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.
32322 Commits
32 Branches
944MB
Tree: c6643fddba
Branches Tags
${ item.name }
Create branch ${ searchTerm }
from 'c6643fddba'
${ noResults }
FFmpeg/libswscale/output.c

1436 lines
52KB
Raw Blame History 

  1. /*

  2.  * Copyright (C) 2001-2003 Michael Niedermayer <michaelni@gmx.at>

  3.  *

  4.  * This file is part of Libav.

  5.  *

  6.  * Libav 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.  * Libav 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 Libav; 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/bswap.h"

  29. #include "libavutil/cpu.h"

  30. #include "libavutil/intreadwrite.h"

  31. #include "libavutil/mathematics.h"

  32. #include "libavutil/pixdesc.h"

  33. #include "config.h"

  34. #include "rgb2rgb.h"

  35. #include "swscale.h"

  36. #include "swscale_internal.h"

  37. 

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

  39. {  1,   3,   1,   3,   1,   3,   1,   3, },

  40. {  2,   0,   2,   0,   2,   0,   2,   0, },

  41. };

  42. 

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

  44. {  6,   2,   6,   2,   6,   2,   6,   2, },

  45. {  0,   4,   0,   4,   0,   4,   0,   4, },

  46. };

  47. 

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

  49. {  8,   4,  11,   7,   8,   4,  11,   7, },

  50. {  2,  14,   1,  13,   2,  14,   1,  13, },

  51. { 10,   6,   9,   5,  10,   6,   9,   5, },

  52. {  0,  12,   3,  15,   0,  12,   3,  15, },

  53. };

  54. 

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

  56. { 17,   9,  23,  15,  16,   8,  22,  14, },

  57. {  5,  29,   3,  27,   4,  28,   2,  26, },

  58. { 21,  13,  19,  11,  20,  12,  18,  10, },

  59. {  0,  24,   6,  30,   1,  25,   7,  31, },

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

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

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

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

  64. };

  65. 

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

  67. {  0,  55,  14,  68,   3,  58,  17,  72, },

  68. { 37,  18,  50,  32,  40,  22,  54,  35, },

  69. {  9,  64,   5,  59,  13,  67,   8,  63, },

  70. { 46,  27,  41,  23,  49,  31,  44,  26, },

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

  72. { 39,  21,  52,  34,  38,  19,  51,  33, },

  73. { 11,  66,   7,  62,  10,  65,   6,  60, },

  74. { 48,  30,  43,  25,  47,  29,  42,  24, },

  75. };

  76. 

  77. #if 1

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

  79. {117,  62, 158, 103, 113,  58, 155, 100, },

  80. { 34, 199,  21, 186,  31, 196,  17, 182, },

  81. {144,  89, 131,  76, 141,  86, 127,  72, },

  82. {  0, 165,  41, 206,  10, 175,  52, 217, },

  83. {110,  55, 151,  96, 120,  65, 162, 107, },

  84. { 28, 193,  14, 179,  38, 203,  24, 189, },

  85. {138,  83, 124,  69, 148,  93, 134,  79, },

  86. {  7, 172,  48, 213,   3, 168,  45, 210, },

  87. };

  88. #elif 1

  89. // tries to correct a gamma of 1.5

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

  91. {  0, 143,  18, 200,   2, 156,  25, 215, },

  92. { 78,  28, 125,  64,  89,  36, 138,  74, },

  93. { 10, 180,   3, 161,  16, 195,   8, 175, },

  94. {109,  51,  93,  38, 121,  60, 105,  47, },

  95. {  1, 152,  23, 210,   0, 147,  20, 205, },

  96. { 85,  33, 134,  71,  81,  30, 130,  67, },

  97. { 14, 190,   6, 171,  12, 185,   5, 166, },

  98. {117,  57, 101,  44, 113,  54,  97,  41, },

  99. };

  100. #elif 1

  101. // tries to correct a gamma of 2.0

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

  103. {  0, 124,   8, 193,   0, 140,  12, 213, },

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

  105. {  3, 168,   1, 145,   6, 187,   3, 162, },

  106. { 86,  31,  70,  21,  99,  39,  82,  28, },

  107. {  0, 134,  11, 206,   0, 129,   9, 200, },

  108. { 62,  17, 114,  48,  58,  16, 109,  45, },

  109. {  5, 181,   2, 157,   4, 175,   1, 151, },

  110. { 95,  36,  78,  26,  90,  34,  74,  24, },

  111. };

  112. #else

  113. // tries to correct a gamma of 2.5

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

  115. {  0, 107,   3, 187,   0, 125,   6, 212, },

  116. { 39,   7,  86,  28,  49,  11, 102,  36, },

  117. {  1, 158,   0, 131,   3, 180,   1, 151, },

  118. { 68,  19,  52,  12,  81,  25,  64,  17, },

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

  120. { 45,   9,  96,  33,  42,   8,  91,  30, },

  121. {  2, 172,   1, 144,   2, 165,   0, 137, },

  122. { 77,  23,  60,  15,  72,  21,  56,  14, },

  123. };

  124. #endif

  125. 

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

  127.     if (big_endian) { \

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

  129.     } else { \

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

  131.     }

  132. 

  133. static av_always_inline void

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

  135.                          int big_endian, int output_bits)

  136. {

  137.     int i;

  138.     int shift = 19 - output_bits;

  139. 

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

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

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

  143.     }

  144. }

  145. 

  146. static av_always_inline void

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

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

  149.                          int big_endian, int output_bits)

  150. {

  151.     int i;

  152.     int shift = 15 + 16 - output_bits;

  153. 

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

  155.         int val = 1 << (30-output_bits);

  156.         int j;

  157. 

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

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

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

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

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

  163.         val -= 0x40000000;

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

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

  166. 

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

  168.     }

  169. }

  170. 

  171. #undef output_pixel

  172. 

