mirror of
https://git.videolan.org/git/ffmpeg.git
synced 2024-08-22 09:15:06 +02:00
![Michael Niedermayer](/assets/img/avatar_default.png)
* qatar/master: (35 commits) flvdec: Do not call parse_keyframes_index with a NULL stream libspeexdec: include system headers before local headers libspeexdec: return meaningful error codes libspeexdec: cosmetics: reindent libspeexdec: decode one frame at a time. swscale: fix signed shift overflows in ff_yuv2rgb_c_init_tables() Move timefilter code from lavf to lavd. mov: add support for hdvd and pgapmetadata atoms mov: rename function _stik, some indentation cosmetics mov: rename function _int8 to remove ambiguity, some indentation cosmetics mov: parse the gnre atom mp3on4: check for allocation failures in decode_init_mp3on4() mp3on4: create a separate flush function for MP3onMP4. mp3on4: ensure that the frame channel count does not exceed the codec channel count. mp3on4: set channel layout mp3on4: fix the output channel order mp3on4: allocate temp buffer with av_malloc() instead of on the stack. mp3on4: copy MPADSPContext from first context to all contexts. fmtconvert: port float_to_int16_interleave() 2-channel x86 inline asm to yasm fmtconvert: port int32_to_float_fmul_scalar() x86 inline asm to yasm ... Conflicts: libavcodec/arm/h264dsp_init_arm.c libavcodec/h264.c libavcodec/h264.h libavcodec/h264_cabac.c libavcodec/h264_cavlc.c libavcodec/h264_ps.c libavcodec/h264dsp_template.c libavcodec/h264idct_template.c libavcodec/h264pred.c libavcodec/h264pred_template.c libavcodec/x86/h264dsp_mmx.c libavdevice/Makefile libavdevice/jack_audio.c libavformat/Makefile libavformat/flvdec.c libavformat/flvenc.c libavutil/pixfmt.h libswscale/utils.c Merged-by: Michael Niedermayer <michaelni@gmx.at>
144 lines
5.6 KiB
C
144 lines
5.6 KiB
C
/*
|
|
* Format Conversion Utils
|
|
* Copyright (c) 2000, 2001 Fabrice Bellard
|
|
* Copyright (c) 2002-2004 Michael Niedermayer <michaelni@gmx.at>
|
|
*
|
|
* This file is part of FFmpeg.
|
|
*
|
|
* FFmpeg is free software; you can redistribute it and/or
|
|
* modify it under the terms of the GNU Lesser General Public
|
|
* License as published by the Free Software Foundation; either
|
|
* version 2.1 of the License, or (at your option) any later version.
|
|
*
|
|
* FFmpeg is distributed in the hope that it will be useful,
|
|
* 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
|
|
* License along with FFmpeg; if not, write to the Free Software
|
|
* Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
|
|
*
|
|
* MMX optimization by Nick Kurshev <nickols_k@mail.ru>
|
|
*/
|
|
|
|
#include "libavutil/cpu.h"
|
|
#include "libavutil/x86_cpu.h"
|
|
#include "libavcodec/fmtconvert.h"
|
|
|
|
#if HAVE_YASM
|
|
|
|
void ff_int32_to_float_fmul_scalar_sse (float *dst, const int *src, float mul, int len);
|
|
void ff_int32_to_float_fmul_scalar_sse2(float *dst, const int *src, float mul, int len);
|
|
|
|
void ff_float_to_int16_3dnow(int16_t *dst, const float *src, long len);
|
|
void ff_float_to_int16_sse (int16_t *dst, const float *src, long len);
|
|
void ff_float_to_int16_sse2 (int16_t *dst, const float *src, long len);
|
|
|
|
void ff_float_to_int16_interleave2_3dnow(int16_t *dst, const float **src, long len);
|
|
void ff_float_to_int16_interleave2_sse (int16_t *dst, const float **src, long len);
|
|
void ff_float_to_int16_interleave2_sse2 (int16_t *dst, const float **src, long len);
|
|
|
|
void ff_float_to_int16_interleave6_sse(int16_t *dst, const float **src, int len);
|
|
void ff_float_to_int16_interleave6_3dnow(int16_t *dst, const float **src, int len);
|
|
void ff_float_to_int16_interleave6_3dn2(int16_t *dst, const float **src, int len);
