xref: /third_party/ffmpeg/libavcodec/vp9dsp_template.c

33 static void vert_4x4_c(uint8_t *_dst, ptrdiff_t stride,
40     stride /= sizeof(pixel);
41     AV_WN4PA(dst + stride * 0, p4);
42     AV_WN4PA(dst + stride * 1, p4);
43     AV_WN4PA(dst + stride * 2, p4);
44     AV_WN4PA(dst + stride * 3, p4);
47 static void vert_8x8_c(uint8_t *_dst, ptrdiff_t stride,
56     stride /= sizeof(pixel);
60         dst += stride;
64 static void vert_16x16_c(uint8_t *_dst, ptrdiff_t stride,
75     stride /= sizeof(pixel);
81         dst += stride;
85 static void vert_32x32_c(uint8_t *_dst, ptrdiff_t stride,
100     stride /= sizeof(pixel);
110         dst += stride;
114 static void hor_4x4_c(uint8_t *_dst, ptrdiff_t stride,
120     stride /= sizeof(pixel);
121     AV_WN4PA(dst + stride * 0, PIXEL_SPLAT_X4(left[3]));
122     AV_WN4PA(dst + stride * 1, PIXEL_SPLAT_X4(left[2]));
123     AV_WN4PA(dst + stride * 2, PIXEL_SPLAT_X4(left[1]));
124     AV_WN4PA(dst + stride * 3, PIXEL_SPLAT_X4(left[0]));
127 static void hor_8x8_c(uint8_t *_dst, ptrdiff_t stride,
134     stride /= sizeof(pixel);
140         dst += stride;
144 static void hor_16x16_c(uint8_t *_dst, ptrdiff_t stride,
151     stride /= sizeof(pixel);
159         dst += stride;
163 static void hor_32x32_c(uint8_t *_dst, ptrdiff_t stride,
170     stride /= sizeof(pixel);
182         dst += stride;
188 static void tm_4x4_c(uint8_t *_dst, ptrdiff_t stride,
196     stride /= sizeof(pixel);
204         dst += stride;
208 static void tm_8x8_c(uint8_t *_dst, ptrdiff_t stride,
216     stride /= sizeof(pixel);
228         dst += stride;
232 static void tm_16x16_c(uint8_t *_dst, ptrdiff_t stride,
240     stride /= sizeof(pixel);
260         dst += stride;
264 static void tm_32x32_c(uint8_t *_dst, ptrdiff_t stride,
272     stride /= sizeof(pixel);
308         dst += stride;
314 static void dc_4x4_c(uint8_t *_dst, ptrdiff_t stride,
323     stride /= sizeof(pixel);
324     AV_WN4PA(dst + stride * 0, dc);
325     AV_WN4PA(dst + stride * 1, dc);
326     AV_WN4PA(dst + stride * 2, dc);
327     AV_WN4PA(dst + stride * 3, dc);
330 static void dc_8x8_c(uint8_t *_dst, ptrdiff_t stride,
342     stride /= sizeof(pixel);
346         dst += stride;
350 static void dc_16x16_c(uint8_t *_dst, ptrdiff_t stride,
364     stride /= sizeof(pixel);
370         dst += stride;
374 static void dc_32x32_c(uint8_t *_dst, ptrdiff_t stride,
393     stride /= sizeof(pixel);
403         dst += stride;
407 static void dc_left_4x4_c(uint8_t *_dst, ptrdiff_t stride,
414     stride /= sizeof(pixel);
415     AV_WN4PA(dst + stride * 0, dc);
416     AV_WN4PA(dst + stride * 1, dc);
417     AV_WN4PA(dst + stride * 2, dc);
418     AV_WN4PA(dst + stride * 3, dc);
421 static void dc_left_8x8_c(uint8_t *_dst, ptrdiff_t stride,
431     stride /= sizeof(pixel);
435         dst += stride;
439 static void dc_left_16x16_c(uint8_t *_dst, ptrdiff_t stride,
450     stride /= sizeof(pixel);
456         dst += stride;
460 static void dc_left_32x32_c(uint8_t *_dst, ptrdiff_t stride,
474     stride /= sizeof(pixel);
484         dst += stride;
488 static void dc_top_4x4_c(uint8_t *_dst, ptrdiff_t stride,
495     stride /= sizeof(pixel);
496     AV_WN4PA(dst + stride * 0, dc);
497     AV_WN4PA(dst + stride * 1, dc);
498     AV_WN4PA(dst + stride * 2, dc);
499     AV_WN4PA(dst + stride * 3, dc);
502 static void dc_top_8x8_c(uint8_t *_dst, ptrdiff_t stride,
512     stride /= sizeof(pixel);
516         dst += stride;
