ref: 2c08c28191d65bb5a54c78567c2f54960b203f6d
parent: 7b8dfcb5a2cfb01ee7a6009d945d06559b564d06
parent: 0cedaa363187e7fbd42f513e6d1547dc385f38ed
author: John Koleszar <jkoleszar@google.com>
date: Wed Nov 7 05:40:37 EST 2012
Merge with upstream experimental changes Include upstream changes (unit test fixes, in particular) into the merged code base. Change-Id: I096f8a9d09e2532fbec0c95d7a995ab22fa54b29
--- a/configure
+++ b/configure
@@ -242,7 +242,6 @@
superblocks
pred_filter
lossless
- newbestrefmv
subpelrefmv
new_mvref
implicit_segmentation
--- a/test/boolcoder_test.cc
+++ /dev/null
@@ -1,90 +1,0 @@
-/*
- * Copyright (c) 2012 The WebM project authors. All Rights Reserved.
- *
- * Use of this source code is governed by a BSD-style license
- * that can be found in the LICENSE file in the root of the source
- * tree. An additional intellectual property rights grant can be found
- * in the file PATENTS. All contributing project authors may
- * be found in the AUTHORS file in the root of the source tree.
- */
-
-extern "C" {-#include "vp8/encoder/boolhuff.h"
-#include "vp8/decoder/dboolhuff.h"
-}
-
-#include <math.h>
-#include <stddef.h>
-#include <stdio.h>
-#include <stdlib.h>
-#include <string.h>
-#include <sys/types.h>
-
-#include "test/acm_random.h"
-#include "third_party/googletest/src/include/gtest/gtest.h"
-#include "vpx/vpx_integer.h"
-
-namespace {-const int num_tests = 10;
-} // namespace
-
-using libvpx_test::ACMRandom;
-
-TEST(VP8, TestBitIO) {- ACMRandom rnd(ACMRandom::DeterministicSeed());
- for (int n = 0; n < num_tests; ++n) {- for (int method = 0; method <= 7; ++method) { // we generate various proba- const int bits_to_test = 1000;
- uint8_t probas[bits_to_test];
-
- for (int i = 0; i < bits_to_test; ++i) {- const int parity = i & 1;
- probas[i] =
- (method == 0) ? 0 : (method == 1) ? 255 :
- (method == 2) ? 128 :
- (method == 3) ? rnd.Rand8() :
- (method == 4) ? (parity ? 0 : 255) :
- // alternate between low and high proba:
- (method == 5) ? (parity ? rnd(128) : 255 - rnd(128)) :
- (method == 6) ?
- (parity ? rnd(64) : 255 - rnd(64)) :
- (parity ? rnd(32) : 255 - rnd(32));
- }
- for (int bit_method = 0; bit_method <= 3; ++bit_method) {- const int random_seed = 6432;
- const int buffer_size = 10000;
- ACMRandom bit_rnd(random_seed);
- BOOL_CODER bw;
- uint8_t bw_buffer[buffer_size];
- vp8_start_encode(&bw, bw_buffer, bw_buffer + buffer_size);
-
- int bit = (bit_method == 0) ? 0 : (bit_method == 1) ? 1 : 0;
- for (int i = 0; i < bits_to_test; ++i) {- if (bit_method == 2) {- bit = (i & 1);
- } else if (bit_method == 3) {- bit = bit_rnd(2);
- }
- vp8_encode_bool(&bw, bit, static_cast<int>(probas[i]));
- }
-
- vp8_stop_encode(&bw);
-
- BOOL_DECODER br;
- vp8dx_start_decode(&br, bw_buffer, buffer_size);
- bit_rnd.Reset(random_seed);
- for (int i = 0; i < bits_to_test; ++i) {- if (bit_method == 2) {- bit = (i & 1);
- } else if (bit_method == 3) {- bit = bit_rnd(2);
- }
- GTEST_ASSERT_EQ(vp8dx_decode_bool(&br, probas[i]), bit)
- << "pos: "<< i << " / " << bits_to_test
- << " bit_method: " << bit_method
- << " method: " << method;
- }
- }
- }
- }
-}
--- a/test/dct16x16_test.cc
+++ b/test/dct16x16_test.cc
@@ -17,7 +17,7 @@
extern "C" {#include "vp9/common/entropy.h"
#include "vp9/common/idct.h"
-#include "vp9/encoder/dct.h"
+#include "vpx_rtcd.h"
}
#include "acm_random.h"
@@ -256,7 +256,7 @@
}
-TEST(VP8Idct16x16Test, AccuracyCheck) {+TEST(VP9Idct16x16Test, AccuracyCheck) {ACMRandom rnd(ACMRandom::DeterministicSeed());
const int count_test_block = 1000;
for (int i = 0; i < count_test_block; ++i) {@@ -271,7 +271,7 @@
reference_16x16_dct_2d(in, out_r);
for (int j = 0; j < 256; j++)
coeff[j] = round(out_r[j]);
- vp8_short_idct16x16_c(coeff, out_c, 32);
+ vp9_short_idct16x16_c(coeff, out_c, 32);
for (int j = 0; j < 256; ++j) {const int diff = out_c[j] - in[j];
const int error = diff * diff;
@@ -280,7 +280,7 @@
<< " at index " << j;
}
- vp8_short_fdct16x16_c(in, out_c, 32);
+ vp9_short_fdct16x16_c(in, out_c, 32);
for (int j = 0; j < 256; ++j) {const double diff = coeff[j] - out_c[j];
const double error = diff * diff;
@@ -291,7 +291,7 @@
}
}
-TEST(VP8Fdct16x16Test, AccuracyCheck) {+TEST(VP9Fdct16x16Test, AccuracyCheck) {ACMRandom rnd(ACMRandom::DeterministicSeed());
int max_error = 0;
double total_error = 0;
@@ -306,8 +306,8 @@
test_input_block[j] = rnd.Rand8() - rnd.Rand8();
const int pitch = 32;
- vp8_short_fdct16x16_c(test_input_block, test_temp_block, pitch);
- vp8_short_idct16x16_c(test_temp_block, test_output_block, pitch);
+ vp9_short_fdct16x16_c(test_input_block, test_temp_block, pitch);
+ vp9_short_idct16x16_c(test_temp_block, test_output_block, pitch);
for (int j = 0; j < 256; ++j) {const int diff = test_input_block[j] - test_output_block[j];
@@ -325,7 +325,7 @@
<< "Error: 16x16 FDCT/IDCT has average roundtrip error > 1/10 per block";
}
-TEST(VP8Fdct16x16Test, CoeffSizeCheck) {+TEST(VP9Fdct16x16Test, CoeffSizeCheck) {ACMRandom rnd(ACMRandom::DeterministicSeed());
const int count_test_block = 1000;
for (int i = 0; i < count_test_block; ++i) {@@ -342,8 +342,8 @@
input_extreme_block[j] = 255;
const int pitch = 32;
- vp8_short_fdct16x16_c(input_block, output_block, pitch);
- vp8_short_fdct16x16_c(input_extreme_block, output_extreme_block, pitch);
+ vp9_short_fdct16x16_c(input_block, output_block, pitch);
+ vp9_short_fdct16x16_c(input_extreme_block, output_extreme_block, pitch);
// The minimum quant value is 4.
for (int j = 0; j < 256; ++j) {--- a/test/fdct4x4_test.cc
+++ b/test/fdct4x4_test.cc
@@ -16,7 +16,7 @@
extern "C" {#include "vp9/common/idct.h"
-#include "vp9/encoder/dct.h"
+#include "vpx_rtcd.h"
}
#include "acm_random.h"
@@ -26,7 +26,7 @@
namespace {-TEST(Vp8FdctTest, SignBiasCheck) {+TEST(Vp9FdctTest, SignBiasCheck) {ACMRandom rnd(ACMRandom::DeterministicSeed());
int16_t test_input_block[16];
int16_t test_output_block[16];
@@ -43,7 +43,7 @@
// TODO(Yaowu): this should be converted to a parameterized test
// to test optimized versions of this function.
- vp8_short_fdct4x4_c(test_input_block, test_output_block, pitch);
+ vp9_short_fdct4x4_c(test_input_block, test_output_block, pitch);
for (int j = 0; j < 16; ++j) {if (test_output_block[j] < 0)
@@ -70,7 +70,7 @@
// TODO(Yaowu): this should be converted to a parameterized test
// to test optimized versions of this function.
- vp8_short_fdct4x4_c(test_input_block, test_output_block, pitch);
+ vp9_short_fdct4x4_c(test_input_block, test_output_block, pitch);
for (int j = 0; j < 16; ++j) {if (test_output_block[j] < 0)
@@ -89,7 +89,7 @@
}
};
-TEST(Vp8FdctTest, RoundTripErrorCheck) {+TEST(Vp9FdctTest, RoundTripErrorCheck) {ACMRandom rnd(ACMRandom::DeterministicSeed());
int max_error = 0;
double total_error = 0;
@@ -106,7 +106,7 @@
// TODO(Yaowu): this should be converted to a parameterized test
// to test optimized versions of this function.
const int pitch = 8;
- vp8_short_fdct4x4_c(test_input_block, test_temp_block, pitch);
+ vp9_short_fdct4x4_c(test_input_block, test_temp_block, pitch);
for (int j = 0; j < 16; ++j) { if(test_temp_block[j] > 0) {@@ -121,7 +121,7 @@
}
// Because the bitstream is not frozen yet, use the idct in the codebase.
- vp8_short_idct4x4llm_c(test_temp_block, test_output_block, pitch);
+ vp9_short_idct4x4llm_c(test_temp_block, test_output_block, pitch);
for (int j = 0; j < 16; ++j) {const int diff = test_input_block[j] - test_output_block[j];
--- a/test/fdct8x8_test.cc
+++ b/test/fdct8x8_test.cc
@@ -15,8 +15,8 @@
#include "third_party/googletest/src/include/gtest/gtest.h"
extern "C" {-#include "vp9/encoder/dct.h"
#include "vp9/common/idct.h"
+#include "vpx_rtcd.h"
}
#include "acm_random.h"
@@ -26,7 +26,7 @@
namespace {-TEST(VP8Fdct8x8Test, SignBiasCheck) {+TEST(VP9Fdct8x8Test, SignBiasCheck) {ACMRandom rnd(ACMRandom::DeterministicSeed());
int16_t test_input_block[64];
int16_t test_output_block[64];
@@ -41,7 +41,7 @@
for (int j = 0; j < 64; ++j)
test_input_block[j] = rnd.Rand8() - rnd.Rand8();
- vp8_short_fdct8x8_c(test_input_block, test_output_block, pitch);
+ vp9_short_fdct8x8_c(test_input_block, test_output_block, pitch);
for (int j = 0; j < 64; ++j) {if (test_output_block[j] < 0)
@@ -66,7 +66,7 @@
for (int j = 0; j < 64; ++j)
test_input_block[j] = (rnd.Rand8() >> 4) - (rnd.Rand8() >> 4);
- vp8_short_fdct8x8_c(test_input_block, test_output_block, pitch);
+ vp9_short_fdct8x8_c(test_input_block, test_output_block, pitch);
for (int j = 0; j < 64; ++j) {if (test_output_block[j] < 0)
@@ -85,7 +85,7 @@
}
};
-TEST(VP8Fdct8x8Test, RoundTripErrorCheck) {+TEST(VP9Fdct8x8Test, RoundTripErrorCheck) {ACMRandom rnd(ACMRandom::DeterministicSeed());
int max_error = 0;
double total_error = 0;
@@ -100,7 +100,7 @@
test_input_block[j] = rnd.Rand8() - rnd.Rand8();
const int pitch = 16;
- vp8_short_fdct8x8_c(test_input_block, test_temp_block, pitch);
+ vp9_short_fdct8x8_c(test_input_block, test_temp_block, pitch);
for (int j = 0; j < 64; ++j){ if(test_temp_block[j] > 0) {test_temp_block[j] += 2;
@@ -112,7 +112,7 @@
test_temp_block[j] *= 4;
}
}
- vp8_short_idct8x8_c(test_temp_block, test_output_block, pitch);
+ vp9_short_idct8x8_c(test_temp_block, test_output_block, pitch);
for (int j = 0; j < 64; ++j) {const int diff = test_input_block[j] - test_output_block[j];
@@ -130,7 +130,7 @@
<< "Error: 8x8 FDCT/IDCT has average roundtrip error > 1/5 per block";
};
-TEST(VP8Fdct8x8Test, ExtremalCheck) {+TEST(VP9Fdct8x8Test, ExtremalCheck) {ACMRandom rnd(ACMRandom::DeterministicSeed());
int max_error = 0;
double total_error = 0;
@@ -145,8 +145,8 @@
test_input_block[j] = rnd.Rand8() % 2 ? 255 : -255;
const int pitch = 16;
- vp8_short_fdct8x8_c(test_input_block, test_temp_block, pitch);
- vp8_short_idct8x8_c(test_temp_block, test_output_block, pitch);
+ vp9_short_fdct8x8_c(test_input_block, test_temp_block, pitch);
+ vp9_short_idct8x8_c(test_temp_block, test_output_block, pitch);
for (int j = 0; j < 64; ++j) {const int diff = test_input_block[j] - test_output_block[j];
--- a/test/idct8x8_test.cc
+++ b/test/idct8x8_test.cc
@@ -15,8 +15,8 @@
#include "third_party/googletest/src/include/gtest/gtest.h"
extern "C" {-#include "vp9/encoder/dct.h"
#include "vp9/common/idct.h"
+#include "vpx_rtcd.h"
}
#include "acm_random.h"
@@ -99,7 +99,7 @@
output[i] = round(out2[i]/32);
}
-TEST(VP8Idct8x8Test, AccuracyCheck) {+TEST(VP9Idct8x8Test, AccuracyCheck) {ACMRandom rnd(ACMRandom::DeterministicSeed());
const int count_test_block = 10000;
for (int i = 0; i < count_test_block; ++i) {@@ -112,7 +112,7 @@
input[j] = rnd.Rand8() - rnd.Rand8();
const int pitch = 16;
- vp8_short_fdct8x8_c(input, output_c, pitch);
+ vp9_short_fdct8x8_c(input, output_c, pitch);
reference_dct_2d(input, output_r);
for (int j = 0; j < 64; ++j) {@@ -140,7 +140,7 @@
reference_dct_2d(input, output_r);
for (int j = 0; j < 64; ++j)
coeff[j] = round(output_r[j]);
- vp8_short_idct8x8_c(coeff, output_c, pitch);
+ vp9_short_idct8x8_c(coeff, output_c, pitch);
for (int j = 0; j < 64; ++j) {const int diff = output_c[j] -input[j];
const int error = diff * diff;
--- a/test/test.mk
+++ b/test/test.mk
@@ -33,9 +33,12 @@
##
ifeq ($(CONFIG_SHARED),)
+## VP8
+ifneq ($(CONFIG_VP8_ENCODER)$(CONFIG_VP8_DECODER),)
+
# These tests require both the encoder and decoder to be built.
ifeq ($(CONFIG_VP8_ENCODER)$(CONFIG_VP8_DECODER),yesyes)
-LIBVPX_TEST_SRCS-yes += boolcoder_test.cc
+LIBVPX_TEST_SRCS-yes += vp8_boolcoder_test.cc
endif
LIBVPX_TEST_SRCS-yes += idctllm_test.cc
@@ -47,13 +50,22 @@
LIBVPX_TEST_SRCS-$(CONFIG_VP8_ENCODER) += subtract_test.cc
LIBVPX_TEST_SRCS-$(CONFIG_VP8_ENCODER) += vp8_fdct4x4_test.cc
-# VP9 tests
+endif # VP8
+
+## VP9
+ifneq ($(CONFIG_VP9_ENCODER)$(CONFIG_VP9_DECODER),)
+
+# These tests require both the encoder and decoder to be built.
+ifeq ($(CONFIG_VP9_ENCODER)$(CONFIG_VP9_DECODER),yesyes)
+LIBVPX_TEST_SRCS-yes += vp9_boolcoder_test.cc
+endif
+
LIBVPX_TEST_SRCS-$(CONFIG_VP9_ENCODER) += fdct4x4_test.cc
LIBVPX_TEST_SRCS-$(CONFIG_VP9_ENCODER) += fdct8x8_test.cc
-LIBVPX_TEST_SRCS-$(CONFIG_VP9_ENCODER) += dct16x16_test.cc
-ifneq ($(CONFIG_VP9_ENCODER)$(CONFIG_VP9_DECODER),)
+#LIBVPX_TEST_SRCS-$(CONFIG_VP9_ENCODER) += dct16x16_test.cc
LIBVPX_TEST_SRCS-yes += idct8x8_test.cc
-endif
+endif # VP9
+
endif
--- /dev/null
+++ b/test/vp8_boolcoder_test.cc
@@ -1,0 +1,90 @@
+/*
+ * Copyright (c) 2012 The WebM project authors. All Rights Reserved.
+ *
+ * Use of this source code is governed by a BSD-style license
+ * that can be found in the LICENSE file in the root of the source
+ * tree. An additional intellectual property rights grant can be found
+ * in the file PATENTS. All contributing project authors may
+ * be found in the AUTHORS file in the root of the source tree.
+ */
+
+extern "C" {+#include "vp8/encoder/boolhuff.h"
+#include "vp8/decoder/dboolhuff.h"
+}
+
+#include <math.h>
+#include <stddef.h>
+#include <stdio.h>
+#include <stdlib.h>
+#include <string.h>
+#include <sys/types.h>
+
+#include "test/acm_random.h"
+#include "third_party/googletest/src/include/gtest/gtest.h"
+#include "vpx/vpx_integer.h"
+
+namespace {+const int num_tests = 10;
+} // namespace
+
+using libvpx_test::ACMRandom;
+
+TEST(VP8, TestBitIO) {+ ACMRandom rnd(ACMRandom::DeterministicSeed());
+ for (int n = 0; n < num_tests; ++n) {+ for (int method = 0; method <= 7; ++method) { // we generate various proba+ const int bits_to_test = 1000;
+ uint8_t probas[bits_to_test];
+
+ for (int i = 0; i < bits_to_test; ++i) {+ const int parity = i & 1;
+ probas[i] =
+ (method == 0) ? 0 : (method == 1) ? 255 :
+ (method == 2) ? 128 :
+ (method == 3) ? rnd.Rand8() :
+ (method == 4) ? (parity ? 0 : 255) :
+ // alternate between low and high proba:
+ (method == 5) ? (parity ? rnd(128) : 255 - rnd(128)) :
+ (method == 6) ?
+ (parity ? rnd(64) : 255 - rnd(64)) :
+ (parity ? rnd(32) : 255 - rnd(32));
+ }
+ for (int bit_method = 0; bit_method <= 3; ++bit_method) {+ const int random_seed = 6432;
+ const int buffer_size = 10000;
+ ACMRandom bit_rnd(random_seed);
+ BOOL_CODER bw;
+ uint8_t bw_buffer[buffer_size];
+ vp8_start_encode(&bw, bw_buffer, bw_buffer + buffer_size);
+
+ int bit = (bit_method == 0) ? 0 : (bit_method == 1) ? 1 : 0;
+ for (int i = 0; i < bits_to_test; ++i) {+ if (bit_method == 2) {+ bit = (i & 1);
+ } else if (bit_method == 3) {+ bit = bit_rnd(2);
+ }
+ vp8_encode_bool(&bw, bit, static_cast<int>(probas[i]));
+ }
+
+ vp8_stop_encode(&bw);
+
+ BOOL_DECODER br;
+ vp8dx_start_decode(&br, bw_buffer, buffer_size);
+ bit_rnd.Reset(random_seed);
+ for (int i = 0; i < bits_to_test; ++i) {+ if (bit_method == 2) {+ bit = (i & 1);
+ } else if (bit_method == 3) {+ bit = bit_rnd(2);
+ }
+ GTEST_ASSERT_EQ(vp8dx_decode_bool(&br, probas[i]), bit)
+ << "pos: "<< i << " / " << bits_to_test
+ << " bit_method: " << bit_method
+ << " method: " << method;
+ }
+ }
+ }
+ }
+}
--- /dev/null
+++ b/test/vp9_boolcoder_test.cc
@@ -1,0 +1,88 @@
+/*
+ * Copyright (c) 2012 The WebM project authors. All Rights Reserved.
+ *
+ * Use of this source code is governed by a BSD-style license
+ * that can be found in the LICENSE file in the root of the source
+ * tree. An additional intellectual property rights grant can be found
+ * in the file PATENTS. All contributing project authors may
+ * be found in the AUTHORS file in the root of the source tree.
