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.
44652 Commits
32 Branches
283MB
Tree: d33908a5fc
Branches Tags
${ item.name }
Create branch ${ searchTerm }
from 'd33908a5fc'
${ noResults }
FFmpeg/libswscale/swscale_unscaled.c

1276 lines
49KB
Raw Blame History 

  1. /*

  2.  * Copyright (C) 2001-2011 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 <inttypes.h>

  22. #include <string.h>

  23. #include <math.h>

  24. #include <stdio.h>

  25. #include "config.h"

  26. #include "swscale.h"

  27. #include "swscale_internal.h"

  28. #include "rgb2rgb.h"

  29. #include "libavutil/intreadwrite.h"

  30. #include "libavutil/cpu.h"

  31. #include "libavutil/avutil.h"

  32. #include "libavutil/mathematics.h"

  33. #include "libavutil/bswap.h"

  34. #include "libavutil/pixdesc.h"

  35. #include "libavutil/avassert.h"

  36. 

  37. #define RGB2YUV_SHIFT 15

  38. #define BY ( (int) (0.114 * 219 / 255 * (1 << RGB2YUV_SHIFT) + 0.5))

  39. #define BV (-(int) (0.081 * 224 / 255 * (1 << RGB2YUV_SHIFT) + 0.5))

  40. #define BU ( (int) (0.500 * 224 / 255 * (1 << RGB2YUV_SHIFT) + 0.5))

  41. #define GY ( (int) (0.587 * 219 / 255 * (1 << RGB2YUV_SHIFT) + 0.5))

  42. #define GV (-(int) (0.419 * 224 / 255 * (1 << RGB2YUV_SHIFT) + 0.5))

  43. #define GU (-(int) (0.331 * 224 / 255 * (1 << RGB2YUV_SHIFT) + 0.5))

  44. #define RY ( (int) (0.299 * 219 / 255 * (1 << RGB2YUV_SHIFT) + 0.5))

  45. #define RV ( (int) (0.500 * 224 / 255 * (1 << RGB2YUV_SHIFT) + 0.5))

  46. #define RU (-(int) (0.169 * 224 / 255 * (1 << RGB2YUV_SHIFT) + 0.5))

  47. 

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

  49. {

  50.   {   0,  1,  0,  1,  0,  1,  0,  1,},

  51.   {   1,  0,  1,  0,  1,  0,  1,  0,},

  52.   {   0,  1,  0,  1,  0,  1,  0,  1,},

  53.   {   1,  0,  1,  0,  1,  0,  1,  0,},

  54.   {   0,  1,  0,  1,  0,  1,  0,  1,},

  55.   {   1,  0,  1,  0,  1,  0,  1,  0,},

  56.   {   0,  1,  0,  1,  0,  1,  0,  1,},

  57.   {   1,  0,  1,  0,  1,  0,  1,  0,},

  58. },{

  59.   {   1,  2,  1,  2,  1,  2,  1,  2,},

  60.   {   3,  0,  3,  0,  3,  0,  3,  0,},

  61.   {   1,  2,  1,  2,  1,  2,  1,  2,},

  62.   {   3,  0,  3,  0,  3,  0,  3,  0,},

  63.   {   1,  2,  1,  2,  1,  2,  1,  2,},

  64.   {   3,  0,  3,  0,  3,  0,  3,  0,},

  65.   {   1,  2,  1,  2,  1,  2,  1,  2,},

  66.   {   3,  0,  3,  0,  3,  0,  3,  0,},

  67. },{

  68.   {   2,  4,  3,  5,  2,  4,  3,  5,},

  69.   {   6,  0,  7,  1,  6,  0,  7,  1,},

  70.   {   3,  5,  2,  4,  3,  5,  2,  4,},

  71.   {   7,  1,  6,  0,  7,  1,  6,  0,},

  72.   {   2,  4,  3,  5,  2,  4,  3,  5,},

  73.   {   6,  0,  7,  1,  6,  0,  7,  1,},

  74.   {   3,  5,  2,  4,  3,  5,  2,  4,},

  75.   {   7,  1,  6,  0,  7,  1,  6,  0,},

  76. },{

  77.   {   4,  8,  7, 11,  4,  8,  7, 11,},

  78.   {  12,  0, 15,  3, 12,  0, 15,  3,},

  79.   {   6, 10,  5,  9,  6, 10,  5,  9,},

  80.   {  14,  2, 13,  1, 14,  2, 13,  1,},

  81.   {   4,  8,  7, 11,  4,  8,  7, 11,},

  82.   {  12,  0, 15,  3, 12,  0, 15,  3,},

  83.   {   6, 10,  5,  9,  6, 10,  5,  9,},

  84.   {  14,  2, 13,  1, 14,  2, 13,  1,},

  85. },{

  86.   {   9, 17, 15, 23,  8, 16, 14, 22,},

  87.   {  25,  1, 31,  7, 24,  0, 30,  6,},

  88.   {  13, 21, 11, 19, 12, 20, 10, 18,},

  89.   {  29,  5, 27,  3, 28,  4, 26,  2,},

  90.   {   8, 16, 14, 22,  9, 17, 15, 23,},

  91.   {  24,  0, 30,  6, 25,  1, 31,  7,},

  92.   {  12, 20, 10, 18, 13, 21, 11, 19,},

  93.   {  28,  4, 26,  2, 29,  5, 27,  3,},

  94. },{

  95.   {  18, 34, 30, 46, 17, 33, 29, 45,},

  96.   {  50,  2, 62, 14, 49,  1, 61, 13,},

  97.   {  26, 42, 22, 38, 25, 41, 21, 37,},

  98.   {  58, 10, 54,  6, 57,  9, 53,  5,},

  99.   {  16, 32, 28, 44, 19, 35, 31, 47,},

  100.   {  48,  0, 60, 12, 51,  3, 63, 15,},

  101.   {  24, 40, 20, 36, 27, 43, 23, 39,},

  102.   {  56,  8, 52,  4, 59, 11, 55,  7,},

  103. },{

  104.   {  18, 34, 30, 46, 17, 33, 29, 45,},

  105.   {  50,  2, 62, 14, 49,  1, 61, 13,},

  106.   {  26, 42, 22, 38, 25, 41, 21, 37,},

  107.   {  58, 10, 54,  6, 57,  9, 53,  5,},

  108.   {  16, 32, 28, 44, 19, 35, 31, 47,},

  109.   {  48,  0, 60, 12, 51,  3, 63, 15,},

  110.   {  24, 40, 20, 36, 27, 43, 23, 39,},

  111.   {  56,  8, 52,  4, 59, 11, 55,  7,},

  112. },{

  113.   {  36, 68, 60, 92, 34, 66, 58, 90,},

  114.   { 100,  4,124, 28, 98,  2,122, 26,},

  115.   {  52, 84, 44, 76, 50, 82, 42, 74,},

  116.   { 116, 20,108, 12,114, 18,106, 10,},

  117.   {  32, 64, 56, 88, 38, 70, 62, 94,},

  118.   {  96,  0,120, 24,102,  6,126, 30,},

  119.   {  48, 80, 40, 72, 54, 86, 46, 78,},

  120.   { 112, 16,104,  8,118, 22,110, 14,},

  121. }};

  122. 

  123. static const uint8_t flat64[8]={64,64,64,64,64,64,64,64};

  124. 

  125. const uint16_t dither_scale[15][16]={

  126. {    2,    3,    3,    5,    5,    5,    5,    5,    5,    5,    5,    5,    5,    5,    5,    5,},

  127. {    2,    3,    7,    7,   13,   13,   25,   25,   25,   25,   25,   25,   25,   25,   25,   25,},

  128. {    3,    3,    4,   15,   15,   29,   57,   57,   57,  113,  113,  113,  113,  113,  113,  113,},

  129. {    3,    4,    4,    5,   31,   31,   61,  121,  241,  241,  241,  241,  481,  481,  481,  481,},

  130. {    3,    4,    5,    5,    6,   63,   63,  125,  249,  497,  993,  993,  993,  993,  993, 1985,},

  131. {    3,    5,    6,    6,    6,    7,  127,  127,  253,  505, 1009, 2017, 4033, 4033, 4033, 4033,},

  132. {    3,    5,    6,    7,    7,    7,    8,  255,  255,  509, 1017, 2033, 4065, 8129,16257,16257,},

  133. {    3,    5,    6,    8,    8,    8,    8,    9,  511,  511, 1021, 2041, 4081, 8161,16321,32641,},

  134. {    3,    5,    7,    8,    9,    9,    9,    9,   10, 1023, 1023, 2045, 4089, 8177,16353,32705,},

  135. {    3,    5,    7,    8,   10,   10,   10,   10,   10,   11, 2047, 2047, 4093, 8185,16369,32737,},

  136. {    3,    5,    7,    8,   10,   11,   11,   11,   11,   11,   12, 4095, 4095, 8189,16377,32753,},

  137. {    3,    5,    7,    9,   10,   12,   12,   12,   12,   12,   12,   13, 8191, 8191,16381,32761,},

  138. {    3,    5,    7,    9,   10,   12,   13,   13,   13,   13,   13,   13,   14,16383,16383,32765,},

  139. {    3,    5,    7,    9,   10,   12,   14,   14,   14,   14,   14,   14,   14,   15,32767,32767,},

  140. {    3,    5,    7,    9,   11,   12,   14,   15,   15,   15,   15,   15,   15,   15,   16,65535,},

  141. };

  142. 