  173. #define output_pixel(pos, val) \

  174.     if (big_endian) { \

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

  176.     } else { \

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

  178.     }

  179. 

  180. static av_always_inline void

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

  182.                          int big_endian, int output_bits)

  183. {

  184.     int i;

  185.     int shift = 15 - output_bits;

  186. 

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

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

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

  190.     }

  191. }

  192. 

  193. static av_always_inline void

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

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

  196.                          int big_endian, int output_bits)

  197. {

  198.     int i;

  199.     int shift = 11 + 16 - output_bits;

  200. 

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

  202.         int val = 1 << (26-output_bits);

  203.         int j;

  204. 

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

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

  207. 

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

  209.     }

  210. }

  211. 

  212. #undef output_pixel

  213. 

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

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

  216.                               uint8_t *dest, int dstW, \

  217.                               const uint8_t *dither, int offset)\

  218. { \

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

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

  221. }\

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

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

  224.                               const uint8_t *dither, int offset)\

  225. { \

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

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

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

  229. }

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

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

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

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

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

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

  236. 

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

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

  239.                            const uint8_t *dither, int offset)

  240. {

  241.     int i;

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

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

  244.         int j;

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

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

  247. 

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

  249.     }

  250. }

  251. 

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

  253.                            const uint8_t *dither, int offset)

  254. {

  255.     int i;

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

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

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

  259.     }

  260. }

  261. 

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

  263.                         const int16_t **chrUSrc, const int16_t **chrVSrc,

  264.                         uint8_t *dest, int chrDstW)

  265. {

  266.     enum PixelFormat dstFormat = c->dstFormat;

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

  268.     int i;

  269. 

  270.     if (dstFormat == PIX_FMT_NV12)

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

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

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

  274.             int j;

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

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

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

  278.             }

  279. 

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

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

  282.         }

  283.     else

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

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

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

  287.             int j;

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

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

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

  291.             }

  292. 

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

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

  295.         }

  296. }

  297. 

  298. #define output_pixel(pos, acc) \

  299.     if (target == PIX_FMT_MONOBLACK) { \

  300.         pos = acc; \

  301.     } else { \

  302.         pos = ~acc; \

  303.     }

  304. 

  305. static av_always_inline void

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

  307.                       const int16_t **lumSrc, int lumFilterSize,

  308.                       const int16_t *chrFilter, const int16_t **chrUSrc,

  309.                       const int16_t **chrVSrc, int chrFilterSize,

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

  311.                       int y, enum PixelFormat target)

  312. {

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

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

  315.     int i;

  316.     unsigned acc = 0;

  317. 

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

  319.         int j;

  320.         int Y1 = 1 << 18;

  321.         int Y2 = 1 << 18;

  322. 

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

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

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

  326.         }

  327.         Y1 >>= 19;

  328.         Y2 >>= 19;

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

  330.             Y1 = av_clip_uint8(Y1);

  331.             Y2 = av_clip_uint8(Y2);

  332.         }

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

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

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

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

  337.         }

  338.     }

  339. }

  340. 

  341. static av_always_inline void

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

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

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

  345.                       int yalpha, int uvalpha, int y,

  346.                       enum PixelFormat target)

  347. {

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

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

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

  351.     int  yalpha1 = 4095 - yalpha;

  352.     int i;

  353. 

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

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

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

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

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

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

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

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

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

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

  364.     }

  365. }

  366. 

  367. static av_always_inline void

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

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

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

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

  372. {

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

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

  375.     int i;

  376. 

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

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

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

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

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

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

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

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

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

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

  387.     }

  388. }

  389. 

  390. #undef output_pixel

  391. 

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

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

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

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

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

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

  398.                                 int y) \

  399. { \

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

  401.                                   chrFilter, chrUSrc, chrVSrc, chrFilterSize, \

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

  403. } \

  404.  \

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

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

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

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

  409. { \

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

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

  412. } \

  413.  \

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

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

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

  417.                                 int uvalpha, int y) \

  418. { \

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

  420.                                   abuf0, dest, dstW, uvalpha, \

  421.                                   y, fmt); \

  422. }

  423. 

  424. YUV2PACKEDWRAPPER(yuv2mono,, white, PIX_FMT_MONOWHITE)

  425. YUV2PACKEDWRAPPER(yuv2mono,, black, PIX_FMT_MONOBLACK)

  426. 

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

  428.     if (target == PIX_FMT_YUYV422) { \

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

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

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

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

  433.     } else { \

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

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

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

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

  438.     }

  439. 

  440. static av_always_inline void

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

  442.                      const int16_t **lumSrc, int lumFilterSize,

  443.                      const int16_t *chrFilter, const int16_t **chrUSrc,

  444.                      const int16_t **chrVSrc, int chrFilterSize,

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

  446.                      int y, enum PixelFormat target)

  447. {

  448.     int i;

  449. 

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

  451.         int j;

  452.         int Y1 = 1 << 18;

  453.         int Y2 = 1 << 18;

  454.         int U  = 1 << 18;

  455.         int V  = 1 << 18;

  456. 

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

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

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

  460.         }

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

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

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

  464.         }

  465.         Y1 >>= 19;

  466.         Y2 >>= 19;

  467.         U  >>= 19;

  468.         V  >>= 19;

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

  470.             Y1 = av_clip_uint8(Y1);

  471.             Y2 = av_clip_uint8(Y2);

  472.             U  = av_clip_uint8(U);

  473.             V  = av_clip_uint8(V);

  474.         }

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

  476.     }

  477. }

  478. 

  479. static av_always_inline void

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

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

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

  483.                      int yalpha, int uvalpha, int y,

  484.                      enum PixelFormat target)

  485. {

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

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

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

  489.     int  yalpha1 = 4095 - yalpha;

  490.     int uvalpha1 = 4095 - uvalpha;

  491.     int i;

  492. 