520 static void dc_top_16x16_c(uint8_t *_dst, ptrdiff_t stride,
531     stride /= sizeof(pixel);
537         dst += stride;
541 static void dc_top_32x32_c(uint8_t *_dst, ptrdiff_t stride,
555     stride /= sizeof(pixel);
565         dst += stride;
571 static void dc_128_4x4_c(uint8_t *_dst, ptrdiff_t stride,
577     stride /= sizeof(pixel);
578     AV_WN4PA(dst + stride * 0, val);
579     AV_WN4PA(dst + stride * 1, val);
580     AV_WN4PA(dst + stride * 2, val);
581     AV_WN4PA(dst + stride * 3, val);
584 static void dc_128_8x8_c(uint8_t *_dst, ptrdiff_t stride,
591     stride /= sizeof(pixel);
595         dst += stride;
599 static void dc_128_16x16_c(uint8_t *_dst, ptrdiff_t stride,
606     stride /= sizeof(pixel);
612         dst += stride;
616 static void dc_128_32x32_c(uint8_t *_dst, ptrdiff_t stride,
623     stride /= sizeof(pixel);
633         dst += stride;
637 static void dc_127_4x4_c(uint8_t *_dst, ptrdiff_t stride,
643     stride /= sizeof(pixel);
644     AV_WN4PA(dst + stride * 0, val);
645     AV_WN4PA(dst + stride * 1, val);
646     AV_WN4PA(dst + stride * 2, val);
647     AV_WN4PA(dst + stride * 3, val);}
649 static void dc_127_8x8_c(uint8_t *_dst, ptrdiff_t stride,
656     stride /= sizeof(pixel);
660         dst += stride;
664 static void dc_127_16x16_c(uint8_t *_dst, ptrdiff_t stride,
671     stride /= sizeof(pixel);
677         dst += stride;
681 static void dc_127_32x32_c(uint8_t *_dst, ptrdiff_t stride,
688     stride /= sizeof(pixel);
698         dst += stride;
702 static void dc_129_4x4_c(uint8_t *_dst, ptrdiff_t stride,
708     stride /= sizeof(pixel);
709     AV_WN4PA(dst + stride * 0, val);
710     AV_WN4PA(dst + stride * 1, val);
711     AV_WN4PA(dst + stride * 2, val);
712     AV_WN4PA(dst + stride * 3, val);
715 static void dc_129_8x8_c(uint8_t *_dst, ptrdiff_t stride,
722     stride /= sizeof(pixel);
726         dst += stride;
730 static void dc_129_16x16_c(uint8_t *_dst, ptrdiff_t stride,
737     stride /= sizeof(pixel);
743         dst += stride;
747 static void dc_129_32x32_c(uint8_t *_dst, ptrdiff_t stride,
754     stride /= sizeof(pixel);
764         dst += stride;
781 #define DST(x, y) dst[(x) + (y) * stride]
783 static void diag_downleft_4x4_c(uint8_t *_dst, ptrdiff_t stride,
791     stride /= sizeof(pixel);
802 static void diag_downleft_##size##x##size##_c(uint8_t *_dst, ptrdiff_t stride, \
810     stride /= sizeof(pixel); \
816         memcpy(dst + j*stride, v + j, (size - 1 - j) * sizeof(pixel)); \
817         memset_bpc(dst + j*stride + size - 1 - j, top[size - 1], j + 1); \
825 static void diag_downright_4x4_c(uint8_t *_dst, ptrdiff_t stride,
834     stride /= sizeof(pixel);
845 static void diag_downright_##size##x##size##_c(uint8_t *_dst, ptrdiff_t stride, \
854     stride /= sizeof(pixel); \
864         memcpy(dst + j*stride, v + size - 1 - j, size * sizeof(pixel)); \
871 static void vert_right_4x4_c(uint8_t *_dst, ptrdiff_t stride,
880     stride /= sizeof(pixel);
894 static void vert_right_##size##x##size##_c(uint8_t *_dst, ptrdiff_t stride, \
903     stride /= sizeof(pixel); \
919         memcpy(dst +  j*2     *stride, ve + size/2 - 1 - j, size * sizeof(pixel)); \
920         memcpy(dst + (j*2 + 1)*stride, vo + size/2 - 1 - j, size * sizeof(pixel)); \
928 static void hor_down_4x4_c(uint8_t *_dst, ptrdiff_t stride,
937     stride /= sizeof(pixel);
951 static void hor_down_##size##x##size##_c(uint8_t *_dst, ptrdiff_t stride, \
