FFmpeg
 All Data Structures Namespaces Files Functions Variables Typedefs Enumerations Enumerator Macros Groups Pages
rgb2rgb_template.c
Go to the documentation of this file.
1 /*
2  * software RGB to RGB converter
3  * pluralize by software PAL8 to RGB converter
4  * software YUV to YUV converter
5  * software YUV to RGB converter
6  * Written by Nick Kurshev.
7  * palette & YUV & runtime CPU stuff by Michael (michaelni@gmx.at)
8  * lot of big-endian byte order fixes by Alex Beregszaszi
9  *
10  * This file is part of FFmpeg.
11  *
12  * FFmpeg is free software; you can redistribute it and/or
13  * modify it under the terms of the GNU Lesser General Public
14  * License as published by the Free Software Foundation; either
15  * version 2.1 of the License, or (at your option) any later version.
16  *
17  * FFmpeg is distributed in the hope that it will be useful,
18  * but WITHOUT ANY WARRANTY; without even the implied warranty of
19  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
20  * Lesser General Public License for more details.
21  *
22  * You should have received a copy of the GNU Lesser General Public
23  * License along with FFmpeg; if not, write to the Free Software
24  * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
25  */
26 
27 #include <stddef.h>
28 
29 #include "libavutil/attributes.h"
30 
31 static inline void rgb24tobgr32_c(const uint8_t *src, uint8_t *dst,
32  int src_size)
33 {
34  uint8_t *dest = dst;
35  const uint8_t *s = src;
36  const uint8_t *end = s + src_size;
37 
38  while (s < end) {
39 #if HAVE_BIGENDIAN
40  /* RGB24 (= R, G, B) -> RGB32 (= A, B, G, R) */
41  *dest++ = 255;
42  *dest++ = s[2];
43  *dest++ = s[1];
44  *dest++ = s[0];
45  s += 3;
46 #else
47  *dest++ = *s++;
48  *dest++ = *s++;
49  *dest++ = *s++;
50  *dest++ = 255;
51 #endif
52  }
53 }
54 
55 static inline void rgb32tobgr24_c(const uint8_t *src, uint8_t *dst,
56  int src_size)
57 {
58  uint8_t *dest = dst;
59  const uint8_t *s = src;
60  const uint8_t *end = s + src_size;
61 
62  while (s < end) {
63 #if HAVE_BIGENDIAN
64  /* RGB32 (= A, B, G, R) -> RGB24 (= R, G, B) */
65  s++;
66  dest[2] = *s++;
67  dest[1] = *s++;
68  dest[0] = *s++;
69  dest += 3;
70 #else
71  *dest++ = *s++;
72  *dest++ = *s++;
73  *dest++ = *s++;
74  s++;
75 #endif
76  }
77 }
78 
79 /*
80  * original by Strepto/Astral
81  * ported to gcc & bugfixed: A'rpi
82  * MMXEXT, 3DNOW optimization by Nick Kurshev
83  * 32-bit C version, and and&add trick by Michael Niedermayer
84  */
85 static inline void rgb15to16_c(const uint8_t *src, uint8_t *dst, int src_size)
86 {
87  register uint8_t *d = dst;
88  register const uint8_t *s = src;
89  register const uint8_t *end = s + src_size;
90  const uint8_t *mm_end = end - 3;
91 
92  while (s < mm_end) {
93  register unsigned x = *((const uint32_t *)s);
94  *((uint32_t *)d) = (x & 0x7FFF7FFF) + (x & 0x7FE07FE0);
95  d += 4;
96  s += 4;
97  }
98  if (s < end) {
99  register unsigned short x = *((const uint16_t *)s);
100  *((uint16_t *)d) = (x & 0x7FFF) + (x & 0x7FE0);
101  }
102 }
103 
104 static inline void rgb16to15_c(const uint8_t *src, uint8_t *dst, int src_size)
105 {
106  register uint8_t *d = dst;
107  register const uint8_t *s = src;
108  register const uint8_t *end = s + src_size;
109  const uint8_t *mm_end = end - 3;
110 
111  while (s < mm_end) {
112  register uint32_t x = *((const uint32_t *)s);
113  *((uint32_t *)d) = ((x >> 1) & 0x7FE07FE0) | (x & 0x001F001F);
114  s += 4;
115  d += 4;
116  }
117  if (s < end) {
118  register uint16_t x = *((const uint16_t *)s);
119  *((uint16_t *)d) = ((x >> 1) & 0x7FE0) | (x & 0x001F);
120  }
121 }
122 
123 static inline void rgb32to16_c(const uint8_t *src, uint8_t *dst, int src_size)
124 {
125  uint16_t *d = (uint16_t *)dst;
126  const uint8_t *s = src;
127  const uint8_t *end = s + src_size;
128 
129  while (s < end) {
130  register int rgb = *(const uint32_t *)s;
131  s += 4;
132  *d++ = ((rgb & 0xFF) >> 3) +
133  ((rgb & 0xFC00) >> 5) +
134  ((rgb & 0xF80000) >> 8);
135  }
136 }
137 
138 static inline void rgb32tobgr16_c(const uint8_t *src, uint8_t *dst,
139  int src_size)
140 {
141  uint16_t *d = (uint16_t *)dst;
142  const uint8_t *s = src;
143  const uint8_t *end = s + src_size;
144 
145  while (s < end) {
146  register int rgb = *(const uint32_t *)s;
147  s += 4;
148  *d++ = ((rgb & 0xF8) << 8) +
149  ((rgb & 0xFC00) >> 5) +
150  ((rgb & 0xF80000) >> 19);
151  }
152 }
153 
154 static inline void rgb32to15_c(const uint8_t *src, uint8_t *dst, int src_size)
155 {
156  uint16_t *d = (uint16_t *)dst;
157  const uint8_t *s = src;
158  const uint8_t *end = s + src_size;
159 
