rgb2rgb.c 15.4 KB
Newer Older
1
/*
2 3 4 5 6 7
 * software RGB to RGB converter
 * pluralize by software PAL8 to RGB converter
 *              software YUV to YUV converter
 *              software YUV to RGB converter
 * Written by Nick Kurshev.
 * palette & YUV & runtime CPU stuff by Michael (michaelni@gmx.at)
8
 *
9 10
 * This file is part of FFmpeg.
 *
11 12 13 14
 * FFmpeg is free software; you can redistribute it and/or
 * modify it under the terms of the GNU Lesser General Public
 * License as published by the Free Software Foundation; either
 * version 2.1 of the License, or (at your option) any later version.
15
 *
16
 * FFmpeg is distributed in the hope that it will be useful,
17
 * but WITHOUT ANY WARRANTY; without even the implied warranty of
18 19
 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
 * Lesser General Public License for more details.
20
 *
21 22
 * You should have received a copy of the GNU Lesser General Public
 * License along with FFmpeg; if not, write to the Free Software
23
 * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
24
 */
25

Nick Kurshev's avatar
Nick Kurshev committed
26
#include <inttypes.h>
27

28
#include "libavutil/attributes.h"
29
#include "libavutil/bswap.h"
30
#include "config.h"
Nick Kurshev's avatar
Nick Kurshev committed
31
#include "rgb2rgb.h"
32
#include "swscale.h"
33
#include "swscale_internal.h"
34

35 36 37
void (*rgb32tobgr24)(const uint8_t *src, uint8_t *dst, int src_size);
void (*rgb32tobgr16)(const uint8_t *src, uint8_t *dst, int src_size);
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 41
void (*rgb24tobgr16)(const uint8_t *src, uint8_t *dst, int src_size);
void (*rgb24tobgr15)(const uint8_t *src, uint8_t *dst, int src_size);
42 43 44
void (*rgb16tobgr24)(const uint8_t *src, uint8_t *dst, int src_size);
void (*rgb15tobgr24)(const uint8_t *src, uint8_t *dst, int src_size);

45 46 47 48
void (*rgb32to16)(const uint8_t *src, uint8_t *dst, int src_size);
void (*rgb32to15)(const uint8_t *src, uint8_t *dst, int src_size);
void (*rgb24to16)(const uint8_t *src, uint8_t *dst, int src_size);
void (*rgb24to15)(const uint8_t *src, uint8_t *dst, int src_size);
49 50 51 52 53
void (*rgb16to32)(const uint8_t *src, uint8_t *dst, int src_size);
void (*rgb16to15)(const uint8_t *src, uint8_t *dst, int src_size);
void (*rgb15to16)(const uint8_t *src, uint8_t *dst, int src_size);
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);
55

