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)
9 * This file is part of FFmpeg.
11 * FFmpeg is free software; you can redistribute it and/or
12 * modify it under the terms of the GNU Lesser General Public
13 * License as published by the Free Software Foundation; either
14 * version 2.1 of the License, or (at your option) any later version.
16 * FFmpeg is distributed in the hope that it will be useful,
17 * but WITHOUT ANY WARRANTY; without even the implied warranty of
18 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
19 * Lesser General Public License for more details.
21 * You should have received a copy of the GNU Lesser General Public
22 * License along with FFmpeg; if not, write to the Free Software
23 * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
28 #include "libavutil/attributes.h"
29 #include "libavutil/bswap.h"
33 #include "swscale_internal.h"
35 void (*rgb32tobgr24)(const uint8_t *src, uint8_t *dst, int src_size);
36 void (*rgb32tobgr16)(const uint8_t *src, uint8_t *dst, int src_size);
37 void (*rgb32tobgr15)(const uint8_t *src, uint8_t *dst, int src_size);
38 void (*rgb24tobgr32)(const uint8_t *src, uint8_t *dst, int src_size);
39 void (*rgb24tobgr24)(const uint8_t *src, uint8_t *dst, int src_size);
40 void (*rgb24tobgr16)(const uint8_t *src, uint8_t *dst, int src_size);
41 void (*rgb24tobgr15)(const uint8_t *src, uint8_t *dst, int src_size);
42 void (*rgb16tobgr24)(const uint8_t *src, uint8_t *dst, int src_size);
43 void (*rgb15tobgr24)(const uint8_t *src, uint8_t *dst, int src_size);
45 void (*rgb32to16)(const uint8_t *src, uint8_t *dst, int src_size);
46 void (*rgb32to15)(const uint8_t *src, uint8_t *dst, int src_size);
47 void (*rgb24to16)(const uint8_t *src, uint8_t *dst, int src_size);
48 void (*rgb24to15)(const uint8_t *src, uint8_t *dst, int src_size);
49 void (*rgb16to32)(const uint8_t *src, uint8_t *dst, int src_size);
50 void (*rgb16to15)(const uint8_t *src, uint8_t *dst, int src_size);
51 void (*rgb15to16)(const uint8_t *src, uint8_t *dst, int src_size);
52 void (*rgb15to32)(const uint8_t *src, uint8_t *dst, int src_size);
54 void (*shuffle_bytes_2103)(const uint8_t *src, uint8_t *dst, int src_size);
56 void (*yv12toyuy2)(const uint8_t *ysrc, const uint8_t *usrc,
57 const uint8_t *vsrc, uint8_t *dst,
58 int width, int height,
59 int lumStride, int chromStride, int dstStride);
60 void (*yv12touyvy)(const uint8_t *ysrc, const uint8_t *usrc,
61 const uint8_t *vsrc, uint8_t *dst,
62 int width, int height,
63 int lumStride, int chromStride, int dstStride);
64 void (*yuv422ptoyuy2)(const uint8_t *ysrc, const uint8_t *usrc,
65 const uint8_t *vsrc, uint8_t *dst,
66 int width, int height,
67 int lumStride, int chromStride, int dstStride);
68 void (*yuv422ptouyvy)(const uint8_t *ysrc, const uint8_t *usrc,
69 const uint8_t *vsrc, uint8_t *dst,
70 int width, int height,
71 int lumStride, int chromStride, int dstStride);
72 void (*yuy2toyv12)(const uint8_t *src, uint8_t *ydst,
