falkTX
/
FFmpeg
mirror of https://github.com/falkTX/FFmpeg.git
1
0
Fork 0
Code Issues 0 Releases 338 Wiki Activity
You can not select more than 25 topics Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.
86080 Commits
32 Branches
541MB
Tree: f61d454ca1
Branches Tags
${ item.name }
Create branch ${ searchTerm }
from 'f61d454ca1'
${ noResults }
FFmpeg/libavcodec/x86/h264dsp_init.c

442 lines
20KB
Raw Blame History 

  1. /*

  2.  * Copyright (c) 2004-2005 Michael Niedermayer, Loren Merritt

  3.  *

  4.  * This file is part of FFmpeg.

  5.  *

  6.  * FFmpeg is free software; you can redistribute it and/or

  7.  * modify it under the terms of the GNU Lesser General Public

  8.  * License as published by the Free Software Foundation; either

  9.  * version 2.1 of the License, or (at your option) any later version.

  10.  *

  11.  * FFmpeg is distributed in the hope that it will be useful,

  12.  * but WITHOUT ANY WARRANTY; without even the implied warranty of

  13.  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU

  14.  * Lesser General Public License for more details.

  15.  *

  16.  * You should have received a copy of the GNU Lesser General Public

  17.  * License along with FFmpeg; if not, write to the Free Software

  18.  * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA

  19.  */

  20. 

  21. #include "libavutil/attributes.h"

  22. #include "libavutil/cpu.h"

  23. #include "libavutil/x86/asm.h"

  24. #include "libavutil/x86/cpu.h"

  25. #include "libavcodec/h264dsp.h"

  26. 

  27. /***********************************/

  28. /* IDCT */

  29. #define IDCT_ADD_FUNC(NUM, DEPTH, OPT)                                  \

  30. void ff_h264_idct ## NUM ## _add_ ## DEPTH ## _ ## OPT(uint8_t *dst,    \

  31.                                                        int16_t *block,  \

  32.                                                        int stride);

  33. 

  34. IDCT_ADD_FUNC(, 8, mmx)

  35. IDCT_ADD_FUNC(, 8, avx)

  36. IDCT_ADD_FUNC(, 10, sse2)

  37. IDCT_ADD_FUNC(_dc, 8, mmxext)

  38. IDCT_ADD_FUNC(_dc, 10, mmxext)

  39. IDCT_ADD_FUNC(8_dc, 8, mmxext)

  40. IDCT_ADD_FUNC(8_dc, 10, sse2)

  41. IDCT_ADD_FUNC(8, 8, mmx)

  42. IDCT_ADD_FUNC(8, 8, sse2)

  43. IDCT_ADD_FUNC(8, 10, sse2)

  44. IDCT_ADD_FUNC(, 10, avx)

  45. IDCT_ADD_FUNC(8_dc, 10, avx)

  46. IDCT_ADD_FUNC(8, 10, avx)

  47. 

  48. 

  49. #define IDCT_ADD_REP_FUNC(NUM, REP, DEPTH, OPT)                         \

  50. void ff_h264_idct ## NUM ## _add ## REP ## _ ## DEPTH ## _ ## OPT       \

  51.     (uint8_t *dst, const int *block_offset,                             \

  52.      int16_t *block, int stride, const uint8_t nnzc[6 * 8]);

  53. 

  54. IDCT_ADD_REP_FUNC(8, 4, 8, mmx)

  55. IDCT_ADD_REP_FUNC(8, 4, 8, mmxext)

  56. IDCT_ADD_REP_FUNC(8, 4, 8, sse2)

  57. IDCT_ADD_REP_FUNC(8, 4, 10, sse2)

  58. IDCT_ADD_REP_FUNC(8, 4, 10, avx)

  59. IDCT_ADD_REP_FUNC(, 16, 8, mmx)

  60. IDCT_ADD_REP_FUNC(, 16, 8, mmxext)

  61. IDCT_ADD_REP_FUNC(, 16, 8, sse2)

  62. IDCT_ADD_REP_FUNC(, 16, 10, sse2)

  63. IDCT_ADD_REP_FUNC(, 16intra, 8, mmx)

  64. IDCT_ADD_REP_FUNC(, 16intra, 8, mmxext)

  65. IDCT_ADD_REP_FUNC(, 16intra, 8, sse2)

  66. IDCT_ADD_REP_FUNC(, 16intra, 10, sse2)

  67. IDCT_ADD_REP_FUNC(, 16, 10, avx)

  68. IDCT_ADD_REP_FUNC(, 16intra, 10, avx)

  69. 

