 
 void vp8_denoiser_set_parameters(VP8_DENOISER *denoiser, int mode);
 
-void vp8_denoiser_denoise_mb(VP8_DENOISER *denoiser,
-                             MACROBLOCK *x,
-                             unsigned int best_sse,
-                             unsigned int zero_mv_sse,
-                             int recon_yoffset,
-                             int recon_uvoffset,
-                             loop_filter_info_n *lfi_n,
-                             int mb_row,
-                             int mb_col,
-                             int block_index,
-                             int consec_zero_last);
+void vp8_denoiser_denoise_mb(VP8_DENOISER *denoiser, MACROBLOCK *x,
+                             unsigned int best_sse, unsigned int zero_mv_sse,
+                             int recon_yoffset, int recon_uvoffset,
+                             loop_filter_info_n *lfi_n, int mb_row, int mb_col,
+                             int block_index, int consec_zero_last);
 
 #ifdef __cplusplus
 }  // extern "C"
--- a/vp8/encoder/encodeframe.c
+++ b/vp8/encoder/encodeframe.c
@@ -8,7 +8,6 @@
  *  be found in the AUTHORS file in the root of the source tree.
  */
 
-
 #include "vpx_config.h"
 #include "vp8_rtcd.h"
 #include "./vpx_dsp_rtcd.h"
@@ -35,31 +34,27 @@
 #endif
 #include "encodeframe.h"
 
-extern void vp8_stuff_mb(VP8_COMP *cpi, MACROBLOCK *x, TOKENEXTRA **t) ;
-extern void vp8_calc_ref_frame_costs(int *ref_frame_cost,
-                                     int prob_intra,
-                                     int prob_last,
-                                     int prob_garf
-                                    );
+extern void vp8_stuff_mb(VP8_COMP *cpi, MACROBLOCK *x, TOKENEXTRA **t);
+extern void vp8_calc_ref_frame_costs(int *ref_frame_cost, int prob_intra,
+                                     int prob_last, int prob_garf);
 extern void vp8_convert_rfct_to_prob(VP8_COMP *const cpi);
 extern void vp8cx_initialize_me_consts(VP8_COMP *cpi, int QIndex);
 extern void vp8_auto_select_speed(VP8_COMP *cpi);
-extern void vp8cx_init_mbrthread_data(VP8_COMP *cpi,
-                                      MACROBLOCK *x,
-                                      MB_ROW_COMP *mbr_ei,
-                                      int count);
-static void adjust_act_zbin( VP8_COMP *cpi, MACROBLOCK *x );
+extern void vp8cx_init_mbrthread_data(VP8_COMP *cpi, MACROBLOCK *x,
+                                      MB_ROW_COMP *mbr_ei, int count);
+static void adjust_act_zbin(VP8_COMP *cpi, MACROBLOCK *x);
 
 #ifdef MODE_STATS
-unsigned int inter_y_modes[10] = {0, 0, 0, 0, 0, 0, 0, 0, 0, 0};
-unsigned int inter_uv_modes[4] = {0, 0, 0, 0};
-unsigned int inter_b_modes[15]  = {0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0};
-unsigned int y_modes[5]   = {0, 0, 0, 0, 0};
-unsigned int uv_modes[4]  = {0, 0, 0, 0};
-unsigned int b_modes[14]  = {0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0};
+unsigned int inter_y_modes[10] = { 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 };
+unsigned int inter_uv_modes[4] = { 0, 0, 0, 0 };
+unsigned int inter_b_modes[15] = {
+  0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0
+};
+unsigned int y_modes[5] = { 0, 0, 0, 0, 0 };
+unsigned int uv_modes[4] = { 0, 0, 0, 0 };
+unsigned int b_modes[14] = { 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 };
 #endif
 
-
 /* activity_avg must be positive, or flat regions could get a zero weight
  *  (infinite lambda), which confounds analysis.
  * This also avoids the need for divide by zero checks in
@@ -72,189 +67,166 @@
  * Eventually this should be replaced by custom no-reference routines,
  *  which will be faster.
  */
-static const unsigned char VP8_VAR_OFFS[16]=
-{
-    128,128,128,128,128,128,128,128,128,128,128,128,128,128,128,128
+static const unsigned char VP8_VAR_OFFS[16] = {
+  128, 128, 128, 128, 128, 128, 128, 128, 128, 128, 128, 128, 128, 128, 128, 128
 };
 
-
 /* Original activity measure from Tim T's code. */
-static unsigned int tt_activity_measure( VP8_COMP *cpi, MACROBLOCK *x )
-{
-    unsigned int act;
-    unsigned int sse;
-    (void)cpi;
-    /* TODO: This could also be done over smaller areas (8x8), but that would
-     *  require extensive changes elsewhere, as lambda is assumed to be fixed
-     *  over an entire MB in most of the code.
-     * Another option is to compute four 8x8 variances, and pick a single
-     *  lambda using a non-linear combination (e.g., the smallest, or second
-     *  smallest, etc.).
-     */
-    act =  vpx_variance16x16(x->src.y_buffer,
-                    x->src.y_stride, VP8_VAR_OFFS, 0, &sse);
-    act = act<<4;
+static unsigned int tt_activity_measure(VP8_COMP *cpi, MACROBLOCK *x) {
+  unsigned int act;
+  unsigned int sse;
+  (void)cpi;
+  /* TODO: This could also be done over smaller areas (8x8), but that would
+   *  require extensive changes elsewhere, as lambda is assumed to be fixed
+   *  over an entire MB in most of the code.
+   * Another option is to compute four 8x8 variances, and pick a single
+   *  lambda using a non-linear combination (e.g., the smallest, or second
+   *  smallest, etc.).
+   */
+  act = vpx_variance16x16(x->src.y_buffer, x->src.y_stride, VP8_VAR_OFFS, 0,
+                          &sse);
+  act = act << 4;
 
-    /* If the region is flat, lower the activity some more. */
-    if (act < 8<<12)
-        act = act < 5<<12 ? act : 5<<12;
+  /* If the region is flat, lower the activity some more. */
+  if (act < 8 << 12) act = act < 5 << 12 ? act : 5 << 12;
 
-    return act;
+  return act;
 }
 
 /* Stub for alternative experimental activity measures. */
-static unsigned int alt_activity_measure( VP8_COMP *cpi,
-                                          MACROBLOCK *x, int use_dc_pred )
-{
-    return vp8_encode_intra(cpi,x, use_dc_pred);
+static unsigned int alt_activity_measure(VP8_COMP *cpi, MACROBLOCK *x,
+                                         int use_dc_pred) {
+  return vp8_encode_intra(cpi, x, use_dc_pred);
 }
 
-
 /* Measure the activity of the current macroblock
  * What we measure here is TBD so abstracted to this function
  */
 #define ALT_ACT_MEASURE 1
-static unsigned int mb_activity_measure( VP8_COMP *cpi, MACROBLOCK *x,
-                                  int mb_row, int mb_col)
-{
-    unsigned int mb_activity;
+static unsigned int mb_activity_measure(VP8_COMP *cpi, MACROBLOCK *x,
+                                        int mb_row, int mb_col) {
+  unsigned int mb_activity;
 
-    if  ( ALT_ACT_MEASURE )
-    {
-        int use_dc_pred = (mb_col || mb_row) && (!mb_col || !mb_row);
+  if (ALT_ACT_MEASURE) {
+    int use_dc_pred = (mb_col || mb_row) && (!mb_col || !mb_row);
 
-        /* Or use and alternative. */
-        mb_activity = alt_activity_measure( cpi, x, use_dc_pred );
-    }
-    else
-    {
-        /* Original activity measure from Tim T's code. */
-        mb_activity = tt_activity_measure( cpi, x );
-    }
+    /* Or use and alternative. */
+    mb_activity = alt_activity_measure(cpi, x, use_dc_pred);
+  } else {
+    /* Original activity measure from Tim T's code. */
+    mb_activity = tt_activity_measure(cpi, x);
+  }
 
-    if ( mb_activity < VP8_ACTIVITY_AVG_MIN )
-        mb_activity = VP8_ACTIVITY_AVG_MIN;
+  if (mb_activity < VP8_ACTIVITY_AVG_MIN) mb_activity = VP8_ACTIVITY_AVG_MIN;
 
-    return mb_activity;
+  return mb_activity;
 }
 
 /* Calculate an "average" mb activity value for the frame */
 #define ACT_MEDIAN 0
-static void calc_av_activity( VP8_COMP *cpi, int64_t activity_sum )
-{
+static void calc_av_activity(VP8_COMP *cpi, int64_t activity_sum) {
 #if ACT_MEDIAN
-    /* Find median: Simple n^2 algorithm for experimentation */
-    {
-        unsigned int median;
-        unsigned int i,j;
-        unsigned int * sortlist;
-        unsigned int tmp;
+  /* Find median: Simple n^2 algorithm for experimentation */
+  {
+    unsigned int median;
+    unsigned int i, j;
+    unsigned int *sortlist;
+    unsigned int tmp;
 
-        /* Create a list to sort to */
-        CHECK_MEM_ERROR(sortlist,
-                        vpx_calloc(sizeof(unsigned int),
-                        cpi->common.MBs));
+    /* Create a list to sort to */
+    CHECK_MEM_ERROR(sortlist,
+                    vpx_calloc(sizeof(unsigned int), cpi->common.MBs));
 
-        /* Copy map to sort list */
-        memcpy( sortlist, cpi->mb_activity_map,
-                sizeof(unsigned int) * cpi->common.MBs );
+    /* Copy map to sort list */
+    memcpy(sortlist, cpi->mb_activity_map,
+           sizeof(unsigned int) * cpi->common.MBs);
 
+    /* Ripple each value down to its correct position */
+    for (i = 1; i < cpi->common.MBs; i++) {
+      for (j = i; j > 0; j--) {
+        if (sortlist[j] < sortlist[j - 1]) {
+          /* Swap values */
+          tmp = sortlist[j - 1];
+          sortlist[j - 1] = sortlist[j];
+          sortlist[j] = tmp;
+        } else
+          break;
+      }
+    }
 
-        /* Ripple each value down to its correct position */
-        for ( i = 1; i < cpi->common.MBs; i ++ )
-        {
-            for ( j = i; j > 0; j -- )
-            {
-                if ( sortlist[j] < sortlist[j-1] )
-                {
-                    /* Swap values */
-                    tmp = sortlist[j-1];
-                    sortlist[j-1] = sortlist[j];
-                    sortlist[j] = tmp;
-                }
-                else
-                    break;
-            }
-        }
+    /* Even number MBs so estimate median as mean of two either side. */
+    median = (1 + sortlist[cpi->common.MBs >> 1] +
+              sortlist[(cpi->common.MBs >> 1) + 1]) >>
+             1;
 
-        /* Even number MBs so estimate median as mean of two either side. */
-        median = ( 1 + sortlist[cpi->common.MBs >> 1] +
-                   sortlist[(cpi->common.MBs >> 1) + 1] ) >> 1;
+    cpi->activity_avg = median;
 
-        cpi->activity_avg = median;
-
-        vpx_free(sortlist);
-    }
+    vpx_free(sortlist);
+  }
 #else
-    /* Simple mean for now */
-    cpi->activity_avg = (unsigned int)(activity_sum/cpi->common.MBs);
+  /* Simple mean for now */
+  cpi->activity_avg = (unsigned int)(activity_sum / cpi->common.MBs);
 #endif
 
-    if (cpi->activity_avg < VP8_ACTIVITY_AVG_MIN)
-        cpi->activity_avg = VP8_ACTIVITY_AVG_MIN;
+  if (cpi->activity_avg < VP8_ACTIVITY_AVG_MIN)
+    cpi->activity_avg = VP8_ACTIVITY_AVG_MIN;
 
-    /* Experimental code: return fixed value normalized for several clips */
-    if  ( ALT_ACT_MEASURE )
-        cpi->activity_avg = 100000;
+  /* Experimental code: return fixed value normalized for several clips */
+  if (ALT_ACT_MEASURE) cpi->activity_avg = 100000;
 }
 
-#define USE_ACT_INDEX   0
-#define OUTPUT_NORM_ACT_STATS   0
+#define USE_ACT_INDEX 0
+#define OUTPUT_NORM_ACT_STATS 0
 
 #if USE_ACT_INDEX
 /* Calculate and activity index for each mb */
-static void calc_activity_index( VP8_COMP *cpi, MACROBLOCK *x )
-{
-    VP8_COMMON *const cm = & cpi->common;
-    int mb_row, mb_col;
+static void calc_activity_index(VP8_COMP *cpi, MACROBLOCK *x) {
+  VP8_COMMON *const cm = &cpi->common;
+  int mb_row, mb_col;
 
-    int64_t act;
-    int64_t a;
-    int64_t b;
+  int64_t act;
+  int64_t a;
+  int64_t b;
 
 #if OUTPUT_NORM_ACT_STATS
-    FILE *f = fopen("norm_act.stt", "a");
-    fprintf(f, "\n%12d\n", cpi->activity_avg );
+  FILE *f = fopen("norm_act.stt", "a");
+  fprintf(f, "\n%12d\n", cpi->activity_avg);
 #endif
 
-    /* Reset pointers to start of activity map */
-    x->mb_activity_ptr = cpi->mb_activity_map;
+  /* Reset pointers to start of activity map */
+  x->mb_activity_ptr = cpi->mb_activity_map;
 
-    /* Calculate normalized mb activity number. */
-    for (mb_row = 0; mb_row < cm->mb_rows; mb_row++)
-    {
-        /* for each macroblock col in image */
-        for (mb_col = 0; mb_col < cm->mb_cols; mb_col++)
-        {
-            /* Read activity from the map */
-            act = *(x->mb_activity_ptr);
+  /* Calculate normalized mb activity number. */
+  for (mb_row = 0; mb_row < cm->mb_rows; mb_row++) {
+    /* for each macroblock col in image */
+    for (mb_col = 0; mb_col < cm->mb_cols; mb_col++) {
+      /* Read activity from the map */
+      act = *(x->mb_activity_ptr);
 
-            /* Calculate a normalized activity number */
-            a = act + 4*cpi->activity_avg;
-            b = 4*act + cpi->activity_avg;
+      /* Calculate a normalized activity number */
+      a = act + 4 * cpi->activity_avg;
+      b = 4 * act + cpi->activity_avg;
 
-            if ( b >= a )
-                *(x->activity_ptr) = (int)((b + (a>>1))/a) - 1;
-            else
-                *(x->activity_ptr) = 1 - (int)((a + (b>>1))/b);
+      if (b >= a)
+        *(x->activity_ptr) = (int)((b + (a >> 1)) / a) - 1;
+      else
+        *(x->activity_ptr) = 1 - (int)((a + (b >> 1)) / b);
 
 #if OUTPUT_NORM_ACT_STATS
-            fprintf(f, " %6d", *(x->mb_activity_ptr));
+      fprintf(f, " %6d", *(x->mb_activity_ptr));
 #endif
-            /* Increment activity map pointers */
-            x->mb_activity_ptr++;
-        }
+      /* Increment activity map pointers */
+      x->mb_activity_ptr++;
+    }
 
 #if OUTPUT_NORM_ACT_STATS
-        fprintf(f, "\n");
+    fprintf(f, "\n");
 #endif
+  }
 
-    }
-
 #if OUTPUT_NORM_ACT_STATS
-    fclose(f);
+  fclose(f);
 #endif
-
 }
 #endif
 
@@ -261,519 +233,468 @@
 /* Loop through all MBs. Note activity of each, average activity and
  * calculate a normalized activity for each
  */
-static void build_activity_map( VP8_COMP *cpi )
-{
-    MACROBLOCK *const x = & cpi->mb;
-    MACROBLOCKD *xd = &x->e_mbd;
-    VP8_COMMON *const cm = & cpi->common;
+static void build_activity_map(VP8_COMP *cpi) {
+  MACROBLOCK *const x = &cpi->mb;
+  MACROBLOCKD *xd = &x->e_mbd;
+  VP8_COMMON *const cm = &cpi->common;
 
 #if ALT_ACT_MEASURE
-    YV12_BUFFER_CONFIG *new_yv12 = &cm->yv12_fb[cm->new_fb_idx];
-    int recon_yoffset;
-    int recon_y_stride = new_yv12->y_stride;
+  YV12_BUFFER_CONFIG *new_yv12 = &cm->yv12_fb[cm->new_fb_idx];
+  int recon_yoffset;
+  int recon_y_stride = new_yv12->y_stride;
 #endif
 
-    int mb_row, mb_col;
-    unsigned int mb_activity;
-    int64_t activity_sum = 0;
+  int mb_row, mb_col;
+  unsigned int mb_activity;
+  int64_t activity_sum = 0;
 
-    /* for each macroblock row in image */
-    for (mb_row = 0; mb_row < cm->mb_rows; mb_row++)
-    {
+  /* for each macroblock row in image */
+  for (mb_row = 0; mb_row < cm->mb_rows; mb_row++) {
 #if ALT_ACT_MEASURE
-        /* reset above block coeffs */
-        xd->up_available = (mb_row != 0);
-        recon_yoffset = (mb_row * recon_y_stride * 16);
+    /* reset above block coeffs */
+    xd->up_available = (mb_row != 0);
+    recon_yoffset = (mb_row * recon_y_stride * 16);
 #endif
-        /* for each macroblock col in image */
-        for (mb_col = 0; mb_col < cm->mb_cols; mb_col++)
-        {
+    /* for each macroblock col in image */
+    for (mb_col = 0; mb_col < cm->mb_cols; mb_col++) {
 #if ALT_ACT_MEASURE
-            xd->dst.y_buffer = new_yv12->y_buffer + recon_yoffset;
-            xd->left_available = (mb_col != 0);
-            recon_yoffset += 16;
+      xd->dst.y_buffer = new_yv12->y_buffer + recon_yoffset;
+      xd->left_available = (mb_col != 0);
+      recon_yoffset += 16;
 #endif
-            /* Copy current mb to a buffer */
-            vp8_copy_mem16x16(x->src.y_buffer, x->src.y_stride, x->thismb, 16);
+      /* Copy current mb to a buffer */
+      vp8_copy_mem16x16(x->src.y_buffer, x->src.y_stride, x->thismb, 16);
 
-            /* measure activity */
-            mb_activity = mb_activity_measure( cpi, x, mb_row, mb_col );
+      /* measure activity */
+      mb_activity = mb_activity_measure(cpi, x, mb_row, mb_col);
 
-            /* Keep frame sum */
-            activity_sum += mb_activity;
+      /* Keep frame sum */
+      activity_sum += mb_activity;
 
-            /* Store MB level activity details. */
-            *x->mb_activity_ptr = mb_activity;
+      /* Store MB level activity details. */
+      *x->mb_activity_ptr = mb_activity;
 
-            /* Increment activity map pointer */
-            x->mb_activity_ptr++;
+      /* Increment activity map pointer */
+      x->mb_activity_ptr++;
 
-            /* adjust to the next column of source macroblocks */
-            x->src.y_buffer += 16;
-        }
+      /* adjust to the next column of source macroblocks */
+      x->src.y_buffer += 16;
+    }
 
+    /* adjust to the next row of mbs */
+    x->src.y_buffer += 16 * x->src.y_stride - 16 * cm->mb_cols;
 
-        /* adjust to the next row of mbs */
-        x->src.y_buffer += 16 * x->src.y_stride - 16 * cm->mb_cols;
-
 #if ALT_ACT_MEASURE
-        /* extend the recon for intra prediction */
-        vp8_extend_mb_row(new_yv12, xd->dst.y_buffer + 16,
-                          xd->dst.u_buffer + 8, xd->dst.v_buffer + 8);
+    /* extend the recon for intra prediction */
+    vp8_extend_mb_row(new_yv12, xd->dst.y_buffer + 16, xd->dst.u_buffer + 8,
+                      xd->dst.v_buffer + 8);
 #endif
+  }
 
-    }
+  /* Calculate an "average" MB activity */
+  calc_av_activity(cpi, activity_sum);
 
-    /* Calculate an "average" MB activity */
-    calc_av_activity(cpi, activity_sum);
-
 #if USE_ACT_INDEX
-    /* Calculate an activity index number of each mb */
-    calc_activity_index( cpi, x );
+  /* Calculate an activity index number of each mb */
+  calc_activity_index(cpi, x);
 #endif
-
 }
 
 /* Macroblock activity masking */
-void vp8_activity_masking(VP8_COMP *cpi, MACROBLOCK *x)
-{
+void vp8_activity_masking(VP8_COMP *cpi, MACROBLOCK *x) {
 #if USE_ACT_INDEX
-    x->rdmult += *(x->mb_activity_ptr) * (x->rdmult >> 2);
-    x->errorperbit = x->rdmult * 100 /(110 * x->rddiv);
-    x->errorperbit += (x->errorperbit==0);
+  x->rdmult += *(x->mb_activity_ptr) * (x->rdmult >> 2);
+  x->errorperbit = x->rdmult * 100 / (110 * x->rddiv);
+  x->errorperbit += (x->errorperbit == 0);
 #else
-    int64_t a;
-    int64_t b;
-    int64_t act = *(x->mb_activity_ptr);
+  int64_t a;
+  int64_t b;
+  int64_t act = *(x->mb_activity_ptr);
 
-    /* Apply the masking to the RD multiplier. */
-    a = act + (2*cpi->activity_avg);
-    b = (2*act) + cpi->activity_avg;
+  /* Apply the masking to the RD multiplier. */
+  a = act + (2 * cpi->activity_avg);
+  b = (2 * act) + cpi->activity_avg;
 
-    x->rdmult = (unsigned int)(((int64_t)x->rdmult*b + (a>>1))/a);
-    x->errorperbit = x->rdmult * 100 /(110 * x->rddiv);
-    x->errorperbit += (x->errorperbit==0);
+  x->rdmult = (unsigned int)(((int64_t)x->rdmult * b + (a >> 1)) / a);
+  x->errorperbit = x->rdmult * 100 / (110 * x->rddiv);
+  x->errorperbit += (x->errorperbit == 0);
 #endif
 
-    /* Activity based Zbin adjustment */
-    adjust_act_zbin(cpi, x);
+  /* Activity based Zbin adjustment */
+  adjust_act_zbin(cpi, x);
 }
 
-static
-void encode_mb_row(VP8_COMP *cpi,
-                   VP8_COMMON *cm,
-                   int mb_row,
-                   MACROBLOCK  *x,
-                   MACROBLOCKD *xd,
-                   TOKENEXTRA **tp,
-                   int *segment_counts,
-                   int *totalrate)
-{
-    int recon_yoffset, recon_uvoffset;
-    int mb_col;
-    int ref_fb_idx = cm->lst_fb_idx;
-    int dst_fb_idx = cm->new_fb_idx;
-    int recon_y_stride = cm->yv12_fb[ref_fb_idx].y_stride;
-    int recon_uv_stride = cm->yv12_fb[ref_fb_idx].uv_stride;
-    int map_index = (mb_row * cpi->common.mb_cols);
+static void encode_mb_row(VP8_COMP *cpi, VP8_COMMON *cm, int mb_row,
+                          MACROBLOCK *x, MACROBLOCKD *xd, TOKENEXTRA **tp,
+                          int *segment_counts, int *totalrate) {
+  int recon_yoffset, recon_uvoffset;
+  int mb_col;
+  int ref_fb_idx = cm->lst_fb_idx;
+  int dst_fb_idx = cm->new_fb_idx;
+  int recon_y_stride = cm->yv12_fb[ref_fb_idx].y_stride;
+  int recon_uv_stride = cm->yv12_fb[ref_fb_idx].uv_stride;
+  int map_index = (mb_row * cpi->common.mb_cols);
 
 #if (CONFIG_REALTIME_ONLY & CONFIG_ONTHEFLY_BITPACKING)
-    const int num_part = (1 << cm->multi_token_partition);
-    TOKENEXTRA * tp_start = cpi->tok;
-    vp8_writer *w;
+  const int num_part = (1 << cm->multi_token_partition);
+  TOKENEXTRA *tp_start = cpi->tok;
+  vp8_writer *w;
 #endif
 
 #if CONFIG_MULTITHREAD
-    const int nsync = cpi->mt_sync_range;
-    const int rightmost_col = cm->mb_cols + nsync;
-    const int *last_row_current_mb_col;
-    int *current_mb_col = &cpi->mt_current_mb_col[mb_row];
+  const int nsync = cpi->mt_sync_range;
+  const int rightmost_col = cm->mb_cols + nsync;
+  const int *last_row_current_mb_col;
+  int *current_mb_col = &cpi->mt_current_mb_col[mb_row];
 
-    if ((cpi->b_multi_threaded != 0) && (mb_row != 0))
-        last_row_current_mb_col = &cpi->mt_current_mb_col[mb_row - 1];
-    else
-        last_row_current_mb_col = &rightmost_col;
+  if ((cpi->b_multi_threaded != 0) && (mb_row != 0))
+    last_row_current_mb_col = &cpi->mt_current_mb_col[mb_row - 1];
+  else
+    last_row_current_mb_col = &rightmost_col;
 #endif
 
 #if (CONFIG_REALTIME_ONLY & CONFIG_ONTHEFLY_BITPACKING)
-    if(num_part > 1)
-        w= &cpi->bc[1 + (mb_row % num_part)];
-    else
-        w = &cpi->bc[1];
+  if (num_part > 1)
+    w = &cpi->bc[1 + (mb_row % num_part)];
+  else
+    w = &cpi->bc[1];
 #endif
 
-    /* reset above block coeffs */
-    xd->above_context = cm->above_context;
+  /* reset above block coeffs */
+  xd->above_context = cm->above_context;
 
-    xd->up_available = (mb_row != 0);
-    recon_yoffset = (mb_row * recon_y_stride * 16);
-    recon_uvoffset = (mb_row * recon_uv_stride * 8);
+  xd->up_available = (mb_row != 0);
+  recon_yoffset = (mb_row * recon_y_stride * 16);
+  recon_uvoffset = (mb_row * recon_uv_stride * 8);
 
-    cpi->tplist[mb_row].start = *tp;
-    /* printf("Main mb_row = %d\n", mb_row); */
+  cpi->tplist[mb_row].start = *tp;
+  /* printf("Main mb_row = %d\n", mb_row); */
 
-    /* Distance of Mb to the top & bottom edges, specified in 1/8th pel
-     * units as they are always compared to values that are in 1/8th pel
-     */
-    xd->mb_to_top_edge = -((mb_row * 16) << 3);
-    xd->mb_to_bottom_edge = ((cm->mb_rows - 1 - mb_row) * 16) << 3;
+  /* Distance of Mb to the top & bottom edges, specified in 1/8th pel
+   * units as they are always compared to values that are in 1/8th pel
+   */
+  xd->mb_to_top_edge = -((mb_row * 16) << 3);
+  xd->mb_to_bottom_edge = ((cm->mb_rows - 1 - mb_row) * 16) << 3;
 
-    /* Set up limit values for vertical motion vector components
-     * to prevent them extending beyond the UMV borders
-     */
-    x->mv_row_min = -((mb_row * 16) + (VP8BORDERINPIXELS - 16));
-    x->mv_row_max = ((cm->mb_rows - 1 - mb_row) * 16)
-                        + (VP8BORDERINPIXELS - 16);
+  /* Set up limit values for vertical motion vector components
+   * to prevent them extending beyond the UMV borders
+   */
+  x->mv_row_min = -((mb_row * 16) + (VP8BORDERINPIXELS - 16));
+  x->mv_row_max = ((cm->mb_rows - 1 - mb_row) * 16) + (VP8BORDERINPIXELS - 16);
 
-    /* Set the mb activity pointer to the start of the row. */
-    x->mb_activity_ptr = &cpi->mb_activity_map[map_index];
+  /* Set the mb activity pointer to the start of the row. */
+  x->mb_activity_ptr = &cpi->mb_activity_map[map_index];
 
-    /* for each macroblock col in image */
-    for (mb_col = 0; mb_col < cm->mb_cols; mb_col++)
-    {
-
-#if  (CONFIG_REALTIME_ONLY & CONFIG_ONTHEFLY_BITPACKING)
-        *tp = cpi->tok;
+  /* for each macroblock col in image */
+  for (mb_col = 0; mb_col < cm->mb_cols; mb_col++) {
+#if (CONFIG_REALTIME_ONLY & CONFIG_ONTHEFLY_BITPACKING)
+    *tp = cpi->tok;
 #endif
-        /* Distance of Mb to the left & right edges, specified in
-         * 1/8th pel units as they are always compared to values
-         * that are in 1/8th pel units
-         */
-        xd->mb_to_left_edge = -((mb_col * 16) << 3);
-        xd->mb_to_right_edge = ((cm->mb_cols - 1 - mb_col) * 16) << 3;
+    /* Distance of Mb to the left & right edges, specified in
+     * 1/8th pel units as they are always compared to values
+     * that are in 1/8th pel units
+     */
+    xd->mb_to_left_edge = -((mb_col * 16) << 3);
+    xd->mb_to_right_edge = ((cm->mb_cols - 1 - mb_col) * 16) << 3;
 
-        /* Set up limit values for horizontal motion vector components
-         * to prevent them extending beyond the UMV borders
-         */
-        x->mv_col_min = -((mb_col * 16) + (VP8BORDERINPIXELS - 16));
-        x->mv_col_max = ((cm->mb_cols - 1 - mb_col) * 16)
-                            + (VP8BORDERINPIXELS - 16);
+    /* Set up limit values for horizontal motion vector components
+     * to prevent them extending beyond the UMV borders
+     */
+    x->mv_col_min = -((mb_col * 16) + (VP8BORDERINPIXELS - 16));
+    x->mv_col_max =
+        ((cm->mb_cols - 1 - mb_col) * 16) + (VP8BORDERINPIXELS - 16);
 
-        xd->dst.y_buffer = cm->yv12_fb[dst_fb_idx].y_buffer + recon_yoffset;
-        xd->dst.u_buffer = cm->yv12_fb[dst_fb_idx].u_buffer + recon_uvoffset;
-        xd->dst.v_buffer = cm->yv12_fb[dst_fb_idx].v_buffer + recon_uvoffset;
-        xd->left_available = (mb_col != 0);
+    xd->dst.y_buffer = cm->yv12_fb[dst_fb_idx].y_buffer + recon_yoffset;
+    xd->dst.u_buffer = cm->yv12_fb[dst_fb_idx].u_buffer + recon_uvoffset;
+    xd->dst.v_buffer = cm->yv12_fb[dst_fb_idx].v_buffer + recon_uvoffset;
+    xd->left_available = (mb_col != 0);
 
-        x->rddiv = cpi->RDDIV;
-        x->rdmult = cpi->RDMULT;
+    x->rddiv = cpi->RDDIV;
+    x->rdmult = cpi->RDMULT;
 
-        /* Copy current mb to a buffer */
-        vp8_copy_mem16x16(x->src.y_buffer, x->src.y_stride, x->thismb, 16);
+    /* Copy current mb to a buffer */
+    vp8_copy_mem16x16(x->src.y_buffer, x->src.y_stride, x->thismb, 16);
 
 #if CONFIG_MULTITHREAD
-        if (cpi->b_multi_threaded != 0) {
-            if (((mb_col - 1) % nsync) == 0) {
-                pthread_mutex_t *mutex = &cpi->pmutex[mb_row];
-                protected_write(mutex, current_mb_col, mb_col - 1);
-            }
+    if (cpi->b_multi_threaded != 0) {
+      if (((mb_col - 1) % nsync) == 0) {
+        pthread_mutex_t *mutex = &cpi->pmutex[mb_row];
+        protected_write(mutex, current_mb_col, mb_col - 1);
+      }
 
-            if (mb_row && !(mb_col & (nsync - 1))) {
-                pthread_mutex_t *mutex = &cpi->pmutex[mb_row-1];
-                sync_read(mutex, mb_col, last_row_current_mb_col, nsync);
-            }
-        }
+      if (mb_row && !(mb_col & (nsync - 1))) {
+        pthread_mutex_t *mutex = &cpi->pmutex[mb_row - 1];
+        sync_read(mutex, mb_col, last_row_current_mb_col, nsync);
+      }
+    }
 #endif
 
-        if(cpi->oxcf.tuning == VP8_TUNE_SSIM)
-            vp8_activity_masking(cpi, x);
+    if (cpi->oxcf.tuning == VP8_TUNE_SSIM) vp8_activity_masking(cpi, x);
 
-        /* Is segmentation enabled */
-        /* MB level adjustment to quantizer */
-        if (xd->segmentation_enabled)
-        {
-            /* Code to set segment id in xd->mbmi.segment_id for current MB
-             * (with range checking)
-             */
-            if (cpi->segmentation_map[map_index+mb_col] <= 3)
-                xd->mode_info_context->mbmi.segment_id = cpi->segmentation_map[map_index+mb_col];
-            else
-                xd->mode_info_context->mbmi.segment_id = 0;
+    /* Is segmentation enabled */
+    /* MB level adjustment to quantizer */
+    if (xd->segmentation_enabled) {
+      /* Code to set segment id in xd->mbmi.segment_id for current MB
+       * (with range checking)
+       */
+      if (cpi->segmentation_map[map_index + mb_col] <= 3)
+        xd->mode_info_context->mbmi.segment_id =
+            cpi->segmentation_map[map_index + mb_col];
+      else
+        xd->mode_info_context->mbmi.segment_id = 0;
 
-            vp8cx_mb_init_quantizer(cpi, x, 1);
-        }
-        else
-            /* Set to Segment 0 by default */
-            xd->mode_info_context->mbmi.segment_id = 0;
+      vp8cx_mb_init_quantizer(cpi, x, 1);
+    } else
+      /* Set to Segment 0 by default */
+      xd->mode_info_context->mbmi.segment_id = 0;
 
-        x->active_ptr = cpi->active_map + map_index + mb_col;
+    x->active_ptr = cpi->active_map + map_index + mb_col;
 
-        if (cm->frame_type == KEY_FRAME)
-        {
-            *totalrate += vp8cx_encode_intra_macroblock(cpi, x, tp);
+    if (cm->frame_type == KEY_FRAME) {
+      *totalrate += vp8cx_encode_intra_macroblock(cpi, x, tp);
 #ifdef MODE_STATS
-            y_modes[xd->mbmi.mode] ++;
+      y_modes[xd->mbmi.mode]++;
 #endif
-        }
-        else
-        {
-            *totalrate += vp8cx_encode_inter_macroblock(cpi, x, tp, recon_yoffset, recon_uvoffset, mb_row, mb_col);
+    } else {
+      *totalrate += vp8cx_encode_inter_macroblock(
+          cpi, x, tp, recon_yoffset, recon_uvoffset, mb_row, mb_col);
 
 #ifdef MODE_STATS
-            inter_y_modes[xd->mbmi.mode] ++;
+      inter_y_modes[xd->mbmi.mode]++;
 
-            if (xd->mbmi.mode == SPLITMV)
-            {
-                int b;
+      if (xd->mbmi.mode == SPLITMV) {
+        int b;
 
-                for (b = 0; b < xd->mbmi.partition_count; b++)
-                {
-                    inter_b_modes[x->partition->bmi[b].mode] ++;
-                }
-            }
+        for (b = 0; b < xd->mbmi.partition_count; b++) {
+          inter_b_modes[x->partition->bmi[b].mode]++;
+        }
+      }
 
 #endif
 
-            // Keep track of how many (consecutive) times a  block is coded
-            // as ZEROMV_LASTREF, for base layer frames.
-            // Reset to 0 if its coded as anything else.
-            if (cpi->current_layer == 0) {
-              if (xd->mode_info_context->mbmi.mode == ZEROMV &&
-                  xd->mode_info_context->mbmi.ref_frame == LAST_FRAME) {
-                // Increment, check for wrap-around.
-                if (cpi->consec_zero_last[map_index+mb_col] < 255)
-                  cpi->consec_zero_last[map_index+mb_col] += 1;
-                if (cpi->consec_zero_last_mvbias[map_index+mb_col] < 255)
-                  cpi->consec_zero_last_mvbias[map_index+mb_col] += 1;
-              } else {
-                cpi->consec_zero_last[map_index+mb_col] = 0;
-                cpi->consec_zero_last_mvbias[map_index+mb_col] = 0;
-              }
-              if (x->zero_last_dot_suppress)
-                cpi->consec_zero_last_mvbias[map_index+mb_col] = 0;
-            }
+      // Keep track of how many (consecutive) times a  block is coded
+      // as ZEROMV_LASTREF, for base layer frames.
+      // Reset to 0 if its coded as anything else.
+      if (cpi->current_layer == 0) {
+        if (xd->mode_info_context->mbmi.mode == ZEROMV &&
+            xd->mode_info_context->mbmi.ref_frame == LAST_FRAME) {
+          // Increment, check for wrap-around.
+          if (cpi->consec_zero_last[map_index + mb_col] < 255)
+            cpi->consec_zero_last[map_index + mb_col] += 1;
+          if (cpi->consec_zero_last_mvbias[map_index + mb_col] < 255)
+            cpi->consec_zero_last_mvbias[map_index + mb_col] += 1;
+        } else {
+          cpi->consec_zero_last[map_index + mb_col] = 0;
+          cpi->consec_zero_last_mvbias[map_index + mb_col] = 0;
+        }
+        if (x->zero_last_dot_suppress)
+          cpi->consec_zero_last_mvbias[map_index + mb_col] = 0;
+      }
 
-            /* Special case code for cyclic refresh
-             * If cyclic update enabled then copy xd->mbmi.segment_id; (which
-             * may have been updated based on mode during
-             * vp8cx_encode_inter_macroblock()) back into the global
-             * segmentation map
-             */
-            if ((cpi->current_layer == 0) &&
-                (cpi->cyclic_refresh_mode_enabled &&
-                 xd->segmentation_enabled))
-            {
-                cpi->segmentation_map[map_index+mb_col] = xd->mode_info_context->mbmi.segment_id;
+      /* Special case code for cyclic refresh
+       * If cyclic update enabled then copy xd->mbmi.segment_id; (which
+       * may have been updated based on mode during
+       * vp8cx_encode_inter_macroblock()) back into the global
+       * segmentation map
+       */
+      if ((cpi->current_layer == 0) &&
+          (cpi->cyclic_refresh_mode_enabled && xd->segmentation_enabled)) {
+        cpi->segmentation_map[map_index + mb_col] =
+            xd->mode_info_context->mbmi.segment_id;
 
-                /* If the block has been refreshed mark it as clean (the
-                 * magnitude of the -ve influences how long it will be before
-                 * we consider another refresh):
-                 * Else if it was coded (last frame 0,0) and has not already
-                 * been refreshed then mark it as a candidate for cleanup
-                 * next time (marked 0) else mark it as dirty (1).
-                 */
-                if (xd->mode_info_context->mbmi.segment_id)
-                    cpi->cyclic_refresh_map[map_index+mb_col] = -1;
-                else if ((xd->mode_info_context->mbmi.mode == ZEROMV) && (xd->mode_info_context->mbmi.ref_frame == LAST_FRAME))
-                {
-                    if (cpi->cyclic_refresh_map[map_index+mb_col] == 1)
-                        cpi->cyclic_refresh_map[map_index+mb_col] = 0;
-                }
-                else
-                    cpi->cyclic_refresh_map[map_index+mb_col] = 1;
+        /* If the block has been refreshed mark it as clean (the
+         * magnitude of the -ve influences how long it will be before
+         * we consider another refresh):
+         * Else if it was coded (last frame 0,0) and has not already
+         * been refreshed then mark it as a candidate for cleanup
+         * next time (marked 0) else mark it as dirty (1).
+         */
+        if (xd->mode_info_context->mbmi.segment_id)
+          cpi->cyclic_refresh_map[map_index + mb_col] = -1;
+        else if ((xd->mode_info_context->mbmi.mode == ZEROMV) &&
+                 (xd->mode_info_context->mbmi.ref_frame == LAST_FRAME)) {
+          if (cpi->cyclic_refresh_map[map_index + mb_col] == 1)
+            cpi->cyclic_refresh_map[map_index + mb_col] = 0;
+        } else
+          cpi->cyclic_refresh_map[map_index + mb_col] = 1;
+      }
+    }
 
-            }
-        }
+    cpi->tplist[mb_row].stop = *tp;
 
-        cpi->tplist[mb_row].stop = *tp;
-
 #if CONFIG_REALTIME_ONLY & CONFIG_ONTHEFLY_BITPACKING
-        /* pack tokens for this MB */
-        {
-            int tok_count = *tp - tp_start;
-            vp8_pack_tokens(w, tp_start, tok_count);
-        }
+    /* pack tokens for this MB */
+    {
+      int tok_count = *tp - tp_start;
+      vp8_pack_tokens(w, tp_start, tok_count);
+    }
 #endif
-        /* Increment pointer into gf usage flags structure. */
-        x->gf_active_ptr++;
+    /* Increment pointer into gf usage flags structure. */
+    x->gf_active_ptr++;
 
-        /* Increment the activity mask pointers. */
-        x->mb_activity_ptr++;
+    /* Increment the activity mask pointers. */
+    x->mb_activity_ptr++;
 
-        /* adjust to the next column of macroblocks */
-        x->src.y_buffer += 16;
-        x->src.u_buffer += 8;
-        x->src.v_buffer += 8;
+    /* adjust to the next column of macroblocks */
+    x->src.y_buffer += 16;
+    x->src.u_buffer += 8;
+    x->src.v_buffer += 8;
 
-        recon_yoffset += 16;
-        recon_uvoffset += 8;
+    recon_yoffset += 16;
+    recon_uvoffset += 8;
 
-        /* Keep track of segment usage */
-        segment_counts[xd->mode_info_context->mbmi.segment_id] ++;
+    /* Keep track of segment usage */
+    segment_counts[xd->mode_info_context->mbmi.segment_id]++;
 
-        /* skip to next mb */
-        xd->mode_info_context++;
-        x->partition_info++;
-        xd->above_context++;
-    }
+    /* skip to next mb */
+    xd->mode_info_context++;
+    x->partition_info++;
+    xd->above_context++;
+  }
 
-    /* extend the recon for intra prediction */
-    vp8_extend_mb_row( &cm->yv12_fb[dst_fb_idx],
-                        xd->dst.y_buffer + 16,
-                        xd->dst.u_buffer + 8,
-                        xd->dst.v_buffer + 8);
+  /* extend the recon for intra prediction */
+  vp8_extend_mb_row(&cm->yv12_fb[dst_fb_idx], xd->dst.y_buffer + 16,
+                    xd->dst.u_buffer + 8, xd->dst.v_buffer + 8);
 
 #if CONFIG_MULTITHREAD
-    if (cpi->b_multi_threaded != 0)
-        protected_write(&cpi->pmutex[mb_row], current_mb_col, rightmost_col);
+  if (cpi->b_multi_threaded != 0)
+    protected_write(&cpi->pmutex[mb_row], current_mb_col, rightmost_col);
 #endif
 
-    /* this is to account for the border */
-    xd->mode_info_context++;
-    x->partition_info++;
+  /* this is to account for the border */
+  xd->mode_info_context++;
+  x->partition_info++;
 }
 
-static void init_encode_frame_mb_context(VP8_COMP *cpi)
-{
-    MACROBLOCK *const x = & cpi->mb;
-    VP8_COMMON *const cm = & cpi->common;
-    MACROBLOCKD *const xd = & x->e_mbd;
+static void init_encode_frame_mb_context(VP8_COMP *cpi) {
+  MACROBLOCK *const x = &cpi->mb;
+  VP8_COMMON *const cm = &cpi->common;
+  MACROBLOCKD *const xd = &x->e_mbd;
 
-    /* GF active flags data structure */
-    x->gf_active_ptr = (signed char *)cpi->gf_active_flags;
+  /* GF active flags data structure */
+  x->gf_active_ptr = (signed char *)cpi->gf_active_flags;
 
-    /* Activity map pointer */
-    x->mb_activity_ptr = cpi->mb_activity_map;
+  /* Activity map pointer */
+  x->mb_activity_ptr = cpi->mb_activity_map;
 
-    x->act_zbin_adj = 0;
+  x->act_zbin_adj = 0;
 
-    x->partition_info = x->pi;
+  x->partition_info = x->pi;
 
-    xd->mode_info_context = cm->mi;
-    xd->mode_info_stride = cm->mode_info_stride;
+  xd->mode_info_context = cm->mi;
+  xd->mode_info_stride = cm->mode_info_stride;
 
-    xd->frame_type = cm->frame_type;
+  xd->frame_type = cm->frame_type;
 
-    /* reset intra mode contexts */
-    if (cm->frame_type == KEY_FRAME)
-        vp8_init_mbmode_probs(cm);
+  /* reset intra mode contexts */
+  if (cm->frame_type == KEY_FRAME) vp8_init_mbmode_probs(cm);
 
-    /* Copy data over into macro block data structures. */
-    x->src = * cpi->Source;
-    xd->pre = cm->yv12_fb[cm->lst_fb_idx];
-    xd->dst = cm->yv12_fb[cm->new_fb_idx];
+  /* Copy data over into macro block data structures. */
+  x->src = *cpi->Source;
+  xd->pre = cm->yv12_fb[cm->lst_fb_idx];
+  xd->dst = cm->yv12_fb[cm->new_fb_idx];
 
-    /* set up frame for intra coded blocks */
-    vp8_setup_intra_recon(&cm->yv12_fb[cm->new_fb_idx]);
+  /* set up frame for intra coded blocks */
+  vp8_setup_intra_recon(&cm->yv12_fb[cm->new_fb_idx]);
 
-    vp8_build_block_offsets(x);
+  vp8_build_block_offsets(x);
 
-    xd->mode_info_context->mbmi.mode = DC_PRED;
-    xd->mode_info_context->mbmi.uv_mode = DC_PRED;
+  xd->mode_info_context->mbmi.mode = DC_PRED;
+  xd->mode_info_context->mbmi.uv_mode = DC_PRED;
 
-    xd->left_context = &cm->left_context;
+  xd->left_context = &cm->left_context;
 
-    x->mvc = cm->fc.mvc;
+  x->mvc = cm->fc.mvc;
 
-    memset(cm->above_context, 0, sizeof(ENTROPY_CONTEXT_PLANES) * cm->mb_cols);
+  memset(cm->above_context, 0, sizeof(ENTROPY_CONTEXT_PLANES) * cm->mb_cols);
 
-    /* Special case treatment when GF and ARF are not sensible options
-     * for reference
-     */
-    if (cpi->ref_frame_flags == VP8_LAST_FRAME)
-        vp8_calc_ref_frame_costs(x->ref_frame_cost,
-                                 cpi->prob_intra_coded,255,128);
-    else if ((cpi->oxcf.number_of_layers > 1) &&
-               (cpi->ref_frame_flags == VP8_GOLD_FRAME))
-        vp8_calc_ref_frame_costs(x->ref_frame_cost,
-                                 cpi->prob_intra_coded,1,255);
-    else if ((cpi->oxcf.number_of_layers > 1) &&
-                (cpi->ref_frame_flags == VP8_ALTR_FRAME))
-        vp8_calc_ref_frame_costs(x->ref_frame_cost,
-                                 cpi->prob_intra_coded,1,1);
-    else
-        vp8_calc_ref_frame_costs(x->ref_frame_cost,
-                                 cpi->prob_intra_coded,
-                                 cpi->prob_last_coded,
-                                 cpi->prob_gf_coded);
+  /* Special case treatment when GF and ARF are not sensible options
+   * for reference
+   */
+  if (cpi->ref_frame_flags == VP8_LAST_FRAME)
+    vp8_calc_ref_frame_costs(x->ref_frame_cost, cpi->prob_intra_coded, 255,
+                             128);
+  else if ((cpi->oxcf.number_of_layers > 1) &&
+           (cpi->ref_frame_flags == VP8_GOLD_FRAME))
+    vp8_calc_ref_frame_costs(x->ref_frame_cost, cpi->prob_intra_coded, 1, 255);
+  else if ((cpi->oxcf.number_of_layers > 1) &&
+           (cpi->ref_frame_flags == VP8_ALTR_FRAME))
+    vp8_calc_ref_frame_costs(x->ref_frame_cost, cpi->prob_intra_coded, 1, 1);
+  else
+    vp8_calc_ref_frame_costs(x->ref_frame_cost, cpi->prob_intra_coded,
+                             cpi->prob_last_coded, cpi->prob_gf_coded);
 
-    xd->fullpixel_mask = 0xffffffff;
-    if(cm->full_pixel)
-        xd->fullpixel_mask = 0xfffffff8;
+  xd->fullpixel_mask = 0xffffffff;
+  if (cm->full_pixel) xd->fullpixel_mask = 0xfffffff8;
 
-    vp8_zero(x->coef_counts);
-    vp8_zero(x->ymode_count);
-    vp8_zero(x->uv_mode_count)
-    x->prediction_error = 0;
-    x->intra_error = 0;
-    vp8_zero(x->count_mb_ref_frame_usage);
+  vp8_zero(x->coef_counts);
+  vp8_zero(x->ymode_count);
+  vp8_zero(x->uv_mode_count) x->prediction_error = 0;
+  x->intra_error = 0;
+  vp8_zero(x->count_mb_ref_frame_usage);
 }
 
 #if CONFIG_MULTITHREAD
-static void sum_coef_counts(MACROBLOCK *x, MACROBLOCK *x_thread)
-{
-    int i = 0;
-    do
-    {
-        int j = 0;
-        do
-        {
-            int k = 0;
-            do
-            {
-                /* at every context */
+static void sum_coef_counts(MACROBLOCK *x, MACROBLOCK *x_thread) {
+  int i = 0;
+  do {
+    int j = 0;
+    do {
+      int k = 0;
+      do {
+        /* at every context */
 
-                /* calc probs and branch cts for this frame only */
-                int t = 0;      /* token/prob index */
+        /* calc probs and branch cts for this frame only */
+        int t = 0; /* token/prob index */
 
-                do
-                {
-                    x->coef_counts [i][j][k][t] +=
-                        x_thread->coef_counts [i][j][k][t];
-                }
-                while (++t < ENTROPY_NODES);
-            }
-            while (++k < PREV_COEF_CONTEXTS);
-        }
-        while (++j < COEF_BANDS);
-    }
-    while (++i < BLOCK_TYPES);
-}
+        do {
+          x->coef_counts[i][j][k][t] += x_thread->coef_counts[i][j][k][t];
+        } while (++t < ENTROPY_NODES);
+      } while (++k < PREV_COEF_CONTEXTS);
+    } while (++j < COEF_BANDS);
+  } while (++i < BLOCK_TYPES);
+}
 #endif  // CONFIG_MULTITHREAD
 
-void vp8_encode_frame(VP8_COMP *cpi)
-{
-    int mb_row;
-    MACROBLOCK *const x = & cpi->mb;
-    VP8_COMMON *const cm = & cpi->common;
-    MACROBLOCKD *const xd = & x->e_mbd;
-    TOKENEXTRA *tp = cpi->tok;
-    int segment_counts[MAX_MB_SEGMENTS];
-    int totalrate;
+void vp8_encode_frame(VP8_COMP *cpi) {
+  int mb_row;
+  MACROBLOCK *const x = &cpi->mb;
+  VP8_COMMON *const cm = &cpi->common;
+  MACROBLOCKD *const xd = &x->e_mbd;
+  TOKENEXTRA *tp = cpi->tok;
+  int segment_counts[MAX_MB_SEGMENTS];
+  int totalrate;
 #if CONFIG_REALTIME_ONLY & CONFIG_ONTHEFLY_BITPACKING
-    BOOL_CODER * bc = &cpi->bc[1]; /* bc[0] is for control partition */
-    const int num_part = (1 << cm->multi_token_partition);
+  BOOL_CODER *bc = &cpi->bc[1]; /* bc[0] is for control partition */
+  const int num_part = (1 << cm->multi_token_partition);
 #endif
 
-    memset(segment_counts, 0, sizeof(segment_counts));
-    totalrate = 0;
+  memset(segment_counts, 0, sizeof(segment_counts));
+  totalrate = 0;
 
-    if (cpi->compressor_speed == 2)
-    {
-        if (cpi->oxcf.cpu_used < 0)
-            cpi->Speed = -(cpi->oxcf.cpu_used);
-        else
-            vp8_auto_select_speed(cpi);
-    }
-
-    /* Functions setup for all frame types so we can use MC in AltRef */
-    if(!cm->use_bilinear_mc_filter)
-    {
-        xd->subpixel_predict        = vp8_sixtap_predict4x4;
-        xd->subpixel_predict8x4     = vp8_sixtap_predict8x4;
-        xd->subpixel_predict8x8     = vp8_sixtap_predict8x8;
-        xd->subpixel_predict16x16   = vp8_sixtap_predict16x16;
-    }
+  if (cpi->compressor_speed == 2) {
+    if (cpi->oxcf.cpu_used < 0)
+      cpi->Speed = -(cpi->oxcf.cpu_used);
     else
-    {
-        xd->subpixel_predict        = vp8_bilinear_predict4x4;
-        xd->subpixel_predict8x4     = vp8_bilinear_predict8x4;
-        xd->subpixel_predict8x8     = vp8_bilinear_predict8x8;
-        xd->subpixel_predict16x16   = vp8_bilinear_predict16x16;
-    }
+      vp8_auto_select_speed(cpi);
+  }
 
-    cpi->mb.skip_true_count = 0;
-    cpi->tok_count = 0;
+  /* Functions setup for all frame types so we can use MC in AltRef */
+  if (!cm->use_bilinear_mc_filter) {
+    xd->subpixel_predict = vp8_sixtap_predict4x4;
+    xd->subpixel_predict8x4 = vp8_sixtap_predict8x4;
+    xd->subpixel_predict8x8 = vp8_sixtap_predict8x8;
+    xd->subpixel_predict16x16 = vp8_sixtap_predict16x16;
+  } else {
+    xd->subpixel_predict = vp8_bilinear_predict4x4;
+    xd->subpixel_predict8x4 = vp8_bilinear_predict8x4;
+    xd->subpixel_predict8x8 = vp8_bilinear_predict8x8;
+    xd->subpixel_predict16x16 = vp8_bilinear_predict16x16;
+  }
 
+  cpi->mb.skip_true_count = 0;
+  cpi->tok_count = 0;
+
 #if 0
     /* Experimental code */
     cpi->frame_distortion = 0;
@@ -780,444 +701,398 @@
     cpi->last_mb_distortion = 0;
 #endif
 
-    xd->mode_info_context = cm->mi;
+  xd->mode_info_context = cm->mi;
 
-    vp8_zero(cpi->mb.MVcount);
+  vp8_zero(cpi->mb.MVcount);
 
-    vp8cx_frame_init_quantizer(cpi);
+  vp8cx_frame_init_quantizer(cpi);
 
-    vp8_initialize_rd_consts(cpi, x,
-                             vp8_dc_quant(cm->base_qindex, cm->y1dc_delta_q));
+  vp8_initialize_rd_consts(cpi, x,
+                           vp8_dc_quant(cm->base_qindex, cm->y1dc_delta_q));
 
-    vp8cx_initialize_me_consts(cpi, cm->base_qindex);
+  vp8cx_initialize_me_consts(cpi, cm->base_qindex);
 
-    if(cpi->oxcf.tuning == VP8_TUNE_SSIM)
-    {
-        /* Initialize encode frame context. */
-        init_encode_frame_mb_context(cpi);
+  if (cpi->oxcf.tuning == VP8_TUNE_SSIM) {
+    /* Initialize encode frame context. */
+    init_encode_frame_mb_context(cpi);
 
-        /* Build a frame level activity map */
-        build_activity_map(cpi);
-    }
+    /* Build a frame level activity map */
+    build_activity_map(cpi);
+  }
 
-    /* re-init encode frame context. */
-    init_encode_frame_mb_context(cpi);
+  /* re-init encode frame context. */
+  init_encode_frame_mb_context(cpi);
 
 #if CONFIG_REALTIME_ONLY & CONFIG_ONTHEFLY_BITPACKING
-    {
-        int i;
-        for(i = 0; i < num_part; i++)
-        {
-            vp8_start_encode(&bc[i], cpi->partition_d[i + 1],
-                    cpi->partition_d_end[i + 1]);
-            bc[i].error = &cm->error;
-        }
+  {
+    int i;
+    for (i = 0; i < num_part; i++) {
+      vp8_start_encode(&bc[i], cpi->partition_d[i + 1],
+                       cpi->partition_d_end[i + 1]);
+      bc[i].error = &cm->error;
     }
+  }
 
 #endif
 
-    {
-        struct vpx_usec_timer  emr_timer;
-        vpx_usec_timer_start(&emr_timer);
+  {
+    struct vpx_usec_timer emr_timer;
+    vpx_usec_timer_start(&emr_timer);
 
 #if CONFIG_MULTITHREAD
-        if (cpi->b_multi_threaded)
-        {
-            int i;
+    if (cpi->b_multi_threaded) {
+      int i;
 
-            vp8cx_init_mbrthread_data(cpi, x, cpi->mb_row_ei,
-                                      cpi->encoding_thread_count);
+      vp8cx_init_mbrthread_data(cpi, x, cpi->mb_row_ei,
+                                cpi->encoding_thread_count);
 
-            for (i = 0; i < cm->mb_rows; i++)
-                cpi->mt_current_mb_col[i] = -1;
+      for (i = 0; i < cm->mb_rows; i++) cpi->mt_current_mb_col[i] = -1;
 
-            for (i = 0; i < cpi->encoding_thread_count; i++)
-            {
-                sem_post(&cpi->h_event_start_encoding[i]);
-            }
+      for (i = 0; i < cpi->encoding_thread_count; i++) {
+        sem_post(&cpi->h_event_start_encoding[i]);
+      }
 
-            for (mb_row = 0; mb_row < cm->mb_rows; mb_row += (cpi->encoding_thread_count + 1))
-            {
-                vp8_zero(cm->left_context)
+      for (mb_row = 0; mb_row < cm->mb_rows;
+           mb_row += (cpi->encoding_thread_count + 1)) {
+        vp8_zero(cm->left_context)
 
 #if CONFIG_REALTIME_ONLY & CONFIG_ONTHEFLY_BITPACKING
-                tp = cpi->tok;
+            tp = cpi->tok;
 #else
-                tp = cpi->tok + mb_row * (cm->mb_cols * 16 * 24);
+            tp = cpi->tok + mb_row * (cm->mb_cols * 16 * 24);
 #endif
 
-                encode_mb_row(cpi, cm, mb_row, x, xd, &tp, segment_counts, &totalrate);
+        encode_mb_row(cpi, cm, mb_row, x, xd, &tp, segment_counts, &totalrate);
 
-                /* adjust to the next row of mbs */
-                x->src.y_buffer += 16 * x->src.y_stride * (cpi->encoding_thread_count + 1) - 16 * cm->mb_cols;
-                x->src.u_buffer +=  8 * x->src.uv_stride * (cpi->encoding_thread_count + 1) - 8 * cm->mb_cols;
-                x->src.v_buffer +=  8 * x->src.uv_stride * (cpi->encoding_thread_count + 1) - 8 * cm->mb_cols;
+        /* adjust to the next row of mbs */
+        x->src.y_buffer +=
+            16 * x->src.y_stride * (cpi->encoding_thread_count + 1) -
+            16 * cm->mb_cols;
+        x->src.u_buffer +=
+            8 * x->src.uv_stride * (cpi->encoding_thread_count + 1) -
+            8 * cm->mb_cols;
+        x->src.v_buffer +=
+            8 * x->src.uv_stride * (cpi->encoding_thread_count + 1) -
+            8 * cm->mb_cols;
 
-                xd->mode_info_context += xd->mode_info_stride * cpi->encoding_thread_count;
-                x->partition_info  += xd->mode_info_stride * cpi->encoding_thread_count;
-                x->gf_active_ptr   += cm->mb_cols * cpi->encoding_thread_count;
+        xd->mode_info_context +=
+            xd->mode_info_stride * cpi->encoding_thread_count;
+        x->partition_info += xd->mode_info_stride * cpi->encoding_thread_count;
+        x->gf_active_ptr += cm->mb_cols * cpi->encoding_thread_count;
 
-                if(mb_row == cm->mb_rows - 1)
-                {
-                    sem_post(&cpi->h_event_end_encoding); /* signal frame encoding end */
-                }
-            }
+        if (mb_row == cm->mb_rows - 1) {
+          sem_post(&cpi->h_event_end_encoding); /* signal frame encoding end */
+        }
+      }
 
-            sem_wait(&cpi->h_event_end_encoding); /* wait for other threads to finish */
+      sem_wait(
+          &cpi->h_event_end_encoding); /* wait for other threads to finish */
 
-            for (mb_row = 0; mb_row < cm->mb_rows; mb_row ++)
-            {
-                cpi->tok_count += (unsigned int)
-                  (cpi->tplist[mb_row].stop - cpi->tplist[mb_row].start);
-            }
+      for (mb_row = 0; mb_row < cm->mb_rows; mb_row++) {
+        cpi->tok_count += (unsigned int)(cpi->tplist[mb_row].stop -
+                                         cpi->tplist[mb_row].start);
+      }
 
-            if (xd->segmentation_enabled)
-            {
-                int j;
+      if (xd->segmentation_enabled) {
+        int j;
 
-                if (xd->segmentation_enabled)
-                {
-                    for (i = 0; i < cpi->encoding_thread_count; i++)
-                    {
-                        for (j = 0; j < 4; j++)
-                            segment_counts[j] += cpi->mb_row_ei[i].segment_counts[j];
-                    }
-                }
-            }
+        if (xd->segmentation_enabled) {
+          for (i = 0; i < cpi->encoding_thread_count; i++) {
+            for (j = 0; j < 4; j++)
+              segment_counts[j] += cpi->mb_row_ei[i].segment_counts[j];
+          }
+        }
+      }
 
-            for (i = 0; i < cpi->encoding_thread_count; i++)
-            {
-                int mode_count;
-                int c_idx;
-                totalrate += cpi->mb_row_ei[i].totalrate;
+      for (i = 0; i < cpi->encoding_thread_count; i++) {
+        int mode_count;
+        int c_idx;
+        totalrate += cpi->mb_row_ei[i].totalrate;
 
-                cpi->mb.skip_true_count += cpi->mb_row_ei[i].mb.skip_true_count;
+        cpi->mb.skip_true_count += cpi->mb_row_ei[i].mb.skip_true_count;
 
-                for(mode_count = 0; mode_count < VP8_YMODES; mode_count++)
-                    cpi->mb.ymode_count[mode_count] +=
-                        cpi->mb_row_ei[i].mb.ymode_count[mode_count];
+        for (mode_count = 0; mode_count < VP8_YMODES; mode_count++)
+          cpi->mb.ymode_count[mode_count] +=
+              cpi->mb_row_ei[i].mb.ymode_count[mode_count];
 
-                for(mode_count = 0; mode_count < VP8_UV_MODES; mode_count++)
-                    cpi->mb.uv_mode_count[mode_count] +=
-                        cpi->mb_row_ei[i].mb.uv_mode_count[mode_count];
+        for (mode_count = 0; mode_count < VP8_UV_MODES; mode_count++)
+          cpi->mb.uv_mode_count[mode_count] +=
+              cpi->mb_row_ei[i].mb.uv_mode_count[mode_count];
 
-                for(c_idx = 0; c_idx < MVvals; c_idx++)
-                {
-                    cpi->mb.MVcount[0][c_idx] +=
-                        cpi->mb_row_ei[i].mb.MVcount[0][c_idx];
-                    cpi->mb.MVcount[1][c_idx] +=
-                        cpi->mb_row_ei[i].mb.MVcount[1][c_idx];
-                }
+        for (c_idx = 0; c_idx < MVvals; c_idx++) {
+          cpi->mb.MVcount[0][c_idx] += cpi->mb_row_ei[i].mb.MVcount[0][c_idx];
+          cpi->mb.MVcount[1][c_idx] += cpi->mb_row_ei[i].mb.MVcount[1][c_idx];
+        }
 
-                cpi->mb.prediction_error +=
-                    cpi->mb_row_ei[i].mb.prediction_error;
-                cpi->mb.intra_error += cpi->mb_row_ei[i].mb.intra_error;
+        cpi->mb.prediction_error += cpi->mb_row_ei[i].mb.prediction_error;
+        cpi->mb.intra_error += cpi->mb_row_ei[i].mb.intra_error;
 
-                for(c_idx = 0; c_idx < MAX_REF_FRAMES; c_idx++)
-                    cpi->mb.count_mb_ref_frame_usage[c_idx] +=
-                        cpi->mb_row_ei[i].mb.count_mb_ref_frame_usage[c_idx];
+        for (c_idx = 0; c_idx < MAX_REF_FRAMES; c_idx++)
+          cpi->mb.count_mb_ref_frame_usage[c_idx] +=
+              cpi->mb_row_ei[i].mb.count_mb_ref_frame_usage[c_idx];
 
-                for(c_idx = 0; c_idx < MAX_ERROR_BINS; c_idx++)
-                    cpi->mb.error_bins[c_idx] +=
-                        cpi->mb_row_ei[i].mb.error_bins[c_idx];
+        for (c_idx = 0; c_idx < MAX_ERROR_BINS; c_idx++)
+          cpi->mb.error_bins[c_idx] += cpi->mb_row_ei[i].mb.error_bins[c_idx];
 
-                /* add up counts for each thread */
-                sum_coef_counts(x, &cpi->mb_row_ei[i].mb);
-            }
+        /* add up counts for each thread */
+        sum_coef_counts(x, &cpi->mb_row_ei[i].mb);
+      }
 
-        }
-        else
+    } else
 #endif  // CONFIG_MULTITHREAD
-        {
+    {
 
-            /* for each macroblock row in image */
-            for (mb_row = 0; mb_row < cm->mb_rows; mb_row++)
-            {
-                vp8_zero(cm->left_context)
+      /* for each macroblock row in image */
+      for (mb_row = 0; mb_row < cm->mb_rows; mb_row++) {
+        vp8_zero(cm->left_context)
 
 #if CONFIG_REALTIME_ONLY & CONFIG_ONTHEFLY_BITPACKING
-                tp = cpi->tok;
+            tp = cpi->tok;
 #endif
 
-                encode_mb_row(cpi, cm, mb_row, x, xd, &tp, segment_counts, &totalrate);
+        encode_mb_row(cpi, cm, mb_row, x, xd, &tp, segment_counts, &totalrate);
 
-                /* adjust to the next row of mbs */
-                x->src.y_buffer += 16 * x->src.y_stride - 16 * cm->mb_cols;
-                x->src.u_buffer += 8 * x->src.uv_stride - 8 * cm->mb_cols;
-                x->src.v_buffer += 8 * x->src.uv_stride - 8 * cm->mb_cols;
-            }
+        /* adjust to the next row of mbs */
+        x->src.y_buffer += 16 * x->src.y_stride - 16 * cm->mb_cols;
+        x->src.u_buffer += 8 * x->src.uv_stride - 8 * cm->mb_cols;
+        x->src.v_buffer += 8 * x->src.uv_stride - 8 * cm->mb_cols;
+      }
 
-            cpi->tok_count = (unsigned int)(tp - cpi->tok);
-        }
+      cpi->tok_count = (unsigned int)(tp - cpi->tok);
+    }
 
 #if CONFIG_REALTIME_ONLY & CONFIG_ONTHEFLY_BITPACKING
-        {
-            int i;
-            for(i = 0; i < num_part; i++)
-            {
-                vp8_stop_encode(&bc[i]);
-                cpi->partition_sz[i+1] = bc[i].pos;
-            }
-        }
+    {
+      int i;
+      for (i = 0; i < num_part; i++) {
+        vp8_stop_encode(&bc[i]);
+        cpi->partition_sz[i + 1] = bc[i].pos;
+      }
+    }
 #endif
 
-        vpx_usec_timer_mark(&emr_timer);
-        cpi->time_encode_mb_row += vpx_usec_timer_elapsed(&emr_timer);
-    }
+    vpx_usec_timer_mark(&emr_timer);
+    cpi->time_encode_mb_row += vpx_usec_timer_elapsed(&emr_timer);
+  }
 
+  // Work out the segment probabilities if segmentation is enabled
+  // and needs to be updated
+  if (xd->segmentation_enabled && xd->update_mb_segmentation_map) {
+    int tot_count;
+    int i;
 
-    // Work out the segment probabilities if segmentation is enabled
-    // and needs to be updated
-    if (xd->segmentation_enabled && xd->update_mb_segmentation_map)
-    {
-        int tot_count;
-        int i;
+    /* Set to defaults */
+    memset(xd->mb_segment_tree_probs, 255, sizeof(xd->mb_segment_tree_probs));
 
-        /* Set to defaults */
-        memset(xd->mb_segment_tree_probs, 255 , sizeof(xd->mb_segment_tree_probs));
+    tot_count = segment_counts[0] + segment_counts[1] + segment_counts[2] +
+                segment_counts[3];
 
-        tot_count = segment_counts[0] + segment_counts[1] + segment_counts[2] + segment_counts[3];
+    if (tot_count) {
+      xd->mb_segment_tree_probs[0] =
+          ((segment_counts[0] + segment_counts[1]) * 255) / tot_count;
 
-        if (tot_count)
-        {
-            xd->mb_segment_tree_probs[0] = ((segment_counts[0] + segment_counts[1]) * 255) / tot_count;
+      tot_count = segment_counts[0] + segment_counts[1];
 
-            tot_count = segment_counts[0] + segment_counts[1];
+      if (tot_count > 0) {
+        xd->mb_segment_tree_probs[1] = (segment_counts[0] * 255) / tot_count;
+      }
 
-            if (tot_count > 0)
-            {
-                xd->mb_segment_tree_probs[1] = (segment_counts[0] * 255) / tot_count;
-            }
+      tot_count = segment_counts[2] + segment_counts[3];
 
-            tot_count = segment_counts[2] + segment_counts[3];
+      if (tot_count > 0)
+        xd->mb_segment_tree_probs[2] = (segment_counts[2] * 255) / tot_count;
 
-            if (tot_count > 0)
-                xd->mb_segment_tree_probs[2] = (segment_counts[2] * 255) / tot_count;
-
-            /* Zero probabilities not allowed */
-            for (i = 0; i < MB_FEATURE_TREE_PROBS; i ++)
-            {
-                if (xd->mb_segment_tree_probs[i] == 0)
-                    xd->mb_segment_tree_probs[i] = 1;
-            }
-        }
+      /* Zero probabilities not allowed */
+      for (i = 0; i < MB_FEATURE_TREE_PROBS; i++) {
+        if (xd->mb_segment_tree_probs[i] == 0) xd->mb_segment_tree_probs[i] = 1;
+      }
     }
+  }
 
-    /* projected_frame_size in units of BYTES */
-    cpi->projected_frame_size = totalrate >> 8;
+  /* projected_frame_size in units of BYTES */
+  cpi->projected_frame_size = totalrate >> 8;
 
-    /* Make a note of the percentage MBs coded Intra. */
-    if (cm->frame_type == KEY_FRAME)
-    {
-        cpi->this_frame_percent_intra = 100;
-    }
-    else
-    {
-        int tot_modes;
+  /* Make a note of the percentage MBs coded Intra. */
+  if (cm->frame_type == KEY_FRAME) {
+    cpi->this_frame_percent_intra = 100;
+  } else {
+    int tot_modes;
 
-        tot_modes = cpi->mb.count_mb_ref_frame_usage[INTRA_FRAME]
-                    + cpi->mb.count_mb_ref_frame_usage[LAST_FRAME]
-                    + cpi->mb.count_mb_ref_frame_usage[GOLDEN_FRAME]
-                    + cpi->mb.count_mb_ref_frame_usage[ALTREF_FRAME];
+    tot_modes = cpi->mb.count_mb_ref_frame_usage[INTRA_FRAME] +
+                cpi->mb.count_mb_ref_frame_usage[LAST_FRAME] +
+                cpi->mb.count_mb_ref_frame_usage[GOLDEN_FRAME] +
+                cpi->mb.count_mb_ref_frame_usage[ALTREF_FRAME];
 
-        if (tot_modes)
-            cpi->this_frame_percent_intra =
-                cpi->mb.count_mb_ref_frame_usage[INTRA_FRAME] * 100 / tot_modes;
+    if (tot_modes)
+      cpi->this_frame_percent_intra =
+          cpi->mb.count_mb_ref_frame_usage[INTRA_FRAME] * 100 / tot_modes;
+  }
 
-    }
-
-#if ! CONFIG_REALTIME_ONLY
-    /* Adjust the projected reference frame usage probability numbers to
-     * reflect what we have just seen. This may be useful when we make
-     * multiple iterations of the recode loop rather than continuing to use
-     * values from the previous frame.
-     */
-    if ((cm->frame_type != KEY_FRAME) && ((cpi->oxcf.number_of_layers > 1) ||
-        (!cm->refresh_alt_ref_frame && !cm->refresh_golden_frame)))
-    {
-      vp8_convert_rfct_to_prob(cpi);
-    }
+#if !CONFIG_REALTIME_ONLY
+  /* Adjust the projected reference frame usage probability numbers to
+   * reflect what we have just seen. This may be useful when we make
+   * multiple iterations of the recode loop rather than continuing to use
+   * values from the previous frame.
+   */
+  if ((cm->frame_type != KEY_FRAME) &&
+      ((cpi->oxcf.number_of_layers > 1) ||
+       (!cm->refresh_alt_ref_frame && !cm->refresh_golden_frame))) {
+    vp8_convert_rfct_to_prob(cpi);
+  }
 #endif
 }
-void vp8_setup_block_ptrs(MACROBLOCK *x)
-{
-    int r, c;
-    int i;
+void vp8_setup_block_ptrs(MACROBLOCK *x) {
+  int r, c;
+  int i;
 
-    for (r = 0; r < 4; r++)
-    {
-        for (c = 0; c < 4; c++)
-        {
-            x->block[r*4+c].src_diff = x->src_diff + r * 4 * 16 + c * 4;
-        }
+  for (r = 0; r < 4; r++) {
+    for (c = 0; c < 4; c++) {
+      x->block[r * 4 + c].src_diff = x->src_diff + r * 4 * 16 + c * 4;
     }
+  }
 
-    for (r = 0; r < 2; r++)
-    {
-        for (c = 0; c < 2; c++)
-        {
-            x->block[16 + r*2+c].src_diff = x->src_diff + 256 + r * 4 * 8 + c * 4;
-        }
+  for (r = 0; r < 2; r++) {
+    for (c = 0; c < 2; c++) {
+      x->block[16 + r * 2 + c].src_diff = x->src_diff + 256 + r * 4 * 8 + c * 4;
     }
+  }
 
-
-    for (r = 0; r < 2; r++)
-    {
-        for (c = 0; c < 2; c++)
-        {
-            x->block[20 + r*2+c].src_diff = x->src_diff + 320 + r * 4 * 8 + c * 4;
-        }
+  for (r = 0; r < 2; r++) {
+    for (c = 0; c < 2; c++) {
+      x->block[20 + r * 2 + c].src_diff = x->src_diff + 320 + r * 4 * 8 + c * 4;
     }
+  }
 
-    x->block[24].src_diff = x->src_diff + 384;
+  x->block[24].src_diff = x->src_diff + 384;
 
-
-    for (i = 0; i < 25; i++)
-    {
-        x->block[i].coeff = x->coeff + i * 16;
-    }
+  for (i = 0; i < 25; i++) {
+    x->block[i].coeff = x->coeff + i * 16;
+  }
 }
 
-void vp8_build_block_offsets(MACROBLOCK *x)
-{
-    int block = 0;
-    int br, bc;
+void vp8_build_block_offsets(MACROBLOCK *x) {
+  int block = 0;
+  int br, bc;
 
-    vp8_build_block_doffsets(&x->e_mbd);
+  vp8_build_block_doffsets(&x->e_mbd);
 
-    /* y blocks */
-    x->thismb_ptr = &x->thismb[0];
-    for (br = 0; br < 4; br++)
-    {
-        for (bc = 0; bc < 4; bc++)
-        {
-            BLOCK *this_block = &x->block[block];
-            this_block->base_src = &x->thismb_ptr;
-            this_block->src_stride = 16;
-            this_block->src = 4 * br * 16 + 4 * bc;
-            ++block;
-        }
+  /* y blocks */
+  x->thismb_ptr = &x->thismb[0];
+  for (br = 0; br < 4; br++) {
+    for (bc = 0; bc < 4; bc++) {
+      BLOCK *this_block = &x->block[block];
+      this_block->base_src = &x->thismb_ptr;
+      this_block->src_stride = 16;
+      this_block->src = 4 * br * 16 + 4 * bc;
+      ++block;
     }
+  }
 
-    /* u blocks */
-    for (br = 0; br < 2; br++)
-    {
-        for (bc = 0; bc < 2; bc++)
-        {
-            BLOCK *this_block = &x->block[block];
-            this_block->base_src = &x->src.u_buffer;
-            this_block->src_stride = x->src.uv_stride;
-            this_block->src = 4 * br * this_block->src_stride + 4 * bc;
-            ++block;
-        }
+  /* u blocks */
+  for (br = 0; br < 2; br++) {
+    for (bc = 0; bc < 2; bc++) {
+      BLOCK *this_block = &x->block[block];
+      this_block->base_src = &x->src.u_buffer;
+      this_block->src_stride = x->src.uv_stride;
+      this_block->src = 4 * br * this_block->src_stride + 4 * bc;
+      ++block;
     }
+  }
 
-    /* v blocks */
-    for (br = 0; br < 2; br++)
-    {
-        for (bc = 0; bc < 2; bc++)
-        {
-            BLOCK *this_block = &x->block[block];
-            this_block->base_src = &x->src.v_buffer;
-            this_block->src_stride = x->src.uv_stride;
-            this_block->src = 4 * br * this_block->src_stride + 4 * bc;
-            ++block;
-        }
+  /* v blocks */
+  for (br = 0; br < 2; br++) {
+    for (bc = 0; bc < 2; bc++) {
+      BLOCK *this_block = &x->block[block];
+      this_block->base_src = &x->src.v_buffer;
+      this_block->src_stride = x->src.uv_stride;
+      this_block->src = 4 * br * this_block->src_stride + 4 * bc;
+      ++block;
     }
+  }
 }
 
-static void sum_intra_stats(VP8_COMP *cpi, MACROBLOCK *x)
-{
-    const MACROBLOCKD *xd = & x->e_mbd;
-    const MB_PREDICTION_MODE m = xd->mode_info_context->mbmi.mode;
-    const MB_PREDICTION_MODE uvm = xd->mode_info_context->mbmi.uv_mode;
+static void sum_intra_stats(VP8_COMP *cpi, MACROBLOCK *x) {
+  const MACROBLOCKD *xd = &x->e_mbd;
+  const MB_PREDICTION_MODE m = xd->mode_info_context->mbmi.mode;
+  const MB_PREDICTION_MODE uvm = xd->mode_info_context->mbmi.uv_mode;
 
 #ifdef MODE_STATS
-    const int is_key = cpi->common.frame_type == KEY_FRAME;
+  const int is_key = cpi->common.frame_type == KEY_FRAME;
 
-    ++ (is_key ? uv_modes : inter_uv_modes)[uvm];
+  ++(is_key ? uv_modes : inter_uv_modes)[uvm];
 
-    if (m == B_PRED)
-    {
-        unsigned int *const bct = is_key ? b_modes : inter_b_modes;
+  if (m == B_PRED) {
+    unsigned int *const bct = is_key ? b_modes : inter_b_modes;
 
-        int b = 0;
+    int b = 0;
 
-        do
-        {
-            ++ bct[xd->block[b].bmi.mode];
-        }
-        while (++b < 16);
-    }
+    do {
+      ++bct[xd->block[b].bmi.mode];
+    } while (++b < 16);
+  }
 
 #else
-    (void)cpi;
+  (void)cpi;
 #endif
 
-    ++x->ymode_count[m];
-    ++x->uv_mode_count[uvm];
-
+  ++x->ymode_count[m];
+  ++x->uv_mode_count[uvm];
 }
 
 /* Experimental stub function to create a per MB zbin adjustment based on
  * some previously calculated measure of MB activity.
  */
-static void adjust_act_zbin( VP8_COMP *cpi, MACROBLOCK *x )
-{
+static void adjust_act_zbin(VP8_COMP *cpi, MACROBLOCK *x) {
 #if USE_ACT_INDEX
-    x->act_zbin_adj = *(x->mb_activity_ptr);
+  x->act_zbin_adj = *(x->mb_activity_ptr);
 #else
-    int64_t a;
-    int64_t b;
-    int64_t act = *(x->mb_activity_ptr);
+  int64_t a;
+  int64_t b;
+  int64_t act = *(x->mb_activity_ptr);
 
-    /* Apply the masking to the RD multiplier. */
-    a = act + 4*cpi->activity_avg;
-    b = 4*act + cpi->activity_avg;
+  /* Apply the masking to the RD multiplier. */
+  a = act + 4 * cpi->activity_avg;
+  b = 4 * act + cpi->activity_avg;
 
-    if ( act > cpi->activity_avg )
-        x->act_zbin_adj = (int)(((int64_t)b + (a>>1))/a) - 1;
-    else
-        x->act_zbin_adj = 1 - (int)(((int64_t)a + (b>>1))/b);
+  if (act > cpi->activity_avg)
+    x->act_zbin_adj = (int)(((int64_t)b + (a >> 1)) / a) - 1;
+  else
+    x->act_zbin_adj = 1 - (int)(((int64_t)a + (b >> 1)) / b);
 #endif
 }
 
 int vp8cx_encode_intra_macroblock(VP8_COMP *cpi, MACROBLOCK *x,
-                                  TOKENEXTRA **t)
-{
-    MACROBLOCKD *xd = &x->e_mbd;
-    int rate;
+                                  TOKENEXTRA **t) {
+  MACROBLOCKD *xd = &x->e_mbd;
+  int rate;
 
-    if (cpi->sf.RD && cpi->compressor_speed != 2)
-        vp8_rd_pick_intra_mode(x, &rate);
-    else
-        vp8_pick_intra_mode(x, &rate);
+  if (cpi->sf.RD && cpi->compressor_speed != 2)
+    vp8_rd_pick_intra_mode(x, &rate);
+  else
+    vp8_pick_intra_mode(x, &rate);
 
-    if(cpi->oxcf.tuning == VP8_TUNE_SSIM)
-    {
-        adjust_act_zbin( cpi, x );
-        vp8_update_zbin_extra(cpi, x);
-    }
+  if (cpi->oxcf.tuning == VP8_TUNE_SSIM) {
+    adjust_act_zbin(cpi, x);
+    vp8_update_zbin_extra(cpi, x);
+  }
 
-    if (x->e_mbd.mode_info_context->mbmi.mode == B_PRED)
-        vp8_encode_intra4x4mby(x);
-    else
-        vp8_encode_intra16x16mby(x);
+  if (x->e_mbd.mode_info_context->mbmi.mode == B_PRED)
+    vp8_encode_intra4x4mby(x);
+  else
+    vp8_encode_intra16x16mby(x);
 
-    vp8_encode_intra16x16mbuv(x);
+  vp8_encode_intra16x16mbuv(x);
 
-    sum_intra_stats(cpi, x);
+  sum_intra_stats(cpi, x);
 
-    vp8_tokenize_mb(cpi, x, t);
+  vp8_tokenize_mb(cpi, x, t);
 
-    if (xd->mode_info_context->mbmi.mode != B_PRED)
-        vp8_inverse_transform_mby(xd);
+  if (xd->mode_info_context->mbmi.mode != B_PRED) vp8_inverse_transform_mby(xd);
 
-    vp8_dequant_idct_add_uv_block
-                    (xd->qcoeff+16*16, xd->dequant_uv,
-                     xd->dst.u_buffer, xd->dst.v_buffer,
-                     xd->dst.uv_stride, xd->eobs+16);
-    return rate;
+  vp8_dequant_idct_add_uv_block(xd->qcoeff + 16 * 16, xd->dequant_uv,
+                                xd->dst.u_buffer, xd->dst.v_buffer,
+                                xd->dst.uv_stride, xd->eobs + 16);
+  return rate;
 }
 #ifdef SPEEDSTATS
 extern int cnt_pm;
@@ -1225,74 +1100,65 @@
 
 extern void vp8_fix_contexts(MACROBLOCKD *x);
 
-int vp8cx_encode_inter_macroblock
-(
-    VP8_COMP *cpi, MACROBLOCK *x, TOKENEXTRA **t,
-    int recon_yoffset, int recon_uvoffset,
-    int mb_row, int mb_col
-)
-{
-    MACROBLOCKD *const xd = &x->e_mbd;
-    int intra_error = 0;
-    int rate;
-    int distortion;
+int vp8cx_encode_inter_macroblock(VP8_COMP *cpi, MACROBLOCK *x, TOKENEXTRA **t,
+                                  int recon_yoffset, int recon_uvoffset,
+                                  int mb_row, int mb_col) {
+  MACROBLOCKD *const xd = &x->e_mbd;
+  int intra_error = 0;
+  int rate;
+  int distortion;
 
-    x->skip = 0;
+  x->skip = 0;
 
-    if (xd->segmentation_enabled)
-        x->encode_breakout = cpi->segment_encode_breakout[xd->mode_info_context->mbmi.segment_id];
-    else
-        x->encode_breakout = cpi->oxcf.encode_breakout;
+  if (xd->segmentation_enabled)
+    x->encode_breakout =
+        cpi->segment_encode_breakout[xd->mode_info_context->mbmi.segment_id];
+  else
+    x->encode_breakout = cpi->oxcf.encode_breakout;
 
 #if CONFIG_TEMPORAL_DENOISING
-    /* Reset the best sse mode/mv for each macroblock. */
-    x->best_reference_frame = INTRA_FRAME;
-    x->best_zeromv_reference_frame = INTRA_FRAME;
-    x->best_sse_inter_mode = 0;
-    x->best_sse_mv.as_int = 0;
-    x->need_to_clamp_best_mvs = 0;
+  /* Reset the best sse mode/mv for each macroblock. */
+  x->best_reference_frame = INTRA_FRAME;
+  x->best_zeromv_reference_frame = INTRA_FRAME;
+  x->best_sse_inter_mode = 0;
+  x->best_sse_mv.as_int = 0;
+  x->need_to_clamp_best_mvs = 0;
 #endif
 
-    if (cpi->sf.RD)
-    {
-        int zbin_mode_boost_enabled = x->zbin_mode_boost_enabled;
+  if (cpi->sf.RD) {
+    int zbin_mode_boost_enabled = x->zbin_mode_boost_enabled;
 
-        /* Are we using the fast quantizer for the mode selection? */
-        if(cpi->sf.use_fastquant_for_pick)
-        {
-            x->quantize_b      = vp8_fast_quantize_b;
+    /* Are we using the fast quantizer for the mode selection? */
+    if (cpi->sf.use_fastquant_for_pick) {
+      x->quantize_b = vp8_fast_quantize_b;
 
-            /* the fast quantizer does not use zbin_extra, so
-             * do not recalculate */
-            x->zbin_mode_boost_enabled = 0;
-        }
-        vp8_rd_pick_inter_mode(cpi, x, recon_yoffset, recon_uvoffset, &rate,
-                               &distortion, &intra_error, mb_row, mb_col);
+      /* the fast quantizer does not use zbin_extra, so
+       * do not recalculate */
+      x->zbin_mode_boost_enabled = 0;
+    }
+    vp8_rd_pick_inter_mode(cpi, x, recon_yoffset, recon_uvoffset, &rate,
+                           &distortion, &intra_error, mb_row, mb_col);
 
-        /* switch back to the regular quantizer for the encode */
-        if (cpi->sf.improved_quant)
-        {
-            x->quantize_b      = vp8_regular_quantize_b;
-        }
+    /* switch back to the regular quantizer for the encode */
+    if (cpi->sf.improved_quant) {
+      x->quantize_b = vp8_regular_quantize_b;
+    }
 
-        /* restore cpi->zbin_mode_boost_enabled */
-        x->zbin_mode_boost_enabled = zbin_mode_boost_enabled;
+    /* restore cpi->zbin_mode_boost_enabled */
+    x->zbin_mode_boost_enabled = zbin_mode_boost_enabled;
 
-    }
-    else
-    {
-        vp8_pick_inter_mode(cpi, x, recon_yoffset, recon_uvoffset, &rate,
-                            &distortion, &intra_error, mb_row, mb_col);
-    }
+  } else {
+    vp8_pick_inter_mode(cpi, x, recon_yoffset, recon_uvoffset, &rate,
+                        &distortion, &intra_error, mb_row, mb_col);
+  }
 
-    x->prediction_error += distortion;
-    x->intra_error += intra_error;
+  x->prediction_error += distortion;
+  x->intra_error += intra_error;
 
-    if(cpi->oxcf.tuning == VP8_TUNE_SSIM)
-    {
-        /* Adjust the zbin based on this MB rate. */
-        adjust_act_zbin( cpi, x );
-    }
+  if (cpi->oxcf.tuning == VP8_TUNE_SSIM) {
+    /* Adjust the zbin based on this MB rate. */
+    adjust_act_zbin(cpi, x);
+  }
 
 #if 0
     /* Experimental RD code */
@@ -1300,125 +1166,104 @@
     cpi->last_mb_distortion = distortion;
 #endif
 
-    /* MB level adjutment to quantizer setup */
-    if (xd->segmentation_enabled)
-    {
-        /* If cyclic update enabled */
-        if (cpi->current_layer == 0 && cpi->cyclic_refresh_mode_enabled)
-        {
-            /* Clear segment_id back to 0 if not coded (last frame 0,0) */
-            if ((xd->mode_info_context->mbmi.segment_id == 1) &&
-                ((xd->mode_info_context->mbmi.ref_frame != LAST_FRAME) || (xd->mode_info_context->mbmi.mode != ZEROMV)))
-            {
-                xd->mode_info_context->mbmi.segment_id = 0;
+  /* MB level adjutment to quantizer setup */
+  if (xd->segmentation_enabled) {
+    /* If cyclic update enabled */
+    if (cpi->current_layer == 0 && cpi->cyclic_refresh_mode_enabled) {
+      /* Clear segment_id back to 0 if not coded (last frame 0,0) */
+      if ((xd->mode_info_context->mbmi.segment_id == 1) &&
+          ((xd->mode_info_context->mbmi.ref_frame != LAST_FRAME) ||
+           (xd->mode_info_context->mbmi.mode != ZEROMV))) {
+        xd->mode_info_context->mbmi.segment_id = 0;
 
-                /* segment_id changed, so update */
-                vp8cx_mb_init_quantizer(cpi, x, 1);
-            }
-        }
+        /* segment_id changed, so update */
+        vp8cx_mb_init_quantizer(cpi, x, 1);
+      }
     }
+  }
 
-    {
-        /* Experimental code.
-         * Special case for gf and arf zeromv modes, for 1 temporal layer.
-         * Increase zbin size to supress noise.
-         */
-        x->zbin_mode_boost = 0;
-        if (x->zbin_mode_boost_enabled)
-        {
-            if ( xd->mode_info_context->mbmi.ref_frame != INTRA_FRAME )
-            {
-                if (xd->mode_info_context->mbmi.mode == ZEROMV)
-                {
-                    if (xd->mode_info_context->mbmi.ref_frame != LAST_FRAME &&
-                        cpi->oxcf.number_of_layers == 1)
-                        x->zbin_mode_boost = GF_ZEROMV_ZBIN_BOOST;
-                    else
-                        x->zbin_mode_boost = LF_ZEROMV_ZBIN_BOOST;
-                }
-                else if (xd->mode_info_context->mbmi.mode == SPLITMV)
-                    x->zbin_mode_boost = 0;
-                else
-                    x->zbin_mode_boost = MV_ZBIN_BOOST;
-            }
-        }
-
-        /* The fast quantizer doesn't use zbin_extra, only do so with
-         * the regular quantizer. */
-        if (cpi->sf.improved_quant)
-            vp8_update_zbin_extra(cpi, x);
+  {
+    /* Experimental code.
+     * Special case for gf and arf zeromv modes, for 1 temporal layer.
+     * Increase zbin size to supress noise.
+     */
+    x->zbin_mode_boost = 0;
+    if (x->zbin_mode_boost_enabled) {
+      if (xd->mode_info_context->mbmi.ref_frame != INTRA_FRAME) {
+        if (xd->mode_info_context->mbmi.mode == ZEROMV) {
+          if (xd->mode_info_context->mbmi.ref_frame != LAST_FRAME &&
+              cpi->oxcf.number_of_layers == 1)
+            x->zbin_mode_boost = GF_ZEROMV_ZBIN_BOOST;
+          else
+            x->zbin_mode_boost = LF_ZEROMV_ZBIN_BOOST;
+        } else if (xd->mode_info_context->mbmi.mode == SPLITMV)
+          x->zbin_mode_boost = 0;
+        else
+          x->zbin_mode_boost = MV_ZBIN_BOOST;
+      }
     }
 
-    x->count_mb_ref_frame_usage[xd->mode_info_context->mbmi.ref_frame] ++;
+    /* The fast quantizer doesn't use zbin_extra, only do so with
+     * the regular quantizer. */
+    if (cpi->sf.improved_quant) vp8_update_zbin_extra(cpi, x);
+  }
 
-    if (xd->mode_info_context->mbmi.ref_frame == INTRA_FRAME)
-    {
-        vp8_encode_intra16x16mbuv(x);
+  x->count_mb_ref_frame_usage[xd->mode_info_context->mbmi.ref_frame]++;
 
-        if (xd->mode_info_context->mbmi.mode == B_PRED)
-        {
-            vp8_encode_intra4x4mby(x);
-        }
-        else
-        {
-            vp8_encode_intra16x16mby(x);
-        }
+  if (xd->mode_info_context->mbmi.ref_frame == INTRA_FRAME) {
+    vp8_encode_intra16x16mbuv(x);
 
-        sum_intra_stats(cpi, x);
+    if (xd->mode_info_context->mbmi.mode == B_PRED) {
+      vp8_encode_intra4x4mby(x);
+    } else {
+      vp8_encode_intra16x16mby(x);
     }
-    else
-    {
-        int ref_fb_idx;
 
-        if (xd->mode_info_context->mbmi.ref_frame == LAST_FRAME)
-            ref_fb_idx = cpi->common.lst_fb_idx;
-        else if (xd->mode_info_context->mbmi.ref_frame == GOLDEN_FRAME)
-            ref_fb_idx = cpi->common.gld_fb_idx;
-        else
-            ref_fb_idx = cpi->common.alt_fb_idx;
+    sum_intra_stats(cpi, x);
+  } else {
+    int ref_fb_idx;
 
-        xd->pre.y_buffer = cpi->common.yv12_fb[ref_fb_idx].y_buffer + recon_yoffset;
-        xd->pre.u_buffer = cpi->common.yv12_fb[ref_fb_idx].u_buffer + recon_uvoffset;
-        xd->pre.v_buffer = cpi->common.yv12_fb[ref_fb_idx].v_buffer + recon_uvoffset;
+    if (xd->mode_info_context->mbmi.ref_frame == LAST_FRAME)
+      ref_fb_idx = cpi->common.lst_fb_idx;
+    else if (xd->mode_info_context->mbmi.ref_frame == GOLDEN_FRAME)
+      ref_fb_idx = cpi->common.gld_fb_idx;
+    else
+      ref_fb_idx = cpi->common.alt_fb_idx;
 
-        if (!x->skip)
-        {
-            vp8_encode_inter16x16(x);
-        }
-        else
-            vp8_build_inter16x16_predictors_mb(xd, xd->dst.y_buffer,
-                                           xd->dst.u_buffer, xd->dst.v_buffer,
-                                           xd->dst.y_stride, xd->dst.uv_stride);
+    xd->pre.y_buffer = cpi->common.yv12_fb[ref_fb_idx].y_buffer + recon_yoffset;
+    xd->pre.u_buffer =
+        cpi->common.yv12_fb[ref_fb_idx].u_buffer + recon_uvoffset;
+    xd->pre.v_buffer =
+        cpi->common.yv12_fb[ref_fb_idx].v_buffer + recon_uvoffset;
 
-    }
+    if (!x->skip) {
+      vp8_encode_inter16x16(x);
+    } else
+      vp8_build_inter16x16_predictors_mb(xd, xd->dst.y_buffer, xd->dst.u_buffer,
+                                         xd->dst.v_buffer, xd->dst.y_stride,
+                                         xd->dst.uv_stride);
+  }
 
-    if (!x->skip)
-    {
-        vp8_tokenize_mb(cpi, x, t);
+  if (!x->skip) {
+    vp8_tokenize_mb(cpi, x, t);
 
-        if (xd->mode_info_context->mbmi.mode != B_PRED)
-            vp8_inverse_transform_mby(xd);
+    if (xd->mode_info_context->mbmi.mode != B_PRED)
+      vp8_inverse_transform_mby(xd);
 
-        vp8_dequant_idct_add_uv_block
-                        (xd->qcoeff+16*16, xd->dequant_uv,
-                         xd->dst.u_buffer, xd->dst.v_buffer,
-                         xd->dst.uv_stride, xd->eobs+16);
-    }
-    else
-    {
-        /* always set mb_skip_coeff as it is needed by the loopfilter */
-        xd->mode_info_context->mbmi.mb_skip_coeff = 1;
+    vp8_dequant_idct_add_uv_block(xd->qcoeff + 16 * 16, xd->dequant_uv,
+                                  xd->dst.u_buffer, xd->dst.v_buffer,
+                                  xd->dst.uv_stride, xd->eobs + 16);
+  } else {
+    /* always set mb_skip_coeff as it is needed by the loopfilter */
+    xd->mode_info_context->mbmi.mb_skip_coeff = 1;
 
-        if (cpi->common.mb_no_coeff_skip)
-        {
-            x->skip_true_count ++;
-            vp8_fix_contexts(xd);
-        }
-        else
-        {
-            vp8_stuff_mb(cpi, x, t);
-        }
+    if (cpi->common.mb_no_coeff_skip) {
+      x->skip_true_count++;
+      vp8_fix_contexts(xd);
+    } else {
+      vp8_stuff_mb(cpi, x, t);
     }
+  }
 
-    return rate;
+  return rate;
 }
--- a/vp8/encoder/encodeframe.h
+++ b/vp8/encoder/encodeframe.h
@@ -22,12 +22,12 @@
 extern void vp8_encode_frame(VP8_COMP *cpi);
 
 extern int vp8cx_encode_inter_macroblock(VP8_COMP *cpi, MACROBLOCK *x,
-        TOKENEXTRA **t,
-        int recon_yoffset, int recon_uvoffset,
-        int mb_row, int mb_col);
+                                         TOKENEXTRA **t, int recon_yoffset,
+                                         int recon_uvoffset, int mb_row,
+                                         int mb_col);
 
 extern int vp8cx_encode_intra_macroblock(VP8_COMP *cpi, MACROBLOCK *x,
-        TOKENEXTRA **t);
+                                         TOKENEXTRA **t);
 #ifdef __cplusplus
 }  // extern "C"
 #endif
--- a/vp8/encoder/encodeintra.c
+++ b/vp8/encoder/encodeintra.c
@@ -8,7 +8,6 @@
  *  be found in the AUTHORS file in the root of the source tree.
  */
 
-
 #include "vpx_config.h"
 #include "vp8_rtcd.h"
 #include "./vpx_dsp_rtcd.h"
@@ -19,122 +18,100 @@
 #include "vp8/common/invtrans.h"
 #include "encodeintra.h"
 
+int vp8_encode_intra(VP8_COMP *cpi, MACROBLOCK *x, int use_dc_pred) {
+  int i;
+  int intra_pred_var = 0;
+  (void)cpi;
 
-int vp8_encode_intra(VP8_COMP *cpi, MACROBLOCK *x, int use_dc_pred)
-{
+  if (use_dc_pred) {
+    x->e_mbd.mode_info_context->mbmi.mode = DC_PRED;
+    x->e_mbd.mode_info_context->mbmi.uv_mode = DC_PRED;
+    x->e_mbd.mode_info_context->mbmi.ref_frame = INTRA_FRAME;
 
-    int i;
-    int intra_pred_var = 0;
-    (void) cpi;
+    vp8_encode_intra16x16mby(x);
 
-    if (use_dc_pred)
-    {
-        x->e_mbd.mode_info_context->mbmi.mode = DC_PRED;
-        x->e_mbd.mode_info_context->mbmi.uv_mode = DC_PRED;
-        x->e_mbd.mode_info_context->mbmi.ref_frame = INTRA_FRAME;
-
-        vp8_encode_intra16x16mby(x);
-
-        vp8_inverse_transform_mby(&x->e_mbd);
+    vp8_inverse_transform_mby(&x->e_mbd);
+  } else {
+    for (i = 0; i < 16; i++) {
+      x->e_mbd.block[i].bmi.as_mode = B_DC_PRED;
+      vp8_encode_intra4x4block(x, i);
     }
-    else
-    {
-        for (i = 0; i < 16; i++)
-        {
-            x->e_mbd.block[i].bmi.as_mode = B_DC_PRED;
-            vp8_encode_intra4x4block(x, i);
-        }
-    }
+  }
 
-    intra_pred_var = vpx_get_mb_ss(x->src_diff);
+  intra_pred_var = vpx_get_mb_ss(x->src_diff);
 
-    return intra_pred_var;
+  return intra_pred_var;
 }
 
-void vp8_encode_intra4x4block(MACROBLOCK *x, int ib)
-{
-    BLOCKD *b = &x->e_mbd.block[ib];
-    BLOCK *be = &x->block[ib];
-    int dst_stride = x->e_mbd.dst.y_stride;
-    unsigned char *dst = x->e_mbd.dst.y_buffer + b->offset;
-    unsigned char *Above = dst - dst_stride;
-    unsigned char *yleft = dst - 1;
-    unsigned char top_left = Above[-1];
+void vp8_encode_intra4x4block(MACROBLOCK *x, int ib) {
+  BLOCKD *b = &x->e_mbd.block[ib];
+  BLOCK *be = &x->block[ib];
+  int dst_stride = x->e_mbd.dst.y_stride;
+  unsigned char *dst = x->e_mbd.dst.y_buffer + b->offset;
+  unsigned char *Above = dst - dst_stride;
+  unsigned char *yleft = dst - 1;
+  unsigned char top_left = Above[-1];
 
-    vp8_intra4x4_predict(Above, yleft, dst_stride, b->bmi.as_mode,
-                         b->predictor, 16, top_left);
+  vp8_intra4x4_predict(Above, yleft, dst_stride, b->bmi.as_mode, b->predictor,
+                       16, top_left);
 
-    vp8_subtract_b(be, b, 16);
+  vp8_subtract_b(be, b, 16);
 
-    x->short_fdct4x4(be->src_diff, be->coeff, 32);
+  x->short_fdct4x4(be->src_diff, be->coeff, 32);
 
-    x->quantize_b(be, b);
+  x->quantize_b(be, b);
 
-    if (*b->eob > 1)
-    {
-      vp8_short_idct4x4llm(b->dqcoeff, b->predictor, 16, dst, dst_stride);
-    }
-    else
-    {
-      vp8_dc_only_idct_add(b->dqcoeff[0], b->predictor, 16, dst, dst_stride);
-    }
+  if (*b->eob > 1) {
+    vp8_short_idct4x4llm(b->dqcoeff, b->predictor, 16, dst, dst_stride);
+  } else {
+    vp8_dc_only_idct_add(b->dqcoeff[0], b->predictor, 16, dst, dst_stride);
+  }
 }
 
-void vp8_encode_intra4x4mby(MACROBLOCK *mb)
-{
-    int i;
+void vp8_encode_intra4x4mby(MACROBLOCK *mb) {
+  int i;
 
-    MACROBLOCKD *xd = &mb->e_mbd;
-    intra_prediction_down_copy(xd, xd->dst.y_buffer - xd->dst.y_stride + 16);
+  MACROBLOCKD *xd = &mb->e_mbd;
+  intra_prediction_down_copy(xd, xd->dst.y_buffer - xd->dst.y_stride + 16);
 
-    for (i = 0; i < 16; i++)
-        vp8_encode_intra4x4block(mb, i);
-    return;
+  for (i = 0; i < 16; i++) vp8_encode_intra4x4block(mb, i);
+  return;
 }
 
-void vp8_encode_intra16x16mby(MACROBLOCK *x)
-{
-    BLOCK *b = &x->block[0];
-    MACROBLOCKD *xd = &x->e_mbd;
+void vp8_encode_intra16x16mby(MACROBLOCK *x) {
+  BLOCK *b = &x->block[0];
+  MACROBLOCKD *xd = &x->e_mbd;
 
-    vp8_build_intra_predictors_mby_s(xd,
-                                         xd->dst.y_buffer - xd->dst.y_stride,
-                                         xd->dst.y_buffer - 1,
-                                         xd->dst.y_stride,
-                                         xd->dst.y_buffer,
-                                         xd->dst.y_stride);
+  vp8_build_intra_predictors_mby_s(xd, xd->dst.y_buffer - xd->dst.y_stride,
+                                   xd->dst.y_buffer - 1, xd->dst.y_stride,
+                                   xd->dst.y_buffer, xd->dst.y_stride);
 
-    vp8_subtract_mby(x->src_diff, *(b->base_src),
-        b->src_stride, xd->dst.y_buffer, xd->dst.y_stride);
+  vp8_subtract_mby(x->src_diff, *(b->base_src), b->src_stride, xd->dst.y_buffer,
+                   xd->dst.y_stride);
 
-    vp8_transform_intra_mby(x);
+  vp8_transform_intra_mby(x);
 
-    vp8_quantize_mby(x);
+  vp8_quantize_mby(x);
 
-    if (x->optimize)
-        vp8_optimize_mby(x);
+  if (x->optimize) vp8_optimize_mby(x);
 }
 
-void vp8_encode_intra16x16mbuv(MACROBLOCK *x)
-{
-    MACROBLOCKD *xd = &x->e_mbd;
+void vp8_encode_intra16x16mbuv(MACROBLOCK *x) {
+  MACROBLOCKD *xd = &x->e_mbd;
 
-    vp8_build_intra_predictors_mbuv_s(xd, xd->dst.u_buffer - xd->dst.uv_stride,
-                                      xd->dst.v_buffer - xd->dst.uv_stride,
-                                      xd->dst.u_buffer - 1,
-                                      xd->dst.v_buffer - 1,
-                                      xd->dst.uv_stride,
-                                      xd->dst.u_buffer, xd->dst.v_buffer,
-                                      xd->dst.uv_stride);
+  vp8_build_intra_predictors_mbuv_s(xd, xd->dst.u_buffer - xd->dst.uv_stride,
+                                    xd->dst.v_buffer - xd->dst.uv_stride,
+                                    xd->dst.u_buffer - 1, xd->dst.v_buffer - 1,
+                                    xd->dst.uv_stride, xd->dst.u_buffer,
+                                    xd->dst.v_buffer, xd->dst.uv_stride);
 
-    vp8_subtract_mbuv(x->src_diff, x->src.u_buffer,
-        x->src.v_buffer, x->src.uv_stride, xd->dst.u_buffer,
-        xd->dst.v_buffer, xd->dst.uv_stride);
+  vp8_subtract_mbuv(x->src_diff, x->src.u_buffer, x->src.v_buffer,
+                    x->src.uv_stride, xd->dst.u_buffer, xd->dst.v_buffer,
+                    xd->dst.uv_stride);
 
-    vp8_transform_mbuv(x);
+  vp8_transform_mbuv(x);
 
-    vp8_quantize_mbuv(x);
+  vp8_quantize_mbuv(x);
 
-    if (x->optimize)
-        vp8_optimize_mbuv(x);
+  if (x->optimize) vp8_optimize_mbuv(x);
 }
--- a/vp8/encoder/encodeintra.h
+++ b/vp8/encoder/encodeintra.h
@@ -8,7 +8,6 @@
  *  be found in the AUTHORS file in the root of the source tree.
  */
 
-
 #ifndef VP8_ENCODER_ENCODEINTRA_H_
 #define VP8_ENCODER_ENCODEINTRA_H_
 #include "onyx_int.h"
--- a/vp8/encoder/encodemb.c
+++ b/vp8/encoder/encodemb.c
@@ -26,13 +26,13 @@
   unsigned char *pred_ptr = bd->predictor;
   int src_stride = be->src_stride;
 
-  vpx_subtract_block(4, 4, diff_ptr, pitch, src_ptr, src_stride,
-                     pred_ptr, pitch);
+  vpx_subtract_block(4, 4, diff_ptr, pitch, src_ptr, src_stride, pred_ptr,
+                     pitch);
 }
 
 void vp8_subtract_mbuv(short *diff, unsigned char *usrc, unsigned char *vsrc,
-                         int src_stride, unsigned char *upred,
-                         unsigned char *vpred, int pred_stride) {
+                       int src_stride, unsigned char *upred,
+                       unsigned char *vpred, int pred_stride) {
   short *udiff = diff + 256;
   short *vdiff = diff + 320;
 
@@ -45,119 +45,90 @@
   vpx_subtract_block(16, 16, diff, 16, src, src_stride, pred, pred_stride);
 }
 
-static void vp8_subtract_mb(MACROBLOCK *x)
-{
-    BLOCK *b = &x->block[0];
+static void vp8_subtract_mb(MACROBLOCK *x) {
+  BLOCK *b = &x->block[0];
 
-    vp8_subtract_mby(x->src_diff, *(b->base_src),
-        b->src_stride, x->e_mbd.dst.y_buffer, x->e_mbd.dst.y_stride);
-    vp8_subtract_mbuv(x->src_diff, x->src.u_buffer,
-        x->src.v_buffer, x->src.uv_stride, x->e_mbd.dst.u_buffer,
-        x->e_mbd.dst.v_buffer, x->e_mbd.dst.uv_stride);
+  vp8_subtract_mby(x->src_diff, *(b->base_src), b->src_stride,
+                   x->e_mbd.dst.y_buffer, x->e_mbd.dst.y_stride);
+  vp8_subtract_mbuv(x->src_diff, x->src.u_buffer, x->src.v_buffer,
+                    x->src.uv_stride, x->e_mbd.dst.u_buffer,
+                    x->e_mbd.dst.v_buffer, x->e_mbd.dst.uv_stride);
 }
 
-static void build_dcblock(MACROBLOCK *x)
-{
-    short *src_diff_ptr = &x->src_diff[384];
-    int i;
+static void build_dcblock(MACROBLOCK *x) {
+  short *src_diff_ptr = &x->src_diff[384];
+  int i;
 
-    for (i = 0; i < 16; i++)
-    {
-        src_diff_ptr[i] = x->coeff[i * 16];
-    }
+  for (i = 0; i < 16; i++) {
+    src_diff_ptr[i] = x->coeff[i * 16];
+  }
 }
 
-void vp8_transform_mbuv(MACROBLOCK *x)
-{
-    int i;
+void vp8_transform_mbuv(MACROBLOCK *x) {
+  int i;
 
-    for (i = 16; i < 24; i += 2)
-    {
-        x->short_fdct8x4(&x->block[i].src_diff[0],
-            &x->block[i].coeff[0], 16);
-    }
+  for (i = 16; i < 24; i += 2) {
+    x->short_fdct8x4(&x->block[i].src_diff[0], &x->block[i].coeff[0], 16);
+  }
 }
 
+void vp8_transform_intra_mby(MACROBLOCK *x) {
+  int i;
 
-void vp8_transform_intra_mby(MACROBLOCK *x)
-{
-    int i;
+  for (i = 0; i < 16; i += 2) {
+    x->short_fdct8x4(&x->block[i].src_diff[0], &x->block[i].coeff[0], 32);
+  }
 
-    for (i = 0; i < 16; i += 2)
-    {
-        x->short_fdct8x4(&x->block[i].src_diff[0],
-            &x->block[i].coeff[0], 32);
-    }
+  /* build dc block from 16 y dc values */
+  build_dcblock(x);
 
-    /* build dc block from 16 y dc values */
-    build_dcblock(x);
-
-    /* do 2nd order transform on the dc block */
-    x->short_walsh4x4(&x->block[24].src_diff[0],
-        &x->block[24].coeff[0], 8);
-
+  /* do 2nd order transform on the dc block */
+  x->short_walsh4x4(&x->block[24].src_diff[0], &x->block[24].coeff[0], 8);
 }
 
+static void transform_mb(MACROBLOCK *x) {
+  int i;
 
-static void transform_mb(MACROBLOCK *x)
-{
-    int i;
+  for (i = 0; i < 16; i += 2) {
+    x->short_fdct8x4(&x->block[i].src_diff[0], &x->block[i].coeff[0], 32);
+  }
 
-    for (i = 0; i < 16; i += 2)
-    {
-        x->short_fdct8x4(&x->block[i].src_diff[0],
-            &x->block[i].coeff[0], 32);
-    }
+  /* build dc block from 16 y dc values */
+  if (x->e_mbd.mode_info_context->mbmi.mode != SPLITMV) build_dcblock(x);
 
-    /* build dc block from 16 y dc values */
-    if (x->e_mbd.mode_info_context->mbmi.mode != SPLITMV)
-        build_dcblock(x);
+  for (i = 16; i < 24; i += 2) {
+    x->short_fdct8x4(&x->block[i].src_diff[0], &x->block[i].coeff[0], 16);
+  }
 
-    for (i = 16; i < 24; i += 2)
-    {
-        x->short_fdct8x4(&x->block[i].src_diff[0],
-            &x->block[i].coeff[0], 16);
-    }
-
-    /* do 2nd order transform on the dc block */
-    if (x->e_mbd.mode_info_context->mbmi.mode != SPLITMV)
-        x->short_walsh4x4(&x->block[24].src_diff[0],
-        &x->block[24].coeff[0], 8);
-
+  /* do 2nd order transform on the dc block */
+  if (x->e_mbd.mode_info_context->mbmi.mode != SPLITMV)
+    x->short_walsh4x4(&x->block[24].src_diff[0], &x->block[24].coeff[0], 8);
 }
 
+static void transform_mby(MACROBLOCK *x) {
+  int i;
 
-static void transform_mby(MACROBLOCK *x)
-{
-    int i;
+  for (i = 0; i < 16; i += 2) {
+    x->short_fdct8x4(&x->block[i].src_diff[0], &x->block[i].coeff[0], 32);
+  }
 
-    for (i = 0; i < 16; i += 2)
-    {
-        x->short_fdct8x4(&x->block[i].src_diff[0],
-            &x->block[i].coeff[0], 32);
-    }
-
-    /* build dc block from 16 y dc values */
-    if (x->e_mbd.mode_info_context->mbmi.mode != SPLITMV)
-    {
-        build_dcblock(x);
-        x->short_walsh4x4(&x->block[24].src_diff[0],
-            &x->block[24].coeff[0], 8);
-    }
+  /* build dc block from 16 y dc values */
+  if (x->e_mbd.mode_info_context->mbmi.mode != SPLITMV) {
+    build_dcblock(x);
+    x->short_walsh4x4(&x->block[24].src_diff[0], &x->block[24].coeff[0], 8);
+  }
 }
 
+#define RDTRUNC(RM, DM, R, D) ((128 + (R) * (RM)) & 0xFF)
 
-
-#define RDTRUNC(RM,DM,R,D) ( (128+(R)*(RM)) & 0xFF )
-
 typedef struct vp8_token_state vp8_token_state;
 
-struct vp8_token_state{
-  int           rate;
-  int           error;
-  signed char   next;
-  signed char   token;
-  short         qc;
+struct vp8_token_state {
+  int rate;
+  int error;
+  signed char next;
+  signed char token;
+  short qc;
 };
 
 /* TODO: experiments to find optimal multiple numbers */
@@ -165,429 +136,374 @@
 #define UV_RD_MULT 2
 #define Y2_RD_MULT 16
 
-static const int plane_rd_mult[4]=
-{
-    Y1_RD_MULT,
-    Y2_RD_MULT,
-    UV_RD_MULT,
-    Y1_RD_MULT
-};
+static const int plane_rd_mult[4] = { Y1_RD_MULT, Y2_RD_MULT, UV_RD_MULT,
+                                      Y1_RD_MULT };
 
-static void optimize_b(MACROBLOCK *mb, int ib, int type,
-                       ENTROPY_CONTEXT *a, ENTROPY_CONTEXT *l)
-{
-    BLOCK *b;
-    BLOCKD *d;
-    vp8_token_state tokens[17][2];
-    unsigned best_mask[2];
-    const short *dequant_ptr;
-    const short *coeff_ptr;
-    short *qcoeff_ptr;
-    short *dqcoeff_ptr;
-    int eob;
-    int i0;
-    int rc;
-    int x;
-    int sz = 0;
-    int next;
-    int rdmult;
-    int rddiv;
-    int final_eob;
-    int rd_cost0;
-    int rd_cost1;
-    int rate0;
-    int rate1;
-    int error0;
-    int error1;
-    int t0;
-    int t1;
-    int best;
-    int band;
-    int pt;
-    int i;
-    int err_mult = plane_rd_mult[type];
+static void optimize_b(MACROBLOCK *mb, int ib, int type, ENTROPY_CONTEXT *a,
+                       ENTROPY_CONTEXT *l) {
+  BLOCK *b;
+  BLOCKD *d;
+  vp8_token_state tokens[17][2];
+  unsigned best_mask[2];
+  const short *dequant_ptr;
+  const short *coeff_ptr;
+  short *qcoeff_ptr;
+  short *dqcoeff_ptr;
+  int eob;
+  int i0;
+  int rc;
+  int x;
+  int sz = 0;
+  int next;
+  int rdmult;
+  int rddiv;
+  int final_eob;
+  int rd_cost0;
+  int rd_cost1;
+  int rate0;
+  int rate1;
+  int error0;
+  int error1;
+  int t0;
+  int t1;
+  int best;
+  int band;
+  int pt;
+  int i;
+  int err_mult = plane_rd_mult[type];
 
-    b = &mb->block[ib];
-    d = &mb->e_mbd.block[ib];
+  b = &mb->block[ib];
+  d = &mb->e_mbd.block[ib];
 
-    dequant_ptr = d->dequant;
-    coeff_ptr = b->coeff;
-    qcoeff_ptr = d->qcoeff;
-    dqcoeff_ptr = d->dqcoeff;
-    i0 = !type;
-    eob = *d->eob;
+  dequant_ptr = d->dequant;
+  coeff_ptr = b->coeff;
+  qcoeff_ptr = d->qcoeff;
+  dqcoeff_ptr = d->dqcoeff;
+  i0 = !type;
+  eob = *d->eob;
 
-    /* Now set up a Viterbi trellis to evaluate alternative roundings. */
-    rdmult = mb->rdmult * err_mult;
-    if(mb->e_mbd.mode_info_context->mbmi.ref_frame==INTRA_FRAME)
-        rdmult = (rdmult * 9)>>4;
+  /* Now set up a Viterbi trellis to evaluate alternative roundings. */
+  rdmult = mb->rdmult * err_mult;
+  if (mb->e_mbd.mode_info_context->mbmi.ref_frame == INTRA_FRAME)
+    rdmult = (rdmult * 9) >> 4;
 
-    rddiv = mb->rddiv;
-    best_mask[0] = best_mask[1] = 0;
-    /* Initialize the sentinel node of the trellis. */
-    tokens[eob][0].rate = 0;
-    tokens[eob][0].error = 0;
-    tokens[eob][0].next = 16;
-    tokens[eob][0].token = DCT_EOB_TOKEN;
-    tokens[eob][0].qc = 0;
-    *(tokens[eob] + 1) = *(tokens[eob] + 0);
-    next = eob;
-    for (i = eob; i-- > i0;)
-    {
-        int base_bits;
-        int d2;
-        int dx;
+  rddiv = mb->rddiv;
+  best_mask[0] = best_mask[1] = 0;
+  /* Initialize the sentinel node of the trellis. */
+  tokens[eob][0].rate = 0;
+  tokens[eob][0].error = 0;
+  tokens[eob][0].next = 16;
+  tokens[eob][0].token = DCT_EOB_TOKEN;
+  tokens[eob][0].qc = 0;
+  *(tokens[eob] + 1) = *(tokens[eob] + 0);
+  next = eob;
+  for (i = eob; i-- > i0;) {
+    int base_bits;
+    int d2;
+    int dx;
 
-        rc = vp8_default_zig_zag1d[i];
-        x = qcoeff_ptr[rc];
-        /* Only add a trellis state for non-zero coefficients. */
-        if (x)
-        {
-            int shortcut=0;
-            error0 = tokens[next][0].error;
-            error1 = tokens[next][1].error;
-            /* Evaluate the first possibility for this state. */
-            rate0 = tokens[next][0].rate;
-            rate1 = tokens[next][1].rate;
-            t0 = (vp8_dct_value_tokens_ptr + x)->Token;
-            /* Consider both possible successor states. */
-            if (next < 16)
-            {
-                band = vp8_coef_bands[i + 1];
-                pt = vp8_prev_token_class[t0];
-                rate0 +=
-                    mb->token_costs[type][band][pt][tokens[next][0].token];
-                rate1 +=
-                    mb->token_costs[type][band][pt][tokens[next][1].token];
-            }
-            rd_cost0 = RDCOST(rdmult, rddiv, rate0, error0);
-            rd_cost1 = RDCOST(rdmult, rddiv, rate1, error1);
-            if (rd_cost0 == rd_cost1)
-            {
-                rd_cost0 = RDTRUNC(rdmult, rddiv, rate0, error0);
-                rd_cost1 = RDTRUNC(rdmult, rddiv, rate1, error1);
-            }
-            /* And pick the best. */
-            best = rd_cost1 < rd_cost0;
-            base_bits = *(vp8_dct_value_cost_ptr + x);
-            dx = dqcoeff_ptr[rc] - coeff_ptr[rc];
-            d2 = dx*dx;
-            tokens[i][0].rate = base_bits + (best ? rate1 : rate0);
-            tokens[i][0].error = d2 + (best ? error1 : error0);
-            tokens[i][0].next = next;
-            tokens[i][0].token = t0;
-            tokens[i][0].qc = x;
-            best_mask[0] |= best << i;
-            /* Evaluate the second possibility for this state. */
-            rate0 = tokens[next][0].rate;
-            rate1 = tokens[next][1].rate;
+    rc = vp8_default_zig_zag1d[i];
+    x = qcoeff_ptr[rc];
+    /* Only add a trellis state for non-zero coefficients. */
+    if (x) {
+      int shortcut = 0;
+      error0 = tokens[next][0].error;
+      error1 = tokens[next][1].error;
+      /* Evaluate the first possibility for this state. */
+      rate0 = tokens[next][0].rate;
+      rate1 = tokens[next][1].rate;
+      t0 = (vp8_dct_value_tokens_ptr + x)->Token;
+      /* Consider both possible successor states. */
+      if (next < 16) {
+        band = vp8_coef_bands[i + 1];
+        pt = vp8_prev_token_class[t0];
+        rate0 += mb->token_costs[type][band][pt][tokens[next][0].token];
+        rate1 += mb->token_costs[type][band][pt][tokens[next][1].token];
+      }
+      rd_cost0 = RDCOST(rdmult, rddiv, rate0, error0);
+      rd_cost1 = RDCOST(rdmult, rddiv, rate1, error1);
+      if (rd_cost0 == rd_cost1) {
+        rd_cost0 = RDTRUNC(rdmult, rddiv, rate0, error0);
+        rd_cost1 = RDTRUNC(rdmult, rddiv, rate1, error1);
+      }
+      /* And pick the best. */
+      best = rd_cost1 < rd_cost0;
+      base_bits = *(vp8_dct_value_cost_ptr + x);
+      dx = dqcoeff_ptr[rc] - coeff_ptr[rc];
+      d2 = dx * dx;
+      tokens[i][0].rate = base_bits + (best ? rate1 : rate0);
+      tokens[i][0].error = d2 + (best ? error1 : error0);
+      tokens[i][0].next = next;
+      tokens[i][0].token = t0;
+      tokens[i][0].qc = x;
+      best_mask[0] |= best << i;
+      /* Evaluate the second possibility for this state. */
+      rate0 = tokens[next][0].rate;
+      rate1 = tokens[next][1].rate;
 
-            if((abs(x)*dequant_ptr[rc]>abs(coeff_ptr[rc])) &&
-               (abs(x)*dequant_ptr[rc]<abs(coeff_ptr[rc])+dequant_ptr[rc]))
-                shortcut = 1;
-            else
-                shortcut = 0;
+      if ((abs(x) * dequant_ptr[rc] > abs(coeff_ptr[rc])) &&
+          (abs(x) * dequant_ptr[rc] < abs(coeff_ptr[rc]) + dequant_ptr[rc]))
+        shortcut = 1;
+      else
+        shortcut = 0;
 
-            if(shortcut)
-            {
-                sz = -(x < 0);
-                x -= 2*sz + 1;
-            }
+      if (shortcut) {
+        sz = -(x < 0);
+        x -= 2 * sz + 1;
+      }
 
-            /* Consider both possible successor states. */
-            if (!x)
-            {
-                /* If we reduced this coefficient to zero, check to see if
-                 *  we need to move the EOB back here.
-                 */
-                t0 = tokens[next][0].token == DCT_EOB_TOKEN ?
-                    DCT_EOB_TOKEN : ZERO_TOKEN;
-                t1 = tokens[next][1].token == DCT_EOB_TOKEN ?
-                    DCT_EOB_TOKEN : ZERO_TOKEN;
-            }
-            else
-            {
-                t0=t1 = (vp8_dct_value_tokens_ptr + x)->Token;
-            }
-            if (next < 16)
-            {
-                band = vp8_coef_bands[i + 1];
-                if(t0!=DCT_EOB_TOKEN)
-                {
-                    pt = vp8_prev_token_class[t0];
-                    rate0 += mb->token_costs[type][band][pt][
-                        tokens[next][0].token];
-                }
-                if(t1!=DCT_EOB_TOKEN)
-                {
-                    pt = vp8_prev_token_class[t1];
-                    rate1 += mb->token_costs[type][band][pt][
-                        tokens[next][1].token];
-                }
-            }
-
-            rd_cost0 = RDCOST(rdmult, rddiv, rate0, error0);
-            rd_cost1 = RDCOST(rdmult, rddiv, rate1, error1);
-            if (rd_cost0 == rd_cost1)
-            {
-                rd_cost0 = RDTRUNC(rdmult, rddiv, rate0, error0);
-                rd_cost1 = RDTRUNC(rdmult, rddiv, rate1, error1);
-            }
-            /* And pick the best. */
-            best = rd_cost1 < rd_cost0;
-            base_bits = *(vp8_dct_value_cost_ptr + x);
-
-            if(shortcut)
-            {
-                dx -= (dequant_ptr[rc] + sz) ^ sz;
-                d2 = dx*dx;
-            }
-            tokens[i][1].rate = base_bits + (best ? rate1 : rate0);
-            tokens[i][1].error = d2 + (best ? error1 : error0);
-            tokens[i][1].next = next;
-            tokens[i][1].token =best?t1:t0;
-            tokens[i][1].qc = x;
-            best_mask[1] |= best << i;
-            /* Finally, make this the new head of the trellis. */
-            next = i;
-        }
-        /* There's no choice to make for a zero coefficient, so we don't
-         *  add a new trellis node, but we do need to update the costs.
+      /* Consider both possible successor states. */
+      if (!x) {
+        /* If we reduced this coefficient to zero, check to see if
+         *  we need to move the EOB back here.
          */
-        else
-        {
-            band = vp8_coef_bands[i + 1];
-            t0 = tokens[next][0].token;
-            t1 = tokens[next][1].token;
-            /* Update the cost of each path if we're past the EOB token. */
-            if (t0 != DCT_EOB_TOKEN)
-            {
-                tokens[next][0].rate += mb->token_costs[type][band][0][t0];
-                tokens[next][0].token = ZERO_TOKEN;
-            }
-            if (t1 != DCT_EOB_TOKEN)
-            {
-                tokens[next][1].rate += mb->token_costs[type][band][0][t1];
-                tokens[next][1].token = ZERO_TOKEN;
-            }
-            /* Don't update next, because we didn't add a new node. */
+        t0 =
+            tokens[next][0].token == DCT_EOB_TOKEN ? DCT_EOB_TOKEN : ZERO_TOKEN;
+        t1 =
+            tokens[next][1].token == DCT_EOB_TOKEN ? DCT_EOB_TOKEN : ZERO_TOKEN;
+      } else {
+        t0 = t1 = (vp8_dct_value_tokens_ptr + x)->Token;
+      }
+      if (next < 16) {
+        band = vp8_coef_bands[i + 1];
+        if (t0 != DCT_EOB_TOKEN) {
+          pt = vp8_prev_token_class[t0];
+          rate0 += mb->token_costs[type][band][pt][tokens[next][0].token];
         }
-    }
+        if (t1 != DCT_EOB_TOKEN) {
+          pt = vp8_prev_token_class[t1];
+          rate1 += mb->token_costs[type][band][pt][tokens[next][1].token];
+        }
+      }
 
-    /* Now pick the best path through the whole trellis. */
-    band = vp8_coef_bands[i + 1];
-    VP8_COMBINEENTROPYCONTEXTS(pt, *a, *l);
-    rate0 = tokens[next][0].rate;
-    rate1 = tokens[next][1].rate;
-    error0 = tokens[next][0].error;
-    error1 = tokens[next][1].error;
-    t0 = tokens[next][0].token;
-    t1 = tokens[next][1].token;
-    rate0 += mb->token_costs[type][band][pt][t0];
-    rate1 += mb->token_costs[type][band][pt][t1];
-    rd_cost0 = RDCOST(rdmult, rddiv, rate0, error0);
-    rd_cost1 = RDCOST(rdmult, rddiv, rate1, error1);
-    if (rd_cost0 == rd_cost1)
-    {
+      rd_cost0 = RDCOST(rdmult, rddiv, rate0, error0);
+      rd_cost1 = RDCOST(rdmult, rddiv, rate1, error1);
+      if (rd_cost0 == rd_cost1) {
         rd_cost0 = RDTRUNC(rdmult, rddiv, rate0, error0);
         rd_cost1 = RDTRUNC(rdmult, rddiv, rate1, error1);
+      }
+      /* And pick the best. */
+      best = rd_cost1 < rd_cost0;
+      base_bits = *(vp8_dct_value_cost_ptr + x);
+
+      if (shortcut) {
+        dx -= (dequant_ptr[rc] + sz) ^ sz;
+        d2 = dx * dx;
+      }
+      tokens[i][1].rate = base_bits + (best ? rate1 : rate0);
+      tokens[i][1].error = d2 + (best ? error1 : error0);
+      tokens[i][1].next = next;
+      tokens[i][1].token = best ? t1 : t0;
+      tokens[i][1].qc = x;
+      best_mask[1] |= best << i;
+      /* Finally, make this the new head of the trellis. */
+      next = i;
     }
-    best = rd_cost1 < rd_cost0;
-    final_eob = i0 - 1;
-    for (i = next; i < eob; i = next)
-    {
-        x = tokens[i][best].qc;
-        if (x)
-            final_eob = i;
-        rc = vp8_default_zig_zag1d[i];
-        qcoeff_ptr[rc] = x;
-        dqcoeff_ptr[rc] = x * dequant_ptr[rc];
-        next = tokens[i][best].next;
-        best = (best_mask[best] >> i) & 1;
+    /* There's no choice to make for a zero coefficient, so we don't
+     *  add a new trellis node, but we do need to update the costs.
+     */
+    else {
+      band = vp8_coef_bands[i + 1];
+      t0 = tokens[next][0].token;
+      t1 = tokens[next][1].token;
+      /* Update the cost of each path if we're past the EOB token. */
+      if (t0 != DCT_EOB_TOKEN) {
+        tokens[next][0].rate += mb->token_costs[type][band][0][t0];
+        tokens[next][0].token = ZERO_TOKEN;
+      }
+      if (t1 != DCT_EOB_TOKEN) {
+        tokens[next][1].rate += mb->token_costs[type][band][0][t1];
+        tokens[next][1].token = ZERO_TOKEN;
+      }
+      /* Don't update next, because we didn't add a new node. */
     }
-    final_eob++;
+  }
 
-    *a = *l = (final_eob != !type);
-    *d->eob = (char)final_eob;
+  /* Now pick the best path through the whole trellis. */
+  band = vp8_coef_bands[i + 1];
+  VP8_COMBINEENTROPYCONTEXTS(pt, *a, *l);
+  rate0 = tokens[next][0].rate;
+  rate1 = tokens[next][1].rate;
+  error0 = tokens[next][0].error;
+  error1 = tokens[next][1].error;
+  t0 = tokens[next][0].token;
+  t1 = tokens[next][1].token;
+  rate0 += mb->token_costs[type][band][pt][t0];
+  rate1 += mb->token_costs[type][band][pt][t1];
+  rd_cost0 = RDCOST(rdmult, rddiv, rate0, error0);
+  rd_cost1 = RDCOST(rdmult, rddiv, rate1, error1);
+  if (rd_cost0 == rd_cost1) {
+    rd_cost0 = RDTRUNC(rdmult, rddiv, rate0, error0);
+    rd_cost1 = RDTRUNC(rdmult, rddiv, rate1, error1);
+  }
+  best = rd_cost1 < rd_cost0;
+  final_eob = i0 - 1;
+  for (i = next; i < eob; i = next) {
+    x = tokens[i][best].qc;
+    if (x) final_eob = i;
+    rc = vp8_default_zig_zag1d[i];
+    qcoeff_ptr[rc] = x;
+    dqcoeff_ptr[rc] = x * dequant_ptr[rc];
+    next = tokens[i][best].next;
+    best = (best_mask[best] >> i) & 1;
+  }
+  final_eob++;
+
+  *a = *l = (final_eob != !type);
+  *d->eob = (char)final_eob;
 }
-static void check_reset_2nd_coeffs(MACROBLOCKD *x, int type,
-                                   ENTROPY_CONTEXT *a, ENTROPY_CONTEXT *l)
-{
-    int sum=0;
-    int i;
-    BLOCKD *bd = &x->block[24];
+static void check_reset_2nd_coeffs(MACROBLOCKD *x, int type, ENTROPY_CONTEXT *a,
+                                   ENTROPY_CONTEXT *l) {
+  int sum = 0;
+  int i;
+  BLOCKD *bd = &x->block[24];
 
-    if(bd->dequant[0]>=35 && bd->dequant[1]>=35)
-        return;
+  if (bd->dequant[0] >= 35 && bd->dequant[1] >= 35) return;
 
-    for(i=0;i<(*bd->eob);i++)
-    {
-        int coef = bd->dqcoeff[vp8_default_zig_zag1d[i]];
-        sum+= (coef>=0)?coef:-coef;
-        if(sum>=35)
-            return;
+  for (i = 0; i < (*bd->eob); i++) {
+    int coef = bd->dqcoeff[vp8_default_zig_zag1d[i]];
+    sum += (coef >= 0) ? coef : -coef;
+    if (sum >= 35) return;
+  }
+  /**************************************************************************
+  our inverse hadamard transform effectively is weighted sum of all 16 inputs
+  with weight either 1 or -1. It has a last stage scaling of (sum+3)>>3. And
+  dc only idct is (dc+4)>>3. So if all the sums are between -35 and 29, the
+  output after inverse wht and idct will be all zero. A sum of absolute value
+  smaller than 35 guarantees all 16 different (+1/-1) weighted sums in wht
+  fall between -35 and +35.
+  **************************************************************************/
+  if (sum < 35) {
+    for (i = 0; i < (*bd->eob); i++) {
+      int rc = vp8_default_zig_zag1d[i];
+      bd->qcoeff[rc] = 0;
+      bd->dqcoeff[rc] = 0;
     }
-    /**************************************************************************
-    our inverse hadamard transform effectively is weighted sum of all 16 inputs
-    with weight either 1 or -1. It has a last stage scaling of (sum+3)>>3. And
-    dc only idct is (dc+4)>>3. So if all the sums are between -35 and 29, the
-    output after inverse wht and idct will be all zero. A sum of absolute value
-    smaller than 35 guarantees all 16 different (+1/-1) weighted sums in wht
-    fall between -35 and +35.
-    **************************************************************************/
-    if(sum < 35)
-    {
-        for(i=0;i<(*bd->eob);i++)
-        {
-            int rc = vp8_default_zig_zag1d[i];
-            bd->qcoeff[rc]=0;
-            bd->dqcoeff[rc]=0;
-        }
-        *bd->eob = 0;
-        *a = *l = (*bd->eob != !type);
-    }
+    *bd->eob = 0;
+    *a = *l = (*bd->eob != !type);
+  }
 }
 
-static void optimize_mb(MACROBLOCK *x)
-{
-    int b;
-    int type;
-    int has_2nd_order;
+static void optimize_mb(MACROBLOCK *x) {
+  int b;
+  int type;
+  int has_2nd_order;
 
-    ENTROPY_CONTEXT_PLANES t_above, t_left;
-    ENTROPY_CONTEXT *ta;
-    ENTROPY_CONTEXT *tl;
+  ENTROPY_CONTEXT_PLANES t_above, t_left;
+  ENTROPY_CONTEXT *ta;
+  ENTROPY_CONTEXT *tl;
 
-    memcpy(&t_above, x->e_mbd.above_context, sizeof(ENTROPY_CONTEXT_PLANES));
-    memcpy(&t_left, x->e_mbd.left_context, sizeof(ENTROPY_CONTEXT_PLANES));
+  memcpy(&t_above, x->e_mbd.above_context, sizeof(ENTROPY_CONTEXT_PLANES));
+  memcpy(&t_left, x->e_mbd.left_context, sizeof(ENTROPY_CONTEXT_PLANES));
 
-    ta = (ENTROPY_CONTEXT *)&t_above;
-    tl = (ENTROPY_CONTEXT *)&t_left;
+  ta = (ENTROPY_CONTEXT *)&t_above;
+  tl = (ENTROPY_CONTEXT *)&t_left;
 
-    has_2nd_order = (x->e_mbd.mode_info_context->mbmi.mode != B_PRED
-        && x->e_mbd.mode_info_context->mbmi.mode != SPLITMV);
-    type = has_2nd_order ? PLANE_TYPE_Y_NO_DC : PLANE_TYPE_Y_WITH_DC;
+  has_2nd_order = (x->e_mbd.mode_info_context->mbmi.mode != B_PRED &&
+                   x->e_mbd.mode_info_context->mbmi.mode != SPLITMV);
+  type = has_2nd_order ? PLANE_TYPE_Y_NO_DC : PLANE_TYPE_Y_WITH_DC;
 
-    for (b = 0; b < 16; b++)
-    {
-        optimize_b(x, b, type,
-            ta + vp8_block2above[b], tl + vp8_block2left[b]);
-    }
+  for (b = 0; b < 16; b++) {
+    optimize_b(x, b, type, ta + vp8_block2above[b], tl + vp8_block2left[b]);
+  }
 
-    for (b = 16; b < 24; b++)
-    {
-        optimize_b(x, b, PLANE_TYPE_UV,
-            ta + vp8_block2above[b], tl + vp8_block2left[b]);
-    }
+  for (b = 16; b < 24; b++) {
+    optimize_b(x, b, PLANE_TYPE_UV, ta + vp8_block2above[b],
+               tl + vp8_block2left[b]);
+  }
 
-    if (has_2nd_order)
-    {
-        b=24;
-        optimize_b(x, b, PLANE_TYPE_Y2,
-            ta + vp8_block2above[b], tl + vp8_block2left[b]);
-        check_reset_2nd_coeffs(&x->e_mbd, PLANE_TYPE_Y2,
-            ta + vp8_block2above[b], tl + vp8_block2left[b]);
-    }
+  if (has_2nd_order) {
+    b = 24;
+    optimize_b(x, b, PLANE_TYPE_Y2, ta + vp8_block2above[b],
+               tl + vp8_block2left[b]);
+    check_reset_2nd_coeffs(&x->e_mbd, PLANE_TYPE_Y2, ta + vp8_block2above[b],
+                           tl + vp8_block2left[b]);
+  }
 }
 
+void vp8_optimize_mby(MACROBLOCK *x) {
+  int b;
+  int type;
+  int has_2nd_order;
 
-void vp8_optimize_mby(MACROBLOCK *x)
-{
-    int b;
-    int type;
-    int has_2nd_order;
+  ENTROPY_CONTEXT_PLANES t_above, t_left;
+  ENTROPY_CONTEXT *ta;
+  ENTROPY_CONTEXT *tl;
 
-    ENTROPY_CONTEXT_PLANES t_above, t_left;
-    ENTROPY_CONTEXT *ta;
-    ENTROPY_CONTEXT *tl;
+  if (!x->e_mbd.above_context) return;
 
-    if (!x->e_mbd.above_context)
-        return;
+  if (!x->e_mbd.left_context) return;
 
-    if (!x->e_mbd.left_context)
-        return;
+  memcpy(&t_above, x->e_mbd.above_context, sizeof(ENTROPY_CONTEXT_PLANES));
+  memcpy(&t_left, x->e_mbd.left_context, sizeof(ENTROPY_CONTEXT_PLANES));
 
-    memcpy(&t_above, x->e_mbd.above_context, sizeof(ENTROPY_CONTEXT_PLANES));
-    memcpy(&t_left, x->e_mbd.left_context, sizeof(ENTROPY_CONTEXT_PLANES));
+  ta = (ENTROPY_CONTEXT *)&t_above;
+  tl = (ENTROPY_CONTEXT *)&t_left;
 
-    ta = (ENTROPY_CONTEXT *)&t_above;
-    tl = (ENTROPY_CONTEXT *)&t_left;
+  has_2nd_order = (x->e_mbd.mode_info_context->mbmi.mode != B_PRED &&
+                   x->e_mbd.mode_info_context->mbmi.mode != SPLITMV);
+  type = has_2nd_order ? PLANE_TYPE_Y_NO_DC : PLANE_TYPE_Y_WITH_DC;
 
-    has_2nd_order = (x->e_mbd.mode_info_context->mbmi.mode != B_PRED
-        && x->e_mbd.mode_info_context->mbmi.mode != SPLITMV);
-    type = has_2nd_order ? PLANE_TYPE_Y_NO_DC : PLANE_TYPE_Y_WITH_DC;
+  for (b = 0; b < 16; b++) {
+    optimize_b(x, b, type, ta + vp8_block2above[b], tl + vp8_block2left[b]);
+  }
 
-    for (b = 0; b < 16; b++)
-    {
-        optimize_b(x, b, type,
-            ta + vp8_block2above[b], tl + vp8_block2left[b]);
-    }
-
-
-    if (has_2nd_order)
-    {
-        b=24;
-        optimize_b(x, b, PLANE_TYPE_Y2,
-            ta + vp8_block2above[b], tl + vp8_block2left[b]);
-        check_reset_2nd_coeffs(&x->e_mbd, PLANE_TYPE_Y2,
-            ta + vp8_block2above[b], tl + vp8_block2left[b]);
-    }
+  if (has_2nd_order) {
+    b = 24;
+    optimize_b(x, b, PLANE_TYPE_Y2, ta + vp8_block2above[b],
+               tl + vp8_block2left[b]);
+    check_reset_2nd_coeffs(&x->e_mbd, PLANE_TYPE_Y2, ta + vp8_block2above[b],
+                           tl + vp8_block2left[b]);
+  }
 }
 
-void vp8_optimize_mbuv(MACROBLOCK *x)
-{
-    int b;
-    ENTROPY_CONTEXT_PLANES t_above, t_left;
-    ENTROPY_CONTEXT *ta;
-    ENTROPY_CONTEXT *tl;
+void vp8_optimize_mbuv(MACROBLOCK *x) {
+  int b;
+  ENTROPY_CONTEXT_PLANES t_above, t_left;
+  ENTROPY_CONTEXT *ta;
+  ENTROPY_CONTEXT *tl;
 
-    if (!x->e_mbd.above_context)
-        return;
+  if (!x->e_mbd.above_context) return;
 
-    if (!x->e_mbd.left_context)
-        return;
+  if (!x->e_mbd.left_context) return;
 
-    memcpy(&t_above, x->e_mbd.above_context, sizeof(ENTROPY_CONTEXT_PLANES));
-    memcpy(&t_left, x->e_mbd.left_context, sizeof(ENTROPY_CONTEXT_PLANES));
+  memcpy(&t_above, x->e_mbd.above_context, sizeof(ENTROPY_CONTEXT_PLANES));
+  memcpy(&t_left, x->e_mbd.left_context, sizeof(ENTROPY_CONTEXT_PLANES));
 
-    ta = (ENTROPY_CONTEXT *)&t_above;
-    tl = (ENTROPY_CONTEXT *)&t_left;
+  ta = (ENTROPY_CONTEXT *)&t_above;
+  tl = (ENTROPY_CONTEXT *)&t_left;
 
-    for (b = 16; b < 24; b++)
-    {
-        optimize_b(x, b, PLANE_TYPE_UV,
-            ta + vp8_block2above[b], tl + vp8_block2left[b]);
-    }
+  for (b = 16; b < 24; b++) {
+    optimize_b(x, b, PLANE_TYPE_UV, ta + vp8_block2above[b],
+               tl + vp8_block2left[b]);
+  }
 }
 
-void vp8_encode_inter16x16(MACROBLOCK *x)
-{
-    vp8_build_inter_predictors_mb(&x->e_mbd);
+void vp8_encode_inter16x16(MACROBLOCK *x) {
+  vp8_build_inter_predictors_mb(&x->e_mbd);
 
-    vp8_subtract_mb(x);
+  vp8_subtract_mb(x);
 
-    transform_mb(x);
+  transform_mb(x);
 
-    vp8_quantize_mb(x);
+  vp8_quantize_mb(x);
 
-    if (x->optimize)
-        optimize_mb(x);
+  if (x->optimize) optimize_mb(x);
 }
 
 /* this funciton is used by first pass only */
-void vp8_encode_inter16x16y(MACROBLOCK *x)
-{
-    BLOCK *b = &x->block[0];
+void vp8_encode_inter16x16y(MACROBLOCK *x) {
+  BLOCK *b = &x->block[0];
 
-    vp8_build_inter16x16_predictors_mby(&x->e_mbd, x->e_mbd.dst.y_buffer,
-                                        x->e_mbd.dst.y_stride);
+  vp8_build_inter16x16_predictors_mby(&x->e_mbd, x->e_mbd.dst.y_buffer,
+                                      x->e_mbd.dst.y_stride);
 
-    vp8_subtract_mby(x->src_diff, *(b->base_src),
-        b->src_stride, x->e_mbd.dst.y_buffer, x->e_mbd.dst.y_stride);
+  vp8_subtract_mby(x->src_diff, *(b->base_src), b->src_stride,
+                   x->e_mbd.dst.y_buffer, x->e_mbd.dst.y_stride);
 
-    transform_mby(x);
+  transform_mby(x);
 
-    vp8_quantize_mby(x);
+  vp8_quantize_mby(x);
 
-    vp8_inverse_transform_mby(&x->e_mbd);
+  vp8_inverse_transform_mby(&x->e_mbd);
 }
--- a/vp8/encoder/encodemb.h
+++ b/vp8/encoder/encodemb.h
@@ -8,7 +8,6 @@
  *  be found in the AUTHORS file in the root of the source tree.
  */
 
-
 #ifndef VP8_ENCODER_ENCODEMB_H_
 #define VP8_ENCODER_ENCODEMB_H_
 
--- a/vp8/encoder/encodemv.c
+++ b/vp8/encoder/encodemv.c
@@ -8,7 +8,6 @@
  *  be found in the AUTHORS file in the root of the source tree.
  */
 
-
 #include "vp8/common/common.h"
 #include "encodemv.h"
 #include "vp8/common/entropymode.h"
@@ -20,54 +19,46 @@
 extern unsigned int active_section;
 #endif
 
-static void encode_mvcomponent(
-    vp8_writer *const w,
-    const int v,
-    const struct mv_context *mvc
-)
-{
-    const vp8_prob *p = mvc->prob;
-    const int x = v < 0 ? -v : v;
+static void encode_mvcomponent(vp8_writer *const w, const int v,
+                               const struct mv_context *mvc) {
+  const vp8_prob *p = mvc->prob;
+  const int x = v < 0 ? -v : v;
 
-    if (x < mvnum_short)     /* Small */
-    {
-        vp8_write(w, 0, p [mvpis_short]);
-        vp8_treed_write(w, vp8_small_mvtree, p + MVPshort, x, 3);
+  if (x < mvnum_short) /* Small */
+  {
+    vp8_write(w, 0, p[mvpis_short]);
+    vp8_treed_write(w, vp8_small_mvtree, p + MVPshort, x, 3);
 
-        if (!x)
-            return;         /* no sign bit */
-    }
-    else                    /* Large */
-    {
-        int i = 0;
+    if (!x) return; /* no sign bit */
+  } else            /* Large */
+  {
+    int i = 0;
 
-        vp8_write(w, 1, p [mvpis_short]);
+    vp8_write(w, 1, p[mvpis_short]);
 
-        do
-            vp8_write(w, (x >> i) & 1, p [MVPbits + i]);
+    do
+      vp8_write(w, (x >> i) & 1, p[MVPbits + i]);
 
-        while (++i < 3);
+    while (++i < 3);
 
-        i = mvlong_width - 1;  /* Skip bit 3, which is sometimes implicit */
+    i = mvlong_width - 1; /* Skip bit 3, which is sometimes implicit */
 
-        do
-            vp8_write(w, (x >> i) & 1, p [MVPbits + i]);
+    do
+      vp8_write(w, (x >> i) & 1, p[MVPbits + i]);
 
-        while (--i > 3);
+    while (--i > 3);
 
-        if (x & 0xFFF0)
-            vp8_write(w, (x >> 3) & 1, p [MVPbits + 3]);
-    }
+    if (x & 0xFFF0) vp8_write(w, (x >> 3) & 1, p[MVPbits + 3]);
+  }
 
-    vp8_write(w, v < 0, p [MVPsign]);
+  vp8_write(w, v < 0, p[MVPsign]);
 }
 #if 0
 static int max_mv_r = 0;
 static int max_mv_c = 0;
 #endif
-void vp8_encode_motion_vector(vp8_writer *w, const MV *mv, const MV_CONTEXT *mvc)
-{
-
+void vp8_encode_motion_vector(vp8_writer *w, const MV *mv,
+                              const MV_CONTEXT *mvc) {
 #if 0
     {
         if (abs(mv->row >> 1) > max_mv_r)
@@ -92,289 +83,250 @@
     }
 #endif
 
-    encode_mvcomponent(w, mv->row >> 1, &mvc[0]);
-    encode_mvcomponent(w, mv->col >> 1, &mvc[1]);
+  encode_mvcomponent(w, mv->row >> 1, &mvc[0]);
+  encode_mvcomponent(w, mv->col >> 1, &mvc[1]);
 }
 
+static unsigned int cost_mvcomponent(const int v,
+                                     const struct mv_context *mvc) {
+  const vp8_prob *p = mvc->prob;
+  const int x = v;
+  unsigned int cost;
 
-static unsigned int cost_mvcomponent(const int v, const struct mv_context *mvc)
-{
-    const vp8_prob *p = mvc->prob;
-    const int x = v;
-    unsigned int cost;
+  if (x < mvnum_short) {
+    cost = vp8_cost_zero(p[mvpis_short]) +
+           vp8_treed_cost(vp8_small_mvtree, p + MVPshort, x, 3);
 
-    if (x < mvnum_short)
-    {
-        cost = vp8_cost_zero(p [mvpis_short])
-               + vp8_treed_cost(vp8_small_mvtree, p + MVPshort, x, 3);
+    if (!x) return cost;
+  } else {
+    int i = 0;
+    cost = vp8_cost_one(p[mvpis_short]);
 
-        if (!x)
-            return cost;
-    }
-    else
-    {
-        int i = 0;
-        cost = vp8_cost_one(p [mvpis_short]);
+    do
+      cost += vp8_cost_bit(p[MVPbits + i], (x >> i) & 1);
 
-        do
-            cost += vp8_cost_bit(p [MVPbits + i], (x >> i) & 1);
+    while (++i < 3);
 
-        while (++i < 3);
+    i = mvlong_width - 1; /* Skip bit 3, which is sometimes implicit */
 
-        i = mvlong_width - 1;  /* Skip bit 3, which is sometimes implicit */
+    do
+      cost += vp8_cost_bit(p[MVPbits + i], (x >> i) & 1);
 
-        do
-            cost += vp8_cost_bit(p [MVPbits + i], (x >> i) & 1);
+    while (--i > 3);
 
-        while (--i > 3);
+    if (x & 0xFFF0) cost += vp8_cost_bit(p[MVPbits + 3], (x >> 3) & 1);
+  }
 
-        if (x & 0xFFF0)
-            cost += vp8_cost_bit(p [MVPbits + 3], (x >> 3) & 1);
-    }
-
-    return cost;   /* + vp8_cost_bit( p [MVPsign], v < 0); */
+  return cost; /* + vp8_cost_bit( p [MVPsign], v < 0); */
 }
 
-void vp8_build_component_cost_table(int *mvcost[2], const MV_CONTEXT *mvc, int mvc_flag[2])
-{
-    int i = 1;
-    unsigned int cost0 = 0;
-    unsigned int cost1 = 0;
+void vp8_build_component_cost_table(int *mvcost[2], const MV_CONTEXT *mvc,
+                                    int mvc_flag[2]) {
+  int i = 1;
+  unsigned int cost0 = 0;
+  unsigned int cost1 = 0;
 
-    vp8_clear_system_state();
+  vp8_clear_system_state();
 
-    i = 1;
+  i = 1;
 
-    if (mvc_flag[0])
-    {
-        mvcost [0] [0] = cost_mvcomponent(0, &mvc[0]);
+  if (mvc_flag[0]) {
+    mvcost[0][0] = cost_mvcomponent(0, &mvc[0]);
 
-        do
-        {
-            cost0 = cost_mvcomponent(i, &mvc[0]);
+    do {
+      cost0 = cost_mvcomponent(i, &mvc[0]);
 
-            mvcost [0] [i] = cost0 + vp8_cost_zero(mvc[0].prob[MVPsign]);
-            mvcost [0] [-i] = cost0 + vp8_cost_one(mvc[0].prob[MVPsign]);
-        }
-        while (++i <= mv_max);
-    }
+      mvcost[0][i] = cost0 + vp8_cost_zero(mvc[0].prob[MVPsign]);
+      mvcost[0][-i] = cost0 + vp8_cost_one(mvc[0].prob[MVPsign]);
+    } while (++i <= mv_max);
+  }
 
-    i = 1;
+  i = 1;
 
-    if (mvc_flag[1])
-    {
-        mvcost [1] [0] = cost_mvcomponent(0, &mvc[1]);
+  if (mvc_flag[1]) {
+    mvcost[1][0] = cost_mvcomponent(0, &mvc[1]);
 
-        do
-        {
-            cost1 = cost_mvcomponent(i, &mvc[1]);
+    do {
+      cost1 = cost_mvcomponent(i, &mvc[1]);
 
-            mvcost [1] [i] = cost1 + vp8_cost_zero(mvc[1].prob[MVPsign]);
-            mvcost [1] [-i] = cost1 + vp8_cost_one(mvc[1].prob[MVPsign]);
-        }
-        while (++i <= mv_max);
-    }
+      mvcost[1][i] = cost1 + vp8_cost_zero(mvc[1].prob[MVPsign]);
+      mvcost[1][-i] = cost1 + vp8_cost_one(mvc[1].prob[MVPsign]);
+    } while (++i <= mv_max);
+  }
 }
 
-
 /* Motion vector probability table update depends on benefit.
  * Small correction allows for the fact that an update to an MV probability
  * may have benefit in subsequent frames as well as the current one.
  */
-#define MV_PROB_UPDATE_CORRECTION   -1
+#define MV_PROB_UPDATE_CORRECTION -1
 
+static void calc_prob(vp8_prob *p, const unsigned int ct[2]) {
+  const unsigned int tot = ct[0] + ct[1];
 
-static void calc_prob(vp8_prob *p, const unsigned int ct[2])
-{
-    const unsigned int tot = ct[0] + ct[1];
-
-    if (tot)
-    {
-        const vp8_prob x = ((ct[0] * 255) / tot) & -2;
-        *p = x ? x : 1;
-    }
+  if (tot) {
+    const vp8_prob x = ((ct[0] * 255) / tot) & -2;
+    *p = x ? x : 1;
+  }
 }
 
-static void update(
-    vp8_writer *const w,
-    const unsigned int ct[2],
-    vp8_prob *const cur_p,
-    const vp8_prob new_p,
-    const vp8_prob update_p,
-    int *updated
-)
-{
-    const int cur_b = vp8_cost_branch(ct, *cur_p);
-    const int new_b = vp8_cost_branch(ct, new_p);
-    const int cost = 7 + MV_PROB_UPDATE_CORRECTION + ((vp8_cost_one(update_p) - vp8_cost_zero(update_p) + 128) >> 8);
+static void update(vp8_writer *const w, const unsigned int ct[2],
+                   vp8_prob *const cur_p, const vp8_prob new_p,
+                   const vp8_prob update_p, int *updated) {
+  const int cur_b = vp8_cost_branch(ct, *cur_p);
+  const int new_b = vp8_cost_branch(ct, new_p);
+  const int cost =
+      7 + MV_PROB_UPDATE_CORRECTION +
+      ((vp8_cost_one(update_p) - vp8_cost_zero(update_p) + 128) >> 8);
 
-    if (cur_b - new_b > cost)
-    {
-        *cur_p = new_p;
-        vp8_write(w, 1, update_p);
-        vp8_write_literal(w, new_p >> 1, 7);
-        *updated = 1;
+  if (cur_b - new_b > cost) {
+    *cur_p = new_p;
+    vp8_write(w, 1, update_p);
+    vp8_write_literal(w, new_p >> 1, 7);
+    *updated = 1;
 
-    }
-    else
-        vp8_write(w, 0, update_p);
+  } else
+    vp8_write(w, 0, update_p);
 }
 
-static void write_component_probs(
-    vp8_writer *const w,
-    struct mv_context *cur_mvc,
-    const struct mv_context *default_mvc_,
-    const struct mv_context *update_mvc,
-    const unsigned int events [MVvals],
-    unsigned int rc,
-    int *updated
-)
-{
-    vp8_prob *Pcur = cur_mvc->prob;
-    const vp8_prob *default_mvc = default_mvc_->prob;
-    const vp8_prob *Pupdate = update_mvc->prob;
-    unsigned int is_short_ct[2], sign_ct[2];
+static void write_component_probs(vp8_writer *const w,
+                                  struct mv_context *cur_mvc,
+                                  const struct mv_context *default_mvc_,
+                                  const struct mv_context *update_mvc,
+                                  const unsigned int events[MVvals],
+                                  unsigned int rc, int *updated) {
+  vp8_prob *Pcur = cur_mvc->prob;
+  const vp8_prob *default_mvc = default_mvc_->prob;
+  const vp8_prob *Pupdate = update_mvc->prob;
+  unsigned int is_short_ct[2], sign_ct[2];
 
-    unsigned int bit_ct [mvlong_width] [2];
+  unsigned int bit_ct[mvlong_width][2];
 
-    unsigned int short_ct  [mvnum_short];
-    unsigned int short_bct [mvnum_short-1] [2];
+  unsigned int short_ct[mvnum_short];
+  unsigned int short_bct[mvnum_short - 1][2];
 
-    vp8_prob Pnew [MVPcount];
+  vp8_prob Pnew[MVPcount];
 
-    (void) rc;
-    vp8_copy_array(Pnew, default_mvc, MVPcount);
+  (void)rc;
+  vp8_copy_array(Pnew, default_mvc, MVPcount);
 
-    vp8_zero(is_short_ct)
-    vp8_zero(sign_ct)
-    vp8_zero(bit_ct)
-    vp8_zero(short_ct)
-    vp8_zero(short_bct)
+  vp8_zero(is_short_ct) vp8_zero(sign_ct) vp8_zero(bit_ct) vp8_zero(short_ct)
+      vp8_zero(short_bct)
 
+  /* j=0 */
+  {
+    const int c = events[mv_max];
 
-    /* j=0 */
-    {
-        const int c = events [mv_max];
+    is_short_ct[0] += c; /* Short vector */
+    short_ct[0] += c;    /* Magnitude distribution */
+  }
 
-        is_short_ct [0] += c;     /* Short vector */
-        short_ct [0] += c;       /* Magnitude distribution */
-    }
+  /* j: 1 ~ mv_max (1023) */
+  {
+    int j = 1;
 
-    /* j: 1 ~ mv_max (1023) */
-    {
-        int j = 1;
+    do {
+      const int c1 = events[mv_max + j]; /* positive */
+      const int c2 = events[mv_max - j]; /* negative */
+      const int c = c1 + c2;
+      int a = j;
 
-        do
-        {
-            const int c1 = events [mv_max + j];  /* positive */
-            const int c2 = events [mv_max - j];  /* negative */
-            const int c  = c1 + c2;
-            int a = j;
+      sign_ct[0] += c1;
+      sign_ct[1] += c2;
 
-            sign_ct [0] += c1;
-            sign_ct [1] += c2;
+      if (a < mvnum_short) {
+        is_short_ct[0] += c; /* Short vector */
+        short_ct[a] += c;    /* Magnitude distribution */
+      } else {
+        int k = mvlong_width - 1;
+        is_short_ct[1] += c; /* Long vector */
 
-            if (a < mvnum_short)
-            {
-                is_short_ct [0] += c;     /* Short vector */
-                short_ct [a] += c;       /* Magnitude distribution */
-            }
-            else
-            {
-                int k = mvlong_width - 1;
-                is_short_ct [1] += c;     /* Long vector */
+        /*  bit 3 not always encoded. */
+        do
+          bit_ct[k][(a >> k) & 1] += c;
 
-                /*  bit 3 not always encoded. */
-                do
-                    bit_ct [k] [(a >> k) & 1] += c;
+        while (--k >= 0);
+      }
+    } while (++j <= mv_max);
+  }
 
-                while (--k >= 0);
-            }
-        }
-        while (++j <= mv_max);
-    }
+  calc_prob(Pnew + mvpis_short, is_short_ct);
 
-    calc_prob(Pnew + mvpis_short, is_short_ct);
+  calc_prob(Pnew + MVPsign, sign_ct);
 
-    calc_prob(Pnew + MVPsign, sign_ct);
+  {
+    vp8_prob p[mvnum_short - 1]; /* actually only need branch ct */
+    int j = 0;
 
-    {
-        vp8_prob p [mvnum_short - 1];    /* actually only need branch ct */
-        int j = 0;
+    vp8_tree_probs_from_distribution(8, vp8_small_mvencodings, vp8_small_mvtree,
+                                     p, short_bct, short_ct, 256, 1);
 
-        vp8_tree_probs_from_distribution(
-            8, vp8_small_mvencodings, vp8_small_mvtree,
-            p, short_bct, short_ct,
-            256, 1
-        );
+    do
+      calc_prob(Pnew + MVPshort + j, short_bct[j]);
 
-        do
-            calc_prob(Pnew + MVPshort + j, short_bct[j]);
+    while (++j < mvnum_short - 1);
+  }
 
-        while (++j < mvnum_short - 1);
-    }
+  {
+    int j = 0;
 
-    {
-        int j = 0;
+    do
+      calc_prob(Pnew + MVPbits + j, bit_ct[j]);
 
-        do
-            calc_prob(Pnew + MVPbits + j, bit_ct[j]);
+    while (++j < mvlong_width);
+  }
 
-        while (++j < mvlong_width);
-    }
+  update(w, is_short_ct, Pcur + mvpis_short, Pnew[mvpis_short], *Pupdate++,
+         updated);
 
-    update(w, is_short_ct, Pcur + mvpis_short, Pnew[mvpis_short], *Pupdate++, updated);
+  update(w, sign_ct, Pcur + MVPsign, Pnew[MVPsign], *Pupdate++, updated);
 
-    update(w, sign_ct, Pcur + MVPsign, Pnew[MVPsign], *Pupdate++, updated);
+  {
+    const vp8_prob *const new_p = Pnew + MVPshort;
+    vp8_prob *const cur_p = Pcur + MVPshort;
 
-    {
-        const vp8_prob *const new_p = Pnew + MVPshort;
-        vp8_prob *const cur_p = Pcur + MVPshort;
+    int j = 0;
 
-        int j = 0;
+    do
 
-        do
+      update(w, short_bct[j], cur_p + j, new_p[j], *Pupdate++, updated);
 
-            update(w, short_bct[j], cur_p + j, new_p[j], *Pupdate++, updated);
+    while (++j < mvnum_short - 1);
+  }
 
-        while (++j < mvnum_short - 1);
-    }
+  {
+    const vp8_prob *const new_p = Pnew + MVPbits;
+    vp8_prob *const cur_p = Pcur + MVPbits;
 
-    {
-        const vp8_prob *const new_p = Pnew + MVPbits;
-        vp8_prob *const cur_p = Pcur + MVPbits;
+    int j = 0;
 
-        int j = 0;
+    do
 
-        do
+      update(w, bit_ct[j], cur_p + j, new_p[j], *Pupdate++, updated);
 
-            update(w, bit_ct[j], cur_p + j, new_p[j], *Pupdate++, updated);
-
-        while (++j < mvlong_width);
-    }
+    while (++j < mvlong_width);
+  }
 }
 
-void vp8_write_mvprobs(VP8_COMP *cpi)
-{
-    vp8_writer *const w  = cpi->bc;
-    MV_CONTEXT *mvc = cpi->common.fc.mvc;
-    int flags[2] = {0, 0};
+void vp8_write_mvprobs(VP8_COMP *cpi) {
+  vp8_writer *const w = cpi->bc;
+  MV_CONTEXT *mvc = cpi->common.fc.mvc;
+  int flags[2] = { 0, 0 };
 #ifdef VP8_ENTROPY_STATS
-    active_section = 4;
+  active_section = 4;
 #endif
-    write_component_probs(
-        w, &mvc[0], &vp8_default_mv_context[0], &vp8_mv_update_probs[0],
-        cpi->mb.MVcount[0], 0, &flags[0]
-    );
-    write_component_probs(
-        w, &mvc[1], &vp8_default_mv_context[1], &vp8_mv_update_probs[1],
-        cpi->mb.MVcount[1], 1, &flags[1]
-    );
+  write_component_probs(w, &mvc[0], &vp8_default_mv_context[0],
+                        &vp8_mv_update_probs[0], cpi->mb.MVcount[0], 0,
+                        &flags[0]);
+  write_component_probs(w, &mvc[1], &vp8_default_mv_context[1],
+                        &vp8_mv_update_probs[1], cpi->mb.MVcount[1], 1,
+                        &flags[1]);
 
-    if (flags[0] || flags[1])
-        vp8_build_component_cost_table(cpi->mb.mvcost, (const MV_CONTEXT *) cpi->common.fc.mvc, flags);
+  if (flags[0] || flags[1])
+    vp8_build_component_cost_table(
+        cpi->mb.mvcost, (const MV_CONTEXT *)cpi->common.fc.mvc, flags);
 
 #ifdef VP8_ENTROPY_STATS
-    active_section = 5;
+  active_section = 5;
 #endif
 }
--- a/vp8/encoder/encodemv.h
+++ b/vp8/encoder/encodemv.h
@@ -8,7 +8,6 @@
  *  be found in the AUTHORS file in the root of the source tree.
  */
 
-
 #ifndef VP8_ENCODER_ENCODEMV_H_
 #define VP8_ENCODER_ENCODEMV_H_
 
@@ -20,7 +19,8 @@
 
 void vp8_write_mvprobs(VP8_COMP *);
 void vp8_encode_motion_vector(vp8_writer *, const MV *, const MV_CONTEXT *);
-void vp8_build_component_cost_table(int *mvcost[2], const MV_CONTEXT *mvc, int mvc_flag[2]);
+void vp8_build_component_cost_table(int *mvcost[2], const MV_CONTEXT *mvc,
+                                    int mvc_flag[2]);
 
 #ifdef __cplusplus
 }  // extern "C"
--- a/vp8/encoder/ethreading.c
+++ b/vp8/encoder/ethreading.c
@@ -17,653 +17,570 @@
 
 #if CONFIG_MULTITHREAD
 
-extern void vp8cx_mb_init_quantizer(VP8_COMP *cpi, MACROBLOCK *x, int ok_to_skip);
+extern void vp8cx_mb_init_quantizer(VP8_COMP *cpi, MACROBLOCK *x,
+                                    int ok_to_skip);
 
-static THREAD_FUNCTION thread_loopfilter(void *p_data)
-{
-    VP8_COMP *cpi = (VP8_COMP *)(((LPFTHREAD_DATA *)p_data)->ptr1);
-    VP8_COMMON *cm = &cpi->common;
+static THREAD_FUNCTION thread_loopfilter(void *p_data) {
+  VP8_COMP *cpi = (VP8_COMP *)(((LPFTHREAD_DATA *)p_data)->ptr1);
+  VP8_COMMON *cm = &cpi->common;
 
-    while (1)
-    {
-        if (protected_read(&cpi->mt_mutex, &cpi->b_multi_threaded) == 0)
-            break;
+  while (1) {
+    if (protected_read(&cpi->mt_mutex, &cpi->b_multi_threaded) == 0) break;
 
-        if (sem_wait(&cpi->h_event_start_lpf) == 0)
-        {
-            /* we're shutting down */
-            if (protected_read(&cpi->mt_mutex, &cpi->b_multi_threaded) == 0)
-                break;
+    if (sem_wait(&cpi->h_event_start_lpf) == 0) {
+      /* we're shutting down */
+      if (protected_read(&cpi->mt_mutex, &cpi->b_multi_threaded) == 0) break;
 
-            vp8_loopfilter_frame(cpi, cm);
+      vp8_loopfilter_frame(cpi, cm);
 
-            sem_post(&cpi->h_event_end_lpf);
-        }
+      sem_post(&cpi->h_event_end_lpf);
     }
+  }
 
-    return 0;
+  return 0;
 }
 
-static
-THREAD_FUNCTION thread_encoding_proc(void *p_data)
-{
-    int ithread = ((ENCODETHREAD_DATA *)p_data)->ithread;
-    VP8_COMP *cpi = (VP8_COMP *)(((ENCODETHREAD_DATA *)p_data)->ptr1);
-    MB_ROW_COMP *mbri = (MB_ROW_COMP *)(((ENCODETHREAD_DATA *)p_data)->ptr2);
-    ENTROPY_CONTEXT_PLANES mb_row_left_context;
+static THREAD_FUNCTION thread_encoding_proc(void *p_data) {
+  int ithread = ((ENCODETHREAD_DATA *)p_data)->ithread;
+  VP8_COMP *cpi = (VP8_COMP *)(((ENCODETHREAD_DATA *)p_data)->ptr1);
+  MB_ROW_COMP *mbri = (MB_ROW_COMP *)(((ENCODETHREAD_DATA *)p_data)->ptr2);
+  ENTROPY_CONTEXT_PLANES mb_row_left_context;
 
-    while (1)
-    {
-        if (protected_read(&cpi->mt_mutex, &cpi->b_multi_threaded) == 0)
-            break;
+  while (1) {
+    if (protected_read(&cpi->mt_mutex, &cpi->b_multi_threaded) == 0) break;
 
-        if (sem_wait(&cpi->h_event_start_encoding[ithread]) == 0)
-        {
-            const int nsync = cpi->mt_sync_range;
-            VP8_COMMON *cm = &cpi->common;
-            int mb_row;
-            MACROBLOCK *x = &mbri->mb;
-            MACROBLOCKD *xd = &x->e_mbd;
-            TOKENEXTRA *tp ;
+    if (sem_wait(&cpi->h_event_start_encoding[ithread]) == 0) {
+      const int nsync = cpi->mt_sync_range;
+      VP8_COMMON *cm = &cpi->common;
+      int mb_row;
+      MACROBLOCK *x = &mbri->mb;
+      MACROBLOCKD *xd = &x->e_mbd;
+      TOKENEXTRA *tp;
 #if CONFIG_REALTIME_ONLY & CONFIG_ONTHEFLY_BITPACKING
-            TOKENEXTRA *tp_start = cpi->tok + (1 + ithread) * (16 * 24);
-            const int num_part = (1 << cm->multi_token_partition);
+      TOKENEXTRA *tp_start = cpi->tok + (1 + ithread) * (16 * 24);
+      const int num_part = (1 << cm->multi_token_partition);
 #endif
 
-            int *segment_counts = mbri->segment_counts;
-            int *totalrate = &mbri->totalrate;
+      int *segment_counts = mbri->segment_counts;
+      int *totalrate = &mbri->totalrate;
 
-            /* we're shutting down */
-            if (protected_read(&cpi->mt_mutex, &cpi->b_multi_threaded) == 0)
-                break;
+      /* we're shutting down */
+      if (protected_read(&cpi->mt_mutex, &cpi->b_multi_threaded) == 0) break;
 
-            xd->mode_info_context = cm->mi + cm->mode_info_stride *
-                (ithread + 1);
-            xd->mode_info_stride = cm->mode_info_stride;
+      xd->mode_info_context = cm->mi + cm->mode_info_stride * (ithread + 1);
+      xd->mode_info_stride = cm->mode_info_stride;
 
-            for (mb_row = ithread + 1; mb_row < cm->mb_rows; mb_row += (cpi->encoding_thread_count + 1))
-            {
+      for (mb_row = ithread + 1; mb_row < cm->mb_rows;
+           mb_row += (cpi->encoding_thread_count + 1)) {
+        int recon_yoffset, recon_uvoffset;
+        int mb_col;
+        int ref_fb_idx = cm->lst_fb_idx;
+        int dst_fb_idx = cm->new_fb_idx;
+        int recon_y_stride = cm->yv12_fb[ref_fb_idx].y_stride;
+        int recon_uv_stride = cm->yv12_fb[ref_fb_idx].uv_stride;
+        int map_index = (mb_row * cm->mb_cols);
+        const int *last_row_current_mb_col;
+        int *current_mb_col = &cpi->mt_current_mb_col[mb_row];
 
-                int recon_yoffset, recon_uvoffset;
-                int mb_col;
-                int ref_fb_idx = cm->lst_fb_idx;
-                int dst_fb_idx = cm->new_fb_idx;
-                int recon_y_stride = cm->yv12_fb[ref_fb_idx].y_stride;
-                int recon_uv_stride = cm->yv12_fb[ref_fb_idx].uv_stride;
-                int map_index = (mb_row * cm->mb_cols);
-                const int *last_row_current_mb_col;
-                int *current_mb_col = &cpi->mt_current_mb_col[mb_row];
-
-#if  (CONFIG_REALTIME_ONLY & CONFIG_ONTHEFLY_BITPACKING)
-                vp8_writer *w = &cpi->bc[1 + (mb_row % num_part)];
+#if (CONFIG_REALTIME_ONLY & CONFIG_ONTHEFLY_BITPACKING)
+        vp8_writer *w = &cpi->bc[1 + (mb_row % num_part)];
 #else
-                tp = cpi->tok + (mb_row * (cm->mb_cols * 16 * 24));
-                cpi->tplist[mb_row].start = tp;
+        tp = cpi->tok + (mb_row * (cm->mb_cols * 16 * 24));
+        cpi->tplist[mb_row].start = tp;
 #endif
 
-                last_row_current_mb_col = &cpi->mt_current_mb_col[mb_row - 1];
+        last_row_current_mb_col = &cpi->mt_current_mb_col[mb_row - 1];
 
-                /* reset above block coeffs */
-                xd->above_context = cm->above_context;
-                xd->left_context = &mb_row_left_context;
+        /* reset above block coeffs */
+        xd->above_context = cm->above_context;
+        xd->left_context = &mb_row_left_context;
 
-                vp8_zero(mb_row_left_context);
+        vp8_zero(mb_row_left_context);
 
-                xd->up_available = (mb_row != 0);
-                recon_yoffset = (mb_row * recon_y_stride * 16);
-                recon_uvoffset = (mb_row * recon_uv_stride * 8);
+        xd->up_available = (mb_row != 0);
+        recon_yoffset = (mb_row * recon_y_stride * 16);
+        recon_uvoffset = (mb_row * recon_uv_stride * 8);
 
-                /* Set the mb activity pointer to the start of the row. */
-                x->mb_activity_ptr = &cpi->mb_activity_map[map_index];
+        /* Set the mb activity pointer to the start of the row. */
+        x->mb_activity_ptr = &cpi->mb_activity_map[map_index];
 
-                /* for each macroblock col in image */
-                for (mb_col = 0; mb_col < cm->mb_cols; mb_col++)
-                {
-                    if (((mb_col - 1) % nsync) == 0) {
-                        pthread_mutex_t *mutex = &cpi->pmutex[mb_row];
-                        protected_write(mutex, current_mb_col, mb_col - 1);
-                    }
+        /* for each macroblock col in image */
+        for (mb_col = 0; mb_col < cm->mb_cols; mb_col++) {
+          if (((mb_col - 1) % nsync) == 0) {
+            pthread_mutex_t *mutex = &cpi->pmutex[mb_row];
+            protected_write(mutex, current_mb_col, mb_col - 1);
+          }
 
-                    if (mb_row && !(mb_col & (nsync - 1))) {
-                      pthread_mutex_t *mutex = &cpi->pmutex[mb_row-1];
-                      sync_read(mutex, mb_col, last_row_current_mb_col, nsync);
-                    }
+          if (mb_row && !(mb_col & (nsync - 1))) {
+            pthread_mutex_t *mutex = &cpi->pmutex[mb_row - 1];
+            sync_read(mutex, mb_col, last_row_current_mb_col, nsync);
+          }
 
 #if CONFIG_REALTIME_ONLY & CONFIG_ONTHEFLY_BITPACKING
-                    tp = tp_start;
+          tp = tp_start;
 #endif
 
-                    /* Distance of Mb to the various image edges.
-                     * These specified to 8th pel as they are always compared
-                     * to values that are in 1/8th pel units
-                     */
-                    xd->mb_to_left_edge = -((mb_col * 16) << 3);
-                    xd->mb_to_right_edge = ((cm->mb_cols - 1 - mb_col) * 16) << 3;
-                    xd->mb_to_top_edge = -((mb_row * 16) << 3);
-                    xd->mb_to_bottom_edge = ((cm->mb_rows - 1 - mb_row) * 16) << 3;
+          /* Distance of Mb to the various image edges.
+           * These specified to 8th pel as they are always compared
+           * to values that are in 1/8th pel units
+           */
+          xd->mb_to_left_edge = -((mb_col * 16) << 3);
+          xd->mb_to_right_edge = ((cm->mb_cols - 1 - mb_col) * 16) << 3;
+          xd->mb_to_top_edge = -((mb_row * 16) << 3);
+          xd->mb_to_bottom_edge = ((cm->mb_rows - 1 - mb_row) * 16) << 3;
 
-                    /* Set up limit values for motion vectors used to prevent
-                     * them extending outside the UMV borders
-                     */
-                    x->mv_col_min = -((mb_col * 16) + (VP8BORDERINPIXELS - 16));
-                    x->mv_col_max = ((cm->mb_cols - 1 - mb_col) * 16) + (VP8BORDERINPIXELS - 16);
-                    x->mv_row_min = -((mb_row * 16) + (VP8BORDERINPIXELS - 16));
-                    x->mv_row_max = ((cm->mb_rows - 1 - mb_row) * 16) + (VP8BORDERINPIXELS - 16);
+          /* Set up limit values for motion vectors used to prevent
+           * them extending outside the UMV borders
+           */
+          x->mv_col_min = -((mb_col * 16) + (VP8BORDERINPIXELS - 16));
+          x->mv_col_max =
+              ((cm->mb_cols - 1 - mb_col) * 16) + (VP8BORDERINPIXELS - 16);
+          x->mv_row_min = -((mb_row * 16) + (VP8BORDERINPIXELS - 16));
+          x->mv_row_max =
+              ((cm->mb_rows - 1 - mb_row) * 16) + (VP8BORDERINPIXELS - 16);
 
-                    xd->dst.y_buffer = cm->yv12_fb[dst_fb_idx].y_buffer + recon_yoffset;
-                    xd->dst.u_buffer = cm->yv12_fb[dst_fb_idx].u_buffer + recon_uvoffset;
-                    xd->dst.v_buffer = cm->yv12_fb[dst_fb_idx].v_buffer + recon_uvoffset;
-                    xd->left_available = (mb_col != 0);
+          xd->dst.y_buffer = cm->yv12_fb[dst_fb_idx].y_buffer + recon_yoffset;
+          xd->dst.u_buffer = cm->yv12_fb[dst_fb_idx].u_buffer + recon_uvoffset;
+          xd->dst.v_buffer = cm->yv12_fb[dst_fb_idx].v_buffer + recon_uvoffset;
+          xd->left_available = (mb_col != 0);
 
-                    x->rddiv = cpi->RDDIV;
-                    x->rdmult = cpi->RDMULT;
+          x->rddiv = cpi->RDDIV;
+          x->rdmult = cpi->RDMULT;
 
-                    /* Copy current mb to a buffer */
-                    vp8_copy_mem16x16(x->src.y_buffer, x->src.y_stride, x->thismb, 16);
+          /* Copy current mb to a buffer */
+          vp8_copy_mem16x16(x->src.y_buffer, x->src.y_stride, x->thismb, 16);
 
-                    if (cpi->oxcf.tuning == VP8_TUNE_SSIM)
-                        vp8_activity_masking(cpi, x);
+          if (cpi->oxcf.tuning == VP8_TUNE_SSIM) vp8_activity_masking(cpi, x);
 
-                    /* Is segmentation enabled */
-                    /* MB level adjustment to quantizer */
-                    if (xd->segmentation_enabled)
-                    {
-                        /* Code to set segment id in xd->mbmi.segment_id for
-                         * current MB (with range checking)
-                         */
-                        if (cpi->segmentation_map[map_index + mb_col] <= 3)
-                            xd->mode_info_context->mbmi.segment_id = cpi->segmentation_map[map_index + mb_col];
-                        else
-                            xd->mode_info_context->mbmi.segment_id = 0;
+          /* Is segmentation enabled */
+          /* MB level adjustment to quantizer */
+          if (xd->segmentation_enabled) {
+            /* Code to set segment id in xd->mbmi.segment_id for
+             * current MB (with range checking)
+             */
+            if (cpi->segmentation_map[map_index + mb_col] <= 3)
+              xd->mode_info_context->mbmi.segment_id =
+                  cpi->segmentation_map[map_index + mb_col];
+            else
+              xd->mode_info_context->mbmi.segment_id = 0;
+
+            vp8cx_mb_init_quantizer(cpi, x, 1);
+          } else
+            /* Set to Segment 0 by default */
+            xd->mode_info_context->mbmi.segment_id = 0;
 
-                        vp8cx_mb_init_quantizer(cpi, x, 1);
-                    }
-                    else
-                        /* Set to Segment 0 by default */
-                        xd->mode_info_context->mbmi.segment_id = 0;
+          x->active_ptr = cpi->active_map + map_index + mb_col;
 
-                    x->active_ptr = cpi->active_map + map_index + mb_col;
-
-                    if (cm->frame_type == KEY_FRAME)
-                    {
-                        *totalrate += vp8cx_encode_intra_macroblock(cpi, x, &tp);
+          if (cm->frame_type == KEY_FRAME) {
+            *totalrate += vp8cx_encode_intra_macroblock(cpi, x, &tp);
 #ifdef MODE_STATS
-                        y_modes[xd->mbmi.mode] ++;
+            y_modes[xd->mbmi.mode]++;
 #endif
-                    }
-                    else
-                    {
-                        *totalrate += vp8cx_encode_inter_macroblock(cpi, x, &tp, recon_yoffset, recon_uvoffset, mb_row, mb_col);
+          } else {
+            *totalrate += vp8cx_encode_inter_macroblock(
+                cpi, x, &tp, recon_yoffset, recon_uvoffset, mb_row, mb_col);
 
 #ifdef MODE_STATS
-                        inter_y_modes[xd->mbmi.mode] ++;
+            inter_y_modes[xd->mbmi.mode]++;
 
-                        if (xd->mbmi.mode == SPLITMV)
-                        {
-                            int b;
+            if (xd->mbmi.mode == SPLITMV) {
+              int b;
 
-                            for (b = 0; b < xd->mbmi.partition_count; b++)
-                            {
-                                inter_b_modes[x->partition->bmi[b].mode] ++;
-                            }
-                        }
+              for (b = 0; b < xd->mbmi.partition_count; b++) {
+                inter_b_modes[x->partition->bmi[b].mode]++;
+              }
+            }
 
 #endif
-                        // Keep track of how many (consecutive) times a  block
-                        // is coded as ZEROMV_LASTREF, for base layer frames.
-                        // Reset to 0 if its coded as anything else.
-                        if (cpi->current_layer == 0) {
-                          if (xd->mode_info_context->mbmi.mode == ZEROMV &&
-                              xd->mode_info_context->mbmi.ref_frame ==
-                                  LAST_FRAME) {
-                            // Increment, check for wrap-around.
-                            if (cpi->consec_zero_last[map_index+mb_col] < 255)
-                              cpi->consec_zero_last[map_index+mb_col] += 1;
-                            if (cpi->consec_zero_last_mvbias[map_index+mb_col] < 255)
-                              cpi->consec_zero_last_mvbias[map_index+mb_col] += 1;
-                          } else {
-                            cpi->consec_zero_last[map_index+mb_col] = 0;
-                            cpi->consec_zero_last_mvbias[map_index+mb_col] = 0;
-                          }
-                          if (x->zero_last_dot_suppress)
-                            cpi->consec_zero_last_mvbias[map_index+mb_col] = 0;
-                        }
+            // Keep track of how many (consecutive) times a  block
+            // is coded as ZEROMV_LASTREF, for base layer frames.
+            // Reset to 0 if its coded as anything else.
+            if (cpi->current_layer == 0) {
+              if (xd->mode_info_context->mbmi.mode == ZEROMV &&
+                  xd->mode_info_context->mbmi.ref_frame == LAST_FRAME) {
+                // Increment, check for wrap-around.
+                if (cpi->consec_zero_last[map_index + mb_col] < 255)
+                  cpi->consec_zero_last[map_index + mb_col] += 1;
+                if (cpi->consec_zero_last_mvbias[map_index + mb_col] < 255)
+                  cpi->consec_zero_last_mvbias[map_index + mb_col] += 1;
+              } else {
+                cpi->consec_zero_last[map_index + mb_col] = 0;
+                cpi->consec_zero_last_mvbias[map_index + mb_col] = 0;
+              }
+              if (x->zero_last_dot_suppress)
+                cpi->consec_zero_last_mvbias[map_index + mb_col] = 0;
+            }
 
-                        /* Special case code for cyclic refresh
-                         * If cyclic update enabled then copy
-                         * xd->mbmi.segment_id; (which may have been updated
-                         * based on mode during
-                         * vp8cx_encode_inter_macroblock()) back into the
-                         * global segmentation map
-                         */
-                        if ((cpi->current_layer == 0) &&
-                            (cpi->cyclic_refresh_mode_enabled &&
-                             xd->segmentation_enabled))
-                        {
-                            const MB_MODE_INFO * mbmi = &xd->mode_info_context->mbmi;
-                            cpi->segmentation_map[map_index + mb_col] = mbmi->segment_id;
-
-                            /* If the block has been refreshed mark it as clean
-                             * (the magnitude of the -ve influences how long it
-                             * will be before we consider another refresh):
-                             * Else if it was coded (last frame 0,0) and has
-                             * not already been refreshed then mark it as a
-                             * candidate for cleanup next time (marked 0) else
-                             * mark it as dirty (1).
-                             */
-                            if (mbmi->segment_id)
-                                cpi->cyclic_refresh_map[map_index + mb_col] = -1;
-                            else if ((mbmi->mode == ZEROMV) && (mbmi->ref_frame == LAST_FRAME))
-                            {
-                                if (cpi->cyclic_refresh_map[map_index + mb_col] == 1)
-                                    cpi->cyclic_refresh_map[map_index + mb_col] = 0;
-                            }
-                            else
-                                cpi->cyclic_refresh_map[map_index + mb_col] = 1;
+            /* Special case code for cyclic refresh
+             * If cyclic update enabled then copy
+             * xd->mbmi.segment_id; (which may have been updated
+             * based on mode during
+             * vp8cx_encode_inter_macroblock()) back into the
+             * global segmentation map
+             */
+            if ((cpi->current_layer == 0) &&
+                (cpi->cyclic_refresh_mode_enabled &&
+                 xd->segmentation_enabled)) {
+              const MB_MODE_INFO *mbmi = &xd->mode_info_context->mbmi;
+              cpi->segmentation_map[map_index + mb_col] = mbmi->segment_id;
+
+              /* If the block has been refreshed mark it as clean
+               * (the magnitude of the -ve influences how long it
+               * will be before we consider another refresh):
+               * Else if it was coded (last frame 0,0) and has
+               * not already been refreshed then mark it as a
+               * candidate for cleanup next time (marked 0) else
+               * mark it as dirty (1).
+               */
+              if (mbmi->segment_id)
+                cpi->cyclic_refresh_map[map_index + mb_col] = -1;
+              else if ((mbmi->mode == ZEROMV) &&
+                       (mbmi->ref_frame == LAST_FRAME)) {
+                if (cpi->cyclic_refresh_map[map_index + mb_col] == 1)
+                  cpi->cyclic_refresh_map[map_index + mb_col] = 0;
+              } else
+                cpi->cyclic_refresh_map[map_index + mb_col] = 1;
+            }
+          }
 
-                        }
-                    }
-
 #if CONFIG_REALTIME_ONLY & CONFIG_ONTHEFLY_BITPACKING
-                    /* pack tokens for this MB */
-                    {
-                        int tok_count = tp - tp_start;
-                        vp8_pack_tokens(w, tp_start, tok_count);
-                    }
+          /* pack tokens for this MB */
+          {
+            int tok_count = tp - tp_start;
+            vp8_pack_tokens(w, tp_start, tok_count);
+          }
 #else
-                    cpi->tplist[mb_row].stop = tp;
+          cpi->tplist[mb_row].stop = tp;
 #endif
-                    /* Increment pointer into gf usage flags structure. */
-                    x->gf_active_ptr++;
+          /* Increment pointer into gf usage flags structure. */
+          x->gf_active_ptr++;
 
-                    /* Increment the activity mask pointers. */
-                    x->mb_activity_ptr++;
+          /* Increment the activity mask pointers. */
+          x->mb_activity_ptr++;
 
-                    /* adjust to the next column of macroblocks */
-                    x->src.y_buffer += 16;
-                    x->src.u_buffer += 8;
-                    x->src.v_buffer += 8;
+          /* adjust to the next column of macroblocks */
+          x->src.y_buffer += 16;
+          x->src.u_buffer += 8;
+          x->src.v_buffer += 8;
 
-                    recon_yoffset += 16;
-                    recon_uvoffset += 8;
+          recon_yoffset += 16;
+          recon_uvoffset += 8;
 
-                    /* Keep track of segment usage */
-                    segment_counts[xd->mode_info_context->mbmi.segment_id]++;
+          /* Keep track of segment usage */
+          segment_counts[xd->mode_info_context->mbmi.segment_id]++;
 
-                    /* skip to next mb */
-                    xd->mode_info_context++;
-                    x->partition_info++;
-                    xd->above_context++;
-                }
+          /* skip to next mb */
+          xd->mode_info_context++;
+          x->partition_info++;
+          xd->above_context++;
+        }
 
-                vp8_extend_mb_row( &cm->yv12_fb[dst_fb_idx],
-                                    xd->dst.y_buffer + 16,
-                                    xd->dst.u_buffer + 8,
-                                    xd->dst.v_buffer + 8);
+        vp8_extend_mb_row(&cm->yv12_fb[dst_fb_idx], xd->dst.y_buffer + 16,
+                          xd->dst.u_buffer + 8, xd->dst.v_buffer + 8);
 
-                protected_write(&cpi->pmutex[mb_row], current_mb_col,
-                                mb_col + nsync);
+        protected_write(&cpi->pmutex[mb_row], current_mb_col, mb_col + nsync);
 
-                /* this is to account for the border */
-                xd->mode_info_context++;
-                x->partition_info++;
+        /* this is to account for the border */
+        xd->mode_info_context++;
+        x->partition_info++;
 
-                x->src.y_buffer += 16 * x->src.y_stride * (cpi->encoding_thread_count + 1) - 16 * cm->mb_cols;
-                x->src.u_buffer += 8 * x->src.uv_stride * (cpi->encoding_thread_count + 1) - 8 * cm->mb_cols;
-                x->src.v_buffer += 8 * x->src.uv_stride * (cpi->encoding_thread_count + 1) - 8 * cm->mb_cols;
+        x->src.y_buffer +=
+            16 * x->src.y_stride * (cpi->encoding_thread_count + 1) -
+            16 * cm->mb_cols;
+        x->src.u_buffer +=
+            8 * x->src.uv_stride * (cpi->encoding_thread_count + 1) -
+            8 * cm->mb_cols;
+        x->src.v_buffer +=
+            8 * x->src.uv_stride * (cpi->encoding_thread_count + 1) -
+            8 * cm->mb_cols;
 
-                xd->mode_info_context += xd->mode_info_stride * cpi->encoding_thread_count;
-                x->partition_info += xd->mode_info_stride * cpi->encoding_thread_count;
-                x->gf_active_ptr   += cm->mb_cols * cpi->encoding_thread_count;
+        xd->mode_info_context +=
+            xd->mode_info_stride * cpi->encoding_thread_count;
+        x->partition_info += xd->mode_info_stride * cpi->encoding_thread_count;
+        x->gf_active_ptr += cm->mb_cols * cpi->encoding_thread_count;
 
-                if (mb_row == cm->mb_rows - 1)
-                {
-                    sem_post(&cpi->h_event_end_encoding); /* signal frame encoding end */
-                }
-            }
+        if (mb_row == cm->mb_rows - 1) {
+          sem_post(&cpi->h_event_end_encoding); /* signal frame encoding end */
         }
+      }
     }
+  }
 
-    /* printf("exit thread %d\n", ithread); */
-    return 0;
+  /* printf("exit thread %d\n", ithread); */
+  return 0;
 }
 
-static void setup_mbby_copy(MACROBLOCK *mbdst, MACROBLOCK *mbsrc)
-{
+static void setup_mbby_copy(MACROBLOCK *mbdst, MACROBLOCK *mbsrc) {
+  MACROBLOCK *x = mbsrc;
+  MACROBLOCK *z = mbdst;
+  int i;
 
-    MACROBLOCK *x = mbsrc;
-    MACROBLOCK *z = mbdst;
-    int i;
+  z->ss = x->ss;
+  z->ss_count = x->ss_count;
+  z->searches_per_step = x->searches_per_step;
+  z->errorperbit = x->errorperbit;
 
-    z->ss               = x->ss;
-    z->ss_count          = x->ss_count;
-    z->searches_per_step  = x->searches_per_step;
-    z->errorperbit      = x->errorperbit;
+  z->sadperbit16 = x->sadperbit16;
+  z->sadperbit4 = x->sadperbit4;
 
-    z->sadperbit16      = x->sadperbit16;
-    z->sadperbit4       = x->sadperbit4;
+  /*
+  z->mv_col_min    = x->mv_col_min;
+  z->mv_col_max    = x->mv_col_max;
+  z->mv_row_min    = x->mv_row_min;
+  z->mv_row_max    = x->mv_row_max;
+  */
 
-    /*
-    z->mv_col_min    = x->mv_col_min;
-    z->mv_col_max    = x->mv_col_max;
-    z->mv_row_min    = x->mv_row_min;
-    z->mv_row_max    = x->mv_row_max;
-    */
+  z->short_fdct4x4 = x->short_fdct4x4;
+  z->short_fdct8x4 = x->short_fdct8x4;
+  z->short_walsh4x4 = x->short_walsh4x4;
+  z->quantize_b = x->quantize_b;
+  z->optimize = x->optimize;
 
-    z->short_fdct4x4     = x->short_fdct4x4;
-    z->short_fdct8x4     = x->short_fdct8x4;
-    z->short_walsh4x4    = x->short_walsh4x4;
-    z->quantize_b        = x->quantize_b;
-    z->optimize          = x->optimize;
+  /*
+  z->mvc              = x->mvc;
+  z->src.y_buffer      = x->src.y_buffer;
+  z->src.u_buffer      = x->src.u_buffer;
+  z->src.v_buffer      = x->src.v_buffer;
+  */
 
-    /*
-    z->mvc              = x->mvc;
-    z->src.y_buffer      = x->src.y_buffer;
-    z->src.u_buffer      = x->src.u_buffer;
-    z->src.v_buffer      = x->src.v_buffer;
-    */
+  z->mvcost[0] = x->mvcost[0];
+  z->mvcost[1] = x->mvcost[1];
+  z->mvsadcost[0] = x->mvsadcost[0];
+  z->mvsadcost[1] = x->mvsadcost[1];
 
-    z->mvcost[0] =  x->mvcost[0];
-    z->mvcost[1] =  x->mvcost[1];
-    z->mvsadcost[0] =  x->mvsadcost[0];
-    z->mvsadcost[1] =  x->mvsadcost[1];
+  z->token_costs = x->token_costs;
+  z->inter_bmode_costs = x->inter_bmode_costs;
+  z->mbmode_cost = x->mbmode_cost;
+  z->intra_uv_mode_cost = x->intra_uv_mode_cost;
+  z->bmode_costs = x->bmode_costs;
 
-    z->token_costs = x->token_costs;
-    z->inter_bmode_costs = x->inter_bmode_costs;
-    z->mbmode_cost = x->mbmode_cost;
-    z->intra_uv_mode_cost = x->intra_uv_mode_cost;
-    z->bmode_costs = x->bmode_costs;
+  for (i = 0; i < 25; i++) {
+    z->block[i].quant = x->block[i].quant;
+    z->block[i].quant_fast = x->block[i].quant_fast;
+    z->block[i].quant_shift = x->block[i].quant_shift;
+    z->block[i].zbin = x->block[i].zbin;
+    z->block[i].zrun_zbin_boost = x->block[i].zrun_zbin_boost;
+    z->block[i].round = x->block[i].round;
+    z->block[i].src_stride = x->block[i].src_stride;
+  }
 
-    for (i = 0; i < 25; i++)
-    {
-        z->block[i].quant           = x->block[i].quant;
-        z->block[i].quant_fast      = x->block[i].quant_fast;
-        z->block[i].quant_shift     = x->block[i].quant_shift;
-        z->block[i].zbin            = x->block[i].zbin;
-        z->block[i].zrun_zbin_boost = x->block[i].zrun_zbin_boost;
-        z->block[i].round           = x->block[i].round;
-        z->block[i].src_stride      = x->block[i].src_stride;
-    }
+  z->q_index = x->q_index;
+  z->act_zbin_adj = x->act_zbin_adj;
+  z->last_act_zbin_adj = x->last_act_zbin_adj;
 
-    z->q_index           = x->q_index;
-    z->act_zbin_adj      = x->act_zbin_adj;
-    z->last_act_zbin_adj = x->last_act_zbin_adj;
+  {
+    MACROBLOCKD *xd = &x->e_mbd;
+    MACROBLOCKD *zd = &z->e_mbd;
 
-    {
-        MACROBLOCKD *xd = &x->e_mbd;
-        MACROBLOCKD *zd = &z->e_mbd;
+    /*
+    zd->mode_info_context = xd->mode_info_context;
+    zd->mode_info        = xd->mode_info;
 
-        /*
-        zd->mode_info_context = xd->mode_info_context;
-        zd->mode_info        = xd->mode_info;
+    zd->mode_info_stride  = xd->mode_info_stride;
+    zd->frame_type       = xd->frame_type;
+    zd->up_available     = xd->up_available   ;
+    zd->left_available   = xd->left_available;
+    zd->left_context     = xd->left_context;
+    zd->last_frame_dc     = xd->last_frame_dc;
+    zd->last_frame_dccons = xd->last_frame_dccons;
+    zd->gold_frame_dc     = xd->gold_frame_dc;
+    zd->gold_frame_dccons = xd->gold_frame_dccons;
+    zd->mb_to_left_edge    = xd->mb_to_left_edge;
+    zd->mb_to_right_edge   = xd->mb_to_right_edge;
+    zd->mb_to_top_edge     = xd->mb_to_top_edge   ;
+    zd->mb_to_bottom_edge  = xd->mb_to_bottom_edge;
+    zd->gf_active_ptr     = xd->gf_active_ptr;
+    zd->frames_since_golden       = xd->frames_since_golden;
+    zd->frames_till_alt_ref_frame   = xd->frames_till_alt_ref_frame;
+    */
+    zd->subpixel_predict = xd->subpixel_predict;
+    zd->subpixel_predict8x4 = xd->subpixel_predict8x4;
+    zd->subpixel_predict8x8 = xd->subpixel_predict8x8;
+    zd->subpixel_predict16x16 = xd->subpixel_predict16x16;
+    zd->segmentation_enabled = xd->segmentation_enabled;
+    zd->mb_segement_abs_delta = xd->mb_segement_abs_delta;
+    memcpy(zd->segment_feature_data, xd->segment_feature_data,
+           sizeof(xd->segment_feature_data));
 
-        zd->mode_info_stride  = xd->mode_info_stride;
-        zd->frame_type       = xd->frame_type;
-        zd->up_available     = xd->up_available   ;
-        zd->left_available   = xd->left_available;
-        zd->left_context     = xd->left_context;
-        zd->last_frame_dc     = xd->last_frame_dc;
-        zd->last_frame_dccons = xd->last_frame_dccons;
-        zd->gold_frame_dc     = xd->gold_frame_dc;
-        zd->gold_frame_dccons = xd->gold_frame_dccons;
-        zd->mb_to_left_edge    = xd->mb_to_left_edge;
-        zd->mb_to_right_edge   = xd->mb_to_right_edge;
-        zd->mb_to_top_edge     = xd->mb_to_top_edge   ;
-        zd->mb_to_bottom_edge  = xd->mb_to_bottom_edge;
-        zd->gf_active_ptr     = xd->gf_active_ptr;
-        zd->frames_since_golden       = xd->frames_since_golden;
-        zd->frames_till_alt_ref_frame   = xd->frames_till_alt_ref_frame;
-        */
-        zd->subpixel_predict         = xd->subpixel_predict;
-        zd->subpixel_predict8x4      = xd->subpixel_predict8x4;
-        zd->subpixel_predict8x8      = xd->subpixel_predict8x8;
-        zd->subpixel_predict16x16    = xd->subpixel_predict16x16;
-        zd->segmentation_enabled     = xd->segmentation_enabled;
-        zd->mb_segement_abs_delta      = xd->mb_segement_abs_delta;
-        memcpy(zd->segment_feature_data, xd->segment_feature_data,
-               sizeof(xd->segment_feature_data));
+    memcpy(zd->dequant_y1_dc, xd->dequant_y1_dc, sizeof(xd->dequant_y1_dc));
+    memcpy(zd->dequant_y1, xd->dequant_y1, sizeof(xd->dequant_y1));
+    memcpy(zd->dequant_y2, xd->dequant_y2, sizeof(xd->dequant_y2));
+    memcpy(zd->dequant_uv, xd->dequant_uv, sizeof(xd->dequant_uv));
 
-        memcpy(zd->dequant_y1_dc, xd->dequant_y1_dc, sizeof(xd->dequant_y1_dc));
-        memcpy(zd->dequant_y1, xd->dequant_y1, sizeof(xd->dequant_y1));
-        memcpy(zd->dequant_y2, xd->dequant_y2, sizeof(xd->dequant_y2));
-        memcpy(zd->dequant_uv, xd->dequant_uv, sizeof(xd->dequant_uv));
-
 #if 1
-        /*TODO:  Remove dequant from BLOCKD.  This is a temporary solution until
-         * the quantizer code uses a passed in pointer to the dequant constants.
-         * This will also require modifications to the x86 and neon assembly.
-         * */
-        for (i = 0; i < 16; i++)
-            zd->block[i].dequant = zd->dequant_y1;
-        for (i = 16; i < 24; i++)
-            zd->block[i].dequant = zd->dequant_uv;
-        zd->block[24].dequant = zd->dequant_y2;
+    /*TODO:  Remove dequant from BLOCKD.  This is a temporary solution until
+     * the quantizer code uses a passed in pointer to the dequant constants.
+     * This will also require modifications to the x86 and neon assembly.
+     * */
+    for (i = 0; i < 16; i++) zd->block[i].dequant = zd->dequant_y1;
+    for (i = 16; i < 24; i++) zd->block[i].dequant = zd->dequant_uv;
+    zd->block[24].dequant = zd->dequant_y2;
 #endif
 
+    memcpy(z->rd_threshes, x->rd_threshes, sizeof(x->rd_threshes));
+    memcpy(z->rd_thresh_mult, x->rd_thresh_mult, sizeof(x->rd_thresh_mult));
 
-        memcpy(z->rd_threshes, x->rd_threshes, sizeof(x->rd_threshes));
-        memcpy(z->rd_thresh_mult, x->rd_thresh_mult, sizeof(x->rd_thresh_mult));
+    z->zbin_over_quant = x->zbin_over_quant;
+    z->zbin_mode_boost_enabled = x->zbin_mode_boost_enabled;
+    z->zbin_mode_boost = x->zbin_mode_boost;
 
-        z->zbin_over_quant = x->zbin_over_quant;
-        z->zbin_mode_boost_enabled = x->zbin_mode_boost_enabled;
-        z->zbin_mode_boost = x->zbin_mode_boost;
-
-        memset(z->error_bins, 0, sizeof(z->error_bins));
-    }
+    memset(z->error_bins, 0, sizeof(z->error_bins));
+  }
 }
 
-void vp8cx_init_mbrthread_data(VP8_COMP *cpi,
-                               MACROBLOCK *x,
-                               MB_ROW_COMP *mbr_ei,
-                               int count
-                              )
-{
+void vp8cx_init_mbrthread_data(VP8_COMP *cpi, MACROBLOCK *x,
+                               MB_ROW_COMP *mbr_ei, int count) {
+  VP8_COMMON *const cm = &cpi->common;
+  MACROBLOCKD *const xd = &x->e_mbd;
+  int i;
 
-    VP8_COMMON *const cm = & cpi->common;
-    MACROBLOCKD *const xd = & x->e_mbd;
-    int i;
+  for (i = 0; i < count; i++) {
+    MACROBLOCK *mb = &mbr_ei[i].mb;
+    MACROBLOCKD *mbd = &mb->e_mbd;
 
-    for (i = 0; i < count; i++)
-    {
-        MACROBLOCK *mb = & mbr_ei[i].mb;
-        MACROBLOCKD *mbd = &mb->e_mbd;
+    mbd->subpixel_predict = xd->subpixel_predict;
+    mbd->subpixel_predict8x4 = xd->subpixel_predict8x4;
+    mbd->subpixel_predict8x8 = xd->subpixel_predict8x8;
+    mbd->subpixel_predict16x16 = xd->subpixel_predict16x16;
+    mb->gf_active_ptr = x->gf_active_ptr;
 
-        mbd->subpixel_predict        = xd->subpixel_predict;
-        mbd->subpixel_predict8x4     = xd->subpixel_predict8x4;
-        mbd->subpixel_predict8x8     = xd->subpixel_predict8x8;
-        mbd->subpixel_predict16x16   = xd->subpixel_predict16x16;
-        mb->gf_active_ptr            = x->gf_active_ptr;
+    memset(mbr_ei[i].segment_counts, 0, sizeof(mbr_ei[i].segment_counts));
+    mbr_ei[i].totalrate = 0;
 
-        memset(mbr_ei[i].segment_counts, 0, sizeof(mbr_ei[i].segment_counts));
-        mbr_ei[i].totalrate = 0;
+    mb->partition_info = x->pi + x->e_mbd.mode_info_stride * (i + 1);
 
-        mb->partition_info = x->pi + x->e_mbd.mode_info_stride * (i + 1);
+    mbd->frame_type = cm->frame_type;
 
-        mbd->frame_type = cm->frame_type;
+    mb->src = *cpi->Source;
+    mbd->pre = cm->yv12_fb[cm->lst_fb_idx];
+    mbd->dst = cm->yv12_fb[cm->new_fb_idx];
 
-        mb->src = * cpi->Source;
-        mbd->pre = cm->yv12_fb[cm->lst_fb_idx];
-        mbd->dst = cm->yv12_fb[cm->new_fb_idx];
+    mb->src.y_buffer += 16 * x->src.y_stride * (i + 1);
+    mb->src.u_buffer += 8 * x->src.uv_stride * (i + 1);
+    mb->src.v_buffer += 8 * x->src.uv_stride * (i + 1);
 
-        mb->src.y_buffer += 16 * x->src.y_stride * (i + 1);
-        mb->src.u_buffer +=  8 * x->src.uv_stride * (i + 1);
-        mb->src.v_buffer +=  8 * x->src.uv_stride * (i + 1);
+    vp8_build_block_offsets(mb);
 
-        vp8_build_block_offsets(mb);
+    mbd->left_context = &cm->left_context;
+    mb->mvc = cm->fc.mvc;
 
-        mbd->left_context = &cm->left_context;
-        mb->mvc = cm->fc.mvc;
+    setup_mbby_copy(&mbr_ei[i].mb, x);
 
-        setup_mbby_copy(&mbr_ei[i].mb, x);
+    mbd->fullpixel_mask = 0xffffffff;
+    if (cm->full_pixel) mbd->fullpixel_mask = 0xfffffff8;
 
-        mbd->fullpixel_mask = 0xffffffff;
-        if(cm->full_pixel)
-            mbd->fullpixel_mask = 0xfffffff8;
-
-        vp8_zero(mb->coef_counts);
-        vp8_zero(x->ymode_count);
-        mb->skip_true_count = 0;
-        vp8_zero(mb->MVcount);
-        mb->prediction_error = 0;
-        mb->intra_error = 0;
-        vp8_zero(mb->count_mb_ref_frame_usage);
-        mb->mbs_tested_so_far = 0;
-        mb->mbs_zero_last_dot_suppress = 0;
-    }
+    vp8_zero(mb->coef_counts);
+    vp8_zero(x->ymode_count);
+    mb->skip_true_count = 0;
+    vp8_zero(mb->MVcount);
+    mb->prediction_error = 0;
+    mb->intra_error = 0;
+    vp8_zero(mb->count_mb_ref_frame_usage);
+    mb->mbs_tested_so_far = 0;
+    mb->mbs_zero_last_dot_suppress = 0;
+  }
 }
 
-int vp8cx_create_encoder_threads(VP8_COMP *cpi)
-{
-    const VP8_COMMON * cm = &cpi->common;
+int vp8cx_create_encoder_threads(VP8_COMP *cpi) {
+  const VP8_COMMON *cm = &cpi->common;
 
-    cpi->b_multi_threaded = 0;
-    cpi->encoding_thread_count = 0;
+  cpi->b_multi_threaded = 0;
+  cpi->encoding_thread_count = 0;
 
-    pthread_mutex_init(&cpi->mt_mutex, NULL);
+  pthread_mutex_init(&cpi->mt_mutex, NULL);
 
-    if (cm->processor_core_count > 1 && cpi->oxcf.multi_threaded > 1)
-    {
-        int ithread;
-        int th_count = cpi->oxcf.multi_threaded - 1;
-        int rc = 0;
+  if (cm->processor_core_count > 1 && cpi->oxcf.multi_threaded > 1) {
+    int ithread;
+    int th_count = cpi->oxcf.multi_threaded - 1;
+    int rc = 0;
 
-        /* don't allocate more threads than cores available */
-        if (cpi->oxcf.multi_threaded > cm->processor_core_count)
-            th_count = cm->processor_core_count - 1;
+    /* don't allocate more threads than cores available */
+    if (cpi->oxcf.multi_threaded > cm->processor_core_count)
+      th_count = cm->processor_core_count - 1;
 
-        /* we have th_count + 1 (main) threads processing one row each */
-        /* no point to have more threads than the sync range allows */
-        if(th_count > ((cm->mb_cols / cpi->mt_sync_range) - 1))
-        {
-            th_count = (cm->mb_cols / cpi->mt_sync_range) - 1;
-        }
+    /* we have th_count + 1 (main) threads processing one row each */
+    /* no point to have more threads than the sync range allows */
+    if (th_count > ((cm->mb_cols / cpi->mt_sync_range) - 1)) {
+      th_count = (cm->mb_cols / cpi->mt_sync_range) - 1;
+    }
 
-        if(th_count == 0)
-            return 0;
+    if (th_count == 0) return 0;
 
-        CHECK_MEM_ERROR(cpi->h_encoding_thread,
-                        vpx_malloc(sizeof(pthread_t) * th_count));
-        CHECK_MEM_ERROR(cpi->h_event_start_encoding,
-                        vpx_malloc(sizeof(sem_t) * th_count));
-        CHECK_MEM_ERROR(cpi->mb_row_ei,
-                        vpx_memalign(32, sizeof(MB_ROW_COMP) * th_count));
-        memset(cpi->mb_row_ei, 0, sizeof(MB_ROW_COMP) * th_count);
-        CHECK_MEM_ERROR(cpi->en_thread_data,
-                        vpx_malloc(sizeof(ENCODETHREAD_DATA) * th_count));
+    CHECK_MEM_ERROR(cpi->h_encoding_thread,
+                    vpx_malloc(sizeof(pthread_t) * th_count));
+    CHECK_MEM_ERROR(cpi->h_event_start_encoding,
+                    vpx_malloc(sizeof(sem_t) * th_count));
+    CHECK_MEM_ERROR(cpi->mb_row_ei,
+                    vpx_memalign(32, sizeof(MB_ROW_COMP) * th_count));
+    memset(cpi->mb_row_ei, 0, sizeof(MB_ROW_COMP) * th_count);
+    CHECK_MEM_ERROR(cpi->en_thread_data,
+                    vpx_malloc(sizeof(ENCODETHREAD_DATA) * th_count));
 
-        sem_init(&cpi->h_event_end_encoding, 0, 0);
+    sem_init(&cpi->h_event_end_encoding, 0, 0);
 
-        cpi->b_multi_threaded = 1;
-        cpi->encoding_thread_count = th_count;
+    cpi->b_multi_threaded = 1;
+    cpi->encoding_thread_count = th_count;
 
-        /*
-        printf("[VP8:] multi_threaded encoding is enabled with %d threads\n\n",
-               (cpi->encoding_thread_count +1));
-        */
+    /*
+    printf("[VP8:] multi_threaded encoding is enabled with %d threads\n\n",
+           (cpi->encoding_thread_count +1));
+    */
 
-        for (ithread = 0; ithread < th_count; ithread++)
-        {
-            ENCODETHREAD_DATA *ethd = &cpi->en_thread_data[ithread];
+    for (ithread = 0; ithread < th_count; ithread++) {
+      ENCODETHREAD_DATA *ethd = &cpi->en_thread_data[ithread];
 
-            /* Setup block ptrs and offsets */
-            vp8_setup_block_ptrs(&cpi->mb_row_ei[ithread].mb);
-            vp8_setup_block_dptrs(&cpi->mb_row_ei[ithread].mb.e_mbd);
+      /* Setup block ptrs and offsets */
+      vp8_setup_block_ptrs(&cpi->mb_row_ei[ithread].mb);
+      vp8_setup_block_dptrs(&cpi->mb_row_ei[ithread].mb.e_mbd);
 
-            sem_init(&cpi->h_event_start_encoding[ithread], 0, 0);
+      sem_init(&cpi->h_event_start_encoding[ithread], 0, 0);
 
-            ethd->ithread = ithread;
-            ethd->ptr1 = (void *)cpi;
-            ethd->ptr2 = (void *)&cpi->mb_row_ei[ithread];
+      ethd->ithread = ithread;
+      ethd->ptr1 = (void *)cpi;
+      ethd->ptr2 = (void *)&cpi->mb_row_ei[ithread];
 
-            rc = pthread_create(&cpi->h_encoding_thread[ithread], 0,
-                                thread_encoding_proc, ethd);
-            if(rc)
-                break;
-        }
+      rc = pthread_create(&cpi->h_encoding_thread[ithread], 0,
+                          thread_encoding_proc, ethd);
+      if (rc) break;
+    }
 
-        if(rc)
-        {
-            /* shutdown other threads */
-            protected_write(&cpi->mt_mutex, &cpi->b_multi_threaded, 0);
-            for(--ithread; ithread >= 0; ithread--)
-            {
-                pthread_join(cpi->h_encoding_thread[ithread], 0);
-                sem_destroy(&cpi->h_event_start_encoding[ithread]);
-            }
-            sem_destroy(&cpi->h_event_end_encoding);
+    if (rc) {
+      /* shutdown other threads */
+      protected_write(&cpi->mt_mutex, &cpi->b_multi_threaded, 0);
+      for (--ithread; ithread >= 0; ithread--) {
+        pthread_join(cpi->h_encoding_thread[ithread], 0);
+        sem_destroy(&cpi->h_event_start_encoding[ithread]);
+      }
+      sem_destroy(&cpi->h_event_end_encoding);
 
-            /* free thread related resources */
-            vpx_free(cpi->h_event_start_encoding);
-            vpx_free(cpi->h_encoding_thread);
-            vpx_free(cpi->mb_row_ei);
-            vpx_free(cpi->en_thread_data);
+      /* free thread related resources */
+      vpx_free(cpi->h_event_start_encoding);
+      vpx_free(cpi->h_encoding_thread);
+      vpx_free(cpi->mb_row_ei);
+      vpx_free(cpi->en_thread_data);
 
-            pthread_mutex_destroy(&cpi->mt_mutex);
+      pthread_mutex_destroy(&cpi->mt_mutex);
 
-            return -1;
-        }
-
-
-        {
-            LPFTHREAD_DATA * lpfthd = &cpi->lpf_thread_data;
-
-            sem_init(&cpi->h_event_start_lpf, 0, 0);
-            sem_init(&cpi->h_event_end_lpf, 0, 0);
-
-            lpfthd->ptr1 = (void *)cpi;
-            rc = pthread_create(&cpi->h_filter_thread, 0, thread_loopfilter,
-                                lpfthd);
-
-            if(rc)
-            {
-                /* shutdown other threads */
-                protected_write(&cpi->mt_mutex, &cpi->b_multi_threaded, 0);
-                for(--ithread; ithread >= 0; ithread--)
-                {
-                    sem_post(&cpi->h_event_start_encoding[ithread]);
-                    pthread_join(cpi->h_encoding_thread[ithread], 0);
-                    sem_destroy(&cpi->h_event_start_encoding[ithread]);
-                }
-                sem_destroy(&cpi->h_event_end_encoding);
-                sem_destroy(&cpi->h_event_end_lpf);
-                sem_destroy(&cpi->h_event_start_lpf);
-
-                /* free thread related resources */
-                vpx_free(cpi->h_event_start_encoding);
-                vpx_free(cpi->h_encoding_thread);
-                vpx_free(cpi->mb_row_ei);
-                vpx_free(cpi->en_thread_data);
-
-                pthread_mutex_destroy(&cpi->mt_mutex);
-
-                return -2;
-            }
-        }
+      return -1;
     }
-    return 0;
-}
 
-void vp8cx_remove_encoder_threads(VP8_COMP *cpi)
-{
-    if (protected_read(&cpi->mt_mutex, &cpi->b_multi_threaded))
     {
-        /* shutdown other threads */
-        protected_write(&cpi->mt_mutex, &cpi->b_multi_threaded, 0);
-        {
-            int i;
+      LPFTHREAD_DATA *lpfthd = &cpi->lpf_thread_data;
 
-            for (i = 0; i < cpi->encoding_thread_count; i++)
-            {
-                sem_post(&cpi->h_event_start_encoding[i]);
-                pthread_join(cpi->h_encoding_thread[i], 0);
+      sem_init(&cpi->h_event_start_lpf, 0, 0);
+      sem_init(&cpi->h_event_end_lpf, 0, 0);
 
-                sem_destroy(&cpi->h_event_start_encoding[i]);
-            }
+      lpfthd->ptr1 = (void *)cpi;
+      rc = pthread_create(&cpi->h_filter_thread, 0, thread_loopfilter, lpfthd);
 
-            sem_post(&cpi->h_event_start_lpf);
-            pthread_join(cpi->h_filter_thread, 0);
+      if (rc) {
+        /* shutdown other threads */
+        protected_write(&cpi->mt_mutex, &cpi->b_multi_threaded, 0);
+        for (--ithread; ithread >= 0; ithread--) {
+          sem_post(&cpi->h_event_start_encoding[ithread]);
+          pthread_join(cpi->h_encoding_thread[ithread], 0);
+          sem_destroy(&cpi->h_event_start_encoding[ithread]);
         }
-
         sem_destroy(&cpi->h_event_end_encoding);
         sem_destroy(&cpi->h_event_end_lpf);
         sem_destroy(&cpi->h_event_start_lpf);
@@ -673,7 +590,44 @@
         vpx_free(cpi->h_encoding_thread);
         vpx_free(cpi->mb_row_ei);
         vpx_free(cpi->en_thread_data);
+
+        pthread_mutex_destroy(&cpi->mt_mutex);
+
+        return -2;
+      }
     }
-    pthread_mutex_destroy(&cpi->mt_mutex);
+  }
+  return 0;
+}
+
+void vp8cx_remove_encoder_threads(VP8_COMP *cpi) {
+  if (protected_read(&cpi->mt_mutex, &cpi->b_multi_threaded)) {
+    /* shutdown other threads */
+    protected_write(&cpi->mt_mutex, &cpi->b_multi_threaded, 0);
+    {
+      int i;
+
+      for (i = 0; i < cpi->encoding_thread_count; i++) {
+        sem_post(&cpi->h_event_start_encoding[i]);
+        pthread_join(cpi->h_encoding_thread[i], 0);
+
+        sem_destroy(&cpi->h_event_start_encoding[i]);
+      }
+
+      sem_post(&cpi->h_event_start_lpf);
+      pthread_join(cpi->h_filter_thread, 0);
+    }
+
+    sem_destroy(&cpi->h_event_end_encoding);
+    sem_destroy(&cpi->h_event_end_lpf);
+    sem_destroy(&cpi->h_event_start_lpf);
+
+    /* free thread related resources */
+    vpx_free(cpi->h_event_start_encoding);
+    vpx_free(cpi->h_encoding_thread);
+    vpx_free(cpi->mb_row_ei);
+    vpx_free(cpi->en_thread_data);
+  }
+  pthread_mutex_destroy(&cpi->mt_mutex);
 }
 #endif
--- a/vp8/encoder/firstpass.c
+++ b/vp8/encoder/firstpass.c
@@ -42,36 +42,33 @@
 
 extern const int vp8_gf_boost_qadjustment[QINDEX_RANGE];
 
-#define IIFACTOR   1.5
+#define IIFACTOR 1.5
 #define IIKFACTOR1 1.40
 #define IIKFACTOR2 1.5
-#define RMAX       14.0
-#define GF_RMAX    48.0
+#define RMAX 14.0
+#define GF_RMAX 48.0
 
 #define KF_MB_INTRA_MIN 300
 #define GF_MB_INTRA_MIN 200
 
-#define DOUBLE_DIVIDE_CHECK(X) ((X)<0?(X)-.000001:(X)+.000001)
+#define DOUBLE_DIVIDE_CHECK(X) ((X) < 0 ? (X)-.000001 : (X) + .000001)
 
-#define POW1 (double)cpi->oxcf.two_pass_vbrbias/100.0
-#define POW2 (double)cpi->oxcf.two_pass_vbrbias/100.0
+#define POW1 (double) cpi->oxcf.two_pass_vbrbias / 100.0
+#define POW2 (double) cpi->oxcf.two_pass_vbrbias / 100.0
 
 #define NEW_BOOST 1
 
-static int vscale_lookup[7] = {0, 1, 1, 2, 2, 3, 3};
-static int hscale_lookup[7] = {0, 0, 1, 1, 2, 2, 3};
+static int vscale_lookup[7] = { 0, 1, 1, 2, 2, 3, 3 };
+static int hscale_lookup[7] = { 0, 0, 1, 1, 2, 2, 3 };
 
-
-static const int cq_level[QINDEX_RANGE] =
-{
-    0,0,1,1,2,3,3,4,4,5,6,6,7,8,8,9,
-    9,10,11,11,12,13,13,14,15,15,16,17,17,18,19,20,
-    20,21,22,22,23,24,24,25,26,27,27,28,29,30,30,31,
-    32,33,33,34,35,36,36,37,38,39,39,40,41,42,42,43,
-    44,45,46,46,47,48,49,50,50,51,52,53,54,55,55,56,
-    57,58,59,60,60,61,62,63,64,65,66,67,67,68,69,70,
-    71,72,73,74,75,75,76,77,78,79,80,81,82,83,84,85,
-    86,86,87,88,89,90,91,92,93,94,95,96,97,98,99,100
+static const int cq_level[QINDEX_RANGE] = {
+  0,  0,  1,  1,  2,  3,  3,  4,  4,  5,  6,  6,  7,  8,  8,  9,  9,  10, 11,
+  11, 12, 13, 13, 14, 15, 15, 16, 17, 17, 18, 19, 20, 20, 21, 22, 22, 23, 24,
+  24, 25, 26, 27, 27, 28, 29, 30, 30, 31, 32, 33, 33, 34, 35, 36, 36, 37, 38,
+  39, 39, 40, 41, 42, 42, 43, 44, 45, 46, 46, 47, 48, 49, 50, 50, 51, 52, 53,
+  54, 55, 55, 56, 57, 58, 59, 60, 60, 61, 62, 63, 64, 65, 66, 67, 67, 68, 69,
+  70, 71, 72, 73, 74, 75, 75, 76, 77, 78, 79, 80, 81, 82, 83, 84, 85, 86, 86,
+  87, 88, 89, 90, 91, 92, 93, 94, 95, 96, 97, 98, 99, 100
 };
 
 static void find_next_key_frame(VP8_COMP *cpi, FIRSTPASS_STATS *this_frame);
@@ -79,828 +76,764 @@
 /* Resets the first pass file to the given position using a relative seek
  * from the current position
  */
-static void reset_fpf_position(VP8_COMP *cpi, FIRSTPASS_STATS *Position)
-{
-    cpi->twopass.stats_in = Position;
+static void reset_fpf_position(VP8_COMP *cpi, FIRSTPASS_STATS *Position) {
+  cpi->twopass.stats_in = Position;
 }
 
-static int lookup_next_frame_stats(VP8_COMP *cpi, FIRSTPASS_STATS *next_frame)
-{
-    if (cpi->twopass.stats_in >= cpi->twopass.stats_in_end)
-        return EOF;
+static int lookup_next_frame_stats(VP8_COMP *cpi, FIRSTPASS_STATS *next_frame) {
+  if (cpi->twopass.stats_in >= cpi->twopass.stats_in_end) return EOF;
 
-    *next_frame = *cpi->twopass.stats_in;
-    return 1;
+  *next_frame = *cpi->twopass.stats_in;
+  return 1;
 }
 
 /* Read frame stats at an offset from the current position */
-static int read_frame_stats( VP8_COMP *cpi,
-                             FIRSTPASS_STATS *frame_stats,
-                             int offset )
-{
-    FIRSTPASS_STATS * fps_ptr = cpi->twopass.stats_in;
+static int read_frame_stats(VP8_COMP *cpi, FIRSTPASS_STATS *frame_stats,
+                            int offset) {
+  FIRSTPASS_STATS *fps_ptr = cpi->twopass.stats_in;
 
-    /* Check legality of offset */
-    if ( offset >= 0 )
-    {
-        if ( &fps_ptr[offset] >= cpi->twopass.stats_in_end )
-             return EOF;
-    }
-    else if ( offset < 0 )
-    {
-        if ( &fps_ptr[offset] < cpi->twopass.stats_in_start )
-             return EOF;
-    }
+  /* Check legality of offset */
+  if (offset >= 0) {
+    if (&fps_ptr[offset] >= cpi->twopass.stats_in_end) return EOF;
+  } else if (offset < 0) {
+    if (&fps_ptr[offset] < cpi->twopass.stats_in_start) return EOF;
+  }
 
-    *frame_stats = fps_ptr[offset];
-    return 1;
+  *frame_stats = fps_ptr[offset];
+  return 1;
 }
 
-static int input_stats(VP8_COMP *cpi, FIRSTPASS_STATS *fps)
-{
-    if (cpi->twopass.stats_in >= cpi->twopass.stats_in_end)
-        return EOF;
+static int input_stats(VP8_COMP *cpi, FIRSTPASS_STATS *fps) {
+  if (cpi->twopass.stats_in >= cpi->twopass.stats_in_end) return EOF;
 
-    *fps = *cpi->twopass.stats_in;
-    cpi->twopass.stats_in =
-         (void*)((char *)cpi->twopass.stats_in + sizeof(FIRSTPASS_STATS));
-    return 1;
+  *fps = *cpi->twopass.stats_in;
+  cpi->twopass.stats_in =
+      (void *)((char *)cpi->twopass.stats_in + sizeof(FIRSTPASS_STATS));
+  return 1;
 }
 
-static void output_stats(const VP8_COMP            *cpi,
+static void output_stats(const VP8_COMP *cpi,
                          struct vpx_codec_pkt_list *pktlist,
-                         FIRSTPASS_STATS            *stats)
-{
-    struct vpx_codec_cx_pkt pkt;
-    (void)cpi;
-    pkt.kind = VPX_CODEC_STATS_PKT;
-    pkt.data.twopass_stats.buf = stats;
-    pkt.data.twopass_stats.sz = sizeof(FIRSTPASS_STATS);
-    vpx_codec_pkt_list_add(pktlist, &pkt);
+                         FIRSTPASS_STATS *stats) {
+  struct vpx_codec_cx_pkt pkt;
+  (void)cpi;
+  pkt.kind = VPX_CODEC_STATS_PKT;
+  pkt.data.twopass_stats.buf = stats;
+  pkt.data.twopass_stats.sz = sizeof(FIRSTPASS_STATS);
+  vpx_codec_pkt_list_add(pktlist, &pkt);
 
 /* TEMP debug code */
 #if OUTPUT_FPF
 
-    {
-        FILE *fpfile;
-        fpfile = fopen("firstpass.stt", "a");
+  {
+    FILE *fpfile;
+    fpfile = fopen("firstpass.stt", "a");
 
-        fprintf(fpfile, "%12.0f %12.0f %12.0f %12.4f %12.4f %12.4f %12.4f"
-                " %12.4f %12.4f %12.4f %12.4f %12.4f %12.4f %12.4f %12.4f"
-                " %12.0f %12.0f %12.4f\n",
-                stats->frame,
-                stats->intra_error,
-                stats->coded_error,
-                stats->ssim_weighted_pred_err,
-                stats->pcnt_inter,
-                stats->pcnt_motion,
-                stats->pcnt_second_ref,
-                stats->pcnt_neutral,
-                stats->MVr,
-                stats->mvr_abs,
-                stats->MVc,
-                stats->mvc_abs,
-                stats->MVrv,
-                stats->MVcv,
-                stats->mv_in_out_count,
-                stats->new_mv_count,
-                stats->count,
-                stats->duration);
-        fclose(fpfile);
-    }
+    fprintf(fpfile,
+            "%12.0f %12.0f %12.0f %12.4f %12.4f %12.4f %12.4f"
+            " %12.4f %12.4f %12.4f %12.4f %12.4f %12.4f %12.4f %12.4f"
+            " %12.0f %12.0f %12.4f\n",
+            stats->frame, stats->intra_error, stats->coded_error,
+            stats->ssim_weighted_pred_err, stats->pcnt_inter,
+            stats->pcnt_motion, stats->pcnt_second_ref, stats->pcnt_neutral,
+            stats->MVr, stats->mvr_abs, stats->MVc, stats->mvc_abs, stats->MVrv,
+            stats->MVcv, stats->mv_in_out_count, stats->new_mv_count,
+            stats->count, stats->duration);
+    fclose(fpfile);
+  }
 #endif
 }
 
-static void zero_stats(FIRSTPASS_STATS *section)
-{
-    section->frame      = 0.0;
-    section->intra_error = 0.0;
-    section->coded_error = 0.0;
-    section->ssim_weighted_pred_err = 0.0;
-    section->pcnt_inter  = 0.0;
-    section->pcnt_motion  = 0.0;
-    section->pcnt_second_ref = 0.0;
-    section->pcnt_neutral = 0.0;
-    section->MVr        = 0.0;
-    section->mvr_abs     = 0.0;
-    section->MVc        = 0.0;
-    section->mvc_abs     = 0.0;
-    section->MVrv       = 0.0;
-    section->MVcv       = 0.0;
-    section->mv_in_out_count  = 0.0;
-    section->new_mv_count = 0.0;
-    section->count      = 0.0;
-    section->duration   = 1.0;
+static void zero_stats(FIRSTPASS_STATS *section) {
+  section->frame = 0.0;
+  section->intra_error = 0.0;
+  section->coded_error = 0.0;
+  section->ssim_weighted_pred_err = 0.0;
+  section->pcnt_inter = 0.0;
+  section->pcnt_motion = 0.0;
+  section->pcnt_second_ref = 0.0;
+  section->pcnt_neutral = 0.0;
+  section->MVr = 0.0;
+  section->mvr_abs = 0.0;
+  section->MVc = 0.0;
+  section->mvc_abs = 0.0;
+  section->MVrv = 0.0;
+  section->MVcv = 0.0;
+  section->mv_in_out_count = 0.0;
+  section->new_mv_count = 0.0;
+  section->count = 0.0;
+  section->duration = 1.0;
 }
 
-static void accumulate_stats(FIRSTPASS_STATS *section, FIRSTPASS_STATS *frame)
-{
-    section->frame += frame->frame;
-    section->intra_error += frame->intra_error;
-    section->coded_error += frame->coded_error;
-    section->ssim_weighted_pred_err += frame->ssim_weighted_pred_err;
-    section->pcnt_inter  += frame->pcnt_inter;
-    section->pcnt_motion += frame->pcnt_motion;
-    section->pcnt_second_ref += frame->pcnt_second_ref;
-    section->pcnt_neutral += frame->pcnt_neutral;
-    section->MVr        += frame->MVr;
-    section->mvr_abs     += frame->mvr_abs;
-    section->MVc        += frame->MVc;
-    section->mvc_abs     += frame->mvc_abs;
-    section->MVrv       += frame->MVrv;
-    section->MVcv       += frame->MVcv;
-    section->mv_in_out_count  += frame->mv_in_out_count;
-    section->new_mv_count += frame->new_mv_count;
-    section->count      += frame->count;
-    section->duration   += frame->duration;
+static void accumulate_stats(FIRSTPASS_STATS *section, FIRSTPASS_STATS *frame) {
+  section->frame += frame->frame;
+  section->intra_error += frame->intra_error;
+  section->coded_error += frame->coded_error;
+  section->ssim_weighted_pred_err += frame->ssim_weighted_pred_err;
+  section->pcnt_inter += frame->pcnt_inter;
+  section->pcnt_motion += frame->pcnt_motion;
+  section->pcnt_second_ref += frame->pcnt_second_ref;
+  section->pcnt_neutral += frame->pcnt_neutral;
+  section->MVr += frame->MVr;
+  section->mvr_abs += frame->mvr_abs;
+  section->MVc += frame->MVc;
+  section->mvc_abs += frame->mvc_abs;
+  section->MVrv += frame->MVrv;
+  section->MVcv += frame->MVcv;
+  section->mv_in_out_count += frame->mv_in_out_count;
+  section->new_mv_count += frame->new_mv_count;
+  section->count += frame->count;
+  section->duration += frame->duration;
 }
 
-static void subtract_stats(FIRSTPASS_STATS *section, FIRSTPASS_STATS *frame)
-{
-    section->frame -= frame->frame;
-    section->intra_error -= frame->intra_error;
-    section->coded_error -= frame->coded_error;
-    section->ssim_weighted_pred_err -= frame->ssim_weighted_pred_err;
-    section->pcnt_inter  -= frame->pcnt_inter;
-    section->pcnt_motion -= frame->pcnt_motion;
-    section->pcnt_second_ref -= frame->pcnt_second_ref;
-    section->pcnt_neutral -= frame->pcnt_neutral;
-    section->MVr        -= frame->MVr;
-    section->mvr_abs     -= frame->mvr_abs;
-    section->MVc        -= frame->MVc;
-    section->mvc_abs     -= frame->mvc_abs;
-    section->MVrv       -= frame->MVrv;
-    section->MVcv       -= frame->MVcv;
-    section->mv_in_out_count  -= frame->mv_in_out_count;
-    section->new_mv_count -= frame->new_mv_count;
-    section->count      -= frame->count;
-    section->duration   -= frame->duration;
+static void subtract_stats(FIRSTPASS_STATS *section, FIRSTPASS_STATS *frame) {
+  section->frame -= frame->frame;
+  section->intra_error -= frame->intra_error;
+  section->coded_error -= frame->coded_error;
+  section->ssim_weighted_pred_err -= frame->ssim_weighted_pred_err;
+  section->pcnt_inter -= frame->pcnt_inter;
+  section->pcnt_motion -= frame->pcnt_motion;
+  section->pcnt_second_ref -= frame->pcnt_second_ref;
+  section->pcnt_neutral -= frame->pcnt_neutral;
+  section->MVr -= frame->MVr;
+  section->mvr_abs -= frame->mvr_abs;
+  section->MVc -= frame->MVc;
+  section->mvc_abs -= frame->mvc_abs;
+  section->MVrv -= frame->MVrv;
+  section->MVcv -= frame->MVcv;
+  section->mv_in_out_count -= frame->mv_in_out_count;
+  section->new_mv_count -= frame->new_mv_count;
+  section->count -= frame->count;
+  section->duration -= frame->duration;
 }
 
-static void avg_stats(FIRSTPASS_STATS *section)
-{
-    if (section->count < 1.0)
-        return;
+static void avg_stats(FIRSTPASS_STATS *section) {
+  if (section->count < 1.0) return;
 
-    section->intra_error /= section->count;
-    section->coded_error /= section->count;
-    section->ssim_weighted_pred_err /= section->count;
-    section->pcnt_inter  /= section->count;
-    section->pcnt_second_ref /= section->count;
-    section->pcnt_neutral /= section->count;
-    section->pcnt_motion /= section->count;
-    section->MVr        /= section->count;
-    section->mvr_abs     /= section->count;
-    section->MVc        /= section->count;
-    section->mvc_abs     /= section->count;
-    section->MVrv       /= section->count;
-    section->MVcv       /= section->count;
-    section->mv_in_out_count   /= section->count;
-    section->duration   /= section->count;
+  section->intra_error /= section->count;
+  section->coded_error /= section->count;
+  section->ssim_weighted_pred_err /= section->count;
+  section->pcnt_inter /= section->count;
+  section->pcnt_second_ref /= section->count;
+  section->pcnt_neutral /= section->count;
+  section->pcnt_motion /= section->count;
+  section->MVr /= section->count;
+  section->mvr_abs /= section->count;
+  section->MVc /= section->count;
+  section->mvc_abs /= section->count;
+  section->MVrv /= section->count;
+  section->MVcv /= section->count;
+  section->mv_in_out_count /= section->count;
+  section->duration /= section->count;
 }
 
 /* Calculate a modified Error used in distributing bits between easier
  * and harder frames
  */
-static double calculate_modified_err(VP8_COMP *cpi, FIRSTPASS_STATS *this_frame)
-{
-    double av_err = ( cpi->twopass.total_stats.ssim_weighted_pred_err /
-                      cpi->twopass.total_stats.count );
-    double this_err = this_frame->ssim_weighted_pred_err;
-    double modified_err;
+static double calculate_modified_err(VP8_COMP *cpi,
+                                     FIRSTPASS_STATS *this_frame) {
+  double av_err = (cpi->twopass.total_stats.ssim_weighted_pred_err /
+                   cpi->twopass.total_stats.count);
+  double this_err = this_frame->ssim_weighted_pred_err;
+  double modified_err;
 
-    if (this_err > av_err)
-        modified_err = av_err * pow((this_err / DOUBLE_DIVIDE_CHECK(av_err)), POW1);
-    else
-        modified_err = av_err * pow((this_err / DOUBLE_DIVIDE_CHECK(av_err)), POW2);
+  if (this_err > av_err)
+    modified_err = av_err * pow((this_err / DOUBLE_DIVIDE_CHECK(av_err)), POW1);
+  else
+    modified_err = av_err * pow((this_err / DOUBLE_DIVIDE_CHECK(av_err)), POW2);
 
-    return modified_err;
+  return modified_err;
 }
 
 static const double weight_table[256] = {
-0.020000, 0.020000, 0.020000, 0.020000, 0.020000, 0.020000, 0.020000, 0.020000,
-0.020000, 0.020000, 0.020000, 0.020000, 0.020000, 0.020000, 0.020000, 0.020000,
-0.020000, 0.020000, 0.020000, 0.020000, 0.020000, 0.020000, 0.020000, 0.020000,
-0.020000, 0.020000, 0.020000, 0.020000, 0.020000, 0.020000, 0.020000, 0.020000,
-0.020000, 0.031250, 0.062500, 0.093750, 0.125000, 0.156250, 0.187500, 0.218750,
-0.250000, 0.281250, 0.312500, 0.343750, 0.375000, 0.406250, 0.437500, 0.468750,
-0.500000, 0.531250, 0.562500, 0.593750, 0.625000, 0.656250, 0.687500, 0.718750,
-0.750000, 0.781250, 0.812500, 0.843750, 0.875000, 0.906250, 0.937500, 0.968750,
-1.000000, 1.000000, 1.000000, 1.000000, 1.000000, 1.000000, 1.000000, 1.000000,
-1.000000, 1.000000, 1.000000, 1.000000, 1.000000, 1.000000, 1.000000, 1.000000,
-1.000000, 1.000000, 1.000000, 1.000000, 1.000000, 1.000000, 1.000000, 1.000000,
-1.000000, 1.000000, 1.000000, 1.000000, 1.000000, 1.000000, 1.000000, 1.000000,
-1.000000, 1.000000, 1.000000, 1.000000, 1.000000, 1.000000, 1.000000, 1.000000,
-1.000000, 1.000000, 1.000000, 1.000000, 1.000000, 1.000000, 1.000000, 1.000000,
-1.000000, 1.000000, 1.000000, 1.000000, 1.000000, 1.000000, 1.000000, 1.000000,
-1.000000, 1.000000, 1.000000, 1.000000, 1.000000, 1.000000, 1.000000, 1.000000,
-1.000000, 1.000000, 1.000000, 1.000000, 1.000000, 1.000000, 1.000000, 1.000000,
-1.000000, 1.000000, 1.000000, 1.000000, 1.000000, 1.000000, 1.000000, 1.000000,
-1.000000, 1.000000, 1.000000, 1.000000, 1.000000, 1.000000, 1.000000, 1.000000,
-1.000000, 1.000000, 1.000000, 1.000000, 1.000000, 1.000000, 1.000000, 1.000000,
-1.000000, 1.000000, 1.000000, 1.000000, 1.000000, 1.000000, 1.000000, 1.000000,
-1.000000, 1.000000, 1.000000, 1.000000, 1.000000, 1.000000, 1.000000, 1.000000,
-1.000000, 1.000000, 1.000000, 1.000000, 1.000000, 1.000000, 1.000000, 1.000000,
-1.000000, 1.000000, 1.000000, 1.000000, 1.000000, 1.000000, 1.000000, 1.000000,
-1.000000, 1.000000, 1.000000, 1.000000, 1.000000, 1.000000, 1.000000, 1.000000,
-1.000000, 1.000000, 1.000000, 1.000000, 1.000000, 1.000000, 1.000000, 1.000000,
-1.000000, 1.000000, 1.000000, 1.000000, 1.000000, 1.000000, 1.000000, 1.000000,
-1.000000, 1.000000, 1.000000, 1.000000, 1.000000, 1.000000, 1.000000, 1.000000,
-1.000000, 1.000000, 1.000000, 1.000000, 1.000000, 1.000000, 1.000000, 1.000000,
-1.000000, 1.000000, 1.000000, 1.000000, 1.000000, 1.000000, 1.000000, 1.000000,
-1.000000, 1.000000, 1.000000, 1.000000, 1.000000, 1.000000, 1.000000, 1.000000,
-1.000000, 1.000000, 1.000000, 1.000000, 1.000000, 1.000000, 1.000000, 1.000000
+  0.020000, 0.020000, 0.020000, 0.020000, 0.020000, 0.020000, 0.020000,
+  0.020000, 0.020000, 0.020000, 0.020000, 0.020000, 0.020000, 0.020000,
+  0.020000, 0.020000, 0.020000, 0.020000, 0.020000, 0.020000, 0.020000,
+  0.020000, 0.020000, 0.020000, 0.020000, 0.020000, 0.020000, 0.020000,
+  0.020000, 0.020000, 0.020000, 0.020000, 0.020000, 0.031250, 0.062500,
+  0.093750, 0.125000, 0.156250, 0.187500, 0.218750, 0.250000, 0.281250,
+  0.312500, 0.343750, 0.375000, 0.406250, 0.437500, 0.468750, 0.500000,
+  0.531250, 0.562500, 0.593750, 0.625000, 0.656250, 0.687500, 0.718750,
+  0.750000, 0.781250, 0.812500, 0.843750, 0.875000, 0.906250, 0.937500,
+  0.968750, 1.000000, 1.000000, 1.000000, 1.000000, 1.000000, 1.000000,
+  1.000000, 1.000000, 1.000000, 1.000000, 1.000000, 1.000000, 1.000000,
+  1.000000, 1.000000, 1.000000, 1.000000, 1.000000, 1.000000, 1.000000,
+  1.000000, 1.000000, 1.000000, 1.000000, 1.000000, 1.000000, 1.000000,
+  1.000000, 1.000000, 1.000000, 1.000000, 1.000000, 1.000000, 1.000000,
+  1.000000, 1.000000, 1.000000, 1.000000, 1.000000, 1.000000, 1.000000,
+  1.000000, 1.000000, 1.000000, 1.000000, 1.000000, 1.000000, 1.000000,
+  1.000000, 1.000000, 1.000000, 1.000000, 1.000000, 1.000000, 1.000000,
+  1.000000, 1.000000, 1.000000, 1.000000, 1.000000, 1.000000, 1.000000,
+  1.000000, 1.000000, 1.000000, 1.000000, 1.000000, 1.000000, 1.000000,
+  1.000000, 1.000000, 1.000000, 1.000000, 1.000000, 1.000000, 1.000000,
+  1.000000, 1.000000, 1.000000, 1.000000, 1.000000, 1.000000, 1.000000,
+  1.000000, 1.000000, 1.000000, 1.000000, 1.000000, 1.000000, 1.000000,
+  1.000000, 1.000000, 1.000000, 1.000000, 1.000000, 1.000000, 1.000000,
+  1.000000, 1.000000, 1.000000, 1.000000, 1.000000, 1.000000, 1.000000,
+  1.000000, 1.000000, 1.000000, 1.000000, 1.000000, 1.000000, 1.000000,
+  1.000000, 1.000000, 1.000000, 1.000000, 1.000000, 1.000000, 1.000000,
+  1.000000, 1.000000, 1.000000, 1.000000, 1.000000, 1.000000, 1.000000,
+  1.000000, 1.000000, 1.000000, 1.000000, 1.000000, 1.000000, 1.000000,
+  1.000000, 1.000000, 1.000000, 1.000000, 1.000000, 1.000000, 1.000000,
+  1.000000, 1.000000, 1.000000, 1.000000, 1.000000, 1.000000, 1.000000,
+  1.000000, 1.000000, 1.000000, 1.000000, 1.000000, 1.000000, 1.000000,
+  1.000000, 1.000000, 1.000000, 1.000000, 1.000000, 1.000000, 1.000000,
+  1.000000, 1.000000, 1.000000, 1.000000, 1.000000, 1.000000, 1.000000,
+  1.000000, 1.000000, 1.000000, 1.000000, 1.000000, 1.000000, 1.000000,
+  1.000000, 1.000000, 1.000000, 1.000000, 1.000000, 1.000000, 1.000000,
+  1.000000, 1.000000, 1.000000, 1.000000, 1.000000, 1.000000, 1.000000,
+  1.000000, 1.000000, 1.000000, 1.000000
 };
 
-static double simple_weight(YV12_BUFFER_CONFIG *source)
-{
-    int i, j;
+static double simple_weight(YV12_BUFFER_CONFIG *source) {
+  int i, j;
 
-    unsigned char *src = source->y_buffer;
-    double sum_weights = 0.0;
+  unsigned char *src = source->y_buffer;
+  double sum_weights = 0.0;
 
-    /* Loop throught the Y plane raw examining levels and creating a weight
-     * for the image
-     */
-    i = source->y_height;
-    do
-    {
-        j = source->y_width;
-        do
-        {
-            sum_weights += weight_table[ *src];
-            src++;
-        }while(--j);
-        src -= source->y_width;
-        src += source->y_stride;
-    }while(--i);
+  /* Loop throught the Y plane raw examining levels and creating a weight
+   * for the image
+   */
+  i = source->y_height;
+  do {
+    j = source->y_width;
+    do {
+      sum_weights += weight_table[*src];
+      src++;
+    } while (--j);
+    src -= source->y_width;
+    src += source->y_stride;
+  } while (--i);
 
-    sum_weights /= (source->y_height * source->y_width);
+  sum_weights /= (source->y_height * source->y_width);
 
-    return sum_weights;
+  return sum_weights;
 }
 
-
 /* This function returns the current per frame maximum bitrate target */
-static int frame_max_bits(VP8_COMP *cpi)
-{
-    /* Max allocation for a single frame based on the max section guidelines
-     * passed in and how many bits are left
-     */
-    int max_bits;
+static int frame_max_bits(VP8_COMP *cpi) {
+  /* Max allocation for a single frame based on the max section guidelines
+   * passed in and how many bits are left
+   */
+  int max_bits;
 
-    /* For CBR we need to also consider buffer fullness.
-     * If we are running below the optimal level then we need to gradually
-     * tighten up on max_bits.
+  /* For CBR we need to also consider buffer fullness.
+   * If we are running below the optimal level then we need to gradually
+   * tighten up on max_bits.
+   */
+  if (cpi->oxcf.end_usage == USAGE_STREAM_FROM_SERVER) {
+    double buffer_fullness_ratio =
+        (double)cpi->buffer_level /
+        DOUBLE_DIVIDE_CHECK((double)cpi->oxcf.optimal_buffer_level);
+
+    /* For CBR base this on the target average bits per frame plus the
+     * maximum sedction rate passed in by the user
      */
-    if (cpi->oxcf.end_usage == USAGE_STREAM_FROM_SERVER)
-    {
-        double buffer_fullness_ratio = (double)cpi->buffer_level / DOUBLE_DIVIDE_CHECK((double)cpi->oxcf.optimal_buffer_level);
+    max_bits = (int)(cpi->av_per_frame_bandwidth *
+                     ((double)cpi->oxcf.two_pass_vbrmax_section / 100.0));
 
-        /* For CBR base this on the target average bits per frame plus the
-         * maximum sedction rate passed in by the user
-         */
-        max_bits = (int)(cpi->av_per_frame_bandwidth * ((double)cpi->oxcf.two_pass_vbrmax_section / 100.0));
+    /* If our buffer is below the optimum level */
+    if (buffer_fullness_ratio < 1.0) {
+      /* The lower of max_bits / 4 or cpi->av_per_frame_bandwidth / 4. */
+      int min_max_bits = ((cpi->av_per_frame_bandwidth >> 2) < (max_bits >> 2))
+                             ? cpi->av_per_frame_bandwidth >> 2
+                             : max_bits >> 2;
 
-        /* If our buffer is below the optimum level */
-        if (buffer_fullness_ratio < 1.0)
-        {
-            /* The lower of max_bits / 4 or cpi->av_per_frame_bandwidth / 4. */
-            int min_max_bits = ((cpi->av_per_frame_bandwidth >> 2) < (max_bits >> 2)) ? cpi->av_per_frame_bandwidth >> 2 : max_bits >> 2;
+      max_bits = (int)(max_bits * buffer_fullness_ratio);
 
-            max_bits = (int)(max_bits * buffer_fullness_ratio);
-
-            /* Lowest value we will set ... which should allow the buffer to
-             * refill.
-             */
-            if (max_bits < min_max_bits)
-                max_bits = min_max_bits;
-        }
+      /* Lowest value we will set ... which should allow the buffer to
+       * refill.
+       */
+      if (max_bits < min_max_bits) max_bits = min_max_bits;
     }
-    /* VBR */
-    else
-    {
-        /* For VBR base this on the bits and frames left plus the
-         * two_pass_vbrmax_section rate passed in by the user
-         */
-        max_bits = (int)(((double)cpi->twopass.bits_left / (cpi->twopass.total_stats.count - (double)cpi->common.current_video_frame)) * ((double)cpi->oxcf.two_pass_vbrmax_section / 100.0));
-    }
+  }
+  /* VBR */
+  else {
+    /* For VBR base this on the bits and frames left plus the
+     * two_pass_vbrmax_section rate passed in by the user
+     */
+    max_bits = (int)(((double)cpi->twopass.bits_left /
+                      (cpi->twopass.total_stats.count -
+                       (double)cpi->common.current_video_frame)) *
+                     ((double)cpi->oxcf.two_pass_vbrmax_section / 100.0));
+  }
 
-    /* Trap case where we are out of bits */
-    if (max_bits < 0)
-        max_bits = 0;
+  /* Trap case where we are out of bits */
+  if (max_bits < 0) max_bits = 0;
 
-    return max_bits;
+  return max_bits;
 }
 
-void vp8_init_first_pass(VP8_COMP *cpi)
-{
-    zero_stats(&cpi->twopass.total_stats);
+void vp8_init_first_pass(VP8_COMP *cpi) {
+  zero_stats(&cpi->twopass.total_stats);
 }
 
-void vp8_end_first_pass(VP8_COMP *cpi)
-{
-    output_stats(cpi, cpi->output_pkt_list, &cpi->twopass.total_stats);
+void vp8_end_first_pass(VP8_COMP *cpi) {
+  output_stats(cpi, cpi->output_pkt_list, &cpi->twopass.total_stats);
 }
 
-static void zz_motion_search( VP8_COMP *cpi, MACROBLOCK * x,
-                              YV12_BUFFER_CONFIG * raw_buffer,
-                              int * raw_motion_err,
-                              YV12_BUFFER_CONFIG * recon_buffer,
-                              int * best_motion_err, int recon_yoffset)
-{
-    MACROBLOCKD * const xd = & x->e_mbd;
-    BLOCK *b = &x->block[0];
-    BLOCKD *d = &x->e_mbd.block[0];
+static void zz_motion_search(VP8_COMP *cpi, MACROBLOCK *x,
+                             YV12_BUFFER_CONFIG *raw_buffer,
+                             int *raw_motion_err,
+                             YV12_BUFFER_CONFIG *recon_buffer,
+                             int *best_motion_err, int recon_yoffset) {
+  MACROBLOCKD *const xd = &x->e_mbd;
+  BLOCK *b = &x->block[0];
+  BLOCKD *d = &x->e_mbd.block[0];
 
-    unsigned char *src_ptr = (*(b->base_src) + b->src);
-    int src_stride = b->src_stride;
-    unsigned char *raw_ptr;
-    int raw_stride = raw_buffer->y_stride;
-    unsigned char *ref_ptr;
-    int ref_stride = x->e_mbd.pre.y_stride;
-    (void)cpi;
+  unsigned char *src_ptr = (*(b->base_src) + b->src);
+  int src_stride = b->src_stride;
+  unsigned char *raw_ptr;
+  int raw_stride = raw_buffer->y_stride;
+  unsigned char *ref_ptr;
+  int ref_stride = x->e_mbd.pre.y_stride;
+  (void)cpi;
 
-    /* Set up pointers for this macro block raw buffer */
-    raw_ptr = (unsigned char *)(raw_buffer->y_buffer + recon_yoffset
-                                + d->offset);
-    vpx_mse16x16(src_ptr, src_stride, raw_ptr, raw_stride,
-                 (unsigned int *)(raw_motion_err));
+  /* Set up pointers for this macro block raw buffer */
+  raw_ptr = (unsigned char *)(raw_buffer->y_buffer + recon_yoffset + d->offset);
+  vpx_mse16x16(src_ptr, src_stride, raw_ptr, raw_stride,
+               (unsigned int *)(raw_motion_err));
 
-    /* Set up pointers for this macro block recon buffer */
-    xd->pre.y_buffer = recon_buffer->y_buffer + recon_yoffset;
-    ref_ptr = (unsigned char *)(xd->pre.y_buffer + d->offset );
-    vpx_mse16x16(src_ptr, src_stride, ref_ptr, ref_stride,
-                 (unsigned int *)(best_motion_err));
+  /* Set up pointers for this macro block recon buffer */
+  xd->pre.y_buffer = recon_buffer->y_buffer + recon_yoffset;
+  ref_ptr = (unsigned char *)(xd->pre.y_buffer + d->offset);
+  vpx_mse16x16(src_ptr, src_stride, ref_ptr, ref_stride,
+               (unsigned int *)(best_motion_err));
 }
 
 static void first_pass_motion_search(VP8_COMP *cpi, MACROBLOCK *x,
                                      int_mv *ref_mv, MV *best_mv,
                                      YV12_BUFFER_CONFIG *recon_buffer,
-                                     int *best_motion_err, int recon_yoffset )
-{
-    MACROBLOCKD *const xd = & x->e_mbd;
-    BLOCK *b = &x->block[0];
-    BLOCKD *d = &x->e_mbd.block[0];
-    int num00;
+                                     int *best_motion_err, int recon_yoffset) {
+  MACROBLOCKD *const xd = &x->e_mbd;
+  BLOCK *b = &x->block[0];
+  BLOCKD *d = &x->e_mbd.block[0];
+  int num00;
 
-    int_mv tmp_mv;
-    int_mv ref_mv_full;
+  int_mv tmp_mv;
+  int_mv ref_mv_full;
 
-    int tmp_err;
-    int step_param = 3; /* Dont search over full range for first pass */
-    int further_steps = (MAX_MVSEARCH_STEPS - 1) - step_param;
-    int n;
-    vp8_variance_fn_ptr_t v_fn_ptr = cpi->fn_ptr[BLOCK_16X16];
-    int new_mv_mode_penalty = 256;
+  int tmp_err;
+  int step_param = 3; /* Dont search over full range for first pass */
+  int further_steps = (MAX_MVSEARCH_STEPS - 1) - step_param;
+  int n;
+  vp8_variance_fn_ptr_t v_fn_ptr = cpi->fn_ptr[BLOCK_16X16];
+  int new_mv_mode_penalty = 256;
 
-    /* override the default variance function to use MSE */
-    v_fn_ptr.vf    = vpx_mse16x16;
+  /* override the default variance function to use MSE */
+  v_fn_ptr.vf = vpx_mse16x16;
 
-    /* Set up pointers for this macro block recon buffer */
-    xd->pre.y_buffer = recon_buffer->y_buffer + recon_yoffset;
+  /* Set up pointers for this macro block recon buffer */
+  xd->pre.y_buffer = recon_buffer->y_buffer + recon_yoffset;
 
-    /* Initial step/diamond search centred on best mv */
-    tmp_mv.as_int = 0;
-    ref_mv_full.as_mv.col = ref_mv->as_mv.col>>3;
-    ref_mv_full.as_mv.row = ref_mv->as_mv.row>>3;
-    tmp_err = cpi->diamond_search_sad(x, b, d, &ref_mv_full, &tmp_mv, step_param,
-                                      x->sadperbit16, &num00, &v_fn_ptr,
-                                      x->mvcost, ref_mv);
-    if ( tmp_err < INT_MAX-new_mv_mode_penalty )
+  /* Initial step/diamond search centred on best mv */
+  tmp_mv.as_int = 0;
+  ref_mv_full.as_mv.col = ref_mv->as_mv.col >> 3;
+  ref_mv_full.as_mv.row = ref_mv->as_mv.row >> 3;
+  tmp_err = cpi->diamond_search_sad(x, b, d, &ref_mv_full, &tmp_mv, step_param,
+                                    x->sadperbit16, &num00, &v_fn_ptr,
+                                    x->mvcost, ref_mv);
+  if (tmp_err < INT_MAX - new_mv_mode_penalty) tmp_err += new_mv_mode_penalty;
+
+  if (tmp_err < *best_motion_err) {
+    *best_motion_err = tmp_err;
+    best_mv->row = tmp_mv.as_mv.row;
+    best_mv->col = tmp_mv.as_mv.col;
+  }
+
+  /* Further step/diamond searches as necessary */
+  n = num00;
+  num00 = 0;
+
+  while (n < further_steps) {
+    n++;
+
+    if (num00)
+      num00--;
+    else {
+      tmp_err = cpi->diamond_search_sad(x, b, d, &ref_mv_full, &tmp_mv,
+                                        step_param + n, x->sadperbit16, &num00,
+                                        &v_fn_ptr, x->mvcost, ref_mv);
+      if (tmp_err < INT_MAX - new_mv_mode_penalty)
         tmp_err += new_mv_mode_penalty;
 
-    if (tmp_err < *best_motion_err)
-    {
+      if (tmp_err < *best_motion_err) {
         *best_motion_err = tmp_err;
         best_mv->row = tmp_mv.as_mv.row;
         best_mv->col = tmp_mv.as_mv.col;
+      }
     }
-
-    /* Further step/diamond searches as necessary */
-    n = num00;
-    num00 = 0;
-
-    while (n < further_steps)
-    {
-        n++;
-
-        if (num00)
-            num00--;
-        else
-        {
-            tmp_err = cpi->diamond_search_sad(x, b, d, &ref_mv_full, &tmp_mv,
-                                              step_param + n, x->sadperbit16,
-                                              &num00, &v_fn_ptr, x->mvcost,
-                                              ref_mv);
-            if ( tmp_err < INT_MAX-new_mv_mode_penalty )
-                tmp_err += new_mv_mode_penalty;
-
-            if (tmp_err < *best_motion_err)
-            {
-                *best_motion_err = tmp_err;
-                best_mv->row = tmp_mv.as_mv.row;
-                best_mv->col = tmp_mv.as_mv.col;
-            }
-        }
-    }
+  }
 }
 
-void vp8_first_pass(VP8_COMP *cpi)
-{
-    int mb_row, mb_col;
-    MACROBLOCK *const x = & cpi->mb;
-    VP8_COMMON *const cm = & cpi->common;
-    MACROBLOCKD *const xd = & x->e_mbd;
+void vp8_first_pass(VP8_COMP *cpi) {
+  int mb_row, mb_col;
+  MACROBLOCK *const x = &cpi->mb;
+  VP8_COMMON *const cm = &cpi->common;
+  MACROBLOCKD *const xd = &x->e_mbd;
 
-    int recon_yoffset, recon_uvoffset;
-    YV12_BUFFER_CONFIG *lst_yv12 = &cm->yv12_fb[cm->lst_fb_idx];
-    YV12_BUFFER_CONFIG *new_yv12 = &cm->yv12_fb[cm->new_fb_idx];
-    YV12_BUFFER_CONFIG *gld_yv12 = &cm->yv12_fb[cm->gld_fb_idx];
-    int recon_y_stride = lst_yv12->y_stride;
-    int recon_uv_stride = lst_yv12->uv_stride;
-    int64_t intra_error = 0;
-    int64_t coded_error = 0;
+  int recon_yoffset, recon_uvoffset;
+  YV12_BUFFER_CONFIG *lst_yv12 = &cm->yv12_fb[cm->lst_fb_idx];
+  YV12_BUFFER_CONFIG *new_yv12 = &cm->yv12_fb[cm->new_fb_idx];
+  YV12_BUFFER_CONFIG *gld_yv12 = &cm->yv12_fb[cm->gld_fb_idx];
+  int recon_y_stride = lst_yv12->y_stride;
+  int recon_uv_stride = lst_yv12->uv_stride;
+  int64_t intra_error = 0;
+  int64_t coded_error = 0;
 
