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.
42813 Commits
32 Branches
367MB
Tree: 7afd42d9f2
Branches Tags
${ item.name }
Create branch ${ searchTerm }
from '7afd42d9f2'
${ noResults }
FFmpeg/libswscale/output.c

1500 lines
55KB
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, const uint8_t, dither_2x2_4)[][8]={

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

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

  42. {  1,   3,   1,   3,   1,   3,   1,   3, },

  43. };

  44. 

  45. DECLARE_ALIGNED(8, const uint8_t, dither_2x2_8)[][8]={

  46. {  6,   2,   6,   2,   6,   2,   6,   2, },

  47. {  0,   4,   0,   4,   0,   4,   0,   4, },

  48. {  6,   2,   6,   2,   6,   2,   6,   2, },

  49. };

  50. 

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

  52. {  8,   4,  11,   7,   8,   4,  11,   7, },

  53. {  2,  14,   1,  13,   2,  14,   1,  13, },

  54. { 10,   6,   9,   5,  10,   6,   9,   5, },

  55. {  0,  12,   3,  15,   0,  12,   3,  15, },

  56. {  8,   4,  11,   7,   8,   4,  11,   7, },

  57. };

  58. 

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

  60. { 17,   9,  23,  15,  16,   8,  22,  14, },

  61. {  5,  29,   3,  27,   4,  28,   2,  26, },

  62. { 21,  13,  19,  11,  20,  12,  18,  10, },

  63. {  0,  24,   6,  30,   1,  25,   7,  31, },

  64. { 16,   8,  22,  14,  17,   9,  23,  15, },

  65. {  4,  28,   2,  26,   5,  29,   3,  27, },

  66. { 20,  12,  18,  10,  21,  13,  19,  11, },

  67. {  1,  25,   7,  31,   0,  24,   6,  30, },

  68. { 17,   9,  23,  15,  16,   8,  22,  14, },

  69. };

  70. 

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

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

  73. { 37,  18,  50,  32,  40,  22,  54,  35, },

  74. {  9,  64,   5,  59,  13,  67,   8,  63, },

  75. { 46,  27,  41,  23,  49,  31,  44,  26, },

  76. {  2,  57,  16,  71,   1,  56,  15,  70, },

  77. { 39,  21,  52,  34,  38,  19,  51,  33, },

  78. { 11,  66,   7,  62,  10,  65,   6,  60, },

  79. { 48,  30,  43,  25,  47,  29,  42,  24, },

  80. {  0,  55,  14,  68,   3,  58,  17,  72, },

  81. };

  82. 

  83. #if 1

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

  85. {117,  62, 158, 103, 113,  58, 155, 100, },

  86. { 34, 199,  21, 186,  31, 196,  17, 182, },

  87. {144,  89, 131,  76, 141,  86, 127,  72, },

  88. {  0, 165,  41, 206,  10, 175,  52, 217, },

  89. {110,  55, 151,  96, 120,  65, 162, 107, },

  90. { 28, 193,  14, 179,  38, 203,  24, 189, },

  91. {138,  83, 124,  69, 148,  93, 134,  79, },

  92. {  7, 172,  48, 213,   3, 168,  45, 210, },

  93. {117,  62, 158, 103, 113,  58, 155, 100, },

  94. };

  95. #elif 1

  96. // tries to correct a gamma of 1.5

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

  98. {  0, 143,  18, 200,   2, 156,  25, 215, },

  99. { 78,  28, 125,  64,  89,  36, 138,  74, },

  100. { 10, 180,   3, 161,  16, 195,   8, 175, },

  101. {109,  51,  93,  38, 121,  60, 105,  47, },

  102. {  1, 152,  23, 210,   0, 147,  20, 205, },

  103. { 85,  33, 134,  71,  81,  30, 130,  67, },

  104. { 14, 190,   6, 171,  12, 185,   5, 166, },

  105. {117,  57, 101,  44, 113,  54,  97,  41, },

  106. {  0, 143,  18, 200,   2, 156,  25, 215, },

  107. };

  108. #elif 1

  109. // tries to correct a gamma of 2.0

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

  111. {  0, 124,   8, 193,   0, 140,  12, 213, },

  112. { 55,  14, 104,  42,  66,  19, 119,  52, },

  113. {  3, 168,   1, 145,   6, 187,   3, 162, },

  114. { 86,  31,  70,  21,  99,  39,  82,  28, },

  115. {  0, 134,  11, 206,   0, 129,   9, 200, },

  116. { 62,  17, 114,  48,  58,  16, 109,  45, },

  117. {  5, 181,   2, 157,   4, 175,   1, 151, },

  118. { 95,  36,  78,  26,  90,  34,  74,  24, },

  119. {  0, 124,   8, 193,   0, 140,  12, 213, },

  120. };

  121. #else

  122. // tries to correct a gamma of 2.5

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

  124. {  0, 107,   3, 187,   0, 125,   6, 212, },

  125. { 39,   7,  86,  28,  49,  11, 102,  36, },

  126. {  1, 158,   0, 131,   3, 180,   1, 151, },

  127. { 68,  19,  52,  12,  81,  25,  64,  17, },

  128. {  0, 119,   5, 203,   0, 113,   4, 195, },

  129. { 45,   9,  96,  33,  42,   8,  91,  30, },

  130. {  2, 172,   1, 144,   2, 165,   0, 137, },

  131. { 77,  23,  60,  15,  72,  21,  56,  14, },

  132. {  0, 107,   3, 187,   0, 125,   6, 212, },

  133. };

  134. #endif

  135. 

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

  137.     if (big_endian) { \

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

  139.     } else { \

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

  141.     }

  142. 

  143. static av_always_inline void

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

  145.                          int big_endian, int output_bits)

  146. {

  147.     int i;

  148.     int shift = 3;

  149.     av_assert0(output_bits == 16);

  150. 

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

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

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

  154.     }

  155. }

  156. 

  157. static av_always_inline void

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

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

  160.                          int big_endian, int output_bits)

  161. {

  162.     int i;

  163.     int shift = 15;

  164.     av_assert0(output_bits == 16);

  165. 

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

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

  168.         int j;

  169. 

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

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

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

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

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

  175.         val -= 0x40000000;

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

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

  178. 

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

  180.     }

  181. }

  182. 

  183. #undef output_pixel

  184. 

  185. #define output_pixel(pos, val) \

  186.     if (big_endian) { \

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

  188.     } else { \

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

  190.     }

  191. 

