2005-06-02 20:45:35 +00:00
|
|
|
/*
|
|
|
|
* Copyright (c) 2004-2005 Michael Niedermayer, Loren Merritt
|
|
|
|
*
|
2006-10-07 15:30:46 +00:00
|
|
|
* This file is part of FFmpeg.
|
|
|
|
*
|
|
|
|
* FFmpeg is free software; you can redistribute it and/or
|
2005-06-02 20:45:35 +00:00
|
|
|
* modify it under the terms of the GNU Lesser General Public
|
|
|
|
* License as published by the Free Software Foundation; either
|
2006-10-07 15:30:46 +00:00
|
|
|
* version 2.1 of the License, or (at your option) any later version.
|
2005-06-02 20:45:35 +00:00
|
|
|
*
|
2006-10-07 15:30:46 +00:00
|
|
|
* FFmpeg is distributed in the hope that it will be useful,
|
2005-06-02 20:45:35 +00:00
|
|
|
* but WITHOUT ANY WARRANTY; without even the implied warranty of
|
|
|
|
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
|
|
|
|
* Lesser General Public License for more details.
|
|
|
|
*
|
|
|
|
* You should have received a copy of the GNU Lesser General Public
|
2006-10-07 15:30:46 +00:00
|
|
|
* License along with FFmpeg; if not, write to the Free Software
|
2006-01-12 22:43:26 +00:00
|
|
|
* Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
|
2005-06-02 20:45:35 +00:00
|
|
|
*/
|
|
|
|
|
2010-09-08 15:07:14 +00:00
|
|
|
#include "libavutil/cpu.h"
|
2010-09-01 20:48:59 +00:00
|
|
|
#include "libavutil/x86_cpu.h"
|
|
|
|
#include "libavcodec/h264dsp.h"
|
2007-11-27 22:23:34 +00:00
|
|
|
#include "dsputil_mmx.h"
|
2005-06-02 20:45:35 +00:00
|
|
|
|
2010-03-06 14:24:59 +00:00
|
|
|
DECLARE_ALIGNED(8, static const uint64_t, ff_pb_3_1 ) = 0x0103010301030103ULL;
|
|
|
|
DECLARE_ALIGNED(8, static const uint64_t, ff_pb_7_3 ) = 0x0307030703070307ULL;
|
2008-07-19 21:53:54 +00:00
|
|
|
|
2005-06-02 20:45:35 +00:00
|
|
|
/***********************************/
|
|
|
|
/* IDCT */
|
|
|
|
|
2010-09-14 13:36:26 +00:00
|
|
|
void ff_h264_idct_add_mmx (uint8_t *dst, int16_t *block, int stride);
|
|
|
|
void ff_h264_idct8_add_mmx (uint8_t *dst, int16_t *block, int stride);
|
|
|
|
void ff_h264_idct8_add_sse2 (uint8_t *dst, int16_t *block, int stride);
|
|
|
|
void ff_h264_idct_dc_add_mmx2 (uint8_t *dst, int16_t *block, int stride);
|
|
|
|
void ff_h264_idct8_dc_add_mmx2(uint8_t *dst, int16_t *block, int stride);
|
|
|
|
|
|
|
|
void ff_h264_idct_add16_mmx (uint8_t *dst, const int *block_offset,
|
|
|
|
DCTELEM *block, int stride, const uint8_t nnzc[6*8]);
|
|
|
|
void ff_h264_idct8_add4_mmx (uint8_t *dst, const int *block_offset,
|
|
|
