2 * Copyright (C) 2001-2011 Michael Niedermayer <michaelni@gmx.at>
4 * This file is part of FFmpeg.
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.
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.
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
27 #include "swscale_internal.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"
37 DECLARE_ALIGNED(8, static const uint8_t, dithers)[8][8][8]={
39 { 0, 1, 0, 1, 0, 1, 0, 1,},
40 { 1, 0, 1, 0, 1, 0, 1, 0,},
41 { 0, 1, 0, 1, 0, 1, 0, 1,},
42 { 1, 0, 1, 0, 1, 0, 1, 0,},
43 { 0, 1, 0, 1, 0, 1, 0, 1,},
44 { 1, 0, 1, 0, 1, 0, 1, 0,},
45 { 0, 1, 0, 1, 0, 1, 0, 1,},
46 { 1, 0, 1, 0, 1, 0, 1, 0,},
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,},
57 { 2, 4, 3, 5, 2, 4, 3, 5,},
58 { 6, 0, 7, 1, 6, 0, 7, 1,},
59 { 3, 5, 2, 4, 3, 5, 2, 4,},
60 { 7, 1, 6, 0, 7, 1, 6, 0,},
61 { 2, 4, 3, 5, 2, 4, 3, 5,},
62 { 6, 0, 7, 1, 6, 0, 7, 1,},
63 { 3, 5, 2, 4, 3, 5, 2, 4,},
64 { 7, 1, 6, 0, 7, 1, 6, 0,},
66 { 4, 8, 7, 11, 4, 8, 7, 11,},
67 { 12, 0, 15, 3, 12, 0, 15, 3,},
68 { 6, 10, 5, 9, 6, 10, 5, 9,},
69 { 14, 2, 13, 1, 14, 2, 13, 1,},
70 { 4, 8, 7, 11, 4, 8, 7, 11,},
71 { 12, 0, 15, 3, 12, 0, 15, 3,},
72 { 6, 10, 5, 9, 6, 10, 5, 9,},
73 { 14, 2, 13, 1, 14, 2, 13, 1,},
75 { 9, 17, 15, 23, 8, 16, 14, 22,},
76 { 25, 1, 31, 7, 24, 0, 30, 6,},
77 { 13, 21, 11, 19, 12, 20, 10, 18,},
78 { 29, 5, 27, 3, 28, 4, 26, 2,},
79 { 8, 16, 14, 22, 9, 17, 15, 23,},
80 { 24, 0, 30, 6, 25, 1, 31, 7,},
81 { 12, 20, 10, 18, 13, 21, 11, 19,},
82 { 28, 4, 26, 2, 29, 5, 27, 3,},
84 { 18, 34, 30, 46, 17, 33, 29, 45,},
85 { 50, 2, 62, 14, 49, 1, 61, 13,},
86 { 26, 42, 22, 38, 25, 41, 21, 37,},
87 { 58, 10, 54, 6, 57, 9, 53, 5,},
88 { 16, 32, 28, 44, 19, 35, 31, 47,},
89 { 48, 0, 60, 12, 51, 3, 63, 15,},
90 { 24, 40, 20, 36, 27, 43, 23, 39,},
91 { 56, 8, 52, 4, 59, 11, 55, 7,},
93 { 18, 34, 30, 46, 17, 33, 29, 45,},
94 { 50, 2, 62, 14, 49, 1, 61, 13,},
95 { 26, 42, 22, 38, 25, 41, 21, 37,},
96 { 58, 10, 54, 6, 57, 9, 53, 5,},
97 { 16, 32, 28, 44, 19, 35, 31, 47,},
98 { 48, 0, 60, 12, 51, 3, 63, 15,},
99 { 24, 40, 20, 36, 27, 43, 23, 39,},
100 { 56, 8, 52, 4, 59, 11, 55, 7,},
102 { 36, 68, 60, 92, 34, 66, 58, 90,},
103 { 100, 4,124, 28, 98, 2,122, 26,},
104 { 52, 84, 44, 76, 50, 82, 42, 74,},
105 { 116, 20,108, 12,114, 18,106, 10,},
106 { 32, 64, 56, 88, 38, 70, 62, 94,},
107 { 96, 0,120, 24,102, 6,126, 30,},
108 { 48, 80, 40, 72, 54, 86, 46, 78,},
109 { 112, 16,104, 8,118, 22,110, 14,},
112 static const uint16_t dither_scale[15][16]={
113 { 2, 3, 3, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5,},
114 { 2, 3, 7, 7, 13, 13, 25, 25, 25, 25, 25, 25, 25, 25, 25, 25,},
115 { 3, 3, 4, 15, 15, 29, 57, 57, 57, 113, 113, 113, 113, 113, 113, 113,},
116 { 3, 4, 4, 5, 31, 31, 61, 121, 241, 241, 241, 241, 481, 481, 481, 481,},
117 { 3, 4, 5, 5, 6, 63, 63, 125, 249, 497, 993, 993, 993, 993, 993, 1985,},
118 { 3, 5, 6, 6, 6, 7, 127, 127, 253, 505, 1009, 2017, 4033, 4033, 4033, 4033,},
119 { 3, 5, 6, 7, 7, 7, 8, 255, 255, 509, 1017, 2033, 4065, 8129,16257,16257,},
120 { 3, 5, 6, 8, 8, 8, 8, 9, 511, 511, 1021, 2041, 4081, 8161,16321,32641,},
121 { 3, 5, 7, 8, 9, 9, 9, 9, 10, 1023, 1023, 2045, 4089, 8177,16353,32705,},
122 { 3, 5, 7, 8, 10, 10, 10, 10, 10, 11, 2047, 2047, 4093, 8185,16369,32737,},
123 { 3, 5, 7, 8, 10, 11, 11, 11, 11, 11, 12, 4095, 4095, 8189,16377,32753,},
124 { 3, 5, 7, 9, 10, 12, 12, 12, 12, 12, 12, 13, 8191, 8191,16381,32761,},
125 { 3, 5, 7, 9, 10, 12, 13, 13, 13, 13, 13, 13, 14,16383,16383,32765,},
126 { 3, 5, 7, 9, 10, 12, 14, 14, 14, 14, 14, 14, 14, 15,32767,32767,},
127 { 3, 5, 7, 9, 11, 12, 14, 15, 15, 15, 15, 15, 15, 15, 16,65535,},
131 static void fillPlane(uint8_t *plane, int stride, int width, int height, int y,
135 uint8_t *ptr = plane + stride * y;
136 for (i = 0; i < height; i++) {
137 memset(ptr, val, width);
142 static void copyPlane(const uint8_t *src, int srcStride,
143 int srcSliceY, int srcSliceH, int width,
144 uint8_t *dst, int dstStride)
146 dst += dstStride * srcSliceY;
147 if (dstStride == srcStride && srcStride > 0) {
148 memcpy(dst, src, srcSliceH * dstStride);
151 for (i = 0; i < srcSliceH; i++) {
152 memcpy(dst, src, width);
159 static int planarToNv12Wrapper(SwsContext *c, const uint8_t *src[],
160 int srcStride[], int srcSliceY,
161 int srcSliceH, uint8_t *dstParam[],
164 uint8_t *dst = dstParam[1] + dstStride[1] * srcSliceY / 2;
166 copyPlane(src[0], srcStride[0], srcSliceY, srcSliceH, c->srcW,
167 dstParam[0], dstStride[0]);
169 if (c->dstFormat == AV_PIX_FMT_NV12)
170 interleaveBytes(src[1], src[2], dst, c->srcW / 2, srcSliceH / 2,
171 srcStride[1], srcStride[2], dstStride[0]);
173 interleaveBytes(src[2], src[1], dst, c->srcW / 2, srcSliceH / 2,
174 srcStride[2], srcStride[1], dstStride[0]);
179 static int planarToYuy2Wrapper(SwsContext *c, const uint8_t *src[],
180 int srcStride[], int srcSliceY, int srcSliceH,
181 uint8_t *dstParam[], int dstStride[])
183 uint8_t *dst = dstParam[0] + dstStride[0] * srcSliceY;
185 yv12toyuy2(src[0], src[1], src[2], dst, c->srcW, srcSliceH, srcStride[0],
186 srcStride[1], dstStride[0]);
191 static int planarToUyvyWrapper(SwsContext *c, const uint8_t *src[],
192 int srcStride[], int srcSliceY, int srcSliceH,
193 uint8_t *dstParam[], int dstStride[])
195 uint8_t *dst = dstParam[0] + dstStride[0] * srcSliceY;
197 yv12touyvy(src[0], src[1], src[2], dst, c->srcW, srcSliceH, srcStride[0],
198 srcStride[1], dstStride[0]);
203 static int yuv422pToYuy2Wrapper(SwsContext *c, const uint8_t *src[],
204 int srcStride[], int srcSliceY, int srcSliceH,
205 uint8_t *dstParam[], int dstStride[])
