vpx/vp9/decoder/idct_blk.c

281 lines
8.8 KiB
C
Raw Normal View History

/*
* 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 "vpx_rtcd.h"
#include "vp9/common/idct.h"
void vp9_dequant_dc_idct_add_y_block_c(short *q, short *dq,
unsigned char *pre,
unsigned char *dst,
int stride, unsigned short *eobs,
short *dc) {
int i, j;
for (i = 0; i < 4; i++) {
for (j = 0; j < 4; j++) {
if (*eobs++ > 1)
vp9_dequant_dc_idct_add_c(q, dq, pre, dst, 16, stride, dc[0]);
else
vp9_dc_only_idct_add_c(dc[0], pre, dst, 16, stride);
q += 16;
pre += 4;
dst += 4;
dc++;
}
pre += 64 - 16;
dst += 4 * stride - 16;
}
}
void vp9_dequant_idct_add_y_block_c(short *q, short *dq,
unsigned char *pre,
unsigned char *dst,
int stride, unsigned short *eobs) {
int i, j;
for (i = 0; i < 4; i++) {
for (j = 0; j < 4; j++) {
if (*eobs++ > 1)
vp9_dequant_idct_add_c(q, dq, pre, dst, 16, stride);
else {
vp9_dc_only_idct_add_c(q[0]*dq[0], pre, dst, 16, stride);
((int *)q)[0] = 0;
}
q += 16;
pre += 4;
dst += 4;
}
pre += 64 - 16;
dst += 4 * stride - 16;
}
}
void vp9_dequant_idct_add_uv_block_c(short *q, short *dq, unsigned char *pre,
unsigned char *dstu, unsigned char *dstv,
int stride, unsigned short *eobs) {
int i, j;
for (i = 0; i < 2; i++) {
for (j = 0; j < 2; j++) {
if (*eobs++ > 1)
vp9_dequant_idct_add_c(q, dq, pre, dstu, 8, stride);
else {
vp9_dc_only_idct_add_c(q[0]*dq[0], pre, dstu, 8, stride);
((int *)q)[0] = 0;
}
q += 16;
pre += 4;
dstu += 4;
}
pre += 32 - 8;
dstu += 4 * stride - 8;
}
for (i = 0; i < 2; i++) {
for (j = 0; j < 2; j++) {
if (*eobs++ > 1)
vp9_dequant_idct_add_c(q, dq, pre, dstv, 8, stride);
else {
vp9_dc_only_idct_add_c(q[0]*dq[0], pre, dstv, 8, stride);
((int *)q)[0] = 0;
}
q += 16;
pre += 4;
dstv += 4;
}
pre += 32 - 8;
dstv += 4 * stride - 8;
}
}
void vp9_dequant_dc_idct_add_y_block_8x8_c(short *q, short *dq,
unsigned char *pre,
unsigned char *dst,
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]);
vp9_dequant_dc_idct_add_8x8_c(&q[128], dq, pre + 8 * 16,
dst + 8 * stride, 16, stride, dc[4]);
vp9_dequant_dc_idct_add_8x8_c(&q[192], dq, pre + 8 * 16 + 8,
dst + 8 * stride + 8, 16, stride, dc[8]);
}
#if CONFIG_SUPERBLOCKS
void vp9_dequant_dc_idct_add_y_block_8x8_inplace_c(short *q, short *dq,
unsigned char *dst,
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,
dst + 8, stride, stride, dc[1]);
vp9_dequant_dc_idct_add_8x8_c(&q[128], dq, dst + 8 * stride,
dst + 8 * stride, stride, stride, dc[4]);
vp9_dequant_dc_idct_add_8x8_c(&q[192], dq, dst + 8 * stride + 8,
dst + 8 * stride + 8, stride, stride, dc[8]);
}
#endif
void vp9_dequant_idct_add_y_block_8x8_c(short *q, short *dq,
unsigned char *pre,
unsigned char *dst,
int stride, unsigned short *eobs,
MACROBLOCKD *xd) {
unsigned char *origdest = dst;
unsigned char *origpred = pre;
vp9_dequant_idct_add_8x8_c(q, dq, pre, dst, 16, stride);
vp9_dequant_idct_add_8x8_c(&q[64], dq, origpred + 8,
origdest + 8, 16, stride);
vp9_dequant_idct_add_8x8_c(&q[128], dq, origpred + 8 * 16,
origdest + 8 * stride, 16, stride);
vp9_dequant_idct_add_8x8_c(&q[192], dq, origpred + 8 * 16 + 8,
origdest + 8 * stride + 8, 16, stride);
}
void vp9_dequant_idct_add_uv_block_8x8_c(short *q, short *dq,
unsigned char *pre,
unsigned char *dstu,
unsigned char *dstv,
int stride, unsigned short *eobs,
MACROBLOCKD *xd) {
vp9_dequant_idct_add_8x8_c(q, dq, pre, dstu, 8, stride);
q += 64;
pre += 64;
vp9_dequant_idct_add_8x8_c(q, dq, pre, dstv, 8, stride);
}
#if CONFIG_SUPERBLOCKS
void vp9_dequant_idct_add_uv_block_8x8_inplace_c(short *q, short *dq,
unsigned char *dstu,
unsigned char *dstv,
int stride,
unsigned short *eobs,
MACROBLOCKD *xd) {
vp9_dequant_idct_add_8x8_c(q, dq, dstu, dstu, stride, stride);
q += 64;
vp9_dequant_idct_add_8x8_c(q, dq, dstv, dstv, stride, stride);
}
#endif