|
DCTELEM *block, int stride, const uint8_t nnzc[6*8]);
|
|
|
|
void ff_h264_idct_add16_mmx2 (uint8_t *dst, const int *block_offset,
|
|
|
|
DCTELEM *block, int stride, const uint8_t nnzc[6*8]);
|
|
|
|
void ff_h264_idct_add16intra_mmx (uint8_t *dst, const int *block_offset,
|
|
|
|
DCTELEM *block, int stride, const uint8_t nnzc[6*8]);
|
|
|
|
void ff_h264_idct_add16intra_mmx2(uint8_t *dst, const int *block_offset,
|
|
|
|
DCTELEM *block, int stride, const uint8_t nnzc[6*8]);
|
|
|
|
void ff_h264_idct8_add4_mmx2 (uint8_t *dst, const int *block_offset,
|
|
|
|
DCTELEM *block, int stride, const uint8_t nnzc[6*8]);
|
|
|
|
void ff_h264_idct8_add4_sse2 (uint8_t *dst, const int *block_offset,
|
|
|
|
DCTELEM *block, int stride, const uint8_t nnzc[6*8]);
|
|
|
|
void ff_h264_idct_add8_mmx (uint8_t **dest, const int *block_offset,
|
|
|
|
DCTELEM *block, int stride, const uint8_t nnzc[6*8]);
|
|
|
|
void ff_h264_idct_add8_mmx2 (uint8_t **dest, const int *block_offset,
|
|
|
|
DCTELEM *block, int stride, const uint8_t nnzc[6*8]);
|
|
|
|
|
|
|
|
void ff_h264_idct_add16_sse2 (uint8_t *dst, const int *block_offset, DCTELEM *block,
|
|
|
|
int stride, const uint8_t nnzc[6*8]);
|
|
|
|
void ff_h264_idct_add16intra_sse2(uint8_t *dst, const int *block_offset, DCTELEM *block,
|
|
|
|
int stride, const uint8_t nnzc[6*8]);
|
|
|
|
void ff_h264_idct_add8_sse2 (uint8_t **dest, const int *block_offset, DCTELEM *block,
|
|
|
|
int stride, const uint8_t nnzc[6*8]);
|
2009-01-03 00:46:17 +00:00
|
|
|
|
2005-06-02 20:45:35 +00:00
|
|
|
/***********************************/
|
|
|
|
/* deblocking */
|
|
|
|
|
2006-08-28 09:33:01 +00:00
|
|
|
static void h264_loop_filter_strength_mmx2( int16_t bS[2][4][4], uint8_t nnz[40], int8_t ref[2][40], int16_t mv[2][40][2],
|
2008-07-19 21:53:54 +00:00
|
|
|
int bidir, int edges, int step, int mask_mv0, int mask_mv1, int field ) {
|
2006-08-28 09:33:01 +00:00
|
|
|
int dir;
|
2008-10-16 13:34:09 +00:00
|
|
|
__asm__ volatile(
|
2010-01-26 17:17:48 +00:00
|
|
|
"movq %0, %%mm7 \n"
|
|
|
|
"movq %1, %%mm6 \n"
|
|
|
|
::"m"(ff_pb_1), "m"(ff_pb_3)
|
2006-08-28 09:33:01 +00:00
|
|
|
);
|
2008-07-19 21:53:54 +00:00
|
|
|
if(field)
|
2008-10-16 13:34:09 +00:00
|
|
|
__asm__ volatile(
|
2010-01-26 17:17:48 +00:00
|
|
|
"movq %0, %%mm6 \n"
|
|
|
|
::"m"(ff_pb_3_1)
|
2008-07-19 21:53:54 +00:00
|
|
|
);
|
2010-01-26 17:17:48 +00:00
|
|
|
__asm__ volatile(
|
|
|
|
"movq %%mm6, %%mm5 \n"
|
|
|
|
"paddb %%mm5, %%mm5 \n"