+ */
+
+#include <math.h>
+#include <stdlib.h>
+#include <string.h>
+
+#include "third_party/googletest/src/include/gtest/gtest.h"
+
+extern "C" {+#include "vp9/encoder/boolhuff.h"
+#include "vp9/decoder/dboolhuff.h"
+}
+
+#include "acm_random.h"
+#include "vpx/vpx_integer.h"
+
+using libvpx_test::ACMRandom;
+
+namespace {+const int num_tests = 10;
+} // namespace
+
+TEST(VP9, TestBitIO) {+ ACMRandom rnd(ACMRandom::DeterministicSeed());
+ for (int n = 0; n < num_tests; ++n) {+ for (int method = 0; method <= 7; ++method) { // we generate various proba+ const int bits_to_test = 1000;
+ uint8_t probas[bits_to_test];
+
+ for (int i = 0; i < bits_to_test; ++i) {+ const int parity = i & 1;
+ probas[i] =
+ (method == 0) ? 0 : (method == 1) ? 255 :
+ (method == 2) ? 128 :
+ (method == 3) ? rnd.Rand8() :
+ (method == 4) ? (parity ? 0 : 255) :
+ // alternate between low and high proba:
+ (method == 5) ? (parity ? rnd(128) : 255 - rnd(128)) :
+ (method == 6) ?
+ (parity ? rnd(64) : 255 - rnd(64)) :
+ (parity ? rnd(32) : 255 - rnd(32));
+ }
+ for (int bit_method = 0; bit_method <= 3; ++bit_method) {+ const int random_seed = 6432;
+ const int buffer_size = 10000;
+ ACMRandom bit_rnd(random_seed);
+ BOOL_CODER bw;
+ uint8_t bw_buffer[buffer_size];
+ vp9_start_encode(&bw, bw_buffer);
+
+ int bit = (bit_method == 0) ? 0 : (bit_method == 1) ? 1 : 0;
+ for (int i = 0; i < bits_to_test; ++i) {+ if (bit_method == 2) {+ bit = (i & 1);
+ } else if (bit_method == 3) {+ bit = bit_rnd(2);
+ }
+ encode_bool(&bw, bit, static_cast<int>(probas[i]));
+ }
+
+ vp9_stop_encode(&bw);
+
+ BOOL_DECODER br;
+ vp9_start_decode(&br, bw_buffer, buffer_size);
+ bit_rnd.Reset(random_seed);
+ for (int i = 0; i < bits_to_test; ++i) {+ if (bit_method == 2) {+ bit = (i & 1);
+ } else if (bit_method == 3) {+ bit = bit_rnd(2);
+ }
+ GTEST_ASSERT_EQ(decode_bool(&br, probas[i]), bit)
+ << "pos: " << i << " / " << bits_to_test
+ << " bit_method: " << bit_method
+ << " method: " << method;
+ }
+ }
+ }
+ }
+}
--- a/vp9/common/blockd.h
+++ b/vp9/common/blockd.h
@@ -44,9 +44,7 @@
/* Segment Feature Masks */
#define SEGMENT_DELTADATA 0
#define SEGMENT_ABSDATA 1
-#if CONFIG_NEWBESTREFMV || CONFIG_NEW_MVREF
#define MAX_MV_REFS 19
-#endif
typedef struct {int r, c;
@@ -216,9 +214,7 @@
MV_REFERENCE_FRAME ref_frame, second_ref_frame;
TX_SIZE txfm_size;
int_mv mv[2]; // for each reference frame used
-#if CONFIG_NEWBESTREFMV || CONFIG_NEW_MVREF
int_mv ref_mvs[MAX_REF_FRAMES][MAX_MV_REFS];
-#endif
SPLITMV_PARTITIONING_TYPE partitioning;
unsigned char mb_skip_coeff; /* does this mb has coefficients at all, 1=no coefficients, 0=need decode tokens */
@@ -280,7 +276,7 @@
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, unsigned short, eobs[25]);
/* 16 Y blocks, 4 U, 4 V, 1 DC 2nd order block, each with 16 entries. */
BLOCKD block[25];
@@ -467,7 +463,10 @@
TX_TYPE tx_type = DCT_DCT;
if (xd->mode_info_context->mbmi.mode == I8X8_PRED &&
xd->q_index < ACTIVE_HT8) {- tx_type = txfm_map(pred_mode_conv(b->bmi.as_mode.first));
+ // TODO(rbultje): MB_PREDICTION_MODE / B_PREDICTION_MODE should be merged
+ // or the relationship otherwise modified to address this type conversion.
+ tx_type = txfm_map(pred_mode_conv(
+ (MB_PREDICTION_MODE)b->bmi.as_mode.first));
}
return tx_type;
}
@@ -483,7 +482,7 @@
static TX_TYPE get_tx_type(const MACROBLOCKD *xd, const BLOCKD *b) {TX_TYPE tx_type = DCT_DCT;
- int ib = (b - xd->block);
+ int ib = (int)(b - xd->block);
if (ib >= 16)
return tx_type;
if (xd->mode_info_context->mbmi.txfm_size == TX_16X16) {--- a/vp9/common/findnearmv.c
+++ b/vp9/common/findnearmv.c
@@ -11,6 +11,7 @@
#include "findnearmv.h"
#include "vp9/common/sadmxn.h"
+#include "vp9/common/subpelvar.h"
#include <limits.h>
const unsigned char vp9_mbsplit_offset[4][16] = {@@ -167,7 +168,6 @@
return p;
}
-#if CONFIG_NEWBESTREFMV
#define SP(x) (((x) & 7) << 1)
unsigned int vp9_sad3x16_c(
const unsigned char *src_ptr,
@@ -186,6 +186,76 @@
return sad_mx_n_c(src_ptr, src_stride, ref_ptr, ref_stride, 16, 3);
}
+#if CONFIG_SUBPELREFMV
+unsigned int vp9_variance2x16_c(const unsigned char *src_ptr,
+ const int source_stride,
+ const unsigned char *ref_ptr,
+ const int recon_stride,
+ unsigned int *sse) {+ unsigned int var;
+ int avg;
+
+ variance(src_ptr, source_stride, ref_ptr, recon_stride, 2, 16, &var, &avg);
+ *sse = var;
+ return (var - ((avg * avg) >> 5));
+}
+
+unsigned int vp9_variance16x2_c(const unsigned char *src_ptr,
+ const int source_stride,
+ const unsigned char *ref_ptr,
+ const int recon_stride,
+ unsigned int *sse) {+ unsigned int var;
+ int avg;
+
+ variance(src_ptr, source_stride, ref_ptr, recon_stride, 16, 2, &var, &avg);
+ *sse = var;
+ return (var - ((avg * avg) >> 5));
+}
+
+unsigned int vp9_sub_pixel_variance16x2_c(const unsigned char *src_ptr,
+ const int src_pixels_per_line,
+ const int xoffset,
+ const int yoffset,
+ const unsigned char *dst_ptr,
+ const int dst_pixels_per_line,
+ unsigned int *sse) {+ unsigned short FData3[16 * 3]; // Temp data bufffer used in filtering
+ unsigned char temp2[20 * 16];
+ const short *HFilter, *VFilter;
+
+ HFilter = vp9_bilinear_filters[xoffset];
+ VFilter = vp9_bilinear_filters[yoffset];
+
+ var_filter_block2d_bil_first_pass(src_ptr, FData3,
+ src_pixels_per_line, 1, 3, 16, HFilter);
+ var_filter_block2d_bil_second_pass(FData3, temp2, 16, 16, 2, 16, VFilter);
+
+ return vp9_variance16x2_c(temp2, 16, dst_ptr, dst_pixels_per_line, sse);
+}
+
+unsigned int vp9_sub_pixel_variance2x16_c(const unsigned char *src_ptr,
+ const int src_pixels_per_line,
+ const int xoffset,
+ const int yoffset,
+ const unsigned char *dst_ptr,
+ const int dst_pixels_per_line,
+ unsigned int *sse) {+ unsigned short FData3[2 * 17]; // Temp data bufffer used in filtering
+ unsigned char temp2[2 * 16];
+ const short *HFilter, *VFilter;
+
+ HFilter = vp9_bilinear_filters[xoffset];
+ VFilter = vp9_bilinear_filters[yoffset];
+
+ var_filter_block2d_bil_first_pass(src_ptr, FData3,
+ src_pixels_per_line, 1, 17, 2, HFilter);
+ var_filter_block2d_bil_second_pass(FData3, temp2, 2, 2, 16, 2, VFilter);
+
+ return vp9_variance2x16_c(temp2, 2, dst_ptr, dst_pixels_per_line, sse);
+}
+#endif
+
/* check a list of motion vectors by sad score using a number rows of pixels
* above and a number cols of pixels in the left to select the one with best
* score to use as ref motion vector
@@ -323,5 +393,3 @@
// Copy back the re-ordered mv list
vpx_memcpy(mvlist, sorted_mvs, sizeof(sorted_mvs));
}
-
-#endif // CONFIG_NEWBESTREFMV
--- a/vp9/common/findnearmv.h
+++ b/vp9/common/findnearmv.h
@@ -18,7 +18,6 @@
#include "treecoder.h"
#include "onyxc_int.h"
-#if CONFIG_NEWBESTREFMV
/* check a list of motion vectors by sad score using a number rows of pixels
* above and a number cols of pixels in the left to select the one with best
* score to use as ref motion vector
@@ -30,7 +29,6 @@
int_mv *best_mv,
int_mv *nearest,
int_mv *near);
-#endif
static void mv_bias(int refmb_ref_frame_sign_bias, int refframe, int_mv *mvp, const int *ref_frame_sign_bias) {MV xmv;
--- a/vp9/common/idctllm.c
+++ b/vp9/common/idctllm.c
@@ -1013,6 +1013,8 @@
}
#endif
+#define TEST_INT_16x16_IDCT 1
+#if !TEST_INT_16x16_IDCT
static const double C1 = 0.995184726672197;
static const double C2 = 0.98078528040323;
static const double C3 = 0.956940335732209;
@@ -1273,3 +1275,235 @@
}
vp9_clear_system_state(); // Make it simd safe : __asm emms;
}
+
+#else
+static const int16_t C1 = 16305;
+static const int16_t C2 = 16069;
+static const int16_t C3 = 15679;
+static const int16_t C4 = 15137;
+static const int16_t C5 = 14449;
+static const int16_t C6 = 13623;
+static const int16_t C7 = 12665;
+static const int16_t C8 = 11585;
+static const int16_t C9 = 10394;
+static const int16_t C10 = 9102;
+static const int16_t C11 = 7723;
+static const int16_t C12 = 6270;
+static const int16_t C13 = 4756;
+static const int16_t C14 = 3196;
+static const int16_t C15 = 1606;
+
+#define INITIAL_SHIFT 2
+#define INITIAL_ROUNDING (1 << (INITIAL_SHIFT - 1))
+#define RIGHT_SHIFT 14
+#define RIGHT_ROUNDING (1 << (RIGHT_SHIFT - 1))
+
+static void butterfly_16x16_idct_1d(int16_t input[16], int16_t output[16],
+ int last_shift_bits) {+ int16_t step[16];
+ int intermediate[16];
+ int temp1, temp2;
+
+ int step1_shift = RIGHT_SHIFT + INITIAL_SHIFT;
+ int step1_rounding = 1 << (step1_shift - 1);
+ int last_rounding = 0;
+
+ if (last_shift_bits > 0)
+ last_rounding = 1 << (last_shift_bits - 1);
+
+ // step 1 and 2
+ step[ 0] = (input[0] + input[8] + INITIAL_ROUNDING) >> INITIAL_SHIFT;
+ step[ 1] = (input[0] - input[8] + INITIAL_ROUNDING) >> INITIAL_SHIFT;
+
+ temp1 = input[4] * C12;
+ temp2 = input[12] * C4;
+ temp1 = (temp1 - temp2 + RIGHT_ROUNDING) >> RIGHT_SHIFT;
+ temp1 *= C8;
+ step[ 2] = (2 * (temp1) + step1_rounding) >> step1_shift;
+
+ temp1 = input[4] * C4;
+ temp2 = input[12] * C12;
+ temp1 = (temp1 + temp2 + RIGHT_ROUNDING) >> RIGHT_SHIFT;
+ temp1 *= C8;
+ step[ 3] = (2 * (temp1) + step1_rounding) >> step1_shift;
+
+ temp1 = input[2] * C8;
+ temp1 = (2 * (temp1) + RIGHT_ROUNDING) >> RIGHT_SHIFT;
+ temp2 = input[6] + input[10];
+ step[ 4] = (temp1 + temp2 + INITIAL_ROUNDING) >> INITIAL_SHIFT;
+ step[ 5] = (temp1 - temp2 + INITIAL_ROUNDING) >> INITIAL_SHIFT;
+
+ temp1 = input[14] * C8;
+ temp1 = (2 * (temp1) + RIGHT_ROUNDING) >> RIGHT_SHIFT;
+ temp2 = input[6] - input[10];
+ step[ 6] = (temp2 - temp1 + INITIAL_ROUNDING) >> INITIAL_SHIFT;
+ step[ 7] = (temp2 + temp1 + INITIAL_ROUNDING) >> INITIAL_SHIFT;
+
+ // for odd input
+ temp1 = input[3] * C12;
+ temp2 = input[13] * C4;
+ temp1 = (temp1 + temp2 + RIGHT_ROUNDING) >> RIGHT_SHIFT;
+ temp1 *= C8;
+ intermediate[ 8] = (2 * (temp1) + RIGHT_ROUNDING) >> RIGHT_SHIFT;
+
+ temp1 = input[3] * C4;
+ temp2 = input[13] * C12;
+ temp2 = (temp2 - temp1 + RIGHT_ROUNDING) >> RIGHT_SHIFT;
+ temp2 *= C8;
+ intermediate[ 9] = (2 * (temp2) + RIGHT_ROUNDING) >> RIGHT_SHIFT;
+
+ intermediate[10] = (2 * (input[9] * C8) + RIGHT_ROUNDING) >> RIGHT_SHIFT;
+ intermediate[11] = input[15] - input[1];
+ intermediate[12] = input[15] + input[1];
+ intermediate[13] = (2 * (input[7] * C8) + RIGHT_ROUNDING) >> RIGHT_SHIFT;
+
+ temp1 = input[11] * C12;
+ temp2 = input[5] * C4;
+ temp2 = (temp2 - temp1 + RIGHT_ROUNDING) >> RIGHT_SHIFT;
+ temp2 *= C8;
+ intermediate[14] = (2 * (temp2) + RIGHT_ROUNDING) >> RIGHT_SHIFT;
+
+ temp1 = input[11] * C4;
+ temp2 = input[5] * C12;
+ temp1 = (temp1 + temp2 + RIGHT_ROUNDING) >> RIGHT_SHIFT;
+ temp1 *= C8;
+ intermediate[15] = (2 * (temp1) + RIGHT_ROUNDING) >> RIGHT_SHIFT;
+
+ step[ 8] = (intermediate[ 8] + intermediate[14] + INITIAL_ROUNDING)
+ >> INITIAL_SHIFT;
+ step[ 9] = (intermediate[ 9] + intermediate[15] + INITIAL_ROUNDING)
+ >> INITIAL_SHIFT;
+ step[10] = (intermediate[10] + intermediate[11] + INITIAL_ROUNDING)
+ >> INITIAL_SHIFT;
+ step[11] = (intermediate[10] - intermediate[11] + INITIAL_ROUNDING)
+ >> INITIAL_SHIFT;
+ step[12] = (intermediate[12] + intermediate[13] + INITIAL_ROUNDING)
+ >> INITIAL_SHIFT;
+ step[13] = (intermediate[12] - intermediate[13] + INITIAL_ROUNDING)
+ >> INITIAL_SHIFT;
+ step[14] = (intermediate[ 8] - intermediate[14] + INITIAL_ROUNDING)
+ >> INITIAL_SHIFT;
+ step[15] = (intermediate[ 9] - intermediate[15] + INITIAL_ROUNDING)
+ >> INITIAL_SHIFT;
+
+ // step 3
+ output[0] = step[ 0] + step[ 3];
+ output[1] = step[ 1] + step[ 2];
+ output[2] = step[ 1] - step[ 2];
+ output[3] = step[ 0] - step[ 3];
+
+ temp1 = step[ 4] * C14;
+ temp2 = step[ 7] * C2;
+ output[4] = (temp1 - temp2 + RIGHT_ROUNDING) >> RIGHT_SHIFT;
+
+ temp1 = step[ 4] * C2;
+ temp2 = step[ 7] * C14;
+ output[7] = (temp1 + temp2 + RIGHT_ROUNDING) >> RIGHT_SHIFT;
+
+ temp1 = step[ 5] * C10;
+ temp2 = step[ 6] * C6;
+ output[5] = (temp1 - temp2 + RIGHT_ROUNDING) >> RIGHT_SHIFT;
+
+ temp1 = step[ 5] * C6;
+ temp2 = step[ 6] * C10;
+ output[6] = (temp1 + temp2 + RIGHT_ROUNDING) >> RIGHT_SHIFT;
+
+ output[8] = step[ 8] + step[11];
+ output[9] = step[ 9] + step[10];
+ output[10] = step[ 9] - step[10];
+ output[11] = step[ 8] - step[11];
+ output[12] = step[12] + step[15];
+ output[13] = step[13] + step[14];
+ output[14] = step[13] - step[14];
+ output[15] = step[12] - step[15];
+
+ // output 4
+ step[ 0] = output[0] + output[7];
+ step[ 1] = output[1] + output[6];
+ step[ 2] = output[2] + output[5];
+ step[ 3] = output[3] + output[4];
+ step[ 4] = output[3] - output[4];
+ step[ 5] = output[2] - output[5];
+ step[ 6] = output[1] - output[6];
+ step[ 7] = output[0] - output[7];
+
+ temp1 = output[8] * C7;
+ temp2 = output[15] * C9;
+ step[ 8] = (temp1 - temp2 + RIGHT_ROUNDING) >> RIGHT_SHIFT;
+
+ temp1 = output[9] * C11;
+ temp2 = output[14] * C5;
+ step[ 9] = (temp1 + temp2 + RIGHT_ROUNDING) >> RIGHT_SHIFT;
+
+ temp1 = output[10] * C3;
+ temp2 = output[13] * C13;
+ step[10] = (temp1 - temp2 + RIGHT_ROUNDING) >> RIGHT_SHIFT;
+
+ temp1 = output[11] * C15;
+ temp2 = output[12] * C1;
+ step[11] = (temp1 + temp2 + RIGHT_ROUNDING) >> RIGHT_SHIFT;
+
+ temp1 = output[11] * C1;
+ temp2 = output[12] * C15;
+ step[12] = (temp2 - temp1 + RIGHT_ROUNDING) >> RIGHT_SHIFT;
+
+ temp1 = output[10] * C13;
+ temp2 = output[13] * C3;
+ step[13] = (temp1 + temp2 + RIGHT_ROUNDING) >> RIGHT_SHIFT;
+
+ temp1 = output[9] * C5;
+ temp2 = output[14] * C11;
+ step[14] = (temp2 - temp1 + RIGHT_ROUNDING) >> RIGHT_SHIFT;
+
+ temp1 = output[8] * C9;
+ temp2 = output[15] * C7;
+ step[15] = (temp1 + temp2 + RIGHT_ROUNDING) >> RIGHT_SHIFT;
+
+ // step 5
+ output[0] = (step[0] + step[15] + last_rounding) >> last_shift_bits;
+ output[1] = (step[1] + step[14] + last_rounding) >> last_shift_bits;
+ output[2] = (step[2] + step[13] + last_rounding) >> last_shift_bits;
+ output[3] = (step[3] + step[12] + last_rounding) >> last_shift_bits;
+ output[4] = (step[4] + step[11] + last_rounding) >> last_shift_bits;
+ output[5] = (step[5] + step[10] + last_rounding) >> last_shift_bits;
+ output[6] = (step[6] + step[ 9] + last_rounding) >> last_shift_bits;
+ output[7] = (step[7] + step[ 8] + last_rounding) >> last_shift_bits;
+
+ output[15] = (step[0] - step[15] + last_rounding) >> last_shift_bits;
+ output[14] = (step[1] - step[14] + last_rounding) >> last_shift_bits;
+ output[13] = (step[2] - step[13] + last_rounding) >> last_shift_bits;
+ output[12] = (step[3] - step[12] + last_rounding) >> last_shift_bits;
+ output[11] = (step[4] - step[11] + last_rounding) >> last_shift_bits;
+ output[10] = (step[5] - step[10] + last_rounding) >> last_shift_bits;
+ output[9] = (step[6] - step[ 9] + last_rounding) >> last_shift_bits;
+ output[8] = (step[7] - step[ 8] + last_rounding) >> last_shift_bits;
+}
+
+void vp9_short_idct16x16_c(int16_t *input, int16_t *output, int pitch) {+ int16_t out[16 * 16];
+ int16_t *outptr = &out[0];
+ const int short_pitch = pitch >> 1;
+ int i, j;
+ int16_t temp_in[16], temp_out[16];
+
+ // First transform rows
+ for (i = 0; i < 16; ++i) {+ butterfly_16x16_idct_1d(input, outptr, 0);
+ input += short_pitch;
+ outptr += 16;
+ }
+
+ // Then transform columns
+ for (i = 0; i < 16; ++i) {+ for (j = 0; j < 16; ++j)
+ temp_in[j] = out[j * 16 + i];
+ butterfly_16x16_idct_1d(temp_in, temp_out, 3);
+ for (j = 0; j < 16; ++j)
+ output[j * 16 + i] = temp_out[j];
+ }
+}
+#undef INITIAL_SHIFT
+#undef INITIAL_ROUNDING
+#undef RIGHT_SHIFT
+#undef RIGHT_ROUNDING
+#endif
--- a/vp9/common/mv.h
+++ b/vp9/common/mv.h
@@ -18,9 +18,9 @@
short col;
} MV;
-typedef union {- 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 */
#endif
--- a/vp9/common/mvref_common.c
+++ b/vp9/common/mvref_common.c
@@ -10,8 +10,6 @@
#include "mvref_common.h"
-#if CONFIG_NEWBESTREFMV
-
#define MVREF_NEIGHBOURS 8
static int mv_ref_search[MVREF_NEIGHBOURS][2] =
{ {0,-1},{-1,0},{-1,-1},{0,-2},{-2,0},{-1,-2},{-2,-1},{-2,-2} };@@ -338,5 +336,3 @@
// Copy over the candidate list.
vpx_memcpy(mv_ref_list, candidate_mvs, sizeof(candidate_mvs));
}
-
-#endif
--- a/vp9/common/mvref_common.h
+++ b/vp9/common/mvref_common.h
@@ -11,8 +11,6 @@
#include "onyxc_int.h"
#include "blockd.h"
-// MR reference entropy header file.