56 57
void (*yv12toyuy2)(const uint8_t *ysrc, const uint8_t *usrc,
                   const uint8_t *vsrc, uint8_t *dst,
58 59
                   int width, int height,
                   int lumStride, int chromStride, int dstStride);
60 61
void (*yv12touyvy)(const uint8_t *ysrc, const uint8_t *usrc,
                   const uint8_t *vsrc, uint8_t *dst,
62 63
                   int width, int height,
                   int lumStride, int chromStride, int dstStride);
64 65
void (*yuv422ptoyuy2)(const uint8_t *ysrc, const uint8_t *usrc,
                      const uint8_t *vsrc, uint8_t *dst,
66 67
                      int width, int height,
                      int lumStride, int chromStride, int dstStride);
68 69
void (*yuv422ptouyvy)(const uint8_t *ysrc, const uint8_t *usrc,
                      const uint8_t *vsrc, uint8_t *dst,
70 71
                      int width, int height,
                      int lumStride, int chromStride, int dstStride);
72 73
void (*yuy2toyv12)(const uint8_t *src, uint8_t *ydst,
                   uint8_t *udst, uint8_t *vdst,
74 75
                   int width, int height,
                   int lumStride, int chromStride, int srcStride);
76 77 78 79 80
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);
81 82
void (*planar2x)(const uint8_t *src, uint8_t *dst, int width, int height,
                 int srcStride, int dstStride);
83
void (*interleaveBytes)(const uint8_t *src1, const uint8_t *src2, uint8_t *dst,
84 85
                        int width, int height, int src1Stride,
                        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 89 90
                    int width, int height,
                    int srcStride1, int srcStride2,
                    int dstStride1, int dstStride2);
91 92
void (*yvu9_to_yuy2)(const uint8_t *src1, const uint8_t *src2,
                     const uint8_t *src3, uint8_t *dst,
93 94 95
                     int width, int height,
                     int srcStride1, int srcStride2,
                     int srcStride3, int dstStride);
96 97
void (*uyvytoyuv420)(uint8_t *ydst, uint8_t *udst, uint8_t *vdst,
                     const uint8_t *src, int width, int height,
98
                     int lumStride, int chromStride, int srcStride);
99 100
void (*uyvytoyuv422)(uint8_t *ydst, uint8_t *udst, uint8_t *vdst,
                     const uint8_t *src, int width, int height,
101
                     int lumStride, int chromStride, int srcStride);
102 103
void (*yuyvtoyuv420)(uint8_t *ydst, uint8_t *udst, uint8_t *vdst,
                     const uint8_t *src, int width, int height,
104
                     int lumStride, int chromStride, int srcStride);
105 106
void (*yuyvtoyuv422)(uint8_t *ydst, uint8_t *udst, uint8_t *vdst,
                     const uint8_t *src, int width, int height,
107
                     int lumStride, int chromStride, int srcStride);
108

109 110 111 112 113 114 115 116 117
#define BY ((int)( 0.098 * (1 << RGB2YUV_SHIFT) + 0.5))
#define BV ((int)(-0.071 * (1 << RGB2YUV_SHIFT) + 0.5))
#define BU ((int)( 0.439 * (1 << RGB2YUV_SHIFT) + 0.5))
#define GY ((int)( 0.504 * (1 << RGB2YUV_SHIFT) + 0.5))
#define GV ((int)(-0.368 * (1 << RGB2YUV_SHIFT) + 0.5))
#define GU ((int)(-0.291 * (1 << RGB2YUV_SHIFT) + 0.5))
#define RY ((int)( 0.257 * (1 << RGB2YUV_SHIFT) + 0.5))
#define RV ((int)( 0.439 * (1 << RGB2YUV_SHIFT) + 0.5))
#define RU ((int)(-0.148 * (1 << RGB2YUV_SHIFT) + 0.5))
118

119
//plain C versions
120 121
#include "rgb2rgb_template.c"

122
/*
123 124
 * RGB15->RGB16 original by Strepto/Astral
 * ported to gcc & bugfixed : A'rpi
125
 * MMXEXT, 3DNOW optimization by Nick Kurshev
126 127
 * 32-bit C version, and and&add trick by Michael Niedermayer
 */
128

129
av_cold void sws_rgb2rgb_init(void)
Ramiro Polla's avatar
Ramiro Polla committed
130
{
131
    rgb2rgb_init_c();
132 133
    if (HAVE_MMX)
        rgb2rgb_init_x86();
134
}
135

136
void rgb32to24(const uint8_t *src, uint8_t *dst, int src_size)
137
{
138 139
    int i, num_pixels = src_size >> 2;

140
    for (i = 0; i < num_pixels; i++) {
Ramiro Polla's avatar
Ramiro Polla committed
141 142
#if HAVE_BIGENDIAN
        /* RGB32 (= A,B,G,R) -> BGR24 (= B,G,R) */
143 144 145
        dst[3 * i + 0] = src[4 * i + 1];
        dst[3 * i + 1] = src[4 * i + 2];
        dst[3 * i + 2] = src[4 * i + 3];
Ramiro Polla's avatar
Ramiro Polla committed
146
#else
147 148 149
        dst[3 * i + 0] = src[4 * i + 2];
        dst[3 * i + 1] = src[4 * i + 1];
        dst[3 * i + 2] = src[4 * i + 0];
Ramiro Polla's avatar
Ramiro Polla committed
150
#endif
151
    }
152 153
}