160  while (s < end) {
161  register int rgb = *(const uint32_t *)s;
162  s += 4;
163  *d++ = ((rgb & 0xFF) >> 3) +
164  ((rgb & 0xF800) >> 6) +
165  ((rgb & 0xF80000) >> 9);
166  }
167 }
168 
169 static inline void rgb32tobgr15_c(const uint8_t *src, uint8_t *dst,
170  int src_size)
171 {
172  uint16_t *d = (uint16_t *)dst;
173  const uint8_t *s = src;
174  const uint8_t *end = s + src_size;
175 
176  while (s < end) {
177  register int rgb = *(const uint32_t *)s;
178  s += 4;
179  *d++ = ((rgb & 0xF8) << 7) +
180  ((rgb & 0xF800) >> 6) +
181  ((rgb & 0xF80000) >> 19);
182  }
183 }
184 
185 static inline void rgb24tobgr16_c(const uint8_t *src, uint8_t *dst,
186  int src_size)
187 {
188  uint16_t *d = (uint16_t *)dst;
189  const uint8_t *s = src;
190  const uint8_t *end = s + src_size;
191 
192  while (s < end) {
193  const int b = *s++;
194  const int g = *s++;
195  const int r = *s++;
196  *d++ = (b >> 3) | ((g & 0xFC) << 3) | ((r & 0xF8) << 8);
197  }
198 }
199 
200 static inline void rgb24to16_c(const uint8_t *src, uint8_t *dst, int src_size)
201 {
202  uint16_t *d = (uint16_t *)dst;
203  const uint8_t *s = src;
204  const uint8_t *end = s + src_size;
205 
206  while (s < end) {
207  const int r = *s++;
208  const int g = *s++;
209  const int b = *s++;
210  *d++ = (b >> 3) | ((g & 0xFC) << 3) | ((r & 0xF8) << 8);
211  }
212 }
213 
214 static inline void rgb24tobgr15_c(const uint8_t *src, uint8_t *dst,
215  int src_size)
216 {
217  uint16_t *d = (uint16_t *)dst;
218  const uint8_t *s = src;
219  const uint8_t *end = s + src_size;
220 
221  while (s < end) {
222  const int b = *s++;
223  const int g = *s++;
224  const int r = *s++;
225  *d++ = (b >> 3) | ((g & 0xF8) << 2) | ((r & 0xF8) << 7);
226  }
227 }
228 
229 static inline void rgb24to15_c(const uint8_t *src, uint8_t *dst, int src_size)
230 {
231  uint16_t *d = (uint16_t *)dst;
232  const uint8_t *s = src;
233  const uint8_t *end = s + src_size;
234 
235  while (s < end) {
236  const int r = *s++;
237  const int g = *s++;
238  const int b = *s++;
239  *d++ = (b >> 3) | ((g & 0xF8) << 2) | ((r & 0xF8) << 7);
240  }
241 }
242 
243 static inline void rgb15tobgr24_c(const uint8_t *src, uint8_t *dst,
244  int src_size)
245 {
246  uint8_t *d = dst;
247  const uint16_t *s = (const uint16_t *)src;
248  const uint16_t *end = s + src_size / 2;
249 
250  while (s < end) {
251  register uint16_t bgr = *s++;
252  *d++ = ((bgr&0x001F)<<3) | ((bgr&0x001F)>> 2);
253  *d++ = ((bgr&0x03E0)>>2) | ((bgr&0x03E0)>> 7);
254  *d++ = ((bgr&0x7C00)>>7) | ((bgr&0x7C00)>>12);
255  }
256 }
257 
258 static inline void rgb16tobgr24_c(const uint8_t *src, uint8_t *dst,
259  int src_size)
260 {
261  uint8_t *d = (uint8_t *)dst;
262  const uint16_t *s = (const uint16_t *)src;
263  const uint16_t *end = s + src_size / 2;
264 
265  while (s < end) {
266  register uint16_t bgr = *s++;
267  *d++ = ((bgr&0x001F)<<3) | ((bgr&0x001F)>> 2);
268  *d++ = ((bgr&0x07E0)>>3) | ((bgr&0x07E0)>> 9);
269  *d++ = ((bgr&0xF800)>>8) | ((bgr&0xF800)>>13);
270  }
271 }
272 
273 static inline void rgb15to32_c(const uint8_t *src, uint8_t *dst, int src_size)
274 {
275  uint8_t *d = dst;
276  const uint16_t *s = (const uint16_t *)src;
277  const uint16_t *end = s + src_size / 2;
278 
279  while (s < end) {
280  register uint16_t bgr = *s++;
281 #if HAVE_BIGENDIAN
282  *d++ = 255;
283  *d++ = ((bgr&0x7C00)>>7) | ((bgr&0x7C00)>>12);
284  *d++ = ((bgr&0x03E0)>>2) | ((bgr&0x03E0)>> 7);
285  *d++ = ((bgr&0x001F)<<3) | ((bgr&0x001F)>> 2);
286 #else
287  *d++ = ((bgr&0x001F)<<3) | ((bgr&0x001F)>> 2);
288  *d++ = ((bgr&0x03E0)>>2) | ((bgr&0x03E0)>> 7);
289  *d++ = ((bgr&0x7C00)>>7) | ((bgr&0x7C00)>>12);
290  *d++ = 255;
291 #endif
292  }
293 }
294 
295 static inline void rgb16to32_c(const uint8_t *src, uint8_t *dst, int src_size)
296 {
297  uint8_t *d = dst;
298  const uint16_t *s = (const uint16_t *)src;
299  const uint16_t *end = s + src_size / 2;
300 
301  while (s < end) {
302  register uint16_t bgr = *s++;
303 #if HAVE_BIGENDIAN
304  *d++ = 255;
305  *d++ = ((bgr&0xF800)>>8) | ((bgr&0xF800)>>13);
306  *d++ = ((bgr&0x07E0)>>3) | ((bgr&0x07E0)>> 9);
307  *d++ = ((bgr&0x001F)<<3) | ((bgr&0x001F)>> 2);
308 #else
309  *d++ = ((bgr&0x001F)<<3) | ((bgr&0x001F)>> 2);
310  *d++ = ((bgr&0x07E0)>>3) | ((bgr&0x07E0)>> 9);
311  *d++ = ((bgr&0xF800)>>8) | ((bgr&0xF800)>>13);
312  *d++ = 255;
313 #endif
314  }
315 }
316 
317 static inline void shuffle_bytes_2103_c(const uint8_t *src, uint8_t *dst,
318  int src_size)
319 {
320  int idx = 15 - src_size;
321  const uint8_t *s = src - idx;
322  uint8_t *d = dst - idx;
323 
324  for (; idx < 15; idx += 4) {
