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.
38550 Commits
32 Branches
522MB
Tree: 2af8f2cea6
Branches Tags
${ item.name }
Create branch ${ searchTerm }
from '2af8f2cea6'
${ noResults }
FFmpeg/libswscale/output.c

1471 lines
54KB
Raw Blame History 

  1. /*

  2.  * Copyright (C) 2001-2012 Michael Niedermayer <michaelni@gmx.at>

  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 <assert.h>

  22. #include <math.h>

  23. #include <stdint.h>

  24. #include <stdio.h>

  25. #include <string.h>

  26. 

  27. #include "libavutil/avutil.h"

  28. #include "libavutil/avassert.h"

  29. #include "libavutil/bswap.h"

  30. #include "libavutil/cpu.h"

  31. #include "libavutil/intreadwrite.h"

  32. #include "libavutil/mathematics.h"

  33. #include "libavutil/pixdesc.h"

  34. #include "config.h"

  35. #include "rgb2rgb.h"

  36. #include "swscale.h"

  37. #include "swscale_internal.h"

  38. 

  39. DECLARE_ALIGNED(8, static const uint8_t, dither_2x2_4)[2][8]={

  40. {  1,   3,   1,   3,   1,   3,   1,   3, },

  41. {  2,   0,   2,   0,   2,   0,   2,   0, },

  42. };

  43. 

  44. DECLARE_ALIGNED(8, static const uint8_t, dither_2x2_8)[2][8]={

  45. {  6,   2,   6,   2,   6,   2,   6,   2, },

  46. {  0,   4,   0,   4,   0,   4,   0,   4, },

  47. };

  48. 

  49. DECLARE_ALIGNED(8, const uint8_t, dither_4x4_16)[4][8]={

  50. {  8,   4,  11,   7,   8,   4,  11,   7, },

  51. {  2,  14,   1,  13,   2,  14,   1,  13, },

  52. { 10,   6,   9,   5,  10,   6,   9,   5, },

  53. {  0,  12,   3,  15,   0,  12,   3,  15, },

  54. };

  55. 

  56. DECLARE_ALIGNED(8, const uint8_t, dither_8x8_32)[8][8]={

  57. { 17,   9,  23,  15,  16,   8,  22,  14, },

  58. {  5,  29,   3,  27,   4,  28,   2,  26, },

  59. { 21,  13,  19,  11,  20,  12,  18,  10, },

  60. {  0,  24,   6,  30,   1,  25,   7,  31, },

  61. { 16,   8,  22,  14,  17,   9,  23,  15, },

  62. {  4,  28,   2,  26,   5,  29,   3,  27, },

  63. { 20,  12,  18,  10,  21,  13,  19,  11, },

  64. {  1,  25,   7,  31,   0,  24,   6,  30, },

  65. };

  66. 

  67. DECLARE_ALIGNED(8, const uint8_t, dither_8x8_73)[8][8]={

  68. {  0,  55,  14,  68,   3,  58,  17,  72, },

  69. { 37,  18,  50,  32,  40,  22,  54,  35, },

  70. {  9,  64,   5,  59,  13,  67,   8,  63, },

  71. { 46,  27,  41,  23,  49,  31,  44,  26, },

  72. {  2,  57,  16,  71,   1,  56,  15,  70, },

  73. { 39,  21,  52,  34,  38,  19,  51,  33, },

  74. { 11,  66,   7,  62,  10,  65,   6,  60, },

  75. { 48,  30,  43,  25,  47,  29,  42,  24, },

  76. };

  77. 

  78. #if 1

  79. DECLARE_ALIGNED(8, const uint8_t, dither_8x8_220)[8][8]={

  80. {117,  62, 158, 103, 113,  58, 155, 100, },

  81. { 34, 199,  21, 186,  31, 196,  17, 182, },

  82. {144,  89, 131,  76, 141,  86, 127,  72, },

  83. {  0, 165,  41, 206,  10, 175,  52, 217, },

  84. {110,  55, 151,  96, 120,  65, 162, 107, },

  85. { 28, 193,  14, 179,  38, 203,  24, 189, },

  86. {138,  83, 124,  69, 148,  93, 134,  79, },

  87. {  7, 172,  48, 213,   3, 168,  45, 210, },

  88. };

  89. #elif 1

  90. // tries to correct a gamma of 1.5

  91. DECLARE_ALIGNED(8, const uint8_t, dither_8x8_220)[8][8]={

  92. {  0, 143,  18, 200,   2, 156,  25, 215, },

  93. { 78,  28, 125,  64,  89,  36, 138,  74, },

  94. { 10, 180,   3, 161,  16, 195,   8, 175, },

  95. {109,  51,  93,  38, 121,  60, 105,  47, },

  96. {  1, 152,  23, 210,   0, 147,  20, 205, },

  97. { 85,  33, 134,  71,  81,  30, 130,  67, },

  98. { 14, 190,   6, 171,  12, 185,   5, 166, },

  99. {117,  57, 101,  44, 113,  54,  97,  41, },

  100. };

  101. #elif 1

  102. // tries to correct a gamma of 2.0

  103. DECLARE_ALIGNED(8, const uint8_t, dither_8x8_220)[8][8]={

  104. {  0, 124,   8, 193,   0, 140,  12, 213, },

  105. { 55,  14, 104,  42,  66,  19, 119,  52, },

  106. {  3, 168,   1, 145,   6, 187,   3, 162, },

  107. { 86,  31,  70,  21,  99,  39,  82,  28, },

  108. {  0, 134,  11, 206,   0, 129,   9, 200, },

  109. { 62,  17, 114,  48,  58,  16, 109,  45, },

  110. {  5, 181,   2, 157,   4, 175,   1, 151, },

  111. { 95,  36,  78,  26,  90,  34,  74,  24, },

  112. };

  113. #else

  114. // tries to correct a gamma of 2.5

  115. DECLARE_ALIGNED(8, const uint8_t, dither_8x8_220)[8][8]={

  116. {  0, 107,   3, 187,   0, 125,   6, 212, },

  117. { 39,   7,  86,  28,  49,  11, 102,  36, },

  118. {  1, 158,   0, 131,   3, 180,   1, 151, },

  119. { 68,  19,  52,  12,  81,  25,  64,  17, },

  120. {  0, 119,   5, 203,   0, 113,   4, 195, },

  121. { 45,   9,  96,  33,  42,   8,  91,  30, },

  122. {  2, 172,   1, 144,   2, 165,   0, 137, },

  123. { 77,  23,  60,  15,  72,  21,  56,  14, },

  124. };

  125. #endif

  126. 

  127. #define output_pixel(pos, val, bias, signedness) \

  128.     if (big_endian) { \

  129.         AV_WB16(pos, bias + av_clip_ ## signedness ## 16(val >> shift)); \

  130.     } else { \

  131.         AV_WL16(pos, bias + av_clip_ ## signedness ## 16(val >> shift)); \

  132.     }

  133. 

  134. static av_always_inline void

  135. yuv2plane1_16_c_template(const int32_t *src, uint16_t *dest, int dstW,

  136.                          int big_endian, int output_bits)

  137. {

  138.     int i;

  139.     int shift = 3;

  140.     av_assert0(output_bits == 16);

  141. 

  142.     for (i = 0; i < dstW; i++) {

  143.         int val = src[i] + (1 << (shift - 1));

  144.         output_pixel(&dest[i], val, 0, uint);

  145.     }

  146. }

  147. 

  148. static av_always_inline void

  149. yuv2planeX_16_c_template(const int16_t *filter, int filterSize,

  150.                          const int32_t **src, uint16_t *dest, int dstW,

  151.                          int big_endian, int output_bits)

  152. {

  153.     int i;

  154.     int shift = 15;

  155.     av_assert0(output_bits == 16);

  156. 

  157.     for (i = 0; i < dstW; i++) {

  158.         int val = 1 << (shift - 1);

  159.         int j;

  160. 

  161.         /* range of val is [0,0x7FFFFFFF], so 31 bits, but with lanczos/spline

  162.          * filters (or anything with negative coeffs, the range can be slightly

  163.          * wider in both directions. To account for this overflow, we subtract

  164.          * a constant so it always fits in the signed range (assuming a

  165.          * reasonable filterSize), and re-add that at the end. */

  166.         val -= 0x40000000;

  167.         for (j = 0; j < filterSize; j++)

  168.             val += src[j][i] * filter[j];

  169. 

  170.         output_pixel(&dest[i], val, 0x8000, int);

  171.     }

  172. }

  173. 

  174. #undef output_pixel

  175. 

  176. #define output_pixel(pos, val) \

  177.     if (big_endian) { \

  178.         AV_WB16(pos, av_clip_uintp2(val >> shift, output_bits)); \

  179.     } else { \

  180.         AV_WL16(pos, av_clip_uintp2(val >> shift, output_bits)); \

  181.     }

  182. 

  183. static av_always_inline void

  184. yuv2plane1_10_c_template(const int16_t *src, uint16_t *dest, int dstW,

  185.                          int big_endian, int output_bits)

  186. {

  187.     int i;

  188.     int shift = 15 - output_bits;

  189. 

  190.     for (i = 0; i < dstW; i++) {

  191.         int val = src[i] + (1 << (shift - 1));

  192.         output_pixel(&dest[i], val);

  193.     }

  194. }

  195. 

  196. static av_always_inline void

  197. yuv2planeX_10_c_template(const int16_t *filter, int filterSize,

  198.                          const int16_t **src, uint16_t *dest, int dstW,

  199.                          int big_endian, int output_bits)

  200. {

  201.     int i;

  202.     int shift = 11 + 16 - output_bits;

  203. 

  204.     for (i = 0; i < dstW; i++) {

  205.         int val = 1 << (shift - 1);

  206.         int j;

  207. 

  208.         for (j = 0; j < filterSize; j++)

  209.             val += src[j][i] * filter[j];

  210. 

  211.         output_pixel(&dest[i], val);

  212.     }

  213. }

  214. 

  215. #undef output_pixel

  216. 