|
|
|
|
:);
|
2008-07-19 21:53:54 +00:00
|
|
|
|
2006-08-28 09:33:01 +00:00
|
|
|
// could do a special case for dir==0 && edges==1, but it only reduces the
|
|
|
|
// average filter time by 1.2%
|
|
|
|
for( dir=1; dir>=0; dir-- ) {
|
2010-01-20 19:23:19 +00:00
|
|
|
const x86_reg d_idx = dir ? -8 : -1;
|
2006-08-28 09:33:01 +00:00
|
|
|
const int mask_mv = dir ? mask_mv1 : mask_mv0;
|
2010-03-06 14:24:59 +00:00
|
|
|
DECLARE_ALIGNED(8, const uint64_t, mask_dir) = dir ? 0 : 0xffffffffffffffffULL;
|
2010-01-26 17:17:48 +00:00
|
|
|
int b_idx, edge;
|
2006-08-28 09:33:01 +00:00
|
|
|
for( b_idx=12, edge=0; edge<edges; edge+=step, b_idx+=8*step ) {
|
2008-10-16 13:34:09 +00:00
|
|
|
__asm__ volatile(
|
2006-08-28 09:33:01 +00:00
|
|
|
"pand %0, %%mm0 \n\t"
|
|
|
|
::"m"(mask_dir)
|
|
|
|
);
|
|
|
|
if(!(mask_mv & edge)) {
|
2010-01-26 17:17:48 +00:00
|
|
|
if(bidir) {
|
2008-10-16 13:34:09 +00:00
|
|
|
__asm__ volatile(
|
2010-01-26 17:17:48 +00:00
|
|
|
"movd (%1,%0), %%mm2 \n"
|
|
|
|
"punpckldq 40(%1,%0), %%mm2 \n" // { ref0[bn], ref1[bn] }
|
|
|
|
"pshufw $0x44, (%1), %%mm0 \n" // { ref0[b], ref0[b] }
|
|
|
|
"pshufw $0x44, 40(%1), %%mm1 \n" // { ref1[b], ref1[b] }
|
|
|
|
"pshufw $0x4E, %%mm2, %%mm3 \n"
|
|
|
|
"psubb %%mm2, %%mm0 \n" // { ref0[b]!=ref0[bn], ref0[b]!=ref1[bn] }
|
|
|
|
"psubb %%mm3, %%mm1 \n" // { ref1[b]!=ref1[bn], ref1[b]!=ref0[bn] }
|
|
|
|
"1: \n"
|
|
|
|
"por %%mm1, %%mm0 \n"
|
|
|
|
"movq (%2,%0,4), %%mm1 \n"
|
|
|
|
"movq 8(%2,%0,4), %%mm2 \n"
|
|
|
|
"movq %%mm1, %%mm3 \n"
|
|
|
|
"movq %%mm2, %%mm4 \n"
|
|
|
|
"psubw (%2), %%mm1 \n"
|
|
|
|
"psubw 8(%2), %%mm2 \n"
|
|
|
|
"psubw 160(%2), %%mm3 \n"
|
|
|
|
"psubw 168(%2), %%mm4 \n"
|
|
|
|
"packsswb %%mm2, %%mm1 \n"
|
|
|
|
"packsswb %%mm4, %%mm3 \n"
|
|
|
|
"paddb %%mm6, %%mm1 \n"
|
|
|
|
"paddb %%mm6, %%mm3 \n"
|
|
|
|
"psubusb %%mm5, %%mm1 \n" // abs(mv[b] - mv[bn]) >= limit
|
|
|
|
"psubusb %%mm5, %%mm3 \n"
|
|
|
|
"packsswb %%mm3, %%mm1 \n"
|
|
|
|
"add $40, %0 \n"
|
|
|
|
"cmp $40, %0 \n"
|
|
|
|
"jl 1b \n"
|
|
|
|
"sub $80, %0 \n"
|
|
|
|
"pshufw $0x4E, %%mm1, %%mm1 \n"
|
|
|
|
"por %%mm1, %%mm0 \n"
|
|
|
|
"pshufw $0x4E, %%mm0, %%mm1 \n"
|
|
|
|
"pminub %%mm1, %%mm0 \n"
|
|
|
|
::"r"(d_idx),
|
|
|
|
"r"(ref[0]+b_idx),
|
|
|
|
"r"(mv[0]+b_idx)
|
2006-08-28 09:33:01 +00:00
|
|
|
);
|
2010-01-26 17:17:48 +00:00
|
|
|
} else {
|
2010-01-19 16:40:36 +00:00
|
|
|
__asm__ volatile(
|
2010-01-26 17:17:48 +00:00
|
|
|
