swscale_unscaled.c
Go to the documentation of this file.
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 <inttypes.h>
22 #include <string.h>
23 #include <math.h>
24 #include <stdio.h>
25 #include "config.h"
26 #include <assert.h>
27 #include "swscale.h"
28 #include "swscale_internal.h"
29 #include "rgb2rgb.h"
30 #include "libavutil/intreadwrite.h"
31 #include "libavutil/cpu.h"
32 #include "libavutil/avutil.h"
33 #include "libavutil/mathematics.h"
34 #include "libavutil/bswap.h"
35 #include "libavutil/pixdesc.h"
36 
37 DECLARE_ALIGNED(8, const uint8_t, dither_8x8_1)[8][8] = {
38  { 0, 1, 0, 1, 0, 1, 0, 1,},
39  { 1, 0, 1, 0, 1, 0, 1, 0,},
40  { 0, 1, 0, 1, 0, 1, 0, 1,},
41  { 1, 0, 1, 0, 1, 0, 1, 0,},
42  { 0, 1, 0, 1, 0, 1, 0, 1,},
43  { 1, 0, 1, 0, 1, 0, 1, 0,},
44  { 0, 1, 0, 1, 0, 1, 0, 1,},
45  { 1, 0, 1, 0, 1, 0, 1, 0,},
46 };
47 DECLARE_ALIGNED(8, const uint8_t, dither_8x8_3)[8][8] = {
48  { 1, 2, 1, 2, 1, 2, 1, 2,},
49  { 3, 0, 3, 0, 3, 0, 3, 0,},
50  { 1, 2, 1, 2, 1, 2, 1, 2,},
51  { 3, 0, 3, 0, 3, 0, 3, 0,},
52  { 1, 2, 1, 2, 1, 2, 1, 2,},
53  { 3, 0, 3, 0, 3, 0, 3, 0,},
54  { 1, 2, 1, 2, 1, 2, 1, 2,},
55  { 3, 0, 3, 0, 3, 0, 3, 0,},
56 };
57 DECLARE_ALIGNED(8, const uint8_t, dither_8x8_64)[8][8] = {
58  { 18, 34, 30, 46, 17, 33, 29, 45,},
59  { 50, 2, 62, 14, 49, 1, 61, 13,},
60  { 26, 42, 22, 38, 25, 41, 21, 37,},
61  { 58, 10, 54, 6, 57, 9, 53, 5,},
62  { 16, 32, 28, 44, 19, 35, 31, 47,},
63  { 48, 0, 60, 12, 51, 3, 63, 15,},
64  { 24, 40, 20, 36, 27, 43, 23, 39,},
65  { 56, 8, 52, 4, 59, 11, 55, 7,},
66 };
67 extern const uint8_t dither_8x8_128[8][8];
68 DECLARE_ALIGNED(8, const uint8_t, dither_8x8_256)[8][8] = {
69  { 72, 136, 120, 184, 68, 132, 116, 180,},
70  { 200, 8, 248, 56, 196, 4, 244, 52,},
71  { 104, 168, 88, 152, 100, 164, 84, 148,},
72  { 232, 40, 216, 24, 228, 36, 212, 20,},
73  { 64, 128, 102, 176, 76, 140, 124, 188,},
74  { 192, 0, 240, 48, 204, 12, 252, 60,},
75  { 96, 160, 80, 144, 108, 172, 92, 156,},
76  { 224, 32, 208, 16, 236, 44, 220, 28,},
77 };
78 
79 #define RGB2YUV_SHIFT 15
80 #define BY ( (int) (0.114 * 219 / 255 * (1 << RGB2YUV_SHIFT) + 0.5))
81 #define BV (-(int) (0.081 * 224 / 255 * (1 << RGB2YUV_SHIFT) + 0.5))
82 #define BU ( (int) (0.500 * 224 / 255 * (1 << RGB2YUV_SHIFT) + 0.5))
83 #define GY ( (int) (0.587 * 219 / 255 * (1 << RGB2YUV_SHIFT) + 0.5))
84 #define GV (-(int) (0.419 * 224 / 255 * (1 << RGB2YUV_SHIFT) + 0.5))
85 #define GU (-(int) (0.331 * 224 / 255 * (1 << RGB2YUV_SHIFT) + 0.5))
86 #define RY ( (int) (0.299 * 219 / 255 * (1 << RGB2YUV_SHIFT) + 0.5))
87 #define RV ( (int) (0.500 * 224 / 255 * (1 << RGB2YUV_SHIFT) + 0.5))
88 #define RU (-(int) (0.169 * 224 / 255 * (1 << RGB2YUV_SHIFT) + 0.5))
89 
90 static void fillPlane(uint8_t *plane, int stride, int width, int height, int y,
91  uint8_t val)
92 {
93  int i;
94  uint8_t *ptr = plane + stride * y;
95  for (i = 0; i < height; i++) {
96  memset(ptr, val, width);
97  ptr += stride;
98  }
99 }
100 
101 static void copyPlane(const uint8_t *src, int srcStride,
102  int srcSliceY, int srcSliceH, int width,
103  uint8_t *dst, int dstStride)
104 {
105  dst += dstStride * srcSliceY;
106  if (dstStride == srcStride && srcStride > 0) {
107  memcpy(dst, src, srcSliceH * dstStride);
108  } else {
109  int i;
110  for (i = 0; i < srcSliceH; i++) {
111  memcpy(dst, src, width);
112  src += srcStride;
113  dst += dstStride;
114  }
115  }
116 }
117 
118 static int planarToNv12Wrapper(SwsContext *c, const uint8_t *src[],
119  int srcStride[], int srcSliceY,
120  int srcSliceH, uint8_t *dstParam[],
121  int dstStride[])
122 {
123  uint8_t *dst = dstParam[1] + dstStride[1] * srcSliceY / 2;
124 
125  copyPlane(src[0], srcStride[0], srcSliceY, srcSliceH, c->srcW,
126  dstParam[0], dstStride[0]);
127 
128  if (c->dstFormat == PIX_FMT_NV12)
129  interleaveBytes(src[1], src[2], dst, c->srcW / 2, srcSliceH / 2,
130  srcStride[1], srcStride[2], dstStride[0]);
131  else
132  interleaveBytes(src[2], src[1], dst, c->srcW / 2, srcSliceH / 2,
133  srcStride[2], srcStride[1], dstStride[0]);
134 
135  return srcSliceH;
136 }
137 
138 static int planarToYuy2Wrapper(SwsContext *c, const uint8_t *src[],
139  int srcStride[], int srcSliceY, int srcSliceH,
140  uint8_t *dstParam[], int dstStride[])
141 {
142  uint8_t *dst = dstParam[0] + dstStride[0] * srcSliceY;
143 