73 uint8_t *udst, uint8_t *vdst,
74 int width, int height,
75 int lumStride, int chromStride, int srcStride);
76 void (*ff_rgb24toyv12)(const uint8_t *src, uint8_t *ydst,
77 uint8_t *udst, uint8_t *vdst,
78 int width, int height,
79 int lumStride, int chromStride, int srcStride,
81 void (*planar2x)(const uint8_t *src, uint8_t *dst, int width, int height,
82 int srcStride, int dstStride);
83 void (*interleaveBytes)(const uint8_t *src1, const uint8_t *src2, uint8_t *dst,
84 int width, int height, int src1Stride,
85 int src2Stride, int dstStride);
86 void (*vu9_to_vu12)(const uint8_t *src1, const uint8_t *src2,
87 uint8_t *dst1, uint8_t *dst2,
88 int width, int height,
89 int srcStride1, int srcStride2,
90 int dstStride1, int dstStride2);
91 void (*yvu9_to_yuy2)(const uint8_t *src1, const uint8_t *src2,
92 const uint8_t *src3, uint8_t *dst,
93 int width, int height,
94 int srcStride1, int srcStride2,
95 int srcStride3, int dstStride);
96 void (*uyvytoyuv420)(uint8_t *ydst, uint8_t *udst, uint8_t *vdst,
97 const uint8_t *src, int width, int height,
98 int lumStride, int chromStride, int srcStride);
99 void (*uyvytoyuv422)(uint8_t *ydst, uint8_t *udst, uint8_t *vdst,
100 const uint8_t *src, int width, int height,
101 int lumStride, int chromStride, int srcStride);
102 void (*yuyvtoyuv420)(uint8_t *ydst, uint8_t *udst, uint8_t *vdst,
103 const uint8_t *src, int width, int height,
104 int lumStride, int chromStride, int srcStride);
105 void (*yuyvtoyuv422)(uint8_t *ydst, uint8_t *udst, uint8_t *vdst,
106 const uint8_t *src, int width, int height,
107 int lumStride, int chromStride, int srcStride);
109 #define BY ((int)( 0.098 * (1 << RGB2YUV_SHIFT) + 0.5))
110 #define BV ((int)(-0.071 * (1 << RGB2YUV_SHIFT) + 0.5))
111 #define BU ((int)( 0.439 * (1 << RGB2YUV_SHIFT) + 0.5))
112 #define GY ((int)( 0.504 * (1 << RGB2YUV_SHIFT) + 0.5))
113 #define GV ((int)(-0.368 * (1 << RGB2YUV_SHIFT) + 0.5))
114 #define GU ((int)(-0.291 * (1 << RGB2YUV_SHIFT) + 0.5))
115 #define RY ((int)( 0.257 * (1 << RGB2YUV_SHIFT) + 0.5))
116 #define RV ((int)( 0.439 * (1 << RGB2YUV_SHIFT) + 0.5))
117 #define RU ((int)(-0.148 * (1 << RGB2YUV_SHIFT) + 0.5))
120 #include "rgb2rgb_template.c"
123 * RGB15->RGB16 original by Strepto/Astral
124 * ported to gcc & bugfixed : A'rpi
125 * MMXEXT, 3DNOW optimization by Nick Kurshev
126 * 32-bit C version, and and&add trick by Michael Niedermayer
129 av_cold void sws_rgb2rgb_init(void)
136 void rgb32to24(const uint8_t *src, uint8_t *dst, int src_size)
138 int i, num_pixels = src_size >> 2;
140 for (i = 0; i < num_pixels; i++) {
142 /* RGB32 (= A,B,G,R) -> BGR24 (= B,G,R) */
143 dst[3 * i + 0] = src[4 * i + 1];
144 dst[3 * i + 1] = src[4 * i + 2];
145 dst[3 * i + 2] = src[4 * i + 3];
147 dst[3 * i + 0] = src[4 * i + 2];
148 dst[3 * i + 1] = src[4 * i + 1];
149 dst[3 * i + 2] = src[4 * i + 0];
154 void rgb24to32(const uint8_t *src, uint8_t *dst, int src_size)