"movd (%1), %%mm0 \n"
|
|
|
|
"psubb (%1,%0), %%mm0 \n" // ref[b] != ref[bn]
|
|
|
|
"movq (%2), %%mm1 \n"
|
|
|
|
"movq 8(%2), %%mm2 \n"
|
|
|
|
"psubw (%2,%0,4), %%mm1 \n"
|
|
|
|
"psubw 8(%2,%0,4), %%mm2 \n"
|
|
|
|
"packsswb %%mm2, %%mm1 \n"
|
|
|
|
"paddb %%mm6, %%mm1 \n"
|
|
|
|
"psubusb %%mm5, %%mm1 \n" // abs(mv[b] - mv[bn]) >= limit
|
|
|
|
"packsswb %%mm1, %%mm1 \n"
|
|
|
|
"por %%mm1, %%mm0 \n"
|
|
|
|
::"r"(d_idx),
|
|
|
|
"r"(ref[0]+b_idx),
|
|
|
|
"r"(mv[0]+b_idx)
|
2010-01-19 16:40:36 +00:00
|
|
|
);
|
|
|
|
}
|
2006-08-28 09:33:01 +00:00
|
|
|
}
|
2008-10-16 13:34:09 +00:00
|
|
|
__asm__ volatile(
|
2010-01-26 17:17:48 +00:00
|
|
|
"movd %0, %%mm1 \n"
|
|
|
|
"por %1, %%mm1 \n" // nnz[b] || nnz[bn]
|
2006-08-28 09:33:01 +00:00
|
|
|
::"m"(nnz[b_idx]),
|
|
|
|
"m"(nnz[b_idx+d_idx])
|
|
|
|
);
|
2008-10-16 13:34:09 +00:00
|
|
|
__asm__ volatile(
|
2010-01-26 17:17:48 +00:00
|
|
|
"pminub %%mm7, %%mm1 \n"
|
|
|
|
"pminub %%mm7, %%mm0 \n"
|
|
|
|
"psllw $1, %%mm1 \n"
|
|
|
|
"pxor %%mm2, %%mm2 \n"
|
|
|
|
"pmaxub %%mm0, %%mm1 \n"
|
|
|
|
"punpcklbw %%mm2, %%mm1 \n"
|
|
|
|
"movq %%mm1, %0 \n"
|
2006-08-28 09:33:01 +00:00
|
|
|
:"=m"(*bS[dir][edge])
|
|
|
|
::"memory"
|
|
|
|
);
|
|
|
|
}
|
|
|
|
edges = 4;
|
|
|
|
step = 1;
|
|
|
|
}
|
2008-10-16 13:34:09 +00:00
|
|
|
__asm__ volatile(
|
2006-08-28 09:33:01 +00:00
|
|
|
"movq (%0), %%mm0 \n\t"
|
|
|
|
"movq 8(%0), %%mm1 \n\t"
|
|
|
|
"movq 16(%0), %%mm2 \n\t"
|
|
|
|
"movq 24(%0), %%mm3 \n\t"
|
|
|
|
TRANSPOSE4(%%mm0, %%mm1, %%mm2, %%mm3, %%mm4)
|
|
|
|
"movq %%mm0, (%0) \n\t"
|
|
|
|
"movq %%mm3, 8(%0) \n\t"
|
|
|
|
"movq %%mm4, 16(%0) \n\t"
|
|
|
|
"movq %%mm2, 24(%0) \n\t"
|
|
|
|
::"r"(bS[0])
|
|
|
|
:"memory"
|
|
|
|
);
|
|
|
|
}
|
2005-06-02 20:45:35 +00:00
|
|
|
|
2010-09-03 16:52:46 +00:00
|
|
|
#define LF_FUNC(DIR, TYPE, OPT) \
|
|
|
|
void ff_x264_deblock_ ## DIR ## _ ## TYPE ## _ ## OPT (uint8_t *pix, int stride, \
|
|
|
|
int alpha, int beta, int8_t *tc0);
|
|
|
|
#define LF_IFUNC(DIR, TYPE, OPT) \
|
|
|
|
void ff_x264_deblock_ ## DIR ## _ ## TYPE ## _ ## OPT (uint8_t *pix, int stride, \
|
|
|
|
int alpha, int beta);
|
|
|
|
|
|
|
|
LF_FUNC (h, chroma, mmxext)
|
|
|
|
LF_IFUNC(h, chroma_intra, mmxext)
|
|
|
|
LF_FUNC (v, chroma, mmxext)
|
|
|
|
LF_IFUNC(v, chroma_intra, mmxext)
|
|
|
|
|
|
|
|
LF_FUNC (h, luma, mmxext)
|
|
|
|
LF_IFUNC(h, luma_intra, mmxext)
|
|
|
|
#if HAVE_YASM && ARCH_X86_32
|
|
|
|
LF_FUNC (v8, luma, mmxext)
|
|
|
|