  70. 

  71. #define IDCT_ADD_REP_FUNC2(NUM, REP, DEPTH, OPT)                      \

  72. void ff_h264_idct ## NUM ## _add ## REP ## _ ## DEPTH ## _ ## OPT     \

  73.     (uint8_t **dst, const int *block_offset,                          \

  74.      int16_t *block, int stride, const uint8_t nnzc[6 * 8]);

  75. 

  76. IDCT_ADD_REP_FUNC2(, 8, 8, mmx)

  77. IDCT_ADD_REP_FUNC2(, 8, 8, mmxext)

  78. IDCT_ADD_REP_FUNC2(, 8, 8, sse2)

  79. IDCT_ADD_REP_FUNC2(, 8, 10, sse2)

  80. IDCT_ADD_REP_FUNC2(, 8, 10, avx)

  81. 

  82. IDCT_ADD_REP_FUNC2(, 8_422, 8, mmx)

  83. 

  84. IDCT_ADD_REP_FUNC2(, 8_422, 10, sse2)

  85. IDCT_ADD_REP_FUNC2(, 8_422, 10, avx)

  86. 

  87. void ff_h264_luma_dc_dequant_idct_mmx(int16_t *output, int16_t *input, int qmul);

  88. void ff_h264_luma_dc_dequant_idct_sse2(int16_t *output, int16_t *input, int qmul);

  89. 

  90. /***********************************/

  91. /* deblocking */

  92. 

  93. void ff_h264_loop_filter_strength_mmxext(int16_t bS[2][4][4], uint8_t nnz[40],

  94.                                          int8_t ref[2][40],

  95.                                          int16_t mv[2][40][2],

  96.                                          int bidir, int edges, int step,

  97.                                          int mask_mv0, int mask_mv1, int field);

  98. 

  99. #define LF_FUNC(DIR, TYPE, DEPTH, OPT)                                        \

  100. void ff_deblock_ ## DIR ## _ ## TYPE ## _ ## DEPTH ## _ ## OPT(uint8_t *pix,  \

  101.                                                                int stride,    \

  102.                                                                int alpha,     \

  103.                                                                int beta,      \

  104.                                                                int8_t *tc0);

  105. #define LF_IFUNC(DIR, TYPE, DEPTH, OPT) \

  106. void ff_deblock_ ## DIR ## _ ## TYPE ## _ ## DEPTH ## _ ## OPT(uint8_t *pix,  \

  107.                                                                int stride,    \

  108.                                                                int alpha,     \

  109.                                                                int beta);

  110. 