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

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

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

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

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

  498. 

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

  500.     }

  501. }

  502. 

  503. static av_always_inline void

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

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

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

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

  508. {

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

  510.     int i;

  511. 

  512.     if (uvalpha < 2048) {

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

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

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

  516.             int U  = ubuf0[i]        >> 7;

  517.             int V  = vbuf0[i]        >> 7;

  518. 

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

  520.         }

  521.     } else {

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

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

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

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

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

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

  528. 

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

  530.         }

  531.     }

  532. }

  533. 

  534. #undef output_pixels

  535. 

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

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

  538. 

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

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

  541. #define output_pixel(pos, val) \

  542.     if (isBE(target)) { \

  543.         AV_WB16(pos, val); \

  544.     } else { \

  545.         AV_WL16(pos, val); \

  546.     }

  547. 

  548. static av_always_inline void

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

  550.                        const int32_t **lumSrc, int lumFilterSize,

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

  552.                        const int32_t **chrVSrc, int chrFilterSize,

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

  554.                        int y, enum PixelFormat target)

  555. {

  556.     int i;

  557. 

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

  559.         int j;

  560.         int Y1 = -0x40000000;

  561.         int Y2 = -0x40000000;

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

  563.         int V  = -128 << 23;

  564.         int R, G, B;

  565. 

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

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

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

  569.         }

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

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

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

  573.         }

  574. 

  575.         // 8bit: 12+15=27; 16-bit: 12+19=31

  576.         Y1 >>= 14; // 10

  577.         Y1 += 0x10000;

  578.         Y2 >>= 14;

  579.         Y2 += 0x10000;

  580.         U  >>= 14;

  581.         V  >>= 14;

  582. 

  583.         // 8bit: 27 -> 17bit, 16bit: 31 - 14 = 17bit

  584.         Y1 -= c->yuv2rgb_y_offset;

  585.         Y2 -= c->yuv2rgb_y_offset;

  586.         Y1 *= c->yuv2rgb_y_coeff;

  587.         Y2 *= c->yuv2rgb_y_coeff;

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

  589.         Y2 += 1 << 13;

  590.         // 8bit: 17 + 13bit = 30bit, 16bit: 17 + 13bit = 30bit

  591. 

  592.         R = V * c->yuv2rgb_v2r_coeff;

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

  594.         B =                            U * c->yuv2rgb_u2b_coeff;

  595. 

  596.         // 8bit: 30 - 22 = 8bit, 16bit: 30bit - 14 = 16bit

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

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

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

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

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

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

  603.         dest += 6;

  604.     }

  605. }

  606. 

  607. static av_always_inline void

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

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

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

  611.                        int yalpha, int uvalpha, int y,

  612.                        enum PixelFormat target)

  613. {

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

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

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

  617.     int  yalpha1 = 4095 - yalpha;

  618.     int uvalpha1 = 4095 - uvalpha;

  619.     int i;

  620. 

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

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

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

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

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

  626.         int R, G, B;

  627. 

  628.         Y1 -= c->yuv2rgb_y_offset;

  629.         Y2 -= c->yuv2rgb_y_offset;

  630.         Y1 *= c->yuv2rgb_y_coeff;

  631.         Y2 *= c->yuv2rgb_y_coeff;

  632.         Y1 += 1 << 13;

  633.         Y2 += 1 << 13;

  634. 

  635.         R = V * c->yuv2rgb_v2r_coeff;

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

  637.         B =                            U * c->yuv2rgb_u2b_coeff;

  638. 

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

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

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

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

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

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

  645.         dest += 6;

  646.     }

  647. }

  648. 

  649. static av_always_inline void

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

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

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

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

  654. {

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

  656.     int i;

  657. 

  658.     if (uvalpha < 2048) {

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

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

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

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

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

  664.             int R, G, B;

  665. 

  666.             Y1 -= c->yuv2rgb_y_offset;

  667.             Y2 -= c->yuv2rgb_y_offset;

  668.             Y1 *= c->yuv2rgb_y_coeff;

  669.             Y2 *= c->yuv2rgb_y_coeff;

  670.             Y1 += 1 << 13;

  671.             Y2 += 1 << 13;

  672. 

  673.             R = V * c->yuv2rgb_v2r_coeff;

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

  675.             B =                            U * c->yuv2rgb_u2b_coeff;

  676. 

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

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

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

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

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

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

  683.             dest += 6;

  684.         }

  685.     } else {

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

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

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

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

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

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

  692.             int R, G, B;

  693. 

  694.             Y1 -= c->yuv2rgb_y_offset;

  695.             Y2 -= c->yuv2rgb_y_offset;

  696.             Y1 *= c->yuv2rgb_y_coeff;

  697.             Y2 *= c->yuv2rgb_y_coeff;

  698.             Y1 += 1 << 13;

  699.             Y2 += 1 << 13;

  700. 

  701.             R = V * c->yuv2rgb_v2r_coeff;

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

  703.             B =                            U * c->yuv2rgb_u2b_coeff;

  704. 

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

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

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

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

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

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

  711.             dest += 6;

  712.         }

  713.     }

  714. }

  715. 

  716. #undef output_pixel

  717. #undef r_b

  718. #undef b_r

  719. 

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

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

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

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

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

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

  726.                         int y) \

  727. { \

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

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

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

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

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

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

  734.                           chrFilter, chrUSrc, chrVSrc, chrFilterSize, \

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

  736. } \

  737.  \

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

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

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

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

  742. { \

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

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

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

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

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

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

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

  750. } \

  751.  \

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

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

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

  755.                         int uvalpha, int y) \

  756. { \

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

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

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

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

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

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

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

  764. }

  765. 

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

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

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

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

  770. 

  771. /*

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

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

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

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

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

  777.  * correct RGB values into the destination buffer.