|
|
|
|
#define ff_float_to_int16_interleave6_sse2 ff_float_to_int16_interleave6_sse
|
|
|
|
#define FLOAT_TO_INT16_INTERLEAVE(cpu) \
|
|
/* gcc pessimizes register allocation if this is in the same function as float_to_int16_interleave_sse2*/\
|
|
static av_noinline void float_to_int16_interleave_misc_##cpu(int16_t *dst, const float **src, long len, int channels){\
|
|
DECLARE_ALIGNED(16, int16_t, tmp)[len];\
|
|
int i,j,c;\
|
|
for(c=0; c<channels; c++){\
|
|
ff_float_to_int16_##cpu(tmp, src[c], len);\
|
|
for(i=0, j=c; i<len; i++, j+=channels)\
|
|
dst[j] = tmp[i];\
|
|
}\
|
|
}\
|
|
\
|
|
static void float_to_int16_interleave_##cpu(int16_t *dst, const float **src, long len, int channels){\
|
|
if(channels==1)\
|
|
ff_float_to_int16_##cpu(dst, src[0], len);\
|
|
else if(channels==2){\
|
|
ff_float_to_int16_interleave2_##cpu(dst, src, len);\
|
|
}else if(channels==6){\
|
|
ff_float_to_int16_interleave6_##cpu(dst, src, len);\
|
|
}else\
|
|
float_to_int16_interleave_misc_##cpu(dst, src, len, channels);\
|
|
}
|
|
|
|
FLOAT_TO_INT16_INTERLEAVE(3dnow)
|
|
FLOAT_TO_INT16_INTERLEAVE(sse)
|
|
FLOAT_TO_INT16_INTERLEAVE(sse2)
|
|
|
|
static void float_to_int16_interleave_3dn2(int16_t *dst, const float **src, long len, int channels){
|
|
if(channels==6)
|
|
ff_float_to_int16_interleave6_3dn2(dst, src, len);
|
|
else
|
|
float_to_int16_interleave_3dnow(dst, src, len, channels);
|
|
}
|
|
|
|
void ff_float_interleave2_mmx(float *dst, const float **src, unsigned int len);
|
|
void ff_float_interleave2_sse(float *dst, const float **src, unsigned int len);
|
|
|
|
void ff_float_interleave6_mmx(float *dst, const float **src, unsigned int len);
|
|
void ff_float_interleave6_sse(float *dst, const float **src, unsigned int len);
|
|
|
|
static void float_interleave_mmx(float *dst, const float **src,
|
|
unsigned int len, int channels)
|
|
{
|
|
if (channels == 2) {
|
|
ff_float_interleave2_mmx(dst, src, len);
|
|
} else if (channels == 6)
|
|
ff_float_interleave6_mmx(dst, src, len);
|
|
else
|
|
ff_float_interleave_c(dst, src, len, channels);
|
|
}
|
|
|
|
static void float_interleave_sse(float *dst, const float **src,
|
|
unsigned int len, int channels)
|
|
{
|
|
if (channels == 2) {
|
|
ff_float_interleave2_sse(dst, src, len);
|
|
} else if (channels == 6)
|
|
ff_float_interleave6_sse(dst, src, len);
|
|
else
|
|
ff_float_interleave_c(dst, src, len, channels);
|
|
}
|
|
#endif
|
|
|
|
void ff_fmt_convert_init_x86(FmtConvertContext *c, AVCodecContext *avctx)
|
|
{
|
|
int mm_flags = av_get_cpu_flags();
|
|
|
|
#if HAVE_YASM
|
|
if (mm_flags & AV_CPU_FLAG_MMX) {
|
|
c->float_interleave = float_interleave_mmx;
|
|
|
|
if (HAVE_AMD3DNOW && mm_flags & AV_CPU_FLAG_3DNOW) {
|
|
if(!(avctx->flags & CODEC_FLAG_BITEXACT)){
|
|
c->float_to_int16 = ff_float_to_int16_3dnow;
|
|
c->float_to_int16_interleave = float_to_int16_interleave_3dnow;
|
|
}
|
|
}
|
|
if (HAVE_AMD3DNOWEXT && mm_flags & AV_CPU_FLAG_3DNOWEXT) {
|
|
if(!(avctx->flags & CODEC_FLAG_BITEXACT)){
|
|
c->float_to_int16_interleave = float_to_int16_interleave_3dn2;
|
|
}
|
|
}
|
|
if (HAVE_SSE && mm_flags & AV_CPU_FLAG_SSE) {
|
|
c->int32_to_float_fmul_scalar = ff_int32_to_float_fmul_scalar_sse;
|
|
c->float_to_int16 = ff_float_to_int16_sse;
|
|
c->float_to_int16_interleave = float_to_int16_interleave_sse;
|
|
c->float_interleave = float_interleave_sse;
|
|
}
|
|
if (HAVE_SSE && mm_flags & AV_CPU_FLAG_SSE2) {
|
|
c->int32_to_float_fmul_scalar = ff_int32_to_float_fmul_scalar_sse2;
|
|
c->float_to_int16 = ff_float_to_int16_sse2;
|
|
c->float_to_int16_interleave = float_to_int16_interleave_sse2;
|
|
}
|
|
}
|
|
#endif
|
|
}
|