  192. static av_always_inline void

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

  194.                          int big_endian, int output_bits)

  195. {

  196.     int i;

  197.     int shift = 15 - output_bits;

  198. 

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

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

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

  202.     }

  203. }

  204. 

  205. static av_always_inline void

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

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

  208.                          int big_endian, int output_bits)

  209. {

  210.     int i;

  211.     int shift = 11 + 16 - output_bits;

  212. 

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

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

  215.         int j;

  216. 

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

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

  219. 

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

  221.     }

  222. }

  223. 

  224. #undef output_pixel

  225. 

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

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

  228.                               uint8_t *dest, int dstW, \

  229.                               const uint8_t *dither, int offset)\

  230. { \

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

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

  233. }\

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

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

  236.                               const uint8_t *dither, int offset)\

  237. { \

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

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

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

  241. }

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

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

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

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

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

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

  248. 

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

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

  251.                            const uint8_t *dither, int offset)

  252. {

  253.     int i;

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

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

  256.         int j;

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

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

  259. 

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

  261.     }

  262. }

  263. 

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

  265.                            const uint8_t *dither, int offset)

  266. {

  267.     int i;

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

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

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

  271.     }

  272. }

  273. 

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

  275.                         const int16_t **chrUSrc, const int16_t **chrVSrc,

  276.                         uint8_t *dest, int chrDstW)

  277. {

  278.     enum PixelFormat dstFormat = c->dstFormat;

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

  280.     int i;

  281. 

  282.     if (dstFormat == PIX_FMT_NV12)

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

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

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

  286.             int j;

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

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

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

  290.             }

  291. 

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

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

  294.         }

  295.     else

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

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

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

  299.             int j;

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

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

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

  303.             }

  304. 

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

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

  307.         }

  308. }

  309. 

  310. #define accumulate_bit(acc, val) \

  311.     acc <<= 1; \

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

  313. #define output_pixel(pos, acc) \

  314.     if (target == PIX_FMT_MONOBLACK) { \

  315.         pos = acc; \

  316.     } else { \

  317.         pos = ~acc; \

  318.     }

  319. 

  320. static av_always_inline void

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

  322.                       const int16_t **lumSrc, int lumFilterSize,

  323.                       const int16_t *chrFilter, const int16_t **chrUSrc,

  324.                       const int16_t **chrVSrc, int chrFilterSize,

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

  326.                       int y, enum PixelFormat target)

  327. {

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

  329.     int i;

  330.     unsigned acc = 0;

  331. 

  332.     for (i = 0; i < dstW; i += 2) {

  333.         int j;

  334.         int Y1 = 1 << 18;

  335.         int Y2 = 1 << 18;

  336. 

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

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

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

  340.         }

  341.         Y1 >>= 19;

  342.         Y2 >>= 19;

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

  344.             Y1 = av_clip_uint8(Y1);

  345.             Y2 = av_clip_uint8(Y2);

  346.         }

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

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

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

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

  351.         }

  352.     }

  353. 

  354.     if (i & 6) {

  355.         output_pixel(*dest, acc);

  356.     }

  357. }

  358. 

  359. static av_always_inline void

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

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

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

  363.                       int yalpha, int uvalpha, int y,

  364.                       enum PixelFormat target)

  365. {

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

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

  368.     int  yalpha1 = 4096 - yalpha;

  369.     int i;

  370. 

  371.     for (i = 0; i < dstW; i += 8) {

  372.         int Y, acc = 0;

  373. 

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

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

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

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

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

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

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

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

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

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

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

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

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

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

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

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

  390. 

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

  392.     }

  393. }

  394. 

  395. static av_always_inline void

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

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

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

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

  400. {

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

  402.     int i;

  403. 

  404.     for (i = 0; i < dstW; i += 8) {

  405.         int acc = 0;

  406. 

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

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

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

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

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

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

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

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

  415. 

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

  417.     }

  418. }

  419. 

  420. #undef output_pixel

  421. #undef accumulate_bit

  422. 

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

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

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

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

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

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

  429.                                 int y) \

  430. { \

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

  432.                                   chrFilter, chrUSrc, chrVSrc, chrFilterSize, \

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

  434. } \

  435.  \

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

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

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

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

  440. { \

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

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

  443. } \

  444.  \

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

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

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

  448.                                 int uvalpha, int y) \

  449. { \

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

  451.                                   abuf0, dest, dstW, uvalpha, \

  452.                                   y, fmt); \

  453. }

  454. 

  455. YUV2PACKEDWRAPPER(yuv2mono,, white, PIX_FMT_MONOWHITE)

  456. YUV2PACKEDWRAPPER(yuv2mono,, black, PIX_FMT_MONOBLACK)

  457. 

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

  459.     if (target == PIX_FMT_YUYV422) { \

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

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

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

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

  464.     } else { \

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

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

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

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

  469.     }

  470. 

  471. static av_always_inline void

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

  473.                      const int16_t **lumSrc, int lumFilterSize,

  474.                      const int16_t *chrFilter, const int16_t **chrUSrc,

  475.                      const int16_t **chrVSrc, int chrFilterSize,

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

  477.                      int y, enum PixelFormat target)

  478. {

  479.     int i;

  480. 

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

  482.         int j;

  483.         int Y1 = 1 << 18;

  484.         int Y2 = 1 << 18;

  485.         int U  = 1 << 18;

  486.         int V  = 1 << 18;

  487. 

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

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

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

  491.         }

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

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

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

  495.         }

  496.         Y1 >>= 19;

  497.         Y2 >>= 19;

  498.         U  >>= 19;

  499.         V  >>= 19;

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

  501.             Y1 = av_clip_uint8(Y1);

  502.             Y2 = av_clip_uint8(Y2);

  503.             U  = av_clip_uint8(U);

  504.             V  = av_clip_uint8(V);

  505.         }

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

  507.     }

  508. }

  509. 

  510. static av_always_inline void

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

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

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

  514.                      int yalpha, int uvalpha, int y,

  515.                      enum PixelFormat target)

  516. {

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

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

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

  520.     int  yalpha1 = 4096 - yalpha;

  521.     int uvalpha1 = 4096 - uvalpha;

  522.     int i;

  523. 

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

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

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

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

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

  529. 

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

  531.             Y1 = av_clip_uint8(Y1);

  532.             Y2 = av_clip_uint8(Y2);

  533.             U  = av_clip_uint8(U);

  534.             V  = av_clip_uint8(V);

  535.         }

  536. 

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

  538.     }

  539. }

  540. 