-#if CONFIG_NEWBESTREFMV
#ifndef __INC_MVREF_COMMON_H
#define __INC_MVREF_COMMON_H
@@ -28,4 +26,3 @@
#endif
-#endif
--- a/vp9/common/postproc.c
+++ b/vp9/common/postproc.c
@@ -12,6 +12,7 @@
#include "vpx_ports/config.h"
#include "vpx_scale/yv12config.h"
#include "postproc.h"
+#include "vp9/common/textblit.h"
#include "vpx_scale/vpxscale.h"
#include "systemdependent.h"
@@ -126,10 +127,6 @@
};
-extern void vp9_blit_text(const char *msg, unsigned char *address,
- const int pitch);
-extern void vp9_blit_line(int x0, int x1, int y0, int y1,
- unsigned char *image, const int pitch);
/****************************************************************************
*/
void vp9_post_proc_down_and_across_c(unsigned char *src_ptr,
--- a/vp9/common/quant_common.c
+++ b/vp9/common/quant_common.c
@@ -30,9 +30,9 @@
last_val = current_val;
ac_val = ac_qlookup[i];
- dc_qlookup[i] = (0.000000305 * ac_val * ac_val * ac_val) +
- (-0.00065 * ac_val * ac_val) +
- (0.9 * ac_val) + 0.5;
+ dc_qlookup[i] = (int)((0.000000305 * ac_val * ac_val * ac_val) +
+ (-0.00065 * ac_val * ac_val) +
+ (0.9 * ac_val) + 0.5);
if (dc_qlookup[i] < ACDC_MIN)
dc_qlookup[i] = ACDC_MIN;
}
--- a/vp9/common/reconintra.h
+++ b/vp9/common/reconintra.h
@@ -13,6 +13,6 @@
#include "blockd.h"
-extern void init_intra_left_above_pixels(MACROBLOCKD *xd);
+extern void vp9_recon_intra_mbuv(MACROBLOCKD *xd);
#endif // __INC_RECONINTRA_H
--- a/vp9/common/rtcd_defs.sh
+++ b/vp9/common/rtcd_defs.sh
@@ -11,6 +11,8 @@
struct macroblock;
struct variance_vtable;
+#define DEC_MVCOSTS int *mvjcost, int *mvcost[2]
+
/* Encoder forward decls */
struct variance_vtable;
union int_mv;
@@ -43,18 +45,24 @@
prototype void vp9_dequantize_b_2x2 "struct blockd *x"
specialize vp9_dequantize_b_2x2
-prototype void vp9_dequant_dc_idct_add_y_block_8x8 "short *q, short *dq, unsigned char *pre, unsigned char *dst, int stride, char *eobs, short *dc, struct macroblockd *xd"
+prototype void vp9_dequant_dc_idct_add_y_block_8x8 "short *q, short *dq, unsigned char *pre, unsigned char *dst, int stride, unsigned short *eobs, short *dc, struct macroblockd *xd"
specialize vp9_dequant_dc_idct_add_y_block_8x8
-prototype void vp9_dequant_idct_add_y_block_8x8 "short *q, short *dq, unsigned char *pre, unsigned char *dst, int stride, char *eobs, struct macroblockd *xd"
+prototype void vp9_dequant_idct_add_y_block_8x8 "short *q, short *dq, unsigned char *pre, unsigned char *dst, int stride, unsigned short *eobs, struct macroblockd *xd"
specialize vp9_dequant_idct_add_y_block_8x8
-prototype void vp9_dequant_idct_add_uv_block_8x8 "short *q, short *dq, unsigned char *pre, unsigned char *dstu, unsigned char *dstv, int stride, char *eobs, struct macroblockd *xd"
+prototype void vp9_dequant_idct_add_uv_block_8x8 "short *q, short *dq, unsigned char *pre, unsigned char *dstu, unsigned char *dstv, int stride, unsigned short *eobs, struct macroblockd *xd"
specialize vp9_dequant_idct_add_uv_block_8x8
prototype void vp9_dequant_idct_add_16x16 "short *input, short *dq, unsigned char *pred, unsigned char *dest, int pitch, int stride"
specialize vp9_dequant_idct_add_16x16
+prototype void vp9_dequant_idct_add_8x8 "short *input, short *dq, unsigned char *pred, unsigned char *dest, int pitch, int stride"
+specialize vp9_dequant_idct_add_8x8
+
+prototype void vp9_dequant_dc_idct_add_8x8 "short *input, short *dq, unsigned char *pred, unsigned char *dest, int pitch, int stride, int Dc"
+specialize vp9_dequant_dc_idct_add_8x8
+
prototype void vp9_dequant_idct_add "short *input, short *dq, unsigned char *pred, unsigned char *dest, int pitch, int stride"
specialize vp9_dequant_idct_add
@@ -61,13 +69,13 @@
prototype void vp9_dequant_dc_idct_add "short *input, short *dq, unsigned char *pred, unsigned char *dest, int pitch, int stride, int Dc"
specialize vp9_dequant_dc_idct_add
-prototype void vp9_dequant_dc_idct_add_y_block "short *q, short *dq, unsigned char *pre, unsigned char *dst, int stride, char *eobs, short *dc"
+prototype void vp9_dequant_dc_idct_add_y_block "short *q, short *dq, unsigned char *pre, unsigned char *dst, int stride, unsigned short *eobs, short *dc"
specialize vp9_dequant_dc_idct_add_y_block mmx
-prototype void vp9_dequant_idct_add_y_block "short *q, short *dq, unsigned char *pre, unsigned char *dst, int stride, char *eobs"
+prototype void vp9_dequant_idct_add_y_block "short *q, short *dq, unsigned char *pre, unsigned char *dst, int stride, unsigned short *eobs"
specialize vp9_dequant_idct_add_y_block mmx
-prototype void vp9_dequant_idct_add_uv_block "short *q, short *dq, unsigned char *pre, unsigned char *dstu, unsigned char *dstv, int stride, char *eobs"
+prototype void vp9_dequant_idct_add_uv_block "short *q, short *dq, unsigned char *pre, unsigned char *dstu, unsigned char *dstv, int stride, unsigned short *eobs"
specialize vp9_dequant_idct_add_uv_block mmx
#
@@ -219,13 +227,11 @@
#
# sad 16x3, 3x16
#
-if [ "$CONFIG_NEWBESTREFMV" = "yes" ]; then
prototype unsigned int vp9_sad16x3 "const unsigned char *src_ptr, int src_stride, const unsigned char *ref_ptr, int ref_stride, int max_sad"
specialize vp9_sad16x3 sse2
prototype unsigned int vp9_sad3x16 "const unsigned char *src_ptr, int src_stride, const unsigned char *ref_ptr, int ref_stride, int max_sad"
specialize vp9_sad3x16 sse2
-fi
#
# Encoder functions below this point.
@@ -368,22 +374,22 @@
prototype void vp9_sad4x4x8 "const unsigned char *src_ptr, int src_stride, const unsigned char *ref_ptr, int ref_stride, unsigned short *sad_array"
specialize vp9_sad4x4x8 sse4
-prototype void vp9_sad32x32x4d "const unsigned char *src_ptr, int src_stride, unsigned char *ref_ptr[], int ref_stride, unsigned int *sad_array"
+prototype void vp9_sad32x32x4d "const unsigned char *src_ptr, int src_stride, const unsigned char* const ref_ptr[], int ref_stride, unsigned int *sad_array"
specialize vp9_sad32x32x4d
-prototype void vp9_sad16x16x4d "const unsigned char *src_ptr, int src_stride, unsigned char *ref_ptr[], int ref_stride, unsigned int *sad_array"
+prototype void vp9_sad16x16x4d "const unsigned char *src_ptr, int src_stride, const unsigned char* const ref_ptr[], int ref_stride, unsigned int *sad_array"
specialize vp9_sad16x16x4d sse3
-prototype void vp9_sad16x8x4d "const unsigned char *src_ptr, int src_stride, unsigned char *ref_ptr[], int ref_stride, unsigned int *sad_array"
+prototype void vp9_sad16x8x4d "const unsigned char *src_ptr, int src_stride, const unsigned char* const ref_ptr[], int ref_stride, unsigned int *sad_array"
specialize vp9_sad16x8x4d sse3
-prototype void vp9_sad8x16x4d "const unsigned char *src_ptr, int src_stride, unsigned char *ref_ptr[], int ref_stride, unsigned int *sad_array"
+prototype void vp9_sad8x16x4d "const unsigned char *src_ptr, int src_stride, const unsigned char* const ref_ptr[], int ref_stride, unsigned int *sad_array"
specialize vp9_sad8x16x4d sse3
-prototype void vp9_sad8x8x4d "const unsigned char *src_ptr, int src_stride, unsigned char *ref_ptr[], int ref_stride, unsigned int *sad_array"
+prototype void vp9_sad8x8x4d "const unsigned char *src_ptr, int src_stride, const unsigned char* const ref_ptr[], int ref_stride, unsigned int *sad_array"
specialize vp9_sad8x8x4d sse3
-prototype void vp9_sad4x4x4d "const unsigned char *src_ptr, int src_stride, unsigned char *ref_ptr[], int ref_stride, unsigned int *sad_array"
+prototype void vp9_sad4x4x4d "const unsigned char *src_ptr, int src_stride, const unsigned char* const ref_ptr[], int ref_stride, unsigned int *sad_array"
specialize vp9_sad4x4x4d sse3
#
@@ -477,6 +483,21 @@
prototype void vp9_short_walsh8x4_x8 "short *InputData, short *OutputData, int pitch"
specialize vp9_short_walsh8x4_x8
+
+#
+# Motion search
+#
+prototype int vp9_full_search_sad "struct macroblock *x, struct block *b, struct blockd *d, union int_mv *ref_mv, int sad_per_bit, int distance, struct variance_vtable *fn_ptr, DEC_MVCOSTS, union int_mv *center_mv"
+specialize vp9_full_search_sad sse3 sse4_1
+vp9_full_search_sad_sse3=vp9_full_search_sadx3
+vp9_full_search_sad_sse4_1=vp9_full_search_sadx8
+
+prototype int vp9_refining_search_sad "struct macroblock *x, struct block *b, struct blockd *d, union int_mv *ref_mv, int sad_per_bit, int distance, struct variance_vtable *fn_ptr, DEC_MVCOSTS, union int_mv *center_mv"
+specialize vp9_refining_search_sad sse3
+vp9_refining_search_sad_sse3=vp9_refining_search_sadx4
+
+prototype int vp9_diamond_search_sad "struct macroblock *x, struct block *b, struct blockd *d, union int_mv *ref_mv, union int_mv *best_mv, int search_param, int sad_per_bit, int *num00, struct variance_vtable *fn_ptr, DEC_MVCOSTS, union int_mv *center_mv"
+vp9_diamond_search_sad_sse3=vp9_diamond_search_sadx4
fi
# end encoder functions
--- /dev/null
+++ b/vp9/common/subpelvar.h
@@ -1,0 +1,147 @@
+/*
+ * Copyright (c) 2010 The WebM project authors. All Rights Reserved.
+ *
+ * Use of this source code is governed by a BSD-style license
+ * that can be found in the LICENSE file in the root of the source
+ * tree. An additional intellectual property rights grant can be found
+ * in the file PATENTS. All contributing project authors may
+ * be found in the AUTHORS file in the root of the source tree.
+ */
+
+
+#include "vp9/common/filter.h"
+
+
+
+static void variance(const unsigned char *src_ptr,
+ int source_stride,
+ const unsigned char *ref_ptr,
+ int recon_stride,
+ int w,
+ int h,
+ unsigned int *sse,
+ int *sum) {+ int i, j;
+ int diff;
+
+ *sum = 0;
+ *sse = 0;
+
+ for (i = 0; i < h; i++) {+ for (j = 0; j < w; j++) {+ diff = src_ptr[j] - ref_ptr[j];
+ *sum += diff;
+ *sse += diff * diff;
+ }
+
+ src_ptr += source_stride;
+ ref_ptr += recon_stride;
+ }
+}
+
+/****************************************************************************
+ *
+ * ROUTINE : filter_block2d_bil_first_pass
+ *
+ * INPUTS : UINT8 *src_ptr : Pointer to source block.
+ * UINT32 src_pixels_per_line : Stride of input block.
+ * UINT32 pixel_step : Offset between filter input samples (see notes).
+ * UINT32 output_height : Input block height.
+ * UINT32 output_width : Input block width.
+ * INT32 *vp9_filter : Array of 2 bi-linear filter taps.
+ *
+ * OUTPUTS : INT32 *output_ptr : Pointer to filtered block.
+ *
+ * RETURNS : void
+ *
+ * FUNCTION : Applies a 1-D 2-tap bi-linear filter to the source block in
+ * either horizontal or vertical direction to produce the
+ * filtered output block. Used to implement first-pass
+ * of 2-D separable filter.
+ *
+ * SPECIAL NOTES : Produces INT32 output to retain precision for next pass.
+ * Two filter taps should sum to VP9_FILTER_WEIGHT.
+ * pixel_step defines whether the filter is applied
+ * horizontally (pixel_step=1) or vertically (pixel_step=stride).
+ * It defines the offset required to move from one input
+ * to the next.
+ *
+ ****************************************************************************/
+static void var_filter_block2d_bil_first_pass(const unsigned char *src_ptr,
+ unsigned short *output_ptr,
+ unsigned int src_pixels_per_line,
+ int pixel_step,
+ unsigned int output_height,
+ unsigned int output_width,
+ const short *vp9_filter) {+ unsigned int i, j;
+
+ for (i = 0; i < output_height; i++) {+ for (j = 0; j < output_width; j++) {+ // Apply bilinear filter
+ output_ptr[j] = (((int)src_ptr[0] * vp9_filter[0]) +
+ ((int)src_ptr[pixel_step] * vp9_filter[1]) +
+ (VP9_FILTER_WEIGHT / 2)) >> VP9_FILTER_SHIFT;
+ src_ptr++;
+ }
+
+ // Next row...
+ src_ptr += src_pixels_per_line - output_width;
+ output_ptr += output_width;
+ }
+}
+
+/****************************************************************************
+ *
+ * ROUTINE : filter_block2d_bil_second_pass
+ *
+ * INPUTS : INT32 *src_ptr : Pointer to source block.
+ * UINT32 src_pixels_per_line : Stride of input block.
+ * UINT32 pixel_step : Offset between filter input samples (see notes).
+ * UINT32 output_height : Input block height.
+ * UINT32 output_width : Input block width.
+ * INT32 *vp9_filter : Array of 2 bi-linear filter taps.
+ *
+ * OUTPUTS : UINT16 *output_ptr : Pointer to filtered block.
+ *
+ * RETURNS : void
+ *
+ * FUNCTION : Applies a 1-D 2-tap bi-linear filter to the source block in
+ * either horizontal or vertical direction to produce the
+ * filtered output 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.
+ * Two filter taps should sum to VP9_FILTER_WEIGHT.
+ * pixel_step defines whether the filter is applied
+ * horizontally (pixel_step=1) or vertically (pixel_step=stride).
+ * It defines the offset required to move from one input
+ * to the next.
+ *
+ ****************************************************************************/
+static void var_filter_block2d_bil_second_pass(const unsigned short *src_ptr,
+ unsigned char *output_ptr,
+ unsigned int src_pixels_per_line,
+ unsigned int pixel_step,
+ unsigned int output_height,
+ unsigned int output_width,
+ const short *vp9_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[0] * vp9_filter[0]) +
+ ((int)src_ptr[pixel_step] * vp9_filter[1]) +
+ (VP9_FILTER_WEIGHT / 2);
+ output_ptr[j] = (unsigned int)(Temp >> VP9_FILTER_SHIFT);
+ src_ptr++;
+ }
+
+ // Next row...
+ src_ptr += src_pixels_per_line - output_width;
+ output_ptr += output_width;
+ }
+}
+
--- a/vp9/common/textblit.c
+++ b/vp9/common/textblit.c
@@ -10,6 +10,7 @@
#include <stdlib.h>
+#include "vp9/common/textblit.h"
void vp9_blit_text(const char *msg, unsigned char *address, const int pitch) {int letter_bitmap;
--- /dev/null
+++ b/vp9/common/textblit.h
@@ -1,0 +1,19 @@
+/*
+ * Copyright (c) 2010 The WebM project authors. All Rights Reserved.
+ *
+ * Use of this source code is governed by a BSD-style license
+ * that can be found in the LICENSE file in the root of the source
+ * tree. An additional intellectual property rights grant can be found
+ * in the file PATENTS. All contributing project authors may
+ * be found in the AUTHORS file in the root of the source tree.
