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 (*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);
80 void (*planar2x)(const uint8_t *src, uint8_t *dst, int width, int height,
81 int srcStride, int dstStride);
82 void (*interleaveBytes)(const uint8_t *src1, const uint8_t *src2, uint8_t *dst,
83 int width, int height, int src1Stride,
84 int src2Stride, int dstStride);
85 void (*vu9_to_vu12)(const uint8_t *src1, const uint8_t *src2,
86 uint8_t *dst1, uint8_t *dst2,
87 int width, int height,
88 int srcStride1, int srcStride2,
89 int dstStride1, int dstStride2);
90 void (*yvu9_to_yuy2)(const uint8_t *src1, const uint8_t *src2,
91 const uint8_t *src3, uint8_t *dst,
92 int width, int height,
93 int srcStride1, int srcStride2,
94 int srcStride3, int dstStride);
95 void (*uyvytoyuv420)(uint8_t *ydst, uint8_t *udst, uint8_t *vdst,
96 const uint8_t *src, int width, int height,
97 int lumStride, int chromStride, int srcStride);
98 void (*uyvytoyuv422)(uint8_t *ydst, uint8_t *udst, uint8_t *vdst,
99 const uint8_t *src, int width, int height,
100 int lumStride, int chromStride, int srcStride);
101 void (*yuyvtoyuv420)(uint8_t *ydst, uint8_t *udst, uint8_t *vdst,
102 const uint8_t *src, int width, int height,
103 int lumStride, int chromStride, int srcStride);
104 void (*yuyvtoyuv422)(uint8_t *ydst, uint8_t *udst, uint8_t *vdst,
105 const uint8_t *src, int width, int height,
106 int lumStride, int chromStride, int srcStride);
108 #define BY ((int)( 0.098 * (1 << RGB2YUV_SHIFT) + 0.5))
109 #define BV ((int)(-0.071 * (1 << RGB2YUV_SHIFT) + 0.5))
110 #define BU ((int)( 0.439 * (1 << RGB2YUV_SHIFT) + 0.5))
111 #define GY ((int)( 0.504 * (1 << RGB2YUV_SHIFT) + 0.5))
112 #define GV ((int)(-0.368 * (1 << RGB2YUV_SHIFT) + 0.5))
113 #define GU ((int)(-0.291 * (1 << RGB2YUV_SHIFT) + 0.5))
114 #define RY ((int)( 0.257 * (1 << RGB2YUV_SHIFT) + 0.5))
115 #define RV ((int)( 0.439 * (1 << RGB2YUV_SHIFT) + 0.5))
116 #define RU ((int)(-0.148 * (1 << RGB2YUV_SHIFT) + 0.5))
119 #include "rgb2rgb_template.c"
122 * RGB15->RGB16 original by Strepto/Astral
123 * ported to gcc & bugfixed : A'rpi
124 * MMXEXT, 3DNOW optimization by Nick Kurshev
125 * 32-bit C version, and and&add trick by Michael Niedermayer
128 av_cold void sws_rgb2rgb_init(void)
135 void rgb32to24(const uint8_t *src, uint8_t *dst, int src_size)
137 int i, num_pixels = src_size >> 2;
139 for (i = 0; i < num_pixels; i++) {
141 /* RGB32 (= A,B,G,R) -> BGR24 (= B,G,R) */
142 dst[3 * i + 0] = src[4 * i + 1];
143 dst[3 * i + 1] = src[4 * i + 2];
144 dst[3 * i + 2] = src[4 * i + 3];
146 dst[3 * i + 0] = src[4 * i + 2];
147 dst[3 * i + 1] = src[4 * i + 1];
148 dst[3 * i + 2] = src[4 * i + 0];
153 void rgb24to32(const uint8_t *src, uint8_t *dst, int src_size)
157 for (i = 0; 3 * i < src_size; i++) {
159 /* RGB24 (= R, G, B) -> BGR32 (= A, R, G, B) */
160 dst[4 * i + 0] = 255;
161 dst[4 * i + 1] = src[3 * i + 0];