  541. static av_always_inline void

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

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

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

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

  546. {

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

  548.     int i;

  549. 

  550.     if (uvalpha < 2048) {

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

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

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

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

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

  556. 

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

  558.                 Y1 = av_clip_uint8(Y1);

  559.                 Y2 = av_clip_uint8(Y2);

  560.                 U  = av_clip_uint8(U);

  561.                 V  = av_clip_uint8(V);

  562.             }

  563. 

  564.             Y1 = av_clip_uint8(Y1);

  565.             Y2 = av_clip_uint8(Y2);

  566.             U  = av_clip_uint8(U);

  567.             V  = av_clip_uint8(V);

  568. 

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

  570.         }

  571.     } else {

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

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

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

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

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

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

  578. 

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

  580.                 Y1 = av_clip_uint8(Y1);

  581.                 Y2 = av_clip_uint8(Y2);

  582.                 U  = av_clip_uint8(U);

  583.                 V  = av_clip_uint8(V);

  584.             }

  585. 

  586.             Y1 = av_clip_uint8(Y1);

  587.             Y2 = av_clip_uint8(Y2);

  588.             U  = av_clip_uint8(U);

  589.             V  = av_clip_uint8(V);

  590. 

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

  592.         }

  593.     }

  594. }

  595. 

  596. #undef output_pixels

  597. 

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

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

  600. 

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

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

  603. #define output_pixel(pos, val) \

  604.     if (isBE(target)) { \

  605.         AV_WB16(pos, val); \

  606.     } else { \

  607.         AV_WL16(pos, val); \

  608.     }

  609. 

  610. static av_always_inline void

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

  612.                        const int32_t **lumSrc, int lumFilterSize,

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

  614.                        const int32_t **chrVSrc, int chrFilterSize,

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

  616.                        int y, enum PixelFormat target)

  617. {

  618.     int i;

  619. 

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

  621.         int j;

  622.         int Y1 = -0x40000000;

  623.         int Y2 = -0x40000000;

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

  625.         int V  = -128 << 23;

  626.         int R, G, B;

  627. 

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

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

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

  631.         }

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

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

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

  635.         }

  636. 

  637.         // 8bit: 12+15=27; 16-bit: 12+19=31

  638.         Y1 >>= 14; // 10

  639.         Y1 += 0x10000;

  640.         Y2 >>= 14;

  641.         Y2 += 0x10000;

  642.         U  >>= 14;

  643.         V  >>= 14;

  644. 

  645.         // 8bit: 27 -> 17bit, 16bit: 31 - 14 = 17bit

  646.         Y1 -= c->yuv2rgb_y_offset;

  647.         Y2 -= c->yuv2rgb_y_offset;

  648.         Y1 *= c->yuv2rgb_y_coeff;

  649.         Y2 *= c->yuv2rgb_y_coeff;

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

  651.         Y2 += 1 << 13;

  652.         // 8bit: 17 + 13bit = 30bit, 16bit: 17 + 13bit = 30bit

  653. 

  654.         R = V * c->yuv2rgb_v2r_coeff;

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

  656.         B =                            U * c->yuv2rgb_u2b_coeff;

  657. 

  658.         // 8bit: 30 - 22 = 8bit, 16bit: 30bit - 14 = 16bit

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

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

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

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

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

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

  665.         dest += 6;

  666.     }

  667. }

  668. 

  669. static av_always_inline void

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

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

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

  673.                        int yalpha, int uvalpha, int y,

  674.                        enum PixelFormat target)

  675. {

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

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

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

  679.     int  yalpha1 = 4096 - yalpha;

  680.     int uvalpha1 = 4096 - uvalpha;

  681.     int i;

  682. 

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

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

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

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

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

  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. }

  710. 

  711. static av_always_inline void

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

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

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

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

  716. {

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

  718.     int i;

  719. 

  720.     if (uvalpha < 2048) {

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

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

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

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

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

  726.             int R, G, B;

  727. 

  728.             Y1 -= c->yuv2rgb_y_offset;

  729.             Y2 -= c->yuv2rgb_y_offset;

  730.             Y1 *= c->yuv2rgb_y_coeff;

  731.             Y2 *= c->yuv2rgb_y_coeff;

  732.             Y1 += 1 << 13;

  733.             Y2 += 1 << 13;

  734. 

  735.             R = V * c->yuv2rgb_v2r_coeff;

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

  737.             B =                            U * c->yuv2rgb_u2b_coeff;

  738. 

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

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

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

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

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

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

  745.             dest += 6;

  746.         }

  747.     } else {

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

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

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

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

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

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

  754.             int R, G, B;

  755. 

  756.             Y1 -= c->yuv2rgb_y_offset;

  757.             Y2 -= c->yuv2rgb_y_offset;

  758.             Y1 *= c->yuv2rgb_y_coeff;

  759.             Y2 *= c->yuv2rgb_y_coeff;

  760.             Y1 += 1 << 13;

  761.             Y2 += 1 << 13;

  762. 

  763.             R = V * c->yuv2rgb_v2r_coeff;

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

  765.             B =                            U * c->yuv2rgb_u2b_coeff;

  766. 

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

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

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

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

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

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

  773.             dest += 6;

  774.         }

  775.     }

  776. }

  777. 

  778. #undef output_pixel

  779. #undef r_b

  780. #undef b_r

  781. 

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

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

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

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

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

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

  788.                         int y) \

  789. { \

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

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

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

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

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

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

  796.                           chrFilter, chrUSrc, chrVSrc, chrFilterSize, \

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

  798. } \

  799.  \

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

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

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

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

  804. { \

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

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

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

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

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

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

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

  812. } \

  813.  \

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

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

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

  817.                         int uvalpha, int y) \

  818. { \

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

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

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

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

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

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

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

  826. }

  827. 

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

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

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

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

  832. 

  833. /*

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

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

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

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

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

  839.  * correct RGB values into the destination buffer.

  840.  */

  841. static av_always_inline void

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

  843.               unsigned A1, unsigned A2,

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

  845.               enum PixelFormat target, int hasAlpha)

  846. {

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

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

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

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

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

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

  853. 

  854. #if CONFIG_SMALL

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

  856. 

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

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

  859. #else

  860.         if (hasAlpha) {

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

  862. 

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

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

  865.         } else {

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

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

  868.         }

  869. #endif

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

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

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

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

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

  875. 

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

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

  878. 

