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.
37634 Commits
32 Branches
325MB
Tree: 18d0a16fc9
Branches Tags
${ item.name }
Create branch ${ searchTerm }
from '18d0a16fc9'
${ noResults }
FFmpeg/libswscale/output.c

1524 lines
56KB
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 output_pixel(pos, val) \

  302.         if (target == PIX_FMT_GRAY16BE) { \

  303.             AV_WB16(pos, val); \

  304.         } else { \

  305.             AV_WL16(pos, val); \

  306.         }

  307. 

  308. static av_always_inline void

  309. yuv2gray16_X_c_template(SwsContext *c, const int16_t *lumFilter,

  310.                         const int32_t **lumSrc, int lumFilterSize,

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

  312.                         const int32_t **chrVSrc, int chrFilterSize,

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

  314.                         int y, enum PixelFormat target)

  315. {

  316.     int i;

  317. 

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

  319.         int j;

  320.         int Y1 = (1 << 14) - 0x40000000;

  321.         int Y2 = (1 << 14) - 0x40000000;

  322. 

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

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

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

  326.         }

  327.         Y1 >>= 15;

  328.         Y2 >>= 15;

  329.         Y1 = av_clip_int16(Y1);

  330.         Y2 = av_clip_int16(Y2);

  331.         output_pixel(&dest[i * 2 + 0], 0x8000 + Y1);

  332.         output_pixel(&dest[i * 2 + 1], 0x8000 + Y2);

  333.     }

  334. }

  335. 

  336. static av_always_inline void

  337. yuv2gray16_2_c_template(SwsContext *c, const int32_t *buf[2],

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

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

  340.                         int yalpha, int uvalpha, int y,

  341.                         enum PixelFormat target)

  342. {

  343.     int  yalpha1 = 4095 - yalpha;

  344.     int i;

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

  346. 

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

  348.         int Y1 = (buf0[i * 2    ] * yalpha1 + buf1[i * 2    ] * yalpha) >> 15;

  349.         int Y2 = (buf0[i * 2 + 1] * yalpha1 + buf1[i * 2 + 1] * yalpha) >> 15;

  350. 

  351.         output_pixel(&dest[i * 2 + 0], Y1);

  352.         output_pixel(&dest[i * 2 + 1], Y2);

  353.     }

  354. }

  355. 

  356. static av_always_inline void

  357. yuv2gray16_1_c_template(SwsContext *c, const int32_t *buf0,

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

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

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

  361. {

  362.     int i;

  363. 

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

  365.         int Y1 = (buf0[i * 2    ]+4)>>3;

  366.         int Y2 = (buf0[i * 2 + 1]+4)>>3;

  367. 

  368.         output_pixel(&dest[i * 2 + 0], Y1);

  369.         output_pixel(&dest[i * 2 + 1], Y2);

  370.     }

  371. }

  372. 

  373. #undef output_pixel

  374. 

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

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

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

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

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

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

  381.                         int y) \

  382. { \

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

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

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

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

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

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

  389.                           chrFilter, chrUSrc, chrVSrc, chrFilterSize, \

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

  391. } \

  392.  \

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

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

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

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

  397. { \

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

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

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

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

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

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

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

  405. } \

  406.  \

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

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

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

  410.                         int uvalpha, int y) \

  411. { \

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

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

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

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

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

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

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

  419. }

  420. 

  421. YUV2PACKED16WRAPPER(yuv2gray16,, LE, PIX_FMT_GRAY16LE)

  422. YUV2PACKED16WRAPPER(yuv2gray16,, BE, PIX_FMT_GRAY16BE)

  423. 

  424. #define output_pixel(pos, acc) \

  425.     if (target == PIX_FMT_MONOBLACK) { \

  426.         pos = acc; \

  427.     } else { \

  428.         pos = ~acc; \

  429.     }

  430. 

  431. static av_always_inline void

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

  433.                       const int16_t **lumSrc, int lumFilterSize,

  434.                       const int16_t *chrFilter, const int16_t **chrUSrc,

  435.                       const int16_t **chrVSrc, int chrFilterSize,

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

  437.                       int y, enum PixelFormat target)

  438. {

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

  440.     uint8_t *g = c->table_gU[128 + YUVRGB_TABLE_HEADROOM] + c->table_gV[128 + YUVRGB_TABLE_HEADROOM];

  441.     int i;

  442.     unsigned acc = 0;

  443. 

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

  445.         int j;

  446.         int Y1 = 1 << 18;

  447.         int Y2 = 1 << 18;

  448. 

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

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

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

  452.         }

  453.         Y1 >>= 19;

  454.         Y2 >>= 19;

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

  456.             Y1 = av_clip_uint8(Y1);

  457.             Y2 = av_clip_uint8(Y2);

  458.         }

  459.         acc += acc + g[Y1 + d128[(i + 0) & 7]];

  460.         acc += acc + g[Y2 + d128[(i + 1) & 7]];

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

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

  463.         }

  464.     }

  465. }

  466. 

  467. static av_always_inline void

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

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

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

  471.                       int yalpha, int uvalpha, int y,

  472.                       enum PixelFormat target)

  473. {

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

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

  476.     uint8_t *g = c->table_gU[128 + YUVRGB_TABLE_HEADROOM] + c->table_gV[128 + YUVRGB_TABLE_HEADROOM];

  477.     int  yalpha1 = 4095 - yalpha;

  478.     int i;

  479. 

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

  481.         int acc =    g[((buf0[i    ] * yalpha1 + buf1[i    ] * yalpha) >> 19) + d128[0]];

  482.         acc += acc + g[((buf0[i + 1] * yalpha1 + buf1[i + 1] * yalpha) >> 19) + d128[1]];

  483.         acc += acc + g[((buf0[i + 2] * yalpha1 + buf1[i + 2] * yalpha) >> 19) + d128[2]];

  484.         acc += acc + g[((buf0[i + 3] * yalpha1 + buf1[i + 3] * yalpha) >> 19) + d128[3]];

  485.         acc += acc + g[((buf0[i + 4] * yalpha1 + buf1[i + 4] * yalpha) >> 19) + d128[4]];

  486.         acc += acc + g[((buf0[i + 5] * yalpha1 + buf1[i + 5] * yalpha) >> 19) + d128[5]];

  487.         acc += acc + g[((buf0[i + 6] * yalpha1 + buf1[i + 6] * yalpha) >> 19) + d128[6]];

  488.         acc += acc + g[((buf0[i + 7] * yalpha1 + buf1[i + 7] * yalpha) >> 19) + d128[7]];

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

  490.     }

  491. }

  492. 

