ed9c66f584
Instead of using the predict buffer, the decoder now writes
the predictor into the recon buffer. For blocks with eob=0,
unnecessary idcts can be eliminated. This gave a performance
boost of ~1.8% for the HD clips used.
Tero: Added needed changes to ARM side and scheduled some
assembly code to prevent interlocks.
Patch Set 6: Merged (I1bcdca7a95aacc3a181b9faa6b10e3a71ee24df3)
into this commit because of similarities in the idct
functions.
Patch Set 7: EC bug fix.
Change-Id: Ie31d90b5d3522e1108163f2ac491e455e3f955e6
128 lines
3.7 KiB
C
128 lines
3.7 KiB
C
/*
|
|
* 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_config.h"
|
|
#include "vp8/common/idct.h"
|
|
#include "vp8/decoder/dequantize.h"
|
|
|
|
void vp8_idct_dequant_dc_0_2x_sse2
|
|
(short *q, short *dq,
|
|
unsigned char *dst, int dst_stride, short *dc);
|
|
void vp8_idct_dequant_dc_full_2x_sse2
|
|
(short *q, short *dq,
|
|
unsigned char *dst, int dst_stride, short *dc);
|
|
|
|
void vp8_idct_dequant_0_2x_sse2
|
|
(short *q, short *dq ,
|
|
unsigned char *dst, int dst_stride);
|
|
void vp8_idct_dequant_full_2x_sse2
|
|
(short *q, short *dq ,
|
|
unsigned char *dst, int dst_stride);
|
|
|
|
void vp8_dequant_dc_idct_add_y_block_sse2
|
|
(short *q, short *dq,
|
|
unsigned char *dst, int stride, char *eobs, short *dc)
|
|
{
|
|
int i;
|
|
|
|
for (i = 0; i < 4; i++)
|
|
{
|
|
if (((short *)(eobs))[0])
|
|
{
|
|
if (((short *)(eobs))[0] & 0xfefe)
|
|
vp8_idct_dequant_dc_full_2x_sse2 (q, dq, dst, stride, dc);
|
|
else
|
|
vp8_idct_dequant_dc_0_2x_sse2 (q, dq, dst, stride, dc);
|
|
}
|
|
|
|
if (((short *)(eobs))[1])
|
|
{
|
|
if (((short *)(eobs))[1] & 0xfefe)
|
|
vp8_idct_dequant_dc_full_2x_sse2 (q+32, dq, dst+8, stride, dc+2);
|
|
else
|
|
vp8_idct_dequant_dc_0_2x_sse2 (q+32, dq, dst+8, stride, dc+2);
|
|
}
|
|
q += 64;
|
|
dc += 4;
|
|
dst += stride*4;
|
|
eobs += 4;
|
|
}
|
|
}
|
|
|
|
void vp8_dequant_idct_add_y_block_sse2
|
|
(short *q, short *dq,
|
|
unsigned char *dst, int stride, char *eobs)
|
|
{
|
|
int i;
|
|
|
|
for (i = 0; i < 4; i++)
|
|
{
|
|
if (((short *)(eobs))[0])
|
|
{
|
|
if (((short *)(eobs))[0] & 0xfefe)
|
|
vp8_idct_dequant_full_2x_sse2 (q, dq, dst, stride);
|
|
else
|
|
vp8_idct_dequant_0_2x_sse2 (q, dq, dst, stride);
|
|
}
|
|
if (((short *)(eobs))[1])
|
|
{
|
|
if (((short *)(eobs))[1] & 0xfefe)
|
|
vp8_idct_dequant_full_2x_sse2 (q+32, dq, dst+8, stride);
|
|
else
|
|
vp8_idct_dequant_0_2x_sse2 (q+32, dq, dst+8, stride);
|
|
}
|
|
q += 64;
|
|
dst += stride*4;
|
|
eobs += 4;
|
|
}
|
|
}
|
|
|
|
void vp8_dequant_idct_add_uv_block_sse2
|
|
(short *q, short *dq,
|
|
unsigned char *dstu, unsigned char *dstv, int stride, char *eobs)
|
|
{
|
|
if (((short *)(eobs))[0])
|
|
{
|
|
if (((short *)(eobs))[0] & 0xfefe)
|
|
vp8_idct_dequant_full_2x_sse2 (q, dq, dstu, stride);
|
|
else
|
|
vp8_idct_dequant_0_2x_sse2 (q, dq, dstu, stride);
|
|
}
|
|
q += 32;
|
|
dstu += stride*4;
|
|
|
|
if (((short *)(eobs))[1])
|
|
{
|
|
if (((short *)(eobs))[1] & 0xfefe)
|
|
vp8_idct_dequant_full_2x_sse2 (q, dq, dstu, stride);
|
|
else
|
|
vp8_idct_dequant_0_2x_sse2 (q, dq, dstu, stride);
|
|
}
|
|
q += 32;
|
|
|
|
if (((short *)(eobs))[2])
|
|
{
|
|
if (((short *)(eobs))[2] & 0xfefe)
|
|
vp8_idct_dequant_full_2x_sse2 (q, dq, dstv, stride);
|
|
else
|
|
vp8_idct_dequant_0_2x_sse2 (q, dq, dstv, stride);
|
|
}
|
|
q += 32;
|
|
dstv += stride*4;
|
|
|
|
if (((short *)(eobs))[3])
|
|
{
|
|
if (((short *)(eobs))[3] & 0xfefe)
|
|
vp8_idct_dequant_full_2x_sse2 (q, dq, dstv, stride);
|
|
else
|
|
vp8_idct_dequant_0_2x_sse2 (q, dq, dstv, stride);
|
|
}
|
|
}
|