325  register unsigned v = *(const uint32_t *)&s[idx], g = v & 0xff00ff00;
326  v &= 0xff00ff;
327  *(uint32_t *)&d[idx] = (v >> 16) + g + (v << 16);
328  }
329 }
330 
331 static inline void shuffle_bytes_0321_c(const uint8_t *src, uint8_t *dst,
332  int src_size)
333 {
334  int idx = 15 - src_size;
335  const uint8_t *s = src - idx;
336  uint8_t *d = dst - idx;
337 
338  for (; idx < 15; idx += 4) {
339  register unsigned v = *(const uint32_t *)&s[idx], g = v & 0x00ff00ff;
340  v &= 0xff00ff00;
341  *(uint32_t *)&d[idx] = (v >> 16) + g + (v << 16);
342  }
343 }
344 
345 static inline void rgb24tobgr24_c(const uint8_t *src, uint8_t *dst, int src_size)
346 {
347  unsigned i;
348 
349  for (i = 0; i < src_size; i += 3) {
350  register uint8_t x = src[i + 2];
351  dst[i + 1] = src[i + 1];
352  dst[i + 2] = src[i + 0];
353  dst[i + 0] = x;
354  }
355 }
356 
357 static inline void yuvPlanartoyuy2_c(const uint8_t *ysrc, const uint8_t *usrc,
358  const uint8_t *vsrc, uint8_t *dst,
359  int width, int height,
360  int lumStride, int chromStride,
361  int dstStride, int vertLumPerChroma)
362 {
363  int y, i;
364  const int chromWidth = width >> 1;
365 
366  for (y = 0; y < height; y++) {
367 #if HAVE_FAST_64BIT
368  uint64_t *ldst = (uint64_t *)dst;
369  const uint8_t *yc = ysrc, *uc = usrc, *vc = vsrc;
370  for (i = 0; i < chromWidth; i += 2) {
371  uint64_t k = yc[0] + (uc[0] << 8) +
372  (yc[1] << 16) + ((unsigned) vc[0] << 24);
373  uint64_t l = yc[2] + (uc[1] << 8) +
374  (yc[3] << 16) + ((unsigned) vc[1] << 24);
375  *ldst++ = k + (l << 32);
376  yc += 4;
377  uc += 2;
378  vc += 2;
379  }
380 
381 #else
382  int *idst = (int32_t *)dst;
383  const uint8_t *yc = ysrc, *uc = usrc, *vc = vsrc;
384 
385  for (i = 0; i < chromWidth; i++) {
386 #if HAVE_BIGENDIAN
387  *idst++ = (yc[0] << 24) + (uc[0] << 16) +
388  (yc[1] << 8) + (vc[0] << 0);
389 #else
390  *idst++ = yc[0] + (uc[0] << 8) +
391  (yc[1] << 16) + (vc[0] << 24);
392 #endif
393  yc += 2;
394  uc++;
395  vc++;
396  }
397 #endif
398  if ((y & (vertLumPerChroma - 1)) == vertLumPerChroma - 1) {
399  usrc += chromStride;
400  vsrc += chromStride;
401  }
402  ysrc += lumStride;
403  dst += dstStride;
404  }
405 }
406 
407 /**
408  * Height should be a multiple of 2 and width should be a multiple of 16.
409  * (If this is a problem for anyone then tell me, and I will fix it.)
410  */
411 static inline void yv12toyuy2_c(const uint8_t *ysrc, const uint8_t *usrc,
412  const uint8_t *vsrc, uint8_t *dst,
413  int width, int height, int lumStride,
414  int chromStride, int dstStride)
415 {
416  //FIXME interpolate chroma
417  yuvPlanartoyuy2_c(ysrc, usrc, vsrc, dst, width, height, lumStride,
418  chromStride, dstStride, 2);
419 }
420 
421 static inline void yuvPlanartouyvy_c(const uint8_t *ysrc, const uint8_t *usrc,
422  const uint8_t *vsrc, uint8_t *dst,
423  int width, int height,
424  int lumStride, int chromStride,
425  int dstStride, int vertLumPerChroma)
426 {
427  int y, i;
428  const int chromWidth = width >> 1;
429 
430  for (y = 0; y < height; y++) {
431 #if HAVE_FAST_64BIT
432  uint64_t *ldst = (uint64_t *)dst;
433  const uint8_t *yc = ysrc, *uc = usrc, *vc = vsrc;
434  for (i = 0; i < chromWidth; i += 2) {
435  uint64_t k = uc[0] + (yc[0] << 8) +
436  (vc[0] << 16) + ((unsigned) yc[1] << 24);
437  uint64_t l = uc[1] + (yc[2] << 8) +
438  (vc[1] << 16) + ((unsigned) yc[3] << 24);
439  *ldst++ = k + (l << 32);
440  yc += 4;
441  uc += 2;
442  vc += 2;
443  }
444 
445 #else
446  int *idst = (int32_t *)dst;
447  const uint8_t *yc = ysrc, *uc = usrc, *vc = vsrc;
448 
449  for (i = 0; i < chromWidth; i++) {
450 #if HAVE_BIGENDIAN
451  *idst++ = (uc[0] << 24) + (yc[0] << 16) +
452  (vc[0] << 8) + (yc[1] << 0);
453 #else
454  *idst++ = uc[0] + (yc[0] << 8) +
455  (vc[0] << 16) + (yc[1] << 24);
456 #endif
457  yc += 2;
458  uc++;
459  vc++;
460  }
461 #endif
462  if ((y & (vertLumPerChroma - 1)) == vertLumPerChroma - 1) {
463  usrc += chromStride;
464  vsrc += chromStride;
465  }
466  ysrc += lumStride;
467  dst += dstStride;
468  }
469 }
470 
471 /**
472  * Height should be a multiple of 2 and width should be a multiple of 16
473  * (If this is a problem for anyone then tell me, and I will fix it.)
474  */
475 static inline void yv12touyvy_c(const uint8_t *ysrc, const uint8_t *usrc,
476  const uint8_t *vsrc, uint8_t *dst,
477  int width, int height, int lumStride,
478  int chromStride, int dstStride)
479 {
480  //FIXME interpolate chroma
481  yuvPlanartouyvy_c(ysrc, usrc, vsrc, dst, width, height, lumStride,
482  chromStride, dstStride, 2);
483 }
484 
485 /**
486  * Width should be a multiple of 16.