  217. #define yuv2NBPS(bits, BE_LE, is_be, template_size, typeX_t) \

  218. static void yuv2plane1_ ## bits ## BE_LE ## _c(const int16_t *src, \

  219.                               uint8_t *dest, int dstW, \

  220.                               const uint8_t *dither, int offset)\

  221. { \

  222.     yuv2plane1_ ## template_size ## _c_template((const typeX_t *) src, \

  223.                          (uint16_t *) dest, dstW, is_be, bits); \

  224. }\

  225. static void yuv2planeX_ ## bits ## BE_LE ## _c(const int16_t *filter, int filterSize, \

  226.                               const int16_t **src, uint8_t *dest, int dstW, \

  227.                               const uint8_t *dither, int offset)\

  228. { \

  229.     yuv2planeX_## template_size ## _c_template(filter, \

  230.                          filterSize, (const typeX_t **) src, \

  231.                          (uint16_t *) dest, dstW, is_be, bits); \

  232. }

  233. yuv2NBPS( 9, BE, 1, 10, int16_t)

  234. yuv2NBPS( 9, LE, 0, 10, int16_t)

  235. yuv2NBPS(10, BE, 1, 10, int16_t)

  236. yuv2NBPS(10, LE, 0, 10, int16_t)

  237. yuv2NBPS(16, BE, 1, 16, int32_t)

  238. yuv2NBPS(16, LE, 0, 16, int32_t)

  239. 

  240. static void yuv2planeX_8_c(const int16_t *filter, int filterSize,

  241.                            const int16_t **src, uint8_t *dest, int dstW,

  242.                            const uint8_t *dither, int offset)

  243. {

  244.     int i;

  245.     for (i=0; i<dstW; i++) {

  246.         int val = dither[(i + offset) & 7] << 12;

  247.         int j;

  248.         for (j=0; j<filterSize; j++)

  249.             val += src[j][i] * filter[j];

  250. 

  251.         dest[i]= av_clip_uint8(val>>19);

  252.     }

  253. }

  254. 

  255. static void yuv2plane1_8_c(const int16_t *src, uint8_t *dest, int dstW,

  256.                            const uint8_t *dither, int offset)

  257. {

  258.     int i;

  259.     for (i=0; i<dstW; i++) {

  260.         int val = (src[i] + dither[(i + offset) & 7]) >> 7;

  261.         dest[i]= av_clip_uint8(val);

  262.     }

  263. }

  264. 

  265. static void yuv2nv12cX_c(SwsContext *c, const int16_t *chrFilter, int chrFilterSize,

  266.                         const int16_t **chrUSrc, const int16_t **chrVSrc,

  267.                         uint8_t *dest, int chrDstW)

  268. {

  269.     enum PixelFormat dstFormat = c->dstFormat;

  270.     const uint8_t *chrDither = c->chrDither8;

  271.     int i;

  272. 

  273.     if (dstFormat == PIX_FMT_NV12)

  274.         for (i=0; i<chrDstW; i++) {

  275.             int u = chrDither[i & 7] << 12;

  276.             int v = chrDither[(i + 3) & 7] << 12;

  277.             int j;

  278.             for (j=0; j<chrFilterSize; j++) {

  279.                 u += chrUSrc[j][i] * chrFilter[j];

  280.                 v += chrVSrc[j][i] * chrFilter[j];

  281.             }

  282. 

  283.             dest[2*i]= av_clip_uint8(u>>19);

  284.             dest[2*i+1]= av_clip_uint8(v>>19);

  285.         }

  286.     else

  287.         for (i=0; i<chrDstW; i++) {

  288.             int u = chrDither[i & 7] << 12;

  289.             int v = chrDither[(i + 3) & 7] << 12;

  290.             int j;

  291.             for (j=0; j<chrFilterSize; j++) {

  292.                 u += chrUSrc[j][i] * chrFilter[j];

  293.                 v += chrVSrc[j][i] * chrFilter[j];

  294.             }

  295. 

  296.             dest[2*i]= av_clip_uint8(v>>19);

  297.             dest[2*i+1]= av_clip_uint8(u>>19);

  298.         }

  299. }

  300. 

  301. #define accumulate_bit(acc, val) \

  302.     acc <<= 1; \

  303.     acc |= (val) >= (128 + 110)

  304. #define output_pixel(pos, acc) \

  305.     if (target == PIX_FMT_MONOBLACK) { \

  306.         pos = acc; \

  307.     } else { \

  308.         pos = ~acc; \

  309.     }

  310. 

  311. static av_always_inline void

  312. yuv2mono_X_c_template(SwsContext *c, const int16_t *lumFilter,

  313.                       const int16_t **lumSrc, int lumFilterSize,

  314.                       const int16_t *chrFilter, const int16_t **chrUSrc,

  315.                       const int16_t **chrVSrc, int chrFilterSize,

  316.                       const int16_t **alpSrc, uint8_t *dest, int dstW,

  317.                       int y, enum PixelFormat target)

  318. {

  319.     const uint8_t * const d128=dither_8x8_220[y&7];

  320.     int i;

  321.     unsigned acc = 0;

  322. 

  323.     for (i = 0; i < dstW - 1; i += 2) {

  324.         int j;

  325.         int Y1 = 1 << 18;

  326.         int Y2 = 1 << 18;

  327. 

  328.         for (j = 0; j < lumFilterSize; j++) {

  329.             Y1 += lumSrc[j][i]   * lumFilter[j];

  330.             Y2 += lumSrc[j][i+1] * lumFilter[j];

  331.         }

  332.         Y1 >>= 19;

  333.         Y2 >>= 19;

  334.         if ((Y1 | Y2) & 0x100) {

  335.             Y1 = av_clip_uint8(Y1);

  336.             Y2 = av_clip_uint8(Y2);

  337.         }

  338.         accumulate_bit(acc, Y1 + d128[(i + 0) & 7]);

  339.         accumulate_bit(acc, Y2 + d128[(i + 1) & 7]);

  340.         if ((i & 7) == 6) {

  341.             output_pixel(*dest++, acc);

  342.         }

  343.     }

  344. }

  345. 

  346. static av_always_inline void

  347. yuv2mono_2_c_template(SwsContext *c, const int16_t *buf[2],

  348.                       const int16_t *ubuf[2], const int16_t *vbuf[2],

  349.                       const int16_t *abuf[2], uint8_t *dest, int dstW,

  350.                       int yalpha, int uvalpha, int y,

  351.                       enum PixelFormat target)

  352. {

  353.     const int16_t *buf0  = buf[0],  *buf1  = buf[1];

  354.     const uint8_t * const d128 = dither_8x8_220[y & 7];

  355.     int  yalpha1 = 4095 - yalpha;

  356.     int i;

  357. 

  358.     for (i = 0; i < dstW - 7; i += 8) {

  359.         int Y, acc = 0;

  360. 

  361.         Y = (buf0[i + 0] * yalpha1 + buf1[i + 0] * yalpha) >> 19;

  362.         accumulate_bit(acc, Y + d128[0]);

  363.         Y = (buf0[i + 1] * yalpha1 + buf1[i + 1] * yalpha) >> 19;

  364.         accumulate_bit(acc, Y + d128[1]);

  365.         Y = (buf0[i + 2] * yalpha1 + buf1[i + 2] * yalpha) >> 19;

  366.         accumulate_bit(acc, Y + d128[2]);

  367.         Y = (buf0[i + 3] * yalpha1 + buf1[i + 3] * yalpha) >> 19;

  368.         accumulate_bit(acc, Y + d128[3]);

  369.         Y = (buf0[i + 4] * yalpha1 + buf1[i + 4] * yalpha) >> 19;

  370.         accumulate_bit(acc, Y + d128[4]);

  371.         Y = (buf0[i + 5] * yalpha1 + buf1[i + 5] * yalpha) >> 19;

  372.         accumulate_bit(acc, Y + d128[5]);

  373.         Y = (buf0[i + 6] * yalpha1 + buf1[i + 6] * yalpha) >> 19;

  374.         accumulate_bit(acc, Y + d128[6]);

  375.         Y = (buf0[i + 7] * yalpha1 + buf1[i + 7] * yalpha) >> 19;

  376.         accumulate_bit(acc, Y + d128[7]);

  377. 

  378.         output_pixel(*dest++, acc);

  379.     }

  380. }

  381. 

  382. static av_always_inline void

  383. yuv2mono_1_c_template(SwsContext *c, const int16_t *buf0,

  384.                       const int16_t *ubuf[2], const int16_t *vbuf[2],

  385.                       const int16_t *abuf0, uint8_t *dest, int dstW,

  386.                       int uvalpha, int y, enum PixelFormat target)

  387. {

  388.     const uint8_t * const d128 = dither_8x8_220[y & 7];

  389.     int i;

  390. 

  391.     for (i = 0; i < dstW - 7; i += 8) {

  392.         int acc = 0;

  393. 

  394.         accumulate_bit(acc, ((buf0[i + 0] + 64) >> 7) + d128[0]);

  395.         accumulate_bit(acc, ((buf0[i + 1] + 64) >> 7) + d128[1]);

  396.         accumulate_bit(acc, ((buf0[i + 2] + 64) >> 7) + d128[2]);

  397.         accumulate_bit(acc, ((buf0[i + 3] + 64) >> 7) + d128[3]);

  398.         accumulate_bit(acc, ((buf0[i + 4] + 64) >> 7) + d128[4]);

  399.         accumulate_bit(acc, ((buf0[i + 5] + 64) >> 7) + d128[5]);

  400.         accumulate_bit(acc, ((buf0[i + 6] + 64) >> 7) + d128[6]);

  401.         accumulate_bit(acc, ((buf0[i + 7] + 64) >> 7) + d128[7]);

  402. 

  403.         output_pixel(*dest++, acc);

  404.     }

  405. }

  406. 

  407. #undef output_pixel

  408. #undef accumulate_bit

  409. 