-    int sum_mvr = 0, sum_mvc = 0;
-    int sum_mvr_abs = 0, sum_mvc_abs = 0;
-    int sum_mvrs = 0, sum_mvcs = 0;
-    int mvcount = 0;
-    int intercount = 0;
-    int second_ref_count = 0;
-    int intrapenalty = 256;
-    int neutral_count = 0;
-    int new_mv_count = 0;
-    int sum_in_vectors = 0;
-    uint32_t lastmv_as_int = 0;
+  int sum_mvr = 0, sum_mvc = 0;
+  int sum_mvr_abs = 0, sum_mvc_abs = 0;
+  int sum_mvrs = 0, sum_mvcs = 0;
+  int mvcount = 0;
+  int intercount = 0;
+  int second_ref_count = 0;
+  int intrapenalty = 256;
+  int neutral_count = 0;
+  int new_mv_count = 0;
+  int sum_in_vectors = 0;
+  uint32_t lastmv_as_int = 0;
 
-    int_mv zero_ref_mv;
+  int_mv zero_ref_mv;
 
-    zero_ref_mv.as_int = 0;
+  zero_ref_mv.as_int = 0;
 
-    vp8_clear_system_state();
+  vp8_clear_system_state();
 
-    x->src = * cpi->Source;
-    xd->pre = *lst_yv12;
-    xd->dst = *new_yv12;
+  x->src = *cpi->Source;
+  xd->pre = *lst_yv12;
+  xd->dst = *new_yv12;
 
-    x->partition_info = x->pi;
+  x->partition_info = x->pi;
 
-    xd->mode_info_context = cm->mi;
+  xd->mode_info_context = cm->mi;
 
-    if(!cm->use_bilinear_mc_filter)
-    {
-         xd->subpixel_predict        = vp8_sixtap_predict4x4;
-         xd->subpixel_predict8x4     = vp8_sixtap_predict8x4;
-         xd->subpixel_predict8x8     = vp8_sixtap_predict8x8;
-         xd->subpixel_predict16x16   = vp8_sixtap_predict16x16;
-     }
-     else
-     {
-         xd->subpixel_predict        = vp8_bilinear_predict4x4;
-         xd->subpixel_predict8x4     = vp8_bilinear_predict8x4;
-         xd->subpixel_predict8x8     = vp8_bilinear_predict8x8;
-         xd->subpixel_predict16x16   = vp8_bilinear_predict16x16;
-     }
+  if (!cm->use_bilinear_mc_filter) {
+    xd->subpixel_predict = vp8_sixtap_predict4x4;
+    xd->subpixel_predict8x4 = vp8_sixtap_predict8x4;
+    xd->subpixel_predict8x8 = vp8_sixtap_predict8x8;
+    xd->subpixel_predict16x16 = vp8_sixtap_predict16x16;
+  } else {
+    xd->subpixel_predict = vp8_bilinear_predict4x4;
+    xd->subpixel_predict8x4 = vp8_bilinear_predict8x4;
+    xd->subpixel_predict8x8 = vp8_bilinear_predict8x8;
+    xd->subpixel_predict16x16 = vp8_bilinear_predict16x16;
+  }
 
-    vp8_build_block_offsets(x);
+  vp8_build_block_offsets(x);
 
-    /* set up frame new frame for intra coded blocks */
-    vp8_setup_intra_recon(new_yv12);
-    vp8cx_frame_init_quantizer(cpi);
+  /* set up frame new frame for intra coded blocks */
+  vp8_setup_intra_recon(new_yv12);
+  vp8cx_frame_init_quantizer(cpi);
 
-    /* Initialise the MV cost table to the defaults */
-    {
-        int flag[2] = {1, 1};
-        vp8_initialize_rd_consts(cpi, x, vp8_dc_quant(cm->base_qindex, cm->y1dc_delta_q));
-        memcpy(cm->fc.mvc, vp8_default_mv_context, sizeof(vp8_default_mv_context));
-        vp8_build_component_cost_table(cpi->mb.mvcost, (const MV_CONTEXT *) cm->fc.mvc, flag);
-    }
+  /* Initialise the MV cost table to the defaults */
+  {
+    int flag[2] = { 1, 1 };
+    vp8_initialize_rd_consts(cpi, x,
+                             vp8_dc_quant(cm->base_qindex, cm->y1dc_delta_q));
+    memcpy(cm->fc.mvc, vp8_default_mv_context, sizeof(vp8_default_mv_context));
+    vp8_build_component_cost_table(cpi->mb.mvcost,
+                                   (const MV_CONTEXT *)cm->fc.mvc, flag);
+  }
 
-    /* for each macroblock row in image */
-    for (mb_row = 0; mb_row < cm->mb_rows; mb_row++)
-    {
-        int_mv best_ref_mv;
+  /* for each macroblock row in image */
+  for (mb_row = 0; mb_row < cm->mb_rows; mb_row++) {
+    int_mv best_ref_mv;
 
-        best_ref_mv.as_int = 0;
+    best_ref_mv.as_int = 0;
 
-        /* reset above block coeffs */
-        xd->up_available = (mb_row != 0);
-        recon_yoffset = (mb_row * recon_y_stride * 16);
-        recon_uvoffset = (mb_row * recon_uv_stride * 8);
+    /* reset above block coeffs */
+    xd->up_available = (mb_row != 0);
+    recon_yoffset = (mb_row * recon_y_stride * 16);
+    recon_uvoffset = (mb_row * recon_uv_stride * 8);
 
-        /* Set up limit values for motion vectors to prevent them extending
-         * outside the UMV borders
-         */
-        x->mv_row_min = -((mb_row * 16) + (VP8BORDERINPIXELS - 16));
-        x->mv_row_max = ((cm->mb_rows - 1 - mb_row) * 16) + (VP8BORDERINPIXELS - 16);
+    /* Set up limit values for motion vectors to prevent them extending
+     * outside the UMV borders
+     */
+    x->mv_row_min = -((mb_row * 16) + (VP8BORDERINPIXELS - 16));
+    x->mv_row_max =
+        ((cm->mb_rows - 1 - mb_row) * 16) + (VP8BORDERINPIXELS - 16);
 
+    /* for each macroblock col in image */
+    for (mb_col = 0; mb_col < cm->mb_cols; mb_col++) {
+      int this_error;
+      int gf_motion_error = INT_MAX;
+      int use_dc_pred = (mb_col || mb_row) && (!mb_col || !mb_row);
 
-        /* for each macroblock col in image */
-        for (mb_col = 0; mb_col < cm->mb_cols; mb_col++)
-        {
-            int this_error;
-            int gf_motion_error = INT_MAX;
-            int use_dc_pred = (mb_col || mb_row) && (!mb_col || !mb_row);
+      xd->dst.y_buffer = new_yv12->y_buffer + recon_yoffset;
+      xd->dst.u_buffer = new_yv12->u_buffer + recon_uvoffset;
+      xd->dst.v_buffer = new_yv12->v_buffer + recon_uvoffset;
+      xd->left_available = (mb_col != 0);
 
-            xd->dst.y_buffer = new_yv12->y_buffer + recon_yoffset;
-            xd->dst.u_buffer = new_yv12->u_buffer + recon_uvoffset;
-            xd->dst.v_buffer = new_yv12->v_buffer + recon_uvoffset;
-            xd->left_available = (mb_col != 0);
+      /* Copy current mb to a buffer */
+      vp8_copy_mem16x16(x->src.y_buffer, x->src.y_stride, x->thismb, 16);
 
-            /* Copy current mb to a buffer */
-            vp8_copy_mem16x16(x->src.y_buffer, x->src.y_stride, x->thismb, 16);
+      /* do intra 16x16 prediction */
+      this_error = vp8_encode_intra(cpi, x, use_dc_pred);
 
-            /* do intra 16x16 prediction */
-            this_error = vp8_encode_intra(cpi, x, use_dc_pred);
+      /* "intrapenalty" below deals with situations where the intra
+       * and inter error scores are very low (eg a plain black frame)
+       * We do not have special cases in first pass for 0,0 and
+       * nearest etc so all inter modes carry an overhead cost
+       * estimate fot the mv. When the error score is very low this
+       * causes us to pick all or lots of INTRA modes and throw lots
+       * of key frames. This penalty adds a cost matching that of a
+       * 0,0 mv to the intra case.
+       */
+      this_error += intrapenalty;
 
-            /* "intrapenalty" below deals with situations where the intra
-             * and inter error scores are very low (eg a plain black frame)
-             * We do not have special cases in first pass for 0,0 and
-             * nearest etc so all inter modes carry an overhead cost
-             * estimate fot the mv. When the error score is very low this
-             * causes us to pick all or lots of INTRA modes and throw lots
-             * of key frames. This penalty adds a cost matching that of a
-             * 0,0 mv to the intra case.
-             */
-            this_error += intrapenalty;
+      /* Cumulative intra error total */
+      intra_error += (int64_t)this_error;
 
-            /* Cumulative intra error total */
-            intra_error += (int64_t)this_error;
+      /* Set up limit values for motion vectors to prevent them
+       * extending outside the UMV borders
+       */
+      x->mv_col_min = -((mb_col * 16) + (VP8BORDERINPIXELS - 16));
+      x->mv_col_max =
+          ((cm->mb_cols - 1 - mb_col) * 16) + (VP8BORDERINPIXELS - 16);
 
-            /* Set up limit values for motion vectors to prevent them
-             * extending outside the UMV borders
-             */
-            x->mv_col_min = -((mb_col * 16) + (VP8BORDERINPIXELS - 16));
-            x->mv_col_max = ((cm->mb_cols - 1 - mb_col) * 16) + (VP8BORDERINPIXELS - 16);
+      /* Other than for the first frame do a motion search */
+      if (cm->current_video_frame > 0) {
+        BLOCKD *d = &x->e_mbd.block[0];
+        MV tmp_mv = { 0, 0 };
+        int tmp_err;
+        int motion_error = INT_MAX;
+        int raw_motion_error = INT_MAX;
 
-            /* Other than for the first frame do a motion search */
-            if (cm->current_video_frame > 0)
-            {
-                BLOCKD *d = &x->e_mbd.block[0];
-                MV tmp_mv = {0, 0};
-                int tmp_err;
-                int motion_error = INT_MAX;
-                int raw_motion_error = INT_MAX;
+        /* Simple 0,0 motion with no mv overhead */
+        zz_motion_search(cpi, x, cpi->last_frame_unscaled_source,
+                         &raw_motion_error, lst_yv12, &motion_error,
+                         recon_yoffset);
+        d->bmi.mv.as_mv.row = 0;
+        d->bmi.mv.as_mv.col = 0;
 
-                /* Simple 0,0 motion with no mv overhead */
-                zz_motion_search( cpi, x, cpi->last_frame_unscaled_source,
-                                  &raw_motion_error, lst_yv12, &motion_error,
-                                  recon_yoffset );
-                d->bmi.mv.as_mv.row = 0;
-                d->bmi.mv.as_mv.col = 0;
+        if (raw_motion_error < cpi->oxcf.encode_breakout)
+          goto skip_motion_search;
 
-                if (raw_motion_error < cpi->oxcf.encode_breakout)
-                    goto skip_motion_search;
+        /* Test last reference frame using the previous best mv as the
+         * starting point (best reference) for the search
+         */
+        first_pass_motion_search(cpi, x, &best_ref_mv, &d->bmi.mv.as_mv,
+                                 lst_yv12, &motion_error, recon_yoffset);
 
-                /* Test last reference frame using the previous best mv as the
-                 * starting point (best reference) for the search
-                 */
-                first_pass_motion_search(cpi, x, &best_ref_mv,
-                                        &d->bmi.mv.as_mv, lst_yv12,
-                                        &motion_error, recon_yoffset);
+        /* If the current best reference mv is not centred on 0,0
+         * then do a 0,0 based search as well
+         */
+        if (best_ref_mv.as_int) {
+          tmp_err = INT_MAX;
+          first_pass_motion_search(cpi, x, &zero_ref_mv, &tmp_mv, lst_yv12,
+                                   &tmp_err, recon_yoffset);
 
-                /* If the current best reference mv is not centred on 0,0
-                 * then do a 0,0 based search as well
-                 */
-                if (best_ref_mv.as_int)
-                {
-                   tmp_err = INT_MAX;
-                   first_pass_motion_search(cpi, x, &zero_ref_mv, &tmp_mv,
-                                     lst_yv12, &tmp_err, recon_yoffset);
+          if (tmp_err < motion_error) {
+            motion_error = tmp_err;
+            d->bmi.mv.as_mv.row = tmp_mv.row;
+            d->bmi.mv.as_mv.col = tmp_mv.col;
+          }
+        }
 
-                   if ( tmp_err < motion_error )
-                   {
-                        motion_error = tmp_err;
-                        d->bmi.mv.as_mv.row = tmp_mv.row;
-                        d->bmi.mv.as_mv.col = tmp_mv.col;
-                   }
-                }
+        /* Experimental search in a second reference frame ((0,0)
+         * based only)
+         */
+        if (cm->current_video_frame > 1) {
+          first_pass_motion_search(cpi, x, &zero_ref_mv, &tmp_mv, gld_yv12,
+                                   &gf_motion_error, recon_yoffset);
 
-                /* Experimental search in a second reference frame ((0,0)
-                 * based only)
-                 */
-                if (cm->current_video_frame > 1)
-                {
-                    first_pass_motion_search(cpi, x, &zero_ref_mv, &tmp_mv, gld_yv12, &gf_motion_error, recon_yoffset);
+          if ((gf_motion_error < motion_error) &&
+              (gf_motion_error < this_error)) {
+            second_ref_count++;
+          }
 
-                    if ((gf_motion_error < motion_error) && (gf_motion_error < this_error))
-                    {
-                        second_ref_count++;
-                    }
+          /* Reset to last frame as reference buffer */
+          xd->pre.y_buffer = lst_yv12->y_buffer + recon_yoffset;
+          xd->pre.u_buffer = lst_yv12->u_buffer + recon_uvoffset;
+          xd->pre.v_buffer = lst_yv12->v_buffer + recon_uvoffset;
+        }
 
-                    /* Reset to last frame as reference buffer */
-                    xd->pre.y_buffer = lst_yv12->y_buffer + recon_yoffset;
-                    xd->pre.u_buffer = lst_yv12->u_buffer + recon_uvoffset;
-                    xd->pre.v_buffer = lst_yv12->v_buffer + recon_uvoffset;
-                }
+      skip_motion_search:
+        /* Intra assumed best */
+        best_ref_mv.as_int = 0;
 
-skip_motion_search:
-                /* Intra assumed best */
-                best_ref_mv.as_int = 0;
+        if (motion_error <= this_error) {
+          /* Keep a count of cases where the inter and intra were
+           * very close and very low. This helps with scene cut
+           * detection for example in cropped clips with black bars
+           * at the sides or top and bottom.
+           */
+          if ((((this_error - intrapenalty) * 9) <= (motion_error * 10)) &&
+              (this_error < (2 * intrapenalty))) {
+            neutral_count++;
+          }
 
-                if (motion_error <= this_error)
-                {
-                    /* Keep a count of cases where the inter and intra were
-                     * very close and very low. This helps with scene cut
-                     * detection for example in cropped clips with black bars
-                     * at the sides or top and bottom.
-                     */
-                    if( (((this_error-intrapenalty) * 9) <=
-                         (motion_error*10)) &&
-                        (this_error < (2*intrapenalty)) )
-                    {
-                        neutral_count++;
-                    }
+          d->bmi.mv.as_mv.row *= 8;
+          d->bmi.mv.as_mv.col *= 8;
+          this_error = motion_error;
+          vp8_set_mbmode_and_mvs(x, NEWMV, &d->bmi.mv);
+          vp8_encode_inter16x16y(x);
+          sum_mvr += d->bmi.mv.as_mv.row;
+          sum_mvr_abs += abs(d->bmi.mv.as_mv.row);
+          sum_mvc += d->bmi.mv.as_mv.col;
+          sum_mvc_abs += abs(d->bmi.mv.as_mv.col);
+          sum_mvrs += d->bmi.mv.as_mv.row * d->bmi.mv.as_mv.row;
+          sum_mvcs += d->bmi.mv.as_mv.col * d->bmi.mv.as_mv.col;
+          intercount++;
 
-                    d->bmi.mv.as_mv.row *= 8;
-                    d->bmi.mv.as_mv.col *= 8;
-                    this_error = motion_error;
-                    vp8_set_mbmode_and_mvs(x, NEWMV, &d->bmi.mv);
-                    vp8_encode_inter16x16y(x);
-                    sum_mvr += d->bmi.mv.as_mv.row;
-                    sum_mvr_abs += abs(d->bmi.mv.as_mv.row);
-                    sum_mvc += d->bmi.mv.as_mv.col;
-                    sum_mvc_abs += abs(d->bmi.mv.as_mv.col);
-                    sum_mvrs += d->bmi.mv.as_mv.row * d->bmi.mv.as_mv.row;
-                    sum_mvcs += d->bmi.mv.as_mv.col * d->bmi.mv.as_mv.col;
-                    intercount++;
+          best_ref_mv.as_int = d->bmi.mv.as_int;
 
-                    best_ref_mv.as_int = d->bmi.mv.as_int;
+          /* Was the vector non-zero */
+          if (d->bmi.mv.as_int) {
+            mvcount++;
 
-                    /* Was the vector non-zero */
-                    if (d->bmi.mv.as_int)
-                    {
-                        mvcount++;
+            /* Was it different from the last non zero vector */
+            if (d->bmi.mv.as_int != lastmv_as_int) new_mv_count++;
+            lastmv_as_int = d->bmi.mv.as_int;
 
-                        /* Was it different from the last non zero vector */
-                        if ( d->bmi.mv.as_int != lastmv_as_int )
-                            new_mv_count++;
-                        lastmv_as_int = d->bmi.mv.as_int;
-
-                        /* Does the Row vector point inwards or outwards */
-                        if (mb_row < cm->mb_rows / 2)
-                        {
-                            if (d->bmi.mv.as_mv.row > 0)
-                                sum_in_vectors--;
-                            else if (d->bmi.mv.as_mv.row < 0)
-                                sum_in_vectors++;
-                        }
-                        else if (mb_row > cm->mb_rows / 2)
-                        {
-                            if (d->bmi.mv.as_mv.row > 0)
-                                sum_in_vectors++;
-                            else if (d->bmi.mv.as_mv.row < 0)
-                                sum_in_vectors--;
-                        }
+            /* Does the Row vector point inwards or outwards */
+            if (mb_row < cm->mb_rows / 2) {
+              if (d->bmi.mv.as_mv.row > 0)
+                sum_in_vectors--;
+              else if (d->bmi.mv.as_mv.row < 0)
+                sum_in_vectors++;
+            } else if (mb_row > cm->mb_rows / 2) {
+              if (d->bmi.mv.as_mv.row > 0)
+                sum_in_vectors++;
+              else if (d->bmi.mv.as_mv.row < 0)
+                sum_in_vectors--;
+            }
 
-                        /* Does the Row vector point inwards or outwards */
-                        if (mb_col < cm->mb_cols / 2)
-                        {
-                            if (d->bmi.mv.as_mv.col > 0)
-                                sum_in_vectors--;
-                            else if (d->bmi.mv.as_mv.col < 0)
-                                sum_in_vectors++;
-                        }
-                        else if (mb_col > cm->mb_cols / 2)
-                        {
-                            if (d->bmi.mv.as_mv.col > 0)
-                                sum_in_vectors++;
-                            else if (d->bmi.mv.as_mv.col < 0)
-                                sum_in_vectors--;
-                        }
-                    }
-                }
+            /* Does the Row vector point inwards or outwards */
+            if (mb_col < cm->mb_cols / 2) {
+              if (d->bmi.mv.as_mv.col > 0)
+                sum_in_vectors--;
+              else if (d->bmi.mv.as_mv.col < 0)
+                sum_in_vectors++;
+            } else if (mb_col > cm->mb_cols / 2) {
+              if (d->bmi.mv.as_mv.col > 0)
+                sum_in_vectors++;
+              else if (d->bmi.mv.as_mv.col < 0)
+                sum_in_vectors--;
             }
-
-            coded_error += (int64_t)this_error;
-
-            /* adjust to the next column of macroblocks */
-            x->src.y_buffer += 16;
-            x->src.u_buffer += 8;
-            x->src.v_buffer += 8;
-
-            recon_yoffset += 16;
-            recon_uvoffset += 8;
+          }
         }
+      }
 
-        /* adjust to the next row of mbs */
-        x->src.y_buffer += 16 * x->src.y_stride - 16 * cm->mb_cols;
-        x->src.u_buffer += 8 * x->src.uv_stride - 8 * cm->mb_cols;
-        x->src.v_buffer += 8 * x->src.uv_stride - 8 * cm->mb_cols;
+      coded_error += (int64_t)this_error;
 
-        /* extend the recon for intra prediction */
-        vp8_extend_mb_row(new_yv12, xd->dst.y_buffer + 16, xd->dst.u_buffer + 8, xd->dst.v_buffer + 8);
-        vp8_clear_system_state();
+      /* adjust to the next column of macroblocks */
+      x->src.y_buffer += 16;
+      x->src.u_buffer += 8;
+      x->src.v_buffer += 8;
+
+      recon_yoffset += 16;
+      recon_uvoffset += 8;
     }
 
+    /* adjust to the next row of mbs */
+    x->src.y_buffer += 16 * x->src.y_stride - 16 * cm->mb_cols;
+    x->src.u_buffer += 8 * x->src.uv_stride - 8 * cm->mb_cols;
+    x->src.v_buffer += 8 * x->src.uv_stride - 8 * cm->mb_cols;
+
+    /* extend the recon for intra prediction */
+    vp8_extend_mb_row(new_yv12, xd->dst.y_buffer + 16, xd->dst.u_buffer + 8,
+                      xd->dst.v_buffer + 8);
     vp8_clear_system_state();
-    {
-        double weight = 0.0;
+  }
 
-        FIRSTPASS_STATS fps;
+  vp8_clear_system_state();
+  {
+    double weight = 0.0;
 
-        fps.frame      = cm->current_video_frame ;
-        fps.intra_error = (double)(intra_error >> 8);
-        fps.coded_error = (double)(coded_error >> 8);
-        weight = simple_weight(cpi->Source);
+    FIRSTPASS_STATS fps;
 
+    fps.frame = cm->current_video_frame;
+    fps.intra_error = (double)(intra_error >> 8);
+    fps.coded_error = (double)(coded_error >> 8);
+    weight = simple_weight(cpi->Source);
 
-        if (weight < 0.1)
-            weight = 0.1;
+    if (weight < 0.1) weight = 0.1;
 
-        fps.ssim_weighted_pred_err = fps.coded_error * weight;
+    fps.ssim_weighted_pred_err = fps.coded_error * weight;
 
-        fps.pcnt_inter  = 0.0;
-        fps.pcnt_motion = 0.0;
-        fps.MVr        = 0.0;
-        fps.mvr_abs     = 0.0;
-        fps.MVc        = 0.0;
-        fps.mvc_abs     = 0.0;
-        fps.MVrv       = 0.0;
-        fps.MVcv       = 0.0;
-        fps.mv_in_out_count  = 0.0;
-        fps.new_mv_count = 0.0;
-        fps.count      = 1.0;
+    fps.pcnt_inter = 0.0;
+    fps.pcnt_motion = 0.0;
+    fps.MVr = 0.0;
+    fps.mvr_abs = 0.0;
+    fps.MVc = 0.0;
+    fps.mvc_abs = 0.0;
+    fps.MVrv = 0.0;
+    fps.MVcv = 0.0;
+    fps.mv_in_out_count = 0.0;
+    fps.new_mv_count = 0.0;
+    fps.count = 1.0;
 
-        fps.pcnt_inter   = 1.0 * (double)intercount / cm->MBs;
-        fps.pcnt_second_ref = 1.0 * (double)second_ref_count / cm->MBs;
-        fps.pcnt_neutral = 1.0 * (double)neutral_count / cm->MBs;
+    fps.pcnt_inter = 1.0 * (double)intercount / cm->MBs;
+    fps.pcnt_second_ref = 1.0 * (double)second_ref_count / cm->MBs;
+    fps.pcnt_neutral = 1.0 * (double)neutral_count / cm->MBs;
 
-        if (mvcount > 0)
-        {
-            fps.MVr = (double)sum_mvr / (double)mvcount;
-            fps.mvr_abs = (double)sum_mvr_abs / (double)mvcount;
-            fps.MVc = (double)sum_mvc / (double)mvcount;
-            fps.mvc_abs = (double)sum_mvc_abs / (double)mvcount;
-            fps.MVrv = ((double)sum_mvrs - (fps.MVr * fps.MVr / (double)mvcount)) / (double)mvcount;
-            fps.MVcv = ((double)sum_mvcs - (fps.MVc * fps.MVc / (double)mvcount)) / (double)mvcount;
-            fps.mv_in_out_count = (double)sum_in_vectors / (double)(mvcount * 2);
-            fps.new_mv_count = new_mv_count;
+    if (mvcount > 0) {
+      fps.MVr = (double)sum_mvr / (double)mvcount;
+      fps.mvr_abs = (double)sum_mvr_abs / (double)mvcount;
+      fps.MVc = (double)sum_mvc / (double)mvcount;
+      fps.mvc_abs = (double)sum_mvc_abs / (double)mvcount;
+      fps.MVrv = ((double)sum_mvrs - (fps.MVr * fps.MVr / (double)mvcount)) /
+                 (double)mvcount;
+      fps.MVcv = ((double)sum_mvcs - (fps.MVc * fps.MVc / (double)mvcount)) /
+                 (double)mvcount;
+      fps.mv_in_out_count = (double)sum_in_vectors / (double)(mvcount * 2);
+      fps.new_mv_count = new_mv_count;
 
-            fps.pcnt_motion = 1.0 * (double)mvcount / cpi->common.MBs;
-        }
-
-        /* TODO:  handle the case when duration is set to 0, or something less
-         * than the full time between subsequent cpi->source_time_stamps
-         */
-        fps.duration = (double)(cpi->source->ts_end
-                       - cpi->source->ts_start);
-
-        /* don't want to do output stats with a stack variable! */
-        memcpy(&cpi->twopass.this_frame_stats,
-               &fps,
-               sizeof(FIRSTPASS_STATS));
-        output_stats(cpi, cpi->output_pkt_list, &cpi->twopass.this_frame_stats);
-        accumulate_stats(&cpi->twopass.total_stats, &fps);
+      fps.pcnt_motion = 1.0 * (double)mvcount / cpi->common.MBs;
     }
 
-    /* Copy the previous Last Frame into the GF buffer if specific
-     * conditions for doing so are met
+    /* TODO:  handle the case when duration is set to 0, or something less
+     * than the full time between subsequent cpi->source_time_stamps
      */
-    if ((cm->current_video_frame > 0) &&
-        (cpi->twopass.this_frame_stats.pcnt_inter > 0.20) &&
-        ((cpi->twopass.this_frame_stats.intra_error /
-          DOUBLE_DIVIDE_CHECK(cpi->twopass.this_frame_stats.coded_error)) >
-         2.0))
-    {
-        vp8_yv12_copy_frame(lst_yv12, gld_yv12);
-    }
+    fps.duration = (double)(cpi->source->ts_end - cpi->source->ts_start);
 
-    /* swap frame pointers so last frame refers to the frame we just
-     * compressed
-     */
-    vp8_swap_yv12_buffer(lst_yv12, new_yv12);
-    vp8_yv12_extend_frame_borders(lst_yv12);
+    /* don't want to do output stats with a stack variable! */
+    memcpy(&cpi->twopass.this_frame_stats, &fps, sizeof(FIRSTPASS_STATS));
+    output_stats(cpi, cpi->output_pkt_list, &cpi->twopass.this_frame_stats);
+    accumulate_stats(&cpi->twopass.total_stats, &fps);
+  }
 
-    /* Special case for the first frame. Copy into the GF buffer as a
-     * second reference.
-     */
-    if (cm->current_video_frame == 0)
-    {
-        vp8_yv12_copy_frame(lst_yv12, gld_yv12);
-    }
+  /* Copy the previous Last Frame into the GF buffer if specific
+   * conditions for doing so are met
+   */
+  if ((cm->current_video_frame > 0) &&
+      (cpi->twopass.this_frame_stats.pcnt_inter > 0.20) &&
+      ((cpi->twopass.this_frame_stats.intra_error /
+        DOUBLE_DIVIDE_CHECK(cpi->twopass.this_frame_stats.coded_error)) >
+       2.0)) {
+    vp8_yv12_copy_frame(lst_yv12, gld_yv12);
+  }
 
+  /* swap frame pointers so last frame refers to the frame we just
+   * compressed
+   */
+  vp8_swap_yv12_buffer(lst_yv12, new_yv12);
+  vp8_yv12_extend_frame_borders(lst_yv12);
 
-    /* use this to see what the first pass reconstruction looks like */
-    if (0)
-    {
-        char filename[512];
-        FILE *recon_file;
-        sprintf(filename, "enc%04d.yuv", (int) cm->current_video_frame);
+  /* Special case for the first frame. Copy into the GF buffer as a
+   * second reference.
+   */
+  if (cm->current_video_frame == 0) {
+    vp8_yv12_copy_frame(lst_yv12, gld_yv12);
+  }
 
-        if (cm->current_video_frame == 0)
-            recon_file = fopen(filename, "wb");
-        else
-            recon_file = fopen(filename, "ab");
+  /* use this to see what the first pass reconstruction looks like */
+  if (0) {
+    char filename[512];
+    FILE *recon_file;
+    sprintf(filename, "enc%04d.yuv", (int)cm->current_video_frame);
 
-        (void) fwrite(lst_yv12->buffer_alloc, lst_yv12->frame_size, 1,
-                      recon_file);
-        fclose(recon_file);
-    }
+    if (cm->current_video_frame == 0)
+      recon_file = fopen(filename, "wb");
+    else
+      recon_file = fopen(filename, "ab");
 
-    cm->current_video_frame++;
+    (void)fwrite(lst_yv12->buffer_alloc, lst_yv12->frame_size, 1, recon_file);
+    fclose(recon_file);
+  }
 
+  cm->current_video_frame++;
 }
 extern const int vp8_bits_per_mb[2][QINDEX_RANGE];
 
@@ -909,545 +842,519 @@
  * Currently simplistic in its assumptions for testing.
  */
 
-static double bitcost( double prob )
-{
+static double bitcost(double prob) {
   if (prob > 0.000122)
     return -log(prob) / log(2.0);
   else
     return 13.0;
 }
-static int64_t estimate_modemvcost(VP8_COMP *cpi,
-                                     FIRSTPASS_STATS * fpstats)
-{
-    int mv_cost;
-    int64_t mode_cost;
+static int64_t estimate_modemvcost(VP8_COMP *cpi, FIRSTPASS_STATS *fpstats) {
+  int mv_cost;
+  int64_t mode_cost;
 
-    double av_pct_inter = fpstats->pcnt_inter / fpstats->count;
-    double av_pct_motion = fpstats->pcnt_motion / fpstats->count;
-    double av_intra = (1.0 - av_pct_inter);
+  double av_pct_inter = fpstats->pcnt_inter / fpstats->count;
+  double av_pct_motion = fpstats->pcnt_motion / fpstats->count;
+  double av_intra = (1.0 - av_pct_inter);
 
-    double zz_cost;
-    double motion_cost;
-    double intra_cost;
+  double zz_cost;
+  double motion_cost;
+  double intra_cost;
 
-    zz_cost = bitcost(av_pct_inter - av_pct_motion);
-    motion_cost = bitcost(av_pct_motion);
-    intra_cost = bitcost(av_intra);
+  zz_cost = bitcost(av_pct_inter - av_pct_motion);
+  motion_cost = bitcost(av_pct_motion);
+  intra_cost = bitcost(av_intra);
 
-    /* Estimate of extra bits per mv overhead for mbs
-     * << 9 is the normalization to the (bits * 512) used in vp8_bits_per_mb
-     */
-    mv_cost = ((int)(fpstats->new_mv_count / fpstats->count) * 8) << 9;
+  /* Estimate of extra bits per mv overhead for mbs
+   * << 9 is the normalization to the (bits * 512) used in vp8_bits_per_mb
+   */
+  mv_cost = ((int)(fpstats->new_mv_count / fpstats->count) * 8) << 9;
 
-    /* Crude estimate of overhead cost from modes
-     * << 9 is the normalization to (bits * 512) used in vp8_bits_per_mb
-     */
-    mode_cost = (int64_t)((((av_pct_inter - av_pct_motion) * zz_cost) +
-                             (av_pct_motion * motion_cost) +
-                             (av_intra * intra_cost)) * cpi->common.MBs) * 512;
+  /* Crude estimate of overhead cost from modes
+   * << 9 is the normalization to (bits * 512) used in vp8_bits_per_mb
+   */
+  mode_cost =
+      (int64_t)((((av_pct_inter - av_pct_motion) * zz_cost) +
+                 (av_pct_motion * motion_cost) + (av_intra * intra_cost)) *
+                cpi->common.MBs) *
+      512;
 
-    return mv_cost + mode_cost;
+  return mv_cost + mode_cost;
 }
 
-static double calc_correction_factor( double err_per_mb,
-                                      double err_devisor,
-                                      double pt_low,
-                                      double pt_high,
-                                      int Q )
-{
-    double power_term;
-    double error_term = err_per_mb / err_devisor;
-    double correction_factor;
+static double calc_correction_factor(double err_per_mb, double err_devisor,
+                                     double pt_low, double pt_high, int Q) {
+  double power_term;
+  double error_term = err_per_mb / err_devisor;
+  double correction_factor;
 
-    /* Adjustment based on Q to power term. */
-    power_term = pt_low + (Q * 0.01);
-    power_term = (power_term > pt_high) ? pt_high : power_term;
+  /* Adjustment based on Q to power term. */
+  power_term = pt_low + (Q * 0.01);
+  power_term = (power_term > pt_high) ? pt_high : power_term;
 
-    /* Adjustments to error term */
-    /* TBD */
+  /* Adjustments to error term */
+  /* TBD */
 
-    /* Calculate correction factor */
-    correction_factor = pow(error_term, power_term);
+  /* Calculate correction factor */
+  correction_factor = pow(error_term, power_term);
 
-    /* Clip range */
-    correction_factor =
-        (correction_factor < 0.05)
-            ? 0.05 : (correction_factor > 5.0) ? 5.0 : correction_factor;
+  /* Clip range */
+  correction_factor = (correction_factor < 0.05)
+                          ? 0.05
+                          : (correction_factor > 5.0) ? 5.0 : correction_factor;
 
-    return correction_factor;
+  return correction_factor;
 }
 
-static int estimate_max_q(VP8_COMP *cpi,
-                          FIRSTPASS_STATS * fpstats,
-                          int section_target_bandwitdh,
-                          int overhead_bits )
-{
-    int Q;
-    int num_mbs = cpi->common.MBs;
-    int target_norm_bits_per_mb;
+static int estimate_max_q(VP8_COMP *cpi, FIRSTPASS_STATS *fpstats,
+                          int section_target_bandwitdh, int overhead_bits) {
+  int Q;
+  int num_mbs = cpi->common.MBs;
+  int target_norm_bits_per_mb;
 
-    double section_err = (fpstats->coded_error / fpstats->count);
-    double err_per_mb = section_err / num_mbs;
-    double err_correction_factor;
-    double speed_correction = 1.0;
-    int overhead_bits_per_mb;
+  double section_err = (fpstats->coded_error / fpstats->count);
+  double err_per_mb = section_err / num_mbs;
+  double err_correction_factor;
+  double speed_correction = 1.0;
+  int overhead_bits_per_mb;
 
-    if (section_target_bandwitdh <= 0)
-        return cpi->twopass.maxq_max_limit;       /* Highest value allowed */
+  if (section_target_bandwitdh <= 0)
+    return cpi->twopass.maxq_max_limit; /* Highest value allowed */
 
-    target_norm_bits_per_mb =
-        (section_target_bandwitdh < (1 << 20))
-            ? (512 * section_target_bandwitdh) / num_mbs
-            : 512 * (section_target_bandwitdh / num_mbs);
+  target_norm_bits_per_mb = (section_target_bandwitdh < (1 << 20))
+                                ? (512 * section_target_bandwitdh) / num_mbs
+                                : 512 * (section_target_bandwitdh / num_mbs);
 
-    /* Calculate a corrective factor based on a rolling ratio of bits spent
-     * vs target bits
-     */
-    if ((cpi->rolling_target_bits > 0) &&
-        (cpi->active_worst_quality < cpi->worst_quality))
-    {
-        double rolling_ratio;
+  /* Calculate a corrective factor based on a rolling ratio of bits spent
+   * vs target bits
+   */
+  if ((cpi->rolling_target_bits > 0) &&
+      (cpi->active_worst_quality < cpi->worst_quality)) {
+    double rolling_ratio;
 
-        rolling_ratio = (double)cpi->rolling_actual_bits /
-                        (double)cpi->rolling_target_bits;
+    rolling_ratio =
+        (double)cpi->rolling_actual_bits / (double)cpi->rolling_target_bits;
 
-        if (rolling_ratio < 0.95)
-            cpi->twopass.est_max_qcorrection_factor -= 0.005;
-        else if (rolling_ratio > 1.05)
-            cpi->twopass.est_max_qcorrection_factor += 0.005;
+    if (rolling_ratio < 0.95)
+      cpi->twopass.est_max_qcorrection_factor -= 0.005;
+    else if (rolling_ratio > 1.05)
+      cpi->twopass.est_max_qcorrection_factor += 0.005;
 
-        cpi->twopass.est_max_qcorrection_factor =
-            (cpi->twopass.est_max_qcorrection_factor < 0.1)
-                ? 0.1
-                : (cpi->twopass.est_max_qcorrection_factor > 10.0)
-                    ? 10.0 : cpi->twopass.est_max_qcorrection_factor;
-    }
+    cpi->twopass.est_max_qcorrection_factor =
+        (cpi->twopass.est_max_qcorrection_factor < 0.1)
+            ? 0.1
+            : (cpi->twopass.est_max_qcorrection_factor > 10.0)
+                  ? 10.0
+                  : cpi->twopass.est_max_qcorrection_factor;
+  }
 
-    /* Corrections for higher compression speed settings
-     * (reduced compression expected)
-     */
-    if ((cpi->compressor_speed == 3) || (cpi->compressor_speed == 1))
-    {
-        if (cpi->oxcf.cpu_used <= 5)
-            speed_correction = 1.04 + (cpi->oxcf.cpu_used * 0.04);
-        else
-            speed_correction = 1.25;
-    }
+  /* Corrections for higher compression speed settings
+   * (reduced compression expected)
+   */
+  if ((cpi->compressor_speed == 3) || (cpi->compressor_speed == 1)) {
+    if (cpi->oxcf.cpu_used <= 5)
+      speed_correction = 1.04 + (cpi->oxcf.cpu_used * 0.04);
+    else
+      speed_correction = 1.25;
+  }
 
-    /* Estimate of overhead bits per mb */
-    /* Correction to overhead bits for min allowed Q. */
-    overhead_bits_per_mb = overhead_bits / num_mbs;
-    overhead_bits_per_mb = (int)(overhead_bits_per_mb *
-                            pow( 0.98, (double)cpi->twopass.maxq_min_limit ));
+  /* Estimate of overhead bits per mb */
+  /* Correction to overhead bits for min allowed Q. */
+  overhead_bits_per_mb = overhead_bits / num_mbs;
+  overhead_bits_per_mb = (int)(overhead_bits_per_mb *
+                               pow(0.98, (double)cpi->twopass.maxq_min_limit));
 
-    /* Try and pick a max Q that will be high enough to encode the
-     * content at the given rate.
-     */
-    for (Q = cpi->twopass.maxq_min_limit; Q < cpi->twopass.maxq_max_limit; Q++)
-    {
-        int bits_per_mb_at_this_q;
+  /* Try and pick a max Q that will be high enough to encode the
+   * content at the given rate.
+   */
+  for (Q = cpi->twopass.maxq_min_limit; Q < cpi->twopass.maxq_max_limit; Q++) {
+    int bits_per_mb_at_this_q;
 
-        /* Error per MB based correction factor */
-        err_correction_factor =
-            calc_correction_factor(err_per_mb, 150.0, 0.40, 0.90, Q);
+    /* Error per MB based correction factor */
+    err_correction_factor =
+        calc_correction_factor(err_per_mb, 150.0, 0.40, 0.90, Q);
 
-        bits_per_mb_at_this_q =
-            vp8_bits_per_mb[INTER_FRAME][Q] + overhead_bits_per_mb;
+    bits_per_mb_at_this_q =
+        vp8_bits_per_mb[INTER_FRAME][Q] + overhead_bits_per_mb;
 
-        bits_per_mb_at_this_q = (int)(.5 + err_correction_factor
-            * speed_correction * cpi->twopass.est_max_qcorrection_factor
-            * cpi->twopass.section_max_qfactor
-            * (double)bits_per_mb_at_this_q);
+    bits_per_mb_at_this_q = (int)(.5 +
+                                  err_correction_factor * speed_correction *
+                                      cpi->twopass.est_max_qcorrection_factor *
+                                      cpi->twopass.section_max_qfactor *
+                                      (double)bits_per_mb_at_this_q);
 
-        /* Mode and motion overhead */
-        /* As Q rises in real encode loop rd code will force overhead down
-         * We make a crude adjustment for this here as *.98 per Q step.
-         */
-        overhead_bits_per_mb = (int)((double)overhead_bits_per_mb * 0.98);
+    /* Mode and motion overhead */
+    /* As Q rises in real encode loop rd code will force overhead down
+     * We make a crude adjustment for this here as *.98 per Q step.
+     */
+    overhead_bits_per_mb = (int)((double)overhead_bits_per_mb * 0.98);
 
-        if (bits_per_mb_at_this_q <= target_norm_bits_per_mb)
-            break;
-    }
+    if (bits_per_mb_at_this_q <= target_norm_bits_per_mb) break;
+  }
 
-    /* Restriction on active max q for constrained quality mode. */
-    if ( (cpi->oxcf.end_usage == USAGE_CONSTRAINED_QUALITY) &&
-         (Q < cpi->cq_target_quality) )
-    {
-        Q = cpi->cq_target_quality;
-    }
+  /* Restriction on active max q for constrained quality mode. */
+  if ((cpi->oxcf.end_usage == USAGE_CONSTRAINED_QUALITY) &&
+      (Q < cpi->cq_target_quality)) {
+    Q = cpi->cq_target_quality;
+  }
 
-    /* Adjust maxq_min_limit and maxq_max_limit limits based on
-     * average q observed in clip for non kf/gf.arf frames
-     * Give average a chance to settle though.
-     */
-    if ( (cpi->ni_frames >
-                  ((int)cpi->twopass.total_stats.count >> 8)) &&
-         (cpi->ni_frames > 150) )
-    {
-        cpi->twopass.maxq_max_limit = ((cpi->ni_av_qi + 32) < cpi->worst_quality)
-                                  ? (cpi->ni_av_qi + 32) : cpi->worst_quality;
-        cpi->twopass.maxq_min_limit = ((cpi->ni_av_qi - 32) > cpi->best_quality)
-                                  ? (cpi->ni_av_qi - 32) : cpi->best_quality;
-    }
+  /* Adjust maxq_min_limit and maxq_max_limit limits based on
+   * average q observed in clip for non kf/gf.arf frames
+   * Give average a chance to settle though.
+   */
+  if ((cpi->ni_frames > ((int)cpi->twopass.total_stats.count >> 8)) &&
+      (cpi->ni_frames > 150)) {
+    cpi->twopass.maxq_max_limit = ((cpi->ni_av_qi + 32) < cpi->worst_quality)
+                                      ? (cpi->ni_av_qi + 32)
+                                      : cpi->worst_quality;
+    cpi->twopass.maxq_min_limit = ((cpi->ni_av_qi - 32) > cpi->best_quality)
+                                      ? (cpi->ni_av_qi - 32)
+                                      : cpi->best_quality;
+  }
 
-    return Q;
+  return Q;
 }
 
 /* For cq mode estimate a cq level that matches the observed
  * complexity and data rate.
  */
-static int estimate_cq( VP8_COMP *cpi,
-                        FIRSTPASS_STATS * fpstats,
-                        int section_target_bandwitdh,
-                        int overhead_bits )
-{
-    int Q;
-    int num_mbs = cpi->common.MBs;
-    int target_norm_bits_per_mb;
+static int estimate_cq(VP8_COMP *cpi, FIRSTPASS_STATS *fpstats,
+                       int section_target_bandwitdh, int overhead_bits) {
+  int Q;
+  int num_mbs = cpi->common.MBs;
+  int target_norm_bits_per_mb;
 
-    double section_err = (fpstats->coded_error / fpstats->count);
-    double err_per_mb = section_err / num_mbs;
-    double err_correction_factor;
-    double speed_correction = 1.0;
-    double clip_iiratio;
-    double clip_iifactor;
-    int overhead_bits_per_mb;
+  double section_err = (fpstats->coded_error / fpstats->count);
+  double err_per_mb = section_err / num_mbs;
+  double err_correction_factor;
+  double speed_correction = 1.0;
+  double clip_iiratio;
+  double clip_iifactor;
+  int overhead_bits_per_mb;
 
-    if (0)
-    {
-        FILE *f = fopen("epmp.stt", "a");
-        fprintf(f, "%10.2f\n", err_per_mb );
-        fclose(f);
-    }
+  if (0) {
+    FILE *f = fopen("epmp.stt", "a");
+    fprintf(f, "%10.2f\n", err_per_mb);
+    fclose(f);
+  }
 
-    target_norm_bits_per_mb = (section_target_bandwitdh < (1 << 20))
-                              ? (512 * section_target_bandwitdh) / num_mbs
-                              : 512 * (section_target_bandwitdh / num_mbs);
+  target_norm_bits_per_mb = (section_target_bandwitdh < (1 << 20))
+                                ? (512 * section_target_bandwitdh) / num_mbs
+                                : 512 * (section_target_bandwitdh / num_mbs);
 
-    /* Estimate of overhead bits per mb */
-    overhead_bits_per_mb = overhead_bits / num_mbs;
+  /* Estimate of overhead bits per mb */
+  overhead_bits_per_mb = overhead_bits / num_mbs;
 
-    /* Corrections for higher compression speed settings
-     * (reduced compression expected)
-     */
-    if ((cpi->compressor_speed == 3) || (cpi->compressor_speed == 1))
-    {
-        if (cpi->oxcf.cpu_used <= 5)
-            speed_correction = 1.04 + (cpi->oxcf.cpu_used * 0.04);
-        else
-            speed_correction = 1.25;
-    }
+  /* Corrections for higher compression speed settings
+   * (reduced compression expected)
+   */
+  if ((cpi->compressor_speed == 3) || (cpi->compressor_speed == 1)) {
+    if (cpi->oxcf.cpu_used <= 5)
+      speed_correction = 1.04 + (cpi->oxcf.cpu_used * 0.04);
+    else
+      speed_correction = 1.25;
+  }
 
-    /* II ratio correction factor for clip as a whole */
-    clip_iiratio = cpi->twopass.total_stats.intra_error /
-                   DOUBLE_DIVIDE_CHECK(cpi->twopass.total_stats.coded_error);
-    clip_iifactor = 1.0 - ((clip_iiratio - 10.0) * 0.025);
-    if (clip_iifactor < 0.80)
-        clip_iifactor = 0.80;
+  /* II ratio correction factor for clip as a whole */
+  clip_iiratio = cpi->twopass.total_stats.intra_error /
+                 DOUBLE_DIVIDE_CHECK(cpi->twopass.total_stats.coded_error);
+  clip_iifactor = 1.0 - ((clip_iiratio - 10.0) * 0.025);
+  if (clip_iifactor < 0.80) clip_iifactor = 0.80;
 
-    /* Try and pick a Q that can encode the content at the given rate. */
-    for (Q = 0; Q < MAXQ; Q++)
-    {
-        int bits_per_mb_at_this_q;
+  /* Try and pick a Q that can encode the content at the given rate. */
+  for (Q = 0; Q < MAXQ; Q++) {
+    int bits_per_mb_at_this_q;
 
-        /* Error per MB based correction factor */
-        err_correction_factor =
-            calc_correction_factor(err_per_mb, 100.0, 0.40, 0.90, Q);
+    /* Error per MB based correction factor */
+    err_correction_factor =
+        calc_correction_factor(err_per_mb, 100.0, 0.40, 0.90, Q);
 
-        bits_per_mb_at_this_q =
-            vp8_bits_per_mb[INTER_FRAME][Q] + overhead_bits_per_mb;
+    bits_per_mb_at_this_q =
+        vp8_bits_per_mb[INTER_FRAME][Q] + overhead_bits_per_mb;
 
-        bits_per_mb_at_this_q =
-            (int)( .5 + err_correction_factor *
-                        speed_correction *
-                        clip_iifactor *
-                        (double)bits_per_mb_at_this_q);
+    bits_per_mb_at_this_q =
+        (int)(.5 +
+              err_correction_factor * speed_correction * clip_iifactor *
+                  (double)bits_per_mb_at_this_q);
 
-        /* Mode and motion overhead */
-        /* As Q rises in real encode loop rd code will force overhead down
-         * We make a crude adjustment for this here as *.98 per Q step.
-         */
-        overhead_bits_per_mb = (int)((double)overhead_bits_per_mb * 0.98);
+    /* Mode and motion overhead */
+    /* As Q rises in real encode loop rd code will force overhead down
+     * We make a crude adjustment for this here as *.98 per Q step.
+     */
+    overhead_bits_per_mb = (int)((double)overhead_bits_per_mb * 0.98);
 
-        if (bits_per_mb_at_this_q <= target_norm_bits_per_mb)
-            break;
-    }
+    if (bits_per_mb_at_this_q <= target_norm_bits_per_mb) break;
+  }
 
-    /* Clip value to range "best allowed to (worst allowed - 1)" */
-    Q = cq_level[Q];
-    if ( Q >= cpi->worst_quality )
-        Q = cpi->worst_quality - 1;
-    if ( Q < cpi->best_quality )
-        Q = cpi->best_quality;
+  /* Clip value to range "best allowed to (worst allowed - 1)" */
+  Q = cq_level[Q];
+  if (Q >= cpi->worst_quality) Q = cpi->worst_quality - 1;
+  if (Q < cpi->best_quality) Q = cpi->best_quality;
 
-    return Q;
+  return Q;
 }
 
-static int estimate_q(VP8_COMP *cpi, double section_err, int section_target_bandwitdh)
-{
-    int Q;
-    int num_mbs = cpi->common.MBs;
-    int target_norm_bits_per_mb;
+static int estimate_q(VP8_COMP *cpi, double section_err,
+                      int section_target_bandwitdh) {
+  int Q;
+  int num_mbs = cpi->common.MBs;
+  int target_norm_bits_per_mb;
 
-    double err_per_mb = section_err / num_mbs;
-    double err_correction_factor;
-    double speed_correction = 1.0;
+  double err_per_mb = section_err / num_mbs;
+  double err_correction_factor;
+  double speed_correction = 1.0;
 
-    target_norm_bits_per_mb = (section_target_bandwitdh < (1 << 20)) ? (512 * section_target_bandwitdh) / num_mbs : 512 * (section_target_bandwitdh / num_mbs);
+  target_norm_bits_per_mb = (section_target_bandwitdh < (1 << 20))
+                                ? (512 * section_target_bandwitdh) / num_mbs
+                                : 512 * (section_target_bandwitdh / num_mbs);
 
-    /* Corrections for higher compression speed settings
-     * (reduced compression expected)
-     */
-    if ((cpi->compressor_speed == 3) || (cpi->compressor_speed == 1))
-    {
-        if (cpi->oxcf.cpu_used <= 5)
-            speed_correction = 1.04 + (cpi->oxcf.cpu_used * 0.04);
-        else
-            speed_correction = 1.25;
-    }
+  /* Corrections for higher compression speed settings
+   * (reduced compression expected)
+   */
+  if ((cpi->compressor_speed == 3) || (cpi->compressor_speed == 1)) {
+    if (cpi->oxcf.cpu_used <= 5)
+      speed_correction = 1.04 + (cpi->oxcf.cpu_used * 0.04);
+    else
+      speed_correction = 1.25;
+  }
 
-    /* Try and pick a Q that can encode the content at the given rate. */
-    for (Q = 0; Q < MAXQ; Q++)
-    {
-        int bits_per_mb_at_this_q;
+  /* Try and pick a Q that can encode the content at the given rate. */
+  for (Q = 0; Q < MAXQ; Q++) {
+    int bits_per_mb_at_this_q;
 
-        /* Error per MB based correction factor */
-        err_correction_factor =
-            calc_correction_factor(err_per_mb, 150.0, 0.40, 0.90, Q);
+    /* Error per MB based correction factor */
+    err_correction_factor =
+        calc_correction_factor(err_per_mb, 150.0, 0.40, 0.90, Q);
 
-        bits_per_mb_at_this_q =
-            (int)( .5 + ( err_correction_factor *
-                          speed_correction *
-                          cpi->twopass.est_max_qcorrection_factor *
-                          (double)vp8_bits_per_mb[INTER_FRAME][Q] / 1.0 ) );
+    bits_per_mb_at_this_q =
+        (int)(.5 + (err_correction_factor * speed_correction *
+                    cpi->twopass.est_max_qcorrection_factor *
+                    (double)vp8_bits_per_mb[INTER_FRAME][Q] / 1.0));
 
-        if (bits_per_mb_at_this_q <= target_norm_bits_per_mb)
-            break;
-    }
+    if (bits_per_mb_at_this_q <= target_norm_bits_per_mb) break;
+  }
 
-    return Q;
+  return Q;
 }
 
 /* Estimate a worst case Q for a KF group */
-static int estimate_kf_group_q(VP8_COMP *cpi, double section_err, int section_target_bandwitdh, double group_iiratio)
-{
-    int Q;
-    int num_mbs = cpi->common.MBs;
-    int target_norm_bits_per_mb = (512 * section_target_bandwitdh) / num_mbs;
-    int bits_per_mb_at_this_q;
+static int estimate_kf_group_q(VP8_COMP *cpi, double section_err,
+                               int section_target_bandwitdh,
+                               double group_iiratio) {
+  int Q;
+  int num_mbs = cpi->common.MBs;
+  int target_norm_bits_per_mb = (512 * section_target_bandwitdh) / num_mbs;
+  int bits_per_mb_at_this_q;
 
-    double err_per_mb = section_err / num_mbs;
-    double err_correction_factor;
-    double speed_correction = 1.0;
-    double current_spend_ratio = 1.0;
+  double err_per_mb = section_err / num_mbs;
+  double err_correction_factor;
+  double speed_correction = 1.0;
+  double current_spend_ratio = 1.0;
 
-    double pow_highq = (POW1 < 0.6) ? POW1 + 0.3 : 0.90;
-    double pow_lowq = (POW1 < 0.7) ? POW1 + 0.1 : 0.80;
+  double pow_highq = (POW1 < 0.6) ? POW1 + 0.3 : 0.90;
+  double pow_lowq = (POW1 < 0.7) ? POW1 + 0.1 : 0.80;
 
-    double iiratio_correction_factor = 1.0;
+  double iiratio_correction_factor = 1.0;
 
-    double combined_correction_factor;
+  double combined_correction_factor;
 
-    /* Trap special case where the target is <= 0 */
-    if (target_norm_bits_per_mb <= 0)
-        return MAXQ * 2;
+  /* Trap special case where the target is <= 0 */
+  if (target_norm_bits_per_mb <= 0) return MAXQ * 2;
 
-    /* Calculate a corrective factor based on a rolling ratio of bits spent
-     *  vs target bits
-     * This is clamped to the range 0.1 to 10.0
-     */
-    if (cpi->long_rolling_target_bits <= 0)
-        current_spend_ratio = 10.0;
-    else
-    {
-        current_spend_ratio = (double)cpi->long_rolling_actual_bits / (double)cpi->long_rolling_target_bits;
-        current_spend_ratio = (current_spend_ratio > 10.0) ? 10.0 : (current_spend_ratio < 0.1) ? 0.1 : current_spend_ratio;
-    }
+  /* Calculate a corrective factor based on a rolling ratio of bits spent
+   *  vs target bits
+   * This is clamped to the range 0.1 to 10.0
+   */
+  if (cpi->long_rolling_target_bits <= 0)
+    current_spend_ratio = 10.0;
+  else {
+    current_spend_ratio = (double)cpi->long_rolling_actual_bits /
+                          (double)cpi->long_rolling_target_bits;
+    current_spend_ratio =
+        (current_spend_ratio > 10.0) ? 10.0 : (current_spend_ratio < 0.1)
+                                                  ? 0.1
+                                                  : current_spend_ratio;
+  }
 
-    /* Calculate a correction factor based on the quality of prediction in
-     * the sequence as indicated by intra_inter error score ratio (IIRatio)
-     * The idea here is to favour subsampling in the hardest sections vs
-     * the easyest.
-     */
-    iiratio_correction_factor = 1.0 - ((group_iiratio - 6.0) * 0.1);
+  /* Calculate a correction factor based on the quality of prediction in
+   * the sequence as indicated by intra_inter error score ratio (IIRatio)
+   * The idea here is to favour subsampling in the hardest sections vs
+   * the easyest.
+   */
+  iiratio_correction_factor = 1.0 - ((group_iiratio - 6.0) * 0.1);
 
-    if (iiratio_correction_factor < 0.5)
-        iiratio_correction_factor = 0.5;
+  if (iiratio_correction_factor < 0.5) iiratio_correction_factor = 0.5;
 
-    /* Corrections for higher compression speed settings
-     * (reduced compression expected)
-     */
-    if ((cpi->compressor_speed == 3) || (cpi->compressor_speed == 1))
-    {
-        if (cpi->oxcf.cpu_used <= 5)
-            speed_correction = 1.04 + (cpi->oxcf.cpu_used * 0.04);
-        else
-            speed_correction = 1.25;
-    }
+  /* Corrections for higher compression speed settings
+   * (reduced compression expected)
+   */
+  if ((cpi->compressor_speed == 3) || (cpi->compressor_speed == 1)) {
+    if (cpi->oxcf.cpu_used <= 5)
+      speed_correction = 1.04 + (cpi->oxcf.cpu_used * 0.04);
+    else
+      speed_correction = 1.25;
+  }
 
-    /* Combine the various factors calculated above */
-    combined_correction_factor = speed_correction * iiratio_correction_factor * current_spend_ratio;
+  /* Combine the various factors calculated above */
+  combined_correction_factor =
+      speed_correction * iiratio_correction_factor * current_spend_ratio;
 
-    /* Try and pick a Q that should be high enough to encode the content at
-     * the given rate.
-     */
-    for (Q = 0; Q < MAXQ; Q++)
-    {
-        /* Error per MB based correction factor */
-        err_correction_factor =
-            calc_correction_factor(err_per_mb, 150.0, pow_lowq, pow_highq, Q);
+  /* Try and pick a Q that should be high enough to encode the content at
+   * the given rate.
+   */
+  for (Q = 0; Q < MAXQ; Q++) {
+    /* Error per MB based correction factor */
+    err_correction_factor =
+        calc_correction_factor(err_per_mb, 150.0, pow_lowq, pow_highq, Q);
 
-        bits_per_mb_at_this_q =
-            (int)(.5 + ( err_correction_factor *
-                         combined_correction_factor *
-                         (double)vp8_bits_per_mb[INTER_FRAME][Q]) );
+    bits_per_mb_at_this_q =
+        (int)(.5 + (err_correction_factor * combined_correction_factor *
+                    (double)vp8_bits_per_mb[INTER_FRAME][Q]));
 
-        if (bits_per_mb_at_this_q <= target_norm_bits_per_mb)
-            break;
-    }
+    if (bits_per_mb_at_this_q <= target_norm_bits_per_mb) break;
+  }
 
-    /* If we could not hit the target even at Max Q then estimate what Q
-     * would have been required
-     */
-    while ((bits_per_mb_at_this_q > target_norm_bits_per_mb)  && (Q < (MAXQ * 2)))
-    {
+  /* If we could not hit the target even at Max Q then estimate what Q
+   * would have been required
+   */
+  while ((bits_per_mb_at_this_q > target_norm_bits_per_mb) &&
+         (Q < (MAXQ * 2))) {
+    bits_per_mb_at_this_q = (int)(0.96 * bits_per_mb_at_this_q);
+    Q++;
+  }
 
-        bits_per_mb_at_this_q = (int)(0.96 * bits_per_mb_at_this_q);
-        Q++;
-    }
+  if (0) {
+    FILE *f = fopen("estkf_q.stt", "a");
+    fprintf(f, "%8d %8d %8d %8.2f %8.3f %8.2f %8.3f %8.3f %8.3f %8d\n",
+            cpi->common.current_video_frame, bits_per_mb_at_this_q,
+            target_norm_bits_per_mb, err_per_mb, err_correction_factor,
+            current_spend_ratio, group_iiratio, iiratio_correction_factor,
+            (double)cpi->buffer_level / (double)cpi->oxcf.optimal_buffer_level,
+            Q);
+    fclose(f);
+  }
 
-    if (0)
-    {
-        FILE *f = fopen("estkf_q.stt", "a");
-        fprintf(f, "%8d %8d %8d %8.2f %8.3f %8.2f %8.3f %8.3f %8.3f %8d\n", cpi->common.current_video_frame, bits_per_mb_at_this_q,
-                target_norm_bits_per_mb, err_per_mb, err_correction_factor,
-                current_spend_ratio, group_iiratio, iiratio_correction_factor,
-                (double)cpi->buffer_level / (double)cpi->oxcf.optimal_buffer_level, Q);
-        fclose(f);
-    }
-
-    return Q;
+  return Q;
 }
 
-void vp8_init_second_pass(VP8_COMP *cpi)
-{
-    FIRSTPASS_STATS this_frame;
-    FIRSTPASS_STATS *start_pos;
+void vp8_init_second_pass(VP8_COMP *cpi) {
+  FIRSTPASS_STATS this_frame;
+  FIRSTPASS_STATS *start_pos;
 
-    double two_pass_min_rate = (double)(cpi->oxcf.target_bandwidth * cpi->oxcf.two_pass_vbrmin_section / 100);
+  double two_pass_min_rate = (double)(cpi->oxcf.target_bandwidth *
+                                      cpi->oxcf.two_pass_vbrmin_section / 100);
 
-    zero_stats(&cpi->twopass.total_stats);
-    zero_stats(&cpi->twopass.total_left_stats);
+  zero_stats(&cpi->twopass.total_stats);
+  zero_stats(&cpi->twopass.total_left_stats);
 
-    if (!cpi->twopass.stats_in_end)
-        return;
+  if (!cpi->twopass.stats_in_end) return;
 
-    cpi->twopass.total_stats = *cpi->twopass.stats_in_end;
-    cpi->twopass.total_left_stats = cpi->twopass.total_stats;
+  cpi->twopass.total_stats = *cpi->twopass.stats_in_end;
+  cpi->twopass.total_left_stats = cpi->twopass.total_stats;
 
-    /* each frame can have a different duration, as the frame rate in the
-     * source isn't guaranteed to be constant.   The frame rate prior to
-     * the first frame encoded in the second pass is a guess.  However the
-     * sum duration is not. Its calculated based on the actual durations of
-     * all frames from the first pass.
-     */
-    vp8_new_framerate(cpi, 10000000.0 * cpi->twopass.total_stats.count / cpi->twopass.total_stats.duration);
+  /* each frame can have a different duration, as the frame rate in the
+   * source isn't guaranteed to be constant.   The frame rate prior to
+   * the first frame encoded in the second pass is a guess.  However the
+   * sum duration is not. Its calculated based on the actual durations of
+   * all frames from the first pass.
+   */
+  vp8_new_framerate(cpi, 10000000.0 * cpi->twopass.total_stats.count /
+                             cpi->twopass.total_stats.duration);
 
-    cpi->output_framerate = cpi->framerate;
-    cpi->twopass.bits_left = (int64_t)(cpi->twopass.total_stats.duration * cpi->oxcf.target_bandwidth / 10000000.0) ;
-    cpi->twopass.bits_left -= (int64_t)(cpi->twopass.total_stats.duration * two_pass_min_rate / 10000000.0);
+  cpi->output_framerate = cpi->framerate;
+  cpi->twopass.bits_left = (int64_t)(cpi->twopass.total_stats.duration *
+                                     cpi->oxcf.target_bandwidth / 10000000.0);
+  cpi->twopass.bits_left -= (int64_t)(cpi->twopass.total_stats.duration *
+                                      two_pass_min_rate / 10000000.0);
 
-    /* Calculate a minimum intra value to be used in determining the IIratio
-     * scores used in the second pass. We have this minimum to make sure
-     * that clips that are static but "low complexity" in the intra domain
-     * are still boosted appropriately for KF/GF/ARF
-     */
-    cpi->twopass.kf_intra_err_min = KF_MB_INTRA_MIN * cpi->common.MBs;
-    cpi->twopass.gf_intra_err_min = GF_MB_INTRA_MIN * cpi->common.MBs;
+  /* Calculate a minimum intra value to be used in determining the IIratio
+   * scores used in the second pass. We have this minimum to make sure
+   * that clips that are static but "low complexity" in the intra domain
+   * are still boosted appropriately for KF/GF/ARF
+   */
+  cpi->twopass.kf_intra_err_min = KF_MB_INTRA_MIN * cpi->common.MBs;
+  cpi->twopass.gf_intra_err_min = GF_MB_INTRA_MIN * cpi->common.MBs;
 
-    /* Scan the first pass file and calculate an average Intra / Inter error
-     * score ratio for the sequence
-     */
-    {
-        double sum_iiratio = 0.0;
-        double IIRatio;
+  /* Scan the first pass file and calculate an average Intra / Inter error
+   * score ratio for the sequence
+   */
+  {
+    double sum_iiratio = 0.0;
+    double IIRatio;
 
-        start_pos = cpi->twopass.stats_in; /* Note starting "file" position */
+    start_pos = cpi->twopass.stats_in; /* Note starting "file" position */
 
-        while (input_stats(cpi, &this_frame) != EOF)
-        {
-            IIRatio = this_frame.intra_error / DOUBLE_DIVIDE_CHECK(this_frame.coded_error);
-            IIRatio = (IIRatio < 1.0) ? 1.0 : (IIRatio > 20.0) ? 20.0 : IIRatio;
-            sum_iiratio += IIRatio;
-        }
-
-        cpi->twopass.avg_iiratio = sum_iiratio / DOUBLE_DIVIDE_CHECK((double)cpi->twopass.total_stats.count);
-
-        /* Reset file position */
-        reset_fpf_position(cpi, start_pos);
+    while (input_stats(cpi, &this_frame) != EOF) {
+      IIRatio =
+          this_frame.intra_error / DOUBLE_DIVIDE_CHECK(this_frame.coded_error);
+      IIRatio = (IIRatio < 1.0) ? 1.0 : (IIRatio > 20.0) ? 20.0 : IIRatio;
+      sum_iiratio += IIRatio;
     }
 
-    /* Scan the first pass file and calculate a modified total error based
-     * upon the bias/power function used to allocate bits
-     */
-    {
-        start_pos = cpi->twopass.stats_in;  /* Note starting "file" position */
+    cpi->twopass.avg_iiratio =
+        sum_iiratio /
+        DOUBLE_DIVIDE_CHECK((double)cpi->twopass.total_stats.count);
 
-        cpi->twopass.modified_error_total = 0.0;
-        cpi->twopass.modified_error_used = 0.0;
+    /* Reset file position */
+    reset_fpf_position(cpi, start_pos);
+  }
 
-        while (input_stats(cpi, &this_frame) != EOF)
-        {
-            cpi->twopass.modified_error_total += calculate_modified_err(cpi, &this_frame);
-        }
-        cpi->twopass.modified_error_left = cpi->twopass.modified_error_total;
+  /* Scan the first pass file and calculate a modified total error based
+   * upon the bias/power function used to allocate bits
+   */
+  {
+    start_pos = cpi->twopass.stats_in; /* Note starting "file" position */
 
-        reset_fpf_position(cpi, start_pos);  /* Reset file position */
+    cpi->twopass.modified_error_total = 0.0;
+    cpi->twopass.modified_error_used = 0.0;
 
+    while (input_stats(cpi, &this_frame) != EOF) {
+      cpi->twopass.modified_error_total +=
+          calculate_modified_err(cpi, &this_frame);
     }
-}
+    cpi->twopass.modified_error_left = cpi->twopass.modified_error_total;
 
-void vp8_end_second_pass(VP8_COMP *cpi)
-{
-  (void)cpi;
+    reset_fpf_position(cpi, start_pos); /* Reset file position */
+  }
 }
 
+void vp8_end_second_pass(VP8_COMP *cpi) { (void)cpi; }
+
 /* This function gives and estimate of how badly we believe the prediction
  * quality is decaying from frame to frame.
  */
-static double get_prediction_decay_rate(VP8_COMP *cpi, FIRSTPASS_STATS *next_frame)
-{
-    double prediction_decay_rate;
-    double motion_decay;
-    double motion_pct = next_frame->pcnt_motion;
-    (void)cpi;
+static double get_prediction_decay_rate(VP8_COMP *cpi,
+                                        FIRSTPASS_STATS *next_frame) {
+  double prediction_decay_rate;
+  double motion_decay;
+  double motion_pct = next_frame->pcnt_motion;
+  (void)cpi;
 
-    /* Initial basis is the % mbs inter coded */
-    prediction_decay_rate = next_frame->pcnt_inter;
+  /* Initial basis is the % mbs inter coded */
+  prediction_decay_rate = next_frame->pcnt_inter;
 
-    /* High % motion -> somewhat higher decay rate */
-    motion_decay = (1.0 - (motion_pct / 20.0));
-    if (motion_decay < prediction_decay_rate)
-        prediction_decay_rate = motion_decay;
+  /* High % motion -> somewhat higher decay rate */
+  motion_decay = (1.0 - (motion_pct / 20.0));
+  if (motion_decay < prediction_decay_rate)
+    prediction_decay_rate = motion_decay;
 
-    /* Adjustment to decay rate based on speed of motion */
-    {
-        double this_mv_rabs;
-        double this_mv_cabs;
-        double distance_factor;
+  /* Adjustment to decay rate based on speed of motion */
+  {
+    double this_mv_rabs;
+    double this_mv_cabs;
+    double distance_factor;
 
-        this_mv_rabs = fabs(next_frame->mvr_abs * motion_pct);
-        this_mv_cabs = fabs(next_frame->mvc_abs * motion_pct);
+    this_mv_rabs = fabs(next_frame->mvr_abs * motion_pct);
+    this_mv_cabs = fabs(next_frame->mvc_abs * motion_pct);
 
-        distance_factor = sqrt((this_mv_rabs * this_mv_rabs) +
-                               (this_mv_cabs * this_mv_cabs)) / 250.0;
-        distance_factor = ((distance_factor > 1.0)
-                                ? 0.0 : (1.0 - distance_factor));
-        if (distance_factor < prediction_decay_rate)
-            prediction_decay_rate = distance_factor;
-    }
+    distance_factor =
+        sqrt((this_mv_rabs * this_mv_rabs) + (this_mv_cabs * this_mv_cabs)) /
+        250.0;
+    distance_factor = ((distance_factor > 1.0) ? 0.0 : (1.0 - distance_factor));
+    if (distance_factor < prediction_decay_rate)
+      prediction_decay_rate = distance_factor;
+  }
 
-    return prediction_decay_rate;
+  return prediction_decay_rate;
 }
 
 /* Function to test for a condition where a complex transition is followed
@@ -1454,47 +1361,38 @@
  * by a static section. For example in slide shows where there is a fade
  * between slides. This is to help with more optimal kf and gf positioning.
  */
-static int detect_transition_to_still(
-    VP8_COMP *cpi,
-    int frame_interval,
-    int still_interval,
-    double loop_decay_rate,
-    double decay_accumulator )
-{
-    int trans_to_still = 0;
+static int detect_transition_to_still(VP8_COMP *cpi, int frame_interval,
+                                      int still_interval,
+                                      double loop_decay_rate,
+                                      double decay_accumulator) {
+  int trans_to_still = 0;
 
-    /* Break clause to detect very still sections after motion
-     * For example a static image after a fade or other transition
-     * instead of a clean scene cut.
-     */
-    if ( (frame_interval > MIN_GF_INTERVAL) &&
-         (loop_decay_rate >= 0.999) &&
-         (decay_accumulator < 0.9) )
-    {
-        int j;
-        FIRSTPASS_STATS * position = cpi->twopass.stats_in;
-        FIRSTPASS_STATS tmp_next_frame;
-        double decay_rate;
+  /* Break clause to detect very still sections after motion
+   * For example a static image after a fade or other transition
+   * instead of a clean scene cut.
+   */
+  if ((frame_interval > MIN_GF_INTERVAL) && (loop_decay_rate >= 0.999) &&
+      (decay_accumulator < 0.9)) {
+    int j;
+    FIRSTPASS_STATS *position = cpi->twopass.stats_in;
+    FIRSTPASS_STATS tmp_next_frame;
+    double decay_rate;
 
-        /* Look ahead a few frames to see if static condition persists... */
-        for ( j = 0; j < still_interval; j++ )
-        {
-            if (EOF == input_stats(cpi, &tmp_next_frame))
-                break;
+    /* Look ahead a few frames to see if static condition persists... */
+    for (j = 0; j < still_interval; j++) {
+      if (EOF == input_stats(cpi, &tmp_next_frame)) break;
 
-            decay_rate = get_prediction_decay_rate(cpi, &tmp_next_frame);
-            if ( decay_rate < 0.999 )
-                break;
-        }
-        /* Reset file position */
-        reset_fpf_position(cpi, position);
-
-        /* Only if it does do we signal a transition to still */
-        if ( j == still_interval )
-            trans_to_still = 1;
+      decay_rate = get_prediction_decay_rate(cpi, &tmp_next_frame);
+      if (decay_rate < 0.999) break;
     }
+    /* Reset file position */
+    reset_fpf_position(cpi, position);
 
-    return trans_to_still;
+    /* Only if it does do we signal a transition to still */
+    if (j == still_interval) trans_to_still = 1;
+  }
+
+  return trans_to_still;
 }
 
 /* This function detects a flash through the high relative pcnt_second_ref
@@ -1501,1577 +1399,1418 @@
  * score in the frame following a flash frame. The offset passed in should
  * reflect this
  */
-static int detect_flash( VP8_COMP *cpi, int offset )
-{
-    FIRSTPASS_STATS next_frame;
+static int detect_flash(VP8_COMP *cpi, int offset) {
+  FIRSTPASS_STATS next_frame;
 
-    int flash_detected = 0;
+  int flash_detected = 0;
 
-    /* Read the frame data. */
-    /* The return is 0 (no flash detected) if not a valid frame */
-    if ( read_frame_stats(cpi, &next_frame, offset) != EOF )
-    {
-        /* What we are looking for here is a situation where there is a
-         * brief break in prediction (such as a flash) but subsequent frames
-         * are reasonably well predicted by an earlier (pre flash) frame.
-         * The recovery after a flash is indicated by a high pcnt_second_ref
-         * comapred to pcnt_inter.
-         */
-        if ( (next_frame.pcnt_second_ref > next_frame.pcnt_inter) &&
-             (next_frame.pcnt_second_ref >= 0.5 ) )
-        {
-            flash_detected = 1;
+  /* Read the frame data. */
+  /* The return is 0 (no flash detected) if not a valid frame */
+  if (read_frame_stats(cpi, &next_frame, offset) != EOF) {
+    /* What we are looking for here is a situation where there is a
+     * brief break in prediction (such as a flash) but subsequent frames
+     * are reasonably well predicted by an earlier (pre flash) frame.
+     * The recovery after a flash is indicated by a high pcnt_second_ref
+     * comapred to pcnt_inter.
+     */
+    if ((next_frame.pcnt_second_ref > next_frame.pcnt_inter) &&
+        (next_frame.pcnt_second_ref >= 0.5)) {
+      flash_detected = 1;
 
-            /*if (1)
-            {
-                FILE *f = fopen("flash.stt", "a");
-                fprintf(f, "%8.0f %6.2f %6.2f\n",
-                    next_frame.frame,
-                    next_frame.pcnt_inter,
-                    next_frame.pcnt_second_ref);
-                fclose(f);
-            }*/
-        }
+      /*if (1)
+      {
+          FILE *f = fopen("flash.stt", "a");
+          fprintf(f, "%8.0f %6.2f %6.2f\n",
+              next_frame.frame,
+              next_frame.pcnt_inter,
+              next_frame.pcnt_second_ref);
+          fclose(f);
+      }*/
     }
+  }
 
-    return flash_detected;
+  return flash_detected;
 }
 
 /* Update the motion related elements to the GF arf boost calculation */
-static void accumulate_frame_motion_stats(
-    VP8_COMP *cpi,
-    FIRSTPASS_STATS * this_frame,
-    double * this_frame_mv_in_out,
-    double * mv_in_out_accumulator,
-    double * abs_mv_in_out_accumulator,
-    double * mv_ratio_accumulator )
-{
-    double this_frame_mvr_ratio;
-    double this_frame_mvc_ratio;
-    double motion_pct;
-    (void)cpi;
+static void accumulate_frame_motion_stats(VP8_COMP *cpi,
+                                          FIRSTPASS_STATS *this_frame,
+                                          double *this_frame_mv_in_out,
+                                          double *mv_in_out_accumulator,
+                                          double *abs_mv_in_out_accumulator,
+                                          double *mv_ratio_accumulator) {
+  double this_frame_mvr_ratio;
+  double this_frame_mvc_ratio;
+  double motion_pct;
+  (void)cpi;
 
-    /* Accumulate motion stats. */
-    motion_pct = this_frame->pcnt_motion;
+  /* Accumulate motion stats. */
+  motion_pct = this_frame->pcnt_motion;
 
-    /* Accumulate Motion In/Out of frame stats */
-    *this_frame_mv_in_out = this_frame->mv_in_out_count * motion_pct;
-    *mv_in_out_accumulator += this_frame->mv_in_out_count * motion_pct;
-    *abs_mv_in_out_accumulator +=
-        fabs(this_frame->mv_in_out_count * motion_pct);
+  /* Accumulate Motion In/Out of frame stats */
+  *this_frame_mv_in_out = this_frame->mv_in_out_count * motion_pct;
+  *mv_in_out_accumulator += this_frame->mv_in_out_count * motion_pct;
+  *abs_mv_in_out_accumulator += fabs(this_frame->mv_in_out_count * motion_pct);
 
-    /* Accumulate a measure of how uniform (or conversely how random)
-     * the motion field is. (A ratio of absmv / mv)
-     */
-    if (motion_pct > 0.05)
-    {
-        this_frame_mvr_ratio = fabs(this_frame->mvr_abs) /
-                               DOUBLE_DIVIDE_CHECK(fabs(this_frame->MVr));
+  /* Accumulate a measure of how uniform (or conversely how random)
+   * the motion field is. (A ratio of absmv / mv)
+   */
+  if (motion_pct > 0.05) {
+    this_frame_mvr_ratio =
+        fabs(this_frame->mvr_abs) / DOUBLE_DIVIDE_CHECK(fabs(this_frame->MVr));
 
-        this_frame_mvc_ratio = fabs(this_frame->mvc_abs) /
-                               DOUBLE_DIVIDE_CHECK(fabs(this_frame->MVc));
+    this_frame_mvc_ratio =
+        fabs(this_frame->mvc_abs) / DOUBLE_DIVIDE_CHECK(fabs(this_frame->MVc));
 
-         *mv_ratio_accumulator +=
-            (this_frame_mvr_ratio < this_frame->mvr_abs)
-                ? (this_frame_mvr_ratio * motion_pct)
-                : this_frame->mvr_abs * motion_pct;
+    *mv_ratio_accumulator += (this_frame_mvr_ratio < this_frame->mvr_abs)
+                                 ? (this_frame_mvr_ratio * motion_pct)
+                                 : this_frame->mvr_abs * motion_pct;
 
-        *mv_ratio_accumulator +=
-            (this_frame_mvc_ratio < this_frame->mvc_abs)
-                ? (this_frame_mvc_ratio * motion_pct)
-                : this_frame->mvc_abs * motion_pct;
-
-    }
+    *mv_ratio_accumulator += (this_frame_mvc_ratio < this_frame->mvc_abs)
+                                 ? (this_frame_mvc_ratio * motion_pct)
+                                 : this_frame->mvc_abs * motion_pct;
+  }
 }
 
 /* Calculate a baseline boost number for the current frame. */
-static double calc_frame_boost(
-    VP8_COMP *cpi,
-    FIRSTPASS_STATS * this_frame,
-    double this_frame_mv_in_out )
-{
-    double frame_boost;
+static double calc_frame_boost(VP8_COMP *cpi, FIRSTPASS_STATS *this_frame,
+                               double this_frame_mv_in_out) {
+  double frame_boost;
 
-    /* Underlying boost factor is based on inter intra error ratio */
-    if (this_frame->intra_error > cpi->twopass.gf_intra_err_min)
-        frame_boost = (IIFACTOR * this_frame->intra_error /
-                      DOUBLE_DIVIDE_CHECK(this_frame->coded_error));
-    else
-        frame_boost = (IIFACTOR * cpi->twopass.gf_intra_err_min /
-                      DOUBLE_DIVIDE_CHECK(this_frame->coded_error));
+  /* Underlying boost factor is based on inter intra error ratio */
+  if (this_frame->intra_error > cpi->twopass.gf_intra_err_min)
+    frame_boost = (IIFACTOR * this_frame->intra_error /
+                   DOUBLE_DIVIDE_CHECK(this_frame->coded_error));
+  else
+    frame_boost = (IIFACTOR * cpi->twopass.gf_intra_err_min /
+                   DOUBLE_DIVIDE_CHECK(this_frame->coded_error));
 
-    /* Increase boost for frames where new data coming into frame
-     * (eg zoom out). Slightly reduce boost if there is a net balance
-     * of motion out of the frame (zoom in).
-     * The range for this_frame_mv_in_out is -1.0 to +1.0
-     */
-    if (this_frame_mv_in_out > 0.0)
-        frame_boost += frame_boost * (this_frame_mv_in_out * 2.0);
-    /* In extreme case boost is halved */
-    else
-        frame_boost += frame_boost * (this_frame_mv_in_out / 2.0);
+  /* Increase boost for frames where new data coming into frame
+   * (eg zoom out). Slightly reduce boost if there is a net balance
+   * of motion out of the frame (zoom in).
+   * The range for this_frame_mv_in_out is -1.0 to +1.0
+   */
+  if (this_frame_mv_in_out > 0.0)
+    frame_boost += frame_boost * (this_frame_mv_in_out * 2.0);
+  /* In extreme case boost is halved */
+  else
+    frame_boost += frame_boost * (this_frame_mv_in_out / 2.0);
 
-    /* Clip to maximum */
-    if (frame_boost > GF_RMAX)
-        frame_boost = GF_RMAX;
+  /* Clip to maximum */
+  if (frame_boost > GF_RMAX) frame_boost = GF_RMAX;
 
-    return frame_boost;
+  return frame_boost;
 }
 
 #if NEW_BOOST
-static int calc_arf_boost(
-    VP8_COMP *cpi,
-    int offset,
-    int f_frames,
-    int b_frames,
-    int *f_boost,
-    int *b_boost )
-{
-    FIRSTPASS_STATS this_frame;
+static int calc_arf_boost(VP8_COMP *cpi, int offset, int f_frames, int b_frames,
+                          int *f_boost, int *b_boost) {
+  FIRSTPASS_STATS this_frame;
 
-    int i;
-    double boost_score = 0.0;
-    double mv_ratio_accumulator = 0.0;
-    double decay_accumulator = 1.0;
-    double this_frame_mv_in_out = 0.0;
-    double mv_in_out_accumulator = 0.0;
-    double abs_mv_in_out_accumulator = 0.0;
-    double r;
-    int flash_detected = 0;
+  int i;
+  double boost_score = 0.0;
+  double mv_ratio_accumulator = 0.0;
+  double decay_accumulator = 1.0;
+  double this_frame_mv_in_out = 0.0;
+  double mv_in_out_accumulator = 0.0;
+  double abs_mv_in_out_accumulator = 0.0;
+  double r;
+  int flash_detected = 0;
 
-    /* Search forward from the proposed arf/next gf position */
-    for ( i = 0; i < f_frames; i++ )
-    {
-        if ( read_frame_stats(cpi, &this_frame, (i+offset)) == EOF )
-            break;
+  /* Search forward from the proposed arf/next gf position */
+  for (i = 0; i < f_frames; i++) {
+    if (read_frame_stats(cpi, &this_frame, (i + offset)) == EOF) break;
 
-        /* Update the motion related elements to the boost calculation */
-        accumulate_frame_motion_stats( cpi, &this_frame,
-            &this_frame_mv_in_out, &mv_in_out_accumulator,
-            &abs_mv_in_out_accumulator, &mv_ratio_accumulator );
+    /* Update the motion related elements to the boost calculation */
+    accumulate_frame_motion_stats(
+        cpi, &this_frame, &this_frame_mv_in_out, &mv_in_out_accumulator,
+        &abs_mv_in_out_accumulator, &mv_ratio_accumulator);
 
-        /* Calculate the baseline boost number for this frame */
-        r = calc_frame_boost( cpi, &this_frame, this_frame_mv_in_out );
+    /* Calculate the baseline boost number for this frame */
+    r = calc_frame_boost(cpi, &this_frame, this_frame_mv_in_out);
 
-        /* We want to discount the the flash frame itself and the recovery
-         * frame that follows as both will have poor scores.
-         */
-        flash_detected = detect_flash(cpi, (i+offset)) ||
-                         detect_flash(cpi, (i+offset+1));
+    /* We want to discount the the flash frame itself and the recovery
+     * frame that follows as both will have poor scores.
+     */
+    flash_detected =
+        detect_flash(cpi, (i + offset)) || detect_flash(cpi, (i + offset + 1));
 
-        /* Cumulative effect of prediction quality decay */
-        if ( !flash_detected )
-        {
-            decay_accumulator =
-                decay_accumulator *
-                get_prediction_decay_rate(cpi, &this_frame);
-            decay_accumulator =
-                decay_accumulator < 0.1 ? 0.1 : decay_accumulator;
-        }
-        boost_score += (decay_accumulator * r);
+    /* Cumulative effect of prediction quality decay */
+    if (!flash_detected) {
+      decay_accumulator =
+          decay_accumulator * get_prediction_decay_rate(cpi, &this_frame);
+      decay_accumulator = decay_accumulator < 0.1 ? 0.1 : decay_accumulator;
+    }
+    boost_score += (decay_accumulator * r);
 
-        /* Break out conditions. */
-        if  ( (!flash_detected) &&
-              ((mv_ratio_accumulator > 100.0) ||
-               (abs_mv_in_out_accumulator > 3.0) ||
-               (mv_in_out_accumulator < -2.0) ) )
-        {
-            break;
-        }
+    /* Break out conditions. */
+    if ((!flash_detected) &&
+        ((mv_ratio_accumulator > 100.0) || (abs_mv_in_out_accumulator > 3.0) ||
+         (mv_in_out_accumulator < -2.0))) {
+      break;
     }
+  }
 
-    *f_boost = (int)(boost_score * 100.0) >> 4;
+  *f_boost = (int)(boost_score * 100.0) >> 4;
 
-    /* Reset for backward looking loop */
-    boost_score = 0.0;
-    mv_ratio_accumulator = 0.0;
-    decay_accumulator = 1.0;
-    this_frame_mv_in_out = 0.0;
-    mv_in_out_accumulator = 0.0;
-    abs_mv_in_out_accumulator = 0.0;
+  /* Reset for backward looking loop */
+  boost_score = 0.0;
+  mv_ratio_accumulator = 0.0;
+  decay_accumulator = 1.0;
+  this_frame_mv_in_out = 0.0;
+  mv_in_out_accumulator = 0.0;
+  abs_mv_in_out_accumulator = 0.0;
 
-    /* Search forward from the proposed arf/next gf position */
-    for ( i = -1; i >= -b_frames; i-- )
-    {
-        if ( read_frame_stats(cpi, &this_frame, (i+offset)) == EOF )
-            break;
+  /* Search forward from the proposed arf/next gf position */
+  for (i = -1; i >= -b_frames; i--) {
+    if (read_frame_stats(cpi, &this_frame, (i + offset)) == EOF) break;
 
-        /* Update the motion related elements to the boost calculation */
-        accumulate_frame_motion_stats( cpi, &this_frame,
-            &this_frame_mv_in_out, &mv_in_out_accumulator,
-            &abs_mv_in_out_accumulator, &mv_ratio_accumulator );
+    /* Update the motion related elements to the boost calculation */
+    accumulate_frame_motion_stats(
+        cpi, &this_frame, &this_frame_mv_in_out, &mv_in_out_accumulator,
+        &abs_mv_in_out_accumulator, &mv_ratio_accumulator);
 
-        /* Calculate the baseline boost number for this frame */
-        r = calc_frame_boost( cpi, &this_frame, this_frame_mv_in_out );
+    /* Calculate the baseline boost number for this frame */
+    r = calc_frame_boost(cpi, &this_frame, this_frame_mv_in_out);
 
-        /* We want to discount the the flash frame itself and the recovery
-         * frame that follows as both will have poor scores.
-         */
-        flash_detected = detect_flash(cpi, (i+offset)) ||
-                         detect_flash(cpi, (i+offset+1));
+    /* We want to discount the the flash frame itself and the recovery
+     * frame that follows as both will have poor scores.
+     */
+    flash_detected =
+        detect_flash(cpi, (i + offset)) || detect_flash(cpi, (i + offset + 1));
 
-        /* Cumulative effect of prediction quality decay */
-        if ( !flash_detected )
-        {
-            decay_accumulator =
-                decay_accumulator *
-                get_prediction_decay_rate(cpi, &this_frame);
-            decay_accumulator =
-                decay_accumulator < 0.1 ? 0.1 : decay_accumulator;
-        }
+    /* Cumulative effect of prediction quality decay */
+    if (!flash_detected) {
+      decay_accumulator =
+          decay_accumulator * get_prediction_decay_rate(cpi, &this_frame);
+      decay_accumulator = decay_accumulator < 0.1 ? 0.1 : decay_accumulator;
+    }
 
-        boost_score += (decay_accumulator * r);
+    boost_score += (decay_accumulator * r);
 
-        /* Break out conditions. */
-        if  ( (!flash_detected) &&
-              ((mv_ratio_accumulator > 100.0) ||
-               (abs_mv_in_out_accumulator > 3.0) ||
-               (mv_in_out_accumulator < -2.0) ) )
-        {
-            break;
-        }
+    /* Break out conditions. */
+    if ((!flash_detected) &&
+        ((mv_ratio_accumulator > 100.0) || (abs_mv_in_out_accumulator > 3.0) ||
+         (mv_in_out_accumulator < -2.0))) {
+      break;
     }
-    *b_boost = (int)(boost_score * 100.0) >> 4;
+  }
+  *b_boost = (int)(boost_score * 100.0) >> 4;
 
-    return (*f_boost + *b_boost);
+  return (*f_boost + *b_boost);
 }
 #endif
 
 /* Analyse and define a gf/arf group . */
-static void define_gf_group(VP8_COMP *cpi, FIRSTPASS_STATS *this_frame)
-{
-    FIRSTPASS_STATS next_frame;
-    FIRSTPASS_STATS *start_pos;
-    int i;
-    double r;
-    double boost_score = 0.0;
-    double old_boost_score = 0.0;
-    double gf_group_err = 0.0;
-    double gf_first_frame_err = 0.0;
-    double mod_frame_err = 0.0;
+static void define_gf_group(VP8_COMP *cpi, FIRSTPASS_STATS *this_frame) {
+  FIRSTPASS_STATS next_frame;
+  FIRSTPASS_STATS *start_pos;
+  int i;
+  double r;
+  double boost_score = 0.0;
+  double old_boost_score = 0.0;
+  double gf_group_err = 0.0;
+  double gf_first_frame_err = 0.0;
+  double mod_frame_err = 0.0;
 
-    double mv_ratio_accumulator = 0.0;
-    double decay_accumulator = 1.0;
+  double mv_ratio_accumulator = 0.0;
+  double decay_accumulator = 1.0;
 
-    double loop_decay_rate = 1.00;          /* Starting decay rate */
+  double loop_decay_rate = 1.00; /* Starting decay rate */
 
-    double this_frame_mv_in_out = 0.0;
-    double mv_in_out_accumulator = 0.0;
-    double abs_mv_in_out_accumulator = 0.0;
-    double mod_err_per_mb_accumulator = 0.0;
+  double this_frame_mv_in_out = 0.0;
+  double mv_in_out_accumulator = 0.0;
+  double abs_mv_in_out_accumulator = 0.0;
+  double mod_err_per_mb_accumulator = 0.0;
 
-    int max_bits = frame_max_bits(cpi);     /* Max for a single frame */
+  int max_bits = frame_max_bits(cpi); /* Max for a single frame */
 
-    unsigned int allow_alt_ref =
-                    cpi->oxcf.play_alternate && cpi->oxcf.lag_in_frames;
+  unsigned int allow_alt_ref =
+      cpi->oxcf.play_alternate && cpi->oxcf.lag_in_frames;
 
-    int alt_boost = 0;
-    int f_boost = 0;
-    int b_boost = 0;
-    int flash_detected;
+  int alt_boost = 0;
+  int f_boost = 0;
+  int b_boost = 0;
+  int flash_detected;
 
-    cpi->twopass.gf_group_bits = 0;
-    cpi->twopass.gf_decay_rate = 0;
+  cpi->twopass.gf_group_bits = 0;
+  cpi->twopass.gf_decay_rate = 0;
 
-    vp8_clear_system_state();
+  vp8_clear_system_state();
 
-    start_pos = cpi->twopass.stats_in;
+  start_pos = cpi->twopass.stats_in;
 
-    memset(&next_frame, 0, sizeof(next_frame)); /* assure clean */
+  memset(&next_frame, 0, sizeof(next_frame)); /* assure clean */
 
-    /* Load stats for the current frame. */
-    mod_frame_err = calculate_modified_err(cpi, this_frame);
+  /* Load stats for the current frame. */
+  mod_frame_err = calculate_modified_err(cpi, this_frame);
 
-    /* Note the error of the frame at the start of the group (this will be
-     * the GF frame error if we code a normal gf
-     */
-    gf_first_frame_err = mod_frame_err;
+  /* Note the error of the frame at the start of the group (this will be
+   * the GF frame error if we code a normal gf
+   */
+  gf_first_frame_err = mod_frame_err;
 
-    /* Special treatment if the current frame is a key frame (which is also
-     * a gf). If it is then its error score (and hence bit allocation) need
-     * to be subtracted out from the calculation for the GF group
-     */
-    if (cpi->common.frame_type == KEY_FRAME)
-        gf_group_err -= gf_first_frame_err;
+  /* Special treatment if the current frame is a key frame (which is also
+   * a gf). If it is then its error score (and hence bit allocation) need
+   * to be subtracted out from the calculation for the GF group
+   */
+  if (cpi->common.frame_type == KEY_FRAME) gf_group_err -= gf_first_frame_err;
 
-    /* Scan forward to try and work out how many frames the next gf group
-     * should contain and what level of boost is appropriate for the GF
-     * or ARF that will be coded with the group
-     */
-    i = 0;
+  /* Scan forward to try and work out how many frames the next gf group
+   * should contain and what level of boost is appropriate for the GF
+   * or ARF that will be coded with the group
+   */
+  i = 0;
 
-    while (((i < cpi->twopass.static_scene_max_gf_interval) ||
-            ((cpi->twopass.frames_to_key - i) < MIN_GF_INTERVAL)) &&
-           (i < cpi->twopass.frames_to_key))
-    {
-        i++;
+  while (((i < cpi->twopass.static_scene_max_gf_interval) ||
+          ((cpi->twopass.frames_to_key - i) < MIN_GF_INTERVAL)) &&
+         (i < cpi->twopass.frames_to_key)) {
+    i++;
 
-        /* Accumulate error score of frames in this gf group */
-        mod_frame_err = calculate_modified_err(cpi, this_frame);
+    /* Accumulate error score of frames in this gf group */
+    mod_frame_err = calculate_modified_err(cpi, this_frame);
 
-        gf_group_err += mod_frame_err;
+    gf_group_err += mod_frame_err;
 
-        mod_err_per_mb_accumulator +=
-            mod_frame_err / DOUBLE_DIVIDE_CHECK((double)cpi->common.MBs);
+    mod_err_per_mb_accumulator +=
+        mod_frame_err / DOUBLE_DIVIDE_CHECK((double)cpi->common.MBs);
 
-        if (EOF == input_stats(cpi, &next_frame))
-            break;
+    if (EOF == input_stats(cpi, &next_frame)) break;
 
-        /* Test for the case where there is a brief flash but the prediction
-         * quality back to an earlier frame is then restored.
-         */
-        flash_detected = detect_flash(cpi, 0);
+    /* Test for the case where there is a brief flash but the prediction
+     * quality back to an earlier frame is then restored.
+     */
+    flash_detected = detect_flash(cpi, 0);
 
-        /* Update the motion related elements to the boost calculation */
-        accumulate_frame_motion_stats( cpi, &next_frame,
-            &this_frame_mv_in_out, &mv_in_out_accumulator,
-            &abs_mv_in_out_accumulator, &mv_ratio_accumulator );
+    /* Update the motion related elements to the boost calculation */
+    accumulate_frame_motion_stats(
+        cpi, &next_frame, &this_frame_mv_in_out, &mv_in_out_accumulator,
+        &abs_mv_in_out_accumulator, &mv_ratio_accumulator);
 
-        /* Calculate a baseline boost number for this frame */
-        r = calc_frame_boost( cpi, &next_frame, this_frame_mv_in_out );
+    /* Calculate a baseline boost number for this frame */
+    r = calc_frame_boost(cpi, &next_frame, this_frame_mv_in_out);
 
-        /* Cumulative effect of prediction quality decay */
-        if ( !flash_detected )
-        {
-            loop_decay_rate = get_prediction_decay_rate(cpi, &next_frame);
-            decay_accumulator = decay_accumulator * loop_decay_rate;
-            decay_accumulator =
-                decay_accumulator < 0.1 ? 0.1 : decay_accumulator;
-        }
-        boost_score += (decay_accumulator * r);
+    /* Cumulative effect of prediction quality decay */
+    if (!flash_detected) {
+      loop_decay_rate = get_prediction_decay_rate(cpi, &next_frame);
+      decay_accumulator = decay_accumulator * loop_decay_rate;
+      decay_accumulator = decay_accumulator < 0.1 ? 0.1 : decay_accumulator;
+    }
+    boost_score += (decay_accumulator * r);
 
-        /* Break clause to detect very still sections after motion
-         * For example a staic image after a fade or other transition.
-         */
-        if ( detect_transition_to_still( cpi, i, 5,
-                                         loop_decay_rate,
-                                         decay_accumulator ) )
-        {
-            allow_alt_ref = 0;
-            boost_score = old_boost_score;
-            break;
-        }
+    /* Break clause to detect very still sections after motion
+     * For example a staic image after a fade or other transition.
+     */
+    if (detect_transition_to_still(cpi, i, 5, loop_decay_rate,
+                                   decay_accumulator)) {
+      allow_alt_ref = 0;
+      boost_score = old_boost_score;
+      break;
+    }
 
-        /* Break out conditions. */
-        if  (
-            /* Break at cpi->max_gf_interval unless almost totally static */
-            (i >= cpi->max_gf_interval && (decay_accumulator < 0.995)) ||
-            (
-                /* Dont break out with a very short interval */
-                (i > MIN_GF_INTERVAL) &&
-                /* Dont break out very close to a key frame */
-                ((cpi->twopass.frames_to_key - i) >= MIN_GF_INTERVAL) &&
-                ((boost_score > 20.0) || (next_frame.pcnt_inter < 0.75)) &&
-                (!flash_detected) &&
-                ((mv_ratio_accumulator > 100.0) ||
-                 (abs_mv_in_out_accumulator > 3.0) ||
-                 (mv_in_out_accumulator < -2.0) ||
-                 ((boost_score - old_boost_score) < 2.0))
-            ) )
-        {
-            boost_score = old_boost_score;
-            break;
-        }
-
-        memcpy(this_frame, &next_frame, sizeof(*this_frame));
-
-        old_boost_score = boost_score;
+    /* Break out conditions. */
+    if (
+        /* Break at cpi->max_gf_interval unless almost totally static */
+        (i >= cpi->max_gf_interval && (decay_accumulator < 0.995)) ||
+        (
+            /* Dont break out with a very short interval */
+            (i > MIN_GF_INTERVAL) &&
+            /* Dont break out very close to a key frame */
+            ((cpi->twopass.frames_to_key - i) >= MIN_GF_INTERVAL) &&
+            ((boost_score > 20.0) || (next_frame.pcnt_inter < 0.75)) &&
+            (!flash_detected) && ((mv_ratio_accumulator > 100.0) ||
+                                  (abs_mv_in_out_accumulator > 3.0) ||
+                                  (mv_in_out_accumulator < -2.0) ||
+                                  ((boost_score - old_boost_score) < 2.0)))) {
+      boost_score = old_boost_score;
+      break;
     }
 
-    cpi->twopass.gf_decay_rate =
-        (i > 0) ? (int)(100.0 * (1.0 - decay_accumulator)) / i : 0;
+    memcpy(this_frame, &next_frame, sizeof(*this_frame));
 
-    /* When using CBR apply additional buffer related upper limits */
-    if (cpi->oxcf.end_usage == USAGE_STREAM_FROM_SERVER)
-    {
-        double max_boost;
+    old_boost_score = boost_score;
+  }
 
-        /* For cbr apply buffer related limits */
-        if (cpi->drop_frames_allowed)
-        {
-            int64_t df_buffer_level = cpi->oxcf.drop_frames_water_mark *
-                                  (cpi->oxcf.optimal_buffer_level / 100);
+  cpi->twopass.gf_decay_rate =
+      (i > 0) ? (int)(100.0 * (1.0 - decay_accumulator)) / i : 0;
 
-            if (cpi->buffer_level > df_buffer_level)
-                max_boost = ((double)((cpi->buffer_level - df_buffer_level) * 2 / 3) * 16.0) / DOUBLE_DIVIDE_CHECK((double)cpi->av_per_frame_bandwidth);
-            else
-                max_boost = 0.0;
-        }
-        else if (cpi->buffer_level > 0)
-        {
-            max_boost = ((double)(cpi->buffer_level * 2 / 3) * 16.0) / DOUBLE_DIVIDE_CHECK((double)cpi->av_per_frame_bandwidth);
-        }
-        else
-        {
-            max_boost = 0.0;
-        }
+  /* When using CBR apply additional buffer related upper limits */
+  if (cpi->oxcf.end_usage == USAGE_STREAM_FROM_SERVER) {
+    double max_boost;
 
-        if (boost_score > max_boost)
-            boost_score = max_boost;
+    /* For cbr apply buffer related limits */
+    if (cpi->drop_frames_allowed) {
+      int64_t df_buffer_level = cpi->oxcf.drop_frames_water_mark *
+                                (cpi->oxcf.optimal_buffer_level / 100);
+
+      if (cpi->buffer_level > df_buffer_level)
+        max_boost =
+            ((double)((cpi->buffer_level - df_buffer_level) * 2 / 3) * 16.0) /
+            DOUBLE_DIVIDE_CHECK((double)cpi->av_per_frame_bandwidth);
+      else
+        max_boost = 0.0;
+    } else if (cpi->buffer_level > 0) {
+      max_boost = ((double)(cpi->buffer_level * 2 / 3) * 16.0) /
+                  DOUBLE_DIVIDE_CHECK((double)cpi->av_per_frame_bandwidth);
+    } else {
+      max_boost = 0.0;
     }
 
-    /* Dont allow conventional gf too near the next kf */
-    if ((cpi->twopass.frames_to_key - i) < MIN_GF_INTERVAL)
-    {
-        while (i < cpi->twopass.frames_to_key)
-        {
-            i++;
+    if (boost_score > max_boost) boost_score = max_boost;
+  }
 
-            if (EOF == input_stats(cpi, this_frame))
-                break;
+  /* Dont allow conventional gf too near the next kf */
+  if ((cpi->twopass.frames_to_key - i) < MIN_GF_INTERVAL) {
+    while (i < cpi->twopass.frames_to_key) {
+      i++;
 
-            if (i < cpi->twopass.frames_to_key)
-            {
-                mod_frame_err = calculate_modified_err(cpi, this_frame);
-                gf_group_err += mod_frame_err;
-            }
-        }
+      if (EOF == input_stats(cpi, this_frame)) break;
+
+      if (i < cpi->twopass.frames_to_key) {
+        mod_frame_err = calculate_modified_err(cpi, this_frame);
+        gf_group_err += mod_frame_err;
+      }
     }
+  }
 
-    cpi->gfu_boost = (int)(boost_score * 100.0) >> 4;
+  cpi->gfu_boost = (int)(boost_score * 100.0) >> 4;
 
 #if NEW_BOOST
-    /* Alterrnative boost calculation for alt ref */
-    alt_boost = calc_arf_boost( cpi, 0, (i-1), (i-1), &f_boost, &b_boost );
+  /* Alterrnative boost calculation for alt ref */
+  alt_boost = calc_arf_boost(cpi, 0, (i - 1), (i - 1), &f_boost, &b_boost);
 #endif
 
-    /* Should we use the alternate refernce frame */
-    if (allow_alt_ref &&
-        (i >= MIN_GF_INTERVAL) &&
-        /* dont use ARF very near next kf */
-        (i <= (cpi->twopass.frames_to_key - MIN_GF_INTERVAL)) &&
+  /* Should we use the alternate refernce frame */
+  if (allow_alt_ref && (i >= MIN_GF_INTERVAL) &&
+      /* dont use ARF very near next kf */
+      (i <= (cpi->twopass.frames_to_key - MIN_GF_INTERVAL)) &&
 #if NEW_BOOST
-        ((next_frame.pcnt_inter > 0.75) ||
-         (next_frame.pcnt_second_ref > 0.5)) &&
-        ((mv_in_out_accumulator / (double)i > -0.2) ||
-         (mv_in_out_accumulator > -2.0)) &&
-        (b_boost > 100) &&
-        (f_boost > 100) )
+      ((next_frame.pcnt_inter > 0.75) || (next_frame.pcnt_second_ref > 0.5)) &&
+      ((mv_in_out_accumulator / (double)i > -0.2) ||
+       (mv_in_out_accumulator > -2.0)) &&
+      (b_boost > 100) && (f_boost > 100))
 #else
-        (next_frame.pcnt_inter > 0.75) &&
-        ((mv_in_out_accumulator / (double)i > -0.2) ||
-         (mv_in_out_accumulator > -2.0)) &&
-        (cpi->gfu_boost > 100) &&
-        (cpi->twopass.gf_decay_rate <=
-            (ARF_DECAY_THRESH + (cpi->gfu_boost / 200))) )
+      (next_frame.pcnt_inter > 0.75) &&
+      ((mv_in_out_accumulator / (double)i > -0.2) ||
+       (mv_in_out_accumulator > -2.0)) &&
+      (cpi->gfu_boost > 100) && (cpi->twopass.gf_decay_rate <=
+                                 (ARF_DECAY_THRESH + (cpi->gfu_boost / 200))))
 #endif
-    {
-        int Boost;
-        int allocation_chunks;
-        int Q = (cpi->oxcf.fixed_q < 0)
-                ? cpi->last_q[INTER_FRAME] : cpi->oxcf.fixed_q;
-        int tmp_q;
-        int arf_frame_bits = 0;
-        int group_bits;
+  {
+    int Boost;
+    int allocation_chunks;
+    int Q =
+        (cpi->oxcf.fixed_q < 0) ? cpi->last_q[INTER_FRAME] : cpi->oxcf.fixed_q;
+    int tmp_q;
+    int arf_frame_bits = 0;
+    int group_bits;
 
 #if NEW_BOOST
-        cpi->gfu_boost = alt_boost;
+    cpi->gfu_boost = alt_boost;
 #endif
 
-        /* Estimate the bits to be allocated to the group as a whole */
-        if ((cpi->twopass.kf_group_bits > 0) &&
-            (cpi->twopass.kf_group_error_left > 0))
-        {
-            group_bits = (int)((double)cpi->twopass.kf_group_bits *
+    /* Estimate the bits to be allocated to the group as a whole */
+    if ((cpi->twopass.kf_group_bits > 0) &&
+        (cpi->twopass.kf_group_error_left > 0)) {
+      group_bits =
+          (int)((double)cpi->twopass.kf_group_bits *
                 (gf_group_err / (double)cpi->twopass.kf_group_error_left));
-        }
-        else
-            group_bits = 0;
+    } else
+      group_bits = 0;
 
-        /* Boost for arf frame */
+/* Boost for arf frame */
 #if NEW_BOOST
-        Boost = (alt_boost * GFQ_ADJUSTMENT) / 100;
+    Boost = (alt_boost * GFQ_ADJUSTMENT) / 100;
 #else
-        Boost = (cpi->gfu_boost * 3 * GFQ_ADJUSTMENT) / (2 * 100);
+    Boost = (cpi->gfu_boost * 3 * GFQ_ADJUSTMENT) / (2 * 100);
 #endif
-        Boost += (i * 50);
+    Boost += (i * 50);
 
-        /* Set max and minimum boost and hence minimum allocation */
-        if (Boost > ((cpi->baseline_gf_interval + 1) * 200))
-            Boost = ((cpi->baseline_gf_interval + 1) * 200);
-        else if (Boost < 125)
-            Boost = 125;
+    /* Set max and minimum boost and hence minimum allocation */
+    if (Boost > ((cpi->baseline_gf_interval + 1) * 200))
+      Boost = ((cpi->baseline_gf_interval + 1) * 200);
+    else if (Boost < 125)
+      Boost = 125;
 
-        allocation_chunks = (i * 100) + Boost;
+    allocation_chunks = (i * 100) + Boost;
 
-        /* Normalize Altboost and allocations chunck down to prevent overflow */
-        while (Boost > 1000)
-        {
-            Boost /= 2;
-            allocation_chunks /= 2;
-        }
+    /* Normalize Altboost and allocations chunck down to prevent overflow */
+    while (Boost > 1000) {
+      Boost /= 2;
+      allocation_chunks /= 2;
+    }
 
-        /* Calculate the number of bits to be spent on the arf based on the
-         * boost number
-         */
-        arf_frame_bits = (int)((double)Boost * (group_bits /
-                               (double)allocation_chunks));
+    /* Calculate the number of bits to be spent on the arf based on the
+     * boost number
+     */
+    arf_frame_bits =
+        (int)((double)Boost * (group_bits / (double)allocation_chunks));
 
-        /* Estimate if there are enough bits available to make worthwhile use
-         * of an arf.
-         */
-        tmp_q = estimate_q(cpi, mod_frame_err, (int)arf_frame_bits);
+    /* Estimate if there are enough bits available to make worthwhile use
+     * of an arf.
+     */
+    tmp_q = estimate_q(cpi, mod_frame_err, (int)arf_frame_bits);
 
-        /* Only use an arf if it is likely we will be able to code
-         * it at a lower Q than the surrounding frames.
-         */
-        if (tmp_q < cpi->worst_quality)
-        {
-            int half_gf_int;
-            int frames_after_arf;
-            int frames_bwd = cpi->oxcf.arnr_max_frames - 1;
-            int frames_fwd = cpi->oxcf.arnr_max_frames - 1;
+    /* Only use an arf if it is likely we will be able to code
+     * it at a lower Q than the surrounding frames.
+     */
+    if (tmp_q < cpi->worst_quality) {
+      int half_gf_int;
+      int frames_after_arf;
+      int frames_bwd = cpi->oxcf.arnr_max_frames - 1;
+      int frames_fwd = cpi->oxcf.arnr_max_frames - 1;
 
-            cpi->source_alt_ref_pending = 1;
+      cpi->source_alt_ref_pending = 1;
 
-            /*
-             * For alt ref frames the error score for the end frame of the
-             * group (the alt ref frame) should not contribute to the group
-             * total and hence the number of bit allocated to the group.
-             * Rather it forms part of the next group (it is the GF at the
-             * start of the next group)
-             * gf_group_err -= mod_frame_err;
-             *
-             * For alt ref frames alt ref frame is technically part of the
-             * GF frame for the next group but we always base the error
-             * calculation and bit allocation on the current group of frames.
-             *
-             * Set the interval till the next gf or arf.
-             * For ARFs this is the number of frames to be coded before the
-             * future frame that is coded as an ARF.
-             * The future frame itself is part of the next group
-             */
-            cpi->baseline_gf_interval = i;
+      /*
+       * For alt ref frames the error score for the end frame of the
+       * group (the alt ref frame) should not contribute to the group
+       * total and hence the number of bit allocated to the group.
+       * Rather it forms part of the next group (it is the GF at the
+       * start of the next group)
+       * gf_group_err -= mod_frame_err;
+       *
+       * For alt ref frames alt ref frame is technically part of the
+       * GF frame for the next group but we always base the error
+       * calculation and bit allocation on the current group of frames.
+       *
+       * Set the interval till the next gf or arf.
+       * For ARFs this is the number of frames to be coded before the
+       * future frame that is coded as an ARF.
+       * The future frame itself is part of the next group
+       */
+      cpi->baseline_gf_interval = i;
 
-            /*
-             * Define the arnr filter width for this group of frames:
-             * We only filter frames that lie within a distance of half
-             * the GF interval from the ARF frame. We also have to trap
-             * cases where the filter extends beyond the end of clip.
-             * Note: this_frame->frame has been updated in the loop
-             * so it now points at the ARF frame.
-             */
-            half_gf_int = cpi->baseline_gf_interval >> 1;
-            frames_after_arf = (int)(cpi->twopass.total_stats.count -
-                               this_frame->frame - 1);
+      /*
+       * Define the arnr filter width for this group of frames:
+       * We only filter frames that lie within a distance of half
+       * the GF interval from the ARF frame. We also have to trap
+       * cases where the filter extends beyond the end of clip.
+       * Note: this_frame->frame has been updated in the loop
+       * so it now points at the ARF frame.
+       */
+      half_gf_int = cpi->baseline_gf_interval >> 1;
+      frames_after_arf =
+          (int)(cpi->twopass.total_stats.count - this_frame->frame - 1);
 
-            switch (cpi->oxcf.arnr_type)
-            {
-            case 1: /* Backward filter */
-                frames_fwd = 0;
-                if (frames_bwd > half_gf_int)
-                    frames_bwd = half_gf_int;
-                break;
+      switch (cpi->oxcf.arnr_type) {
+        case 1: /* Backward filter */
+          frames_fwd = 0;
+          if (frames_bwd > half_gf_int) frames_bwd = half_gf_int;
+          break;
 
-            case 2: /* Forward filter */
-                if (frames_fwd > half_gf_int)
-                    frames_fwd = half_gf_int;
-                if (frames_fwd > frames_after_arf)
-                    frames_fwd = frames_after_arf;
-                frames_bwd = 0;
-                break;
+        case 2: /* Forward filter */
+          if (frames_fwd > half_gf_int) frames_fwd = half_gf_int;
+          if (frames_fwd > frames_after_arf) frames_fwd = frames_after_arf;
+          frames_bwd = 0;
+          break;
 
-            case 3: /* Centered filter */
-            default:
-                frames_fwd >>= 1;
-                if (frames_fwd > frames_after_arf)
-                    frames_fwd = frames_after_arf;
-                if (frames_fwd > half_gf_int)
-                    frames_fwd = half_gf_int;
+        case 3: /* Centered filter */
+        default:
+          frames_fwd >>= 1;
+          if (frames_fwd > frames_after_arf) frames_fwd = frames_after_arf;
+          if (frames_fwd > half_gf_int) frames_fwd = half_gf_int;
 
-                frames_bwd = frames_fwd;
+          frames_bwd = frames_fwd;
 
-                /* For even length filter there is one more frame backward
-                 * than forward: e.g. len=6 ==> bbbAff, len=7 ==> bbbAfff.
-                 */
-                if (frames_bwd < half_gf_int)
-                    frames_bwd += (cpi->oxcf.arnr_max_frames+1) & 0x1;
-                break;
-            }
+          /* For even length filter there is one more frame backward
+           * than forward: e.g. len=6 ==> bbbAff, len=7 ==> bbbAfff.
+           */
+          if (frames_bwd < half_gf_int)
+            frames_bwd += (cpi->oxcf.arnr_max_frames + 1) & 0x1;
+          break;
+      }
 
-            cpi->active_arnr_frames = frames_bwd + 1 + frames_fwd;
-        }
-        else
-        {
-            cpi->source_alt_ref_pending = 0;
-            cpi->baseline_gf_interval = i;
-        }
+      cpi->active_arnr_frames = frames_bwd + 1 + frames_fwd;
+    } else {
+      cpi->source_alt_ref_pending = 0;
+      cpi->baseline_gf_interval = i;
     }
-    else
-    {
-        cpi->source_alt_ref_pending = 0;
-        cpi->baseline_gf_interval = i;
-    }
+  } else {
+    cpi->source_alt_ref_pending = 0;
+    cpi->baseline_gf_interval = i;
+  }
 
-    /*
-     * Now decide how many bits should be allocated to the GF group as  a
-     * proportion of those remaining in the kf group.
-     * The final key frame group in the clip is treated as a special case
-     * where cpi->twopass.kf_group_bits is tied to cpi->twopass.bits_left.
-     * This is also important for short clips where there may only be one
-     * key frame.
-     */
-    if (cpi->twopass.frames_to_key >= (int)(cpi->twopass.total_stats.count -
-                                            cpi->common.current_video_frame))
-    {
-        cpi->twopass.kf_group_bits =
-            (cpi->twopass.bits_left > 0) ? cpi->twopass.bits_left : 0;
-    }
+  /*
+   * Now decide how many bits should be allocated to the GF group as  a
+   * proportion of those remaining in the kf group.
+   * The final key frame group in the clip is treated as a special case
+   * where cpi->twopass.kf_group_bits is tied to cpi->twopass.bits_left.
+   * This is also important for short clips where there may only be one
+   * key frame.
+   */
+  if (cpi->twopass.frames_to_key >=
+      (int)(cpi->twopass.total_stats.count - cpi->common.current_video_frame)) {
+    cpi->twopass.kf_group_bits =
+        (cpi->twopass.bits_left > 0) ? cpi->twopass.bits_left : 0;
+  }
 
-    /* Calculate the bits to be allocated to the group as a whole */
-    if ((cpi->twopass.kf_group_bits > 0) &&
-        (cpi->twopass.kf_group_error_left > 0))
-    {
-        cpi->twopass.gf_group_bits =
-            (int64_t)(cpi->twopass.kf_group_bits *
-                      (gf_group_err / cpi->twopass.kf_group_error_left));
-    }
-    else
-        cpi->twopass.gf_group_bits = 0;
-
+  /* Calculate the bits to be allocated to the group as a whole */
+  if ((cpi->twopass.kf_group_bits > 0) &&
+      (cpi->twopass.kf_group_error_left > 0)) {
     cpi->twopass.gf_group_bits =
-        (cpi->twopass.gf_group_bits < 0)
-            ? 0
-            : (cpi->twopass.gf_group_bits > cpi->twopass.kf_group_bits)
-                ? cpi->twopass.kf_group_bits : cpi->twopass.gf_group_bits;
+        (int64_t)(cpi->twopass.kf_group_bits *
+                  (gf_group_err / cpi->twopass.kf_group_error_left));
+  } else
+    cpi->twopass.gf_group_bits = 0;
 
-    /* Clip cpi->twopass.gf_group_bits based on user supplied data rate
-     * variability limit (cpi->oxcf.two_pass_vbrmax_section)
-     */
-    if (cpi->twopass.gf_group_bits >
-        (int64_t)max_bits * cpi->baseline_gf_interval)
-        cpi->twopass.gf_group_bits =
-            (int64_t)max_bits * cpi->baseline_gf_interval;
+  cpi->twopass.gf_group_bits =
+      (cpi->twopass.gf_group_bits < 0)
+          ? 0
+          : (cpi->twopass.gf_group_bits > cpi->twopass.kf_group_bits)
+                ? cpi->twopass.kf_group_bits
+                : cpi->twopass.gf_group_bits;
 
-    /* Reset the file position */
-    reset_fpf_position(cpi, start_pos);
+  /* Clip cpi->twopass.gf_group_bits based on user supplied data rate
+   * variability limit (cpi->oxcf.two_pass_vbrmax_section)
+   */
+  if (cpi->twopass.gf_group_bits >
+      (int64_t)max_bits * cpi->baseline_gf_interval)
+    cpi->twopass.gf_group_bits = (int64_t)max_bits * cpi->baseline_gf_interval;
 
-    /* Update the record of error used so far (only done once per gf group) */
-    cpi->twopass.modified_error_used += gf_group_err;
+  /* Reset the file position */
+  reset_fpf_position(cpi, start_pos);
 
-    /* Assign  bits to the arf or gf. */
-    for (i = 0; i <= (cpi->source_alt_ref_pending && cpi->common.frame_type != KEY_FRAME); i++) {
-        int Boost;
-        int allocation_chunks;
-        int Q = (cpi->oxcf.fixed_q < 0) ? cpi->last_q[INTER_FRAME] : cpi->oxcf.fixed_q;
-        int gf_bits;
+  /* Update the record of error used so far (only done once per gf group) */
+  cpi->twopass.modified_error_used += gf_group_err;
 
-        /* For ARF frames */
-        if (cpi->source_alt_ref_pending && i == 0)
-        {
+  /* Assign  bits to the arf or gf. */
+  for (i = 0; i <= (cpi->source_alt_ref_pending &&
+                    cpi->common.frame_type != KEY_FRAME);
+       i++) {
+    int Boost;
+    int allocation_chunks;
+    int Q =
+        (cpi->oxcf.fixed_q < 0) ? cpi->last_q[INTER_FRAME] : cpi->oxcf.fixed_q;
+    int gf_bits;
+
+    /* For ARF frames */
+    if (cpi->source_alt_ref_pending && i == 0) {
 #if NEW_BOOST
-            Boost = (alt_boost * GFQ_ADJUSTMENT) / 100;
+      Boost = (alt_boost * GFQ_ADJUSTMENT) / 100;
 #else
-            Boost = (cpi->gfu_boost * 3 * GFQ_ADJUSTMENT) / (2 * 100);
+      Boost = (cpi->gfu_boost * 3 * GFQ_ADJUSTMENT) / (2 * 100);
 #endif
-            Boost += (cpi->baseline_gf_interval * 50);
+      Boost += (cpi->baseline_gf_interval * 50);
 
-            /* Set max and minimum boost and hence minimum allocation */
-            if (Boost > ((cpi->baseline_gf_interval + 1) * 200))
-                Boost = ((cpi->baseline_gf_interval + 1) * 200);
-            else if (Boost < 125)
-                Boost = 125;
+      /* Set max and minimum boost and hence minimum allocation */
+      if (Boost > ((cpi->baseline_gf_interval + 1) * 200))
+        Boost = ((cpi->baseline_gf_interval + 1) * 200);
+      else if (Boost < 125)
+        Boost = 125;
 
-            allocation_chunks =
-                ((cpi->baseline_gf_interval + 1) * 100) + Boost;
-        }
-        /* Else for standard golden frames */
-        else
-        {
-            /* boost based on inter / intra ratio of subsequent frames */
-            Boost = (cpi->gfu_boost * GFQ_ADJUSTMENT) / 100;
+      allocation_chunks = ((cpi->baseline_gf_interval + 1) * 100) + Boost;
+    }
+    /* Else for standard golden frames */
+    else {
+      /* boost based on inter / intra ratio of subsequent frames */
+      Boost = (cpi->gfu_boost * GFQ_ADJUSTMENT) / 100;
 
-            /* Set max and minimum boost and hence minimum allocation */
-            if (Boost > (cpi->baseline_gf_interval * 150))
-                Boost = (cpi->baseline_gf_interval * 150);
-            else if (Boost < 125)
-                Boost = 125;
+      /* Set max and minimum boost and hence minimum allocation */
+      if (Boost > (cpi->baseline_gf_interval * 150))
+        Boost = (cpi->baseline_gf_interval * 150);
+      else if (Boost < 125)
+        Boost = 125;
 
-            allocation_chunks =
-                (cpi->baseline_gf_interval * 100) + (Boost - 100);
-        }
+      allocation_chunks = (cpi->baseline_gf_interval * 100) + (Boost - 100);
+    }
 
-        /* Normalize Altboost and allocations chunck down to prevent overflow */
-        while (Boost > 1000)
-        {
-            Boost /= 2;
-            allocation_chunks /= 2;
-        }
+    /* Normalize Altboost and allocations chunck down to prevent overflow */
+    while (Boost > 1000) {
+      Boost /= 2;
+      allocation_chunks /= 2;
+    }
 
-        /* Calculate the number of bits to be spent on the gf or arf based on
-         * the boost number
-         */
-        gf_bits = (int)((double)Boost *
-                        (cpi->twopass.gf_group_bits /
-                         (double)allocation_chunks));
+    /* Calculate the number of bits to be spent on the gf or arf based on
+     * the boost number
+     */
+    gf_bits = (int)((double)Boost *
+                    (cpi->twopass.gf_group_bits / (double)allocation_chunks));
 
-        /* If the frame that is to be boosted is simpler than the average for
-         * the gf/arf group then use an alternative calculation
-         * based on the error score of the frame itself
-         */
-        if (mod_frame_err < gf_group_err / (double)cpi->baseline_gf_interval)
-        {
-            double  alt_gf_grp_bits;
-            int     alt_gf_bits;
+    /* If the frame that is to be boosted is simpler than the average for
+     * the gf/arf group then use an alternative calculation
+     * based on the error score of the frame itself
+     */
+    if (mod_frame_err < gf_group_err / (double)cpi->baseline_gf_interval) {
+      double alt_gf_grp_bits;
+      int alt_gf_bits;
 
-            alt_gf_grp_bits =
-                (double)cpi->twopass.kf_group_bits  *
-                (mod_frame_err * (double)cpi->baseline_gf_interval) /
-                DOUBLE_DIVIDE_CHECK((double)cpi->twopass.kf_group_error_left);
+      alt_gf_grp_bits =
+          (double)cpi->twopass.kf_group_bits *
+          (mod_frame_err * (double)cpi->baseline_gf_interval) /
+          DOUBLE_DIVIDE_CHECK((double)cpi->twopass.kf_group_error_left);
 
-            alt_gf_bits = (int)((double)Boost * (alt_gf_grp_bits /
-                                                 (double)allocation_chunks));
+      alt_gf_bits =
+          (int)((double)Boost * (alt_gf_grp_bits / (double)allocation_chunks));
 
-            if (gf_bits > alt_gf_bits)
-            {
-                gf_bits = alt_gf_bits;
-            }
-        }
-        /* Else if it is harder than other frames in the group make sure it at
-         * least receives an allocation in keeping with its relative error
-         * score, otherwise it may be worse off than an "un-boosted" frame
-         */
-        else
-        {
-            int alt_gf_bits =
-                (int)((double)cpi->twopass.kf_group_bits *
-                      mod_frame_err /
-                      DOUBLE_DIVIDE_CHECK((double)cpi->twopass.kf_group_error_left));
+      if (gf_bits > alt_gf_bits) {
+        gf_bits = alt_gf_bits;
+      }
+    }
+    /* Else if it is harder than other frames in the group make sure it at
+     * least receives an allocation in keeping with its relative error
+     * score, otherwise it may be worse off than an "un-boosted" frame
+     */
+    else {
+      int alt_gf_bits =
+          (int)((double)cpi->twopass.kf_group_bits * mod_frame_err /
+                DOUBLE_DIVIDE_CHECK((double)cpi->twopass.kf_group_error_left));
 
-            if (alt_gf_bits > gf_bits)
-            {
-                gf_bits = alt_gf_bits;
-            }
-        }
+      if (alt_gf_bits > gf_bits) {
+        gf_bits = alt_gf_bits;
+      }
+    }
 
-        /* Apply an additional limit for CBR */
-        if (cpi->oxcf.end_usage == USAGE_STREAM_FROM_SERVER)
-        {
-            if (cpi->twopass.gf_bits > (int)(cpi->buffer_level >> 1))
-                cpi->twopass.gf_bits = (int)(cpi->buffer_level >> 1);
-        }
+    /* Apply an additional limit for CBR */
+    if (cpi->oxcf.end_usage == USAGE_STREAM_FROM_SERVER) {
+      if (cpi->twopass.gf_bits > (int)(cpi->buffer_level >> 1))
+        cpi->twopass.gf_bits = (int)(cpi->buffer_level >> 1);
+    }
 
-        /* Dont allow a negative value for gf_bits */
-        if (gf_bits < 0)
-            gf_bits = 0;
+    /* Dont allow a negative value for gf_bits */
+    if (gf_bits < 0) gf_bits = 0;
 
-        /* Add in minimum for a frame */
-        gf_bits += cpi->min_frame_bandwidth;
+    /* Add in minimum for a frame */
+    gf_bits += cpi->min_frame_bandwidth;
 
-        if (i == 0)
-        {
-            cpi->twopass.gf_bits = gf_bits;
-        }
-        if (i == 1 || (!cpi->source_alt_ref_pending && (cpi->common.frame_type != KEY_FRAME)))
-        {
-            /* Per frame bit target for this frame */
-            cpi->per_frame_bandwidth = gf_bits;
-        }
+    if (i == 0) {
+      cpi->twopass.gf_bits = gf_bits;
     }
+    if (i == 1 || (!cpi->source_alt_ref_pending &&
+                   (cpi->common.frame_type != KEY_FRAME))) {
+      /* Per frame bit target for this frame */
+      cpi->per_frame_bandwidth = gf_bits;
+    }
+  }
 
-    {
-        /* Adjust KF group bits and error remainin */
-        cpi->twopass.kf_group_error_left -= (int64_t)gf_group_err;
-        cpi->twopass.kf_group_bits -= cpi->twopass.gf_group_bits;
+  {
+    /* Adjust KF group bits and error remainin */
+    cpi->twopass.kf_group_error_left -= (int64_t)gf_group_err;
+    cpi->twopass.kf_group_bits -= cpi->twopass.gf_group_bits;
 
-        if (cpi->twopass.kf_group_bits < 0)
-            cpi->twopass.kf_group_bits = 0;
+    if (cpi->twopass.kf_group_bits < 0) cpi->twopass.kf_group_bits = 0;
 
-        /* Note the error score left in the remaining frames of the group.
-         * For normal GFs we want to remove the error score for the first
-         * frame of the group (except in Key frame case where this has
-         * already happened)
-         */
-        if (!cpi->source_alt_ref_pending && cpi->common.frame_type != KEY_FRAME)
-            cpi->twopass.gf_group_error_left = (int)(gf_group_err -
-                                                     gf_first_frame_err);
-        else
-            cpi->twopass.gf_group_error_left = (int) gf_group_err;
+    /* Note the error score left in the remaining frames of the group.
+     * For normal GFs we want to remove the error score for the first
+     * frame of the group (except in Key frame case where this has
+     * already happened)
+     */
+    if (!cpi->source_alt_ref_pending && cpi->common.frame_type != KEY_FRAME)
+      cpi->twopass.gf_group_error_left =
+          (int)(gf_group_err - gf_first_frame_err);
+    else
+      cpi->twopass.gf_group_error_left = (int)gf_group_err;
 
-        cpi->twopass.gf_group_bits -= cpi->twopass.gf_bits - cpi->min_frame_bandwidth;
+    cpi->twopass.gf_group_bits -=
+        cpi->twopass.gf_bits - cpi->min_frame_bandwidth;
 
-        if (cpi->twopass.gf_group_bits < 0)
-            cpi->twopass.gf_group_bits = 0;
+    if (cpi->twopass.gf_group_bits < 0) cpi->twopass.gf_group_bits = 0;
 
-        /* This condition could fail if there are two kfs very close together
-         * despite (MIN_GF_INTERVAL) and would cause a devide by 0 in the
-         * calculation of cpi->twopass.alt_extra_bits.
-         */
-        if ( cpi->baseline_gf_interval >= 3 )
-        {
+    /* This condition could fail if there are two kfs very close together
+     * despite (MIN_GF_INTERVAL) and would cause a devide by 0 in the
+     * calculation of cpi->twopass.alt_extra_bits.
+     */
+    if (cpi->baseline_gf_interval >= 3) {
 #if NEW_BOOST
-            int boost = (cpi->source_alt_ref_pending)
-                        ? b_boost : cpi->gfu_boost;
+      int boost = (cpi->source_alt_ref_pending) ? b_boost : cpi->gfu_boost;
 #else
-            int boost = cpi->gfu_boost;
+      int boost = cpi->gfu_boost;
 #endif
-            if ( boost >= 150 )
-            {
-                int pct_extra;
+      if (boost >= 150) {
+        int pct_extra;
 
-                pct_extra = (boost - 100) / 50;
-                pct_extra = (pct_extra > 20) ? 20 : pct_extra;
+        pct_extra = (boost - 100) / 50;
+        pct_extra = (pct_extra > 20) ? 20 : pct_extra;
 
-                cpi->twopass.alt_extra_bits =
-                    (int)(cpi->twopass.gf_group_bits * pct_extra) / 100;
-                cpi->twopass.gf_group_bits -= cpi->twopass.alt_extra_bits;
-                cpi->twopass.alt_extra_bits /=
-                    ((cpi->baseline_gf_interval-1)>>1);
-            }
-            else
-                cpi->twopass.alt_extra_bits = 0;
-        }
-        else
-            cpi->twopass.alt_extra_bits = 0;
-    }
+        cpi->twopass.alt_extra_bits =
+            (int)(cpi->twopass.gf_group_bits * pct_extra) / 100;
+        cpi->twopass.gf_group_bits -= cpi->twopass.alt_extra_bits;
+        cpi->twopass.alt_extra_bits /= ((cpi->baseline_gf_interval - 1) >> 1);
+      } else
+        cpi->twopass.alt_extra_bits = 0;
+    } else
+      cpi->twopass.alt_extra_bits = 0;
+  }
 
-    /* Adjustments based on a measure of complexity of the section */
-    if (cpi->common.frame_type != KEY_FRAME)
-    {
-        FIRSTPASS_STATS sectionstats;
-        double Ratio;
+  /* Adjustments based on a measure of complexity of the section */
+  if (cpi->common.frame_type != KEY_FRAME) {
+    FIRSTPASS_STATS sectionstats;
+    double Ratio;
 
-        zero_stats(&sectionstats);
-        reset_fpf_position(cpi, start_pos);
+    zero_stats(&sectionstats);
+    reset_fpf_position(cpi, start_pos);
 
-        for (i = 0 ; i < cpi->baseline_gf_interval ; i++)
-        {
-            input_stats(cpi, &next_frame);
-            accumulate_stats(&sectionstats, &next_frame);
-        }
+    for (i = 0; i < cpi->baseline_gf_interval; i++) {
+      input_stats(cpi, &next_frame);
+      accumulate_stats(&sectionstats, &next_frame);
+    }
 
-        avg_stats(&sectionstats);
+    avg_stats(&sectionstats);
 
-        cpi->twopass.section_intra_rating = (unsigned int)
-            (sectionstats.intra_error /
-            DOUBLE_DIVIDE_CHECK(sectionstats.coded_error));
+    cpi->twopass.section_intra_rating =
+        (unsigned int)(sectionstats.intra_error /
+                       DOUBLE_DIVIDE_CHECK(sectionstats.coded_error));
 
-        Ratio = sectionstats.intra_error / DOUBLE_DIVIDE_CHECK(sectionstats.coded_error);
-        cpi->twopass.section_max_qfactor = 1.0 - ((Ratio - 10.0) * 0.025);
+    Ratio = sectionstats.intra_error /
+            DOUBLE_DIVIDE_CHECK(sectionstats.coded_error);
+    cpi->twopass.section_max_qfactor = 1.0 - ((Ratio - 10.0) * 0.025);
 
-        if (cpi->twopass.section_max_qfactor < 0.80)
-            cpi->twopass.section_max_qfactor = 0.80;
+    if (cpi->twopass.section_max_qfactor < 0.80)
+      cpi->twopass.section_max_qfactor = 0.80;
 
-        reset_fpf_position(cpi, start_pos);
-    }
+    reset_fpf_position(cpi, start_pos);
+  }
 }
 
-/* Allocate bits to a normal frame that is neither a gf an arf or a key frame. */
-static void assign_std_frame_bits(VP8_COMP *cpi, FIRSTPASS_STATS *this_frame)
-{
-    int    target_frame_size;
+/* Allocate bits to a normal frame that is neither a gf an arf or a key frame.
+ */
+static void assign_std_frame_bits(VP8_COMP *cpi, FIRSTPASS_STATS *this_frame) {
+  int target_frame_size;
 
-    double modified_err;
-    double err_fraction;
+  double modified_err;
+  double err_fraction;
 
-    int max_bits = frame_max_bits(cpi);  /* Max for a single frame */
+  int max_bits = frame_max_bits(cpi); /* Max for a single frame */
 
-    /* Calculate modified prediction error used in bit allocation */
-    modified_err = calculate_modified_err(cpi, this_frame);
+  /* Calculate modified prediction error used in bit allocation */
+  modified_err = calculate_modified_err(cpi, this_frame);
 
-    /* What portion of the remaining GF group error is used by this frame */
-    if (cpi->twopass.gf_group_error_left > 0)
-        err_fraction = modified_err / cpi->twopass.gf_group_error_left;
-    else
-        err_fraction = 0.0;
+  /* What portion of the remaining GF group error is used by this frame */
+  if (cpi->twopass.gf_group_error_left > 0)
+    err_fraction = modified_err / cpi->twopass.gf_group_error_left;
+  else
+    err_fraction = 0.0;
 
-    /* How many of those bits available for allocation should we give it? */
-    target_frame_size = (int)((double)cpi->twopass.gf_group_bits * err_fraction);
+  /* How many of those bits available for allocation should we give it? */
+  target_frame_size = (int)((double)cpi->twopass.gf_group_bits * err_fraction);
 
-    /* Clip to target size to 0 - max_bits (or cpi->twopass.gf_group_bits)
-     * at the top end.
-     */
-    if (target_frame_size < 0)
-        target_frame_size = 0;
-    else
-    {
-        if (target_frame_size > max_bits)
-            target_frame_size = max_bits;
+  /* Clip to target size to 0 - max_bits (or cpi->twopass.gf_group_bits)
+   * at the top end.
+   */
+  if (target_frame_size < 0)
+    target_frame_size = 0;
+  else {
+    if (target_frame_size > max_bits) target_frame_size = max_bits;
 
-        if (target_frame_size > cpi->twopass.gf_group_bits)
-            target_frame_size = (int)cpi->twopass.gf_group_bits;
-    }
+    if (target_frame_size > cpi->twopass.gf_group_bits)
+      target_frame_size = (int)cpi->twopass.gf_group_bits;
+  }
 
-    /* Adjust error and bits remaining */
-    cpi->twopass.gf_group_error_left -= (int)modified_err;
-    cpi->twopass.gf_group_bits -= target_frame_size;
+  /* Adjust error and bits remaining */
+  cpi->twopass.gf_group_error_left -= (int)modified_err;
+  cpi->twopass.gf_group_bits -= target_frame_size;
 
-    if (cpi->twopass.gf_group_bits < 0)
-        cpi->twopass.gf_group_bits = 0;
+  if (cpi->twopass.gf_group_bits < 0) cpi->twopass.gf_group_bits = 0;
 
-    /* Add in the minimum number of bits that is set aside for every frame. */
-    target_frame_size += cpi->min_frame_bandwidth;
+  /* Add in the minimum number of bits that is set aside for every frame. */
+  target_frame_size += cpi->min_frame_bandwidth;
 
-    /* Every other frame gets a few extra bits */
-    if ( (cpi->frames_since_golden & 0x01) &&
-         (cpi->frames_till_gf_update_due > 0) )
-    {
-        target_frame_size += cpi->twopass.alt_extra_bits;
-    }
+  /* Every other frame gets a few extra bits */
+  if ((cpi->frames_since_golden & 0x01) &&
+      (cpi->frames_till_gf_update_due > 0)) {
+    target_frame_size += cpi->twopass.alt_extra_bits;
+  }
 
-    /* Per frame bit target for this frame */
-    cpi->per_frame_bandwidth = target_frame_size;
+  /* Per frame bit target for this frame */
+  cpi->per_frame_bandwidth = target_frame_size;
 }
 
-void vp8_second_pass(VP8_COMP *cpi)
-{
-    int tmp_q;
-    int frames_left = (int)(cpi->twopass.total_stats.count - cpi->common.current_video_frame);
+void vp8_second_pass(VP8_COMP *cpi) {
+  int tmp_q;
+  int frames_left =
+      (int)(cpi->twopass.total_stats.count - cpi->common.current_video_frame);
 
-    FIRSTPASS_STATS this_frame;
-    FIRSTPASS_STATS this_frame_copy;
+  FIRSTPASS_STATS this_frame;
+  FIRSTPASS_STATS this_frame_copy;
 
-    double this_frame_intra_error;
-    double this_frame_coded_error;
+  double this_frame_intra_error;
+  double this_frame_coded_error;
 
-    int overhead_bits;
+  int overhead_bits;
 
-    vp8_zero(this_frame);
+  vp8_zero(this_frame);
 
-    if (!cpi->twopass.stats_in)
-    {
-        return ;
-    }
+  if (!cpi->twopass.stats_in) {
+    return;
+  }
 
-    vp8_clear_system_state();
+  vp8_clear_system_state();
 
-    if (EOF == input_stats(cpi, &this_frame))
-        return;
+  if (EOF == input_stats(cpi, &this_frame)) return;
 
-    this_frame_intra_error = this_frame.intra_error;
-    this_frame_coded_error = this_frame.coded_error;
+  this_frame_intra_error = this_frame.intra_error;
+  this_frame_coded_error = this_frame.coded_error;
 
-    /* keyframe and section processing ! */
-    if (cpi->twopass.frames_to_key == 0)
-    {
-        /* Define next KF group and assign bits to it */
-        memcpy(&this_frame_copy, &this_frame, sizeof(this_frame));
-        find_next_key_frame(cpi, &this_frame_copy);
+  /* keyframe and section processing ! */
+  if (cpi->twopass.frames_to_key == 0) {
+    /* Define next KF group and assign bits to it */
+    memcpy(&this_frame_copy, &this_frame, sizeof(this_frame));
+    find_next_key_frame(cpi, &this_frame_copy);
 
-        /* Special case: Error error_resilient_mode mode does not make much
-         * sense for two pass but with its current meaning this code is
-         * designed to stop outlandish behaviour if someone does set it when
-         * using two pass. It effectively disables GF groups. This is
-         * temporary code until we decide what should really happen in this
-         * case.
-         */
-        if (cpi->oxcf.error_resilient_mode)
-        {
-            cpi->twopass.gf_group_bits = cpi->twopass.kf_group_bits;
-            cpi->twopass.gf_group_error_left =
-                                  (int)cpi->twopass.kf_group_error_left;
-            cpi->baseline_gf_interval = cpi->twopass.frames_to_key;
-            cpi->frames_till_gf_update_due = cpi->baseline_gf_interval;
-            cpi->source_alt_ref_pending = 0;
-        }
-
+    /* Special case: Error error_resilient_mode mode does not make much
+     * sense for two pass but with its current meaning this code is
+     * designed to stop outlandish behaviour if someone does set it when
+     * using two pass. It effectively disables GF groups. This is
+     * temporary code until we decide what should really happen in this
+     * case.
+     */
+    if (cpi->oxcf.error_resilient_mode) {
+      cpi->twopass.gf_group_bits = cpi->twopass.kf_group_bits;
+      cpi->twopass.gf_group_error_left = (int)cpi->twopass.kf_group_error_left;
+      cpi->baseline_gf_interval = cpi->twopass.frames_to_key;
+      cpi->frames_till_gf_update_due = cpi->baseline_gf_interval;
+      cpi->source_alt_ref_pending = 0;
     }
+  }
 
-    /* Is this a GF / ARF (Note that a KF is always also a GF) */
-    if (cpi->frames_till_gf_update_due == 0)
-    {
-        /* Define next gf group and assign bits to it */
-        memcpy(&this_frame_copy, &this_frame, sizeof(this_frame));
-        define_gf_group(cpi, &this_frame_copy);
+  /* Is this a GF / ARF (Note that a KF is always also a GF) */
+  if (cpi->frames_till_gf_update_due == 0) {
+    /* Define next gf group and assign bits to it */
+    memcpy(&this_frame_copy, &this_frame, sizeof(this_frame));
+    define_gf_group(cpi, &this_frame_copy);
 
-        /* If we are going to code an altref frame at the end of the group
-         * and the current frame is not a key frame.... If the previous
-         * group used an arf this frame has already benefited from that arf
-         * boost and it should not be given extra bits If the previous
-         * group was NOT coded using arf we may want to apply some boost to
-         * this GF as well
-         */
-        if (cpi->source_alt_ref_pending && (cpi->common.frame_type != KEY_FRAME))
-        {
-            /* Assign a standard frames worth of bits from those allocated
-             * to the GF group
-             */
-            int bak = cpi->per_frame_bandwidth;
-            memcpy(&this_frame_copy, &this_frame, sizeof(this_frame));
-            assign_std_frame_bits(cpi, &this_frame_copy);
-            cpi->per_frame_bandwidth = bak;
-        }
+    /* If we are going to code an altref frame at the end of the group
+     * and the current frame is not a key frame.... If the previous
+     * group used an arf this frame has already benefited from that arf
+     * boost and it should not be given extra bits If the previous
+     * group was NOT coded using arf we may want to apply some boost to
+     * this GF as well
+     */
+    if (cpi->source_alt_ref_pending && (cpi->common.frame_type != KEY_FRAME)) {
+      /* Assign a standard frames worth of bits from those allocated
+       * to the GF group
+       */
+      int bak = cpi->per_frame_bandwidth;
+      memcpy(&this_frame_copy, &this_frame, sizeof(this_frame));
+      assign_std_frame_bits(cpi, &this_frame_copy);
+      cpi->per_frame_bandwidth = bak;
     }
+  }
 
-    /* Otherwise this is an ordinary frame */
-    else
-    {
-        /* Special case: Error error_resilient_mode mode does not make much
-         * sense for two pass but with its current meaning but this code is
-         * designed to stop outlandish behaviour if someone does set it
-         * when using two pass. It effectively disables GF groups. This is
-         * temporary code till we decide what should really happen in this
-         * case.
-         */
-        if (cpi->oxcf.error_resilient_mode)
-        {
-            cpi->frames_till_gf_update_due = cpi->twopass.frames_to_key;
+  /* Otherwise this is an ordinary frame */
+  else {
+    /* Special case: Error error_resilient_mode mode does not make much
+     * sense for two pass but with its current meaning but this code is
+     * designed to stop outlandish behaviour if someone does set it
+     * when using two pass. It effectively disables GF groups. This is
+     * temporary code till we decide what should really happen in this
+     * case.
+     */
+    if (cpi->oxcf.error_resilient_mode) {
+      cpi->frames_till_gf_update_due = cpi->twopass.frames_to_key;
 
-            if (cpi->common.frame_type != KEY_FRAME)
-            {
-                /* Assign bits from those allocated to the GF group */
-                memcpy(&this_frame_copy, &this_frame, sizeof(this_frame));
-                assign_std_frame_bits(cpi, &this_frame_copy);
-            }
-        }
-        else
-        {
-            /* Assign bits from those allocated to the GF group */
-            memcpy(&this_frame_copy, &this_frame, sizeof(this_frame));
-            assign_std_frame_bits(cpi, &this_frame_copy);
-        }
+      if (cpi->common.frame_type != KEY_FRAME) {
+        /* Assign bits from those allocated to the GF group */
+        memcpy(&this_frame_copy, &this_frame, sizeof(this_frame));
+        assign_std_frame_bits(cpi, &this_frame_copy);
+      }
+    } else {
+      /* Assign bits from those allocated to the GF group */
+      memcpy(&this_frame_copy, &this_frame, sizeof(this_frame));
+      assign_std_frame_bits(cpi, &this_frame_copy);
     }
+  }
 
-    /* Keep a globally available copy of this and the next frame's iiratio. */
-    cpi->twopass.this_iiratio = (unsigned int)(this_frame_intra_error /
-                        DOUBLE_DIVIDE_CHECK(this_frame_coded_error));
-    {
-        FIRSTPASS_STATS next_frame;
-        if ( lookup_next_frame_stats(cpi, &next_frame) != EOF )
-        {
-            cpi->twopass.next_iiratio = (unsigned int)(next_frame.intra_error /
-                                DOUBLE_DIVIDE_CHECK(next_frame.coded_error));
-        }
+  /* Keep a globally available copy of this and the next frame's iiratio. */
+  cpi->twopass.this_iiratio =
+      (unsigned int)(this_frame_intra_error /
+                     DOUBLE_DIVIDE_CHECK(this_frame_coded_error));
+  {
+    FIRSTPASS_STATS next_frame;
+    if (lookup_next_frame_stats(cpi, &next_frame) != EOF) {
+      cpi->twopass.next_iiratio =
+          (unsigned int)(next_frame.intra_error /
+                         DOUBLE_DIVIDE_CHECK(next_frame.coded_error));
     }
+  }
 
-    /* Set nominal per second bandwidth for this frame */
-    cpi->target_bandwidth = (int)
-    (cpi->per_frame_bandwidth * cpi->output_framerate);
-    if (cpi->target_bandwidth < 0)
-        cpi->target_bandwidth = 0;
+  /* Set nominal per second bandwidth for this frame */
+  cpi->target_bandwidth =
+      (int)(cpi->per_frame_bandwidth * cpi->output_framerate);
+  if (cpi->target_bandwidth < 0) cpi->target_bandwidth = 0;
 
+  /* Account for mv, mode and other overheads. */
+  overhead_bits = (int)estimate_modemvcost(cpi, &cpi->twopass.total_left_stats);
 
-    /* Account for mv, mode and other overheads. */
-    overhead_bits = (int)estimate_modemvcost(
-                        cpi, &cpi->twopass.total_left_stats );
+  /* Special case code for first frame. */
+  if (cpi->common.current_video_frame == 0) {
+    cpi->twopass.est_max_qcorrection_factor = 1.0;
 
-    /* Special case code for first frame. */
-    if (cpi->common.current_video_frame == 0)
-    {
-        cpi->twopass.est_max_qcorrection_factor = 1.0;
+    /* Set a cq_level in constrained quality mode. */
+    if (cpi->oxcf.end_usage == USAGE_CONSTRAINED_QUALITY) {
+      int est_cq;
 
-        /* Set a cq_level in constrained quality mode. */
-        if ( cpi->oxcf.end_usage == USAGE_CONSTRAINED_QUALITY )
-        {
-            int est_cq;
+      est_cq = estimate_cq(cpi, &cpi->twopass.total_left_stats,
+                           (int)(cpi->twopass.bits_left / frames_left),
+                           overhead_bits);
 
-            est_cq =
-                estimate_cq( cpi,
-                             &cpi->twopass.total_left_stats,
-                             (int)(cpi->twopass.bits_left / frames_left),
-                             overhead_bits );
+      cpi->cq_target_quality = cpi->oxcf.cq_level;
+      if (est_cq > cpi->cq_target_quality) cpi->cq_target_quality = est_cq;
+    }
 
-            cpi->cq_target_quality = cpi->oxcf.cq_level;
-            if ( est_cq > cpi->cq_target_quality )
-                cpi->cq_target_quality = est_cq;
-        }
+    /* guess at maxq needed in 2nd pass */
+    cpi->twopass.maxq_max_limit = cpi->worst_quality;
+    cpi->twopass.maxq_min_limit = cpi->best_quality;
 
-        /* guess at maxq needed in 2nd pass */
-        cpi->twopass.maxq_max_limit = cpi->worst_quality;
-        cpi->twopass.maxq_min_limit = cpi->best_quality;
+    tmp_q = estimate_max_q(cpi, &cpi->twopass.total_left_stats,
+                           (int)(cpi->twopass.bits_left / frames_left),
+                           overhead_bits);
 
-        tmp_q = estimate_max_q(
-                    cpi,
-                    &cpi->twopass.total_left_stats,
-                    (int)(cpi->twopass.bits_left / frames_left),
-                    overhead_bits );
+    /* Limit the maxq value returned subsequently.
+     * This increases the risk of overspend or underspend if the initial
+     * estimate for the clip is bad, but helps prevent excessive
+     * variation in Q, especially near the end of a clip
+     * where for example a small overspend may cause Q to crash
+     */
+    cpi->twopass.maxq_max_limit =
+        ((tmp_q + 32) < cpi->worst_quality) ? (tmp_q + 32) : cpi->worst_quality;
+    cpi->twopass.maxq_min_limit =
+        ((tmp_q - 32) > cpi->best_quality) ? (tmp_q - 32) : cpi->best_quality;
 
-        /* Limit the maxq value returned subsequently.
-         * This increases the risk of overspend or underspend if the initial
-         * estimate for the clip is bad, but helps prevent excessive
-         * variation in Q, especially near the end of a clip
-         * where for example a small overspend may cause Q to crash
-         */
-        cpi->twopass.maxq_max_limit = ((tmp_q + 32) < cpi->worst_quality)
-                                  ? (tmp_q + 32) : cpi->worst_quality;
-        cpi->twopass.maxq_min_limit = ((tmp_q - 32) > cpi->best_quality)
-                                  ? (tmp_q - 32) : cpi->best_quality;
+    cpi->active_worst_quality = tmp_q;
+    cpi->ni_av_qi = tmp_q;
+  }
 
-        cpi->active_worst_quality         = tmp_q;
-        cpi->ni_av_qi                     = tmp_q;
-    }
+  /* The last few frames of a clip almost always have to few or too many
+   * bits and for the sake of over exact rate control we dont want to make
+   * radical adjustments to the allowed quantizer range just to use up a
+   * few surplus bits or get beneath the target rate.
+   */
+  else if ((cpi->common.current_video_frame <
+            (((unsigned int)cpi->twopass.total_stats.count * 255) >> 8)) &&
+           ((cpi->common.current_video_frame + cpi->baseline_gf_interval) <
+            (unsigned int)cpi->twopass.total_stats.count)) {
+    if (frames_left < 1) frames_left = 1;
 
-    /* The last few frames of a clip almost always have to few or too many
-     * bits and for the sake of over exact rate control we dont want to make
-     * radical adjustments to the allowed quantizer range just to use up a
-     * few surplus bits or get beneath the target rate.
-     */
-    else if ( (cpi->common.current_video_frame <
-                 (((unsigned int)cpi->twopass.total_stats.count * 255)>>8)) &&
-              ((cpi->common.current_video_frame + cpi->baseline_gf_interval) <
-                 (unsigned int)cpi->twopass.total_stats.count) )
-    {
-        if (frames_left < 1)
-            frames_left = 1;
+    tmp_q = estimate_max_q(cpi, &cpi->twopass.total_left_stats,
+                           (int)(cpi->twopass.bits_left / frames_left),
+                           overhead_bits);
 
-        tmp_q = estimate_max_q(
-                    cpi,
-                    &cpi->twopass.total_left_stats,
-                    (int)(cpi->twopass.bits_left / frames_left),
-                    overhead_bits );
+    /* Move active_worst_quality but in a damped way */
+    if (tmp_q > cpi->active_worst_quality)
+      cpi->active_worst_quality++;
+    else if (tmp_q < cpi->active_worst_quality)
+      cpi->active_worst_quality--;
 
-        /* Move active_worst_quality but in a damped way */
-        if (tmp_q > cpi->active_worst_quality)
-            cpi->active_worst_quality ++;
-        else if (tmp_q < cpi->active_worst_quality)
-            cpi->active_worst_quality --;
+    cpi->active_worst_quality =
+        ((cpi->active_worst_quality * 3) + tmp_q + 2) / 4;
+  }
 
-        cpi->active_worst_quality =
-            ((cpi->active_worst_quality * 3) + tmp_q + 2) / 4;
-    }
+  cpi->twopass.frames_to_key--;
 
-    cpi->twopass.frames_to_key --;
-
-    /* Update the total stats remaining sturcture */
-    subtract_stats(&cpi->twopass.total_left_stats, &this_frame );
+  /* Update the total stats remaining sturcture */
+  subtract_stats(&cpi->twopass.total_left_stats, &this_frame);
 }
 
+static int test_candidate_kf(VP8_COMP *cpi, FIRSTPASS_STATS *last_frame,
+                             FIRSTPASS_STATS *this_frame,
+                             FIRSTPASS_STATS *next_frame) {
+  int is_viable_kf = 0;
 
-static int test_candidate_kf(VP8_COMP *cpi,  FIRSTPASS_STATS *last_frame, FIRSTPASS_STATS *this_frame, FIRSTPASS_STATS *next_frame)
-{
-    int is_viable_kf = 0;
+  /* Does the frame satisfy the primary criteria of a key frame
+   *      If so, then examine how well it predicts subsequent frames
+   */
+  if ((this_frame->pcnt_second_ref < 0.10) &&
+      (next_frame->pcnt_second_ref < 0.10) &&
+      ((this_frame->pcnt_inter < 0.05) ||
+       (((this_frame->pcnt_inter - this_frame->pcnt_neutral) < .25) &&
+        ((this_frame->intra_error /
+          DOUBLE_DIVIDE_CHECK(this_frame->coded_error)) < 2.5) &&
+        ((fabs(last_frame->coded_error - this_frame->coded_error) /
+              DOUBLE_DIVIDE_CHECK(this_frame->coded_error) >
+          .40) ||
+         (fabs(last_frame->intra_error - this_frame->intra_error) /
+              DOUBLE_DIVIDE_CHECK(this_frame->intra_error) >
+          .40) ||
+         ((next_frame->intra_error /
+           DOUBLE_DIVIDE_CHECK(next_frame->coded_error)) > 3.5))))) {
+    int i;
+    FIRSTPASS_STATS *start_pos;
 
-    /* Does the frame satisfy the primary criteria of a key frame
-     *      If so, then examine how well it predicts subsequent frames
-     */
-    if ((this_frame->pcnt_second_ref < 0.10) &&
-        (next_frame->pcnt_second_ref < 0.10) &&
-        ((this_frame->pcnt_inter < 0.05) ||
-         (
-             ((this_frame->pcnt_inter - this_frame->pcnt_neutral) < .25) &&
-             ((this_frame->intra_error / DOUBLE_DIVIDE_CHECK(this_frame->coded_error)) < 2.5) &&
-             ((fabs(last_frame->coded_error - this_frame->coded_error) / DOUBLE_DIVIDE_CHECK(this_frame->coded_error) > .40) ||
-              (fabs(last_frame->intra_error - this_frame->intra_error) / DOUBLE_DIVIDE_CHECK(this_frame->intra_error) > .40) ||
-              ((next_frame->intra_error / DOUBLE_DIVIDE_CHECK(next_frame->coded_error)) > 3.5)
-             )
-         )
-        )
-       )
-    {
-        int i;
-        FIRSTPASS_STATS *start_pos;
+    FIRSTPASS_STATS local_next_frame;
 
-        FIRSTPASS_STATS local_next_frame;
+    double boost_score = 0.0;
+    double old_boost_score = 0.0;
+    double decay_accumulator = 1.0;
+    double next_iiratio;
 
-        double boost_score = 0.0;
-        double old_boost_score = 0.0;
-        double decay_accumulator = 1.0;
-        double next_iiratio;
+    memcpy(&local_next_frame, next_frame, sizeof(*next_frame));
 
-        memcpy(&local_next_frame, next_frame, sizeof(*next_frame));
+    /* Note the starting file position so we can reset to it */
+    start_pos = cpi->twopass.stats_in;
 
-        /* Note the starting file position so we can reset to it */
-        start_pos = cpi->twopass.stats_in;
+    /* Examine how well the key frame predicts subsequent frames */
+    for (i = 0; i < 16; i++) {
+      next_iiratio = (IIKFACTOR1 * local_next_frame.intra_error /
+                      DOUBLE_DIVIDE_CHECK(local_next_frame.coded_error));
 
-        /* Examine how well the key frame predicts subsequent frames */
-        for (i = 0 ; i < 16; i++)
-        {
-            next_iiratio = (IIKFACTOR1 * local_next_frame.intra_error / DOUBLE_DIVIDE_CHECK(local_next_frame.coded_error)) ;
+      if (next_iiratio > RMAX) next_iiratio = RMAX;
 
-            if (next_iiratio > RMAX)
-                next_iiratio = RMAX;
+      /* Cumulative effect of decay in prediction quality */
+      if (local_next_frame.pcnt_inter > 0.85)
+        decay_accumulator = decay_accumulator * local_next_frame.pcnt_inter;
+      else
+        decay_accumulator =
+            decay_accumulator * ((0.85 + local_next_frame.pcnt_inter) / 2.0);
 
-            /* Cumulative effect of decay in prediction quality */
-            if (local_next_frame.pcnt_inter > 0.85)
-                decay_accumulator = decay_accumulator * local_next_frame.pcnt_inter;
-            else
-                decay_accumulator = decay_accumulator * ((0.85 + local_next_frame.pcnt_inter) / 2.0);
+      /* Keep a running total */
+      boost_score += (decay_accumulator * next_iiratio);
 
-            /* Keep a running total */
-            boost_score += (decay_accumulator * next_iiratio);
+      /* Test various breakout clauses */
+      if ((local_next_frame.pcnt_inter < 0.05) || (next_iiratio < 1.5) ||
+          (((local_next_frame.pcnt_inter - local_next_frame.pcnt_neutral) <
+            0.20) &&
+           (next_iiratio < 3.0)) ||
+          ((boost_score - old_boost_score) < 0.5) ||
+          (local_next_frame.intra_error < 200)) {
+        break;
+      }
 
-            /* Test various breakout clauses */
-            if ((local_next_frame.pcnt_inter < 0.05) ||
-                (next_iiratio < 1.5) ||
-                (((local_next_frame.pcnt_inter -
-                   local_next_frame.pcnt_neutral) < 0.20) &&
-                 (next_iiratio < 3.0)) ||
-                ((boost_score - old_boost_score) < 0.5) ||
-                (local_next_frame.intra_error < 200)
-               )
-            {
-                break;
-            }
+      old_boost_score = boost_score;
 
-            old_boost_score = boost_score;
+      /* Get the next frame details */
+      if (EOF == input_stats(cpi, &local_next_frame)) break;
+    }
 
-            /* Get the next frame details */
-            if (EOF == input_stats(cpi, &local_next_frame))
-                break;
-        }
+    /* If there is tolerable prediction for at least the next 3 frames
+     * then break out else discard this pottential key frame and move on
+     */
+    if (boost_score > 5.0 && (i > 3))
+      is_viable_kf = 1;
+    else {
+      /* Reset the file position */
+      reset_fpf_position(cpi, start_pos);
 
-        /* If there is tolerable prediction for at least the next 3 frames
-         * then break out else discard this pottential key frame and move on
-         */
-        if (boost_score > 5.0 && (i > 3))
-            is_viable_kf = 1;
-        else
-        {
-            /* Reset the file position */
-            reset_fpf_position(cpi, start_pos);
-
-            is_viable_kf = 0;
-        }
+      is_viable_kf = 0;
     }
+  }
 
-    return is_viable_kf;
+  return is_viable_kf;
 }
-static void find_next_key_frame(VP8_COMP *cpi, FIRSTPASS_STATS *this_frame)
-{
-    int i,j;
-    FIRSTPASS_STATS last_frame;
-    FIRSTPASS_STATS first_frame;
-    FIRSTPASS_STATS next_frame;
-    FIRSTPASS_STATS *start_position;
+static void find_next_key_frame(VP8_COMP *cpi, FIRSTPASS_STATS *this_frame) {
+  int i, j;
+  FIRSTPASS_STATS last_frame;
+  FIRSTPASS_STATS first_frame;
+  FIRSTPASS_STATS next_frame;
+  FIRSTPASS_STATS *start_position;
 
-    double decay_accumulator = 1.0;
-    double boost_score = 0;
-    double old_boost_score = 0.0;
-    double loop_decay_rate;
+  double decay_accumulator = 1.0;
+  double boost_score = 0;
+  double old_boost_score = 0.0;
+  double loop_decay_rate;
 
-    double kf_mod_err = 0.0;
-    double kf_group_err = 0.0;
-    double kf_group_intra_err = 0.0;
-    double kf_group_coded_err = 0.0;
-    double recent_loop_decay[8] = {1.0,1.0,1.0,1.0,1.0,1.0,1.0,1.0};
+  double kf_mod_err = 0.0;
+  double kf_group_err = 0.0;
+  double kf_group_intra_err = 0.0;
+  double kf_group_coded_err = 0.0;
+  double recent_loop_decay[8] = { 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0 };
 
-    memset(&next_frame, 0, sizeof(next_frame));
+  memset(&next_frame, 0, sizeof(next_frame));
 
-    vp8_clear_system_state();
-    start_position = cpi->twopass.stats_in;
+  vp8_clear_system_state();
+  start_position = cpi->twopass.stats_in;
 
-    cpi->common.frame_type = KEY_FRAME;
+  cpi->common.frame_type = KEY_FRAME;
 
-    /* is this a forced key frame by interval */
-    cpi->this_key_frame_forced = cpi->next_key_frame_forced;
+  /* is this a forced key frame by interval */
+  cpi->this_key_frame_forced = cpi->next_key_frame_forced;
 
-    /* Clear the alt ref active flag as this can never be active on a key
-     * frame
-     */
-    cpi->source_alt_ref_active = 0;
+  /* Clear the alt ref active flag as this can never be active on a key
+   * frame
+   */
+  cpi->source_alt_ref_active = 0;
 
-    /* Kf is always a gf so clear frames till next gf counter */
-    cpi->frames_till_gf_update_due = 0;
+  /* Kf is always a gf so clear frames till next gf counter */
+  cpi->frames_till_gf_update_due = 0;
 
-    cpi->twopass.frames_to_key = 1;
+  cpi->twopass.frames_to_key = 1;
 
-    /* Take a copy of the initial frame details */
-    memcpy(&first_frame, this_frame, sizeof(*this_frame));
+  /* Take a copy of the initial frame details */
+  memcpy(&first_frame, this_frame, sizeof(*this_frame));
 
-    cpi->twopass.kf_group_bits = 0;
-    cpi->twopass.kf_group_error_left = 0;
+  cpi->twopass.kf_group_bits = 0;
+  cpi->twopass.kf_group_error_left = 0;
 
-    kf_mod_err = calculate_modified_err(cpi, this_frame);
+  kf_mod_err = calculate_modified_err(cpi, this_frame);
 
-    /* find the next keyframe */
-    i = 0;
-    while (cpi->twopass.stats_in < cpi->twopass.stats_in_end)
-    {
-        /* Accumulate kf group error */
-        kf_group_err += calculate_modified_err(cpi, this_frame);
+  /* find the next keyframe */
+  i = 0;
+  while (cpi->twopass.stats_in < cpi->twopass.stats_in_end) {
+    /* Accumulate kf group error */
+    kf_group_err += calculate_modified_err(cpi, this_frame);
 
-        /* These figures keep intra and coded error counts for all frames
-         * including key frames in the group. The effect of the key frame
-         * itself can be subtracted out using the first_frame data
-         * collected above
-         */
-        kf_group_intra_err += this_frame->intra_error;
-        kf_group_coded_err += this_frame->coded_error;
+    /* These figures keep intra and coded error counts for all frames
+     * including key frames in the group. The effect of the key frame
+     * itself can be subtracted out using the first_frame data
+     * collected above
+     */
+    kf_group_intra_err += this_frame->intra_error;
+    kf_group_coded_err += this_frame->coded_error;
 
-        /* Load the next frame's stats. */
-        memcpy(&last_frame, this_frame, sizeof(*this_frame));
-        input_stats(cpi, this_frame);
+    /* Load the next frame's stats. */
+    memcpy(&last_frame, this_frame, sizeof(*this_frame));
+    input_stats(cpi, this_frame);
 
-        /* Provided that we are not at the end of the file... */
-        if (cpi->oxcf.auto_key
-            && lookup_next_frame_stats(cpi, &next_frame) != EOF)
-        {
-            /* Normal scene cut check */
-            if ( ( i >= MIN_GF_INTERVAL ) &&
-                 test_candidate_kf(cpi, &last_frame, this_frame, &next_frame) )
-            {
-                break;
-            }
+    /* Provided that we are not at the end of the file... */
+    if (cpi->oxcf.auto_key &&
+        lookup_next_frame_stats(cpi, &next_frame) != EOF) {
+      /* Normal scene cut check */
+      if ((i >= MIN_GF_INTERVAL) &&
+          test_candidate_kf(cpi, &last_frame, this_frame, &next_frame)) {
+        break;
+      }
 
-            /* How fast is prediction quality decaying */
-            loop_decay_rate = get_prediction_decay_rate(cpi, &next_frame);
+      /* How fast is prediction quality decaying */
+      loop_decay_rate = get_prediction_decay_rate(cpi, &next_frame);
 
-            /* We want to know something about the recent past... rather than
-             * as used elsewhere where we are concened with decay in prediction
-             * quality since the last GF or KF.
-             */
-            recent_loop_decay[i%8] = loop_decay_rate;
-            decay_accumulator = 1.0;
-            for (j = 0; j < 8; j++)
-            {
-                decay_accumulator = decay_accumulator * recent_loop_decay[j];
-            }
+      /* We want to know something about the recent past... rather than
+       * as used elsewhere where we are concened with decay in prediction
+       * quality since the last GF or KF.
+       */
+      recent_loop_decay[i % 8] = loop_decay_rate;
+      decay_accumulator = 1.0;
+      for (j = 0; j < 8; j++) {
+        decay_accumulator = decay_accumulator * recent_loop_decay[j];
+      }
 
-            /* Special check for transition or high motion followed by a
-             * static scene.
-             */
-            if ( detect_transition_to_still( cpi, i,
-                                             ((int)(cpi->key_frame_frequency) -
-                                              (int)i),
-                                             loop_decay_rate,
-                                             decay_accumulator ) )
-            {
-                break;
-            }
+      /* Special check for transition or high motion followed by a
+       * static scene.
+       */
+      if (detect_transition_to_still(cpi, i,
+                                     ((int)(cpi->key_frame_frequency) - (int)i),
+                                     loop_decay_rate, decay_accumulator)) {
+        break;
+      }
 
+      /* Step on to the next frame */
+      cpi->twopass.frames_to_key++;
 
-            /* Step on to the next frame */
-            cpi->twopass.frames_to_key ++;
+      /* If we don't have a real key frame within the next two
+       * forcekeyframeevery intervals then break out of the loop.
+       */
+      if (cpi->twopass.frames_to_key >= 2 * (int)cpi->key_frame_frequency)
+        break;
+    } else
+      cpi->twopass.frames_to_key++;
 
-            /* If we don't have a real key frame within the next two
-             * forcekeyframeevery intervals then break out of the loop.
-             */
-            if (cpi->twopass.frames_to_key >= 2 *(int)cpi->key_frame_frequency)
-                break;
-        } else
-            cpi->twopass.frames_to_key ++;
+    i++;
+  }
 
-        i++;
-    }
+  /* If there is a max kf interval set by the user we must obey it.
+   * We already breakout of the loop above at 2x max.
+   * This code centers the extra kf if the actual natural
+   * interval is between 1x and 2x
+   */
+  if (cpi->oxcf.auto_key &&
+      cpi->twopass.frames_to_key > (int)cpi->key_frame_frequency) {
+    FIRSTPASS_STATS *current_pos = cpi->twopass.stats_in;
+    FIRSTPASS_STATS tmp_frame;
 
-    /* If there is a max kf interval set by the user we must obey it.
-     * We already breakout of the loop above at 2x max.
-     * This code centers the extra kf if the actual natural
-     * interval is between 1x and 2x
-     */
-    if (cpi->oxcf.auto_key
-        && cpi->twopass.frames_to_key > (int)cpi->key_frame_frequency )
-    {
-        FIRSTPASS_STATS *current_pos = cpi->twopass.stats_in;
-        FIRSTPASS_STATS tmp_frame;
+    cpi->twopass.frames_to_key /= 2;
 
-        cpi->twopass.frames_to_key /= 2;
+    /* Copy first frame details */
+    memcpy(&tmp_frame, &first_frame, sizeof(first_frame));
 
-        /* Copy first frame details */
-        memcpy(&tmp_frame, &first_frame, sizeof(first_frame));
+    /* Reset to the start of the group */
+    reset_fpf_position(cpi, start_position);
 
-        /* Reset to the start of the group */
-        reset_fpf_position(cpi, start_position);
+    kf_group_err = 0;
+    kf_group_intra_err = 0;
+    kf_group_coded_err = 0;
 
-        kf_group_err = 0;
-        kf_group_intra_err = 0;
-        kf_group_coded_err = 0;
+    /* Rescan to get the correct error data for the forced kf group */
+    for (i = 0; i < cpi->twopass.frames_to_key; i++) {
+      /* Accumulate kf group errors */
+      kf_group_err += calculate_modified_err(cpi, &tmp_frame);
+      kf_group_intra_err += tmp_frame.intra_error;
+      kf_group_coded_err += tmp_frame.coded_error;
 
-        /* Rescan to get the correct error data for the forced kf group */
-        for( i = 0; i < cpi->twopass.frames_to_key; i++ )
-        {
-            /* Accumulate kf group errors */
-            kf_group_err += calculate_modified_err(cpi, &tmp_frame);
-            kf_group_intra_err += tmp_frame.intra_error;
-            kf_group_coded_err += tmp_frame.coded_error;
+      /* Load a the next frame's stats */
+      input_stats(cpi, &tmp_frame);
+    }
 
-            /* Load a the next frame's stats */
-            input_stats(cpi, &tmp_frame);
-        }
+    /* Reset to the start of the group */
+    reset_fpf_position(cpi, current_pos);
 
-        /* Reset to the start of the group */
-        reset_fpf_position(cpi, current_pos);
+    cpi->next_key_frame_forced = 1;
+  } else
+    cpi->next_key_frame_forced = 0;
 
-        cpi->next_key_frame_forced = 1;
-    }
-    else
-        cpi->next_key_frame_forced = 0;
+  /* Special case for the last frame of the file */
+  if (cpi->twopass.stats_in >= cpi->twopass.stats_in_end) {
+    /* Accumulate kf group error */
+    kf_group_err += calculate_modified_err(cpi, this_frame);
 
-    /* Special case for the last frame of the file */
-    if (cpi->twopass.stats_in >= cpi->twopass.stats_in_end)
-    {
-        /* Accumulate kf group error */
-        kf_group_err += calculate_modified_err(cpi, this_frame);
+    /* These figures keep intra and coded error counts for all frames
+     * including key frames in the group. The effect of the key frame
+     * itself can be subtracted out using the first_frame data
+     * collected above
+     */
+    kf_group_intra_err += this_frame->intra_error;
+    kf_group_coded_err += this_frame->coded_error;
+  }
 
-        /* These figures keep intra and coded error counts for all frames
-         * including key frames in the group. The effect of the key frame
-         * itself can be subtracted out using the first_frame data
-         * collected above
-         */
-        kf_group_intra_err += this_frame->intra_error;
-        kf_group_coded_err += this_frame->coded_error;
-    }
+  /* Calculate the number of bits that should be assigned to the kf group. */
+  if ((cpi->twopass.bits_left > 0) &&
+      (cpi->twopass.modified_error_left > 0.0)) {
+    /* Max for a single normal frame (not key frame) */
+    int max_bits = frame_max_bits(cpi);
 
-    /* Calculate the number of bits that should be assigned to the kf group. */
-    if ((cpi->twopass.bits_left > 0) && (cpi->twopass.modified_error_left > 0.0))
-    {
-        /* Max for a single normal frame (not key frame) */
-        int max_bits = frame_max_bits(cpi);
+    /* Maximum bits for the kf group */
+    int64_t max_grp_bits;
 
-        /* Maximum bits for the kf group */
-        int64_t max_grp_bits;
+    /* Default allocation based on bits left and relative
+     * complexity of the section
+     */
+    cpi->twopass.kf_group_bits =
+        (int64_t)(cpi->twopass.bits_left *
+                  (kf_group_err / cpi->twopass.modified_error_left));
 
-        /* Default allocation based on bits left and relative
-         * complexity of the section
-         */
-        cpi->twopass.kf_group_bits = (int64_t)( cpi->twopass.bits_left *
-                                          ( kf_group_err /
-                                            cpi->twopass.modified_error_left ));
+    /* Clip based on maximum per frame rate defined by the user. */
+    max_grp_bits = (int64_t)max_bits * (int64_t)cpi->twopass.frames_to_key;
+    if (cpi->twopass.kf_group_bits > max_grp_bits)
+      cpi->twopass.kf_group_bits = max_grp_bits;
 
-        /* Clip based on maximum per frame rate defined by the user. */
-        max_grp_bits = (int64_t)max_bits * (int64_t)cpi->twopass.frames_to_key;
-        if (cpi->twopass.kf_group_bits > max_grp_bits)
-            cpi->twopass.kf_group_bits = max_grp_bits;
+    /* Additional special case for CBR if buffer is getting full. */
+    if (cpi->oxcf.end_usage == USAGE_STREAM_FROM_SERVER) {
+      int64_t opt_buffer_lvl = cpi->oxcf.optimal_buffer_level;
+      int64_t buffer_lvl = cpi->buffer_level;
 
-        /* Additional special case for CBR if buffer is getting full. */
-        if (cpi->oxcf.end_usage == USAGE_STREAM_FROM_SERVER)
-        {
-            int64_t opt_buffer_lvl = cpi->oxcf.optimal_buffer_level;
-            int64_t buffer_lvl = cpi->buffer_level;
+      /* If the buffer is near or above the optimal and this kf group is
+       * not being allocated much then increase the allocation a bit.
+       */
+      if (buffer_lvl >= opt_buffer_lvl) {
+        int64_t high_water_mark =
+            (opt_buffer_lvl + cpi->oxcf.maximum_buffer_size) >> 1;
 
-            /* If the buffer is near or above the optimal and this kf group is
-             * not being allocated much then increase the allocation a bit.
-             */
-            if (buffer_lvl >= opt_buffer_lvl)
-            {
-                int64_t high_water_mark = (opt_buffer_lvl +
-                                       cpi->oxcf.maximum_buffer_size) >> 1;
+        int64_t av_group_bits;
 
-                int64_t av_group_bits;
+        /* Av bits per frame * number of frames */
+        av_group_bits = (int64_t)cpi->av_per_frame_bandwidth *
+                        (int64_t)cpi->twopass.frames_to_key;
 
-                /* Av bits per frame * number of frames */
-                av_group_bits = (int64_t)cpi->av_per_frame_bandwidth *
-                                (int64_t)cpi->twopass.frames_to_key;
+        /* We are at or above the maximum. */
+        if (cpi->buffer_level >= high_water_mark) {
+          int64_t min_group_bits;
 
-                /* We are at or above the maximum. */
-                if (cpi->buffer_level >= high_water_mark)
-                {
-                    int64_t min_group_bits;
+          min_group_bits =
+              av_group_bits + (int64_t)(buffer_lvl - high_water_mark);
 
-                    min_group_bits = av_group_bits +
-                                     (int64_t)(buffer_lvl -
-                                                 high_water_mark);
+          if (cpi->twopass.kf_group_bits < min_group_bits)
+            cpi->twopass.kf_group_bits = min_group_bits;
+        }
+        /* We are above optimal but below the maximum */
+        else if (cpi->twopass.kf_group_bits < av_group_bits) {
+          int64_t bits_below_av = av_group_bits - cpi->twopass.kf_group_bits;
 
-                    if (cpi->twopass.kf_group_bits < min_group_bits)
-                        cpi->twopass.kf_group_bits = min_group_bits;
-                }
-                /* We are above optimal but below the maximum */
-                else if (cpi->twopass.kf_group_bits < av_group_bits)
-                {
-                    int64_t bits_below_av = av_group_bits -
-                                              cpi->twopass.kf_group_bits;
-
-                    cpi->twopass.kf_group_bits +=
-                       (int64_t)((double)bits_below_av *
-                                   (double)(buffer_lvl - opt_buffer_lvl) /
-                                   (double)(high_water_mark - opt_buffer_lvl));
-                }
-            }
+          cpi->twopass.kf_group_bits += (int64_t)(
+              (double)bits_below_av * (double)(buffer_lvl - opt_buffer_lvl) /
+              (double)(high_water_mark - opt_buffer_lvl));
         }
+      }
     }
-    else
-        cpi->twopass.kf_group_bits = 0;
+  } else
+    cpi->twopass.kf_group_bits = 0;
 
-    /* Reset the first pass file position */
-    reset_fpf_position(cpi, start_position);
+  /* Reset the first pass file position */
+  reset_fpf_position(cpi, start_position);
 
-    /* determine how big to make this keyframe based on how well the
-     * subsequent frames use inter blocks
-     */
-    decay_accumulator = 1.0;
-    boost_score = 0.0;
+  /* determine how big to make this keyframe based on how well the
+   * subsequent frames use inter blocks
+   */
+  decay_accumulator = 1.0;
+  boost_score = 0.0;
 
-    for (i = 0 ; i < cpi->twopass.frames_to_key ; i++)
-    {
-        double r;
+  for (i = 0; i < cpi->twopass.frames_to_key; i++) {
+    double r;
 
-        if (EOF == input_stats(cpi, &next_frame))
-            break;
+    if (EOF == input_stats(cpi, &next_frame)) break;
 
-        if (next_frame.intra_error > cpi->twopass.kf_intra_err_min)
-            r = (IIKFACTOR2 * next_frame.intra_error /
-                     DOUBLE_DIVIDE_CHECK(next_frame.coded_error));
-        else
-            r = (IIKFACTOR2 * cpi->twopass.kf_intra_err_min /
-                     DOUBLE_DIVIDE_CHECK(next_frame.coded_error));
+    if (next_frame.intra_error > cpi->twopass.kf_intra_err_min)
+      r = (IIKFACTOR2 * next_frame.intra_error /
+           DOUBLE_DIVIDE_CHECK(next_frame.coded_error));
+    else
+      r = (IIKFACTOR2 * cpi->twopass.kf_intra_err_min /
+           DOUBLE_DIVIDE_CHECK(next_frame.coded_error));
 
-        if (r > RMAX)
-            r = RMAX;
+    if (r > RMAX) r = RMAX;
 
-        /* How fast is prediction quality decaying */
-        loop_decay_rate = get_prediction_decay_rate(cpi, &next_frame);
+    /* How fast is prediction quality decaying */
+    loop_decay_rate = get_prediction_decay_rate(cpi, &next_frame);
 
-        decay_accumulator = decay_accumulator * loop_decay_rate;
-        decay_accumulator = decay_accumulator < 0.1 ? 0.1 : decay_accumulator;
+    decay_accumulator = decay_accumulator * loop_decay_rate;
+    decay_accumulator = decay_accumulator < 0.1 ? 0.1 : decay_accumulator;
 
-        boost_score += (decay_accumulator * r);
+    boost_score += (decay_accumulator * r);
 
-        if ((i > MIN_GF_INTERVAL) &&
-            ((boost_score - old_boost_score) < 1.0))
-        {
-            break;
-        }
-
-        old_boost_score = boost_score;
+    if ((i > MIN_GF_INTERVAL) && ((boost_score - old_boost_score) < 1.0)) {
+      break;
     }
 
-    if (1)
-    {
-        FIRSTPASS_STATS sectionstats;
-        double Ratio;
+    old_boost_score = boost_score;
+  }
 
-        zero_stats(&sectionstats);
-        reset_fpf_position(cpi, start_position);
+  if (1) {
+    FIRSTPASS_STATS sectionstats;
+    double Ratio;
 
-        for (i = 0 ; i < cpi->twopass.frames_to_key ; i++)
-        {
-            input_stats(cpi, &next_frame);
-            accumulate_stats(&sectionstats, &next_frame);
-        }
+    zero_stats(&sectionstats);
+    reset_fpf_position(cpi, start_position);
 
-        avg_stats(&sectionstats);
+    for (i = 0; i < cpi->twopass.frames_to_key; i++) {
+      input_stats(cpi, &next_frame);
+      accumulate_stats(&sectionstats, &next_frame);
+    }
 
-        cpi->twopass.section_intra_rating = (unsigned int)
-            (sectionstats.intra_error
-            / DOUBLE_DIVIDE_CHECK(sectionstats.coded_error));
+    avg_stats(&sectionstats);
 
-        Ratio = sectionstats.intra_error / DOUBLE_DIVIDE_CHECK(sectionstats.coded_error);
-        cpi->twopass.section_max_qfactor = 1.0 - ((Ratio - 10.0) * 0.025);
+    cpi->twopass.section_intra_rating =
+        (unsigned int)(sectionstats.intra_error /
+                       DOUBLE_DIVIDE_CHECK(sectionstats.coded_error));
 
-        if (cpi->twopass.section_max_qfactor < 0.80)
-            cpi->twopass.section_max_qfactor = 0.80;
-    }
+    Ratio = sectionstats.intra_error /
+            DOUBLE_DIVIDE_CHECK(sectionstats.coded_error);
+    cpi->twopass.section_max_qfactor = 1.0 - ((Ratio - 10.0) * 0.025);
 
-    /* When using CBR apply additional buffer fullness related upper limits */
-    if (cpi->oxcf.end_usage == USAGE_STREAM_FROM_SERVER)
-    {
-        double max_boost;
+    if (cpi->twopass.section_max_qfactor < 0.80)
+      cpi->twopass.section_max_qfactor = 0.80;
+  }
 
-        if (cpi->drop_frames_allowed)
-        {
-            int df_buffer_level = (int)(cpi->oxcf.drop_frames_water_mark
-                                  * (cpi->oxcf.optimal_buffer_level / 100));
+  /* When using CBR apply additional buffer fullness related upper limits */
+  if (cpi->oxcf.end_usage == USAGE_STREAM_FROM_SERVER) {
+    double max_boost;
 
-            if (cpi->buffer_level > df_buffer_level)
-                max_boost = ((double)((cpi->buffer_level - df_buffer_level) * 2 / 3) * 16.0) / DOUBLE_DIVIDE_CHECK((double)cpi->av_per_frame_bandwidth);
-            else
-                max_boost = 0.0;
-        }
-        else if (cpi->buffer_level > 0)
-        {
-            max_boost = ((double)(cpi->buffer_level * 2 / 3) * 16.0) / DOUBLE_DIVIDE_CHECK((double)cpi->av_per_frame_bandwidth);
-        }
-        else
-        {
-            max_boost = 0.0;
-        }
+    if (cpi->drop_frames_allowed) {
+      int df_buffer_level = (int)(cpi->oxcf.drop_frames_water_mark *
+                                  (cpi->oxcf.optimal_buffer_level / 100));
 
-        if (boost_score > max_boost)
-            boost_score = max_boost;
+      if (cpi->buffer_level > df_buffer_level)
+        max_boost =
+            ((double)((cpi->buffer_level - df_buffer_level) * 2 / 3) * 16.0) /
+            DOUBLE_DIVIDE_CHECK((double)cpi->av_per_frame_bandwidth);
+      else
+        max_boost = 0.0;
+    } else if (cpi->buffer_level > 0) {
+      max_boost = ((double)(cpi->buffer_level * 2 / 3) * 16.0) /
+                  DOUBLE_DIVIDE_CHECK((double)cpi->av_per_frame_bandwidth);
+    } else {
+      max_boost = 0.0;
     }
 
-    /* Reset the first pass file position */
-    reset_fpf_position(cpi, start_position);
+    if (boost_score > max_boost) boost_score = max_boost;
+  }
 
-    /* Work out how many bits to allocate for the key frame itself */
-    if (1)
-    {
-        int kf_boost = (int)boost_score;
-        int allocation_chunks;
-        int Counter = cpi->twopass.frames_to_key;
-        int alt_kf_bits;
-        YV12_BUFFER_CONFIG *lst_yv12 = &cpi->common.yv12_fb[cpi->common.lst_fb_idx];
-        /* Min boost based on kf interval */
+  /* Reset the first pass file position */
+  reset_fpf_position(cpi, start_position);
+
+  /* Work out how many bits to allocate for the key frame itself */
+  if (1) {
+    int kf_boost = (int)boost_score;
+    int allocation_chunks;
+    int Counter = cpi->twopass.frames_to_key;
+    int alt_kf_bits;
+    YV12_BUFFER_CONFIG *lst_yv12 = &cpi->common.yv12_fb[cpi->common.lst_fb_idx];
+/* Min boost based on kf interval */
 #if 0
 
         while ((kf_boost < 48) && (Counter > 0))
@@ -3082,291 +2821,285 @@
 
 #endif
 
-        if (kf_boost < 48)
-        {
-            kf_boost += ((Counter + 1) >> 1);
+    if (kf_boost < 48) {
+      kf_boost += ((Counter + 1) >> 1);
 
-            if (kf_boost > 48) kf_boost = 48;
-        }
+      if (kf_boost > 48) kf_boost = 48;
+    }
 
-        /* bigger frame sizes need larger kf boosts, smaller frames smaller
-         * boosts...
-         */
-        if ((lst_yv12->y_width * lst_yv12->y_height) > (320 * 240))
-            kf_boost += 2 * (lst_yv12->y_width * lst_yv12->y_height) / (320 * 240);
-        else if ((lst_yv12->y_width * lst_yv12->y_height) < (320 * 240))
-            kf_boost -= 4 * (320 * 240) / (lst_yv12->y_width * lst_yv12->y_height);
+    /* bigger frame sizes need larger kf boosts, smaller frames smaller
+     * boosts...
+     */
+    if ((lst_yv12->y_width * lst_yv12->y_height) > (320 * 240))
+      kf_boost += 2 * (lst_yv12->y_width * lst_yv12->y_height) / (320 * 240);
+    else if ((lst_yv12->y_width * lst_yv12->y_height) < (320 * 240))
+      kf_boost -= 4 * (320 * 240) / (lst_yv12->y_width * lst_yv12->y_height);
 
-        /* Min KF boost */
-        kf_boost = (int)((double)kf_boost * 100.0) >> 4; /* Scale 16 to 100 */
-        if (kf_boost < 250)
-            kf_boost = 250;
+    /* Min KF boost */
+    kf_boost = (int)((double)kf_boost * 100.0) >> 4; /* Scale 16 to 100 */
+    if (kf_boost < 250) kf_boost = 250;
 
-        /*
-         * We do three calculations for kf size.
-         * The first is based on the error score for the whole kf group.
-         * The second (optionaly) on the key frames own error if this is
-         * smaller than the average for the group.
-         * The final one insures that the frame receives at least the
-         * allocation it would have received based on its own error score vs
-         * the error score remaining
-         * Special case if the sequence appears almost totaly static
-         * as measured by the decay accumulator. In this case we want to
-         * spend almost all of the bits on the key frame.
-         * cpi->twopass.frames_to_key-1 because key frame itself is taken
-         * care of by kf_boost.
-         */
-        if ( decay_accumulator >= 0.99 )
-        {
-            allocation_chunks =
-                ((cpi->twopass.frames_to_key - 1) * 10) + kf_boost;
-        }
-        else
-        {
-            allocation_chunks =
-                ((cpi->twopass.frames_to_key - 1) * 100) + kf_boost;
-        }
+    /*
+     * We do three calculations for kf size.
+     * The first is based on the error score for the whole kf group.
+     * The second (optionaly) on the key frames own error if this is
+     * smaller than the average for the group.
+     * The final one insures that the frame receives at least the
+     * allocation it would have received based on its own error score vs
+     * the error score remaining
+     * Special case if the sequence appears almost totaly static
+     * as measured by the decay accumulator. In this case we want to
+     * spend almost all of the bits on the key frame.
+     * cpi->twopass.frames_to_key-1 because key frame itself is taken
+     * care of by kf_boost.
+     */
+    if (decay_accumulator >= 0.99) {
+      allocation_chunks = ((cpi->twopass.frames_to_key - 1) * 10) + kf_boost;
+    } else {
+      allocation_chunks = ((cpi->twopass.frames_to_key - 1) * 100) + kf_boost;
+    }
 
-        /* Normalize Altboost and allocations chunck down to prevent overflow */
-        while (kf_boost > 1000)
-        {
-            kf_boost /= 2;
-            allocation_chunks /= 2;
-        }
+    /* Normalize Altboost and allocations chunck down to prevent overflow */
+    while (kf_boost > 1000) {
+      kf_boost /= 2;
+      allocation_chunks /= 2;
+    }
 
-        cpi->twopass.kf_group_bits = (cpi->twopass.kf_group_bits < 0) ? 0 : cpi->twopass.kf_group_bits;
+    cpi->twopass.kf_group_bits =
+        (cpi->twopass.kf_group_bits < 0) ? 0 : cpi->twopass.kf_group_bits;
 
-        /* Calculate the number of bits to be spent on the key frame */
-        cpi->twopass.kf_bits  = (int)((double)kf_boost * ((double)cpi->twopass.kf_group_bits / (double)allocation_chunks));
+    /* Calculate the number of bits to be spent on the key frame */
+    cpi->twopass.kf_bits =
+        (int)((double)kf_boost *
+              ((double)cpi->twopass.kf_group_bits / (double)allocation_chunks));
 
-        /* Apply an additional limit for CBR */
-        if (cpi->oxcf.end_usage == USAGE_STREAM_FROM_SERVER)
-        {
-            if (cpi->twopass.kf_bits > (int)((3 * cpi->buffer_level) >> 2))
-                cpi->twopass.kf_bits = (int)((3 * cpi->buffer_level) >> 2);
-        }
+    /* Apply an additional limit for CBR */
+    if (cpi->oxcf.end_usage == USAGE_STREAM_FROM_SERVER) {
+      if (cpi->twopass.kf_bits > (int)((3 * cpi->buffer_level) >> 2))
+        cpi->twopass.kf_bits = (int)((3 * cpi->buffer_level) >> 2);
+    }
 
-        /* If the key frame is actually easier than the average for the
-         * kf group (which does sometimes happen... eg a blank intro frame)
-         * Then use an alternate calculation based on the kf error score
-         * which should give a smaller key frame.
-         */
-        if (kf_mod_err < kf_group_err / cpi->twopass.frames_to_key)
-        {
-            double  alt_kf_grp_bits =
-                        ((double)cpi->twopass.bits_left *
-                         (kf_mod_err * (double)cpi->twopass.frames_to_key) /
-                         DOUBLE_DIVIDE_CHECK(cpi->twopass.modified_error_left));
+    /* If the key frame is actually easier than the average for the
+     * kf group (which does sometimes happen... eg a blank intro frame)
+     * Then use an alternate calculation based on the kf error score
+     * which should give a smaller key frame.
+     */
+    if (kf_mod_err < kf_group_err / cpi->twopass.frames_to_key) {
+      double alt_kf_grp_bits =
+          ((double)cpi->twopass.bits_left *
+           (kf_mod_err * (double)cpi->twopass.frames_to_key) /
+           DOUBLE_DIVIDE_CHECK(cpi->twopass.modified_error_left));
 
-            alt_kf_bits = (int)((double)kf_boost *
-                                (alt_kf_grp_bits / (double)allocation_chunks));
+      alt_kf_bits = (int)((double)kf_boost *
+                          (alt_kf_grp_bits / (double)allocation_chunks));
 
-            if (cpi->twopass.kf_bits > alt_kf_bits)
-            {
-                cpi->twopass.kf_bits = alt_kf_bits;
-            }
-        }
-        /* Else if it is much harder than other frames in the group make sure
-         * it at least receives an allocation in keeping with its relative
-         * error score
-         */
-        else
-        {
-            alt_kf_bits =
-                (int)((double)cpi->twopass.bits_left *
-                      (kf_mod_err /
-                       DOUBLE_DIVIDE_CHECK(cpi->twopass.modified_error_left)));
+      if (cpi->twopass.kf_bits > alt_kf_bits) {
+        cpi->twopass.kf_bits = alt_kf_bits;
+      }
+    }
+    /* Else if it is much harder than other frames in the group make sure
+     * it at least receives an allocation in keeping with its relative
+     * error score
+     */
+    else {
+      alt_kf_bits = (int)((double)cpi->twopass.bits_left *
+                          (kf_mod_err / DOUBLE_DIVIDE_CHECK(
+                                            cpi->twopass.modified_error_left)));
 
-            if (alt_kf_bits > cpi->twopass.kf_bits)
-            {
-                cpi->twopass.kf_bits = alt_kf_bits;
-            }
-        }
+      if (alt_kf_bits > cpi->twopass.kf_bits) {
+        cpi->twopass.kf_bits = alt_kf_bits;
+      }
+    }
 
-        cpi->twopass.kf_group_bits -= cpi->twopass.kf_bits;
-        /* Add in the minimum frame allowance */
-        cpi->twopass.kf_bits += cpi->min_frame_bandwidth;
+    cpi->twopass.kf_group_bits -= cpi->twopass.kf_bits;
+    /* Add in the minimum frame allowance */
+    cpi->twopass.kf_bits += cpi->min_frame_bandwidth;
 
-        /* Peer frame bit target for this frame */
-        cpi->per_frame_bandwidth = cpi->twopass.kf_bits;
+    /* Peer frame bit target for this frame */
+    cpi->per_frame_bandwidth = cpi->twopass.kf_bits;
 
-        /* Convert to a per second bitrate */
-        cpi->target_bandwidth = (int)(cpi->twopass.kf_bits *
-                                      cpi->output_framerate);
-    }
+    /* Convert to a per second bitrate */
+    cpi->target_bandwidth = (int)(cpi->twopass.kf_bits * cpi->output_framerate);
+  }
 
-    /* Note the total error score of the kf group minus the key frame itself */
-    cpi->twopass.kf_group_error_left = (int)(kf_group_err - kf_mod_err);
+  /* Note the total error score of the kf group minus the key frame itself */
+  cpi->twopass.kf_group_error_left = (int)(kf_group_err - kf_mod_err);
 
-    /* Adjust the count of total modified error left. The count of bits left
-     * is adjusted elsewhere based on real coded frame sizes
-     */
-    cpi->twopass.modified_error_left -= kf_group_err;
+  /* Adjust the count of total modified error left. The count of bits left
+   * is adjusted elsewhere based on real coded frame sizes
+   */
+  cpi->twopass.modified_error_left -= kf_group_err;
 
-    if (cpi->oxcf.allow_spatial_resampling)
-    {
-        int resample_trigger = 0;
-        int last_kf_resampled = 0;
-        int kf_q;
-        int scale_val = 0;
-        int hr, hs, vr, vs;
-        int new_width = cpi->oxcf.Width;
-        int new_height = cpi->oxcf.Height;
+  if (cpi->oxcf.allow_spatial_resampling) {
+    int resample_trigger = 0;
+    int last_kf_resampled = 0;
+    int kf_q;
+    int scale_val = 0;
+    int hr, hs, vr, vs;
+    int new_width = cpi->oxcf.Width;
+    int new_height = cpi->oxcf.Height;
 
-        int projected_buffer_level;
-        int tmp_q;
+    int projected_buffer_level;
+    int tmp_q;
 
-        double projected_bits_perframe;
-        double group_iiratio = (kf_group_intra_err - first_frame.intra_error) / (kf_group_coded_err - first_frame.coded_error);
-        double err_per_frame = kf_group_err / cpi->twopass.frames_to_key;
-        double bits_per_frame;
-        double av_bits_per_frame;
-        double effective_size_ratio;
+    double projected_bits_perframe;
+    double group_iiratio = (kf_group_intra_err - first_frame.intra_error) /
+                           (kf_group_coded_err - first_frame.coded_error);
+    double err_per_frame = kf_group_err / cpi->twopass.frames_to_key;
+    double bits_per_frame;
+    double av_bits_per_frame;
+    double effective_size_ratio;
 
-        if ((cpi->common.Width != cpi->oxcf.Width) || (cpi->common.Height != cpi->oxcf.Height))
-            last_kf_resampled = 1;
+    if ((cpi->common.Width != cpi->oxcf.Width) ||
+        (cpi->common.Height != cpi->oxcf.Height))
+      last_kf_resampled = 1;
 
-        /* Set back to unscaled by defaults */
-        cpi->common.horiz_scale = NORMAL;
-        cpi->common.vert_scale = NORMAL;
+    /* Set back to unscaled by defaults */
+    cpi->common.horiz_scale = NORMAL;
+    cpi->common.vert_scale = NORMAL;
 
-        /* Calculate Average bits per frame. */
-        av_bits_per_frame = cpi->oxcf.target_bandwidth / DOUBLE_DIVIDE_CHECK((double)cpi->framerate);
+    /* Calculate Average bits per frame. */
+    av_bits_per_frame = cpi->oxcf.target_bandwidth /
+                        DOUBLE_DIVIDE_CHECK((double)cpi->framerate);
 
-        /* CBR... Use the clip average as the target for deciding resample */
-        if (cpi->oxcf.end_usage == USAGE_STREAM_FROM_SERVER)
-        {
-            bits_per_frame = av_bits_per_frame;
-        }
+    /* CBR... Use the clip average as the target for deciding resample */
+    if (cpi->oxcf.end_usage == USAGE_STREAM_FROM_SERVER) {
+      bits_per_frame = av_bits_per_frame;
+    }
 
-        /* In VBR we want to avoid downsampling in easy section unless we
-         * are under extreme pressure So use the larger of target bitrate
-         * for this section or average bitrate for sequence
-         */
-        else
-        {
-            /* This accounts for how hard the section is... */
-            bits_per_frame = (double)
-                (cpi->twopass.kf_group_bits / cpi->twopass.frames_to_key);
+    /* In VBR we want to avoid downsampling in easy section unless we
+     * are under extreme pressure So use the larger of target bitrate
+     * for this section or average bitrate for sequence
+     */
+    else {
+      /* This accounts for how hard the section is... */
+      bits_per_frame =
+          (double)(cpi->twopass.kf_group_bits / cpi->twopass.frames_to_key);
 
-            /* Dont turn to resampling in easy sections just because they
-             * have been assigned a small number of bits
-             */
-            if (bits_per_frame < av_bits_per_frame)
-                bits_per_frame = av_bits_per_frame;
-        }
+      /* Dont turn to resampling in easy sections just because they
+       * have been assigned a small number of bits
+       */
+      if (bits_per_frame < av_bits_per_frame)
+        bits_per_frame = av_bits_per_frame;
+    }
 
-        /* bits_per_frame should comply with our minimum */
-        if (bits_per_frame < (cpi->oxcf.target_bandwidth * cpi->oxcf.two_pass_vbrmin_section / 100))
-            bits_per_frame = (cpi->oxcf.target_bandwidth * cpi->oxcf.two_pass_vbrmin_section / 100);
+    /* bits_per_frame should comply with our minimum */
+    if (bits_per_frame <
+        (cpi->oxcf.target_bandwidth * cpi->oxcf.two_pass_vbrmin_section / 100))
+      bits_per_frame = (cpi->oxcf.target_bandwidth *
+                        cpi->oxcf.two_pass_vbrmin_section / 100);
 
-        /* Work out if spatial resampling is necessary */
-        kf_q = estimate_kf_group_q(cpi, err_per_frame,
-                                  (int)bits_per_frame, group_iiratio);
+    /* Work out if spatial resampling is necessary */
+    kf_q = estimate_kf_group_q(cpi, err_per_frame, (int)bits_per_frame,
+                               group_iiratio);
 
-        /* If we project a required Q higher than the maximum allowed Q then
-         * make a guess at the actual size of frames in this section
-         */
-        projected_bits_perframe = bits_per_frame;
-        tmp_q = kf_q;
+    /* If we project a required Q higher than the maximum allowed Q then
+     * make a guess at the actual size of frames in this section
+     */
+    projected_bits_perframe = bits_per_frame;
+    tmp_q = kf_q;
 
-        while (tmp_q > cpi->worst_quality)
-        {
-            projected_bits_perframe *= 1.04;
-            tmp_q--;
-        }
+    while (tmp_q > cpi->worst_quality) {
+      projected_bits_perframe *= 1.04;
+      tmp_q--;
+    }
 
-        /* Guess at buffer level at the end of the section */
-        projected_buffer_level = (int)
-                    (cpi->buffer_level - (int)
-                    ((projected_bits_perframe - av_bits_per_frame) *
+    /* Guess at buffer level at the end of the section */
+    projected_buffer_level =
+        (int)(cpi->buffer_level -
+              (int)((projected_bits_perframe - av_bits_per_frame) *
                     cpi->twopass.frames_to_key));
 
-        if (0)
-        {
-            FILE *f = fopen("Subsamle.stt", "a");
-            fprintf(f, " %8d %8d %8d %8d %12.0f %8d %8d %8d\n",  cpi->common.current_video_frame, kf_q, cpi->common.horiz_scale, cpi->common.vert_scale,  kf_group_err / cpi->twopass.frames_to_key, (int)(cpi->twopass.kf_group_bits / cpi->twopass.frames_to_key), new_height, new_width);
-            fclose(f);
-        }
+    if (0) {
+      FILE *f = fopen("Subsamle.stt", "a");
+      fprintf(f, " %8d %8d %8d %8d %12.0f %8d %8d %8d\n",
+              cpi->common.current_video_frame, kf_q, cpi->common.horiz_scale,
+              cpi->common.vert_scale, kf_group_err / cpi->twopass.frames_to_key,
+              (int)(cpi->twopass.kf_group_bits / cpi->twopass.frames_to_key),
+              new_height, new_width);
+      fclose(f);
+    }
 
-        /* The trigger for spatial resampling depends on the various
-         * parameters such as whether we are streaming (CBR) or VBR.
-         */
-        if (cpi->oxcf.end_usage == USAGE_STREAM_FROM_SERVER)
-        {
-            /* Trigger resample if we are projected to fall below down
-             * sample level or resampled last time and are projected to
-             * remain below the up sample level
-             */
-            if ((projected_buffer_level < (cpi->oxcf.resample_down_water_mark * cpi->oxcf.optimal_buffer_level / 100)) ||
-                (last_kf_resampled && (projected_buffer_level < (cpi->oxcf.resample_up_water_mark * cpi->oxcf.optimal_buffer_level / 100))))
-                resample_trigger = 1;
-            else
-                resample_trigger = 0;
-        }
-        else
-        {
-            int64_t clip_bits = (int64_t)(cpi->twopass.total_stats.count * cpi->oxcf.target_bandwidth / DOUBLE_DIVIDE_CHECK((double)cpi->framerate));
-            int64_t over_spend = cpi->oxcf.starting_buffer_level - cpi->buffer_level;
+    /* The trigger for spatial resampling depends on the various
+     * parameters such as whether we are streaming (CBR) or VBR.
+     */
+    if (cpi->oxcf.end_usage == USAGE_STREAM_FROM_SERVER) {
+      /* Trigger resample if we are projected to fall below down
+       * sample level or resampled last time and are projected to
+       * remain below the up sample level
+       */
+      if ((projected_buffer_level < (cpi->oxcf.resample_down_water_mark *
+                                     cpi->oxcf.optimal_buffer_level / 100)) ||
+          (last_kf_resampled &&
+           (projected_buffer_level < (cpi->oxcf.resample_up_water_mark *
+                                      cpi->oxcf.optimal_buffer_level / 100))))
+        resample_trigger = 1;
+      else
+        resample_trigger = 0;
+    } else {
+      int64_t clip_bits = (int64_t)(
+          cpi->twopass.total_stats.count * cpi->oxcf.target_bandwidth /
+          DOUBLE_DIVIDE_CHECK((double)cpi->framerate));
+      int64_t over_spend = cpi->oxcf.starting_buffer_level - cpi->buffer_level;
 
-            /* If triggered last time the threshold for triggering again is
-             * reduced:
-             *
-             * Projected Q higher than allowed and Overspend > 5% of total
-             * bits
-             */
-            if ((last_kf_resampled && (kf_q > cpi->worst_quality)) ||
-                ((kf_q > cpi->worst_quality) &&
-                 (over_spend > clip_bits / 20)))
-                resample_trigger = 1;
-            else
-                resample_trigger = 0;
+      /* If triggered last time the threshold for triggering again is
+       * reduced:
+       *
+       * Projected Q higher than allowed and Overspend > 5% of total
+       * bits
+       */
+      if ((last_kf_resampled && (kf_q > cpi->worst_quality)) ||
+          ((kf_q > cpi->worst_quality) && (over_spend > clip_bits / 20)))
+        resample_trigger = 1;
+      else
+        resample_trigger = 0;
+    }
 
-        }
+    if (resample_trigger) {
+      while ((kf_q >= cpi->worst_quality) && (scale_val < 6)) {
+        scale_val++;
 
-        if (resample_trigger)
-        {
-            while ((kf_q >= cpi->worst_quality) && (scale_val < 6))
-            {
-                scale_val ++;
+        cpi->common.vert_scale = vscale_lookup[scale_val];
+        cpi->common.horiz_scale = hscale_lookup[scale_val];
 
-                cpi->common.vert_scale   = vscale_lookup[scale_val];
-                cpi->common.horiz_scale  = hscale_lookup[scale_val];
+        Scale2Ratio(cpi->common.horiz_scale, &hr, &hs);
+        Scale2Ratio(cpi->common.vert_scale, &vr, &vs);
 
-                Scale2Ratio(cpi->common.horiz_scale, &hr, &hs);
-                Scale2Ratio(cpi->common.vert_scale, &vr, &vs);
+        new_width = ((hs - 1) + (cpi->oxcf.Width * hr)) / hs;
+        new_height = ((vs - 1) + (cpi->oxcf.Height * vr)) / vs;
 
-                new_width = ((hs - 1) + (cpi->oxcf.Width * hr)) / hs;
-                new_height = ((vs - 1) + (cpi->oxcf.Height * vr)) / vs;
+        /* Reducing the area to 1/4 does not reduce the complexity
+         * (err_per_frame) to 1/4... effective_sizeratio attempts
+         * to provide a crude correction for this
+         */
+        effective_size_ratio = (double)(new_width * new_height) /
+                               (double)(cpi->oxcf.Width * cpi->oxcf.Height);
+        effective_size_ratio = (1.0 + (3.0 * effective_size_ratio)) / 4.0;
 
-                /* Reducing the area to 1/4 does not reduce the complexity
-                 * (err_per_frame) to 1/4... effective_sizeratio attempts
-                 * to provide a crude correction for this
-                 */
-                effective_size_ratio = (double)(new_width * new_height) / (double)(cpi->oxcf.Width * cpi->oxcf.Height);
-                effective_size_ratio = (1.0 + (3.0 * effective_size_ratio)) / 4.0;
+        /* Now try again and see what Q we get with the smaller
+         * image size
+         */
+        kf_q = estimate_kf_group_q(cpi, err_per_frame * effective_size_ratio,
+                                   (int)bits_per_frame, group_iiratio);
 
-                /* Now try again and see what Q we get with the smaller
-                 * image size
-                 */
-                kf_q = estimate_kf_group_q(cpi,
-                                          err_per_frame * effective_size_ratio,
-                                          (int)bits_per_frame, group_iiratio);
-
-                if (0)
-                {
-                    FILE *f = fopen("Subsamle.stt", "a");
-                    fprintf(f, "******** %8d %8d %8d %12.0f %8d %8d %8d\n",  kf_q, cpi->common.horiz_scale, cpi->common.vert_scale,  kf_group_err / cpi->twopass.frames_to_key, (int)(cpi->twopass.kf_group_bits / cpi->twopass.frames_to_key), new_height, new_width);
-                    fclose(f);
-                }
-            }
+        if (0) {
+          FILE *f = fopen("Subsamle.stt", "a");
+          fprintf(
+              f, "******** %8d %8d %8d %12.0f %8d %8d %8d\n", kf_q,
+              cpi->common.horiz_scale, cpi->common.vert_scale,
+              kf_group_err / cpi->twopass.frames_to_key,
+              (int)(cpi->twopass.kf_group_bits / cpi->twopass.frames_to_key),
+              new_height, new_width);
+          fclose(f);
         }
+      }
+    }
 
-        if ((cpi->common.Width != new_width) || (cpi->common.Height != new_height))
-        {
-            cpi->common.Width = new_width;
-            cpi->common.Height = new_height;
-            vp8_alloc_compressor_data(cpi);
-        }
+    if ((cpi->common.Width != new_width) ||
+        (cpi->common.Height != new_height)) {
+      cpi->common.Width = new_width;
+      cpi->common.Height = new_height;
+      vp8_alloc_compressor_data(cpi);
     }
+  }
 }
--- a/vp8/encoder/firstpass.h
+++ b/vp8/encoder/firstpass.h
@@ -8,7 +8,6 @@
  *  be found in the AUTHORS file in the root of the source tree.
  */
 
-
 #ifndef VP8_ENCODER_FIRSTPASS_H_
 #define VP8_ENCODER_FIRSTPASS_H_
 
--- a/vp8/encoder/lookahead.c
+++ b/vp8/encoder/lookahead.c
@@ -13,219 +13,167 @@
 #include "lookahead.h"
 #include "vp8/common/extend.h"
 
-#define MAX_LAG_BUFFERS (CONFIG_REALTIME_ONLY? 1 : 25)
+#define MAX_LAG_BUFFERS (CONFIG_REALTIME_ONLY ? 1 : 25)
 
-struct lookahead_ctx
-{
-    unsigned int max_sz;         /* Absolute size of the queue */
-    unsigned int sz;             /* Number of buffers currently in the queue */
-    unsigned int read_idx;       /* Read index */
-    unsigned int write_idx;      /* Write index */
-    struct lookahead_entry *buf; /* Buffer list */
+struct lookahead_ctx {
+  unsigned int max_sz;         /* Absolute size of the queue */
+  unsigned int sz;             /* Number of buffers currently in the queue */
+  unsigned int read_idx;       /* Read index */
+  unsigned int write_idx;      /* Write index */
+  struct lookahead_entry *buf; /* Buffer list */
 };
 
-
 /* Return the buffer at the given absolute index and increment the index */
-static struct lookahead_entry *
-pop(struct lookahead_ctx *ctx,
-    unsigned int         *idx)
-{
-    unsigned int            index = *idx;
-    struct lookahead_entry *buf = ctx->buf + index;
+static struct lookahead_entry *pop(struct lookahead_ctx *ctx,
+                                   unsigned int *idx) {
+  unsigned int index = *idx;
+  struct lookahead_entry *buf = ctx->buf + index;
 
-    assert(index < ctx->max_sz);
-    if(++index >= ctx->max_sz)
-        index -= ctx->max_sz;
-    *idx = index;
-    return buf;
+  assert(index < ctx->max_sz);
+  if (++index >= ctx->max_sz) index -= ctx->max_sz;
+  *idx = index;
+  return buf;
 }
 
+void vp8_lookahead_destroy(struct lookahead_ctx *ctx) {
+  if (ctx) {
+    if (ctx->buf) {
+      unsigned int i;
 
-void
-vp8_lookahead_destroy(struct lookahead_ctx *ctx)
-{
-    if(ctx)
-    {
-        if(ctx->buf)
-        {
-            unsigned int i;
-
-            for(i = 0; i < ctx->max_sz; i++)
-                vp8_yv12_de_alloc_frame_buffer(&ctx->buf[i].img);
-            free(ctx->buf);
-        }
-        free(ctx);
+      for (i = 0; i < ctx->max_sz; i++)
+        vp8_yv12_de_alloc_frame_buffer(&ctx->buf[i].img);
+      free(ctx->buf);
     }
+    free(ctx);
+  }
 }
 
+struct lookahead_ctx *vp8_lookahead_init(unsigned int width,
+                                         unsigned int height,
+                                         unsigned int depth) {
+  struct lookahead_ctx *ctx = NULL;
+  unsigned int i;
 
-struct lookahead_ctx*
-vp8_lookahead_init(unsigned int width,
-                   unsigned int height,
-                   unsigned int depth)
-{
-    struct lookahead_ctx *ctx = NULL;
-    unsigned int i;
+  /* Clamp the lookahead queue depth */
+  if (depth < 1)
+    depth = 1;
+  else if (depth > MAX_LAG_BUFFERS)
+    depth = MAX_LAG_BUFFERS;
 
-    /* Clamp the lookahead queue depth */
-    if(depth < 1)
-        depth = 1;
-    else if(depth > MAX_LAG_BUFFERS)
-        depth = MAX_LAG_BUFFERS;
+  /* Keep last frame in lookahead buffer by increasing depth by 1.*/
+  depth += 1;
 
-    /* Keep last frame in lookahead buffer by increasing depth by 1.*/
-    depth += 1;
+  /* Align the buffer dimensions */
+  width = (width + 15) & ~15;
+  height = (height + 15) & ~15;
 
-    /* Align the buffer dimensions */
-    width = (width + 15) & ~15;
-    height = (height + 15) & ~15;
-
-    /* Allocate the lookahead structures */
-    ctx = calloc(1, sizeof(*ctx));
-    if(ctx)
-    {
-        ctx->max_sz = depth;
-        ctx->buf = calloc(depth, sizeof(*ctx->buf));
-        if(!ctx->buf)
-            goto bail;
-        for(i=0; i<depth; i++)
-            if (vp8_yv12_alloc_frame_buffer(&ctx->buf[i].img,
-                                            width, height, VP8BORDERINPIXELS))
-                goto bail;
-    }
-    return ctx;
+  /* Allocate the lookahead structures */
+  ctx = calloc(1, sizeof(*ctx));
+  if (ctx) {
+    ctx->max_sz = depth;
+    ctx->buf = calloc(depth, sizeof(*ctx->buf));
+    if (!ctx->buf) goto bail;
+    for (i = 0; i < depth; i++)
+      if (vp8_yv12_alloc_frame_buffer(&ctx->buf[i].img, width, height,
+                                      VP8BORDERINPIXELS))
+        goto bail;
+  }
+  return ctx;
 bail:
-    vp8_lookahead_destroy(ctx);
-    return NULL;
+  vp8_lookahead_destroy(ctx);
+  return NULL;
 }
 
+int vp8_lookahead_push(struct lookahead_ctx *ctx, YV12_BUFFER_CONFIG *src,
+                       int64_t ts_start, int64_t ts_end, unsigned int flags,
+                       unsigned char *active_map) {
+  struct lookahead_entry *buf;
+  int row, col, active_end;
+  int mb_rows = (src->y_height + 15) >> 4;
+  int mb_cols = (src->y_width + 15) >> 4;
 
-int
-vp8_lookahead_push(struct lookahead_ctx *ctx,
-                   YV12_BUFFER_CONFIG   *src,
-                   int64_t               ts_start,
-                   int64_t               ts_end,
-                   unsigned int          flags,
-                   unsigned char        *active_map)
-{
-    struct lookahead_entry* buf;
-    int row, col, active_end;
-    int mb_rows = (src->y_height + 15) >> 4;
-    int mb_cols = (src->y_width + 15) >> 4;
+  if (ctx->sz + 2 > ctx->max_sz) return 1;
+  ctx->sz++;
+  buf = pop(ctx, &ctx->write_idx);
 
-    if(ctx->sz + 2 > ctx->max_sz)
-        return 1;
-    ctx->sz++;
-    buf = pop(ctx, &ctx->write_idx);
+  /* Only do this partial copy if the following conditions are all met:
+   * 1. Lookahead queue has has size of 1.
+   * 2. Active map is provided.
+   * 3. This is not a key frame, golden nor altref frame.
+   */
+  if (ctx->max_sz == 1 && active_map && !flags) {
+    for (row = 0; row < mb_rows; ++row) {
+      col = 0;
 
-    /* Only do this partial copy if the following conditions are all met:
-     * 1. Lookahead queue has has size of 1.
-     * 2. Active map is provided.
-     * 3. This is not a key frame, golden nor altref frame.
-     */
-    if (ctx->max_sz == 1 && active_map && !flags)
-    {
-        for (row = 0; row < mb_rows; ++row)
-        {
-            col = 0;
+      while (1) {
+        /* Find the first active macroblock in this row. */
+        for (; col < mb_cols; ++col) {
+          if (active_map[col]) break;
+        }
 
-            while (1)
-            {
-                /* Find the first active macroblock in this row. */
-                for (; col < mb_cols; ++col)
-                {
-                    if (active_map[col])
-                        break;
-                }
+        /* No more active macroblock in this row. */
+        if (col == mb_cols) break;
 
-                /* No more active macroblock in this row. */
-                if (col == mb_cols)
-                    break;
+        /* Find the end of active region in this row. */
+        active_end = col;
 
-                /* Find the end of active region in this row. */
-                active_end = col;
+        for (; active_end < mb_cols; ++active_end) {
+          if (!active_map[active_end]) break;
+        }
 
-                for (; active_end < mb_cols; ++active_end)
-                {
-                    if (!active_map[active_end])
-                        break;
-                }
+        /* Only copy this active region. */
+        vp8_copy_and_extend_frame_with_rect(src, &buf->img, row << 4, col << 4,
+                                            16, (active_end - col) << 4);
 
-                /* Only copy this active region. */
-                vp8_copy_and_extend_frame_with_rect(src, &buf->img,
-                                                    row << 4,
-                                                    col << 4, 16,
-                                                    (active_end - col) << 4);
+        /* Start again from the end of this active region. */
+        col = active_end;
+      }
 
-                /* Start again from the end of this active region. */
-                col = active_end;
-            }
-
-            active_map += mb_cols;
-        }
+      active_map += mb_cols;
     }
-    else
-    {
-        vp8_copy_and_extend_frame(src, &buf->img);
-    }
-    buf->ts_start = ts_start;
-    buf->ts_end = ts_end;
-    buf->flags = flags;
-    return 0;
+  } else {
+    vp8_copy_and_extend_frame(src, &buf->img);
+  }
+  buf->ts_start = ts_start;
+  buf->ts_end = ts_end;
+  buf->flags = flags;
+  return 0;
 }
 
+struct lookahead_entry *vp8_lookahead_pop(struct lookahead_ctx *ctx,
+                                          int drain) {
+  struct lookahead_entry *buf = NULL;
 
-struct lookahead_entry*
-vp8_lookahead_pop(struct lookahead_ctx *ctx,
-                  int                   drain)
-{
-    struct lookahead_entry* buf = NULL;
-
-    assert(ctx != NULL);
-    if(ctx->sz && (drain || ctx->sz == ctx->max_sz - 1))
-    {
-        buf = pop(ctx, &ctx->read_idx);
-        ctx->sz--;
-    }
-    return buf;
+  assert(ctx != NULL);
+  if (ctx->sz && (drain || ctx->sz == ctx->max_sz - 1)) {
+    buf = pop(ctx, &ctx->read_idx);
+    ctx->sz--;
+  }
+  return buf;
 }
 
+struct lookahead_entry *vp8_lookahead_peek(struct lookahead_ctx *ctx,
+                                           unsigned int index, int direction) {
+  struct lookahead_entry *buf = NULL;
 
-struct lookahead_entry*
-vp8_lookahead_peek(struct lookahead_ctx *ctx,
-                   unsigned int          index,
-                   int                   direction)
-{
-    struct lookahead_entry* buf = NULL;
-
-    if (direction == PEEK_FORWARD)
-    {
-        assert(index < ctx->max_sz - 1);
-        if(index < ctx->sz)
-        {
-            index += ctx->read_idx;
-            if(index >= ctx->max_sz)
-                index -= ctx->max_sz;
-            buf = ctx->buf + index;
-        }
+  if (direction == PEEK_FORWARD) {
+    assert(index < ctx->max_sz - 1);
+    if (index < ctx->sz) {
+      index += ctx->read_idx;
+      if (index >= ctx->max_sz) index -= ctx->max_sz;
+      buf = ctx->buf + index;
     }
-    else if (direction == PEEK_BACKWARD)
-    {
-        assert(index == 1);
+  } else if (direction == PEEK_BACKWARD) {
+    assert(index == 1);
 
-        if(ctx->read_idx == 0)
-            index = ctx->max_sz - 1;
-        else
-            index = ctx->read_idx - index;
-        buf = ctx->buf + index;
-    }
+    if (ctx->read_idx == 0)
+      index = ctx->max_sz - 1;
+    else
+      index = ctx->read_idx - index;
+    buf = ctx->buf + index;
+  }
 
-    return buf;
+  return buf;
 }
 
-
-unsigned int
-vp8_lookahead_depth(struct lookahead_ctx *ctx)
-{
-    return ctx->sz;
-}
+unsigned int vp8_lookahead_depth(struct lookahead_ctx *ctx) { return ctx->sz; }
--- a/vp8/encoder/lookahead.h
+++ b/vp8/encoder/lookahead.h
@@ -16,15 +16,13 @@
 extern "C" {
 #endif
 
-struct lookahead_entry
-{
-    YV12_BUFFER_CONFIG  img;
-    int64_t             ts_start;
-    int64_t             ts_end;
-    unsigned int        flags;
+struct lookahead_entry {
+  YV12_BUFFER_CONFIG img;
+  int64_t ts_start;
+  int64_t ts_end;
+  unsigned int flags;
 };
 
-
 struct lookahead_ctx;
 
 /**\brief Initializes the lookahead stage
@@ -34,18 +32,15 @@
  *
  *
  */
-struct lookahead_ctx* vp8_lookahead_init(unsigned int width,
+struct lookahead_ctx *vp8_lookahead_init(unsigned int width,
                                          unsigned int height,
-                                         unsigned int depth
-                                         );
+                                         unsigned int depth);
 
-
 /**\brief Destroys the lookahead stage
  *
  */
 void vp8_lookahead_destroy(struct lookahead_ctx *ctx);
 
-
 /**\brief Enqueue a source buffer
  *
  * This function will copy the source image into a new framebuffer with
@@ -61,15 +56,10 @@
  * \param[in] flags       Flags set on this frame
  * \param[in] active_map  Map that specifies which macroblock is active
  */
-int
-vp8_lookahead_push(struct lookahead_ctx *ctx,
-                   YV12_BUFFER_CONFIG   *src,
-                   int64_t               ts_start,
-                   int64_t               ts_end,
-                   unsigned int          flags,
-                   unsigned char        *active_map);
+int vp8_lookahead_push(struct lookahead_ctx *ctx, YV12_BUFFER_CONFIG *src,
+                       int64_t ts_start, int64_t ts_end, unsigned int flags,
+                       unsigned char *active_map);
 
-
 /**\brief Get the next source buffer to encode
  *
  *
@@ -81,12 +71,9 @@
  * \retval NULL, if drain not set and queue not of the configured depth
  *
  */
-struct lookahead_entry*
-vp8_lookahead_pop(struct lookahead_ctx *ctx,
-                  int                   drain);
+struct lookahead_entry *vp8_lookahead_pop(struct lookahead_ctx *ctx, int drain);
 
-
-#define PEEK_FORWARD   1
+#define PEEK_FORWARD 1
 #define PEEK_BACKWARD -1
 /**\brief Get a future source buffer to encode
  *
@@ -96,19 +83,14 @@
  * \retval NULL, if no buffer exists at the specified index
  *
  */
-struct lookahead_entry*
-vp8_lookahead_peek(struct lookahead_ctx *ctx,
-                   unsigned int          index,
-                   int                   direction);
+struct lookahead_entry *vp8_lookahead_peek(struct lookahead_ctx *ctx,
+                                           unsigned int index, int direction);
 
-
 /**\brief Get the number of frames currently in the lookahead queue
  *
  * \param[in] ctx       Pointer to the lookahead context
  */
-unsigned int
-vp8_lookahead_depth(struct lookahead_ctx *ctx);
-
+unsigned int vp8_lookahead_depth(struct lookahead_ctx *ctx);
 
 #ifdef __cplusplus
 }  // extern "C"
--- a/vp8/encoder/mcomp.c
+++ b/vp8/encoder/mcomp.c
@@ -8,7 +8,6 @@
  *  be found in the AUTHORS file in the root of the source tree.
  */
 
-
 #include "./vp8_rtcd.h"
 #include "./vpx_dsp_rtcd.h"
 #include "onyx_int.h"
@@ -23,161 +22,159 @@
 #include "vpx_dsp/vpx_dsp_common.h"
 
 #ifdef VP8_ENTROPY_STATS
-static int mv_ref_ct [31] [4] [2];
-static int mv_mode_cts [4] [2];
+static int mv_ref_ct[31][4][2];
+static int mv_mode_cts[4][2];
 #endif
 
-int vp8_mv_bit_cost(int_mv *mv, int_mv *ref, int *mvcost[2], int Weight)
-{
-    /* MV costing is based on the distribution of vectors in the previous
-     * frame and as such will tend to over state the cost of vectors. In
-     * addition coding a new vector can have a knock on effect on the cost
-     * of subsequent vectors and the quality of prediction from NEAR and
-     * NEAREST for subsequent blocks. The "Weight" parameter allows, to a
-     * limited extent, for some account to be taken of these factors.
-     */
-    return ((mvcost[0][(mv->as_mv.row - ref->as_mv.row) >> 1] + mvcost[1][(mv->as_mv.col - ref->as_mv.col) >> 1]) * Weight) >> 7;
+int vp8_mv_bit_cost(int_mv *mv, int_mv *ref, int *mvcost[2], int Weight) {
+  /* MV costing is based on the distribution of vectors in the previous
+   * frame and as such will tend to over state the cost of vectors. In
+   * addition coding a new vector can have a knock on effect on the cost
+   * of subsequent vectors and the quality of prediction from NEAR and
+   * NEAREST for subsequent blocks. The "Weight" parameter allows, to a
+   * limited extent, for some account to be taken of these factors.
+   */
+  return ((mvcost[0][(mv->as_mv.row - ref->as_mv.row) >> 1] +
+           mvcost[1][(mv->as_mv.col - ref->as_mv.col) >> 1]) *
+          Weight) >>
+         7;
 }
 
-static int mv_err_cost(int_mv *mv, int_mv *ref, int *mvcost[2], int error_per_bit)
-{
-    /* Ignore mv costing if mvcost is NULL */
-    if (mvcost)
-        return ((mvcost[0][(mv->as_mv.row - ref->as_mv.row) >> 1] +
-                 mvcost[1][(mv->as_mv.col - ref->as_mv.col) >> 1])
-                 * error_per_bit + 128) >> 8;
-    return 0;
+static int mv_err_cost(int_mv *mv, int_mv *ref, int *mvcost[2],
+                       int error_per_bit) {
+  /* Ignore mv costing if mvcost is NULL */
+  if (mvcost)
+    return ((mvcost[0][(mv->as_mv.row - ref->as_mv.row) >> 1] +
+             mvcost[1][(mv->as_mv.col - ref->as_mv.col) >> 1]) *
+                error_per_bit +
+            128) >>
+           8;
+  return 0;
 }
 
-static int mvsad_err_cost(int_mv *mv, int_mv *ref, int *mvsadcost[2], int error_per_bit)
-{
-    /* Calculate sad error cost on full pixel basis. */
-    /* Ignore mv costing if mvsadcost is NULL */
-    if (mvsadcost)
-        return ((mvsadcost[0][(mv->as_mv.row - ref->as_mv.row)] +
-                 mvsadcost[1][(mv->as_mv.col - ref->as_mv.col)])
-                * error_per_bit + 128) >> 8;
-    return 0;
+static int mvsad_err_cost(int_mv *mv, int_mv *ref, int *mvsadcost[2],
+                          int error_per_bit) {
+  /* Calculate sad error cost on full pixel basis. */
+  /* Ignore mv costing if mvsadcost is NULL */
+  if (mvsadcost)
+    return ((mvsadcost[0][(mv->as_mv.row - ref->as_mv.row)] +
+             mvsadcost[1][(mv->as_mv.col - ref->as_mv.col)]) *
+                error_per_bit +
+            128) >>
+           8;
+  return 0;
 }
 
-void vp8_init_dsmotion_compensation(MACROBLOCK *x, int stride)
-{
-    int Len;
-    int search_site_count = 0;
+void vp8_init_dsmotion_compensation(MACROBLOCK *x, int stride) {
+  int Len;
+  int search_site_count = 0;
 
+  /* Generate offsets for 4 search sites per step. */
+  Len = MAX_FIRST_STEP;
+  x->ss[search_site_count].mv.col = 0;
+  x->ss[search_site_count].mv.row = 0;
+  x->ss[search_site_count].offset = 0;
+  search_site_count++;
 
-    /* Generate offsets for 4 search sites per step. */
-    Len = MAX_FIRST_STEP;
+  while (Len > 0) {
+    /* Compute offsets for search sites. */
     x->ss[search_site_count].mv.col = 0;
-    x->ss[search_site_count].mv.row = 0;
-    x->ss[search_site_count].offset = 0;
+    x->ss[search_site_count].mv.row = -Len;
+    x->ss[search_site_count].offset = -Len * stride;
     search_site_count++;
 
-    while (Len > 0)
-    {
+    /* Compute offsets for search sites. */
+    x->ss[search_site_count].mv.col = 0;
+    x->ss[search_site_count].mv.row = Len;
+    x->ss[search_site_count].offset = Len * stride;
+    search_site_count++;
 
-        /* Compute offsets for search sites. */
-        x->ss[search_site_count].mv.col = 0;
-        x->ss[search_site_count].mv.row = -Len;
-        x->ss[search_site_count].offset = -Len * stride;
-        search_site_count++;
+    /* Compute offsets for search sites. */
+    x->ss[search_site_count].mv.col = -Len;
+    x->ss[search_site_count].mv.row = 0;
+    x->ss[search_site_count].offset = -Len;
+    search_site_count++;
 
-        /* Compute offsets for search sites. */
-        x->ss[search_site_count].mv.col = 0;
-        x->ss[search_site_count].mv.row = Len;
-        x->ss[search_site_count].offset = Len * stride;
-        search_site_count++;
+    /* Compute offsets for search sites. */
+    x->ss[search_site_count].mv.col = Len;
+    x->ss[search_site_count].mv.row = 0;
+    x->ss[search_site_count].offset = Len;
+    search_site_count++;
 
-        /* Compute offsets for search sites. */
-        x->ss[search_site_count].mv.col = -Len;
-        x->ss[search_site_count].mv.row = 0;
-        x->ss[search_site_count].offset = -Len;
-        search_site_count++;
+    /* Contract. */
+    Len /= 2;
+  }
 
-        /* Compute offsets for search sites. */
-        x->ss[search_site_count].mv.col = Len;
-        x->ss[search_site_count].mv.row = 0;
-        x->ss[search_site_count].offset = Len;
-        search_site_count++;
+  x->ss_count = search_site_count;
+  x->searches_per_step = 4;
+}
 
-        /* Contract. */
-        Len /= 2;
-    }
+void vp8_init3smotion_compensation(MACROBLOCK *x, int stride) {
+  int Len;
+  int search_site_count = 0;
 
-    x->ss_count = search_site_count;
-    x->searches_per_step = 4;
-}
+  /* Generate offsets for 8 search sites per step. */
+  Len = MAX_FIRST_STEP;
+  x->ss[search_site_count].mv.col = 0;
+  x->ss[search_site_count].mv.row = 0;
+  x->ss[search_site_count].offset = 0;
+  search_site_count++;
 
-void vp8_init3smotion_compensation(MACROBLOCK *x, int stride)
-{
-    int Len;
-    int search_site_count = 0;
+  while (Len > 0) {
+    /* Compute offsets for search sites. */
+    x->ss[search_site_count].mv.col = 0;
+    x->ss[search_site_count].mv.row = -Len;
+    x->ss[search_site_count].offset = -Len * stride;
+    search_site_count++;
 
-    /* Generate offsets for 8 search sites per step. */
-    Len = MAX_FIRST_STEP;
+    /* Compute offsets for search sites. */
     x->ss[search_site_count].mv.col = 0;
+    x->ss[search_site_count].mv.row = Len;
+    x->ss[search_site_count].offset = Len * stride;
+    search_site_count++;
+
+    /* Compute offsets for search sites. */
+    x->ss[search_site_count].mv.col = -Len;
     x->ss[search_site_count].mv.row = 0;
-    x->ss[search_site_count].offset = 0;
+    x->ss[search_site_count].offset = -Len;
     search_site_count++;
 
-    while (Len > 0)
-    {
+    /* Compute offsets for search sites. */
+    x->ss[search_site_count].mv.col = Len;
+    x->ss[search_site_count].mv.row = 0;
+    x->ss[search_site_count].offset = Len;
+    search_site_count++;
 
-        /* Compute offsets for search sites. */
-        x->ss[search_site_count].mv.col = 0;
-        x->ss[search_site_count].mv.row = -Len;
-        x->ss[search_site_count].offset = -Len * stride;
-        search_site_count++;
+    /* Compute offsets for search sites. */
+    x->ss[search_site_count].mv.col = -Len;
+    x->ss[search_site_count].mv.row = -Len;
+    x->ss[search_site_count].offset = -Len * stride - Len;
+    search_site_count++;
 
-        /* Compute offsets for search sites. */
-        x->ss[search_site_count].mv.col = 0;
-        x->ss[search_site_count].mv.row = Len;
-        x->ss[search_site_count].offset = Len * stride;
-        search_site_count++;
+    /* Compute offsets for search sites. */
+    x->ss[search_site_count].mv.col = Len;
+    x->ss[search_site_count].mv.row = -Len;
+    x->ss[search_site_count].offset = -Len * stride + Len;
+    search_site_count++;
 
-        /* Compute offsets for search sites. */
-        x->ss[search_site_count].mv.col = -Len;
-        x->ss[search_site_count].mv.row = 0;
-        x->ss[search_site_count].offset = -Len;
-        search_site_count++;
+    /* Compute offsets for search sites. */
+    x->ss[search_site_count].mv.col = -Len;
+    x->ss[search_site_count].mv.row = Len;
+    x->ss[search_site_count].offset = Len * stride - Len;
+    search_site_count++;
 
-        /* Compute offsets for search sites. */
-        x->ss[search_site_count].mv.col = Len;
-        x->ss[search_site_count].mv.row = 0;
-        x->ss[search_site_count].offset = Len;
-        search_site_count++;
+    /* Compute offsets for search sites. */
+    x->ss[search_site_count].mv.col = Len;
+    x->ss[search_site_count].mv.row = Len;
+    x->ss[search_site_count].offset = Len * stride + Len;
+    search_site_count++;
 
-        /* Compute offsets for search sites. */
-        x->ss[search_site_count].mv.col = -Len;
-        x->ss[search_site_count].mv.row = -Len;
-        x->ss[search_site_count].offset = -Len * stride - Len;
-        search_site_count++;
+    /* Contract. */
+    Len /= 2;
+  }
 
-        /* Compute offsets for search sites. */
-        x->ss[search_site_count].mv.col = Len;
-        x->ss[search_site_count].mv.row = -Len;
-        x->ss[search_site_count].offset = -Len * stride + Len;
-        search_site_count++;
-
-        /* Compute offsets for search sites. */
-        x->ss[search_site_count].mv.col = -Len;
-        x->ss[search_site_count].mv.row = Len;
-        x->ss[search_site_count].offset = Len * stride - Len;
-        search_site_count++;
-
-        /* Compute offsets for search sites. */
-        x->ss[search_site_count].mv.col = Len;
-        x->ss[search_site_count].mv.row = Len;
-        x->ss[search_site_count].offset = Len * stride + Len;
-        search_site_count++;
-
-
-        /* Contract. */
-        Len /= 2;
-    }
-
-    x->ss_count = search_site_count;
-    x->searches_per_step = 8;
+  x->ss_count = search_site_count;
+  x->searches_per_step = 8;
 }
 
 /*
@@ -191,18 +188,33 @@
  */
 
 /* estimated cost of a motion vector (r,c) */
-#define MVC(r,c) (mvcost ? ((mvcost[0][(r)-rr] + mvcost[1][(c) - rc]) * error_per_bit + 128 )>>8 : 0)
+#define MVC(r, c)                                                             \
+  (mvcost                                                                     \
+       ? ((mvcost[0][(r)-rr] + mvcost[1][(c)-rc]) * error_per_bit + 128) >> 8 \
+       : 0)
 /* pointer to predictor base of a motionvector */
-#define PRE(r,c) (y + (((r)>>2) * y_stride + ((c)>>2) -(offset)))
+#define PRE(r, c) (y + (((r) >> 2) * y_stride + ((c) >> 2) - (offset)))
 /* convert motion vector component to offset for svf calc */
-#define SP(x) (((x)&3)<<1)
+#define SP(x) (((x)&3) << 1)
 /* returns subpixel variance error function. */
-#define DIST(r,c) vfp->svf( PRE(r,c), y_stride, SP(c),SP(r), z,b->src_stride,&sse)
-#define IFMVCV(r,c,s,e) if ( c >= minc && c <= maxc && r >= minr && r <= maxr) s else e;
+#define DIST(r, c) \
+  vfp->svf(PRE(r, c), y_stride, SP(c), SP(r), z, b->src_stride, &sse)
+#define IFMVCV(r, c, s, e) \
+  if (c >= minc && c <= maxc && r >= minr && r <= maxr) s else e;
 /* returns distortion + motion vector cost */
-#define ERR(r,c) (MVC(r,c)+DIST(r,c))
+#define ERR(r, c) (MVC(r, c) + DIST(r, c))
 /* checks if (r,c) has better score than previous best */
-#define CHECK_BETTER(v,r,c) IFMVCV(r,c,{thismse = DIST(r,c); if((v = (MVC(r,c)+thismse)) < besterr) { besterr = v; br=r; bc=c; *distortion = thismse; *sse1 = sse; }}, v=UINT_MAX;)
+#define CHECK_BETTER(v, r, c)                    \
+  IFMVCV(r, c, {                                 \
+    thismse = DIST(r, c);                        \
+    if ((v = (MVC(r, c) + thismse)) < besterr) { \
+      besterr = v;                               \
+      br = r;                                    \
+      bc = c;                                    \
+      *distortion = thismse;                     \
+      *sse1 = sse;                               \
+    }                                            \
+  }, v = UINT_MAX;)
 
 int vp8_find_best_sub_pixel_step_iteratively(MACROBLOCK *x, BLOCK *b, BLOCKD *d,
                                              int_mv *bestmv, int_mv *ref_mv,
@@ -209,150 +221,135 @@
                                              int error_per_bit,
                                              const vp8_variance_fn_ptr_t *vfp,
                                              int *mvcost[2], int *distortion,
-                                             unsigned int *sse1)
-{
-    unsigned char *z = (*(b->base_src) + b->src);
+                                             unsigned int *sse1) {
+  unsigned char *z = (*(b->base_src) + b->src);
 
-    int rr = ref_mv->as_mv.row >> 1, rc = ref_mv->as_mv.col >> 1;
-    int br = bestmv->as_mv.row * 4, bc = bestmv->as_mv.col * 4;
-    int tr = br, tc = bc;
-    unsigned int besterr;
-    unsigned int left, right, up, down, diag;
-    unsigned int sse;
-    unsigned int whichdir;
-    unsigned int halfiters = 4;
-    unsigned int quarteriters = 4;
-    int thismse;
+  int rr = ref_mv->as_mv.row >> 1, rc = ref_mv->as_mv.col >> 1;
+  int br = bestmv->as_mv.row * 4, bc = bestmv->as_mv.col * 4;
+  int tr = br, tc = bc;
+  unsigned int besterr;
+  unsigned int left, right, up, down, diag;
+  unsigned int sse;
+  unsigned int whichdir;
+  unsigned int halfiters = 4;
+  unsigned int quarteriters = 4;
+  int thismse;
 
-    int minc = VPXMAX(x->mv_col_min * 4,
-                      (ref_mv->as_mv.col >> 1) - ((1 << mvlong_width) - 1));
-    int maxc = VPXMIN(x->mv_col_max * 4,
-                      (ref_mv->as_mv.col >> 1) + ((1 << mvlong_width) - 1));
-    int minr = VPXMAX(x->mv_row_min * 4,
-                      (ref_mv->as_mv.row >> 1) - ((1 << mvlong_width) - 1));
-    int maxr = VPXMIN(x->mv_row_max * 4,
-                      (ref_mv->as_mv.row >> 1) + ((1 << mvlong_width) - 1));
+  int minc = VPXMAX(x->mv_col_min * 4,
+                    (ref_mv->as_mv.col >> 1) - ((1 << mvlong_width) - 1));
+  int maxc = VPXMIN(x->mv_col_max * 4,
+                    (ref_mv->as_mv.col >> 1) + ((1 << mvlong_width) - 1));
+  int minr = VPXMAX(x->mv_row_min * 4,
+                    (ref_mv->as_mv.row >> 1) - ((1 << mvlong_width) - 1));
+  int maxr = VPXMIN(x->mv_row_max * 4,
+                    (ref_mv->as_mv.row >> 1) + ((1 << mvlong_width) - 1));
 
-    int y_stride;
-    int offset;
-    int pre_stride = x->e_mbd.pre.y_stride;
-    unsigned char *base_pre = x->e_mbd.pre.y_buffer;
+  int y_stride;
+  int offset;
+  int pre_stride = x->e_mbd.pre.y_stride;
+  unsigned char *base_pre = x->e_mbd.pre.y_buffer;
 
-
 #if ARCH_X86 || ARCH_X86_64
-    MACROBLOCKD *xd = &x->e_mbd;
-    unsigned char *y_0 = base_pre + d->offset + (bestmv->as_mv.row) * pre_stride + bestmv->as_mv.col;
-    unsigned char *y;
-    int buf_r1, buf_r2, buf_c1;
+  MACROBLOCKD *xd = &x->e_mbd;
+  unsigned char *y_0 = base_pre + d->offset + (bestmv->as_mv.row) * pre_stride +
+                       bestmv->as_mv.col;
+  unsigned char *y;
+  int buf_r1, buf_r2, buf_c1;
 
-    /* Clamping to avoid out-of-range data access */
-    buf_r1 = ((bestmv->as_mv.row - 3) < x->mv_row_min)?(bestmv->as_mv.row - x->mv_row_min):3;
-    buf_r2 = ((bestmv->as_mv.row + 3) > x->mv_row_max)?(x->mv_row_max - bestmv->as_mv.row):3;
-    buf_c1 = ((bestmv->as_mv.col - 3) < x->mv_col_min)?(bestmv->as_mv.col - x->mv_col_min):3;
-    y_stride = 32;
+  /* Clamping to avoid out-of-range data access */
+  buf_r1 = ((bestmv->as_mv.row - 3) < x->mv_row_min)
+               ? (bestmv->as_mv.row - x->mv_row_min)
+               : 3;
+  buf_r2 = ((bestmv->as_mv.row + 3) > x->mv_row_max)
+               ? (x->mv_row_max - bestmv->as_mv.row)
+               : 3;
+  buf_c1 = ((bestmv->as_mv.col - 3) < x->mv_col_min)
+               ? (bestmv->as_mv.col - x->mv_col_min)
+               : 3;
+  y_stride = 32;
 
-    /* Copy to intermediate buffer before searching. */
-    vfp->copymem(y_0 - buf_c1 - pre_stride*buf_r1, pre_stride, xd->y_buf, y_stride, 16+buf_r1+buf_r2);
-    y = xd->y_buf + y_stride*buf_r1 +buf_c1;
+  /* Copy to intermediate buffer before searching. */
+  vfp->copymem(y_0 - buf_c1 - pre_stride * buf_r1, pre_stride, xd->y_buf,
+               y_stride, 16 + buf_r1 + buf_r2);
+  y = xd->y_buf + y_stride * buf_r1 + buf_c1;
 #else
-    unsigned char *y = base_pre + d->offset + (bestmv->as_mv.row) * pre_stride + bestmv->as_mv.col;
-    y_stride = pre_stride;
+  unsigned char *y = base_pre + d->offset + (bestmv->as_mv.row) * pre_stride +
+                     bestmv->as_mv.col;
+  y_stride = pre_stride;
 #endif
 
-    offset = (bestmv->as_mv.row) * y_stride + bestmv->as_mv.col;
+  offset = (bestmv->as_mv.row) * y_stride + bestmv->as_mv.col;
 
-    /* central mv */
-    bestmv->as_mv.row *= 8;
-    bestmv->as_mv.col *= 8;
+  /* central mv */
+  bestmv->as_mv.row *= 8;
+  bestmv->as_mv.col *= 8;
 
-    /* calculate central point error */
-    besterr = vfp->vf(y, y_stride, z, b->src_stride, sse1);
-    *distortion = besterr;
-    besterr += mv_err_cost(bestmv, ref_mv, mvcost, error_per_bit);
+  /* calculate central point error */
+  besterr = vfp->vf(y, y_stride, z, b->src_stride, sse1);
+  *distortion = besterr;
+  besterr += mv_err_cost(bestmv, ref_mv, mvcost, error_per_bit);
 
-    /* TODO: Each subsequent iteration checks at least one point in common
-     * with the last iteration could be 2 ( if diag selected)
-     */
-    while (--halfiters)
-    {
-        /* 1/2 pel */
-        CHECK_BETTER(left, tr, tc - 2);
-        CHECK_BETTER(right, tr, tc + 2);
-        CHECK_BETTER(up, tr - 2, tc);
-        CHECK_BETTER(down, tr + 2, tc);
+  /* TODO: Each subsequent iteration checks at least one point in common
+   * with the last iteration could be 2 ( if diag selected)
+   */
+  while (--halfiters) {
+    /* 1/2 pel */
+    CHECK_BETTER(left, tr, tc - 2);
+    CHECK_BETTER(right, tr, tc + 2);
+    CHECK_BETTER(up, tr - 2, tc);
+    CHECK_BETTER(down, tr + 2, tc);
 
-        whichdir = (left < right ? 0 : 1) + (up < down ? 0 : 2);
+    whichdir = (left < right ? 0 : 1) + (up < down ? 0 : 2);
 
-        switch (whichdir)
-        {
-        case 0:
-            CHECK_BETTER(diag, tr - 2, tc - 2);
-            break;
-        case 1:
-            CHECK_BETTER(diag, tr - 2, tc + 2);
-            break;
-        case 2:
-            CHECK_BETTER(diag, tr + 2, tc - 2);
-            break;
-        case 3:
-            CHECK_BETTER(diag, tr + 2, tc + 2);
-            break;
-        }
-
-        /* no reason to check the same one again. */
-        if (tr == br && tc == bc)
-            break;
-
-        tr = br;
-        tc = bc;
+    switch (whichdir) {
+      case 0: CHECK_BETTER(diag, tr - 2, tc - 2); break;
+      case 1: CHECK_BETTER(diag, tr - 2, tc + 2); break;
+      case 2: CHECK_BETTER(diag, tr + 2, tc - 2); break;
+      case 3: CHECK_BETTER(diag, tr + 2, tc + 2); break;
     }
 
-    /* TODO: Each subsequent iteration checks at least one point in common
-     * with the last iteration could be 2 ( if diag selected)
-     */
+    /* no reason to check the same one again. */
+    if (tr == br && tc == bc) break;
 
-    /* 1/4 pel */
-    while (--quarteriters)
-    {
-        CHECK_BETTER(left, tr, tc - 1);
-        CHECK_BETTER(right, tr, tc + 1);
-        CHECK_BETTER(up, tr - 1, tc);
-        CHECK_BETTER(down, tr + 1, tc);
+    tr = br;
+    tc = bc;
+  }
 
-        whichdir = (left < right ? 0 : 1) + (up < down ? 0 : 2);
+  /* TODO: Each subsequent iteration checks at least one point in common
+   * with the last iteration could be 2 ( if diag selected)
+   */
 
-        switch (whichdir)
-        {
-        case 0:
-            CHECK_BETTER(diag, tr - 1, tc - 1);
-            break;
-        case 1:
-            CHECK_BETTER(diag, tr - 1, tc + 1);
-            break;
-        case 2:
-            CHECK_BETTER(diag, tr + 1, tc - 1);
-            break;
-        case 3:
-            CHECK_BETTER(diag, tr + 1, tc + 1);
-            break;
-        }
+  /* 1/4 pel */
+  while (--quarteriters) {
+    CHECK_BETTER(left, tr, tc - 1);
+    CHECK_BETTER(right, tr, tc + 1);
+    CHECK_BETTER(up, tr - 1, tc);
+    CHECK_BETTER(down, tr + 1, tc);
 
-        /* no reason to check the same one again. */
-        if (tr == br && tc == bc)
-            break;
+    whichdir = (left < right ? 0 : 1) + (up < down ? 0 : 2);
 
-        tr = br;
-        tc = bc;
+    switch (whichdir) {
+      case 0: CHECK_BETTER(diag, tr - 1, tc - 1); break;
+      case 1: CHECK_BETTER(diag, tr - 1, tc + 1); break;
+      case 2: CHECK_BETTER(diag, tr + 1, tc - 1); break;
+      case 3: CHECK_BETTER(diag, tr + 1, tc + 1); break;
     }
 
-    bestmv->as_mv.row = br * 2;
-    bestmv->as_mv.col = bc * 2;
+    /* no reason to check the same one again. */
+    if (tr == br && tc == bc) break;
 
-    if ((abs(bestmv->as_mv.col - ref_mv->as_mv.col) > (MAX_FULL_PEL_VAL<<3)) ||
-        (abs(bestmv->as_mv.row - ref_mv->as_mv.row) > (MAX_FULL_PEL_VAL<<3)))
-        return INT_MAX;
+    tr = br;
+    tc = bc;
+  }
 
-    return besterr;
+  bestmv->as_mv.row = br * 2;
+  bestmv->as_mv.col = bc * 2;
+
+  if ((abs(bestmv->as_mv.col - ref_mv->as_mv.col) > (MAX_FULL_PEL_VAL << 3)) ||
+      (abs(bestmv->as_mv.row - ref_mv->as_mv.row) > (MAX_FULL_PEL_VAL << 3)))
+    return INT_MAX;
+
+  return besterr;
 }
 #undef MVC
 #undef PRE
@@ -367,310 +364,289 @@
                                  int error_per_bit,
                                  const vp8_variance_fn_ptr_t *vfp,
                                  int *mvcost[2], int *distortion,
-                                 unsigned int *sse1)
-{
-    int bestmse = INT_MAX;
-    int_mv startmv;
-    int_mv this_mv;
-    unsigned char *z = (*(b->base_src) + b->src);
-    int left, right, up, down, diag;
-    unsigned int sse;
-    int whichdir ;
-    int thismse;
-    int y_stride;
-    int pre_stride = x->e_mbd.pre.y_stride;
-    unsigned char *base_pre = x->e_mbd.pre.y_buffer;
+                                 unsigned int *sse1) {
+  int bestmse = INT_MAX;
+  int_mv startmv;
+  int_mv this_mv;
+  unsigned char *z = (*(b->base_src) + b->src);
+  int left, right, up, down, diag;
+  unsigned int sse;
+  int whichdir;
+  int thismse;
+  int y_stride;
+  int pre_stride = x->e_mbd.pre.y_stride;
+  unsigned char *base_pre = x->e_mbd.pre.y_buffer;
 
 #if ARCH_X86 || ARCH_X86_64
-    MACROBLOCKD *xd = &x->e_mbd;
-    unsigned char *y_0 = base_pre + d->offset + (bestmv->as_mv.row) * pre_stride + bestmv->as_mv.col;
-    unsigned char *y;
+  MACROBLOCKD *xd = &x->e_mbd;
+  unsigned char *y_0 = base_pre + d->offset + (bestmv->as_mv.row) * pre_stride +
+                       bestmv->as_mv.col;
+  unsigned char *y;
 
-    y_stride = 32;
-    /* Copy 18 rows x 32 cols area to intermediate buffer before searching. */
-     vfp->copymem(y_0 - 1 - pre_stride, pre_stride, xd->y_buf, y_stride, 18);
-     y = xd->y_buf + y_stride + 1;
+  y_stride = 32;
+  /* Copy 18 rows x 32 cols area to intermediate buffer before searching. */
+  vfp->copymem(y_0 - 1 - pre_stride, pre_stride, xd->y_buf, y_stride, 18);
+  y = xd->y_buf + y_stride + 1;
 #else
-     unsigned char *y = base_pre + d->offset + (bestmv->as_mv.row) * pre_stride + bestmv->as_mv.col;
-     y_stride = pre_stride;
+  unsigned char *y = base_pre + d->offset + (bestmv->as_mv.row) * pre_stride +
+                     bestmv->as_mv.col;
+  y_stride = pre_stride;
 #endif
 
-    /* central mv */
-    bestmv->as_mv.row *= 8;
-    bestmv->as_mv.col *= 8;
-    startmv = *bestmv;
+  /* central mv */
+  bestmv->as_mv.row *= 8;
+  bestmv->as_mv.col *= 8;
+  startmv = *bestmv;
 
-    /* calculate central point error */
-    bestmse = vfp->vf(y, y_stride, z, b->src_stride, sse1);
-    *distortion = bestmse;
-    bestmse += mv_err_cost(bestmv, ref_mv, mvcost, error_per_bit);
+  /* calculate central point error */
+  bestmse = vfp->vf(y, y_stride, z, b->src_stride, sse1);
+  *distortion = bestmse;
+  bestmse += mv_err_cost(bestmv, ref_mv, mvcost, error_per_bit);
 
-    /* go left then right and check error */
-    this_mv.as_mv.row = startmv.as_mv.row;
-    this_mv.as_mv.col = ((startmv.as_mv.col - 8) | 4);
-    thismse = vfp->svf_halfpix_h(y - 1, y_stride, z, b->src_stride, &sse);
-    left = thismse + mv_err_cost(&this_mv, ref_mv, mvcost, error_per_bit);
+  /* go left then right and check error */
+  this_mv.as_mv.row = startmv.as_mv.row;
+  this_mv.as_mv.col = ((startmv.as_mv.col - 8) | 4);
+  thismse = vfp->svf_halfpix_h(y - 1, y_stride, z, b->src_stride, &sse);
+  left = thismse + mv_err_cost(&this_mv, ref_mv, mvcost, error_per_bit);
 
-    if (left < bestmse)
-    {
-        *bestmv = this_mv;
-        bestmse = left;
-        *distortion = thismse;
-        *sse1 = sse;
-    }
+  if (left < bestmse) {
+    *bestmv = this_mv;
+    bestmse = left;
+    *distortion = thismse;
+    *sse1 = sse;
+  }
 
-    this_mv.as_mv.col += 8;
-    thismse = vfp->svf_halfpix_h(y, y_stride, z, b->src_stride, &sse);
-    right = thismse + mv_err_cost(&this_mv, ref_mv, mvcost, error_per_bit);
+  this_mv.as_mv.col += 8;
+  thismse = vfp->svf_halfpix_h(y, y_stride, z, b->src_stride, &sse);
+  right = thismse + mv_err_cost(&this_mv, ref_mv, mvcost, error_per_bit);
 
-    if (right < bestmse)
-    {
-        *bestmv = this_mv;
-        bestmse = right;
-        *distortion = thismse;
-        *sse1 = sse;
-    }
+  if (right < bestmse) {
+    *bestmv = this_mv;
+    bestmse = right;
+    *distortion = thismse;
+    *sse1 = sse;
+  }
 
-    /* go up then down and check error */
-    this_mv.as_mv.col = startmv.as_mv.col;
-    this_mv.as_mv.row = ((startmv.as_mv.row - 8) | 4);
-    thismse =  vfp->svf_halfpix_v(y - y_stride, y_stride, z, b->src_stride, &sse);
-    up = thismse + mv_err_cost(&this_mv, ref_mv, mvcost, error_per_bit);
+  /* go up then down and check error */
+  this_mv.as_mv.col = startmv.as_mv.col;
+  this_mv.as_mv.row = ((startmv.as_mv.row - 8) | 4);
+  thismse = vfp->svf_halfpix_v(y - y_stride, y_stride, z, b->src_stride, &sse);
+  up = thismse + mv_err_cost(&this_mv, ref_mv, mvcost, error_per_bit);
 
-    if (up < bestmse)
-    {
-        *bestmv = this_mv;
-        bestmse = up;
-        *distortion = thismse;
-        *sse1 = sse;
-    }
+  if (up < bestmse) {
+    *bestmv = this_mv;
+    bestmse = up;
+    *distortion = thismse;
+    *sse1 = sse;
+  }
 
-    this_mv.as_mv.row += 8;
-    thismse = vfp->svf_halfpix_v(y, y_stride, z, b->src_stride, &sse);
-    down = thismse + mv_err_cost(&this_mv, ref_mv, mvcost, error_per_bit);
+  this_mv.as_mv.row += 8;
+  thismse = vfp->svf_halfpix_v(y, y_stride, z, b->src_stride, &sse);
+  down = thismse + mv_err_cost(&this_mv, ref_mv, mvcost, error_per_bit);
 
-    if (down < bestmse)
-    {
-        *bestmv = this_mv;
-        bestmse = down;
-        *distortion = thismse;
-        *sse1 = sse;
-    }
+  if (down < bestmse) {
+    *bestmv = this_mv;
+    bestmse = down;
+    *distortion = thismse;
+    *sse1 = sse;
+  }
 
+  /* now check 1 more diagonal */
+  whichdir = (left < right ? 0 : 1) + (up < down ? 0 : 2);
+  this_mv = startmv;
 
-    /* now check 1 more diagonal */
-    whichdir = (left < right ? 0 : 1) + (up < down ? 0 : 2);
-    this_mv = startmv;
-
-    switch (whichdir)
-    {
+  switch (whichdir) {
     case 0:
-        this_mv.as_mv.col = (this_mv.as_mv.col - 8) | 4;
-        this_mv.as_mv.row = (this_mv.as_mv.row - 8) | 4;
-        thismse = vfp->svf_halfpix_hv(y - 1 - y_stride, y_stride, z, b->src_stride, &sse);
-        break;
+      this_mv.as_mv.col = (this_mv.as_mv.col - 8) | 4;
+      this_mv.as_mv.row = (this_mv.as_mv.row - 8) | 4;
+      thismse = vfp->svf_halfpix_hv(y - 1 - y_stride, y_stride, z,
+                                    b->src_stride, &sse);
+      break;
     case 1:
-        this_mv.as_mv.col += 4;
-        this_mv.as_mv.row = (this_mv.as_mv.row - 8) | 4;
-        thismse = vfp->svf_halfpix_hv(y - y_stride, y_stride, z, b->src_stride, &sse);
-        break;
+      this_mv.as_mv.col += 4;
+      this_mv.as_mv.row = (this_mv.as_mv.row - 8) | 4;
+      thismse =
+          vfp->svf_halfpix_hv(y - y_stride, y_stride, z, b->src_stride, &sse);
+      break;
     case 2:
-        this_mv.as_mv.col = (this_mv.as_mv.col - 8) | 4;
-        this_mv.as_mv.row += 4;
-        thismse = vfp->svf_halfpix_hv(y - 1, y_stride, z, b->src_stride, &sse);
-        break;
+      this_mv.as_mv.col = (this_mv.as_mv.col - 8) | 4;
+      this_mv.as_mv.row += 4;
+      thismse = vfp->svf_halfpix_hv(y - 1, y_stride, z, b->src_stride, &sse);
+      break;
     case 3:
     default:
-        this_mv.as_mv.col += 4;
-        this_mv.as_mv.row += 4;
-        thismse = vfp->svf_halfpix_hv(y, y_stride, z, b->src_stride, &sse);
-        break;
-    }
+      this_mv.as_mv.col += 4;
+      this_mv.as_mv.row += 4;
+      thismse = vfp->svf_halfpix_hv(y, y_stride, z, b->src_stride, &sse);
+      break;
+  }
 
-    diag = thismse + mv_err_cost(&this_mv, ref_mv, mvcost, error_per_bit);
+  diag = thismse + mv_err_cost(&this_mv, ref_mv, mvcost, error_per_bit);
 
-    if (diag < bestmse)
-    {
-        *bestmv = this_mv;
-        bestmse = diag;
-        *distortion = thismse;
-        *sse1 = sse;
-    }
+  if (diag < bestmse) {
+    *bestmv = this_mv;
+    bestmse = diag;
+    *distortion = thismse;
+    *sse1 = sse;
+  }
 
+  /* time to check quarter pels. */
+  if (bestmv->as_mv.row < startmv.as_mv.row) y -= y_stride;
 