487  */
488 static inline void yuv422ptouyvy_c(const uint8_t *ysrc, const uint8_t *usrc,
489  const uint8_t *vsrc, uint8_t *dst,
490  int width, int height, int lumStride,
491  int chromStride, int dstStride)
492 {
493  yuvPlanartouyvy_c(ysrc, usrc, vsrc, dst, width, height, lumStride,
494  chromStride, dstStride, 1);
495 }
496 
497 /**
498  * Width should be a multiple of 16.
499  */
500 static inline void yuv422ptoyuy2_c(const uint8_t *ysrc, const uint8_t *usrc,
501  const uint8_t *vsrc, uint8_t *dst,
502  int width, int height, int lumStride,
503  int chromStride, int dstStride)
504 {
505  yuvPlanartoyuy2_c(ysrc, usrc, vsrc, dst, width, height, lumStride,
506  chromStride, dstStride, 1);
507 }
508 
509 /**
510  * Height should be a multiple of 2 and width should be a multiple of 16.
511  * (If this is a problem for anyone then tell me, and I will fix it.)
512  */
513 static inline void yuy2toyv12_c(const uint8_t *src, uint8_t *ydst,
514  uint8_t *udst, uint8_t *vdst,
515  int width, int height, int lumStride,
516  int chromStride, int srcStride)
517 {
518  int y;
519  const int chromWidth = width >> 1;
520 
521  for (y = 0; y < height; y += 2) {
522  int i;
523  for (i = 0; i < chromWidth; i++) {
524  ydst[2 * i + 0] = src[4 * i + 0];
525  udst[i] = src[4 * i + 1];
526  ydst[2 * i + 1] = src[4 * i + 2];
527  vdst[i] = src[4 * i + 3];
528  }
529  ydst += lumStride;
530  src += srcStride;
531 
532  for (i = 0; i < chromWidth; i++) {
533  ydst[2 * i + 0] = src[4 * i + 0];
534  ydst[2 * i + 1] = src[4 * i + 2];
535  }
536  udst += chromStride;
537  vdst += chromStride;
538  ydst += lumStride;
539  src += srcStride;
540  }
541 }
542 
543 static inline void planar2x_c(const uint8_t *src, uint8_t *dst, int srcWidth,
544  int srcHeight, int srcStride, int dstStride)
545 {
546  int x, y;
547 
548  dst[0] = src[0];
549 
550  // first line
551  for (x = 0; x < srcWidth - 1; x++) {
552  dst[2 * x + 1] = (3 * src[x] + src[x + 1]) >> 2;
553  dst[2 * x + 2] = (src[x] + 3 * src[x + 1]) >> 2;
554  }
555  dst[2 * srcWidth - 1] = src[srcWidth - 1];
556 
557  dst += dstStride;
558 
559  for (y = 1; y < srcHeight; y++) {
560  const int mmxSize = 1;
561 
562  dst[0] = (src[0] * 3 + src[srcStride]) >> 2;
563  dst[dstStride] = (src[0] + 3 * src[srcStride]) >> 2;
564 
565  for (x = mmxSize - 1; x < srcWidth - 1; x++) {
566  dst[2 * x + 1] = (src[x + 0] * 3 + src[x + srcStride + 1]) >> 2;
567  dst[2 * x + dstStride + 2] = (src[x + 0] + 3 * src[x + srcStride + 1]) >> 2;
568  dst[2 * x + dstStride + 1] = (src[x + 1] + 3 * src[x + srcStride]) >> 2;
569  dst[2 * x + 2] = (src[x + 1] * 3 + src[x + srcStride]) >> 2;
570  }
571  dst[srcWidth * 2 - 1] = (src[srcWidth - 1] * 3 + src[srcWidth - 1 + srcStride]) >> 2;
572  dst[srcWidth * 2 - 1 + dstStride] = (src[srcWidth - 1] + 3 * src[srcWidth - 1 + srcStride]) >> 2;
573 
574  dst += dstStride * 2;
575  src += srcStride;
576  }
577 
578  // last line
579  dst[0] = src[0];
580 
581  for (x = 0; x < srcWidth - 1; x++) {
582  dst[2 * x + 1] = (src[x] * 3 + src[x + 1]) >> 2;
583  dst[2 * x + 2] = (src[x] + 3 * src[x + 1]) >> 2;
584  }
585  dst[2 * srcWidth - 1] = src[srcWidth - 1];
586 }
587 
588 /**
589  * Height should be a multiple of 2 and width should be a multiple of 16.
590  * (If this is a problem for anyone then tell me, and I will fix it.)
591  * Chrominance data is only taken from every second line, others are ignored.
592  * FIXME: Write HQ version.
593  */
594 static inline void uyvytoyv12_c(const uint8_t *src, uint8_t *ydst,
595  uint8_t *udst, uint8_t *vdst,
596  int width, int height, int lumStride,
597  int chromStride, int srcStride)
598 {
599  int y;
600  const int chromWidth = width >> 1;
601 
602  for (y = 0; y < height; y += 2) {
603  int i;
604  for (i = 0; i < chromWidth; i++) {
605  udst[i] = src[4 * i + 0];
606  ydst[2 * i + 0] = src[4 * i + 1];
607  vdst[i] = src[4 * i + 2];
608  ydst[2 * i + 1] = src[4 * i + 3];
609  }
610  ydst += lumStride;
611  src += srcStride;
612 
613  for (i = 0; i < chromWidth; i++) {
614  ydst[2 * i + 0] = src[4 * i + 1];
615  ydst[2 * i + 1] = src[4 * i + 3];
616  }
617  udst += chromStride;
618  vdst += chromStride;
619  ydst += lumStride;
620  src += srcStride;
621  }
622 }
623 
624 /**
625  * Height should be a multiple of 2 and width should be a multiple of 2.
626  * (If this is a problem for anyone then tell me, and I will fix it.)
627  * Chrominance data is only taken from every second line,
628  * others are ignored in the C version.
629  * FIXME: Write HQ version.