  493. static av_always_inline void

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

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

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

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

  498. {

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

  500.     uint8_t *g = c->table_gU[128 + YUVRGB_TABLE_HEADROOM] + c->table_gV[128 + YUVRGB_TABLE_HEADROOM];

  501.     int i;

  502. 

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

  504.         int acc =    g[(buf0[i    ] >> 7) + d128[0]];

  505.         acc += acc + g[(buf0[i + 1] >> 7) + d128[1]];

  506.         acc += acc + g[(buf0[i + 2] >> 7) + d128[2]];

  507.         acc += acc + g[(buf0[i + 3] >> 7) + d128[3]];

  508.         acc += acc + g[(buf0[i + 4] >> 7) + d128[4]];

  509.         acc += acc + g[(buf0[i + 5] >> 7) + d128[5]];

  510.         acc += acc + g[(buf0[i + 6] >> 7) + d128[6]];

  511.         acc += acc + g[(buf0[i + 7] >> 7) + d128[7]];

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

  513.     }

  514. }

  515. 

  516. #undef output_pixel

  517. 

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

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

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

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

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

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

  524.                                 int y) \

  525. { \

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

  527.                                   chrFilter, chrUSrc, chrVSrc, chrFilterSize, \

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

  529. } \

  530.  \

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

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

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

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

  535. { \

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

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

  538. } \

  539.  \

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

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

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

  543.                                 int uvalpha, int y) \

  544. { \

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

  546.                                   abuf0, dest, dstW, uvalpha, \

  547.                                   y, fmt); \

  548. }

  549. 

  550. YUV2PACKEDWRAPPER(yuv2mono,, white, PIX_FMT_MONOWHITE)

  551. YUV2PACKEDWRAPPER(yuv2mono,, black, PIX_FMT_MONOBLACK)

  552. 

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

  554.     if (target == PIX_FMT_YUYV422) { \

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

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

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

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

  559.     } else { \

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

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

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

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

  564.     }

  565. 

  566. static av_always_inline void

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

  568.                      const int16_t **lumSrc, int lumFilterSize,

  569.                      const int16_t *chrFilter, const int16_t **chrUSrc,

  570.                      const int16_t **chrVSrc, int chrFilterSize,

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

  572.                      int y, enum PixelFormat target)

  573. {

  574.     int i;

  575. 

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

  577.         int j;

  578.         int Y1 = 1 << 18;

  579.         int Y2 = 1 << 18;

  580.         int U  = 1 << 18;

  581.         int V  = 1 << 18;

  582. 

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

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

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

  586.         }

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

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

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

  590.         }

  591.         Y1 >>= 19;

  592.         Y2 >>= 19;

  593.         U  >>= 19;

  594.         V  >>= 19;

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

  596.             Y1 = av_clip_uint8(Y1);

  597.             Y2 = av_clip_uint8(Y2);

  598.             U  = av_clip_uint8(U);

  599.             V  = av_clip_uint8(V);

  600.         }

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

  602.     }

  603. }

  604. 

  605. static av_always_inline void

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

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

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

  609.                      int yalpha, int uvalpha, int y,

  610.                      enum PixelFormat target)

  611. {

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

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

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

  615.     int  yalpha1 = 4095 - yalpha;

  616.     int uvalpha1 = 4095 - uvalpha;

  617.     int i;

  618. 

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

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

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

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

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

  624. 

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

  626.     }

  627. }

  628. 

  629. static av_always_inline void

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

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

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

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

  634. {

  635.     const int16_t *ubuf0 = ubuf[0], *ubuf1 = ubuf[1],

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

  637.     int i;

  638. 

  639.     if (uvalpha < 2048) {

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

  641.             int Y1 = buf0[i * 2]     >> 7;

  642.             int Y2 = buf0[i * 2 + 1] >> 7;

  643.             int U  = ubuf1[i]        >> 7;

  644.             int V  = vbuf1[i]        >> 7;

  645. 

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

  647.         }

  648.     } else {

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

  650.             int Y1 =  buf0[i * 2]          >> 7;

  651.             int Y2 =  buf0[i * 2 + 1]      >> 7;

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

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

  654. 

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

  656.         }

  657.     }

  658. }

  659. 

  660. #undef output_pixels

  661. 

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

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

  664. 

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

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

  667. #define output_pixel(pos, val) \

  668.     if (isBE(target)) { \

  669.         AV_WB16(pos, val); \

  670.     } else { \

  671.         AV_WL16(pos, val); \

  672.     }

  673. 

  674. static av_always_inline void

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

  676.                        const int32_t **lumSrc, int lumFilterSize,

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

  678.                        const int32_t **chrVSrc, int chrFilterSize,

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

  680.                        int y, enum PixelFormat target)

  681. {

  682.     int i;

  683. 

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

  685.         int j;

  686.         int Y1 = -0x40000000;

  687.         int Y2 = -0x40000000;

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

  689.         int V  = -128 << 23;

  690.         int R, G, B;

  691. 

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

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

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

  695.         }

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

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

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

  699.         }

  700. 

  701.         // 8bit: 12+15=27; 16-bit: 12+19=31

  702.         Y1 >>= 14; // 10

  703.         Y1 += 0x10000;

  704.         Y2 >>= 14;

  705.         Y2 += 0x10000;

  706.         U  >>= 14;

  707.         V  >>= 14;

  708. 

  709.         // 8bit: 27 -> 17bit, 16bit: 31 - 14 = 17bit

  710.         Y1 -= c->yuv2rgb_y_offset;

  711.         Y2 -= c->yuv2rgb_y_offset;

  712.         Y1 *= c->yuv2rgb_y_coeff;

  713.         Y2 *= c->yuv2rgb_y_coeff;

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

  715.         Y2 += 1 << 13;

  716.         // 8bit: 17 + 13bit = 30bit, 16bit: 17 + 13bit = 30bit

  717. 