+ */
+
+#ifndef __INC_TEXTBLIT_H
+#define __INC_TEXTBLIT_H
+
+extern void vp9_blit_text(const char *msg, unsigned char *address,
+ const int pitch);
+extern void vp9_blit_line(int x0, int x1, int y0, int y1,
+ unsigned char *image, const int pitch);
+
+#endif // __INC_TEXTBLIT_H
--- a/vp9/common/x86/loopfilter_x86.c
+++ b/vp9/common/x86/loopfilter_x86.c
@@ -97,13 +97,17 @@
DECLARE_ALIGNED(16, unsigned char, flat_oq1[16]);
DECLARE_ALIGNED(16, unsigned char, flat_oq0[16]);
__m128i mask, hev, flat;
- __m128i thresh, limit, blimit;
const __m128i zero = _mm_set1_epi16(0);
__m128i p4, p3, p2, p1, p0, q0, q1, q2, q3, q4;
-
- thresh = _mm_shuffle_epi32(_mm_cvtsi32_si128(_thresh[0] * 0x01010101), 0);
- limit = _mm_shuffle_epi32(_mm_cvtsi32_si128(_limit[0] * 0x01010101), 0);
- blimit = _mm_shuffle_epi32(_mm_cvtsi32_si128(_blimit[0] * 0x01010101), 0);
+ const unsigned int extended_thresh = _thresh[0] * 0x01010101u;
+ const unsigned int extended_limit = _limit[0] * 0x01010101u;
+ const unsigned int extended_blimit = _blimit[0] * 0x01010101u;
+ const __m128i thresh =
+ _mm_shuffle_epi32(_mm_cvtsi32_si128((int)extended_thresh), 0);
+ const __m128i limit =
+ _mm_shuffle_epi32(_mm_cvtsi32_si128((int)extended_limit), 0);
+ const __m128i blimit =
+ _mm_shuffle_epi32(_mm_cvtsi32_si128((int)extended_blimit), 0);
p4 = _mm_loadu_si128((__m128i *)(s - 5 * p));
p3 = _mm_loadu_si128((__m128i *)(s - 4 * p));
--- a/vp9/common/x86/sadmxn_x86.c
+++ b/vp9/common/x86/sadmxn_x86.c
@@ -13,9 +13,6 @@
#include "./vpx_rtcd.h"
-#if CONFIG_NEWBESTREFMV
-
-
#if HAVE_SSE2
unsigned int vp9_sad16x3_sse2(
const unsigned char *src_ptr,
@@ -89,4 +86,3 @@
#endif
-#endif // CONFIG_NEWBESTREFMV
--- a/vp9/decoder/arm/armv6/idct_blk_v6.c
+++ b/vp9/decoder/arm/armv6/idct_blk_v6.c
@@ -12,9 +12,10 @@
#include "vp9/common/idct.h"
#include "vp9/decoder/dequantize.h"
-void vp8_dequant_dc_idct_add_y_block_v6
-(short *q, short *dq, unsigned char *pre,
- unsigned char *dst, int stride, char *eobs, short *dc) {+void vp8_dequant_dc_idct_add_y_block_v6(short *q, short *dq,
+ unsigned char *pre,
+ unsigned char *dst, int stride,
+ unsigned short *eobs, short *dc) {int i;
for (i = 0; i < 4; i++) {@@ -46,9 +47,9 @@
}
}
-void vp8_dequant_idct_add_y_block_v6
-(short *q, short *dq, unsigned char *pre,
- unsigned char *dst, int stride, char *eobs) {+void vp8_dequant_idct_add_y_block_v6(short *q, short *dq, unsigned char *pre,
+ unsigned char *dst, int stride,
+ unsigned short *eobs) {int i;
for (i = 0; i < 4; i++) {@@ -87,9 +88,9 @@
}
}
-void vp8_dequant_idct_add_uv_block_v6
-(short *q, short *dq, unsigned char *pre,
- unsigned char *dstu, unsigned char *dstv, int stride, char *eobs) {+void vp8_dequant_idct_add_uv_block_v6(short *q, short *dq, unsigned char *pre,
+ unsigned char *dstu, unsigned char *dstv,
+ int stride, unsigned short *eobs) {int i;
for (i = 0; i < 2; i++) {--- a/vp9/decoder/arm/neon/idct_blk_neon.c
+++ b/vp9/decoder/arm/neon/idct_blk_neon.c
@@ -27,9 +27,10 @@
(short *q, short dq, unsigned char *pre, int pitch,
unsigned char *dst, int stride);
-void vp8_dequant_dc_idct_add_y_block_neon
-(short *q, short *dq, unsigned char *pre,
- unsigned char *dst, int stride, char *eobs, short *dc) {+void vp8_dequant_dc_idct_add_y_block_neon(short *q, short *dq,
+ unsigned char *pre,
+ unsigned char *dst, int stride,
+ unsigned short *eobs, short *dc) {int i;
for (i = 0; i < 4; i++) {@@ -51,9 +52,9 @@
}
}
-void vp8_dequant_idct_add_y_block_neon
-(short *q, short *dq, unsigned char *pre,
- unsigned char *dst, int stride, char *eobs) {+void vp8_dequant_idct_add_y_block_neon(short *q, short *dq, unsigned char *pre,
+ unsigned char *dst, int stride,
+ unsigned short *eobs) {int i;
for (i = 0; i < 4; i++) {@@ -74,9 +75,11 @@
}
}
-void vp8_dequant_idct_add_uv_block_neon
-(short *q, short *dq, unsigned char *pre,
- unsigned char *dstu, unsigned char *dstv, int stride, char *eobs) {+void vp8_dequant_idct_add_uv_block_neon(short *q, short *dq,
+ unsigned char *pre,
+ unsigned char *dstu,
+ unsigned char *dstv, int stride,
+ unsigned short *eobs) {if (((short *)eobs)[0] & 0xfefe)
idct_dequant_full_2x_neon(q, dq, pre, dstu, 8, stride);
else
--- a/vp9/decoder/dboolhuff.h
+++ b/vp9/decoder/dboolhuff.h
@@ -56,7 +56,7 @@
{ \int shift = VP9_BD_VALUE_SIZE - 8 - ((_count) + 8); \
int loop_end, x; \
- size_t bits_left = ((_bufend)-(_bufptr))*CHAR_BIT; \
+ int bits_left = (int)(((_bufend)-(_bufptr))*CHAR_BIT); \
\
x = shift + CHAR_BIT - bits_left; \
loop_end = 0; \
--- a/vp9/decoder/decodemv.c
+++ b/vp9/decoder/decodemv.c
@@ -654,7 +654,7 @@
const int mis = pbi->common.mode_info_stride;
MACROBLOCKD *const xd = &pbi->mb;
- int_mv *const mv = &mbmi->mv;
+ int_mv *const mv = &mbmi->mv[0];
int mb_to_left_edge;
int mb_to_right_edge;
int mb_to_top_edge;
@@ -712,17 +712,13 @@
int_mv nearest_second, nearby_second, best_mv_second;
vp9_prob mv_ref_p [VP9_MVREFS - 1];
-#if CONFIG_NEWBESTREFMV
int recon_y_stride, recon_yoffset;
int recon_uv_stride, recon_uvoffset;
-#endif
vp9_find_near_mvs(xd, mi,
prev_mi,
&nearest, &nearby, &best_mv, rct,
mbmi->ref_frame, cm->ref_frame_sign_bias);
-
-#if CONFIG_NEWBESTREFMV
{int ref_fb_idx;
MV_REFERENCE_FRAME ref_frame = mbmi->ref_frame;
@@ -755,7 +751,6 @@
mbmi->ref_mvs[ref_frame],
&best_mv, &nearest, &nearby);
}
-#endif
vp9_mv_ref_probs(&pbi->common, mv_ref_p, rct);
@@ -808,7 +803,6 @@
mbmi->second_ref_frame = mbmi->ref_frame + 1;
if (mbmi->second_ref_frame == 4)
mbmi->second_ref_frame = 1;
-#if CONFIG_NEWBESTREFMV
if (mbmi->second_ref_frame) {int second_ref_fb_idx;
/* Select the appropriate reference frame for this MB */
@@ -845,13 +839,7 @@
&nearest_second,
&nearby_second);
}
-#else
- vp9_find_near_mvs(xd, mi, prev_mi,
- &nearest_second, &nearby_second, &best_mv_second,
- rct,
- mbmi->second_ref_frame,
- pbi->common.ref_frame_sign_bias);
-#endif
+
} else {mbmi->second_ref_frame = 0;
}
@@ -1172,7 +1160,7 @@
}
}
-void vp9_decode_mode_mvs_init(VP9D_COMP *pbi, BOOL_DECODER* const bc) {+void vp9_decode_mode_mvs_init(VP9D_COMP* const pbi, BOOL_DECODER* const bc) {VP9_COMMON *cm = &pbi->common;
vpx_memset(cm->mbskip_pred_probs, 0, sizeof(cm->mbskip_pred_probs));
@@ -1184,8 +1172,8 @@
mb_mode_mv_init(pbi, bc);
}
-void vp9_decode_mb_mode_mv(VP9D_COMP *pbi,
- MACROBLOCKD *xd,
+void vp9_decode_mb_mode_mv(VP9D_COMP* const pbi,
+ MACROBLOCKD* const xd,
int mb_row,
int mb_col,
BOOL_DECODER* const bc) {--- a/vp9/decoder/decodframe.c
+++ b/vp9/decoder/decodframe.c
@@ -14,6 +14,7 @@
#include "vp9/common/reconintra.h"
#include "vp9/common/reconintra4x4.h"
#include "vp9/common/reconinter.h"
+#include "vp9/decoder/decodframe.h"
#include "detokenize.h"
#include "vp9/common/invtrans.h"
#include "vp9/common/alloccommon.h"
@@ -728,7 +729,8 @@
"%d length", 1);
}
- if (vp9_start_decode(bool_decoder, partition, partition_size))
+ if (vp9_start_decode(bool_decoder,
+ partition, (unsigned int)partition_size))
vpx_internal_error(&pc->error, VPX_CODEC_MEM_ERROR,
"Failed to allocate bool decoder %d", 1);
}
@@ -985,7 +987,8 @@
init_frame(pbi);
- if (vp9_start_decode(&header_bc, data, first_partition_length_in_bytes))
+ if (vp9_start_decode(&header_bc, data,
+ (unsigned int)first_partition_length_in_bytes))
vpx_internal_error(&pc->error, VPX_CODEC_MEM_ERROR,
"Failed to allocate bool decoder 0");
if (pc->frame_type == KEY_FRAME) {--- /dev/null
+++ b/vp9/decoder/decodframe.h
@@ -1,0 +1,19 @@
+/*
+ * Copyright (c) 2010 The WebM project authors. All Rights Reserved.
+ *
+ * Use of this source code is governed by a BSD-style license
+ * that can be found in the LICENSE file in the root of the source
+ * tree. An additional intellectual property rights grant can be found
+ * in the file PATENTS. All contributing project authors may
+ * be found in the AUTHORS file in the root of the source tree.
+ */
+
+
+#ifndef __INC_DECODFRAME_H
+#define __INC_DECODFRAME_H
+
+struct VP9Decompressor;
+
+extern void vp9_init_de_quantizer(struct VP9Decompressor *pbi);
+
+#endif // __INC_DECODFRAME_H
--- a/vp9/decoder/dequantize.c
+++ b/vp9/decoder/dequantize.c
@@ -9,23 +9,11 @@
*/
-#include "vpx_ports/config.h"
+#include "vpx_rtcd.h"
#include "dequantize.h"
#include "vp9/common/idct.h"
#include "vpx_mem/vpx_mem.h"
#include "onyxd_int.h"
-
-extern void vp9_short_idct4x4llm_c(short *input, short *output, int pitch);
-extern void vp9_short_idct4x4llm_1_c(short *input, short *output, int pitch);
-extern void vp9_short_idct8x8_c(short *input, short *output, int pitch);
-extern void vp9_short_idct8x8_1_c(short *input, short *output, int pitch);
-
-#if CONFIG_LOSSLESS
-extern void vp9_short_inv_walsh4x4_x8_c(short *input, short *output,
- int pitch);
-extern void vp9_short_inv_walsh4x4_1_x8_c(short *input, short *output,
- int pitch);
-#endif
#ifdef DEC_DEBUG
extern int dec_debug;
--- a/vp9/decoder/dequantize.h
+++ b/vp9/decoder/dequantize.h
@@ -25,17 +25,20 @@
extern void vp9_dequant_dc_idct_add_y_block_lossless_c(short *q, short *dq,
unsigned char *pre,
unsigned char *dst,
- int stride, char *eobs,
+ int stride,
+ unsigned short *eobs,
short *dc);
extern void vp9_dequant_idct_add_y_block_lossless_c(short *q, short *dq,
unsigned char *pre,
unsigned char *dst,
- int stride, char *eobs);
+ int stride,
+ unsigned short *eobs);
extern void vp9_dequant_idct_add_uv_block_lossless_c(short *q, short *dq,
unsigned char *pre,
unsigned char *dst_u,
unsigned char *dst_v,
- int stride, char *eobs);
+ int stride,
+ unsigned short *eobs);
#endif
typedef void (*vp9_dequant_idct_add_fn_t)(short *input, short *dq,
@@ -44,12 +47,13 @@
unsigned char *pred, unsigned char *output, int pitch, int stride, int dc);
typedef void(*vp9_dequant_dc_idct_add_y_block_fn_t)(short *q, short *dq,
- unsigned char *pre, unsigned char *dst, int stride, char *eobs, short *dc);
+ unsigned char *pre, unsigned char *dst, int stride, unsigned short *eobs,
+ short *dc);
typedef void(*vp9_dequant_idct_add_y_block_fn_t)(short *q, short *dq,
- unsigned char *pre, unsigned char *dst, int stride, char *eobs);
+ unsigned char *pre, unsigned char *dst, int stride, unsigned short *eobs);
typedef void(*vp9_dequant_idct_add_uv_block_fn_t)(short *q, short *dq,
unsigned char *pre, unsigned char *dst_u, unsigned char *dst_v, int stride,
- char *eobs);
+ unsigned short *eobs);
void vp9_ht_dequant_idct_add_c(TX_TYPE tx_type, short *input, short *dq,
unsigned char *pred, unsigned char *dest,
@@ -66,12 +70,14 @@
#if CONFIG_SUPERBLOCKS
void vp9_dequant_dc_idct_add_y_block_8x8_inplace_c(short *q, short *dq,
unsigned char *dst,
- int stride, char *eobs,
+ int stride,
+ unsigned short *eobs,
short *dc, MACROBLOCKD *xd);
void vp9_dequant_idct_add_uv_block_8x8_inplace_c(short *q, short *dq,
unsigned char *dstu,
unsigned char *dstv,
- int stride, char *eobs,
+ int stride,
+ unsigned short *eobs,
MACROBLOCKD *xd);
#endif
--- a/vp9/decoder/detokenize.c
+++ b/vp9/decoder/detokenize.c
@@ -98,7 +98,7 @@
static const unsigned char cat6_prob[14] =
{ 254, 254, 252, 249, 243, 230, 196, 177, 153, 140, 133, 130, 129, 0 };-void vp9_reset_mb_tokens_context(MACROBLOCKD *xd) {+void vp9_reset_mb_tokens_context(MACROBLOCKD* const xd) {/* Clear entropy contexts for Y2 blocks */
if ((xd->mode_info_context->mbmi.mode != B_PRED &&
xd->mode_info_context->mbmi.mode != I8X8_PRED &&
@@ -414,12 +414,13 @@
return c;
}
-int vp9_decode_mb_tokens_16x16(VP9D_COMP *pbi, MACROBLOCKD *xd,
+int vp9_decode_mb_tokens_16x16(VP9D_COMP* const pbi,
+ MACROBLOCKD* const xd,
BOOL_DECODER* const bc) {ENTROPY_CONTEXT* const A = (ENTROPY_CONTEXT *)xd->above_context;
ENTROPY_CONTEXT* const L = (ENTROPY_CONTEXT *)xd->left_context;
- char* const eobs = xd->eobs;
+ unsigned short* const eobs = xd->eobs;
PLANE_TYPE type;
int c, i, eobtotal = 0, seg_eob;
const int segment_id = xd->mode_info_context->mbmi.segment_id;
@@ -477,12 +478,13 @@
return eobtotal;
}
-int vp9_decode_mb_tokens_8x8(VP9D_COMP *pbi, MACROBLOCKD *xd,
+int vp9_decode_mb_tokens_8x8(VP9D_COMP* const pbi,
+ MACROBLOCKD* const xd,
BOOL_DECODER* const bc) {ENTROPY_CONTEXT *const A = (ENTROPY_CONTEXT *)xd->above_context;
ENTROPY_CONTEXT *const L = (ENTROPY_CONTEXT *)xd->left_context;
- char *const eobs = xd->eobs;
+ unsigned short *const eobs = xd->eobs;
PLANE_TYPE type;
int c, i, eobtotal = 0, seg_eob;
const int segment_id = xd->mode_info_context->mbmi.segment_id;
@@ -571,12 +573,13 @@
}
-int vp9_decode_mb_tokens(VP9D_COMP *dx, MACROBLOCKD *xd,
+int vp9_decode_mb_tokens(VP9D_COMP* const dx,
+ MACROBLOCKD* const xd,
BOOL_DECODER* const bc) {ENTROPY_CONTEXT *const A = (ENTROPY_CONTEXT *)xd->above_context;
ENTROPY_CONTEXT *const L = (ENTROPY_CONTEXT *)xd->left_context;
- char *const eobs = xd->eobs;
+ unsigned short *const eobs = xd->eobs;
const int *scan = vp9_default_zig_zag1d;
PLANE_TYPE type;
int c, i, eobtotal = 0, seg_eob = 16;
--- a/vp9/decoder/idct_blk.c
+++ b/vp9/decoder/idct_blk.c
@@ -8,30 +8,13 @@
* be found in the AUTHORS file in the root of the source tree.
*/
-#include "vpx_ports/config.h"
+#include "vpx_rtcd.h"
#include "vp9/common/idct.h"
-#include "dequantize.h"
-void vp9_dequant_dc_idct_add_c(short *input, short *dq, unsigned char *pred,
- unsigned char *dest, int pitch, int stride,
- int Dc);
-void vp9_dequant_idct_add_c(short *input, short *dq, unsigned char *pred,
- unsigned char *dest, int pitch, int stride);
-void vp9_dc_only_idct_add_c(short input_dc, unsigned char *pred_ptr,
- unsigned char *dst_ptr, int pitch, int stride);
-#if CONFIG_LOSSLESS
-void vp9_dequant_idct_add_lossless_c(short *input, short *dq,
- unsigned char *pred, unsigned char *dest,
- int pitch, int stride);
-void vp9_dc_only_idct_add_lossless_c(short input_dc, unsigned char *pred_ptr,
- unsigned char *dst_ptr,
- int pitch, int stride);
-#endif
-
void vp9_dequant_dc_idct_add_y_block_c(short *q, short *dq,
unsigned char *pre,
unsigned char *dst,
- int stride, char *eobs,
+ int stride, unsigned short *eobs,
short *dc) {int i, j;
@@ -56,7 +39,7 @@
void vp9_dequant_idct_add_y_block_c(short *q, short *dq,
unsigned char *pre,
unsigned char *dst,
- int stride, char *eobs) {+ int stride, unsigned short *eobs) {int i, j;
for (i = 0; i < 4; i++) {@@ -80,7 +63,7 @@
void vp9_dequant_idct_add_uv_block_c(short *q, short *dq, unsigned char *pre,
unsigned char *dstu, unsigned char *dstv,
- int stride, char *eobs) {+ int stride, unsigned short *eobs) {int i, j;
for (i = 0; i < 2; i++) {@@ -124,7 +107,8 @@
void vp9_dequant_dc_idct_add_y_block_8x8_c(short *q, short *dq,
unsigned char *pre,
unsigned char *dst,
- int stride, char *eobs, short *dc,
+ int stride, unsigned short *eobs,
+ short *dc,
MACROBLOCKD *xd) {vp9_dequant_dc_idct_add_8x8_c(q, dq, pre, dst, 16, stride, dc[0]);
vp9_dequant_dc_idct_add_8x8_c(&q[64], dq, pre + 8, dst + 8, 16, stride, dc[1]);
@@ -137,7 +121,8 @@
#if CONFIG_SUPERBLOCKS
void vp9_dequant_dc_idct_add_y_block_8x8_inplace_c(short *q, short *dq,
unsigned char *dst,
- int stride, char *eobs,
+ int stride,
+ unsigned short *eobs,
short *dc, MACROBLOCKD *xd) {vp9_dequant_dc_idct_add_8x8_c(q, dq, dst, dst, stride, stride, dc[0]);
vp9_dequant_dc_idct_add_8x8_c(&q[64], dq, dst + 8,
@@ -152,7 +137,7 @@
void vp9_dequant_idct_add_y_block_8x8_c(short *q, short *dq,
unsigned char *pre,
unsigned char *dst,
- int stride, char *eobs,
+ int stride, unsigned short *eobs,
MACROBLOCKD *xd) {unsigned char *origdest = dst;
unsigned char *origpred = pre;
@@ -170,7 +155,7 @@
unsigned char *pre,
unsigned char *dstu,
unsigned char *dstv,
- int stride, char *eobs,
+ int stride, unsigned short *eobs,
MACROBLOCKD *xd) {vp9_dequant_idct_add_8x8_c(q, dq, pre, dstu, 8, stride);
@@ -184,7 +169,8 @@
void vp9_dequant_idct_add_uv_block_8x8_inplace_c(short *q, short *dq,
unsigned char *dstu,
unsigned char *dstv,
- int stride, char *eobs,
+ int stride,
+ unsigned short *eobs,
MACROBLOCKD *xd) {vp9_dequant_idct_add_8x8_c(q, dq, dstu, dstu, stride, stride);
@@ -198,7 +184,8 @@
void vp9_dequant_dc_idct_add_y_block_lossless_c(short *q, short *dq,
unsigned char *pre,
unsigned char *dst,
- int stride, char *eobs,
+ int stride,
+ unsigned short *eobs,
short *dc) {int i, j;
@@ -223,7 +210,7 @@
void vp9_dequant_idct_add_y_block_lossless_c(short *q, short *dq,
unsigned char *pre,
unsigned char *dst,
- int stride, char *eobs) {+ int stride, unsigned short *eobs) {int i, j;
for (i = 0; i < 4; i++) {@@ -249,7 +236,8 @@
unsigned char *pre,
unsigned char *dstu,
unsigned char *dstv,
- int stride, char *eobs) {+ int stride,
+ unsigned short *eobs) {int i, j;
for (i = 0; i < 2; i++) {--- a/vp9/decoder/onyxd_if.c
+++ b/vp9/decoder/onyxd_if.c
@@ -26,12 +26,12 @@
#include "vpx_scale/vpxscale.h"
#include "vp9/common/systemdependent.h"
#include "vpx_ports/vpx_timer.h"
+#include "vp9/decoder/decodframe.h"
#include "detokenize.h"
#if ARCH_ARM
#include "vpx_ports/arm.h"
#endif
-extern void vp9_init_de_quantizer(VP9D_COMP *pbi);
static int get_free_fb(VP9_COMMON *cm);
static void ref_cnt_fb(int *buf, int *idx, int new_idx);
--- a/vp9/decoder/x86/idct_blk_mmx.c
+++ b/vp9/decoder/x86/idct_blk_mmx.c
@@ -15,7 +15,8 @@
void vp9_dequant_dc_idct_add_y_block_mmx(short *q, short *dq,
unsigned char *pre,
unsigned char *dst,
- int stride, char *eobs, short *dc) {+ int stride, unsigned short *eobs,
+ short *dc) {int i;
for (i = 0; i < 4; i++) {@@ -53,7 +54,7 @@
void vp9_dequant_idct_add_y_block_mmx(short *q, short *dq,
unsigned char *pre,
unsigned char *dst,
- int stride, char *eobs) {+ int stride, unsigned short *eobs) {int i;
for (i = 0; i < 4; i++) {@@ -96,7 +97,7 @@
unsigned char *pre,
unsigned char *dstu,
unsigned char *dstv,
- int stride, char *eobs) {+ int stride, unsigned short *eobs) {int i;
for (i = 0; i < 2; i++) {--- a/vp9/decoder/x86/idct_blk_sse2.c
+++ b/vp9/decoder/x86/idct_blk_sse2.c
@@ -31,7 +31,8 @@
void vp9_dequant_dc_idct_add_y_block_sse2(short *q, short *dq,
unsigned char *pre,
unsigned char *dst,
- int stride, char *eobs, short *dc) {+ int stride, unsigned short *eobs,
+ short *dc) {int i;
for (i = 0; i < 4; i++) {@@ -57,7 +58,7 @@
void vp9_dequant_idct_add_y_block_sse2(short *q, short *dq,
unsigned char *pre, unsigned char *dst,
- int stride, char *eobs) {+ int stride, unsigned short *eobs) {int i;
for (i = 0; i < 4; i++) {@@ -82,7 +83,7 @@
unsigned char *pre,
unsigned char *dstu,
unsigned char *dstv,
- int stride, char *eobs) {+ int stride, unsigned short *eobs) {if (((short *)(eobs))[0] & 0xfefe)
vp9_idct_dequant_full_2x_sse2(q, dq, pre, dstu, stride, 8);
else
--- a/vp9/encoder/bitstream.c
+++ b/vp9/encoder/bitstream.c
@@ -29,10 +29,7 @@
#include "vp9/common/entropy.h"
#include "vp9/encoder/encodemv.h"
#include "vp9/common/entropymv.h"
-
-#if CONFIG_NEWBESTREFMV
#include "vp9/common/mvref_common.h"
-#endif
#if defined(SECTIONBITS_OUTPUT)
unsigned __int64 Sectionbits[500];
@@ -186,7 +183,6 @@
void vp9_update_skip_probs(VP9_COMP *cpi) {VP9_COMMON *const pc = &cpi->common;
- int prob_skip_false[3] = {0, 0, 0};int k;
for (k = 0; k < MBSKIP_CONTEXTS; ++k) {@@ -218,7 +214,6 @@
static void update_refpred_stats(VP9_COMP *cpi) {VP9_COMMON *const cm = &cpi->common;
int i;
- int tot_count;
vp9_prob new_pred_probs[PREDICTION_PROBS];
int old_cost, new_cost;
@@ -884,7 +879,6 @@
}
static void pack_inter_mode_mvs(VP9_COMP *const cpi, vp9_writer *const bc) {- int i;
VP9_COMMON *const pc = &cpi->common;
const nmv_context *nmvc = &pc->fc.nmvc;
MACROBLOCK *x = &cpi->mb;
@@ -1062,9 +1056,8 @@
// Only used for context just now and soon to be deprecated.