  778.  */

  779. static av_always_inline void

  780. yuv2rgb_write(uint8_t *_dest, int i, unsigned Y1, unsigned Y2,

  781.               unsigned A1, unsigned A2,

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

  783.               enum PixelFormat target, int hasAlpha)

  784. {

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

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

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

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

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

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

  791. 

  792. #if CONFIG_SMALL

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

  794. 

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

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

  797. #else

  798.         if (hasAlpha) {

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

  800. 

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

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

  803.         } else {

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

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

  806.         }

  807. #endif

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

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

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

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

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

  813. 

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

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

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

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

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

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

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

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

  822. #undef r_b

  823. #undef b_r

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

  825.                target == PIX_FMT_RGB555 || target == PIX_FMT_BGR555 ||

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

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

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

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

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

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

  832. 

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

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

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

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

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

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

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

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

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

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

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

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

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

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

  847.         } else {

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

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

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

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

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

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

  854.         }

  855. 

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

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

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

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

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

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

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

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

  864. 

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

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

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

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

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

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

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

  872.         } else {

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

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

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

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

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

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

  879.         }

  880. 

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

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

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

  884.         } else {

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

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

  887.         }

  888.     }

  889. }

  890. 

  891. static av_always_inline void

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

  893.                      const int16_t **lumSrc, int lumFilterSize,

  894.                      const int16_t *chrFilter, const int16_t **chrUSrc,

  895.                      const int16_t **chrVSrc, int chrFilterSize,

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

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

  898. {

  899.     int i;

  900. 

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

  902.         int j, A1, A2;

  903.         int Y1 = 1 << 18;

  904.         int Y2 = 1 << 18;

  905.         int U  = 1 << 18;

  906.         int V  = 1 << 18;

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

  908. 

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

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

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

  912.         }

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

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

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

  916.         }

  917.         Y1 >>= 19;

  918.         Y2 >>= 19;

  919.         U  >>= 19;

  920.         V  >>= 19;

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

  922.             Y1 = av_clip_uint8(Y1);

  923.             Y2 = av_clip_uint8(Y2);

  924.             U  = av_clip_uint8(U);

  925.             V  = av_clip_uint8(V);

  926.         }

  927.         if (hasAlpha) {

  928.             A1 = 1 << 18;

  929.             A2 = 1 << 18;

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

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

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

  933.             }

  934.             A1 >>= 19;

  935.             A2 >>= 19;

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

  937.                 A1 = av_clip_uint8(A1);

  938.                 A2 = av_clip_uint8(A2);

  939.             }

  940.         }

  941. 

  942.         /* FIXME fix tables so that clipping is not needed and then use _NOCLIP*/

  943.         r =  c->table_rV[V];

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

  945.         b =  c->table_bU[U];

  946. 

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

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

  949.     }

  950. }

  951. 

  952. static av_always_inline void

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

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

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

  956.                      int yalpha, int uvalpha, int y,

  957.                      enum PixelFormat target, int hasAlpha)

  958. {

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

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

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

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

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

  964.     int  yalpha1 = 4095 - yalpha;

  965.     int uvalpha1 = 4095 - uvalpha;

  966.     int i;

  967. 

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

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

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

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

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

  973.         int A1, A2;

  974.         const void *r =  c->table_rV[V],

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

  976.                    *b =  c->table_bU[U];

  977. 

  978.         if (hasAlpha) {

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

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

  981.         }

  982. 

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

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

  985.     }

  986. }

  987. 

  988. static av_always_inline void

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

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

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

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

  993.                      int hasAlpha)

  994. {

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

  996.     int i;

  997. 

  998.     if (uvalpha < 2048) {

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

  1000.             int Y1 = buf0[i * 2]     >> 7;

  1001.             int Y2 = buf0[i * 2 + 1] >> 7;

  1002.             int U  = ubuf0[i]        >> 7;

  1003.             int V  = vbuf0[i]        >> 7;

  1004.             int A1, A2;

  1005.             const void *r =  c->table_rV[V],

  1006.                        *g = (c->table_gU[U] + c->table_gV[V]),

  1007.                        *b =  c->table_bU[U];

  1008. 

  1009.             if (hasAlpha) {

  1010.                 A1 = abuf0[i * 2    ] >> 7;

  1011.                 A2 = abuf0[i * 2 + 1] >> 7;

  1012.             }

  1013. 

  1014.             yuv2rgb_write(dest, i, Y1, Y2, hasAlpha ? A1 : 0, hasAlpha ? A2 : 0,

  1015.                           r, g, b, y, target, hasAlpha);

  1016.         }

  1017.     } else {

  1018.         const int16_t *ubuf1 = ubuf[1], *vbuf1 = vbuf[1];

  1019.         for (i = 0; i < (dstW >> 1); i++) {

  1020.             int Y1 =  buf0[i * 2]          >> 7;

  1021.             int Y2 =  buf0[i * 2 + 1]      >> 7;

  1022.             int U  = (ubuf0[i] + ubuf1[i]) >> 8;

  1023.             int V  = (vbuf0[i] + vbuf1[i]) >> 8;

  1024.             int A1, A2;

  1025.             const void *r =  c->table_rV[V],

  1026.                        *g = (c->table_gU[U] + c->table_gV[V]),

  1027.                        *b =  c->table_bU[U];

  1028. 

  1029.             if (hasAlpha) {

  1030.                 A1 = abuf0[i * 2    ] >> 7;

  1031.                 A2 = abuf0[i * 2 + 1] >> 7;

  1032.             }

  1033. 

  1034.             yuv2rgb_write(dest, i, Y1, Y2, hasAlpha ? A1 : 0, hasAlpha ? A2 : 0,

  1035.                           r, g, b, y, target, hasAlpha);

  1036.         }

  1037.     }

  1038. }

  1039. 