  410. #define YUV2PACKEDWRAPPER(name, base, ext, fmt) \

  411. static void name ## ext ## _X_c(SwsContext *c, const int16_t *lumFilter, \

  412.                                 const int16_t **lumSrc, int lumFilterSize, \

  413.                                 const int16_t *chrFilter, const int16_t **chrUSrc, \

  414.                                 const int16_t **chrVSrc, int chrFilterSize, \

  415.                                 const int16_t **alpSrc, uint8_t *dest, int dstW, \

  416.                                 int y) \

  417. { \

  418.     name ## base ## _X_c_template(c, lumFilter, lumSrc, lumFilterSize, \

  419.                                   chrFilter, chrUSrc, chrVSrc, chrFilterSize, \

  420.                                   alpSrc, dest, dstW, y, fmt); \

  421. } \

  422.  \

  423. static void name ## ext ## _2_c(SwsContext *c, const int16_t *buf[2], \

  424.                                 const int16_t *ubuf[2], const int16_t *vbuf[2], \

  425.                                 const int16_t *abuf[2], uint8_t *dest, int dstW, \

  426.                                 int yalpha, int uvalpha, int y) \

  427. { \

  428.     name ## base ## _2_c_template(c, buf, ubuf, vbuf, abuf, \

  429.                                   dest, dstW, yalpha, uvalpha, y, fmt); \

  430. } \

  431.  \

  432. static void name ## ext ## _1_c(SwsContext *c, const int16_t *buf0, \

  433.                                 const int16_t *ubuf[2], const int16_t *vbuf[2], \

  434.                                 const int16_t *abuf0, uint8_t *dest, int dstW, \

  435.                                 int uvalpha, int y) \

  436. { \

  437.     name ## base ## _1_c_template(c, buf0, ubuf, vbuf, \

  438.                                   abuf0, dest, dstW, uvalpha, \

  439.                                   y, fmt); \

  440. }

  441. 

  442. YUV2PACKEDWRAPPER(yuv2mono,, white, PIX_FMT_MONOWHITE)

  443. YUV2PACKEDWRAPPER(yuv2mono,, black, PIX_FMT_MONOBLACK)

  444. 

  445. #define output_pixels(pos, Y1, U, Y2, V) \

  446.     if (target == PIX_FMT_YUYV422) { \

  447.         dest[pos + 0] = Y1; \

  448.         dest[pos + 1] = U;  \

  449.         dest[pos + 2] = Y2; \

  450.         dest[pos + 3] = V;  \

  451.     } else { \

  452.         dest[pos + 0] = U;  \

  453.         dest[pos + 1] = Y1; \

  454.         dest[pos + 2] = V;  \

  455.         dest[pos + 3] = Y2; \

  456.     }

  457. 

  458. static av_always_inline void

  459. yuv2422_X_c_template(SwsContext *c, const int16_t *lumFilter,

  460.                      const int16_t **lumSrc, int lumFilterSize,

  461.                      const int16_t *chrFilter, const int16_t **chrUSrc,

  462.                      const int16_t **chrVSrc, int chrFilterSize,

  463.                      const int16_t **alpSrc, uint8_t *dest, int dstW,

  464.                      int y, enum PixelFormat target)

  465. {

  466.     int i;

  467. 

  468.     for (i = 0; i < (dstW >> 1); i++) {

  469.         int j;

  470.         int Y1 = 1 << 18;

  471.         int Y2 = 1 << 18;

  472.         int U  = 1 << 18;

  473.         int V  = 1 << 18;

  474. 

  475.         for (j = 0; j < lumFilterSize; j++) {

  476.             Y1 += lumSrc[j][i * 2]     * lumFilter[j];

  477.             Y2 += lumSrc[j][i * 2 + 1] * lumFilter[j];

  478.         }

  479.         for (j = 0; j < chrFilterSize; j++) {

  480.             U += chrUSrc[j][i] * chrFilter[j];

  481.             V += chrVSrc[j][i] * chrFilter[j];

  482.         }

  483.         Y1 >>= 19;

  484.         Y2 >>= 19;

  485.         U  >>= 19;

  486.         V  >>= 19;

  487.         if ((Y1 | Y2 | U | V) & 0x100) {

  488.             Y1 = av_clip_uint8(Y1);

  489.             Y2 = av_clip_uint8(Y2);

  490.             U  = av_clip_uint8(U);

  491.             V  = av_clip_uint8(V);

  492.         }

  493.         output_pixels(4*i, Y1, U, Y2, V);

  494.     }

  495. }

  496. 

  497. static av_always_inline void

  498. yuv2422_2_c_template(SwsContext *c, const int16_t *buf[2],

  499.                      const int16_t *ubuf[2], const int16_t *vbuf[2],

  500.                      const int16_t *abuf[2], uint8_t *dest, int dstW,

  501.                      int yalpha, int uvalpha, int y,

  502.                      enum PixelFormat target)

  503. {

  504.     const int16_t *buf0  = buf[0],  *buf1  = buf[1],

  505.                   *ubuf0 = ubuf[0], *ubuf1 = ubuf[1],

  506.                   *vbuf0 = vbuf[0], *vbuf1 = vbuf[1];

  507.     int  yalpha1 = 4095 - yalpha;

  508.     int uvalpha1 = 4095 - uvalpha;

  509.     int i;

  510. 

  511.     for (i = 0; i < (dstW >> 1); i++) {

  512.         int Y1 = (buf0[i * 2]     * yalpha1  + buf1[i * 2]     * yalpha)  >> 19;

  513.         int Y2 = (buf0[i * 2 + 1] * yalpha1  + buf1[i * 2 + 1] * yalpha)  >> 19;

  514.         int U  = (ubuf0[i]        * uvalpha1 + ubuf1[i]        * uvalpha) >> 19;

  515.         int V  = (vbuf0[i]        * uvalpha1 + vbuf1[i]        * uvalpha) >> 19;

  516. 

  517.         if ((Y1 | Y2 | U | V) & 0x100) {

  518.             Y1 = av_clip_uint8(Y1);

  519.             Y2 = av_clip_uint8(Y2);

  520.             U  = av_clip_uint8(U);

  521.             V  = av_clip_uint8(V);

  522.         }

  523. 

  524.         output_pixels(i * 4, Y1, U, Y2, V);

  525.     }

  526. }

  527. 

  528. static av_always_inline void

  529. yuv2422_1_c_template(SwsContext *c, const int16_t *buf0,

  530.                      const int16_t *ubuf[2], const int16_t *vbuf[2],

  531.                      const int16_t *abuf0, uint8_t *dest, int dstW,

  532.                      int uvalpha, int y, enum PixelFormat target)

  533. {

  534.     const int16_t *ubuf0 = ubuf[0], *vbuf0 = vbuf[0];

  535.     int i;

  536. 

  537.     if (uvalpha < 2048) {

  538.         for (i = 0; i < (dstW >> 1); i++) {

  539.             int Y1 = (buf0[i * 2    ]+64) >> 7;

  540.             int Y2 = (buf0[i * 2 + 1]+64) >> 7;

  541.             int U  = (ubuf0[i]       +64) >> 7;

  542.             int V  = (vbuf0[i]       +64) >> 7;

  543. 

  544.             if ((Y1 | Y2 | U | V) & 0x100) {

  545.                 Y1 = av_clip_uint8(Y1);

  546.                 Y2 = av_clip_uint8(Y2);

  547.                 U  = av_clip_uint8(U);

  548.                 V  = av_clip_uint8(V);

  549.             }

  550. 

  551.             output_pixels(i * 4, Y1, U, Y2, V);

  552.         }

  553.     } else {

  554.         const int16_t *ubuf1 = ubuf[1], *vbuf1 = vbuf[1];

  555.         for (i = 0; i < (dstW >> 1); i++) {

  556.             int Y1 = (buf0[i * 2    ]    + 64) >> 7;

  557.             int Y2 = (buf0[i * 2 + 1]    + 64) >> 7;

  558.             int U  = (ubuf0[i] + ubuf1[i]+128) >> 8;

  559.             int V  = (vbuf0[i] + vbuf1[i]+128) >> 8;

  560. 

  561.             if ((Y1 | Y2 | U | V) & 0x100) {

  562.                 Y1 = av_clip_uint8(Y1);

  563.                 Y2 = av_clip_uint8(Y2);

  564.                 U  = av_clip_uint8(U);

  565.                 V  = av_clip_uint8(V);

  566.             }

  567. 

  568.             output_pixels(i * 4, Y1, U, Y2, V);

  569.         }

  570.     }

  571. }

  572. 

  573. #undef output_pixels

  574. 

  575. YUV2PACKEDWRAPPER(yuv2, 422, yuyv422, PIX_FMT_YUYV422)

  576. YUV2PACKEDWRAPPER(yuv2, 422, uyvy422, PIX_FMT_UYVY422)

  577. 

  578. #define R_B ((target == PIX_FMT_RGB48LE || target == PIX_FMT_RGB48BE) ? R : B)

  579. #define B_R ((target == PIX_FMT_RGB48LE || target == PIX_FMT_RGB48BE) ? B : R)

  580. #define output_pixel(pos, val) \

  581.     if (isBE(target)) { \

  582.         AV_WB16(pos, val); \

  583.     } else { \

  584.         AV_WL16(pos, val); \

  585.     }

  586. 

  587. static av_always_inline void

  588. yuv2rgb48_X_c_template(SwsContext *c, const int16_t *lumFilter,

  589.                        const int32_t **lumSrc, int lumFilterSize,

  590.                        const int16_t *chrFilter, const int32_t **chrUSrc,

  591.                        const int32_t **chrVSrc, int chrFilterSize,

  592.                        const int32_t **alpSrc, uint16_t *dest, int dstW,

  593.                        int y, enum PixelFormat target)

  594. {

  595.     int i;

  596. 

  597.     for (i = 0; i < (dstW >> 1); i++) {

  598.         int j;

  599.         int Y1 = -0x40000000;

  600.         int Y2 = -0x40000000;

  601.         int U  = -128 << 23; // 19

  602.         int V  = -128 << 23;

  603.         int R, G, B;

  604. 

  605.         for (j = 0; j < lumFilterSize; j++) {

  606.             Y1 += lumSrc[j][i * 2]     * lumFilter[j];

  607.             Y2 += lumSrc[j][i * 2 + 1] * lumFilter[j];

  608.         }

  609.         for (j = 0; j < chrFilterSize; j++) {

  610.             U += chrUSrc[j][i] * chrFilter[j];

  611.             V += chrVSrc[j][i] * chrFilter[j];

  612.         }

  613. 

  614.         // 8bit: 12+15=27; 16-bit: 12+19=31

  615.         Y1 >>= 14; // 10

  616.         Y1 += 0x10000;

  617.         Y2 >>= 14;

  618.         Y2 += 0x10000;

  619.         U  >>= 14;

  620.         V  >>= 14;

  621. 

  622.         // 8bit: 27 -> 17bit, 16bit: 31 - 14 = 17bit

  623.         Y1 -= c->yuv2rgb_y_offset;

  624.         Y2 -= c->yuv2rgb_y_offset;

  625.         Y1 *= c->yuv2rgb_y_coeff;

  626.         Y2 *= c->yuv2rgb_y_coeff;

  627.         Y1 += 1 << 13; // 21

  628.         Y2 += 1 << 13;

  629.         // 8bit: 17 + 13bit = 30bit, 16bit: 17 + 13bit = 30bit

  630. 