  718.         R = V * c->yuv2rgb_v2r_coeff;

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

  720.         B =                            U * c->yuv2rgb_u2b_coeff;

  721. 

  722.         // 8bit: 30 - 22 = 8bit, 16bit: 30bit - 14 = 16bit

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

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

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

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

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

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

  729.         dest += 6;

  730.     }

  731. }

  732. 

  733. static av_always_inline void

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

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

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

  737.                        int yalpha, int uvalpha, int y,

  738.                        enum PixelFormat target)

  739. {

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

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

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

  743.     int  yalpha1 = 4095 - yalpha;

  744.     int uvalpha1 = 4095 - uvalpha;

  745.     int i;

  746. 

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

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

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

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

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

  752.         int R, G, B;

  753. 

  754.         Y1 -= c->yuv2rgb_y_offset;

  755.         Y2 -= c->yuv2rgb_y_offset;

  756.         Y1 *= c->yuv2rgb_y_coeff;

  757.         Y2 *= c->yuv2rgb_y_coeff;

  758.         Y1 += 1 << 13;

  759.         Y2 += 1 << 13;

  760. 

  761.         R = V * c->yuv2rgb_v2r_coeff;

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

  763.         B =                            U * c->yuv2rgb_u2b_coeff;

  764. 

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

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

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

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

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

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

  771.         dest += 6;

  772.     }

  773. }

  774. 

  775. static av_always_inline void

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

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

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

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

  780. {

  781.     const int32_t *ubuf0 = ubuf[0], *ubuf1 = ubuf[1],

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

  783.     int i;

  784. 

  785.     if (uvalpha < 2048) {

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

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

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

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

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

  791.             int R, G, B;

  792. 

  793.             Y1 -= c->yuv2rgb_y_offset;

  794.             Y2 -= c->yuv2rgb_y_offset;

  795.             Y1 *= c->yuv2rgb_y_coeff;

  796.             Y2 *= c->yuv2rgb_y_coeff;

  797.             Y1 += 1 << 13;

  798.             Y2 += 1 << 13;

  799. 

  800.             R = V * c->yuv2rgb_v2r_coeff;

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

  802.             B =                            U * c->yuv2rgb_u2b_coeff;

  803. 

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

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

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

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

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

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

  810.             dest += 6;

  811.         }

  812.     } else {

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

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

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

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

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

  818.             int R, G, B;

  819. 

  820.             Y1 -= c->yuv2rgb_y_offset;

  821.             Y2 -= c->yuv2rgb_y_offset;

  822.             Y1 *= c->yuv2rgb_y_coeff;

  823.             Y2 *= c->yuv2rgb_y_coeff;

  824.             Y1 += 1 << 13;

  825.             Y2 += 1 << 13;

  826. 

  827.             R = V * c->yuv2rgb_v2r_coeff;

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

  829.             B =                            U * c->yuv2rgb_u2b_coeff;

  830. 

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

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

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

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

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

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

  837.             dest += 6;

  838.         }

  839.     }

  840. }

  841. 

  842. #undef output_pixel

  843. #undef r_b

  844. #undef b_r

  845. 

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

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

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

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

  850. 

  851. /*

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

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

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

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

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

  857.  * correct RGB values into the destination buffer.

  858.  */

  859. static av_always_inline void

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

  861.               unsigned A1, unsigned A2,

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

  863.               enum PixelFormat target, int hasAlpha)

  864. {

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

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

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

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

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

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

  871. 

  872. #if CONFIG_SMALL

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

  874. 

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

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

  877. #else

  878.         if (hasAlpha) {

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

  880. 

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

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

  883.         } else {

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

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

  886.         }

  887. #endif

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

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

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

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

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

  893. 

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

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

  896. 

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

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

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

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

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

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

  903. #undef r_b

  904. #undef b_r

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

  906.                target == PIX_FMT_RGB555 || target == PIX_FMT_BGR555 ||

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

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

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

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

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

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

  913. 

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

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

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

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

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

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

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

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

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

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

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

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

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

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

  928.         } else {

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

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

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

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

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

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

  935.         }

  936. 

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

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

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

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

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

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

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

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

  945. 

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

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

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

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

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

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

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

  953.         } else {

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

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

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

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

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

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

  960.         }

  961. 

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

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

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

  965.         } else {

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

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

  968.         }

  969.     }

  970. }

  971. 

  972. static av_always_inline void

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

  974.                      const int16_t **lumSrc, int lumFilterSize,

  975.                      const int16_t *chrFilter, const int16_t **chrUSrc,

  976.                      const int16_t **chrVSrc, int chrFilterSize,

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

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

  979. {

  980.     int i;

  981. 

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

  983.         int j;

  984.         int Y1 = 1 << 18;

  985.         int Y2 = 1 << 18;

  986.         int U  = 1 << 18;

  987.         int V  = 1 << 18;

  988.         int av_unused A1, A2;

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

  990. 

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

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

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

  994.         }

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

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

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

  998.         }

  999.         Y1 >>= 19;

  1000.         Y2 >>= 19;

  1001.         U  >>= 19;

  1002.         V  >>= 19;

  1003.         if (hasAlpha) {

  1004.             A1 = 1 << 18;

  1005.             A2 = 1 << 18;

  1006.             for (j = 0; j < lumFilterSize; j++) {

  1007.                 A1 += alpSrc[j][i * 2    ] * lumFilter[j];

  1008.                 A2 += alpSrc[j][i * 2 + 1] * lumFilter[j];

  1009.             }

  1010.             A1 >>= 19;

  1011.             A2 >>= 19;

  1012.             if ((A1 | A2) & 0x100) {

  1013.                 A1 = av_clip_uint8(A1);

  1014.                 A2 = av_clip_uint8(A2);

  1015.             }

  1016.         }

  1017. 

  1018.         r =  c->table_rV[V + YUVRGB_TABLE_HEADROOM];

  1019.         g = (c->table_gU[U + YUVRGB_TABLE_HEADROOM] + c->table_gV[V + YUVRGB_TABLE_HEADROOM]);

  1020.         b =  c->table_bU[U + YUVRGB_TABLE_HEADROOM];

  1021. 