  143. 

  144. static void fillPlane(uint8_t *plane, int stride, int width, int height, int y,

  145.                       uint8_t val)

  146. {

  147.     int i;

  148.     uint8_t *ptr = plane + stride * y;

  149.     for (i = 0; i < height; i++) {

  150.         memset(ptr, val, width);

  151.         ptr += stride;

  152.     }

  153. }

  154. 

  155. static void fillPlane16(uint8_t *plane, int stride, int width, int height, int y,

  156.                       int alpha, int bits)

  157. {

  158.     int i, j;

  159.     uint8_t *ptr = plane + stride * y;

  160.     int v = alpha ? -1 : (1<<bits);

  161.     for (i = 0; i < height; i++) {

  162.         for (j = 0; j < width; j++) {

  163.             AV_WN16(ptr+2*j, v);

  164.         }

  165.         ptr += stride;

  166.     }

  167. }

  168. 

  169. static void copyPlane(const uint8_t *src, int srcStride,

  170.                       int srcSliceY, int srcSliceH, int width,

  171.                       uint8_t *dst, int dstStride)

  172. {

  173.     dst += dstStride * srcSliceY;

  174.     if (dstStride == srcStride && srcStride > 0) {

  175.         memcpy(dst, src, srcSliceH * dstStride);

  176.     } else {

  177.         int i;

  178.         for (i = 0; i < srcSliceH; i++) {

  179.             memcpy(dst, src, width);

  180.             src += srcStride;

  181.             dst += dstStride;

  182.         }

  183.     }

  184. }

  185. 

  186. static int planarToNv12Wrapper(SwsContext *c, const uint8_t *src[],

  187.                                int srcStride[], int srcSliceY,

  188.                                int srcSliceH, uint8_t *dstParam[],

  189.                                int dstStride[])

  190. {

  191.     uint8_t *dst = dstParam[1] + dstStride[1] * srcSliceY / 2;

  192. 

  193.     copyPlane(src[0], srcStride[0], srcSliceY, srcSliceH, c->srcW,

  194.               dstParam[0], dstStride[0]);

  195. 

  196.     if (c->dstFormat == PIX_FMT_NV12)

  197.         interleaveBytes(src[1], src[2], dst, c->srcW / 2, srcSliceH / 2,

  198.                         srcStride[1], srcStride[2], dstStride[0]);

  199.     else

  200.         interleaveBytes(src[2], src[1], dst, c->srcW / 2, srcSliceH / 2,

  201.                         srcStride[2], srcStride[1], dstStride[0]);

  202. 

  203.     return srcSliceH;

  204. }

  205. 

  206. static int planarToYuy2Wrapper(SwsContext *c, const uint8_t *src[],

  207.                                int srcStride[], int srcSliceY, int srcSliceH,

  208.                                uint8_t *dstParam[], int dstStride[])

  209. {

  210.     uint8_t *dst = dstParam[0] + dstStride[0] * srcSliceY;

  211. 

  212.     yv12toyuy2(src[0], src[1], src[2], dst, c->srcW, srcSliceH, srcStride[0],

  213.                srcStride[1], dstStride[0]);

  214. 

  215.     return srcSliceH;

  216. }

  217. 

  218. static int planarToUyvyWrapper(SwsContext *c, const uint8_t *src[],

  219.                                int srcStride[], int srcSliceY, int srcSliceH,

  220.                                uint8_t *dstParam[], int dstStride[])

  221. {

  222.     uint8_t *dst = dstParam[0] + dstStride[0] * srcSliceY;

  223. 

  224.     yv12touyvy(src[0], src[1], src[2], dst, c->srcW, srcSliceH, srcStride[0],

  225.                srcStride[1], dstStride[0]);

  226. 

  227.     return srcSliceH;

  228. }

  229. 

  230. static int yuv422pToYuy2Wrapper(SwsContext *c, const uint8_t *src[],

  231.                                 int srcStride[], int srcSliceY, int srcSliceH,

  232.                                 uint8_t *dstParam[], int dstStride[])

  233. {

  234.     uint8_t *dst = dstParam[0] + dstStride[0] * srcSliceY;

  235. 

  236.     yuv422ptoyuy2(src[0], src[1], src[2], dst, c->srcW, srcSliceH, srcStride[0],

  237.                   srcStride[1], dstStride[0]);

  238. 

  239.     return srcSliceH;

  240. }

  241. 

  242. static int yuv422pToUyvyWrapper(SwsContext *c, const uint8_t *src[],

  243.                                 int srcStride[], int srcSliceY, int srcSliceH,

  244.                                 uint8_t *dstParam[], int dstStride[])

  245. {

  246.     uint8_t *dst = dstParam[0] + dstStride[0] * srcSliceY;

  247. 

  248.     yuv422ptouyvy(src[0], src[1], src[2], dst, c->srcW, srcSliceH, srcStride[0],

  249.                   srcStride[1], dstStride[0]);

  250. 

  251.     return srcSliceH;

  252. }

  253. 

  254. static int yuyvToYuv420Wrapper(SwsContext *c, const uint8_t *src[],

  255.                                int srcStride[], int srcSliceY, int srcSliceH,

  256.                                uint8_t *dstParam[], int dstStride[])

  257. {

  258.     uint8_t *ydst = dstParam[0] + dstStride[0] * srcSliceY;

  259.     uint8_t *udst = dstParam[1] + dstStride[1] * srcSliceY / 2;

  260.     uint8_t *vdst = dstParam[2] + dstStride[2] * srcSliceY / 2;

  261. 

  262.     yuyvtoyuv420(ydst, udst, vdst, src[0], c->srcW, srcSliceH, dstStride[0],

  263.                  dstStride[1], srcStride[0]);

  264. 

  265.     if (dstParam[3])

  266.         fillPlane(dstParam[3], dstStride[3], c->srcW, srcSliceH, srcSliceY, 255);

  267. 

  268.     return srcSliceH;

  269. }

  270. 

  271. static int yuyvToYuv422Wrapper(SwsContext *c, const uint8_t *src[],

  272.                                int srcStride[], int srcSliceY, int srcSliceH,

  273.                                uint8_t *dstParam[], int dstStride[])

  274. {

  275.     uint8_t *ydst = dstParam[0] + dstStride[0] * srcSliceY;

  276.     uint8_t *udst = dstParam[1] + dstStride[1] * srcSliceY;

  277.     uint8_t *vdst = dstParam[2] + dstStride[2] * srcSliceY;

  278. 

  279.     yuyvtoyuv422(ydst, udst, vdst, src[0], c->srcW, srcSliceH, dstStride[0],

  280.                  dstStride[1], srcStride[0]);

  281. 

  282.     return srcSliceH;

  283. }

  284. 

  285. static int uyvyToYuv420Wrapper(SwsContext *c, const uint8_t *src[],

  286.                                int srcStride[], int srcSliceY, int srcSliceH,

  287.                                uint8_t *dstParam[], int dstStride[])

  288. {

  289.     uint8_t *ydst = dstParam[0] + dstStride[0] * srcSliceY;

  290.     uint8_t *udst = dstParam[1] + dstStride[1] * srcSliceY / 2;

  291.     uint8_t *vdst = dstParam[2] + dstStride[2] * srcSliceY / 2;

  292. 

  293.     uyvytoyuv420(ydst, udst, vdst, src[0], c->srcW, srcSliceH, dstStride[0],

  294.                  dstStride[1], srcStride[0]);

  295. 

  296.     if (dstParam[3])

  297.         fillPlane(dstParam[3], dstStride[3], c->srcW, srcSliceH, srcSliceY, 255);

  298. 

  299.     return srcSliceH;

  300. }

  301. 

  302. static int uyvyToYuv422Wrapper(SwsContext *c, const uint8_t *src[],

  303.                                int srcStride[], int srcSliceY, int srcSliceH,

  304.                                uint8_t *dstParam[], int dstStride[])

  305. {

  306.     uint8_t *ydst = dstParam[0] + dstStride[0] * srcSliceY;

  307.     uint8_t *udst = dstParam[1] + dstStride[1] * srcSliceY;

  308.     uint8_t *vdst = dstParam[2] + dstStride[2] * srcSliceY;

  309. 

  310.     uyvytoyuv422(ydst, udst, vdst, src[0], c->srcW, srcSliceH, dstStride[0],

  311.                  dstStride[1], srcStride[0]);

  312. 

  313.     return srcSliceH;

  314. }

  315. 