  1040. #define YUV2RGBWRAPPERX(name, base, ext, fmt, hasAlpha) \

  1041. static void name ## ext ## _X_c(SwsContext *c, const int16_t *lumFilter, \

  1042.                                 const int16_t **lumSrc, int lumFilterSize, \

  1043.                                 const int16_t *chrFilter, const int16_t **chrUSrc, \

  1044.                                 const int16_t **chrVSrc, int chrFilterSize, \

  1045.                                 const int16_t **alpSrc, uint8_t *dest, int dstW, \

  1046.                                 int y) \

  1047. { \

  1048.     name ## base ## _X_c_template(c, lumFilter, lumSrc, lumFilterSize, \

  1049.                                   chrFilter, chrUSrc, chrVSrc, chrFilterSize, \

  1050.                                   alpSrc, dest, dstW, y, fmt, hasAlpha); \

  1051. }

  1052. #define YUV2RGBWRAPPER(name, base, ext, fmt, hasAlpha) \

  1053. YUV2RGBWRAPPERX(name, base, ext, fmt, hasAlpha) \

  1054. static void name ## ext ## _2_c(SwsContext *c, const int16_t *buf[2], \

  1055.                                 const int16_t *ubuf[2], const int16_t *vbuf[2], \

  1056.                                 const int16_t *abuf[2], uint8_t *dest, int dstW, \

  1057.                                 int yalpha, int uvalpha, int y) \

  1058. { \

  1059.     name ## base ## _2_c_template(c, buf, ubuf, vbuf, abuf, \

  1060.                                   dest, dstW, yalpha, uvalpha, y, fmt, hasAlpha); \

  1061. } \

  1062.  \

  1063. static void name ## ext ## _1_c(SwsContext *c, const int16_t *buf0, \

  1064.                                 const int16_t *ubuf[2], const int16_t *vbuf[2], \

  1065.                                 const int16_t *abuf0, uint8_t *dest, int dstW, \

  1066.                                 int uvalpha, int y) \

  1067. { \

  1068.     name ## base ## _1_c_template(c, buf0, ubuf, vbuf, abuf0, dest, \

  1069.                                   dstW, uvalpha, y, fmt, hasAlpha); \

  1070. }

  1071. 

  1072. #if CONFIG_SMALL

  1073. YUV2RGBWRAPPER(yuv2rgb,,  32_1,  PIX_FMT_RGB32_1,   CONFIG_SWSCALE_ALPHA && c->alpPixBuf)

  1074. YUV2RGBWRAPPER(yuv2rgb,,  32,    PIX_FMT_RGB32,     CONFIG_SWSCALE_ALPHA && c->alpPixBuf)

  1075. #else

  1076. #if CONFIG_SWSCALE_ALPHA

  1077. YUV2RGBWRAPPER(yuv2rgb,, a32_1,  PIX_FMT_RGB32_1,   1)

  1078. YUV2RGBWRAPPER(yuv2rgb,, a32,    PIX_FMT_RGB32,     1)

  1079. #endif

  1080. YUV2RGBWRAPPER(yuv2rgb,, x32_1,  PIX_FMT_RGB32_1,   0)

  1081. YUV2RGBWRAPPER(yuv2rgb,, x32,    PIX_FMT_RGB32,     0)

  1082. #endif

  1083. YUV2RGBWRAPPER(yuv2, rgb, rgb24, PIX_FMT_RGB24,   0)

  1084. YUV2RGBWRAPPER(yuv2, rgb, bgr24, PIX_FMT_BGR24,   0)

  1085. YUV2RGBWRAPPER(yuv2rgb,,  16,    PIX_FMT_RGB565,    0)

  1086. YUV2RGBWRAPPER(yuv2rgb,,  15,    PIX_FMT_RGB555,    0)

  1087. YUV2RGBWRAPPER(yuv2rgb,,  12,    PIX_FMT_RGB444,    0)

  1088. YUV2RGBWRAPPER(yuv2rgb,,   8,    PIX_FMT_RGB8,      0)

  1089. YUV2RGBWRAPPER(yuv2rgb,,   4,    PIX_FMT_RGB4,      0)

  1090. YUV2RGBWRAPPER(yuv2rgb,,   4b,   PIX_FMT_RGB4_BYTE, 0)

  1091. 

  1092. static av_always_inline void

  1093. yuv2rgb_full_X_c_template(SwsContext *c, const int16_t *lumFilter,

  1094.                           const int16_t **lumSrc, int lumFilterSize,

  1095.                           const int16_t *chrFilter, const int16_t **chrUSrc,

  1096.                           const int16_t **chrVSrc, int chrFilterSize,

  1097.                           const int16_t **alpSrc, uint8_t *dest,

  1098.                           int dstW, int y, enum PixelFormat target, int hasAlpha)

  1099. {

  1100.     int i;

  1101.     int step = (target == PIX_FMT_RGB24 || target == PIX_FMT_BGR24) ? 3 : 4;

  1102. 

  1103.     for (i = 0; i < dstW; i++) {

  1104.         int j;

  1105.         int Y = 0;

  1106.         int U = -128 << 19;

  1107.         int V = -128 << 19;

  1108.         int R, G, B, A;

  1109. 