vp9_find_near_mvs(xd, m, prev_m, &n1, &n2, &best_mv, ct,
rf, cpi->common.ref_frame_sign_bias);
-#if CONFIG_NEWBESTREFMV
+
best_mv.as_int = mi->ref_mvs[rf][0].as_int;
-#endif
vp9_mv_ref_probs(&cpi->common, mv_ref_p, ct);
@@ -1124,10 +1117,8 @@
mi->second_ref_frame,
cpi->common.ref_frame_sign_bias);
-#if CONFIG_NEWBESTREFMV
best_second_mv.as_int =
mi->ref_mvs[mi->second_ref_frame][0].as_int;
-#endif
}
// does the feature use compound prediction or not
@@ -1315,7 +1306,6 @@
const MODE_INFO *m,
int mode_info_stride,
vp9_writer *const bc) {- const int mis = mode_info_stride;
int ym;
int segment_id;
@@ -1331,6 +1321,7 @@
(vp9_get_segdata(xd, segment_id, SEG_LVL_EOB) != 0))) {int skip_coeff = m->mbmi.mb_skip_coeff;
#if CONFIG_SUPERBLOCKS
+ const int mis = mode_info_stride;
if (m->mbmi.encoded_as_sb) {skip_coeff &= m[1].mbmi.mb_skip_coeff;
skip_coeff &= m[mis].mbmi.mb_skip_coeff;
--- a/vp9/encoder/block.h
+++ b/vp9/encoder/block.h
@@ -70,9 +70,7 @@
PARTITION_INFO partition_info;
int_mv best_ref_mv;
int_mv second_best_ref_mv;
-#if CONFIG_NEWBESTREFMV || CONFIG_NEW_MVREF
int_mv ref_mvs[MAX_REF_FRAMES][MAX_MV_REFS];
-#endif
int rate;
int distortion;
int64_t intra_error;
--- a/vp9/encoder/dct.c
+++ b/vp9/encoder/dct.c
@@ -901,6 +901,8 @@
}
#endif
+#define TEST_INT_16x16_DCT 1
+#if !TEST_INT_16x16_DCT
static const double C1 = 0.995184726672197;
static const double C2 = 0.98078528040323;
static const double C3 = 0.956940335732209;
@@ -1107,3 +1109,225 @@
}
vp9_clear_system_state(); // Make it simd safe : __asm emms;
}
+
+#else
+static const int16_t C1 = 16305;
+static const int16_t C2 = 16069;
+static const int16_t C3 = 15679;
+static const int16_t C4 = 15137;
+static const int16_t C5 = 14449;
+static const int16_t C6 = 13623;
+static const int16_t C7 = 12665;
+static const int16_t C8 = 11585;
+static const int16_t C9 = 10394;
+static const int16_t C10 = 9102;
+static const int16_t C11 = 7723;
+static const int16_t C12 = 6270;
+static const int16_t C13 = 4756;
+static const int16_t C14 = 3196;
+static const int16_t C15 = 1606;
+
+#define RIGHT_SHIFT 14
+#define ROUNDING (1 << (RIGHT_SHIFT - 1))
+
+static void dct16x16_1d(int16_t input[16], int16_t output[16],
+ int last_shift_bits) {+ int16_t step[16];
+ int intermediate[16];
+ int temp1, temp2;
+ int final_shift = RIGHT_SHIFT;
+ int final_rounding = ROUNDING;
+ int output_shift = 0;
+ int output_rounding = 0;
+
+ final_shift += last_shift_bits;
+ if (final_shift > 0)
+ final_rounding = 1 << (final_shift - 1);
+
+ output_shift += last_shift_bits;
+ if (output_shift > 0)
+ output_rounding = 1 << (output_shift - 1);
+
+ // step 1
+ step[ 0] = input[0] + input[15];
+ step[ 1] = input[1] + input[14];
+ step[ 2] = input[2] + input[13];
+ step[ 3] = input[3] + input[12];
+ step[ 4] = input[4] + input[11];
+ step[ 5] = input[5] + input[10];
+ step[ 6] = input[6] + input[ 9];
+ step[ 7] = input[7] + input[ 8];
+ step[ 8] = input[7] - input[ 8];
+ step[ 9] = input[6] - input[ 9];
+ step[10] = input[5] - input[10];
+ step[11] = input[4] - input[11];
+ step[12] = input[3] - input[12];
+ step[13] = input[2] - input[13];
+ step[14] = input[1] - input[14];
+ step[15] = input[0] - input[15];
+
+ // step 2
+ output[0] = step[0] + step[7];
+ output[1] = step[1] + step[6];
+ output[2] = step[2] + step[5];
+ output[3] = step[3] + step[4];
+ output[4] = step[3] - step[4];
+ output[5] = step[2] - step[5];
+ output[6] = step[1] - step[6];
+ output[7] = step[0] - step[7];
+
+ temp1 = step[ 8] * C7;
+ temp2 = step[15] * C9;
+ output[ 8] = (temp1 + temp2 + ROUNDING) >> RIGHT_SHIFT;
+
+ temp1 = step[ 9] * C11;
+ temp2 = step[14] * C5;
+ output[ 9] = (temp1 - temp2 + ROUNDING) >> RIGHT_SHIFT;
+
+ temp1 = step[10] * C3;
+ temp2 = step[13] * C13;
+ output[10] = (temp1 + temp2 + ROUNDING) >> RIGHT_SHIFT;
+
+ temp1 = step[11] * C15;
+ temp2 = step[12] * C1;
+ output[11] = (temp1 - temp2 + ROUNDING) >> RIGHT_SHIFT;
+
+ temp1 = step[11] * C1;
+ temp2 = step[12] * C15;
+ output[12] = (temp2 + temp1 + ROUNDING) >> RIGHT_SHIFT;
+
+ temp1 = step[10] * C13;
+ temp2 = step[13] * C3;
+ output[13] = (temp2 - temp1 + ROUNDING) >> RIGHT_SHIFT;
+
+ temp1 = step[ 9] * C5;
+ temp2 = step[14] * C11;
+ output[14] = (temp2 + temp1 + ROUNDING) >> RIGHT_SHIFT;
+
+ temp1 = step[ 8] * C9;
+ temp2 = step[15] * C7;
+ output[15] = (temp2 - temp1 + ROUNDING) >> RIGHT_SHIFT;
+
+ // step 3
+ step[ 0] = output[0] + output[3];
+ step[ 1] = output[1] + output[2];
+ step[ 2] = output[1] - output[2];
+ step[ 3] = output[0] - output[3];
+
+ temp1 = output[4] * C14;
+ temp2 = output[7] * C2;
+ step[ 4] = (temp1 + temp2 + ROUNDING) >> RIGHT_SHIFT;
+
+ temp1 = output[5] * C10;
+ temp2 = output[6] * C6;
+ step[ 5] = (temp1 + temp2 + ROUNDING) >> RIGHT_SHIFT;
+
+ temp1 = output[5] * C6;
+ temp2 = output[6] * C10;
+ step[ 6] = (temp2 - temp1 + ROUNDING) >> RIGHT_SHIFT;
+
+ temp1 = output[4] * C2;
+ temp2 = output[7] * C14;
+ step[ 7] = (temp2 - temp1 + ROUNDING) >> RIGHT_SHIFT;
+
+ step[ 8] = output[ 8] + output[11];
+ step[ 9] = output[ 9] + output[10];
+ step[10] = output[ 9] - output[10];
+ step[11] = output[ 8] - output[11];
+
+ step[12] = output[12] + output[15];
+ step[13] = output[13] + output[14];
+ step[14] = output[13] - output[14];
+ step[15] = output[12] - output[15];
+
+ // step 4
+ output[ 0] = (step[ 0] + step[ 1] + output_rounding) >> output_shift;
+ output[ 8] = (step[ 0] - step[ 1] + output_rounding) >> output_shift;
+
+ temp1 = step[2] * C12;
+ temp2 = step[3] * C4;
+ temp1 = (temp1 + temp2 + final_rounding) >> final_shift;
+ output[ 4] = (2 * (temp1 * C8) + ROUNDING) >> RIGHT_SHIFT;
+
+ temp1 = step[2] * C4;
+ temp2 = step[3] * C12;
+ temp1 = (temp2 - temp1 + final_rounding) >> final_shift;
+ output[12] = (2 * (temp1 * C8) + ROUNDING) >> RIGHT_SHIFT;
+
+ output[ 2] = (2 * ((step[4] + step[ 5]) * C8) + final_rounding)
+ >> final_shift;
+ output[14] = (2 * ((step[7] - step[ 6]) * C8) + final_rounding)
+ >> final_shift;
+
+ temp1 = step[4] - step[5];
+ temp2 = step[6] + step[7];
+ output[ 6] = (temp1 + temp2 + output_rounding) >> output_shift;
+ output[10] = (temp1 - temp2 + output_rounding) >> output_shift;
+
+ intermediate[8] = step[8] + step[14];
+ intermediate[9] = step[9] + step[15];
+
+ temp1 = intermediate[8] * C12;
+ temp2 = intermediate[9] * C4;
+ temp1 = (temp1 - temp2 + final_rounding) >> final_shift;
+ output[3] = (2 * (temp1 * C8) + ROUNDING) >> RIGHT_SHIFT;
+
+ temp1 = intermediate[8] * C4;
+ temp2 = intermediate[9] * C12;
+ temp1 = (temp2 + temp1 + final_rounding) >> final_shift;
+ output[13] = (2 * (temp1 * C8) + ROUNDING) >> RIGHT_SHIFT;
+
+ output[ 9] = (2 * ((step[10] + step[11]) * C8) + final_rounding)
+ >> final_shift;
+
+ intermediate[11] = step[10] - step[11];
+ intermediate[12] = step[12] + step[13];
+ intermediate[13] = step[12] - step[13];
+ intermediate[14] = step[ 8] - step[14];
+ intermediate[15] = step[ 9] - step[15];
+
+ output[15] = (intermediate[11] + intermediate[12] + output_rounding)
+ >> output_shift;
+ output[ 1] = -(intermediate[11] - intermediate[12] + output_rounding)
+ >> output_shift;
+
+ output[ 7] = (2 * (intermediate[13] * C8) + final_rounding) >> final_shift;
+
+ temp1 = intermediate[14] * C12;
+ temp2 = intermediate[15] * C4;
+ temp1 = (temp1 - temp2 + final_rounding) >> final_shift;
+ output[11] = (-2 * (temp1 * C8) + ROUNDING) >> RIGHT_SHIFT;
+
+ temp1 = intermediate[14] * C4;
+ temp2 = intermediate[15] * C12;
+ temp1 = (temp2 + temp1 + final_rounding) >> final_shift;
+ output[ 5] = (2 * (temp1 * C8) + ROUNDING) >> RIGHT_SHIFT;
+}
+
+void vp9_short_fdct16x16_c(int16_t *input, int16_t *out, int pitch) {+ int shortpitch = pitch >> 1;
+ int i, j;
+ int16_t output[256];
+ int16_t *outptr = &output[0];
+
+ // First transform columns
+ for (i = 0; i < 16; i++) {+ int16_t temp_in[16];
+ int16_t temp_out[16];
+ for (j = 0; j < 16; j++)
+ temp_in[j] = input[j * shortpitch + i];
+ dct16x16_1d(temp_in, temp_out, 0);
+ for (j = 0; j < 16; j++)
+ output[j * 16 + i] = temp_out[j];
+ }
+
+ // Then transform rows
+ for (i = 0; i < 16; ++i) {+ dct16x16_1d(outptr, out, 1);
+ outptr += 16;
+ out += 16;
+ }
+}
+#undef RIGHT_SHIFT
+#undef ROUNDING
+#endif
--- a/vp9/encoder/encodeframe.c
+++ b/vp9/encoder/encodeframe.c
@@ -10,6 +10,7 @@
#include "vpx_ports/config.h"
+#include "vp9/encoder/encodeframe.h"
#include "encodemb.h"
#include "encodemv.h"
#include "vp9/common/common.h"
@@ -27,6 +28,7 @@
#include "vp9/common/findnearmv.h"
#include "vp9/common/reconintra.h"
#include "vp9/common/seg_common.h"
+#include "vp9/encoder/tokenize.h"
#include "vpx_rtcd.h"
#include <stdio.h>
#include <math.h>
@@ -34,13 +36,10 @@
#include "vp9/common/subpixel.h"
#include "vpx_ports/vpx_timer.h"
#include "vp9/common/pred_common.h"
+#include "vp9/common/mvref_common.h"
#define DBG_PRNT_SEGMAP 0
-#if CONFIG_NEWBESTREFMV
-#include "vp9/common/mvref_common.h"
-#endif
-
#if CONFIG_RUNTIME_CPU_DETECT
#define RTCD(x) &cpi->common.rtcd.x
#define IF_RTCD(x) (x)
@@ -54,36 +53,20 @@
int mb_row_debug, mb_col_debug;
#endif
-extern void vp9_initialize_me_consts(VP9_COMP *cpi, int QIndex);
+static void encode_inter_macroblock(VP9_COMP *cpi, MACROBLOCK *x,
+ TOKENEXTRA **t, int recon_yoffset,
+ int recon_uvoffset, int output_enabled);
-extern void vp9_auto_select_speed(VP9_COMP *cpi);
+static void encode_inter_superblock(VP9_COMP *cpi, MACROBLOCK *x,
+ TOKENEXTRA **t, int recon_yoffset,
+ int recon_uvoffset, int mb_col, int mb_row);
-int64_t vp9_rd_pick_inter_mode_sb(VP9_COMP *cpi, MACROBLOCK *x,
- int recon_yoffset, int recon_uvoffset,
- int *returnrate, int *returndistortion);
+static void encode_intra_macro_block(VP9_COMP *cpi, MACROBLOCK *x,
+ TOKENEXTRA **t, int output_enabled);
-extern void vp9_pick_mode_inter_macroblock(VP9_COMP *cpi, MACROBLOCK *x,
- int recon_yoffset,
- int recon_uvoffset, int *r, int *d);
+static void encode_intra_super_block(VP9_COMP *cpi, MACROBLOCK *x,
+ TOKENEXTRA **t, int mb_col);
-void vp9_build_block_offsets(MACROBLOCK *x);
-
-void vp9_setup_block_ptrs(MACROBLOCK *x);
-
-void vp9_encode_inter_macroblock(VP9_COMP *cpi, MACROBLOCK *x, TOKENEXTRA **t,
- int recon_yoffset, int recon_uvoffset,
- int output_enabled);
-
-void vp9_encode_inter_superblock(VP9_COMP *cpi, MACROBLOCK *x, TOKENEXTRA **t,
- int recon_yoffset, int recon_uvoffset,
- int mb_col, int mb_row);
-
-void vp9_encode_intra_macro_block(VP9_COMP *cpi, MACROBLOCK *x,
- TOKENEXTRA **t, int output_enabled);
-
-void vp9_encode_intra_super_block(VP9_COMP *cpi, MACROBLOCK *x,
- TOKENEXTRA **t, int mb_col);
-
static void adjust_act_zbin(VP9_COMP *cpi, MACROBLOCK *x);
#ifdef MODE_STATS
@@ -634,7 +617,7 @@
*totaldist += d;
// Dummy encode, do not do the tokenization
- vp9_encode_intra_macro_block(cpi, x, tp, 0);
+ encode_intra_macro_block(cpi, x, tp, 0);
// Note the encoder may have changed the segment_id
// Save the coding context
@@ -661,8 +644,8 @@
*totaldist += d;
// Dummy encode, do not do the tokenization
- vp9_encode_inter_macroblock(cpi, x, tp,
- recon_yoffset, recon_uvoffset, 0);
+ encode_inter_macroblock(cpi, x, tp,
+ recon_yoffset, recon_uvoffset, 0);
seg_id = mbmi->segment_id;
if (cpi->mb.e_mbd.segmentation_enabled && seg_id == 0) {@@ -992,10 +975,10 @@
if (cm->frame_type == KEY_FRAME) {#if CONFIG_SUPERBLOCKS
if (xd->mode_info_context->mbmi.encoded_as_sb)
- vp9_encode_intra_super_block(cpi, x, tp, mb_col);
+ encode_intra_super_block(cpi, x, tp, mb_col);
else
#endif
- vp9_encode_intra_macro_block(cpi, x, tp, 1);
+ encode_intra_macro_block(cpi, x, tp, 1);
// Note the encoder may have changed the segment_id
#ifdef MODE_STATS
@@ -1018,12 +1001,12 @@
#if CONFIG_SUPERBLOCKS
if (xd->mode_info_context->mbmi.encoded_as_sb)
- vp9_encode_inter_superblock(cpi, x, tp, recon_yoffset, recon_uvoffset,
- mb_col, mb_row);
+ encode_inter_superblock(cpi, x, tp, recon_yoffset, recon_uvoffset,
+ mb_col, mb_row);
else
#endif
- vp9_encode_inter_macroblock(cpi, x, tp,
- recon_yoffset, recon_uvoffset, 1);
+ encode_inter_macroblock(cpi, x, tp,
+ recon_yoffset, recon_uvoffset, 1);
// Note the encoder may have changed the segment_id
#ifdef MODE_STATS
@@ -1404,7 +1387,7 @@
x->src.v_buffer += 16 * x->src.uv_stride - 8 * offset;
}
- cpi->tok_count = tp - cpi->tok;
+ cpi->tok_count = (unsigned int)(tp - cpi->tok);
}
vpx_usec_timer_mark(&emr_timer);
@@ -1592,7 +1575,7 @@
encode_frame_internal(cpi);
for (i = 0; i < NB_PREDICTION_TYPES; ++i) {- const int diff = cpi->rd_comp_pred_diff[i] / cpi->common.MBs;
+ const int diff = (int)(cpi->rd_comp_pred_diff[i] / cpi->common.MBs);
cpi->rd_prediction_type_threshes[frame_type][i] += diff;
cpi->rd_prediction_type_threshes[frame_type][i] >>= 1;
}
@@ -1602,7 +1585,7 @@
int diff;
if (i == TX_MODE_SELECT)
pd -= RDCOST(cpi->mb.rdmult, cpi->mb.rddiv, 2048 * (TX_SIZE_MAX - 1), 0);
- diff = pd / cpi->common.MBs;
+ diff = (int)(pd / cpi->common.MBs);
cpi->rd_tx_select_threshes[frame_type][i] += diff;
cpi->rd_tx_select_threshes[frame_type][i] /= 2;
}
@@ -1851,10 +1834,8 @@
}
}
-void vp9_encode_intra_super_block(VP9_COMP *cpi,
- MACROBLOCK *x,
- TOKENEXTRA **t,
- int mb_col) {+static void encode_intra_super_block(VP9_COMP *cpi, MACROBLOCK *x,
+ TOKENEXTRA **t, int mb_col) {const int output_enabled = 1;
int n;
MACROBLOCKD *xd = &x->e_mbd;
@@ -1931,10 +1912,8 @@
}
#endif /* CONFIG_SUPERBLOCKS */
-void vp9_encode_intra_macro_block(VP9_COMP *cpi,
- MACROBLOCK *x,
- TOKENEXTRA **t,
- int output_enabled) {+static void encode_intra_macro_block(VP9_COMP *cpi, MACROBLOCK *x,
+ TOKENEXTRA **t, int output_enabled) {MB_MODE_INFO * mbmi = &x->e_mbd.mode_info_context->mbmi;
if ((cpi->oxcf.tuning == VP8_TUNE_SSIM) && output_enabled) {adjust_act_zbin(cpi, x);
@@ -1977,18 +1956,13 @@
} else {mbmi->txfm_size = TX_4X4;
}
- }
-#if CONFIG_NEWBESTREFMV
- else
+ } else {vp9_tokenize_mb(cpi, &x->e_mbd, t, 1);
-#endif
+ }
}
-
-extern void vp9_fix_contexts(MACROBLOCKD *xd);
-
-void vp9_encode_inter_macroblock(VP9_COMP *cpi, MACROBLOCK *x,
- TOKENEXTRA **t, int recon_yoffset,
- int recon_uvoffset, int output_enabled) {+static void encode_inter_macroblock(VP9_COMP *cpi, MACROBLOCK *x,
+ TOKENEXTRA **t, int recon_yoffset,
+ int recon_uvoffset, int output_enabled) {VP9_COMMON *cm = &cpi->common;
MACROBLOCKD *const xd = &x->e_mbd;
MB_MODE_INFO * mbmi = &xd->mode_info_context->mbmi;
@@ -2172,9 +2146,10 @@
}
#if CONFIG_SUPERBLOCKS
-void vp9_encode_inter_superblock(VP9_COMP *cpi, MACROBLOCK *x, TOKENEXTRA **t,
- int recon_yoffset, int recon_uvoffset,
- int mb_col, int mb_row) {+static void encode_inter_superblock(VP9_COMP *cpi, MACROBLOCK *x,
+ TOKENEXTRA **t, int recon_yoffset,
+ int recon_uvoffset,
+ int mb_col, int mb_row) {const int output_enabled = 1;
VP9_COMMON *cm = &cpi->common;
MACROBLOCKD *xd = &x->e_mbd;
--- /dev/null
+++ b/vp9/encoder/encodeframe.h
@@ -1,0 +1,21 @@
+/*
+ * Copyright (c) 2010 The WebM project authors. All Rights Reserved.