-    /* time to check quarter pels. */
-    if (bestmv->as_mv.row < startmv.as_mv.row)
-        y -= y_stride;
+  if (bestmv->as_mv.col < startmv.as_mv.col) y--;
 
-    if (bestmv->as_mv.col < startmv.as_mv.col)
-        y--;
+  startmv = *bestmv;
 
-    startmv = *bestmv;
+  /* go left then right and check error */
+  this_mv.as_mv.row = startmv.as_mv.row;
 
+  if (startmv.as_mv.col & 7) {
+    this_mv.as_mv.col = startmv.as_mv.col - 2;
+    thismse = vfp->svf(y, y_stride, this_mv.as_mv.col & 7,
+                       this_mv.as_mv.row & 7, z, b->src_stride, &sse);
+  } else {
+    this_mv.as_mv.col = (startmv.as_mv.col - 8) | 6;
+    thismse = vfp->svf(y - 1, y_stride, 6, this_mv.as_mv.row & 7, z,
+                       b->src_stride, &sse);
+  }
 
+  left = thismse + mv_err_cost(&this_mv, ref_mv, mvcost, error_per_bit);
 
-    /* go left then right and check error */
-    this_mv.as_mv.row = startmv.as_mv.row;
+  if (left < bestmse) {
+    *bestmv = this_mv;
+    bestmse = left;
+    *distortion = thismse;
+    *sse1 = sse;
+  }
 
-    if (startmv.as_mv.col & 7)
-    {
-        this_mv.as_mv.col = startmv.as_mv.col - 2;
-        thismse = vfp->svf(y, y_stride, this_mv.as_mv.col & 7, this_mv.as_mv.row & 7, z, b->src_stride, &sse);
-    }
-    else
-    {
-        this_mv.as_mv.col = (startmv.as_mv.col - 8) | 6;
-        thismse = vfp->svf(y - 1, y_stride, 6, this_mv.as_mv.row & 7, z, b->src_stride, &sse);
-    }
+  this_mv.as_mv.col += 4;
+  thismse = vfp->svf(y, y_stride, this_mv.as_mv.col & 7, this_mv.as_mv.row & 7,
+                     z, b->src_stride, &sse);
+  right = thismse + mv_err_cost(&this_mv, ref_mv, mvcost, error_per_bit);
 
-    left = thismse + mv_err_cost(&this_mv, ref_mv, mvcost, error_per_bit);
+  if (right < bestmse) {
+    *bestmv = this_mv;
+    bestmse = right;
+    *distortion = thismse;
+    *sse1 = sse;
+  }
 
-    if (left < bestmse)
-    {
-        *bestmv = this_mv;
-        bestmse = left;
-        *distortion = thismse;
-        *sse1 = sse;
-    }
+  /* go up then down and check error */
+  this_mv.as_mv.col = startmv.as_mv.col;
 
-    this_mv.as_mv.col += 4;
-    thismse = vfp->svf(y, y_stride, this_mv.as_mv.col & 7, this_mv.as_mv.row & 7, z, b->src_stride, &sse);
-    right = thismse + mv_err_cost(&this_mv, ref_mv, mvcost, error_per_bit);
+  if (startmv.as_mv.row & 7) {
+    this_mv.as_mv.row = startmv.as_mv.row - 2;
+    thismse = vfp->svf(y, y_stride, this_mv.as_mv.col & 7,
+                       this_mv.as_mv.row & 7, z, b->src_stride, &sse);
+  } else {
+    this_mv.as_mv.row = (startmv.as_mv.row - 8) | 6;
+    thismse = vfp->svf(y - y_stride, y_stride, this_mv.as_mv.col & 7, 6, z,
+                       b->src_stride, &sse);
+  }
 
-    if (right < bestmse)
-    {
-        *bestmv = this_mv;
-        bestmse = right;
-        *distortion = thismse;
-        *sse1 = sse;
-    }
+  up = thismse + mv_err_cost(&this_mv, ref_mv, mvcost, error_per_bit);
 
-    /* go up then down and check error */
-    this_mv.as_mv.col = startmv.as_mv.col;
+  if (up < bestmse) {
+    *bestmv = this_mv;
+    bestmse = up;
+    *distortion = thismse;
+    *sse1 = sse;
+  }
 
-    if (startmv.as_mv.row & 7)
-    {
-        this_mv.as_mv.row = startmv.as_mv.row - 2;
-        thismse = vfp->svf(y, y_stride, this_mv.as_mv.col & 7, this_mv.as_mv.row & 7, z, b->src_stride, &sse);
-    }
-    else
-    {
-        this_mv.as_mv.row = (startmv.as_mv.row - 8) | 6;
-        thismse = vfp->svf(y - y_stride, y_stride, this_mv.as_mv.col & 7, 6, z, b->src_stride, &sse);
-    }
+  this_mv.as_mv.row += 4;
+  thismse = vfp->svf(y, y_stride, this_mv.as_mv.col & 7, this_mv.as_mv.row & 7,
+                     z, b->src_stride, &sse);
+  down = thismse + mv_err_cost(&this_mv, ref_mv, mvcost, error_per_bit);
 
-    up = thismse + mv_err_cost(&this_mv, ref_mv, mvcost, error_per_bit);
+  if (down < bestmse) {
+    *bestmv = this_mv;
+    bestmse = down;
+    *distortion = thismse;
+    *sse1 = sse;
+  }
 
-    if (up < bestmse)
-    {
-        *bestmv = this_mv;
-        bestmse = up;
-        *distortion = thismse;
-        *sse1 = sse;
-    }
+  /* now check 1 more diagonal */
+  whichdir = (left < right ? 0 : 1) + (up < down ? 0 : 2);
 
-    this_mv.as_mv.row += 4;
-    thismse = vfp->svf(y, y_stride, this_mv.as_mv.col & 7, this_mv.as_mv.row & 7, z, b->src_stride, &sse);
-    down = thismse + mv_err_cost(&this_mv, ref_mv, mvcost, error_per_bit);
+  this_mv = startmv;
 
-    if (down < bestmse)
-    {
-        *bestmv = this_mv;
-        bestmse = down;
-        *distortion = thismse;
-        *sse1 = sse;
-    }
-
-
-    /* now check 1 more diagonal */
-    whichdir = (left < right ? 0 : 1) + (up < down ? 0 : 2);
-
-    this_mv = startmv;
-
-    switch (whichdir)
-    {
+  switch (whichdir) {
     case 0:
 
-        if (startmv.as_mv.row & 7)
-        {
-            this_mv.as_mv.row -= 2;
+      if (startmv.as_mv.row & 7) {
+        this_mv.as_mv.row -= 2;
 
-            if (startmv.as_mv.col & 7)
-            {
-                this_mv.as_mv.col -= 2;
-                thismse = vfp->svf(y, y_stride, this_mv.as_mv.col & 7, this_mv.as_mv.row & 7, z, b->src_stride, &sse);
-            }
-            else
-            {
-                this_mv.as_mv.col = (startmv.as_mv.col - 8) | 6;
-                thismse = vfp->svf(y - 1, y_stride, 6, this_mv.as_mv.row & 7, z, b->src_stride, &sse);;
-            }
+        if (startmv.as_mv.col & 7) {
+          this_mv.as_mv.col -= 2;
+          thismse = vfp->svf(y, y_stride, this_mv.as_mv.col & 7,
+                             this_mv.as_mv.row & 7, z, b->src_stride, &sse);
+        } else {
+          this_mv.as_mv.col = (startmv.as_mv.col - 8) | 6;
+          thismse = vfp->svf(y - 1, y_stride, 6, this_mv.as_mv.row & 7, z,
+                             b->src_stride, &sse);
+          ;
         }
-        else
-        {
-            this_mv.as_mv.row = (startmv.as_mv.row - 8) | 6;
+      } else {
+        this_mv.as_mv.row = (startmv.as_mv.row - 8) | 6;
 
-            if (startmv.as_mv.col & 7)
-            {
-                this_mv.as_mv.col -= 2;
-                thismse = vfp->svf(y - y_stride, y_stride, this_mv.as_mv.col & 7, 6, z, b->src_stride, &sse);
-            }
-            else
-            {
-                this_mv.as_mv.col = (startmv.as_mv.col - 8) | 6;
-                thismse = vfp->svf(y - y_stride - 1, y_stride, 6, 6, z, b->src_stride, &sse);
-            }
+        if (startmv.as_mv.col & 7) {
+          this_mv.as_mv.col -= 2;
+          thismse = vfp->svf(y - y_stride, y_stride, this_mv.as_mv.col & 7, 6,
+                             z, b->src_stride, &sse);
+        } else {
+          this_mv.as_mv.col = (startmv.as_mv.col - 8) | 6;
+          thismse = vfp->svf(y - y_stride - 1, y_stride, 6, 6, z, b->src_stride,
+                             &sse);
         }
+      }
 
-        break;
+      break;
     case 1:
-        this_mv.as_mv.col += 2;
+      this_mv.as_mv.col += 2;
 
-        if (startmv.as_mv.row & 7)
-        {
-            this_mv.as_mv.row -= 2;
-            thismse = vfp->svf(y, y_stride, this_mv.as_mv.col & 7, this_mv.as_mv.row & 7, z, b->src_stride, &sse);
-        }
-        else
-        {
-            this_mv.as_mv.row = (startmv.as_mv.row - 8) | 6;
-            thismse = vfp->svf(y - y_stride, y_stride, this_mv.as_mv.col & 7, 6, z, b->src_stride, &sse);
-        }
+      if (startmv.as_mv.row & 7) {
+        this_mv.as_mv.row -= 2;
+        thismse = vfp->svf(y, y_stride, this_mv.as_mv.col & 7,
+                           this_mv.as_mv.row & 7, z, b->src_stride, &sse);
+      } else {
+        this_mv.as_mv.row = (startmv.as_mv.row - 8) | 6;
+        thismse = vfp->svf(y - y_stride, y_stride, this_mv.as_mv.col & 7, 6, z,
+                           b->src_stride, &sse);
+      }
 
-        break;
+      break;
     case 2:
-        this_mv.as_mv.row += 2;
+      this_mv.as_mv.row += 2;
 
-        if (startmv.as_mv.col & 7)
-        {
-            this_mv.as_mv.col -= 2;
-            thismse = vfp->svf(y, y_stride, this_mv.as_mv.col & 7, this_mv.as_mv.row & 7, z, b->src_stride, &sse);
-        }
-        else
-        {
-            this_mv.as_mv.col = (startmv.as_mv.col - 8) | 6;
-            thismse = vfp->svf(y - 1, y_stride, 6, this_mv.as_mv.row & 7, z, b->src_stride, &sse);
-        }
+      if (startmv.as_mv.col & 7) {
+        this_mv.as_mv.col -= 2;
+        thismse = vfp->svf(y, y_stride, this_mv.as_mv.col & 7,
+                           this_mv.as_mv.row & 7, z, b->src_stride, &sse);
+      } else {
+        this_mv.as_mv.col = (startmv.as_mv.col - 8) | 6;
+        thismse = vfp->svf(y - 1, y_stride, 6, this_mv.as_mv.row & 7, z,
+                           b->src_stride, &sse);
+      }
 
-        break;
+      break;
     case 3:
-        this_mv.as_mv.col += 2;
-        this_mv.as_mv.row += 2;
-        thismse = vfp->svf(y, y_stride,  this_mv.as_mv.col & 7, this_mv.as_mv.row & 7, z, b->src_stride, &sse);
-        break;
-    }
+      this_mv.as_mv.col += 2;
+      this_mv.as_mv.row += 2;
+      thismse = vfp->svf(y, y_stride, this_mv.as_mv.col & 7,
+                         this_mv.as_mv.row & 7, z, b->src_stride, &sse);
+      break;
+  }
 
-    diag = thismse + mv_err_cost(&this_mv, ref_mv, mvcost, error_per_bit);
+  diag = thismse + mv_err_cost(&this_mv, ref_mv, mvcost, error_per_bit);
 
-    if (diag < bestmse)
-    {
-        *bestmv = this_mv;
-        bestmse = diag;
-        *distortion = thismse;
-        *sse1 = sse;
-    }
+  if (diag < bestmse) {
+    *bestmv = this_mv;
+    bestmse = diag;
+    *distortion = thismse;
+    *sse1 = sse;
+  }
 
-    return bestmse;
+  return bestmse;
 }
 
 int vp8_find_best_half_pixel_step(MACROBLOCK *x, BLOCK *b, BLOCKD *d,
@@ -678,1355 +654,1238 @@
                                   int error_per_bit,
                                   const vp8_variance_fn_ptr_t *vfp,
                                   int *mvcost[2], int *distortion,
-                                  unsigned int *sse1)
-{
-    int bestmse = INT_MAX;
-    int_mv startmv;
-    int_mv this_mv;
-    unsigned char *z = (*(b->base_src) + b->src);
-    int left, right, up, down, diag;
-    unsigned int sse;
-    int whichdir ;
-    int thismse;
-    int y_stride;
-    int pre_stride = x->e_mbd.pre.y_stride;
-    unsigned char *base_pre = x->e_mbd.pre.y_buffer;
+                                  unsigned int *sse1) {
+  int bestmse = INT_MAX;
+  int_mv startmv;
+  int_mv this_mv;
+  unsigned char *z = (*(b->base_src) + b->src);
+  int left, right, up, down, diag;
+  unsigned int sse;
+  int whichdir;
+  int thismse;
+  int y_stride;
+  int pre_stride = x->e_mbd.pre.y_stride;
+  unsigned char *base_pre = x->e_mbd.pre.y_buffer;
 
 #if ARCH_X86 || ARCH_X86_64
-    MACROBLOCKD *xd = &x->e_mbd;
-    unsigned char *y_0 = base_pre + d->offset + (bestmv->as_mv.row) * pre_stride + bestmv->as_mv.col;
-    unsigned char *y;
+  MACROBLOCKD *xd = &x->e_mbd;
+  unsigned char *y_0 = base_pre + d->offset + (bestmv->as_mv.row) * pre_stride +
+                       bestmv->as_mv.col;
+  unsigned char *y;
 
-    y_stride = 32;
-    /* Copy 18 rows x 32 cols area to intermediate buffer before searching. */
-    vfp->copymem(y_0 - 1 - pre_stride, pre_stride, xd->y_buf, y_stride, 18);
-    y = xd->y_buf + y_stride + 1;
+  y_stride = 32;
+  /* Copy 18 rows x 32 cols area to intermediate buffer before searching. */
+  vfp->copymem(y_0 - 1 - pre_stride, pre_stride, xd->y_buf, y_stride, 18);
+  y = xd->y_buf + y_stride + 1;
 #else
-    unsigned char *y = base_pre + d->offset + (bestmv->as_mv.row) * pre_stride + bestmv->as_mv.col;
-    y_stride = pre_stride;
+  unsigned char *y = base_pre + d->offset + (bestmv->as_mv.row) * pre_stride +
+                     bestmv->as_mv.col;
+  y_stride = pre_stride;
 #endif
 
-    /* central mv */
-    bestmv->as_mv.row *= 8;
-    bestmv->as_mv.col *= 8;
-    startmv = *bestmv;
+  /* central mv */
+  bestmv->as_mv.row *= 8;
+  bestmv->as_mv.col *= 8;
+  startmv = *bestmv;
 
-    /* calculate central point error */
-    bestmse = vfp->vf(y, y_stride, z, b->src_stride, sse1);
-    *distortion = bestmse;
-    bestmse += mv_err_cost(bestmv, ref_mv, mvcost, error_per_bit);
+  /* calculate central point error */
+  bestmse = vfp->vf(y, y_stride, z, b->src_stride, sse1);
+  *distortion = bestmse;
+  bestmse += mv_err_cost(bestmv, ref_mv, mvcost, error_per_bit);
 
-    /* go left then right and check error */
-    this_mv.as_mv.row = startmv.as_mv.row;
-    this_mv.as_mv.col = ((startmv.as_mv.col - 8) | 4);
-    thismse = vfp->svf_halfpix_h(y - 1, y_stride, z, b->src_stride, &sse);
-    left = thismse + mv_err_cost(&this_mv, ref_mv, mvcost, error_per_bit);
+  /* go left then right and check error */
+  this_mv.as_mv.row = startmv.as_mv.row;
+  this_mv.as_mv.col = ((startmv.as_mv.col - 8) | 4);
+  thismse = vfp->svf_halfpix_h(y - 1, y_stride, z, b->src_stride, &sse);
+  left = thismse + mv_err_cost(&this_mv, ref_mv, mvcost, error_per_bit);
 
-    if (left < bestmse)
-    {
-        *bestmv = this_mv;
-        bestmse = left;
-        *distortion = thismse;
-        *sse1 = sse;
-    }
+  if (left < bestmse) {
+    *bestmv = this_mv;
+    bestmse = left;
+    *distortion = thismse;
+    *sse1 = sse;
+  }
 
-    this_mv.as_mv.col += 8;
-    thismse = vfp->svf_halfpix_h(y, y_stride, z, b->src_stride, &sse);
-    right = thismse + mv_err_cost(&this_mv, ref_mv, mvcost, error_per_bit);
+  this_mv.as_mv.col += 8;
+  thismse = vfp->svf_halfpix_h(y, y_stride, z, b->src_stride, &sse);
+  right = thismse + mv_err_cost(&this_mv, ref_mv, mvcost, error_per_bit);
 
-    if (right < bestmse)
-    {
-        *bestmv = this_mv;
-        bestmse = right;
-        *distortion = thismse;
-        *sse1 = sse;
-    }
+  if (right < bestmse) {
+    *bestmv = this_mv;
+    bestmse = right;
+    *distortion = thismse;
+    *sse1 = sse;
+  }
 
-    /* go up then down and check error */
-    this_mv.as_mv.col = startmv.as_mv.col;
-    this_mv.as_mv.row = ((startmv.as_mv.row - 8) | 4);
-    thismse = vfp->svf_halfpix_v(y - y_stride, y_stride, z, b->src_stride, &sse);
-    up = thismse + mv_err_cost(&this_mv, ref_mv, mvcost, error_per_bit);
+  /* go up then down and check error */
+  this_mv.as_mv.col = startmv.as_mv.col;
+  this_mv.as_mv.row = ((startmv.as_mv.row - 8) | 4);
+  thismse = vfp->svf_halfpix_v(y - y_stride, y_stride, z, b->src_stride, &sse);
+  up = thismse + mv_err_cost(&this_mv, ref_mv, mvcost, error_per_bit);
 
-    if (up < bestmse)
-    {
-        *bestmv = this_mv;
-        bestmse = up;
-        *distortion = thismse;
-        *sse1 = sse;
-    }
+  if (up < bestmse) {
+    *bestmv = this_mv;
+    bestmse = up;
+    *distortion = thismse;
+    *sse1 = sse;
+  }
 
-    this_mv.as_mv.row += 8;
-    thismse = vfp->svf_halfpix_v(y, y_stride, z, b->src_stride, &sse);
-    down = thismse + mv_err_cost(&this_mv, ref_mv, mvcost, error_per_bit);
+  this_mv.as_mv.row += 8;
+  thismse = vfp->svf_halfpix_v(y, y_stride, z, b->src_stride, &sse);
+  down = thismse + mv_err_cost(&this_mv, ref_mv, mvcost, error_per_bit);
 
-    if (down < bestmse)
-    {
-        *bestmv = this_mv;
-        bestmse = down;
-        *distortion = thismse;
-        *sse1 = sse;
-    }
+  if (down < bestmse) {
+    *bestmv = this_mv;
+    bestmse = down;
+    *distortion = thismse;
+    *sse1 = sse;
+  }
 
-    /* now check 1 more diagonal - */
-    whichdir = (left < right ? 0 : 1) + (up < down ? 0 : 2);
-    this_mv = startmv;
+  /* now check 1 more diagonal - */
+  whichdir = (left < right ? 0 : 1) + (up < down ? 0 : 2);
+  this_mv = startmv;
 
-    switch (whichdir)
-    {
+  switch (whichdir) {
     case 0:
-        this_mv.as_mv.col = (this_mv.as_mv.col - 8) | 4;
-        this_mv.as_mv.row = (this_mv.as_mv.row - 8) | 4;
-        thismse = vfp->svf_halfpix_hv(y - 1 - y_stride, y_stride, z, b->src_stride, &sse);
-        break;
+      this_mv.as_mv.col = (this_mv.as_mv.col - 8) | 4;
+      this_mv.as_mv.row = (this_mv.as_mv.row - 8) | 4;
+      thismse = vfp->svf_halfpix_hv(y - 1 - y_stride, y_stride, z,
+                                    b->src_stride, &sse);
+      break;
     case 1:
-        this_mv.as_mv.col += 4;
-        this_mv.as_mv.row = (this_mv.as_mv.row - 8) | 4;
-        thismse = vfp->svf_halfpix_hv(y - y_stride, y_stride, z, b->src_stride, &sse);
-        break;
+      this_mv.as_mv.col += 4;
+      this_mv.as_mv.row = (this_mv.as_mv.row - 8) | 4;
+      thismse =
+          vfp->svf_halfpix_hv(y - y_stride, y_stride, z, b->src_stride, &sse);
+      break;
     case 2:
-        this_mv.as_mv.col = (this_mv.as_mv.col - 8) | 4;
-        this_mv.as_mv.row += 4;
-        thismse = vfp->svf_halfpix_hv(y - 1, y_stride, z, b->src_stride, &sse);
-        break;
+      this_mv.as_mv.col = (this_mv.as_mv.col - 8) | 4;
+      this_mv.as_mv.row += 4;
+      thismse = vfp->svf_halfpix_hv(y - 1, y_stride, z, b->src_stride, &sse);
+      break;
     case 3:
     default:
-        this_mv.as_mv.col += 4;
-        this_mv.as_mv.row += 4;
-        thismse = vfp->svf_halfpix_hv(y, y_stride, z, b->src_stride, &sse);
-        break;
-    }
+      this_mv.as_mv.col += 4;
+      this_mv.as_mv.row += 4;
+      thismse = vfp->svf_halfpix_hv(y, y_stride, z, b->src_stride, &sse);
+      break;
+  }
 
-    diag = thismse + mv_err_cost(&this_mv, ref_mv, mvcost, error_per_bit);
+  diag = thismse + mv_err_cost(&this_mv, ref_mv, mvcost, error_per_bit);
 
-    if (diag < bestmse)
-    {
-        *bestmv = this_mv;
-        bestmse = diag;
-        *distortion = thismse;
-        *sse1 = sse;
-    }
+  if (diag < bestmse) {
+    *bestmv = this_mv;
+    bestmse = diag;
+    *distortion = thismse;
+    *sse1 = sse;
+  }
 
-    return bestmse;
+  return bestmse;
 }
 
-#define CHECK_BOUNDS(range) \
-{\
-    all_in = 1;\
-    all_in &= ((br-range) >= x->mv_row_min);\
-    all_in &= ((br+range) <= x->mv_row_max);\
-    all_in &= ((bc-range) >= x->mv_col_min);\
-    all_in &= ((bc+range) <= x->mv_col_max);\
-}
+#define CHECK_BOUNDS(range)                    \
+  {                                            \
+    all_in = 1;                                \
+    all_in &= ((br - range) >= x->mv_row_min); \
+    all_in &= ((br + range) <= x->mv_row_max); \
+    all_in &= ((bc - range) >= x->mv_col_min); \
+    all_in &= ((bc + range) <= x->mv_col_max); \
+  }
 
-#define CHECK_POINT \
-{\
-    if (this_mv.as_mv.col < x->mv_col_min) continue;\
-    if (this_mv.as_mv.col > x->mv_col_max) continue;\
-    if (this_mv.as_mv.row < x->mv_row_min) continue;\
-    if (this_mv.as_mv.row > x->mv_row_max) continue;\
-}
+#define CHECK_POINT                                  \
+  {                                                  \
+    if (this_mv.as_mv.col < x->mv_col_min) continue; \
+    if (this_mv.as_mv.col > x->mv_col_max) continue; \
+    if (this_mv.as_mv.row < x->mv_row_min) continue; \
+    if (this_mv.as_mv.row > x->mv_row_max) continue; \
+  }
 
-#define CHECK_BETTER \
-{\
-    if (thissad < bestsad)\
-    {\
-        thissad += mvsad_err_cost(&this_mv, &fcenter_mv, mvsadcost, sad_per_bit);\
-        if (thissad < bestsad)\
-        {\
-            bestsad = thissad;\
-            best_site = i;\
-        }\
-    }\
-}
+#define CHECK_BETTER                                                     \
+  {                                                                      \
+    if (thissad < bestsad) {                                             \
+      thissad +=                                                         \
+          mvsad_err_cost(&this_mv, &fcenter_mv, mvsadcost, sad_per_bit); \
+      if (thissad < bestsad) {                                           \
+        bestsad = thissad;                                               \
+        best_site = i;                                                   \
+      }                                                                  \
+    }                                                                    \
+  }
 
-static const MV next_chkpts[6][3] =
-{
-    {{ -2, 0}, { -1, -2}, {1, -2}},
-    {{ -1, -2}, {1, -2}, {2, 0}},
-    {{1, -2}, {2, 0}, {1, 2}},
-    {{2, 0}, {1, 2}, { -1, 2}},
-    {{1, 2}, { -1, 2}, { -2, 0}},
-    {{ -1, 2}, { -2, 0}, { -1, -2}}
+static const MV next_chkpts[6][3] = {
+  { { -2, 0 }, { -1, -2 }, { 1, -2 } }, { { -1, -2 }, { 1, -2 }, { 2, 0 } },
+  { { 1, -2 }, { 2, 0 }, { 1, 2 } },    { { 2, 0 }, { 1, 2 }, { -1, 2 } },
+  { { 1, 2 }, { -1, 2 }, { -2, 0 } },   { { -1, 2 }, { -2, 0 }, { -1, -2 } }
 };
 
-int vp8_hex_search
-(
-    MACROBLOCK *x,
-    BLOCK *b,
-    BLOCKD *d,
-    int_mv *ref_mv,
-    int_mv *best_mv,
-    int search_param,
-    int sad_per_bit,
-    const vp8_variance_fn_ptr_t *vfp,
-    int *mvsadcost[2],
-    int *mvcost[2],
-    int_mv *center_mv
-)
-{
-    MV hex[6] = { { -1, -2}, {1, -2}, {2, 0}, {1, 2}, { -1, 2}, { -2, 0} } ;
-    MV neighbors[4] = {{0, -1}, { -1, 0}, {1, 0}, {0, 1}} ;
-    int i, j;
+int vp8_hex_search(MACROBLOCK *x, BLOCK *b, BLOCKD *d, int_mv *ref_mv,
+                   int_mv *best_mv, int search_param, int sad_per_bit,
+                   const vp8_variance_fn_ptr_t *vfp, int *mvsadcost[2],
+                   int *mvcost[2], int_mv *center_mv) {
+  MV hex[6] = {
+    { -1, -2 }, { 1, -2 }, { 2, 0 }, { 1, 2 }, { -1, 2 }, { -2, 0 }
+  };
+  MV neighbors[4] = { { 0, -1 }, { -1, 0 }, { 1, 0 }, { 0, 1 } };
+  int i, j;
 
-    unsigned char *what = (*(b->base_src) + b->src);
-    int what_stride = b->src_stride;
-    int pre_stride = x->e_mbd.pre.y_stride;
-    unsigned char *base_pre = x->e_mbd.pre.y_buffer;
+  unsigned char *what = (*(b->base_src) + b->src);
+  int what_stride = b->src_stride;
+  int pre_stride = x->e_mbd.pre.y_stride;
+  unsigned char *base_pre = x->e_mbd.pre.y_buffer;
 
-    int in_what_stride = pre_stride;
-    int br, bc;
-    int_mv this_mv;
-    unsigned int bestsad;
-    unsigned int thissad;
-    unsigned char *base_offset;
-    unsigned char *this_offset;
-    int k = -1;
-    int all_in;
-    int best_site = -1;
-    int hex_range = 127;
-    int dia_range = 8;
+  int in_what_stride = pre_stride;
+  int br, bc;
+  int_mv this_mv;
+  unsigned int bestsad;
+  unsigned int thissad;
+  unsigned char *base_offset;
+  unsigned char *this_offset;
+  int k = -1;
+  int all_in;
+  int best_site = -1;
+  int hex_range = 127;
+  int dia_range = 8;
 
-    int_mv fcenter_mv;
-    fcenter_mv.as_mv.row = center_mv->as_mv.row >> 3;
-    fcenter_mv.as_mv.col = center_mv->as_mv.col >> 3;
+  int_mv fcenter_mv;
+  fcenter_mv.as_mv.row = center_mv->as_mv.row >> 3;
+  fcenter_mv.as_mv.col = center_mv->as_mv.col >> 3;
 
-    (void)mvcost;
+  (void)mvcost;
 
-    /* adjust ref_mv to make sure it is within MV range */
-    vp8_clamp_mv(ref_mv, x->mv_col_min, x->mv_col_max, x->mv_row_min, x->mv_row_max);
-    br = ref_mv->as_mv.row;
-    bc = ref_mv->as_mv.col;
+  /* adjust ref_mv to make sure it is within MV range */
+  vp8_clamp_mv(ref_mv, x->mv_col_min, x->mv_col_max, x->mv_row_min,
+               x->mv_row_max);
+  br = ref_mv->as_mv.row;
+  bc = ref_mv->as_mv.col;
 
-    /* Work out the start point for the search */
-    base_offset = (unsigned char *)(base_pre + d->offset);
-    this_offset = base_offset + (br * (pre_stride)) + bc;
-    this_mv.as_mv.row = br;
-    this_mv.as_mv.col = bc;
-    bestsad = vfp->sdf(what, what_stride, this_offset, in_what_stride)
-            + mvsad_err_cost(&this_mv, &fcenter_mv, mvsadcost, sad_per_bit);
+  /* Work out the start point for the search */
+  base_offset = (unsigned char *)(base_pre + d->offset);
+  this_offset = base_offset + (br * (pre_stride)) + bc;
+  this_mv.as_mv.row = br;
+  this_mv.as_mv.col = bc;
+  bestsad = vfp->sdf(what, what_stride, this_offset, in_what_stride) +
+            mvsad_err_cost(&this_mv, &fcenter_mv, mvsadcost, sad_per_bit);
 
 #if CONFIG_MULTI_RES_ENCODING
-    /* Lower search range based on prediction info */
-    if (search_param >= 6) goto cal_neighbors;
-    else if (search_param >= 5) hex_range = 4;
-    else if (search_param >= 4) hex_range = 6;
-    else if (search_param >= 3) hex_range = 15;
-    else if (search_param >= 2) hex_range = 31;
-    else if (search_param >= 1) hex_range = 63;
+  /* Lower search range based on prediction info */
+  if (search_param >= 6)
+    goto cal_neighbors;
+  else if (search_param >= 5)
+    hex_range = 4;
+  else if (search_param >= 4)
+    hex_range = 6;
+  else if (search_param >= 3)
+    hex_range = 15;
+  else if (search_param >= 2)
+    hex_range = 31;
+  else if (search_param >= 1)
+    hex_range = 63;
 
-    dia_range = 8;
+  dia_range = 8;
 #else
-    (void)search_param;
+  (void)search_param;
 #endif
 
-    /* hex search */
-    CHECK_BOUNDS(2)
+  /* hex search */
+  CHECK_BOUNDS(2)
 
-    if(all_in)
-    {
-        for (i = 0; i < 6; i++)
-        {
-            this_mv.as_mv.row = br + hex[i].row;
-            this_mv.as_mv.col = bc + hex[i].col;
-            this_offset = base_offset + (this_mv.as_mv.row * in_what_stride) + this_mv.as_mv.col;
-            thissad = vfp->sdf(what, what_stride, this_offset, in_what_stride);
-            CHECK_BETTER
-        }
-    }else
-    {
-        for (i = 0; i < 6; i++)
-        {
-            this_mv.as_mv.row = br + hex[i].row;
-            this_mv.as_mv.col = bc + hex[i].col;
-            CHECK_POINT
-            this_offset = base_offset + (this_mv.as_mv.row * in_what_stride) + this_mv.as_mv.col;
-            thissad = vfp->sdf(what, what_stride, this_offset, in_what_stride);
-            CHECK_BETTER
-        }
+  if (all_in) {
+    for (i = 0; i < 6; i++) {
+      this_mv.as_mv.row = br + hex[i].row;
+      this_mv.as_mv.col = bc + hex[i].col;
+      this_offset = base_offset + (this_mv.as_mv.row * in_what_stride) +
+                    this_mv.as_mv.col;
+      thissad = vfp->sdf(what, what_stride, this_offset, in_what_stride);
+      CHECK_BETTER
     }
-
-    if (best_site == -1)
-        goto cal_neighbors;
-    else
-    {
-        br += hex[best_site].row;
-        bc += hex[best_site].col;
-        k = best_site;
+  } else {
+    for (i = 0; i < 6; i++) {
+      this_mv.as_mv.row = br + hex[i].row;
+      this_mv.as_mv.col = bc + hex[i].col;
+      CHECK_POINT
+      this_offset = base_offset + (this_mv.as_mv.row * in_what_stride) +
+                    this_mv.as_mv.col;
+      thissad = vfp->sdf(what, what_stride, this_offset, in_what_stride);
+      CHECK_BETTER
     }
+  }
 
-    for (j = 1; j < hex_range; j++)
-    {
-        best_site = -1;
-        CHECK_BOUNDS(2)
+  if (best_site == -1)
+    goto cal_neighbors;
+  else {
+    br += hex[best_site].row;
+    bc += hex[best_site].col;
+    k = best_site;
+  }
 
-        if(all_in)
-        {
-            for (i = 0; i < 3; i++)
-            {
-                this_mv.as_mv.row = br + next_chkpts[k][i].row;
-                this_mv.as_mv.col = bc + next_chkpts[k][i].col;
-                this_offset = base_offset + (this_mv.as_mv.row * (in_what_stride)) + this_mv.as_mv.col;
-                thissad = vfp->sdf(what, what_stride, this_offset, in_what_stride);
-                CHECK_BETTER
-            }
-        }else
-        {
-            for (i = 0; i < 3; i++)
-            {
-                this_mv.as_mv.row = br + next_chkpts[k][i].row;
-                this_mv.as_mv.col = bc + next_chkpts[k][i].col;
-                CHECK_POINT
-                this_offset = base_offset + (this_mv.as_mv.row * (in_what_stride)) + this_mv.as_mv.col;
-                thissad = vfp->sdf(what, what_stride, this_offset, in_what_stride);
-                CHECK_BETTER
-            }
-        }
+  for (j = 1; j < hex_range; j++) {
+    best_site = -1;
+    CHECK_BOUNDS(2)
 
-        if (best_site == -1)
-            break;
-        else
-        {
-            br += next_chkpts[k][best_site].row;
-            bc += next_chkpts[k][best_site].col;
-            k += 5 + best_site;
-            if (k >= 12) k -= 12;
-            else if (k >= 6) k -= 6;
-        }
+    if (all_in) {
+      for (i = 0; i < 3; i++) {
+        this_mv.as_mv.row = br + next_chkpts[k][i].row;
+        this_mv.as_mv.col = bc + next_chkpts[k][i].col;
+        this_offset = base_offset + (this_mv.as_mv.row * (in_what_stride)) +
+                      this_mv.as_mv.col;
+        thissad = vfp->sdf(what, what_stride, this_offset, in_what_stride);
+        CHECK_BETTER
+      }
+    } else {
+      for (i = 0; i < 3; i++) {
+        this_mv.as_mv.row = br + next_chkpts[k][i].row;
+        this_mv.as_mv.col = bc + next_chkpts[k][i].col;
+        CHECK_POINT
+        this_offset = base_offset + (this_mv.as_mv.row * (in_what_stride)) +
+                      this_mv.as_mv.col;
+        thissad = vfp->sdf(what, what_stride, this_offset, in_what_stride);
+        CHECK_BETTER
+      }
     }
 
-    /* check 4 1-away neighbors */
+    if (best_site == -1)
+      break;
+    else {
+      br += next_chkpts[k][best_site].row;
+      bc += next_chkpts[k][best_site].col;
+      k += 5 + best_site;
+      if (k >= 12)
+        k -= 12;
+      else if (k >= 6)
+        k -= 6;
+    }
+  }
+
+/* check 4 1-away neighbors */
 cal_neighbors:
-    for (j = 0; j < dia_range; j++)
-    {
-        best_site = -1;
-        CHECK_BOUNDS(1)
+  for (j = 0; j < dia_range; j++) {
+    best_site = -1;
+    CHECK_BOUNDS(1)
 
-        if(all_in)
-        {
-            for (i = 0; i < 4; i++)
-            {
-                this_mv.as_mv.row = br + neighbors[i].row;
-                this_mv.as_mv.col = bc + neighbors[i].col;
-                this_offset = base_offset + (this_mv.as_mv.row * (in_what_stride)) + this_mv.as_mv.col;
-                thissad = vfp->sdf(what, what_stride, this_offset, in_what_stride);
-                CHECK_BETTER
-            }
-        }else
-        {
-            for (i = 0; i < 4; i++)
-            {
-                this_mv.as_mv.row = br + neighbors[i].row;
-                this_mv.as_mv.col = bc + neighbors[i].col;
-                CHECK_POINT
-                this_offset = base_offset + (this_mv.as_mv.row * (in_what_stride)) + this_mv.as_mv.col;
-                thissad = vfp->sdf(what, what_stride, this_offset, in_what_stride);
-                CHECK_BETTER
-            }
-        }
+    if (all_in) {
+      for (i = 0; i < 4; i++) {
+        this_mv.as_mv.row = br + neighbors[i].row;
+        this_mv.as_mv.col = bc + neighbors[i].col;
+        this_offset = base_offset + (this_mv.as_mv.row * (in_what_stride)) +
+                      this_mv.as_mv.col;
+        thissad = vfp->sdf(what, what_stride, this_offset, in_what_stride);
+        CHECK_BETTER
+      }
+    } else {
+      for (i = 0; i < 4; i++) {
+        this_mv.as_mv.row = br + neighbors[i].row;
+        this_mv.as_mv.col = bc + neighbors[i].col;
+        CHECK_POINT
+        this_offset = base_offset + (this_mv.as_mv.row * (in_what_stride)) +
+                      this_mv.as_mv.col;
+        thissad = vfp->sdf(what, what_stride, this_offset, in_what_stride);
+        CHECK_BETTER
+      }
+    }
 
-        if (best_site == -1)
-            break;
-        else
-        {
-            br += neighbors[best_site].row;
-            bc += neighbors[best_site].col;
-        }
+    if (best_site == -1)
+      break;
+    else {
+      br += neighbors[best_site].row;
+      bc += neighbors[best_site].col;
     }
+  }
 
-    best_mv->as_mv.row = br;
-    best_mv->as_mv.col = bc;
+  best_mv->as_mv.row = br;
+  best_mv->as_mv.col = bc;
 
-    return bestsad;
+  return bestsad;
 }
 #undef CHECK_BOUNDS
 #undef CHECK_POINT
 #undef CHECK_BETTER
 
-int vp8_diamond_search_sad_c
-(
-    MACROBLOCK *x,
-    BLOCK *b,
-    BLOCKD *d,
-    int_mv *ref_mv,
-    int_mv *best_mv,
-    int search_param,
-    int sad_per_bit,
-    int *num00,
-    vp8_variance_fn_ptr_t *fn_ptr,
-    int *mvcost[2],
-    int_mv *center_mv
-)
-{
-    int i, j, step;
+int vp8_diamond_search_sad_c(MACROBLOCK *x, BLOCK *b, BLOCKD *d, int_mv *ref_mv,
+                             int_mv *best_mv, int search_param, int sad_per_bit,
+                             int *num00, vp8_variance_fn_ptr_t *fn_ptr,
+                             int *mvcost[2], int_mv *center_mv) {
+  int i, j, step;
 
-    unsigned char *what = (*(b->base_src) + b->src);
-    int what_stride = b->src_stride;
-    unsigned char *in_what;
-    int pre_stride = x->e_mbd.pre.y_stride;
-    unsigned char *base_pre = x->e_mbd.pre.y_buffer;
-    int in_what_stride = pre_stride;
-    unsigned char *best_address;
+  unsigned char *what = (*(b->base_src) + b->src);
+  int what_stride = b->src_stride;
+  unsigned char *in_what;
+  int pre_stride = x->e_mbd.pre.y_stride;
+  unsigned char *base_pre = x->e_mbd.pre.y_buffer;
+  int in_what_stride = pre_stride;
+  unsigned char *best_address;
 
-    int tot_steps;
-    int_mv this_mv;
+  int tot_steps;
+  int_mv this_mv;
 
-    unsigned int bestsad;
-    unsigned int thissad;
-    int best_site = 0;
-    int last_site = 0;
+  unsigned int bestsad;
+  unsigned int thissad;
+  int best_site = 0;
+  int last_site = 0;
 
-    int ref_row;
-    int ref_col;
-    int this_row_offset;
-    int this_col_offset;
-    search_site *ss;
+  int ref_row;
+  int ref_col;
+  int this_row_offset;
+  int this_col_offset;
+  search_site *ss;
 
-    unsigned char *check_here;
+  unsigned char *check_here;
 
-    int *mvsadcost[2];
-    int_mv fcenter_mv;
+  int *mvsadcost[2];
+  int_mv fcenter_mv;
 
-    mvsadcost[0] = x->mvsadcost[0];
-    mvsadcost[1] = x->mvsadcost[1];
-    fcenter_mv.as_mv.row = center_mv->as_mv.row >> 3;
-    fcenter_mv.as_mv.col = center_mv->as_mv.col >> 3;
+  mvsadcost[0] = x->mvsadcost[0];
+  mvsadcost[1] = x->mvsadcost[1];
+  fcenter_mv.as_mv.row = center_mv->as_mv.row >> 3;
+  fcenter_mv.as_mv.col = center_mv->as_mv.col >> 3;
 
-    vp8_clamp_mv(ref_mv, x->mv_col_min, x->mv_col_max, x->mv_row_min, x->mv_row_max);
-    ref_row = ref_mv->as_mv.row;
-    ref_col = ref_mv->as_mv.col;
-    *num00 = 0;
-    best_mv->as_mv.row = ref_row;
-    best_mv->as_mv.col = ref_col;
+  vp8_clamp_mv(ref_mv, x->mv_col_min, x->mv_col_max, x->mv_row_min,
+               x->mv_row_max);
+  ref_row = ref_mv->as_mv.row;
+  ref_col = ref_mv->as_mv.col;
+  *num00 = 0;
+  best_mv->as_mv.row = ref_row;
+  best_mv->as_mv.col = ref_col;
 
-    /* Work out the start point for the search */
-    in_what = (unsigned char *)(base_pre + d->offset + (ref_row * pre_stride) + ref_col);
-    best_address = in_what;
+  /* Work out the start point for the search */
+  in_what = (unsigned char *)(base_pre + d->offset + (ref_row * pre_stride) +
+                              ref_col);
+  best_address = in_what;
 
-    /* Check the starting position */
-    bestsad = fn_ptr->sdf(what, what_stride, in_what, in_what_stride)
-            + mvsad_err_cost(best_mv, &fcenter_mv, mvsadcost, sad_per_bit);
+  /* Check the starting position */
+  bestsad = fn_ptr->sdf(what, what_stride, in_what, in_what_stride) +
+            mvsad_err_cost(best_mv, &fcenter_mv, mvsadcost, sad_per_bit);
 
-    /* search_param determines the length of the initial step and hence
-     * the number of iterations 0 = initial step (MAX_FIRST_STEP) pel :
-     * 1 = (MAX_FIRST_STEP/2) pel, 2 = (MAX_FIRST_STEP/4) pel... etc.
-     */
-    ss = &x->ss[search_param * x->searches_per_step];
-    tot_steps = (x->ss_count / x->searches_per_step) - search_param;
+  /* search_param determines the length of the initial step and hence
+   * the number of iterations 0 = initial step (MAX_FIRST_STEP) pel :
+   * 1 = (MAX_FIRST_STEP/2) pel, 2 = (MAX_FIRST_STEP/4) pel... etc.
+   */
+  ss = &x->ss[search_param * x->searches_per_step];
+  tot_steps = (x->ss_count / x->searches_per_step) - search_param;
 
-    i = 1;
+  i = 1;
 
-    for (step = 0; step < tot_steps ; step++)
-    {
-        for (j = 0 ; j < x->searches_per_step ; j++)
-        {
-            /* Trap illegal vectors */
-            this_row_offset = best_mv->as_mv.row + ss[i].mv.row;
-            this_col_offset = best_mv->as_mv.col + ss[i].mv.col;
+  for (step = 0; step < tot_steps; step++) {
+    for (j = 0; j < x->searches_per_step; j++) {
+      /* Trap illegal vectors */
+      this_row_offset = best_mv->as_mv.row + ss[i].mv.row;
+      this_col_offset = best_mv->as_mv.col + ss[i].mv.col;
 
-            if ((this_col_offset > x->mv_col_min) && (this_col_offset < x->mv_col_max) &&
-            (this_row_offset > x->mv_row_min) && (this_row_offset < x->mv_row_max))
+      if ((this_col_offset > x->mv_col_min) &&
+          (this_col_offset < x->mv_col_max) &&
+          (this_row_offset > x->mv_row_min) &&
+          (this_row_offset < x->mv_row_max))
 
-            {
-                check_here = ss[i].offset + best_address;
-                thissad = fn_ptr->sdf(what, what_stride, check_here, in_what_stride);
+      {
+        check_here = ss[i].offset + best_address;
+        thissad = fn_ptr->sdf(what, what_stride, check_here, in_what_stride);
 
-                if (thissad < bestsad)
-                {
-                    this_mv.as_mv.row = this_row_offset;
-                    this_mv.as_mv.col = this_col_offset;
-                    thissad += mvsad_err_cost(&this_mv, &fcenter_mv,
-                                              mvsadcost, sad_per_bit);
+        if (thissad < bestsad) {
+          this_mv.as_mv.row = this_row_offset;
+          this_mv.as_mv.col = this_col_offset;
+          thissad +=
+              mvsad_err_cost(&this_mv, &fcenter_mv, mvsadcost, sad_per_bit);
 
-                    if (thissad < bestsad)
-                    {
-                        bestsad = thissad;
-                        best_site = i;
-                    }
-                }
-            }
-
-            i++;
+          if (thissad < bestsad) {
+            bestsad = thissad;
+            best_site = i;
+          }
         }
+      }
 
-        if (best_site != last_site)
-        {
-            best_mv->as_mv.row += ss[best_site].mv.row;
-            best_mv->as_mv.col += ss[best_site].mv.col;
-            best_address += ss[best_site].offset;
-            last_site = best_site;
-        }
-        else if (best_address == in_what)
-            (*num00)++;
+      i++;
     }
 
-    this_mv.as_mv.row = best_mv->as_mv.row << 3;
-    this_mv.as_mv.col = best_mv->as_mv.col << 3;
+    if (best_site != last_site) {
+      best_mv->as_mv.row += ss[best_site].mv.row;
+      best_mv->as_mv.col += ss[best_site].mv.col;
+      best_address += ss[best_site].offset;
+      last_site = best_site;
+    } else if (best_address == in_what)
+      (*num00)++;
+  }
 
-    return fn_ptr->vf(what, what_stride, best_address, in_what_stride, &thissad)
-           + mv_err_cost(&this_mv, center_mv, mvcost, x->errorperbit);
+  this_mv.as_mv.row = best_mv->as_mv.row << 3;
+  this_mv.as_mv.col = best_mv->as_mv.col << 3;
+
+  return fn_ptr->vf(what, what_stride, best_address, in_what_stride, &thissad) +
+         mv_err_cost(&this_mv, center_mv, mvcost, x->errorperbit);
 }
 
-int vp8_diamond_search_sadx4
-(
-    MACROBLOCK *x,
-    BLOCK *b,
-    BLOCKD *d,
-    int_mv *ref_mv,
-    int_mv *best_mv,
-    int search_param,
-    int sad_per_bit,
-    int *num00,
-    vp8_variance_fn_ptr_t *fn_ptr,
-    int *mvcost[2],
-    int_mv *center_mv
-)
-{
-    int i, j, step;
+int vp8_diamond_search_sadx4(MACROBLOCK *x, BLOCK *b, BLOCKD *d, int_mv *ref_mv,
+                             int_mv *best_mv, int search_param, int sad_per_bit,
+                             int *num00, vp8_variance_fn_ptr_t *fn_ptr,
+                             int *mvcost[2], int_mv *center_mv) {
+  int i, j, step;
 
-    unsigned char *what = (*(b->base_src) + b->src);
-    int what_stride = b->src_stride;
-    unsigned char *in_what;
-    int pre_stride = x->e_mbd.pre.y_stride;
-    unsigned char *base_pre = x->e_mbd.pre.y_buffer;
-    int in_what_stride = pre_stride;
-    unsigned char *best_address;
+  unsigned char *what = (*(b->base_src) + b->src);
+  int what_stride = b->src_stride;
+  unsigned char *in_what;
+  int pre_stride = x->e_mbd.pre.y_stride;
+  unsigned char *base_pre = x->e_mbd.pre.y_buffer;
+  int in_what_stride = pre_stride;
+  unsigned char *best_address;
 
-    int tot_steps;
-    int_mv this_mv;
+  int tot_steps;
+  int_mv this_mv;
 
-    unsigned int bestsad;
-    unsigned int thissad;
-    int best_site = 0;
-    int last_site = 0;
+  unsigned int bestsad;
+  unsigned int thissad;
+  int best_site = 0;
+  int last_site = 0;
 
-    int ref_row;
-    int ref_col;
-    int this_row_offset;
-    int this_col_offset;
-    search_site *ss;
+  int ref_row;
+  int ref_col;
+  int this_row_offset;
+  int this_col_offset;
+  search_site *ss;
 
-    unsigned char *check_here;
+  unsigned char *check_here;
 
-    int *mvsadcost[2];
-    int_mv fcenter_mv;
+  int *mvsadcost[2];
+  int_mv fcenter_mv;
 
-    mvsadcost[0] = x->mvsadcost[0];
-    mvsadcost[1] = x->mvsadcost[1];
-    fcenter_mv.as_mv.row = center_mv->as_mv.row >> 3;
-    fcenter_mv.as_mv.col = center_mv->as_mv.col >> 3;
+  mvsadcost[0] = x->mvsadcost[0];
+  mvsadcost[1] = x->mvsadcost[1];
+  fcenter_mv.as_mv.row = center_mv->as_mv.row >> 3;
+  fcenter_mv.as_mv.col = center_mv->as_mv.col >> 3;
 
-    vp8_clamp_mv(ref_mv, x->mv_col_min, x->mv_col_max, x->mv_row_min, x->mv_row_max);
-    ref_row = ref_mv->as_mv.row;
-    ref_col = ref_mv->as_mv.col;
-    *num00 = 0;
-    best_mv->as_mv.row = ref_row;
-    best_mv->as_mv.col = ref_col;
+  vp8_clamp_mv(ref_mv, x->mv_col_min, x->mv_col_max, x->mv_row_min,
+               x->mv_row_max);
+  ref_row = ref_mv->as_mv.row;
+  ref_col = ref_mv->as_mv.col;
+  *num00 = 0;
+  best_mv->as_mv.row = ref_row;
+  best_mv->as_mv.col = ref_col;
 
-    /* Work out the start point for the search */
-    in_what = (unsigned char *)(base_pre + d->offset + (ref_row * pre_stride) + ref_col);
-    best_address = in_what;
+  /* Work out the start point for the search */
+  in_what = (unsigned char *)(base_pre + d->offset + (ref_row * pre_stride) +
+                              ref_col);
+  best_address = in_what;
 
-    /* Check the starting position */
-    bestsad = fn_ptr->sdf(what, what_stride, in_what, in_what_stride)
-            + mvsad_err_cost(best_mv, &fcenter_mv, mvsadcost, sad_per_bit);
+  /* Check the starting position */
+  bestsad = fn_ptr->sdf(what, what_stride, in_what, in_what_stride) +
+            mvsad_err_cost(best_mv, &fcenter_mv, mvsadcost, sad_per_bit);
 
-    /* search_param determines the length of the initial step and hence the
-     * number of iterations 0 = initial step (MAX_FIRST_STEP) pel : 1 =
-     * (MAX_FIRST_STEP/2) pel, 2 = (MAX_FIRST_STEP/4) pel... etc.
-     */
-    ss = &x->ss[search_param * x->searches_per_step];
-    tot_steps = (x->ss_count / x->searches_per_step) - search_param;
+  /* search_param determines the length of the initial step and hence the
+   * number of iterations 0 = initial step (MAX_FIRST_STEP) pel : 1 =
+   * (MAX_FIRST_STEP/2) pel, 2 = (MAX_FIRST_STEP/4) pel... etc.
+   */
+  ss = &x->ss[search_param * x->searches_per_step];
+  tot_steps = (x->ss_count / x->searches_per_step) - search_param;
 
-    i = 1;
+  i = 1;
 
-    for (step = 0; step < tot_steps ; step++)
-    {
-        int all_in = 1, t;
+  for (step = 0; step < tot_steps; step++) {
+    int all_in = 1, t;
 
-        /* To know if all neighbor points are within the bounds, 4 bounds
-         * checking are enough instead of checking 4 bounds for each
-         * points.
-         */
-        all_in &= ((best_mv->as_mv.row + ss[i].mv.row)> x->mv_row_min);
-        all_in &= ((best_mv->as_mv.row + ss[i+1].mv.row) < x->mv_row_max);
-        all_in &= ((best_mv->as_mv.col + ss[i+2].mv.col) > x->mv_col_min);
-        all_in &= ((best_mv->as_mv.col + ss[i+3].mv.col) < x->mv_col_max);
+    /* To know if all neighbor points are within the bounds, 4 bounds
+     * checking are enough instead of checking 4 bounds for each
+     * points.
+     */
+    all_in &= ((best_mv->as_mv.row + ss[i].mv.row) > x->mv_row_min);
+    all_in &= ((best_mv->as_mv.row + ss[i + 1].mv.row) < x->mv_row_max);
+    all_in &= ((best_mv->as_mv.col + ss[i + 2].mv.col) > x->mv_col_min);
+    all_in &= ((best_mv->as_mv.col + ss[i + 3].mv.col) < x->mv_col_max);
 
-        if (all_in)
-        {
-            unsigned int sad_array[4];
+    if (all_in) {
+      unsigned int sad_array[4];
 
-            for (j = 0 ; j < x->searches_per_step ; j += 4)
-            {
-                const unsigned char *block_offset[4];
+      for (j = 0; j < x->searches_per_step; j += 4) {
+        const unsigned char *block_offset[4];
 
-                for (t = 0; t < 4; t++)
-                    block_offset[t] = ss[i+t].offset + best_address;
+        for (t = 0; t < 4; t++)
+          block_offset[t] = ss[i + t].offset + best_address;
 
-                fn_ptr->sdx4df(what, what_stride, block_offset, in_what_stride, sad_array);
+        fn_ptr->sdx4df(what, what_stride, block_offset, in_what_stride,
+                       sad_array);
 
-                for (t = 0; t < 4; t++, i++)
-                {
-                    if (sad_array[t] < bestsad)
-                    {
-                        this_mv.as_mv.row = best_mv->as_mv.row + ss[i].mv.row;
-                        this_mv.as_mv.col = best_mv->as_mv.col + ss[i].mv.col;
-                        sad_array[t] += mvsad_err_cost(&this_mv, &fcenter_mv,
-                                                       mvsadcost, sad_per_bit);
+        for (t = 0; t < 4; t++, i++) {
+          if (sad_array[t] < bestsad) {
+            this_mv.as_mv.row = best_mv->as_mv.row + ss[i].mv.row;
+            this_mv.as_mv.col = best_mv->as_mv.col + ss[i].mv.col;
+            sad_array[t] +=
+                mvsad_err_cost(&this_mv, &fcenter_mv, mvsadcost, sad_per_bit);
 
-                        if (sad_array[t] < bestsad)
-                        {
-                            bestsad = sad_array[t];
-                            best_site = i;
-                        }
-                    }
-                }
+            if (sad_array[t] < bestsad) {
+              bestsad = sad_array[t];
+              best_site = i;
             }
+          }
         }
-        else
-        {
-            for (j = 0 ; j < x->searches_per_step ; j++)
-            {
-                /* Trap illegal vectors */
-                this_row_offset = best_mv->as_mv.row + ss[i].mv.row;
-                this_col_offset = best_mv->as_mv.col + ss[i].mv.col;
+      }
+    } else {
+      for (j = 0; j < x->searches_per_step; j++) {
+        /* Trap illegal vectors */
+        this_row_offset = best_mv->as_mv.row + ss[i].mv.row;
+        this_col_offset = best_mv->as_mv.col + ss[i].mv.col;
 
-                if ((this_col_offset > x->mv_col_min) && (this_col_offset < x->mv_col_max) &&
-                (this_row_offset > x->mv_row_min) && (this_row_offset < x->mv_row_max))
-                {
-                    check_here = ss[i].offset + best_address;
-                    thissad = fn_ptr->sdf(what, what_stride, check_here, in_what_stride);
+        if ((this_col_offset > x->mv_col_min) &&
+            (this_col_offset < x->mv_col_max) &&
+            (this_row_offset > x->mv_row_min) &&
+            (this_row_offset < x->mv_row_max)) {
+          check_here = ss[i].offset + best_address;
+          thissad = fn_ptr->sdf(what, what_stride, check_here, in_what_stride);
 
-                    if (thissad < bestsad)
-                    {
-                        this_mv.as_mv.row = this_row_offset;
-                        this_mv.as_mv.col = this_col_offset;
-                        thissad += mvsad_err_cost(&this_mv, &fcenter_mv,
-                                                  mvsadcost, sad_per_bit);
+          if (thissad < bestsad) {
+            this_mv.as_mv.row = this_row_offset;
+            this_mv.as_mv.col = this_col_offset;
+            thissad +=
+                mvsad_err_cost(&this_mv, &fcenter_mv, mvsadcost, sad_per_bit);
 
-                        if (thissad < bestsad)
-                        {
-                            bestsad = thissad;
-                            best_site = i;
-                        }
-                    }
-                }
-                i++;
+            if (thissad < bestsad) {
+              bestsad = thissad;
+              best_site = i;
             }
+          }
         }
-
-        if (best_site != last_site)
-        {
-            best_mv->as_mv.row += ss[best_site].mv.row;
-            best_mv->as_mv.col += ss[best_site].mv.col;
-            best_address += ss[best_site].offset;
-            last_site = best_site;
-        }
-        else if (best_address == in_what)
-            (*num00)++;
+        i++;
+      }
     }
 
-    this_mv.as_mv.row = best_mv->as_mv.row * 8;
-    this_mv.as_mv.col = best_mv->as_mv.col * 8;
+    if (best_site != last_site) {
+      best_mv->as_mv.row += ss[best_site].mv.row;
+      best_mv->as_mv.col += ss[best_site].mv.col;
+      best_address += ss[best_site].offset;
+      last_site = best_site;
+    } else if (best_address == in_what)
+      (*num00)++;
+  }
 
-    return fn_ptr->vf(what, what_stride, best_address, in_what_stride, &thissad)
-           + mv_err_cost(&this_mv, center_mv, mvcost, x->errorperbit);
+  this_mv.as_mv.row = best_mv->as_mv.row * 8;
+  this_mv.as_mv.col = best_mv->as_mv.col * 8;
+
+  return fn_ptr->vf(what, what_stride, best_address, in_what_stride, &thissad) +
+         mv_err_cost(&this_mv, center_mv, mvcost, x->errorperbit);
 }
 
 int vp8_full_search_sad_c(MACROBLOCK *x, BLOCK *b, BLOCKD *d, int_mv *ref_mv,
-                        int sad_per_bit, int distance,
-                        vp8_variance_fn_ptr_t *fn_ptr, int *mvcost[2],
-                        int_mv *center_mv)
-{
-    unsigned char *what = (*(b->base_src) + b->src);
-    int what_stride = b->src_stride;
-    unsigned char *in_what;
-    int pre_stride = x->e_mbd.pre.y_stride;
-    unsigned char *base_pre = x->e_mbd.pre.y_buffer;
-    int in_what_stride = pre_stride;
-    int mv_stride = pre_stride;
-    unsigned char *bestaddress;
-    int_mv *best_mv = &d->bmi.mv;
-    int_mv this_mv;
-    unsigned int bestsad;
-    unsigned int thissad;
-    int r, c;
+                          int sad_per_bit, int distance,
+                          vp8_variance_fn_ptr_t *fn_ptr, int *mvcost[2],
+                          int_mv *center_mv) {
+  unsigned char *what = (*(b->base_src) + b->src);
+  int what_stride = b->src_stride;
+  unsigned char *in_what;
+  int pre_stride = x->e_mbd.pre.y_stride;
+  unsigned char *base_pre = x->e_mbd.pre.y_buffer;
+  int in_what_stride = pre_stride;
+  int mv_stride = pre_stride;
+  unsigned char *bestaddress;
+  int_mv *best_mv = &d->bmi.mv;
+  int_mv this_mv;
+  unsigned int bestsad;
+  unsigned int thissad;
+  int r, c;
 
-    unsigned char *check_here;
+  unsigned char *check_here;
 
-    int ref_row = ref_mv->as_mv.row;
-    int ref_col = ref_mv->as_mv.col;
+  int ref_row = ref_mv->as_mv.row;
+  int ref_col = ref_mv->as_mv.col;
 
-    int row_min = ref_row - distance;
-    int row_max = ref_row + distance;
-    int col_min = ref_col - distance;
-    int col_max = ref_col + distance;
+  int row_min = ref_row - distance;
+  int row_max = ref_row + distance;
+  int col_min = ref_col - distance;
+  int col_max = ref_col + distance;
 
-    int *mvsadcost[2];
-    int_mv fcenter_mv;
+  int *mvsadcost[2];
+  int_mv fcenter_mv;
 
-    mvsadcost[0] = x->mvsadcost[0];
-    mvsadcost[1] = x->mvsadcost[1];
-    fcenter_mv.as_mv.row = center_mv->as_mv.row >> 3;
-    fcenter_mv.as_mv.col = center_mv->as_mv.col >> 3;
+  mvsadcost[0] = x->mvsadcost[0];
+  mvsadcost[1] = x->mvsadcost[1];
+  fcenter_mv.as_mv.row = center_mv->as_mv.row >> 3;
+  fcenter_mv.as_mv.col = center_mv->as_mv.col >> 3;
 
-    /* Work out the mid point for the search */
-    in_what = base_pre + d->offset;
-    bestaddress = in_what + (ref_row * pre_stride) + ref_col;
+  /* Work out the mid point for the search */
+  in_what = base_pre + d->offset;
+  bestaddress = in_what + (ref_row * pre_stride) + ref_col;
 
-    best_mv->as_mv.row = ref_row;
-    best_mv->as_mv.col = ref_col;
+  best_mv->as_mv.row = ref_row;
+  best_mv->as_mv.col = ref_col;
 
-    /* Baseline value at the centre */
-    bestsad = fn_ptr->sdf(what, what_stride, bestaddress, in_what_stride)
-            + mvsad_err_cost(best_mv, &fcenter_mv, mvsadcost, sad_per_bit);
+  /* Baseline value at the centre */
+  bestsad = fn_ptr->sdf(what, what_stride, bestaddress, in_what_stride) +
+            mvsad_err_cost(best_mv, &fcenter_mv, mvsadcost, sad_per_bit);
 
-    /* Apply further limits to prevent us looking using vectors that
-     * stretch beyiond the UMV border
-     */
-    if (col_min < x->mv_col_min)
-        col_min = x->mv_col_min;
+  /* Apply further limits to prevent us looking using vectors that
+   * stretch beyiond the UMV border
+   */
+  if (col_min < x->mv_col_min) col_min = x->mv_col_min;
 
-    if (col_max > x->mv_col_max)
-        col_max = x->mv_col_max;
+  if (col_max > x->mv_col_max) col_max = x->mv_col_max;
 
-    if (row_min < x->mv_row_min)
-        row_min = x->mv_row_min;
+  if (row_min < x->mv_row_min) row_min = x->mv_row_min;
 
-    if (row_max > x->mv_row_max)
-        row_max = x->mv_row_max;
+  if (row_max > x->mv_row_max) row_max = x->mv_row_max;
 
-    for (r = row_min; r < row_max ; r++)
-    {
-        this_mv.as_mv.row = r;
-        check_here = r * mv_stride + in_what + col_min;
+  for (r = row_min; r < row_max; r++) {
+    this_mv.as_mv.row = r;
+    check_here = r * mv_stride + in_what + col_min;
 
-        for (c = col_min; c < col_max; c++)
-        {
-            thissad = fn_ptr->sdf(what, what_stride, check_here, in_what_stride);
+    for (c = col_min; c < col_max; c++) {
+      thissad = fn_ptr->sdf(what, what_stride, check_here, in_what_stride);
 
-            this_mv.as_mv.col = c;
-            thissad += mvsad_err_cost(&this_mv, &fcenter_mv,
-                                      mvsadcost, sad_per_bit);
+      this_mv.as_mv.col = c;
+      thissad += mvsad_err_cost(&this_mv, &fcenter_mv, mvsadcost, sad_per_bit);
 
-            if (thissad < bestsad)
-            {
-                bestsad = thissad;
-                best_mv->as_mv.row = r;
-                best_mv->as_mv.col = c;
-                bestaddress = check_here;
-            }
+      if (thissad < bestsad) {
+        bestsad = thissad;
+        best_mv->as_mv.row = r;
+        best_mv->as_mv.col = c;
+        bestaddress = check_here;
+      }
 
-            check_here++;
-        }
+      check_here++;
     }
+  }
 
-    this_mv.as_mv.row = best_mv->as_mv.row << 3;
-    this_mv.as_mv.col = best_mv->as_mv.col << 3;
+  this_mv.as_mv.row = best_mv->as_mv.row << 3;
+  this_mv.as_mv.col = best_mv->as_mv.col << 3;
 
-    return fn_ptr->vf(what, what_stride, bestaddress, in_what_stride, &thissad)
-           + mv_err_cost(&this_mv, center_mv, mvcost, x->errorperbit);
+  return fn_ptr->vf(what, what_stride, bestaddress, in_what_stride, &thissad) +
+         mv_err_cost(&this_mv, center_mv, mvcost, x->errorperbit);
 }
 
 int vp8_full_search_sadx3(MACROBLOCK *x, BLOCK *b, BLOCKD *d, int_mv *ref_mv,
                           int sad_per_bit, int distance,
                           vp8_variance_fn_ptr_t *fn_ptr, int *mvcost[2],
-                          int_mv *center_mv)
-{
-    unsigned char *what = (*(b->base_src) + b->src);
-    int what_stride = b->src_stride;
-    unsigned char *in_what;
-    int pre_stride = x->e_mbd.pre.y_stride;
-    unsigned char *base_pre = x->e_mbd.pre.y_buffer;
-    int in_what_stride = pre_stride;
-    int mv_stride = pre_stride;
-    unsigned char *bestaddress;
-    int_mv *best_mv = &d->bmi.mv;
-    int_mv this_mv;
-    unsigned int bestsad;
-    unsigned int thissad;
-    int r, c;
+                          int_mv *center_mv) {
+  unsigned char *what = (*(b->base_src) + b->src);
+  int what_stride = b->src_stride;
+  unsigned char *in_what;
+  int pre_stride = x->e_mbd.pre.y_stride;
+  unsigned char *base_pre = x->e_mbd.pre.y_buffer;
+  int in_what_stride = pre_stride;
+  int mv_stride = pre_stride;
+  unsigned char *bestaddress;
+  int_mv *best_mv = &d->bmi.mv;
+  int_mv this_mv;
+  unsigned int bestsad;
+  unsigned int thissad;
+  int r, c;
 
-    unsigned char *check_here;
+  unsigned char *check_here;
 
-    int ref_row = ref_mv->as_mv.row;
-    int ref_col = ref_mv->as_mv.col;
+  int ref_row = ref_mv->as_mv.row;
+  int ref_col = ref_mv->as_mv.col;
 
-    int row_min = ref_row - distance;
-    int row_max = ref_row + distance;
-    int col_min = ref_col - distance;
-    int col_max = ref_col + distance;
+  int row_min = ref_row - distance;
+  int row_max = ref_row + distance;
+  int col_min = ref_col - distance;
+  int col_max = ref_col + distance;
 
-    unsigned int sad_array[3];
+  unsigned int sad_array[3];
 
-    int *mvsadcost[2];
-    int_mv fcenter_mv;
+  int *mvsadcost[2];
+  int_mv fcenter_mv;
 
-    mvsadcost[0] = x->mvsadcost[0];
-    mvsadcost[1] = x->mvsadcost[1];
-    fcenter_mv.as_mv.row = center_mv->as_mv.row >> 3;
-    fcenter_mv.as_mv.col = center_mv->as_mv.col >> 3;
+  mvsadcost[0] = x->mvsadcost[0];
+  mvsadcost[1] = x->mvsadcost[1];
+  fcenter_mv.as_mv.row = center_mv->as_mv.row >> 3;
+  fcenter_mv.as_mv.col = center_mv->as_mv.col >> 3;
 
-    /* Work out the mid point for the search */
-    in_what = base_pre + d->offset;
-    bestaddress = in_what + (ref_row * pre_stride) + ref_col;
+  /* Work out the mid point for the search */
+  in_what = base_pre + d->offset;
+  bestaddress = in_what + (ref_row * pre_stride) + ref_col;
 
-    best_mv->as_mv.row = ref_row;
-    best_mv->as_mv.col = ref_col;
+  best_mv->as_mv.row = ref_row;
+  best_mv->as_mv.col = ref_col;
 
-    /* Baseline value at the centre */
-    bestsad = fn_ptr->sdf(what, what_stride, bestaddress, in_what_stride)
-            + mvsad_err_cost(best_mv, &fcenter_mv, mvsadcost, sad_per_bit);
+  /* Baseline value at the centre */
+  bestsad = fn_ptr->sdf(what, what_stride, bestaddress, in_what_stride) +
+            mvsad_err_cost(best_mv, &fcenter_mv, mvsadcost, sad_per_bit);
 
-    /* Apply further limits to prevent us looking using vectors that stretch
-     * beyond the UMV border
-     */
-    if (col_min < x->mv_col_min)
-        col_min = x->mv_col_min;
+  /* Apply further limits to prevent us looking using vectors that stretch
+   * beyond the UMV border
+   */
+  if (col_min < x->mv_col_min) col_min = x->mv_col_min;
 
-    if (col_max > x->mv_col_max)
-        col_max = x->mv_col_max;
+  if (col_max > x->mv_col_max) col_max = x->mv_col_max;
 
-    if (row_min < x->mv_row_min)
-        row_min = x->mv_row_min;
+  if (row_min < x->mv_row_min) row_min = x->mv_row_min;
 
-    if (row_max > x->mv_row_max)
-        row_max = x->mv_row_max;
+  if (row_max > x->mv_row_max) row_max = x->mv_row_max;
 
-    for (r = row_min; r < row_max ; r++)
-    {
-        this_mv.as_mv.row = r;
-        check_here = r * mv_stride + in_what + col_min;
-        c = col_min;
+  for (r = row_min; r < row_max; r++) {
+    this_mv.as_mv.row = r;
+    check_here = r * mv_stride + in_what + col_min;
+    c = col_min;
 
-        while ((c + 2) < col_max)
-        {
-            int i;
+    while ((c + 2) < col_max) {
+      int i;
 
-            fn_ptr->sdx3f(what, what_stride, check_here, in_what_stride, sad_array);
+      fn_ptr->sdx3f(what, what_stride, check_here, in_what_stride, sad_array);
 
-            for (i = 0; i < 3; i++)
-            {
-                thissad = sad_array[i];
+      for (i = 0; i < 3; i++) {
+        thissad = sad_array[i];
 
-                if (thissad < bestsad)
-                {
-                    this_mv.as_mv.col = c;
-                    thissad += mvsad_err_cost(&this_mv, &fcenter_mv,
-                                              mvsadcost, sad_per_bit);
+        if (thissad < bestsad) {
+          this_mv.as_mv.col = c;
+          thissad +=
+              mvsad_err_cost(&this_mv, &fcenter_mv, mvsadcost, sad_per_bit);
 
-                    if (thissad < bestsad)
-                    {
-                        bestsad = thissad;
-                        best_mv->as_mv.row = r;
-                        best_mv->as_mv.col = c;
-                        bestaddress = check_here;
-                    }
-                }
-
-                check_here++;
-                c++;
-            }
+          if (thissad < bestsad) {
+            bestsad = thissad;
+            best_mv->as_mv.row = r;
+            best_mv->as_mv.col = c;
+            bestaddress = check_here;
+          }
         }
 
-        while (c < col_max)
-        {
-            thissad = fn_ptr->sdf(what, what_stride, check_here, in_what_stride);
+        check_here++;
+        c++;
+      }
+    }
 
-            if (thissad < bestsad)
-            {
-                this_mv.as_mv.col = c;
-                thissad += mvsad_err_cost(&this_mv, &fcenter_mv,
-                                          mvsadcost, sad_per_bit);
+    while (c < col_max) {
+      thissad = fn_ptr->sdf(what, what_stride, check_here, in_what_stride);
 
-                if (thissad < bestsad)
-                {
-                    bestsad = thissad;
-                    best_mv->as_mv.row = r;
-                    best_mv->as_mv.col = c;
-                    bestaddress = check_here;
-                }
-            }
+      if (thissad < bestsad) {
+        this_mv.as_mv.col = c;
+        thissad +=
+            mvsad_err_cost(&this_mv, &fcenter_mv, mvsadcost, sad_per_bit);
 
-            check_here ++;
-            c ++;
+        if (thissad < bestsad) {
+          bestsad = thissad;
+          best_mv->as_mv.row = r;
+          best_mv->as_mv.col = c;
+          bestaddress = check_here;
         }
+      }
 
+      check_here++;
+      c++;
     }
+  }
 
-    this_mv.as_mv.row = best_mv->as_mv.row << 3;
-    this_mv.as_mv.col = best_mv->as_mv.col << 3;
+  this_mv.as_mv.row = best_mv->as_mv.row << 3;
+  this_mv.as_mv.col = best_mv->as_mv.col << 3;
 
-    return fn_ptr->vf(what, what_stride, bestaddress, in_what_stride, &thissad)
-           + mv_err_cost(&this_mv, center_mv, mvcost, x->errorperbit);
+  return fn_ptr->vf(what, what_stride, bestaddress, in_what_stride, &thissad) +
+         mv_err_cost(&this_mv, center_mv, mvcost, x->errorperbit);
 }
 
 int vp8_full_search_sadx8(MACROBLOCK *x, BLOCK *b, BLOCKD *d, int_mv *ref_mv,
                           int sad_per_bit, int distance,
                           vp8_variance_fn_ptr_t *fn_ptr, int *mvcost[2],
-                          int_mv *center_mv)
-{
-    unsigned char *what = (*(b->base_src) + b->src);
-    int what_stride = b->src_stride;
-    int pre_stride = x->e_mbd.pre.y_stride;
-    unsigned char *base_pre = x->e_mbd.pre.y_buffer;
-    unsigned char *in_what;
-    int in_what_stride = pre_stride;
-    int mv_stride = pre_stride;
-    unsigned char *bestaddress;
-    int_mv *best_mv = &d->bmi.mv;
-    int_mv this_mv;
-    unsigned int bestsad;
-    unsigned int thissad;
-    int r, c;
+                          int_mv *center_mv) {
+  unsigned char *what = (*(b->base_src) + b->src);
+  int what_stride = b->src_stride;
+  int pre_stride = x->e_mbd.pre.y_stride;
+  unsigned char *base_pre = x->e_mbd.pre.y_buffer;
+  unsigned char *in_what;
+  int in_what_stride = pre_stride;
+  int mv_stride = pre_stride;
+  unsigned char *bestaddress;
+  int_mv *best_mv = &d->bmi.mv;
+  int_mv this_mv;
+  unsigned int bestsad;
+  unsigned int thissad;
+  int r, c;
 
-    unsigned char *check_here;
+  unsigned char *check_here;
 
-    int ref_row = ref_mv->as_mv.row;
-    int ref_col = ref_mv->as_mv.col;
+  int ref_row = ref_mv->as_mv.row;
+  int ref_col = ref_mv->as_mv.col;
 
-    int row_min = ref_row - distance;
-    int row_max = ref_row + distance;
-    int col_min = ref_col - distance;
-    int col_max = ref_col + distance;
+  int row_min = ref_row - distance;
+  int row_max = ref_row + distance;
+  int col_min = ref_col - distance;
+  int col_max = ref_col + distance;
 
-    DECLARE_ALIGNED(16, unsigned int, sad_array8[8]);
-    unsigned int sad_array[3];
+  DECLARE_ALIGNED(16, unsigned int, sad_array8[8]);
+  unsigned int sad_array[3];
 
-    int *mvsadcost[2];
-    int_mv fcenter_mv;
+  int *mvsadcost[2];
+  int_mv fcenter_mv;
 
-    mvsadcost[0] = x->mvsadcost[0];
-    mvsadcost[1] = x->mvsadcost[1];
-    fcenter_mv.as_mv.row = center_mv->as_mv.row >> 3;
-    fcenter_mv.as_mv.col = center_mv->as_mv.col >> 3;
+  mvsadcost[0] = x->mvsadcost[0];
+  mvsadcost[1] = x->mvsadcost[1];
+  fcenter_mv.as_mv.row = center_mv->as_mv.row >> 3;
+  fcenter_mv.as_mv.col = center_mv->as_mv.col >> 3;
 
-    /* Work out the mid point for the search */
-    in_what = base_pre + d->offset;
-    bestaddress = in_what + (ref_row * pre_stride) + ref_col;
+  /* Work out the mid point for the search */
+  in_what = base_pre + d->offset;
+  bestaddress = in_what + (ref_row * pre_stride) + ref_col;
 
-    best_mv->as_mv.row = ref_row;
-    best_mv->as_mv.col = ref_col;
+  best_mv->as_mv.row = ref_row;
+  best_mv->as_mv.col = ref_col;
 
-    /* Baseline value at the centre */
-    bestsad = fn_ptr->sdf(what, what_stride, bestaddress, in_what_stride)
-            + mvsad_err_cost(best_mv, &fcenter_mv, mvsadcost, sad_per_bit);
+  /* Baseline value at the centre */
+  bestsad = fn_ptr->sdf(what, what_stride, bestaddress, in_what_stride) +
+            mvsad_err_cost(best_mv, &fcenter_mv, mvsadcost, sad_per_bit);
 
-    /* Apply further limits to prevent us looking using vectors that stretch
-     * beyond the UMV border
-     */
-    if (col_min < x->mv_col_min)
-        col_min = x->mv_col_min;
+  /* Apply further limits to prevent us looking using vectors that stretch
+   * beyond the UMV border
+   */
+  if (col_min < x->mv_col_min) col_min = x->mv_col_min;
 
-    if (col_max > x->mv_col_max)
-        col_max = x->mv_col_max;
+  if (col_max > x->mv_col_max) col_max = x->mv_col_max;
 
-    if (row_min < x->mv_row_min)
-        row_min = x->mv_row_min;
+  if (row_min < x->mv_row_min) row_min = x->mv_row_min;
 
-    if (row_max > x->mv_row_max)
-        row_max = x->mv_row_max;
+  if (row_max > x->mv_row_max) row_max = x->mv_row_max;
 
-    for (r = row_min; r < row_max ; r++)
-    {
-        this_mv.as_mv.row = r;
-        check_here = r * mv_stride + in_what + col_min;
-        c = col_min;
+  for (r = row_min; r < row_max; r++) {
+    this_mv.as_mv.row = r;
+    check_here = r * mv_stride + in_what + col_min;
+    c = col_min;
 
-        while ((c + 7) < col_max)
-        {
-            int i;
+    while ((c + 7) < col_max) {
+      int i;
 
-            fn_ptr->sdx8f(what, what_stride, check_here, in_what_stride, sad_array8);
+      fn_ptr->sdx8f(what, what_stride, check_here, in_what_stride, sad_array8);
 
-            for (i = 0; i < 8; i++)
-            {
-                thissad = sad_array8[i];
+      for (i = 0; i < 8; i++) {
+        thissad = sad_array8[i];
 
-                if (thissad < bestsad)
-                {
-                    this_mv.as_mv.col = c;
-                    thissad += mvsad_err_cost(&this_mv, &fcenter_mv,
-                                              mvsadcost, sad_per_bit);
+        if (thissad < bestsad) {
+          this_mv.as_mv.col = c;
+          thissad +=
+              mvsad_err_cost(&this_mv, &fcenter_mv, mvsadcost, sad_per_bit);
 
-                    if (thissad < bestsad)
-                    {
-                        bestsad = thissad;
-                        best_mv->as_mv.row = r;
-                        best_mv->as_mv.col = c;
-                        bestaddress = check_here;
-                    }
-                }
-
-                check_here++;
-                c++;
-            }
+          if (thissad < bestsad) {
+            bestsad = thissad;
+            best_mv->as_mv.row = r;
+            best_mv->as_mv.col = c;
+            bestaddress = check_here;
+          }
         }
 
-        while ((c + 2) < col_max)
-        {
-            int i;
+        check_here++;
+        c++;
+      }
+    }
 
-            fn_ptr->sdx3f(what, what_stride, check_here , in_what_stride, sad_array);
+    while ((c + 2) < col_max) {
+      int i;
 
-            for (i = 0; i < 3; i++)
-            {
-                thissad = sad_array[i];
+      fn_ptr->sdx3f(what, what_stride, check_here, in_what_stride, sad_array);
 
-                if (thissad < bestsad)
-                {
-                    this_mv.as_mv.col = c;
-                    thissad  += mvsad_err_cost(&this_mv, &fcenter_mv,
-                        mvsadcost, sad_per_bit);
+      for (i = 0; i < 3; i++) {
+        thissad = sad_array[i];
 
-                    if (thissad < bestsad)
-                    {
-                        bestsad = thissad;
-                        best_mv->as_mv.row = r;
-                        best_mv->as_mv.col = c;
-                        bestaddress = check_here;
-                    }
-                }
+        if (thissad < bestsad) {
+          this_mv.as_mv.col = c;
+          thissad +=
+              mvsad_err_cost(&this_mv, &fcenter_mv, mvsadcost, sad_per_bit);
 
-                check_here++;
-                c++;
-            }
+          if (thissad < bestsad) {
+            bestsad = thissad;
+            best_mv->as_mv.row = r;
+            best_mv->as_mv.col = c;
+            bestaddress = check_here;
+          }
         }
 
-        while (c < col_max)
-        {
-            thissad = fn_ptr->sdf(what, what_stride, check_here , in_what_stride);
+        check_here++;
+        c++;
+      }
+    }
 
-            if (thissad < bestsad)
-            {
-                this_mv.as_mv.col = c;
-                thissad  += mvsad_err_cost(&this_mv, &fcenter_mv,
-                    mvsadcost, sad_per_bit);
+    while (c < col_max) {
+      thissad = fn_ptr->sdf(what, what_stride, check_here, in_what_stride);
 
-                if (thissad < bestsad)
-                {
-                    bestsad = thissad;
-                    best_mv->as_mv.row = r;
-                    best_mv->as_mv.col = c;
-                    bestaddress = check_here;
-                }
-            }
+      if (thissad < bestsad) {
+        this_mv.as_mv.col = c;
+        thissad +=
+            mvsad_err_cost(&this_mv, &fcenter_mv, mvsadcost, sad_per_bit);
 
-            check_here ++;
-            c ++;
+        if (thissad < bestsad) {
+          bestsad = thissad;
+          best_mv->as_mv.row = r;
+          best_mv->as_mv.col = c;
+          bestaddress = check_here;
         }
+      }
+
+      check_here++;
+      c++;
     }
+  }
 
-    this_mv.as_mv.row = best_mv->as_mv.row * 8;
-    this_mv.as_mv.col = best_mv->as_mv.col * 8;
+  this_mv.as_mv.row = best_mv->as_mv.row * 8;
+  this_mv.as_mv.col = best_mv->as_mv.col * 8;
 
-    return fn_ptr->vf(what, what_stride, bestaddress, in_what_stride, &thissad)
-           + mv_err_cost(&this_mv, center_mv, mvcost, x->errorperbit);
+  return fn_ptr->vf(what, what_stride, bestaddress, in_what_stride, &thissad) +
+         mv_err_cost(&this_mv, center_mv, mvcost, x->errorperbit);
 }
 
-int vp8_refining_search_sad_c(MACROBLOCK *x, BLOCK *b, BLOCKD *d, int_mv *ref_mv,
-                            int error_per_bit, int search_range,
-                            vp8_variance_fn_ptr_t *fn_ptr, int *mvcost[2],
-                            int_mv *center_mv)
-{
-    MV neighbors[4] = {{-1, 0}, {0, -1}, {0, 1}, {1, 0}};
-    int i, j;
-    short this_row_offset, this_col_offset;
+int vp8_refining_search_sad_c(MACROBLOCK *x, BLOCK *b, BLOCKD *d,
+                              int_mv *ref_mv, int error_per_bit,
+                              int search_range, vp8_variance_fn_ptr_t *fn_ptr,
+                              int *mvcost[2], int_mv *center_mv) {
+  MV neighbors[4] = { { -1, 0 }, { 0, -1 }, { 0, 1 }, { 1, 0 } };
+  int i, j;
+  short this_row_offset, this_col_offset;
 
-    int what_stride = b->src_stride;
-    int pre_stride = x->e_mbd.pre.y_stride;
-    unsigned char *base_pre = x->e_mbd.pre.y_buffer;
-    int in_what_stride = pre_stride;
-    unsigned char *what = (*(b->base_src) + b->src);
-    unsigned char *best_address = (unsigned char *)(base_pre + d->offset +
-        (ref_mv->as_mv.row * pre_stride) + ref_mv->as_mv.col);
-    unsigned char *check_here;
-    int_mv this_mv;
-    unsigned int bestsad;
-    unsigned int thissad;
+  int what_stride = b->src_stride;
+  int pre_stride = x->e_mbd.pre.y_stride;
+  unsigned char *base_pre = x->e_mbd.pre.y_buffer;
+  int in_what_stride = pre_stride;
+  unsigned char *what = (*(b->base_src) + b->src);
+  unsigned char *best_address =
+      (unsigned char *)(base_pre + d->offset +
+                        (ref_mv->as_mv.row * pre_stride) + ref_mv->as_mv.col);
+  unsigned char *check_here;
+  int_mv this_mv;
+  unsigned int bestsad;
+  unsigned int thissad;
 
-    int *mvsadcost[2];
-    int_mv fcenter_mv;
+  int *mvsadcost[2];
+  int_mv fcenter_mv;
 
-    mvsadcost[0] = x->mvsadcost[0];
-    mvsadcost[1] = x->mvsadcost[1];
-    fcenter_mv.as_mv.row = center_mv->as_mv.row >> 3;
-    fcenter_mv.as_mv.col = center_mv->as_mv.col >> 3;
+  mvsadcost[0] = x->mvsadcost[0];
+  mvsadcost[1] = x->mvsadcost[1];
+  fcenter_mv.as_mv.row = center_mv->as_mv.row >> 3;
+  fcenter_mv.as_mv.col = center_mv->as_mv.col >> 3;
 
-    bestsad = fn_ptr->sdf(what, what_stride, best_address, in_what_stride)
-            + mvsad_err_cost(ref_mv, &fcenter_mv, mvsadcost, error_per_bit);
+  bestsad = fn_ptr->sdf(what, what_stride, best_address, in_what_stride) +
+            mvsad_err_cost(ref_mv, &fcenter_mv, mvsadcost, error_per_bit);
 
-    for (i=0; i<search_range; i++)
-    {
-        int best_site = -1;
+  for (i = 0; i < search_range; i++) {
+    int best_site = -1;
 
-        for (j = 0 ; j < 4 ; j++)
-        {
-            this_row_offset = ref_mv->as_mv.row + neighbors[j].row;
-            this_col_offset = ref_mv->as_mv.col + neighbors[j].col;
+    for (j = 0; j < 4; j++) {
+      this_row_offset = ref_mv->as_mv.row + neighbors[j].row;
+      this_col_offset = ref_mv->as_mv.col + neighbors[j].col;
 
-            if ((this_col_offset > x->mv_col_min) && (this_col_offset < x->mv_col_max) &&
-            (this_row_offset > x->mv_row_min) && (this_row_offset < x->mv_row_max))
-            {
-                check_here = (neighbors[j].row)*in_what_stride + neighbors[j].col + best_address;
-                thissad = fn_ptr->sdf(what, what_stride, check_here , in_what_stride);
+      if ((this_col_offset > x->mv_col_min) &&
+          (this_col_offset < x->mv_col_max) &&
+          (this_row_offset > x->mv_row_min) &&
+          (this_row_offset < x->mv_row_max)) {
+        check_here = (neighbors[j].row) * in_what_stride + neighbors[j].col +
+                     best_address;
+        thissad = fn_ptr->sdf(what, what_stride, check_here, in_what_stride);
 
-                if (thissad < bestsad)
-                {
-                    this_mv.as_mv.row = this_row_offset;
-                    this_mv.as_mv.col = this_col_offset;
-                    thissad += mvsad_err_cost(&this_mv, &fcenter_mv, mvsadcost, error_per_bit);
+        if (thissad < bestsad) {
+          this_mv.as_mv.row = this_row_offset;
+          this_mv.as_mv.col = this_col_offset;
+          thissad +=
+              mvsad_err_cost(&this_mv, &fcenter_mv, mvsadcost, error_per_bit);
 
-                    if (thissad < bestsad)
-                    {
-                        bestsad = thissad;
-                        best_site = j;
-                    }
-                }
-            }
+          if (thissad < bestsad) {
+            bestsad = thissad;
+            best_site = j;
+          }
         }
+      }
+    }
 
-        if (best_site == -1)
-            break;
-        else
-        {
-            ref_mv->as_mv.row += neighbors[best_site].row;
-            ref_mv->as_mv.col += neighbors[best_site].col;
-            best_address += (neighbors[best_site].row)*in_what_stride + neighbors[best_site].col;
-        }
+    if (best_site == -1)
+      break;
+    else {
+      ref_mv->as_mv.row += neighbors[best_site].row;
+      ref_mv->as_mv.col += neighbors[best_site].col;
+      best_address += (neighbors[best_site].row) * in_what_stride +
+                      neighbors[best_site].col;
     }
+  }
 
-    this_mv.as_mv.row = ref_mv->as_mv.row << 3;
-    this_mv.as_mv.col = ref_mv->as_mv.col << 3;
+  this_mv.as_mv.row = ref_mv->as_mv.row << 3;
+  this_mv.as_mv.col = ref_mv->as_mv.col << 3;
 
-    return fn_ptr->vf(what, what_stride, best_address, in_what_stride, &thissad)
-           + mv_err_cost(&this_mv, center_mv, mvcost, x->errorperbit);
+  return fn_ptr->vf(what, what_stride, best_address, in_what_stride, &thissad) +
+         mv_err_cost(&this_mv, center_mv, mvcost, x->errorperbit);
 }
 
 int vp8_refining_search_sadx4(MACROBLOCK *x, BLOCK *b, BLOCKD *d,
                               int_mv *ref_mv, int error_per_bit,
                               int search_range, vp8_variance_fn_ptr_t *fn_ptr,
-                              int *mvcost[2], int_mv *center_mv)
-{
-    MV neighbors[4] = {{-1, 0}, {0, -1}, {0, 1}, {1, 0}};
-    int i, j;
-    short this_row_offset, this_col_offset;
+                              int *mvcost[2], int_mv *center_mv) {
+  MV neighbors[4] = { { -1, 0 }, { 0, -1 }, { 0, 1 }, { 1, 0 } };
+  int i, j;
+  short this_row_offset, this_col_offset;
 
-    int what_stride = b->src_stride;
-    int pre_stride = x->e_mbd.pre.y_stride;
-    unsigned char *base_pre = x->e_mbd.pre.y_buffer;
-    int in_what_stride = pre_stride;
-    unsigned char *what = (*(b->base_src) + b->src);
-    unsigned char *best_address = (unsigned char *)(base_pre + d->offset +
-        (ref_mv->as_mv.row * pre_stride) + ref_mv->as_mv.col);
-    unsigned char *check_here;
-    int_mv this_mv;
-    unsigned int bestsad;
-    unsigned int thissad;
+  int what_stride = b->src_stride;
+  int pre_stride = x->e_mbd.pre.y_stride;
+  unsigned char *base_pre = x->e_mbd.pre.y_buffer;
+  int in_what_stride = pre_stride;
+  unsigned char *what = (*(b->base_src) + b->src);
+  unsigned char *best_address =
+      (unsigned char *)(base_pre + d->offset +
+                        (ref_mv->as_mv.row * pre_stride) + ref_mv->as_mv.col);
+  unsigned char *check_here;
+  int_mv this_mv;
+  unsigned int bestsad;
+  unsigned int thissad;
 
-    int *mvsadcost[2];
-    int_mv fcenter_mv;
+  int *mvsadcost[2];
+  int_mv fcenter_mv;
 
-    mvsadcost[0] = x->mvsadcost[0];
-    mvsadcost[1] = x->mvsadcost[1];
-    fcenter_mv.as_mv.row = center_mv->as_mv.row >> 3;
-    fcenter_mv.as_mv.col = center_mv->as_mv.col >> 3;
+  mvsadcost[0] = x->mvsadcost[0];
+  mvsadcost[1] = x->mvsadcost[1];
+  fcenter_mv.as_mv.row = center_mv->as_mv.row >> 3;
+  fcenter_mv.as_mv.col = center_mv->as_mv.col >> 3;
 
-    bestsad = fn_ptr->sdf(what, what_stride, best_address, in_what_stride)
-            + mvsad_err_cost(ref_mv, &fcenter_mv, mvsadcost, error_per_bit);
+  bestsad = fn_ptr->sdf(what, what_stride, best_address, in_what_stride) +
+            mvsad_err_cost(ref_mv, &fcenter_mv, mvsadcost, error_per_bit);
 
-    for (i=0; i<search_range; i++)
-    {
-        int best_site = -1;
-        int all_in = 1;
+  for (i = 0; i < search_range; i++) {
+    int best_site = -1;
+    int all_in = 1;
 
-        all_in &= ((ref_mv->as_mv.row - 1) > x->mv_row_min);
-        all_in &= ((ref_mv->as_mv.row + 1) < x->mv_row_max);
-        all_in &= ((ref_mv->as_mv.col - 1) > x->mv_col_min);
-        all_in &= ((ref_mv->as_mv.col + 1) < x->mv_col_max);
+    all_in &= ((ref_mv->as_mv.row - 1) > x->mv_row_min);
+    all_in &= ((ref_mv->as_mv.row + 1) < x->mv_row_max);
+    all_in &= ((ref_mv->as_mv.col - 1) > x->mv_col_min);
+    all_in &= ((ref_mv->as_mv.col + 1) < x->mv_col_max);
 
-        if(all_in)
-        {
-            unsigned int sad_array[4];
-            const unsigned char *block_offset[4];
-            block_offset[0] = best_address - in_what_stride;
-            block_offset[1] = best_address - 1;
-            block_offset[2] = best_address + 1;
-            block_offset[3] = best_address + in_what_stride;
+    if (all_in) {
+      unsigned int sad_array[4];
+      const unsigned char *block_offset[4];
+      block_offset[0] = best_address - in_what_stride;
+      block_offset[1] = best_address - 1;
+      block_offset[2] = best_address + 1;
+      block_offset[3] = best_address + in_what_stride;
 
-            fn_ptr->sdx4df(what, what_stride, block_offset, in_what_stride, sad_array);
+      fn_ptr->sdx4df(what, what_stride, block_offset, in_what_stride,
+                     sad_array);
 
-            for (j = 0; j < 4; j++)
-            {
-                if (sad_array[j] < bestsad)
-                {
-                    this_mv.as_mv.row = ref_mv->as_mv.row + neighbors[j].row;
-                    this_mv.as_mv.col = ref_mv->as_mv.col + neighbors[j].col;
-                    sad_array[j] += mvsad_err_cost(&this_mv, &fcenter_mv, mvsadcost, error_per_bit);
+      for (j = 0; j < 4; j++) {
+        if (sad_array[j] < bestsad) {
+          this_mv.as_mv.row = ref_mv->as_mv.row + neighbors[j].row;
+          this_mv.as_mv.col = ref_mv->as_mv.col + neighbors[j].col;
+          sad_array[j] +=
+              mvsad_err_cost(&this_mv, &fcenter_mv, mvsadcost, error_per_bit);
 
-                    if (sad_array[j] < bestsad)
-                    {
-                        bestsad = sad_array[j];
-                        best_site = j;
-                    }
-                }
-            }
+          if (sad_array[j] < bestsad) {
+            bestsad = sad_array[j];
+            best_site = j;
+          }
         }
-        else
-        {
-            for (j = 0 ; j < 4 ; j++)
-            {
-                this_row_offset = ref_mv->as_mv.row + neighbors[j].row;
-                this_col_offset = ref_mv->as_mv.col + neighbors[j].col;
+      }
+    } else {
+      for (j = 0; j < 4; j++) {
+        this_row_offset = ref_mv->as_mv.row + neighbors[j].row;
+        this_col_offset = ref_mv->as_mv.col + neighbors[j].col;
 
-                if ((this_col_offset > x->mv_col_min) && (this_col_offset < x->mv_col_max) &&
-                (this_row_offset > x->mv_row_min) && (this_row_offset < x->mv_row_max))
-                {
-                    check_here = (neighbors[j].row)*in_what_stride + neighbors[j].col + best_address;
-                    thissad = fn_ptr->sdf(what, what_stride, check_here , in_what_stride);
+        if ((this_col_offset > x->mv_col_min) &&
+            (this_col_offset < x->mv_col_max) &&
+            (this_row_offset > x->mv_row_min) &&
+            (this_row_offset < x->mv_row_max)) {
+          check_here = (neighbors[j].row) * in_what_stride + neighbors[j].col +
+                       best_address;
+          thissad = fn_ptr->sdf(what, what_stride, check_here, in_what_stride);
 
-                    if (thissad < bestsad)
-                    {
-                        this_mv.as_mv.row = this_row_offset;
-                        this_mv.as_mv.col = this_col_offset;
-                        thissad += mvsad_err_cost(&this_mv, &fcenter_mv, mvsadcost, error_per_bit);
+          if (thissad < bestsad) {
+            this_mv.as_mv.row = this_row_offset;
+            this_mv.as_mv.col = this_col_offset;
+            thissad +=
+                mvsad_err_cost(&this_mv, &fcenter_mv, mvsadcost, error_per_bit);
 
-                        if (thissad < bestsad)
-                        {
-                            bestsad = thissad;
-                            best_site = j;
-                        }
-                    }
-                }
+            if (thissad < bestsad) {
+              bestsad = thissad;
+              best_site = j;
             }
+          }
         }
+      }
+    }
 
-        if (best_site == -1)
-            break;
-        else
-        {
-            ref_mv->as_mv.row += neighbors[best_site].row;
-            ref_mv->as_mv.col += neighbors[best_site].col;
-            best_address += (neighbors[best_site].row)*in_what_stride + neighbors[best_site].col;
-        }
+    if (best_site == -1)
+      break;
+    else {
+      ref_mv->as_mv.row += neighbors[best_site].row;
+      ref_mv->as_mv.col += neighbors[best_site].col;
+      best_address += (neighbors[best_site].row) * in_what_stride +
+                      neighbors[best_site].col;
     }
+  }
 
-    this_mv.as_mv.row = ref_mv->as_mv.row * 8;
-    this_mv.as_mv.col = ref_mv->as_mv.col * 8;
+  this_mv.as_mv.row = ref_mv->as_mv.row * 8;
+  this_mv.as_mv.col = ref_mv->as_mv.col * 8;
 
-    return fn_ptr->vf(what, what_stride, best_address, in_what_stride, &thissad)
-           + mv_err_cost(&this_mv, center_mv, mvcost, x->errorperbit);
+  return fn_ptr->vf(what, what_stride, best_address, in_what_stride, &thissad) +
+         mv_err_cost(&this_mv, center_mv, mvcost, x->errorperbit);
 }
 
 #ifdef VP8_ENTROPY_STATS
-void print_mode_context(void)
-{
-    FILE *f = fopen("modecont.c", "w");
-    int i, j;
+void print_mode_context(void) {
+  FILE *f = fopen("modecont.c", "w");
+  int i, j;
 
-    fprintf(f, "#include \"entropy.h\"\n");
-    fprintf(f, "const int vp8_mode_contexts[6][4] =\n");
-    fprintf(f, "{\n");
+  fprintf(f, "#include \"entropy.h\"\n");
+  fprintf(f, "const int vp8_mode_contexts[6][4] =\n");
+  fprintf(f, "{\n");
 
-    for (j = 0; j < 6; j++)
-    {
-        fprintf(f, "  { /* %d */\n", j);
-        fprintf(f, "    ");
+  for (j = 0; j < 6; j++) {
+    fprintf(f, "  { /* %d */\n", j);
+    fprintf(f, "    ");
 
-        for (i = 0; i < 4; i++)
-        {
-            int overal_prob;
-            int this_prob;
-            int count;
+    for (i = 0; i < 4; i++) {
+      int overal_prob;
+      int this_prob;
+      int count;
 
-            /* Overall probs */
-            count = mv_mode_cts[i][0] + mv_mode_cts[i][1];
+      /* Overall probs */
+      count = mv_mode_cts[i][0] + mv_mode_cts[i][1];
 
-            if (count)
-                overal_prob = 256 * mv_mode_cts[i][0] / count;
-            else
-                overal_prob = 128;
+      if (count)
+        overal_prob = 256 * mv_mode_cts[i][0] / count;
+      else
+        overal_prob = 128;
 
-            if (overal_prob == 0)
-                overal_prob = 1;
+      if (overal_prob == 0) overal_prob = 1;
 
-            /* context probs */
-            count = mv_ref_ct[j][i][0] + mv_ref_ct[j][i][1];
+      /* context probs */
+      count = mv_ref_ct[j][i][0] + mv_ref_ct[j][i][1];
 
-            if (count)
-                this_prob = 256 * mv_ref_ct[j][i][0] / count;
-            else
-                this_prob = 128;
+      if (count)
+        this_prob = 256 * mv_ref_ct[j][i][0] / count;
+      else
+        this_prob = 128;
 
-            if (this_prob == 0)
-                this_prob = 1;
+      if (this_prob == 0) this_prob = 1;
 
-            fprintf(f, "%5d, ", this_prob);
-        }
-
-        fprintf(f, "  },\n");
+      fprintf(f, "%5d, ", this_prob);
     }
 
-    fprintf(f, "};\n");
-    fclose(f);
+    fprintf(f, "  },\n");
+  }
+
+  fprintf(f, "};\n");
+  fclose(f);
 }
 
 /* MV ref count VP8_ENTROPY_STATS stats code */
 #ifdef VP8_ENTROPY_STATS
-void init_mv_ref_counts()
-{
-    memset(mv_ref_ct, 0, sizeof(mv_ref_ct));
-    memset(mv_mode_cts, 0, sizeof(mv_mode_cts));
+void init_mv_ref_counts() {
+  memset(mv_ref_ct, 0, sizeof(mv_ref_ct));
+  memset(mv_mode_cts, 0, sizeof(mv_mode_cts));
 }
 
-void accum_mv_refs(MB_PREDICTION_MODE m, const int ct[4])
-{
-    if (m == ZEROMV)
-    {
-        ++mv_ref_ct [ct[0]] [0] [0];
-        ++mv_mode_cts[0][0];
-    }
-    else
-    {
-        ++mv_ref_ct [ct[0]] [0] [1];
-        ++mv_mode_cts[0][1];
+void accum_mv_refs(MB_PREDICTION_MODE m, const int ct[4]) {
+  if (m == ZEROMV) {
+    ++mv_ref_ct[ct[0]][0][0];
+    ++mv_mode_cts[0][0];
+  } else {
+    ++mv_ref_ct[ct[0]][0][1];
+    ++mv_mode_cts[0][1];
 
-        if (m == NEARESTMV)
-        {
-            ++mv_ref_ct [ct[1]] [1] [0];
-            ++mv_mode_cts[1][0];
-        }
-        else
-        {
-            ++mv_ref_ct [ct[1]] [1] [1];
-            ++mv_mode_cts[1][1];
+    if (m == NEARESTMV) {
+      ++mv_ref_ct[ct[1]][1][0];
+      ++mv_mode_cts[1][0];
+    } else {
+      ++mv_ref_ct[ct[1]][1][1];
+      ++mv_mode_cts[1][1];
 
-            if (m == NEARMV)
-            {
-                ++mv_ref_ct [ct[2]] [2] [0];
-                ++mv_mode_cts[2][0];
-            }
-            else
-            {
-                ++mv_ref_ct [ct[2]] [2] [1];
-                ++mv_mode_cts[2][1];
+      if (m == NEARMV) {
+        ++mv_ref_ct[ct[2]][2][0];
+        ++mv_mode_cts[2][0];
+      } else {
+        ++mv_ref_ct[ct[2]][2][1];
+        ++mv_mode_cts[2][1];
 
-                if (m == NEWMV)
-                {
-                    ++mv_ref_ct [ct[3]] [3] [0];
-                    ++mv_mode_cts[3][0];
-                }
-                else
-                {
-                    ++mv_ref_ct [ct[3]] [3] [1];
-                    ++mv_mode_cts[3][1];
-                }
-            }
+        if (m == NEWMV) {
+          ++mv_ref_ct[ct[3]][3][0];
+          ++mv_mode_cts[3][0];
+        } else {
+          ++mv_ref_ct[ct[3]][3][1];
+          ++mv_mode_cts[3][1];
         }
+      }
     }
+  }
 }
 
-#endif/* END MV ref count VP8_ENTROPY_STATS stats code */
+#endif /* END MV ref count VP8_ENTROPY_STATS stats code */
 
 #endif
--- a/vp8/encoder/mcomp.h
+++ b/vp8/encoder/mcomp.h
@@ -8,7 +8,6 @@
  *  be found in the AUTHORS file in the root of the source tree.
  */
 
-
 #ifndef VP8_ENCODER_MCOMP_H_
 #define VP8_ENCODER_MCOMP_H_
 
@@ -24,7 +23,6 @@
 extern void accum_mv_refs(MB_PREDICTION_MODE, const int near_mv_ref_cts[4]);
 #endif
 
-
 /* The maximum number of steps in a step search given the largest allowed
  * initial step
  */
@@ -34,79 +32,47 @@
 #define MAX_FULL_PEL_VAL ((1 << (MAX_MVSEARCH_STEPS)) - 1)
 
 /* Maximum size of the first step in full pel units */
-#define MAX_FIRST_STEP (1 << (MAX_MVSEARCH_STEPS-1))
+#define MAX_FIRST_STEP (1 << (MAX_MVSEARCH_STEPS - 1))
 
 extern void print_mode_context(void);
 extern int vp8_mv_bit_cost(int_mv *mv, int_mv *ref, int *mvcost[2], int Weight);
 extern void vp8_init_dsmotion_compensation(MACROBLOCK *x, int stride);
-extern void vp8_init3smotion_compensation(MACROBLOCK *x,  int stride);
+extern void vp8_init3smotion_compensation(MACROBLOCK *x, int stride);
 
+extern int vp8_hex_search(MACROBLOCK *x, BLOCK *b, BLOCKD *d, int_mv *ref_mv,
+                          int_mv *best_mv, int search_param, int error_per_bit,
+                          const vp8_variance_fn_ptr_t *vf, int *mvsadcost[2],
+                          int *mvcost[2], int_mv *center_mv);
 
-extern int vp8_hex_search
-(
-    MACROBLOCK *x,
-    BLOCK *b,
-    BLOCKD *d,
-    int_mv *ref_mv,
-    int_mv *best_mv,
-    int search_param,
-    int error_per_bit,
-    const vp8_variance_fn_ptr_t *vf,
-    int *mvsadcost[2],
-    int *mvcost[2],
-    int_mv *center_mv
-);
+typedef int(fractional_mv_step_fp)(MACROBLOCK *x, BLOCK *b, BLOCKD *d,
+                                   int_mv *bestmv, int_mv *ref_mv,
+                                   int error_per_bit,
+                                   const vp8_variance_fn_ptr_t *vfp,
+                                   int *mvcost[2], int *distortion,
+                                   unsigned int *sse);
 
-typedef int (fractional_mv_step_fp)
-    (MACROBLOCK *x, BLOCK *b, BLOCKD *d, int_mv *bestmv, int_mv *ref_mv,
-     int error_per_bit, const vp8_variance_fn_ptr_t *vfp, int *mvcost[2],
-     int *distortion, unsigned int *sse);
-
 extern fractional_mv_step_fp vp8_find_best_sub_pixel_step_iteratively;
 extern fractional_mv_step_fp vp8_find_best_sub_pixel_step;
 extern fractional_mv_step_fp vp8_find_best_half_pixel_step;
 extern fractional_mv_step_fp vp8_skip_fractional_mv_step;
 
-typedef int (*vp8_full_search_fn_t)
-    (
-     MACROBLOCK *x,
-     BLOCK *b,
-     BLOCKD *d,
-     int_mv *ref_mv,
-     int sad_per_bit,
-     int distance,
-     vp8_variance_fn_ptr_t *fn_ptr,
-     int *mvcost[2],
-     int_mv *center_mv
-    );
+typedef int (*vp8_full_search_fn_t)(MACROBLOCK *x, BLOCK *b, BLOCKD *d,
+                                    int_mv *ref_mv, int sad_per_bit,
+                                    int distance, vp8_variance_fn_ptr_t *fn_ptr,
+                                    int *mvcost[2], int_mv *center_mv);
 
-typedef int (*vp8_refining_search_fn_t)
-    (
-     MACROBLOCK *x,
-     BLOCK *b,
-     BLOCKD *d,
-     int_mv *ref_mv,
-     int sad_per_bit,
-     int distance,
-     vp8_variance_fn_ptr_t *fn_ptr,
-     int *mvcost[2],
-     int_mv *center_mv
-    );
+typedef int (*vp8_refining_search_fn_t)(MACROBLOCK *x, BLOCK *b, BLOCKD *d,
+                                        int_mv *ref_mv, int sad_per_bit,
+                                        int distance,
+                                        vp8_variance_fn_ptr_t *fn_ptr,
+                                        int *mvcost[2], int_mv *center_mv);
 
-typedef int (*vp8_diamond_search_fn_t)
-    (
-     MACROBLOCK *x,
-     BLOCK *b,
-     BLOCKD *d,
-     int_mv *ref_mv,
-     int_mv *best_mv,
-     int search_param,
-     int sad_per_bit,
-     int *num00,
-     vp8_variance_fn_ptr_t *fn_ptr,
-     int *mvcost[2],
-     int_mv *center_mv
-    );
+typedef int (*vp8_diamond_search_fn_t)(MACROBLOCK *x, BLOCK *b, BLOCKD *d,
+                                       int_mv *ref_mv, int_mv *best_mv,
+                                       int search_param, int sad_per_bit,
+                                       int *num00,
+                                       vp8_variance_fn_ptr_t *fn_ptr,
+                                       int *mvcost[2], int_mv *center_mv);
 
 #ifdef __cplusplus
 }  // extern "C"
--- a/vp8/encoder/mips/msa/dct_msa.c
+++ b/vp8/encoder/mips/msa/dct_msa.c
@@ -11,189 +11,186 @@
 #include "./vp8_rtcd.h"
 #include "vp8/common/mips/msa/vp8_macros_msa.h"
 
-#define TRANSPOSE4x4_H(in0, in1, in2, in3, out0, out1, out2, out3)  \
-{                                                                   \
-    v8i16 s0_m, s1_m, tp0_m, tp1_m, tp2_m, tp3_m;                   \
-                                                                    \
-    ILVR_H2_SH(in2, in0, in3, in1, s0_m, s1_m);                     \
-    ILVRL_H2_SH(s1_m, s0_m, tp0_m, tp1_m);                          \
-    ILVL_H2_SH(in2, in0, in3, in1, s0_m, s1_m);                     \
-    ILVRL_H2_SH(s1_m, s0_m, tp2_m, tp3_m);                          \
-    PCKEV_D2_SH(tp2_m, tp0_m, tp3_m, tp1_m, out0, out2);            \
-    PCKOD_D2_SH(tp2_m, tp0_m, tp3_m, tp1_m, out1, out3);            \
-}
+#define TRANSPOSE4x4_H(in0, in1, in2, in3, out0, out1, out2, out3) \
+  {                                                                \
+    v8i16 s0_m, s1_m, tp0_m, tp1_m, tp2_m, tp3_m;                  \
+                                                                   \
+    ILVR_H2_SH(in2, in0, in3, in1, s0_m, s1_m);                    \
+    ILVRL_H2_SH(s1_m, s0_m, tp0_m, tp1_m);                         \
+    ILVL_H2_SH(in2, in0, in3, in1, s0_m, s1_m);                    \
+    ILVRL_H2_SH(s1_m, s0_m, tp2_m, tp3_m);                         \
+    PCKEV_D2_SH(tp2_m, tp0_m, tp3_m, tp1_m, out0, out2);           \
+    PCKOD_D2_SH(tp2_m, tp0_m, tp3_m, tp1_m, out1, out3);           \
+  }
 
-#define SET_DOTP_VALUES(coeff, val0, val1, val2, const1, const2)    \
-{                                                                   \
-    v8i16 tmp0_m;                                                   \
-                                                                    \
-    SPLATI_H3_SH(coeff, val0, val1, val2, tmp0_m, const1, const2);  \
-    ILVEV_H2_SH(tmp0_m, const1, const2, tmp0_m, const1, const2);    \
-}
+#define SET_DOTP_VALUES(coeff, val0, val1, val2, const1, const2)   \
+  {                                                                \
+    v8i16 tmp0_m;                                                  \
+                                                                   \
+    SPLATI_H3_SH(coeff, val0, val1, val2, tmp0_m, const1, const2); \
+    ILVEV_H2_SH(tmp0_m, const1, const2, tmp0_m, const1, const2);   \
+  }
 
-#define RET_1_IF_NZERO_H(in0)       \
-({                                  \
-    v8i16 tmp0_m;                   \
-    v8i16 one_m = __msa_ldi_h(1);   \
-                                    \
-    tmp0_m = __msa_ceqi_h(in0, 0);  \
-    tmp0_m = tmp0_m ^ 255;          \
-    tmp0_m = one_m & tmp0_m;        \
-                                    \
-    tmp0_m;                         \
-})
+#define RET_1_IF_NZERO_H(in0)      \
+  ({                               \
+    v8i16 tmp0_m;                  \
+    v8i16 one_m = __msa_ldi_h(1);  \
+                                   \
+    tmp0_m = __msa_ceqi_h(in0, 0); \
+    tmp0_m = tmp0_m ^ 255;         \
+    tmp0_m = one_m & tmp0_m;       \
+                                   \
+    tmp0_m;                        \
+  })
 
-#define RET_1_IF_NZERO_W(in0)       \
-({                                  \
-    v4i32 tmp0_m;                   \
-    v4i32 one_m = __msa_ldi_w(1);   \
-                                    \
-    tmp0_m = __msa_ceqi_w(in0, 0);  \
-    tmp0_m = tmp0_m ^ 255;          \
-    tmp0_m = one_m & tmp0_m;        \
-                                    \
-    tmp0_m;                         \
-})
+#define RET_1_IF_NZERO_W(in0)      \
+  ({                               \
+    v4i32 tmp0_m;                  \
+    v4i32 one_m = __msa_ldi_w(1);  \
+                                   \
+    tmp0_m = __msa_ceqi_w(in0, 0); \
+    tmp0_m = tmp0_m ^ 255;         \
+    tmp0_m = one_m & tmp0_m;       \
+                                   \
+    tmp0_m;                        \
+  })
 
-#define RET_1_IF_NEG_W(in0)           \
-({                                    \
-    v4i32 tmp0_m;                     \
-                                      \
-    v4i32 one_m = __msa_ldi_w(1);     \
-    tmp0_m = __msa_clti_s_w(in0, 0);  \
-    tmp0_m = one_m & tmp0_m;          \
-                                      \
-    tmp0_m;                           \
-})
+#define RET_1_IF_NEG_W(in0)          \
+  ({                                 \
+    v4i32 tmp0_m;                    \
+                                     \
+    v4i32 one_m = __msa_ldi_w(1);    \
+    tmp0_m = __msa_clti_s_w(in0, 0); \
+    tmp0_m = one_m & tmp0_m;         \
+                                     \
+    tmp0_m;                          \
+  })
 
-void vp8_short_fdct4x4_msa(int16_t *input, int16_t *output, int32_t pitch)
-{
-    v8i16 in0, in1, in2, in3;
-    v8i16 temp0, temp1;
-    v8i16 const0, const1;
-    v8i16 coeff = { 2217, 5352, -5352, 14500, 7500, 12000, 25000, 26000 };
-    v4i32 out0, out1, out2, out3;
-    v8i16 zero = { 0 };
+void vp8_short_fdct4x4_msa(int16_t *input, int16_t *output, int32_t pitch) {
+  v8i16 in0, in1, in2, in3;
+  v8i16 temp0, temp1;
+  v8i16 const0, const1;
+  v8i16 coeff = { 2217, 5352, -5352, 14500, 7500, 12000, 25000, 26000 };
+  v4i32 out0, out1, out2, out3;
+  v8i16 zero = { 0 };
 
-    LD_SH4(input, pitch / 2, in0, in1, in2, in3);
-    TRANSPOSE4x4_SH_SH(in0, in1, in2, in3, in0, in1, in2, in3);
+  LD_SH4(input, pitch / 2, in0, in1, in2, in3);
+  TRANSPOSE4x4_SH_SH(in0, in1, in2, in3, in0, in1, in2, in3);
 
-    BUTTERFLY_4(in0, in1, in2, in3, temp0, temp1, in1, in3);
-    SLLI_4V(temp0, temp1, in1, in3, 3);
-    in0 = temp0 + temp1;
-    in2 = temp0 - temp1;
-    SET_DOTP_VALUES(coeff, 0, 1, 2, const0, const1);
-    temp0 = __msa_ilvr_h(in3, in1);
-    in1 = __msa_splati_h(coeff, 3);
-    out0 = (v4i32)__msa_ilvev_h(zero, in1);
-    coeff = __msa_ilvl_h(zero, coeff);
-    out1 = __msa_splati_w((v4i32)coeff, 0);
-    DPADD_SH2_SW(temp0, temp0, const0, const1, out0, out1);
-    out0 >>= 12;
-    out1 >>= 12;
-    PCKEV_H2_SH(out0, out0, out1, out1, in1, in3);
-    TRANSPOSE4x4_SH_SH(in0, in1, in2, in3, in0, in1, in2, in3);
+  BUTTERFLY_4(in0, in1, in2, in3, temp0, temp1, in1, in3);
+  SLLI_4V(temp0, temp1, in1, in3, 3);
+  in0 = temp0 + temp1;
+  in2 = temp0 - temp1;
+  SET_DOTP_VALUES(coeff, 0, 1, 2, const0, const1);
+  temp0 = __msa_ilvr_h(in3, in1);
+  in1 = __msa_splati_h(coeff, 3);
+  out0 = (v4i32)__msa_ilvev_h(zero, in1);
+  coeff = __msa_ilvl_h(zero, coeff);
+  out1 = __msa_splati_w((v4i32)coeff, 0);
+  DPADD_SH2_SW(temp0, temp0, const0, const1, out0, out1);
+  out0 >>= 12;
+  out1 >>= 12;
+  PCKEV_H2_SH(out0, out0, out1, out1, in1, in3);
+  TRANSPOSE4x4_SH_SH(in0, in1, in2, in3, in0, in1, in2, in3);
 
-    BUTTERFLY_4(in0, in1, in2, in3, temp0, temp1, in1, in3);
-    in0 = temp0 + temp1 + 7;
-    in2 = temp0 - temp1 + 7;
-    in0 >>= 4;
-    in2 >>= 4;
-    ILVR_H2_SW(zero, in0, zero, in2, out0, out2);
-    temp1 = RET_1_IF_NZERO_H(in3);
-    ILVR_H2_SH(zero, temp1, in3, in1, temp1, temp0);
-    SPLATI_W2_SW(coeff, 2, out3, out1);
-    out3 += out1;
-    out1 = __msa_splati_w((v4i32)coeff, 1);
-    DPADD_SH2_SW(temp0, temp0, const0, const1, out1, out3);
-    out1 >>= 16;
-    out3 >>= 16;
-    out1 += (v4i32)temp1;
-    PCKEV_H2_SH(out1, out0, out3, out2, in0, in2);
-    ST_SH2(in0, in2, output, 8);
+  BUTTERFLY_4(in0, in1, in2, in3, temp0, temp1, in1, in3);
+  in0 = temp0 + temp1 + 7;
+  in2 = temp0 - temp1 + 7;
+  in0 >>= 4;
+  in2 >>= 4;
+  ILVR_H2_SW(zero, in0, zero, in2, out0, out2);
+  temp1 = RET_1_IF_NZERO_H(in3);
+  ILVR_H2_SH(zero, temp1, in3, in1, temp1, temp0);
+  SPLATI_W2_SW(coeff, 2, out3, out1);
+  out3 += out1;
+  out1 = __msa_splati_w((v4i32)coeff, 1);
+  DPADD_SH2_SW(temp0, temp0, const0, const1, out1, out3);
+  out1 >>= 16;
+  out3 >>= 16;
+  out1 += (v4i32)temp1;
+  PCKEV_H2_SH(out1, out0, out3, out2, in0, in2);
+  ST_SH2(in0, in2, output, 8);
 }
 
-void vp8_short_fdct8x4_msa(int16_t *input, int16_t *output, int32_t pitch)
-{
-    v8i16 in0, in1, in2, in3;
-    v8i16 temp0, temp1, tmp0, tmp1;
-    v8i16 const0, const1, const2;
-    v8i16 coeff = { 2217, 5352, -5352, 14500, 7500, 12000, 25000, 26000 };
-    v8i16 zero = { 0 };
-    v4i32 vec0_w, vec1_w, vec2_w, vec3_w;
+void vp8_short_fdct8x4_msa(int16_t *input, int16_t *output, int32_t pitch) {
+  v8i16 in0, in1, in2, in3;
+  v8i16 temp0, temp1, tmp0, tmp1;
+  v8i16 const0, const1, const2;
+  v8i16 coeff = { 2217, 5352, -5352, 14500, 7500, 12000, 25000, 26000 };
+  v8i16 zero = { 0 };
+  v4i32 vec0_w, vec1_w, vec2_w, vec3_w;
 
-    LD_SH4(input, pitch / 2, in0, in1, in2, in3);
-    TRANSPOSE4x4_H(in0, in1, in2, in3, in0, in1, in2, in3);
+  LD_SH4(input, pitch / 2, in0, in1, in2, in3);
+  TRANSPOSE4x4_H(in0, in1, in2, in3, in0, in1, in2, in3);
 
-    BUTTERFLY_4(in0, in1, in2, in3, temp0, temp1, in1, in3);
-    SLLI_4V(temp0, temp1, in1, in3, 3);
-    in0 = temp0 + temp1;
-    in2 = temp0 - temp1;
-    SET_DOTP_VALUES(coeff, 0, 1, 2, const1, const2);
-    temp0 = __msa_splati_h(coeff, 3);
-    vec1_w = (v4i32)__msa_ilvev_h(zero, temp0);
-    coeff = __msa_ilvl_h(zero, coeff);
-    vec3_w = __msa_splati_w((v4i32)coeff, 0);
-    ILVRL_H2_SH(in3, in1, tmp1, tmp0);
-    vec0_w = vec1_w;
-    vec2_w = vec3_w;
-    DPADD_SH4_SW(tmp1, tmp0, tmp1, tmp0, const1, const1, const2, const2,
-                 vec0_w, vec1_w, vec2_w, vec3_w);
-    SRA_4V(vec1_w, vec0_w, vec3_w, vec2_w, 12);
-    PCKEV_H2_SH(vec1_w, vec0_w, vec3_w, vec2_w, in1, in3);
-    TRANSPOSE4x4_H(in0, in1, in2, in3, in0, in1, in2, in3);
+  BUTTERFLY_4(in0, in1, in2, in3, temp0, temp1, in1, in3);
+  SLLI_4V(temp0, temp1, in1, in3, 3);
+  in0 = temp0 + temp1;
+  in2 = temp0 - temp1;
+  SET_DOTP_VALUES(coeff, 0, 1, 2, const1, const2);
+  temp0 = __msa_splati_h(coeff, 3);
+  vec1_w = (v4i32)__msa_ilvev_h(zero, temp0);
+  coeff = __msa_ilvl_h(zero, coeff);
+  vec3_w = __msa_splati_w((v4i32)coeff, 0);
+  ILVRL_H2_SH(in3, in1, tmp1, tmp0);
+  vec0_w = vec1_w;
+  vec2_w = vec3_w;
+  DPADD_SH4_SW(tmp1, tmp0, tmp1, tmp0, const1, const1, const2, const2, vec0_w,
+               vec1_w, vec2_w, vec3_w);
+  SRA_4V(vec1_w, vec0_w, vec3_w, vec2_w, 12);
+  PCKEV_H2_SH(vec1_w, vec0_w, vec3_w, vec2_w, in1, in3);
+  TRANSPOSE4x4_H(in0, in1, in2, in3, in0, in1, in2, in3);
 
-    BUTTERFLY_4(in0, in1, in2, in3, temp0, temp1, in1, in3);
-    in0 = temp0 + temp1 + 7;
-    in2 = temp0 - temp1 + 7;
-    in0 >>= 4;
-    in2 >>= 4;
-    SPLATI_W2_SW(coeff, 2, vec3_w, vec1_w);
-    vec3_w += vec1_w;
-    vec1_w = __msa_splati_w((v4i32)coeff, 1);
-    const0 = RET_1_IF_NZERO_H(in3);
-    ILVRL_H2_SH(in3, in1, tmp1, tmp0);
-    vec0_w = vec1_w;
-    vec2_w = vec3_w;
-    DPADD_SH4_SW(tmp1, tmp0, tmp1, tmp0, const1, const1, const2, const2,
-                 vec0_w, vec1_w, vec2_w, vec3_w);
-    SRA_4V(vec1_w, vec0_w, vec3_w, vec2_w, 16);
-    PCKEV_H2_SH(vec1_w, vec0_w, vec3_w, vec2_w, in1, in3);
-    in1 += const0;
-    PCKEV_D2_SH(in1, in0, in3, in2, temp0, temp1);
-    ST_SH2(temp0, temp1, output, 8);
+  BUTTERFLY_4(in0, in1, in2, in3, temp0, temp1, in1, in3);
+  in0 = temp0 + temp1 + 7;
+  in2 = temp0 - temp1 + 7;
+  in0 >>= 4;
+  in2 >>= 4;
+  SPLATI_W2_SW(coeff, 2, vec3_w, vec1_w);
+  vec3_w += vec1_w;
+  vec1_w = __msa_splati_w((v4i32)coeff, 1);
+  const0 = RET_1_IF_NZERO_H(in3);
+  ILVRL_H2_SH(in3, in1, tmp1, tmp0);
+  vec0_w = vec1_w;
+  vec2_w = vec3_w;
+  DPADD_SH4_SW(tmp1, tmp0, tmp1, tmp0, const1, const1, const2, const2, vec0_w,
+               vec1_w, vec2_w, vec3_w);
+  SRA_4V(vec1_w, vec0_w, vec3_w, vec2_w, 16);
+  PCKEV_H2_SH(vec1_w, vec0_w, vec3_w, vec2_w, in1, in3);
+  in1 += const0;
+  PCKEV_D2_SH(in1, in0, in3, in2, temp0, temp1);
+  ST_SH2(temp0, temp1, output, 8);
 
-    PCKOD_D2_SH(in1, in0, in3, in2, in0, in2);
-    ST_SH2(in0, in2, output + 16, 8);
+  PCKOD_D2_SH(in1, in0, in3, in2, in0, in2);
+  ST_SH2(in0, in2, output + 16, 8);
 }
 
-void vp8_short_walsh4x4_msa(int16_t *input, int16_t *output, int32_t pitch)
-{
-    v8i16 in0_h, in1_h, in2_h, in3_h;
-    v4i32 in0_w, in1_w, in2_w, in3_w, temp0, temp1, temp2, temp3;
+void vp8_short_walsh4x4_msa(int16_t *input, int16_t *output, int32_t pitch) {
+  v8i16 in0_h, in1_h, in2_h, in3_h;
+  v4i32 in0_w, in1_w, in2_w, in3_w, temp0, temp1, temp2, temp3;
 
-    LD_SH4(input, pitch / 2, in0_h, in1_h, in2_h, in3_h);
-    TRANSPOSE4x4_SH_SH(in0_h, in1_h, in2_h, in3_h, in0_h, in1_h, in2_h, in3_h);
+  LD_SH4(input, pitch / 2, in0_h, in1_h, in2_h, in3_h);
+  TRANSPOSE4x4_SH_SH(in0_h, in1_h, in2_h, in3_h, in0_h, in1_h, in2_h, in3_h);
 
-    UNPCK_R_SH_SW(in0_h, in0_w);
-    UNPCK_R_SH_SW(in1_h, in1_w);
-    UNPCK_R_SH_SW(in2_h, in2_w);
-    UNPCK_R_SH_SW(in3_h, in3_w);
-    BUTTERFLY_4(in0_w, in1_w, in3_w, in2_w, temp0, temp3, temp2, temp1);
-    SLLI_4V(temp0, temp1, temp2, temp3, 2);
-    BUTTERFLY_4(temp0, temp1, temp2, temp3, in0_w, in1_w, in2_w, in3_w);
-    temp0 = RET_1_IF_NZERO_W(temp0);
-    in0_w += temp0;
-    TRANSPOSE4x4_SW_SW(in0_w, in1_w, in2_w, in3_w, in0_w, in1_w, in2_w, in3_w);
+  UNPCK_R_SH_SW(in0_h, in0_w);
+  UNPCK_R_SH_SW(in1_h, in1_w);
+  UNPCK_R_SH_SW(in2_h, in2_w);
+  UNPCK_R_SH_SW(in3_h, in3_w);
+  BUTTERFLY_4(in0_w, in1_w, in3_w, in2_w, temp0, temp3, temp2, temp1);
+  SLLI_4V(temp0, temp1, temp2, temp3, 2);
+  BUTTERFLY_4(temp0, temp1, temp2, temp3, in0_w, in1_w, in2_w, in3_w);
+  temp0 = RET_1_IF_NZERO_W(temp0);
+  in0_w += temp0;
+  TRANSPOSE4x4_SW_SW(in0_w, in1_w, in2_w, in3_w, in0_w, in1_w, in2_w, in3_w);
 
-    BUTTERFLY_4(in0_w, in1_w, in3_w, in2_w, temp0, temp3, temp2, temp1);
-    BUTTERFLY_4(temp0, temp1, temp2, temp3, in0_w, in1_w, in2_w, in3_w);
-    in0_w += RET_1_IF_NEG_W(in0_w);
-    in1_w += RET_1_IF_NEG_W(in1_w);
-    in2_w += RET_1_IF_NEG_W(in2_w);
-    in3_w += RET_1_IF_NEG_W(in3_w);
-    ADD4(in0_w, 3, in1_w, 3, in2_w, 3, in3_w, 3, in0_w, in1_w, in2_w, in3_w);
-    SRA_4V(in0_w, in1_w, in2_w, in3_w, 3);
-    PCKEV_H2_SH(in1_w, in0_w, in3_w, in2_w, in0_h, in1_h);
-    ST_SH2(in0_h, in1_h, output, 8);
+  BUTTERFLY_4(in0_w, in1_w, in3_w, in2_w, temp0, temp3, temp2, temp1);
+  BUTTERFLY_4(temp0, temp1, temp2, temp3, in0_w, in1_w, in2_w, in3_w);
+  in0_w += RET_1_IF_NEG_W(in0_w);
+  in1_w += RET_1_IF_NEG_W(in1_w);
+  in2_w += RET_1_IF_NEG_W(in2_w);
+  in3_w += RET_1_IF_NEG_W(in3_w);
+  ADD4(in0_w, 3, in1_w, 3, in2_w, 3, in3_w, 3, in0_w, in1_w, in2_w, in3_w);
+  SRA_4V(in0_w, in1_w, in2_w, in3_w, 3);
+  PCKEV_H2_SH(in1_w, in0_w, in3_w, in2_w, in0_h, in1_h);
+  ST_SH2(in0_h, in1_h, output, 8);
 }
--- a/vp8/encoder/mips/msa/denoising_msa.c
+++ b/vp8/encoder/mips/msa/denoising_msa.c
@@ -16,609 +16,553 @@
 int32_t vp8_denoiser_filter_msa(uint8_t *mc_running_avg_y_ptr,
                                 int32_t mc_avg_y_stride,
                                 uint8_t *running_avg_y_ptr,
-                                int32_t avg_y_stride,
-                                uint8_t *sig_ptr, int32_t sig_stride,
-                                uint32_t motion_magnitude,
-                                int32_t increase_denoising)
-{
-    uint8_t *running_avg_y_start = running_avg_y_ptr;
-    uint8_t *sig_start = sig_ptr;
-    int32_t cnt = 0;
-    int32_t sum_diff = 0;
-    int32_t shift_inc1 = 3;
-    int32_t delta = 0;
-    int32_t sum_diff_thresh;
-    v16u8 src0, src1, src2, src3, src4, src5, src6, src7;
-    v16u8 src8, src9, src10, src11, src12, src13, src14, src15;
-    v16u8 mc_running_avg_y0, running_avg_y, sig0;
-    v16u8 mc_running_avg_y1, running_avg_y1, sig1;
-    v16u8 coeff0, coeff1;
-    v8i16 diff0, diff1, abs_diff0, abs_diff1, abs_diff_neg0, abs_diff_neg1;
-    v8i16 adjust0, adjust1, adjust2, adjust3;
-    v8i16 shift_inc1_vec = { 0 };
-    v8i16 col_sum0 = { 0 };
-    v8i16 col_sum1 = { 0 };
-    v8i16 col_sum2 = { 0 };
-    v8i16 col_sum3 = { 0 };
-    v8i16 temp0_h, temp1_h, temp2_h, temp3_h, cmp, delta_vec;
-    v4i32 temp0_w;
-    v2i64 temp0_d, temp1_d;
-    v8i16 zero = { 0 };
-    v8i16 one = __msa_ldi_h(1);
-    v8i16 four = __msa_ldi_h(4);
-    v8i16 val_127 = __msa_ldi_h(127);
-    v8i16 adj_val = { 6, 4, 3, 0, -6, -4, -3, 0 };
+                                int32_t avg_y_stride, uint8_t *sig_ptr,
+                                int32_t sig_stride, uint32_t motion_magnitude,
+                                int32_t increase_denoising) {
+  uint8_t *running_avg_y_start = running_avg_y_ptr;
+  uint8_t *sig_start = sig_ptr;
+  int32_t cnt = 0;
+  int32_t sum_diff = 0;
+  int32_t shift_inc1 = 3;
+  int32_t delta = 0;
+  int32_t sum_diff_thresh;
+  v16u8 src0, src1, src2, src3, src4, src5, src6, src7;
+  v16u8 src8, src9, src10, src11, src12, src13, src14, src15;
+  v16u8 mc_running_avg_y0, running_avg_y, sig0;
+  v16u8 mc_running_avg_y1, running_avg_y1, sig1;
+  v16u8 coeff0, coeff1;
+  v8i16 diff0, diff1, abs_diff0, abs_diff1, abs_diff_neg0, abs_diff_neg1;
+  v8i16 adjust0, adjust1, adjust2, adjust3;
+  v8i16 shift_inc1_vec = { 0 };
+  v8i16 col_sum0 = { 0 };
+  v8i16 col_sum1 = { 0 };
+  v8i16 col_sum2 = { 0 };
+  v8i16 col_sum3 = { 0 };
+  v8i16 temp0_h, temp1_h, temp2_h, temp3_h, cmp, delta_vec;
+  v4i32 temp0_w;
+  v2i64 temp0_d, temp1_d;
+  v8i16 zero = { 0 };
+  v8i16 one = __msa_ldi_h(1);
+  v8i16 four = __msa_ldi_h(4);
+  v8i16 val_127 = __msa_ldi_h(127);
+  v8i16 adj_val = { 6, 4, 3, 0, -6, -4, -3, 0 };
 
-    if (motion_magnitude <= MOTION_MAGNITUDE_THRESHOLD)
-    {
-        adj_val = __msa_add_a_h(adj_val, one);
-        if (increase_denoising)
-        {
-            adj_val = __msa_add_a_h(adj_val, one);
-            shift_inc1 = 4;
-        }
-
-        temp0_h = zero - adj_val;
-        adj_val = (v8i16)__msa_ilvev_d((v2i64)temp0_h, (v2i64)adj_val);
+  if (motion_magnitude <= MOTION_MAGNITUDE_THRESHOLD) {
+    adj_val = __msa_add_a_h(adj_val, one);
+    if (increase_denoising) {
+      adj_val = __msa_add_a_h(adj_val, one);
+      shift_inc1 = 4;
     }
 
-    adj_val = __msa_insert_h(adj_val, 3, cnt);
-    adj_val = __msa_insert_h(adj_val, 7, cnt);
-    shift_inc1_vec = __msa_fill_h(shift_inc1);
+    temp0_h = zero - adj_val;
+    adj_val = (v8i16)__msa_ilvev_d((v2i64)temp0_h, (v2i64)adj_val);
+  }
 
-    for (cnt = 8; cnt--;)
-    {
-        v8i16 mask0 = { 0 };
-        v8i16 mask1 = { 0 };
+  adj_val = __msa_insert_h(adj_val, 3, cnt);
+  adj_val = __msa_insert_h(adj_val, 7, cnt);
+  shift_inc1_vec = __msa_fill_h(shift_inc1);
 
+  for (cnt = 8; cnt--;) {
+    v8i16 mask0 = { 0 };
+    v8i16 mask1 = { 0 };
+
+    mc_running_avg_y0 = LD_UB(mc_running_avg_y_ptr);
+    sig0 = LD_UB(sig_ptr);
+    sig_ptr += sig_stride;
+    mc_running_avg_y_ptr += mc_avg_y_stride;
+
+    mc_running_avg_y1 = LD_UB(mc_running_avg_y_ptr);
+    sig1 = LD_UB(sig_ptr);
+
+    ILVRL_B2_UB(mc_running_avg_y0, sig0, coeff0, coeff1);
+    HSUB_UB2_SH(coeff0, coeff1, diff0, diff1);
+    abs_diff0 = __msa_add_a_h(diff0, zero);
+    abs_diff1 = __msa_add_a_h(diff1, zero);
+    cmp = __msa_clei_s_h(abs_diff0, 15);
+    cmp = cmp & one;
+    mask0 += cmp;
+    cmp = __msa_clei_s_h(abs_diff0, 7);
+    cmp = cmp & one;
+    mask0 += cmp;
+    cmp = abs_diff0 < shift_inc1_vec;
+    cmp = cmp & one;
+    mask0 += cmp;
+    cmp = __msa_clei_s_h(abs_diff1, 15);
+    cmp = cmp & one;
+    mask1 += cmp;
+    cmp = __msa_clei_s_h(abs_diff1, 7);
+    cmp = cmp & one;
+    mask1 += cmp;
+    cmp = abs_diff1 < shift_inc1_vec;
+    cmp = cmp & one;
+    mask1 += cmp;
+    temp0_h = __msa_clei_s_h(diff0, 0);
+    temp0_h = temp0_h & four;
+    mask0 += temp0_h;
+    temp1_h = __msa_clei_s_h(diff1, 0);
+    temp1_h = temp1_h & four;
+    mask1 += temp1_h;
+    VSHF_H2_SH(adj_val, adj_val, adj_val, adj_val, mask0, mask1, adjust0,
+               adjust1);
+    temp2_h = __msa_ceqi_h(adjust0, 0);
+    temp3_h = __msa_ceqi_h(adjust1, 0);
+    adjust0 = (v8i16)__msa_bmnz_v((v16u8)adjust0, (v16u8)diff0, (v16u8)temp2_h);
+    adjust1 = (v8i16)__msa_bmnz_v((v16u8)adjust1, (v16u8)diff1, (v16u8)temp3_h);
+    ADD2(col_sum0, adjust0, col_sum1, adjust1, col_sum0, col_sum1);
+    UNPCK_UB_SH(sig0, temp0_h, temp1_h);
+    ADD2(temp0_h, adjust0, temp1_h, adjust1, temp0_h, temp1_h);
+    MAXI_SH2_SH(temp0_h, temp1_h, 0);
+    SAT_UH2_SH(temp0_h, temp1_h, 7);
+    temp2_h = (v8i16)__msa_pckev_b((v16i8)temp3_h, (v16i8)temp2_h);
+    running_avg_y = (v16u8)__msa_pckev_b((v16i8)temp1_h, (v16i8)temp0_h);
+    running_avg_y =
+        __msa_bmnz_v(running_avg_y, mc_running_avg_y0, (v16u8)temp2_h);
+    ST_UB(running_avg_y, running_avg_y_ptr);
+    running_avg_y_ptr += avg_y_stride;
+
+    mask0 = zero;
+    mask1 = zero;
+    ILVRL_B2_UB(mc_running_avg_y1, sig1, coeff0, coeff1);
+    HSUB_UB2_SH(coeff0, coeff1, diff0, diff1);
+    abs_diff0 = __msa_add_a_h(diff0, zero);
+    abs_diff1 = __msa_add_a_h(diff1, zero);
+    cmp = __msa_clei_s_h(abs_diff0, 15);
+    cmp = cmp & one;
+    mask0 += cmp;
+    cmp = __msa_clei_s_h(abs_diff0, 7);
+    cmp = cmp & one;
+    mask0 += cmp;
+    cmp = abs_diff0 < shift_inc1_vec;
+    cmp = cmp & one;
+    mask0 += cmp;
+    cmp = __msa_clei_s_h(abs_diff1, 15);
+    cmp = cmp & one;
+    mask1 += cmp;
+    cmp = __msa_clei_s_h(abs_diff1, 7);
+    cmp = cmp & one;
+    mask1 += cmp;
+    cmp = abs_diff1 < shift_inc1_vec;
+    cmp = cmp & one;
+    mask1 += cmp;
+    temp0_h = __msa_clei_s_h(diff0, 0);
+    temp0_h = temp0_h & four;
+    mask0 += temp0_h;
+    temp1_h = __msa_clei_s_h(diff1, 0);
+    temp1_h = temp1_h & four;
+    mask1 += temp1_h;
+    VSHF_H2_SH(adj_val, adj_val, adj_val, adj_val, mask0, mask1, adjust0,
+               adjust1);
+    temp2_h = __msa_ceqi_h(adjust0, 0);
+    temp3_h = __msa_ceqi_h(adjust1, 0);
+    adjust0 = (v8i16)__msa_bmnz_v((v16u8)adjust0, (v16u8)diff0, (v16u8)temp2_h);
+    adjust1 = (v8i16)__msa_bmnz_v((v16u8)adjust1, (v16u8)diff1, (v16u8)temp3_h);
+    ADD2(col_sum0, adjust0, col_sum1, adjust1, col_sum0, col_sum1);
+    UNPCK_UB_SH(sig1, temp0_h, temp1_h);
+    ADD2(temp0_h, adjust0, temp1_h, adjust1, temp0_h, temp1_h);
+    MAXI_SH2_SH(temp0_h, temp1_h, 0);
+    SAT_UH2_SH(temp0_h, temp1_h, 7);
+    temp2_h = (v8i16)__msa_pckev_b((v16i8)temp3_h, (v16i8)temp2_h);
+    running_avg_y = (v16u8)__msa_pckev_b((v16i8)temp1_h, (v16i8)temp0_h);
+    running_avg_y =
+        __msa_bmnz_v(running_avg_y, mc_running_avg_y1, (v16u8)temp2_h);
+    ST_UB(running_avg_y, running_avg_y_ptr);
+    sig_ptr += sig_stride;
+    mc_running_avg_y_ptr += mc_avg_y_stride;
+    running_avg_y_ptr += avg_y_stride;
+  }
+
+  col_sum0 = __msa_min_s_h(col_sum0, val_127);
+  col_sum1 = __msa_min_s_h(col_sum1, val_127);
+  temp0_h = col_sum0 + col_sum1;
+  temp0_w = __msa_hadd_s_w(temp0_h, temp0_h);
+  temp0_d = __msa_hadd_s_d(temp0_w, temp0_w);
+  temp1_d = __msa_splati_d(temp0_d, 1);
+  temp0_d += temp1_d;
+  sum_diff = __msa_copy_s_w((v4i32)temp0_d, 0);
+  sig_ptr -= sig_stride * 16;
+  mc_running_avg_y_ptr -= mc_avg_y_stride * 16;
+  running_avg_y_ptr -= avg_y_stride * 16;
+
+  if (increase_denoising) {
+    sum_diff_thresh = SUM_DIFF_THRESHOLD_HIGH;
+  }
+
+  if (abs(sum_diff) > sum_diff_thresh) {
+    delta = ((abs(sum_diff) - sum_diff_thresh) >> 8) + 1;
+    delta_vec = __msa_fill_h(delta);
+    if (delta < 4) {
+      for (cnt = 8; cnt--;) {
+        running_avg_y = LD_UB(running_avg_y_ptr);
         mc_running_avg_y0 = LD_UB(mc_running_avg_y_ptr);
         sig0 = LD_UB(sig_ptr);
         sig_ptr += sig_stride;
         mc_running_avg_y_ptr += mc_avg_y_stride;
-
+        running_avg_y_ptr += avg_y_stride;
         mc_running_avg_y1 = LD_UB(mc_running_avg_y_ptr);
         sig1 = LD_UB(sig_ptr);
-
+        running_avg_y1 = LD_UB(running_avg_y_ptr);
         ILVRL_B2_UB(mc_running_avg_y0, sig0, coeff0, coeff1);
         HSUB_UB2_SH(coeff0, coeff1, diff0, diff1);
         abs_diff0 = __msa_add_a_h(diff0, zero);
         abs_diff1 = __msa_add_a_h(diff1, zero);
-        cmp = __msa_clei_s_h(abs_diff0, 15);
-        cmp = cmp & one;
-        mask0 += cmp;
-        cmp = __msa_clei_s_h(abs_diff0, 7);
-        cmp = cmp & one;
-        mask0 += cmp;
-        cmp = abs_diff0 < shift_inc1_vec;
-        cmp = cmp & one;
-        mask0 += cmp;
-        cmp = __msa_clei_s_h(abs_diff1, 15);
-        cmp = cmp & one;
-        mask1 += cmp;
-        cmp = __msa_clei_s_h(abs_diff1, 7);
-        cmp = cmp & one;
-        mask1 += cmp;
-        cmp = abs_diff1 < shift_inc1_vec;
-        cmp = cmp & one;
-        mask1 += cmp;
+        temp0_h = abs_diff0 < delta_vec;
+        temp1_h = abs_diff1 < delta_vec;
+        abs_diff0 = (v8i16)__msa_bmz_v((v16u8)abs_diff0, (v16u8)delta_vec,
+                                       (v16u8)temp0_h);
+        abs_diff1 = (v8i16)__msa_bmz_v((v16u8)abs_diff1, (v16u8)delta_vec,
+                                       (v16u8)temp1_h);
+        SUB2(zero, abs_diff0, zero, abs_diff1, abs_diff_neg0, abs_diff_neg1);
+        abs_diff_neg0 = zero - abs_diff0;
+        abs_diff_neg1 = zero - abs_diff1;
         temp0_h = __msa_clei_s_h(diff0, 0);
-        temp0_h = temp0_h & four;
-        mask0 += temp0_h;
         temp1_h = __msa_clei_s_h(diff1, 0);
-        temp1_h = temp1_h & four;
-        mask1 += temp1_h;
-        VSHF_H2_SH(adj_val, adj_val, adj_val, adj_val, mask0, mask1, adjust0,
-                   adjust1);
-        temp2_h = __msa_ceqi_h(adjust0, 0);
-        temp3_h = __msa_ceqi_h(adjust1, 0);
-        adjust0 = (v8i16)__msa_bmnz_v((v16u8)adjust0, (v16u8)diff0,
-                                     (v16u8)temp2_h);
-        adjust1 = (v8i16)__msa_bmnz_v((v16u8)adjust1, (v16u8)diff1,
-                                     (v16u8)temp3_h);
-        ADD2(col_sum0, adjust0, col_sum1, adjust1, col_sum0, col_sum1);
-        UNPCK_UB_SH(sig0, temp0_h, temp1_h);
-        ADD2(temp0_h, adjust0, temp1_h, adjust1, temp0_h, temp1_h);
-        MAXI_SH2_SH(temp0_h, temp1_h, 0);
-        SAT_UH2_SH(temp0_h, temp1_h, 7);
-        temp2_h = (v8i16)__msa_pckev_b((v16i8)temp3_h, (v16i8)temp2_h);
-        running_avg_y = (v16u8)__msa_pckev_b((v16i8)temp1_h, (v16i8)temp0_h);
-        running_avg_y = __msa_bmnz_v(running_avg_y, mc_running_avg_y0,
-                                     (v16u8)temp2_h);
-        ST_UB(running_avg_y, running_avg_y_ptr);
-        running_avg_y_ptr += avg_y_stride;
-
-        mask0 = zero;
-        mask1 = zero;
+        adjust0 = (v8i16)__msa_bmnz_v((v16u8)abs_diff0, (v16u8)abs_diff_neg0,
+                                      (v16u8)temp0_h);
+        adjust1 = (v8i16)__msa_bmnz_v((v16u8)abs_diff1, (v16u8)abs_diff_neg1,
+                                      (v16u8)temp1_h);
+        ILVRL_B2_SH(zero, running_avg_y, temp2_h, temp3_h);
+        ADD2(temp2_h, adjust0, temp3_h, adjust1, adjust2, adjust3);
+        MAXI_SH2_SH(adjust2, adjust3, 0);
+        SAT_UH2_SH(adjust2, adjust3, 7);
+        temp0_h = __msa_ceqi_h(diff0, 0);
+        temp1_h = __msa_ceqi_h(diff1, 0);
+        adjust2 =
+            (v8i16)__msa_bmz_v((v16u8)adjust2, (v16u8)temp2_h, (v16u8)temp0_h);
+        adjust3 =
+            (v8i16)__msa_bmz_v((v16u8)adjust3, (v16u8)temp3_h, (v16u8)temp1_h);
+        adjust0 =
+            (v8i16)__msa_bmnz_v((v16u8)adjust0, (v16u8)zero, (v16u8)temp0_h);
+        adjust1 =
+            (v8i16)__msa_bmnz_v((v16u8)adjust1, (v16u8)zero, (v16u8)temp1_h);
+        ADD2(col_sum2, adjust0, col_sum3, adjust1, col_sum2, col_sum3);
+        running_avg_y = (v16u8)__msa_pckev_b((v16i8)adjust3, (v16i8)adjust2);
+        ST_UB(running_avg_y, running_avg_y_ptr - avg_y_stride);
         ILVRL_B2_UB(mc_running_avg_y1, sig1, coeff0, coeff1);
         HSUB_UB2_SH(coeff0, coeff1, diff0, diff1);
         abs_diff0 = __msa_add_a_h(diff0, zero);
         abs_diff1 = __msa_add_a_h(diff1, zero);
-        cmp = __msa_clei_s_h(abs_diff0, 15);
-        cmp = cmp & one;
-        mask0 += cmp;
-        cmp = __msa_clei_s_h(abs_diff0, 7);
-        cmp = cmp & one;
-        mask0 += cmp;
-        cmp = abs_diff0 < shift_inc1_vec;
-        cmp = cmp & one;
-        mask0 += cmp;
-        cmp = __msa_clei_s_h(abs_diff1, 15);
-        cmp = cmp & one;
-        mask1 += cmp;
-        cmp = __msa_clei_s_h(abs_diff1, 7);
-        cmp = cmp & one;
-        mask1 += cmp;
-        cmp = abs_diff1 < shift_inc1_vec;
-        cmp = cmp & one;
-        mask1 += cmp;
+        temp0_h = abs_diff0 < delta_vec;
+        temp1_h = abs_diff1 < delta_vec;
+        abs_diff0 = (v8i16)__msa_bmz_v((v16u8)abs_diff0, (v16u8)delta_vec,
+                                       (v16u8)temp0_h);
+        abs_diff1 = (v8i16)__msa_bmz_v((v16u8)abs_diff1, (v16u8)delta_vec,
+                                       (v16u8)temp1_h);
+        SUB2(zero, abs_diff0, zero, abs_diff1, abs_diff_neg0, abs_diff_neg1);
         temp0_h = __msa_clei_s_h(diff0, 0);
-        temp0_h = temp0_h & four;
-        mask0 += temp0_h;
         temp1_h = __msa_clei_s_h(diff1, 0);
-        temp1_h = temp1_h & four;
-        mask1 += temp1_h;
-        VSHF_H2_SH(adj_val, adj_val, adj_val, adj_val, mask0, mask1, adjust0,
-                   adjust1);
-        temp2_h = __msa_ceqi_h(adjust0, 0);
-        temp3_h = __msa_ceqi_h(adjust1, 0);
-        adjust0 = (v8i16)__msa_bmnz_v((v16u8)adjust0, (v16u8)diff0,
-                                      (v16u8)temp2_h);
-        adjust1 = (v8i16)__msa_bmnz_v((v16u8)adjust1, (v16u8)diff1,
-                                      (v16u8)temp3_h);
-        ADD2(col_sum0, adjust0, col_sum1, adjust1, col_sum0, col_sum1);
-        UNPCK_UB_SH(sig1, temp0_h, temp1_h);
-        ADD2(temp0_h, adjust0, temp1_h, adjust1, temp0_h, temp1_h);
-        MAXI_SH2_SH(temp0_h, temp1_h, 0);
-        SAT_UH2_SH(temp0_h, temp1_h, 7);
-        temp2_h = (v8i16)__msa_pckev_b((v16i8)temp3_h, (v16i8)temp2_h);
-        running_avg_y = (v16u8)__msa_pckev_b((v16i8)temp1_h, (v16i8)temp0_h);
-        running_avg_y = __msa_bmnz_v(running_avg_y, mc_running_avg_y1,
-                                     (v16u8)temp2_h);
+        adjust0 = (v8i16)__msa_bmnz_v((v16u8)abs_diff0, (v16u8)abs_diff_neg0,
+                                      (v16u8)temp0_h);
+        adjust1 = (v8i16)__msa_bmnz_v((v16u8)abs_diff1, (v16u8)abs_diff_neg1,
+                                      (v16u8)temp1_h);
+        ILVRL_H2_SH(zero, running_avg_y1, temp2_h, temp3_h);
+        ADD2(temp2_h, adjust0, temp3_h, adjust1, adjust2, adjust3);
+        MAXI_SH2_SH(adjust2, adjust3, 0);
+        SAT_UH2_SH(adjust2, adjust3, 7);
+        temp0_h = __msa_ceqi_h(diff0, 0);
+        temp1_h = __msa_ceqi_h(diff1, 0);
+        adjust2 =
+            (v8i16)__msa_bmz_v((v16u8)adjust2, (v16u8)temp2_h, (v16u8)temp0_h);
+        adjust3 =
+            (v8i16)__msa_bmz_v((v16u8)adjust3, (v16u8)temp3_h, (v16u8)temp1_h);
+        adjust0 =
+            (v8i16)__msa_bmz_v((v16u8)adjust0, (v16u8)zero, (v16u8)temp0_h);
+        adjust1 =
+            (v8i16)__msa_bmz_v((v16u8)adjust1, (v16u8)zero, (v16u8)temp1_h);
+        ADD2(col_sum2, adjust0, col_sum3, adjust1, col_sum2, col_sum3);
+        running_avg_y = (v16u8)__msa_pckev_b((v16i8)adjust3, (v16i8)adjust2);
         ST_UB(running_avg_y, running_avg_y_ptr);
-        sig_ptr += sig_stride;
-        mc_running_avg_y_ptr += mc_avg_y_stride;
         running_avg_y_ptr += avg_y_stride;
+      }
+
+      col_sum2 = __msa_min_s_h(col_sum2, val_127);
+      col_sum3 = __msa_min_s_h(col_sum3, val_127);
+      temp0_h = col_sum2 + col_sum3;
+      temp0_w = __msa_hadd_s_w(temp0_h, temp0_h);
+      temp0_d = __msa_hadd_s_d(temp0_w, temp0_w);
+      temp1_d = __msa_splati_d(temp0_d, 1);
+      temp0_d += (v2i64)temp1_d;
+      sum_diff = __msa_copy_s_w((v4i32)temp0_d, 0);
+      if (abs(sum_diff) > SUM_DIFF_THRESHOLD) {
+        return COPY_BLOCK;
+      }
+    } else {
+      return COPY_BLOCK;
     }
+  }
 
-    col_sum0 = __msa_min_s_h(col_sum0, val_127);
-    col_sum1 = __msa_min_s_h(col_sum1, val_127);
-    temp0_h = col_sum0 + col_sum1;
-    temp0_w = __msa_hadd_s_w(temp0_h, temp0_h);
-    temp0_d = __msa_hadd_s_d(temp0_w, temp0_w);
-    temp1_d = __msa_splati_d(temp0_d, 1);
-    temp0_d += temp1_d;
-    sum_diff = __msa_copy_s_w((v4i32)temp0_d, 0);
-    sig_ptr -= sig_stride * 16;
-    mc_running_avg_y_ptr -= mc_avg_y_stride * 16;
-    running_avg_y_ptr -= avg_y_stride * 16;
+  LD_UB8(sig_start, sig_stride, src0, src1, src2, src3, src4, src5, src6, src7);
+  sig_start += (8 * sig_stride);
+  LD_UB8(sig_start, sig_stride, src8, src9, src10, src11, src12, src13, src14,
+         src15);
 
-    if (increase_denoising)
-    {
-        sum_diff_thresh = SUM_DIFF_THRESHOLD_HIGH;
-    }
+  ST_UB8(src0, src1, src2, src3, src4, src5, src6, src7, running_avg_y_start,
+         avg_y_stride);
+  running_avg_y_start += (8 * avg_y_stride);
+  ST_UB8(src8, src9, src10, src11, src12, src13, src14, src15,
+         running_avg_y_start, avg_y_stride);
 
-    if (abs(sum_diff) > sum_diff_thresh)
-    {
-        delta = ((abs(sum_diff) - sum_diff_thresh) >> 8) + 1;
-        delta_vec = __msa_fill_h(delta);
-        if (delta < 4)
-        {
-            for (cnt = 8; cnt--;)
-            {
-                running_avg_y = LD_UB(running_avg_y_ptr);
-                mc_running_avg_y0 = LD_UB(mc_running_avg_y_ptr);
-                sig0 = LD_UB(sig_ptr);
-                sig_ptr += sig_stride;
-                mc_running_avg_y_ptr += mc_avg_y_stride;
-                running_avg_y_ptr += avg_y_stride;
-                mc_running_avg_y1 = LD_UB(mc_running_avg_y_ptr);
-                sig1 = LD_UB(sig_ptr);
-                running_avg_y1 = LD_UB(running_avg_y_ptr);
-                ILVRL_B2_UB(mc_running_avg_y0, sig0, coeff0, coeff1);
-                HSUB_UB2_SH(coeff0, coeff1, diff0, diff1);
-                abs_diff0 = __msa_add_a_h(diff0, zero);
-                abs_diff1 = __msa_add_a_h(diff1, zero);
-                temp0_h = abs_diff0 < delta_vec;
-                temp1_h = abs_diff1 < delta_vec;
-                abs_diff0 = (v8i16)__msa_bmz_v((v16u8)abs_diff0,
-                                               (v16u8)delta_vec,
-                                               (v16u8)temp0_h);
-                abs_diff1 = (v8i16)__msa_bmz_v((v16u8)abs_diff1,
-                                               (v16u8)delta_vec,
-                                               (v16u8)temp1_h);
-                SUB2(zero, abs_diff0, zero, abs_diff1, abs_diff_neg0,
-                     abs_diff_neg1);
-                abs_diff_neg0 = zero - abs_diff0;
-                abs_diff_neg1 = zero - abs_diff1;
-                temp0_h = __msa_clei_s_h(diff0, 0);
-                temp1_h = __msa_clei_s_h(diff1, 0);
-                adjust0 = (v8i16)__msa_bmnz_v((v16u8)abs_diff0,
-                                              (v16u8)abs_diff_neg0,
-                                              (v16u8)temp0_h);
-                adjust1 = (v8i16)__msa_bmnz_v((v16u8)abs_diff1,
-                                              (v16u8)abs_diff_neg1,
-                                              (v16u8)temp1_h);
-                ILVRL_B2_SH(zero, running_avg_y, temp2_h, temp3_h);
-                ADD2(temp2_h, adjust0, temp3_h, adjust1, adjust2, adjust3);
-                MAXI_SH2_SH(adjust2, adjust3, 0);
-                SAT_UH2_SH(adjust2, adjust3, 7);
-                temp0_h = __msa_ceqi_h(diff0, 0);
-                temp1_h = __msa_ceqi_h(diff1, 0);
-                adjust2 = (v8i16)__msa_bmz_v((v16u8)adjust2, (v16u8)temp2_h,
-                                             (v16u8)temp0_h);
-                adjust3 = (v8i16)__msa_bmz_v((v16u8)adjust3, (v16u8)temp3_h,
-                                             (v16u8)temp1_h);
-                adjust0 = (v8i16)__msa_bmnz_v((v16u8)adjust0, (v16u8)zero,
-                                              (v16u8)temp0_h);
-                adjust1 = (v8i16)__msa_bmnz_v((v16u8)adjust1, (v16u8)zero,
-                                              (v16u8)temp1_h);
-                ADD2(col_sum2, adjust0, col_sum3, adjust1, col_sum2, col_sum3);
-                running_avg_y = (v16u8)__msa_pckev_b((v16i8)adjust3,
-                                                     (v16i8)adjust2);
-                ST_UB(running_avg_y, running_avg_y_ptr - avg_y_stride);
-                ILVRL_B2_UB(mc_running_avg_y1, sig1, coeff0, coeff1);
-                HSUB_UB2_SH(coeff0, coeff1, diff0, diff1);
-                abs_diff0 = __msa_add_a_h(diff0, zero);
-                abs_diff1 = __msa_add_a_h(diff1, zero);
-                temp0_h = abs_diff0 < delta_vec;
-                temp1_h = abs_diff1 < delta_vec;
-                abs_diff0 = (v8i16)__msa_bmz_v((v16u8)abs_diff0,
-                                               (v16u8)delta_vec,
-                                               (v16u8)temp0_h);
-                abs_diff1 = (v8i16)__msa_bmz_v((v16u8)abs_diff1,
-                                               (v16u8)delta_vec,
-                                               (v16u8)temp1_h);
-                SUB2(zero, abs_diff0, zero, abs_diff1, abs_diff_neg0,
-                     abs_diff_neg1);
-                temp0_h = __msa_clei_s_h(diff0, 0);
-                temp1_h = __msa_clei_s_h(diff1, 0);
-                adjust0 = (v8i16)__msa_bmnz_v((v16u8)abs_diff0,
-                                              (v16u8)abs_diff_neg0,
-                                              (v16u8)temp0_h);
-                adjust1 = (v8i16)__msa_bmnz_v((v16u8)abs_diff1,
-                                              (v16u8)abs_diff_neg1,
-                                              (v16u8)temp1_h);
-                ILVRL_H2_SH(zero, running_avg_y1, temp2_h, temp3_h);
-                ADD2(temp2_h, adjust0, temp3_h, adjust1, adjust2, adjust3);
-                MAXI_SH2_SH(adjust2, adjust3, 0);
-                SAT_UH2_SH(adjust2, adjust3, 7);
-                temp0_h = __msa_ceqi_h(diff0, 0);
-                temp1_h = __msa_ceqi_h(diff1, 0);
-                adjust2 = (v8i16)__msa_bmz_v((v16u8)adjust2, (v16u8)temp2_h,
-                                             (v16u8)temp0_h);
-                adjust3 = (v8i16)__msa_bmz_v((v16u8)adjust3, (v16u8)temp3_h,
-                                             (v16u8)temp1_h);
-                adjust0 = (v8i16)__msa_bmz_v((v16u8)adjust0, (v16u8)zero,
-                                             (v16u8)temp0_h);
-                adjust1 = (v8i16)__msa_bmz_v((v16u8)adjust1, (v16u8)zero,
-                                             (v16u8)temp1_h);
-                ADD2(col_sum2, adjust0, col_sum3, adjust1, col_sum2, col_sum3);
-                running_avg_y = (v16u8)__msa_pckev_b((v16i8)adjust3,
-                                                     (v16i8)adjust2);
-                ST_UB(running_avg_y, running_avg_y_ptr);
-                running_avg_y_ptr += avg_y_stride;
-            }
+  return FILTER_BLOCK;
+}
 
-            col_sum2 = __msa_min_s_h(col_sum2, val_127);
-            col_sum3 = __msa_min_s_h(col_sum3, val_127);
-            temp0_h = col_sum2 + col_sum3;
-            temp0_w = __msa_hadd_s_w(temp0_h, temp0_h);
-            temp0_d = __msa_hadd_s_d(temp0_w, temp0_w);
-            temp1_d = __msa_splati_d(temp0_d, 1);
-            temp0_d += (v2i64)temp1_d;
-            sum_diff = __msa_copy_s_w((v4i32)temp0_d, 0);
-            if (abs(sum_diff) > SUM_DIFF_THRESHOLD)
-            {
-                return COPY_BLOCK;
-            }
-        }
-        else
-        {
-            return COPY_BLOCK;
-        }
-    }
+int32_t vp8_denoiser_filter_uv_msa(
+    uint8_t *mc_running_avg_y_ptr, int32_t mc_avg_y_stride,
+    uint8_t *running_avg_y_ptr, int32_t avg_y_stride, uint8_t *sig_ptr,
+    int32_t sig_stride, uint32_t motion_magnitude, int32_t increase_denoising) {
+  uint8_t *running_avg_y_start = running_avg_y_ptr;
+  uint8_t *sig_start = sig_ptr;
+  int32_t cnt = 0;
+  int32_t sum_diff = 0;
+  int32_t shift_inc1 = 3;
+  int32_t delta = 0;
+  int32_t sum_block = 0;
+  int32_t sum_diff_thresh;
+  int64_t dst0, dst1, src0, src1, src2, src3;
+  v16u8 mc_running_avg_y0, running_avg_y, sig0;
+  v16u8 mc_running_avg_y1, running_avg_y1, sig1;
+  v16u8 sig2, sig3, sig4, sig5, sig6, sig7;
+  v16u8 coeff0;
+  v8i16 diff0, abs_diff0, abs_diff_neg0;
+  v8i16 adjust0, adjust2;
+  v8i16 shift_inc1_vec = { 0 };
+  v8i16 col_sum0 = { 0 };
+  v8i16 temp0_h, temp2_h, cmp, delta_vec;
+  v4i32 temp0_w;
+  v2i64 temp0_d, temp1_d;
+  v16i8 zero = { 0 };
+  v8i16 one = __msa_ldi_h(1);
+  v8i16 four = __msa_ldi_h(4);
+  v8i16 adj_val = { 6, 4, 3, 0, -6, -4, -3, 0 };
 
-    LD_UB8(sig_start, sig_stride, src0, src1, src2, src3, src4, src5, src6,
-           src7);
-    sig_start += (8 * sig_stride);
-    LD_UB8(sig_start, sig_stride, src8, src9, src10, src11, src12, src13,
-           src14, src15);
+  sig0 = LD_UB(sig_ptr);
+  sig_ptr += sig_stride;
+  temp0_h = (v8i16)__msa_ilvr_b(zero, (v16i8)sig0);
+  sig1 = LD_UB(sig_ptr);
+  sig_ptr += sig_stride;
+  temp0_h += (v8i16)__msa_ilvr_b(zero, (v16i8)sig1);
+  sig2 = LD_UB(sig_ptr);
+  sig_ptr += sig_stride;
+  temp0_h += (v8i16)__msa_ilvr_b(zero, (v16i8)sig2);
+  sig3 = LD_UB(sig_ptr);
+  sig_ptr += sig_stride;
+  temp0_h += (v8i16)__msa_ilvr_b(zero, (v16i8)sig3);
+  sig4 = LD_UB(sig_ptr);
+  sig_ptr += sig_stride;
+  temp0_h += (v8i16)__msa_ilvr_b(zero, (v16i8)sig4);
+  sig5 = LD_UB(sig_ptr);
+  sig_ptr += sig_stride;
+  temp0_h += (v8i16)__msa_ilvr_b(zero, (v16i8)sig5);
+  sig6 = LD_UB(sig_ptr);
+  sig_ptr += sig_stride;
+  temp0_h += (v8i16)__msa_ilvr_b(zero, (v16i8)sig6);
+  sig7 = LD_UB(sig_ptr);
+  sig_ptr += sig_stride;
+  temp0_h += (v8i16)__msa_ilvr_b(zero, (v16i8)sig7);
+  temp0_w = __msa_hadd_s_w(temp0_h, temp0_h);
+  temp0_d = __msa_hadd_s_d(temp0_w, temp0_w);
+  temp1_d = __msa_splati_d(temp0_d, 1);
+  temp0_d += temp1_d;
+  sum_block = __msa_copy_s_w((v4i32)temp0_d, 0);
+  sig_ptr -= sig_stride * 8;
 
-    ST_UB8(src0, src1, src2, src3, src4, src5, src6, src7, running_avg_y_start,
-           avg_y_stride);
-    running_avg_y_start += (8 * avg_y_stride);
-    ST_UB8(src8, src9, src10, src11, src12, src13, src14, src15,
-           running_avg_y_start, avg_y_stride);
+  if (abs(sum_block - (128 * 8 * 8)) < SUM_DIFF_FROM_AVG_THRESH_UV) {
+    return COPY_BLOCK;
+  }
 
-    return FILTER_BLOCK;
-}
+  if (motion_magnitude <= MOTION_MAGNITUDE_THRESHOLD) {
+    adj_val = __msa_add_a_h(adj_val, one);
 
-int32_t vp8_denoiser_filter_uv_msa(uint8_t *mc_running_avg_y_ptr,
-                                   int32_t mc_avg_y_stride,
-                                   uint8_t *running_avg_y_ptr,
-                                   int32_t avg_y_stride,
-                                   uint8_t *sig_ptr,
-                                   int32_t sig_stride,
-                                   uint32_t motion_magnitude,
-                                   int32_t increase_denoising)
-{
-    uint8_t *running_avg_y_start = running_avg_y_ptr;
-    uint8_t *sig_start = sig_ptr;
-    int32_t cnt = 0;
-    int32_t sum_diff = 0;
-    int32_t shift_inc1 = 3;
-    int32_t delta = 0;
-    int32_t sum_block = 0;
-    int32_t sum_diff_thresh;
-    int64_t dst0, dst1, src0, src1, src2, src3;
-    v16u8 mc_running_avg_y0, running_avg_y, sig0;
-    v16u8 mc_running_avg_y1, running_avg_y1, sig1;
-    v16u8 sig2, sig3, sig4, sig5, sig6, sig7;
-    v16u8 coeff0;
-    v8i16 diff0, abs_diff0, abs_diff_neg0;
-    v8i16 adjust0, adjust2;
-    v8i16 shift_inc1_vec = { 0 };
-    v8i16 col_sum0 = { 0 };
-    v8i16 temp0_h, temp2_h, cmp, delta_vec;
-    v4i32 temp0_w;
-    v2i64 temp0_d, temp1_d;
-    v16i8 zero = { 0 };
-    v8i16 one = __msa_ldi_h(1);
-    v8i16 four = __msa_ldi_h(4);
-    v8i16 adj_val = { 6, 4, 3, 0, -6, -4, -3, 0 };
+    if (increase_denoising) {
+      adj_val = __msa_add_a_h(adj_val, one);
+      shift_inc1 = 4;
+    }
 
+    temp0_h = (v8i16)zero - adj_val;
+    adj_val = (v8i16)__msa_ilvev_d((v2i64)temp0_h, (v2i64)adj_val);
+  }
 
+  adj_val = __msa_insert_h(adj_val, 3, cnt);
+  adj_val = __msa_insert_h(adj_val, 7, cnt);
+  shift_inc1_vec = __msa_fill_h(shift_inc1);
+  for (cnt = 4; cnt--;) {
+    v8i16 mask0 = { 0 };
+    mc_running_avg_y0 = LD_UB(mc_running_avg_y_ptr);
     sig0 = LD_UB(sig_ptr);
     sig_ptr += sig_stride;
-    temp0_h = (v8i16)__msa_ilvr_b(zero, (v16i8)sig0);
+    mc_running_avg_y_ptr += mc_avg_y_stride;
+    mc_running_avg_y1 = LD_UB(mc_running_avg_y_ptr);
     sig1 = LD_UB(sig_ptr);
-    sig_ptr += sig_stride;
-    temp0_h += (v8i16)__msa_ilvr_b(zero, (v16i8)sig1);
-    sig2 = LD_UB(sig_ptr);
-    sig_ptr += sig_stride;
-    temp0_h += (v8i16)__msa_ilvr_b(zero, (v16i8)sig2);
-    sig3 = LD_UB(sig_ptr);
-    sig_ptr += sig_stride;
-    temp0_h += (v8i16)__msa_ilvr_b(zero, (v16i8)sig3);
-    sig4 = LD_UB(sig_ptr);
-    sig_ptr += sig_stride;
-    temp0_h += (v8i16)__msa_ilvr_b(zero, (v16i8)sig4);
-    sig5 = LD_UB(sig_ptr);
-    sig_ptr += sig_stride;
-    temp0_h += (v8i16)__msa_ilvr_b(zero, (v16i8)sig5);
-    sig6 = LD_UB(sig_ptr);
-    sig_ptr += sig_stride;
-    temp0_h += (v8i16)__msa_ilvr_b(zero, (v16i8)sig6);
-    sig7 = LD_UB(sig_ptr);
-    sig_ptr += sig_stride;
-    temp0_h += (v8i16)__msa_ilvr_b(zero, (v16i8)sig7);
-    temp0_w = __msa_hadd_s_w(temp0_h, temp0_h);
-    temp0_d = __msa_hadd_s_d(temp0_w, temp0_w);
-    temp1_d = __msa_splati_d(temp0_d, 1);
-    temp0_d += temp1_d;
-    sum_block = __msa_copy_s_w((v4i32)temp0_d, 0);
-    sig_ptr -= sig_stride * 8;
+    coeff0 = (v16u8)__msa_ilvr_b((v16i8)mc_running_avg_y0, (v16i8)sig0);
+    diff0 = __msa_hsub_u_h(coeff0, coeff0);
+    abs_diff0 = __msa_add_a_h(diff0, (v8i16)zero);
+    cmp = __msa_clei_s_h(abs_diff0, 15);
+    cmp = cmp & one;
+    mask0 += cmp;
+    cmp = __msa_clei_s_h(abs_diff0, 7);
+    cmp = cmp & one;
+    mask0 += cmp;
+    cmp = abs_diff0 < shift_inc1_vec;
+    cmp = cmp & one;
+    mask0 += cmp;
+    temp0_h = __msa_clei_s_h(diff0, 0);
+    temp0_h = temp0_h & four;
+    mask0 += temp0_h;
+    adjust0 = __msa_vshf_h(mask0, adj_val, adj_val);
+    temp2_h = __msa_ceqi_h(adjust0, 0);
+    adjust0 = (v8i16)__msa_bmnz_v((v16u8)adjust0, (v16u8)diff0, (v16u8)temp2_h);
+    col_sum0 += adjust0;
+    temp0_h = (v8i16)__msa_ilvr_b(zero, (v16i8)sig0);
+    temp0_h += adjust0;
+    temp0_h = __msa_maxi_s_h(temp0_h, 0);
+    temp0_h = (v8i16)__msa_sat_u_h((v8u16)temp0_h, 7);
+    temp2_h = (v8i16)__msa_pckev_b((v16i8)temp2_h, (v16i8)temp2_h);
+    running_avg_y = (v16u8)__msa_pckev_b((v16i8)temp0_h, (v16i8)temp0_h);
+    running_avg_y =
+        __msa_bmnz_v(running_avg_y, mc_running_avg_y0, (v16u8)temp2_h);
+    dst0 = __msa_copy_s_d((v2i64)running_avg_y, 0);
+    SD(dst0, running_avg_y_ptr);
+    running_avg_y_ptr += avg_y_stride;
 
-    if (abs(sum_block - (128 * 8 * 8)) < SUM_DIFF_FROM_AVG_THRESH_UV)
-    {
-        return COPY_BLOCK;
-    }
+    mask0 = __msa_ldi_h(0);
+    coeff0 = (v16u8)__msa_ilvr_b((v16i8)mc_running_avg_y1, (v16i8)sig1);
+    diff0 = __msa_hsub_u_h(coeff0, coeff0);
+    abs_diff0 = __msa_add_a_h(diff0, (v8i16)zero);
+    cmp = __msa_clei_s_h(abs_diff0, 15);
+    cmp = cmp & one;
+    mask0 += cmp;
+    cmp = __msa_clei_s_h(abs_diff0, 7);
+    cmp = cmp & one;
+    mask0 += cmp;
+    cmp = abs_diff0 < shift_inc1_vec;
+    cmp = cmp & one;
+    mask0 += cmp;
+    temp0_h = __msa_clei_s_h(diff0, 0);
+    temp0_h = temp0_h & four;
+    mask0 += temp0_h;
+    adjust0 = __msa_vshf_h(mask0, adj_val, adj_val);
+    temp2_h = __msa_ceqi_h(adjust0, 0);
+    adjust0 = (v8i16)__msa_bmnz_v((v16u8)adjust0, (v16u8)diff0, (v16u8)temp2_h);
+    col_sum0 += adjust0;
+    temp0_h = (v8i16)__msa_ilvr_b(zero, (v16i8)sig1);
+    temp0_h += adjust0;
+    temp0_h = __msa_maxi_s_h(temp0_h, 0);
+    temp0_h = (v8i16)__msa_sat_u_h((v8u16)temp0_h, 7);
 
-    if (motion_magnitude <= MOTION_MAGNITUDE_THRESHOLD)
-    {
-        adj_val = __msa_add_a_h(adj_val, one);
+    temp2_h = (v8i16)__msa_pckev_b((v16i8)temp2_h, (v16i8)temp2_h);
+    running_avg_y = (v16u8)__msa_pckev_b((v16i8)temp0_h, (v16i8)temp0_h);
+    running_avg_y =
+        __msa_bmnz_v(running_avg_y, mc_running_avg_y1, (v16u8)temp2_h);
+    dst1 = __msa_copy_s_d((v2i64)running_avg_y, 0);
+    SD(dst1, running_avg_y_ptr);
 
-        if (increase_denoising)
-        {
-            adj_val = __msa_add_a_h(adj_val, one);
-            shift_inc1 = 4;
-        }
+    sig_ptr += sig_stride;
+    mc_running_avg_y_ptr += mc_avg_y_stride;
+    running_avg_y_ptr += avg_y_stride;
+  }
 
-        temp0_h = (v8i16)zero - adj_val;
-        adj_val = (v8i16)__msa_ilvev_d((v2i64)temp0_h, (v2i64)adj_val);
-    }
+  temp0_h = col_sum0;
+  temp0_w = __msa_hadd_s_w(temp0_h, temp0_h);
+  temp0_d = __msa_hadd_s_d(temp0_w, temp0_w);
+  temp1_d = __msa_splati_d(temp0_d, 1);
+  temp0_d += temp1_d;
+  sum_diff = __msa_copy_s_w((v4i32)temp0_d, 0);
+  sig_ptr -= sig_stride * 8;
+  mc_running_avg_y_ptr -= mc_avg_y_stride * 8;
+  running_avg_y_ptr -= avg_y_stride * 8;
+  sum_diff_thresh = SUM_DIFF_THRESHOLD_UV;
 
-    adj_val = __msa_insert_h(adj_val, 3, cnt);
-    adj_val = __msa_insert_h(adj_val, 7, cnt);
-    shift_inc1_vec = __msa_fill_h(shift_inc1);
-    for (cnt = 4; cnt--;)
-    {
-        v8i16 mask0 = { 0 };
+  if (increase_denoising) {
+    sum_diff_thresh = SUM_DIFF_THRESHOLD_HIGH_UV;
+  }
+
+  if (abs(sum_diff) > sum_diff_thresh) {
+    delta = ((abs(sum_diff) - sum_diff_thresh) >> 8) + 1;
+    delta_vec = __msa_fill_h(delta);
+    if (delta < 4) {
+      for (cnt = 4; cnt--;) {
+        running_avg_y = LD_UB(running_avg_y_ptr);
         mc_running_avg_y0 = LD_UB(mc_running_avg_y_ptr);
         sig0 = LD_UB(sig_ptr);
+        /* Update pointers for next iteration. */
         sig_ptr += sig_stride;
         mc_running_avg_y_ptr += mc_avg_y_stride;
+        running_avg_y_ptr += avg_y_stride;
+
         mc_running_avg_y1 = LD_UB(mc_running_avg_y_ptr);
         sig1 = LD_UB(sig_ptr);
+        running_avg_y1 = LD_UB(running_avg_y_ptr);
+
         coeff0 = (v16u8)__msa_ilvr_b((v16i8)mc_running_avg_y0, (v16i8)sig0);
         diff0 = __msa_hsub_u_h(coeff0, coeff0);
         abs_diff0 = __msa_add_a_h(diff0, (v8i16)zero);
-        cmp = __msa_clei_s_h(abs_diff0, 15);
-        cmp = cmp & one;
-        mask0 += cmp;
-        cmp = __msa_clei_s_h(abs_diff0, 7);
-        cmp = cmp & one;
-        mask0 += cmp;
-        cmp = abs_diff0 < shift_inc1_vec;
-        cmp = cmp & one;
-        mask0 += cmp;
+        temp0_h = delta_vec < abs_diff0;
+        abs_diff0 = (v8i16)__msa_bmnz_v((v16u8)abs_diff0, (v16u8)delta_vec,
+                                        (v16u8)temp0_h);
+        abs_diff_neg0 = (v8i16)zero - abs_diff0;
         temp0_h = __msa_clei_s_h(diff0, 0);
-        temp0_h = temp0_h & four;
-        mask0 += temp0_h;
-        adjust0 = __msa_vshf_h(mask0, adj_val, adj_val);
-        temp2_h = __msa_ceqi_h(adjust0, 0);
-        adjust0 = (v8i16)__msa_bmnz_v((v16u8)adjust0, (v16u8)diff0,
-                                      (v16u8)temp2_h);
+        adjust0 = (v8i16)__msa_bmz_v((v16u8)abs_diff0, (v16u8)abs_diff_neg0,
+                                     (v16u8)temp0_h);
+        temp2_h = (v8i16)__msa_ilvr_b(zero, (v16i8)running_avg_y);
+        adjust2 = temp2_h + adjust0;
+        adjust2 = __msa_maxi_s_h(adjust2, 0);
+        adjust2 = (v8i16)__msa_sat_u_h((v8u16)adjust2, 7);
+        temp0_h = __msa_ceqi_h(diff0, 0);
+        adjust2 =
+            (v8i16)__msa_bmnz_v((v16u8)adjust2, (v16u8)temp2_h, (v16u8)temp0_h);
+        adjust0 =
+            (v8i16)__msa_bmnz_v((v16u8)adjust0, (v16u8)zero, (v16u8)temp0_h);
         col_sum0 += adjust0;
-        temp0_h = (v8i16)__msa_ilvr_b(zero, (v16i8)sig0);
-        temp0_h += adjust0;
-        temp0_h = __msa_maxi_s_h(temp0_h, 0);
-        temp0_h = (v8i16)__msa_sat_u_h((v8u16)temp0_h, 7);
-        temp2_h = (v8i16)__msa_pckev_b((v16i8)temp2_h, (v16i8)temp2_h);
-        running_avg_y = (v16u8)__msa_pckev_b((v16i8)temp0_h, (v16i8)temp0_h);
-        running_avg_y = __msa_bmnz_v(running_avg_y, mc_running_avg_y0,
-                                     (v16u8)temp2_h);
-        dst0 = __msa_copy_s_d((v2i64)running_avg_y,  0);
-        SD(dst0, running_avg_y_ptr);
-        running_avg_y_ptr += avg_y_stride;
+        running_avg_y = (v16u8)__msa_pckev_b((v16i8)adjust2, (v16i8)adjust2);
+        dst0 = __msa_copy_s_d((v2i64)running_avg_y, 0);
+        SD(dst0, running_avg_y_ptr - avg_y_stride);
 
-        mask0 = __msa_ldi_h(0);
         coeff0 = (v16u8)__msa_ilvr_b((v16i8)mc_running_avg_y1, (v16i8)sig1);
         diff0 = __msa_hsub_u_h(coeff0, coeff0);
         abs_diff0 = __msa_add_a_h(diff0, (v8i16)zero);
-        cmp = __msa_clei_s_h(abs_diff0, 15);
-        cmp = cmp & one;
-        mask0 += cmp;
-        cmp = __msa_clei_s_h(abs_diff0, 7);
-        cmp = cmp & one;
-        mask0 += cmp;
-        cmp = abs_diff0 < shift_inc1_vec;
-        cmp = cmp & one;
-        mask0 += cmp;
+        temp0_h = delta_vec < abs_diff0;
+        abs_diff0 = (v8i16)__msa_bmnz_v((v16u8)abs_diff0, (v16u8)delta_vec,
+                                        (v16u8)temp0_h);
+        abs_diff_neg0 = (v8i16)zero - abs_diff0;
         temp0_h = __msa_clei_s_h(diff0, 0);
-        temp0_h = temp0_h & four;
-        mask0 += temp0_h;
-        adjust0 = __msa_vshf_h(mask0, adj_val, adj_val);
-        temp2_h = __msa_ceqi_h(adjust0, 0);
-        adjust0 = (v8i16)__msa_bmnz_v((v16u8)adjust0, (v16u8)diff0,
-                                      (v16u8)temp2_h);
+        adjust0 = (v8i16)__msa_bmz_v((v16u8)abs_diff0, (v16u8)abs_diff_neg0,
+                                     (v16u8)temp0_h);
+        temp2_h = (v8i16)__msa_ilvr_b(zero, (v16i8)running_avg_y1);
+        adjust2 = temp2_h + adjust0;
+        adjust2 = __msa_maxi_s_h(adjust2, 0);
+        adjust2 = (v8i16)__msa_sat_u_h((v8u16)adjust2, 7);
+        temp0_h = __msa_ceqi_h(diff0, 0);
+        adjust2 =
+            (v8i16)__msa_bmnz_v((v16u8)adjust2, (v16u8)temp2_h, (v16u8)temp0_h);
+        adjust0 =
+            (v8i16)__msa_bmnz_v((v16u8)adjust0, (v16u8)zero, (v16u8)temp0_h);
         col_sum0 += adjust0;
-        temp0_h = (v8i16)__msa_ilvr_b(zero, (v16i8)sig1);
-        temp0_h += adjust0;
-        temp0_h = __msa_maxi_s_h(temp0_h, 0);
-        temp0_h = (v8i16)__msa_sat_u_h((v8u16)temp0_h, 7);
-
-        temp2_h = (v8i16)__msa_pckev_b((v16i8)temp2_h, (v16i8)temp2_h);
-        running_avg_y = (v16u8)__msa_pckev_b((v16i8)temp0_h, (v16i8)temp0_h);
-        running_avg_y = __msa_bmnz_v(running_avg_y, mc_running_avg_y1,
-                                     (v16u8)temp2_h);
+        running_avg_y = (v16u8)__msa_pckev_b((v16i8)adjust2, (v16i8)adjust2);
         dst1 = __msa_copy_s_d((v2i64)running_avg_y, 0);
         SD(dst1, running_avg_y_ptr);
-
-        sig_ptr += sig_stride;
-        mc_running_avg_y_ptr += mc_avg_y_stride;
         running_avg_y_ptr += avg_y_stride;
-    }
+      }
 
-    temp0_h = col_sum0;
-    temp0_w = __msa_hadd_s_w(temp0_h, temp0_h);
-    temp0_d = __msa_hadd_s_d(temp0_w, temp0_w);
-    temp1_d = __msa_splati_d(temp0_d, 1);
-    temp0_d += temp1_d;
-    sum_diff = __msa_copy_s_w((v4i32)temp0_d, 0);
-    sig_ptr -= sig_stride * 8;
-    mc_running_avg_y_ptr -= mc_avg_y_stride * 8;
-    running_avg_y_ptr -= avg_y_stride * 8;
-    sum_diff_thresh = SUM_DIFF_THRESHOLD_UV;
+      temp0_h = col_sum0;
+      temp0_w = __msa_hadd_s_w(temp0_h, temp0_h);
+      temp0_d = __msa_hadd_s_d(temp0_w, temp0_w);
+      temp1_d = __msa_splati_d(temp0_d, 1);
+      temp0_d += temp1_d;
+      sum_diff = __msa_copy_s_w((v4i32)temp0_d, 0);
 
-    if (increase_denoising)
-    {
-        sum_diff_thresh = SUM_DIFF_THRESHOLD_HIGH_UV;
+      if (abs(sum_diff) > sum_diff_thresh) {
+        return COPY_BLOCK;
+      }
+    } else {
+      return COPY_BLOCK;
     }
+  }
 
-    if (abs(sum_diff) > sum_diff_thresh)
-    {
-        delta = ((abs(sum_diff) - sum_diff_thresh) >> 8) + 1;
-        delta_vec = __msa_fill_h(delta);
-        if (delta < 4)
-        {
-            for (cnt = 4; cnt--;)
-            {
-                running_avg_y = LD_UB(running_avg_y_ptr);
-                mc_running_avg_y0 = LD_UB(mc_running_avg_y_ptr);
-                sig0 = LD_UB(sig_ptr);
-                /* Update pointers for next iteration. */
-                sig_ptr += sig_stride;
-                mc_running_avg_y_ptr += mc_avg_y_stride;
-                running_avg_y_ptr += avg_y_stride;
+  LD4(sig_start, sig_stride, src0, src1, src2, src3);
+  sig_start += (4 * sig_stride);
+  SD4(src0, src1, src2, src3, running_avg_y_start, avg_y_stride);
+  running_avg_y_start += (4 * avg_y_stride);
 
-                mc_running_avg_y1 = LD_UB(mc_running_avg_y_ptr);
-                sig1 = LD_UB(sig_ptr);
-                running_avg_y1 = LD_UB(running_avg_y_ptr);
+  LD4(sig_start, sig_stride, src0, src1, src2, src3);
+  SD4(src0, src1, src2, src3, running_avg_y_start, avg_y_stride);
 
-                coeff0 = (v16u8)__msa_ilvr_b((v16i8)mc_running_avg_y0,
-                                             (v16i8)sig0);
-                diff0 = __msa_hsub_u_h(coeff0, coeff0);
-                abs_diff0 = __msa_add_a_h(diff0, (v8i16)zero);
-                temp0_h = delta_vec < abs_diff0;
-                abs_diff0 = (v8i16)__msa_bmnz_v((v16u8)abs_diff0,
-                                                (v16u8)delta_vec,
-                                                (v16u8)temp0_h);
-                abs_diff_neg0 = (v8i16)zero - abs_diff0;
-                temp0_h = __msa_clei_s_h(diff0, 0);
-                adjust0 = (v8i16)__msa_bmz_v((v16u8)abs_diff0,
-                                             (v16u8)abs_diff_neg0,
-                                             (v16u8)temp0_h);
-                temp2_h = (v8i16)__msa_ilvr_b(zero, (v16i8)running_avg_y);
-                adjust2 = temp2_h + adjust0;
-                adjust2 = __msa_maxi_s_h(adjust2, 0);
-                adjust2 = (v8i16)__msa_sat_u_h((v8u16)adjust2, 7);
-                temp0_h = __msa_ceqi_h(diff0, 0);
-                adjust2 = (v8i16)__msa_bmnz_v((v16u8)adjust2, (v16u8)temp2_h,
-                                              (v16u8)temp0_h);
-                adjust0 = (v8i16)__msa_bmnz_v((v16u8)adjust0, (v16u8)zero,
-                                              (v16u8)temp0_h);
-                col_sum0 += adjust0;
-                running_avg_y = (v16u8)__msa_pckev_b((v16i8)adjust2,
-                                                     (v16i8)adjust2);
-                dst0 = __msa_copy_s_d((v2i64)running_avg_y, 0);
-                SD(dst0, running_avg_y_ptr - avg_y_stride);
-
-                coeff0 = (v16u8)__msa_ilvr_b((v16i8)mc_running_avg_y1,
-                                             (v16i8)sig1);
-                diff0 = __msa_hsub_u_h(coeff0, coeff0);
-                abs_diff0 = __msa_add_a_h(diff0, (v8i16)zero);
-                temp0_h = delta_vec < abs_diff0;
-                abs_diff0 = (v8i16)__msa_bmnz_v((v16u8)abs_diff0,
-                                                (v16u8)delta_vec,
-                                                (v16u8)temp0_h);
-                abs_diff_neg0 = (v8i16)zero - abs_diff0;
-                temp0_h = __msa_clei_s_h(diff0, 0);
-                adjust0 = (v8i16)__msa_bmz_v((v16u8)abs_diff0,
-                                             (v16u8)abs_diff_neg0,
-                                             (v16u8)temp0_h);
-                temp2_h = (v8i16)__msa_ilvr_b(zero, (v16i8)running_avg_y1);
-                adjust2 = temp2_h + adjust0;
-                adjust2 = __msa_maxi_s_h(adjust2, 0);
-                adjust2 = (v8i16)__msa_sat_u_h((v8u16)adjust2, 7);
-                temp0_h = __msa_ceqi_h(diff0, 0);
-                adjust2 = (v8i16)__msa_bmnz_v((v16u8)adjust2, (v16u8)temp2_h,
-                                              (v16u8)temp0_h);
-                adjust0 = (v8i16)__msa_bmnz_v((v16u8)adjust0, (v16u8)zero,
-                                              (v16u8)temp0_h);
-                col_sum0 += adjust0;
-                running_avg_y = (v16u8)__msa_pckev_b((v16i8)adjust2,
-                                                     (v16i8)adjust2);
-                dst1 = __msa_copy_s_d((v2i64)running_avg_y, 0);
-                SD(dst1, running_avg_y_ptr);
-                running_avg_y_ptr += avg_y_stride;
-            }
-
-            temp0_h = col_sum0;
-            temp0_w = __msa_hadd_s_w(temp0_h, temp0_h);
-            temp0_d = __msa_hadd_s_d(temp0_w, temp0_w);
-            temp1_d = __msa_splati_d(temp0_d, 1);
-            temp0_d += temp1_d;
-            sum_diff = __msa_copy_s_w((v4i32)temp0_d, 0);
-
-            if (abs(sum_diff) > sum_diff_thresh)
-            {
-                return COPY_BLOCK;
-            }
-        }
-        else
-        {
-            return COPY_BLOCK;
-        }
-    }
-
-    LD4(sig_start, sig_stride, src0, src1, src2, src3);
-    sig_start += (4 * sig_stride);
-    SD4(src0, src1, src2, src3, running_avg_y_start, avg_y_stride);
-    running_avg_y_start += (4 * avg_y_stride);
-
-    LD4(sig_start, sig_stride, src0, src1, src2, src3);
-    SD4(src0, src1, src2, src3, running_avg_y_start, avg_y_stride);
-
-    return FILTER_BLOCK;
+  return FILTER_BLOCK;
 }
--- a/vp8/encoder/mips/msa/encodeopt_msa.c
+++ b/vp8/encoder/mips/msa/encodeopt_msa.c
@@ -12,163 +12,156 @@
 #include "vp8/common/mips/msa/vp8_macros_msa.h"
 #include "vp8/encoder/block.h"
 
-int32_t vp8_block_error_msa(int16_t *coeff_ptr, int16_t *dq_coeff_ptr)
-{
-    int32_t err = 0;
-    uint32_t loop_cnt;
-    v8i16 coeff, dq_coeff, coeff0, coeff1;
-    v4i32 diff0, diff1;
-    v2i64 err0 = { 0 };
-    v2i64 err1 = { 0 };
+int32_t vp8_block_error_msa(int16_t *coeff_ptr, int16_t *dq_coeff_ptr) {
+  int32_t err = 0;
+  uint32_t loop_cnt;
+  v8i16 coeff, dq_coeff, coeff0, coeff1;
+  v4i32 diff0, diff1;
+  v2i64 err0 = { 0 };
+  v2i64 err1 = { 0 };
 
-    for (loop_cnt = 2; loop_cnt--;)
-    {
-        coeff = LD_SH(coeff_ptr);
-        dq_coeff = LD_SH(dq_coeff_ptr);
-        ILVRL_H2_SH(coeff, dq_coeff, coeff0, coeff1);
-        HSUB_UH2_SW(coeff0, coeff1, diff0, diff1);
-        DPADD_SD2_SD(diff0, diff1, err0, err1);
-        coeff_ptr += 8;
-        dq_coeff_ptr += 8;
-    }
+  for (loop_cnt = 2; loop_cnt--;) {
+    coeff = LD_SH(coeff_ptr);
+    dq_coeff = LD_SH(dq_coeff_ptr);
+    ILVRL_H2_SH(coeff, dq_coeff, coeff0, coeff1);
+    HSUB_UH2_SW(coeff0, coeff1, diff0, diff1);
+    DPADD_SD2_SD(diff0, diff1, err0, err1);
+    coeff_ptr += 8;
+    dq_coeff_ptr += 8;
+  }
 
-    err0 += __msa_splati_d(err0, 1);
-    err1 += __msa_splati_d(err1, 1);
-    err = __msa_copy_s_d(err0, 0);
-    err += __msa_copy_s_d(err1, 0);
+  err0 += __msa_splati_d(err0, 1);
+  err1 += __msa_splati_d(err1, 1);
+  err = __msa_copy_s_d(err0, 0);
+  err += __msa_copy_s_d(err1, 0);
 
-    return err;
+  return err;
 }
 
-int32_t vp8_mbblock_error_msa(MACROBLOCK *mb, int32_t dc)
-{
-    BLOCK *be;
-    BLOCKD *bd;
-    int16_t *coeff_ptr, *dq_coeff_ptr;
-    int32_t err = 0;
-    uint32_t loop_cnt;
-    v8i16 coeff, coeff0, coeff1, coeff2, coeff3, coeff4;
-    v8i16 dq_coeff, dq_coeff2, dq_coeff3, dq_coeff4;
-    v4i32 diff0, diff1;
-    v2i64 err0, err1;
-    v16u8 zero  = { 0 };
-    v16u8 mask0 = (v16u8)__msa_ldi_b(255);
+int32_t vp8_mbblock_error_msa(MACROBLOCK *mb, int32_t dc) {
+  BLOCK *be;
+  BLOCKD *bd;
+  int16_t *coeff_ptr, *dq_coeff_ptr;
+  int32_t err = 0;
+  uint32_t loop_cnt;
+  v8i16 coeff, coeff0, coeff1, coeff2, coeff3, coeff4;
+  v8i16 dq_coeff, dq_coeff2, dq_coeff3, dq_coeff4;
+  v4i32 diff0, diff1;
+  v2i64 err0, err1;
+  v16u8 zero = { 0 };
+  v16u8 mask0 = (v16u8)__msa_ldi_b(255);
 
-    if (1 == dc)
-    {
-        mask0 = (v16u8)__msa_insve_w((v4i32)mask0, 0, (v4i32)zero);
-    }
+  if (1 == dc) {
+    mask0 = (v16u8)__msa_insve_w((v4i32)mask0, 0, (v4i32)zero);
+  }
 
-    for (loop_cnt = 0; loop_cnt < 8; loop_cnt++)
-    {
-        be = &mb->block[2 * loop_cnt];
-        bd = &mb->e_mbd.block[2 * loop_cnt];
-        coeff_ptr = be->coeff;
-        dq_coeff_ptr = bd->dqcoeff;
-        coeff = LD_SH(coeff_ptr);
-        dq_coeff = LD_SH(dq_coeff_ptr);
-        coeff_ptr += 8;
-        dq_coeff_ptr += 8;
-        coeff2 = LD_SH(coeff_ptr);
-        dq_coeff2 = LD_SH(dq_coeff_ptr);
-        be = &mb->block[2 * loop_cnt + 1];
-        bd = &mb->e_mbd.block[2 * loop_cnt + 1];
-        coeff_ptr = be->coeff;
-        dq_coeff_ptr = bd->dqcoeff;
-        coeff3 = LD_SH(coeff_ptr);
-        dq_coeff3 = LD_SH(dq_coeff_ptr);
-        coeff_ptr += 8;
-        dq_coeff_ptr += 8;
-        coeff4 = LD_SH(coeff_ptr);
-        dq_coeff4 = LD_SH(dq_coeff_ptr);
-        ILVRL_H2_SH(coeff, dq_coeff, coeff0, coeff1);
-        HSUB_UH2_SW(coeff0, coeff1, diff0, diff1);
-        diff0 = (v4i32)__msa_bmnz_v(zero, (v16u8)diff0, mask0);
-        DOTP_SW2_SD(diff0, diff1, diff0, diff1, err0, err1);
-        ILVRL_H2_SH(coeff2, dq_coeff2, coeff0, coeff1);
-        HSUB_UH2_SW(coeff0, coeff1, diff0, diff1);
-        DPADD_SD2_SD(diff0, diff1, err0, err1);
-        err0 += __msa_splati_d(err0, 1);
-        err1 += __msa_splati_d(err1, 1);
-        err += __msa_copy_s_d(err0, 0);
-        err += __msa_copy_s_d(err1, 0);
+  for (loop_cnt = 0; loop_cnt < 8; loop_cnt++) {
+    be = &mb->block[2 * loop_cnt];
+    bd = &mb->e_mbd.block[2 * loop_cnt];
+    coeff_ptr = be->coeff;
+    dq_coeff_ptr = bd->dqcoeff;
+    coeff = LD_SH(coeff_ptr);
+    dq_coeff = LD_SH(dq_coeff_ptr);
+    coeff_ptr += 8;
+    dq_coeff_ptr += 8;
+    coeff2 = LD_SH(coeff_ptr);
+    dq_coeff2 = LD_SH(dq_coeff_ptr);
+    be = &mb->block[2 * loop_cnt + 1];
+    bd = &mb->e_mbd.block[2 * loop_cnt + 1];
+    coeff_ptr = be->coeff;
+    dq_coeff_ptr = bd->dqcoeff;
+    coeff3 = LD_SH(coeff_ptr);
+    dq_coeff3 = LD_SH(dq_coeff_ptr);
+    coeff_ptr += 8;
+    dq_coeff_ptr += 8;
+    coeff4 = LD_SH(coeff_ptr);
+    dq_coeff4 = LD_SH(dq_coeff_ptr);
+    ILVRL_H2_SH(coeff, dq_coeff, coeff0, coeff1);
+    HSUB_UH2_SW(coeff0, coeff1, diff0, diff1);
+    diff0 = (v4i32)__msa_bmnz_v(zero, (v16u8)diff0, mask0);
+    DOTP_SW2_SD(diff0, diff1, diff0, diff1, err0, err1);
+    ILVRL_H2_SH(coeff2, dq_coeff2, coeff0, coeff1);
+    HSUB_UH2_SW(coeff0, coeff1, diff0, diff1);
+    DPADD_SD2_SD(diff0, diff1, err0, err1);
+    err0 += __msa_splati_d(err0, 1);
+    err1 += __msa_splati_d(err1, 1);
+    err += __msa_copy_s_d(err0, 0);
+    err += __msa_copy_s_d(err1, 0);
 
-        ILVRL_H2_SH(coeff3, dq_coeff3, coeff0, coeff1);
-        HSUB_UH2_SW(coeff0, coeff1, diff0, diff1);
-        diff0 = (v4i32)__msa_bmnz_v(zero, (v16u8)diff0, mask0);
-        DOTP_SW2_SD(diff0, diff1, diff0, diff1, err0, err1);
-        ILVRL_H2_SH(coeff4, dq_coeff4, coeff0, coeff1);
-        HSUB_UH2_SW(coeff0, coeff1, diff0, diff1);
-        DPADD_SD2_SD(diff0, diff1, err0, err1);
-        err0 += __msa_splati_d(err0, 1);
-        err1 += __msa_splati_d(err1, 1);
-        err += __msa_copy_s_d(err0, 0);
-        err += __msa_copy_s_d(err1, 0);
-    }
+    ILVRL_H2_SH(coeff3, dq_coeff3, coeff0, coeff1);
+    HSUB_UH2_SW(coeff0, coeff1, diff0, diff1);
+    diff0 = (v4i32)__msa_bmnz_v(zero, (v16u8)diff0, mask0);
+    DOTP_SW2_SD(diff0, diff1, diff0, diff1, err0, err1);
+    ILVRL_H2_SH(coeff4, dq_coeff4, coeff0, coeff1);
+    HSUB_UH2_SW(coeff0, coeff1, diff0, diff1);
+    DPADD_SD2_SD(diff0, diff1, err0, err1);
+    err0 += __msa_splati_d(err0, 1);
+    err1 += __msa_splati_d(err1, 1);
+    err += __msa_copy_s_d(err0, 0);
+    err += __msa_copy_s_d(err1, 0);
+  }
 
-    return err;
+  return err;
 }
 
-int32_t vp8_mbuverror_msa(MACROBLOCK *mb)
-{
-    BLOCK *be;
-    BLOCKD *bd;
-    int16_t *coeff_ptr, *dq_coeff_ptr;
-    int32_t err = 0;
-    uint32_t loop_cnt;
-    v8i16 coeff, coeff0, coeff1, coeff2, coeff3, coeff4;
-    v8i16 dq_coeff, dq_coeff2, dq_coeff3, dq_coeff4;
-    v4i32 diff0, diff1;
-    v2i64 err0, err1, err_dup0, err_dup1;
+int32_t vp8_mbuverror_msa(MACROBLOCK *mb) {
+  BLOCK *be;
+  BLOCKD *bd;
+  int16_t *coeff_ptr, *dq_coeff_ptr;
+  int32_t err = 0;
+  uint32_t loop_cnt;
+  v8i16 coeff, coeff0, coeff1, coeff2, coeff3, coeff4;
+  v8i16 dq_coeff, dq_coeff2, dq_coeff3, dq_coeff4;
+  v4i32 diff0, diff1;
+  v2i64 err0, err1, err_dup0, err_dup1;
 
-    for (loop_cnt = 16; loop_cnt < 24; loop_cnt += 2)
-    {
-        be = &mb->block[loop_cnt];
-        bd = &mb->e_mbd.block[loop_cnt];
-        coeff_ptr = be->coeff;
-        dq_coeff_ptr = bd->dqcoeff;
-        coeff = LD_SH(coeff_ptr);
-        dq_coeff = LD_SH(dq_coeff_ptr);
-        coeff_ptr += 8;
-        dq_coeff_ptr += 8;
-        coeff2 = LD_SH(coeff_ptr);
-        dq_coeff2 = LD_SH(dq_coeff_ptr);
-        be = &mb->block[loop_cnt + 1];
-        bd = &mb->e_mbd.block[loop_cnt + 1];
-        coeff_ptr = be->coeff;
-        dq_coeff_ptr = bd->dqcoeff;
-        coeff3 = LD_SH(coeff_ptr);
-        dq_coeff3 = LD_SH(dq_coeff_ptr);
-        coeff_ptr += 8;
-        dq_coeff_ptr += 8;
-        coeff4 = LD_SH(coeff_ptr);
-        dq_coeff4 = LD_SH(dq_coeff_ptr);
+  for (loop_cnt = 16; loop_cnt < 24; loop_cnt += 2) {
+    be = &mb->block[loop_cnt];
+    bd = &mb->e_mbd.block[loop_cnt];
+    coeff_ptr = be->coeff;
+    dq_coeff_ptr = bd->dqcoeff;
+    coeff = LD_SH(coeff_ptr);
+    dq_coeff = LD_SH(dq_coeff_ptr);
+    coeff_ptr += 8;
+    dq_coeff_ptr += 8;
+    coeff2 = LD_SH(coeff_ptr);
+    dq_coeff2 = LD_SH(dq_coeff_ptr);
+    be = &mb->block[loop_cnt + 1];
+    bd = &mb->e_mbd.block[loop_cnt + 1];
+    coeff_ptr = be->coeff;
+    dq_coeff_ptr = bd->dqcoeff;
+    coeff3 = LD_SH(coeff_ptr);
+    dq_coeff3 = LD_SH(dq_coeff_ptr);
+    coeff_ptr += 8;
+    dq_coeff_ptr += 8;
+    coeff4 = LD_SH(coeff_ptr);
+    dq_coeff4 = LD_SH(dq_coeff_ptr);
 
-        ILVRL_H2_SH(coeff, dq_coeff, coeff0, coeff1);
-        HSUB_UH2_SW(coeff0, coeff1, diff0, diff1);
-        DOTP_SW2_SD(diff0, diff1, diff0, diff1, err0, err1);
+    ILVRL_H2_SH(coeff, dq_coeff, coeff0, coeff1);
+    HSUB_UH2_SW(coeff0, coeff1, diff0, diff1);
+    DOTP_SW2_SD(diff0, diff1, diff0, diff1, err0, err1);
 
-        ILVRL_H2_SH(coeff2, dq_coeff2, coeff0, coeff1);
-        HSUB_UH2_SW(coeff0, coeff1, diff0, diff1);
-        DPADD_SD2_SD(diff0, diff1, err0, err1);
-        err_dup0 = __msa_splati_d(err0, 1);
-        err_dup1 = __msa_splati_d(err1, 1);
-        ADD2(err0, err_dup0, err1, err_dup1, err0, err1);
-        err += __msa_copy_s_d(err0, 0);
-        err += __msa_copy_s_d(err1, 0);
+    ILVRL_H2_SH(coeff2, dq_coeff2, coeff0, coeff1);
+    HSUB_UH2_SW(coeff0, coeff1, diff0, diff1);
+    DPADD_SD2_SD(diff0, diff1, err0, err1);
+    err_dup0 = __msa_splati_d(err0, 1);
+    err_dup1 = __msa_splati_d(err1, 1);
+    ADD2(err0, err_dup0, err1, err_dup1, err0, err1);
+    err += __msa_copy_s_d(err0, 0);
+    err += __msa_copy_s_d(err1, 0);
 
-        ILVRL_H2_SH(coeff3, dq_coeff3, coeff0, coeff1);
-        HSUB_UH2_SW(coeff0, coeff1, diff0, diff1);
-        DOTP_SW2_SD(diff0, diff1, diff0, diff1, err0, err1);
-        ILVRL_H2_SH(coeff4, dq_coeff4, coeff0, coeff1);
-        HSUB_UH2_SW(coeff0, coeff1, diff0, diff1);
-        DPADD_SD2_SD(diff0, diff1, err0, err1);
-        err_dup0 = __msa_splati_d(err0, 1);
-        err_dup1 = __msa_splati_d(err1, 1);
-        ADD2(err0, err_dup0, err1, err_dup1, err0, err1);
-        err += __msa_copy_s_d(err0, 0);
-        err += __msa_copy_s_d(err1, 0);
-    }
+    ILVRL_H2_SH(coeff3, dq_coeff3, coeff0, coeff1);
+    HSUB_UH2_SW(coeff0, coeff1, diff0, diff1);
+    DOTP_SW2_SD(diff0, diff1, diff0, diff1, err0, err1);
+    ILVRL_H2_SH(coeff4, dq_coeff4, coeff0, coeff1);
+    HSUB_UH2_SW(coeff0, coeff1, diff0, diff1);
+    DPADD_SD2_SD(diff0, diff1, err0, err1);
+    err_dup0 = __msa_splati_d(err0, 1);
+    err_dup1 = __msa_splati_d(err1, 1);
+    ADD2(err0, err_dup0, err1, err_dup1, err0, err1);
+    err += __msa_copy_s_d(err0, 0);
+    err += __msa_copy_s_d(err1, 0);
+  }
 
-    return err;
+  return err;
 }
--- a/vp8/encoder/mips/msa/quantize_msa.c
+++ b/vp8/encoder/mips/msa/quantize_msa.c
@@ -15,232 +15,199 @@
 static int8_t fast_quantize_b_msa(int16_t *coeff_ptr, int16_t *zbin,
                                   int16_t *round, int16_t *quant,
                                   int16_t *de_quant, int16_t *q_coeff,
-                                  int16_t *dq_coeff)
-{
-    int32_t cnt, eob;
-    v16i8 inv_zig_zag = { 0, 1, 5, 6, 2, 4, 7, 12,
-                          3, 8, 11, 13, 9, 10, 14, 15 };
-    v8i16 round0, round1;
-    v8i16 sign_z0, sign_z1;
-    v8i16 q_coeff0, q_coeff1;
-    v8i16 x0, x1, de_quant0, de_quant1;
-    v8i16 coeff0, coeff1, z0, z1;
-    v8i16 quant0, quant1, quant2, quant3;
-    v8i16 zero = { 0 };
-    v8i16 inv_zig_zag0, inv_zig_zag1;
-    v8i16 zigzag_mask0 = { 0, 1, 4, 8, 5, 2, 3, 6 };
-    v8i16 zigzag_mask1 = { 9, 12, 13, 10, 7, 11, 14, 15 };
-    v8i16 temp0_h, temp1_h, temp2_h, temp3_h;
-    v4i32 temp0_w, temp1_w, temp2_w, temp3_w;
+                                  int16_t *dq_coeff) {
+  int32_t cnt, eob;
+  v16i8 inv_zig_zag = { 0, 1, 5, 6, 2, 4, 7, 12, 3, 8, 11, 13, 9, 10, 14, 15 };
+  v8i16 round0, round1;
+  v8i16 sign_z0, sign_z1;
+  v8i16 q_coeff0, q_coeff1;
+  v8i16 x0, x1, de_quant0, de_quant1;
+  v8i16 coeff0, coeff1, z0, z1;
+  v8i16 quant0, quant1, quant2, quant3;
+  v8i16 zero = { 0 };
+  v8i16 inv_zig_zag0, inv_zig_zag1;
+  v8i16 zigzag_mask0 = { 0, 1, 4, 8, 5, 2, 3, 6 };
+  v8i16 zigzag_mask1 = { 9, 12, 13, 10, 7, 11, 14, 15 };
+  v8i16 temp0_h, temp1_h, temp2_h, temp3_h;
+  v4i32 temp0_w, temp1_w, temp2_w, temp3_w;
 
-    ILVRL_B2_SH(zero, inv_zig_zag, inv_zig_zag0, inv_zig_zag1);
-    eob = -1;
-    LD_SH2(coeff_ptr, 8, coeff0, coeff1);
-    VSHF_H2_SH(coeff0, coeff1, coeff0, coeff1, zigzag_mask0, zigzag_mask1,
-               z0, z1);
-    LD_SH2(round, 8, coeff0, coeff1);
-    VSHF_H2_SH(coeff0, coeff1, coeff0, coeff1, zigzag_mask0, zigzag_mask1,
-               round0, round1);
-    LD_SH2(quant, 8, coeff0, coeff1);
-    VSHF_H2_SH(coeff0, coeff1, coeff0, coeff1, zigzag_mask0, zigzag_mask1,
-               quant0, quant2);
-    sign_z0 = z0 >> 15;
-    sign_z1 = z1 >> 15;
-    x0 = __msa_add_a_h(z0, zero);
-    x1 = __msa_add_a_h(z1, zero);
-    ILVL_H2_SH(quant0, quant0, quant2, quant2, quant1, quant3);
-    ILVR_H2_SH(quant0, quant0, quant2, quant2, quant0, quant2);
-    ILVL_H2_SH(round0, x0, round1, x1, temp1_h, temp3_h);
-    ILVR_H2_SH(round0, x0, round1, x1, temp0_h, temp2_h);
-    DOTP_SH4_SW(temp0_h, temp1_h, temp2_h, temp3_h, quant0, quant1, quant2,
-                quant3, temp0_w, temp1_w, temp2_w, temp3_w);
-    SRA_4V(temp0_w, temp1_w, temp2_w, temp3_w, 16);
-    PCKEV_H2_SH(temp1_w, temp0_w, temp3_w, temp2_w, x0, x1);
-    x0 = x0 ^ sign_z0;
-    x1 = x1 ^ sign_z1;
-    SUB2(x0, sign_z0, x1, sign_z1, x0, x1);
-    VSHF_H2_SH(x0, x1, x0, x1, inv_zig_zag0, inv_zig_zag1, q_coeff0, q_coeff1);
-    ST_SH2(q_coeff0, q_coeff1, q_coeff, 8);
-    LD_SH2(de_quant, 8, de_quant0, de_quant1);
-    q_coeff0 *= de_quant0;
-    q_coeff1 *= de_quant1;
-    ST_SH2(q_coeff0, q_coeff1, dq_coeff, 8);
+  ILVRL_B2_SH(zero, inv_zig_zag, inv_zig_zag0, inv_zig_zag1);
+  eob = -1;
+  LD_SH2(coeff_ptr, 8, coeff0, coeff1);
+  VSHF_H2_SH(coeff0, coeff1, coeff0, coeff1, zigzag_mask0, zigzag_mask1, z0,
+             z1);
+  LD_SH2(round, 8, coeff0, coeff1);
+  VSHF_H2_SH(coeff0, coeff1, coeff0, coeff1, zigzag_mask0, zigzag_mask1, round0,
+             round1);
+  LD_SH2(quant, 8, coeff0, coeff1);
+  VSHF_H2_SH(coeff0, coeff1, coeff0, coeff1, zigzag_mask0, zigzag_mask1, quant0,
+             quant2);
+  sign_z0 = z0 >> 15;
+  sign_z1 = z1 >> 15;
+  x0 = __msa_add_a_h(z0, zero);
+  x1 = __msa_add_a_h(z1, zero);
+  ILVL_H2_SH(quant0, quant0, quant2, quant2, quant1, quant3);
+  ILVR_H2_SH(quant0, quant0, quant2, quant2, quant0, quant2);
+  ILVL_H2_SH(round0, x0, round1, x1, temp1_h, temp3_h);
+  ILVR_H2_SH(round0, x0, round1, x1, temp0_h, temp2_h);
+  DOTP_SH4_SW(temp0_h, temp1_h, temp2_h, temp3_h, quant0, quant1, quant2,
+              quant3, temp0_w, temp1_w, temp2_w, temp3_w);
+  SRA_4V(temp0_w, temp1_w, temp2_w, temp3_w, 16);
+  PCKEV_H2_SH(temp1_w, temp0_w, temp3_w, temp2_w, x0, x1);
+  x0 = x0 ^ sign_z0;
+  x1 = x1 ^ sign_z1;
+  SUB2(x0, sign_z0, x1, sign_z1, x0, x1);
+  VSHF_H2_SH(x0, x1, x0, x1, inv_zig_zag0, inv_zig_zag1, q_coeff0, q_coeff1);
+  ST_SH2(q_coeff0, q_coeff1, q_coeff, 8);
+  LD_SH2(de_quant, 8, de_quant0, de_quant1);
+  q_coeff0 *= de_quant0;
+  q_coeff1 *= de_quant1;
+  ST_SH2(q_coeff0, q_coeff1, dq_coeff, 8);
 
-    for (cnt = 0; cnt < 16; ++cnt)
-    {
-        if ((cnt <= 7) && (x1[7 - cnt] != 0))
-        {
-            eob = (15 - cnt);
-            break;
-        }
+  for (cnt = 0; cnt < 16; ++cnt) {
+    if ((cnt <= 7) && (x1[7 - cnt] != 0)) {
+      eob = (15 - cnt);
+      break;
+    }
 
-        if ((cnt > 7) && (x0[7 - (cnt - 8)] != 0))
-        {
-            eob = (7 - (cnt - 8));
-            break;
-        }
+    if ((cnt > 7) && (x0[7 - (cnt - 8)] != 0)) {
+      eob = (7 - (cnt - 8));
+      break;
     }
+  }
 
-    return (int8_t)(eob + 1);
+  return (int8_t)(eob + 1);
 }
 
-static int8_t exact_regular_quantize_b_msa(int16_t *zbin_boost,
-                                           int16_t *coeff_ptr,
-                                           int16_t *zbin,
-                                           int16_t *round,
-                                           int16_t *quant,
-                                           int16_t *quant_shift,
-                                           int16_t *de_quant,
-                                           int16_t zbin_oq_in,
-                                           int16_t *q_coeff,
-                                           int16_t *dq_coeff)
-{
-    int32_t cnt, eob;
-    int16_t *boost_temp = zbin_boost;
-    v16i8 inv_zig_zag = { 0, 1, 5, 6, 2, 4, 7, 12,
-                          3, 8, 11, 13, 9, 10, 14, 15 };
-    v8i16 round0, round1;
-    v8i16 sign_z0, sign_z1;
-    v8i16 q_coeff0, q_coeff1;
-    v8i16 z_bin0, z_bin1, zbin_o_q;
-    v8i16 x0, x1, sign_x0, sign_x1, de_quant0, de_quant1;
-    v8i16 coeff0, coeff1, z0, z1;
-    v8i16 quant0, quant1, quant2, quant3;
-    v8i16 zero = { 0 };
-    v8i16 inv_zig_zag0, inv_zig_zag1;
-    v8i16 zigzag_mask0 = { 0, 1, 4, 8, 5, 2, 3, 6 };
-    v8i16 zigzag_mask1 = { 9, 12, 13, 10, 7, 11, 14, 15 };
-    v8i16 temp0_h, temp1_h, temp2_h, temp3_h;
-    v4i32 temp0_w, temp1_w, temp2_w, temp3_w;
+static int8_t exact_regular_quantize_b_msa(
+    int16_t *zbin_boost, int16_t *coeff_ptr, int16_t *zbin, int16_t *round,
+    int16_t *quant, int16_t *quant_shift, int16_t *de_quant, int16_t zbin_oq_in,
+    int16_t *q_coeff, int16_t *dq_coeff) {
+  int32_t cnt, eob;
+  int16_t *boost_temp = zbin_boost;
+  v16i8 inv_zig_zag = { 0, 1, 5, 6, 2, 4, 7, 12, 3, 8, 11, 13, 9, 10, 14, 15 };
+  v8i16 round0, round1;
+  v8i16 sign_z0, sign_z1;
+  v8i16 q_coeff0, q_coeff1;
+  v8i16 z_bin0, z_bin1, zbin_o_q;
+  v8i16 x0, x1, sign_x0, sign_x1, de_quant0, de_quant1;
+  v8i16 coeff0, coeff1, z0, z1;
+  v8i16 quant0, quant1, quant2, quant3;
+  v8i16 zero = { 0 };
+  v8i16 inv_zig_zag0, inv_zig_zag1;
+  v8i16 zigzag_mask0 = { 0, 1, 4, 8, 5, 2, 3, 6 };
+  v8i16 zigzag_mask1 = { 9, 12, 13, 10, 7, 11, 14, 15 };
+  v8i16 temp0_h, temp1_h, temp2_h, temp3_h;
+  v4i32 temp0_w, temp1_w, temp2_w, temp3_w;
 
-    ILVRL_B2_SH(zero, inv_zig_zag, inv_zig_zag0, inv_zig_zag1);
-    zbin_o_q = __msa_fill_h(zbin_oq_in);
-    eob = -1;
-    LD_SH2(coeff_ptr, 8, coeff0, coeff1);
-    VSHF_H2_SH(coeff0, coeff1, coeff0, coeff1, zigzag_mask0, zigzag_mask1,
-               z0, z1);
-    LD_SH2(round, 8, coeff0, coeff1);
-    VSHF_H2_SH(coeff0, coeff1, coeff0, coeff1, zigzag_mask0, zigzag_mask1,
-               round0, round1);
-    LD_SH2(quant, 8, coeff0, coeff1);
-    VSHF_H2_SH(coeff0, coeff1, coeff0, coeff1, zigzag_mask0, zigzag_mask1,
-               quant0, quant2);
-    LD_SH2(zbin, 8, coeff0, coeff1);
-    VSHF_H2_SH(coeff0, coeff1, coeff0, coeff1, zigzag_mask0, zigzag_mask1,
-               z_bin0, z_bin1);
-    sign_z0 = z0 >> 15;
-    sign_z1 = z1 >> 15;
-    x0 = __msa_add_a_h(z0, zero);
-    x1 = __msa_add_a_h(z1, zero);
-    SUB2(x0, z_bin0, x1, z_bin1, z_bin0, z_bin1);
-    SUB2(z_bin0, zbin_o_q, z_bin1, zbin_o_q, z_bin0, z_bin1);
-    ILVL_H2_SH(quant0, quant0, quant2, quant2, quant1, quant3);
-    ILVR_H2_SH(quant0, quant0, quant2, quant2, quant0, quant2);
-    ILVL_H2_SH(round0, x0, round1, x1, temp1_h, temp3_h);
-    ILVR_H2_SH(round0, x0, round1, x1, temp0_h, temp2_h);
-    DOTP_SH4_SW(temp0_h, temp1_h, temp2_h, temp3_h, quant0, quant1, quant2,
-                quant3, temp0_w, temp1_w, temp2_w, temp3_w);
-    SRA_4V(temp0_w, temp1_w, temp2_w, temp3_w, 16);
-    PCKEV_H2_SH(temp1_w, temp0_w, temp3_w, temp2_w, temp0_h, temp2_h);
-    LD_SH2(quant_shift, 8, coeff0, coeff1);
-    VSHF_H2_SH(coeff0, coeff1, coeff0, coeff1, zigzag_mask0, zigzag_mask1,
-               quant0, quant2);
-    ILVL_H2_SH(quant0, quant0, quant2, quant2, quant1, quant3);
-    ILVR_H2_SH(quant0, quant0, quant2, quant2, quant0, quant2);
-    ADD2(x0, round0, x1, round1, x0, x1);
-    ILVL_H2_SH(temp0_h, x0, temp2_h, x1, temp1_h, temp3_h);
-    ILVR_H2_SH(temp0_h, x0, temp2_h, x1, temp0_h, temp2_h);
-    DOTP_SH4_SW(temp0_h, temp1_h, temp2_h, temp3_h, quant0, quant1, quant2,
-                quant3, temp0_w, temp1_w, temp2_w, temp3_w);
-    SRA_4V(temp0_w, temp1_w, temp2_w, temp3_w, 16);
-    PCKEV_H2_SH(temp1_w, temp0_w, temp3_w, temp2_w, x0, x1);
-    sign_x0 = x0 ^ sign_z0;
-    sign_x1 = x1 ^ sign_z1;
-    SUB2(sign_x0, sign_z0, sign_x1, sign_z1, sign_x0, sign_x1);
-    for (cnt = 0; cnt < 16; ++cnt)
-    {
-        if (cnt <= 7)
-        {
-            if (boost_temp[0] <= z_bin0[cnt])
-            {
-                if (x0[cnt])
-                {
-                    eob = cnt;
-                    boost_temp = zbin_boost;
-                }
-                else
-                {
-                    boost_temp++;
-                }
-            }
-            else
-            {
-                sign_x0[cnt] = 0;
-                boost_temp++;
-            }
+  ILVRL_B2_SH(zero, inv_zig_zag, inv_zig_zag0, inv_zig_zag1);
+  zbin_o_q = __msa_fill_h(zbin_oq_in);
+  eob = -1;
+  LD_SH2(coeff_ptr, 8, coeff0, coeff1);
+  VSHF_H2_SH(coeff0, coeff1, coeff0, coeff1, zigzag_mask0, zigzag_mask1, z0,
+             z1);
+  LD_SH2(round, 8, coeff0, coeff1);
+  VSHF_H2_SH(coeff0, coeff1, coeff0, coeff1, zigzag_mask0, zigzag_mask1, round0,
+             round1);
+  LD_SH2(quant, 8, coeff0, coeff1);
+  VSHF_H2_SH(coeff0, coeff1, coeff0, coeff1, zigzag_mask0, zigzag_mask1, quant0,
+             quant2);
+  LD_SH2(zbin, 8, coeff0, coeff1);
+  VSHF_H2_SH(coeff0, coeff1, coeff0, coeff1, zigzag_mask0, zigzag_mask1, z_bin0,
+             z_bin1);
+  sign_z0 = z0 >> 15;
+  sign_z1 = z1 >> 15;
+  x0 = __msa_add_a_h(z0, zero);
+  x1 = __msa_add_a_h(z1, zero);
+  SUB2(x0, z_bin0, x1, z_bin1, z_bin0, z_bin1);
+  SUB2(z_bin0, zbin_o_q, z_bin1, zbin_o_q, z_bin0, z_bin1);
+  ILVL_H2_SH(quant0, quant0, quant2, quant2, quant1, quant3);
+  ILVR_H2_SH(quant0, quant0, quant2, quant2, quant0, quant2);
+  ILVL_H2_SH(round0, x0, round1, x1, temp1_h, temp3_h);
+  ILVR_H2_SH(round0, x0, round1, x1, temp0_h, temp2_h);
+  DOTP_SH4_SW(temp0_h, temp1_h, temp2_h, temp3_h, quant0, quant1, quant2,
+              quant3, temp0_w, temp1_w, temp2_w, temp3_w);
+  SRA_4V(temp0_w, temp1_w, temp2_w, temp3_w, 16);
+  PCKEV_H2_SH(temp1_w, temp0_w, temp3_w, temp2_w, temp0_h, temp2_h);
+  LD_SH2(quant_shift, 8, coeff0, coeff1);
+  VSHF_H2_SH(coeff0, coeff1, coeff0, coeff1, zigzag_mask0, zigzag_mask1, quant0,
+             quant2);
+  ILVL_H2_SH(quant0, quant0, quant2, quant2, quant1, quant3);
+  ILVR_H2_SH(quant0, quant0, quant2, quant2, quant0, quant2);
+  ADD2(x0, round0, x1, round1, x0, x1);
+  ILVL_H2_SH(temp0_h, x0, temp2_h, x1, temp1_h, temp3_h);
+  ILVR_H2_SH(temp0_h, x0, temp2_h, x1, temp0_h, temp2_h);
+  DOTP_SH4_SW(temp0_h, temp1_h, temp2_h, temp3_h, quant0, quant1, quant2,
+              quant3, temp0_w, temp1_w, temp2_w, temp3_w);
+  SRA_4V(temp0_w, temp1_w, temp2_w, temp3_w, 16);
+  PCKEV_H2_SH(temp1_w, temp0_w, temp3_w, temp2_w, x0, x1);
+  sign_x0 = x0 ^ sign_z0;
+  sign_x1 = x1 ^ sign_z1;
+  SUB2(sign_x0, sign_z0, sign_x1, sign_z1, sign_x0, sign_x1);
+  for (cnt = 0; cnt < 16; ++cnt) {
+    if (cnt <= 7) {
+      if (boost_temp[0] <= z_bin0[cnt]) {
+        if (x0[cnt]) {
+          eob = cnt;
+          boost_temp = zbin_boost;
+        } else {
+          boost_temp++;
         }
-        else
-        {
-            if (boost_temp[0] <= z_bin1[cnt - 8])
-            {
-                if (x1[cnt - 8])
-                {
-                    eob = cnt;
-                    boost_temp = zbin_boost;
-                }
-                else
-                {
-                    boost_temp++;
-                }
-            }
-            else
-            {
-                sign_x1[cnt - 8] = 0;
-                boost_temp++;
-            }
+      } else {
+        sign_x0[cnt] = 0;
+        boost_temp++;
+      }
+    } else {
+      if (boost_temp[0] <= z_bin1[cnt - 8]) {
+        if (x1[cnt - 8]) {
+          eob = cnt;
+          boost_temp = zbin_boost;
+        } else {
+          boost_temp++;
         }
+      } else {
+        sign_x1[cnt - 8] = 0;
+        boost_temp++;
+      }
     }
+  }
 
-    VSHF_H2_SH(sign_x0, sign_x1, sign_x0, sign_x1, inv_zig_zag0, inv_zig_zag1,
-               q_coeff0, q_coeff1);
-    ST_SH2(q_coeff0, q_coeff1, q_coeff, 8);
-    LD_SH2(de_quant, 8, de_quant0, de_quant1);
-    MUL2(de_quant0, q_coeff0, de_quant1, q_coeff1, de_quant0, de_quant1);
-    ST_SH2(de_quant0, de_quant1, dq_coeff, 8);
+  VSHF_H2_SH(sign_x0, sign_x1, sign_x0, sign_x1, inv_zig_zag0, inv_zig_zag1,
+             q_coeff0, q_coeff1);
+  ST_SH2(q_coeff0, q_coeff1, q_coeff, 8);
+  LD_SH2(de_quant, 8, de_quant0, de_quant1);
+  MUL2(de_quant0, q_coeff0, de_quant1, q_coeff1, de_quant0, de_quant1);
+  ST_SH2(de_quant0, de_quant1, dq_coeff, 8);
 
-    return (int8_t)(eob + 1);
+  return (int8_t)(eob + 1);
 }
 
-void vp8_fast_quantize_b_msa(BLOCK *b, BLOCKD *d)
-{
-    int16_t *coeff_ptr = b->coeff;
-    int16_t *zbin_ptr = b->zbin;
-    int16_t *round_ptr = b->round;
-    int16_t *quant_ptr = b->quant_fast;
-    int16_t *qcoeff_ptr = d->qcoeff;
-    int16_t *dqcoeff_ptr = d->dqcoeff;
-    int16_t *dequant_ptr = d->dequant;
+void vp8_fast_quantize_b_msa(BLOCK *b, BLOCKD *d) {
+  int16_t *coeff_ptr = b->coeff;
+  int16_t *zbin_ptr = b->zbin;
+  int16_t *round_ptr = b->round;
+  int16_t *quant_ptr = b->quant_fast;
+  int16_t *qcoeff_ptr = d->qcoeff;
+  int16_t *dqcoeff_ptr = d->dqcoeff;
+  int16_t *dequant_ptr = d->dequant;
 
-    *d->eob = fast_quantize_b_msa(coeff_ptr, zbin_ptr, round_ptr, quant_ptr,
-                                  dequant_ptr, qcoeff_ptr, dqcoeff_ptr);
+  *d->eob = fast_quantize_b_msa(coeff_ptr, zbin_ptr, round_ptr, quant_ptr,
+                                dequant_ptr, qcoeff_ptr, dqcoeff_ptr);
 }
 
-void vp8_regular_quantize_b_msa(BLOCK *b, BLOCKD *d)
-{
-    int16_t *zbin_boost_ptr = b->zrun_zbin_boost;
-    int16_t *coeff_ptr = b->coeff;
-    int16_t *zbin_ptr = b->zbin;
-    int16_t *round_ptr = b->round;
-    int16_t *quant_ptr = b->quant;
-    int16_t *quant_shift_ptr = b->quant_shift;
-    int16_t *qcoeff_ptr = d->qcoeff;
-    int16_t *dqcoeff_ptr = d->dqcoeff;
-    int16_t *dequant_ptr = d->dequant;
-    int16_t zbin_oq_value = b->zbin_extra;
+void vp8_regular_quantize_b_msa(BLOCK *b, BLOCKD *d) {
+  int16_t *zbin_boost_ptr = b->zrun_zbin_boost;
+  int16_t *coeff_ptr = b->coeff;
+  int16_t *zbin_ptr = b->zbin;
+  int16_t *round_ptr = b->round;
+  int16_t *quant_ptr = b->quant;
+  int16_t *quant_shift_ptr = b->quant_shift;
+  int16_t *qcoeff_ptr = d->qcoeff;
+  int16_t *dqcoeff_ptr = d->dqcoeff;
+  int16_t *dequant_ptr = d->dequant;
+  int16_t zbin_oq_value = b->zbin_extra;
 
-    *d->eob = exact_regular_quantize_b_msa(zbin_boost_ptr, coeff_ptr,
-                                           zbin_ptr, round_ptr,
-                                           quant_ptr, quant_shift_ptr,
-                                           dequant_ptr, zbin_oq_value,
-                                           qcoeff_ptr, dqcoeff_ptr);
+  *d->eob = exact_regular_quantize_b_msa(
+      zbin_boost_ptr, coeff_ptr, zbin_ptr, round_ptr, quant_ptr,
+      quant_shift_ptr, dequant_ptr, zbin_oq_value, qcoeff_ptr, dqcoeff_ptr);
 }
--- a/vp8/encoder/mips/msa/temporal_filter_msa.c
+++ b/vp8/encoder/mips/msa/temporal_filter_msa.c
@@ -11,293 +11,274 @@
 #include "./vp8_rtcd.h"
 #include "vp8/common/mips/msa/vp8_macros_msa.h"
 
-static void temporal_filter_apply_16size_msa(uint8_t *frame1_ptr,
-                                             uint32_t stride,
-                                             uint8_t *frame2_ptr,
-                                             int32_t strength_in,
-                                             int32_t filter_wt_in,
-                                             uint32_t *acc, uint16_t *cnt)
-{
-    uint32_t row;
-    v16i8 frame1_0_b, frame1_1_b, frame2_0_b, frame2_1_b;
-    v16u8 frame_l, frame_h;
-    v16i8 zero = { 0 };
-    v8i16 frame2_0_h, frame2_1_h, mod0_h, mod1_h;
-    v8i16 diff0, diff1, cnt0, cnt1;
-    v4i32 const3, const16, filter_wt, strength;
-    v4i32 mod0_w, mod1_w, mod2_w, mod3_w;
-    v4i32 diff0_r, diff0_l, diff1_r, diff1_l;
-    v4i32 frame2_0, frame2_1, frame2_2, frame2_3;
-    v4i32 acc0, acc1, acc2, acc3;
+static void temporal_filter_apply_16size_msa(
+    uint8_t *frame1_ptr, uint32_t stride, uint8_t *frame2_ptr,
+    int32_t strength_in, int32_t filter_wt_in, uint32_t *acc, uint16_t *cnt) {
+  uint32_t row;
+  v16i8 frame1_0_b, frame1_1_b, frame2_0_b, frame2_1_b;
+  v16u8 frame_l, frame_h;
+  v16i8 zero = { 0 };
+  v8i16 frame2_0_h, frame2_1_h, mod0_h, mod1_h;
+  v8i16 diff0, diff1, cnt0, cnt1;
+  v4i32 const3, const16, filter_wt, strength;
+  v4i32 mod0_w, mod1_w, mod2_w, mod3_w;
+  v4i32 diff0_r, diff0_l, diff1_r, diff1_l;
+  v4i32 frame2_0, frame2_1, frame2_2, frame2_3;
+  v4i32 acc0, acc1, acc2, acc3;
 
-    filter_wt = __msa_fill_w(filter_wt_in);
-    strength = __msa_fill_w(strength_in);
-    const3 = __msa_ldi_w(3);
-    const16 = __msa_ldi_w(16);
+  filter_wt = __msa_fill_w(filter_wt_in);
+  strength = __msa_fill_w(strength_in);
+  const3 = __msa_ldi_w(3);
+  const16 = __msa_ldi_w(16);
 
-    for (row = 8; row--;)
-    {
-        frame1_0_b = LD_SB(frame1_ptr);
-        frame2_0_b = LD_SB(frame2_ptr);
-        frame1_ptr += stride;
-        frame2_ptr += 16;
-        frame1_1_b = LD_SB(frame1_ptr);
-        frame2_1_b = LD_SB(frame2_ptr);
-        LD_SW2(acc, 4, acc0, acc1);
-        LD_SW2(acc + 8, 4, acc2, acc3);
-        LD_SH2(cnt, 8, cnt0, cnt1);
-        ILVRL_B2_UB(frame1_0_b, frame2_0_b, frame_l, frame_h);
-        HSUB_UB2_SH(frame_l, frame_h, diff0, diff1);
-        UNPCK_SH_SW(diff0, diff0_r, diff0_l);
-        UNPCK_SH_SW(diff1, diff1_r, diff1_l);
-        MUL4(diff0_r, diff0_r, diff0_l, diff0_l, diff1_r, diff1_r, diff1_l,
-             diff1_l, mod0_w, mod1_w, mod2_w, mod3_w);
-        MUL4(mod0_w, const3, mod1_w, const3, mod2_w, const3, mod3_w, const3,
-             mod0_w, mod1_w, mod2_w, mod3_w);
-        SRAR_W4_SW(mod0_w, mod1_w, mod2_w, mod3_w, strength);
-        diff0_r = (mod0_w < const16);
-        diff0_l = (mod1_w < const16);
-        diff1_r = (mod2_w < const16);
-        diff1_l = (mod3_w < const16);
-        SUB4(const16, mod0_w, const16, mod1_w, const16, mod2_w, const16, mod3_w,
-             mod0_w, mod1_w, mod2_w, mod3_w);
-        mod0_w = diff0_r & mod0_w;
-        mod1_w = diff0_l & mod1_w;
-        mod2_w = diff1_r & mod2_w;
-        mod3_w = diff1_l & mod3_w;
-        MUL4(mod0_w, filter_wt, mod1_w, filter_wt, mod2_w, filter_wt, mod3_w,
-             filter_wt, mod0_w, mod1_w, mod2_w, mod3_w);
-        PCKEV_H2_SH(mod1_w, mod0_w, mod3_w, mod2_w, mod0_h, mod1_h)
-        ADD2(mod0_h, cnt0, mod1_h, cnt1, mod0_h, mod1_h);
-        ST_SH2(mod0_h, mod1_h, cnt, 8);
-        cnt += 16;
-        ILVRL_B2_SH(zero, frame2_0_b, frame2_0_h, frame2_1_h);
-        UNPCK_SH_SW(frame2_0_h, frame2_0, frame2_1);
-        UNPCK_SH_SW(frame2_1_h, frame2_2, frame2_3);
-        MUL4(mod0_w, frame2_0, mod1_w, frame2_1, mod2_w, frame2_2, mod3_w,
-             frame2_3, mod0_w, mod1_w, mod2_w, mod3_w);
-        ADD4(mod0_w, acc0, mod1_w, acc1, mod2_w, acc2, mod3_w, acc3,
-             mod0_w, mod1_w, mod2_w, mod3_w);
-        ST_SW2(mod0_w, mod1_w, acc, 4);
-        ST_SW2(mod2_w, mod3_w, acc + 8, 4);
-        acc += 16;
-        LD_SW2(acc, 4, acc0, acc1);
-        LD_SW2(acc + 8, 4, acc2, acc3);
-        LD_SH2(cnt, 8, cnt0, cnt1);
-        ILVRL_B2_UB(frame1_1_b, frame2_1_b, frame_l, frame_h);
-        HSUB_UB2_SH(frame_l, frame_h, diff0, diff1);
-        UNPCK_SH_SW(diff0, diff0_r, diff0_l);
-        UNPCK_SH_SW(diff1, diff1_r, diff1_l);
-        MUL4(diff0_r, diff0_r, diff0_l, diff0_l, diff1_r, diff1_r, diff1_l,
-             diff1_l, mod0_w, mod1_w, mod2_w, mod3_w);
-        MUL4(mod0_w, const3, mod1_w, const3, mod2_w, const3, mod3_w, const3,
-             mod0_w, mod1_w, mod2_w, mod3_w);
-        SRAR_W4_SW(mod0_w, mod1_w, mod2_w, mod3_w, strength);
-        diff0_r = (mod0_w < const16);
-        diff0_l = (mod1_w < const16);
-        diff1_r = (mod2_w < const16);
-        diff1_l = (mod3_w < const16);
-        SUB4(const16, mod0_w, const16, mod1_w, const16, mod2_w, const16, mod3_w,
-             mod0_w, mod1_w, mod2_w, mod3_w);
-        mod0_w = diff0_r & mod0_w;
-        mod1_w = diff0_l & mod1_w;
-        mod2_w = diff1_r & mod2_w;
-        mod3_w = diff1_l & mod3_w;
-        MUL4(mod0_w, filter_wt, mod1_w, filter_wt, mod2_w, filter_wt, mod3_w,
-             filter_wt, mod0_w, mod1_w, mod2_w, mod3_w);
-        PCKEV_H2_SH(mod1_w, mod0_w, mod3_w, mod2_w, mod0_h, mod1_h);
-        ADD2(mod0_h, cnt0, mod1_h, cnt1, mod0_h, mod1_h);
-        ST_SH2(mod0_h, mod1_h, cnt, 8);
-        cnt += 16;
+  for (row = 8; row--;) {
+    frame1_0_b = LD_SB(frame1_ptr);
+    frame2_0_b = LD_SB(frame2_ptr);
+    frame1_ptr += stride;
+    frame2_ptr += 16;
+    frame1_1_b = LD_SB(frame1_ptr);
+    frame2_1_b = LD_SB(frame2_ptr);
+    LD_SW2(acc, 4, acc0, acc1);
+    LD_SW2(acc + 8, 4, acc2, acc3);
+    LD_SH2(cnt, 8, cnt0, cnt1);
+    ILVRL_B2_UB(frame1_0_b, frame2_0_b, frame_l, frame_h);
+    HSUB_UB2_SH(frame_l, frame_h, diff0, diff1);
+    UNPCK_SH_SW(diff0, diff0_r, diff0_l);
+    UNPCK_SH_SW(diff1, diff1_r, diff1_l);
+    MUL4(diff0_r, diff0_r, diff0_l, diff0_l, diff1_r, diff1_r, diff1_l, diff1_l,
+         mod0_w, mod1_w, mod2_w, mod3_w);
+    MUL4(mod0_w, const3, mod1_w, const3, mod2_w, const3, mod3_w, const3, mod0_w,
+         mod1_w, mod2_w, mod3_w);
+    SRAR_W4_SW(mod0_w, mod1_w, mod2_w, mod3_w, strength);
+    diff0_r = (mod0_w < const16);
+    diff0_l = (mod1_w < const16);
+    diff1_r = (mod2_w < const16);
+    diff1_l = (mod3_w < const16);
+    SUB4(const16, mod0_w, const16, mod1_w, const16, mod2_w, const16, mod3_w,
+         mod0_w, mod1_w, mod2_w, mod3_w);
+    mod0_w = diff0_r & mod0_w;
+    mod1_w = diff0_l & mod1_w;
+    mod2_w = diff1_r & mod2_w;
+    mod3_w = diff1_l & mod3_w;
+    MUL4(mod0_w, filter_wt, mod1_w, filter_wt, mod2_w, filter_wt, mod3_w,
+         filter_wt, mod0_w, mod1_w, mod2_w, mod3_w);
+    PCKEV_H2_SH(mod1_w, mod0_w, mod3_w, mod2_w, mod0_h, mod1_h)
+    ADD2(mod0_h, cnt0, mod1_h, cnt1, mod0_h, mod1_h);
+    ST_SH2(mod0_h, mod1_h, cnt, 8);
+    cnt += 16;
+    ILVRL_B2_SH(zero, frame2_0_b, frame2_0_h, frame2_1_h);
+    UNPCK_SH_SW(frame2_0_h, frame2_0, frame2_1);
+    UNPCK_SH_SW(frame2_1_h, frame2_2, frame2_3);
+    MUL4(mod0_w, frame2_0, mod1_w, frame2_1, mod2_w, frame2_2, mod3_w, frame2_3,
+         mod0_w, mod1_w, mod2_w, mod3_w);
+    ADD4(mod0_w, acc0, mod1_w, acc1, mod2_w, acc2, mod3_w, acc3, mod0_w, mod1_w,
+         mod2_w, mod3_w);
+    ST_SW2(mod0_w, mod1_w, acc, 4);
+    ST_SW2(mod2_w, mod3_w, acc + 8, 4);
+    acc += 16;
+    LD_SW2(acc, 4, acc0, acc1);
+    LD_SW2(acc + 8, 4, acc2, acc3);
+    LD_SH2(cnt, 8, cnt0, cnt1);
+    ILVRL_B2_UB(frame1_1_b, frame2_1_b, frame_l, frame_h);
+    HSUB_UB2_SH(frame_l, frame_h, diff0, diff1);
+    UNPCK_SH_SW(diff0, diff0_r, diff0_l);
+    UNPCK_SH_SW(diff1, diff1_r, diff1_l);
+    MUL4(diff0_r, diff0_r, diff0_l, diff0_l, diff1_r, diff1_r, diff1_l, diff1_l,
+         mod0_w, mod1_w, mod2_w, mod3_w);
+    MUL4(mod0_w, const3, mod1_w, const3, mod2_w, const3, mod3_w, const3, mod0_w,
+         mod1_w, mod2_w, mod3_w);
+    SRAR_W4_SW(mod0_w, mod1_w, mod2_w, mod3_w, strength);
+    diff0_r = (mod0_w < const16);
+    diff0_l = (mod1_w < const16);
+    diff1_r = (mod2_w < const16);
+    diff1_l = (mod3_w < const16);
+    SUB4(const16, mod0_w, const16, mod1_w, const16, mod2_w, const16, mod3_w,
+         mod0_w, mod1_w, mod2_w, mod3_w);
+    mod0_w = diff0_r & mod0_w;
+    mod1_w = diff0_l & mod1_w;
+    mod2_w = diff1_r & mod2_w;
+    mod3_w = diff1_l & mod3_w;
+    MUL4(mod0_w, filter_wt, mod1_w, filter_wt, mod2_w, filter_wt, mod3_w,
+         filter_wt, mod0_w, mod1_w, mod2_w, mod3_w);
+    PCKEV_H2_SH(mod1_w, mod0_w, mod3_w, mod2_w, mod0_h, mod1_h);
+    ADD2(mod0_h, cnt0, mod1_h, cnt1, mod0_h, mod1_h);
+    ST_SH2(mod0_h, mod1_h, cnt, 8);
+    cnt += 16;
 
-        UNPCK_UB_SH(frame2_1_b, frame2_0_h, frame2_1_h);
-        UNPCK_SH_SW(frame2_0_h, frame2_0, frame2_1);
-        UNPCK_SH_SW(frame2_1_h, frame2_2, frame2_3);
-        MUL4(mod0_w, frame2_0, mod1_w, frame2_1, mod2_w, frame2_2, mod3_w,
-             frame2_3, mod0_w, mod1_w, mod2_w, mod3_w);
-        ADD4(mod0_w, acc0, mod1_w, acc1, mod2_w, acc2, mod3_w, acc3,
-             mod0_w, mod1_w, mod2_w, mod3_w);
-        ST_SW2(mod0_w, mod1_w, acc, 4);
-        ST_SW2(mod2_w, mod3_w, acc + 8, 4);
-        acc += 16;
-        frame1_ptr += stride;
-        frame2_ptr += 16;
-    }
+    UNPCK_UB_SH(frame2_1_b, frame2_0_h, frame2_1_h);
+    UNPCK_SH_SW(frame2_0_h, frame2_0, frame2_1);
+    UNPCK_SH_SW(frame2_1_h, frame2_2, frame2_3);
+    MUL4(mod0_w, frame2_0, mod1_w, frame2_1, mod2_w, frame2_2, mod3_w, frame2_3,
+         mod0_w, mod1_w, mod2_w, mod3_w);
+    ADD4(mod0_w, acc0, mod1_w, acc1, mod2_w, acc2, mod3_w, acc3, mod0_w, mod1_w,
+         mod2_w, mod3_w);
+    ST_SW2(mod0_w, mod1_w, acc, 4);
+    ST_SW2(mod2_w, mod3_w, acc + 8, 4);
+    acc += 16;
+    frame1_ptr += stride;
+    frame2_ptr += 16;
+  }
 }
 
-static void temporal_filter_apply_8size_msa(uint8_t *frame1_ptr,
-                                            uint32_t stride,
-                                            uint8_t *frame2_ptr,
-                                            int32_t strength_in,
-                                            int32_t filter_wt_in,
-                                            uint32_t *acc, uint16_t *cnt)
-{
-    uint32_t row;
-    uint64_t f0, f1, f2, f3, f4, f5, f6, f7;
-    v16i8 frame1 = { 0 };
-    v16i8 frame2 = { 0 };
-    v16i8 frame3 = { 0 };
-    v16i8 frame4 = { 0 };
-    v16u8 frame_l, frame_h;
-    v8i16 frame2_0_h, frame2_1_h, mod0_h, mod1_h;
-    v8i16 diff0, diff1, cnt0, cnt1;
-    v4i32 const3, const16;
-    v4i32 filter_wt, strength;
-    v4i32 mod0_w, mod1_w, mod2_w, mod3_w;
-    v4i32 diff0_r, diff0_l, diff1_r, diff1_l;
-    v4i32 frame2_0, frame2_1, frame2_2, frame2_3;
-    v4i32 acc0, acc1, acc2, acc3;
+static void temporal_filter_apply_8size_msa(
+    uint8_t *frame1_ptr, uint32_t stride, uint8_t *frame2_ptr,
+    int32_t strength_in, int32_t filter_wt_in, uint32_t *acc, uint16_t *cnt) {
+  uint32_t row;
+  uint64_t f0, f1, f2, f3, f4, f5, f6, f7;
+  v16i8 frame1 = { 0 };
+  v16i8 frame2 = { 0 };
+  v16i8 frame3 = { 0 };
+  v16i8 frame4 = { 0 };
+  v16u8 frame_l, frame_h;
+  v8i16 frame2_0_h, frame2_1_h, mod0_h, mod1_h;
+  v8i16 diff0, diff1, cnt0, cnt1;
+  v4i32 const3, const16;
+  v4i32 filter_wt, strength;
+  v4i32 mod0_w, mod1_w, mod2_w, mod3_w;
+  v4i32 diff0_r, diff0_l, diff1_r, diff1_l;
+  v4i32 frame2_0, frame2_1, frame2_2, frame2_3;
+  v4i32 acc0, acc1, acc2, acc3;
 
-    filter_wt = __msa_fill_w(filter_wt_in);
-    strength = __msa_fill_w(strength_in);
-    const3 = __msa_ldi_w(3);
-    const16 = __msa_ldi_w(16);
+  filter_wt = __msa_fill_w(filter_wt_in);
+  strength = __msa_fill_w(strength_in);
+  const3 = __msa_ldi_w(3);
+  const16 = __msa_ldi_w(16);
 
-    for (row = 2; row--;)
-    {
-        LD2(frame1_ptr, stride, f0, f1);
-        frame1_ptr += (2 * stride);
-        LD2(frame2_ptr, 8, f2, f3);
-        frame2_ptr += 16;
-        LD2(frame1_ptr, stride, f4, f5);
-        frame1_ptr += (2 * stride);
-        LD2(frame2_ptr, 8, f6, f7);
-        frame2_ptr += 16;
+  for (row = 2; row--;) {
+    LD2(frame1_ptr, stride, f0, f1);
+    frame1_ptr += (2 * stride);
+    LD2(frame2_ptr, 8, f2, f3);
+    frame2_ptr += 16;
+    LD2(frame1_ptr, stride, f4, f5);
+    frame1_ptr += (2 * stride);
+    LD2(frame2_ptr, 8, f6, f7);
+    frame2_ptr += 16;
 
-        LD_SW2(acc, 4, acc0, acc1);
-        LD_SW2(acc + 8, 4, acc2, acc3);
-        LD_SH2(cnt, 8, cnt0, cnt1);
-        INSERT_D2_SB(f0, f1, frame1);
-        INSERT_D2_SB(f2, f3, frame2);
-        INSERT_D2_SB(f4, f5, frame3);
-        INSERT_D2_SB(f6, f7, frame4);
-        ILVRL_B2_UB(frame1, frame2, frame_l, frame_h);
-        HSUB_UB2_SH(frame_l, frame_h, diff0, diff1);
-        UNPCK_SH_SW(diff0, diff0_r, diff0_l);
-        UNPCK_SH_SW(diff1, diff1_r, diff1_l);
-        MUL4(diff0_r, diff0_r, diff0_l, diff0_l, diff1_r, diff1_r, diff1_l,
-             diff1_l, mod0_w, mod1_w, mod2_w, mod3_w);
-        MUL4(mod0_w, const3, mod1_w, const3, mod2_w, const3, mod3_w, const3,
-             mod0_w, mod1_w, mod2_w, mod3_w);
-        SRAR_W4_SW(mod0_w, mod1_w, mod2_w, mod3_w, strength);
-        diff0_r = (mod0_w < const16);
-        diff0_l = (mod1_w < const16);
-        diff1_r = (mod2_w < const16);
-        diff1_l = (mod3_w < const16);
-        SUB4(const16, mod0_w, const16, mod1_w, const16, mod2_w, const16, mod3_w,
-             mod0_w, mod1_w, mod2_w, mod3_w);
-        mod0_w = diff0_r & mod0_w;
-        mod1_w = diff0_l & mod1_w;
-        mod2_w = diff1_r & mod2_w;
-        mod3_w = diff1_l & mod3_w;
-        MUL4(mod0_w, filter_wt, mod1_w, filter_wt, mod2_w, filter_wt, mod3_w,
-             filter_wt, mod0_w, mod1_w, mod2_w, mod3_w);
-        PCKEV_H2_SH(mod1_w, mod0_w, mod3_w, mod2_w, mod0_h, mod1_h);
-        ADD2(mod0_h, cnt0, mod1_h, cnt1, mod0_h, mod1_h);
-        ST_SH2(mod0_h, mod1_h, cnt, 8);
-        cnt += 16;
+    LD_SW2(acc, 4, acc0, acc1);
+    LD_SW2(acc + 8, 4, acc2, acc3);
+    LD_SH2(cnt, 8, cnt0, cnt1);
+    INSERT_D2_SB(f0, f1, frame1);
+    INSERT_D2_SB(f2, f3, frame2);
+    INSERT_D2_SB(f4, f5, frame3);
+    INSERT_D2_SB(f6, f7, frame4);
+    ILVRL_B2_UB(frame1, frame2, frame_l, frame_h);
+    HSUB_UB2_SH(frame_l, frame_h, diff0, diff1);
+    UNPCK_SH_SW(diff0, diff0_r, diff0_l);
+    UNPCK_SH_SW(diff1, diff1_r, diff1_l);
+    MUL4(diff0_r, diff0_r, diff0_l, diff0_l, diff1_r, diff1_r, diff1_l, diff1_l,
+         mod0_w, mod1_w, mod2_w, mod3_w);
+    MUL4(mod0_w, const3, mod1_w, const3, mod2_w, const3, mod3_w, const3, mod0_w,
+         mod1_w, mod2_w, mod3_w);
+    SRAR_W4_SW(mod0_w, mod1_w, mod2_w, mod3_w, strength);
+    diff0_r = (mod0_w < const16);
+    diff0_l = (mod1_w < const16);
+    diff1_r = (mod2_w < const16);
+    diff1_l = (mod3_w < const16);
+    SUB4(const16, mod0_w, const16, mod1_w, const16, mod2_w, const16, mod3_w,
+         mod0_w, mod1_w, mod2_w, mod3_w);
+    mod0_w = diff0_r & mod0_w;
+    mod1_w = diff0_l & mod1_w;
+    mod2_w = diff1_r & mod2_w;
+    mod3_w = diff1_l & mod3_w;
+    MUL4(mod0_w, filter_wt, mod1_w, filter_wt, mod2_w, filter_wt, mod3_w,
+         filter_wt, mod0_w, mod1_w, mod2_w, mod3_w);
+    PCKEV_H2_SH(mod1_w, mod0_w, mod3_w, mod2_w, mod0_h, mod1_h);
+    ADD2(mod0_h, cnt0, mod1_h, cnt1, mod0_h, mod1_h);
+    ST_SH2(mod0_h, mod1_h, cnt, 8);
+    cnt += 16;
 
-        UNPCK_UB_SH(frame2, frame2_0_h, frame2_1_h);
-        UNPCK_SH_SW(frame2_0_h, frame2_0, frame2_1);
-        UNPCK_SH_SW(frame2_1_h, frame2_2, frame2_3);
-        MUL4(mod0_w, frame2_0, mod1_w, frame2_1, mod2_w, frame2_2, mod3_w,
-             frame2_3, mod0_w, mod1_w, mod2_w, mod3_w);
-        ADD4(mod0_w, acc0, mod1_w, acc1, mod2_w, acc2, mod3_w, acc3,
-             mod0_w, mod1_w, mod2_w, mod3_w);
-        ST_SW2(mod0_w, mod1_w, acc, 4);
-        ST_SW2(mod2_w, mod3_w, acc + 8, 4);
-        acc += 16;
+    UNPCK_UB_SH(frame2, frame2_0_h, frame2_1_h);
+    UNPCK_SH_SW(frame2_0_h, frame2_0, frame2_1);
+    UNPCK_SH_SW(frame2_1_h, frame2_2, frame2_3);
+    MUL4(mod0_w, frame2_0, mod1_w, frame2_1, mod2_w, frame2_2, mod3_w, frame2_3,
+         mod0_w, mod1_w, mod2_w, mod3_w);
+    ADD4(mod0_w, acc0, mod1_w, acc1, mod2_w, acc2, mod3_w, acc3, mod0_w, mod1_w,
+         mod2_w, mod3_w);
+    ST_SW2(mod0_w, mod1_w, acc, 4);
+    ST_SW2(mod2_w, mod3_w, acc + 8, 4);
+    acc += 16;
 
-        LD_SW2(acc, 4, acc0, acc1);
-        LD_SW2(acc + 8, 4, acc2, acc3);
-        LD_SH2(cnt, 8, cnt0, cnt1);
-        ILVRL_B2_UB(frame3, frame4, frame_l, frame_h);
-        HSUB_UB2_SH(frame_l, frame_h, diff0, diff1);
-        UNPCK_SH_SW(diff0, diff0_r, diff0_l);
-        UNPCK_SH_SW(diff1, diff1_r, diff1_l);
-        MUL4(diff0_r, diff0_r, diff0_l, diff0_l, diff1_r, diff1_r, diff1_l,
-             diff1_l, mod0_w, mod1_w, mod2_w, mod3_w);
-        MUL4(mod0_w, const3, mod1_w, const3, mod2_w, const3, mod3_w, const3,
-             mod0_w, mod1_w, mod2_w, mod3_w);
-        SRAR_W4_SW(mod0_w, mod1_w, mod2_w, mod3_w, strength);
-        diff0_r = (mod0_w < const16);
-        diff0_l = (mod1_w < const16);
-        diff1_r = (mod2_w < const16);
-        diff1_l = (mod3_w < const16);
-        SUB4(const16, mod0_w, const16, mod1_w, const16, mod2_w, const16, mod3_w,
-             mod0_w, mod1_w, mod2_w, mod3_w);
-        mod0_w = diff0_r & mod0_w;
-        mod1_w = diff0_l & mod1_w;
-        mod2_w = diff1_r & mod2_w;
-        mod3_w = diff1_l & mod3_w;
-        MUL4(mod0_w, filter_wt, mod1_w, filter_wt, mod2_w, filter_wt, mod3_w,
-             filter_wt, mod0_w, mod1_w, mod2_w, mod3_w);
-        PCKEV_H2_SH(mod1_w, mod0_w, mod3_w, mod2_w, mod0_h, mod1_h);
-        ADD2(mod0_h, cnt0, mod1_h, cnt1, mod0_h, mod1_h);
-        ST_SH2(mod0_h, mod1_h, cnt, 8);
-        cnt += 16;
+    LD_SW2(acc, 4, acc0, acc1);
+    LD_SW2(acc + 8, 4, acc2, acc3);
+    LD_SH2(cnt, 8, cnt0, cnt1);
+    ILVRL_B2_UB(frame3, frame4, frame_l, frame_h);
+    HSUB_UB2_SH(frame_l, frame_h, diff0, diff1);
+    UNPCK_SH_SW(diff0, diff0_r, diff0_l);
+    UNPCK_SH_SW(diff1, diff1_r, diff1_l);
+    MUL4(diff0_r, diff0_r, diff0_l, diff0_l, diff1_r, diff1_r, diff1_l, diff1_l,
+         mod0_w, mod1_w, mod2_w, mod3_w);
+    MUL4(mod0_w, const3, mod1_w, const3, mod2_w, const3, mod3_w, const3, mod0_w,
+         mod1_w, mod2_w, mod3_w);
+    SRAR_W4_SW(mod0_w, mod1_w, mod2_w, mod3_w, strength);
+    diff0_r = (mod0_w < const16);
+    diff0_l = (mod1_w < const16);
+    diff1_r = (mod2_w < const16);
+    diff1_l = (mod3_w < const16);
+    SUB4(const16, mod0_w, const16, mod1_w, const16, mod2_w, const16, mod3_w,
+         mod0_w, mod1_w, mod2_w, mod3_w);
+    mod0_w = diff0_r & mod0_w;
+    mod1_w = diff0_l & mod1_w;
+    mod2_w = diff1_r & mod2_w;
+    mod3_w = diff1_l & mod3_w;
+    MUL4(mod0_w, filter_wt, mod1_w, filter_wt, mod2_w, filter_wt, mod3_w,
+         filter_wt, mod0_w, mod1_w, mod2_w, mod3_w);
+    PCKEV_H2_SH(mod1_w, mod0_w, mod3_w, mod2_w, mod0_h, mod1_h);
+    ADD2(mod0_h, cnt0, mod1_h, cnt1, mod0_h, mod1_h);
+    ST_SH2(mod0_h, mod1_h, cnt, 8);
+    cnt += 16;
 