static void ff_x264_deblock_v_luma_mmxext(uint8_t *pix, int stride, int alpha, int beta, int8_t *tc0)
|
|
|
|
{
|
|
|
|
if((tc0[0] & tc0[1]) >= 0)
|
|
|
|
ff_x264_deblock_v8_luma_mmxext(pix+0, stride, alpha, beta, tc0);
|
|
|
|
if((tc0[2] & tc0[3]) >= 0)
|
|
|
|
ff_x264_deblock_v8_luma_mmxext(pix+8, stride, alpha, beta, tc0+2);
|
|
|
|
}
|
|
|
|
LF_IFUNC(v8, luma_intra, mmxext)
|
|
|
|
static void ff_x264_deblock_v_luma_intra_mmxext(uint8_t *pix, int stride, int alpha, int beta)
|
|
|
|
{
|
|
|
|
ff_x264_deblock_v8_luma_intra_mmxext(pix+0, stride, alpha, beta);
|
|
|
|
ff_x264_deblock_v8_luma_intra_mmxext(pix+8, stride, alpha, beta);
|
|
|
|
}
|
|
|
|
#endif
|
|
|
|
|
|
|
|
LF_FUNC (h, luma, sse2)
|
|
|
|
LF_IFUNC(h, luma_intra, sse2)
|
|
|
|
LF_FUNC (v, luma, sse2)
|
|
|
|
LF_IFUNC(v, luma_intra, sse2)
|
|
|
|
|
2005-10-09 23:38:52 +00:00
|
|
|
/***********************************/
|
|
|
|
/* weighted prediction */
|
|
|
|
|
2010-09-01 20:56:16 +00:00
|
|
|
#define H264_WEIGHT(W, H, OPT) \
|
|
|
|
void ff_h264_weight_ ## W ## x ## H ## _ ## OPT(uint8_t *dst, \
|
|
|
|
int stride, int log2_denom, int weight, int offset);
|
|
|
|
|
|
|
|
#define H264_BIWEIGHT(W, H, OPT) \
|
|
|
|
void ff_h264_biweight_ ## W ## x ## H ## _ ## OPT(uint8_t *dst, \
|
|
|
|
uint8_t *src, int stride, int log2_denom, int weightd, \
|
|
|
|
int weights, int offset);
|
|
|
|
|
|
|
|
#define H264_BIWEIGHT_MMX(W,H) \
|
|
|
|
H264_WEIGHT (W, H, mmx2) \
|
|
|
|
H264_BIWEIGHT(W, H, mmx2)
|
|
|
|
|
|
|
|
#define H264_BIWEIGHT_MMX_SSE(W,H) \
|
|
|
|
H264_BIWEIGHT_MMX(W, H) \
|
|
|
|
H264_WEIGHT (W, H, sse2) \
|
|
|
|
H264_BIWEIGHT (W, H, sse2) \
|
|
|
|
H264_BIWEIGHT (W, H, ssse3)
|
|
|
|
|
|
|
|
H264_BIWEIGHT_MMX_SSE(16, 16)
|
|
|
|
H264_BIWEIGHT_MMX_SSE(16, 8)
|
|
|
|
H264_BIWEIGHT_MMX_SSE( 8, 16)
|
|
|
|
H264_BIWEIGHT_MMX_SSE( 8, 8)
|
|
|
|
H264_BIWEIGHT_MMX_SSE( 8, 4)
|
|
|
|
H264_BIWEIGHT_MMX ( 4, 8)
|
|
|
|
H264_BIWEIGHT_MMX ( 4, 4)
|
|
|
|
H264_BIWEIGHT_MMX ( 4, 2)
|
2010-09-01 20:48:59 +00:00
|
|
|
|
|
|
|
void ff_h264dsp_init_x86(H264DSPContext *c)
|
|
|
|
{
|
2010-09-08 15:07:14 +00:00
|
|
|
int mm_flags = av_get_cpu_flags();
|
2010-09-01 20:48:59 +00:00
|
|
|
|
2010-09-14 13:36:26 +00:00
|
|
|
if (mm_flags & AV_CPU_FLAG_MMX2) {
|
|
|
|
c->h264_loop_filter_strength= h264_loop_filter_strength_mmx2;
|
|
|
|
}
|
|
|
|
#if HAVE_YASM
|
2010-09-04 09:59:08 +00:00
|
|
|
if (mm_flags & AV_CPU_FLAG_MMX) {
|
2010-09-01 20:48:59 +00:00
|
|
|
c->h264_idct_dc_add=