630  */
631 void ff_rgb24toyv12_c(const uint8_t *src, uint8_t *ydst, uint8_t *udst,
632  uint8_t *vdst, int width, int height, int lumStride,
633  int chromStride, int srcStride, int32_t *rgb2yuv)
634 {
635  int32_t ry = rgb2yuv[RY_IDX], gy = rgb2yuv[GY_IDX], by = rgb2yuv[BY_IDX];
636  int32_t ru = rgb2yuv[RU_IDX], gu = rgb2yuv[GU_IDX], bu = rgb2yuv[BU_IDX];
637  int32_t rv = rgb2yuv[RV_IDX], gv = rgb2yuv[GV_IDX], bv = rgb2yuv[BV_IDX];
638  int y;
639  const int chromWidth = width >> 1;
640 
641  for (y = 0; y < height; y += 2) {
642  int i;
643  for (i = 0; i < chromWidth; i++) {
644  unsigned int b = src[6 * i + 0];
645  unsigned int g = src[6 * i + 1];
646  unsigned int r = src[6 * i + 2];
647 
648  unsigned int Y = ((ry * r + gy * g + by * b) >> RGB2YUV_SHIFT) + 16;
649  unsigned int V = ((rv * r + gv * g + bv * b) >> RGB2YUV_SHIFT) + 128;
650  unsigned int U = ((ru * r + gu * g + bu * b) >> RGB2YUV_SHIFT) + 128;
651 
652  udst[i] = U;
653  vdst[i] = V;
654  ydst[2 * i] = Y;
655 
656  b = src[6 * i + 3];
657  g = src[6 * i + 4];
658  r = src[6 * i + 5];
659 
660  Y = ((ry * r + gy * g + by * b) >> RGB2YUV_SHIFT) + 16;
661  ydst[2 * i + 1] = Y;
662  }
663  ydst += lumStride;
664  src += srcStride;
665 
666  if (y+1 == height)
667  break;
668 
669  for (i = 0; i < chromWidth; i++) {
670  unsigned int b = src[6 * i + 0];
671  unsigned int g = src[6 * i + 1];
672  unsigned int r = src[6 * i + 2];
673 
674  unsigned int Y = ((ry * r + gy * g + by * b) >> RGB2YUV_SHIFT) + 16;
675 
676  ydst[2 * i] = Y;
677 
678  b = src[6 * i + 3];
679  g = src[6 * i + 4];
680  r = src[6 * i + 5];
681 
682  Y = ((ry * r + gy * g + by * b) >> RGB2YUV_SHIFT) + 16;
683  ydst[2 * i + 1] = Y;
684  }
685  udst += chromStride;
686  vdst += chromStride;
687  ydst += lumStride;
688  src += srcStride;
689  }
690 }
691 
692 static void interleaveBytes_c(const uint8_t *src1, const uint8_t *src2,
693  uint8_t *dest, int width, int height,
694  int src1Stride, int src2Stride, int dstStride)
695 {
696  int h;
697 
698  for (h = 0; h < height; h++) {
699  int w;
700  for (w = 0; w < width; w++) {
701  dest[2 * w + 0] = src1[w];
702  dest[2 * w + 1] = src2[w];
703  }
704  dest += dstStride;
705  src1 += src1Stride;
706  src2 += src2Stride;
707  }
708 }
709 
710 static void deinterleaveBytes_c(const uint8_t *src, uint8_t *dst1, uint8_t *dst2,
711  int width, int height, int srcStride,
712  int dst1Stride, int dst2Stride)
713 {
714  int h;
715 
716  for (h = 0; h < height; h++) {
717  int w;
718  for (w = 0; w < width; w++) {
719  dst1[w] = src[2 * w + 0];
720  dst2[w] = src[2 * w + 1];
721  }
722  src += srcStride;
723  dst1 += dst1Stride;
724  dst2 += dst2Stride;
725  }
726 }
727 
728 static inline void vu9_to_vu12_c(const uint8_t *src1, const uint8_t *src2,
729  uint8_t *dst1, uint8_t *dst2,
730  int width, int height,
731  int srcStride1, int srcStride2,
732  int dstStride1, int dstStride2)
733 {
734  int x, y;
735  int w = width / 2;
736  int h = height / 2;
737 
738  for (y = 0; y < h; y++) {
739  const uint8_t *s1 = src1 + srcStride1 * (y >> 1);
740  uint8_t *d = dst1 + dstStride1 * y;
741  for (x = 0; x < w; x++)
742  d[2 * x] = d[2 * x + 1] = s1[x];
743  }
744  for (y = 0; y < h; y++) {
745  const uint8_t *s2 = src2 + srcStride2 * (y >> 1);
746  uint8_t *d = dst2 + dstStride2 * y;
747  for (x = 0; x < w; x++)
748  d[2 * x] = d[2 * x + 1] = s2[x];
749  }
750 }
751 
752 static inline void yvu9_to_yuy2_c(const uint8_t *src1, const uint8_t *src2,
753  const uint8_t *src3, uint8_t *dst,
754  int width, int height,
755  int srcStride1, int srcStride2,
756  int srcStride3, int dstStride)
757 {
758  int x, y;
759  int w = width / 2;
760  int h = height;
761 
762  for (y = 0; y < h; y++) {
763  const uint8_t *yp = src1 + srcStride1 * y;
764  const uint8_t *up = src2 + srcStride2 * (y >> 2);
765  const uint8_t *vp = src3 + srcStride3 * (y >> 2);
766  uint8_t *d = dst + dstStride * y;
767  for (x = 0; x < w; x++) {
768  const int x2 = x << 2;
769  d[8 * x + 0] = yp[x2];
770  d[8 * x + 1] = up[x];
771  d[8 * x + 2] = yp[x2 + 1];
772  d[8 * x + 3] = vp[x];
773  d[8 * x + 4] = yp[x2 + 2];
774  d[8 * x + 5] = up[x];
775  d[8 * x + 6] = yp[x2 + 3];
776  d[8 * x + 7] = vp[x];
777  }
778  }
779 }
780 
781 static void extract_even_c(const uint8_t *src, uint8_t *dst, int count)
782 {
783  dst += count;
784  src += count * 2;
785  count = -count;
786  while (count < 0) {
787  dst[count] = src[2 * count];
788  count++;
789  }
790 }
791 
792 static void extract_even2_c(const uint8_t *src, uint8_t *dst0, uint8_t *dst1,
793  int count)
794 {
795  dst0 += count;
796  dst1 += count;
797  src += count * 4;
798  count = -count;
799  while (count < 0) {
800  dst0[count] = src[4 * count + 0];
801  dst1[count] = src[4 * count + 2];
802  count++;
803  }
804 }
805 