  1022.         yuv2rgb_write(dest, i, Y1, Y2, hasAlpha ? A1 : 0, hasAlpha ? A2 : 0,

  1023.                       r, g, b, y, target, hasAlpha);

  1024.     }

  1025. }

  1026. 

  1027. static av_always_inline void

  1028. yuv2rgb_2_c_template(SwsContext *c, const int16_t *buf[2],

  1029.                      const int16_t *ubuf[2], const int16_t *vbuf[2],

  1030.                      const int16_t *abuf[2], uint8_t *dest, int dstW,

  1031.                      int yalpha, int uvalpha, int y,

  1032.                      enum PixelFormat target, int hasAlpha)

  1033. {

  1034.     const int16_t *buf0  = buf[0],  *buf1  = buf[1],

  1035.                   *ubuf0 = ubuf[0], *ubuf1 = ubuf[1],

  1036.                   *vbuf0 = vbuf[0], *vbuf1 = vbuf[1],

  1037.                   *abuf0 = hasAlpha ? abuf[0] : NULL,

  1038.                   *abuf1 = hasAlpha ? abuf[1] : NULL;

  1039.     int  yalpha1 = 4095 - yalpha;

  1040.     int uvalpha1 = 4095 - uvalpha;

  1041.     int i;

  1042. 

  1043.     for (i = 0; i < (dstW >> 1); i++) {

  1044.         int Y1 = (buf0[i * 2]     * yalpha1  + buf1[i * 2]     * yalpha)  >> 19;

  1045.         int Y2 = (buf0[i * 2 + 1] * yalpha1  + buf1[i * 2 + 1] * yalpha)  >> 19;

  1046.         int U  = (ubuf0[i]        * uvalpha1 + ubuf1[i]        * uvalpha) >> 19;

  1047.         int V  = (vbuf0[i]        * uvalpha1 + vbuf1[i]        * uvalpha) >> 19;

  1048.         int A1, A2;

  1049.         const void *r =  c->table_rV[V + YUVRGB_TABLE_HEADROOM],

  1050.                    *g = (c->table_gU[U + YUVRGB_TABLE_HEADROOM] + c->table_gV[V + YUVRGB_TABLE_HEADROOM]),

  1051.                    *b =  c->table_bU[U + YUVRGB_TABLE_HEADROOM];

  1052. 

  1053.         if (hasAlpha) {

  1054.             A1 = (abuf0[i * 2    ] * yalpha1 + abuf1[i * 2    ] * yalpha) >> 19;

  1055.             A2 = (abuf0[i * 2 + 1] * yalpha1 + abuf1[i * 2 + 1] * yalpha) >> 19;

  1056.         }

  1057. 

  1058.         yuv2rgb_write(dest, i, Y1, Y2, hasAlpha ? A1 : 0, hasAlpha ? A2 : 0,

  1059.                       r, g, b, y, target, hasAlpha);

  1060.     }

  1061. }

  1062. 

  1063. static av_always_inline void

  1064. yuv2rgb_1_c_template(SwsContext *c, const int16_t *buf0,

  1065.                      const int16_t *ubuf[2], const int16_t *vbuf[2],

  1066.                      const int16_t *abuf0, uint8_t *dest, int dstW,

  1067.                      int uvalpha, int y, enum PixelFormat target,

  1068.                      int hasAlpha)

  1069. {

  1070.     const int16_t *ubuf0 = ubuf[0], *ubuf1 = ubuf[1],

  1071.                   *vbuf0 = vbuf[0], *vbuf1 = vbuf[1];

  1072.     int i;

  1073. 

  1074.     if (uvalpha < 2048) {

  1075.         for (i = 0; i < (dstW >> 1); i++) {

  1076.             int Y1 = buf0[i * 2]     >> 7;

  1077.             int Y2 = buf0[i * 2 + 1] >> 7;

  1078.             int U  = ubuf1[i]        >> 7;

  1079.             int V  = vbuf1[i]        >> 7;

  1080.             int A1, A2;

  1081.             const void *r =  c->table_rV[V + YUVRGB_TABLE_HEADROOM],

  1082.                        *g = (c->table_gU[U + YUVRGB_TABLE_HEADROOM] + c->table_gV[V + YUVRGB_TABLE_HEADROOM]),

  1083.                        *b =  c->table_bU[U + YUVRGB_TABLE_HEADROOM];

  1084. 

  1085.             if (hasAlpha) {

  1086.                 A1 = abuf0[i * 2    ] >> 7;

  1087.                 A2 = abuf0[i * 2 + 1] >> 7;

  1088.             }

  1089. 

  1090.             yuv2rgb_write(dest, i, Y1, Y2, hasAlpha ? A1 : 0, hasAlpha ? A2 : 0,

  1091.                           r, g, b, y, target, hasAlpha);

  1092.         }

  1093.     } else {

  1094.         for (i = 0; i < (dstW >> 1); i++) {

  1095.             int Y1 =  buf0[i * 2]          >> 7;

  1096.             int Y2 =  buf0[i * 2 + 1]      >> 7;

  1097.             int U  = (ubuf0[i] + ubuf1[i]) >> 8;

  1098.             int V  = (vbuf0[i] + vbuf1[i]) >> 8;

  1099.             int A1, A2;

  1100.             const void *r =  c->table_rV[V + YUVRGB_TABLE_HEADROOM],

  1101.                        *g = (c->table_gU[U + YUVRGB_TABLE_HEADROOM] + c->table_gV[V + YUVRGB_TABLE_HEADROOM]),

  1102.                        *b =  c->table_bU[U + YUVRGB_TABLE_HEADROOM];

  1103. 

  1104.             if (hasAlpha) {

  1105.                 A1 = abuf0[i * 2    ] >> 7;

  1106.                 A2 = abuf0[i * 2 + 1] >> 7;

  1107.             }

  1108. 

  1109.             yuv2rgb_write(dest, i, Y1, Y2, hasAlpha ? A1 : 0, hasAlpha ? A2 : 0,

  1110.                           r, g, b, y, target, hasAlpha);

  1111.         }

  1112.     }

  1113. }

  1114. 

