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.
37884 Commits
32 Branches
353MB
Tree: cba4e2cbbc
Branches Tags
${ item.name }
Create branch ${ searchTerm }
from 'cba4e2cbbc'
${ noResults }
FFmpeg/libswscale/output.c

1456 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 output_pixel(pos, acc) \

  302.     if (target == PIX_FMT_MONOBLACK) { \

  303.         pos = acc; \

  304.     } else { \

  305.         pos = ~acc; \

  306.     }

  307. 

  308. static av_always_inline void

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

  310.                       const int16_t **lumSrc, int lumFilterSize,

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

  312.                       const int16_t **chrVSrc, int chrFilterSize,

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

  314.                       int y, enum PixelFormat target)

  315. {

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

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

  318.     int i;

  319.     unsigned acc = 0;

  320. 

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

  322.         int j;

  323.         int Y1 = 1 << 18;

  324.         int Y2 = 1 << 18;

  325. 

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

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

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

  329.         }

  330.         Y1 >>= 19;

  331.         Y2 >>= 19;

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

  333.             Y1 = av_clip_uint8(Y1);

  334.             Y2 = av_clip_uint8(Y2);

  335.         }

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

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

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

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

  340.         }

  341.     }

  342. }

  343. 

  344. static av_always_inline void

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

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

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

  348.                       int yalpha, int uvalpha, int y,

  349.                       enum PixelFormat target)

  350. {

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

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

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

  354.     int  yalpha1 = 4095 - yalpha;

  355.     int i;

  356. 

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

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

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

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

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

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

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

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

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

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

  367.     }

  368. }

  369. 

  370. static av_always_inline void

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

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

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

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

  375. {

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

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

  378.     int i;

  379. 

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

  381.         int acc =    g[((buf0[i    ] + 64) >> 7) + d128[0]];

  382.         acc += acc + g[((buf0[i + 1] + 64) >> 7) + d128[1]];

  383.         acc += acc + g[((buf0[i + 2] + 64) >> 7) + d128[2]];

  384.         acc += acc + g[((buf0[i + 3] + 64) >> 7) + d128[3]];

  385.         acc += acc + g[((buf0[i + 4] + 64) >> 7) + d128[4]];

  386.         acc += acc + g[((buf0[i + 5] + 64) >> 7) + d128[5]];

  387.         acc += acc + g[((buf0[i + 6] + 64) >> 7) + d128[6]];

  388.         acc += acc + g[((buf0[i + 7] + 64) >> 7) + d128[7]];

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

  390.     }

  391. }

  392. 

  393. #undef output_pixel

  394. 

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

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

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

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

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

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

  401.                                 int y) \

  402. { \

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

  404.                                   chrFilter, chrUSrc, chrVSrc, chrFilterSize, \

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

  406. } \

  407.  \

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

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

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

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

  412. { \

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

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

  415. } \

  416.  \

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

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

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

  420.                                 int uvalpha, int y) \

  421. { \

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

  423.                                   abuf0, dest, dstW, uvalpha, \

  424.                                   y, fmt); \

  425. }

  426. 

  427. YUV2PACKEDWRAPPER(yuv2mono,, white, PIX_FMT_MONOWHITE)

  428. YUV2PACKEDWRAPPER(yuv2mono,, black, PIX_FMT_MONOBLACK)

  429. 

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

  431.     if (target == PIX_FMT_YUYV422) { \

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

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

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

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

  436.     } else { \

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

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

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

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

  441.     }

  442. 

  443. static av_always_inline void

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

  445.                      const int16_t **lumSrc, int lumFilterSize,

  446.                      const int16_t *chrFilter, const int16_t **chrUSrc,

  447.                      const int16_t **chrVSrc, int chrFilterSize,

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

  449.                      int y, enum PixelFormat target)

  450. {

  451.     int i;

  452. 

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

  454.         int j;

  455.         int Y1 = 1 << 18;

  456.         int Y2 = 1 << 18;

  457.         int U  = 1 << 18;

  458.         int V  = 1 << 18;

  459. 

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

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

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

  463.         }

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

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

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

  467.         }

  468.         Y1 >>= 19;

  469.         Y2 >>= 19;

  470.         U  >>= 19;

  471.         V  >>= 19;

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

  473.             Y1 = av_clip_uint8(Y1);

  474.             Y2 = av_clip_uint8(Y2);

  475.             U  = av_clip_uint8(U);

  476.             V  = av_clip_uint8(V);

  477.         }

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

  479.     }

  480. }

  481. 

  482. static av_always_inline void

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

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

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

  486.                      int yalpha, int uvalpha, int y,

  487.                      enum PixelFormat target)

  488. {

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

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

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

  492.     int  yalpha1 = 4095 - yalpha;

  493.     int uvalpha1 = 4095 - uvalpha;

  494.     int i;

  495. 

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

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

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

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

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

  501. 

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

  503.             Y1 = av_clip_uint8(Y1);

  504.             Y2 = av_clip_uint8(Y2);

  505.             U  = av_clip_uint8(U);

  506.             V  = av_clip_uint8(V);

  507.         }

  508. 

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

  510.     }

  511. }

  512. 

  513. static av_always_inline void

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

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

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

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

  518. {

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

  520.     int i;

  521. 