  316. static void gray8aToPacked32(const uint8_t *src, uint8_t *dst, int num_pixels,

  317.                              const uint8_t *palette)

  318. {

  319.     int i;

  320.     for (i = 0; i < num_pixels; i++)

  321.         ((uint32_t *) dst)[i] = ((const uint32_t *) palette)[src[i << 1]] | (src[(i << 1) + 1] << 24);

  322. }

  323. 

  324. static void gray8aToPacked32_1(const uint8_t *src, uint8_t *dst, int num_pixels,

  325.                                const uint8_t *palette)

  326. {

  327.     int i;

  328. 

  329.     for (i = 0; i < num_pixels; i++)

  330.         ((uint32_t *) dst)[i] = ((const uint32_t *) palette)[src[i << 1]] | src[(i << 1) + 1];

  331. }

  332. 

  333. static void gray8aToPacked24(const uint8_t *src, uint8_t *dst, int num_pixels,

  334.                              const uint8_t *palette)

  335. {

  336.     int i;

  337. 

  338.     for (i = 0; i < num_pixels; i++) {

  339.         //FIXME slow?

  340.         dst[0] = palette[src[i << 1] * 4 + 0];

  341.         dst[1] = palette[src[i << 1] * 4 + 1];

  342.         dst[2] = palette[src[i << 1] * 4 + 2];

  343.         dst += 3;

  344.     }

  345. }

  346. 

  347. static int packed_16bpc_bswap(SwsContext *c, const uint8_t *src[],

  348.                               int srcStride[], int srcSliceY, int srcSliceH,

  349.                               uint8_t *dst[], int dstStride[])

  350. {

  351.     int i, j;

  352.     int srcstr = srcStride[0] >> 1;

  353.     int dststr = dstStride[0] >> 1;

  354.     uint16_t       *dstPtr =       (uint16_t *) dst[0];

  355.     const uint16_t *srcPtr = (const uint16_t *) src[0];

  356.     int min_stride         = FFMIN(srcstr, dststr);

  357. 

  358.     for (i = 0; i < srcSliceH; i++) {

  359.         for (j = 0; j < min_stride; j++) {

  360.             dstPtr[j] = av_bswap16(srcPtr[j]);

  361.         }

  362.         srcPtr += srcstr;

  363.         dstPtr += dststr;

  364.     }

  365. 

  366.     return srcSliceH;

  367. }

  368. 

  369. static int palToRgbWrapper(SwsContext *c, const uint8_t *src[], int srcStride[],

  370.                            int srcSliceY, int srcSliceH, uint8_t *dst[],

  371.                            int dstStride[])

  372. {

  373.     const enum PixelFormat srcFormat = c->srcFormat;

  374.     const enum PixelFormat dstFormat = c->dstFormat;

  375.     void (*conv)(const uint8_t *src, uint8_t *dst, int num_pixels,

  376.                  const uint8_t *palette) = NULL;

  377.     int i;

  378.     uint8_t *dstPtr = dst[0] + dstStride[0] * srcSliceY;

  379.     const uint8_t *srcPtr = src[0];

  380. 

  381.     if (srcFormat == PIX_FMT_GRAY8A) {

  382.         switch (dstFormat) {

  383.         case PIX_FMT_RGB32  : conv = gray8aToPacked32; break;

  384.         case PIX_FMT_BGR32  : conv = gray8aToPacked32; break;

  385.         case PIX_FMT_BGR32_1: conv = gray8aToPacked32_1; break;

  386.         case PIX_FMT_RGB32_1: conv = gray8aToPacked32_1; break;

  387.         case PIX_FMT_RGB24  : conv = gray8aToPacked24; break;

  388.         case PIX_FMT_BGR24  : conv = gray8aToPacked24; break;

  389.         }

  390.     } else if (usePal(srcFormat)) {

  391.         switch (dstFormat) {

  392.         case PIX_FMT_RGB32  : conv = sws_convertPalette8ToPacked32; break;

  393.         case PIX_FMT_BGR32  : conv = sws_convertPalette8ToPacked32; break;

  394.         case PIX_FMT_BGR32_1: conv = sws_convertPalette8ToPacked32; break;

  395.         case PIX_FMT_RGB32_1: conv = sws_convertPalette8ToPacked32; break;

  396.         case PIX_FMT_RGB24  : conv = sws_convertPalette8ToPacked24; break;

  397.         case PIX_FMT_BGR24  : conv = sws_convertPalette8ToPacked24; break;

  398.         }

  399.     }

  400. 

  401.     if (!conv)

  402.         av_log(c, AV_LOG_ERROR, "internal error %s -> %s converter\n",

  403.                av_get_pix_fmt_name(srcFormat), av_get_pix_fmt_name(dstFormat));

  404.     else {

  405.         for (i = 0; i < srcSliceH; i++) {

  406.             conv(srcPtr, dstPtr, c->srcW, (uint8_t *) c->pal_rgb);

  407.             srcPtr += srcStride[0];

  408.             dstPtr += dstStride[0];

  409.         }

  410.     }

  411. 

  412.     return srcSliceH;

  413. }

  414. 

  415. static void gbr24ptopacked24(const uint8_t *src[], int srcStride[],

  416.                              uint8_t *dst, int dstStride, int srcSliceH,

  417.                              int width)

  418. {

  419.     int x, h, i;

  420.     for (h = 0; h < srcSliceH; h++) {

  421.         uint8_t *dest = dst + dstStride * h;

  422.         for (x = 0; x < width; x++) {

  423.             *dest++ = src[0][x];

  424.             *dest++ = src[1][x];

  425.             *dest++ = src[2][x];

  426.         }

  427. 

  428.         for (i = 0; i < 3; i++)

  429.             src[i] += srcStride[i];

  430.     }

  431. }

  432. 

  433. static void gbr24ptopacked32(const uint8_t *src[], int srcStride[],

  434.                              uint8_t *dst, int dstStride, int srcSliceH,

  435.                              int alpha_first, int width)

  436. {

  437.     int x, h, i;

  438.     for (h = 0; h < srcSliceH; h++) {

  439.         uint8_t *dest = dst + dstStride * h;

  440. 

  441.         if (alpha_first) {

  442.             for (x = 0; x < width; x++) {

  443.                 *dest++ = 0xff;

  444.                 *dest++ = src[0][x];

  445.                 *dest++ = src[1][x];

  446.                 *dest++ = src[2][x];

  447.             }

  448.         } else {

  449.             for (x = 0; x < width; x++) {

  450.                 *dest++ = src[0][x];

  451.                 *dest++ = src[1][x];

  452.                 *dest++ = src[2][x];

  453.                 *dest++ = 0xff;

  454.             }

  455.         }

  456. 

  457.         for (i = 0; i < 3; i++)

  458.             src[i] += srcStride[i];

  459.     }

  460. }

  461. 

  462. static int planarRgbToRgbWrapper(SwsContext *c, const uint8_t *src[],

  463.                                  int srcStride[], int srcSliceY, int srcSliceH,

  464.                                  uint8_t *dst[], int dstStride[])

  465. {

  466.     int alpha_first = 0;

  467.     if (c->srcFormat != PIX_FMT_GBRP) {

  468.         av_log(c, AV_LOG_ERROR, "unsupported planar RGB conversion %s -> %s\n",

  469.                av_get_pix_fmt_name(c->srcFormat),

  470.                av_get_pix_fmt_name(c->dstFormat));

  471.         return srcSliceH;

  472.     }

  473. 

  474.     switch (c->dstFormat) {

  475.     case PIX_FMT_BGR24:

  476.         gbr24ptopacked24((const uint8_t *[]) { src[1], src[0], src[2] },

  477.                          (int []) { srcStride[1], srcStride[0], srcStride[2] },

  478.                          dst[0] + srcSliceY * dstStride[0], dstStride[0],

  479.                          srcSliceH, c->srcW);

  480.         break;

  481. 

  482.     case PIX_FMT_RGB24:

  483.         gbr24ptopacked24((const uint8_t *[]) { src[2], src[0], src[1] },

  484.                          (int []) { srcStride[2], srcStride[0], srcStride[1] },

  485.                          dst[0] + srcSliceY * dstStride[0], dstStride[0],

  486.                          srcSliceH, c->srcW);

  487.         break;

  488. 

  489.     case PIX_FMT_ARGB:

  490.         alpha_first = 1;

  491.     case PIX_FMT_RGBA:

  492.         gbr24ptopacked32((const uint8_t *[]) { src[2], src[0], src[1] },

  493.                          (int []) { srcStride[2], srcStride[0], srcStride[1] },

  494.                          dst[0] + srcSliceY * dstStride[0], dstStride[0],

  495.                          srcSliceH, alpha_first, c->srcW);

  496.         break;

  497. 

  498.     case PIX_FMT_ABGR:

  499.         alpha_first = 1;

  500.     case PIX_FMT_BGRA:

  501.         gbr24ptopacked32((const uint8_t *[]) { src[1], src[0], src[2] },

  502.                          (int []) { srcStride[1], srcStride[0], srcStride[2] },

  503.                          dst[0] + srcSliceY * dstStride[0], dstStride[0],

  504.                          srcSliceH, alpha_first, c->srcW);

  505.         break;

  506. 