  631.         R = V * c->yuv2rgb_v2r_coeff;

  632.         G = V * c->yuv2rgb_v2g_coeff + U * c->yuv2rgb_u2g_coeff;

  633.         B =                            U * c->yuv2rgb_u2b_coeff;

  634. 

  635.         // 8bit: 30 - 22 = 8bit, 16bit: 30bit - 14 = 16bit

  636.         output_pixel(&dest[0], av_clip_uintp2(R_B + Y1, 30) >> 14);

  637.         output_pixel(&dest[1], av_clip_uintp2(  G + Y1, 30) >> 14);

  638.         output_pixel(&dest[2], av_clip_uintp2(B_R + Y1, 30) >> 14);

  639.         output_pixel(&dest[3], av_clip_uintp2(R_B + Y2, 30) >> 14);

  640.         output_pixel(&dest[4], av_clip_uintp2(  G + Y2, 30) >> 14);

  641.         output_pixel(&dest[5], av_clip_uintp2(B_R + Y2, 30) >> 14);

  642.         dest += 6;

  643.     }

  644. }

  645. 

  646. static av_always_inline void

  647. yuv2rgb48_2_c_template(SwsContext *c, const int32_t *buf[2],

  648.                        const int32_t *ubuf[2], const int32_t *vbuf[2],

  649.                        const int32_t *abuf[2], uint16_t *dest, int dstW,

  650.                        int yalpha, int uvalpha, int y,

  651.                        enum PixelFormat target)

  652. {

  653.     const int32_t *buf0  = buf[0],  *buf1  = buf[1],

  654.                   *ubuf0 = ubuf[0], *ubuf1 = ubuf[1],

  655.                   *vbuf0 = vbuf[0], *vbuf1 = vbuf[1];

  656.     int  yalpha1 = 4095 - yalpha;

  657.     int uvalpha1 = 4095 - uvalpha;

  658.     int i;

  659. 

  660.     for (i = 0; i < (dstW >> 1); i++) {

  661.         int Y1 = (buf0[i * 2]     * yalpha1  + buf1[i * 2]     * yalpha) >> 14;

  662.         int Y2 = (buf0[i * 2 + 1] * yalpha1  + buf1[i * 2 + 1] * yalpha) >> 14;

  663.         int U  = (ubuf0[i]        * uvalpha1 + ubuf1[i]        * uvalpha + (-128 << 23)) >> 14;

  664.         int V  = (vbuf0[i]        * uvalpha1 + vbuf1[i]        * uvalpha + (-128 << 23)) >> 14;

  665.         int R, G, B;

  666. 

  667.         Y1 -= c->yuv2rgb_y_offset;

  668.         Y2 -= c->yuv2rgb_y_offset;

  669.         Y1 *= c->yuv2rgb_y_coeff;

  670.         Y2 *= c->yuv2rgb_y_coeff;

  671.         Y1 += 1 << 13;

  672.         Y2 += 1 << 13;

  673. 

  674.         R = V * c->yuv2rgb_v2r_coeff;

  675.         G = V * c->yuv2rgb_v2g_coeff + U * c->yuv2rgb_u2g_coeff;

  676.         B =                            U * c->yuv2rgb_u2b_coeff;

  677. 

  678.         output_pixel(&dest[0], av_clip_uintp2(R_B + Y1, 30) >> 14);

  679.         output_pixel(&dest[1], av_clip_uintp2(  G + Y1, 30) >> 14);

  680.         output_pixel(&dest[2], av_clip_uintp2(B_R + Y1, 30) >> 14);

  681.         output_pixel(&dest[3], av_clip_uintp2(R_B + Y2, 30) >> 14);

  682.         output_pixel(&dest[4], av_clip_uintp2(  G + Y2, 30) >> 14);

  683.         output_pixel(&dest[5], av_clip_uintp2(B_R + Y2, 30) >> 14);

  684.         dest += 6;

  685.     }

  686. }

  687. 

  688. static av_always_inline void

  689. yuv2rgb48_1_c_template(SwsContext *c, const int32_t *buf0,

  690.                        const int32_t *ubuf[2], const int32_t *vbuf[2],

  691.                        const int32_t *abuf0, uint16_t *dest, int dstW,

  692.                        int uvalpha, int y, enum PixelFormat target)

  693. {

  694.     const int32_t *ubuf0 = ubuf[0], *vbuf0 = vbuf[0];

  695.     int i;

  696. 

  697.     if (uvalpha < 2048) {

  698.         for (i = 0; i < (dstW >> 1); i++) {

  699.             int Y1 = (buf0[i * 2]    ) >> 2;

  700.             int Y2 = (buf0[i * 2 + 1]) >> 2;

  701.             int U  = (ubuf0[i] + (-128 << 11)) >> 2;

  702.             int V  = (vbuf0[i] + (-128 << 11)) >> 2;

  703.             int R, G, B;

  704. 

  705.             Y1 -= c->yuv2rgb_y_offset;

  706.             Y2 -= c->yuv2rgb_y_offset;

  707.             Y1 *= c->yuv2rgb_y_coeff;

  708.             Y2 *= c->yuv2rgb_y_coeff;

  709.             Y1 += 1 << 13;

  710.             Y2 += 1 << 13;

  711. 

  712.             R = V * c->yuv2rgb_v2r_coeff;

  713.             G = V * c->yuv2rgb_v2g_coeff + U * c->yuv2rgb_u2g_coeff;

  714.             B =                            U * c->yuv2rgb_u2b_coeff;

  715. 

  716.             output_pixel(&dest[0], av_clip_uintp2(R_B + Y1, 30) >> 14);

  717.             output_pixel(&dest[1], av_clip_uintp2(  G + Y1, 30) >> 14);

  718.             output_pixel(&dest[2], av_clip_uintp2(B_R + Y1, 30) >> 14);

  719.             output_pixel(&dest[3], av_clip_uintp2(R_B + Y2, 30) >> 14);

  720.             output_pixel(&dest[4], av_clip_uintp2(  G + Y2, 30) >> 14);

  721.             output_pixel(&dest[5], av_clip_uintp2(B_R + Y2, 30) >> 14);

  722.             dest += 6;

  723.         }

  724.     } else {

  725.         const int32_t *ubuf1 = ubuf[1], *vbuf1 = vbuf[1];

  726.         for (i = 0; i < (dstW >> 1); i++) {

  727.             int Y1 = (buf0[i * 2]    ) >> 2;

  728.             int Y2 = (buf0[i * 2 + 1]) >> 2;

  729.             int U  = (ubuf0[i] + ubuf1[i] + (-128 << 12)) >> 3;

  730.             int V  = (vbuf0[i] + vbuf1[i] + (-128 << 12)) >> 3;

  731.             int R, G, B;

  732. 

  733.             Y1 -= c->yuv2rgb_y_offset;

  734.             Y2 -= c->yuv2rgb_y_offset;

  735.             Y1 *= c->yuv2rgb_y_coeff;

  736.             Y2 *= c->yuv2rgb_y_coeff;

  737.             Y1 += 1 << 13;

  738.             Y2 += 1 << 13;

  739. 

  740.             R = V * c->yuv2rgb_v2r_coeff;

  741.             G = V * c->yuv2rgb_v2g_coeff + U * c->yuv2rgb_u2g_coeff;

  742.             B =                            U * c->yuv2rgb_u2b_coeff;

  743. 

  744.             output_pixel(&dest[0], av_clip_uintp2(R_B + Y1, 30) >> 14);

  745.             output_pixel(&dest[1], av_clip_uintp2(  G + Y1, 30) >> 14);

  746.             output_pixel(&dest[2], av_clip_uintp2(B_R + Y1, 30) >> 14);

  747.             output_pixel(&dest[3], av_clip_uintp2(R_B + Y2, 30) >> 14);

  748.             output_pixel(&dest[4], av_clip_uintp2(  G + Y2, 30) >> 14);

  749.             output_pixel(&dest[5], av_clip_uintp2(B_R + Y2, 30) >> 14);

  750.             dest += 6;

  751.         }

  752.     }

  753. }

  754. 

  755. #undef output_pixel

  756. #undef r_b

  757. #undef b_r

  758. 

  759. #define YUV2PACKED16WRAPPER(name, base, ext, fmt) \

  760. static void name ## ext ## _X_c(SwsContext *c, const int16_t *lumFilter, \

  761.                         const int16_t **_lumSrc, int lumFilterSize, \

  762.                         const int16_t *chrFilter, const int16_t **_chrUSrc, \

  763.                         const int16_t **_chrVSrc, int chrFilterSize, \

  764.                         const int16_t **_alpSrc, uint8_t *_dest, int dstW, \

  765.                         int y) \

  766. { \

  767.     const int32_t **lumSrc  = (const int32_t **) _lumSrc, \

  768.                   **chrUSrc = (const int32_t **) _chrUSrc, \

  769.                   **chrVSrc = (const int32_t **) _chrVSrc, \

  770.                   **alpSrc  = (const int32_t **) _alpSrc; \

  771.     uint16_t *dest = (uint16_t *) _dest; \

  772.     name ## base ## _X_c_template(c, lumFilter, lumSrc, lumFilterSize, \

  773.                           chrFilter, chrUSrc, chrVSrc, chrFilterSize, \

  774.                           alpSrc, dest, dstW, y, fmt); \

  775. } \

  776.  \

  777. static void name ## ext ## _2_c(SwsContext *c, const int16_t *_buf[2], \

  778.                         const int16_t *_ubuf[2], const int16_t *_vbuf[2], \

  779.                         const int16_t *_abuf[2], uint8_t *_dest, int dstW, \

  780.                         int yalpha, int uvalpha, int y) \

  781. { \

  782.     const int32_t **buf  = (const int32_t **) _buf, \

  783.                   **ubuf = (const int32_t **) _ubuf, \

  784.                   **vbuf = (const int32_t **) _vbuf, \

  785.                   **abuf = (const int32_t **) _abuf; \

  786.     uint16_t *dest = (uint16_t *) _dest; \

  787.     name ## base ## _2_c_template(c, buf, ubuf, vbuf, abuf, \

  788.                           dest, dstW, yalpha, uvalpha, y, fmt); \

  789. } \

  790.  \

  791. static void name ## ext ## _1_c(SwsContext *c, const int16_t *_buf0, \

  792.                         const int16_t *_ubuf[2], const int16_t *_vbuf[2], \

  793.                         const int16_t *_abuf0, uint8_t *_dest, int dstW, \

  794.                         int uvalpha, int y) \

  795. { \

  796.     const int32_t *buf0  = (const int32_t *)  _buf0, \

  797.                  **ubuf  = (const int32_t **) _ubuf, \

  798.                  **vbuf  = (const int32_t **) _vbuf, \

  799.                   *abuf0 = (const int32_t *)  _abuf0; \

  800.     uint16_t *dest = (uint16_t *) _dest; \

  801.     name ## base ## _1_c_template(c, buf0, ubuf, vbuf, abuf0, dest, \

  802.                                   dstW, uvalpha, y, fmt); \

  803. }

  804. 