  522.     if (uvalpha < 2048) {

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

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

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

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

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

  528. 

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

  530.                 Y1 = av_clip_uint8(Y1);

  531.                 Y2 = av_clip_uint8(Y2);

  532.                 U  = av_clip_uint8(U);

  533.                 V  = av_clip_uint8(V);

  534.             }

  535. 

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

  537.         }

  538.     } else {

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

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

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

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

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

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

  545. 

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

  547.                 Y1 = av_clip_uint8(Y1);

  548.                 Y2 = av_clip_uint8(Y2);

  549.                 U  = av_clip_uint8(U);

  550.                 V  = av_clip_uint8(V);

  551.             }

  552. 

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

  554.         }

  555.     }

  556. }

  557. 

  558. #undef output_pixels

  559. 

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

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

  562. 

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

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

  565. #define output_pixel(pos, val) \

  566.     if (isBE(target)) { \

  567.         AV_WB16(pos, val); \

  568.     } else { \

  569.         AV_WL16(pos, val); \

  570.     }

  571. 

  572. static av_always_inline void

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

  574.                        const int32_t **lumSrc, int lumFilterSize,

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

  576.                        const int32_t **chrVSrc, int chrFilterSize,

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

  578.                        int y, enum PixelFormat target)

  579. {

  580.     int i;

  581. 

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

  583.         int j;

  584.         int Y1 = -0x40000000;

  585.         int Y2 = -0x40000000;

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

  587.         int V  = -128 << 23;

  588.         int R, G, B;

  589. 

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

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

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

  593.         }

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

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

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

  597.         }

  598. 

  599.         // 8bit: 12+15=27; 16-bit: 12+19=31

  600.         Y1 >>= 14; // 10

  601.         Y1 += 0x10000;

  602.         Y2 >>= 14;

  603.         Y2 += 0x10000;

  604.         U  >>= 14;

  605.         V  >>= 14;

  606. 

  607.         // 8bit: 27 -> 17bit, 16bit: 31 - 14 = 17bit

  608.         Y1 -= c->yuv2rgb_y_offset;

  609.         Y2 -= c->yuv2rgb_y_offset;

  610.         Y1 *= c->yuv2rgb_y_coeff;

  611.         Y2 *= c->yuv2rgb_y_coeff;

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

  613.         Y2 += 1 << 13;

  614.         // 8bit: 17 + 13bit = 30bit, 16bit: 17 + 13bit = 30bit

  615. 

  616.         R = V * c->yuv2rgb_v2r_coeff;

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

  618.         B =                            U * c->yuv2rgb_u2b_coeff;

  619. 

  620.         // 8bit: 30 - 22 = 8bit, 16bit: 30bit - 14 = 16bit

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

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

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

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

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

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

  627.         dest += 6;

  628.     }

  629. }

  630. 

  631. static av_always_inline void

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

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

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

  635.                        int yalpha, int uvalpha, int y,

  636.                        enum PixelFormat target)

  637. {

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

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

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

  641.     int  yalpha1 = 4095 - yalpha;

  642.     int uvalpha1 = 4095 - uvalpha;

  643.     int i;

  644. 

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

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

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

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

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

  650.         int R, G, B;

  651. 

  652.         Y1 -= c->yuv2rgb_y_offset;

  653.         Y2 -= c->yuv2rgb_y_offset;

  654.         Y1 *= c->yuv2rgb_y_coeff;

  655.         Y2 *= c->yuv2rgb_y_coeff;

  656.         Y1 += 1 << 13;

  657.         Y2 += 1 << 13;

  658. 

  659.         R = V * c->yuv2rgb_v2r_coeff;

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

  661.         B =                            U * c->yuv2rgb_u2b_coeff;

  662. 

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

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

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

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

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

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

  669.         dest += 6;

  670.     }

  671. }

  672. 

  673. static av_always_inline void

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

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

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

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

  678. {

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

  680.     int i;

  681. 

  682.     if (uvalpha < 2048) {

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

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

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

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

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

  688.             int R, G, B;

  689. 

  690.             Y1 -= c->yuv2rgb_y_offset;

  691.             Y2 -= c->yuv2rgb_y_offset;

  692.             Y1 *= c->yuv2rgb_y_coeff;

  693.             Y2 *= c->yuv2rgb_y_coeff;

  694.             Y1 += 1 << 13;

  695.             Y2 += 1 << 13;

  696. 

  697.             R = V * c->yuv2rgb_v2r_coeff;

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

  699.             B =                            U * c->yuv2rgb_u2b_coeff;

  700. 

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

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

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

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

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

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

  707.             dest += 6;

  708.         }

  709.     } else {

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

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

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

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

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

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

  716.             int R, G, B;

  717. 

  718.             Y1 -= c->yuv2rgb_y_offset;

  719.             Y2 -= c->yuv2rgb_y_offset;

  720.             Y1 *= c->yuv2rgb_y_coeff;

  721.             Y2 *= c->yuv2rgb_y_coeff;

  722.             Y1 += 1 << 13;

  723.             Y2 += 1 << 13;

  724. 

  725.             R = V * c->yuv2rgb_v2r_coeff;

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

  727.             B =                            U * c->yuv2rgb_u2b_coeff;

  728. 