  507.     default:

  508.         av_log(c, AV_LOG_ERROR,

  509.                "unsupported planar RGB conversion %s -> %s\n",

  510.                av_get_pix_fmt_name(c->srcFormat),

  511.                av_get_pix_fmt_name(c->dstFormat));

  512.     }

  513. 

  514.     return srcSliceH;

  515. }

  516. 

  517. #define isRGBA32(x) (            \

  518.            (x) == PIX_FMT_ARGB   \

  519.         || (x) == PIX_FMT_RGBA   \

  520.         || (x) == PIX_FMT_BGRA   \

  521.         || (x) == PIX_FMT_ABGR   \

  522.         )

  523. 

  524. #define isRGBA64(x) (                \

  525.            (x) == PIX_FMT_RGBA64LE   \

  526.         || (x) == PIX_FMT_RGBA64BE   \

  527.         || (x) == PIX_FMT_BGRA64LE   \

  528.         || (x) == PIX_FMT_BGRA64BE   \

  529.         )

  530. 

  531. #define isRGB48(x) (                \

  532.            (x) == PIX_FMT_RGB48LE   \

  533.         || (x) == PIX_FMT_RGB48BE   \

  534.         || (x) == PIX_FMT_BGR48LE   \

  535.         || (x) == PIX_FMT_BGR48BE   \

  536.         )

  537. 

  538. /* {RGB,BGR}{15,16,24,32,32_1} -> {RGB,BGR}{15,16,24,32} */

  539. typedef void (* rgbConvFn) (const uint8_t *, uint8_t *, int);

  540. static rgbConvFn findRgbConvFn(SwsContext *c)

  541. {

  542.     const enum PixelFormat srcFormat = c->srcFormat;

  543.     const enum PixelFormat dstFormat = c->dstFormat;

  544.     const int srcId = c->srcFormatBpp;

  545.     const int dstId = c->dstFormatBpp;

  546.     rgbConvFn conv = NULL;

  547. 

  548. #define IS_NOT_NE(bpp, fmt) \

  549.     (((bpp + 7) >> 3) == 2 && \

  550.      (!(av_pix_fmt_descriptors[fmt].flags & PIX_FMT_BE) != !HAVE_BIGENDIAN))

  551. 

  552. #define CONV_IS(src, dst) (srcFormat == PIX_FMT_##src && dstFormat == PIX_FMT_##dst)

  553. 

  554.     if (isRGBA32(srcFormat) && isRGBA32(dstFormat)) {

  555.         if (     CONV_IS(ABGR, RGBA)

  556.               || CONV_IS(ARGB, BGRA)

  557.               || CONV_IS(BGRA, ARGB)

  558.               || CONV_IS(RGBA, ABGR)) conv = shuffle_bytes_3210;

  559.         else if (CONV_IS(ABGR, ARGB)

  560.               || CONV_IS(ARGB, ABGR)) conv = shuffle_bytes_0321;

  561.         else if (CONV_IS(ABGR, BGRA)

  562.               || CONV_IS(ARGB, RGBA)) conv = shuffle_bytes_1230;

  563.         else if (CONV_IS(BGRA, RGBA)

  564.               || CONV_IS(RGBA, BGRA)) conv = shuffle_bytes_2103;

  565.         else if (CONV_IS(BGRA, ABGR)

  566.               || CONV_IS(RGBA, ARGB)) conv = shuffle_bytes_3012;

  567.     } else if (isRGB48(srcFormat) && isRGB48(dstFormat)) {

  568.         if      (CONV_IS(RGB48LE, BGR48LE)

  569.               || CONV_IS(BGR48LE, RGB48LE)

  570.               || CONV_IS(RGB48BE, BGR48BE)

  571.               || CONV_IS(BGR48BE, RGB48BE)) conv = rgb48tobgr48_nobswap;

  572.         else if (CONV_IS(RGB48LE, BGR48BE)

  573.               || CONV_IS(BGR48LE, RGB48BE)

  574.               || CONV_IS(RGB48BE, BGR48LE)

  575.               || CONV_IS(BGR48BE, RGB48LE)) conv = rgb48tobgr48_bswap;

  576.     } else if (isRGBA64(srcFormat) && isRGB48(dstFormat)) {

  577.         if      (CONV_IS(RGBA64LE, BGR48LE)

  578.               || CONV_IS(BGRA64LE, RGB48LE)

  579.               || CONV_IS(RGBA64BE, BGR48BE)

  580.               || CONV_IS(BGRA64BE, RGB48BE)) conv = rgb64tobgr48_nobswap;

  581.         else if (CONV_IS(RGBA64LE, BGR48BE)

  582.               || CONV_IS(BGRA64LE, RGB48BE)

  583.               || CONV_IS(RGBA64BE, BGR48LE)

  584.               || CONV_IS(BGRA64BE, RGB48LE)) conv = rgb64tobgr48_bswap;

  585.         else if (CONV_IS(RGBA64LE, RGB48LE)

  586.               || CONV_IS(BGRA64LE, BGR48LE)

  587.               || CONV_IS(RGBA64BE, RGB48BE)

  588.               || CONV_IS(BGRA64BE, BGR48BE)) conv = rgb64to48_nobswap;

  589.         else if (CONV_IS(RGBA64LE, RGB48BE)

  590.               || CONV_IS(BGRA64LE, BGR48BE)

  591.               || CONV_IS(RGBA64BE, RGB48LE)

  592.               || CONV_IS(BGRA64BE, BGR48LE)) conv = rgb64to48_bswap;

  593.     } else

  594.     /* BGR -> BGR */

  595.     if ((isBGRinInt(srcFormat) && isBGRinInt(dstFormat)) ||

  596.         (isRGBinInt(srcFormat) && isRGBinInt(dstFormat))) {

  597.         switch (srcId | (dstId << 16)) {

  598.         case 0x000F000C: conv = rgb12to15; break;

  599.         case 0x000F0010: conv = rgb16to15; break;

  600.         case 0x000F0018: conv = rgb24to15; break;

  601.         case 0x000F0020: conv = rgb32to15; break;

  602.         case 0x0010000F: conv = rgb15to16; break;

  603.         case 0x00100018: conv = rgb24to16; break;

  604.         case 0x00100020: conv = rgb32to16; break;

  605.         case 0x0018000F: conv = rgb15to24; break;

  606.         case 0x00180010: conv = rgb16to24; break;

  607.         case 0x00180020: conv = rgb32to24; break;

  608.         case 0x0020000F: conv = rgb15to32; break;

  609.         case 0x00200010: conv = rgb16to32; break;

  610.         case 0x00200018: conv = rgb24to32; break;

  611.         }

  612.     } else if ((isBGRinInt(srcFormat) && isRGBinInt(dstFormat)) ||

  613.                (isRGBinInt(srcFormat) && isBGRinInt(dstFormat))) {

  614.         switch (srcId | (dstId << 16)) {

  615.         case 0x000C000C: conv = rgb12tobgr12; break;

  616.         case 0x000F000F: conv = rgb15tobgr15; break;

  617.         case 0x000F0010: conv = rgb16tobgr15; break;

  618.         case 0x000F0018: conv = rgb24tobgr15; break;

  619.         case 0x000F0020: conv = rgb32tobgr15; break;

  620.         case 0x0010000F: conv = rgb15tobgr16; break;

  621.         case 0x00100010: conv = rgb16tobgr16; break;

  622.         case 0x00100018: conv = rgb24tobgr16; break;

  623.         case 0x00100020: conv = rgb32tobgr16; break;

  624.         case 0x0018000F: conv = rgb15tobgr24; break;

  625.         case 0x00180010: conv = rgb16tobgr24; break;

  626.         case 0x00180018: conv = rgb24tobgr24; break;

  627.         case 0x00180020: conv = rgb32tobgr24; break;

  628.         case 0x0020000F: conv = rgb15tobgr32; break;

  629.         case 0x00200010: conv = rgb16tobgr32; break;

  630.         case 0x00200018: conv = rgb24tobgr32; break;

  631.         }

  632.     }

  633. 

  634.     return conv;

  635. }

  636. 

  637. /* {RGB,BGR}{15,16,24,32,32_1} -> {RGB,BGR}{15,16,24,32} */

  638. static int rgbToRgbWrapper(SwsContext *c, const uint8_t *src[], int srcStride[],

  639.                            int srcSliceY, int srcSliceH, uint8_t *dst[],

  640.                            int dstStride[])

  641. 

  642. {

  643.     const enum PixelFormat srcFormat = c->srcFormat;

  644.     const enum PixelFormat dstFormat = c->dstFormat;

  645.     const int srcBpp = (c->srcFormatBpp + 7) >> 3;

  646.     const int dstBpp = (c->dstFormatBpp + 7) >> 3;

  647.     rgbConvFn conv = findRgbConvFn(c);

  648. 