144  yv12toyuy2(src[0], src[1], src[2], dst, c->srcW, srcSliceH, srcStride[0],
145  srcStride[1], dstStride[0]);
146 
147  return srcSliceH;
148 }
149 
150 static int planarToUyvyWrapper(SwsContext *c, const uint8_t *src[],
151  int srcStride[], int srcSliceY, int srcSliceH,
152  uint8_t *dstParam[], int dstStride[])
153 {
154  uint8_t *dst = dstParam[0] + dstStride[0] * srcSliceY;
155 
156  yv12touyvy(src[0], src[1], src[2], dst, c->srcW, srcSliceH, srcStride[0],
157  srcStride[1], dstStride[0]);
158 
159  return srcSliceH;
160 }
161 
162 static int yuv422pToYuy2Wrapper(SwsContext *c, const uint8_t *src[],
163  int srcStride[], int srcSliceY, int srcSliceH,
164  uint8_t *dstParam[], int dstStride[])
165 {
166  uint8_t *dst = dstParam[0] + dstStride[0] * srcSliceY;
167 
168  yuv422ptoyuy2(src[0], src[1], src[2], dst, c->srcW, srcSliceH, srcStride[0],
169  srcStride[1], dstStride[0]);
170 
171  return srcSliceH;
172 }
173 
174 static int yuv422pToUyvyWrapper(SwsContext *c, const uint8_t *src[],
175  int srcStride[], int srcSliceY, int srcSliceH,
176  uint8_t *dstParam[], int dstStride[])
177 {
178  uint8_t *dst = dstParam[0] + dstStride[0] * srcSliceY;
179 
180  yuv422ptouyvy(src[0], src[1], src[2], dst, c->srcW, srcSliceH, srcStride[0],
181  srcStride[1], dstStride[0]);
182 
183  return srcSliceH;
184 }
185 
186 static int yuyvToYuv420Wrapper(SwsContext *c, const uint8_t *src[],
187  int srcStride[], int srcSliceY, int srcSliceH,
188  uint8_t *dstParam[], int dstStride[])
189 {
190  uint8_t *ydst = dstParam[0] + dstStride[0] * srcSliceY;
191  uint8_t *udst = dstParam[1] + dstStride[1] * srcSliceY / 2;
192  uint8_t *vdst = dstParam[2] + dstStride[2] * srcSliceY / 2;
193 
194  yuyvtoyuv420(ydst, udst, vdst, src[0], c->srcW, srcSliceH, dstStride[0],
195  dstStride[1], srcStride[0]);
196 
197  if (dstParam[3])
198  fillPlane(dstParam[3], dstStride[3], c->srcW, srcSliceH, srcSliceY, 255);
199 
200  return srcSliceH;
201 }
202 
203 static int yuyvToYuv422Wrapper(SwsContext *c, const uint8_t *src[],
204  int srcStride[], int srcSliceY, int srcSliceH,
205  uint8_t *dstParam[], int dstStride[])
206 {
207  uint8_t *ydst = dstParam[0] + dstStride[0] * srcSliceY;
208  uint8_t *udst = dstParam[1] + dstStride[1] * srcSliceY;
209  uint8_t *vdst = dstParam[2] + dstStride[2] * srcSliceY;
210 
211  yuyvtoyuv422(ydst, udst, vdst, src[0], c->srcW, srcSliceH, dstStride[0],
212  dstStride[1], srcStride[0]);
213 
214  return srcSliceH;
215 }
216 
217 static int uyvyToYuv420Wrapper(SwsContext *c, const uint8_t *src[],
218  int srcStride[], int srcSliceY, int srcSliceH,
219  uint8_t *dstParam[], int dstStride[])
220 {
221  uint8_t *ydst = dstParam[0] + dstStride[0] * srcSliceY;
222  uint8_t *udst = dstParam[1] + dstStride[1] * srcSliceY / 2;
223  uint8_t *vdst = dstParam[2] + dstStride[2] * srcSliceY / 2;
224 
225  uyvytoyuv420(ydst, udst, vdst, src[0], c->srcW, srcSliceH, dstStride[0],
226  dstStride[1], srcStride[0]);
227 
228  if (dstParam[3])
229  fillPlane(dstParam[3], dstStride[3], c->srcW, srcSliceH, srcSliceY, 255);
230 
231  return srcSliceH;
232 }
233 
234 static int uyvyToYuv422Wrapper(SwsContext *c, const uint8_t *src[],
235  int srcStride[], int srcSliceY, int srcSliceH,
236  uint8_t *dstParam[], int dstStride[])
237 {
238  uint8_t *ydst = dstParam[0] + dstStride[0] * srcSliceY;
239  uint8_t *udst = dstParam[1] + dstStride[1] * srcSliceY;
240  uint8_t *vdst = dstParam[2] + dstStride[2] * srcSliceY;
241 
242  uyvytoyuv422(ydst, udst, vdst, src[0], c->srcW, srcSliceH, dstStride[0],
243  dstStride[1], srcStride[0]);
244 
245  return srcSliceH;
246 }
247 
248 static void gray8aToPacked32(const uint8_t *src, uint8_t *dst, int num_pixels,
249  const uint8_t *palette)
250 {
251  int i;
252  for (i = 0; i < num_pixels; i++)
253  ((uint32_t *) dst)[i] = ((const uint32_t *) palette)[src[i << 1]] | (src[(i << 1) + 1] << 24);
254 }
255 
256 static void gray8aToPacked32_1(const uint8_t *src, uint8_t *dst, int num_pixels,
257  const uint8_t *palette)
258 {
259  int i;
260 
261  for (i = 0; i < num_pixels; i++)
262  ((uint32_t *) dst)[i] = ((const uint32_t *) palette)[src[i << 1]] | src[(i << 1) + 1];
263 }
264 
265 static void gray8aToPacked24(const uint8_t *src, uint8_t *dst, int num_pixels,
266  const uint8_t *palette)
267 {
268  int i;
269 
270  for (i = 0; i < num_pixels; i++) {
271  //FIXME slow?