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

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

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

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

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

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

  735.             dest += 6;

  736.         }

  737.     }

  738. }

  739. 

  740. #undef output_pixel

  741. #undef r_b

  742. #undef b_r

  743. 

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

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

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

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

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

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

  750.                         int y) \

  751. { \

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

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

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

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

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

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

  758.                           chrFilter, chrUSrc, chrVSrc, chrFilterSize, \

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

  760. } \

  761.  \

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

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

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

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

  766. { \

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

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

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

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

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

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

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

  774. } \

  775.  \

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

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

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

  779.                         int uvalpha, int y) \

  780. { \

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

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

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

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

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

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

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

  788. }

  789. 

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

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

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

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

  794. 

  795. /*

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

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

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

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

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

  801.  * correct RGB values into the destination buffer.

  802.  */

  803. static av_always_inline void

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

  805.               unsigned A1, unsigned A2,

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

  807.               enum PixelFormat target, int hasAlpha)

  808. {

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

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

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

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

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

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

  815. 

  816. #if CONFIG_SMALL

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

  818. 

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

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

  821. #else

  822.         if (hasAlpha) {

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

  824. 

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

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

  827.         } else {

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

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

  830.         }

  831. #endif

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

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

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

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

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

  837. 

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

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

  840. 

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

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

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

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

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

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

  847. #undef r_b

  848. #undef b_r

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

  850.                target == PIX_FMT_RGB555 || target == PIX_FMT_BGR555 ||

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

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

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

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

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

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

  857. 

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

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

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

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

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

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

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

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

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

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

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

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

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

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

  872.         } else {

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

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

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

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

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

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

  879.         }

  880. 

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

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

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

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

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

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

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

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

  889. 

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

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

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

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

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

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

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

  897.         } else {

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

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

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

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

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

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

  904.         }

  905. 

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

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

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

  909.         } else {

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

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

  912.         }

  913.     }

  914. }

  915. 

  916. static av_always_inline void

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

  918.                      const int16_t **lumSrc, int lumFilterSize,

  919.                      const int16_t *chrFilter, const int16_t **chrUSrc,

  920.                      const int16_t **chrVSrc, int chrFilterSize,

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

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

  923. {

  924.     int i;

  925. 

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

  927.         int j, A1, A2;

  928.         int Y1 = 1 << 18;

  929.         int Y2 = 1 << 18;

  930.         int U  = 1 << 18;

  931.         int V  = 1 << 18;

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

  933. 

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

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

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

  937.         }

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

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

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

  941.         }

  942.         Y1 >>= 19;

  943.         Y2 >>= 19;

  944.         U  >>= 19;

  945.         V  >>= 19;

  946.         if (hasAlpha) {

  947.             A1 = 1 << 18;

  948.             A2 = 1 << 18;

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

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

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

  952.             }

  953.             A1 >>= 19;

  954.             A2 >>= 19;

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

  956.                 A1 = av_clip_uint8(A1);

  957.                 A2 = av_clip_uint8(A2);

  958.             }

  959.         }

  960. 

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

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

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

  964. 

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

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

  967.     }

  968. }

  969. 

  970. static av_always_inline void

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

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

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

  974.                      int yalpha, int uvalpha, int y,

  975.                      enum PixelFormat target, int hasAlpha)

  976. {

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

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

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

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

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

  982.     int  yalpha1 = 4095 - yalpha;

  983.     int uvalpha1 = 4095 - uvalpha;

  984.     int i;

  985. 

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

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

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

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

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

  991.         int A1, A2;

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

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

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

  995. 

  996.         if (hasAlpha) {

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

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

  999.         }

  1000. 

  1001.         yuv2rgb_write(dest, i, Y1, Y2, hasAlpha ? A1 : 0, hasAlpha ? A2 : 0,

  1002.                       r, g, b, y, target, hasAlpha);

  1003.     }

  1004. }

  1005. 

  1006. static av_always_inline void

  1007. yuv2rgb_1_c_template(SwsContext *c, const int16_t *buf0,

  1008.                      const int16_t *ubuf[2], const int16_t *vbuf[2],

  1009.                      const int16_t *abuf0, uint8_t *dest, int dstW,

  1010.                      int uvalpha, int y, enum PixelFormat target,

  1011.                      int hasAlpha)

  1012. {

  1013.     const int16_t *ubuf0 = ubuf[0], *vbuf0 = vbuf[0];

  1014.     int i;

  1015. 

  1016.     if (uvalpha < 2048) {

  1017.         for (i = 0; i < (dstW >> 1); i++) {

  1018.             int Y1 = (buf0[i * 2    ] + 64) >> 7;

  1019.             int Y2 = (buf0[i * 2 + 1] + 64) >> 7;

  1020.             int U  = (ubuf0[i]        + 64) >> 7;

  1021.             int V  = (vbuf0[i]        + 64) >> 7;

  1022.             int A1, A2;

  1023.             const void *r =  c->table_rV[V + YUVRGB_TABLE_HEADROOM],

  1024.                        *g = (c->table_gU[U + YUVRGB_TABLE_HEADROOM] + c->table_gV[V + YUVRGB_TABLE_HEADROOM]),

  1025.                        *b =  c->table_bU[U + YUVRGB_TABLE_HEADROOM];

  1026. 