  649.     if (!conv) {

  650.         av_log(c, AV_LOG_ERROR, "internal error %s -> %s converter\n",

  651.                av_get_pix_fmt_name(srcFormat), av_get_pix_fmt_name(dstFormat));

  652.     } else {

  653.         const uint8_t *srcPtr = src[0];

  654.               uint8_t *dstPtr = dst[0];

  655.         int src_bswap = IS_NOT_NE(c->srcFormatBpp, srcFormat);

  656.         int dst_bswap = IS_NOT_NE(c->dstFormatBpp, dstFormat);

  657. 

  658.         if ((srcFormat == PIX_FMT_RGB32_1 || srcFormat == PIX_FMT_BGR32_1) &&

  659.             !isRGBA32(dstFormat))

  660.             srcPtr += ALT32_CORR;

  661. 

  662.         if ((dstFormat == PIX_FMT_RGB32_1 || dstFormat == PIX_FMT_BGR32_1) &&

  663.             !isRGBA32(srcFormat))

  664.             dstPtr += ALT32_CORR;

  665. 

  666.         if (dstStride[0] * srcBpp == srcStride[0] * dstBpp && srcStride[0] > 0 &&

  667.             !(srcStride[0] % srcBpp) && !dst_bswap && !src_bswap)

  668.             conv(srcPtr, dstPtr + dstStride[0] * srcSliceY,

  669.                  srcSliceH * srcStride[0]);

  670.         else {

  671.             int i, j;

  672.             dstPtr += dstStride[0] * srcSliceY;

  673. 

  674.             for (i = 0; i < srcSliceH; i++) {

  675.                 if(src_bswap) {

  676.                     for(j=0; j<c->srcW; j++)

  677.                         ((uint16_t*)c->formatConvBuffer)[j] = av_bswap16(((uint16_t*)srcPtr)[j]);

  678.                     conv(c->formatConvBuffer, dstPtr, c->srcW * srcBpp);

  679.                 }else

  680.                     conv(srcPtr, dstPtr, c->srcW * srcBpp);

  681.                 if(dst_bswap)

  682.                     for(j=0; j<c->srcW; j++)

  683.                         ((uint16_t*)dstPtr)[j] = av_bswap16(((uint16_t*)dstPtr)[j]);

  684.                 srcPtr += srcStride[0];

  685.                 dstPtr += dstStride[0];

  686.             }

  687.         }

  688.     }

  689.     return srcSliceH;

  690. }

  691. 

  692. static int bgr24ToYv12Wrapper(SwsContext *c, const uint8_t *src[],

  693.                               int srcStride[], int srcSliceY, int srcSliceH,

  694.                               uint8_t *dst[], int dstStride[])

  695. {

  696.     rgb24toyv12(

  697.         src[0],

  698.         dst[0] +  srcSliceY       * dstStride[0],

  699.         dst[1] + (srcSliceY >> 1) * dstStride[1],

  700.         dst[2] + (srcSliceY >> 1) * dstStride[2],

  701.         c->srcW, srcSliceH,

  702.         dstStride[0], dstStride[1], srcStride[0]);

  703.     if (dst[3])

  704.         fillPlane(dst[3], dstStride[3], c->srcW, srcSliceH, srcSliceY, 255);

  705.     return srcSliceH;

  706. }

  707. 

  708. static int yvu9ToYv12Wrapper(SwsContext *c, const uint8_t *src[],

  709.                              int srcStride[], int srcSliceY, int srcSliceH,

  710.                              uint8_t *dst[], int dstStride[])

  711. {

  712.     copyPlane(src[0], srcStride[0], srcSliceY, srcSliceH, c->srcW,

  713.               dst[0], dstStride[0]);

  714. 

  715.     planar2x(src[1], dst[1] + dstStride[1] * (srcSliceY >> 1), c->chrSrcW,

  716.              srcSliceH >> 2, srcStride[1], dstStride[1]);

  717.     planar2x(src[2], dst[2] + dstStride[2] * (srcSliceY >> 1), c->chrSrcW,

  718.              srcSliceH >> 2, srcStride[2], dstStride[2]);

  719.     if (dst[3])

  720.         fillPlane(dst[3], dstStride[3], c->srcW, srcSliceH, srcSliceY, 255);

  721.     return srcSliceH;

  722. }

  723. 

  724. /* unscaled copy like stuff (assumes nearly identical formats) */

  725. static int packedCopyWrapper(SwsContext *c, const uint8_t *src[],

  726.                              int srcStride[], int srcSliceY, int srcSliceH,

  727.                              uint8_t *dst[], int dstStride[])

  728. {

  729.     if (dstStride[0] == srcStride[0] && srcStride[0] > 0)

  730.         memcpy(dst[0] + dstStride[0] * srcSliceY, src[0], srcSliceH * dstStride[0]);

  731.     else {

  732.         int i;

  733.         const uint8_t *srcPtr = src[0];

  734.         uint8_t *dstPtr = dst[0] + dstStride[0] * srcSliceY;

  735.         int length = 0;

  736. 

  737.         /* universal length finder */

  738.         while (length + c->srcW <= FFABS(dstStride[0]) &&

  739.                length + c->srcW <= FFABS(srcStride[0]))

  740.             length += c->srcW;

  741.         av_assert1(length != 0);

  742. 

  743.         for (i = 0; i < srcSliceH; i++) {

  744.             memcpy(dstPtr, srcPtr, length);

  745.             srcPtr += srcStride[0];

  746.             dstPtr += dstStride[0];

  747.         }

  748.     }

  749.     return srcSliceH;

  750. }

  751. 

  752. #define DITHER_COPY(dst, dstStride, src, srcStride, bswap, dbswap)\

  753.     uint16_t scale= dither_scale[dst_depth-1][src_depth-1];\

  754.     int shift= src_depth-dst_depth + dither_scale[src_depth-2][dst_depth-1];\

  755.     for (i = 0; i < height; i++) {\

  756.         const uint8_t *dither= dithers[src_depth-9][i&7];\

  757.         for (j = 0; j < length-7; j+=8){\

  758.             dst[j+0] = dbswap((bswap(src[j+0]) + dither[0])*scale>>shift);\

  759.             dst[j+1] = dbswap((bswap(src[j+1]) + dither[1])*scale>>shift);\

  760.             dst[j+2] = dbswap((bswap(src[j+2]) + dither[2])*scale>>shift);\

  761.             dst[j+3] = dbswap((bswap(src[j+3]) + dither[3])*scale>>shift);\

  762.             dst[j+4] = dbswap((bswap(src[j+4]) + dither[4])*scale>>shift);\

  763.             dst[j+5] = dbswap((bswap(src[j+5]) + dither[5])*scale>>shift);\

  764.             dst[j+6] = dbswap((bswap(src[j+6]) + dither[6])*scale>>shift);\

  765.             dst[j+7] = dbswap((bswap(src[j+7]) + dither[7])*scale>>shift);\

  766.         }\

  767.         for (; j < length; j++)\

  768.             dst[j] = dbswap((bswap(src[j]) + dither[j&7])*scale>>shift);\

  769.         dst += dstStride;\

  770.         src += srcStride;\

  771.     }

  772. 

  773. static int planarCopyWrapper(SwsContext *c, const uint8_t *src[],

  774.                              int srcStride[], int srcSliceY, int srcSliceH,

  775.                              uint8_t *dst[], int dstStride[])

  776. {

  777.     int plane, i, j;

  778.     for (plane = 0; plane < 4; plane++) {

  779.         int length = (plane == 0 || plane == 3) ? c->srcW  : -((-c->srcW  ) >> c->chrDstHSubSample);

  780.         int y =      (plane == 0 || plane == 3) ? srcSliceY: -((-srcSliceY) >> c->chrDstVSubSample);

  781.         int height = (plane == 0 || plane == 3) ? srcSliceH: -((-srcSliceH) >> c->chrDstVSubSample);

  782.         const uint8_t *srcPtr = src[plane];

  783.         uint8_t *dstPtr = dst[plane] + dstStride[plane] * y;

  784.         int shiftonly= plane==1 || plane==2 || (!c->srcRange && plane==0);

  785. 

  786.         if (!dst[plane])

  787.             continue;

  788.         // ignore palette for GRAY8

  789.         if (plane == 1 && !dst[2]) continue;

  790.         if (!src[plane] || (plane == 1 && !src[2])) {

  791.             if (is16BPS(c->dstFormat) || isNBPS(c->dstFormat)) {

  792.                 fillPlane16(dst[plane], dstStride[plane], length, height, y,

  793.                         plane == 3, av_pix_fmt_descriptors[c->dstFormat].comp[plane].depth_minus1);

  794.             } else {

  795.                 fillPlane(dst[plane], dstStride[plane], length, height, y,

  796.                         (plane == 3) ? 255 : 128);

  797.             }

  798.         } else {

  799.             if(isNBPS(c->srcFormat) || isNBPS(c->dstFormat)

  800.                || (is16BPS(c->srcFormat) != is16BPS(c->dstFormat))

  801.             ) {

  802.                 const int src_depth = av_pix_fmt_descriptors[c->srcFormat].comp[plane].depth_minus1 + 1;

  803.                 const int dst_depth = av_pix_fmt_descriptors[c->dstFormat].comp[plane].depth_minus1 + 1;

  804.                 const uint16_t *srcPtr2 = (const uint16_t *) srcPtr;

  805.                 uint16_t *dstPtr2 = (uint16_t*)dstPtr;

  806. 