207 uint8_t *dst = dstParam[0] + dstStride[0] * srcSliceY;
209 yuv422ptoyuy2(src[0], src[1], src[2], dst, c->srcW, srcSliceH, srcStride[0],
210 srcStride[1], dstStride[0]);
215 static int yuv422pToUyvyWrapper(SwsContext *c, const uint8_t *src[],
216 int srcStride[], int srcSliceY, int srcSliceH,
217 uint8_t *dstParam[], int dstStride[])
219 uint8_t *dst = dstParam[0] + dstStride[0] * srcSliceY;
221 yuv422ptouyvy(src[0], src[1], src[2], dst, c->srcW, srcSliceH, srcStride[0],
222 srcStride[1], dstStride[0]);
227 static int yuyvToYuv420Wrapper(SwsContext *c, const uint8_t *src[],
228 int srcStride[], int srcSliceY, int srcSliceH,
229 uint8_t *dstParam[], int dstStride[])
231 uint8_t *ydst = dstParam[0] + dstStride[0] * srcSliceY;
232 uint8_t *udst = dstParam[1] + dstStride[1] * srcSliceY / 2;
233 uint8_t *vdst = dstParam[2] + dstStride[2] * srcSliceY / 2;
235 yuyvtoyuv420(ydst, udst, vdst, src[0], c->srcW, srcSliceH, dstStride[0],
236 dstStride[1], srcStride[0]);
239 fillPlane(dstParam[3], dstStride[3], c->srcW, srcSliceH, srcSliceY, 255);
244 static int yuyvToYuv422Wrapper(SwsContext *c, const uint8_t *src[],
245 int srcStride[], int srcSliceY, int srcSliceH,
246 uint8_t *dstParam[], int dstStride[])
248 uint8_t *ydst = dstParam[0] + dstStride[0] * srcSliceY;
249 uint8_t *udst = dstParam[1] + dstStride[1] * srcSliceY;
250 uint8_t *vdst = dstParam[2] + dstStride[2] * srcSliceY;
252 yuyvtoyuv422(ydst, udst, vdst, src[0], c->srcW, srcSliceH, dstStride[0],
253 dstStride[1], srcStride[0]);
258 static int uyvyToYuv420Wrapper(SwsContext *c, const uint8_t *src[],
259 int srcStride[], int srcSliceY, int srcSliceH,
260 uint8_t *dstParam[], int dstStride[])
262 uint8_t *ydst = dstParam[0] + dstStride[0] * srcSliceY;
263 uint8_t *udst = dstParam[1] + dstStride[1] * srcSliceY / 2;
264 uint8_t *vdst = dstParam[2] + dstStride[2] * srcSliceY / 2;
266 uyvytoyuv420(ydst, udst, vdst, src[0], c->srcW, srcSliceH, dstStride[0],
267 dstStride[1], srcStride[0]);
270 fillPlane(dstParam[3], dstStride[3], c->srcW, srcSliceH, srcSliceY, 255);
275 static int uyvyToYuv422Wrapper(SwsContext *c, const uint8_t *src[],
276 int srcStride[], int srcSliceY, int srcSliceH,
277 uint8_t *dstParam[], int dstStride[])
279 uint8_t *ydst = dstParam[0] + dstStride[0] * srcSliceY;
280 uint8_t *udst = dstParam[1] + dstStride[1] * srcSliceY;
281 uint8_t *vdst = dstParam[2] + dstStride[2] * srcSliceY;
283 uyvytoyuv422(ydst, udst, vdst, src[0], c->srcW, srcSliceH, dstStride[0],
284 dstStride[1], srcStride[0]);
289 static void gray8aToPacked32(const uint8_t *src, uint8_t *dst, int num_pixels,
290 const uint8_t *palette)
293 for (i = 0; i < num_pixels; i++)
294 ((uint32_t *) dst)[i] = ((const uint32_t *) palette)[src[i << 1]] | (src[(i << 1) + 1] << 24);
297 static void gray8aToPacked32_1(const uint8_t *src, uint8_t *dst, int num_pixels,
298 const uint8_t *palette)
302 for (i = 0; i < num_pixels; i++)
303 ((uint32_t *) dst)[i] = ((const uint32_t *) palette)[src[i << 1]] | src[(i << 1) + 1];
306 static void gray8aToPacked24(const uint8_t *src, uint8_t *dst, int num_pixels,
307 const uint8_t *palette)
311 for (i = 0; i < num_pixels; i++) {
313 dst[0] = palette[src[i << 1] * 4 + 0];
314 dst[1] = palette[src[i << 1] * 4 + 1];
315 dst[2] = palette[src[i << 1] * 4 + 2];
320 static int packed_16bpc_bswap(SwsContext *c, const uint8_t *src[],
321 int srcStride[], int srcSliceY, int srcSliceH,
322 uint8_t *dst[], int dstStride[])
326 for (p = 0; p < 4; p++) {
327 int srcstr = srcStride[p] / 2;
328 int dststr = dstStride[p] / 2;
329 uint16_t *dstPtr = (uint16_t *) dst[p];
330 const uint16_t *srcPtr = (const uint16_t *) src[p];
331 int min_stride = FFMIN(FFABS(srcstr), FFABS(dststr));
332 if(!dstPtr || !srcPtr)
334 for (i = 0; i < (srcSliceH >> c->chrDstVSubSample); i++) {
335 for (j = 0; j < min_stride; j++) {
336 dstPtr[j] = av_bswap16(srcPtr[j]);
346 static int palToRgbWrapper(SwsContext *c, const uint8_t *src[], int srcStride[],
347 int srcSliceY, int srcSliceH, uint8_t *dst[],
350 const enum AVPixelFormat srcFormat = c->srcFormat;
351 const enum AVPixelFormat dstFormat = c->dstFormat;
352 void (*conv)(const uint8_t *src, uint8_t *dst, int num_pixels,
353 const uint8_t *palette) = NULL;
355 uint8_t *dstPtr = dst[0] + dstStride[0] * srcSliceY;
356 const uint8_t *srcPtr = src[0];
358 if (srcFormat == AV_PIX_FMT_GRAY8A) {
360 case AV_PIX_FMT_RGB32 : conv = gray8aToPacked32; break;
361 case AV_PIX_FMT_BGR32 : conv = gray8aToPacked32; break;
362 case AV_PIX_FMT_BGR32_1: conv = gray8aToPacked32_1; break;
363 case AV_PIX_FMT_RGB32_1: conv = gray8aToPacked32_1; break;
364 case AV_PIX_FMT_RGB24 : conv = gray8aToPacked24; break;
365 case AV_PIX_FMT_BGR24 : conv = gray8aToPacked24; break;
367 } else if (usePal(srcFormat)) {
369 case AV_PIX_FMT_RGB32 : conv = sws_convertPalette8ToPacked32; break;
370 case AV_PIX_FMT_BGR32 : conv = sws_convertPalette8ToPacked32; break;
371 case AV_PIX_FMT_BGR32_1: conv = sws_convertPalette8ToPacked32; break;
372 case AV_PIX_FMT_RGB32_1: conv = sws_convertPalette8ToPacked32; break;
373 case AV_PIX_FMT_RGB24 : conv = sws_convertPalette8ToPacked24; break;
374 case AV_PIX_FMT_BGR24 : conv = sws_convertPalette8ToPacked24; break;
379 av_log(c, AV_LOG_ERROR, "internal error %s -> %s converter\n",
380 av_get_pix_fmt_name(srcFormat), av_get_pix_fmt_name(dstFormat));
382 for (i = 0; i < srcSliceH; i++) {
383 conv(srcPtr, dstPtr, c->srcW, (uint8_t *) c->pal_rgb);
384 srcPtr += srcStride[0];
385 dstPtr += dstStride[0];
392 static void gbr16ptopacked16(const uint16_t *src[], int srcStride[],
393 uint8_t *dst, int dstStride, int srcSliceH,
394 int alpha, int swap, int bpp, int width)