  1027.             if (hasAlpha) {

  1028.                 A1 = (abuf0[i * 2    ] + 64) >> 7;

  1029.                 A2 = (abuf0[i * 2 + 1] + 64) >> 7;

  1030.             }

  1031. 

  1032.             yuv2rgb_write(dest, i, Y1, Y2, hasAlpha ? A1 : 0, hasAlpha ? A2 : 0,

  1033.                           r, g, b, y, target, hasAlpha);

  1034.         }

  1035.     } else {

  1036.         const int16_t *ubuf1 = ubuf[1], *vbuf1 = vbuf[1];

  1037.         for (i = 0; i < (dstW >> 1); i++) {

  1038.             int Y1 = (buf0[i * 2    ]     +  64) >> 7;

  1039.             int Y2 = (buf0[i * 2 + 1]     +  64) >> 7;

  1040.             int U  = (ubuf0[i] + ubuf1[i] + 128) >> 8;

  1041.             int V  = (vbuf0[i] + vbuf1[i] + 128) >> 8;

  1042.             int A1, A2;

  1043.             const void *r =  c->table_rV[V + YUVRGB_TABLE_HEADROOM],

  1044.                        *g = (c->table_gU[U + YUVRGB_TABLE_HEADROOM] + c->table_gV[V + YUVRGB_TABLE_HEADROOM]),

  1045.                        *b =  c->table_bU[U + YUVRGB_TABLE_HEADROOM];

  1046. 

  1047.             if (hasAlpha) {

  1048.                 A1 = (abuf0[i * 2    ] + 64) >> 7;

  1049.                 A2 = (abuf0[i * 2 + 1] + 64) >> 7;

  1050.             }

  1051. 

  1052.             yuv2rgb_write(dest, i, Y1, Y2, hasAlpha ? A1 : 0, hasAlpha ? A2 : 0,

  1053.                           r, g, b, y, target, hasAlpha);

  1054.         }

  1055.     }

  1056. }

  1057. 

  1058. #define YUV2RGBWRAPPERX(name, base, ext, fmt, hasAlpha) \

  1059. static void name ## ext ## _X_c(SwsContext *c, const int16_t *lumFilter, \

  1060.                                 const int16_t **lumSrc, int lumFilterSize, \

  1061.                                 const int16_t *chrFilter, const int16_t **chrUSrc, \

  1062.                                 const int16_t **chrVSrc, int chrFilterSize, \

  1063.                                 const int16_t **alpSrc, uint8_t *dest, int dstW, \

  1064.                                 int y) \

  1065. { \

  1066.     name ## base ## _X_c_template(c, lumFilter, lumSrc, lumFilterSize, \

  1067.                                   chrFilter, chrUSrc, chrVSrc, chrFilterSize, \

  1068.                                   alpSrc, dest, dstW, y, fmt, hasAlpha); \

  1069. }

  1070. #define YUV2RGBWRAPPER(name, base, ext, fmt, hasAlpha) \

  1071. YUV2RGBWRAPPERX(name, base, ext, fmt, hasAlpha) \

  1072. static void name ## ext ## _2_c(SwsContext *c, const int16_t *buf[2], \

  1073.                                 const int16_t *ubuf[2], const int16_t *vbuf[2], \

  1074.                                 const int16_t *abuf[2], uint8_t *dest, int dstW, \

  1075.                                 int yalpha, int uvalpha, int y) \

  1076. { \

  1077.     name ## base ## _2_c_template(c, buf, ubuf, vbuf, abuf, \

  1078.                                   dest, dstW, yalpha, uvalpha, y, fmt, hasAlpha); \

  1079. } \

  1080.  \

  1081. static void name ## ext ## _1_c(SwsContext *c, const int16_t *buf0, \

  1082.                                 const int16_t *ubuf[2], const int16_t *vbuf[2], \

  1083.                                 const int16_t *abuf0, uint8_t *dest, int dstW, \

  1084.                                 int uvalpha, int y) \

  1085. { \

  1086.     name ## base ## _1_c_template(c, buf0, ubuf, vbuf, abuf0, dest, \

  1087.                                   dstW, uvalpha, y, fmt, hasAlpha); \

  1088. }

  1089. 