162 dst[4 * i + 2] = src[3 * i + 1];
163 dst[4 * i + 3] = src[3 * i + 2];
165 dst[4 * i + 0] = src[3 * i + 2];
166 dst[4 * i + 1] = src[3 * i + 1];
167 dst[4 * i + 2] = src[3 * i + 0];
168 dst[4 * i + 3] = 255;
173 void rgb16tobgr32(const uint8_t *src, uint8_t *dst, int src_size)
176 const uint16_t *s = (const uint16_t *)src;
177 const uint16_t *end = s + src_size / 2;
180 register uint16_t bgr = *s++;
183 *d++ = ((bgr&0x001F)<<3) | ((bgr&0x001F)>> 2);
184 *d++ = ((bgr&0x07E0)>>3) | ((bgr&0x07E0)>> 9);
185 *d++ = ((bgr&0xF800)>>8) | ((bgr&0xF800)>>13);
187 *d++ = ((bgr&0xF800)>>8) | ((bgr&0xF800)>>13);
188 *d++ = ((bgr&0x07E0)>>3) | ((bgr&0x07E0)>> 9);
189 *d++ = ((bgr&0x001F)<<3) | ((bgr&0x001F)>> 2);
195 void rgb12to15(const uint8_t *src, uint8_t *dst, int src_size)
197 uint16_t rgb, r, g, b;
198 uint16_t *d = (uint16_t *)dst;
199 const uint16_t *s = (const uint16_t *)src;
200 const uint16_t *end = s + src_size / 2;
207 r = (r << 3) | ((r & 0x800) >> 1);
208 g = (g << 2) | ((g & 0x080) >> 2);
209 b = (b << 1) | ( b >> 3);
214 void rgb16to24(const uint8_t *src, uint8_t *dst, int src_size)
217 const uint16_t *s = (const uint16_t *)src;
218 const uint16_t *end = s + src_size / 2;
221 register uint16_t bgr = *s++;
222 *d++ = ((bgr&0xF800)>>8) | ((bgr&0xF800)>>13);
223 *d++ = ((bgr&0x07E0)>>3) | ((bgr&0x07E0)>> 9);
224 *d++ = ((bgr&0x001F)<<3) | ((bgr&0x001F)>> 2);
228 void rgb16tobgr16(const uint8_t *src, uint8_t *dst, int src_size)
230 int i, num_pixels = src_size >> 1;
232 for (i = 0; i < num_pixels; i++) {
233 unsigned rgb = ((const uint16_t *)src)[i];
234 ((uint16_t *)dst)[i] = (rgb >> 11) | (rgb & 0x7E0) | (rgb << 11);
238 void rgb16tobgr15(const uint8_t *src, uint8_t *dst, int src_size)
240 int i, num_pixels = src_size >> 1;
242 for (i = 0; i < num_pixels; i++) {
243 unsigned rgb = ((const uint16_t *)src)[i];
244 ((uint16_t *)dst)[i] = (rgb >> 11) | ((rgb & 0x7C0) >> 1) | ((rgb & 0x1F) << 10);
248 void rgb15tobgr32(const uint8_t *src, uint8_t *dst, int src_size)
251 const uint16_t *s = (const uint16_t *)src;
252 const uint16_t *end = s + src_size / 2;
255 register uint16_t bgr = *s++;
258 *d++ = ((bgr&0x001F)<<3) | ((bgr&0x001F)>> 2);
259 *d++ = ((bgr&0x03E0)>>2) | ((bgr&0x03E0)>> 7);
260 *d++ = ((bgr&0x7C00)>>7) | ((bgr&0x7C00)>>12);
262 *d++ = ((bgr&0x7C00)>>7) | ((bgr&0x7C00)>>12);
263 *d++ = ((bgr&0x03E0)>>2) | ((bgr&0x03E0)>> 7);
264 *d++ = ((bgr&0x001F)<<3) | ((bgr&0x001F)>> 2);
270 void rgb15to24(const uint8_t *src, uint8_t *dst, int src_size)
273 const uint16_t *s = (const uint16_t *)src;
274 const uint16_t *end = s + src_size / 2;
277 register uint16_t bgr = *s++;
278 *d++ = ((bgr&0x7C00)>>7) | ((bgr&0x7C00)>>12);
279 *d++ = ((bgr&0x03E0)>>2) | ((bgr&0x03E0)>> 7);
280 *d++ = ((bgr&0x001F)<<3) | ((bgr&0x001F)>> 2);
284 void rgb15tobgr16(const uint8_t *src, uint8_t *dst, int src_size)
286 int i, num_pixels = src_size >> 1;