  1115. #define YUV2RGBWRAPPERX(name, base, ext, fmt, hasAlpha) \

  1116. static void name ## ext ## _X_c(SwsContext *c, const int16_t *lumFilter, \

  1117.                                 const int16_t **lumSrc, int lumFilterSize, \

  1118.                                 const int16_t *chrFilter, const int16_t **chrUSrc, \

  1119.                                 const int16_t **chrVSrc, int chrFilterSize, \

  1120.                                 const int16_t **alpSrc, uint8_t *dest, int dstW, \

  1121.                                 int y) \

  1122. { \

  1123.     name ## base ## _X_c_template(c, lumFilter, lumSrc, lumFilterSize, \

  1124.                                   chrFilter, chrUSrc, chrVSrc, chrFilterSize, \

  1125.                                   alpSrc, dest, dstW, y, fmt, hasAlpha); \

  1126. }

  1127. #define YUV2RGBWRAPPER(name, base, ext, fmt, hasAlpha) \

  1128. YUV2RGBWRAPPERX(name, base, ext, fmt, hasAlpha) \

  1129. static void name ## ext ## _2_c(SwsContext *c, const int16_t *buf[2], \

  1130.                                 const int16_t *ubuf[2], const int16_t *vbuf[2], \

  1131.                                 const int16_t *abuf[2], uint8_t *dest, int dstW, \

  1132.                                 int yalpha, int uvalpha, int y) \

  1133. { \

  1134.     name ## base ## _2_c_template(c, buf, ubuf, vbuf, abuf, \

  1135.                                   dest, dstW, yalpha, uvalpha, y, fmt, hasAlpha); \

  1136. } \

  1137.  \

  1138. static void name ## ext ## _1_c(SwsContext *c, const int16_t *buf0, \

  1139.                                 const int16_t *ubuf[2], const int16_t *vbuf[2], \

  1140.                                 const int16_t *abuf0, uint8_t *dest, int dstW, \

  1141.                                 int uvalpha, int y) \

  1142. { \

  1143.     name ## base ## _1_c_template(c, buf0, ubuf, vbuf, abuf0, dest, \

  1144.                                   dstW, uvalpha, y, fmt, hasAlpha); \

  1145. }

  1146. 

  1147. #if CONFIG_SMALL

  1148. YUV2RGBWRAPPER(yuv2rgb,,  32_1,  PIX_FMT_RGB32_1,   CONFIG_SWSCALE_ALPHA && c->alpPixBuf)

  1149. YUV2RGBWRAPPER(yuv2rgb,,  32,    PIX_FMT_RGB32,     CONFIG_SWSCALE_ALPHA && c->alpPixBuf)

  1150. #else

  1151. #if CONFIG_SWSCALE_ALPHA

  1152. YUV2RGBWRAPPER(yuv2rgb,, a32_1,  PIX_FMT_RGB32_1,   1)

  1153. YUV2RGBWRAPPER(yuv2rgb,, a32,    PIX_FMT_RGB32,     1)

  1154. #endif

  1155. YUV2RGBWRAPPER(yuv2rgb,, x32_1,  PIX_FMT_RGB32_1,   0)

  1156. YUV2RGBWRAPPER(yuv2rgb,, x32,    PIX_FMT_RGB32,     0)

  1157. #endif

  1158. YUV2RGBWRAPPER(yuv2, rgb, rgb24, PIX_FMT_RGB24,   0)

  1159. YUV2RGBWRAPPER(yuv2, rgb, bgr24, PIX_FMT_BGR24,   0)

  1160. YUV2RGBWRAPPER(yuv2rgb,,  16,    PIX_FMT_RGB565,    0)

  1161. YUV2RGBWRAPPER(yuv2rgb,,  15,    PIX_FMT_RGB555,    0)

  1162. YUV2RGBWRAPPER(yuv2rgb,,  12,    PIX_FMT_RGB444,    0)

  1163. YUV2RGBWRAPPER(yuv2rgb,,   8,    PIX_FMT_RGB8,      0)

  1164. YUV2RGBWRAPPER(yuv2rgb,,   4,    PIX_FMT_RGB4,      0)

  1165. YUV2RGBWRAPPER(yuv2rgb,,   4b,   PIX_FMT_RGB4_BYTE, 0)

  1166. 

  1167. static av_always_inline void

  1168. yuv2rgb_full_X_c_template(SwsContext *c, const int16_t *lumFilter,

  1169.                           const int16_t **lumSrc, int lumFilterSize,

  1170.                           const int16_t *chrFilter, const int16_t **chrUSrc,

  1171.                           const int16_t **chrVSrc, int chrFilterSize,

  1172.                           const int16_t **alpSrc, uint8_t *dest,

  1173.                           int dstW, int y, enum PixelFormat target, int hasAlpha)

  1174. {

  1175.     int i;

  1176.     int step = (target == PIX_FMT_RGB24 || target == PIX_FMT_BGR24) ? 3 : 4;

  1177. 

  1178.     for (i = 0; i < dstW; i++) {

  1179.         int j;

  1180.         int Y = 1<<9;

  1181.         int U = (1<<9)-(128 << 19);

  1182.         int V = (1<<9)-(128 << 19);

  1183.         int av_unused A;

  1184.         int R, G, B;

  1185. 

  1186.         for (j = 0; j < lumFilterSize; j++) {

  1187.             Y += lumSrc[j][i] * lumFilter[j];

  1188.         }

  1189.         for (j = 0; j < chrFilterSize; j++) {

  1190.             U += chrUSrc[j][i] * chrFilter[j];

  1191.             V += chrVSrc[j][i] * chrFilter[j];

  1192.         }

  1193.         Y >>= 10;

  1194.         U >>= 10;

  1195.         V >>= 10;

  1196.         if (hasAlpha) {

  1197.             A = 1 << 18;

  1198.             for (j = 0; j < lumFilterSize; j++) {

  1199.                 A += alpSrc[j][i] * lumFilter[j];

  1200.             }

  1201.             A >>= 19;

  1202.             if (A & 0x100)

  1203.                 A = av_clip_uint8(A);

  1204.         }

  1205.         Y -= c->yuv2rgb_y_offset;

  1206.         Y *= c->yuv2rgb_y_coeff;

  1207.         Y += 1 << 21;

  1208.         R = Y + V*c->yuv2rgb_v2r_coeff;

  1209.         G = Y + V*c->yuv2rgb_v2g_coeff + U*c->yuv2rgb_u2g_coeff;

  1210.         B = Y +                          U*c->yuv2rgb_u2b_coeff;

  1211.         if ((R | G | B) & 0xC0000000) {

  1212.             R = av_clip_uintp2(R, 30);

  1213.             G = av_clip_uintp2(G, 30);

  1214.             B = av_clip_uintp2(B, 30);

  1215.         }

  1216. 