  1110.         for (j = 0; j < lumFilterSize; j++) {

  1111.             Y += lumSrc[j][i] * lumFilter[j];

  1112.         }

  1113.         for (j = 0; j < chrFilterSize; j++) {

  1114.             U += chrUSrc[j][i] * chrFilter[j];

  1115.             V += chrVSrc[j][i] * chrFilter[j];

  1116.         }

  1117.         Y >>= 10;

  1118.         U >>= 10;

  1119.         V >>= 10;

  1120.         if (hasAlpha) {

  1121.             A = 1 << 21;

  1122.             for (j = 0; j < lumFilterSize; j++) {

  1123.                 A += alpSrc[j][i] * lumFilter[j];

  1124.             }

  1125.             A >>= 19;

  1126.             if (A & 0x100)

  1127.                 A = av_clip_uint8(A);

  1128.         }

  1129.         Y -= c->yuv2rgb_y_offset;

  1130.         Y *= c->yuv2rgb_y_coeff;

  1131.         Y += 1 << 21;

  1132.         R = Y + V*c->yuv2rgb_v2r_coeff;

  1133.         G = Y + V*c->yuv2rgb_v2g_coeff + U*c->yuv2rgb_u2g_coeff;

  1134.         B = Y +                          U*c->yuv2rgb_u2b_coeff;

  1135.         if ((R | G | B) & 0xC0000000) {

  1136.             R = av_clip_uintp2(R, 30);

  1137.             G = av_clip_uintp2(G, 30);

  1138.             B = av_clip_uintp2(B, 30);

  1139.         }

  1140. 

  1141.         switch(target) {

  1142.         case PIX_FMT_ARGB:

  1143.             dest[0] = hasAlpha ? A : 255;

  1144.             dest[1] = R >> 22;

  1145.             dest[2] = G >> 22;

  1146.             dest[3] = B >> 22;

  1147.             break;

  1148.         case PIX_FMT_RGB24:

  1149.             dest[0] = R >> 22;

  1150.             dest[1] = G >> 22;

  1151.             dest[2] = B >> 22;

  1152.             break;

  1153.         case PIX_FMT_RGBA:

  1154.             dest[0] = R >> 22;

  1155.             dest[1] = G >> 22;

  1156.             dest[2] = B >> 22;

  1157.             dest[3] = hasAlpha ? A : 255;

  1158.             break;

  1159.         case PIX_FMT_ABGR:

  1160.             dest[0] = hasAlpha ? A : 255;

  1161.             dest[1] = B >> 22;

  1162.             dest[2] = G >> 22;

  1163.             dest[3] = R >> 22;

  1164.             dest += 4;

  1165.             break;

  1166.         case PIX_FMT_BGR24:

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

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

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

  1170.             break;

  1171.         case PIX_FMT_BGRA:

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

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

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

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

  1176.             break;

  1177.         }

  1178.         dest += step;

  1179.     }

  1180. }

  1181. 

  1182. #if CONFIG_SMALL

  1183. YUV2RGBWRAPPERX(yuv2, rgb_full, bgra32_full, PIX_FMT_BGRA,  CONFIG_SWSCALE_ALPHA && c->alpPixBuf)

  1184. YUV2RGBWRAPPERX(yuv2, rgb_full, abgr32_full, PIX_FMT_ABGR,  CONFIG_SWSCALE_ALPHA && c->alpPixBuf)

  1185. YUV2RGBWRAPPERX(yuv2, rgb_full, rgba32_full, PIX_FMT_RGBA,  CONFIG_SWSCALE_ALPHA && c->alpPixBuf)

  1186. YUV2RGBWRAPPERX(yuv2, rgb_full, argb32_full, PIX_FMT_ARGB,  CONFIG_SWSCALE_ALPHA && c->alpPixBuf)

  1187. #else

  1188. #if CONFIG_SWSCALE_ALPHA

  1189. YUV2RGBWRAPPERX(yuv2, rgb_full, bgra32_full, PIX_FMT_BGRA,  1)

  1190. YUV2RGBWRAPPERX(yuv2, rgb_full, abgr32_full, PIX_FMT_ABGR,  1)

  1191. YUV2RGBWRAPPERX(yuv2, rgb_full, rgba32_full, PIX_FMT_RGBA,  1)

  1192. YUV2RGBWRAPPERX(yuv2, rgb_full, argb32_full, PIX_FMT_ARGB,  1)

  1193. #endif

  1194. YUV2RGBWRAPPERX(yuv2, rgb_full, bgrx32_full, PIX_FMT_BGRA,  0)

  1195. YUV2RGBWRAPPERX(yuv2, rgb_full, xbgr32_full, PIX_FMT_ABGR,  0)

  1196. YUV2RGBWRAPPERX(yuv2, rgb_full, rgbx32_full, PIX_FMT_RGBA,  0)

  1197. YUV2RGBWRAPPERX(yuv2, rgb_full, xrgb32_full, PIX_FMT_ARGB,  0)

  1198. #endif

  1199. YUV2RGBWRAPPERX(yuv2, rgb_full, bgr24_full,  PIX_FMT_BGR24, 0)

  1200. YUV2RGBWRAPPERX(yuv2, rgb_full, rgb24_full,  PIX_FMT_RGB24, 0)

  1201. 

  1202. void ff_sws_init_output_funcs(SwsContext *c,

  1203.                               yuv2planar1_fn *yuv2plane1,

  1204.                               yuv2planarX_fn *yuv2planeX,

  1205.                               yuv2interleavedX_fn *yuv2nv12cX,

  1206.                               yuv2packed1_fn *yuv2packed1,

  1207.                               yuv2packed2_fn *yuv2packed2,

  1208.                               yuv2packedX_fn *yuv2packedX)

  1209. {

  1210.     enum PixelFormat dstFormat = c->dstFormat;

  1211. 

  1212.     if (is16BPS(dstFormat)) {

  1213.         *yuv2planeX = isBE(dstFormat) ? yuv2planeX_16BE_c  : yuv2planeX_16LE_c;

  1214.         *yuv2plane1 = isBE(dstFormat) ? yuv2plane1_16BE_c  : yuv2plane1_16LE_c;

  1215.     } else if (is9_OR_10BPS(dstFormat)) {

  1216.         if (av_pix_fmt_descriptors[dstFormat].comp[0].depth_minus1 == 8) {

  1217.             *yuv2planeX = isBE(dstFormat) ? yuv2planeX_9BE_c  : yuv2planeX_9LE_c;

  1218.             *yuv2plane1 = isBE(dstFormat) ? yuv2plane1_9BE_c  : yuv2plane1_9LE_c;

  1219.         } else {

  1220.             *yuv2planeX = isBE(dstFormat) ? yuv2planeX_10BE_c  : yuv2planeX_10LE_c;

  1221.             *yuv2plane1 = isBE(dstFormat) ? yuv2plane1_10BE_c  : yuv2plane1_10LE_c;

  1222.         }

  1223.     } else {

  1224.         *yuv2plane1 = yuv2plane1_8_c;

  1225.         *yuv2planeX = yuv2planeX_8_c;

  1226.         if (dstFormat == PIX_FMT_NV12 || dstFormat == PIX_FMT_NV21)

  1227.             *yuv2nv12cX = yuv2nv12cX_c;

  1228.     }

  1229. 