-        UNPCK_UB_SH(frame4, frame2_0_h, frame2_1_h);
-        UNPCK_SH_SW(frame2_0_h, frame2_0, frame2_1);
-        UNPCK_SH_SW(frame2_1_h, frame2_2, frame2_3);
-        MUL4(mod0_w, frame2_0, mod1_w, frame2_1, mod2_w, frame2_2, mod3_w,
-             frame2_3, mod0_w, mod1_w, mod2_w, mod3_w);
-        ADD4(mod0_w, acc0, mod1_w, acc1, mod2_w, acc2, mod3_w, acc3,
-             mod0_w, mod1_w, mod2_w, mod3_w);
-        ST_SW2(mod0_w, mod1_w, acc, 4);
-        ST_SW2(mod2_w, mod3_w, acc + 8, 4);
-        acc += 16;
-    }
+    UNPCK_UB_SH(frame4, frame2_0_h, frame2_1_h);
+    UNPCK_SH_SW(frame2_0_h, frame2_0, frame2_1);
+    UNPCK_SH_SW(frame2_1_h, frame2_2, frame2_3);
+    MUL4(mod0_w, frame2_0, mod1_w, frame2_1, mod2_w, frame2_2, mod3_w, frame2_3,
+         mod0_w, mod1_w, mod2_w, mod3_w);
+    ADD4(mod0_w, acc0, mod1_w, acc1, mod2_w, acc2, mod3_w, acc3, mod0_w, mod1_w,
+         mod2_w, mod3_w);
+    ST_SW2(mod0_w, mod1_w, acc, 4);
+    ST_SW2(mod2_w, mod3_w, acc + 8, 4);
+    acc += 16;
+  }
 }
 
 void vp8_temporal_filter_apply_msa(uint8_t *frame1, uint32_t stride,
                                    uint8_t *frame2, uint32_t block_size,
-                                   int32_t strength,  int32_t filter_weight,
-                                   uint32_t *accumulator, uint16_t *count)
-{
-    if (8 == block_size)
-    {
-        temporal_filter_apply_8size_msa(frame1, stride, frame2, strength,
-                                        filter_weight, accumulator, count);
-    }
-    else if (16 == block_size)
-    {
-        temporal_filter_apply_16size_msa(frame1, stride, frame2, strength,
-                                         filter_weight, accumulator, count);
-    }
-    else
-    {
-        uint32_t i, j, k;
-        int32_t modifier;
-        int32_t byte = 0;
-        const int32_t rounding = strength > 0 ? 1 << (strength - 1) : 0;
+                                   int32_t strength, int32_t filter_weight,
+                                   uint32_t *accumulator, uint16_t *count) {
+  if (8 == block_size) {
+    temporal_filter_apply_8size_msa(frame1, stride, frame2, strength,
+                                    filter_weight, accumulator, count);
+  } else if (16 == block_size) {
+    temporal_filter_apply_16size_msa(frame1, stride, frame2, strength,
+                                     filter_weight, accumulator, count);
+  } else {
+    uint32_t i, j, k;
+    int32_t modifier;
+    int32_t byte = 0;
+    const int32_t rounding = strength > 0 ? 1 << (strength - 1) : 0;
 
-        for (i = 0, k = 0; i < block_size; ++i)
-        {
-            for (j = 0; j < block_size; ++j, ++k)
-            {
-                int src_byte = frame1[byte];
-                int pixel_value = *frame2++;
+    for (i = 0, k = 0; i < block_size; ++i) {
+      for (j = 0; j < block_size; ++j, ++k) {
+        int src_byte = frame1[byte];
+        int pixel_value = *frame2++;
 
-                modifier = src_byte - pixel_value;
-                modifier *= modifier;
-                modifier *= 3;
-                modifier += rounding;
-                modifier >>= strength;
+        modifier = src_byte - pixel_value;
+        modifier *= modifier;
+        modifier *= 3;
+        modifier += rounding;
+        modifier >>= strength;
 
-                if (modifier > 16)
-                    modifier = 16;
+        if (modifier > 16) modifier = 16;
 
-                modifier = 16 - modifier;
-                modifier *= filter_weight;
+        modifier = 16 - modifier;
+        modifier *= filter_weight;
 
-                count[k] += modifier;
-                accumulator[k] += modifier * pixel_value;
+        count[k] += modifier;
+        accumulator[k] += modifier * pixel_value;
 
-                byte++;
-            }
+        byte++;
+      }
 
-            byte += stride - block_size;
-        }
+      byte += stride - block_size;
     }
+  }
 }
--- a/vp8/encoder/modecosts.c
+++ b/vp8/encoder/modecosts.c
@@ -8,7 +8,6 @@
  *  be found in the AUTHORS file in the root of the source tree.
  */
 
-
 #include "vp8/common/blockd.h"
 #include "modecosts.h"
 #include "onyx_int.h"
@@ -15,41 +14,35 @@
 #include "treewriter.h"
 #include "vp8/common/entropymode.h"
 
+void vp8_init_mode_costs(VP8_COMP *c) {
+  VP8_COMMON *x = &c->common;
+  struct rd_costs_struct *rd_costs = &c->rd_costs;
 
-void vp8_init_mode_costs(VP8_COMP *c)
-{
-    VP8_COMMON *x = &c->common;
-    struct rd_costs_struct *rd_costs = &c->rd_costs;
+  {
+    const vp8_tree_p T = vp8_bmode_tree;
 
-    {
-        const vp8_tree_p T = vp8_bmode_tree;
+    int i = 0;
 
-        int i = 0;
+    do {
+      int j = 0;
 
-        do
-        {
-            int j = 0;
+      do {
+        vp8_cost_tokens(rd_costs->bmode_costs[i][j], vp8_kf_bmode_prob[i][j],
+                        T);
+      } while (++j < VP8_BINTRAMODES);
+    } while (++i < VP8_BINTRAMODES);
 
-            do
-            {
-                vp8_cost_tokens(rd_costs->bmode_costs[i][j],
-                                vp8_kf_bmode_prob[i][j], T);
-            }
-            while (++j < VP8_BINTRAMODES);
-        }
-        while (++i < VP8_BINTRAMODES);
+    vp8_cost_tokens(rd_costs->inter_bmode_costs, x->fc.bmode_prob, T);
+  }
+  vp8_cost_tokens(rd_costs->inter_bmode_costs, x->fc.sub_mv_ref_prob,
+                  vp8_sub_mv_ref_tree);
 
-        vp8_cost_tokens(rd_costs->inter_bmode_costs, x->fc.bmode_prob, T);
-    }
-    vp8_cost_tokens(rd_costs->inter_bmode_costs, x->fc.sub_mv_ref_prob,
-                    vp8_sub_mv_ref_tree);
+  vp8_cost_tokens(rd_costs->mbmode_cost[1], x->fc.ymode_prob, vp8_ymode_tree);
+  vp8_cost_tokens(rd_costs->mbmode_cost[0], vp8_kf_ymode_prob,
+                  vp8_kf_ymode_tree);
 
-    vp8_cost_tokens(rd_costs->mbmode_cost[1], x->fc.ymode_prob, vp8_ymode_tree);
-    vp8_cost_tokens(rd_costs->mbmode_cost[0], vp8_kf_ymode_prob,
-                    vp8_kf_ymode_tree);
-
-    vp8_cost_tokens(rd_costs->intra_uv_mode_cost[1], x->fc.uv_mode_prob,
-                    vp8_uv_mode_tree);
-    vp8_cost_tokens(rd_costs->intra_uv_mode_cost[0], vp8_kf_uv_mode_prob,
-                    vp8_uv_mode_tree);
+  vp8_cost_tokens(rd_costs->intra_uv_mode_cost[1], x->fc.uv_mode_prob,
+                  vp8_uv_mode_tree);
+  vp8_cost_tokens(rd_costs->intra_uv_mode_cost[0], vp8_kf_uv_mode_prob,
+                  vp8_uv_mode_tree);
 }
--- a/vp8/encoder/modecosts.h
+++ b/vp8/encoder/modecosts.h
@@ -8,7 +8,6 @@
  *  be found in the AUTHORS file in the root of the source tree.
  */
 
-
 #ifndef VP8_ENCODER_MODECOSTS_H_
 #define VP8_ENCODER_MODECOSTS_H_
 
--- a/vp8/encoder/mr_dissim.c
+++ b/vp8/encoder/mr_dissim.c
@@ -8,7 +8,6 @@
  *  be found in the AUTHORS file in the root of the source tree.
  */
 
-
 #include <limits.h>
 #include "vpx_config.h"
 #include "onyx_int.h"
@@ -18,223 +17,199 @@
 #include "rdopt.h"
 #include "vp8/common/common.h"
 
-void vp8_cal_low_res_mb_cols(VP8_COMP *cpi)
-{
-    int low_res_w;
+void vp8_cal_low_res_mb_cols(VP8_COMP *cpi) {
+  int low_res_w;
 
-    /* Support arbitrary down-sampling factor */
-    unsigned int iw = cpi->oxcf.Width*cpi->oxcf.mr_down_sampling_factor.den
-                      + cpi->oxcf.mr_down_sampling_factor.num - 1;
+  /* Support arbitrary down-sampling factor */
+  unsigned int iw = cpi->oxcf.Width * cpi->oxcf.mr_down_sampling_factor.den +
+                    cpi->oxcf.mr_down_sampling_factor.num - 1;
 
-    low_res_w = iw/cpi->oxcf.mr_down_sampling_factor.num;
-    cpi->mr_low_res_mb_cols = ((low_res_w + 15) >> 4);
+  low_res_w = iw / cpi->oxcf.mr_down_sampling_factor.num;
+  cpi->mr_low_res_mb_cols = ((low_res_w + 15) >> 4);
 }
 
-#define GET_MV(x)    \
-if(x->mbmi.ref_frame !=INTRA_FRAME)   \
-{   \
-    mvx[cnt] = x->mbmi.mv.as_mv.row;  \
-    mvy[cnt] = x->mbmi.mv.as_mv.col;  \
-    cnt++;    \
-}
+#define GET_MV(x)                         \
+  if (x->mbmi.ref_frame != INTRA_FRAME) { \
+    mvx[cnt] = x->mbmi.mv.as_mv.row;      \
+    mvy[cnt] = x->mbmi.mv.as_mv.col;      \
+    cnt++;                                \
+  }
 
-#define GET_MV_SIGN(x)    \
-if(x->mbmi.ref_frame !=INTRA_FRAME)   \
-{   \
-    mvx[cnt] = x->mbmi.mv.as_mv.row;  \
-    mvy[cnt] = x->mbmi.mv.as_mv.col;  \
-    if (cm->ref_frame_sign_bias[x->mbmi.ref_frame]  \
-        != cm->ref_frame_sign_bias[tmp->mbmi.ref_frame])  \
-    {  \
-        mvx[cnt] *= -1;   \
-        mvy[cnt] *= -1;   \
-    }  \
-    cnt++;  \
-}
+#define GET_MV_SIGN(x)                                  \
+  if (x->mbmi.ref_frame != INTRA_FRAME) {               \
+    mvx[cnt] = x->mbmi.mv.as_mv.row;                    \
+    mvy[cnt] = x->mbmi.mv.as_mv.col;                    \
+    if (cm->ref_frame_sign_bias[x->mbmi.ref_frame] !=   \
+        cm->ref_frame_sign_bias[tmp->mbmi.ref_frame]) { \
+      mvx[cnt] *= -1;                                   \
+      mvy[cnt] *= -1;                                   \
+    }                                                   \
+    cnt++;                                              \
+  }
 
-void vp8_cal_dissimilarity(VP8_COMP *cpi)
-{
-    VP8_COMMON *cm = &cpi->common;
-    int i;
+void vp8_cal_dissimilarity(VP8_COMP *cpi) {
+  VP8_COMMON *cm = &cpi->common;
+  int i;
 
-    /* Note: The first row & first column in mip are outside the frame, which
-     * were initialized to all 0.(ref_frame, mode, mv...)
-     * Their ref_frame = 0 means they won't be counted in the following
-     * calculation.
+  /* Note: The first row & first column in mip are outside the frame, which
+   * were initialized to all 0.(ref_frame, mode, mv...)
+   * Their ref_frame = 0 means they won't be counted in the following
+   * calculation.
+   */
+  if (cpi->oxcf.mr_total_resolutions > 1 &&
+      cpi->oxcf.mr_encoder_id < (cpi->oxcf.mr_total_resolutions - 1)) {
+    /* Store info for show/no-show frames for supporting alt_ref.
+     * If parent frame is alt_ref, child has one too.
      */
-    if (cpi->oxcf.mr_total_resolutions >1
-        && cpi->oxcf.mr_encoder_id < (cpi->oxcf.mr_total_resolutions - 1))
-    {
-        /* Store info for show/no-show frames for supporting alt_ref.
-         * If parent frame is alt_ref, child has one too.
-         */
-        LOWER_RES_FRAME_INFO* store_info
-                      = (LOWER_RES_FRAME_INFO*)cpi->oxcf.mr_low_res_mode_info;
+    LOWER_RES_FRAME_INFO *store_info =
+        (LOWER_RES_FRAME_INFO *)cpi->oxcf.mr_low_res_mode_info;
 
-        store_info->frame_type = cm->frame_type;
+    store_info->frame_type = cm->frame_type;
 
-        if(cm->frame_type != KEY_FRAME)
-        {
-            store_info->is_frame_dropped = 0;
-            for (i = 1; i < MAX_REF_FRAMES; i++)
-                store_info->low_res_ref_frames[i] = cpi->current_ref_frames[i];
-        }
+    if (cm->frame_type != KEY_FRAME) {
+      store_info->is_frame_dropped = 0;
+      for (i = 1; i < MAX_REF_FRAMES; i++)
+        store_info->low_res_ref_frames[i] = cpi->current_ref_frames[i];
+    }
 
-        if(cm->frame_type != KEY_FRAME)
-        {
-            int mb_row;
-            int mb_col;
-            /* Point to beginning of allocated MODE_INFO arrays. */
-            MODE_INFO *tmp = cm->mip + cm->mode_info_stride;
-            LOWER_RES_MB_INFO* store_mode_info = store_info->mb_info;
+    if (cm->frame_type != KEY_FRAME) {
+      int mb_row;
+      int mb_col;
+      /* Point to beginning of allocated MODE_INFO arrays. */
+      MODE_INFO *tmp = cm->mip + cm->mode_info_stride;
+      LOWER_RES_MB_INFO *store_mode_info = store_info->mb_info;
 
-            for (mb_row = 0; mb_row < cm->mb_rows; mb_row ++)
-            {
-                tmp++;
-                for (mb_col = 0; mb_col < cm->mb_cols; mb_col ++)
-                {
-                    int dissim = INT_MAX;
+      for (mb_row = 0; mb_row < cm->mb_rows; mb_row++) {
+        tmp++;
+        for (mb_col = 0; mb_col < cm->mb_cols; mb_col++) {
+          int dissim = INT_MAX;
 
-                    if(tmp->mbmi.ref_frame !=INTRA_FRAME)
-                    {
-                        int              mvx[8];
-                        int              mvy[8];
-                        int              mmvx;
-                        int              mmvy;
-                        int              cnt=0;
-                        const MODE_INFO *here = tmp;
-                        const MODE_INFO *above = here - cm->mode_info_stride;
-                        const MODE_INFO *left = here - 1;
-                        const MODE_INFO *aboveleft = above - 1;
-                        const MODE_INFO *aboveright = NULL;
-                        const MODE_INFO *right = NULL;
-                        const MODE_INFO *belowleft = NULL;
-                        const MODE_INFO *below = NULL;
-                        const MODE_INFO *belowright = NULL;
+          if (tmp->mbmi.ref_frame != INTRA_FRAME) {
+            int mvx[8];
+            int mvy[8];
+            int mmvx;
+            int mmvy;
+            int cnt = 0;
+            const MODE_INFO *here = tmp;
+            const MODE_INFO *above = here - cm->mode_info_stride;
+            const MODE_INFO *left = here - 1;
+            const MODE_INFO *aboveleft = above - 1;
+            const MODE_INFO *aboveright = NULL;
+            const MODE_INFO *right = NULL;
+            const MODE_INFO *belowleft = NULL;
+            const MODE_INFO *below = NULL;
+            const MODE_INFO *belowright = NULL;
 
-                        /* If alternate reference frame is used, we have to
-                         * check sign of MV. */
-                        if(cpi->oxcf.play_alternate)
-                        {
-                            /* Gather mv of neighboring MBs */
-                            GET_MV_SIGN(above)
-                            GET_MV_SIGN(left)
-                            GET_MV_SIGN(aboveleft)
+            /* If alternate reference frame is used, we have to
+             * check sign of MV. */
+            if (cpi->oxcf.play_alternate) {
+              /* Gather mv of neighboring MBs */
+              GET_MV_SIGN(above)
+              GET_MV_SIGN(left)
+              GET_MV_SIGN(aboveleft)
 
-                            if(mb_col < (cm->mb_cols-1))
-                            {
-                                right = here + 1;
-                                aboveright = above + 1;
-                                GET_MV_SIGN(right)
-                                GET_MV_SIGN(aboveright)
-                            }
+              if (mb_col < (cm->mb_cols - 1)) {
+                right = here + 1;
+                aboveright = above + 1;
+                GET_MV_SIGN(right)
+                GET_MV_SIGN(aboveright)
+              }
 
-                            if(mb_row < (cm->mb_rows-1))
-                            {
-                                below = here + cm->mode_info_stride;
-                                belowleft = below - 1;
-                                GET_MV_SIGN(below)
-                                GET_MV_SIGN(belowleft)
-                            }
+              if (mb_row < (cm->mb_rows - 1)) {
+                below = here + cm->mode_info_stride;
+                belowleft = below - 1;
+                GET_MV_SIGN(below)
+                GET_MV_SIGN(belowleft)
+              }
 
-                            if(mb_col < (cm->mb_cols-1)
-                                && mb_row < (cm->mb_rows-1))
-                            {
-                                belowright = below + 1;
-                                GET_MV_SIGN(belowright)
-                            }
-                        }else
-                        {
-                            /* No alt_ref and gather mv of neighboring MBs */
-                            GET_MV(above)
-                            GET_MV(left)
-                            GET_MV(aboveleft)
+              if (mb_col < (cm->mb_cols - 1) && mb_row < (cm->mb_rows - 1)) {
+                belowright = below + 1;
+                GET_MV_SIGN(belowright)
+              }
+            } else {
+              /* No alt_ref and gather mv of neighboring MBs */
+              GET_MV(above)
+              GET_MV(left)
+              GET_MV(aboveleft)
 
-                            if(mb_col < (cm->mb_cols-1))
-                            {
-                                right = here + 1;
-                                aboveright = above + 1;
-                                GET_MV(right)
-                                GET_MV(aboveright)
-                            }
+              if (mb_col < (cm->mb_cols - 1)) {
+                right = here + 1;
+                aboveright = above + 1;
+                GET_MV(right)
+                GET_MV(aboveright)
+              }
 
-                            if(mb_row < (cm->mb_rows-1))
-                            {
-                                below = here + cm->mode_info_stride;
-                                belowleft = below - 1;
-                                GET_MV(below)
-                                GET_MV(belowleft)
-                            }
+              if (mb_row < (cm->mb_rows - 1)) {
+                below = here + cm->mode_info_stride;
+                belowleft = below - 1;
+                GET_MV(below)
+                GET_MV(belowleft)
+              }
 
-                            if(mb_col < (cm->mb_cols-1)
-                                && mb_row < (cm->mb_rows-1))
-                            {
-                                belowright = below + 1;
-                                GET_MV(belowright)
-                            }
-                        }
+              if (mb_col < (cm->mb_cols - 1) && mb_row < (cm->mb_rows - 1)) {
+                belowright = below + 1;
+                GET_MV(belowright)
+              }
+            }
 
-                        if (cnt > 0)
-                        {
-                            int max_mvx = mvx[0];
-                            int min_mvx = mvx[0];
-                            int max_mvy = mvy[0];
-                            int min_mvy = mvy[0];
-                            int i;
+            if (cnt > 0) {
+              int max_mvx = mvx[0];
+              int min_mvx = mvx[0];
+              int max_mvy = mvy[0];
+              int min_mvy = mvy[0];
+              int i;
 
-                            if (cnt > 1)
-                            {
-                                for (i=1; i< cnt; i++)
-                                {
-                                    if (mvx[i] > max_mvx) max_mvx = mvx[i];
-                                    else if (mvx[i] < min_mvx) min_mvx = mvx[i];
-                                    if (mvy[i] > max_mvy) max_mvy = mvy[i];
-                                    else if (mvy[i] < min_mvy) min_mvy = mvy[i];
-                                }
-                            }
+              if (cnt > 1) {
+                for (i = 1; i < cnt; i++) {
+                  if (mvx[i] > max_mvx)
+                    max_mvx = mvx[i];
+                  else if (mvx[i] < min_mvx)
+                    min_mvx = mvx[i];
+                  if (mvy[i] > max_mvy)
+                    max_mvy = mvy[i];
+                  else if (mvy[i] < min_mvy)
+                    min_mvy = mvy[i];
+                }
+              }
+
+              mmvx = VPXMAX(abs(min_mvx - here->mbmi.mv.as_mv.row),
+                            abs(max_mvx - here->mbmi.mv.as_mv.row));
+              mmvy = VPXMAX(abs(min_mvy - here->mbmi.mv.as_mv.col),
+                            abs(max_mvy - here->mbmi.mv.as_mv.col));
+              dissim = VPXMAX(mmvx, mmvy);
+            }
+          }
 
-                            mmvx = VPXMAX(
-                                abs(min_mvx - here->mbmi.mv.as_mv.row),
-                                abs(max_mvx - here->mbmi.mv.as_mv.row));
-                            mmvy = VPXMAX(
-                                abs(min_mvy - here->mbmi.mv.as_mv.col),
-                                abs(max_mvy - here->mbmi.mv.as_mv.col));
-                            dissim = VPXMAX(mmvx, mmvy);
-                        }
-                    }
-
-                    /* Store mode info for next resolution encoding */
-                    store_mode_info->mode = tmp->mbmi.mode;
-                    store_mode_info->ref_frame = tmp->mbmi.ref_frame;
-                    store_mode_info->mv.as_int = tmp->mbmi.mv.as_int;
-                    store_mode_info->dissim = dissim;
-                    tmp++;
-                    store_mode_info++;
-                }
-            }
-        }
-    }
-}
+          /* Store mode info for next resolution encoding */
+          store_mode_info->mode = tmp->mbmi.mode;
+          store_mode_info->ref_frame = tmp->mbmi.ref_frame;
+          store_mode_info->mv.as_int = tmp->mbmi.mv.as_int;
+          store_mode_info->dissim = dissim;
+          tmp++;
+          store_mode_info++;
+        }
+      }
+    }
+  }
+}
 
 /* This function is called only when this frame is dropped at current
    resolution level. */
-void vp8_store_drop_frame_info(VP8_COMP *cpi)
-{
-    /* If the frame is dropped in lower-resolution encoding, this information
-       is passed to higher resolution level so that the encoder knows there
-       is no mode & motion info available.
+void vp8_store_drop_frame_info(VP8_COMP *cpi) {
+  /* If the frame is dropped in lower-resolution encoding, this information
+     is passed to higher resolution level so that the encoder knows there
+     is no mode & motion info available.
+   */
+  if (cpi->oxcf.mr_total_resolutions > 1 &&
+      cpi->oxcf.mr_encoder_id < (cpi->oxcf.mr_total_resolutions - 1)) {
+    /* Store info for show/no-show frames for supporting alt_ref.
+     * If parent frame is alt_ref, child has one too.
      */
-    if (cpi->oxcf.mr_total_resolutions >1
-        && cpi->oxcf.mr_encoder_id < (cpi->oxcf.mr_total_resolutions - 1))
-    {
-        /* Store info for show/no-show frames for supporting alt_ref.
-         * If parent frame is alt_ref, child has one too.
-         */
-        LOWER_RES_FRAME_INFO* store_info
-                      = (LOWER_RES_FRAME_INFO*)cpi->oxcf.mr_low_res_mode_info;
+    LOWER_RES_FRAME_INFO *store_info =
+        (LOWER_RES_FRAME_INFO *)cpi->oxcf.mr_low_res_mode_info;
 
-        /* Set frame_type to be INTER_FRAME since we won't drop key frame. */
-        store_info->frame_type = INTER_FRAME;
-        store_info->is_frame_dropped = 1;
-    }
+    /* Set frame_type to be INTER_FRAME since we won't drop key frame. */
+    store_info->frame_type = INTER_FRAME;
+    store_info->is_frame_dropped = 1;
+  }
 }
--- a/vp8/encoder/mr_dissim.h
+++ b/vp8/encoder/mr_dissim.h
@@ -8,7 +8,6 @@
  *  be found in the AUTHORS file in the root of the source tree.
  */
 
-
 #ifndef VP8_ENCODER_MR_DISSIM_H_
 #define VP8_ENCODER_MR_DISSIM_H_
 #include "vpx_config.h"
--- a/vp8/encoder/onyx_if.c
+++ b/vp8/encoder/onyx_if.c
@@ -8,7 +8,6 @@
  *  be found in the AUTHORS file in the root of the source tree.
  */
 
-
 #include "vpx_config.h"
 #include "./vpx_scale_rtcd.h"
 #include "./vpx_dsp_rtcd.h"
@@ -56,7 +55,9 @@
 extern void vp8cx_set_alt_lf_level(VP8_COMP *cpi, int filt_val);
 extern void vp8cx_pick_filter_level(YV12_BUFFER_CONFIG *sd, VP8_COMP *cpi);
 
-extern void vp8_deblock_frame(YV12_BUFFER_CONFIG *source, YV12_BUFFER_CONFIG *post, int filt_lvl, int low_var_thresh, int flag);
+extern void vp8_deblock_frame(YV12_BUFFER_CONFIG *source,
+                              YV12_BUFFER_CONFIG *post, int filt_lvl,
+                              int low_var_thresh, int flag);
 extern void print_parms(VP8_CONFIG *ocf, char *filenam);
 extern unsigned int vp8_get_processor_freq();
 extern void print_tree_update_probs();
@@ -78,7 +79,6 @@
 #include "vpx_dsp/ssim.h"
 #endif
 
-
 #ifdef OUTPUT_YUV_SRC
 FILE *yuv_file;
 #endif
@@ -97,13 +97,13 @@
 extern int skip_false_count;
 #endif
 
-
 #ifdef VP8_ENTROPY_STATS
 extern int intra_mode_stats[10][10][10];
 #endif
 
 #ifdef SPEEDSTATS
-unsigned int frames_at_speed[16] = {0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0};
+unsigned int frames_at_speed[16] = { 0, 0, 0, 0, 0, 0, 0, 0,
+                                     0, 0, 0, 0, 0, 0, 0, 0 };
 unsigned int tot_pm = 0;
 unsigned int cnt_pm = 0;
 unsigned int tot_ef = 0;
@@ -112,12 +112,12 @@
 
 #ifdef MODE_STATS
 extern unsigned __int64 Sectionbits[50];
-extern int y_modes[5]  ;
-extern int uv_modes[4] ;
-extern int b_modes[10]  ;
+extern int y_modes[5];
+extern int uv_modes[4];
+extern int b_modes[10];
 
-extern int inter_y_modes[10] ;
-extern int inter_uv_modes[4] ;
+extern int inter_y_modes[10];
+extern int inter_uv_modes[4];
 extern unsigned int inter_b_modes[15];
 #endif
 
@@ -129,71 +129,59 @@
 extern const int vp8cx_base_skip_false_prob[128];
 
 /* Tables relating active max Q to active min Q */
-static const unsigned char kf_low_motion_minq[QINDEX_RANGE] =
-{
-    0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,
-    0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,
-    0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,
-    0,0,0,0,1,1,1,1,1,1,1,1,2,2,2,2,
-    3,3,3,3,3,3,4,4,4,5,5,5,5,5,6,6,
-    6,6,7,7,8,8,8,8,9,9,10,10,10,10,11,11,
-    11,11,12,12,13,13,13,13,14,14,15,15,15,15,16,16,
-    16,16,17,17,18,18,18,18,19,20,20,21,21,22,23,23
+static const unsigned char kf_low_motion_minq[QINDEX_RANGE] = {
+  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,
+  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,
+  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  1,  1,  1,  1,  1,
+  1,  1,  1,  2,  2,  2,  2,  3,  3,  3,  3,  3,  3,  4,  4,  4,  5,  5,  5,
+  5,  5,  6,  6,  6,  6,  7,  7,  8,  8,  8,  8,  9,  9,  10, 10, 10, 10, 11,
+  11, 11, 11, 12, 12, 13, 13, 13, 13, 14, 14, 15, 15, 15, 15, 16, 16, 16, 16,
+  17, 17, 18, 18, 18, 18, 19, 20, 20, 21, 21, 22, 23, 23
 };
-static const unsigned char kf_high_motion_minq[QINDEX_RANGE] =
-{
-    0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,
-    0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,
-    1,1,1,1,1,1,1,1,2,2,2,2,3,3,3,3,
-    3,3,3,3,4,4,4,4,5,5,5,5,5,5,6,6,
-    6,6,7,7,8,8,8,8,9,9,10,10,10,10,11,11,
-    11,11,12,12,13,13,13,13,14,14,15,15,15,15,16,16,
-    16,16,17,17,18,18,18,18,19,19,20,20,20,20,21,21,
-    21,21,22,22,23,23,24,25,25,26,26,27,28,28,29,30
+static const unsigned char kf_high_motion_minq[QINDEX_RANGE] = {
+  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,
+  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  1,  1,  1,  1,  1,  1,
+  1,  1,  2,  2,  2,  2,  3,  3,  3,  3,  3,  3,  3,  3,  4,  4,  4,  4,  5,
+  5,  5,  5,  5,  5,  6,  6,  6,  6,  7,  7,  8,  8,  8,  8,  9,  9,  10, 10,
+  10, 10, 11, 11, 11, 11, 12, 12, 13, 13, 13, 13, 14, 14, 15, 15, 15, 15, 16,
+  16, 16, 16, 17, 17, 18, 18, 18, 18, 19, 19, 20, 20, 20, 20, 21, 21, 21, 21,
+  22, 22, 23, 23, 24, 25, 25, 26, 26, 27, 28, 28, 29, 30
 };
-static const unsigned char gf_low_motion_minq[QINDEX_RANGE] =
-{
-    0,0,0,0,1,1,1,1,1,1,1,1,2,2,2,2,
-    3,3,3,3,4,4,4,4,5,5,5,5,6,6,6,6,
-    7,7,7,7,8,8,8,8,9,9,9,9,10,10,10,10,
-    11,11,12,12,13,13,14,14,15,15,16,16,17,17,18,18,
-    19,19,20,20,21,21,22,22,23,23,24,24,25,25,26,26,
-    27,27,28,28,29,29,30,30,31,31,32,32,33,33,34,34,
-    35,35,36,36,37,37,38,38,39,39,40,40,41,41,42,42,
-    43,44,45,46,47,48,49,50,51,52,53,54,55,56,57,58
+static const unsigned char gf_low_motion_minq[QINDEX_RANGE] = {
+  0,  0,  0,  0,  1,  1,  1,  1,  1,  1,  1,  1,  2,  2,  2,  2,  3,  3,  3,
+  3,  4,  4,  4,  4,  5,  5,  5,  5,  6,  6,  6,  6,  7,  7,  7,  7,  8,  8,
+  8,  8,  9,  9,  9,  9,  10, 10, 10, 10, 11, 11, 12, 12, 13, 13, 14, 14, 15,
+  15, 16, 16, 17, 17, 18, 18, 19, 19, 20, 20, 21, 21, 22, 22, 23, 23, 24, 24,
+  25, 25, 26, 26, 27, 27, 28, 28, 29, 29, 30, 30, 31, 31, 32, 32, 33, 33, 34,
+  34, 35, 35, 36, 36, 37, 37, 38, 38, 39, 39, 40, 40, 41, 41, 42, 42, 43, 44,
+  45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58
 };
-static const unsigned char gf_mid_motion_minq[QINDEX_RANGE] =
-{
-    0,0,0,0,1,1,1,1,1,1,2,2,3,3,3,4,
-    4,4,5,5,5,6,6,6,7,7,7,8,8,8,9,9,
-    9,10,10,10,10,11,11,11,12,12,12,12,13,13,13,14,
-    14,14,15,15,16,16,17,17,18,18,19,19,20,20,21,21,
-    22,22,23,23,24,24,25,25,26,26,27,27,28,28,29,29,
-    30,30,31,31,32,32,33,33,34,34,35,35,36,36,37,37,
-    38,39,39,40,40,41,41,42,42,43,43,44,45,46,47,48,
-    49,50,51,52,53,54,55,56,57,58,59,60,61,62,63,64
+static const unsigned char gf_mid_motion_minq[QINDEX_RANGE] = {
+  0,  0,  0,  0,  1,  1,  1,  1,  1,  1,  2,  2,  3,  3,  3,  4,  4,  4,  5,
+  5,  5,  6,  6,  6,  7,  7,  7,  8,  8,  8,  9,  9,  9,  10, 10, 10, 10, 11,
+  11, 11, 12, 12, 12, 12, 13, 13, 13, 14, 14, 14, 15, 15, 16, 16, 17, 17, 18,
+  18, 19, 19, 20, 20, 21, 21, 22, 22, 23, 23, 24, 24, 25, 25, 26, 26, 27, 27,
+  28, 28, 29, 29, 30, 30, 31, 31, 32, 32, 33, 33, 34, 34, 35, 35, 36, 36, 37,
+  37, 38, 39, 39, 40, 40, 41, 41, 42, 42, 43, 43, 44, 45, 46, 47, 48, 49, 50,
+  51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63, 64
 };
-static const unsigned char gf_high_motion_minq[QINDEX_RANGE] =
-{
-    0,0,0,0,1,1,1,1,1,2,2,2,3,3,3,4,
-    4,4,5,5,5,6,6,6,7,7,7,8,8,8,9,9,
-    9,10,10,10,11,11,12,12,13,13,14,14,15,15,16,16,
-    17,17,18,18,19,19,20,20,21,21,22,22,23,23,24,24,
-    25,25,26,26,27,27,28,28,29,29,30,30,31,31,32,32,
-    33,33,34,34,35,35,36,36,37,37,38,38,39,39,40,40,
-    41,41,42,42,43,44,45,46,47,48,49,50,51,52,53,54,
-    55,56,57,58,59,60,62,64,66,68,70,72,74,76,78,80
+static const unsigned char gf_high_motion_minq[QINDEX_RANGE] = {
+  0,  0,  0,  0,  1,  1,  1,  1,  1,  2,  2,  2,  3,  3,  3,  4,  4,  4,  5,
+  5,  5,  6,  6,  6,  7,  7,  7,  8,  8,  8,  9,  9,  9,  10, 10, 10, 11, 11,
+  12, 12, 13, 13, 14, 14, 15, 15, 16, 16, 17, 17, 18, 18, 19, 19, 20, 20, 21,
+  21, 22, 22, 23, 23, 24, 24, 25, 25, 26, 26, 27, 27, 28, 28, 29, 29, 30, 30,
+  31, 31, 32, 32, 33, 33, 34, 34, 35, 35, 36, 36, 37, 37, 38, 38, 39, 39, 40,
+  40, 41, 41, 42, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56,
+  57, 58, 59, 60, 62, 64, 66, 68, 70, 72, 74, 76, 78, 80
 };
-static const unsigned char inter_minq[QINDEX_RANGE] =
-{
-    0,0,1,1,2,3,3,4,4,5,6,6,7,8,8,9,
-    9,10,11,11,12,13,13,14,15,15,16,17,17,18,19,20,
-    20,21,22,22,23,24,24,25,26,27,27,28,29,30,30,31,
-    32,33,33,34,35,36,36,37,38,39,39,40,41,42,42,43,
-    44,45,46,46,47,48,49,50,50,51,52,53,54,55,55,56,
-    57,58,59,60,60,61,62,63,64,65,66,67,67,68,69,70,
-    71,72,73,74,75,75,76,77,78,79,80,81,82,83,84,85,
-    86,86,87,88,89,90,91,92,93,94,95,96,97,98,99,100
+static const unsigned char inter_minq[QINDEX_RANGE] = {
+  0,  0,  1,  1,  2,  3,  3,  4,  4,  5,  6,  6,  7,  8,  8,  9,  9,  10, 11,
+  11, 12, 13, 13, 14, 15, 15, 16, 17, 17, 18, 19, 20, 20, 21, 22, 22, 23, 24,
+  24, 25, 26, 27, 27, 28, 29, 30, 30, 31, 32, 33, 33, 34, 35, 36, 36, 37, 38,
+  39, 39, 40, 41, 42, 42, 43, 44, 45, 46, 46, 47, 48, 49, 50, 50, 51, 52, 53,
+  54, 55, 55, 56, 57, 58, 59, 60, 60, 61, 62, 63, 64, 65, 66, 67, 67, 68, 69,
+  70, 71, 72, 73, 74, 75, 75, 76, 77, 78, 79, 80, 81, 82, 83, 84, 85, 86, 86,
+  87, 88, 89, 90, 91, 92, 93, 94, 95, 96, 97, 98, 99, 100
 };
 
 #ifdef PACKET_TESTING
@@ -200,145 +188,135 @@
 extern FILE *vpxlogc;
 #endif
 
-static void save_layer_context(VP8_COMP *cpi)
-{
-    LAYER_CONTEXT *lc = &cpi->layer_context[cpi->current_layer];
+static void save_layer_context(VP8_COMP *cpi) {
+  LAYER_CONTEXT *lc = &cpi->layer_context[cpi->current_layer];
 
-    /* Save layer dependent coding state */
-    lc->target_bandwidth                 = cpi->target_bandwidth;
-    lc->starting_buffer_level            = cpi->oxcf.starting_buffer_level;
-    lc->optimal_buffer_level             = cpi->oxcf.optimal_buffer_level;
-    lc->maximum_buffer_size              = cpi->oxcf.maximum_buffer_size;
-    lc->starting_buffer_level_in_ms      = cpi->oxcf.starting_buffer_level_in_ms;
-    lc->optimal_buffer_level_in_ms       = cpi->oxcf.optimal_buffer_level_in_ms;
-    lc->maximum_buffer_size_in_ms        = cpi->oxcf.maximum_buffer_size_in_ms;
-    lc->buffer_level                     = cpi->buffer_level;
-    lc->bits_off_target                  = cpi->bits_off_target;
-    lc->total_actual_bits                = cpi->total_actual_bits;
-    lc->worst_quality                    = cpi->worst_quality;
-    lc->active_worst_quality             = cpi->active_worst_quality;
-    lc->best_quality                     = cpi->best_quality;
-    lc->active_best_quality              = cpi->active_best_quality;
-    lc->ni_av_qi                         = cpi->ni_av_qi;
-    lc->ni_tot_qi                        = cpi->ni_tot_qi;
-    lc->ni_frames                        = cpi->ni_frames;
-    lc->avg_frame_qindex                 = cpi->avg_frame_qindex;
-    lc->rate_correction_factor           = cpi->rate_correction_factor;
-    lc->key_frame_rate_correction_factor = cpi->key_frame_rate_correction_factor;
-    lc->gf_rate_correction_factor        = cpi->gf_rate_correction_factor;
-    lc->zbin_over_quant                  = cpi->mb.zbin_over_quant;
-    lc->inter_frame_target               = cpi->inter_frame_target;
-    lc->total_byte_count                 = cpi->total_byte_count;
-    lc->filter_level                     = cpi->common.filter_level;
+  /* Save layer dependent coding state */
+  lc->target_bandwidth = cpi->target_bandwidth;
+  lc->starting_buffer_level = cpi->oxcf.starting_buffer_level;
+  lc->optimal_buffer_level = cpi->oxcf.optimal_buffer_level;
+  lc->maximum_buffer_size = cpi->oxcf.maximum_buffer_size;
+  lc->starting_buffer_level_in_ms = cpi->oxcf.starting_buffer_level_in_ms;
+  lc->optimal_buffer_level_in_ms = cpi->oxcf.optimal_buffer_level_in_ms;
+  lc->maximum_buffer_size_in_ms = cpi->oxcf.maximum_buffer_size_in_ms;
+  lc->buffer_level = cpi->buffer_level;
+  lc->bits_off_target = cpi->bits_off_target;
+  lc->total_actual_bits = cpi->total_actual_bits;
+  lc->worst_quality = cpi->worst_quality;
+  lc->active_worst_quality = cpi->active_worst_quality;
+  lc->best_quality = cpi->best_quality;
+  lc->active_best_quality = cpi->active_best_quality;
+  lc->ni_av_qi = cpi->ni_av_qi;
+  lc->ni_tot_qi = cpi->ni_tot_qi;
+  lc->ni_frames = cpi->ni_frames;
+  lc->avg_frame_qindex = cpi->avg_frame_qindex;
+  lc->rate_correction_factor = cpi->rate_correction_factor;
+  lc->key_frame_rate_correction_factor = cpi->key_frame_rate_correction_factor;
+  lc->gf_rate_correction_factor = cpi->gf_rate_correction_factor;
+  lc->zbin_over_quant = cpi->mb.zbin_over_quant;
+  lc->inter_frame_target = cpi->inter_frame_target;
+  lc->total_byte_count = cpi->total_byte_count;
+  lc->filter_level = cpi->common.filter_level;
 
-    lc->last_frame_percent_intra         = cpi->last_frame_percent_intra;
+  lc->last_frame_percent_intra = cpi->last_frame_percent_intra;
 
-    memcpy (lc->count_mb_ref_frame_usage,
-            cpi->mb.count_mb_ref_frame_usage,
-            sizeof(cpi->mb.count_mb_ref_frame_usage));
+  memcpy(lc->count_mb_ref_frame_usage, cpi->mb.count_mb_ref_frame_usage,
+         sizeof(cpi->mb.count_mb_ref_frame_usage));
 }
 
-static void restore_layer_context(VP8_COMP *cpi, const int layer)
-{
-    LAYER_CONTEXT *lc = &cpi->layer_context[layer];
+static void restore_layer_context(VP8_COMP *cpi, const int layer) {
+  LAYER_CONTEXT *lc = &cpi->layer_context[layer];
 
-    /* Restore layer dependent coding state */
-    cpi->current_layer                    = layer;
-    cpi->target_bandwidth                 = lc->target_bandwidth;
-    cpi->oxcf.target_bandwidth            = lc->target_bandwidth;
-    cpi->oxcf.starting_buffer_level       = lc->starting_buffer_level;
-    cpi->oxcf.optimal_buffer_level        = lc->optimal_buffer_level;
-    cpi->oxcf.maximum_buffer_size         = lc->maximum_buffer_size;
-    cpi->oxcf.starting_buffer_level_in_ms = lc->starting_buffer_level_in_ms;
-    cpi->oxcf.optimal_buffer_level_in_ms  = lc->optimal_buffer_level_in_ms;
-    cpi->oxcf.maximum_buffer_size_in_ms   = lc->maximum_buffer_size_in_ms;
-    cpi->buffer_level                     = lc->buffer_level;
-    cpi->bits_off_target                  = lc->bits_off_target;
-    cpi->total_actual_bits                = lc->total_actual_bits;
-    cpi->active_worst_quality             = lc->active_worst_quality;
-    cpi->active_best_quality              = lc->active_best_quality;
-    cpi->ni_av_qi                         = lc->ni_av_qi;
-    cpi->ni_tot_qi                        = lc->ni_tot_qi;
-    cpi->ni_frames                        = lc->ni_frames;
-    cpi->avg_frame_qindex                 = lc->avg_frame_qindex;
-    cpi->rate_correction_factor           = lc->rate_correction_factor;
-    cpi->key_frame_rate_correction_factor = lc->key_frame_rate_correction_factor;
-    cpi->gf_rate_correction_factor        = lc->gf_rate_correction_factor;
-    cpi->mb.zbin_over_quant                  = lc->zbin_over_quant;
-    cpi->inter_frame_target               = lc->inter_frame_target;
-    cpi->total_byte_count                 = lc->total_byte_count;
-    cpi->common.filter_level              = lc->filter_level;
+  /* Restore layer dependent coding state */
+  cpi->current_layer = layer;
+  cpi->target_bandwidth = lc->target_bandwidth;
+  cpi->oxcf.target_bandwidth = lc->target_bandwidth;
+  cpi->oxcf.starting_buffer_level = lc->starting_buffer_level;
+  cpi->oxcf.optimal_buffer_level = lc->optimal_buffer_level;
+  cpi->oxcf.maximum_buffer_size = lc->maximum_buffer_size;
+  cpi->oxcf.starting_buffer_level_in_ms = lc->starting_buffer_level_in_ms;
+  cpi->oxcf.optimal_buffer_level_in_ms = lc->optimal_buffer_level_in_ms;
+  cpi->oxcf.maximum_buffer_size_in_ms = lc->maximum_buffer_size_in_ms;
+  cpi->buffer_level = lc->buffer_level;
+  cpi->bits_off_target = lc->bits_off_target;
+  cpi->total_actual_bits = lc->total_actual_bits;
+  cpi->active_worst_quality = lc->active_worst_quality;
+  cpi->active_best_quality = lc->active_best_quality;
+  cpi->ni_av_qi = lc->ni_av_qi;
+  cpi->ni_tot_qi = lc->ni_tot_qi;
+  cpi->ni_frames = lc->ni_frames;
+  cpi->avg_frame_qindex = lc->avg_frame_qindex;
+  cpi->rate_correction_factor = lc->rate_correction_factor;
+  cpi->key_frame_rate_correction_factor = lc->key_frame_rate_correction_factor;
+  cpi->gf_rate_correction_factor = lc->gf_rate_correction_factor;
+  cpi->mb.zbin_over_quant = lc->zbin_over_quant;
+  cpi->inter_frame_target = lc->inter_frame_target;
+  cpi->total_byte_count = lc->total_byte_count;
+  cpi->common.filter_level = lc->filter_level;
 
-    cpi->last_frame_percent_intra         = lc->last_frame_percent_intra;
+  cpi->last_frame_percent_intra = lc->last_frame_percent_intra;
 
-    memcpy (cpi->mb.count_mb_ref_frame_usage,
-            lc->count_mb_ref_frame_usage,
-            sizeof(cpi->mb.count_mb_ref_frame_usage));
+  memcpy(cpi->mb.count_mb_ref_frame_usage, lc->count_mb_ref_frame_usage,
+         sizeof(cpi->mb.count_mb_ref_frame_usage));
 }
 
-static int rescale(int val, int num, int denom)
-{
-    int64_t llnum = num;
-    int64_t llden = denom;
-    int64_t llval = val;
+static int rescale(int val, int num, int denom) {
+  int64_t llnum = num;
+  int64_t llden = denom;
+  int64_t llval = val;
 
-    return (int)(llval * llnum / llden);
+  return (int)(llval * llnum / llden);
 }
 
-static void init_temporal_layer_context(VP8_COMP *cpi,
-                                        VP8_CONFIG *oxcf,
+static void init_temporal_layer_context(VP8_COMP *cpi, VP8_CONFIG *oxcf,
                                         const int layer,
-                                        double prev_layer_framerate)
-{
-    LAYER_CONTEXT *lc = &cpi->layer_context[layer];
+                                        double prev_layer_framerate) {
+  LAYER_CONTEXT *lc = &cpi->layer_context[layer];
 
-    lc->framerate = cpi->output_framerate / cpi->oxcf.rate_decimator[layer];
-    lc->target_bandwidth = cpi->oxcf.target_bitrate[layer] * 1000;
+  lc->framerate = cpi->output_framerate / cpi->oxcf.rate_decimator[layer];
+  lc->target_bandwidth = cpi->oxcf.target_bitrate[layer] * 1000;
 
-    lc->starting_buffer_level_in_ms = oxcf->starting_buffer_level;
-    lc->optimal_buffer_level_in_ms  = oxcf->optimal_buffer_level;
-    lc->maximum_buffer_size_in_ms   = oxcf->maximum_buffer_size;
+  lc->starting_buffer_level_in_ms = oxcf->starting_buffer_level;
+  lc->optimal_buffer_level_in_ms = oxcf->optimal_buffer_level;
+  lc->maximum_buffer_size_in_ms = oxcf->maximum_buffer_size;
 
-    lc->starting_buffer_level =
-        rescale((int)(oxcf->starting_buffer_level),
-                lc->target_bandwidth, 1000);
+  lc->starting_buffer_level =
+      rescale((int)(oxcf->starting_buffer_level), lc->target_bandwidth, 1000);
 
-    if (oxcf->optimal_buffer_level == 0)
-      lc->optimal_buffer_level = lc->target_bandwidth / 8;
-    else
-      lc->optimal_buffer_level =
-          rescale((int)(oxcf->optimal_buffer_level),
-                  lc->target_bandwidth, 1000);
+  if (oxcf->optimal_buffer_level == 0)
+    lc->optimal_buffer_level = lc->target_bandwidth / 8;
+  else
+    lc->optimal_buffer_level =
+        rescale((int)(oxcf->optimal_buffer_level), lc->target_bandwidth, 1000);
 
-    if (oxcf->maximum_buffer_size == 0)
-      lc->maximum_buffer_size = lc->target_bandwidth / 8;
-    else
-      lc->maximum_buffer_size =
-          rescale((int)(oxcf->maximum_buffer_size),
-                  lc->target_bandwidth, 1000);
+  if (oxcf->maximum_buffer_size == 0)
+    lc->maximum_buffer_size = lc->target_bandwidth / 8;
+  else
+    lc->maximum_buffer_size =
+        rescale((int)(oxcf->maximum_buffer_size), lc->target_bandwidth, 1000);
 
-    /* Work out the average size of a frame within this layer */
-    if (layer > 0)
-      lc->avg_frame_size_for_layer =
-          (int)((cpi->oxcf.target_bitrate[layer] -
-                cpi->oxcf.target_bitrate[layer-1]) * 1000 /
-                (lc->framerate - prev_layer_framerate));
+  /* Work out the average size of a frame within this layer */
+  if (layer > 0)
+    lc->avg_frame_size_for_layer =
+        (int)((cpi->oxcf.target_bitrate[layer] -
+               cpi->oxcf.target_bitrate[layer - 1]) *
+              1000 / (lc->framerate - prev_layer_framerate));
 
-     lc->active_worst_quality         = cpi->oxcf.worst_allowed_q;
-     lc->active_best_quality          = cpi->oxcf.best_allowed_q;
-     lc->avg_frame_qindex             = cpi->oxcf.worst_allowed_q;
+  lc->active_worst_quality = cpi->oxcf.worst_allowed_q;
+  lc->active_best_quality = cpi->oxcf.best_allowed_q;
+  lc->avg_frame_qindex = cpi->oxcf.worst_allowed_q;
 
-     lc->buffer_level                 = lc->starting_buffer_level;
-     lc->bits_off_target              = lc->starting_buffer_level;
+  lc->buffer_level = lc->starting_buffer_level;
+  lc->bits_off_target = lc->starting_buffer_level;
 
-     lc->total_actual_bits                 = 0;
-     lc->ni_av_qi                          = 0;
-     lc->ni_tot_qi                         = 0;
-     lc->ni_frames                         = 0;
-     lc->rate_correction_factor            = 1.0;
-     lc->key_frame_rate_correction_factor  = 1.0;
-     lc->gf_rate_correction_factor         = 1.0;
-     lc->inter_frame_target                = 0;
+  lc->total_actual_bits = 0;
+  lc->ni_av_qi = 0;
+  lc->ni_tot_qi = 0;
+  lc->ni_frames = 0;
+  lc->rate_correction_factor = 1.0;
+  lc->key_frame_rate_correction_factor = 1.0;
+  lc->gf_rate_correction_factor = 1.0;
+  lc->inter_frame_target = 0;
 }
 
 // Upon a run-time change in temporal layers, reset the layer context parameters
@@ -345,179 +323,165 @@
 // for any "new" layers. For "existing" layers, let them inherit the parameters
 // from the previous layer state (at the same layer #). In future we may want
 // to better map the previous layer state(s) to the "new" ones.
-static void reset_temporal_layer_change(VP8_COMP *cpi,
-                                        VP8_CONFIG *oxcf,
-                                        const int prev_num_layers)
-{
-    int i;
-    double prev_layer_framerate = 0;
-    const int curr_num_layers = cpi->oxcf.number_of_layers;
-    // If the previous state was 1 layer, get current layer context from cpi.
-    // We need this to set the layer context for the new layers below.
-    if (prev_num_layers == 1)
-    {
-        cpi->current_layer = 0;
-        save_layer_context(cpi);
+static void reset_temporal_layer_change(VP8_COMP *cpi, VP8_CONFIG *oxcf,
+                                        const int prev_num_layers) {
+  int i;
+  double prev_layer_framerate = 0;
+  const int curr_num_layers = cpi->oxcf.number_of_layers;
+  // If the previous state was 1 layer, get current layer context from cpi.
+  // We need this to set the layer context for the new layers below.
+  if (prev_num_layers == 1) {
+    cpi->current_layer = 0;
+    save_layer_context(cpi);
+  }
+  for (i = 0; i < curr_num_layers; i++) {
+    LAYER_CONTEXT *lc = &cpi->layer_context[i];
+    if (i >= prev_num_layers) {
+      init_temporal_layer_context(cpi, oxcf, i, prev_layer_framerate);
     }
-    for (i = 0; i < curr_num_layers; i++)
-    {
-        LAYER_CONTEXT *lc = &cpi->layer_context[i];
-        if (i >= prev_num_layers)
-        {
-           init_temporal_layer_context(cpi, oxcf, i, prev_layer_framerate);
-        }
-        // The initial buffer levels are set based on their starting levels.
-        // We could set the buffer levels based on the previous state (normalized
-        // properly by the layer bandwidths) but we would need to keep track of
-        // the previous set of layer bandwidths (i.e., target_bitrate[i])
-        // before the layer change. For now, reset to the starting levels.
-        lc->buffer_level = cpi->oxcf.starting_buffer_level_in_ms *
-                           cpi->oxcf.target_bitrate[i];
-        lc->bits_off_target = lc->buffer_level;
-        // TDOD(marpan): Should we set the rate_correction_factor and
-        // active_worst/best_quality to values derived from the previous layer
-        // state (to smooth-out quality dips/rate fluctuation at transition)?
+    // The initial buffer levels are set based on their starting levels.
+    // We could set the buffer levels based on the previous state (normalized
+    // properly by the layer bandwidths) but we would need to keep track of
+    // the previous set of layer bandwidths (i.e., target_bitrate[i])
+    // before the layer change. For now, reset to the starting levels.
+    lc->buffer_level =
+        cpi->oxcf.starting_buffer_level_in_ms * cpi->oxcf.target_bitrate[i];
+    lc->bits_off_target = lc->buffer_level;
+    // TDOD(marpan): Should we set the rate_correction_factor and
+    // active_worst/best_quality to values derived from the previous layer
+    // state (to smooth-out quality dips/rate fluctuation at transition)?
 
-        // We need to treat the 1 layer case separately: oxcf.target_bitrate[i]
-        // is not set for 1 layer, and the restore_layer_context/save_context()
-        // are not called in the encoding loop, so we need to call it here to
-        // pass the layer context state to |cpi|.
-        if (curr_num_layers == 1)
-        {
-            lc->target_bandwidth = cpi->oxcf.target_bandwidth;
-            lc->buffer_level = cpi->oxcf.starting_buffer_level_in_ms *
-                               lc->target_bandwidth  / 1000;
-            lc->bits_off_target = lc->buffer_level;
-            restore_layer_context(cpi, 0);
-        }
-        prev_layer_framerate = cpi->output_framerate /
-                               cpi->oxcf.rate_decimator[i];
+    // We need to treat the 1 layer case separately: oxcf.target_bitrate[i]
+    // is not set for 1 layer, and the restore_layer_context/save_context()
+    // are not called in the encoding loop, so we need to call it here to
+    // pass the layer context state to |cpi|.
+    if (curr_num_layers == 1) {
+      lc->target_bandwidth = cpi->oxcf.target_bandwidth;
+      lc->buffer_level =
+          cpi->oxcf.starting_buffer_level_in_ms * lc->target_bandwidth / 1000;
+      lc->bits_off_target = lc->buffer_level;
+      restore_layer_context(cpi, 0);
     }
+    prev_layer_framerate = cpi->output_framerate / cpi->oxcf.rate_decimator[i];
+  }
 }
 
-static void setup_features(VP8_COMP *cpi)
-{
-    // If segmentation enabled set the update flags
-    if ( cpi->mb.e_mbd.segmentation_enabled )
-    {
-        cpi->mb.e_mbd.update_mb_segmentation_map = 1;
-        cpi->mb.e_mbd.update_mb_segmentation_data = 1;
-    }
-    else
-    {
-        cpi->mb.e_mbd.update_mb_segmentation_map = 0;
-        cpi->mb.e_mbd.update_mb_segmentation_data = 0;
-    }
+static void setup_features(VP8_COMP *cpi) {
+  // If segmentation enabled set the update flags
+  if (cpi->mb.e_mbd.segmentation_enabled) {
+    cpi->mb.e_mbd.update_mb_segmentation_map = 1;
+    cpi->mb.e_mbd.update_mb_segmentation_data = 1;
+  } else {
+    cpi->mb.e_mbd.update_mb_segmentation_map = 0;
+    cpi->mb.e_mbd.update_mb_segmentation_data = 0;
+  }
 
-    cpi->mb.e_mbd.mode_ref_lf_delta_enabled = 0;
-    cpi->mb.e_mbd.mode_ref_lf_delta_update = 0;
-    memset(cpi->mb.e_mbd.ref_lf_deltas, 0, sizeof(cpi->mb.e_mbd.ref_lf_deltas));
-    memset(cpi->mb.e_mbd.mode_lf_deltas, 0, sizeof(cpi->mb.e_mbd.mode_lf_deltas));
-    memset(cpi->mb.e_mbd.last_ref_lf_deltas, 0, sizeof(cpi->mb.e_mbd.ref_lf_deltas));
-    memset(cpi->mb.e_mbd.last_mode_lf_deltas, 0, sizeof(cpi->mb.e_mbd.mode_lf_deltas));
+  cpi->mb.e_mbd.mode_ref_lf_delta_enabled = 0;
+  cpi->mb.e_mbd.mode_ref_lf_delta_update = 0;
+  memset(cpi->mb.e_mbd.ref_lf_deltas, 0, sizeof(cpi->mb.e_mbd.ref_lf_deltas));
+  memset(cpi->mb.e_mbd.mode_lf_deltas, 0, sizeof(cpi->mb.e_mbd.mode_lf_deltas));
+  memset(cpi->mb.e_mbd.last_ref_lf_deltas, 0,
+         sizeof(cpi->mb.e_mbd.ref_lf_deltas));
+  memset(cpi->mb.e_mbd.last_mode_lf_deltas, 0,
+         sizeof(cpi->mb.e_mbd.mode_lf_deltas));
 
-    set_default_lf_deltas(cpi);
-
+  set_default_lf_deltas(cpi);
 }
 
-
 static void dealloc_raw_frame_buffers(VP8_COMP *cpi);
 
-void vp8_initialize_enc(void)
-{
-    static volatile int init_done = 0;
+void vp8_initialize_enc(void) {
+  static volatile int init_done = 0;
 
-    if (!init_done) {
-        vpx_dsp_rtcd();
-        vp8_init_intra_predictors();
-        init_done = 1;
-    }
+  if (!init_done) {
+    vpx_dsp_rtcd();
+    vp8_init_intra_predictors();
+    init_done = 1;
+  }
 }
 
-static void dealloc_compressor_data(VP8_COMP *cpi)
-{
-    vpx_free(cpi->tplist);
-    cpi->tplist = NULL;
+static void dealloc_compressor_data(VP8_COMP *cpi) {
+  vpx_free(cpi->tplist);
+  cpi->tplist = NULL;
 
-    /* Delete last frame MV storage buffers */
-    vpx_free(cpi->lfmv);
-    cpi->lfmv = 0;
+  /* Delete last frame MV storage buffers */
+  vpx_free(cpi->lfmv);
+  cpi->lfmv = 0;
 
-    vpx_free(cpi->lf_ref_frame_sign_bias);
-    cpi->lf_ref_frame_sign_bias = 0;
+  vpx_free(cpi->lf_ref_frame_sign_bias);
+  cpi->lf_ref_frame_sign_bias = 0;
 
-    vpx_free(cpi->lf_ref_frame);
-    cpi->lf_ref_frame = 0;
+  vpx_free(cpi->lf_ref_frame);
+  cpi->lf_ref_frame = 0;
 
-    /* Delete sementation map */
-    vpx_free(cpi->segmentation_map);
-    cpi->segmentation_map = 0;
+  /* Delete sementation map */
+  vpx_free(cpi->segmentation_map);
+  cpi->segmentation_map = 0;
 
-    vpx_free(cpi->active_map);
-    cpi->active_map = 0;
+  vpx_free(cpi->active_map);
+  cpi->active_map = 0;
 
-    vp8_de_alloc_frame_buffers(&cpi->common);
+  vp8_de_alloc_frame_buffers(&cpi->common);
 
-    vp8_yv12_de_alloc_frame_buffer(&cpi->pick_lf_lvl_frame);
-    vp8_yv12_de_alloc_frame_buffer(&cpi->scaled_source);
-    dealloc_raw_frame_buffers(cpi);
+  vp8_yv12_de_alloc_frame_buffer(&cpi->pick_lf_lvl_frame);
+  vp8_yv12_de_alloc_frame_buffer(&cpi->scaled_source);
+  dealloc_raw_frame_buffers(cpi);
 
-    vpx_free(cpi->tok);
-    cpi->tok = 0;
+  vpx_free(cpi->tok);
+  cpi->tok = 0;
 
-    /* Structure used to monitor GF usage */
-    vpx_free(cpi->gf_active_flags);
-    cpi->gf_active_flags = 0;
+  /* Structure used to monitor GF usage */
+  vpx_free(cpi->gf_active_flags);
+  cpi->gf_active_flags = 0;
 
-    /* Activity mask based per mb zbin adjustments */
-    vpx_free(cpi->mb_activity_map);
-    cpi->mb_activity_map = 0;
+  /* Activity mask based per mb zbin adjustments */
+  vpx_free(cpi->mb_activity_map);
+  cpi->mb_activity_map = 0;
 
-    vpx_free(cpi->mb.pip);
-    cpi->mb.pip = 0;
+  vpx_free(cpi->mb.pip);
+  cpi->mb.pip = 0;
 
 #if CONFIG_MULTITHREAD
-    /* De-allocate mutex */
-    if (cpi->pmutex != NULL) {
-        VP8_COMMON *const pc = &cpi->common;
-        int i;
+  /* De-allocate mutex */
+  if (cpi->pmutex != NULL) {
+    VP8_COMMON *const pc = &cpi->common;
+    int i;
 
-        for (i = 0; i < pc->mb_rows; i++) {
-            pthread_mutex_destroy(&cpi->pmutex[i]);
-        }
-        vpx_free(cpi->pmutex);
-        cpi->pmutex = NULL;
+    for (i = 0; i < pc->mb_rows; i++) {
+      pthread_mutex_destroy(&cpi->pmutex[i]);
     }
+    vpx_free(cpi->pmutex);
+    cpi->pmutex = NULL;
+  }
 
-    vpx_free(cpi->mt_current_mb_col);
-    cpi->mt_current_mb_col = NULL;
+  vpx_free(cpi->mt_current_mb_col);
+  cpi->mt_current_mb_col = NULL;
 #endif
 }
 
-static void enable_segmentation(VP8_COMP *cpi)
-{
-    /* Set the appropriate feature bit */
-    cpi->mb.e_mbd.segmentation_enabled = 1;
-    cpi->mb.e_mbd.update_mb_segmentation_map = 1;
-    cpi->mb.e_mbd.update_mb_segmentation_data = 1;
+static void enable_segmentation(VP8_COMP *cpi) {
+  /* Set the appropriate feature bit */
+  cpi->mb.e_mbd.segmentation_enabled = 1;
+  cpi->mb.e_mbd.update_mb_segmentation_map = 1;
+  cpi->mb.e_mbd.update_mb_segmentation_data = 1;
 }
-static void disable_segmentation(VP8_COMP *cpi)
-{
-    /* Clear the appropriate feature bit */
-    cpi->mb.e_mbd.segmentation_enabled = 0;
+static void disable_segmentation(VP8_COMP *cpi) {
+  /* Clear the appropriate feature bit */
+  cpi->mb.e_mbd.segmentation_enabled = 0;
 }
 
 /* Valid values for a segment are 0 to 3
  * Segmentation map is arrange as [Rows][Columns]
  */
-static void set_segmentation_map(VP8_COMP *cpi, unsigned char *segmentation_map)
-{
-    /* Copy in the new segmentation map */
-    memcpy(cpi->segmentation_map, segmentation_map, (cpi->common.mb_rows * cpi->common.mb_cols));
+static void set_segmentation_map(VP8_COMP *cpi,
+                                 unsigned char *segmentation_map) {
+  /* Copy in the new segmentation map */
+  memcpy(cpi->segmentation_map, segmentation_map,
+         (cpi->common.mb_rows * cpi->common.mb_cols));
 
-    /* Signal that the map should be updated. */
-    cpi->mb.e_mbd.update_mb_segmentation_map = 1;
-    cpi->mb.e_mbd.update_mb_segmentation_data = 1;
+  /* Signal that the map should be updated. */
+  cpi->mb.e_mbd.update_mb_segmentation_map = 1;
+  cpi->mb.e_mbd.update_mb_segmentation_data = 1;
 }
 
 /* The values given for each segment can be either deltas (from the default
@@ -532,1305 +496,1156 @@
  * abs_delta = SEGMENT_ABSDATA (use the absolute values given).
  *
  */
-static void set_segment_data(VP8_COMP *cpi, signed char *feature_data, unsigned char abs_delta)
-{
-    cpi->mb.e_mbd.mb_segement_abs_delta = abs_delta;
-    memcpy(cpi->segment_feature_data, feature_data, sizeof(cpi->segment_feature_data));
+static void set_segment_data(VP8_COMP *cpi, signed char *feature_data,
+                             unsigned char abs_delta) {
+  cpi->mb.e_mbd.mb_segement_abs_delta = abs_delta;
+  memcpy(cpi->segment_feature_data, feature_data,
+         sizeof(cpi->segment_feature_data));
 }
 
-
 /* A simple function to cyclically refresh the background at a lower Q */
-static void cyclic_background_refresh(VP8_COMP *cpi, int Q, int lf_adjustment)
-{
-    unsigned char *seg_map = cpi->segmentation_map;
-    signed char feature_data[MB_LVL_MAX][MAX_MB_SEGMENTS];
-    int i;
-    int block_count = cpi->cyclic_refresh_mode_max_mbs_perframe;
-    int mbs_in_frame = cpi->common.mb_rows * cpi->common.mb_cols;
+static void cyclic_background_refresh(VP8_COMP *cpi, int Q, int lf_adjustment) {
+  unsigned char *seg_map = cpi->segmentation_map;
+  signed char feature_data[MB_LVL_MAX][MAX_MB_SEGMENTS];
+  int i;
+  int block_count = cpi->cyclic_refresh_mode_max_mbs_perframe;
+  int mbs_in_frame = cpi->common.mb_rows * cpi->common.mb_cols;
 
-    cpi->cyclic_refresh_q = Q / 2;
+  cpi->cyclic_refresh_q = Q / 2;
 
-    if (cpi->oxcf.screen_content_mode) {
-      // Modify quality ramp-up based on Q. Above some Q level, increase the
-      // number of blocks to be refreshed, and reduce it below the thredhold.
-      // Turn-off under certain conditions (i.e., away from key frame, and if
-      // we are at good quality (low Q) and most of the blocks were skipped-encoded
-      // in previous frame.
-      int qp_thresh = (cpi->oxcf.screen_content_mode == 2) ? 80 : 100;
-      if (Q >= qp_thresh) {
-        cpi->cyclic_refresh_mode_max_mbs_perframe =
-            (cpi->common.mb_rows * cpi->common.mb_cols) / 10;
-      } else if (cpi->frames_since_key > 250 &&
-                 Q < 20 &&
-                 cpi->mb.skip_true_count > (int)(0.95 * mbs_in_frame)) {
-        cpi->cyclic_refresh_mode_max_mbs_perframe = 0;
-      } else {
-        cpi->cyclic_refresh_mode_max_mbs_perframe =
-            (cpi->common.mb_rows * cpi->common.mb_cols) / 20;
-      }
-      block_count = cpi->cyclic_refresh_mode_max_mbs_perframe;
+  if (cpi->oxcf.screen_content_mode) {
+    // Modify quality ramp-up based on Q. Above some Q level, increase the
+    // number of blocks to be refreshed, and reduce it below the thredhold.
+    // Turn-off under certain conditions (i.e., away from key frame, and if
+    // we are at good quality (low Q) and most of the blocks were
+    // skipped-encoded
+    // in previous frame.
+    int qp_thresh = (cpi->oxcf.screen_content_mode == 2) ? 80 : 100;
+    if (Q >= qp_thresh) {
+      cpi->cyclic_refresh_mode_max_mbs_perframe =
+          (cpi->common.mb_rows * cpi->common.mb_cols) / 10;
+    } else if (cpi->frames_since_key > 250 && Q < 20 &&
+               cpi->mb.skip_true_count > (int)(0.95 * mbs_in_frame)) {
+      cpi->cyclic_refresh_mode_max_mbs_perframe = 0;
+    } else {
+      cpi->cyclic_refresh_mode_max_mbs_perframe =
+          (cpi->common.mb_rows * cpi->common.mb_cols) / 20;
     }
+    block_count = cpi->cyclic_refresh_mode_max_mbs_perframe;
+  }
 
-    // Set every macroblock to be eligible for update.
-    // For key frame this will reset seg map to 0.
-    memset(cpi->segmentation_map, 0, mbs_in_frame);
+  // Set every macroblock to be eligible for update.
+  // For key frame this will reset seg map to 0.
+  memset(cpi->segmentation_map, 0, mbs_in_frame);
 
-    if (cpi->common.frame_type != KEY_FRAME && block_count > 0)
-    {
-        /* Cycle through the macro_block rows */
-        /* MB loop to set local segmentation map */
-        i = cpi->cyclic_refresh_mode_index;
-        assert(i < mbs_in_frame);
-        do
-        {
-          /* If the MB is as a candidate for clean up then mark it for
-           * possible boost/refresh (segment 1) The segment id may get
-           * reset to 0 later if the MB gets coded anything other than
-           * last frame 0,0 as only (last frame 0,0) MBs are eligable for
-           * refresh : that is to say Mbs likely to be background blocks.
-           */
-          if (cpi->cyclic_refresh_map[i] == 0)
-          {
-              seg_map[i] = 1;
-              block_count --;
-          }
-          else if (cpi->cyclic_refresh_map[i] < 0)
-              cpi->cyclic_refresh_map[i]++;
+  if (cpi->common.frame_type != KEY_FRAME && block_count > 0) {
+    /* Cycle through the macro_block rows */
+    /* MB loop to set local segmentation map */
+    i = cpi->cyclic_refresh_mode_index;
+    assert(i < mbs_in_frame);
+    do {
+      /* If the MB is as a candidate for clean up then mark it for
+       * possible boost/refresh (segment 1) The segment id may get
+       * reset to 0 later if the MB gets coded anything other than
+       * last frame 0,0 as only (last frame 0,0) MBs are eligable for
+       * refresh : that is to say Mbs likely to be background blocks.
+       */
+      if (cpi->cyclic_refresh_map[i] == 0) {
+        seg_map[i] = 1;
+        block_count--;
+      } else if (cpi->cyclic_refresh_map[i] < 0)
+        cpi->cyclic_refresh_map[i]++;
 
-          i++;
-          if (i == mbs_in_frame)
-              i = 0;
+      i++;
+      if (i == mbs_in_frame) i = 0;
 
-        }
-        while(block_count && i != cpi->cyclic_refresh_mode_index);
+    } while (block_count && i != cpi->cyclic_refresh_mode_index);
 
-        cpi->cyclic_refresh_mode_index = i;
+    cpi->cyclic_refresh_mode_index = i;
 
 #if CONFIG_TEMPORAL_DENOISING
-        if (cpi->oxcf.noise_sensitivity > 0) {
-          if (cpi->denoiser.denoiser_mode == kDenoiserOnYUVAggressive &&
-              Q < (int)cpi->denoiser.denoise_pars.qp_thresh &&
-              (cpi->frames_since_key >
-               2 * cpi->denoiser.denoise_pars.consec_zerolast)) {
-            // Under aggressive denoising, use segmentation to turn off loop
-            // filter below some qp thresh. The filter is reduced for all
-            // blocks that have been encoded as ZEROMV LAST x frames in a row,
-            // where x is set by cpi->denoiser.denoise_pars.consec_zerolast.
-            // This is to avoid "dot" artifacts that can occur from repeated
-            // loop filtering on noisy input source.
-            cpi->cyclic_refresh_q = Q;
-            // lf_adjustment = -MAX_LOOP_FILTER;
-            lf_adjustment = -40;
-            for (i = 0; i < mbs_in_frame; ++i) {
-              seg_map[i] = (cpi->consec_zero_last[i] >
-                            cpi->denoiser.denoise_pars.consec_zerolast) ? 1 : 0;
-            }
-          }
+    if (cpi->oxcf.noise_sensitivity > 0) {
+      if (cpi->denoiser.denoiser_mode == kDenoiserOnYUVAggressive &&
+          Q < (int)cpi->denoiser.denoise_pars.qp_thresh &&
+          (cpi->frames_since_key >
+           2 * cpi->denoiser.denoise_pars.consec_zerolast)) {
+        // Under aggressive denoising, use segmentation to turn off loop
+        // filter below some qp thresh. The filter is reduced for all
+        // blocks that have been encoded as ZEROMV LAST x frames in a row,
+        // where x is set by cpi->denoiser.denoise_pars.consec_zerolast.
+        // This is to avoid "dot" artifacts that can occur from repeated
+        // loop filtering on noisy input source.
+        cpi->cyclic_refresh_q = Q;
+        // lf_adjustment = -MAX_LOOP_FILTER;
+        lf_adjustment = -40;
+        for (i = 0; i < mbs_in_frame; ++i) {
+          seg_map[i] = (cpi->consec_zero_last[i] >
+                        cpi->denoiser.denoise_pars.consec_zerolast)
+                           ? 1
+                           : 0;
         }
-#endif
+      }
     }
+#endif
+  }
 
-    /* Activate segmentation. */
-    cpi->mb.e_mbd.update_mb_segmentation_map = 1;
-    cpi->mb.e_mbd.update_mb_segmentation_data = 1;
-    enable_segmentation(cpi);
+  /* Activate segmentation. */
+  cpi->mb.e_mbd.update_mb_segmentation_map = 1;
+  cpi->mb.e_mbd.update_mb_segmentation_data = 1;
+  enable_segmentation(cpi);
 
-    /* Set up the quant segment data */
-    feature_data[MB_LVL_ALT_Q][0] = 0;
-    feature_data[MB_LVL_ALT_Q][1] = (cpi->cyclic_refresh_q - Q);
-    feature_data[MB_LVL_ALT_Q][2] = 0;
-    feature_data[MB_LVL_ALT_Q][3] = 0;
+  /* Set up the quant segment data */
+  feature_data[MB_LVL_ALT_Q][0] = 0;
+  feature_data[MB_LVL_ALT_Q][1] = (cpi->cyclic_refresh_q - Q);
+  feature_data[MB_LVL_ALT_Q][2] = 0;
+  feature_data[MB_LVL_ALT_Q][3] = 0;
 
-    /* Set up the loop segment data */
-    feature_data[MB_LVL_ALT_LF][0] = 0;
-    feature_data[MB_LVL_ALT_LF][1] = lf_adjustment;
-    feature_data[MB_LVL_ALT_LF][2] = 0;
-    feature_data[MB_LVL_ALT_LF][3] = 0;
+  /* Set up the loop segment data */
+  feature_data[MB_LVL_ALT_LF][0] = 0;
+  feature_data[MB_LVL_ALT_LF][1] = lf_adjustment;
+  feature_data[MB_LVL_ALT_LF][2] = 0;
+  feature_data[MB_LVL_ALT_LF][3] = 0;
 
-    /* Initialise the feature data structure */
-    set_segment_data(cpi, &feature_data[0][0], SEGMENT_DELTADATA);
-
+  /* Initialise the feature data structure */
+  set_segment_data(cpi, &feature_data[0][0], SEGMENT_DELTADATA);
 }
 
-static void set_default_lf_deltas(VP8_COMP *cpi)
-{
-    cpi->mb.e_mbd.mode_ref_lf_delta_enabled = 1;
-    cpi->mb.e_mbd.mode_ref_lf_delta_update = 1;
+static void set_default_lf_deltas(VP8_COMP *cpi) {
+  cpi->mb.e_mbd.mode_ref_lf_delta_enabled = 1;
+  cpi->mb.e_mbd.mode_ref_lf_delta_update = 1;
 
-    memset(cpi->mb.e_mbd.ref_lf_deltas, 0, sizeof(cpi->mb.e_mbd.ref_lf_deltas));
-    memset(cpi->mb.e_mbd.mode_lf_deltas, 0, sizeof(cpi->mb.e_mbd.mode_lf_deltas));
+  memset(cpi->mb.e_mbd.ref_lf_deltas, 0, sizeof(cpi->mb.e_mbd.ref_lf_deltas));
+  memset(cpi->mb.e_mbd.mode_lf_deltas, 0, sizeof(cpi->mb.e_mbd.mode_lf_deltas));
 
-    /* Test of ref frame deltas */
-    cpi->mb.e_mbd.ref_lf_deltas[INTRA_FRAME] = 2;
-    cpi->mb.e_mbd.ref_lf_deltas[LAST_FRAME] = 0;
-    cpi->mb.e_mbd.ref_lf_deltas[GOLDEN_FRAME] = -2;
-    cpi->mb.e_mbd.ref_lf_deltas[ALTREF_FRAME] = -2;
+  /* Test of ref frame deltas */
+  cpi->mb.e_mbd.ref_lf_deltas[INTRA_FRAME] = 2;
+  cpi->mb.e_mbd.ref_lf_deltas[LAST_FRAME] = 0;
+  cpi->mb.e_mbd.ref_lf_deltas[GOLDEN_FRAME] = -2;
+  cpi->mb.e_mbd.ref_lf_deltas[ALTREF_FRAME] = -2;
 
-    cpi->mb.e_mbd.mode_lf_deltas[0] = 4;               /* BPRED */
+  cpi->mb.e_mbd.mode_lf_deltas[0] = 4; /* BPRED */
 
-    if(cpi->oxcf.Mode == MODE_REALTIME)
-      cpi->mb.e_mbd.mode_lf_deltas[1] = -12;              /* Zero */
-    else
-      cpi->mb.e_mbd.mode_lf_deltas[1] = -2;              /* Zero */
+  if (cpi->oxcf.Mode == MODE_REALTIME)
+    cpi->mb.e_mbd.mode_lf_deltas[1] = -12; /* Zero */
+  else
+    cpi->mb.e_mbd.mode_lf_deltas[1] = -2; /* Zero */
 
-    cpi->mb.e_mbd.mode_lf_deltas[2] = 2;               /* New mv */
-    cpi->mb.e_mbd.mode_lf_deltas[3] = 4;               /* Split mv */
+  cpi->mb.e_mbd.mode_lf_deltas[2] = 2; /* New mv */
+  cpi->mb.e_mbd.mode_lf_deltas[3] = 4; /* Split mv */
 }
 
 /* Convenience macros for mapping speed and mode into a continuous
  * range
  */
-#define GOOD(x) (x+1)
-#define RT(x) (x+7)
+#define GOOD(x) (x + 1)
+#define RT(x) (x + 7)
 
-static int speed_map(int speed, const int *map)
-{
-    int res;
+static int speed_map(int speed, const int *map) {
+  int res;
 
-    do
-    {
-        res = *map++;
-    } while(speed >= *map++);
-    return res;
+  do {
+    res = *map++;
+  } while (speed >= *map++);
+  return res;
 }
 
 static const int thresh_mult_map_znn[] = {
-    /* map common to zero, nearest, and near */
-    0, GOOD(2), 1500, GOOD(3), 2000, RT(0), 1000, RT(2), 2000, INT_MAX
+  /* map common to zero, nearest, and near */
+  0, GOOD(2), 1500, GOOD(3), 2000, RT(0), 1000, RT(2), 2000, INT_MAX
 };
 
-static const int thresh_mult_map_vhpred[] = {
-    1000, GOOD(2), 1500, GOOD(3), 2000, RT(0), 1000, RT(1), 2000,
-    RT(7), INT_MAX, INT_MAX
-};
+static const int thresh_mult_map_vhpred[] = { 1000, GOOD(2), 1500,    GOOD(3),
+                                              2000, RT(0),   1000,    RT(1),
+                                              2000, RT(7),   INT_MAX, INT_MAX };
 
-static const int thresh_mult_map_bpred[] = {
-    2000, GOOD(0), 2500, GOOD(2), 5000, GOOD(3), 7500, RT(0), 2500, RT(1), 5000,
-    RT(6), INT_MAX, INT_MAX
-};
+static const int thresh_mult_map_bpred[] = { 2000,    GOOD(0), 2500, GOOD(2),
+                                             5000,    GOOD(3), 7500, RT(0),
+                                             2500,    RT(1),   5000, RT(6),
+                                             INT_MAX, INT_MAX };
 
-static const int thresh_mult_map_tm[] = {
-    1000, GOOD(2), 1500, GOOD(3), 2000, RT(0), 0, RT(1), 1000, RT(2), 2000,
-    RT(7), INT_MAX, INT_MAX
-};
+static const int thresh_mult_map_tm[] = { 1000,    GOOD(2), 1500, GOOD(3),
+                                          2000,    RT(0),   0,    RT(1),
+                                          1000,    RT(2),   2000, RT(7),
+                                          INT_MAX, INT_MAX };
 
-static const int thresh_mult_map_new1[] = {
-    1000, GOOD(2), 2000, RT(0), 2000, INT_MAX
-};
+static const int thresh_mult_map_new1[] = { 1000,  GOOD(2), 2000,
+                                            RT(0), 2000,    INT_MAX };
 
-static const int thresh_mult_map_new2[] = {
-    1000, GOOD(2), 2000, GOOD(3), 2500, GOOD(5), 4000, RT(0), 2000, RT(2), 2500,
-    RT(5), 4000, INT_MAX
-};
+static const int thresh_mult_map_new2[] = { 1000, GOOD(2), 2000, GOOD(3),
+                                            2500, GOOD(5), 4000, RT(0),
+                                            2000, RT(2),   2500, RT(5),
+                                            4000, INT_MAX };
 
 static const int thresh_mult_map_split1[] = {
-    2500, GOOD(0), 1700, GOOD(2), 10000, GOOD(3), 25000, GOOD(4), INT_MAX,
-    RT(0), 5000, RT(1), 10000, RT(2), 25000, RT(3), INT_MAX, INT_MAX
+  2500,  GOOD(0), 1700,  GOOD(2), 10000, GOOD(3), 25000, GOOD(4), INT_MAX,
+  RT(0), 5000,    RT(1), 10000,   RT(2), 25000,   RT(3), INT_MAX, INT_MAX
 };
 
 static const int thresh_mult_map_split2[] = {
-    5000, GOOD(0), 4500, GOOD(2), 20000, GOOD(3), 50000, GOOD(4), INT_MAX,
-    RT(0), 10000, RT(1), 20000, RT(2), 50000, RT(3), INT_MAX, INT_MAX
+  5000,  GOOD(0), 4500,  GOOD(2), 20000, GOOD(3), 50000, GOOD(4), INT_MAX,
+  RT(0), 10000,   RT(1), 20000,   RT(2), 50000,   RT(3), INT_MAX, INT_MAX
 };
 
 static const int mode_check_freq_map_zn2[] = {
-    /* {zero,nearest}{2,3} */
-    0, RT(10), 1<<1, RT(11), 1<<2, RT(12), 1<<3, INT_MAX
+  /* {zero,nearest}{2,3} */
+  0, RT(10), 1 << 1, RT(11), 1 << 2, RT(12), 1 << 3, INT_MAX
 };
 
 static const int mode_check_freq_map_vhbpred[] = {
-    0, GOOD(5), 2, RT(0), 0, RT(3), 2, RT(5), 4, INT_MAX
+  0, GOOD(5), 2, RT(0), 0, RT(3), 2, RT(5), 4, INT_MAX
 };
 
 static const int mode_check_freq_map_near2[] = {
-    0, GOOD(5), 2, RT(0), 0, RT(3), 2, RT(10), 1<<2, RT(11), 1<<3, RT(12), 1<<4,
-    INT_MAX
+  0,      GOOD(5), 2,      RT(0),  0,      RT(3),  2,
+  RT(10), 1 << 2,  RT(11), 1 << 3, RT(12), 1 << 4, INT_MAX
 };
 
 static const int mode_check_freq_map_new1[] = {
-    0, RT(10), 1<<1, RT(11), 1<<2, RT(12), 1<<3, INT_MAX
+  0, RT(10), 1 << 1, RT(11), 1 << 2, RT(12), 1 << 3, INT_MAX
 };
 
-static const int mode_check_freq_map_new2[] = {
-    0, GOOD(5), 4, RT(0), 0, RT(3), 4, RT(10), 1<<3, RT(11), 1<<4, RT(12), 1<<5,
-    INT_MAX
-};
+static const int mode_check_freq_map_new2[] = { 0,      GOOD(5), 4,      RT(0),
+                                                0,      RT(3),   4,      RT(10),
+                                                1 << 3, RT(11),  1 << 4, RT(12),
+                                                1 << 5, INT_MAX };
 
 static const int mode_check_freq_map_split1[] = {
-    0, GOOD(2), 2, GOOD(3), 7, RT(1), 2, RT(2), 7, INT_MAX
+  0, GOOD(2), 2, GOOD(3), 7, RT(1), 2, RT(2), 7, INT_MAX
 };
 
 static const int mode_check_freq_map_split2[] = {
-    0, GOOD(1), 2, GOOD(2), 4, GOOD(3), 15, RT(1), 4, RT(2), 15, INT_MAX
+  0, GOOD(1), 2, GOOD(2), 4, GOOD(3), 15, RT(1), 4, RT(2), 15, INT_MAX
 };
 
-void vp8_set_speed_features(VP8_COMP *cpi)
-{
-    SPEED_FEATURES *sf = &cpi->sf;
-    int Mode = cpi->compressor_speed;
-    int Speed = cpi->Speed;
-    int i;
-    VP8_COMMON *cm = &cpi->common;
-    int last_improved_quant = sf->improved_quant;
-    int ref_frames;
+void vp8_set_speed_features(VP8_COMP *cpi) {
+  SPEED_FEATURES *sf = &cpi->sf;
+  int Mode = cpi->compressor_speed;
+  int Speed = cpi->Speed;
+  int i;
+  VP8_COMMON *cm = &cpi->common;
+  int last_improved_quant = sf->improved_quant;
+  int ref_frames;
 
-    /* Initialise default mode frequency sampling variables */
-    for (i = 0; i < MAX_MODES; i ++)
-    {
-        cpi->mode_check_freq[i] = 0;
-    }
+  /* Initialise default mode frequency sampling variables */
+  for (i = 0; i < MAX_MODES; i++) {
+    cpi->mode_check_freq[i] = 0;
+  }
 
-    cpi->mb.mbs_tested_so_far = 0;
-    cpi->mb.mbs_zero_last_dot_suppress = 0;
+  cpi->mb.mbs_tested_so_far = 0;
+  cpi->mb.mbs_zero_last_dot_suppress = 0;
 
-    /* best quality defaults */
-    sf->RD = 1;
-    sf->search_method = NSTEP;
-    sf->improved_quant = 1;
-    sf->improved_dct = 1;
-    sf->auto_filter = 1;
-    sf->recode_loop = 1;
-    sf->quarter_pixel_search = 1;
-    sf->half_pixel_search = 1;
-    sf->iterative_sub_pixel = 1;
-    sf->optimize_coefficients = 1;
-    sf->use_fastquant_for_pick = 0;
-    sf->no_skip_block4x4_search = 1;
+  /* best quality defaults */
+  sf->RD = 1;
+  sf->search_method = NSTEP;
+  sf->improved_quant = 1;
+  sf->improved_dct = 1;
+  sf->auto_filter = 1;
+  sf->recode_loop = 1;
+  sf->quarter_pixel_search = 1;
+  sf->half_pixel_search = 1;
+  sf->iterative_sub_pixel = 1;
+  sf->optimize_coefficients = 1;
+  sf->use_fastquant_for_pick = 0;
+  sf->no_skip_block4x4_search = 1;
 
-    sf->first_step = 0;
-    sf->max_step_search_steps = MAX_MVSEARCH_STEPS;
-    sf->improved_mv_pred = 1;
+  sf->first_step = 0;
+  sf->max_step_search_steps = MAX_MVSEARCH_STEPS;
+  sf->improved_mv_pred = 1;
 
-    /* default thresholds to 0 */
-    for (i = 0; i < MAX_MODES; i++)
-        sf->thresh_mult[i] = 0;
+  /* default thresholds to 0 */
+  for (i = 0; i < MAX_MODES; i++) sf->thresh_mult[i] = 0;
 
-    /* Count enabled references */
-    ref_frames = 1;
-    if (cpi->ref_frame_flags & VP8_LAST_FRAME)
-        ref_frames++;
-    if (cpi->ref_frame_flags & VP8_GOLD_FRAME)
-        ref_frames++;
-    if (cpi->ref_frame_flags & VP8_ALTR_FRAME)
-        ref_frames++;
+  /* Count enabled references */
+  ref_frames = 1;
+  if (cpi->ref_frame_flags & VP8_LAST_FRAME) ref_frames++;
+  if (cpi->ref_frame_flags & VP8_GOLD_FRAME) ref_frames++;
+  if (cpi->ref_frame_flags & VP8_ALTR_FRAME) ref_frames++;
 
-    /* Convert speed to continuous range, with clamping */
-    if (Mode == 0)
-        Speed = 0;
-    else if (Mode == 2)
-        Speed = RT(Speed);
-    else
-    {
-        if (Speed > 5)
-            Speed = 5;
-        Speed = GOOD(Speed);
-    }
+  /* Convert speed to continuous range, with clamping */
+  if (Mode == 0)
+    Speed = 0;
+  else if (Mode == 2)
+    Speed = RT(Speed);
+  else {
+    if (Speed > 5) Speed = 5;
+    Speed = GOOD(Speed);
+  }
 
-    sf->thresh_mult[THR_ZERO1] =
-    sf->thresh_mult[THR_NEAREST1] =
-    sf->thresh_mult[THR_NEAR1] =
-    sf->thresh_mult[THR_DC] = 0; /* always */
+  sf->thresh_mult[THR_ZERO1] = sf->thresh_mult[THR_NEAREST1] =
+      sf->thresh_mult[THR_NEAR1] = sf->thresh_mult[THR_DC] = 0; /* always */
 
-    sf->thresh_mult[THR_ZERO2] =
-    sf->thresh_mult[THR_ZERO3] =
-    sf->thresh_mult[THR_NEAREST2] =
-    sf->thresh_mult[THR_NEAREST3] =
-    sf->thresh_mult[THR_NEAR2]  =
-    sf->thresh_mult[THR_NEAR3]  = speed_map(Speed, thresh_mult_map_znn);
+  sf->thresh_mult[THR_ZERO2] = sf->thresh_mult[THR_ZERO3] =
+      sf->thresh_mult[THR_NEAREST2] = sf->thresh_mult[THR_NEAREST3] =
+          sf->thresh_mult[THR_NEAR2] = sf->thresh_mult[THR_NEAR3] =
+              speed_map(Speed, thresh_mult_map_znn);
 
-    sf->thresh_mult[THR_V_PRED] =
-    sf->thresh_mult[THR_H_PRED] = speed_map(Speed, thresh_mult_map_vhpred);
-    sf->thresh_mult[THR_B_PRED] = speed_map(Speed, thresh_mult_map_bpred);
-    sf->thresh_mult[THR_TM]     = speed_map(Speed, thresh_mult_map_tm);
-    sf->thresh_mult[THR_NEW1]   = speed_map(Speed, thresh_mult_map_new1);
-    sf->thresh_mult[THR_NEW2]   =
-    sf->thresh_mult[THR_NEW3]   = speed_map(Speed, thresh_mult_map_new2);
-    sf->thresh_mult[THR_SPLIT1] = speed_map(Speed, thresh_mult_map_split1);
-    sf->thresh_mult[THR_SPLIT2] =
-    sf->thresh_mult[THR_SPLIT3] = speed_map(Speed, thresh_mult_map_split2);
+  sf->thresh_mult[THR_V_PRED] = sf->thresh_mult[THR_H_PRED] =
+      speed_map(Speed, thresh_mult_map_vhpred);
+  sf->thresh_mult[THR_B_PRED] = speed_map(Speed, thresh_mult_map_bpred);
+  sf->thresh_mult[THR_TM] = speed_map(Speed, thresh_mult_map_tm);
+  sf->thresh_mult[THR_NEW1] = speed_map(Speed, thresh_mult_map_new1);
+  sf->thresh_mult[THR_NEW2] = sf->thresh_mult[THR_NEW3] =
+      speed_map(Speed, thresh_mult_map_new2);
+  sf->thresh_mult[THR_SPLIT1] = speed_map(Speed, thresh_mult_map_split1);
+  sf->thresh_mult[THR_SPLIT2] = sf->thresh_mult[THR_SPLIT3] =
+      speed_map(Speed, thresh_mult_map_split2);
 
-    // Special case for temporal layers.
-    // Reduce the thresholds for zero/nearest/near for GOLDEN, if GOLDEN is
-    // used as second reference. We don't modify thresholds for ALTREF case
-    // since ALTREF is usually used as long-term reference in temporal layers.
-    if ((cpi->Speed <= 6) &&
-        (cpi->oxcf.number_of_layers > 1) &&
-        (cpi->ref_frame_flags & VP8_LAST_FRAME) &&
-        (cpi->ref_frame_flags & VP8_GOLD_FRAME)) {
-      if (cpi->closest_reference_frame == GOLDEN_FRAME) {
-        sf->thresh_mult[THR_ZERO2] =  sf->thresh_mult[THR_ZERO2] >> 3;
-        sf->thresh_mult[THR_NEAREST2] = sf->thresh_mult[THR_NEAREST2] >> 3;
-        sf->thresh_mult[THR_NEAR2]  = sf->thresh_mult[THR_NEAR2] >> 3;
-      } else {
-        sf->thresh_mult[THR_ZERO2] =  sf->thresh_mult[THR_ZERO2] >> 1;
-        sf->thresh_mult[THR_NEAREST2] = sf->thresh_mult[THR_NEAREST2] >> 1;
-        sf->thresh_mult[THR_NEAR2]  = sf->thresh_mult[THR_NEAR2] >> 1;
-      }
+  // Special case for temporal layers.
+  // Reduce the thresholds for zero/nearest/near for GOLDEN, if GOLDEN is
+  // used as second reference. We don't modify thresholds for ALTREF case
+  // since ALTREF is usually used as long-term reference in temporal layers.
+  if ((cpi->Speed <= 6) && (cpi->oxcf.number_of_layers > 1) &&
+      (cpi->ref_frame_flags & VP8_LAST_FRAME) &&
+      (cpi->ref_frame_flags & VP8_GOLD_FRAME)) {
+    if (cpi->closest_reference_frame == GOLDEN_FRAME) {
+      sf->thresh_mult[THR_ZERO2] = sf->thresh_mult[THR_ZERO2] >> 3;
+      sf->thresh_mult[THR_NEAREST2] = sf->thresh_mult[THR_NEAREST2] >> 3;
+      sf->thresh_mult[THR_NEAR2] = sf->thresh_mult[THR_NEAR2] >> 3;
+    } else {
+      sf->thresh_mult[THR_ZERO2] = sf->thresh_mult[THR_ZERO2] >> 1;
+      sf->thresh_mult[THR_NEAREST2] = sf->thresh_mult[THR_NEAREST2] >> 1;
+      sf->thresh_mult[THR_NEAR2] = sf->thresh_mult[THR_NEAR2] >> 1;
     }
+  }
 
-    cpi->mode_check_freq[THR_ZERO1] =
-    cpi->mode_check_freq[THR_NEAREST1] =
-    cpi->mode_check_freq[THR_NEAR1] =
-    cpi->mode_check_freq[THR_TM]     =
-    cpi->mode_check_freq[THR_DC] = 0; /* always */
+  cpi->mode_check_freq[THR_ZERO1] = cpi->mode_check_freq[THR_NEAREST1] =
+      cpi->mode_check_freq[THR_NEAR1] = cpi->mode_check_freq[THR_TM] =
+          cpi->mode_check_freq[THR_DC] = 0; /* always */
 
-    cpi->mode_check_freq[THR_ZERO2] =
-    cpi->mode_check_freq[THR_ZERO3] =
-    cpi->mode_check_freq[THR_NEAREST2] =
-    cpi->mode_check_freq[THR_NEAREST3] = speed_map(Speed,
-                                                   mode_check_freq_map_zn2);
+  cpi->mode_check_freq[THR_ZERO2] = cpi->mode_check_freq[THR_ZERO3] =
+      cpi->mode_check_freq[THR_NEAREST2] = cpi->mode_check_freq[THR_NEAREST3] =
+          speed_map(Speed, mode_check_freq_map_zn2);
 
-    cpi->mode_check_freq[THR_NEAR2]  =
-    cpi->mode_check_freq[THR_NEAR3]  = speed_map(Speed,
-                                                 mode_check_freq_map_near2);
+  cpi->mode_check_freq[THR_NEAR2] = cpi->mode_check_freq[THR_NEAR3] =
+      speed_map(Speed, mode_check_freq_map_near2);
 
-    cpi->mode_check_freq[THR_V_PRED] =
-    cpi->mode_check_freq[THR_H_PRED] =
-    cpi->mode_check_freq[THR_B_PRED] = speed_map(Speed,
-                                                 mode_check_freq_map_vhbpred);
-    cpi->mode_check_freq[THR_NEW1]   = speed_map(Speed,
-                                                 mode_check_freq_map_new1);
-    cpi->mode_check_freq[THR_NEW2]   =
-    cpi->mode_check_freq[THR_NEW3]   = speed_map(Speed,
-                                                 mode_check_freq_map_new2);
-    cpi->mode_check_freq[THR_SPLIT1] = speed_map(Speed,
-                                                 mode_check_freq_map_split1);
-    cpi->mode_check_freq[THR_SPLIT2] =
-    cpi->mode_check_freq[THR_SPLIT3] = speed_map(Speed,
-                                                 mode_check_freq_map_split2);
-    Speed = cpi->Speed;
-    switch (Mode)
-    {
+  cpi->mode_check_freq[THR_V_PRED] = cpi->mode_check_freq[THR_H_PRED] =
+      cpi->mode_check_freq[THR_B_PRED] =
+          speed_map(Speed, mode_check_freq_map_vhbpred);
+  cpi->mode_check_freq[THR_NEW1] = speed_map(Speed, mode_check_freq_map_new1);
+  cpi->mode_check_freq[THR_NEW2] = cpi->mode_check_freq[THR_NEW3] =
+      speed_map(Speed, mode_check_freq_map_new2);
+  cpi->mode_check_freq[THR_SPLIT1] =
+      speed_map(Speed, mode_check_freq_map_split1);
+  cpi->mode_check_freq[THR_SPLIT2] = cpi->mode_check_freq[THR_SPLIT3] =
+      speed_map(Speed, mode_check_freq_map_split2);
+  Speed = cpi->Speed;
+  switch (Mode) {
 #if !CONFIG_REALTIME_ONLY
     case 0: /* best quality mode */
-        sf->first_step = 0;
-        sf->max_step_search_steps = MAX_MVSEARCH_STEPS;
-        break;
+      sf->first_step = 0;
+      sf->max_step_search_steps = MAX_MVSEARCH_STEPS;
+      break;
     case 1:
     case 3:
-        if (Speed > 0)
-        {
-            /* Disable coefficient optimization above speed 0 */
-            sf->optimize_coefficients = 0;
-            sf->use_fastquant_for_pick = 1;
-            sf->no_skip_block4x4_search = 0;
+      if (Speed > 0) {
+        /* Disable coefficient optimization above speed 0 */
+        sf->optimize_coefficients = 0;
+        sf->use_fastquant_for_pick = 1;
+        sf->no_skip_block4x4_search = 0;
 
-            sf->first_step = 1;
-        }
+        sf->first_step = 1;
+      }
 
-        if (Speed > 2)
-        {
-            sf->improved_quant = 0;
-            sf->improved_dct = 0;
+      if (Speed > 2) {
+        sf->improved_quant = 0;
+        sf->improved_dct = 0;
 
-            /* Only do recode loop on key frames, golden frames and
-             * alt ref frames
-             */
-            sf->recode_loop = 2;
+        /* Only do recode loop on key frames, golden frames and
+         * alt ref frames
+         */
+        sf->recode_loop = 2;
+      }
 
-        }
+      if (Speed > 3) {
+        sf->auto_filter = 1;
+        sf->recode_loop = 0; /* recode loop off */
+        sf->RD = 0;          /* Turn rd off */
+      }
 
-        if (Speed > 3)
-        {
-            sf->auto_filter = 1;
-            sf->recode_loop = 0; /* recode loop off */
-            sf->RD = 0;         /* Turn rd off */
+      if (Speed > 4) {
+        sf->auto_filter = 0; /* Faster selection of loop filter */
+      }
 
-        }
-
-        if (Speed > 4)
-        {
-            sf->auto_filter = 0;  /* Faster selection of loop filter */
-        }
-
-        break;
+      break;
 #endif
     case 2:
-        sf->optimize_coefficients = 0;
-        sf->recode_loop = 0;
-        sf->auto_filter = 1;
-        sf->iterative_sub_pixel = 1;
-        sf->search_method = NSTEP;
+      sf->optimize_coefficients = 0;
+      sf->recode_loop = 0;
+      sf->auto_filter = 1;
+      sf->iterative_sub_pixel = 1;
+      sf->search_method = NSTEP;
 
-        if (Speed > 0)
-        {
-            sf->improved_quant = 0;
-            sf->improved_dct = 0;
+      if (Speed > 0) {
+        sf->improved_quant = 0;
+        sf->improved_dct = 0;
 
-            sf->use_fastquant_for_pick = 1;
-            sf->no_skip_block4x4_search = 0;
-            sf->first_step = 1;
-        }
+        sf->use_fastquant_for_pick = 1;
+        sf->no_skip_block4x4_search = 0;
+        sf->first_step = 1;
+      }
 
-        if (Speed > 2)
-            sf->auto_filter = 0;  /* Faster selection of loop filter */
+      if (Speed > 2) sf->auto_filter = 0; /* Faster selection of loop filter */
 
-        if (Speed > 3)
-        {
-            sf->RD = 0;
-            sf->auto_filter = 1;
-        }
+      if (Speed > 3) {
+        sf->RD = 0;
+        sf->auto_filter = 1;
+      }
 
-        if (Speed > 4)
-        {
-            sf->auto_filter = 0;  /* Faster selection of loop filter */
-            sf->search_method = HEX;
-            sf->iterative_sub_pixel = 0;
-        }
+      if (Speed > 4) {
+        sf->auto_filter = 0; /* Faster selection of loop filter */
+        sf->search_method = HEX;
+        sf->iterative_sub_pixel = 0;
+      }
 
-        if (Speed > 6)
-        {
-            unsigned int sum = 0;
-            unsigned int total_mbs = cm->MBs;
-            int thresh;
-            unsigned int total_skip;
+      if (Speed > 6) {
+        unsigned int sum = 0;
+        unsigned int total_mbs = cm->MBs;
+        int thresh;
+        unsigned int total_skip;
 
-            int min = 2000;
+        int min = 2000;
 
-            if (cpi->oxcf.encode_breakout > 2000)
-                min = cpi->oxcf.encode_breakout;
+        if (cpi->oxcf.encode_breakout > 2000) min = cpi->oxcf.encode_breakout;
 
-            min >>= 7;
+        min >>= 7;
 
-            for (i = 0; i < min; i++)
-            {
-                sum += cpi->mb.error_bins[i];
-            }
+        for (i = 0; i < min; i++) {
+          sum += cpi->mb.error_bins[i];
+        }
 
-            total_skip = sum;
-            sum = 0;
+        total_skip = sum;
+        sum = 0;
 
-            /* i starts from 2 to make sure thresh started from 2048 */
-            for (; i < 1024; i++)
-            {
-                sum += cpi->mb.error_bins[i];
+        /* i starts from 2 to make sure thresh started from 2048 */
+        for (; i < 1024; i++) {
+          sum += cpi->mb.error_bins[i];
 
-                if (10 * sum >= (unsigned int)(cpi->Speed - 6)*(total_mbs - total_skip))
-                    break;
-            }
+          if (10 * sum >=
+              (unsigned int)(cpi->Speed - 6) * (total_mbs - total_skip))
+            break;
+        }
 
-            i--;
-            thresh = (i << 7);
+        i--;
+        thresh = (i << 7);
 
-            if (thresh < 2000)
-                thresh = 2000;
+        if (thresh < 2000) thresh = 2000;
 
-            if (ref_frames > 1)
-            {
-                sf->thresh_mult[THR_NEW1 ] = thresh;
-                sf->thresh_mult[THR_NEAREST1  ] = thresh >> 1;
-                sf->thresh_mult[THR_NEAR1     ] = thresh >> 1;
-            }
+        if (ref_frames > 1) {
+          sf->thresh_mult[THR_NEW1] = thresh;
+          sf->thresh_mult[THR_NEAREST1] = thresh >> 1;
+          sf->thresh_mult[THR_NEAR1] = thresh >> 1;
+        }
 
-            if (ref_frames > 2)
-            {
-                sf->thresh_mult[THR_NEW2] = thresh << 1;
-                sf->thresh_mult[THR_NEAREST2 ] = thresh;
-                sf->thresh_mult[THR_NEAR2    ] = thresh;
-            }
+        if (ref_frames > 2) {
+          sf->thresh_mult[THR_NEW2] = thresh << 1;
+          sf->thresh_mult[THR_NEAREST2] = thresh;
+          sf->thresh_mult[THR_NEAR2] = thresh;
+        }
 
-            if (ref_frames > 3)
-            {
-                sf->thresh_mult[THR_NEW3] = thresh << 1;
-                sf->thresh_mult[THR_NEAREST3 ] = thresh;
-                sf->thresh_mult[THR_NEAR3    ] = thresh;
-            }
-
-            sf->improved_mv_pred = 0;
+        if (ref_frames > 3) {
+          sf->thresh_mult[THR_NEW3] = thresh << 1;
+          sf->thresh_mult[THR_NEAREST3] = thresh;
+          sf->thresh_mult[THR_NEAR3] = thresh;
         }
 
-        if (Speed > 8)
-            sf->quarter_pixel_search = 0;
+        sf->improved_mv_pred = 0;
+      }
 
-        if(cm->version == 0)
-        {
-            cm->filter_type = NORMAL_LOOPFILTER;
+      if (Speed > 8) sf->quarter_pixel_search = 0;
 
-            if (Speed >= 14)
-                cm->filter_type = SIMPLE_LOOPFILTER;
-        }
-        else
-        {
-            cm->filter_type = SIMPLE_LOOPFILTER;
-        }
+      if (cm->version == 0) {
+        cm->filter_type = NORMAL_LOOPFILTER;
 
-        /* This has a big hit on quality. Last resort */
-        if (Speed >= 15)
-            sf->half_pixel_search = 0;
+        if (Speed >= 14) cm->filter_type = SIMPLE_LOOPFILTER;
+      } else {
+        cm->filter_type = SIMPLE_LOOPFILTER;
+      }
 
-        memset(cpi->mb.error_bins, 0, sizeof(cpi->mb.error_bins));
+      /* This has a big hit on quality. Last resort */
+      if (Speed >= 15) sf->half_pixel_search = 0;
 
-    }; /* switch */
+      memset(cpi->mb.error_bins, 0, sizeof(cpi->mb.error_bins));
 
-    /* Slow quant, dct and trellis not worthwhile for first pass
-     * so make sure they are always turned off.
-     */
-    if ( cpi->pass == 1 )
-    {
-        sf->improved_quant = 0;
-        sf->optimize_coefficients = 0;
-        sf->improved_dct = 0;
-    }
+  }; /* switch */
 
-    if (cpi->sf.search_method == NSTEP)
-    {
-        vp8_init3smotion_compensation(&cpi->mb, cm->yv12_fb[cm->lst_fb_idx].y_stride);
-    }
-    else if (cpi->sf.search_method == DIAMOND)
-    {
-        vp8_init_dsmotion_compensation(&cpi->mb, cm->yv12_fb[cm->lst_fb_idx].y_stride);
-    }
+  /* Slow quant, dct and trellis not worthwhile for first pass
+   * so make sure they are always turned off.
+   */
+  if (cpi->pass == 1) {
+    sf->improved_quant = 0;
+    sf->optimize_coefficients = 0;
+    sf->improved_dct = 0;
+  }
 
-    if (cpi->sf.improved_dct)
-    {
-        cpi->mb.short_fdct8x4 = vp8_short_fdct8x4;
-        cpi->mb.short_fdct4x4 = vp8_short_fdct4x4;
-    }
-    else
-    {
-        /* No fast FDCT defined for any platform at this time. */
-        cpi->mb.short_fdct8x4 = vp8_short_fdct8x4;
-        cpi->mb.short_fdct4x4 = vp8_short_fdct4x4;
-    }
+  if (cpi->sf.search_method == NSTEP) {
+    vp8_init3smotion_compensation(&cpi->mb,
+                                  cm->yv12_fb[cm->lst_fb_idx].y_stride);
+  } else if (cpi->sf.search_method == DIAMOND) {
+    vp8_init_dsmotion_compensation(&cpi->mb,
+                                   cm->yv12_fb[cm->lst_fb_idx].y_stride);
+  }
 
-    cpi->mb.short_walsh4x4 = vp8_short_walsh4x4;
+  if (cpi->sf.improved_dct) {
+    cpi->mb.short_fdct8x4 = vp8_short_fdct8x4;
+    cpi->mb.short_fdct4x4 = vp8_short_fdct4x4;
+  } else {
+    /* No fast FDCT defined for any platform at this time. */
+    cpi->mb.short_fdct8x4 = vp8_short_fdct8x4;
+    cpi->mb.short_fdct4x4 = vp8_short_fdct4x4;
+  }
 
-    if (cpi->sf.improved_quant)
-    {
-        cpi->mb.quantize_b      = vp8_regular_quantize_b;
-    }
-    else
-    {
-        cpi->mb.quantize_b      = vp8_fast_quantize_b;
-    }
-    if (cpi->sf.improved_quant != last_improved_quant)
-        vp8cx_init_quantizer(cpi);
+  cpi->mb.short_walsh4x4 = vp8_short_walsh4x4;
 
-    if (cpi->sf.iterative_sub_pixel == 1)
-    {
-        cpi->find_fractional_mv_step = vp8_find_best_sub_pixel_step_iteratively;
-    }
-    else if (cpi->sf.quarter_pixel_search)
-    {
-        cpi->find_fractional_mv_step = vp8_find_best_sub_pixel_step;
-    }
-    else if (cpi->sf.half_pixel_search)
-    {
-        cpi->find_fractional_mv_step = vp8_find_best_half_pixel_step;
-    }
-    else
-    {
-        cpi->find_fractional_mv_step = vp8_skip_fractional_mv_step;
-    }
+  if (cpi->sf.improved_quant) {
+    cpi->mb.quantize_b = vp8_regular_quantize_b;
+  } else {
+    cpi->mb.quantize_b = vp8_fast_quantize_b;
+  }
+  if (cpi->sf.improved_quant != last_improved_quant) vp8cx_init_quantizer(cpi);
 
-    if (cpi->sf.optimize_coefficients == 1 && cpi->pass!=1)
-        cpi->mb.optimize = 1;
-    else
-        cpi->mb.optimize = 0;
+  if (cpi->sf.iterative_sub_pixel == 1) {
+    cpi->find_fractional_mv_step = vp8_find_best_sub_pixel_step_iteratively;
+  } else if (cpi->sf.quarter_pixel_search) {
+    cpi->find_fractional_mv_step = vp8_find_best_sub_pixel_step;
+  } else if (cpi->sf.half_pixel_search) {
+    cpi->find_fractional_mv_step = vp8_find_best_half_pixel_step;
+  } else {
+    cpi->find_fractional_mv_step = vp8_skip_fractional_mv_step;
+  }
 
-    if (cpi->common.full_pixel)
-        cpi->find_fractional_mv_step = vp8_skip_fractional_mv_step;
+  if (cpi->sf.optimize_coefficients == 1 && cpi->pass != 1)
+    cpi->mb.optimize = 1;
+  else
+    cpi->mb.optimize = 0;
 
+  if (cpi->common.full_pixel)
+    cpi->find_fractional_mv_step = vp8_skip_fractional_mv_step;
+
 #ifdef SPEEDSTATS
-    frames_at_speed[cpi->Speed]++;
+  frames_at_speed[cpi->Speed]++;
 #endif
 }
 #undef GOOD
 #undef RT
 
-static void alloc_raw_frame_buffers(VP8_COMP *cpi)
-{
+static void alloc_raw_frame_buffers(VP8_COMP *cpi) {
 #if VP8_TEMPORAL_ALT_REF
-    int width = (cpi->oxcf.Width + 15) & ~15;
-    int height = (cpi->oxcf.Height + 15) & ~15;
+  int width = (cpi->oxcf.Width + 15) & ~15;
+  int height = (cpi->oxcf.Height + 15) & ~15;
 #endif
 
-    cpi->lookahead = vp8_lookahead_init(cpi->oxcf.Width, cpi->oxcf.Height,
-                                        cpi->oxcf.lag_in_frames);
-    if(!cpi->lookahead)
-        vpx_internal_error(&cpi->common.error, VPX_CODEC_MEM_ERROR,
-                           "Failed to allocate lag buffers");
+  cpi->lookahead = vp8_lookahead_init(cpi->oxcf.Width, cpi->oxcf.Height,
+                                      cpi->oxcf.lag_in_frames);
+  if (!cpi->lookahead)
+    vpx_internal_error(&cpi->common.error, VPX_CODEC_MEM_ERROR,
+                       "Failed to allocate lag buffers");
 
 #if VP8_TEMPORAL_ALT_REF
 
-    if (vp8_yv12_alloc_frame_buffer(&cpi->alt_ref_buffer,
-                                    width, height, VP8BORDERINPIXELS))
-        vpx_internal_error(&cpi->common.error, VPX_CODEC_MEM_ERROR,
-                           "Failed to allocate altref buffer");
+  if (vp8_yv12_alloc_frame_buffer(&cpi->alt_ref_buffer, width, height,
+                                  VP8BORDERINPIXELS))
+    vpx_internal_error(&cpi->common.error, VPX_CODEC_MEM_ERROR,
+                       "Failed to allocate altref buffer");
 
 #endif
 }
 
-
-static void dealloc_raw_frame_buffers(VP8_COMP *cpi)
-{
+static void dealloc_raw_frame_buffers(VP8_COMP *cpi) {
 #if VP8_TEMPORAL_ALT_REF
-    vp8_yv12_de_alloc_frame_buffer(&cpi->alt_ref_buffer);
+  vp8_yv12_de_alloc_frame_buffer(&cpi->alt_ref_buffer);
 #endif
-    vp8_lookahead_destroy(cpi->lookahead);
+  vp8_lookahead_destroy(cpi->lookahead);
 }
 
+static int vp8_alloc_partition_data(VP8_COMP *cpi) {
+  vpx_free(cpi->mb.pip);
 
-static int vp8_alloc_partition_data(VP8_COMP *cpi)
-{
-        vpx_free(cpi->mb.pip);
+  cpi->mb.pip =
+      vpx_calloc((cpi->common.mb_cols + 1) * (cpi->common.mb_rows + 1),
+                 sizeof(PARTITION_INFO));
+  if (!cpi->mb.pip) return 1;
 
-    cpi->mb.pip = vpx_calloc((cpi->common.mb_cols + 1) *
-                                (cpi->common.mb_rows + 1),
-                                sizeof(PARTITION_INFO));
-    if(!cpi->mb.pip)
-        return 1;
+  cpi->mb.pi = cpi->mb.pip + cpi->common.mode_info_stride + 1;
 
-    cpi->mb.pi = cpi->mb.pip + cpi->common.mode_info_stride + 1;
-
-    return 0;
+  return 0;
 }
 
-void vp8_alloc_compressor_data(VP8_COMP *cpi)
-{
-    VP8_COMMON *cm = & cpi->common;
+void vp8_alloc_compressor_data(VP8_COMP *cpi) {
+  VP8_COMMON *cm = &cpi->common;
 
-    int width = cm->Width;
-    int height = cm->Height;
+  int width = cm->Width;
+  int height = cm->Height;
 #if CONFIG_MULTITHREAD
-    int prev_mb_rows = cm->mb_rows;
+  int prev_mb_rows = cm->mb_rows;
 #endif
 
-    if (vp8_alloc_frame_buffers(cm, width, height))
-        vpx_internal_error(&cpi->common.error, VPX_CODEC_MEM_ERROR,
-                           "Failed to allocate frame buffers");
+  if (vp8_alloc_frame_buffers(cm, width, height))
+    vpx_internal_error(&cpi->common.error, VPX_CODEC_MEM_ERROR,
+                       "Failed to allocate frame buffers");
 
-    if (vp8_alloc_partition_data(cpi))
-        vpx_internal_error(&cpi->common.error, VPX_CODEC_MEM_ERROR,
-                           "Failed to allocate partition data");
+  if (vp8_alloc_partition_data(cpi))
+    vpx_internal_error(&cpi->common.error, VPX_CODEC_MEM_ERROR,
+                       "Failed to allocate partition data");
 
+  if ((width & 0xf) != 0) width += 16 - (width & 0xf);
 
-    if ((width & 0xf) != 0)
-        width += 16 - (width & 0xf);
+  if ((height & 0xf) != 0) height += 16 - (height & 0xf);
 
-    if ((height & 0xf) != 0)
-        height += 16 - (height & 0xf);
+  if (vp8_yv12_alloc_frame_buffer(&cpi->pick_lf_lvl_frame, width, height,
+                                  VP8BORDERINPIXELS))
+    vpx_internal_error(&cpi->common.error, VPX_CODEC_MEM_ERROR,
+                       "Failed to allocate last frame buffer");
 
+  if (vp8_yv12_alloc_frame_buffer(&cpi->scaled_source, width, height,
+                                  VP8BORDERINPIXELS))
+    vpx_internal_error(&cpi->common.error, VPX_CODEC_MEM_ERROR,
+                       "Failed to allocate scaled source buffer");
 
-    if (vp8_yv12_alloc_frame_buffer(&cpi->pick_lf_lvl_frame,
-                                    width, height, VP8BORDERINPIXELS))
-        vpx_internal_error(&cpi->common.error, VPX_CODEC_MEM_ERROR,
-                           "Failed to allocate last frame buffer");
+  vpx_free(cpi->tok);
 
-    if (vp8_yv12_alloc_frame_buffer(&cpi->scaled_source,
-                                    width, height, VP8BORDERINPIXELS))
-        vpx_internal_error(&cpi->common.error, VPX_CODEC_MEM_ERROR,
-                           "Failed to allocate scaled source buffer");
-
-    vpx_free(cpi->tok);
-
-    {
+  {
 #if CONFIG_REALTIME_ONLY & CONFIG_ONTHEFLY_BITPACKING
-        unsigned int tokens = 8 * 24 * 16; /* one MB for each thread */
+    unsigned int tokens = 8 * 24 * 16; /* one MB for each thread */
 #else
-        unsigned int tokens = cm->mb_rows * cm->mb_cols * 24 * 16;
+    unsigned int tokens = cm->mb_rows * cm->mb_cols * 24 * 16;
 #endif
-        CHECK_MEM_ERROR(cpi->tok, vpx_calloc(tokens, sizeof(*cpi->tok)));
-    }
+    CHECK_MEM_ERROR(cpi->tok, vpx_calloc(tokens, sizeof(*cpi->tok)));
+  }
 
-    /* Data used for real time vc mode to see if gf needs refreshing */
-    cpi->zeromv_count = 0;
+  /* Data used for real time vc mode to see if gf needs refreshing */
+  cpi->zeromv_count = 0;
 
+  /* Structures used to monitor GF usage */
+  vpx_free(cpi->gf_active_flags);
+  CHECK_MEM_ERROR(
+      cpi->gf_active_flags,
+      vpx_calloc(sizeof(*cpi->gf_active_flags), cm->mb_rows * cm->mb_cols));
+  cpi->gf_active_count = cm->mb_rows * cm->mb_cols;
 
-    /* Structures used to monitor GF usage */
-    vpx_free(cpi->gf_active_flags);
-    CHECK_MEM_ERROR(cpi->gf_active_flags,
-                    vpx_calloc(sizeof(*cpi->gf_active_flags),
-                    cm->mb_rows * cm->mb_cols));
-    cpi->gf_active_count = cm->mb_rows * cm->mb_cols;
+  vpx_free(cpi->mb_activity_map);
+  CHECK_MEM_ERROR(
+      cpi->mb_activity_map,
+      vpx_calloc(sizeof(*cpi->mb_activity_map), cm->mb_rows * cm->mb_cols));
 
-    vpx_free(cpi->mb_activity_map);
-    CHECK_MEM_ERROR(cpi->mb_activity_map,
-                    vpx_calloc(sizeof(*cpi->mb_activity_map),
-                    cm->mb_rows * cm->mb_cols));
+  /* allocate memory for storing last frame's MVs for MV prediction. */
+  vpx_free(cpi->lfmv);
+  CHECK_MEM_ERROR(cpi->lfmv, vpx_calloc((cm->mb_rows + 2) * (cm->mb_cols + 2),
+                                        sizeof(*cpi->lfmv)));
+  vpx_free(cpi->lf_ref_frame_sign_bias);
+  CHECK_MEM_ERROR(cpi->lf_ref_frame_sign_bias,
+                  vpx_calloc((cm->mb_rows + 2) * (cm->mb_cols + 2),
+                             sizeof(*cpi->lf_ref_frame_sign_bias)));
+  vpx_free(cpi->lf_ref_frame);
+  CHECK_MEM_ERROR(cpi->lf_ref_frame,
+                  vpx_calloc((cm->mb_rows + 2) * (cm->mb_cols + 2),
+                             sizeof(*cpi->lf_ref_frame)));
 
-    /* allocate memory for storing last frame's MVs for MV prediction. */
-    vpx_free(cpi->lfmv);
-    CHECK_MEM_ERROR(cpi->lfmv, vpx_calloc((cm->mb_rows+2) * (cm->mb_cols+2),
-                    sizeof(*cpi->lfmv)));
-    vpx_free(cpi->lf_ref_frame_sign_bias);
-    CHECK_MEM_ERROR(cpi->lf_ref_frame_sign_bias,
-                    vpx_calloc((cm->mb_rows+2) * (cm->mb_cols+2),
-                    sizeof(*cpi->lf_ref_frame_sign_bias)));
-    vpx_free(cpi->lf_ref_frame);
-    CHECK_MEM_ERROR(cpi->lf_ref_frame,
-                    vpx_calloc((cm->mb_rows+2) * (cm->mb_cols+2),
-                    sizeof(*cpi->lf_ref_frame)));
+  /* Create the encoder segmentation map and set all entries to 0 */
+  vpx_free(cpi->segmentation_map);
+  CHECK_MEM_ERROR(
+      cpi->segmentation_map,
+      vpx_calloc(cm->mb_rows * cm->mb_cols, sizeof(*cpi->segmentation_map)));
+  cpi->cyclic_refresh_mode_index = 0;
+  vpx_free(cpi->active_map);
+  CHECK_MEM_ERROR(cpi->active_map, vpx_calloc(cm->mb_rows * cm->mb_cols,
+                                              sizeof(*cpi->active_map)));
+  memset(cpi->active_map, 1, (cm->mb_rows * cm->mb_cols));
 
-    /* Create the encoder segmentation map and set all entries to 0 */
-    vpx_free(cpi->segmentation_map);
-    CHECK_MEM_ERROR(cpi->segmentation_map,
-                    vpx_calloc(cm->mb_rows * cm->mb_cols,
-                    sizeof(*cpi->segmentation_map)));
-    cpi->cyclic_refresh_mode_index = 0;
-    vpx_free(cpi->active_map);
-    CHECK_MEM_ERROR(cpi->active_map,
-                    vpx_calloc(cm->mb_rows * cm->mb_cols,
-                    sizeof(*cpi->active_map)));
-    memset(cpi->active_map , 1, (cm->mb_rows * cm->mb_cols));
-
 #if CONFIG_MULTITHREAD
-    if (width < 640)
-        cpi->mt_sync_range = 1;
-    else if (width <= 1280)
-        cpi->mt_sync_range = 4;
-    else if (width <= 2560)
-        cpi->mt_sync_range = 8;
-    else
-        cpi->mt_sync_range = 16;
+  if (width < 640)
+    cpi->mt_sync_range = 1;
+  else if (width <= 1280)
+    cpi->mt_sync_range = 4;
+  else if (width <= 2560)
+    cpi->mt_sync_range = 8;
+  else
+    cpi->mt_sync_range = 16;
 
-    if (cpi->oxcf.multi_threaded > 1)
-    {
-        int i;
+  if (cpi->oxcf.multi_threaded > 1) {
+    int i;
 
-        /* De-allocate and re-allocate mutex */
-        if (cpi->pmutex != NULL) {
-            for (i = 0; i < prev_mb_rows; i++) {
-                pthread_mutex_destroy(&cpi->pmutex[i]);
-            }
-            vpx_free(cpi->pmutex);
-            cpi->pmutex = NULL;
-        }
+    /* De-allocate and re-allocate mutex */
+    if (cpi->pmutex != NULL) {
+      for (i = 0; i < prev_mb_rows; i++) {
+        pthread_mutex_destroy(&cpi->pmutex[i]);
+      }
+      vpx_free(cpi->pmutex);
+      cpi->pmutex = NULL;
+    }
 
-        CHECK_MEM_ERROR(cpi->pmutex, vpx_malloc(sizeof(*cpi->pmutex) *
-                                                cm->mb_rows));
-        if (cpi->pmutex) {
-            for (i = 0; i < cm->mb_rows; i++) {
-                pthread_mutex_init(&cpi->pmutex[i], NULL);
-            }
-        }
+    CHECK_MEM_ERROR(cpi->pmutex,
+                    vpx_malloc(sizeof(*cpi->pmutex) * cm->mb_rows));
+    if (cpi->pmutex) {
+      for (i = 0; i < cm->mb_rows; i++) {
+        pthread_mutex_init(&cpi->pmutex[i], NULL);
+      }
+    }
 
-        vpx_free(cpi->mt_current_mb_col);
-        CHECK_MEM_ERROR(cpi->mt_current_mb_col,
+    vpx_free(cpi->mt_current_mb_col);
+    CHECK_MEM_ERROR(cpi->mt_current_mb_col,
                     vpx_malloc(sizeof(*cpi->mt_current_mb_col) * cm->mb_rows));
-    }
+  }
 
 #endif
 
-    vpx_free(cpi->tplist);
-    CHECK_MEM_ERROR(cpi->tplist, vpx_malloc(sizeof(TOKENLIST) * cm->mb_rows));
+  vpx_free(cpi->tplist);
+  CHECK_MEM_ERROR(cpi->tplist, vpx_malloc(sizeof(TOKENLIST) * cm->mb_rows));
 
 #if CONFIG_TEMPORAL_DENOISING
-    if (cpi->oxcf.noise_sensitivity > 0) {
-      vp8_denoiser_free(&cpi->denoiser);
-      if (vp8_denoiser_allocate(&cpi->denoiser, width, height,
-                                cm->mb_rows, cm->mb_cols,
-                                cpi->oxcf.noise_sensitivity))
-          vpx_internal_error(&cpi->common.error, VPX_CODEC_MEM_ERROR,
-                             "Failed to allocate denoiser");
-    }
+  if (cpi->oxcf.noise_sensitivity > 0) {
+    vp8_denoiser_free(&cpi->denoiser);
+    if (vp8_denoiser_allocate(&cpi->denoiser, width, height, cm->mb_rows,
+                              cm->mb_cols, cpi->oxcf.noise_sensitivity))
+      vpx_internal_error(&cpi->common.error, VPX_CODEC_MEM_ERROR,
+                         "Failed to allocate denoiser");
+  }
 #endif
 }
 
-
 /* Quant MOD */
-static const int q_trans[] =
-{
-    0,   1,  2,  3,  4,  5,  7,  8,
-    9,  10, 12, 13, 15, 17, 18, 19,
-    20,  21, 23, 24, 25, 26, 27, 28,
-    29,  30, 31, 33, 35, 37, 39, 41,
-    43,  45, 47, 49, 51, 53, 55, 57,
-    59,  61, 64, 67, 70, 73, 76, 79,
-    82,  85, 88, 91, 94, 97, 100, 103,
-    106, 109, 112, 115, 118, 121, 124, 127,
+static const int q_trans[] = {
+  0,  1,  2,  3,  4,  5,  7,   8,   9,   10,  12,  13,  15,  17,  18,  19,
+  20, 21, 23, 24, 25, 26, 27,  28,  29,  30,  31,  33,  35,  37,  39,  41,
+  43, 45, 47, 49, 51, 53, 55,  57,  59,  61,  64,  67,  70,  73,  76,  79,
+  82, 85, 88, 91, 94, 97, 100, 103, 106, 109, 112, 115, 118, 121, 124, 127,
 };
 
-int vp8_reverse_trans(int x)
-{
-    int i;
+int vp8_reverse_trans(int x) {
+  int i;
 
-    for (i = 0; i < 64; i++)
-        if (q_trans[i] >= x)
-            return i;
+  for (i = 0; i < 64; i++)
+    if (q_trans[i] >= x) return i;
 
-    return 63;
+  return 63;
 }
-void vp8_new_framerate(VP8_COMP *cpi, double framerate)
-{
-    if(framerate < .1)
-        framerate = 30;
+void vp8_new_framerate(VP8_COMP *cpi, double framerate) {
+  if (framerate < .1) framerate = 30;
 
-    cpi->framerate              = framerate;
-    cpi->output_framerate       = framerate;
-    cpi->per_frame_bandwidth    = (int)(cpi->oxcf.target_bandwidth /
-                                  cpi->output_framerate);
-    cpi->av_per_frame_bandwidth = cpi->per_frame_bandwidth;
-    cpi->min_frame_bandwidth    = (int)(cpi->av_per_frame_bandwidth *
-                                  cpi->oxcf.two_pass_vbrmin_section / 100);
+  cpi->framerate = framerate;
+  cpi->output_framerate = framerate;
+  cpi->per_frame_bandwidth =
+      (int)(cpi->oxcf.target_bandwidth / cpi->output_framerate);
+  cpi->av_per_frame_bandwidth = cpi->per_frame_bandwidth;
+  cpi->min_frame_bandwidth = (int)(cpi->av_per_frame_bandwidth *
+                                   cpi->oxcf.two_pass_vbrmin_section / 100);
 
-    /* Set Maximum gf/arf interval */
-    cpi->max_gf_interval = ((int)(cpi->output_framerate / 2.0) + 2);
+  /* Set Maximum gf/arf interval */
+  cpi->max_gf_interval = ((int)(cpi->output_framerate / 2.0) + 2);
 
-    if(cpi->max_gf_interval < 12)
-        cpi->max_gf_interval = 12;
+  if (cpi->max_gf_interval < 12) cpi->max_gf_interval = 12;
 
-    /* Extended interval for genuinely static scenes */
-    cpi->twopass.static_scene_max_gf_interval = cpi->key_frame_frequency >> 1;
+  /* Extended interval for genuinely static scenes */
+  cpi->twopass.static_scene_max_gf_interval = cpi->key_frame_frequency >> 1;
 
-     /* Special conditions when altr ref frame enabled in lagged compress mode */
-    if (cpi->oxcf.play_alternate && cpi->oxcf.lag_in_frames)
-    {
-        if (cpi->max_gf_interval > cpi->oxcf.lag_in_frames - 1)
-            cpi->max_gf_interval = cpi->oxcf.lag_in_frames - 1;
+  /* Special conditions when altr ref frame enabled in lagged compress mode */
+  if (cpi->oxcf.play_alternate && cpi->oxcf.lag_in_frames) {
+    if (cpi->max_gf_interval > cpi->oxcf.lag_in_frames - 1)
+      cpi->max_gf_interval = cpi->oxcf.lag_in_frames - 1;
 
-        if (cpi->twopass.static_scene_max_gf_interval > cpi->oxcf.lag_in_frames - 1)
-            cpi->twopass.static_scene_max_gf_interval = cpi->oxcf.lag_in_frames - 1;
-    }
+    if (cpi->twopass.static_scene_max_gf_interval > cpi->oxcf.lag_in_frames - 1)
+      cpi->twopass.static_scene_max_gf_interval = cpi->oxcf.lag_in_frames - 1;
+  }
 
-    if ( cpi->max_gf_interval > cpi->twopass.static_scene_max_gf_interval )
-        cpi->max_gf_interval = cpi->twopass.static_scene_max_gf_interval;
+  if (cpi->max_gf_interval > cpi->twopass.static_scene_max_gf_interval)
+    cpi->max_gf_interval = cpi->twopass.static_scene_max_gf_interval;
 }
 
+static void init_config(VP8_COMP *cpi, VP8_CONFIG *oxcf) {
+  VP8_COMMON *cm = &cpi->common;
 
-static void init_config(VP8_COMP *cpi, VP8_CONFIG *oxcf)
-{
-    VP8_COMMON *cm = &cpi->common;
+  cpi->oxcf = *oxcf;
 
-    cpi->oxcf = *oxcf;
+  cpi->auto_gold = 1;
+  cpi->auto_adjust_gold_quantizer = 1;
 
-    cpi->auto_gold = 1;
-    cpi->auto_adjust_gold_quantizer = 1;
+  cm->version = oxcf->Version;
+  vp8_setup_version(cm);
 
-    cm->version = oxcf->Version;
-    vp8_setup_version(cm);
+  /* Frame rate is not available on the first frame, as it's derived from
+   * the observed timestamps. The actual value used here doesn't matter
+   * too much, as it will adapt quickly.
+   */
+  if (oxcf->timebase.num > 0) {
+    cpi->framerate =
+        (double)(oxcf->timebase.den) / (double)(oxcf->timebase.num);
+  } else {
+    cpi->framerate = 30;
+  }
 
-    /* Frame rate is not available on the first frame, as it's derived from
-     * the observed timestamps. The actual value used here doesn't matter
-     * too much, as it will adapt quickly.
-     */
-    if (oxcf->timebase.num > 0) {
-      cpi->framerate = (double)(oxcf->timebase.den) /
-                       (double)(oxcf->timebase.num);
-    } else {
-      cpi->framerate = 30;
-    }
+  /* If the reciprocal of the timebase seems like a reasonable framerate,
+   * then use that as a guess, otherwise use 30.
+   */
+  if (cpi->framerate > 180) cpi->framerate = 30;
 
-    /* If the reciprocal of the timebase seems like a reasonable framerate,
-     * then use that as a guess, otherwise use 30.
-     */
-    if (cpi->framerate > 180)
-        cpi->framerate = 30;
+  cpi->ref_framerate = cpi->framerate;
 
-    cpi->ref_framerate = cpi->framerate;
+  cpi->ref_frame_flags = VP8_ALTR_FRAME | VP8_GOLD_FRAME | VP8_LAST_FRAME;
 
-    cpi->ref_frame_flags = VP8_ALTR_FRAME | VP8_GOLD_FRAME | VP8_LAST_FRAME;
+  cm->refresh_golden_frame = 0;
+  cm->refresh_last_frame = 1;
+  cm->refresh_entropy_probs = 1;
 
-    cm->refresh_golden_frame = 0;
-    cm->refresh_last_frame = 1;
-    cm->refresh_entropy_probs = 1;
+  /* change includes all joint functionality */
+  vp8_change_config(cpi, oxcf);
 
-    /* change includes all joint functionality */
-    vp8_change_config(cpi, oxcf);
+  /* Initialize active best and worst q and average q values. */
+  cpi->active_worst_quality = cpi->oxcf.worst_allowed_q;
+  cpi->active_best_quality = cpi->oxcf.best_allowed_q;
+  cpi->avg_frame_qindex = cpi->oxcf.worst_allowed_q;
 
-    /* Initialize active best and worst q and average q values. */
-    cpi->active_worst_quality         = cpi->oxcf.worst_allowed_q;
-    cpi->active_best_quality          = cpi->oxcf.best_allowed_q;
-    cpi->avg_frame_qindex             = cpi->oxcf.worst_allowed_q;
+  /* Initialise the starting buffer levels */
+  cpi->buffer_level = cpi->oxcf.starting_buffer_level;
+  cpi->bits_off_target = cpi->oxcf.starting_buffer_level;
 
-    /* Initialise the starting buffer levels */
-    cpi->buffer_level                 = cpi->oxcf.starting_buffer_level;
-    cpi->bits_off_target              = cpi->oxcf.starting_buffer_level;
+  cpi->rolling_target_bits = cpi->av_per_frame_bandwidth;
+  cpi->rolling_actual_bits = cpi->av_per_frame_bandwidth;
+  cpi->long_rolling_target_bits = cpi->av_per_frame_bandwidth;
+  cpi->long_rolling_actual_bits = cpi->av_per_frame_bandwidth;
 
-    cpi->rolling_target_bits          = cpi->av_per_frame_bandwidth;
-    cpi->rolling_actual_bits          = cpi->av_per_frame_bandwidth;
-    cpi->long_rolling_target_bits     = cpi->av_per_frame_bandwidth;
-    cpi->long_rolling_actual_bits     = cpi->av_per_frame_bandwidth;
+  cpi->total_actual_bits = 0;
+  cpi->total_target_vs_actual = 0;
 
-    cpi->total_actual_bits            = 0;
-    cpi->total_target_vs_actual       = 0;
+  /* Temporal scalabilty */
+  if (cpi->oxcf.number_of_layers > 1) {
+    unsigned int i;
+    double prev_layer_framerate = 0;
 
-    /* Temporal scalabilty */
-    if (cpi->oxcf.number_of_layers > 1)
-    {
-        unsigned int i;
-        double prev_layer_framerate=0;
-
-        for (i=0; i<cpi->oxcf.number_of_layers; i++)
-        {
-            init_temporal_layer_context(cpi, oxcf, i, prev_layer_framerate);
-            prev_layer_framerate = cpi->output_framerate /
-                                   cpi->oxcf.rate_decimator[i];
-        }
+    for (i = 0; i < cpi->oxcf.number_of_layers; i++) {
+      init_temporal_layer_context(cpi, oxcf, i, prev_layer_framerate);
+      prev_layer_framerate =
+          cpi->output_framerate / cpi->oxcf.rate_decimator[i];
     }
+  }
 
 #if VP8_TEMPORAL_ALT_REF
-    {
-        int i;
+  {
+    int i;
 
-        cpi->fixed_divide[0] = 0;
+    cpi->fixed_divide[0] = 0;
 
-        for (i = 1; i < 512; i++)
-            cpi->fixed_divide[i] = 0x80000 / i;
-    }
+    for (i = 1; i < 512; i++) cpi->fixed_divide[i] = 0x80000 / i;
+  }
 #endif
 }
 
-static void update_layer_contexts (VP8_COMP *cpi)
-{
-    VP8_CONFIG *oxcf = &cpi->oxcf;
+static void update_layer_contexts(VP8_COMP *cpi) {
+  VP8_CONFIG *oxcf = &cpi->oxcf;
 
-    /* Update snapshots of the layer contexts to reflect new parameters */
-    if (oxcf->number_of_layers > 1)
-    {
-        unsigned int i;
-        double prev_layer_framerate=0;
+  /* Update snapshots of the layer contexts to reflect new parameters */
+  if (oxcf->number_of_layers > 1) {
+    unsigned int i;
+    double prev_layer_framerate = 0;
 
-        assert(oxcf->number_of_layers <= VPX_TS_MAX_LAYERS);
-        for (i = 0; i < oxcf->number_of_layers && i < VPX_TS_MAX_LAYERS; ++i)
-        {
-            LAYER_CONTEXT *lc = &cpi->layer_context[i];
+    assert(oxcf->number_of_layers <= VPX_TS_MAX_LAYERS);
+    for (i = 0; i < oxcf->number_of_layers && i < VPX_TS_MAX_LAYERS; ++i) {
+      LAYER_CONTEXT *lc = &cpi->layer_context[i];
 
-            lc->framerate =
-                cpi->ref_framerate / oxcf->rate_decimator[i];
-            lc->target_bandwidth = oxcf->target_bitrate[i] * 1000;
+      lc->framerate = cpi->ref_framerate / oxcf->rate_decimator[i];
+      lc->target_bandwidth = oxcf->target_bitrate[i] * 1000;
 
-            lc->starting_buffer_level = rescale(
-                          (int)oxcf->starting_buffer_level_in_ms,
-                          lc->target_bandwidth, 1000);
+      lc->starting_buffer_level = rescale(
+          (int)oxcf->starting_buffer_level_in_ms, lc->target_bandwidth, 1000);
 
-            if (oxcf->optimal_buffer_level == 0)
-                lc->optimal_buffer_level = lc->target_bandwidth / 8;
-            else
-                lc->optimal_buffer_level = rescale(
-                          (int)oxcf->optimal_buffer_level_in_ms,
-                          lc->target_bandwidth, 1000);
+      if (oxcf->optimal_buffer_level == 0)
+        lc->optimal_buffer_level = lc->target_bandwidth / 8;
+      else
+        lc->optimal_buffer_level = rescale(
+            (int)oxcf->optimal_buffer_level_in_ms, lc->target_bandwidth, 1000);
 
-            if (oxcf->maximum_buffer_size == 0)
-                lc->maximum_buffer_size = lc->target_bandwidth / 8;
-            else
-                lc->maximum_buffer_size = rescale(
-                          (int)oxcf->maximum_buffer_size_in_ms,
-                          lc->target_bandwidth, 1000);
+      if (oxcf->maximum_buffer_size == 0)
+        lc->maximum_buffer_size = lc->target_bandwidth / 8;
+      else
+        lc->maximum_buffer_size = rescale((int)oxcf->maximum_buffer_size_in_ms,
+                                          lc->target_bandwidth, 1000);
 
-            /* Work out the average size of a frame within this layer */
-            if (i > 0)
-                lc->avg_frame_size_for_layer =
-                   (int)((oxcf->target_bitrate[i] -
-                          oxcf->target_bitrate[i-1]) * 1000 /
-                          (lc->framerate - prev_layer_framerate));
+      /* Work out the average size of a frame within this layer */
+      if (i > 0)
+        lc->avg_frame_size_for_layer =
+            (int)((oxcf->target_bitrate[i] - oxcf->target_bitrate[i - 1]) *
+                  1000 / (lc->framerate - prev_layer_framerate));
 
-            prev_layer_framerate = lc->framerate;
-        }
+      prev_layer_framerate = lc->framerate;
     }
+  }
 }
 
-void vp8_change_config(VP8_COMP *cpi, VP8_CONFIG *oxcf)
-{
-    VP8_COMMON *cm = &cpi->common;
-    int last_w, last_h;
-    unsigned int prev_number_of_layers;
+void vp8_change_config(VP8_COMP *cpi, VP8_CONFIG *oxcf) {
+  VP8_COMMON *cm = &cpi->common;
+  int last_w, last_h;
+  unsigned int prev_number_of_layers;
 
-    if (!cpi)
-        return;
+  if (!cpi) return;
 
-    if (!oxcf)
-        return;
+  if (!oxcf) return;
 
-    if (cm->version != oxcf->Version)
-    {
-        cm->version = oxcf->Version;
-        vp8_setup_version(cm);
-    }
+  if (cm->version != oxcf->Version) {
+    cm->version = oxcf->Version;
+    vp8_setup_version(cm);
+  }
 
-    last_w = cpi->oxcf.Width;
-    last_h = cpi->oxcf.Height;
-    prev_number_of_layers = cpi->oxcf.number_of_layers;
+  last_w = cpi->oxcf.Width;
+  last_h = cpi->oxcf.Height;
+  prev_number_of_layers = cpi->oxcf.number_of_layers;
 
-    cpi->oxcf = *oxcf;
+  cpi->oxcf = *oxcf;
 
-    switch (cpi->oxcf.Mode)
-    {
-
+  switch (cpi->oxcf.Mode) {
     case MODE_REALTIME:
-        cpi->pass = 0;
-        cpi->compressor_speed = 2;
+      cpi->pass = 0;
+      cpi->compressor_speed = 2;
 
-        if (cpi->oxcf.cpu_used < -16)
-        {
-            cpi->oxcf.cpu_used = -16;
-        }
+      if (cpi->oxcf.cpu_used < -16) {
+        cpi->oxcf.cpu_used = -16;
+      }
 
-        if (cpi->oxcf.cpu_used > 16)
-            cpi->oxcf.cpu_used = 16;
+      if (cpi->oxcf.cpu_used > 16) cpi->oxcf.cpu_used = 16;
 
-        break;
+      break;
 
     case MODE_GOODQUALITY:
-        cpi->pass = 0;
-        cpi->compressor_speed = 1;
+      cpi->pass = 0;
+      cpi->compressor_speed = 1;
 
-        if (cpi->oxcf.cpu_used < -5)
-        {
-            cpi->oxcf.cpu_used = -5;
-        }
+      if (cpi->oxcf.cpu_used < -5) {
+        cpi->oxcf.cpu_used = -5;
+      }
 
-        if (cpi->oxcf.cpu_used > 5)
-            cpi->oxcf.cpu_used = 5;
+      if (cpi->oxcf.cpu_used > 5) cpi->oxcf.cpu_used = 5;
 
-        break;
+      break;
 
     case MODE_BESTQUALITY:
-        cpi->pass = 0;
-        cpi->compressor_speed = 0;
-        break;
+      cpi->pass = 0;
+      cpi->compressor_speed = 0;
+      break;
 
     case MODE_FIRSTPASS:
-        cpi->pass = 1;
-        cpi->compressor_speed = 1;
-        break;
+      cpi->pass = 1;
+      cpi->compressor_speed = 1;
+      break;
     case MODE_SECONDPASS:
-        cpi->pass = 2;
-        cpi->compressor_speed = 1;
+      cpi->pass = 2;
+      cpi->compressor_speed = 1;
 
-        if (cpi->oxcf.cpu_used < -5)
-        {
-            cpi->oxcf.cpu_used = -5;
-        }
+      if (cpi->oxcf.cpu_used < -5) {
+        cpi->oxcf.cpu_used = -5;
+      }
 
-        if (cpi->oxcf.cpu_used > 5)
-            cpi->oxcf.cpu_used = 5;
+      if (cpi->oxcf.cpu_used > 5) cpi->oxcf.cpu_used = 5;
 
-        break;
+      break;
     case MODE_SECONDPASS_BEST:
-        cpi->pass = 2;
-        cpi->compressor_speed = 0;
-        break;
-    }
+      cpi->pass = 2;
+      cpi->compressor_speed = 0;
+      break;
+  }
 
-    if (cpi->pass == 0)
-        cpi->auto_worst_q = 1;
+  if (cpi->pass == 0) cpi->auto_worst_q = 1;
 
-    cpi->oxcf.worst_allowed_q = q_trans[oxcf->worst_allowed_q];
-    cpi->oxcf.best_allowed_q = q_trans[oxcf->best_allowed_q];
-    cpi->oxcf.cq_level = q_trans[cpi->oxcf.cq_level];
+  cpi->oxcf.worst_allowed_q = q_trans[oxcf->worst_allowed_q];
+  cpi->oxcf.best_allowed_q = q_trans[oxcf->best_allowed_q];
+  cpi->oxcf.cq_level = q_trans[cpi->oxcf.cq_level];
 
-    if (oxcf->fixed_q >= 0)
-    {
-        if (oxcf->worst_allowed_q < 0)
-            cpi->oxcf.fixed_q = q_trans[0];
-        else
-            cpi->oxcf.fixed_q = q_trans[oxcf->worst_allowed_q];
+  if (oxcf->fixed_q >= 0) {
+    if (oxcf->worst_allowed_q < 0)
+      cpi->oxcf.fixed_q = q_trans[0];
+    else
+      cpi->oxcf.fixed_q = q_trans[oxcf->worst_allowed_q];
 
-        if (oxcf->alt_q < 0)
-            cpi->oxcf.alt_q = q_trans[0];
-        else
-            cpi->oxcf.alt_q = q_trans[oxcf->alt_q];
+    if (oxcf->alt_q < 0)
+      cpi->oxcf.alt_q = q_trans[0];
+    else
+      cpi->oxcf.alt_q = q_trans[oxcf->alt_q];
 
-        if (oxcf->key_q < 0)
-            cpi->oxcf.key_q = q_trans[0];
-        else
-            cpi->oxcf.key_q = q_trans[oxcf->key_q];
+    if (oxcf->key_q < 0)
+      cpi->oxcf.key_q = q_trans[0];
+    else
+      cpi->oxcf.key_q = q_trans[oxcf->key_q];
 
-        if (oxcf->gold_q < 0)
-            cpi->oxcf.gold_q = q_trans[0];
-        else
-            cpi->oxcf.gold_q = q_trans[oxcf->gold_q];
+    if (oxcf->gold_q < 0)
+      cpi->oxcf.gold_q = q_trans[0];
+    else
+      cpi->oxcf.gold_q = q_trans[oxcf->gold_q];
+  }
 
-    }
+  cpi->baseline_gf_interval =
+      cpi->oxcf.alt_freq ? cpi->oxcf.alt_freq : DEFAULT_GF_INTERVAL;
 
-    cpi->baseline_gf_interval =
-        cpi->oxcf.alt_freq ? cpi->oxcf.alt_freq : DEFAULT_GF_INTERVAL;
-
 #if (CONFIG_REALTIME_ONLY & CONFIG_ONTHEFLY_BITPACKING)
-    cpi->oxcf.token_partitions = 3;
+  cpi->oxcf.token_partitions = 3;
 #endif
 
-    if (cpi->oxcf.token_partitions >= 0 && cpi->oxcf.token_partitions <= 3)
-        cm->multi_token_partition =
-            (TOKEN_PARTITION) cpi->oxcf.token_partitions;
+  if (cpi->oxcf.token_partitions >= 0 && cpi->oxcf.token_partitions <= 3)
+    cm->multi_token_partition = (TOKEN_PARTITION)cpi->oxcf.token_partitions;
 
-    setup_features(cpi);
+  setup_features(cpi);
 
-    {
-        int i;
+  {
+    int i;
 
-        for (i = 0; i < MAX_MB_SEGMENTS; i++)
-            cpi->segment_encode_breakout[i] = cpi->oxcf.encode_breakout;
-    }
+    for (i = 0; i < MAX_MB_SEGMENTS; i++)
+      cpi->segment_encode_breakout[i] = cpi->oxcf.encode_breakout;
+  }
 
-    /* At the moment the first order values may not be > MAXQ */
-    if (cpi->oxcf.fixed_q > MAXQ)
-        cpi->oxcf.fixed_q = MAXQ;
+  /* At the moment the first order values may not be > MAXQ */
+  if (cpi->oxcf.fixed_q > MAXQ) cpi->oxcf.fixed_q = MAXQ;
 
-    /* local file playback mode == really big buffer */
-    if (cpi->oxcf.end_usage == USAGE_LOCAL_FILE_PLAYBACK)
-    {
-        cpi->oxcf.starting_buffer_level       = 60000;
-        cpi->oxcf.optimal_buffer_level        = 60000;
-        cpi->oxcf.maximum_buffer_size         = 240000;
-        cpi->oxcf.starting_buffer_level_in_ms = 60000;
-        cpi->oxcf.optimal_buffer_level_in_ms  = 60000;
-        cpi->oxcf.maximum_buffer_size_in_ms   = 240000;
-    }
+  /* local file playback mode == really big buffer */
+  if (cpi->oxcf.end_usage == USAGE_LOCAL_FILE_PLAYBACK) {
+    cpi->oxcf.starting_buffer_level = 60000;
+    cpi->oxcf.optimal_buffer_level = 60000;
+    cpi->oxcf.maximum_buffer_size = 240000;
+    cpi->oxcf.starting_buffer_level_in_ms = 60000;
+    cpi->oxcf.optimal_buffer_level_in_ms = 60000;
+    cpi->oxcf.maximum_buffer_size_in_ms = 240000;
+  }
 
-    /* Convert target bandwidth from Kbit/s to Bit/s */
-    cpi->oxcf.target_bandwidth       *= 1000;
+  /* Convert target bandwidth from Kbit/s to Bit/s */
+  cpi->oxcf.target_bandwidth *= 1000;
 
-    cpi->oxcf.starting_buffer_level =
-        rescale((int)cpi->oxcf.starting_buffer_level,
-                cpi->oxcf.target_bandwidth, 1000);
+  cpi->oxcf.starting_buffer_level = rescale(
+      (int)cpi->oxcf.starting_buffer_level, cpi->oxcf.target_bandwidth, 1000);
 
-    /* Set or reset optimal and maximum buffer levels. */
-    if (cpi->oxcf.optimal_buffer_level == 0)
-        cpi->oxcf.optimal_buffer_level = cpi->oxcf.target_bandwidth / 8;
-    else
-        cpi->oxcf.optimal_buffer_level =
-            rescale((int)cpi->oxcf.optimal_buffer_level,
-                    cpi->oxcf.target_bandwidth, 1000);
+  /* Set or reset optimal and maximum buffer levels. */
+  if (cpi->oxcf.optimal_buffer_level == 0)
+    cpi->oxcf.optimal_buffer_level = cpi->oxcf.target_bandwidth / 8;
+  else
+    cpi->oxcf.optimal_buffer_level = rescale(
+        (int)cpi->oxcf.optimal_buffer_level, cpi->oxcf.target_bandwidth, 1000);
 
-    if (cpi->oxcf.maximum_buffer_size == 0)
-        cpi->oxcf.maximum_buffer_size = cpi->oxcf.target_bandwidth / 8;
-    else
-        cpi->oxcf.maximum_buffer_size =
-            rescale((int)cpi->oxcf.maximum_buffer_size,
-                    cpi->oxcf.target_bandwidth, 1000);
-    // Under a configuration change, where maximum_buffer_size may change,
-    // keep buffer level clipped to the maximum allowed buffer size.
-    if (cpi->bits_off_target > cpi->oxcf.maximum_buffer_size) {
-      cpi->bits_off_target = cpi->oxcf.maximum_buffer_size;
-      cpi->buffer_level = cpi->bits_off_target;
-    }
+  if (cpi->oxcf.maximum_buffer_size == 0)
+    cpi->oxcf.maximum_buffer_size = cpi->oxcf.target_bandwidth / 8;
+  else
+    cpi->oxcf.maximum_buffer_size = rescale((int)cpi->oxcf.maximum_buffer_size,
+                                            cpi->oxcf.target_bandwidth, 1000);
+  // Under a configuration change, where maximum_buffer_size may change,
+  // keep buffer level clipped to the maximum allowed buffer size.
+  if (cpi->bits_off_target > cpi->oxcf.maximum_buffer_size) {
+    cpi->bits_off_target = cpi->oxcf.maximum_buffer_size;
+    cpi->buffer_level = cpi->bits_off_target;
+  }
 
-    /* Set up frame rate and related parameters rate control values. */
-    vp8_new_framerate(cpi, cpi->framerate);
+  /* Set up frame rate and related parameters rate control values. */
+  vp8_new_framerate(cpi, cpi->framerate);
 
-    /* Set absolute upper and lower quality limits */
-    cpi->worst_quality               = cpi->oxcf.worst_allowed_q;
-    cpi->best_quality                = cpi->oxcf.best_allowed_q;
+  /* Set absolute upper and lower quality limits */
+  cpi->worst_quality = cpi->oxcf.worst_allowed_q;
+  cpi->best_quality = cpi->oxcf.best_allowed_q;
 
-    /* active values should only be modified if out of new range */
-    if (cpi->active_worst_quality > cpi->oxcf.worst_allowed_q)
-    {
-      cpi->active_worst_quality = cpi->oxcf.worst_allowed_q;
-    }
-    /* less likely */
-    else if (cpi->active_worst_quality < cpi->oxcf.best_allowed_q)
-    {
-      cpi->active_worst_quality = cpi->oxcf.best_allowed_q;
-    }
-    if (cpi->active_best_quality < cpi->oxcf.best_allowed_q)
-    {
-      cpi->active_best_quality = cpi->oxcf.best_allowed_q;
-    }
-    /* less likely */
-    else if (cpi->active_best_quality > cpi->oxcf.worst_allowed_q)
-    {
-      cpi->active_best_quality = cpi->oxcf.worst_allowed_q;
-    }
+  /* active values should only be modified if out of new range */
+  if (cpi->active_worst_quality > cpi->oxcf.worst_allowed_q) {
+    cpi->active_worst_quality = cpi->oxcf.worst_allowed_q;
+  }
+  /* less likely */
+  else if (cpi->active_worst_quality < cpi->oxcf.best_allowed_q) {
+    cpi->active_worst_quality = cpi->oxcf.best_allowed_q;
+  }
+  if (cpi->active_best_quality < cpi->oxcf.best_allowed_q) {
+    cpi->active_best_quality = cpi->oxcf.best_allowed_q;
+  }
+  /* less likely */
+  else if (cpi->active_best_quality > cpi->oxcf.worst_allowed_q) {
+    cpi->active_best_quality = cpi->oxcf.worst_allowed_q;
+  }
 
-    cpi->buffered_mode = cpi->oxcf.optimal_buffer_level > 0;
+  cpi->buffered_mode = cpi->oxcf.optimal_buffer_level > 0;
 
-    cpi->cq_target_quality = cpi->oxcf.cq_level;
+  cpi->cq_target_quality = cpi->oxcf.cq_level;
 
-    /* Only allow dropped frames in buffered mode */
-    cpi->drop_frames_allowed = cpi->oxcf.allow_df && cpi->buffered_mode;
+  /* Only allow dropped frames in buffered mode */
+  cpi->drop_frames_allowed = cpi->oxcf.allow_df && cpi->buffered_mode;
 
-    cpi->target_bandwidth = cpi->oxcf.target_bandwidth;
+  cpi->target_bandwidth = cpi->oxcf.target_bandwidth;
 
-    // Check if the number of temporal layers has changed, and if so reset the
-    // pattern counter and set/initialize the temporal layer context for the
-    // new layer configuration.
-    if (cpi->oxcf.number_of_layers != prev_number_of_layers)
-    {
-        // If the number of temporal layers are changed we must start at the
-        // base of the pattern cycle, so set the layer id to 0 and reset
-        // the temporal pattern counter.
-        if (cpi->temporal_layer_id > 0) {
-          cpi->temporal_layer_id = 0;
-        }
-        cpi->temporal_pattern_counter = 0;
-        reset_temporal_layer_change(cpi, oxcf, prev_number_of_layers);
+  // Check if the number of temporal layers has changed, and if so reset the
+  // pattern counter and set/initialize the temporal layer context for the
+  // new layer configuration.
+  if (cpi->oxcf.number_of_layers != prev_number_of_layers) {
+    // If the number of temporal layers are changed we must start at the
+    // base of the pattern cycle, so set the layer id to 0 and reset
+    // the temporal pattern counter.
+    if (cpi->temporal_layer_id > 0) {
+      cpi->temporal_layer_id = 0;
     }
+    cpi->temporal_pattern_counter = 0;
+    reset_temporal_layer_change(cpi, oxcf, prev_number_of_layers);
+  }
 
-    if (!cpi->initial_width)
-    {
-        cpi->initial_width = cpi->oxcf.Width;
-        cpi->initial_height = cpi->oxcf.Height;
-    }
+  if (!cpi->initial_width) {
+    cpi->initial_width = cpi->oxcf.Width;
+    cpi->initial_height = cpi->oxcf.Height;
+  }
 
-    cm->Width       = cpi->oxcf.Width;
-    cm->Height      = cpi->oxcf.Height;
-    assert(cm->Width <= cpi->initial_width);
-    assert(cm->Height <= cpi->initial_height);
+  cm->Width = cpi->oxcf.Width;
+  cm->Height = cpi->oxcf.Height;
+  assert(cm->Width <= cpi->initial_width);
+  assert(cm->Height <= cpi->initial_height);
 
-    /* TODO(jkoleszar): if an internal spatial resampling is active,
-     * and we downsize the input image, maybe we should clear the
-     * internal scale immediately rather than waiting for it to
-     * correct.
-     */
+  /* TODO(jkoleszar): if an internal spatial resampling is active,
+   * and we downsize the input image, maybe we should clear the
+   * internal scale immediately rather than waiting for it to
+   * correct.
+   */
 
-    /* VP8 sharpness level mapping 0-7 (vs 0-10 in general VPx dialogs) */
-    if (cpi->oxcf.Sharpness > 7)
-        cpi->oxcf.Sharpness = 7;
+  /* VP8 sharpness level mapping 0-7 (vs 0-10 in general VPx dialogs) */
+  if (cpi->oxcf.Sharpness > 7) cpi->oxcf.Sharpness = 7;
 
-    cm->sharpness_level = cpi->oxcf.Sharpness;
+  cm->sharpness_level = cpi->oxcf.Sharpness;
 
-    if (cm->horiz_scale != NORMAL || cm->vert_scale != NORMAL)
-    {
-        int UNINITIALIZED_IS_SAFE(hr), UNINITIALIZED_IS_SAFE(hs);
-        int UNINITIALIZED_IS_SAFE(vr), UNINITIALIZED_IS_SAFE(vs);
+  if (cm->horiz_scale != NORMAL || cm->vert_scale != NORMAL) {
+    int UNINITIALIZED_IS_SAFE(hr), UNINITIALIZED_IS_SAFE(hs);
+    int UNINITIALIZED_IS_SAFE(vr), UNINITIALIZED_IS_SAFE(vs);
 
-        Scale2Ratio(cm->horiz_scale, &hr, &hs);
-        Scale2Ratio(cm->vert_scale, &vr, &vs);
+    Scale2Ratio(cm->horiz_scale, &hr, &hs);
+    Scale2Ratio(cm->vert_scale, &vr, &vs);
 
-        /* always go to the next whole number */
-        cm->Width = (hs - 1 + cpi->oxcf.Width * hr) / hs;
-        cm->Height = (vs - 1 + cpi->oxcf.Height * vr) / vs;
-    }
+    /* always go to the next whole number */
+    cm->Width = (hs - 1 + cpi->oxcf.Width * hr) / hs;
+    cm->Height = (vs - 1 + cpi->oxcf.Height * vr) / vs;
+  }
 
-    if (last_w != cpi->oxcf.Width || last_h != cpi->oxcf.Height)
-        cpi->force_next_frame_intra = 1;
+  if (last_w != cpi->oxcf.Width || last_h != cpi->oxcf.Height)
+    cpi->force_next_frame_intra = 1;
 
-    if (((cm->Width + 15) & ~15) != cm->yv12_fb[cm->lst_fb_idx].y_width ||
-        ((cm->Height + 15) & ~15) != cm->yv12_fb[cm->lst_fb_idx].y_height ||
-        cm->yv12_fb[cm->lst_fb_idx].y_width == 0)
-    {
-        dealloc_raw_frame_buffers(cpi);
-        alloc_raw_frame_buffers(cpi);
-        vp8_alloc_compressor_data(cpi);
-    }
+  if (((cm->Width + 15) & ~15) != cm->yv12_fb[cm->lst_fb_idx].y_width ||
+      ((cm->Height + 15) & ~15) != cm->yv12_fb[cm->lst_fb_idx].y_height ||
+      cm->yv12_fb[cm->lst_fb_idx].y_width == 0) {
+    dealloc_raw_frame_buffers(cpi);
+    alloc_raw_frame_buffers(cpi);
+    vp8_alloc_compressor_data(cpi);
+  }
 
-    if (cpi->oxcf.fixed_q >= 0)
-    {
-        cpi->last_q[0] = cpi->oxcf.fixed_q;
-        cpi->last_q[1] = cpi->oxcf.fixed_q;
-    }
+  if (cpi->oxcf.fixed_q >= 0) {
+    cpi->last_q[0] = cpi->oxcf.fixed_q;
+    cpi->last_q[1] = cpi->oxcf.fixed_q;
+  }
 
-    cpi->Speed = cpi->oxcf.cpu_used;
+  cpi->Speed = cpi->oxcf.cpu_used;
 
-    /* force to allowlag to 0 if lag_in_frames is 0; */
-    if (cpi->oxcf.lag_in_frames == 0)
-    {
-        cpi->oxcf.allow_lag = 0;
-    }
-    /* Limit on lag buffers as these are not currently dynamically allocated */
-    else if (cpi->oxcf.lag_in_frames > MAX_LAG_BUFFERS)
-        cpi->oxcf.lag_in_frames = MAX_LAG_BUFFERS;
+  /* force to allowlag to 0 if lag_in_frames is 0; */
+  if (cpi->oxcf.lag_in_frames == 0) {
+    cpi->oxcf.allow_lag = 0;
+  }
+  /* Limit on lag buffers as these are not currently dynamically allocated */
+  else if (cpi->oxcf.lag_in_frames > MAX_LAG_BUFFERS)
+    cpi->oxcf.lag_in_frames = MAX_LAG_BUFFERS;
 
-    /* YX Temp */
-    cpi->alt_ref_source = NULL;
-    cpi->is_src_frame_alt_ref = 0;
+  /* YX Temp */
+  cpi->alt_ref_source = NULL;
+  cpi->is_src_frame_alt_ref = 0;
 
 #if CONFIG_TEMPORAL_DENOISING
-    if (cpi->oxcf.noise_sensitivity)
-    {
-      if (!cpi->denoiser.yv12_mc_running_avg.buffer_alloc)
-      {
-        int width = (cpi->oxcf.Width + 15) & ~15;
-        int height = (cpi->oxcf.Height + 15) & ~15;
-        if (vp8_denoiser_allocate(&cpi->denoiser, width, height,
-                                  cm->mb_rows, cm->mb_cols,
-                                  cpi->oxcf.noise_sensitivity))
-            vpx_internal_error(&cpi->common.error, VPX_CODEC_MEM_ERROR,
-                               "Failed to allocate denoiser");
-      }
+  if (cpi->oxcf.noise_sensitivity) {
+    if (!cpi->denoiser.yv12_mc_running_avg.buffer_alloc) {
+      int width = (cpi->oxcf.Width + 15) & ~15;
+      int height = (cpi->oxcf.Height + 15) & ~15;
+      if (vp8_denoiser_allocate(&cpi->denoiser, width, height, cm->mb_rows,
+                                cm->mb_cols, cpi->oxcf.noise_sensitivity))
+        vpx_internal_error(&cpi->common.error, VPX_CODEC_MEM_ERROR,
+                           "Failed to allocate denoiser");
     }
+  }
 #endif
 
 #if 0
@@ -1838,95 +1653,90 @@
     cpi->frame_distortion = 0;
     cpi->last_frame_distortion = 0;
 #endif
-
 }
 
 #ifndef M_LOG2_E
 #define M_LOG2_E 0.693147180559945309417
 #endif
-#define log2f(x) (log (x) / (float) M_LOG2_E)
+#define log2f(x) (log(x) / (float)M_LOG2_E)
 
-static void cal_mvsadcosts(int *mvsadcost[2])
-{
-    int i = 1;
+static void cal_mvsadcosts(int *mvsadcost[2]) {
+  int i = 1;
 
-    mvsadcost [0] [0] = 300;
-    mvsadcost [1] [0] = 300;
+  mvsadcost[0][0] = 300;
+  mvsadcost[1][0] = 300;
 
-    do
-    {
-        double z = 256 * (2 * (log2f(8 * i) + .6));
-        mvsadcost [0][i] = (int) z;
-        mvsadcost [1][i] = (int) z;
-        mvsadcost [0][-i] = (int) z;
-        mvsadcost [1][-i] = (int) z;
-    }
-    while (++i <= mvfp_max);
+  do {
+    double z = 256 * (2 * (log2f(8 * i) + .6));
+    mvsadcost[0][i] = (int)z;
+    mvsadcost[1][i] = (int)z;
+    mvsadcost[0][-i] = (int)z;
+    mvsadcost[1][-i] = (int)z;
+  } while (++i <= mvfp_max);
 }
 
-struct VP8_COMP* vp8_create_compressor(VP8_CONFIG *oxcf)
-{
-    int i;
+struct VP8_COMP *vp8_create_compressor(VP8_CONFIG *oxcf) {
+  int i;
 
-    VP8_COMP *cpi;
-    VP8_COMMON *cm;
+  VP8_COMP *cpi;
+  VP8_COMMON *cm;
 
-    cpi = vpx_memalign(32, sizeof(VP8_COMP));
-    /* Check that the CPI instance is valid */
-    if (!cpi)
-        return 0;
+  cpi = vpx_memalign(32, sizeof(VP8_COMP));
+  /* Check that the CPI instance is valid */
+  if (!cpi) return 0;
 
-    cm = &cpi->common;
+  cm = &cpi->common;
 
-    memset(cpi, 0, sizeof(VP8_COMP));
+  memset(cpi, 0, sizeof(VP8_COMP));
 
-    if (setjmp(cm->error.jmp))
-    {
-        cpi->common.error.setjmp = 0;
-        vp8_remove_compressor(&cpi);
-        return 0;
-    }
+  if (setjmp(cm->error.jmp)) {
+    cpi->common.error.setjmp = 0;
+    vp8_remove_compressor(&cpi);
+    return 0;
+  }
 
-    cpi->common.error.setjmp = 1;
+  cpi->common.error.setjmp = 1;
 
-    CHECK_MEM_ERROR(cpi->mb.ss, vpx_calloc(sizeof(search_site), (MAX_MVSEARCH_STEPS * 8) + 1));
+  CHECK_MEM_ERROR(cpi->mb.ss, vpx_calloc(sizeof(search_site),
+                                         (MAX_MVSEARCH_STEPS * 8) + 1));
 
-    vp8_create_common(&cpi->common);
+  vp8_create_common(&cpi->common);
 
-    init_config(cpi, oxcf);
+  init_config(cpi, oxcf);
 
-    memcpy(cpi->base_skip_false_prob, vp8cx_base_skip_false_prob, sizeof(vp8cx_base_skip_false_prob));
-    cpi->common.current_video_frame   = 0;
-    cpi->temporal_pattern_counter     = 0;
-    cpi->temporal_layer_id            = -1;
-    cpi->kf_overspend_bits            = 0;
-    cpi->kf_bitrate_adjustment        = 0;
-    cpi->frames_till_gf_update_due      = 0;
-    cpi->gf_overspend_bits            = 0;
-    cpi->non_gf_bitrate_adjustment     = 0;
-    cpi->prob_last_coded              = 128;
-    cpi->prob_gf_coded                = 128;
-    cpi->prob_intra_coded             = 63;
+  memcpy(cpi->base_skip_false_prob, vp8cx_base_skip_false_prob,
+         sizeof(vp8cx_base_skip_false_prob));
+  cpi->common.current_video_frame = 0;
+  cpi->temporal_pattern_counter = 0;
+  cpi->temporal_layer_id = -1;
+  cpi->kf_overspend_bits = 0;
+  cpi->kf_bitrate_adjustment = 0;
+  cpi->frames_till_gf_update_due = 0;
+  cpi->gf_overspend_bits = 0;
+  cpi->non_gf_bitrate_adjustment = 0;
+  cpi->prob_last_coded = 128;
+  cpi->prob_gf_coded = 128;
+  cpi->prob_intra_coded = 63;
 
-    /* Prime the recent reference frame usage counters.
-     * Hereafter they will be maintained as a sort of moving average
-     */
-    cpi->recent_ref_frame_usage[INTRA_FRAME]  = 1;
-    cpi->recent_ref_frame_usage[LAST_FRAME]   = 1;
-    cpi->recent_ref_frame_usage[GOLDEN_FRAME] = 1;
-    cpi->recent_ref_frame_usage[ALTREF_FRAME] = 1;
+  /* Prime the recent reference frame usage counters.
+   * Hereafter they will be maintained as a sort of moving average
+   */
+  cpi->recent_ref_frame_usage[INTRA_FRAME] = 1;
+  cpi->recent_ref_frame_usage[LAST_FRAME] = 1;
+  cpi->recent_ref_frame_usage[GOLDEN_FRAME] = 1;
+  cpi->recent_ref_frame_usage[ALTREF_FRAME] = 1;
 
-    /* Set reference frame sign bias for ALTREF frame to 1 (for now) */
-    cpi->common.ref_frame_sign_bias[ALTREF_FRAME] = 1;
+  /* Set reference frame sign bias for ALTREF frame to 1 (for now) */
+  cpi->common.ref_frame_sign_bias[ALTREF_FRAME] = 1;
 
-    cpi->twopass.gf_decay_rate = 0;
-    cpi->baseline_gf_interval = DEFAULT_GF_INTERVAL;
+  cpi->twopass.gf_decay_rate = 0;
+  cpi->baseline_gf_interval = DEFAULT_GF_INTERVAL;
 
-    cpi->gold_is_last = 0 ;
-    cpi->alt_is_last  = 0 ;
-    cpi->gold_is_alt  = 0 ;
+  cpi->gold_is_last = 0;
+  cpi->alt_is_last = 0;
+  cpi->gold_is_alt = 0;
 
-    cpi->active_map_enabled = 0;
+  cpi->active_map_enabled = 0;
 
 #if 0
     /* Experimental code for lagged and one pass */
@@ -1951,107 +1761,105 @@
     }
 #endif
 
-    cpi->mse_source_denoised = 0;
+  cpi->mse_source_denoised = 0;
 
-    /* Should we use the cyclic refresh method.
-     * Currently this is tied to error resilliant mode
-     */
-    cpi->cyclic_refresh_mode_enabled = cpi->oxcf.error_resilient_mode;
-    cpi->cyclic_refresh_mode_max_mbs_perframe = (cpi->common.mb_rows * cpi->common.mb_cols) / 7;
-    if (cpi->oxcf.number_of_layers == 1) {
-        cpi->cyclic_refresh_mode_max_mbs_perframe =
-            (cpi->common.mb_rows * cpi->common.mb_cols) / 20;
-    } else if (cpi->oxcf.number_of_layers == 2) {
-        cpi->cyclic_refresh_mode_max_mbs_perframe =
-            (cpi->common.mb_rows * cpi->common.mb_cols) / 10;
-    }
-    cpi->cyclic_refresh_mode_index = 0;
-    cpi->cyclic_refresh_q = 32;
+  /* Should we use the cyclic refresh method.
+   * Currently this is tied to error resilliant mode
+   */
+  cpi->cyclic_refresh_mode_enabled = cpi->oxcf.error_resilient_mode;
+  cpi->cyclic_refresh_mode_max_mbs_perframe =
+      (cpi->common.mb_rows * cpi->common.mb_cols) / 7;
+  if (cpi->oxcf.number_of_layers == 1) {
+    cpi->cyclic_refresh_mode_max_mbs_perframe =
+        (cpi->common.mb_rows * cpi->common.mb_cols) / 20;
+  } else if (cpi->oxcf.number_of_layers == 2) {
+    cpi->cyclic_refresh_mode_max_mbs_perframe =
+        (cpi->common.mb_rows * cpi->common.mb_cols) / 10;
+  }
+  cpi->cyclic_refresh_mode_index = 0;
+  cpi->cyclic_refresh_q = 32;
 
-    if (cpi->cyclic_refresh_mode_enabled)
-    {
-        CHECK_MEM_ERROR(cpi->cyclic_refresh_map, vpx_calloc((cpi->common.mb_rows * cpi->common.mb_cols), 1));
-    }
-    else
-        cpi->cyclic_refresh_map = (signed char *) NULL;
-
-    CHECK_MEM_ERROR(cpi->consec_zero_last,
-                    vpx_calloc(cm->mb_rows * cm->mb_cols, 1));
-    CHECK_MEM_ERROR(cpi->consec_zero_last_mvbias,
+  if (cpi->cyclic_refresh_mode_enabled) {
+    CHECK_MEM_ERROR(cpi->cyclic_refresh_map,
                     vpx_calloc((cpi->common.mb_rows * cpi->common.mb_cols), 1));
+  } else
+    cpi->cyclic_refresh_map = (signed char *)NULL;
 
+  CHECK_MEM_ERROR(cpi->consec_zero_last,
+                  vpx_calloc(cm->mb_rows * cm->mb_cols, 1));
+  CHECK_MEM_ERROR(cpi->consec_zero_last_mvbias,
+                  vpx_calloc((cpi->common.mb_rows * cpi->common.mb_cols), 1));
+
 #ifdef VP8_ENTROPY_STATS
-    init_context_counters();
+  init_context_counters();
 #endif
 
-    /*Initialize the feed-forward activity masking.*/
-    cpi->activity_avg = 90<<12;
+  /*Initialize the feed-forward activity masking.*/
+  cpi->activity_avg = 90 << 12;
 
-    /* Give a sensible default for the first frame. */
-    cpi->frames_since_key = 8;
-    cpi->key_frame_frequency = cpi->oxcf.key_freq;
-    cpi->this_key_frame_forced = 0;
-    cpi->next_key_frame_forced = 0;
+  /* Give a sensible default for the first frame. */
+  cpi->frames_since_key = 8;
+  cpi->key_frame_frequency = cpi->oxcf.key_freq;
+  cpi->this_key_frame_forced = 0;
+  cpi->next_key_frame_forced = 0;
 
-    cpi->source_alt_ref_pending = 0;
-    cpi->source_alt_ref_active = 0;
-    cpi->common.refresh_alt_ref_frame = 0;
+  cpi->source_alt_ref_pending = 0;
+  cpi->source_alt_ref_active = 0;
+  cpi->common.refresh_alt_ref_frame = 0;
 
-    cpi->force_maxqp = 0;
+  cpi->force_maxqp = 0;
 
-    cpi->b_calculate_psnr = CONFIG_INTERNAL_STATS;
+  cpi->b_calculate_psnr = CONFIG_INTERNAL_STATS;
 #if CONFIG_INTERNAL_STATS
-    cpi->b_calculate_ssimg = 0;
+  cpi->b_calculate_ssimg = 0;
 
-    cpi->count = 0;
-    cpi->bytes = 0;
+  cpi->count = 0;
+  cpi->bytes = 0;
 
-    if (cpi->b_calculate_psnr)
-    {
-        cpi->total_sq_error = 0.0;
-        cpi->total_sq_error2 = 0.0;
-        cpi->total_y = 0.0;
-        cpi->total_u = 0.0;
-        cpi->total_v = 0.0;
-        cpi->total = 0.0;
-        cpi->totalp_y = 0.0;
-        cpi->totalp_u = 0.0;
-        cpi->totalp_v = 0.0;
-        cpi->totalp = 0.0;
-        cpi->tot_recode_hits = 0;
-        cpi->summed_quality = 0;
-        cpi->summed_weights = 0;
-    }
+  if (cpi->b_calculate_psnr) {
+    cpi->total_sq_error = 0.0;
+    cpi->total_sq_error2 = 0.0;
+    cpi->total_y = 0.0;
+    cpi->total_u = 0.0;
+    cpi->total_v = 0.0;
+    cpi->total = 0.0;
+    cpi->totalp_y = 0.0;
+    cpi->totalp_u = 0.0;
+    cpi->totalp_v = 0.0;
+    cpi->totalp = 0.0;
+    cpi->tot_recode_hits = 0;
+    cpi->summed_quality = 0;
+    cpi->summed_weights = 0;
+  }
 
 #endif
 
-    cpi->first_time_stamp_ever = 0x7FFFFFFF;
+  cpi->first_time_stamp_ever = 0x7FFFFFFF;
 
-    cpi->frames_till_gf_update_due      = 0;
-    cpi->key_frame_count              = 1;
+  cpi->frames_till_gf_update_due = 0;
+  cpi->key_frame_count = 1;
 
-    cpi->ni_av_qi                     = cpi->oxcf.worst_allowed_q;
-    cpi->ni_tot_qi                    = 0;
-    cpi->ni_frames                   = 0;
-    cpi->total_byte_count             = 0;
+  cpi->ni_av_qi = cpi->oxcf.worst_allowed_q;
+  cpi->ni_tot_qi = 0;
+  cpi->ni_frames = 0;
+  cpi->total_byte_count = 0;
 
-    cpi->drop_frame                  = 0;
+  cpi->drop_frame = 0;
 
-    cpi->rate_correction_factor         = 1.0;
-    cpi->key_frame_rate_correction_factor = 1.0;
-    cpi->gf_rate_correction_factor  = 1.0;
-    cpi->twopass.est_max_qcorrection_factor  = 1.0;
+  cpi->rate_correction_factor = 1.0;
+  cpi->key_frame_rate_correction_factor = 1.0;
+  cpi->gf_rate_correction_factor = 1.0;
+  cpi->twopass.est_max_qcorrection_factor = 1.0;
 
-    for (i = 0; i < KEY_FRAME_CONTEXT; i++)
-    {
-        cpi->prior_key_frame_distance[i] = (int)cpi->output_framerate;
-    }
+  for (i = 0; i < KEY_FRAME_CONTEXT; i++) {
+    cpi->prior_key_frame_distance[i] = (int)cpi->output_framerate;
+  }
 
 #ifdef OUTPUT_YUV_SRC
-    yuv_file = fopen("bd.yuv", "ab");
+  yuv_file = fopen("bd.yuv", "ab");
 #endif
 #ifdef OUTPUT_YUV_DENOISED
-    yuv_denoised_file = fopen("denoised.yuv", "ab");
+  yuv_denoised_file = fopen("denoised.yuv", "ab");
 #endif
 
 #if 0
@@ -2059,379 +1867,350 @@
     kf_list = fopen("kf_list.stt", "w");
 #endif
 
-    cpi->output_pkt_list = oxcf->output_pkt_list;
+  cpi->output_pkt_list = oxcf->output_pkt_list;
 
 #if !CONFIG_REALTIME_ONLY
 
-    if (cpi->pass == 1)
-    {
-        vp8_init_first_pass(cpi);
-    }
-    else if (cpi->pass == 2)
-    {
-        size_t packet_sz = sizeof(FIRSTPASS_STATS);
-        int packets = (int)(oxcf->two_pass_stats_in.sz / packet_sz);
+  if (cpi->pass == 1) {
+    vp8_init_first_pass(cpi);
+  } else if (cpi->pass == 2) {
+    size_t packet_sz = sizeof(FIRSTPASS_STATS);
+    int packets = (int)(oxcf->two_pass_stats_in.sz / packet_sz);
 
-        cpi->twopass.stats_in_start = oxcf->two_pass_stats_in.buf;
-        cpi->twopass.stats_in = cpi->twopass.stats_in_start;
-        cpi->twopass.stats_in_end = (void*)((char *)cpi->twopass.stats_in
-                            + (packets - 1) * packet_sz);
-        vp8_init_second_pass(cpi);
-    }
+    cpi->twopass.stats_in_start = oxcf->two_pass_stats_in.buf;
+    cpi->twopass.stats_in = cpi->twopass.stats_in_start;
+    cpi->twopass.stats_in_end =
+        (void *)((char *)cpi->twopass.stats_in + (packets - 1) * packet_sz);
+    vp8_init_second_pass(cpi);
+  }
 
 #endif
 
-    if (cpi->compressor_speed == 2)
-    {
-        cpi->avg_encode_time      = 0;
-        cpi->avg_pick_mode_time    = 0;
-    }
+  if (cpi->compressor_speed == 2) {
+    cpi->avg_encode_time = 0;
+    cpi->avg_pick_mode_time = 0;
+  }
 
-    vp8_set_speed_features(cpi);
+  vp8_set_speed_features(cpi);
 
-    /* Set starting values of RD threshold multipliers (128 = *1) */
-    for (i = 0; i < MAX_MODES; i++)
-    {
-        cpi->mb.rd_thresh_mult[i] = 128;
-    }
+  /* Set starting values of RD threshold multipliers (128 = *1) */
+  for (i = 0; i < MAX_MODES; i++) {
+    cpi->mb.rd_thresh_mult[i] = 128;
+  }
 
 #ifdef VP8_ENTROPY_STATS
-    init_mv_ref_counts();
+  init_mv_ref_counts();
 #endif
 
 #if CONFIG_MULTITHREAD
-    if(vp8cx_create_encoder_threads(cpi))
-    {
-        vp8_remove_compressor(&cpi);
-        return 0;
-    }
+  if (vp8cx_create_encoder_threads(cpi)) {
+    vp8_remove_compressor(&cpi);
+    return 0;
+  }
 #endif
 
-    cpi->fn_ptr[BLOCK_16X16].sdf            = vpx_sad16x16;
-    cpi->fn_ptr[BLOCK_16X16].vf             = vpx_variance16x16;
-    cpi->fn_ptr[BLOCK_16X16].svf            = vpx_sub_pixel_variance16x16;
-    cpi->fn_ptr[BLOCK_16X16].svf_halfpix_h  = vpx_variance_halfpixvar16x16_h;
-    cpi->fn_ptr[BLOCK_16X16].svf_halfpix_v  = vpx_variance_halfpixvar16x16_v;
-    cpi->fn_ptr[BLOCK_16X16].svf_halfpix_hv = vpx_variance_halfpixvar16x16_hv;
-    cpi->fn_ptr[BLOCK_16X16].sdx3f          = vpx_sad16x16x3;
-    cpi->fn_ptr[BLOCK_16X16].sdx8f          = vpx_sad16x16x8;
-    cpi->fn_ptr[BLOCK_16X16].sdx4df         = vpx_sad16x16x4d;
+  cpi->fn_ptr[BLOCK_16X16].sdf = vpx_sad16x16;
+  cpi->fn_ptr[BLOCK_16X16].vf = vpx_variance16x16;
+  cpi->fn_ptr[BLOCK_16X16].svf = vpx_sub_pixel_variance16x16;
+  cpi->fn_ptr[BLOCK_16X16].svf_halfpix_h = vpx_variance_halfpixvar16x16_h;
+  cpi->fn_ptr[BLOCK_16X16].svf_halfpix_v = vpx_variance_halfpixvar16x16_v;
+  cpi->fn_ptr[BLOCK_16X16].svf_halfpix_hv = vpx_variance_halfpixvar16x16_hv;
+  cpi->fn_ptr[BLOCK_16X16].sdx3f = vpx_sad16x16x3;
+  cpi->fn_ptr[BLOCK_16X16].sdx8f = vpx_sad16x16x8;
+  cpi->fn_ptr[BLOCK_16X16].sdx4df = vpx_sad16x16x4d;
 
-    cpi->fn_ptr[BLOCK_16X8].sdf            = vpx_sad16x8;
-    cpi->fn_ptr[BLOCK_16X8].vf             = vpx_variance16x8;
-    cpi->fn_ptr[BLOCK_16X8].svf            = vpx_sub_pixel_variance16x8;
-    cpi->fn_ptr[BLOCK_16X8].svf_halfpix_h  = NULL;
-    cpi->fn_ptr[BLOCK_16X8].svf_halfpix_v  = NULL;
-    cpi->fn_ptr[BLOCK_16X8].svf_halfpix_hv = NULL;
-    cpi->fn_ptr[BLOCK_16X8].sdx3f          = vpx_sad16x8x3;
-    cpi->fn_ptr[BLOCK_16X8].sdx8f          = vpx_sad16x8x8;
-    cpi->fn_ptr[BLOCK_16X8].sdx4df         = vpx_sad16x8x4d;
+  cpi->fn_ptr[BLOCK_16X8].sdf = vpx_sad16x8;
+  cpi->fn_ptr[BLOCK_16X8].vf = vpx_variance16x8;
+  cpi->fn_ptr[BLOCK_16X8].svf = vpx_sub_pixel_variance16x8;
+  cpi->fn_ptr[BLOCK_16X8].svf_halfpix_h = NULL;
+  cpi->fn_ptr[BLOCK_16X8].svf_halfpix_v = NULL;
+  cpi->fn_ptr[BLOCK_16X8].svf_halfpix_hv = NULL;
+  cpi->fn_ptr[BLOCK_16X8].sdx3f = vpx_sad16x8x3;
+  cpi->fn_ptr[BLOCK_16X8].sdx8f = vpx_sad16x8x8;
+  cpi->fn_ptr[BLOCK_16X8].sdx4df = vpx_sad16x8x4d;
 
-    cpi->fn_ptr[BLOCK_8X16].sdf            = vpx_sad8x16;
-    cpi->fn_ptr[BLOCK_8X16].vf             = vpx_variance8x16;
-    cpi->fn_ptr[BLOCK_8X16].svf            = vpx_sub_pixel_variance8x16;
-    cpi->fn_ptr[BLOCK_8X16].svf_halfpix_h  = NULL;
-    cpi->fn_ptr[BLOCK_8X16].svf_halfpix_v  = NULL;
-    cpi->fn_ptr[BLOCK_8X16].svf_halfpix_hv = NULL;
-    cpi->fn_ptr[BLOCK_8X16].sdx3f          = vpx_sad8x16x3;
-    cpi->fn_ptr[BLOCK_8X16].sdx8f          = vpx_sad8x16x8;
-    cpi->fn_ptr[BLOCK_8X16].sdx4df         = vpx_sad8x16x4d;
+  cpi->fn_ptr[BLOCK_8X16].sdf = vpx_sad8x16;
+  cpi->fn_ptr[BLOCK_8X16].vf = vpx_variance8x16;
+  cpi->fn_ptr[BLOCK_8X16].svf = vpx_sub_pixel_variance8x16;
+  cpi->fn_ptr[BLOCK_8X16].svf_halfpix_h = NULL;
+  cpi->fn_ptr[BLOCK_8X16].svf_halfpix_v = NULL;
+  cpi->fn_ptr[BLOCK_8X16].svf_halfpix_hv = NULL;
+  cpi->fn_ptr[BLOCK_8X16].sdx3f = vpx_sad8x16x3;
+  cpi->fn_ptr[BLOCK_8X16].sdx8f = vpx_sad8x16x8;
+  cpi->fn_ptr[BLOCK_8X16].sdx4df = vpx_sad8x16x4d;
 
-    cpi->fn_ptr[BLOCK_8X8].sdf            = vpx_sad8x8;
-    cpi->fn_ptr[BLOCK_8X8].vf             = vpx_variance8x8;
-    cpi->fn_ptr[BLOCK_8X8].svf            = vpx_sub_pixel_variance8x8;
-    cpi->fn_ptr[BLOCK_8X8].svf_halfpix_h  = NULL;
-    cpi->fn_ptr[BLOCK_8X8].svf_halfpix_v  = NULL;
-    cpi->fn_ptr[BLOCK_8X8].svf_halfpix_hv = NULL;
-    cpi->fn_ptr[BLOCK_8X8].sdx3f          = vpx_sad8x8x3;
-    cpi->fn_ptr[BLOCK_8X8].sdx8f          = vpx_sad8x8x8;
-    cpi->fn_ptr[BLOCK_8X8].sdx4df         = vpx_sad8x8x4d;
+  cpi->fn_ptr[BLOCK_8X8].sdf = vpx_sad8x8;
+  cpi->fn_ptr[BLOCK_8X8].vf = vpx_variance8x8;
+  cpi->fn_ptr[BLOCK_8X8].svf = vpx_sub_pixel_variance8x8;
+  cpi->fn_ptr[BLOCK_8X8].svf_halfpix_h = NULL;
+  cpi->fn_ptr[BLOCK_8X8].svf_halfpix_v = NULL;
+  cpi->fn_ptr[BLOCK_8X8].svf_halfpix_hv = NULL;
+  cpi->fn_ptr[BLOCK_8X8].sdx3f = vpx_sad8x8x3;
+  cpi->fn_ptr[BLOCK_8X8].sdx8f = vpx_sad8x8x8;
+  cpi->fn_ptr[BLOCK_8X8].sdx4df = vpx_sad8x8x4d;
 
-    cpi->fn_ptr[BLOCK_4X4].sdf            = vpx_sad4x4;
-    cpi->fn_ptr[BLOCK_4X4].vf             = vpx_variance4x4;
-    cpi->fn_ptr[BLOCK_4X4].svf            = vpx_sub_pixel_variance4x4;
-    cpi->fn_ptr[BLOCK_4X4].svf_halfpix_h  = NULL;
-    cpi->fn_ptr[BLOCK_4X4].svf_halfpix_v  = NULL;
-    cpi->fn_ptr[BLOCK_4X4].svf_halfpix_hv = NULL;
-    cpi->fn_ptr[BLOCK_4X4].sdx3f          = vpx_sad4x4x3;
-    cpi->fn_ptr[BLOCK_4X4].sdx8f          = vpx_sad4x4x8;
-    cpi->fn_ptr[BLOCK_4X4].sdx4df         = vpx_sad4x4x4d;
+  cpi->fn_ptr[BLOCK_4X4].sdf = vpx_sad4x4;
+  cpi->fn_ptr[BLOCK_4X4].vf = vpx_variance4x4;
+  cpi->fn_ptr[BLOCK_4X4].svf = vpx_sub_pixel_variance4x4;
+  cpi->fn_ptr[BLOCK_4X4].svf_halfpix_h = NULL;
+  cpi->fn_ptr[BLOCK_4X4].svf_halfpix_v = NULL;
+  cpi->fn_ptr[BLOCK_4X4].svf_halfpix_hv = NULL;
+  cpi->fn_ptr[BLOCK_4X4].sdx3f = vpx_sad4x4x3;
+  cpi->fn_ptr[BLOCK_4X4].sdx8f = vpx_sad4x4x8;
+  cpi->fn_ptr[BLOCK_4X4].sdx4df = vpx_sad4x4x4d;
 
 #if ARCH_X86 || ARCH_X86_64
-    cpi->fn_ptr[BLOCK_16X16].copymem      = vp8_copy32xn;
-    cpi->fn_ptr[BLOCK_16X8].copymem       = vp8_copy32xn;
-    cpi->fn_ptr[BLOCK_8X16].copymem       = vp8_copy32xn;
-    cpi->fn_ptr[BLOCK_8X8].copymem        = vp8_copy32xn;
-    cpi->fn_ptr[BLOCK_4X4].copymem        = vp8_copy32xn;
+  cpi->fn_ptr[BLOCK_16X16].copymem = vp8_copy32xn;
+  cpi->fn_ptr[BLOCK_16X8].copymem = vp8_copy32xn;
+  cpi->fn_ptr[BLOCK_8X16].copymem = vp8_copy32xn;
+  cpi->fn_ptr[BLOCK_8X8].copymem = vp8_copy32xn;
+  cpi->fn_ptr[BLOCK_4X4].copymem = vp8_copy32xn;
 #endif
 
-    cpi->full_search_sad = vp8_full_search_sad;
-    cpi->diamond_search_sad = vp8_diamond_search_sad;
-    cpi->refining_search_sad = vp8_refining_search_sad;
+  cpi->full_search_sad = vp8_full_search_sad;
+  cpi->diamond_search_sad = vp8_diamond_search_sad;
+  cpi->refining_search_sad = vp8_refining_search_sad;
 
-    /* make sure frame 1 is okay */
-    cpi->mb.error_bins[0] = cpi->common.MBs;
+  /* make sure frame 1 is okay */
+  cpi->mb.error_bins[0] = cpi->common.MBs;
 
-    /* vp8cx_init_quantizer() is first called here. Add check in
-     * vp8cx_frame_init_quantizer() so that vp8cx_init_quantizer is only
-     * called later when needed. This will avoid unnecessary calls of
-     * vp8cx_init_quantizer() for every frame.
-     */
-    vp8cx_init_quantizer(cpi);
+  /* vp8cx_init_quantizer() is first called here. Add check in
+   * vp8cx_frame_init_quantizer() so that vp8cx_init_quantizer is only
+   * called later when needed. This will avoid unnecessary calls of
+   * vp8cx_init_quantizer() for every frame.
+   */
+  vp8cx_init_quantizer(cpi);
 
-    vp8_loop_filter_init(cm);
+  vp8_loop_filter_init(cm);
 
-    cpi->common.error.setjmp = 0;
+  cpi->common.error.setjmp = 0;
 
 #if CONFIG_MULTI_RES_ENCODING
 
-    /* Calculate # of MBs in a row in lower-resolution level image. */
-    if (cpi->oxcf.mr_encoder_id > 0)
-        vp8_cal_low_res_mb_cols(cpi);
+  /* Calculate # of MBs in a row in lower-resolution level image. */
+  if (cpi->oxcf.mr_encoder_id > 0) vp8_cal_low_res_mb_cols(cpi);
 
 #endif
 
-    /* setup RD costs to MACROBLOCK struct */
+  /* setup RD costs to MACROBLOCK struct */
 
-    cpi->mb.mvcost[0] = &cpi->rd_costs.mvcosts[0][mv_max+1];
-    cpi->mb.mvcost[1] = &cpi->rd_costs.mvcosts[1][mv_max+1];
-    cpi->mb.mvsadcost[0] = &cpi->rd_costs.mvsadcosts[0][mvfp_max+1];
-    cpi->mb.mvsadcost[1] = &cpi->rd_costs.mvsadcosts[1][mvfp_max+1];
+  cpi->mb.mvcost[0] = &cpi->rd_costs.mvcosts[0][mv_max + 1];
+  cpi->mb.mvcost[1] = &cpi->rd_costs.mvcosts[1][mv_max + 1];
+  cpi->mb.mvsadcost[0] = &cpi->rd_costs.mvsadcosts[0][mvfp_max + 1];
+  cpi->mb.mvsadcost[1] = &cpi->rd_costs.mvsadcosts[1][mvfp_max + 1];
 
-    cal_mvsadcosts(cpi->mb.mvsadcost);
+  cal_mvsadcosts(cpi->mb.mvsadcost);
 
-    cpi->mb.mbmode_cost = cpi->rd_costs.mbmode_cost;
-    cpi->mb.intra_uv_mode_cost = cpi->rd_costs.intra_uv_mode_cost;
-    cpi->mb.bmode_costs = cpi->rd_costs.bmode_costs;
-    cpi->mb.inter_bmode_costs = cpi->rd_costs.inter_bmode_costs;
-    cpi->mb.token_costs = cpi->rd_costs.token_costs;
+  cpi->mb.mbmode_cost = cpi->rd_costs.mbmode_cost;
+  cpi->mb.intra_uv_mode_cost = cpi->rd_costs.intra_uv_mode_cost;
+  cpi->mb.bmode_costs = cpi->rd_costs.bmode_costs;
+  cpi->mb.inter_bmode_costs = cpi->rd_costs.inter_bmode_costs;
+  cpi->mb.token_costs = cpi->rd_costs.token_costs;
 
-    /* setup block ptrs & offsets */
-    vp8_setup_block_ptrs(&cpi->mb);
-    vp8_setup_block_dptrs(&cpi->mb.e_mbd);
+  /* setup block ptrs & offsets */
+  vp8_setup_block_ptrs(&cpi->mb);
+  vp8_setup_block_dptrs(&cpi->mb.e_mbd);
 
-    return  cpi;
+  return cpi;
 }
 
+void vp8_remove_compressor(VP8_COMP **ptr) {
+  VP8_COMP *cpi = *ptr;
 
-void vp8_remove_compressor(VP8_COMP **ptr)
-{
-    VP8_COMP *cpi = *ptr;
+  if (!cpi) return;
 
-    if (!cpi)
-        return;
-
-    if (cpi && (cpi->common.current_video_frame > 0))
-    {
+  if (cpi && (cpi->common.current_video_frame > 0)) {
 #if !CONFIG_REALTIME_ONLY
 
-        if (cpi->pass == 2)
-        {
-            vp8_end_second_pass(cpi);
-        }
+    if (cpi->pass == 2) {
+      vp8_end_second_pass(cpi);
+    }
 
 #endif
 
 #ifdef VP8_ENTROPY_STATS
-        print_context_counters();
-        print_tree_update_probs();
-        print_mode_context();
+    print_context_counters();
+    print_tree_update_probs();
+    print_mode_context();
 #endif
 
 #if CONFIG_INTERNAL_STATS
 
-        if (cpi->pass != 1)
-        {
-            FILE *f = fopen("opsnr.stt", "a");
-            double time_encoded = (cpi->last_end_time_stamp_seen
-                                   - cpi->first_time_stamp_ever) / 10000000.000;
-            double total_encode_time = (cpi->time_receive_data +
-                                            cpi->time_compress_data) / 1000.000;
-            double dr = (double)cpi->bytes * 8.0 / 1000.0 / time_encoded;
-            const double target_rate = (double)cpi->oxcf.target_bandwidth / 1000;
-            const double rate_err = ((100.0 * (dr - target_rate)) / target_rate);
+    if (cpi->pass != 1) {
+      FILE *f = fopen("opsnr.stt", "a");
+      double time_encoded =
+          (cpi->last_end_time_stamp_seen - cpi->first_time_stamp_ever) /
+          10000000.000;
+      double total_encode_time =
+          (cpi->time_receive_data + cpi->time_compress_data) / 1000.000;
+      double dr = (double)cpi->bytes * 8.0 / 1000.0 / time_encoded;
+      const double target_rate = (double)cpi->oxcf.target_bandwidth / 1000;
+      const double rate_err = ((100.0 * (dr - target_rate)) / target_rate);
 
-            if (cpi->b_calculate_psnr)
-            {
-                if (cpi->oxcf.number_of_layers > 1)
-                {
-                    int i;
+      if (cpi->b_calculate_psnr) {
+        if (cpi->oxcf.number_of_layers > 1) {
+          int i;
 
-                    fprintf(f, "Layer\tBitrate\tAVGPsnr\tGLBPsnr\tAVPsnrP\t"
-                               "GLPsnrP\tVPXSSIM\t\n");
-                    for (i=0; i<(int)cpi->oxcf.number_of_layers; i++)
-                    {
-                        double dr = (double)cpi->bytes_in_layer[i] *
-                                              8.0 / 1000.0  / time_encoded;
-                        double samples = 3.0 / 2 * cpi->frames_in_layer[i] *
-                                         cpi->common.Width * cpi->common.Height;
-                        double total_psnr =
-                            vpx_sse_to_psnr(samples, 255.0,
-                                            cpi->total_error2[i]);
-                        double total_psnr2 =
-                            vpx_sse_to_psnr(samples, 255.0,
-                                            cpi->total_error2_p[i]);
-                        double total_ssim = 100 * pow(cpi->sum_ssim[i] /
-                                                      cpi->sum_weights[i], 8.0);
+          fprintf(f,
+                  "Layer\tBitrate\tAVGPsnr\tGLBPsnr\tAVPsnrP\t"
+                  "GLPsnrP\tVPXSSIM\t\n");
+          for (i = 0; i < (int)cpi->oxcf.number_of_layers; i++) {
+            double dr =
+                (double)cpi->bytes_in_layer[i] * 8.0 / 1000.0 / time_encoded;
+            double samples = 3.0 / 2 * cpi->frames_in_layer[i] *
+                             cpi->common.Width * cpi->common.Height;
+            double total_psnr =
+                vpx_sse_to_psnr(samples, 255.0, cpi->total_error2[i]);
+            double total_psnr2 =
+                vpx_sse_to_psnr(samples, 255.0, cpi->total_error2_p[i]);
+            double total_ssim =
+                100 * pow(cpi->sum_ssim[i] / cpi->sum_weights[i], 8.0);
 
-                        fprintf(f, "%5d\t%7.3f\t%7.3f\t%7.3f\t%7.3f\t"
-                                   "%7.3f\t%7.3f\n",
-                                   i, dr,
-                                   cpi->sum_psnr[i] / cpi->frames_in_layer[i],
-                                   total_psnr,
-                                   cpi->sum_psnr_p[i] / cpi->frames_in_layer[i],
-                                   total_psnr2, total_ssim);
-                    }
-                }
-                else
-                {
-                    double samples = 3.0 / 2 * cpi->count *
-                                     cpi->common.Width * cpi->common.Height;
-                    double total_psnr = vpx_sse_to_psnr(samples, 255.0,
-                                                        cpi->total_sq_error);
-                    double total_psnr2 = vpx_sse_to_psnr(samples, 255.0,
-                                                         cpi->total_sq_error2);
-                    double total_ssim = 100 * pow(cpi->summed_quality /
-                                                      cpi->summed_weights, 8.0);
+            fprintf(f,
+                    "%5d\t%7.3f\t%7.3f\t%7.3f\t%7.3f\t"
+                    "%7.3f\t%7.3f\n",
+                    i, dr, cpi->sum_psnr[i] / cpi->frames_in_layer[i],
+                    total_psnr, cpi->sum_psnr_p[i] / cpi->frames_in_layer[i],
+                    total_psnr2, total_ssim);
+          }
+        } else {
+          double samples =
+              3.0 / 2 * cpi->count * cpi->common.Width * cpi->common.Height;
+          double total_psnr =
+              vpx_sse_to_psnr(samples, 255.0, cpi->total_sq_error);
+          double total_psnr2 =
+              vpx_sse_to_psnr(samples, 255.0, cpi->total_sq_error2);
+          double total_ssim =
+              100 * pow(cpi->summed_quality / cpi->summed_weights, 8.0);
 
-                    fprintf(f, "Bitrate\tAVGPsnr\tGLBPsnr\tAVPsnrP\t"
-                               "GLPsnrP\tVPXSSIM\t  Time(us)  Rc-Err "
-                               "Abs Err\n");
-                    fprintf(f, "%7.3f\t%7.3f\t%7.3f\t%7.3f\t%7.3f\t"
-                               "%7.3f\t%8.0f %7.2f %7.2f\n",
-                               dr, cpi->total / cpi->count, total_psnr,
-                               cpi->totalp / cpi->count, total_psnr2,
-                               total_ssim, total_encode_time,
-                               rate_err, fabs(rate_err));
-                }
-            }
-            fclose(f);
+          fprintf(f,
+                  "Bitrate\tAVGPsnr\tGLBPsnr\tAVPsnrP\t"
+                  "GLPsnrP\tVPXSSIM\t  Time(us)  Rc-Err "
+                  "Abs Err\n");
+          fprintf(f,
+                  "%7.3f\t%7.3f\t%7.3f\t%7.3f\t%7.3f\t"
+                  "%7.3f\t%8.0f %7.2f %7.2f\n",
+                  dr, cpi->total / cpi->count, total_psnr,
+                  cpi->totalp / cpi->count, total_psnr2, total_ssim,
+                  total_encode_time, rate_err, fabs(rate_err));
+        }
+      }
+      fclose(f);
 #if 0
             f = fopen("qskip.stt", "a");
             fprintf(f, "minq:%d -maxq:%d skiptrue:skipfalse = %d:%d\n", cpi->oxcf.best_allowed_q, cpi->oxcf.worst_allowed_q, skiptruecount, skipfalsecount);
             fclose(f);
 #endif
+    }
 
-        }
-
 #endif
 
-
 #ifdef SPEEDSTATS
 
-        if (cpi->compressor_speed == 2)
-        {
-            int i;
-            FILE *f = fopen("cxspeed.stt", "a");
-            cnt_pm /= cpi->common.MBs;
+    if (cpi->compressor_speed == 2) {
+      int i;
+      FILE *f = fopen("cxspeed.stt", "a");
+      cnt_pm /= cpi->common.MBs;
 
-            for (i = 0; i < 16; i++)
-                fprintf(f, "%5d", frames_at_speed[i]);
+      for (i = 0; i < 16; i++) fprintf(f, "%5d", frames_at_speed[i]);
 
-            fprintf(f, "\n");
-            fclose(f);
-        }
+      fprintf(f, "\n");
+      fclose(f);
+    }
 
 #endif
 
-
 #ifdef MODE_STATS
-        {
-            extern int count_mb_seg[4];
-            FILE *f = fopen("modes.stt", "a");
-            double dr = (double)cpi->framerate * (double)bytes * (double)8 / (double)count / (double)1000 ;
-            fprintf(f, "intra_mode in Intra Frames:\n");
-            fprintf(f, "Y: %8d, %8d, %8d, %8d, %8d\n", y_modes[0], y_modes[1], y_modes[2], y_modes[3], y_modes[4]);
-            fprintf(f, "UV:%8d, %8d, %8d, %8d\n", uv_modes[0], uv_modes[1], uv_modes[2], uv_modes[3]);
-            fprintf(f, "B: ");
-            {
-                int i;
+    {
+      extern int count_mb_seg[4];
+      FILE *f = fopen("modes.stt", "a");
+      double dr = (double)cpi->framerate * (double)bytes * (double)8 /
+                  (double)count / (double)1000;
+      fprintf(f, "intra_mode in Intra Frames:\n");
+      fprintf(f, "Y: %8d, %8d, %8d, %8d, %8d\n", y_modes[0], y_modes[1],
+              y_modes[2], y_modes[3], y_modes[4]);
+      fprintf(f, "UV:%8d, %8d, %8d, %8d\n", uv_modes[0], uv_modes[1],
+              uv_modes[2], uv_modes[3]);
+      fprintf(f, "B: ");
+      {
+        int i;
 
-                for (i = 0; i < 10; i++)
-                    fprintf(f, "%8d, ", b_modes[i]);
+        for (i = 0; i < 10; i++) fprintf(f, "%8d, ", b_modes[i]);
 
-                fprintf(f, "\n");
+        fprintf(f, "\n");
+      }
 
-            }
+      fprintf(f, "Modes in Inter Frames:\n");
+      fprintf(f, "Y: %8d, %8d, %8d, %8d, %8d, %8d, %8d, %8d, %8d, %8d\n",
+              inter_y_modes[0], inter_y_modes[1], inter_y_modes[2],
+              inter_y_modes[3], inter_y_modes[4], inter_y_modes[5],
+              inter_y_modes[6], inter_y_modes[7], inter_y_modes[8],
+              inter_y_modes[9]);
+      fprintf(f, "UV:%8d, %8d, %8d, %8d\n", inter_uv_modes[0],
+              inter_uv_modes[1], inter_uv_modes[2], inter_uv_modes[3]);
+      fprintf(f, "B: ");
+      {
+        int i;
 
-            fprintf(f, "Modes in Inter Frames:\n");
-            fprintf(f, "Y: %8d, %8d, %8d, %8d, %8d, %8d, %8d, %8d, %8d, %8d\n",
-                    inter_y_modes[0], inter_y_modes[1], inter_y_modes[2], inter_y_modes[3], inter_y_modes[4],
-                    inter_y_modes[5], inter_y_modes[6], inter_y_modes[7], inter_y_modes[8], inter_y_modes[9]);
-            fprintf(f, "UV:%8d, %8d, %8d, %8d\n", inter_uv_modes[0], inter_uv_modes[1], inter_uv_modes[2], inter_uv_modes[3]);
-            fprintf(f, "B: ");
-            {
-                int i;
+        for (i = 0; i < 15; i++) fprintf(f, "%8d, ", inter_b_modes[i]);
 
-                for (i = 0; i < 15; i++)
-                    fprintf(f, "%8d, ", inter_b_modes[i]);
+        fprintf(f, "\n");
+      }
+      fprintf(f, "P:%8d, %8d, %8d, %8d\n", count_mb_seg[0], count_mb_seg[1],
+              count_mb_seg[2], count_mb_seg[3]);
+      fprintf(f, "PB:%8d, %8d, %8d, %8d\n", inter_b_modes[LEFT4X4],
+              inter_b_modes[ABOVE4X4], inter_b_modes[ZERO4X4],
+              inter_b_modes[NEW4X4]);
 
-                fprintf(f, "\n");
-
-            }
-            fprintf(f, "P:%8d, %8d, %8d, %8d\n", count_mb_seg[0], count_mb_seg[1], count_mb_seg[2], count_mb_seg[3]);
-            fprintf(f, "PB:%8d, %8d, %8d, %8d\n", inter_b_modes[LEFT4X4], inter_b_modes[ABOVE4X4], inter_b_modes[ZERO4X4], inter_b_modes[NEW4X4]);
-
-
-
-            fclose(f);
-        }
+      fclose(f);
+    }
 #endif
 
 #ifdef VP8_ENTROPY_STATS
-        {
-            int i, j, k;
-            FILE *fmode = fopen("modecontext.c", "w");
-
-            fprintf(fmode, "\n#include \"entropymode.h\"\n\n");
-            fprintf(fmode, "const unsigned int vp8_kf_default_bmode_counts ");
-            fprintf(fmode, "[VP8_BINTRAMODES] [VP8_BINTRAMODES] [VP8_BINTRAMODES] =\n{\n");
-
-            for (i = 0; i < 10; i++)
-            {
-
-                fprintf(fmode, "    { /* Above Mode :  %d */\n", i);
-
-                for (j = 0; j < 10; j++)
-                {
-
-                    fprintf(fmode, "        {");
-
-                    for (k = 0; k < 10; k++)
-                    {
-                        if (!intra_mode_stats[i][j][k])
-                            fprintf(fmode, " %5d, ", 1);
-                        else
-                            fprintf(fmode, " %5d, ", intra_mode_stats[i][j][k]);
-                    }
+    {
+      int i, j, k;
+      FILE *fmode = fopen("modecontext.c", "w");
 
-                    fprintf(fmode, "}, /* left_mode %d */\n", j);
+      fprintf(fmode, "\n#include \"entropymode.h\"\n\n");
+      fprintf(fmode, "const unsigned int vp8_kf_default_bmode_counts ");
+      fprintf(fmode,
+              "[VP8_BINTRAMODES] [VP8_BINTRAMODES] [VP8_BINTRAMODES] =\n{\n");
 
-                }
+      for (i = 0; i < 10; i++) {
+        fprintf(fmode, "    { /* Above Mode :  %d */\n", i);
 
-                fprintf(fmode, "    },\n");
+        for (j = 0; j < 10; j++) {
+          fprintf(fmode, "        {");
 
-            }
+          for (k = 0; k < 10; k++) {
+            if (!intra_mode_stats[i][j][k])
+              fprintf(fmode, " %5d, ", 1);
+            else
+              fprintf(fmode, " %5d, ", intra_mode_stats[i][j][k]);
+          }
 
-            fprintf(fmode, "};\n");
-            fclose(fmode);
+          fprintf(fmode, "}, /* left_mode %d */\n", j);
         }
-#endif
 
+        fprintf(fmode, "    },\n");
+      }
 
+      fprintf(fmode, "};\n");
+      fclose(fmode);
+    }
+#endif
+
 #if defined(SECTIONBITS_OUTPUT)
 
-        if (0)
-        {
-            int i;
-            FILE *f = fopen("tokenbits.stt", "a");
+    if (0) {
+      int i;
+      FILE *f = fopen("tokenbits.stt", "a");
 
-            for (i = 0; i < 28; i++)
-                fprintf(f, "%8d", (int)(Sectionbits[i] / 256));
+      for (i = 0; i < 28; i++) fprintf(f, "%8d", (int)(Sectionbits[i] / 256));
 
-            fprintf(f, "\n");
-            fclose(f);
-        }
+      fprintf(f, "\n");
+      fclose(f);
+    }
 
 #endif
 
@@ -2442,32 +2221,31 @@
             printf("%6d %10ld %10ld %10ld %10ld\n", cpi->common.current_video_frame, cpi->time_receive_data / 1000, cpi->time_encode_mb_row / 1000, cpi->time_compress_data / 1000, (cpi->time_receive_data + cpi->time_compress_data) / 1000);
         }
 #endif
+  }
 
-    }
-
 #if CONFIG_MULTITHREAD
-    vp8cx_remove_encoder_threads(cpi);
+  vp8cx_remove_encoder_threads(cpi);
 #endif
 
 #if CONFIG_TEMPORAL_DENOISING
-    vp8_denoiser_free(&cpi->denoiser);
+  vp8_denoiser_free(&cpi->denoiser);
 #endif
-    dealloc_compressor_data(cpi);
-    vpx_free(cpi->mb.ss);
-    vpx_free(cpi->tok);
-    vpx_free(cpi->cyclic_refresh_map);
-    vpx_free(cpi->consec_zero_last);
-    vpx_free(cpi->consec_zero_last_mvbias);
+  dealloc_compressor_data(cpi);
+  vpx_free(cpi->mb.ss);
+  vpx_free(cpi->tok);
+  vpx_free(cpi->cyclic_refresh_map);
+  vpx_free(cpi->consec_zero_last);
+  vpx_free(cpi->consec_zero_last_mvbias);
 
-    vp8_remove_common(&cpi->common);
-    vpx_free(cpi);
-    *ptr = 0;
+  vp8_remove_common(&cpi->common);
+  vpx_free(cpi);
+  *ptr = 0;
 
 #ifdef OUTPUT_YUV_SRC
-    fclose(yuv_file);
+  fclose(yuv_file);
 #endif
 #ifdef OUTPUT_YUV_DENOISED
-    fclose(yuv_denoised_file);
+  fclose(yuv_denoised_file);
 #endif
 
 #if 0
@@ -2482,579 +2260,508 @@
         fclose(kf_list);
 
 #endif
-
 }
 
-
 static uint64_t calc_plane_error(unsigned char *orig, int orig_stride,
                                  unsigned char *recon, int recon_stride,
-                                 unsigned int cols, unsigned int rows)
-{
-    unsigned int row, col;
-    uint64_t total_sse = 0;
-    int diff;
+                                 unsigned int cols, unsigned int rows) {
+  unsigned int row, col;
+  uint64_t total_sse = 0;
+  int diff;
 
-    for (row = 0; row + 16 <= rows; row += 16)
-    {
-        for (col = 0; col + 16 <= cols; col += 16)
-        {
-            unsigned int sse;
+  for (row = 0; row + 16 <= rows; row += 16) {
+    for (col = 0; col + 16 <= cols; col += 16) {
+      unsigned int sse;
 
-            vpx_mse16x16(orig + col, orig_stride,
-                                            recon + col, recon_stride,
-                                            &sse);
-            total_sse += sse;
-        }
+      vpx_mse16x16(orig + col, orig_stride, recon + col, recon_stride, &sse);
+      total_sse += sse;
+    }
 
-        /* Handle odd-sized width */
-        if (col < cols)
-        {
-            unsigned int   border_row, border_col;
-            unsigned char *border_orig = orig;
-            unsigned char *border_recon = recon;
+    /* Handle odd-sized width */
+    if (col < cols) {
+      unsigned int border_row, border_col;
+      unsigned char *border_orig = orig;
+      unsigned char *border_recon = recon;
 
-            for (border_row = 0; border_row < 16; border_row++)
-            {
-                for (border_col = col; border_col < cols; border_col++)
-                {
-                    diff = border_orig[border_col] - border_recon[border_col];
-                    total_sse += diff * diff;
-                }
-
-                border_orig += orig_stride;
-                border_recon += recon_stride;
-            }
+      for (border_row = 0; border_row < 16; border_row++) {
+        for (border_col = col; border_col < cols; border_col++) {
+          diff = border_orig[border_col] - border_recon[border_col];
+          total_sse += diff * diff;
         }
 
-        orig += orig_stride * 16;
-        recon += recon_stride * 16;
+        border_orig += orig_stride;
+        border_recon += recon_stride;
+      }
     }
 
-    /* Handle odd-sized height */
-    for (; row < rows; row++)
-    {
-        for (col = 0; col < cols; col++)
-        {
-            diff = orig[col] - recon[col];
-            total_sse += diff * diff;
-        }
+    orig += orig_stride * 16;
+    recon += recon_stride * 16;
+  }
 
-        orig += orig_stride;
-        recon += recon_stride;
+  /* Handle odd-sized height */
+  for (; row < rows; row++) {
+    for (col = 0; col < cols; col++) {
+      diff = orig[col] - recon[col];
+      total_sse += diff * diff;
     }
 
-    vp8_clear_system_state();
-    return total_sse;
+    orig += orig_stride;
+    recon += recon_stride;
+  }
+
+  vp8_clear_system_state();
+  return total_sse;
 }
 
+static void generate_psnr_packet(VP8_COMP *cpi) {
+  YV12_BUFFER_CONFIG *orig = cpi->Source;
+  YV12_BUFFER_CONFIG *recon = cpi->common.frame_to_show;
+  struct vpx_codec_cx_pkt pkt;
+  uint64_t sse;
+  int i;
+  unsigned int width = cpi->common.Width;
+  unsigned int height = cpi->common.Height;
 
-static void generate_psnr_packet(VP8_COMP *cpi)
-{
-    YV12_BUFFER_CONFIG      *orig = cpi->Source;
-    YV12_BUFFER_CONFIG      *recon = cpi->common.frame_to_show;
-    struct vpx_codec_cx_pkt  pkt;
-    uint64_t                 sse;
-    int                      i;
-    unsigned int             width = cpi->common.Width;
-    unsigned int             height = cpi->common.Height;
+  pkt.kind = VPX_CODEC_PSNR_PKT;
+  sse = calc_plane_error(orig->y_buffer, orig->y_stride, recon->y_buffer,
+                         recon->y_stride, width, height);
+  pkt.data.psnr.sse[0] = sse;
+  pkt.data.psnr.sse[1] = sse;
+  pkt.data.psnr.samples[0] = width * height;
+  pkt.data.psnr.samples[1] = width * height;
 
-    pkt.kind = VPX_CODEC_PSNR_PKT;
-    sse = calc_plane_error(orig->y_buffer, orig->y_stride,
-                           recon->y_buffer, recon->y_stride,
-                           width, height);
-    pkt.data.psnr.sse[0] = sse;
-    pkt.data.psnr.sse[1] = sse;
-    pkt.data.psnr.samples[0] = width * height;
-    pkt.data.psnr.samples[1] = width * height;
+  width = (width + 1) / 2;
+  height = (height + 1) / 2;
 
-    width = (width + 1) / 2;
-    height = (height + 1) / 2;
+  sse = calc_plane_error(orig->u_buffer, orig->uv_stride, recon->u_buffer,
+                         recon->uv_stride, width, height);
+  pkt.data.psnr.sse[0] += sse;
+  pkt.data.psnr.sse[2] = sse;
+  pkt.data.psnr.samples[0] += width * height;
+  pkt.data.psnr.samples[2] = width * height;
 
-    sse = calc_plane_error(orig->u_buffer, orig->uv_stride,
-                           recon->u_buffer, recon->uv_stride,
-                           width, height);
-    pkt.data.psnr.sse[0] += sse;
-    pkt.data.psnr.sse[2] = sse;
-    pkt.data.psnr.samples[0] += width * height;
-    pkt.data.psnr.samples[2] = width * height;
+  sse = calc_plane_error(orig->v_buffer, orig->uv_stride, recon->v_buffer,
+                         recon->uv_stride, width, height);
+  pkt.data.psnr.sse[0] += sse;
+  pkt.data.psnr.sse[3] = sse;
+  pkt.data.psnr.samples[0] += width * height;
+  pkt.data.psnr.samples[3] = width * height;
 
-    sse = calc_plane_error(orig->v_buffer, orig->uv_stride,
-                           recon->v_buffer, recon->uv_stride,
-                           width, height);
-    pkt.data.psnr.sse[0] += sse;
-    pkt.data.psnr.sse[3] = sse;
-    pkt.data.psnr.samples[0] += width * height;
-    pkt.data.psnr.samples[3] = width * height;
+  for (i = 0; i < 4; i++)
+    pkt.data.psnr.psnr[i] = vpx_sse_to_psnr(pkt.data.psnr.samples[i], 255.0,
+                                            (double)(pkt.data.psnr.sse[i]));
 
-    for (i = 0; i < 4; i++)
-        pkt.data.psnr.psnr[i] = vpx_sse_to_psnr(pkt.data.psnr.samples[i], 255.0,
-                                                (double)(pkt.data.psnr.sse[i]));
-
-    vpx_codec_pkt_list_add(cpi->output_pkt_list, &pkt);
+  vpx_codec_pkt_list_add(cpi->output_pkt_list, &pkt);
 }
 
+int vp8_use_as_reference(VP8_COMP *cpi, int ref_frame_flags) {
+  if (ref_frame_flags > 7) return -1;
 
-int vp8_use_as_reference(VP8_COMP *cpi, int ref_frame_flags)
-{
-    if (ref_frame_flags > 7)
-        return -1 ;
-
-    cpi->ref_frame_flags = ref_frame_flags;
-    return 0;
+  cpi->ref_frame_flags = ref_frame_flags;
+  return 0;
 }
-int vp8_update_reference(VP8_COMP *cpi, int ref_frame_flags)
-{
-    if (ref_frame_flags > 7)
-        return -1 ;
+int vp8_update_reference(VP8_COMP *cpi, int ref_frame_flags) {
+  if (ref_frame_flags > 7) return -1;
 
-    cpi->common.refresh_golden_frame = 0;
-    cpi->common.refresh_alt_ref_frame = 0;
-    cpi->common.refresh_last_frame   = 0;
+  cpi->common.refresh_golden_frame = 0;
+  cpi->common.refresh_alt_ref_frame = 0;
+  cpi->common.refresh_last_frame = 0;
 
-    if (ref_frame_flags & VP8_LAST_FRAME)
-        cpi->common.refresh_last_frame = 1;
+  if (ref_frame_flags & VP8_LAST_FRAME) cpi->common.refresh_last_frame = 1;
 
-    if (ref_frame_flags & VP8_GOLD_FRAME)
-        cpi->common.refresh_golden_frame = 1;
+  if (ref_frame_flags & VP8_GOLD_FRAME) cpi->common.refresh_golden_frame = 1;
 
-    if (ref_frame_flags & VP8_ALTR_FRAME)
-        cpi->common.refresh_alt_ref_frame = 1;
+  if (ref_frame_flags & VP8_ALTR_FRAME) cpi->common.refresh_alt_ref_frame = 1;
 
-    return 0;
+  return 0;
 }
 
-int vp8_get_reference(VP8_COMP *cpi, enum vpx_ref_frame_type ref_frame_flag, YV12_BUFFER_CONFIG *sd)
-{
-    VP8_COMMON *cm = &cpi->common;
-    int ref_fb_idx;
+int vp8_get_reference(VP8_COMP *cpi, enum vpx_ref_frame_type ref_frame_flag,
+                      YV12_BUFFER_CONFIG *sd) {
+  VP8_COMMON *cm = &cpi->common;
+  int ref_fb_idx;
 
-    if (ref_frame_flag == VP8_LAST_FRAME)
-        ref_fb_idx = cm->lst_fb_idx;
-    else if (ref_frame_flag == VP8_GOLD_FRAME)
-        ref_fb_idx = cm->gld_fb_idx;
-    else if (ref_frame_flag == VP8_ALTR_FRAME)
-        ref_fb_idx = cm->alt_fb_idx;
-    else
-        return -1;
+  if (ref_frame_flag == VP8_LAST_FRAME)
+    ref_fb_idx = cm->lst_fb_idx;
+  else if (ref_frame_flag == VP8_GOLD_FRAME)
+    ref_fb_idx = cm->gld_fb_idx;
+  else if (ref_frame_flag == VP8_ALTR_FRAME)
+    ref_fb_idx = cm->alt_fb_idx;
+  else
+    return -1;
 
-    vp8_yv12_copy_frame(&cm->yv12_fb[ref_fb_idx], sd);
+  vp8_yv12_copy_frame(&cm->yv12_fb[ref_fb_idx], sd);
 
-    return 0;
+  return 0;
 }
-int vp8_set_reference(VP8_COMP *cpi, enum vpx_ref_frame_type ref_frame_flag, YV12_BUFFER_CONFIG *sd)
-{
-    VP8_COMMON *cm = &cpi->common;
+int vp8_set_reference(VP8_COMP *cpi, enum vpx_ref_frame_type ref_frame_flag,
+                      YV12_BUFFER_CONFIG *sd) {
+  VP8_COMMON *cm = &cpi->common;
 
-    int ref_fb_idx;
+  int ref_fb_idx;
 
-    if (ref_frame_flag == VP8_LAST_FRAME)
-        ref_fb_idx = cm->lst_fb_idx;
-    else if (ref_frame_flag == VP8_GOLD_FRAME)
-        ref_fb_idx = cm->gld_fb_idx;
-    else if (ref_frame_flag == VP8_ALTR_FRAME)
-        ref_fb_idx = cm->alt_fb_idx;
-    else
-        return -1;
+  if (ref_frame_flag == VP8_LAST_FRAME)
+    ref_fb_idx = cm->lst_fb_idx;
+  else if (ref_frame_flag == VP8_GOLD_FRAME)
+    ref_fb_idx = cm->gld_fb_idx;
+  else if (ref_frame_flag == VP8_ALTR_FRAME)
+    ref_fb_idx = cm->alt_fb_idx;
+  else
+    return -1;
 
-    vp8_yv12_copy_frame(sd, &cm->yv12_fb[ref_fb_idx]);
+  vp8_yv12_copy_frame(sd, &cm->yv12_fb[ref_fb_idx]);
 
-    return 0;
+  return 0;
 }
-int vp8_update_entropy(VP8_COMP *cpi, int update)
-{
-    VP8_COMMON *cm = &cpi->common;
-    cm->refresh_entropy_probs = update;
+int vp8_update_entropy(VP8_COMP *cpi, int update) {
+  VP8_COMMON *cm = &cpi->common;
+  cm->refresh_entropy_probs = update;
 
-    return 0;
+  return 0;
 }
 
-
 #if defined(OUTPUT_YUV_SRC) || defined(OUTPUT_YUV_DENOISED)
-void vp8_write_yuv_frame(FILE *yuv_file, YV12_BUFFER_CONFIG *s)
-{
-    unsigned char *src = s->y_buffer;
-    int h = s->y_height;
+void vp8_write_yuv_frame(FILE *yuv_file, YV12_BUFFER_CONFIG *s) {
+  unsigned char *src = s->y_buffer;
+  int h = s->y_height;
 
-    do
-    {
-        fwrite(src, s->y_width, 1,  yuv_file);
-        src += s->y_stride;
-    }
-    while (--h);
+  do {
+    fwrite(src, s->y_width, 1, yuv_file);
+    src += s->y_stride;
+  } while (--h);
 
-    src = s->u_buffer;
-    h = s->uv_height;
+  src = s->u_buffer;
+  h = s->uv_height;
 
-    do
-    {
-        fwrite(src, s->uv_width, 1,  yuv_file);
-        src += s->uv_stride;
-    }
-    while (--h);
+  do {
+    fwrite(src, s->uv_width, 1, yuv_file);
+    src += s->uv_stride;
+  } while (--h);
 
-    src = s->v_buffer;
-    h = s->uv_height;
+  src = s->v_buffer;
+  h = s->uv_height;
 
-    do
-    {
-        fwrite(src, s->uv_width, 1, yuv_file);
-        src += s->uv_stride;
-    }
-    while (--h);
+  do {
+    fwrite(src, s->uv_width, 1, yuv_file);
+    src += s->uv_stride;
+  } while (--h);
 }
 #endif
 
-static void scale_and_extend_source(YV12_BUFFER_CONFIG *sd, VP8_COMP *cpi)
-{
-    VP8_COMMON *cm = &cpi->common;
+static void scale_and_extend_source(YV12_BUFFER_CONFIG *sd, VP8_COMP *cpi) {
+  VP8_COMMON *cm = &cpi->common;
 
-    /* are we resizing the image */
-    if (cm->horiz_scale != 0 || cm->vert_scale != 0)
-    {
+  /* are we resizing the image */
+  if (cm->horiz_scale != 0 || cm->vert_scale != 0) {
 #if CONFIG_SPATIAL_RESAMPLING
-        int UNINITIALIZED_IS_SAFE(hr), UNINITIALIZED_IS_SAFE(hs);
-        int UNINITIALIZED_IS_SAFE(vr), UNINITIALIZED_IS_SAFE(vs);
-        int tmp_height;
+    int UNINITIALIZED_IS_SAFE(hr), UNINITIALIZED_IS_SAFE(hs);
+    int UNINITIALIZED_IS_SAFE(vr), UNINITIALIZED_IS_SAFE(vs);
+    int tmp_height;
 
-        if (cm->vert_scale == 3)
-            tmp_height = 9;
-        else
-            tmp_height = 11;
+    if (cm->vert_scale == 3)
+      tmp_height = 9;
+    else
+      tmp_height = 11;
 
-        Scale2Ratio(cm->horiz_scale, &hr, &hs);
-        Scale2Ratio(cm->vert_scale, &vr, &vs);
+    Scale2Ratio(cm->horiz_scale, &hr, &hs);
+    Scale2Ratio(cm->vert_scale, &vr, &vs);
 
-        vpx_scale_frame(sd, &cpi->scaled_source, cm->temp_scale_frame.y_buffer,
-                        tmp_height, hs, hr, vs, vr, 0);
+    vpx_scale_frame(sd, &cpi->scaled_source, cm->temp_scale_frame.y_buffer,
+                    tmp_height, hs, hr, vs, vr, 0);
 
-        vp8_yv12_extend_frame_borders(&cpi->scaled_source);
-        cpi->Source = &cpi->scaled_source;
+    vp8_yv12_extend_frame_borders(&cpi->scaled_source);
+    cpi->Source = &cpi->scaled_source;
 #endif
-    }
-    else
-        cpi->Source = sd;
+  } else
+    cpi->Source = sd;
 }
 
-
-static int resize_key_frame(VP8_COMP *cpi)
-{
+static int resize_key_frame(VP8_COMP *cpi) {
 #if CONFIG_SPATIAL_RESAMPLING
-    VP8_COMMON *cm = &cpi->common;
+  VP8_COMMON *cm = &cpi->common;
 
-    /* Do we need to apply resampling for one pass cbr.
-     * In one pass this is more limited than in two pass cbr.
-     * The test and any change is only made once per key frame sequence.
+  /* Do we need to apply resampling for one pass cbr.
+   * In one pass this is more limited than in two pass cbr.
+   * The test and any change is only made once per key frame sequence.
+   */
+  if (cpi->oxcf.allow_spatial_resampling &&
+      (cpi->oxcf.end_usage == USAGE_STREAM_FROM_SERVER)) {
+    int UNINITIALIZED_IS_SAFE(hr), UNINITIALIZED_IS_SAFE(hs);
+    int UNINITIALIZED_IS_SAFE(vr), UNINITIALIZED_IS_SAFE(vs);
+    int new_width, new_height;
+
+    /* If we are below the resample DOWN watermark then scale down a
+     * notch.
      */
-    if (cpi->oxcf.allow_spatial_resampling && (cpi->oxcf.end_usage == USAGE_STREAM_FROM_SERVER))
-    {
-        int UNINITIALIZED_IS_SAFE(hr), UNINITIALIZED_IS_SAFE(hs);
-        int UNINITIALIZED_IS_SAFE(vr), UNINITIALIZED_IS_SAFE(vs);
-        int new_width, new_height;
+    if (cpi->buffer_level < (cpi->oxcf.resample_down_water_mark *
+                             cpi->oxcf.optimal_buffer_level / 100)) {
+      cm->horiz_scale =
+          (cm->horiz_scale < ONETWO) ? cm->horiz_scale + 1 : ONETWO;
+      cm->vert_scale = (cm->vert_scale < ONETWO) ? cm->vert_scale + 1 : ONETWO;
+    }
+    /* Should we now start scaling back up */
+    else if (cpi->buffer_level > (cpi->oxcf.resample_up_water_mark *
+                                  cpi->oxcf.optimal_buffer_level / 100)) {
+      cm->horiz_scale =
+          (cm->horiz_scale > NORMAL) ? cm->horiz_scale - 1 : NORMAL;
+      cm->vert_scale = (cm->vert_scale > NORMAL) ? cm->vert_scale - 1 : NORMAL;
+    }
 
-        /* If we are below the resample DOWN watermark then scale down a
-         * notch.
-         */
-        if (cpi->buffer_level < (cpi->oxcf.resample_down_water_mark * cpi->oxcf.optimal_buffer_level / 100))
-        {
-            cm->horiz_scale = (cm->horiz_scale < ONETWO) ? cm->horiz_scale + 1 : ONETWO;
-            cm->vert_scale = (cm->vert_scale < ONETWO) ? cm->vert_scale + 1 : ONETWO;
-        }
-        /* Should we now start scaling back up */
-        else if (cpi->buffer_level > (cpi->oxcf.resample_up_water_mark * cpi->oxcf.optimal_buffer_level / 100))
-        {
-            cm->horiz_scale = (cm->horiz_scale > NORMAL) ? cm->horiz_scale - 1 : NORMAL;
-            cm->vert_scale = (cm->vert_scale > NORMAL) ? cm->vert_scale - 1 : NORMAL;
-        }
+    /* Get the new height and width */
+    Scale2Ratio(cm->horiz_scale, &hr, &hs);
+    Scale2Ratio(cm->vert_scale, &vr, &vs);
+    new_width = ((hs - 1) + (cpi->oxcf.Width * hr)) / hs;
+    new_height = ((vs - 1) + (cpi->oxcf.Height * vr)) / vs;
 
-        /* Get the new height and width */
-        Scale2Ratio(cm->horiz_scale, &hr, &hs);
-        Scale2Ratio(cm->vert_scale, &vr, &vs);
-        new_width = ((hs - 1) + (cpi->oxcf.Width * hr)) / hs;
-        new_height = ((vs - 1) + (cpi->oxcf.Height * vr)) / vs;
-
-        /* If the image size has changed we need to reallocate the buffers
-         * and resample the source image
-         */
-        if ((cm->Width != new_width) || (cm->Height != new_height))
-        {
-            cm->Width = new_width;
-            cm->Height = new_height;
-            vp8_alloc_compressor_data(cpi);
-            scale_and_extend_source(cpi->un_scaled_source, cpi);
-            return 1;
-        }
+    /* If the image size has changed we need to reallocate the buffers
+     * and resample the source image
+     */
+    if ((cm->Width != new_width) || (cm->Height != new_height)) {
+      cm->Width = new_width;
+      cm->Height = new_height;
+      vp8_alloc_compressor_data(cpi);
+      scale_and_extend_source(cpi->un_scaled_source, cpi);
+      return 1;
     }
+  }
 
 #endif
-    return 0;
+  return 0;
 }
 
+static void update_alt_ref_frame_stats(VP8_COMP *cpi) {
+  VP8_COMMON *cm = &cpi->common;
 
-static void update_alt_ref_frame_stats(VP8_COMP *cpi)
-{
-    VP8_COMMON *cm = &cpi->common;
+  /* Select an interval before next GF or altref */
+  if (!cpi->auto_gold) cpi->frames_till_gf_update_due = DEFAULT_GF_INTERVAL;
 
-    /* Select an interval before next GF or altref */
-    if (!cpi->auto_gold)
-        cpi->frames_till_gf_update_due = DEFAULT_GF_INTERVAL;
+  if ((cpi->pass != 2) && cpi->frames_till_gf_update_due) {
+    cpi->current_gf_interval = cpi->frames_till_gf_update_due;
 
-    if ((cpi->pass != 2) && cpi->frames_till_gf_update_due)
-    {
-        cpi->current_gf_interval = cpi->frames_till_gf_update_due;
+    /* Set the bits per frame that we should try and recover in
+     * subsequent inter frames to account for the extra GF spend...
+     * note that his does not apply for GF updates that occur
+     * coincident with a key frame as the extra cost of key frames is
+     * dealt with elsewhere.
+     */
+    cpi->gf_overspend_bits += cpi->projected_frame_size;
+    cpi->non_gf_bitrate_adjustment =
+        cpi->gf_overspend_bits / cpi->frames_till_gf_update_due;
+  }
 
-        /* Set the bits per frame that we should try and recover in
-         * subsequent inter frames to account for the extra GF spend...
-         * note that his does not apply for GF updates that occur
-         * coincident with a key frame as the extra cost of key frames is
-         * dealt with elsewhere.
-         */
-        cpi->gf_overspend_bits += cpi->projected_frame_size;
-        cpi->non_gf_bitrate_adjustment = cpi->gf_overspend_bits / cpi->frames_till_gf_update_due;
-    }
+  /* Update data structure that monitors level of reference to last GF */
+  memset(cpi->gf_active_flags, 1, (cm->mb_rows * cm->mb_cols));
+  cpi->gf_active_count = cm->mb_rows * cm->mb_cols;
 
-    /* Update data structure that monitors level of reference to last GF */
-    memset(cpi->gf_active_flags, 1, (cm->mb_rows * cm->mb_cols));
-    cpi->gf_active_count = cm->mb_rows * cm->mb_cols;
+  /* this frame refreshes means next frames don't unless specified by user */
+  cpi->frames_since_golden = 0;
 
-    /* this frame refreshes means next frames don't unless specified by user */
-    cpi->frames_since_golden = 0;
+  /* Clear the alternate reference update pending flag. */
+  cpi->source_alt_ref_pending = 0;
 
-    /* Clear the alternate reference update pending flag. */
-    cpi->source_alt_ref_pending = 0;
-
-    /* Set the alternate reference frame active flag */
-    cpi->source_alt_ref_active = 1;
-
-
+  /* Set the alternate reference frame active flag */
+  cpi->source_alt_ref_active = 1;
 }
-static void update_golden_frame_stats(VP8_COMP *cpi)
-{
-    VP8_COMMON *cm = &cpi->common;
+static void update_golden_frame_stats(VP8_COMP *cpi) {
+  VP8_COMMON *cm = &cpi->common;
 
-    /* Update the Golden frame usage counts. */
-    if (cm->refresh_golden_frame)
-    {
-        /* Select an interval before next GF */
-        if (!cpi->auto_gold)
-            cpi->frames_till_gf_update_due = DEFAULT_GF_INTERVAL;
+  /* Update the Golden frame usage counts. */
+  if (cm->refresh_golden_frame) {
+    /* Select an interval before next GF */
+    if (!cpi->auto_gold) cpi->frames_till_gf_update_due = DEFAULT_GF_INTERVAL;
 
-        if ((cpi->pass != 2) && (cpi->frames_till_gf_update_due > 0))
-        {
-            cpi->current_gf_interval = cpi->frames_till_gf_update_due;
+    if ((cpi->pass != 2) && (cpi->frames_till_gf_update_due > 0)) {
+      cpi->current_gf_interval = cpi->frames_till_gf_update_due;
 
-            /* Set the bits per frame that we should try and recover in
-             * subsequent inter frames to account for the extra GF spend...
-             * note that his does not apply for GF updates that occur
-             * coincident with a key frame as the extra cost of key frames
-             * is dealt with elsewhere.
-             */
-            if ((cm->frame_type != KEY_FRAME) && !cpi->source_alt_ref_active)
-            {
-                /* Calcluate GF bits to be recovered
-                 * Projected size - av frame bits available for inter
-                 * frames for clip as a whole
-                 */
-                cpi->gf_overspend_bits += (cpi->projected_frame_size - cpi->inter_frame_target);
-            }
+      /* Set the bits per frame that we should try and recover in
+       * subsequent inter frames to account for the extra GF spend...
+       * note that his does not apply for GF updates that occur
+       * coincident with a key frame as the extra cost of key frames
+       * is dealt with elsewhere.
+       */
+      if ((cm->frame_type != KEY_FRAME) && !cpi->source_alt_ref_active) {
+        /* Calcluate GF bits to be recovered
+         * Projected size - av frame bits available for inter
+         * frames for clip as a whole
+         */
+        cpi->gf_overspend_bits +=
+            (cpi->projected_frame_size - cpi->inter_frame_target);
+      }
 
-            cpi->non_gf_bitrate_adjustment = cpi->gf_overspend_bits / cpi->frames_till_gf_update_due;
+      cpi->non_gf_bitrate_adjustment =
+          cpi->gf_overspend_bits / cpi->frames_till_gf_update_due;
+    }
 
-        }
+    /* Update data structure that monitors level of reference to last GF */
+    memset(cpi->gf_active_flags, 1, (cm->mb_rows * cm->mb_cols));
+    cpi->gf_active_count = cm->mb_rows * cm->mb_cols;
 
-        /* Update data structure that monitors level of reference to last GF */
-        memset(cpi->gf_active_flags, 1, (cm->mb_rows * cm->mb_cols));
-        cpi->gf_active_count = cm->mb_rows * cm->mb_cols;
+    /* this frame refreshes means next frames don't unless specified by
+     * user
+     */
+    cm->refresh_golden_frame = 0;
+    cpi->frames_since_golden = 0;
 
-        /* this frame refreshes means next frames don't unless specified by
-         * user
-         */
-        cm->refresh_golden_frame = 0;
-        cpi->frames_since_golden = 0;
+    cpi->recent_ref_frame_usage[INTRA_FRAME] = 1;
+    cpi->recent_ref_frame_usage[LAST_FRAME] = 1;
+    cpi->recent_ref_frame_usage[GOLDEN_FRAME] = 1;
+    cpi->recent_ref_frame_usage[ALTREF_FRAME] = 1;
 
-        cpi->recent_ref_frame_usage[INTRA_FRAME] = 1;
-        cpi->recent_ref_frame_usage[LAST_FRAME] = 1;
-        cpi->recent_ref_frame_usage[GOLDEN_FRAME] = 1;
-        cpi->recent_ref_frame_usage[ALTREF_FRAME] = 1;
+    /* ******** Fixed Q test code only ************ */
+    /* If we are going to use the ALT reference for the next group of
+     * frames set a flag to say so.
+     */
+    if (cpi->oxcf.fixed_q >= 0 && cpi->oxcf.play_alternate &&
+        !cpi->common.refresh_alt_ref_frame) {
+      cpi->source_alt_ref_pending = 1;
+      cpi->frames_till_gf_update_due = cpi->baseline_gf_interval;
+    }
 
-        /* ******** Fixed Q test code only ************ */
-        /* If we are going to use the ALT reference for the next group of
-         * frames set a flag to say so.
-         */
-        if (cpi->oxcf.fixed_q >= 0 &&
-            cpi->oxcf.play_alternate && !cpi->common.refresh_alt_ref_frame)
-        {
-            cpi->source_alt_ref_pending = 1;
-            cpi->frames_till_gf_update_due = cpi->baseline_gf_interval;
-        }
+    if (!cpi->source_alt_ref_pending) cpi->source_alt_ref_active = 0;
 
-        if (!cpi->source_alt_ref_pending)
-            cpi->source_alt_ref_active = 0;
+    /* Decrement count down till next gf */
+    if (cpi->frames_till_gf_update_due > 0) cpi->frames_till_gf_update_due--;
 
-        /* Decrement count down till next gf */
-        if (cpi->frames_till_gf_update_due > 0)
-            cpi->frames_till_gf_update_due--;
+  } else if (!cpi->common.refresh_alt_ref_frame) {
+    /* Decrement count down till next gf */
+    if (cpi->frames_till_gf_update_due > 0) cpi->frames_till_gf_update_due--;
 
-    }
-    else if (!cpi->common.refresh_alt_ref_frame)
-    {
-        /* Decrement count down till next gf */
-        if (cpi->frames_till_gf_update_due > 0)
-            cpi->frames_till_gf_update_due--;
+    if (cpi->frames_till_alt_ref_frame) cpi->frames_till_alt_ref_frame--;
 
-        if (cpi->frames_till_alt_ref_frame)
-            cpi->frames_till_alt_ref_frame --;
+    cpi->frames_since_golden++;
 
-        cpi->frames_since_golden ++;
-
-        if (cpi->frames_since_golden > 1)
-        {
-            cpi->recent_ref_frame_usage[INTRA_FRAME] +=
-                cpi->mb.count_mb_ref_frame_usage[INTRA_FRAME];
-            cpi->recent_ref_frame_usage[LAST_FRAME] +=
-                cpi->mb.count_mb_ref_frame_usage[LAST_FRAME];
-            cpi->recent_ref_frame_usage[GOLDEN_FRAME] +=
-                cpi->mb.count_mb_ref_frame_usage[GOLDEN_FRAME];
-            cpi->recent_ref_frame_usage[ALTREF_FRAME] +=
-                cpi->mb.count_mb_ref_frame_usage[ALTREF_FRAME];
-        }
+    if (cpi->frames_since_golden > 1) {
+      cpi->recent_ref_frame_usage[INTRA_FRAME] +=
+          cpi->mb.count_mb_ref_frame_usage[INTRA_FRAME];
+      cpi->recent_ref_frame_usage[LAST_FRAME] +=
+          cpi->mb.count_mb_ref_frame_usage[LAST_FRAME];
+      cpi->recent_ref_frame_usage[GOLDEN_FRAME] +=
+          cpi->mb.count_mb_ref_frame_usage[GOLDEN_FRAME];
+      cpi->recent_ref_frame_usage[ALTREF_FRAME] +=
+          cpi->mb.count_mb_ref_frame_usage[ALTREF_FRAME];
     }
+  }
 }
 
 /* This function updates the reference frame probability estimates that
  * will be used during mode selection
  */
-static void update_rd_ref_frame_probs(VP8_COMP *cpi)
-{
-    VP8_COMMON *cm = &cpi->common;
+static void update_rd_ref_frame_probs(VP8_COMP *cpi) {
+  VP8_COMMON *cm = &cpi->common;
 
-    const int *const rfct = cpi->mb.count_mb_ref_frame_usage;
-    const int rf_intra = rfct[INTRA_FRAME];
-    const int rf_inter = rfct[LAST_FRAME] + rfct[GOLDEN_FRAME] + rfct[ALTREF_FRAME];
+  const int *const rfct = cpi->mb.count_mb_ref_frame_usage;
+  const int rf_intra = rfct[INTRA_FRAME];
+  const int rf_inter =
+      rfct[LAST_FRAME] + rfct[GOLDEN_FRAME] + rfct[ALTREF_FRAME];
 
-    if (cm->frame_type == KEY_FRAME)
-    {
-        cpi->prob_intra_coded = 255;
-        cpi->prob_last_coded  = 128;
-        cpi->prob_gf_coded  = 128;
-    }
-    else if (!(rf_intra + rf_inter))
-    {
-        cpi->prob_intra_coded = 63;
-        cpi->prob_last_coded  = 128;
-        cpi->prob_gf_coded    = 128;
-    }
+  if (cm->frame_type == KEY_FRAME) {
+    cpi->prob_intra_coded = 255;
+    cpi->prob_last_coded = 128;
+    cpi->prob_gf_coded = 128;
+  } else if (!(rf_intra + rf_inter)) {
+    cpi->prob_intra_coded = 63;
+    cpi->prob_last_coded = 128;
+    cpi->prob_gf_coded = 128;
+  }
 
-    /* update reference frame costs since we can do better than what we got
-     * last frame.
-     */
-    if (cpi->oxcf.number_of_layers == 1)
-    {
-        if (cpi->common.refresh_alt_ref_frame)
-        {
-            cpi->prob_intra_coded += 40;
-            if (cpi->prob_intra_coded > 255)
-                cpi->prob_intra_coded = 255;
-            cpi->prob_last_coded = 200;
-            cpi->prob_gf_coded = 1;
-        }
-        else if (cpi->frames_since_golden == 0)
-        {
-            cpi->prob_last_coded = 214;
-        }
-        else if (cpi->frames_since_golden == 1)
-        {
-            cpi->prob_last_coded = 192;
-            cpi->prob_gf_coded = 220;
-        }
-        else if (cpi->source_alt_ref_active)
-        {
-            cpi->prob_gf_coded -= 20;
+  /* update reference frame costs since we can do better than what we got
+   * last frame.
+   */
+  if (cpi->oxcf.number_of_layers == 1) {
+    if (cpi->common.refresh_alt_ref_frame) {
+      cpi->prob_intra_coded += 40;
+      if (cpi->prob_intra_coded > 255) cpi->prob_intra_coded = 255;
+      cpi->prob_last_coded = 200;
+      cpi->prob_gf_coded = 1;
+    } else if (cpi->frames_since_golden == 0) {
+      cpi->prob_last_coded = 214;
+    } else if (cpi->frames_since_golden == 1) {
+      cpi->prob_last_coded = 192;
+      cpi->prob_gf_coded = 220;
+    } else if (cpi->source_alt_ref_active) {
+      cpi->prob_gf_coded -= 20;
 
-            if (cpi->prob_gf_coded < 10)
-                cpi->prob_gf_coded = 10;
-        }
-        if (!cpi->source_alt_ref_active)
-            cpi->prob_gf_coded = 255;
+      if (cpi->prob_gf_coded < 10) cpi->prob_gf_coded = 10;
     }
+    if (!cpi->source_alt_ref_active) cpi->prob_gf_coded = 255;
+  }
 }
 
-
 #if !CONFIG_REALTIME_ONLY
 /* 1 = key, 0 = inter */
-static int decide_key_frame(VP8_COMP *cpi)
-{
-    VP8_COMMON *cm = &cpi->common;
+static int decide_key_frame(VP8_COMP *cpi) {
+  VP8_COMMON *cm = &cpi->common;
 
-    int code_key_frame = 0;
+  int code_key_frame = 0;
 
-    cpi->kf_boost = 0;
+  cpi->kf_boost = 0;
 
-    if (cpi->Speed > 11)
-        return 0;
+  if (cpi->Speed > 11) return 0;
 
-    /* Clear down mmx registers */
-    vp8_clear_system_state();
+  /* Clear down mmx registers */
+  vp8_clear_system_state();
 
-    if ((cpi->compressor_speed == 2) && (cpi->Speed >= 5) && (cpi->sf.RD == 0))
-    {
-        double change = 1.0 * abs((int)(cpi->mb.intra_error -
-            cpi->last_intra_error)) / (1 + cpi->last_intra_error);
-        double change2 = 1.0 * abs((int)(cpi->mb.prediction_error -
-            cpi->last_prediction_error)) / (1 + cpi->last_prediction_error);
-        double minerror = cm->MBs * 256;
+  if ((cpi->compressor_speed == 2) && (cpi->Speed >= 5) && (cpi->sf.RD == 0)) {
+    double change = 1.0 *
+                    abs((int)(cpi->mb.intra_error - cpi->last_intra_error)) /
+                    (1 + cpi->last_intra_error);
+    double change2 =
+        1.0 *
+        abs((int)(cpi->mb.prediction_error - cpi->last_prediction_error)) /
+        (1 + cpi->last_prediction_error);
+    double minerror = cm->MBs * 256;
 
-        cpi->last_intra_error = cpi->mb.intra_error;
-        cpi->last_prediction_error = cpi->mb.prediction_error;
+    cpi->last_intra_error = cpi->mb.intra_error;
+    cpi->last_prediction_error = cpi->mb.prediction_error;
 
-        if (10 * cpi->mb.intra_error / (1 + cpi->mb.prediction_error) < 15
-            && cpi->mb.prediction_error > minerror
-            && (change > .25 || change2 > .25))
-        {
-            /*(change > 1.4 || change < .75)&& cpi->this_frame_percent_intra > cpi->last_frame_percent_intra + 3*/
-            return 1;
-        }
-
-        return 0;
-
+    if (10 * cpi->mb.intra_error / (1 + cpi->mb.prediction_error) < 15 &&
+        cpi->mb.prediction_error > minerror &&
+        (change > .25 || change2 > .25)) {
+      /*(change > 1.4 || change < .75)&& cpi->this_frame_percent_intra >
+       * cpi->last_frame_percent_intra + 3*/
+      return 1;
     }
 
-    /* If the following are true we might as well code a key frame */
-    if (((cpi->this_frame_percent_intra == 100) &&
-         (cpi->this_frame_percent_intra > (cpi->last_frame_percent_intra + 2))) ||
-        ((cpi->this_frame_percent_intra > 95) &&
-         (cpi->this_frame_percent_intra >= (cpi->last_frame_percent_intra + 5))))
-    {
-        code_key_frame = 1;
-    }
-    /* in addition if the following are true and this is not a golden frame
-     * then code a key frame Note that on golden frames there often seems
-     * to be a pop in intra useage anyway hence this restriction is
-     * designed to prevent spurious key frames. The Intra pop needs to be
-     * investigated.
-     */
-    else if (((cpi->this_frame_percent_intra > 60) &&
-              (cpi->this_frame_percent_intra > (cpi->last_frame_percent_intra * 2))) ||
-             ((cpi->this_frame_percent_intra > 75) &&
-              (cpi->this_frame_percent_intra > (cpi->last_frame_percent_intra * 3 / 2))) ||
-             ((cpi->this_frame_percent_intra > 90) &&
-              (cpi->this_frame_percent_intra > (cpi->last_frame_percent_intra + 10))))
-    {
-        if (!cm->refresh_golden_frame)
-            code_key_frame = 1;
-    }
+    return 0;
+  }
 
-    return code_key_frame;
+  /* If the following are true we might as well code a key frame */
+  if (((cpi->this_frame_percent_intra == 100) &&
+       (cpi->this_frame_percent_intra > (cpi->last_frame_percent_intra + 2))) ||
+      ((cpi->this_frame_percent_intra > 95) &&
+       (cpi->this_frame_percent_intra >=
+        (cpi->last_frame_percent_intra + 5)))) {
+    code_key_frame = 1;
+  }
+  /* in addition if the following are true and this is not a golden frame
+   * then code a key frame Note that on golden frames there often seems
+   * to be a pop in intra useage anyway hence this restriction is
+   * designed to prevent spurious key frames. The Intra pop needs to be
+   * investigated.
+   */
+  else if (((cpi->this_frame_percent_intra > 60) &&
+            (cpi->this_frame_percent_intra >
+             (cpi->last_frame_percent_intra * 2))) ||
+           ((cpi->this_frame_percent_intra > 75) &&
+            (cpi->this_frame_percent_intra >
+             (cpi->last_frame_percent_intra * 3 / 2))) ||
+           ((cpi->this_frame_percent_intra > 90) &&
+            (cpi->this_frame_percent_intra >
+             (cpi->last_frame_percent_intra + 10)))) {
+    if (!cm->refresh_golden_frame) code_key_frame = 1;
+  }
 
+  return code_key_frame;
 }
 
-static void Pass1Encode(VP8_COMP *cpi, unsigned long *size, unsigned char *dest, unsigned int *frame_flags)
-{
-    (void) size;
-    (void) dest;
-    (void) frame_flags;
-    vp8_set_quantizer(cpi, 26);
+static void Pass1Encode(VP8_COMP *cpi, unsigned long *size, unsigned char *dest,
+                        unsigned int *frame_flags) {
+  (void)size;
+  (void)dest;
+  (void)frame_flags;
+  vp8_set_quantizer(cpi, 26);
 
-    vp8_first_pass(cpi);
+  vp8_first_pass(cpi);
 }
 #endif
 
@@ -3096,255 +2803,211 @@
 /* Function to test for conditions that indeicate we should loop
  * back and recode a frame.
  */
-static int recode_loop_test( VP8_COMP *cpi,
-                              int high_limit, int low_limit,
-                              int q, int maxq, int minq )
-{
-    int force_recode = 0;
-    VP8_COMMON *cm = &cpi->common;
+static int recode_loop_test(VP8_COMP *cpi, int high_limit, int low_limit, int q,
+                            int maxq, int minq) {
+  int force_recode = 0;
+  VP8_COMMON *cm = &cpi->common;
 
-    /* Is frame recode allowed at all
-     * Yes if either recode mode 1 is selected or mode two is selcted
-     * and the frame is a key frame. golden frame or alt_ref_frame
-     */
-    if ( (cpi->sf.recode_loop == 1) ||
-         ( (cpi->sf.recode_loop == 2) &&
-           ( (cm->frame_type == KEY_FRAME) ||
-             cm->refresh_golden_frame ||
-             cm->refresh_alt_ref_frame ) ) )
-    {
-        /* General over and under shoot tests */
-        if ( ((cpi->projected_frame_size > high_limit) && (q < maxq)) ||
-             ((cpi->projected_frame_size < low_limit) && (q > minq)) )
-        {
-            force_recode = 1;
-        }
-        /* Special Constrained quality tests */
-        else if (cpi->oxcf.end_usage == USAGE_CONSTRAINED_QUALITY)
-        {
-            /* Undershoot and below auto cq level */
-            if ( (q > cpi->cq_target_quality) &&
-                 (cpi->projected_frame_size <
-                     ((cpi->this_frame_target * 7) >> 3)))
-            {
-                force_recode = 1;
-            }
-            /* Severe undershoot and between auto and user cq level */
-            else if ( (q > cpi->oxcf.cq_level) &&
-                      (cpi->projected_frame_size < cpi->min_frame_bandwidth) &&
-                      (cpi->active_best_quality > cpi->oxcf.cq_level))
-            {
-                force_recode = 1;
-                cpi->active_best_quality = cpi->oxcf.cq_level;
-            }
-        }
+  /* Is frame recode allowed at all
+   * Yes if either recode mode 1 is selected or mode two is selcted
+   * and the frame is a key frame. golden frame or alt_ref_frame
+   */
+  if ((cpi->sf.recode_loop == 1) ||
+      ((cpi->sf.recode_loop == 2) &&
+       ((cm->frame_type == KEY_FRAME) || cm->refresh_golden_frame ||
+        cm->refresh_alt_ref_frame))) {
+    /* General over and under shoot tests */
+    if (((cpi->projected_frame_size > high_limit) && (q < maxq)) ||
+        ((cpi->projected_frame_size < low_limit) && (q > minq))) {
+      force_recode = 1;
     }
+    /* Special Constrained quality tests */
+    else if (cpi->oxcf.end_usage == USAGE_CONSTRAINED_QUALITY) {
+      /* Undershoot and below auto cq level */
+      if ((q > cpi->cq_target_quality) &&
+          (cpi->projected_frame_size < ((cpi->this_frame_target * 7) >> 3))) {
+        force_recode = 1;
+      }
+      /* Severe undershoot and between auto and user cq level */
+      else if ((q > cpi->oxcf.cq_level) &&
+               (cpi->projected_frame_size < cpi->min_frame_bandwidth) &&
+               (cpi->active_best_quality > cpi->oxcf.cq_level)) {
+        force_recode = 1;
+        cpi->active_best_quality = cpi->oxcf.cq_level;
+      }
+    }
+  }
 
-    return force_recode;
+  return force_recode;
 }
 #endif  // !CONFIG_REALTIME_ONLY
 
-static void update_reference_frames(VP8_COMP *cpi)
-{
-    VP8_COMMON *cm = &cpi->common;
-    YV12_BUFFER_CONFIG *yv12_fb = cm->yv12_fb;
+static void update_reference_frames(VP8_COMP *cpi) {
+  VP8_COMMON *cm = &cpi->common;
+  YV12_BUFFER_CONFIG *yv12_fb = cm->yv12_fb;
 
-    /* At this point the new frame has been encoded.
-     * If any buffer copy / swapping is signaled it should be done here.
-     */
+  /* At this point the new frame has been encoded.
+   * If any buffer copy / swapping is signaled it should be done here.
+   */
 