397 int src_alpha = src[3] != NULL;
398 int scale_high = 16 - bpp, scale_low = (bpp - 8) * 2;
399 for (h = 0; h < srcSliceH; h++) {
400 uint16_t *dest = (uint16_t *)(dst + dstStride * h);
405 if (alpha && !src_alpha) {
406 for (x = 0; x < width; x++) {
407 component = av_bswap16(src[0][x]);
408 *dest++ = av_bswap16(component << scale_high | component >> scale_low);
409 component = av_bswap16(src[1][x]);
410 *dest++ = av_bswap16(component << scale_high | component >> scale_low);
411 component = av_bswap16(src[2][x]);
412 *dest++ = av_bswap16(component << scale_high | component >> scale_low);
415 } else if (alpha && src_alpha) {
416 for (x = 0; x < width; x++) {
417 component = av_bswap16(src[0][x]);
418 *dest++ = av_bswap16(component << scale_high | component >> scale_low);
419 component = av_bswap16(src[1][x]);
420 *dest++ = av_bswap16(component << scale_high | component >> scale_low);
421 component = av_bswap16(src[2][x]);
422 *dest++ = av_bswap16(component << scale_high | component >> scale_low);
423 component = av_bswap16(src[3][x]);
424 *dest++ = av_bswap16(component << scale_high | component >> scale_low);
427 for (x = 0; x < width; x++) {
428 component = av_bswap16(src[0][x]);
429 *dest++ = av_bswap16(component << scale_high | component >> scale_low);
430 component = av_bswap16(src[1][x]);
431 *dest++ = av_bswap16(component << scale_high | component >> scale_low);
432 component = av_bswap16(src[2][x]);
433 *dest++ = av_bswap16(component << scale_high | component >> scale_low);
438 if (alpha && !src_alpha) {
439 for (x = 0; x < width; x++) {
440 *dest++ = av_bswap16(src[0][x] << scale_high | src[0][x] >> scale_low);
441 *dest++ = av_bswap16(src[1][x] << scale_high | src[1][x] >> scale_low);
442 *dest++ = av_bswap16(src[2][x] << scale_high | src[2][x] >> scale_low);
445 } else if (alpha && src_alpha) {
446 for (x = 0; x < width; x++) {
447 *dest++ = av_bswap16(src[0][x] << scale_high | src[0][x] >> scale_low);
448 *dest++ = av_bswap16(src[1][x] << scale_high | src[1][x] >> scale_low);
449 *dest++ = av_bswap16(src[2][x] << scale_high | src[2][x] >> scale_low);
450 *dest++ = av_bswap16(src[3][x] << scale_high | src[3][x] >> scale_low);
453 for (x = 0; x < width; x++) {
454 *dest++ = av_bswap16(src[0][x] << scale_high | src[0][x] >> scale_low);
455 *dest++ = av_bswap16(src[1][x] << scale_high | src[1][x] >> scale_low);
456 *dest++ = av_bswap16(src[2][x] << scale_high | src[2][x] >> scale_low);
461 if (alpha && !src_alpha) {
462 for (x = 0; x < width; x++) {
463 *dest++ = av_bswap16(src[0][x]) << scale_high | av_bswap16(src[0][x]) >> scale_low;
464 *dest++ = av_bswap16(src[1][x]) << scale_high | av_bswap16(src[1][x]) >> scale_low;
465 *dest++ = av_bswap16(src[2][x]) << scale_high | av_bswap16(src[2][x]) >> scale_low;
468 } else if (alpha && src_alpha) {
469 for (x = 0; x < width; x++) {
470 *dest++ = av_bswap16(src[0][x]) << scale_high | av_bswap16(src[0][x]) >> scale_low;
471 *dest++ = av_bswap16(src[1][x]) << scale_high | av_bswap16(src[1][x]) >> scale_low;
472 *dest++ = av_bswap16(src[2][x]) << scale_high | av_bswap16(src[2][x]) >> scale_low;
473 *dest++ = av_bswap16(src[3][x]) << scale_high | av_bswap16(src[3][x]) >> scale_low;
476 for (x = 0; x < width; x++) {
477 *dest++ = av_bswap16(src[0][x]) << scale_high | av_bswap16(src[0][x]) >> scale_low;
478 *dest++ = av_bswap16(src[1][x]) << scale_high | av_bswap16(src[1][x]) >> scale_low;
479 *dest++ = av_bswap16(src[2][x]) << scale_high | av_bswap16(src[2][x]) >> scale_low;
484 if (alpha && !src_alpha) {
485 for (x = 0; x < width; x++) {
486 *dest++ = src[0][x] << scale_high | src[0][x] >> scale_low;
487 *dest++ = src[1][x] << scale_high | src[1][x] >> scale_low;
488 *dest++ = src[2][x] << scale_high | src[2][x] >> scale_low;
491 } else if (alpha && src_alpha) {
492 for (x = 0; x < width; x++) {
493 *dest++ = src[0][x] << scale_high | src[0][x] >> scale_low;
494 *dest++ = src[1][x] << scale_high | src[1][x] >> scale_low;
495 *dest++ = src[2][x] << scale_high | src[2][x] >> scale_low;
496 *dest++ = src[3][x] << scale_high | src[3][x] >> scale_low;
499 for (x = 0; x < width; x++) {
500 *dest++ = src[0][x] << scale_high | src[0][x] >> scale_low;
501 *dest++ = src[1][x] << scale_high | src[1][x] >> scale_low;
502 *dest++ = src[2][x] << scale_high | src[2][x] >> scale_low;
506 for (i = 0; i < 3 + src_alpha; i++)
507 src[i] += srcStride[i] >> 1;
511 static int planarRgb16ToRgb16Wrapper(SwsContext *c, const uint8_t *src[],
512 int srcStride[], int srcSliceY, int srcSliceH,
513 uint8_t *dst[], int dstStride[])
515 const uint16_t *src102[] = { (uint16_t *)src[1], (uint16_t *)src[0], (uint16_t *)src[2], (uint16_t *)src[3] };
516 const uint16_t *src201[] = { (uint16_t *)src[2], (uint16_t *)src[0], (uint16_t *)src[1], (uint16_t *)src[3] };
517 int stride102[] = { srcStride[1], srcStride[0], srcStride[2], srcStride[3] };
518 int stride201[] = { srcStride[2], srcStride[0], srcStride[1], srcStride[3] };
519 const AVPixFmtDescriptor *src_format = av_pix_fmt_desc_get(c->srcFormat);
520 const AVPixFmtDescriptor *dst_format = av_pix_fmt_desc_get(c->dstFormat);
521 int bits_per_sample = src_format->comp[0].depth_minus1 + 1;
523 if ( HAVE_BIGENDIAN && !(src_format->flags & AV_PIX_FMT_FLAG_BE) ||
524 !HAVE_BIGENDIAN && src_format->flags & AV_PIX_FMT_FLAG_BE)
526 if ( HAVE_BIGENDIAN && !(dst_format->flags & AV_PIX_FMT_FLAG_BE) ||
527 !HAVE_BIGENDIAN && dst_format->flags & AV_PIX_FMT_FLAG_BE)
530 if ((src_format->flags & (AV_PIX_FMT_FLAG_PLANAR | AV_PIX_FMT_FLAG_RGB)) !=
531 (AV_PIX_FMT_FLAG_PLANAR | AV_PIX_FMT_FLAG_RGB) ||
532 bits_per_sample <= 8) {