  1217.         switch(target) {

  1218.         case PIX_FMT_ARGB:

  1219.             dest[0] = hasAlpha ? A : 255;

  1220.             dest[1] = R >> 22;

  1221.             dest[2] = G >> 22;

  1222.             dest[3] = B >> 22;

  1223.             break;

  1224.         case PIX_FMT_RGB24:

  1225.             dest[0] = R >> 22;

  1226.             dest[1] = G >> 22;

  1227.             dest[2] = B >> 22;

  1228.             break;

  1229.         case PIX_FMT_RGBA:

  1230.             dest[0] = R >> 22;

  1231.             dest[1] = G >> 22;

  1232.             dest[2] = B >> 22;

  1233.             dest[3] = hasAlpha ? A : 255;

  1234.             break;

  1235.         case PIX_FMT_ABGR:

  1236.             dest[0] = hasAlpha ? A : 255;

  1237.             dest[1] = B >> 22;

  1238.             dest[2] = G >> 22;

  1239.             dest[3] = R >> 22;

  1240.             break;

  1241.         case PIX_FMT_BGR24:

  1242.             dest[0] = B >> 22;

  1243.             dest[1] = G >> 22;

  1244.             dest[2] = R >> 22;

  1245.             break;

  1246.         case PIX_FMT_BGRA:

  1247.             dest[0] = B >> 22;

  1248.             dest[1] = G >> 22;

  1249.             dest[2] = R >> 22;

  1250.             dest[3] = hasAlpha ? A : 255;

  1251.             break;

  1252.         }

  1253.         dest += step;

  1254.     }

  1255. }

  1256. 

  1257. #if CONFIG_SMALL

  1258. YUV2RGBWRAPPERX(yuv2, rgb_full, bgra32_full, PIX_FMT_BGRA,  CONFIG_SWSCALE_ALPHA && c->alpPixBuf)

  1259. YUV2RGBWRAPPERX(yuv2, rgb_full, abgr32_full, PIX_FMT_ABGR,  CONFIG_SWSCALE_ALPHA && c->alpPixBuf)

  1260. YUV2RGBWRAPPERX(yuv2, rgb_full, rgba32_full, PIX_FMT_RGBA,  CONFIG_SWSCALE_ALPHA && c->alpPixBuf)

  1261. YUV2RGBWRAPPERX(yuv2, rgb_full, argb32_full, PIX_FMT_ARGB,  CONFIG_SWSCALE_ALPHA && c->alpPixBuf)

  1262. #else

  1263. #if CONFIG_SWSCALE_ALPHA

  1264. YUV2RGBWRAPPERX(yuv2, rgb_full, bgra32_full, PIX_FMT_BGRA,  1)

  1265. YUV2RGBWRAPPERX(yuv2, rgb_full, abgr32_full, PIX_FMT_ABGR,  1)

  1266. YUV2RGBWRAPPERX(yuv2, rgb_full, rgba32_full, PIX_FMT_RGBA,  1)

  1267. YUV2RGBWRAPPERX(yuv2, rgb_full, argb32_full, PIX_FMT_ARGB,  1)

  1268. #endif

  1269. YUV2RGBWRAPPERX(yuv2, rgb_full, bgrx32_full, PIX_FMT_BGRA,  0)

  1270. YUV2RGBWRAPPERX(yuv2, rgb_full, xbgr32_full, PIX_FMT_ABGR,  0)

  1271. YUV2RGBWRAPPERX(yuv2, rgb_full, rgbx32_full, PIX_FMT_RGBA,  0)

  1272. YUV2RGBWRAPPERX(yuv2, rgb_full, xrgb32_full, PIX_FMT_ARGB,  0)

  1273. #endif

  1274. YUV2RGBWRAPPERX(yuv2, rgb_full, bgr24_full,  PIX_FMT_BGR24, 0)

  1275. YUV2RGBWRAPPERX(yuv2, rgb_full, rgb24_full,  PIX_FMT_RGB24, 0)

  1276. 

  1277. void ff_sws_init_output_funcs(SwsContext *c,

  1278.                               yuv2planar1_fn *yuv2plane1,

  1279.                               yuv2planarX_fn *yuv2planeX,

  1280.                               yuv2interleavedX_fn *yuv2nv12cX,

  1281.                               yuv2packed1_fn *yuv2packed1,

  1282.                               yuv2packed2_fn *yuv2packed2,

  1283.                               yuv2packedX_fn *yuv2packedX)

  1284. {

  1285.     enum PixelFormat dstFormat = c->dstFormat;

  1286. 

  1287.     if (is16BPS(dstFormat)) {

  1288.         *yuv2planeX = isBE(dstFormat) ? yuv2planeX_16BE_c  : yuv2planeX_16LE_c;

  1289.         *yuv2plane1 = isBE(dstFormat) ? yuv2plane1_16BE_c  : yuv2plane1_16LE_c;

  1290.     } else if (is9_OR_10BPS(dstFormat)) {

  1291.         if (av_pix_fmt_descriptors[dstFormat].comp[0].depth_minus1 == 8) {

  1292.             *yuv2planeX = isBE(dstFormat) ? yuv2planeX_9BE_c  : yuv2planeX_9LE_c;

  1293.             *yuv2plane1 = isBE(dstFormat) ? yuv2plane1_9BE_c  : yuv2plane1_9LE_c;

  1294.         } else {

  1295.             *yuv2planeX = isBE(dstFormat) ? yuv2planeX_10BE_c  : yuv2planeX_10LE_c;

  1296.             *yuv2plane1 = isBE(dstFormat) ? yuv2plane1_10BE_c  : yuv2plane1_10LE_c;

  1297.         }

  1298.     } else {

  1299.         *yuv2plane1 = yuv2plane1_8_c;

  1300.         *yuv2planeX = yuv2planeX_8_c;

  1301.         if (dstFormat == PIX_FMT_NV12 || dstFormat == PIX_FMT_NV21)

  1302.             *yuv2nv12cX = yuv2nv12cX_c;

  1303.     }

  1304. 