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

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

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

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

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

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

  885. #undef r_b

  886. #undef b_r

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

  888.                target == PIX_FMT_RGB555 || target == PIX_FMT_BGR555 ||

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

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

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

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

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

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

  895. 

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

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

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

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

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

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

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

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

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

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

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

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

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

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

  910.         } else {

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

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

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

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

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

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

  917.         }

  918. 

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

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

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

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

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

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

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

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

  927. 

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

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

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

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

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

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

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

  935.         } else {

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

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

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

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

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

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

  942.         }

  943. 

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

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

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

  947.         } else {

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

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

  950.         }

  951.     }

  952. }

  953. 

  954. static av_always_inline void

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

  956.                      const int16_t **lumSrc, int lumFilterSize,

  957.                      const int16_t *chrFilter, const int16_t **chrUSrc,

  958.                      const int16_t **chrVSrc, int chrFilterSize,

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

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

  961. {

  962.     int i;

  963. 

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

  965.         int j, A1, A2;

  966.         int Y1 = 1 << 18;

  967.         int Y2 = 1 << 18;

  968.         int U  = 1 << 18;

  969.         int V  = 1 << 18;

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

  971. 

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

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

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

  975.         }

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

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

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

  979.         }

  980.         Y1 >>= 19;

  981.         Y2 >>= 19;

  982.         U  >>= 19;

  983.         V  >>= 19;

  984.         if (hasAlpha) {

  985.             A1 = 1 << 18;

  986.             A2 = 1 << 18;

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

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

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

  990.             }

  991.             A1 >>= 19;

  992.             A2 >>= 19;

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

  994.                 A1 = av_clip_uint8(A1);

  995.                 A2 = av_clip_uint8(A2);

  996.             }

  997.         }

  998. 

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

  1000.         g = (c->table_gU[U + YUVRGB_TABLE_HEADROOM] + c->table_gV[V + YUVRGB_TABLE_HEADROOM]);

  1001.         b =  c->table_bU[U + YUVRGB_TABLE_HEADROOM];

  1002. 

  1003.         yuv2rgb_write(dest, i, Y1, Y2, hasAlpha ? A1 : 0, hasAlpha ? A2 : 0,

  1004.                       r, g, b, y, target, hasAlpha);

  1005.     }

  1006. }

  1007. 

  1008. static av_always_inline void

  1009. yuv2rgb_2_c_template(SwsContext *c, const int16_t *buf[2],

  1010.                      const int16_t *ubuf[2], const int16_t *vbuf[2],

  1011.                      const int16_t *abuf[2], uint8_t *dest, int dstW,

  1012.                      int yalpha, int uvalpha, int y,

  1013.                      enum PixelFormat target, int hasAlpha)

  1014. {

  1015.     const int16_t *buf0  = buf[0],  *buf1  = buf[1],

  1016.                   *ubuf0 = ubuf[0], *ubuf1 = ubuf[1],

  1017.                   *vbuf0 = vbuf[0], *vbuf1 = vbuf[1],

  1018.                   *abuf0 = hasAlpha ? abuf[0] : NULL,

  1019.                   *abuf1 = hasAlpha ? abuf[1] : NULL;

  1020.     int  yalpha1 = 4096 - yalpha;

  1021.     int uvalpha1 = 4096 - uvalpha;

  1022.     int i;

  1023. 

  1024.     for (i = 0; i < ((dstW + 1) >> 1); i++) {

  1025.         int Y1 = (buf0[i * 2]     * yalpha1  + buf1[i * 2]     * yalpha)  >> 19;

  1026.         int Y2 = (buf0[i * 2 + 1] * yalpha1  + buf1[i * 2 + 1] * yalpha)  >> 19;

  1027.         int U  = (ubuf0[i]        * uvalpha1 + ubuf1[i]        * uvalpha) >> 19;

  1028.         int V  = (vbuf0[i]        * uvalpha1 + vbuf1[i]        * uvalpha) >> 19;

  1029.         int A1, A2;

  1030.         const void *r =  c->table_rV[V + YUVRGB_TABLE_HEADROOM],

  1031.                    *g = (c->table_gU[U + YUVRGB_TABLE_HEADROOM] + c->table_gV[V + YUVRGB_TABLE_HEADROOM]),

  1032.                    *b =  c->table_bU[U + YUVRGB_TABLE_HEADROOM];

  1033. 

  1034.         if (hasAlpha) {

  1035.             A1 = (abuf0[i * 2    ] * yalpha1 + abuf1[i * 2    ] * yalpha) >> 19;

  1036.             A2 = (abuf0[i * 2 + 1] * yalpha1 + abuf1[i * 2 + 1] * yalpha) >> 19;

  1037.             A1 = av_clip_uint8(A1);

  1038.             A2 = av_clip_uint8(A2);

  1039.         }

  1040. 

  1041.         yuv2rgb_write(dest, i, Y1, Y2, hasAlpha ? A1 : 0, hasAlpha ? A2 : 0,

  1042.                       r, g, b, y, target, hasAlpha);

  1043.     }

  1044. }

  1045. 

  1046. static av_always_inline void

  1047. yuv2rgb_1_c_template(SwsContext *c, const int16_t *buf0,

  1048.                      const int16_t *ubuf[2], const int16_t *vbuf[2],

  1049.                      const int16_t *abuf0, uint8_t *dest, int dstW,

  1050.                      int uvalpha, int y, enum PixelFormat target,

  1051.                      int hasAlpha)

  1052. {

  1053.     const int16_t *ubuf0 = ubuf[0], *vbuf0 = vbuf[0];

  1054.     int i;

  1055. 

  1056.     if (uvalpha < 2048) {

  1057.         for (i = 0; i < ((dstW + 1) >> 1); i++) {

  1058.             int Y1 = (buf0[i * 2    ] + 64) >> 7;

  1059.             int Y2 = (buf0[i * 2 + 1] + 64) >> 7;

  1060.             int U  = (ubuf0[i]        + 64) >> 7;

  1061.             int V  = (vbuf0[i]        + 64) >> 7;

  1062.             int A1, A2;

  1063.             const void *r =  c->table_rV[V + YUVRGB_TABLE_HEADROOM],

  1064.                        *g = (c->table_gU[U + YUVRGB_TABLE_HEADROOM] + c->table_gV[V + YUVRGB_TABLE_HEADROOM]),

  1065.                        *b =  c->table_bU[U + YUVRGB_TABLE_HEADROOM];

  1066. 