288 for (i = 0; i < num_pixels; i++) {
289 unsigned rgb = ((const uint16_t *)src)[i];
290 ((uint16_t *)dst)[i] = ((rgb & 0x7C00) >> 10) | ((rgb & 0x3E0) << 1) | (rgb << 11);
294 void rgb15tobgr15(const uint8_t *src, uint8_t *dst, int src_size)
296 int i, num_pixels = src_size >> 1;
298 for (i = 0; i < num_pixels; i++) {
299 unsigned rgb = ((const uint16_t *)src)[i];
300 unsigned br = rgb & 0x7C1F;
301 ((uint16_t *)dst)[i] = (br >> 10) | (rgb & 0x3E0) | (br << 10);
305 void rgb12tobgr12(const uint8_t *src, uint8_t *dst, int src_size)
307 uint16_t *d = (uint16_t *)dst;
308 uint16_t *s = (uint16_t *)src;
309 int i, num_pixels = src_size >> 1;
311 for (i = 0; i < num_pixels; i++) {
313 d[i] = (rgb << 8 | rgb & 0xF0 | rgb >> 8) & 0xFFF;
318 #define DEFINE_SHUFFLE_BYTES(a, b, c, d) \
319 void shuffle_bytes_ ## a ## b ## c ## d(const uint8_t *src, \
320 uint8_t *dst, int src_size) \
324 for (i = 0; i < src_size; i += 4) { \
325 dst[i + 0] = src[i + a]; \
326 dst[i + 1] = src[i + b]; \
327 dst[i + 2] = src[i + c]; \
328 dst[i + 3] = src[i + d]; \
332 DEFINE_SHUFFLE_BYTES(0, 3, 2, 1)
333 DEFINE_SHUFFLE_BYTES(1, 2, 3, 0)
334 DEFINE_SHUFFLE_BYTES(3, 0, 1, 2)
335 DEFINE_SHUFFLE_BYTES(3, 2, 1, 0)
337 #define DEFINE_RGB48TOBGR48(need_bswap, swap) \
338 void rgb48tobgr48_ ## need_bswap(const uint8_t *src, \
339 uint8_t *dst, int src_size) \
341 uint16_t *d = (uint16_t *)dst; \
342 uint16_t *s = (uint16_t *)src; \
343 int i, num_pixels = src_size >> 1; \
345 for (i = 0; i < num_pixels; i += 3) { \
346 d[i ] = swap ? av_bswap16(s[i + 2]) : s[i + 2]; \
347 d[i + 1] = swap ? av_bswap16(s[i + 1]) : s[i + 1]; \
348 d[i + 2] = swap ? av_bswap16(s[i ]) : s[i ]; \
352 DEFINE_RGB48TOBGR48(nobswap, 0)
353 DEFINE_RGB48TOBGR48(bswap, 1)
355 #define DEFINE_RGB64TOBGR48(need_bswap, swap) \
356 void rgb64tobgr48_ ## need_bswap(const uint8_t *src, \
357 uint8_t *dst, int src_size) \
359 uint16_t *d = (uint16_t *)dst; \
360 uint16_t *s = (uint16_t *)src; \
361 int i, num_pixels = src_size >> 3; \
363 for (i = 0; i < num_pixels; i++) { \
364 d[3 * i ] = swap ? av_bswap16(s[4 * i + 2]) : s[4 * i + 2]; \
365 d[3 * i + 1] = swap ? av_bswap16(s[4 * i + 1]) : s[4 * i + 1]; \
366 d[3 * i + 2] = swap ? av_bswap16(s[4 * i ]) : s[4 * i ]; \
370 DEFINE_RGB64TOBGR48(nobswap, 0)
371 DEFINE_RGB64TOBGR48(bswap, 1)
373 #define DEFINE_RGB64TO48(need_bswap, swap) \
374 void rgb64to48_ ## need_bswap(const uint8_t *src, \
375 uint8_t *dst, int src_size) \
377 uint16_t *d = (uint16_t *)dst; \
378 uint16_t *s = (uint16_t *)src; \
379 int i, num_pixels = src_size >> 3; \
381 for (i = 0; i < num_pixels; i++) { \
382 d[3 * i ] = swap ? av_bswap16(s[4 * i ]) : s[4 * i ]; \
383 d[3 * i + 1] = swap ? av_bswap16(s[4 * i + 1]) : s[4 * i + 1]; \
384 d[3 * i + 2] = swap ? av_bswap16(s[4 * i + 2]) : s[4 * i + 2]; \
388 DEFINE_RGB64TO48(nobswap, 0)
389 DEFINE_RGB64TO48(bswap, 1)