  111. #define LF_FUNCS(type, depth)                   \

  112. LF_FUNC(h,  chroma,          depth, mmxext)     \

  113. LF_IFUNC(h, chroma_intra,    depth, mmxext)     \

  114. LF_FUNC(h,  chroma422,       depth, mmxext)     \

  115. LF_IFUNC(h, chroma422_intra, depth, mmxext)     \

  116. LF_FUNC(v,  chroma,          depth, mmxext)     \

  117. LF_IFUNC(v, chroma_intra,    depth, mmxext)     \

  118. LF_FUNC(h,  luma,            depth, mmxext)     \

  119. LF_IFUNC(h, luma_intra,      depth, mmxext)     \

  120. LF_FUNC(h,  luma,            depth, sse2)       \

  121. LF_IFUNC(h, luma_intra,      depth, sse2)       \

  122. LF_FUNC(v,  luma,            depth, sse2)       \

  123. LF_IFUNC(v, luma_intra,      depth, sse2)       \

  124. LF_FUNC(h,  chroma,          depth, sse2)       \

  125. LF_IFUNC(h, chroma_intra,    depth, sse2)       \

  126. LF_FUNC(h,  chroma422,       depth, sse2)       \

  127. LF_IFUNC(h, chroma422_intra, depth, sse2)       \

  128. LF_FUNC(v,  chroma,          depth, sse2)       \

  129. LF_IFUNC(v, chroma_intra,    depth, sse2)       \

  130. LF_FUNC(h,  luma,            depth, avx)        \

  131. LF_IFUNC(h, luma_intra,      depth, avx)        \

  132. LF_FUNC(v,  luma,            depth, avx)        \

  133. LF_IFUNC(v, luma_intra,      depth, avx)        \

  134. LF_FUNC(h,  chroma,          depth, avx)        \

  135. LF_IFUNC(h, chroma_intra,    depth, avx)        \

  136. LF_FUNC(h,  chroma422,       depth, avx)        \

  137. LF_IFUNC(h, chroma422_intra, depth, avx)        \

  138. LF_FUNC(v,  chroma,          depth, avx)        \

  139. LF_IFUNC(v, chroma_intra,    depth, avx)

  140. 

  141. LF_FUNC(h, luma_mbaff, 8, sse2)

  142. LF_FUNC(h, luma_mbaff, 8, avx)

  143. 

  144. LF_FUNCS(uint8_t,   8)

  145. LF_FUNCS(uint16_t, 10)

  146. 

  147. #if ARCH_X86_32 && HAVE_MMXEXT_EXTERNAL

  148. LF_FUNC(v8, luma, 8, mmxext)

  149. static void deblock_v_luma_8_mmxext(uint8_t *pix, int stride, int alpha,

  150.                                     int beta, int8_t *tc0)

  151. {

  152.     if ((tc0[0] & tc0[1]) >= 0)

  153.         ff_deblock_v8_luma_8_mmxext(pix + 0, stride, alpha, beta, tc0);

  154.     if ((tc0[2] & tc0[3]) >= 0)

  155.         ff_deblock_v8_luma_8_mmxext(pix + 8, stride, alpha, beta, tc0 + 2);

  156. }

  157. LF_IFUNC(v8, luma_intra, 8, mmxext)

  158. static void deblock_v_luma_intra_8_mmxext(uint8_t *pix, int stride,

  159.                                           int alpha, int beta)

  160. {

  161.     ff_deblock_v8_luma_intra_8_mmxext(pix + 0, stride, alpha, beta);

  162.     ff_deblock_v8_luma_intra_8_mmxext(pix + 8, stride, alpha, beta);

  163. }

  164. #endif /* ARCH_X86_32 && HAVE_MMXEXT_EXTERNAL */

  165. 

  166. LF_FUNC(v,  luma,       10, mmxext)

  167. LF_IFUNC(v, luma_intra, 10, mmxext)

  168. 

  169. /***********************************/

  170. /* weighted prediction */

  171. 

  172. #define H264_WEIGHT(W, OPT)                                             \

  173. void ff_h264_weight_ ## W ## _ ## OPT(uint8_t *dst, ptrdiff_t stride,   \

  174.                                       int height, int log2_denom,       \

  175.                                       int weight, int offset);

  176. 

  177. #define H264_BIWEIGHT(W, OPT)                                           \

  178. void ff_h264_biweight_ ## W ## _ ## OPT(uint8_t *dst, uint8_t *src,     \

  179.                                         ptrdiff_t stride, int height,   \

  180.                                         int log2_denom, int weightd,    \

  181.                                         int weights, int offset);

  182. 

  183. #define H264_BIWEIGHT_MMX(W)                    \

  184.     H264_WEIGHT(W, mmxext)                      \

  185.     H264_BIWEIGHT(W, mmxext)

  186. 

  187. #define H264_BIWEIGHT_MMX_SSE(W)                \

  188.     H264_BIWEIGHT_MMX(W)                        \

  189.     H264_WEIGHT(W, sse2)                        \

  190.     H264_BIWEIGHT(W, sse2)                      \

  191.     H264_BIWEIGHT(W, ssse3)

  192. 

  193. H264_BIWEIGHT_MMX_SSE(16)

  194. H264_BIWEIGHT_MMX_SSE(8)

  195. H264_BIWEIGHT_MMX(4)

  196. 

  197. #define H264_WEIGHT_10(W, DEPTH, OPT)                                   \

  198. void ff_h264_weight_ ## W ## _ ## DEPTH ## _ ## OPT(uint8_t *dst,       \

  199.                                                     ptrdiff_t stride,   \

  200.                                                     int height,         \

  201.                                                     int log2_denom,     \

  202.                                                     int weight,         \

  203.                                                     int offset);

  204. 