  1305.     if(c->flags & SWS_FULL_CHR_H_INT) {

  1306.         switch (dstFormat) {

  1307.             case PIX_FMT_RGBA:

  1308. #if CONFIG_SMALL

  1309.                 *yuv2packedX = yuv2rgba32_full_X_c;

  1310. #else

  1311. #if CONFIG_SWSCALE_ALPHA

  1312.                 if (c->alpPixBuf) {

  1313.                     *yuv2packedX = yuv2rgba32_full_X_c;

  1314.                 } else

  1315. #endif /* CONFIG_SWSCALE_ALPHA */

  1316.                 {

  1317.                     *yuv2packedX = yuv2rgbx32_full_X_c;

  1318.                 }

  1319. #endif /* !CONFIG_SMALL */

  1320.                 break;

  1321.             case PIX_FMT_ARGB:

  1322. #if CONFIG_SMALL

  1323.                 *yuv2packedX = yuv2argb32_full_X_c;

  1324. #else

  1325. #if CONFIG_SWSCALE_ALPHA

  1326.                 if (c->alpPixBuf) {

  1327.                     *yuv2packedX = yuv2argb32_full_X_c;

  1328.                 } else

  1329. #endif /* CONFIG_SWSCALE_ALPHA */

  1330.                 {

  1331.                     *yuv2packedX = yuv2xrgb32_full_X_c;

  1332.                 }

  1333. #endif /* !CONFIG_SMALL */

  1334.                 break;

  1335.             case PIX_FMT_BGRA:

  1336. #if CONFIG_SMALL

  1337.                 *yuv2packedX = yuv2bgra32_full_X_c;

  1338. #else

  1339. #if CONFIG_SWSCALE_ALPHA

  1340.                 if (c->alpPixBuf) {

  1341.                     *yuv2packedX = yuv2bgra32_full_X_c;

  1342.                 } else

  1343. #endif /* CONFIG_SWSCALE_ALPHA */

  1344.                 {

  1345.                     *yuv2packedX = yuv2bgrx32_full_X_c;

  1346.                 }

  1347. #endif /* !CONFIG_SMALL */

  1348.                 break;

  1349.             case PIX_FMT_ABGR:

  1350. #if CONFIG_SMALL

  1351.                 *yuv2packedX = yuv2abgr32_full_X_c;

  1352. #else

  1353. #if CONFIG_SWSCALE_ALPHA

  1354.                 if (c->alpPixBuf) {

  1355.                     *yuv2packedX = yuv2abgr32_full_X_c;

  1356.                 } else

  1357. #endif /* CONFIG_SWSCALE_ALPHA */

  1358.                 {

  1359.                     *yuv2packedX = yuv2xbgr32_full_X_c;

  1360.                 }

  1361. #endif /* !CONFIG_SMALL */

  1362.                 break;

  1363.             case PIX_FMT_RGB24:

  1364.             *yuv2packedX = yuv2rgb24_full_X_c;

  1365.             break;

  1366.         case PIX_FMT_BGR24:

  1367.             *yuv2packedX = yuv2bgr24_full_X_c;

  1368.             break;

  1369.         }

  1370.         if(!*yuv2packedX)

  1371.             goto YUV_PACKED;

  1372.     } else {

  1373.         YUV_PACKED:

  1374.         switch (dstFormat) {

  1375.         case PIX_FMT_RGB48LE:

  1376.             *yuv2packed1 = yuv2rgb48le_1_c;

  1377.             *yuv2packed2 = yuv2rgb48le_2_c;

  1378.             *yuv2packedX = yuv2rgb48le_X_c;

  1379.             break;

  1380.         case PIX_FMT_RGB48BE:

  1381.             *yuv2packed1 = yuv2rgb48be_1_c;

  1382.             *yuv2packed2 = yuv2rgb48be_2_c;

  1383.             *yuv2packedX = yuv2rgb48be_X_c;

  1384.             break;

  1385.         case PIX_FMT_BGR48LE:

  1386.             *yuv2packed1 = yuv2bgr48le_1_c;

  1387.             *yuv2packed2 = yuv2bgr48le_2_c;

  1388.             *yuv2packedX = yuv2bgr48le_X_c;

  1389.             break;

  1390.         case PIX_FMT_BGR48BE:

  1391.             *yuv2packed1 = yuv2bgr48be_1_c;

  1392.             *yuv2packed2 = yuv2bgr48be_2_c;

  1393.             *yuv2packedX = yuv2bgr48be_X_c;

  1394.             break;

  1395.         case PIX_FMT_RGB32:

  1396.         case PIX_FMT_BGR32:

  1397. #if CONFIG_SMALL

  1398.             *yuv2packed1 = yuv2rgb32_1_c;

  1399.             *yuv2packed2 = yuv2rgb32_2_c;

  1400.             *yuv2packedX = yuv2rgb32_X_c;

  1401. #else

  1402. #if CONFIG_SWSCALE_ALPHA

  1403.                 if (c->alpPixBuf) {

  1404.                     *yuv2packed1 = yuv2rgba32_1_c;

  1405.                     *yuv2packed2 = yuv2rgba32_2_c;

  1406.                     *yuv2packedX = yuv2rgba32_X_c;

  1407.                 } else

  1408. #endif /* CONFIG_SWSCALE_ALPHA */

  1409.                 {

  1410.                     *yuv2packed1 = yuv2rgbx32_1_c;

  1411.                     *yuv2packed2 = yuv2rgbx32_2_c;

  1412.                     *yuv2packedX = yuv2rgbx32_X_c;