  1230.     if(c->flags & SWS_FULL_CHR_H_INT) {

  1231.         switch (dstFormat) {

  1232.             case PIX_FMT_RGBA:

  1233. #if CONFIG_SMALL

  1234.                 *yuv2packedX = yuv2rgba32_full_X_c;

  1235. #else

  1236. #if CONFIG_SWSCALE_ALPHA

  1237.                 if (c->alpPixBuf) {

  1238.                     *yuv2packedX = yuv2rgba32_full_X_c;

  1239.                 } else

  1240. #endif /* CONFIG_SWSCALE_ALPHA */

  1241.                 {

  1242.                     *yuv2packedX = yuv2rgbx32_full_X_c;

  1243.                 }

  1244. #endif /* !CONFIG_SMALL */

  1245.                 break;

  1246.             case PIX_FMT_ARGB:

  1247. #if CONFIG_SMALL

  1248.                 *yuv2packedX = yuv2argb32_full_X_c;

  1249. #else

  1250. #if CONFIG_SWSCALE_ALPHA

  1251.                 if (c->alpPixBuf) {

  1252.                     *yuv2packedX = yuv2argb32_full_X_c;

  1253.                 } else

  1254. #endif /* CONFIG_SWSCALE_ALPHA */

  1255.                 {

  1256.                     *yuv2packedX = yuv2xrgb32_full_X_c;

  1257.                 }

  1258. #endif /* !CONFIG_SMALL */

  1259.                 break;

  1260.             case PIX_FMT_BGRA:

  1261. #if CONFIG_SMALL

  1262.                 *yuv2packedX = yuv2bgra32_full_X_c;

  1263. #else

  1264. #if CONFIG_SWSCALE_ALPHA

  1265.                 if (c->alpPixBuf) {

  1266.                     *yuv2packedX = yuv2bgra32_full_X_c;

  1267.                 } else

  1268. #endif /* CONFIG_SWSCALE_ALPHA */

  1269.                 {

  1270.                     *yuv2packedX = yuv2bgrx32_full_X_c;

  1271.                 }

  1272. #endif /* !CONFIG_SMALL */

  1273.                 break;

  1274.             case PIX_FMT_ABGR:

  1275. #if CONFIG_SMALL

  1276.                 *yuv2packedX = yuv2abgr32_full_X_c;

  1277. #else

  1278. #if CONFIG_SWSCALE_ALPHA

  1279.                 if (c->alpPixBuf) {

  1280.                     *yuv2packedX = yuv2abgr32_full_X_c;

  1281.                 } else

  1282. #endif /* CONFIG_SWSCALE_ALPHA */

  1283.                 {

  1284.                     *yuv2packedX = yuv2xbgr32_full_X_c;

  1285.                 }

  1286. #endif /* !CONFIG_SMALL */

  1287.                 break;

  1288.             case PIX_FMT_RGB24:

  1289.             *yuv2packedX = yuv2rgb24_full_X_c;

  1290.             break;

  1291.         case PIX_FMT_BGR24:

  1292.             *yuv2packedX = yuv2bgr24_full_X_c;

  1293.             break;

  1294.         }

  1295.     } else {

  1296.         switch (dstFormat) {

  1297.         case PIX_FMT_RGB48LE:

  1298.             *yuv2packed1 = yuv2rgb48le_1_c;

  1299.             *yuv2packed2 = yuv2rgb48le_2_c;

  1300.             *yuv2packedX = yuv2rgb48le_X_c;

  1301.             break;

  1302.         case PIX_FMT_RGB48BE:

  1303.             *yuv2packed1 = yuv2rgb48be_1_c;

  1304.             *yuv2packed2 = yuv2rgb48be_2_c;

  1305.             *yuv2packedX = yuv2rgb48be_X_c;

  1306.             break;

  1307.         case PIX_FMT_BGR48LE:

  1308.             *yuv2packed1 = yuv2bgr48le_1_c;

  1309.             *yuv2packed2 = yuv2bgr48le_2_c;

  1310.             *yuv2packedX = yuv2bgr48le_X_c;

  1311.             break;

  1312.         case PIX_FMT_BGR48BE:

  1313.             *yuv2packed1 = yuv2bgr48be_1_c;

  1314.             *yuv2packed2 = yuv2bgr48be_2_c;

  1315.             *yuv2packedX = yuv2bgr48be_X_c;

  1316.             break;

  1317.         case PIX_FMT_RGB32:

  1318.         case PIX_FMT_BGR32:

  1319. #if CONFIG_SMALL

  1320.             *yuv2packed1 = yuv2rgb32_1_c;

  1321.             *yuv2packed2 = yuv2rgb32_2_c;

  1322.             *yuv2packedX = yuv2rgb32_X_c;

  1323. #else

  1324. #if CONFIG_SWSCALE_ALPHA

  1325.                 if (c->alpPixBuf) {

  1326.                     *yuv2packed1 = yuv2rgba32_1_c;

  1327.                     *yuv2packed2 = yuv2rgba32_2_c;

  1328.                     *yuv2packedX = yuv2rgba32_X_c;