272  dst[0] = palette[src[i << 1] * 4 + 0];
273  dst[1] = palette[src[i << 1] * 4 + 1];
274  dst[2] = palette[src[i << 1] * 4 + 2];
275  dst += 3;
276  }
277 }
278 
279 static int packed_16bpc_bswap(SwsContext *c, const uint8_t *src[],
280  int srcStride[], int srcSliceY, int srcSliceH,
281  uint8_t *dst[], int dstStride[])
282 {
283  int i, j;
284  int srcstr = srcStride[0] >> 1;
285  int dststr = dstStride[0] >> 1;
286  uint16_t *dstPtr = (uint16_t *) dst[0];
287  const uint16_t *srcPtr = (const uint16_t *) src[0];
288  int min_stride = FFMIN(srcstr, dststr);
289 
290  for (i = 0; i < srcSliceH; i++) {
291  for (j = 0; j < min_stride; j++) {
292  dstPtr[j] = av_bswap16(srcPtr[j]);
293  }
294  srcPtr += srcstr;
295  dstPtr += dststr;
296  }
297 
298  return srcSliceH;
299 }
300 
301 static int palToRgbWrapper(SwsContext *c, const uint8_t *src[], int srcStride[],
302  int srcSliceY, int srcSliceH, uint8_t *dst[],
303  int dstStride[])
304 {
305  const enum PixelFormat srcFormat = c->srcFormat;
306  const enum PixelFormat dstFormat = c->dstFormat;
307  void (*conv)(const uint8_t *src, uint8_t *dst, int num_pixels,
308  const uint8_t *palette) = NULL;
309  int i;
310  uint8_t *dstPtr = dst[0] + dstStride[0] * srcSliceY;
311  const uint8_t *srcPtr = src[0];
312 
313  if (srcFormat == PIX_FMT_Y400A) {
314  switch (dstFormat) {
315  case PIX_FMT_RGB32 : conv = gray8aToPacked32; break;
316  case PIX_FMT_BGR32 : conv = gray8aToPacked32; break;
317  case PIX_FMT_BGR32_1: conv = gray8aToPacked32_1; break;
318  case PIX_FMT_RGB32_1: conv = gray8aToPacked32_1; break;
319  case PIX_FMT_RGB24 : conv = gray8aToPacked24; break;
320  case PIX_FMT_BGR24 : conv = gray8aToPacked24; break;
321  }
322  } else if (usePal(srcFormat)) {
323  switch (dstFormat) {
324  case PIX_FMT_RGB32 : conv = sws_convertPalette8ToPacked32; break;
325  case PIX_FMT_BGR32 : conv = sws_convertPalette8ToPacked32; break;
328  case PIX_FMT_RGB24 : conv = sws_convertPalette8ToPacked24; break;
329  case PIX_FMT_BGR24 : conv = sws_convertPalette8ToPacked24; break;
330  }
331  }
332 
333  if (!conv)
334  av_log(c, AV_LOG_ERROR, "internal error %s -> %s converter\n",
335  sws_format_name(srcFormat), sws_format_name(dstFormat));
336  else {
337  for (i = 0; i < srcSliceH; i++) {
338  conv(srcPtr, dstPtr, c->srcW, (uint8_t *) c->pal_rgb);
339  srcPtr += srcStride[0];
340  dstPtr += dstStride[0];
341  }
342  }
343 
344  return srcSliceH;
345 }
346 
347 #define isRGBA32(x) ( \
348  (x) == PIX_FMT_ARGB \
349  || (x) == PIX_FMT_RGBA \
350  || (x) == PIX_FMT_BGRA \
351  || (x) == PIX_FMT_ABGR \
352  )
353 
354 /* {RGB,BGR}{15,16,24,32,32_1} -> {RGB,BGR}{15,16,24,32} */
355 typedef void (* rgbConvFn) (const uint8_t *, uint8_t *, int);
357 {
358  const enum PixelFormat srcFormat = c->srcFormat;
359  const enum PixelFormat dstFormat = c->dstFormat;
360  const int srcId = c->srcFormatBpp;
361  const int dstId = c->dstFormatBpp;
362  rgbConvFn conv = NULL;
363 
364 #define IS_NOT_NE(bpp, fmt) \
365  (((bpp + 7) >> 3) == 2 && \
366  (!(av_pix_fmt_descriptors[fmt].flags & PIX_FMT_BE) != !HAVE_BIGENDIAN))
367 
368  /* if this is non-native rgb444/555/565, don't handle it here. */
369  if (IS_NOT_NE(srcId, srcFormat) || IS_NOT_NE(dstId, dstFormat))
370  return NULL;
371 
372 #define CONV_IS(src, dst) (srcFormat == PIX_FMT_##src && dstFormat == PIX_FMT_##dst)
373 
374  if (isRGBA32(srcFormat) && isRGBA32(dstFormat)) {
375  if ( CONV_IS(ABGR, RGBA)
376  || CONV_IS(ARGB, BGRA)
377  || CONV_IS(BGRA, ARGB)
378  || CONV_IS(RGBA, ABGR)) conv = shuffle_bytes_3210;
379  else if (CONV_IS(ABGR, ARGB)
380  || CONV_IS(ARGB, ABGR)) conv = shuffle_bytes_0321;
381  else if (CONV_IS(ABGR, BGRA)
382  || CONV_IS(ARGB, RGBA)) conv = shuffle_bytes_1230;
383  else if (CONV_IS(BGRA, RGBA)
384  || CONV_IS(RGBA, BGRA)) conv = shuffle_bytes_2103;
385  else if (CONV_IS(BGRA, ABGR)
386  || CONV_IS(RGBA, ARGB)) conv = shuffle_bytes_3012;
387  } else
388  /* BGR -> BGR */
389  if ((isBGRinInt(srcFormat) && isBGRinInt(dstFormat)) ||
390  (isRGBinInt(srcFormat) && isRGBinInt(dstFormat))) {
391  switch (srcId | (dstId << 16)) {
392  case 0x000F000C: conv = rgb12to15; break;
393  case 0x000F0010: conv = rgb16to15; break;
394  case 0x000F0018: conv = rgb24to15; break;
395  case 0x000F0020: conv = rgb32to15; break;
396  case 0x0010000F: conv = rgb15to16; break;