-    if (cm->frame_type == KEY_FRAME)
-    {
-        yv12_fb[cm->new_fb_idx].flags |= VP8_GOLD_FRAME | VP8_ALTR_FRAME ;
+  if (cm->frame_type == KEY_FRAME) {
+    yv12_fb[cm->new_fb_idx].flags |= VP8_GOLD_FRAME | VP8_ALTR_FRAME;
 
-        yv12_fb[cm->gld_fb_idx].flags &= ~VP8_GOLD_FRAME;
-        yv12_fb[cm->alt_fb_idx].flags &= ~VP8_ALTR_FRAME;
+    yv12_fb[cm->gld_fb_idx].flags &= ~VP8_GOLD_FRAME;
+    yv12_fb[cm->alt_fb_idx].flags &= ~VP8_ALTR_FRAME;
 
-        cm->alt_fb_idx = cm->gld_fb_idx = cm->new_fb_idx;
+    cm->alt_fb_idx = cm->gld_fb_idx = cm->new_fb_idx;
 
-        cpi->current_ref_frames[GOLDEN_FRAME] = cm->current_video_frame;
-        cpi->current_ref_frames[ALTREF_FRAME] = cm->current_video_frame;
-    }
-    else    /* For non key frames */
-    {
-        if (cm->refresh_alt_ref_frame)
-        {
-            assert(!cm->copy_buffer_to_arf);
+    cpi->current_ref_frames[GOLDEN_FRAME] = cm->current_video_frame;
+    cpi->current_ref_frames[ALTREF_FRAME] = cm->current_video_frame;
+  } else /* For non key frames */
+  {
+    if (cm->refresh_alt_ref_frame) {
+      assert(!cm->copy_buffer_to_arf);
 
-            cm->yv12_fb[cm->new_fb_idx].flags |= VP8_ALTR_FRAME;
-            cm->yv12_fb[cm->alt_fb_idx].flags &= ~VP8_ALTR_FRAME;
-            cm->alt_fb_idx = cm->new_fb_idx;
+      cm->yv12_fb[cm->new_fb_idx].flags |= VP8_ALTR_FRAME;
+      cm->yv12_fb[cm->alt_fb_idx].flags &= ~VP8_ALTR_FRAME;
+      cm->alt_fb_idx = cm->new_fb_idx;
 
-            cpi->current_ref_frames[ALTREF_FRAME] = cm->current_video_frame;
-        }
-        else if (cm->copy_buffer_to_arf)
-        {
-            assert(!(cm->copy_buffer_to_arf & ~0x3));
+      cpi->current_ref_frames[ALTREF_FRAME] = cm->current_video_frame;
+    } else if (cm->copy_buffer_to_arf) {
+      assert(!(cm->copy_buffer_to_arf & ~0x3));
 
-            if (cm->copy_buffer_to_arf == 1)
-            {
-                if(cm->alt_fb_idx != cm->lst_fb_idx)
-                {
-                    yv12_fb[cm->lst_fb_idx].flags |= VP8_ALTR_FRAME;
-                    yv12_fb[cm->alt_fb_idx].flags &= ~VP8_ALTR_FRAME;
-                    cm->alt_fb_idx = cm->lst_fb_idx;
+      if (cm->copy_buffer_to_arf == 1) {
+        if (cm->alt_fb_idx != cm->lst_fb_idx) {
+          yv12_fb[cm->lst_fb_idx].flags |= VP8_ALTR_FRAME;
+          yv12_fb[cm->alt_fb_idx].flags &= ~VP8_ALTR_FRAME;
+          cm->alt_fb_idx = cm->lst_fb_idx;
 
-                    cpi->current_ref_frames[ALTREF_FRAME] =
-                        cpi->current_ref_frames[LAST_FRAME];
-                }
-            }
-            else /* if (cm->copy_buffer_to_arf == 2) */
-            {
-                if(cm->alt_fb_idx != cm->gld_fb_idx)
-                {
-                    yv12_fb[cm->gld_fb_idx].flags |= VP8_ALTR_FRAME;
-                    yv12_fb[cm->alt_fb_idx].flags &= ~VP8_ALTR_FRAME;
-                    cm->alt_fb_idx = cm->gld_fb_idx;
+          cpi->current_ref_frames[ALTREF_FRAME] =
+              cpi->current_ref_frames[LAST_FRAME];
+        }
+      } else /* if (cm->copy_buffer_to_arf == 2) */
+      {
+        if (cm->alt_fb_idx != cm->gld_fb_idx) {
+          yv12_fb[cm->gld_fb_idx].flags |= VP8_ALTR_FRAME;
+          yv12_fb[cm->alt_fb_idx].flags &= ~VP8_ALTR_FRAME;
+          cm->alt_fb_idx = cm->gld_fb_idx;
 
-                    cpi->current_ref_frames[ALTREF_FRAME] =
-                        cpi->current_ref_frames[GOLDEN_FRAME];
-                }
-            }
+          cpi->current_ref_frames[ALTREF_FRAME] =
+              cpi->current_ref_frames[GOLDEN_FRAME];
         }
+      }
+    }
 
-        if (cm->refresh_golden_frame)
-        {
-            assert(!cm->copy_buffer_to_gf);
+    if (cm->refresh_golden_frame) {
+      assert(!cm->copy_buffer_to_gf);
 
-            cm->yv12_fb[cm->new_fb_idx].flags |= VP8_GOLD_FRAME;
-            cm->yv12_fb[cm->gld_fb_idx].flags &= ~VP8_GOLD_FRAME;
-            cm->gld_fb_idx = cm->new_fb_idx;
+      cm->yv12_fb[cm->new_fb_idx].flags |= VP8_GOLD_FRAME;
+      cm->yv12_fb[cm->gld_fb_idx].flags &= ~VP8_GOLD_FRAME;
+      cm->gld_fb_idx = cm->new_fb_idx;
 
-            cpi->current_ref_frames[GOLDEN_FRAME] = cm->current_video_frame;
-        }
-        else if (cm->copy_buffer_to_gf)
-        {
-            assert(!(cm->copy_buffer_to_arf & ~0x3));
+      cpi->current_ref_frames[GOLDEN_FRAME] = cm->current_video_frame;
+    } else if (cm->copy_buffer_to_gf) {
+      assert(!(cm->copy_buffer_to_arf & ~0x3));
 
-            if (cm->copy_buffer_to_gf == 1)
-            {
-                if(cm->gld_fb_idx != cm->lst_fb_idx)
-                {
-                    yv12_fb[cm->lst_fb_idx].flags |= VP8_GOLD_FRAME;
-                    yv12_fb[cm->gld_fb_idx].flags &= ~VP8_GOLD_FRAME;
-                    cm->gld_fb_idx = cm->lst_fb_idx;
+      if (cm->copy_buffer_to_gf == 1) {
+        if (cm->gld_fb_idx != cm->lst_fb_idx) {
+          yv12_fb[cm->lst_fb_idx].flags |= VP8_GOLD_FRAME;
+          yv12_fb[cm->gld_fb_idx].flags &= ~VP8_GOLD_FRAME;
+          cm->gld_fb_idx = cm->lst_fb_idx;
 
-                    cpi->current_ref_frames[GOLDEN_FRAME] =
-                        cpi->current_ref_frames[LAST_FRAME];
-                }
-            }
-            else /* if (cm->copy_buffer_to_gf == 2) */
-            {
-                if(cm->alt_fb_idx != cm->gld_fb_idx)
-                {
-                    yv12_fb[cm->alt_fb_idx].flags |= VP8_GOLD_FRAME;
-                    yv12_fb[cm->gld_fb_idx].flags &= ~VP8_GOLD_FRAME;
-                    cm->gld_fb_idx = cm->alt_fb_idx;
+          cpi->current_ref_frames[GOLDEN_FRAME] =
+              cpi->current_ref_frames[LAST_FRAME];
+        }
+      } else /* if (cm->copy_buffer_to_gf == 2) */
+      {
+        if (cm->alt_fb_idx != cm->gld_fb_idx) {
+          yv12_fb[cm->alt_fb_idx].flags |= VP8_GOLD_FRAME;
+          yv12_fb[cm->gld_fb_idx].flags &= ~VP8_GOLD_FRAME;
+          cm->gld_fb_idx = cm->alt_fb_idx;
 
-                    cpi->current_ref_frames[GOLDEN_FRAME] =
-                        cpi->current_ref_frames[ALTREF_FRAME];
-                }
-            }
+          cpi->current_ref_frames[GOLDEN_FRAME] =
+              cpi->current_ref_frames[ALTREF_FRAME];
         }
+      }
     }
+  }
 
-    if (cm->refresh_last_frame)
-    {
-        cm->yv12_fb[cm->new_fb_idx].flags |= VP8_LAST_FRAME;
-        cm->yv12_fb[cm->lst_fb_idx].flags &= ~VP8_LAST_FRAME;
-        cm->lst_fb_idx = cm->new_fb_idx;
+  if (cm->refresh_last_frame) {
+    cm->yv12_fb[cm->new_fb_idx].flags |= VP8_LAST_FRAME;
+    cm->yv12_fb[cm->lst_fb_idx].flags &= ~VP8_LAST_FRAME;
+    cm->lst_fb_idx = cm->new_fb_idx;
 
-        cpi->current_ref_frames[LAST_FRAME] = cm->current_video_frame;
-    }
+    cpi->current_ref_frames[LAST_FRAME] = cm->current_video_frame;
+  }
 
 #if CONFIG_TEMPORAL_DENOISING
-    if (cpi->oxcf.noise_sensitivity)
+  if (cpi->oxcf.noise_sensitivity) {
+    /* we shouldn't have to keep multiple copies as we know in advance which
+     * buffer we should start - for now to get something up and running
+     * I've chosen to copy the buffers
+     */
+    if (cm->frame_type == KEY_FRAME) {
+      int i;
+      for (i = LAST_FRAME; i < MAX_REF_FRAMES; ++i)
+        vp8_yv12_copy_frame(cpi->Source, &cpi->denoiser.yv12_running_avg[i]);
+    } else /* For non key frames */
     {
-        /* we shouldn't have to keep multiple copies as we know in advance which
-         * buffer we should start - for now to get something up and running
-         * I've chosen to copy the buffers
-         */
-        if (cm->frame_type == KEY_FRAME)
-        {
-            int i;
-            for (i = LAST_FRAME; i < MAX_REF_FRAMES; ++i)
-              vp8_yv12_copy_frame(cpi->Source,
-                                  &cpi->denoiser.yv12_running_avg[i]);
-        }
-        else /* For non key frames */
-        {
-            vp8_yv12_extend_frame_borders(
-                    &cpi->denoiser.yv12_running_avg[INTRA_FRAME]);
+      vp8_yv12_extend_frame_borders(
+          &cpi->denoiser.yv12_running_avg[INTRA_FRAME]);
 
-            if (cm->refresh_alt_ref_frame || cm->copy_buffer_to_arf)
-            {
-                vp8_yv12_copy_frame(
-                        &cpi->denoiser.yv12_running_avg[INTRA_FRAME],
-                        &cpi->denoiser.yv12_running_avg[ALTREF_FRAME]);
-            }
-            if (cm->refresh_golden_frame || cm->copy_buffer_to_gf)
-            {
-                vp8_yv12_copy_frame(
-                        &cpi->denoiser.yv12_running_avg[INTRA_FRAME],
-                        &cpi->denoiser.yv12_running_avg[GOLDEN_FRAME]);
-            }
-            if(cm->refresh_last_frame)
-            {
-                vp8_yv12_copy_frame(
-                        &cpi->denoiser.yv12_running_avg[INTRA_FRAME],
-                        &cpi->denoiser.yv12_running_avg[LAST_FRAME]);
-            }
-        }
-        if (cpi->oxcf.noise_sensitivity == 4)
-          vp8_yv12_copy_frame(cpi->Source, &cpi->denoiser.yv12_last_source);
-
+      if (cm->refresh_alt_ref_frame || cm->copy_buffer_to_arf) {
+        vp8_yv12_copy_frame(&cpi->denoiser.yv12_running_avg[INTRA_FRAME],
+                            &cpi->denoiser.yv12_running_avg[ALTREF_FRAME]);
+      }
+      if (cm->refresh_golden_frame || cm->copy_buffer_to_gf) {
+        vp8_yv12_copy_frame(&cpi->denoiser.yv12_running_avg[INTRA_FRAME],
+                            &cpi->denoiser.yv12_running_avg[GOLDEN_FRAME]);
+      }
+      if (cm->refresh_last_frame) {
+        vp8_yv12_copy_frame(&cpi->denoiser.yv12_running_avg[INTRA_FRAME],
+                            &cpi->denoiser.yv12_running_avg[LAST_FRAME]);
+      }
     }
+    if (cpi->oxcf.noise_sensitivity == 4)
+      vp8_yv12_copy_frame(cpi->Source, &cpi->denoiser.yv12_last_source);
+  }
 #endif
-
 }
 
 static int measure_square_diff_partial(YV12_BUFFER_CONFIG *source,
                                        YV12_BUFFER_CONFIG *dest,
-                                       VP8_COMP *cpi)
-    {
-        int i, j;
-        int Total = 0;
-        int num_blocks = 0;
-        int skip = 2;
-        int min_consec_zero_last = 10;
-        int tot_num_blocks = (source->y_height * source->y_width) >> 8;
-        unsigned char *src = source->y_buffer;
-        unsigned char *dst = dest->y_buffer;
+                                       VP8_COMP *cpi) {
+  int i, j;
+  int Total = 0;
+  int num_blocks = 0;
+  int skip = 2;
+  int min_consec_zero_last = 10;
+  int tot_num_blocks = (source->y_height * source->y_width) >> 8;
+  unsigned char *src = source->y_buffer;
+  unsigned char *dst = dest->y_buffer;
 
-        /* Loop through the Y plane, every |skip| blocks along rows and colmumns,
-         * summing the square differences, and only for blocks that have been
-         * zero_last mode at least |x| frames in a row.
-         */
-        for (i = 0; i < source->y_height; i += 16 * skip)
-        {
-            int block_index_row = (i >> 4) * cpi->common.mb_cols;
-            for (j = 0; j < source->y_width; j += 16 * skip)
-            {
-                int index = block_index_row + (j >> 4);
-                if (cpi->consec_zero_last[index] >= min_consec_zero_last) {
-                  unsigned int sse;
-                  Total += vpx_mse16x16(src + j,
-                                        source->y_stride,
-                                        dst + j, dest->y_stride,
-                                        &sse);
-                  num_blocks++;
-                }
-            }
-            src += 16 * skip * source->y_stride;
-            dst += 16 * skip * dest->y_stride;
-        }
-        // Only return non-zero if we have at least ~1/16 samples for estimate.
-        if (num_blocks > (tot_num_blocks >> 4)) {
-        return (Total / num_blocks);
-        } else {
-          return 0;
-        }
+  /* Loop through the Y plane, every |skip| blocks along rows and colmumns,
+   * summing the square differences, and only for blocks that have been
+   * zero_last mode at least |x| frames in a row.
+   */
+  for (i = 0; i < source->y_height; i += 16 * skip) {
+    int block_index_row = (i >> 4) * cpi->common.mb_cols;
+    for (j = 0; j < source->y_width; j += 16 * skip) {
+      int index = block_index_row + (j >> 4);
+      if (cpi->consec_zero_last[index] >= min_consec_zero_last) {
+        unsigned int sse;
+        Total += vpx_mse16x16(src + j, source->y_stride, dst + j,
+                              dest->y_stride, &sse);
+        num_blocks++;
+      }
     }
+    src += 16 * skip * source->y_stride;
+    dst += 16 * skip * dest->y_stride;
+  }
+  // Only return non-zero if we have at least ~1/16 samples for estimate.
+  if (num_blocks > (tot_num_blocks >> 4)) {
+    return (Total / num_blocks);
+  } else {
+    return 0;
+  }
+}
 
 #if CONFIG_TEMPORAL_DENOISING
 static void process_denoiser_mode_change(VP8_COMP *cpi) {
@@ -3372,9 +3035,9 @@
   int ystride = cpi->Source->y_stride;
   unsigned char *src = cpi->Source->y_buffer;
   unsigned char *dst = cpi->denoiser.yv12_last_source.y_buffer;
-  static const unsigned char const_source[16] = {
-      128, 128, 128, 128, 128, 128, 128, 128, 128, 128, 128, 128, 128,
-      128, 128, 128};
+  static const unsigned char const_source[16] = { 128, 128, 128, 128, 128, 128,
+                                                  128, 128, 128, 128, 128, 128,
+                                                  128, 128, 128, 128 };
   int bandwidth = (int)(cpi->target_bandwidth);
   // For temporal layers, use full bandwidth (top layer).
   if (cpi->oxcf.number_of_layers > 1) {
@@ -3392,23 +3055,16 @@
       int index = block_index_row + (j >> 4);
       if (cpi->consec_zero_last[index] >= min_consec_zero_last) {
         unsigned int sse;
-        const unsigned int var = vpx_variance16x16(src + j,
-                                                   ystride,
-                                                   dst + j,
-                                                   ystride,
-                                                   &sse);
+        const unsigned int var =
+            vpx_variance16x16(src + j, ystride, dst + j, ystride, &sse);
         // Only consider this block as valid for noise measurement
         // if the sum_diff average of the current and previous frame
         // is small (to avoid effects from lighting change).
         if ((sse - var) < 128) {
           unsigned int sse2;
-          const unsigned int act = vpx_variance16x16(src + j,
-                                                     ystride,
-                                                     const_source,
-                                                     0,
-                                                     &sse2);
-          if (act > 0)
-            total += sse / act;
+          const unsigned int act =
+              vpx_variance16x16(src + j, ystride, const_source, 0, &sse2);
+          if (act > 0) total += sse / act;
           num_blocks++;
         }
       }
@@ -3421,8 +3077,7 @@
   // Only consider this frame as valid sample if we have computed nmse over
   // at least ~1/16 blocks, and Total > 0 (Total == 0 can happen if the
   // application inputs duplicate frames, or contrast is all zero).
-  if (total > 0 &&
-      (num_blocks > (tot_num_blocks >> 4))) {
+  if (total > 0 && (num_blocks > (tot_num_blocks >> 4))) {
     // Update the recursive mean square source_diff.
     total = (total << 8) / num_blocks;
     if (cpi->denoiser.nmse_source_diff_count == 0) {
@@ -3431,10 +3086,10 @@
       cpi->denoiser.qp_avg = cm->base_qindex;
     } else {
       // For subsequent samples, use average with weight ~1/4 for new sample.
-      cpi->denoiser.nmse_source_diff = (int)((total +
-          3 * cpi->denoiser.nmse_source_diff) >> 2);
-      cpi->denoiser.qp_avg = (int)((cm->base_qindex +
-          3 * cpi->denoiser.qp_avg) >> 2);
+      cpi->denoiser.nmse_source_diff =
+          (int)((total + 3 * cpi->denoiser.nmse_source_diff) >> 2);
+      cpi->denoiser.qp_avg =
+          (int)((cm->base_qindex + 3 * cpi->denoiser.qp_avg) >> 2);
     }
     cpi->denoiser.nmse_source_diff_count++;
   }
@@ -3444,7 +3099,7 @@
     // Check for going up: from normal to aggressive mode.
     if ((cpi->denoiser.denoiser_mode == kDenoiserOnYUV) &&
         (cpi->denoiser.nmse_source_diff >
-        cpi->denoiser.threshold_aggressive_mode) &&
+         cpi->denoiser.threshold_aggressive_mode) &&
         (cpi->denoiser.qp_avg < cpi->denoiser.qp_threshold_up &&
          bandwidth > cpi->denoiser.bitrate_threshold)) {
       vp8_denoiser_set_parameters(&cpi->denoiser, kDenoiserOnYUVAggressive);
@@ -3451,11 +3106,11 @@
     } else {
       // Check for going down: from aggressive to normal mode.
       if (((cpi->denoiser.denoiser_mode == kDenoiserOnYUVAggressive) &&
-          (cpi->denoiser.nmse_source_diff <
-          cpi->denoiser.threshold_aggressive_mode)) ||
+           (cpi->denoiser.nmse_source_diff <
+            cpi->denoiser.threshold_aggressive_mode)) ||
           ((cpi->denoiser.denoiser_mode == kDenoiserOnYUVAggressive) &&
-          (cpi->denoiser.qp_avg > cpi->denoiser.qp_threshold_down ||
-           bandwidth < cpi->denoiser.bitrate_threshold))) {
+           (cpi->denoiser.qp_avg > cpi->denoiser.qp_threshold_down ||
+            bandwidth < cpi->denoiser.bitrate_threshold))) {
         vp8_denoiser_set_parameters(&cpi->denoiser, kDenoiserOnYUV);
       }
     }
@@ -3467,313 +3122,290 @@
 }
 #endif
 
-void vp8_loopfilter_frame(VP8_COMP *cpi, VP8_COMMON *cm)
-{
-    const FRAME_TYPE frame_type = cm->frame_type;
+void vp8_loopfilter_frame(VP8_COMP *cpi, VP8_COMMON *cm) {
+  const FRAME_TYPE frame_type = cm->frame_type;
 
-    int update_any_ref_buffers = 1;
-    if (cpi->common.refresh_last_frame == 0 &&
-        cpi->common.refresh_golden_frame == 0 &&
-        cpi->common.refresh_alt_ref_frame == 0) {
-        update_any_ref_buffers = 0;
-    }
+  int update_any_ref_buffers = 1;
+  if (cpi->common.refresh_last_frame == 0 &&
+      cpi->common.refresh_golden_frame == 0 &&
+      cpi->common.refresh_alt_ref_frame == 0) {
+    update_any_ref_buffers = 0;
+  }
 
-    if (cm->no_lpf)
-    {
-        cm->filter_level = 0;
-    }
-    else
-    {
-        struct vpx_usec_timer timer;
+  if (cm->no_lpf) {
+    cm->filter_level = 0;
+  } else {
+    struct vpx_usec_timer timer;
 
-        vp8_clear_system_state();
+    vp8_clear_system_state();
 
-        vpx_usec_timer_start(&timer);
-        if (cpi->sf.auto_filter == 0) {
+    vpx_usec_timer_start(&timer);
+    if (cpi->sf.auto_filter == 0) {
 #if CONFIG_TEMPORAL_DENOISING
-            if (cpi->oxcf.noise_sensitivity && cm->frame_type != KEY_FRAME) {
-                // Use the denoised buffer for selecting base loop filter level.
-                // Denoised signal for current frame is stored in INTRA_FRAME.
-                // No denoising on key frames.
-                vp8cx_pick_filter_level_fast(
-                    &cpi->denoiser.yv12_running_avg[INTRA_FRAME], cpi);
-            } else {
-                vp8cx_pick_filter_level_fast(cpi->Source, cpi);
-            }
+      if (cpi->oxcf.noise_sensitivity && cm->frame_type != KEY_FRAME) {
+        // Use the denoised buffer for selecting base loop filter level.
+        // Denoised signal for current frame is stored in INTRA_FRAME.
+        // No denoising on key frames.
+        vp8cx_pick_filter_level_fast(
+            &cpi->denoiser.yv12_running_avg[INTRA_FRAME], cpi);
+      } else {
+        vp8cx_pick_filter_level_fast(cpi->Source, cpi);
+      }
 #else
-            vp8cx_pick_filter_level_fast(cpi->Source, cpi);
+      vp8cx_pick_filter_level_fast(cpi->Source, cpi);
 #endif
-        } else {
+    } else {
 #if CONFIG_TEMPORAL_DENOISING
-            if (cpi->oxcf.noise_sensitivity && cm->frame_type != KEY_FRAME) {
-                // Use the denoised buffer for selecting base loop filter level.
-                // Denoised signal for current frame is stored in INTRA_FRAME.
-                // No denoising on key frames.
-                vp8cx_pick_filter_level(
-                    &cpi->denoiser.yv12_running_avg[INTRA_FRAME], cpi);
-            } else {
-                vp8cx_pick_filter_level(cpi->Source, cpi);
-            }
+      if (cpi->oxcf.noise_sensitivity && cm->frame_type != KEY_FRAME) {
+        // Use the denoised buffer for selecting base loop filter level.
+        // Denoised signal for current frame is stored in INTRA_FRAME.
+        // No denoising on key frames.
+        vp8cx_pick_filter_level(&cpi->denoiser.yv12_running_avg[INTRA_FRAME],
+                                cpi);
+      } else {
+        vp8cx_pick_filter_level(cpi->Source, cpi);
+      }
 #else
-            vp8cx_pick_filter_level(cpi->Source, cpi);
+      vp8cx_pick_filter_level(cpi->Source, cpi);
 #endif
-        }
+    }
 
-
-        if (cm->filter_level > 0)
-        {
-            vp8cx_set_alt_lf_level(cpi, cm->filter_level);
-        }
-
-        vpx_usec_timer_mark(&timer);
-        cpi->time_pick_lpf += vpx_usec_timer_elapsed(&timer);
+    if (cm->filter_level > 0) {
+      vp8cx_set_alt_lf_level(cpi, cm->filter_level);
     }
 
+    vpx_usec_timer_mark(&timer);
+    cpi->time_pick_lpf += vpx_usec_timer_elapsed(&timer);
+  }
+
 #if CONFIG_MULTITHREAD
-    if (cpi->b_multi_threaded)
-        sem_post(&cpi->h_event_end_lpf); /* signal that we have set filter_level */
+  if (cpi->b_multi_threaded)
+    sem_post(&cpi->h_event_end_lpf); /* signal that we have set filter_level */
 #endif
 
-    // No need to apply loop-filter if the encoded frame does not update
-    // any reference buffers.
-    if (cm->filter_level > 0 && update_any_ref_buffers)
-    {
-        vp8_loop_filter_frame(cm, &cpi->mb.e_mbd, frame_type);
-    }
+  // No need to apply loop-filter if the encoded frame does not update
+  // any reference buffers.
+  if (cm->filter_level > 0 && update_any_ref_buffers) {
+    vp8_loop_filter_frame(cm, &cpi->mb.e_mbd, frame_type);
+  }
 
-    vp8_yv12_extend_frame_borders(cm->frame_to_show);
-
+  vp8_yv12_extend_frame_borders(cm->frame_to_show);
 }
 
-static void encode_frame_to_data_rate
-(
-    VP8_COMP *cpi,
-    unsigned long *size,
-    unsigned char *dest,
-    unsigned char* dest_end,
-    unsigned int *frame_flags
-)
-{
-    int Q;
-    int frame_over_shoot_limit;
-    int frame_under_shoot_limit;
+static void encode_frame_to_data_rate(VP8_COMP *cpi, unsigned long *size,
+                                      unsigned char *dest,
+                                      unsigned char *dest_end,
+                                      unsigned int *frame_flags) {
+  int Q;
+  int frame_over_shoot_limit;
+  int frame_under_shoot_limit;
 
-    int Loop = 0;
-    int loop_count;
+  int Loop = 0;
+  int loop_count;
 
-    VP8_COMMON *cm = &cpi->common;
-    int active_worst_qchanged = 0;
+  VP8_COMMON *cm = &cpi->common;
+  int active_worst_qchanged = 0;
 
 #if !CONFIG_REALTIME_ONLY
-    int q_low;
-    int q_high;
-    int zbin_oq_high;
-    int zbin_oq_low = 0;
-    int top_index;
-    int bottom_index;
-    int overshoot_seen = 0;
-    int undershoot_seen = 0;
+  int q_low;
+  int q_high;
+  int zbin_oq_high;
+  int zbin_oq_low = 0;
+  int top_index;
+  int bottom_index;
+  int overshoot_seen = 0;
+  int undershoot_seen = 0;
 #endif
 
-    int drop_mark = (int)(cpi->oxcf.drop_frames_water_mark *
-                          cpi->oxcf.optimal_buffer_level / 100);
-    int drop_mark75 = drop_mark * 2 / 3;
-    int drop_mark50 = drop_mark / 4;
-    int drop_mark25 = drop_mark / 8;
+  int drop_mark = (int)(cpi->oxcf.drop_frames_water_mark *
+                        cpi->oxcf.optimal_buffer_level / 100);
+  int drop_mark75 = drop_mark * 2 / 3;
+  int drop_mark50 = drop_mark / 4;
+  int drop_mark25 = drop_mark / 8;
 
+  /* Clear down mmx registers to allow floating point in what follows */
+  vp8_clear_system_state();
 
-    /* Clear down mmx registers to allow floating point in what follows */
-    vp8_clear_system_state();
+  if (cpi->force_next_frame_intra) {
+    cm->frame_type = KEY_FRAME; /* delayed intra frame */
+    cpi->force_next_frame_intra = 0;
+  }
 
-    if(cpi->force_next_frame_intra)
-    {
-        cm->frame_type = KEY_FRAME;  /* delayed intra frame */
-        cpi->force_next_frame_intra = 0;
-    }
-
-    /* For an alt ref frame in 2 pass we skip the call to the second pass
-     * function that sets the target bandwidth
-     */
+/* For an alt ref frame in 2 pass we skip the call to the second pass
+ * function that sets the target bandwidth
+ */
 #if !CONFIG_REALTIME_ONLY
 
-    if (cpi->pass == 2)
-    {
-        if (cpi->common.refresh_alt_ref_frame)
-        {
-            /* Per frame bit target for the alt ref frame */
-            cpi->per_frame_bandwidth = cpi->twopass.gf_bits;
-            /* per second target bitrate */
-            cpi->target_bandwidth = (int)(cpi->twopass.gf_bits *
-                                          cpi->output_framerate);
-        }
+  if (cpi->pass == 2) {
+    if (cpi->common.refresh_alt_ref_frame) {
+      /* Per frame bit target for the alt ref frame */
+      cpi->per_frame_bandwidth = cpi->twopass.gf_bits;
+      /* per second target bitrate */
+      cpi->target_bandwidth =
+          (int)(cpi->twopass.gf_bits * cpi->output_framerate);
     }
-    else
+  } else
 #endif
-        cpi->per_frame_bandwidth  = (int)(cpi->target_bandwidth / cpi->output_framerate);
+    cpi->per_frame_bandwidth =
+        (int)(cpi->target_bandwidth / cpi->output_framerate);
 
-    /* Default turn off buffer to buffer copying */
-    cm->copy_buffer_to_gf = 0;
-    cm->copy_buffer_to_arf = 0;
+  /* Default turn off buffer to buffer copying */
+  cm->copy_buffer_to_gf = 0;
+  cm->copy_buffer_to_arf = 0;
 
-    /* Clear zbin over-quant value and mode boost values. */
-    cpi->mb.zbin_over_quant = 0;
-    cpi->mb.zbin_mode_boost = 0;
+  /* Clear zbin over-quant value and mode boost values. */
+  cpi->mb.zbin_over_quant = 0;
+  cpi->mb.zbin_mode_boost = 0;
 
-    /* Enable or disable mode based tweaking of the zbin
-     * For 2 Pass Only used where GF/ARF prediction quality
-     * is above a threshold
-     */
-    cpi->mb.zbin_mode_boost_enabled = 1;
-    if (cpi->pass == 2)
-    {
-        if ( cpi->gfu_boost <= 400 )
-        {
-            cpi->mb.zbin_mode_boost_enabled = 0;
-        }
+  /* Enable or disable mode based tweaking of the zbin
+   * For 2 Pass Only used where GF/ARF prediction quality
+   * is above a threshold
+   */
+  cpi->mb.zbin_mode_boost_enabled = 1;
+  if (cpi->pass == 2) {
+    if (cpi->gfu_boost <= 400) {
+      cpi->mb.zbin_mode_boost_enabled = 0;
     }
+  }
 
-    /* Current default encoder behaviour for the altref sign bias */
-    if (cpi->source_alt_ref_active)
-        cpi->common.ref_frame_sign_bias[ALTREF_FRAME] = 1;
-    else
-        cpi->common.ref_frame_sign_bias[ALTREF_FRAME] = 0;
+  /* Current default encoder behaviour for the altref sign bias */
+  if (cpi->source_alt_ref_active)
+    cpi->common.ref_frame_sign_bias[ALTREF_FRAME] = 1;
+  else
+    cpi->common.ref_frame_sign_bias[ALTREF_FRAME] = 0;
 
-    /* Check to see if a key frame is signaled
-     * For two pass with auto key frame enabled cm->frame_type may already
-     * be set, but not for one pass.
-     */
-    if ((cm->current_video_frame == 0) ||
-        (cm->frame_flags & FRAMEFLAGS_KEY) ||
-        (cpi->oxcf.auto_key && (cpi->frames_since_key % cpi->key_frame_frequency == 0)))
-    {
-        /* Key frame from VFW/auto-keyframe/first frame */
-        cm->frame_type = KEY_FRAME;
+  /* Check to see if a key frame is signaled
+   * For two pass with auto key frame enabled cm->frame_type may already
+   * be set, but not for one pass.
+   */
+  if ((cm->current_video_frame == 0) || (cm->frame_flags & FRAMEFLAGS_KEY) ||
+      (cpi->oxcf.auto_key &&
+       (cpi->frames_since_key % cpi->key_frame_frequency == 0))) {
+    /* Key frame from VFW/auto-keyframe/first frame */
+    cm->frame_type = KEY_FRAME;
 #if CONFIG_TEMPORAL_DENOISING
-        if (cpi->oxcf.noise_sensitivity == 4) {
-          // For adaptive mode, reset denoiser to normal mode on key frame.
-          vp8_denoiser_set_parameters(&cpi->denoiser, kDenoiserOnYUV);
-        }
-#endif
+    if (cpi->oxcf.noise_sensitivity == 4) {
+      // For adaptive mode, reset denoiser to normal mode on key frame.
+      vp8_denoiser_set_parameters(&cpi->denoiser, kDenoiserOnYUV);
     }
+#endif
+  }
 
 #if CONFIG_MULTI_RES_ENCODING
-    if (cpi->oxcf.mr_total_resolutions > 1) {
-      LOWER_RES_FRAME_INFO* low_res_frame_info
-         = (LOWER_RES_FRAME_INFO*)cpi->oxcf.mr_low_res_mode_info;
+  if (cpi->oxcf.mr_total_resolutions > 1) {
+    LOWER_RES_FRAME_INFO *low_res_frame_info =
+        (LOWER_RES_FRAME_INFO *)cpi->oxcf.mr_low_res_mode_info;
 
-      if (cpi->oxcf.mr_encoder_id) {
+    if (cpi->oxcf.mr_encoder_id) {
+      // TODO(marpan): This constraint shouldn't be needed, as we would like
+      // to allow for key frame setting (forced or periodic) defined per
+      // spatial layer. For now, keep this in.
+      cm->frame_type = low_res_frame_info->frame_type;
 
-        // TODO(marpan): This constraint shouldn't be needed, as we would like
-        // to allow for key frame setting (forced or periodic) defined per
-        // spatial layer. For now, keep this in.
-        cm->frame_type = low_res_frame_info->frame_type;
+      // Check if lower resolution is available for motion vector reuse.
+      if (cm->frame_type != KEY_FRAME) {
+        cpi->mr_low_res_mv_avail = 1;
+        cpi->mr_low_res_mv_avail &= !(low_res_frame_info->is_frame_dropped);
 
-        // Check if lower resolution is available for motion vector reuse.
-        if(cm->frame_type != KEY_FRAME)
-        {
-          cpi->mr_low_res_mv_avail = 1;
-          cpi->mr_low_res_mv_avail &= !(low_res_frame_info->is_frame_dropped);
+        if (cpi->ref_frame_flags & VP8_LAST_FRAME)
+          cpi->mr_low_res_mv_avail &=
+              (cpi->current_ref_frames[LAST_FRAME] ==
+               low_res_frame_info->low_res_ref_frames[LAST_FRAME]);
 
-          if (cpi->ref_frame_flags & VP8_LAST_FRAME)
-              cpi->mr_low_res_mv_avail &= (cpi->current_ref_frames[LAST_FRAME]
-                       == low_res_frame_info->low_res_ref_frames[LAST_FRAME]);
+        if (cpi->ref_frame_flags & VP8_GOLD_FRAME)
+          cpi->mr_low_res_mv_avail &=
+              (cpi->current_ref_frames[GOLDEN_FRAME] ==
+               low_res_frame_info->low_res_ref_frames[GOLDEN_FRAME]);
 
-          if (cpi->ref_frame_flags & VP8_GOLD_FRAME)
-              cpi->mr_low_res_mv_avail &= (cpi->current_ref_frames[GOLDEN_FRAME]
-                       == low_res_frame_info->low_res_ref_frames[GOLDEN_FRAME]);
-
-          // Don't use altref to determine whether low res is available.
-          // TODO (marpan): Should we make this type of condition on a
-          // per-reference frame basis?
-          /*
-          if (cpi->ref_frame_flags & VP8_ALTR_FRAME)
-              cpi->mr_low_res_mv_avail &= (cpi->current_ref_frames[ALTREF_FRAME]
-                       == low_res_frame_info->low_res_ref_frames[ALTREF_FRAME]);
-          */
-        }
+        // Don't use altref to determine whether low res is available.
+        // TODO (marpan): Should we make this type of condition on a
+        // per-reference frame basis?
+        /*
+        if (cpi->ref_frame_flags & VP8_ALTR_FRAME)
+            cpi->mr_low_res_mv_avail &= (cpi->current_ref_frames[ALTREF_FRAME]
+                     == low_res_frame_info->low_res_ref_frames[ALTREF_FRAME]);
+        */
       }
+    }
 
-      // On a key frame: For the lowest resolution, keep track of the key frame
-      // counter value. For the higher resolutions, reset the current video
-      // frame counter to that of the lowest resolution.
-      // This is done to the handle the case where we may stop/start encoding
-      // higher layer(s). The restart-encoding of higher layer is only signaled
-      // by a key frame for now.
-      // TODO (marpan): Add flag to indicate restart-encoding of higher layer.
-      if (cm->frame_type == KEY_FRAME) {
-        if (cpi->oxcf.mr_encoder_id) {
-          // If the initial starting value of the buffer level is zero (this can
-          // happen because we may have not started encoding this higher stream),
-          // then reset it to non-zero value based on |starting_buffer_level|.
-          if (cpi->common.current_video_frame == 0 && cpi->buffer_level == 0) {
-            unsigned int i;
-            cpi->bits_off_target = cpi->oxcf.starting_buffer_level;
-            cpi->buffer_level = cpi->oxcf.starting_buffer_level;
-            for (i = 0; i < cpi->oxcf.number_of_layers; i++) {
-              LAYER_CONTEXT *lc = &cpi->layer_context[i];
-              lc->bits_off_target = lc->starting_buffer_level;
-              lc->buffer_level = lc->starting_buffer_level;
-            }
+    // On a key frame: For the lowest resolution, keep track of the key frame
+    // counter value. For the higher resolutions, reset the current video
+    // frame counter to that of the lowest resolution.
+    // This is done to the handle the case where we may stop/start encoding
+    // higher layer(s). The restart-encoding of higher layer is only signaled
+    // by a key frame for now.
+    // TODO (marpan): Add flag to indicate restart-encoding of higher layer.
+    if (cm->frame_type == KEY_FRAME) {
+      if (cpi->oxcf.mr_encoder_id) {
+        // If the initial starting value of the buffer level is zero (this can
+        // happen because we may have not started encoding this higher stream),
+        // then reset it to non-zero value based on |starting_buffer_level|.
+        if (cpi->common.current_video_frame == 0 && cpi->buffer_level == 0) {
+          unsigned int i;
+          cpi->bits_off_target = cpi->oxcf.starting_buffer_level;
+          cpi->buffer_level = cpi->oxcf.starting_buffer_level;
+          for (i = 0; i < cpi->oxcf.number_of_layers; i++) {
+            LAYER_CONTEXT *lc = &cpi->layer_context[i];
+            lc->bits_off_target = lc->starting_buffer_level;
+            lc->buffer_level = lc->starting_buffer_level;
           }
-          cpi->common.current_video_frame =
-              low_res_frame_info->key_frame_counter_value;
-        } else {
-          low_res_frame_info->key_frame_counter_value =
-              cpi->common.current_video_frame;
         }
+        cpi->common.current_video_frame =
+            low_res_frame_info->key_frame_counter_value;
+      } else {
+        low_res_frame_info->key_frame_counter_value =
+            cpi->common.current_video_frame;
       }
-
     }
+  }
 #endif
 
-    // Find the reference frame closest to the current frame.
-    cpi->closest_reference_frame = LAST_FRAME;
-    if(cm->frame_type != KEY_FRAME) {
-      int i;
-      MV_REFERENCE_FRAME closest_ref = INTRA_FRAME;
-      if (cpi->ref_frame_flags & VP8_LAST_FRAME) {
-        closest_ref = LAST_FRAME;
-      } else if (cpi->ref_frame_flags & VP8_GOLD_FRAME) {
-        closest_ref = GOLDEN_FRAME;
-      } else if (cpi->ref_frame_flags & VP8_ALTR_FRAME) {
-        closest_ref = ALTREF_FRAME;
-      }
-      for(i = 1; i <= 3; i++) {
-        vpx_ref_frame_type_t ref_frame_type = (vpx_ref_frame_type_t)
-            ((i == 3) ? 4 : i);
-        if (cpi->ref_frame_flags & ref_frame_type) {
-          if ((cm->current_video_frame - cpi->current_ref_frames[i]) <
-              (cm->current_video_frame - cpi->current_ref_frames[closest_ref])) {
-            closest_ref = i;
-          }
+  // Find the reference frame closest to the current frame.
+  cpi->closest_reference_frame = LAST_FRAME;
+  if (cm->frame_type != KEY_FRAME) {
+    int i;
+    MV_REFERENCE_FRAME closest_ref = INTRA_FRAME;
+    if (cpi->ref_frame_flags & VP8_LAST_FRAME) {
+      closest_ref = LAST_FRAME;
+    } else if (cpi->ref_frame_flags & VP8_GOLD_FRAME) {
+      closest_ref = GOLDEN_FRAME;
+    } else if (cpi->ref_frame_flags & VP8_ALTR_FRAME) {
+      closest_ref = ALTREF_FRAME;
+    }
+    for (i = 1; i <= 3; i++) {
+      vpx_ref_frame_type_t ref_frame_type =
+          (vpx_ref_frame_type_t)((i == 3) ? 4 : i);
+      if (cpi->ref_frame_flags & ref_frame_type) {
+        if ((cm->current_video_frame - cpi->current_ref_frames[i]) <
+            (cm->current_video_frame - cpi->current_ref_frames[closest_ref])) {
+          closest_ref = i;
         }
       }
-      cpi->closest_reference_frame = closest_ref;
     }
+    cpi->closest_reference_frame = closest_ref;
+  }
 
-    /* Set various flags etc to special state if it is a key frame */
-    if (cm->frame_type == KEY_FRAME)
-    {
-        int i;
+  /* Set various flags etc to special state if it is a key frame */
+  if (cm->frame_type == KEY_FRAME) {
+    int i;
 
-        // Set the loop filter deltas and segmentation map update
-        setup_features(cpi);
+    // Set the loop filter deltas and segmentation map update
+    setup_features(cpi);
 
-        /* The alternate reference frame cannot be active for a key frame */
-        cpi->source_alt_ref_active = 0;
+    /* The alternate reference frame cannot be active for a key frame */
+    cpi->source_alt_ref_active = 0;
 
-        /* Reset the RD threshold multipliers to default of * 1 (128) */
-        for (i = 0; i < MAX_MODES; i++)
-        {
-            cpi->mb.rd_thresh_mult[i] = 128;
-        }
-
-        // Reset the zero_last counter to 0 on key frame.
-        memset(cpi->consec_zero_last, 0, cm->mb_rows * cm->mb_cols);
-        memset(cpi->consec_zero_last_mvbias, 0,
-               (cpi->common.mb_rows * cpi->common.mb_cols));
+    /* Reset the RD threshold multipliers to default of * 1 (128) */
+    for (i = 0; i < MAX_MODES; i++) {
+      cpi->mb.rd_thresh_mult[i] = 128;
     }
 
+    // Reset the zero_last counter to 0 on key frame.
+    memset(cpi->consec_zero_last, 0, cm->mb_rows * cm->mb_cols);
+    memset(cpi->consec_zero_last_mvbias, 0,
+           (cpi->common.mb_rows * cpi->common.mb_cols));
+  }
+
 #if 0
     /* Experimental code for lagged compress and one pass
      * Initialise one_pass GF frames stats
@@ -3794,486 +3426,424 @@
     }
 #endif
 
-    update_rd_ref_frame_probs(cpi);
+  update_rd_ref_frame_probs(cpi);
 
-    if (cpi->drop_frames_allowed)
-    {
-        /* The reset to decimation 0 is only done here for one pass.
-         * Once it is set two pass leaves decimation on till the next kf.
-         */
-        if ((cpi->buffer_level > drop_mark) && (cpi->decimation_factor > 0))
-            cpi->decimation_factor --;
+  if (cpi->drop_frames_allowed) {
+    /* The reset to decimation 0 is only done here for one pass.
+     * Once it is set two pass leaves decimation on till the next kf.
+     */
+    if ((cpi->buffer_level > drop_mark) && (cpi->decimation_factor > 0))
+      cpi->decimation_factor--;
 
-        if (cpi->buffer_level > drop_mark75 && cpi->decimation_factor > 0)
-            cpi->decimation_factor = 1;
+    if (cpi->buffer_level > drop_mark75 && cpi->decimation_factor > 0)
+      cpi->decimation_factor = 1;
 
-        else if (cpi->buffer_level < drop_mark25 && (cpi->decimation_factor == 2 || cpi->decimation_factor == 3))
-        {
-            cpi->decimation_factor = 3;
-        }
-        else if (cpi->buffer_level < drop_mark50 && (cpi->decimation_factor == 1 || cpi->decimation_factor == 2))
-        {
-            cpi->decimation_factor = 2;
-        }
-        else if (cpi->buffer_level < drop_mark75 && (cpi->decimation_factor == 0 || cpi->decimation_factor == 1))
-        {
-            cpi->decimation_factor = 1;
-        }
+    else if (cpi->buffer_level < drop_mark25 &&
+             (cpi->decimation_factor == 2 || cpi->decimation_factor == 3)) {
+      cpi->decimation_factor = 3;
+    } else if (cpi->buffer_level < drop_mark50 &&
+               (cpi->decimation_factor == 1 || cpi->decimation_factor == 2)) {
+      cpi->decimation_factor = 2;
+    } else if (cpi->buffer_level < drop_mark75 &&
+               (cpi->decimation_factor == 0 || cpi->decimation_factor == 1)) {
+      cpi->decimation_factor = 1;
     }
+  }
 
-    /* The following decimates the frame rate according to a regular
-     * pattern (i.e. to 1/2 or 2/3 frame rate) This can be used to help
-     * prevent buffer under-run in CBR mode. Alternatively it might be
-     * desirable in some situations to drop frame rate but throw more bits
-     * at each frame.
-     *
-     * Note that dropping a key frame can be problematic if spatial
-     * resampling is also active
+  /* The following decimates the frame rate according to a regular
+   * pattern (i.e. to 1/2 or 2/3 frame rate) This can be used to help
+   * prevent buffer under-run in CBR mode. Alternatively it might be
+   * desirable in some situations to drop frame rate but throw more bits
+   * at each frame.
+   *
+   * Note that dropping a key frame can be problematic if spatial
+   * resampling is also active
+   */
+  if (cpi->decimation_factor > 0) {
+    switch (cpi->decimation_factor) {
+      case 1:
+        cpi->per_frame_bandwidth = cpi->per_frame_bandwidth * 3 / 2;
+        break;
+      case 2:
+        cpi->per_frame_bandwidth = cpi->per_frame_bandwidth * 5 / 4;
+        break;
+      case 3:
+        cpi->per_frame_bandwidth = cpi->per_frame_bandwidth * 5 / 4;
+        break;
+    }
+
+    /* Note that we should not throw out a key frame (especially when
+     * spatial resampling is enabled).
      */
-    if (cpi->decimation_factor > 0)
-    {
-        switch (cpi->decimation_factor)
-        {
-        case 1:
-            cpi->per_frame_bandwidth  = cpi->per_frame_bandwidth * 3 / 2;
-            break;
-        case 2:
-            cpi->per_frame_bandwidth  = cpi->per_frame_bandwidth * 5 / 4;
-            break;
-        case 3:
-            cpi->per_frame_bandwidth  = cpi->per_frame_bandwidth * 5 / 4;
-            break;
-        }
+    if (cm->frame_type == KEY_FRAME) {
+      cpi->decimation_count = cpi->decimation_factor;
+    } else if (cpi->decimation_count > 0) {
+      cpi->decimation_count--;
 
-        /* Note that we should not throw out a key frame (especially when
-         * spatial resampling is enabled).
-         */
-        if (cm->frame_type == KEY_FRAME)
-        {
-            cpi->decimation_count = cpi->decimation_factor;
-        }
-        else if (cpi->decimation_count > 0)
-        {
-            cpi->decimation_count --;
+      cpi->bits_off_target += cpi->av_per_frame_bandwidth;
+      if (cpi->bits_off_target > cpi->oxcf.maximum_buffer_size)
+        cpi->bits_off_target = cpi->oxcf.maximum_buffer_size;
 
-            cpi->bits_off_target += cpi->av_per_frame_bandwidth;
-            if (cpi->bits_off_target > cpi->oxcf.maximum_buffer_size)
-                cpi->bits_off_target = cpi->oxcf.maximum_buffer_size;
-
 #if CONFIG_MULTI_RES_ENCODING
-            vp8_store_drop_frame_info(cpi);
+      vp8_store_drop_frame_info(cpi);
 #endif
 
-            cm->current_video_frame++;
-            cpi->frames_since_key++;
-            // We advance the temporal pattern for dropped frames.
-            cpi->temporal_pattern_counter++;
+      cm->current_video_frame++;
+      cpi->frames_since_key++;
+      // We advance the temporal pattern for dropped frames.
+      cpi->temporal_pattern_counter++;
 
 #if CONFIG_INTERNAL_STATS
-            cpi->count ++;
+      cpi->count++;
 #endif
 
-            cpi->buffer_level = cpi->bits_off_target;
+      cpi->buffer_level = cpi->bits_off_target;
 
-            if (cpi->oxcf.number_of_layers > 1)
-            {
-                unsigned int i;
+      if (cpi->oxcf.number_of_layers > 1) {
+        unsigned int i;
 
-                /* Propagate bits saved by dropping the frame to higher
-                 * layers
-                 */
-                for (i=cpi->current_layer+1; i<cpi->oxcf.number_of_layers; i++)
-                {
-                    LAYER_CONTEXT *lc = &cpi->layer_context[i];
-                    lc->bits_off_target += (int)(lc->target_bandwidth /
-                                                 lc->framerate);
-                    if (lc->bits_off_target > lc->maximum_buffer_size)
-                        lc->bits_off_target = lc->maximum_buffer_size;
-                    lc->buffer_level = lc->bits_off_target;
-                }
-            }
-
-            return;
+        /* Propagate bits saved by dropping the frame to higher
+         * layers
+         */
+        for (i = cpi->current_layer + 1; i < cpi->oxcf.number_of_layers; i++) {
+          LAYER_CONTEXT *lc = &cpi->layer_context[i];
+          lc->bits_off_target += (int)(lc->target_bandwidth / lc->framerate);
+          if (lc->bits_off_target > lc->maximum_buffer_size)
+            lc->bits_off_target = lc->maximum_buffer_size;
+          lc->buffer_level = lc->bits_off_target;
         }
-        else
-            cpi->decimation_count = cpi->decimation_factor;
-    }
-    else
-        cpi->decimation_count = 0;
+      }
 
-    /* Decide how big to make the frame */
-    if (!vp8_pick_frame_size(cpi))
-    {
-        /*TODO: 2 drop_frame and return code could be put together. */
+      return;
+    } else
+      cpi->decimation_count = cpi->decimation_factor;
+  } else
+    cpi->decimation_count = 0;
+
+  /* Decide how big to make the frame */
+  if (!vp8_pick_frame_size(cpi)) {
+/*TODO: 2 drop_frame and return code could be put together. */
 #if CONFIG_MULTI_RES_ENCODING
-        vp8_store_drop_frame_info(cpi);
+    vp8_store_drop_frame_info(cpi);
 #endif
-        cm->current_video_frame++;
-        cpi->frames_since_key++;
-        // We advance the temporal pattern for dropped frames.
-        cpi->temporal_pattern_counter++;
-        return;
-    }
+    cm->current_video_frame++;
+    cpi->frames_since_key++;
+    // We advance the temporal pattern for dropped frames.
+    cpi->temporal_pattern_counter++;
+    return;
+  }
 
-    /* Reduce active_worst_allowed_q for CBR if our buffer is getting too full.
-     * This has a knock on effect on active best quality as well.
-     * For CBR if the buffer reaches its maximum level then we can no longer
-     * save up bits for later frames so we might as well use them up
-     * on the current frame.
-     */
-    if ((cpi->oxcf.end_usage == USAGE_STREAM_FROM_SERVER) &&
-        (cpi->buffer_level >= cpi->oxcf.optimal_buffer_level) && cpi->buffered_mode)
-    {
-        /* Max adjustment is 1/4 */
-        int Adjustment = cpi->active_worst_quality / 4;
+  /* Reduce active_worst_allowed_q for CBR if our buffer is getting too full.
+   * This has a knock on effect on active best quality as well.
+   * For CBR if the buffer reaches its maximum level then we can no longer
+   * save up bits for later frames so we might as well use them up
+   * on the current frame.
+   */
+  if ((cpi->oxcf.end_usage == USAGE_STREAM_FROM_SERVER) &&
+      (cpi->buffer_level >= cpi->oxcf.optimal_buffer_level) &&
+      cpi->buffered_mode) {
+    /* Max adjustment is 1/4 */
+    int Adjustment = cpi->active_worst_quality / 4;
 
-        if (Adjustment)
-        {
-            int buff_lvl_step;
+    if (Adjustment) {
+      int buff_lvl_step;
 
-            if (cpi->buffer_level < cpi->oxcf.maximum_buffer_size)
-            {
-                buff_lvl_step = (int)
-                                ((cpi->oxcf.maximum_buffer_size -
-                                  cpi->oxcf.optimal_buffer_level) /
-                                  Adjustment);
+      if (cpi->buffer_level < cpi->oxcf.maximum_buffer_size) {
+        buff_lvl_step = (int)((cpi->oxcf.maximum_buffer_size -
+                               cpi->oxcf.optimal_buffer_level) /
+                              Adjustment);
 
-                if (buff_lvl_step)
-                    Adjustment = (int)
-                                 ((cpi->buffer_level -
-                                 cpi->oxcf.optimal_buffer_level) /
-                                 buff_lvl_step);
-                else
-                    Adjustment = 0;
-            }
+        if (buff_lvl_step)
+          Adjustment =
+              (int)((cpi->buffer_level - cpi->oxcf.optimal_buffer_level) /
+                    buff_lvl_step);
+        else
+          Adjustment = 0;
+      }
 
-            cpi->active_worst_quality -= Adjustment;
+      cpi->active_worst_quality -= Adjustment;
 
-            if(cpi->active_worst_quality < cpi->active_best_quality)
-                cpi->active_worst_quality = cpi->active_best_quality;
-        }
+      if (cpi->active_worst_quality < cpi->active_best_quality)
+        cpi->active_worst_quality = cpi->active_best_quality;
     }
+  }
 
-    /* Set an active best quality and if necessary active worst quality
-     * There is some odd behavior for one pass here that needs attention.
-     */
-    if ( (cpi->pass == 2) || (cpi->ni_frames > 150))
-    {
-        vp8_clear_system_state();
+  /* Set an active best quality and if necessary active worst quality
+   * There is some odd behavior for one pass here that needs attention.
+   */
+  if ((cpi->pass == 2) || (cpi->ni_frames > 150)) {
+    vp8_clear_system_state();
 
-        Q = cpi->active_worst_quality;
+    Q = cpi->active_worst_quality;
 
-        if ( cm->frame_type == KEY_FRAME )
-        {
-            if ( cpi->pass == 2 )
-            {
-                if (cpi->gfu_boost > 600)
-                   cpi->active_best_quality = kf_low_motion_minq[Q];
-                else
-                   cpi->active_best_quality = kf_high_motion_minq[Q];
+    if (cm->frame_type == KEY_FRAME) {
+      if (cpi->pass == 2) {
+        if (cpi->gfu_boost > 600)
+          cpi->active_best_quality = kf_low_motion_minq[Q];
+        else
+          cpi->active_best_quality = kf_high_motion_minq[Q];
 
-                /* Special case for key frames forced because we have reached
-                 * the maximum key frame interval. Here force the Q to a range
-                 * based on the ambient Q to reduce the risk of popping
-                 */
-                if ( cpi->this_key_frame_forced )
-                {
-                    if ( cpi->active_best_quality > cpi->avg_frame_qindex * 7/8)
-                        cpi->active_best_quality = cpi->avg_frame_qindex * 7/8;
-                    else if ( cpi->active_best_quality < cpi->avg_frame_qindex >> 2 )
-                        cpi->active_best_quality = cpi->avg_frame_qindex >> 2;
-                }
-            }
-            /* One pass more conservative */
-            else
-               cpi->active_best_quality = kf_high_motion_minq[Q];
+        /* Special case for key frames forced because we have reached
+         * the maximum key frame interval. Here force the Q to a range
+         * based on the ambient Q to reduce the risk of popping
+         */
+        if (cpi->this_key_frame_forced) {
+          if (cpi->active_best_quality > cpi->avg_frame_qindex * 7 / 8)
+            cpi->active_best_quality = cpi->avg_frame_qindex * 7 / 8;
+          else if (cpi->active_best_quality<cpi->avg_frame_qindex>> 2)
+            cpi->active_best_quality = cpi->avg_frame_qindex >> 2;
         }
+      }
+      /* One pass more conservative */
+      else
+        cpi->active_best_quality = kf_high_motion_minq[Q];
+    }
 
-        else if (cpi->oxcf.number_of_layers==1 &&
-                (cm->refresh_golden_frame || cpi->common.refresh_alt_ref_frame))
-        {
-            /* Use the lower of cpi->active_worst_quality and recent
-             * average Q as basis for GF/ARF Q limit unless last frame was
-             * a key frame.
-             */
-            if ( (cpi->frames_since_key > 1) &&
-               (cpi->avg_frame_qindex < cpi->active_worst_quality) )
-            {
-                Q = cpi->avg_frame_qindex;
-            }
+    else if (cpi->oxcf.number_of_layers == 1 &&
+             (cm->refresh_golden_frame || cpi->common.refresh_alt_ref_frame)) {
+      /* Use the lower of cpi->active_worst_quality and recent
+       * average Q as basis for GF/ARF Q limit unless last frame was
+       * a key frame.
+       */
+      if ((cpi->frames_since_key > 1) &&
+          (cpi->avg_frame_qindex < cpi->active_worst_quality)) {
+        Q = cpi->avg_frame_qindex;
+      }
 
-            /* For constrained quality dont allow Q less than the cq level */
-            if ( (cpi->oxcf.end_usage == USAGE_CONSTRAINED_QUALITY) &&
-                 (Q < cpi->cq_target_quality) )
-            {
-                Q = cpi->cq_target_quality;
-            }
+      /* For constrained quality dont allow Q less than the cq level */
+      if ((cpi->oxcf.end_usage == USAGE_CONSTRAINED_QUALITY) &&
+          (Q < cpi->cq_target_quality)) {
+        Q = cpi->cq_target_quality;
+      }
 
-            if ( cpi->pass == 2 )
-            {
-                if ( cpi->gfu_boost > 1000 )
-                    cpi->active_best_quality = gf_low_motion_minq[Q];
-                else if ( cpi->gfu_boost < 400 )
-                    cpi->active_best_quality = gf_high_motion_minq[Q];
-                else
-                    cpi->active_best_quality = gf_mid_motion_minq[Q];
-
-                /* Constrained quality use slightly lower active best. */
-                if ( cpi->oxcf.end_usage == USAGE_CONSTRAINED_QUALITY )
-                {
-                    cpi->active_best_quality =
-                        cpi->active_best_quality * 15/16;
-                }
-            }
-            /* One pass more conservative */
-            else
-                cpi->active_best_quality = gf_high_motion_minq[Q];
-        }
+      if (cpi->pass == 2) {
+        if (cpi->gfu_boost > 1000)
+          cpi->active_best_quality = gf_low_motion_minq[Q];
+        else if (cpi->gfu_boost < 400)
+          cpi->active_best_quality = gf_high_motion_minq[Q];
         else
-        {
-            cpi->active_best_quality = inter_minq[Q];
+          cpi->active_best_quality = gf_mid_motion_minq[Q];
 
-            /* For the constant/constrained quality mode we dont want
-             * q to fall below the cq level.
-             */
-            if ((cpi->oxcf.end_usage == USAGE_CONSTRAINED_QUALITY) &&
-                (cpi->active_best_quality < cpi->cq_target_quality) )
-            {
-                /* If we are strongly undershooting the target rate in the last
-                 * frames then use the user passed in cq value not the auto
-                 * cq value.
-                 */
-                if ( cpi->rolling_actual_bits < cpi->min_frame_bandwidth )
-                    cpi->active_best_quality = cpi->oxcf.cq_level;
-                else
-                    cpi->active_best_quality = cpi->cq_target_quality;
-            }
+        /* Constrained quality use slightly lower active best. */
+        if (cpi->oxcf.end_usage == USAGE_CONSTRAINED_QUALITY) {
+          cpi->active_best_quality = cpi->active_best_quality * 15 / 16;
         }
+      }
+      /* One pass more conservative */
+      else
+        cpi->active_best_quality = gf_high_motion_minq[Q];
+    } else {
+      cpi->active_best_quality = inter_minq[Q];
 
-        /* If CBR and the buffer is as full then it is reasonable to allow
-         * higher quality on the frames to prevent bits just going to waste.
+      /* For the constant/constrained quality mode we dont want
+       * q to fall below the cq level.
+       */
+      if ((cpi->oxcf.end_usage == USAGE_CONSTRAINED_QUALITY) &&
+          (cpi->active_best_quality < cpi->cq_target_quality)) {
+        /* If we are strongly undershooting the target rate in the last
+         * frames then use the user passed in cq value not the auto
+         * cq value.
          */
-        if (cpi->oxcf.end_usage == USAGE_STREAM_FROM_SERVER)
-        {
-            /* Note that the use of >= here elliminates the risk of a devide
-             * by 0 error in the else if clause
-             */
-            if (cpi->buffer_level >= cpi->oxcf.maximum_buffer_size)
-                cpi->active_best_quality = cpi->best_quality;
+        if (cpi->rolling_actual_bits < cpi->min_frame_bandwidth)
+          cpi->active_best_quality = cpi->oxcf.cq_level;
+        else
+          cpi->active_best_quality = cpi->cq_target_quality;
+      }
+    }
 
-            else if (cpi->buffer_level > cpi->oxcf.optimal_buffer_level)
-            {
-                int Fraction = (int)
-                  (((cpi->buffer_level - cpi->oxcf.optimal_buffer_level) * 128)
-                  / (cpi->oxcf.maximum_buffer_size -
-                  cpi->oxcf.optimal_buffer_level));
-                int min_qadjustment = ((cpi->active_best_quality -
-                                        cpi->best_quality) * Fraction) / 128;
+    /* If CBR and the buffer is as full then it is reasonable to allow
+     * higher quality on the frames to prevent bits just going to waste.
+     */
+    if (cpi->oxcf.end_usage == USAGE_STREAM_FROM_SERVER) {
+      /* Note that the use of >= here elliminates the risk of a devide
+       * by 0 error in the else if clause
+       */
+      if (cpi->buffer_level >= cpi->oxcf.maximum_buffer_size)
+        cpi->active_best_quality = cpi->best_quality;
 
-                cpi->active_best_quality -= min_qadjustment;
-            }
-        }
+      else if (cpi->buffer_level > cpi->oxcf.optimal_buffer_level) {
+        int Fraction =
+            (int)(((cpi->buffer_level - cpi->oxcf.optimal_buffer_level) * 128) /
+                  (cpi->oxcf.maximum_buffer_size -
+                   cpi->oxcf.optimal_buffer_level));
+        int min_qadjustment =
+            ((cpi->active_best_quality - cpi->best_quality) * Fraction) / 128;
+
+        cpi->active_best_quality -= min_qadjustment;
+      }
     }
-    /* Make sure constrained quality mode limits are adhered to for the first
-     * few frames of one pass encodes
-     */
-    else if (cpi->oxcf.end_usage == USAGE_CONSTRAINED_QUALITY)
-    {
-        if ( (cm->frame_type == KEY_FRAME) ||
-             cm->refresh_golden_frame || cpi->common.refresh_alt_ref_frame )
-        {
-             cpi->active_best_quality = cpi->best_quality;
-        }
-        else if (cpi->active_best_quality < cpi->cq_target_quality)
-        {
-            cpi->active_best_quality = cpi->cq_target_quality;
-        }
+  }
+  /* Make sure constrained quality mode limits are adhered to for the first
+   * few frames of one pass encodes
+   */
+  else if (cpi->oxcf.end_usage == USAGE_CONSTRAINED_QUALITY) {
+    if ((cm->frame_type == KEY_FRAME) || cm->refresh_golden_frame ||
+        cpi->common.refresh_alt_ref_frame) {
+      cpi->active_best_quality = cpi->best_quality;
+    } else if (cpi->active_best_quality < cpi->cq_target_quality) {
+      cpi->active_best_quality = cpi->cq_target_quality;
     }
+  }
 
-    /* Clip the active best and worst quality values to limits */
-    if (cpi->active_worst_quality > cpi->worst_quality)
-        cpi->active_worst_quality = cpi->worst_quality;
+  /* Clip the active best and worst quality values to limits */
+  if (cpi->active_worst_quality > cpi->worst_quality)
+    cpi->active_worst_quality = cpi->worst_quality;
 
-    if (cpi->active_best_quality < cpi->best_quality)
-        cpi->active_best_quality = cpi->best_quality;
+  if (cpi->active_best_quality < cpi->best_quality)
+    cpi->active_best_quality = cpi->best_quality;
 
-    if ( cpi->active_worst_quality < cpi->active_best_quality )
-        cpi->active_worst_quality = cpi->active_best_quality;
+  if (cpi->active_worst_quality < cpi->active_best_quality)
+    cpi->active_worst_quality = cpi->active_best_quality;
 
-    /* Determine initial Q to try */
-    Q = vp8_regulate_q(cpi, cpi->this_frame_target);
+  /* Determine initial Q to try */
+  Q = vp8_regulate_q(cpi, cpi->this_frame_target);
 
 #if !CONFIG_REALTIME_ONLY
 
-    /* Set highest allowed value for Zbin over quant */
-    if (cm->frame_type == KEY_FRAME)
-        zbin_oq_high = 0;
-    else if ((cpi->oxcf.number_of_layers == 1) && ((cm->refresh_alt_ref_frame ||
-              (cm->refresh_golden_frame && !cpi->source_alt_ref_active))))
-    {
-          zbin_oq_high = 16;
-    }
-    else
-        zbin_oq_high = ZBIN_OQ_MAX;
+  /* Set highest allowed value for Zbin over quant */
+  if (cm->frame_type == KEY_FRAME)
+    zbin_oq_high = 0;
+  else if ((cpi->oxcf.number_of_layers == 1) &&
+           ((cm->refresh_alt_ref_frame ||
+             (cm->refresh_golden_frame && !cpi->source_alt_ref_active)))) {
+    zbin_oq_high = 16;
+  } else
+    zbin_oq_high = ZBIN_OQ_MAX;
 #endif
 
-    /* Setup background Q adjustment for error resilient mode.
-     * For multi-layer encodes only enable this for the base layer.
-    */
-    if (cpi->cyclic_refresh_mode_enabled)
-    {
-      // Special case for screen_content_mode with golden frame updates.
-      int disable_cr_gf = (cpi->oxcf.screen_content_mode == 2 &&
-                           cm->refresh_golden_frame);
-      if (cpi->current_layer == 0 && cpi->force_maxqp == 0 && !disable_cr_gf)
-        cyclic_background_refresh(cpi, Q, 0);
-      else
-        disable_segmentation(cpi);
-    }
+  /* Setup background Q adjustment for error resilient mode.
+   * For multi-layer encodes only enable this for the base layer.
+  */
+  if (cpi->cyclic_refresh_mode_enabled) {
+    // Special case for screen_content_mode with golden frame updates.
+    int disable_cr_gf =
+        (cpi->oxcf.screen_content_mode == 2 && cm->refresh_golden_frame);
+    if (cpi->current_layer == 0 && cpi->force_maxqp == 0 && !disable_cr_gf)
+      cyclic_background_refresh(cpi, Q, 0);
+    else
+      disable_segmentation(cpi);
+  }
 
-    vp8_compute_frame_size_bounds(cpi, &frame_under_shoot_limit, &frame_over_shoot_limit);
+  vp8_compute_frame_size_bounds(cpi, &frame_under_shoot_limit,
+                                &frame_over_shoot_limit);
 
 #if !CONFIG_REALTIME_ONLY
-    /* Limit Q range for the adaptive loop. */
-    bottom_index = cpi->active_best_quality;
-    top_index    = cpi->active_worst_quality;
-    q_low  = cpi->active_best_quality;
-    q_high = cpi->active_worst_quality;
+  /* Limit Q range for the adaptive loop. */
+  bottom_index = cpi->active_best_quality;
+  top_index = cpi->active_worst_quality;
+  q_low = cpi->active_best_quality;
+  q_high = cpi->active_worst_quality;
 #endif
 
-    vp8_save_coding_context(cpi);
+  vp8_save_coding_context(cpi);
 
-    loop_count = 0;
+  loop_count = 0;
 
-    scale_and_extend_source(cpi->un_scaled_source, cpi);
+  scale_and_extend_source(cpi->un_scaled_source, cpi);
 
 #if CONFIG_TEMPORAL_DENOISING && CONFIG_POSTPROC
-    // Option to apply spatial blur under the aggressive or adaptive
-    // (temporal denoising) mode.
-    if (cpi->oxcf.noise_sensitivity >= 3) {
-      if (cpi->denoiser.denoise_pars.spatial_blur != 0) {
-        vp8_de_noise(cm, cpi->Source, cpi->Source,
-            cpi->denoiser.denoise_pars.spatial_blur, 1, 0, 0);
-      }
+  // Option to apply spatial blur under the aggressive or adaptive
+  // (temporal denoising) mode.
+  if (cpi->oxcf.noise_sensitivity >= 3) {
+    if (cpi->denoiser.denoise_pars.spatial_blur != 0) {
+      vp8_de_noise(cm, cpi->Source, cpi->Source,
+                   cpi->denoiser.denoise_pars.spatial_blur, 1, 0, 0);
     }
+  }
 #endif
 
 #if !(CONFIG_REALTIME_ONLY) && CONFIG_POSTPROC && !(CONFIG_TEMPORAL_DENOISING)
 
-    if (cpi->oxcf.noise_sensitivity > 0)
-    {
-        unsigned char *src;
-        int l = 0;
+  if (cpi->oxcf.noise_sensitivity > 0) {
+    unsigned char *src;
+    int l = 0;
 
-        switch (cpi->oxcf.noise_sensitivity)
-        {
-        case 1:
-            l = 20;
-            break;
-        case 2:
-            l = 40;
-            break;
-        case 3:
-            l = 60;
-            break;
-        case 4:
-            l = 80;
-            break;
-        case 5:
-            l = 100;
-            break;
-        case 6:
-            l = 150;
-            break;
-        }
+    switch (cpi->oxcf.noise_sensitivity) {
+      case 1: l = 20; break;
+      case 2: l = 40; break;
+      case 3: l = 60; break;
+      case 4: l = 80; break;
+      case 5: l = 100; break;
+      case 6: l = 150; break;
+    }
 
+    if (cm->frame_type == KEY_FRAME) {
+      vp8_de_noise(cm, cpi->Source, cpi->Source, l, 1, 0, 1);
+    } else {
+      vp8_de_noise(cm, cpi->Source, cpi->Source, l, 1, 0, 1);
 
-        if (cm->frame_type == KEY_FRAME)
-        {
-            vp8_de_noise(cm, cpi->Source, cpi->Source, l , 1,  0, 1);
-        }
-        else
-        {
-            vp8_de_noise(cm, cpi->Source, cpi->Source, l , 1,  0, 1);
+      src = cpi->Source->y_buffer;
 
-            src = cpi->Source->y_buffer;
-
-            if (cpi->Source->y_stride < 0)
-            {
-                src += cpi->Source->y_stride * (cpi->Source->y_height - 1);
-            }
-        }
+      if (cpi->Source->y_stride < 0) {
+        src += cpi->Source->y_stride * (cpi->Source->y_height - 1);
+      }
     }
+  }
 
 #endif
 
-
 #ifdef OUTPUT_YUV_SRC
-    vp8_write_yuv_frame(yuv_file, cpi->Source);
+  vp8_write_yuv_frame(yuv_file, cpi->Source);
 #endif
 
-    do
-    {
-        vp8_clear_system_state();
+  do {
+    vp8_clear_system_state();
 
-        vp8_set_quantizer(cpi, Q);
+    vp8_set_quantizer(cpi, Q);
 
-        /* setup skip prob for costing in mode/mv decision */
-        if (cpi->common.mb_no_coeff_skip)
-        {
-            cpi->prob_skip_false = cpi->base_skip_false_prob[Q];
+    /* setup skip prob for costing in mode/mv decision */
+    if (cpi->common.mb_no_coeff_skip) {
+      cpi->prob_skip_false = cpi->base_skip_false_prob[Q];
 
-            if (cm->frame_type != KEY_FRAME)
-            {
-                if (cpi->common.refresh_alt_ref_frame)
-                {
-                    if (cpi->last_skip_false_probs[2] != 0)
-                        cpi->prob_skip_false = cpi->last_skip_false_probs[2];
+      if (cm->frame_type != KEY_FRAME) {
+        if (cpi->common.refresh_alt_ref_frame) {
+          if (cpi->last_skip_false_probs[2] != 0)
+            cpi->prob_skip_false = cpi->last_skip_false_probs[2];
 
-                    /*
-                                        if(cpi->last_skip_false_probs[2]!=0 && abs(Q- cpi->last_skip_probs_q[2])<=16 )
-                       cpi->prob_skip_false = cpi->last_skip_false_probs[2];
-                                        else if (cpi->last_skip_false_probs[2]!=0)
-                       cpi->prob_skip_false = (cpi->last_skip_false_probs[2]  + cpi->prob_skip_false ) / 2;
-                       */
-                }
-                else if (cpi->common.refresh_golden_frame)
-                {
-                    if (cpi->last_skip_false_probs[1] != 0)
-                        cpi->prob_skip_false = cpi->last_skip_false_probs[1];
+          /*
+                              if(cpi->last_skip_false_probs[2]!=0 && abs(Q-
+             cpi->last_skip_probs_q[2])<=16 )
+             cpi->prob_skip_false = cpi->last_skip_false_probs[2];
+                              else if (cpi->last_skip_false_probs[2]!=0)
+             cpi->prob_skip_false = (cpi->last_skip_false_probs[2]  +
+             cpi->prob_skip_false ) / 2;
+             */
+        } else if (cpi->common.refresh_golden_frame) {
+          if (cpi->last_skip_false_probs[1] != 0)
+            cpi->prob_skip_false = cpi->last_skip_false_probs[1];
 
-                    /*
-                                        if(cpi->last_skip_false_probs[1]!=0 && abs(Q- cpi->last_skip_probs_q[1])<=16 )
-                       cpi->prob_skip_false = cpi->last_skip_false_probs[1];
-                                        else if (cpi->last_skip_false_probs[1]!=0)
-                       cpi->prob_skip_false = (cpi->last_skip_false_probs[1]  + cpi->prob_skip_false ) / 2;
-                       */
-                }
-                else
-                {
-                    if (cpi->last_skip_false_probs[0] != 0)
-                        cpi->prob_skip_false = cpi->last_skip_false_probs[0];
+          /*
+                              if(cpi->last_skip_false_probs[1]!=0 && abs(Q-
+             cpi->last_skip_probs_q[1])<=16 )
+             cpi->prob_skip_false = cpi->last_skip_false_probs[1];
+                              else if (cpi->last_skip_false_probs[1]!=0)
+             cpi->prob_skip_false = (cpi->last_skip_false_probs[1]  +
+             cpi->prob_skip_false ) / 2;
+             */
+        } else {
+          if (cpi->last_skip_false_probs[0] != 0)
+            cpi->prob_skip_false = cpi->last_skip_false_probs[0];
 
-                    /*
-                    if(cpi->last_skip_false_probs[0]!=0 && abs(Q- cpi->last_skip_probs_q[0])<=16 )
-                        cpi->prob_skip_false = cpi->last_skip_false_probs[0];
-                    else if(cpi->last_skip_false_probs[0]!=0)
-                        cpi->prob_skip_false = (cpi->last_skip_false_probs[0]  + cpi->prob_skip_false ) / 2;
-                        */
-                }
+          /*
+          if(cpi->last_skip_false_probs[0]!=0 && abs(Q-
+          cpi->last_skip_probs_q[0])<=16 )
+              cpi->prob_skip_false = cpi->last_skip_false_probs[0];
+          else if(cpi->last_skip_false_probs[0]!=0)
+              cpi->prob_skip_false = (cpi->last_skip_false_probs[0]  +
+          cpi->prob_skip_false ) / 2;
+              */
+        }
 
-                /* as this is for cost estimate, let's make sure it does not
-                 * go extreme eitehr way
-                 */
-                if (cpi->prob_skip_false < 5)
-                    cpi->prob_skip_false = 5;
+        /* as this is for cost estimate, let's make sure it does not
+         * go extreme eitehr way
+         */
+        if (cpi->prob_skip_false < 5) cpi->prob_skip_false = 5;
 
-                if (cpi->prob_skip_false > 250)
-                    cpi->prob_skip_false = 250;
+        if (cpi->prob_skip_false > 250) cpi->prob_skip_false = 250;
 
-                if (cpi->oxcf.number_of_layers == 1 && cpi->is_src_frame_alt_ref)
-                    cpi->prob_skip_false = 1;
-            }
+        if (cpi->oxcf.number_of_layers == 1 && cpi->is_src_frame_alt_ref)
+          cpi->prob_skip_false = 1;
+      }
 
 #if 0
 
@@ -4285,350 +3855,321 @@
             }
 
 #endif
+    }
 
+    if (cm->frame_type == KEY_FRAME) {
+      if (resize_key_frame(cpi)) {
+        /* If the frame size has changed, need to reset Q, quantizer,
+         * and background refresh.
+         */
+        Q = vp8_regulate_q(cpi, cpi->this_frame_target);
+        if (cpi->cyclic_refresh_mode_enabled) {
+          if (cpi->current_layer == 0)
+            cyclic_background_refresh(cpi, Q, 0);
+          else
+            disable_segmentation(cpi);
         }
+        // Reset the zero_last counter to 0 on key frame.
+        memset(cpi->consec_zero_last, 0, cm->mb_rows * cm->mb_cols);
+        memset(cpi->consec_zero_last_mvbias, 0,
+               (cpi->common.mb_rows * cpi->common.mb_cols));
+        vp8_set_quantizer(cpi, Q);
+      }
 
-        if (cm->frame_type == KEY_FRAME)
-        {
-            if(resize_key_frame(cpi))
-            {
-              /* If the frame size has changed, need to reset Q, quantizer,
-               * and background refresh.
-               */
-              Q = vp8_regulate_q(cpi, cpi->this_frame_target);
-              if (cpi->cyclic_refresh_mode_enabled)
-              {
-                if (cpi->current_layer==0)
-                  cyclic_background_refresh(cpi, Q, 0);
-                else
-                  disable_segmentation(cpi);
-              }
-              // Reset the zero_last counter to 0 on key frame.
-              memset(cpi->consec_zero_last, 0, cm->mb_rows * cm->mb_cols);
-              memset(cpi->consec_zero_last_mvbias, 0,
-                     (cpi->common.mb_rows * cpi->common.mb_cols));
-              vp8_set_quantizer(cpi, Q);
-            }
+      vp8_setup_key_frame(cpi);
+    }
 
-            vp8_setup_key_frame(cpi);
-        }
-
 #if CONFIG_REALTIME_ONLY & CONFIG_ONTHEFLY_BITPACKING
-        {
-            if(cpi->oxcf.error_resilient_mode)
-                cm->refresh_entropy_probs = 0;
+    {
+      if (cpi->oxcf.error_resilient_mode) cm->refresh_entropy_probs = 0;
 
-            if (cpi->oxcf.error_resilient_mode & VPX_ERROR_RESILIENT_PARTITIONS)
-            {
-                if (cm->frame_type == KEY_FRAME)
-                    cm->refresh_entropy_probs = 1;
-            }
+      if (cpi->oxcf.error_resilient_mode & VPX_ERROR_RESILIENT_PARTITIONS) {
+        if (cm->frame_type == KEY_FRAME) cm->refresh_entropy_probs = 1;
+      }
 
-            if (cm->refresh_entropy_probs == 0)
-            {
-                /* save a copy for later refresh */
-                memcpy(&cm->lfc, &cm->fc, sizeof(cm->fc));
-            }
+      if (cm->refresh_entropy_probs == 0) {
+        /* save a copy for later refresh */
+        memcpy(&cm->lfc, &cm->fc, sizeof(cm->fc));
+      }
 
-            vp8_update_coef_context(cpi);
+      vp8_update_coef_context(cpi);
 
-            vp8_update_coef_probs(cpi);
+      vp8_update_coef_probs(cpi);
 
-            /* transform / motion compensation build reconstruction frame
-             * +pack coef partitions
-             */
-            vp8_encode_frame(cpi);
+      /* transform / motion compensation build reconstruction frame
+       * +pack coef partitions
+       */
+      vp8_encode_frame(cpi);
 
-            /* cpi->projected_frame_size is not needed for RT mode */
-        }
+      /* cpi->projected_frame_size is not needed for RT mode */
+    }
 #else
-        /* transform / motion compensation build reconstruction frame */
-        vp8_encode_frame(cpi);
+    /* transform / motion compensation build reconstruction frame */
+    vp8_encode_frame(cpi);
 
-        if (cpi->oxcf.screen_content_mode == 2) {
-          if (vp8_drop_encodedframe_overshoot(cpi, Q))
-            return;
-        }
+    if (cpi->oxcf.screen_content_mode == 2) {
+      if (vp8_drop_encodedframe_overshoot(cpi, Q)) return;
+    }
 
-        cpi->projected_frame_size -= vp8_estimate_entropy_savings(cpi);
-        cpi->projected_frame_size = (cpi->projected_frame_size > 0) ? cpi->projected_frame_size : 0;
+    cpi->projected_frame_size -= vp8_estimate_entropy_savings(cpi);
+    cpi->projected_frame_size =
+        (cpi->projected_frame_size > 0) ? cpi->projected_frame_size : 0;
 #endif
-        vp8_clear_system_state();
+    vp8_clear_system_state();
 
-        /* Test to see if the stats generated for this frame indicate that
-         * we should have coded a key frame (assuming that we didn't)!
-         */
+    /* Test to see if the stats generated for this frame indicate that
+     * we should have coded a key frame (assuming that we didn't)!
+     */
 
-        if (cpi->pass != 2 && cpi->oxcf.auto_key && cm->frame_type != KEY_FRAME
-            && cpi->compressor_speed != 2)
-        {
+    if (cpi->pass != 2 && cpi->oxcf.auto_key && cm->frame_type != KEY_FRAME &&
+        cpi->compressor_speed != 2) {
 #if !CONFIG_REALTIME_ONLY
-            if (decide_key_frame(cpi))
-            {
-                /* Reset all our sizing numbers and recode */
-                cm->frame_type = KEY_FRAME;
+      if (decide_key_frame(cpi)) {
+        /* Reset all our sizing numbers and recode */
+        cm->frame_type = KEY_FRAME;
 
-                vp8_pick_frame_size(cpi);
+        vp8_pick_frame_size(cpi);
 
-                /* Clear the Alt reference frame active flag when we have
-                 * a key frame
-                 */
-                cpi->source_alt_ref_active = 0;
+        /* Clear the Alt reference frame active flag when we have
+         * a key frame
+         */
+        cpi->source_alt_ref_active = 0;
 
-                // Set the loop filter deltas and segmentation map update
-                setup_features(cpi);
+        // Set the loop filter deltas and segmentation map update
+        setup_features(cpi);
 
-                vp8_restore_coding_context(cpi);
+        vp8_restore_coding_context(cpi);
 
-                Q = vp8_regulate_q(cpi, cpi->this_frame_target);
+        Q = vp8_regulate_q(cpi, cpi->this_frame_target);
 
-                vp8_compute_frame_size_bounds(cpi, &frame_under_shoot_limit, &frame_over_shoot_limit);
+        vp8_compute_frame_size_bounds(cpi, &frame_under_shoot_limit,
+                                      &frame_over_shoot_limit);
 
-                /* Limit Q range for the adaptive loop. */
-                bottom_index = cpi->active_best_quality;
-                top_index    = cpi->active_worst_quality;
-                q_low  = cpi->active_best_quality;
-                q_high = cpi->active_worst_quality;
+        /* Limit Q range for the adaptive loop. */
+        bottom_index = cpi->active_best_quality;
+        top_index = cpi->active_worst_quality;
+        q_low = cpi->active_best_quality;
+        q_high = cpi->active_worst_quality;
 
-                loop_count++;
-                Loop = 1;
+        loop_count++;
+        Loop = 1;
 
-                continue;
-            }
+        continue;
+      }
 #endif
-        }
+    }
 
-        vp8_clear_system_state();
+    vp8_clear_system_state();
 
-        if (frame_over_shoot_limit == 0)
-            frame_over_shoot_limit = 1;
+    if (frame_over_shoot_limit == 0) frame_over_shoot_limit = 1;
 
-        /* Are we are overshooting and up against the limit of active max Q. */
-        if (((cpi->pass != 2) || (cpi->oxcf.end_usage == USAGE_STREAM_FROM_SERVER)) &&
-            (Q == cpi->active_worst_quality)                     &&
-            (cpi->active_worst_quality < cpi->worst_quality)      &&
-            (cpi->projected_frame_size > frame_over_shoot_limit))
-        {
-            int over_size_percent = ((cpi->projected_frame_size - frame_over_shoot_limit) * 100) / frame_over_shoot_limit;
+    /* Are we are overshooting and up against the limit of active max Q. */
+    if (((cpi->pass != 2) ||
+         (cpi->oxcf.end_usage == USAGE_STREAM_FROM_SERVER)) &&
+        (Q == cpi->active_worst_quality) &&
+        (cpi->active_worst_quality < cpi->worst_quality) &&
+        (cpi->projected_frame_size > frame_over_shoot_limit)) {
+      int over_size_percent =
+          ((cpi->projected_frame_size - frame_over_shoot_limit) * 100) /
+          frame_over_shoot_limit;
 
-            /* If so is there any scope for relaxing it */
-            while ((cpi->active_worst_quality < cpi->worst_quality) && (over_size_percent > 0))
-            {
-                cpi->active_worst_quality++;
-                /* Assume 1 qstep = about 4% on frame size. */
-                over_size_percent = (int)(over_size_percent * 0.96);
-            }
+      /* If so is there any scope for relaxing it */
+      while ((cpi->active_worst_quality < cpi->worst_quality) &&
+             (over_size_percent > 0)) {
+        cpi->active_worst_quality++;
+        /* Assume 1 qstep = about 4% on frame size. */
+        over_size_percent = (int)(over_size_percent * 0.96);
+      }
 #if !CONFIG_REALTIME_ONLY
-            top_index = cpi->active_worst_quality;
+      top_index = cpi->active_worst_quality;
 #endif  // !CONFIG_REALTIME_ONLY
-            /* If we have updated the active max Q do not call
-             * vp8_update_rate_correction_factors() this loop.
-             */
-            active_worst_qchanged = 1;
-        }
-        else
-            active_worst_qchanged = 0;
+        /* If we have updated the active max Q do not call
+         * vp8_update_rate_correction_factors() this loop.
+         */
+      active_worst_qchanged = 1;
+    } else
+      active_worst_qchanged = 0;
 
 #if !CONFIG_REALTIME_ONLY
-        /* Special case handling for forced key frames */
-        if ( (cm->frame_type == KEY_FRAME) && cpi->this_key_frame_forced )
-        {
-            int last_q = Q;
-            int kf_err = vp8_calc_ss_err(cpi->Source,
-                                         &cm->yv12_fb[cm->new_fb_idx]);
+    /* Special case handling for forced key frames */
+    if ((cm->frame_type == KEY_FRAME) && cpi->this_key_frame_forced) {
+      int last_q = Q;
+      int kf_err = vp8_calc_ss_err(cpi->Source, &cm->yv12_fb[cm->new_fb_idx]);
 
-            /* The key frame is not good enough */
-            if ( kf_err > ((cpi->ambient_err * 7) >> 3) )
-            {
-                /* Lower q_high */
-                q_high = (Q > q_low) ? (Q - 1) : q_low;
+      /* The key frame is not good enough */
+      if (kf_err > ((cpi->ambient_err * 7) >> 3)) {
+        /* Lower q_high */
+        q_high = (Q > q_low) ? (Q - 1) : q_low;
 
-                /* Adjust Q */
-                Q = (q_high + q_low) >> 1;
-            }
-            /* The key frame is much better than the previous frame */
-            else if ( kf_err < (cpi->ambient_err >> 1) )
-            {
-                /* Raise q_low */
-                q_low = (Q < q_high) ? (Q + 1) : q_high;
+        /* Adjust Q */
+        Q = (q_high + q_low) >> 1;
+      }
+      /* The key frame is much better than the previous frame */
+      else if (kf_err < (cpi->ambient_err >> 1)) {
+        /* Raise q_low */
+        q_low = (Q < q_high) ? (Q + 1) : q_high;
 
-                /* Adjust Q */
-                Q = (q_high + q_low + 1) >> 1;
-            }
+        /* Adjust Q */
+        Q = (q_high + q_low + 1) >> 1;
+      }
 
-            /* Clamp Q to upper and lower limits: */
-            if (Q > q_high)
-                Q = q_high;
-            else if (Q < q_low)
-                Q = q_low;
+      /* Clamp Q to upper and lower limits: */
+      if (Q > q_high)
+        Q = q_high;
+      else if (Q < q_low)
+        Q = q_low;
 
-            Loop = Q != last_q;
-        }
+      Loop = Q != last_q;
+    }
 
-        /* Is the projected frame size out of range and are we allowed
-         * to attempt to recode.
-         */
-        else if ( recode_loop_test( cpi,
-                               frame_over_shoot_limit, frame_under_shoot_limit,
-                               Q, top_index, bottom_index ) )
-        {
-            int last_q = Q;
-            int Retries = 0;
+    /* Is the projected frame size out of range and are we allowed
+     * to attempt to recode.
+     */
+    else if (recode_loop_test(cpi, frame_over_shoot_limit,
+                              frame_under_shoot_limit, Q, top_index,
+                              bottom_index)) {
+      int last_q = Q;
+      int Retries = 0;
 
-            /* Frame size out of permitted range. Update correction factor
-             * & compute new Q to try...
-             */
+      /* Frame size out of permitted range. Update correction factor
+       * & compute new Q to try...
+       */
 
-            /* Frame is too large */
-            if (cpi->projected_frame_size > cpi->this_frame_target)
-            {
-                /* Raise Qlow as to at least the current value */
-                q_low = (Q < q_high) ? (Q + 1) : q_high;
+      /* Frame is too large */
+      if (cpi->projected_frame_size > cpi->this_frame_target) {
+        /* Raise Qlow as to at least the current value */
+        q_low = (Q < q_high) ? (Q + 1) : q_high;
 
-                /* If we are using over quant do the same for zbin_oq_low */
-                if (cpi->mb.zbin_over_quant > 0)
-                    zbin_oq_low = (cpi->mb.zbin_over_quant < zbin_oq_high) ?
-                        (cpi->mb.zbin_over_quant + 1) : zbin_oq_high;
+        /* If we are using over quant do the same for zbin_oq_low */
+        if (cpi->mb.zbin_over_quant > 0)
+          zbin_oq_low = (cpi->mb.zbin_over_quant < zbin_oq_high)
+                            ? (cpi->mb.zbin_over_quant + 1)
+                            : zbin_oq_high;
 
-                if (undershoot_seen)
-                {
-                    /* Update rate_correction_factor unless
-                     * cpi->active_worst_quality has changed.
-                     */
-                    if (!active_worst_qchanged)
-                        vp8_update_rate_correction_factors(cpi, 1);
+        if (undershoot_seen) {
+          /* Update rate_correction_factor unless
+           * cpi->active_worst_quality has changed.
+           */
+          if (!active_worst_qchanged)
+            vp8_update_rate_correction_factors(cpi, 1);
 
-                    Q = (q_high + q_low + 1) / 2;
+          Q = (q_high + q_low + 1) / 2;
 
-                    /* Adjust cpi->zbin_over_quant (only allowed when Q
-                     * is max)
-                     */
-                    if (Q < MAXQ)
-                        cpi->mb.zbin_over_quant = 0;
-                    else
-                    {
-                        zbin_oq_low = (cpi->mb.zbin_over_quant < zbin_oq_high) ?
-                            (cpi->mb.zbin_over_quant + 1) : zbin_oq_high;
-                        cpi->mb.zbin_over_quant =
-                            (zbin_oq_high + zbin_oq_low) / 2;
-                    }
-                }
-                else
-                {
-                    /* Update rate_correction_factor unless
-                     * cpi->active_worst_quality has changed.
-                     */
-                    if (!active_worst_qchanged)
-                        vp8_update_rate_correction_factors(cpi, 0);
+          /* Adjust cpi->zbin_over_quant (only allowed when Q
+           * is max)
+           */
+          if (Q < MAXQ)
+            cpi->mb.zbin_over_quant = 0;
+          else {
+            zbin_oq_low = (cpi->mb.zbin_over_quant < zbin_oq_high)
+                              ? (cpi->mb.zbin_over_quant + 1)
+                              : zbin_oq_high;
+            cpi->mb.zbin_over_quant = (zbin_oq_high + zbin_oq_low) / 2;
+          }
+        } else {
+          /* Update rate_correction_factor unless
+           * cpi->active_worst_quality has changed.
+           */
+          if (!active_worst_qchanged)
+            vp8_update_rate_correction_factors(cpi, 0);
 
-                    Q = vp8_regulate_q(cpi, cpi->this_frame_target);
+          Q = vp8_regulate_q(cpi, cpi->this_frame_target);
 
-                    while (((Q < q_low) ||
-                        (cpi->mb.zbin_over_quant < zbin_oq_low)) &&
-                        (Retries < 10))
-                    {
-                        vp8_update_rate_correction_factors(cpi, 0);
-                        Q = vp8_regulate_q(cpi, cpi->this_frame_target);
-                        Retries ++;
-                    }
-                }
+          while (((Q < q_low) || (cpi->mb.zbin_over_quant < zbin_oq_low)) &&
+                 (Retries < 10)) {
+            vp8_update_rate_correction_factors(cpi, 0);
+            Q = vp8_regulate_q(cpi, cpi->this_frame_target);
+            Retries++;
+          }
+        }
 
-                overshoot_seen = 1;
-            }
-            /* Frame is too small */
-            else
-            {
-                if (cpi->mb.zbin_over_quant == 0)
-                    /* Lower q_high if not using over quant */
-                    q_high = (Q > q_low) ? (Q - 1) : q_low;
-                else
-                    /* else lower zbin_oq_high */
-                    zbin_oq_high = (cpi->mb.zbin_over_quant > zbin_oq_low) ?
-                        (cpi->mb.zbin_over_quant - 1) : zbin_oq_low;
+        overshoot_seen = 1;
+      }
+      /* Frame is too small */
+      else {
+        if (cpi->mb.zbin_over_quant == 0)
+          /* Lower q_high if not using over quant */
+          q_high = (Q > q_low) ? (Q - 1) : q_low;
+        else
+          /* else lower zbin_oq_high */
+          zbin_oq_high = (cpi->mb.zbin_over_quant > zbin_oq_low)
+                             ? (cpi->mb.zbin_over_quant - 1)
+                             : zbin_oq_low;
 
-                if (overshoot_seen)
-                {
-                    /* Update rate_correction_factor unless
-                     * cpi->active_worst_quality has changed.
-                     */
-                    if (!active_worst_qchanged)
-                        vp8_update_rate_correction_factors(cpi, 1);
+        if (overshoot_seen) {
+          /* Update rate_correction_factor unless
+           * cpi->active_worst_quality has changed.
+           */
+          if (!active_worst_qchanged)
+            vp8_update_rate_correction_factors(cpi, 1);
 
-                    Q = (q_high + q_low) / 2;
+          Q = (q_high + q_low) / 2;
 
-                    /* Adjust cpi->zbin_over_quant (only allowed when Q
-                     * is max)
-                     */
-                    if (Q < MAXQ)
-                        cpi->mb.zbin_over_quant = 0;
-                    else
-                        cpi->mb.zbin_over_quant =
-                            (zbin_oq_high + zbin_oq_low) / 2;
-                }
-                else
-                {
-                    /* Update rate_correction_factor unless
-                     * cpi->active_worst_quality has changed.
-                     */
-                    if (!active_worst_qchanged)
-                        vp8_update_rate_correction_factors(cpi, 0);
+          /* Adjust cpi->zbin_over_quant (only allowed when Q
+           * is max)
+           */
+          if (Q < MAXQ)
+            cpi->mb.zbin_over_quant = 0;
+          else
+            cpi->mb.zbin_over_quant = (zbin_oq_high + zbin_oq_low) / 2;
+        } else {
+          /* Update rate_correction_factor unless
+           * cpi->active_worst_quality has changed.
+           */
+          if (!active_worst_qchanged)
+            vp8_update_rate_correction_factors(cpi, 0);
 
-                    Q = vp8_regulate_q(cpi, cpi->this_frame_target);
+          Q = vp8_regulate_q(cpi, cpi->this_frame_target);
 
-                    /* Special case reset for qlow for constrained quality.
-                     * This should only trigger where there is very substantial
-                     * undershoot on a frame and the auto cq level is above
-                     * the user passsed in value.
-                     */
-                    if ( (cpi->oxcf.end_usage == USAGE_CONSTRAINED_QUALITY) &&
-                         (Q < q_low) )
-                    {
-                        q_low = Q;
-                    }
+          /* Special case reset for qlow for constrained quality.
+           * This should only trigger where there is very substantial
+           * undershoot on a frame and the auto cq level is above
+           * the user passsed in value.
+           */
+          if ((cpi->oxcf.end_usage == USAGE_CONSTRAINED_QUALITY) &&
+              (Q < q_low)) {
+            q_low = Q;
+          }
 
-                    while (((Q > q_high) ||
-                        (cpi->mb.zbin_over_quant > zbin_oq_high)) &&
-                        (Retries < 10))
-                    {
-                        vp8_update_rate_correction_factors(cpi, 0);
-                        Q = vp8_regulate_q(cpi, cpi->this_frame_target);
-                        Retries ++;
-                    }
-                }
+          while (((Q > q_high) || (cpi->mb.zbin_over_quant > zbin_oq_high)) &&
+                 (Retries < 10)) {
+            vp8_update_rate_correction_factors(cpi, 0);
+            Q = vp8_regulate_q(cpi, cpi->this_frame_target);
+            Retries++;
+          }
+        }
 
-                undershoot_seen = 1;
-            }
+        undershoot_seen = 1;
+      }
 
-            /* Clamp Q to upper and lower limits: */
-            if (Q > q_high)
-                Q = q_high;
-            else if (Q < q_low)
-                Q = q_low;
+      /* Clamp Q to upper and lower limits: */
+      if (Q > q_high)
+        Q = q_high;
+      else if (Q < q_low)
+        Q = q_low;
 
-            /* Clamp cpi->zbin_over_quant */
-            cpi->mb.zbin_over_quant = (cpi->mb.zbin_over_quant < zbin_oq_low) ?
-                zbin_oq_low : (cpi->mb.zbin_over_quant > zbin_oq_high) ?
-                    zbin_oq_high : cpi->mb.zbin_over_quant;
+      /* Clamp cpi->zbin_over_quant */
+      cpi->mb.zbin_over_quant = (cpi->mb.zbin_over_quant < zbin_oq_low)
+                                    ? zbin_oq_low
+                                    : (cpi->mb.zbin_over_quant > zbin_oq_high)
+                                          ? zbin_oq_high
+                                          : cpi->mb.zbin_over_quant;
 
-            Loop = Q != last_q;
-        }
-        else
+      Loop = Q != last_q;
+    } else
 #endif
-            Loop = 0;
+      Loop = 0;
 
-        if (cpi->is_src_frame_alt_ref)
-            Loop = 0;
+    if (cpi->is_src_frame_alt_ref) Loop = 0;
 
-        if (Loop == 1)
-        {
-            vp8_restore_coding_context(cpi);
-            loop_count++;
+    if (Loop == 1) {
+      vp8_restore_coding_context(cpi);
+      loop_count++;
 #if CONFIG_INTERNAL_STATS
-            cpi->tot_recode_hits++;
+      cpi->tot_recode_hits++;
 #endif
-        }
     }
-    while (Loop == 1);
+  } while (Loop == 1);
 
 #if 0
     /* Experimental code for lagged and one pass
@@ -4641,86 +4182,81 @@
     }
 #endif
 
-    /* Special case code to reduce pulsing when key frames are forced at a
-     * fixed interval. Note the reconstruction error if it is the frame before
-     * the force key frame
-     */
-    if ( cpi->next_key_frame_forced && (cpi->twopass.frames_to_key == 0) )
-    {
-        cpi->ambient_err = vp8_calc_ss_err(cpi->Source,
-                                           &cm->yv12_fb[cm->new_fb_idx]);
-    }
+  /* Special case code to reduce pulsing when key frames are forced at a
+   * fixed interval. Note the reconstruction error if it is the frame before
+   * the force key frame
+   */
+  if (cpi->next_key_frame_forced && (cpi->twopass.frames_to_key == 0)) {
+    cpi->ambient_err =
+        vp8_calc_ss_err(cpi->Source, &cm->yv12_fb[cm->new_fb_idx]);
+  }
 
-    /* This frame's MVs are saved and will be used in next frame's MV predictor.
-     * Last frame has one more line(add to bottom) and one more column(add to
-     * right) than cm->mip. The edge elements are initialized to 0.
-     */
+/* This frame's MVs are saved and will be used in next frame's MV predictor.
+ * Last frame has one more line(add to bottom) and one more column(add to
+ * right) than cm->mip. The edge elements are initialized to 0.
+ */
 #if CONFIG_MULTI_RES_ENCODING
-    if(!cpi->oxcf.mr_encoder_id && cm->show_frame)
+  if (!cpi->oxcf.mr_encoder_id && cm->show_frame)
 #else
-    if(cm->show_frame)   /* do not save for altref frame */
+  if (cm->show_frame) /* do not save for altref frame */
 #endif
-    {
-        int mb_row;
-        int mb_col;
-        /* Point to beginning of allocated MODE_INFO arrays. */
-        MODE_INFO *tmp = cm->mip;
+  {
+    int mb_row;
+    int mb_col;
+    /* Point to beginning of allocated MODE_INFO arrays. */
+    MODE_INFO *tmp = cm->mip;
 
-        if(cm->frame_type != KEY_FRAME)
-        {
-            for (mb_row = 0; mb_row < cm->mb_rows+1; mb_row ++)
-            {
-                for (mb_col = 0; mb_col < cm->mb_cols+1; mb_col ++)
-                {
-                    if(tmp->mbmi.ref_frame != INTRA_FRAME)
-                        cpi->lfmv[mb_col + mb_row*(cm->mode_info_stride+1)].as_int = tmp->mbmi.mv.as_int;
+    if (cm->frame_type != KEY_FRAME) {
+      for (mb_row = 0; mb_row < cm->mb_rows + 1; mb_row++) {
+        for (mb_col = 0; mb_col < cm->mb_cols + 1; mb_col++) {
+          if (tmp->mbmi.ref_frame != INTRA_FRAME)
+            cpi->lfmv[mb_col + mb_row * (cm->mode_info_stride + 1)].as_int =
+                tmp->mbmi.mv.as_int;
 
-                    cpi->lf_ref_frame_sign_bias[mb_col + mb_row*(cm->mode_info_stride+1)] = cm->ref_frame_sign_bias[tmp->mbmi.ref_frame];
-                    cpi->lf_ref_frame[mb_col + mb_row*(cm->mode_info_stride+1)] = tmp->mbmi.ref_frame;
-                    tmp++;
-                }
-            }
+          cpi->lf_ref_frame_sign_bias[mb_col +
+                                      mb_row * (cm->mode_info_stride + 1)] =
+              cm->ref_frame_sign_bias[tmp->mbmi.ref_frame];
+          cpi->lf_ref_frame[mb_col + mb_row * (cm->mode_info_stride + 1)] =
+              tmp->mbmi.ref_frame;
+          tmp++;
         }
+      }
     }
+  }
 
-    /* Count last ref frame 0,0 usage on current encoded frame. */
-    {
-        int mb_row;
-        int mb_col;
-        /* Point to beginning of MODE_INFO arrays. */
-        MODE_INFO *tmp = cm->mi;
+  /* Count last ref frame 0,0 usage on current encoded frame. */
+  {
+    int mb_row;
+    int mb_col;
+    /* Point to beginning of MODE_INFO arrays. */
+    MODE_INFO *tmp = cm->mi;
 
-        cpi->zeromv_count = 0;
+    cpi->zeromv_count = 0;
 
-        if(cm->frame_type != KEY_FRAME)
-        {
-            for (mb_row = 0; mb_row < cm->mb_rows; mb_row ++)
-            {
-                for (mb_col = 0; mb_col < cm->mb_cols; mb_col ++)
-                {
-                    if (tmp->mbmi.mode == ZEROMV &&
-                       tmp->mbmi.ref_frame == LAST_FRAME)
-                        cpi->zeromv_count++;
-                    tmp++;
-                }
-                tmp++;
-            }
+    if (cm->frame_type != KEY_FRAME) {
+      for (mb_row = 0; mb_row < cm->mb_rows; mb_row++) {
+        for (mb_col = 0; mb_col < cm->mb_cols; mb_col++) {
+          if (tmp->mbmi.mode == ZEROMV && tmp->mbmi.ref_frame == LAST_FRAME)
+            cpi->zeromv_count++;
+          tmp++;
         }
+        tmp++;
+      }
     }
+  }
 
 #if CONFIG_MULTI_RES_ENCODING
-    vp8_cal_dissimilarity(cpi);
+  vp8_cal_dissimilarity(cpi);
 #endif
 
-    /* Update the GF useage maps.
-     * This is done after completing the compression of a frame when all
-     * modes etc. are finalized but before loop filter
-     */
-    if (cpi->oxcf.number_of_layers == 1)
-        vp8_update_gf_useage_maps(cpi, cm, &cpi->mb);
+  /* Update the GF useage maps.
+   * This is done after completing the compression of a frame when all
+   * modes etc. are finalized but before loop filter
+   */
+  if (cpi->oxcf.number_of_layers == 1)
+    vp8_update_gf_useage_maps(cpi, cm, &cpi->mb);
 
-    if (cm->frame_type == KEY_FRAME)
-        cm->refresh_last_frame = 1;
+  if (cm->frame_type == KEY_FRAME) cm->refresh_last_frame = 1;
 
 #if 0
     {
@@ -4730,265 +4266,247 @@
     }
 #endif
 
-    /* For inter frames the current default behavior is that when
-     * cm->refresh_golden_frame is set we copy the old GF over to the ARF buffer
-     * This is purely an encoder decision at present.
-     */
-    if (!cpi->oxcf.error_resilient_mode && cm->refresh_golden_frame)
-        cm->copy_buffer_to_arf  = 2;
-    else
-        cm->copy_buffer_to_arf  = 0;
+  /* For inter frames the current default behavior is that when
+   * cm->refresh_golden_frame is set we copy the old GF over to the ARF buffer
+   * This is purely an encoder decision at present.
+   */
+  if (!cpi->oxcf.error_resilient_mode && cm->refresh_golden_frame)
+    cm->copy_buffer_to_arf = 2;
+  else
+    cm->copy_buffer_to_arf = 0;
 
-    cm->frame_to_show = &cm->yv12_fb[cm->new_fb_idx];
+  cm->frame_to_show = &cm->yv12_fb[cm->new_fb_idx];
 
 #if CONFIG_TEMPORAL_DENOISING
-    // Get some measure of the amount of noise, by measuring the (partial) mse
-    // between source and denoised buffer, for y channel. Partial refers to
-    // computing the sse for a sub-sample of the frame (i.e., skip x blocks along row/column),
-    // and only for blocks in that set that are consecutive ZEROMV_LAST mode.
-    // Do this every ~8 frames, to further reduce complexity.
-    // TODO(marpan): Keep this for now for the case cpi->oxcf.noise_sensitivity < 4,
-    // should be removed in favor of the process_denoiser_mode_change() function below.
-    if (cpi->oxcf.noise_sensitivity > 0 &&
-       cpi->oxcf.noise_sensitivity < 4 &&
-       !cpi->oxcf.screen_content_mode &&
-       cpi->frames_since_key%8 == 0 &&
-       cm->frame_type != KEY_FRAME) {
-       cpi->mse_source_denoised = measure_square_diff_partial(
-           &cpi->denoiser.yv12_running_avg[INTRA_FRAME], cpi->Source, cpi);
-    }
+  // Get some measure of the amount of noise, by measuring the (partial) mse
+  // between source and denoised buffer, for y channel. Partial refers to
+  // computing the sse for a sub-sample of the frame (i.e., skip x blocks along
+  // row/column),
+  // and only for blocks in that set that are consecutive ZEROMV_LAST mode.
+  // Do this every ~8 frames, to further reduce complexity.
+  // TODO(marpan): Keep this for now for the case cpi->oxcf.noise_sensitivity <
+  // 4,
+  // should be removed in favor of the process_denoiser_mode_change() function
+  // below.
+  if (cpi->oxcf.noise_sensitivity > 0 && cpi->oxcf.noise_sensitivity < 4 &&
+      !cpi->oxcf.screen_content_mode && cpi->frames_since_key % 8 == 0 &&
+      cm->frame_type != KEY_FRAME) {
+    cpi->mse_source_denoised = measure_square_diff_partial(
+        &cpi->denoiser.yv12_running_avg[INTRA_FRAME], cpi->Source, cpi);
+  }
 
-    // For the adaptive denoising mode (noise_sensitivity == 4), sample the mse
-    // of source diff (between current and previous frame), and determine if we
-    // should switch the denoiser mode. Sampling refers to computing the mse for
-    // a sub-sample of the frame (i.e., skip x blocks along row/column), and
-    // only for blocks in that set that have used ZEROMV LAST, along with some
-    // constraint on the sum diff between blocks. This process is called every
-    // ~8 frames, to further reduce complexity.
-    if (cpi->oxcf.noise_sensitivity == 4 &&
-        !cpi->oxcf.screen_content_mode &&
-        cpi->frames_since_key % 8 == 0 &&
-        cm->frame_type != KEY_FRAME) {
-      process_denoiser_mode_change(cpi);
-    }
+  // For the adaptive denoising mode (noise_sensitivity == 4), sample the mse
+  // of source diff (between current and previous frame), and determine if we
+  // should switch the denoiser mode. Sampling refers to computing the mse for
+  // a sub-sample of the frame (i.e., skip x blocks along row/column), and
+  // only for blocks in that set that have used ZEROMV LAST, along with some
+  // constraint on the sum diff between blocks. This process is called every
+  // ~8 frames, to further reduce complexity.
+  if (cpi->oxcf.noise_sensitivity == 4 && !cpi->oxcf.screen_content_mode &&
+      cpi->frames_since_key % 8 == 0 && cm->frame_type != KEY_FRAME) {
+    process_denoiser_mode_change(cpi);
+  }
 #endif
 
 #if CONFIG_MULTITHREAD
-    if (cpi->b_multi_threaded)
-    {
-        /* start loopfilter in separate thread */
-        sem_post(&cpi->h_event_start_lpf);
-    }
-    else
+  if (cpi->b_multi_threaded) {
+    /* start loopfilter in separate thread */
+    sem_post(&cpi->h_event_start_lpf);
+  } else
 #endif
-    {
-        vp8_loopfilter_frame(cpi, cm);
-    }
+  {
+    vp8_loopfilter_frame(cpi, cm);
+  }
 
-    update_reference_frames(cpi);
+  update_reference_frames(cpi);
 
 #ifdef OUTPUT_YUV_DENOISED
-    vp8_write_yuv_frame(yuv_denoised_file,
-                        &cpi->denoiser.yv12_running_avg[INTRA_FRAME]);
+  vp8_write_yuv_frame(yuv_denoised_file,
+                      &cpi->denoiser.yv12_running_avg[INTRA_FRAME]);
 #endif
 
 #if !(CONFIG_REALTIME_ONLY & CONFIG_ONTHEFLY_BITPACKING)
-    if (cpi->oxcf.error_resilient_mode)
-    {
-        cm->refresh_entropy_probs = 0;
-    }
+  if (cpi->oxcf.error_resilient_mode) {
+    cm->refresh_entropy_probs = 0;
+  }
 #endif
 
 #if CONFIG_MULTITHREAD
-    /* wait that filter_level is picked so that we can continue with stream packing */
-    if (cpi->b_multi_threaded)
-        sem_wait(&cpi->h_event_end_lpf);
+  /* wait that filter_level is picked so that we can continue with stream
+   * packing */
+  if (cpi->b_multi_threaded) sem_wait(&cpi->h_event_end_lpf);
 #endif
 
-    /* build the bitstream */
-    vp8_pack_bitstream(cpi, dest, dest_end, size);
+  /* build the bitstream */
+  vp8_pack_bitstream(cpi, dest, dest_end, size);
 
 #if CONFIG_MULTITHREAD
-    /* wait for the lpf thread done */
-    if (cpi->b_multi_threaded)
-    {
-        sem_wait(&cpi->h_event_end_lpf);
-    }
+  /* wait for the lpf thread done */
+  if (cpi->b_multi_threaded) {
+    sem_wait(&cpi->h_event_end_lpf);
+  }
 #endif
 
-    /* Move storing frame_type out of the above loop since it is also
-     * needed in motion search besides loopfilter */
-    cm->last_frame_type = cm->frame_type;
+  /* Move storing frame_type out of the above loop since it is also
+   * needed in motion search besides loopfilter */
+  cm->last_frame_type = cm->frame_type;
 
-    /* Update rate control heuristics */
-    cpi->total_byte_count += (*size);
-    cpi->projected_frame_size = (*size) << 3;
+  /* Update rate control heuristics */
+  cpi->total_byte_count += (*size);
+  cpi->projected_frame_size = (*size) << 3;
 
-    if (cpi->oxcf.number_of_layers > 1)
-    {
-        unsigned int i;
-        for (i=cpi->current_layer+1; i<cpi->oxcf.number_of_layers; i++)
-          cpi->layer_context[i].total_byte_count += (*size);
-    }
+  if (cpi->oxcf.number_of_layers > 1) {
+    unsigned int i;
+    for (i = cpi->current_layer + 1; i < cpi->oxcf.number_of_layers; i++)
+      cpi->layer_context[i].total_byte_count += (*size);
+  }
 
-    if (!active_worst_qchanged)
-        vp8_update_rate_correction_factors(cpi, 2);
+  if (!active_worst_qchanged) vp8_update_rate_correction_factors(cpi, 2);
 
-    cpi->last_q[cm->frame_type] = cm->base_qindex;
+  cpi->last_q[cm->frame_type] = cm->base_qindex;
 
-    if (cm->frame_type == KEY_FRAME)
-    {
-        vp8_adjust_key_frame_context(cpi);
-    }
+  if (cm->frame_type == KEY_FRAME) {
+    vp8_adjust_key_frame_context(cpi);
+  }
 
-    /* Keep a record of ambient average Q. */
-    if (cm->frame_type != KEY_FRAME)
-        cpi->avg_frame_qindex = (2 + 3 * cpi->avg_frame_qindex + cm->base_qindex) >> 2;
+  /* Keep a record of ambient average Q. */
+  if (cm->frame_type != KEY_FRAME)
+    cpi->avg_frame_qindex =
+        (2 + 3 * cpi->avg_frame_qindex + cm->base_qindex) >> 2;
 
-    /* Keep a record from which we can calculate the average Q excluding
-     * GF updates and key frames
+  /* Keep a record from which we can calculate the average Q excluding
+   * GF updates and key frames
+   */
+  if ((cm->frame_type != KEY_FRAME) &&
+      ((cpi->oxcf.number_of_layers > 1) ||
+       (!cm->refresh_golden_frame && !cm->refresh_alt_ref_frame))) {
+    cpi->ni_frames++;
+
+    /* Calculate the average Q for normal inter frames (not key or GFU
+     * frames).
      */
-    if ((cm->frame_type != KEY_FRAME) && ((cpi->oxcf.number_of_layers > 1) ||
-        (!cm->refresh_golden_frame && !cm->refresh_alt_ref_frame)))
-    {
-        cpi->ni_frames++;
+    if (cpi->pass == 2) {
+      cpi->ni_tot_qi += Q;
+      cpi->ni_av_qi = (cpi->ni_tot_qi / cpi->ni_frames);
+    } else {
+      /* Damp value for first few frames */
+      if (cpi->ni_frames > 150) {
+        cpi->ni_tot_qi += Q;
+        cpi->ni_av_qi = (cpi->ni_tot_qi / cpi->ni_frames);
+      }
+      /* For one pass, early in the clip ... average the current frame Q
+       * value with the worstq entered by the user as a dampening measure
+       */
+      else {
+        cpi->ni_tot_qi += Q;
+        cpi->ni_av_qi =
+            ((cpi->ni_tot_qi / cpi->ni_frames) + cpi->worst_quality + 1) / 2;
+      }
 
-        /* Calculate the average Q for normal inter frames (not key or GFU
-         * frames).
-         */
-        if ( cpi->pass == 2 )
-        {
-            cpi->ni_tot_qi += Q;
-            cpi->ni_av_qi = (cpi->ni_tot_qi / cpi->ni_frames);
-        }
-        else
-        {
-            /* Damp value for first few frames */
-            if (cpi->ni_frames > 150 )
-            {
-                cpi->ni_tot_qi += Q;
-                cpi->ni_av_qi = (cpi->ni_tot_qi / cpi->ni_frames);
-            }
-            /* For one pass, early in the clip ... average the current frame Q
-             * value with the worstq entered by the user as a dampening measure
-             */
-            else
-            {
-                cpi->ni_tot_qi += Q;
-                cpi->ni_av_qi = ((cpi->ni_tot_qi / cpi->ni_frames) + cpi->worst_quality + 1) / 2;
-            }
-
-            /* If the average Q is higher than what was used in the last
-             * frame (after going through the recode loop to keep the frame
-             * size within range) then use the last frame value - 1. The -1
-             * is designed to stop Q and hence the data rate, from
-             * progressively falling away during difficult sections, but at
-             * the same time reduce the number of itterations around the
-             * recode loop.
-             */
-            if (Q > cpi->ni_av_qi)
-                cpi->ni_av_qi = Q - 1;
-        }
+      /* If the average Q is higher than what was used in the last
+       * frame (after going through the recode loop to keep the frame
+       * size within range) then use the last frame value - 1. The -1
+       * is designed to stop Q and hence the data rate, from
+       * progressively falling away during difficult sections, but at
+       * the same time reduce the number of itterations around the
+       * recode loop.
+       */
+      if (Q > cpi->ni_av_qi) cpi->ni_av_qi = Q - 1;
     }
+  }
 
-    /* Update the buffer level variable. */
-    /* Non-viewable frames are a special case and are treated as pure overhead. */
-    if ( !cm->show_frame )
-        cpi->bits_off_target -= cpi->projected_frame_size;
-    else
-        cpi->bits_off_target += cpi->av_per_frame_bandwidth - cpi->projected_frame_size;
+  /* Update the buffer level variable. */
+  /* Non-viewable frames are a special case and are treated as pure overhead. */
+  if (!cm->show_frame)
+    cpi->bits_off_target -= cpi->projected_frame_size;
+  else
+    cpi->bits_off_target +=
+        cpi->av_per_frame_bandwidth - cpi->projected_frame_size;
 
-    /* Clip the buffer level to the maximum specified buffer size */
-    if (cpi->bits_off_target > cpi->oxcf.maximum_buffer_size)
-        cpi->bits_off_target = cpi->oxcf.maximum_buffer_size;
+  /* Clip the buffer level to the maximum specified buffer size */
+  if (cpi->bits_off_target > cpi->oxcf.maximum_buffer_size)
+    cpi->bits_off_target = cpi->oxcf.maximum_buffer_size;
 
-    // If the frame dropper is not enabled, don't let the buffer level go below
-    // some threshold, given here by -|maximum_buffer_size|. For now we only do
-    // this for screen content input.
-    if (cpi->drop_frames_allowed == 0 && cpi->oxcf.screen_content_mode &&
-        cpi->bits_off_target < -cpi->oxcf.maximum_buffer_size)
-        cpi->bits_off_target = -cpi->oxcf.maximum_buffer_size;
+  // If the frame dropper is not enabled, don't let the buffer level go below
+  // some threshold, given here by -|maximum_buffer_size|. For now we only do
+  // this for screen content input.
+  if (cpi->drop_frames_allowed == 0 && cpi->oxcf.screen_content_mode &&
+      cpi->bits_off_target < -cpi->oxcf.maximum_buffer_size)
+    cpi->bits_off_target = -cpi->oxcf.maximum_buffer_size;
 
-    /* Rolling monitors of whether we are over or underspending used to
-     * help regulate min and Max Q in two pass.
-     */
-    cpi->rolling_target_bits = ((cpi->rolling_target_bits * 3) + cpi->this_frame_target + 2) / 4;
-    cpi->rolling_actual_bits = ((cpi->rolling_actual_bits * 3) + cpi->projected_frame_size + 2) / 4;
-    cpi->long_rolling_target_bits = ((cpi->long_rolling_target_bits * 31) + cpi->this_frame_target + 16) / 32;
-    cpi->long_rolling_actual_bits = ((cpi->long_rolling_actual_bits * 31) + cpi->projected_frame_size + 16) / 32;
+  /* Rolling monitors of whether we are over or underspending used to
+   * help regulate min and Max Q in two pass.
+   */
+  cpi->rolling_target_bits =
+      ((cpi->rolling_target_bits * 3) + cpi->this_frame_target + 2) / 4;
+  cpi->rolling_actual_bits =
+      ((cpi->rolling_actual_bits * 3) + cpi->projected_frame_size + 2) / 4;
+  cpi->long_rolling_target_bits =
+      ((cpi->long_rolling_target_bits * 31) + cpi->this_frame_target + 16) / 32;
+  cpi->long_rolling_actual_bits =
+      ((cpi->long_rolling_actual_bits * 31) + cpi->projected_frame_size + 16) /
+      32;
 
-    /* Actual bits spent */
-    cpi->total_actual_bits += cpi->projected_frame_size;
+  /* Actual bits spent */
+  cpi->total_actual_bits += cpi->projected_frame_size;
 
-    /* Debug stats */
-    cpi->total_target_vs_actual += (cpi->this_frame_target - cpi->projected_frame_size);
+  /* Debug stats */
+  cpi->total_target_vs_actual +=
+      (cpi->this_frame_target - cpi->projected_frame_size);
 
-    cpi->buffer_level = cpi->bits_off_target;
+  cpi->buffer_level = cpi->bits_off_target;
 
-    /* Propagate values to higher temporal layers */
-    if (cpi->oxcf.number_of_layers > 1)
-    {
-        unsigned int i;
+  /* Propagate values to higher temporal layers */
+  if (cpi->oxcf.number_of_layers > 1) {
+    unsigned int i;
 
-        for (i=cpi->current_layer+1; i<cpi->oxcf.number_of_layers; i++)
-        {
-            LAYER_CONTEXT *lc = &cpi->layer_context[i];
-            int bits_off_for_this_layer =
-               (int)(lc->target_bandwidth / lc->framerate -
-                     cpi->projected_frame_size);
+    for (i = cpi->current_layer + 1; i < cpi->oxcf.number_of_layers; i++) {
+      LAYER_CONTEXT *lc = &cpi->layer_context[i];
+      int bits_off_for_this_layer = (int)(lc->target_bandwidth / lc->framerate -
+                                          cpi->projected_frame_size);
 
-            lc->bits_off_target += bits_off_for_this_layer;
+      lc->bits_off_target += bits_off_for_this_layer;
 
-            /* Clip buffer level to maximum buffer size for the layer */
-            if (lc->bits_off_target > lc->maximum_buffer_size)
-                lc->bits_off_target = lc->maximum_buffer_size;
+      /* Clip buffer level to maximum buffer size for the layer */
+      if (lc->bits_off_target > lc->maximum_buffer_size)
+        lc->bits_off_target = lc->maximum_buffer_size;
 
-            lc->total_actual_bits += cpi->projected_frame_size;
-            lc->total_target_vs_actual += bits_off_for_this_layer;
-            lc->buffer_level = lc->bits_off_target;
-        }
+      lc->total_actual_bits += cpi->projected_frame_size;
+      lc->total_target_vs_actual += bits_off_for_this_layer;
+      lc->buffer_level = lc->bits_off_target;
     }
+  }
 
-    /* Update bits left to the kf and gf groups to account for overshoot
-     * or undershoot on these frames
-     */
-    if (cm->frame_type == KEY_FRAME)
-    {
-        cpi->twopass.kf_group_bits += cpi->this_frame_target - cpi->projected_frame_size;
+  /* Update bits left to the kf and gf groups to account for overshoot
+   * or undershoot on these frames
+   */
+  if (cm->frame_type == KEY_FRAME) {
+    cpi->twopass.kf_group_bits +=
+        cpi->this_frame_target - cpi->projected_frame_size;
 
-        if (cpi->twopass.kf_group_bits < 0)
-            cpi->twopass.kf_group_bits = 0 ;
-    }
-    else if (cm->refresh_golden_frame || cm->refresh_alt_ref_frame)
-    {
-        cpi->twopass.gf_group_bits += cpi->this_frame_target - cpi->projected_frame_size;
+    if (cpi->twopass.kf_group_bits < 0) cpi->twopass.kf_group_bits = 0;
+  } else if (cm->refresh_golden_frame || cm->refresh_alt_ref_frame) {
+    cpi->twopass.gf_group_bits +=
+        cpi->this_frame_target - cpi->projected_frame_size;
 
-        if (cpi->twopass.gf_group_bits < 0)
-            cpi->twopass.gf_group_bits = 0 ;
-    }
+    if (cpi->twopass.gf_group_bits < 0) cpi->twopass.gf_group_bits = 0;
+  }
 
-    if (cm->frame_type != KEY_FRAME)
-    {
-        if (cpi->common.refresh_alt_ref_frame)
-        {
-            cpi->last_skip_false_probs[2] = cpi->prob_skip_false;
-            cpi->last_skip_probs_q[2] = cm->base_qindex;
-        }
-        else if (cpi->common.refresh_golden_frame)
-        {
-            cpi->last_skip_false_probs[1] = cpi->prob_skip_false;
-            cpi->last_skip_probs_q[1] = cm->base_qindex;
-        }
-        else
-        {
-            cpi->last_skip_false_probs[0] = cpi->prob_skip_false;
-            cpi->last_skip_probs_q[0] = cm->base_qindex;
+  if (cm->frame_type != KEY_FRAME) {
+    if (cpi->common.refresh_alt_ref_frame) {
+      cpi->last_skip_false_probs[2] = cpi->prob_skip_false;
+      cpi->last_skip_probs_q[2] = cm->base_qindex;
+    } else if (cpi->common.refresh_golden_frame) {
+      cpi->last_skip_false_probs[1] = cpi->prob_skip_false;
+      cpi->last_skip_probs_q[1] = cm->base_qindex;
+    } else {
+      cpi->last_skip_false_probs[0] = cpi->prob_skip_false;
+      cpi->last_skip_probs_q[0] = cm->base_qindex;
 
-            /* update the baseline */
-            cpi->base_skip_false_prob[cm->base_qindex] = cpi->prob_skip_false;
-
-        }
+      /* update the baseline */
+      cpi->base_skip_false_prob[cm->base_qindex] = cpi->prob_skip_false;
     }
+  }
 
 #if 0 && CONFIG_INTERNAL_STATS
     {
@@ -5055,98 +4573,83 @@
 
 #endif
 
-    if (cm->refresh_golden_frame == 1)
-        cm->frame_flags = cm->frame_flags | FRAMEFLAGS_GOLDEN;
-    else
-        cm->frame_flags = cm->frame_flags&~FRAMEFLAGS_GOLDEN;
+  if (cm->refresh_golden_frame == 1)
+    cm->frame_flags = cm->frame_flags | FRAMEFLAGS_GOLDEN;
+  else
+    cm->frame_flags = cm->frame_flags & ~FRAMEFLAGS_GOLDEN;
 
-    if (cm->refresh_alt_ref_frame == 1)
-        cm->frame_flags = cm->frame_flags | FRAMEFLAGS_ALTREF;
-    else
-        cm->frame_flags = cm->frame_flags&~FRAMEFLAGS_ALTREF;
+  if (cm->refresh_alt_ref_frame == 1)
+    cm->frame_flags = cm->frame_flags | FRAMEFLAGS_ALTREF;
+  else
+    cm->frame_flags = cm->frame_flags & ~FRAMEFLAGS_ALTREF;
 
+  if (cm->refresh_last_frame & cm->refresh_golden_frame) /* both refreshed */
+    cpi->gold_is_last = 1;
+  else if (cm->refresh_last_frame ^ cm->refresh_golden_frame)
+    /* 1 refreshed but not the other */
+    cpi->gold_is_last = 0;
 
-    if (cm->refresh_last_frame & cm->refresh_golden_frame)
-        /* both refreshed */
-        cpi->gold_is_last = 1;
-    else if (cm->refresh_last_frame ^ cm->refresh_golden_frame)
-        /* 1 refreshed but not the other */
-        cpi->gold_is_last = 0;
+  if (cm->refresh_last_frame & cm->refresh_alt_ref_frame) /* both refreshed */
+    cpi->alt_is_last = 1;
+  else if (cm->refresh_last_frame ^ cm->refresh_alt_ref_frame)
+    /* 1 refreshed but not the other */
+    cpi->alt_is_last = 0;
 
-    if (cm->refresh_last_frame & cm->refresh_alt_ref_frame)
-        /* both refreshed */
-        cpi->alt_is_last = 1;
-    else if (cm->refresh_last_frame ^ cm->refresh_alt_ref_frame)
-        /* 1 refreshed but not the other */
-        cpi->alt_is_last = 0;
+  if (cm->refresh_alt_ref_frame & cm->refresh_golden_frame) /* both refreshed */
+    cpi->gold_is_alt = 1;
+  else if (cm->refresh_alt_ref_frame ^ cm->refresh_golden_frame)
+    /* 1 refreshed but not the other */
+    cpi->gold_is_alt = 0;
 
-    if (cm->refresh_alt_ref_frame & cm->refresh_golden_frame)
-        /* both refreshed */
-        cpi->gold_is_alt = 1;
-    else if (cm->refresh_alt_ref_frame ^ cm->refresh_golden_frame)
-        /* 1 refreshed but not the other */
-        cpi->gold_is_alt = 0;
+  cpi->ref_frame_flags = VP8_ALTR_FRAME | VP8_GOLD_FRAME | VP8_LAST_FRAME;
 
-    cpi->ref_frame_flags = VP8_ALTR_FRAME | VP8_GOLD_FRAME | VP8_LAST_FRAME;
+  if (cpi->gold_is_last) cpi->ref_frame_flags &= ~VP8_GOLD_FRAME;
 
-    if (cpi->gold_is_last)
-        cpi->ref_frame_flags &= ~VP8_GOLD_FRAME;
+  if (cpi->alt_is_last) cpi->ref_frame_flags &= ~VP8_ALTR_FRAME;
 
-    if (cpi->alt_is_last)
-        cpi->ref_frame_flags &= ~VP8_ALTR_FRAME;
+  if (cpi->gold_is_alt) cpi->ref_frame_flags &= ~VP8_ALTR_FRAME;
 
-    if (cpi->gold_is_alt)
-        cpi->ref_frame_flags &= ~VP8_ALTR_FRAME;
-
-
-    if (!cpi->oxcf.error_resilient_mode)
-    {
-        if (cpi->oxcf.play_alternate && cm->refresh_alt_ref_frame && (cm->frame_type != KEY_FRAME))
-            /* Update the alternate reference frame stats as appropriate. */
-            update_alt_ref_frame_stats(cpi);
-        else
-            /* Update the Golden frame stats as appropriate. */
-            update_golden_frame_stats(cpi);
-    }
-
-    if (cm->frame_type == KEY_FRAME)
-    {
-        /* Tell the caller that the frame was coded as a key frame */
-        *frame_flags = cm->frame_flags | FRAMEFLAGS_KEY;
-
-        /* As this frame is a key frame  the next defaults to an inter frame. */
-        cm->frame_type = INTER_FRAME;
-
-        cpi->last_frame_percent_intra = 100;
-    }
+  if (!cpi->oxcf.error_resilient_mode) {
+    if (cpi->oxcf.play_alternate && cm->refresh_alt_ref_frame &&
+        (cm->frame_type != KEY_FRAME))
+      /* Update the alternate reference frame stats as appropriate. */
+      update_alt_ref_frame_stats(cpi);
     else
-    {
-        *frame_flags = cm->frame_flags&~FRAMEFLAGS_KEY;
+      /* Update the Golden frame stats as appropriate. */
+      update_golden_frame_stats(cpi);
+  }
 
-        cpi->last_frame_percent_intra = cpi->this_frame_percent_intra;
-    }
+  if (cm->frame_type == KEY_FRAME) {
+    /* Tell the caller that the frame was coded as a key frame */
+    *frame_flags = cm->frame_flags | FRAMEFLAGS_KEY;
 
-    /* Clear the one shot update flags for segmentation map and mode/ref
-     * loop filter deltas.
-     */
-    cpi->mb.e_mbd.update_mb_segmentation_map = 0;
-    cpi->mb.e_mbd.update_mb_segmentation_data = 0;
-    cpi->mb.e_mbd.mode_ref_lf_delta_update = 0;
+    /* As this frame is a key frame  the next defaults to an inter frame. */
+    cm->frame_type = INTER_FRAME;
 
+    cpi->last_frame_percent_intra = 100;
+  } else {
+    *frame_flags = cm->frame_flags & ~FRAMEFLAGS_KEY;
 
-    /* Dont increment frame counters if this was an altref buffer update
-     * not a real frame
-     */
-    if (cm->show_frame)
-    {
-        cm->current_video_frame++;
-        cpi->frames_since_key++;
-        cpi->temporal_pattern_counter++;
-    }
+    cpi->last_frame_percent_intra = cpi->this_frame_percent_intra;
+  }
 
-    /* reset to normal state now that we are done. */
+  /* Clear the one shot update flags for segmentation map and mode/ref
+   * loop filter deltas.
+   */
+  cpi->mb.e_mbd.update_mb_segmentation_map = 0;
+  cpi->mb.e_mbd.update_mb_segmentation_data = 0;
+  cpi->mb.e_mbd.mode_ref_lf_delta_update = 0;
 
+  /* Dont increment frame counters if this was an altref buffer update
+   * not a real frame
+   */
+  if (cm->show_frame) {
+    cm->current_video_frame++;
+    cpi->frames_since_key++;
+    cpi->temporal_pattern_counter++;
+  }
 
+/* reset to normal state now that we are done. */
 
 #if 0
     {
@@ -5160,327 +4663,285 @@
     }
 #endif
 
-    /* DEBUG */
-    /* vp8_write_yuv_frame("encoder_recon.yuv", cm->frame_to_show); */
-
-
+  /* DEBUG */
+  /* vp8_write_yuv_frame("encoder_recon.yuv", cm->frame_to_show); */
 }
 #if !CONFIG_REALTIME_ONLY
-static void Pass2Encode(VP8_COMP *cpi, unsigned long *size, unsigned char *dest, unsigned char * dest_end, unsigned int *frame_flags)
-{
+static void Pass2Encode(VP8_COMP *cpi, unsigned long *size, unsigned char *dest,
+                        unsigned char *dest_end, unsigned int *frame_flags) {
+  if (!cpi->common.refresh_alt_ref_frame) vp8_second_pass(cpi);
 
-    if (!cpi->common.refresh_alt_ref_frame)
-        vp8_second_pass(cpi);
+  encode_frame_to_data_rate(cpi, size, dest, dest_end, frame_flags);
+  cpi->twopass.bits_left -= 8 * (int)(*size);
 
-    encode_frame_to_data_rate(cpi, size, dest, dest_end, frame_flags);
-    cpi->twopass.bits_left -= 8 * (int)(*size);
-
-    if (!cpi->common.refresh_alt_ref_frame)
-    {
-        double two_pass_min_rate = (double)(cpi->oxcf.target_bandwidth
-            *cpi->oxcf.two_pass_vbrmin_section / 100);
-        cpi->twopass.bits_left += (int64_t)(two_pass_min_rate / cpi->framerate);
-    }
+  if (!cpi->common.refresh_alt_ref_frame) {
+    double two_pass_min_rate =
+        (double)(cpi->oxcf.target_bandwidth *
+                 cpi->oxcf.two_pass_vbrmin_section / 100);
+    cpi->twopass.bits_left += (int64_t)(two_pass_min_rate / cpi->framerate);
+  }
 }
 #endif
 
-int vp8_receive_raw_frame(VP8_COMP *cpi, unsigned int frame_flags, YV12_BUFFER_CONFIG *sd, int64_t time_stamp, int64_t end_time)
-{
-    struct vpx_usec_timer  timer;
-    int                    res = 0;
+int vp8_receive_raw_frame(VP8_COMP *cpi, unsigned int frame_flags,
+                          YV12_BUFFER_CONFIG *sd, int64_t time_stamp,
+                          int64_t end_time) {
+  struct vpx_usec_timer timer;
+  int res = 0;
 
-    vpx_usec_timer_start(&timer);
+  vpx_usec_timer_start(&timer);
 
-    /* Reinit the lookahead buffer if the frame size changes */
-    if (sd->y_width != cpi->oxcf.Width || sd->y_height != cpi->oxcf.Height)
-    {
-        assert(cpi->oxcf.lag_in_frames < 2);
-        dealloc_raw_frame_buffers(cpi);
-        alloc_raw_frame_buffers(cpi);
-    }
+  /* Reinit the lookahead buffer if the frame size changes */
+  if (sd->y_width != cpi->oxcf.Width || sd->y_height != cpi->oxcf.Height) {
+    assert(cpi->oxcf.lag_in_frames < 2);
+    dealloc_raw_frame_buffers(cpi);
+    alloc_raw_frame_buffers(cpi);
+  }
 
-    if(vp8_lookahead_push(cpi->lookahead, sd, time_stamp, end_time,
-                          frame_flags, cpi->active_map_enabled ? cpi->active_map : NULL))
-        res = -1;
-    vpx_usec_timer_mark(&timer);
-    cpi->time_receive_data += vpx_usec_timer_elapsed(&timer);
+  if (vp8_lookahead_push(cpi->lookahead, sd, time_stamp, end_time, frame_flags,
+                         cpi->active_map_enabled ? cpi->active_map : NULL))
+    res = -1;
+  vpx_usec_timer_mark(&timer);
+  cpi->time_receive_data += vpx_usec_timer_elapsed(&timer);
 
-    return res;
+  return res;
 }
 
+static int frame_is_reference(const VP8_COMP *cpi) {
+  const VP8_COMMON *cm = &cpi->common;
+  const MACROBLOCKD *xd = &cpi->mb.e_mbd;
 
-static int frame_is_reference(const VP8_COMP *cpi)
-{
-    const VP8_COMMON *cm = &cpi->common;
-    const MACROBLOCKD *xd = &cpi->mb.e_mbd;
-
-    return cm->frame_type == KEY_FRAME || cm->refresh_last_frame
-           || cm->refresh_golden_frame || cm->refresh_alt_ref_frame
-           || cm->copy_buffer_to_gf || cm->copy_buffer_to_arf
-           || cm->refresh_entropy_probs
-           || xd->mode_ref_lf_delta_update
-           || xd->update_mb_segmentation_map || xd->update_mb_segmentation_data;
+  return cm->frame_type == KEY_FRAME || cm->refresh_last_frame ||
+         cm->refresh_golden_frame || cm->refresh_alt_ref_frame ||
+         cm->copy_buffer_to_gf || cm->copy_buffer_to_arf ||
+         cm->refresh_entropy_probs || xd->mode_ref_lf_delta_update ||
+         xd->update_mb_segmentation_map || xd->update_mb_segmentation_data;
 }
 
+int vp8_get_compressed_data(VP8_COMP *cpi, unsigned int *frame_flags,
+                            unsigned long *size, unsigned char *dest,
+                            unsigned char *dest_end, int64_t *time_stamp,
+                            int64_t *time_end, int flush) {
+  VP8_COMMON *cm;
+  struct vpx_usec_timer tsctimer;
+  struct vpx_usec_timer ticktimer;
+  struct vpx_usec_timer cmptimer;
+  YV12_BUFFER_CONFIG *force_src_buffer = NULL;
 
-int vp8_get_compressed_data(VP8_COMP *cpi, unsigned int *frame_flags, unsigned long *size, unsigned char *dest, unsigned char *dest_end, int64_t *time_stamp, int64_t *time_end, int flush)
-{
-    VP8_COMMON *cm;
-    struct vpx_usec_timer  tsctimer;
-    struct vpx_usec_timer  ticktimer;
-    struct vpx_usec_timer  cmptimer;
-    YV12_BUFFER_CONFIG    *force_src_buffer = NULL;
+  if (!cpi) return -1;
 
-    if (!cpi)
-        return -1;
+  cm = &cpi->common;
 
-    cm = &cpi->common;
+  if (setjmp(cpi->common.error.jmp)) {
+    cpi->common.error.setjmp = 0;
+    vp8_clear_system_state();
+    return VPX_CODEC_CORRUPT_FRAME;
+  }
 
-    if (setjmp(cpi->common.error.jmp))
-    {
-        cpi->common.error.setjmp = 0;
-        vp8_clear_system_state();
-        return VPX_CODEC_CORRUPT_FRAME;
-    }
+  cpi->common.error.setjmp = 1;
 
-    cpi->common.error.setjmp = 1;
+  vpx_usec_timer_start(&cmptimer);
 
-    vpx_usec_timer_start(&cmptimer);
+  cpi->source = NULL;
 
-    cpi->source = NULL;
-
 #if !CONFIG_REALTIME_ONLY
-    /* Should we code an alternate reference frame */
-    if (cpi->oxcf.error_resilient_mode == 0 &&
-        cpi->oxcf.play_alternate &&
-        cpi->source_alt_ref_pending)
-    {
-        if ((cpi->source = vp8_lookahead_peek(cpi->lookahead,
-                                              cpi->frames_till_gf_update_due,
-                                              PEEK_FORWARD)))
-        {
-            cpi->alt_ref_source = cpi->source;
-            if (cpi->oxcf.arnr_max_frames > 0)
-            {
-                vp8_temporal_filter_prepare_c(cpi,
-                                              cpi->frames_till_gf_update_due);
-                force_src_buffer = &cpi->alt_ref_buffer;
-            }
-            cpi->frames_till_alt_ref_frame = cpi->frames_till_gf_update_due;
-            cm->refresh_alt_ref_frame = 1;
-            cm->refresh_golden_frame = 0;
-            cm->refresh_last_frame = 0;
-            cm->show_frame = 0;
-            /* Clear Pending alt Ref flag. */
-            cpi->source_alt_ref_pending = 0;
-            cpi->is_src_frame_alt_ref = 0;
-        }
+  /* Should we code an alternate reference frame */
+  if (cpi->oxcf.error_resilient_mode == 0 && cpi->oxcf.play_alternate &&
+      cpi->source_alt_ref_pending) {
+    if ((cpi->source = vp8_lookahead_peek(
+             cpi->lookahead, cpi->frames_till_gf_update_due, PEEK_FORWARD))) {
+      cpi->alt_ref_source = cpi->source;
+      if (cpi->oxcf.arnr_max_frames > 0) {
+        vp8_temporal_filter_prepare_c(cpi, cpi->frames_till_gf_update_due);
+        force_src_buffer = &cpi->alt_ref_buffer;
+      }
+      cpi->frames_till_alt_ref_frame = cpi->frames_till_gf_update_due;
+      cm->refresh_alt_ref_frame = 1;
+      cm->refresh_golden_frame = 0;
+      cm->refresh_last_frame = 0;
+      cm->show_frame = 0;
+      /* Clear Pending alt Ref flag. */
+      cpi->source_alt_ref_pending = 0;
+      cpi->is_src_frame_alt_ref = 0;
     }
+  }
 #endif
 
-    if (!cpi->source)
-    {
-        /* Read last frame source if we are encoding first pass. */
-        if (cpi->pass == 1 && cm->current_video_frame > 0)
-        {
-            if((cpi->last_source = vp8_lookahead_peek(cpi->lookahead, 1,
-                                                      PEEK_BACKWARD)) == NULL)
-              return -1;
-        }
+  if (!cpi->source) {
+    /* Read last frame source if we are encoding first pass. */
+    if (cpi->pass == 1 && cm->current_video_frame > 0) {
+      if ((cpi->last_source =
+               vp8_lookahead_peek(cpi->lookahead, 1, PEEK_BACKWARD)) == NULL)
+        return -1;
+    }
 
+    if ((cpi->source = vp8_lookahead_pop(cpi->lookahead, flush))) {
+      cm->show_frame = 1;
 
-        if ((cpi->source = vp8_lookahead_pop(cpi->lookahead, flush)))
-        {
-            cm->show_frame = 1;
+      cpi->is_src_frame_alt_ref =
+          cpi->alt_ref_source && (cpi->source == cpi->alt_ref_source);
 
-            cpi->is_src_frame_alt_ref = cpi->alt_ref_source
-                                        && (cpi->source == cpi->alt_ref_source);
-
-            if(cpi->is_src_frame_alt_ref)
-                cpi->alt_ref_source = NULL;
-        }
+      if (cpi->is_src_frame_alt_ref) cpi->alt_ref_source = NULL;
     }
+  }
 
-    if (cpi->source)
-    {
-        cpi->Source = force_src_buffer ? force_src_buffer : &cpi->source->img;
-        cpi->un_scaled_source = cpi->Source;
-        *time_stamp = cpi->source->ts_start;
-        *time_end = cpi->source->ts_end;
-        *frame_flags = cpi->source->flags;
+  if (cpi->source) {
+    cpi->Source = force_src_buffer ? force_src_buffer : &cpi->source->img;
+    cpi->un_scaled_source = cpi->Source;
+    *time_stamp = cpi->source->ts_start;
+    *time_end = cpi->source->ts_end;
+    *frame_flags = cpi->source->flags;
 
-        if (cpi->pass == 1 && cm->current_video_frame > 0)
-        {
-            cpi->last_frame_unscaled_source = &cpi->last_source->img;
-        }
+    if (cpi->pass == 1 && cm->current_video_frame > 0) {
+      cpi->last_frame_unscaled_source = &cpi->last_source->img;
     }
-    else
-    {
-        *size = 0;
+  } else {
+    *size = 0;
 #if !CONFIG_REALTIME_ONLY
 
-        if (flush && cpi->pass == 1 && !cpi->twopass.first_pass_done)
-        {
-            vp8_end_first_pass(cpi);    /* get last stats packet */
-            cpi->twopass.first_pass_done = 1;
-        }
+    if (flush && cpi->pass == 1 && !cpi->twopass.first_pass_done) {
+      vp8_end_first_pass(cpi); /* get last stats packet */
+      cpi->twopass.first_pass_done = 1;
+    }
 
 #endif
 
-        return -1;
-    }
+    return -1;
+  }
 
-    if (cpi->source->ts_start < cpi->first_time_stamp_ever)
-    {
-        cpi->first_time_stamp_ever = cpi->source->ts_start;
-        cpi->last_end_time_stamp_seen = cpi->source->ts_start;
-    }
+  if (cpi->source->ts_start < cpi->first_time_stamp_ever) {
+    cpi->first_time_stamp_ever = cpi->source->ts_start;
+    cpi->last_end_time_stamp_seen = cpi->source->ts_start;
+  }
 
-    /* adjust frame rates based on timestamps given */
-    if (cm->show_frame)
-    {
-        int64_t this_duration;
-        int step = 0;
+  /* adjust frame rates based on timestamps given */
+  if (cm->show_frame) {
+    int64_t this_duration;
+    int step = 0;
 
-        if (cpi->source->ts_start == cpi->first_time_stamp_ever)
-        {
-            this_duration = cpi->source->ts_end - cpi->source->ts_start;
-            step = 1;
-        }
-        else
-        {
-            int64_t last_duration;
+    if (cpi->source->ts_start == cpi->first_time_stamp_ever) {
+      this_duration = cpi->source->ts_end - cpi->source->ts_start;
+      step = 1;
+    } else {
+      int64_t last_duration;
 
-            this_duration = cpi->source->ts_end - cpi->last_end_time_stamp_seen;
-            last_duration = cpi->last_end_time_stamp_seen
-                            - cpi->last_time_stamp_seen;
-            /* do a step update if the duration changes by 10% */
-            if (last_duration)
-                step = (int)(((this_duration - last_duration) *
-                            10 / last_duration));
-        }
+      this_duration = cpi->source->ts_end - cpi->last_end_time_stamp_seen;
+      last_duration = cpi->last_end_time_stamp_seen - cpi->last_time_stamp_seen;
+      /* do a step update if the duration changes by 10% */
+      if (last_duration)
+        step = (int)(((this_duration - last_duration) * 10 / last_duration));
+    }
 
-        if (this_duration)
-        {
-            if (step)
-                cpi->ref_framerate = 10000000.0 / this_duration;
-            else
-            {
-                double avg_duration, interval;
+    if (this_duration) {
+      if (step)
+        cpi->ref_framerate = 10000000.0 / this_duration;
+      else {
+        double avg_duration, interval;
 
-                /* Average this frame's rate into the last second's average
-                 * frame rate. If we haven't seen 1 second yet, then average
-                 * over the whole interval seen.
-                 */
-                interval = (double)(cpi->source->ts_end -
-                                    cpi->first_time_stamp_ever);
-                if(interval > 10000000.0)
-                    interval = 10000000;
+        /* Average this frame's rate into the last second's average
+         * frame rate. If we haven't seen 1 second yet, then average
+         * over the whole interval seen.
+         */
+        interval = (double)(cpi->source->ts_end - cpi->first_time_stamp_ever);
+        if (interval > 10000000.0) interval = 10000000;
 
-                avg_duration = 10000000.0 / cpi->ref_framerate;
-                avg_duration *= (interval - avg_duration + this_duration);
-                avg_duration /= interval;
+        avg_duration = 10000000.0 / cpi->ref_framerate;
+        avg_duration *= (interval - avg_duration + this_duration);
+        avg_duration /= interval;
 
-                cpi->ref_framerate = 10000000.0 / avg_duration;
-            }
+        cpi->ref_framerate = 10000000.0 / avg_duration;
+      }
 #if CONFIG_MULTI_RES_ENCODING
-            if (cpi->oxcf.mr_total_resolutions > 1) {
-              LOWER_RES_FRAME_INFO* low_res_frame_info = (LOWER_RES_FRAME_INFO*)
-                  cpi->oxcf.mr_low_res_mode_info;
-              // Frame rate should be the same for all spatial layers in
-              // multi-res-encoding (simulcast), so we constrain the frame for
-              // higher layers to be that of lowest resolution. This is needed
-              // as he application may decide to skip encoding a high layer and
-              // then start again, in which case a big jump in time-stamps will
-              // be received for that high layer, which will yield an incorrect
-              // frame rate (from time-stamp adjustment in above calculation).
-              if (cpi->oxcf.mr_encoder_id) {
-                 cpi->ref_framerate = low_res_frame_info->low_res_framerate;
-              }
-              else {
-                // Keep track of frame rate for lowest resolution.
-                low_res_frame_info->low_res_framerate = cpi->ref_framerate;
-              }
-            }
+      if (cpi->oxcf.mr_total_resolutions > 1) {
+        LOWER_RES_FRAME_INFO *low_res_frame_info =
+            (LOWER_RES_FRAME_INFO *)cpi->oxcf.mr_low_res_mode_info;
+        // Frame rate should be the same for all spatial layers in
+        // multi-res-encoding (simulcast), so we constrain the frame for
+        // higher layers to be that of lowest resolution. This is needed
+        // as he application may decide to skip encoding a high layer and
+        // then start again, in which case a big jump in time-stamps will
+        // be received for that high layer, which will yield an incorrect
+        // frame rate (from time-stamp adjustment in above calculation).
+        if (cpi->oxcf.mr_encoder_id) {
+          cpi->ref_framerate = low_res_frame_info->low_res_framerate;
+        } else {
+          // Keep track of frame rate for lowest resolution.
+          low_res_frame_info->low_res_framerate = cpi->ref_framerate;
+        }
+      }
 #endif
-            if (cpi->oxcf.number_of_layers > 1)
-            {
-                unsigned int i;
+      if (cpi->oxcf.number_of_layers > 1) {
+        unsigned int i;
 
-                /* Update frame rates for each layer */
-                assert(cpi->oxcf.number_of_layers <= VPX_TS_MAX_LAYERS);
-                for (i = 0; i < cpi->oxcf.number_of_layers &&
-                     i < VPX_TS_MAX_LAYERS; ++i)
-                {
-                    LAYER_CONTEXT *lc = &cpi->layer_context[i];
-                    lc->framerate = cpi->ref_framerate /
-                                    cpi->oxcf.rate_decimator[i];
-                }
-            }
-            else
-                vp8_new_framerate(cpi, cpi->ref_framerate);
+        /* Update frame rates for each layer */
+        assert(cpi->oxcf.number_of_layers <= VPX_TS_MAX_LAYERS);
+        for (i = 0; i < cpi->oxcf.number_of_layers && i < VPX_TS_MAX_LAYERS;
+             ++i) {
+          LAYER_CONTEXT *lc = &cpi->layer_context[i];
+          lc->framerate = cpi->ref_framerate / cpi->oxcf.rate_decimator[i];
         }
-
-        cpi->last_time_stamp_seen = cpi->source->ts_start;
-        cpi->last_end_time_stamp_seen = cpi->source->ts_end;
+      } else
+        vp8_new_framerate(cpi, cpi->ref_framerate);
     }
 
-    if (cpi->oxcf.number_of_layers > 1)
-    {
-        int layer;
+    cpi->last_time_stamp_seen = cpi->source->ts_start;
+    cpi->last_end_time_stamp_seen = cpi->source->ts_end;
+  }
 
-        update_layer_contexts (cpi);
+  if (cpi->oxcf.number_of_layers > 1) {
+    int layer;
 
-        /* Restore layer specific context & set frame rate */
-        if (cpi->temporal_layer_id >= 0) {
-          layer = cpi->temporal_layer_id;
-        } else {
-          layer = cpi->oxcf.layer_id[
-                  cpi->temporal_pattern_counter % cpi->oxcf.periodicity];
-        }
-        restore_layer_context (cpi, layer);
-        vp8_new_framerate(cpi, cpi->layer_context[layer].framerate);
-    }
+    update_layer_contexts(cpi);
 
-    if (cpi->compressor_speed == 2)
-    {
-        vpx_usec_timer_start(&tsctimer);
-        vpx_usec_timer_start(&ticktimer);
+    /* Restore layer specific context & set frame rate */
+    if (cpi->temporal_layer_id >= 0) {
+      layer = cpi->temporal_layer_id;
+    } else {
+      layer =
+          cpi->oxcf
+              .layer_id[cpi->temporal_pattern_counter % cpi->oxcf.periodicity];
     }
+    restore_layer_context(cpi, layer);
+    vp8_new_framerate(cpi, cpi->layer_context[layer].framerate);
+  }
 
-    cpi->lf_zeromv_pct = (cpi->zeromv_count * 100)/cm->MBs;
+  if (cpi->compressor_speed == 2) {
+    vpx_usec_timer_start(&tsctimer);
+    vpx_usec_timer_start(&ticktimer);
+  }
 
+  cpi->lf_zeromv_pct = (cpi->zeromv_count * 100) / cm->MBs;
+
 #if CONFIG_REALTIME_ONLY & CONFIG_ONTHEFLY_BITPACKING
-    {
-        int i;
-        const int num_part = (1 << cm->multi_token_partition);
-        /* the available bytes in dest */
-        const unsigned long dest_size = dest_end - dest;
-        const int tok_part_buff_size = (dest_size * 9) / (10 * num_part);
+  {
+    int i;
+    const int num_part = (1 << cm->multi_token_partition);
+    /* the available bytes in dest */
+    const unsigned long dest_size = dest_end - dest;
+    const int tok_part_buff_size = (dest_size * 9) / (10 * num_part);
 
-        unsigned char *dp = dest;
+    unsigned char *dp = dest;
 
-        cpi->partition_d[0] = dp;
-        dp += dest_size/10;         /* reserve 1/10 for control partition */
-        cpi->partition_d_end[0] = dp;
+    cpi->partition_d[0] = dp;
+    dp += dest_size / 10; /* reserve 1/10 for control partition */
+    cpi->partition_d_end[0] = dp;
 
-        for(i = 0; i < num_part; i++)
-        {
-            cpi->partition_d[i + 1] = dp;
-            dp += tok_part_buff_size;
-            cpi->partition_d_end[i + 1] = dp;
-        }
+    for (i = 0; i < num_part; i++) {
+      cpi->partition_d[i + 1] = dp;
+      dp += tok_part_buff_size;
+      cpi->partition_d_end[i + 1] = dp;
     }
+  }
 #endif
 
-    /* start with a 0 size frame */
-    *size = 0;
+  /* start with a 0 size frame */
+  *size = 0;
 
-    /* Clear down mmx registers */
-    vp8_clear_system_state();
+  /* Clear down mmx registers */
+  vp8_clear_system_state();
 
-    cm->frame_type = INTER_FRAME;
-    cm->frame_flags = *frame_flags;
+  cm->frame_type = INTER_FRAME;
+  cm->frame_flags = *frame_flags;
 
 #if 0
 
@@ -5496,211 +4957,185 @@
     }
 
 #endif
-    /* find a free buffer for the new frame */
-    {
-        int i = 0;
-        for(; i < NUM_YV12_BUFFERS; i++)
-        {
-            if(!cm->yv12_fb[i].flags)
-            {
-                cm->new_fb_idx = i;
-                break;
-            }
-        }
-
-        assert(i < NUM_YV12_BUFFERS );
+  /* find a free buffer for the new frame */
+  {
+    int i = 0;
+    for (; i < NUM_YV12_BUFFERS; i++) {
+      if (!cm->yv12_fb[i].flags) {
+        cm->new_fb_idx = i;
+        break;
+      }
     }
+
+    assert(i < NUM_YV12_BUFFERS);
+  }
 #if !CONFIG_REALTIME_ONLY
 
-    if (cpi->pass == 1)
-    {
-        Pass1Encode(cpi, size, dest, frame_flags);
-    }
-    else if (cpi->pass == 2)
-    {
-        Pass2Encode(cpi, size, dest, dest_end, frame_flags);
-    }
-    else
+  if (cpi->pass == 1) {
+    Pass1Encode(cpi, size, dest, frame_flags);
+  } else if (cpi->pass == 2) {
+    Pass2Encode(cpi, size, dest, dest_end, frame_flags);
+  } else
 #endif
-        encode_frame_to_data_rate(cpi, size, dest, dest_end, frame_flags);
+    encode_frame_to_data_rate(cpi, size, dest, dest_end, frame_flags);
 
-    if (cpi->compressor_speed == 2)
-    {
-        unsigned int duration, duration2;
-        vpx_usec_timer_mark(&tsctimer);
-        vpx_usec_timer_mark(&ticktimer);
+  if (cpi->compressor_speed == 2) {
+    unsigned int duration, duration2;
+    vpx_usec_timer_mark(&tsctimer);
+    vpx_usec_timer_mark(&ticktimer);
 
-        duration = (int)(vpx_usec_timer_elapsed(&ticktimer));
-        duration2 = (unsigned int)((double)duration / 2);
+    duration = (int)(vpx_usec_timer_elapsed(&ticktimer));
+    duration2 = (unsigned int)((double)duration / 2);
 
-        if (cm->frame_type != KEY_FRAME)
-        {
-            if (cpi->avg_encode_time == 0)
-                cpi->avg_encode_time = duration;
-            else
-                cpi->avg_encode_time = (7 * cpi->avg_encode_time + duration) >> 3;
-        }
-
-        if (duration2)
-        {
-            {
-
-                if (cpi->avg_pick_mode_time == 0)
-                    cpi->avg_pick_mode_time = duration2;
-                else
-                    cpi->avg_pick_mode_time = (7 * cpi->avg_pick_mode_time + duration2) >> 3;
-            }
-        }
-
+    if (cm->frame_type != KEY_FRAME) {
+      if (cpi->avg_encode_time == 0)
+        cpi->avg_encode_time = duration;
+      else
+        cpi->avg_encode_time = (7 * cpi->avg_encode_time + duration) >> 3;
     }
 
-    if (cm->refresh_entropy_probs == 0)
-    {
-        memcpy(&cm->fc, &cm->lfc, sizeof(cm->fc));
+    if (duration2) {
+      {
+        if (cpi->avg_pick_mode_time == 0)
+          cpi->avg_pick_mode_time = duration2;
+        else
+          cpi->avg_pick_mode_time =
+              (7 * cpi->avg_pick_mode_time + duration2) >> 3;
+      }
     }
+  }
 
-    /* Save the contexts separately for alt ref, gold and last. */
-    /* (TODO jbb -> Optimize this with pointers to avoid extra copies. ) */
-    if(cm->refresh_alt_ref_frame)
-        memcpy(&cpi->lfc_a, &cm->fc, sizeof(cm->fc));
+  if (cm->refresh_entropy_probs == 0) {
+    memcpy(&cm->fc, &cm->lfc, sizeof(cm->fc));
+  }
 
-    if(cm->refresh_golden_frame)
-        memcpy(&cpi->lfc_g, &cm->fc, sizeof(cm->fc));
+  /* Save the contexts separately for alt ref, gold and last. */
+  /* (TODO jbb -> Optimize this with pointers to avoid extra copies. ) */
+  if (cm->refresh_alt_ref_frame) memcpy(&cpi->lfc_a, &cm->fc, sizeof(cm->fc));
 
-    if(cm->refresh_last_frame)
-        memcpy(&cpi->lfc_n, &cm->fc, sizeof(cm->fc));
+  if (cm->refresh_golden_frame) memcpy(&cpi->lfc_g, &cm->fc, sizeof(cm->fc));
 
-    /* if its a dropped frame honor the requests on subsequent frames */
-    if (*size > 0)
-    {
-        cpi->droppable = !frame_is_reference(cpi);
+  if (cm->refresh_last_frame) memcpy(&cpi->lfc_n, &cm->fc, sizeof(cm->fc));
 
-        /* return to normal state */
-        cm->refresh_entropy_probs = 1;
-        cm->refresh_alt_ref_frame = 0;
-        cm->refresh_golden_frame = 0;
-        cm->refresh_last_frame = 1;
-        cm->frame_type = INTER_FRAME;
+  /* if its a dropped frame honor the requests on subsequent frames */
+  if (*size > 0) {
+    cpi->droppable = !frame_is_reference(cpi);
 
-    }
+    /* return to normal state */
+    cm->refresh_entropy_probs = 1;
+    cm->refresh_alt_ref_frame = 0;
+    cm->refresh_golden_frame = 0;
+    cm->refresh_last_frame = 1;
+    cm->frame_type = INTER_FRAME;
+  }
 
-    /* Save layer specific state */
-    if (cpi->oxcf.number_of_layers > 1)
-        save_layer_context (cpi);
+  /* Save layer specific state */
+  if (cpi->oxcf.number_of_layers > 1) save_layer_context(cpi);
 
-    vpx_usec_timer_mark(&cmptimer);
-    cpi->time_compress_data += vpx_usec_timer_elapsed(&cmptimer);
+  vpx_usec_timer_mark(&cmptimer);
+  cpi->time_compress_data += vpx_usec_timer_elapsed(&cmptimer);
 
-    if (cpi->b_calculate_psnr && cpi->pass != 1 && cm->show_frame)
-    {
-        generate_psnr_packet(cpi);
-    }
+  if (cpi->b_calculate_psnr && cpi->pass != 1 && cm->show_frame) {
+    generate_psnr_packet(cpi);
+  }
 
 #if CONFIG_INTERNAL_STATS
 
-    if (cpi->pass != 1)
-    {
-        cpi->bytes += *size;
+  if (cpi->pass != 1) {
+    cpi->bytes += *size;
 
-        if (cm->show_frame)
-        {
-            cpi->common.show_frame_mi = cpi->common.mi;
-            cpi->count ++;
+    if (cm->show_frame) {
+      cpi->common.show_frame_mi = cpi->common.mi;
+      cpi->count++;
 
-            if (cpi->b_calculate_psnr)
-            {
-                uint64_t ye,ue,ve;
-                double frame_psnr;
-                YV12_BUFFER_CONFIG      *orig = cpi->Source;
-                YV12_BUFFER_CONFIG      *recon = cpi->common.frame_to_show;
-                unsigned int y_width = cpi->common.Width;
-                unsigned int y_height = cpi->common.Height;
-                unsigned int uv_width = (y_width + 1) / 2;
-                unsigned int uv_height = (y_height + 1) / 2;
-                int y_samples = y_height * y_width;
-                int uv_samples = uv_height * uv_width;
-                int t_samples = y_samples + 2 * uv_samples;
-                double sq_error;
+      if (cpi->b_calculate_psnr) {
+        uint64_t ye, ue, ve;
+        double frame_psnr;
+        YV12_BUFFER_CONFIG *orig = cpi->Source;
+        YV12_BUFFER_CONFIG *recon = cpi->common.frame_to_show;
+        unsigned int y_width = cpi->common.Width;
+        unsigned int y_height = cpi->common.Height;
+        unsigned int uv_width = (y_width + 1) / 2;
+        unsigned int uv_height = (y_height + 1) / 2;
+        int y_samples = y_height * y_width;
+        int uv_samples = uv_height * uv_width;
+        int t_samples = y_samples + 2 * uv_samples;
+        double sq_error;
 
-                ye = calc_plane_error(orig->y_buffer, orig->y_stride,
-                  recon->y_buffer, recon->y_stride, y_width, y_height);
+        ye = calc_plane_error(orig->y_buffer, orig->y_stride, recon->y_buffer,
+                              recon->y_stride, y_width, y_height);
 
-                ue = calc_plane_error(orig->u_buffer, orig->uv_stride,
-                  recon->u_buffer, recon->uv_stride, uv_width, uv_height);
+        ue = calc_plane_error(orig->u_buffer, orig->uv_stride, recon->u_buffer,
+                              recon->uv_stride, uv_width, uv_height);
 
-                ve = calc_plane_error(orig->v_buffer, orig->uv_stride,
-                  recon->v_buffer, recon->uv_stride, uv_width, uv_height);
+        ve = calc_plane_error(orig->v_buffer, orig->uv_stride, recon->v_buffer,
+                              recon->uv_stride, uv_width, uv_height);
 
-                sq_error = (double)(ye + ue + ve);
+        sq_error = (double)(ye + ue + ve);
 
-                frame_psnr = vpx_sse_to_psnr(t_samples, 255.0, sq_error);
+        frame_psnr = vpx_sse_to_psnr(t_samples, 255.0, sq_error);
 
-                cpi->total_y += vpx_sse_to_psnr(y_samples, 255.0, (double)ye);
-                cpi->total_u += vpx_sse_to_psnr(uv_samples, 255.0, (double)ue);
-                cpi->total_v += vpx_sse_to_psnr(uv_samples, 255.0, (double)ve);
-                cpi->total_sq_error += sq_error;
-                cpi->total  += frame_psnr;
+        cpi->total_y += vpx_sse_to_psnr(y_samples, 255.0, (double)ye);
+        cpi->total_u += vpx_sse_to_psnr(uv_samples, 255.0, (double)ue);
+        cpi->total_v += vpx_sse_to_psnr(uv_samples, 255.0, (double)ve);
+        cpi->total_sq_error += sq_error;
+        cpi->total += frame_psnr;
 #if CONFIG_POSTPROC
-                {
-                    YV12_BUFFER_CONFIG      *pp = &cm->post_proc_buffer;
-                    double sq_error2;
-                    double frame_psnr2, frame_ssim2 = 0;
-                    double weight = 0;
+        {
+          YV12_BUFFER_CONFIG *pp = &cm->post_proc_buffer;
+          double sq_error2;
+          double frame_psnr2, frame_ssim2 = 0;
+          double weight = 0;
 
-                    vp8_deblock(cm, cm->frame_to_show, &cm->post_proc_buffer, cm->filter_level * 10 / 6, 1, 0);
-                    vp8_clear_system_state();
+          vp8_deblock(cm, cm->frame_to_show, &cm->post_proc_buffer,
+                      cm->filter_level * 10 / 6, 1, 0);
+          vp8_clear_system_state();
 
-                    ye = calc_plane_error(orig->y_buffer, orig->y_stride,
-                      pp->y_buffer, pp->y_stride, y_width, y_height);
+          ye = calc_plane_error(orig->y_buffer, orig->y_stride, pp->y_buffer,
+                                pp->y_stride, y_width, y_height);
 
-                    ue = calc_plane_error(orig->u_buffer, orig->uv_stride,
-                      pp->u_buffer, pp->uv_stride, uv_width, uv_height);
+          ue = calc_plane_error(orig->u_buffer, orig->uv_stride, pp->u_buffer,
+                                pp->uv_stride, uv_width, uv_height);
 
-                    ve = calc_plane_error(orig->v_buffer, orig->uv_stride,
-                      pp->v_buffer, pp->uv_stride, uv_width, uv_height);
+          ve = calc_plane_error(orig->v_buffer, orig->uv_stride, pp->v_buffer,
+                                pp->uv_stride, uv_width, uv_height);
 
-                    sq_error2 = (double)(ye + ue + ve);
+          sq_error2 = (double)(ye + ue + ve);
 
-                    frame_psnr2 = vpx_sse_to_psnr(t_samples, 255.0, sq_error2);
+          frame_psnr2 = vpx_sse_to_psnr(t_samples, 255.0, sq_error2);
 
-                    cpi->totalp_y += vpx_sse_to_psnr(y_samples,
-                                                     255.0, (double)ye);
-                    cpi->totalp_u += vpx_sse_to_psnr(uv_samples,
-                                                     255.0, (double)ue);
-                    cpi->totalp_v += vpx_sse_to_psnr(uv_samples,
-                                                     255.0, (double)ve);
-                    cpi->total_sq_error2 += sq_error2;
-                    cpi->totalp  += frame_psnr2;
+          cpi->totalp_y += vpx_sse_to_psnr(y_samples, 255.0, (double)ye);
+          cpi->totalp_u += vpx_sse_to_psnr(uv_samples, 255.0, (double)ue);
+          cpi->totalp_v += vpx_sse_to_psnr(uv_samples, 255.0, (double)ve);
+          cpi->total_sq_error2 += sq_error2;
+          cpi->totalp += frame_psnr2;
 
-                    frame_ssim2 = vpx_calc_ssim(cpi->Source,
-                      &cm->post_proc_buffer, &weight);
+          frame_ssim2 =
+              vpx_calc_ssim(cpi->Source, &cm->post_proc_buffer, &weight);
 
-                    cpi->summed_quality += frame_ssim2 * weight;
-                    cpi->summed_weights += weight;
+          cpi->summed_quality += frame_ssim2 * weight;
+          cpi->summed_weights += weight;
 
-                    if (cpi->oxcf.number_of_layers > 1)
-                    {
-                         unsigned int i;
+          if (cpi->oxcf.number_of_layers > 1) {
+            unsigned int i;
 
-                         for (i=cpi->current_layer;
-                                       i<cpi->oxcf.number_of_layers; i++)
-                         {
-                             cpi->frames_in_layer[i]++;
+            for (i = cpi->current_layer; i < cpi->oxcf.number_of_layers; i++) {
+              cpi->frames_in_layer[i]++;
 
-                             cpi->bytes_in_layer[i] += *size;
-                             cpi->sum_psnr[i]       += frame_psnr;
-                             cpi->sum_psnr_p[i]     += frame_psnr2;
-                             cpi->total_error2[i]   += sq_error;
-                             cpi->total_error2_p[i] += sq_error2;
-                             cpi->sum_ssim[i]       += frame_ssim2 * weight;
-                             cpi->sum_weights[i]    += weight;
-                         }
-                    }
-                }
-#endif
+              cpi->bytes_in_layer[i] += *size;
+              cpi->sum_psnr[i] += frame_psnr;
+              cpi->sum_psnr_p[i] += frame_psnr2;
+              cpi->total_error2[i] += sq_error;
+              cpi->total_error2_p[i] += sq_error2;
+              cpi->sum_ssim[i] += frame_ssim2 * weight;
+              cpi->sum_weights[i] += weight;
             }
+          }
         }
+#endif
+      }
     }
+  }
 
 #if 0
 
@@ -5727,176 +5162,157 @@
 #endif
 #endif
 
-    cpi->common.error.setjmp = 0;
+  cpi->common.error.setjmp = 0;
 
-    return 0;
+  return 0;
 }
 
-int vp8_get_preview_raw_frame(VP8_COMP *cpi, YV12_BUFFER_CONFIG *dest, vp8_ppflags_t *flags)
-{
-    if (cpi->common.refresh_alt_ref_frame)
-        return -1;
-    else
-    {
-        int ret;
+int vp8_get_preview_raw_frame(VP8_COMP *cpi, YV12_BUFFER_CONFIG *dest,
+                              vp8_ppflags_t *flags) {
+  if (cpi->common.refresh_alt_ref_frame)
+    return -1;
+  else {
+    int ret;
 
 #if CONFIG_POSTPROC
-        cpi->common.show_frame_mi = cpi->common.mi;
-        ret = vp8_post_proc_frame(&cpi->common, dest, flags);
+    cpi->common.show_frame_mi = cpi->common.mi;
+    ret = vp8_post_proc_frame(&cpi->common, dest, flags);
 #else
-        (void)flags;
+    (void)flags;
 
-        if (cpi->common.frame_to_show)
-        {
-            *dest = *cpi->common.frame_to_show;
-            dest->y_width = cpi->common.Width;
-            dest->y_height = cpi->common.Height;
-            dest->uv_height = cpi->common.Height / 2;
-            ret = 0;
-        }
-        else
-        {
-            ret = -1;
-        }
+    if (cpi->common.frame_to_show) {
+      *dest = *cpi->common.frame_to_show;
+      dest->y_width = cpi->common.Width;
+      dest->y_height = cpi->common.Height;
+      dest->uv_height = cpi->common.Height / 2;
+      ret = 0;
+    } else {
+      ret = -1;
+    }
 
 #endif
-        vp8_clear_system_state();
-        return ret;
-    }
+    vp8_clear_system_state();
+    return ret;
+  }
 }
 
-int vp8_set_roimap(VP8_COMP *cpi, unsigned char *map, unsigned int rows, unsigned int cols, int delta_q[4], int delta_lf[4], unsigned int threshold[4])
-{
-    signed char feature_data[MB_LVL_MAX][MAX_MB_SEGMENTS];
-    int internal_delta_q[MAX_MB_SEGMENTS];
-    const int range = 63;
-    int i;
+int vp8_set_roimap(VP8_COMP *cpi, unsigned char *map, unsigned int rows,
+                   unsigned int cols, int delta_q[4], int delta_lf[4],
+                   unsigned int threshold[4]) {
+  signed char feature_data[MB_LVL_MAX][MAX_MB_SEGMENTS];
+  int internal_delta_q[MAX_MB_SEGMENTS];
+  const int range = 63;
+  int i;
 
-    // This method is currently incompatible with the cyclic refresh method
-    if ( cpi->cyclic_refresh_mode_enabled )
-        return -1;
+  // This method is currently incompatible with the cyclic refresh method
+  if (cpi->cyclic_refresh_mode_enabled) return -1;
 
-    // Check number of rows and columns match
-    if (cpi->common.mb_rows != (int)rows || cpi->common.mb_cols != (int)cols)
-        return -1;
+  // Check number of rows and columns match
+  if (cpi->common.mb_rows != (int)rows || cpi->common.mb_cols != (int)cols)
+    return -1;
 
-    // Range check the delta Q values and convert the external Q range values
-    // to internal ones.
-    if ( (abs(delta_q[0]) > range) || (abs(delta_q[1]) > range) ||
-         (abs(delta_q[2]) > range) || (abs(delta_q[3]) > range) )
-        return -1;
+  // Range check the delta Q values and convert the external Q range values
+  // to internal ones.
+  if ((abs(delta_q[0]) > range) || (abs(delta_q[1]) > range) ||
+      (abs(delta_q[2]) > range) || (abs(delta_q[3]) > range))
+    return -1;
 
-    // Range check the delta lf values
-    if ( (abs(delta_lf[0]) > range) || (abs(delta_lf[1]) > range) ||
-         (abs(delta_lf[2]) > range) || (abs(delta_lf[3]) > range) )
-        return -1;
+  // Range check the delta lf values
+  if ((abs(delta_lf[0]) > range) || (abs(delta_lf[1]) > range) ||
+      (abs(delta_lf[2]) > range) || (abs(delta_lf[3]) > range))
+    return -1;
 
-    if (!map)
-    {
-        disable_segmentation(cpi);
-        return 0;
-    }
+  if (!map) {
+    disable_segmentation(cpi);
+    return 0;
+  }
 
-    // Translate the external delta q values to internal values.
-    for ( i = 0; i < MAX_MB_SEGMENTS; i++ )
-        internal_delta_q[i] =
-            ( delta_q[i] >= 0 ) ? q_trans[delta_q[i]] : -q_trans[-delta_q[i]];
+  // Translate the external delta q values to internal values.
+  for (i = 0; i < MAX_MB_SEGMENTS; i++)
+    internal_delta_q[i] =
+        (delta_q[i] >= 0) ? q_trans[delta_q[i]] : -q_trans[-delta_q[i]];
 
-    /* Set the segmentation Map */
-    set_segmentation_map(cpi, map);
+  /* Set the segmentation Map */
+  set_segmentation_map(cpi, map);
 
-    /* Activate segmentation. */
-    enable_segmentation(cpi);
+  /* Activate segmentation. */
+  enable_segmentation(cpi);
 
-    /* Set up the quant segment data */
-    feature_data[MB_LVL_ALT_Q][0] = internal_delta_q[0];
-    feature_data[MB_LVL_ALT_Q][1] = internal_delta_q[1];
-    feature_data[MB_LVL_ALT_Q][2] = internal_delta_q[2];
-    feature_data[MB_LVL_ALT_Q][3] = internal_delta_q[3];
+  /* Set up the quant segment data */
+  feature_data[MB_LVL_ALT_Q][0] = internal_delta_q[0];
+  feature_data[MB_LVL_ALT_Q][1] = internal_delta_q[1];
+  feature_data[MB_LVL_ALT_Q][2] = internal_delta_q[2];
+  feature_data[MB_LVL_ALT_Q][3] = internal_delta_q[3];
 
-    /* Set up the loop segment data s */
-    feature_data[MB_LVL_ALT_LF][0] = delta_lf[0];
-    feature_data[MB_LVL_ALT_LF][1] = delta_lf[1];
-    feature_data[MB_LVL_ALT_LF][2] = delta_lf[2];
-    feature_data[MB_LVL_ALT_LF][3] = delta_lf[3];
+  /* Set up the loop segment data s */
+  feature_data[MB_LVL_ALT_LF][0] = delta_lf[0];
+  feature_data[MB_LVL_ALT_LF][1] = delta_lf[1];
+  feature_data[MB_LVL_ALT_LF][2] = delta_lf[2];
+  feature_data[MB_LVL_ALT_LF][3] = delta_lf[3];
 
-    cpi->segment_encode_breakout[0] = threshold[0];
-    cpi->segment_encode_breakout[1] = threshold[1];
-    cpi->segment_encode_breakout[2] = threshold[2];
-    cpi->segment_encode_breakout[3] = threshold[3];
+  cpi->segment_encode_breakout[0] = threshold[0];
+  cpi->segment_encode_breakout[1] = threshold[1];
+  cpi->segment_encode_breakout[2] = threshold[2];
+  cpi->segment_encode_breakout[3] = threshold[3];
 
-    /* Initialise the feature data structure */
-    set_segment_data(cpi, &feature_data[0][0], SEGMENT_DELTADATA);
+  /* Initialise the feature data structure */
+  set_segment_data(cpi, &feature_data[0][0], SEGMENT_DELTADATA);
 
-    return 0;
+  return 0;
 }
 
-int vp8_set_active_map(VP8_COMP *cpi, unsigned char *map, unsigned int rows, unsigned int cols)
-{
-    if ((int)rows == cpi->common.mb_rows && (int)cols == cpi->common.mb_cols)
-    {
-        if (map)
-        {
-            memcpy(cpi->active_map, map, rows * cols);
-            cpi->active_map_enabled = 1;
-        }
-        else
-            cpi->active_map_enabled = 0;
+int vp8_set_active_map(VP8_COMP *cpi, unsigned char *map, unsigned int rows,
+                       unsigned int cols) {
+  if ((int)rows == cpi->common.mb_rows && (int)cols == cpi->common.mb_cols) {
+    if (map) {
+      memcpy(cpi->active_map, map, rows * cols);
+      cpi->active_map_enabled = 1;
+    } else
+      cpi->active_map_enabled = 0;
 
-        return 0;
-    }
-    else
-    {
-        return -1 ;
-    }
+    return 0;
+  } else {
+    return -1;
+  }
 }
 
-int vp8_set_internal_size(VP8_COMP *cpi, VPX_SCALING horiz_mode, VPX_SCALING vert_mode)
-{
-    if (horiz_mode <= ONETWO)
-        cpi->common.horiz_scale = horiz_mode;
-    else
-        return -1;
+int vp8_set_internal_size(VP8_COMP *cpi, VPX_SCALING horiz_mode,
+                          VPX_SCALING vert_mode) {
+  if (horiz_mode <= ONETWO)
+    cpi->common.horiz_scale = horiz_mode;
+  else
+    return -1;
 
-    if (vert_mode <= ONETWO)
-        cpi->common.vert_scale  = vert_mode;
-    else
-        return -1;
+  if (vert_mode <= ONETWO)
+    cpi->common.vert_scale = vert_mode;
+  else
+    return -1;
 
-    return 0;
+  return 0;
 }
 
+int vp8_calc_ss_err(YV12_BUFFER_CONFIG *source, YV12_BUFFER_CONFIG *dest) {
+  int i, j;
+  int Total = 0;
 
+  unsigned char *src = source->y_buffer;
+  unsigned char *dst = dest->y_buffer;
 
-int vp8_calc_ss_err(YV12_BUFFER_CONFIG *source, YV12_BUFFER_CONFIG *dest)
-{
-    int i, j;
-    int Total = 0;
-
-    unsigned char *src = source->y_buffer;
-    unsigned char *dst = dest->y_buffer;
-
-    /* Loop through the Y plane raw and reconstruction data summing
-     * (square differences)
-     */
-    for (i = 0; i < source->y_height; i += 16)
-    {
-        for (j = 0; j < source->y_width; j += 16)
-        {
-            unsigned int sse;
-            Total += vpx_mse16x16(src + j, source->y_stride,
-                                  dst + j, dest->y_stride, &sse);
-        }
-
-        src += 16 * source->y_stride;
-        dst += 16 * dest->y_stride;
+  /* Loop through the Y plane raw and reconstruction data summing
+   * (square differences)
+   */
+  for (i = 0; i < source->y_height; i += 16) {
+    for (j = 0; j < source->y_width; j += 16) {
+      unsigned int sse;
+      Total += vpx_mse16x16(src + j, source->y_stride, dst + j, dest->y_stride,
+                            &sse);
     }
 
-    return Total;
-}
+    src += 16 * source->y_stride;
+    dst += 16 * dest->y_stride;
+  }
 
-
-int vp8_get_quantizer(VP8_COMP *cpi)
-{
-    return cpi->common.base_qindex;
+  return Total;
 }
+
+int vp8_get_quantizer(VP8_COMP *cpi) { return cpi->common.base_qindex; }
--- a/vp8/encoder/onyx_int.h
+++ b/vp8/encoder/onyx_int.h
@@ -8,7 +8,6 @@
  *  be found in the AUTHORS file in the root of the source tree.
  */
 
-
 #ifndef VP8_ENCODER_ONYX_INT_H_
 #define VP8_ENCODER_ONYX_INT_H_
 
@@ -37,24 +36,23 @@
 extern "C" {
 #endif
 
-#define MIN_GF_INTERVAL             4
-#define DEFAULT_GF_INTERVAL         7
+#define MIN_GF_INTERVAL 4
+#define DEFAULT_GF_INTERVAL 7
 
 #define KEY_FRAME_CONTEXT 5
 
-#define MAX_LAG_BUFFERS (CONFIG_REALTIME_ONLY? 1 : 25)
+#define MAX_LAG_BUFFERS (CONFIG_REALTIME_ONLY ? 1 : 25)
 
-#define AF_THRESH   25
-#define AF_THRESH2  100
+#define AF_THRESH 25
+#define AF_THRESH2 100
 #define ARF_DECAY_THRESH 12
 
+#define MIN_THRESHMULT 32
+#define MAX_THRESHMULT 512
 
-#define MIN_THRESHMULT  32
-#define MAX_THRESHMULT  512
-
 #define GF_ZEROMV_ZBIN_BOOST 12
 #define LF_ZEROMV_ZBIN_BOOST 6
-#define MV_ZBIN_BOOST        4
+#define MV_ZBIN_BOOST 4
 #define ZBIN_OQ_MAX 192
 
 #if !(CONFIG_REALTIME_ONLY)
@@ -61,399 +59,377 @@
 #define VP8_TEMPORAL_ALT_REF 1
 #endif
 
-typedef struct
-{
-    int kf_indicated;
-    unsigned int frames_since_key;
-    unsigned int frames_since_golden;
-    int filter_level;
-    int frames_till_gf_update_due;
-    int recent_ref_frame_usage[MAX_REF_FRAMES];
+typedef struct {
+  int kf_indicated;
+  unsigned int frames_since_key;
+  unsigned int frames_since_golden;
+  int filter_level;
+  int frames_till_gf_update_due;
+  int recent_ref_frame_usage[MAX_REF_FRAMES];
 
-    MV_CONTEXT mvc[2];
-    int mvcosts[2][MVvals+1];
+  MV_CONTEXT mvc[2];
+  int mvcosts[2][MVvals + 1];
 
 #ifdef MODE_STATS
-    int y_modes[5];
-    int uv_modes[4];
-    int b_modes[10];
-    int inter_y_modes[10];
-    int inter_uv_modes[4];
-    int inter_b_modes[10];
+  int y_modes[5];
+  int uv_modes[4];
+  int b_modes[10];
+  int inter_y_modes[10];
+  int inter_uv_modes[4];
+  int inter_b_modes[10];
 #endif
 
-    vp8_prob ymode_prob[4], uv_mode_prob[3];   /* interframe intra mode probs */
-    vp8_prob kf_ymode_prob[4], kf_uv_mode_prob[3];   /* keyframe "" */
+  vp8_prob ymode_prob[4], uv_mode_prob[3]; /* interframe intra mode probs */
+  vp8_prob kf_ymode_prob[4], kf_uv_mode_prob[3]; /* keyframe "" */
 
-    int ymode_count[5], uv_mode_count[4];  /* intra MB type cts this frame */
+  int ymode_count[5], uv_mode_count[4]; /* intra MB type cts this frame */
 
-    int count_mb_ref_frame_usage[MAX_REF_FRAMES];
+  int count_mb_ref_frame_usage[MAX_REF_FRAMES];
 
-    int this_frame_percent_intra;
-    int last_frame_percent_intra;
+  int this_frame_percent_intra;
+  int last_frame_percent_intra;
 
-
 } CODING_CONTEXT;
 
-typedef struct
-{
-    double frame;
-    double intra_error;
-    double coded_error;
-    double ssim_weighted_pred_err;
-    double pcnt_inter;
-    double pcnt_motion;
-    double pcnt_second_ref;
-    double pcnt_neutral;
-    double MVr;
-    double mvr_abs;
-    double MVc;
-    double mvc_abs;
-    double MVrv;
-    double MVcv;
-    double mv_in_out_count;
-    double new_mv_count;
-    double duration;
-    double count;
-}
-FIRSTPASS_STATS;
+typedef struct {
+  double frame;
+  double intra_error;
+  double coded_error;
+  double ssim_weighted_pred_err;
+  double pcnt_inter;
+  double pcnt_motion;
+  double pcnt_second_ref;
+  double pcnt_neutral;
+  double MVr;
+  double mvr_abs;
+  double MVc;
+  double mvc_abs;
+  double MVrv;
+  double MVcv;
+  double mv_in_out_count;
+  double new_mv_count;
+  double duration;
+  double count;
+} FIRSTPASS_STATS;
 
-typedef struct
-{
-    int frames_so_far;
-    double frame_intra_error;
-    double frame_coded_error;
-    double frame_pcnt_inter;
-    double frame_pcnt_motion;
-    double frame_mvr;
-    double frame_mvr_abs;
-    double frame_mvc;
-    double frame_mvc_abs;
+typedef struct {
+  int frames_so_far;
+  double frame_intra_error;
+  double frame_coded_error;
+  double frame_pcnt_inter;
+  double frame_pcnt_motion;
+  double frame_mvr;
+  double frame_mvr_abs;
+  double frame_mvc;
+  double frame_mvc_abs;
 
 } ONEPASS_FRAMESTATS;
 
+typedef enum {
+  THR_ZERO1 = 0,
+  THR_DC = 1,
 
-typedef enum
-{
-    THR_ZERO1          = 0,
-    THR_DC             = 1,
+  THR_NEAREST1 = 2,
+  THR_NEAR1 = 3,
 
-    THR_NEAREST1       = 2,
-    THR_NEAR1          = 3,
+  THR_ZERO2 = 4,
+  THR_NEAREST2 = 5,
 
-    THR_ZERO2          = 4,
-    THR_NEAREST2       = 5,
+  THR_ZERO3 = 6,
+  THR_NEAREST3 = 7,
 
-    THR_ZERO3          = 6,
-    THR_NEAREST3       = 7,
+  THR_NEAR2 = 8,
+  THR_NEAR3 = 9,
 
-    THR_NEAR2          = 8,
-    THR_NEAR3          = 9,
+  THR_V_PRED = 10,
+  THR_H_PRED = 11,
+  THR_TM = 12,
 
-    THR_V_PRED         = 10,
-    THR_H_PRED         = 11,
-    THR_TM             = 12,
+  THR_NEW1 = 13,
+  THR_NEW2 = 14,
+  THR_NEW3 = 15,
 
-    THR_NEW1           = 13,
-    THR_NEW2           = 14,
-    THR_NEW3           = 15,
+  THR_SPLIT1 = 16,
+  THR_SPLIT2 = 17,
+  THR_SPLIT3 = 18,
 
-    THR_SPLIT1         = 16,
-    THR_SPLIT2         = 17,
-    THR_SPLIT3         = 18,
+  THR_B_PRED = 19
+} THR_MODES;
 
-    THR_B_PRED         = 19
-}
-THR_MODES;
+typedef enum { DIAMOND = 0, NSTEP = 1, HEX = 2 } SEARCH_METHODS;
 
-typedef enum
-{
-    DIAMOND = 0,
-    NSTEP = 1,
-    HEX = 2
-} SEARCH_METHODS;
+typedef struct {
+  int RD;
+  SEARCH_METHODS search_method;
+  int improved_quant;
+  int improved_dct;
+  int auto_filter;
+  int recode_loop;
+  int iterative_sub_pixel;
+  int half_pixel_search;
+  int quarter_pixel_search;
+  int thresh_mult[MAX_MODES];
+  int max_step_search_steps;
+  int first_step;
+  int optimize_coefficients;
 
-typedef struct
-{
-    int RD;
-    SEARCH_METHODS search_method;
-    int improved_quant;
-    int improved_dct;
-    int auto_filter;
-    int recode_loop;
-    int iterative_sub_pixel;
-    int half_pixel_search;
-    int quarter_pixel_search;
-    int thresh_mult[MAX_MODES];
-    int max_step_search_steps;
-    int first_step;
-    int optimize_coefficients;
+  int use_fastquant_for_pick;
+  int no_skip_block4x4_search;
+  int improved_mv_pred;
 
-    int use_fastquant_for_pick;
-    int no_skip_block4x4_search;
-    int improved_mv_pred;
-
 } SPEED_FEATURES;
 
-typedef struct
-{
-    MACROBLOCK  mb;
-    int segment_counts[MAX_MB_SEGMENTS];
-    int totalrate;
+typedef struct {
+  MACROBLOCK mb;
+  int segment_counts[MAX_MB_SEGMENTS];
+  int totalrate;
 } MB_ROW_COMP;
 
-typedef struct
-{
-    TOKENEXTRA *start;
-    TOKENEXTRA *stop;
+typedef struct {
+  TOKENEXTRA *start;
+  TOKENEXTRA *stop;
 } TOKENLIST;
 
-typedef struct
-{
-    int ithread;
-    void *ptr1;
-    void *ptr2;
+typedef struct {
+  int ithread;
+  void *ptr1;
+  void *ptr2;
 } ENCODETHREAD_DATA;
-typedef struct
-{
-    int ithread;
-    void *ptr1;
+typedef struct {
+  int ithread;
+  void *ptr1;
 } LPFTHREAD_DATA;
 
-enum
-{
-    BLOCK_16X8,
-    BLOCK_8X16,
-    BLOCK_8X8,
-    BLOCK_4X4,
-    BLOCK_16X16,
-    BLOCK_MAX_SEGMENTS
+enum {
+  BLOCK_16X8,
+  BLOCK_8X16,
+  BLOCK_8X8,
+  BLOCK_4X4,
+  BLOCK_16X16,
+  BLOCK_MAX_SEGMENTS
 };
 
-typedef struct
-{
-    /* Layer configuration */
-    double framerate;
-    int target_bandwidth;
+typedef struct {
+  /* Layer configuration */
+  double framerate;
+  int target_bandwidth;
 
-    /* Layer specific coding parameters */
-    int64_t starting_buffer_level;
-    int64_t optimal_buffer_level;
-    int64_t maximum_buffer_size;
-    int64_t starting_buffer_level_in_ms;
-    int64_t optimal_buffer_level_in_ms;
-    int64_t maximum_buffer_size_in_ms;
+  /* Layer specific coding parameters */
+  int64_t starting_buffer_level;
+  int64_t optimal_buffer_level;
+  int64_t maximum_buffer_size;
+  int64_t starting_buffer_level_in_ms;
+  int64_t optimal_buffer_level_in_ms;
+  int64_t maximum_buffer_size_in_ms;
 
-    int avg_frame_size_for_layer;
+  int avg_frame_size_for_layer;
 
-    int64_t buffer_level;
-    int64_t bits_off_target;
+  int64_t buffer_level;
+  int64_t bits_off_target;
 
-    int64_t total_actual_bits;
-    int total_target_vs_actual;
+  int64_t total_actual_bits;
+  int total_target_vs_actual;
 
-    int worst_quality;
-    int active_worst_quality;
-    int best_quality;
-    int active_best_quality;
+  int worst_quality;
+  int active_worst_quality;
+  int best_quality;
+  int active_best_quality;
 
-    int ni_av_qi;
-    int ni_tot_qi;
-    int ni_frames;
-    int avg_frame_qindex;
+  int ni_av_qi;
+  int ni_tot_qi;
+  int ni_frames;
+  int avg_frame_qindex;
 
-    double rate_correction_factor;
-    double key_frame_rate_correction_factor;
-    double gf_rate_correction_factor;
+  double rate_correction_factor;
+  double key_frame_rate_correction_factor;
+  double gf_rate_correction_factor;
 
-    int zbin_over_quant;
+  int zbin_over_quant;
 
-    int inter_frame_target;
-    int64_t total_byte_count;
+  int inter_frame_target;
+  int64_t total_byte_count;
 
-    int filter_level;
+  int filter_level;
 
-    int last_frame_percent_intra;
+  int last_frame_percent_intra;
 
-    int count_mb_ref_frame_usage[MAX_REF_FRAMES];
+  int count_mb_ref_frame_usage[MAX_REF_FRAMES];
 
 } LAYER_CONTEXT;
 
-typedef struct VP8_COMP
-{
+typedef struct VP8_COMP {
+  DECLARE_ALIGNED(16, short, Y1quant[QINDEX_RANGE][16]);
+  DECLARE_ALIGNED(16, short, Y1quant_shift[QINDEX_RANGE][16]);
+  DECLARE_ALIGNED(16, short, Y1zbin[QINDEX_RANGE][16]);
+  DECLARE_ALIGNED(16, short, Y1round[QINDEX_RANGE][16]);
 
-    DECLARE_ALIGNED(16, short, Y1quant[QINDEX_RANGE][16]);
-    DECLARE_ALIGNED(16, short, Y1quant_shift[QINDEX_RANGE][16]);
-    DECLARE_ALIGNED(16, short, Y1zbin[QINDEX_RANGE][16]);
-    DECLARE_ALIGNED(16, short, Y1round[QINDEX_RANGE][16]);
+  DECLARE_ALIGNED(16, short, Y2quant[QINDEX_RANGE][16]);
+  DECLARE_ALIGNED(16, short, Y2quant_shift[QINDEX_RANGE][16]);
+  DECLARE_ALIGNED(16, short, Y2zbin[QINDEX_RANGE][16]);
+  DECLARE_ALIGNED(16, short, Y2round[QINDEX_RANGE][16]);
 
-    DECLARE_ALIGNED(16, short, Y2quant[QINDEX_RANGE][16]);
-    DECLARE_ALIGNED(16, short, Y2quant_shift[QINDEX_RANGE][16]);
-    DECLARE_ALIGNED(16, short, Y2zbin[QINDEX_RANGE][16]);
-    DECLARE_ALIGNED(16, short, Y2round[QINDEX_RANGE][16]);
+  DECLARE_ALIGNED(16, short, UVquant[QINDEX_RANGE][16]);
+  DECLARE_ALIGNED(16, short, UVquant_shift[QINDEX_RANGE][16]);
+  DECLARE_ALIGNED(16, short, UVzbin[QINDEX_RANGE][16]);
+  DECLARE_ALIGNED(16, short, UVround[QINDEX_RANGE][16]);
 
-    DECLARE_ALIGNED(16, short, UVquant[QINDEX_RANGE][16]);
-    DECLARE_ALIGNED(16, short, UVquant_shift[QINDEX_RANGE][16]);
-    DECLARE_ALIGNED(16, short, UVzbin[QINDEX_RANGE][16]);
-    DECLARE_ALIGNED(16, short, UVround[QINDEX_RANGE][16]);
+  DECLARE_ALIGNED(16, short, zrun_zbin_boost_y1[QINDEX_RANGE][16]);
+  DECLARE_ALIGNED(16, short, zrun_zbin_boost_y2[QINDEX_RANGE][16]);
+  DECLARE_ALIGNED(16, short, zrun_zbin_boost_uv[QINDEX_RANGE][16]);
+  DECLARE_ALIGNED(16, short, Y1quant_fast[QINDEX_RANGE][16]);
+  DECLARE_ALIGNED(16, short, Y2quant_fast[QINDEX_RANGE][16]);
+  DECLARE_ALIGNED(16, short, UVquant_fast[QINDEX_RANGE][16]);
 
-    DECLARE_ALIGNED(16, short, zrun_zbin_boost_y1[QINDEX_RANGE][16]);
-    DECLARE_ALIGNED(16, short, zrun_zbin_boost_y2[QINDEX_RANGE][16]);
-    DECLARE_ALIGNED(16, short, zrun_zbin_boost_uv[QINDEX_RANGE][16]);
-    DECLARE_ALIGNED(16, short, Y1quant_fast[QINDEX_RANGE][16]);
-    DECLARE_ALIGNED(16, short, Y2quant_fast[QINDEX_RANGE][16]);
-    DECLARE_ALIGNED(16, short, UVquant_fast[QINDEX_RANGE][16]);
+  MACROBLOCK mb;
+  VP8_COMMON common;
+  vp8_writer bc[9]; /* one boolcoder for each partition */
 
+  VP8_CONFIG oxcf;
 
-    MACROBLOCK mb;
-    VP8_COMMON common;
-    vp8_writer bc[9]; /* one boolcoder for each partition */
+  struct lookahead_ctx *lookahead;
+  struct lookahead_entry *source;
+  struct lookahead_entry *alt_ref_source;
+  struct lookahead_entry *last_source;
 
-    VP8_CONFIG oxcf;
+  YV12_BUFFER_CONFIG *Source;
+  YV12_BUFFER_CONFIG *un_scaled_source;
+  YV12_BUFFER_CONFIG scaled_source;
+  YV12_BUFFER_CONFIG *last_frame_unscaled_source;
 
-    struct lookahead_ctx    *lookahead;
-    struct lookahead_entry  *source;
-    struct lookahead_entry  *alt_ref_source;
-    struct lookahead_entry  *last_source;
+  unsigned int frames_till_alt_ref_frame;
+  /* frame in src_buffers has been identified to be encoded as an alt ref */
+  int source_alt_ref_pending;
+  /* an alt ref frame has been encoded and is usable */
+  int source_alt_ref_active;
+  /* source of frame to encode is an exact copy of an alt ref frame */
+  int is_src_frame_alt_ref;
 
-    YV12_BUFFER_CONFIG *Source;
-    YV12_BUFFER_CONFIG *un_scaled_source;
-    YV12_BUFFER_CONFIG scaled_source;
-    YV12_BUFFER_CONFIG *last_frame_unscaled_source;
+  /* golden frame same as last frame ( short circuit gold searches) */
+  int gold_is_last;
+  /* Alt reference frame same as last ( short circuit altref search) */
+  int alt_is_last;
+  /* don't do both alt and gold search ( just do gold). */
+  int gold_is_alt;
 
-    unsigned int frames_till_alt_ref_frame;
-    /* frame in src_buffers has been identified to be encoded as an alt ref */
-    int source_alt_ref_pending;
-    /* an alt ref frame has been encoded and is usable */
-    int source_alt_ref_active;
-    /* source of frame to encode is an exact copy of an alt ref frame */
-    int is_src_frame_alt_ref;
+  YV12_BUFFER_CONFIG pick_lf_lvl_frame;
 
-    /* golden frame same as last frame ( short circuit gold searches) */
-    int gold_is_last;
-    /* Alt reference frame same as last ( short circuit altref search) */
-    int alt_is_last;
-    /* don't do both alt and gold search ( just do gold). */
-    int gold_is_alt;
+  TOKENEXTRA *tok;
+  unsigned int tok_count;
 
-    YV12_BUFFER_CONFIG pick_lf_lvl_frame;
+  unsigned int frames_since_key;
+  unsigned int key_frame_frequency;
+  unsigned int this_key_frame_forced;
+  unsigned int next_key_frame_forced;
 
-    TOKENEXTRA *tok;
-    unsigned int tok_count;
+  /* Ambient reconstruction err target for force key frames */
+  int ambient_err;
 
+  unsigned int mode_check_freq[MAX_MODES];
 
-    unsigned int frames_since_key;
-    unsigned int key_frame_frequency;
-    unsigned int this_key_frame_forced;
-    unsigned int next_key_frame_forced;
+  int rd_baseline_thresh[MAX_MODES];
 
-    /* Ambient reconstruction err target for force key frames */
-    int ambient_err;
+  int RDMULT;
+  int RDDIV;
 
-    unsigned int mode_check_freq[MAX_MODES];
+  CODING_CONTEXT coding_context;
 
-    int rd_baseline_thresh[MAX_MODES];
+  /* Rate targetting variables */
+  int64_t last_prediction_error;
+  int64_t last_intra_error;
 
-    int RDMULT;
-    int RDDIV ;
+  int this_frame_target;
+  int projected_frame_size;
+  int last_q[2]; /* Separate values for Intra/Inter */
 
-    CODING_CONTEXT coding_context;
+  double rate_correction_factor;
+  double key_frame_rate_correction_factor;
+  double gf_rate_correction_factor;
 
-    /* Rate targetting variables */
-    int64_t last_prediction_error;
-    int64_t last_intra_error;
+  int frames_since_golden;
+  /* Count down till next GF */
+  int frames_till_gf_update_due;
 
-    int this_frame_target;
-    int projected_frame_size;
-    int last_q[2];                   /* Separate values for Intra/Inter */
+  /* GF interval chosen when we coded the last GF */
+  int current_gf_interval;
 
-    double rate_correction_factor;
-    double key_frame_rate_correction_factor;
-    double gf_rate_correction_factor;
+  /* Total bits overspent becasue of GF boost (cumulative) */
+  int gf_overspend_bits;
 
-    int frames_since_golden;
-    /* Count down till next GF */
-    int frames_till_gf_update_due;
+  /* Used in the few frames following a GF to recover the extra bits
+   * spent in that GF
+   */
+  int non_gf_bitrate_adjustment;
 
-    /* GF interval chosen when we coded the last GF */
-    int current_gf_interval;
+  /* Extra bits spent on key frames that need to be recovered */
+  int kf_overspend_bits;
 
-    /* Total bits overspent becasue of GF boost (cumulative) */
-    int gf_overspend_bits;
+  /* Current number of bit s to try and recover on each inter frame. */
+  int kf_bitrate_adjustment;
+  int max_gf_interval;
+  int baseline_gf_interval;
+  int active_arnr_frames;
 
-    /* Used in the few frames following a GF to recover the extra bits
-     * spent in that GF
-     */
-    int non_gf_bitrate_adjustment;
+  int64_t key_frame_count;
+  int prior_key_frame_distance[KEY_FRAME_CONTEXT];
+  /* Current section per frame bandwidth target */
+  int per_frame_bandwidth;
+  /* Average frame size target for clip */
+  int av_per_frame_bandwidth;
+  /* Minimum allocation that should be used for any frame */
+  int min_frame_bandwidth;
+  int inter_frame_target;
+  double output_framerate;
+  int64_t last_time_stamp_seen;
+  int64_t last_end_time_stamp_seen;
+  int64_t first_time_stamp_ever;
 
-    /* Extra bits spent on key frames that need to be recovered */
-    int kf_overspend_bits;
+  int ni_av_qi;
+  int ni_tot_qi;
+  int ni_frames;
+  int avg_frame_qindex;
 
-    /* Current number of bit s to try and recover on each inter frame. */
-    int kf_bitrate_adjustment;
-    int max_gf_interval;
-    int baseline_gf_interval;
-    int active_arnr_frames;
+  int64_t total_byte_count;
 
-    int64_t key_frame_count;
-    int prior_key_frame_distance[KEY_FRAME_CONTEXT];
-    /* Current section per frame bandwidth target */
-    int per_frame_bandwidth;
-    /* Average frame size target for clip */
-    int av_per_frame_bandwidth;
-    /* Minimum allocation that should be used for any frame */
-    int min_frame_bandwidth;
-    int inter_frame_target;
-    double output_framerate;
-    int64_t last_time_stamp_seen;
-    int64_t last_end_time_stamp_seen;
-    int64_t first_time_stamp_ever;
+  int buffered_mode;
 
-    int ni_av_qi;
-    int ni_tot_qi;
-    int ni_frames;
-    int avg_frame_qindex;
+  double framerate;
+  double ref_framerate;
+  int64_t buffer_level;
+  int64_t bits_off_target;
 
-    int64_t total_byte_count;
+  int rolling_target_bits;
+  int rolling_actual_bits;
 
-    int buffered_mode;
+  int long_rolling_target_bits;
+  int long_rolling_actual_bits;
 
-    double framerate;
-    double ref_framerate;
-    int64_t buffer_level;
-    int64_t bits_off_target;
+  int64_t total_actual_bits;
+  int total_target_vs_actual; /* debug stats */
 
-    int rolling_target_bits;
-    int rolling_actual_bits;
+  int worst_quality;
+  int active_worst_quality;
+  int best_quality;
+  int active_best_quality;
 
-    int long_rolling_target_bits;
-    int long_rolling_actual_bits;
+  int cq_target_quality;
 
-    int64_t total_actual_bits;
-    int total_target_vs_actual; /* debug stats */
+  int drop_frames_allowed; /* Are we permitted to drop frames? */
+  int drop_frame;          /* Drop this frame? */
 
-    int worst_quality;
-    int active_worst_quality;
-    int best_quality;
-    int active_best_quality;
+  vp8_prob frame_coef_probs[BLOCK_TYPES][COEF_BANDS][PREV_COEF_CONTEXTS]
+                           [ENTROPY_NODES];
+  char update_probs[BLOCK_TYPES][COEF_BANDS][PREV_COEF_CONTEXTS][ENTROPY_NODES];
 
-    int cq_target_quality;
+  unsigned int frame_branch_ct[BLOCK_TYPES][COEF_BANDS][PREV_COEF_CONTEXTS]
+                              [ENTROPY_NODES][2];
 
-    int drop_frames_allowed; /* Are we permitted to drop frames? */
-    int drop_frame;          /* Drop this frame? */
+  int gfu_boost;
+  int kf_boost;
+  int last_boost;
 
-    vp8_prob frame_coef_probs [BLOCK_TYPES] [COEF_BANDS] [PREV_COEF_CONTEXTS] [ENTROPY_NODES];
-    char update_probs [BLOCK_TYPES] [COEF_BANDS] [PREV_COEF_CONTEXTS] [ENTROPY_NODES];
+  int target_bandwidth;
+  struct vpx_codec_pkt_list *output_pkt_list;
 
-    unsigned int frame_branch_ct [BLOCK_TYPES] [COEF_BANDS] [PREV_COEF_CONTEXTS] [ENTROPY_NODES][2];
-
-    int gfu_boost;
-    int kf_boost;
-    int last_boost;
-
-    int target_bandwidth;
-    struct vpx_codec_pkt_list  *output_pkt_list;
-
 #if 0
     /* Experimental code for lagged and one pass */
     ONEPASS_FRAMESTATS one_pass_frame_stats[MAX_LAG_BUFFERS];
@@ -460,255 +436,250 @@
     int one_pass_frame_index;
 #endif
 
-    int decimation_factor;
-    int decimation_count;
+  int decimation_factor;
+  int decimation_count;
 
-    /* for real time encoding */
-    int avg_encode_time;     /* microsecond */
-    int avg_pick_mode_time;  /* microsecond */
-    int Speed;
-    int compressor_speed;
+  /* for real time encoding */
+  int avg_encode_time;    /* microsecond */
+  int avg_pick_mode_time; /* microsecond */
+  int Speed;
+  int compressor_speed;
 
-    int auto_gold;
-    int auto_adjust_gold_quantizer;
-    int auto_worst_q;
-    int cpu_used;
-    int pass;
+  int auto_gold;
+  int auto_adjust_gold_quantizer;
+  int auto_worst_q;
+  int cpu_used;
+  int pass;
 
+  int prob_intra_coded;
+  int prob_last_coded;
+  int prob_gf_coded;
+  int prob_skip_false;
+  int last_skip_false_probs[3];
+  int last_skip_probs_q[3];
+  int recent_ref_frame_usage[MAX_REF_FRAMES];
 
-    int prob_intra_coded;
-    int prob_last_coded;
-    int prob_gf_coded;
-    int prob_skip_false;
-    int last_skip_false_probs[3];
-    int last_skip_probs_q[3];
-    int recent_ref_frame_usage[MAX_REF_FRAMES];
+  int this_frame_percent_intra;
+  int last_frame_percent_intra;
 
-    int this_frame_percent_intra;
-    int last_frame_percent_intra;
+  int ref_frame_flags;
 
-    int ref_frame_flags;
+  SPEED_FEATURES sf;
 
-    SPEED_FEATURES sf;
+  /* Count ZEROMV on all reference frames. */
+  int zeromv_count;
+  int lf_zeromv_pct;
 
-    /* Count ZEROMV on all reference frames. */
-    int zeromv_count;
-    int lf_zeromv_pct;
+  unsigned char *segmentation_map;
+  signed char segment_feature_data[MB_LVL_MAX][MAX_MB_SEGMENTS];
+  int segment_encode_breakout[MAX_MB_SEGMENTS];
 
-    unsigned char *segmentation_map;
-    signed char segment_feature_data[MB_LVL_MAX][MAX_MB_SEGMENTS];
-    int  segment_encode_breakout[MAX_MB_SEGMENTS];
+  unsigned char *active_map;
+  unsigned int active_map_enabled;
 
-    unsigned char *active_map;
-    unsigned int active_map_enabled;
+  /* Video conferencing cyclic refresh mode flags. This is a mode
+   * designed to clean up the background over time in live encoding
+   * scenarious. It uses segmentation.
+   */
+  int cyclic_refresh_mode_enabled;
+  int cyclic_refresh_mode_max_mbs_perframe;
+  int cyclic_refresh_mode_index;
+  int cyclic_refresh_q;
+  signed char *cyclic_refresh_map;
+  // Count on how many (consecutive) times a macroblock uses ZER0MV_LAST.
+  unsigned char *consec_zero_last;
+  // Counter that is reset when a block is checked for a mode-bias against
+  // ZEROMV_LASTREF.
+  unsigned char *consec_zero_last_mvbias;
 
-    /* Video conferencing cyclic refresh mode flags. This is a mode
-     * designed to clean up the background over time in live encoding
-     * scenarious. It uses segmentation.
-     */
-    int cyclic_refresh_mode_enabled;
-    int cyclic_refresh_mode_max_mbs_perframe;
-    int cyclic_refresh_mode_index;
-    int cyclic_refresh_q;
-    signed char *cyclic_refresh_map;
-    // Count on how many (consecutive) times a macroblock uses ZER0MV_LAST.
-    unsigned char *consec_zero_last;
-    // Counter that is reset when a block is checked for a mode-bias against
-    // ZEROMV_LASTREF.
-    unsigned char *consec_zero_last_mvbias;
+  // Frame counter for the temporal pattern. Counter is rest when the temporal
+  // layers are changed dynamically (run-time change).
+  unsigned int temporal_pattern_counter;
+  // Temporal layer id.
+  int temporal_layer_id;
 
-    // Frame counter for the temporal pattern. Counter is rest when the temporal
-    // layers are changed dynamically (run-time change).
-    unsigned int temporal_pattern_counter;
-    // Temporal layer id.
-    int temporal_layer_id;
+  // Measure of average squared difference between source and denoised signal.
+  int mse_source_denoised;
 
-    // Measure of average squared difference between source and denoised signal.
-    int mse_source_denoised;
+  int force_maxqp;
 
-    int force_maxqp;
-
 #if CONFIG_MULTITHREAD
-    /* multithread data */
-    pthread_mutex_t *pmutex;
-    pthread_mutex_t mt_mutex;           /* mutex for b_multi_threaded */
-    int * mt_current_mb_col;
-    int mt_sync_range;
-    int b_multi_threaded;
-    int encoding_thread_count;
+  /* multithread data */
+  pthread_mutex_t *pmutex;
+  pthread_mutex_t mt_mutex; /* mutex for b_multi_threaded */
+  int *mt_current_mb_col;
+  int mt_sync_range;
+  int b_multi_threaded;
+  int encoding_thread_count;
 
-    pthread_t *h_encoding_thread;
-    pthread_t h_filter_thread;
+  pthread_t *h_encoding_thread;
+  pthread_t h_filter_thread;
 
-    MB_ROW_COMP *mb_row_ei;
-    ENCODETHREAD_DATA *en_thread_data;
-    LPFTHREAD_DATA lpf_thread_data;
+  MB_ROW_COMP *mb_row_ei;
+  ENCODETHREAD_DATA *en_thread_data;
+  LPFTHREAD_DATA lpf_thread_data;
 
-    /* events */
-    sem_t *h_event_start_encoding;
-    sem_t h_event_end_encoding;
-    sem_t h_event_start_lpf;
-    sem_t h_event_end_lpf;
+  /* events */
+  sem_t *h_event_start_encoding;
+  sem_t h_event_end_encoding;
+  sem_t h_event_start_lpf;
+  sem_t h_event_end_lpf;
 #endif
 
-    TOKENLIST *tplist;
-    unsigned int partition_sz[MAX_PARTITIONS];
-    unsigned char *partition_d[MAX_PARTITIONS];
-    unsigned char *partition_d_end[MAX_PARTITIONS];
+  TOKENLIST *tplist;
+  unsigned int partition_sz[MAX_PARTITIONS];
+  unsigned char *partition_d[MAX_PARTITIONS];
+  unsigned char *partition_d_end[MAX_PARTITIONS];
 
+  fractional_mv_step_fp *find_fractional_mv_step;
+  vp8_full_search_fn_t full_search_sad;
+  vp8_refining_search_fn_t refining_search_sad;
+  vp8_diamond_search_fn_t diamond_search_sad;
+  vp8_variance_fn_ptr_t fn_ptr[BLOCK_MAX_SEGMENTS];
+  uint64_t time_receive_data;
+  uint64_t time_compress_data;
+  uint64_t time_pick_lpf;
+  uint64_t time_encode_mb_row;
 
-    fractional_mv_step_fp *find_fractional_mv_step;
-    vp8_full_search_fn_t full_search_sad;
-    vp8_refining_search_fn_t refining_search_sad;
-    vp8_diamond_search_fn_t diamond_search_sad;
-    vp8_variance_fn_ptr_t fn_ptr[BLOCK_MAX_SEGMENTS];
-    uint64_t time_receive_data;
-    uint64_t time_compress_data;
-    uint64_t time_pick_lpf;
-    uint64_t time_encode_mb_row;
+  int base_skip_false_prob[128];
 
-    int base_skip_false_prob[128];
+  FRAME_CONTEXT lfc_n; /* last frame entropy */
+  FRAME_CONTEXT lfc_a; /* last alt ref entropy */
+  FRAME_CONTEXT lfc_g; /* last gold ref entropy */
 
-    FRAME_CONTEXT lfc_n; /* last frame entropy */
-    FRAME_CONTEXT lfc_a; /* last alt ref entropy */
-    FRAME_CONTEXT lfc_g; /* last gold ref entropy */
+  struct twopass_rc {
+    unsigned int section_intra_rating;
+    double section_max_qfactor;
+    unsigned int next_iiratio;
+    unsigned int this_iiratio;
+    FIRSTPASS_STATS total_stats;
+    FIRSTPASS_STATS this_frame_stats;
+    FIRSTPASS_STATS *stats_in, *stats_in_end, *stats_in_start;
+    FIRSTPASS_STATS total_left_stats;
+    int first_pass_done;
+    int64_t bits_left;
+    int64_t clip_bits_total;
+    double avg_iiratio;
+    double modified_error_total;
+    double modified_error_used;
+    double modified_error_left;
+    double kf_intra_err_min;
+    double gf_intra_err_min;
+    int frames_to_key;
+    int maxq_max_limit;
+    int maxq_min_limit;
+    int gf_decay_rate;
+    int static_scene_max_gf_interval;
+    int kf_bits;
+    /* Remaining error from uncoded frames in a gf group. */
+    int gf_group_error_left;
+    /* Projected total bits available for a key frame group of frames */
+    int64_t kf_group_bits;
+    /* Error score of frames still to be coded in kf group */
+    int64_t kf_group_error_left;
+    /* Projected Bits available for a group including 1 GF or ARF */
+    int64_t gf_group_bits;
+    /* Bits for the golden frame or ARF */
+    int gf_bits;
+    int alt_extra_bits;
+    double est_max_qcorrection_factor;
+  } twopass;
 
-
-    struct twopass_rc
-    {
-        unsigned int section_intra_rating;
-        double section_max_qfactor;
-        unsigned int next_iiratio;
-        unsigned int this_iiratio;
-        FIRSTPASS_STATS total_stats;
-        FIRSTPASS_STATS this_frame_stats;
-        FIRSTPASS_STATS *stats_in, *stats_in_end, *stats_in_start;
-        FIRSTPASS_STATS total_left_stats;
-        int first_pass_done;
-        int64_t bits_left;
-        int64_t clip_bits_total;
-        double avg_iiratio;
-        double modified_error_total;
-        double modified_error_used;
-        double modified_error_left;
-        double kf_intra_err_min;
-        double gf_intra_err_min;
-        int frames_to_key;
-        int maxq_max_limit;
-        int maxq_min_limit;
-        int gf_decay_rate;
-        int static_scene_max_gf_interval;
-        int kf_bits;
-        /* Remaining error from uncoded frames in a gf group. */
-        int gf_group_error_left;
-        /* Projected total bits available for a key frame group of frames */
-        int64_t kf_group_bits;
-        /* Error score of frames still to be coded in kf group */
-        int64_t kf_group_error_left;
-        /* Projected Bits available for a group including 1 GF or ARF */
-        int64_t gf_group_bits;
-        /* Bits for the golden frame or ARF */
-        int gf_bits;
-        int alt_extra_bits;
-        double est_max_qcorrection_factor;
-    } twopass;
-
 #if VP8_TEMPORAL_ALT_REF
-    YV12_BUFFER_CONFIG alt_ref_buffer;
-    YV12_BUFFER_CONFIG *frames[MAX_LAG_BUFFERS];
-    int fixed_divide[512];
+  YV12_BUFFER_CONFIG alt_ref_buffer;
+  YV12_BUFFER_CONFIG *frames[MAX_LAG_BUFFERS];
+  int fixed_divide[512];
 #endif
 
 #if CONFIG_INTERNAL_STATS
-    int    count;
-    double total_y;
-    double total_u;
-    double total_v;
-    double total ;
-    double total_sq_error;
-    double totalp_y;
-    double totalp_u;
-    double totalp_v;
-    double totalp;
-    double total_sq_error2;
-    int    bytes;
-    double summed_quality;
-    double summed_weights;
-    unsigned int tot_recode_hits;
+  int count;
+  double total_y;
+  double total_u;
+  double total_v;
+  double total;
+  double total_sq_error;
+  double totalp_y;
+  double totalp_u;
+  double totalp_v;
+  double totalp;
+  double total_sq_error2;
+  int bytes;
+  double summed_quality;
+  double summed_weights;
+  unsigned int tot_recode_hits;
 
-    int b_calculate_ssimg;
+  int b_calculate_ssimg;
 #endif
-    int b_calculate_psnr;
+  int b_calculate_psnr;
 
-    /* Per MB activity measurement */
-    unsigned int activity_avg;
-    unsigned int * mb_activity_map;
+  /* Per MB activity measurement */
+  unsigned int activity_avg;
+  unsigned int *mb_activity_map;
 
-    /* Record of which MBs still refer to last golden frame either
-     * directly or through 0,0
-     */
-    unsigned char *gf_active_flags;
-    int gf_active_count;
+  /* Record of which MBs still refer to last golden frame either
+   * directly or through 0,0
+   */
+  unsigned char *gf_active_flags;
+  int gf_active_count;
 
-    int output_partition;
+  int output_partition;
 
-    /* Store last frame's MV info for next frame MV prediction */
-    int_mv *lfmv;
-    int *lf_ref_frame_sign_bias;
-    int *lf_ref_frame;
+  /* Store last frame's MV info for next frame MV prediction */
+  int_mv *lfmv;
+  int *lf_ref_frame_sign_bias;
+  int *lf_ref_frame;
 
-    /* force next frame to intra when kf_auto says so */
-    int force_next_frame_intra;
+  /* force next frame to intra when kf_auto says so */
+  int force_next_frame_intra;
 
-    int droppable;
+  int droppable;
 
-    int initial_width;
-    int initial_height;
+  int initial_width;
+  int initial_height;
 
 #if CONFIG_TEMPORAL_DENOISING
-    VP8_DENOISER denoiser;
+  VP8_DENOISER denoiser;
 #endif
 
-    /* Coding layer state variables */
-    unsigned int current_layer;
-    LAYER_CONTEXT layer_context[VPX_TS_MAX_LAYERS];
+  /* Coding layer state variables */
+  unsigned int current_layer;
+  LAYER_CONTEXT layer_context[VPX_TS_MAX_LAYERS];
 
-    int64_t frames_in_layer[VPX_TS_MAX_LAYERS];
-    int64_t bytes_in_layer[VPX_TS_MAX_LAYERS];
-    double sum_psnr[VPX_TS_MAX_LAYERS];
-    double sum_psnr_p[VPX_TS_MAX_LAYERS];
-    double total_error2[VPX_TS_MAX_LAYERS];
-    double total_error2_p[VPX_TS_MAX_LAYERS];
-    double sum_ssim[VPX_TS_MAX_LAYERS];
-    double sum_weights[VPX_TS_MAX_LAYERS];
+  int64_t frames_in_layer[VPX_TS_MAX_LAYERS];
+  int64_t bytes_in_layer[VPX_TS_MAX_LAYERS];
+  double sum_psnr[VPX_TS_MAX_LAYERS];
+  double sum_psnr_p[VPX_TS_MAX_LAYERS];
+  double total_error2[VPX_TS_MAX_LAYERS];
+  double total_error2_p[VPX_TS_MAX_LAYERS];
+  double sum_ssim[VPX_TS_MAX_LAYERS];
+  double sum_weights[VPX_TS_MAX_LAYERS];
 
-    double total_ssimg_y_in_layer[VPX_TS_MAX_LAYERS];
-    double total_ssimg_u_in_layer[VPX_TS_MAX_LAYERS];
-    double total_ssimg_v_in_layer[VPX_TS_MAX_LAYERS];
-    double total_ssimg_all_in_layer[VPX_TS_MAX_LAYERS];
+  double total_ssimg_y_in_layer[VPX_TS_MAX_LAYERS];
+  double total_ssimg_u_in_layer[VPX_TS_MAX_LAYERS];
+  double total_ssimg_v_in_layer[VPX_TS_MAX_LAYERS];
+  double total_ssimg_all_in_layer[VPX_TS_MAX_LAYERS];
 
 #if CONFIG_MULTI_RES_ENCODING
-    /* Number of MBs per row at lower-resolution level */
-    int    mr_low_res_mb_cols;
-    /* Indicate if lower-res mv info is available */
-    unsigned char  mr_low_res_mv_avail;
+  /* Number of MBs per row at lower-resolution level */
+  int mr_low_res_mb_cols;
+  /* Indicate if lower-res mv info is available */
+  unsigned char mr_low_res_mv_avail;
 #endif
-    /* The frame number of each reference frames */
-    unsigned int current_ref_frames[MAX_REF_FRAMES];
-    // Closest reference frame to current frame.
-    MV_REFERENCE_FRAME closest_reference_frame;
+  /* The frame number of each reference frames */
+  unsigned int current_ref_frames[MAX_REF_FRAMES];
+  // Closest reference frame to current frame.
+  MV_REFERENCE_FRAME closest_reference_frame;
 
-    struct rd_costs_struct
-    {
-        int mvcosts[2][MVvals+1];
-        int mvsadcosts[2][MVfpvals+1];
-        int mbmode_cost[2][MB_MODE_COUNT];
-        int intra_uv_mode_cost[2][MB_MODE_COUNT];
-        int bmode_costs[10][10][10];
-        int inter_bmode_costs[B_MODE_COUNT];
-        int token_costs[BLOCK_TYPES][COEF_BANDS]
-        [PREV_COEF_CONTEXTS][MAX_ENTROPY_TOKENS];
-    } rd_costs;
+  struct rd_costs_struct {
+    int mvcosts[2][MVvals + 1];
+    int mvsadcosts[2][MVfpvals + 1];
+    int mbmode_cost[2][MB_MODE_COUNT];
+    int intra_uv_mode_cost[2][MB_MODE_COUNT];
+    int bmode_costs[10][10][10];
+    int inter_bmode_costs[B_MODE_COUNT];
+    int token_costs[BLOCK_TYPES][COEF_BANDS][PREV_COEF_CONTEXTS]
+                   [MAX_ENTROPY_TOKENS];
+  } rd_costs;
 } VP8_COMP;
 
 void vp8_initialize_enc(void);
@@ -726,20 +697,22 @@
 void vp8_set_speed_features(VP8_COMP *cpi);
 
 #if CONFIG_DEBUG
-#define CHECK_MEM_ERROR(lval,expr) do {\
-        lval = (expr); \
-        if(!lval) \
-            vpx_internal_error(&cpi->common.error, VPX_CODEC_MEM_ERROR,\
-                               "Failed to allocate "#lval" at %s:%d", \
-                               __FILE__,__LINE__);\
-    } while(0)
+#define CHECK_MEM_ERROR(lval, expr)                                         \
+  do {                                                                      \
+    lval = (expr);                                                          \
+    if (!lval)                                                              \
+      vpx_internal_error(&cpi->common.error, VPX_CODEC_MEM_ERROR,           \
+                         "Failed to allocate " #lval " at %s:%d", __FILE__, \
+                         __LINE__);                                         \
+  } while (0)
 #else
-#define CHECK_MEM_ERROR(lval,expr) do {\
-        lval = (expr); \
-        if(!lval) \
-            vpx_internal_error(&cpi->common.error, VPX_CODEC_MEM_ERROR,\
-                               "Failed to allocate "#lval);\
-    } while(0)
+#define CHECK_MEM_ERROR(lval, expr)                               \
+  do {                                                            \
+    lval = (expr);                                                \
+    if (!lval)                                                    \
+      vpx_internal_error(&cpi->common.error, VPX_CODEC_MEM_ERROR, \
+                         "Failed to allocate " #lval);            \
+  } while (0)
 #endif
 #ifdef __cplusplus
 }  // extern "C"
--- a/vp8/encoder/pickinter.c
+++ b/vp8/encoder/pickinter.c
@@ -8,7 +8,6 @@
  *  be found in the AUTHORS file in the root of the source tree.
  */
 
-
 #include <limits.h>
 #include "vpx_config.h"
 #include "./vpx_dsp_rtcd.h"
@@ -47,11 +46,14 @@
 // skin color classifier is defined.
 