|
|
|
|
c->h264_idct_add= ff_h264_idct_add_mmx;
|
|
|
|
c->h264_idct8_dc_add=
|
|
|
|
c->h264_idct8_add= ff_h264_idct8_add_mmx;
|
|
|
|
|
|
|
|
c->h264_idct_add16 = ff_h264_idct_add16_mmx;
|
|
|
|
c->h264_idct8_add4 = ff_h264_idct8_add4_mmx;
|
|
|
|
c->h264_idct_add8 = ff_h264_idct_add8_mmx;
|
|
|
|
c->h264_idct_add16intra= ff_h264_idct_add16intra_mmx;
|
|
|
|
|
2010-09-04 09:59:08 +00:00
|
|
|
if (mm_flags & AV_CPU_FLAG_MMX2) {
|
2010-09-01 20:48:59 +00:00
|
|
|
c->h264_idct_dc_add= ff_h264_idct_dc_add_mmx2;
|
|
|
|
c->h264_idct8_dc_add= ff_h264_idct8_dc_add_mmx2;
|
|
|
|
c->h264_idct_add16 = ff_h264_idct_add16_mmx2;
|
|
|
|
c->h264_idct8_add4 = ff_h264_idct8_add4_mmx2;
|
|
|
|
c->h264_idct_add8 = ff_h264_idct_add8_mmx2;
|
|
|
|
c->h264_idct_add16intra= ff_h264_idct_add16intra_mmx2;
|
|
|
|
|
2010-09-03 16:52:46 +00:00
|
|
|
c->h264_v_loop_filter_chroma= ff_x264_deblock_v_chroma_mmxext;
|
|
|
|
c->h264_h_loop_filter_chroma= ff_x264_deblock_h_chroma_mmxext;
|
|
|
|
c->h264_v_loop_filter_chroma_intra= ff_x264_deblock_v_chroma_intra_mmxext;
|
|
|
|
c->h264_h_loop_filter_chroma_intra= ff_x264_deblock_h_chroma_intra_mmxext;
|
2010-09-01 20:56:16 +00:00
|
|
|
#if ARCH_X86_32
|
2010-09-03 16:52:46 +00:00
|
|
|
c->h264_v_loop_filter_luma= ff_x264_deblock_v_luma_mmxext;
|
|
|
|
c->h264_h_loop_filter_luma= ff_x264_deblock_h_luma_mmxext;
|
2010-09-01 20:56:16 +00:00
|
|
|
c->h264_v_loop_filter_luma_intra = ff_x264_deblock_v_luma_intra_mmxext;
|
|
|
|
c->h264_h_loop_filter_luma_intra = ff_x264_deblock_h_luma_intra_mmxext;
|
|
|
|
#endif
|
2010-09-01 20:48:59 +00:00
|
|
|
c->weight_h264_pixels_tab[0]= ff_h264_weight_16x16_mmx2;
|
|
|
|
c->weight_h264_pixels_tab[1]= ff_h264_weight_16x8_mmx2;
|
|
|
|
c->weight_h264_pixels_tab[2]= ff_h264_weight_8x16_mmx2;
|
|
|
|
c->weight_h264_pixels_tab[3]= ff_h264_weight_8x8_mmx2;
|
|
|
|
c->weight_h264_pixels_tab[4]= ff_h264_weight_8x4_mmx2;
|
|
|
|
c->weight_h264_pixels_tab[5]= ff_h264_weight_4x8_mmx2;
|
|
|
|
c->weight_h264_pixels_tab[6]= ff_h264_weight_4x4_mmx2;
|
|
|
|
c->weight_h264_pixels_tab[7]= ff_h264_weight_4x2_mmx2;
|
|
|
|
|
|
|
|
c->biweight_h264_pixels_tab[0]= ff_h264_biweight_16x16_mmx2;
|
|
|
|
c->biweight_h264_pixels_tab[1]= ff_h264_biweight_16x8_mmx2;
|
|
|
|
c->biweight_h264_pixels_tab[2]= ff_h264_biweight_8x16_mmx2;
|
|
|
|
c->biweight_h264_pixels_tab[3]= ff_h264_biweight_8x8_mmx2;
|
|
|
|
c->biweight_h264_pixels_tab[4]= ff_h264_biweight_8x4_mmx2;
|
|
|
|