  805. YUV2PACKED16WRAPPER(yuv2, rgb48, rgb48be, PIX_FMT_RGB48BE)

  806. YUV2PACKED16WRAPPER(yuv2, rgb48, rgb48le, PIX_FMT_RGB48LE)

  807. YUV2PACKED16WRAPPER(yuv2, rgb48, bgr48be, PIX_FMT_BGR48BE)

  808. YUV2PACKED16WRAPPER(yuv2, rgb48, bgr48le, PIX_FMT_BGR48LE)

  809. 

  810. /*

  811.  * Write out 2 RGB pixels in the target pixel format. This function takes a

  812.  * R/G/B LUT as generated by ff_yuv2rgb_c_init_tables(), which takes care of

  813.  * things like endianness conversion and shifting. The caller takes care of

  814.  * setting the correct offset in these tables from the chroma (U/V) values.

  815.  * This function then uses the luminance (Y1/Y2) values to write out the

  816.  * correct RGB values into the destination buffer.

  817.  */

  818. static av_always_inline void

  819. yuv2rgb_write(uint8_t *_dest, int i, int Y1, int Y2,

  820.               unsigned A1, unsigned A2,

  821.               const void *_r, const void *_g, const void *_b, int y,

  822.               enum PixelFormat target, int hasAlpha)

  823. {

  824.     if (target == PIX_FMT_ARGB || target == PIX_FMT_RGBA ||

  825.         target == PIX_FMT_ABGR || target == PIX_FMT_BGRA) {

  826.         uint32_t *dest = (uint32_t *) _dest;

  827.         const uint32_t *r = (const uint32_t *) _r;

  828.         const uint32_t *g = (const uint32_t *) _g;

  829.         const uint32_t *b = (const uint32_t *) _b;

  830. 

  831. #if CONFIG_SMALL

  832.         int sh = hasAlpha ? ((target == PIX_FMT_RGB32_1 || target == PIX_FMT_BGR32_1) ? 0 : 24) : 0;

  833. 

  834.         dest[i * 2 + 0] = r[Y1] + g[Y1] + b[Y1] + (hasAlpha ? A1 << sh : 0);

  835.         dest[i * 2 + 1] = r[Y2] + g[Y2] + b[Y2] + (hasAlpha ? A2 << sh : 0);

  836. #else

  837.         if (hasAlpha) {

  838.             int sh = (target == PIX_FMT_RGB32_1 || target == PIX_FMT_BGR32_1) ? 0 : 24;

  839. 

  840.             dest[i * 2 + 0] = r[Y1] + g[Y1] + b[Y1] + (A1 << sh);

  841.             dest[i * 2 + 1] = r[Y2] + g[Y2] + b[Y2] + (A2 << sh);

  842.         } else {

  843.             dest[i * 2 + 0] = r[Y1] + g[Y1] + b[Y1];

  844.             dest[i * 2 + 1] = r[Y2] + g[Y2] + b[Y2];

  845.         }

  846. #endif

  847.     } else if (target == PIX_FMT_RGB24 || target == PIX_FMT_BGR24) {

  848.         uint8_t *dest = (uint8_t *) _dest;

  849.         const uint8_t *r = (const uint8_t *) _r;

  850.         const uint8_t *g = (const uint8_t *) _g;

  851.         const uint8_t *b = (const uint8_t *) _b;

  852. 

  853. #define r_b ((target == PIX_FMT_RGB24) ? r : b)

  854. #define b_r ((target == PIX_FMT_RGB24) ? b : r)

  855. 

  856.         dest[i * 6 + 0] = r_b[Y1];

  857.         dest[i * 6 + 1] =   g[Y1];

  858.         dest[i * 6 + 2] = b_r[Y1];

  859.         dest[i * 6 + 3] = r_b[Y2];

  860.         dest[i * 6 + 4] =   g[Y2];

  861.         dest[i * 6 + 5] = b_r[Y2];

  862. #undef r_b

  863. #undef b_r

  864.     } else if (target == PIX_FMT_RGB565 || target == PIX_FMT_BGR565 ||

  865.                target == PIX_FMT_RGB555 || target == PIX_FMT_BGR555 ||

  866.                target == PIX_FMT_RGB444 || target == PIX_FMT_BGR444) {

  867.         uint16_t *dest = (uint16_t *) _dest;

  868.         const uint16_t *r = (const uint16_t *) _r;

  869.         const uint16_t *g = (const uint16_t *) _g;

  870.         const uint16_t *b = (const uint16_t *) _b;

  871.         int dr1, dg1, db1, dr2, dg2, db2;

  872. 

  873.         if (target == PIX_FMT_RGB565 || target == PIX_FMT_BGR565) {

  874.             dr1 = dither_2x2_8[ y & 1     ][0];

  875.             dg1 = dither_2x2_4[ y & 1     ][0];

  876.             db1 = dither_2x2_8[(y & 1) ^ 1][0];

  877.             dr2 = dither_2x2_8[ y & 1     ][1];

  878.             dg2 = dither_2x2_4[ y & 1     ][1];

  879.             db2 = dither_2x2_8[(y & 1) ^ 1][1];

  880.         } else if (target == PIX_FMT_RGB555 || target == PIX_FMT_BGR555) {

  881.             dr1 = dither_2x2_8[ y & 1     ][0];

  882.             dg1 = dither_2x2_8[ y & 1     ][1];

  883.             db1 = dither_2x2_8[(y & 1) ^ 1][0];

  884.             dr2 = dither_2x2_8[ y & 1     ][1];

  885.             dg2 = dither_2x2_8[ y & 1     ][0];

  886.             db2 = dither_2x2_8[(y & 1) ^ 1][1];

  887.         } else {

  888.             dr1 = dither_4x4_16[ y & 3     ][0];

  889.             dg1 = dither_4x4_16[ y & 3     ][1];

  890.             db1 = dither_4x4_16[(y & 3) ^ 3][0];

  891.             dr2 = dither_4x4_16[ y & 3     ][1];

  892.             dg2 = dither_4x4_16[ y & 3     ][0];

  893.             db2 = dither_4x4_16[(y & 3) ^ 3][1];

  894.         }

  895. 

  896.         dest[i * 2 + 0] = r[Y1 + dr1] + g[Y1 + dg1] + b[Y1 + db1];

  897.         dest[i * 2 + 1] = r[Y2 + dr2] + g[Y2 + dg2] + b[Y2 + db2];

  898.     } else /* 8/4-bit */ {

  899.         uint8_t *dest = (uint8_t *) _dest;

  900.         const uint8_t *r = (const uint8_t *) _r;

  901.         const uint8_t *g = (const uint8_t *) _g;

  902.         const uint8_t *b = (const uint8_t *) _b;

  903.         int dr1, dg1, db1, dr2, dg2, db2;

  904. 

  905.         if (target == PIX_FMT_RGB8 || target == PIX_FMT_BGR8) {

  906.             const uint8_t * const d64 = dither_8x8_73[y & 7];

  907.             const uint8_t * const d32 = dither_8x8_32[y & 7];

  908.             dr1 = dg1 = d32[(i * 2 + 0) & 7];

  909.             db1 =       d64[(i * 2 + 0) & 7];

  910.             dr2 = dg2 = d32[(i * 2 + 1) & 7];

  911.             db2 =       d64[(i * 2 + 1) & 7];

  912.         } else {

  913.             const uint8_t * const d64  = dither_8x8_73 [y & 7];

  914.             const uint8_t * const d128 = dither_8x8_220[y & 7];

  915.             dr1 = db1 = d128[(i * 2 + 0) & 7];

  916.             dg1 =        d64[(i * 2 + 0) & 7];

  917.             dr2 = db2 = d128[(i * 2 + 1) & 7];

  918.             dg2 =        d64[(i * 2 + 1) & 7];

  919.         }

  920. 

  921.         if (target == PIX_FMT_RGB4 || target == PIX_FMT_BGR4) {

  922.             dest[i] = r[Y1 + dr1] + g[Y1 + dg1] + b[Y1 + db1] +

  923.                     ((r[Y2 + dr2] + g[Y2 + dg2] + b[Y2 + db2]) << 4);

  924.         } else {

  925.             dest[i * 2 + 0] = r[Y1 + dr1] + g[Y1 + dg1] + b[Y1 + db1];

  926.             dest[i * 2 + 1] = r[Y2 + dr2] + g[Y2 + dg2] + b[Y2 + db2];

  927.         }

  928.     }

  929. }

  930. 

  931. static av_always_inline void

  932. yuv2rgb_X_c_template(SwsContext *c, const int16_t *lumFilter,

  933.                      const int16_t **lumSrc, int lumFilterSize,

  934.                      const int16_t *chrFilter, const int16_t **chrUSrc,

  935.                      const int16_t **chrVSrc, int chrFilterSize,

  936.                      const int16_t **alpSrc, uint8_t *dest, int dstW,

  937.                      int y, enum PixelFormat target, int hasAlpha)

  938. {

  939.     int i;

  940. 

  941.     for (i = 0; i < (dstW >> 1); i++) {

  942.         int j, A1, A2;

  943.         int Y1 = 1 << 18;

  944.         int Y2 = 1 << 18;

  945.         int U  = 1 << 18;

  946.         int V  = 1 << 18;

  947.         const void *r, *g, *b;

  948. 