Add lossless compression mode. This commit adds lossless compression capability to the experimental branch. The lossless experiment can be enabled using --enable-lossless in configure. When the experiment is enabled, the encoder will use lossless compression mode by command line option --lossless, and the decoder automatically recognizes a losslessly encoded clip and decodes accordingly. To achieve the lossless coding, this commit has changed the following: 1. To encode at lossless mode, encoder forces the use of unit quantizer, i.e, Q 0, where effective quantization is 1. Encoder also disables the usage of 8x8 transform and allows only 4x4 transform; 2. At Q 0, the first order 4x4 DCT/IDCT have been switched over to a pair of forward and inverse Walsh-Hadamard Transform (http://goo.gl/EIsfy), with proper scaling applied to match the range of the original 4x4 DCT/IDCT pair; 3. At Q 0, the second order remains to use the previous walsh-hadamard transform pair. However, to maintain the reversibility in second order transform at Q 0, scaling down is applied to first order DC coefficients prior to forward transform, and scaling up is applied to the second order output prior to quantization. Symmetric upscaling and downscaling are added around inverse second order transform; 4. At lossless mode, encoder also disables a number of minor features to ensure no loss is introduced, these features includes: a. Trellis quantization optimization b. Loop filtering c. Aggressive zero-binning, rounding and zero-bin boosting d. Mode based zero-bin boosting Lossless coding test was performed on all clips within the derf set, to verify that the commit has achieved lossless compression for all clips. The average compression ratio is around 2.57 to 1. (http://goo.gl/dEShs) Change-Id: Ia3aba7dd09df40dd590f93b9aba134defbc64e34
2012-06-14 04:03:31 +02:00
#if CONFIG_LOSSLESS
void vp9_dequant_dc_idct_add_y_block_lossless_c(short *q, short *dq,
unsigned char *pre,
unsigned char *dst,
int stride,
unsigned short *eobs,
short *dc) {
int i, j;
for (i = 0; i < 4; i++) {
for (j = 0; j < 4; j++) {
if (*eobs++ > 1)
vp9_dequant_dc_idct_add_lossless_c(q, dq, pre, dst, 16, stride, dc[0]);
else
vp9_dc_only_inv_walsh_add_c(dc[0], pre, dst, 16, stride);
q += 16;
pre += 4;
dst += 4;
dc++;
Add lossless compression mode. This commit adds lossless compression capability to the experimental branch. The lossless experiment can be enabled using --enable-lossless in configure. When the experiment is enabled, the encoder will use lossless compression mode by command line option --lossless, and the decoder automatically recognizes a losslessly encoded clip and decodes accordingly. To achieve the lossless coding, this commit has changed the following: 1. To encode at lossless mode, encoder forces the use of unit quantizer, i.e, Q 0, where effective quantization is 1. Encoder also disables the usage of 8x8 transform and allows only 4x4 transform; 2. At Q 0, the first order 4x4 DCT/IDCT have been switched over to a pair of forward and inverse Walsh-Hadamard Transform (http://goo.gl/EIsfy), with proper scaling applied to match the range of the original 4x4 DCT/IDCT pair; 3. At Q 0, the second order remains to use the previous walsh-hadamard transform pair. However, to maintain the reversibility in second order transform at Q 0, scaling down is applied to first order DC coefficients prior to forward transform, and scaling up is applied to the second order output prior to quantization. Symmetric upscaling and downscaling are added around inverse second order transform; 4. At lossless mode, encoder also disables a number of minor features to ensure no loss is introduced, these features includes: a. Trellis quantization optimization b. Loop filtering c. Aggressive zero-binning, rounding and zero-bin boosting d. Mode based zero-bin boosting Lossless coding test was performed on all clips within the derf set, to verify that the commit has achieved lossless compression for all clips. The average compression ratio is around 2.57 to 1. (http://goo.gl/dEShs) Change-Id: Ia3aba7dd09df40dd590f93b9aba134defbc64e34
2012-06-14 04:03:31 +02:00
}
pre += 64 - 16;
dst += 4 * stride - 16;
}