158 for (i = 0; 3 * i < src_size; i++) {
160 /* RGB24 (= R, G, B) -> BGR32 (= A, R, G, B) */
161 dst[4 * i + 0] = 255;
162 dst[4 * i + 1] = src[3 * i + 0];
163 dst[4 * i + 2] = src[3 * i + 1];
164 dst[4 * i + 3] = src[3 * i + 2];
166 dst[4 * i + 0] = src[3 * i + 2];
167 dst[4 * i + 1] = src[3 * i + 1];
168 dst[4 * i + 2] = src[3 * i + 0];
169 dst[4 * i + 3] = 255;
174 void rgb16tobgr32(const uint8_t *src, uint8_t *dst, int src_size)
177 const uint16_t *s = (const uint16_t *)src;
178 const uint16_t *end = s + src_size / 2;
181 register uint16_t bgr = *s++;
184 *d++ = ((bgr&0x001F)<<3) | ((bgr&0x001F)>> 2);
185 *d++ = ((bgr&0x07E0)>>3) | ((bgr&0x07E0)>> 9);
186 *d++ = ((bgr&0xF800)>>8) | ((bgr&0xF800)>>13);
188 *d++ = ((bgr&0xF800)>>8) | ((bgr&0xF800)>>13);
189 *d++ = ((bgr&0x07E0)>>3) | ((bgr&0x07E0)>> 9);
190 *d++ = ((bgr&0x001F)<<3) | ((bgr&0x001F)>> 2);
196 void rgb12to15(const uint8_t *src, uint8_t *dst, int src_size)
198 uint16_t rgb, r, g, b;
199 uint16_t *d = (uint16_t *)dst;
200 const uint16_t *s = (const uint16_t *)src;
201 const uint16_t *end = s + src_size / 2;
208 r = (r << 3) | ((r & 0x800) >> 1);
209 g = (g << 2) | ((g & 0x080) >> 2);
210 b = (b << 1) | ( b >> 3);
215 void rgb16to24(const uint8_t *src, uint8_t *dst, int src_size)
218 const uint16_t *s = (const uint16_t *)src;
219 const uint16_t *end = s + src_size / 2;
222 register uint16_t bgr = *s++;
223 *d++ = ((bgr&0xF800)>>8) | ((bgr&0xF800)>>13);
224 *d++ = ((bgr&0x07E0)>>3) | ((bgr&0x07E0)>> 9);
225 *d++ = ((bgr&0x001F)<<3) | ((bgr&0x001F)>> 2);
229 void rgb16tobgr16(const uint8_t *src, uint8_t *dst, int src_size)
231 int i, num_pixels = src_size >> 1;
233 for (i = 0; i < num_pixels; i++) {
234 unsigned rgb = ((const uint16_t *)src)[i];
235 ((uint16_t *)dst)[i] = (rgb >> 11) | (rgb & 0x7E0) | (rgb << 11);
239 void rgb16tobgr15(const uint8_t *src, uint8_t *dst, int src_size)
241 int i, num_pixels = src_size >> 1;
243 for (i = 0; i < num_pixels; i++) {
244 unsigned rgb = ((const uint16_t *)src)[i];
245 ((uint16_t *)dst)[i] = (rgb >> 11) | ((rgb & 0x7C0) >> 1) | ((rgb & 0x1F) << 10);
249 void rgb15tobgr32(const uint8_t *src, uint8_t *dst, int src_size)
252 const uint16_t *s = (const uint16_t *)src;
253 const uint16_t *end = s + src_size / 2;
256 register uint16_t bgr = *s++;
259 *d++ = ((bgr&0x001F)<<3) | ((bgr&0x001F)>> 2);
260 *d++ = ((bgr&0x03E0)>>2) | ((bgr&0x03E0)>> 7);
261 *d++ = ((bgr&0x7C00)>>7) | ((bgr&0x7C00)>>12);
263 *d++ = ((bgr&0x7C00)>>7) | ((bgr&0x7C00)>>12);
264 *d++ = ((bgr&0x03E0)>>2) | ((bgr&0x03E0)>> 7);
265 *d++ = ((bgr&0x001F)<<3) | ((bgr&0x001F)>> 2);
271 void rgb15to24(const uint8_t *src, uint8_t *dst, int src_size)
274 const uint16_t *s = (const uint16_t *)src;
275 const uint16_t *end = s + src_size / 2;
278 register uint16_t bgr = *s++;
279 *d++ = ((bgr&0x7C00)>>7) | ((bgr&0x7C00)>>12);
280 *d++ = ((bgr&0x03E0)>>2) | ((bgr&0x03E0)>> 7);
281 *d++ = ((bgr&0x001F)<<3) | ((bgr&0x001F)>> 2);
