diff options
author | Jason Garrett-Glaser | 2011-01-14 21:34:25 +0000 |
---|---|---|
committer | Jason Garrett-Glaser | 2011-01-14 21:34:25 +0000 |
commit | 19fb234e4af1ff9f58ff2fdd604ac6f6bb87ad6b (patch) | |
tree | 220be84d79d9c771c1afeab43fdd2aaa82fea01d /libavcodec/h264dsp.h | |
parent | 6c18f1cda2e2b2471ebf75d30d552cb0cb61b6ad (diff) |
H.264: split luma dc idct out and implement MMX/SSE2 versions
About 2.5x the speed.
NOTE: the way that the asm code handles large qmuls is a bit suboptimal.
If x264-style dequant was used (separate shift and qmul values), it might
be possible to get some extra speed.
Originally committed as revision 26336 to svn://svn.ffmpeg.org/ffmpeg/trunk
Diffstat (limited to 'libavcodec/h264dsp.h')
-rw-r--r-- | libavcodec/h264dsp.h | 2 |
1 files changed, 2 insertions, 0 deletions
diff --git a/libavcodec/h264dsp.h b/libavcodec/h264dsp.h index 3d32a9c18f..9f16480d56 100644 --- a/libavcodec/h264dsp.h +++ b/libavcodec/h264dsp.h @@ -65,11 +65,13 @@ typedef struct H264DSPContext{ void (*h264_idct8_add)(uint8_t *dst/*align 8*/, DCTELEM *block/*align 16*/, int stride); void (*h264_idct_dc_add)(uint8_t *dst/*align 4*/, DCTELEM *block/*align 16*/, int stride); void (*h264_idct8_dc_add)(uint8_t *dst/*align 8*/, DCTELEM *block/*align 16*/, int stride); + void (*h264_dct)(DCTELEM block[4][4]); void (*h264_idct_add16)(uint8_t *dst/*align 16*/, const int *blockoffset, DCTELEM *block/*align 16*/, int stride, const uint8_t nnzc[6*8]); void (*h264_idct8_add4)(uint8_t *dst/*align 16*/, const int *blockoffset, DCTELEM *block/*align 16*/, int stride, const uint8_t nnzc[6*8]); void (*h264_idct_add8)(uint8_t **dst/*align 16*/, const int *blockoffset, DCTELEM *block/*align 16*/, int stride, const uint8_t nnzc[6*8]); void (*h264_idct_add16intra)(uint8_t *dst/*align 16*/, const int *blockoffset, DCTELEM *block/*align 16*/, int stride, const uint8_t nnzc[6*8]); + void (*h264_luma_dc_dequant_idct)(DCTELEM *output, DCTELEM *input/*align 16*/, int qmul); }H264DSPContext; void ff_h264dsp_init(H264DSPContext *c); |