  1090. #if CONFIG_SMALL

  1091. YUV2RGBWRAPPER(yuv2rgb,,  32_1,  PIX_FMT_RGB32_1,   CONFIG_SWSCALE_ALPHA && c->alpPixBuf)

  1092. YUV2RGBWRAPPER(yuv2rgb,,  32,    PIX_FMT_RGB32,     CONFIG_SWSCALE_ALPHA && c->alpPixBuf)

  1093. #else

  1094. #if CONFIG_SWSCALE_ALPHA

  1095. YUV2RGBWRAPPER(yuv2rgb,, a32_1,  PIX_FMT_RGB32_1,   1)

  1096. YUV2RGBWRAPPER(yuv2rgb,, a32,    PIX_FMT_RGB32,     1)

  1097. #endif

  1098. YUV2RGBWRAPPER(yuv2rgb,, x32_1,  PIX_FMT_RGB32_1,   0)

  1099. YUV2RGBWRAPPER(yuv2rgb,, x32,    PIX_FMT_RGB32,     0)

  1100. #endif

  1101. YUV2RGBWRAPPER(yuv2, rgb, rgb24, PIX_FMT_RGB24,   0)

  1102. YUV2RGBWRAPPER(yuv2, rgb, bgr24, PIX_FMT_BGR24,   0)

  1103. YUV2RGBWRAPPER(yuv2rgb,,  16,    PIX_FMT_RGB565,    0)

  1104. YUV2RGBWRAPPER(yuv2rgb,,  15,    PIX_FMT_RGB555,    0)

  1105. YUV2RGBWRAPPER(yuv2rgb,,  12,    PIX_FMT_RGB444,    0)

  1106. YUV2RGBWRAPPER(yuv2rgb,,   8,    PIX_FMT_RGB8,      0)

  1107. YUV2RGBWRAPPER(yuv2rgb,,   4,    PIX_FMT_RGB4,      0)

  1108. YUV2RGBWRAPPER(yuv2rgb,,   4b,   PIX_FMT_RGB4_BYTE, 0)

  1109. 

  1110. static av_always_inline void

  1111. yuv2rgb_full_X_c_template(SwsContext *c, const int16_t *lumFilter,

  1112.                           const int16_t **lumSrc, int lumFilterSize,

  1113.                           const int16_t *chrFilter, const int16_t **chrUSrc,

  1114.                           const int16_t **chrVSrc, int chrFilterSize,

  1115.                           const int16_t **alpSrc, uint8_t *dest,

  1116.                           int dstW, int y, enum PixelFormat target, int hasAlpha)

  1117. {

  1118.     int i;

  1119.     int step = (target == PIX_FMT_RGB24 || target == PIX_FMT_BGR24) ? 3 : 4;

  1120. 

  1121.     for (i = 0; i < dstW; i++) {

  1122.         int j;

  1123.         int Y = 1<<9;

  1124.         int U = (1<<9)-(128 << 19);

  1125.         int V = (1<<9)-(128 << 19);

  1126.         int R, G, B, A;

  1127. 

  1128.         for (j = 0; j < lumFilterSize; j++) {

  1129.             Y += lumSrc[j][i] * lumFilter[j];

  1130.         }

  1131.         for (j = 0; j < chrFilterSize; j++) {

  1132.             U += chrUSrc[j][i] * chrFilter[j];

  1133.             V += chrVSrc[j][i] * chrFilter[j];

  1134.         }

  1135.         Y >>= 10;

  1136.         U >>= 10;

  1137.         V >>= 10;

  1138.         if (hasAlpha) {

  1139.             A = 1 << 18;

  1140.             for (j = 0; j < lumFilterSize; j++) {

  1141.                 A += alpSrc[j][i] * lumFilter[j];

  1142.             }

  1143.             A >>= 19;

  1144.             if (A & 0x100)

  1145.                 A = av_clip_uint8(A);

  1146.         }

  1147.         Y -= c->yuv2rgb_y_offset;

  1148.         Y *= c->yuv2rgb_y_coeff;

  1149.         Y += 1 << 21;

  1150.         R = Y + V*c->yuv2rgb_v2r_coeff;

  1151.         G = Y + V*c->yuv2rgb_v2g_coeff + U*c->yuv2rgb_u2g_coeff;

  1152.         B = Y +                          U*c->yuv2rgb_u2b_coeff;

  1153.         if ((R | G | B) & 0xC0000000) {

  1154.             R = av_clip_uintp2(R, 30);

  1155.             G = av_clip_uintp2(G, 30);

  1156.             B = av_clip_uintp2(B, 30);

  1157.         }

  1158. 

  1159.         switch(target) {

  1160.         case PIX_FMT_ARGB:

  1161.             dest[0] = hasAlpha ? A : 255;

  1162.             dest[1] = R >> 22;

  1163.             dest[2] = G >> 22;

  1164.             dest[3] = B >> 22;

  1165.             break;

  1166.         case PIX_FMT_RGB24:

  1167.             dest[0] = R >> 22;

  1168.             dest[1] = G >> 22;

  1169.             dest[2] = B >> 22;

  1170.             break;

  1171.         case PIX_FMT_RGBA:

  1172.             dest[0] = R >> 22;

  1173.             dest[1] = G >> 22;

  1174.             dest[2] = B >> 22;

  1175.             dest[3] = hasAlpha ? A : 255;

  1176.             break;

  1177.         case PIX_FMT_ABGR:

  1178.             dest[0] = hasAlpha ? A : 255;

  1179.             dest[1] = B >> 22;

  1180.             dest[2] = G >> 22;

  1181.             dest[3] = R >> 22;

  1182.             break;

  1183.         case PIX_FMT_BGR24:

  1184.             dest[0] = B >> 22;

  1185.             dest[1] = G >> 22;

  1186.             dest[2] = R >> 22;

  1187.             break;

  1188.         case PIX_FMT_BGRA:

  1189.             dest[0] = B >> 22;

  1190.             dest[1] = G >> 22;

  1191.             dest[2] = R >> 22;

  1192.             dest[3] = hasAlpha ? A : 255;

  1193.             break;

  1194.         }

  1195.         dest += step;

  1196.     }

  1197. }

  1198. 