  1067.             if (hasAlpha) {

  1068.                 A1 = abuf0[i * 2    ] * 255 + 16384 >> 15;

  1069.                 A2 = abuf0[i * 2 + 1] * 255 + 16384 >> 15;

  1070.                 A1 = av_clip_uint8(A1);

  1071.                 A2 = av_clip_uint8(A2);

  1072.             }

  1073. 

  1074.             yuv2rgb_write(dest, i, Y1, Y2, hasAlpha ? A1 : 0, hasAlpha ? A2 : 0,

  1075.                           r, g, b, y, target, hasAlpha);

  1076.         }

  1077.     } else {

  1078.         const int16_t *ubuf1 = ubuf[1], *vbuf1 = vbuf[1];

  1079.         for (i = 0; i < ((dstW + 1) >> 1); i++) {

  1080.             int Y1 = (buf0[i * 2    ]     +  64) >> 7;

  1081.             int Y2 = (buf0[i * 2 + 1]     +  64) >> 7;

  1082.             int U  = (ubuf0[i] + ubuf1[i] + 128) >> 8;

  1083.             int V  = (vbuf0[i] + vbuf1[i] + 128) >> 8;

  1084.             int A1, A2;

  1085.             const void *r =  c->table_rV[V + YUVRGB_TABLE_HEADROOM],

  1086.                        *g = (c->table_gU[U + YUVRGB_TABLE_HEADROOM] + c->table_gV[V + YUVRGB_TABLE_HEADROOM]),

  1087.                        *b =  c->table_bU[U + YUVRGB_TABLE_HEADROOM];

  1088. 

  1089.             if (hasAlpha) {

  1090.                 A1 = (abuf0[i * 2    ] + 64) >> 7;

  1091.                 A2 = (abuf0[i * 2 + 1] + 64) >> 7;

  1092.                 A1 = av_clip_uint8(A1);

  1093.                 A2 = av_clip_uint8(A2);

  1094.             }

  1095. 

  1096.             yuv2rgb_write(dest, i, Y1, Y2, hasAlpha ? A1 : 0, hasAlpha ? A2 : 0,

  1097.                           r, g, b, y, target, hasAlpha);

  1098.         }

  1099.     }

  1100. }

  1101. 

  1102. #define YUV2RGBWRAPPERX(name, base, ext, fmt, hasAlpha) \

  1103. static void name ## ext ## _X_c(SwsContext *c, const int16_t *lumFilter, \

  1104.                                 const int16_t **lumSrc, int lumFilterSize, \

  1105.                                 const int16_t *chrFilter, const int16_t **chrUSrc, \

  1106.                                 const int16_t **chrVSrc, int chrFilterSize, \

  1107.                                 const int16_t **alpSrc, uint8_t *dest, int dstW, \

  1108.                                 int y) \

  1109. { \

  1110.     name ## base ## _X_c_template(c, lumFilter, lumSrc, lumFilterSize, \

  1111.                                   chrFilter, chrUSrc, chrVSrc, chrFilterSize, \

  1112.                                   alpSrc, dest, dstW, y, fmt, hasAlpha); \

  1113. }

  1114. #define YUV2RGBWRAPPER(name, base, ext, fmt, hasAlpha) \

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

  1116. static void name ## ext ## _2_c(SwsContext *c, const int16_t *buf[2], \

  1117.                                 const int16_t *ubuf[2], const int16_t *vbuf[2], \

  1118.                                 const int16_t *abuf[2], uint8_t *dest, int dstW, \

  1119.                                 int yalpha, int uvalpha, int y) \

  1120. { \

  1121.     name ## base ## _2_c_template(c, buf, ubuf, vbuf, abuf, \

  1122.                                   dest, dstW, yalpha, uvalpha, y, fmt, hasAlpha); \

  1123. } \

  1124.  \

  1125. static void name ## ext ## _1_c(SwsContext *c, const int16_t *buf0, \

  1126.                                 const int16_t *ubuf[2], const int16_t *vbuf[2], \

  1127.                                 const int16_t *abuf0, uint8_t *dest, int dstW, \

  1128.                                 int uvalpha, int y) \

  1129. { \

  1130.     name ## base ## _1_c_template(c, buf0, ubuf, vbuf, abuf0, dest, \

  1131.                                   dstW, uvalpha, y, fmt, hasAlpha); \

  1132. }

  1133. 

  1134. #if CONFIG_SMALL

  1135. YUV2RGBWRAPPER(yuv2rgb,,  32_1,  PIX_FMT_RGB32_1,   CONFIG_SWSCALE_ALPHA && c->alpPixBuf)

  1136. YUV2RGBWRAPPER(yuv2rgb,,  32,    PIX_FMT_RGB32,     CONFIG_SWSCALE_ALPHA && c->alpPixBuf)

  1137. #else

  1138. #if CONFIG_SWSCALE_ALPHA

  1139. YUV2RGBWRAPPER(yuv2rgb,, a32_1,  PIX_FMT_RGB32_1,   1)

  1140. YUV2RGBWRAPPER(yuv2rgb,, a32,    PIX_FMT_RGB32,     1)

  1141. #endif

  1142. YUV2RGBWRAPPER(yuv2rgb,, x32_1,  PIX_FMT_RGB32_1,   0)

  1143. YUV2RGBWRAPPER(yuv2rgb,, x32,    PIX_FMT_RGB32,     0)

  1144. #endif

  1145. YUV2RGBWRAPPER(yuv2, rgb, rgb24, PIX_FMT_RGB24,   0)

  1146. YUV2RGBWRAPPER(yuv2, rgb, bgr24, PIX_FMT_BGR24,   0)

  1147. YUV2RGBWRAPPER(yuv2rgb,,  16,    PIX_FMT_RGB565,    0)

  1148. YUV2RGBWRAPPER(yuv2rgb,,  15,    PIX_FMT_RGB555,    0)

  1149. YUV2RGBWRAPPER(yuv2rgb,,  12,    PIX_FMT_RGB444,    0)

  1150. YUV2RGBWRAPPER(yuv2rgb,,   8,    PIX_FMT_RGB8,      0)

  1151. YUV2RGBWRAPPER(yuv2rgb,,   4,    PIX_FMT_RGB4,      0)

  1152. YUV2RGBWRAPPER(yuv2rgb,,   4b,   PIX_FMT_RGB4_BYTE, 0)

  1153. 

