diff options
| author | Ronald S. Bultje <rsbultje@gmail.com> | 2010-07-03 19:26:30 +0000 |
|---|---|---|
| committer | Ronald S. Bultje <rsbultje@gmail.com> | 2010-07-03 19:26:30 +0000 |
| commit | f2a30bd84071eeb5000dd916ec16418851686254 (patch) | |
| tree | 408effe7732216b9fe4fde2177d6ec213aabc157 /libavcodec/x86/vp8dsp-init.c | |
| parent | ea28e81faa0b211a985f5aa520c60334aaf081cc (diff) | |
| download | ffmpeg-f2a30bd84071eeb5000dd916ec16418851686254.tar.gz | |
Simple H/V loopfilter for VP8 in MMX, MMX2 and SSE2 (yay for yasm macros).
Originally committed as revision 24029 to svn://svn.ffmpeg.org/ffmpeg/trunk
Diffstat (limited to 'libavcodec/x86/vp8dsp-init.c')
| -rw-r--r-- | libavcodec/x86/vp8dsp-init.c | 16 |
1 files changed, 16 insertions, 0 deletions
diff --git a/libavcodec/x86/vp8dsp-init.c b/libavcodec/x86/vp8dsp-init.c index 698d394278..9cd72b6955 100644 --- a/libavcodec/x86/vp8dsp-init.c +++ b/libavcodec/x86/vp8dsp-init.c @@ -222,6 +222,13 @@ extern void ff_vp8_idct_dc_add_mmx(uint8_t *dst, DCTELEM block[16], int stride); extern void ff_vp8_idct_dc_add_sse4(uint8_t *dst, DCTELEM block[16], int stride); extern void ff_vp8_luma_dc_wht_mmxext(DCTELEM block[4][4][16], DCTELEM dc[16]); extern void ff_vp8_idct_add_mmx(uint8_t *dst, DCTELEM block[16], int stride); + +extern void ff_vp8_v_loop_filter_simple_mmx (uint8_t *dst, int stride, int flim); +extern void ff_vp8_v_loop_filter_simple_mmxext(uint8_t *dst, int stride, int flim); +extern void ff_vp8_v_loop_filter_simple_sse2 (uint8_t *dst, int stride, int flim); +extern void ff_vp8_h_loop_filter_simple_mmx (uint8_t *dst, int stride, int flim); +extern void ff_vp8_h_loop_filter_simple_mmxext(uint8_t *dst, int stride, int flim); +extern void ff_vp8_h_loop_filter_simple_sse2 (uint8_t *dst, int stride, int flim); #endif #define VP8_LUMA_MC_FUNC(IDX, SIZE, OPT) \ @@ -260,6 +267,9 @@ av_cold void ff_vp8dsp_init_x86(VP8DSPContext* c) c->put_vp8_bilinear_pixels_tab[0][0][0] = ff_put_vp8_pixels16_mmx; c->put_vp8_epel_pixels_tab[1][0][0] = c->put_vp8_bilinear_pixels_tab[1][0][0] = ff_put_vp8_pixels8_mmx; + + c->vp8_v_loop_filter_simple = ff_vp8_v_loop_filter_simple_mmx; + c->vp8_h_loop_filter_simple = ff_vp8_h_loop_filter_simple_mmx; } /* note that 4-tap width=16 functions are missing because w=16 @@ -272,6 +282,9 @@ av_cold void ff_vp8dsp_init_x86(VP8DSPContext* c) VP8_BILINEAR_MC_FUNC(0, 16, mmxext); VP8_BILINEAR_MC_FUNC(1, 8, mmxext); VP8_BILINEAR_MC_FUNC(2, 4, mmxext); + + c->vp8_v_loop_filter_simple = ff_vp8_v_loop_filter_simple_mmxext; + c->vp8_h_loop_filter_simple = ff_vp8_h_loop_filter_simple_mmxext; } if (mm_flags & FF_MM_SSE) { @@ -284,6 +297,9 @@ av_cold void ff_vp8dsp_init_x86(VP8DSPContext* c) VP8_MC_FUNC(1, 8, sse2); VP8_BILINEAR_MC_FUNC(0, 16, sse2); VP8_BILINEAR_MC_FUNC(1, 8, sse2); + + c->vp8_v_loop_filter_simple = ff_vp8_v_loop_filter_simple_sse2; + c->vp8_h_loop_filter_simple = ff_vp8_h_loop_filter_simple_sse2; } if (mm_flags & FF_MM_SSSE3) { |