806 static void extract_even2avg_c(const uint8_t *src0, const uint8_t *src1,
807  uint8_t *dst0, uint8_t *dst1, int count)
808 {
809  dst0 += count;
810  dst1 += count;
811  src0 += count * 4;
812  src1 += count * 4;
813  count = -count;
814  while (count < 0) {
815  dst0[count] = (src0[4 * count + 0] + src1[4 * count + 0]) >> 1;
816  dst1[count] = (src0[4 * count + 2] + src1[4 * count + 2]) >> 1;
817  count++;
818  }
819 }
820 
821 static void extract_odd2_c(const uint8_t *src, uint8_t *dst0, uint8_t *dst1,
822  int count)
823 {
824  dst0 += count;
825  dst1 += count;
826  src += count * 4;
827  count = -count;
828  src++;
829  while (count < 0) {
830  dst0[count] = src[4 * count + 0];
831  dst1[count] = src[4 * count + 2];
832  count++;
833  }
834 }
835 
836 static void extract_odd2avg_c(const uint8_t *src0, const uint8_t *src1,
837  uint8_t *dst0, uint8_t *dst1, int count)
838 {
839  dst0 += count;
840  dst1 += count;
841  src0 += count * 4;
842  src1 += count * 4;
843  count = -count;
844  src0++;
845  src1++;
846  while (count < 0) {
847  dst0[count] = (src0[4 * count + 0] + src1[4 * count + 0]) >> 1;
848  dst1[count] = (src0[4 * count + 2] + src1[4 * count + 2]) >> 1;
849  count++;
850  }
851 }
852 
853 static void yuyvtoyuv420_c(uint8_t *ydst, uint8_t *udst, uint8_t *vdst,
854  const uint8_t *src, int width, int height,
855  int lumStride, int chromStride, int srcStride)
856 {
857  int y;
858  const int chromWidth = AV_CEIL_RSHIFT(width, 1);
859 
860  for (y = 0; y < height; y++) {
861  extract_even_c(src, ydst, width);
862  if (y & 1) {
863  extract_odd2avg_c(src - srcStride, src, udst, vdst, chromWidth);
864  udst += chromStride;
865  vdst += chromStride;
866  }
867 
868  src += srcStride;
869  ydst += lumStride;
870  }
871 }
872 
873 static void yuyvtoyuv422_c(uint8_t *ydst, uint8_t *udst, uint8_t *vdst,
874  const uint8_t *src, int width, int height,
875  int lumStride, int chromStride, int srcStride)
876 {
877  int y;
878  const int chromWidth = AV_CEIL_RSHIFT(width, 1);
879 
880  for (y = 0; y < height; y++) {
881  extract_even_c(src, ydst, width);
882  extract_odd2_c(src, udst, vdst, chromWidth);
883 
884  src += srcStride;
885  ydst += lumStride;
886  udst += chromStride;
887  vdst += chromStride;
888  }
889 }
890 
891 static void uyvytoyuv420_c(uint8_t *ydst, uint8_t *udst, uint8_t *vdst,
892  const uint8_t *src, int width, int height,
893  int lumStride, int chromStride, int srcStride)
894 {
895  int y;
896  const int chromWidth = AV_CEIL_RSHIFT(width, 1);
897 
898  for (y = 0; y < height; y++) {
899  extract_even_c(src + 1, ydst, width);
900  if (y & 1) {
901  extract_even2avg_c(src - srcStride, src, udst, vdst, chromWidth);
902  udst += chromStride;
903  vdst += chromStride;
904  }
905 
906  src += srcStride;
907  ydst += lumStride;
908  }
909 }
910 
911 static void uyvytoyuv422_c(uint8_t *ydst, uint8_t *udst, uint8_t *vdst,
912  const uint8_t *src, int width, int height,
913  int lumStride, int chromStride, int srcStride)
914 {
915  int y;
916  const int chromWidth = AV_CEIL_RSHIFT(width, 1);
917 
918  for (y = 0; y < height; y++) {
919  extract_even_c(src + 1, ydst, width);
920  extract_even2_c(src, udst, vdst, chromWidth);
921 
922  src += srcStride;
923  ydst += lumStride;
924  udst += chromStride;
925  vdst += chromStride;
926  }
927 }
928 
929 static av_cold void rgb2rgb_init_c(void)
930 {
946 #if HAVE_BIGENDIAN
949 #else
952 #endif
966 
971 }
static void uyvytoyuv420_c(uint8_t *ydst, uint8_t *udst, uint8_t *vdst, const uint8_t *src, int width, int height, int lumStride, int chromStride, int srcStride)
static void uyvytoyv12_c(const uint8_t *src, uint8_t *ydst, uint8_t *udst, uint8_t *vdst, int width, int height, int lumStride, int chromStride, int srcStride)
Height should be a multiple of 2 and width should be a multiple of 16.
const char * s
Definition: avisynth_c.h:768
static void yuv422ptoyuy2_c(const uint8_t *ysrc, const uint8_t *usrc, const uint8_t *vsrc, uint8_t *dst, int width, int height, int lumStride, int chromStride, int dstStride)
Width should be a multiple of 16.
#define RV_IDX
#define BV_IDX
static void extract_odd2avg_c(const uint8_t *src0, const uint8_t *src1, uint8_t *dst0, uint8_t *dst1, int count)
static void fn() rgb2yuv(uint8_t *_yuv[3], const ptrdiff_t yuv_stride[3], int16_t *rgb[3], ptrdiff_t s, int w, int h, const int16_t rgb2yuv_coeffs[3][3][8], const int16_t yuv_offset[8])
static void yuyvtoyuv420_c(uint8_t *ydst, uint8_t *udst, uint8_t *vdst, const uint8_t *src, int width, int height, int lumStride, int chromStride, int srcStride)
void(* ff_rgb24toyv12)(const uint8_t *src, uint8_t *ydst, uint8_t *udst, uint8_t *vdst, int width, int height, int lumStride, int chromStride, int srcStride, int32_t *rgb2yuv)
Height should be a multiple of 2 and width should be a multiple of 2.