397  case 0x00100018: conv = rgb24to16; break;
398  case 0x00100020: conv = rgb32to16; break;
399  case 0x0018000F: conv = rgb15to24; break;
400  case 0x00180010: conv = rgb16to24; break;
401  case 0x00180020: conv = rgb32to24; break;
402  case 0x0020000F: conv = rgb15to32; break;
403  case 0x00200010: conv = rgb16to32; break;
404  case 0x00200018: conv = rgb24to32; break;
405  }
406  } else if ((isBGRinInt(srcFormat) && isRGBinInt(dstFormat)) ||
407  (isRGBinInt(srcFormat) && isBGRinInt(dstFormat))) {
408  switch (srcId | (dstId << 16)) {
409  case 0x000C000C: conv = rgb12tobgr12; break;
410  case 0x000F000F: conv = rgb15tobgr15; break;
411  case 0x000F0010: conv = rgb16tobgr15; break;
412  case 0x000F0018: conv = rgb24tobgr15; break;
413  case 0x000F0020: conv = rgb32tobgr15; break;
414  case 0x0010000F: conv = rgb15tobgr16; break;
415  case 0x00100010: conv = rgb16tobgr16; break;
416  case 0x00100018: conv = rgb24tobgr16; break;
417  case 0x00100020: conv = rgb32tobgr16; break;
418  case 0x0018000F: conv = rgb15tobgr24; break;
419  case 0x00180010: conv = rgb16tobgr24; break;
420  case 0x00180018: conv = rgb24tobgr24; break;
421  case 0x00180020: conv = rgb32tobgr24; break;
422  case 0x0020000F: conv = rgb15tobgr32; break;
423  case 0x00200010: conv = rgb16tobgr32; break;
424  case 0x00200018: conv = rgb24tobgr32; break;
425  }
426  }
427 
428  return conv;
429 }
430 
431 /* {RGB,BGR}{15,16,24,32,32_1} -> {RGB,BGR}{15,16,24,32} */
432 static int rgbToRgbWrapper(SwsContext *c, const uint8_t *src[], int srcStride[],
433  int srcSliceY, int srcSliceH, uint8_t *dst[],
434  int dstStride[])
435 
436 {
437  const enum PixelFormat srcFormat = c->srcFormat;
438  const enum PixelFormat dstFormat = c->dstFormat;
439  const int srcBpp = (c->srcFormatBpp + 7) >> 3;
440  const int dstBpp = (c->dstFormatBpp + 7) >> 3;
441  rgbConvFn conv = findRgbConvFn(c);
442 
443  if (!conv) {
444  av_log(c, AV_LOG_ERROR, "internal error %s -> %s converter\n",
445  sws_format_name(srcFormat), sws_format_name(dstFormat));
446  } else {
447  const uint8_t *srcPtr = src[0];
448  uint8_t *dstPtr = dst[0];
449  if ((srcFormat == PIX_FMT_RGB32_1 || srcFormat == PIX_FMT_BGR32_1) &&
450  !isRGBA32(dstFormat))
451  srcPtr += ALT32_CORR;
452 
453  if ((dstFormat == PIX_FMT_RGB32_1 || dstFormat == PIX_FMT_BGR32_1) &&
454  !isRGBA32(srcFormat))
455  dstPtr += ALT32_CORR;
456 
457  if (dstStride[0] * srcBpp == srcStride[0] * dstBpp && srcStride[0] > 0 &&
458  !(srcStride[0] % srcBpp))
459  conv(srcPtr, dstPtr + dstStride[0] * srcSliceY,
460  srcSliceH * srcStride[0]);
461  else {
462  int i;
463  dstPtr += dstStride[0] * srcSliceY;
464 
465  for (i = 0; i < srcSliceH; i++) {
466  conv(srcPtr, dstPtr, c->srcW * srcBpp);
467  srcPtr += srcStride[0];
468  dstPtr += dstStride[0];
469  }
470  }
471  }
472  return srcSliceH;
473 }
474 
475 static int bgr24ToYv12Wrapper(SwsContext *c, const uint8_t *src[],
476  int srcStride[], int srcSliceY, int srcSliceH,
477  uint8_t *dst[], int dstStride[])
478 {
479  rgb24toyv12(
480  src[0],
481  dst[0] + srcSliceY * dstStride[0],
482  dst[1] + (srcSliceY >> 1) * dstStride[1],
483  dst[2] + (srcSliceY >> 1) * dstStride[2],
484  c->srcW, srcSliceH,
485  dstStride[0], dstStride[1], srcStride[0]);
486  if (dst[3])
487  fillPlane(dst[3], dstStride[3], c->srcW, srcSliceH, srcSliceY, 255);
488  return srcSliceH;
489 }
490 
491 static int yvu9ToYv12Wrapper(SwsContext *c, const uint8_t *src[],
492  int srcStride[], int srcSliceY, int srcSliceH,
493  uint8_t *dst[], int dstStride[])
494 {
495  copyPlane(src[0], srcStride[0], srcSliceY, srcSliceH, c->srcW,
496  dst[0], dstStride[0]);
497 
498  planar2x(src[1], dst[1] + dstStride[1] * (srcSliceY >> 1), c->chrSrcW,
499  srcSliceH >> 2, srcStride[1], dstStride[1]);
500  planar2x(src[2], dst[2] + dstStride[2] * (srcSliceY >> 1), c->chrSrcW,
501  srcSliceH >> 2, srcStride[2], dstStride[2]);
502  if (dst[3])
503  fillPlane(dst[3], dstStride[3], c->srcW, srcSliceH, srcSliceY, 255);
504  return srcSliceH;
505 }
506 
507 /* unscaled copy like stuff (assumes nearly identical formats) */
508 static int packedCopyWrapper(SwsContext *c, const uint8_t *src[],
509  int srcStride[], int srcSliceY, int srcSliceH,
510  uint8_t *dst[], int dstStride[])
511 {
512  if (dstStride[0] == srcStride[0] && srcStride[0] > 0)
513  memcpy(dst[0] + dstStride[0] * srcSliceY, src[0], srcSliceH * dstStride[0]);