533 av_log(c, AV_LOG_ERROR, "unsupported planar RGB conversion %s -> %s\n",
534 src_format->name, dst_format->name);
537 switch (c->dstFormat) {
538 case AV_PIX_FMT_BGR48LE:
539 case AV_PIX_FMT_BGR48BE:
540 gbr16ptopacked16(src102, stride102,
541 dst[0] + srcSliceY * dstStride[0], dstStride[0],
542 srcSliceH, 0, swap, bits_per_sample, c->srcW);
544 case AV_PIX_FMT_RGB48LE:
545 case AV_PIX_FMT_RGB48BE:
546 gbr16ptopacked16(src201, stride201,
547 dst[0] + srcSliceY * dstStride[0], dstStride[0],
548 srcSliceH, 0, swap, bits_per_sample, c->srcW);
550 case AV_PIX_FMT_RGBA64LE:
551 case AV_PIX_FMT_RGBA64BE:
552 gbr16ptopacked16(src201, stride201,
553 dst[0] + srcSliceY * dstStride[0], dstStride[0],
554 srcSliceH, 1, swap, bits_per_sample, c->srcW);
556 case AV_PIX_FMT_BGRA64LE:
557 case AV_PIX_FMT_BGRA64BE:
558 gbr16ptopacked16(src102, stride102,
559 dst[0] + srcSliceY * dstStride[0], dstStride[0],
560 srcSliceH, 1, swap, bits_per_sample, c->srcW);
563 av_log(c, AV_LOG_ERROR,
564 "unsupported planar RGB conversion %s -> %s\n",
565 src_format->name, dst_format->name);
571 static void gbr24ptopacked24(const uint8_t *src[], int srcStride[],
572 uint8_t *dst, int dstStride, int srcSliceH,
576 for (h = 0; h < srcSliceH; h++) {
577 uint8_t *dest = dst + dstStride * h;
578 for (x = 0; x < width; x++) {
584 for (i = 0; i < 3; i++)
585 src[i] += srcStride[i];
589 static void gbr24ptopacked32(const uint8_t *src[], int srcStride[],
590 uint8_t *dst, int dstStride, int srcSliceH,
591 int alpha_first, int width)
594 for (h = 0; h < srcSliceH; h++) {
595 uint8_t *dest = dst + dstStride * h;
598 for (x = 0; x < width; x++) {
605 for (x = 0; x < width; x++) {
613 for (i = 0; i < 3; i++)
614 src[i] += srcStride[i];
618 static int planarRgbToRgbWrapper(SwsContext *c, const uint8_t *src[],
619 int srcStride[], int srcSliceY, int srcSliceH,
620 uint8_t *dst[], int dstStride[])
623 const uint8_t *src102[] = { src[1], src[0], src[2] };
624 const uint8_t *src201[] = { src[2], src[0], src[1] };
625 int stride102[] = { srcStride[1], srcStride[0], srcStride[2] };
626 int stride201[] = { srcStride[2], srcStride[0], srcStride[1] };
628 if (c->srcFormat != AV_PIX_FMT_GBRP) {
629 av_log(c, AV_LOG_ERROR, "unsupported planar RGB conversion %s -> %s\n",
630 av_get_pix_fmt_name(c->srcFormat),
631 av_get_pix_fmt_name(c->dstFormat));
635 switch (c->dstFormat) {
636 case AV_PIX_FMT_BGR24:
637 gbr24ptopacked24(src102, stride102,
638 dst[0] + srcSliceY * dstStride[0], dstStride[0],
642 case AV_PIX_FMT_RGB24:
643 gbr24ptopacked24(src201, stride201,
644 dst[0] + srcSliceY * dstStride[0], dstStride[0],
648 case AV_PIX_FMT_ARGB:
650 case AV_PIX_FMT_RGBA:
651 gbr24ptopacked32(src201, stride201,
652 dst[0] + srcSliceY * dstStride[0], dstStride[0],
653 srcSliceH, alpha_first, c->srcW);
656 case AV_PIX_FMT_ABGR:
658 case AV_PIX_FMT_BGRA:
659 gbr24ptopacked32(src102, stride102,
660 dst[0] + srcSliceY * dstStride[0], dstStride[0],
661 srcSliceH, alpha_first, c->srcW);
665 av_log(c, AV_LOG_ERROR,
666 "unsupported planar RGB conversion %s -> %s\n",
667 av_get_pix_fmt_name(c->srcFormat),
668 av_get_pix_fmt_name(c->dstFormat));
674 static int planarRgbToplanarRgbWrapper(SwsContext *c, const uint8_t *src[],
675 int srcStride[], int srcSliceY, int srcSliceH,
676 uint8_t *dst[], int dstStride[])
678 copyPlane(src[0], srcStride[0], srcSliceY, srcSliceH, c->srcW,
679 dst[0], dstStride[0]);
680 copyPlane(src[1], srcStride[1], srcSliceY, srcSliceH, c->srcW,
681 dst[1], dstStride[1]);
682 copyPlane(src[2], srcStride[2], srcSliceY, srcSliceH, c->srcW,
683 dst[2], dstStride[2]);
685 fillPlane(dst[3], dstStride[3], c->srcW, srcSliceH, srcSliceY, 255);
690 static void packedtogbr24p(const uint8_t *src, int srcStride,
691 uint8_t *dst[], int dstStride[], int srcSliceH,
692 int alpha_first, int inc_size, int width)
704 for (h = 0; h < srcSliceH; h++) {
705 for (x = 0; x < width; x++) {
712 src += srcStride - width * inc_size;
713 dest[0] += dstStride[0];
714 dest[1] += dstStride[1];
715 dest[2] += dstStride[2];
719 static int rgbToPlanarRgbWrapper(SwsContext *c, const uint8_t *src[],
720 int srcStride[], int srcSliceY, int srcSliceH,
721 uint8_t *dst[], int dstStride[])
724 int stride102[] = { dstStride[1], dstStride[0], dstStride[2] };
725 int stride201[] = { dstStride[2], dstStride[0], dstStride[1] };
726 uint8_t *dst102[] = { dst[1] + srcSliceY * dstStride[1],
727 dst[0] + srcSliceY * dstStride[0],
728 dst[2] + srcSliceY * dstStride[2] };
729 uint8_t *dst201[] = { dst[2] + srcSliceY * dstStride[2],
730 dst[0] + srcSliceY * dstStride[0],
731 dst[1] + srcSliceY * dstStride[1] };
733 switch (c->srcFormat) {
734 case AV_PIX_FMT_RGB24:
735 packedtogbr24p((const uint8_t *) src[0], srcStride[0], dst201,
736 stride201, srcSliceH, alpha_first, 3, c->srcW);
738 case AV_PIX_FMT_BGR24:
739 packedtogbr24p((const uint8_t *) src[0], srcStride[0], dst102,
740 stride102, srcSliceH, alpha_first, 3, c->srcW);
742 case AV_PIX_FMT_ARGB:
744 case AV_PIX_FMT_RGBA:
745 packedtogbr24p((const uint8_t *) src[0], srcStride[0], dst201,
746 stride201, srcSliceH, alpha_first, 4, c->srcW);
748 case AV_PIX_FMT_ABGR:
750 case AV_PIX_FMT_BGRA:
751 packedtogbr24p((const uint8_t *) src[0], srcStride[0], dst102,
752 stride102, srcSliceH, alpha_first, 4, c->srcW);
755 av_log(c, AV_LOG_ERROR,
756 "unsupported planar RGB conversion %s -> %s\n",
757 av_get_pix_fmt_name(c->srcFormat),
758 av_get_pix_fmt_name(c->dstFormat));
764 #define isRGBA32(x) ( \
765 (x) == AV_PIX_FMT_ARGB \
766 || (x) == AV_PIX_FMT_RGBA \
767 || (x) == AV_PIX_FMT_BGRA \
768 || (x) == AV_PIX_FMT_ABGR \
771 #define isRGBA64(x) ( \