Add lossless compression mode. This commit adds lossless compression capability to the experimental branch. The lossless experiment can be enabled using --enable-lossless in configure. When the experiment is enabled, the encoder will use lossless compression mode by command line option --lossless, and the decoder automatically recognizes a losslessly encoded clip and decodes accordingly. To achieve the lossless coding, this commit has changed the following: 1. To encode at lossless mode, encoder forces the use of unit quantizer, i.e, Q 0, where effective quantization is 1. Encoder also disables the usage of 8x8 transform and allows only 4x4 transform; 2. At Q 0, the first order 4x4 DCT/IDCT have been switched over to a pair of forward and inverse Walsh-Hadamard Transform (http://goo.gl/EIsfy), with proper scaling applied to match the range of the original 4x4 DCT/IDCT pair; 3. At Q 0, the second order remains to use the previous walsh-hadamard transform pair. However, to maintain the reversibility in second order transform at Q 0, scaling down is applied to first order DC coefficients prior to forward transform, and scaling up is applied to the second order output prior to quantization. Symmetric upscaling and downscaling are added around inverse second order transform; 4. At lossless mode, encoder also disables a number of minor features to ensure no loss is introduced, these features includes: a. Trellis quantization optimization b. Loop filtering c. Aggressive zero-binning, rounding and zero-bin boosting d. Mode based zero-bin boosting Lossless coding test was performed on all clips within the derf set, to verify that the commit has achieved lossless compression for all clips. The average compression ratio is around 2.57 to 1. (http://goo.gl/dEShs) Change-Id: Ia3aba7dd09df40dd590f93b9aba134defbc64e34
2012-06-14 04:03:31 +02:00
}
void vp9_dequant_idct_add_y_block_lossless_c(short *q, short *dq,
unsigned char *pre,
unsigned char *dst,
int stride, unsigned short *eobs) {
int i, j;
for (i = 0; i < 4; i++) {
for (j = 0; j < 4; j++) {
if (*eobs++ > 1)
vp9_dequant_idct_add_lossless_c(q, dq, pre, dst, 16, stride);
else {
vp9_dc_only_inv_walsh_add_c(q[0]*dq[0], pre, dst, 16, stride);
((int *)q)[0] = 0;
}
q += 16;
pre += 4;
dst += 4;
Add lossless compression mode. This commit adds lossless compression capability to the experimental branch. The lossless experiment can be enabled using --enable-lossless in configure. When the experiment is enabled, the encoder will use lossless compression mode by command line option --lossless, and the decoder automatically recognizes a losslessly encoded clip and decodes accordingly. To achieve the lossless coding, this commit has changed the following: 1. To encode at lossless mode, encoder forces the use of unit quantizer, i.e, Q 0, where effective quantization is 1. Encoder also disables the usage of 8x8 transform and allows only 4x4 transform; 2. At Q 0, the first order 4x4 DCT/IDCT have been switched over to a pair of forward and inverse Walsh-Hadamard Transform (http://goo.gl/EIsfy), with proper scaling applied to match the range of the original 4x4 DCT/IDCT pair; 3. At Q 0, the second order remains to use the previous walsh-hadamard transform pair. However, to maintain the reversibility in second order transform at Q 0, scaling down is applied to first order DC coefficients prior to forward transform, and scaling up is applied to the second order output prior to quantization. Symmetric upscaling and downscaling are added around inverse second order transform; 4. At lossless mode, encoder also disables a number of minor features to ensure no loss is introduced, these features includes: a. Trellis quantization optimization b. Loop filtering c. Aggressive zero-binning, rounding and zero-bin boosting d. Mode based zero-bin boosting Lossless coding test was performed on all clips within the derf set, to verify that the commit has achieved lossless compression for all clips. The average compression ratio is around 2.57 to 1. (http://goo.gl/dEShs) Change-Id: Ia3aba7dd09df40dd590f93b9aba134defbc64e34
2012-06-14 04:03:31 +02:00
}
pre += 64 - 16;
dst += 4 * stride - 16;
}