960     stride /= sizeof(pixel); \
972         memcpy(dst + j*stride, v + size*2 - 2 - j*2, size * sizeof(pixel)); \
979 static void vert_left_4x4_c(uint8_t *_dst, ptrdiff_t stride,
987     stride /= sizeof(pixel);
1001 static void vert_left_##size##x##size##_c(uint8_t *_dst, ptrdiff_t stride, \
1009     stride /= sizeof(pixel); \
1018         memcpy(dst +  j*2      * stride, ve + j, (size - j - 1) * sizeof(pixel)); \
1019         memset_bpc(dst +  j*2      * stride + size - j - 1, top[size - 1], j + 1); \
1020         memcpy(dst + (j*2 + 1) * stride, vo + j, (size - j - 1) * sizeof(pixel)); \
1021         memset_bpc(dst + (j*2 + 1) * stride + size - j - 1, top[size - 1], j + 1); \
1029 static void hor_up_4x4_c(uint8_t *_dst, ptrdiff_t stride,
1036     stride /= sizeof(pixel);
1047 static void hor_up_##size##x##size##_c(uint8_t *_dst, ptrdiff_t stride, \
1055     stride /= sizeof(pixel); \
1064         memcpy(dst + j*stride, v + j*2, size * sizeof(pixel)); \
1066         memcpy(dst + j*stride, v + j*2, (size*2 - 2 - j*2) * sizeof(pixel)); \
1067         memset_bpc(dst + j*stride + size*2 - 2 - j*2, left[size - 1], \
1125                                                     ptrdiff_t stride, \
1132     stride /= sizeof(pixel); \
1139                 dst[j * stride] = av_clip_pixel(dst[j * stride] + \
1154             dst[j * stride] = av_clip_pixel(dst[j * stride] + \
1168 #define IN(x) ((dctint) in[(x) * stride])
1170 static av_always_inline void idct4_1d(const dctcoef *in, ptrdiff_t stride,
1186 static av_always_inline void iadst4_1d(const dctcoef *in, ptrdiff_t stride,
1204 static av_always_inline void idct8_1d(const dctcoef *in, ptrdiff_t stride,
1240 static av_always_inline void iadst8_1d(const dctcoef *in, ptrdiff_t stride,
1286 static av_always_inline void idct16_1d(const dctcoef *in, ptrdiff_t stride,
1374 static av_always_inline void iadst16_1d(const dctcoef *in, ptrdiff_t stride,
1479 static av_always_inline void idct32_1d(const dctcoef *in, ptrdiff_t stride,
1687 static av_always_inline void iwht4_1d(const dctcoef *in, ptrdiff_t stride,
1861                                            ptrdiff_t stride, \
1865     stride /= sizeof(pixel); \
1870 lf_8_fn(h, wd, stride, 1) \
1871 lf_8_fn(v, wd, 1, stride)
1882                                         ptrdiff_t stride, \
1885     loop_filter_##dir##_16_8_c(dst, stride, E, I, H); \
1886     loop_filter_##dir##_16_8_c(dst + 8 * stridea, stride, E, I, H); \
1889 lf_16_fn(h, stride)
1896                                                   ptrdiff_t stride, \
1899     loop_filter_##dir##_##wd1##_8_c(dst, stride, E & 0xff, I & 0xff, H & 0xff); \
1900     loop_filter_##dir##_##wd2##_8_c(dst + 8 * stridea, stride, E >> 8, I >> 8, H >> 8); \
1904 lf_mix_fn(h, wd1, wd2, stride) \
1994 #define FILTER_8TAP(src, x, F, stride) \
1995     av_clip_pixel((F[0] * src[x + -3 * stride] + \
1996                    F[1] * src[x + -2 * stride] + \
1997                    F[2] * src[x + -1 * stride] + \
1998                    F[3] * src[x + +0 * stride] + \
1999                    F[4] * src[x + +1 * stride] + \
2000                    F[5] * src[x + +2 * stride] + \
2001                    F[6] * src[x + +3 * stride] + \
2002                    F[7] * src[x + +4 * stride] + 64) >> 7)
2118 #define FILTER_BILIN(src, x, mxy, stride) \
2119     (src[x] + ((mxy * (src[x + stride] - src[x]) + 8) >> 4))

Indexes created Thu Nov 07 10:32:03 CST 2024