  1199. #if CONFIG_SMALL

  1200. YUV2RGBWRAPPERX(yuv2, rgb_full, bgra32_full, PIX_FMT_BGRA,  CONFIG_SWSCALE_ALPHA && c->alpPixBuf)

  1201. YUV2RGBWRAPPERX(yuv2, rgb_full, abgr32_full, PIX_FMT_ABGR,  CONFIG_SWSCALE_ALPHA && c->alpPixBuf)

  1202. YUV2RGBWRAPPERX(yuv2, rgb_full, rgba32_full, PIX_FMT_RGBA,  CONFIG_SWSCALE_ALPHA && c->alpPixBuf)

  1203. YUV2RGBWRAPPERX(yuv2, rgb_full, argb32_full, PIX_FMT_ARGB,  CONFIG_SWSCALE_ALPHA && c->alpPixBuf)

  1204. #else

  1205. #if CONFIG_SWSCALE_ALPHA

  1206. YUV2RGBWRAPPERX(yuv2, rgb_full, bgra32_full, PIX_FMT_BGRA,  1)

  1207. YUV2RGBWRAPPERX(yuv2, rgb_full, abgr32_full, PIX_FMT_ABGR,  1)

  1208. YUV2RGBWRAPPERX(yuv2, rgb_full, rgba32_full, PIX_FMT_RGBA,  1)

  1209. YUV2RGBWRAPPERX(yuv2, rgb_full, argb32_full, PIX_FMT_ARGB,  1)

  1210. #endif

  1211. YUV2RGBWRAPPERX(yuv2, rgb_full, bgrx32_full, PIX_FMT_BGRA,  0)

  1212. YUV2RGBWRAPPERX(yuv2, rgb_full, xbgr32_full, PIX_FMT_ABGR,  0)

  1213. YUV2RGBWRAPPERX(yuv2, rgb_full, rgbx32_full, PIX_FMT_RGBA,  0)

  1214. YUV2RGBWRAPPERX(yuv2, rgb_full, xrgb32_full, PIX_FMT_ARGB,  0)

  1215. #endif

  1216. YUV2RGBWRAPPERX(yuv2, rgb_full, bgr24_full,  PIX_FMT_BGR24, 0)

  1217. YUV2RGBWRAPPERX(yuv2, rgb_full, rgb24_full,  PIX_FMT_RGB24, 0)

  1218. 

  1219. void ff_sws_init_output_funcs(SwsContext *c,

  1220.                               yuv2planar1_fn *yuv2plane1,

  1221.                               yuv2planarX_fn *yuv2planeX,

  1222.                               yuv2interleavedX_fn *yuv2nv12cX,

  1223.                               yuv2packed1_fn *yuv2packed1,

  1224.                               yuv2packed2_fn *yuv2packed2,

  1225.                               yuv2packedX_fn *yuv2packedX)

  1226. {

  1227.     enum PixelFormat dstFormat = c->dstFormat;

  1228. 

  1229.     if (is16BPS(dstFormat)) {

  1230.         *yuv2planeX = isBE(dstFormat) ? yuv2planeX_16BE_c  : yuv2planeX_16LE_c;

  1231.         *yuv2plane1 = isBE(dstFormat) ? yuv2plane1_16BE_c  : yuv2plane1_16LE_c;

  1232.     } else if (is9_OR_10BPS(dstFormat)) {

  1233.         if (av_pix_fmt_descriptors[dstFormat].comp[0].depth_minus1 == 8) {

  1234.             *yuv2planeX = isBE(dstFormat) ? yuv2planeX_9BE_c  : yuv2planeX_9LE_c;

  1235.             *yuv2plane1 = isBE(dstFormat) ? yuv2plane1_9BE_c  : yuv2plane1_9LE_c;

  1236.         } else {

  1237.             *yuv2planeX = isBE(dstFormat) ? yuv2planeX_10BE_c  : yuv2planeX_10LE_c;

  1238.             *yuv2plane1 = isBE(dstFormat) ? yuv2plane1_10BE_c  : yuv2plane1_10LE_c;

  1239.         }

  1240.     } else {

  1241.         *yuv2plane1 = yuv2plane1_8_c;

  1242.         *yuv2planeX = yuv2planeX_8_c;

  1243.         if (dstFormat == PIX_FMT_NV12 || dstFormat == PIX_FMT_NV21)

  1244.             *yuv2nv12cX = yuv2nv12cX_c;

  1245.     }

  1246. 