514  else {
515  int i;
516  const uint8_t *srcPtr = src[0];
517  uint8_t *dstPtr = dst[0] + dstStride[0] * srcSliceY;
518  int length = 0;
519 
520  /* universal length finder */
521  while (length + c->srcW <= FFABS(dstStride[0]) &&
522  length + c->srcW <= FFABS(srcStride[0]))
523  length += c->srcW;
524  assert(length != 0);
525 
526  for (i = 0; i < srcSliceH; i++) {
527  memcpy(dstPtr, srcPtr, length);
528  srcPtr += srcStride[0];
529  dstPtr += dstStride[0];
530  }
531  }
532  return srcSliceH;
533 }
534 
535 #define clip9(x) av_clip_uintp2(x, 9)
536 #define clip10(x) av_clip_uintp2(x, 10)
537 #define DITHER_COPY(dst, dstStride, wfunc, src, srcStride, rfunc, dithers, shift, clip) \
538  for (i = 0; i < height; i++) { \
539  const uint8_t *dither = dithers[i & 7]; \
540  for (j = 0; j < length - 7; j += 8) { \
541  wfunc(&dst[j + 0], clip((rfunc(&src[j + 0]) + dither[0]) >> shift)); \
542  wfunc(&dst[j + 1], clip((rfunc(&src[j + 1]) + dither[1]) >> shift)); \
543  wfunc(&dst[j + 2], clip((rfunc(&src[j + 2]) + dither[2]) >> shift)); \
544  wfunc(&dst[j + 3], clip((rfunc(&src[j + 3]) + dither[3]) >> shift)); \
545  wfunc(&dst[j + 4], clip((rfunc(&src[j + 4]) + dither[4]) >> shift)); \
546  wfunc(&dst[j + 5], clip((rfunc(&src[j + 5]) + dither[5]) >> shift)); \
547  wfunc(&dst[j + 6], clip((rfunc(&src[j + 6]) + dither[6]) >> shift)); \
548  wfunc(&dst[j + 7], clip((rfunc(&src[j + 7]) + dither[7]) >> shift)); \
549  } \
550  for (; j < length; j++) \
551  wfunc(&dst[j], (rfunc(&src[j]) + dither[j & 7]) >> shift); \
552  dst += dstStride; \
553  src += srcStride; \
554  }
555 
556 static int planarCopyWrapper(SwsContext *c, const uint8_t *src[],
557  int srcStride[], int srcSliceY, int srcSliceH,
558  uint8_t *dst[], int dstStride[])
559 {
560  int plane, i, j;
561  for (plane = 0; plane < 4; plane++) {
562  int length = (plane == 0 || plane == 3) ? c->srcW : -((-c->srcW ) >> c->chrDstHSubSample);
563  int y = (plane == 0 || plane == 3) ? srcSliceY: -((-srcSliceY) >> c->chrDstVSubSample);
564  int height = (plane == 0 || plane == 3) ? srcSliceH: -((-srcSliceH) >> c->chrDstVSubSample);
565  const uint8_t *srcPtr = src[plane];
566  uint8_t *dstPtr = dst[plane] + dstStride[plane] * y;
567 
568  if (!dst[plane])
569  continue;
570  // ignore palette for GRAY8
571  if (plane == 1 && !dst[2]) continue;
572  if (!src[plane] || (plane == 1 && !src[2])) {
573  if (is16BPS(c->dstFormat))
574  length *= 2;
575  fillPlane(dst[plane], dstStride[plane], length, height, y,
576  (plane == 3) ? 255 : 128);
577  } else {
578  if (is9_OR_10BPS(c->srcFormat)) {
579  const int src_depth = av_pix_fmt_descriptors[c->srcFormat].comp[plane].depth_minus1 + 1;
580  const int dst_depth = av_pix_fmt_descriptors[c->dstFormat].comp[plane].depth_minus1 + 1;
581  const uint16_t *srcPtr2 = (const uint16_t *) srcPtr;
582 
583  if (is16BPS(c->dstFormat)) {
584  uint16_t *dstPtr2 = (uint16_t *) dstPtr;
585 #define COPY9_OR_10TO16(rfunc, wfunc) \
586  for (i = 0; i < height; i++) { \
587  for (j = 0; j < length; j++) { \
588  int srcpx = rfunc(&srcPtr2[j]); \
589  wfunc(&dstPtr2[j], (srcpx << (16 - src_depth)) | (srcpx >> (2 * src_depth - 16))); \
590  } \
591  dstPtr2 += dstStride[plane] / 2; \
592  srcPtr2 += srcStride[plane] / 2; \
593  }
594  if (isBE(c->dstFormat)) {
595  if (isBE(c->srcFormat)) {
597  } else {
599  }
600  } else {
601  if (isBE(c->srcFormat)) {
603  } else {
605  }
606  }
607  } else if (is9_OR_10BPS(c->dstFormat)) {
608  uint16_t *dstPtr2 = (uint16_t *) dstPtr;
609 #define COPY9_OR_10TO9_OR_10(loop) \
610  for (i = 0; i < height; i++) { \
611  for (j = 0; j < length; j++) { \
612  loop; \
613  } \
614  dstPtr2 += dstStride[plane] / 2; \
615  srcPtr2 += srcStride[plane] / 2; \
616  }
617 #define COPY9_OR_10TO9_OR_10_2(rfunc, wfunc) \
618  if (dst_depth > src_depth) { \
619  COPY9_OR_10TO9_OR_10(int srcpx = rfunc(&srcPtr2[j]); \
620  wfunc(&dstPtr2[j], (srcpx << 1) | (srcpx >> 9))); \
621  } else if (dst_depth < src_depth) { \
622  DITHER_COPY(dstPtr2, dstStride[plane] / 2, wfunc, \
623  srcPtr2, srcStride[plane] / 2, rfunc, \
624  dither_8x8_1, 1, clip9); \
625  } else { \
626  COPY9_OR_10TO9_OR_10(wfunc(&dstPtr2[j], rfunc(&srcPtr2[j]))); \
627  }
628  if (isBE(c->dstFormat)) {
629  if (isBE(c->srcFormat)) {
631  } else {
633  }
634  } else {
635  if (isBE(c->srcFormat)) {