  807.                 if (dst_depth == 8) {

  808.                     if(isBE(c->srcFormat) == HAVE_BIGENDIAN){

  809.                         DITHER_COPY(dstPtr, dstStride[plane], srcPtr2, srcStride[plane]/2, , )

  810.                     } else {

  811.                         DITHER_COPY(dstPtr, dstStride[plane], srcPtr2, srcStride[plane]/2, av_bswap16, )

  812.                     }

  813.                 } else if (src_depth == 8) {

  814.                     for (i = 0; i < height; i++) {

  815.                         #define COPY816(w)\

  816.                         if(shiftonly){\

  817.                             for (j = 0; j < length; j++)\

  818.                                 w(&dstPtr2[j], srcPtr[j]<<(dst_depth-8));\

  819.                         }else{\

  820.                             for (j = 0; j < length; j++)\

  821.                                 w(&dstPtr2[j], (srcPtr[j]<<(dst_depth-8)) |\

  822.                                                (srcPtr[j]>>(2*8-dst_depth)));\

  823.                         }

  824.                         if(isBE(c->dstFormat)){

  825.                             COPY816(AV_WB16)

  826.                         } else {

  827.                             COPY816(AV_WL16)

  828.                         }

  829.                         dstPtr2 += dstStride[plane]/2;

  830.                         srcPtr  += srcStride[plane];

  831.                     }

  832.                 } else if (src_depth <= dst_depth) {

  833.                     int orig_length = length;

  834.                     for (i = 0; i < height; i++) {

  835.                         if(isBE(c->srcFormat) == HAVE_BIGENDIAN &&

  836.                            isBE(c->dstFormat) == HAVE_BIGENDIAN &&

  837.                            shiftonly) {

  838.                              unsigned shift = dst_depth - src_depth;

  839.                              length = orig_length;

  840. #if HAVE_FAST_64BIT

  841. #define FAST_COPY_UP(shift) \

  842.     for (j = 0; j < length - 3; j += 4) { \

  843.         uint64_t v = AV_RN64A(srcPtr2 + j); \

  844.         AV_WN64A(dstPtr2 + j, v << shift); \

  845.     } \

  846.     length &= 3;

  847. #else

  848. #define FAST_COPY_UP(shift) \

  849.     for (j = 0; j < length - 1; j += 2) { \

  850.         uint32_t v = AV_RN32A(srcPtr2 + j); \

  851.         AV_WN32A(dstPtr2 + j, v << shift); \

  852.     } \

  853.     length &= 1;

  854. #endif

  855.                              switch (shift)

  856.                              {

  857.                              case 6: FAST_COPY_UP(6); break;

  858.                              case 7: FAST_COPY_UP(7); break;

  859.                              }

  860.                         }

  861. #define COPY_UP(r,w) \

  862.     if(shiftonly){\

  863.         for (j = 0; j < length; j++){ \

  864.             unsigned int v= r(&srcPtr2[j]);\

  865.             w(&dstPtr2[j], v<<(dst_depth-src_depth));\

  866.         }\

  867.     }else{\

  868.         for (j = 0; j < length; j++){ \

  869.             unsigned int v= r(&srcPtr2[j]);\

  870.             w(&dstPtr2[j], (v<<(dst_depth-src_depth)) | \

  871.                         (v>>(2*src_depth-dst_depth)));\

  872.         }\

  873.     }

  874.                         if(isBE(c->srcFormat)){

  875.                             if(isBE(c->dstFormat)){

  876.                                 COPY_UP(AV_RB16, AV_WB16)

  877.                             } else {

  878.                                 COPY_UP(AV_RB16, AV_WL16)

  879.                             }

  880.                         } else {

  881.                             if(isBE(c->dstFormat)){

  882.                                 COPY_UP(AV_RL16, AV_WB16)

  883.                             } else {

  884.                                 COPY_UP(AV_RL16, AV_WL16)

  885.                             }

  886.                         }

  887.                         dstPtr2 += dstStride[plane]/2;

  888.                         srcPtr2 += srcStride[plane]/2;

  889.                     }

  890.                 } else {

  891.                     if(isBE(c->srcFormat) == HAVE_BIGENDIAN){

  892.                         if(isBE(c->dstFormat) == HAVE_BIGENDIAN){

  893.                             DITHER_COPY(dstPtr2, dstStride[plane]/2, srcPtr2, srcStride[plane]/2, , )

  894.                         } else {

  895.                             DITHER_COPY(dstPtr2, dstStride[plane]/2, srcPtr2, srcStride[plane]/2, , av_bswap16)

  896.                         }

  897.                     }else{

  898.                         if(isBE(c->dstFormat) == HAVE_BIGENDIAN){

  899.                             DITHER_COPY(dstPtr2, dstStride[plane]/2, srcPtr2, srcStride[plane]/2, av_bswap16, )

  900.                         } else {

  901.                             DITHER_COPY(dstPtr2, dstStride[plane]/2, srcPtr2, srcStride[plane]/2, av_bswap16, av_bswap16)

  902.                         }

  903.                     }

  904.                 }

  905.             } else if (is16BPS(c->srcFormat) && is16BPS(c->dstFormat) &&

  906.                       isBE(c->srcFormat) != isBE(c->dstFormat)) {

  907. 

  908.                 for (i = 0; i < height; i++) {

  909.                     for (j = 0; j < length; j++)

  910.                         ((uint16_t *) dstPtr)[j] = av_bswap16(((const uint16_t *) srcPtr)[j]);

  911.                     srcPtr += srcStride[plane];

  912.                     dstPtr += dstStride[plane];

  913.                 }

  914.             } else if (dstStride[plane] == srcStride[plane] &&

  915.                        srcStride[plane] > 0 && srcStride[plane] == length) {

  916.                 memcpy(dst[plane] + dstStride[plane] * y, src[plane],

  917.                        height * dstStride[plane]);

  918.             } else {

  919.                 if (is16BPS(c->srcFormat) && is16BPS(c->dstFormat))

  920.                     length *= 2;

  921.                 else if (!av_pix_fmt_descriptors[c->srcFormat].comp[0].depth_minus1)

  922.                     length >>= 3; // monowhite/black

  923.                 for (i = 0; i < height; i++) {

  924.                     memcpy(dstPtr, srcPtr, length);

  925.                     srcPtr += srcStride[plane];

  926.                     dstPtr += dstStride[plane];

  927.                 }

  928.             }

  929.         }

  930.     }

  931.     return srcSliceH;

  932. }

  933. 

  934. 

  935. #define IS_DIFFERENT_ENDIANESS(src_fmt, dst_fmt, pix_fmt)          \

  936.     ((src_fmt == pix_fmt ## BE && dst_fmt == pix_fmt ## LE) ||     \

  937.      (src_fmt == pix_fmt ## LE && dst_fmt == pix_fmt ## BE))

  938. 

  939. 

  940. void ff_get_unscaled_swscale(SwsContext *c)

  941. {

  942.     const enum PixelFormat srcFormat = c->srcFormat;

  943.     const enum PixelFormat dstFormat = c->dstFormat;

  944.     const int flags = c->flags;

  945.     const int dstH = c->dstH;

  946.     int needsDither;

  947. 

  948.     needsDither = isAnyRGB(dstFormat) &&

  949.             c->dstFormatBpp < 24 &&

  950.            (c->dstFormatBpp < c->srcFormatBpp || (!isAnyRGB(srcFormat)));

  951. 

  952.     /* yv12_to_nv12 */

  953.     if ((srcFormat == PIX_FMT_YUV420P || srcFormat == PIX_FMT_YUVA420P) &&

  954.         (dstFormat == PIX_FMT_NV12 || dstFormat == PIX_FMT_NV21)) {

  955.         c->swScale = planarToNv12Wrapper;

  956.     }

  957.     /* yuv2bgr */

  958.     if ((srcFormat == PIX_FMT_YUV420P || srcFormat == PIX_FMT_YUV422P ||

  959.          srcFormat == PIX_FMT_YUVA420P) && isAnyRGB(dstFormat) &&

  960.         !(flags & SWS_ACCURATE_RND) && !(dstH & 1)) {

  961.         c->swScale = ff_yuv2rgb_get_func_ptr(c);

  962.     }

  963. 

  964.     if (srcFormat == PIX_FMT_YUV410P &&

  965.         (dstFormat == PIX_FMT_YUV420P || dstFormat == PIX_FMT_YUVA420P) &&

  966.         !(flags & SWS_BITEXACT)) {

  967.         c->swScale = yvu9ToYv12Wrapper;

  968.     }

  969. 