772 (x) == AV_PIX_FMT_RGBA64LE \
773 || (x) == AV_PIX_FMT_RGBA64BE \
774 || (x) == AV_PIX_FMT_BGRA64LE \
775 || (x) == AV_PIX_FMT_BGRA64BE \
778 #define isRGB48(x) ( \
779 (x) == AV_PIX_FMT_RGB48LE \
780 || (x) == AV_PIX_FMT_RGB48BE \
781 || (x) == AV_PIX_FMT_BGR48LE \
782 || (x) == AV_PIX_FMT_BGR48BE \
785 /* {RGB,BGR}{15,16,24,32,32_1} -> {RGB,BGR}{15,16,24,32} */
786 typedef void (* rgbConvFn) (const uint8_t *, uint8_t *, int);
787 static rgbConvFn findRgbConvFn(SwsContext *c)
789 const enum AVPixelFormat srcFormat = c->srcFormat;
790 const enum AVPixelFormat dstFormat = c->dstFormat;
791 const int srcId = c->srcFormatBpp;
792 const int dstId = c->dstFormatBpp;
793 rgbConvFn conv = NULL;
795 #define IS_NOT_NE(bpp, desc) \
796 (((bpp + 7) >> 3) == 2 && \
797 (!(desc->flags & AV_PIX_FMT_FLAG_BE) != !HAVE_BIGENDIAN))
799 #define CONV_IS(src, dst) (srcFormat == AV_PIX_FMT_##src && dstFormat == AV_PIX_FMT_##dst)
801 if (isRGBA32(srcFormat) && isRGBA32(dstFormat)) {
802 if ( CONV_IS(ABGR, RGBA)
803 || CONV_IS(ARGB, BGRA)
804 || CONV_IS(BGRA, ARGB)
805 || CONV_IS(RGBA, ABGR)) conv = shuffle_bytes_3210;
806 else if (CONV_IS(ABGR, ARGB)
807 || CONV_IS(ARGB, ABGR)) conv = shuffle_bytes_0321;
808 else if (CONV_IS(ABGR, BGRA)
809 || CONV_IS(ARGB, RGBA)) conv = shuffle_bytes_1230;
810 else if (CONV_IS(BGRA, RGBA)
811 || CONV_IS(RGBA, BGRA)) conv = shuffle_bytes_2103;
812 else if (CONV_IS(BGRA, ABGR)
813 || CONV_IS(RGBA, ARGB)) conv = shuffle_bytes_3012;
814 } else if (isRGB48(srcFormat) && isRGB48(dstFormat)) {
815 if (CONV_IS(RGB48LE, BGR48LE)
816 || CONV_IS(BGR48LE, RGB48LE)
817 || CONV_IS(RGB48BE, BGR48BE)
818 || CONV_IS(BGR48BE, RGB48BE)) conv = rgb48tobgr48_nobswap;
819 else if (CONV_IS(RGB48LE, BGR48BE)
820 || CONV_IS(BGR48LE, RGB48BE)
821 || CONV_IS(RGB48BE, BGR48LE)
822 || CONV_IS(BGR48BE, RGB48LE)) conv = rgb48tobgr48_bswap;
823 } else if (isRGBA64(srcFormat) && isRGB48(dstFormat)) {
824 if (CONV_IS(RGBA64LE, BGR48LE)
825 || CONV_IS(BGRA64LE, RGB48LE)
826 || CONV_IS(RGBA64BE, BGR48BE)
827 || CONV_IS(BGRA64BE, RGB48BE)) conv = rgb64tobgr48_nobswap;
828 else if (CONV_IS(RGBA64LE, BGR48BE)
829 || CONV_IS(BGRA64LE, RGB48BE)
830 || CONV_IS(RGBA64BE, BGR48LE)
831 || CONV_IS(BGRA64BE, RGB48LE)) conv = rgb64tobgr48_bswap;
832 else if (CONV_IS(RGBA64LE, RGB48LE)
833 || CONV_IS(BGRA64LE, BGR48LE)
834 || CONV_IS(RGBA64BE, RGB48BE)
835 || CONV_IS(BGRA64BE, BGR48BE)) conv = rgb64to48_nobswap;
836 else if (CONV_IS(RGBA64LE, RGB48BE)
837 || CONV_IS(BGRA64LE, BGR48BE)
838 || CONV_IS(RGBA64BE, RGB48LE)
839 || CONV_IS(BGRA64BE, BGR48LE)) conv = rgb64to48_bswap;
842 if ((isBGRinInt(srcFormat) && isBGRinInt(dstFormat)) ||
843 (isRGBinInt(srcFormat) && isRGBinInt(dstFormat))) {
844 switch (srcId | (dstId << 16)) {
845 case 0x000F000C: conv = rgb12to15; break;
846 case 0x000F0010: conv = rgb16to15; break;
847 case 0x000F0018: conv = rgb24to15; break;
848 case 0x000F0020: conv = rgb32to15; break;
849 case 0x0010000F: conv = rgb15to16; break;
850 case 0x00100018: conv = rgb24to16; break;
851 case 0x00100020: conv = rgb32to16; break;
852 case 0x0018000F: conv = rgb15to24; break;
853 case 0x00180010: conv = rgb16to24; break;
854 case 0x00180020: conv = rgb32to24; break;
855 case 0x0020000F: conv = rgb15to32; break;
856 case 0x00200010: conv = rgb16to32; break;
857 case 0x00200018: conv = rgb24to32; break;
859 } else if ((isBGRinInt(srcFormat) && isRGBinInt(dstFormat)) ||
860 (isRGBinInt(srcFormat) && isBGRinInt(dstFormat))) {
861 switch (srcId | (dstId << 16)) {
862 case 0x000C000C: conv = rgb12tobgr12; break;
863 case 0x000F000F: conv = rgb15tobgr15; break;
864 case 0x000F0010: conv = rgb16tobgr15; break;
865 case 0x000F0018: conv = rgb24tobgr15; break;
866 case 0x000F0020: conv = rgb32tobgr15; break;
867 case 0x0010000F: conv = rgb15tobgr16; break;
868 case 0x00100010: conv = rgb16tobgr16; break;
869 case 0x00100018: conv = rgb24tobgr16; break;
870 case 0x00100020: conv = rgb32tobgr16; break;
871 case 0x0018000F: conv = rgb15tobgr24; break;
872 case 0x00180010: conv = rgb16tobgr24; break;
873 case 0x00180018: conv = rgb24tobgr24; break;
874 case 0x00180020: conv = rgb32tobgr24; break;
875 case 0x0020000F: conv = rgb15tobgr32; break;
876 case 0x00200010: conv = rgb16tobgr32; break;
877 case 0x00200018: conv = rgb24tobgr32; break;
881 if ((dstFormat == AV_PIX_FMT_RGB32_1 || dstFormat == AV_PIX_FMT_BGR32_1) && !isRGBA32(srcFormat) && ALT32_CORR<0)
887 /* {RGB,BGR}{15,16,24,32,32_1} -> {RGB,BGR}{15,16,24,32} */
888 static int rgbToRgbWrapper(SwsContext *c, const uint8_t *src[], int srcStride[],
889 int srcSliceY, int srcSliceH, uint8_t *dst[],
893 const enum AVPixelFormat srcFormat = c->srcFormat;
894 const enum AVPixelFormat dstFormat = c->dstFormat;
895 const AVPixFmtDescriptor *desc_src = av_pix_fmt_desc_get(c->srcFormat);
896 const AVPixFmtDescriptor *desc_dst = av_pix_fmt_desc_get(c->dstFormat);
897 const int srcBpp = (c->srcFormatBpp + 7) >> 3;
898 const int dstBpp = (c->dstFormatBpp + 7) >> 3;
899 rgbConvFn conv = findRgbConvFn(c);
902 av_log(c, AV_LOG_ERROR, "internal error %s -> %s converter\n",
903 av_get_pix_fmt_name(srcFormat), av_get_pix_fmt_name(dstFormat));
905 const uint8_t *srcPtr = src[0];
906 uint8_t *dstPtr = dst[0];
907 int src_bswap = IS_NOT_NE(c->srcFormatBpp, desc_src);
908 int dst_bswap = IS_NOT_NE(c->dstFormatBpp, desc_dst);
910 if ((srcFormat == AV_PIX_FMT_RGB32_1 || srcFormat == AV_PIX_FMT_BGR32_1) &&
911 !isRGBA32(dstFormat))
912 srcPtr += ALT32_CORR;
914 if ((dstFormat == AV_PIX_FMT_RGB32_1 || dstFormat == AV_PIX_FMT_BGR32_1) &&
915 !isRGBA32(srcFormat)) {
917 av_assert0(ALT32_CORR == 1);
918 for (i = 0; i < srcSliceH; i++)