  205. #define H264_BIWEIGHT_10(W, DEPTH, OPT)                                 \

  206. void ff_h264_biweight_ ## W ## _ ## DEPTH ## _ ## OPT(uint8_t *dst,     \

  207.                                                       uint8_t *src,     \

  208.                                                       ptrdiff_t stride, \

  209.                                                       int height,       \

  210.                                                       int log2_denom,   \

  211.                                                       int weightd,      \

  212.                                                       int weights,      \

  213.                                                       int offset);

  214. 

  215. #define H264_BIWEIGHT_10_SSE(W, DEPTH)          \

  216.     H264_WEIGHT_10(W, DEPTH, sse2)              \

  217.     H264_WEIGHT_10(W, DEPTH, sse4)              \

  218.     H264_BIWEIGHT_10(W, DEPTH, sse2)            \

  219.     H264_BIWEIGHT_10(W, DEPTH, sse4)

  220. 

  221. H264_BIWEIGHT_10_SSE(16, 10)

  222. H264_BIWEIGHT_10_SSE(8,  10)

  223. H264_BIWEIGHT_10_SSE(4,  10)

  224. 

  225. av_cold void ff_h264dsp_init_x86(H264DSPContext *c, const int bit_depth,

  226.                                  const int chroma_format_idc)

  227. {

  228. #if HAVE_YASM

  229.     int cpu_flags = av_get_cpu_flags();

  230. 

  231.     if (EXTERNAL_MMXEXT(cpu_flags) && chroma_format_idc <= 1)

  232.         c->h264_loop_filter_strength = ff_h264_loop_filter_strength_mmxext;

  233. 

  234.     if (bit_depth == 8) {

  235.         if (EXTERNAL_MMX(cpu_flags)) {

  236.             c->h264_idct_dc_add   =

  237.             c->h264_idct_add      = ff_h264_idct_add_8_mmx;

  238.             c->h264_idct8_dc_add  =

  239.             c->h264_idct8_add     = ff_h264_idct8_add_8_mmx;

  240. 

  241.             c->h264_idct_add16 = ff_h264_idct_add16_8_mmx;

  242.             c->h264_idct8_add4 = ff_h264_idct8_add4_8_mmx;

  243.             if (chroma_format_idc <= 1) {

  244.                 c->h264_idct_add8 = ff_h264_idct_add8_8_mmx;

  245.             } else {

  246.                 c->h264_idct_add8 = ff_h264_idct_add8_422_8_mmx;

  247.             }

  248.             c->h264_idct_add16intra = ff_h264_idct_add16intra_8_mmx;

  249.             if (cpu_flags & AV_CPU_FLAG_CMOV)

  250.                 c->h264_luma_dc_dequant_idct = ff_h264_luma_dc_dequant_idct_mmx;

  251.         }

  252.         if (EXTERNAL_MMXEXT(cpu_flags)) {

  253.             c->h264_idct_dc_add  = ff_h264_idct_dc_add_8_mmxext;

  254.             c->h264_idct8_dc_add = ff_h264_idct8_dc_add_8_mmxext;

  255.             c->h264_idct_add16   = ff_h264_idct_add16_8_mmxext;

  256.             c->h264_idct8_add4   = ff_h264_idct8_add4_8_mmxext;

  257.             if (chroma_format_idc <= 1)

  258.                 c->h264_idct_add8 = ff_h264_idct_add8_8_mmxext;

  259.             c->h264_idct_add16intra = ff_h264_idct_add16intra_8_mmxext;

  260. 

  261.             c->h264_v_loop_filter_chroma       = ff_deblock_v_chroma_8_mmxext;

  262.             c->h264_v_loop_filter_chroma_intra = ff_deblock_v_chroma_intra_8_mmxext;

  263.             if (chroma_format_idc <= 1) {

  264.                 c->h264_h_loop_filter_chroma       = ff_deblock_h_chroma_8_mmxext;

  265.                 c->h264_h_loop_filter_chroma_intra = ff_deblock_h_chroma_intra_8_mmxext;

  266.             } else {

  267.                 c->h264_h_loop_filter_chroma = ff_deblock_h_chroma422_8_mmxext;