Add lossless compression mode. This commit adds lossless compression capability to the experimental branch. The lossless experiment can be enabled using --enable-lossless in configure. When the experiment is enabled, the encoder will use lossless compression mode by command line option --lossless, and the decoder automatically recognizes a losslessly encoded clip and decodes accordingly. To achieve the lossless coding, this commit has changed the following: 1. To encode at lossless mode, encoder forces the use of unit quantizer, i.e, Q 0, where effective quantization is 1. Encoder also disables the usage of 8x8 transform and allows only 4x4 transform; 2. At Q 0, the first order 4x4 DCT/IDCT have been switched over to a pair of forward and inverse Walsh-Hadamard Transform (http://goo.gl/EIsfy), with proper scaling applied to match the range of the original 4x4 DCT/IDCT pair; 3. At Q 0, the second order remains to use the previous walsh-hadamard transform pair. However, to maintain the reversibility in second order transform at Q 0, scaling down is applied to first order DC coefficients prior to forward transform, and scaling up is applied to the second order output prior to quantization. Symmetric upscaling and downscaling are added around inverse second order transform; 4. At lossless mode, encoder also disables a number of minor features to ensure no loss is introduced, these features includes: a. Trellis quantization optimization b. Loop filtering c. Aggressive zero-binning, rounding and zero-bin boosting d. Mode based zero-bin boosting Lossless coding test was performed on all clips within the derf set, to verify that the commit has achieved lossless compression for all clips. The average compression ratio is around 2.57 to 1. (http://goo.gl/dEShs) Change-Id: Ia3aba7dd09df40dd590f93b9aba134defbc64e34
2012-06-14 04:03:31 +02:00
}
void vp9_dequant_idct_add_uv_block_lossless_c(short *q, short *dq,
unsigned char *pre,
unsigned char *dstu,
unsigned char *dstv,
int stride,
unsigned short *eobs) {
int i, j;
for (i = 0; i < 2; i++) {
for (j = 0; j < 2; j++) {
if (*eobs++ > 1)
vp9_dequant_idct_add_lossless_c(q, dq, pre, dstu, 8, stride);
else {
vp9_dc_only_inv_walsh_add_c(q[0]*dq[0], pre, dstu, 8, stride);
((int *)q)[0] = 0;
}
q += 16;
pre += 4;
dstu += 4;
Add lossless compression mode. This commit adds lossless compression capability to the experimental branch. The lossless experiment can be enabled using --enable-lossless in configure. When the experiment is enabled, the encoder will use lossless compression mode by command line option --lossless, and the decoder automatically recognizes a losslessly encoded clip and decodes accordingly. To achieve the lossless coding, this commit has changed the following: 1. To encode at lossless mode, encoder forces the use of unit quantizer, i.e, Q 0, where effective quantization is 1. Encoder also disables the usage of 8x8 transform and allows only 4x4 transform; 2. At Q 0, the first order 4x4 DCT/IDCT have been switched over to a pair of forward and inverse Walsh-Hadamard Transform (http://goo.gl/EIsfy), with proper scaling applied to match the range of the original 4x4 DCT/IDCT pair; 3. At Q 0, the second order remains to use the previous walsh-hadamard transform pair. However, to maintain the reversibility in second order transform at Q 0, scaling down is applied to first order DC coefficients prior to forward transform, and scaling up is applied to the second order output prior to quantization. Symmetric upscaling and downscaling are added around inverse second order transform; 4. At lossless mode, encoder also disables a number of minor features to ensure no loss is introduced, these features includes: a. Trellis quantization optimization b. Loop filtering c. Aggressive zero-binning, rounding and zero-bin boosting d. Mode based zero-bin boosting Lossless coding test was performed on all clips within the derf set, to verify that the commit has achieved lossless compression for all clips. The average compression ratio is around 2.57 to 1. (http://goo.gl/dEShs) Change-Id: Ia3aba7dd09df40dd590f93b9aba134defbc64e34
2012-06-14 04:03:31 +02:00
}