637  } else {
639  }
640  }
641  } else {
642 #define W8(a, b) { *(a) = (b); }
643 #define COPY9_OR_10TO8(rfunc) \
644  if (src_depth == 9) { \
645  DITHER_COPY(dstPtr, dstStride[plane], W8, \
646  srcPtr2, srcStride[plane] / 2, rfunc, \
647  dither_8x8_1, 1, av_clip_uint8); \
648  } else { \
649  DITHER_COPY(dstPtr, dstStride[plane], W8, \
650  srcPtr2, srcStride[plane] / 2, rfunc, \
651  dither_8x8_3, 2, av_clip_uint8); \
652  }
653  if (isBE(c->srcFormat)) {
655  } else {
657  }
658  }
659  } else if (is9_OR_10BPS(c->dstFormat)) {
660  const int dst_depth = av_pix_fmt_descriptors[c->dstFormat].comp[plane].depth_minus1 + 1;
661  uint16_t *dstPtr2 = (uint16_t *) dstPtr;
662 
663  if (is16BPS(c->srcFormat)) {
664  const uint16_t *srcPtr2 = (const uint16_t *) srcPtr;
665 #define COPY16TO9_OR_10(rfunc, wfunc) \
666  if (dst_depth == 9) { \
667  DITHER_COPY(dstPtr2, dstStride[plane] / 2, wfunc, \
668  srcPtr2, srcStride[plane] / 2, rfunc, \
669  dither_8x8_128, 7, clip9); \
670  } else { \
671  DITHER_COPY(dstPtr2, dstStride[plane] / 2, wfunc, \
672  srcPtr2, srcStride[plane] / 2, rfunc, \
673  dither_8x8_64, 6, clip10); \
674  }
675  if (isBE(c->dstFormat)) {
676  if (isBE(c->srcFormat)) {
678  } else {
680  }
681  } else {
682  if (isBE(c->srcFormat)) {
684  } else {
686  }
687  }
688  } else /* 8bit */ {
689 #define COPY8TO9_OR_10(wfunc) \
690  for (i = 0; i < height; i++) { \
691  for (j = 0; j < length; j++) { \
692  const int srcpx = srcPtr[j]; \
693  wfunc(&dstPtr2[j], (srcpx << (dst_depth - 8)) | (srcpx >> (16 - dst_depth))); \
694  } \
695  dstPtr2 += dstStride[plane] / 2; \
696  srcPtr += srcStride[plane]; \
697  }
698  if (isBE(c->dstFormat)) {
700  } else {
702  }
703  }
704  } else if (is16BPS(c->srcFormat) && !is16BPS(c->dstFormat)) {
705  const uint16_t *srcPtr2 = (const uint16_t *) srcPtr;
706 #define COPY16TO8(rfunc) \
707  DITHER_COPY(dstPtr, dstStride[plane], W8, \
708  srcPtr2, srcStride[plane] / 2, rfunc, \
709  dither_8x8_256, 8, av_clip_uint8);
710  if (isBE(c->srcFormat)) {
712  } else {
714  }
715  } else if (!is16BPS(c->srcFormat) && is16BPS(c->dstFormat)) {
716  for (i = 0; i < height; i++) {
717  for (j = 0; j < length; j++) {
718  dstPtr[ j << 1 ] = srcPtr[j];
719  dstPtr[(j << 1) + 1] = srcPtr[j];
720  }
721  srcPtr += srcStride[plane];
722  dstPtr += dstStride[plane];
723  }
724  } else if (is16BPS(c->srcFormat) && is16BPS(c->dstFormat) &&
725  isBE(c->srcFormat) != isBE(c->dstFormat)) {
726 
727  for (i = 0; i < height; i++) {
728  for (j = 0; j < length; j++)
729  ((uint16_t *) dstPtr)[j] = av_bswap16(((const uint16_t *) srcPtr)[j]);
730  srcPtr += srcStride[plane];
731  dstPtr += dstStride[plane];
732  }
733  } else if (dstStride[plane] == srcStride[plane] &&
734  srcStride[plane] > 0 && srcStride[plane] == length) {
735  memcpy(dst[plane] + dstStride[plane] * y, src[plane],
736  height * dstStride[plane]);
737  } else {
738  if (is16BPS(c->srcFormat) && is16BPS(c->dstFormat))
739  length *= 2;
741  length >>= 3; // monowhite/black
742  for (i = 0; i < height; i++) {
743  memcpy(dstPtr, srcPtr, length);
744  srcPtr += srcStride[plane];
745  dstPtr += dstStride[plane];
746  }
747  }
748  }
749  }
750  return srcSliceH;
751 }
752 
753 
754 #define IS_DIFFERENT_ENDIANESS(src_fmt, dst_fmt, pix_fmt) \
755  ((src_fmt == pix_fmt ## BE && dst_fmt == pix_fmt ## LE) || \
756  (src_fmt == pix_fmt ## LE && dst_fmt == pix_fmt ## BE))
757 
758 
760 {
761  const enum PixelFormat srcFormat = c->srcFormat;
762  const enum PixelFormat dstFormat = c->dstFormat;
763  const int flags = c->flags;
764  const int dstH = c->dstH;
765  int needsDither;
766 
767  needsDither = isAnyRGB(dstFormat) &&
768  c->dstFormatBpp < 24 &&
769  (c->dstFormatBpp < c->srcFormatBpp || (!isAnyRGB(srcFormat)));
770 
771  /* yv12_to_nv12 */
772  if ((srcFormat == PIX_FMT_YUV420P || srcFormat == PIX_FMT_YUVA420P) &&
773  (dstFormat == PIX_FMT_NV12 || dstFormat == PIX_FMT_NV21)) {
775  }
776  /* yuv2bgr */
777  if ((srcFormat == PIX_FMT_YUV420P || srcFormat == PIX_FMT_YUV422P ||
778  srcFormat == PIX_FMT_YUVA420P) && isAnyRGB(dstFormat) &&
779  !(flags & SWS_ACCURATE_RND) && !(dstH & 1)) {
781  }
782 
783  if (srcFormat == PIX_FMT_YUV410P &&
784  (dstFormat == PIX_FMT_YUV420P || dstFormat == PIX_FMT_YUVA420P) &&
785  !(flags & SWS_BITEXACT)) {
787  }
788 
789  /* bgr24toYV12 */