  268.                 c->h264_h_loop_filter_chroma_intra = ff_deblock_h_chroma422_intra_8_mmxext;

  269.             }

  270. #if ARCH_X86_32 && HAVE_MMXEXT_EXTERNAL

  271.             c->h264_v_loop_filter_luma       = deblock_v_luma_8_mmxext;

  272.             c->h264_h_loop_filter_luma       = ff_deblock_h_luma_8_mmxext;

  273.             c->h264_v_loop_filter_luma_intra = deblock_v_luma_intra_8_mmxext;

  274.             c->h264_h_loop_filter_luma_intra = ff_deblock_h_luma_intra_8_mmxext;

  275. #endif /* ARCH_X86_32 && HAVE_MMXEXT_EXTERNAL */

  276.             c->weight_h264_pixels_tab[0] = ff_h264_weight_16_mmxext;

  277.             c->weight_h264_pixels_tab[1] = ff_h264_weight_8_mmxext;

  278.             c->weight_h264_pixels_tab[2] = ff_h264_weight_4_mmxext;

  279. 

  280.             c->biweight_h264_pixels_tab[0] = ff_h264_biweight_16_mmxext;

  281.             c->biweight_h264_pixels_tab[1] = ff_h264_biweight_8_mmxext;

  282.             c->biweight_h264_pixels_tab[2] = ff_h264_biweight_4_mmxext;

  283.         }

  284.         if (EXTERNAL_SSE2(cpu_flags)) {

  285.             c->h264_idct8_add  = ff_h264_idct8_add_8_sse2;

  286. 

  287.             c->h264_idct_add16 = ff_h264_idct_add16_8_sse2;

  288.             c->h264_idct8_add4 = ff_h264_idct8_add4_8_sse2;

  289.             if (chroma_format_idc <= 1)

  290.                 c->h264_idct_add8 = ff_h264_idct_add8_8_sse2;

  291.             c->h264_idct_add16intra      = ff_h264_idct_add16intra_8_sse2;

  292.             c->h264_luma_dc_dequant_idct = ff_h264_luma_dc_dequant_idct_sse2;

  293. 

  294.             c->weight_h264_pixels_tab[0] = ff_h264_weight_16_sse2;

  295.             c->weight_h264_pixels_tab[1] = ff_h264_weight_8_sse2;

  296. 

  297.             c->biweight_h264_pixels_tab[0] = ff_h264_biweight_16_sse2;

  298.             c->biweight_h264_pixels_tab[1] = ff_h264_biweight_8_sse2;

  299. 

  300.             c->h264_v_loop_filter_luma       = ff_deblock_v_luma_8_sse2;

  301.             c->h264_h_loop_filter_luma       = ff_deblock_h_luma_8_sse2;

  302.             c->h264_v_loop_filter_luma_intra = ff_deblock_v_luma_intra_8_sse2;

  303.             c->h264_h_loop_filter_luma_intra = ff_deblock_h_luma_intra_8_sse2;

  304. 

  305. #if ARCH_X86_64

  306.             c->h264_h_loop_filter_luma_mbaff = ff_deblock_h_luma_mbaff_8_sse2;

  307. #endif

  308. 

  309.             c->h264_v_loop_filter_chroma       = ff_deblock_v_chroma_8_sse2;

  310.             c->h264_v_loop_filter_chroma_intra = ff_deblock_v_chroma_intra_8_sse2;

  311.             if (chroma_format_idc <= 1) {

  312.                 c->h264_h_loop_filter_chroma       = ff_deblock_h_chroma_8_sse2;

  313.                 c->h264_h_loop_filter_chroma_intra = ff_deblock_h_chroma_intra_8_sse2;

  314.             } else {

  315.                 c->h264_h_loop_filter_chroma       = ff_deblock_h_chroma422_8_sse2;

  316.                 c->h264_h_loop_filter_chroma_intra = ff_deblock_h_chroma422_intra_8_sse2;

  317.             }

  318.         }

  319.         if (EXTERNAL_SSSE3(cpu_flags)) {

  320.             c->biweight_h264_pixels_tab[0] = ff_h264_biweight_16_ssse3;

  321.             c->biweight_h264_pixels_tab[1] = ff_h264_biweight_8_ssse3;