 // Fixed-point skin color model parameters.
-static const int skin_mean[5][2] =
-    {{7463, 9614}, {6400, 10240}, {7040, 10240}, {8320, 9280}, {6800, 9614}};
-static const int skin_inv_cov[4] = {4107, 1663, 1663, 2157};  // q16
-static const int skin_threshold[6] = {1570636, 1400000, 800000, 800000, 800000,
-    800000};  // q18
+static const int skin_mean[5][2] = { { 7463, 9614 },
+                                     { 6400, 10240 },
+                                     { 7040, 10240 },
+                                     { 8320, 9280 },
+                                     { 6800, 9614 } };
+static const int skin_inv_cov[4] = { 4107, 1663, 1663, 2157 };  // q16
+static const int skin_threshold[6] = { 1570636, 1400000, 800000,
+                                       800000,  800000,  800000 };  // q18
 
 // Evaluates the Mahalanobis distance measure for the input CbCr values.
 static int evaluate_skin_color_difference(int cb, int cr, int idx) {
@@ -66,62 +68,51 @@
   const int cb_diff_q2 = (cb_diff_q12 + (1 << 9)) >> 10;
   const int cbcr_diff_q2 = (cbcr_diff_q12 + (1 << 9)) >> 10;
   const int cr_diff_q2 = (cr_diff_q12 + (1 << 9)) >> 10;
-  const int skin_diff = skin_inv_cov[0] * cb_diff_q2 +
-      skin_inv_cov[1] * cbcr_diff_q2 +
-      skin_inv_cov[2] * cbcr_diff_q2 +
-      skin_inv_cov[3] * cr_diff_q2;
+  const int skin_diff =
+      skin_inv_cov[0] * cb_diff_q2 + skin_inv_cov[1] * cbcr_diff_q2 +
+      skin_inv_cov[2] * cbcr_diff_q2 + skin_inv_cov[3] * cr_diff_q2;
   return skin_diff;
 }
 
 // Checks if the input yCbCr values corresponds to skin color.
-static int is_skin_color(int y, int cb, int cr, int consec_zeromv)
-{
-  if (y < 40 || y > 220)
-  {
+static int is_skin_color(int y, int cb, int cr, int consec_zeromv) {
+  if (y < 40 || y > 220) {
     return 0;
-  }
-  else
-  {
-    if (MODEL_MODE == 0)
-    {
+  } else {
+    if (MODEL_MODE == 0) {
       return (evaluate_skin_color_difference(cb, cr, 0) < skin_threshold[0]);
-    }
-    else
-    {
+    } else {
       int i = 0;
       // No skin if block has been zero motion for long consecutive time.
-      if (consec_zeromv > 60)
-        return 0;
+      if (consec_zeromv > 60) return 0;
       // Exit on grey.
-       if (cb == 128 && cr == 128)
-         return 0;
-       // Exit on very strong cb.
-       if (cb > 150 && cr < 110)
-         return 0;
-       for (; i < 5; i++) {
-         int skin_color_diff = evaluate_skin_color_difference(cb, cr, i);
-         if (skin_color_diff < skin_threshold[i + 1]) {
-            if (y < 60 && skin_color_diff > 3 * (skin_threshold[i + 1] >> 2))
-              return 0;
-            else if (consec_zeromv > 25 &&
-                     skin_color_diff > (skin_threshold[i + 1] >> 1))
-              return 0;
-            else
-             return 1;
-         }
-         // Exit if difference is much large than the threshold.
-         if (skin_color_diff > (skin_threshold[i + 1] << 3)) {
-           return 0;
-         }
-       }
+      if (cb == 128 && cr == 128) return 0;
+      // Exit on very strong cb.
+      if (cb > 150 && cr < 110) return 0;
+      for (; i < 5; i++) {
+        int skin_color_diff = evaluate_skin_color_difference(cb, cr, i);
+        if (skin_color_diff < skin_threshold[i + 1]) {
+          if (y < 60 && skin_color_diff > 3 * (skin_threshold[i + 1] >> 2))
+            return 0;
+          else if (consec_zeromv > 25 &&
+                   skin_color_diff > (skin_threshold[i + 1] >> 1))
+            return 0;
+          else
+            return 1;
+        }
+        // Exit if difference is much large than the threshold.
+        if (skin_color_diff > (skin_threshold[i + 1] << 3)) {
+          return 0;
+        }
+      }
       return 0;
     }
   }
 }
 
-static int macroblock_corner_grad(unsigned char* signal, int stride,
-                                  int offsetx, int offsety, int sgnx, int sgny)
-{
+static int macroblock_corner_grad(unsigned char *signal, int stride,
+                                  int offsetx, int offsety, int sgnx,
+                                  int sgny) {
   int y1 = signal[offsetx * stride + offsety];
   int y2 = signal[offsetx * stride + offsety + sgny];
   int y3 = signal[(offsetx + sgnx) * stride + offsety];
@@ -129,15 +120,10 @@
   return VPXMAX(VPXMAX(abs(y1 - y2), abs(y1 - y3)), abs(y1 - y4));
 }
 
-static int check_dot_artifact_candidate(VP8_COMP *cpi,
-                                        MACROBLOCK *x,
-                                        unsigned char *target_last,
-                                        int stride,
-                                        unsigned char* last_ref,
-                                        int mb_row,
-                                        int mb_col,
-                                        int channel)
-{
+static int check_dot_artifact_candidate(VP8_COMP *cpi, MACROBLOCK *x,
+                                        unsigned char *target_last, int stride,
+                                        unsigned char *last_ref, int mb_row,
+                                        int mb_col, int channel) {
   int threshold1 = 6;
   int threshold2 = 3;
   unsigned int max_num = (cpi->common.MBs) / 10;
@@ -150,8 +136,7 @@
   if (channel > 0) {
     shift = 7;
   }
-  if (cpi->oxcf.number_of_layers > 1)
-  {
+  if (cpi->oxcf.number_of_layers > 1) {
     num_frames = 20;
   }
   x->zero_last_dot_suppress = 0;
@@ -163,8 +148,7 @@
   if (cpi->current_layer == 0 &&
       cpi->consec_zero_last_mvbias[index] > num_frames &&
       x->mbs_zero_last_dot_suppress < max_num &&
-      !cpi->oxcf.screen_content_mode)
-  {
+      !cpi->oxcf.screen_content_mode) {
     // If this block is checked here, label it so we don't check it again until
     // ~|x| framaes later.
     x->zero_last_dot_suppress = 1;
@@ -175,16 +159,14 @@
     // Top-left:
     grad_last = macroblock_corner_grad(last_ref, stride, 0, 0, 1, 1);
     grad_source = macroblock_corner_grad(target_last, stride, 0, 0, 1, 1);
-    if (grad_last >= threshold1 && grad_source <= threshold2)
-    {
-       x->mbs_zero_last_dot_suppress++;
-       return 1;
+    if (grad_last >= threshold1 && grad_source <= threshold2) {
+      x->mbs_zero_last_dot_suppress++;
+      return 1;
     }
     // Top-right:
     grad_last = macroblock_corner_grad(last_ref, stride, 0, shift, 1, -1);
     grad_source = macroblock_corner_grad(target_last, stride, 0, shift, 1, -1);
-    if (grad_last >= threshold1 && grad_source <= threshold2)
-    {
+    if (grad_last >= threshold1 && grad_source <= threshold2) {
       x->mbs_zero_last_dot_suppress++;
       return 1;
     }
@@ -191,16 +173,15 @@
     // Bottom-left:
     grad_last = macroblock_corner_grad(last_ref, stride, shift, 0, -1, 1);
     grad_source = macroblock_corner_grad(target_last, stride, shift, 0, -1, 1);
-    if (grad_last >= threshold1 && grad_source <= threshold2)
-    {
+    if (grad_last >= threshold1 && grad_source <= threshold2) {
       x->mbs_zero_last_dot_suppress++;
       return 1;
     }
     // Bottom-right:
     grad_last = macroblock_corner_grad(last_ref, stride, shift, shift, -1, -1);
-    grad_source = macroblock_corner_grad(target_last, stride, shift, shift, -1, -1);
-    if (grad_last >= threshold1 && grad_source <= threshold2)
-    {
+    grad_source =
+        macroblock_corner_grad(target_last, stride, shift, shift, -1, -1);
+    if (grad_last >= threshold1 && grad_source <= threshold2) {
       x->mbs_zero_last_dot_suppress++;
       return 1;
     }
@@ -214,1394 +195,1217 @@
                                 int error_per_bit,
                                 const vp8_variance_fn_ptr_t *vfp,
                                 int *mvcost[2], int *distortion,
-                                unsigned int *sse)
-{
-    (void) b;
-    (void) d;
-    (void) ref_mv;
-    (void) error_per_bit;
-    (void) vfp;
-    (void) mb;
-    (void) mvcost;
-    (void) distortion;
-    (void) sse;
-    bestmv->as_mv.row <<= 3;
-    bestmv->as_mv.col <<= 3;
-    return 0;
+                                unsigned int *sse) {
+  (void)b;
+  (void)d;
+  (void)ref_mv;
+  (void)error_per_bit;
+  (void)vfp;
+  (void)mb;
+  (void)mvcost;
+  (void)distortion;
+  (void)sse;
+  bestmv->as_mv.row <<= 3;
+  bestmv->as_mv.col <<= 3;
+  return 0;
 }
 
+int vp8_get_inter_mbpred_error(MACROBLOCK *mb, const vp8_variance_fn_ptr_t *vfp,
+                               unsigned int *sse, int_mv this_mv) {
+  BLOCK *b = &mb->block[0];
+  BLOCKD *d = &mb->e_mbd.block[0];
+  unsigned char *what = (*(b->base_src) + b->src);
+  int what_stride = b->src_stride;
+  int pre_stride = mb->e_mbd.pre.y_stride;
+  unsigned char *in_what = mb->e_mbd.pre.y_buffer + d->offset;
+  int in_what_stride = pre_stride;
+  int xoffset = this_mv.as_mv.col & 7;
+  int yoffset = this_mv.as_mv.row & 7;
 
-int vp8_get_inter_mbpred_error(MACROBLOCK *mb,
-                                  const vp8_variance_fn_ptr_t *vfp,
-                                  unsigned int *sse,
-                                  int_mv this_mv)
-{
+  in_what += (this_mv.as_mv.row >> 3) * pre_stride + (this_mv.as_mv.col >> 3);
 
-    BLOCK *b = &mb->block[0];
-    BLOCKD *d = &mb->e_mbd.block[0];
-    unsigned char *what = (*(b->base_src) + b->src);
-    int what_stride = b->src_stride;
-    int pre_stride = mb->e_mbd.pre.y_stride;
-    unsigned char *in_what = mb->e_mbd.pre.y_buffer + d->offset ;
-    int in_what_stride = pre_stride;
-    int xoffset = this_mv.as_mv.col & 7;
-    int yoffset = this_mv.as_mv.row & 7;
-
-    in_what += (this_mv.as_mv.row >> 3) * pre_stride + (this_mv.as_mv.col >> 3);
-
-    if (xoffset | yoffset)
-    {
-        return vfp->svf(in_what, in_what_stride, xoffset, yoffset, what, what_stride, sse);
-    }
-    else
-    {
-        return vfp->vf(what, what_stride, in_what, in_what_stride, sse);
-    }
-
+  if (xoffset | yoffset) {
+    return vfp->svf(in_what, in_what_stride, xoffset, yoffset, what,
+                    what_stride, sse);
+  } else {
+    return vfp->vf(what, what_stride, in_what, in_what_stride, sse);
+  }
 }
 
-static int get_prediction_error(BLOCK *be, BLOCKD *b)
-{
-    unsigned char *sptr;
-    unsigned char *dptr;
-    sptr = (*(be->base_src) + be->src);
-    dptr = b->predictor;
+static int get_prediction_error(BLOCK *be, BLOCKD *b) {
+  unsigned char *sptr;
+  unsigned char *dptr;
+  sptr = (*(be->base_src) + be->src);
+  dptr = b->predictor;
 
-    return vpx_get4x4sse_cs(sptr, be->src_stride, dptr, 16);
-
+  return vpx_get4x4sse_cs(sptr, be->src_stride, dptr, 16);
 }
 
-static int pick_intra4x4block(
-    MACROBLOCK *x,
-    int ib,
-    B_PREDICTION_MODE *best_mode,
-    const int *mode_costs,
+static int pick_intra4x4block(MACROBLOCK *x, int ib,
+                              B_PREDICTION_MODE *best_mode,
+                              const int *mode_costs,
 
-    int *bestrate,
-    int *bestdistortion)
-{
+                              int *bestrate, int *bestdistortion) {
+  BLOCKD *b = &x->e_mbd.block[ib];
+  BLOCK *be = &x->block[ib];
+  int dst_stride = x->e_mbd.dst.y_stride;
+  unsigned char *dst = x->e_mbd.dst.y_buffer + b->offset;
+  B_PREDICTION_MODE mode;
+  int best_rd = INT_MAX;
+  int rate;
+  int distortion;
 
-    BLOCKD *b = &x->e_mbd.block[ib];
-    BLOCK *be = &x->block[ib];
-    int dst_stride = x->e_mbd.dst.y_stride;
-    unsigned char *dst = x->e_mbd.dst.y_buffer + b->offset;
-    B_PREDICTION_MODE mode;
-    int best_rd = INT_MAX;
-    int rate;
-    int distortion;
+  unsigned char *Above = dst - dst_stride;
+  unsigned char *yleft = dst - 1;
+  unsigned char top_left = Above[-1];
 
-    unsigned char *Above = dst - dst_stride;
-    unsigned char *yleft = dst - 1;
-    unsigned char top_left = Above[-1];
+  for (mode = B_DC_PRED; mode <= B_HE_PRED; mode++) {
+    int this_rd;
 
-    for (mode = B_DC_PRED; mode <= B_HE_PRED; mode++)
-    {
-        int this_rd;
+    rate = mode_costs[mode];
 
-        rate = mode_costs[mode];
+    vp8_intra4x4_predict(Above, yleft, dst_stride, mode, b->predictor, 16,
+                         top_left);
+    distortion = get_prediction_error(be, b);
+    this_rd = RDCOST(x->rdmult, x->rddiv, rate, distortion);
 
-        vp8_intra4x4_predict(Above, yleft, dst_stride, mode,
-                             b->predictor, 16, top_left);
-        distortion = get_prediction_error(be, b);
-        this_rd = RDCOST(x->rdmult, x->rddiv, rate, distortion);
-
-        if (this_rd < best_rd)
-        {
-            *bestrate = rate;
-            *bestdistortion = distortion;
-            best_rd = this_rd;
-            *best_mode = mode;
-        }
+    if (this_rd < best_rd) {
+      *bestrate = rate;
+      *bestdistortion = distortion;
+      best_rd = this_rd;
+      *best_mode = mode;
     }
+  }
 
-    b->bmi.as_mode = *best_mode;
-    vp8_encode_intra4x4block(x, ib);
-    return best_rd;
+  b->bmi.as_mode = *best_mode;
+  vp8_encode_intra4x4block(x, ib);
+  return best_rd;
 }
 
+static int pick_intra4x4mby_modes(MACROBLOCK *mb, int *Rate, int *best_dist) {
+  MACROBLOCKD *const xd = &mb->e_mbd;
+  int i;
+  int cost = mb->mbmode_cost[xd->frame_type][B_PRED];
+  int error;
+  int distortion = 0;
+  const int *bmode_costs;
 
-static int pick_intra4x4mby_modes
-(
-    MACROBLOCK *mb,
-    int *Rate,
-    int *best_dist
-)
-{
-    MACROBLOCKD *const xd = &mb->e_mbd;
-    int i;
-    int cost = mb->mbmode_cost [xd->frame_type] [B_PRED];
-    int error;
-    int distortion = 0;
-    const int *bmode_costs;
+  intra_prediction_down_copy(xd, xd->dst.y_buffer - xd->dst.y_stride + 16);
 
-    intra_prediction_down_copy(xd, xd->dst.y_buffer - xd->dst.y_stride + 16);
+  bmode_costs = mb->inter_bmode_costs;
 
-    bmode_costs = mb->inter_bmode_costs;
+  for (i = 0; i < 16; i++) {
+    MODE_INFO *const mic = xd->mode_info_context;
+    const int mis = xd->mode_info_stride;
 
-    for (i = 0; i < 16; i++)
-    {
-        MODE_INFO *const mic = xd->mode_info_context;
-        const int mis = xd->mode_info_stride;
+    B_PREDICTION_MODE UNINITIALIZED_IS_SAFE(best_mode);
+    int UNINITIALIZED_IS_SAFE(r), UNINITIALIZED_IS_SAFE(d);
 
-        B_PREDICTION_MODE UNINITIALIZED_IS_SAFE(best_mode);
-        int UNINITIALIZED_IS_SAFE(r), UNINITIALIZED_IS_SAFE(d);
+    if (mb->e_mbd.frame_type == KEY_FRAME) {
+      const B_PREDICTION_MODE A = above_block_mode(mic, i, mis);
+      const B_PREDICTION_MODE L = left_block_mode(mic, i);
 
-        if (mb->e_mbd.frame_type == KEY_FRAME)
-        {
-            const B_PREDICTION_MODE A = above_block_mode(mic, i, mis);
-            const B_PREDICTION_MODE L = left_block_mode(mic, i);
+      bmode_costs = mb->bmode_costs[A][L];
+    }
 
-            bmode_costs  = mb->bmode_costs[A][L];
-        }
+    pick_intra4x4block(mb, i, &best_mode, bmode_costs, &r, &d);
 
+    cost += r;
+    distortion += d;
+    mic->bmi[i].as_mode = best_mode;
 
-        pick_intra4x4block(mb, i, &best_mode, bmode_costs, &r, &d);
+    /* Break out case where we have already exceeded best so far value
+     * that was passed in
+     */
+    if (distortion > *best_dist) break;
+  }
 
-        cost += r;
-        distortion += d;
-        mic->bmi[i].as_mode = best_mode;
+  *Rate = cost;
 
-        /* Break out case where we have already exceeded best so far value
-         * that was passed in
-         */
-        if (distortion > *best_dist)
-            break;
-    }
+  if (i == 16) {
+    *best_dist = distortion;
+    error = RDCOST(mb->rdmult, mb->rddiv, cost, distortion);
+  } else {
+    *best_dist = INT_MAX;
+    error = INT_MAX;
+  }
 
-    *Rate = cost;
-
-    if (i == 16)
-    {
-        *best_dist = distortion;
-        error = RDCOST(mb->rdmult, mb->rddiv, cost, distortion);
-    }
-    else
-    {
-        *best_dist = INT_MAX;
-        error = INT_MAX;
-    }
-
-    return error;
+  return error;
 }
 
-static void pick_intra_mbuv_mode(MACROBLOCK *mb)
-{
+static void pick_intra_mbuv_mode(MACROBLOCK *mb) {
+  MACROBLOCKD *x = &mb->e_mbd;
+  unsigned char *uabove_row = x->dst.u_buffer - x->dst.uv_stride;
+  unsigned char *vabove_row = x->dst.v_buffer - x->dst.uv_stride;
+  unsigned char *usrc_ptr = (mb->block[16].src + *mb->block[16].base_src);
+  unsigned char *vsrc_ptr = (mb->block[20].src + *mb->block[20].base_src);
+  int uvsrc_stride = mb->block[16].src_stride;
+  unsigned char uleft_col[8];
+  unsigned char vleft_col[8];
+  unsigned char utop_left = uabove_row[-1];
+  unsigned char vtop_left = vabove_row[-1];
+  int i, j;
+  int expected_udc;
+  int expected_vdc;
+  int shift;
+  int Uaverage = 0;
+  int Vaverage = 0;
+  int diff;
+  int pred_error[4] = { 0, 0, 0, 0 }, best_error = INT_MAX;
+  MB_PREDICTION_MODE UNINITIALIZED_IS_SAFE(best_mode);
 
-    MACROBLOCKD *x = &mb->e_mbd;
-    unsigned char *uabove_row = x->dst.u_buffer - x->dst.uv_stride;
-    unsigned char *vabove_row = x->dst.v_buffer - x->dst.uv_stride;
-    unsigned char *usrc_ptr = (mb->block[16].src + *mb->block[16].base_src);
-    unsigned char *vsrc_ptr = (mb->block[20].src + *mb->block[20].base_src);
-    int uvsrc_stride = mb->block[16].src_stride;
-    unsigned char uleft_col[8];
-    unsigned char vleft_col[8];
-    unsigned char utop_left = uabove_row[-1];
-    unsigned char vtop_left = vabove_row[-1];
-    int i, j;
-    int expected_udc;
-    int expected_vdc;
-    int shift;
-    int Uaverage = 0;
-    int Vaverage = 0;
-    int diff;
-    int pred_error[4] = {0, 0, 0, 0}, best_error = INT_MAX;
-    MB_PREDICTION_MODE UNINITIALIZED_IS_SAFE(best_mode);
+  for (i = 0; i < 8; i++) {
+    uleft_col[i] = x->dst.u_buffer[i * x->dst.uv_stride - 1];
+    vleft_col[i] = x->dst.v_buffer[i * x->dst.uv_stride - 1];
+  }
 
+  if (!x->up_available && !x->left_available) {
+    expected_udc = 128;
+    expected_vdc = 128;
+  } else {
+    shift = 2;
 
-    for (i = 0; i < 8; i++)
-    {
-        uleft_col[i] = x->dst.u_buffer [i* x->dst.uv_stride -1];
-        vleft_col[i] = x->dst.v_buffer [i* x->dst.uv_stride -1];
-    }
+    if (x->up_available) {
+      for (i = 0; i < 8; i++) {
+        Uaverage += uabove_row[i];
+        Vaverage += vabove_row[i];
+      }
 
-    if (!x->up_available && !x->left_available)
-    {
-        expected_udc = 128;
-        expected_vdc = 128;
+      shift++;
     }
-    else
-    {
-        shift = 2;
 
-        if (x->up_available)
-        {
+    if (x->left_available) {
+      for (i = 0; i < 8; i++) {
+        Uaverage += uleft_col[i];
+        Vaverage += vleft_col[i];
+      }
 
-            for (i = 0; i < 8; i++)
-            {
-                Uaverage += uabove_row[i];
-                Vaverage += vabove_row[i];
-            }
-
-            shift ++;
-
-        }
-
-        if (x->left_available)
-        {
-            for (i = 0; i < 8; i++)
-            {
-                Uaverage += uleft_col[i];
-                Vaverage += vleft_col[i];
-            }
-
-            shift ++;
-
-        }
-
-        expected_udc = (Uaverage + (1 << (shift - 1))) >> shift;
-        expected_vdc = (Vaverage + (1 << (shift - 1))) >> shift;
+      shift++;
     }
 
+    expected_udc = (Uaverage + (1 << (shift - 1))) >> shift;
+    expected_vdc = (Vaverage + (1 << (shift - 1))) >> shift;
+  }
 
-    for (i = 0; i < 8; i++)
-    {
-        for (j = 0; j < 8; j++)
-        {
+  for (i = 0; i < 8; i++) {
+    for (j = 0; j < 8; j++) {
+      int predu = uleft_col[i] + uabove_row[j] - utop_left;
+      int predv = vleft_col[i] + vabove_row[j] - vtop_left;
+      int u_p, v_p;
 
-            int predu = uleft_col[i] + uabove_row[j] - utop_left;
-            int predv = vleft_col[i] + vabove_row[j] - vtop_left;
-            int u_p, v_p;
+      u_p = usrc_ptr[j];
+      v_p = vsrc_ptr[j];
 
-            u_p = usrc_ptr[j];
-            v_p = vsrc_ptr[j];
+      if (predu < 0) predu = 0;
 
-            if (predu < 0)
-                predu = 0;
+      if (predu > 255) predu = 255;
 
-            if (predu > 255)
-                predu = 255;
+      if (predv < 0) predv = 0;
 
-            if (predv < 0)
-                predv = 0;
+      if (predv > 255) predv = 255;
 
-            if (predv > 255)
-                predv = 255;
+      diff = u_p - expected_udc;
+      pred_error[DC_PRED] += diff * diff;
+      diff = v_p - expected_vdc;
+      pred_error[DC_PRED] += diff * diff;
 
+      diff = u_p - uabove_row[j];
+      pred_error[V_PRED] += diff * diff;
+      diff = v_p - vabove_row[j];
+      pred_error[V_PRED] += diff * diff;
 
-            diff = u_p - expected_udc;
-            pred_error[DC_PRED] += diff * diff;
-            diff = v_p - expected_vdc;
-            pred_error[DC_PRED] += diff * diff;
+      diff = u_p - uleft_col[i];
+      pred_error[H_PRED] += diff * diff;
+      diff = v_p - vleft_col[i];
+      pred_error[H_PRED] += diff * diff;
 
+      diff = u_p - predu;
+      pred_error[TM_PRED] += diff * diff;
+      diff = v_p - predv;
+      pred_error[TM_PRED] += diff * diff;
+    }
 
-            diff = u_p - uabove_row[j];
-            pred_error[V_PRED] += diff * diff;
-            diff = v_p - vabove_row[j];
-            pred_error[V_PRED] += diff * diff;
+    usrc_ptr += uvsrc_stride;
+    vsrc_ptr += uvsrc_stride;
 
-
-            diff = u_p - uleft_col[i];
-            pred_error[H_PRED] += diff * diff;
-            diff = v_p - vleft_col[i];
-            pred_error[H_PRED] += diff * diff;
-
-
-            diff = u_p - predu;
-            pred_error[TM_PRED] += diff * diff;
-            diff = v_p - predv;
-            pred_error[TM_PRED] += diff * diff;
-
-
-        }
-
-        usrc_ptr += uvsrc_stride;
-        vsrc_ptr += uvsrc_stride;
-
-        if (i == 3)
-        {
-            usrc_ptr = (mb->block[18].src + *mb->block[18].base_src);
-            vsrc_ptr = (mb->block[22].src + *mb->block[22].base_src);
-        }
-
-
-
+    if (i == 3) {
+      usrc_ptr = (mb->block[18].src + *mb->block[18].base_src);
+      vsrc_ptr = (mb->block[22].src + *mb->block[22].base_src);
     }
+  }
 
-
-    for (i = DC_PRED; i <= TM_PRED; i++)
-    {
-        if (best_error > pred_error[i])
-        {
-            best_error = pred_error[i];
-            best_mode = (MB_PREDICTION_MODE)i;
-        }
+  for (i = DC_PRED; i <= TM_PRED; i++) {
+    if (best_error > pred_error[i]) {
+      best_error = pred_error[i];
+      best_mode = (MB_PREDICTION_MODE)i;
     }
+  }
 
-
-    mb->e_mbd.mode_info_context->mbmi.uv_mode = best_mode;
-
+  mb->e_mbd.mode_info_context->mbmi.uv_mode = best_mode;
 }
 
-static void update_mvcount(MACROBLOCK *x, int_mv *best_ref_mv)
-{
-    MACROBLOCKD *xd = &x->e_mbd;
-    /* Split MV modes currently not supported when RD is nopt enabled,
-     * therefore, only need to modify MVcount in NEWMV mode. */
-    if (xd->mode_info_context->mbmi.mode == NEWMV)
-    {
-        x->MVcount[0][mv_max+((xd->mode_info_context->mbmi.mv.as_mv.row -
-                                      best_ref_mv->as_mv.row) >> 1)]++;
-        x->MVcount[1][mv_max+((xd->mode_info_context->mbmi.mv.as_mv.col -
-                                      best_ref_mv->as_mv.col) >> 1)]++;
-    }
+static void update_mvcount(MACROBLOCK *x, int_mv *best_ref_mv) {
+  MACROBLOCKD *xd = &x->e_mbd;
+  /* Split MV modes currently not supported when RD is nopt enabled,
+   * therefore, only need to modify MVcount in NEWMV mode. */
+  if (xd->mode_info_context->mbmi.mode == NEWMV) {
+    x->MVcount[0][mv_max + ((xd->mode_info_context->mbmi.mv.as_mv.row -
+                             best_ref_mv->as_mv.row) >>
+                            1)]++;
+    x->MVcount[1][mv_max + ((xd->mode_info_context->mbmi.mv.as_mv.col -
+                             best_ref_mv->as_mv.col) >>
+                            1)]++;
+  }
 }
 
-
 #if CONFIG_MULTI_RES_ENCODING
-static
-void get_lower_res_motion_info(VP8_COMP *cpi, MACROBLOCKD *xd, int *dissim,
-                               int *parent_ref_frame,
-                               MB_PREDICTION_MODE *parent_mode,
-                               int_mv *parent_ref_mv, int mb_row, int mb_col)
-{
-    LOWER_RES_MB_INFO* store_mode_info
-                          = ((LOWER_RES_FRAME_INFO*)cpi->oxcf.mr_low_res_mode_info)->mb_info;
-    unsigned int parent_mb_index;
+static void get_lower_res_motion_info(VP8_COMP *cpi, MACROBLOCKD *xd,
+                                      int *dissim, int *parent_ref_frame,
+                                      MB_PREDICTION_MODE *parent_mode,
+                                      int_mv *parent_ref_mv, int mb_row,
+                                      int mb_col) {
+  LOWER_RES_MB_INFO *store_mode_info =
+      ((LOWER_RES_FRAME_INFO *)cpi->oxcf.mr_low_res_mode_info)->mb_info;
+  unsigned int parent_mb_index;
 
-    /* Consider different down_sampling_factor.  */
-    {
-        /* TODO: Removed the loop that supports special down_sampling_factor
-         * such as 2, 4, 8. Will revisit it if needed.
-         * Should also try using a look-up table to see if it helps
-         * performance. */
-        int parent_mb_row, parent_mb_col;
+  /* Consider different down_sampling_factor.  */
+  {
+    /* TODO: Removed the loop that supports special down_sampling_factor
+     * such as 2, 4, 8. Will revisit it if needed.
+     * Should also try using a look-up table to see if it helps
+     * performance. */
+    int parent_mb_row, parent_mb_col;
 
-        parent_mb_row = mb_row*cpi->oxcf.mr_down_sampling_factor.den
-                    /cpi->oxcf.mr_down_sampling_factor.num;
-        parent_mb_col = mb_col*cpi->oxcf.mr_down_sampling_factor.den
-                    /cpi->oxcf.mr_down_sampling_factor.num;
-        parent_mb_index = parent_mb_row*cpi->mr_low_res_mb_cols + parent_mb_col;
-    }
+    parent_mb_row = mb_row * cpi->oxcf.mr_down_sampling_factor.den /
+                    cpi->oxcf.mr_down_sampling_factor.num;
+    parent_mb_col = mb_col * cpi->oxcf.mr_down_sampling_factor.den /
+                    cpi->oxcf.mr_down_sampling_factor.num;
+    parent_mb_index = parent_mb_row * cpi->mr_low_res_mb_cols + parent_mb_col;
+  }
 
-    /* Read lower-resolution mode & motion result from memory.*/
-    *parent_ref_frame = store_mode_info[parent_mb_index].ref_frame;
-    *parent_mode =  store_mode_info[parent_mb_index].mode;
-    *dissim = store_mode_info[parent_mb_index].dissim;
+  /* Read lower-resolution mode & motion result from memory.*/
+  *parent_ref_frame = store_mode_info[parent_mb_index].ref_frame;
+  *parent_mode = store_mode_info[parent_mb_index].mode;
+  *dissim = store_mode_info[parent_mb_index].dissim;
 
-    /* For highest-resolution encoder, adjust dissim value. Lower its quality
-     * for good performance. */
-    if (cpi->oxcf.mr_encoder_id == (cpi->oxcf.mr_total_resolutions - 1))
-        *dissim>>=1;
+  /* For highest-resolution encoder, adjust dissim value. Lower its quality
+   * for good performance. */
+  if (cpi->oxcf.mr_encoder_id == (cpi->oxcf.mr_total_resolutions - 1))
+    *dissim >>= 1;
 
-    if(*parent_ref_frame != INTRA_FRAME)
-    {
-        /* Consider different down_sampling_factor.
-         * The result can be rounded to be more precise, but it takes more time.
-         */
-        (*parent_ref_mv).as_mv.row = store_mode_info[parent_mb_index].mv.as_mv.row
-                                  *cpi->oxcf.mr_down_sampling_factor.num
-                                  /cpi->oxcf.mr_down_sampling_factor.den;
-        (*parent_ref_mv).as_mv.col = store_mode_info[parent_mb_index].mv.as_mv.col
-                                  *cpi->oxcf.mr_down_sampling_factor.num
-                                  /cpi->oxcf.mr_down_sampling_factor.den;
+  if (*parent_ref_frame != INTRA_FRAME) {
+    /* Consider different down_sampling_factor.
+     * The result can be rounded to be more precise, but it takes more time.
+     */
+    (*parent_ref_mv).as_mv.row = store_mode_info[parent_mb_index].mv.as_mv.row *
+                                 cpi->oxcf.mr_down_sampling_factor.num /
+                                 cpi->oxcf.mr_down_sampling_factor.den;
+    (*parent_ref_mv).as_mv.col = store_mode_info[parent_mb_index].mv.as_mv.col *
+                                 cpi->oxcf.mr_down_sampling_factor.num /
+                                 cpi->oxcf.mr_down_sampling_factor.den;
 
-        vp8_clamp_mv2(parent_ref_mv, xd);
-    }
+    vp8_clamp_mv2(parent_ref_mv, xd);
+  }
 }
 #endif
 
-static void check_for_encode_breakout(unsigned int sse, MACROBLOCK* x)
-{
-    MACROBLOCKD *xd = &x->e_mbd;
+static void check_for_encode_breakout(unsigned int sse, MACROBLOCK *x) {
+  MACROBLOCKD *xd = &x->e_mbd;
 
-    unsigned int threshold = (xd->block[0].dequant[1]
-        * xd->block[0].dequant[1] >>4);
+  unsigned int threshold =
+      (xd->block[0].dequant[1] * xd->block[0].dequant[1] >> 4);
 
-    if(threshold < x->encode_breakout)
-        threshold = x->encode_breakout;
+  if (threshold < x->encode_breakout) threshold = x->encode_breakout;
 
-    if (sse < threshold )
-    {
-        /* Check u and v to make sure skip is ok */
-        unsigned int sse2 = 0;
+  if (sse < threshold) {
+    /* Check u and v to make sure skip is ok */
+    unsigned int sse2 = 0;
 
-        sse2 = VP8_UVSSE(x);
+    sse2 = VP8_UVSSE(x);
 
-        if (sse2 * 2 < x->encode_breakout)
-            x->skip = 1;
-        else
-            x->skip = 0;
-    }
+    if (sse2 * 2 < x->encode_breakout)
+      x->skip = 1;
+    else
+      x->skip = 0;
+  }
 }
 
-static int evaluate_inter_mode(unsigned int* sse, int rate2, int* distortion2,
-                               VP8_COMP *cpi, MACROBLOCK *x, int rd_adj)
-{
-    MB_PREDICTION_MODE this_mode = x->e_mbd.mode_info_context->mbmi.mode;
-    int_mv mv = x->e_mbd.mode_info_context->mbmi.mv;
-    int this_rd;
-    int denoise_aggressive = 0;
-    /* Exit early and don't compute the distortion if this macroblock
-     * is marked inactive. */
-    if (cpi->active_map_enabled && x->active_ptr[0] == 0)
-    {
-        *sse = 0;
-        *distortion2 = 0;
-        x->skip = 1;
-        return INT_MAX;
-    }
+static int evaluate_inter_mode(unsigned int *sse, int rate2, int *distortion2,
+                               VP8_COMP *cpi, MACROBLOCK *x, int rd_adj) {
+  MB_PREDICTION_MODE this_mode = x->e_mbd.mode_info_context->mbmi.mode;
+  int_mv mv = x->e_mbd.mode_info_context->mbmi.mv;
+  int this_rd;
+  int denoise_aggressive = 0;
+  /* Exit early and don't compute the distortion if this macroblock
+   * is marked inactive. */
+  if (cpi->active_map_enabled && x->active_ptr[0] == 0) {
+    *sse = 0;
+    *distortion2 = 0;
+    x->skip = 1;
+    return INT_MAX;
+  }
 
-    if((this_mode != NEWMV) ||
-        !(cpi->sf.half_pixel_search) || cpi->common.full_pixel==1)
-        *distortion2 = vp8_get_inter_mbpred_error(x,
-                                              &cpi->fn_ptr[BLOCK_16X16],
-                                              sse, mv);
+  if ((this_mode != NEWMV) || !(cpi->sf.half_pixel_search) ||
+      cpi->common.full_pixel == 1)
+    *distortion2 =
+        vp8_get_inter_mbpred_error(x, &cpi->fn_ptr[BLOCK_16X16], sse, mv);
 
-    this_rd = RDCOST(x->rdmult, x->rddiv, rate2, *distortion2);
+  this_rd = RDCOST(x->rdmult, x->rddiv, rate2, *distortion2);
 
 #if CONFIG_TEMPORAL_DENOISING
-    if (cpi->oxcf.noise_sensitivity > 0) {
-      denoise_aggressive =
+  if (cpi->oxcf.noise_sensitivity > 0) {
+    denoise_aggressive =
         (cpi->denoiser.denoiser_mode == kDenoiserOnYUVAggressive) ? 1 : 0;
-    }
+  }
 #endif
 
-    // Adjust rd for ZEROMV and LAST, if LAST is the closest reference frame.
-    // TODO: We should also add condition on distance of closest to current.
-    if(!cpi->oxcf.screen_content_mode &&
-       this_mode == ZEROMV &&
-       x->e_mbd.mode_info_context->mbmi.ref_frame == LAST_FRAME &&
-       (denoise_aggressive || (cpi->closest_reference_frame == LAST_FRAME)))
-    {
-        // No adjustment if block is considered to be skin area.
-        if(x->is_skin)
-            rd_adj = 100;
+  // Adjust rd for ZEROMV and LAST, if LAST is the closest reference frame.
+  // TODO: We should also add condition on distance of closest to current.
+  if (!cpi->oxcf.screen_content_mode && this_mode == ZEROMV &&
+      x->e_mbd.mode_info_context->mbmi.ref_frame == LAST_FRAME &&
+      (denoise_aggressive || (cpi->closest_reference_frame == LAST_FRAME))) {
+    // No adjustment if block is considered to be skin area.
+    if (x->is_skin) rd_adj = 100;
 
-        this_rd = ((int64_t)this_rd) * rd_adj / 100;
-    }
+    this_rd = ((int64_t)this_rd) * rd_adj / 100;
+  }
 
-    check_for_encode_breakout(*sse, x);
-    return this_rd;
+  check_for_encode_breakout(*sse, x);
+  return this_rd;
 }
 
 static void calculate_zeromv_rd_adjustment(VP8_COMP *cpi, MACROBLOCK *x,
-                                    int *rd_adjustment)
-{
-    MODE_INFO *mic = x->e_mbd.mode_info_context;
-    int_mv mv_l, mv_a, mv_al;
-    int local_motion_check = 0;
+                                           int *rd_adjustment) {
+  MODE_INFO *mic = x->e_mbd.mode_info_context;
+  int_mv mv_l, mv_a, mv_al;
+  int local_motion_check = 0;
 
-    if (cpi->lf_zeromv_pct > 40)
-    {
-        /* left mb */
-        mic -= 1;
-        mv_l = mic->mbmi.mv;
+  if (cpi->lf_zeromv_pct > 40) {
+    /* left mb */
+    mic -= 1;
+    mv_l = mic->mbmi.mv;
 
-        if (mic->mbmi.ref_frame != INTRA_FRAME)
-            if( abs(mv_l.as_mv.row) < 8 && abs(mv_l.as_mv.col) < 8)
-                local_motion_check++;
+    if (mic->mbmi.ref_frame != INTRA_FRAME)
+      if (abs(mv_l.as_mv.row) < 8 && abs(mv_l.as_mv.col) < 8)
+        local_motion_check++;
 
-        /* above-left mb */
-        mic -= x->e_mbd.mode_info_stride;
-        mv_al = mic->mbmi.mv;
+    /* above-left mb */
+    mic -= x->e_mbd.mode_info_stride;
+    mv_al = mic->mbmi.mv;
 
-        if (mic->mbmi.ref_frame != INTRA_FRAME)
-            if( abs(mv_al.as_mv.row) < 8 && abs(mv_al.as_mv.col) < 8)
-                local_motion_check++;
+    if (mic->mbmi.ref_frame != INTRA_FRAME)
+      if (abs(mv_al.as_mv.row) < 8 && abs(mv_al.as_mv.col) < 8)
+        local_motion_check++;
 
-        /* above mb */
-        mic += 1;
-        mv_a = mic->mbmi.mv;
+    /* above mb */
+    mic += 1;
+    mv_a = mic->mbmi.mv;
 
-        if (mic->mbmi.ref_frame != INTRA_FRAME)
-            if( abs(mv_a.as_mv.row) < 8 && abs(mv_a.as_mv.col) < 8)
-                local_motion_check++;
+    if (mic->mbmi.ref_frame != INTRA_FRAME)
+      if (abs(mv_a.as_mv.row) < 8 && abs(mv_a.as_mv.col) < 8)
+        local_motion_check++;
 
-        if (((!x->e_mbd.mb_to_top_edge || !x->e_mbd.mb_to_left_edge)
-            && local_motion_check >0) ||  local_motion_check >2 )
-            *rd_adjustment = 80;
-        else if (local_motion_check > 0)
-            *rd_adjustment = 90;
-    }
+    if (((!x->e_mbd.mb_to_top_edge || !x->e_mbd.mb_to_left_edge) &&
+         local_motion_check > 0) ||
+        local_motion_check > 2)
+      *rd_adjustment = 80;
+    else if (local_motion_check > 0)
+      *rd_adjustment = 90;
+  }
 }
 
 void vp8_pick_inter_mode(VP8_COMP *cpi, MACROBLOCK *x, int recon_yoffset,
                          int recon_uvoffset, int *returnrate,
                          int *returndistortion, int *returnintra, int mb_row,
-                         int mb_col)
-{
-    BLOCK *b = &x->block[0];
-    BLOCKD *d = &x->e_mbd.block[0];
-    MACROBLOCKD *xd = &x->e_mbd;
-    MB_MODE_INFO best_mbmode;
+                         int mb_col) {
+  BLOCK *b = &x->block[0];
+  BLOCKD *d = &x->e_mbd.block[0];
+  MACROBLOCKD *xd = &x->e_mbd;
+  MB_MODE_INFO best_mbmode;
 
-    int_mv best_ref_mv_sb[2];
-    int_mv mode_mv_sb[2][MB_MODE_COUNT];
-    int_mv best_ref_mv;
-    int_mv *mode_mv;
-    MB_PREDICTION_MODE this_mode;
-    int num00;
-    int mdcounts[4];
-    int best_rd = INT_MAX;
-    int rd_adjustment = 100;
-    int best_intra_rd = INT_MAX;
-    int mode_index;
-    int rate;
-    int rate2;
-    int distortion2;
-    int bestsme = INT_MAX;
-    int best_mode_index = 0;
-    unsigned int sse = UINT_MAX, best_rd_sse = UINT_MAX;
+  int_mv best_ref_mv_sb[2];
+  int_mv mode_mv_sb[2][MB_MODE_COUNT];
+  int_mv best_ref_mv;
+  int_mv *mode_mv;
+  MB_PREDICTION_MODE this_mode;
+  int num00;
+  int mdcounts[4];
+  int best_rd = INT_MAX;
+  int rd_adjustment = 100;
+  int best_intra_rd = INT_MAX;
+  int mode_index;
+  int rate;
+  int rate2;
+  int distortion2;
+  int bestsme = INT_MAX;
+  int best_mode_index = 0;
+  unsigned int sse = UINT_MAX, best_rd_sse = UINT_MAX;
 #if CONFIG_TEMPORAL_DENOISING
-    unsigned int zero_mv_sse = UINT_MAX, best_sse = UINT_MAX;
+  unsigned int zero_mv_sse = UINT_MAX, best_sse = UINT_MAX;
 #endif
 
-    int sf_improved_mv_pred = cpi->sf.improved_mv_pred;
+  int sf_improved_mv_pred = cpi->sf.improved_mv_pred;
 
 #if CONFIG_MULTI_RES_ENCODING
-    int dissim = INT_MAX;
-    int parent_ref_frame = 0;
-    int_mv parent_ref_mv;
-    MB_PREDICTION_MODE parent_mode = 0;
-    int parent_ref_valid = 0;
+  int dissim = INT_MAX;
+  int parent_ref_frame = 0;
+  int_mv parent_ref_mv;
+  MB_PREDICTION_MODE parent_mode = 0;
+  int parent_ref_valid = 0;
 #endif
 
-    int_mv mvp;
+  int_mv mvp;
 
-    int near_sadidx[8] = {0, 1, 2, 3, 4, 5, 6, 7};
-    int saddone=0;
-    /* search range got from mv_pred(). It uses step_param levels. (0-7) */
-    int sr=0;
+  int near_sadidx[8] = { 0, 1, 2, 3, 4, 5, 6, 7 };
+  int saddone = 0;
+  /* search range got from mv_pred(). It uses step_param levels. (0-7) */
+  int sr = 0;
 
-    unsigned char *plane[4][3];
-    int ref_frame_map[4];
-    int sign_bias = 0;
-    int dot_artifact_candidate = 0;
-    get_predictor_pointers(cpi, plane, recon_yoffset, recon_uvoffset);
+  unsigned char *plane[4][3];
+  int ref_frame_map[4];
+  int sign_bias = 0;
+  int dot_artifact_candidate = 0;
+  get_predictor_pointers(cpi, plane, recon_yoffset, recon_uvoffset);
 
-    // If the current frame is using LAST as a reference, check for
-    // biasing the mode selection for dot artifacts.
-    if (cpi->ref_frame_flags & VP8_LAST_FRAME) {
-      unsigned char* target_y = x->src.y_buffer;
-      unsigned char* target_u = x->block[16].src + *x->block[16].base_src;
-      unsigned char* target_v = x->block[20].src + *x->block[20].base_src;
-      int stride = x->src.y_stride;
-      int stride_uv = x->block[16].src_stride;
+  // If the current frame is using LAST as a reference, check for
+  // biasing the mode selection for dot artifacts.
+  if (cpi->ref_frame_flags & VP8_LAST_FRAME) {
+    unsigned char *target_y = x->src.y_buffer;
+    unsigned char *target_u = x->block[16].src + *x->block[16].base_src;
+    unsigned char *target_v = x->block[20].src + *x->block[20].base_src;
+    int stride = x->src.y_stride;
+    int stride_uv = x->block[16].src_stride;
 #if CONFIG_TEMPORAL_DENOISING
-      if (cpi->oxcf.noise_sensitivity) {
-        const int uv_denoise = (cpi->oxcf.noise_sensitivity >= 2) ? 1 : 0;
-        target_y =
-            cpi->denoiser.yv12_running_avg[LAST_FRAME].y_buffer + recon_yoffset;
-        stride = cpi->denoiser.yv12_running_avg[LAST_FRAME].y_stride;
-        if (uv_denoise) {
-          target_u =
-              cpi->denoiser.yv12_running_avg[LAST_FRAME].u_buffer +
-                  recon_uvoffset;
-          target_v =
-              cpi->denoiser.yv12_running_avg[LAST_FRAME].v_buffer +
-                  recon_uvoffset;
-          stride_uv = cpi->denoiser.yv12_running_avg[LAST_FRAME].uv_stride;
-        }
+    if (cpi->oxcf.noise_sensitivity) {
+      const int uv_denoise = (cpi->oxcf.noise_sensitivity >= 2) ? 1 : 0;
+      target_y =
+          cpi->denoiser.yv12_running_avg[LAST_FRAME].y_buffer + recon_yoffset;
+      stride = cpi->denoiser.yv12_running_avg[LAST_FRAME].y_stride;
+      if (uv_denoise) {
+        target_u = cpi->denoiser.yv12_running_avg[LAST_FRAME].u_buffer +
+                   recon_uvoffset;
+        target_v = cpi->denoiser.yv12_running_avg[LAST_FRAME].v_buffer +
+                   recon_uvoffset;
+        stride_uv = cpi->denoiser.yv12_running_avg[LAST_FRAME].uv_stride;
       }
+    }
 #endif
-      dot_artifact_candidate =
-          check_dot_artifact_candidate(cpi, x, target_y, stride,
-              plane[LAST_FRAME][0], mb_row, mb_col, 0);
-      // If not found in Y channel, check UV channel.
+    dot_artifact_candidate = check_dot_artifact_candidate(
+        cpi, x, target_y, stride, plane[LAST_FRAME][0], mb_row, mb_col, 0);
+    // If not found in Y channel, check UV channel.
+    if (!dot_artifact_candidate) {
+      dot_artifact_candidate = check_dot_artifact_candidate(
+          cpi, x, target_u, stride_uv, plane[LAST_FRAME][1], mb_row, mb_col, 1);
       if (!dot_artifact_candidate) {
-        dot_artifact_candidate =
-            check_dot_artifact_candidate(cpi, x, target_u, stride_uv,
-                plane[LAST_FRAME][1], mb_row, mb_col, 1);
-        if (!dot_artifact_candidate) {
-          dot_artifact_candidate =
-              check_dot_artifact_candidate(cpi, x, target_v, stride_uv,
-                  plane[LAST_FRAME][2], mb_row, mb_col, 2);
-        }
+        dot_artifact_candidate = check_dot_artifact_candidate(
+            cpi, x, target_v, stride_uv, plane[LAST_FRAME][2], mb_row, mb_col,
+            2);
       }
     }
+  }
 
 #if CONFIG_MULTI_RES_ENCODING
-    // |parent_ref_valid| will be set here if potentially we can do mv resue for
-    // this higher resol (|cpi->oxcf.mr_encoder_id| > 0) frame.
-    // |parent_ref_valid| may be reset depending on |parent_ref_frame| for
-    // the current macroblock below.
-    parent_ref_valid = cpi->oxcf.mr_encoder_id && cpi->mr_low_res_mv_avail;
-    if (parent_ref_valid)
-    {
-        int parent_ref_flag;
+  // |parent_ref_valid| will be set here if potentially we can do mv resue for
+  // this higher resol (|cpi->oxcf.mr_encoder_id| > 0) frame.
+  // |parent_ref_valid| may be reset depending on |parent_ref_frame| for
+  // the current macroblock below.
+  parent_ref_valid = cpi->oxcf.mr_encoder_id && cpi->mr_low_res_mv_avail;
+  if (parent_ref_valid) {
+    int parent_ref_flag;
 
-        get_lower_res_motion_info(cpi, xd, &dissim, &parent_ref_frame,
-                                  &parent_mode, &parent_ref_mv, mb_row, mb_col);
+    get_lower_res_motion_info(cpi, xd, &dissim, &parent_ref_frame, &parent_mode,
+                              &parent_ref_mv, mb_row, mb_col);
 
-        /* TODO(jkoleszar): The references available (ref_frame_flags) to the
-         * lower res encoder should match those available to this encoder, but
-         * there seems to be a situation where this mismatch can happen in the
-         * case of frame dropping and temporal layers. For example,
-         * GOLD being disallowed in ref_frame_flags, but being returned as
-         * parent_ref_frame.
-         *
-         * In this event, take the conservative approach of disabling the
-         * lower res info for this MB.
-         */
+    /* TODO(jkoleszar): The references available (ref_frame_flags) to the
+     * lower res encoder should match those available to this encoder, but
+     * there seems to be a situation where this mismatch can happen in the
+     * case of frame dropping and temporal layers. For example,
+     * GOLD being disallowed in ref_frame_flags, but being returned as
+     * parent_ref_frame.
+     *
+     * In this event, take the conservative approach of disabling the
+     * lower res info for this MB.
+     */
 
-        parent_ref_flag = 0;
-        // Note availability for mv reuse is only based on last and golden.
-        if (parent_ref_frame == LAST_FRAME)
-            parent_ref_flag = (cpi->ref_frame_flags & VP8_LAST_FRAME);
-        else if (parent_ref_frame == GOLDEN_FRAME)
-            parent_ref_flag = (cpi->ref_frame_flags & VP8_GOLD_FRAME);
+    parent_ref_flag = 0;
+    // Note availability for mv reuse is only based on last and golden.
+    if (parent_ref_frame == LAST_FRAME)
+      parent_ref_flag = (cpi->ref_frame_flags & VP8_LAST_FRAME);
+    else if (parent_ref_frame == GOLDEN_FRAME)
+      parent_ref_flag = (cpi->ref_frame_flags & VP8_GOLD_FRAME);
 
-        //assert(!parent_ref_frame || parent_ref_flag);
+    // assert(!parent_ref_frame || parent_ref_flag);
 
-        // If |parent_ref_frame| did not match either last or golden then
-        // shut off mv reuse.
-        if (parent_ref_frame && !parent_ref_flag)
-            parent_ref_valid = 0;
+    // If |parent_ref_frame| did not match either last or golden then
+    // shut off mv reuse.
+    if (parent_ref_frame && !parent_ref_flag) parent_ref_valid = 0;
 
-        // Don't do mv reuse since we want to allow for another mode besides
-        // ZEROMV_LAST to remove dot artifact.
-        if (dot_artifact_candidate)
-          parent_ref_valid = 0;
-    }
+    // Don't do mv reuse since we want to allow for another mode besides
+    // ZEROMV_LAST to remove dot artifact.
+    if (dot_artifact_candidate) parent_ref_valid = 0;
+  }
 #endif
 
-    // Check if current macroblock is in skin area.
-    {
+  // Check if current macroblock is in skin area.
+  {
     const int y = (x->src.y_buffer[7 * x->src.y_stride + 7] +
-        x->src.y_buffer[7 * x->src.y_stride + 8] +
-        x->src.y_buffer[8 * x->src.y_stride + 7] +
-        x->src.y_buffer[8 * x->src.y_stride + 8]) >> 2;
+                   x->src.y_buffer[7 * x->src.y_stride + 8] +
+                   x->src.y_buffer[8 * x->src.y_stride + 7] +
+                   x->src.y_buffer[8 * x->src.y_stride + 8]) >>
+                  2;
     const int cb = (x->src.u_buffer[3 * x->src.uv_stride + 3] +
-        x->src.u_buffer[3 * x->src.uv_stride + 4] +
-        x->src.u_buffer[4 * x->src.uv_stride + 3] +
-        x->src.u_buffer[4 * x->src.uv_stride + 4]) >> 2;
+                    x->src.u_buffer[3 * x->src.uv_stride + 4] +
+                    x->src.u_buffer[4 * x->src.uv_stride + 3] +
+                    x->src.u_buffer[4 * x->src.uv_stride + 4]) >>
+                   2;
     const int cr = (x->src.v_buffer[3 * x->src.uv_stride + 3] +
-        x->src.v_buffer[3 * x->src.uv_stride + 4] +
-        x->src.v_buffer[4 * x->src.uv_stride + 3] +
-        x->src.v_buffer[4 * x->src.uv_stride + 4]) >> 2;
+                    x->src.v_buffer[3 * x->src.uv_stride + 4] +
+                    x->src.v_buffer[4 * x->src.uv_stride + 3] +
+                    x->src.v_buffer[4 * x->src.uv_stride + 4]) >>
+                   2;
     x->is_skin = 0;
     if (!cpi->oxcf.screen_content_mode) {
       int block_index = mb_row * cpi->common.mb_cols + mb_col;
       x->is_skin = is_skin_color(y, cb, cr, cpi->consec_zero_last[block_index]);
     }
-    }
+  }
 #if CONFIG_TEMPORAL_DENOISING
-    if (cpi->oxcf.noise_sensitivity) {
-      // Under aggressive denoising mode, should we use skin map to reduce denoiser
-      // and ZEROMV bias? Will need to revisit the accuracy of this detection for
-      // very noisy input. For now keep this as is (i.e., don't turn it off).
-      // if (cpi->denoiser.denoiser_mode == kDenoiserOnYUVAggressive)
-      //   x->is_skin = 0;
-    }
+  if (cpi->oxcf.noise_sensitivity) {
+    // Under aggressive denoising mode, should we use skin map to reduce
+    // denoiser
+    // and ZEROMV bias? Will need to revisit the accuracy of this detection for
+    // very noisy input. For now keep this as is (i.e., don't turn it off).
+    // if (cpi->denoiser.denoiser_mode == kDenoiserOnYUVAggressive)
+    //   x->is_skin = 0;
+  }
 #endif
 
-    mode_mv = mode_mv_sb[sign_bias];
-    best_ref_mv.as_int = 0;
-    memset(mode_mv_sb, 0, sizeof(mode_mv_sb));
-    memset(&best_mbmode, 0, sizeof(best_mbmode));
+  mode_mv = mode_mv_sb[sign_bias];
+  best_ref_mv.as_int = 0;
+  memset(mode_mv_sb, 0, sizeof(mode_mv_sb));
+  memset(&best_mbmode, 0, sizeof(best_mbmode));
 
-    /* Setup search priorities */
+/* Setup search priorities */
 #if CONFIG_MULTI_RES_ENCODING
-    if (parent_ref_valid && parent_ref_frame && dissim < 8)
-    {
-        ref_frame_map[0] = -1;
-        ref_frame_map[1] = parent_ref_frame;
-        ref_frame_map[2] = -1;
-        ref_frame_map[3] = -1;
-    } else
+  if (parent_ref_valid && parent_ref_frame && dissim < 8) {
+    ref_frame_map[0] = -1;
+    ref_frame_map[1] = parent_ref_frame;
+    ref_frame_map[2] = -1;
+    ref_frame_map[3] = -1;
+  } else
 #endif
     get_reference_search_order(cpi, ref_frame_map);
 
-    /* Check to see if there is at least 1 valid reference frame that we need
-     * to calculate near_mvs.
-     */
-    if (ref_frame_map[1] > 0)
-    {
-        sign_bias = vp8_find_near_mvs_bias(&x->e_mbd,
-                                           x->e_mbd.mode_info_context,
-                                           mode_mv_sb,
-                                           best_ref_mv_sb,
-                                           mdcounts,
-                                           ref_frame_map[1],
-                                           cpi->common.ref_frame_sign_bias);
+  /* Check to see if there is at least 1 valid reference frame that we need
+   * to calculate near_mvs.
+   */
+  if (ref_frame_map[1] > 0) {
+    sign_bias = vp8_find_near_mvs_bias(
+        &x->e_mbd, x->e_mbd.mode_info_context, mode_mv_sb, best_ref_mv_sb,
+        mdcounts, ref_frame_map[1], cpi->common.ref_frame_sign_bias);
 
-        mode_mv = mode_mv_sb[sign_bias];
-        best_ref_mv.as_int = best_ref_mv_sb[sign_bias].as_int;
-    }
+    mode_mv = mode_mv_sb[sign_bias];
+    best_ref_mv.as_int = best_ref_mv_sb[sign_bias].as_int;
+  }
 
-    /* Count of the number of MBs tested so far this frame */
-    x->mbs_tested_so_far++;
+  /* Count of the number of MBs tested so far this frame */
+  x->mbs_tested_so_far++;
 
-    *returnintra = INT_MAX;
-    x->skip = 0;
+  *returnintra = INT_MAX;
+  x->skip = 0;
 
-    x->e_mbd.mode_info_context->mbmi.ref_frame = INTRA_FRAME;
+  x->e_mbd.mode_info_context->mbmi.ref_frame = INTRA_FRAME;
 
-    /* If the frame has big static background and current MB is in low
-    *  motion area, its mode decision is biased to ZEROMV mode.
-    *  No adjustment if cpu_used is <= -12 (i.e., cpi->Speed >= 12).
-    *  At such speed settings, ZEROMV is already heavily favored.
-    */
-    if (cpi->Speed < 12) {
-      calculate_zeromv_rd_adjustment(cpi, x, &rd_adjustment);
-    }
+  /* If the frame has big static background and current MB is in low
+  *  motion area, its mode decision is biased to ZEROMV mode.
+  *  No adjustment if cpu_used is <= -12 (i.e., cpi->Speed >= 12).
+  *  At such speed settings, ZEROMV is already heavily favored.
+  */
+  if (cpi->Speed < 12) {
+    calculate_zeromv_rd_adjustment(cpi, x, &rd_adjustment);
+  }
 
 #if CONFIG_TEMPORAL_DENOISING
-    if (cpi->oxcf.noise_sensitivity) {
-      rd_adjustment = (int)(rd_adjustment *
-          cpi->denoiser.denoise_pars.pickmode_mv_bias / 100);
-    }
+  if (cpi->oxcf.noise_sensitivity) {
+    rd_adjustment = (int)(rd_adjustment *
+                          cpi->denoiser.denoise_pars.pickmode_mv_bias / 100);
+  }
 #endif
 
-    if (dot_artifact_candidate)
-    {
-        // Bias against ZEROMV_LAST mode.
-        rd_adjustment = 150;
-    }
+  if (dot_artifact_candidate) {
+    // Bias against ZEROMV_LAST mode.
+    rd_adjustment = 150;
+  }
 
+  /* if we encode a new mv this is important
+   * find the best new motion vector
+   */
+  for (mode_index = 0; mode_index < MAX_MODES; mode_index++) {
+    int frame_cost;
+    int this_rd = INT_MAX;
+    int this_ref_frame = ref_frame_map[vp8_ref_frame_order[mode_index]];
 
-    /* if we encode a new mv this is important
-     * find the best new motion vector
-     */
-    for (mode_index = 0; mode_index < MAX_MODES; mode_index++)
-    {
-        int frame_cost;
-        int this_rd = INT_MAX;
-        int this_ref_frame = ref_frame_map[vp8_ref_frame_order[mode_index]];
+    if (best_rd <= x->rd_threshes[mode_index]) continue;
 
-        if (best_rd <= x->rd_threshes[mode_index])
-            continue;
+    if (this_ref_frame < 0) continue;
 
-        if (this_ref_frame < 0)
-            continue;
+    x->e_mbd.mode_info_context->mbmi.ref_frame = this_ref_frame;
 
-        x->e_mbd.mode_info_context->mbmi.ref_frame = this_ref_frame;
+    /* everything but intra */
+    if (x->e_mbd.mode_info_context->mbmi.ref_frame) {
+      x->e_mbd.pre.y_buffer = plane[this_ref_frame][0];
+      x->e_mbd.pre.u_buffer = plane[this_ref_frame][1];
+      x->e_mbd.pre.v_buffer = plane[this_ref_frame][2];
 
-        /* everything but intra */
-        if (x->e_mbd.mode_info_context->mbmi.ref_frame)
-        {
-            x->e_mbd.pre.y_buffer = plane[this_ref_frame][0];
-            x->e_mbd.pre.u_buffer = plane[this_ref_frame][1];
-            x->e_mbd.pre.v_buffer = plane[this_ref_frame][2];
+      if (sign_bias != cpi->common.ref_frame_sign_bias[this_ref_frame]) {
+        sign_bias = cpi->common.ref_frame_sign_bias[this_ref_frame];
+        mode_mv = mode_mv_sb[sign_bias];
+        best_ref_mv.as_int = best_ref_mv_sb[sign_bias].as_int;
+      }
 
-            if (sign_bias != cpi->common.ref_frame_sign_bias[this_ref_frame])
-            {
-                sign_bias = cpi->common.ref_frame_sign_bias[this_ref_frame];
-                mode_mv = mode_mv_sb[sign_bias];
-                best_ref_mv.as_int = best_ref_mv_sb[sign_bias].as_int;
-            }
-
 #if CONFIG_MULTI_RES_ENCODING
-            if (parent_ref_valid)
-            {
-                if (vp8_mode_order[mode_index] == NEARESTMV &&
-                    mode_mv[NEARESTMV].as_int ==0)
-                    continue;
-                if (vp8_mode_order[mode_index] == NEARMV &&
-                    mode_mv[NEARMV].as_int ==0)
-                    continue;
+      if (parent_ref_valid) {
+        if (vp8_mode_order[mode_index] == NEARESTMV &&
+            mode_mv[NEARESTMV].as_int == 0)
+          continue;
+        if (vp8_mode_order[mode_index] == NEARMV && mode_mv[NEARMV].as_int == 0)
+          continue;
 
-                if (vp8_mode_order[mode_index] == NEWMV && parent_mode == ZEROMV
-                    && best_ref_mv.as_int==0)
-                    continue;
-                else if(vp8_mode_order[mode_index] == NEWMV && dissim==0
-                    && best_ref_mv.as_int==parent_ref_mv.as_int)
-                    continue;
-            }
+        if (vp8_mode_order[mode_index] == NEWMV && parent_mode == ZEROMV &&
+            best_ref_mv.as_int == 0)
+          continue;
+        else if (vp8_mode_order[mode_index] == NEWMV && dissim == 0 &&
+                 best_ref_mv.as_int == parent_ref_mv.as_int)
+          continue;
+      }
 #endif
-        }
+    }
 
-        /* Check to see if the testing frequency for this mode is at its max
-         * If so then prevent it from being tested and increase the threshold
-         * for its testing */
-        if (x->mode_test_hit_counts[mode_index] &&
-                                         (cpi->mode_check_freq[mode_index] > 1))
-        {
-            if (x->mbs_tested_so_far <= (cpi->mode_check_freq[mode_index] *
-                                         x->mode_test_hit_counts[mode_index]))
-            {
-                /* Increase the threshold for coding this mode to make it less
-                 * likely to be chosen */
-                x->rd_thresh_mult[mode_index] += 4;
+    /* Check to see if the testing frequency for this mode is at its max
+     * If so then prevent it from being tested and increase the threshold
+     * for its testing */
+    if (x->mode_test_hit_counts[mode_index] &&
+        (cpi->mode_check_freq[mode_index] > 1)) {
+      if (x->mbs_tested_so_far <= (cpi->mode_check_freq[mode_index] *
+                                   x->mode_test_hit_counts[mode_index])) {
+        /* Increase the threshold for coding this mode to make it less
+         * likely to be chosen */
+        x->rd_thresh_mult[mode_index] += 4;
 
-                if (x->rd_thresh_mult[mode_index] > MAX_THRESHMULT)
-                    x->rd_thresh_mult[mode_index] = MAX_THRESHMULT;
+        if (x->rd_thresh_mult[mode_index] > MAX_THRESHMULT)
+          x->rd_thresh_mult[mode_index] = MAX_THRESHMULT;
 
-                x->rd_threshes[mode_index] =
-                                 (cpi->rd_baseline_thresh[mode_index] >> 7) *
-                                 x->rd_thresh_mult[mode_index];
-                continue;
-            }
-        }
+        x->rd_threshes[mode_index] =
+            (cpi->rd_baseline_thresh[mode_index] >> 7) *
+            x->rd_thresh_mult[mode_index];
+        continue;
+      }
+    }
 
-        /* We have now reached the point where we are going to test the current
-         * mode so increment the counter for the number of times it has been
-         * tested */
-        x->mode_test_hit_counts[mode_index] ++;
+    /* We have now reached the point where we are going to test the current
+     * mode so increment the counter for the number of times it has been
+     * tested */
+    x->mode_test_hit_counts[mode_index]++;
 
-        rate2 = 0;
-        distortion2 = 0;
+    rate2 = 0;
+    distortion2 = 0;
 
-        this_mode = vp8_mode_order[mode_index];
+    this_mode = vp8_mode_order[mode_index];
 
-        x->e_mbd.mode_info_context->mbmi.mode = this_mode;
-        x->e_mbd.mode_info_context->mbmi.uv_mode = DC_PRED;
+    x->e_mbd.mode_info_context->mbmi.mode = this_mode;
+    x->e_mbd.mode_info_context->mbmi.uv_mode = DC_PRED;
 
-        /* Work out the cost assosciated with selecting the reference frame */
-        frame_cost =
-            x->ref_frame_cost[x->e_mbd.mode_info_context->mbmi.ref_frame];
-        rate2 += frame_cost;
+    /* Work out the cost assosciated with selecting the reference frame */
+    frame_cost = x->ref_frame_cost[x->e_mbd.mode_info_context->mbmi.ref_frame];
+    rate2 += frame_cost;
 
-        /* Only consider ZEROMV/ALTREF_FRAME for alt ref frame,
-         * unless ARNR filtering is enabled in which case we want
-         * an unfiltered alternative */
-        if (cpi->is_src_frame_alt_ref && (cpi->oxcf.arnr_max_frames == 0))
-        {
-            if (this_mode != ZEROMV ||
-                x->e_mbd.mode_info_context->mbmi.ref_frame != ALTREF_FRAME)
-                continue;
-        }
+    /* Only consider ZEROMV/ALTREF_FRAME for alt ref frame,
+     * unless ARNR filtering is enabled in which case we want
+     * an unfiltered alternative */
+    if (cpi->is_src_frame_alt_ref && (cpi->oxcf.arnr_max_frames == 0)) {
+      if (this_mode != ZEROMV ||
+          x->e_mbd.mode_info_context->mbmi.ref_frame != ALTREF_FRAME)
+        continue;
+    }
 
-        switch (this_mode)
-        {
-        case B_PRED:
-            /* Pass best so far to pick_intra4x4mby_modes to use as breakout */
-            distortion2 = best_rd_sse;
-            pick_intra4x4mby_modes(x, &rate, &distortion2);
+    switch (this_mode) {
+      case B_PRED:
+        /* Pass best so far to pick_intra4x4mby_modes to use as breakout */
+        distortion2 = best_rd_sse;
+        pick_intra4x4mby_modes(x, &rate, &distortion2);
 
-            if (distortion2 == INT_MAX)
-            {
-                this_rd = INT_MAX;
-            }
-            else
-            {
-                rate2 += rate;
-                distortion2 = vpx_variance16x16(
-                                    *(b->base_src), b->src_stride,
-                                    x->e_mbd.predictor, 16, &sse);
-                this_rd = RDCOST(x->rdmult, x->rddiv, rate2, distortion2);
+        if (distortion2 == INT_MAX) {
+          this_rd = INT_MAX;
+        } else {
+          rate2 += rate;
+          distortion2 = vpx_variance16x16(*(b->base_src), b->src_stride,
+                                          x->e_mbd.predictor, 16, &sse);
+          this_rd = RDCOST(x->rdmult, x->rddiv, rate2, distortion2);
 
-                if (this_rd < best_intra_rd)
-                {
-                    best_intra_rd = this_rd;
-                    *returnintra = distortion2;
-                }
-            }
+          if (this_rd < best_intra_rd) {
+            best_intra_rd = this_rd;
+            *returnintra = distortion2;
+          }
+        }
 
-            break;
+        break;
 
-        case SPLITMV:
+      case SPLITMV:
 
-            /* Split MV modes currently not supported when RD is not enabled. */
-            break;
+        /* Split MV modes currently not supported when RD is not enabled. */
+        break;
 
-        case DC_PRED:
-        case V_PRED:
-        case H_PRED:
-        case TM_PRED:
-            vp8_build_intra_predictors_mby_s(xd,
-                                             xd->dst.y_buffer - xd->dst.y_stride,
-                                             xd->dst.y_buffer - 1,
-                                             xd->dst.y_stride,
-                                             xd->predictor,
-                                             16);
-            distortion2 = vpx_variance16x16
-                                          (*(b->base_src), b->src_stride,
-                                          x->e_mbd.predictor, 16, &sse);
-            rate2 += x->mbmode_cost[x->e_mbd.frame_type][x->e_mbd.mode_info_context->mbmi.mode];
-            this_rd = RDCOST(x->rdmult, x->rddiv, rate2, distortion2);
+      case DC_PRED:
+      case V_PRED:
+      case H_PRED:
+      case TM_PRED:
+        vp8_build_intra_predictors_mby_s(
+            xd, xd->dst.y_buffer - xd->dst.y_stride, xd->dst.y_buffer - 1,
+            xd->dst.y_stride, xd->predictor, 16);
+        distortion2 = vpx_variance16x16(*(b->base_src), b->src_stride,
+                                        x->e_mbd.predictor, 16, &sse);
+        rate2 += x->mbmode_cost[x->e_mbd.frame_type][x->e_mbd.mode_info_context
+                                                         ->mbmi.mode];
+        this_rd = RDCOST(x->rdmult, x->rddiv, rate2, distortion2);
 
-            if (this_rd < best_intra_rd)
-            {
-                best_intra_rd = this_rd;
-                *returnintra = distortion2;
-            }
-            break;
+        if (this_rd < best_intra_rd) {
+          best_intra_rd = this_rd;
+          *returnintra = distortion2;
+        }
+        break;
 
-        case NEWMV:
-        {
-            int thissme;
-            int step_param;
-            int further_steps;
-            int n = 0;
-            int sadpb = x->sadperbit16;
-            int_mv mvp_full;
+      case NEWMV: {
+        int thissme;
+        int step_param;
+        int further_steps;
+        int n = 0;
+        int sadpb = x->sadperbit16;
+        int_mv mvp_full;
 
-            int col_min = ((best_ref_mv.as_mv.col+7)>>3) - MAX_FULL_PEL_VAL;
-            int row_min = ((best_ref_mv.as_mv.row+7)>>3) - MAX_FULL_PEL_VAL;
-            int col_max = (best_ref_mv.as_mv.col>>3)
-                         + MAX_FULL_PEL_VAL;
-            int row_max = (best_ref_mv.as_mv.row>>3)
-                         + MAX_FULL_PEL_VAL;
+        int col_min = ((best_ref_mv.as_mv.col + 7) >> 3) - MAX_FULL_PEL_VAL;
+        int row_min = ((best_ref_mv.as_mv.row + 7) >> 3) - MAX_FULL_PEL_VAL;
+        int col_max = (best_ref_mv.as_mv.col >> 3) + MAX_FULL_PEL_VAL;
+        int row_max = (best_ref_mv.as_mv.row >> 3) + MAX_FULL_PEL_VAL;
 
-            int tmp_col_min = x->mv_col_min;
-            int tmp_col_max = x->mv_col_max;
-            int tmp_row_min = x->mv_row_min;
-            int tmp_row_max = x->mv_row_max;
+        int tmp_col_min = x->mv_col_min;
+        int tmp_col_max = x->mv_col_max;
+        int tmp_row_min = x->mv_row_min;
+        int tmp_row_max = x->mv_row_max;
 
-            int speed_adjust = (cpi->Speed > 5) ? ((cpi->Speed >= 8)? 3 : 2) : 1;
+        int speed_adjust = (cpi->Speed > 5) ? ((cpi->Speed >= 8) ? 3 : 2) : 1;
 
-            /* Further step/diamond searches as necessary */
-            step_param = cpi->sf.first_step + speed_adjust;
+        /* Further step/diamond searches as necessary */
+        step_param = cpi->sf.first_step + speed_adjust;
 
 #if CONFIG_MULTI_RES_ENCODING
-            /* If lower-res frame is not available for mv reuse (because of
-               frame dropping or different temporal layer pattern), then higher
-               resol encoder does motion search without any previous knowledge.
-               Also, since last frame motion info is not stored, then we can not
-               use improved_mv_pred. */
-            if (cpi->oxcf.mr_encoder_id)
-                sf_improved_mv_pred = 0;
+        /* If lower-res frame is not available for mv reuse (because of
+           frame dropping or different temporal layer pattern), then higher
+           resol encoder does motion search without any previous knowledge.
+           Also, since last frame motion info is not stored, then we can not
+           use improved_mv_pred. */
+        if (cpi->oxcf.mr_encoder_id) sf_improved_mv_pred = 0;
 
-            // Only use parent MV as predictor if this candidate reference frame
-            // (|this_ref_frame|) is equal to |parent_ref_frame|.
-            if (parent_ref_valid && (parent_ref_frame == this_ref_frame))
-            {
-                /* Use parent MV as predictor. Adjust search range
-                 * accordingly.
-                 */
-                mvp.as_int = parent_ref_mv.as_int;
-                mvp_full.as_mv.col = parent_ref_mv.as_mv.col>>3;
-                mvp_full.as_mv.row = parent_ref_mv.as_mv.row>>3;
+        // Only use parent MV as predictor if this candidate reference frame
+        // (|this_ref_frame|) is equal to |parent_ref_frame|.
+        if (parent_ref_valid && (parent_ref_frame == this_ref_frame)) {
+          /* Use parent MV as predictor. Adjust search range
+           * accordingly.
+           */
+          mvp.as_int = parent_ref_mv.as_int;
+          mvp_full.as_mv.col = parent_ref_mv.as_mv.col >> 3;
+          mvp_full.as_mv.row = parent_ref_mv.as_mv.row >> 3;
 
-                if(dissim <=32) step_param += 3;
-                else if(dissim <=128) step_param += 2;
-                else step_param += 1;
-            }else
+          if (dissim <= 32)
+            step_param += 3;
+          else if (dissim <= 128)
+            step_param += 2;
+          else
+            step_param += 1;
+        } else
 #endif
-            {
-                if(sf_improved_mv_pred)
-                {
-                    if(!saddone)
-                    {
-                        vp8_cal_sad(cpi,xd,x, recon_yoffset ,&near_sadidx[0] );
-                        saddone = 1;
-                    }
+        {
+          if (sf_improved_mv_pred) {
+            if (!saddone) {
+              vp8_cal_sad(cpi, xd, x, recon_yoffset, &near_sadidx[0]);
+              saddone = 1;
+            }
 
-                    vp8_mv_pred(cpi, &x->e_mbd, x->e_mbd.mode_info_context,
-                                &mvp,x->e_mbd.mode_info_context->mbmi.ref_frame,
-                                cpi->common.ref_frame_sign_bias, &sr,
-                                &near_sadidx[0]);
+            vp8_mv_pred(cpi, &x->e_mbd, x->e_mbd.mode_info_context, &mvp,
+                        x->e_mbd.mode_info_context->mbmi.ref_frame,
+                        cpi->common.ref_frame_sign_bias, &sr, &near_sadidx[0]);
 
-                    sr += speed_adjust;
-                    /* adjust search range according to sr from mv prediction */
-                    if(sr > step_param)
-                        step_param = sr;
+            sr += speed_adjust;
+            /* adjust search range according to sr from mv prediction */
+            if (sr > step_param) step_param = sr;
 
-                    mvp_full.as_mv.col = mvp.as_mv.col>>3;
-                    mvp_full.as_mv.row = mvp.as_mv.row>>3;
-                }else
-                {
-                    mvp.as_int = best_ref_mv.as_int;
-                    mvp_full.as_mv.col = best_ref_mv.as_mv.col>>3;
-                    mvp_full.as_mv.row = best_ref_mv.as_mv.row>>3;
-                }
-            }
+            mvp_full.as_mv.col = mvp.as_mv.col >> 3;
+            mvp_full.as_mv.row = mvp.as_mv.row >> 3;
+          } else {
+            mvp.as_int = best_ref_mv.as_int;
+            mvp_full.as_mv.col = best_ref_mv.as_mv.col >> 3;
+            mvp_full.as_mv.row = best_ref_mv.as_mv.row >> 3;
+          }
+        }
 
 #if CONFIG_MULTI_RES_ENCODING
-            if (parent_ref_valid && (parent_ref_frame == this_ref_frame) &&
-                dissim <= 2 &&
-                VPXMAX(abs(best_ref_mv.as_mv.row - parent_ref_mv.as_mv.row),
-                       abs(best_ref_mv.as_mv.col - parent_ref_mv.as_mv.col)) <=
-                    4)
-            {
-                d->bmi.mv.as_int = mvp_full.as_int;
-                mode_mv[NEWMV].as_int = mvp_full.as_int;
+        if (parent_ref_valid && (parent_ref_frame == this_ref_frame) &&
+            dissim <= 2 &&
+            VPXMAX(abs(best_ref_mv.as_mv.row - parent_ref_mv.as_mv.row),
+                   abs(best_ref_mv.as_mv.col - parent_ref_mv.as_mv.col)) <= 4) {
+          d->bmi.mv.as_int = mvp_full.as_int;
+          mode_mv[NEWMV].as_int = mvp_full.as_int;
 
-                cpi->find_fractional_mv_step(x, b, d, &d->bmi.mv, &best_ref_mv,
-                                             x->errorperbit,
-                                             &cpi->fn_ptr[BLOCK_16X16],
-                                             cpi->mb.mvcost,
-                                             &distortion2,&sse);
-            }else
+          cpi->find_fractional_mv_step(
+              x, b, d, &d->bmi.mv, &best_ref_mv, x->errorperbit,
+              &cpi->fn_ptr[BLOCK_16X16], cpi->mb.mvcost, &distortion2, &sse);
+        } else
 #endif
-            {
-                /* Get intersection of UMV window and valid MV window to
-                 * reduce # of checks in diamond search. */
-                if (x->mv_col_min < col_min )
-                    x->mv_col_min = col_min;
-                if (x->mv_col_max > col_max )
-                    x->mv_col_max = col_max;
-                if (x->mv_row_min < row_min )
-                    x->mv_row_min = row_min;
-                if (x->mv_row_max > row_max )
-                    x->mv_row_max = row_max;
+        {
+          /* Get intersection of UMV window and valid MV window to
+           * reduce # of checks in diamond search. */
+          if (x->mv_col_min < col_min) x->mv_col_min = col_min;
+          if (x->mv_col_max > col_max) x->mv_col_max = col_max;
+          if (x->mv_row_min < row_min) x->mv_row_min = row_min;
+          if (x->mv_row_max > row_max) x->mv_row_max = row_max;
 
-                further_steps = (cpi->Speed >= 8)?
-                           0: (cpi->sf.max_step_search_steps - 1 - step_param);
+          further_steps =
+              (cpi->Speed >= 8)
+                  ? 0
+                  : (cpi->sf.max_step_search_steps - 1 - step_param);
 
-                if (cpi->sf.search_method == HEX)
-                {
+          if (cpi->sf.search_method == HEX) {
 #if CONFIG_MULTI_RES_ENCODING
-                /* TODO: In higher-res pick_inter_mode, step_param is used to
-                 * modify hex search range. Here, set step_param to 0 not to
-                 * change the behavior in lowest-resolution encoder.
-                 * Will improve it later.
-                 */
-                /* Set step_param to 0 to ensure large-range motion search
-                 * when mv reuse if not valid (i.e. |parent_ref_valid| = 0),
-                 * or if this candidate reference frame (|this_ref_frame|) is
-                 * not equal to |parent_ref_frame|.
-                 */
-                if (!parent_ref_valid || (parent_ref_frame != this_ref_frame))
-                    step_param = 0;
+            /* TODO: In higher-res pick_inter_mode, step_param is used to
+             * modify hex search range. Here, set step_param to 0 not to
+             * change the behavior in lowest-resolution encoder.
+             * Will improve it later.
+             */
+            /* Set step_param to 0 to ensure large-range motion search
+             * when mv reuse if not valid (i.e. |parent_ref_valid| = 0),
+             * or if this candidate reference frame (|this_ref_frame|) is
+             * not equal to |parent_ref_frame|.
+             */
+            if (!parent_ref_valid || (parent_ref_frame != this_ref_frame))
+              step_param = 0;
 #endif
-                    bestsme = vp8_hex_search(x, b, d, &mvp_full, &d->bmi.mv,
-                                          step_param, sadpb,
-                                          &cpi->fn_ptr[BLOCK_16X16],
-                                          x->mvsadcost, x->mvcost, &best_ref_mv);
-                    mode_mv[NEWMV].as_int = d->bmi.mv.as_int;
-                }
-                else
-                {
-                    bestsme = cpi->diamond_search_sad(x, b, d, &mvp_full,
-                                          &d->bmi.mv, step_param, sadpb, &num00,
-                                          &cpi->fn_ptr[BLOCK_16X16],
-                                          x->mvcost, &best_ref_mv);
-                    mode_mv[NEWMV].as_int = d->bmi.mv.as_int;
+            bestsme = vp8_hex_search(x, b, d, &mvp_full, &d->bmi.mv, step_param,
+                                     sadpb, &cpi->fn_ptr[BLOCK_16X16],
+                                     x->mvsadcost, x->mvcost, &best_ref_mv);
+            mode_mv[NEWMV].as_int = d->bmi.mv.as_int;
+          } else {
+            bestsme = cpi->diamond_search_sad(
+                x, b, d, &mvp_full, &d->bmi.mv, step_param, sadpb, &num00,
+                &cpi->fn_ptr[BLOCK_16X16], x->mvcost, &best_ref_mv);
+            mode_mv[NEWMV].as_int = d->bmi.mv.as_int;
 
-                    /* Further step/diamond searches as necessary */
-                    n = num00;
-                    num00 = 0;
+            /* Further step/diamond searches as necessary */
+            n = num00;
+            num00 = 0;
 
-                    while (n < further_steps)
-                    {
-                        n++;
+            while (n < further_steps) {
+              n++;
 
-                        if (num00)
-                            num00--;
-                        else
-                        {
-                            thissme =
-                            cpi->diamond_search_sad(x, b, d, &mvp_full,
-                                                    &d->bmi.mv,
-                                                    step_param + n,
-                                                    sadpb, &num00,
-                                                    &cpi->fn_ptr[BLOCK_16X16],
-                                                    x->mvcost, &best_ref_mv);
-                            if (thissme < bestsme)
-                            {
-                                bestsme = thissme;
-                                mode_mv[NEWMV].as_int = d->bmi.mv.as_int;
-                            }
-                            else
-                            {
-                                d->bmi.mv.as_int = mode_mv[NEWMV].as_int;
-                            }
-                        }
-                    }
+              if (num00)
+                num00--;
+              else {
+                thissme = cpi->diamond_search_sad(
+                    x, b, d, &mvp_full, &d->bmi.mv, step_param + n, sadpb,
+                    &num00, &cpi->fn_ptr[BLOCK_16X16], x->mvcost, &best_ref_mv);
+                if (thissme < bestsme) {
+                  bestsme = thissme;
+                  mode_mv[NEWMV].as_int = d->bmi.mv.as_int;
+                } else {
+                  d->bmi.mv.as_int = mode_mv[NEWMV].as_int;
                 }
-
-                x->mv_col_min = tmp_col_min;
-                x->mv_col_max = tmp_col_max;
-                x->mv_row_min = tmp_row_min;
-                x->mv_row_max = tmp_row_max;
-
-                if (bestsme < INT_MAX)
-                    cpi->find_fractional_mv_step(x, b, d, &d->bmi.mv,
-                                             &best_ref_mv, x->errorperbit,
-                                             &cpi->fn_ptr[BLOCK_16X16],
-                                             cpi->mb.mvcost,
-                                             &distortion2,&sse);
+              }
             }
+          }
 
-            mode_mv[NEWMV].as_int = d->bmi.mv.as_int;
-            // The clamp below is not necessary from the perspective
-            // of VP8 bitstream, but is added to improve ChromeCast
-            // mirroring's robustness. Please do not remove.
-            vp8_clamp_mv2(&mode_mv[this_mode], xd);
-            /* mv cost; */
-            rate2 += vp8_mv_bit_cost(&mode_mv[NEWMV], &best_ref_mv,
-                                     cpi->mb.mvcost, 128);
+          x->mv_col_min = tmp_col_min;
+          x->mv_col_max = tmp_col_max;
+          x->mv_row_min = tmp_row_min;
+          x->mv_row_max = tmp_row_max;
+
+          if (bestsme < INT_MAX)
+            cpi->find_fractional_mv_step(
+                x, b, d, &d->bmi.mv, &best_ref_mv, x->errorperbit,
+                &cpi->fn_ptr[BLOCK_16X16], cpi->mb.mvcost, &distortion2, &sse);
         }
 
-        case NEARESTMV:
-        case NEARMV:
-            if (mode_mv[this_mode].as_int == 0)
-                continue;
+        mode_mv[NEWMV].as_int = d->bmi.mv.as_int;
+        // The clamp below is not necessary from the perspective
+        // of VP8 bitstream, but is added to improve ChromeCast
+        // mirroring's robustness. Please do not remove.
+        vp8_clamp_mv2(&mode_mv[this_mode], xd);
+        /* mv cost; */
+        rate2 +=
+            vp8_mv_bit_cost(&mode_mv[NEWMV], &best_ref_mv, cpi->mb.mvcost, 128);
+      }
 
-        case ZEROMV:
+      case NEARESTMV:
+      case NEARMV:
+        if (mode_mv[this_mode].as_int == 0) continue;
 
-            /* Trap vectors that reach beyond the UMV borders
-             * Note that ALL New MV, Nearest MV Near MV and Zero MV code drops
-             * through to this point because of the lack of break statements
-             * in the previous two cases.
-             */
-            if (((mode_mv[this_mode].as_mv.row >> 3) < x->mv_row_min) ||
-                ((mode_mv[this_mode].as_mv.row >> 3) > x->mv_row_max) ||
-                ((mode_mv[this_mode].as_mv.col >> 3) < x->mv_col_min) ||
-                ((mode_mv[this_mode].as_mv.col >> 3) > x->mv_col_max))
-                continue;
+      case ZEROMV:
 
-            rate2 += vp8_cost_mv_ref(this_mode, mdcounts);
-            x->e_mbd.mode_info_context->mbmi.mv.as_int =
-                                                    mode_mv[this_mode].as_int;
-            this_rd = evaluate_inter_mode(&sse, rate2, &distortion2, cpi, x,
-                                          rd_adjustment);
+        /* Trap vectors that reach beyond the UMV borders
+         * Note that ALL New MV, Nearest MV Near MV and Zero MV code drops
+         * through to this point because of the lack of break statements
+         * in the previous two cases.
+         */
+        if (((mode_mv[this_mode].as_mv.row >> 3) < x->mv_row_min) ||
+            ((mode_mv[this_mode].as_mv.row >> 3) > x->mv_row_max) ||
+            ((mode_mv[this_mode].as_mv.col >> 3) < x->mv_col_min) ||
+            ((mode_mv[this_mode].as_mv.col >> 3) > x->mv_col_max))
+          continue;
 
-            break;
-        default:
-            break;
-        }
+        rate2 += vp8_cost_mv_ref(this_mode, mdcounts);
+        x->e_mbd.mode_info_context->mbmi.mv.as_int = mode_mv[this_mode].as_int;
+        this_rd = evaluate_inter_mode(&sse, rate2, &distortion2, cpi, x,
+                                      rd_adjustment);
 
+        break;
+      default: break;
+    }
+
 #if CONFIG_TEMPORAL_DENOISING
-        if (cpi->oxcf.noise_sensitivity)
-        {
-            /* Store for later use by denoiser. */
-            // Dont' denoise with GOLDEN OR ALTREF is they are old reference
-            // frames (greater than MAX_GF_ARF_DENOISE_RANGE frames in past).
-            int skip_old_reference = ((this_ref_frame != LAST_FRAME) &&
-                (cpi->common.current_video_frame -
-                 cpi->current_ref_frames[this_ref_frame] >
-                 MAX_GF_ARF_DENOISE_RANGE)) ? 1 : 0;
-            if (this_mode == ZEROMV && sse < zero_mv_sse &&
-                !skip_old_reference)
-            {
-                zero_mv_sse = sse;
-                x->best_zeromv_reference_frame =
-                        x->e_mbd.mode_info_context->mbmi.ref_frame;
-            }
+    if (cpi->oxcf.noise_sensitivity) {
+      /* Store for later use by denoiser. */
+      // Dont' denoise with GOLDEN OR ALTREF is they are old reference
+      // frames (greater than MAX_GF_ARF_DENOISE_RANGE frames in past).
+      int skip_old_reference = ((this_ref_frame != LAST_FRAME) &&
+                                (cpi->common.current_video_frame -
+                                     cpi->current_ref_frames[this_ref_frame] >
+                                 MAX_GF_ARF_DENOISE_RANGE))
+                                   ? 1
+                                   : 0;
+      if (this_mode == ZEROMV && sse < zero_mv_sse && !skip_old_reference) {
+        zero_mv_sse = sse;
+        x->best_zeromv_reference_frame =
+            x->e_mbd.mode_info_context->mbmi.ref_frame;
+      }
 
-            // Store the best NEWMV in x for later use in the denoiser.
-            if (x->e_mbd.mode_info_context->mbmi.mode == NEWMV &&
-                sse < best_sse && !skip_old_reference)
-            {
-                best_sse = sse;
-                x->best_sse_inter_mode = NEWMV;
-                x->best_sse_mv = x->e_mbd.mode_info_context->mbmi.mv;
-                x->need_to_clamp_best_mvs =
-                    x->e_mbd.mode_info_context->mbmi.need_to_clamp_mvs;
-                x->best_reference_frame =
-                    x->e_mbd.mode_info_context->mbmi.ref_frame;
-            }
-        }
+      // Store the best NEWMV in x for later use in the denoiser.
+      if (x->e_mbd.mode_info_context->mbmi.mode == NEWMV && sse < best_sse &&
+          !skip_old_reference) {
+        best_sse = sse;
+        x->best_sse_inter_mode = NEWMV;
+        x->best_sse_mv = x->e_mbd.mode_info_context->mbmi.mv;
+        x->need_to_clamp_best_mvs =
+            x->e_mbd.mode_info_context->mbmi.need_to_clamp_mvs;
+        x->best_reference_frame = x->e_mbd.mode_info_context->mbmi.ref_frame;
+      }
+    }
 #endif
 
-        if (this_rd < best_rd || x->skip)
-        {
-            /* Note index of best mode */
-            best_mode_index = mode_index;
+    if (this_rd < best_rd || x->skip) {
+      /* Note index of best mode */
+      best_mode_index = mode_index;
 
-            *returnrate = rate2;
-            *returndistortion = distortion2;
-            best_rd_sse = sse;
-            best_rd = this_rd;
-            memcpy(&best_mbmode, &x->e_mbd.mode_info_context->mbmi,
-                   sizeof(MB_MODE_INFO));
+      *returnrate = rate2;
+      *returndistortion = distortion2;
+      best_rd_sse = sse;
+      best_rd = this_rd;
+      memcpy(&best_mbmode, &x->e_mbd.mode_info_context->mbmi,
+             sizeof(MB_MODE_INFO));
 
-            /* Testing this mode gave rise to an improvement in best error
-             * score. Lower threshold a bit for next time
-             */
-            x->rd_thresh_mult[mode_index] =
-                     (x->rd_thresh_mult[mode_index] >= (MIN_THRESHMULT + 2)) ?
-                     x->rd_thresh_mult[mode_index] - 2 : MIN_THRESHMULT;
-            x->rd_threshes[mode_index] =
-                                   (cpi->rd_baseline_thresh[mode_index] >> 7) *
+      /* Testing this mode gave rise to an improvement in best error
+       * score. Lower threshold a bit for next time
+       */
+      x->rd_thresh_mult[mode_index] =
+          (x->rd_thresh_mult[mode_index] >= (MIN_THRESHMULT + 2))
+              ? x->rd_thresh_mult[mode_index] - 2
+              : MIN_THRESHMULT;
+      x->rd_threshes[mode_index] = (cpi->rd_baseline_thresh[mode_index] >> 7) *
                                    x->rd_thresh_mult[mode_index];
-        }
+    }
 
-        /* If the mode did not help improve the best error case then raise the
-         * threshold for testing that mode next time around.
-         */
-        else
-        {
-            x->rd_thresh_mult[mode_index] += 4;
+    /* If the mode did not help improve the best error case then raise the
+     * threshold for testing that mode next time around.
+     */
+    else {
+      x->rd_thresh_mult[mode_index] += 4;
 
-            if (x->rd_thresh_mult[mode_index] > MAX_THRESHMULT)
-                x->rd_thresh_mult[mode_index] = MAX_THRESHMULT;
+      if (x->rd_thresh_mult[mode_index] > MAX_THRESHMULT)
+        x->rd_thresh_mult[mode_index] = MAX_THRESHMULT;
 
-            x->rd_threshes[mode_index] =
-                         (cpi->rd_baseline_thresh[mode_index] >> 7) *
-                         x->rd_thresh_mult[mode_index];
-        }
-
-        if (x->skip)
-            break;
+      x->rd_threshes[mode_index] = (cpi->rd_baseline_thresh[mode_index] >> 7) *
+                                   x->rd_thresh_mult[mode_index];
     }
 
-    /* Reduce the activation RD thresholds for the best choice mode */
-    if ((cpi->rd_baseline_thresh[best_mode_index] > 0) && (cpi->rd_baseline_thresh[best_mode_index] < (INT_MAX >> 2)))
-    {
-        int best_adjustment = (x->rd_thresh_mult[best_mode_index] >> 3);
+    if (x->skip) break;
+  }
 
-        x->rd_thresh_mult[best_mode_index] =
-                        (x->rd_thresh_mult[best_mode_index]
-                        >= (MIN_THRESHMULT + best_adjustment)) ?
-                        x->rd_thresh_mult[best_mode_index] - best_adjustment :
-                        MIN_THRESHMULT;
-        x->rd_threshes[best_mode_index] =
-                        (cpi->rd_baseline_thresh[best_mode_index] >> 7) *
-                        x->rd_thresh_mult[best_mode_index];
-    }
+  /* Reduce the activation RD thresholds for the best choice mode */
+  if ((cpi->rd_baseline_thresh[best_mode_index] > 0) &&
+      (cpi->rd_baseline_thresh[best_mode_index] < (INT_MAX >> 2))) {
+    int best_adjustment = (x->rd_thresh_mult[best_mode_index] >> 3);
+
+    x->rd_thresh_mult[best_mode_index] =
+        (x->rd_thresh_mult[best_mode_index] >=
+         (MIN_THRESHMULT + best_adjustment))
+            ? x->rd_thresh_mult[best_mode_index] - best_adjustment
+            : MIN_THRESHMULT;
+    x->rd_threshes[best_mode_index] =
+        (cpi->rd_baseline_thresh[best_mode_index] >> 7) *
+        x->rd_thresh_mult[best_mode_index];
+  }
 
+  {
+    int this_rdbin = (*returndistortion >> 7);
 
-    {
-        int this_rdbin = (*returndistortion >> 7);
+    if (this_rdbin >= 1024) {
+      this_rdbin = 1023;
+    }
 
-        if (this_rdbin >= 1024)
-        {
-            this_rdbin = 1023;
-        }
+    x->error_bins[this_rdbin]++;
+  }
 
-        x->error_bins[this_rdbin] ++;
+#if CONFIG_TEMPORAL_DENOISING
+  if (cpi->oxcf.noise_sensitivity) {
+    int block_index = mb_row * cpi->common.mb_cols + mb_col;
+    int reevaluate = 0;
+    int is_noisy = 0;
+    if (x->best_sse_inter_mode == DC_PRED) {
+      /* No best MV found. */
+      x->best_sse_inter_mode = best_mbmode.mode;
+      x->best_sse_mv = best_mbmode.mv;
+      x->need_to_clamp_best_mvs = best_mbmode.need_to_clamp_mvs;
+      x->best_reference_frame = best_mbmode.ref_frame;
+      best_sse = best_rd_sse;
     }
+    // For non-skin blocks that have selected ZEROMV for this current frame,
+    // and have been selecting ZEROMV_LAST (on the base layer frame) at
+    // least |x~20| consecutive past frames in a row, label the block for
+    // possible increase in denoising strength. We also condition this
+    // labeling on there being significant denoising in the scene
+    if (cpi->oxcf.noise_sensitivity == 4) {
+      if (cpi->denoiser.nmse_source_diff >
+          70 * cpi->denoiser.threshold_aggressive_mode / 100)
+        is_noisy = 1;
+    } else {
+      if (cpi->mse_source_denoised > 1000) is_noisy = 1;
+    }
+    x->increase_denoising = 0;
+    if (!x->is_skin && x->best_sse_inter_mode == ZEROMV &&
+        (x->best_reference_frame == LAST_FRAME ||
+         x->best_reference_frame == cpi->closest_reference_frame) &&
+        cpi->consec_zero_last[block_index] >= 20 && is_noisy) {
+      x->increase_denoising = 1;
+    }
+    x->denoise_zeromv = 0;
+    vp8_denoiser_denoise_mb(&cpi->denoiser, x, best_sse, zero_mv_sse,
+                            recon_yoffset, recon_uvoffset, &cpi->common.lf_info,
+                            mb_row, mb_col, block_index,
+                            cpi->consec_zero_last_mvbias[block_index]);
 
-#if CONFIG_TEMPORAL_DENOISING
-    if (cpi->oxcf.noise_sensitivity)
-    {
-        int block_index = mb_row * cpi->common.mb_cols + mb_col;
-        int reevaluate = 0;
-        int is_noisy = 0;
-        if (x->best_sse_inter_mode == DC_PRED)
-        {
-            /* No best MV found. */
-            x->best_sse_inter_mode = best_mbmode.mode;
-            x->best_sse_mv = best_mbmode.mv;
-            x->need_to_clamp_best_mvs = best_mbmode.need_to_clamp_mvs;
-            x->best_reference_frame = best_mbmode.ref_frame;
-            best_sse = best_rd_sse;
-        }
-        // For non-skin blocks that have selected ZEROMV for this current frame,
-        // and have been selecting ZEROMV_LAST (on the base layer frame) at
-        // least |x~20| consecutive past frames in a row, label the block for
-        // possible increase in denoising strength. We also condition this
-        // labeling on there being significant denoising in the scene
-        if  (cpi->oxcf.noise_sensitivity == 4) {
-          if (cpi->denoiser.nmse_source_diff >
-              70 * cpi->denoiser.threshold_aggressive_mode / 100)
-            is_noisy = 1;
-        } else {
-          if (cpi->mse_source_denoised > 1000)
-            is_noisy = 1;
-        }
-        x->increase_denoising = 0;
-        if (!x->is_skin &&
-            x->best_sse_inter_mode == ZEROMV &&
-            (x->best_reference_frame == LAST_FRAME ||
-            x->best_reference_frame == cpi->closest_reference_frame) &&
-            cpi->consec_zero_last[block_index] >= 20 &&
-            is_noisy) {
-            x->increase_denoising = 1;
-        }
-        x->denoise_zeromv = 0;
-        vp8_denoiser_denoise_mb(&cpi->denoiser, x, best_sse, zero_mv_sse,
-                                recon_yoffset, recon_uvoffset,
-                                &cpi->common.lf_info, mb_row, mb_col,
-                                block_index,
-                                cpi->consec_zero_last_mvbias[block_index]);
+    // Reevaluate ZEROMV after denoising: for large noise content
+    // (i.e., cpi->mse_source_denoised is above threshold), do this for all
+    // blocks that did not pick ZEROMV as best mode but are using ZEROMV
+    // for denoising. Otherwise, always re-evaluate for blocks that picked
+    // INTRA mode as best mode.
+    // Avoid blocks that have been biased against ZERO_LAST
+    // (i.e., dot artifact candidate blocks).
+    reevaluate = (best_mbmode.ref_frame == INTRA_FRAME) ||
+                 (best_mbmode.mode != ZEROMV && x->denoise_zeromv &&
+                  cpi->mse_source_denoised > 2000);
+    if (!dot_artifact_candidate && reevaluate &&
+        x->best_zeromv_reference_frame != INTRA_FRAME) {
+      int this_rd = 0;
+      int this_ref_frame = x->best_zeromv_reference_frame;
+      rd_adjustment = 100;
+      rate2 =
+          x->ref_frame_cost[this_ref_frame] + vp8_cost_mv_ref(ZEROMV, mdcounts);
+      distortion2 = 0;
 
-        // Reevaluate ZEROMV after denoising: for large noise content
-        // (i.e., cpi->mse_source_denoised is above threshold), do this for all
-        // blocks that did not pick ZEROMV as best mode but are using ZEROMV
-        // for denoising. Otherwise, always re-evaluate for blocks that picked
-        // INTRA mode as best mode.
-        // Avoid blocks that have been biased against ZERO_LAST
-        // (i.e., dot artifact candidate blocks).
-        reevaluate = (best_mbmode.ref_frame == INTRA_FRAME) ||
-                     (best_mbmode.mode != ZEROMV &&
-                      x->denoise_zeromv &&
-                      cpi->mse_source_denoised > 2000);
-        if (!dot_artifact_candidate &&
-            reevaluate &&
-            x->best_zeromv_reference_frame != INTRA_FRAME)
-        {
-            int this_rd = 0;
-            int this_ref_frame = x->best_zeromv_reference_frame;
-            rd_adjustment = 100;
-            rate2 = x->ref_frame_cost[this_ref_frame] +
-                    vp8_cost_mv_ref(ZEROMV, mdcounts);
-            distortion2 = 0;
+      /* set up the proper prediction buffers for the frame */
+      x->e_mbd.mode_info_context->mbmi.ref_frame = this_ref_frame;
+      x->e_mbd.pre.y_buffer = plane[this_ref_frame][0];
+      x->e_mbd.pre.u_buffer = plane[this_ref_frame][1];
+      x->e_mbd.pre.v_buffer = plane[this_ref_frame][2];
 
-            /* set up the proper prediction buffers for the frame */
-            x->e_mbd.mode_info_context->mbmi.ref_frame = this_ref_frame;
-            x->e_mbd.pre.y_buffer = plane[this_ref_frame][0];
-            x->e_mbd.pre.u_buffer = plane[this_ref_frame][1];
-            x->e_mbd.pre.v_buffer = plane[this_ref_frame][2];
+      x->e_mbd.mode_info_context->mbmi.mode = ZEROMV;
+      x->e_mbd.mode_info_context->mbmi.uv_mode = DC_PRED;
+      x->e_mbd.mode_info_context->mbmi.mv.as_int = 0;
+      this_rd =
+          evaluate_inter_mode(&sse, rate2, &distortion2, cpi, x, rd_adjustment);
 
-            x->e_mbd.mode_info_context->mbmi.mode = ZEROMV;
-            x->e_mbd.mode_info_context->mbmi.uv_mode = DC_PRED;
-            x->e_mbd.mode_info_context->mbmi.mv.as_int = 0;
-            this_rd = evaluate_inter_mode(&sse, rate2, &distortion2, cpi, x,
-                                          rd_adjustment);
-
-            if (this_rd < best_rd)
-            {
-                memcpy(&best_mbmode, &x->e_mbd.mode_info_context->mbmi,
-                       sizeof(MB_MODE_INFO));
-            }
-        }
-
+      if (this_rd < best_rd) {
+        memcpy(&best_mbmode, &x->e_mbd.mode_info_context->mbmi,
+               sizeof(MB_MODE_INFO));
+      }
     }
+  }
 #endif
 
-    if (cpi->is_src_frame_alt_ref &&
-        (best_mbmode.mode != ZEROMV || best_mbmode.ref_frame != ALTREF_FRAME))
-    {
-        x->e_mbd.mode_info_context->mbmi.mode = ZEROMV;
-        x->e_mbd.mode_info_context->mbmi.ref_frame = ALTREF_FRAME;
-        x->e_mbd.mode_info_context->mbmi.mv.as_int = 0;
-        x->e_mbd.mode_info_context->mbmi.uv_mode = DC_PRED;
-        x->e_mbd.mode_info_context->mbmi.mb_skip_coeff =
-                                        (cpi->common.mb_no_coeff_skip);
-        x->e_mbd.mode_info_context->mbmi.partitioning = 0;
+  if (cpi->is_src_frame_alt_ref &&
+      (best_mbmode.mode != ZEROMV || best_mbmode.ref_frame != ALTREF_FRAME)) {
+    x->e_mbd.mode_info_context->mbmi.mode = ZEROMV;
+    x->e_mbd.mode_info_context->mbmi.ref_frame = ALTREF_FRAME;
+    x->e_mbd.mode_info_context->mbmi.mv.as_int = 0;
+    x->e_mbd.mode_info_context->mbmi.uv_mode = DC_PRED;
+    x->e_mbd.mode_info_context->mbmi.mb_skip_coeff =
+        (cpi->common.mb_no_coeff_skip);
+    x->e_mbd.mode_info_context->mbmi.partitioning = 0;
 
-        return;
-    }
+    return;
+  }
 
-    /* set to the best mb mode, this copy can be skip if x->skip since it
-     * already has the right content */
-    if (!x->skip)
-        memcpy(&x->e_mbd.mode_info_context->mbmi, &best_mbmode,
-               sizeof(MB_MODE_INFO));
+  /* set to the best mb mode, this copy can be skip if x->skip since it
+   * already has the right content */
+  if (!x->skip)
+    memcpy(&x->e_mbd.mode_info_context->mbmi, &best_mbmode,
+           sizeof(MB_MODE_INFO));
 
-    if (best_mbmode.mode <= B_PRED)
-    {
-        /* set mode_info_context->mbmi.uv_mode */
-        pick_intra_mbuv_mode(x);
-    }
+  if (best_mbmode.mode <= B_PRED) {
+    /* set mode_info_context->mbmi.uv_mode */
+    pick_intra_mbuv_mode(x);
+  }
 
-    if (sign_bias
-      != cpi->common.ref_frame_sign_bias[xd->mode_info_context->mbmi.ref_frame])
-        best_ref_mv.as_int = best_ref_mv_sb[!sign_bias].as_int;
+  if (sign_bias !=
+      cpi->common.ref_frame_sign_bias[xd->mode_info_context->mbmi.ref_frame])
+    best_ref_mv.as_int = best_ref_mv_sb[!sign_bias].as_int;
 
-    update_mvcount(x, &best_ref_mv);
+  update_mvcount(x, &best_ref_mv);
 }
 
-void vp8_pick_intra_mode(MACROBLOCK *x, int *rate_)
-{
-    int error4x4, error16x16 = INT_MAX;
-    int rate, best_rate = 0, distortion, best_sse;
-    MB_PREDICTION_MODE mode, best_mode = DC_PRED;
-    int this_rd;
-    unsigned int sse;
-    BLOCK *b = &x->block[0];
-    MACROBLOCKD *xd = &x->e_mbd;
+void vp8_pick_intra_mode(MACROBLOCK *x, int *rate_) {
+  int error4x4, error16x16 = INT_MAX;
+  int rate, best_rate = 0, distortion, best_sse;
+  MB_PREDICTION_MODE mode, best_mode = DC_PRED;
+  int this_rd;
+  unsigned int sse;
+  BLOCK *b = &x->block[0];
+  MACROBLOCKD *xd = &x->e_mbd;
 
-    xd->mode_info_context->mbmi.ref_frame = INTRA_FRAME;
+  xd->mode_info_context->mbmi.ref_frame = INTRA_FRAME;
 
-    pick_intra_mbuv_mode(x);
+  pick_intra_mbuv_mode(x);
 
-    for (mode = DC_PRED; mode <= TM_PRED; mode ++)
-    {
-        xd->mode_info_context->mbmi.mode = mode;
-        vp8_build_intra_predictors_mby_s(xd,
-                                         xd->dst.y_buffer - xd->dst.y_stride,
-                                         xd->dst.y_buffer - 1,
-                                         xd->dst.y_stride,
-                                         xd->predictor,
-                                         16);
-        distortion = vpx_variance16x16
-            (*(b->base_src), b->src_stride, xd->predictor, 16, &sse);
-        rate = x->mbmode_cost[xd->frame_type][mode];
-        this_rd = RDCOST(x->rdmult, x->rddiv, rate, distortion);
+  for (mode = DC_PRED; mode <= TM_PRED; mode++) {
+    xd->mode_info_context->mbmi.mode = mode;
+    vp8_build_intra_predictors_mby_s(xd, xd->dst.y_buffer - xd->dst.y_stride,
+                                     xd->dst.y_buffer - 1, xd->dst.y_stride,
+                                     xd->predictor, 16);
+    distortion = vpx_variance16x16(*(b->base_src), b->src_stride, xd->predictor,
+                                   16, &sse);
+    rate = x->mbmode_cost[xd->frame_type][mode];
+    this_rd = RDCOST(x->rdmult, x->rddiv, rate, distortion);
 
-        if (error16x16 > this_rd)
-        {
-            error16x16 = this_rd;
-            best_mode = mode;
-            best_sse = sse;
-            best_rate = rate;
-        }
+    if (error16x16 > this_rd) {
+      error16x16 = this_rd;
+      best_mode = mode;
+      best_sse = sse;
+      best_rate = rate;
     }
-    xd->mode_info_context->mbmi.mode = best_mode;
+  }
+  xd->mode_info_context->mbmi.mode = best_mode;
 
-    error4x4 = pick_intra4x4mby_modes(x, &rate,
-                                      &best_sse);
-    if (error4x4 < error16x16)
-    {
-        xd->mode_info_context->mbmi.mode = B_PRED;
-        best_rate = rate;
-    }
+  error4x4 = pick_intra4x4mby_modes(x, &rate, &best_sse);
+  if (error4x4 < error16x16) {
+    xd->mode_info_context->mbmi.mode = B_PRED;
+    best_rate = rate;
+  }
 
-    *rate_ = best_rate;
+  *rate_ = best_rate;
 }
--- a/vp8/encoder/pickinter.h
+++ b/vp8/encoder/pickinter.h
@@ -8,7 +8,6 @@
  *  be found in the AUTHORS file in the root of the source tree.
  */
 
-
 #ifndef VP8_ENCODER_PICKINTER_H_
 #define VP8_ENCODER_PICKINTER_H_
 #include "vpx_config.h"
@@ -26,8 +25,7 @@
 
 extern int vp8_get_inter_mbpred_error(MACROBLOCK *mb,
                                       const vp8_variance_fn_ptr_t *vfp,
-                                      unsigned int *sse,
-                                      int_mv this_mv);
+                                      unsigned int *sse, int_mv this_mv);
 #ifdef __cplusplus
 }  // extern "C"
 #endif
--- a/vp8/encoder/picklpf.c
+++ b/vp8/encoder/picklpf.c
@@ -8,7 +8,6 @@
  *  be found in the AUTHORS file in the root of the source tree.
  */
 
-
 #include "./vpx_dsp_rtcd.h"
 #include "./vpx_scale_rtcd.h"
 #include "vp8/common/onyxc_int.h"
@@ -22,386 +21,360 @@
 #include "vpx_ports/arm.h"
 #endif
 
-extern int vp8_calc_ss_err(YV12_BUFFER_CONFIG *source, YV12_BUFFER_CONFIG *dest);
+extern int vp8_calc_ss_err(YV12_BUFFER_CONFIG *source,
+                           YV12_BUFFER_CONFIG *dest);
 
 static void yv12_copy_partial_frame(YV12_BUFFER_CONFIG *src_ybc,
-                                    YV12_BUFFER_CONFIG *dst_ybc)
-{
-    unsigned char *src_y, *dst_y;
-    int yheight;
-    int ystride;
-    int yoffset;
-    int linestocopy;
+                                    YV12_BUFFER_CONFIG *dst_ybc) {
+  unsigned char *src_y, *dst_y;
+  int yheight;
+  int ystride;
+  int yoffset;
+  int linestocopy;
 
-    yheight  = src_ybc->y_height;
-    ystride  = src_ybc->y_stride;
+  yheight = src_ybc->y_height;
+  ystride = src_ybc->y_stride;
 
-    /* number of MB rows to use in partial filtering */
-    linestocopy = (yheight >> 4) / PARTIAL_FRAME_FRACTION;
-    linestocopy = linestocopy ? linestocopy << 4 : 16;     /* 16 lines per MB */
+  /* number of MB rows to use in partial filtering */
+  linestocopy = (yheight >> 4) / PARTIAL_FRAME_FRACTION;
+  linestocopy = linestocopy ? linestocopy << 4 : 16; /* 16 lines per MB */
 
-    /* Copy extra 4 so that full filter context is available if filtering done
-     * on the copied partial frame and not original. Partial filter does mb
-     * filtering for top row also, which can modify3 pixels above.
-     */
-    linestocopy += 4;
-    /* partial image starts at ~middle of frame (macroblock border)*/
-    yoffset  = ystride * (((yheight >> 5) * 16) - 4);
-    src_y = src_ybc->y_buffer + yoffset;
-    dst_y = dst_ybc->y_buffer + yoffset;
+  /* Copy extra 4 so that full filter context is available if filtering done
+   * on the copied partial frame and not original. Partial filter does mb
+   * filtering for top row also, which can modify3 pixels above.
+   */
+  linestocopy += 4;
+  /* partial image starts at ~middle of frame (macroblock border)*/
+  yoffset = ystride * (((yheight >> 5) * 16) - 4);
+  src_y = src_ybc->y_buffer + yoffset;
+  dst_y = dst_ybc->y_buffer + yoffset;
 
-    memcpy(dst_y, src_y, ystride * linestocopy);
+  memcpy(dst_y, src_y, ystride * linestocopy);
 }
 
 static int calc_partial_ssl_err(YV12_BUFFER_CONFIG *source,
-                                YV12_BUFFER_CONFIG *dest)
-{
-    int i, j;
-    int Total = 0;
-    int srcoffset, dstoffset;
-    unsigned char *src = source->y_buffer;
-    unsigned char *dst = dest->y_buffer;
+                                YV12_BUFFER_CONFIG *dest) {
+  int i, j;
+  int Total = 0;
+  int srcoffset, dstoffset;
+  unsigned char *src = source->y_buffer;
+  unsigned char *dst = dest->y_buffer;
 
-    int linestocopy;
+  int linestocopy;
 
-    /* number of MB rows to use in partial filtering */
-    linestocopy = (source->y_height >> 4) / PARTIAL_FRAME_FRACTION;
-    linestocopy = linestocopy ? linestocopy << 4 : 16;     /* 16 lines per MB */
+  /* number of MB rows to use in partial filtering */
+  linestocopy = (source->y_height >> 4) / PARTIAL_FRAME_FRACTION;
+  linestocopy = linestocopy ? linestocopy << 4 : 16; /* 16 lines per MB */
 
+  /* partial image starts at ~middle of frame (macroblock border)*/
+  srcoffset = source->y_stride * ((dest->y_height >> 5) * 16);
+  dstoffset = dest->y_stride * ((dest->y_height >> 5) * 16);
 
-    /* partial image starts at ~middle of frame (macroblock border)*/
-    srcoffset = source->y_stride * ((dest->y_height >> 5) * 16);
-    dstoffset = dest->y_stride   * ((dest->y_height >> 5) * 16);
+  src += srcoffset;
+  dst += dstoffset;
 
-    src += srcoffset;
-    dst += dstoffset;
-
-    /* Loop through the Y plane raw and reconstruction data summing
-     * (square differences)
-     */
-    for (i = 0; i < linestocopy; i += 16)
-    {
-        for (j = 0; j < source->y_width; j += 16)
-        {
-            unsigned int sse;
-            Total += vpx_mse16x16(src + j, source->y_stride,
-                                                     dst + j, dest->y_stride,
-                                                     &sse);
-        }
-
-        src += 16 * source->y_stride;
-        dst += 16 * dest->y_stride;
+  /* Loop through the Y plane raw and reconstruction data summing
+   * (square differences)
+   */
+  for (i = 0; i < linestocopy; i += 16) {
+    for (j = 0; j < source->y_width; j += 16) {
+      unsigned int sse;
+      Total += vpx_mse16x16(src + j, source->y_stride, dst + j, dest->y_stride,
+                            &sse);
     }
 
-    return Total;
+    src += 16 * source->y_stride;
+    dst += 16 * dest->y_stride;
+  }
+
+  return Total;
 }
 
 /* Enforce a minimum filter level based upon baseline Q */
-static int get_min_filter_level(VP8_COMP *cpi, int base_qindex)
-{
-    int min_filter_level;
+static int get_min_filter_level(VP8_COMP *cpi, int base_qindex) {
+  int min_filter_level;
 
-    if (cpi->source_alt_ref_active && cpi->common.refresh_golden_frame &&
-        !cpi->common.refresh_alt_ref_frame)
-        min_filter_level = 0;
+  if (cpi->source_alt_ref_active && cpi->common.refresh_golden_frame &&
+      !cpi->common.refresh_alt_ref_frame)
+    min_filter_level = 0;
+  else {
+    if (base_qindex <= 6)
+      min_filter_level = 0;
+    else if (base_qindex <= 16)
+      min_filter_level = 1;
     else
-    {
-        if (base_qindex <= 6)
-            min_filter_level = 0;
-        else if (base_qindex <= 16)
-            min_filter_level = 1;
-        else
-            min_filter_level = (base_qindex / 8);
-    }
+      min_filter_level = (base_qindex / 8);
+  }
 
-    return min_filter_level;
+  return min_filter_level;
 }
 
 /* Enforce a maximum filter level based upon baseline Q */
-static int get_max_filter_level(VP8_COMP *cpi, int base_qindex)
-{
-    /* PGW August 2006: Highest filter values almost always a bad idea */
+static int get_max_filter_level(VP8_COMP *cpi, int base_qindex) {
+  /* PGW August 2006: Highest filter values almost always a bad idea */
 
-    /* jbb chg: 20100118 - not so any more with this overquant stuff allow
-     * high values with lots of intra coming in.
-     */
-    int max_filter_level = MAX_LOOP_FILTER;
-    (void)base_qindex;
+  /* jbb chg: 20100118 - not so any more with this overquant stuff allow
+   * high values with lots of intra coming in.
+   */
+  int max_filter_level = MAX_LOOP_FILTER;
+  (void)base_qindex;
 
-    if (cpi->twopass.section_intra_rating > 8)
-        max_filter_level = MAX_LOOP_FILTER * 3 / 4;
+  if (cpi->twopass.section_intra_rating > 8)
+    max_filter_level = MAX_LOOP_FILTER * 3 / 4;
 
-    return max_filter_level;
+  return max_filter_level;
 }
 
-void vp8cx_pick_filter_level_fast(YV12_BUFFER_CONFIG *sd, VP8_COMP *cpi)
-{
-    VP8_COMMON *cm = &cpi->common;
+void vp8cx_pick_filter_level_fast(YV12_BUFFER_CONFIG *sd, VP8_COMP *cpi) {
+  VP8_COMMON *cm = &cpi->common;
 
-    int best_err = 0;
-    int filt_err = 0;
-    int min_filter_level = get_min_filter_level(cpi, cm->base_qindex);
-    int max_filter_level = get_max_filter_level(cpi, cm->base_qindex);
-    int filt_val;
-    int best_filt_val;
-    YV12_BUFFER_CONFIG * saved_frame = cm->frame_to_show;
+  int best_err = 0;
+  int filt_err = 0;
+  int min_filter_level = get_min_filter_level(cpi, cm->base_qindex);
+  int max_filter_level = get_max_filter_level(cpi, cm->base_qindex);
+  int filt_val;
+  int best_filt_val;
+  YV12_BUFFER_CONFIG *saved_frame = cm->frame_to_show;
 
-    /* Replace unfiltered frame buffer with a new one */
-    cm->frame_to_show = &cpi->pick_lf_lvl_frame;
+  /* Replace unfiltered frame buffer with a new one */
+  cm->frame_to_show = &cpi->pick_lf_lvl_frame;
 
-    if (cm->frame_type == KEY_FRAME)
-        cm->sharpness_level = 0;
-    else
-        cm->sharpness_level = cpi->oxcf.Sharpness;
+  if (cm->frame_type == KEY_FRAME)
+    cm->sharpness_level = 0;
+  else
+    cm->sharpness_level = cpi->oxcf.Sharpness;
 
-    if (cm->sharpness_level != cm->last_sharpness_level)
-    {
-        vp8_loop_filter_update_sharpness(&cm->lf_info, cm->sharpness_level);
-        cm->last_sharpness_level = cm->sharpness_level;
-    }
+  if (cm->sharpness_level != cm->last_sharpness_level) {
+    vp8_loop_filter_update_sharpness(&cm->lf_info, cm->sharpness_level);
+    cm->last_sharpness_level = cm->sharpness_level;
+  }
 
-    /* Start the search at the previous frame filter level unless it is
-     * now out of range.
-     */
-    if (cm->filter_level < min_filter_level)
-        cm->filter_level = min_filter_level;
-    else if (cm->filter_level > max_filter_level)
-        cm->filter_level = max_filter_level;
+  /* Start the search at the previous frame filter level unless it is
+   * now out of range.
+   */
+  if (cm->filter_level < min_filter_level)
+    cm->filter_level = min_filter_level;
+  else if (cm->filter_level > max_filter_level)
+    cm->filter_level = max_filter_level;
 
-    filt_val = cm->filter_level;
-    best_filt_val = filt_val;
+  filt_val = cm->filter_level;
+  best_filt_val = filt_val;
 
-    /* Get the err using the previous frame's filter value. */
+  /* Get the err using the previous frame's filter value. */
 
-    /* Copy the unfiltered / processed recon buffer to the new buffer */
-    yv12_copy_partial_frame(saved_frame, cm->frame_to_show);
-    vp8_loop_filter_partial_frame(cm, &cpi->mb.e_mbd, filt_val);
+  /* Copy the unfiltered / processed recon buffer to the new buffer */
+  yv12_copy_partial_frame(saved_frame, cm->frame_to_show);
+  vp8_loop_filter_partial_frame(cm, &cpi->mb.e_mbd, filt_val);
 
-    best_err = calc_partial_ssl_err(sd, cm->frame_to_show);
+  best_err = calc_partial_ssl_err(sd, cm->frame_to_show);
 
-    filt_val -= 1 + (filt_val > 10);
+  filt_val -= 1 + (filt_val > 10);
 
-    /* Search lower filter levels */
-    while (filt_val >= min_filter_level)
-    {
-        /* Apply the loop filter */
-        yv12_copy_partial_frame(saved_frame, cm->frame_to_show);
-        vp8_loop_filter_partial_frame(cm, &cpi->mb.e_mbd, filt_val);
+  /* Search lower filter levels */
+  while (filt_val >= min_filter_level) {
+    /* Apply the loop filter */
+    yv12_copy_partial_frame(saved_frame, cm->frame_to_show);
+    vp8_loop_filter_partial_frame(cm, &cpi->mb.e_mbd, filt_val);
 
-        /* Get the err for filtered frame */
-        filt_err = calc_partial_ssl_err(sd, cm->frame_to_show);
+    /* Get the err for filtered frame */
+    filt_err = calc_partial_ssl_err(sd, cm->frame_to_show);
 
-        /* Update the best case record or exit loop. */
-        if (filt_err < best_err)
-        {
-            best_err = filt_err;
-            best_filt_val = filt_val;
-        }
-        else
-            break;
+    /* Update the best case record or exit loop. */
+    if (filt_err < best_err) {
+      best_err = filt_err;
+      best_filt_val = filt_val;
+    } else
+      break;
 
-        /* Adjust filter level */
-        filt_val -= 1 + (filt_val > 10);
-    }
+    /* Adjust filter level */
+    filt_val -= 1 + (filt_val > 10);
+  }
 
-    /* Search up (note that we have already done filt_val = cm->filter_level) */
-    filt_val = cm->filter_level + 1 + (filt_val > 10);
+  /* Search up (note that we have already done filt_val = cm->filter_level) */
+  filt_val = cm->filter_level + 1 + (filt_val > 10);
 
-    if (best_filt_val == cm->filter_level)
-    {
-        /* Resist raising filter level for very small gains */
-        best_err -= (best_err >> 10);
+  if (best_filt_val == cm->filter_level) {
+    /* Resist raising filter level for very small gains */
+    best_err -= (best_err >> 10);
 
-        while (filt_val < max_filter_level)
-        {
-            /* Apply the loop filter */
-            yv12_copy_partial_frame(saved_frame, cm->frame_to_show);
+    while (filt_val < max_filter_level) {
+      /* Apply the loop filter */
+      yv12_copy_partial_frame(saved_frame, cm->frame_to_show);
 
-            vp8_loop_filter_partial_frame(cm, &cpi->mb.e_mbd, filt_val);
+      vp8_loop_filter_partial_frame(cm, &cpi->mb.e_mbd, filt_val);
 
-            /* Get the err for filtered frame */
-            filt_err = calc_partial_ssl_err(sd, cm->frame_to_show);
+      /* Get the err for filtered frame */
+      filt_err = calc_partial_ssl_err(sd, cm->frame_to_show);
 
-            /* Update the best case record or exit loop. */
-            if (filt_err < best_err)
-            {
-                /* Do not raise filter level if improvement is < 1 part
-                 * in 4096
-                 */
-                best_err = filt_err - (filt_err >> 10);
+      /* Update the best case record or exit loop. */
+      if (filt_err < best_err) {
+        /* Do not raise filter level if improvement is < 1 part
+         * in 4096
+         */
+        best_err = filt_err - (filt_err >> 10);
 
-                best_filt_val = filt_val;
-            }
-            else
-                break;
+        best_filt_val = filt_val;
+      } else
+        break;
 
-            /* Adjust filter level */
-            filt_val += 1 + (filt_val > 10);
-        }
+      /* Adjust filter level */
+      filt_val += 1 + (filt_val > 10);
     }
+  }
 
-    cm->filter_level = best_filt_val;
+  cm->filter_level = best_filt_val;
 
-    if (cm->filter_level < min_filter_level)
-        cm->filter_level = min_filter_level;
+  if (cm->filter_level < min_filter_level) cm->filter_level = min_filter_level;
 
-    if (cm->filter_level > max_filter_level)
-        cm->filter_level = max_filter_level;
+  if (cm->filter_level > max_filter_level) cm->filter_level = max_filter_level;
 
-    /* restore unfiltered frame pointer */
-    cm->frame_to_show = saved_frame;
+  /* restore unfiltered frame pointer */
+  cm->frame_to_show = saved_frame;
 }
 
 /* Stub function for now Alt LF not used */
-void vp8cx_set_alt_lf_level(VP8_COMP *cpi, int filt_val)
-{
-    MACROBLOCKD *mbd = &cpi->mb.e_mbd;
-    (void) filt_val;
+void vp8cx_set_alt_lf_level(VP8_COMP *cpi, int filt_val) {
+  MACROBLOCKD *mbd = &cpi->mb.e_mbd;
+  (void)filt_val;
 
-    mbd->segment_feature_data[MB_LVL_ALT_LF][0] = cpi->segment_feature_data[MB_LVL_ALT_LF][0];
-    mbd->segment_feature_data[MB_LVL_ALT_LF][1] = cpi->segment_feature_data[MB_LVL_ALT_LF][1];
-    mbd->segment_feature_data[MB_LVL_ALT_LF][2] = cpi->segment_feature_data[MB_LVL_ALT_LF][2];
-    mbd->segment_feature_data[MB_LVL_ALT_LF][3] = cpi->segment_feature_data[MB_LVL_ALT_LF][3];
+  mbd->segment_feature_data[MB_LVL_ALT_LF][0] =
+      cpi->segment_feature_data[MB_LVL_ALT_LF][0];
+  mbd->segment_feature_data[MB_LVL_ALT_LF][1] =
+      cpi->segment_feature_data[MB_LVL_ALT_LF][1];
+  mbd->segment_feature_data[MB_LVL_ALT_LF][2] =
+      cpi->segment_feature_data[MB_LVL_ALT_LF][2];
+  mbd->segment_feature_data[MB_LVL_ALT_LF][3] =
+      cpi->segment_feature_data[MB_LVL_ALT_LF][3];
 }
 
-void vp8cx_pick_filter_level(YV12_BUFFER_CONFIG *sd, VP8_COMP *cpi)
-{
-    VP8_COMMON *cm = &cpi->common;
+void vp8cx_pick_filter_level(YV12_BUFFER_CONFIG *sd, VP8_COMP *cpi) {
+  VP8_COMMON *cm = &cpi->common;
 
-    int best_err = 0;
-    int filt_err = 0;
-    int min_filter_level = get_min_filter_level(cpi, cm->base_qindex);
-    int max_filter_level = get_max_filter_level(cpi, cm->base_qindex);
+  int best_err = 0;
+  int filt_err = 0;
+  int min_filter_level = get_min_filter_level(cpi, cm->base_qindex);
+  int max_filter_level = get_max_filter_level(cpi, cm->base_qindex);
 
-    int filter_step;
-    int filt_high = 0;
-    int filt_mid;
-    int filt_low = 0;
-    int filt_best;
-    int filt_direction = 0;
+  int filter_step;
+  int filt_high = 0;
+  int filt_mid;
+  int filt_low = 0;
+  int filt_best;
+  int filt_direction = 0;
 
-    /* Bias against raising loop filter and in favor of lowering it */
-    int Bias = 0;
+  /* Bias against raising loop filter and in favor of lowering it */
+  int Bias = 0;
 
-    int ss_err[MAX_LOOP_FILTER + 1];
+  int ss_err[MAX_LOOP_FILTER + 1];
 
-    YV12_BUFFER_CONFIG * saved_frame = cm->frame_to_show;
+  YV12_BUFFER_CONFIG *saved_frame = cm->frame_to_show;
 
-    memset(ss_err, 0, sizeof(ss_err));
+  memset(ss_err, 0, sizeof(ss_err));
 
-    /* Replace unfiltered frame buffer with a new one */
-    cm->frame_to_show = &cpi->pick_lf_lvl_frame;
+  /* Replace unfiltered frame buffer with a new one */
+  cm->frame_to_show = &cpi->pick_lf_lvl_frame;
 
-    if (cm->frame_type == KEY_FRAME)
-        cm->sharpness_level = 0;
-    else
-        cm->sharpness_level = cpi->oxcf.Sharpness;
+  if (cm->frame_type == KEY_FRAME)
+    cm->sharpness_level = 0;
+  else
+    cm->sharpness_level = cpi->oxcf.Sharpness;
 
-    /* Start the search at the previous frame filter level unless it is
-     * now out of range.
-     */
-    filt_mid = cm->filter_level;
+  /* Start the search at the previous frame filter level unless it is
+   * now out of range.
+   */
+  filt_mid = cm->filter_level;
 
-    if (filt_mid < min_filter_level)
-        filt_mid = min_filter_level;
-    else if (filt_mid > max_filter_level)
-        filt_mid = max_filter_level;
+  if (filt_mid < min_filter_level)
+    filt_mid = min_filter_level;
+  else if (filt_mid > max_filter_level)
+    filt_mid = max_filter_level;
 
-    /* Define the initial step size */
-    filter_step = (filt_mid < 16) ? 4 : filt_mid / 4;
+  /* Define the initial step size */
+  filter_step = (filt_mid < 16) ? 4 : filt_mid / 4;
 
-    /* Get baseline error score */
+  /* Get baseline error score */
 
-    /* Copy the unfiltered / processed recon buffer to the new buffer */
-    vpx_yv12_copy_y(saved_frame, cm->frame_to_show);
+  /* Copy the unfiltered / processed recon buffer to the new buffer */
+  vpx_yv12_copy_y(saved_frame, cm->frame_to_show);
 
-    vp8cx_set_alt_lf_level(cpi, filt_mid);
-    vp8_loop_filter_frame_yonly(cm, &cpi->mb.e_mbd, filt_mid);
+  vp8cx_set_alt_lf_level(cpi, filt_mid);
+  vp8_loop_filter_frame_yonly(cm, &cpi->mb.e_mbd, filt_mid);
 
-    best_err = vp8_calc_ss_err(sd, cm->frame_to_show);
+  best_err = vp8_calc_ss_err(sd, cm->frame_to_show);
 
-    ss_err[filt_mid] = best_err;
+  ss_err[filt_mid] = best_err;
 
-    filt_best = filt_mid;
+  filt_best = filt_mid;
 
-    while (filter_step > 0)
-    {
-        Bias = (best_err >> (15 - (filt_mid / 8))) * filter_step;
+  while (filter_step > 0) {
+    Bias = (best_err >> (15 - (filt_mid / 8))) * filter_step;
 
-        if (cpi->twopass.section_intra_rating < 20)
-            Bias = Bias * cpi->twopass.section_intra_rating / 20;
+    if (cpi->twopass.section_intra_rating < 20)
+      Bias = Bias * cpi->twopass.section_intra_rating / 20;
 
-        filt_high = ((filt_mid + filter_step) > max_filter_level) ? max_filter_level : (filt_mid + filter_step);
-        filt_low = ((filt_mid - filter_step) < min_filter_level) ? min_filter_level : (filt_mid - filter_step);
+    filt_high = ((filt_mid + filter_step) > max_filter_level)
+                    ? max_filter_level
+                    : (filt_mid + filter_step);
+    filt_low = ((filt_mid - filter_step) < min_filter_level)
+                   ? min_filter_level
+                   : (filt_mid - filter_step);
 
-        if ((filt_direction <= 0) && (filt_low != filt_mid))
-        {
-            if(ss_err[filt_low] == 0)
-            {
-                /* Get Low filter error score */
-                vpx_yv12_copy_y(saved_frame, cm->frame_to_show);
-                vp8cx_set_alt_lf_level(cpi, filt_low);
-                vp8_loop_filter_frame_yonly(cm, &cpi->mb.e_mbd, filt_low);
+    if ((filt_direction <= 0) && (filt_low != filt_mid)) {
+      if (ss_err[filt_low] == 0) {
+        /* Get Low filter error score */
+        vpx_yv12_copy_y(saved_frame, cm->frame_to_show);
+        vp8cx_set_alt_lf_level(cpi, filt_low);
+        vp8_loop_filter_frame_yonly(cm, &cpi->mb.e_mbd, filt_low);
 
-                filt_err = vp8_calc_ss_err(sd, cm->frame_to_show);
-                ss_err[filt_low] = filt_err;
-            }
-            else
-                filt_err = ss_err[filt_low];
+        filt_err = vp8_calc_ss_err(sd, cm->frame_to_show);
+        ss_err[filt_low] = filt_err;
+      } else
+        filt_err = ss_err[filt_low];
 
-            /* If value is close to the best so far then bias towards a
-             * lower loop filter value.
-             */
-            if ((filt_err - Bias) < best_err)
-            {
-                /* Was it actually better than the previous best? */
-                if (filt_err < best_err)
-                    best_err = filt_err;
+      /* If value is close to the best so far then bias towards a
+       * lower loop filter value.
+       */
+      if ((filt_err - Bias) < best_err) {
+        /* Was it actually better than the previous best? */
+        if (filt_err < best_err) best_err = filt_err;
 
-                filt_best = filt_low;
-            }
-        }
+        filt_best = filt_low;
+      }
+    }
 
-        /* Now look at filt_high */
-        if ((filt_direction >= 0) && (filt_high != filt_mid))
-        {
-            if(ss_err[filt_high] == 0)
-            {
-                vpx_yv12_copy_y(saved_frame, cm->frame_to_show);
-                vp8cx_set_alt_lf_level(cpi, filt_high);
-                vp8_loop_filter_frame_yonly(cm, &cpi->mb.e_mbd, filt_high);
+    /* Now look at filt_high */
+    if ((filt_direction >= 0) && (filt_high != filt_mid)) {
+      if (ss_err[filt_high] == 0) {
+        vpx_yv12_copy_y(saved_frame, cm->frame_to_show);
+        vp8cx_set_alt_lf_level(cpi, filt_high);
+        vp8_loop_filter_frame_yonly(cm, &cpi->mb.e_mbd, filt_high);
 
-                filt_err = vp8_calc_ss_err(sd, cm->frame_to_show);
-                ss_err[filt_high] = filt_err;
-            }
-            else
-                filt_err = ss_err[filt_high];
+        filt_err = vp8_calc_ss_err(sd, cm->frame_to_show);
+        ss_err[filt_high] = filt_err;
+      } else
+        filt_err = ss_err[filt_high];
 
-            /* Was it better than the previous best? */
-            if (filt_err < (best_err - Bias))
-            {
-                best_err = filt_err;
-                filt_best = filt_high;
-            }
-        }
+      /* Was it better than the previous best? */
+      if (filt_err < (best_err - Bias)) {
+        best_err = filt_err;
+        filt_best = filt_high;
+      }
+    }
 
-        /* Half the step distance if the best filter value was the same
-         * as last time
-         */
-        if (filt_best == filt_mid)
-        {
-            filter_step = filter_step / 2;
-            filt_direction = 0;
-        }
-        else
-        {
-            filt_direction = (filt_best < filt_mid) ? -1 : 1;
-            filt_mid = filt_best;
-        }
+    /* Half the step distance if the best filter value was the same
+     * as last time
+     */
+    if (filt_best == filt_mid) {
+      filter_step = filter_step / 2;
+      filt_direction = 0;
+    } else {
+      filt_direction = (filt_best < filt_mid) ? -1 : 1;
+      filt_mid = filt_best;
     }
+  }
 
-    cm->filter_level = filt_best;
+  cm->filter_level = filt_best;
 
-    /* restore unfiltered frame pointer */
-    cm->frame_to_show = saved_frame;
+  /* restore unfiltered frame pointer */
+  cm->frame_to_show = saved_frame;
 }
--- a/vp8/encoder/quantize.h
+++ b/vp8/encoder/quantize.h
@@ -8,7 +8,6 @@
  *  be found in the AUTHORS file in the root of the source tree.
  */
 
-
 #ifndef VP8_ENCODER_QUANTIZE_H_
 #define VP8_ENCODER_QUANTIZE_H_
 
@@ -24,7 +23,8 @@
 extern void vp8_set_quantizer(struct VP8_COMP *cpi, int Q);
 extern void vp8cx_frame_init_quantizer(struct VP8_COMP *cpi);
 extern void vp8_update_zbin_extra(struct VP8_COMP *cpi, struct macroblock *x);
-extern void vp8cx_mb_init_quantizer(struct VP8_COMP *cpi, struct macroblock *x, int ok_to_skip);
+extern void vp8cx_mb_init_quantizer(struct VP8_COMP *cpi, struct macroblock *x,
+                                    int ok_to_skip);
 extern void vp8cx_init_quantizer(struct VP8_COMP *cpi);
 
 #ifdef __cplusplus
--- a/vp8/encoder/ratectrl.c
+++ b/vp8/encoder/ratectrl.c
@@ -8,7 +8,6 @@
  *  be found in the AUTHORS file in the root of the source tree.
  */
 
-
 #include <stdlib.h>
 #include <stdio.h>
 #include <string.h>
@@ -24,14 +23,11 @@
 #include "encodemv.h"
 #include "vpx_dsp/vpx_dsp_common.h"
 
+#define MIN_BPB_FACTOR 0.01
+#define MAX_BPB_FACTOR 50
 
-#define MIN_BPB_FACTOR          0.01
-#define MAX_BPB_FACTOR          50
-
 extern const MB_PREDICTION_MODE vp8_mode_order[MAX_MODES];
 
-
-
 #ifdef MODE_STATS
 extern int y_modes[5];
 extern int uv_modes[4];
@@ -43,94 +39,73 @@
 #endif
 
 /* Bits Per MB at different Q (Multiplied by 512) */
-#define BPER_MB_NORMBITS    9
+#define BPER_MB_NORMBITS 9
 
 /* Work in progress recalibration of baseline rate tables based on
  * the assumption that bits per mb is inversely proportional to the
  * quantizer value.
  */
-const int vp8_bits_per_mb[2][QINDEX_RANGE] =
-{
-    /* Intra case 450000/Qintra */
-    {
-        1125000,900000, 750000, 642857, 562500, 500000, 450000, 450000,
-        409090, 375000, 346153, 321428, 300000, 281250, 264705, 264705,
-        250000, 236842, 225000, 225000, 214285, 214285, 204545, 204545,
-        195652, 195652, 187500, 180000, 180000, 173076, 166666, 160714,
-        155172, 150000, 145161, 140625, 136363, 132352, 128571, 125000,
-        121621, 121621, 118421, 115384, 112500, 109756, 107142, 104651,
-        102272, 100000, 97826,  97826,  95744,  93750,  91836,  90000,
-        88235,  86538,  84905,  83333,  81818,  80357,  78947,  77586,
-        76271,  75000,  73770,  72580,  71428,  70312,  69230,  68181,
-        67164,  66176,  65217,  64285,  63380,  62500,  61643,  60810,
-        60000,  59210,  59210,  58441,  57692,  56962,  56250,  55555,
-        54878,  54216,  53571,  52941,  52325,  51724,  51136,  50561,
-        49450,  48387,  47368,  46875,  45918,  45000,  44554,  44117,
-        43269,  42452,  41666,  40909,  40178,  39473,  38793,  38135,
-        36885,  36290,  35714,  35156,  34615,  34090,  33582,  33088,
-        32608,  32142,  31468,  31034,  30405,  29801,  29220,  28662,
-    },
-    /* Inter case 285000/Qinter */
-    {
-        712500, 570000, 475000, 407142, 356250, 316666, 285000, 259090,
-        237500, 219230, 203571, 190000, 178125, 167647, 158333, 150000,
-        142500, 135714, 129545, 123913, 118750, 114000, 109615, 105555,
-        101785, 98275,  95000,  91935,  89062,  86363,  83823,  81428,
-        79166,  77027,  75000,  73076,  71250,  69512,  67857,  66279,
-        64772,  63333,  61956,  60638,  59375,  58163,  57000,  55882,
-        54807,  53773,  52777,  51818,  50892,  50000,  49137,  47500,
-        45967,  44531,  43181,  41911,  40714,  39583,  38513,  37500,
-        36538,  35625,  34756,  33928,  33139,  32386,  31666,  30978,
-        30319,  29687,  29081,  28500,  27941,  27403,  26886,  26388,
-        25909,  25446,  25000,  24568,  23949,  23360,  22800,  22265,
-        21755,  21268,  20802,  20357,  19930,  19520,  19127,  18750,
-        18387,  18037,  17701,  17378,  17065,  16764,  16473,  16101,
-        15745,  15405,  15079,  14766,  14467,  14179,  13902,  13636,
-        13380,  13133,  12895,  12666,  12445,  12179,  11924,  11632,
-        11445,  11220,  11003,  10795,  10594,  10401,  10215,  10035,
-    }
+const int vp8_bits_per_mb[2][QINDEX_RANGE] = {
+  /* Intra case 450000/Qintra */
+  {
+      1125000, 900000, 750000, 642857, 562500, 500000, 450000, 450000, 409090,
+      375000, 346153, 321428, 300000, 281250, 264705, 264705, 250000, 236842,
+      225000, 225000, 214285, 214285, 204545, 204545, 195652, 195652, 187500,
+      180000, 180000, 173076, 166666, 160714, 155172, 150000, 145161, 140625,
+      136363, 132352, 128571, 125000, 121621, 121621, 118421, 115384, 112500,
+      109756, 107142, 104651, 102272, 100000, 97826, 97826, 95744, 93750, 91836,
+      90000, 88235, 86538, 84905, 83333, 81818, 80357, 78947, 77586, 76271,
+      75000, 73770, 72580, 71428, 70312, 69230, 68181, 67164, 66176, 65217,
+      64285, 63380, 62500, 61643, 60810, 60000, 59210, 59210, 58441, 57692,
+      56962, 56250, 55555, 54878, 54216, 53571, 52941, 52325, 51724, 51136,
+      50561, 49450, 48387, 47368, 46875, 45918, 45000, 44554, 44117, 43269,
+      42452, 41666, 40909, 40178, 39473, 38793, 38135, 36885, 36290, 35714,
+      35156, 34615, 34090, 33582, 33088, 32608, 32142, 31468, 31034, 30405,
+      29801, 29220, 28662,
+  },
+  /* Inter case 285000/Qinter */
+  {
+      712500, 570000, 475000, 407142, 356250, 316666, 285000, 259090, 237500,
+      219230, 203571, 190000, 178125, 167647, 158333, 150000, 142500, 135714,
+      129545, 123913, 118750, 114000, 109615, 105555, 101785, 98275, 95000,
+      91935, 89062, 86363, 83823, 81428, 79166, 77027, 75000, 73076, 71250,
+      69512, 67857, 66279, 64772, 63333, 61956, 60638, 59375, 58163, 57000,
+      55882, 54807, 53773, 52777, 51818, 50892, 50000, 49137, 47500, 45967,
+      44531, 43181, 41911, 40714, 39583, 38513, 37500, 36538, 35625, 34756,
+      33928, 33139, 32386, 31666, 30978, 30319, 29687, 29081, 28500, 27941,
+      27403, 26886, 26388, 25909, 25446, 25000, 24568, 23949, 23360, 22800,
+      22265, 21755, 21268, 20802, 20357, 19930, 19520, 19127, 18750, 18387,
+      18037, 17701, 17378, 17065, 16764, 16473, 16101, 15745, 15405, 15079,
+      14766, 14467, 14179, 13902, 13636, 13380, 13133, 12895, 12666, 12445,
+      12179, 11924, 11632, 11445, 11220, 11003, 10795, 10594, 10401, 10215,
+      10035,
+  }
 };
 
-static const int kf_boost_qadjustment[QINDEX_RANGE] =
-{
-    128, 129, 130, 131, 132, 133, 134, 135,
-    136, 137, 138, 139, 140, 141, 142, 143,
-    144, 145, 146, 147, 148, 149, 150, 151,
-    152, 153, 154, 155, 156, 157, 158, 159,
-    160, 161, 162, 163, 164, 165, 166, 167,
-    168, 169, 170, 171, 172, 173, 174, 175,
-    176, 177, 178, 179, 180, 181, 182, 183,
-    184, 185, 186, 187, 188, 189, 190, 191,
-    192, 193, 194, 195, 196, 197, 198, 199,
-    200, 200, 201, 201, 202, 203, 203, 203,
-    204, 204, 205, 205, 206, 206, 207, 207,
-    208, 208, 209, 209, 210, 210, 211, 211,
-    212, 212, 213, 213, 214, 214, 215, 215,
-    216, 216, 217, 217, 218, 218, 219, 219,
-    220, 220, 220, 220, 220, 220, 220, 220,
-    220, 220, 220, 220, 220, 220, 220, 220,
+static const int kf_boost_qadjustment[QINDEX_RANGE] = {
+  128, 129, 130, 131, 132, 133, 134, 135, 136, 137, 138, 139, 140, 141, 142,
+  143, 144, 145, 146, 147, 148, 149, 150, 151, 152, 153, 154, 155, 156, 157,
+  158, 159, 160, 161, 162, 163, 164, 165, 166, 167, 168, 169, 170, 171, 172,
+  173, 174, 175, 176, 177, 178, 179, 180, 181, 182, 183, 184, 185, 186, 187,
+  188, 189, 190, 191, 192, 193, 194, 195, 196, 197, 198, 199, 200, 200, 201,
+  201, 202, 203, 203, 203, 204, 204, 205, 205, 206, 206, 207, 207, 208, 208,
+  209, 209, 210, 210, 211, 211, 212, 212, 213, 213, 214, 214, 215, 215, 216,
+  216, 217, 217, 218, 218, 219, 219, 220, 220, 220, 220, 220, 220, 220, 220,
+  220, 220, 220, 220, 220, 220, 220, 220,
 };
 
 /* #define GFQ_ADJUSTMENT (Q+100) */
 #define GFQ_ADJUSTMENT vp8_gf_boost_qadjustment[Q]
-const int vp8_gf_boost_qadjustment[QINDEX_RANGE] =
-{
-    80, 82, 84, 86, 88, 90, 92, 94,
-    96, 97, 98, 99, 100, 101, 102, 103,
-    104, 105, 106, 107, 108, 109, 110, 111,
-    112, 113, 114, 115, 116, 117, 118, 119,
-    120, 121, 122, 123, 124, 125, 126, 127,
-    128, 129, 130, 131, 132, 133, 134, 135,
-    136, 137, 138, 139, 140, 141, 142, 143,
-    144, 145, 146, 147, 148, 149, 150, 151,
-    152, 153, 154, 155, 156, 157, 158, 159,
-    160, 161, 162, 163, 164, 165, 166, 167,
-    168, 169, 170, 171, 172, 173, 174, 175,
-    176, 177, 178, 179, 180, 181, 182, 183,
-    184, 184, 185, 185, 186, 186, 187, 187,
-    188, 188, 189, 189, 190, 190, 191, 191,
-    192, 192, 193, 193, 194, 194, 194, 194,
-    195, 195, 196, 196, 197, 197, 198, 198
+const int vp8_gf_boost_qadjustment[QINDEX_RANGE] = {
+  80,  82,  84,  86,  88,  90,  92,  94,  96,  97,  98,  99,  100, 101, 102,
+  103, 104, 105, 106, 107, 108, 109, 110, 111, 112, 113, 114, 115, 116, 117,
+  118, 119, 120, 121, 122, 123, 124, 125, 126, 127, 128, 129, 130, 131, 132,
+  133, 134, 135, 136, 137, 138, 139, 140, 141, 142, 143, 144, 145, 146, 147,
+  148, 149, 150, 151, 152, 153, 154, 155, 156, 157, 158, 159, 160, 161, 162,
+  163, 164, 165, 166, 167, 168, 169, 170, 171, 172, 173, 174, 175, 176, 177,
+  178, 179, 180, 181, 182, 183, 184, 184, 185, 185, 186, 186, 187, 187, 188,
+  188, 189, 189, 190, 190, 191, 191, 192, 192, 193, 193, 194, 194, 194, 194,
+  195, 195, 196, 196, 197, 197, 198, 198
 };
 
 /*
@@ -155,273 +130,231 @@
 };
 */
 
-static const int kf_gf_boost_qlimits[QINDEX_RANGE] =
-{
-    150, 155, 160, 165, 170, 175, 180, 185,
-    190, 195, 200, 205, 210, 215, 220, 225,
-    230, 235, 240, 245, 250, 255, 260, 265,
-    270, 275, 280, 285, 290, 295, 300, 305,
-    310, 320, 330, 340, 350, 360, 370, 380,
-    390, 400, 410, 420, 430, 440, 450, 460,
-    470, 480, 490, 500, 510, 520, 530, 540,
-    550, 560, 570, 580, 590, 600, 600, 600,
-    600, 600, 600, 600, 600, 600, 600, 600,
-    600, 600, 600, 600, 600, 600, 600, 600,
-    600, 600, 600, 600, 600, 600, 600, 600,
-    600, 600, 600, 600, 600, 600, 600, 600,
-    600, 600, 600, 600, 600, 600, 600, 600,
-    600, 600, 600, 600, 600, 600, 600, 600,
-    600, 600, 600, 600, 600, 600, 600, 600,
-    600, 600, 600, 600, 600, 600, 600, 600,
+static const int kf_gf_boost_qlimits[QINDEX_RANGE] = {
+  150, 155, 160, 165, 170, 175, 180, 185, 190, 195, 200, 205, 210, 215, 220,
+  225, 230, 235, 240, 245, 250, 255, 260, 265, 270, 275, 280, 285, 290, 295,
+  300, 305, 310, 320, 330, 340, 350, 360, 370, 380, 390, 400, 410, 420, 430,
+  440, 450, 460, 470, 480, 490, 500, 510, 520, 530, 540, 550, 560, 570, 580,
+  590, 600, 600, 600, 600, 600, 600, 600, 600, 600, 600, 600, 600, 600, 600,
+  600, 600, 600, 600, 600, 600, 600, 600, 600, 600, 600, 600, 600, 600, 600,
+  600, 600, 600, 600, 600, 600, 600, 600, 600, 600, 600, 600, 600, 600, 600,
+  600, 600, 600, 600, 600, 600, 600, 600, 600, 600, 600, 600, 600, 600, 600,
+  600, 600, 600, 600, 600, 600, 600, 600,
 };
 
-static const int gf_adjust_table[101] =
-{
-    100,
-    115, 130, 145, 160, 175, 190, 200, 210, 220, 230,
-    240, 260, 270, 280, 290, 300, 310, 320, 330, 340,
-    350, 360, 370, 380, 390, 400, 400, 400, 400, 400,
-    400, 400, 400, 400, 400, 400, 400, 400, 400, 400,
-    400, 400, 400, 400, 400, 400, 400, 400, 400, 400,
-    400, 400, 400, 400, 400, 400, 400, 400, 400, 400,
-    400, 400, 400, 400, 400, 400, 400, 400, 400, 400,
-    400, 400, 400, 400, 400, 400, 400, 400, 400, 400,
-    400, 400, 400, 400, 400, 400, 400, 400, 400, 400,
-    400, 400, 400, 400, 400, 400, 400, 400, 400, 400,
+static const int gf_adjust_table[101] = {
+  100, 115, 130, 145, 160, 175, 190, 200, 210, 220, 230, 240, 260, 270, 280,
+  290, 300, 310, 320, 330, 340, 350, 360, 370, 380, 390, 400, 400, 400, 400,
+  400, 400, 400, 400, 400, 400, 400, 400, 400, 400, 400, 400, 400, 400, 400,
+  400, 400, 400, 400, 400, 400, 400, 400, 400, 400, 400, 400, 400, 400, 400,
+  400, 400, 400, 400, 400, 400, 400, 400, 400, 400, 400, 400, 400, 400, 400,
+  400, 400, 400, 400, 400, 400, 400, 400, 400, 400, 400, 400, 400, 400, 400,
+  400, 400, 400, 400, 400, 400, 400, 400, 400, 400, 400,
 };
 
-static const int gf_intra_usage_adjustment[20] =
-{
-    125, 120, 115, 110, 105, 100,  95,  85,  80,  75,
-    70,  65,  60,  55,  50,  50,  50,  50,  50,  50,
+static const int gf_intra_usage_adjustment[20] = {
+  125, 120, 115, 110, 105, 100, 95, 85, 80, 75,
+  70,  65,  60,  55,  50,  50,  50, 50, 50, 50,
 };
 
-static const int gf_interval_table[101] =
-{
-    7,
-    7, 7, 7, 7, 7, 7, 7, 7, 7, 7,
-    7, 7, 7, 7, 7, 7, 7, 7, 7, 7,
-    7, 7, 7, 7, 7, 7, 7, 7, 7, 7,
-    8, 8, 8, 8, 8, 8, 8, 8, 8, 8,
-    8, 8, 8, 8, 8, 8, 8, 8, 8, 8,
-    9, 9, 9, 9, 9, 9, 9, 9, 9, 9,
-    9, 9, 9, 9, 9, 9, 9, 9, 9, 9,
-    10, 10, 10, 10, 10, 10, 10, 10, 10, 10,
-    10, 10, 10, 10, 10, 10, 10, 10, 10, 10,
-    11, 11, 11, 11, 11, 11, 11, 11, 11, 11,
+static const int gf_interval_table[101] = {
+  7,  7,  7,  7,  7,  7,  7,  7,  7,  7,  7,  7,  7,  7,  7,  7,  7,
+  7,  7,  7,  7,  7,  7,  7,  7,  7,  7,  7,  7,  7,  7,  8,  8,  8,
+  8,  8,  8,  8,  8,  8,  8,  8,  8,  8,  8,  8,  8,  8,  8,  8,  8,
+  9,  9,  9,  9,  9,  9,  9,  9,  9,  9,  9,  9,  9,  9,  9,  9,  9,
+  9,  9,  9,  10, 10, 10, 10, 10, 10, 10, 10, 10, 10, 10, 10, 10, 10,
+  10, 10, 10, 10, 10, 10, 11, 11, 11, 11, 11, 11, 11, 11, 11, 11,
 };
 
-static const unsigned int prior_key_frame_weight[KEY_FRAME_CONTEXT] = { 1, 2, 3, 4, 5 };
+static const unsigned int prior_key_frame_weight[KEY_FRAME_CONTEXT] = { 1, 2, 3,
+                                                                        4, 5 };
 
+void vp8_save_coding_context(VP8_COMP *cpi) {
+  CODING_CONTEXT *const cc = &cpi->coding_context;
 
-void vp8_save_coding_context(VP8_COMP *cpi)
-{
-    CODING_CONTEXT *const cc = & cpi->coding_context;
+  /* Stores a snapshot of key state variables which can subsequently be
+   * restored with a call to vp8_restore_coding_context. These functions are
+   * intended for use in a re-code loop in vp8_compress_frame where the
+   * quantizer value is adjusted between loop iterations.
+   */
 
-    /* Stores a snapshot of key state variables which can subsequently be
-     * restored with a call to vp8_restore_coding_context. These functions are
-     * intended for use in a re-code loop in vp8_compress_frame where the
-     * quantizer value is adjusted between loop iterations.
-     */
+  cc->frames_since_key = cpi->frames_since_key;
+  cc->filter_level = cpi->common.filter_level;
+  cc->frames_till_gf_update_due = cpi->frames_till_gf_update_due;
+  cc->frames_since_golden = cpi->frames_since_golden;
 
-    cc->frames_since_key          = cpi->frames_since_key;
-    cc->filter_level             = cpi->common.filter_level;
-    cc->frames_till_gf_update_due   = cpi->frames_till_gf_update_due;
-    cc->frames_since_golden       = cpi->frames_since_golden;
+  vp8_copy(cc->mvc, cpi->common.fc.mvc);
+  vp8_copy(cc->mvcosts, cpi->rd_costs.mvcosts);
 
-    vp8_copy(cc->mvc,      cpi->common.fc.mvc);
-    vp8_copy(cc->mvcosts,  cpi->rd_costs.mvcosts);
+  vp8_copy(cc->ymode_prob, cpi->common.fc.ymode_prob);
+  vp8_copy(cc->uv_mode_prob, cpi->common.fc.uv_mode_prob);
 
-    vp8_copy(cc->ymode_prob,   cpi->common.fc.ymode_prob);
-    vp8_copy(cc->uv_mode_prob,  cpi->common.fc.uv_mode_prob);
+  vp8_copy(cc->ymode_count, cpi->mb.ymode_count);
+  vp8_copy(cc->uv_mode_count, cpi->mb.uv_mode_count);
 
-    vp8_copy(cc->ymode_count, cpi->mb.ymode_count);
-    vp8_copy(cc->uv_mode_count, cpi->mb.uv_mode_count);
-
-
-    /* Stats */
+/* Stats */
 #ifdef MODE_STATS
-    vp8_copy(cc->y_modes,       y_modes);
-    vp8_copy(cc->uv_modes,      uv_modes);
-    vp8_copy(cc->b_modes,       b_modes);
-    vp8_copy(cc->inter_y_modes,  inter_y_modes);
-    vp8_copy(cc->inter_uv_modes, inter_uv_modes);
-    vp8_copy(cc->inter_b_modes,  inter_b_modes);
+  vp8_copy(cc->y_modes, y_modes);
+  vp8_copy(cc->uv_modes, uv_modes);
+  vp8_copy(cc->b_modes, b_modes);
+  vp8_copy(cc->inter_y_modes, inter_y_modes);
+  vp8_copy(cc->inter_uv_modes, inter_uv_modes);
+  vp8_copy(cc->inter_b_modes, inter_b_modes);
 #endif
 
-    cc->this_frame_percent_intra = cpi->this_frame_percent_intra;
+  cc->this_frame_percent_intra = cpi->this_frame_percent_intra;
 }
 
+void vp8_restore_coding_context(VP8_COMP *cpi) {
+  CODING_CONTEXT *const cc = &cpi->coding_context;
 
-void vp8_restore_coding_context(VP8_COMP *cpi)
-{
-    CODING_CONTEXT *const cc = & cpi->coding_context;
+  /* Restore key state variables to the snapshot state stored in the
+   * previous call to vp8_save_coding_context.
+   */
 
-    /* Restore key state variables to the snapshot state stored in the
-     * previous call to vp8_save_coding_context.
-     */
+  cpi->frames_since_key = cc->frames_since_key;
+  cpi->common.filter_level = cc->filter_level;
+  cpi->frames_till_gf_update_due = cc->frames_till_gf_update_due;
+  cpi->frames_since_golden = cc->frames_since_golden;
 
-    cpi->frames_since_key         =   cc->frames_since_key;
-    cpi->common.filter_level     =   cc->filter_level;
-    cpi->frames_till_gf_update_due  =   cc->frames_till_gf_update_due;
-    cpi->frames_since_golden       =   cc->frames_since_golden;
+  vp8_copy(cpi->common.fc.mvc, cc->mvc);
 
-    vp8_copy(cpi->common.fc.mvc, cc->mvc);
+  vp8_copy(cpi->rd_costs.mvcosts, cc->mvcosts);
 
-    vp8_copy(cpi->rd_costs.mvcosts, cc->mvcosts);
+  vp8_copy(cpi->common.fc.ymode_prob, cc->ymode_prob);
+  vp8_copy(cpi->common.fc.uv_mode_prob, cc->uv_mode_prob);
 
-    vp8_copy(cpi->common.fc.ymode_prob,   cc->ymode_prob);
-    vp8_copy(cpi->common.fc.uv_mode_prob,  cc->uv_mode_prob);
+  vp8_copy(cpi->mb.ymode_count, cc->ymode_count);
+  vp8_copy(cpi->mb.uv_mode_count, cc->uv_mode_count);
 
-    vp8_copy(cpi->mb.ymode_count, cc->ymode_count);
-    vp8_copy(cpi->mb.uv_mode_count, cc->uv_mode_count);
-
-    /* Stats */
+/* Stats */
 #ifdef MODE_STATS
-    vp8_copy(y_modes, cc->y_modes);
-    vp8_copy(uv_modes, cc->uv_modes);
-    vp8_copy(b_modes, cc->b_modes);
-    vp8_copy(inter_y_modes, cc->inter_y_modes);
-    vp8_copy(inter_uv_modes, cc->inter_uv_modes);
-    vp8_copy(inter_b_modes, cc->inter_b_modes);
+  vp8_copy(y_modes, cc->y_modes);
+  vp8_copy(uv_modes, cc->uv_modes);
+  vp8_copy(b_modes, cc->b_modes);
+  vp8_copy(inter_y_modes, cc->inter_y_modes);
+  vp8_copy(inter_uv_modes, cc->inter_uv_modes);
+  vp8_copy(inter_b_modes, cc->inter_b_modes);
 #endif
 
-
-    cpi->this_frame_percent_intra = cc->this_frame_percent_intra;
+  cpi->this_frame_percent_intra = cc->this_frame_percent_intra;
 }
 
+void vp8_setup_key_frame(VP8_COMP *cpi) {
+  /* Setup for Key frame: */
 
-void vp8_setup_key_frame(VP8_COMP *cpi)
-{
-    /* Setup for Key frame: */
+  vp8_default_coef_probs(&cpi->common);
 
-    vp8_default_coef_probs(& cpi->common);
+  memcpy(cpi->common.fc.mvc, vp8_default_mv_context,
+         sizeof(vp8_default_mv_context));
+  {
+    int flag[2] = { 1, 1 };
+    vp8_build_component_cost_table(
+        cpi->mb.mvcost, (const MV_CONTEXT *)cpi->common.fc.mvc, flag);
+  }
 
-    memcpy(cpi->common.fc.mvc, vp8_default_mv_context, sizeof(vp8_default_mv_context));
-    {
-        int flag[2] = {1, 1};
-        vp8_build_component_cost_table(cpi->mb.mvcost, (const MV_CONTEXT *) cpi->common.fc.mvc, flag);
-    }
+  /* Make sure we initialize separate contexts for altref,gold, and normal.
+   * TODO shouldn't need 3 different copies of structure to do this!
+   */
+  memcpy(&cpi->lfc_a, &cpi->common.fc, sizeof(cpi->common.fc));
+  memcpy(&cpi->lfc_g, &cpi->common.fc, sizeof(cpi->common.fc));
+  memcpy(&cpi->lfc_n, &cpi->common.fc, sizeof(cpi->common.fc));
 
-    /* Make sure we initialize separate contexts for altref,gold, and normal.
-     * TODO shouldn't need 3 different copies of structure to do this!
-     */
-    memcpy(&cpi->lfc_a, &cpi->common.fc, sizeof(cpi->common.fc));
-    memcpy(&cpi->lfc_g, &cpi->common.fc, sizeof(cpi->common.fc));
-    memcpy(&cpi->lfc_n, &cpi->common.fc, sizeof(cpi->common.fc));
+  cpi->common.filter_level = cpi->common.base_qindex * 3 / 8;
 
-    cpi->common.filter_level = cpi->common.base_qindex * 3 / 8 ;
+  /* Provisional interval before next GF */
+  if (cpi->auto_gold)
+    cpi->frames_till_gf_update_due = cpi->baseline_gf_interval;
+  else
+    cpi->frames_till_gf_update_due = DEFAULT_GF_INTERVAL;
 
-    /* Provisional interval before next GF */
-    if (cpi->auto_gold)
-        cpi->frames_till_gf_update_due = cpi->baseline_gf_interval;
-    else
-        cpi->frames_till_gf_update_due = DEFAULT_GF_INTERVAL;
-
-    cpi->common.refresh_golden_frame = 1;
-    cpi->common.refresh_alt_ref_frame = 1;
+  cpi->common.refresh_golden_frame = 1;
+  cpi->common.refresh_alt_ref_frame = 1;
 }
 
-
 static int estimate_bits_at_q(int frame_kind, int Q, int MBs,
-                              double correction_factor)
-{
-    int Bpm = (int)(.5 + correction_factor * vp8_bits_per_mb[frame_kind][Q]);
+                              double correction_factor) {
+  int Bpm = (int)(.5 + correction_factor * vp8_bits_per_mb[frame_kind][Q]);
 
-    /* Attempt to retain reasonable accuracy without overflow. The cutoff is
-     * chosen such that the maximum product of Bpm and MBs fits 31 bits. The
-     * largest Bpm takes 20 bits.
-     */
-    if (MBs > (1 << 11))
-        return (Bpm >> BPER_MB_NORMBITS) * MBs;
-    else
-        return (Bpm * MBs) >> BPER_MB_NORMBITS;
+  /* Attempt to retain reasonable accuracy without overflow. The cutoff is
+   * chosen such that the maximum product of Bpm and MBs fits 31 bits. The
+   * largest Bpm takes 20 bits.
+   */
+  if (MBs > (1 << 11))
+    return (Bpm >> BPER_MB_NORMBITS) * MBs;
+  else
+    return (Bpm * MBs) >> BPER_MB_NORMBITS;
 }
 
+static void calc_iframe_target_size(VP8_COMP *cpi) {
+  /* boost defaults to half second */
+  int kf_boost;
+  uint64_t target;
 
-static void calc_iframe_target_size(VP8_COMP *cpi)
-{
-    /* boost defaults to half second */
-    int kf_boost;
-    uint64_t target;
+  /* Clear down mmx registers to allow floating point in what follows */
+  vp8_clear_system_state();
 
-    /* Clear down mmx registers to allow floating point in what follows */
-    vp8_clear_system_state();
+  if (cpi->oxcf.fixed_q >= 0) {
+    int Q = cpi->oxcf.key_q;
 
-    if (cpi->oxcf.fixed_q >= 0)
-    {
-        int Q = cpi->oxcf.key_q;
+    target = estimate_bits_at_q(INTRA_FRAME, Q, cpi->common.MBs,
+                                cpi->key_frame_rate_correction_factor);
+  } else if (cpi->pass == 2) {
+    /* New Two pass RC */
+    target = cpi->per_frame_bandwidth;
+  }
+  /* First Frame is a special case */
+  else if (cpi->common.current_video_frame == 0) {
+    /* 1 Pass there is no information on which to base size so use
+     * bandwidth per second * fraction of the initial buffer
+     * level
+     */
+    target = cpi->oxcf.starting_buffer_level / 2;
 
-        target = estimate_bits_at_q(INTRA_FRAME, Q, cpi->common.MBs,
-                                    cpi->key_frame_rate_correction_factor);
-    }
-    else if (cpi->pass == 2)
-    {
-        /* New Two pass RC */
-        target = cpi->per_frame_bandwidth;
-    }
-    /* First Frame is a special case */
-    else if (cpi->common.current_video_frame == 0)
-    {
-        /* 1 Pass there is no information on which to base size so use
-         * bandwidth per second * fraction of the initial buffer
-         * level
-         */
-        target = cpi->oxcf.starting_buffer_level / 2;
+    if (target > cpi->oxcf.target_bandwidth * 3 / 2)
+      target = cpi->oxcf.target_bandwidth * 3 / 2;
+  } else {
+    /* if this keyframe was forced, use a more recent Q estimate */
+    int Q = (cpi->common.frame_flags & FRAMEFLAGS_KEY) ? cpi->avg_frame_qindex
+                                                       : cpi->ni_av_qi;
 
-        if(target > cpi->oxcf.target_bandwidth * 3 / 2)
-            target = cpi->oxcf.target_bandwidth * 3 / 2;
+    int initial_boost = 32; /* |3.0 * per_frame_bandwidth| */
+    /* Boost depends somewhat on frame rate: only used for 1 layer case. */
+    if (cpi->oxcf.number_of_layers == 1) {
+      kf_boost = VPXMAX(initial_boost, (int)(2 * cpi->output_framerate - 16));
+    } else {
+      /* Initial factor: set target size to: |3.0 * per_frame_bandwidth|. */
+      kf_boost = initial_boost;
     }
-    else
-    {
-        /* if this keyframe was forced, use a more recent Q estimate */
-        int Q = (cpi->common.frame_flags & FRAMEFLAGS_KEY)
-                ? cpi->avg_frame_qindex : cpi->ni_av_qi;
 
-        int initial_boost = 32; /* |3.0 * per_frame_bandwidth| */
-        /* Boost depends somewhat on frame rate: only used for 1 layer case. */
-        if (cpi->oxcf.number_of_layers == 1) {
-          kf_boost = VPXMAX(initial_boost,
-                            (int)(2 * cpi->output_framerate - 16));
-        }
-        else {
-          /* Initial factor: set target size to: |3.0 * per_frame_bandwidth|. */
-          kf_boost = initial_boost;
-        }
+    /* adjustment up based on q: this factor ranges from ~1.2 to 2.2. */
+    kf_boost = kf_boost * kf_boost_qadjustment[Q] / 100;
 
-        /* adjustment up based on q: this factor ranges from ~1.2 to 2.2. */
-        kf_boost = kf_boost * kf_boost_qadjustment[Q] / 100;
+    /* frame separation adjustment ( down) */
+    if (cpi->frames_since_key < cpi->output_framerate / 2)
+      kf_boost =
+          (int)(kf_boost * cpi->frames_since_key / (cpi->output_framerate / 2));
 
-        /* frame separation adjustment ( down) */
-        if (cpi->frames_since_key  < cpi->output_framerate / 2)
-            kf_boost = (int)(kf_boost
-                       * cpi->frames_since_key / (cpi->output_framerate / 2));
+    /* Minimal target size is |2* per_frame_bandwidth|. */
+    if (kf_boost < 16) kf_boost = 16;
 
-        /* Minimal target size is |2* per_frame_bandwidth|. */
-        if (kf_boost < 16)
-            kf_boost = 16;
+    target = ((16 + kf_boost) * cpi->per_frame_bandwidth) >> 4;
+  }
 
-        target = ((16 + kf_boost) * cpi->per_frame_bandwidth) >> 4;
-    }
+  if (cpi->oxcf.rc_max_intra_bitrate_pct) {
+    unsigned int max_rate =
+        cpi->per_frame_bandwidth * cpi->oxcf.rc_max_intra_bitrate_pct / 100;
 
+    if (target > max_rate) target = max_rate;
+  }
 
-    if (cpi->oxcf.rc_max_intra_bitrate_pct)
-    {
-        unsigned int max_rate = cpi->per_frame_bandwidth
-                                * cpi->oxcf.rc_max_intra_bitrate_pct / 100;
+  cpi->this_frame_target = (int)target;
 
-        if (target > max_rate)
-            target = max_rate;
-    }
+  /* TODO: if we separate rate targeting from Q targetting, move this.
+   * Reset the active worst quality to the baseline value for key frames.
+   */
+  if (cpi->pass != 2) cpi->active_worst_quality = cpi->worst_quality;
 
-    cpi->this_frame_target = (int)target;
-
-    /* TODO: if we separate rate targeting from Q targetting, move this.
-     * Reset the active worst quality to the baseline value for key frames.
-     */
-    if (cpi->pass != 2)
-        cpi->active_worst_quality = cpi->worst_quality;
-
 #if 0
     {
         FILE *f;
@@ -435,40 +368,38 @@
 #endif
 }
 
-
 /* Do the best we can to define the parameters for the next GF based on what
  * information we have available.
  */
-static void calc_gf_params(VP8_COMP *cpi)
-{
-    int Q = (cpi->oxcf.fixed_q < 0) ? cpi->last_q[INTER_FRAME] : cpi->oxcf.fixed_q;
-    int Boost = 0;
+static void calc_gf_params(VP8_COMP *cpi) {
+  int Q =
+      (cpi->oxcf.fixed_q < 0) ? cpi->last_q[INTER_FRAME] : cpi->oxcf.fixed_q;
+  int Boost = 0;
 
-    int gf_frame_useage = 0;      /* Golden frame useage since last GF */
-    int tot_mbs = cpi->recent_ref_frame_usage[INTRA_FRAME]  +
-                  cpi->recent_ref_frame_usage[LAST_FRAME]   +
-                  cpi->recent_ref_frame_usage[GOLDEN_FRAME] +
-                  cpi->recent_ref_frame_usage[ALTREF_FRAME];
+  int gf_frame_useage = 0; /* Golden frame useage since last GF */
+  int tot_mbs = cpi->recent_ref_frame_usage[INTRA_FRAME] +
+                cpi->recent_ref_frame_usage[LAST_FRAME] +
+                cpi->recent_ref_frame_usage[GOLDEN_FRAME] +
+                cpi->recent_ref_frame_usage[ALTREF_FRAME];
 
-    int pct_gf_active = (100 * cpi->gf_active_count) / (cpi->common.mb_rows * cpi->common.mb_cols);
+  int pct_gf_active = (100 * cpi->gf_active_count) /
+                      (cpi->common.mb_rows * cpi->common.mb_cols);
 
-    if (tot_mbs)
-        gf_frame_useage = (cpi->recent_ref_frame_usage[GOLDEN_FRAME] + cpi->recent_ref_frame_usage[ALTREF_FRAME]) * 100 / tot_mbs;
+  if (tot_mbs)
+    gf_frame_useage = (cpi->recent_ref_frame_usage[GOLDEN_FRAME] +
+                       cpi->recent_ref_frame_usage[ALTREF_FRAME]) *
+                      100 / tot_mbs;
 
-    if (pct_gf_active > gf_frame_useage)
-        gf_frame_useage = pct_gf_active;
+  if (pct_gf_active > gf_frame_useage) gf_frame_useage = pct_gf_active;
 
-    /* Not two pass */
-    if (cpi->pass != 2)
-    {
-        /* Single Pass lagged mode: TBD */
-        if (0)
-        {
-        }
+  /* Not two pass */
+  if (cpi->pass != 2) {
+    /* Single Pass lagged mode: TBD */
+    if (0) {
+    }
 
-        /* Single Pass compression: Has to use current and historical data */
-        else
-        {
+    /* Single Pass compression: Has to use current and historical data */
+    else {
 #if 0
             /* Experimental code */
             int index = cpi->one_pass_frame_index;
@@ -517,430 +448,379 @@
             */
 #else
 
-            /*************************************************************/
-            /* OLD code */
+      /*************************************************************/
+      /* OLD code */
 
-            /* Adjust boost based upon ambient Q */
-            Boost = GFQ_ADJUSTMENT;
+      /* Adjust boost based upon ambient Q */
+      Boost = GFQ_ADJUSTMENT;
 
-            /* Adjust based upon most recently measure intra useage */
-            Boost = Boost * gf_intra_usage_adjustment[(cpi->this_frame_percent_intra < 15) ? cpi->this_frame_percent_intra : 14] / 100;
+      /* Adjust based upon most recently measure intra useage */
+      Boost = Boost *
+              gf_intra_usage_adjustment[(cpi->this_frame_percent_intra < 15)
+                                            ? cpi->this_frame_percent_intra
+                                            : 14] /
+              100;
 
-            /* Adjust gf boost based upon GF usage since last GF */
-            Boost = Boost * gf_adjust_table[gf_frame_useage] / 100;
+      /* Adjust gf boost based upon GF usage since last GF */
+      Boost = Boost * gf_adjust_table[gf_frame_useage] / 100;
 #endif
-        }
+    }
 
-        /* golden frame boost without recode loop often goes awry.  be
-         * safe by keeping numbers down.
-         */
-        if (!cpi->sf.recode_loop)
-        {
-            if (cpi->compressor_speed == 2)
-                Boost = Boost / 2;
-        }
+    /* golden frame boost without recode loop often goes awry.  be
+     * safe by keeping numbers down.
+     */
+    if (!cpi->sf.recode_loop) {
+      if (cpi->compressor_speed == 2) Boost = Boost / 2;
+    }
 
-        /* Apply an upper limit based on Q for 1 pass encodes */
-        if (Boost > kf_gf_boost_qlimits[Q] && (cpi->pass == 0))
-            Boost = kf_gf_boost_qlimits[Q];
+    /* Apply an upper limit based on Q for 1 pass encodes */
+    if (Boost > kf_gf_boost_qlimits[Q] && (cpi->pass == 0))
+      Boost = kf_gf_boost_qlimits[Q];
 
-        /* Apply lower limits to boost. */
-        else if (Boost < 110)
-            Boost = 110;
+    /* Apply lower limits to boost. */
+    else if (Boost < 110)
+      Boost = 110;
 
-        /* Note the boost used */
-        cpi->last_boost = Boost;
+    /* Note the boost used */
+    cpi->last_boost = Boost;
+  }
 
-    }
-
-    /* Estimate next interval
-     * This is updated once the real frame size/boost is known.
-     */
-    if (cpi->oxcf.fixed_q == -1)
+  /* Estimate next interval
+   * This is updated once the real frame size/boost is known.
+   */
+  if (cpi->oxcf.fixed_q == -1) {
+    if (cpi->pass == 2) /* 2 Pass */
     {
-        if (cpi->pass == 2)         /* 2 Pass */
-        {
-            cpi->frames_till_gf_update_due = cpi->baseline_gf_interval;
-        }
-        else                            /* 1 Pass */
-        {
-            cpi->frames_till_gf_update_due = cpi->baseline_gf_interval;
+      cpi->frames_till_gf_update_due = cpi->baseline_gf_interval;
+    } else /* 1 Pass */
+    {
+      cpi->frames_till_gf_update_due = cpi->baseline_gf_interval;
 
-            if (cpi->last_boost > 750)
-                cpi->frames_till_gf_update_due++;
+      if (cpi->last_boost > 750) cpi->frames_till_gf_update_due++;
 
-            if (cpi->last_boost > 1000)
-                cpi->frames_till_gf_update_due++;
+      if (cpi->last_boost > 1000) cpi->frames_till_gf_update_due++;
 
-            if (cpi->last_boost > 1250)
-                cpi->frames_till_gf_update_due++;
+      if (cpi->last_boost > 1250) cpi->frames_till_gf_update_due++;
 
-            if (cpi->last_boost >= 1500)
-                cpi->frames_till_gf_update_due ++;
+      if (cpi->last_boost >= 1500) cpi->frames_till_gf_update_due++;
 
-            if (gf_interval_table[gf_frame_useage] > cpi->frames_till_gf_update_due)
-                cpi->frames_till_gf_update_due = gf_interval_table[gf_frame_useage];
+      if (gf_interval_table[gf_frame_useage] > cpi->frames_till_gf_update_due)
+        cpi->frames_till_gf_update_due = gf_interval_table[gf_frame_useage];
 
-            if (cpi->frames_till_gf_update_due > cpi->max_gf_interval)
-                cpi->frames_till_gf_update_due = cpi->max_gf_interval;
-        }
+      if (cpi->frames_till_gf_update_due > cpi->max_gf_interval)
+        cpi->frames_till_gf_update_due = cpi->max_gf_interval;
     }
-    else
-        cpi->frames_till_gf_update_due = cpi->baseline_gf_interval;
+  } else
+    cpi->frames_till_gf_update_due = cpi->baseline_gf_interval;
 
-    /* ARF on or off */
-    if (cpi->pass != 2)
-    {
-        /* For now Alt ref is not allowed except in 2 pass modes. */
-        cpi->source_alt_ref_pending = 0;
+  /* ARF on or off */
+  if (cpi->pass != 2) {
+    /* For now Alt ref is not allowed except in 2 pass modes. */
+    cpi->source_alt_ref_pending = 0;
 
-        /*if ( cpi->oxcf.fixed_q == -1)
-        {
-            if ( cpi->oxcf.play_alternate && (cpi->last_boost > (100 + (AF_THRESH*cpi->frames_till_gf_update_due)) ) )
-                cpi->source_alt_ref_pending = 1;
-            else
-                cpi->source_alt_ref_pending = 0;
-        }*/
-    }
+    /*if ( cpi->oxcf.fixed_q == -1)
+    {
+        if ( cpi->oxcf.play_alternate && (cpi->last_boost > (100 +
+    (AF_THRESH*cpi->frames_till_gf_update_due)) ) )
+            cpi->source_alt_ref_pending = 1;
+        else
+            cpi->source_alt_ref_pending = 0;
+    }*/
+  }
 }
 
+static void calc_pframe_target_size(VP8_COMP *cpi) {
+  int min_frame_target;
+  int old_per_frame_bandwidth = cpi->per_frame_bandwidth;
 
-static void calc_pframe_target_size(VP8_COMP *cpi)
-{
-    int min_frame_target;
-    int old_per_frame_bandwidth = cpi->per_frame_bandwidth;
+  if (cpi->current_layer > 0)
+    cpi->per_frame_bandwidth =
+        cpi->layer_context[cpi->current_layer].avg_frame_size_for_layer;
 
-    if ( cpi->current_layer > 0)
-        cpi->per_frame_bandwidth =
-            cpi->layer_context[cpi->current_layer].avg_frame_size_for_layer;
+  min_frame_target = 0;
 
-    min_frame_target = 0;
+  if (cpi->pass == 2) {
+    min_frame_target = cpi->min_frame_bandwidth;
 
-    if (cpi->pass == 2)
-    {
-        min_frame_target = cpi->min_frame_bandwidth;
+    if (min_frame_target < (cpi->av_per_frame_bandwidth >> 5))
+      min_frame_target = cpi->av_per_frame_bandwidth >> 5;
+  } else if (min_frame_target < cpi->per_frame_bandwidth / 4)
+    min_frame_target = cpi->per_frame_bandwidth / 4;
 
-        if (min_frame_target < (cpi->av_per_frame_bandwidth >> 5))
-            min_frame_target = cpi->av_per_frame_bandwidth >> 5;
+  /* Special alt reference frame case */
+  if ((cpi->common.refresh_alt_ref_frame) &&
+      (cpi->oxcf.number_of_layers == 1)) {
+    if (cpi->pass == 2) {
+      /* Per frame bit target for the alt ref frame */
+      cpi->per_frame_bandwidth = cpi->twopass.gf_bits;
+      cpi->this_frame_target = cpi->per_frame_bandwidth;
     }
-    else if (min_frame_target < cpi->per_frame_bandwidth / 4)
-        min_frame_target = cpi->per_frame_bandwidth / 4;
 
+    /* One Pass ??? TBD */
+  }
 
-    /* Special alt reference frame case */
-    if((cpi->common.refresh_alt_ref_frame) && (cpi->oxcf.number_of_layers == 1))
-    {
-        if (cpi->pass == 2)
-        {
-            /* Per frame bit target for the alt ref frame */
-            cpi->per_frame_bandwidth = cpi->twopass.gf_bits;
-            cpi->this_frame_target = cpi->per_frame_bandwidth;
-        }
-
-        /* One Pass ??? TBD */
+  /* Normal frames (gf,and inter) */
+  else {
+    /* 2 pass */
+    if (cpi->pass == 2) {
+      cpi->this_frame_target = cpi->per_frame_bandwidth;
     }
+    /* 1 pass */
+    else {
+      int Adjustment;
+      /* Make rate adjustment to recover bits spent in key frame
+       * Test to see if the key frame inter data rate correction
+       * should still be in force
+       */
+      if (cpi->kf_overspend_bits > 0) {
+        Adjustment = (cpi->kf_bitrate_adjustment <= cpi->kf_overspend_bits)
+                         ? cpi->kf_bitrate_adjustment
+                         : cpi->kf_overspend_bits;
 
-    /* Normal frames (gf,and inter) */
-    else
-    {
-        /* 2 pass */
-        if (cpi->pass == 2)
-        {
-            cpi->this_frame_target = cpi->per_frame_bandwidth;
-        }
-        /* 1 pass */
-        else
-        {
-            int Adjustment;
-            /* Make rate adjustment to recover bits spent in key frame
-             * Test to see if the key frame inter data rate correction
-             * should still be in force
-             */
-            if (cpi->kf_overspend_bits > 0)
-            {
-                Adjustment = (cpi->kf_bitrate_adjustment <= cpi->kf_overspend_bits) ? cpi->kf_bitrate_adjustment : cpi->kf_overspend_bits;
+        if (Adjustment > (cpi->per_frame_bandwidth - min_frame_target))
+          Adjustment = (cpi->per_frame_bandwidth - min_frame_target);
 
-                if (Adjustment > (cpi->per_frame_bandwidth - min_frame_target))
-                    Adjustment = (cpi->per_frame_bandwidth - min_frame_target);
+        cpi->kf_overspend_bits -= Adjustment;
 
-                cpi->kf_overspend_bits -= Adjustment;
+        /* Calculate an inter frame bandwidth target for the next
+         * few frames designed to recover any extra bits spent on
+         * the key frame.
+         */
+        cpi->this_frame_target = cpi->per_frame_bandwidth - Adjustment;
 
-                /* Calculate an inter frame bandwidth target for the next
-                 * few frames designed to recover any extra bits spent on
-                 * the key frame.
-                 */
-                cpi->this_frame_target = cpi->per_frame_bandwidth - Adjustment;
+        if (cpi->this_frame_target < min_frame_target)
+          cpi->this_frame_target = min_frame_target;
+      } else
+        cpi->this_frame_target = cpi->per_frame_bandwidth;
 
-                if (cpi->this_frame_target < min_frame_target)
-                    cpi->this_frame_target = min_frame_target;
-            }
-            else
-                cpi->this_frame_target = cpi->per_frame_bandwidth;
+      /* If appropriate make an adjustment to recover bits spent on a
+       * recent GF
+       */
+      if ((cpi->gf_overspend_bits > 0) &&
+          (cpi->this_frame_target > min_frame_target)) {
+        Adjustment = (cpi->non_gf_bitrate_adjustment <= cpi->gf_overspend_bits)
+                         ? cpi->non_gf_bitrate_adjustment
+                         : cpi->gf_overspend_bits;
 
-            /* If appropriate make an adjustment to recover bits spent on a
-             * recent GF
-             */
-            if ((cpi->gf_overspend_bits > 0) && (cpi->this_frame_target > min_frame_target))
-            {
-                Adjustment = (cpi->non_gf_bitrate_adjustment <= cpi->gf_overspend_bits) ? cpi->non_gf_bitrate_adjustment : cpi->gf_overspend_bits;
+        if (Adjustment > (cpi->this_frame_target - min_frame_target))
+          Adjustment = (cpi->this_frame_target - min_frame_target);
 
-                if (Adjustment > (cpi->this_frame_target - min_frame_target))
-                    Adjustment = (cpi->this_frame_target - min_frame_target);
+        cpi->gf_overspend_bits -= Adjustment;
+        cpi->this_frame_target -= Adjustment;
+      }
 
-                cpi->gf_overspend_bits -= Adjustment;
-                cpi->this_frame_target -= Adjustment;
-            }
+      /* Apply small + and - boosts for non gf frames */
+      if ((cpi->last_boost > 150) && (cpi->frames_till_gf_update_due > 0) &&
+          (cpi->current_gf_interval >= (MIN_GF_INTERVAL << 1))) {
+        /* % Adjustment limited to the range 1% to 10% */
+        Adjustment = (cpi->last_boost - 100) >> 5;
 
-            /* Apply small + and - boosts for non gf frames */
-            if ((cpi->last_boost > 150) && (cpi->frames_till_gf_update_due > 0) &&
-                (cpi->current_gf_interval >= (MIN_GF_INTERVAL << 1)))
-            {
-                /* % Adjustment limited to the range 1% to 10% */
-                Adjustment = (cpi->last_boost - 100) >> 5;
+        if (Adjustment < 1)
+          Adjustment = 1;
+        else if (Adjustment > 10)
+          Adjustment = 10;
 
-                if (Adjustment < 1)
-                    Adjustment = 1;
-                else if (Adjustment > 10)
-                    Adjustment = 10;
+        /* Convert to bits */
+        Adjustment = (cpi->this_frame_target * Adjustment) / 100;
 
-                /* Convert to bits */
-                Adjustment = (cpi->this_frame_target * Adjustment) / 100;
+        if (Adjustment > (cpi->this_frame_target - min_frame_target))
+          Adjustment = (cpi->this_frame_target - min_frame_target);
 
-                if (Adjustment > (cpi->this_frame_target - min_frame_target))
-                    Adjustment = (cpi->this_frame_target - min_frame_target);
-
-                if (cpi->frames_since_golden == (cpi->current_gf_interval >> 1))
-                {
-                    Adjustment = (cpi->current_gf_interval - 1) * Adjustment;
-                    // Limit adjustment to 10% of current target.
-                    if (Adjustment > (10 * cpi->this_frame_target) / 100)
-                        Adjustment = (10 * cpi->this_frame_target) / 100;
-                    cpi->this_frame_target += Adjustment;
-                }
-                else
-                    cpi->this_frame_target -= Adjustment;
-            }
-        }
+        if (cpi->frames_since_golden == (cpi->current_gf_interval >> 1)) {
+          Adjustment = (cpi->current_gf_interval - 1) * Adjustment;
+          // Limit adjustment to 10% of current target.
+          if (Adjustment > (10 * cpi->this_frame_target) / 100)
+            Adjustment = (10 * cpi->this_frame_target) / 100;
+          cpi->this_frame_target += Adjustment;
+        } else
+          cpi->this_frame_target -= Adjustment;
+      }
     }
+  }
 
-    /* Sanity check that the total sum of adjustments is not above the
-     * maximum allowed That is that having allowed for KF and GF penalties
-     * we have not pushed the current interframe target to low. If the
-     * adjustment we apply here is not capable of recovering all the extra
-     * bits we have spent in the KF or GF then the remainder will have to
-     * be recovered over a longer time span via other buffer / rate control
-     * mechanisms.
-     */
-    if (cpi->this_frame_target < min_frame_target)
-        cpi->this_frame_target = min_frame_target;
+  /* Sanity check that the total sum of adjustments is not above the
+   * maximum allowed That is that having allowed for KF and GF penalties
+   * we have not pushed the current interframe target to low. If the
+   * adjustment we apply here is not capable of recovering all the extra
+   * bits we have spent in the KF or GF then the remainder will have to
+   * be recovered over a longer time span via other buffer / rate control
+   * mechanisms.
+   */
+  if (cpi->this_frame_target < min_frame_target)
+    cpi->this_frame_target = min_frame_target;
 
-    if (!cpi->common.refresh_alt_ref_frame)
-        /* Note the baseline target data rate for this inter frame. */
-        cpi->inter_frame_target = cpi->this_frame_target;
+  if (!cpi->common.refresh_alt_ref_frame)
+    /* Note the baseline target data rate for this inter frame. */
+    cpi->inter_frame_target = cpi->this_frame_target;
 
-    /* One Pass specific code */
-    if (cpi->pass == 0)
-    {
-        /* Adapt target frame size with respect to any buffering constraints: */
-        if (cpi->buffered_mode)
-        {
-            int one_percent_bits = (int)
-                (1 + cpi->oxcf.optimal_buffer_level / 100);
+  /* One Pass specific code */
+  if (cpi->pass == 0) {
+    /* Adapt target frame size with respect to any buffering constraints: */
+    if (cpi->buffered_mode) {
+      int one_percent_bits = (int)(1 + cpi->oxcf.optimal_buffer_level / 100);
 
-            if ((cpi->buffer_level < cpi->oxcf.optimal_buffer_level) ||
-                (cpi->bits_off_target < cpi->oxcf.optimal_buffer_level))
-            {
-                int percent_low = 0;
+      if ((cpi->buffer_level < cpi->oxcf.optimal_buffer_level) ||
+          (cpi->bits_off_target < cpi->oxcf.optimal_buffer_level)) {
+        int percent_low = 0;
 
-                /* Decide whether or not we need to adjust the frame data
-                 * rate target.
-                 *
-                 * If we are are below the optimal buffer fullness level
-                 * and adherence to buffering constraints is important to
-                 * the end usage then adjust the per frame target.
-                 */
-                if ((cpi->oxcf.end_usage == USAGE_STREAM_FROM_SERVER) &&
-                    (cpi->buffer_level < cpi->oxcf.optimal_buffer_level))
-                {
-                    percent_low = (int)
-                        ((cpi->oxcf.optimal_buffer_level - cpi->buffer_level) /
-                        one_percent_bits);
-                }
-                /* Are we overshooting the long term clip data rate... */
-                else if (cpi->bits_off_target < 0)
-                {
-                    /* Adjust per frame data target downwards to compensate. */
-                    percent_low = (int)(100 * -cpi->bits_off_target /
-                                       (cpi->total_byte_count * 8));
-                }
+        /* Decide whether or not we need to adjust the frame data
+         * rate target.
+         *
+         * If we are are below the optimal buffer fullness level
+         * and adherence to buffering constraints is important to
+         * the end usage then adjust the per frame target.
+         */
+        if ((cpi->oxcf.end_usage == USAGE_STREAM_FROM_SERVER) &&
+            (cpi->buffer_level < cpi->oxcf.optimal_buffer_level)) {
+          percent_low =
+              (int)((cpi->oxcf.optimal_buffer_level - cpi->buffer_level) /
+                    one_percent_bits);
+        }
+        /* Are we overshooting the long term clip data rate... */
+        else if (cpi->bits_off_target < 0) {
+          /* Adjust per frame data target downwards to compensate. */
+          percent_low =
+              (int)(100 * -cpi->bits_off_target / (cpi->total_byte_count * 8));
+        }
 
-                if (percent_low > cpi->oxcf.under_shoot_pct)
-                    percent_low = cpi->oxcf.under_shoot_pct;
-                else if (percent_low < 0)
-                    percent_low = 0;
+        if (percent_low > cpi->oxcf.under_shoot_pct)
+          percent_low = cpi->oxcf.under_shoot_pct;
+        else if (percent_low < 0)
+          percent_low = 0;
 
-                /* lower the target bandwidth for this frame. */
-                cpi->this_frame_target -=
-                        (cpi->this_frame_target * percent_low) / 200;
+        /* lower the target bandwidth for this frame. */
+        cpi->this_frame_target -= (cpi->this_frame_target * percent_low) / 200;
 
-                /* Are we using allowing control of active_worst_allowed_q
-                 * according to buffer level.
-                 */
-                if (cpi->auto_worst_q && cpi->ni_frames > 150)
-                {
-                    int64_t critical_buffer_level;
-
-                    /* For streaming applications the most important factor is
-                     * cpi->buffer_level as this takes into account the
-                     * specified short term buffering constraints. However,
-                     * hitting the long term clip data rate target is also
-                     * important.
-                     */
-                    if (cpi->oxcf.end_usage == USAGE_STREAM_FROM_SERVER)
-                    {
-                        /* Take the smaller of cpi->buffer_level and
-                         * cpi->bits_off_target
-                         */
-                        critical_buffer_level =
-                            (cpi->buffer_level < cpi->bits_off_target)
-                            ? cpi->buffer_level : cpi->bits_off_target;
-                    }
-                    /* For local file playback short term buffering constraints
-                     * are less of an issue
-                     */
-                    else
-                    {
-                        /* Consider only how we are doing for the clip as a
-                         * whole
-                         */
-                        critical_buffer_level = cpi->bits_off_target;
-                    }
+        /* Are we using allowing control of active_worst_allowed_q
+         * according to buffer level.
+         */
+        if (cpi->auto_worst_q && cpi->ni_frames > 150) {
+          int64_t critical_buffer_level;
 
-                    /* Set the active worst quality based upon the selected
-                     * buffer fullness number.
-                     */
-                    if (critical_buffer_level < cpi->oxcf.optimal_buffer_level)
-                    {
-                        if ( critical_buffer_level >
-                             (cpi->oxcf.optimal_buffer_level >> 2) )
-                        {
-                            int64_t qadjustment_range =
-                                      cpi->worst_quality - cpi->ni_av_qi;
-                            int64_t above_base =
-                                      (critical_buffer_level -
-                                       (cpi->oxcf.optimal_buffer_level >> 2));
-
-                            /* Step active worst quality down from
-                             * cpi->ni_av_qi when (critical_buffer_level ==
-                             * cpi->optimal_buffer_level) to
-                             * cpi->worst_quality when
-                             * (critical_buffer_level ==
-                             *     cpi->optimal_buffer_level >> 2)
-                             */
-                            cpi->active_worst_quality =
-                                cpi->worst_quality -
-                                (int)((qadjustment_range * above_base) /
-                                 (cpi->oxcf.optimal_buffer_level*3>>2));
-                        }
-                        else
-                        {
-                            cpi->active_worst_quality = cpi->worst_quality;
-                        }
-                    }
-                    else
-                    {
-                        cpi->active_worst_quality = cpi->ni_av_qi;
-                    }
-                }
-                else
-                {
-                    cpi->active_worst_quality = cpi->worst_quality;
-                }
-            }
-            else
-            {
-                int percent_high = 0;
+          /* For streaming applications the most important factor is
+           * cpi->buffer_level as this takes into account the
+           * specified short term buffering constraints. However,
+           * hitting the long term clip data rate target is also
+           * important.
+           */
+          if (cpi->oxcf.end_usage == USAGE_STREAM_FROM_SERVER) {
+            /* Take the smaller of cpi->buffer_level and
+             * cpi->bits_off_target
+             */
+            critical_buffer_level = (cpi->buffer_level < cpi->bits_off_target)
+                                        ? cpi->buffer_level
+                                        : cpi->bits_off_target;
+          }
+          /* For local file playback short term buffering constraints
+           * are less of an issue
+           */
+          else {
+            /* Consider only how we are doing for the clip as a
+             * whole
+             */
+            critical_buffer_level = cpi->bits_off_target;
+          }
 
-                if ((cpi->oxcf.end_usage == USAGE_STREAM_FROM_SERVER)
-                     && (cpi->buffer_level > cpi->oxcf.optimal_buffer_level))
-                {
-                    percent_high = (int)((cpi->buffer_level
-                                    - cpi->oxcf.optimal_buffer_level)
-                                   / one_percent_bits);
-                }
-                else if (cpi->bits_off_target > cpi->oxcf.optimal_buffer_level)
-                {
-                    percent_high = (int)((100 * cpi->bits_off_target)
-                                         / (cpi->total_byte_count * 8));
-                }
+          /* Set the active worst quality based upon the selected
+           * buffer fullness number.
+           */
+          if (critical_buffer_level < cpi->oxcf.optimal_buffer_level) {
+            if (critical_buffer_level > (cpi->oxcf.optimal_buffer_level >> 2)) {
+              int64_t qadjustment_range = cpi->worst_quality - cpi->ni_av_qi;
+              int64_t above_base = (critical_buffer_level -
+                                    (cpi->oxcf.optimal_buffer_level >> 2));
 
-                if (percent_high > cpi->oxcf.over_shoot_pct)
-                    percent_high = cpi->oxcf.over_shoot_pct;
-                else if (percent_high < 0)
-                    percent_high = 0;
-
-                cpi->this_frame_target += (cpi->this_frame_target *
-                                          percent_high) / 200;
-
-                /* Are we allowing control of active_worst_allowed_q according
-                 * to buffer level.
-                 */
-                if (cpi->auto_worst_q && cpi->ni_frames > 150)
-                {
-                    /* When using the relaxed buffer model stick to the
-                     * user specified value
-                     */
-                    cpi->active_worst_quality = cpi->ni_av_qi;
-                }
-                else
-                {
-                    cpi->active_worst_quality = cpi->worst_quality;
-                }
+              /* Step active worst quality down from
+               * cpi->ni_av_qi when (critical_buffer_level ==
+               * cpi->optimal_buffer_level) to
+               * cpi->worst_quality when
+               * (critical_buffer_level ==
+               *     cpi->optimal_buffer_level >> 2)
+               */
+              cpi->active_worst_quality =
+                  cpi->worst_quality -
+                  (int)((qadjustment_range * above_base) /
+                        (cpi->oxcf.optimal_buffer_level * 3 >> 2));
+            } else {
+              cpi->active_worst_quality = cpi->worst_quality;
             }
+          } else {
+            cpi->active_worst_quality = cpi->ni_av_qi;
+          }
+        } else {
+          cpi->active_worst_quality = cpi->worst_quality;
+        }
+      } else {
+        int percent_high = 0;
 
-            /* Set active_best_quality to prevent quality rising too high */
-            cpi->active_best_quality = cpi->best_quality;
+        if ((cpi->oxcf.end_usage == USAGE_STREAM_FROM_SERVER) &&
+            (cpi->buffer_level > cpi->oxcf.optimal_buffer_level)) {
+          percent_high =
+              (int)((cpi->buffer_level - cpi->oxcf.optimal_buffer_level) /
+                    one_percent_bits);
+        } else if (cpi->bits_off_target > cpi->oxcf.optimal_buffer_level) {
+          percent_high =
+              (int)((100 * cpi->bits_off_target) / (cpi->total_byte_count * 8));
+        }
 
-            /* Worst quality obviously must not be better than best quality */
-            if (cpi->active_worst_quality <= cpi->active_best_quality)
-                cpi->active_worst_quality = cpi->active_best_quality + 1;
+        if (percent_high > cpi->oxcf.over_shoot_pct)
+          percent_high = cpi->oxcf.over_shoot_pct;
+        else if (percent_high < 0)
+          percent_high = 0;
 
-            if(cpi->active_worst_quality > 127)
-                cpi->active_worst_quality = 127;
-        }
-        /* Unbuffered mode (eg. video conferencing) */
-        else
-        {
-            /* Set the active worst quality */
-            cpi->active_worst_quality = cpi->worst_quality;
-        }
+        cpi->this_frame_target += (cpi->this_frame_target * percent_high) / 200;
 
-        /* Special trap for constrained quality mode
-         * "active_worst_quality" may never drop below cq level
-         * for any frame type.
+        /* Are we allowing control of active_worst_allowed_q according
+         * to buffer level.
          */
-        if ( cpi->oxcf.end_usage == USAGE_CONSTRAINED_QUALITY &&
-             cpi->active_worst_quality < cpi->cq_target_quality)
-        {
-            cpi->active_worst_quality = cpi->cq_target_quality;
+        if (cpi->auto_worst_q && cpi->ni_frames > 150) {
+          /* When using the relaxed buffer model stick to the
+           * user specified value
+           */
+          cpi->active_worst_quality = cpi->ni_av_qi;
+        } else {
+          cpi->active_worst_quality = cpi->worst_quality;
         }
+      }
+
+      /* Set active_best_quality to prevent quality rising too high */
+      cpi->active_best_quality = cpi->best_quality;
+
+      /* Worst quality obviously must not be better than best quality */
+      if (cpi->active_worst_quality <= cpi->active_best_quality)
+        cpi->active_worst_quality = cpi->active_best_quality + 1;
+
+      if (cpi->active_worst_quality > 127) cpi->active_worst_quality = 127;
     }
+    /* Unbuffered mode (eg. video conferencing) */
+    else {
+      /* Set the active worst quality */
+      cpi->active_worst_quality = cpi->worst_quality;
+    }
 
-    /* Test to see if we have to drop a frame
-     * The auto-drop frame code is only used in buffered mode.
-     * In unbufferd mode (eg vide conferencing) the descision to
-     * code or drop a frame is made outside the codec in response to real
-     * world comms or buffer considerations.
+    /* Special trap for constrained quality mode
+     * "active_worst_quality" may never drop below cq level
+     * for any frame type.
      */
-    if (cpi->drop_frames_allowed &&
-        (cpi->oxcf.end_usage == USAGE_STREAM_FROM_SERVER) &&
-        ((cpi->common.frame_type != KEY_FRAME)))
-    {
-        /* Check for a buffer underun-crisis in which case we have to drop
-         * a frame
-         */
-        if ((cpi->buffer_level < 0))
-        {
+    if (cpi->oxcf.end_usage == USAGE_CONSTRAINED_QUALITY &&
+        cpi->active_worst_quality < cpi->cq_target_quality) {
+      cpi->active_worst_quality = cpi->cq_target_quality;
+    }
+  }
+
+  /* Test to see if we have to drop a frame
+   * The auto-drop frame code is only used in buffered mode.
+   * In unbufferd mode (eg vide conferencing) the descision to
+   * code or drop a frame is made outside the codec in response to real
+   * world comms or buffer considerations.
+   */
+  if (cpi->drop_frames_allowed &&
+      (cpi->oxcf.end_usage == USAGE_STREAM_FROM_SERVER) &&
+      ((cpi->common.frame_type != KEY_FRAME))) {
+    /* Check for a buffer underun-crisis in which case we have to drop
+     * a frame
+     */
+    if ((cpi->buffer_level < 0)) {
 #if 0
             FILE *f = fopen("dec.stt", "a");
             fprintf(f, "%10d %10d %10d %10d ***** BUFFER EMPTY\n",
@@ -949,64 +829,64 @@
                     (cpi->buffer_level * 100) / cpi->oxcf.optimal_buffer_level);
             fclose(f);
 #endif
-            cpi->drop_frame = 1;
+      cpi->drop_frame = 1;
 
-            /* Update the buffer level variable. */
-            cpi->bits_off_target += cpi->av_per_frame_bandwidth;
-            if (cpi->bits_off_target > cpi->oxcf.maximum_buffer_size)
-              cpi->bits_off_target = (int)cpi->oxcf.maximum_buffer_size;
-            cpi->buffer_level = cpi->bits_off_target;
+      /* Update the buffer level variable. */
+      cpi->bits_off_target += cpi->av_per_frame_bandwidth;
+      if (cpi->bits_off_target > cpi->oxcf.maximum_buffer_size)
+        cpi->bits_off_target = (int)cpi->oxcf.maximum_buffer_size;
+      cpi->buffer_level = cpi->bits_off_target;
 
-            if (cpi->oxcf.number_of_layers > 1) {
-              unsigned int i;
+      if (cpi->oxcf.number_of_layers > 1) {
+        unsigned int i;
 
-              // Propagate bits saved by dropping the frame to higher layers.
-              for (i = cpi->current_layer + 1; i < cpi->oxcf.number_of_layers;
-                  i++) {
-                LAYER_CONTEXT *lc = &cpi->layer_context[i];
-                lc->bits_off_target += (int)(lc->target_bandwidth /
-                                             lc->framerate);
-                if (lc->bits_off_target > lc->maximum_buffer_size)
-                  lc->bits_off_target = lc->maximum_buffer_size;
-                lc->buffer_level = lc->bits_off_target;
-              }
-            }
+        // Propagate bits saved by dropping the frame to higher layers.
+        for (i = cpi->current_layer + 1; i < cpi->oxcf.number_of_layers; i++) {
+          LAYER_CONTEXT *lc = &cpi->layer_context[i];
+          lc->bits_off_target += (int)(lc->target_bandwidth / lc->framerate);
+          if (lc->bits_off_target > lc->maximum_buffer_size)
+            lc->bits_off_target = lc->maximum_buffer_size;
+          lc->buffer_level = lc->bits_off_target;
         }
+      }
     }
+  }
 
-    /* Adjust target frame size for Golden Frames: */
-    if (cpi->oxcf.error_resilient_mode == 0 &&
-        (cpi->frames_till_gf_update_due == 0) && !cpi->drop_frame)
-    {
-        int Q = (cpi->oxcf.fixed_q < 0) ? cpi->last_q[INTER_FRAME] : cpi->oxcf.fixed_q;
+  /* Adjust target frame size for Golden Frames: */
+  if (cpi->oxcf.error_resilient_mode == 0 &&
+      (cpi->frames_till_gf_update_due == 0) && !cpi->drop_frame) {
+    int Q =
+        (cpi->oxcf.fixed_q < 0) ? cpi->last_q[INTER_FRAME] : cpi->oxcf.fixed_q;
 
-        int gf_frame_useage = 0;      /* Golden frame useage since last GF */
-        int tot_mbs = cpi->recent_ref_frame_usage[INTRA_FRAME]  +
-                      cpi->recent_ref_frame_usage[LAST_FRAME]   +
-                      cpi->recent_ref_frame_usage[GOLDEN_FRAME] +
-                      cpi->recent_ref_frame_usage[ALTREF_FRAME];
+    int gf_frame_useage = 0; /* Golden frame useage since last GF */
+    int tot_mbs = cpi->recent_ref_frame_usage[INTRA_FRAME] +
+                  cpi->recent_ref_frame_usage[LAST_FRAME] +
+                  cpi->recent_ref_frame_usage[GOLDEN_FRAME] +
+                  cpi->recent_ref_frame_usage[ALTREF_FRAME];
 
-        int pct_gf_active = (100 * cpi->gf_active_count) / (cpi->common.mb_rows * cpi->common.mb_cols);
+    int pct_gf_active = (100 * cpi->gf_active_count) /
+                        (cpi->common.mb_rows * cpi->common.mb_cols);
 
-        if (tot_mbs)
-            gf_frame_useage = (cpi->recent_ref_frame_usage[GOLDEN_FRAME] + cpi->recent_ref_frame_usage[ALTREF_FRAME]) * 100 / tot_mbs;
+    if (tot_mbs)
+      gf_frame_useage = (cpi->recent_ref_frame_usage[GOLDEN_FRAME] +
+                         cpi->recent_ref_frame_usage[ALTREF_FRAME]) *
+                        100 / tot_mbs;
 
-        if (pct_gf_active > gf_frame_useage)
-            gf_frame_useage = pct_gf_active;
+    if (pct_gf_active > gf_frame_useage) gf_frame_useage = pct_gf_active;
 
-        /* Is a fixed manual GF frequency being used */
-        if (cpi->auto_gold)
-        {
-            /* For one pass throw a GF if recent frame intra useage is
-             * low or the GF useage is high
-             */
-            if ((cpi->pass == 0) && (cpi->this_frame_percent_intra < 15 || gf_frame_useage >= 5))
-                cpi->common.refresh_golden_frame = 1;
+    /* Is a fixed manual GF frequency being used */
+    if (cpi->auto_gold) {
+      /* For one pass throw a GF if recent frame intra useage is
+       * low or the GF useage is high
+       */
+      if ((cpi->pass == 0) &&
+          (cpi->this_frame_percent_intra < 15 || gf_frame_useage >= 5))
+        cpi->common.refresh_golden_frame = 1;
 
-            /* Two pass GF descision */
-            else if (cpi->pass == 2)
-                cpi->common.refresh_golden_frame = 1;
-        }
+      /* Two pass GF descision */
+      else if (cpi->pass == 2)
+        cpi->common.refresh_golden_frame = 1;
+    }
 
 #if 0
 
@@ -1023,8 +903,7 @@
 
 #endif
 
-        if (cpi->common.refresh_golden_frame == 1)
-        {
+    if (cpi->common.refresh_golden_frame == 1) {
 #if 0
 
             if (0)
@@ -1038,533 +917,461 @@
 
 #endif
 
-            if (cpi->auto_adjust_gold_quantizer)
-            {
-                calc_gf_params(cpi);
-            }
+      if (cpi->auto_adjust_gold_quantizer) {
+        calc_gf_params(cpi);
+      }
 
-            /* If we are using alternate ref instead of gf then do not apply the
-             * boost It will instead be applied to the altref update Jims
-             * modified boost
+      /* If we are using alternate ref instead of gf then do not apply the
+       * boost It will instead be applied to the altref update Jims
+       * modified boost
+       */
+      if (!cpi->source_alt_ref_active) {
+        if (cpi->oxcf.fixed_q < 0) {
+          if (cpi->pass == 2) {
+            /* The spend on the GF is defined in the two pass
+             * code for two pass encodes
              */
-            if (!cpi->source_alt_ref_active)
-            {
-                if (cpi->oxcf.fixed_q < 0)
-                {
-                    if (cpi->pass == 2)
-                    {
-                        /* The spend on the GF is defined in the two pass
-                         * code for two pass encodes
-                         */
-                        cpi->this_frame_target = cpi->per_frame_bandwidth;
-                    }
-                    else
-                    {
-                        int Boost = cpi->last_boost;
-                        int frames_in_section = cpi->frames_till_gf_update_due + 1;
-                        int allocation_chunks = (frames_in_section * 100) + (Boost - 100);
-                        int bits_in_section = cpi->inter_frame_target * frames_in_section;
+            cpi->this_frame_target = cpi->per_frame_bandwidth;
+          } else {
+            int Boost = cpi->last_boost;
+            int frames_in_section = cpi->frames_till_gf_update_due + 1;
+            int allocation_chunks = (frames_in_section * 100) + (Boost - 100);
+            int bits_in_section = cpi->inter_frame_target * frames_in_section;
 
-                        /* Normalize Altboost and allocations chunck down to
-                         * prevent overflow
-                         */
-                        while (Boost > 1000)
-                        {
-                            Boost /= 2;
-                            allocation_chunks /= 2;
-                        }
-
-                        /* Avoid loss of precision but avoid overflow */
-                        if ((bits_in_section >> 7) > allocation_chunks)
-                            cpi->this_frame_target = Boost * (bits_in_section / allocation_chunks);
-                        else
-                            cpi->this_frame_target = (Boost * bits_in_section) / allocation_chunks;
-                    }
-                }
-                else
-                    cpi->this_frame_target =
-                        (estimate_bits_at_q(1, Q, cpi->common.MBs, 1.0)
-                         * cpi->last_boost) / 100;
-
-            }
-            /* If there is an active ARF at this location use the minimum
-             * bits on this frame even if it is a contructed arf.
-             * The active maximum quantizer insures that an appropriate
-             * number of bits will be spent if needed for contstructed ARFs.
+            /* Normalize Altboost and allocations chunck down to
+             * prevent overflow
              */
-            else
-            {
-                cpi->this_frame_target = 0;
+            while (Boost > 1000) {
+              Boost /= 2;
+              allocation_chunks /= 2;
             }
 
-            cpi->current_gf_interval = cpi->frames_till_gf_update_due;
+            /* Avoid loss of precision but avoid overflow */
+            if ((bits_in_section >> 7) > allocation_chunks)
+              cpi->this_frame_target =
+                  Boost * (bits_in_section / allocation_chunks);
+            else
+              cpi->this_frame_target =
+                  (Boost * bits_in_section) / allocation_chunks;
+          }
+        } else
+          cpi->this_frame_target =
+              (estimate_bits_at_q(1, Q, cpi->common.MBs, 1.0) *
+               cpi->last_boost) /
+              100;
 
-        }
+      }
+      /* If there is an active ARF at this location use the minimum
+       * bits on this frame even if it is a contructed arf.
+       * The active maximum quantizer insures that an appropriate
+       * number of bits will be spent if needed for contstructed ARFs.
+       */
+      else {
+        cpi->this_frame_target = 0;
+      }
+
+      cpi->current_gf_interval = cpi->frames_till_gf_update_due;
     }
+  }
 
-    cpi->per_frame_bandwidth = old_per_frame_bandwidth;
+  cpi->per_frame_bandwidth = old_per_frame_bandwidth;
 }
 
+void vp8_update_rate_correction_factors(VP8_COMP *cpi, int damp_var) {
+  int Q = cpi->common.base_qindex;
+  int correction_factor = 100;
+  double rate_correction_factor;
+  double adjustment_limit;
 
-void vp8_update_rate_correction_factors(VP8_COMP *cpi, int damp_var)
-{
-    int    Q = cpi->common.base_qindex;
-    int    correction_factor = 100;
-    double rate_correction_factor;
-    double adjustment_limit;
+  int projected_size_based_on_q = 0;
 
-    int    projected_size_based_on_q = 0;
+  /* Clear down mmx registers to allow floating point in what follows */
+  vp8_clear_system_state();
 
-    /* Clear down mmx registers to allow floating point in what follows */
-    vp8_clear_system_state();
-
-    if (cpi->common.frame_type == KEY_FRAME)
-    {
-        rate_correction_factor = cpi->key_frame_rate_correction_factor;
-    }
+  if (cpi->common.frame_type == KEY_FRAME) {
+    rate_correction_factor = cpi->key_frame_rate_correction_factor;
+  } else {
+    if (cpi->oxcf.number_of_layers == 1 &&
+        (cpi->common.refresh_alt_ref_frame || cpi->common.refresh_golden_frame))
+      rate_correction_factor = cpi->gf_rate_correction_factor;
     else
-    {
-        if (cpi->oxcf.number_of_layers == 1 &&
-           (cpi->common.refresh_alt_ref_frame ||
-            cpi->common.refresh_golden_frame))
-            rate_correction_factor = cpi->gf_rate_correction_factor;
-        else
-            rate_correction_factor = cpi->rate_correction_factor;
-    }
+      rate_correction_factor = cpi->rate_correction_factor;
+  }
 
-    /* Work out how big we would have expected the frame to be at this Q
-     * given the current correction factor. Stay in double to avoid int
-     * overflow when values are large
-     */
-    projected_size_based_on_q = (int)(((.5 + rate_correction_factor * vp8_bits_per_mb[cpi->common.frame_type][Q]) * cpi->common.MBs) / (1 << BPER_MB_NORMBITS));
+  /* Work out how big we would have expected the frame to be at this Q
+   * given the current correction factor. Stay in double to avoid int
+   * overflow when values are large
+   */
+  projected_size_based_on_q =
+      (int)(((.5 +
+              rate_correction_factor *
+                  vp8_bits_per_mb[cpi->common.frame_type][Q]) *
+             cpi->common.MBs) /
+            (1 << BPER_MB_NORMBITS));
 
-    /* Make some allowance for cpi->zbin_over_quant */
-    if (cpi->mb.zbin_over_quant > 0)
-    {
-        int Z = cpi->mb.zbin_over_quant;
-        double Factor = 0.99;
-        double factor_adjustment = 0.01 / 256.0;
+  /* Make some allowance for cpi->zbin_over_quant */
+  if (cpi->mb.zbin_over_quant > 0) {
+    int Z = cpi->mb.zbin_over_quant;
+    double Factor = 0.99;
+    double factor_adjustment = 0.01 / 256.0;
 
-        while (Z > 0)
-        {
-            Z --;
-            projected_size_based_on_q =
-                (int)(Factor * projected_size_based_on_q);
-            Factor += factor_adjustment;
+    while (Z > 0) {
+      Z--;
+      projected_size_based_on_q = (int)(Factor * projected_size_based_on_q);
+      Factor += factor_adjustment;
 
-            if (Factor  >= 0.999)
-                Factor = 0.999;
-        }
+      if (Factor >= 0.999) Factor = 0.999;
     }
+  }
 
-    /* Work out a size correction factor. */
-    if (projected_size_based_on_q > 0)
-        correction_factor = (100 * cpi->projected_frame_size) / projected_size_based_on_q;
+  /* Work out a size correction factor. */
+  if (projected_size_based_on_q > 0)
+    correction_factor =
+        (100 * cpi->projected_frame_size) / projected_size_based_on_q;
 
-    /* More heavily damped adjustment used if we have been oscillating
-     * either side of target
-     */
-    switch (damp_var)
-    {
-    case 0:
-        adjustment_limit = 0.75;
-        break;
-    case 1:
-        adjustment_limit = 0.375;
-        break;
+  /* More heavily damped adjustment used if we have been oscillating
+   * either side of target
+   */
+  switch (damp_var) {
+    case 0: adjustment_limit = 0.75; break;
+    case 1: adjustment_limit = 0.375; break;
     case 2:
-    default:
-        adjustment_limit = 0.25;
-        break;
-    }
+    default: adjustment_limit = 0.25; break;
+  }
 
-    if (correction_factor > 102)
-    {
-        /* We are not already at the worst allowable quality */
-        correction_factor = (int)(100.5 + ((correction_factor - 100) * adjustment_limit));
-        rate_correction_factor = ((rate_correction_factor * correction_factor) / 100);
+  if (correction_factor > 102) {
+    /* We are not already at the worst allowable quality */
+    correction_factor =
+        (int)(100.5 + ((correction_factor - 100) * adjustment_limit));
+    rate_correction_factor =
+        ((rate_correction_factor * correction_factor) / 100);
 
-        /* Keep rate_correction_factor within limits */
-        if (rate_correction_factor > MAX_BPB_FACTOR)
-            rate_correction_factor = MAX_BPB_FACTOR;
-    }
-    else if (correction_factor < 99)
-    {
-        /* We are not already at the best allowable quality */
-        correction_factor = (int)(100.5 - ((100 - correction_factor) * adjustment_limit));
-        rate_correction_factor = ((rate_correction_factor * correction_factor) / 100);
+    /* Keep rate_correction_factor within limits */
+    if (rate_correction_factor > MAX_BPB_FACTOR)
+      rate_correction_factor = MAX_BPB_FACTOR;
+  } else if (correction_factor < 99) {
+    /* We are not already at the best allowable quality */
+    correction_factor =
+        (int)(100.5 - ((100 - correction_factor) * adjustment_limit));
+    rate_correction_factor =
+        ((rate_correction_factor * correction_factor) / 100);
 
-        /* Keep rate_correction_factor within limits */
-        if (rate_correction_factor < MIN_BPB_FACTOR)
-            rate_correction_factor = MIN_BPB_FACTOR;
-    }
+    /* Keep rate_correction_factor within limits */
+    if (rate_correction_factor < MIN_BPB_FACTOR)
+      rate_correction_factor = MIN_BPB_FACTOR;
+  }
 
-    if (cpi->common.frame_type == KEY_FRAME)
-        cpi->key_frame_rate_correction_factor = rate_correction_factor;
+  if (cpi->common.frame_type == KEY_FRAME)
+    cpi->key_frame_rate_correction_factor = rate_correction_factor;
+  else {
+    if (cpi->oxcf.number_of_layers == 1 &&
+        (cpi->common.refresh_alt_ref_frame || cpi->common.refresh_golden_frame))
+      cpi->gf_rate_correction_factor = rate_correction_factor;
     else
-    {
-        if (cpi->oxcf.number_of_layers == 1 &&
-           (cpi->common.refresh_alt_ref_frame ||
-            cpi->common.refresh_golden_frame))
-            cpi->gf_rate_correction_factor = rate_correction_factor;
-        else
-            cpi->rate_correction_factor = rate_correction_factor;
-    }
+      cpi->rate_correction_factor = rate_correction_factor;
+  }
 }
 
+int vp8_regulate_q(VP8_COMP *cpi, int target_bits_per_frame) {
+  int Q = cpi->active_worst_quality;
 
-int vp8_regulate_q(VP8_COMP *cpi, int target_bits_per_frame)
-{
-    int Q = cpi->active_worst_quality;
+  if (cpi->force_maxqp == 1) {
+    cpi->active_worst_quality = cpi->worst_quality;
+    return cpi->worst_quality;
+  }
 
-    if (cpi->force_maxqp == 1) {
-      cpi->active_worst_quality = cpi->worst_quality;
-      return cpi->worst_quality;
-    }
+  /* Reset Zbin OQ value */
+  cpi->mb.zbin_over_quant = 0;
 
-    /* Reset Zbin OQ value */
-    cpi->mb.zbin_over_quant = 0;
+  if (cpi->oxcf.fixed_q >= 0) {
+    Q = cpi->oxcf.fixed_q;
 
-    if (cpi->oxcf.fixed_q >= 0)
-    {
-        Q = cpi->oxcf.fixed_q;
+    if (cpi->common.frame_type == KEY_FRAME) {
+      Q = cpi->oxcf.key_q;
+    } else if (cpi->oxcf.number_of_layers == 1 &&
+               cpi->common.refresh_alt_ref_frame) {
+      Q = cpi->oxcf.alt_q;
+    } else if (cpi->oxcf.number_of_layers == 1 &&
+               cpi->common.refresh_golden_frame) {
+      Q = cpi->oxcf.gold_q;
+    }
+  } else {
+    int i;
+    int last_error = INT_MAX;
+    int target_bits_per_mb;
+    int bits_per_mb_at_this_q;
+    double correction_factor;
 
-        if (cpi->common.frame_type == KEY_FRAME)
-        {
-            Q = cpi->oxcf.key_q;
-        }
-        else if (cpi->oxcf.number_of_layers == 1 &&
-            cpi->common.refresh_alt_ref_frame)
-        {
-            Q = cpi->oxcf.alt_q;
-        }
-        else if (cpi->oxcf.number_of_layers == 1  &&
-            cpi->common.refresh_golden_frame)
-        {
-            Q = cpi->oxcf.gold_q;
-        }
+    /* Select the appropriate correction factor based upon type of frame. */
+    if (cpi->common.frame_type == KEY_FRAME)
+      correction_factor = cpi->key_frame_rate_correction_factor;
+    else {
+      if (cpi->oxcf.number_of_layers == 1 &&
+          (cpi->common.refresh_alt_ref_frame ||
+           cpi->common.refresh_golden_frame))
+        correction_factor = cpi->gf_rate_correction_factor;
+      else
+        correction_factor = cpi->rate_correction_factor;
     }
+
+    /* Calculate required scaling factor based on target frame size and
+     * size of frame produced using previous Q
+     */
+    if (target_bits_per_frame >= (INT_MAX >> BPER_MB_NORMBITS))
+      /* Case where we would overflow int */
+      target_bits_per_mb = (target_bits_per_frame / cpi->common.MBs)
+                           << BPER_MB_NORMBITS;
     else
-    {
-        int i;
-        int last_error = INT_MAX;
-        int target_bits_per_mb;
-        int bits_per_mb_at_this_q;
-        double correction_factor;
+      target_bits_per_mb =
+          (target_bits_per_frame << BPER_MB_NORMBITS) / cpi->common.MBs;
 
-        /* Select the appropriate correction factor based upon type of frame. */
-        if (cpi->common.frame_type == KEY_FRAME)
-            correction_factor = cpi->key_frame_rate_correction_factor;
-        else
-        {
-            if (cpi->oxcf.number_of_layers == 1 &&
-               (cpi->common.refresh_alt_ref_frame ||
-                cpi->common.refresh_golden_frame))
-                correction_factor = cpi->gf_rate_correction_factor;
-            else
-                correction_factor = cpi->rate_correction_factor;
-        }
+    i = cpi->active_best_quality;
 
-        /* Calculate required scaling factor based on target frame size and
-         * size of frame produced using previous Q
-         */
-        if (target_bits_per_frame >= (INT_MAX >> BPER_MB_NORMBITS))
-            /* Case where we would overflow int */
-            target_bits_per_mb = (target_bits_per_frame / cpi->common.MBs) << BPER_MB_NORMBITS;
+    do {
+      bits_per_mb_at_this_q =
+          (int)(.5 +
+                correction_factor * vp8_bits_per_mb[cpi->common.frame_type][i]);
+
+      if (bits_per_mb_at_this_q <= target_bits_per_mb) {
+        if ((target_bits_per_mb - bits_per_mb_at_this_q) <= last_error)
+          Q = i;
         else
-            target_bits_per_mb = (target_bits_per_frame << BPER_MB_NORMBITS) / cpi->common.MBs;
+          Q = i - 1;
 
-        i = cpi->active_best_quality;
+        break;
+      } else
+        last_error = bits_per_mb_at_this_q - target_bits_per_mb;
+    } while (++i <= cpi->active_worst_quality);
 
-        do
-        {
-            bits_per_mb_at_this_q = (int)(.5 + correction_factor * vp8_bits_per_mb[cpi->common.frame_type][i]);
+    /* If we are at MAXQ then enable Q over-run which seeks to claw
+     * back additional bits through things like the RD multiplier
+     * and zero bin size.
+     */
+    if (Q >= MAXQ) {
+      int zbin_oqmax;
 
-            if (bits_per_mb_at_this_q <= target_bits_per_mb)
-            {
-                if ((target_bits_per_mb - bits_per_mb_at_this_q) <= last_error)
-                    Q = i;
-                else
-                    Q = i - 1;
+      double Factor = 0.99;
+      double factor_adjustment = 0.01 / 256.0;
 
-                break;
-            }
-            else
-                last_error = bits_per_mb_at_this_q - target_bits_per_mb;
-        }
-        while (++i <= cpi->active_worst_quality);
-
-
-        /* If we are at MAXQ then enable Q over-run which seeks to claw
-         * back additional bits through things like the RD multiplier
-         * and zero bin size.
-         */
-        if (Q >= MAXQ)
-        {
-            int zbin_oqmax;
-
-            double Factor = 0.99;
-            double factor_adjustment = 0.01 / 256.0;
-
-            if (cpi->common.frame_type == KEY_FRAME)
-                zbin_oqmax = 0;
-            else if (cpi->oxcf.number_of_layers == 1 &&
-                (cpi->common.refresh_alt_ref_frame ||
+      if (cpi->common.frame_type == KEY_FRAME)
+        zbin_oqmax = 0;
+      else if (cpi->oxcf.number_of_layers == 1 &&
+               (cpi->common.refresh_alt_ref_frame ||
                 (cpi->common.refresh_golden_frame &&
                  !cpi->source_alt_ref_active)))
-                zbin_oqmax = 16;
-            else
-                zbin_oqmax = ZBIN_OQ_MAX;
+        zbin_oqmax = 16;
+      else
+        zbin_oqmax = ZBIN_OQ_MAX;
 
-            /*{
-                double Factor = (double)target_bits_per_mb/(double)bits_per_mb_at_this_q;
-                double Oq;
+      /*{
+          double Factor =
+      (double)target_bits_per_mb/(double)bits_per_mb_at_this_q;
+          double Oq;
 
-                Factor = Factor/1.2683;
+          Factor = Factor/1.2683;
 
-                Oq = pow( Factor, (1.0/-0.165) );
+          Oq = pow( Factor, (1.0/-0.165) );
 
-                if ( Oq > zbin_oqmax )
-                    Oq = zbin_oqmax;
+          if ( Oq > zbin_oqmax )
+              Oq = zbin_oqmax;
 
-                cpi->zbin_over_quant = (int)Oq;
-            }*/
+          cpi->zbin_over_quant = (int)Oq;
+      }*/
 
-            /* Each incrment in the zbin is assumed to have a fixed effect
-             * on bitrate. This is not of course true. The effect will be
-             * highly clip dependent and may well have sudden steps. The
-             * idea here is to acheive higher effective quantizers than the
-             * normal maximum by expanding the zero bin and hence
-             * decreasing the number of low magnitude non zero coefficients.
-             */
-            while (cpi->mb.zbin_over_quant < zbin_oqmax)
-            {
-                cpi->mb.zbin_over_quant ++;
+      /* Each incrment in the zbin is assumed to have a fixed effect
+       * on bitrate. This is not of course true. The effect will be
+       * highly clip dependent and may well have sudden steps. The
+       * idea here is to acheive higher effective quantizers than the
+       * normal maximum by expanding the zero bin and hence
+       * decreasing the number of low magnitude non zero coefficients.
+       */
+      while (cpi->mb.zbin_over_quant < zbin_oqmax) {
+        cpi->mb.zbin_over_quant++;
 
-                if (cpi->mb.zbin_over_quant > zbin_oqmax)
-                    cpi->mb.zbin_over_quant = zbin_oqmax;
+        if (cpi->mb.zbin_over_quant > zbin_oqmax)
+          cpi->mb.zbin_over_quant = zbin_oqmax;
 
-                /* Adjust bits_per_mb_at_this_q estimate */
-                bits_per_mb_at_this_q = (int)(Factor * bits_per_mb_at_this_q);
-                Factor += factor_adjustment;
+        /* Adjust bits_per_mb_at_this_q estimate */
+        bits_per_mb_at_this_q = (int)(Factor * bits_per_mb_at_this_q);
+        Factor += factor_adjustment;
 
-                if (Factor  >= 0.999)
-                    Factor = 0.999;
+        if (Factor >= 0.999) Factor = 0.999;
 
-                /* Break out if we get down to the target rate */
-                if (bits_per_mb_at_this_q <= target_bits_per_mb)
-                    break;
-            }
-
-        }
+        /* Break out if we get down to the target rate */
+        if (bits_per_mb_at_this_q <= target_bits_per_mb) break;
+      }
     }
+  }
 
-    return Q;
+  return Q;
 }
 
+static int estimate_keyframe_frequency(VP8_COMP *cpi) {
+  int i;
 
-static int estimate_keyframe_frequency(VP8_COMP *cpi)
-{
-    int i;
+  /* Average key frame frequency */
+  int av_key_frame_frequency = 0;
 
-    /* Average key frame frequency */
-    int av_key_frame_frequency = 0;
-
-    /* First key frame at start of sequence is a special case. We have no
-     * frequency data.
+  /* First key frame at start of sequence is a special case. We have no
+   * frequency data.
+   */
+  if (cpi->key_frame_count == 1) {
+    /* Assume a default of 1 kf every 2 seconds, or the max kf interval,
+     * whichever is smaller.
      */
-    if (cpi->key_frame_count == 1)
-    {
-        /* Assume a default of 1 kf every 2 seconds, or the max kf interval,
-         * whichever is smaller.
-         */
-        int key_freq = cpi->oxcf.key_freq>0 ? cpi->oxcf.key_freq : 1;
-        av_key_frame_frequency = 1 + (int)cpi->output_framerate * 2;
+    int key_freq = cpi->oxcf.key_freq > 0 ? cpi->oxcf.key_freq : 1;
+    av_key_frame_frequency = 1 + (int)cpi->output_framerate * 2;
 
-        if (cpi->oxcf.auto_key && av_key_frame_frequency > key_freq)
-            av_key_frame_frequency = key_freq;
+    if (cpi->oxcf.auto_key && av_key_frame_frequency > key_freq)
+      av_key_frame_frequency = key_freq;
 
-        cpi->prior_key_frame_distance[KEY_FRAME_CONTEXT - 1]
-            = av_key_frame_frequency;
-    }
-    else
-    {
-        unsigned int total_weight = 0;
-        int last_kf_interval =
-                (cpi->frames_since_key > 0) ? cpi->frames_since_key : 1;
+    cpi->prior_key_frame_distance[KEY_FRAME_CONTEXT - 1] =
+        av_key_frame_frequency;
+  } else {
+    unsigned int total_weight = 0;
+    int last_kf_interval =
+        (cpi->frames_since_key > 0) ? cpi->frames_since_key : 1;
 
-        /* reset keyframe context and calculate weighted average of last
-         * KEY_FRAME_CONTEXT keyframes
-         */
-        for (i = 0; i < KEY_FRAME_CONTEXT; i++)
-        {
-            if (i < KEY_FRAME_CONTEXT - 1)
-                cpi->prior_key_frame_distance[i]
-                    = cpi->prior_key_frame_distance[i+1];
-            else
-                cpi->prior_key_frame_distance[i] = last_kf_interval;
+    /* reset keyframe context and calculate weighted average of last
+     * KEY_FRAME_CONTEXT keyframes
+     */
+    for (i = 0; i < KEY_FRAME_CONTEXT; i++) {
+      if (i < KEY_FRAME_CONTEXT - 1)
+        cpi->prior_key_frame_distance[i] = cpi->prior_key_frame_distance[i + 1];
+      else
+        cpi->prior_key_frame_distance[i] = last_kf_interval;
 
-            av_key_frame_frequency += prior_key_frame_weight[i]
-                                      * cpi->prior_key_frame_distance[i];
-            total_weight += prior_key_frame_weight[i];
-        }
-
-        av_key_frame_frequency  /= total_weight;
-
+      av_key_frame_frequency +=
+          prior_key_frame_weight[i] * cpi->prior_key_frame_distance[i];
+      total_weight += prior_key_frame_weight[i];
     }
-    // TODO (marpan): Given the checks above, |av_key_frame_frequency|
-    // should always be above 0. But for now we keep the sanity check in.
-    if (av_key_frame_frequency == 0)
-        av_key_frame_frequency = 1;
-    return av_key_frame_frequency;
+
+    av_key_frame_frequency /= total_weight;
+  }
+  // TODO (marpan): Given the checks above, |av_key_frame_frequency|
+  // should always be above 0. But for now we keep the sanity check in.
+  if (av_key_frame_frequency == 0) av_key_frame_frequency = 1;
+  return av_key_frame_frequency;
 }
 
+void vp8_adjust_key_frame_context(VP8_COMP *cpi) {
+  /* Clear down mmx registers to allow floating point in what follows */
+  vp8_clear_system_state();
 
-void vp8_adjust_key_frame_context(VP8_COMP *cpi)
-{
-    /* Clear down mmx registers to allow floating point in what follows */
-    vp8_clear_system_state();
+  /* Do we have any key frame overspend to recover? */
+  /* Two-pass overspend handled elsewhere. */
+  if ((cpi->pass != 2) &&
+      (cpi->projected_frame_size > cpi->per_frame_bandwidth)) {
+    int overspend;
 
-    /* Do we have any key frame overspend to recover? */
-    /* Two-pass overspend handled elsewhere. */
-    if ((cpi->pass != 2)
-         && (cpi->projected_frame_size > cpi->per_frame_bandwidth))
-    {
-        int overspend;
+    /* Update the count of key frame overspend to be recovered in
+     * subsequent frames. A portion of the KF overspend is treated as gf
+     * overspend (and hence recovered more quickly) as the kf is also a
+     * gf. Otherwise the few frames following each kf tend to get more
+     * bits allocated than those following other gfs.
+     */
+    overspend = (cpi->projected_frame_size - cpi->per_frame_bandwidth);
 
-        /* Update the count of key frame overspend to be recovered in
-         * subsequent frames. A portion of the KF overspend is treated as gf
-         * overspend (and hence recovered more quickly) as the kf is also a
-         * gf. Otherwise the few frames following each kf tend to get more
-         * bits allocated than those following other gfs.
-         */
-        overspend = (cpi->projected_frame_size - cpi->per_frame_bandwidth);
-
-        if (cpi->oxcf.number_of_layers > 1)
-            cpi->kf_overspend_bits += overspend;
-        else
-        {
-            cpi->kf_overspend_bits += overspend * 7 / 8;
-            cpi->gf_overspend_bits += overspend * 1 / 8;
-        }
-
-        /* Work out how much to try and recover per frame. */
-        cpi->kf_bitrate_adjustment = cpi->kf_overspend_bits
-                                     / estimate_keyframe_frequency(cpi);
+    if (cpi->oxcf.number_of_layers > 1)
+      cpi->kf_overspend_bits += overspend;
+    else {
+      cpi->kf_overspend_bits += overspend * 7 / 8;
+      cpi->gf_overspend_bits += overspend * 1 / 8;
     }
 
-    cpi->frames_since_key = 0;
-    cpi->key_frame_count++;
+    /* Work out how much to try and recover per frame. */
+    cpi->kf_bitrate_adjustment =
+        cpi->kf_overspend_bits / estimate_keyframe_frequency(cpi);
+  }
+
+  cpi->frames_since_key = 0;
+  cpi->key_frame_count++;
 }
 
-
-void vp8_compute_frame_size_bounds(VP8_COMP *cpi, int *frame_under_shoot_limit, int *frame_over_shoot_limit)
-{
-    /* Set-up bounds on acceptable frame size: */
-    if (cpi->oxcf.fixed_q >= 0)
-    {
-        /* Fixed Q scenario: frame size never outranges target
-         * (there is no target!)
-         */
-        *frame_under_shoot_limit = 0;
-        *frame_over_shoot_limit  = INT_MAX;
-    }
-    else
-    {
-        if (cpi->common.frame_type == KEY_FRAME)
-        {
-            *frame_over_shoot_limit  = cpi->this_frame_target * 9 / 8;
-            *frame_under_shoot_limit = cpi->this_frame_target * 7 / 8;
+void vp8_compute_frame_size_bounds(VP8_COMP *cpi, int *frame_under_shoot_limit,
+                                   int *frame_over_shoot_limit) {
+  /* Set-up bounds on acceptable frame size: */
+  if (cpi->oxcf.fixed_q >= 0) {
+    /* Fixed Q scenario: frame size never outranges target
+     * (there is no target!)
+     */
+    *frame_under_shoot_limit = 0;
+    *frame_over_shoot_limit = INT_MAX;
+  } else {
+    if (cpi->common.frame_type == KEY_FRAME) {
+      *frame_over_shoot_limit = cpi->this_frame_target * 9 / 8;
+      *frame_under_shoot_limit = cpi->this_frame_target * 7 / 8;
+    } else {
+      if (cpi->oxcf.number_of_layers > 1 || cpi->common.refresh_alt_ref_frame ||
+          cpi->common.refresh_golden_frame) {
+        *frame_over_shoot_limit = cpi->this_frame_target * 9 / 8;
+        *frame_under_shoot_limit = cpi->this_frame_target * 7 / 8;
+      } else {
+        /* For CBR take buffer fullness into account */
+        if (cpi->oxcf.end_usage == USAGE_STREAM_FROM_SERVER) {
+          if (cpi->buffer_level >= ((cpi->oxcf.optimal_buffer_level +
+                                     cpi->oxcf.maximum_buffer_size) >>
+                                    1)) {
+            /* Buffer is too full so relax overshoot and tighten
+             * undershoot
+             */
+            *frame_over_shoot_limit = cpi->this_frame_target * 12 / 8;
+            *frame_under_shoot_limit = cpi->this_frame_target * 6 / 8;
+          } else if (cpi->buffer_level <=
+                     (cpi->oxcf.optimal_buffer_level >> 1)) {
+            /* Buffer is too low so relax undershoot and tighten
+             * overshoot
+             */
+            *frame_over_shoot_limit = cpi->this_frame_target * 10 / 8;
+            *frame_under_shoot_limit = cpi->this_frame_target * 4 / 8;
+          } else {
+            *frame_over_shoot_limit = cpi->this_frame_target * 11 / 8;
+            *frame_under_shoot_limit = cpi->this_frame_target * 5 / 8;
+          }
         }
-        else
-        {
-            if (cpi->oxcf.number_of_layers > 1 ||
-                cpi->common.refresh_alt_ref_frame ||
-                cpi->common.refresh_golden_frame)
-            {
-                *frame_over_shoot_limit  = cpi->this_frame_target * 9 / 8;
-                *frame_under_shoot_limit = cpi->this_frame_target * 7 / 8;
-            }
-            else
-            {
-                /* For CBR take buffer fullness into account */
-                if (cpi->oxcf.end_usage == USAGE_STREAM_FROM_SERVER)
-                {
-                    if (cpi->buffer_level >= ((cpi->oxcf.optimal_buffer_level + cpi->oxcf.maximum_buffer_size) >> 1))
-                    {
-                        /* Buffer is too full so relax overshoot and tighten
-                         * undershoot
-                         */
-                        *frame_over_shoot_limit  = cpi->this_frame_target * 12 / 8;
-                        *frame_under_shoot_limit = cpi->this_frame_target * 6 / 8;
-                    }
-                    else if (cpi->buffer_level <= (cpi->oxcf.optimal_buffer_level >> 1))
-                    {
-                        /* Buffer is too low so relax undershoot and tighten
-                         * overshoot
-                         */
-                        *frame_over_shoot_limit  = cpi->this_frame_target * 10 / 8;
-                        *frame_under_shoot_limit = cpi->this_frame_target * 4 / 8;
-                    }
-                    else
-                    {
-                        *frame_over_shoot_limit  = cpi->this_frame_target * 11 / 8;
-                        *frame_under_shoot_limit = cpi->this_frame_target * 5 / 8;
-                    }
-                }
-                /* VBR and CQ mode */
-                /* Note that tighter restrictions here can help quality
-                 * but hurt encode speed
-                 */
-                else
-                {
-                    /* Stron overshoot limit for constrained quality */
-                    if (cpi->oxcf.end_usage == USAGE_CONSTRAINED_QUALITY)
-                    {
-                        *frame_over_shoot_limit  = cpi->this_frame_target * 11 / 8;
-                        *frame_under_shoot_limit = cpi->this_frame_target * 2 / 8;
-                    }
-                    else
-                    {
-                        *frame_over_shoot_limit  = cpi->this_frame_target * 11 / 8;
-                        *frame_under_shoot_limit = cpi->this_frame_target * 5 / 8;
-                    }
-                }
-            }
+        /* VBR and CQ mode */
+        /* Note that tighter restrictions here can help quality
+         * but hurt encode speed
+         */
+        else {
+          /* Stron overshoot limit for constrained quality */
+          if (cpi->oxcf.end_usage == USAGE_CONSTRAINED_QUALITY) {
+            *frame_over_shoot_limit = cpi->this_frame_target * 11 / 8;
+            *frame_under_shoot_limit = cpi->this_frame_target * 2 / 8;
+          } else {
+            *frame_over_shoot_limit = cpi->this_frame_target * 11 / 8;
+            *frame_under_shoot_limit = cpi->this_frame_target * 5 / 8;
+          }
         }
-
-        /* For very small rate targets where the fractional adjustment
-         * (eg * 7/8) may be tiny make sure there is at least a minimum
-         * range.
-         */
-        *frame_over_shoot_limit += 200;
-        *frame_under_shoot_limit -= 200;
-        if ( *frame_under_shoot_limit < 0 )
-            *frame_under_shoot_limit = 0;
-
+      }
     }
+
+    /* For very small rate targets where the fractional adjustment
+     * (eg * 7/8) may be tiny make sure there is at least a minimum
+     * range.
+     */
+    *frame_over_shoot_limit += 200;
+    *frame_under_shoot_limit -= 200;
+    if (*frame_under_shoot_limit < 0) *frame_under_shoot_limit = 0;
+  }
 }
 
-
 /* return of 0 means drop frame */
-int vp8_pick_frame_size(VP8_COMP *cpi)
-{
-    VP8_COMMON *cm = &cpi->common;
+int vp8_pick_frame_size(VP8_COMP *cpi) {
+  VP8_COMMON *cm = &cpi->common;
 
-    if (cm->frame_type == KEY_FRAME)
-        calc_iframe_target_size(cpi);
-    else
-    {
-        calc_pframe_target_size(cpi);
+  if (cm->frame_type == KEY_FRAME)
+    calc_iframe_target_size(cpi);
+  else {
+    calc_pframe_target_size(cpi);
 
-        /* Check if we're dropping the frame: */
-        if (cpi->drop_frame)
-        {
-            cpi->drop_frame = 0;
-            return 0;
-        }
+    /* Check if we're dropping the frame: */
+    if (cpi->drop_frame) {
+      cpi->drop_frame = 0;
+      return 0;
     }
-    return 1;
+  }
+  return 1;
 }
 // If this just encoded frame (mcomp/transform/quant, but before loopfilter and
 // pack_bitstream) has large overshoot, and was not being encoded close to the
@@ -1573,10 +1380,8 @@
 // TODO(marpan): Should do this exit condition during the encode_frame
 // (i.e., halfway during the encoding of the frame) to save cycles.
 int vp8_drop_encodedframe_overshoot(VP8_COMP *cpi, int Q) {
-  if (cpi->pass == 0 &&
-      cpi->oxcf.end_usage == USAGE_STREAM_FROM_SERVER &&
-      cpi->drop_frames_allowed == 0 &&
-      cpi->common.frame_type != KEY_FRAME) {
+  if (cpi->pass == 0 && cpi->oxcf.end_usage == USAGE_STREAM_FROM_SERVER &&
+      cpi->drop_frames_allowed == 0 && cpi->common.frame_type != KEY_FRAME) {
     // Note: the "projected_frame_size" from encode_frame() only gives estimate
     // of mode/motion vector rate (in non-rd mode): so below we only require
     // that projected_frame_size is somewhat greater than per-frame-bandwidth,
@@ -1590,8 +1395,7 @@
     // residual error over 16x16 block. Should add QP dependence on threshold?
     int thresh_pred_err_mb = (256 << 4);
     int pred_err_mb = (int)(cpi->mb.prediction_error / cpi->common.MBs);
-    if (Q < thresh_qp &&
-        cpi->projected_frame_size > thresh_rate &&
+    if (Q < thresh_qp && cpi->projected_frame_size > thresh_rate &&
         pred_err_mb > thresh_pred_err_mb) {
       double new_correction_factor;
       const int target_size = cpi->av_per_frame_bandwidth;
@@ -1612,13 +1416,14 @@
       // frame because the QP/rate_correction_factor may have been too low
       // before the drop and then takes too long to come up.
       if (target_size >= (INT_MAX >> BPER_MB_NORMBITS))
-        target_bits_per_mb =
-            (target_size / cpi->common.MBs) << BPER_MB_NORMBITS;
+        target_bits_per_mb = (target_size / cpi->common.MBs)
+                             << BPER_MB_NORMBITS;
       else
         target_bits_per_mb =
             (target_size << BPER_MB_NORMBITS) / cpi->common.MBs;
       // Rate correction factor based on target_size_per_mb and max_QP.
-      new_correction_factor = (double)target_bits_per_mb /
+      new_correction_factor =
+          (double)target_bits_per_mb /
           (double)vp8_bits_per_mb[INTER_FRAME][cpi->worst_quality];
       if (new_correction_factor > cpi->rate_correction_factor)
         cpi->rate_correction_factor =
--- a/vp8/encoder/ratectrl.h
+++ b/vp8/encoder/ratectrl.h
@@ -8,7 +8,6 @@
  *  be found in the AUTHORS file in the root of the source tree.
  */
 
-
 #ifndef VP8_ENCODER_RATECTRL_H_
 #define VP8_ENCODER_RATECTRL_H_
 
@@ -25,7 +24,9 @@
 extern void vp8_update_rate_correction_factors(VP8_COMP *cpi, int damp_var);
 extern int vp8_regulate_q(VP8_COMP *cpi, int target_bits_per_frame);
 extern void vp8_adjust_key_frame_context(VP8_COMP *cpi);
-extern void vp8_compute_frame_size_bounds(VP8_COMP *cpi, int *frame_under_shoot_limit, int *frame_over_shoot_limit);
+extern void vp8_compute_frame_size_bounds(VP8_COMP *cpi,
+                                          int *frame_under_shoot_limit,
+                                          int *frame_over_shoot_limit);
 
 /* return of 0 means drop frame */
 extern int vp8_pick_frame_size(VP8_COMP *cpi);
--- a/vp8/encoder/rdopt.c
+++ b/vp8/encoder/rdopt.c
@@ -8,7 +8,6 @@
  *  be found in the AUTHORS file in the root of the source tree.
  */
 
-
 #include <stdio.h>
 #include <math.h>
 #include <limits.h>
@@ -40,19 +39,17 @@
 #endif
 extern void vp8_update_zbin_extra(VP8_COMP *cpi, MACROBLOCK *x);
 
-#define MAXF(a,b)            (((a) > (b)) ? (a) : (b))
+#define MAXF(a, b) (((a) > (b)) ? (a) : (b))
 
-typedef struct rate_distortion_struct
-{
-    int rate2;
-    int rate_y;
-    int rate_uv;
-    int distortion2;
-    int distortion_uv;
+typedef struct rate_distortion_struct {
+  int rate2;
+  int rate_y;
+  int rate_uv;
+  int distortion2;
+  int distortion_uv;
 } RATE_DISTORTION;
 
-typedef struct best_mode_struct
-{
+typedef struct best_mode_struct {
   int yrd;
   int rd;
   int intra_rd;
@@ -61,290 +58,202 @@
   PARTITION_INFO partition;
 } BEST_MODE;
 
-static const int auto_speed_thresh[17] =
-{
-    1000,
-    200,
-    150,
-    130,
-    150,
-    125,
-    120,
-    115,
-    115,
-    115,
-    115,
-    115,
-    115,
-    115,
-    115,
-    115,
-    105
-};
+static const int auto_speed_thresh[17] = { 1000, 200, 150, 130, 150, 125,
+                                           120,  115, 115, 115, 115, 115,
+                                           115,  115, 115, 115, 105 };
 
-const MB_PREDICTION_MODE vp8_mode_order[MAX_MODES] =
-{
-    ZEROMV,
-    DC_PRED,
+const MB_PREDICTION_MODE vp8_mode_order[MAX_MODES] = {
+  ZEROMV,    DC_PRED,
 
-    NEARESTMV,
-    NEARMV,
+  NEARESTMV, NEARMV,
 
-    ZEROMV,
-    NEARESTMV,
+  ZEROMV,    NEARESTMV,
 
-    ZEROMV,
-    NEARESTMV,
+  ZEROMV,    NEARESTMV,
 
-    NEARMV,
-    NEARMV,
+  NEARMV,    NEARMV,
 
-    V_PRED,
-    H_PRED,
-    TM_PRED,
+  V_PRED,    H_PRED,    TM_PRED,
 
-    NEWMV,
-    NEWMV,
-    NEWMV,
+  NEWMV,     NEWMV,     NEWMV,
 
-    SPLITMV,
-    SPLITMV,
-    SPLITMV,
+  SPLITMV,   SPLITMV,   SPLITMV,
 
-    B_PRED,
+  B_PRED,
 };
 
 /* This table determines the search order in reference frame priority order,
  * which may not necessarily match INTRA,LAST,GOLDEN,ARF
  */
-const int vp8_ref_frame_order[MAX_MODES] =
-{
-    1,
-    0,
+const int vp8_ref_frame_order[MAX_MODES] = {
+  1, 0,
 
-    1,
-    1,
+  1, 1,
 
-    2,
-    2,
+  2, 2,
 
-    3,
-    3,
+  3, 3,
 
-    2,
-    3,
+  2, 3,
 
-    0,
-    0,
-    0,
+  0, 0, 0,
 
-    1,
-    2,
-    3,
+  1, 2, 3,
 
-    1,
-    2,
-    3,
+  1, 2, 3,
 
-    0,
+  0,
 };
 
 static void fill_token_costs(
     int c[BLOCK_TYPES][COEF_BANDS][PREV_COEF_CONTEXTS][MAX_ENTROPY_TOKENS],
-    const vp8_prob p[BLOCK_TYPES][COEF_BANDS][PREV_COEF_CONTEXTS][ENTROPY_NODES]
-)
-{
-    int i, j, k;
+    const vp8_prob
+        p[BLOCK_TYPES][COEF_BANDS][PREV_COEF_CONTEXTS][ENTROPY_NODES]) {
+  int i, j, k;
 
+  for (i = 0; i < BLOCK_TYPES; i++)
+    for (j = 0; j < COEF_BANDS; j++)
+      for (k = 0; k < PREV_COEF_CONTEXTS; k++)
 
-    for (i = 0; i < BLOCK_TYPES; i++)
-        for (j = 0; j < COEF_BANDS; j++)
-            for (k = 0; k < PREV_COEF_CONTEXTS; k++)
-
-                /* check for pt=0 and band > 1 if block type 0
-                 * and 0 if blocktype 1
-                 */
-                if (k == 0 && j > (i == 0))
-                    vp8_cost_tokens2(c[i][j][k], p [i][j][k], vp8_coef_tree, 2);
-                else
-                    vp8_cost_tokens(c[i][j][k], p [i][j][k], vp8_coef_tree);
+        /* check for pt=0 and band > 1 if block type 0
+         * and 0 if blocktype 1
+         */
+        if (k == 0 && j > (i == 0))
+          vp8_cost_tokens2(c[i][j][k], p[i][j][k], vp8_coef_tree, 2);
+        else
+          vp8_cost_tokens(c[i][j][k], p[i][j][k], vp8_coef_tree);
 }
 
-static const int rd_iifactor[32] =
-{
-    4, 4, 3, 2, 1, 0, 0, 0,
-    0, 0, 0, 0, 0, 0, 0, 0,
-    0, 0, 0, 0, 0, 0, 0, 0,
-    0, 0, 0, 0, 0, 0, 0, 0
-};
+static const int rd_iifactor[32] = { 4, 4, 3, 2, 1, 0, 0, 0, 0, 0, 0,
+                                     0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
+                                     0, 0, 0, 0, 0, 0, 0, 0, 0, 0 };
 
 /* values are now correlated to quantizer */
-static const int sad_per_bit16lut[QINDEX_RANGE] =
-{
-    2,  2,  2,  2,  2,  2,  2,  2,
-    2,  2,  2,  2,  2,  2,  2,  2,
-    3,  3,  3,  3,  3,  3,  3,  3,
-    3,  3,  3,  3,  3,  3,  4,  4,
-    4,  4,  4,  4,  4,  4,  4,  4,
-    4,  4,  5,  5,  5,  5,  5,  5,
-    5,  5,  5,  5,  5,  5,  6,  6,
-    6,  6,  6,  6,  6,  6,  6,  6,
-    6,  6,  7,  7,  7,  7,  7,  7,
-    7,  7,  7,  7,  7,  7,  8,  8,
-    8,  8,  8,  8,  8,  8,  8,  8,
-    8,  8,  9,  9,  9,  9,  9,  9,
-    9,  9,  9,  9,  9,  9,  10, 10,
-    10, 10, 10, 10, 10, 10, 11, 11,
-    11, 11, 11, 11, 12, 12, 12, 12,
-    12, 12, 13, 13, 13, 13, 14, 14
+static const int sad_per_bit16lut[QINDEX_RANGE] = {
+  2,  2,  2,  2,  2,  2,  2,  2,  2,  2,  2,  2,  2,  2,  2,  2,  3,  3,  3,
+  3,  3,  3,  3,  3,  3,  3,  3,  3,  3,  3,  4,  4,  4,  4,  4,  4,  4,  4,
+  4,  4,  4,  4,  5,  5,  5,  5,  5,  5,  5,  5,  5,  5,  5,  5,  6,  6,  6,
+  6,  6,  6,  6,  6,  6,  6,  6,  6,  7,  7,  7,  7,  7,  7,  7,  7,  7,  7,
+  7,  7,  8,  8,  8,  8,  8,  8,  8,  8,  8,  8,  8,  8,  9,  9,  9,  9,  9,
+  9,  9,  9,  9,  9,  9,  9,  10, 10, 10, 10, 10, 10, 10, 10, 11, 11, 11, 11,
+  11, 11, 12, 12, 12, 12, 12, 12, 13, 13, 13, 13, 14, 14
 };
-static const int sad_per_bit4lut[QINDEX_RANGE] =
-{
-    2,  2,  2,  2,  2,  2,  3,  3,
-    3,  3,  3,  3,  3,  3,  3,  3,
-    3,  3,  3,  3,  4,  4,  4,  4,
-    4,  4,  4,  4,  4,  4,  5,  5,
-    5,  5,  5,  5,  6,  6,  6,  6,
-    6,  6,  6,  6,  6,  6,  6,  6,
-    7,  7,  7,  7,  7,  7,  7,  7,
-    7,  7,  7,  7,  7,  8,  8,  8,
-    8,  8,  9,  9,  9,  9,  9,  9,
-    10, 10, 10, 10, 10, 10, 10, 10,
-    11, 11, 11, 11, 11, 11, 11, 11,
-    12, 12, 12, 12, 12, 12, 12, 12,
-    13, 13, 13, 13, 13, 13, 13, 14,
-    14, 14, 14, 14, 15, 15, 15, 15,
-    16, 16, 16, 16, 17, 17, 17, 18,
-    18, 18, 19, 19, 19, 20, 20, 20,
+static const int sad_per_bit4lut[QINDEX_RANGE] = {
+  2,  2,  2,  2,  2,  2,  3,  3,  3,  3,  3,  3,  3,  3,  3,  3,  3,  3,  3,
+  3,  4,  4,  4,  4,  4,  4,  4,  4,  4,  4,  5,  5,  5,  5,  5,  5,  6,  6,
+  6,  6,  6,  6,  6,  6,  6,  6,  6,  6,  7,  7,  7,  7,  7,  7,  7,  7,  7,
+  7,  7,  7,  7,  8,  8,  8,  8,  8,  9,  9,  9,  9,  9,  9,  10, 10, 10, 10,
+  10, 10, 10, 10, 11, 11, 11, 11, 11, 11, 11, 11, 12, 12, 12, 12, 12, 12, 12,
+  12, 13, 13, 13, 13, 13, 13, 13, 14, 14, 14, 14, 14, 15, 15, 15, 15, 16, 16,
+  16, 16, 17, 17, 17, 18, 18, 18, 19, 19, 19, 20, 20, 20,
 };
 
-void vp8cx_initialize_me_consts(VP8_COMP *cpi, int QIndex)
-{
-    cpi->mb.sadperbit16 =  sad_per_bit16lut[QIndex];
-    cpi->mb.sadperbit4  =  sad_per_bit4lut[QIndex];
+void vp8cx_initialize_me_consts(VP8_COMP *cpi, int QIndex) {
+  cpi->mb.sadperbit16 = sad_per_bit16lut[QIndex];
+  cpi->mb.sadperbit4 = sad_per_bit4lut[QIndex];
 }
 
-void vp8_initialize_rd_consts(VP8_COMP *cpi, MACROBLOCK *x, int Qvalue)
-{
-    int q;
-    int i;
-    double capped_q = (Qvalue < 160) ? (double)Qvalue : 160.0;
-    double rdconst = 2.80;
+void vp8_initialize_rd_consts(VP8_COMP *cpi, MACROBLOCK *x, int Qvalue) {
+  int q;
+  int i;
+  double capped_q = (Qvalue < 160) ? (double)Qvalue : 160.0;
+  double rdconst = 2.80;
 
-    vp8_clear_system_state();
+  vp8_clear_system_state();
 
-    /* Further tests required to see if optimum is different
-     * for key frames, golden frames and arf frames.
-     */
-    cpi->RDMULT = (int)(rdconst * (capped_q * capped_q));
+  /* Further tests required to see if optimum is different
+   * for key frames, golden frames and arf frames.
+   */
+  cpi->RDMULT = (int)(rdconst * (capped_q * capped_q));
 
-    /* Extend rate multiplier along side quantizer zbin increases */
-    if (cpi->mb.zbin_over_quant  > 0)
-    {
-        double oq_factor;
-        double modq;
+  /* Extend rate multiplier along side quantizer zbin increases */
+  if (cpi->mb.zbin_over_quant > 0) {
+    double oq_factor;
+    double modq;
 
-        /* Experimental code using the same basic equation as used for Q above
-         * The units of cpi->mb.zbin_over_quant are 1/128 of Q bin size
-         */
-        oq_factor = 1.0 + ((double)0.0015625 * cpi->mb.zbin_over_quant);
-        modq = (int)((double)capped_q * oq_factor);
-        cpi->RDMULT = (int)(rdconst * (modq * modq));
-    }
+    /* Experimental code using the same basic equation as used for Q above
+     * The units of cpi->mb.zbin_over_quant are 1/128 of Q bin size
+     */
+    oq_factor = 1.0 + ((double)0.0015625 * cpi->mb.zbin_over_quant);
+    modq = (int)((double)capped_q * oq_factor);
+    cpi->RDMULT = (int)(rdconst * (modq * modq));
+  }
 
-    if (cpi->pass == 2 && (cpi->common.frame_type != KEY_FRAME))
-    {
-        if (cpi->twopass.next_iiratio > 31)
-            cpi->RDMULT += (cpi->RDMULT * rd_iifactor[31]) >> 4;
-        else
-            cpi->RDMULT +=
-                (cpi->RDMULT * rd_iifactor[cpi->twopass.next_iiratio]) >> 4;
-    }
+  if (cpi->pass == 2 && (cpi->common.frame_type != KEY_FRAME)) {
+    if (cpi->twopass.next_iiratio > 31)
+      cpi->RDMULT += (cpi->RDMULT * rd_iifactor[31]) >> 4;
+    else
+      cpi->RDMULT +=
+          (cpi->RDMULT * rd_iifactor[cpi->twopass.next_iiratio]) >> 4;
+  }
 
-    cpi->mb.errorperbit = (cpi->RDMULT / 110);
-    cpi->mb.errorperbit += (cpi->mb.errorperbit==0);
+  cpi->mb.errorperbit = (cpi->RDMULT / 110);
+  cpi->mb.errorperbit += (cpi->mb.errorperbit == 0);
 
-    vp8_set_speed_features(cpi);
+  vp8_set_speed_features(cpi);
 
-    for (i = 0; i < MAX_MODES; i++)
-    {
-        x->mode_test_hit_counts[i] = 0;
-    }
+  for (i = 0; i < MAX_MODES; i++) {
+    x->mode_test_hit_counts[i] = 0;
+  }
 
-    q = (int)pow(Qvalue, 1.25);
+  q = (int)pow(Qvalue, 1.25);
 
-    if (q < 8)
-        q = 8;
+  if (q < 8) q = 8;
 
-    if (cpi->RDMULT > 1000)
-    {
-        cpi->RDDIV = 1;
-        cpi->RDMULT /= 100;
+  if (cpi->RDMULT > 1000) {
+    cpi->RDDIV = 1;
+    cpi->RDMULT /= 100;
 
-        for (i = 0; i < MAX_MODES; i++)
-        {
-            if (cpi->sf.thresh_mult[i] < INT_MAX)
-            {
-                x->rd_threshes[i] = cpi->sf.thresh_mult[i] * q / 100;
-            }
-            else
-            {
-                x->rd_threshes[i] = INT_MAX;
-            }
+    for (i = 0; i < MAX_MODES; i++) {
+      if (cpi->sf.thresh_mult[i] < INT_MAX) {
+        x->rd_threshes[i] = cpi->sf.thresh_mult[i] * q / 100;
+      } else {
+        x->rd_threshes[i] = INT_MAX;
+      }
 
-            cpi->rd_baseline_thresh[i] = x->rd_threshes[i];
-        }
+      cpi->rd_baseline_thresh[i] = x->rd_threshes[i];
     }
-    else
-    {
-        cpi->RDDIV = 100;
+  } else {
+    cpi->RDDIV = 100;
 
-        for (i = 0; i < MAX_MODES; i++)
-        {
-            if (cpi->sf.thresh_mult[i] < (INT_MAX / q))
-            {
-                x->rd_threshes[i] = cpi->sf.thresh_mult[i] * q;
-            }
-            else
-            {
-                x->rd_threshes[i] = INT_MAX;
-            }
+    for (i = 0; i < MAX_MODES; i++) {
+      if (cpi->sf.thresh_mult[i] < (INT_MAX / q)) {
+        x->rd_threshes[i] = cpi->sf.thresh_mult[i] * q;
+      } else {
+        x->rd_threshes[i] = INT_MAX;
+      }
 
-            cpi->rd_baseline_thresh[i] = x->rd_threshes[i];
-        }
+      cpi->rd_baseline_thresh[i] = x->rd_threshes[i];
     }
+  }
 
-    {
-      /* build token cost array for the type of frame we have now */
-      FRAME_CONTEXT *l = &cpi->lfc_n;
+  {
+    /* build token cost array for the type of frame we have now */
+    FRAME_CONTEXT *l = &cpi->lfc_n;
 
-      if(cpi->common.refresh_alt_ref_frame)
-          l = &cpi->lfc_a;
-      else if(cpi->common.refresh_golden_frame)
-          l = &cpi->lfc_g;
+    if (cpi->common.refresh_alt_ref_frame)
+      l = &cpi->lfc_a;
+    else if (cpi->common.refresh_golden_frame)
+      l = &cpi->lfc_g;
 
-      fill_token_costs(
-          cpi->mb.token_costs,
-          (const vp8_prob( *)[8][3][11]) l->coef_probs
-      );
-      /*
-      fill_token_costs(
-          cpi->mb.token_costs,
-          (const vp8_prob( *)[8][3][11]) cpi->common.fc.coef_probs);
-      */
+    fill_token_costs(cpi->mb.token_costs,
+                     (const vp8_prob(*)[8][3][11])l->coef_probs);
+    /*
+    fill_token_costs(
+        cpi->mb.token_costs,
+        (const vp8_prob( *)[8][3][11]) cpi->common.fc.coef_probs);
+    */
 
-
-      /* TODO make these mode costs depend on last,alt or gold too.  (jbb) */
-      vp8_init_mode_costs(cpi);
-    }
-
+    /* TODO make these mode costs depend on last,alt or gold too.  (jbb) */
+    vp8_init_mode_costs(cpi);
+  }
 }
 
-void vp8_auto_select_speed(VP8_COMP *cpi)
-{
-    int milliseconds_for_compress = (int)(1000000 / cpi->framerate);
+void vp8_auto_select_speed(VP8_COMP *cpi) {
+  int milliseconds_for_compress = (int)(1000000 / cpi->framerate);
 
-    milliseconds_for_compress = milliseconds_for_compress * (16 - cpi->oxcf.cpu_used) / 16;
+  milliseconds_for_compress =
+      milliseconds_for_compress * (16 - cpi->oxcf.cpu_used) / 16;
 
 #if 0
 
@@ -360,728 +269,641 @@
 
 #endif
 
-    if (cpi->avg_pick_mode_time < milliseconds_for_compress && (cpi->avg_encode_time - cpi->avg_pick_mode_time) < milliseconds_for_compress)
-    {
-        if (cpi->avg_pick_mode_time == 0)
-        {
-            cpi->Speed = 4;
+  if (cpi->avg_pick_mode_time < milliseconds_for_compress &&
+      (cpi->avg_encode_time - cpi->avg_pick_mode_time) <
+          milliseconds_for_compress) {
+    if (cpi->avg_pick_mode_time == 0) {
+      cpi->Speed = 4;
+    } else {
+      if (milliseconds_for_compress * 100 < cpi->avg_encode_time * 95) {
+        cpi->Speed += 2;
+        cpi->avg_pick_mode_time = 0;
+        cpi->avg_encode_time = 0;
+
+        if (cpi->Speed > 16) {
+          cpi->Speed = 16;
         }
-        else
-        {
-            if (milliseconds_for_compress * 100 < cpi->avg_encode_time * 95)
-            {
-                cpi->Speed          += 2;
-                cpi->avg_pick_mode_time = 0;
-                cpi->avg_encode_time = 0;
+      }
 
-                if (cpi->Speed > 16)
-                {
-                    cpi->Speed = 16;
-                }
-            }
+      if (milliseconds_for_compress * 100 >
+          cpi->avg_encode_time * auto_speed_thresh[cpi->Speed]) {
+        cpi->Speed -= 1;
+        cpi->avg_pick_mode_time = 0;
+        cpi->avg_encode_time = 0;
 
-            if (milliseconds_for_compress * 100 > cpi->avg_encode_time * auto_speed_thresh[cpi->Speed])
-            {
-                cpi->Speed          -= 1;
-                cpi->avg_pick_mode_time = 0;
-                cpi->avg_encode_time = 0;
-
-                /* In real-time mode, cpi->speed is in [4, 16]. */
-                if (cpi->Speed < 4)
-                {
-                    cpi->Speed = 4;
-                }
-            }
+        /* In real-time mode, cpi->speed is in [4, 16]. */
+        if (cpi->Speed < 4) {
+          cpi->Speed = 4;
         }
+      }
     }
-    else
-    {
-        cpi->Speed += 4;
+  } else {
+    cpi->Speed += 4;
 
-        if (cpi->Speed > 16)
-            cpi->Speed = 16;
+    if (cpi->Speed > 16) cpi->Speed = 16;
 
-
-        cpi->avg_pick_mode_time = 0;
-        cpi->avg_encode_time = 0;
-    }
+    cpi->avg_pick_mode_time = 0;
+    cpi->avg_encode_time = 0;
+  }
 }
 
-int vp8_block_error_c(short *coeff, short *dqcoeff)
-{
-    int i;
-    int error = 0;
+int vp8_block_error_c(short *coeff, short *dqcoeff) {
+  int i;
+  int error = 0;
 
-    for (i = 0; i < 16; i++)
-    {
-        int this_diff = coeff[i] - dqcoeff[i];
-        error += this_diff * this_diff;
-    }
+  for (i = 0; i < 16; i++) {
+    int this_diff = coeff[i] - dqcoeff[i];
+    error += this_diff * this_diff;
+  }
 
-    return error;
+  return error;
 }
 
-int vp8_mbblock_error_c(MACROBLOCK *mb, int dc)
-{
-    BLOCK  *be;
-    BLOCKD *bd;
-    int i, j;
-    int berror, error = 0;
+int vp8_mbblock_error_c(MACROBLOCK *mb, int dc) {
+  BLOCK *be;
+  BLOCKD *bd;
+  int i, j;
+  int berror, error = 0;
 
-    for (i = 0; i < 16; i++)
-    {
-        be = &mb->block[i];
-        bd = &mb->e_mbd.block[i];
+  for (i = 0; i < 16; i++) {
+    be = &mb->block[i];
+    bd = &mb->e_mbd.block[i];
 
-        berror = 0;
+    berror = 0;
 
-        for (j = dc; j < 16; j++)
-        {
-            int this_diff = be->coeff[j] - bd->dqcoeff[j];
-            berror += this_diff * this_diff;
-        }
-
-        error += berror;
+    for (j = dc; j < 16; j++) {
+      int this_diff = be->coeff[j] - bd->dqcoeff[j];
+      berror += this_diff * this_diff;
     }
 
-    return error;
+    error += berror;
+  }
+
+  return error;
 }
 
-int vp8_mbuverror_c(MACROBLOCK *mb)
-{
+int vp8_mbuverror_c(MACROBLOCK *mb) {
+  BLOCK *be;
+  BLOCKD *bd;
 
-    BLOCK  *be;
-    BLOCKD *bd;
+  int i;
+  int error = 0;
 
+  for (i = 16; i < 24; i++) {
+    be = &mb->block[i];
+    bd = &mb->e_mbd.block[i];
 
-    int i;
-    int error = 0;
+    error += vp8_block_error_c(be->coeff, bd->dqcoeff);
+  }
 
-    for (i = 16; i < 24; i++)
-    {
-        be = &mb->block[i];
-        bd = &mb->e_mbd.block[i];
-
-        error += vp8_block_error_c(be->coeff, bd->dqcoeff);
-    }
-
-    return error;
+  return error;
 }
 
-int VP8_UVSSE(MACROBLOCK *x)
-{
-    unsigned char *uptr, *vptr;
-    unsigned char *upred_ptr = (*(x->block[16].base_src) + x->block[16].src);
-    unsigned char *vpred_ptr = (*(x->block[20].base_src) + x->block[20].src);
-    int uv_stride = x->block[16].src_stride;
+int VP8_UVSSE(MACROBLOCK *x) {
+  unsigned char *uptr, *vptr;
+  unsigned char *upred_ptr = (*(x->block[16].base_src) + x->block[16].src);
+  unsigned char *vpred_ptr = (*(x->block[20].base_src) + x->block[20].src);
+  int uv_stride = x->block[16].src_stride;
 
-    unsigned int sse1 = 0;
-    unsigned int sse2 = 0;
-    int mv_row = x->e_mbd.mode_info_context->mbmi.mv.as_mv.row;
-    int mv_col = x->e_mbd.mode_info_context->mbmi.mv.as_mv.col;
-    int offset;
-    int pre_stride = x->e_mbd.pre.uv_stride;
+  unsigned int sse1 = 0;
+  unsigned int sse2 = 0;
+  int mv_row = x->e_mbd.mode_info_context->mbmi.mv.as_mv.row;
+  int mv_col = x->e_mbd.mode_info_context->mbmi.mv.as_mv.col;
+  int offset;
+  int pre_stride = x->e_mbd.pre.uv_stride;
 
-    if (mv_row < 0)
-        mv_row -= 1;
-    else
-        mv_row += 1;
+  if (mv_row < 0)
+    mv_row -= 1;
+  else
+    mv_row += 1;
 
-    if (mv_col < 0)
-        mv_col -= 1;
-    else
-        mv_col += 1;
+  if (mv_col < 0)
+    mv_col -= 1;
+  else
+    mv_col += 1;
 
-    mv_row /= 2;
-    mv_col /= 2;
+  mv_row /= 2;
+  mv_col /= 2;
 
-    offset = (mv_row >> 3) * pre_stride + (mv_col >> 3);
-    uptr = x->e_mbd.pre.u_buffer + offset;
-    vptr = x->e_mbd.pre.v_buffer + offset;
+  offset = (mv_row >> 3) * pre_stride + (mv_col >> 3);
+  uptr = x->e_mbd.pre.u_buffer + offset;
+  vptr = x->e_mbd.pre.v_buffer + offset;
 
-    if ((mv_row | mv_col) & 7)
-    {
-        vpx_sub_pixel_variance8x8(uptr, pre_stride,
-            mv_col & 7, mv_row & 7, upred_ptr, uv_stride, &sse2);
-        vpx_sub_pixel_variance8x8(vptr, pre_stride,
-            mv_col & 7, mv_row & 7, vpred_ptr, uv_stride, &sse1);
-        sse2 += sse1;
-    }
-    else
-    {
-        vpx_variance8x8(uptr, pre_stride,
-            upred_ptr, uv_stride, &sse2);
-        vpx_variance8x8(vptr, pre_stride,
-            vpred_ptr, uv_stride, &sse1);
-        sse2 += sse1;
-    }
-    return sse2;
-
+  if ((mv_row | mv_col) & 7) {
+    vpx_sub_pixel_variance8x8(uptr, pre_stride, mv_col & 7, mv_row & 7,
+                              upred_ptr, uv_stride, &sse2);
+    vpx_sub_pixel_variance8x8(vptr, pre_stride, mv_col & 7, mv_row & 7,
+                              vpred_ptr, uv_stride, &sse1);
+    sse2 += sse1;
+  } else {
+    vpx_variance8x8(uptr, pre_stride, upred_ptr, uv_stride, &sse2);
+    vpx_variance8x8(vptr, pre_stride, vpred_ptr, uv_stride, &sse1);
+    sse2 += sse1;
+  }
+  return sse2;
 }
 
-static int cost_coeffs(MACROBLOCK *mb, BLOCKD *b, int type, ENTROPY_CONTEXT *a, ENTROPY_CONTEXT *l)
-{
-    int c = !type;              /* start at coef 0, unless Y with Y2 */
-    int eob = (int)(*b->eob);
-    int pt ;    /* surrounding block/prev coef predictor */
-    int cost = 0;
-    short *qcoeff_ptr = b->qcoeff;
+static int cost_coeffs(MACROBLOCK *mb, BLOCKD *b, int type, ENTROPY_CONTEXT *a,
+                       ENTROPY_CONTEXT *l) {
+  int c = !type; /* start at coef 0, unless Y with Y2 */
+  int eob = (int)(*b->eob);
+  int pt; /* surrounding block/prev coef predictor */
+  int cost = 0;
+  short *qcoeff_ptr = b->qcoeff;
 
-    VP8_COMBINEENTROPYCONTEXTS(pt, *a, *l);
+  VP8_COMBINEENTROPYCONTEXTS(pt, *a, *l);
 
-    assert(eob <= 16);
-    for (; c < eob; c++)
-    {
-        const int v = qcoeff_ptr[vp8_default_zig_zag1d[c]];
-        const int t = vp8_dct_value_tokens_ptr[v].Token;
-        cost += mb->token_costs [type] [vp8_coef_bands[c]] [pt] [t];
-        cost += vp8_dct_value_cost_ptr[v];
-        pt = vp8_prev_token_class[t];
-    }
+  assert(eob <= 16);
+  for (; c < eob; c++) {
+    const int v = qcoeff_ptr[vp8_default_zig_zag1d[c]];
+    const int t = vp8_dct_value_tokens_ptr[v].Token;
+    cost += mb->token_costs[type][vp8_coef_bands[c]][pt][t];
+    cost += vp8_dct_value_cost_ptr[v];
+    pt = vp8_prev_token_class[t];
+  }
 
-    if (c < 16)
-        cost += mb->token_costs [type] [vp8_coef_bands[c]] [pt] [DCT_EOB_TOKEN];
+  if (c < 16)
+    cost += mb->token_costs[type][vp8_coef_bands[c]][pt][DCT_EOB_TOKEN];
 
-    pt = (c != !type); /* is eob first coefficient; */
-    *a = *l = pt;
+  pt = (c != !type); /* is eob first coefficient; */
+  *a = *l = pt;
 
-    return cost;
+  return cost;
 }
 
-static int vp8_rdcost_mby(MACROBLOCK *mb)
-{
-    int cost = 0;
-    int b;
-    MACROBLOCKD *x = &mb->e_mbd;
-    ENTROPY_CONTEXT_PLANES t_above, t_left;
-    ENTROPY_CONTEXT *ta;
-    ENTROPY_CONTEXT *tl;
+static int vp8_rdcost_mby(MACROBLOCK *mb) {
+  int cost = 0;
+  int b;
+  MACROBLOCKD *x = &mb->e_mbd;
+  ENTROPY_CONTEXT_PLANES t_above, t_left;
+  ENTROPY_CONTEXT *ta;
+  ENTROPY_CONTEXT *tl;
 
-    memcpy(&t_above, mb->e_mbd.above_context, sizeof(ENTROPY_CONTEXT_PLANES));
-    memcpy(&t_left, mb->e_mbd.left_context, sizeof(ENTROPY_CONTEXT_PLANES));
+  memcpy(&t_above, mb->e_mbd.above_context, sizeof(ENTROPY_CONTEXT_PLANES));
+  memcpy(&t_left, mb->e_mbd.left_context, sizeof(ENTROPY_CONTEXT_PLANES));
 
-    ta = (ENTROPY_CONTEXT *)&t_above;
-    tl = (ENTROPY_CONTEXT *)&t_left;
+  ta = (ENTROPY_CONTEXT *)&t_above;
+  tl = (ENTROPY_CONTEXT *)&t_left;
 
-    for (b = 0; b < 16; b++)
-        cost += cost_coeffs(mb, x->block + b, PLANE_TYPE_Y_NO_DC,
-                    ta + vp8_block2above[b], tl + vp8_block2left[b]);
+  for (b = 0; b < 16; b++)
+    cost += cost_coeffs(mb, x->block + b, PLANE_TYPE_Y_NO_DC,
+                        ta + vp8_block2above[b], tl + vp8_block2left[b]);
 
-    cost += cost_coeffs(mb, x->block + 24, PLANE_TYPE_Y2,
-                ta + vp8_block2above[24], tl + vp8_block2left[24]);
+  cost += cost_coeffs(mb, x->block + 24, PLANE_TYPE_Y2,
+                      ta + vp8_block2above[24], tl + vp8_block2left[24]);
 
-    return cost;
+  return cost;
 }
 
-static void macro_block_yrd( MACROBLOCK *mb,
-                             int *Rate,
-                             int *Distortion)
-{
-    int b;
-    MACROBLOCKD *const x = &mb->e_mbd;
-    BLOCK   *const mb_y2 = mb->block + 24;
-    BLOCKD *const x_y2  = x->block + 24;
-    short *Y2DCPtr = mb_y2->src_diff;
-    BLOCK *beptr;
-    int d;
+static void macro_block_yrd(MACROBLOCK *mb, int *Rate, int *Distortion) {
+  int b;
+  MACROBLOCKD *const x = &mb->e_mbd;
+  BLOCK *const mb_y2 = mb->block + 24;
+  BLOCKD *const x_y2 = x->block + 24;
+  short *Y2DCPtr = mb_y2->src_diff;
+  BLOCK *beptr;
+  int d;
 
-    vp8_subtract_mby( mb->src_diff, *(mb->block[0].base_src),
-        mb->block[0].src_stride,  mb->e_mbd.predictor, 16);
+  vp8_subtract_mby(mb->src_diff, *(mb->block[0].base_src),
+                   mb->block[0].src_stride, mb->e_mbd.predictor, 16);
 
-    /* Fdct and building the 2nd order block */
-    for (beptr = mb->block; beptr < mb->block + 16; beptr += 2)
-    {
-        mb->short_fdct8x4(beptr->src_diff, beptr->coeff, 32);
-        *Y2DCPtr++ = beptr->coeff[0];
-        *Y2DCPtr++ = beptr->coeff[16];
-    }
+  /* Fdct and building the 2nd order block */
+  for (beptr = mb->block; beptr < mb->block + 16; beptr += 2) {
+    mb->short_fdct8x4(beptr->src_diff, beptr->coeff, 32);
+    *Y2DCPtr++ = beptr->coeff[0];
+    *Y2DCPtr++ = beptr->coeff[16];
+  }
 
-    /* 2nd order fdct */
-    mb->short_walsh4x4(mb_y2->src_diff, mb_y2->coeff, 8);
+  /* 2nd order fdct */
+  mb->short_walsh4x4(mb_y2->src_diff, mb_y2->coeff, 8);
 
-    /* Quantization */
-    for (b = 0; b < 16; b++)
-    {
-        mb->quantize_b(&mb->block[b], &mb->e_mbd.block[b]);
-    }
+  /* Quantization */
+  for (b = 0; b < 16; b++) {
+    mb->quantize_b(&mb->block[b], &mb->e_mbd.block[b]);
+  }
 
-    /* DC predication and Quantization of 2nd Order block */
-    mb->quantize_b(mb_y2, x_y2);
+  /* DC predication and Quantization of 2nd Order block */
+  mb->quantize_b(mb_y2, x_y2);
 
-    /* Distortion */
-    d = vp8_mbblock_error(mb, 1) << 2;
-    d += vp8_block_error(mb_y2->coeff, x_y2->dqcoeff);
+  /* Distortion */
+  d = vp8_mbblock_error(mb, 1) << 2;
+  d += vp8_block_error(mb_y2->coeff, x_y2->dqcoeff);
 
-    *Distortion = (d >> 4);
+  *Distortion = (d >> 4);
 
-    /* rate */
-    *Rate = vp8_rdcost_mby(mb);
+  /* rate */
+  *Rate = vp8_rdcost_mby(mb);
 }
 
-static void copy_predictor(unsigned char *dst, const unsigned char *predictor)
-{
-    const unsigned int *p = (const unsigned int *)predictor;
-    unsigned int *d = (unsigned int *)dst;
-    d[0] = p[0];
-    d[4] = p[4];
-    d[8] = p[8];
-    d[12] = p[12];
+static void copy_predictor(unsigned char *dst, const unsigned char *predictor) {
+  const unsigned int *p = (const unsigned int *)predictor;
+  unsigned int *d = (unsigned int *)dst;
+  d[0] = p[0];
+  d[4] = p[4];
+  d[8] = p[8];
+  d[12] = p[12];
 }
-static int rd_pick_intra4x4block(
-    MACROBLOCK *x,
-    BLOCK *be,
-    BLOCKD *b,
-    B_PREDICTION_MODE *best_mode,
-    const int *bmode_costs,
-    ENTROPY_CONTEXT *a,
-    ENTROPY_CONTEXT *l,
+static int rd_pick_intra4x4block(MACROBLOCK *x, BLOCK *be, BLOCKD *b,
+                                 B_PREDICTION_MODE *best_mode,
+                                 const int *bmode_costs, ENTROPY_CONTEXT *a,
+                                 ENTROPY_CONTEXT *l,
 
-    int *bestrate,
-    int *bestratey,
-    int *bestdistortion)
-{
-    B_PREDICTION_MODE mode;
-    int best_rd = INT_MAX;
-    int rate = 0;
-    int distortion;
+                                 int *bestrate, int *bestratey,
+                                 int *bestdistortion) {
+  B_PREDICTION_MODE mode;
+  int best_rd = INT_MAX;
+  int rate = 0;
+  int distortion;
 
-    ENTROPY_CONTEXT ta = *a, tempa = *a;
-    ENTROPY_CONTEXT tl = *l, templ = *l;
-    /*
-     * The predictor buffer is a 2d buffer with a stride of 16.  Create
-     * a temp buffer that meets the stride requirements, but we are only
-     * interested in the left 4x4 block
-     * */
-    DECLARE_ALIGNED(16, unsigned char,  best_predictor[16*4]);
-    DECLARE_ALIGNED(16, short, best_dqcoeff[16]);
-    int dst_stride = x->e_mbd.dst.y_stride;
-    unsigned char *dst = x->e_mbd.dst.y_buffer + b->offset;
+  ENTROPY_CONTEXT ta = *a, tempa = *a;
+  ENTROPY_CONTEXT tl = *l, templ = *l;
+  /*
+   * The predictor buffer is a 2d buffer with a stride of 16.  Create
+   * a temp buffer that meets the stride requirements, but we are only
+   * interested in the left 4x4 block
+   * */
+  DECLARE_ALIGNED(16, unsigned char, best_predictor[16 * 4]);
+  DECLARE_ALIGNED(16, short, best_dqcoeff[16]);
+  int dst_stride = x->e_mbd.dst.y_stride;
+  unsigned char *dst = x->e_mbd.dst.y_buffer + b->offset;
 
-    unsigned char *Above = dst - dst_stride;
-    unsigned char *yleft = dst - 1;
-    unsigned char top_left = Above[-1];
+  unsigned char *Above = dst - dst_stride;
+  unsigned char *yleft = dst - 1;
+  unsigned char top_left = Above[-1];
 
-    for (mode = B_DC_PRED; mode <= B_HU_PRED; mode++)
-    {
-        int this_rd;
-        int ratey;
+  for (mode = B_DC_PRED; mode <= B_HU_PRED; mode++) {
+    int this_rd;
+    int ratey;
 
-        rate = bmode_costs[mode];
+    rate = bmode_costs[mode];
 
-        vp8_intra4x4_predict(Above, yleft, dst_stride, mode,
-                             b->predictor, 16, top_left);
-        vp8_subtract_b(be, b, 16);
-        x->short_fdct4x4(be->src_diff, be->coeff, 32);
-        x->quantize_b(be, b);
+    vp8_intra4x4_predict(Above, yleft, dst_stride, mode, b->predictor, 16,
+                         top_left);
+    vp8_subtract_b(be, b, 16);
+    x->short_fdct4x4(be->src_diff, be->coeff, 32);
+    x->quantize_b(be, b);
 
-        tempa = ta;
-        templ = tl;
+    tempa = ta;
+    templ = tl;
 
-        ratey = cost_coeffs(x, b, PLANE_TYPE_Y_WITH_DC, &tempa, &templ);
-        rate += ratey;
-        distortion = vp8_block_error(be->coeff, b->dqcoeff) >> 2;
+    ratey = cost_coeffs(x, b, PLANE_TYPE_Y_WITH_DC, &tempa, &templ);
+    rate += ratey;
+    distortion = vp8_block_error(be->coeff, b->dqcoeff) >> 2;
 
-        this_rd = RDCOST(x->rdmult, x->rddiv, rate, distortion);
+    this_rd = RDCOST(x->rdmult, x->rddiv, rate, distortion);
 
-        if (this_rd < best_rd)
-        {
-            *bestrate = rate;
-            *bestratey = ratey;
-            *bestdistortion = distortion;
-            best_rd = this_rd;
-            *best_mode = mode;
-            *a = tempa;
-            *l = templ;
-            copy_predictor(best_predictor, b->predictor);
-            memcpy(best_dqcoeff, b->dqcoeff, 32);
-        }
+    if (this_rd < best_rd) {
+      *bestrate = rate;
+      *bestratey = ratey;
+      *bestdistortion = distortion;
+      best_rd = this_rd;
+      *best_mode = mode;
+      *a = tempa;
+      *l = templ;
+      copy_predictor(best_predictor, b->predictor);
+      memcpy(best_dqcoeff, b->dqcoeff, 32);
     }
-    b->bmi.as_mode = *best_mode;
+  }
+  b->bmi.as_mode = *best_mode;
 
-    vp8_short_idct4x4llm(best_dqcoeff, best_predictor, 16, dst, dst_stride);
+  vp8_short_idct4x4llm(best_dqcoeff, best_predictor, 16, dst, dst_stride);
 
-    return best_rd;
+  return best_rd;
 }
 
-static int rd_pick_intra4x4mby_modes(MACROBLOCK *mb, int *Rate,
-                                     int *rate_y, int *Distortion, int best_rd)
-{
-    MACROBLOCKD *const xd = &mb->e_mbd;
-    int i;
-    int cost = mb->mbmode_cost [xd->frame_type] [B_PRED];
-    int distortion = 0;
-    int tot_rate_y = 0;
-    int64_t total_rd = 0;
-    ENTROPY_CONTEXT_PLANES t_above, t_left;
-    ENTROPY_CONTEXT *ta;
-    ENTROPY_CONTEXT *tl;
-    const int *bmode_costs;
+static int rd_pick_intra4x4mby_modes(MACROBLOCK *mb, int *Rate, int *rate_y,
+                                     int *Distortion, int best_rd) {
+  MACROBLOCKD *const xd = &mb->e_mbd;
+  int i;
+  int cost = mb->mbmode_cost[xd->frame_type][B_PRED];
+  int distortion = 0;
+  int tot_rate_y = 0;
+  int64_t total_rd = 0;
+  ENTROPY_CONTEXT_PLANES t_above, t_left;
+  ENTROPY_CONTEXT *ta;
+  ENTROPY_CONTEXT *tl;
+  const int *bmode_costs;
 
-    memcpy(&t_above, mb->e_mbd.above_context, sizeof(ENTROPY_CONTEXT_PLANES));
-    memcpy(&t_left, mb->e_mbd.left_context, sizeof(ENTROPY_CONTEXT_PLANES));
+  memcpy(&t_above, mb->e_mbd.above_context, sizeof(ENTROPY_CONTEXT_PLANES));
+  memcpy(&t_left, mb->e_mbd.left_context, sizeof(ENTROPY_CONTEXT_PLANES));
 
-    ta = (ENTROPY_CONTEXT *)&t_above;
-    tl = (ENTROPY_CONTEXT *)&t_left;
+  ta = (ENTROPY_CONTEXT *)&t_above;
+  tl = (ENTROPY_CONTEXT *)&t_left;
 
-    intra_prediction_down_copy(xd, xd->dst.y_buffer - xd->dst.y_stride + 16);
+  intra_prediction_down_copy(xd, xd->dst.y_buffer - xd->dst.y_stride + 16);
 
-    bmode_costs = mb->inter_bmode_costs;
+  bmode_costs = mb->inter_bmode_costs;
 
-    for (i = 0; i < 16; i++)
-    {
-        MODE_INFO *const mic = xd->mode_info_context;
-        const int mis = xd->mode_info_stride;
-        B_PREDICTION_MODE UNINITIALIZED_IS_SAFE(best_mode);
-        int UNINITIALIZED_IS_SAFE(r), UNINITIALIZED_IS_SAFE(ry), UNINITIALIZED_IS_SAFE(d);
+  for (i = 0; i < 16; i++) {
+    MODE_INFO *const mic = xd->mode_info_context;
+    const int mis = xd->mode_info_stride;
+    B_PREDICTION_MODE UNINITIALIZED_IS_SAFE(best_mode);
+    int UNINITIALIZED_IS_SAFE(r), UNINITIALIZED_IS_SAFE(ry),
+        UNINITIALIZED_IS_SAFE(d);
 
-        if (mb->e_mbd.frame_type == KEY_FRAME)
-        {
-            const B_PREDICTION_MODE A = above_block_mode(mic, i, mis);
-            const B_PREDICTION_MODE L = left_block_mode(mic, i);
+    if (mb->e_mbd.frame_type == KEY_FRAME) {
+      const B_PREDICTION_MODE A = above_block_mode(mic, i, mis);
+      const B_PREDICTION_MODE L = left_block_mode(mic, i);
 