  1247.     if(c->flags & SWS_FULL_CHR_H_INT) {

  1248.         switch (dstFormat) {

  1249.             case PIX_FMT_RGBA:

  1250. #if CONFIG_SMALL

  1251.                 *yuv2packedX = yuv2rgba32_full_X_c;

  1252. #else

  1253. #if CONFIG_SWSCALE_ALPHA

  1254.                 if (c->alpPixBuf) {

  1255.                     *yuv2packedX = yuv2rgba32_full_X_c;

  1256.                 } else

  1257. #endif /* CONFIG_SWSCALE_ALPHA */

  1258.                 {

  1259.                     *yuv2packedX = yuv2rgbx32_full_X_c;

  1260.                 }

  1261. #endif /* !CONFIG_SMALL */

  1262.                 break;

  1263.             case PIX_FMT_ARGB:

  1264. #if CONFIG_SMALL

  1265.                 *yuv2packedX = yuv2argb32_full_X_c;

  1266. #else

  1267. #if CONFIG_SWSCALE_ALPHA

  1268.                 if (c->alpPixBuf) {

  1269.                     *yuv2packedX = yuv2argb32_full_X_c;

  1270.                 } else

  1271. #endif /* CONFIG_SWSCALE_ALPHA */

  1272.                 {

  1273.                     *yuv2packedX = yuv2xrgb32_full_X_c;

  1274.                 }

  1275. #endif /* !CONFIG_SMALL */

  1276.                 break;

  1277.             case PIX_FMT_BGRA:

  1278. #if CONFIG_SMALL

  1279.                 *yuv2packedX = yuv2bgra32_full_X_c;

  1280. #else

  1281. #if CONFIG_SWSCALE_ALPHA

  1282.                 if (c->alpPixBuf) {

  1283.                     *yuv2packedX = yuv2bgra32_full_X_c;

  1284.                 } else

  1285. #endif /* CONFIG_SWSCALE_ALPHA */

  1286.                 {

  1287.                     *yuv2packedX = yuv2bgrx32_full_X_c;

  1288.                 }

  1289. #endif /* !CONFIG_SMALL */

  1290.                 break;

  1291.             case PIX_FMT_ABGR:

  1292. #if CONFIG_SMALL

  1293.                 *yuv2packedX = yuv2abgr32_full_X_c;

  1294. #else

  1295. #if CONFIG_SWSCALE_ALPHA

  1296.                 if (c->alpPixBuf) {

  1297.                     *yuv2packedX = yuv2abgr32_full_X_c;

  1298.                 } else

  1299. #endif /* CONFIG_SWSCALE_ALPHA */

  1300.                 {

  1301.                     *yuv2packedX = yuv2xbgr32_full_X_c;

  1302.                 }

  1303. #endif /* !CONFIG_SMALL */

  1304.                 break;

  1305.             case PIX_FMT_RGB24:

  1306.             *yuv2packedX = yuv2rgb24_full_X_c;

  1307.             break;

  1308.         case PIX_FMT_BGR24:

  1309.             *yuv2packedX = yuv2bgr24_full_X_c;

  1310.             break;

  1311.         }

  1312.         if(!*yuv2packedX)

  1313.             goto YUV_PACKED;

  1314.     } else {

  1315.         YUV_PACKED:

  1316.         switch (dstFormat) {

  1317.         case PIX_FMT_RGB48LE:

  1318.             *yuv2packed1 = yuv2rgb48le_1_c;

  1319.             *yuv2packed2 = yuv2rgb48le_2_c;

  1320.             *yuv2packedX = yuv2rgb48le_X_c;

  1321.             break;

  1322.         case PIX_FMT_RGB48BE:

  1323.             *yuv2packed1 = yuv2rgb48be_1_c;

  1324.             *yuv2packed2 = yuv2rgb48be_2_c;

  1325.             *yuv2packedX = yuv2rgb48be_X_c;

  1326.             break;

  1327.         case PIX_FMT_BGR48LE:

  1328.             *yuv2packed1 = yuv2bgr48le_1_c;

  1329.             *yuv2packed2 = yuv2bgr48le_2_c;

  1330.             *yuv2packedX = yuv2bgr48le_X_c;

  1331.             break;

  1332.         case PIX_FMT_BGR48BE:

  1333.             *yuv2packed1 = yuv2bgr48be_1_c;

  1334.             *yuv2packed2 = yuv2bgr48be_2_c;

  1335.             *yuv2packedX = yuv2bgr48be_X_c;

  1336.             break;

  1337.         case PIX_FMT_RGB32:

  1338.         case PIX_FMT_BGR32:

  1339. #if CONFIG_SMALL

  1340.             *yuv2packed1 = yuv2rgb32_1_c;

  1341.             *yuv2packed2 = yuv2rgb32_2_c;

  1342.             *yuv2packedX = yuv2rgb32_X_c;

  1343. #else

  1344. #if CONFIG_SWSCALE_ALPHA

  1345.                 if (c->alpPixBuf) {

  1346.                     *yuv2packed1 = yuv2rgba32_1_c;

  1347.                     *yuv2packed2 = yuv2rgba32_2_c;

  1348.                     *yuv2packedX = yuv2rgba32_X_c;

  1349.                 } else

  1350. #endif /* CONFIG_SWSCALE_ALPHA */

  1351.                 {

  1352.                     *yuv2packed1 = yuv2rgbx32_1_c;

  1353.                     *yuv2packed2 = yuv2rgbx32_2_c;

  1354.                     *yuv2packedX = yuv2rgbx32_X_c;

  1355.                 }

  1356. #endif /* !CONFIG_SMALL */

  1357.             break;

  1358.         case PIX_FMT_RGB32_1:

  1359.         case PIX_FMT_BGR32_1:

  1360. #if CONFIG_SMALL

  1361.                 *yuv2packed1 = yuv2rgb32_1_1_c;