+ *
+ * Use of this source code is governed by a BSD-style license
+ * that can be found in the LICENSE file in the root of the source
+ * tree. An additional intellectual property rights grant can be found
+ * in the file PATENTS. All contributing project authors may
+ * be found in the AUTHORS file in the root of the source tree.
+ */
+
+
+#ifndef __INC_ENCODEFRAME_H
+#define __INC_ENCODEFRAME_H
+
+struct macroblock;
+
+extern void vp9_build_block_offsets(struct macroblock *x);
+
+extern void vp9_setup_block_ptrs(struct macroblock *x);
+
+#endif // __INC_ENCODEFRAME_H
--- a/vp9/encoder/encodemv.c
+++ b/vp9/encoder/encodemv.c
@@ -168,7 +168,7 @@
if (cur_b - mod_b - cost > 0) {return cur_b - mod_b - cost;
} else {- return -vp9_cost_zero(upd_p);
+ return 0 - vp9_cost_zero(upd_p);
}
}
--- a/vp9/encoder/encodemv.h
+++ b/vp9/encoder/encodemv.h
@@ -18,11 +18,11 @@
void vp9_encode_nmv(vp9_writer* const w, const MV* const mv,
const MV* const ref, const nmv_context* const mvctx);
void vp9_encode_nmv_fp(vp9_writer* const w, const MV* const mv,
- const MV* const ref, const nmv_context *mvctx,
+ const MV* const ref, const nmv_context* const mvctx,
int usehp);
void vp9_build_nmv_cost_table(int *mvjoint,
int *mvcost[2],
- const nmv_context *mvctx,
+ const nmv_context* const mvctx,
int usehp,
int mvc_flag_v,
int mvc_flag_h);
--- a/vp9/encoder/firstpass.c
+++ b/vp9/encoder/firstpass.c
@@ -18,6 +18,7 @@
#include "mcomp.h"
#include "firstpass.h"
#include "vpx_scale/vpxscale.h"
+#include "vp9/encoder/encodeframe.h"
#include "encodemb.h"
#include "vp9/common/extend.h"
#include "vp9/common/systemdependent.h"
@@ -24,6 +25,7 @@
#include "vpx_mem/vpx_mem.h"
#include "vp9/common/swapyv12buffer.h"
#include <stdio.h>
+#include "vp9/encoder/quantize.h"
#include "rdopt.h"
#include "ratectrl.h"
#include "vp9/common/quant_common.h"
@@ -38,17 +40,6 @@
#define IF_RTCD(x) NULL
#endif
-extern void vp9_build_block_offsets(MACROBLOCK *x);
-
-extern void vp9_setup_block_ptrs(MACROBLOCK *x);
-
-extern void vp9_frame_init_quantizer(VP9_COMP *cpi);
-
-extern void vp9_set_mbmode_and_mvs(MACROBLOCK *x, MB_PREDICTION_MODE mb,
- int_mv *mv);
-
-extern void vp9_alloc_compressor_data(VP9_COMP *cpi);
-
#define IIFACTOR 12.5
#define IIKFACTOR1 12.5
#define IIKFACTOR2 15.0
@@ -705,9 +696,9 @@
FIRSTPASS_STATS fps;
fps.frame = cm->current_video_frame;
- fps.intra_error = intra_error >> 8;
- fps.coded_error = coded_error >> 8;
- fps.sr_coded_error = sr_coded_error >> 8;
+ fps.intra_error = (double)(intra_error >> 8);
+ fps.coded_error = (double)(coded_error >> 8);
+ fps.sr_coded_error = (double)(sr_coded_error >> 8);
weight = simple_weight(cpi->Source);
@@ -747,8 +738,8 @@
// TODO: handle the case when duration is set to 0, or something less
// than the full time between subsequent cpi->source_time_stamp s .
- fps.duration = cpi->source->ts_end
- - cpi->source->ts_start;
+ 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,
@@ -910,7 +901,7 @@
double err_per_mb = section_err / num_mbs;
double err_correction_factor;
double speed_correction = 1.0;
- int overhead_bits_per_mb;
+ double overhead_bits_per_mb;
if (section_target_bandwitdh <= 0)
return cpi->twopass.maxq_max_limit; // Highest value allowed
@@ -985,7 +976,7 @@
err_correction_factor = 5.0;
bits_per_mb_at_this_q =
- vp9_bits_per_mb(INTER_FRAME, Q) + overhead_bits_per_mb;
+ vp9_bits_per_mb(INTER_FRAME, Q) + (int)overhead_bits_per_mb;
bits_per_mb_at_this_q = (int)(.5 + err_correction_factor *
(double)bits_per_mb_at_this_q);
@@ -1012,7 +1003,7 @@
// Give average a chance to settle though.
// PGW TODO.. This code is broken for the extended Q range
if ((cpi->ni_frames >
- ((unsigned int)cpi->twopass.total_stats->count >> 8)) &&
+ ((int)cpi->twopass.total_stats->count >> 8)) &&
(cpi->ni_frames > 150)) {adjust_maxq_qrange(cpi);
}
@@ -1038,7 +1029,7 @@
double speed_correction = 1.0;
double clip_iiratio;
double clip_iifactor;
- int overhead_bits_per_mb;
+ double overhead_bits_per_mb;
target_norm_bits_per_mb = (section_target_bandwitdh < (1 << 20))
@@ -1091,7 +1082,7 @@
err_correction_factor = 5.0;
bits_per_mb_at_this_q =
- vp9_bits_per_mb(INTER_FRAME, Q) + overhead_bits_per_mb;
+ vp9_bits_per_mb(INTER_FRAME, Q) + (int)overhead_bits_per_mb;
bits_per_mb_at_this_q = (int)(.5 + err_correction_factor *
(double)bits_per_mb_at_this_q);
@@ -1426,7 +1417,7 @@
calc_frame_boost(cpi, &this_frame, this_frame_mv_in_out));
}
- *f_boost = boost_score;
+ *f_boost = (int)boost_score;
// Reset for backward looking loop
boost_score = 0.0;
@@ -1464,7 +1455,7 @@
calc_frame_boost(cpi, &this_frame, this_frame_mv_in_out));
}
- *b_boost = boost_score;
+ *b_boost = (int)boost_score;
arf_boost = (*f_boost + *b_boost);
if (arf_boost < ((b_frames + f_frames) * 20))
@@ -1486,8 +1477,8 @@
// 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 = cpi->twopass.total_stats->count -
- this_frame->frame - 1;
+ frames_after_arf = (int)(cpi->twopass.total_stats->count -
+ this_frame->frame - 1);
switch (cpi->oxcf.arnr_type) {case 1: // Backward filter
@@ -1710,7 +1701,7 @@
(cpi->twopass.kf_group_error_left > 0)) {cpi->twopass.gf_group_bits =
(int)((double)cpi->twopass.kf_group_bits *
- (gf_group_err / (double)cpi->twopass.kf_group_error_left));
+ (gf_group_err / cpi->twopass.kf_group_error_left));
} else
cpi->twopass.gf_group_bits = 0;
@@ -1776,7 +1767,7 @@
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);
+ DOUBLE_DIVIDE_CHECK(cpi->twopass.kf_group_error_left);
alt_gf_bits = (int)((double)boost * (alt_gf_grp_bits /
(double)allocation_chunks));
@@ -1792,7 +1783,7 @@
int alt_gf_bits =
(int)((double)cpi->twopass.kf_group_bits *
mod_frame_err /
- DOUBLE_DIVIDE_CHECK((double)cpi->twopass.kf_group_error_left));
+ DOUBLE_DIVIDE_CHECK(cpi->twopass.kf_group_error_left));
if (alt_gf_bits > gf_bits) {gf_bits = alt_gf_bits;
@@ -1815,7 +1806,7 @@
{// Adjust KF group bits and error remainin
- cpi->twopass.kf_group_error_left -= gf_group_err;
+ 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)
@@ -1826,9 +1817,10 @@
// 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 = gf_group_err - gf_first_frame_err;
+ cpi->twopass.gf_group_error_left = (int64_t)(gf_group_err
+ - gf_first_frame_err);
else
- cpi->twopass.gf_group_error_left = gf_group_err;
+ cpi->twopass.gf_group_error_left = (int64_t)gf_group_err;
cpi->twopass.gf_group_bits -= cpi->twopass.gf_bits - cpi->min_frame_bandwidth;
@@ -1848,8 +1840,8 @@
pct_extra = (boost - 100) / 50;
pct_extra = (pct_extra > 20) ? 20 : pct_extra;
- cpi->twopass.alt_extra_bits =
- (cpi->twopass.gf_group_bits * pct_extra) / 100;
+ 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);
@@ -1872,9 +1864,9 @@
avg_stats(§ionstats);
- cpi->twopass.section_intra_rating =
- sectionstats.intra_error /
- DOUBLE_DIVIDE_CHECK(sectionstats.coded_error);
+ cpi->twopass.section_intra_rating = (int)
+ (sectionstats.intra_error /
+ DOUBLE_DIVIDE_CHECK(sectionstats.coded_error));
reset_fpf_position(cpi, start_pos);
}
@@ -1907,10 +1899,11 @@
target_frame_size = max_bits;
if (target_frame_size > cpi->twopass.gf_group_bits)
- target_frame_size = cpi->twopass.gf_group_bits;
+ target_frame_size = (int)cpi->twopass.gf_group_bits;
}
- cpi->twopass.gf_group_error_left -= modified_err; // Adjust error remaining
+ // Adjust error remaining
+ cpi->twopass.gf_group_error_left -= (int64_t)modified_err;
cpi->twopass.gf_group_bits -= target_frame_size; // Adjust bits remaining
if (cpi->twopass.gf_group_bits < 0)
@@ -2019,25 +2012,26 @@
}
// Keep a globally available copy of this and the next frame's iiratio.
- cpi->twopass.this_iiratio = this_frame_intra_error /
- DOUBLE_DIVIDE_CHECK(this_frame_coded_error);
+ cpi->twopass.this_iiratio = (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 = next_frame.intra_error /
- DOUBLE_DIVIDE_CHECK(next_frame.coded_error);
+ cpi->twopass.next_iiratio = (int)(next_frame.intra_error /
+ DOUBLE_DIVIDE_CHECK(next_frame.coded_error));
}
}
// Set nominal per second bandwidth for this frame
- cpi->target_bandwidth = cpi->per_frame_bandwidth * cpi->output_frame_rate;
+ cpi->target_bandwidth = (int)(cpi->per_frame_bandwidth
+ * cpi->output_frame_rate);
if (cpi->target_bandwidth < 0)
cpi->target_bandwidth = 0;
// Account for mv, mode and other overheads.
- overhead_bits = estimate_modemvcost(
- cpi, cpi->twopass.total_left_stats);
+ overhead_bits = (int)estimate_modemvcost(
+ cpi, cpi->twopass.total_left_stats);
// Special case code for first frame.