790  if (srcFormat == PIX_FMT_BGR24 &&
791  (dstFormat == PIX_FMT_YUV420P || dstFormat == PIX_FMT_YUVA420P) &&
792  !(flags & SWS_ACCURATE_RND))
794 
795  /* RGB/BGR -> RGB/BGR (no dither needed forms) */
796  if (isAnyRGB(srcFormat) && isAnyRGB(dstFormat) && findRgbConvFn(c)
797  && (!needsDither || (c->flags&(SWS_FAST_BILINEAR|SWS_POINT))))
799 
800  /* bswap 16 bits per pixel/component packed formats */
801  if (IS_DIFFERENT_ENDIANESS(srcFormat, dstFormat, PIX_FMT_BGR444) ||
802  IS_DIFFERENT_ENDIANESS(srcFormat, dstFormat, PIX_FMT_BGR48) ||
803  IS_DIFFERENT_ENDIANESS(srcFormat, dstFormat, PIX_FMT_BGR555) ||
804  IS_DIFFERENT_ENDIANESS(srcFormat, dstFormat, PIX_FMT_BGR565) ||
805  IS_DIFFERENT_ENDIANESS(srcFormat, dstFormat, PIX_FMT_GRAY16) ||
806  IS_DIFFERENT_ENDIANESS(srcFormat, dstFormat, PIX_FMT_RGB444) ||
807  IS_DIFFERENT_ENDIANESS(srcFormat, dstFormat, PIX_FMT_RGB48) ||
808  IS_DIFFERENT_ENDIANESS(srcFormat, dstFormat, PIX_FMT_RGB555) ||
809  IS_DIFFERENT_ENDIANESS(srcFormat, dstFormat, PIX_FMT_RGB565))
811 
812  if ((usePal(srcFormat) && (
813  dstFormat == PIX_FMT_RGB32 ||
814  dstFormat == PIX_FMT_RGB32_1 ||
815  dstFormat == PIX_FMT_RGB24 ||
816  dstFormat == PIX_FMT_BGR32 ||
817  dstFormat == PIX_FMT_BGR32_1 ||
818  dstFormat == PIX_FMT_BGR24)))
820 
821  if (srcFormat == PIX_FMT_YUV422P) {
822  if (dstFormat == PIX_FMT_YUYV422)
824  else if (dstFormat == PIX_FMT_UYVY422)
826  }
827 
828  /* LQ converters if -sws 0 or -sws 4*/
829  if (c->flags&(SWS_FAST_BILINEAR|SWS_POINT)) {
830  /* yv12_to_yuy2 */
831  if (srcFormat == PIX_FMT_YUV420P || srcFormat == PIX_FMT_YUVA420P) {
832  if (dstFormat == PIX_FMT_YUYV422)
834  else if (dstFormat == PIX_FMT_UYVY422)
836  }
837  }
838  if (srcFormat == PIX_FMT_YUYV422 &&
839  (dstFormat == PIX_FMT_YUV420P || dstFormat == PIX_FMT_YUVA420P))
841  if (srcFormat == PIX_FMT_UYVY422 &&
842  (dstFormat == PIX_FMT_YUV420P || dstFormat == PIX_FMT_YUVA420P))
844  if (srcFormat == PIX_FMT_YUYV422 && dstFormat == PIX_FMT_YUV422P)
846  if (srcFormat == PIX_FMT_UYVY422 && dstFormat == PIX_FMT_YUV422P)
848 
849  /* simple copy */
850  if ( srcFormat == dstFormat ||
851  (srcFormat == PIX_FMT_YUVA420P && dstFormat == PIX_FMT_YUV420P) ||
852  (srcFormat == PIX_FMT_YUV420P && dstFormat == PIX_FMT_YUVA420P) ||
853  (isPlanarYUV(srcFormat) && isGray(dstFormat)) ||
854  (isPlanarYUV(dstFormat) && isGray(srcFormat)) ||
855  (isGray(dstFormat) && isGray(srcFormat)) ||
856  (isPlanarYUV(srcFormat) && isPlanarYUV(dstFormat) &&
859  dstFormat != PIX_FMT_NV12 && dstFormat != PIX_FMT_NV21 &&
860  srcFormat != PIX_FMT_NV12 && srcFormat != PIX_FMT_NV21))
861  {
862  if (isPacked(c->srcFormat))
864  else /* Planar YUV or gray */
866  }
867 
868  if (ARCH_BFIN)
870  if (HAVE_ALTIVEC)
872 }
873 
874 static void reset_ptr(const uint8_t *src[], int format)
875 {
876  if (!isALPHA(format))
877  src[3] = NULL;
878  if (!isPlanar(format)) {
879  src[3] = src[2] = NULL;
880 
881  if (!usePal(format))
882  src[1] = NULL;
883  }
884 }
885 
886 static int check_image_pointers(uint8_t *data[4], enum PixelFormat pix_fmt,
887  const int linesizes[4])
888 {
890  int i;
891 
892  for (i = 0; i < 4; i++) {
893  int plane = desc->comp[i].plane;
894  if (!data[plane] || !linesizes[plane])
895  return 0;
896  }
897 
898  return 1;
899 }
900 
906  const uint8_t * const srcSlice[],
907  const int srcStride[], int srcSliceY,
908  int srcSliceH, uint8_t *const dst[],
909  const int dstStride[])
910 {
911  int i;
912  const uint8_t *src2[4] = { srcSlice[0], srcSlice[1], srcSlice[2], srcSlice[3] };
913  uint8_t *dst2[4] = { dst[0], dst[1], dst[2], dst[3] };
914 
915  // do not mess up sliceDir if we have a "trailing" 0-size slice
916  if (srcSliceH == 0)
917  return 0;
918 
919  if (!check_image_pointers(srcSlice, c->srcFormat, srcStride)) {
920  av_log(c, AV_LOG_ERROR, "bad src image pointers\n");
921  return 0;
922  }
923  if (!check_image_pointers(dst, c->dstFormat, dstStride)) {
924  av_log(c, AV_LOG_ERROR, "bad dst image pointers\n");
925  return 0;
926  }
927 
928  if (c->sliceDir == 0 && srcSliceY != 0 && srcSliceY + srcSliceH != c->srcH) {
929  av_log(c, AV_LOG_ERROR, "Slices start in the middle!\n");
930  return 0;
931  }
932  if (c->sliceDir == 0) {
933  if (srcSliceY == 0) c->sliceDir = 1; else c->sliceDir = -1;
934  }
935 
936  if (usePal(c->srcFormat)) {
937  for (i = 0; i < 256; i++) {