919 dstPtr[dstStride[0] * (srcSliceY + i)] = 255;
920 dstPtr += ALT32_CORR;
923 if (dstStride[0] * srcBpp == srcStride[0] * dstBpp && srcStride[0] > 0 &&
924 !(srcStride[0] % srcBpp) && !dst_bswap && !src_bswap)
925 conv(srcPtr, dstPtr + dstStride[0] * srcSliceY,
926 srcSliceH * srcStride[0]);
929 dstPtr += dstStride[0] * srcSliceY;
931 for (i = 0; i < srcSliceH; i++) {
933 for(j=0; j<c->srcW; j++)
934 ((uint16_t*)c->formatConvBuffer)[j] = av_bswap16(((uint16_t*)srcPtr)[j]);
935 conv(c->formatConvBuffer, dstPtr, c->srcW * srcBpp);
937 conv(srcPtr, dstPtr, c->srcW * srcBpp);
939 for(j=0; j<c->srcW; j++)
940 ((uint16_t*)dstPtr)[j] = av_bswap16(((uint16_t*)dstPtr)[j]);
941 srcPtr += srcStride[0];
942 dstPtr += dstStride[0];
949 static int bgr24ToYv12Wrapper(SwsContext *c, const uint8_t *src[],
950 int srcStride[], int srcSliceY, int srcSliceH,
951 uint8_t *dst[], int dstStride[])
955 dst[0] + srcSliceY * dstStride[0],
956 dst[1] + (srcSliceY >> 1) * dstStride[1],
957 dst[2] + (srcSliceY >> 1) * dstStride[2],
959 dstStride[0], dstStride[1], srcStride[0],
960 c->input_rgb2yuv_table);
962 fillPlane(dst[3], dstStride[3], c->srcW, srcSliceH, srcSliceY, 255);
966 static int yvu9ToYv12Wrapper(SwsContext *c, const uint8_t *src[],
967 int srcStride[], int srcSliceY, int srcSliceH,
968 uint8_t *dst[], int dstStride[])
970 copyPlane(src[0], srcStride[0], srcSliceY, srcSliceH, c->srcW,
971 dst[0], dstStride[0]);
973 planar2x(src[1], dst[1] + dstStride[1] * (srcSliceY >> 1), c->chrSrcW,
974 srcSliceH >> 2, srcStride[1], dstStride[1]);
975 planar2x(src[2], dst[2] + dstStride[2] * (srcSliceY >> 1), c->chrSrcW,
976 srcSliceH >> 2, srcStride[2], dstStride[2]);
978 fillPlane(dst[3], dstStride[3], c->srcW, srcSliceH, srcSliceY, 255);
982 /* unscaled copy like stuff (assumes nearly identical formats) */
983 static int packedCopyWrapper(SwsContext *c, const uint8_t *src[],
984 int srcStride[], int srcSliceY, int srcSliceH,
985 uint8_t *dst[], int dstStride[])
987 if (dstStride[0] == srcStride[0] && srcStride[0] > 0)
988 memcpy(dst[0] + dstStride[0] * srcSliceY, src[0], srcSliceH * dstStride[0]);
991 const uint8_t *srcPtr = src[0];
992 uint8_t *dstPtr = dst[0] + dstStride[0] * srcSliceY;
995 /* universal length finder */
996 while (length + c->srcW <= FFABS(dstStride[0]) &&
997 length + c->srcW <= FFABS(srcStride[0]))
999 av_assert1(length != 0);
1001 for (i = 0; i < srcSliceH; i++) {
1002 memcpy(dstPtr, srcPtr, length);
1003 srcPtr += srcStride[0];
1004 dstPtr += dstStride[0];
1010 #define DITHER_COPY(dst, dstStride, src, srcStride, bswap, dbswap)\
1011 uint16_t scale= dither_scale[dst_depth-1][src_depth-1];\
1012 int shift= src_depth-dst_depth + dither_scale[src_depth-2][dst_depth-1];\
1013 for (i = 0; i < height; i++) {\
1014 const uint8_t *dither= dithers[src_depth-9][i&7];\
1015 for (j = 0; j < length-7; j+=8){\
1016 dst[j+0] = dbswap((bswap(src[j+0]) + dither[0])*scale>>shift);\
1017 dst[j+1] = dbswap((bswap(src[j+1]) + dither[1])*scale>>shift);\
1018 dst[j+2] = dbswap((bswap(src[j+2]) + dither[2])*scale>>shift);\
1019 dst[j+3] = dbswap((bswap(src[j+3]) + dither[3])*scale>>shift);\
1020 dst[j+4] = dbswap((bswap(src[j+4]) + dither[4])*scale>>shift);\
1021 dst[j+5] = dbswap((bswap(src[j+5]) + dither[5])*scale>>shift);\
1022 dst[j+6] = dbswap((bswap(src[j+6]) + dither[6])*scale>>shift);\
1023 dst[j+7] = dbswap((bswap(src[j+7]) + dither[7])*scale>>shift);\
1025 for (; j < length; j++)\
1026 dst[j] = dbswap((bswap(src[j]) + dither[j&7])*scale>>shift);\
1031 static int planarCopyWrapper(SwsContext *c, const uint8_t *src[],
1032 int srcStride[], int srcSliceY, int srcSliceH,
1033 uint8_t *dst[], int dstStride[])
1035 const AVPixFmtDescriptor *desc_src = av_pix_fmt_desc_get(c->srcFormat);
1036 const AVPixFmtDescriptor *desc_dst = av_pix_fmt_desc_get(c->dstFormat);
1038 for (plane = 0; plane < 4; plane++) {
1039 int length = (plane == 0 || plane == 3) ? c->srcW : FF_CEIL_RSHIFT(c->srcW, c->chrDstHSubSample);
1040 int y = (plane == 0 || plane == 3) ? srcSliceY: FF_CEIL_RSHIFT(srcSliceY, c->chrDstVSubSample);
1041 int height = (plane == 0 || plane == 3) ? srcSliceH: FF_CEIL_RSHIFT(srcSliceH, c->chrDstVSubSample);
1042 const uint8_t *srcPtr = src[plane];
1043 uint8_t *dstPtr = dst[plane] + dstStride[plane] * y;
1044 int shiftonly= plane==1 || plane==2 || (!c->srcRange && plane==0);
1048 // ignore palette for GRAY8
1049 if (plane == 1 && !dst[2]) continue;
1050 if (!src[plane] || (plane == 1 && !src[2])) {
1051 if (is16BPS(c->dstFormat) || isNBPS(c->dstFormat)) {
1052 fillPlane16(dst[plane], dstStride[plane], length, height, y,
1053 plane == 3, desc_dst->comp[plane].depth_minus1,
1054 isBE(c->dstFormat));
1056 fillPlane(dst[plane], dstStride[plane], length, height, y,
1057 (plane == 3) ? 255 : 128);
1060 if(isNBPS(c->srcFormat) || isNBPS(c->dstFormat)
1061 || (is16BPS(c->srcFormat) != is16BPS(c->dstFormat))
1063 const int src_depth = desc_src->comp[plane].depth_minus1 + 1;
1064 const int dst_depth = desc_dst->comp[plane].depth_minus1 + 1;
1065 const uint16_t *srcPtr2 = (const uint16_t *) srcPtr;
1066 uint16_t *dstPtr2 = (uint16_t*)dstPtr;
1068 if (dst_depth == 8) {
1069 if(isBE(c->srcFormat) == HAVE_BIGENDIAN){
1070 DITHER_COPY(dstPtr, dstStride[plane], srcPtr2, srcStride[plane]/2, , )
1072 DITHER_COPY(dstPtr, dstStride[plane], srcPtr2, srcStride[plane]/2, av_bswap16, )
1074 } else if (src_depth == 8) {
1075 for (i = 0; i < height; i++) {
1078 for (j = 0; j < length; j++)\
1079 w(&dstPtr2[j], srcPtr[j]<<(dst_depth-8));\
1081 for (j = 0; j < length; j++)\
1082 w(&dstPtr2[j], (srcPtr[j]<<(dst_depth-8)) |\