  1154. static av_always_inline void

  1155. yuv2rgb_full_X_c_template(SwsContext *c, const int16_t *lumFilter,

  1156.                           const int16_t **lumSrc, int lumFilterSize,

  1157.                           const int16_t *chrFilter, const int16_t **chrUSrc,

  1158.                           const int16_t **chrVSrc, int chrFilterSize,

  1159.                           const int16_t **alpSrc, uint8_t *dest,

  1160.                           int dstW, int y, enum PixelFormat target, int hasAlpha)

  1161. {

  1162.     int i;

  1163.     int step = (target == PIX_FMT_RGB24 || target == PIX_FMT_BGR24) ? 3 : 4;

  1164. 

  1165.     for (i = 0; i < dstW; i++) {

  1166.         int j;

  1167.         int Y = 1<<9;

  1168.         int U = (1<<9)-(128 << 19);

  1169.         int V = (1<<9)-(128 << 19);

  1170.         int R, G, B, A;

  1171. 

  1172.         for (j = 0; j < lumFilterSize; j++) {

  1173.             Y += lumSrc[j][i] * lumFilter[j];

  1174.         }

  1175.         for (j = 0; j < chrFilterSize; j++) {

  1176.             U += chrUSrc[j][i] * chrFilter[j];

  1177.             V += chrVSrc[j][i] * chrFilter[j];

  1178.         }

  1179.         Y >>= 10;

  1180.         U >>= 10;

  1181.         V >>= 10;

  1182.         if (hasAlpha) {

  1183.             A = 1 << 18;

  1184.             for (j = 0; j < lumFilterSize; j++) {

  1185.                 A += alpSrc[j][i] * lumFilter[j];

  1186.             }

  1187.             A >>= 19;

  1188.             if (A & 0x100)

  1189.                 A = av_clip_uint8(A);

  1190.         }

  1191.         Y -= c->yuv2rgb_y_offset;

  1192.         Y *= c->yuv2rgb_y_coeff;

  1193.         Y += 1 << 21;

  1194.         R = Y + V*c->yuv2rgb_v2r_coeff;

  1195.         G = Y + V*c->yuv2rgb_v2g_coeff + U*c->yuv2rgb_u2g_coeff;

  1196.         B = Y +                          U*c->yuv2rgb_u2b_coeff;

  1197.         if ((R | G | B) & 0xC0000000) {

  1198.             R = av_clip_uintp2(R, 30);

  1199.             G = av_clip_uintp2(G, 30);

  1200.             B = av_clip_uintp2(B, 30);

  1201.         }

  1202. 

  1203.         switch(target) {

  1204.         case PIX_FMT_ARGB:

  1205.             dest[0] = hasAlpha ? A : 255;

  1206.             dest[1] = R >> 22;

  1207.             dest[2] = G >> 22;

  1208.             dest[3] = B >> 22;

  1209.             break;

  1210.         case PIX_FMT_RGB24:

  1211.             dest[0] = R >> 22;

  1212.             dest[1] = G >> 22;

  1213.             dest[2] = B >> 22;

  1214.             break;

  1215.         case PIX_FMT_RGBA:

  1216.             dest[0] = R >> 22;

  1217.             dest[1] = G >> 22;

  1218.             dest[2] = B >> 22;

  1219.             dest[3] = hasAlpha ? A : 255;

  1220.             break;

  1221.         case PIX_FMT_ABGR:

  1222.             dest[0] = hasAlpha ? A : 255;

  1223.             dest[1] = B >> 22;

  1224.             dest[2] = G >> 22;

  1225.             dest[3] = R >> 22;

  1226.             break;

  1227.         case PIX_FMT_BGR24:

  1228.             dest[0] = B >> 22;

  1229.             dest[1] = G >> 22;

  1230.             dest[2] = R >> 22;

  1231.             break;

  1232.         case PIX_FMT_BGRA:

  1233.             dest[0] = B >> 22;

  1234.             dest[1] = G >> 22;

  1235.             dest[2] = R >> 22;

  1236.             dest[3] = hasAlpha ? A : 255;

  1237.             break;

  1238.         }

  1239.         dest += step;

  1240.     }

  1241. }

  1242. 

  1243. #if CONFIG_SMALL

  1244. YUV2RGBWRAPPERX(yuv2, rgb_full, bgra32_full, PIX_FMT_BGRA,  CONFIG_SWSCALE_ALPHA && c->alpPixBuf)

  1245. YUV2RGBWRAPPERX(yuv2, rgb_full, abgr32_full, PIX_FMT_ABGR,  CONFIG_SWSCALE_ALPHA && c->alpPixBuf)

  1246. YUV2RGBWRAPPERX(yuv2, rgb_full, rgba32_full, PIX_FMT_RGBA,  CONFIG_SWSCALE_ALPHA && c->alpPixBuf)

  1247. YUV2RGBWRAPPERX(yuv2, rgb_full, argb32_full, PIX_FMT_ARGB,  CONFIG_SWSCALE_ALPHA && c->alpPixBuf)

  1248. #else

  1249. #if CONFIG_SWSCALE_ALPHA

  1250. YUV2RGBWRAPPERX(yuv2, rgb_full, bgra32_full, PIX_FMT_BGRA,  1)

  1251. YUV2RGBWRAPPERX(yuv2, rgb_full, abgr32_full, PIX_FMT_ABGR,  1)

  1252. YUV2RGBWRAPPERX(yuv2, rgb_full, rgba32_full, PIX_FMT_RGBA,  1)

  1253. YUV2RGBWRAPPERX(yuv2, rgb_full, argb32_full, PIX_FMT_ARGB,  1)

  1254. #endif

  1255. YUV2RGBWRAPPERX(yuv2, rgb_full, bgrx32_full, PIX_FMT_BGRA,  0)

  1256. YUV2RGBWRAPPERX(yuv2, rgb_full, xbgr32_full, PIX_FMT_ABGR,  0)

  1257. YUV2RGBWRAPPERX(yuv2, rgb_full, rgbx32_full, PIX_FMT_RGBA,  0)

  1258. YUV2RGBWRAPPERX(yuv2, rgb_full, xrgb32_full, PIX_FMT_ARGB,  0)

  1259. #endif

  1260. YUV2RGBWRAPPERX(yuv2, rgb_full, bgr24_full,  PIX_FMT_BGR24, 0)

  1261. YUV2RGBWRAPPERX(yuv2, rgb_full, rgb24_full,  PIX_FMT_RGB24, 0)

  1262. 