Definition: rgb2rgb.c:77
void(* planar2x)(const uint8_t *src, uint8_t *dst, int width, int height, int srcStride, int dstStride)
Definition: rgb2rgb.c:82
const char * g
Definition: vf_curves.c:112
static void planar2x_c(const uint8_t *src, uint8_t *dst, int srcWidth, int srcHeight, int srcStride, int dstStride)
static void rgb16to32_c(const uint8_t *src, uint8_t *dst, int src_size)
const char * b
Definition: vf_curves.c:113
#define GV_IDX
void(* yuyvtoyuv422)(uint8_t *ydst, uint8_t *udst, uint8_t *vdst, const uint8_t *src, int width, int height, int lumStride, int chromStride, int srcStride)
Definition: rgb2rgb.c:109
void(* yuv422ptouyvy)(const uint8_t *ysrc, const uint8_t *usrc, const uint8_t *vsrc, uint8_t *dst, int width, int height, int lumStride, int chromStride, int dstStride)
Width should be a multiple of 16.
Definition: rgb2rgb.c:69
static void rgb24tobgr32_c(const uint8_t *src, uint8_t *dst, int src_size)
void(* rgb32to16)(const uint8_t *src, uint8_t *dst, int src_size)
Definition: rgb2rgb.c:45
static void rgb15tobgr24_c(const uint8_t *src, uint8_t *dst, int src_size)
static void vu9_to_vu12_c(const uint8_t *src1, const uint8_t *src2, uint8_t *dst1, uint8_t *dst2, int width, int height, int srcStride1, int srcStride2, int dstStride1, int dstStride2)
static void yuy2toyv12_c(const uint8_t *src, uint8_t *ydst, uint8_t *udst, uint8_t *vdst, int width, int height, int lumStride, int chromStride, int srcStride)
Height should be a multiple of 2 and width should be a multiple of 16.
Macro definitions for various function/variable attributes.
void(* rgb16to15)(const uint8_t *src, uint8_t *dst, int src_size)
Definition: rgb2rgb.c:50
void(* uyvytoyuv420)(uint8_t *ydst, uint8_t *udst, uint8_t *vdst, const uint8_t *src, int width, int height, int lumStride, int chromStride, int srcStride)
Definition: rgb2rgb.c:100
void(* uyvytoyuv422)(uint8_t *ydst, uint8_t *udst, uint8_t *vdst, const uint8_t *src, int width, int height, int lumStride, int chromStride, int srcStride)
Definition: rgb2rgb.c:103
static void rgb16tobgr24_c(const uint8_t *src, uint8_t *dst, int src_size)
#define RU_IDX
static void rgb24tobgr16_c(const uint8_t *src, uint8_t *dst, int src_size)
uint8_t
#define av_cold
Definition: attributes.h:82
void(* yuv422ptoyuy2)(const uint8_t *ysrc, const uint8_t *usrc, const uint8_t *vsrc, uint8_t *dst, int width, int height, int lumStride, int chromStride, int dstStride)
Width should be a multiple of 16.
Definition: rgb2rgb.c:65
#define Y
Definition: vf_boxblur.c:76
static av_cold int end(AVCodecContext *avctx)
Definition: avrndec.c:90
static void extract_even_c(const uint8_t *src, uint8_t *dst, int count)
void ff_rgb24toyv12_c(const uint8_t *src, uint8_t *ydst, uint8_t *udst, uint8_t *vdst, int width, int height, int lumStride, int chromStride, int srcStride, int32_t *rgb2yuv)
Height should be a multiple of 2 and width should be a multiple of 2.
void(* rgb15tobgr24)(const uint8_t *src, uint8_t *dst, int src_size)
Definition: rgb2rgb.c:43
#define height
static void rgb16to15_c(const uint8_t *src, uint8_t *dst, int src_size)
static void yuvPlanartouyvy_c(const uint8_t *ysrc, const uint8_t *usrc, const uint8_t *vsrc, uint8_t *dst, int width, int height, int lumStride, int chromStride, int dstStride, int vertLumPerChroma)
void(* yv12touyvy)(const uint8_t *ysrc, const uint8_t *usrc, const uint8_t *vsrc, uint8_t *dst, int width, int height, int lumStride, int chromStride, int dstStride)
Height should be a multiple of 2 and width should be a multiple of 16.
Definition: rgb2rgb.c:61
#define U(x)
Definition: vp56_arith.h:37
#define s2
Definition: regdef.h:39
void(* interleaveBytes)(const uint8_t *src1, const uint8_t *src2, uint8_t *dst, int width, int height, int src1Stride, int src2Stride, int dstStride)
Definition: rgb2rgb.c:84
static av_cold void rgb2rgb_init_c(void)
void(* rgb16to32)(const uint8_t *src, uint8_t *dst, int src_size)
Definition: rgb2rgb.c:49
void(* rgb24tobgr32)(const uint8_t *src, uint8_t *dst, int src_size)
Definition: rgb2rgb.c:38
const char * r
Definition: vf_curves.c:111
static void yv12toyuy2_c(const uint8_t *ysrc, const uint8_t *usrc, const uint8_t *vsrc, uint8_t *dst, int width, int height, int lumStride, int chromStride, int dstStride)
Height should be a multiple of 2 and width should be a multiple of 16.
void(* rgb24to15)(const uint8_t *src, uint8_t *dst, int src_size)
Definition: rgb2rgb.c:48
static void interleaveBytes_c(const uint8_t *src1, const uint8_t *src2, uint8_t *dest, int width, int height, int src1Stride, int src2Stride, int dstStride)
GLsizei count
Definition: opengl_enc.c:109
static void rgb24to16_c(const uint8_t *src, uint8_t *dst, int src_size)
static void extract_even2_c(const uint8_t *src, uint8_t *dst0, uint8_t *dst1, int count)
#define RGB2YUV_SHIFT
static void yv12touyvy_c(const uint8_t *ysrc, const uint8_t *usrc, const uint8_t *vsrc, uint8_t *dst, int width, int height, int lumStride, int chromStride, int dstStride)
Height should be a multiple of 2 and width should be a multiple of 16 (If this is a problem for anyon...