  949.         for (j = 0; j < lumFilterSize; j++) {

  950.             Y1 += lumSrc[j][i * 2]     * lumFilter[j];

  951.             Y2 += lumSrc[j][i * 2 + 1] * lumFilter[j];

  952.         }

  953.         for (j = 0; j < chrFilterSize; j++) {

  954.             U += chrUSrc[j][i] * chrFilter[j];

  955.             V += chrVSrc[j][i] * chrFilter[j];

  956.         }

  957.         Y1 >>= 19;

  958.         Y2 >>= 19;

  959.         U  >>= 19;

  960.         V  >>= 19;

  961.         if (hasAlpha) {

  962.             A1 = 1 << 18;

  963.             A2 = 1 << 18;

  964.             for (j = 0; j < lumFilterSize; j++) {

  965.                 A1 += alpSrc[j][i * 2    ] * lumFilter[j];

  966.                 A2 += alpSrc[j][i * 2 + 1] * lumFilter[j];

  967.             }

  968.             A1 >>= 19;

  969.             A2 >>= 19;

  970.             if ((A1 | A2) & 0x100) {

  971.                 A1 = av_clip_uint8(A1);

  972.                 A2 = av_clip_uint8(A2);

  973.             }

  974.         }

  975. 

  976.         r =  c->table_rV[V + YUVRGB_TABLE_HEADROOM];

  977.         g = (c->table_gU[U + YUVRGB_TABLE_HEADROOM] + c->table_gV[V + YUVRGB_TABLE_HEADROOM]);

  978.         b =  c->table_bU[U + YUVRGB_TABLE_HEADROOM];

  979. 

  980.         yuv2rgb_write(dest, i, Y1, Y2, hasAlpha ? A1 : 0, hasAlpha ? A2 : 0,

  981.                       r, g, b, y, target, hasAlpha);

  982.     }

  983. }

  984. 

  985. static av_always_inline void

  986. yuv2rgb_2_c_template(SwsContext *c, const int16_t *buf[2],

  987.                      const int16_t *ubuf[2], const int16_t *vbuf[2],

  988.                      const int16_t *abuf[2], uint8_t *dest, int dstW,

  989.                      int yalpha, int uvalpha, int y,

  990.                      enum PixelFormat target, int hasAlpha)

  991. {

  992.     const int16_t *buf0  = buf[0],  *buf1  = buf[1],

  993.                   *ubuf0 = ubuf[0], *ubuf1 = ubuf[1],

  994.                   *vbuf0 = vbuf[0], *vbuf1 = vbuf[1],

  995.                   *abuf0 = hasAlpha ? abuf[0] : NULL,

  996.                   *abuf1 = hasAlpha ? abuf[1] : NULL;

  997.     int  yalpha1 = 4095 - yalpha;

  998.     int uvalpha1 = 4095 - uvalpha;

  999.     int i;

  1000. 

  1001.     for (i = 0; i < (dstW >> 1); i++) {

  1002.         int Y1 = (buf0[i * 2]     * yalpha1  + buf1[i * 2]     * yalpha)  >> 19;

  1003.         int Y2 = (buf0[i * 2 + 1] * yalpha1  + buf1[i * 2 + 1] * yalpha)  >> 19;

  1004.         int U  = (ubuf0[i]        * uvalpha1 + ubuf1[i]        * uvalpha) >> 19;

  1005.         int V  = (vbuf0[i]        * uvalpha1 + vbuf1[i]        * uvalpha) >> 19;

  1006.         int A1, A2;

  1007.         const void *r =  c->table_rV[V + YUVRGB_TABLE_HEADROOM],

  1008.                    *g = (c->table_gU[U + YUVRGB_TABLE_HEADROOM] + c->table_gV[V + YUVRGB_TABLE_HEADROOM]),

  1009.                    *b =  c->table_bU[U + YUVRGB_TABLE_HEADROOM];

  1010. 

  1011.         if (hasAlpha) {

  1012.             A1 = (abuf0[i * 2    ] * yalpha1 + abuf1[i * 2    ] * yalpha) >> 19;

  1013.             A2 = (abuf0[i * 2 + 1] * yalpha1 + abuf1[i * 2 + 1] * yalpha) >> 19;

  1014.         }

  1015. 

  1016.         yuv2rgb_write(dest, i, Y1, Y2, hasAlpha ? A1 : 0, hasAlpha ? A2 : 0,

  1017.                       r, g, b, y, target, hasAlpha);

  1018.     }

  1019. }

  1020. 

  1021. static av_always_inline void

  1022. yuv2rgb_1_c_template(SwsContext *c, const int16_t *buf0,

  1023.                      const int16_t *ubuf[2], const int16_t *vbuf[2],

  1024.                      const int16_t *abuf0, uint8_t *dest, int dstW,

  1025.                      int uvalpha, int y, enum PixelFormat target,

  1026.                      int hasAlpha)

  1027. {

  1028.     const int16_t *ubuf0 = ubuf[0], *vbuf0 = vbuf[0];

  1029.     int i;

  1030. 

  1031.     if (uvalpha < 2048) {

  1032.         for (i = 0; i < (dstW >> 1); i++) {

  1033.             int Y1 = (buf0[i * 2    ] + 64) >> 7;

  1034.             int Y2 = (buf0[i * 2 + 1] + 64) >> 7;

  1035.             int U  = (ubuf0[i]        + 64) >> 7;

  1036.             int V  = (vbuf0[i]        + 64) >> 7;

  1037.             int A1, A2;

  1038.             const void *r =  c->table_rV[V + YUVRGB_TABLE_HEADROOM],

  1039.                        *g = (c->table_gU[U + YUVRGB_TABLE_HEADROOM] + c->table_gV[V + YUVRGB_TABLE_HEADROOM]),

  1040.                        *b =  c->table_bU[U + YUVRGB_TABLE_HEADROOM];

  1041. 

  1042.             if (hasAlpha) {

  1043.                 A1 = abuf0[i * 2    ] * 255 + 16384 >> 15;

  1044.                 A2 = abuf0[i * 2 + 1] * 255 + 16384 >> 15;

  1045.             }

  1046. 

  1047.             yuv2rgb_write(dest, i, Y1, Y2, hasAlpha ? A1 : 0, hasAlpha ? A2 : 0,

  1048.                           r, g, b, y, target, hasAlpha);

  1049.         }

  1050.     } else {

  1051.         const int16_t *ubuf1 = ubuf[1], *vbuf1 = vbuf[1];

  1052.         for (i = 0; i < (dstW >> 1); i++) {

  1053.             int Y1 = (buf0[i * 2    ]     +  64) >> 7;

  1054.             int Y2 = (buf0[i * 2 + 1]     +  64) >> 7;

  1055.             int U  = (ubuf0[i] + ubuf1[i] + 128) >> 8;

  1056.             int V  = (vbuf0[i] + vbuf1[i] + 128) >> 8;

  1057.             int A1, A2;

  1058.             const void *r =  c->table_rV[V + YUVRGB_TABLE_HEADROOM],

  1059.                        *g = (c->table_gU[U + YUVRGB_TABLE_HEADROOM] + c->table_gV[V + YUVRGB_TABLE_HEADROOM]),

  1060.                        *b =  c->table_bU[U + YUVRGB_TABLE_HEADROOM];

  1061. 

  1062.             if (hasAlpha) {

  1063.                 A1 = (abuf0[i * 2    ] + 64) >> 7;

  1064.                 A2 = (abuf0[i * 2 + 1] + 64) >> 7;

  1065.             }

  1066. 

  1067.             yuv2rgb_write(dest, i, Y1, Y2, hasAlpha ? A1 : 0, hasAlpha ? A2 : 0,

  1068.                           r, g, b, y, target, hasAlpha);

  1069.         }

  1070.     }

  1071. }

  1072. 

  1073. #define YUV2RGBWRAPPERX(name, base, ext, fmt, hasAlpha) \

  1074. static void name ## ext ## _X_c(SwsContext *c, const int16_t *lumFilter, \

  1075.                                 const int16_t **lumSrc, int lumFilterSize, \

  1076.                                 const int16_t *chrFilter, const int16_t **chrUSrc, \

  1077.                                 const int16_t **chrVSrc, int chrFilterSize, \

  1078.                                 const int16_t **alpSrc, uint8_t *dest, int dstW, \

  1079.                                 int y) \

  1080. { \

  1081.     name ## base ## _X_c_template(c, lumFilter, lumSrc, lumFilterSize, \

  1082.                                   chrFilter, chrUSrc, chrVSrc, chrFilterSize, \

  1083.                                   alpSrc, dest, dstW, y, fmt, hasAlpha); \

  1084. }

  1085. #define YUV2RGBWRAPPER(name, base, ext, fmt, hasAlpha) \

  1086. YUV2RGBWRAPPERX(name, base, ext, fmt, hasAlpha) \

  1087. static void name ## ext ## _2_c(SwsContext *c, const int16_t *buf[2], \

  1088.                                 const int16_t *ubuf[2], const int16_t *vbuf[2], \

  1089.                                 const int16_t *abuf[2], uint8_t *dest, int dstW, \

  1090.                                 int yalpha, int uvalpha, int y) \

  1091. { \

  1092.     name ## base ## _2_c_template(c, buf, ubuf, vbuf, abuf, \

  1093.                                   dest, dstW, yalpha, uvalpha, y, fmt, hasAlpha); \

  1094. } \

  1095.  \

  1096. static void name ## ext ## _1_c(SwsContext *c, const int16_t *buf0, \

  1097.                                 const int16_t *ubuf[2], const int16_t *vbuf[2], \

  1098.                                 const int16_t *abuf0, uint8_t *dest, int dstW, \

  1099.                                 int uvalpha, int y) \

  1100. { \

  1101.     name ## base ## _1_c_template(c, buf0, ubuf, vbuf, abuf0, dest, \

  1102.                                   dstW, uvalpha, y, fmt, hasAlpha); \

  1103. }

  1104. 