1083 (srcPtr[j]>>(2*8-dst_depth)));\
1085 if(isBE(c->dstFormat)){
1090 dstPtr2 += dstStride[plane]/2;
1091 srcPtr += srcStride[plane];
1093 } else if (src_depth <= dst_depth) {
1094 int orig_length = length;
1095 for (i = 0; i < height; i++) {
1096 if(isBE(c->srcFormat) == HAVE_BIGENDIAN &&
1097 isBE(c->dstFormat) == HAVE_BIGENDIAN &&
1099 unsigned shift = dst_depth - src_depth;
1100 length = orig_length;
1102 #define FAST_COPY_UP(shift) \
1103 for (j = 0; j < length - 3; j += 4) { \
1104 uint64_t v = AV_RN64A(srcPtr2 + j); \
1105 AV_WN64A(dstPtr2 + j, v << shift); \
1109 #define FAST_COPY_UP(shift) \
1110 for (j = 0; j < length - 1; j += 2) { \
1111 uint32_t v = AV_RN32A(srcPtr2 + j); \
1112 AV_WN32A(dstPtr2 + j, v << shift); \
1118 case 6: FAST_COPY_UP(6); break;
1119 case 7: FAST_COPY_UP(7); break;
1122 #define COPY_UP(r,w) \
1124 for (j = 0; j < length; j++){ \
1125 unsigned int v= r(&srcPtr2[j]);\
1126 w(&dstPtr2[j], v<<(dst_depth-src_depth));\
1129 for (j = 0; j < length; j++){ \
1130 unsigned int v= r(&srcPtr2[j]);\
1131 w(&dstPtr2[j], (v<<(dst_depth-src_depth)) | \
1132 (v>>(2*src_depth-dst_depth)));\
1135 if(isBE(c->srcFormat)){
1136 if(isBE(c->dstFormat)){
1137 COPY_UP(AV_RB16, AV_WB16)
1139 COPY_UP(AV_RB16, AV_WL16)
1142 if(isBE(c->dstFormat)){
1143 COPY_UP(AV_RL16, AV_WB16)
1145 COPY_UP(AV_RL16, AV_WL16)
1148 dstPtr2 += dstStride[plane]/2;
1149 srcPtr2 += srcStride[plane]/2;
1152 if(isBE(c->srcFormat) == HAVE_BIGENDIAN){
1153 if(isBE(c->dstFormat) == HAVE_BIGENDIAN){
1154 DITHER_COPY(dstPtr2, dstStride[plane]/2, srcPtr2, srcStride[plane]/2, , )
1156 DITHER_COPY(dstPtr2, dstStride[plane]/2, srcPtr2, srcStride[plane]/2, , av_bswap16)
1159 if(isBE(c->dstFormat) == HAVE_BIGENDIAN){
1160 DITHER_COPY(dstPtr2, dstStride[plane]/2, srcPtr2, srcStride[plane]/2, av_bswap16, )
1162 DITHER_COPY(dstPtr2, dstStride[plane]/2, srcPtr2, srcStride[plane]/2, av_bswap16, av_bswap16)
1166 } else if (is16BPS(c->srcFormat) && is16BPS(c->dstFormat) &&
1167 isBE(c->srcFormat) != isBE(c->dstFormat)) {
1169 for (i = 0; i < height; i++) {
1170 for (j = 0; j < length; j++)
1171 ((uint16_t *) dstPtr)[j] = av_bswap16(((const uint16_t *) srcPtr)[j]);
1172 srcPtr += srcStride[plane];
1173 dstPtr += dstStride[plane];
1175 } else if (dstStride[plane] == srcStride[plane] &&
1176 srcStride[plane] > 0 && srcStride[plane] == length) {
1177 memcpy(dst[plane] + dstStride[plane] * y, src[plane],
1178 height * dstStride[plane]);
1180 if (is16BPS(c->srcFormat) && is16BPS(c->dstFormat))
1182 else if (!desc_src->comp[0].depth_minus1)
1183 length >>= 3; // monowhite/black
1184 for (i = 0; i < height; i++) {
1185 memcpy(dstPtr, srcPtr, length);
1186 srcPtr += srcStride[plane];
1187 dstPtr += dstStride[plane];
1196 #define IS_DIFFERENT_ENDIANESS(src_fmt, dst_fmt, pix_fmt) \
1197 ((src_fmt == pix_fmt ## BE && dst_fmt == pix_fmt ## LE) || \
1198 (src_fmt == pix_fmt ## LE && dst_fmt == pix_fmt ## BE))
1201 void ff_get_unscaled_swscale(SwsContext *c)
1203 const enum AVPixelFormat srcFormat = c->srcFormat;
1204 const enum AVPixelFormat dstFormat = c->dstFormat;
1205 const int flags = c->flags;
1206 const int dstH = c->dstH;
1209 needsDither = isAnyRGB(dstFormat) &&
1210 c->dstFormatBpp < 24 &&
1211 (c->dstFormatBpp < c->srcFormatBpp || (!isAnyRGB(srcFormat)));
1214 if ((srcFormat == AV_PIX_FMT_YUV420P || srcFormat == AV_PIX_FMT_YUVA420P) &&
1215 (dstFormat == AV_PIX_FMT_NV12 || dstFormat == AV_PIX_FMT_NV21)) {
1216 c->swscale = planarToNv12Wrapper;
1219 if ((srcFormat == AV_PIX_FMT_YUV420P || srcFormat == AV_PIX_FMT_YUV422P ||
1220 srcFormat == AV_PIX_FMT_YUVA420P) && isAnyRGB(dstFormat) &&
1221 !(flags & SWS_ACCURATE_RND) && (c->dither == SWS_DITHER_BAYER || c->dither == SWS_DITHER_AUTO) && !(dstH & 1)) {
1222 c->swscale = ff_yuv2rgb_get_func_ptr(c);
1225 if (srcFormat == AV_PIX_FMT_YUV410P &&
1226 (dstFormat == AV_PIX_FMT_YUV420P || dstFormat == AV_PIX_FMT_YUVA420P) &&
1227 !(flags & SWS_BITEXACT)) {
1228 c->swscale = yvu9ToYv12Wrapper;
1232 if (srcFormat == AV_PIX_FMT_BGR24 &&
1233 (dstFormat == AV_PIX_FMT_YUV420P || dstFormat == AV_PIX_FMT_YUVA420P) &&
1234 !(flags & SWS_ACCURATE_RND))
1235 c->swscale = bgr24ToYv12Wrapper;
1237 /* RGB/BGR -> RGB/BGR (no dither needed forms) */
1238 if (isAnyRGB(srcFormat) && isAnyRGB(dstFormat) && findRgbConvFn(c)
1239 && (!needsDither || (c->flags&(SWS_FAST_BILINEAR|SWS_POINT))))
1240 c->swscale = rgbToRgbWrapper;
1242 if ((srcFormat == AV_PIX_FMT_GBRP && dstFormat == AV_PIX_FMT_GBRAP) ||
1243 (srcFormat == AV_PIX_FMT_GBRAP && dstFormat == AV_PIX_FMT_GBRP))
1244 c->swscale = planarRgbToplanarRgbWrapper;
1246 #define isByteRGB(f) ( \
1247 f == AV_PIX_FMT_RGB32 || \
1248 f == AV_PIX_FMT_RGB32_1 || \
1249 f == AV_PIX_FMT_RGB24 || \
1250 f == AV_PIX_FMT_BGR32 || \
1251 f == AV_PIX_FMT_BGR32_1 || \
1252 f == AV_PIX_FMT_BGR24)
1254 if (srcFormat == AV_PIX_FMT_GBRP && isPlanar(srcFormat) && isByteRGB(dstFormat))
1255 c->swscale = planarRgbToRgbWrapper;
1257 if ((srcFormat == AV_PIX_FMT_GBRP9LE || srcFormat == AV_PIX_FMT_GBRP9BE ||
1258 srcFormat == AV_PIX_FMT_GBRP16LE || srcFormat == AV_PIX_FMT_GBRP16BE ||
1259 srcFormat == AV_PIX_FMT_GBRP10LE || srcFormat == AV_PIX_FMT_GBRP10BE ||
1260 srcFormat == AV_PIX_FMT_GBRP12LE || srcFormat == AV_PIX_FMT_GBRP12BE ||
1261 srcFormat == AV_PIX_FMT_GBRP14LE || srcFormat == AV_PIX_FMT_GBRP14BE ||
1262 srcFormat == AV_PIX_FMT_GBRAP16LE || srcFormat == AV_PIX_FMT_GBRAP16BE) &&
1263 (dstFormat == AV_PIX_FMT_RGB48LE || dstFormat == AV_PIX_FMT_RGB48BE ||
1264 dstFormat == AV_PIX_FMT_BGR48LE || dstFormat == AV_PIX_FMT_BGR48BE ||
1265 dstFormat == AV_PIX_FMT_RGBA64LE || dstFormat == AV_PIX_FMT_RGBA64BE ||