  1362.                 *yuv2packed2 = yuv2rgb32_1_2_c;

  1363.                 *yuv2packedX = yuv2rgb32_1_X_c;

  1364. #else

  1365. #if CONFIG_SWSCALE_ALPHA

  1366.                 if (c->alpPixBuf) {

  1367.                     *yuv2packed1 = yuv2rgba32_1_1_c;

  1368.                     *yuv2packed2 = yuv2rgba32_1_2_c;

  1369.                     *yuv2packedX = yuv2rgba32_1_X_c;

  1370.                 } else

  1371. #endif /* CONFIG_SWSCALE_ALPHA */

  1372.                 {

  1373.                     *yuv2packed1 = yuv2rgbx32_1_1_c;

  1374.                     *yuv2packed2 = yuv2rgbx32_1_2_c;

  1375.                     *yuv2packedX = yuv2rgbx32_1_X_c;

  1376.                 }

  1377. #endif /* !CONFIG_SMALL */

  1378.                 break;

  1379.         case PIX_FMT_RGB24:

  1380.             *yuv2packed1 = yuv2rgb24_1_c;

  1381.             *yuv2packed2 = yuv2rgb24_2_c;

  1382.             *yuv2packedX = yuv2rgb24_X_c;

  1383.             break;

  1384.         case PIX_FMT_BGR24:

  1385.             *yuv2packed1 = yuv2bgr24_1_c;

  1386.             *yuv2packed2 = yuv2bgr24_2_c;

  1387.             *yuv2packedX = yuv2bgr24_X_c;

  1388.             break;

  1389.         case PIX_FMT_RGB565LE:

  1390.         case PIX_FMT_RGB565BE:

  1391.         case PIX_FMT_BGR565LE:

  1392.         case PIX_FMT_BGR565BE:

  1393.             *yuv2packed1 = yuv2rgb16_1_c;

  1394.             *yuv2packed2 = yuv2rgb16_2_c;

  1395.             *yuv2packedX = yuv2rgb16_X_c;

  1396.             break;

  1397.         case PIX_FMT_RGB555LE:

  1398.         case PIX_FMT_RGB555BE:

  1399.         case PIX_FMT_BGR555LE:

  1400.         case PIX_FMT_BGR555BE:

  1401.             *yuv2packed1 = yuv2rgb15_1_c;

  1402.             *yuv2packed2 = yuv2rgb15_2_c;

  1403.             *yuv2packedX = yuv2rgb15_X_c;

  1404.             break;

  1405.         case PIX_FMT_RGB444LE:

  1406.         case PIX_FMT_RGB444BE:

  1407.         case PIX_FMT_BGR444LE:

  1408.         case PIX_FMT_BGR444BE:

  1409.             *yuv2packed1 = yuv2rgb12_1_c;

  1410.             *yuv2packed2 = yuv2rgb12_2_c;

  1411.             *yuv2packedX = yuv2rgb12_X_c;

  1412.             break;

  1413.         case PIX_FMT_RGB8:

  1414.         case PIX_FMT_BGR8:

  1415.             *yuv2packed1 = yuv2rgb8_1_c;

  1416.             *yuv2packed2 = yuv2rgb8_2_c;

  1417.             *yuv2packedX = yuv2rgb8_X_c;

  1418.             break;

  1419.         case PIX_FMT_RGB4:

  1420.         case PIX_FMT_BGR4:

  1421.             *yuv2packed1 = yuv2rgb4_1_c;

  1422.             *yuv2packed2 = yuv2rgb4_2_c;

  1423.             *yuv2packedX = yuv2rgb4_X_c;

  1424.             break;

  1425.         case PIX_FMT_RGB4_BYTE:

  1426.         case PIX_FMT_BGR4_BYTE:

  1427.             *yuv2packed1 = yuv2rgb4b_1_c;

  1428.             *yuv2packed2 = yuv2rgb4b_2_c;

  1429.             *yuv2packedX = yuv2rgb4b_X_c;

  1430.             break;

  1431.         }

  1432.     }

  1433.     switch (dstFormat) {

  1434.     case PIX_FMT_MONOWHITE:

  1435.         *yuv2packed1 = yuv2monowhite_1_c;

  1436.         *yuv2packed2 = yuv2monowhite_2_c;

  1437.         *yuv2packedX = yuv2monowhite_X_c;

  1438.         break;

  1439.     case PIX_FMT_MONOBLACK:

  1440.         *yuv2packed1 = yuv2monoblack_1_c;

  1441.         *yuv2packed2 = yuv2monoblack_2_c;

  1442.         *yuv2packedX = yuv2monoblack_X_c;

  1443.         break;

  1444.     case PIX_FMT_YUYV422:

  1445.         *yuv2packed1 = yuv2yuyv422_1_c;

  1446.         *yuv2packed2 = yuv2yuyv422_2_c;

  1447.         *yuv2packedX = yuv2yuyv422_X_c;

  1448.         break;

  1449.     case PIX_FMT_UYVY422:

  1450.         *yuv2packed1 = yuv2uyvy422_1_c;

  1451.         *yuv2packed2 = yuv2uyvy422_2_c;

  1452.         *yuv2packedX = yuv2uyvy422_X_c;

  1453.         break;

  1454.     }

  1455. }