  1263. void ff_sws_init_output_funcs(SwsContext *c,

  1264.                               yuv2planar1_fn *yuv2plane1,

  1265.                               yuv2planarX_fn *yuv2planeX,

  1266.                               yuv2interleavedX_fn *yuv2nv12cX,

  1267.                               yuv2packed1_fn *yuv2packed1,

  1268.                               yuv2packed2_fn *yuv2packed2,

  1269.                               yuv2packedX_fn *yuv2packedX)

  1270. {

  1271.     enum PixelFormat dstFormat = c->dstFormat;

  1272. 

  1273.     if (is16BPS(dstFormat)) {

  1274.         *yuv2planeX = isBE(dstFormat) ? yuv2planeX_16BE_c  : yuv2planeX_16LE_c;

  1275.         *yuv2plane1 = isBE(dstFormat) ? yuv2plane1_16BE_c  : yuv2plane1_16LE_c;

  1276.     } else if (is9_OR_10BPS(dstFormat)) {

  1277.         if (av_pix_fmt_descriptors[dstFormat].comp[0].depth_minus1 == 8) {

  1278.             *yuv2planeX = isBE(dstFormat) ? yuv2planeX_9BE_c  : yuv2planeX_9LE_c;

  1279.             *yuv2plane1 = isBE(dstFormat) ? yuv2plane1_9BE_c  : yuv2plane1_9LE_c;

  1280.         } else {

  1281.             *yuv2planeX = isBE(dstFormat) ? yuv2planeX_10BE_c  : yuv2planeX_10LE_c;

  1282.             *yuv2plane1 = isBE(dstFormat) ? yuv2plane1_10BE_c  : yuv2plane1_10LE_c;

  1283.         }

  1284.     } else {

  1285.         *yuv2plane1 = yuv2plane1_8_c;

  1286.         *yuv2planeX = yuv2planeX_8_c;

  1287.         if (dstFormat == PIX_FMT_NV12 || dstFormat == PIX_FMT_NV21)

  1288.             *yuv2nv12cX = yuv2nv12cX_c;

  1289.     }

  1290. 

  1291.     if(c->flags & SWS_FULL_CHR_H_INT) {

  1292.         switch (dstFormat) {

  1293.             case PIX_FMT_RGBA:

  1294. #if CONFIG_SMALL

  1295.                 *yuv2packedX = yuv2rgba32_full_X_c;

  1296. #else

  1297. #if CONFIG_SWSCALE_ALPHA

  1298.                 if (c->alpPixBuf) {

  1299.                     *yuv2packedX = yuv2rgba32_full_X_c;

  1300.                 } else

  1301. #endif /* CONFIG_SWSCALE_ALPHA */

  1302.                 {

  1303.                     *yuv2packedX = yuv2rgbx32_full_X_c;

  1304.                 }

  1305. #endif /* !CONFIG_SMALL */

  1306.                 break;

  1307.             case PIX_FMT_ARGB:

  1308. #if CONFIG_SMALL

  1309.                 *yuv2packedX = yuv2argb32_full_X_c;

  1310. #else

  1311. #if CONFIG_SWSCALE_ALPHA

  1312.                 if (c->alpPixBuf) {

  1313.                     *yuv2packedX = yuv2argb32_full_X_c;

  1314.                 } else

  1315. #endif /* CONFIG_SWSCALE_ALPHA */

  1316.                 {

  1317.                     *yuv2packedX = yuv2xrgb32_full_X_c;

  1318.                 }

  1319. #endif /* !CONFIG_SMALL */

  1320.                 break;

  1321.             case PIX_FMT_BGRA:

  1322. #if CONFIG_SMALL

  1323.                 *yuv2packedX = yuv2bgra32_full_X_c;

  1324. #else

  1325. #if CONFIG_SWSCALE_ALPHA

  1326.                 if (c->alpPixBuf) {

  1327.                     *yuv2packedX = yuv2bgra32_full_X_c;

  1328.                 } else

  1329. #endif /* CONFIG_SWSCALE_ALPHA */

  1330.                 {

  1331.                     *yuv2packedX = yuv2bgrx32_full_X_c;

  1332.                 }

  1333. #endif /* !CONFIG_SMALL */

  1334.                 break;

  1335.             case PIX_FMT_ABGR:

  1336. #if CONFIG_SMALL

  1337.                 *yuv2packedX = yuv2abgr32_full_X_c;

  1338. #else

  1339. #if CONFIG_SWSCALE_ALPHA

  1340.                 if (c->alpPixBuf) {

  1341.                     *yuv2packedX = yuv2abgr32_full_X_c;

  1342.                 } else

  1343. #endif /* CONFIG_SWSCALE_ALPHA */

  1344.                 {

  1345.                     *yuv2packedX = yuv2xbgr32_full_X_c;

  1346.                 }

  1347. #endif /* !CONFIG_SMALL */

  1348.                 break;

  1349.             case PIX_FMT_RGB24:

  1350.             *yuv2packedX = yuv2rgb24_full_X_c;

  1351.             break;

  1352.         case PIX_FMT_BGR24:

  1353.             *yuv2packedX = yuv2bgr24_full_X_c;

  1354.             break;

  1355.         }

  1356.         if(!*yuv2packedX)

  1357.             goto YUV_PACKED;

  1358.     } else {

  1359.         YUV_PACKED:

  1360.         switch (dstFormat) {

  1361.         case PIX_FMT_RGB48LE:

  1362.             *yuv2packed1 = yuv2rgb48le_1_c;

  1363.             *yuv2packed2 = yuv2rgb48le_2_c;

  1364.             *yuv2packedX = yuv2rgb48le_X_c;

  1365.             break;

  1366.         case PIX_FMT_RGB48BE:

  1367.             *yuv2packed1 = yuv2rgb48be_1_c;

  1368.             *yuv2packed2 = yuv2rgb48be_2_c;

  1369.             *yuv2packedX = yuv2rgb48be_X_c;

  1370.             break;

  1371.         case PIX_FMT_BGR48LE:

  1372.             *yuv2packed1 = yuv2bgr48le_1_c;

  1373.             *yuv2packed2 = yuv2bgr48le_2_c;

  1374.             *yuv2packedX = yuv2bgr48le_X_c;

  1375.             break;

  1376.         case PIX_FMT_BGR48BE:

  1377.             *yuv2packed1 = yuv2bgr48be_1_c;

  1378.             *yuv2packed2 = yuv2bgr48be_2_c;

  1379.             *yuv2packedX = yuv2bgr48be_X_c;

  1380.             break;

  1381.         case PIX_FMT_RGB32:

  1382.         case PIX_FMT_BGR32:

  1383. #if CONFIG_SMALL

  1384.             *yuv2packed1 = yuv2rgb32_1_c;

  1385.             *yuv2packed2 = yuv2rgb32_2_c;

  1386.             *yuv2packedX = yuv2rgb32_X_c;

  1387. #else

  1388. #if CONFIG_SWSCALE_ALPHA

  1389.                 if (c->alpPixBuf) {

  1390.                     *yuv2packed1 = yuv2rgba32_1_c;

  1391.                     *yuv2packed2 = yuv2rgba32_2_c;

  1392.                     *yuv2packedX = yuv2rgba32_X_c;

  1393.                 } else

  1394. #endif /* CONFIG_SWSCALE_ALPHA */

  1395.                 {

  1396.                     *yuv2packed1 = yuv2rgbx32_1_c;

  1397.                     *yuv2packed2 = yuv2rgbx32_2_c;

  1398.                     *yuv2packedX = yuv2rgbx32_X_c;

  1399.                 }

  1400. #endif /* !CONFIG_SMALL */

  1401.             break;

  1402.         case PIX_FMT_RGB32_1:

  1403.         case PIX_FMT_BGR32_1:

  1404. #if CONFIG_SMALL

  1405.                 *yuv2packed1 = yuv2rgb32_1_1_c;

  1406.                 *yuv2packed2 = yuv2rgb32_1_2_c;

  1407.                 *yuv2packedX = yuv2rgb32_1_X_c;

  1408. #else

  1409. #if CONFIG_SWSCALE_ALPHA

  1410.                 if (c->alpPixBuf) {

  1411.                     *yuv2packed1 = yuv2rgba32_1_1_c;

  1412.                     *yuv2packed2 = yuv2rgba32_1_2_c;

  1413.                     *yuv2packedX = yuv2rgba32_1_X_c;

  1414.                 } else

  1415. #endif /* CONFIG_SWSCALE_ALPHA */

  1416.                 {

  1417.                     *yuv2packed1 = yuv2rgbx32_1_1_c;

  1418.                     *yuv2packed2 = yuv2rgbx32_1_2_c;

  1419.                     *yuv2packedX = yuv2rgbx32_1_X_c;

  1420.                 }

  1421. #endif /* !CONFIG_SMALL */

  1422.                 break;

  1423.         case PIX_FMT_RGB24:

  1424.             *yuv2packed1 = yuv2rgb24_1_c;

  1425.             *yuv2packed2 = yuv2rgb24_2_c;

  1426.             *yuv2packedX = yuv2rgb24_X_c;

  1427.             break;

  1428.         case PIX_FMT_BGR24:

  1429.             *yuv2packed1 = yuv2bgr24_1_c;

  1430.             *yuv2packed2 = yuv2bgr24_2_c;

  1431.             *yuv2packedX = yuv2bgr24_X_c;

  1432.             break;

  1433.         case PIX_FMT_RGB565LE:

  1434.         case PIX_FMT_RGB565BE:

  1435.         case PIX_FMT_BGR565LE:

  1436.         case PIX_FMT_BGR565BE:

  1437.             *yuv2packed1 = yuv2rgb16_1_c;

  1438.             *yuv2packed2 = yuv2rgb16_2_c;

  1439.             *yuv2packedX = yuv2rgb16_X_c;

  1440.             break;

  1441.         case PIX_FMT_RGB555LE:

  1442.         case PIX_FMT_RGB555BE:

  1443.         case PIX_FMT_BGR555LE:

  1444.         case PIX_FMT_BGR555BE:

  1445.             *yuv2packed1 = yuv2rgb15_1_c;

  1446.             *yuv2packed2 = yuv2rgb15_2_c;

  1447.             *yuv2packedX = yuv2rgb15_X_c;

  1448.             break;

  1449.         case PIX_FMT_RGB444LE:

  1450.         case PIX_FMT_RGB444BE:

  1451.         case PIX_FMT_BGR444LE:

  1452.         case PIX_FMT_BGR444BE:

  1453.             *yuv2packed1 = yuv2rgb12_1_c;

  1454.             *yuv2packed2 = yuv2rgb12_2_c;

  1455.             *yuv2packedX = yuv2rgb12_X_c;

  1456.             break;

  1457.         case PIX_FMT_RGB8:

  1458.         case PIX_FMT_BGR8:

  1459.             *yuv2packed1 = yuv2rgb8_1_c;

  1460.             *yuv2packed2 = yuv2rgb8_2_c;

  1461.             *yuv2packedX = yuv2rgb8_X_c;

  1462.             break;

  1463.         case PIX_FMT_RGB4:

  1464.         case PIX_FMT_BGR4:

  1465.             *yuv2packed1 = yuv2rgb4_1_c;

  1466.             *yuv2packed2 = yuv2rgb4_2_c;

  1467.             *yuv2packedX = yuv2rgb4_X_c;

  1468.             break;

  1469.         case PIX_FMT_RGB4_BYTE:

  1470.         case PIX_FMT_BGR4_BYTE:

  1471.             *yuv2packed1 = yuv2rgb4b_1_c;

  1472.             *yuv2packed2 = yuv2rgb4b_2_c;

  1473.             *yuv2packedX = yuv2rgb4b_X_c;

  1474.             break;

  1475.         }

  1476.     }

  1477.     switch (dstFormat) {

  1478.     case PIX_FMT_MONOWHITE:

  1479.         *yuv2packed1 = yuv2monowhite_1_c;

  1480.         *yuv2packed2 = yuv2monowhite_2_c;

  1481.         *yuv2packedX = yuv2monowhite_X_c;

  1482.         break;

  1483.     case PIX_FMT_MONOBLACK:

  1484.         *yuv2packed1 = yuv2monoblack_1_c;

  1485.         *yuv2packed2 = yuv2monoblack_2_c;

  1486.         *yuv2packedX = yuv2monoblack_X_c;

  1487.         break;

  1488.     case PIX_FMT_YUYV422:

  1489.         *yuv2packed1 = yuv2yuyv422_1_c;

  1490.         *yuv2packed2 = yuv2yuyv422_2_c;

  1491.         *yuv2packedX = yuv2yuyv422_X_c;

  1492.         break;

  1493.     case PIX_FMT_UYVY422:

  1494.         *yuv2packed1 = yuv2uyvy422_1_c;

  1495.         *yuv2packed2 = yuv2uyvy422_2_c;

  1496.         *yuv2packedX = yuv2uyvy422_X_c;

  1497.         break;

  1498.     }

  1499. }