if (cpi->common.current_video_frame == 0) {@@ -2427,9 +2421,9 @@
avg_stats(§ionstats);
- cpi->twopass.section_intra_rating =
- sectionstats.intra_error
- / DOUBLE_DIVIDE_CHECK(sectionstats.coded_error);
+ cpi->twopass.section_intra_rating = (int)
+ (sectionstats.intra_error
+ / DOUBLE_DIVIDE_CHECK(sectionstats.coded_error));
}
// Reset the first pass file position
@@ -2437,7 +2431,7 @@
// Work out how many bits to allocate for the key frame itself
if (1) {- int kf_boost = boost_score;
+ int kf_boost = (int)boost_score;
int allocation_chunks;
int alt_kf_bits;
@@ -2519,10 +2513,14 @@
}
cpi->twopass.kf_group_bits -= cpi->twopass.kf_bits;
- cpi->twopass.kf_bits += cpi->min_frame_bandwidth; // Add in the minimum frame allowance
+ // Add in the minimum frame allowance
+ cpi->twopass.kf_bits += cpi->min_frame_bandwidth;
- cpi->per_frame_bandwidth = cpi->twopass.kf_bits; // Peer frame bit target for this frame
- cpi->target_bandwidth = cpi->twopass.kf_bits * cpi->output_frame_rate; // Convert to a per second bitrate
+ // 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_frame_rate);
}
// Note the total error score of the kf group minus the key frame itself
--- a/vp9/encoder/generic/csystemdependent.c
+++ b/vp9/encoder/generic/csystemdependent.c
@@ -27,10 +27,6 @@
void vp9_cmachine_specific_config(VP9_COMP *cpi) {#if CONFIG_RUNTIME_CPU_DETECT
cpi->rtcd.common = &cpi->common.rtcd;
-
- cpi->rtcd.search.full_search = vp9_full_search_sad;
- cpi->rtcd.search.refining_search = vp9_refining_search_sad;
- cpi->rtcd.search.diamond_search = vp9_diamond_search_sad;
cpi->rtcd.temporal.apply = vp9_temporal_filter_apply_c;
#endif
--- a/vp9/encoder/lookahead.c
+++ b/vp9/encoder/lookahead.c
@@ -43,7 +43,7 @@
vp9_lookahead_destroy(struct lookahead_ctx *ctx) { if (ctx) { if (ctx->buf) {- int i;
+ unsigned int i;
for (i = 0; i < ctx->max_sz; i++)
vp8_yv12_de_alloc_frame_buffer(&ctx->buf[i].img);
@@ -59,7 +59,6 @@
unsigned int height,
unsigned int depth) {struct lookahead_ctx *ctx = NULL;
- int i;
/* Clamp the lookahead queue depth */
if (depth < 1)
@@ -74,6 +73,7 @@
/* Allocate the lookahead structures */
ctx = calloc(1, sizeof(*ctx));
if (ctx) {+ unsigned int i;
ctx->max_sz = depth;
ctx->buf = calloc(depth, sizeof(*ctx->buf));
if (!ctx->buf)
@@ -175,9 +175,9 @@
struct lookahead_entry *buf = NULL;
assert(index < ctx->max_sz);
- if (index < ctx->sz) {+ if (index < (int)ctx->sz) {index += ctx->read_idx;
- if (index >= ctx->max_sz)
+ if (index >= (int)ctx->max_sz)
index -= ctx->max_sz;
buf = ctx->buf + index;
}
--- a/vp9/encoder/mbgraph.c
+++ b/vp9/encoder/mbgraph.c
@@ -139,7 +139,7 @@
// If the current best reference mv is not centred on 0,0 then do a 0,0 based search as well
if (ref_mv->as_int) {- int tmp_err;
+ unsigned int tmp_err;
int_mv zero_ref_mv, tmp_mv;
zero_ref_mv.as_int = 0;
@@ -202,7 +202,7 @@
MACROBLOCK *const x = &cpi->mb;
MACROBLOCKD *const xd = &x->e_mbd;
MB_PREDICTION_MODE best_mode = -1, mode;
- int best_err = INT_MAX;
+ unsigned int best_err = INT_MAX;
// calculate SATD for each intra prediction mode;
// we're intentionally not doing 4x4, we just want a rough estimate
@@ -449,7 +449,7 @@
// being a GF - so exit if we don't look ahead beyond that
if (n_frames <= cpi->frames_till_gf_update_due)
return;
- if (n_frames > cpi->common.frames_till_alt_ref_frame)
+ if (n_frames > (int)cpi->common.frames_till_alt_ref_frame)
n_frames = cpi->common.frames_till_alt_ref_frame;
if (n_frames > MAX_LAG_BUFFERS)
n_frames = MAX_LAG_BUFFERS;
--- a/vp9/encoder/mcomp.c
+++ b/vp9/encoder/mcomp.c
@@ -1255,11 +1255,11 @@
#undef CHECK_POINT
#undef CHECK_BETTER
-int vp9_diamond_search_sad(MACROBLOCK *x, BLOCK *b, BLOCKD *d,
- int_mv *ref_mv, int_mv *best_mv,
- int search_param, int sad_per_bit, int *num00,
- vp9_variance_fn_ptr_t *fn_ptr, DEC_MVCOSTS,
- int_mv *center_mv) {+int vp9_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,
+ vp9_variance_fn_ptr_t *fn_ptr, DEC_MVCOSTS,
+ int_mv *center_mv) {int i, j, step;
unsigned char *what = (*(b->base_src) + b->src);
@@ -1380,7 +1380,7 @@
int tot_steps;
int_mv this_mv;
- int bestsad = INT_MAX;
+ unsigned int bestsad = INT_MAX;
int best_site = 0;
int last_site = 0;
@@ -1568,10 +1568,10 @@
return bestsme;
}
-int vp9_full_search_sad(MACROBLOCK *x, BLOCK *b, BLOCKD *d, int_mv *ref_mv,
- int sad_per_bit, int distance,
- vp9_variance_fn_ptr_t *fn_ptr, DEC_MVCOSTS,
- int_mv *center_mv) {+int vp9_full_search_sad_c(MACROBLOCK *x, BLOCK *b, BLOCKD *d, int_mv *ref_mv,
+ int sad_per_bit, int distance,
+ vp9_variance_fn_ptr_t *fn_ptr, DEC_MVCOSTS,
+ int_mv *center_mv) {unsigned char *what = (*(b->base_src) + b->src);
int what_stride = b->src_stride;
unsigned char *in_what;
@@ -1674,7 +1674,7 @@
unsigned char *bestaddress;
int_mv *best_mv = &d->bmi.as_mv.first;
int_mv this_mv;
- int bestsad = INT_MAX;
+ unsigned int bestsad = INT_MAX;
int r, c;
unsigned char *check_here;
@@ -1802,7 +1802,7 @@
unsigned char *bestaddress;
int_mv *best_mv = &d->bmi.as_mv.first;
int_mv this_mv;
- int bestsad = INT_MAX;
+ unsigned int bestsad = INT_MAX;
int r, c;
unsigned char *check_here;
@@ -1942,11 +1942,10 @@
else
return INT_MAX;
}
-
-int vp9_refining_search_sad(MACROBLOCK *x, BLOCK *b, BLOCKD *d, int_mv *ref_mv,
- int error_per_bit, int search_range,
- vp9_variance_fn_ptr_t *fn_ptr, DEC_MVCOSTS,
- int_mv *center_mv) {+int vp9_refining_search_sad_c(MACROBLOCK *x, BLOCK *b, BLOCKD *d,
+ int_mv *ref_mv, int error_per_bit,
+ int search_range, vp9_variance_fn_ptr_t *fn_ptr,
+ DEC_MVCOSTS, 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;
--- a/vp9/encoder/mcomp.h
+++ b/vp9/encoder/mcomp.h
@@ -68,87 +68,24 @@
extern fractional_mv_step_fp vp9_find_best_sub_pixel_step;
extern fractional_mv_step_fp vp9_find_best_half_pixel_step;
-#define prototype_full_search_sad(sym)\
- int (sym)\
- (\
- MACROBLOCK *x, \
- BLOCK *b, \
- BLOCKD *d, \
- int_mv *ref_mv, \
- int sad_per_bit, \
- int distance, \
- vp9_variance_fn_ptr_t *fn_ptr, \
- DEC_MVSADCOSTS, \
- int_mv *center_mv \
- )
+typedef int (*vp9_full_search_fn_t)(MACROBLOCK *x, BLOCK *b, BLOCKD *d,
+ int_mv *ref_mv, int sad_per_bit,
+ int distance, vp9_variance_fn_ptr_t *fn_ptr,
+ DEC_MVCOSTS, int_mv *center_mv);
-#define prototype_refining_search_sad(sym)\
- int (sym)\
- (\
- MACROBLOCK *x, \
- BLOCK *b, \
- BLOCKD *d, \
- int_mv *ref_mv, \
- int sad_per_bit, \
- int distance, \
- vp9_variance_fn_ptr_t *fn_ptr, \
- DEC_MVSADCOSTS, \
- int_mv *center_mv \
- )
+typedef int (*vp9_refining_search_fn_t)(MACROBLOCK *x, BLOCK *b, BLOCKD *d,
+ int_mv *ref_mv, int sad_per_bit,
+ int distance,
+ vp9_variance_fn_ptr_t *fn_ptr,
+ DEC_MVCOSTS, int_mv *center_mv);
-#define prototype_diamond_search_sad(sym)\
- int (sym)\
- (\
- MACROBLOCK *x, \
- BLOCK *b, \
- BLOCKD *d, \
- int_mv *ref_mv, \
- int_mv *best_mv, \
- int search_param, \
- int sad_per_bit, \
- int *num00, \
- vp9_variance_fn_ptr_t *fn_ptr, \
- DEC_MVSADCOSTS, \
- int_mv *center_mv \
- )
+typedef int (*vp9_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,
+ vp9_variance_fn_ptr_t *fn_ptr,
+ DEC_MVCOSTS, int_mv *center_mv);
-#if ARCH_X86 || ARCH_X86_64
-#include "x86/mcomp_x86.h"
-#endif
-
-typedef prototype_full_search_sad(*vp9_full_search_fn_t);
-extern prototype_full_search_sad(vp9_full_search_sad);
-extern prototype_full_search_sad(vp9_full_search_sadx3);
-extern prototype_full_search_sad(vp9_full_search_sadx8);
-
-typedef prototype_refining_search_sad(*vp9_refining_search_fn_t);
-extern prototype_refining_search_sad(vp9_refining_search_sad);
-extern prototype_refining_search_sad(vp9_refining_search_sadx4);
-
-typedef prototype_diamond_search_sad(*vp9_diamond_search_fn_t);
-extern prototype_diamond_search_sad(vp9_diamond_search_sad);
-extern prototype_diamond_search_sad(vp9_diamond_search_sadx4);
-
-#ifndef vp9_search_full_search
-#define vp9_search_full_search vp9_full_search_sad
-#endif
-extern prototype_full_search_sad(vp9_search_full_search);
-
-#ifndef vp9_search_refining_search
-#define vp9_search_refining_search vp9_refining_search_sad
-#endif
-extern prototype_refining_search_sad(vp9_search_refining_search);
-
-#ifndef vp9_search_diamond_search
-#define vp9_search_diamond_search vp9_diamond_search_sad
-#endif
-extern prototype_diamond_search_sad(vp9_search_diamond_search);
-
-typedef struct {- prototype_full_search_sad(*full_search);
- prototype_refining_search_sad(*refining_search);
- prototype_diamond_search_sad(*diamond_search);
-} vp9_search_rtcd_vtable_t;
#if CONFIG_RUNTIME_CPU_DETECT
#define SEARCH_INVOKE(ctx,fn) (ctx)->fn
--- a/vp9/encoder/onyx_if.c
+++ b/vp9/encoder/onyx_if.c
@@ -36,12 +36,11 @@
#include "vp9/common/pred_common.h"
#include "vp9/encoder/rdopt.h"
#include "bitstream.h"
+#include "vp9/encoder/picklpf.h"
#include "ratectrl.h"
-
-#if CONFIG_NEWBESTREFMV
#include "vp9/common/mvref_common.h"
-#endif
+
#if ARCH_ARM
#include "vpx_ports/arm.h"
#endif
@@ -58,18 +57,8 @@
#define RTCD(x) NULL
#endif
-extern void vp9_pick_filter_level_fast(YV12_BUFFER_CONFIG *sd, VP9_COMP *cpi);
-
-extern void vp9_set_alt_lf_level(VP9_COMP *cpi, int filt_val);
-
-extern void vp9_pick_filter_level(YV12_BUFFER_CONFIG *sd, VP9_COMP *cpi);
-
extern void vp9_cmachine_specific_config(VP9_COMP *cpi);
-extern void vp9_deblock_frame(YV12_BUFFER_CONFIG *source,
- YV12_BUFFER_CONFIG *post,
- int filt_lvl, int low_var_thresh, int flag);
-
extern void print_tree_update_probs();
#if HAVE_ARMV7
@@ -80,10 +69,6 @@
YV12_BUFFER_CONFIG *dst_ybc);
#endif
-int vp9_calc_ss_err(YV12_BUFFER_CONFIG *source, YV12_BUFFER_CONFIG *dest);
-
-extern void vp9_temporal_filter_prepare_c(VP9_COMP *cpi, int distance);
-
static void set_default_lf_deltas(VP9_COMP *cpi);
#define DEFAULT_INTERP_FILTER EIGHTTAP /* SWITCHABLE for better performance */
@@ -257,7 +242,7 @@
skip_prob = 255;
base_skip_false_prob[i][1] = skip_prob;
- skip_prob = t * 0.75;
+ skip_prob = t * 3 / 4;
if (skip_prob < 1)
skip_prob = 1;
else if (skip_prob > 255)
@@ -264,7 +249,7 @@
skip_prob = 255;
base_skip_false_prob[i][2] = skip_prob;
- skip_prob = t * 1.25;
+ skip_prob = t * 5 / 4;
if (skip_prob < 1)
skip_prob = 1;
else if (skip_prob > 255)
@@ -1413,7 +1398,7 @@
int64_t llden = denom;
int64_t llval = val;
- return llval * llnum / llden;
+ return (int)(llval * llnum / llden);
}
@@ -1925,7 +1910,7 @@
vp9_init_first_pass(cpi);
} else if (cpi->pass == 2) {size_t packet_sz = sizeof(FIRSTPASS_STATS);
- int packets = oxcf->two_pass_stats_in.sz / packet_sz;
+ 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;
@@ -1989,9 +1974,9 @@
cpi->fn_ptr[BLOCK_4X4].copymem = vp9_copy32xn;
#endif
- cpi->full_search_sad = SEARCH_INVOKE(&cpi->rtcd.search, full_search);
- cpi->diamond_search_sad = SEARCH_INVOKE(&cpi->rtcd.search, diamond_search);
- cpi->refining_search_sad = SEARCH_INVOKE(&cpi->rtcd.search, refining_search);
+ cpi->full_search_sad = vp9_full_search_sad;
+ cpi->diamond_search_sad = vp9_diamond_search_sad;
+ cpi->refining_search_sad = vp9_refining_search_sad;
// make sure frame 1 is okay
cpi->error_bins[0] = cpi->common.MBs;
@@ -2351,7 +2336,7 @@
for (i = 0; i < 4; i++)
pkt.data.psnr.psnr[i] = vp9_mse2psnr(pkt.data.psnr.samples[i], 255.0,
- pkt.data.psnr.sse[i]);
+ (double)pkt.data.psnr.sse[i]);
vpx_codec_pkt_list_add(cpi->output_pkt_list, &pkt);
}
@@ -2917,7 +2902,9 @@
// pass function that sets the target bandwidth so must set it here
if (cpi->common.refresh_alt_ref_frame) {cpi->per_frame_bandwidth = cpi->twopass.gf_bits; // Per frame bit target for the alt ref frame
- cpi->target_bandwidth = cpi->twopass.gf_bits * cpi->output_frame_rate; // per second target bitrate
+ // per second target bitrate
+ cpi->target_bandwidth = (int)(cpi->twopass.gf_bits *
+ cpi->output_frame_rate);
}
// Default turn off buffer to buffer copying
@@ -4119,7 +4106,7 @@
- cpi->last_time_stamp_seen;
// do a step update if the duration changes by 10%
if (last_duration)
- step = ((this_duration - last_duration) * 10 / last_duration);
+ step = (int)((this_duration - last_duration) * 10 / last_duration);
}
if (this_duration) {@@ -4132,7 +4119,8 @@
* frame rate. If we haven't seen 1 second yet, then average
* over the whole interval seen.
*/
- interval = cpi->source->ts_end - cpi->first_time_stamp_ever;
+ interval = (double)(cpi->source->ts_end
+ - cpi->first_time_stamp_ever);
if (interval > 10000000.0)
interval = 10000000;
@@ -4234,17 +4222,17 @@
int y_samples = orig->y_height * orig->y_width;
int uv_samples = orig->uv_height * orig->uv_width;
int t_samples = y_samples + 2 * uv_samples;
- int64_t sq_error;
+ double sq_error;
- ye = calc_plane_error(orig->y_buffer, orig->y_stride,
+ ye = (double)calc_plane_error(orig->y_buffer, orig->y_stride,
recon->y_buffer, recon->y_stride, orig->y_width,
orig->y_height);
- ue = calc_plane_error(orig->u_buffer, orig->uv_stride,
+ ue = (double)calc_plane_error(orig->u_buffer, orig->uv_stride,
recon->u_buffer, recon->uv_stride, orig->uv_width,
orig->uv_height);
- ve = calc_plane_error(orig->v_buffer, orig->uv_stride,
+ ve = (double)calc_plane_error(orig->v_buffer, orig->uv_stride,
recon->v_buffer, recon->uv_stride, orig->uv_width,
orig->uv_height);
@@ -4265,15 +4253,15 @@
#endif
vp9_clear_system_state();
- ye = calc_plane_error(orig->y_buffer, orig->y_stride,
+ ye = (double)calc_plane_error(orig->y_buffer, orig->y_stride,
pp->y_buffer, pp->y_stride, orig->y_width,
orig->y_height);
- ue = calc_plane_error(orig->u_buffer, orig->uv_stride,
+ ue = (double)calc_plane_error(orig->u_buffer, orig->uv_stride,
pp->u_buffer, pp->uv_stride, orig->uv_width,
orig->uv_height);
- ve = calc_plane_error(orig->v_buffer, orig->uv_stride,
+ ve = (double)calc_plane_error(orig->v_buffer, orig->uv_stride,
pp->v_buffer, pp->uv_stride, orig->uv_width,
orig->uv_height);
--- a/vp9/encoder/onyx_int.h
+++ b/vp9/encoder/onyx_int.h
@@ -349,7 +349,6 @@
typedef struct VP9_ENCODER_RTCD {VP9_COMMON_RTCD *common;
- vp9_search_rtcd_vtable_t search;
vp9_temporal_rtcd_vtable_t temporal;
} VP9_ENCODER_RTCD;
@@ -665,7 +664,8 @@
int maxq_min_limit;
int static_scene_max_gf_interval;
int kf_bits;
- int gf_group_error_left; // Remaining error from uncoded frames in a gf group. Two pass use only
+ // Remaining error from uncoded frames in a gf group. Two pass use only
+ int64_t gf_group_error_left;
// Projected total bits available for a key frame group of frames
int64_t kf_group_bits;
@@ -673,8 +673,10 @@
// Error score of frames still to be coded in kf group
int64_t kf_group_error_left;
- int gf_group_bits; // Projected Bits available for a group of frames including 1 GF or ARF
- int gf_bits; // Bits for the golden frame or ARF - 2 pass only
+ // Projected Bits available for a group of frames including 1 GF or ARF
+ int64_t gf_group_bits;
+ // Bits for the golden frame or ARF - 2 pass only
+ int gf_bits;
int alt_extra_bits;
int sr_update_lag;
@@ -764,10 +766,12 @@
void vp9_activity_masking(VP9_COMP *cpi, MACROBLOCK *x);
-void vp9_tokenize_mb(VP9_COMP *, MACROBLOCKD *, TOKENEXTRA **, int dry_run);
-void vp9_stuff_mb(VP9_COMP *cpi, MACROBLOCKD *xd, TOKENEXTRA **t, int dry_run);
-
void vp9_set_speed_features(VP9_COMP *cpi);
+
+extern int vp9_calc_ss_err(YV12_BUFFER_CONFIG *source,
+ YV12_BUFFER_CONFIG *dest);
+
+extern void vp9_alloc_compressor_data(VP9_COMP *cpi);
#if CONFIG_DEBUG
#define CHECK_MEM_ERROR(lval,expr) do {\--- a/vp9/encoder/picklpf.c
+++ b/vp9/encoder/picklpf.c
@@ -11,6 +11,7 @@
#include "vp9/common/onyxc_int.h"
#include "onyx_int.h"
+#include "vp9/encoder/picklpf.h"
#include "quantize.h"
#include "vpx_mem/vpx_mem.h"
#include "vpx_scale/vpxscale.h"
@@ -20,8 +21,6 @@
#include "vpx_ports/arm.h"
#endif
-extern int vp9_calc_ss_err(YV12_BUFFER_CONFIG *source,
- YV12_BUFFER_CONFIG *dest);
#if HAVE_ARMV7
extern void vp8_yv12_copy_frame_yonly_no_extend_frame_borders_neon(YV12_BUFFER_CONFIG *src_ybc, YV12_BUFFER_CONFIG *dst_ybc);
#endif
--- /dev/null
+++ b/vp9/encoder/picklpf.h
@@ -1,0 +1,26 @@
+/*
+ * Copyright (c) 2010 The WebM project authors. All Rights Reserved.
+ *
+ * Use of this source code is governed by a BSD-style license
+ * that can be found in the LICENSE file in the root of the source
+ * tree. An additional intellectual property rights grant can be found
+ * in the file PATENTS. All contributing project authors may
+ * be found in the AUTHORS file in the root of the source tree.