-            bmode_costs  = mb->bmode_costs[A][L];
-        }
+      bmode_costs = mb->bmode_costs[A][L];
+    }
 
-        total_rd += rd_pick_intra4x4block(
-            mb, mb->block + i, xd->block + i, &best_mode, bmode_costs,
-            ta + vp8_block2above[i],
-            tl + vp8_block2left[i], &r, &ry, &d);
+    total_rd += rd_pick_intra4x4block(
+        mb, mb->block + i, xd->block + i, &best_mode, bmode_costs,
+        ta + vp8_block2above[i], tl + vp8_block2left[i], &r, &ry, &d);
 
-        cost += r;
-        distortion += d;
-        tot_rate_y += ry;
+    cost += r;
+    distortion += d;
+    tot_rate_y += ry;
 
-        mic->bmi[i].as_mode = best_mode;
+    mic->bmi[i].as_mode = best_mode;
 
-        if(total_rd >= (int64_t)best_rd)
-            break;
-    }
+    if (total_rd >= (int64_t)best_rd) break;
+  }
 
-    if(total_rd >= (int64_t)best_rd)
-        return INT_MAX;
+  if (total_rd >= (int64_t)best_rd) return INT_MAX;
 
-    *Rate = cost;
-    *rate_y = tot_rate_y;
-    *Distortion = distortion;
+  *Rate = cost;
+  *rate_y = tot_rate_y;
+  *Distortion = distortion;
 
-    return RDCOST(mb->rdmult, mb->rddiv, cost, distortion);
+  return RDCOST(mb->rdmult, mb->rddiv, cost, distortion);
 }
 
+static int rd_pick_intra16x16mby_mode(MACROBLOCK *x, int *Rate, int *rate_y,
+                                      int *Distortion) {
+  MB_PREDICTION_MODE mode;
+  MB_PREDICTION_MODE UNINITIALIZED_IS_SAFE(mode_selected);
+  int rate, ratey;
+  int distortion;
+  int best_rd = INT_MAX;
+  int this_rd;
+  MACROBLOCKD *xd = &x->e_mbd;
 
-static int rd_pick_intra16x16mby_mode(MACROBLOCK *x,
-                                      int *Rate,
-                                      int *rate_y,
-                                      int *Distortion)
-{
-    MB_PREDICTION_MODE mode;
-    MB_PREDICTION_MODE UNINITIALIZED_IS_SAFE(mode_selected);
-    int rate, ratey;
-    int distortion;
-    int best_rd = INT_MAX;
-    int this_rd;
-    MACROBLOCKD *xd = &x->e_mbd;
+  /* Y Search for 16x16 intra prediction mode */
+  for (mode = DC_PRED; mode <= TM_PRED; mode++) {
+    xd->mode_info_context->mbmi.mode = mode;
 
-    /* Y Search for 16x16 intra prediction mode */
-    for (mode = DC_PRED; mode <= TM_PRED; mode++)
-    {
-        xd->mode_info_context->mbmi.mode = mode;
+    vp8_build_intra_predictors_mby_s(xd, xd->dst.y_buffer - xd->dst.y_stride,
+                                     xd->dst.y_buffer - 1, xd->dst.y_stride,
+                                     xd->predictor, 16);
 
-        vp8_build_intra_predictors_mby_s(xd,
-                                         xd->dst.y_buffer - xd->dst.y_stride,
-                                         xd->dst.y_buffer - 1,
-                                         xd->dst.y_stride,
-                                         xd->predictor,
-                                         16);
+    macro_block_yrd(x, &ratey, &distortion);
+    rate = ratey +
+           x->mbmode_cost[xd->frame_type][xd->mode_info_context->mbmi.mode];
 
-        macro_block_yrd(x, &ratey, &distortion);
-        rate = ratey + x->mbmode_cost[xd->frame_type]
-                                     [xd->mode_info_context->mbmi.mode];
+    this_rd = RDCOST(x->rdmult, x->rddiv, rate, distortion);
 
-        this_rd = RDCOST(x->rdmult, x->rddiv, rate, distortion);
-
-        if (this_rd < best_rd)
-        {
-            mode_selected = mode;
-            best_rd = this_rd;
-            *Rate = rate;
-            *rate_y = ratey;
-            *Distortion = distortion;
-        }
+    if (this_rd < best_rd) {
+      mode_selected = mode;
+      best_rd = this_rd;
+      *Rate = rate;
+      *rate_y = ratey;
+      *Distortion = distortion;
     }
+  }
 
-    xd->mode_info_context->mbmi.mode = mode_selected;
-    return best_rd;
+  xd->mode_info_context->mbmi.mode = mode_selected;
+  return best_rd;
 }
 
-static int rd_cost_mbuv(MACROBLOCK *mb)
-{
-    int b;
-    int cost = 0;
-    MACROBLOCKD *x = &mb->e_mbd;
-    ENTROPY_CONTEXT_PLANES t_above, t_left;
-    ENTROPY_CONTEXT *ta;
-    ENTROPY_CONTEXT *tl;
+static int rd_cost_mbuv(MACROBLOCK *mb) {
+  int b;
+  int cost = 0;
+  MACROBLOCKD *x = &mb->e_mbd;
+  ENTROPY_CONTEXT_PLANES t_above, t_left;
+  ENTROPY_CONTEXT *ta;
+  ENTROPY_CONTEXT *tl;
 
-    memcpy(&t_above, mb->e_mbd.above_context, sizeof(ENTROPY_CONTEXT_PLANES));
-    memcpy(&t_left, mb->e_mbd.left_context, sizeof(ENTROPY_CONTEXT_PLANES));
+  memcpy(&t_above, mb->e_mbd.above_context, sizeof(ENTROPY_CONTEXT_PLANES));
+  memcpy(&t_left, mb->e_mbd.left_context, sizeof(ENTROPY_CONTEXT_PLANES));
 
-    ta = (ENTROPY_CONTEXT *)&t_above;
-    tl = (ENTROPY_CONTEXT *)&t_left;
+  ta = (ENTROPY_CONTEXT *)&t_above;
+  tl = (ENTROPY_CONTEXT *)&t_left;
 
-    for (b = 16; b < 24; b++)
-        cost += cost_coeffs(mb, x->block + b, PLANE_TYPE_UV,
-                    ta + vp8_block2above[b], tl + vp8_block2left[b]);
+  for (b = 16; b < 24; b++)
+    cost += cost_coeffs(mb, x->block + b, PLANE_TYPE_UV,
+                        ta + vp8_block2above[b], tl + vp8_block2left[b]);
 
-    return cost;
+  return cost;
 }
 
-
 static int rd_inter16x16_uv(VP8_COMP *cpi, MACROBLOCK *x, int *rate,
-                            int *distortion, int fullpixel)
-{
-    (void)cpi;
-    (void)fullpixel;
+                            int *distortion, int fullpixel) {
+  (void)cpi;
+  (void)fullpixel;
 
-    vp8_build_inter16x16_predictors_mbuv(&x->e_mbd);
-    vp8_subtract_mbuv(x->src_diff,
-        x->src.u_buffer, x->src.v_buffer, x->src.uv_stride,
-        &x->e_mbd.predictor[256], &x->e_mbd.predictor[320], 8);
+  vp8_build_inter16x16_predictors_mbuv(&x->e_mbd);
+  vp8_subtract_mbuv(x->src_diff, x->src.u_buffer, x->src.v_buffer,
+                    x->src.uv_stride, &x->e_mbd.predictor[256],
+                    &x->e_mbd.predictor[320], 8);
 
-    vp8_transform_mbuv(x);
-    vp8_quantize_mbuv(x);
+  vp8_transform_mbuv(x);
+  vp8_quantize_mbuv(x);
 
-    *rate       = rd_cost_mbuv(x);
-    *distortion = vp8_mbuverror(x) / 4;
+  *rate = rd_cost_mbuv(x);
+  *distortion = vp8_mbuverror(x) / 4;
 
-    return RDCOST(x->rdmult, x->rddiv, *rate, *distortion);
+  return RDCOST(x->rdmult, x->rddiv, *rate, *distortion);
 }
 
 static int rd_inter4x4_uv(VP8_COMP *cpi, MACROBLOCK *x, int *rate,
-                          int *distortion, int fullpixel)
-{
-    (void)cpi;
-    (void)fullpixel;
+                          int *distortion, int fullpixel) {
+  (void)cpi;
+  (void)fullpixel;
 
-    vp8_build_inter4x4_predictors_mbuv(&x->e_mbd);
-    vp8_subtract_mbuv(x->src_diff,
-        x->src.u_buffer, x->src.v_buffer, x->src.uv_stride,
-        &x->e_mbd.predictor[256], &x->e_mbd.predictor[320], 8);
+  vp8_build_inter4x4_predictors_mbuv(&x->e_mbd);
+  vp8_subtract_mbuv(x->src_diff, x->src.u_buffer, x->src.v_buffer,
+                    x->src.uv_stride, &x->e_mbd.predictor[256],
+                    &x->e_mbd.predictor[320], 8);
 
-    vp8_transform_mbuv(x);
-    vp8_quantize_mbuv(x);
+  vp8_transform_mbuv(x);
+  vp8_quantize_mbuv(x);
 
-    *rate       = rd_cost_mbuv(x);
-    *distortion = vp8_mbuverror(x) / 4;
+  *rate = rd_cost_mbuv(x);
+  *distortion = vp8_mbuverror(x) / 4;
 
-    return RDCOST(x->rdmult, x->rddiv, *rate, *distortion);
+  return RDCOST(x->rdmult, x->rddiv, *rate, *distortion);
 }
 
 static void rd_pick_intra_mbuv_mode(MACROBLOCK *x, int *rate,
-                                    int *rate_tokenonly, int *distortion)
-{
-    MB_PREDICTION_MODE mode;
-    MB_PREDICTION_MODE UNINITIALIZED_IS_SAFE(mode_selected);
-    int best_rd = INT_MAX;
-    int UNINITIALIZED_IS_SAFE(d), UNINITIALIZED_IS_SAFE(r);
-    int rate_to;
-    MACROBLOCKD *xd = &x->e_mbd;
+                                    int *rate_tokenonly, int *distortion) {
+  MB_PREDICTION_MODE mode;
+  MB_PREDICTION_MODE UNINITIALIZED_IS_SAFE(mode_selected);
+  int best_rd = INT_MAX;
+  int UNINITIALIZED_IS_SAFE(d), UNINITIALIZED_IS_SAFE(r);
+  int rate_to;
+  MACROBLOCKD *xd = &x->e_mbd;
 
-    for (mode = DC_PRED; mode <= TM_PRED; mode++)
-    {
-        int this_rate;
-        int this_distortion;
-        int this_rd;
+  for (mode = DC_PRED; mode <= TM_PRED; mode++) {
+    int this_rate;
+    int this_distortion;
+    int this_rd;
 
-        xd->mode_info_context->mbmi.uv_mode = mode;
+    xd->mode_info_context->mbmi.uv_mode = mode;
 
-        vp8_build_intra_predictors_mbuv_s(xd,
-                                          xd->dst.u_buffer - xd->dst.uv_stride,
-                                          xd->dst.v_buffer - xd->dst.uv_stride,
-                                          xd->dst.u_buffer - 1,
-                                          xd->dst.v_buffer - 1,
-                                          xd->dst.uv_stride,
-                                          &xd->predictor[256], &xd->predictor[320],
-                                          8);
+    vp8_build_intra_predictors_mbuv_s(
+        xd, xd->dst.u_buffer - xd->dst.uv_stride,
+        xd->dst.v_buffer - xd->dst.uv_stride, xd->dst.u_buffer - 1,
+        xd->dst.v_buffer - 1, xd->dst.uv_stride, &xd->predictor[256],
+        &xd->predictor[320], 8);
 
+    vp8_subtract_mbuv(x->src_diff, x->src.u_buffer, x->src.v_buffer,
+                      x->src.uv_stride, &xd->predictor[256],
+                      &xd->predictor[320], 8);
+    vp8_transform_mbuv(x);
+    vp8_quantize_mbuv(x);
 
-        vp8_subtract_mbuv(x->src_diff,
-                      x->src.u_buffer, x->src.v_buffer, x->src.uv_stride,
-                      &xd->predictor[256], &xd->predictor[320], 8);
-        vp8_transform_mbuv(x);
-        vp8_quantize_mbuv(x);
+    rate_to = rd_cost_mbuv(x);
+    this_rate = rate_to +
+                x->intra_uv_mode_cost[xd->frame_type][xd->mode_info_context
+                                                          ->mbmi.uv_mode];
 
-        rate_to = rd_cost_mbuv(x);
-        this_rate = rate_to + x->intra_uv_mode_cost[xd->frame_type][xd->mode_info_context->mbmi.uv_mode];
+    this_distortion = vp8_mbuverror(x) / 4;
 
-        this_distortion = vp8_mbuverror(x) / 4;
+    this_rd = RDCOST(x->rdmult, x->rddiv, this_rate, this_distortion);
 
-        this_rd = RDCOST(x->rdmult, x->rddiv, this_rate, this_distortion);
-
-        if (this_rd < best_rd)
-        {
-            best_rd = this_rd;
-            d = this_distortion;
-            r = this_rate;
-            *rate_tokenonly = rate_to;
-            mode_selected = mode;
-        }
+    if (this_rd < best_rd) {
+      best_rd = this_rd;
+      d = this_distortion;
+      r = this_rate;
+      *rate_tokenonly = rate_to;
+      mode_selected = mode;
     }
+  }
 
-    *rate = r;
-    *distortion = d;
+  *rate = r;
+  *distortion = d;
 
-    xd->mode_info_context->mbmi.uv_mode = mode_selected;
+  xd->mode_info_context->mbmi.uv_mode = mode_selected;
 }
 
-int vp8_cost_mv_ref(MB_PREDICTION_MODE m, const int near_mv_ref_ct[4])
-{
-    vp8_prob p [VP8_MVREFS-1];
-    assert(NEARESTMV <= m  &&  m <= SPLITMV);
-    vp8_mv_ref_probs(p, near_mv_ref_ct);
-    return vp8_cost_token(vp8_mv_ref_tree, p,
-                          vp8_mv_ref_encoding_array + (m - NEARESTMV));
+int vp8_cost_mv_ref(MB_PREDICTION_MODE m, const int near_mv_ref_ct[4]) {
+  vp8_prob p[VP8_MVREFS - 1];
+  assert(NEARESTMV <= m && m <= SPLITMV);
+  vp8_mv_ref_probs(p, near_mv_ref_ct);
+  return vp8_cost_token(vp8_mv_ref_tree, p,
+                        vp8_mv_ref_encoding_array + (m - NEARESTMV));
 }
 
-void vp8_set_mbmode_and_mvs(MACROBLOCK *x, MB_PREDICTION_MODE mb, int_mv *mv)
-{
-    x->e_mbd.mode_info_context->mbmi.mode = mb;
-    x->e_mbd.mode_info_context->mbmi.mv.as_int = mv->as_int;
+void vp8_set_mbmode_and_mvs(MACROBLOCK *x, MB_PREDICTION_MODE mb, int_mv *mv) {
+  x->e_mbd.mode_info_context->mbmi.mode = mb;
+  x->e_mbd.mode_info_context->mbmi.mv.as_int = mv->as_int;
 }
 
-static int labels2mode(
-    MACROBLOCK *x,
-    int const *labelings, int which_label,
-    B_PREDICTION_MODE this_mode,
-    int_mv *this_mv, int_mv *best_ref_mv,
-    int *mvcost[2]
-)
-{
-    MACROBLOCKD *const xd = & x->e_mbd;
-    MODE_INFO *const mic = xd->mode_info_context;
-    const int mis = xd->mode_info_stride;
+static int labels2mode(MACROBLOCK *x, int const *labelings, int which_label,
+                       B_PREDICTION_MODE this_mode, int_mv *this_mv,
+                       int_mv *best_ref_mv, int *mvcost[2]) {
+  MACROBLOCKD *const xd = &x->e_mbd;
+  MODE_INFO *const mic = xd->mode_info_context;
+  const int mis = xd->mode_info_stride;
 
-    int cost = 0;
-    int thismvcost = 0;
+  int cost = 0;
+  int thismvcost = 0;
 
-    /* We have to be careful retrieving previously-encoded motion vectors.
-       Ones from this macroblock have to be pulled from the BLOCKD array
-       as they have not yet made it to the bmi array in our MB_MODE_INFO. */
+  /* We have to be careful retrieving previously-encoded motion vectors.
+     Ones from this macroblock have to be pulled from the BLOCKD array
+     as they have not yet made it to the bmi array in our MB_MODE_INFO. */
 
-    int i = 0;
+  int i = 0;
 
-    do
-    {
-        BLOCKD *const d = xd->block + i;
-        const int row = i >> 2,  col = i & 3;
+  do {
+    BLOCKD *const d = xd->block + i;
+    const int row = i >> 2, col = i & 3;
 
-        B_PREDICTION_MODE m;
+    B_PREDICTION_MODE m;
 
-        if (labelings[i] != which_label)
-            continue;
+    if (labelings[i] != which_label) continue;
 
-        if (col  &&  labelings[i] == labelings[i-1])
-            m = LEFT4X4;
-        else if (row  &&  labelings[i] == labelings[i-4])
-            m = ABOVE4X4;
-        else
-        {
-            /* the only time we should do costing for new motion vector
-             * or mode is when we are on a new label  (jbb May 08, 2007)
-             */
-            switch (m = this_mode)
-            {
-            case NEW4X4 :
-                thismvcost  = vp8_mv_bit_cost(this_mv, best_ref_mv, mvcost, 102);
-                break;
-            case LEFT4X4:
-                this_mv->as_int = col ? d[-1].bmi.mv.as_int : left_block_mv(mic, i);
-                break;
-            case ABOVE4X4:
-                this_mv->as_int = row ? d[-4].bmi.mv.as_int : above_block_mv(mic, i, mis);
-                break;
-            case ZERO4X4:
-                this_mv->as_int = 0;
-                break;
-            default:
-                break;
-            }
+    if (col && labelings[i] == labelings[i - 1])
+      m = LEFT4X4;
+    else if (row && labelings[i] == labelings[i - 4])
+      m = ABOVE4X4;
+    else {
+      /* the only time we should do costing for new motion vector
+       * or mode is when we are on a new label  (jbb May 08, 2007)
+       */
+      switch (m = this_mode) {
+        case NEW4X4:
+          thismvcost = vp8_mv_bit_cost(this_mv, best_ref_mv, mvcost, 102);
+          break;
+        case LEFT4X4:
+          this_mv->as_int = col ? d[-1].bmi.mv.as_int : left_block_mv(mic, i);
+          break;
+        case ABOVE4X4:
+          this_mv->as_int =
+              row ? d[-4].bmi.mv.as_int : above_block_mv(mic, i, mis);
+          break;
+        case ZERO4X4: this_mv->as_int = 0; break;
+        default: break;
+      }
 
-            if (m == ABOVE4X4)  /* replace above with left if same */
-            {
-                int_mv left_mv;
+      if (m == ABOVE4X4) /* replace above with left if same */
+      {
+        int_mv left_mv;
 
-                left_mv.as_int = col ? d[-1].bmi.mv.as_int :
-                                        left_block_mv(mic, i);
+        left_mv.as_int = col ? d[-1].bmi.mv.as_int : left_block_mv(mic, i);
 
-                if (left_mv.as_int == this_mv->as_int)
-                    m = LEFT4X4;
-            }
+        if (left_mv.as_int == this_mv->as_int) m = LEFT4X4;
+      }
 
-            cost = x->inter_bmode_costs[ m];
-        }
+      cost = x->inter_bmode_costs[m];
+    }
 
-        d->bmi.mv.as_int = this_mv->as_int;
+    d->bmi.mv.as_int = this_mv->as_int;
 
-        x->partition_info->bmi[i].mode = m;
-        x->partition_info->bmi[i].mv.as_int = this_mv->as_int;
+    x->partition_info->bmi[i].mode = m;
+    x->partition_info->bmi[i].mv.as_int = this_mv->as_int;
 
-    }
-    while (++i < 16);
+  } while (++i < 16);
 
-    cost += thismvcost ;
-    return cost;
+  cost += thismvcost;
+  return cost;
 }
 
 static int rdcost_mbsegment_y(MACROBLOCK *mb, const int *labels,
                               int which_label, ENTROPY_CONTEXT *ta,
-                              ENTROPY_CONTEXT *tl)
-{
-    int cost = 0;
-    int b;
-    MACROBLOCKD *x = &mb->e_mbd;
+                              ENTROPY_CONTEXT *tl) {
+  int cost = 0;
+  int b;
+  MACROBLOCKD *x = &mb->e_mbd;
 
-    for (b = 0; b < 16; b++)
-        if (labels[ b] == which_label)
-            cost += cost_coeffs(mb, x->block + b, PLANE_TYPE_Y_WITH_DC,
-                                ta + vp8_block2above[b],
-                                tl + vp8_block2left[b]);
+  for (b = 0; b < 16; b++)
+    if (labels[b] == which_label)
+      cost += cost_coeffs(mb, x->block + b, PLANE_TYPE_Y_WITH_DC,
+                          ta + vp8_block2above[b], tl + vp8_block2left[b]);
 
-    return cost;
-
+  return cost;
 }
-static unsigned int vp8_encode_inter_mb_segment(MACROBLOCK *x, int const *labels, int which_label)
-{
-    int i;
-    unsigned int distortion = 0;
-    int pre_stride = x->e_mbd.pre.y_stride;
-    unsigned char *base_pre = x->e_mbd.pre.y_buffer;
+static unsigned int vp8_encode_inter_mb_segment(MACROBLOCK *x,
+                                                int const *labels,
+                                                int which_label) {
+  int i;
+  unsigned int distortion = 0;
+  int pre_stride = x->e_mbd.pre.y_stride;
+  unsigned char *base_pre = x->e_mbd.pre.y_buffer;
 
+  for (i = 0; i < 16; i++) {
+    if (labels[i] == which_label) {
+      BLOCKD *bd = &x->e_mbd.block[i];
+      BLOCK *be = &x->block[i];
 
-    for (i = 0; i < 16; i++)
-    {
-        if (labels[i] == which_label)
-        {
-            BLOCKD *bd = &x->e_mbd.block[i];
-            BLOCK *be = &x->block[i];
+      vp8_build_inter_predictors_b(bd, 16, base_pre, pre_stride,
+                                   x->e_mbd.subpixel_predict);
+      vp8_subtract_b(be, bd, 16);
+      x->short_fdct4x4(be->src_diff, be->coeff, 32);
+      x->quantize_b(be, bd);
 
-            vp8_build_inter_predictors_b(bd, 16, base_pre, pre_stride, x->e_mbd.subpixel_predict);
-            vp8_subtract_b(be, bd, 16);
-            x->short_fdct4x4(be->src_diff, be->coeff, 32);
-            x->quantize_b(be, bd);
-
-            distortion += vp8_block_error(be->coeff, bd->dqcoeff);
-        }
+      distortion += vp8_block_error(be->coeff, bd->dqcoeff);
     }
+  }
 
-    return distortion;
+  return distortion;
 }
 
+static const unsigned int segmentation_to_sseshift[4] = { 3, 3, 2, 0 };
 
-static const unsigned int segmentation_to_sseshift[4] = {3, 3, 2, 0};
-
-
-typedef struct
-{
+typedef struct {
   int_mv *ref_mv;
   int_mv mvp;
 
@@ -1102,1545 +924,1427 @@
 
 } BEST_SEG_INFO;
 
+static void rd_check_segment(VP8_COMP *cpi, MACROBLOCK *x, BEST_SEG_INFO *bsi,
+                             unsigned int segmentation) {
+  int i;
+  int const *labels;
+  int br = 0;
+  int bd = 0;
+  B_PREDICTION_MODE this_mode;
 
-static void rd_check_segment(VP8_COMP *cpi, MACROBLOCK *x,
-                             BEST_SEG_INFO *bsi, unsigned int segmentation)
-{
-    int i;
-    int const *labels;
-    int br = 0;
-    int bd = 0;
-    B_PREDICTION_MODE this_mode;
+  int label_count;
+  int this_segment_rd = 0;
+  int label_mv_thresh;
+  int rate = 0;
+  int sbr = 0;
+  int sbd = 0;
+  int segmentyrate = 0;
 
+  vp8_variance_fn_ptr_t *v_fn_ptr;
 
-    int label_count;
-    int this_segment_rd = 0;
-    int label_mv_thresh;
-    int rate = 0;
-    int sbr = 0;
-    int sbd = 0;
-    int segmentyrate = 0;
+  ENTROPY_CONTEXT_PLANES t_above, t_left;
+  ENTROPY_CONTEXT *ta;
+  ENTROPY_CONTEXT *tl;
+  ENTROPY_CONTEXT_PLANES t_above_b, t_left_b;
+  ENTROPY_CONTEXT *ta_b;
+  ENTROPY_CONTEXT *tl_b;
 
-    vp8_variance_fn_ptr_t *v_fn_ptr;
+  memcpy(&t_above, x->e_mbd.above_context, sizeof(ENTROPY_CONTEXT_PLANES));
+  memcpy(&t_left, x->e_mbd.left_context, sizeof(ENTROPY_CONTEXT_PLANES));
 
-    ENTROPY_CONTEXT_PLANES t_above, t_left;
-    ENTROPY_CONTEXT *ta;
-    ENTROPY_CONTEXT *tl;
-    ENTROPY_CONTEXT_PLANES t_above_b, t_left_b;
-    ENTROPY_CONTEXT *ta_b;
-    ENTROPY_CONTEXT *tl_b;
+  ta = (ENTROPY_CONTEXT *)&t_above;
+  tl = (ENTROPY_CONTEXT *)&t_left;
+  ta_b = (ENTROPY_CONTEXT *)&t_above_b;
+  tl_b = (ENTROPY_CONTEXT *)&t_left_b;
 
-    memcpy(&t_above, x->e_mbd.above_context, sizeof(ENTROPY_CONTEXT_PLANES));
-    memcpy(&t_left, x->e_mbd.left_context, sizeof(ENTROPY_CONTEXT_PLANES));
+  br = 0;
+  bd = 0;
 
-    ta = (ENTROPY_CONTEXT *)&t_above;
-    tl = (ENTROPY_CONTEXT *)&t_left;
-    ta_b = (ENTROPY_CONTEXT *)&t_above_b;
-    tl_b = (ENTROPY_CONTEXT *)&t_left_b;
+  v_fn_ptr = &cpi->fn_ptr[segmentation];
+  labels = vp8_mbsplits[segmentation];
+  label_count = vp8_mbsplit_count[segmentation];
 
-    br = 0;
-    bd = 0;
+  /* 64 makes this threshold really big effectively making it so that we
+   * very rarely check mvs on segments.   setting this to 1 would make mv
+   * thresh roughly equal to what it is for macroblocks
+   */
+  label_mv_thresh = 1 * bsi->mvthresh / label_count;
 
-    v_fn_ptr = &cpi->fn_ptr[segmentation];
-    labels = vp8_mbsplits[segmentation];
-    label_count = vp8_mbsplit_count[segmentation];
+  /* Segmentation method overheads */
+  rate = vp8_cost_token(vp8_mbsplit_tree, vp8_mbsplit_probs,
+                        vp8_mbsplit_encodings + segmentation);
+  rate += vp8_cost_mv_ref(SPLITMV, bsi->mdcounts);
+  this_segment_rd += RDCOST(x->rdmult, x->rddiv, rate, 0);
+  br += rate;
 
-    /* 64 makes this threshold really big effectively making it so that we
-     * very rarely check mvs on segments.   setting this to 1 would make mv
-     * thresh roughly equal to what it is for macroblocks
-     */
-    label_mv_thresh = 1 * bsi->mvthresh / label_count ;
+  for (i = 0; i < label_count; i++) {
+    int_mv mode_mv[B_MODE_COUNT];
+    int best_label_rd = INT_MAX;
+    B_PREDICTION_MODE mode_selected = ZERO4X4;
+    int bestlabelyrate = 0;
 
-    /* Segmentation method overheads */
-    rate = vp8_cost_token(vp8_mbsplit_tree, vp8_mbsplit_probs, vp8_mbsplit_encodings + segmentation);
-    rate += vp8_cost_mv_ref(SPLITMV, bsi->mdcounts);
-    this_segment_rd += RDCOST(x->rdmult, x->rddiv, rate, 0);
-    br += rate;
+    /* search for the best motion vector on this segment */
+    for (this_mode = LEFT4X4; this_mode <= NEW4X4; this_mode++) {
+      int this_rd;
+      int distortion;
+      int labelyrate;
+      ENTROPY_CONTEXT_PLANES t_above_s, t_left_s;
+      ENTROPY_CONTEXT *ta_s;
+      ENTROPY_CONTEXT *tl_s;
 
-    for (i = 0; i < label_count; i++)
-    {
-        int_mv mode_mv[B_MODE_COUNT];
-        int best_label_rd = INT_MAX;
-        B_PREDICTION_MODE mode_selected = ZERO4X4;
-        int bestlabelyrate = 0;
+      memcpy(&t_above_s, &t_above, sizeof(ENTROPY_CONTEXT_PLANES));
+      memcpy(&t_left_s, &t_left, sizeof(ENTROPY_CONTEXT_PLANES));
 
-        /* search for the best motion vector on this segment */
-        for (this_mode = LEFT4X4; this_mode <= NEW4X4 ; this_mode ++)
-        {
-            int this_rd;
-            int distortion;
-            int labelyrate;
-            ENTROPY_CONTEXT_PLANES t_above_s, t_left_s;
-            ENTROPY_CONTEXT *ta_s;
-            ENTROPY_CONTEXT *tl_s;
+      ta_s = (ENTROPY_CONTEXT *)&t_above_s;
+      tl_s = (ENTROPY_CONTEXT *)&t_left_s;
 
-            memcpy(&t_above_s, &t_above, sizeof(ENTROPY_CONTEXT_PLANES));
-            memcpy(&t_left_s, &t_left, sizeof(ENTROPY_CONTEXT_PLANES));
+      if (this_mode == NEW4X4) {
+        int sseshift;
+        int num00;
+        int step_param = 0;
+        int further_steps;
+        int n;
+        int thissme;
+        int bestsme = INT_MAX;
+        int_mv temp_mv;
+        BLOCK *c;
+        BLOCKD *e;
 
-            ta_s = (ENTROPY_CONTEXT *)&t_above_s;
-            tl_s = (ENTROPY_CONTEXT *)&t_left_s;
+        /* Is the best so far sufficiently good that we cant justify
+         * doing a new motion search.
+         */
+        if (best_label_rd < label_mv_thresh) break;
 
-            if (this_mode == NEW4X4)
-            {
-                int sseshift;
-                int num00;
-                int step_param = 0;
-                int further_steps;
-                int n;
-                int thissme;
-                int bestsme = INT_MAX;
-                int_mv  temp_mv;
-                BLOCK *c;
-                BLOCKD *e;
+        if (cpi->compressor_speed) {
+          if (segmentation == BLOCK_8X16 || segmentation == BLOCK_16X8) {
+            bsi->mvp.as_int = bsi->sv_mvp[i].as_int;
+            if (i == 1 && segmentation == BLOCK_16X8)
+              bsi->mvp.as_int = bsi->sv_mvp[2].as_int;
 
-                /* Is the best so far sufficiently good that we cant justify
-                 * doing a new motion search.
-                 */
-                if (best_label_rd < label_mv_thresh)
-                    break;
+            step_param = bsi->sv_istep[i];
+          }
 
-                if(cpi->compressor_speed)
-                {
-                    if (segmentation == BLOCK_8X16 || segmentation == BLOCK_16X8)
-                    {
-                        bsi->mvp.as_int = bsi->sv_mvp[i].as_int;
-                        if (i==1 && segmentation == BLOCK_16X8)
-                          bsi->mvp.as_int = bsi->sv_mvp[2].as_int;
+          /* use previous block's result as next block's MV
+           * predictor.
+           */
+          if (segmentation == BLOCK_4X4 && i > 0) {
+            bsi->mvp.as_int = x->e_mbd.block[i - 1].bmi.mv.as_int;
+            if (i == 4 || i == 8 || i == 12)
+              bsi->mvp.as_int = x->e_mbd.block[i - 4].bmi.mv.as_int;
+            step_param = 2;
+          }
+        }
 
-                        step_param = bsi->sv_istep[i];
-                    }
+        further_steps = (MAX_MVSEARCH_STEPS - 1) - step_param;
 
-                    /* use previous block's result as next block's MV
-                     * predictor.
-                     */
-                    if (segmentation == BLOCK_4X4 && i>0)
-                    {
-                        bsi->mvp.as_int = x->e_mbd.block[i-1].bmi.mv.as_int;
-                        if (i==4 || i==8 || i==12)
-                            bsi->mvp.as_int = x->e_mbd.block[i-4].bmi.mv.as_int;
-                        step_param = 2;
-                    }
-                }
+        {
+          int sadpb = x->sadperbit4;
+          int_mv mvp_full;
 
-                further_steps = (MAX_MVSEARCH_STEPS - 1) - step_param;
+          mvp_full.as_mv.row = bsi->mvp.as_mv.row >> 3;
+          mvp_full.as_mv.col = bsi->mvp.as_mv.col >> 3;
 
-                {
-                    int sadpb = x->sadperbit4;
-                    int_mv mvp_full;
+          /* find first label */
+          n = vp8_mbsplit_offset[segmentation][i];
 
-                    mvp_full.as_mv.row = bsi->mvp.as_mv.row >>3;
-                    mvp_full.as_mv.col = bsi->mvp.as_mv.col >>3;
+          c = &x->block[n];
+          e = &x->e_mbd.block[n];
 
-                    /* find first label */
-                    n = vp8_mbsplit_offset[segmentation][i];
+          {
+            bestsme = cpi->diamond_search_sad(
+                x, c, e, &mvp_full, &mode_mv[NEW4X4], step_param, sadpb, &num00,
+                v_fn_ptr, x->mvcost, bsi->ref_mv);
 
-                    c = &x->block[n];
-                    e = &x->e_mbd.block[n];
+            n = num00;
+            num00 = 0;
 
-                    {
-                        bestsme = cpi->diamond_search_sad(x, c, e, &mvp_full,
-                                                &mode_mv[NEW4X4], step_param,
-                                                sadpb, &num00, v_fn_ptr,
-                                                x->mvcost, bsi->ref_mv);
+            while (n < further_steps) {
+              n++;
 
-                        n = num00;
-                        num00 = 0;
-
-                        while (n < further_steps)
-                        {
-                            n++;
-
-                            if (num00)
-                                num00--;
-                            else
-                            {
-                                thissme = cpi->diamond_search_sad(x, c, e,
-                                                    &mvp_full, &temp_mv,
-                                                    step_param + n, sadpb,
-                                                    &num00, v_fn_ptr,
-                                                    x->mvcost, bsi->ref_mv);
+              if (num00)
+                num00--;
+              else {
+                thissme = cpi->diamond_search_sad(
+                    x, c, e, &mvp_full, &temp_mv, step_param + n, sadpb, &num00,
+                    v_fn_ptr, x->mvcost, bsi->ref_mv);
 
-                                if (thissme < bestsme)
-                                {
-                                    bestsme = thissme;
-                                    mode_mv[NEW4X4].as_int = temp_mv.as_int;
-                                }
-                            }
-                        }
-                    }
+                if (thissme < bestsme) {
+                  bestsme = thissme;
+                  mode_mv[NEW4X4].as_int = temp_mv.as_int;
+                }
+              }
+            }
+          }
 
-                    sseshift = segmentation_to_sseshift[segmentation];
+          sseshift = segmentation_to_sseshift[segmentation];
 
-                    /* Should we do a full search (best quality only) */
-                    if ((cpi->compressor_speed == 0) && (bestsme >> sseshift) > 4000)
-                    {
-                        /* Check if mvp_full is within the range. */
-                        vp8_clamp_mv(&mvp_full, x->mv_col_min, x->mv_col_max, x->mv_row_min, x->mv_row_max);
+          /* Should we do a full search (best quality only) */
+          if ((cpi->compressor_speed == 0) && (bestsme >> sseshift) > 4000) {
+            /* Check if mvp_full is within the range. */
+            vp8_clamp_mv(&mvp_full, x->mv_col_min, x->mv_col_max, x->mv_row_min,
+                         x->mv_row_max);
 
-                        thissme = cpi->full_search_sad(x, c, e, &mvp_full,
-                                                       sadpb, 16, v_fn_ptr,
-                                                       x->mvcost, bsi->ref_mv);
+            thissme = cpi->full_search_sad(x, c, e, &mvp_full, sadpb, 16,
+                                           v_fn_ptr, x->mvcost, bsi->ref_mv);
 
-                        if (thissme < bestsme)
-                        {
-                            bestsme = thissme;
-                            mode_mv[NEW4X4].as_int = e->bmi.mv.as_int;
-                        }
-                        else
-                        {
-                            /* The full search result is actually worse so
-                             * re-instate the previous best vector
-                             */
-                            e->bmi.mv.as_int = mode_mv[NEW4X4].as_int;
-                        }
-                    }
-                }
+            if (thissme < bestsme) {
+              bestsme = thissme;
+              mode_mv[NEW4X4].as_int = e->bmi.mv.as_int;
+            } else {
+              /* The full search result is actually worse so
+               * re-instate the previous best vector
+               */
+              e->bmi.mv.as_int = mode_mv[NEW4X4].as_int;
+            }
+          }
+        }
+
+        if (bestsme < INT_MAX) {
+          int disto;
+          unsigned int sse;
+          cpi->find_fractional_mv_step(x, c, e, &mode_mv[NEW4X4], bsi->ref_mv,
+                                       x->errorperbit, v_fn_ptr, x->mvcost,
+                                       &disto, &sse);
+        }
+      } /* NEW4X4 */
 
-                if (bestsme < INT_MAX)
-                {
-                    int disto;
-                    unsigned int sse;
-                    cpi->find_fractional_mv_step(x, c, e, &mode_mv[NEW4X4],
-                        bsi->ref_mv, x->errorperbit, v_fn_ptr, x->mvcost,
-                        &disto, &sse);
-                }
-            } /* NEW4X4 */
+      rate = labels2mode(x, labels, i, this_mode, &mode_mv[this_mode],
+                         bsi->ref_mv, x->mvcost);
 
-            rate = labels2mode(x, labels, i, this_mode, &mode_mv[this_mode],
-                               bsi->ref_mv, x->mvcost);
+      /* Trap vectors that reach beyond the UMV borders */
+      if (((mode_mv[this_mode].as_mv.row >> 3) < x->mv_row_min) ||
+          ((mode_mv[this_mode].as_mv.row >> 3) > x->mv_row_max) ||
+          ((mode_mv[this_mode].as_mv.col >> 3) < x->mv_col_min) ||
+          ((mode_mv[this_mode].as_mv.col >> 3) > x->mv_col_max)) {
+        continue;
+      }
 
-            /* Trap vectors that reach beyond the UMV borders */
-            if (((mode_mv[this_mode].as_mv.row >> 3) < x->mv_row_min) || ((mode_mv[this_mode].as_mv.row >> 3) > x->mv_row_max) ||
-                ((mode_mv[this_mode].as_mv.col >> 3) < x->mv_col_min) || ((mode_mv[this_mode].as_mv.col >> 3) > x->mv_col_max))
-            {
-                continue;
-            }
+      distortion = vp8_encode_inter_mb_segment(x, labels, i) / 4;
 
-            distortion = vp8_encode_inter_mb_segment(x, labels, i) / 4;
+      labelyrate = rdcost_mbsegment_y(x, labels, i, ta_s, tl_s);
+      rate += labelyrate;
 
-            labelyrate = rdcost_mbsegment_y(x, labels, i, ta_s, tl_s);
-            rate += labelyrate;
+      this_rd = RDCOST(x->rdmult, x->rddiv, rate, distortion);
 
-            this_rd = RDCOST(x->rdmult, x->rddiv, rate, distortion);
+      if (this_rd < best_label_rd) {
+        sbr = rate;
+        sbd = distortion;
+        bestlabelyrate = labelyrate;
+        mode_selected = this_mode;
+        best_label_rd = this_rd;
 
-            if (this_rd < best_label_rd)
-            {
-                sbr = rate;
-                sbd = distortion;
-                bestlabelyrate = labelyrate;
-                mode_selected = this_mode;
-                best_label_rd = this_rd;
+        memcpy(ta_b, ta_s, sizeof(ENTROPY_CONTEXT_PLANES));
+        memcpy(tl_b, tl_s, sizeof(ENTROPY_CONTEXT_PLANES));
+      }
+    } /*for each 4x4 mode*/
 
-                memcpy(ta_b, ta_s, sizeof(ENTROPY_CONTEXT_PLANES));
-                memcpy(tl_b, tl_s, sizeof(ENTROPY_CONTEXT_PLANES));
+    memcpy(ta, ta_b, sizeof(ENTROPY_CONTEXT_PLANES));
+    memcpy(tl, tl_b, sizeof(ENTROPY_CONTEXT_PLANES));
 
-            }
-        } /*for each 4x4 mode*/
+    labels2mode(x, labels, i, mode_selected, &mode_mv[mode_selected],
+                bsi->ref_mv, x->mvcost);
 
-        memcpy(ta, ta_b, sizeof(ENTROPY_CONTEXT_PLANES));
-        memcpy(tl, tl_b, sizeof(ENTROPY_CONTEXT_PLANES));
+    br += sbr;
+    bd += sbd;
+    segmentyrate += bestlabelyrate;
+    this_segment_rd += best_label_rd;
 
-        labels2mode(x, labels, i, mode_selected, &mode_mv[mode_selected],
-                    bsi->ref_mv, x->mvcost);
+    if (this_segment_rd >= bsi->segment_rd) break;
 
-        br += sbr;
-        bd += sbd;
-        segmentyrate += bestlabelyrate;
-        this_segment_rd += best_label_rd;
+  } /* for each label */
 
-        if (this_segment_rd >= bsi->segment_rd)
-            break;
+  if (this_segment_rd < bsi->segment_rd) {
+    bsi->r = br;
+    bsi->d = bd;
+    bsi->segment_yrate = segmentyrate;
+    bsi->segment_rd = this_segment_rd;
+    bsi->segment_num = segmentation;
 
-    } /* for each label */
-
-    if (this_segment_rd < bsi->segment_rd)
-    {
-        bsi->r = br;
-        bsi->d = bd;
-        bsi->segment_yrate = segmentyrate;
-        bsi->segment_rd = this_segment_rd;
-        bsi->segment_num = segmentation;
-
-        /* store everything needed to come back to this!! */
-        for (i = 0; i < 16; i++)
-        {
-            bsi->mvs[i].as_mv = x->partition_info->bmi[i].mv.as_mv;
-            bsi->modes[i] = x->partition_info->bmi[i].mode;
-            bsi->eobs[i] = x->e_mbd.eobs[i];
-        }
+    /* store everything needed to come back to this!! */
+    for (i = 0; i < 16; i++) {
+      bsi->mvs[i].as_mv = x->partition_info->bmi[i].mv.as_mv;
+      bsi->modes[i] = x->partition_info->bmi[i].mode;
+      bsi->eobs[i] = x->e_mbd.eobs[i];
     }
+  }
 }
 
-static
-void vp8_cal_step_param(int sr, int *sp)
-{
-    int step = 0;
+static void vp8_cal_step_param(int sr, int *sp) {
+  int step = 0;
 
-    if (sr > MAX_FIRST_STEP) sr = MAX_FIRST_STEP;
-    else if (sr < 1) sr = 1;
+  if (sr > MAX_FIRST_STEP)
+    sr = MAX_FIRST_STEP;
+  else if (sr < 1)
+    sr = 1;
 
-    while (sr>>=1)
-        step++;
+  while (sr >>= 1) step++;
 
-    *sp = MAX_MVSEARCH_STEPS - 1 - step;
+  *sp = MAX_MVSEARCH_STEPS - 1 - step;
 }
 
 static int vp8_rd_pick_best_mbsegmentation(VP8_COMP *cpi, MACROBLOCK *x,
                                            int_mv *best_ref_mv, int best_rd,
                                            int *mdcounts, int *returntotrate,
-                                           int *returnyrate, int *returndistortion,
-                                           int mvthresh)
-{
-    int i;
-    BEST_SEG_INFO bsi;
+                                           int *returnyrate,
+                                           int *returndistortion,
+                                           int mvthresh) {
+  int i;
+  BEST_SEG_INFO bsi;
 
-    memset(&bsi, 0, sizeof(bsi));
+  memset(&bsi, 0, sizeof(bsi));
 
-    bsi.segment_rd = best_rd;
-    bsi.ref_mv = best_ref_mv;
-    bsi.mvp.as_int = best_ref_mv->as_int;
-    bsi.mvthresh = mvthresh;
-    bsi.mdcounts = mdcounts;
+  bsi.segment_rd = best_rd;
+  bsi.ref_mv = best_ref_mv;
+  bsi.mvp.as_int = best_ref_mv->as_int;
+  bsi.mvthresh = mvthresh;
+  bsi.mdcounts = mdcounts;
 
-    for(i = 0; i < 16; i++)
-    {
-        bsi.modes[i] = ZERO4X4;
-    }
+  for (i = 0; i < 16; i++) {
+    bsi.modes[i] = ZERO4X4;
+  }
 
-    if(cpi->compressor_speed == 0)
-    {
-        /* for now, we will keep the original segmentation order
-           when in best quality mode */
-        rd_check_segment(cpi, x, &bsi, BLOCK_16X8);
-        rd_check_segment(cpi, x, &bsi, BLOCK_8X16);
-        rd_check_segment(cpi, x, &bsi, BLOCK_8X8);
-        rd_check_segment(cpi, x, &bsi, BLOCK_4X4);
-    }
-    else
-    {
-        int sr;
+  if (cpi->compressor_speed == 0) {
+    /* for now, we will keep the original segmentation order
+       when in best quality mode */
+    rd_check_segment(cpi, x, &bsi, BLOCK_16X8);
+    rd_check_segment(cpi, x, &bsi, BLOCK_8X16);
+    rd_check_segment(cpi, x, &bsi, BLOCK_8X8);
+    rd_check_segment(cpi, x, &bsi, BLOCK_4X4);
+  } else {
+    int sr;
 
-        rd_check_segment(cpi, x, &bsi, BLOCK_8X8);
+    rd_check_segment(cpi, x, &bsi, BLOCK_8X8);
 
-        if (bsi.segment_rd < best_rd)
-        {
-            int col_min = ((best_ref_mv->as_mv.col+7)>>3) - MAX_FULL_PEL_VAL;
-            int row_min = ((best_ref_mv->as_mv.row+7)>>3) - MAX_FULL_PEL_VAL;
-            int col_max = (best_ref_mv->as_mv.col>>3) + MAX_FULL_PEL_VAL;
-            int row_max = (best_ref_mv->as_mv.row>>3) + MAX_FULL_PEL_VAL;
+    if (bsi.segment_rd < best_rd) {
+      int col_min = ((best_ref_mv->as_mv.col + 7) >> 3) - MAX_FULL_PEL_VAL;
+      int row_min = ((best_ref_mv->as_mv.row + 7) >> 3) - MAX_FULL_PEL_VAL;
+      int col_max = (best_ref_mv->as_mv.col >> 3) + MAX_FULL_PEL_VAL;
+      int row_max = (best_ref_mv->as_mv.row >> 3) + MAX_FULL_PEL_VAL;
 
-            int tmp_col_min = x->mv_col_min;
-            int tmp_col_max = x->mv_col_max;
-            int tmp_row_min = x->mv_row_min;
-            int tmp_row_max = x->mv_row_max;
+      int tmp_col_min = x->mv_col_min;
+      int tmp_col_max = x->mv_col_max;
+      int tmp_row_min = x->mv_row_min;
+      int tmp_row_max = x->mv_row_max;
 
-            /* Get intersection of UMV window and valid MV window to reduce # of checks in diamond search. */
-            if (x->mv_col_min < col_min )
-                x->mv_col_min = col_min;
-            if (x->mv_col_max > col_max )
-                x->mv_col_max = col_max;
-            if (x->mv_row_min < row_min )
-                x->mv_row_min = row_min;
-            if (x->mv_row_max > row_max )
-                x->mv_row_max = row_max;
+      /* Get intersection of UMV window and valid MV window to reduce # of
+       * checks in diamond search. */
+      if (x->mv_col_min < col_min) x->mv_col_min = col_min;
+      if (x->mv_col_max > col_max) x->mv_col_max = col_max;
+      if (x->mv_row_min < row_min) x->mv_row_min = row_min;
+      if (x->mv_row_max > row_max) x->mv_row_max = row_max;
 
-            /* Get 8x8 result */
-            bsi.sv_mvp[0].as_int = bsi.mvs[0].as_int;
-            bsi.sv_mvp[1].as_int = bsi.mvs[2].as_int;
-            bsi.sv_mvp[2].as_int = bsi.mvs[8].as_int;
-            bsi.sv_mvp[3].as_int = bsi.mvs[10].as_int;
+      /* Get 8x8 result */
+      bsi.sv_mvp[0].as_int = bsi.mvs[0].as_int;
+      bsi.sv_mvp[1].as_int = bsi.mvs[2].as_int;
+      bsi.sv_mvp[2].as_int = bsi.mvs[8].as_int;
+      bsi.sv_mvp[3].as_int = bsi.mvs[10].as_int;
 
-            /* Use 8x8 result as 16x8/8x16's predictor MV. Adjust search range according to the closeness of 2 MV. */
-            /* block 8X16 */
-            {
-                sr = MAXF((abs(bsi.sv_mvp[0].as_mv.row - bsi.sv_mvp[2].as_mv.row))>>3, (abs(bsi.sv_mvp[0].as_mv.col - bsi.sv_mvp[2].as_mv.col))>>3);
-                vp8_cal_step_param(sr, &bsi.sv_istep[0]);
+      /* Use 8x8 result as 16x8/8x16's predictor MV. Adjust search range
+       * according to the closeness of 2 MV. */
+      /* block 8X16 */
+      {
+        sr =
+            MAXF((abs(bsi.sv_mvp[0].as_mv.row - bsi.sv_mvp[2].as_mv.row)) >> 3,
+                 (abs(bsi.sv_mvp[0].as_mv.col - bsi.sv_mvp[2].as_mv.col)) >> 3);
+        vp8_cal_step_param(sr, &bsi.sv_istep[0]);
 
-                sr = MAXF((abs(bsi.sv_mvp[1].as_mv.row - bsi.sv_mvp[3].as_mv.row))>>3, (abs(bsi.sv_mvp[1].as_mv.col - bsi.sv_mvp[3].as_mv.col))>>3);
-                vp8_cal_step_param(sr, &bsi.sv_istep[1]);
+        sr =
+            MAXF((abs(bsi.sv_mvp[1].as_mv.row - bsi.sv_mvp[3].as_mv.row)) >> 3,
+                 (abs(bsi.sv_mvp[1].as_mv.col - bsi.sv_mvp[3].as_mv.col)) >> 3);
+        vp8_cal_step_param(sr, &bsi.sv_istep[1]);
 
-                rd_check_segment(cpi, x, &bsi, BLOCK_8X16);
-            }
+        rd_check_segment(cpi, x, &bsi, BLOCK_8X16);
+      }
 
-            /* block 16X8 */
-            {
-                sr = MAXF((abs(bsi.sv_mvp[0].as_mv.row - bsi.sv_mvp[1].as_mv.row))>>3, (abs(bsi.sv_mvp[0].as_mv.col - bsi.sv_mvp[1].as_mv.col))>>3);
-                vp8_cal_step_param(sr, &bsi.sv_istep[0]);
+      /* block 16X8 */
+      {
+        sr =
+            MAXF((abs(bsi.sv_mvp[0].as_mv.row - bsi.sv_mvp[1].as_mv.row)) >> 3,
+                 (abs(bsi.sv_mvp[0].as_mv.col - bsi.sv_mvp[1].as_mv.col)) >> 3);
+        vp8_cal_step_param(sr, &bsi.sv_istep[0]);
 
-                sr = MAXF((abs(bsi.sv_mvp[2].as_mv.row - bsi.sv_mvp[3].as_mv.row))>>3, (abs(bsi.sv_mvp[2].as_mv.col - bsi.sv_mvp[3].as_mv.col))>>3);
-                vp8_cal_step_param(sr, &bsi.sv_istep[1]);
+        sr =
+            MAXF((abs(bsi.sv_mvp[2].as_mv.row - bsi.sv_mvp[3].as_mv.row)) >> 3,
+                 (abs(bsi.sv_mvp[2].as_mv.col - bsi.sv_mvp[3].as_mv.col)) >> 3);
+        vp8_cal_step_param(sr, &bsi.sv_istep[1]);
 
-                rd_check_segment(cpi, x, &bsi, BLOCK_16X8);
-            }
+        rd_check_segment(cpi, x, &bsi, BLOCK_16X8);
+      }
 
-            /* If 8x8 is better than 16x8/8x16, then do 4x4 search */
-            /* Not skip 4x4 if speed=0 (good quality) */
-            if (cpi->sf.no_skip_block4x4_search || bsi.segment_num == BLOCK_8X8)  /* || (sv_segment_rd8x8-bsi.segment_rd) < sv_segment_rd8x8>>5) */
-            {
-                bsi.mvp.as_int = bsi.sv_mvp[0].as_int;
-                rd_check_segment(cpi, x, &bsi, BLOCK_4X4);
-            }
+      /* If 8x8 is better than 16x8/8x16, then do 4x4 search */
+      /* Not skip 4x4 if speed=0 (good quality) */
+      if (cpi->sf.no_skip_block4x4_search || bsi.segment_num == BLOCK_8X8)
+      /* || (sv_segment_rd8x8-bsi.segment_rd) < sv_segment_rd8x8>>5) */
+      {
+        bsi.mvp.as_int = bsi.sv_mvp[0].as_int;
+        rd_check_segment(cpi, x, &bsi, BLOCK_4X4);
+      }
 
-            /* restore UMV window */
-            x->mv_col_min = tmp_col_min;
-            x->mv_col_max = tmp_col_max;
-            x->mv_row_min = tmp_row_min;
-            x->mv_row_max = tmp_row_max;
-        }
+      /* restore UMV window */
+      x->mv_col_min = tmp_col_min;
+      x->mv_col_max = tmp_col_max;
+      x->mv_row_min = tmp_row_min;
+      x->mv_row_max = tmp_row_max;
     }
+  }
 
-    /* set it to the best */
-    for (i = 0; i < 16; i++)
-    {
-        BLOCKD *bd = &x->e_mbd.block[i];
+  /* set it to the best */
+  for (i = 0; i < 16; i++) {
+    BLOCKD *bd = &x->e_mbd.block[i];
 
-        bd->bmi.mv.as_int = bsi.mvs[i].as_int;
-        *bd->eob = bsi.eobs[i];
-    }
+    bd->bmi.mv.as_int = bsi.mvs[i].as_int;
+    *bd->eob = bsi.eobs[i];
+  }
 
-    *returntotrate = bsi.r;
-    *returndistortion = bsi.d;
-    *returnyrate = bsi.segment_yrate;
+  *returntotrate = bsi.r;
+  *returndistortion = bsi.d;
+  *returnyrate = bsi.segment_yrate;
 
-    /* save partitions */
-    x->e_mbd.mode_info_context->mbmi.partitioning = bsi.segment_num;
-    x->partition_info->count = vp8_mbsplit_count[bsi.segment_num];
+  /* save partitions */
+  x->e_mbd.mode_info_context->mbmi.partitioning = bsi.segment_num;
+  x->partition_info->count = vp8_mbsplit_count[bsi.segment_num];
 
-    for (i = 0; i < x->partition_info->count; i++)
-    {
-        int j;
+  for (i = 0; i < x->partition_info->count; i++) {
+    int j;
 
-        j = vp8_mbsplit_offset[bsi.segment_num][i];
+    j = vp8_mbsplit_offset[bsi.segment_num][i];
 
-        x->partition_info->bmi[i].mode = bsi.modes[j];
-        x->partition_info->bmi[i].mv.as_mv = bsi.mvs[j].as_mv;
-    }
-    /*
-     * used to set x->e_mbd.mode_info_context->mbmi.mv.as_int
-     */
-    x->partition_info->bmi[15].mv.as_int = bsi.mvs[15].as_int;
+    x->partition_info->bmi[i].mode = bsi.modes[j];
+    x->partition_info->bmi[i].mv.as_mv = bsi.mvs[j].as_mv;
+  }
+  /*
+   * used to set x->e_mbd.mode_info_context->mbmi.mv.as_int
+   */
+  x->partition_info->bmi[15].mv.as_int = bsi.mvs[15].as_int;
 
-    return bsi.segment_rd;
+  return bsi.segment_rd;
 }
 
 /* The improved MV prediction */
-void vp8_mv_pred
-(
-    VP8_COMP *cpi,
-    MACROBLOCKD *xd,
-    const MODE_INFO *here,
-    int_mv *mvp,
-    int refframe,
-    int *ref_frame_sign_bias,
-    int *sr,
-    int near_sadidx[]
-)
-{
-    const MODE_INFO *above = here - xd->mode_info_stride;
-    const MODE_INFO *left = here - 1;
-    const MODE_INFO *aboveleft = above - 1;
-    int_mv           near_mvs[8];
-    int              near_ref[8];
-    int_mv           mv;
-    int              vcnt=0;
-    int              find=0;
-    int              mb_offset;
+void vp8_mv_pred(VP8_COMP *cpi, MACROBLOCKD *xd, const MODE_INFO *here,
+                 int_mv *mvp, int refframe, int *ref_frame_sign_bias, int *sr,
+                 int near_sadidx[]) {
+  const MODE_INFO *above = here - xd->mode_info_stride;
+  const MODE_INFO *left = here - 1;
+  const MODE_INFO *aboveleft = above - 1;
+  int_mv near_mvs[8];
+  int near_ref[8];
+  int_mv mv;
+  int vcnt = 0;
+  int find = 0;
+  int mb_offset;
 
-    int              mvx[8];
-    int              mvy[8];
-    int              i;
+  int mvx[8];
+  int mvy[8];
+  int i;
 
-    mv.as_int = 0;
+  mv.as_int = 0;
 
-    if(here->mbmi.ref_frame != INTRA_FRAME)
-    {
-        near_mvs[0].as_int = near_mvs[1].as_int = near_mvs[2].as_int = near_mvs[3].as_int = near_mvs[4].as_int = near_mvs[5].as_int = near_mvs[6].as_int = near_mvs[7].as_int = 0;
-        near_ref[0] = near_ref[1] = near_ref[2] = near_ref[3] = near_ref[4] = near_ref[5] = near_ref[6] = near_ref[7] = 0;
+  if (here->mbmi.ref_frame != INTRA_FRAME) {
+    near_mvs[0].as_int = near_mvs[1].as_int = near_mvs[2].as_int =
+        near_mvs[3].as_int = near_mvs[4].as_int = near_mvs[5].as_int =
+            near_mvs[6].as_int = near_mvs[7].as_int = 0;
+    near_ref[0] = near_ref[1] = near_ref[2] = near_ref[3] = near_ref[4] =
+        near_ref[5] = near_ref[6] = near_ref[7] = 0;
 
-        /* read in 3 nearby block's MVs from current frame as prediction
-         * candidates.
-         */
-        if (above->mbmi.ref_frame != INTRA_FRAME)
-        {
-            near_mvs[vcnt].as_int = above->mbmi.mv.as_int;
-            mv_bias(ref_frame_sign_bias[above->mbmi.ref_frame], refframe, &near_mvs[vcnt], ref_frame_sign_bias);
-            near_ref[vcnt] =  above->mbmi.ref_frame;
-        }
-        vcnt++;
-        if (left->mbmi.ref_frame != INTRA_FRAME)
-        {
-            near_mvs[vcnt].as_int = left->mbmi.mv.as_int;
-            mv_bias(ref_frame_sign_bias[left->mbmi.ref_frame], refframe, &near_mvs[vcnt], ref_frame_sign_bias);
-            near_ref[vcnt] =  left->mbmi.ref_frame;
-        }
-        vcnt++;
-        if (aboveleft->mbmi.ref_frame != INTRA_FRAME)
-        {
-            near_mvs[vcnt].as_int = aboveleft->mbmi.mv.as_int;
-            mv_bias(ref_frame_sign_bias[aboveleft->mbmi.ref_frame], refframe, &near_mvs[vcnt], ref_frame_sign_bias);
-            near_ref[vcnt] =  aboveleft->mbmi.ref_frame;
-        }
-        vcnt++;
+    /* read in 3 nearby block's MVs from current frame as prediction
+     * candidates.
+     */
+    if (above->mbmi.ref_frame != INTRA_FRAME) {
+      near_mvs[vcnt].as_int = above->mbmi.mv.as_int;
+      mv_bias(ref_frame_sign_bias[above->mbmi.ref_frame], refframe,
+              &near_mvs[vcnt], ref_frame_sign_bias);
+      near_ref[vcnt] = above->mbmi.ref_frame;
+    }
+    vcnt++;
+    if (left->mbmi.ref_frame != INTRA_FRAME) {
+      near_mvs[vcnt].as_int = left->mbmi.mv.as_int;
+      mv_bias(ref_frame_sign_bias[left->mbmi.ref_frame], refframe,
+              &near_mvs[vcnt], ref_frame_sign_bias);
+      near_ref[vcnt] = left->mbmi.ref_frame;
+    }
+    vcnt++;
+    if (aboveleft->mbmi.ref_frame != INTRA_FRAME) {
+      near_mvs[vcnt].as_int = aboveleft->mbmi.mv.as_int;
+      mv_bias(ref_frame_sign_bias[aboveleft->mbmi.ref_frame], refframe,
+              &near_mvs[vcnt], ref_frame_sign_bias);
+      near_ref[vcnt] = aboveleft->mbmi.ref_frame;
+    }
+    vcnt++;
 
-        /* read in 5 nearby block's MVs from last frame. */
-        if(cpi->common.last_frame_type != KEY_FRAME)
-        {
-            mb_offset = (-xd->mb_to_top_edge/128 + 1) * (xd->mode_info_stride +1) + (-xd->mb_to_left_edge/128 +1) ;
+    /* read in 5 nearby block's MVs from last frame. */
+    if (cpi->common.last_frame_type != KEY_FRAME) {
+      mb_offset = (-xd->mb_to_top_edge / 128 + 1) * (xd->mode_info_stride + 1) +
+                  (-xd->mb_to_left_edge / 128 + 1);
 
-            /* current in last frame */
-            if (cpi->lf_ref_frame[mb_offset] != INTRA_FRAME)
-            {
-                near_mvs[vcnt].as_int = cpi->lfmv[mb_offset].as_int;
-                mv_bias(cpi->lf_ref_frame_sign_bias[mb_offset], refframe, &near_mvs[vcnt], ref_frame_sign_bias);
-                near_ref[vcnt] =  cpi->lf_ref_frame[mb_offset];
-            }
-            vcnt++;
+      /* current in last frame */
+      if (cpi->lf_ref_frame[mb_offset] != INTRA_FRAME) {
+        near_mvs[vcnt].as_int = cpi->lfmv[mb_offset].as_int;
+        mv_bias(cpi->lf_ref_frame_sign_bias[mb_offset], refframe,
+                &near_mvs[vcnt], ref_frame_sign_bias);
+        near_ref[vcnt] = cpi->lf_ref_frame[mb_offset];
+      }
+      vcnt++;
 
-            /* above in last frame */
-            if (cpi->lf_ref_frame[mb_offset - xd->mode_info_stride-1] != INTRA_FRAME)
-            {
-                near_mvs[vcnt].as_int = cpi->lfmv[mb_offset - xd->mode_info_stride-1].as_int;
-                mv_bias(cpi->lf_ref_frame_sign_bias[mb_offset - xd->mode_info_stride-1], refframe, &near_mvs[vcnt], ref_frame_sign_bias);
-                near_ref[vcnt] =  cpi->lf_ref_frame[mb_offset - xd->mode_info_stride-1];
-            }
-            vcnt++;
+      /* above in last frame */
+      if (cpi->lf_ref_frame[mb_offset - xd->mode_info_stride - 1] !=
+          INTRA_FRAME) {
+        near_mvs[vcnt].as_int =
+            cpi->lfmv[mb_offset - xd->mode_info_stride - 1].as_int;
+        mv_bias(
+            cpi->lf_ref_frame_sign_bias[mb_offset - xd->mode_info_stride - 1],
+            refframe, &near_mvs[vcnt], ref_frame_sign_bias);
+        near_ref[vcnt] =
+            cpi->lf_ref_frame[mb_offset - xd->mode_info_stride - 1];
+      }
+      vcnt++;
 
-            /* left in last frame */
-            if (cpi->lf_ref_frame[mb_offset-1] != INTRA_FRAME)
-            {
-                near_mvs[vcnt].as_int = cpi->lfmv[mb_offset -1].as_int;
-                mv_bias(cpi->lf_ref_frame_sign_bias[mb_offset -1], refframe, &near_mvs[vcnt], ref_frame_sign_bias);
-                near_ref[vcnt] =  cpi->lf_ref_frame[mb_offset - 1];
-            }
-            vcnt++;
+      /* left in last frame */
+      if (cpi->lf_ref_frame[mb_offset - 1] != INTRA_FRAME) {
+        near_mvs[vcnt].as_int = cpi->lfmv[mb_offset - 1].as_int;
+        mv_bias(cpi->lf_ref_frame_sign_bias[mb_offset - 1], refframe,
+                &near_mvs[vcnt], ref_frame_sign_bias);
+        near_ref[vcnt] = cpi->lf_ref_frame[mb_offset - 1];
+      }
+      vcnt++;
 
-            /* right in last frame */
-            if (cpi->lf_ref_frame[mb_offset +1] != INTRA_FRAME)
-            {
-                near_mvs[vcnt].as_int = cpi->lfmv[mb_offset +1].as_int;
-                mv_bias(cpi->lf_ref_frame_sign_bias[mb_offset +1], refframe, &near_mvs[vcnt], ref_frame_sign_bias);
-                near_ref[vcnt] =  cpi->lf_ref_frame[mb_offset +1];
-            }
-            vcnt++;
+      /* right in last frame */
+      if (cpi->lf_ref_frame[mb_offset + 1] != INTRA_FRAME) {
+        near_mvs[vcnt].as_int = cpi->lfmv[mb_offset + 1].as_int;
+        mv_bias(cpi->lf_ref_frame_sign_bias[mb_offset + 1], refframe,
+                &near_mvs[vcnt], ref_frame_sign_bias);
+        near_ref[vcnt] = cpi->lf_ref_frame[mb_offset + 1];
+      }
+      vcnt++;
 
-            /* below in last frame */
-            if (cpi->lf_ref_frame[mb_offset + xd->mode_info_stride +1] != INTRA_FRAME)
-            {
-                near_mvs[vcnt].as_int = cpi->lfmv[mb_offset + xd->mode_info_stride +1].as_int;
-                mv_bias(cpi->lf_ref_frame_sign_bias[mb_offset + xd->mode_info_stride +1], refframe, &near_mvs[vcnt], ref_frame_sign_bias);
-                near_ref[vcnt] =  cpi->lf_ref_frame[mb_offset + xd->mode_info_stride +1];
-            }
-            vcnt++;
-        }
+      /* below in last frame */
+      if (cpi->lf_ref_frame[mb_offset + xd->mode_info_stride + 1] !=
+          INTRA_FRAME) {
+        near_mvs[vcnt].as_int =
+            cpi->lfmv[mb_offset + xd->mode_info_stride + 1].as_int;
+        mv_bias(
+            cpi->lf_ref_frame_sign_bias[mb_offset + xd->mode_info_stride + 1],
+            refframe, &near_mvs[vcnt], ref_frame_sign_bias);
+        near_ref[vcnt] =
+            cpi->lf_ref_frame[mb_offset + xd->mode_info_stride + 1];
+      }
+      vcnt++;
+    }
 
-        for(i=0; i< vcnt; i++)
-        {
-            if(near_ref[near_sadidx[i]] != INTRA_FRAME)
-            {
-                if(here->mbmi.ref_frame == near_ref[near_sadidx[i]])
-                {
-                    mv.as_int = near_mvs[near_sadidx[i]].as_int;
-                    find = 1;
-                    if (i < 3)
-                        *sr = 3;
-                    else
-                        *sr = 2;
-                    break;
-                }
-            }
+    for (i = 0; i < vcnt; i++) {
+      if (near_ref[near_sadidx[i]] != INTRA_FRAME) {
+        if (here->mbmi.ref_frame == near_ref[near_sadidx[i]]) {
+          mv.as_int = near_mvs[near_sadidx[i]].as_int;
+          find = 1;
+          if (i < 3)
+            *sr = 3;
+          else
+            *sr = 2;
+          break;
         }
+      }
+    }
 
-        if(!find)
-        {
-            for(i=0; i<vcnt; i++)
-            {
-                mvx[i] = near_mvs[i].as_mv.row;
-                mvy[i] = near_mvs[i].as_mv.col;
-            }
+    if (!find) {
+      for (i = 0; i < vcnt; i++) {
+        mvx[i] = near_mvs[i].as_mv.row;
+        mvy[i] = near_mvs[i].as_mv.col;
+      }
 
-            insertsortmv(mvx, vcnt);
-            insertsortmv(mvy, vcnt);
-            mv.as_mv.row = mvx[vcnt/2];
-            mv.as_mv.col = mvy[vcnt/2];
+      insertsortmv(mvx, vcnt);
+      insertsortmv(mvy, vcnt);
+      mv.as_mv.row = mvx[vcnt / 2];
+      mv.as_mv.col = mvy[vcnt / 2];
 
-            /* sr is set to 0 to allow calling function to decide the search
-             * range.
-             */
-            *sr = 0;
-        }
+      /* sr is set to 0 to allow calling function to decide the search
+       * range.
+       */
+      *sr = 0;
     }
+  }
 
-    /* Set up return values */
-    mvp->as_int = mv.as_int;
-    vp8_clamp_mv2(mvp, xd);
+  /* Set up return values */
+  mvp->as_int = mv.as_int;
+  vp8_clamp_mv2(mvp, xd);
 }
 
-void vp8_cal_sad(VP8_COMP *cpi, MACROBLOCKD *xd, MACROBLOCK *x, int recon_yoffset, int near_sadidx[])
-{
-    /* near_sad indexes:
-     *   0-cf above, 1-cf left, 2-cf aboveleft,
-     *   3-lf current, 4-lf above, 5-lf left, 6-lf right, 7-lf below
-     */
-    int near_sad[8] = {0};
-    BLOCK *b = &x->block[0];
-    unsigned char *src_y_ptr = *(b->base_src);
+void vp8_cal_sad(VP8_COMP *cpi, MACROBLOCKD *xd, MACROBLOCK *x,
+                 int recon_yoffset, int near_sadidx[]) {
+  /* near_sad indexes:
+   *   0-cf above, 1-cf left, 2-cf aboveleft,
+   *   3-lf current, 4-lf above, 5-lf left, 6-lf right, 7-lf below
+   */
+  int near_sad[8] = { 0 };
+  BLOCK *b = &x->block[0];
+  unsigned char *src_y_ptr = *(b->base_src);
 
-    /* calculate sad for current frame 3 nearby MBs. */
-    if( xd->mb_to_top_edge==0 && xd->mb_to_left_edge ==0)
-    {
-        near_sad[0] = near_sad[1] = near_sad[2] = INT_MAX;
-    }else if(xd->mb_to_top_edge==0)
-    {   /* only has left MB for sad calculation. */
-        near_sad[0] = near_sad[2] = INT_MAX;
-        near_sad[1] = cpi->fn_ptr[BLOCK_16X16].sdf(src_y_ptr, b->src_stride, xd->dst.y_buffer - 16,xd->dst.y_stride);
-    }else if(xd->mb_to_left_edge ==0)
-    {   /* only has left MB for sad calculation. */
-        near_sad[1] = near_sad[2] = INT_MAX;
-        near_sad[0] = cpi->fn_ptr[BLOCK_16X16].sdf(src_y_ptr, b->src_stride, xd->dst.y_buffer - xd->dst.y_stride *16,xd->dst.y_stride);
-    }else
-    {
-        near_sad[0] = cpi->fn_ptr[BLOCK_16X16].sdf(src_y_ptr, b->src_stride, xd->dst.y_buffer - xd->dst.y_stride *16,xd->dst.y_stride);
-        near_sad[1] = cpi->fn_ptr[BLOCK_16X16].sdf(src_y_ptr, b->src_stride, xd->dst.y_buffer - 16,xd->dst.y_stride);
-        near_sad[2] = cpi->fn_ptr[BLOCK_16X16].sdf(src_y_ptr, b->src_stride, xd->dst.y_buffer - xd->dst.y_stride *16 -16,xd->dst.y_stride);
-    }
+  /* calculate sad for current frame 3 nearby MBs. */
+  if (xd->mb_to_top_edge == 0 && xd->mb_to_left_edge == 0) {
+    near_sad[0] = near_sad[1] = near_sad[2] = INT_MAX;
+  } else if (xd->mb_to_top_edge ==
+             0) { /* only has left MB for sad calculation. */
+    near_sad[0] = near_sad[2] = INT_MAX;
+    near_sad[1] = cpi->fn_ptr[BLOCK_16X16].sdf(
+        src_y_ptr, b->src_stride, xd->dst.y_buffer - 16, xd->dst.y_stride);
+  } else if (xd->mb_to_left_edge ==
+             0) { /* only has left MB for sad calculation. */
+    near_sad[1] = near_sad[2] = INT_MAX;
+    near_sad[0] = cpi->fn_ptr[BLOCK_16X16].sdf(
+        src_y_ptr, b->src_stride, xd->dst.y_buffer - xd->dst.y_stride * 16,
+        xd->dst.y_stride);
+  } else {
+    near_sad[0] = cpi->fn_ptr[BLOCK_16X16].sdf(
+        src_y_ptr, b->src_stride, xd->dst.y_buffer - xd->dst.y_stride * 16,
+        xd->dst.y_stride);
+    near_sad[1] = cpi->fn_ptr[BLOCK_16X16].sdf(
+        src_y_ptr, b->src_stride, xd->dst.y_buffer - 16, xd->dst.y_stride);
+    near_sad[2] = cpi->fn_ptr[BLOCK_16X16].sdf(
+        src_y_ptr, b->src_stride, xd->dst.y_buffer - xd->dst.y_stride * 16 - 16,
+        xd->dst.y_stride);
+  }
 
-    if(cpi->common.last_frame_type != KEY_FRAME)
-    {
-        /* calculate sad for last frame 5 nearby MBs. */
-        unsigned char *pre_y_buffer = cpi->common.yv12_fb[cpi->common.lst_fb_idx].y_buffer + recon_yoffset;
-        int pre_y_stride = cpi->common.yv12_fb[cpi->common.lst_fb_idx].y_stride;
+  if (cpi->common.last_frame_type != KEY_FRAME) {
+    /* calculate sad for last frame 5 nearby MBs. */
+    unsigned char *pre_y_buffer =
+        cpi->common.yv12_fb[cpi->common.lst_fb_idx].y_buffer + recon_yoffset;
+    int pre_y_stride = cpi->common.yv12_fb[cpi->common.lst_fb_idx].y_stride;
 
-        if(xd->mb_to_top_edge==0) near_sad[4] = INT_MAX;
-        if(xd->mb_to_left_edge ==0) near_sad[5] = INT_MAX;
-        if(xd->mb_to_right_edge ==0) near_sad[6] = INT_MAX;
-        if(xd->mb_to_bottom_edge==0) near_sad[7] = INT_MAX;
+    if (xd->mb_to_top_edge == 0) near_sad[4] = INT_MAX;
+    if (xd->mb_to_left_edge == 0) near_sad[5] = INT_MAX;
+    if (xd->mb_to_right_edge == 0) near_sad[6] = INT_MAX;
+    if (xd->mb_to_bottom_edge == 0) near_sad[7] = INT_MAX;
 
-        if(near_sad[4] != INT_MAX)
-            near_sad[4] = cpi->fn_ptr[BLOCK_16X16].sdf(src_y_ptr, b->src_stride, pre_y_buffer - pre_y_stride *16, pre_y_stride);
-        if(near_sad[5] != INT_MAX)
-            near_sad[5] = cpi->fn_ptr[BLOCK_16X16].sdf(src_y_ptr, b->src_stride, pre_y_buffer - 16, pre_y_stride);
-        near_sad[3] = cpi->fn_ptr[BLOCK_16X16].sdf(src_y_ptr, b->src_stride, pre_y_buffer, pre_y_stride);
-        if(near_sad[6] != INT_MAX)
-            near_sad[6] = cpi->fn_ptr[BLOCK_16X16].sdf(src_y_ptr, b->src_stride, pre_y_buffer + 16, pre_y_stride);
-        if(near_sad[7] != INT_MAX)
-            near_sad[7] = cpi->fn_ptr[BLOCK_16X16].sdf(src_y_ptr, b->src_stride, pre_y_buffer + pre_y_stride *16, pre_y_stride);
-    }
+    if (near_sad[4] != INT_MAX)
+      near_sad[4] = cpi->fn_ptr[BLOCK_16X16].sdf(
+          src_y_ptr, b->src_stride, pre_y_buffer - pre_y_stride * 16,
+          pre_y_stride);
+    if (near_sad[5] != INT_MAX)
+      near_sad[5] = cpi->fn_ptr[BLOCK_16X16].sdf(
+          src_y_ptr, b->src_stride, pre_y_buffer - 16, pre_y_stride);
+    near_sad[3] = cpi->fn_ptr[BLOCK_16X16].sdf(src_y_ptr, b->src_stride,
+                                               pre_y_buffer, pre_y_stride);
+    if (near_sad[6] != INT_MAX)
+      near_sad[6] = cpi->fn_ptr[BLOCK_16X16].sdf(
+          src_y_ptr, b->src_stride, pre_y_buffer + 16, pre_y_stride);
+    if (near_sad[7] != INT_MAX)
+      near_sad[7] = cpi->fn_ptr[BLOCK_16X16].sdf(
+          src_y_ptr, b->src_stride, pre_y_buffer + pre_y_stride * 16,
+          pre_y_stride);
+  }
 
-    if(cpi->common.last_frame_type != KEY_FRAME)
-    {
-        insertsortsad(near_sad, near_sadidx, 8);
-    }else
-    {
-        insertsortsad(near_sad, near_sadidx, 3);
-    }
+  if (cpi->common.last_frame_type != KEY_FRAME) {
+    insertsortsad(near_sad, near_sadidx, 8);
+  } else {
+    insertsortsad(near_sad, near_sadidx, 3);
+  }
 }
 
-static void rd_update_mvcount(MACROBLOCK *x, int_mv *best_ref_mv)
-{
-    if (x->e_mbd.mode_info_context->mbmi.mode == SPLITMV)
-    {
-        int i;
+static void rd_update_mvcount(MACROBLOCK *x, int_mv *best_ref_mv) {
+  if (x->e_mbd.mode_info_context->mbmi.mode == SPLITMV) {
+    int i;
 
-        for (i = 0; i < x->partition_info->count; i++)
-        {
-            if (x->partition_info->bmi[i].mode == NEW4X4)
-            {
-                x->MVcount[0][mv_max+((x->partition_info->bmi[i].mv.as_mv.row
-                                          - best_ref_mv->as_mv.row) >> 1)]++;
-                x->MVcount[1][mv_max+((x->partition_info->bmi[i].mv.as_mv.col
-                                          - best_ref_mv->as_mv.col) >> 1)]++;
-            }
-        }
+    for (i = 0; i < x->partition_info->count; i++) {
+      if (x->partition_info->bmi[i].mode == NEW4X4) {
+        x->MVcount[0][mv_max + ((x->partition_info->bmi[i].mv.as_mv.row -
+                                 best_ref_mv->as_mv.row) >>
+                                1)]++;
+        x->MVcount[1][mv_max + ((x->partition_info->bmi[i].mv.as_mv.col -
+                                 best_ref_mv->as_mv.col) >>
+                                1)]++;
+      }
     }
-    else if (x->e_mbd.mode_info_context->mbmi.mode == NEWMV)
-    {
-        x->MVcount[0][mv_max+((x->e_mbd.mode_info_context->mbmi.mv.as_mv.row
-                                          - best_ref_mv->as_mv.row) >> 1)]++;
-        x->MVcount[1][mv_max+((x->e_mbd.mode_info_context->mbmi.mv.as_mv.col
-                                          - best_ref_mv->as_mv.col) >> 1)]++;
-    }
+  } else if (x->e_mbd.mode_info_context->mbmi.mode == NEWMV) {
+    x->MVcount[0][mv_max + ((x->e_mbd.mode_info_context->mbmi.mv.as_mv.row -
+                             best_ref_mv->as_mv.row) >>
+                            1)]++;
+    x->MVcount[1][mv_max + ((x->e_mbd.mode_info_context->mbmi.mv.as_mv.col -
+                             best_ref_mv->as_mv.col) >>
+                            1)]++;
+  }
 }
 
-static int evaluate_inter_mode_rd(int mdcounts[4],
-                                  RATE_DISTORTION* rd,
-                                  int* disable_skip,
-                                  VP8_COMP *cpi, MACROBLOCK *x)
-{
-    MB_PREDICTION_MODE this_mode = x->e_mbd.mode_info_context->mbmi.mode;
-    BLOCK *b = &x->block[0];
-    MACROBLOCKD *xd = &x->e_mbd;
-    int distortion;
-    vp8_build_inter16x16_predictors_mby(&x->e_mbd, x->e_mbd.predictor, 16);
+static int evaluate_inter_mode_rd(int mdcounts[4], RATE_DISTORTION *rd,
+                                  int *disable_skip, VP8_COMP *cpi,
+                                  MACROBLOCK *x) {
+  MB_PREDICTION_MODE this_mode = x->e_mbd.mode_info_context->mbmi.mode;
+  BLOCK *b = &x->block[0];
+  MACROBLOCKD *xd = &x->e_mbd;
+  int distortion;
+  vp8_build_inter16x16_predictors_mby(&x->e_mbd, x->e_mbd.predictor, 16);
 
-    if (cpi->active_map_enabled && x->active_ptr[0] == 0) {
-        x->skip = 1;
-    }
-    else if (x->encode_breakout)
-    {
-        unsigned int sse;
-        unsigned int var;
-        unsigned int threshold = (xd->block[0].dequant[1]
-                    * xd->block[0].dequant[1] >>4);
+  if (cpi->active_map_enabled && x->active_ptr[0] == 0) {
+    x->skip = 1;
+  } else if (x->encode_breakout) {
+    unsigned int sse;
+    unsigned int var;
+    unsigned int threshold =
+        (xd->block[0].dequant[1] * xd->block[0].dequant[1] >> 4);
 
-        if(threshold < x->encode_breakout)
-            threshold = x->encode_breakout;
+    if (threshold < x->encode_breakout) threshold = x->encode_breakout;
 
-        var = vpx_variance16x16
-                (*(b->base_src), b->src_stride,
-                x->e_mbd.predictor, 16, &sse);
+    var = vpx_variance16x16(*(b->base_src), b->src_stride, x->e_mbd.predictor,
+                            16, &sse);
 
-        if (sse < threshold)
-        {
-             unsigned int q2dc = xd->block[24].dequant[0];
-            /* If theres is no codeable 2nd order dc
-               or a very small uniform pixel change change */
-            if ((sse - var < q2dc * q2dc >>4) ||
-                (sse /2 > var && sse-var < 64))
-            {
-                /* Check u and v to make sure skip is ok */
-                unsigned int sse2 = VP8_UVSSE(x);
-                if (sse2 * 2 < threshold)
-                {
-                    x->skip = 1;
-                    rd->distortion2 = sse + sse2;
-                    rd->rate2 = 500;
+    if (sse < threshold) {
+      unsigned int q2dc = xd->block[24].dequant[0];
+      /* If theres is no codeable 2nd order dc
+         or a very small uniform pixel change change */
+      if ((sse - var<q2dc * q2dc>> 4) || (sse / 2 > var && sse - var < 64)) {
+        /* Check u and v to make sure skip is ok */
+        unsigned int sse2 = VP8_UVSSE(x);
+        if (sse2 * 2 < threshold) {
+          x->skip = 1;
+          rd->distortion2 = sse + sse2;
+          rd->rate2 = 500;
 
-                    /* for best_yrd calculation */
-                    rd->rate_uv = 0;
-                    rd->distortion_uv = sse2;
+          /* for best_yrd calculation */
+          rd->rate_uv = 0;
+          rd->distortion_uv = sse2;
 
-                    *disable_skip = 1;
-                    return RDCOST(x->rdmult, x->rddiv, rd->rate2,
-                                  rd->distortion2);
-                }
-            }
+          *disable_skip = 1;
+          return RDCOST(x->rdmult, x->rddiv, rd->rate2, rd->distortion2);
         }
+      }
     }
+  }
 
+  /* Add in the Mv/mode cost */
+  rd->rate2 += vp8_cost_mv_ref(this_mode, mdcounts);
 
-    /* Add in the Mv/mode cost */
-    rd->rate2 += vp8_cost_mv_ref(this_mode, mdcounts);
+  /* Y cost and distortion */
+  macro_block_yrd(x, &rd->rate_y, &distortion);
+  rd->rate2 += rd->rate_y;
+  rd->distortion2 += distortion;
 
-    /* Y cost and distortion */
-    macro_block_yrd(x, &rd->rate_y, &distortion);
-    rd->rate2 += rd->rate_y;
-    rd->distortion2 += distortion;
-
-    /* UV cost and distortion */
-    rd_inter16x16_uv(cpi, x, &rd->rate_uv, &rd->distortion_uv,
-                     cpi->common.full_pixel);
-    rd->rate2 += rd->rate_uv;
-    rd->distortion2 += rd->distortion_uv;
-    return INT_MAX;
+  /* UV cost and distortion */
+  rd_inter16x16_uv(cpi, x, &rd->rate_uv, &rd->distortion_uv,
+                   cpi->common.full_pixel);
+  rd->rate2 += rd->rate_uv;
+  rd->distortion2 += rd->distortion_uv;
+  return INT_MAX;
 }
 
-static int calculate_final_rd_costs(int this_rd,
-                                    RATE_DISTORTION* rd,
-                                    int* other_cost,
-                                    int disable_skip,
-                                    int uv_intra_tteob,
-                                    int intra_rd_penalty,
-                                    VP8_COMP *cpi, MACROBLOCK *x)
-{
-    MB_PREDICTION_MODE this_mode = x->e_mbd.mode_info_context->mbmi.mode;
+static int calculate_final_rd_costs(int this_rd, RATE_DISTORTION *rd,
+                                    int *other_cost, int disable_skip,
+                                    int uv_intra_tteob, int intra_rd_penalty,
+                                    VP8_COMP *cpi, MACROBLOCK *x) {
+  MB_PREDICTION_MODE this_mode = x->e_mbd.mode_info_context->mbmi.mode;
 
-    /* Where skip is allowable add in the default per mb cost for the no
-     * skip case. where we then decide to skip we have to delete this and
-     * replace it with the cost of signalling a skip
-     */
-    if (cpi->common.mb_no_coeff_skip)
-    {
-        *other_cost += vp8_cost_bit(cpi->prob_skip_false, 0);
-        rd->rate2 += *other_cost;
-    }
+  /* Where skip is allowable add in the default per mb cost for the no
+   * skip case. where we then decide to skip we have to delete this and
+   * replace it with the cost of signalling a skip
+   */
+  if (cpi->common.mb_no_coeff_skip) {
+    *other_cost += vp8_cost_bit(cpi->prob_skip_false, 0);
+    rd->rate2 += *other_cost;
+  }
 
-    /* Estimate the reference frame signaling cost and add it
-     * to the rolling cost variable.
+  /* Estimate the reference frame signaling cost and add it
+   * to the rolling cost variable.
+   */
+  rd->rate2 += x->ref_frame_cost[x->e_mbd.mode_info_context->mbmi.ref_frame];
+
+  if (!disable_skip) {
+    /* Test for the condition where skip block will be activated
+     * because there are no non zero coefficients and make any
+     * necessary adjustment for rate
      */
-    rd->rate2 +=
-        x->ref_frame_cost[x->e_mbd.mode_info_context->mbmi.ref_frame];
+    if (cpi->common.mb_no_coeff_skip) {
+      int i;
+      int tteob;
+      int has_y2_block = (this_mode != SPLITMV && this_mode != B_PRED);
 
-    if (!disable_skip)
-    {
-        /* Test for the condition where skip block will be activated
-         * because there are no non zero coefficients and make any
-         * necessary adjustment for rate
-         */
-        if (cpi->common.mb_no_coeff_skip)
-        {
-            int i;
-            int tteob;
-            int has_y2_block = (this_mode!=SPLITMV && this_mode!=B_PRED);
+      tteob = 0;
+      if (has_y2_block) tteob += x->e_mbd.eobs[24];
 
-            tteob = 0;
-            if(has_y2_block)
-                tteob += x->e_mbd.eobs[24];
+      for (i = 0; i < 16; i++) tteob += (x->e_mbd.eobs[i] > has_y2_block);
 
-            for (i = 0; i < 16; i++)
-                tteob += (x->e_mbd.eobs[i] > has_y2_block);
+      if (x->e_mbd.mode_info_context->mbmi.ref_frame) {
+        for (i = 16; i < 24; i++) tteob += x->e_mbd.eobs[i];
+      } else
+        tteob += uv_intra_tteob;
 
-            if (x->e_mbd.mode_info_context->mbmi.ref_frame)
-            {
-                for (i = 16; i < 24; i++)
-                    tteob += x->e_mbd.eobs[i];
-            }
-            else
-                tteob += uv_intra_tteob;
+      if (tteob == 0) {
+        rd->rate2 -= (rd->rate_y + rd->rate_uv);
+        /* for best_yrd calculation */
+        rd->rate_uv = 0;
 
-            if (tteob == 0)
-            {
-                rd->rate2 -= (rd->rate_y + rd->rate_uv);
-                /* for best_yrd calculation */
-                rd->rate_uv = 0;
+        /* Back out no skip flag costing and add in skip flag costing */
+        if (cpi->prob_skip_false) {
+          int prob_skip_cost;
 
-                /* Back out no skip flag costing and add in skip flag costing */
-                if (cpi->prob_skip_false)
-                {
-                    int prob_skip_cost;
-
-                    prob_skip_cost = vp8_cost_bit(cpi->prob_skip_false, 1);
-                    prob_skip_cost -=
-                        (int)vp8_cost_bit(cpi->prob_skip_false, 0);
-                    rd->rate2 += prob_skip_cost;
-                    *other_cost += prob_skip_cost;
-                }
-            }
+          prob_skip_cost = vp8_cost_bit(cpi->prob_skip_false, 1);
+          prob_skip_cost -= (int)vp8_cost_bit(cpi->prob_skip_false, 0);
+          rd->rate2 += prob_skip_cost;
+          *other_cost += prob_skip_cost;
         }
-        /* Calculate the final RD estimate for this mode */
-        this_rd = RDCOST(x->rdmult, x->rddiv, rd->rate2, rd->distortion2);
-        if (this_rd < INT_MAX && x->e_mbd.mode_info_context->mbmi.ref_frame
-                                 == INTRA_FRAME)
-            this_rd += intra_rd_penalty;
+      }
     }
-    return this_rd;
+    /* Calculate the final RD estimate for this mode */
+    this_rd = RDCOST(x->rdmult, x->rddiv, rd->rate2, rd->distortion2);
+    if (this_rd < INT_MAX &&
+        x->e_mbd.mode_info_context->mbmi.ref_frame == INTRA_FRAME)
+      this_rd += intra_rd_penalty;
+  }
+  return this_rd;
 }
 
-static void update_best_mode(BEST_MODE* best_mode, int this_rd,
-                             RATE_DISTORTION* rd, int other_cost, MACROBLOCK *x)
-{
-    MB_PREDICTION_MODE this_mode = x->e_mbd.mode_info_context->mbmi.mode;
+static void update_best_mode(BEST_MODE *best_mode, int this_rd,
+                             RATE_DISTORTION *rd, int other_cost,
+                             MACROBLOCK *x) {
+  MB_PREDICTION_MODE this_mode = x->e_mbd.mode_info_context->mbmi.mode;
 
-    other_cost +=
-    x->ref_frame_cost[x->e_mbd.mode_info_context->mbmi.ref_frame];
+  other_cost += x->ref_frame_cost[x->e_mbd.mode_info_context->mbmi.ref_frame];
 
-    /* Calculate the final y RD estimate for this mode */
-    best_mode->yrd = RDCOST(x->rdmult, x->rddiv, (rd->rate2-rd->rate_uv-other_cost),
-                      (rd->distortion2-rd->distortion_uv));
+  /* Calculate the final y RD estimate for this mode */
+  best_mode->yrd =
+      RDCOST(x->rdmult, x->rddiv, (rd->rate2 - rd->rate_uv - other_cost),
+             (rd->distortion2 - rd->distortion_uv));
 
-    best_mode->rd = this_rd;
-    memcpy(&best_mode->mbmode, &x->e_mbd.mode_info_context->mbmi, sizeof(MB_MODE_INFO));
-    memcpy(&best_mode->partition, x->partition_info, sizeof(PARTITION_INFO));
+  best_mode->rd = this_rd;
+  memcpy(&best_mode->mbmode, &x->e_mbd.mode_info_context->mbmi,
+         sizeof(MB_MODE_INFO));
+  memcpy(&best_mode->partition, x->partition_info, sizeof(PARTITION_INFO));
 
-    if ((this_mode == B_PRED) || (this_mode == SPLITMV))
-    {
-        int i;
-        for (i = 0; i < 16; i++)
-        {
-            best_mode->bmodes[i] = x->e_mbd.block[i].bmi;
-        }
+  if ((this_mode == B_PRED) || (this_mode == SPLITMV)) {
+    int i;
+    for (i = 0; i < 16; i++) {
+      best_mode->bmodes[i] = x->e_mbd.block[i].bmi;
     }
+  }
 }
 
 void vp8_rd_pick_inter_mode(VP8_COMP *cpi, MACROBLOCK *x, int recon_yoffset,
                             int recon_uvoffset, int *returnrate,
-                            int *returndistortion, int *returnintra,
-                            int mb_row, int mb_col)
-{
-    BLOCK *b = &x->block[0];
-    BLOCKD *d = &x->e_mbd.block[0];
-    MACROBLOCKD *xd = &x->e_mbd;
-    int_mv best_ref_mv_sb[2];
-    int_mv mode_mv_sb[2][MB_MODE_COUNT];
-    int_mv best_ref_mv;
-    int_mv *mode_mv;
-    MB_PREDICTION_MODE this_mode;
-    int num00;
-    int best_mode_index = 0;
-    BEST_MODE best_mode;
+                            int *returndistortion, int *returnintra, int mb_row,
+                            int mb_col) {
+  BLOCK *b = &x->block[0];
+  BLOCKD *d = &x->e_mbd.block[0];
+  MACROBLOCKD *xd = &x->e_mbd;
+  int_mv best_ref_mv_sb[2];
+  int_mv mode_mv_sb[2][MB_MODE_COUNT];
+  int_mv best_ref_mv;
+  int_mv *mode_mv;
+  MB_PREDICTION_MODE this_mode;
+  int num00;
+  int best_mode_index = 0;
+  BEST_MODE best_mode;
 
-    int i;
-    int mode_index;
-    int mdcounts[4];
-    int rate;
-    RATE_DISTORTION rd;
-    int uv_intra_rate, uv_intra_distortion, uv_intra_rate_tokenonly;
-    int uv_intra_tteob = 0;
-    int uv_intra_done = 0;
+  int i;
+  int mode_index;
+  int mdcounts[4];
+  int rate;
+  RATE_DISTORTION rd;
+  int uv_intra_rate, uv_intra_distortion, uv_intra_rate_tokenonly;
+  int uv_intra_tteob = 0;
+  int uv_intra_done = 0;
 
-    MB_PREDICTION_MODE uv_intra_mode = 0;
-    int_mv mvp;
-    int near_sadidx[8] = {0, 1, 2, 3, 4, 5, 6, 7};
-    int saddone=0;
-    /* search range got from mv_pred(). It uses step_param levels. (0-7) */
-    int sr=0;
+  MB_PREDICTION_MODE uv_intra_mode = 0;
+  int_mv mvp;
+  int near_sadidx[8] = { 0, 1, 2, 3, 4, 5, 6, 7 };
+  int saddone = 0;
+  /* search range got from mv_pred(). It uses step_param levels. (0-7) */
+  int sr = 0;
 
-    unsigned char *plane[4][3];
-    int ref_frame_map[4];
-    int sign_bias = 0;
+  unsigned char *plane[4][3];
+  int ref_frame_map[4];
+  int sign_bias = 0;
 
-    int intra_rd_penalty =  10* vp8_dc_quant(cpi->common.base_qindex,
-                                             cpi->common.y1dc_delta_q);
+  int intra_rd_penalty =
+      10 * vp8_dc_quant(cpi->common.base_qindex, cpi->common.y1dc_delta_q);
 
 #if CONFIG_TEMPORAL_DENOISING
-    unsigned int zero_mv_sse = UINT_MAX, best_sse = UINT_MAX,
-            best_rd_sse = UINT_MAX;
+  unsigned int zero_mv_sse = UINT_MAX, best_sse = UINT_MAX,
+               best_rd_sse = UINT_MAX;
 #endif
 
+  mode_mv = mode_mv_sb[sign_bias];
+  best_ref_mv.as_int = 0;
+  best_mode.rd = INT_MAX;
+  best_mode.yrd = INT_MAX;
+  best_mode.intra_rd = INT_MAX;
+  memset(mode_mv_sb, 0, sizeof(mode_mv_sb));
+  memset(&best_mode.mbmode, 0, sizeof(best_mode.mbmode));
+  memset(&best_mode.bmodes, 0, sizeof(best_mode.bmodes));
+
+  /* Setup search priorities */
+  get_reference_search_order(cpi, ref_frame_map);
+
+  /* Check to see if there is at least 1 valid reference frame that we need
+   * to calculate near_mvs.
+   */
+  if (ref_frame_map[1] > 0) {
+    sign_bias = vp8_find_near_mvs_bias(
+        &x->e_mbd, x->e_mbd.mode_info_context, mode_mv_sb, best_ref_mv_sb,
+        mdcounts, ref_frame_map[1], cpi->common.ref_frame_sign_bias);
+
     mode_mv = mode_mv_sb[sign_bias];
-    best_ref_mv.as_int = 0;
-    best_mode.rd = INT_MAX;
-    best_mode.yrd = INT_MAX;
-    best_mode.intra_rd = INT_MAX;
-    memset(mode_mv_sb, 0, sizeof(mode_mv_sb));
-    memset(&best_mode.mbmode, 0, sizeof(best_mode.mbmode));
-    memset(&best_mode.bmodes, 0, sizeof(best_mode.bmodes));
+    best_ref_mv.as_int = best_ref_mv_sb[sign_bias].as_int;
+  }
 
-    /* Setup search priorities */
-    get_reference_search_order(cpi, ref_frame_map);
+  get_predictor_pointers(cpi, plane, recon_yoffset, recon_uvoffset);
 
-    /* Check to see if there is at least 1 valid reference frame that we need
-     * to calculate near_mvs.
-     */
-    if (ref_frame_map[1] > 0)
-    {
-        sign_bias = vp8_find_near_mvs_bias(&x->e_mbd,
-                                           x->e_mbd.mode_info_context,
-                                           mode_mv_sb,
-                                           best_ref_mv_sb,
-                                           mdcounts,
-                                           ref_frame_map[1],
-                                           cpi->common.ref_frame_sign_bias);
+  *returnintra = INT_MAX;
+  /* Count of the number of MBs tested so far this frame */
+  x->mbs_tested_so_far++;
 
-        mode_mv = mode_mv_sb[sign_bias];
-        best_ref_mv.as_int = best_ref_mv_sb[sign_bias].as_int;
-    }
+  x->skip = 0;
 
-    get_predictor_pointers(cpi, plane, recon_yoffset, recon_uvoffset);
+  for (mode_index = 0; mode_index < MAX_MODES; mode_index++) {
+    int this_rd = INT_MAX;
+    int disable_skip = 0;
+    int other_cost = 0;
+    int this_ref_frame = ref_frame_map[vp8_ref_frame_order[mode_index]];
 
-    *returnintra = INT_MAX;
-    /* Count of the number of MBs tested so far this frame */
-    x->mbs_tested_so_far++;
+    /* Test best rd so far against threshold for trying this mode. */
+    if (best_mode.rd <= x->rd_threshes[mode_index]) continue;
 
-    x->skip = 0;
+    if (this_ref_frame < 0) continue;
 
-    for (mode_index = 0; mode_index < MAX_MODES; mode_index++)
-    {
-        int this_rd = INT_MAX;
-        int disable_skip = 0;
-        int other_cost = 0;
-        int this_ref_frame = ref_frame_map[vp8_ref_frame_order[mode_index]];
+    /* These variables hold are rolling total cost and distortion for
+     * this mode
+     */
+    rd.rate2 = 0;
+    rd.distortion2 = 0;
 
-        /* Test best rd so far against threshold for trying this mode. */
-        if (best_mode.rd <= x->rd_threshes[mode_index])
-            continue;
+    this_mode = vp8_mode_order[mode_index];
 
-        if (this_ref_frame < 0)
-            continue;
+    x->e_mbd.mode_info_context->mbmi.mode = this_mode;
+    x->e_mbd.mode_info_context->mbmi.ref_frame = this_ref_frame;
 
-        /* These variables hold are rolling total cost and distortion for
-         * this mode
-         */
-        rd.rate2 = 0;
-        rd.distortion2 = 0;
+    /* Only consider ZEROMV/ALTREF_FRAME for alt ref frame,
+     * unless ARNR filtering is enabled in which case we want
+     * an unfiltered alternative
+     */
+    if (cpi->is_src_frame_alt_ref && (cpi->oxcf.arnr_max_frames == 0)) {
+      if (this_mode != ZEROMV ||
+          x->e_mbd.mode_info_context->mbmi.ref_frame != ALTREF_FRAME)
+        continue;
+    }
 
-        this_mode = vp8_mode_order[mode_index];
+    /* everything but intra */
+    if (x->e_mbd.mode_info_context->mbmi.ref_frame) {
+      x->e_mbd.pre.y_buffer = plane[this_ref_frame][0];
+      x->e_mbd.pre.u_buffer = plane[this_ref_frame][1];
+      x->e_mbd.pre.v_buffer = plane[this_ref_frame][2];
 
-        x->e_mbd.mode_info_context->mbmi.mode = this_mode;
-        x->e_mbd.mode_info_context->mbmi.ref_frame = this_ref_frame;
+      if (sign_bias != cpi->common.ref_frame_sign_bias[this_ref_frame]) {
+        sign_bias = cpi->common.ref_frame_sign_bias[this_ref_frame];
+        mode_mv = mode_mv_sb[sign_bias];
+        best_ref_mv.as_int = best_ref_mv_sb[sign_bias].as_int;
+      }
+    }
 
-        /* Only consider ZEROMV/ALTREF_FRAME for alt ref frame,
-         * unless ARNR filtering is enabled in which case we want
-         * an unfiltered alternative
+    /* Check to see if the testing frequency for this mode is at its
+     * max If so then prevent it from being tested and increase the
+     * threshold for its testing
+     */
+    if (x->mode_test_hit_counts[mode_index] &&
+        (cpi->mode_check_freq[mode_index] > 1)) {
+      if (x->mbs_tested_so_far <= cpi->mode_check_freq[mode_index] *
+                                      x->mode_test_hit_counts[mode_index]) {
+        /* Increase the threshold for coding this mode to make it
+         * less likely to be chosen
          */
-        if (cpi->is_src_frame_alt_ref && (cpi->oxcf.arnr_max_frames == 0))
-        {
-            if (this_mode != ZEROMV || x->e_mbd.mode_info_context->mbmi.ref_frame != ALTREF_FRAME)
-                continue;
-        }
+        x->rd_thresh_mult[mode_index] += 4;
 
-        /* everything but intra */
-        if (x->e_mbd.mode_info_context->mbmi.ref_frame)
-        {
-            x->e_mbd.pre.y_buffer = plane[this_ref_frame][0];
-            x->e_mbd.pre.u_buffer = plane[this_ref_frame][1];
-            x->e_mbd.pre.v_buffer = plane[this_ref_frame][2];
+        if (x->rd_thresh_mult[mode_index] > MAX_THRESHMULT)
+          x->rd_thresh_mult[mode_index] = MAX_THRESHMULT;
 
-            if (sign_bias != cpi->common.ref_frame_sign_bias[this_ref_frame])
-            {
-                sign_bias = cpi->common.ref_frame_sign_bias[this_ref_frame];
-                mode_mv = mode_mv_sb[sign_bias];
-                best_ref_mv.as_int = best_ref_mv_sb[sign_bias].as_int;
-            }
-        }
+        x->rd_threshes[mode_index] =
+            (cpi->rd_baseline_thresh[mode_index] >> 7) *
+            x->rd_thresh_mult[mode_index];
 
-        /* Check to see if the testing frequency for this mode is at its
-         * max If so then prevent it from being tested and increase the
-         * threshold for its testing
-         */
-        if (x->mode_test_hit_counts[mode_index] && (cpi->mode_check_freq[mode_index] > 1))
-        {
-            if (x->mbs_tested_so_far  <= cpi->mode_check_freq[mode_index] * x->mode_test_hit_counts[mode_index])
-            {
-                /* Increase the threshold for coding this mode to make it
-                 * less likely to be chosen
-                 */
-                x->rd_thresh_mult[mode_index] += 4;
+        continue;
+      }
+    }
 
-                if (x->rd_thresh_mult[mode_index] > MAX_THRESHMULT)
-                    x->rd_thresh_mult[mode_index] = MAX_THRESHMULT;
+    /* We have now reached the point where we are going to test the
+     * current mode so increment the counter for the number of times
+     * it has been tested
+     */
+    x->mode_test_hit_counts[mode_index]++;
 
-                x->rd_threshes[mode_index] =
-                    (cpi->rd_baseline_thresh[mode_index] >> 7) *
-                    x->rd_thresh_mult[mode_index];
+    /* Experimental code. Special case for gf and arf zeromv modes.
+     * Increase zbin size to supress noise
+     */
+    if (x->zbin_mode_boost_enabled) {
+      if (this_ref_frame == INTRA_FRAME)
+        x->zbin_mode_boost = 0;
+      else {
+        if (vp8_mode_order[mode_index] == ZEROMV) {
+          if (this_ref_frame != LAST_FRAME)
+            x->zbin_mode_boost = GF_ZEROMV_ZBIN_BOOST;
+          else
+            x->zbin_mode_boost = LF_ZEROMV_ZBIN_BOOST;
+        } else if (vp8_mode_order[mode_index] == SPLITMV)
+          x->zbin_mode_boost = 0;
+        else
+          x->zbin_mode_boost = MV_ZBIN_BOOST;
+      }
 
-                continue;
-            }
-        }
+      vp8_update_zbin_extra(cpi, x);
+    }
 
-        /* We have now reached the point where we are going to test the
-         * current mode so increment the counter for the number of times
-         * it has been tested
-         */
-        x->mode_test_hit_counts[mode_index] ++;
+    if (!uv_intra_done && this_ref_frame == INTRA_FRAME) {
+      rd_pick_intra_mbuv_mode(x, &uv_intra_rate, &uv_intra_rate_tokenonly,
+                              &uv_intra_distortion);
+      uv_intra_mode = x->e_mbd.mode_info_context->mbmi.uv_mode;
 
-        /* Experimental code. Special case for gf and arf zeromv modes.
-         * Increase zbin size to supress noise
-         */
-        if (x->zbin_mode_boost_enabled)
-        {
-            if ( this_ref_frame == INTRA_FRAME )
-                x->zbin_mode_boost = 0;
-            else
-            {
-                if (vp8_mode_order[mode_index] == ZEROMV)
-                {
-                    if (this_ref_frame != LAST_FRAME)
-                        x->zbin_mode_boost = GF_ZEROMV_ZBIN_BOOST;
-                    else
-                        x->zbin_mode_boost = LF_ZEROMV_ZBIN_BOOST;
-                }
-                else if (vp8_mode_order[mode_index] == SPLITMV)
-                    x->zbin_mode_boost = 0;
-                else
-                    x->zbin_mode_boost = MV_ZBIN_BOOST;
-            }
+      /*
+       * Total of the eobs is used later to further adjust rate2. Since uv
+       * block's intra eobs will be overwritten when we check inter modes,
+       * we need to save uv_intra_tteob here.
+       */
+      for (i = 16; i < 24; i++) uv_intra_tteob += x->e_mbd.eobs[i];
 
-            vp8_update_zbin_extra(cpi, x);
-        }
+      uv_intra_done = 1;
+    }
 
-        if(!uv_intra_done && this_ref_frame == INTRA_FRAME)
-        {
-            rd_pick_intra_mbuv_mode(x, &uv_intra_rate,
-                                    &uv_intra_rate_tokenonly,
-                                    &uv_intra_distortion);
-            uv_intra_mode = x->e_mbd.mode_info_context->mbmi.uv_mode;
+    switch (this_mode) {
+      case B_PRED: {
+        int tmp_rd;
 
-            /*
-             * Total of the eobs is used later to further adjust rate2. Since uv
-             * block's intra eobs will be overwritten when we check inter modes,
-             * we need to save uv_intra_tteob here.
-             */
-            for (i = 16; i < 24; i++)
-                uv_intra_tteob += x->e_mbd.eobs[i];
+        /* Note the rate value returned here includes the cost of
+         * coding the BPRED mode: x->mbmode_cost[x->e_mbd.frame_type][BPRED]
+         */
+        int distortion;
+        tmp_rd = rd_pick_intra4x4mby_modes(x, &rate, &rd.rate_y, &distortion,
+                                           best_mode.yrd);
+        rd.rate2 += rate;
+        rd.distortion2 += distortion;
 
-            uv_intra_done = 1;
+        if (tmp_rd < best_mode.yrd) {
+          rd.rate2 += uv_intra_rate;
+          rd.rate_uv = uv_intra_rate_tokenonly;
+          rd.distortion2 += uv_intra_distortion;
+          rd.distortion_uv = uv_intra_distortion;
+        } else {
+          this_rd = INT_MAX;
+          disable_skip = 1;
         }
+      } break;
 
-        switch (this_mode)
-        {
-        case B_PRED:
-        {
-            int tmp_rd;
+      case SPLITMV: {
+        int tmp_rd;
+        int this_rd_thresh;
+        int distortion;
 
-            /* Note the rate value returned here includes the cost of
-             * coding the BPRED mode: x->mbmode_cost[x->e_mbd.frame_type][BPRED]
-             */
-            int distortion;
-            tmp_rd = rd_pick_intra4x4mby_modes(x, &rate, &rd.rate_y, &distortion, best_mode.yrd);
-            rd.rate2 += rate;
-            rd.distortion2 += distortion;
+        this_rd_thresh = (vp8_ref_frame_order[mode_index] == 1)
+                             ? x->rd_threshes[THR_NEW1]
+                             : x->rd_threshes[THR_NEW3];
+        this_rd_thresh = (vp8_ref_frame_order[mode_index] == 2)
+                             ? x->rd_threshes[THR_NEW2]
+                             : this_rd_thresh;
 
-            if(tmp_rd < best_mode.yrd)
-            {
-                rd.rate2 += uv_intra_rate;
-                rd.rate_uv = uv_intra_rate_tokenonly;
-                rd.distortion2 += uv_intra_distortion;
-                rd.distortion_uv = uv_intra_distortion;
-            }
-            else
-            {
-                this_rd = INT_MAX;
-                disable_skip = 1;
-            }
+        tmp_rd = vp8_rd_pick_best_mbsegmentation(
+            cpi, x, &best_ref_mv, best_mode.yrd, mdcounts, &rate, &rd.rate_y,
+            &distortion, this_rd_thresh);
+
+        rd.rate2 += rate;
+        rd.distortion2 += distortion;
+
+        /* If even the 'Y' rd value of split is higher than best so far
+         * then dont bother looking at UV
+         */
+        if (tmp_rd < best_mode.yrd) {
+          /* Now work out UV cost and add it in */
+          rd_inter4x4_uv(cpi, x, &rd.rate_uv, &rd.distortion_uv,
+                         cpi->common.full_pixel);
+          rd.rate2 += rd.rate_uv;
+          rd.distortion2 += rd.distortion_uv;
+        } else {
+          this_rd = INT_MAX;
+          disable_skip = 1;
         }
-        break;
+      } break;
+      case DC_PRED:
+      case V_PRED:
+      case H_PRED:
+      case TM_PRED: {
+        int distortion;
+        x->e_mbd.mode_info_context->mbmi.ref_frame = INTRA_FRAME;
 
-        case SPLITMV:
-        {
-            int tmp_rd;
-            int this_rd_thresh;
-            int distortion;
+        vp8_build_intra_predictors_mby_s(
+            xd, xd->dst.y_buffer - xd->dst.y_stride, xd->dst.y_buffer - 1,
+            xd->dst.y_stride, xd->predictor, 16);
+        macro_block_yrd(x, &rd.rate_y, &distortion);
+        rd.rate2 += rd.rate_y;
+        rd.distortion2 += distortion;
+        rd.rate2 +=
+            x->mbmode_cost[x->e_mbd.frame_type][x->e_mbd.mode_info_context->mbmi
+                                                    .mode];
+        rd.rate2 += uv_intra_rate;
+        rd.rate_uv = uv_intra_rate_tokenonly;
+        rd.distortion2 += uv_intra_distortion;
+        rd.distortion_uv = uv_intra_distortion;
+      } break;
 
-            this_rd_thresh = (vp8_ref_frame_order[mode_index] == 1) ?
-                x->rd_threshes[THR_NEW1] : x->rd_threshes[THR_NEW3];
-            this_rd_thresh = (vp8_ref_frame_order[mode_index] == 2) ?
-                x->rd_threshes[THR_NEW2] : this_rd_thresh;
+      case NEWMV: {
+        int thissme;
+        int bestsme = INT_MAX;
+        int step_param = cpi->sf.first_step;
+        int further_steps;
+        int n;
+        /* If last step (1-away) of n-step search doesn't pick the center point
+           as the best match, we will do a final 1-away diamond refining search
+        */
+        int do_refine = 1;
 
-            tmp_rd = vp8_rd_pick_best_mbsegmentation(cpi, x, &best_ref_mv,
-                                                     best_mode.yrd, mdcounts,
-                                                     &rate, &rd.rate_y, &distortion, this_rd_thresh) ;
+        int sadpb = x->sadperbit16;
+        int_mv mvp_full;
 
-            rd.rate2 += rate;
-            rd.distortion2 += distortion;
+        int col_min = ((best_ref_mv.as_mv.col + 7) >> 3) - MAX_FULL_PEL_VAL;
+        int row_min = ((best_ref_mv.as_mv.row + 7) >> 3) - MAX_FULL_PEL_VAL;
+        int col_max = (best_ref_mv.as_mv.col >> 3) + MAX_FULL_PEL_VAL;
+        int row_max = (best_ref_mv.as_mv.row >> 3) + MAX_FULL_PEL_VAL;
 
-            /* If even the 'Y' rd value of split is higher than best so far
-             * then dont bother looking at UV
-             */
-            if (tmp_rd < best_mode.yrd)
-            {
-                /* Now work out UV cost and add it in */
-                rd_inter4x4_uv(cpi, x, &rd.rate_uv, &rd.distortion_uv, cpi->common.full_pixel);
-                rd.rate2 += rd.rate_uv;
-                rd.distortion2 += rd.distortion_uv;
-            }
-            else
-            {
-                this_rd = INT_MAX;
-                disable_skip = 1;
-            }
-        }
-        break;
-        case DC_PRED:
-        case V_PRED:
-        case H_PRED:
-        case TM_PRED:
-        {
-            int distortion;
-            x->e_mbd.mode_info_context->mbmi.ref_frame = INTRA_FRAME;
+        int tmp_col_min = x->mv_col_min;
+        int tmp_col_max = x->mv_col_max;
+        int tmp_row_min = x->mv_row_min;
+        int tmp_row_max = x->mv_row_max;
 
-            vp8_build_intra_predictors_mby_s(xd,
-                                             xd->dst.y_buffer - xd->dst.y_stride,
-                                             xd->dst.y_buffer - 1,
-                                             xd->dst.y_stride,
-                                             xd->predictor,
-                                             16);
-            macro_block_yrd(x, &rd.rate_y, &distortion) ;
-            rd.rate2 += rd.rate_y;
-            rd.distortion2 += distortion;
-            rd.rate2 += x->mbmode_cost[x->e_mbd.frame_type][x->e_mbd.mode_info_context->mbmi.mode];
-            rd.rate2 += uv_intra_rate;
-            rd.rate_uv = uv_intra_rate_tokenonly;
-            rd.distortion2 += uv_intra_distortion;
-            rd.distortion_uv = uv_intra_distortion;
+        if (!saddone) {
+          vp8_cal_sad(cpi, xd, x, recon_yoffset, &near_sadidx[0]);
+          saddone = 1;
         }
-        break;
 
-        case NEWMV:
-        {
-            int thissme;
-            int bestsme = INT_MAX;
-            int step_param = cpi->sf.first_step;
-            int further_steps;
-            int n;
-            int do_refine=1;   /* If last step (1-away) of n-step search doesn't pick the center point as the best match,
-                                  we will do a final 1-away diamond refining search  */
+        vp8_mv_pred(cpi, &x->e_mbd, x->e_mbd.mode_info_context, &mvp,
+                    x->e_mbd.mode_info_context->mbmi.ref_frame,
+                    cpi->common.ref_frame_sign_bias, &sr, &near_sadidx[0]);
 
-            int sadpb = x->sadperbit16;
-            int_mv mvp_full;
+        mvp_full.as_mv.col = mvp.as_mv.col >> 3;
+        mvp_full.as_mv.row = mvp.as_mv.row >> 3;
 
-            int col_min = ((best_ref_mv.as_mv.col+7)>>3) - MAX_FULL_PEL_VAL;
-            int row_min = ((best_ref_mv.as_mv.row+7)>>3) - MAX_FULL_PEL_VAL;
-            int col_max = (best_ref_mv.as_mv.col>>3) + MAX_FULL_PEL_VAL;
-            int row_max = (best_ref_mv.as_mv.row>>3) + MAX_FULL_PEL_VAL;
+        /* Get intersection of UMV window and valid MV window to
+         * reduce # of checks in diamond search.
+         */
+        if (x->mv_col_min < col_min) x->mv_col_min = col_min;
+        if (x->mv_col_max > col_max) x->mv_col_max = col_max;
+        if (x->mv_row_min < row_min) x->mv_row_min = row_min;
+        if (x->mv_row_max > row_max) x->mv_row_max = row_max;
 
-            int tmp_col_min = x->mv_col_min;
-            int tmp_col_max = x->mv_col_max;
-            int tmp_row_min = x->mv_row_min;
-            int tmp_row_max = x->mv_row_max;
+        /* adjust search range according to sr from mv prediction */
+        if (sr > step_param) step_param = sr;
 
-            if(!saddone)
-            {
-                vp8_cal_sad(cpi,xd,x, recon_yoffset ,&near_sadidx[0] );
-                saddone = 1;
-            }
+        /* Initial step/diamond search */
+        {
+          bestsme = cpi->diamond_search_sad(
+              x, b, d, &mvp_full, &d->bmi.mv, step_param, sadpb, &num00,
+              &cpi->fn_ptr[BLOCK_16X16], x->mvcost, &best_ref_mv);
+          mode_mv[NEWMV].as_int = d->bmi.mv.as_int;
 
-            vp8_mv_pred(cpi, &x->e_mbd, x->e_mbd.mode_info_context, &mvp,
-                        x->e_mbd.mode_info_context->mbmi.ref_frame, cpi->common.ref_frame_sign_bias, &sr, &near_sadidx[0]);
+          /* Further step/diamond searches as necessary */
+          further_steps = (cpi->sf.max_step_search_steps - 1) - step_param;
 
-            mvp_full.as_mv.col = mvp.as_mv.col>>3;
-            mvp_full.as_mv.row = mvp.as_mv.row>>3;
+          n = num00;
+          num00 = 0;
 
-            /* Get intersection of UMV window and valid MV window to
-             * reduce # of checks in diamond search.
-             */
-            if (x->mv_col_min < col_min )
-                x->mv_col_min = col_min;
-            if (x->mv_col_max > col_max )
-                x->mv_col_max = col_max;
-            if (x->mv_row_min < row_min )
-                x->mv_row_min = row_min;
-            if (x->mv_row_max > row_max )
-                x->mv_row_max = row_max;
+          /* If there won't be more n-step search, check to see if refining
+           * search is needed. */
+          if (n > further_steps) do_refine = 0;
 
-            /* adjust search range according to sr from mv prediction */
-            if(sr > step_param)
-                step_param = sr;
+          while (n < further_steps) {
+            n++;
 
-            /* Initial step/diamond search */
-            {
-                bestsme = cpi->diamond_search_sad(x, b, d, &mvp_full, &d->bmi.mv,
-                                        step_param, sadpb, &num00,
-                                        &cpi->fn_ptr[BLOCK_16X16],
-                                        x->mvcost, &best_ref_mv);
-                mode_mv[NEWMV].as_int = d->bmi.mv.as_int;
+            if (num00)
+              num00--;
+            else {
+              thissme = cpi->diamond_search_sad(
+                  x, b, d, &mvp_full, &d->bmi.mv, step_param + n, sadpb, &num00,
+                  &cpi->fn_ptr[BLOCK_16X16], x->mvcost, &best_ref_mv);
 
-                /* Further step/diamond searches as necessary */
-                further_steps = (cpi->sf.max_step_search_steps - 1) - step_param;
+              /* check to see if refining search is needed. */
+              if (num00 > (further_steps - n)) do_refine = 0;
 
-                n = num00;
-                num00 = 0;
+              if (thissme < bestsme) {
+                bestsme = thissme;
+                mode_mv[NEWMV].as_int = d->bmi.mv.as_int;
+              } else {
+                d->bmi.mv.as_int = mode_mv[NEWMV].as_int;
+              }
+            }
+          }
+        }
 
-                /* If there won't be more n-step search, check to see if refining search is needed. */
-                if (n > further_steps)
-                    do_refine = 0;
+        /* final 1-away diamond refining search */
+        if (do_refine == 1) {
+          int search_range;
 
-                while (n < further_steps)
-                {
-                    n++;
+          search_range = 8;
 
-                    if (num00)
-                        num00--;
-                    else
-                    {
-                        thissme = cpi->diamond_search_sad(x, b, d, &mvp_full,
-                                    &d->bmi.mv, step_param + n, sadpb, &num00,
-                                    &cpi->fn_ptr[BLOCK_16X16], x->mvcost,
-                                    &best_ref_mv);
+          thissme = cpi->refining_search_sad(
+              x, b, d, &d->bmi.mv, sadpb, search_range,
+              &cpi->fn_ptr[BLOCK_16X16], x->mvcost, &best_ref_mv);
 
-                        /* check to see if refining search is needed. */
-                        if (num00 > (further_steps-n))
-                            do_refine = 0;
-
-                        if (thissme < bestsme)
-                        {
-                            bestsme = thissme;
-                            mode_mv[NEWMV].as_int = d->bmi.mv.as_int;
-                        }
-                        else
-                        {
-                            d->bmi.mv.as_int = mode_mv[NEWMV].as_int;
-                        }
-                    }
-                }
-            }
+          if (thissme < bestsme) {
+            bestsme = thissme;
+            mode_mv[NEWMV].as_int = d->bmi.mv.as_int;
+          } else {
+            d->bmi.mv.as_int = mode_mv[NEWMV].as_int;
+          }
+        }
 
-            /* final 1-away diamond refining search */
-            if (do_refine == 1)
-            {
-                int search_range;
+        x->mv_col_min = tmp_col_min;
+        x->mv_col_max = tmp_col_max;
+        x->mv_row_min = tmp_row_min;
+        x->mv_row_max = tmp_row_max;
 
-                search_range = 8;
+        if (bestsme < INT_MAX) {
+          int dis; /* TODO: use dis in distortion calculation later. */
+          unsigned int sse;
+          cpi->find_fractional_mv_step(
+              x, b, d, &d->bmi.mv, &best_ref_mv, x->errorperbit,
+              &cpi->fn_ptr[BLOCK_16X16], x->mvcost, &dis, &sse);
+        }
 
-                thissme = cpi->refining_search_sad(x, b, d, &d->bmi.mv, sadpb,
-                                       search_range, &cpi->fn_ptr[BLOCK_16X16],
-                                       x->mvcost, &best_ref_mv);
+        mode_mv[NEWMV].as_int = d->bmi.mv.as_int;
 
-                if (thissme < bestsme)
-                {
-                    bestsme = thissme;
-                    mode_mv[NEWMV].as_int = d->bmi.mv.as_int;
-                }
-                else
-                {
-                    d->bmi.mv.as_int = mode_mv[NEWMV].as_int;
-                }
-            }
+        /* Add the new motion vector cost to our rolling cost variable */
+        rd.rate2 +=
+            vp8_mv_bit_cost(&mode_mv[NEWMV], &best_ref_mv, x->mvcost, 96);
+      }
 
-            x->mv_col_min = tmp_col_min;
-            x->mv_col_max = tmp_col_max;
-            x->mv_row_min = tmp_row_min;
-            x->mv_row_max = tmp_row_max;
+      case NEARESTMV:
+      case NEARMV:
+        /* Clip "next_nearest" so that it does not extend to far out
+         * of image
+         */
+        vp8_clamp_mv2(&mode_mv[this_mode], xd);
 
-            if (bestsme < INT_MAX)
-            {
-                int dis; /* TODO: use dis in distortion calculation later. */
-                unsigned int sse;
-                cpi->find_fractional_mv_step(x, b, d, &d->bmi.mv, &best_ref_mv,
-                                             x->errorperbit,
-                                             &cpi->fn_ptr[BLOCK_16X16],
-                                             x->mvcost, &dis, &sse);
-            }
+        /* Do not bother proceeding if the vector (from newmv, nearest
+         * or near) is 0,0 as this should then be coded using the zeromv
+         * mode.
+         */
+        if (((this_mode == NEARMV) || (this_mode == NEARESTMV)) &&
+            (mode_mv[this_mode].as_int == 0))
+          continue;
 
-            mode_mv[NEWMV].as_int = d->bmi.mv.as_int;
+      case ZEROMV:
 
-            /* Add the new motion vector cost to our rolling cost variable */
-            rd.rate2 += vp8_mv_bit_cost(&mode_mv[NEWMV], &best_ref_mv, x->mvcost, 96);
-        }
+        /* Trap vectors that reach beyond the UMV borders
+         * Note that ALL New MV, Nearest MV Near MV and Zero MV code
+         * drops through to this point because of the lack of break
+         * statements in the previous two cases.
+         */
+        if (((mode_mv[this_mode].as_mv.row >> 3) < x->mv_row_min) ||
+            ((mode_mv[this_mode].as_mv.row >> 3) > x->mv_row_max) ||
+            ((mode_mv[this_mode].as_mv.col >> 3) < x->mv_col_min) ||
+            ((mode_mv[this_mode].as_mv.col >> 3) > x->mv_col_max))
+          continue;
 
-        case NEARESTMV:
-        case NEARMV:
-            /* Clip "next_nearest" so that it does not extend to far out
-             * of image
-             */
-            vp8_clamp_mv2(&mode_mv[this_mode], xd);
+        vp8_set_mbmode_and_mvs(x, this_mode, &mode_mv[this_mode]);
+        this_rd = evaluate_inter_mode_rd(mdcounts, &rd, &disable_skip, cpi, x);
+        break;
 
-            /* Do not bother proceeding if the vector (from newmv, nearest
-             * or near) is 0,0 as this should then be coded using the zeromv
-             * mode.
-             */
-            if (((this_mode == NEARMV) || (this_mode == NEARESTMV)) && (mode_mv[this_mode].as_int == 0))
-                continue;
+      default: break;
+    }
 
-        case ZEROMV:
+    this_rd =
+        calculate_final_rd_costs(this_rd, &rd, &other_cost, disable_skip,
+                                 uv_intra_tteob, intra_rd_penalty, cpi, x);
 
-            /* Trap vectors that reach beyond the UMV borders
-             * Note that ALL New MV, Nearest MV Near MV and Zero MV code
-             * drops through to this point because of the lack of break
-             * statements in the previous two cases.
-             */
-            if (((mode_mv[this_mode].as_mv.row >> 3) < x->mv_row_min) || ((mode_mv[this_mode].as_mv.row >> 3) > x->mv_row_max) ||
-                ((mode_mv[this_mode].as_mv.col >> 3) < x->mv_col_min) || ((mode_mv[this_mode].as_mv.col >> 3) > x->mv_col_max))
-                continue;
-
-            vp8_set_mbmode_and_mvs(x, this_mode, &mode_mv[this_mode]);
-            this_rd = evaluate_inter_mode_rd(mdcounts, &rd,
-                                             &disable_skip, cpi, x);
-            break;
-
-        default:
-            break;
-        }
-
-        this_rd = calculate_final_rd_costs(this_rd, &rd, &other_cost,
-                                           disable_skip, uv_intra_tteob,
-                                           intra_rd_penalty, cpi, x);
-
-        /* Keep record of best intra distortion */
-        if ((x->e_mbd.mode_info_context->mbmi.ref_frame == INTRA_FRAME) &&
-            (this_rd < best_mode.intra_rd) )
-        {
-          best_mode.intra_rd = this_rd;
-            *returnintra = rd.distortion2 ;
-        }
+    /* Keep record of best intra distortion */
+    if ((x->e_mbd.mode_info_context->mbmi.ref_frame == INTRA_FRAME) &&
+        (this_rd < best_mode.intra_rd)) {
+      best_mode.intra_rd = this_rd;
+      *returnintra = rd.distortion2;
+    }
 #if CONFIG_TEMPORAL_DENOISING
-        if (cpi->oxcf.noise_sensitivity)
-        {
-            unsigned int sse;
-            vp8_get_inter_mbpred_error(x,&cpi->fn_ptr[BLOCK_16X16],&sse,
-                                   mode_mv[this_mode]);
+    if (cpi->oxcf.noise_sensitivity) {
+      unsigned int sse;
+      vp8_get_inter_mbpred_error(x, &cpi->fn_ptr[BLOCK_16X16], &sse,
+                                 mode_mv[this_mode]);
 
-            if (sse < best_rd_sse)
-                best_rd_sse = sse;
+      if (sse < best_rd_sse) best_rd_sse = sse;
 
-            /* Store for later use by denoiser. */
-            if (this_mode == ZEROMV && sse < zero_mv_sse )
-            {
-                zero_mv_sse = sse;
-                x->best_zeromv_reference_frame =
-                        x->e_mbd.mode_info_context->mbmi.ref_frame;
-            }
+      /* Store for later use by denoiser. */
+      if (this_mode == ZEROMV && sse < zero_mv_sse) {
+        zero_mv_sse = sse;
+        x->best_zeromv_reference_frame =
+            x->e_mbd.mode_info_context->mbmi.ref_frame;
+      }
 
-            /* Store the best NEWMV in x for later use in the denoiser. */
-            if (x->e_mbd.mode_info_context->mbmi.mode == NEWMV &&
-                    sse < best_sse)
-            {
-                best_sse = sse;
-                vp8_get_inter_mbpred_error(x,&cpi->fn_ptr[BLOCK_16X16],&best_sse,
-                                       mode_mv[this_mode]);
-                x->best_sse_inter_mode = NEWMV;
-                x->best_sse_mv = x->e_mbd.mode_info_context->mbmi.mv;
-                x->need_to_clamp_best_mvs =
-                    x->e_mbd.mode_info_context->mbmi.need_to_clamp_mvs;
-                x->best_reference_frame =
-                    x->e_mbd.mode_info_context->mbmi.ref_frame;
-            }
-        }
+      /* Store the best NEWMV in x for later use in the denoiser. */
+      if (x->e_mbd.mode_info_context->mbmi.mode == NEWMV && sse < best_sse) {
+        best_sse = sse;
+        vp8_get_inter_mbpred_error(x, &cpi->fn_ptr[BLOCK_16X16], &best_sse,
+                                   mode_mv[this_mode]);
+        x->best_sse_inter_mode = NEWMV;
+        x->best_sse_mv = x->e_mbd.mode_info_context->mbmi.mv;
+        x->need_to_clamp_best_mvs =
+            x->e_mbd.mode_info_context->mbmi.need_to_clamp_mvs;
+        x->best_reference_frame = x->e_mbd.mode_info_context->mbmi.ref_frame;
+      }
+    }
 #endif
 
-        /* Did this mode help.. i.i is it the new best mode */
-        if (this_rd < best_mode.rd || x->skip)
-        {
-            /* Note index of best mode so far */
-            best_mode_index = mode_index;
-            *returnrate = rd.rate2;
-            *returndistortion = rd.distortion2;
-            if (this_mode <= B_PRED)
-            {
-                x->e_mbd.mode_info_context->mbmi.uv_mode = uv_intra_mode;
-                /* required for left and above block mv */
-                x->e_mbd.mode_info_context->mbmi.mv.as_int = 0;
-            }
-            update_best_mode(&best_mode, this_rd, &rd, other_cost, x);
+    /* Did this mode help.. i.i is it the new best mode */
+    if (this_rd < best_mode.rd || x->skip) {
+      /* Note index of best mode so far */
+      best_mode_index = mode_index;
+      *returnrate = rd.rate2;
+      *returndistortion = rd.distortion2;
+      if (this_mode <= B_PRED) {
+        x->e_mbd.mode_info_context->mbmi.uv_mode = uv_intra_mode;
+        /* required for left and above block mv */
+        x->e_mbd.mode_info_context->mbmi.mv.as_int = 0;
+      }
+      update_best_mode(&best_mode, this_rd, &rd, other_cost, x);
 
+      /* Testing this mode gave rise to an improvement in best error
+       * score. Lower threshold a bit for next time
+       */
+      x->rd_thresh_mult[mode_index] =
+          (x->rd_thresh_mult[mode_index] >= (MIN_THRESHMULT + 2))
+              ? x->rd_thresh_mult[mode_index] - 2
+              : MIN_THRESHMULT;
+    }
 
-            /* Testing this mode gave rise to an improvement in best error
-             * score. Lower threshold a bit for next time
-             */
-            x->rd_thresh_mult[mode_index] =
-                (x->rd_thresh_mult[mode_index] >= (MIN_THRESHMULT + 2)) ?
-                    x->rd_thresh_mult[mode_index] - 2 : MIN_THRESHMULT;
-        }
+    /* If the mode did not help improve the best error case then raise
+     * the threshold for testing that mode next time around.
+     */
+    else {
+      x->rd_thresh_mult[mode_index] += 4;
 
-        /* If the mode did not help improve the best error case then raise
-         * the threshold for testing that mode next time around.
-         */
-        else
-        {
-            x->rd_thresh_mult[mode_index] += 4;
+      if (x->rd_thresh_mult[mode_index] > MAX_THRESHMULT)
+        x->rd_thresh_mult[mode_index] = MAX_THRESHMULT;
+    }
+    x->rd_threshes[mode_index] = (cpi->rd_baseline_thresh[mode_index] >> 7) *
+                                 x->rd_thresh_mult[mode_index];
 
-            if (x->rd_thresh_mult[mode_index] > MAX_THRESHMULT)
-                x->rd_thresh_mult[mode_index] = MAX_THRESHMULT;
-        }
-        x->rd_threshes[mode_index] =
-            (cpi->rd_baseline_thresh[mode_index] >> 7) *
-                x->rd_thresh_mult[mode_index];
+    if (x->skip) break;
+  }
 
-        if (x->skip)
-            break;
+  /* Reduce the activation RD thresholds for the best choice mode */
+  if ((cpi->rd_baseline_thresh[best_mode_index] > 0) &&
+      (cpi->rd_baseline_thresh[best_mode_index] < (INT_MAX >> 2))) {
+    int best_adjustment = (x->rd_thresh_mult[best_mode_index] >> 2);
 
-    }
+    x->rd_thresh_mult[best_mode_index] =
+        (x->rd_thresh_mult[best_mode_index] >=
+         (MIN_THRESHMULT + best_adjustment))
+            ? x->rd_thresh_mult[best_mode_index] - best_adjustment
+            : MIN_THRESHMULT;
+    x->rd_threshes[best_mode_index] =
+        (cpi->rd_baseline_thresh[best_mode_index] >> 7) *
+        x->rd_thresh_mult[best_mode_index];
+  }
 
-    /* Reduce the activation RD thresholds for the best choice mode */
-    if ((cpi->rd_baseline_thresh[best_mode_index] > 0) && (cpi->rd_baseline_thresh[best_mode_index] < (INT_MAX >> 2)))
-    {
-        int best_adjustment = (x->rd_thresh_mult[best_mode_index] >> 2);
-
-        x->rd_thresh_mult[best_mode_index] =
-            (x->rd_thresh_mult[best_mode_index] >=
-                (MIN_THRESHMULT + best_adjustment)) ?
-                    x->rd_thresh_mult[best_mode_index] - best_adjustment :
-                    MIN_THRESHMULT;
-        x->rd_threshes[best_mode_index] =
-            (cpi->rd_baseline_thresh[best_mode_index] >> 7) *
-                x->rd_thresh_mult[best_mode_index];
+#if CONFIG_TEMPORAL_DENOISING
+  if (cpi->oxcf.noise_sensitivity) {
+    int block_index = mb_row * cpi->common.mb_cols + mb_col;
+    if (x->best_sse_inter_mode == DC_PRED) {
+      /* No best MV found. */
+      x->best_sse_inter_mode = best_mode.mbmode.mode;
+      x->best_sse_mv = best_mode.mbmode.mv;
+      x->need_to_clamp_best_mvs = best_mode.mbmode.need_to_clamp_mvs;
+      x->best_reference_frame = best_mode.mbmode.ref_frame;
+      best_sse = best_rd_sse;
     }
+    vp8_denoiser_denoise_mb(&cpi->denoiser, x, best_sse, zero_mv_sse,
+                            recon_yoffset, recon_uvoffset, &cpi->common.lf_info,
+                            mb_row, mb_col, block_index, 0);
 
-#if CONFIG_TEMPORAL_DENOISING
-    if (cpi->oxcf.noise_sensitivity)
-    {
-        int block_index = mb_row * cpi->common.mb_cols + mb_col;
-        if (x->best_sse_inter_mode == DC_PRED)
-        {
-            /* No best MV found. */
-            x->best_sse_inter_mode = best_mode.mbmode.mode;
-            x->best_sse_mv = best_mode.mbmode.mv;
-            x->need_to_clamp_best_mvs = best_mode.mbmode.need_to_clamp_mvs;
-            x->best_reference_frame = best_mode.mbmode.ref_frame;
-            best_sse = best_rd_sse;
-        }
-        vp8_denoiser_denoise_mb(&cpi->denoiser, x, best_sse, zero_mv_sse,
-                                recon_yoffset, recon_uvoffset,
-                                &cpi->common.lf_info, mb_row, mb_col,
-                                block_index, 0);
+    /* Reevaluate ZEROMV after denoising. */
+    if (best_mode.mbmode.ref_frame == INTRA_FRAME &&
+        x->best_zeromv_reference_frame != INTRA_FRAME) {
+      int this_rd = INT_MAX;
+      int disable_skip = 0;
+      int other_cost = 0;
+      int this_ref_frame = x->best_zeromv_reference_frame;
+      rd.rate2 =
+          x->ref_frame_cost[this_ref_frame] + vp8_cost_mv_ref(ZEROMV, mdcounts);
+      rd.distortion2 = 0;
 
-        /* Reevaluate ZEROMV after denoising. */
-        if (best_mode.mbmode.ref_frame == INTRA_FRAME &&
-            x->best_zeromv_reference_frame != INTRA_FRAME)
-        {
-            int this_rd = INT_MAX;
-            int disable_skip = 0;
-            int other_cost = 0;
-            int this_ref_frame = x->best_zeromv_reference_frame;
-            rd.rate2 = x->ref_frame_cost[this_ref_frame] +
-                    vp8_cost_mv_ref(ZEROMV, mdcounts);
-            rd.distortion2 = 0;
+      /* set up the proper prediction buffers for the frame */
+      x->e_mbd.mode_info_context->mbmi.ref_frame = this_ref_frame;
+      x->e_mbd.pre.y_buffer = plane[this_ref_frame][0];
+      x->e_mbd.pre.u_buffer = plane[this_ref_frame][1];
+      x->e_mbd.pre.v_buffer = plane[this_ref_frame][2];
 
-            /* set up the proper prediction buffers for the frame */
-            x->e_mbd.mode_info_context->mbmi.ref_frame = this_ref_frame;
-            x->e_mbd.pre.y_buffer = plane[this_ref_frame][0];
-            x->e_mbd.pre.u_buffer = plane[this_ref_frame][1];
-            x->e_mbd.pre.v_buffer = plane[this_ref_frame][2];
+      x->e_mbd.mode_info_context->mbmi.mode = ZEROMV;
+      x->e_mbd.mode_info_context->mbmi.uv_mode = DC_PRED;
+      x->e_mbd.mode_info_context->mbmi.mv.as_int = 0;
 
-            x->e_mbd.mode_info_context->mbmi.mode = ZEROMV;
-            x->e_mbd.mode_info_context->mbmi.uv_mode = DC_PRED;
-            x->e_mbd.mode_info_context->mbmi.mv.as_int = 0;
-
-            this_rd = evaluate_inter_mode_rd(mdcounts, &rd, &disable_skip, cpi, x);
-            this_rd = calculate_final_rd_costs(this_rd, &rd, &other_cost,
-                                               disable_skip, uv_intra_tteob,
-                                               intra_rd_penalty, cpi, x);
-            if (this_rd < best_mode.rd || x->skip)
-            {
-                *returnrate = rd.rate2;
-                *returndistortion = rd.distortion2;
-                update_best_mode(&best_mode, this_rd, &rd, other_cost, x);
-            }
-        }
-
+      this_rd = evaluate_inter_mode_rd(mdcounts, &rd, &disable_skip, cpi, x);
+      this_rd =
+          calculate_final_rd_costs(this_rd, &rd, &other_cost, disable_skip,
+                                   uv_intra_tteob, intra_rd_penalty, cpi, x);
+      if (this_rd < best_mode.rd || x->skip) {
+        *returnrate = rd.rate2;
+        *returndistortion = rd.distortion2;
+        update_best_mode(&best_mode, this_rd, &rd, other_cost, x);
+      }
     }
+  }
 #endif
 
-    if (cpi->is_src_frame_alt_ref &&
-        (best_mode.mbmode.mode != ZEROMV || best_mode.mbmode.ref_frame != ALTREF_FRAME))
-    {
-        x->e_mbd.mode_info_context->mbmi.mode = ZEROMV;
-        x->e_mbd.mode_info_context->mbmi.ref_frame = ALTREF_FRAME;
-        x->e_mbd.mode_info_context->mbmi.mv.as_int = 0;
-        x->e_mbd.mode_info_context->mbmi.uv_mode = DC_PRED;
-        x->e_mbd.mode_info_context->mbmi.mb_skip_coeff =
-                                        (cpi->common.mb_no_coeff_skip);
-        x->e_mbd.mode_info_context->mbmi.partitioning = 0;
-        return;
-    }
+  if (cpi->is_src_frame_alt_ref &&
+      (best_mode.mbmode.mode != ZEROMV ||
+       best_mode.mbmode.ref_frame != ALTREF_FRAME)) {
+    x->e_mbd.mode_info_context->mbmi.mode = ZEROMV;
+    x->e_mbd.mode_info_context->mbmi.ref_frame = ALTREF_FRAME;
+    x->e_mbd.mode_info_context->mbmi.mv.as_int = 0;
+    x->e_mbd.mode_info_context->mbmi.uv_mode = DC_PRED;
+    x->e_mbd.mode_info_context->mbmi.mb_skip_coeff =
+        (cpi->common.mb_no_coeff_skip);
+    x->e_mbd.mode_info_context->mbmi.partitioning = 0;
+    return;
+  }
 
+  /* macroblock modes */
+  memcpy(&x->e_mbd.mode_info_context->mbmi, &best_mode.mbmode,
+         sizeof(MB_MODE_INFO));
 
-    /* macroblock modes */
-    memcpy(&x->e_mbd.mode_info_context->mbmi, &best_mode.mbmode, sizeof(MB_MODE_INFO));
+  if (best_mode.mbmode.mode == B_PRED) {
+    for (i = 0; i < 16; i++)
+      xd->mode_info_context->bmi[i].as_mode = best_mode.bmodes[i].as_mode;
+  }
 
-    if (best_mode.mbmode.mode == B_PRED)
-    {
-        for (i = 0; i < 16; i++)
-            xd->mode_info_context->bmi[i].as_mode = best_mode.bmodes[i].as_mode;
-    }
+  if (best_mode.mbmode.mode == SPLITMV) {
+    for (i = 0; i < 16; i++)
+      xd->mode_info_context->bmi[i].mv.as_int = best_mode.bmodes[i].mv.as_int;
 
-    if (best_mode.mbmode.mode == SPLITMV)
-    {
-        for (i = 0; i < 16; i++)
-            xd->mode_info_context->bmi[i].mv.as_int = best_mode.bmodes[i].mv.as_int;
+    memcpy(x->partition_info, &best_mode.partition, sizeof(PARTITION_INFO));
 
-        memcpy(x->partition_info, &best_mode.partition, sizeof(PARTITION_INFO));
+    x->e_mbd.mode_info_context->mbmi.mv.as_int =
+        x->partition_info->bmi[15].mv.as_int;
+  }
 
-        x->e_mbd.mode_info_context->mbmi.mv.as_int =
-                                      x->partition_info->bmi[15].mv.as_int;
-    }
+  if (sign_bias !=
+      cpi->common.ref_frame_sign_bias[xd->mode_info_context->mbmi.ref_frame])
+    best_ref_mv.as_int = best_ref_mv_sb[!sign_bias].as_int;
 
-    if (sign_bias
-        != cpi->common.ref_frame_sign_bias[xd->mode_info_context->mbmi.ref_frame])
-        best_ref_mv.as_int = best_ref_mv_sb[!sign_bias].as_int;
-
-    rd_update_mvcount(x, &best_ref_mv);
+  rd_update_mvcount(x, &best_ref_mv);
 }
 
-void vp8_rd_pick_intra_mode(MACROBLOCK *x, int *rate_)
-{
-    int error4x4, error16x16;
-    int rate4x4, rate16x16 = 0, rateuv;
-    int dist4x4, dist16x16, distuv;
-    int rate;
-    int rate4x4_tokenonly = 0;
-    int rate16x16_tokenonly = 0;
-    int rateuv_tokenonly = 0;
+void vp8_rd_pick_intra_mode(MACROBLOCK *x, int *rate_) {
+  int error4x4, error16x16;
+  int rate4x4, rate16x16 = 0, rateuv;
+  int dist4x4, dist16x16, distuv;
+  int rate;
+  int rate4x4_tokenonly = 0;
+  int rate16x16_tokenonly = 0;
+  int rateuv_tokenonly = 0;
 
-    x->e_mbd.mode_info_context->mbmi.ref_frame = INTRA_FRAME;
+  x->e_mbd.mode_info_context->mbmi.ref_frame = INTRA_FRAME;
 
-    rd_pick_intra_mbuv_mode(x, &rateuv, &rateuv_tokenonly, &distuv);
-    rate = rateuv;
+  rd_pick_intra_mbuv_mode(x, &rateuv, &rateuv_tokenonly, &distuv);
+  rate = rateuv;
 
-    error16x16 = rd_pick_intra16x16mby_mode(x, &rate16x16, &rate16x16_tokenonly,
-                                            &dist16x16);
+  error16x16 = rd_pick_intra16x16mby_mode(x, &rate16x16, &rate16x16_tokenonly,
+                                          &dist16x16);
 
-    error4x4 = rd_pick_intra4x4mby_modes(x, &rate4x4, &rate4x4_tokenonly,
-                                         &dist4x4, error16x16);
+  error4x4 = rd_pick_intra4x4mby_modes(x, &rate4x4, &rate4x4_tokenonly,
+                                       &dist4x4, error16x16);
 
-    if (error4x4 < error16x16)
-    {
-        x->e_mbd.mode_info_context->mbmi.mode = B_PRED;
-        rate += rate4x4;
-    }
-    else
-    {
-        rate += rate16x16;
-    }
+  if (error4x4 < error16x16) {
+    x->e_mbd.mode_info_context->mbmi.mode = B_PRED;
+    rate += rate4x4;
+  } else {
+    rate += rate16x16;
+  }
 
-    *rate_ = rate;
+  *rate_ = rate;
 }
--- a/vp8/encoder/rdopt.h
+++ b/vp8/encoder/rdopt.h
@@ -8,7 +8,6 @@
  *  be found in the AUTHORS file in the root of the source tree.
  */
 
-
 #ifndef VP8_ENCODER_RDOPT_H_
 #define VP8_ENCODER_RDOPT_H_
 
@@ -18,57 +17,47 @@
 extern "C" {
 #endif
 
-#define RDCOST(RM,DM,R,D) ( ((128+(R)*(RM)) >> 8) + (DM)*(D) )
+#define RDCOST(RM, DM, R, D) (((128 + (R) * (RM)) >> 8) + (DM) * (D))
 
-static INLINE void insertsortmv(int arr[], int len)
-{
-    int i, j, k;
+static INLINE void insertsortmv(int arr[], int len) {
+  int i, j, k;
 
-    for ( i = 1 ; i <= len-1 ; i++ )
-    {
-        for ( j = 0 ; j < i ; j++ )
-        {
-            if ( arr[j] > arr[i] )
-            {
-                int temp;
+  for (i = 1; i <= len - 1; i++) {
+    for (j = 0; j < i; j++) {
+      if (arr[j] > arr[i]) {
+        int temp;
 
-                temp = arr[i];
+        temp = arr[i];
 
-                for ( k = i; k >j; k--)
-                    arr[k] = arr[k - 1] ;
+        for (k = i; k > j; k--) arr[k] = arr[k - 1];
 
-                arr[j] = temp ;
-            }
-        }
+        arr[j] = temp;
+      }
     }
+  }
 }
 
-static INLINE void insertsortsad(int arr[],int idx[], int len)
-{
-    int i, j, k;
+static INLINE void insertsortsad(int arr[], int idx[], int len) {
+  int i, j, k;
 
-    for ( i = 1 ; i <= len-1 ; i++ )
-    {
-        for ( j = 0 ; j < i ; j++ )
-        {
-            if ( arr[j] > arr[i] )
-            {
-                int temp, tempi;
+  for (i = 1; i <= len - 1; i++) {
+    for (j = 0; j < i; j++) {
+      if (arr[j] > arr[i]) {
+        int temp, tempi;
 
-                temp = arr[i];
-                tempi = idx[i];
+        temp = arr[i];
+        tempi = idx[i];
 
-                for ( k = i; k >j; k--)
-                {
-                    arr[k] = arr[k - 1] ;
-                    idx[k] = idx[k - 1];
-                }
-
-                arr[j] = temp ;
-                idx[j] = tempi;
-            }
+        for (k = i; k > j; k--) {
+          arr[k] = arr[k - 1];
+          idx[k] = idx[k - 1];
         }
+
+        arr[j] = temp;
+        idx[j] = tempi;
+      }
     }
+  }
 }
 
 extern void vp8_initialize_rd_consts(VP8_COMP *cpi, MACROBLOCK *x, int Qvalue);
@@ -78,66 +67,48 @@
                                    int *returnintra, int mb_row, int mb_col);
 extern void vp8_rd_pick_intra_mode(MACROBLOCK *x, int *rate);
 
-
 static INLINE void get_plane_pointers(const YV12_BUFFER_CONFIG *fb,
                                       unsigned char *plane[3],
                                       unsigned int recon_yoffset,
-                                      unsigned int recon_uvoffset)
-{
-    plane[0] = fb->y_buffer + recon_yoffset;
-    plane[1] = fb->u_buffer + recon_uvoffset;
-    plane[2] = fb->v_buffer + recon_uvoffset;
+                                      unsigned int recon_uvoffset) {
+  plane[0] = fb->y_buffer + recon_yoffset;
+  plane[1] = fb->u_buffer + recon_uvoffset;
+  plane[2] = fb->v_buffer + recon_uvoffset;
 }
 
-
 static INLINE void get_predictor_pointers(const VP8_COMP *cpi,
                                           unsigned char *plane[4][3],
                                           unsigned int recon_yoffset,
-                                          unsigned int recon_uvoffset)
-{
-    if (cpi->ref_frame_flags & VP8_LAST_FRAME)
-        get_plane_pointers(&cpi->common.yv12_fb[cpi->common.lst_fb_idx],
-                           plane[LAST_FRAME], recon_yoffset, recon_uvoffset);
+                                          unsigned int recon_uvoffset) {
+  if (cpi->ref_frame_flags & VP8_LAST_FRAME)
+    get_plane_pointers(&cpi->common.yv12_fb[cpi->common.lst_fb_idx],
+                       plane[LAST_FRAME], recon_yoffset, recon_uvoffset);
 
-    if (cpi->ref_frame_flags & VP8_GOLD_FRAME)
-        get_plane_pointers(&cpi->common.yv12_fb[cpi->common.gld_fb_idx],
-                           plane[GOLDEN_FRAME], recon_yoffset, recon_uvoffset);
+  if (cpi->ref_frame_flags & VP8_GOLD_FRAME)
+    get_plane_pointers(&cpi->common.yv12_fb[cpi->common.gld_fb_idx],
+                       plane[GOLDEN_FRAME], recon_yoffset, recon_uvoffset);
 
-    if (cpi->ref_frame_flags & VP8_ALTR_FRAME)
-        get_plane_pointers(&cpi->common.yv12_fb[cpi->common.alt_fb_idx],
-                           plane[ALTREF_FRAME], recon_yoffset, recon_uvoffset);
+  if (cpi->ref_frame_flags & VP8_ALTR_FRAME)
+    get_plane_pointers(&cpi->common.yv12_fb[cpi->common.alt_fb_idx],
+                       plane[ALTREF_FRAME], recon_yoffset, recon_uvoffset);
 }
 
-
 static INLINE void get_reference_search_order(const VP8_COMP *cpi,
-                                              int ref_frame_map[4])
-{
-    int i=0;
+                                              int ref_frame_map[4]) {
+  int i = 0;
 
-    ref_frame_map[i++] = INTRA_FRAME;
-    if (cpi->ref_frame_flags & VP8_LAST_FRAME)
-        ref_frame_map[i++] = LAST_FRAME;
-    if (cpi->ref_frame_flags & VP8_GOLD_FRAME)
-        ref_frame_map[i++] = GOLDEN_FRAME;
-    if (cpi->ref_frame_flags & VP8_ALTR_FRAME)
-        ref_frame_map[i++] = ALTREF_FRAME;
-    for(; i<4; i++)
-        ref_frame_map[i] = -1;
+  ref_frame_map[i++] = INTRA_FRAME;
+  if (cpi->ref_frame_flags & VP8_LAST_FRAME) ref_frame_map[i++] = LAST_FRAME;
+  if (cpi->ref_frame_flags & VP8_GOLD_FRAME) ref_frame_map[i++] = GOLDEN_FRAME;
+  if (cpi->ref_frame_flags & VP8_ALTR_FRAME) ref_frame_map[i++] = ALTREF_FRAME;
+  for (; i < 4; i++) ref_frame_map[i] = -1;
 }
 
-
-extern void vp8_mv_pred
-(
-    VP8_COMP *cpi,
-    MACROBLOCKD *xd,
-    const MODE_INFO *here,
-    int_mv *mvp,
-    int refframe,
-    int *ref_frame_sign_bias,
-    int *sr,
-    int near_sadidx[]
-);
-void vp8_cal_sad(VP8_COMP *cpi, MACROBLOCKD *xd, MACROBLOCK *x, int recon_yoffset, int near_sadidx[]);
+extern void vp8_mv_pred(VP8_COMP *cpi, MACROBLOCKD *xd, const MODE_INFO *here,
+                        int_mv *mvp, int refframe, int *ref_frame_sign_bias,
+                        int *sr, int near_sadidx[]);
+void vp8_cal_sad(VP8_COMP *cpi, MACROBLOCKD *xd, MACROBLOCK *x,
+                 int recon_yoffset, int near_sadidx[]);
 int VP8_UVSSE(MACROBLOCK *x);
 int vp8_cost_mv_ref(MB_PREDICTION_MODE m, const int near_mv_ref_ct[4]);
 void vp8_set_mbmode_and_mvs(MACROBLOCK *x, MB_PREDICTION_MODE mb, int_mv *mv);
--- a/vp8/encoder/segmentation.c
+++ b/vp8/encoder/segmentation.c
@@ -8,59 +8,48 @@
  *  be found in the AUTHORS file in the root of the source tree.
  */
 
-
 #include "segmentation.h"
 #include "vpx_mem/vpx_mem.h"
 
-void vp8_update_gf_useage_maps(VP8_COMP *cpi, VP8_COMMON *cm, MACROBLOCK *x)
-{
-    int mb_row, mb_col;
+void vp8_update_gf_useage_maps(VP8_COMP *cpi, VP8_COMMON *cm, MACROBLOCK *x) {
+  int mb_row, mb_col;
 
-    MODE_INFO *this_mb_mode_info = cm->mi;
+  MODE_INFO *this_mb_mode_info = cm->mi;
 
-    x->gf_active_ptr = (signed char *)cpi->gf_active_flags;
+  x->gf_active_ptr = (signed char *)cpi->gf_active_flags;
 
-    if ((cm->frame_type == KEY_FRAME) || (cm->refresh_golden_frame))
-    {
-        /* Reset Gf useage monitors */
-        memset(cpi->gf_active_flags, 1, (cm->mb_rows * cm->mb_cols));
-        cpi->gf_active_count = cm->mb_rows * cm->mb_cols;
-    }
-    else
-    {
-        /* for each macroblock row in image */
-        for (mb_row = 0; mb_row < cm->mb_rows; mb_row++)
-        {
-            /* for each macroblock col in image */
-            for (mb_col = 0; mb_col < cm->mb_cols; mb_col++)
-            {
+  if ((cm->frame_type == KEY_FRAME) || (cm->refresh_golden_frame)) {
+    /* Reset Gf useage monitors */
+    memset(cpi->gf_active_flags, 1, (cm->mb_rows * cm->mb_cols));
+    cpi->gf_active_count = cm->mb_rows * cm->mb_cols;
+  } else {
+    /* for each macroblock row in image */
+    for (mb_row = 0; mb_row < cm->mb_rows; mb_row++) {
+      /* for each macroblock col in image */
+      for (mb_col = 0; mb_col < cm->mb_cols; mb_col++) {
+        /* If using golden then set GF active flag if not already set.
+         * If using last frame 0,0 mode then leave flag as it is
+         * else if using non 0,0 motion or intra modes then clear
+         * flag if it is currently set
+         */
+        if ((this_mb_mode_info->mbmi.ref_frame == GOLDEN_FRAME) ||
+            (this_mb_mode_info->mbmi.ref_frame == ALTREF_FRAME)) {
+          if (*(x->gf_active_ptr) == 0) {
+            *(x->gf_active_ptr) = 1;
+            cpi->gf_active_count++;
+          }
+        } else if ((this_mb_mode_info->mbmi.mode != ZEROMV) &&
+                   *(x->gf_active_ptr)) {
+          *(x->gf_active_ptr) = 0;
+          cpi->gf_active_count--;
+        }
 
-                /* If using golden then set GF active flag if not already set.
-                 * If using last frame 0,0 mode then leave flag as it is
-                 * else if using non 0,0 motion or intra modes then clear
-                 * flag if it is currently set
-                 */
-                if ((this_mb_mode_info->mbmi.ref_frame == GOLDEN_FRAME) || (this_mb_mode_info->mbmi.ref_frame == ALTREF_FRAME))
-                {
-                    if (*(x->gf_active_ptr) == 0)
-                    {
-                        *(x->gf_active_ptr) = 1;
-                        cpi->gf_active_count ++;
-                    }
-                }
-                else if ((this_mb_mode_info->mbmi.mode != ZEROMV) && *(x->gf_active_ptr))
-                {
-                    *(x->gf_active_ptr) = 0;
-                    cpi->gf_active_count--;
-                }
+        x->gf_active_ptr++;  /* Step onto next entry */
+        this_mb_mode_info++; /* skip to next mb */
+      }
 
-                x->gf_active_ptr++;          /* Step onto next entry */
-                this_mb_mode_info++;         /* skip to next mb */
-
-            }
-
-            /* this is to account for the border */
-            this_mb_mode_info++;
-        }
+      /* this is to account for the border */
+      this_mb_mode_info++;
     }
+  }
 }
--- a/vp8/encoder/segmentation.h
+++ b/vp8/encoder/segmentation.h
@@ -19,7 +19,8 @@
 extern "C" {
 #endif
 
-extern void vp8_update_gf_useage_maps(VP8_COMP *cpi, VP8_COMMON *cm, MACROBLOCK *x);
+extern void vp8_update_gf_useage_maps(VP8_COMP *cpi, VP8_COMMON *cm,
+                                      MACROBLOCK *x);
 
 #ifdef __cplusplus
 }  // extern "C"
--- a/vp8/encoder/temporal_filter.c
+++ b/vp8/encoder/temporal_filter.c
@@ -8,7 +8,6 @@
  *  be found in the AUTHORS file in the root of the source tree.
  */
 
-
 #include "vp8/common/onyxc_int.h"
 #include "onyx_int.h"
 #include "vp8/common/systemdependent.h"
@@ -29,493 +28,399 @@
 #include <math.h>
 #include <limits.h>
 
-#define ALT_REF_MC_ENABLED 1    /* dis/enable MC in AltRef filtering */
-#define ALT_REF_SUBPEL_ENABLED 1 /* dis/enable subpel in MC AltRef filtering */
+#define ALT_REF_MC_ENABLED 1     /* toggle MC in AltRef filtering */
+#define ALT_REF_SUBPEL_ENABLED 1 /* toggle subpel in MC AltRef filtering */
 
 #if VP8_TEMPORAL_ALT_REF
 
-static void vp8_temporal_filter_predictors_mb_c
-(
-    MACROBLOCKD *x,
-    unsigned char *y_mb_ptr,
-    unsigned char *u_mb_ptr,
-    unsigned char *v_mb_ptr,
-    int stride,
-    int mv_row,
-    int mv_col,
-    unsigned char *pred
-)
-{
-    int offset;
-    unsigned char *yptr, *uptr, *vptr;
+static void vp8_temporal_filter_predictors_mb_c(
+    MACROBLOCKD *x, unsigned char *y_mb_ptr, unsigned char *u_mb_ptr,
+    unsigned char *v_mb_ptr, int stride, int mv_row, int mv_col,
+    unsigned char *pred) {
+  int offset;
+  unsigned char *yptr, *uptr, *vptr;
 
-    /* Y */
-    yptr = y_mb_ptr + (mv_row >> 3) * stride + (mv_col >> 3);
+  /* Y */
+  yptr = y_mb_ptr + (mv_row >> 3) * stride + (mv_col >> 3);
 
-    if ((mv_row | mv_col) & 7)
-    {
-        x->subpixel_predict16x16(yptr, stride,
-                                    mv_col & 7, mv_row & 7, &pred[0], 16);
-    }
-    else
-    {
-        vp8_copy_mem16x16(yptr, stride, &pred[0], 16);
-    }
+  if ((mv_row | mv_col) & 7) {
+    x->subpixel_predict16x16(yptr, stride, mv_col & 7, mv_row & 7, &pred[0],
+                             16);
+  } else {
+    vp8_copy_mem16x16(yptr, stride, &pred[0], 16);
+  }
 
-    /* U & V */
-    mv_row >>= 1;
-    mv_col >>= 1;
-    stride = (stride + 1) >> 1;
-    offset = (mv_row >> 3) * stride + (mv_col >> 3);
-    uptr = u_mb_ptr + offset;
-    vptr = v_mb_ptr + offset;
+  /* U & V */
+  mv_row >>= 1;
+  mv_col >>= 1;
+  stride = (stride + 1) >> 1;
+  offset = (mv_row >> 3) * stride + (mv_col >> 3);
+  uptr = u_mb_ptr + offset;
+  vptr = v_mb_ptr + offset;
 
-    if ((mv_row | mv_col) & 7)
-    {
-        x->subpixel_predict8x8(uptr, stride,
-                            mv_col & 7, mv_row & 7, &pred[256], 8);
-        x->subpixel_predict8x8(vptr, stride,
-                            mv_col & 7, mv_row & 7, &pred[320], 8);
-    }
-    else
-    {
-        vp8_copy_mem8x8(uptr, stride, &pred[256], 8);
-        vp8_copy_mem8x8(vptr, stride, &pred[320], 8);
-    }
+  if ((mv_row | mv_col) & 7) {
+    x->subpixel_predict8x8(uptr, stride, mv_col & 7, mv_row & 7, &pred[256], 8);
+    x->subpixel_predict8x8(vptr, stride, mv_col & 7, mv_row & 7, &pred[320], 8);
+  } else {
+    vp8_copy_mem8x8(uptr, stride, &pred[256], 8);
+    vp8_copy_mem8x8(vptr, stride, &pred[320], 8);
+  }
 }
-void vp8_temporal_filter_apply_c
-(
-    unsigned char *frame1,
-    unsigned int stride,
-    unsigned char *frame2,
-    unsigned int block_size,
-    int strength,
-    int filter_weight,
-    unsigned int *accumulator,
-    unsigned short *count
-)
-{
-    unsigned int i, j, k;
-    int modifier;
-    int byte = 0;
-    const int rounding = strength > 0 ? 1 << (strength - 1) : 0;
+void vp8_temporal_filter_apply_c(unsigned char *frame1, unsigned int stride,
+                                 unsigned char *frame2, unsigned int block_size,
+                                 int strength, int filter_weight,
+                                 unsigned int *accumulator,
+                                 unsigned short *count) {
+  unsigned int i, j, k;
+  int modifier;
+  int byte = 0;
+  const int rounding = strength > 0 ? 1 << (strength - 1) : 0;
 
-    for (i = 0,k = 0; i < block_size; i++)
-    {
-        for (j = 0; j < block_size; j++, k++)
-        {
+  for (i = 0, k = 0; i < block_size; i++) {
+    for (j = 0; j < block_size; j++, k++) {
+      int src_byte = frame1[byte];
+      int pixel_value = *frame2++;
 
-            int src_byte = frame1[byte];
-            int pixel_value = *frame2++;
+      modifier = src_byte - pixel_value;
+      /* This is an integer approximation of:
+       * float coeff = (3.0 * modifer * modifier) / pow(2, strength);
+       * modifier =  (int)roundf(coeff > 16 ? 0 : 16-coeff);
+       */
+      modifier *= modifier;
+      modifier *= 3;
+      modifier += rounding;
+      modifier >>= strength;
 
-            modifier   = src_byte - pixel_value;
-            /* This is an integer approximation of:
-             * float coeff = (3.0 * modifer * modifier) / pow(2, strength);
-             * modifier =  (int)roundf(coeff > 16 ? 0 : 16-coeff);
-             */
-            modifier  *= modifier;
-            modifier  *= 3;
-            modifier  += rounding;
-            modifier >>= strength;
+      if (modifier > 16) modifier = 16;
 
-            if (modifier > 16)
-                modifier = 16;
+      modifier = 16 - modifier;
+      modifier *= filter_weight;
 
-            modifier = 16 - modifier;
-            modifier *= filter_weight;
+      count[k] += modifier;
+      accumulator[k] += modifier * pixel_value;
 
-            count[k] += modifier;
-            accumulator[k] += modifier * pixel_value;
-
-            byte++;
-        }
-
-        byte += stride - block_size;
+      byte++;
     }
+
+    byte += stride - block_size;
+  }
 }
 
 #if ALT_REF_MC_ENABLED
 
-static int vp8_temporal_filter_find_matching_mb_c
-(
-    VP8_COMP *cpi,
-    YV12_BUFFER_CONFIG *arf_frame,
-    YV12_BUFFER_CONFIG *frame_ptr,
-    int mb_offset,
-    int error_thresh
-)
-{
-    MACROBLOCK *x = &cpi->mb;
-    int step_param;
-    int sadpb = x->sadperbit16;
-    int bestsme = INT_MAX;
+static int vp8_temporal_filter_find_matching_mb_c(VP8_COMP *cpi,
+                                                  YV12_BUFFER_CONFIG *arf_frame,
+                                                  YV12_BUFFER_CONFIG *frame_ptr,
+                                                  int mb_offset,
+                                                  int error_thresh) {
+  MACROBLOCK *x = &cpi->mb;
+  int step_param;
+  int sadpb = x->sadperbit16;
+  int bestsme = INT_MAX;
 
-    BLOCK *b = &x->block[0];
-    BLOCKD *d = &x->e_mbd.block[0];
-    int_mv best_ref_mv1;
-    int_mv best_ref_mv1_full; /* full-pixel value of best_ref_mv1 */
+  BLOCK *b = &x->block[0];
+  BLOCKD *d = &x->e_mbd.block[0];
+  int_mv best_ref_mv1;
+  int_mv best_ref_mv1_full; /* full-pixel value of best_ref_mv1 */
 
-    /* Save input state */
-    unsigned char **base_src = b->base_src;
-    int src = b->src;
-    int src_stride = b->src_stride;
-    unsigned char *base_pre = x->e_mbd.pre.y_buffer;
-    int pre = d->offset;
-    int pre_stride = x->e_mbd.pre.y_stride;
+  /* Save input state */
+  unsigned char **base_src = b->base_src;
+  int src = b->src;
+  int src_stride = b->src_stride;
+  unsigned char *base_pre = x->e_mbd.pre.y_buffer;
+  int pre = d->offset;
+  int pre_stride = x->e_mbd.pre.y_stride;
 
-    (void)error_thresh;
+  (void)error_thresh;
 
-    best_ref_mv1.as_int = 0;
-    best_ref_mv1_full.as_mv.col = best_ref_mv1.as_mv.col >>3;
-    best_ref_mv1_full.as_mv.row = best_ref_mv1.as_mv.row >>3;
+  best_ref_mv1.as_int = 0;
+  best_ref_mv1_full.as_mv.col = best_ref_mv1.as_mv.col >> 3;
+  best_ref_mv1_full.as_mv.row = best_ref_mv1.as_mv.row >> 3;
 
-    /* Setup frame pointers */
-    b->base_src = &arf_frame->y_buffer;
-    b->src_stride = arf_frame->y_stride;
-    b->src = mb_offset;
+  /* Setup frame pointers */
+  b->base_src = &arf_frame->y_buffer;
+  b->src_stride = arf_frame->y_stride;
+  b->src = mb_offset;
 
-    x->e_mbd.pre.y_buffer = frame_ptr->y_buffer;
-    x->e_mbd.pre.y_stride = frame_ptr->y_stride;
-    d->offset = mb_offset;
+  x->e_mbd.pre.y_buffer = frame_ptr->y_buffer;
+  x->e_mbd.pre.y_stride = frame_ptr->y_stride;
+  d->offset = mb_offset;
 
-    /* Further step/diamond searches as necessary */
-    if (cpi->Speed < 8)
-    {
-        step_param = cpi->sf.first_step + (cpi->Speed > 5);
-    }
-    else
-    {
-        step_param = cpi->sf.first_step + 2;
-    }
+  /* Further step/diamond searches as necessary */
+  if (cpi->Speed < 8) {
+    step_param = cpi->sf.first_step + (cpi->Speed > 5);
+  } else {
+    step_param = cpi->sf.first_step + 2;
+  }
 
-    /* TODO Check that the 16x16 vf & sdf are selected here */
-    /* Ignore mv costing by sending NULL cost arrays */
-    bestsme = vp8_hex_search(x, b, d, &best_ref_mv1_full, &d->bmi.mv,
-                             step_param, sadpb,
-                             &cpi->fn_ptr[BLOCK_16X16],
-                             NULL, NULL, &best_ref_mv1);
+  /* TODO Check that the 16x16 vf & sdf are selected here */
+  /* Ignore mv costing by sending NULL cost arrays */
+  bestsme =
+      vp8_hex_search(x, b, d, &best_ref_mv1_full, &d->bmi.mv, step_param, sadpb,
+                     &cpi->fn_ptr[BLOCK_16X16], NULL, NULL, &best_ref_mv1);
 
 #if ALT_REF_SUBPEL_ENABLED
-    /* Try sub-pixel MC? */
-    {
-        int distortion;
-        unsigned int sse;
-        /* Ignore mv costing by sending NULL cost array */
-        bestsme = cpi->find_fractional_mv_step(x, b, d,
-                                               &d->bmi.mv,
-                                               &best_ref_mv1,
-                                               x->errorperbit,
-                                               &cpi->fn_ptr[BLOCK_16X16],
-                                               NULL, &distortion, &sse);
-    }
+  /* Try sub-pixel MC? */
+  {
+    int distortion;
+    unsigned int sse;
+    /* Ignore mv costing by sending NULL cost array */
+    bestsme = cpi->find_fractional_mv_step(
+        x, b, d, &d->bmi.mv, &best_ref_mv1, x->errorperbit,
+        &cpi->fn_ptr[BLOCK_16X16], NULL, &distortion, &sse);
+  }
 #endif
 
-    /* Save input state */
-    b->base_src = base_src;
-    b->src = src;
-    b->src_stride = src_stride;
-    x->e_mbd.pre.y_buffer = base_pre;
-    d->offset = pre;
-    x->e_mbd.pre.y_stride = pre_stride;
+  /* Save input state */
+  b->base_src = base_src;
+  b->src = src;
+  b->src_stride = src_stride;
+  x->e_mbd.pre.y_buffer = base_pre;
+  d->offset = pre;
+  x->e_mbd.pre.y_stride = pre_stride;
 
-    return bestsme;
+  return bestsme;
 }
 #endif
 
-static void vp8_temporal_filter_iterate_c
-(
-    VP8_COMP *cpi,
-    int frame_count,
-    int alt_ref_index,
-    int strength
-)
-{
-    int byte;
-    int frame;
-    int mb_col, mb_row;
-    unsigned int filter_weight;
-    int mb_cols = cpi->common.mb_cols;
-    int mb_rows = cpi->common.mb_rows;
-    int mb_y_offset = 0;
-    int mb_uv_offset = 0;
-    DECLARE_ALIGNED(16, unsigned int, accumulator[16*16 + 8*8 + 8*8]);
-    DECLARE_ALIGNED(16, unsigned short, count[16*16 + 8*8 + 8*8]);
-    MACROBLOCKD *mbd = &cpi->mb.e_mbd;
-    YV12_BUFFER_CONFIG *f = cpi->frames[alt_ref_index];
-    unsigned char *dst1, *dst2;
-    DECLARE_ALIGNED(16, unsigned char,  predictor[16*16 + 8*8 + 8*8]);
+static void vp8_temporal_filter_iterate_c(VP8_COMP *cpi, int frame_count,
+                                          int alt_ref_index, int strength) {
+  int byte;
+  int frame;
+  int mb_col, mb_row;
+  unsigned int filter_weight;
+  int mb_cols = cpi->common.mb_cols;
+  int mb_rows = cpi->common.mb_rows;
+  int mb_y_offset = 0;
+  int mb_uv_offset = 0;
+  DECLARE_ALIGNED(16, unsigned int, accumulator[16 * 16 + 8 * 8 + 8 * 8]);
+  DECLARE_ALIGNED(16, unsigned short, count[16 * 16 + 8 * 8 + 8 * 8]);
+  MACROBLOCKD *mbd = &cpi->mb.e_mbd;
+  YV12_BUFFER_CONFIG *f = cpi->frames[alt_ref_index];
+  unsigned char *dst1, *dst2;
+  DECLARE_ALIGNED(16, unsigned char, predictor[16 * 16 + 8 * 8 + 8 * 8]);
 
-    /* Save input state */
-    unsigned char *y_buffer = mbd->pre.y_buffer;
-    unsigned char *u_buffer = mbd->pre.u_buffer;
-    unsigned char *v_buffer = mbd->pre.v_buffer;
+  /* Save input state */
+  unsigned char *y_buffer = mbd->pre.y_buffer;
+  unsigned char *u_buffer = mbd->pre.u_buffer;
+  unsigned char *v_buffer = mbd->pre.v_buffer;
 
-    for (mb_row = 0; mb_row < mb_rows; mb_row++)
-    {
+  for (mb_row = 0; mb_row < mb_rows; mb_row++) {
 #if ALT_REF_MC_ENABLED
-        /* Source frames are extended to 16 pixels.  This is different than
-         *  L/A/G reference frames that have a border of 32 (VP8BORDERINPIXELS)
-         * A 6 tap filter is used for motion search.  This requires 2 pixels
-         *  before and 3 pixels after.  So the largest Y mv on a border would
-         *  then be 16 - 3.  The UV blocks are half the size of the Y and
-         *  therefore only extended by 8.  The largest mv that a UV block
-         *  can support is 8 - 3.  A UV mv is half of a Y mv.
-         *  (16 - 3) >> 1 == 6 which is greater than 8 - 3.
-         * To keep the mv in play for both Y and UV planes the max that it
-         *  can be on a border is therefore 16 - 5.
-         */
-        cpi->mb.mv_row_min = -((mb_row * 16) + (16 - 5));
-        cpi->mb.mv_row_max = ((cpi->common.mb_rows - 1 - mb_row) * 16)
-                                + (16 - 5);
+    /* Source frames are extended to 16 pixels.  This is different than
+     *  L/A/G reference frames that have a border of 32 (VP8BORDERINPIXELS)
+     * A 6 tap filter is used for motion search.  This requires 2 pixels
+     *  before and 3 pixels after.  So the largest Y mv on a border would
+     *  then be 16 - 3.  The UV blocks are half the size of the Y and
+     *  therefore only extended by 8.  The largest mv that a UV block
+     *  can support is 8 - 3.  A UV mv is half of a Y mv.
+     *  (16 - 3) >> 1 == 6 which is greater than 8 - 3.
+     * To keep the mv in play for both Y and UV planes the max that it
+     *  can be on a border is therefore 16 - 5.
+     */
+    cpi->mb.mv_row_min = -((mb_row * 16) + (16 - 5));
+    cpi->mb.mv_row_max = ((cpi->common.mb_rows - 1 - mb_row) * 16) + (16 - 5);
 #endif
 
-        for (mb_col = 0; mb_col < mb_cols; mb_col++)
-        {
-            int i, j, k;
-            int stride;
+    for (mb_col = 0; mb_col < mb_cols; mb_col++) {
+      int i, j, k;
+      int stride;
 
-            memset(accumulator, 0, 384*sizeof(unsigned int));
-            memset(count, 0, 384*sizeof(unsigned short));
+      memset(accumulator, 0, 384 * sizeof(unsigned int));
+      memset(count, 0, 384 * sizeof(unsigned short));
 
 #if ALT_REF_MC_ENABLED
-            cpi->mb.mv_col_min = -((mb_col * 16) + (16 - 5));
-            cpi->mb.mv_col_max = ((cpi->common.mb_cols - 1 - mb_col) * 16)
-                                    + (16 - 5);
+      cpi->mb.mv_col_min = -((mb_col * 16) + (16 - 5));
+      cpi->mb.mv_col_max = ((cpi->common.mb_cols - 1 - mb_col) * 16) + (16 - 5);
 #endif
 
-            for (frame = 0; frame < frame_count; frame++)
-            {
-                if (cpi->frames[frame] == NULL)
-                    continue;
+      for (frame = 0; frame < frame_count; frame++) {
+        if (cpi->frames[frame] == NULL) continue;
 
-                mbd->block[0].bmi.mv.as_mv.row = 0;
-                mbd->block[0].bmi.mv.as_mv.col = 0;
+        mbd->block[0].bmi.mv.as_mv.row = 0;
+        mbd->block[0].bmi.mv.as_mv.col = 0;
 
-                if (frame == alt_ref_index)
-                {
-                    filter_weight = 2;
-                }
-                else
-                {
-                    int err = 0;
+        if (frame == alt_ref_index) {
+          filter_weight = 2;
+        } else {
+          int err = 0;
 #if ALT_REF_MC_ENABLED
-#define THRESH_LOW   10000
-#define THRESH_HIGH  20000
-                    /* Find best match in this frame by MC */
-                    err = vp8_temporal_filter_find_matching_mb_c
-                              (cpi,
-                               cpi->frames[alt_ref_index],
-                               cpi->frames[frame],
-                               mb_y_offset,
-                               THRESH_LOW);
+#define THRESH_LOW 10000
+#define THRESH_HIGH 20000
+          /* Find best match in this frame by MC */
+          err = vp8_temporal_filter_find_matching_mb_c(
+              cpi, cpi->frames[alt_ref_index], cpi->frames[frame], mb_y_offset,
+              THRESH_LOW);
 #endif
-                    /* Assign higher weight to matching MB if it's error
-                     * score is lower. If not applying MC default behavior
-                     * is to weight all MBs equal.
-                     */
-                    filter_weight = err<THRESH_LOW
-                                       ? 2 : err<THRESH_HIGH ? 1 : 0;
-                }
+          /* Assign higher weight to matching MB if it's error
+           * score is lower. If not applying MC default behavior
+           * is to weight all MBs equal.
+           */
+          filter_weight = err < THRESH_LOW ? 2 : err < THRESH_HIGH ? 1 : 0;
+        }
 
-                if (filter_weight != 0)
-                {
-                    /* Construct the predictors */
-                    vp8_temporal_filter_predictors_mb_c
-                        (mbd,
-                         cpi->frames[frame]->y_buffer + mb_y_offset,
-                         cpi->frames[frame]->u_buffer + mb_uv_offset,
-                         cpi->frames[frame]->v_buffer + mb_uv_offset,
-                         cpi->frames[frame]->y_stride,
-                         mbd->block[0].bmi.mv.as_mv.row,
-                         mbd->block[0].bmi.mv.as_mv.col,
-                         predictor);
+        if (filter_weight != 0) {
+          /* Construct the predictors */
+          vp8_temporal_filter_predictors_mb_c(
+              mbd, cpi->frames[frame]->y_buffer + mb_y_offset,
+              cpi->frames[frame]->u_buffer + mb_uv_offset,
+              cpi->frames[frame]->v_buffer + mb_uv_offset,
+              cpi->frames[frame]->y_stride, mbd->block[0].bmi.mv.as_mv.row,
+              mbd->block[0].bmi.mv.as_mv.col, predictor);
 
-                    /* Apply the filter (YUV) */
-                    vp8_temporal_filter_apply
-                        (f->y_buffer + mb_y_offset,
-                         f->y_stride,
-                         predictor,
-                         16,
-                         strength,
-                         filter_weight,
-                         accumulator,
-                         count);
+          /* Apply the filter (YUV) */
+          vp8_temporal_filter_apply(f->y_buffer + mb_y_offset, f->y_stride,
+                                    predictor, 16, strength, filter_weight,
+                                    accumulator, count);
 
-                    vp8_temporal_filter_apply
-                        (f->u_buffer + mb_uv_offset,
-                         f->uv_stride,
-                         predictor + 256,
-                         8,
-                         strength,
-                         filter_weight,
-                         accumulator + 256,
-                         count + 256);
+          vp8_temporal_filter_apply(f->u_buffer + mb_uv_offset, f->uv_stride,
+                                    predictor + 256, 8, strength, filter_weight,
+                                    accumulator + 256, count + 256);
 
-                    vp8_temporal_filter_apply
-                        (f->v_buffer + mb_uv_offset,
-                         f->uv_stride,
-                         predictor + 320,
-                         8,
-                         strength,
-                         filter_weight,
-                         accumulator + 320,
-                         count + 320);
-                }
-            }
+          vp8_temporal_filter_apply(f->v_buffer + mb_uv_offset, f->uv_stride,
+                                    predictor + 320, 8, strength, filter_weight,
+                                    accumulator + 320, count + 320);
+        }
+      }
 
-            /* Normalize filter output to produce AltRef frame */
-            dst1 = cpi->alt_ref_buffer.y_buffer;
-            stride = cpi->alt_ref_buffer.y_stride;
-            byte = mb_y_offset;
-            for (i = 0,k = 0; i < 16; i++)
-            {
-                for (j = 0; j < 16; j++, k++)
-                {
-                    unsigned int pval = accumulator[k] + (count[k] >> 1);
-                    pval *= cpi->fixed_divide[count[k]];
-                    pval >>= 19;
+      /* Normalize filter output to produce AltRef frame */
+      dst1 = cpi->alt_ref_buffer.y_buffer;
+      stride = cpi->alt_ref_buffer.y_stride;
+      byte = mb_y_offset;
+      for (i = 0, k = 0; i < 16; i++) {
+        for (j = 0; j < 16; j++, k++) {
+          unsigned int pval = accumulator[k] + (count[k] >> 1);
+          pval *= cpi->fixed_divide[count[k]];
+          pval >>= 19;
 
-                    dst1[byte] = (unsigned char)pval;
+          dst1[byte] = (unsigned char)pval;
 
-                    /* move to next pixel */
-                    byte++;
-                }
+          /* move to next pixel */
+          byte++;
+        }
 
-                byte += stride - 16;
-            }
+        byte += stride - 16;
+      }
 
-            dst1 = cpi->alt_ref_buffer.u_buffer;
-            dst2 = cpi->alt_ref_buffer.v_buffer;
-            stride = cpi->alt_ref_buffer.uv_stride;
-            byte = mb_uv_offset;
-            for (i = 0,k = 256; i < 8; i++)
-            {
-                for (j = 0; j < 8; j++, k++)
-                {
-                    int m=k+64;
+      dst1 = cpi->alt_ref_buffer.u_buffer;
+      dst2 = cpi->alt_ref_buffer.v_buffer;
+      stride = cpi->alt_ref_buffer.uv_stride;
+      byte = mb_uv_offset;
+      for (i = 0, k = 256; i < 8; i++) {
+        for (j = 0; j < 8; j++, k++) {
+          int m = k + 64;
 
-                    /* U */
-                    unsigned int pval = accumulator[k] + (count[k] >> 1);
-                    pval *= cpi->fixed_divide[count[k]];
-                    pval >>= 19;
-                    dst1[byte] = (unsigned char)pval;
+          /* U */
+          unsigned int pval = accumulator[k] + (count[k] >> 1);
+          pval *= cpi->fixed_divide[count[k]];
+          pval >>= 19;
+          dst1[byte] = (unsigned char)pval;
 
-                    /* V */
-                    pval = accumulator[m] + (count[m] >> 1);
-                    pval *= cpi->fixed_divide[count[m]];
-                    pval >>= 19;
-                    dst2[byte] = (unsigned char)pval;
+          /* V */
+          pval = accumulator[m] + (count[m] >> 1);
+          pval *= cpi->fixed_divide[count[m]];
+          pval >>= 19;
+          dst2[byte] = (unsigned char)pval;
 
-                    /* move to next pixel */
-                    byte++;
-                }
-
-                byte += stride - 8;
-            }
-
-            mb_y_offset += 16;
-            mb_uv_offset += 8;
+          /* move to next pixel */
+          byte++;
         }
 
-        mb_y_offset += 16*(f->y_stride-mb_cols);
-        mb_uv_offset += 8*(f->uv_stride-mb_cols);
+        byte += stride - 8;
+      }
+
+      mb_y_offset += 16;
+      mb_uv_offset += 8;
     }
 
-    /* Restore input state */
-    mbd->pre.y_buffer = y_buffer;
-    mbd->pre.u_buffer = u_buffer;
-    mbd->pre.v_buffer = v_buffer;
+    mb_y_offset += 16 * (f->y_stride - mb_cols);
+    mb_uv_offset += 8 * (f->uv_stride - mb_cols);
+  }
+
+  /* Restore input state */
+  mbd->pre.y_buffer = y_buffer;
+  mbd->pre.u_buffer = u_buffer;
+  mbd->pre.v_buffer = v_buffer;
 }
 
-void vp8_temporal_filter_prepare_c
-(
-    VP8_COMP *cpi,
-    int distance
-)
-{
-    int frame = 0;
+void vp8_temporal_filter_prepare_c(VP8_COMP *cpi, int distance) {
+  int frame = 0;
 
-    int num_frames_backward = 0;
-    int num_frames_forward = 0;
-    int frames_to_blur_backward = 0;
-    int frames_to_blur_forward = 0;
-    int frames_to_blur = 0;
-    int start_frame = 0;
+  int num_frames_backward = 0;
+  int num_frames_forward = 0;
+  int frames_to_blur_backward = 0;
+  int frames_to_blur_forward = 0;
+  int frames_to_blur = 0;
+  int start_frame = 0;
 
-    int strength = cpi->oxcf.arnr_strength;
+  int strength = cpi->oxcf.arnr_strength;
 
-    int blur_type = cpi->oxcf.arnr_type;
+  int blur_type = cpi->oxcf.arnr_type;
 
-    int max_frames = cpi->active_arnr_frames;
+  int max_frames = cpi->active_arnr_frames;
 
-    num_frames_backward = distance;
-    num_frames_forward = vp8_lookahead_depth(cpi->lookahead)
-                         - (num_frames_backward + 1);
+  num_frames_backward = distance;
+  num_frames_forward =
+      vp8_lookahead_depth(cpi->lookahead) - (num_frames_backward + 1);
 
-    switch (blur_type)
-    {
+  switch (blur_type) {
     case 1:
-        /* Backward Blur */
+      /* Backward Blur */
 
-        frames_to_blur_backward = num_frames_backward;
+      frames_to_blur_backward = num_frames_backward;
 
-        if (frames_to_blur_backward >= max_frames)
-            frames_to_blur_backward = max_frames - 1;
+      if (frames_to_blur_backward >= max_frames)
+        frames_to_blur_backward = max_frames - 1;
 
-        frames_to_blur = frames_to_blur_backward + 1;
-        break;
+      frames_to_blur = frames_to_blur_backward + 1;
+      break;
 
     case 2:
-        /* Forward Blur */
+      /* Forward Blur */
 
-        frames_to_blur_forward = num_frames_forward;
+      frames_to_blur_forward = num_frames_forward;
 
-        if (frames_to_blur_forward >= max_frames)
-            frames_to_blur_forward = max_frames - 1;
+      if (frames_to_blur_forward >= max_frames)
+        frames_to_blur_forward = max_frames - 1;
 
-        frames_to_blur = frames_to_blur_forward + 1;
-        break;
+      frames_to_blur = frames_to_blur_forward + 1;
+      break;
 
     case 3:
     default:
-        /* Center Blur */
-        frames_to_blur_forward = num_frames_forward;
-        frames_to_blur_backward = num_frames_backward;
+      /* Center Blur */
+      frames_to_blur_forward = num_frames_forward;
+      frames_to_blur_backward = num_frames_backward;
 
-        if (frames_to_blur_forward > frames_to_blur_backward)
-            frames_to_blur_forward = frames_to_blur_backward;
+      if (frames_to_blur_forward > frames_to_blur_backward)
+        frames_to_blur_forward = frames_to_blur_backward;
 
-        if (frames_to_blur_backward > frames_to_blur_forward)
-            frames_to_blur_backward = frames_to_blur_forward;
+      if (frames_to_blur_backward > frames_to_blur_forward)
+        frames_to_blur_backward = frames_to_blur_forward;
 
-        /* When max_frames is even we have 1 more frame backward than forward */
-        if (frames_to_blur_forward > (max_frames - 1) / 2)
-            frames_to_blur_forward = ((max_frames - 1) / 2);
+      /* When max_frames is even we have 1 more frame backward than forward */
+      if (frames_to_blur_forward > (max_frames - 1) / 2)
+        frames_to_blur_forward = ((max_frames - 1) / 2);
 
-        if (frames_to_blur_backward > (max_frames / 2))
-            frames_to_blur_backward = (max_frames / 2);
+      if (frames_to_blur_backward > (max_frames / 2))
+        frames_to_blur_backward = (max_frames / 2);
 
-        frames_to_blur = frames_to_blur_backward + frames_to_blur_forward + 1;
-        break;
-    }
+      frames_to_blur = frames_to_blur_backward + frames_to_blur_forward + 1;
+      break;
+  }
 
-    start_frame = distance + frames_to_blur_forward;
+  start_frame = distance + frames_to_blur_forward;
 
-    /* Setup frame pointers, NULL indicates frame not included in filter */
-    memset(cpi->frames, 0, max_frames*sizeof(YV12_BUFFER_CONFIG *));
-    for (frame = 0; frame < frames_to_blur; frame++)
-    {
-        int which_buffer =  start_frame - frame;
-        struct lookahead_entry* buf = vp8_lookahead_peek(cpi->lookahead,
-                                                         which_buffer,
-                                                         PEEK_FORWARD);
-        cpi->frames[frames_to_blur-1-frame] = &buf->img;
-    }
+  /* Setup frame pointers, NULL indicates frame not included in filter */
+  memset(cpi->frames, 0, max_frames * sizeof(YV12_BUFFER_CONFIG *));
+  for (frame = 0; frame < frames_to_blur; frame++) {
+    int which_buffer = start_frame - frame;
+    struct lookahead_entry *buf =
+        vp8_lookahead_peek(cpi->lookahead, which_buffer, PEEK_FORWARD);
+    cpi->frames[frames_to_blur - 1 - frame] = &buf->img;
+  }
 
-    vp8_temporal_filter_iterate_c (
-        cpi,
-        frames_to_blur,
-        frames_to_blur_backward,
-        strength );
+  vp8_temporal_filter_iterate_c(cpi, frames_to_blur, frames_to_blur_backward,
+                                strength);
 }
 #endif
--- a/vp8/encoder/tokenize.c
+++ b/vp8/encoder/tokenize.c
@@ -8,7 +8,6 @@
  *  be found in the AUTHORS file in the root of the source tree.
  */
 
-
 #include <math.h>
 #include <stdio.h>
 #include <string.h>
@@ -21,16 +20,17 @@
    compressions, then generating context.c = initial stats. */
 
 #ifdef VP8_ENTROPY_STATS
-_int64 context_counters[BLOCK_TYPES] [COEF_BANDS] [PREV_COEF_CONTEXTS] [MAX_ENTROPY_TOKENS];
+_int64 context_counters[BLOCK_TYPES][COEF_BANDS][PREV_COEF_CONTEXTS]
+                       [MAX_ENTROPY_TOKENS];
 #endif
-void vp8_stuff_mb(VP8_COMP *cpi, MACROBLOCK *x, TOKENEXTRA **t) ;
+void vp8_stuff_mb(VP8_COMP *cpi, MACROBLOCK *x, TOKENEXTRA **t);
 void vp8_fix_contexts(MACROBLOCKD *x);
 
 #include "dct_value_tokens.h"
 #include "dct_value_cost.h"
 
-const TOKENVALUE *const vp8_dct_value_tokens_ptr = dct_value_tokens +
-        DCT_MAX_VALUE;
+const TOKENVALUE *const vp8_dct_value_tokens_ptr =
+    dct_value_tokens + DCT_MAX_VALUE;
 const short *const vp8_dct_value_cost_ptr = dct_value_cost + DCT_MAX_VALUE;
 
 #if 0
@@ -73,7 +73,8 @@
             t[i].Extra = eb;
         }
 
-        // initialize the cost for extra bits for all possible coefficient value.
+        // initialize the cost for extra bits for all possible coefficient
+value.
         {
             int cost = 0;
             const vp8_extra_bit_struct *p = vp8_extra_bits + t[i].Token;
@@ -84,7 +85,8 @@
                 const int Length = p->Len;
 
                 if (Length)
-                    cost += vp8_treed_cost(p->tree, p->prob, extra >> 1, Length);
+                    cost += vp8_treed_cost(p->tree, p->prob, extra >> 1,
+Length);
 
                 cost += vp8_cost_bit(vp8_prob_half, extra & 1); // sign
                 dct_value_cost[i + DCT_MAX_VALUE] = cost;
@@ -100,509 +102,435 @@
 }
 */
 
-static void tokenize2nd_order_b
-(
-    MACROBLOCK *x,
-    TOKENEXTRA **tp,
-    VP8_COMP *cpi
-)
-{
-    MACROBLOCKD *xd = &x->e_mbd;
-    int pt;             /* near block/prev token context index */
-    int c;              /* start at DC */
-    TOKENEXTRA *t = *tp;/* store tokens starting here */
-    const BLOCKD *b;
-    const short *qcoeff_ptr;
-    ENTROPY_CONTEXT * a;
-    ENTROPY_CONTEXT * l;
-    int band, rc, v, token;
-    int eob;
+static void tokenize2nd_order_b(MACROBLOCK *x, TOKENEXTRA **tp, VP8_COMP *cpi) {
+  MACROBLOCKD *xd = &x->e_mbd;
+  int pt;              /* near block/prev token context index */
+  int c;               /* start at DC */
+  TOKENEXTRA *t = *tp; /* store tokens starting here */
+  const BLOCKD *b;
+  const short *qcoeff_ptr;
+  ENTROPY_CONTEXT *a;
+  ENTROPY_CONTEXT *l;
+  int band, rc, v, token;
+  int eob;
 
-    b = xd->block + 24;
-    qcoeff_ptr = b->qcoeff;
-    a = (ENTROPY_CONTEXT *)xd->above_context + 8;
-    l = (ENTROPY_CONTEXT *)xd->left_context + 8;
-    eob = xd->eobs[24];
-    VP8_COMBINEENTROPYCONTEXTS(pt, *a, *l);
+  b = xd->block + 24;
+  qcoeff_ptr = b->qcoeff;
+  a = (ENTROPY_CONTEXT *)xd->above_context + 8;
+  l = (ENTROPY_CONTEXT *)xd->left_context + 8;
+  eob = xd->eobs[24];
+  VP8_COMBINEENTROPYCONTEXTS(pt, *a, *l);
 
-    if(!eob)
-    {
-        /* c = band for this case */
-        t->Token = DCT_EOB_TOKEN;
-        t->context_tree = cpi->common.fc.coef_probs [1] [0] [pt];
-        t->skip_eob_node = 0;
-
-        ++x->coef_counts       [1] [0] [pt] [DCT_EOB_TOKEN];
-        t++;
-        *tp = t;
-        *a = *l = 0;
-        return;
-    }
-
-    v = qcoeff_ptr[0];
-    t->Extra = vp8_dct_value_tokens_ptr[v].Extra;
-    token    = vp8_dct_value_tokens_ptr[v].Token;
-    t->Token = token;
-
-    t->context_tree = cpi->common.fc.coef_probs [1] [0] [pt];
+  if (!eob) {
+    /* c = band for this case */
+    t->Token = DCT_EOB_TOKEN;
+    t->context_tree = cpi->common.fc.coef_probs[1][0][pt];
     t->skip_eob_node = 0;
-    ++x->coef_counts       [1] [0] [pt] [token];
-    pt = vp8_prev_token_class[token];
+
+    ++x->coef_counts[1][0][pt][DCT_EOB_TOKEN];
     t++;
-    c = 1;
+    *tp = t;
+    *a = *l = 0;
+    return;
+  }
 
-    for (; c < eob; c++)
-    {
-        rc = vp8_default_zig_zag1d[c];
-        band = vp8_coef_bands[c];
-        v = qcoeff_ptr[rc];
+  v = qcoeff_ptr[0];
+  t->Extra = vp8_dct_value_tokens_ptr[v].Extra;
+  token = vp8_dct_value_tokens_ptr[v].Token;
+  t->Token = token;
 
-        t->Extra = vp8_dct_value_tokens_ptr[v].Extra;
-        token    = vp8_dct_value_tokens_ptr[v].Token;
+  t->context_tree = cpi->common.fc.coef_probs[1][0][pt];
+  t->skip_eob_node = 0;
+  ++x->coef_counts[1][0][pt][token];
+  pt = vp8_prev_token_class[token];
+  t++;
+  c = 1;
 
-        t->Token = token;
-        t->context_tree = cpi->common.fc.coef_probs [1] [band] [pt];
+  for (; c < eob; c++) {
+    rc = vp8_default_zig_zag1d[c];
+    band = vp8_coef_bands[c];
+    v = qcoeff_ptr[rc];
 
-        t->skip_eob_node = ((pt == 0));
+    t->Extra = vp8_dct_value_tokens_ptr[v].Extra;
+    token = vp8_dct_value_tokens_ptr[v].Token;
 
-        ++x->coef_counts       [1] [band] [pt] [token];
+    t->Token = token;
+    t->context_tree = cpi->common.fc.coef_probs[1][band][pt];
 
-        pt = vp8_prev_token_class[token];
-        t++;
-    }
-    if (c < 16)
-    {
-        band = vp8_coef_bands[c];
-        t->Token = DCT_EOB_TOKEN;
-        t->context_tree = cpi->common.fc.coef_probs [1] [band] [pt];
+    t->skip_eob_node = ((pt == 0));
 
-        t->skip_eob_node = 0;
+    ++x->coef_counts[1][band][pt][token];
 
-        ++x->coef_counts       [1] [band] [pt] [DCT_EOB_TOKEN];
+    pt = vp8_prev_token_class[token];
+    t++;
+  }
+  if (c < 16) {
+    band = vp8_coef_bands[c];
+    t->Token = DCT_EOB_TOKEN;
+    t->context_tree = cpi->common.fc.coef_probs[1][band][pt];
 
-        t++;
-    }
+    t->skip_eob_node = 0;
 
-    *tp = t;
-    *a = *l = 1;
+    ++x->coef_counts[1][band][pt][DCT_EOB_TOKEN];
 
+    t++;
+  }
+
+  *tp = t;
+  *a = *l = 1;
 }
 
-static void tokenize1st_order_b
-(
-    MACROBLOCK *x,
-    TOKENEXTRA **tp,
-    int type,           /* which plane: 0=Y no DC, 1=Y2, 2=UV, 3=Y with DC */
-    VP8_COMP *cpi
-)
-{
-    MACROBLOCKD *xd = &x->e_mbd;
-    unsigned int block;
-    const BLOCKD *b;
-    int pt;             /* near block/prev token context index */
-    int c;
-    int token;
-    TOKENEXTRA *t = *tp;/* store tokens starting here */
-    const short *qcoeff_ptr;
-    ENTROPY_CONTEXT * a;
-    ENTROPY_CONTEXT * l;
-    int band, rc, v;
-    int tmp1, tmp2;
+static void tokenize1st_order_b(
+    MACROBLOCK *x, TOKENEXTRA **tp,
+    int type, /* which plane: 0=Y no DC, 1=Y2, 2=UV, 3=Y with DC */
+    VP8_COMP *cpi) {
+  MACROBLOCKD *xd = &x->e_mbd;
+  unsigned int block;
+  const BLOCKD *b;
+  int pt; /* near block/prev token context index */
+  int c;
+  int token;
+  TOKENEXTRA *t = *tp; /* store tokens starting here */
+  const short *qcoeff_ptr;
+  ENTROPY_CONTEXT *a;
+  ENTROPY_CONTEXT *l;
+  int band, rc, v;
+  int tmp1, tmp2;
 
-    b = xd->block;
-    /* Luma */
-    for (block = 0; block < 16; block++, b++)
-    {
-        const int eob = *b->eob;
-        tmp1 = vp8_block2above[block];
-        tmp2 = vp8_block2left[block];
-        qcoeff_ptr = b->qcoeff;
-        a = (ENTROPY_CONTEXT *)xd->above_context + tmp1;
-        l = (ENTROPY_CONTEXT *)xd->left_context + tmp2;
+  b = xd->block;
+  /* Luma */
+  for (block = 0; block < 16; block++, b++) {
+    const int eob = *b->eob;
+    tmp1 = vp8_block2above[block];
+    tmp2 = vp8_block2left[block];
+    qcoeff_ptr = b->qcoeff;
+    a = (ENTROPY_CONTEXT *)xd->above_context + tmp1;
+    l = (ENTROPY_CONTEXT *)xd->left_context + tmp2;
 
-        VP8_COMBINEENTROPYCONTEXTS(pt, *a, *l);
+    VP8_COMBINEENTROPYCONTEXTS(pt, *a, *l);
 
-        c = type ? 0 : 1;
+    c = type ? 0 : 1;
 
-        if(c >= eob)
-        {
-            /* c = band for this case */
-            t->Token = DCT_EOB_TOKEN;
-            t->context_tree = cpi->common.fc.coef_probs [type] [c] [pt];
-            t->skip_eob_node = 0;
+    if (c >= eob) {
+      /* c = band for this case */
+      t->Token = DCT_EOB_TOKEN;
+      t->context_tree = cpi->common.fc.coef_probs[type][c][pt];
+      t->skip_eob_node = 0;
 
-            ++x->coef_counts       [type] [c] [pt] [DCT_EOB_TOKEN];
-            t++;
-            *tp = t;
-            *a = *l = 0;
-            continue;
-        }
+      ++x->coef_counts[type][c][pt][DCT_EOB_TOKEN];
+      t++;
+      *tp = t;
+      *a = *l = 0;
+      continue;
+    }
 
-        v = qcoeff_ptr[c];
+    v = qcoeff_ptr[c];
 
-        t->Extra = vp8_dct_value_tokens_ptr[v].Extra;
-        token    = vp8_dct_value_tokens_ptr[v].Token;
-        t->Token = token;
+    t->Extra = vp8_dct_value_tokens_ptr[v].Extra;
+    token = vp8_dct_value_tokens_ptr[v].Token;
+    t->Token = token;
 
-        t->context_tree = cpi->common.fc.coef_probs [type] [c] [pt];
-        t->skip_eob_node = 0;
-        ++x->coef_counts       [type] [c] [pt] [token];
-        pt = vp8_prev_token_class[token];
-        t++;
-        c++;
+    t->context_tree = cpi->common.fc.coef_probs[type][c][pt];
+    t->skip_eob_node = 0;
+    ++x->coef_counts[type][c][pt][token];
+    pt = vp8_prev_token_class[token];
+    t++;
+    c++;
 
-        assert(eob <= 16);
-        for (; c < eob; c++)
-        {
-            rc = vp8_default_zig_zag1d[c];
-            band = vp8_coef_bands[c];
-            v = qcoeff_ptr[rc];
+    assert(eob <= 16);
+    for (; c < eob; c++) {
+      rc = vp8_default_zig_zag1d[c];
+      band = vp8_coef_bands[c];
+      v = qcoeff_ptr[rc];
 
-            t->Extra = vp8_dct_value_tokens_ptr[v].Extra;
-            token    = vp8_dct_value_tokens_ptr[v].Token;
+      t->Extra = vp8_dct_value_tokens_ptr[v].Extra;
+      token = vp8_dct_value_tokens_ptr[v].Token;
 
-            t->Token = token;
-            t->context_tree = cpi->common.fc.coef_probs [type] [band] [pt];
+      t->Token = token;
+      t->context_tree = cpi->common.fc.coef_probs[type][band][pt];
 
-            t->skip_eob_node = (pt == 0);
-            ++x->coef_counts       [type] [band] [pt] [token];
+      t->skip_eob_node = (pt == 0);
+      ++x->coef_counts[type][band][pt][token];
 
-            pt = vp8_prev_token_class[token];
-            t++;
-        }
-        if (c < 16)
-        {
-            band = vp8_coef_bands[c];
-            t->Token = DCT_EOB_TOKEN;
-            t->context_tree = cpi->common.fc.coef_probs [type] [band] [pt];
+      pt = vp8_prev_token_class[token];
+      t++;
+    }
+    if (c < 16) {
+      band = vp8_coef_bands[c];
+      t->Token = DCT_EOB_TOKEN;
+      t->context_tree = cpi->common.fc.coef_probs[type][band][pt];
 
-            t->skip_eob_node = 0;
-            ++x->coef_counts       [type] [band] [pt] [DCT_EOB_TOKEN];
+      t->skip_eob_node = 0;
+      ++x->coef_counts[type][band][pt][DCT_EOB_TOKEN];
 
-            t++;
-        }
-        *tp = t;
-        *a = *l = 1;
+      t++;
     }
+    *tp = t;
+    *a = *l = 1;
+  }
 
-    /* Chroma */
-    for (block = 16; block < 24; block++, b++)
-    {
-        const int eob = *b->eob;
-        tmp1 = vp8_block2above[block];
-        tmp2 = vp8_block2left[block];
-        qcoeff_ptr = b->qcoeff;
-        a = (ENTROPY_CONTEXT *)xd->above_context + tmp1;
-        l = (ENTROPY_CONTEXT *)xd->left_context + tmp2;
+  /* Chroma */
+  for (block = 16; block < 24; block++, b++) {
+    const int eob = *b->eob;
+    tmp1 = vp8_block2above[block];
+    tmp2 = vp8_block2left[block];
+    qcoeff_ptr = b->qcoeff;
+    a = (ENTROPY_CONTEXT *)xd->above_context + tmp1;
+    l = (ENTROPY_CONTEXT *)xd->left_context + tmp2;
 
-        VP8_COMBINEENTROPYCONTEXTS(pt, *a, *l);
+    VP8_COMBINEENTROPYCONTEXTS(pt, *a, *l);
 
-        if(!eob)
-        {
-            /* c = band for this case */
-            t->Token = DCT_EOB_TOKEN;
-            t->context_tree = cpi->common.fc.coef_probs [2] [0] [pt];
-            t->skip_eob_node = 0;
+    if (!eob) {
+      /* c = band for this case */
+      t->Token = DCT_EOB_TOKEN;
+      t->context_tree = cpi->common.fc.coef_probs[2][0][pt];
+      t->skip_eob_node = 0;
 
-            ++x->coef_counts       [2] [0] [pt] [DCT_EOB_TOKEN];
-            t++;
-            *tp = t;
-            *a = *l = 0;
-            continue;
-        }
+      ++x->coef_counts[2][0][pt][DCT_EOB_TOKEN];
+      t++;
+      *tp = t;
+      *a = *l = 0;
+      continue;
+    }
 
-        v = qcoeff_ptr[0];
+    v = qcoeff_ptr[0];
 
-        t->Extra = vp8_dct_value_tokens_ptr[v].Extra;
-        token    = vp8_dct_value_tokens_ptr[v].Token;
-        t->Token = token;
+    t->Extra = vp8_dct_value_tokens_ptr[v].Extra;
+    token = vp8_dct_value_tokens_ptr[v].Token;
+    t->Token = token;
 
-        t->context_tree = cpi->common.fc.coef_probs [2] [0] [pt];
-        t->skip_eob_node = 0;
-        ++x->coef_counts       [2] [0] [pt] [token];
-        pt = vp8_prev_token_class[token];
-        t++;
-        c = 1;
+    t->context_tree = cpi->common.fc.coef_probs[2][0][pt];
+    t->skip_eob_node = 0;
+    ++x->coef_counts[2][0][pt][token];
+    pt = vp8_prev_token_class[token];
+    t++;
+    c = 1;
 
-        assert(eob <= 16);
-        for (; c < eob; c++)
-        {
-            rc = vp8_default_zig_zag1d[c];
-            band = vp8_coef_bands[c];
-            v = qcoeff_ptr[rc];
+    assert(eob <= 16);
+    for (; c < eob; c++) {
+      rc = vp8_default_zig_zag1d[c];
+      band = vp8_coef_bands[c];
+      v = qcoeff_ptr[rc];
 
-            t->Extra = vp8_dct_value_tokens_ptr[v].Extra;
-            token    = vp8_dct_value_tokens_ptr[v].Token;
+      t->Extra = vp8_dct_value_tokens_ptr[v].Extra;
+      token = vp8_dct_value_tokens_ptr[v].Token;
 
-            t->Token = token;
-            t->context_tree = cpi->common.fc.coef_probs [2] [band] [pt];
+      t->Token = token;
+      t->context_tree = cpi->common.fc.coef_probs[2][band][pt];
 
-            t->skip_eob_node = (pt == 0);
+      t->skip_eob_node = (pt == 0);
 
-            ++x->coef_counts       [2] [band] [pt] [token];
+      ++x->coef_counts[2][band][pt][token];
 
-            pt = vp8_prev_token_class[token];
-            t++;
-        }
-        if (c < 16)
-        {
-            band = vp8_coef_bands[c];
-            t->Token = DCT_EOB_TOKEN;
-            t->context_tree = cpi->common.fc.coef_probs [2] [band] [pt];
+      pt = vp8_prev_token_class[token];
+      t++;
+    }
+    if (c < 16) {
+      band = vp8_coef_bands[c];
+      t->Token = DCT_EOB_TOKEN;
+      t->context_tree = cpi->common.fc.coef_probs[2][band][pt];
 
-            t->skip_eob_node = 0;
+      t->skip_eob_node = 0;
 
-            ++x->coef_counts       [2] [band] [pt] [DCT_EOB_TOKEN];
+      ++x->coef_counts[2][band][pt][DCT_EOB_TOKEN];
 
-            t++;
-        }
-        *tp = t;
-        *a = *l = 1;
+      t++;
     }
+    *tp = t;
+    *a = *l = 1;
+  }
 }
 
+static int mb_is_skippable(MACROBLOCKD *x, int has_y2_block) {
+  int skip = 1;
+  int i = 0;
 
-static int mb_is_skippable(MACROBLOCKD *x, int has_y2_block)
-{
-    int skip = 1;
-    int i = 0;
+  if (has_y2_block) {
+    for (i = 0; i < 16; i++) skip &= (x->eobs[i] < 2);
+  }
 
-    if (has_y2_block)
-    {
-        for (i = 0; i < 16; i++)
-            skip &= (x->eobs[i] < 2);
-    }
+  for (; i < 24 + has_y2_block; i++) skip &= (!x->eobs[i]);
 
-    for (; i < 24 + has_y2_block; i++)
-        skip &= (!x->eobs[i]);
-
-    return skip;
+  return skip;
 }
 
+void vp8_tokenize_mb(VP8_COMP *cpi, MACROBLOCK *x, TOKENEXTRA **t) {
+  MACROBLOCKD *xd = &x->e_mbd;
+  int plane_type;
+  int has_y2_block;
 
-void vp8_tokenize_mb(VP8_COMP *cpi, MACROBLOCK *x, TOKENEXTRA **t)
-{
-    MACROBLOCKD *xd = &x->e_mbd;
-    int plane_type;
-    int has_y2_block;
+  has_y2_block = (xd->mode_info_context->mbmi.mode != B_PRED &&
+                  xd->mode_info_context->mbmi.mode != SPLITMV);
 
-    has_y2_block = (xd->mode_info_context->mbmi.mode != B_PRED
-                    && xd->mode_info_context->mbmi.mode != SPLITMV);
-
-    xd->mode_info_context->mbmi.mb_skip_coeff =
-        mb_is_skippable(xd, has_y2_block);
-    if (xd->mode_info_context->mbmi.mb_skip_coeff)
-    {
-        if (!cpi->common.mb_no_coeff_skip)
-        {
-            vp8_stuff_mb(cpi, x, t);
-        }
-        else
-        {
-            vp8_fix_contexts(xd);
-            x->skip_true_count++;
-        }
-
-        return;
+  xd->mode_info_context->mbmi.mb_skip_coeff = mb_is_skippable(xd, has_y2_block);
+  if (xd->mode_info_context->mbmi.mb_skip_coeff) {
+    if (!cpi->common.mb_no_coeff_skip) {
+      vp8_stuff_mb(cpi, x, t);
+    } else {
+      vp8_fix_contexts(xd);
+      x->skip_true_count++;
     }
 
-    plane_type = 3;
-    if(has_y2_block)
-    {
-        tokenize2nd_order_b(x, t, cpi);
-        plane_type = 0;
-    }
+    return;
+  }
 
-    tokenize1st_order_b(x, t, plane_type, cpi);
+  plane_type = 3;
+  if (has_y2_block) {
+    tokenize2nd_order_b(x, t, cpi);
+    plane_type = 0;
+  }
+
+  tokenize1st_order_b(x, t, plane_type, cpi);
 }
 
-
 #ifdef VP8_ENTROPY_STATS
 
-void init_context_counters(void)
-{
-    memset(context_counters, 0, sizeof(context_counters));
+void init_context_counters(void) {
+  memset(context_counters, 0, sizeof(context_counters));
 }
 
-void print_context_counters()
-{
+void print_context_counters() {
+  int type, band, pt, t;
 
-    int type, band, pt, t;
+  FILE *const f = fopen("context.c", "w");
 
-    FILE *const f = fopen("context.c", "w");
+  fprintf(f, "#include \"entropy.h\"\n");
 
-    fprintf(f, "#include \"entropy.h\"\n");
+  fprintf(f, "\n/* *** GENERATED FILE: DO NOT EDIT *** */\n\n");
 
-    fprintf(f, "\n/* *** GENERATED FILE: DO NOT EDIT *** */\n\n");
+  fprintf(f,
+          "int Contexts[BLOCK_TYPES] [COEF_BANDS] [PREV_COEF_CONTEXTS] "
+          "[MAX_ENTROPY_TOKENS];\n\n");
 
-    fprintf(f, "int Contexts[BLOCK_TYPES] [COEF_BANDS] [PREV_COEF_CONTEXTS] [MAX_ENTROPY_TOKENS];\n\n");
+  fprintf(f,
+          "const int default_contexts[BLOCK_TYPES] [COEF_BANDS] "
+          "[PREV_COEF_CONTEXTS] [MAX_ENTROPY_TOKENS] = {");
 
-    fprintf(f, "const int default_contexts[BLOCK_TYPES] [COEF_BANDS] [PREV_COEF_CONTEXTS] [MAX_ENTROPY_TOKENS] = {");
+#define Comma(X) (X ? "," : "")
 
-# define Comma( X) (X? ",":"")
+  type = 0;
 
-    type = 0;
+  do {
+    fprintf(f, "%s\n  { /* block Type %d */", Comma(type), type);
 
-    do
-    {
-        fprintf(f, "%s\n  { /* block Type %d */", Comma(type), type);
+    band = 0;
 
-        band = 0;
+    do {
+      fprintf(f, "%s\n    { /* Coeff Band %d */", Comma(band), band);
 
-        do
-        {
-            fprintf(f, "%s\n    { /* Coeff Band %d */", Comma(band), band);
+      pt = 0;
 
-            pt = 0;
+      do {
+        fprintf(f, "%s\n      {", Comma(pt));
 
-            do
-            {
-                fprintf(f, "%s\n      {", Comma(pt));
+        t = 0;
 
-                t = 0;
+        do {
+          const _int64 x = context_counters[type][band][pt][t];
+          const int y = (int)x;
 
-                do
-                {
-                    const _int64 x = context_counters [type] [band] [pt] [t];
-                    const int y = (int) x;
+          assert(x == (_int64)y); /* no overflow handling yet */
+          fprintf(f, "%s %d", Comma(t), y);
 
-                    assert(x == (_int64) y);  /* no overflow handling yet */
-                    fprintf(f, "%s %d", Comma(t), y);
+        } while (++t < MAX_ENTROPY_TOKENS);
 
-                }
-                while (++t < MAX_ENTROPY_TOKENS);
+        fprintf(f, "}");
+      } while (++pt < PREV_COEF_CONTEXTS);
 
-                fprintf(f, "}");
-            }
-            while (++pt < PREV_COEF_CONTEXTS);
+      fprintf(f, "\n    }");
 
-            fprintf(f, "\n    }");
+    } while (++band < COEF_BANDS);
 
-        }
-        while (++band < COEF_BANDS);
+    fprintf(f, "\n  }");
+  } while (++type < BLOCK_TYPES);
 
-        fprintf(f, "\n  }");
-    }
-    while (++type < BLOCK_TYPES);
-
-    fprintf(f, "\n};\n");
-    fclose(f);
+  fprintf(f, "\n};\n");
+  fclose(f);
 }
 #endif
 
+static void stuff2nd_order_b(TOKENEXTRA **tp, ENTROPY_CONTEXT *a,
+                             ENTROPY_CONTEXT *l, VP8_COMP *cpi, MACROBLOCK *x) {
+  int pt;              /* near block/prev token context index */
+  TOKENEXTRA *t = *tp; /* store tokens starting here */
+  VP8_COMBINEENTROPYCONTEXTS(pt, *a, *l);
 
-static void stuff2nd_order_b
-(
-    TOKENEXTRA **tp,
-    ENTROPY_CONTEXT *a,
-    ENTROPY_CONTEXT *l,
-    VP8_COMP *cpi,
-    MACROBLOCK *x
-)
-{
-    int pt; /* near block/prev token context index */
-    TOKENEXTRA *t = *tp;        /* store tokens starting here */
-    VP8_COMBINEENTROPYCONTEXTS(pt, *a, *l);
+  t->Token = DCT_EOB_TOKEN;
+  t->context_tree = cpi->common.fc.coef_probs[1][0][pt];
+  t->skip_eob_node = 0;
+  ++x->coef_counts[1][0][pt][DCT_EOB_TOKEN];
+  ++t;
 
-    t->Token = DCT_EOB_TOKEN;
-    t->context_tree = cpi->common.fc.coef_probs [1] [0] [pt];
-    t->skip_eob_node = 0;
-    ++x->coef_counts       [1] [0] [pt] [DCT_EOB_TOKEN];
-    ++t;
-
-    *tp = t;
-    pt = 0;
-    *a = *l = pt;
+  *tp = t;
+  pt = 0;
+  *a = *l = pt;
 }
 
-static void stuff1st_order_b
-(
-    TOKENEXTRA **tp,
-    ENTROPY_CONTEXT *a,
-    ENTROPY_CONTEXT *l,
-    int type,
-    VP8_COMP *cpi,
-    MACROBLOCK *x
-)
-{
-    int pt; /* near block/prev token context index */
-    int band;
-    TOKENEXTRA *t = *tp;        /* store tokens starting here */
-    VP8_COMBINEENTROPYCONTEXTS(pt, *a, *l);
-    band = type ? 0 : 1;
-    t->Token = DCT_EOB_TOKEN;
-    t->context_tree = cpi->common.fc.coef_probs [type] [band] [pt];
-    t->skip_eob_node = 0;
-    ++x->coef_counts       [type] [band] [pt] [DCT_EOB_TOKEN];
-    ++t;
-    *tp = t;
-    pt = 0; /* 0 <-> all coeff data is zero */
-    *a = *l = pt;
+static void stuff1st_order_b(TOKENEXTRA **tp, ENTROPY_CONTEXT *a,
+                             ENTROPY_CONTEXT *l, int type, VP8_COMP *cpi,
+                             MACROBLOCK *x) {
+  int pt; /* near block/prev token context index */
+  int band;
+  TOKENEXTRA *t = *tp; /* store tokens starting here */
+  VP8_COMBINEENTROPYCONTEXTS(pt, *a, *l);
+  band = type ? 0 : 1;
+  t->Token = DCT_EOB_TOKEN;
+  t->context_tree = cpi->common.fc.coef_probs[type][band][pt];
+  t->skip_eob_node = 0;
+  ++x->coef_counts[type][band][pt][DCT_EOB_TOKEN];
+  ++t;
+  *tp = t;
+  pt = 0; /* 0 <-> all coeff data is zero */
+  *a = *l = pt;
 }
 
-static
-void stuff1st_order_buv
-(
-    TOKENEXTRA **tp,
-    ENTROPY_CONTEXT *a,
-    ENTROPY_CONTEXT *l,
-    VP8_COMP *cpi,
-    MACROBLOCK *x
-)
-{
-    int pt; /* near block/prev token context index */
-    TOKENEXTRA *t = *tp;        /* store tokens starting here */
-    VP8_COMBINEENTROPYCONTEXTS(pt, *a, *l);
+static void stuff1st_order_buv(TOKENEXTRA **tp, ENTROPY_CONTEXT *a,
+                               ENTROPY_CONTEXT *l, VP8_COMP *cpi,
+                               MACROBLOCK *x) {
+  int pt;              /* near block/prev token context index */
+  TOKENEXTRA *t = *tp; /* store tokens starting here */
+  VP8_COMBINEENTROPYCONTEXTS(pt, *a, *l);
 
-    t->Token = DCT_EOB_TOKEN;
-    t->context_tree = cpi->common.fc.coef_probs [2] [0] [pt];
-    t->skip_eob_node = 0;
-    ++x->coef_counts[2] [0] [pt] [DCT_EOB_TOKEN];
-    ++t;
-    *tp = t;
-    pt = 0; /* 0 <-> all coeff data is zero */
-    *a = *l = pt;
+  t->Token = DCT_EOB_TOKEN;
+  t->context_tree = cpi->common.fc.coef_probs[2][0][pt];
+  t->skip_eob_node = 0;
+  ++x->coef_counts[2][0][pt][DCT_EOB_TOKEN];
+  ++t;
+  *tp = t;
+  pt = 0; /* 0 <-> all coeff data is zero */
+  *a = *l = pt;
 }
 
-void vp8_stuff_mb(VP8_COMP *cpi, MACROBLOCK *x, TOKENEXTRA **t)
-{
-    MACROBLOCKD *xd = &x->e_mbd;
-    ENTROPY_CONTEXT * A = (ENTROPY_CONTEXT *)xd->above_context;
-    ENTROPY_CONTEXT * L = (ENTROPY_CONTEXT *)xd->left_context;
-    int plane_type;
-    int b;
-    plane_type = 3;
-    if((xd->mode_info_context->mbmi.mode != B_PRED
-                        && xd->mode_info_context->mbmi.mode != SPLITMV))
-    {
-        stuff2nd_order_b(t,
-                     A + vp8_block2above[24], L + vp8_block2left[24], cpi, x);
-        plane_type = 0;
-    }
+void vp8_stuff_mb(VP8_COMP *cpi, MACROBLOCK *x, TOKENEXTRA **t) {
+  MACROBLOCKD *xd = &x->e_mbd;
+  ENTROPY_CONTEXT *A = (ENTROPY_CONTEXT *)xd->above_context;
+  ENTROPY_CONTEXT *L = (ENTROPY_CONTEXT *)xd->left_context;
+  int plane_type;
+  int b;
+  plane_type = 3;
+  if ((xd->mode_info_context->mbmi.mode != B_PRED &&
+       xd->mode_info_context->mbmi.mode != SPLITMV)) {
+    stuff2nd_order_b(t, A + vp8_block2above[24], L + vp8_block2left[24], cpi,
+                     x);
+    plane_type = 0;
+  }
 
-    for (b = 0; b < 16; b++)
-        stuff1st_order_b(t,
-                         A + vp8_block2above[b],
-                         L + vp8_block2left[b], plane_type, cpi, x);
+  for (b = 0; b < 16; b++)
+    stuff1st_order_b(t, A + vp8_block2above[b], L + vp8_block2left[b],
+                     plane_type, cpi, x);
 
-    for (b = 16; b < 24; b++)
-        stuff1st_order_buv(t,
-                           A + vp8_block2above[b],
-                           L + vp8_block2left[b], cpi, x);
-
+  for (b = 16; b < 24; b++)
+    stuff1st_order_buv(t, A + vp8_block2above[b], L + vp8_block2left[b], cpi,
+                       x);
 }
-void vp8_fix_contexts(MACROBLOCKD *x)
-{
-    /* Clear entropy contexts for Y2 blocks */
-    if (x->mode_info_context->mbmi.mode != B_PRED && x->mode_info_context->mbmi.mode != SPLITMV)
-    {
-        memset(x->above_context, 0, sizeof(ENTROPY_CONTEXT_PLANES));
-        memset(x->left_context, 0, sizeof(ENTROPY_CONTEXT_PLANES));
-    }
-    else
-    {
-        memset(x->above_context, 0, sizeof(ENTROPY_CONTEXT_PLANES)-1);
-        memset(x->left_context, 0, sizeof(ENTROPY_CONTEXT_PLANES)-1);
-    }
-
+void vp8_fix_contexts(MACROBLOCKD *x) {
+  /* Clear entropy contexts for Y2 blocks */
+  if (x->mode_info_context->mbmi.mode != B_PRED &&
+      x->mode_info_context->mbmi.mode != SPLITMV) {
+    memset(x->above_context, 0, sizeof(ENTROPY_CONTEXT_PLANES));
+    memset(x->left_context, 0, sizeof(ENTROPY_CONTEXT_PLANES));
+  } else {
+    memset(x->above_context, 0, sizeof(ENTROPY_CONTEXT_PLANES) - 1);
+    memset(x->left_context, 0, sizeof(ENTROPY_CONTEXT_PLANES) - 1);
+  }
 }
--- a/vp8/encoder/tokenize.h
+++ b/vp8/encoder/tokenize.h
@@ -8,7 +8,6 @@
  *  be found in the AUTHORS file in the root of the source tree.
  */
 
-
 #ifndef VP8_ENCODER_TOKENIZE_H_
 #define VP8_ENCODER_TOKENIZE_H_
 
@@ -21,18 +20,16 @@
 
 void vp8_tokenize_initialize();
 
-typedef struct
-{
-    short Token;
-    short Extra;
+typedef struct {
+  short Token;
+  short Extra;
 } TOKENVALUE;
 
-typedef struct
-{
-    const vp8_prob *context_tree;
-    short           Extra;
-    unsigned char   Token;
-    unsigned char   skip_eob_node;
+typedef struct {
+  const vp8_prob *context_tree;
+  short Extra;
+  unsigned char Token;
+  unsigned char skip_eob_node;
 } TOKENEXTRA;
 
 int rd_cost_mby(MACROBLOCKD *);
@@ -41,7 +38,8 @@
 void init_context_counters();
 void print_context_counters();
 
-extern _int64 context_counters[BLOCK_TYPES] [COEF_BANDS] [PREV_COEF_CONTEXTS] [MAX_ENTROPY_TOKENS];
+extern _int64 context_counters[BLOCK_TYPES][COEF_BANDS][PREV_COEF_CONTEXTS]
+                              [MAX_ENTROPY_TOKENS];
 #endif
 
 extern const short *const vp8_dct_value_cost_ptr;
--- a/vp8/encoder/treewriter.c
+++ b/vp8/encoder/treewriter.c
@@ -8,36 +8,25 @@
  *  be found in the AUTHORS file in the root of the source tree.
  */
 
-
 #include "treewriter.h"
 
-static void cost(
-    int *const C,
-    vp8_tree T,
-    const vp8_prob *const P,
-    int i,
-    int c
-)
-{
-    const vp8_prob p = P [i>>1];
+static void cost(int *const C, vp8_tree T, const vp8_prob *const P, int i,
+                 int c) {
+  const vp8_prob p = P[i >> 1];
 
-    do
-    {
-        const vp8_tree_index j = T[i];
-        const int d = c + vp8_cost_bit(p, i & 1);
+  do {
+    const vp8_tree_index j = T[i];
+    const int d = c + vp8_cost_bit(p, i & 1);
 
-        if (j <= 0)
-            C[-j] = d;
-        else
-            cost(C, T, P, j, d);
-    }
-    while (++i & 1);
+    if (j <= 0)
+      C[-j] = d;
+    else
+      cost(C, T, P, j, d);
+  } while (++i & 1);
 }
-void vp8_cost_tokens(int *c, const vp8_prob *p, vp8_tree t)
-{
-    cost(c, t, p, 0, 0);
+void vp8_cost_tokens(int *c, const vp8_prob *p, vp8_tree t) {
+  cost(c, t, p, 0, 0);
 }
-void vp8_cost_tokens2(int *c, const vp8_prob *p, vp8_tree t,int start)
-{
-    cost(c, t, p, start, 0);
+void vp8_cost_tokens2(int *c, const vp8_prob *p, vp8_tree t, int start) {
+  cost(c, t, p, start, 0);
 }
--- a/vp8/encoder/treewriter.h
+++ b/vp8/encoder/treewriter.h
@@ -8,7 +8,6 @@
  *  be found in the AUTHORS file in the root of the source tree.
  */
 
-
 #ifndef VP8_ENCODER_TREEWRITER_H_
 #define VP8_ENCODER_TREEWRITER_H_
 
@@ -18,7 +17,7 @@
 #include "./vpx_config.h"
 #include "vp8/common/treecoder.h"
 
-#include "boolhuff.h"       /* for now */
+#include "boolhuff.h" /* for now */
 
 #ifdef __cplusplus
 extern "C" {
@@ -28,105 +27,75 @@
 
 #define vp8_write vp8_encode_bool
 #define vp8_write_literal vp8_encode_value
-#define vp8_write_bit( W, V) vp8_write( W, V, vp8_prob_half)
+#define vp8_write_bit(W, V) vp8_write(W, V, vp8_prob_half)
 
 #define vp8bc_write vp8bc_write_bool
 #define vp8bc_write_literal vp8bc_write_bits
-#define vp8bc_write_bit( W, V) vp8bc_write_bits( W, V, 1)
+#define vp8bc_write_bit(W, V) vp8bc_write_bits(W, V, 1)
 
-
 /* Approximate length of an encoded bool in 256ths of a bit at given prob */
 
-#define vp8_cost_zero( x) ( vp8_prob_cost[x])
-#define vp8_cost_one( x)  vp8_cost_zero( vp8_complement(x))
+#define vp8_cost_zero(x) (vp8_prob_cost[x])
+#define vp8_cost_one(x) vp8_cost_zero(vp8_complement(x))
 
-#define vp8_cost_bit( x, b) vp8_cost_zero( (b)?  vp8_complement(x) : (x) )
+#define vp8_cost_bit(x, b) vp8_cost_zero((b) ? vp8_complement(x) : (x))
 
 /* VP8BC version is scaled by 2^20 rather than 2^8; see bool_coder.h */
 
-
 /* Both of these return bits, not scaled bits. */
 
-static INLINE unsigned int vp8_cost_branch(const unsigned int ct[2], vp8_prob p)
-{
-    /* Imitate existing calculation */
+static INLINE unsigned int vp8_cost_branch(const unsigned int ct[2],
+                                           vp8_prob p) {
+  /* Imitate existing calculation */
 
-    return ((ct[0] * vp8_cost_zero(p))
-            + (ct[1] * vp8_cost_one(p))) >> 8;
+  return ((ct[0] * vp8_cost_zero(p)) + (ct[1] * vp8_cost_one(p))) >> 8;
 }
 
 /* Small functions to write explicit values and tokens, as well as
    estimate their lengths. */
 
-static void vp8_treed_write
-(
-    vp8_writer *const w,
-    vp8_tree t,
-    const vp8_prob *const p,
-    int v,
-    int n               /* number of bits in v, assumed nonzero */
-)
-{
-    vp8_tree_index i = 0;
+static void vp8_treed_write(vp8_writer *const w, vp8_tree t,
+                            const vp8_prob *const p, int v,
+                            int n /* number of bits in v, assumed nonzero */
+                            ) {
+  vp8_tree_index i = 0;
 
-    do
-    {
-        const int b = (v >> --n) & 1;
-        vp8_write(w, b, p[i>>1]);
-        i = t[i+b];
-    }
-    while (n);
+  do {
+    const int b = (v >> --n) & 1;
+    vp8_write(w, b, p[i >> 1]);
+    i = t[i + b];
+  } while (n);
 }
-static INLINE void vp8_write_token
-(
-    vp8_writer *const w,
-    vp8_tree t,
-    const vp8_prob *const p,
-    vp8_token *const x
-)
-{
-    vp8_treed_write(w, t, p, x->value, x->Len);
+static INLINE void vp8_write_token(vp8_writer *const w, vp8_tree t,
+                                   const vp8_prob *const p,
+                                   vp8_token *const x) {
+  vp8_treed_write(w, t, p, x->value, x->Len);
 }
 
-static int vp8_treed_cost(
-    vp8_tree t,
-    const vp8_prob *const p,
-    int v,
-    int n               /* number of bits in v, assumed nonzero */
-)
-{
-    int c = 0;
-    vp8_tree_index i = 0;
+static int vp8_treed_cost(vp8_tree t, const vp8_prob *const p, int v,
+                          int n /* number of bits in v, assumed nonzero */
+                          ) {
+  int c = 0;
+  vp8_tree_index i = 0;
 
-    do
-    {
-        const int b = (v >> --n) & 1;
-        c += vp8_cost_bit(p[i>>1], b);
-        i = t[i+b];
-    }
-    while (n);
+  do {
+    const int b = (v >> --n) & 1;
+    c += vp8_cost_bit(p[i >> 1], b);
+    i = t[i + b];
+  } while (n);
 
-    return c;
+  return c;
 }
-static INLINE int vp8_cost_token
-(
-    vp8_tree t,
-    const vp8_prob *const p,
-    vp8_token *const x
-)
-{
-    return vp8_treed_cost(t, p, x->value, x->Len);
+static INLINE int vp8_cost_token(vp8_tree t, const vp8_prob *const p,
+                                 vp8_token *const x) {
+  return vp8_treed_cost(t, p, x->value, x->Len);
 }
 
 /* Fill array of costs for all possible token values. */
 
-void vp8_cost_tokens(
-    int *Costs, const vp8_prob *, vp8_tree
-);
+void vp8_cost_tokens(int *Costs, const vp8_prob *, vp8_tree);
 
-void vp8_cost_tokens2(
-    int *Costs, const vp8_prob *, vp8_tree, int
-);
+void vp8_cost_tokens2(int *Costs, const vp8_prob *, vp8_tree, int);
 
 #ifdef __cplusplus
 }  // extern "C"
--- a/vp8/encoder/vp8_quantize.c
+++ b/vp8/encoder/vp8_quantize.c
@@ -8,7 +8,6 @@
  *  be found in the AUTHORS file in the root of the source tree.
  */
 
-
 #include <math.h>
 #include "vpx_mem/vpx_mem.h"
 
@@ -16,568 +15,475 @@
 #include "vp8/encoder/quantize.h"
 #include "vp8/common/quant_common.h"
 
-void vp8_fast_quantize_b_c(BLOCK *b, BLOCKD *d)
-{
-    int i, rc, eob;
-    int x, y, z, sz;
-    short *coeff_ptr   = b->coeff;
-    short *round_ptr   = b->round;
-    short *quant_ptr   = b->quant_fast;
-    short *qcoeff_ptr  = d->qcoeff;
-    short *dqcoeff_ptr = d->dqcoeff;
-    short *dequant_ptr = d->dequant;
+void vp8_fast_quantize_b_c(BLOCK *b, BLOCKD *d) {
+  int i, rc, eob;
+  int x, y, z, sz;
+  short *coeff_ptr = b->coeff;
+  short *round_ptr = b->round;
+  short *quant_ptr = b->quant_fast;
+  short *qcoeff_ptr = d->qcoeff;
+  short *dqcoeff_ptr = d->dqcoeff;
+  short *dequant_ptr = d->dequant;
 
-    eob = -1;
-    for (i = 0; i < 16; i++)
-    {
-        rc   = vp8_default_zig_zag1d[i];
-        z    = coeff_ptr[rc];
+  eob = -1;
+  for (i = 0; i < 16; i++) {
+    rc = vp8_default_zig_zag1d[i];
+    z = coeff_ptr[rc];
 
-        sz = (z >> 31);                              /* sign of z */
-        x  = (z ^ sz) - sz;                          /* x = abs(z) */
+    sz = (z >> 31);    /* sign of z */
+    x = (z ^ sz) - sz; /* x = abs(z) */
 
-        y  = ((x + round_ptr[rc]) * quant_ptr[rc]) >> 16; /* quantize (x) */
-        x  = (y ^ sz) - sz;                          /* get the sign back */
-        qcoeff_ptr[rc] = x;                          /* write to destination */
-        dqcoeff_ptr[rc] = x * dequant_ptr[rc];       /* dequantized value */
+    y = ((x + round_ptr[rc]) * quant_ptr[rc]) >> 16; /* quantize (x) */
+    x = (y ^ sz) - sz;                               /* get the sign back */
+    qcoeff_ptr[rc] = x;                              /* write to destination */
+    dqcoeff_ptr[rc] = x * dequant_ptr[rc];           /* dequantized value */
 
-        if (y)
-        {
-            eob = i;                                 /* last nonzero coeffs */
-        }
+    if (y) {
+      eob = i; /* last nonzero coeffs */
     }
-    *d->eob = (char)(eob + 1);
+  }
+  *d->eob = (char)(eob + 1);
 }
 
-void vp8_regular_quantize_b_c(BLOCK *b, BLOCKD *d)
-{
-    int i, rc, eob;
-    int zbin;
-    int x, y, z, sz;
-    short *zbin_boost_ptr  = b->zrun_zbin_boost;
-    short *coeff_ptr       = b->coeff;
-    short *zbin_ptr        = b->zbin;
-    short *round_ptr       = b->round;
-    short *quant_ptr       = b->quant;
-    short *quant_shift_ptr = b->quant_shift;
-    short *qcoeff_ptr      = d->qcoeff;
-    short *dqcoeff_ptr     = d->dqcoeff;
-    short *dequant_ptr     = d->dequant;
-    short zbin_oq_value    = b->zbin_extra;
+void vp8_regular_quantize_b_c(BLOCK *b, BLOCKD *d) {
+  int i, rc, eob;
+  int zbin;
+  int x, y, z, sz;
+  short *zbin_boost_ptr = b->zrun_zbin_boost;
+  short *coeff_ptr = b->coeff;
+  short *zbin_ptr = b->zbin;
+  short *round_ptr = b->round;
+  short *quant_ptr = b->quant;
+  short *quant_shift_ptr = b->quant_shift;
+  short *qcoeff_ptr = d->qcoeff;
+  short *dqcoeff_ptr = d->dqcoeff;
+  short *dequant_ptr = d->dequant;
+  short zbin_oq_value = b->zbin_extra;
 
-    memset(qcoeff_ptr, 0, 32);
-    memset(dqcoeff_ptr, 0, 32);
+  memset(qcoeff_ptr, 0, 32);
+  memset(dqcoeff_ptr, 0, 32);
 
-    eob = -1;
+  eob = -1;
 
-    for (i = 0; i < 16; i++)
-    {
-        rc   = vp8_default_zig_zag1d[i];
-        z    = coeff_ptr[rc];
+  for (i = 0; i < 16; i++) {
+    rc = vp8_default_zig_zag1d[i];
+    z = coeff_ptr[rc];
 
-        zbin = zbin_ptr[rc] + *zbin_boost_ptr + zbin_oq_value;
+    zbin = zbin_ptr[rc] + *zbin_boost_ptr + zbin_oq_value;
 
-        zbin_boost_ptr ++;
-        sz = (z >> 31);                              /* sign of z */
-        x  = (z ^ sz) - sz;                          /* x = abs(z) */
+    zbin_boost_ptr++;
+    sz = (z >> 31);    /* sign of z */
+    x = (z ^ sz) - sz; /* x = abs(z) */
 
-        if (x >= zbin)
-        {
-            x += round_ptr[rc];
-            y  = ((((x * quant_ptr[rc]) >> 16) + x)
-                 * quant_shift_ptr[rc]) >> 16;       /* quantize (x) */
-            x  = (y ^ sz) - sz;                      /* get the sign back */
-            qcoeff_ptr[rc]  = x;                     /* write to destination */
-            dqcoeff_ptr[rc] = x * dequant_ptr[rc];   /* dequantized value */
+    if (x >= zbin) {
+      x += round_ptr[rc];
+      y = ((((x * quant_ptr[rc]) >> 16) + x) * quant_shift_ptr[rc]) >>
+          16;                                /* quantize (x) */
+      x = (y ^ sz) - sz;                     /* get the sign back */
+      qcoeff_ptr[rc] = x;                    /* write to destination */
+      dqcoeff_ptr[rc] = x * dequant_ptr[rc]; /* dequantized value */
 
-            if (y)
-            {
-                eob = i;                             /* last nonzero coeffs */
-                zbin_boost_ptr = b->zrun_zbin_boost; /* reset zero runlength */
-            }
-        }
+      if (y) {
+        eob = i;                             /* last nonzero coeffs */
+        zbin_boost_ptr = b->zrun_zbin_boost; /* reset zero runlength */
+      }
     }
+  }
 
-    *d->eob = (char)(eob + 1);
+  *d->eob = (char)(eob + 1);
 }
 
-void vp8_quantize_mby(MACROBLOCK *x)
-{
-    int i;
-    int has_2nd_order = (x->e_mbd.mode_info_context->mbmi.mode != B_PRED
-        && x->e_mbd.mode_info_context->mbmi.mode != SPLITMV);
+void vp8_quantize_mby(MACROBLOCK *x) {
+  int i;
+  int has_2nd_order = (x->e_mbd.mode_info_context->mbmi.mode != B_PRED &&
+                       x->e_mbd.mode_info_context->mbmi.mode != SPLITMV);
 
-    for (i = 0; i < 16; i++)
-        x->quantize_b(&x->block[i], &x->e_mbd.block[i]);
+  for (i = 0; i < 16; i++) x->quantize_b(&x->block[i], &x->e_mbd.block[i]);
 
-    if(has_2nd_order)
-        x->quantize_b(&x->block[24], &x->e_mbd.block[24]);
+  if (has_2nd_order) x->quantize_b(&x->block[24], &x->e_mbd.block[24]);
 }
 
-void vp8_quantize_mb(MACROBLOCK *x)
-{
-    int i;
-    int has_2nd_order=(x->e_mbd.mode_info_context->mbmi.mode != B_PRED
-        && x->e_mbd.mode_info_context->mbmi.mode != SPLITMV);
+void vp8_quantize_mb(MACROBLOCK *x) {
+  int i;
+  int has_2nd_order = (x->e_mbd.mode_info_context->mbmi.mode != B_PRED &&
+                       x->e_mbd.mode_info_context->mbmi.mode != SPLITMV);
 
-    for (i = 0; i < 24+has_2nd_order; i++)
-        x->quantize_b(&x->block[i], &x->e_mbd.block[i]);
+  for (i = 0; i < 24 + has_2nd_order; i++)
+    x->quantize_b(&x->block[i], &x->e_mbd.block[i]);
 }
 
+void vp8_quantize_mbuv(MACROBLOCK *x) {
+  int i;
 
-void vp8_quantize_mbuv(MACROBLOCK *x)
-{
-    int i;
-
-    for (i = 16; i < 24; i++)
-        x->quantize_b(&x->block[i], &x->e_mbd.block[i]);
+  for (i = 16; i < 24; i++) x->quantize_b(&x->block[i], &x->e_mbd.block[i]);
 }
 
-static const int qrounding_factors[129] =
-{
-    48, 48, 48, 48, 48, 48, 48, 48,
-    48, 48, 48, 48, 48, 48, 48, 48,
-    48, 48, 48, 48, 48, 48, 48, 48,
-    48, 48, 48, 48, 48, 48, 48, 48,
-    48, 48, 48, 48, 48, 48, 48, 48,
-    48, 48, 48, 48, 48, 48, 48, 48,
-    48, 48, 48, 48, 48, 48, 48, 48,
-    48, 48, 48, 48, 48, 48, 48, 48,
-    48, 48, 48, 48, 48, 48, 48, 48,
-    48, 48, 48, 48, 48, 48, 48, 48,
-    48, 48, 48, 48, 48, 48, 48, 48,
-    48, 48, 48, 48, 48, 48, 48, 48,
-    48, 48, 48, 48, 48, 48, 48, 48,
-    48, 48, 48, 48, 48, 48, 48, 48,
-    48, 48, 48, 48, 48, 48, 48, 48,
-    48, 48, 48, 48, 48, 48, 48, 48,
-    48
+static const int qrounding_factors[129] = {
+  48, 48, 48, 48, 48, 48, 48, 48, 48, 48, 48, 48, 48, 48, 48, 48, 48, 48, 48,
+  48, 48, 48, 48, 48, 48, 48, 48, 48, 48, 48, 48, 48, 48, 48, 48, 48, 48, 48,
+  48, 48, 48, 48, 48, 48, 48, 48, 48, 48, 48, 48, 48, 48, 48, 48, 48, 48, 48,
+  48, 48, 48, 48, 48, 48, 48, 48, 48, 48, 48, 48, 48, 48, 48, 48, 48, 48, 48,
+  48, 48, 48, 48, 48, 48, 48, 48, 48, 48, 48, 48, 48, 48, 48, 48, 48, 48, 48,
+  48, 48, 48, 48, 48, 48, 48, 48, 48, 48, 48, 48, 48, 48, 48, 48, 48, 48, 48,
+  48, 48, 48, 48, 48, 48, 48, 48, 48, 48, 48, 48, 48, 48, 48
 };
 
-
-static const int qzbin_factors[129] =
-{
-    84, 84, 84, 84, 84, 84, 84, 84,
-    84, 84, 84, 84, 84, 84, 84, 84,
-    84, 84, 84, 84, 84, 84, 84, 84,
-    84, 84, 84, 84, 84, 84, 84, 84,
-    84, 84, 84, 84, 84, 84, 84, 84,
-    84, 84, 84, 84, 84, 84, 84, 84,
-    80, 80, 80, 80, 80, 80, 80, 80,
-    80, 80, 80, 80, 80, 80, 80, 80,
-    80, 80, 80, 80, 80, 80, 80, 80,
-    80, 80, 80, 80, 80, 80, 80, 80,
-    80, 80, 80, 80, 80, 80, 80, 80,
-    80, 80, 80, 80, 80, 80, 80, 80,
-    80, 80, 80, 80, 80, 80, 80, 80,
-    80, 80, 80, 80, 80, 80, 80, 80,
-    80, 80, 80, 80, 80, 80, 80, 80,
-    80, 80, 80, 80, 80, 80, 80, 80,
-    80
+static const int qzbin_factors[129] = {
+  84, 84, 84, 84, 84, 84, 84, 84, 84, 84, 84, 84, 84, 84, 84, 84, 84, 84, 84,
+  84, 84, 84, 84, 84, 84, 84, 84, 84, 84, 84, 84, 84, 84, 84, 84, 84, 84, 84,
+  84, 84, 84, 84, 84, 84, 84, 84, 84, 84, 80, 80, 80, 80, 80, 80, 80, 80, 80,
+  80, 80, 80, 80, 80, 80, 80, 80, 80, 80, 80, 80, 80, 80, 80, 80, 80, 80, 80,
+  80, 80, 80, 80, 80, 80, 80, 80, 80, 80, 80, 80, 80, 80, 80, 80, 80, 80, 80,
+  80, 80, 80, 80, 80, 80, 80, 80, 80, 80, 80, 80, 80, 80, 80, 80, 80, 80, 80,
+  80, 80, 80, 80, 80, 80, 80, 80, 80, 80, 80, 80, 80, 80, 80
 };
 
-
-static const int qrounding_factors_y2[129] =
-{
-    48, 48, 48, 48, 48, 48, 48, 48,
-    48, 48, 48, 48, 48, 48, 48, 48,
-    48, 48, 48, 48, 48, 48, 48, 48,
-    48, 48, 48, 48, 48, 48, 48, 48,
-    48, 48, 48, 48, 48, 48, 48, 48,
-    48, 48, 48, 48, 48, 48, 48, 48,
-    48, 48, 48, 48, 48, 48, 48, 48,
-    48, 48, 48, 48, 48, 48, 48, 48,
-    48, 48, 48, 48, 48, 48, 48, 48,
-    48, 48, 48, 48, 48, 48, 48, 48,
-    48, 48, 48, 48, 48, 48, 48, 48,
-    48, 48, 48, 48, 48, 48, 48, 48,
-    48, 48, 48, 48, 48, 48, 48, 48,
-    48, 48, 48, 48, 48, 48, 48, 48,
-    48, 48, 48, 48, 48, 48, 48, 48,
-    48, 48, 48, 48, 48, 48, 48, 48,
-    48
+static const int qrounding_factors_y2[129] = {
+  48, 48, 48, 48, 48, 48, 48, 48, 48, 48, 48, 48, 48, 48, 48, 48, 48, 48, 48,
+  48, 48, 48, 48, 48, 48, 48, 48, 48, 48, 48, 48, 48, 48, 48, 48, 48, 48, 48,
+  48, 48, 48, 48, 48, 48, 48, 48, 48, 48, 48, 48, 48, 48, 48, 48, 48, 48, 48,
+  48, 48, 48, 48, 48, 48, 48, 48, 48, 48, 48, 48, 48, 48, 48, 48, 48, 48, 48,
+  48, 48, 48, 48, 48, 48, 48, 48, 48, 48, 48, 48, 48, 48, 48, 48, 48, 48, 48,
+  48, 48, 48, 48, 48, 48, 48, 48, 48, 48, 48, 48, 48, 48, 48, 48, 48, 48, 48,
+  48, 48, 48, 48, 48, 48, 48, 48, 48, 48, 48, 48, 48, 48, 48
 };
 
-
-static const int qzbin_factors_y2[129] =
-{
-    84, 84, 84, 84, 84, 84, 84, 84,
-    84, 84, 84, 84, 84, 84, 84, 84,
-    84, 84, 84, 84, 84, 84, 84, 84,
-    84, 84, 84, 84, 84, 84, 84, 84,
-    84, 84, 84, 84, 84, 84, 84, 84,
-    84, 84, 84, 84, 84, 84, 84, 84,
-    80, 80, 80, 80, 80, 80, 80, 80,
-    80, 80, 80, 80, 80, 80, 80, 80,
-    80, 80, 80, 80, 80, 80, 80, 80,
-    80, 80, 80, 80, 80, 80, 80, 80,
-    80, 80, 80, 80, 80, 80, 80, 80,
-    80, 80, 80, 80, 80, 80, 80, 80,
-    80, 80, 80, 80, 80, 80, 80, 80,
-    80, 80, 80, 80, 80, 80, 80, 80,
-    80, 80, 80, 80, 80, 80, 80, 80,
-    80, 80, 80, 80, 80, 80, 80, 80,
-    80
+static const int qzbin_factors_y2[129] = {
+  84, 84, 84, 84, 84, 84, 84, 84, 84, 84, 84, 84, 84, 84, 84, 84, 84, 84, 84,
+  84, 84, 84, 84, 84, 84, 84, 84, 84, 84, 84, 84, 84, 84, 84, 84, 84, 84, 84,
+  84, 84, 84, 84, 84, 84, 84, 84, 84, 84, 80, 80, 80, 80, 80, 80, 80, 80, 80,
+  80, 80, 80, 80, 80, 80, 80, 80, 80, 80, 80, 80, 80, 80, 80, 80, 80, 80, 80,
+  80, 80, 80, 80, 80, 80, 80, 80, 80, 80, 80, 80, 80, 80, 80, 80, 80, 80, 80,
+  80, 80, 80, 80, 80, 80, 80, 80, 80, 80, 80, 80, 80, 80, 80, 80, 80, 80, 80,
+  80, 80, 80, 80, 80, 80, 80, 80, 80, 80, 80, 80, 80, 80, 80
 };
 
-
-static void invert_quant(int improved_quant, short *quant,
-                         short *shift, short d)
-{
-    if(improved_quant)
-    {
-        unsigned t;
-        int l, m;
-        t = d;
-        for(l = 0; t > 1; l++)
-            t>>=1;
-        m = 1 + (1<<(16+l))/d;
-        *quant = (short)(m - (1<<16));
-        *shift = l;
-        /* use multiplication and constant shift by 16 */
-        *shift = 1 << (16 - *shift);
-    }
-    else
-    {
-        *quant = (1 << 16) / d;
-        *shift = 0;
-        /* use multiplication and constant shift by 16 */
-        *shift = 1 << (16 - *shift);
-    }
+static void invert_quant(int improved_quant, short *quant, short *shift,
+                         short d) {
+  if (improved_quant) {
+    unsigned t;
+    int l, m;
+    t = d;
+    for (l = 0; t > 1; l++) t >>= 1;
+    m = 1 + (1 << (16 + l)) / d;
+    *quant = (short)(m - (1 << 16));
+    *shift = l;
+    /* use multiplication and constant shift by 16 */
+    *shift = 1 << (16 - *shift);
+  } else {
+    *quant = (1 << 16) / d;
+    *shift = 0;
+    /* use multiplication and constant shift by 16 */
+    *shift = 1 << (16 - *shift);
+  }
 }
 
+void vp8cx_init_quantizer(VP8_COMP *cpi) {
+  int i;
+  int quant_val;
+  int Q;
 
-void vp8cx_init_quantizer(VP8_COMP *cpi)
-{
-    int i;
-    int quant_val;
-    int Q;
+  int zbin_boost[16] = { 0,  0,  8,  10, 12, 14, 16, 20,
+                         24, 28, 32, 36, 40, 44, 44, 44 };
 
-    int zbin_boost[16] = {0, 0, 8, 10, 12, 14, 16, 20, 24, 28, 32, 36, 40, 44,
-                          44, 44};
+  for (Q = 0; Q < QINDEX_RANGE; Q++) {
+    /* dc values */
+    quant_val = vp8_dc_quant(Q, cpi->common.y1dc_delta_q);
+    cpi->Y1quant_fast[Q][0] = (1 << 16) / quant_val;
+    invert_quant(cpi->sf.improved_quant, cpi->Y1quant[Q] + 0,
+                 cpi->Y1quant_shift[Q] + 0, quant_val);
+    cpi->Y1zbin[Q][0] = ((qzbin_factors[Q] * quant_val) + 64) >> 7;
+    cpi->Y1round[Q][0] = (qrounding_factors[Q] * quant_val) >> 7;
+    cpi->common.Y1dequant[Q][0] = quant_val;
+    cpi->zrun_zbin_boost_y1[Q][0] = (quant_val * zbin_boost[0]) >> 7;
 
-    for (Q = 0; Q < QINDEX_RANGE; Q++)
-    {
-        /* dc values */
-        quant_val = vp8_dc_quant(Q, cpi->common.y1dc_delta_q);
-        cpi->Y1quant_fast[Q][0] = (1 << 16) / quant_val;
-        invert_quant(cpi->sf.improved_quant, cpi->Y1quant[Q] + 0,
-                     cpi->Y1quant_shift[Q] + 0, quant_val);
-        cpi->Y1zbin[Q][0] = ((qzbin_factors[Q] * quant_val) + 64) >> 7;
-        cpi->Y1round[Q][0] = (qrounding_factors[Q] * quant_val) >> 7;
-        cpi->common.Y1dequant[Q][0] = quant_val;
-        cpi->zrun_zbin_boost_y1[Q][0] = (quant_val * zbin_boost[0]) >> 7;
+    quant_val = vp8_dc2quant(Q, cpi->common.y2dc_delta_q);
+    cpi->Y2quant_fast[Q][0] = (1 << 16) / quant_val;
+    invert_quant(cpi->sf.improved_quant, cpi->Y2quant[Q] + 0,
+                 cpi->Y2quant_shift[Q] + 0, quant_val);
+    cpi->Y2zbin[Q][0] = ((qzbin_factors_y2[Q] * quant_val) + 64) >> 7;
+    cpi->Y2round[Q][0] = (qrounding_factors_y2[Q] * quant_val) >> 7;
+    cpi->common.Y2dequant[Q][0] = quant_val;
+    cpi->zrun_zbin_boost_y2[Q][0] = (quant_val * zbin_boost[0]) >> 7;
 
-        quant_val = vp8_dc2quant(Q, cpi->common.y2dc_delta_q);
-        cpi->Y2quant_fast[Q][0] = (1 << 16) / quant_val;
-        invert_quant(cpi->sf.improved_quant, cpi->Y2quant[Q] + 0,
-                     cpi->Y2quant_shift[Q] + 0, quant_val);
-        cpi->Y2zbin[Q][0] = ((qzbin_factors_y2[Q] * quant_val) + 64) >> 7;
-        cpi->Y2round[Q][0] = (qrounding_factors_y2[Q] * quant_val) >> 7;
-        cpi->common.Y2dequant[Q][0] = quant_val;
-        cpi->zrun_zbin_boost_y2[Q][0] = (quant_val * zbin_boost[0]) >> 7;
+    quant_val = vp8_dc_uv_quant(Q, cpi->common.uvdc_delta_q);
+    cpi->UVquant_fast[Q][0] = (1 << 16) / quant_val;
+    invert_quant(cpi->sf.improved_quant, cpi->UVquant[Q] + 0,
+                 cpi->UVquant_shift[Q] + 0, quant_val);
+    cpi->UVzbin[Q][0] = ((qzbin_factors[Q] * quant_val) + 64) >> 7;
+    ;
+    cpi->UVround[Q][0] = (qrounding_factors[Q] * quant_val) >> 7;
+    cpi->common.UVdequant[Q][0] = quant_val;
+    cpi->zrun_zbin_boost_uv[Q][0] = (quant_val * zbin_boost[0]) >> 7;
 
-        quant_val = vp8_dc_uv_quant(Q, cpi->common.uvdc_delta_q);
-        cpi->UVquant_fast[Q][0] = (1 << 16) / quant_val;
-        invert_quant(cpi->sf.improved_quant, cpi->UVquant[Q] + 0,
-                     cpi->UVquant_shift[Q] + 0, quant_val);
-        cpi->UVzbin[Q][0] = ((qzbin_factors[Q] * quant_val) + 64) >> 7;;
-        cpi->UVround[Q][0] = (qrounding_factors[Q] * quant_val) >> 7;
-        cpi->common.UVdequant[Q][0] = quant_val;
-        cpi->zrun_zbin_boost_uv[Q][0] = (quant_val * zbin_boost[0]) >> 7;
+    /* all the ac values = ; */
+    quant_val = vp8_ac_yquant(Q);
+    cpi->Y1quant_fast[Q][1] = (1 << 16) / quant_val;
+    invert_quant(cpi->sf.improved_quant, cpi->Y1quant[Q] + 1,
+                 cpi->Y1quant_shift[Q] + 1, quant_val);
+    cpi->Y1zbin[Q][1] = ((qzbin_factors[Q] * quant_val) + 64) >> 7;
+    cpi->Y1round[Q][1] = (qrounding_factors[Q] * quant_val) >> 7;
+    cpi->common.Y1dequant[Q][1] = quant_val;
+    cpi->zrun_zbin_boost_y1[Q][1] = (quant_val * zbin_boost[1]) >> 7;
 
-        /* all the ac values = ; */
-        quant_val = vp8_ac_yquant(Q);
-        cpi->Y1quant_fast[Q][1] = (1 << 16) / quant_val;
-        invert_quant(cpi->sf.improved_quant, cpi->Y1quant[Q] + 1,
-                     cpi->Y1quant_shift[Q] + 1, quant_val);
-        cpi->Y1zbin[Q][1] = ((qzbin_factors[Q] * quant_val) + 64) >> 7;
-        cpi->Y1round[Q][1] = (qrounding_factors[Q] * quant_val) >> 7;
-        cpi->common.Y1dequant[Q][1] = quant_val;
-        cpi->zrun_zbin_boost_y1[Q][1] = (quant_val * zbin_boost[1]) >> 7;
+    quant_val = vp8_ac2quant(Q, cpi->common.y2ac_delta_q);
+    cpi->Y2quant_fast[Q][1] = (1 << 16) / quant_val;
+    invert_quant(cpi->sf.improved_quant, cpi->Y2quant[Q] + 1,
+                 cpi->Y2quant_shift[Q] + 1, quant_val);
+    cpi->Y2zbin[Q][1] = ((qzbin_factors_y2[Q] * quant_val) + 64) >> 7;
+    cpi->Y2round[Q][1] = (qrounding_factors_y2[Q] * quant_val) >> 7;
+    cpi->common.Y2dequant[Q][1] = quant_val;
+    cpi->zrun_zbin_boost_y2[Q][1] = (quant_val * zbin_boost[1]) >> 7;
 
-        quant_val = vp8_ac2quant(Q, cpi->common.y2ac_delta_q);
-        cpi->Y2quant_fast[Q][1] = (1 << 16) / quant_val;
-        invert_quant(cpi->sf.improved_quant, cpi->Y2quant[Q] + 1,
-                     cpi->Y2quant_shift[Q] + 1, quant_val);
-        cpi->Y2zbin[Q][1] = ((qzbin_factors_y2[Q] * quant_val) + 64) >> 7;
-        cpi->Y2round[Q][1] = (qrounding_factors_y2[Q] * quant_val) >> 7;
-        cpi->common.Y2dequant[Q][1] = quant_val;
-        cpi->zrun_zbin_boost_y2[Q][1] = (quant_val * zbin_boost[1]) >> 7;
+    quant_val = vp8_ac_uv_quant(Q, cpi->common.uvac_delta_q);
+    cpi->UVquant_fast[Q][1] = (1 << 16) / quant_val;
+    invert_quant(cpi->sf.improved_quant, cpi->UVquant[Q] + 1,
+                 cpi->UVquant_shift[Q] + 1, quant_val);
+    cpi->UVzbin[Q][1] = ((qzbin_factors[Q] * quant_val) + 64) >> 7;
+    cpi->UVround[Q][1] = (qrounding_factors[Q] * quant_val) >> 7;
+    cpi->common.UVdequant[Q][1] = quant_val;
+    cpi->zrun_zbin_boost_uv[Q][1] = (quant_val * zbin_boost[1]) >> 7;
 
-        quant_val = vp8_ac_uv_quant(Q, cpi->common.uvac_delta_q);
-        cpi->UVquant_fast[Q][1] = (1 << 16) / quant_val;
-        invert_quant(cpi->sf.improved_quant, cpi->UVquant[Q] + 1,
-                     cpi->UVquant_shift[Q] + 1, quant_val);
-        cpi->UVzbin[Q][1] = ((qzbin_factors[Q] * quant_val) + 64) >> 7;
-        cpi->UVround[Q][1] = (qrounding_factors[Q] * quant_val) >> 7;
-        cpi->common.UVdequant[Q][1] = quant_val;
-        cpi->zrun_zbin_boost_uv[Q][1] = (quant_val * zbin_boost[1]) >> 7;
+    for (i = 2; i < 16; i++) {
+      cpi->Y1quant_fast[Q][i] = cpi->Y1quant_fast[Q][1];
+      cpi->Y1quant[Q][i] = cpi->Y1quant[Q][1];
+      cpi->Y1quant_shift[Q][i] = cpi->Y1quant_shift[Q][1];
+      cpi->Y1zbin[Q][i] = cpi->Y1zbin[Q][1];
+      cpi->Y1round[Q][i] = cpi->Y1round[Q][1];
+      cpi->zrun_zbin_boost_y1[Q][i] =
+          (cpi->common.Y1dequant[Q][1] * zbin_boost[i]) >> 7;
 
-        for (i = 2; i < 16; i++)
-        {
-            cpi->Y1quant_fast[Q][i] = cpi->Y1quant_fast[Q][1];
-            cpi->Y1quant[Q][i] = cpi->Y1quant[Q][1];
-            cpi->Y1quant_shift[Q][i] = cpi->Y1quant_shift[Q][1];
-            cpi->Y1zbin[Q][i] = cpi->Y1zbin[Q][1];
-            cpi->Y1round[Q][i] = cpi->Y1round[Q][1];
-            cpi->zrun_zbin_boost_y1[Q][i] = (cpi->common.Y1dequant[Q][1] *
-                                             zbin_boost[i]) >> 7;
+      cpi->Y2quant_fast[Q][i] = cpi->Y2quant_fast[Q][1];
+      cpi->Y2quant[Q][i] = cpi->Y2quant[Q][1];
+      cpi->Y2quant_shift[Q][i] = cpi->Y2quant_shift[Q][1];
+      cpi->Y2zbin[Q][i] = cpi->Y2zbin[Q][1];
+      cpi->Y2round[Q][i] = cpi->Y2round[Q][1];
+      cpi->zrun_zbin_boost_y2[Q][i] =
+          (cpi->common.Y2dequant[Q][1] * zbin_boost[i]) >> 7;
 
-            cpi->Y2quant_fast[Q][i] = cpi->Y2quant_fast[Q][1];
-            cpi->Y2quant[Q][i] = cpi->Y2quant[Q][1];
-            cpi->Y2quant_shift[Q][i] = cpi->Y2quant_shift[Q][1];
-            cpi->Y2zbin[Q][i] = cpi->Y2zbin[Q][1];
-            cpi->Y2round[Q][i] = cpi->Y2round[Q][1];
-            cpi->zrun_zbin_boost_y2[Q][i] = (cpi->common.Y2dequant[Q][1] *
-                                             zbin_boost[i]) >> 7;
-
-            cpi->UVquant_fast[Q][i] = cpi->UVquant_fast[Q][1];
-            cpi->UVquant[Q][i] = cpi->UVquant[Q][1];
-            cpi->UVquant_shift[Q][i] = cpi->UVquant_shift[Q][1];
-            cpi->UVzbin[Q][i] = cpi->UVzbin[Q][1];
-            cpi->UVround[Q][i] = cpi->UVround[Q][1];
-            cpi->zrun_zbin_boost_uv[Q][i] = (cpi->common.UVdequant[Q][1] *
-                                             zbin_boost[i]) >> 7;
-        }
+      cpi->UVquant_fast[Q][i] = cpi->UVquant_fast[Q][1];
+      cpi->UVquant[Q][i] = cpi->UVquant[Q][1];
+      cpi->UVquant_shift[Q][i] = cpi->UVquant_shift[Q][1];
+      cpi->UVzbin[Q][i] = cpi->UVzbin[Q][1];
+      cpi->UVround[Q][i] = cpi->UVround[Q][1];
+      cpi->zrun_zbin_boost_uv[Q][i] =
+          (cpi->common.UVdequant[Q][1] * zbin_boost[i]) >> 7;
     }
+  }
 }
 
-#define ZBIN_EXTRA_Y \
-    (( cpi->common.Y1dequant[QIndex][1] *  \
-    ( x->zbin_over_quant +  \
-      x->zbin_mode_boost +  \
-      x->act_zbin_adj ) ) >> 7)
+#define ZBIN_EXTRA_Y                                                \
+  ((cpi->common.Y1dequant[QIndex][1] *                              \
+    (x->zbin_over_quant + x->zbin_mode_boost + x->act_zbin_adj)) >> \
+   7)
 
-#define ZBIN_EXTRA_UV \
-    (( cpi->common.UVdequant[QIndex][1] *  \
-    ( x->zbin_over_quant +  \
-      x->zbin_mode_boost +  \
-      x->act_zbin_adj ) ) >> 7)
+#define ZBIN_EXTRA_UV                                               \
+  ((cpi->common.UVdequant[QIndex][1] *                              \
+    (x->zbin_over_quant + x->zbin_mode_boost + x->act_zbin_adj)) >> \
+   7)
 
-#define ZBIN_EXTRA_Y2 \
-    (( cpi->common.Y2dequant[QIndex][1] *  \
-    ( (x->zbin_over_quant / 2) +  \
-       x->zbin_mode_boost +  \
-       x->act_zbin_adj ) ) >> 7)
+#define ZBIN_EXTRA_Y2                                                     \
+  ((cpi->common.Y2dequant[QIndex][1] *                                    \
+    ((x->zbin_over_quant / 2) + x->zbin_mode_boost + x->act_zbin_adj)) >> \
+   7)
 
-void vp8cx_mb_init_quantizer(VP8_COMP *cpi, MACROBLOCK *x, int ok_to_skip)
-{
-    int i;
-    int QIndex;
-    MACROBLOCKD *xd = &x->e_mbd;
-    int zbin_extra;
+void vp8cx_mb_init_quantizer(VP8_COMP *cpi, MACROBLOCK *x, int ok_to_skip) {
+  int i;
+  int QIndex;
+  MACROBLOCKD *xd = &x->e_mbd;
+  int zbin_extra;
 
-    /* Select the baseline MB Q index. */
-    if (xd->segmentation_enabled)
-    {
-        /* Abs Value */
-        if (xd->mb_segement_abs_delta == SEGMENT_ABSDATA)
-            QIndex = xd->segment_feature_data[MB_LVL_ALT_Q][xd->mode_info_context->mbmi.segment_id];
-        /* Delta Value */
-        else
-        {
-            QIndex = cpi->common.base_qindex + xd->segment_feature_data[MB_LVL_ALT_Q][xd->mode_info_context->mbmi.segment_id];
-            /* Clamp to valid range */
-            QIndex = (QIndex >= 0) ? ((QIndex <= MAXQ) ? QIndex : MAXQ) : 0;
-        }
+  /* Select the baseline MB Q index. */
+  if (xd->segmentation_enabled) {
+    /* Abs Value */
+    if (xd->mb_segement_abs_delta == SEGMENT_ABSDATA)
+      QIndex = xd->segment_feature_data[MB_LVL_ALT_Q][xd->mode_info_context
+                                                          ->mbmi.segment_id];
+    /* Delta Value */
+    else {
+      QIndex = cpi->common.base_qindex +
+               xd->segment_feature_data[MB_LVL_ALT_Q][xd->mode_info_context
+                                                          ->mbmi.segment_id];
+      /* Clamp to valid range */
+      QIndex = (QIndex >= 0) ? ((QIndex <= MAXQ) ? QIndex : MAXQ) : 0;
     }
-    else
-        QIndex = cpi->common.base_qindex;
+  } else
+    QIndex = cpi->common.base_qindex;
 
-    /* This initialization should be called at least once. Use ok_to_skip to
-     * decide if it is ok to skip.
-     * Before encoding a frame, this function is always called with ok_to_skip
-     * =0, which means no skiping of calculations. The "last" values are
-     * initialized at that time.
-     */
-    if (!ok_to_skip || QIndex != x->q_index)
-    {
+  /* This initialization should be called at least once. Use ok_to_skip to
+   * decide if it is ok to skip.
+   * Before encoding a frame, this function is always called with ok_to_skip
+   * =0, which means no skiping of calculations. The "last" values are
+   * initialized at that time.
+   */
+  if (!ok_to_skip || QIndex != x->q_index) {
+    xd->dequant_y1_dc[0] = 1;
+    xd->dequant_y1[0] = cpi->common.Y1dequant[QIndex][0];
+    xd->dequant_y2[0] = cpi->common.Y2dequant[QIndex][0];
+    xd->dequant_uv[0] = cpi->common.UVdequant[QIndex][0];
 
-        xd->dequant_y1_dc[0] = 1;
-        xd->dequant_y1[0] = cpi->common.Y1dequant[QIndex][0];
-        xd->dequant_y2[0] = cpi->common.Y2dequant[QIndex][0];
-        xd->dequant_uv[0] = cpi->common.UVdequant[QIndex][0];
-
-        for (i = 1; i < 16; i++)
-        {
-            xd->dequant_y1_dc[i] =
-            xd->dequant_y1[i] = cpi->common.Y1dequant[QIndex][1];
-            xd->dequant_y2[i] = cpi->common.Y2dequant[QIndex][1];
-            xd->dequant_uv[i] = cpi->common.UVdequant[QIndex][1];
-        }
+    for (i = 1; i < 16; i++) {
+      xd->dequant_y1_dc[i] = xd->dequant_y1[i] =
+          cpi->common.Y1dequant[QIndex][1];
+      xd->dequant_y2[i] = cpi->common.Y2dequant[QIndex][1];
+      xd->dequant_uv[i] = cpi->common.UVdequant[QIndex][1];
+    }
 #if 1
-        /*TODO:  Remove dequant from BLOCKD.  This is a temporary solution until
-         * the quantizer code uses a passed in pointer to the dequant constants.
-         * This will also require modifications to the x86 and neon assembly.
-         * */
-        for (i = 0; i < 16; i++)
-            x->e_mbd.block[i].dequant = xd->dequant_y1;
-        for (i = 16; i < 24; i++)
-            x->e_mbd.block[i].dequant = xd->dequant_uv;
-        x->e_mbd.block[24].dequant = xd->dequant_y2;
+    /*TODO:  Remove dequant from BLOCKD.  This is a temporary solution until
+     * the quantizer code uses a passed in pointer to the dequant constants.
+     * This will also require modifications to the x86 and neon assembly.
+     * */
+    for (i = 0; i < 16; i++) x->e_mbd.block[i].dequant = xd->dequant_y1;
+    for (i = 16; i < 24; i++) x->e_mbd.block[i].dequant = xd->dequant_uv;
+    x->e_mbd.block[24].dequant = xd->dequant_y2;
 #endif
 
-        /* Y */
-        zbin_extra = ZBIN_EXTRA_Y;
+    /* Y */
+    zbin_extra = ZBIN_EXTRA_Y;
 
-        for (i = 0; i < 16; i++)
-        {
-            x->block[i].quant = cpi->Y1quant[QIndex];
-            x->block[i].quant_fast = cpi->Y1quant_fast[QIndex];
-            x->block[i].quant_shift = cpi->Y1quant_shift[QIndex];
-            x->block[i].zbin = cpi->Y1zbin[QIndex];
-            x->block[i].round = cpi->Y1round[QIndex];
-            x->block[i].zrun_zbin_boost = cpi->zrun_zbin_boost_y1[QIndex];
-            x->block[i].zbin_extra = (short)zbin_extra;
-        }
+    for (i = 0; i < 16; i++) {
+      x->block[i].quant = cpi->Y1quant[QIndex];
+      x->block[i].quant_fast = cpi->Y1quant_fast[QIndex];
+      x->block[i].quant_shift = cpi->Y1quant_shift[QIndex];
+      x->block[i].zbin = cpi->Y1zbin[QIndex];
+      x->block[i].round = cpi->Y1round[QIndex];
+      x->block[i].zrun_zbin_boost = cpi->zrun_zbin_boost_y1[QIndex];
+      x->block[i].zbin_extra = (short)zbin_extra;
+    }
 
-        /* UV */
-        zbin_extra = ZBIN_EXTRA_UV;
+    /* UV */
+    zbin_extra = ZBIN_EXTRA_UV;
 
-        for (i = 16; i < 24; i++)
-        {
-            x->block[i].quant = cpi->UVquant[QIndex];
-            x->block[i].quant_fast = cpi->UVquant_fast[QIndex];
-            x->block[i].quant_shift = cpi->UVquant_shift[QIndex];
-            x->block[i].zbin = cpi->UVzbin[QIndex];
-            x->block[i].round = cpi->UVround[QIndex];
-            x->block[i].zrun_zbin_boost = cpi->zrun_zbin_boost_uv[QIndex];
-            x->block[i].zbin_extra = (short)zbin_extra;
-        }
-
-        /* Y2 */
-        zbin_extra = ZBIN_EXTRA_Y2;
-
-        x->block[24].quant_fast = cpi->Y2quant_fast[QIndex];
-        x->block[24].quant = cpi->Y2quant[QIndex];
-        x->block[24].quant_shift = cpi->Y2quant_shift[QIndex];
-        x->block[24].zbin = cpi->Y2zbin[QIndex];
-        x->block[24].round = cpi->Y2round[QIndex];
-        x->block[24].zrun_zbin_boost = cpi->zrun_zbin_boost_y2[QIndex];
-        x->block[24].zbin_extra = (short)zbin_extra;
-
-        /* save this macroblock QIndex for vp8_update_zbin_extra() */
-        x->q_index = QIndex;
-
-        x->last_zbin_over_quant = x->zbin_over_quant;
-        x->last_zbin_mode_boost = x->zbin_mode_boost;
-        x->last_act_zbin_adj = x->act_zbin_adj;
-
-
-
+    for (i = 16; i < 24; i++) {
+      x->block[i].quant = cpi->UVquant[QIndex];
+      x->block[i].quant_fast = cpi->UVquant_fast[QIndex];
+      x->block[i].quant_shift = cpi->UVquant_shift[QIndex];
+      x->block[i].zbin = cpi->UVzbin[QIndex];
+      x->block[i].round = cpi->UVround[QIndex];
+      x->block[i].zrun_zbin_boost = cpi->zrun_zbin_boost_uv[QIndex];
+      x->block[i].zbin_extra = (short)zbin_extra;
     }
-    else if(x->last_zbin_over_quant != x->zbin_over_quant
-            || x->last_zbin_mode_boost != x->zbin_mode_boost
-            || x->last_act_zbin_adj != x->act_zbin_adj)
-    {
-        /* Y */
-        zbin_extra = ZBIN_EXTRA_Y;
 
-        for (i = 0; i < 16; i++)
-            x->block[i].zbin_extra = (short)zbin_extra;
+    /* Y2 */
+    zbin_extra = ZBIN_EXTRA_Y2;
 
-        /* UV */
-        zbin_extra = ZBIN_EXTRA_UV;
+    x->block[24].quant_fast = cpi->Y2quant_fast[QIndex];
+    x->block[24].quant = cpi->Y2quant[QIndex];
+    x->block[24].quant_shift = cpi->Y2quant_shift[QIndex];
+    x->block[24].zbin = cpi->Y2zbin[QIndex];
+    x->block[24].round = cpi->Y2round[QIndex];
+    x->block[24].zrun_zbin_boost = cpi->zrun_zbin_boost_y2[QIndex];
+    x->block[24].zbin_extra = (short)zbin_extra;
 
-        for (i = 16; i < 24; i++)
-            x->block[i].zbin_extra = (short)zbin_extra;
+    /* save this macroblock QIndex for vp8_update_zbin_extra() */
+    x->q_index = QIndex;
 
-        /* Y2 */
-        zbin_extra = ZBIN_EXTRA_Y2;
-        x->block[24].zbin_extra = (short)zbin_extra;
+    x->last_zbin_over_quant = x->zbin_over_quant;
+    x->last_zbin_mode_boost = x->zbin_mode_boost;
+    x->last_act_zbin_adj = x->act_zbin_adj;
 
-        x->last_zbin_over_quant = x->zbin_over_quant;
-        x->last_zbin_mode_boost = x->zbin_mode_boost;
-        x->last_act_zbin_adj = x->act_zbin_adj;
-    }
-}
-
-void vp8_update_zbin_extra(VP8_COMP *cpi, MACROBLOCK *x)
-{
-    int i;
-    int QIndex = x->q_index;
-    int zbin_extra;
-
+  } else if (x->last_zbin_over_quant != x->zbin_over_quant ||
+             x->last_zbin_mode_boost != x->zbin_mode_boost ||
+             x->last_act_zbin_adj != x->act_zbin_adj) {
     /* Y */
     zbin_extra = ZBIN_EXTRA_Y;
 
-    for (i = 0; i < 16; i++)
-        x->block[i].zbin_extra = (short)zbin_extra;
+    for (i = 0; i < 16; i++) x->block[i].zbin_extra = (short)zbin_extra;
 
     /* UV */
     zbin_extra = ZBIN_EXTRA_UV;
 
-    for (i = 16; i < 24; i++)
-        x->block[i].zbin_extra = (short)zbin_extra;
+    for (i = 16; i < 24; i++) x->block[i].zbin_extra = (short)zbin_extra;
 
     /* Y2 */
     zbin_extra = ZBIN_EXTRA_Y2;
     x->block[24].zbin_extra = (short)zbin_extra;
+
+    x->last_zbin_over_quant = x->zbin_over_quant;
+    x->last_zbin_mode_boost = x->zbin_mode_boost;
+    x->last_act_zbin_adj = x->act_zbin_adj;
+  }
 }
+
+void vp8_update_zbin_extra(VP8_COMP *cpi, MACROBLOCK *x) {
+  int i;
+  int QIndex = x->q_index;
+  int zbin_extra;
+
+  /* Y */
+  zbin_extra = ZBIN_EXTRA_Y;
+
+  for (i = 0; i < 16; i++) x->block[i].zbin_extra = (short)zbin_extra;
+
+  /* UV */
+  zbin_extra = ZBIN_EXTRA_UV;
+
+  for (i = 16; i < 24; i++) x->block[i].zbin_extra = (short)zbin_extra;
+
+  /* Y2 */
+  zbin_extra = ZBIN_EXTRA_Y2;
+  x->block[24].zbin_extra = (short)zbin_extra;
+}
 #undef ZBIN_EXTRA_Y
 #undef ZBIN_EXTRA_UV
 #undef ZBIN_EXTRA_Y2
 
-void vp8cx_frame_init_quantizer(VP8_COMP *cpi)
-{
-    /* Clear Zbin mode boost for default case */
-    cpi->mb.zbin_mode_boost = 0;
+void vp8cx_frame_init_quantizer(VP8_COMP *cpi) {
+  /* Clear Zbin mode boost for default case */
+  cpi->mb.zbin_mode_boost = 0;
 
-    /* MB level quantizer setup */
-    vp8cx_mb_init_quantizer(cpi, &cpi->mb, 0);
+  /* MB level quantizer setup */
+  vp8cx_mb_init_quantizer(cpi, &cpi->mb, 0);
 }
 
+void vp8_set_quantizer(struct VP8_COMP *cpi, int Q) {
+  VP8_COMMON *cm = &cpi->common;
+  MACROBLOCKD *mbd = &cpi->mb.e_mbd;
+  int update = 0;
+  int new_delta_q;
+  int new_uv_delta_q;
+  cm->base_qindex = Q;
 
-void vp8_set_quantizer(struct VP8_COMP *cpi, int Q)
-{
-    VP8_COMMON *cm = &cpi->common;
-    MACROBLOCKD *mbd = &cpi->mb.e_mbd;
-    int update = 0;
-    int new_delta_q;
-    int new_uv_delta_q;
-    cm->base_qindex = Q;
+  /* if any of the delta_q values are changing update flag has to be set */
+  /* currently only y2dc_delta_q may change */
 
-    /* if any of the delta_q values are changing update flag has to be set */
-    /* currently only y2dc_delta_q may change */
+  cm->y1dc_delta_q = 0;
+  cm->y2ac_delta_q = 0;
 
-    cm->y1dc_delta_q = 0;
-    cm->y2ac_delta_q = 0;
+  if (Q < 4) {
+    new_delta_q = 4 - Q;
+  } else
+    new_delta_q = 0;
 
-    if (Q < 4)
-    {
-        new_delta_q = 4-Q;
-    }
-    else
-        new_delta_q = 0;
+  update |= cm->y2dc_delta_q != new_delta_q;
+  cm->y2dc_delta_q = new_delta_q;
 
-    update |= cm->y2dc_delta_q != new_delta_q;
-    cm->y2dc_delta_q = new_delta_q;
-
-    new_uv_delta_q = 0;
-    // For screen content, lower the q value for UV channel. For now, select
-    // conservative delta; same delta for dc and ac, and decrease it with lower
-    // Q, and set to 0 below some threshold. May want to condition this in
-    // future on the variance/energy in UV channel.
-    if (cpi->oxcf.screen_content_mode && Q > 40) {
-      new_uv_delta_q = -(int)(0.15 * Q);
-      // Check range: magnitude of delta is 4 bits.
-      if (new_uv_delta_q < -15) {
-        new_uv_delta_q = -15;
-      }
+  new_uv_delta_q = 0;
+  // For screen content, lower the q value for UV channel. For now, select
+  // conservative delta; same delta for dc and ac, and decrease it with lower
+  // Q, and set to 0 below some threshold. May want to condition this in
+  // future on the variance/energy in UV channel.
+  if (cpi->oxcf.screen_content_mode && Q > 40) {
+    new_uv_delta_q = -(int)(0.15 * Q);
+    // Check range: magnitude of delta is 4 bits.
+    if (new_uv_delta_q < -15) {
+      new_uv_delta_q = -15;
     }
-    update |= cm->uvdc_delta_q != new_uv_delta_q;
-    cm->uvdc_delta_q = new_uv_delta_q;
-    cm->uvac_delta_q = new_uv_delta_q;
+  }
+  update |= cm->uvdc_delta_q != new_uv_delta_q;
+  cm->uvdc_delta_q = new_uv_delta_q;
+  cm->uvac_delta_q = new_uv_delta_q;
 
-    /* Set Segment specific quatizers */
-    mbd->segment_feature_data[MB_LVL_ALT_Q][0] = cpi->segment_feature_data[MB_LVL_ALT_Q][0];
-    mbd->segment_feature_data[MB_LVL_ALT_Q][1] = cpi->segment_feature_data[MB_LVL_ALT_Q][1];
-    mbd->segment_feature_data[MB_LVL_ALT_Q][2] = cpi->segment_feature_data[MB_LVL_ALT_Q][2];
-    mbd->segment_feature_data[MB_LVL_ALT_Q][3] = cpi->segment_feature_data[MB_LVL_ALT_Q][3];
+  /* Set Segment specific quatizers */
+  mbd->segment_feature_data[MB_LVL_ALT_Q][0] =
+      cpi->segment_feature_data[MB_LVL_ALT_Q][0];
+  mbd->segment_feature_data[MB_LVL_ALT_Q][1] =
+      cpi->segment_feature_data[MB_LVL_ALT_Q][1];
+  mbd->segment_feature_data[MB_LVL_ALT_Q][2] =
+      cpi->segment_feature_data[MB_LVL_ALT_Q][2];
+  mbd->segment_feature_data[MB_LVL_ALT_Q][3] =
+      cpi->segment_feature_data[MB_LVL_ALT_Q][3];
 
-    /* quantizer has to be reinitialized for any delta_q changes */
-    if(update)
-        vp8cx_init_quantizer(cpi);
-
+  /* quantizer has to be reinitialized for any delta_q changes */
+  if (update) vp8cx_init_quantizer(cpi);
 }
--- a/vp8/encoder/x86/denoising_sse2.c
+++ b/vp8/encoder/x86/denoising_sse2.c
@@ -20,16 +20,16 @@
 /* Compute the sum of all pixel differences of this MB. */
 static INLINE unsigned int abs_sum_diff_16x1(__m128i acc_diff) {
   const __m128i k_1 = _mm_set1_epi16(1);
-  const __m128i acc_diff_lo = _mm_srai_epi16(
-      _mm_unpacklo_epi8(acc_diff, acc_diff), 8);
-  const __m128i acc_diff_hi = _mm_srai_epi16(
-      _mm_unpackhi_epi8(acc_diff, acc_diff), 8);
+  const __m128i acc_diff_lo =
+      _mm_srai_epi16(_mm_unpacklo_epi8(acc_diff, acc_diff), 8);
+  const __m128i acc_diff_hi =
+      _mm_srai_epi16(_mm_unpackhi_epi8(acc_diff, acc_diff), 8);
   const __m128i acc_diff_16 = _mm_add_epi16(acc_diff_lo, acc_diff_hi);
   const __m128i hg_fe_dc_ba = _mm_madd_epi16(acc_diff_16, k_1);
-  const __m128i hgfe_dcba = _mm_add_epi32(hg_fe_dc_ba,
-                                          _mm_srli_si128(hg_fe_dc_ba, 8));
-  const __m128i hgfedcba = _mm_add_epi32(hgfe_dcba,
-                                         _mm_srli_si128(hgfe_dcba, 4));
+  const __m128i hgfe_dcba =
+      _mm_add_epi32(hg_fe_dc_ba, _mm_srli_si128(hg_fe_dc_ba, 8));
+  const __m128i hgfedcba =
+      _mm_add_epi32(hgfe_dcba, _mm_srli_si128(hgfe_dcba, 4));
   unsigned int sum_diff = abs(_mm_cvtsi128_si32(hgfedcba));
 
   return sum_diff;
@@ -36,344 +36,337 @@
 }
 
 int vp8_denoiser_filter_sse2(unsigned char *mc_running_avg_y,
-                             int mc_avg_y_stride,
-                             unsigned char *running_avg_y, int avg_y_stride,
-                             unsigned char *sig, int sig_stride,
-                             unsigned int motion_magnitude,
-                             int increase_denoising)
-{
-    unsigned char *running_avg_y_start = running_avg_y;
-    unsigned char *sig_start = sig;
-    unsigned int sum_diff_thresh;
-    int r;
-    int shift_inc  = (increase_denoising &&
-        motion_magnitude <= MOTION_MAGNITUDE_THRESHOLD) ? 1 : 0;
-    __m128i acc_diff = _mm_setzero_si128();
-    const __m128i k_0 = _mm_setzero_si128();
-    const __m128i k_4 = _mm_set1_epi8(4 + shift_inc);
-    const __m128i k_8 = _mm_set1_epi8(8);
-    const __m128i k_16 = _mm_set1_epi8(16);
-    /* Modify each level's adjustment according to motion_magnitude. */
-    const __m128i l3 = _mm_set1_epi8(
-                       (motion_magnitude <= MOTION_MAGNITUDE_THRESHOLD) ?
-                        7 + shift_inc : 6);
-    /* Difference between level 3 and level 2 is 2. */
-    const __m128i l32 = _mm_set1_epi8(2);
-    /* Difference between level 2 and level 1 is 1. */
-    const __m128i l21 = _mm_set1_epi8(1);
+                             int mc_avg_y_stride, unsigned char *running_avg_y,
+                             int avg_y_stride, unsigned char *sig,
+                             int sig_stride, unsigned int motion_magnitude,
+                             int increase_denoising) {
+  unsigned char *running_avg_y_start = running_avg_y;
+  unsigned char *sig_start = sig;
+  unsigned int sum_diff_thresh;
+  int r;
+  int shift_inc =
+      (increase_denoising && motion_magnitude <= MOTION_MAGNITUDE_THRESHOLD)
+          ? 1
+          : 0;
+  __m128i acc_diff = _mm_setzero_si128();
+  const __m128i k_0 = _mm_setzero_si128();
+  const __m128i k_4 = _mm_set1_epi8(4 + shift_inc);
+  const __m128i k_8 = _mm_set1_epi8(8);
+  const __m128i k_16 = _mm_set1_epi8(16);
+  /* Modify each level's adjustment according to motion_magnitude. */
+  const __m128i l3 = _mm_set1_epi8(
+      (motion_magnitude <= MOTION_MAGNITUDE_THRESHOLD) ? 7 + shift_inc : 6);
+  /* Difference between level 3 and level 2 is 2. */
+  const __m128i l32 = _mm_set1_epi8(2);
+  /* Difference between level 2 and level 1 is 1. */
+  const __m128i l21 = _mm_set1_epi8(1);
 
-    for (r = 0; r < 16; ++r)
-    {
-        /* Calculate differences */
-        const __m128i v_sig = _mm_loadu_si128((__m128i *)(&sig[0]));
-        const __m128i v_mc_running_avg_y = _mm_loadu_si128(
-                                           (__m128i *)(&mc_running_avg_y[0]));
-        __m128i v_running_avg_y;
-        const __m128i pdiff = _mm_subs_epu8(v_mc_running_avg_y, v_sig);
-        const __m128i ndiff = _mm_subs_epu8(v_sig, v_mc_running_avg_y);
-        /* Obtain the sign. FF if diff is negative. */
-        const __m128i diff_sign = _mm_cmpeq_epi8(pdiff, k_0);
-        /* Clamp absolute difference to 16 to be used to get mask. Doing this
-         * allows us to use _mm_cmpgt_epi8, which operates on signed byte. */
-        const __m128i clamped_absdiff = _mm_min_epu8(
-                                        _mm_or_si128(pdiff, ndiff), k_16);
-        /* Get masks for l2 l1 and l0 adjustments */
-        const __m128i mask2 = _mm_cmpgt_epi8(k_16, clamped_absdiff);
-        const __m128i mask1 = _mm_cmpgt_epi8(k_8, clamped_absdiff);
-        const __m128i mask0 = _mm_cmpgt_epi8(k_4, clamped_absdiff);
-        /* Get adjustments for l2, l1, and l0 */
-        __m128i adj2 = _mm_and_si128(mask2, l32);
-        const __m128i adj1 = _mm_and_si128(mask1, l21);
-        const __m128i adj0 = _mm_and_si128(mask0, clamped_absdiff);
-        __m128i adj,  padj, nadj;
+  for (r = 0; r < 16; ++r) {
+    /* Calculate differences */
+    const __m128i v_sig = _mm_loadu_si128((__m128i *)(&sig[0]));
+    const __m128i v_mc_running_avg_y =
+        _mm_loadu_si128((__m128i *)(&mc_running_avg_y[0]));
+    __m128i v_running_avg_y;
+    const __m128i pdiff = _mm_subs_epu8(v_mc_running_avg_y, v_sig);
+    const __m128i ndiff = _mm_subs_epu8(v_sig, v_mc_running_avg_y);
+    /* Obtain the sign. FF if diff is negative. */
+    const __m128i diff_sign = _mm_cmpeq_epi8(pdiff, k_0);
+    /* Clamp absolute difference to 16 to be used to get mask. Doing this
+     * allows us to use _mm_cmpgt_epi8, which operates on signed byte. */
+    const __m128i clamped_absdiff =
+        _mm_min_epu8(_mm_or_si128(pdiff, ndiff), k_16);
+    /* Get masks for l2 l1 and l0 adjustments */
+    const __m128i mask2 = _mm_cmpgt_epi8(k_16, clamped_absdiff);
+    const __m128i mask1 = _mm_cmpgt_epi8(k_8, clamped_absdiff);
+    const __m128i mask0 = _mm_cmpgt_epi8(k_4, clamped_absdiff);
+    /* Get adjustments for l2, l1, and l0 */
+    __m128i adj2 = _mm_and_si128(mask2, l32);
+    const __m128i adj1 = _mm_and_si128(mask1, l21);
+    const __m128i adj0 = _mm_and_si128(mask0, clamped_absdiff);
+    __m128i adj, padj, nadj;
 
-        /* Combine the adjustments and get absolute adjustments. */
-        adj2 = _mm_add_epi8(adj2, adj1);
-        adj = _mm_sub_epi8(l3, adj2);
-        adj = _mm_andnot_si128(mask0, adj);
-        adj = _mm_or_si128(adj, adj0);
+    /* Combine the adjustments and get absolute adjustments. */
+    adj2 = _mm_add_epi8(adj2, adj1);
+    adj = _mm_sub_epi8(l3, adj2);
+    adj = _mm_andnot_si128(mask0, adj);
+    adj = _mm_or_si128(adj, adj0);
 
-        /* Restore the sign and get positive and negative adjustments. */
-        padj = _mm_andnot_si128(diff_sign, adj);
-        nadj = _mm_and_si128(diff_sign, adj);
+    /* Restore the sign and get positive and negative adjustments. */
+    padj = _mm_andnot_si128(diff_sign, adj);
+    nadj = _mm_and_si128(diff_sign, adj);
 
-        /* Calculate filtered value. */
-        v_running_avg_y = _mm_adds_epu8(v_sig, padj);
-        v_running_avg_y = _mm_subs_epu8(v_running_avg_y, nadj);
-        _mm_storeu_si128((__m128i *)running_avg_y, v_running_avg_y);
+    /* Calculate filtered value. */
+    v_running_avg_y = _mm_adds_epu8(v_sig, padj);
+    v_running_avg_y = _mm_subs_epu8(v_running_avg_y, nadj);
+    _mm_storeu_si128((__m128i *)running_avg_y, v_running_avg_y);
 
-        /* Adjustments <=7, and each element in acc_diff can fit in signed
-         * char.
-         */
-        acc_diff = _mm_adds_epi8(acc_diff, padj);
-        acc_diff = _mm_subs_epi8(acc_diff, nadj);
+    /* Adjustments <=7, and each element in acc_diff can fit in signed
+     * char.
+     */
+    acc_diff = _mm_adds_epi8(acc_diff, padj);
+    acc_diff = _mm_subs_epi8(acc_diff, nadj);
 
-        /* Update pointers for next iteration. */
-        sig += sig_stride;
-        mc_running_avg_y += mc_avg_y_stride;
-        running_avg_y += avg_y_stride;
-    }
+    /* Update pointers for next iteration. */
+    sig += sig_stride;
+    mc_running_avg_y += mc_avg_y_stride;
+    running_avg_y += avg_y_stride;
+  }
 
-    {
-        /* Compute the sum of all pixel differences of this MB. */
-        unsigned int abs_sum_diff = abs_sum_diff_16x1(acc_diff);
-        sum_diff_thresh = SUM_DIFF_THRESHOLD;
-        if (increase_denoising) sum_diff_thresh = SUM_DIFF_THRESHOLD_HIGH;
-        if (abs_sum_diff > sum_diff_thresh) {
-          // Before returning to copy the block (i.e., apply no denoising),
-          // check if we can still apply some (weaker) temporal filtering to
-          // this block, that would otherwise not be denoised at all. Simplest
-          // is to apply an additional adjustment to running_avg_y to bring it
-          // closer to sig. The adjustment is capped by a maximum delta, and
-          // chosen such that in most cases the resulting sum_diff will be
-          // within the acceptable range given by sum_diff_thresh.
+  {
+    /* Compute the sum of all pixel differences of this MB. */
+    unsigned int abs_sum_diff = abs_sum_diff_16x1(acc_diff);
+    sum_diff_thresh = SUM_DIFF_THRESHOLD;
+    if (increase_denoising) sum_diff_thresh = SUM_DIFF_THRESHOLD_HIGH;
+    if (abs_sum_diff > sum_diff_thresh) {
+      // Before returning to copy the block (i.e., apply no denoising),
+      // check if we can still apply some (weaker) temporal filtering to
+      // this block, that would otherwise not be denoised at all. Simplest
+      // is to apply an additional adjustment to running_avg_y to bring it
+      // closer to sig. The adjustment is capped by a maximum delta, and
+      // chosen such that in most cases the resulting sum_diff will be
+      // within the acceptable range given by sum_diff_thresh.
 
