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_0321)(const uint8_t *src, uint8_t *dst, int src_size);
55 void (*shuffle_bytes_2103)(const uint8_t *src, uint8_t *dst, int src_size);
57 void (*yv12toyuy2)(const uint8_t *ysrc, const uint8_t *usrc,
58 const uint8_t *vsrc, uint8_t *dst,
59 int width, int height,
60 int lumStride, int chromStride, int dstStride);
61 void (*yv12touyvy)(const uint8_t *ysrc, const uint8_t *usrc,
62 const uint8_t *vsrc, uint8_t *dst,
63 int width, int height,
64 int lumStride, int chromStride, int dstStride);
65 void (*yuv422ptoyuy2)(const uint8_t *ysrc, const uint8_t *usrc,
66 const uint8_t *vsrc, uint8_t *dst,
67 int width, int height,
68 int lumStride, int chromStride, int dstStride);
69 void (*yuv422ptouyvy)(const uint8_t *ysrc, const uint8_t *usrc,
70 const uint8_t *vsrc, uint8_t *dst,
71 int width, int height,
72 int lumStride, int chromStride, int dstStride);
73 void (*yuy2toyv12)(const uint8_t *src, uint8_t *ydst,
74 uint8_t *udst, uint8_t *vdst,
75 int width, int height,
76 int lumStride, int chromStride, int srcStride);
77 void (*ff_rgb24toyv12)(const uint8_t *src, uint8_t *ydst,
78 uint8_t *udst, uint8_t *vdst,
79 int width, int height,
80 int lumStride, int chromStride, int srcStride,
82 void (*planar2x)(const uint8_t *src, uint8_t *dst, int width, int height,
83 int srcStride, int dstStride);
84 void (*interleaveBytes)(const uint8_t *src1, const uint8_t *src2, uint8_t *dst,
85 int width, int height, int src1Stride,
86 int src2Stride, int dstStride);
87 void (*deinterleaveBytes)(const uint8_t *src, uint8_t *dst1, uint8_t *dst2,
88 int width, int height, int srcStride,
89 int dst1Stride, int dst2Stride);
90 void (*vu9_to_vu12)(const uint8_t *src1, const uint8_t *src2,
91 uint8_t *dst1, uint8_t *dst2,
92 int width, int height,
93 int srcStride1, int srcStride2,
94 int dstStride1, int dstStride2);
95 void (*yvu9_to_yuy2)(const uint8_t *src1, const uint8_t *src2,
96 const uint8_t *src3, uint8_t *dst,
97 int width, int height,
98 int srcStride1, int srcStride2,
99 int srcStride3, int dstStride);
100 void (*uyvytoyuv420)(uint8_t *ydst, uint8_t *udst, uint8_t *vdst,
101 const uint8_t *src, int width, int height,
102 int lumStride, int chromStride, int srcStride);
103 void (*uyvytoyuv422)(uint8_t *ydst, uint8_t *udst, uint8_t *vdst,
104 const uint8_t *src, int width, int height,
105 int lumStride, int chromStride, int srcStride);
106 void (*yuyvtoyuv420)(uint8_t *ydst, uint8_t *udst, uint8_t *vdst,
107 const uint8_t *src, int width, int height,
108 int lumStride, int chromStride, int srcStride);
109 void (*yuyvtoyuv422)(uint8_t *ydst, uint8_t *udst, uint8_t *vdst,
110 const uint8_t *src, int width, int height,
111 int lumStride, int chromStride, int srcStride);
113 #define BY ((int)( 0.098 * (1 << RGB2YUV_SHIFT) + 0.5))
114 #define BV ((int)(-0.071 * (1 << RGB2YUV_SHIFT) + 0.5))
115 #define BU ((int)( 0.439 * (1 << RGB2YUV_SHIFT) + 0.5))
116 #define GY ((int)( 0.504 * (1 << RGB2YUV_SHIFT) + 0.5))
117 #define GV ((int)(-0.368 * (1 << RGB2YUV_SHIFT) + 0.5))
118 #define GU ((int)(-0.291 * (1 << RGB2YUV_SHIFT) + 0.5))
119 #define RY ((int)( 0.257 * (1 << RGB2YUV_SHIFT) + 0.5))
120 #define RV ((int)( 0.439 * (1 << RGB2YUV_SHIFT) + 0.5))
121 #define RU ((int)(-0.148 * (1 << RGB2YUV_SHIFT) + 0.5))
124 #include "rgb2rgb_template.c"