285 void rgb15tobgr16(const uint8_t *src, uint8_t *dst, int src_size)
287 int i, num_pixels = src_size >> 1;
289 for (i = 0; i < num_pixels; i++) {
290 unsigned rgb = ((const uint16_t *)src)[i];
291 ((uint16_t *)dst)[i] = ((rgb & 0x7C00) >> 10) | ((rgb & 0x3E0) << 1) | (rgb << 11);
295 void rgb15tobgr15(const uint8_t *src, uint8_t *dst, int src_size)
297 int i, num_pixels = src_size >> 1;
299 for (i = 0; i < num_pixels; i++) {
300 unsigned rgb = ((const uint16_t *)src)[i];
301 unsigned br = rgb & 0x7C1F;
302 ((uint16_t *)dst)[i] = (br >> 10) | (rgb & 0x3E0) | (br << 10);
306 void rgb12tobgr12(const uint8_t *src, uint8_t *dst, int src_size)
308 uint16_t *d = (uint16_t *)dst;
309 uint16_t *s = (uint16_t *)src;
310 int i, num_pixels = src_size >> 1;
312 for (i = 0; i < num_pixels; i++) {
314 d[i] = (rgb << 8 | rgb & 0xF0 | rgb >> 8) & 0xFFF;
319 #define DEFINE_SHUFFLE_BYTES(a, b, c, d) \
320 void shuffle_bytes_ ## a ## b ## c ## d(const uint8_t *src, \
321 uint8_t *dst, int src_size) \
325 for (i = 0; i < src_size; i += 4) { \
326 dst[i + 0] = src[i + a]; \
327 dst[i + 1] = src[i + b]; \
328 dst[i + 2] = src[i + c]; \
329 dst[i + 3] = src[i + d]; \
333 DEFINE_SHUFFLE_BYTES(0, 3, 2, 1)
334 DEFINE_SHUFFLE_BYTES(1, 2, 3, 0)
335 DEFINE_SHUFFLE_BYTES(3, 0, 1, 2)
336 DEFINE_SHUFFLE_BYTES(3, 2, 1, 0)
338 #define DEFINE_RGB48TOBGR48(need_bswap, swap) \
339 void rgb48tobgr48_ ## need_bswap(const uint8_t *src, \
340 uint8_t *dst, int src_size) \
342 uint16_t *d = (uint16_t *)dst; \
343 uint16_t *s = (uint16_t *)src; \
344 int i, num_pixels = src_size >> 1; \
346 for (i = 0; i < num_pixels; i += 3) { \
347 d[i ] = swap ? av_bswap16(s[i + 2]) : s[i + 2]; \
348 d[i + 1] = swap ? av_bswap16(s[i + 1]) : s[i + 1]; \
349 d[i + 2] = swap ? av_bswap16(s[i ]) : s[i ]; \
353 DEFINE_RGB48TOBGR48(nobswap, 0)
354 DEFINE_RGB48TOBGR48(bswap, 1)
356 #define DEFINE_RGB64TOBGR48(need_bswap, swap) \
357 void rgb64tobgr48_ ## need_bswap(const uint8_t *src, \
358 uint8_t *dst, int src_size) \
360 uint16_t *d = (uint16_t *)dst; \
361 uint16_t *s = (uint16_t *)src; \
362 int i, num_pixels = src_size >> 3; \
364 for (i = 0; i < num_pixels; i++) { \
365 d[3 * i ] = swap ? av_bswap16(s[4 * i + 2]) : s[4 * i + 2]; \
366 d[3 * i + 1] = swap ? av_bswap16(s[4 * i + 1]) : s[4 * i + 1]; \
367 d[3 * i + 2] = swap ? av_bswap16(s[4 * i ]) : s[4 * i ]; \
371 DEFINE_RGB64TOBGR48(nobswap, 0)
372 DEFINE_RGB64TOBGR48(bswap, 1)
374 #define DEFINE_RGB64TO48(need_bswap, swap) \
375 void rgb64to48_ ## need_bswap(const uint8_t *src, \
376 uint8_t *dst, int src_size) \
378 uint16_t *d = (uint16_t *)dst; \
379 uint16_t *s = (uint16_t *)src; \
380 int i, num_pixels = src_size >> 3; \
382 for (i = 0; i < num_pixels; i++) { \
383 d[3 * i ] = swap ? av_bswap16(s[4 * i ]) : s[4 * i ]; \
384 d[3 * i + 1] = swap ? av_bswap16(s[4 * i + 1]) : s[4 * i + 1]; \
385 d[3 * i + 2] = swap ? av_bswap16(s[4 * i + 2]) : s[4 * i + 2]; \
389 DEFINE_RGB64TO48(nobswap, 0)
390 DEFINE_RGB64TO48(bswap, 1)