1266 dstFormat == AV_PIX_FMT_BGRA64LE || dstFormat == AV_PIX_FMT_BGRA64BE))
1267 c->swscale = planarRgb16ToRgb16Wrapper;
1269 if (av_pix_fmt_desc_get(srcFormat)->comp[0].depth_minus1 == 7 &&
1270 isPackedRGB(srcFormat) && dstFormat == AV_PIX_FMT_GBRP)
1271 c->swscale = rgbToPlanarRgbWrapper;
1273 /* bswap 16 bits per pixel/component packed formats */
1274 if (IS_DIFFERENT_ENDIANESS(srcFormat, dstFormat, AV_PIX_FMT_BGR444) ||
1275 IS_DIFFERENT_ENDIANESS(srcFormat, dstFormat, AV_PIX_FMT_BGR48) ||
1276 IS_DIFFERENT_ENDIANESS(srcFormat, dstFormat, AV_PIX_FMT_BGRA64) ||
1277 IS_DIFFERENT_ENDIANESS(srcFormat, dstFormat, AV_PIX_FMT_BGR555) ||
1278 IS_DIFFERENT_ENDIANESS(srcFormat, dstFormat, AV_PIX_FMT_BGR565) ||
1279 IS_DIFFERENT_ENDIANESS(srcFormat, dstFormat, AV_PIX_FMT_GRAY16) ||
1280 IS_DIFFERENT_ENDIANESS(srcFormat, dstFormat, AV_PIX_FMT_GBRP9) ||
1281 IS_DIFFERENT_ENDIANESS(srcFormat, dstFormat, AV_PIX_FMT_GBRP10) ||
1282 IS_DIFFERENT_ENDIANESS(srcFormat, dstFormat, AV_PIX_FMT_GBRP12) ||
1283 IS_DIFFERENT_ENDIANESS(srcFormat, dstFormat, AV_PIX_FMT_GBRP14) ||
1284 IS_DIFFERENT_ENDIANESS(srcFormat, dstFormat, AV_PIX_FMT_GBRP16) ||
1285 IS_DIFFERENT_ENDIANESS(srcFormat, dstFormat, AV_PIX_FMT_GBRAP16) ||
1286 IS_DIFFERENT_ENDIANESS(srcFormat, dstFormat, AV_PIX_FMT_RGB444) ||
1287 IS_DIFFERENT_ENDIANESS(srcFormat, dstFormat, AV_PIX_FMT_RGB48) ||
1288 IS_DIFFERENT_ENDIANESS(srcFormat, dstFormat, AV_PIX_FMT_RGBA64) ||
1289 IS_DIFFERENT_ENDIANESS(srcFormat, dstFormat, AV_PIX_FMT_RGB555) ||
1290 IS_DIFFERENT_ENDIANESS(srcFormat, dstFormat, AV_PIX_FMT_RGB565) ||
1291 IS_DIFFERENT_ENDIANESS(srcFormat, dstFormat, AV_PIX_FMT_XYZ12) ||
1292 IS_DIFFERENT_ENDIANESS(srcFormat, dstFormat, AV_PIX_FMT_YUV420P9) ||
1293 IS_DIFFERENT_ENDIANESS(srcFormat, dstFormat, AV_PIX_FMT_YUV420P10) ||
1294 IS_DIFFERENT_ENDIANESS(srcFormat, dstFormat, AV_PIX_FMT_YUV420P12) ||
1295 IS_DIFFERENT_ENDIANESS(srcFormat, dstFormat, AV_PIX_FMT_YUV420P14) ||
1296 IS_DIFFERENT_ENDIANESS(srcFormat, dstFormat, AV_PIX_FMT_YUV420P16) ||
1297 IS_DIFFERENT_ENDIANESS(srcFormat, dstFormat, AV_PIX_FMT_YUV422P9) ||
1298 IS_DIFFERENT_ENDIANESS(srcFormat, dstFormat, AV_PIX_FMT_YUV422P10) ||
1299 IS_DIFFERENT_ENDIANESS(srcFormat, dstFormat, AV_PIX_FMT_YUV422P12) ||
1300 IS_DIFFERENT_ENDIANESS(srcFormat, dstFormat, AV_PIX_FMT_YUV422P14) ||
1301 IS_DIFFERENT_ENDIANESS(srcFormat, dstFormat, AV_PIX_FMT_YUV422P16) ||
1302 IS_DIFFERENT_ENDIANESS(srcFormat, dstFormat, AV_PIX_FMT_YUV444P9) ||
1303 IS_DIFFERENT_ENDIANESS(srcFormat, dstFormat, AV_PIX_FMT_YUV444P10) ||
1304 IS_DIFFERENT_ENDIANESS(srcFormat, dstFormat, AV_PIX_FMT_YUV444P12) ||
1305 IS_DIFFERENT_ENDIANESS(srcFormat, dstFormat, AV_PIX_FMT_YUV444P14) ||
1306 IS_DIFFERENT_ENDIANESS(srcFormat, dstFormat, AV_PIX_FMT_YUV444P16))
1307 c->swscale = packed_16bpc_bswap;
1309 if (usePal(srcFormat) && isByteRGB(dstFormat))
1310 c->swscale = palToRgbWrapper;
1312 if (srcFormat == AV_PIX_FMT_YUV422P) {
1313 if (dstFormat == AV_PIX_FMT_YUYV422)
1314 c->swscale = yuv422pToYuy2Wrapper;
1315 else if (dstFormat == AV_PIX_FMT_UYVY422)
1316 c->swscale = yuv422pToUyvyWrapper;
1319 /* LQ converters if -sws 0 or -sws 4*/
1320 if (c->flags&(SWS_FAST_BILINEAR|SWS_POINT)) {
1322 if (srcFormat == AV_PIX_FMT_YUV420P || srcFormat == AV_PIX_FMT_YUVA420P) {
1323 if (dstFormat == AV_PIX_FMT_YUYV422)
1324 c->swscale = planarToYuy2Wrapper;
1325 else if (dstFormat == AV_PIX_FMT_UYVY422)
1326 c->swscale = planarToUyvyWrapper;
1329 if (srcFormat == AV_PIX_FMT_YUYV422 &&
1330 (dstFormat == AV_PIX_FMT_YUV420P || dstFormat == AV_PIX_FMT_YUVA420P))
1331 c->swscale = yuyvToYuv420Wrapper;
1332 if (srcFormat == AV_PIX_FMT_UYVY422 &&
1333 (dstFormat == AV_PIX_FMT_YUV420P || dstFormat == AV_PIX_FMT_YUVA420P))
1334 c->swscale = uyvyToYuv420Wrapper;
1335 if (srcFormat == AV_PIX_FMT_YUYV422 && dstFormat == AV_PIX_FMT_YUV422P)
1336 c->swscale = yuyvToYuv422Wrapper;
1337 if (srcFormat == AV_PIX_FMT_UYVY422 && dstFormat == AV_PIX_FMT_YUV422P)
1338 c->swscale = uyvyToYuv422Wrapper;
1340 #define isPlanarGray(x) (isGray(x) && (x) != AV_PIX_FMT_GRAY8A)
1342 if ( srcFormat == dstFormat ||
1343 (srcFormat == AV_PIX_FMT_YUVA420P && dstFormat == AV_PIX_FMT_YUV420P) ||
1344 (srcFormat == AV_PIX_FMT_YUV420P && dstFormat == AV_PIX_FMT_YUVA420P) ||
1345 (isPlanarYUV(srcFormat) && isPlanarGray(dstFormat)) ||
1346 (isPlanarYUV(dstFormat) && isPlanarGray(srcFormat)) ||
1347 (isPlanarGray(dstFormat) && isPlanarGray(srcFormat)) ||
1348 (isPlanarYUV(srcFormat) && isPlanarYUV(dstFormat) &&
1349 c->chrDstHSubSample == c->chrSrcHSubSample &&
1350 c->chrDstVSubSample == c->chrSrcVSubSample &&
1351 dstFormat != AV_PIX_FMT_NV12 && dstFormat != AV_PIX_FMT_NV21 &&
1352 srcFormat != AV_PIX_FMT_NV12 && srcFormat != AV_PIX_FMT_NV21))
1354 if (isPacked(c->srcFormat))
1355 c->swscale = packedCopyWrapper;
1356 else /* Planar YUV or gray */
1357 c->swscale = planarCopyWrapper;
1361 ff_get_unscaled_swscale_bfin(c);
1363 ff_get_unscaled_swscale_ppc(c);
1366 /* Convert the palette to the same packed 32-bit format as the palette */
1367 void sws_convertPalette8ToPacked32(const uint8_t *src, uint8_t *dst,
1368 int num_pixels, const uint8_t *palette)
1372 for (i = 0; i < num_pixels; i++)
1373 ((uint32_t *) dst)[i] = ((const uint32_t *) palette)[src[i]];
1376 /* Palette format: ABCD -> dst format: ABC */
1377 void sws_convertPalette8ToPacked24(const uint8_t *src, uint8_t *dst,
1378 int num_pixels, const uint8_t *palette)
1382 for (i = 0; i < num_pixels; i++) {
1384 dst[0] = palette[src[i] * 4 + 0];
1385 dst[1] = palette[src[i] * 4 + 1];
1386 dst[2] = palette[src[i] * 4 + 2];