  322.         }

  323.         if (EXTERNAL_AVX(cpu_flags)) {

  324.             c->h264_v_loop_filter_luma       = ff_deblock_v_luma_8_avx;

  325.             c->h264_h_loop_filter_luma       = ff_deblock_h_luma_8_avx;

  326.             c->h264_v_loop_filter_luma_intra = ff_deblock_v_luma_intra_8_avx;

  327.             c->h264_h_loop_filter_luma_intra = ff_deblock_h_luma_intra_8_avx;

  328. #if ARCH_X86_64

  329.             c->h264_h_loop_filter_luma_mbaff = ff_deblock_h_luma_mbaff_8_avx;

  330. #endif

  331. 

  332.             c->h264_v_loop_filter_chroma       = ff_deblock_v_chroma_8_avx;

  333.             c->h264_v_loop_filter_chroma_intra = ff_deblock_v_chroma_intra_8_avx;

  334.             if (chroma_format_idc <= 1) {

  335.                 c->h264_h_loop_filter_chroma       = ff_deblock_h_chroma_8_avx;

  336.                 c->h264_h_loop_filter_chroma_intra = ff_deblock_h_chroma_intra_8_avx;

  337.             } else {

  338.                 c->h264_h_loop_filter_chroma       = ff_deblock_h_chroma422_8_avx;

  339.                 c->h264_h_loop_filter_chroma_intra = ff_deblock_h_chroma422_intra_8_avx;

  340.             }

  341. 

  342.             c->h264_idct_add        = ff_h264_idct_add_8_avx;

  343.         }

  344.     } else if (bit_depth == 10) {

  345.         if (EXTERNAL_MMXEXT(cpu_flags)) {

  346. #if ARCH_X86_32

  347.             c->h264_v_loop_filter_chroma       = ff_deblock_v_chroma_10_mmxext;

  348.             c->h264_v_loop_filter_chroma_intra = ff_deblock_v_chroma_intra_10_mmxext;

  349.             if (chroma_format_idc <= 1) {

  350.                 c->h264_h_loop_filter_chroma = ff_deblock_h_chroma_10_mmxext;

  351.             } else {

  352.                 c->h264_h_loop_filter_chroma = ff_deblock_h_chroma422_10_mmxext;

  353.             }

  354.             c->h264_v_loop_filter_luma         = ff_deblock_v_luma_10_mmxext;

  355.             c->h264_h_loop_filter_luma         = ff_deblock_h_luma_10_mmxext;

  356.             c->h264_v_loop_filter_luma_intra   = ff_deblock_v_luma_intra_10_mmxext;

  357.             c->h264_h_loop_filter_luma_intra   = ff_deblock_h_luma_intra_10_mmxext;

  358. #endif /* ARCH_X86_32 */

  359.             c->h264_idct_dc_add = ff_h264_idct_dc_add_10_mmxext;

  360.         }

  361.         if (EXTERNAL_SSE2(cpu_flags)) {

  362.             c->h264_idct_add     = ff_h264_idct_add_10_sse2;

  363.             c->h264_idct8_dc_add = ff_h264_idct8_dc_add_10_sse2;

  364. 

  365.             c->h264_idct_add16 = ff_h264_idct_add16_10_sse2;

  366.             if (chroma_format_idc <= 1) {

  367.                 c->h264_idct_add8 = ff_h264_idct_add8_10_sse2;

  368.             } else {

  369.                 c->h264_idct_add8 = ff_h264_idct_add8_422_10_sse2;

  370.             }

  371.             c->h264_idct_add16intra = ff_h264_idct_add16intra_10_sse2;

  372. #if HAVE_ALIGNED_STACK

  373.             c->h264_idct8_add  = ff_h264_idct8_add_10_sse2;

  374.             c->h264_idct8_add4 = ff_h264_idct8_add4_10_sse2;

  375. #endif /* HAVE_ALIGNED_STACK */

  376. 

  377.             c->weight_h264_pixels_tab[0] = ff_h264_weight_16_10_sse2;

  378.             c->weight_h264_pixels_tab[1] = ff_h264_weight_8_10_sse2;

  379.             c->weight_h264_pixels_tab[2] = ff_h264_weight_4_10_sse2;

  380. 