  970.     /* bgr24toYV12 */

  971.     if (srcFormat == PIX_FMT_BGR24 &&

  972.         (dstFormat == PIX_FMT_YUV420P || dstFormat == PIX_FMT_YUVA420P) &&

  973.         !(flags & SWS_ACCURATE_RND))

  974.         c->swScale = bgr24ToYv12Wrapper;

  975. 

  976.     /* RGB/BGR -> RGB/BGR (no dither needed forms) */

  977.     if (isAnyRGB(srcFormat) && isAnyRGB(dstFormat) && findRgbConvFn(c)

  978.         && (!needsDither || (c->flags&(SWS_FAST_BILINEAR|SWS_POINT))))

  979.         c->swScale= rgbToRgbWrapper;

  980. 

  981. #define isByteRGB(f) (\

  982.         f == PIX_FMT_RGB32   ||\

  983.         f == PIX_FMT_RGB32_1 ||\

  984.         f == PIX_FMT_RGB24   ||\

  985.         f == PIX_FMT_BGR32   ||\

  986.         f == PIX_FMT_BGR32_1 ||\

  987.         f == PIX_FMT_BGR24)

  988. 

  989.     if (isAnyRGB(srcFormat) && isPlanar(srcFormat) && isByteRGB(dstFormat))

  990.         c->swScale = planarRgbToRgbWrapper;

  991. 

  992.     /* bswap 16 bits per pixel/component packed formats */

  993.     if (IS_DIFFERENT_ENDIANESS(srcFormat, dstFormat, PIX_FMT_BGR444) ||

  994.         IS_DIFFERENT_ENDIANESS(srcFormat, dstFormat, PIX_FMT_BGR48)  ||

  995.         IS_DIFFERENT_ENDIANESS(srcFormat, dstFormat, PIX_FMT_BGRA64) ||

  996.         IS_DIFFERENT_ENDIANESS(srcFormat, dstFormat, PIX_FMT_BGR555) ||

  997.         IS_DIFFERENT_ENDIANESS(srcFormat, dstFormat, PIX_FMT_BGR565) ||

  998.         IS_DIFFERENT_ENDIANESS(srcFormat, dstFormat, PIX_FMT_GRAY16) ||

  999.         IS_DIFFERENT_ENDIANESS(srcFormat, dstFormat, PIX_FMT_RGB444) ||

  1000.         IS_DIFFERENT_ENDIANESS(srcFormat, dstFormat, PIX_FMT_RGB48)  ||

  1001.         IS_DIFFERENT_ENDIANESS(srcFormat, dstFormat, PIX_FMT_RGBA64) ||

  1002.         IS_DIFFERENT_ENDIANESS(srcFormat, dstFormat, PIX_FMT_RGB555) ||

  1003.         IS_DIFFERENT_ENDIANESS(srcFormat, dstFormat, PIX_FMT_RGB565))

  1004.         c->swScale = packed_16bpc_bswap;

  1005. 

  1006.     if (usePal(srcFormat) && isByteRGB(dstFormat))

  1007.         c->swScale = palToRgbWrapper;

  1008. 

  1009.     if (srcFormat == PIX_FMT_YUV422P) {

  1010.         if (dstFormat == PIX_FMT_YUYV422)

  1011.             c->swScale = yuv422pToYuy2Wrapper;

  1012.         else if (dstFormat == PIX_FMT_UYVY422)

  1013.             c->swScale = yuv422pToUyvyWrapper;

  1014.     }

  1015. 

  1016.     /* LQ converters if -sws 0 or -sws 4*/

  1017.     if (c->flags&(SWS_FAST_BILINEAR|SWS_POINT)) {

  1018.         /* yv12_to_yuy2 */

  1019.         if (srcFormat == PIX_FMT_YUV420P || srcFormat == PIX_FMT_YUVA420P) {

  1020.             if (dstFormat == PIX_FMT_YUYV422)

  1021.                 c->swScale = planarToYuy2Wrapper;

  1022.             else if (dstFormat == PIX_FMT_UYVY422)

  1023.                 c->swScale = planarToUyvyWrapper;

  1024.         }

  1025.     }

  1026.     if (srcFormat == PIX_FMT_YUYV422 &&

  1027.        (dstFormat == PIX_FMT_YUV420P || dstFormat == PIX_FMT_YUVA420P))

  1028.         c->swScale = yuyvToYuv420Wrapper;

  1029.     if (srcFormat == PIX_FMT_UYVY422 &&

  1030.        (dstFormat == PIX_FMT_YUV420P || dstFormat == PIX_FMT_YUVA420P))

  1031.         c->swScale = uyvyToYuv420Wrapper;

  1032.     if (srcFormat == PIX_FMT_YUYV422 && dstFormat == PIX_FMT_YUV422P)

  1033.         c->swScale = yuyvToYuv422Wrapper;

  1034.     if (srcFormat == PIX_FMT_UYVY422 && dstFormat == PIX_FMT_YUV422P)

  1035.         c->swScale = uyvyToYuv422Wrapper;

  1036. 

  1037. #define isPlanarGray(x) (isGray(x) && (x) != PIX_FMT_GRAY8A)

  1038.     /* simple copy */

  1039.     if ( srcFormat == dstFormat ||

  1040.         (srcFormat == PIX_FMT_YUVA420P && dstFormat == PIX_FMT_YUV420P) ||

  1041.         (srcFormat == PIX_FMT_YUV420P && dstFormat == PIX_FMT_YUVA420P) ||

  1042.         (isPlanarYUV(srcFormat) && isPlanarGray(dstFormat)) ||

  1043.         (isPlanarYUV(dstFormat) && isPlanarGray(srcFormat)) ||

  1044.         (isPlanarGray(dstFormat) && isPlanarGray(srcFormat)) ||

  1045.         (isPlanarYUV(srcFormat) && isPlanarYUV(dstFormat) &&

  1046.          c->chrDstHSubSample == c->chrSrcHSubSample &&

  1047.          c->chrDstVSubSample == c->chrSrcVSubSample &&

  1048.          dstFormat != PIX_FMT_NV12 && dstFormat != PIX_FMT_NV21 &&

  1049.          srcFormat != PIX_FMT_NV12 && srcFormat != PIX_FMT_NV21))

  1050.     {

  1051.         if (isPacked(c->srcFormat))

  1052.             c->swScale = packedCopyWrapper;

  1053.         else /* Planar YUV or gray */

  1054.             c->swScale = planarCopyWrapper;

  1055.     }

  1056. 

  1057.     if (ARCH_BFIN)

  1058.         ff_bfin_get_unscaled_swscale(c);

  1059.     if (HAVE_ALTIVEC)

  1060.         ff_swscale_get_unscaled_altivec(c);

  1061. }

  1062. 

  1063. static void reset_ptr(const uint8_t *src[], int format)

  1064. {

  1065.     if (!isALPHA(format))

  1066.         src[3] = NULL;

  1067.     if (!isPlanar(format)) {

  1068.         src[3] = src[2] = NULL;

  1069. 

  1070.         if (!usePal(format))

  1071.             src[1] = NULL;

  1072.     }

  1073. }

  1074. 

  1075. static int check_image_pointers(const uint8_t * const data[4], enum PixelFormat pix_fmt,

  1076.                                 const int linesizes[4])

  1077. {

  1078.     const AVPixFmtDescriptor *desc = &av_pix_fmt_descriptors[pix_fmt];

  1079.     int i;

  1080. 

  1081.     for (i = 0; i < 4; i++) {

  1082.         int plane = desc->comp[i].plane;

  1083.         if (!data[plane] || !linesizes[plane])

  1084.             return 0;

  1085.     }

  1086. 

  1087.     return 1;

  1088. }

  1089. 

  1090. /**

  1091.  * swscale wrapper, so we don't need to export the SwsContext.

  1092.  * Assumes planar YUV to be in YUV order instead of YVU.

  1093.  */

  1094. int attribute_align_arg sws_scale(struct SwsContext *c,

  1095.                                   const uint8_t * const srcSlice[],

  1096.                                   const int srcStride[], int srcSliceY,

  1097.                                   int srcSliceH, uint8_t *const dst[],

  1098.                                   const int dstStride[])

  1099. {

  1100.     int i, ret;

  1101.     const uint8_t *src2[4] = { srcSlice[0], srcSlice[1], srcSlice[2], srcSlice[3] };

  1102.     uint8_t *dst2[4] = { dst[0], dst[1], dst[2], dst[3] };

  1103.     uint8_t *rgb0_tmp = NULL;

  1104. 

  1105.     // do not mess up sliceDir if we have a "trailing" 0-size slice

  1106.     if (srcSliceH == 0)

  1107.         return 0;

  1108. 

  1109.     if (!check_image_pointers(srcSlice, c->srcFormat, srcStride)) {

  1110.         av_log(c, AV_LOG_ERROR, "bad src image pointers\n");

  1111.         return 0;

  1112.     }

  1113.     if (!check_image_pointers((const uint8_t* const*)dst, c->dstFormat, dstStride)) {

  1114.         av_log(c, AV_LOG_ERROR, "bad dst image pointers\n");

  1115.         return 0;

  1116.     }

  1117. 