  1105. #if CONFIG_SMALL

  1106. YUV2RGBWRAPPER(yuv2rgb,,  32_1,  PIX_FMT_RGB32_1,   CONFIG_SWSCALE_ALPHA && c->alpPixBuf)

  1107. YUV2RGBWRAPPER(yuv2rgb,,  32,    PIX_FMT_RGB32,     CONFIG_SWSCALE_ALPHA && c->alpPixBuf)

  1108. #else

  1109. #if CONFIG_SWSCALE_ALPHA

  1110. YUV2RGBWRAPPER(yuv2rgb,, a32_1,  PIX_FMT_RGB32_1,   1)

  1111. YUV2RGBWRAPPER(yuv2rgb,, a32,    PIX_FMT_RGB32,     1)

  1112. #endif

  1113. YUV2RGBWRAPPER(yuv2rgb,, x32_1,  PIX_FMT_RGB32_1,   0)

  1114. YUV2RGBWRAPPER(yuv2rgb,, x32,    PIX_FMT_RGB32,     0)

  1115. #endif

  1116. YUV2RGBWRAPPER(yuv2, rgb, rgb24, PIX_FMT_RGB24,   0)

  1117. YUV2RGBWRAPPER(yuv2, rgb, bgr24, PIX_FMT_BGR24,   0)

  1118. YUV2RGBWRAPPER(yuv2rgb,,  16,    PIX_FMT_RGB565,    0)

  1119. YUV2RGBWRAPPER(yuv2rgb,,  15,    PIX_FMT_RGB555,    0)

  1120. YUV2RGBWRAPPER(yuv2rgb,,  12,    PIX_FMT_RGB444,    0)

  1121. YUV2RGBWRAPPER(yuv2rgb,,   8,    PIX_FMT_RGB8,      0)

  1122. YUV2RGBWRAPPER(yuv2rgb,,   4,    PIX_FMT_RGB4,      0)

  1123. YUV2RGBWRAPPER(yuv2rgb,,   4b,   PIX_FMT_RGB4_BYTE, 0)

  1124. 

  1125. static av_always_inline void

  1126. yuv2rgb_full_X_c_template(SwsContext *c, const int16_t *lumFilter,

  1127.                           const int16_t **lumSrc, int lumFilterSize,

  1128.                           const int16_t *chrFilter, const int16_t **chrUSrc,

  1129.                           const int16_t **chrVSrc, int chrFilterSize,

  1130.                           const int16_t **alpSrc, uint8_t *dest,

  1131.                           int dstW, int y, enum PixelFormat target, int hasAlpha)

  1132. {

  1133.     int i;

  1134.     int step = (target == PIX_FMT_RGB24 || target == PIX_FMT_BGR24) ? 3 : 4;

  1135. 

  1136.     for (i = 0; i < dstW; i++) {

  1137.         int j;

  1138.         int Y = 1<<9;

  1139.         int U = (1<<9)-(128 << 19);

  1140.         int V = (1<<9)-(128 << 19);

  1141.         int R, G, B, A;

  1142. 

  1143.         for (j = 0; j < lumFilterSize; j++) {

  1144.             Y += lumSrc[j][i] * lumFilter[j];

  1145.         }

  1146.         for (j = 0; j < chrFilterSize; j++) {

  1147.             U += chrUSrc[j][i] * chrFilter[j];

  1148.             V += chrVSrc[j][i] * chrFilter[j];

  1149.         }

  1150.         Y >>= 10;

  1151.         U >>= 10;

  1152.         V >>= 10;

  1153.         if (hasAlpha) {

  1154.             A = 1 << 18;

  1155.             for (j = 0; j < lumFilterSize; j++) {

  1156.                 A += alpSrc[j][i] * lumFilter[j];

  1157.             }

  1158.             A >>= 19;

  1159.             if (A & 0x100)

  1160.                 A = av_clip_uint8(A);

  1161.         }

  1162.         Y -= c->yuv2rgb_y_offset;

  1163.         Y *= c->yuv2rgb_y_coeff;

  1164.         Y += 1 << 21;

  1165.         R = Y + V*c->yuv2rgb_v2r_coeff;

  1166.         G = Y + V*c->yuv2rgb_v2g_coeff + U*c->yuv2rgb_u2g_coeff;

  1167.         B = Y +                          U*c->yuv2rgb_u2b_coeff;

  1168.         if ((R | G | B) & 0xC0000000) {

  1169.             R = av_clip_uintp2(R, 30);

  1170.             G = av_clip_uintp2(G, 30);

  1171.             B = av_clip_uintp2(B, 30);

  1172.         }

  1173. 

  1174.         switch(target) {

  1175.         case PIX_FMT_ARGB:

  1176.             dest[0] = hasAlpha ? A : 255;

  1177.             dest[1] = R >> 22;

  1178.             dest[2] = G >> 22;

  1179.             dest[3] = B >> 22;

  1180.             break;

  1181.         case PIX_FMT_RGB24:

  1182.             dest[0] = R >> 22;

  1183.             dest[1] = G >> 22;

  1184.             dest[2] = B >> 22;

  1185.             break;

  1186.         case PIX_FMT_RGBA:

  1187.             dest[0] = R >> 22;

  1188.             dest[1] = G >> 22;

  1189.             dest[2] = B >> 22;

  1190.             dest[3] = hasAlpha ? A : 255;

  1191.             break;

  1192.         case PIX_FMT_ABGR:

  1193.             dest[0] = hasAlpha ? A : 255;

  1194.             dest[1] = B >> 22;

  1195.             dest[2] = G >> 22;

  1196.             dest[3] = R >> 22;

  1197.             break;

  1198.         case PIX_FMT_BGR24:

  1199.             dest[0] = B >> 22;

  1200.             dest[1] = G >> 22;

  1201.             dest[2] = R >> 22;

  1202.             break;

  1203.         case PIX_FMT_BGRA:

  1204.             dest[0] = B >> 22;

  1205.             dest[1] = G >> 22;

  1206.             dest[2] = R >> 22;

  1207.             dest[3] = hasAlpha ? A : 255;

  1208.             break;

  1209.         }

  1210.         dest += step;

  1211.     }

  1212. }

  1213. 

  1214. #if CONFIG_SMALL

  1215. YUV2RGBWRAPPERX(yuv2, rgb_full, bgra32_full, PIX_FMT_BGRA,  CONFIG_SWSCALE_ALPHA && c->alpPixBuf)

  1216. YUV2RGBWRAPPERX(yuv2, rgb_full, abgr32_full, PIX_FMT_ABGR,  CONFIG_SWSCALE_ALPHA && c->alpPixBuf)

  1217. YUV2RGBWRAPPERX(yuv2, rgb_full, rgba32_full, PIX_FMT_RGBA,  CONFIG_SWSCALE_ALPHA && c->alpPixBuf)

  1218. YUV2RGBWRAPPERX(yuv2, rgb_full, argb32_full, PIX_FMT_ARGB,  CONFIG_SWSCALE_ALPHA && c->alpPixBuf)

  1219. #else

  1220. #if CONFIG_SWSCALE_ALPHA

  1221. YUV2RGBWRAPPERX(yuv2, rgb_full, bgra32_full, PIX_FMT_BGRA,  1)

  1222. YUV2RGBWRAPPERX(yuv2, rgb_full, abgr32_full, PIX_FMT_ABGR,  1)

  1223. YUV2RGBWRAPPERX(yuv2, rgb_full, rgba32_full, PIX_FMT_RGBA,  1)

  1224. YUV2RGBWRAPPERX(yuv2, rgb_full, argb32_full, PIX_FMT_ARGB,  1)

  1225. #endif

  1226. YUV2RGBWRAPPERX(yuv2, rgb_full, bgrx32_full, PIX_FMT_BGRA,  0)

  1227. YUV2RGBWRAPPERX(yuv2, rgb_full, xbgr32_full, PIX_FMT_ABGR,  0)

  1228. YUV2RGBWRAPPERX(yuv2, rgb_full, rgbx32_full, PIX_FMT_RGBA,  0)

  1229. YUV2RGBWRAPPERX(yuv2, rgb_full, xrgb32_full, PIX_FMT_ARGB,  0)

  1230. #endif

  1231. YUV2RGBWRAPPERX(yuv2, rgb_full, bgr24_full,  PIX_FMT_BGR24, 0)

  1232. YUV2RGBWRAPPERX(yuv2, rgb_full, rgb24_full,  PIX_FMT_RGB24, 0)

  1233. 

  1234. void ff_sws_init_output_funcs(SwsContext *c,

  1235.                               yuv2planar1_fn *yuv2plane1,

  1236.                               yuv2planarX_fn *yuv2planeX,

  1237.                               yuv2interleavedX_fn *yuv2nv12cX,

  1238.                               yuv2packed1_fn *yuv2packed1,

  1239.                               yuv2packed2_fn *yuv2packed2,

  1240.                               yuv2packedX_fn *yuv2packedX)

  1241. {

  1242.     enum PixelFormat dstFormat = c->dstFormat;

  1243. 

  1244.     if (is16BPS(dstFormat)) {

  1245.         *yuv2planeX = isBE(dstFormat) ? yuv2planeX_16BE_c  : yuv2planeX_16LE_c;

  1246.         *yuv2plane1 = isBE(dstFormat) ? yuv2plane1_16BE_c  : yuv2plane1_16LE_c;

  1247.     } else if (is9_OR_10BPS(dstFormat)) {

  1248.         if (av_pix_fmt_descriptors[dstFormat].comp[0].depth_minus1 == 8) {

  1249.             *yuv2planeX = isBE(dstFormat) ? yuv2planeX_9BE_c  : yuv2planeX_9LE_c;

  1250.             *yuv2plane1 = isBE(dstFormat) ? yuv2plane1_9BE_c  : yuv2plane1_9LE_c;

  1251.         } else {

  1252.             *yuv2planeX = isBE(dstFormat) ? yuv2planeX_10BE_c  : yuv2planeX_10LE_c;

  1253.             *yuv2plane1 = isBE(dstFormat) ? yuv2plane1_10BE_c  : yuv2plane1_10LE_c;

  1254.         }

  1255.     } else {

  1256.         *yuv2plane1 = yuv2plane1_8_c;

  1257.         *yuv2planeX = yuv2planeX_8_c;

  1258.         if (dstFormat == PIX_FMT_NV12 || dstFormat == PIX_FMT_NV21)

  1259.             *yuv2nv12cX = yuv2nv12cX_c;

  1260.     }

  1261. 