+ */
+
+
+#ifndef __INC_PICKLPF_H
+#define __INC_PICKLPF_H
+
+struct yv12_buffer_config;
+struct VP9_COMP;
+
+extern void vp9_pick_filter_level_fast(struct yv12_buffer_config *sd,
+ struct VP9_COMP *cpi);
+
+extern void vp9_set_alt_lf_level(struct VP9_COMP *cpi, int filt_val);
+
+extern void vp9_pick_filter_level(struct yv12_buffer_config *sd,
+ struct VP9_COMP *cpi);
+
+#endif // __INC_PICKLPF_H
--- a/vp9/encoder/ratectrl.c
+++ b/vp9/encoder/ratectrl.c
@@ -311,8 +311,8 @@
target = cpi->per_frame_bandwidth;
if (cpi->oxcf.rc_max_intra_bitrate_pct) {- unsigned int max_rate = cpi->per_frame_bandwidth
- * cpi->oxcf.rc_max_intra_bitrate_pct / 100;
+ int max_rate = cpi->per_frame_bandwidth
+ * cpi->oxcf.rc_max_intra_bitrate_pct / 100;
if (target > max_rate)
target = max_rate;
--- a/vp9/encoder/rdopt.c
+++ b/vp9/encoder/rdopt.c
@@ -41,9 +41,7 @@
#include "vp9/common/pred_common.h"
#include "vp9/common/entropy.h"
#include "vpx_rtcd.h"
-#if CONFIG_NEWBESTREFMV
#include "vp9/common/mvref_common.h"
-#endif
#if CONFIG_RUNTIME_CPU_DETECT
#define IF_RTCD(x) (x)
@@ -51,9 +49,6 @@
#define IF_RTCD(x) NULL
#endif
-extern void vp9_mb_init_quantizer(VP9_COMP *cpi, MACROBLOCK *x);
-extern void vp9_update_zbin_extra(VP9_COMP *cpi, MACROBLOCK *x);
-
#define MAXF(a,b) (((a) > (b)) ? (a) : (b))
#define INVALID_MV 0x80008000
@@ -391,59 +386,6 @@
}
}
-void vp9_auto_select_speed(VP9_COMP *cpi) {- int milliseconds_for_compress = (int)(1000000 / cpi->oxcf.frame_rate);
-
- milliseconds_for_compress = milliseconds_for_compress * (16 - cpi->oxcf.cpu_used) / 16;
-
- /*
- // this is done during parameter valid check
- if( cpi->oxcf.cpu_used > 16)
- cpi->oxcf.cpu_used = 16;
- if( cpi->oxcf.cpu_used < -16)
- cpi->oxcf.cpu_used = -16;
- */
-
- 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;
- }
- }
-
- 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) { // if ( cpi->Speed < 0 )- cpi->Speed = 4; // cpi->Speed = 0;
- }
- }
- }
- } else {- cpi->Speed += 4;
-
- if (cpi->Speed > 16)
- cpi->Speed = 16;
-
-
- cpi->avg_pick_mode_time = 0;
- cpi->avg_encode_time = 0;
- }
-}
-
int vp9_block_error_c(short *coeff, short *dqcoeff, int block_size) {int i, error = 0;
@@ -613,7 +555,7 @@
default_eob = 64;
if (type == PLANE_TYPE_Y_WITH_DC) {BLOCKD *bb;
- int ib = (b - xd->block);
+ int ib = (int)(b - xd->block);
if (ib < 16) {ib = (ib & 8) + ((ib & 4) >> 1);
bb = xd->block + ib;
@@ -707,9 +649,6 @@
BLOCK *beptr;
int d;
- vp9_subtract_mby(mb->src_diff, *(mb->block[0].base_src), xd->predictor,
- mb->block[0].src_stride);
-
// Fdct and building the 2nd order block
for (beptr = mb->block; beptr < mb->block + 16; beptr += 2) {mb->vp9_short_fdct8x4(beptr->src_diff, beptr->coeff, 32);
@@ -778,9 +717,6 @@
BLOCKD *const x_y2 = xd->block + 24;
int d;
- vp9_subtract_mby(mb->src_diff, *(mb->block[0].base_src), xd->predictor,
- mb->block[0].src_stride);
-
vp9_transform_mby_8x8(mb);
vp9_quantize_mby_8x8(mb);
@@ -827,9 +763,6 @@
BLOCK *be = &mb->block[0];
TX_TYPE tx_type;
- vp9_subtract_mby(mb->src_diff, *(mb->block[0].base_src), mb->e_mbd.predictor,
- mb->block[0].src_stride);
-
tx_type = get_tx_type_16x16(xd, b);
if (tx_type != DCT_DCT) {vp9_fht(be->src_diff, 32, be->coeff, tx_type, 16);
@@ -866,7 +799,9 @@
int d16x16, r16x16, r16x16s, s16x16;
int64_t rd16x16, rd16x16s;
- // FIXME don't do sub x3
+ vp9_subtract_mby(x->src_diff, *(x->block[0].base_src), xd->predictor,
+ x->block[0].src_stride);
+
if (skip_prob == 0)
skip_prob = 1;
s0 = vp9_cost_bit(skip_prob, 0);
@@ -1243,7 +1178,7 @@
cost += vp9_cost_bit(128, allow_comp);
#endif
*Rate = cost;
- *rate_y += tot_rate_y;
+ *rate_y = tot_rate_y;
*Distortion = distortion;
return RDCOST(mb->rdmult, mb->rddiv, cost, distortion);
@@ -1579,7 +1514,7 @@
#endif
}
*Rate = cost;
- *rate_y += tot_rate_y;
+ *rate_y = tot_rate_y;
*Distortion = distortion;
return RDCOST(mb->rdmult, mb->rddiv, cost, distortion);
}
@@ -2767,7 +2702,7 @@
if (mbmi->second_ref_frame)
x->partition_info->bmi[15].second_mv.as_int = bsi.second_mvs[15].as_int;
- return bsi.segment_rd;
+ return (int)(bsi.segment_rd);
}
/* Order arr in increasing order, original position stored in idx */
@@ -3251,7 +3186,6 @@
MACROBLOCKD *xd = &x->e_mbd;
MB_MODE_INFO * mbmi = &xd->mode_info_context->mbmi;
-
vp9_find_near_mvs(xd, xd->mode_info_context,
xd->prev_mode_info_context,
&frame_nearest_mv[frame_type], &frame_near_mv[frame_type],
@@ -3262,7 +3196,6 @@
u_buffer[frame_type] = yv12->u_buffer + recon_uvoffset;
v_buffer[frame_type] = yv12->v_buffer + recon_uvoffset;
-#if CONFIG_NEWBESTREFMV
vp9_find_mv_refs(xd, xd->mode_info_context,
xd->prev_mode_info_context,
frame_type,
@@ -3275,7 +3208,6 @@
&frame_best_ref_mv[frame_type],
&frame_nearest_mv[frame_type],
&frame_near_mv[frame_type]);
-#endif
}
static int64_t handle_inter_mode(VP9_COMP *cpi, MACROBLOCK *x,
@@ -3453,7 +3385,7 @@
#endif
}
- if (sse < threshold) {+ if ((int)sse < threshold) {unsigned int q2dc = xd->block[24].dequant[0];
/* If there is no codeable 2nd order dc
or a very small uniform pixel change change */
@@ -3527,10 +3459,10 @@
return this_rd; // if 0, this will be re-calculated by caller
}
-void vp9_rd_pick_inter_mode(VP9_COMP *cpi, MACROBLOCK *x,
- int recon_yoffset, int recon_uvoffset,
- int *returnrate, int *returndistortion,
- int64_t *returnintra) {+static void rd_pick_inter_mode(VP9_COMP *cpi, MACROBLOCK *x,
+ int recon_yoffset, int recon_uvoffset,
+ int *returnrate, int *returndistortion,
+ int64_t *returnintra) {VP9_COMMON *cm = &cpi->common;
MACROBLOCKD *xd = &x->e_mbd;
union b_mode_info best_bmodes[16];
@@ -3936,7 +3868,7 @@
second_ref, best_yrd, mdcounts,
&rate, &rate_y, &distortion,
&skippable,
- this_rd_thresh, seg_mvs,
+ (int)this_rd_thresh, seg_mvs,
txfm_cache);
rate2 += rate;
distortion2 += distortion;
@@ -4153,7 +4085,9 @@
for (i = 0; i < NB_TXFM_MODES; i++) {int64_t adj_rd;
if (this_mode != B_PRED) {- adj_rd = this_rd + txfm_cache[i] - txfm_cache[cm->txfm_mode];
+ const int64_t txfm_mode_diff =
+ txfm_cache[i] - txfm_cache[cm->txfm_mode];
+ adj_rd = this_rd + txfm_mode_diff;
} else {adj_rd = this_rd;
}
@@ -4268,11 +4202,12 @@
}
end:
- store_coding_context(x, &x->mb_context[xd->mb_index], best_mode_index, &best_partition,
- &frame_best_ref_mv[xd->mode_info_context->mbmi.ref_frame],
- &frame_best_ref_mv[xd->mode_info_context->mbmi.second_ref_frame],
- best_pred_diff[0], best_pred_diff[1], best_pred_diff[2],
- best_txfm_diff);
+ store_coding_context(x, &x->mb_context[xd->mb_index],
+ best_mode_index, &best_partition,
+ &frame_best_ref_mv[xd->mode_info_context->mbmi.ref_frame],
+ &frame_best_ref_mv[xd->mode_info_context->mbmi.second_ref_frame],
+ (int)best_pred_diff[0], (int)best_pred_diff[1], (int)best_pred_diff[2],
+ best_txfm_diff);
}
#if CONFIG_SUPERBLOCKS
@@ -4834,8 +4769,8 @@
{int zbin_mode_boost_enabled = cpi->zbin_mode_boost_enabled;
- vp9_rd_pick_inter_mode(cpi, x, recon_yoffset, recon_uvoffset, &rate,
- &distortion, &intra_error);
+ rd_pick_inter_mode(cpi, x, recon_yoffset, recon_uvoffset, &rate,
+ &distortion, &intra_error);
/* restore cpi->zbin_mode_boost_enabled */
cpi->zbin_mode_boost_enabled = zbin_mode_boost_enabled;
--- a/vp9/encoder/rdopt.h
+++ b/vp9/encoder/rdopt.h
@@ -17,10 +17,7 @@
extern void vp9_initialize_rd_consts(VP9_COMP *cpi, int Qvalue);
-extern void vp9_rd_pick_inter_mode(VP9_COMP *cpi, MACROBLOCK *x,
- int recon_yoffset, int recon_uvoffset,
- int *returnrate, int *returndistortion,
- int64_t *returnintra);
+extern void vp9_initialize_me_consts(VP9_COMP *cpi, int QIndex);
extern void vp9_rd_pick_intra_mode(VP9_COMP *cpi, MACROBLOCK *x,
int *r, int *d);
@@ -27,6 +24,14 @@
extern void vp9_rd_pick_intra_mode_sb(VP9_COMP *cpi, MACROBLOCK *x,
int *r, int *d);
+
+extern void vp9_pick_mode_inter_macroblock(VP9_COMP *cpi, MACROBLOCK *x,
+ int recon_yoffset,
+ int recon_uvoffset, int *r, int *d);
+
+extern int64_t vp9_rd_pick_inter_mode_sb(VP9_COMP *cpi, MACROBLOCK *x,
+ int recon_yoffset, int recon_uvoffset,
+ int *returnrate, int *returndist);
extern void vp9_mv_pred(VP9_COMP *cpi, MACROBLOCKD *xd,
const MODE_INFO *here, int_mv *mvp,
--- a/vp9/encoder/temporal_filter.h
+++ b/vp9/encoder/temporal_filter.h
@@ -44,4 +44,8 @@
#define TEMPORAL_INVOKE(ctx,fn) vp9_temporal_filter_##fn
#endif
+struct VP9_COMP;
+
+extern void vp9_temporal_filter_prepare_c(struct VP9_COMP *cpi, int distance);
+
#endif // __INC_TEMPORAL_FILTER_H
--- a/vp9/encoder/tokenize.c
+++ b/vp9/encoder/tokenize.c
@@ -48,9 +48,6 @@
[PREV_COEF_CONTEXTS][ENTROPY_NODES] [2];
#endif /* ENTROPY_STATS */
-void vp9_stuff_mb(VP9_COMP *cpi, MACROBLOCKD *xd, TOKENEXTRA **t, int dry_run);
-void vp9_fix_contexts(MACROBLOCKD *xd);
-
static TOKENVALUE dct_value_tokens[DCT_MAX_VALUE * 2];
const TOKENVALUE *vp9_dct_value_tokens_ptr;
static int dct_value_cost[DCT_MAX_VALUE * 2];
--- a/vp9/encoder/tokenize.h
+++ b/vp9/encoder/tokenize.h
@@ -29,13 +29,21 @@
unsigned char skip_eob_node;
} TOKENEXTRA;
-int rd_cost_mby(MACROBLOCKD *);
-
extern int vp9_mby_is_skippable_4x4(MACROBLOCKD *xd, int has_y2_block);
extern int vp9_mbuv_is_skippable_4x4(MACROBLOCKD *xd);
extern int vp9_mby_is_skippable_8x8(MACROBLOCKD *xd, int has_y2_block);
extern int vp9_mbuv_is_skippable_8x8(MACROBLOCKD *xd);
extern int vp9_mby_is_skippable_16x16(MACROBLOCKD *xd);
+
+struct VP9_COMP;
+
+extern void vp9_tokenize_mb(struct VP9_COMP *cpi, MACROBLOCKD *xd,
+ TOKENEXTRA **t, int dry_run);
+
+extern void vp9_stuff_mb(struct VP9_COMP *cpi, MACROBLOCKD *xd,
+ TOKENEXTRA **t, int dry_run);
+
+extern void vp9_fix_contexts(MACROBLOCKD *xd);
#ifdef ENTROPY_STATS
void init_context_counters();
--- a/vp9/encoder/variance_c.c
+++ b/vp9/encoder/variance_c.c
@@ -11,6 +11,7 @@
#include "variance.h"
#include "vp9/common/filter.h"
+#include "vp9/common/subpelvar.h"
unsigned int vp9_get_mb_ss_c(const short *src_ptr) {@@ -24,32 +25,7 @@
}
-static void variance(const unsigned char *src_ptr,
- int source_stride,
- const unsigned char *ref_ptr,
- int recon_stride,
- int w,
- int h,
- unsigned int *sse,
- int *sum) {- int i, j;
- int diff;
- *sum = 0;
- *sse = 0;
-
- for (i = 0; i < h; i++) {- for (j = 0; j < w; j++) {- diff = src_ptr[j] - ref_ptr[j];
- *sum += diff;
- *sse += diff * diff;
- }
-
- src_ptr += source_stride;
- ref_ptr += recon_stride;
- }
-}
-
#if CONFIG_SUPERBLOCKS
unsigned int vp9_variance32x32_c(const unsigned char *src_ptr,
int source_stride,
@@ -146,113 +122,6 @@
}
-/****************************************************************************
- *
- * ROUTINE : filter_block2d_bil_first_pass
- *
- * INPUTS : UINT8 *src_ptr : Pointer to source block.
- * UINT32 src_pixels_per_line : Stride of input block.
- * UINT32 pixel_step : Offset between filter input samples (see notes).
- * UINT32 output_height : Input block height.
- * UINT32 output_width : Input block width.
- * INT32 *vp9_filter : Array of 2 bi-linear filter taps.
- *
- * OUTPUTS : INT32 *output_ptr : Pointer to filtered block.
- *
- * RETURNS : void
- *
- * FUNCTION : Applies a 1-D 2-tap bi-linear filter to the source block in
- * either horizontal or vertical direction to produce the
- * filtered output block. Used to implement first-pass
- * of 2-D separable filter.
- *
- * SPECIAL NOTES : Produces INT32 output to retain precision for next pass.
- * Two filter taps should sum to VP9_FILTER_WEIGHT.
- * pixel_step defines whether the filter is applied
- * horizontally (pixel_step=1) or vertically (pixel_step=stride).
- * It defines the offset required to move from one input
- * to the next.
- *
- ****************************************************************************/
-static void var_filter_block2d_bil_first_pass(const unsigned char *src_ptr,
- unsigned short *output_ptr,
- unsigned int src_pixels_per_line,
- int pixel_step,
- unsigned int output_height,
- unsigned int output_width,
- const short *vp9_filter) {- unsigned int i, j;
-
- for (i = 0; i < output_height; i++) {- for (j = 0; j < output_width; j++) {- // Apply bilinear filter
- output_ptr[j] = (((int)src_ptr[0] * vp9_filter[0]) +
- ((int)src_ptr[pixel_step] * vp9_filter[1]) +
- (VP9_FILTER_WEIGHT / 2)) >> VP9_FILTER_SHIFT;
- src_ptr++;
- }
-
- // Next row...
- src_ptr += src_pixels_per_line - output_width;
- output_ptr += output_width;
- }
-}
-
-/****************************************************************************
- *
- * ROUTINE : filter_block2d_bil_second_pass
- *
- * INPUTS : INT32 *src_ptr : Pointer to source block.
- * UINT32 src_pixels_per_line : Stride of input block.
- * UINT32 pixel_step : Offset between filter input samples (see notes).
- * UINT32 output_height : Input block height.
- * UINT32 output_width : Input block width.
- * INT32 *vp9_filter : Array of 2 bi-linear filter taps.
- *
- * OUTPUTS : UINT16 *output_ptr : Pointer to filtered block.
- *
- * RETURNS : void
- *
- * FUNCTION : Applies a 1-D 2-tap bi-linear filter to the source block in
- * either horizontal or vertical direction to produce the
- * filtered output 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.
- * Two filter taps should sum to VP9_FILTER_WEIGHT.
- * pixel_step defines whether the filter is applied
- * horizontally (pixel_step=1) or vertically (pixel_step=stride).
- * It defines the offset required to move from one input
- * to the next.
- *
- ****************************************************************************/
-static void var_filter_block2d_bil_second_pass(const unsigned short *src_ptr,
- unsigned char *output_ptr,
- unsigned int src_pixels_per_line,
- unsigned int pixel_step,
- unsigned int output_height,
- unsigned int output_width,
- const short *vp9_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[0] * vp9_filter[0]) +
- ((int)src_ptr[pixel_step] * vp9_filter[1]) +
- (VP9_FILTER_WEIGHT / 2);
- output_ptr[j] = (unsigned int)(Temp >> VP9_FILTER_SHIFT);
- src_ptr++;
- }
-
- // Next row...
- src_ptr += src_pixels_per_line - output_width;
- output_ptr += output_width;
- }
-}
-
-
unsigned int vp9_sub_pixel_variance4x4_c(const unsigned char *src_ptr,
int src_pixels_per_line,
int xoffset,
@@ -469,72 +338,3 @@
return vp9_variance8x16_c(temp2, 8, dst_ptr, dst_pixels_per_line, sse);
}
-#if CONFIG_NEWBESTREFMV
-unsigned int vp9_variance2x16_c(const unsigned char *src_ptr,
- const int source_stride,
- const unsigned char *ref_ptr,
- const int recon_stride,
- unsigned int *sse) {- unsigned int var;
- int avg;
-
- variance(src_ptr, source_stride, ref_ptr, recon_stride, 2, 16, &var, &avg);
- *sse = var;
- return (var - ((avg * avg) >> 5));
-}
-
-unsigned int vp9_variance16x2_c(const unsigned char *src_ptr,
- const int source_stride,
- const unsigned char *ref_ptr,
- const int recon_stride,
- unsigned int *sse) {- unsigned int var;
- int avg;
-
- variance(src_ptr, source_stride, ref_ptr, recon_stride, 16, 2, &var, &avg);
- *sse = var;
- return (var - ((avg * avg) >> 5));
-}
-
-unsigned int vp9_sub_pixel_variance16x2_c(const unsigned char *src_ptr,
- const int src_pixels_per_line,
- const int xoffset,
- const int yoffset,
- const unsigned char *dst_ptr,
- const int dst_pixels_per_line,
- unsigned int *sse) {- unsigned short FData3[16 * 3]; // Temp data bufffer used in filtering
- unsigned char temp2[20 * 16];
- const short *HFilter, *VFilter;
-
- HFilter = vp9_bilinear_filters[xoffset];
- VFilter = vp9_bilinear_filters[yoffset];
-
- var_filter_block2d_bil_first_pass(src_ptr, FData3,
- src_pixels_per_line, 1, 3, 16, HFilter);
- var_filter_block2d_bil_second_pass(FData3, temp2, 16, 16, 2, 16, VFilter);
-
- return vp9_variance16x2_c(temp2, 16, dst_ptr, dst_pixels_per_line, sse);
-}
-
-unsigned int vp9_sub_pixel_variance2x16_c(const unsigned char *src_ptr,
- const int src_pixels_per_line,
- const int xoffset,
- const int yoffset,
- const unsigned char *dst_ptr,
- const int dst_pixels_per_line,
- unsigned int *sse) {- unsigned short FData3[2 * 17]; // Temp data bufffer used in filtering
- unsigned char temp2[2 * 16];
- const short *HFilter, *VFilter;
-
- HFilter = vp9_bilinear_filters[xoffset];
- VFilter = vp9_bilinear_filters[yoffset];
-
- var_filter_block2d_bil_first_pass(src_ptr, FData3,
- src_pixels_per_line, 1, 17, 2, HFilter);
- var_filter_block2d_bil_second_pass(FData3, temp2, 2, 2, 16, 2, VFilter);
-
- return vp9_variance2x16_c(temp2, 2, dst_ptr, dst_pixels_per_line, sse);
-}
-#endif
--- a/vp9/encoder/x86/x86_csystemdependent.c
+++ b/vp9/encoder/x86/x86_csystemdependent.c
@@ -95,20 +95,6 @@
}
#endif
-#if HAVE_SSE3
- if (flags & HAS_SSE3) {- cpi->rtcd.search.full_search = vp9_full_search_sadx3;
- cpi->rtcd.search.diamond_search = vp9_diamond_search_sadx4;
- cpi->rtcd.search.refining_search = vp9_refining_search_sadx4;
- }
-#endif
-
-
-#if HAVE_SSE4_1
- if (flags & HAS_SSE4_1) {- cpi->rtcd.search.full_search = vp9_full_search_sadx8;
- }
-#endif
#endif
}
--- a/vp9/vp9_common.mk
+++ b/vp9/vp9_common.mk
@@ -53,6 +53,7 @@
VP9_COMMON_SRCS-yes += common/rtcd.c
VP9_COMMON_SRCS-yes += common/rtcd_defs.sh
VP9_COMMON_SRCS-yes += common/sadmxn.h
+VP9_COMMON_SRCS-yes += common/subpelvar.h
VP9_COMMON_SRCS-yes += common/seg_common.h
VP9_COMMON_SRCS-yes += common/seg_common.c
VP9_COMMON_SRCS-yes += common/setupintrarecon.h
--- a/vp9/vp9_cx_iface.c
+++ b/vp9/vp9_cx_iface.c
@@ -179,7 +179,7 @@
if (cfg->g_pass == VPX_RC_LAST_PASS) {size_t packet_sz = sizeof(FIRSTPASS_STATS);
- int n_packets = cfg->rc_twopass_stats_in.sz / packet_sz;
+ int n_packets = (int)(cfg->rc_twopass_stats_in.sz / packet_sz);
FIRSTPASS_STATS *stats;
if (!cfg->rc_twopass_stats_in.buf)
@@ -698,9 +698,9 @@
pkt.data.frame.pts =
(dst_time_stamp * ctx->cfg.g_timebase.den + round)
/ ctx->cfg.g_timebase.num / 10000000;
- pkt.data.frame.duration =
- (delta * 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)
--- a/vpx_scale/vpx_scale.mk
+++ b/vpx_scale/vpx_scale.mk
@@ -5,7 +5,6 @@
SCALE_SRCS-yes += generic/vpxscale.c
SCALE_SRCS-yes += generic/yv12config.c
SCALE_SRCS-yes += generic/yv12extend.c
-SCALE_SRCS-yes += generic/yv12extend_generic.h
SCALE_SRCS-$(CONFIG_SPATIAL_RESAMPLING) += generic/gen_scalers.c
#neon
--
⑨