  1413.                 }

  1414. #endif /* !CONFIG_SMALL */

  1415.             break;

  1416.         case PIX_FMT_RGB32_1:

  1417.         case PIX_FMT_BGR32_1:

  1418. #if CONFIG_SMALL

  1419.                 *yuv2packed1 = yuv2rgb32_1_1_c;

  1420.                 *yuv2packed2 = yuv2rgb32_1_2_c;

  1421.                 *yuv2packedX = yuv2rgb32_1_X_c;

  1422. #else

  1423. #if CONFIG_SWSCALE_ALPHA

  1424.                 if (c->alpPixBuf) {

  1425.                     *yuv2packed1 = yuv2rgba32_1_1_c;

  1426.                     *yuv2packed2 = yuv2rgba32_1_2_c;

  1427.                     *yuv2packedX = yuv2rgba32_1_X_c;

  1428.                 } else

  1429. #endif /* CONFIG_SWSCALE_ALPHA */

  1430.                 {

  1431.                     *yuv2packed1 = yuv2rgbx32_1_1_c;

  1432.                     *yuv2packed2 = yuv2rgbx32_1_2_c;

  1433.                     *yuv2packedX = yuv2rgbx32_1_X_c;

  1434.                 }

  1435. #endif /* !CONFIG_SMALL */

  1436.                 break;

  1437.         case PIX_FMT_RGB24:

  1438.             *yuv2packed1 = yuv2rgb24_1_c;

  1439.             *yuv2packed2 = yuv2rgb24_2_c;

  1440.             *yuv2packedX = yuv2rgb24_X_c;

  1441.             break;

  1442.         case PIX_FMT_BGR24:

  1443.             *yuv2packed1 = yuv2bgr24_1_c;

  1444.             *yuv2packed2 = yuv2bgr24_2_c;

  1445.             *yuv2packedX = yuv2bgr24_X_c;

  1446.             break;

  1447.         case PIX_FMT_RGB565LE:

  1448.         case PIX_FMT_RGB565BE:

  1449.         case PIX_FMT_BGR565LE:

  1450.         case PIX_FMT_BGR565BE:

  1451.             *yuv2packed1 = yuv2rgb16_1_c;

  1452.             *yuv2packed2 = yuv2rgb16_2_c;

  1453.             *yuv2packedX = yuv2rgb16_X_c;

  1454.             break;

  1455.         case PIX_FMT_RGB555LE:

  1456.         case PIX_FMT_RGB555BE:

  1457.         case PIX_FMT_BGR555LE:

  1458.         case PIX_FMT_BGR555BE:

  1459.             *yuv2packed1 = yuv2rgb15_1_c;

  1460.             *yuv2packed2 = yuv2rgb15_2_c;

  1461.             *yuv2packedX = yuv2rgb15_X_c;

  1462.             break;

  1463.         case PIX_FMT_RGB444LE:

  1464.         case PIX_FMT_RGB444BE:

  1465.         case PIX_FMT_BGR444LE:

  1466.         case PIX_FMT_BGR444BE:

  1467.             *yuv2packed1 = yuv2rgb12_1_c;

  1468.             *yuv2packed2 = yuv2rgb12_2_c;

  1469.             *yuv2packedX = yuv2rgb12_X_c;

  1470.             break;

  1471.         case PIX_FMT_RGB8:

  1472.         case PIX_FMT_BGR8:

  1473.             *yuv2packed1 = yuv2rgb8_1_c;

  1474.             *yuv2packed2 = yuv2rgb8_2_c;

  1475.             *yuv2packedX = yuv2rgb8_X_c;

  1476.             break;

  1477.         case PIX_FMT_RGB4:

  1478.         case PIX_FMT_BGR4:

  1479.             *yuv2packed1 = yuv2rgb4_1_c;

  1480.             *yuv2packed2 = yuv2rgb4_2_c;

  1481.             *yuv2packedX = yuv2rgb4_X_c;

  1482.             break;

  1483.         case PIX_FMT_RGB4_BYTE:

  1484.         case PIX_FMT_BGR4_BYTE:

  1485.             *yuv2packed1 = yuv2rgb4b_1_c;

  1486.             *yuv2packed2 = yuv2rgb4b_2_c;

  1487.             *yuv2packedX = yuv2rgb4b_X_c;

  1488.             break;

  1489.         }

  1490.     }

  1491.     switch (dstFormat) {

  1492.     case PIX_FMT_GRAY16BE:

  1493.         *yuv2packed1 = yuv2gray16BE_1_c;

  1494.         *yuv2packed2 = yuv2gray16BE_2_c;

  1495.         *yuv2packedX = yuv2gray16BE_X_c;

  1496.         break;

  1497.     case PIX_FMT_GRAY16LE:

  1498.         *yuv2packed1 = yuv2gray16LE_1_c;

  1499.         *yuv2packed2 = yuv2gray16LE_2_c;

  1500.         *yuv2packedX = yuv2gray16LE_X_c;

  1501.         break;

  1502.     case PIX_FMT_MONOWHITE:

  1503.         *yuv2packed1 = yuv2monowhite_1_c;

  1504.         *yuv2packed2 = yuv2monowhite_2_c;

  1505.         *yuv2packedX = yuv2monowhite_X_c;

  1506.         break;

  1507.     case PIX_FMT_MONOBLACK:

  1508.         *yuv2packed1 = yuv2monoblack_1_c;

  1509.         *yuv2packed2 = yuv2monoblack_2_c;

  1510.         *yuv2packedX = yuv2monoblack_X_c;

  1511.         break;

  1512.     case PIX_FMT_YUYV422:

  1513.         *yuv2packed1 = yuv2yuyv422_1_c;

  1514.         *yuv2packed2 = yuv2yuyv422_2_c;

  1515.         *yuv2packedX = yuv2yuyv422_X_c;

  1516.         break;

  1517.     case PIX_FMT_UYVY422:

  1518.         *yuv2packed1 = yuv2uyvy422_1_c;

  1519.         *yuv2packed2 = yuv2uyvy422_2_c;

  1520.         *yuv2packedX = yuv2uyvy422_X_c;

  1521.         break;

  1522.     }

  1523. }