-          // The delta is set by the excess of absolute pixel diff over the
-          // threshold.
-          int delta = ((abs_sum_diff - sum_diff_thresh) >> 8) + 1;
-          // Only apply the adjustment for max delta up to 3.
-          if (delta < 4) {
-            const __m128i k_delta = _mm_set1_epi8(delta);
-            sig -= sig_stride * 16;
-            mc_running_avg_y -= mc_avg_y_stride * 16;
-            running_avg_y -= avg_y_stride * 16;
-            for (r = 0; r < 16; ++r) {
-              __m128i v_running_avg_y =
-                  _mm_loadu_si128((__m128i *)(&running_avg_y[0]));
-              // Calculate differences.
-              const __m128i v_sig = _mm_loadu_si128((__m128i *)(&sig[0]));
-              const __m128i v_mc_running_avg_y =
-                  _mm_loadu_si128((__m128i *)(&mc_running_avg_y[0]));
-              const __m128i pdiff = _mm_subs_epu8(v_mc_running_avg_y, v_sig);
-              const __m128i ndiff = _mm_subs_epu8(v_sig, v_mc_running_avg_y);
-              // Obtain the sign. FF if diff is negative.
-              const __m128i diff_sign = _mm_cmpeq_epi8(pdiff, k_0);
-              // Clamp absolute difference to delta to get the adjustment.
-              const __m128i adj =
-                  _mm_min_epu8(_mm_or_si128(pdiff, ndiff), k_delta);
-              // Restore the sign and get positive and negative adjustments.
-              __m128i padj, nadj;
-              padj = _mm_andnot_si128(diff_sign, adj);
-              nadj = _mm_and_si128(diff_sign, adj);
-              // Calculate filtered value.
-              v_running_avg_y = _mm_subs_epu8(v_running_avg_y, padj);
-              v_running_avg_y = _mm_adds_epu8(v_running_avg_y, nadj);
-             _mm_storeu_si128((__m128i *)running_avg_y, v_running_avg_y);
+      // The delta is set by the excess of absolute pixel diff over the
+      // threshold.
+      int delta = ((abs_sum_diff - sum_diff_thresh) >> 8) + 1;
+      // Only apply the adjustment for max delta up to 3.
+      if (delta < 4) {
+        const __m128i k_delta = _mm_set1_epi8(delta);
+        sig -= sig_stride * 16;
+        mc_running_avg_y -= mc_avg_y_stride * 16;
+        running_avg_y -= avg_y_stride * 16;
+        for (r = 0; r < 16; ++r) {
+          __m128i v_running_avg_y =
+              _mm_loadu_si128((__m128i *)(&running_avg_y[0]));
+          // Calculate differences.
+          const __m128i v_sig = _mm_loadu_si128((__m128i *)(&sig[0]));
+          const __m128i v_mc_running_avg_y =
+              _mm_loadu_si128((__m128i *)(&mc_running_avg_y[0]));
+          const __m128i pdiff = _mm_subs_epu8(v_mc_running_avg_y, v_sig);
+          const __m128i ndiff = _mm_subs_epu8(v_sig, v_mc_running_avg_y);
+          // Obtain the sign. FF if diff is negative.
+          const __m128i diff_sign = _mm_cmpeq_epi8(pdiff, k_0);
+          // Clamp absolute difference to delta to get the adjustment.
+          const __m128i adj = _mm_min_epu8(_mm_or_si128(pdiff, ndiff), k_delta);
+          // Restore the sign and get positive and negative adjustments.
+          __m128i padj, nadj;
+          padj = _mm_andnot_si128(diff_sign, adj);
+          nadj = _mm_and_si128(diff_sign, adj);
+          // Calculate filtered value.
+          v_running_avg_y = _mm_subs_epu8(v_running_avg_y, padj);
+          v_running_avg_y = _mm_adds_epu8(v_running_avg_y, nadj);
+          _mm_storeu_si128((__m128i *)running_avg_y, v_running_avg_y);
 
-             // Accumulate the adjustments.
-             acc_diff = _mm_subs_epi8(acc_diff, padj);
-             acc_diff = _mm_adds_epi8(acc_diff, nadj);
+          // Accumulate the adjustments.
+          acc_diff = _mm_subs_epi8(acc_diff, padj);
+          acc_diff = _mm_adds_epi8(acc_diff, nadj);
 
-             // Update pointers for next iteration.
-             sig += sig_stride;
-             mc_running_avg_y += mc_avg_y_stride;
-             running_avg_y += avg_y_stride;
-            }
-            abs_sum_diff = abs_sum_diff_16x1(acc_diff);
-            if (abs_sum_diff > sum_diff_thresh) {
-              return COPY_BLOCK;
-            }
-          } else {
-            return COPY_BLOCK;
-          }
+          // Update pointers for next iteration.
+          sig += sig_stride;
+          mc_running_avg_y += mc_avg_y_stride;
+          running_avg_y += avg_y_stride;
         }
+        abs_sum_diff = abs_sum_diff_16x1(acc_diff);
+        if (abs_sum_diff > sum_diff_thresh) {
+          return COPY_BLOCK;
+        }
+      } else {
+        return COPY_BLOCK;
+      }
     }
+  }
 
-    vp8_copy_mem16x16(running_avg_y_start, avg_y_stride, sig_start, sig_stride);
-    return FILTER_BLOCK;
+  vp8_copy_mem16x16(running_avg_y_start, avg_y_stride, sig_start, sig_stride);
+  return FILTER_BLOCK;
 }
 
 int vp8_denoiser_filter_uv_sse2(unsigned char *mc_running_avg,
-                             int mc_avg_stride,
-                             unsigned char *running_avg, int avg_stride,
-                             unsigned char *sig, int sig_stride,
-                             unsigned int motion_magnitude,
-                             int increase_denoising) {
-    unsigned char *running_avg_start = running_avg;
-    unsigned char *sig_start = sig;
-    unsigned int sum_diff_thresh;
-    int r;
-    int shift_inc  = (increase_denoising &&
-        motion_magnitude <= MOTION_MAGNITUDE_THRESHOLD_UV) ? 1 : 0;
-    __m128i acc_diff = _mm_setzero_si128();
-    const __m128i k_0 = _mm_setzero_si128();
-    const __m128i k_4 = _mm_set1_epi8(4 + shift_inc);
-    const __m128i k_8 = _mm_set1_epi8(8);
-    const __m128i k_16 = _mm_set1_epi8(16);
-    /* Modify each level's adjustment according to motion_magnitude. */
-    const __m128i l3 = _mm_set1_epi8(
-                       (motion_magnitude <= MOTION_MAGNITUDE_THRESHOLD_UV) ?
-                        7 + shift_inc : 6);
-    /* Difference between level 3 and level 2 is 2. */
-    const __m128i l32 = _mm_set1_epi8(2);
-    /* Difference between level 2 and level 1 is 1. */
-    const __m128i l21 = _mm_set1_epi8(1);
+                                int mc_avg_stride, unsigned char *running_avg,
+                                int avg_stride, unsigned char *sig,
+                                int sig_stride, unsigned int motion_magnitude,
+                                int increase_denoising) {
+  unsigned char *running_avg_start = running_avg;
+  unsigned char *sig_start = sig;
+  unsigned int sum_diff_thresh;
+  int r;
+  int shift_inc =
+      (increase_denoising && motion_magnitude <= MOTION_MAGNITUDE_THRESHOLD_UV)
+          ? 1
+          : 0;
+  __m128i acc_diff = _mm_setzero_si128();
+  const __m128i k_0 = _mm_setzero_si128();
+  const __m128i k_4 = _mm_set1_epi8(4 + shift_inc);
+  const __m128i k_8 = _mm_set1_epi8(8);
+  const __m128i k_16 = _mm_set1_epi8(16);
+  /* Modify each level's adjustment according to motion_magnitude. */
+  const __m128i l3 = _mm_set1_epi8(
+      (motion_magnitude <= MOTION_MAGNITUDE_THRESHOLD_UV) ? 7 + shift_inc : 6);
+  /* Difference between level 3 and level 2 is 2. */
+  const __m128i l32 = _mm_set1_epi8(2);
+  /* Difference between level 2 and level 1 is 1. */
+  const __m128i l21 = _mm_set1_epi8(1);
 
-    {
-      const __m128i k_1 = _mm_set1_epi16(1);
-      __m128i vec_sum_block = _mm_setzero_si128();
+  {
+    const __m128i k_1 = _mm_set1_epi16(1);
+    __m128i vec_sum_block = _mm_setzero_si128();
 
-      // Avoid denoising color signal if its close to average level.
-      for (r = 0; r < 8; ++r) {
-        const __m128i v_sig = _mm_loadl_epi64((__m128i *)(&sig[0]));
-        const __m128i v_sig_unpack = _mm_unpacklo_epi8(v_sig, k_0);
-        vec_sum_block = _mm_add_epi16(vec_sum_block, v_sig_unpack);
-        sig += sig_stride;
+    // Avoid denoising color signal if its close to average level.
+    for (r = 0; r < 8; ++r) {
+      const __m128i v_sig = _mm_loadl_epi64((__m128i *)(&sig[0]));
+      const __m128i v_sig_unpack = _mm_unpacklo_epi8(v_sig, k_0);
+      vec_sum_block = _mm_add_epi16(vec_sum_block, v_sig_unpack);
+      sig += sig_stride;
+    }
+    sig -= sig_stride * 8;
+    {
+      const __m128i hg_fe_dc_ba = _mm_madd_epi16(vec_sum_block, k_1);
+      const __m128i hgfe_dcba =
+          _mm_add_epi32(hg_fe_dc_ba, _mm_srli_si128(hg_fe_dc_ba, 8));
+      const __m128i hgfedcba =
+          _mm_add_epi32(hgfe_dcba, _mm_srli_si128(hgfe_dcba, 4));
+      const int sum_block = _mm_cvtsi128_si32(hgfedcba);
+      if (abs(sum_block - (128 * 8 * 8)) < SUM_DIFF_FROM_AVG_THRESH_UV) {
+        return COPY_BLOCK;
       }
-      sig -= sig_stride * 8;
-      {
-        const __m128i hg_fe_dc_ba = _mm_madd_epi16(vec_sum_block, k_1);
-        const __m128i hgfe_dcba = _mm_add_epi32(hg_fe_dc_ba,
-                                                _mm_srli_si128(hg_fe_dc_ba, 8));
-        const __m128i hgfedcba = _mm_add_epi32(hgfe_dcba,
-                                               _mm_srli_si128(hgfe_dcba, 4));
-        const int sum_block = _mm_cvtsi128_si32(hgfedcba);
-        if (abs(sum_block - (128 * 8 * 8)) < SUM_DIFF_FROM_AVG_THRESH_UV) {
-          return COPY_BLOCK;
-        }
-      }
     }
+  }
 
-    for (r = 0; r < 4; ++r) {
-        /* Calculate differences */
-        const __m128i v_sig_low = _mm_castpd_si128(
-            _mm_load_sd((double *)(&sig[0])));
-        const __m128i v_sig = _mm_castpd_si128(
-            _mm_loadh_pd(_mm_castsi128_pd(v_sig_low),
-                         (double *)(&sig[sig_stride])));
-        const __m128i v_mc_running_avg_low = _mm_castpd_si128(
-            _mm_load_sd((double *)(&mc_running_avg[0])));
-        const __m128i v_mc_running_avg = _mm_castpd_si128(
-            _mm_loadh_pd(_mm_castsi128_pd(v_mc_running_avg_low),
-                         (double *)(&mc_running_avg[mc_avg_stride])));
-        const __m128i pdiff = _mm_subs_epu8(v_mc_running_avg, v_sig);
-        const __m128i ndiff = _mm_subs_epu8(v_sig, v_mc_running_avg);
-        /* Obtain the sign. FF if diff is negative. */
-        const __m128i diff_sign = _mm_cmpeq_epi8(pdiff, k_0);
-        /* Clamp absolute difference to 16 to be used to get mask. Doing this
-         * allows us to use _mm_cmpgt_epi8, which operates on signed byte. */
-        const __m128i clamped_absdiff = _mm_min_epu8(
-                                        _mm_or_si128(pdiff, ndiff), k_16);
-        /* Get masks for l2 l1 and l0 adjustments */
-        const __m128i mask2 = _mm_cmpgt_epi8(k_16, clamped_absdiff);
-        const __m128i mask1 = _mm_cmpgt_epi8(k_8, clamped_absdiff);
-        const __m128i mask0 = _mm_cmpgt_epi8(k_4, clamped_absdiff);
-        /* Get adjustments for l2, l1, and l0 */
-        __m128i adj2 = _mm_and_si128(mask2, l32);
-        const __m128i adj1 = _mm_and_si128(mask1, l21);
-        const __m128i adj0 = _mm_and_si128(mask0, clamped_absdiff);
-        __m128i adj,  padj, nadj;
-        __m128i v_running_avg;
+  for (r = 0; r < 4; ++r) {
+    /* Calculate differences */
+    const __m128i v_sig_low =
+        _mm_castpd_si128(_mm_load_sd((double *)(&sig[0])));
+    const __m128i v_sig = _mm_castpd_si128(_mm_loadh_pd(
+        _mm_castsi128_pd(v_sig_low), (double *)(&sig[sig_stride])));
+    const __m128i v_mc_running_avg_low =
+        _mm_castpd_si128(_mm_load_sd((double *)(&mc_running_avg[0])));
+    const __m128i v_mc_running_avg = _mm_castpd_si128(
+        _mm_loadh_pd(_mm_castsi128_pd(v_mc_running_avg_low),
+                     (double *)(&mc_running_avg[mc_avg_stride])));
+    const __m128i pdiff = _mm_subs_epu8(v_mc_running_avg, v_sig);
+    const __m128i ndiff = _mm_subs_epu8(v_sig, v_mc_running_avg);
+    /* Obtain the sign. FF if diff is negative. */
+    const __m128i diff_sign = _mm_cmpeq_epi8(pdiff, k_0);
+    /* Clamp absolute difference to 16 to be used to get mask. Doing this
+     * allows us to use _mm_cmpgt_epi8, which operates on signed byte. */
+    const __m128i clamped_absdiff =
+        _mm_min_epu8(_mm_or_si128(pdiff, ndiff), k_16);
+    /* Get masks for l2 l1 and l0 adjustments */
+    const __m128i mask2 = _mm_cmpgt_epi8(k_16, clamped_absdiff);
+    const __m128i mask1 = _mm_cmpgt_epi8(k_8, clamped_absdiff);
+    const __m128i mask0 = _mm_cmpgt_epi8(k_4, clamped_absdiff);
+    /* Get adjustments for l2, l1, and l0 */
+    __m128i adj2 = _mm_and_si128(mask2, l32);
+    const __m128i adj1 = _mm_and_si128(mask1, l21);
+    const __m128i adj0 = _mm_and_si128(mask0, clamped_absdiff);
+    __m128i adj, padj, nadj;
+    __m128i v_running_avg;
 
-        /* Combine the adjustments and get absolute adjustments. */
-        adj2 = _mm_add_epi8(adj2, adj1);
-        adj = _mm_sub_epi8(l3, adj2);
-        adj = _mm_andnot_si128(mask0, adj);
-        adj = _mm_or_si128(adj, adj0);
+    /* Combine the adjustments and get absolute adjustments. */
+    adj2 = _mm_add_epi8(adj2, adj1);
+    adj = _mm_sub_epi8(l3, adj2);
+    adj = _mm_andnot_si128(mask0, adj);
+    adj = _mm_or_si128(adj, adj0);
 
-        /* Restore the sign and get positive and negative adjustments. */
-        padj = _mm_andnot_si128(diff_sign, adj);
-        nadj = _mm_and_si128(diff_sign, adj);
+    /* Restore the sign and get positive and negative adjustments. */
+    padj = _mm_andnot_si128(diff_sign, adj);
+    nadj = _mm_and_si128(diff_sign, adj);
 
-        /* Calculate filtered value. */
-        v_running_avg = _mm_adds_epu8(v_sig, padj);
-        v_running_avg = _mm_subs_epu8(v_running_avg, nadj);
+    /* Calculate filtered value. */
+    v_running_avg = _mm_adds_epu8(v_sig, padj);
+    v_running_avg = _mm_subs_epu8(v_running_avg, nadj);
 
-        _mm_storel_pd((double *)&running_avg[0],
-                      _mm_castsi128_pd(v_running_avg));
-        _mm_storeh_pd((double *)&running_avg[avg_stride],
-                      _mm_castsi128_pd(v_running_avg));
+    _mm_storel_pd((double *)&running_avg[0], _mm_castsi128_pd(v_running_avg));
+    _mm_storeh_pd((double *)&running_avg[avg_stride],
+                  _mm_castsi128_pd(v_running_avg));
 
-        /* Adjustments <=7, and each element in acc_diff can fit in signed
-         * char.
-         */
-        acc_diff = _mm_adds_epi8(acc_diff, padj);
-        acc_diff = _mm_subs_epi8(acc_diff, nadj);
+    /* Adjustments <=7, and each element in acc_diff can fit in signed
+     * char.
+     */
+    acc_diff = _mm_adds_epi8(acc_diff, padj);
+    acc_diff = _mm_subs_epi8(acc_diff, nadj);
 
-        /* Update pointers for next iteration. */
-        sig += sig_stride * 2;
-        mc_running_avg += mc_avg_stride * 2;
-        running_avg += avg_stride * 2;
-    }
+    /* Update pointers for next iteration. */
+    sig += sig_stride * 2;
+    mc_running_avg += mc_avg_stride * 2;
+    running_avg += avg_stride * 2;
+  }
 
-    {
-        unsigned int abs_sum_diff = abs_sum_diff_16x1(acc_diff);
-        sum_diff_thresh = SUM_DIFF_THRESHOLD_UV;
-        if (increase_denoising) sum_diff_thresh = SUM_DIFF_THRESHOLD_HIGH_UV;
-        if (abs_sum_diff > sum_diff_thresh) {
-          // Before returning to copy the block (i.e., apply no denoising),
-          // check if we can still apply some (weaker) temporal filtering to
-          // this block, that would otherwise not be denoised at all. Simplest
-          // is to apply an additional adjustment to running_avg_y to bring it
-          // closer to sig. The adjustment is capped by a maximum delta, and
-          // chosen such that in most cases the resulting sum_diff will be
-          // within the acceptable range given by sum_diff_thresh.
+  {
+    unsigned int abs_sum_diff = abs_sum_diff_16x1(acc_diff);
+    sum_diff_thresh = SUM_DIFF_THRESHOLD_UV;
+    if (increase_denoising) sum_diff_thresh = SUM_DIFF_THRESHOLD_HIGH_UV;
+    if (abs_sum_diff > sum_diff_thresh) {
+      // Before returning to copy the block (i.e., apply no denoising),
+      // check if we can still apply some (weaker) temporal filtering to
+      // this block, that would otherwise not be denoised at all. Simplest
+      // is to apply an additional adjustment to running_avg_y to bring it
+      // closer to sig. The adjustment is capped by a maximum delta, and
+      // chosen such that in most cases the resulting sum_diff will be
+      // within the acceptable range given by sum_diff_thresh.
 
-          // The delta is set by the excess of absolute pixel diff over the
-          // threshold.
-          int delta = ((abs_sum_diff - sum_diff_thresh) >> 8) + 1;
-          // Only apply the adjustment for max delta up to 3.
-          if (delta < 4) {
-            const __m128i k_delta = _mm_set1_epi8(delta);
-            sig -= sig_stride * 8;
-            mc_running_avg -= mc_avg_stride * 8;
-            running_avg -= avg_stride * 8;
-            for (r = 0; r < 4; ++r) {
-              // Calculate differences.
-              const __m128i v_sig_low = _mm_castpd_si128(
-                  _mm_load_sd((double *)(&sig[0])));
-              const __m128i v_sig = _mm_castpd_si128(
-                  _mm_loadh_pd(_mm_castsi128_pd(v_sig_low),
-                               (double *)(&sig[sig_stride])));
-              const __m128i v_mc_running_avg_low = _mm_castpd_si128(
-                  _mm_load_sd((double *)(&mc_running_avg[0])));
-              const __m128i v_mc_running_avg = _mm_castpd_si128(
-                  _mm_loadh_pd(_mm_castsi128_pd(v_mc_running_avg_low),
-                               (double *)(&mc_running_avg[mc_avg_stride])));
-              const __m128i pdiff = _mm_subs_epu8(v_mc_running_avg, v_sig);
-              const __m128i ndiff = _mm_subs_epu8(v_sig, v_mc_running_avg);
-              // Obtain the sign. FF if diff is negative.
-              const __m128i diff_sign = _mm_cmpeq_epi8(pdiff, k_0);
-              // Clamp absolute difference to delta to get the adjustment.
-              const __m128i adj =
-                  _mm_min_epu8(_mm_or_si128(pdiff, ndiff), k_delta);
-              // Restore the sign and get positive and negative adjustments.
-              __m128i padj, nadj;
-              const __m128i v_running_avg_low = _mm_castpd_si128(
-                  _mm_load_sd((double *)(&running_avg[0])));
-              __m128i v_running_avg = _mm_castpd_si128(
-                  _mm_loadh_pd(_mm_castsi128_pd(v_running_avg_low),
-                               (double *)(&running_avg[avg_stride])));
-              padj = _mm_andnot_si128(diff_sign, adj);
-              nadj = _mm_and_si128(diff_sign, adj);
-              // Calculate filtered value.
-              v_running_avg = _mm_subs_epu8(v_running_avg, padj);
-              v_running_avg = _mm_adds_epu8(v_running_avg, nadj);
+      // The delta is set by the excess of absolute pixel diff over the
+      // threshold.
+      int delta = ((abs_sum_diff - sum_diff_thresh) >> 8) + 1;
+      // Only apply the adjustment for max delta up to 3.
+      if (delta < 4) {
+        const __m128i k_delta = _mm_set1_epi8(delta);
+        sig -= sig_stride * 8;
+        mc_running_avg -= mc_avg_stride * 8;
+        running_avg -= avg_stride * 8;
+        for (r = 0; r < 4; ++r) {
+          // Calculate differences.
+          const __m128i v_sig_low =
+              _mm_castpd_si128(_mm_load_sd((double *)(&sig[0])));
+          const __m128i v_sig = _mm_castpd_si128(_mm_loadh_pd(
+              _mm_castsi128_pd(v_sig_low), (double *)(&sig[sig_stride])));
+          const __m128i v_mc_running_avg_low =
+              _mm_castpd_si128(_mm_load_sd((double *)(&mc_running_avg[0])));
+          const __m128i v_mc_running_avg = _mm_castpd_si128(
+              _mm_loadh_pd(_mm_castsi128_pd(v_mc_running_avg_low),
+                           (double *)(&mc_running_avg[mc_avg_stride])));
+          const __m128i pdiff = _mm_subs_epu8(v_mc_running_avg, v_sig);
+          const __m128i ndiff = _mm_subs_epu8(v_sig, v_mc_running_avg);
+          // Obtain the sign. FF if diff is negative.
+          const __m128i diff_sign = _mm_cmpeq_epi8(pdiff, k_0);
+          // Clamp absolute difference to delta to get the adjustment.
+          const __m128i adj = _mm_min_epu8(_mm_or_si128(pdiff, ndiff), k_delta);
+          // Restore the sign and get positive and negative adjustments.
+          __m128i padj, nadj;
+          const __m128i v_running_avg_low =
+              _mm_castpd_si128(_mm_load_sd((double *)(&running_avg[0])));
+          __m128i v_running_avg = _mm_castpd_si128(
+              _mm_loadh_pd(_mm_castsi128_pd(v_running_avg_low),
+                           (double *)(&running_avg[avg_stride])));
+          padj = _mm_andnot_si128(diff_sign, adj);
+          nadj = _mm_and_si128(diff_sign, adj);
+          // Calculate filtered value.
+          v_running_avg = _mm_subs_epu8(v_running_avg, padj);
+          v_running_avg = _mm_adds_epu8(v_running_avg, nadj);
 
-              _mm_storel_pd((double *)&running_avg[0],
-                            _mm_castsi128_pd(v_running_avg));
-              _mm_storeh_pd((double *)&running_avg[avg_stride],
-                            _mm_castsi128_pd(v_running_avg));
+          _mm_storel_pd((double *)&running_avg[0],
+                        _mm_castsi128_pd(v_running_avg));
+          _mm_storeh_pd((double *)&running_avg[avg_stride],
+                        _mm_castsi128_pd(v_running_avg));
 
-             // Accumulate the adjustments.
-             acc_diff = _mm_subs_epi8(acc_diff, padj);
-             acc_diff = _mm_adds_epi8(acc_diff, nadj);
+          // Accumulate the adjustments.
+          acc_diff = _mm_subs_epi8(acc_diff, padj);
+          acc_diff = _mm_adds_epi8(acc_diff, nadj);
 
-             // Update pointers for next iteration.
-             sig += sig_stride * 2;
-             mc_running_avg += mc_avg_stride * 2;
-             running_avg += avg_stride * 2;
-            }
-            abs_sum_diff = abs_sum_diff_16x1(acc_diff);
-            if (abs_sum_diff > sum_diff_thresh) {
-              return COPY_BLOCK;
-            }
-          } else {
-            return COPY_BLOCK;
-          }
+          // Update pointers for next iteration.
+          sig += sig_stride * 2;
+          mc_running_avg += mc_avg_stride * 2;
+          running_avg += avg_stride * 2;
         }
+        abs_sum_diff = abs_sum_diff_16x1(acc_diff);
+        if (abs_sum_diff > sum_diff_thresh) {
+          return COPY_BLOCK;
+        }
+      } else {
+        return COPY_BLOCK;
+      }
     }
+  }
 
-    vp8_copy_mem8x8(running_avg_start, avg_stride, sig_start, sig_stride);
-    return FILTER_BLOCK;
+  vp8_copy_mem8x8(running_avg_start, avg_stride, sig_start, sig_stride);
+  return FILTER_BLOCK;
 }
--- a/vp8/encoder/x86/quantize_sse4.c
+++ b/vp8/encoder/x86/quantize_sse4.c
@@ -8,7 +8,6 @@
  *  be found in the AUTHORS file in the root of the source tree.
  */
 
-
 #include <smmintrin.h> /* SSE4.1 */
 
 #include "./vp8_rtcd.h"
@@ -15,114 +14,113 @@
 #include "vp8/encoder/block.h"
 #include "vp8/common/entropy.h" /* vp8_default_inv_zig_zag */
 
-#define SELECT_EOB(i, z, x, y, q) \
-    do { \
-        short boost = *zbin_boost_ptr; \
-        short x_z = _mm_extract_epi16(x, z); \
-        short y_z = _mm_extract_epi16(y, z); \
-        int cmp = (x_z < boost) | (y_z == 0); \
-        zbin_boost_ptr++; \
-        if (cmp) \
-            break; \
-        q = _mm_insert_epi16(q, y_z, z); \
-        eob = i; \
-        zbin_boost_ptr = b->zrun_zbin_boost; \
-    } while (0)
+#define SELECT_EOB(i, z, x, y, q)         \
+  do {                                    \
+    short boost = *zbin_boost_ptr;        \
+    short x_z = _mm_extract_epi16(x, z);  \
+    short y_z = _mm_extract_epi16(y, z);  \
+    int cmp = (x_z < boost) | (y_z == 0); \
+    zbin_boost_ptr++;                     \
+    if (cmp) break;                       \
+    q = _mm_insert_epi16(q, y_z, z);      \
+    eob = i;                              \
+    zbin_boost_ptr = b->zrun_zbin_boost;  \
+  } while (0)
 
 void vp8_regular_quantize_b_sse4_1(BLOCK *b, BLOCKD *d) {
-    char eob = 0;
-    short *zbin_boost_ptr  = b->zrun_zbin_boost;
+  char eob = 0;
+  short *zbin_boost_ptr = b->zrun_zbin_boost;
 
-    __m128i sz0, x0, sz1, x1, y0, y1, x_minus_zbin0, x_minus_zbin1,
-            dqcoeff0, dqcoeff1;
-    __m128i quant_shift0 = _mm_load_si128((__m128i *)(b->quant_shift));
-    __m128i quant_shift1 = _mm_load_si128((__m128i *)(b->quant_shift + 8));
-    __m128i z0 = _mm_load_si128((__m128i *)(b->coeff));
-    __m128i z1 = _mm_load_si128((__m128i *)(b->coeff+8));
-    __m128i zbin_extra = _mm_cvtsi32_si128(b->zbin_extra);
-    __m128i zbin0 = _mm_load_si128((__m128i *)(b->zbin));
-    __m128i zbin1 = _mm_load_si128((__m128i *)(b->zbin + 8));
-    __m128i round0 = _mm_load_si128((__m128i *)(b->round));
-    __m128i round1 = _mm_load_si128((__m128i *)(b->round + 8));
-    __m128i quant0 = _mm_load_si128((__m128i *)(b->quant));
-    __m128i quant1 = _mm_load_si128((__m128i *)(b->quant + 8));
-    __m128i dequant0 = _mm_load_si128((__m128i *)(d->dequant));
-    __m128i dequant1 = _mm_load_si128((__m128i *)(d->dequant + 8));
-    __m128i qcoeff0 = _mm_setzero_si128();
-    __m128i qcoeff1 = _mm_setzero_si128();
+  __m128i sz0, x0, sz1, x1, y0, y1, x_minus_zbin0, x_minus_zbin1, dqcoeff0,
+      dqcoeff1;
+  __m128i quant_shift0 = _mm_load_si128((__m128i *)(b->quant_shift));
+  __m128i quant_shift1 = _mm_load_si128((__m128i *)(b->quant_shift + 8));
+  __m128i z0 = _mm_load_si128((__m128i *)(b->coeff));
+  __m128i z1 = _mm_load_si128((__m128i *)(b->coeff + 8));
+  __m128i zbin_extra = _mm_cvtsi32_si128(b->zbin_extra);
+  __m128i zbin0 = _mm_load_si128((__m128i *)(b->zbin));
+  __m128i zbin1 = _mm_load_si128((__m128i *)(b->zbin + 8));
+  __m128i round0 = _mm_load_si128((__m128i *)(b->round));
+  __m128i round1 = _mm_load_si128((__m128i *)(b->round + 8));
+  __m128i quant0 = _mm_load_si128((__m128i *)(b->quant));
+  __m128i quant1 = _mm_load_si128((__m128i *)(b->quant + 8));
+  __m128i dequant0 = _mm_load_si128((__m128i *)(d->dequant));
+  __m128i dequant1 = _mm_load_si128((__m128i *)(d->dequant + 8));
+  __m128i qcoeff0 = _mm_setzero_si128();
+  __m128i qcoeff1 = _mm_setzero_si128();
 
-    /* Duplicate to all lanes. */
-    zbin_extra = _mm_shufflelo_epi16(zbin_extra, 0);
-    zbin_extra = _mm_unpacklo_epi16(zbin_extra, zbin_extra);
+  /* Duplicate to all lanes. */
+  zbin_extra = _mm_shufflelo_epi16(zbin_extra, 0);
+  zbin_extra = _mm_unpacklo_epi16(zbin_extra, zbin_extra);
 
-    /* Sign of z: z >> 15 */
-    sz0 = _mm_srai_epi16(z0, 15);
-    sz1 = _mm_srai_epi16(z1, 15);
+  /* Sign of z: z >> 15 */
+  sz0 = _mm_srai_epi16(z0, 15);
+  sz1 = _mm_srai_epi16(z1, 15);
 
-    /* x = abs(z): (z ^ sz) - sz */
-    x0 = _mm_xor_si128(z0, sz0);
-    x1 = _mm_xor_si128(z1, sz1);
-    x0 = _mm_sub_epi16(x0, sz0);
-    x1 = _mm_sub_epi16(x1, sz1);
+  /* x = abs(z): (z ^ sz) - sz */
+  x0 = _mm_xor_si128(z0, sz0);
+  x1 = _mm_xor_si128(z1, sz1);
+  x0 = _mm_sub_epi16(x0, sz0);
+  x1 = _mm_sub_epi16(x1, sz1);
 
-    /* zbin[] + zbin_extra */
-    zbin0 = _mm_add_epi16(zbin0, zbin_extra);
-    zbin1 = _mm_add_epi16(zbin1, zbin_extra);
+  /* zbin[] + zbin_extra */
+  zbin0 = _mm_add_epi16(zbin0, zbin_extra);
+  zbin1 = _mm_add_epi16(zbin1, zbin_extra);
 
-    /* In C x is compared to zbin where zbin = zbin[] + boost + extra. Rebalance
-     * the equation because boost is the only value which can change:
-     * x - (zbin[] + extra) >= boost */
-    x_minus_zbin0 = _mm_sub_epi16(x0, zbin0);
-    x_minus_zbin1 = _mm_sub_epi16(x1, zbin1);
+  /* In C x is compared to zbin where zbin = zbin[] + boost + extra. Rebalance
+   * the equation because boost is the only value which can change:
+   * x - (zbin[] + extra) >= boost */
+  x_minus_zbin0 = _mm_sub_epi16(x0, zbin0);
+  x_minus_zbin1 = _mm_sub_epi16(x1, zbin1);
 
-    /* All the remaining calculations are valid whether they are done now with
-     * simd or later inside the loop one at a time. */
-    x0 = _mm_add_epi16(x0, round0);
-    x1 = _mm_add_epi16(x1, round1);
+  /* All the remaining calculations are valid whether they are done now with
+   * simd or later inside the loop one at a time. */
+  x0 = _mm_add_epi16(x0, round0);
+  x1 = _mm_add_epi16(x1, round1);
 
-    y0 = _mm_mulhi_epi16(x0, quant0);
-    y1 = _mm_mulhi_epi16(x1, quant1);
+  y0 = _mm_mulhi_epi16(x0, quant0);
+  y1 = _mm_mulhi_epi16(x1, quant1);
 
-    y0 = _mm_add_epi16(y0, x0);
-    y1 = _mm_add_epi16(y1, x1);
+  y0 = _mm_add_epi16(y0, x0);
+  y1 = _mm_add_epi16(y1, x1);
 
-    /* Instead of shifting each value independently we convert the scaling
-     * factor with 1 << (16 - shift) so we can use multiply/return high half. */
-    y0 = _mm_mulhi_epi16(y0, quant_shift0);
-    y1 = _mm_mulhi_epi16(y1, quant_shift1);
+  /* Instead of shifting each value independently we convert the scaling
+   * factor with 1 << (16 - shift) so we can use multiply/return high half. */
+  y0 = _mm_mulhi_epi16(y0, quant_shift0);
+  y1 = _mm_mulhi_epi16(y1, quant_shift1);
 
-    /* Return the sign: (y ^ sz) - sz */
-    y0 = _mm_xor_si128(y0, sz0);
-    y1 = _mm_xor_si128(y1, sz1);
-    y0 = _mm_sub_epi16(y0, sz0);
-    y1 = _mm_sub_epi16(y1, sz1);
+  /* Return the sign: (y ^ sz) - sz */
+  y0 = _mm_xor_si128(y0, sz0);
+  y1 = _mm_xor_si128(y1, sz1);
+  y0 = _mm_sub_epi16(y0, sz0);
+  y1 = _mm_sub_epi16(y1, sz1);
 
-    /* The loop gets unrolled anyway. Avoid the vp8_default_zig_zag1d lookup. */
-    SELECT_EOB(1, 0, x_minus_zbin0, y0, qcoeff0);
-    SELECT_EOB(2, 1, x_minus_zbin0, y0, qcoeff0);
-    SELECT_EOB(3, 4, x_minus_zbin0, y0, qcoeff0);
-    SELECT_EOB(4, 0, x_minus_zbin1, y1, qcoeff1);
-    SELECT_EOB(5, 5, x_minus_zbin0, y0, qcoeff0);
-    SELECT_EOB(6, 2, x_minus_zbin0, y0, qcoeff0);
-    SELECT_EOB(7, 3, x_minus_zbin0, y0, qcoeff0);
-    SELECT_EOB(8, 6, x_minus_zbin0, y0, qcoeff0);
-    SELECT_EOB(9, 1, x_minus_zbin1, y1, qcoeff1);
-    SELECT_EOB(10, 4, x_minus_zbin1, y1, qcoeff1);
-    SELECT_EOB(11, 5, x_minus_zbin1, y1, qcoeff1);
-    SELECT_EOB(12, 2, x_minus_zbin1, y1, qcoeff1);
-    SELECT_EOB(13, 7, x_minus_zbin0, y0, qcoeff0);
-    SELECT_EOB(14, 3, x_minus_zbin1, y1, qcoeff1);
-    SELECT_EOB(15, 6, x_minus_zbin1, y1, qcoeff1);
-    SELECT_EOB(16, 7, x_minus_zbin1, y1, qcoeff1);
+  /* The loop gets unrolled anyway. Avoid the vp8_default_zig_zag1d lookup. */
+  SELECT_EOB(1, 0, x_minus_zbin0, y0, qcoeff0);
+  SELECT_EOB(2, 1, x_minus_zbin0, y0, qcoeff0);
+  SELECT_EOB(3, 4, x_minus_zbin0, y0, qcoeff0);
+  SELECT_EOB(4, 0, x_minus_zbin1, y1, qcoeff1);
+  SELECT_EOB(5, 5, x_minus_zbin0, y0, qcoeff0);
+  SELECT_EOB(6, 2, x_minus_zbin0, y0, qcoeff0);
+  SELECT_EOB(7, 3, x_minus_zbin0, y0, qcoeff0);
+  SELECT_EOB(8, 6, x_minus_zbin0, y0, qcoeff0);
+  SELECT_EOB(9, 1, x_minus_zbin1, y1, qcoeff1);
+  SELECT_EOB(10, 4, x_minus_zbin1, y1, qcoeff1);
+  SELECT_EOB(11, 5, x_minus_zbin1, y1, qcoeff1);
+  SELECT_EOB(12, 2, x_minus_zbin1, y1, qcoeff1);
+  SELECT_EOB(13, 7, x_minus_zbin0, y0, qcoeff0);
+  SELECT_EOB(14, 3, x_minus_zbin1, y1, qcoeff1);
+  SELECT_EOB(15, 6, x_minus_zbin1, y1, qcoeff1);
+  SELECT_EOB(16, 7, x_minus_zbin1, y1, qcoeff1);
 
-    _mm_store_si128((__m128i *)(d->qcoeff), qcoeff0);
-    _mm_store_si128((__m128i *)(d->qcoeff + 8), qcoeff1);
+  _mm_store_si128((__m128i *)(d->qcoeff), qcoeff0);
+  _mm_store_si128((__m128i *)(d->qcoeff + 8), qcoeff1);
 
-    dqcoeff0 = _mm_mullo_epi16(qcoeff0, dequant0);
-    dqcoeff1 = _mm_mullo_epi16(qcoeff1, dequant1);
+  dqcoeff0 = _mm_mullo_epi16(qcoeff0, dequant0);
+  dqcoeff1 = _mm_mullo_epi16(qcoeff1, dequant1);
 
-    _mm_store_si128((__m128i *)(d->dqcoeff), dqcoeff0);
-    _mm_store_si128((__m128i *)(d->dqcoeff + 8), dqcoeff1);
+  _mm_store_si128((__m128i *)(d->dqcoeff), dqcoeff0);
+  _mm_store_si128((__m128i *)(d->dqcoeff + 8), dqcoeff1);
 
-    *d->eob = eob;
+  *d->eob = eob;
 }
--- a/vp8/encoder/x86/quantize_ssse3.c
+++ b/vp8/encoder/x86/quantize_ssse3.c
@@ -20,8 +20,7 @@
   unsigned long eob;
   _BitScanReverse(&eob, mask);
   eob++;
-  if (mask == 0)
-    eob = 0;
+  if (mask == 0) eob = 0;
   return eob;
 }
 #else
@@ -28,13 +27,12 @@
 static int bsr(int mask) {
   int eob;
 #if defined(__GNUC__) && __GNUC__
-  __asm__ __volatile__("bsr %1, %0" : "=r" (eob) : "r" (mask) : "flags");
+  __asm__ __volatile__("bsr %1, %0" : "=r"(eob) : "r"(mask) : "flags");
 #elif defined(__SUNPRO_C) || defined(__SUNPRO_CC)
-  asm volatile("bsr %1, %0" : "=r" (eob) : "r" (mask) : "flags");
+  asm volatile("bsr %1, %0" : "=r"(eob) : "r"(mask) : "flags");
 #endif
   eob++;
-  if (mask == 0)
-    eob = 0;
+  if (mask == 0) eob = 0;
   return eob;
 }
 #endif
@@ -53,8 +51,9 @@
 
   __m128i sz0, sz1, x, x0, x1, y0, y1, zeros, abs0, abs1;
 
-  DECLARE_ALIGNED(16, const uint8_t, pshufb_zig_zag_mask[16]) =
-    { 0, 1, 4, 8, 5, 2, 3, 6, 9, 12, 13, 10, 7, 11, 14, 15 };
+  DECLARE_ALIGNED(16, const uint8_t, pshufb_zig_zag_mask[16]) = {
+    0, 1, 4, 8, 5, 2, 3, 6, 9, 12, 13, 10, 7, 11, 14, 15
+  };
   __m128i zig_zag = _mm_load_si128((const __m128i *)pshufb_zig_zag_mask);
 
   /* sign of z: z >> 15 */
--- a/vp8/encoder/x86/vp8_enc_stubs_mmx.c
+++ b/vp8/encoder/x86/vp8_enc_stubs_mmx.c
@@ -8,7 +8,6 @@
  *  be found in the AUTHORS file in the root of the source tree.
  */
 
-
 #include "vpx_config.h"
 #include "vp8_rtcd.h"
 #include "vpx_ports/x86.h"
@@ -15,10 +14,9 @@
 #include "vp8/encoder/block.h"
 
 void vp8_short_fdct4x4_mmx(short *input, short *output, int pitch);
-void vp8_short_fdct8x4_mmx(short *input, short *output, int pitch)
-{
-    vp8_short_fdct4x4_mmx(input,   output,    pitch);
-    vp8_short_fdct4x4_mmx(input + 4, output + 16, pitch);
+void vp8_short_fdct8x4_mmx(short *input, short *output, int pitch) {
+  vp8_short_fdct4x4_mmx(input, output, pitch);
+  vp8_short_fdct4x4_mmx(input + 4, output + 16, pitch);
 }
 
 int vp8_fast_quantize_b_impl_mmx(short *coeff_ptr, short *zbin_ptr,
@@ -25,43 +23,32 @@
                                  short *qcoeff_ptr, short *dequant_ptr,
                                  const short *scan_mask, short *round_ptr,
                                  short *quant_ptr, short *dqcoeff_ptr);
-void vp8_fast_quantize_b_mmx(BLOCK *b, BLOCKD *d)
-{
-    const short *scan_mask   = vp8_default_zig_zag_mask;
-    short *coeff_ptr   = b->coeff;
-    short *zbin_ptr    = b->zbin;
-    short *round_ptr   = b->round;
-    short *quant_ptr   = b->quant_fast;
-    short *qcoeff_ptr  = d->qcoeff;
-    short *dqcoeff_ptr = d->dqcoeff;
-    short *dequant_ptr = d->dequant;
+void vp8_fast_quantize_b_mmx(BLOCK *b, BLOCKD *d) {
+  const short *scan_mask = vp8_default_zig_zag_mask;
+  short *coeff_ptr = b->coeff;
+  short *zbin_ptr = b->zbin;
+  short *round_ptr = b->round;
+  short *quant_ptr = b->quant_fast;
+  short *qcoeff_ptr = d->qcoeff;
+  short *dqcoeff_ptr = d->dqcoeff;
+  short *dequant_ptr = d->dequant;
 
-    *d->eob = (char)vp8_fast_quantize_b_impl_mmx(
-                                                 coeff_ptr,
-                                                 zbin_ptr,
-                                                 qcoeff_ptr,
-                                                 dequant_ptr,
-                                                 scan_mask,
+  *d->eob = (char)vp8_fast_quantize_b_impl_mmx(
+      coeff_ptr, zbin_ptr, qcoeff_ptr, dequant_ptr, scan_mask,
 
-                                                 round_ptr,
-                                                 quant_ptr,
-                                                 dqcoeff_ptr
-                                                 );
+      round_ptr, quant_ptr, dqcoeff_ptr);
 }
 
 int vp8_mbblock_error_mmx_impl(short *coeff_ptr, short *dcoef_ptr, int dc);
-int vp8_mbblock_error_mmx(MACROBLOCK *mb, int dc)
-{
-    short *coeff_ptr =  mb->block[0].coeff;
-    short *dcoef_ptr =  mb->e_mbd.block[0].dqcoeff;
-    return vp8_mbblock_error_mmx_impl(coeff_ptr, dcoef_ptr, dc);
+int vp8_mbblock_error_mmx(MACROBLOCK *mb, int dc) {
+  short *coeff_ptr = mb->block[0].coeff;
+  short *dcoef_ptr = mb->e_mbd.block[0].dqcoeff;
+  return vp8_mbblock_error_mmx_impl(coeff_ptr, dcoef_ptr, dc);
 }
 
 int vp8_mbuverror_mmx_impl(short *s_ptr, short *d_ptr);
-int vp8_mbuverror_mmx(MACROBLOCK *mb)
-{
-    short *s_ptr = &mb->coeff[256];
-    short *d_ptr = &mb->e_mbd.dqcoeff[256];
-    return vp8_mbuverror_mmx_impl(s_ptr, d_ptr);
+int vp8_mbuverror_mmx(MACROBLOCK *mb) {
+  short *s_ptr = &mb->coeff[256];
+  short *d_ptr = &mb->e_mbd.dqcoeff[256];
+  return vp8_mbuverror_mmx_impl(s_ptr, d_ptr);
 }
-
--- a/vp8/encoder/x86/vp8_enc_stubs_sse2.c
+++ b/vp8/encoder/x86/vp8_enc_stubs_sse2.c
@@ -8,7 +8,6 @@
  *  be found in the AUTHORS file in the root of the source tree.
  */
 
-
 #include "vpx_config.h"
 #include "vp8_rtcd.h"
 #include "vpx_ports/x86.h"
@@ -15,18 +14,15 @@
 #include "vp8/encoder/block.h"
 
 int vp8_mbblock_error_xmm_impl(short *coeff_ptr, short *dcoef_ptr, int dc);
-int vp8_mbblock_error_xmm(MACROBLOCK *mb, int dc)
-{
-    short *coeff_ptr =  mb->block[0].coeff;
-    short *dcoef_ptr =  mb->e_mbd.block[0].dqcoeff;
-    return vp8_mbblock_error_xmm_impl(coeff_ptr, dcoef_ptr, dc);
+int vp8_mbblock_error_xmm(MACROBLOCK *mb, int dc) {
+  short *coeff_ptr = mb->block[0].coeff;
+  short *dcoef_ptr = mb->e_mbd.block[0].dqcoeff;
+  return vp8_mbblock_error_xmm_impl(coeff_ptr, dcoef_ptr, dc);
 }
 
 int vp8_mbuverror_xmm_impl(short *s_ptr, short *d_ptr);
-int vp8_mbuverror_xmm(MACROBLOCK *mb)
-{
-    short *s_ptr = &mb->coeff[256];
-    short *d_ptr = &mb->e_mbd.dqcoeff[256];
-    return vp8_mbuverror_xmm_impl(s_ptr, d_ptr);
+int vp8_mbuverror_xmm(MACROBLOCK *mb) {
+  short *s_ptr = &mb->coeff[256];
+  short *d_ptr = &mb->e_mbd.dqcoeff[256];
+  return vp8_mbuverror_xmm_impl(s_ptr, d_ptr);
 }
-
--- a/vp8/encoder/x86/vp8_quantize_sse2.c
+++ b/vp8/encoder/x86/vp8_quantize_sse2.c
@@ -8,7 +8,6 @@
  *  be found in the AUTHORS file in the root of the source tree.
  */
 
-
 #include "vpx_config.h"
 #include "vp8_rtcd.h"
 #include "vpx_ports/x86.h"
@@ -16,134 +15,131 @@
 #include "vp8/encoder/block.h"
 #include "vp8/common/entropy.h" /* vp8_default_inv_zig_zag */
 
-#include <mmintrin.h> /* MMX */
+#include <mmintrin.h>  /* MMX */
 #include <xmmintrin.h> /* SSE */
 #include <emmintrin.h> /* SSE2 */
 
-#define SELECT_EOB(i, z) \
-    do { \
-        short boost = *zbin_boost_ptr; \
-        int cmp = (x[z] < boost) | (y[z] == 0); \
-        zbin_boost_ptr++; \
-        if (cmp) \
-            break; \
-        qcoeff_ptr[z] = y[z]; \
-        eob = i; \
-        zbin_boost_ptr = b->zrun_zbin_boost; \
-    } while (0)
+#define SELECT_EOB(i, z)                    \
+  do {                                      \
+    short boost = *zbin_boost_ptr;          \
+    int cmp = (x[z] < boost) | (y[z] == 0); \
+    zbin_boost_ptr++;                       \
+    if (cmp) break;                         \
+    qcoeff_ptr[z] = y[z];                   \
+    eob = i;                                \
+    zbin_boost_ptr = b->zrun_zbin_boost;    \
+  } while (0)
 
-void vp8_regular_quantize_b_sse2(BLOCK *b, BLOCKD *d)
-{
-    char eob = 0;
-    short *zbin_boost_ptr;
-    short *qcoeff_ptr      = d->qcoeff;
-    DECLARE_ALIGNED(16, short, x[16]);
-    DECLARE_ALIGNED(16, short, y[16]);
+void vp8_regular_quantize_b_sse2(BLOCK *b, BLOCKD *d) {
+  char eob = 0;
+  short *zbin_boost_ptr;
+  short *qcoeff_ptr = d->qcoeff;
+  DECLARE_ALIGNED(16, short, x[16]);
+  DECLARE_ALIGNED(16, short, y[16]);
 
-    __m128i sz0, x0, sz1, x1, y0, y1, x_minus_zbin0, x_minus_zbin1;
-    __m128i quant_shift0 = _mm_load_si128((__m128i *)(b->quant_shift));
-    __m128i quant_shift1 = _mm_load_si128((__m128i *)(b->quant_shift + 8));
-    __m128i z0 = _mm_load_si128((__m128i *)(b->coeff));
-    __m128i z1 = _mm_load_si128((__m128i *)(b->coeff+8));
-    __m128i zbin_extra = _mm_cvtsi32_si128(b->zbin_extra);
-    __m128i zbin0 = _mm_load_si128((__m128i *)(b->zbin));
-    __m128i zbin1 = _mm_load_si128((__m128i *)(b->zbin + 8));
-    __m128i round0 = _mm_load_si128((__m128i *)(b->round));
-    __m128i round1 = _mm_load_si128((__m128i *)(b->round + 8));
-    __m128i quant0 = _mm_load_si128((__m128i *)(b->quant));
-    __m128i quant1 = _mm_load_si128((__m128i *)(b->quant + 8));
-    __m128i dequant0 = _mm_load_si128((__m128i *)(d->dequant));
-    __m128i dequant1 = _mm_load_si128((__m128i *)(d->dequant + 8));
+  __m128i sz0, x0, sz1, x1, y0, y1, x_minus_zbin0, x_minus_zbin1;
+  __m128i quant_shift0 = _mm_load_si128((__m128i *)(b->quant_shift));
+  __m128i quant_shift1 = _mm_load_si128((__m128i *)(b->quant_shift + 8));
+  __m128i z0 = _mm_load_si128((__m128i *)(b->coeff));
+  __m128i z1 = _mm_load_si128((__m128i *)(b->coeff + 8));
+  __m128i zbin_extra = _mm_cvtsi32_si128(b->zbin_extra);
+  __m128i zbin0 = _mm_load_si128((__m128i *)(b->zbin));
+  __m128i zbin1 = _mm_load_si128((__m128i *)(b->zbin + 8));
+  __m128i round0 = _mm_load_si128((__m128i *)(b->round));
+  __m128i round1 = _mm_load_si128((__m128i *)(b->round + 8));
+  __m128i quant0 = _mm_load_si128((__m128i *)(b->quant));
+  __m128i quant1 = _mm_load_si128((__m128i *)(b->quant + 8));
+  __m128i dequant0 = _mm_load_si128((__m128i *)(d->dequant));
+  __m128i dequant1 = _mm_load_si128((__m128i *)(d->dequant + 8));
 
-    memset(qcoeff_ptr, 0, 32);
+  memset(qcoeff_ptr, 0, 32);
 
-    /* Duplicate to all lanes. */
-    zbin_extra = _mm_shufflelo_epi16(zbin_extra, 0);
-    zbin_extra = _mm_unpacklo_epi16(zbin_extra, zbin_extra);
+  /* Duplicate to all lanes. */
+  zbin_extra = _mm_shufflelo_epi16(zbin_extra, 0);
+  zbin_extra = _mm_unpacklo_epi16(zbin_extra, zbin_extra);
 
-    /* Sign of z: z >> 15 */
-    sz0 = _mm_srai_epi16(z0, 15);
-    sz1 = _mm_srai_epi16(z1, 15);
+  /* Sign of z: z >> 15 */
+  sz0 = _mm_srai_epi16(z0, 15);
+  sz1 = _mm_srai_epi16(z1, 15);
 
-    /* x = abs(z): (z ^ sz) - sz */
-    x0 = _mm_xor_si128(z0, sz0);
-    x1 = _mm_xor_si128(z1, sz1);
-    x0 = _mm_sub_epi16(x0, sz0);
-    x1 = _mm_sub_epi16(x1, sz1);
+  /* x = abs(z): (z ^ sz) - sz */
+  x0 = _mm_xor_si128(z0, sz0);
+  x1 = _mm_xor_si128(z1, sz1);
+  x0 = _mm_sub_epi16(x0, sz0);
+  x1 = _mm_sub_epi16(x1, sz1);
 
-    /* zbin[] + zbin_extra */
-    zbin0 = _mm_add_epi16(zbin0, zbin_extra);
-    zbin1 = _mm_add_epi16(zbin1, zbin_extra);
+  /* zbin[] + zbin_extra */
+  zbin0 = _mm_add_epi16(zbin0, zbin_extra);
+  zbin1 = _mm_add_epi16(zbin1, zbin_extra);
 
-    /* In C x is compared to zbin where zbin = zbin[] + boost + extra. Rebalance
-     * the equation because boost is the only value which can change:
-     * x - (zbin[] + extra) >= boost */
-    x_minus_zbin0 = _mm_sub_epi16(x0, zbin0);
-    x_minus_zbin1 = _mm_sub_epi16(x1, zbin1);
+  /* In C x is compared to zbin where zbin = zbin[] + boost + extra. Rebalance
+   * the equation because boost is the only value which can change:
+   * x - (zbin[] + extra) >= boost */
+  x_minus_zbin0 = _mm_sub_epi16(x0, zbin0);
+  x_minus_zbin1 = _mm_sub_epi16(x1, zbin1);
 
-    _mm_store_si128((__m128i *)(x), x_minus_zbin0);
-    _mm_store_si128((__m128i *)(x + 8), x_minus_zbin1);
+  _mm_store_si128((__m128i *)(x), x_minus_zbin0);
+  _mm_store_si128((__m128i *)(x + 8), x_minus_zbin1);
 
-    /* All the remaining calculations are valid whether they are done now with
-     * simd or later inside the loop one at a time. */
-    x0 = _mm_add_epi16(x0, round0);
-    x1 = _mm_add_epi16(x1, round1);
+  /* All the remaining calculations are valid whether they are done now with
+   * simd or later inside the loop one at a time. */
+  x0 = _mm_add_epi16(x0, round0);
+  x1 = _mm_add_epi16(x1, round1);
 
-    y0 = _mm_mulhi_epi16(x0, quant0);
-    y1 = _mm_mulhi_epi16(x1, quant1);
+  y0 = _mm_mulhi_epi16(x0, quant0);
+  y1 = _mm_mulhi_epi16(x1, quant1);
 
-    y0 = _mm_add_epi16(y0, x0);
-    y1 = _mm_add_epi16(y1, x1);
+  y0 = _mm_add_epi16(y0, x0);
+  y1 = _mm_add_epi16(y1, x1);
 
-    /* Instead of shifting each value independently we convert the scaling
-     * factor with 1 << (16 - shift) so we can use multiply/return high half. */
-    y0 = _mm_mulhi_epi16(y0, quant_shift0);
-    y1 = _mm_mulhi_epi16(y1, quant_shift1);
+  /* Instead of shifting each value independently we convert the scaling
+   * factor with 1 << (16 - shift) so we can use multiply/return high half. */
+  y0 = _mm_mulhi_epi16(y0, quant_shift0);
+  y1 = _mm_mulhi_epi16(y1, quant_shift1);
 
-    /* Return the sign: (y ^ sz) - sz */
-    y0 = _mm_xor_si128(y0, sz0);
-    y1 = _mm_xor_si128(y1, sz1);
-    y0 = _mm_sub_epi16(y0, sz0);
-    y1 = _mm_sub_epi16(y1, sz1);
+  /* Return the sign: (y ^ sz) - sz */
+  y0 = _mm_xor_si128(y0, sz0);
+  y1 = _mm_xor_si128(y1, sz1);
+  y0 = _mm_sub_epi16(y0, sz0);
+  y1 = _mm_sub_epi16(y1, sz1);
 
-    _mm_store_si128((__m128i *)(y), y0);
-    _mm_store_si128((__m128i *)(y + 8), y1);
+  _mm_store_si128((__m128i *)(y), y0);
+  _mm_store_si128((__m128i *)(y + 8), y1);
 
-    zbin_boost_ptr = b->zrun_zbin_boost;
+  zbin_boost_ptr = b->zrun_zbin_boost;
 
-    /* The loop gets unrolled anyway. Avoid the vp8_default_zig_zag1d lookup. */
-    SELECT_EOB(1, 0);
-    SELECT_EOB(2, 1);
-    SELECT_EOB(3, 4);
-    SELECT_EOB(4, 8);
-    SELECT_EOB(5, 5);
-    SELECT_EOB(6, 2);
-    SELECT_EOB(7, 3);
-    SELECT_EOB(8, 6);
-    SELECT_EOB(9, 9);
-    SELECT_EOB(10, 12);
-    SELECT_EOB(11, 13);
-    SELECT_EOB(12, 10);
-    SELECT_EOB(13, 7);
-    SELECT_EOB(14, 11);
-    SELECT_EOB(15, 14);
-    SELECT_EOB(16, 15);
+  /* The loop gets unrolled anyway. Avoid the vp8_default_zig_zag1d lookup. */
+  SELECT_EOB(1, 0);
+  SELECT_EOB(2, 1);
+  SELECT_EOB(3, 4);
+  SELECT_EOB(4, 8);
+  SELECT_EOB(5, 5);
+  SELECT_EOB(6, 2);
+  SELECT_EOB(7, 3);
+  SELECT_EOB(8, 6);
+  SELECT_EOB(9, 9);
+  SELECT_EOB(10, 12);
+  SELECT_EOB(11, 13);
+  SELECT_EOB(12, 10);
+  SELECT_EOB(13, 7);
+  SELECT_EOB(14, 11);
+  SELECT_EOB(15, 14);
+  SELECT_EOB(16, 15);
 
-    y0 = _mm_load_si128((__m128i *)(d->qcoeff));
-    y1 = _mm_load_si128((__m128i *)(d->qcoeff + 8));
+  y0 = _mm_load_si128((__m128i *)(d->qcoeff));
+  y1 = _mm_load_si128((__m128i *)(d->qcoeff + 8));
 
-    /* dqcoeff = qcoeff * dequant */
-    y0 = _mm_mullo_epi16(y0, dequant0);
-    y1 = _mm_mullo_epi16(y1, dequant1);
+  /* dqcoeff = qcoeff * dequant */
+  y0 = _mm_mullo_epi16(y0, dequant0);
+  y1 = _mm_mullo_epi16(y1, dequant1);
 
-    _mm_store_si128((__m128i *)(d->dqcoeff), y0);
-    _mm_store_si128((__m128i *)(d->dqcoeff + 8), y1);
+  _mm_store_si128((__m128i *)(d->dqcoeff), y0);
+  _mm_store_si128((__m128i *)(d->dqcoeff + 8), y1);
 
-    *d->eob = eob;
+  *d->eob = eob;
 }
 
-void vp8_fast_quantize_b_sse2(BLOCK *b, BLOCKD *d)
-{
+void vp8_fast_quantize_b_sse2(BLOCK *b, BLOCKD *d) {
   __m128i z0 = _mm_load_si128((__m128i *)(b->coeff));
   __m128i z1 = _mm_load_si128((__m128i *)(b->coeff + 8));
   __m128i round0 = _mm_load_si128((__m128i *)(b->round));
@@ -152,8 +148,10 @@
   __m128i quant_fast1 = _mm_load_si128((__m128i *)(b->quant_fast + 8));
   __m128i dequant0 = _mm_load_si128((__m128i *)(d->dequant));
   __m128i dequant1 = _mm_load_si128((__m128i *)(d->dequant + 8));
-  __m128i inv_zig_zag0 = _mm_load_si128((const __m128i *)(vp8_default_inv_zig_zag));
-  __m128i inv_zig_zag1 = _mm_load_si128((const __m128i *)(vp8_default_inv_zig_zag + 8));
+  __m128i inv_zig_zag0 =
+      _mm_load_si128((const __m128i *)(vp8_default_inv_zig_zag));
+  __m128i inv_zig_zag1 =
+      _mm_load_si128((const __m128i *)(vp8_default_inv_zig_zag + 8));
 
   __m128i sz0, sz1, x0, x1, y0, y1, xdq0, xdq1, zeros, ones;
 
@@ -210,17 +208,17 @@
   x0 = _mm_max_epi16(x0, x1);
 
   /* now down to 8 */
-  x1 = _mm_shuffle_epi32(x0, 0xE); // 0b00001110
+  x1 = _mm_shuffle_epi32(x0, 0xE);  // 0b00001110
 
   x0 = _mm_max_epi16(x0, x1);
 
   /* only 4 left */
-  x1 = _mm_shufflelo_epi16(x0, 0xE); // 0b00001110
+  x1 = _mm_shufflelo_epi16(x0, 0xE);  // 0b00001110
 
   x0 = _mm_max_epi16(x0, x1);
 
   /* okay, just 2! */
-  x1 = _mm_shufflelo_epi16(x0, 0x1); // 0b00000001
+  x1 = _mm_shufflelo_epi16(x0, 0x1);  // 0b00000001
 
   x0 = _mm_max_epi16(x0, x1);
 
--- a/vp8/vp8_cx_iface.c
+++ b/vp8/vp8_cx_iface.c
@@ -8,7 +8,6 @@
  *  be found in the AUTHORS file in the root of the source tree.
  */
 
-
 #include "./vpx_config.h"
 #include "./vp8_rtcd.h"
 #include "./vpx_dsp_rtcd.h"
@@ -26,481 +25,447 @@
 #include <stdlib.h>
 #include <string.h>
 
-struct vp8_extracfg
-{
-    struct vpx_codec_pkt_list *pkt_list;
-    int                         cpu_used;                    /** available cpu percentage in 1/16*/
-    unsigned int                enable_auto_alt_ref;           /** if encoder decides to uses alternate reference frame */
-    unsigned int                noise_sensitivity;
-    unsigned int                Sharpness;
-    unsigned int                static_thresh;
-    unsigned int                token_partitions;
-    unsigned int                arnr_max_frames;    /* alt_ref Noise Reduction Max Frame Count */
-    unsigned int                arnr_strength;    /* alt_ref Noise Reduction Strength */
-    unsigned int                arnr_type;        /* alt_ref filter type */
-    vp8e_tuning                 tuning;
-    unsigned int                cq_level;         /* constrained quality level */
-    unsigned int                rc_max_intra_bitrate_pct;
-    unsigned int                screen_content_mode;
-
+struct vp8_extracfg {
+  struct vpx_codec_pkt_list *pkt_list;
+  int cpu_used; /** available cpu percentage in 1/16*/
+  /** if encoder decides to uses alternate reference frame */
+  unsigned int enable_auto_alt_ref;
+  unsigned int noise_sensitivity;
+  unsigned int Sharpness;
+  unsigned int static_thresh;
+  unsigned int token_partitions;
+  unsigned int arnr_max_frames; /* alt_ref Noise Reduction Max Frame Count */
+  unsigned int arnr_strength;   /* alt_ref Noise Reduction Strength */
+  unsigned int arnr_type;       /* alt_ref filter type */
+  vp8e_tuning tuning;
+  unsigned int cq_level; /* constrained quality level */
+  unsigned int rc_max_intra_bitrate_pct;
+  unsigned int screen_content_mode;
 };
 
 static struct vp8_extracfg default_extracfg = {
   NULL,
 #if !(CONFIG_REALTIME_ONLY)
-  0,                          /* cpu_used      */
+  0, /* cpu_used      */
 #else
-  4,                          /* cpu_used      */
+  4, /* cpu_used      */
 #endif
-  0,                          /* enable_auto_alt_ref */
-  0,                          /* noise_sensitivity */
-  0,                          /* Sharpness */
-  0,                          /* static_thresh */
+  0, /* enable_auto_alt_ref */
+  0, /* noise_sensitivity */
+  0, /* Sharpness */
+  0, /* static_thresh */
 #if (CONFIG_REALTIME_ONLY & CONFIG_ONTHEFLY_BITPACKING)
   VP8_EIGHT_TOKENPARTITION,
 #else
-  VP8_ONE_TOKENPARTITION,     /* token_partitions */
+  VP8_ONE_TOKENPARTITION, /* token_partitions */
 #endif
-  0,                          /* arnr_max_frames */
-  3,                          /* arnr_strength */
-  3,                          /* arnr_type*/
-  0,                          /* tuning*/
-  10,                         /* cq_level */
-  0,                          /* rc_max_intra_bitrate_pct */
-  0,                          /* screen_content_mode */
+  0,  /* arnr_max_frames */
+  3,  /* arnr_strength */
+  3,  /* arnr_type*/
+  0,  /* tuning*/
+  10, /* cq_level */
+  0,  /* rc_max_intra_bitrate_pct */
+  0,  /* screen_content_mode */
 };
 
-struct vpx_codec_alg_priv
-{
-    vpx_codec_priv_t        base;
-    vpx_codec_enc_cfg_t     cfg;
-    struct vp8_extracfg     vp8_cfg;
-    VP8_CONFIG              oxcf;
-    struct VP8_COMP        *cpi;
-    unsigned char          *cx_data;
-    unsigned int            cx_data_sz;
-    vpx_image_t             preview_img;
-    unsigned int            next_frame_flag;
-    vp8_postproc_cfg_t      preview_ppcfg;
-    /* pkt_list size depends on the maximum number of lagged frames allowed. */
-    vpx_codec_pkt_list_decl(64) pkt_list;
-    unsigned int                fixed_kf_cntr;
-    vpx_enc_frame_flags_t   control_frame_flags;
+struct vpx_codec_alg_priv {
+  vpx_codec_priv_t base;
+  vpx_codec_enc_cfg_t cfg;
+  struct vp8_extracfg vp8_cfg;
+  VP8_CONFIG oxcf;
+  struct VP8_COMP *cpi;
+  unsigned char *cx_data;
+  unsigned int cx_data_sz;
+  vpx_image_t preview_img;
+  unsigned int next_frame_flag;
+  vp8_postproc_cfg_t preview_ppcfg;
+  /* pkt_list size depends on the maximum number of lagged frames allowed. */
+  vpx_codec_pkt_list_decl(64) pkt_list;
+  unsigned int fixed_kf_cntr;
+  vpx_enc_frame_flags_t control_frame_flags;
 };
 
+static vpx_codec_err_t update_error_state(
+    vpx_codec_alg_priv_t *ctx, const struct vpx_internal_error_info *error) {
+  vpx_codec_err_t res;
 
-static vpx_codec_err_t
-update_error_state(vpx_codec_alg_priv_t                 *ctx,
-                   const struct vpx_internal_error_info *error)
-{
-    vpx_codec_err_t res;
+  if ((res = error->error_code))
+    ctx->base.err_detail = error->has_detail ? error->detail : NULL;
 
-    if ((res = error->error_code))
-        ctx->base.err_detail = error->has_detail
-                               ? error->detail
-                               : NULL;
-
-    return res;
+  return res;
 }
 
-
 #undef ERROR
-#define ERROR(str) do {\
-        ctx->base.err_detail = str;\
-        return VPX_CODEC_INVALID_PARAM;\
-    } while(0)
+#define ERROR(str)                  \
+  do {                              \
+    ctx->base.err_detail = str;     \
+    return VPX_CODEC_INVALID_PARAM; \
+  } while (0)
 
-#define RANGE_CHECK(p,memb,lo,hi) do {\
-        if(!(((p)->memb == lo || (p)->memb > (lo)) && (p)->memb <= hi)) \
-            ERROR(#memb " out of range ["#lo".."#hi"]");\
-    } while(0)
+#define RANGE_CHECK(p, memb, lo, hi)                                 \
+  do {                                                               \
+    if (!(((p)->memb == lo || (p)->memb > (lo)) && (p)->memb <= hi)) \
+      ERROR(#memb " out of range [" #lo ".." #hi "]");               \
+  } while (0)
 
-#define RANGE_CHECK_HI(p,memb,hi) do {\
-        if(!((p)->memb <= (hi))) \
-            ERROR(#memb " out of range [.."#hi"]");\
-    } while(0)
+#define RANGE_CHECK_HI(p, memb, hi)                                     \
+  do {                                                                  \
+    if (!((p)->memb <= (hi))) ERROR(#memb " out of range [.." #hi "]"); \
+  } while (0)
 
-#define RANGE_CHECK_LO(p,memb,lo) do {\
-        if(!((p)->memb >= (lo))) \
-            ERROR(#memb " out of range ["#lo"..]");\
-    } while(0)
+#define RANGE_CHECK_LO(p, memb, lo)                                     \
+  do {                                                                  \
+    if (!((p)->memb >= (lo))) ERROR(#memb " out of range [" #lo "..]"); \
+  } while (0)
 
-#define RANGE_CHECK_BOOL(p,memb) do {\
-        if(!!((p)->memb) != (p)->memb) ERROR(#memb " expected boolean");\
-    } while(0)
+#define RANGE_CHECK_BOOL(p, memb)                                     \
+  do {                                                                \
+    if (!!((p)->memb) != (p)->memb) ERROR(#memb " expected boolean"); \
+  } while (0)
 
-static vpx_codec_err_t validate_config(vpx_codec_alg_priv_t      *ctx,
+static vpx_codec_err_t validate_config(vpx_codec_alg_priv_t *ctx,
                                        const vpx_codec_enc_cfg_t *cfg,
                                        const struct vp8_extracfg *vp8_cfg,
-                                       int                        finalize)
-{
-    RANGE_CHECK(cfg, g_w,                   1, 16383); /* 14 bits available */
-    RANGE_CHECK(cfg, g_h,                   1, 16383); /* 14 bits available */
-    RANGE_CHECK(cfg, g_timebase.den,        1, 1000000000);
-    RANGE_CHECK(cfg, g_timebase.num,        1, 1000000000);
-    RANGE_CHECK_HI(cfg, g_profile,          3);
-    RANGE_CHECK_HI(cfg, rc_max_quantizer,   63);
-    RANGE_CHECK_HI(cfg, rc_min_quantizer,   cfg->rc_max_quantizer);
-    RANGE_CHECK_HI(cfg, g_threads,          64);
+                                       int finalize) {
+  RANGE_CHECK(cfg, g_w, 1, 16383); /* 14 bits available */
+  RANGE_CHECK(cfg, g_h, 1, 16383); /* 14 bits available */
+  RANGE_CHECK(cfg, g_timebase.den, 1, 1000000000);
+  RANGE_CHECK(cfg, g_timebase.num, 1, 1000000000);
+  RANGE_CHECK_HI(cfg, g_profile, 3);
+  RANGE_CHECK_HI(cfg, rc_max_quantizer, 63);
+  RANGE_CHECK_HI(cfg, rc_min_quantizer, cfg->rc_max_quantizer);
+  RANGE_CHECK_HI(cfg, g_threads, 64);
 #if CONFIG_REALTIME_ONLY
-    RANGE_CHECK_HI(cfg, g_lag_in_frames,    0);
+  RANGE_CHECK_HI(cfg, g_lag_in_frames, 0);
 #elif CONFIG_MULTI_RES_ENCODING
-    if (ctx->base.enc.total_encoders > 1)
-        RANGE_CHECK_HI(cfg, g_lag_in_frames,    0);
+  if (ctx->base.enc.total_encoders > 1) RANGE_CHECK_HI(cfg, g_lag_in_frames, 0);
 #else
-    RANGE_CHECK_HI(cfg, g_lag_in_frames,    25);
+  RANGE_CHECK_HI(cfg, g_lag_in_frames, 25);
 #endif
-    RANGE_CHECK(cfg, rc_end_usage,          VPX_VBR, VPX_Q);
-    RANGE_CHECK_HI(cfg, rc_undershoot_pct,  1000);
-    RANGE_CHECK_HI(cfg, rc_overshoot_pct,   1000);
-    RANGE_CHECK_HI(cfg, rc_2pass_vbr_bias_pct, 100);
-    RANGE_CHECK(cfg, kf_mode,               VPX_KF_DISABLED, VPX_KF_AUTO);
+  RANGE_CHECK(cfg, rc_end_usage, VPX_VBR, VPX_Q);
+  RANGE_CHECK_HI(cfg, rc_undershoot_pct, 1000);
+  RANGE_CHECK_HI(cfg, rc_overshoot_pct, 1000);
+  RANGE_CHECK_HI(cfg, rc_2pass_vbr_bias_pct, 100);
+  RANGE_CHECK(cfg, kf_mode, VPX_KF_DISABLED, VPX_KF_AUTO);
 
 /* TODO: add spatial re-sampling support and frame dropping in
  * multi-res-encoder.*/
 #if CONFIG_MULTI_RES_ENCODING
-    if (ctx->base.enc.total_encoders > 1)
-        RANGE_CHECK_HI(cfg, rc_resize_allowed,     0);
+  if (ctx->base.enc.total_encoders > 1)
+    RANGE_CHECK_HI(cfg, rc_resize_allowed, 0);
 #else
-    RANGE_CHECK_BOOL(cfg, rc_resize_allowed);
+  RANGE_CHECK_BOOL(cfg, rc_resize_allowed);
 #endif
-    RANGE_CHECK_HI(cfg, rc_dropframe_thresh,   100);
-    RANGE_CHECK_HI(cfg, rc_resize_up_thresh,   100);
-    RANGE_CHECK_HI(cfg, rc_resize_down_thresh, 100);
+  RANGE_CHECK_HI(cfg, rc_dropframe_thresh, 100);
+  RANGE_CHECK_HI(cfg, rc_resize_up_thresh, 100);
+  RANGE_CHECK_HI(cfg, rc_resize_down_thresh, 100);
 
 #if CONFIG_REALTIME_ONLY
-    RANGE_CHECK(cfg,        g_pass,         VPX_RC_ONE_PASS, VPX_RC_ONE_PASS);
+  RANGE_CHECK(cfg, g_pass, VPX_RC_ONE_PASS, VPX_RC_ONE_PASS);
 #elif CONFIG_MULTI_RES_ENCODING
-    if (ctx->base.enc.total_encoders > 1)
-        RANGE_CHECK(cfg,    g_pass,         VPX_RC_ONE_PASS, VPX_RC_ONE_PASS);
+  if (ctx->base.enc.total_encoders > 1)
+    RANGE_CHECK(cfg, g_pass, VPX_RC_ONE_PASS, VPX_RC_ONE_PASS);
 #else
-    RANGE_CHECK(cfg,        g_pass,         VPX_RC_ONE_PASS, VPX_RC_LAST_PASS);
+  RANGE_CHECK(cfg, g_pass, VPX_RC_ONE_PASS, VPX_RC_LAST_PASS);
 #endif
 
-    /* VP8 does not support a lower bound on the keyframe interval in
-     * automatic keyframe placement mode.
-     */
-    if (cfg->kf_mode != VPX_KF_DISABLED && cfg->kf_min_dist != cfg->kf_max_dist
-        && cfg->kf_min_dist > 0)
-        ERROR("kf_min_dist not supported in auto mode, use 0 "
-              "or kf_max_dist instead.");
+  /* VP8 does not support a lower bound on the keyframe interval in
+   * automatic keyframe placement mode.
+   */
+  if (cfg->kf_mode != VPX_KF_DISABLED && cfg->kf_min_dist != cfg->kf_max_dist &&
+      cfg->kf_min_dist > 0)
+    ERROR(
+        "kf_min_dist not supported in auto mode, use 0 "
+        "or kf_max_dist instead.");
 
-    RANGE_CHECK_BOOL(vp8_cfg,               enable_auto_alt_ref);
-    RANGE_CHECK(vp8_cfg, cpu_used,           -16, 16);
+  RANGE_CHECK_BOOL(vp8_cfg, enable_auto_alt_ref);
+  RANGE_CHECK(vp8_cfg, cpu_used, -16, 16);
 
 #if CONFIG_REALTIME_ONLY && !CONFIG_TEMPORAL_DENOISING
-    RANGE_CHECK(vp8_cfg, noise_sensitivity,  0, 0);
+  RANGE_CHECK(vp8_cfg, noise_sensitivity, 0, 0);
 #else
-    RANGE_CHECK_HI(vp8_cfg, noise_sensitivity,  6);
+  RANGE_CHECK_HI(vp8_cfg, noise_sensitivity, 6);
 #endif
 
-    RANGE_CHECK(vp8_cfg, token_partitions,   VP8_ONE_TOKENPARTITION,
-                VP8_EIGHT_TOKENPARTITION);
-    RANGE_CHECK_HI(vp8_cfg, Sharpness,       7);
-    RANGE_CHECK(vp8_cfg, arnr_max_frames, 0, 15);
-    RANGE_CHECK_HI(vp8_cfg, arnr_strength,   6);
-    RANGE_CHECK(vp8_cfg, arnr_type,       1, 3);
-    RANGE_CHECK(vp8_cfg, cq_level, 0, 63);
-    RANGE_CHECK_HI(vp8_cfg, screen_content_mode, 2);
-    if (finalize && (cfg->rc_end_usage == VPX_CQ || cfg->rc_end_usage == VPX_Q))
-        RANGE_CHECK(vp8_cfg, cq_level,
-                    cfg->rc_min_quantizer, cfg->rc_max_quantizer);
+  RANGE_CHECK(vp8_cfg, token_partitions, VP8_ONE_TOKENPARTITION,
+              VP8_EIGHT_TOKENPARTITION);
+  RANGE_CHECK_HI(vp8_cfg, Sharpness, 7);
+  RANGE_CHECK(vp8_cfg, arnr_max_frames, 0, 15);
+  RANGE_CHECK_HI(vp8_cfg, arnr_strength, 6);
+  RANGE_CHECK(vp8_cfg, arnr_type, 1, 3);
+  RANGE_CHECK(vp8_cfg, cq_level, 0, 63);
+  RANGE_CHECK_HI(vp8_cfg, screen_content_mode, 2);
+  if (finalize && (cfg->rc_end_usage == VPX_CQ || cfg->rc_end_usage == VPX_Q))
+    RANGE_CHECK(vp8_cfg, cq_level, cfg->rc_min_quantizer,
+                cfg->rc_max_quantizer);
 
 #if !(CONFIG_REALTIME_ONLY)
-    if (cfg->g_pass == VPX_RC_LAST_PASS)
-    {
-        size_t           packet_sz = sizeof(FIRSTPASS_STATS);
-        int              n_packets = (int)(cfg->rc_twopass_stats_in.sz /
-                                          packet_sz);
-        FIRSTPASS_STATS *stats;
+  if (cfg->g_pass == VPX_RC_LAST_PASS) {
+    size_t packet_sz = sizeof(FIRSTPASS_STATS);
+    int n_packets = (int)(cfg->rc_twopass_stats_in.sz / packet_sz);
+    FIRSTPASS_STATS *stats;
 
-        if (!cfg->rc_twopass_stats_in.buf)
-            ERROR("rc_twopass_stats_in.buf not set.");
+    if (!cfg->rc_twopass_stats_in.buf)
+      ERROR("rc_twopass_stats_in.buf not set.");
 
-        if (cfg->rc_twopass_stats_in.sz % packet_sz)
-            ERROR("rc_twopass_stats_in.sz indicates truncated packet.");
+    if (cfg->rc_twopass_stats_in.sz % packet_sz)
+      ERROR("rc_twopass_stats_in.sz indicates truncated packet.");
 
-        if (cfg->rc_twopass_stats_in.sz < 2 * packet_sz)
-            ERROR("rc_twopass_stats_in requires at least two packets.");
+    if (cfg->rc_twopass_stats_in.sz < 2 * packet_sz)
+      ERROR("rc_twopass_stats_in requires at least two packets.");
 
-        stats = (void*)((char *)cfg->rc_twopass_stats_in.buf
-                + (n_packets - 1) * packet_sz);
+    stats = (void *)((char *)cfg->rc_twopass_stats_in.buf +
+                     (n_packets - 1) * packet_sz);
 
-        if ((int)(stats->count + 0.5) != n_packets - 1)
-            ERROR("rc_twopass_stats_in missing EOS stats packet");
-    }
+    if ((int)(stats->count + 0.5) != n_packets - 1)
+      ERROR("rc_twopass_stats_in missing EOS stats packet");
+  }
 #endif
 
-    RANGE_CHECK(cfg, ts_number_layers, 1, 5);
+  RANGE_CHECK(cfg, ts_number_layers, 1, 5);
 
-    if (cfg->ts_number_layers > 1)
-    {
-        unsigned int i;
-        RANGE_CHECK_HI(cfg, ts_periodicity, 16);
+  if (cfg->ts_number_layers > 1) {
+    unsigned int i;
+    RANGE_CHECK_HI(cfg, ts_periodicity, 16);
 
-        for (i=1; i<cfg->ts_number_layers; i++)
-            if (cfg->ts_target_bitrate[i] <= cfg->ts_target_bitrate[i-1] &&
-                cfg->rc_target_bitrate > 0)
-                ERROR("ts_target_bitrate entries are not strictly increasing");
+    for (i = 1; i < cfg->ts_number_layers; i++)
+      if (cfg->ts_target_bitrate[i] <= cfg->ts_target_bitrate[i - 1] &&
+          cfg->rc_target_bitrate > 0)
+        ERROR("ts_target_bitrate entries are not strictly increasing");
 
-        RANGE_CHECK(cfg, ts_rate_decimator[cfg->ts_number_layers-1], 1, 1);
-        for (i=cfg->ts_number_layers-2; i>0; i--)
-            if (cfg->ts_rate_decimator[i-1] != 2*cfg->ts_rate_decimator[i])
-                ERROR("ts_rate_decimator factors are not powers of 2");
+    RANGE_CHECK(cfg, ts_rate_decimator[cfg->ts_number_layers - 1], 1, 1);
+    for (i = cfg->ts_number_layers - 2; i > 0; i--)
+      if (cfg->ts_rate_decimator[i - 1] != 2 * cfg->ts_rate_decimator[i])
+        ERROR("ts_rate_decimator factors are not powers of 2");
 
-        RANGE_CHECK_HI(cfg, ts_layer_id[i], cfg->ts_number_layers-1);
-    }
+    RANGE_CHECK_HI(cfg, ts_layer_id[i], cfg->ts_number_layers - 1);
+  }
 
 #if (CONFIG_REALTIME_ONLY & CONFIG_ONTHEFLY_BITPACKING)
-    if(cfg->g_threads > (1 << vp8_cfg->token_partitions))
-        ERROR("g_threads cannot be bigger than number of token partitions");
+  if (cfg->g_threads > (1 << vp8_cfg->token_partitions))
+    ERROR("g_threads cannot be bigger than number of token partitions");
 #endif
 
-    return VPX_CODEC_OK;
+  return VPX_CODEC_OK;
 }
 
-
 static vpx_codec_err_t validate_img(vpx_codec_alg_priv_t *ctx,
-                                    const vpx_image_t    *img)
-{
-    switch (img->fmt)
-    {
+                                    const vpx_image_t *img) {
+  switch (img->fmt) {
     case VPX_IMG_FMT_YV12:
     case VPX_IMG_FMT_I420:
     case VPX_IMG_FMT_VPXI420:
-    case VPX_IMG_FMT_VPXYV12:
-        break;
+    case VPX_IMG_FMT_VPXYV12: break;
     default:
-        ERROR("Invalid image format. Only YV12 and I420 images are supported");
-    }
+      ERROR("Invalid image format. Only YV12 and I420 images are supported");
+  }
 
-    if ((img->d_w != ctx->cfg.g_w) || (img->d_h != ctx->cfg.g_h))
-        ERROR("Image size must match encoder init configuration size");
+  if ((img->d_w != ctx->cfg.g_w) || (img->d_h != ctx->cfg.g_h))
+    ERROR("Image size must match encoder init configuration size");
 
-    return VPX_CODEC_OK;
+  return VPX_CODEC_OK;
 }
 
-
 static vpx_codec_err_t set_vp8e_config(VP8_CONFIG *oxcf,
                                        vpx_codec_enc_cfg_t cfg,
                                        struct vp8_extracfg vp8_cfg,
-                                       vpx_codec_priv_enc_mr_cfg_t *mr_cfg)
-{
-    oxcf->multi_threaded         = cfg.g_threads;
-    oxcf->Version               = cfg.g_profile;
+                                       vpx_codec_priv_enc_mr_cfg_t *mr_cfg) {
+  oxcf->multi_threaded = cfg.g_threads;
+  oxcf->Version = cfg.g_profile;
 
-    oxcf->Width                 = cfg.g_w;
-    oxcf->Height                = cfg.g_h;
-    oxcf->timebase              = cfg.g_timebase;
+  oxcf->Width = cfg.g_w;
+  oxcf->Height = cfg.g_h;
+  oxcf->timebase = cfg.g_timebase;
 
-    oxcf->error_resilient_mode = cfg.g_error_resilient;
+  oxcf->error_resilient_mode = cfg.g_error_resilient;
 
-    switch (cfg.g_pass)
-    {
-    case VPX_RC_ONE_PASS:
-        oxcf->Mode = MODE_BESTQUALITY;
-        break;
-    case VPX_RC_FIRST_PASS:
-        oxcf->Mode = MODE_FIRSTPASS;
-        break;
-    case VPX_RC_LAST_PASS:
-        oxcf->Mode = MODE_SECONDPASS_BEST;
-        break;
-    }
+  switch (cfg.g_pass) {
+    case VPX_RC_ONE_PASS: oxcf->Mode = MODE_BESTQUALITY; break;
+    case VPX_RC_FIRST_PASS: oxcf->Mode = MODE_FIRSTPASS; break;
+    case VPX_RC_LAST_PASS: oxcf->Mode = MODE_SECONDPASS_BEST; break;
+  }
 
-    if (cfg.g_pass == VPX_RC_FIRST_PASS || cfg.g_pass == VPX_RC_ONE_PASS)
-    {
-        oxcf->allow_lag     = 0;
-        oxcf->lag_in_frames = 0;
-    }
-    else
-    {
-        oxcf->allow_lag     = (cfg.g_lag_in_frames) > 0;
-        oxcf->lag_in_frames = cfg.g_lag_in_frames;
-    }
+  if (cfg.g_pass == VPX_RC_FIRST_PASS || cfg.g_pass == VPX_RC_ONE_PASS) {
+    oxcf->allow_lag = 0;
+    oxcf->lag_in_frames = 0;
+  } else {
+    oxcf->allow_lag = (cfg.g_lag_in_frames) > 0;
+    oxcf->lag_in_frames = cfg.g_lag_in_frames;
+  }
 
-    oxcf->allow_df               = (cfg.rc_dropframe_thresh > 0);
-    oxcf->drop_frames_water_mark   = cfg.rc_dropframe_thresh;
+  oxcf->allow_df = (cfg.rc_dropframe_thresh > 0);
+  oxcf->drop_frames_water_mark = cfg.rc_dropframe_thresh;
 
-    oxcf->allow_spatial_resampling = cfg.rc_resize_allowed;
-    oxcf->resample_up_water_mark   = cfg.rc_resize_up_thresh;
-    oxcf->resample_down_water_mark = cfg.rc_resize_down_thresh;
+  oxcf->allow_spatial_resampling = cfg.rc_resize_allowed;
+  oxcf->resample_up_water_mark = cfg.rc_resize_up_thresh;
+  oxcf->resample_down_water_mark = cfg.rc_resize_down_thresh;
 
-    if (cfg.rc_end_usage == VPX_VBR) {
-      oxcf->end_usage = USAGE_LOCAL_FILE_PLAYBACK;
-    } else if (cfg.rc_end_usage == VPX_CBR) {
-      oxcf->end_usage = USAGE_STREAM_FROM_SERVER;
-    } else if (cfg.rc_end_usage == VPX_CQ) {
-      oxcf->end_usage = USAGE_CONSTRAINED_QUALITY;
-    } else if (cfg.rc_end_usage == VPX_Q) {
-      oxcf->end_usage = USAGE_CONSTANT_QUALITY;
-    }
+  if (cfg.rc_end_usage == VPX_VBR) {
+    oxcf->end_usage = USAGE_LOCAL_FILE_PLAYBACK;
+  } else if (cfg.rc_end_usage == VPX_CBR) {
+    oxcf->end_usage = USAGE_STREAM_FROM_SERVER;
+  } else if (cfg.rc_end_usage == VPX_CQ) {
+    oxcf->end_usage = USAGE_CONSTRAINED_QUALITY;
+  } else if (cfg.rc_end_usage == VPX_Q) {
+    oxcf->end_usage = USAGE_CONSTANT_QUALITY;
+  }
 
-    oxcf->target_bandwidth         = cfg.rc_target_bitrate;
-    oxcf->rc_max_intra_bitrate_pct = vp8_cfg.rc_max_intra_bitrate_pct;
+  oxcf->target_bandwidth = cfg.rc_target_bitrate;
+  oxcf->rc_max_intra_bitrate_pct = vp8_cfg.rc_max_intra_bitrate_pct;
 
-    oxcf->best_allowed_q           = cfg.rc_min_quantizer;
-    oxcf->worst_allowed_q          = cfg.rc_max_quantizer;
-    oxcf->cq_level                 = vp8_cfg.cq_level;
-    oxcf->fixed_q = -1;
+  oxcf->best_allowed_q = cfg.rc_min_quantizer;
+  oxcf->worst_allowed_q = cfg.rc_max_quantizer;
+  oxcf->cq_level = vp8_cfg.cq_level;
+  oxcf->fixed_q = -1;
 
-    oxcf->under_shoot_pct          = cfg.rc_undershoot_pct;
-    oxcf->over_shoot_pct           = cfg.rc_overshoot_pct;
+  oxcf->under_shoot_pct = cfg.rc_undershoot_pct;
+  oxcf->over_shoot_pct = cfg.rc_overshoot_pct;
 
-    oxcf->maximum_buffer_size_in_ms   = cfg.rc_buf_sz;
-    oxcf->starting_buffer_level_in_ms = cfg.rc_buf_initial_sz;
-    oxcf->optimal_buffer_level_in_ms  = cfg.rc_buf_optimal_sz;
+  oxcf->maximum_buffer_size_in_ms = cfg.rc_buf_sz;
+  oxcf->starting_buffer_level_in_ms = cfg.rc_buf_initial_sz;
+  oxcf->optimal_buffer_level_in_ms = cfg.rc_buf_optimal_sz;
 
-    oxcf->maximum_buffer_size      = cfg.rc_buf_sz;
-    oxcf->starting_buffer_level    = cfg.rc_buf_initial_sz;
-    oxcf->optimal_buffer_level     = cfg.rc_buf_optimal_sz;
+  oxcf->maximum_buffer_size = cfg.rc_buf_sz;
+  oxcf->starting_buffer_level = cfg.rc_buf_initial_sz;
+  oxcf->optimal_buffer_level = cfg.rc_buf_optimal_sz;
 
-    oxcf->two_pass_vbrbias         = cfg.rc_2pass_vbr_bias_pct;
-    oxcf->two_pass_vbrmin_section  = cfg.rc_2pass_vbr_minsection_pct;
-    oxcf->two_pass_vbrmax_section  = cfg.rc_2pass_vbr_maxsection_pct;
+  oxcf->two_pass_vbrbias = cfg.rc_2pass_vbr_bias_pct;
+  oxcf->two_pass_vbrmin_section = cfg.rc_2pass_vbr_minsection_pct;
+  oxcf->two_pass_vbrmax_section = cfg.rc_2pass_vbr_maxsection_pct;
 
-    oxcf->auto_key                 = cfg.kf_mode == VPX_KF_AUTO
-                                       && cfg.kf_min_dist != cfg.kf_max_dist;
-    oxcf->key_freq                 = cfg.kf_max_dist;
+  oxcf->auto_key =
+      cfg.kf_mode == VPX_KF_AUTO && cfg.kf_min_dist != cfg.kf_max_dist;
+  oxcf->key_freq = cfg.kf_max_dist;
 
-    oxcf->number_of_layers         = cfg.ts_number_layers;
-    oxcf->periodicity              = cfg.ts_periodicity;
+  oxcf->number_of_layers = cfg.ts_number_layers;
+  oxcf->periodicity = cfg.ts_periodicity;
 
-    if (oxcf->number_of_layers > 1)
-    {
-        memcpy (oxcf->target_bitrate, cfg.ts_target_bitrate,
-                sizeof(cfg.ts_target_bitrate));
-        memcpy (oxcf->rate_decimator, cfg.ts_rate_decimator,
-                sizeof(cfg.ts_rate_decimator));
-        memcpy (oxcf->layer_id, cfg.ts_layer_id, sizeof(cfg.ts_layer_id));
-    }
+  if (oxcf->number_of_layers > 1) {
+    memcpy(oxcf->target_bitrate, cfg.ts_target_bitrate,
+           sizeof(cfg.ts_target_bitrate));
+    memcpy(oxcf->rate_decimator, cfg.ts_rate_decimator,
+           sizeof(cfg.ts_rate_decimator));
+    memcpy(oxcf->layer_id, cfg.ts_layer_id, sizeof(cfg.ts_layer_id));
+  }
 
 #if CONFIG_MULTI_RES_ENCODING
-    /* When mr_cfg is NULL, oxcf->mr_total_resolutions and oxcf->mr_encoder_id
-     * are both memset to 0, which ensures the correct logic under this
-     * situation.
-     */
-    if(mr_cfg)
-    {
-        oxcf->mr_total_resolutions        = mr_cfg->mr_total_resolutions;
-        oxcf->mr_encoder_id               = mr_cfg->mr_encoder_id;
-        oxcf->mr_down_sampling_factor.num = mr_cfg->mr_down_sampling_factor.num;
-        oxcf->mr_down_sampling_factor.den = mr_cfg->mr_down_sampling_factor.den;
-        oxcf->mr_low_res_mode_info        = mr_cfg->mr_low_res_mode_info;
-    }
+  /* When mr_cfg is NULL, oxcf->mr_total_resolutions and oxcf->mr_encoder_id
+   * are both memset to 0, which ensures the correct logic under this
+   * situation.
+   */
+  if (mr_cfg) {
+    oxcf->mr_total_resolutions = mr_cfg->mr_total_resolutions;
+    oxcf->mr_encoder_id = mr_cfg->mr_encoder_id;
+    oxcf->mr_down_sampling_factor.num = mr_cfg->mr_down_sampling_factor.num;
+    oxcf->mr_down_sampling_factor.den = mr_cfg->mr_down_sampling_factor.den;
+    oxcf->mr_low_res_mode_info = mr_cfg->mr_low_res_mode_info;
+  }
 #else
-    (void)mr_cfg;
+  (void)mr_cfg;
 #endif
 
-    oxcf->cpu_used               = vp8_cfg.cpu_used;
-    oxcf->encode_breakout        = vp8_cfg.static_thresh;
-    oxcf->play_alternate         = vp8_cfg.enable_auto_alt_ref;
-    oxcf->noise_sensitivity      = vp8_cfg.noise_sensitivity;
-    oxcf->Sharpness              = vp8_cfg.Sharpness;
-    oxcf->token_partitions       = vp8_cfg.token_partitions;
+  oxcf->cpu_used = vp8_cfg.cpu_used;
+  oxcf->encode_breakout = vp8_cfg.static_thresh;
+  oxcf->play_alternate = vp8_cfg.enable_auto_alt_ref;
+  oxcf->noise_sensitivity = vp8_cfg.noise_sensitivity;
+  oxcf->Sharpness = vp8_cfg.Sharpness;
+  oxcf->token_partitions = vp8_cfg.token_partitions;
 
-    oxcf->two_pass_stats_in      = cfg.rc_twopass_stats_in;
-    oxcf->output_pkt_list        = vp8_cfg.pkt_list;
+  oxcf->two_pass_stats_in = cfg.rc_twopass_stats_in;
+  oxcf->output_pkt_list = vp8_cfg.pkt_list;
 
-    oxcf->arnr_max_frames        = vp8_cfg.arnr_max_frames;
-    oxcf->arnr_strength          = vp8_cfg.arnr_strength;
-    oxcf->arnr_type              = vp8_cfg.arnr_type;
+  oxcf->arnr_max_frames = vp8_cfg.arnr_max_frames;
+  oxcf->arnr_strength = vp8_cfg.arnr_strength;
+  oxcf->arnr_type = vp8_cfg.arnr_type;
 
-    oxcf->tuning                 = vp8_cfg.tuning;
+  oxcf->tuning = vp8_cfg.tuning;
 
-    oxcf->screen_content_mode    = vp8_cfg.screen_content_mode;
+  oxcf->screen_content_mode = vp8_cfg.screen_content_mode;
 
-    /*
-        printf("Current VP8 Settings: \n");
-        printf("target_bandwidth: %d\n", oxcf->target_bandwidth);
-        printf("noise_sensitivity: %d\n", oxcf->noise_sensitivity);
-        printf("Sharpness: %d\n",    oxcf->Sharpness);
-        printf("cpu_used: %d\n",  oxcf->cpu_used);
-        printf("Mode: %d\n",     oxcf->Mode);
-        printf("auto_key: %d\n",  oxcf->auto_key);
-        printf("key_freq: %d\n", oxcf->key_freq);
-        printf("end_usage: %d\n", oxcf->end_usage);
-        printf("under_shoot_pct: %d\n", oxcf->under_shoot_pct);
-        printf("over_shoot_pct: %d\n", oxcf->over_shoot_pct);
-        printf("starting_buffer_level: %d\n", oxcf->starting_buffer_level);
-        printf("optimal_buffer_level: %d\n",  oxcf->optimal_buffer_level);
-        printf("maximum_buffer_size: %d\n", oxcf->maximum_buffer_size);
-        printf("fixed_q: %d\n",  oxcf->fixed_q);
-        printf("worst_allowed_q: %d\n", oxcf->worst_allowed_q);
-        printf("best_allowed_q: %d\n", oxcf->best_allowed_q);
-        printf("allow_spatial_resampling: %d\n",  oxcf->allow_spatial_resampling);
-        printf("resample_down_water_mark: %d\n", oxcf->resample_down_water_mark);
-        printf("resample_up_water_mark: %d\n", oxcf->resample_up_water_mark);
-        printf("allow_df: %d\n", oxcf->allow_df);
-        printf("drop_frames_water_mark: %d\n", oxcf->drop_frames_water_mark);
-        printf("two_pass_vbrbias: %d\n",  oxcf->two_pass_vbrbias);
-        printf("two_pass_vbrmin_section: %d\n", oxcf->two_pass_vbrmin_section);
-        printf("two_pass_vbrmax_section: %d\n", oxcf->two_pass_vbrmax_section);
-        printf("allow_lag: %d\n", oxcf->allow_lag);
-        printf("lag_in_frames: %d\n", oxcf->lag_in_frames);
-        printf("play_alternate: %d\n", oxcf->play_alternate);
-        printf("Version: %d\n", oxcf->Version);
-        printf("multi_threaded: %d\n",   oxcf->multi_threaded);
-        printf("encode_breakout: %d\n", oxcf->encode_breakout);
-    */
-    return VPX_CODEC_OK;
+  /*
+      printf("Current VP8 Settings: \n");
+      printf("target_bandwidth: %d\n", oxcf->target_bandwidth);
+      printf("noise_sensitivity: %d\n", oxcf->noise_sensitivity);
+      printf("Sharpness: %d\n",    oxcf->Sharpness);
+      printf("cpu_used: %d\n",  oxcf->cpu_used);
+      printf("Mode: %d\n",     oxcf->Mode);
+      printf("auto_key: %d\n",  oxcf->auto_key);
+      printf("key_freq: %d\n", oxcf->key_freq);
+      printf("end_usage: %d\n", oxcf->end_usage);
+      printf("under_shoot_pct: %d\n", oxcf->under_shoot_pct);
+      printf("over_shoot_pct: %d\n", oxcf->over_shoot_pct);
+      printf("starting_buffer_level: %d\n", oxcf->starting_buffer_level);
+      printf("optimal_buffer_level: %d\n",  oxcf->optimal_buffer_level);
+      printf("maximum_buffer_size: %d\n", oxcf->maximum_buffer_size);
+      printf("fixed_q: %d\n",  oxcf->fixed_q);
+      printf("worst_allowed_q: %d\n", oxcf->worst_allowed_q);
+      printf("best_allowed_q: %d\n", oxcf->best_allowed_q);
+      printf("allow_spatial_resampling: %d\n",  oxcf->allow_spatial_resampling);
+      printf("resample_down_water_mark: %d\n", oxcf->resample_down_water_mark);
+      printf("resample_up_water_mark: %d\n", oxcf->resample_up_water_mark);
+      printf("allow_df: %d\n", oxcf->allow_df);
+      printf("drop_frames_water_mark: %d\n", oxcf->drop_frames_water_mark);
+      printf("two_pass_vbrbias: %d\n",  oxcf->two_pass_vbrbias);
+      printf("two_pass_vbrmin_section: %d\n", oxcf->two_pass_vbrmin_section);
+      printf("two_pass_vbrmax_section: %d\n", oxcf->two_pass_vbrmax_section);
+      printf("allow_lag: %d\n", oxcf->allow_lag);
+      printf("lag_in_frames: %d\n", oxcf->lag_in_frames);
+      printf("play_alternate: %d\n", oxcf->play_alternate);
+      printf("Version: %d\n", oxcf->Version);
+      printf("multi_threaded: %d\n",   oxcf->multi_threaded);
+      printf("encode_breakout: %d\n", oxcf->encode_breakout);
+  */
+  return VPX_CODEC_OK;
 }
 
-static vpx_codec_err_t vp8e_set_config(vpx_codec_alg_priv_t       *ctx,
-                                       const vpx_codec_enc_cfg_t  *cfg)
-{
-    vpx_codec_err_t res;
+static vpx_codec_err_t vp8e_set_config(vpx_codec_alg_priv_t *ctx,
+                                       const vpx_codec_enc_cfg_t *cfg) {
+  vpx_codec_err_t res;
 
-    if (cfg->g_w != ctx->cfg.g_w || cfg->g_h != ctx->cfg.g_h)
-    {
-        if (cfg->g_lag_in_frames > 1 || cfg->g_pass != VPX_RC_ONE_PASS)
-            ERROR("Cannot change width or height after initialization");
-        if ((ctx->cpi->initial_width && (int)cfg->g_w > ctx->cpi->initial_width) ||
-            (ctx->cpi->initial_height && (int)cfg->g_h > ctx->cpi->initial_height))
-            ERROR("Cannot increase width or height larger than their initial values");
-    }
+  if (cfg->g_w != ctx->cfg.g_w || cfg->g_h != ctx->cfg.g_h) {
+    if (cfg->g_lag_in_frames > 1 || cfg->g_pass != VPX_RC_ONE_PASS)
+      ERROR("Cannot change width or height after initialization");
+    if ((ctx->cpi->initial_width && (int)cfg->g_w > ctx->cpi->initial_width) ||
+        (ctx->cpi->initial_height && (int)cfg->g_h > ctx->cpi->initial_height))
+      ERROR("Cannot increase width or height larger than their initial values");
+  }
 
-    /* Prevent increasing lag_in_frames. This check is stricter than it needs
-     * to be -- the limit is not increasing past the first lag_in_frames
-     * value, but we don't track the initial config, only the last successful
-     * config.
-     */
-    if ((cfg->g_lag_in_frames > ctx->cfg.g_lag_in_frames))
-        ERROR("Cannot increase lag_in_frames");
+  /* Prevent increasing lag_in_frames. This check is stricter than it needs
+   * to be -- the limit is not increasing past the first lag_in_frames
+   * value, but we don't track the initial config, only the last successful
+   * config.
+   */
+  if ((cfg->g_lag_in_frames > ctx->cfg.g_lag_in_frames))
+    ERROR("Cannot increase lag_in_frames");
 
-    res = validate_config(ctx, cfg, &ctx->vp8_cfg, 0);
+  res = validate_config(ctx, cfg, &ctx->vp8_cfg, 0);
 
-    if (!res)
-    {
-        ctx->cfg = *cfg;
-        set_vp8e_config(&ctx->oxcf, ctx->cfg, ctx->vp8_cfg, NULL);
-        vp8_change_config(ctx->cpi, &ctx->oxcf);
-    }
+  if (!res) {
+    ctx->cfg = *cfg;
+    set_vp8e_config(&ctx->oxcf, ctx->cfg, ctx->vp8_cfg, NULL);
+    vp8_change_config(ctx->cpi, &ctx->oxcf);
+  }
 
-    return res;
+  return res;
 }
 
-static vpx_codec_err_t get_quantizer(vpx_codec_alg_priv_t *ctx, va_list args)
-{
+static vpx_codec_err_t get_quantizer(vpx_codec_alg_priv_t *ctx, va_list args) {
   int *const arg = va_arg(args, int *);
-  if (arg == NULL)
-    return VPX_CODEC_INVALID_PARAM;
+  if (arg == NULL) return VPX_CODEC_INVALID_PARAM;
   *arg = vp8_get_quantizer(ctx->cpi);
   return VPX_CODEC_OK;
 }
 
-static vpx_codec_err_t get_quantizer64(vpx_codec_alg_priv_t *ctx, va_list args)
-{
+static vpx_codec_err_t get_quantizer64(vpx_codec_alg_priv_t *ctx,
+                                       va_list args) {
   int *const arg = va_arg(args, int *);
-  if (arg == NULL)
-    return VPX_CODEC_INVALID_PARAM;
+  if (arg == NULL) return VPX_CODEC_INVALID_PARAM;
   *arg = vp8_reverse_trans(vp8_get_quantizer(ctx->cpi));
   return VPX_CODEC_OK;
 }
 
 static vpx_codec_err_t update_extracfg(vpx_codec_alg_priv_t *ctx,
-                                       const struct vp8_extracfg *extra_cfg)
-{
+                                       const struct vp8_extracfg *extra_cfg) {
   const vpx_codec_err_t res = validate_config(ctx, &ctx->cfg, extra_cfg, 0);
   if (res == VPX_CODEC_OK) {
     ctx->vp8_cfg = *extra_cfg;
@@ -510,8 +475,7 @@
   return res;
 }
 
-static vpx_codec_err_t set_cpu_used(vpx_codec_alg_priv_t *ctx, va_list args)
-{
+static vpx_codec_err_t set_cpu_used(vpx_codec_alg_priv_t *ctx, va_list args) {
   struct vp8_extracfg extra_cfg = ctx->vp8_cfg;
   extra_cfg.cpu_used = CAST(VP8E_SET_CPUUSED, args);
   return update_extracfg(ctx, &extra_cfg);
@@ -518,8 +482,7 @@
 }
 
 static vpx_codec_err_t set_enable_auto_alt_ref(vpx_codec_alg_priv_t *ctx,
-                                               va_list args)
-{
+                                               va_list args) {
   struct vp8_extracfg extra_cfg = ctx->vp8_cfg;
   extra_cfg.enable_auto_alt_ref = CAST(VP8E_SET_ENABLEAUTOALTREF, args);
   return update_extracfg(ctx, &extra_cfg);
@@ -526,15 +489,13 @@
 }
 
 static vpx_codec_err_t set_noise_sensitivity(vpx_codec_alg_priv_t *ctx,
-                                             va_list args)
-{
+                                             va_list args) {
   struct vp8_extracfg extra_cfg = ctx->vp8_cfg;
   extra_cfg.noise_sensitivity = CAST(VP8E_SET_NOISE_SENSITIVITY, args);
   return update_extracfg(ctx, &extra_cfg);
 }
 
-static vpx_codec_err_t set_sharpness(vpx_codec_alg_priv_t *ctx, va_list args)
-{
+static vpx_codec_err_t set_sharpness(vpx_codec_alg_priv_t *ctx, va_list args) {
   struct vp8_extracfg extra_cfg = ctx->vp8_cfg;
   extra_cfg.Sharpness = CAST(VP8E_SET_SHARPNESS, args);
   return update_extracfg(ctx, &extra_cfg);
@@ -541,8 +502,7 @@
 }
 
 static vpx_codec_err_t set_static_thresh(vpx_codec_alg_priv_t *ctx,
-                                         va_list args)
-{
+                                         va_list args) {
   struct vp8_extracfg extra_cfg = ctx->vp8_cfg;
   extra_cfg.static_thresh = CAST(VP8E_SET_STATIC_THRESHOLD, args);
   return update_extracfg(ctx, &extra_cfg);
@@ -549,8 +509,7 @@
 }
 
 static vpx_codec_err_t set_token_partitions(vpx_codec_alg_priv_t *ctx,
-                                            va_list args)
-{
+                                            va_list args) {
   struct vp8_extracfg extra_cfg = ctx->vp8_cfg;
   extra_cfg.token_partitions = CAST(VP8E_SET_TOKEN_PARTITIONS, args);
   return update_extracfg(ctx, &extra_cfg);
@@ -557,8 +516,7 @@
 }
 
 static vpx_codec_err_t set_arnr_max_frames(vpx_codec_alg_priv_t *ctx,
-                                           va_list args)
-{
+                                           va_list args) {
   struct vp8_extracfg extra_cfg = ctx->vp8_cfg;
   extra_cfg.arnr_max_frames = CAST(VP8E_SET_ARNR_MAXFRAMES, args);
   return update_extracfg(ctx, &extra_cfg);
@@ -565,29 +523,25 @@
 }
 
 static vpx_codec_err_t set_arnr_strength(vpx_codec_alg_priv_t *ctx,
-                                         va_list args)
-{
+                                         va_list args) {
   struct vp8_extracfg extra_cfg = ctx->vp8_cfg;
   extra_cfg.arnr_strength = CAST(VP8E_SET_ARNR_STRENGTH, args);
   return update_extracfg(ctx, &extra_cfg);
 }
 
-static vpx_codec_err_t set_arnr_type(vpx_codec_alg_priv_t *ctx, va_list args)
-{
+static vpx_codec_err_t set_arnr_type(vpx_codec_alg_priv_t *ctx, va_list args) {
   struct vp8_extracfg extra_cfg = ctx->vp8_cfg;
   extra_cfg.arnr_type = CAST(VP8E_SET_ARNR_TYPE, args);
   return update_extracfg(ctx, &extra_cfg);
 }
 
-static vpx_codec_err_t set_tuning(vpx_codec_alg_priv_t *ctx, va_list args)
-{
+static vpx_codec_err_t set_tuning(vpx_codec_alg_priv_t *ctx, va_list args) {
   struct vp8_extracfg extra_cfg = ctx->vp8_cfg;
   extra_cfg.tuning = CAST(VP8E_SET_TUNING, args);
   return update_extracfg(ctx, &extra_cfg);
 }
 
-static vpx_codec_err_t set_cq_level(vpx_codec_alg_priv_t *ctx, va_list args)
-{
+static vpx_codec_err_t set_cq_level(vpx_codec_alg_priv_t *ctx, va_list args) {
   struct vp8_extracfg extra_cfg = ctx->vp8_cfg;
   extra_cfg.cq_level = CAST(VP8E_SET_CQ_LEVEL, args);
   return update_extracfg(ctx, &extra_cfg);
@@ -594,8 +548,7 @@
 }
 
 static vpx_codec_err_t set_rc_max_intra_bitrate_pct(vpx_codec_alg_priv_t *ctx,
-                                                    va_list args)
-{
+                                                    va_list args) {
   struct vp8_extracfg extra_cfg = ctx->vp8_cfg;
   extra_cfg.rc_max_intra_bitrate_pct =
       CAST(VP8E_SET_MAX_INTRA_BITRATE_PCT, args);
@@ -603,776 +556,677 @@
 }
 
 static vpx_codec_err_t set_screen_content_mode(vpx_codec_alg_priv_t *ctx,
-                                               va_list args)
-{
+                                               va_list args) {
   struct vp8_extracfg extra_cfg = ctx->vp8_cfg;
-  extra_cfg.screen_content_mode =
-      CAST(VP8E_SET_SCREEN_CONTENT_MODE, args);
+  extra_cfg.screen_content_mode = CAST(VP8E_SET_SCREEN_CONTENT_MODE, args);
   return update_extracfg(ctx, &extra_cfg);
 }
 
 static vpx_codec_err_t vp8e_mr_alloc_mem(const vpx_codec_enc_cfg_t *cfg,
-                                        void **mem_loc)
-{
-    vpx_codec_err_t res = 0;
+                                         void **mem_loc) {
+  vpx_codec_err_t res = 0;
 
 #if CONFIG_MULTI_RES_ENCODING
-    LOWER_RES_FRAME_INFO *shared_mem_loc;
-    int mb_rows = ((cfg->g_w + 15) >>4);
-    int mb_cols = ((cfg->g_h + 15) >>4);
+  LOWER_RES_FRAME_INFO *shared_mem_loc;
+  int mb_rows = ((cfg->g_w + 15) >> 4);
+  int mb_cols = ((cfg->g_h + 15) >> 4);
 
-    shared_mem_loc = calloc(1, sizeof(LOWER_RES_FRAME_INFO));
-    if(!shared_mem_loc)
-    {
-        res = VPX_CODEC_MEM_ERROR;
-    }
+  shared_mem_loc = calloc(1, sizeof(LOWER_RES_FRAME_INFO));
+  if (!shared_mem_loc) {
+    res = VPX_CODEC_MEM_ERROR;
+  }
 
-    shared_mem_loc->mb_info = calloc(mb_rows*mb_cols, sizeof(LOWER_RES_MB_INFO));
-    if(!(shared_mem_loc->mb_info))
-    {
-        res = VPX_CODEC_MEM_ERROR;
-    }
-    else
-    {
-        *mem_loc = (void *)shared_mem_loc;
-        res = VPX_CODEC_OK;
-    }
+  shared_mem_loc->mb_info =
+      calloc(mb_rows * mb_cols, sizeof(LOWER_RES_MB_INFO));
+  if (!(shared_mem_loc->mb_info)) {
+    res = VPX_CODEC_MEM_ERROR;
+  } else {
+    *mem_loc = (void *)shared_mem_loc;
+    res = VPX_CODEC_OK;
+  }
 #else
-    (void)cfg;
-    (void)mem_loc;
+  (void)cfg;
+  (void)mem_loc;
 #endif
-    return res;
+  return res;
 }
 
 static vpx_codec_err_t vp8e_init(vpx_codec_ctx_t *ctx,
-                                 vpx_codec_priv_enc_mr_cfg_t *mr_cfg)
-{
-    vpx_codec_err_t        res = VPX_CODEC_OK;
+                                 vpx_codec_priv_enc_mr_cfg_t *mr_cfg) {
+  vpx_codec_err_t res = VPX_CODEC_OK;
 
+  vp8_rtcd();
+  vpx_dsp_rtcd();
+  vpx_scale_rtcd();
 
-    vp8_rtcd();
-    vpx_dsp_rtcd();
-    vpx_scale_rtcd();
+  if (!ctx->priv) {
+    struct vpx_codec_alg_priv *priv =
+        (struct vpx_codec_alg_priv *)vpx_calloc(1, sizeof(*priv));
 
-    if (!ctx->priv)
-    {
-        struct vpx_codec_alg_priv *priv =
-            (struct vpx_codec_alg_priv *)vpx_calloc(1, sizeof(*priv));
+    if (!priv) {
+      return VPX_CODEC_MEM_ERROR;
+    }
 
-        if (!priv)
-        {
-            return VPX_CODEC_MEM_ERROR;
-        }
+    ctx->priv = (vpx_codec_priv_t *)priv;
+    ctx->priv->init_flags = ctx->init_flags;
 
-        ctx->priv = (vpx_codec_priv_t *)priv;
-        ctx->priv->init_flags = ctx->init_flags;
+    if (ctx->config.enc) {
+      /* Update the reference to the config structure to an
+       * internal copy.
+       */
+      priv->cfg = *ctx->config.enc;
+      ctx->config.enc = &priv->cfg;
+    }
 
-        if (ctx->config.enc)
-        {
-            /* Update the reference to the config structure to an
-             * internal copy.
-             */
-            priv->cfg = *ctx->config.enc;
-            ctx->config.enc = &priv->cfg;
-        }
+    priv->vp8_cfg = default_extracfg;
+    priv->vp8_cfg.pkt_list = &priv->pkt_list.head;
 
-        priv->vp8_cfg = default_extracfg;
-        priv->vp8_cfg.pkt_list = &priv->pkt_list.head;
+    priv->cx_data_sz = priv->cfg.g_w * priv->cfg.g_h * 3 / 2 * 2;
 
-        priv->cx_data_sz = priv->cfg.g_w * priv->cfg.g_h * 3 / 2 * 2;
+    if (priv->cx_data_sz < 32768) priv->cx_data_sz = 32768;
 
-        if (priv->cx_data_sz < 32768) priv->cx_data_sz = 32768;
+    priv->cx_data = malloc(priv->cx_data_sz);
 
-        priv->cx_data = malloc(priv->cx_data_sz);
+    if (!priv->cx_data) {
+      return VPX_CODEC_MEM_ERROR;
+    }
 
-        if (!priv->cx_data)
-        {
-            return VPX_CODEC_MEM_ERROR;
-        }
+    if (mr_cfg)
+      ctx->priv->enc.total_encoders = mr_cfg->mr_total_resolutions;
+    else
+      ctx->priv->enc.total_encoders = 1;
 
-        if(mr_cfg)
-            ctx->priv->enc.total_encoders   = mr_cfg->mr_total_resolutions;
-        else
-            ctx->priv->enc.total_encoders   = 1;
+    once(vp8_initialize_enc);
 
-        once(vp8_initialize_enc);
+    res = validate_config(priv, &priv->cfg, &priv->vp8_cfg, 0);
 
-        res = validate_config(priv, &priv->cfg, &priv->vp8_cfg, 0);
-
-        if (!res)
-        {
-            set_vp8e_config(&priv->oxcf, priv->cfg, priv->vp8_cfg, mr_cfg);
-            priv->cpi = vp8_create_compressor(&priv->oxcf);
-            if (!priv->cpi)
-                res = VPX_CODEC_MEM_ERROR;
-        }
+    if (!res) {
+      set_vp8e_config(&priv->oxcf, priv->cfg, priv->vp8_cfg, mr_cfg);
+      priv->cpi = vp8_create_compressor(&priv->oxcf);
+      if (!priv->cpi) res = VPX_CODEC_MEM_ERROR;
     }
+  }
 
-    return res;
+  return res;
 }
 
-static vpx_codec_err_t vp8e_destroy(vpx_codec_alg_priv_t *ctx)
-{
+static vpx_codec_err_t vp8e_destroy(vpx_codec_alg_priv_t *ctx) {
 #if CONFIG_MULTI_RES_ENCODING
-    /* Free multi-encoder shared memory */
-    if (ctx->oxcf.mr_total_resolutions > 0 && (ctx->oxcf.mr_encoder_id == ctx->oxcf.mr_total_resolutions-1))
-    {
-        LOWER_RES_FRAME_INFO *shared_mem_loc = (LOWER_RES_FRAME_INFO *)ctx->oxcf.mr_low_res_mode_info;
-        free(shared_mem_loc->mb_info);
-        free(ctx->oxcf.mr_low_res_mode_info);
-    }
+  /* Free multi-encoder shared memory */
+  if (ctx->oxcf.mr_total_resolutions > 0 &&
+      (ctx->oxcf.mr_encoder_id == ctx->oxcf.mr_total_resolutions - 1)) {
+    LOWER_RES_FRAME_INFO *shared_mem_loc =
+        (LOWER_RES_FRAME_INFO *)ctx->oxcf.mr_low_res_mode_info;
+    free(shared_mem_loc->mb_info);
+    free(ctx->oxcf.mr_low_res_mode_info);
+  }
 #endif
 
-    free(ctx->cx_data);
-    vp8_remove_compressor(&ctx->cpi);
-    vpx_free(ctx);
-    return VPX_CODEC_OK;
+  free(ctx->cx_data);
+  vp8_remove_compressor(&ctx->cpi);
+  vpx_free(ctx);
+  return VPX_CODEC_OK;
 }
 
-static vpx_codec_err_t image2yuvconfig(const vpx_image_t   *img,
-                                       YV12_BUFFER_CONFIG  *yv12)
-{
-    const int y_w = img->d_w;
-    const int y_h = img->d_h;
-    const int uv_w = (img->d_w + 1) / 2;
-    const int uv_h = (img->d_h + 1) / 2;
-    vpx_codec_err_t        res = VPX_CODEC_OK;
-    yv12->y_buffer = img->planes[VPX_PLANE_Y];
-    yv12->u_buffer = img->planes[VPX_PLANE_U];
-    yv12->v_buffer = img->planes[VPX_PLANE_V];
+static vpx_codec_err_t image2yuvconfig(const vpx_image_t *img,
+                                       YV12_BUFFER_CONFIG *yv12) {
+  const int y_w = img->d_w;
+  const int y_h = img->d_h;
+  const int uv_w = (img->d_w + 1) / 2;
+  const int uv_h = (img->d_h + 1) / 2;
+  vpx_codec_err_t res = VPX_CODEC_OK;
+  yv12->y_buffer = img->planes[VPX_PLANE_Y];
+  yv12->u_buffer = img->planes[VPX_PLANE_U];
+  yv12->v_buffer = img->planes[VPX_PLANE_V];
 
-    yv12->y_crop_width  = y_w;
-    yv12->y_crop_height = y_h;
-    yv12->y_width  = y_w;
-    yv12->y_height = y_h;
-    yv12->uv_crop_width = uv_w;
-    yv12->uv_crop_height = uv_h;
-    yv12->uv_width = uv_w;
-    yv12->uv_height = uv_h;
+  yv12->y_crop_width = y_w;
+  yv12->y_crop_height = y_h;
+  yv12->y_width = y_w;
+  yv12->y_height = y_h;
+  yv12->uv_crop_width = uv_w;
+  yv12->uv_crop_height = uv_h;
+  yv12->uv_width = uv_w;
+  yv12->uv_height = uv_h;
 
-    yv12->y_stride = img->stride[VPX_PLANE_Y];
-    yv12->uv_stride = img->stride[VPX_PLANE_U];
+  yv12->y_stride = img->stride[VPX_PLANE_Y];
+  yv12->uv_stride = img->stride[VPX_PLANE_U];
 
-    yv12->border  = (img->stride[VPX_PLANE_Y] - img->w) / 2;
-    return res;
+  yv12->border = (img->stride[VPX_PLANE_Y] - img->w) / 2;
+  return res;
 }
 
-static void pick_quickcompress_mode(vpx_codec_alg_priv_t  *ctx,
-                                    unsigned long          duration,
-                                    unsigned long          deadline)
-{
-    int new_qc;
+static void pick_quickcompress_mode(vpx_codec_alg_priv_t *ctx,
+                                    unsigned long duration,
+                                    unsigned long deadline) {
+  int new_qc;
 
 #if !(CONFIG_REALTIME_ONLY)
-    /* Use best quality mode if no deadline is given. */
-    new_qc = MODE_BESTQUALITY;
+  /* Use best quality mode if no deadline is given. */
+  new_qc = MODE_BESTQUALITY;
 
-    if (deadline)
-    {
-        uint64_t     duration_us;
+  if (deadline) {
+    uint64_t duration_us;
 
-        /* Convert duration parameter from stream timebase to microseconds */
-        duration_us = (uint64_t)duration * 1000000
-                      * (uint64_t)ctx->cfg.g_timebase.num
-                      / (uint64_t)ctx->cfg.g_timebase.den;
+    /* Convert duration parameter from stream timebase to microseconds */
+    duration_us = (uint64_t)duration * 1000000 *
+                  (uint64_t)ctx->cfg.g_timebase.num /
+                  (uint64_t)ctx->cfg.g_timebase.den;
 
-        /* If the deadline is more that the duration this frame is to be shown,
-         * use good quality mode. Otherwise use realtime mode.
-         */
-        new_qc = (deadline > duration_us) ? MODE_GOODQUALITY : MODE_REALTIME;
-    }
+    /* If the deadline is more that the duration this frame is to be shown,
+     * use good quality mode. Otherwise use realtime mode.
+     */
+    new_qc = (deadline > duration_us) ? MODE_GOODQUALITY : MODE_REALTIME;
+  }
 
 #else
-    (void)duration;
-    new_qc = MODE_REALTIME;
+  (void)duration;
+  new_qc = MODE_REALTIME;
 #endif
 
-    if (deadline == VPX_DL_REALTIME)
-        new_qc = MODE_REALTIME;
-    else if (ctx->cfg.g_pass == VPX_RC_FIRST_PASS)
-        new_qc = MODE_FIRSTPASS;
-    else if (ctx->cfg.g_pass == VPX_RC_LAST_PASS)
-        new_qc = (new_qc == MODE_BESTQUALITY)
-                 ? MODE_SECONDPASS_BEST
-                 : MODE_SECONDPASS;
+  if (deadline == VPX_DL_REALTIME)
+    new_qc = MODE_REALTIME;
+  else if (ctx->cfg.g_pass == VPX_RC_FIRST_PASS)
+    new_qc = MODE_FIRSTPASS;
+  else if (ctx->cfg.g_pass == VPX_RC_LAST_PASS)
+    new_qc =
+        (new_qc == MODE_BESTQUALITY) ? MODE_SECONDPASS_BEST : MODE_SECONDPASS;
 
-    if (ctx->oxcf.Mode != new_qc)
-    {
-        ctx->oxcf.Mode = new_qc;
-        vp8_change_config(ctx->cpi, &ctx->oxcf);
-    }
+  if (ctx->oxcf.Mode != new_qc) {
+    ctx->oxcf.Mode = new_qc;
+    vp8_change_config(ctx->cpi, &ctx->oxcf);
+  }
 }
 
 static vpx_codec_err_t set_reference_and_update(vpx_codec_alg_priv_t *ctx,
-                                                int flags)
-{
+                                                int flags) {
+  /* Handle Flags */
+  if (((flags & VP8_EFLAG_NO_UPD_GF) && (flags & VP8_EFLAG_FORCE_GF)) ||
+      ((flags & VP8_EFLAG_NO_UPD_ARF) && (flags & VP8_EFLAG_FORCE_ARF))) {
+    ctx->base.err_detail = "Conflicting flags.";
+    return VPX_CODEC_INVALID_PARAM;
+  }
 
-    /* Handle Flags */
-    if (((flags & VP8_EFLAG_NO_UPD_GF) && (flags & VP8_EFLAG_FORCE_GF))
-        || ((flags & VP8_EFLAG_NO_UPD_ARF) && (flags & VP8_EFLAG_FORCE_ARF)))
-    {
-        ctx->base.err_detail = "Conflicting flags.";
-        return VPX_CODEC_INVALID_PARAM;
-    }
+  if (flags &
+      (VP8_EFLAG_NO_REF_LAST | VP8_EFLAG_NO_REF_GF | VP8_EFLAG_NO_REF_ARF)) {
+    int ref = 7;
 
-    if (flags & (VP8_EFLAG_NO_REF_LAST | VP8_EFLAG_NO_REF_GF
-                 | VP8_EFLAG_NO_REF_ARF))
-    {
-        int ref = 7;
+    if (flags & VP8_EFLAG_NO_REF_LAST) ref ^= VP8_LAST_FRAME;
 
-        if (flags & VP8_EFLAG_NO_REF_LAST)
-            ref ^= VP8_LAST_FRAME;
+    if (flags & VP8_EFLAG_NO_REF_GF) ref ^= VP8_GOLD_FRAME;
 
-        if (flags & VP8_EFLAG_NO_REF_GF)
-            ref ^= VP8_GOLD_FRAME;
+    if (flags & VP8_EFLAG_NO_REF_ARF) ref ^= VP8_ALTR_FRAME;
 
-        if (flags & VP8_EFLAG_NO_REF_ARF)
-            ref ^= VP8_ALTR_FRAME;
+    vp8_use_as_reference(ctx->cpi, ref);
+  }
 
-        vp8_use_as_reference(ctx->cpi, ref);
-    }
+  if (flags &
+      (VP8_EFLAG_NO_UPD_LAST | VP8_EFLAG_NO_UPD_GF | VP8_EFLAG_NO_UPD_ARF |
+       VP8_EFLAG_FORCE_GF | VP8_EFLAG_FORCE_ARF)) {
+    int upd = 7;
 
-    if (flags & (VP8_EFLAG_NO_UPD_LAST | VP8_EFLAG_NO_UPD_GF
-                 | VP8_EFLAG_NO_UPD_ARF | VP8_EFLAG_FORCE_GF
-                 | VP8_EFLAG_FORCE_ARF))
-    {
-        int upd = 7;
+    if (flags & VP8_EFLAG_NO_UPD_LAST) upd ^= VP8_LAST_FRAME;
 
-        if (flags & VP8_EFLAG_NO_UPD_LAST)
-            upd ^= VP8_LAST_FRAME;
+    if (flags & VP8_EFLAG_NO_UPD_GF) upd ^= VP8_GOLD_FRAME;
 
-        if (flags & VP8_EFLAG_NO_UPD_GF)
-            upd ^= VP8_GOLD_FRAME;
+    if (flags & VP8_EFLAG_NO_UPD_ARF) upd ^= VP8_ALTR_FRAME;
 
-        if (flags & VP8_EFLAG_NO_UPD_ARF)
-            upd ^= VP8_ALTR_FRAME;
+    vp8_update_reference(ctx->cpi, upd);
+  }
 
-        vp8_update_reference(ctx->cpi, upd);
-    }
+  if (flags & VP8_EFLAG_NO_UPD_ENTROPY) {
+    vp8_update_entropy(ctx->cpi, 0);
+  }
 
-    if (flags & VP8_EFLAG_NO_UPD_ENTROPY)
-    {
-        vp8_update_entropy(ctx->cpi, 0);
-    }
-
-    return VPX_CODEC_OK;
+  return VPX_CODEC_OK;
 }
 
-static vpx_codec_err_t vp8e_encode(vpx_codec_alg_priv_t  *ctx,
-                                   const vpx_image_t     *img,
-                                   vpx_codec_pts_t        pts,
-                                   unsigned long          duration,
-                                   vpx_enc_frame_flags_t  flags,
-                                   unsigned long          deadline)
-{
-    vpx_codec_err_t res = VPX_CODEC_OK;
+static vpx_codec_err_t vp8e_encode(vpx_codec_alg_priv_t *ctx,
+                                   const vpx_image_t *img, vpx_codec_pts_t pts,
+                                   unsigned long duration,
+                                   vpx_enc_frame_flags_t flags,
+                                   unsigned long deadline) {
+  vpx_codec_err_t res = VPX_CODEC_OK;
 
-    if (!ctx->cfg.rc_target_bitrate)
-        return res;
+  if (!ctx->cfg.rc_target_bitrate) return res;
 
-    if (img)
-        res = validate_img(ctx, img);
+  if (img) res = validate_img(ctx, img);
 
-    if (!res)
-        res = validate_config(ctx, &ctx->cfg, &ctx->vp8_cfg, 1);
+  if (!res) res = validate_config(ctx, &ctx->cfg, &ctx->vp8_cfg, 1);
 
-    pick_quickcompress_mode(ctx, duration, deadline);
-    vpx_codec_pkt_list_init(&ctx->pkt_list);
+  pick_quickcompress_mode(ctx, duration, deadline);
+  vpx_codec_pkt_list_init(&ctx->pkt_list);
 
-    // If no flags are set in the encode call, then use the frame flags as
-    // defined via the control function: vp8e_set_frame_flags.
-    if (!flags) {
-        flags = ctx->control_frame_flags;
-    }
-    ctx->control_frame_flags = 0;
+  // If no flags are set in the encode call, then use the frame flags as
+  // defined via the control function: vp8e_set_frame_flags.
+  if (!flags) {
+    flags = ctx->control_frame_flags;
+  }
+  ctx->control_frame_flags = 0;
 
-    if (!res)
-        res = set_reference_and_update(ctx, flags);
+  if (!res) res = set_reference_and_update(ctx, flags);
 
-    /* Handle fixed keyframe intervals */
-    if (ctx->cfg.kf_mode == VPX_KF_AUTO
-        && ctx->cfg.kf_min_dist == ctx->cfg.kf_max_dist)
-    {
-        if (++ctx->fixed_kf_cntr > ctx->cfg.kf_min_dist)
-        {
-            flags |= VPX_EFLAG_FORCE_KF;
-            ctx->fixed_kf_cntr = 1;
-        }
+  /* Handle fixed keyframe intervals */
+  if (ctx->cfg.kf_mode == VPX_KF_AUTO &&
+      ctx->cfg.kf_min_dist == ctx->cfg.kf_max_dist) {
+    if (++ctx->fixed_kf_cntr > ctx->cfg.kf_min_dist) {
+      flags |= VPX_EFLAG_FORCE_KF;
+      ctx->fixed_kf_cntr = 1;
     }
+  }
 
-    /* Initialize the encoder instance on the first frame*/
-    if (!res && ctx->cpi)
-    {
-        unsigned int lib_flags;
-        YV12_BUFFER_CONFIG sd;
-        int64_t dst_time_stamp, dst_end_time_stamp;
-        unsigned long size, cx_data_sz;
-        unsigned char *cx_data;
-        unsigned char *cx_data_end;
-        int comp_data_state = 0;
+  /* Initialize the encoder instance on the first frame*/
+  if (!res && ctx->cpi) {
+    unsigned int lib_flags;
+    YV12_BUFFER_CONFIG sd;
+    int64_t dst_time_stamp, dst_end_time_stamp;
+    unsigned long size, cx_data_sz;
+    unsigned char *cx_data;
+    unsigned char *cx_data_end;
+    int comp_data_state = 0;
 
-        /* Set up internal flags */
-        if (ctx->base.init_flags & VPX_CODEC_USE_PSNR)
-            ((VP8_COMP *)ctx->cpi)->b_calculate_psnr = 1;
+    /* Set up internal flags */
+    if (ctx->base.init_flags & VPX_CODEC_USE_PSNR)
+      ((VP8_COMP *)ctx->cpi)->b_calculate_psnr = 1;
 
-        if (ctx->base.init_flags & VPX_CODEC_USE_OUTPUT_PARTITION)
-            ((VP8_COMP *)ctx->cpi)->output_partition = 1;
+    if (ctx->base.init_flags & VPX_CODEC_USE_OUTPUT_PARTITION)
+      ((VP8_COMP *)ctx->cpi)->output_partition = 1;
 
-        /* Convert API flags to internal codec lib flags */
-        lib_flags = (flags & VPX_EFLAG_FORCE_KF) ? FRAMEFLAGS_KEY : 0;
+    /* Convert API flags to internal codec lib flags */
+    lib_flags = (flags & VPX_EFLAG_FORCE_KF) ? FRAMEFLAGS_KEY : 0;
 
-        /* vp8 use 10,000,000 ticks/second as time stamp */
-        dst_time_stamp    = pts * 10000000 * ctx->cfg.g_timebase.num / ctx->cfg.g_timebase.den;
-        dst_end_time_stamp = (pts + duration) * 10000000 * ctx->cfg.g_timebase.num / ctx->cfg.g_timebase.den;
+    /* vp8 use 10,000,000 ticks/second as time stamp */
+    dst_time_stamp =
+        pts * 10000000 * ctx->cfg.g_timebase.num / ctx->cfg.g_timebase.den;
+    dst_end_time_stamp = (pts + duration) * 10000000 * ctx->cfg.g_timebase.num /
+                         ctx->cfg.g_timebase.den;
 
-        if (img != NULL)
-        {
-            res = image2yuvconfig(img, &sd);
+    if (img != NULL) {
+      res = image2yuvconfig(img, &sd);
 
-            if (vp8_receive_raw_frame(ctx->cpi, ctx->next_frame_flag | lib_flags,
-                                      &sd, dst_time_stamp, dst_end_time_stamp))
-            {
-                VP8_COMP *cpi = (VP8_COMP *)ctx->cpi;
-                res = update_error_state(ctx, &cpi->common.error);
-            }
+      if (vp8_receive_raw_frame(ctx->cpi, ctx->next_frame_flag | lib_flags, &sd,
+                                dst_time_stamp, dst_end_time_stamp)) {
+        VP8_COMP *cpi = (VP8_COMP *)ctx->cpi;
+        res = update_error_state(ctx, &cpi->common.error);
+      }
 
-            /* reset for next frame */
-            ctx->next_frame_flag = 0;
-        }
+      /* reset for next frame */
+      ctx->next_frame_flag = 0;
+    }
 
-        cx_data = ctx->cx_data;
-        cx_data_sz = ctx->cx_data_sz;
-        cx_data_end = ctx->cx_data + cx_data_sz;
-        lib_flags = 0;
+    cx_data = ctx->cx_data;
+    cx_data_sz = ctx->cx_data_sz;
+    cx_data_end = ctx->cx_data + cx_data_sz;
+    lib_flags = 0;
 
-        while (cx_data_sz >= ctx->cx_data_sz / 2)
-        {
-            comp_data_state = vp8_get_compressed_data(ctx->cpi,
-                                                  &lib_flags,
-                                                  &size,
-                                                  cx_data,
-                                                  cx_data_end,
-                                                  &dst_time_stamp,
-                                                  &dst_end_time_stamp,
-                                                  !img);
+    while (cx_data_sz >= ctx->cx_data_sz / 2) {
+      comp_data_state = vp8_get_compressed_data(
+          ctx->cpi, &lib_flags, &size, cx_data, cx_data_end, &dst_time_stamp,
+          &dst_end_time_stamp, !img);
 
-            if(comp_data_state == VPX_CODEC_CORRUPT_FRAME)
-                return VPX_CODEC_CORRUPT_FRAME;
-            else if(comp_data_state == -1)
-                break;
+      if (comp_data_state == VPX_CODEC_CORRUPT_FRAME)
+        return VPX_CODEC_CORRUPT_FRAME;
+      else if (comp_data_state == -1)
+        break;
 
-            if (size)
-            {
-                vpx_codec_pts_t    round, delta;
-                vpx_codec_cx_pkt_t pkt;
-                VP8_COMP *cpi = (VP8_COMP *)ctx->cpi;
+      if (size) {
+        vpx_codec_pts_t round, delta;
+        vpx_codec_cx_pkt_t pkt;
+        VP8_COMP *cpi = (VP8_COMP *)ctx->cpi;
 
-                /* Add the frame packet to the list of returned packets. */
-                round = (vpx_codec_pts_t)10000000
-                        * ctx->cfg.g_timebase.num / 2 - 1;
-                delta = (dst_end_time_stamp - dst_time_stamp);
-                pkt.kind = VPX_CODEC_CX_FRAME_PKT;
-                pkt.data.frame.pts =
-                    (dst_time_stamp * ctx->cfg.g_timebase.den + round)
-                    / ctx->cfg.g_timebase.num / 10000000;
-                pkt.data.frame.duration = (unsigned long)
-                    ((delta * ctx->cfg.g_timebase.den + round)
-                    / ctx->cfg.g_timebase.num / 10000000);
-                pkt.data.frame.flags = lib_flags << 16;
+        /* Add the frame packet to the list of returned packets. */
+        round = (vpx_codec_pts_t)10000000 * ctx->cfg.g_timebase.num / 2 - 1;
+        delta = (dst_end_time_stamp - dst_time_stamp);
+        pkt.kind = VPX_CODEC_CX_FRAME_PKT;
+        pkt.data.frame.pts =
+            (dst_time_stamp * ctx->cfg.g_timebase.den + round) /
+            ctx->cfg.g_timebase.num / 10000000;
+        pkt.data.frame.duration =
+            (unsigned long)((delta * ctx->cfg.g_timebase.den + round) /
+                            ctx->cfg.g_timebase.num / 10000000);
+        pkt.data.frame.flags = lib_flags << 16;
 
-                if (lib_flags & FRAMEFLAGS_KEY)
-                    pkt.data.frame.flags |= VPX_FRAME_IS_KEY;
+        if (lib_flags & FRAMEFLAGS_KEY)
+          pkt.data.frame.flags |= VPX_FRAME_IS_KEY;
 
-                if (!cpi->common.show_frame)
-                {
-                    pkt.data.frame.flags |= VPX_FRAME_IS_INVISIBLE;
+        if (!cpi->common.show_frame) {
+          pkt.data.frame.flags |= VPX_FRAME_IS_INVISIBLE;
 
-                    /* This timestamp should be as close as possible to the
-                     * prior PTS so that if a decoder uses pts to schedule when
-                     * to do this, we start right after last frame was decoded.
-                     * Invisible frames have no duration.
-                     */
-                    pkt.data.frame.pts = ((cpi->last_time_stamp_seen
-                        * ctx->cfg.g_timebase.den + round)
-                        / ctx->cfg.g_timebase.num / 10000000) + 1;
-                    pkt.data.frame.duration = 0;
-                }
+          /* This timestamp should be as close as possible to the
+           * prior PTS so that if a decoder uses pts to schedule when
+           * to do this, we start right after last frame was decoded.
+           * Invisible frames have no duration.
+           */
+          pkt.data.frame.pts =
+              ((cpi->last_time_stamp_seen * ctx->cfg.g_timebase.den + round) /
+               ctx->cfg.g_timebase.num / 10000000) +
+              1;
+          pkt.data.frame.duration = 0;
+        }
 
-                if (cpi->droppable)
-                    pkt.data.frame.flags |= VPX_FRAME_IS_DROPPABLE;
+        if (cpi->droppable) pkt.data.frame.flags |= VPX_FRAME_IS_DROPPABLE;
 
-                if (cpi->output_partition)
-                {
-                    int i;
-                    const int num_partitions =
-                            (1 << cpi->common.multi_token_partition) + 1;
+        if (cpi->output_partition) {
+          int i;
+          const int num_partitions =
+              (1 << cpi->common.multi_token_partition) + 1;
 
-                    pkt.data.frame.flags |= VPX_FRAME_IS_FRAGMENT;
+          pkt.data.frame.flags |= VPX_FRAME_IS_FRAGMENT;
 
-                    for (i = 0; i < num_partitions; ++i)
-                    {
+          for (i = 0; i < num_partitions; ++i) {
 #if CONFIG_REALTIME_ONLY & CONFIG_ONTHEFLY_BITPACKING
-                        pkt.data.frame.buf = cpi->partition_d[i];
+            pkt.data.frame.buf = cpi->partition_d[i];
 #else
-                        pkt.data.frame.buf = cx_data;
-                        cx_data += cpi->partition_sz[i];
-                        cx_data_sz -= cpi->partition_sz[i];
+            pkt.data.frame.buf = cx_data;
+            cx_data += cpi->partition_sz[i];
+            cx_data_sz -= cpi->partition_sz[i];
 #endif
-                        pkt.data.frame.sz = cpi->partition_sz[i];
-                        pkt.data.frame.partition_id = i;
-                        /* don't set the fragment bit for the last partition */
-                        if (i == (num_partitions - 1))
-                            pkt.data.frame.flags &= ~VPX_FRAME_IS_FRAGMENT;
-                        vpx_codec_pkt_list_add(&ctx->pkt_list.head, &pkt);
-                    }
+            pkt.data.frame.sz = cpi->partition_sz[i];
+            pkt.data.frame.partition_id = i;
+            /* don't set the fragment bit for the last partition */
+            if (i == (num_partitions - 1))
+              pkt.data.frame.flags &= ~VPX_FRAME_IS_FRAGMENT;
+            vpx_codec_pkt_list_add(&ctx->pkt_list.head, &pkt);
+          }
 #if CONFIG_REALTIME_ONLY & CONFIG_ONTHEFLY_BITPACKING
-                    /* In lagged mode the encoder can buffer multiple frames.
-                     * We don't want this in partitioned output because
-                     * partitions are spread all over the output buffer.
-                     * So, force an exit!
-                     */
-                    cx_data_sz -= ctx->cx_data_sz / 2;
+          /* In lagged mode the encoder can buffer multiple frames.
+           * We don't want this in partitioned output because
+           * partitions are spread all over the output buffer.
+           * So, force an exit!
+           */
+          cx_data_sz -= ctx->cx_data_sz / 2;
 #endif
-                }
-                else
-                {
-                    pkt.data.frame.buf = cx_data;
-                    pkt.data.frame.sz  = size;
-                    pkt.data.frame.partition_id = -1;
-                    vpx_codec_pkt_list_add(&ctx->pkt_list.head, &pkt);
-                    cx_data += size;
-                    cx_data_sz -= size;
-                }
-            }
+        } else {
+          pkt.data.frame.buf = cx_data;
+          pkt.data.frame.sz = size;
+          pkt.data.frame.partition_id = -1;
+          vpx_codec_pkt_list_add(&ctx->pkt_list.head, &pkt);
+          cx_data += size;
+          cx_data_sz -= size;
         }
+      }
     }
+  }
 
-    return res;
+  return res;
 }
 
-
-static const vpx_codec_cx_pkt_t *vp8e_get_cxdata(vpx_codec_alg_priv_t  *ctx,
-        vpx_codec_iter_t      *iter)
-{
-    return vpx_codec_pkt_list_get(&ctx->pkt_list.head, iter);
+static const vpx_codec_cx_pkt_t *vp8e_get_cxdata(vpx_codec_alg_priv_t *ctx,
+                                                 vpx_codec_iter_t *iter) {
+  return vpx_codec_pkt_list_get(&ctx->pkt_list.head, iter);
 }
 
 static vpx_codec_err_t vp8e_set_reference(vpx_codec_alg_priv_t *ctx,
-                                          va_list args)
-{
-    vpx_ref_frame_t *data = va_arg(args, vpx_ref_frame_t *);
+                                          va_list args) {
+  vpx_ref_frame_t *data = va_arg(args, vpx_ref_frame_t *);
 
-    if (data)
-    {
-        vpx_ref_frame_t *frame = (vpx_ref_frame_t *)data;
-        YV12_BUFFER_CONFIG sd;
+  if (data) {
+    vpx_ref_frame_t *frame = (vpx_ref_frame_t *)data;
+    YV12_BUFFER_CONFIG sd;
 
-        image2yuvconfig(&frame->img, &sd);
-        vp8_set_reference(ctx->cpi, frame->frame_type, &sd);
-        return VPX_CODEC_OK;
-    }
-    else
-        return VPX_CODEC_INVALID_PARAM;
-
+    image2yuvconfig(&frame->img, &sd);
+    vp8_set_reference(ctx->cpi, frame->frame_type, &sd);
+    return VPX_CODEC_OK;
+  } else
+    return VPX_CODEC_INVALID_PARAM;
 }
 
 static vpx_codec_err_t vp8e_get_reference(vpx_codec_alg_priv_t *ctx,
-                                          va_list args)
-{
+                                          va_list args) {
+  vpx_ref_frame_t *data = va_arg(args, vpx_ref_frame_t *);
 
-    vpx_ref_frame_t *data = va_arg(args, vpx_ref_frame_t *);
+  if (data) {
+    vpx_ref_frame_t *frame = (vpx_ref_frame_t *)data;
+    YV12_BUFFER_CONFIG sd;
 
-    if (data)
-    {
-        vpx_ref_frame_t *frame = (vpx_ref_frame_t *)data;
-        YV12_BUFFER_CONFIG sd;
-
-        image2yuvconfig(&frame->img, &sd);
-        vp8_get_reference(ctx->cpi, frame->frame_type, &sd);
-        return VPX_CODEC_OK;
-    }
-    else
-        return VPX_CODEC_INVALID_PARAM;
+    image2yuvconfig(&frame->img, &sd);
+    vp8_get_reference(ctx->cpi, frame->frame_type, &sd);
+    return VPX_CODEC_OK;
+  } else
+    return VPX_CODEC_INVALID_PARAM;
 }
 
 static vpx_codec_err_t vp8e_set_previewpp(vpx_codec_alg_priv_t *ctx,
-                                          va_list args)
-{
+                                          va_list args) {
 #if CONFIG_POSTPROC
-    vp8_postproc_cfg_t *data = va_arg(args, vp8_postproc_cfg_t *);
+  vp8_postproc_cfg_t *data = va_arg(args, vp8_postproc_cfg_t *);
 
-    if (data)
-    {
-        ctx->preview_ppcfg = *((vp8_postproc_cfg_t *)data);
-        return VPX_CODEC_OK;
-    }
-    else
-        return VPX_CODEC_INVALID_PARAM;
+  if (data) {
+    ctx->preview_ppcfg = *((vp8_postproc_cfg_t *)data);
+    return VPX_CODEC_OK;
+  } else
+    return VPX_CODEC_INVALID_PARAM;
 #else
-    (void)ctx;
-    (void)args;
-    return VPX_CODEC_INCAPABLE;
+  (void)ctx;
+  (void)args;
+  return VPX_CODEC_INCAPABLE;
 #endif
 }
 
+static vpx_image_t *vp8e_get_preview(vpx_codec_alg_priv_t *ctx) {
+  YV12_BUFFER_CONFIG sd;
+  vp8_ppflags_t flags;
+  vp8_zero(flags);
 
-static vpx_image_t *vp8e_get_preview(vpx_codec_alg_priv_t *ctx)
-{
+  if (ctx->preview_ppcfg.post_proc_flag) {
+    flags.post_proc_flag = ctx->preview_ppcfg.post_proc_flag;
+    flags.deblocking_level = ctx->preview_ppcfg.deblocking_level;
+    flags.noise_level = ctx->preview_ppcfg.noise_level;
+  }
 
-    YV12_BUFFER_CONFIG sd;
-    vp8_ppflags_t flags;
-    vp8_zero(flags);
+  if (0 == vp8_get_preview_raw_frame(ctx->cpi, &sd, &flags)) {
+    /*
+    vpx_img_wrap(&ctx->preview_img, VPX_IMG_FMT_YV12,
+        sd.y_width + 2*VP8BORDERINPIXELS,
+        sd.y_height + 2*VP8BORDERINPIXELS,
+        1,
+        sd.buffer_alloc);
+    vpx_img_set_rect(&ctx->preview_img,
+        VP8BORDERINPIXELS, VP8BORDERINPIXELS,
+        sd.y_width, sd.y_height);
+        */
 
-    if (ctx->preview_ppcfg.post_proc_flag)
-    {
-        flags.post_proc_flag        = ctx->preview_ppcfg.post_proc_flag;
-        flags.deblocking_level      = ctx->preview_ppcfg.deblocking_level;
-        flags.noise_level           = ctx->preview_ppcfg.noise_level;
-    }
+    ctx->preview_img.bps = 12;
+    ctx->preview_img.planes[VPX_PLANE_Y] = sd.y_buffer;
+    ctx->preview_img.planes[VPX_PLANE_U] = sd.u_buffer;
+    ctx->preview_img.planes[VPX_PLANE_V] = sd.v_buffer;
 
-    if (0 == vp8_get_preview_raw_frame(ctx->cpi, &sd, &flags))
-    {
+    ctx->preview_img.fmt = VPX_IMG_FMT_I420;
+    ctx->preview_img.x_chroma_shift = 1;
+    ctx->preview_img.y_chroma_shift = 1;
 
-        /*
-        vpx_img_wrap(&ctx->preview_img, VPX_IMG_FMT_YV12,
-            sd.y_width + 2*VP8BORDERINPIXELS,
-            sd.y_height + 2*VP8BORDERINPIXELS,
-            1,
-            sd.buffer_alloc);
-        vpx_img_set_rect(&ctx->preview_img,
-            VP8BORDERINPIXELS, VP8BORDERINPIXELS,
-            sd.y_width, sd.y_height);
-            */
+    ctx->preview_img.d_w = sd.y_width;
+    ctx->preview_img.d_h = sd.y_height;
+    ctx->preview_img.stride[VPX_PLANE_Y] = sd.y_stride;
+    ctx->preview_img.stride[VPX_PLANE_U] = sd.uv_stride;
+    ctx->preview_img.stride[VPX_PLANE_V] = sd.uv_stride;
+    ctx->preview_img.w = sd.y_width;
+    ctx->preview_img.h = sd.y_height;
 
-        ctx->preview_img.bps = 12;
-        ctx->preview_img.planes[VPX_PLANE_Y] = sd.y_buffer;
-        ctx->preview_img.planes[VPX_PLANE_U] = sd.u_buffer;
-        ctx->preview_img.planes[VPX_PLANE_V] = sd.v_buffer;
-
-        ctx->preview_img.fmt = VPX_IMG_FMT_I420;
-        ctx->preview_img.x_chroma_shift = 1;
-        ctx->preview_img.y_chroma_shift = 1;
-
-        ctx->preview_img.d_w = sd.y_width;
-        ctx->preview_img.d_h = sd.y_height;
-        ctx->preview_img.stride[VPX_PLANE_Y] = sd.y_stride;
-        ctx->preview_img.stride[VPX_PLANE_U] = sd.uv_stride;
-        ctx->preview_img.stride[VPX_PLANE_V] = sd.uv_stride;
-        ctx->preview_img.w   = sd.y_width;
-        ctx->preview_img.h   = sd.y_height;
-
-        return &ctx->preview_img;
-    }
-    else
-        return NULL;
+    return &ctx->preview_img;
+  } else
+    return NULL;
 }
 
 static vpx_codec_err_t vp8e_set_frame_flags(vpx_codec_alg_priv_t *ctx,
-                                            va_list args)
-{
-    int frame_flags = va_arg(args, int);
-    ctx->control_frame_flags = frame_flags;
-    return set_reference_and_update(ctx, frame_flags);
+                                            va_list args) {
+  int frame_flags = va_arg(args, int);
+  ctx->control_frame_flags = frame_flags;
+  return set_reference_and_update(ctx, frame_flags);
 }
 
 static vpx_codec_err_t vp8e_set_temporal_layer_id(vpx_codec_alg_priv_t *ctx,
-                                                  va_list args)
-{
-    int layer_id = va_arg(args, int);
-    if (layer_id < 0 || layer_id >= (int)ctx->cfg.ts_number_layers) {
-      return VPX_CODEC_INVALID_PARAM;
-    }
-    ctx->cpi->temporal_layer_id = layer_id;
-    return VPX_CODEC_OK;
+                                                  va_list args) {
+  int layer_id = va_arg(args, int);
+  if (layer_id < 0 || layer_id >= (int)ctx->cfg.ts_number_layers) {
+    return VPX_CODEC_INVALID_PARAM;
+  }
+  ctx->cpi->temporal_layer_id = layer_id;
+  return VPX_CODEC_OK;
 }
 
 static vpx_codec_err_t vp8e_set_roi_map(vpx_codec_alg_priv_t *ctx,
-                                        va_list args)
-{
-    vpx_roi_map_t *data = va_arg(args, vpx_roi_map_t *);
+                                        va_list args) {
+  vpx_roi_map_t *data = va_arg(args, vpx_roi_map_t *);
 
-    if (data)
-    {
-        vpx_roi_map_t *roi = (vpx_roi_map_t *)data;
+  if (data) {
+    vpx_roi_map_t *roi = (vpx_roi_map_t *)data;
 
-        if (!vp8_set_roimap(ctx->cpi, roi->roi_map, roi->rows, roi->cols, roi->delta_q, roi->delta_lf, roi->static_threshold))
-            return VPX_CODEC_OK;
-        else
-            return VPX_CODEC_INVALID_PARAM;
-    }
+    if (!vp8_set_roimap(ctx->cpi, roi->roi_map, roi->rows, roi->cols,
+                        roi->delta_q, roi->delta_lf, roi->static_threshold))
+      return VPX_CODEC_OK;
     else
-        return VPX_CODEC_INVALID_PARAM;
+      return VPX_CODEC_INVALID_PARAM;
+  } else
+    return VPX_CODEC_INVALID_PARAM;
 }
 
-
 static vpx_codec_err_t vp8e_set_activemap(vpx_codec_alg_priv_t *ctx,
-                                          va_list args)
-{
-    vpx_active_map_t *data = va_arg(args, vpx_active_map_t *);
+                                          va_list args) {
+  vpx_active_map_t *data = va_arg(args, vpx_active_map_t *);
 
-    if (data)
-    {
+  if (data) {
+    vpx_active_map_t *map = (vpx_active_map_t *)data;
 
-        vpx_active_map_t *map = (vpx_active_map_t *)data;
-
-        if (!vp8_set_active_map(ctx->cpi, map->active_map, map->rows, map->cols))
-            return VPX_CODEC_OK;
-        else
-            return VPX_CODEC_INVALID_PARAM;
-    }
+    if (!vp8_set_active_map(ctx->cpi, map->active_map, map->rows, map->cols))
+      return VPX_CODEC_OK;
     else
-        return VPX_CODEC_INVALID_PARAM;
+      return VPX_CODEC_INVALID_PARAM;
+  } else
+    return VPX_CODEC_INVALID_PARAM;
 }
 
 static vpx_codec_err_t vp8e_set_scalemode(vpx_codec_alg_priv_t *ctx,
-                                          va_list args)
-{
+                                          va_list args) {
+  vpx_scaling_mode_t *data = va_arg(args, vpx_scaling_mode_t *);
 
-    vpx_scaling_mode_t *data =  va_arg(args, vpx_scaling_mode_t *);
+  if (data) {
+    int res;
+    vpx_scaling_mode_t scalemode = *(vpx_scaling_mode_t *)data;
+    res = vp8_set_internal_size(ctx->cpi, (VPX_SCALING)scalemode.h_scaling_mode,
+                                (VPX_SCALING)scalemode.v_scaling_mode);
 
-    if (data)
-    {
-        int res;
-        vpx_scaling_mode_t scalemode = *(vpx_scaling_mode_t *)data ;
-        res = vp8_set_internal_size(ctx->cpi,
-                                    (VPX_SCALING)scalemode.h_scaling_mode,
-                                    (VPX_SCALING)scalemode.v_scaling_mode);
-
-        if (!res)
-        {
-            /*force next frame a key frame to effect scaling mode */
-            ctx->next_frame_flag |= FRAMEFLAGS_KEY;
-            return VPX_CODEC_OK;
-        }
-        else
-            return VPX_CODEC_INVALID_PARAM;
-    }
-    else
-        return VPX_CODEC_INVALID_PARAM;
+    if (!res) {
+      /*force next frame a key frame to effect scaling mode */
+      ctx->next_frame_flag |= FRAMEFLAGS_KEY;
+      return VPX_CODEC_OK;
+    } else
+      return VPX_CODEC_INVALID_PARAM;
+  } else
+    return VPX_CODEC_INVALID_PARAM;
 }
 
-
-static vpx_codec_ctrl_fn_map_t vp8e_ctf_maps[] =
-{
-    {VP8_SET_REFERENCE,                 vp8e_set_reference},
-    {VP8_COPY_REFERENCE,                vp8e_get_reference},
-    {VP8_SET_POSTPROC,                  vp8e_set_previewpp},
-    {VP8E_SET_FRAME_FLAGS,              vp8e_set_frame_flags},
-    {VP8E_SET_TEMPORAL_LAYER_ID,        vp8e_set_temporal_layer_id},
-    {VP8E_SET_ROI_MAP,                  vp8e_set_roi_map},
-    {VP8E_SET_ACTIVEMAP,                vp8e_set_activemap},
-    {VP8E_SET_SCALEMODE,                vp8e_set_scalemode},
-    {VP8E_SET_CPUUSED,                  set_cpu_used},
-    {VP8E_SET_NOISE_SENSITIVITY,        set_noise_sensitivity},
-    {VP8E_SET_ENABLEAUTOALTREF,         set_enable_auto_alt_ref},
-    {VP8E_SET_SHARPNESS,                set_sharpness},
-    {VP8E_SET_STATIC_THRESHOLD,         set_static_thresh},
-    {VP8E_SET_TOKEN_PARTITIONS,         set_token_partitions},
-    {VP8E_GET_LAST_QUANTIZER,           get_quantizer},
-    {VP8E_GET_LAST_QUANTIZER_64,        get_quantizer64},
-    {VP8E_SET_ARNR_MAXFRAMES,           set_arnr_max_frames},
-    {VP8E_SET_ARNR_STRENGTH ,           set_arnr_strength},
-    {VP8E_SET_ARNR_TYPE     ,           set_arnr_type},
-    {VP8E_SET_TUNING,                   set_tuning},
-    {VP8E_SET_CQ_LEVEL,                 set_cq_level},
-    {VP8E_SET_MAX_INTRA_BITRATE_PCT,    set_rc_max_intra_bitrate_pct},
-    {VP8E_SET_SCREEN_CONTENT_MODE,      set_screen_content_mode},
-    { -1, NULL},
+static vpx_codec_ctrl_fn_map_t vp8e_ctf_maps[] = {
+  { VP8_SET_REFERENCE, vp8e_set_reference },
+  { VP8_COPY_REFERENCE, vp8e_get_reference },
+  { VP8_SET_POSTPROC, vp8e_set_previewpp },
+  { VP8E_SET_FRAME_FLAGS, vp8e_set_frame_flags },
+  { VP8E_SET_TEMPORAL_LAYER_ID, vp8e_set_temporal_layer_id },
+  { VP8E_SET_ROI_MAP, vp8e_set_roi_map },
+  { VP8E_SET_ACTIVEMAP, vp8e_set_activemap },
+  { VP8E_SET_SCALEMODE, vp8e_set_scalemode },
+  { VP8E_SET_CPUUSED, set_cpu_used },
+  { VP8E_SET_NOISE_SENSITIVITY, set_noise_sensitivity },
+  { VP8E_SET_ENABLEAUTOALTREF, set_enable_auto_alt_ref },
+  { VP8E_SET_SHARPNESS, set_sharpness },
+  { VP8E_SET_STATIC_THRESHOLD, set_static_thresh },
+  { VP8E_SET_TOKEN_PARTITIONS, set_token_partitions },
+  { VP8E_GET_LAST_QUANTIZER, get_quantizer },
+  { VP8E_GET_LAST_QUANTIZER_64, get_quantizer64 },
+  { VP8E_SET_ARNR_MAXFRAMES, set_arnr_max_frames },
+  { VP8E_SET_ARNR_STRENGTH, set_arnr_strength },
+  { VP8E_SET_ARNR_TYPE, set_arnr_type },
+  { VP8E_SET_TUNING, set_tuning },
+  { VP8E_SET_CQ_LEVEL, set_cq_level },
+  { VP8E_SET_MAX_INTRA_BITRATE_PCT, set_rc_max_intra_bitrate_pct },
+  { VP8E_SET_SCREEN_CONTENT_MODE, set_screen_content_mode },
+  { -1, NULL },
 };
 
-static vpx_codec_enc_cfg_map_t vp8e_usage_cfg_map[] =
-{
+static vpx_codec_enc_cfg_map_t vp8e_usage_cfg_map[] = {
+  { 0,
     {
-    0,
-    {
-        0,                  /* g_usage */
-        0,                  /* g_threads */
-        0,                  /* g_profile */
+        0, /* g_usage */
+        0, /* g_threads */
+        0, /* g_profile */
 
-        320,                /* g_width */
-        240,                /* g_height */
-        VPX_BITS_8,         /* g_bit_depth */
-        8,                  /* g_input_bit_depth */
+        320,        /* g_width */
+        240,        /* g_height */
+        VPX_BITS_8, /* g_bit_depth */
+        8,          /* g_input_bit_depth */
 
-        {1, 30},            /* g_timebase */
+        { 1, 30 }, /* g_timebase */
 
-        0,                  /* g_error_resilient */
+        0, /* g_error_resilient */
 
-        VPX_RC_ONE_PASS,    /* g_pass */
+        VPX_RC_ONE_PASS, /* g_pass */
 
-        0,                  /* g_lag_in_frames */
+        0, /* g_lag_in_frames */
 
-        0,                  /* rc_dropframe_thresh */
-        0,                  /* rc_resize_allowed */
-        1,                  /* rc_scaled_width */
-        1,                  /* rc_scaled_height */
-        60,                 /* rc_resize_down_thresold */
-        30,                 /* rc_resize_up_thresold */
+        0,  /* rc_dropframe_thresh */
+        0,  /* rc_resize_allowed */
+        1,  /* rc_scaled_width */
+        1,  /* rc_scaled_height */
+        60, /* rc_resize_down_thresold */
+        30, /* rc_resize_up_thresold */
 
-        VPX_VBR,            /* rc_end_usage */
-        {NULL, 0},          /* rc_twopass_stats_in */
-        {NULL, 0},          /* rc_firstpass_mb_stats_in */
-        256,                /* rc_target_bandwidth */
-        4,                  /* rc_min_quantizer */
-        63,                 /* rc_max_quantizer */
-        100,                /* rc_undershoot_pct */
-        100,                /* rc_overshoot_pct */
+        VPX_VBR,     /* rc_end_usage */
+        { NULL, 0 }, /* rc_twopass_stats_in */
+        { NULL, 0 }, /* rc_firstpass_mb_stats_in */
+        256,         /* rc_target_bandwidth */
+        4,           /* rc_min_quantizer */
+        63,          /* rc_max_quantizer */
+        100,         /* rc_undershoot_pct */
+        100,         /* rc_overshoot_pct */
 
-        6000,               /* rc_max_buffer_size */
-        4000,               /* rc_buffer_initial_size; */
-        5000,               /* rc_buffer_optimal_size; */
+        6000, /* rc_max_buffer_size */
+        4000, /* rc_buffer_initial_size; */
+        5000, /* rc_buffer_optimal_size; */
 
-        50,                 /* rc_two_pass_vbrbias  */
-        0,                  /* rc_two_pass_vbrmin_section */
-        400,                /* rc_two_pass_vbrmax_section */
+        50,  /* rc_two_pass_vbrbias  */
+        0,   /* rc_two_pass_vbrmin_section */
+        400, /* rc_two_pass_vbrmax_section */
 
         /* keyframing settings (kf) */
-        VPX_KF_AUTO,        /* g_kfmode*/
-        0,                  /* kf_min_dist */
-        128,                /* kf_max_dist */
+        VPX_KF_AUTO, /* g_kfmode*/
+        0,           /* kf_min_dist */
+        128,         /* kf_max_dist */
 
         VPX_SS_DEFAULT_LAYERS, /* ss_number_layers */
-        {0},
-        {0},                /* ss_target_bitrate */
-        1,                  /* ts_number_layers */
-        {0},                /* ts_target_bitrate */
-        {0},                /* ts_rate_decimator */
-        0,                  /* ts_periodicity */
-        {0},                /* ts_layer_id */
-        {0},                /* layer_target_bitrate */
-        0                   /* temporal_layering_mode */
-    }},
+        { 0 },
+        { 0 }, /* ss_target_bitrate */
+        1,     /* ts_number_layers */
+        { 0 }, /* ts_target_bitrate */
+        { 0 }, /* ts_rate_decimator */
+        0,     /* ts_periodicity */
+        { 0 }, /* ts_layer_id */
+        { 0 }, /* layer_target_bitrate */
+        0      /* temporal_layering_mode */
+    } },
 };
 
-
 #ifndef VERSION_STRING
 #define VERSION_STRING
 #endif
-CODEC_INTERFACE(vpx_codec_vp8_cx) =
-{
-    "WebM Project VP8 Encoder" VERSION_STRING,
-    VPX_CODEC_INTERNAL_ABI_VERSION,
-    VPX_CODEC_CAP_ENCODER | VPX_CODEC_CAP_PSNR |
-    VPX_CODEC_CAP_OUTPUT_PARTITION,
-    /* vpx_codec_caps_t          caps; */
-    vp8e_init,          /* vpx_codec_init_fn_t       init; */
-    vp8e_destroy,       /* vpx_codec_destroy_fn_t    destroy; */
-    vp8e_ctf_maps,      /* vpx_codec_ctrl_fn_map_t  *ctrl_maps; */
-    {
-        NULL,    /* vpx_codec_peek_si_fn_t    peek_si; */
-        NULL,    /* vpx_codec_get_si_fn_t     get_si; */
-        NULL,    /* vpx_codec_decode_fn_t     decode; */
-        NULL,    /* vpx_codec_frame_get_fn_t  frame_get; */
-        NULL,    /* vpx_codec_set_fb_fn_t     set_fb_fn; */
-    },
-    {
-        1,                  /* 1 cfg map */
-        vp8e_usage_cfg_map, /* vpx_codec_enc_cfg_map_t    cfg_maps; */
-        vp8e_encode,        /* vpx_codec_encode_fn_t      encode; */
-        vp8e_get_cxdata,    /* vpx_codec_get_cx_data_fn_t   get_cx_data; */
-        vp8e_set_config,
-        NULL,
-        vp8e_get_preview,
-        vp8e_mr_alloc_mem,
-    } /* encoder functions */
+CODEC_INTERFACE(vpx_codec_vp8_cx) = {
+  "WebM Project VP8 Encoder" VERSION_STRING,
+  VPX_CODEC_INTERNAL_ABI_VERSION,
+  VPX_CODEC_CAP_ENCODER | VPX_CODEC_CAP_PSNR | VPX_CODEC_CAP_OUTPUT_PARTITION,
+  /* vpx_codec_caps_t          caps; */
+  vp8e_init,     /* vpx_codec_init_fn_t       init; */
+  vp8e_destroy,  /* vpx_codec_destroy_fn_t    destroy; */
+  vp8e_ctf_maps, /* vpx_codec_ctrl_fn_map_t  *ctrl_maps; */
+  {
+      NULL, /* vpx_codec_peek_si_fn_t    peek_si; */
+      NULL, /* vpx_codec_get_si_fn_t     get_si; */
+      NULL, /* vpx_codec_decode_fn_t     decode; */
+      NULL, /* vpx_codec_frame_get_fn_t  frame_get; */
+      NULL, /* vpx_codec_set_fb_fn_t     set_fb_fn; */
+  },
+  {
+      1,                  /* 1 cfg map */
+      vp8e_usage_cfg_map, /* vpx_codec_enc_cfg_map_t    cfg_maps; */
+      vp8e_encode,        /* vpx_codec_encode_fn_t      encode; */
+      vp8e_get_cxdata,    /* vpx_codec_get_cx_data_fn_t   get_cx_data; */
+      vp8e_set_config, NULL, vp8e_get_preview, vp8e_mr_alloc_mem,
+  } /* encoder functions */
 };
--- a/vp8/vp8_dx_iface.c
+++ b/vp8/vp8_dx_iface.c
@@ -8,7 +8,6 @@
  *  be found in the AUTHORS file in the root of the source tree.
  */
 
-
 #include <assert.h>
 #include <stdlib.h>
 #include <string.h>
@@ -31,120 +30,111 @@
 #include "decoder/decoderthreading.h"
 
 #define VP8_CAP_POSTPROC (CONFIG_POSTPROC ? VPX_CODEC_CAP_POSTPROC : 0)
-#define VP8_CAP_ERROR_CONCEALMENT (CONFIG_ERROR_CONCEALMENT ? \
-                                    VPX_CODEC_CAP_ERROR_CONCEALMENT : 0)
+#define VP8_CAP_ERROR_CONCEALMENT \
+  (CONFIG_ERROR_CONCEALMENT ? VPX_CODEC_CAP_ERROR_CONCEALMENT : 0)
 
-typedef vpx_codec_stream_info_t  vp8_stream_info_t;
+typedef vpx_codec_stream_info_t vp8_stream_info_t;
 
 /* Structures for handling memory allocations */
-typedef enum
-{
-    VP8_SEG_ALG_PRIV     = 256,
-    VP8_SEG_MAX
-} mem_seg_id_t;
-#define NELEMENTS(x) ((int)(sizeof(x)/sizeof(x[0])))
+typedef enum { VP8_SEG_ALG_PRIV = 256, VP8_SEG_MAX } mem_seg_id_t;
+#define NELEMENTS(x) ((int)(sizeof(x) / sizeof(x[0])))
 
-struct vpx_codec_alg_priv
-{
-    vpx_codec_priv_t        base;
-    vpx_codec_dec_cfg_t     cfg;
-    vp8_stream_info_t       si;
-    int                     decoder_init;
-    int                     postproc_cfg_set;
-    vp8_postproc_cfg_t      postproc_cfg;
+struct vpx_codec_alg_priv {
+  vpx_codec_priv_t base;
+  vpx_codec_dec_cfg_t cfg;
+  vp8_stream_info_t si;
+  int decoder_init;
+  int postproc_cfg_set;
+  vp8_postproc_cfg_t postproc_cfg;
 #if CONFIG_POSTPROC_VISUALIZER
-    unsigned int            dbg_postproc_flag;
-    int                     dbg_color_ref_frame_flag;
-    int                     dbg_color_mb_modes_flag;
-    int                     dbg_color_b_modes_flag;
-    int                     dbg_display_mv_flag;
+  unsigned int dbg_postproc_flag;
+  int dbg_color_ref_frame_flag;
+  int dbg_color_mb_modes_flag;
+  int dbg_color_b_modes_flag;
+  int dbg_display_mv_flag;
 #endif
-    vpx_decrypt_cb          decrypt_cb;
-    void                    *decrypt_state;
-    vpx_image_t             img;
-    int                     img_setup;
-    struct frame_buffers    yv12_frame_buffers;
-    void                    *user_priv;
-    FRAGMENT_DATA           fragments;
+  vpx_decrypt_cb decrypt_cb;
+  void *decrypt_state;
+  vpx_image_t img;
+  int img_setup;
+  struct frame_buffers yv12_frame_buffers;
+  void *user_priv;
+  FRAGMENT_DATA fragments;
 };
 
-static int vp8_init_ctx(vpx_codec_ctx_t *ctx)
-{
-    vpx_codec_alg_priv_t *priv =
-        (vpx_codec_alg_priv_t *)vpx_calloc(1, sizeof(*priv));
-    if (!priv) return 1;
+static int vp8_init_ctx(vpx_codec_ctx_t *ctx) {
+  vpx_codec_alg_priv_t *priv =
+      (vpx_codec_alg_priv_t *)vpx_calloc(1, sizeof(*priv));
+  if (!priv) return 1;
 
-    ctx->priv = (vpx_codec_priv_t *)priv;
-    ctx->priv->init_flags = ctx->init_flags;
+  ctx->priv = (vpx_codec_priv_t *)priv;
+  ctx->priv->init_flags = ctx->init_flags;
 
-    priv->si.sz = sizeof(priv->si);
-    priv->decrypt_cb = NULL;
-    priv->decrypt_state = NULL;
+  priv->si.sz = sizeof(priv->si);
+  priv->decrypt_cb = NULL;
+  priv->decrypt_state = NULL;
 
-    if (ctx->config.dec)
-    {
-        /* Update the reference to the config structure to an internal copy. */
-        priv->cfg = *ctx->config.dec;
-        ctx->config.dec = &priv->cfg;
-    }
+  if (ctx->config.dec) {
+    /* Update the reference to the config structure to an internal copy. */
+    priv->cfg = *ctx->config.dec;
+    ctx->config.dec = &priv->cfg;
+  }
 
-    return 0;
+  return 0;
 }
 
 static vpx_codec_err_t vp8_init(vpx_codec_ctx_t *ctx,
-                                vpx_codec_priv_enc_mr_cfg_t *data)
-{
-    vpx_codec_err_t res = VPX_CODEC_OK;
-    vpx_codec_alg_priv_t *priv = NULL;
-    (void) data;
+                                vpx_codec_priv_enc_mr_cfg_t *data) {
+  vpx_codec_err_t res = VPX_CODEC_OK;
+  vpx_codec_alg_priv_t *priv = NULL;
+  (void)data;
 
-    vp8_rtcd();
-    vpx_dsp_rtcd();
-    vpx_scale_rtcd();
+  vp8_rtcd();
+  vpx_dsp_rtcd();
+  vpx_scale_rtcd();
 
-    /* This function only allocates space for the vpx_codec_alg_priv_t
-     * structure. More memory may be required at the time the stream
-     * information becomes known.
-     */
-    if (!ctx->priv) {
-      if (vp8_init_ctx(ctx)) return VPX_CODEC_MEM_ERROR;
-      priv = (vpx_codec_alg_priv_t *)ctx->priv;
+  /* This function only allocates space for the vpx_codec_alg_priv_t
+   * structure. More memory may be required at the time the stream
+   * information becomes known.
+   */
+  if (!ctx->priv) {
+    if (vp8_init_ctx(ctx)) return VPX_CODEC_MEM_ERROR;
+    priv = (vpx_codec_alg_priv_t *)ctx->priv;
 
-      /* initialize number of fragments to zero */
-      priv->fragments.count = 0;
-      /* is input fragments enabled? */
-      priv->fragments.enabled =
-          (priv->base.init_flags & VPX_CODEC_USE_INPUT_FRAGMENTS);
+    /* initialize number of fragments to zero */
+    priv->fragments.count = 0;
+    /* is input fragments enabled? */
+    priv->fragments.enabled =
+        (priv->base.init_flags & VPX_CODEC_USE_INPUT_FRAGMENTS);
 
-      /*post processing level initialized to do nothing */
-    } else {
-      priv = (vpx_codec_alg_priv_t *)ctx->priv;
-    }
+    /*post processing level initialized to do nothing */
+  } else {
+    priv = (vpx_codec_alg_priv_t *)ctx->priv;
+  }
 
-    priv->yv12_frame_buffers.use_frame_threads =
-        (ctx->priv->init_flags & VPX_CODEC_USE_FRAME_THREADING);
+  priv->yv12_frame_buffers.use_frame_threads =
+      (ctx->priv->init_flags & VPX_CODEC_USE_FRAME_THREADING);
 
-    /* for now, disable frame threading */
-    priv->yv12_frame_buffers.use_frame_threads = 0;
+  /* for now, disable frame threading */
+  priv->yv12_frame_buffers.use_frame_threads = 0;
 
-    if (priv->yv12_frame_buffers.use_frame_threads &&
-        ((ctx->priv->init_flags & VPX_CODEC_USE_ERROR_CONCEALMENT) ||
-         (ctx->priv->init_flags & VPX_CODEC_USE_INPUT_FRAGMENTS))) {
-      /* row-based threading, error concealment, and input fragments will
-       * not be supported when using frame-based threading */
-      res = VPX_CODEC_INVALID_PARAM;
-    }
+  if (priv->yv12_frame_buffers.use_frame_threads &&
+      ((ctx->priv->init_flags & VPX_CODEC_USE_ERROR_CONCEALMENT) ||
+       (ctx->priv->init_flags & VPX_CODEC_USE_INPUT_FRAGMENTS))) {
+    /* row-based threading, error concealment, and input fragments will
+     * not be supported when using frame-based threading */
+    res = VPX_CODEC_INVALID_PARAM;
+  }
 
-    return res;
+  return res;
 }
 
-static vpx_codec_err_t vp8_destroy(vpx_codec_alg_priv_t *ctx)
-{
-    vp8_remove_decoder_instances(&ctx->yv12_frame_buffers);
+static vpx_codec_err_t vp8_destroy(vpx_codec_alg_priv_t *ctx) {
+  vp8_remove_decoder_instances(&ctx->yv12_frame_buffers);
 
-    vpx_free(ctx);
+  vpx_free(ctx);
 
-    return VPX_CODEC_OK;
+  return VPX_CODEC_OK;
 }
 
 static vpx_codec_err_t vp8_peek_si_internal(const uint8_t *data,
@@ -151,508 +141,444 @@
                                             unsigned int data_sz,
                                             vpx_codec_stream_info_t *si,
                                             vpx_decrypt_cb decrypt_cb,
-                                            void *decrypt_state)
-{
-    vpx_codec_err_t res = VPX_CODEC_OK;
+                                            void *decrypt_state) {
+  vpx_codec_err_t res = VPX_CODEC_OK;
 
-    assert(data != NULL);
+  assert(data != NULL);
 
-    if(data + data_sz <= data)
-    {
-        res = VPX_CODEC_INVALID_PARAM;
+  if (data + data_sz <= data) {
+    res = VPX_CODEC_INVALID_PARAM;
+  } else {
+    /* Parse uncompresssed part of key frame header.
+     * 3 bytes:- including version, frame type and an offset
+     * 3 bytes:- sync code (0x9d, 0x01, 0x2a)
+     * 4 bytes:- including image width and height in the lowest 14 bits
+     *           of each 2-byte value.
+     */
+    uint8_t clear_buffer[10];
+    const uint8_t *clear = data;
+    if (decrypt_cb) {
+      int n = VPXMIN(sizeof(clear_buffer), data_sz);
+      decrypt_cb(decrypt_state, data, clear_buffer, n);
+      clear = clear_buffer;
     }
-    else
+    si->is_kf = 0;
+
+    if (data_sz >= 10 && !(clear[0] & 0x01)) /* I-Frame */
     {
-        /* Parse uncompresssed part of key frame header.
-         * 3 bytes:- including version, frame type and an offset
-         * 3 bytes:- sync code (0x9d, 0x01, 0x2a)
-         * 4 bytes:- including image width and height in the lowest 14 bits
-         *           of each 2-byte value.
-         */
-        uint8_t clear_buffer[10];
-        const uint8_t *clear = data;
-        if (decrypt_cb)
-        {
-            int n = VPXMIN(sizeof(clear_buffer), data_sz);
-            decrypt_cb(decrypt_state, data, clear_buffer, n);
-            clear = clear_buffer;
-        }
-        si->is_kf = 0;
+      si->is_kf = 1;
 
-        if (data_sz >= 10 && !(clear[0] & 0x01))  /* I-Frame */
-        {
-            si->is_kf = 1;
+      /* vet via sync code */
+      if (clear[3] != 0x9d || clear[4] != 0x01 || clear[5] != 0x2a)
+        return VPX_CODEC_UNSUP_BITSTREAM;
 
-            /* vet via sync code */
-            if (clear[3] != 0x9d || clear[4] != 0x01 || clear[5] != 0x2a)
-                return VPX_CODEC_UNSUP_BITSTREAM;
+      si->w = (clear[6] | (clear[7] << 8)) & 0x3fff;
+      si->h = (clear[8] | (clear[9] << 8)) & 0x3fff;
 
-            si->w = (clear[6] | (clear[7] << 8)) & 0x3fff;
-            si->h = (clear[8] | (clear[9] << 8)) & 0x3fff;
-
-            /*printf("w=%d, h=%d\n", si->w, si->h);*/
-            if (!(si->h | si->w))
-                res = VPX_CODEC_UNSUP_BITSTREAM;
-        }
-        else
-        {
-            res = VPX_CODEC_UNSUP_BITSTREAM;
-        }
+      /*printf("w=%d, h=%d\n", si->w, si->h);*/
+      if (!(si->h | si->w)) res = VPX_CODEC_UNSUP_BITSTREAM;
+    } else {
+      res = VPX_CODEC_UNSUP_BITSTREAM;
     }
+  }
 
-    return res;
+  return res;
 }
 
-static vpx_codec_err_t vp8_peek_si(const uint8_t *data,
-                                   unsigned int data_sz,
+static vpx_codec_err_t vp8_peek_si(const uint8_t *data, unsigned int data_sz,
                                    vpx_codec_stream_info_t *si) {
-    return vp8_peek_si_internal(data, data_sz, si, NULL, NULL);
+  return vp8_peek_si_internal(data, data_sz, si, NULL, NULL);
 }
 
-static vpx_codec_err_t vp8_get_si(vpx_codec_alg_priv_t    *ctx,
-                                  vpx_codec_stream_info_t *si)
-{
+static vpx_codec_err_t vp8_get_si(vpx_codec_alg_priv_t *ctx,
+                                  vpx_codec_stream_info_t *si) {
+  unsigned int sz;
 
-    unsigned int sz;
+  if (si->sz >= sizeof(vp8_stream_info_t))
+    sz = sizeof(vp8_stream_info_t);
+  else
+    sz = sizeof(vpx_codec_stream_info_t);
 
-    if (si->sz >= sizeof(vp8_stream_info_t))
-        sz = sizeof(vp8_stream_info_t);
-    else
-        sz = sizeof(vpx_codec_stream_info_t);
+  memcpy(si, &ctx->si, sz);
+  si->sz = sz;
 
-    memcpy(si, &ctx->si, sz);
-    si->sz = sz;
-
-    return VPX_CODEC_OK;
+  return VPX_CODEC_OK;
 }
 
+static vpx_codec_err_t update_error_state(
+    vpx_codec_alg_priv_t *ctx, const struct vpx_internal_error_info *error) {
+  vpx_codec_err_t res;
 
-static vpx_codec_err_t
-update_error_state(vpx_codec_alg_priv_t                 *ctx,
-                   const struct vpx_internal_error_info *error)
-{
-    vpx_codec_err_t res;
+  if ((res = error->error_code))
+    ctx->base.err_detail = error->has_detail ? error->detail : NULL;
 
-    if ((res = error->error_code))
-        ctx->base.err_detail = error->has_detail
-                               ? error->detail
-                               : NULL;
-
-    return res;
+  return res;
 }
 
-static void yuvconfig2image(vpx_image_t               *img,
-                            const YV12_BUFFER_CONFIG  *yv12,
-                            void                      *user_priv)
-{
-    /** vpx_img_wrap() doesn't allow specifying independent strides for
-      * the Y, U, and V planes, nor other alignment adjustments that
-      * might be representable by a YV12_BUFFER_CONFIG, so we just
-      * initialize all the fields.*/
-    img->fmt = VPX_IMG_FMT_I420;
-    img->w = yv12->y_stride;
-    img->h = (yv12->y_height + 2 * VP8BORDERINPIXELS + 15) & ~15;
-    img->d_w = img->r_w = yv12->y_width;
-    img->d_h = img->r_h = yv12->y_height;
-    img->x_chroma_shift = 1;
-    img->y_chroma_shift = 1;
-    img->planes[VPX_PLANE_Y] = yv12->y_buffer;
-    img->planes[VPX_PLANE_U] = yv12->u_buffer;
-    img->planes[VPX_PLANE_V] = yv12->v_buffer;
-    img->planes[VPX_PLANE_ALPHA] = NULL;
-    img->stride[VPX_PLANE_Y] = yv12->y_stride;
-    img->stride[VPX_PLANE_U] = yv12->uv_stride;
-    img->stride[VPX_PLANE_V] = yv12->uv_stride;
-    img->stride[VPX_PLANE_ALPHA] = yv12->y_stride;
-    img->bit_depth = 8;
-    img->bps = 12;
-    img->user_priv = user_priv;
-    img->img_data = yv12->buffer_alloc;
-    img->img_data_owner = 0;
-    img->self_allocd = 0;
+static void yuvconfig2image(vpx_image_t *img, const YV12_BUFFER_CONFIG *yv12,
+                            void *user_priv) {
+  /** vpx_img_wrap() doesn't allow specifying independent strides for
+    * the Y, U, and V planes, nor other alignment adjustments that
+    * might be representable by a YV12_BUFFER_CONFIG, so we just
+    * initialize all the fields.*/
+  img->fmt = VPX_IMG_FMT_I420;
+  img->w = yv12->y_stride;
+  img->h = (yv12->y_height + 2 * VP8BORDERINPIXELS + 15) & ~15;
+  img->d_w = img->r_w = yv12->y_width;
+  img->d_h = img->r_h = yv12->y_height;
+  img->x_chroma_shift = 1;
+  img->y_chroma_shift = 1;
+  img->planes[VPX_PLANE_Y] = yv12->y_buffer;
+  img->planes[VPX_PLANE_U] = yv12->u_buffer;
+  img->planes[VPX_PLANE_V] = yv12->v_buffer;
+  img->planes[VPX_PLANE_ALPHA] = NULL;
+  img->stride[VPX_PLANE_Y] = yv12->y_stride;
+  img->stride[VPX_PLANE_U] = yv12->uv_stride;
+  img->stride[VPX_PLANE_V] = yv12->uv_stride;
+  img->stride[VPX_PLANE_ALPHA] = yv12->y_stride;
+  img->bit_depth = 8;
+  img->bps = 12;
+  img->user_priv = user_priv;
+  img->img_data = yv12->buffer_alloc;
+  img->img_data_owner = 0;
+  img->self_allocd = 0;
 }
 
-static int
-update_fragments(vpx_codec_alg_priv_t  *ctx,
-                 const uint8_t         *data,
-                 unsigned int           data_sz,
-                 vpx_codec_err_t       *res)
-{
-    *res = VPX_CODEC_OK;
+static int update_fragments(vpx_codec_alg_priv_t *ctx, const uint8_t *data,
+                            unsigned int data_sz, vpx_codec_err_t *res) {
+  *res = VPX_CODEC_OK;
 
-    if (ctx->fragments.count == 0)
-    {
-        /* New frame, reset fragment pointers and sizes */
-        memset((void*)ctx->fragments.ptrs, 0, sizeof(ctx->fragments.ptrs));
-        memset(ctx->fragments.sizes, 0, sizeof(ctx->fragments.sizes));
+  if (ctx->fragments.count == 0) {
+    /* New frame, reset fragment pointers and sizes */
+    memset((void *)ctx->fragments.ptrs, 0, sizeof(ctx->fragments.ptrs));
+    memset(ctx->fragments.sizes, 0, sizeof(ctx->fragments.sizes));
+  }
+  if (ctx->fragments.enabled && !(data == NULL && data_sz == 0)) {
+    /* Store a pointer to this fragment and return. We haven't
+     * received the complete frame yet, so we will wait with decoding.
+     */
+    ctx->fragments.ptrs[ctx->fragments.count] = data;
+    ctx->fragments.sizes[ctx->fragments.count] = data_sz;
+    ctx->fragments.count++;
+    if (ctx->fragments.count > (1 << EIGHT_PARTITION) + 1) {
+      ctx->fragments.count = 0;
+      *res = VPX_CODEC_INVALID_PARAM;
+      return -1;
     }
-    if (ctx->fragments.enabled && !(data == NULL && data_sz == 0))
-    {
-        /* Store a pointer to this fragment and return. We haven't
-         * received the complete frame yet, so we will wait with decoding.
-         */
-        ctx->fragments.ptrs[ctx->fragments.count] = data;
-        ctx->fragments.sizes[ctx->fragments.count] = data_sz;
-        ctx->fragments.count++;
-        if (ctx->fragments.count > (1 << EIGHT_PARTITION) + 1)
-        {
-            ctx->fragments.count = 0;
-            *res = VPX_CODEC_INVALID_PARAM;
-            return -1;
-        }
-        return 0;
-    }
+    return 0;
+  }
 
-    if (!ctx->fragments.enabled && (data == NULL && data_sz == 0))
-    {
-        return 0;
-    }
+  if (!ctx->fragments.enabled && (data == NULL && data_sz == 0)) {
+    return 0;
+  }
 
-    if (!ctx->fragments.enabled)
-    {
-        ctx->fragments.ptrs[0] = data;
-        ctx->fragments.sizes[0] = data_sz;
-        ctx->fragments.count = 1;
-    }
+  if (!ctx->fragments.enabled) {
+    ctx->fragments.ptrs[0] = data;
+    ctx->fragments.sizes[0] = data_sz;
+    ctx->fragments.count = 1;
+  }
 
-    return 1;
+  return 1;
 }
 
-static vpx_codec_err_t vp8_decode(vpx_codec_alg_priv_t  *ctx,
-                                  const uint8_t         *data,
-                                  unsigned int            data_sz,
-                                  void                    *user_priv,
-                                  long                    deadline)
-{
-    vpx_codec_err_t res = VPX_CODEC_OK;
-    unsigned int resolution_change = 0;
-    unsigned int w, h;
+static vpx_codec_err_t vp8_decode(vpx_codec_alg_priv_t *ctx,
+                                  const uint8_t *data, unsigned int data_sz,
+                                  void *user_priv, long deadline) {
+  vpx_codec_err_t res = VPX_CODEC_OK;
+  unsigned int resolution_change = 0;
+  unsigned int w, h;
 
-    if (!ctx->fragments.enabled && (data == NULL && data_sz == 0))
-    {
-        return 0;
-    }
+  if (!ctx->fragments.enabled && (data == NULL && data_sz == 0)) {
+    return 0;
+  }
 
-    /* Update the input fragment data */
-    if(update_fragments(ctx, data, data_sz, &res) <= 0)
-        return res;
+  /* Update the input fragment data */
+  if (update_fragments(ctx, data, data_sz, &res) <= 0) return res;
 
-    /* Determine the stream parameters. Note that we rely on peek_si to
-     * validate that we have a buffer that does not wrap around the top
-     * of the heap.
-     */
-    w = ctx->si.w;
-    h = ctx->si.h;
+  /* Determine the stream parameters. Note that we rely on peek_si to
+   * validate that we have a buffer that does not wrap around the top
+   * of the heap.
+   */
+  w = ctx->si.w;
+  h = ctx->si.h;
 
-    res = vp8_peek_si_internal(ctx->fragments.ptrs[0], ctx->fragments.sizes[0],
-                               &ctx->si, ctx->decrypt_cb, ctx->decrypt_state);
+  res = vp8_peek_si_internal(ctx->fragments.ptrs[0], ctx->fragments.sizes[0],
+                             &ctx->si, ctx->decrypt_cb, ctx->decrypt_state);
 
-    if((res == VPX_CODEC_UNSUP_BITSTREAM) && !ctx->si.is_kf)
-    {
-        /* the peek function returns an error for non keyframes, however for
-         * this case, it is not an error */
-        res = VPX_CODEC_OK;
-    }
+  if ((res == VPX_CODEC_UNSUP_BITSTREAM) && !ctx->si.is_kf) {
+    /* the peek function returns an error for non keyframes, however for
+     * this case, it is not an error */
+    res = VPX_CODEC_OK;
+  }
 
-    if(!ctx->decoder_init && !ctx->si.is_kf)
-        res = VPX_CODEC_UNSUP_BITSTREAM;
+  if (!ctx->decoder_init && !ctx->si.is_kf) res = VPX_CODEC_UNSUP_BITSTREAM;
 
-    if ((ctx->si.h != h) || (ctx->si.w != w))
-        resolution_change = 1;
+  if ((ctx->si.h != h) || (ctx->si.w != w)) resolution_change = 1;
 
-    /* Initialize the decoder instance on the first frame*/
-    if (!res && !ctx->decoder_init)
-    {
-      VP8D_CONFIG oxcf;
+  /* Initialize the decoder instance on the first frame*/
+  if (!res && !ctx->decoder_init) {
+    VP8D_CONFIG oxcf;
 
-      oxcf.Width = ctx->si.w;
-      oxcf.Height = ctx->si.h;
-      oxcf.Version = 9;
-      oxcf.postprocess = 0;
-      oxcf.max_threads = ctx->cfg.threads;
-      oxcf.error_concealment =
-          (ctx->base.init_flags & VPX_CODEC_USE_ERROR_CONCEALMENT);
+    oxcf.Width = ctx->si.w;
+    oxcf.Height = ctx->si.h;
+    oxcf.Version = 9;
+    oxcf.postprocess = 0;
+    oxcf.max_threads = ctx->cfg.threads;
+    oxcf.error_concealment =
+        (ctx->base.init_flags & VPX_CODEC_USE_ERROR_CONCEALMENT);
 
-      /* If postprocessing was enabled by the application and a
-       * configuration has not been provided, default it.
-       */
-       if (!ctx->postproc_cfg_set
-           && (ctx->base.init_flags & VPX_CODEC_USE_POSTPROC)) {
-         ctx->postproc_cfg.post_proc_flag =
-             VP8_DEBLOCK | VP8_DEMACROBLOCK | VP8_MFQE;
-         ctx->postproc_cfg.deblocking_level = 4;
-         ctx->postproc_cfg.noise_level = 0;
-       }
-
-       res = vp8_create_decoder_instances(&ctx->yv12_frame_buffers, &oxcf);
-       ctx->decoder_init = 1;
-    }
-
-    /* Set these even if already initialized.  The caller may have changed the
-     * decrypt config between frames.
+    /* If postprocessing was enabled by the application and a
+     * configuration has not been provided, default it.
      */
-    if (ctx->decoder_init) {
-      ctx->yv12_frame_buffers.pbi[0]->decrypt_cb = ctx->decrypt_cb;
-      ctx->yv12_frame_buffers.pbi[0]->decrypt_state = ctx->decrypt_state;
+    if (!ctx->postproc_cfg_set &&
+        (ctx->base.init_flags & VPX_CODEC_USE_POSTPROC)) {
+      ctx->postproc_cfg.post_proc_flag =
+          VP8_DEBLOCK | VP8_DEMACROBLOCK | VP8_MFQE;
+      ctx->postproc_cfg.deblocking_level = 4;
+      ctx->postproc_cfg.noise_level = 0;
     }
 
-    if (!res)
-    {
-        VP8D_COMP *pbi = ctx->yv12_frame_buffers.pbi[0];
-        if (resolution_change)
-        {
-            VP8_COMMON *const pc = & pbi->common;
-            MACROBLOCKD *const xd  = & pbi->mb;
+    res = vp8_create_decoder_instances(&ctx->yv12_frame_buffers, &oxcf);
+    ctx->decoder_init = 1;
+  }
+
+  /* Set these even if already initialized.  The caller may have changed the
+   * decrypt config between frames.
+   */
+  if (ctx->decoder_init) {
+    ctx->yv12_frame_buffers.pbi[0]->decrypt_cb = ctx->decrypt_cb;
+    ctx->yv12_frame_buffers.pbi[0]->decrypt_state = ctx->decrypt_state;
+  }
+
+  if (!res) {
+    VP8D_COMP *pbi = ctx->yv12_frame_buffers.pbi[0];
+    if (resolution_change) {
+      VP8_COMMON *const pc = &pbi->common;
+      MACROBLOCKD *const xd = &pbi->mb;
 #if CONFIG_MULTITHREAD
-            int i;
+      int i;
 #endif
-            pc->Width = ctx->si.w;
-            pc->Height = ctx->si.h;
-            {
-                int prev_mb_rows = pc->mb_rows;
+      pc->Width = ctx->si.w;
+      pc->Height = ctx->si.h;
+      {
+        int prev_mb_rows = pc->mb_rows;
 
-                if (setjmp(pbi->common.error.jmp))
-                {
-                    pbi->common.error.setjmp = 0;
-                    vp8_clear_system_state();
-                    /* same return value as used in vp8dx_receive_compressed_data */
-                    return -1;
-                }
+        if (setjmp(pbi->common.error.jmp)) {
+          pbi->common.error.setjmp = 0;
+          vp8_clear_system_state();
+          /* same return value as used in vp8dx_receive_compressed_data */
+          return -1;
+        }
 
-                pbi->common.error.setjmp = 1;
+        pbi->common.error.setjmp = 1;
 
-                if (pc->Width <= 0)
-                {
-                    pc->Width = w;
-                    vpx_internal_error(&pc->error, VPX_CODEC_CORRUPT_FRAME,
-                                       "Invalid frame width");
-                }
+        if (pc->Width <= 0) {
+          pc->Width = w;
+          vpx_internal_error(&pc->error, VPX_CODEC_CORRUPT_FRAME,
+                             "Invalid frame width");
+        }
 
-                if (pc->Height <= 0)
-                {
-                    pc->Height = h;
-                    vpx_internal_error(&pc->error, VPX_CODEC_CORRUPT_FRAME,
-                                       "Invalid frame height");
-                }
+        if (pc->Height <= 0) {
+          pc->Height = h;
+          vpx_internal_error(&pc->error, VPX_CODEC_CORRUPT_FRAME,
+                             "Invalid frame height");
+        }
 
-                if (vp8_alloc_frame_buffers(pc, pc->Width, pc->Height))
-                    vpx_internal_error(&pc->error, VPX_CODEC_MEM_ERROR,
-                                       "Failed to allocate frame buffers");
+        if (vp8_alloc_frame_buffers(pc, pc->Width, pc->Height))
+          vpx_internal_error(&pc->error, VPX_CODEC_MEM_ERROR,
+                             "Failed to allocate frame buffers");
 
-                xd->pre = pc->yv12_fb[pc->lst_fb_idx];
-                xd->dst = pc->yv12_fb[pc->new_fb_idx];
+        xd->pre = pc->yv12_fb[pc->lst_fb_idx];
+        xd->dst = pc->yv12_fb[pc->new_fb_idx];
 
 #if CONFIG_MULTITHREAD
-                for (i = 0; i < pbi->allocated_decoding_thread_count; i++)
-                {
-                    pbi->mb_row_di[i].mbd.dst = pc->yv12_fb[pc->new_fb_idx];
-                    vp8_build_block_doffsets(&pbi->mb_row_di[i].mbd);
-                }
+        for (i = 0; i < pbi->allocated_decoding_thread_count; i++) {
+          pbi->mb_row_di[i].mbd.dst = pc->yv12_fb[pc->new_fb_idx];
+          vp8_build_block_doffsets(&pbi->mb_row_di[i].mbd);
+        }
 #endif
-                vp8_build_block_doffsets(&pbi->mb);
+        vp8_build_block_doffsets(&pbi->mb);
 
-                /* allocate memory for last frame MODE_INFO array */
+/* allocate memory for last frame MODE_INFO array */
 #if CONFIG_ERROR_CONCEALMENT
 
-                if (pbi->ec_enabled)
-                {
-                    /* old prev_mip was released by vp8_de_alloc_frame_buffers()
-                     * called in vp8_alloc_frame_buffers() */
-                    pc->prev_mip = vpx_calloc(
-                                       (pc->mb_cols + 1) * (pc->mb_rows + 1),
-                                       sizeof(MODE_INFO));
+        if (pbi->ec_enabled) {
+          /* old prev_mip was released by vp8_de_alloc_frame_buffers()
+           * called in vp8_alloc_frame_buffers() */
+          pc->prev_mip = vpx_calloc((pc->mb_cols + 1) * (pc->mb_rows + 1),
+                                    sizeof(MODE_INFO));
 
-                    if (!pc->prev_mip)
-                    {
-                        vp8_de_alloc_frame_buffers(pc);
-                        vpx_internal_error(&pc->error, VPX_CODEC_MEM_ERROR,
-                                           "Failed to allocate"
-                                           "last frame MODE_INFO array");
-                    }
+          if (!pc->prev_mip) {
+            vp8_de_alloc_frame_buffers(pc);
+            vpx_internal_error(&pc->error, VPX_CODEC_MEM_ERROR,
+                               "Failed to allocate"
+                               "last frame MODE_INFO array");
+          }
 
-                    pc->prev_mi = pc->prev_mip + pc->mode_info_stride + 1;
+          pc->prev_mi = pc->prev_mip + pc->mode_info_stride + 1;
 
-                    if (vp8_alloc_overlap_lists(pbi))
-                        vpx_internal_error(&pc->error, VPX_CODEC_MEM_ERROR,
-                                           "Failed to allocate overlap lists "
-                                           "for error concealment");
-                }
+          if (vp8_alloc_overlap_lists(pbi))
+            vpx_internal_error(&pc->error, VPX_CODEC_MEM_ERROR,
+                               "Failed to allocate overlap lists "
+                               "for error concealment");
+        }
 
 #endif
 
 #if CONFIG_MULTITHREAD
-                if (pbi->b_multithreaded_rd)
-                    vp8mt_alloc_temp_buffers(pbi, pc->Width, prev_mb_rows);
+        if (pbi->b_multithreaded_rd)
+          vp8mt_alloc_temp_buffers(pbi, pc->Width, prev_mb_rows);
 #else
-                (void)prev_mb_rows;
+        (void)prev_mb_rows;
 #endif
-            }
+      }
 
-            pbi->common.error.setjmp = 0;
+      pbi->common.error.setjmp = 0;
 
-            /* required to get past the first get_free_fb() call */
-            pbi->common.fb_idx_ref_cnt[0] = 0;
-        }
+      /* required to get past the first get_free_fb() call */
+      pbi->common.fb_idx_ref_cnt[0] = 0;
+    }
 
-        /* update the pbi fragment data */
-        pbi->fragments = ctx->fragments;
+    /* update the pbi fragment data */
+    pbi->fragments = ctx->fragments;
 
-        ctx->user_priv = user_priv;
-        if (vp8dx_receive_compressed_data(pbi, data_sz, data, deadline))
-        {
-            res = update_error_state(ctx, &pbi->common.error);
-        }
-
-        /* get ready for the next series of fragments */
-        ctx->fragments.count = 0;
+    ctx->user_priv = user_priv;
+    if (vp8dx_receive_compressed_data(pbi, data_sz, data, deadline)) {
+      res = update_error_state(ctx, &pbi->common.error);
     }
 
-    return res;
+    /* get ready for the next series of fragments */
+    ctx->fragments.count = 0;
+  }
+
+  return res;
 }
 
-static vpx_image_t *vp8_get_frame(vpx_codec_alg_priv_t  *ctx,
-                                  vpx_codec_iter_t      *iter)
-{
-    vpx_image_t *img = NULL;
+static vpx_image_t *vp8_get_frame(vpx_codec_alg_priv_t *ctx,
+                                  vpx_codec_iter_t *iter) {
+  vpx_image_t *img = NULL;
 
-    /* iter acts as a flip flop, so an image is only returned on the first
-     * call to get_frame.
-     */
-    if (!(*iter) && ctx->yv12_frame_buffers.pbi[0])
-    {
-        YV12_BUFFER_CONFIG sd;
-        int64_t time_stamp = 0, time_end_stamp = 0;
-        vp8_ppflags_t flags;
-        vp8_zero(flags);
+  /* iter acts as a flip flop, so an image is only returned on the first
+   * call to get_frame.
+   */
+  if (!(*iter) && ctx->yv12_frame_buffers.pbi[0]) {
+    YV12_BUFFER_CONFIG sd;
+    int64_t time_stamp = 0, time_end_stamp = 0;
+    vp8_ppflags_t flags;
+    vp8_zero(flags);
 
-        if (ctx->base.init_flags & VPX_CODEC_USE_POSTPROC)
-        {
-            flags.post_proc_flag= ctx->postproc_cfg.post_proc_flag
+    if (ctx->base.init_flags & VPX_CODEC_USE_POSTPROC) {
+      flags.post_proc_flag =
+          ctx->postproc_cfg.post_proc_flag
 #if CONFIG_POSTPROC_VISUALIZER
 
-                                | ((ctx->dbg_color_ref_frame_flag != 0) ? VP8D_DEBUG_CLR_FRM_REF_BLKS : 0)
-                                | ((ctx->dbg_color_mb_modes_flag != 0) ? VP8D_DEBUG_CLR_BLK_MODES : 0)
-                                | ((ctx->dbg_color_b_modes_flag != 0) ? VP8D_DEBUG_CLR_BLK_MODES : 0)
-                                | ((ctx->dbg_display_mv_flag != 0) ? VP8D_DEBUG_DRAW_MV : 0)
+          | ((ctx->dbg_color_ref_frame_flag != 0) ? VP8D_DEBUG_CLR_FRM_REF_BLKS
+                                                  : 0) |
+          ((ctx->dbg_color_mb_modes_flag != 0) ? VP8D_DEBUG_CLR_BLK_MODES : 0) |
+          ((ctx->dbg_color_b_modes_flag != 0) ? VP8D_DEBUG_CLR_BLK_MODES : 0) |
+          ((ctx->dbg_display_mv_flag != 0) ? VP8D_DEBUG_DRAW_MV : 0)
 #endif
-                                ;
-            flags.deblocking_level      = ctx->postproc_cfg.deblocking_level;
-            flags.noise_level           = ctx->postproc_cfg.noise_level;
+          ;
+      flags.deblocking_level = ctx->postproc_cfg.deblocking_level;
+      flags.noise_level = ctx->postproc_cfg.noise_level;
 #if CONFIG_POSTPROC_VISUALIZER
-            flags.display_ref_frame_flag= ctx->dbg_color_ref_frame_flag;
-            flags.display_mb_modes_flag = ctx->dbg_color_mb_modes_flag;
-            flags.display_b_modes_flag  = ctx->dbg_color_b_modes_flag;
-            flags.display_mv_flag       = ctx->dbg_display_mv_flag;
+      flags.display_ref_frame_flag = ctx->dbg_color_ref_frame_flag;
+      flags.display_mb_modes_flag = ctx->dbg_color_mb_modes_flag;
+      flags.display_b_modes_flag = ctx->dbg_color_b_modes_flag;
+      flags.display_mv_flag = ctx->dbg_display_mv_flag;
 #endif
-        }
+    }
 
-        if (0 == vp8dx_get_raw_frame(ctx->yv12_frame_buffers.pbi[0], &sd,
-                                     &time_stamp, &time_end_stamp, &flags))
-        {
-            yuvconfig2image(&ctx->img, &sd, ctx->user_priv);
+    if (0 == vp8dx_get_raw_frame(ctx->yv12_frame_buffers.pbi[0], &sd,
+                                 &time_stamp, &time_end_stamp, &flags)) {
+      yuvconfig2image(&ctx->img, &sd, ctx->user_priv);
 
-            img = &ctx->img;
-            *iter = img;
-        }
+      img = &ctx->img;
+      *iter = img;
     }
+  }
 
-    return img;
+  return img;
 }
 
-static vpx_codec_err_t image2yuvconfig(const vpx_image_t   *img,
-                                       YV12_BUFFER_CONFIG  *yv12)
-{
-    const int y_w = img->d_w;
-    const int y_h = img->d_h;
-    const int uv_w = (img->d_w + 1) / 2;
-    const int uv_h = (img->d_h + 1) / 2;
-    vpx_codec_err_t        res = VPX_CODEC_OK;
-    yv12->y_buffer = img->planes[VPX_PLANE_Y];
-    yv12->u_buffer = img->planes[VPX_PLANE_U];
-    yv12->v_buffer = img->planes[VPX_PLANE_V];
+static vpx_codec_err_t image2yuvconfig(const vpx_image_t *img,
+                                       YV12_BUFFER_CONFIG *yv12) {
+  const int y_w = img->d_w;
+  const int y_h = img->d_h;
+  const int uv_w = (img->d_w + 1) / 2;
+  const int uv_h = (img->d_h + 1) / 2;
+  vpx_codec_err_t res = VPX_CODEC_OK;
+  yv12->y_buffer = img->planes[VPX_PLANE_Y];
+  yv12->u_buffer = img->planes[VPX_PLANE_U];
+  yv12->v_buffer = img->planes[VPX_PLANE_V];
 
-    yv12->y_crop_width  = y_w;
-    yv12->y_crop_height = y_h;
-    yv12->y_width  = y_w;
-    yv12->y_height = y_h;
-    yv12->uv_crop_width = uv_w;
-    yv12->uv_crop_height = uv_h;
-    yv12->uv_width = uv_w;
-    yv12->uv_height = uv_h;
+  yv12->y_crop_width = y_w;
+  yv12->y_crop_height = y_h;
+  yv12->y_width = y_w;
+  yv12->y_height = y_h;
+  yv12->uv_crop_width = uv_w;
+  yv12->uv_crop_height = uv_h;
+  yv12->uv_width = uv_w;
+  yv12->uv_height = uv_h;
 
-    yv12->y_stride = img->stride[VPX_PLANE_Y];
-    yv12->uv_stride = img->stride[VPX_PLANE_U];
+  yv12->y_stride = img->stride[VPX_PLANE_Y];
+  yv12->uv_stride = img->stride[VPX_PLANE_U];
 
-    yv12->border  = (img->stride[VPX_PLANE_Y] - img->d_w) / 2;
-    return res;
+  yv12->border = (img->stride[VPX_PLANE_Y] - img->d_w) / 2;
+  return res;
 }
 
-
 static vpx_codec_err_t vp8_set_reference(vpx_codec_alg_priv_t *ctx,
-                                         va_list args)
-{
+                                         va_list args) {
+  vpx_ref_frame_t *data = va_arg(args, vpx_ref_frame_t *);
 
-    vpx_ref_frame_t *data = va_arg(args, vpx_ref_frame_t *);
+  if (data && !ctx->yv12_frame_buffers.use_frame_threads) {
+    vpx_ref_frame_t *frame = (vpx_ref_frame_t *)data;
+    YV12_BUFFER_CONFIG sd;
 
-    if (data && !ctx->yv12_frame_buffers.use_frame_threads)
-    {
-        vpx_ref_frame_t *frame = (vpx_ref_frame_t *)data;
-        YV12_BUFFER_CONFIG sd;
+    image2yuvconfig(&frame->img, &sd);
 
-        image2yuvconfig(&frame->img, &sd);
-
-        return vp8dx_set_reference(ctx->yv12_frame_buffers.pbi[0],
-                                   frame->frame_type, &sd);
-    }
-    else
-        return VPX_CODEC_INVALID_PARAM;
-
+    return vp8dx_set_reference(ctx->yv12_frame_buffers.pbi[0],
+                               frame->frame_type, &sd);
+  } else
+    return VPX_CODEC_INVALID_PARAM;
 }
 
 static vpx_codec_err_t vp8_get_reference(vpx_codec_alg_priv_t *ctx,
-                                         va_list args)
-{
+                                         va_list args) {
+  vpx_ref_frame_t *data = va_arg(args, vpx_ref_frame_t *);
 
-    vpx_ref_frame_t *data = va_arg(args, vpx_ref_frame_t *);
+  if (data && !ctx->yv12_frame_buffers.use_frame_threads) {
+    vpx_ref_frame_t *frame = (vpx_ref_frame_t *)data;
+    YV12_BUFFER_CONFIG sd;
 
-    if (data && !ctx->yv12_frame_buffers.use_frame_threads)
-    {
-        vpx_ref_frame_t *frame = (vpx_ref_frame_t *)data;
-        YV12_BUFFER_CONFIG sd;
+    image2yuvconfig(&frame->img, &sd);
 
-        image2yuvconfig(&frame->img, &sd);
-
-        return vp8dx_get_reference(ctx->yv12_frame_buffers.pbi[0],
-                                   frame->frame_type, &sd);
-    }
-    else
-        return VPX_CODEC_INVALID_PARAM;
-
+    return vp8dx_get_reference(ctx->yv12_frame_buffers.pbi[0],
+                               frame->frame_type, &sd);
+  } else
+    return VPX_CODEC_INVALID_PARAM;
 }
 
 static vpx_codec_err_t vp8_set_postproc(vpx_codec_alg_priv_t *ctx,
-                                        va_list args)
-{
+                                        va_list args) {
 #if CONFIG_POSTPROC
-    vp8_postproc_cfg_t *data = va_arg(args, vp8_postproc_cfg_t *);
+  vp8_postproc_cfg_t *data = va_arg(args, vp8_postproc_cfg_t *);
 
-    if (data)
-    {
-        ctx->postproc_cfg_set = 1;
-        ctx->postproc_cfg = *((vp8_postproc_cfg_t *)data);
-        return VPX_CODEC_OK;
-    }
-    else
-        return VPX_CODEC_INVALID_PARAM;
+  if (data) {
+    ctx->postproc_cfg_set = 1;
+    ctx->postproc_cfg = *((vp8_postproc_cfg_t *)data);
+    return VPX_CODEC_OK;
+  } else
+    return VPX_CODEC_INVALID_PARAM;
 
 #else
-    (void)ctx;
-    (void)args;
-    return VPX_CODEC_INCAPABLE;
+  (void)ctx;
+  (void)args;
+  return VPX_CODEC_INCAPABLE;
 #endif
 }
 
-
 static vpx_codec_err_t vp8_set_dbg_color_ref_frame(vpx_codec_alg_priv_t *ctx,
                                                    va_list args) {
 #if CONFIG_POSTPROC_VISUALIZER && CONFIG_POSTPROC
@@ -702,127 +628,109 @@
 }
 
 static vpx_codec_err_t vp8_get_last_ref_updates(vpx_codec_alg_priv_t *ctx,
-                                                va_list args)
-{
-    int *update_info = va_arg(args, int *);
+                                                va_list args) {
+  int *update_info = va_arg(args, int *);
 
-    if (update_info && !ctx->yv12_frame_buffers.use_frame_threads)
-    {
-        VP8D_COMP *pbi = (VP8D_COMP *)ctx->yv12_frame_buffers.pbi[0];
+  if (update_info && !ctx->yv12_frame_buffers.use_frame_threads) {
+    VP8D_COMP *pbi = (VP8D_COMP *)ctx->yv12_frame_buffers.pbi[0];
 
-        *update_info = pbi->common.refresh_alt_ref_frame * (int) VP8_ALTR_FRAME
-            + pbi->common.refresh_golden_frame * (int) VP8_GOLD_FRAME
-            + pbi->common.refresh_last_frame * (int) VP8_LAST_FRAME;
+    *update_info = pbi->common.refresh_alt_ref_frame * (int)VP8_ALTR_FRAME +
+                   pbi->common.refresh_golden_frame * (int)VP8_GOLD_FRAME +
+                   pbi->common.refresh_last_frame * (int)VP8_LAST_FRAME;
 
-        return VPX_CODEC_OK;
-    }
-    else
-        return VPX_CODEC_INVALID_PARAM;
+    return VPX_CODEC_OK;
+  } else
+    return VPX_CODEC_INVALID_PARAM;
 }
 
-extern int vp8dx_references_buffer( VP8_COMMON *oci, int ref_frame );
+extern int vp8dx_references_buffer(VP8_COMMON *oci, int ref_frame);
 static vpx_codec_err_t vp8_get_last_ref_frame(vpx_codec_alg_priv_t *ctx,
-                                              va_list args)
-{
-    int *ref_info = va_arg(args, int *);
+                                              va_list args) {
+  int *ref_info = va_arg(args, int *);
 
-    if (ref_info && !ctx->yv12_frame_buffers.use_frame_threads)
-    {
-        VP8D_COMP *pbi = (VP8D_COMP *)ctx->yv12_frame_buffers.pbi[0];
-        VP8_COMMON *oci = &pbi->common;
-        *ref_info =
-            (vp8dx_references_buffer( oci, ALTREF_FRAME )?VP8_ALTR_FRAME:0) |
-            (vp8dx_references_buffer( oci, GOLDEN_FRAME )?VP8_GOLD_FRAME:0) |
-            (vp8dx_references_buffer( oci, LAST_FRAME )?VP8_LAST_FRAME:0);
+  if (ref_info && !ctx->yv12_frame_buffers.use_frame_threads) {
+    VP8D_COMP *pbi = (VP8D_COMP *)ctx->yv12_frame_buffers.pbi[0];
+    VP8_COMMON *oci = &pbi->common;
+    *ref_info =
+        (vp8dx_references_buffer(oci, ALTREF_FRAME) ? VP8_ALTR_FRAME : 0) |
+        (vp8dx_references_buffer(oci, GOLDEN_FRAME) ? VP8_GOLD_FRAME : 0) |
+        (vp8dx_references_buffer(oci, LAST_FRAME) ? VP8_LAST_FRAME : 0);
 
-        return VPX_CODEC_OK;
-    }
-    else
-        return VPX_CODEC_INVALID_PARAM;
+    return VPX_CODEC_OK;
+  } else
+    return VPX_CODEC_INVALID_PARAM;
 }
 
 static vpx_codec_err_t vp8_get_frame_corrupted(vpx_codec_alg_priv_t *ctx,
-                                               va_list args)
-{
+                                               va_list args) {
+  int *corrupted = va_arg(args, int *);
+  VP8D_COMP *pbi = (VP8D_COMP *)ctx->yv12_frame_buffers.pbi[0];
 
-    int *corrupted = va_arg(args, int *);
-    VP8D_COMP *pbi = (VP8D_COMP *)ctx->yv12_frame_buffers.pbi[0];
-
-    if (corrupted && pbi)
-    {
-        const YV12_BUFFER_CONFIG *const frame = pbi->common.frame_to_show;
-        if (frame == NULL) return VPX_CODEC_ERROR;
-        *corrupted = frame->corrupted;
-        return VPX_CODEC_OK;
-    }
-    else
-        return VPX_CODEC_INVALID_PARAM;
-
+  if (corrupted && pbi) {
+    const YV12_BUFFER_CONFIG *const frame = pbi->common.frame_to_show;
+    if (frame == NULL) return VPX_CODEC_ERROR;
+    *corrupted = frame->corrupted;
+    return VPX_CODEC_OK;
+  } else
+    return VPX_CODEC_INVALID_PARAM;
 }
 
 static vpx_codec_err_t vp8_set_decryptor(vpx_codec_alg_priv_t *ctx,
-                                         va_list args)
-{
-    vpx_decrypt_init *init = va_arg(args, vpx_decrypt_init *);
+                                         va_list args) {
+  vpx_decrypt_init *init = va_arg(args, vpx_decrypt_init *);
 
-    if (init)
-    {
-        ctx->decrypt_cb = init->decrypt_cb;
-        ctx->decrypt_state = init->decrypt_state;
-    }
-    else
-    {
-        ctx->decrypt_cb = NULL;
-        ctx->decrypt_state = NULL;
-    }
-    return VPX_CODEC_OK;
+  if (init) {
+    ctx->decrypt_cb = init->decrypt_cb;
+    ctx->decrypt_state = init->decrypt_state;
+  } else {
+    ctx->decrypt_cb = NULL;
+    ctx->decrypt_state = NULL;
+  }
+  return VPX_CODEC_OK;
 }
 
-vpx_codec_ctrl_fn_map_t vp8_ctf_maps[] =
-{
-    {VP8_SET_REFERENCE,             vp8_set_reference},
-    {VP8_COPY_REFERENCE,            vp8_get_reference},
-    {VP8_SET_POSTPROC,              vp8_set_postproc},
-    {VP8_SET_DBG_COLOR_REF_FRAME,   vp8_set_dbg_color_ref_frame},
-    {VP8_SET_DBG_COLOR_MB_MODES,    vp8_set_dbg_color_mb_modes},
-    {VP8_SET_DBG_COLOR_B_MODES,     vp8_set_dbg_color_b_modes},
-    {VP8_SET_DBG_DISPLAY_MV,        vp8_set_dbg_display_mv},
-    {VP8D_GET_LAST_REF_UPDATES,     vp8_get_last_ref_updates},
-    {VP8D_GET_FRAME_CORRUPTED,      vp8_get_frame_corrupted},
-    {VP8D_GET_LAST_REF_USED,        vp8_get_last_ref_frame},
-    {VPXD_SET_DECRYPTOR,            vp8_set_decryptor},
-    { -1, NULL},
+vpx_codec_ctrl_fn_map_t vp8_ctf_maps[] = {
+  { VP8_SET_REFERENCE, vp8_set_reference },
+  { VP8_COPY_REFERENCE, vp8_get_reference },
+  { VP8_SET_POSTPROC, vp8_set_postproc },
+  { VP8_SET_DBG_COLOR_REF_FRAME, vp8_set_dbg_color_ref_frame },
+  { VP8_SET_DBG_COLOR_MB_MODES, vp8_set_dbg_color_mb_modes },
+  { VP8_SET_DBG_COLOR_B_MODES, vp8_set_dbg_color_b_modes },
+  { VP8_SET_DBG_DISPLAY_MV, vp8_set_dbg_display_mv },
+  { VP8D_GET_LAST_REF_UPDATES, vp8_get_last_ref_updates },
+  { VP8D_GET_FRAME_CORRUPTED, vp8_get_frame_corrupted },
+  { VP8D_GET_LAST_REF_USED, vp8_get_last_ref_frame },
+  { VPXD_SET_DECRYPTOR, vp8_set_decryptor },
+  { -1, NULL },
 };
 
-
 #ifndef VERSION_STRING
 #define VERSION_STRING
 #endif
-CODEC_INTERFACE(vpx_codec_vp8_dx) =
-{
-    "WebM Project VP8 Decoder" VERSION_STRING,
-    VPX_CODEC_INTERNAL_ABI_VERSION,
-    VPX_CODEC_CAP_DECODER | VP8_CAP_POSTPROC | VP8_CAP_ERROR_CONCEALMENT |
-    VPX_CODEC_CAP_INPUT_FRAGMENTS,
-    /* vpx_codec_caps_t          caps; */
-    vp8_init,         /* vpx_codec_init_fn_t       init; */
-    vp8_destroy,      /* vpx_codec_destroy_fn_t    destroy; */
-    vp8_ctf_maps,     /* vpx_codec_ctrl_fn_map_t  *ctrl_maps; */
-    {
-        vp8_peek_si,      /* vpx_codec_peek_si_fn_t    peek_si; */
-        vp8_get_si,       /* vpx_codec_get_si_fn_t     get_si; */
-        vp8_decode,       /* vpx_codec_decode_fn_t     decode; */
-        vp8_get_frame,    /* vpx_codec_frame_get_fn_t  frame_get; */
-        NULL,
-    },
-    { /* encoder functions */
-        0,
-        NULL,  /* vpx_codec_enc_cfg_map_t */
-        NULL,  /* vpx_codec_encode_fn_t */
-        NULL,  /* vpx_codec_get_cx_data_fn_t */
-        NULL,  /* vpx_codec_enc_config_set_fn_t */
-        NULL,  /* vpx_codec_get_global_headers_fn_t */
-        NULL,  /* vpx_codec_get_preview_frame_fn_t */
-        NULL   /* vpx_codec_enc_mr_get_mem_loc_fn_t */
-    }
+CODEC_INTERFACE(vpx_codec_vp8_dx) = {
+  "WebM Project VP8 Decoder" VERSION_STRING,
+  VPX_CODEC_INTERNAL_ABI_VERSION,
+  VPX_CODEC_CAP_DECODER | VP8_CAP_POSTPROC | VP8_CAP_ERROR_CONCEALMENT |
+      VPX_CODEC_CAP_INPUT_FRAGMENTS,
+  /* vpx_codec_caps_t          caps; */
+  vp8_init,     /* vpx_codec_init_fn_t       init; */
+  vp8_destroy,  /* vpx_codec_destroy_fn_t    destroy; */
+  vp8_ctf_maps, /* vpx_codec_ctrl_fn_map_t  *ctrl_maps; */
+  {
+      vp8_peek_si,   /* vpx_codec_peek_si_fn_t    peek_si; */
+      vp8_get_si,    /* vpx_codec_get_si_fn_t     get_si; */
+      vp8_decode,    /* vpx_codec_decode_fn_t     decode; */
+      vp8_get_frame, /* vpx_codec_frame_get_fn_t  frame_get; */
+      NULL,
+  },
+  {
+      /* encoder functions */
+      0, NULL, /* vpx_codec_enc_cfg_map_t */
+      NULL,    /* vpx_codec_encode_fn_t */
+      NULL,    /* vpx_codec_get_cx_data_fn_t */
+      NULL,    /* vpx_codec_enc_config_set_fn_t */
+      NULL,    /* vpx_codec_get_global_headers_fn_t */
+      NULL,    /* vpx_codec_get_preview_frame_fn_t */
+      NULL     /* vpx_codec_enc_mr_get_mem_loc_fn_t */
+  }
 };
-- 
⑨