  381.             c->biweight_h264_pixels_tab[0] = ff_h264_biweight_16_10_sse2;

  382.             c->biweight_h264_pixels_tab[1] = ff_h264_biweight_8_10_sse2;

  383.             c->biweight_h264_pixels_tab[2] = ff_h264_biweight_4_10_sse2;

  384. 

  385.             c->h264_v_loop_filter_chroma       = ff_deblock_v_chroma_10_sse2;

  386.             c->h264_v_loop_filter_chroma_intra = ff_deblock_v_chroma_intra_10_sse2;

  387.             if (chroma_format_idc <= 1) {

  388.                 c->h264_h_loop_filter_chroma = ff_deblock_h_chroma_10_sse2;

  389.             } else {

  390.                 c->h264_h_loop_filter_chroma = ff_deblock_h_chroma422_10_sse2;

  391.             }

  392. #if HAVE_ALIGNED_STACK

  393.             c->h264_v_loop_filter_luma       = ff_deblock_v_luma_10_sse2;

  394.             c->h264_h_loop_filter_luma       = ff_deblock_h_luma_10_sse2;

  395.             c->h264_v_loop_filter_luma_intra = ff_deblock_v_luma_intra_10_sse2;

  396.             c->h264_h_loop_filter_luma_intra = ff_deblock_h_luma_intra_10_sse2;

  397. #endif /* HAVE_ALIGNED_STACK */

  398.         }

  399.         if (EXTERNAL_SSE4(cpu_flags)) {

  400.             c->weight_h264_pixels_tab[0] = ff_h264_weight_16_10_sse4;

  401.             c->weight_h264_pixels_tab[1] = ff_h264_weight_8_10_sse4;

  402.             c->weight_h264_pixels_tab[2] = ff_h264_weight_4_10_sse4;

  403. 

  404.             c->biweight_h264_pixels_tab[0] = ff_h264_biweight_16_10_sse4;

  405.             c->biweight_h264_pixels_tab[1] = ff_h264_biweight_8_10_sse4;

  406.             c->biweight_h264_pixels_tab[2] = ff_h264_biweight_4_10_sse4;

  407.         }

  408.         if (EXTERNAL_AVX(cpu_flags)) {

  409.             c->h264_idct_dc_add  =

  410.             c->h264_idct_add     = ff_h264_idct_add_10_avx;

  411.             c->h264_idct8_dc_add = ff_h264_idct8_dc_add_10_avx;

  412. 

  413.             c->h264_idct_add16 = ff_h264_idct_add16_10_avx;

  414.             if (chroma_format_idc <= 1) {

  415.                 c->h264_idct_add8 = ff_h264_idct_add8_10_avx;

  416.             } else {

  417.                 c->h264_idct_add8 = ff_h264_idct_add8_422_10_avx;

  418.             }

  419.             c->h264_idct_add16intra = ff_h264_idct_add16intra_10_avx;

  420. #if HAVE_ALIGNED_STACK

  421.             c->h264_idct8_add  = ff_h264_idct8_add_10_avx;

  422.             c->h264_idct8_add4 = ff_h264_idct8_add4_10_avx;

  423. #endif /* HAVE_ALIGNED_STACK */

  424. 

  425.             c->h264_v_loop_filter_chroma       = ff_deblock_v_chroma_10_avx;

  426.             c->h264_v_loop_filter_chroma_intra = ff_deblock_v_chroma_intra_10_avx;

  427.             if (chroma_format_idc <= 1) {

  428.                 c->h264_h_loop_filter_chroma = ff_deblock_h_chroma_10_avx;

  429.             } else {

  430.                 c->h264_h_loop_filter_chroma = ff_deblock_h_chroma422_10_avx;

  431.             }

  432. #if HAVE_ALIGNED_STACK

  433.             c->h264_v_loop_filter_luma         = ff_deblock_v_luma_10_avx;

  434.             c->h264_h_loop_filter_luma         = ff_deblock_h_luma_10_avx;

  435.             c->h264_v_loop_filter_luma_intra   = ff_deblock_v_luma_intra_10_avx;

  436.             c->h264_h_loop_filter_luma_intra   = ff_deblock_h_luma_intra_10_avx;

  437. #endif /* HAVE_ALIGNED_STACK */

  438.         }

  439.     }

  440. #endif

  441. }