  1329.                 } else

  1330. #endif /* CONFIG_SWSCALE_ALPHA */

  1331.                 {

  1332.                     *yuv2packed1 = yuv2rgbx32_1_c;

  1333.                     *yuv2packed2 = yuv2rgbx32_2_c;

  1334.                     *yuv2packedX = yuv2rgbx32_X_c;

  1335.                 }

  1336. #endif /* !CONFIG_SMALL */

  1337.             break;

  1338.         case PIX_FMT_RGB32_1:

  1339.         case PIX_FMT_BGR32_1:

  1340. #if CONFIG_SMALL

  1341.                 *yuv2packed1 = yuv2rgb32_1_1_c;

  1342.                 *yuv2packed2 = yuv2rgb32_1_2_c;

  1343.                 *yuv2packedX = yuv2rgb32_1_X_c;

  1344. #else

  1345. #if CONFIG_SWSCALE_ALPHA

  1346.                 if (c->alpPixBuf) {

  1347.                     *yuv2packed1 = yuv2rgba32_1_1_c;

  1348.                     *yuv2packed2 = yuv2rgba32_1_2_c;

  1349.                     *yuv2packedX = yuv2rgba32_1_X_c;

  1350.                 } else

  1351. #endif /* CONFIG_SWSCALE_ALPHA */

  1352.                 {

  1353.                     *yuv2packed1 = yuv2rgbx32_1_1_c;

  1354.                     *yuv2packed2 = yuv2rgbx32_1_2_c;

  1355.                     *yuv2packedX = yuv2rgbx32_1_X_c;

  1356.                 }

  1357. #endif /* !CONFIG_SMALL */

  1358.                 break;

  1359.         case PIX_FMT_RGB24:

  1360.             *yuv2packed1 = yuv2rgb24_1_c;

  1361.             *yuv2packed2 = yuv2rgb24_2_c;

  1362.             *yuv2packedX = yuv2rgb24_X_c;

  1363.             break;

  1364.         case PIX_FMT_BGR24:

  1365.             *yuv2packed1 = yuv2bgr24_1_c;

  1366.             *yuv2packed2 = yuv2bgr24_2_c;

  1367.             *yuv2packedX = yuv2bgr24_X_c;

  1368.             break;

  1369.         case PIX_FMT_RGB565LE:

  1370.         case PIX_FMT_RGB565BE:

  1371.         case PIX_FMT_BGR565LE:

  1372.         case PIX_FMT_BGR565BE:

  1373.             *yuv2packed1 = yuv2rgb16_1_c;

  1374.             *yuv2packed2 = yuv2rgb16_2_c;

  1375.             *yuv2packedX = yuv2rgb16_X_c;

  1376.             break;

  1377.         case PIX_FMT_RGB555LE:

  1378.         case PIX_FMT_RGB555BE:

  1379.         case PIX_FMT_BGR555LE:

  1380.         case PIX_FMT_BGR555BE:

  1381.             *yuv2packed1 = yuv2rgb15_1_c;

  1382.             *yuv2packed2 = yuv2rgb15_2_c;

  1383.             *yuv2packedX = yuv2rgb15_X_c;

  1384.             break;

  1385.         case PIX_FMT_RGB444LE:

  1386.         case PIX_FMT_RGB444BE:

  1387.         case PIX_FMT_BGR444LE:

  1388.         case PIX_FMT_BGR444BE:

  1389.             *yuv2packed1 = yuv2rgb12_1_c;

  1390.             *yuv2packed2 = yuv2rgb12_2_c;

  1391.             *yuv2packedX = yuv2rgb12_X_c;

  1392.             break;

  1393.         case PIX_FMT_RGB8:

  1394.         case PIX_FMT_BGR8:

  1395.             *yuv2packed1 = yuv2rgb8_1_c;

  1396.             *yuv2packed2 = yuv2rgb8_2_c;

  1397.             *yuv2packedX = yuv2rgb8_X_c;

  1398.             break;

  1399.         case PIX_FMT_RGB4:

  1400.         case PIX_FMT_BGR4:

  1401.             *yuv2packed1 = yuv2rgb4_1_c;

  1402.             *yuv2packed2 = yuv2rgb4_2_c;

  1403.             *yuv2packedX = yuv2rgb4_X_c;

  1404.             break;

  1405.         case PIX_FMT_RGB4_BYTE:

  1406.         case PIX_FMT_BGR4_BYTE:

  1407.             *yuv2packed1 = yuv2rgb4b_1_c;

  1408.             *yuv2packed2 = yuv2rgb4b_2_c;

  1409.             *yuv2packedX = yuv2rgb4b_X_c;

  1410.             break;

  1411.         }

  1412.     }

  1413.     switch (dstFormat) {

  1414.     case PIX_FMT_MONOWHITE:

  1415.         *yuv2packed1 = yuv2monowhite_1_c;

  1416.         *yuv2packed2 = yuv2monowhite_2_c;

  1417.         *yuv2packedX = yuv2monowhite_X_c;

  1418.         break;

  1419.     case PIX_FMT_MONOBLACK:

  1420.         *yuv2packed1 = yuv2monoblack_1_c;

  1421.         *yuv2packed2 = yuv2monoblack_2_c;

  1422.         *yuv2packedX = yuv2monoblack_X_c;

  1423.         break;

  1424.     case PIX_FMT_YUYV422:

  1425.         *yuv2packed1 = yuv2yuyv422_1_c;

  1426.         *yuv2packed2 = yuv2yuyv422_2_c;

  1427.         *yuv2packedX = yuv2yuyv422_X_c;

  1428.         break;

  1429.     case PIX_FMT_UYVY422:

  1430.         *yuv2packed1 = yuv2uyvy422_1_c;

  1431.         *yuv2packed2 = yuv2uyvy422_2_c;

  1432.         *yuv2packedX = yuv2uyvy422_X_c;

  1433.         break;

  1434.     }

  1435. }