127 * RGB15->RGB16 original by Strepto/Astral
128 * ported to gcc & bugfixed : A'rpi
129 * MMXEXT, 3DNOW optimization by Nick Kurshev
130 * 32-bit C version, and and&add trick by Michael Niedermayer
133 av_cold void ff_sws_rgb2rgb_init(void)
140 void rgb32to24(const uint8_t *src, uint8_t *dst, int src_size)
142 int i, num_pixels = src_size >> 2;
144 for (i = 0; i < num_pixels; i++) {
146 /* RGB32 (= A,B,G,R) -> BGR24 (= B,G,R) */
147 dst[3 * i + 0] = src[4 * i + 1];
148 dst[3 * i + 1] = src[4 * i + 2];
149 dst[3 * i + 2] = src[4 * i + 3];
151 dst[3 * i + 0] = src[4 * i + 2];
152 dst[3 * i + 1] = src[4 * i + 1];
153 dst[3 * i + 2] = src[4 * i + 0];
158 void rgb24to32(const uint8_t *src, uint8_t *dst, int src_size)
162 for (i = 0; 3 * i < src_size; i++) {
164 /* RGB24 (= R, G, B) -> BGR32 (= A, R, G, B) */
165 dst[4 * i + 0] = 255;
166 dst[4 * i + 1] = src[3 * i + 0];
167 dst[4 * i + 2] = src[3 * i + 1];
168 dst[4 * i + 3] = src[3 * i + 2];
170 dst[4 * i + 0] = src[3 * i + 2];
171 dst[4 * i + 1] = src[3 * i + 1];
172 dst[4 * i + 2] = src[3 * i + 0];
173 dst[4 * i + 3] = 255;
178 void rgb16tobgr32(const uint8_t *src, uint8_t *dst, int src_size)
181 const uint16_t *s = (const uint16_t *)src;
182 const uint16_t *end = s + src_size / 2;
185 register uint16_t bgr = *s++;
188 *d++ = ((bgr&0x001F)<<3) | ((bgr&0x001F)>> 2);
189 *d++ = ((bgr&0x07E0)>>3) | ((bgr&0x07E0)>> 9);
190 *d++ = ((bgr&0xF800)>>8) | ((bgr&0xF800)>>13);
192 *d++ = ((bgr&0xF800)>>8) | ((bgr&0xF800)>>13);
193 *d++ = ((bgr&0x07E0)>>3) | ((bgr&0x07E0)>> 9);
194 *d++ = ((bgr&0x001F)<<3) | ((bgr&0x001F)>> 2);
200 void rgb12to15(const uint8_t *src, uint8_t *dst, int src_size)
202 uint16_t rgb, r, g, b;
203 uint16_t *d = (uint16_t *)dst;
204 const uint16_t *s = (const uint16_t *)src;
205 const uint16_t *end = s + src_size / 2;
212 r = (r << 3) | ((r & 0x800) >> 1);
213 g = (g << 2) | ((g & 0x080) >> 2);
214 b = (b << 1) | ( b >> 3);
219 void rgb16to24(const uint8_t *src, uint8_t *dst, int src_size)
222 const uint16_t *s = (const uint16_t *)src;
223 const uint16_t *end = s + src_size / 2;
226 register uint16_t bgr = *s++;
227 *d++ = ((bgr&0xF800)>>8) | ((bgr&0xF800)>>13);
228 *d++ = ((bgr&0x07E0)>>3) | ((bgr&0x07E0)>> 9);
229 *d++ = ((bgr&0x001F)<<3) | ((bgr&0x001F)>> 2);
233 void rgb16tobgr16(const uint8_t *src, uint8_t *dst, int src_size)
235 int i, num_pixels = src_size >> 1;
237 for (i = 0; i < num_pixels; i++) {
238 unsigned rgb = ((const uint16_t *)src)[i];
239 ((uint16_t *)dst)[i] = (rgb >> 11) | (rgb & 0x7E0) | (rgb << 11);
243 void rgb16tobgr15(const uint8_t *src, uint8_t *dst, int src_size)
245 int i, num_pixels = src_size >> 1;
247 for (i = 0; i < num_pixels; i++) {
248 unsigned rgb = ((const uint16_t *)src)[i];
249 ((uint16_t *)dst)[i] = (rgb >> 11) | ((rgb & 0x7C0) >> 1) | ((rgb & 0x1F) << 10);
253 void rgb15tobgr32(const uint8_t *src, uint8_t *dst, int src_size)
256 const uint16_t *s = (const uint16_t *)src;
257 const uint16_t *end = s + src_size / 2;
260 register uint16_t bgr = *s++;
263 *d++ = ((bgr&0x001F)<<3) | ((bgr&0x001F)>> 2);
264 *d++ = ((bgr&0x03E0)>>2) | ((bgr&0x03E0)>> 7);
265 *d++ = ((bgr&0x7C00)>>7) | ((bgr&0x7C00)>>12);
267 *d++ = ((bgr&0x7C00)>>7) | ((bgr&0x7C00)>>12);
268 *d++ = ((bgr&0x03E0)>>2) | ((bgr&0x03E0)>> 7);
269 *d++ = ((bgr&0x001F)<<3) | ((bgr&0x001F)>> 2);