  1118.     if (c->sliceDir == 0 && srcSliceY != 0 && srcSliceY + srcSliceH != c->srcH) {

  1119.         av_log(c, AV_LOG_ERROR, "Slices start in the middle!\n");

  1120.         return 0;

  1121.     }

  1122.     if (c->sliceDir == 0) {

  1123.         if (srcSliceY == 0) c->sliceDir = 1; else c->sliceDir = -1;

  1124.     }

  1125. 

  1126.     if (usePal(c->srcFormat)) {

  1127.         for (i = 0; i < 256; i++) {

  1128.             int p, r, g, b, y, u, v, a = 0xff;

  1129.             if (c->srcFormat == PIX_FMT_PAL8) {

  1130.                 p = ((const uint32_t *)(srcSlice[1]))[i];

  1131.                 a = (p >> 24) & 0xFF;

  1132.                 r = (p >> 16) & 0xFF;

  1133.                 g = (p >>  8) & 0xFF;

  1134.                 b =  p        & 0xFF;

  1135.             } else if (c->srcFormat == PIX_FMT_RGB8) {

  1136.                 r = ( i >> 5     ) * 36;

  1137.                 g = ((i >> 2) & 7) * 36;

  1138.                 b = ( i       & 3) * 85;

  1139.             } else if (c->srcFormat == PIX_FMT_BGR8) {

  1140.                 b = ( i >> 6     ) * 85;

  1141.                 g = ((i >> 3) & 7) * 36;

  1142.                 r = ( i       & 7) * 36;

  1143.             } else if (c->srcFormat == PIX_FMT_RGB4_BYTE) {

  1144.                 r = ( i >> 3     ) * 255;

  1145.                 g = ((i >> 1) & 3) * 85;

  1146.                 b = ( i       & 1) * 255;

  1147.             } else if (c->srcFormat == PIX_FMT_GRAY8 || c->srcFormat == PIX_FMT_GRAY8A) {

  1148.                 r = g = b = i;

  1149.             } else {

  1150.                 av_assert1(c->srcFormat == PIX_FMT_BGR4_BYTE);

  1151.                 b = ( i >> 3     ) * 255;

  1152.                 g = ((i >> 1) & 3) * 85;

  1153.                 r = ( i       & 1) * 255;

  1154.             }

  1155.             y = av_clip_uint8((RY * r + GY * g + BY * b + ( 33 << (RGB2YUV_SHIFT - 1))) >> RGB2YUV_SHIFT);

  1156.             u = av_clip_uint8((RU * r + GU * g + BU * b + (257 << (RGB2YUV_SHIFT - 1))) >> RGB2YUV_SHIFT);

  1157.             v = av_clip_uint8((RV * r + GV * g + BV * b + (257 << (RGB2YUV_SHIFT - 1))) >> RGB2YUV_SHIFT);

  1158.             c->pal_yuv[i]= y + (u<<8) + (v<<16) + (a<<24);

  1159. 

  1160.             switch (c->dstFormat) {

  1161.             case PIX_FMT_BGR32:

  1162. #if !HAVE_BIGENDIAN

  1163.             case PIX_FMT_RGB24:

  1164. #endif

  1165.                 c->pal_rgb[i]=  r + (g<<8) + (b<<16) + (a<<24);

  1166.                 break;

  1167.             case PIX_FMT_BGR32_1:

  1168. #if HAVE_BIGENDIAN

  1169.             case PIX_FMT_BGR24:

  1170. #endif

  1171.                 c->pal_rgb[i]= a + (r<<8) + (g<<16) + (b<<24);

  1172.                 break;

  1173.             case PIX_FMT_RGB32_1:

  1174. #if HAVE_BIGENDIAN

  1175.             case PIX_FMT_RGB24:

  1176. #endif

  1177.                 c->pal_rgb[i]= a + (b<<8) + (g<<16) + (r<<24);

  1178.                 break;

  1179.             case PIX_FMT_RGB32:

  1180. #if !HAVE_BIGENDIAN

  1181.             case PIX_FMT_BGR24:

  1182. #endif

  1183.             default:

  1184.                 c->pal_rgb[i]=  b + (g<<8) + (r<<16) + (a<<24);

  1185.             }

  1186.         }

  1187.     }

  1188. 

  1189.     if (c->src0Alpha && !c->dst0Alpha && isALPHA(c->dstFormat)) {

  1190.         uint8_t *base;

  1191.         int x,y;

  1192.         rgb0_tmp = av_malloc(FFABS(srcStride[0]) * srcSliceH + 32);

  1193.         base = srcStride[0] < 0 ? rgb0_tmp - srcStride[0] * (srcSliceH-1) : rgb0_tmp;

  1194.         for (y=0; y<srcSliceH; y++){

  1195.             memcpy(base + srcStride[0]*y, src2[0] + srcStride[0]*y, 4*c->srcW);

  1196.             for (x=c->src0Alpha-1; x<4*c->srcW; x+=4) {

  1197.                 base[ srcStride[0]*y + x] = 0xFF;

  1198.             }

  1199.         }

  1200.         src2[0] = base;

  1201.     }

  1202. 

  1203.     // copy strides, so they can safely be modified

  1204.     if (c->sliceDir == 1) {

  1205.         // slices go from top to bottom

  1206.         int srcStride2[4] = { srcStride[0], srcStride[1], srcStride[2],

  1207.                               srcStride[3] };

  1208.         int dstStride2[4] = { dstStride[0], dstStride[1], dstStride[2],

  1209.                               dstStride[3] };

  1210. 

  1211.         reset_ptr(src2, c->srcFormat);

  1212.         reset_ptr((void*)dst2, c->dstFormat);

  1213. 

  1214.         /* reset slice direction at end of frame */

  1215.         if (srcSliceY + srcSliceH == c->srcH)

  1216.             c->sliceDir = 0;

  1217. 

  1218.         ret = c->swScale(c, src2, srcStride2, srcSliceY, srcSliceH, dst2,

  1219.                           dstStride2);

  1220.     } else {

  1221.         // slices go from bottom to top => we flip the image internally

  1222.         int srcStride2[4] = { -srcStride[0], -srcStride[1], -srcStride[2],

  1223.                               -srcStride[3] };

  1224.         int dstStride2[4] = { -dstStride[0], -dstStride[1], -dstStride[2],

  1225.                               -dstStride[3] };

  1226. 

  1227.         src2[0] += (srcSliceH - 1) * srcStride[0];

  1228.         if (!usePal(c->srcFormat))

  1229.             src2[1] += ((srcSliceH >> c->chrSrcVSubSample) - 1) * srcStride[1];

  1230.         src2[2] += ((srcSliceH >> c->chrSrcVSubSample) - 1) * srcStride[2];

  1231.         src2[3] += (srcSliceH - 1) * srcStride[3];

  1232.         dst2[0] += ( c->dstH                         - 1) * dstStride[0];

  1233.         dst2[1] += ((c->dstH >> c->chrDstVSubSample) - 1) * dstStride[1];

  1234.         dst2[2] += ((c->dstH >> c->chrDstVSubSample) - 1) * dstStride[2];

  1235.         dst2[3] += ( c->dstH                         - 1) * dstStride[3];

  1236. 

  1237.         reset_ptr(src2, c->srcFormat);

  1238.         reset_ptr((void*)dst2, c->dstFormat);

  1239. 

  1240.         /* reset slice direction at end of frame */

  1241.         if (!srcSliceY)

  1242.             c->sliceDir = 0;

  1243. 

  1244.         ret = c->swScale(c, src2, srcStride2, c->srcH-srcSliceY-srcSliceH,

  1245.                           srcSliceH, dst2, dstStride2);

  1246.     }

  1247. 

  1248.     av_free(rgb0_tmp);

  1249.     return ret;

  1250. }

  1251. 

  1252. /* Convert the palette to the same packed 32-bit format as the palette */

  1253. void sws_convertPalette8ToPacked32(const uint8_t *src, uint8_t *dst,

  1254.                                    int num_pixels, const uint8_t *palette)

  1255. {

  1256.     int i;

  1257. 

  1258.     for (i = 0; i < num_pixels; i++)

  1259.         ((uint32_t *) dst)[i] = ((const uint32_t *) palette)[src[i]];

  1260. }

  1261. 

  1262. /* Palette format: ABCD -> dst format: ABC */

  1263. void sws_convertPalette8ToPacked24(const uint8_t *src, uint8_t *dst,

  1264.                                    int num_pixels, const uint8_t *palette)

  1265. {

  1266.     int i;

  1267. 

  1268.     for (i = 0; i < num_pixels; i++) {

  1269.         //FIXME slow?

  1270.         dst[0] = palette[src[i] * 4 + 0];

  1271.         dst[1] = palette[src[i] * 4 + 1];

  1272.         dst[2] = palette[src[i] * 4 + 2];

  1273.         dst += 3;

  1274.     }

  1275. }