938  int p, r, g, b, y, u, v;
939  if (c->srcFormat == PIX_FMT_PAL8) {
940  p = ((const uint32_t *)(srcSlice[1]))[i];
941  r = (p >> 16) & 0xFF;
942  g = (p >> 8) & 0xFF;
943  b = p & 0xFF;
944  } else if (c->srcFormat == PIX_FMT_RGB8) {
945  r = ( i >> 5 ) * 36;
946  g = ((i >> 2) & 7) * 36;
947  b = ( i & 3) * 85;
948  } else if (c->srcFormat == PIX_FMT_BGR8) {
949  b = ( i >> 6 ) * 85;
950  g = ((i >> 3) & 7) * 36;
951  r = ( i & 7) * 36;
952  } else if (c->srcFormat == PIX_FMT_RGB4_BYTE) {
953  r = ( i >> 3 ) * 255;
954  g = ((i >> 1) & 3) * 85;
955  b = ( i & 1) * 255;
956  } else if (c->srcFormat == PIX_FMT_GRAY8 ||
957  c->srcFormat == PIX_FMT_Y400A) {
958  r = g = b = i;
959  } else {
960  assert(c->srcFormat == PIX_FMT_BGR4_BYTE);
961  b = ( i >> 3 ) * 255;
962  g = ((i >> 1) & 3) * 85;
963  r = ( i & 1) * 255;
964  }
965  y = av_clip_uint8((RY * r + GY * g + BY * b + ( 33 << (RGB2YUV_SHIFT - 1))) >> RGB2YUV_SHIFT);
966  u = av_clip_uint8((RU * r + GU * g + BU * b + (257 << (RGB2YUV_SHIFT - 1))) >> RGB2YUV_SHIFT);
967  v = av_clip_uint8((RV * r + GV * g + BV * b + (257 << (RGB2YUV_SHIFT - 1))) >> RGB2YUV_SHIFT);
968  c->pal_yuv[i] = y + (u << 8) + (v << 16);
969 
970  switch (c->dstFormat) {
971  case PIX_FMT_BGR32:
972 #if !HAVE_BIGENDIAN
973  case PIX_FMT_RGB24:
974 #endif
975  c->pal_rgb[i] = r + (g << 8) + (b << 16);
976  break;
977  case PIX_FMT_BGR32_1:
978 #if HAVE_BIGENDIAN
979  case PIX_FMT_BGR24:
980 #endif
981  c->pal_rgb[i] = (r + (g << 8) + (b << 16)) << 8;
982  break;
983  case PIX_FMT_RGB32_1:
984 #if HAVE_BIGENDIAN
985  case PIX_FMT_RGB24:
986 #endif
987  c->pal_rgb[i] = (b + (g << 8) + (r << 16)) << 8;
988  break;
989  case PIX_FMT_RGB32:
990 #if !HAVE_BIGENDIAN
991  case PIX_FMT_BGR24:
992 #endif
993  default:
994  c->pal_rgb[i] = b + (g << 8) + (r << 16);
995  }
996  }
997  }
998 
999  // copy strides, so they can safely be modified
1000  if (c->sliceDir == 1) {
1001  // slices go from top to bottom
1002  int srcStride2[4] = { srcStride[0], srcStride[1], srcStride[2],
1003  srcStride[3] };
1004  int dstStride2[4] = { dstStride[0], dstStride[1], dstStride[2],
1005  dstStride[3] };
1006 
1007  reset_ptr(src2, c->srcFormat);
1008  reset_ptr((const uint8_t **) dst2, c->dstFormat);
1009 
1010  /* reset slice direction at end of frame */
1011  if (srcSliceY + srcSliceH == c->srcH)
1012  c->sliceDir = 0;
1013 
1014  return c->swScale(c, src2, srcStride2, srcSliceY, srcSliceH, dst2,
1015  dstStride2);
1016  } else {
1017  // slices go from bottom to top => we flip the image internally
1018  int srcStride2[4] = { -srcStride[0], -srcStride[1], -srcStride[2],
1019  -srcStride[3] };
1020  int dstStride2[4] = { -dstStride[0], -dstStride[1], -dstStride[2],
1021  -dstStride[3] };
1022 
1023  src2[0] += (srcSliceH - 1) * srcStride[0];
1024  if (!usePal(c->srcFormat))
1025  src2[1] += ((srcSliceH >> c->chrSrcVSubSample) - 1) * srcStride[1];
1026  src2[2] += ((srcSliceH >> c->chrSrcVSubSample) - 1) * srcStride[2];
1027  src2[3] += (srcSliceH - 1) * srcStride[3];
1028  dst2[0] += ( c->dstH - 1) * dstStride[0];
1029  dst2[1] += ((c->dstH >> c->chrDstVSubSample) - 1) * dstStride[1];
1030  dst2[2] += ((c->dstH >> c->chrDstVSubSample) - 1) * dstStride[2];
1031  dst2[3] += ( c->dstH - 1) * dstStride[3];
1032 
1033  reset_ptr(src2, c->srcFormat);
1034  reset_ptr((const uint8_t **) dst2, c->dstFormat);
1035 
1036  /* reset slice direction at end of frame */
1037  if (!srcSliceY)
1038  c->sliceDir = 0;
1039 
1040  return c->swScale(c, src2, srcStride2, c->srcH-srcSliceY-srcSliceH,
1041  srcSliceH, dst2, dstStride2);
1042  }
1043 }
1044 
1045 /* Convert the palette to the same packed 32-bit format as the palette */
1046 void sws_convertPalette8ToPacked32(const uint8_t *src, uint8_t *dst,
1047  int num_pixels, const uint8_t *palette)
1048 {
1049  int i;
1050 
1051  for (i = 0; i < num_pixels; i++)
1052  ((uint32_t *) dst)[i] = ((const uint32_t *) palette)[src[i]];
1053 }
1054 
1055 /* Palette format: ABCD -> dst format: ABC */
1056 void sws_convertPalette8ToPacked24(const uint8_t *src, uint8_t *dst,
1057  int num_pixels, const uint8_t *palette)
1058 {
1059  int i;
1060 
1061  for (i = 0; i < num_pixels; i++) {
1062  //FIXME slow?
1063  dst[0] = palette[src[i] * 4 + 0];
1064  dst[1] = palette[src[i] * 4 + 1];
1065  dst[2] = palette[src[i] * 4 + 2];
1066  dst += 3;
1067  }
1068 }