275 void rgb15to24(const uint8_t *src, uint8_t *dst, int src_size)
278 const uint16_t *s = (const uint16_t *)src;
279 const uint16_t *end = s + src_size / 2;
282 register uint16_t bgr = *s++;
283 *d++ = ((bgr&0x7C00)>>7) | ((bgr&0x7C00)>>12);
284 *d++ = ((bgr&0x03E0)>>2) | ((bgr&0x03E0)>> 7);
285 *d++ = ((bgr&0x001F)<<3) | ((bgr&0x001F)>> 2);
289 void rgb15tobgr16(const uint8_t *src, uint8_t *dst, int src_size)
291 int i, num_pixels = src_size >> 1;
293 for (i = 0; i < num_pixels; i++) {
294 unsigned rgb = ((const uint16_t *)src)[i];
295 ((uint16_t *)dst)[i] = ((rgb & 0x7C00) >> 10) | ((rgb & 0x3E0) << 1) | (rgb << 11);
299 void rgb15tobgr15(const uint8_t *src, uint8_t *dst, int src_size)
301 int i, num_pixels = src_size >> 1;
303 for (i = 0; i < num_pixels; i++) {
304 unsigned rgb = ((const uint16_t *)src)[i];
305 unsigned br = rgb & 0x7C1F;
306 ((uint16_t *)dst)[i] = (br >> 10) | (rgb & 0x3E0) | (br << 10);
310 void rgb12tobgr12(const uint8_t *src, uint8_t *dst, int src_size)
312 uint16_t *d = (uint16_t *)dst;
313 uint16_t *s = (uint16_t *)src;
314 int i, num_pixels = src_size >> 1;
316 for (i = 0; i < num_pixels; i++) {
318 d[i] = (rgb << 8 | rgb & 0xF0 | rgb >> 8) & 0xFFF;
323 #define DEFINE_SHUFFLE_BYTES(a, b, c, d) \
324 void shuffle_bytes_ ## a ## b ## c ## d(const uint8_t *src, \
325 uint8_t *dst, int src_size) \
329 for (i = 0; i < src_size; i += 4) { \
330 dst[i + 0] = src[i + a]; \
331 dst[i + 1] = src[i + b]; \
332 dst[i + 2] = src[i + c]; \
333 dst[i + 3] = src[i + d]; \
337 DEFINE_SHUFFLE_BYTES(1, 2, 3, 0)
338 DEFINE_SHUFFLE_BYTES(3, 0, 1, 2)
339 DEFINE_SHUFFLE_BYTES(3, 2, 1, 0)
341 #define DEFINE_RGB48TOBGR48(need_bswap, swap) \
342 void rgb48tobgr48_ ## need_bswap(const uint8_t *src, \
343 uint8_t *dst, int src_size) \
345 uint16_t *d = (uint16_t *)dst; \
346 uint16_t *s = (uint16_t *)src; \
347 int i, num_pixels = src_size >> 1; \
349 for (i = 0; i < num_pixels; i += 3) { \
350 d[i ] = swap ? av_bswap16(s[i + 2]) : s[i + 2]; \
351 d[i + 1] = swap ? av_bswap16(s[i + 1]) : s[i + 1]; \
352 d[i + 2] = swap ? av_bswap16(s[i ]) : s[i ]; \
356 DEFINE_RGB48TOBGR48(nobswap, 0)
357 DEFINE_RGB48TOBGR48(bswap, 1)
359 #define DEFINE_RGB64TOBGR48(need_bswap, swap) \
360 void rgb64tobgr48_ ## need_bswap(const uint8_t *src, \
361 uint8_t *dst, int src_size) \
363 uint16_t *d = (uint16_t *)dst; \
364 uint16_t *s = (uint16_t *)src; \
365 int i, num_pixels = src_size >> 3; \
367 for (i = 0; i < num_pixels; i++) { \
368 d[3 * i ] = swap ? av_bswap16(s[4 * i + 2]) : s[4 * i + 2]; \
369 d[3 * i + 1] = swap ? av_bswap16(s[4 * i + 1]) : s[4 * i + 1]; \
370 d[3 * i + 2] = swap ? av_bswap16(s[4 * i ]) : s[4 * i ]; \
374 DEFINE_RGB64TOBGR48(nobswap, 0)
375 DEFINE_RGB64TOBGR48(bswap, 1)
377 #define DEFINE_RGB64TO48(need_bswap, swap) \
378 void rgb64to48_ ## need_bswap(const uint8_t *src, \
379 uint8_t *dst, int src_size) \
381 uint16_t *d = (uint16_t *)dst; \
382 uint16_t *s = (uint16_t *)src; \
383 int i, num_pixels = src_size >> 3; \
385 for (i = 0; i < num_pixels; i++) { \
386 d[3 * i ] = swap ? av_bswap16(s[4 * i ]) : s[4 * i ]; \
387 d[3 * i + 1] = swap ? av_bswap16(s[4 * i + 1]) : s[4 * i + 1]; \
388 d[3 * i + 2] = swap ? av_bswap16(s[4 * i + 2]) : s[4 * i + 2]; \
392 DEFINE_RGB64TO48(nobswap, 0)
393 DEFINE_RGB64TO48(bswap, 1)