pre += 32 - 8;
dstu += 4 * stride - 8;
}
for (i = 0; i < 2; i++) {
for (j = 0; j < 2; j++) {
if (*eobs++ > 1)
vp9_dequant_idct_add_lossless_c(q, dq, pre, dstv, 8, stride);
else {
vp9_dc_only_inv_walsh_add_c(q[0]*dq[0], pre, dstv, 8, stride);
((int *)q)[0] = 0;
}
q += 16;
pre += 4;
dstv += 4;
Add lossless compression mode. This commit adds lossless compression capability to the experimental branch. The lossless experiment can be enabled using --enable-lossless in configure. When the experiment is enabled, the encoder will use lossless compression mode by command line option --lossless, and the decoder automatically recognizes a losslessly encoded clip and decodes accordingly. To achieve the lossless coding, this commit has changed the following: 1. To encode at lossless mode, encoder forces the use of unit quantizer, i.e, Q 0, where effective quantization is 1. Encoder also disables the usage of 8x8 transform and allows only 4x4 transform; 2. At Q 0, the first order 4x4 DCT/IDCT have been switched over to a pair of forward and inverse Walsh-Hadamard Transform (http://goo.gl/EIsfy), with proper scaling applied to match the range of the original 4x4 DCT/IDCT pair; 3. At Q 0, the second order remains to use the previous walsh-hadamard transform pair. However, to maintain the reversibility in second order transform at Q 0, scaling down is applied to first order DC coefficients prior to forward transform, and scaling up is applied to the second order output prior to quantization. Symmetric upscaling and downscaling are added around inverse second order transform; 4. At lossless mode, encoder also disables a number of minor features to ensure no loss is introduced, these features includes: a. Trellis quantization optimization b. Loop filtering c. Aggressive zero-binning, rounding and zero-bin boosting d. Mode based zero-bin boosting Lossless coding test was performed on all clips within the derf set, to verify that the commit has achieved lossless compression for all clips. The average compression ratio is around 2.57 to 1. (http://goo.gl/dEShs) Change-Id: Ia3aba7dd09df40dd590f93b9aba134defbc64e34
2012-06-14 04:03:31 +02:00
}
pre += 32 - 8;
dstv += 4 * stride - 8;
}
Add lossless compression mode. This commit adds lossless compression capability to the experimental branch. The lossless experiment can be enabled using --enable-lossless in configure. When the experiment is enabled, the encoder will use lossless compression mode by command line option --lossless, and the decoder automatically recognizes a losslessly encoded clip and decodes accordingly. To achieve the lossless coding, this commit has changed the following: 1. To encode at lossless mode, encoder forces the use of unit quantizer, i.e, Q 0, where effective quantization is 1. Encoder also disables the usage of 8x8 transform and allows only 4x4 transform; 2. At Q 0, the first order 4x4 DCT/IDCT have been switched over to a pair of forward and inverse Walsh-Hadamard Transform (http://goo.gl/EIsfy), with proper scaling applied to match the range of the original 4x4 DCT/IDCT pair; 3. At Q 0, the second order remains to use the previous walsh-hadamard transform pair. However, to maintain the reversibility in second order transform at Q 0, scaling down is applied to first order DC coefficients prior to forward transform, and scaling up is applied to the second order output prior to quantization. Symmetric upscaling and downscaling are added around inverse second order transform; 4. At lossless mode, encoder also disables a number of minor features to ensure no loss is introduced, these features includes: a. Trellis quantization optimization b. Loop filtering c. Aggressive zero-binning, rounding and zero-bin boosting d. Mode based zero-bin boosting Lossless coding test was performed on all clips within the derf set, to verify that the commit has achieved lossless compression for all clips. The average compression ratio is around 2.57 to 1. (http://goo.gl/dEShs) Change-Id: Ia3aba7dd09df40dd590f93b9aba134defbc64e34
2012-06-14 04:03:31 +02:00
}
#endif