#define width
static void shuffle_bytes_0321_c(const uint8_t *src, uint8_t *dst, int src_size)
void(* rgb32tobgr15)(const uint8_t *src, uint8_t *dst, int src_size)
Definition: rgb2rgb.c:37
static void rgb24tobgr24_c(const uint8_t *src, uint8_t *dst, int src_size)
int32_t
#define RY_IDX
#define src
Definition: vp9dsp.c:530
static void rgb32tobgr24_c(const uint8_t *src, uint8_t *dst, int src_size)
static void uyvytoyuv422_c(uint8_t *ydst, uint8_t *udst, uint8_t *vdst, const uint8_t *src, int width, int height, int lumStride, int chromStride, int srcStride)
void(* yuy2toyv12)(const uint8_t *src, uint8_t *ydst, uint8_t *udst, uint8_t *vdst, int width, int height, int lumStride, int chromStride, int srcStride)
Height should be a multiple of 2 and width should be a multiple of 16.
Definition: rgb2rgb.c:73
static void yuyvtoyuv422_c(uint8_t *ydst, uint8_t *udst, uint8_t *vdst, const uint8_t *src, int width, int height, int lumStride, int chromStride, int srcStride)
void(* rgb15to32)(const uint8_t *src, uint8_t *dst, int src_size)
Definition: rgb2rgb.c:52
#define src1
Definition: h264pred.c:139
static void yuv422ptouyvy_c(const uint8_t *ysrc, const uint8_t *usrc, const uint8_t *vsrc, uint8_t *dst, int width, int height, int lumStride, int chromStride, int dstStride)
Width should be a multiple of 16.
static void deinterleaveBytes_c(const uint8_t *src, uint8_t *dst1, uint8_t *dst2, int width, int height, int srcStride, int dst1Stride, int dst2Stride)
void(* rgb15to16)(const uint8_t *src, uint8_t *dst, int src_size)
Definition: rgb2rgb.c:51
static void yvu9_to_yuy2_c(const uint8_t *src1, const uint8_t *src2, const uint8_t *src3, uint8_t *dst, int width, int height, int srcStride1, int srcStride2, int srcStride3, int dstStride)
void(* vu9_to_vu12)(const uint8_t *src1, const uint8_t *src2, uint8_t *dst1, uint8_t *dst2, int width, int height, int srcStride1, int srcStride2, int dstStride1, int dstStride2)
Definition: rgb2rgb.c:90
static void rgb15to32_c(const uint8_t *src, uint8_t *dst, int src_size)
static void rgb15to16_c(const uint8_t *src, uint8_t *dst, int src_size)
void(* rgb32to15)(const uint8_t *src, uint8_t *dst, int src_size)
Definition: rgb2rgb.c:46
void(* shuffle_bytes_2103)(const uint8_t *src, uint8_t *dst, int src_size)
Definition: rgb2rgb.c:55
#define src0
Definition: h264pred.c:138
#define s1
Definition: regdef.h:38
static void rgb32to16_c(const uint8_t *src, uint8_t *dst, int src_size)
void(* shuffle_bytes_0321)(const uint8_t *src, uint8_t *dst, int src_size)
Definition: rgb2rgb.c:54
#define GY_IDX
void(* rgb24tobgr15)(const uint8_t *src, uint8_t *dst, int src_size)
Definition: rgb2rgb.c:41
static void extract_even2avg_c(const uint8_t *src0, const uint8_t *src1, uint8_t *dst0, uint8_t *dst1, int count)
static void extract_odd2_c(const uint8_t *src, uint8_t *dst0, uint8_t *dst1, int count)
void(* yv12toyuy2)(const uint8_t *ysrc, const uint8_t *usrc, const uint8_t *vsrc, uint8_t *dst, int width, int height, int lumStride, int chromStride, int dstStride)
Height should be a multiple of 2 and width should be a multiple of 16.
Definition: rgb2rgb.c:57
static void rgb32to15_c(const uint8_t *src, uint8_t *dst, int src_size)
static void yuvPlanartoyuy2_c(const uint8_t *ysrc, const uint8_t *usrc, const uint8_t *vsrc, uint8_t *dst, int width, int height, int lumStride, int chromStride, int dstStride, int vertLumPerChroma)
static void rgb24tobgr15_c(const uint8_t *src, uint8_t *dst, int src_size)
void(* yuyvtoyuv420)(uint8_t *ydst, uint8_t *udst, uint8_t *vdst, const uint8_t *src, int width, int height, int lumStride, int chromStride, int srcStride)
Definition: rgb2rgb.c:106
void(* rgb24tobgr16)(const uint8_t *src, uint8_t *dst, int src_size)
Definition: rgb2rgb.c:40
static void rgb32tobgr15_c(const uint8_t *src, uint8_t *dst, int src_size)
void(* yvu9_to_yuy2)(const uint8_t *src1, const uint8_t *src2, const uint8_t *src3, uint8_t *dst, int width, int height, int srcStride1, int srcStride2, int srcStride3, int dstStride)
Definition: rgb2rgb.c:95
void(* rgb32tobgr24)(const uint8_t *src, uint8_t *dst, int src_size)
Definition: rgb2rgb.c:35
void(* rgb24to16)(const uint8_t *src, uint8_t *dst, int src_size)
Definition: rgb2rgb.c:47
void(* rgb16tobgr24)(const uint8_t *src, uint8_t *dst, int src_size)
Definition: rgb2rgb.c:42
void(* rgb32tobgr16)(const uint8_t *src, uint8_t *dst, int src_size)
Definition: rgb2rgb.c:36
static void shuffle_bytes_2103_c(const uint8_t *src, uint8_t *dst, int src_size)
static void rgb24to15_c(const uint8_t *src, uint8_t *dst, int src_size)
void(* deinterleaveBytes)(const uint8_t *src, uint8_t *dst1, uint8_t *dst2, int width, int height, int srcStride, int dst1Stride, int dst2Stride)
Definition: rgb2rgb.c:87
#define GU_IDX
#define BY_IDX
static void rgb32tobgr16_c(const uint8_t *src, uint8_t *dst, int src_size)
#define BU_IDX
void(* rgb24tobgr24)(const uint8_t *src, uint8_t *dst, int src_size)
Definition: rgb2rgb.c:39
#define V
Definition: avdct.c:30
#define AV_CEIL_RSHIFT(a, b)
Definition: common.h:58