c->biweight_h264_pixels_tab[5]= ff_h264_biweight_4x8_mmx2;
|
|
|
|
c->biweight_h264_pixels_tab[6]= ff_h264_biweight_4x4_mmx2;
|
|
|
|
c->biweight_h264_pixels_tab[7]= ff_h264_biweight_4x2_mmx2;
|
|
|
|
|
2010-09-04 09:59:08 +00:00
|
|
|
if (mm_flags&AV_CPU_FLAG_SSE2) {
|
2010-09-14 13:36:26 +00:00
|
|
|
c->h264_idct8_add = ff_h264_idct8_add_sse2;
|
|
|
|
c->h264_idct8_add4= ff_h264_idct8_add4_sse2;
|
|
|
|
|
2010-09-01 20:56:16 +00:00
|
|
|
c->weight_h264_pixels_tab[0]= ff_h264_weight_16x16_sse2;
|
|
|
|
c->weight_h264_pixels_tab[1]= ff_h264_weight_16x8_sse2;
|
|
|
|
c->weight_h264_pixels_tab[2]= ff_h264_weight_8x16_sse2;
|
|
|
|
c->weight_h264_pixels_tab[3]= ff_h264_weight_8x8_sse2;
|
|
|
|
c->weight_h264_pixels_tab[4]= ff_h264_weight_8x4_sse2;
|
|
|
|
|
2010-09-01 20:48:59 +00:00
|
|
|
c->biweight_h264_pixels_tab[0]= ff_h264_biweight_16x16_sse2;
|
2010-09-01 20:56:16 +00:00
|
|
|
c->biweight_h264_pixels_tab[1]= ff_h264_biweight_16x8_sse2;
|
|
|
|
c->biweight_h264_pixels_tab[2]= ff_h264_biweight_8x16_sse2;
|
2010-09-01 20:48:59 +00:00
|
|
|
c->biweight_h264_pixels_tab[3]= ff_h264_biweight_8x8_sse2;
|
2010-09-01 20:56:16 +00:00
|
|
|
c->biweight_h264_pixels_tab[4]= ff_h264_biweight_8x4_sse2;
|
|
|
|
|
2010-09-01 20:48:59 +00:00
|
|
|
c->h264_v_loop_filter_luma = ff_x264_deblock_v_luma_sse2;
|
|
|
|
c->h264_h_loop_filter_luma = ff_x264_deblock_h_luma_sse2;
|
|
|
|
c->h264_v_loop_filter_luma_intra = ff_x264_deblock_v_luma_intra_sse2;
|
|
|
|
c->h264_h_loop_filter_luma_intra = ff_x264_deblock_h_luma_intra_sse2;
|
2010-09-18 20:44:32 +00:00
|
|
|
|
2010-09-01 20:48:59 +00:00
|
|
|
c->h264_idct_add16 = ff_h264_idct_add16_sse2;
|
|
|
|
c->h264_idct_add8 = ff_h264_idct_add8_sse2;
|
|
|
|
c->h264_idct_add16intra = ff_h264_idct_add16intra_sse2;
|
|
|
|
}
|
2010-09-04 09:59:08 +00:00
|
|
|
if (mm_flags&AV_CPU_FLAG_SSSE3) {
|
2010-09-01 20:48:59 +00:00
|
|
|
c->biweight_h264_pixels_tab[0]= ff_h264_biweight_16x16_ssse3;
|
2010-09-01 20:56:16 +00:00
|
|
|
c->biweight_h264_pixels_tab[1]= ff_h264_biweight_16x8_ssse3;
|
|
|
|
c->biweight_h264_pixels_tab[2]= ff_h264_biweight_8x16_ssse3;
|
2010-09-01 20:48:59 +00:00
|
|
|
c->biweight_h264_pixels_tab[3]= ff_h264_biweight_8x8_ssse3;
|
2010-09-01 20:56:16 +00:00
|
|
|
c->biweight_h264_pixels_tab[4]= ff_h264_biweight_8x4_ssse3;
|
2010-09-01 20:48:59 +00:00
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
2010-09-14 13:36:26 +00:00
|
|
|
#endif
|
2010-09-01 20:48:59 +00:00
|
|
|
}
|