  1262.     if(c->flags & SWS_FULL_CHR_H_INT) {

  1263.         switch (dstFormat) {

  1264.             case PIX_FMT_RGBA:

  1265. #if CONFIG_SMALL

  1266.                 *yuv2packedX = yuv2rgba32_full_X_c;

  1267. #else

  1268. #if CONFIG_SWSCALE_ALPHA

  1269.                 if (c->alpPixBuf) {

  1270.                     *yuv2packedX = yuv2rgba32_full_X_c;

  1271.                 } else

  1272. #endif /* CONFIG_SWSCALE_ALPHA */

  1273.                 {

  1274.                     *yuv2packedX = yuv2rgbx32_full_X_c;

  1275.                 }

  1276. #endif /* !CONFIG_SMALL */

  1277.                 break;

  1278.             case PIX_FMT_ARGB:

  1279. #if CONFIG_SMALL

  1280.                 *yuv2packedX = yuv2argb32_full_X_c;

  1281. #else

  1282. #if CONFIG_SWSCALE_ALPHA

  1283.                 if (c->alpPixBuf) {

  1284.                     *yuv2packedX = yuv2argb32_full_X_c;

  1285.                 } else

  1286. #endif /* CONFIG_SWSCALE_ALPHA */

  1287.                 {

  1288.                     *yuv2packedX = yuv2xrgb32_full_X_c;

  1289.                 }

  1290. #endif /* !CONFIG_SMALL */

  1291.                 break;

  1292.             case PIX_FMT_BGRA:

  1293. #if CONFIG_SMALL

  1294.                 *yuv2packedX = yuv2bgra32_full_X_c;

  1295. #else

  1296. #if CONFIG_SWSCALE_ALPHA

  1297.                 if (c->alpPixBuf) {

  1298.                     *yuv2packedX = yuv2bgra32_full_X_c;

  1299.                 } else

  1300. #endif /* CONFIG_SWSCALE_ALPHA */

  1301.                 {

  1302.                     *yuv2packedX = yuv2bgrx32_full_X_c;

  1303.                 }

  1304. #endif /* !CONFIG_SMALL */

  1305.                 break;

  1306.             case PIX_FMT_ABGR:

  1307. #if CONFIG_SMALL

  1308.                 *yuv2packedX = yuv2abgr32_full_X_c;

  1309. #else

  1310. #if CONFIG_SWSCALE_ALPHA

  1311.                 if (c->alpPixBuf) {

  1312.                     *yuv2packedX = yuv2abgr32_full_X_c;

  1313.                 } else

  1314. #endif /* CONFIG_SWSCALE_ALPHA */

  1315.                 {

  1316.                     *yuv2packedX = yuv2xbgr32_full_X_c;

  1317.                 }

  1318. #endif /* !CONFIG_SMALL */

  1319.                 break;

  1320.             case PIX_FMT_RGB24:

  1321.             *yuv2packedX = yuv2rgb24_full_X_c;

  1322.             break;

  1323.         case PIX_FMT_BGR24:

  1324.             *yuv2packedX = yuv2bgr24_full_X_c;

  1325.             break;

  1326.         }

  1327.         if(!*yuv2packedX)

  1328.             goto YUV_PACKED;

  1329.     } else {

  1330.         YUV_PACKED:

  1331.         switch (dstFormat) {

  1332.         case PIX_FMT_RGB48LE:

  1333.             *yuv2packed1 = yuv2rgb48le_1_c;

  1334.             *yuv2packed2 = yuv2rgb48le_2_c;

  1335.             *yuv2packedX = yuv2rgb48le_X_c;

  1336.             break;

  1337.         case PIX_FMT_RGB48BE:

  1338.             *yuv2packed1 = yuv2rgb48be_1_c;

  1339.             *yuv2packed2 = yuv2rgb48be_2_c;

  1340.             *yuv2packedX = yuv2rgb48be_X_c;

  1341.             break;

  1342.         case PIX_FMT_BGR48LE:

  1343.             *yuv2packed1 = yuv2bgr48le_1_c;

  1344.             *yuv2packed2 = yuv2bgr48le_2_c;

  1345.             *yuv2packedX = yuv2bgr48le_X_c;

  1346.             break;

  1347.         case PIX_FMT_BGR48BE:

  1348.             *yuv2packed1 = yuv2bgr48be_1_c;

  1349.             *yuv2packed2 = yuv2bgr48be_2_c;

  1350.             *yuv2packedX = yuv2bgr48be_X_c;

  1351.             break;

  1352.         case PIX_FMT_RGB32:

  1353.         case PIX_FMT_BGR32:

  1354. #if CONFIG_SMALL

  1355.             *yuv2packed1 = yuv2rgb32_1_c;

  1356.             *yuv2packed2 = yuv2rgb32_2_c;

  1357.             *yuv2packedX = yuv2rgb32_X_c;

  1358. #else

  1359. #if CONFIG_SWSCALE_ALPHA

  1360.                 if (c->alpPixBuf) {

  1361.                     *yuv2packed1 = yuv2rgba32_1_c;

  1362.                     *yuv2packed2 = yuv2rgba32_2_c;

  1363.                     *yuv2packedX = yuv2rgba32_X_c;

  1364.                 } else

  1365. #endif /* CONFIG_SWSCALE_ALPHA */

  1366.                 {

  1367.                     *yuv2packed1 = yuv2rgbx32_1_c;

  1368.                     *yuv2packed2 = yuv2rgbx32_2_c;

  1369.                     *yuv2packedX = yuv2rgbx32_X_c;

  1370.                 }

  1371. #endif /* !CONFIG_SMALL */

  1372.             break;

  1373.         case PIX_FMT_RGB32_1:

  1374.         case PIX_FMT_BGR32_1:

  1375. #if CONFIG_SMALL

  1376.                 *yuv2packed1 = yuv2rgb32_1_1_c;

  1377.                 *yuv2packed2 = yuv2rgb32_1_2_c;

  1378.                 *yuv2packedX = yuv2rgb32_1_X_c;

  1379. #else

  1380. #if CONFIG_SWSCALE_ALPHA

  1381.                 if (c->alpPixBuf) {

  1382.                     *yuv2packed1 = yuv2rgba32_1_1_c;

  1383.                     *yuv2packed2 = yuv2rgba32_1_2_c;

  1384.                     *yuv2packedX = yuv2rgba32_1_X_c;

  1385.                 } else

  1386. #endif /* CONFIG_SWSCALE_ALPHA */

  1387.                 {

  1388.                     *yuv2packed1 = yuv2rgbx32_1_1_c;

  1389.                     *yuv2packed2 = yuv2rgbx32_1_2_c;

  1390.                     *yuv2packedX = yuv2rgbx32_1_X_c;

  1391.                 }

  1392. #endif /* !CONFIG_SMALL */

  1393.                 break;

  1394.         case PIX_FMT_RGB24:

  1395.             *yuv2packed1 = yuv2rgb24_1_c;

  1396.             *yuv2packed2 = yuv2rgb24_2_c;

  1397.             *yuv2packedX = yuv2rgb24_X_c;

  1398.             break;

  1399.         case PIX_FMT_BGR24:

  1400.             *yuv2packed1 = yuv2bgr24_1_c;

  1401.             *yuv2packed2 = yuv2bgr24_2_c;

  1402.             *yuv2packedX = yuv2bgr24_X_c;

  1403.             break;

  1404.         case PIX_FMT_RGB565LE:

  1405.         case PIX_FMT_RGB565BE:

  1406.         case PIX_FMT_BGR565LE:

  1407.         case PIX_FMT_BGR565BE:

  1408.             *yuv2packed1 = yuv2rgb16_1_c;

  1409.             *yuv2packed2 = yuv2rgb16_2_c;

  1410.             *yuv2packedX = yuv2rgb16_X_c;

  1411.             break;

  1412.         case PIX_FMT_RGB555LE:

  1413.         case PIX_FMT_RGB555BE:

  1414.         case PIX_FMT_BGR555LE:

  1415.         case PIX_FMT_BGR555BE:

  1416.             *yuv2packed1 = yuv2rgb15_1_c;

  1417.             *yuv2packed2 = yuv2rgb15_2_c;

  1418.             *yuv2packedX = yuv2rgb15_X_c;

  1419.             break;

  1420.         case PIX_FMT_RGB444LE:

  1421.         case PIX_FMT_RGB444BE:

  1422.         case PIX_FMT_BGR444LE:

  1423.         case PIX_FMT_BGR444BE:

  1424.             *yuv2packed1 = yuv2rgb12_1_c;

  1425.             *yuv2packed2 = yuv2rgb12_2_c;

  1426.             *yuv2packedX = yuv2rgb12_X_c;

  1427.             break;

  1428.         case PIX_FMT_RGB8:

  1429.         case PIX_FMT_BGR8:

  1430.             *yuv2packed1 = yuv2rgb8_1_c;

  1431.             *yuv2packed2 = yuv2rgb8_2_c;

  1432.             *yuv2packedX = yuv2rgb8_X_c;

  1433.             break;

  1434.         case PIX_FMT_RGB4:

  1435.         case PIX_FMT_BGR4:

  1436.             *yuv2packed1 = yuv2rgb4_1_c;

  1437.             *yuv2packed2 = yuv2rgb4_2_c;

  1438.             *yuv2packedX = yuv2rgb4_X_c;

  1439.             break;

  1440.         case PIX_FMT_RGB4_BYTE:

  1441.         case PIX_FMT_BGR4_BYTE:

  1442.             *yuv2packed1 = yuv2rgb4b_1_c;

  1443.             *yuv2packed2 = yuv2rgb4b_2_c;

  1444.             *yuv2packedX = yuv2rgb4b_X_c;

  1445.             break;

  1446.         }

  1447.     }

  1448.     switch (dstFormat) {

  1449.     case PIX_FMT_MONOWHITE:

  1450.         *yuv2packed1 = yuv2monowhite_1_c;

  1451.         *yuv2packed2 = yuv2monowhite_2_c;

  1452.         *yuv2packedX = yuv2monowhite_X_c;

  1453.         break;

  1454.     case PIX_FMT_MONOBLACK:

  1455.         *yuv2packed1 = yuv2monoblack_1_c;

  1456.         *yuv2packed2 = yuv2monoblack_2_c;

  1457.         *yuv2packedX = yuv2monoblack_X_c;

  1458.         break;

  1459.     case PIX_FMT_YUYV422:

  1460.         *yuv2packed1 = yuv2yuyv422_1_c;

  1461.         *yuv2packed2 = yuv2yuyv422_2_c;

  1462.         *yuv2packedX = yuv2yuyv422_X_c;

  1463.         break;

  1464.     case PIX_FMT_UYVY422:

  1465.         *yuv2packed1 = yuv2uyvy422_1_c;

  1466.         *yuv2packed2 = yuv2uyvy422_2_c;

  1467.         *yuv2packedX = yuv2uyvy422_X_c;

  1468.         break;

  1469.     }

  1470. }