154
void rgb24to32(const uint8_t *src, uint8_t *dst, int src_size)
155
{
156
    int i;
157 158

    for (i = 0; 3 * i < src_size; i++) {
Ramiro Polla's avatar
Ramiro Polla committed
159
#if HAVE_BIGENDIAN
160 161 162 163 164
        /* RGB24 (= R, G, B) -> BGR32 (= A, R, G, B) */
        dst[4 * i + 0] = 255;
        dst[4 * i + 1] = src[3 * i + 0];
        dst[4 * i + 2] = src[3 * i + 1];
        dst[4 * i + 3] = src[3 * i + 2];
Ramiro Polla's avatar
Ramiro Polla committed
165
#else
166 167 168 169
        dst[4 * i + 0] = src[3 * i + 2];
        dst[4 * i + 1] = src[3 * i + 1];
        dst[4 * i + 2] = src[3 * i + 0];
        dst[4 * i + 3] = 255;
Ramiro Polla's avatar
Ramiro Polla committed
170
#endif
171
    }
172 173
}

174
void rgb16tobgr32(const uint8_t *src, uint8_t *dst, int src_size)
175
{
176 177
    uint8_t *d          = dst;
    const uint16_t *s   = (const uint16_t *)src;
178 179
    const uint16_t *end = s + src_size / 2;

Ramiro Polla's avatar
Ramiro Polla committed
180
    while (s < end) {
181
        register uint16_t bgr = *s++;
Ramiro Polla's avatar
Ramiro Polla committed
182 183
#if HAVE_BIGENDIAN
        *d++ = 255;
184 185
        *d++ = ((bgr&0x001F)<<3) | ((bgr&0x001F)>> 2);
        *d++ = ((bgr&0x07E0)>>3) | ((bgr&0x07E0)>> 9);
186
        *d++ = ((bgr&0xF800)>>8) | ((bgr&0xF800)>>13);
Ramiro Polla's avatar
Ramiro Polla committed
187
#else
188
        *d++ = ((bgr&0xF800)>>8) | ((bgr&0xF800)>>13);
189 190
        *d++ = ((bgr&0x07E0)>>3) | ((bgr&0x07E0)>> 9);
        *d++ = ((bgr&0x001F)<<3) | ((bgr&0x001F)>> 2);
Ramiro Polla's avatar
Ramiro Polla committed
191 192
        *d++ = 255;
#endif
193
    }
194 195
}

Paul B Mahol's avatar
Paul B Mahol committed
196 197 198
void rgb12to15(const uint8_t *src, uint8_t *dst, int src_size)
{
    uint16_t rgb, r, g, b;
199 200
    uint16_t *d         = (uint16_t *)dst;
    const uint16_t *s   = (const uint16_t *)src;
201 202
    const uint16_t *end = s + src_size / 2;

Paul B Mahol's avatar
Paul B Mahol committed
203
    while (s < end) {
204 205 206 207 208 209 210
        rgb  = *s++;
        r    = rgb & 0xF00;
        g    = rgb & 0x0F0;
        b    = rgb & 0x00F;
        r    = (r << 3) | ((r & 0x800) >> 1);
        g    = (g << 2) | ((g & 0x080) >> 2);
        b    = (b << 1) | ( b          >> 3);
Paul B Mahol's avatar
Paul B Mahol committed
211 212 213 214
        *d++ = r | g | b;
    }
}

215
void rgb16to24(const uint8_t *src, uint8_t *dst, int src_size)
216
{
217 218
    uint8_t *d          = dst;
    const uint16_t *s   = (const uint16_t *)src;
219 220
    const uint16_t *end = s + src_size / 2;

Ramiro Polla's avatar
Ramiro Polla committed
221
    while (s < end) {
222
        register uint16_t bgr = *s++;
223
        *d++ = ((bgr&0xF800)>>8) | ((bgr&0xF800)>>13);
224 225
        *d++ = ((bgr&0x07E0)>>3) | ((bgr&0x07E0)>> 9);
        *d++ = ((bgr&0x001F)<<3) | ((bgr&0x001F)>> 2);
226
    }
227 228
}

229
void rgb16tobgr16(const uint8_t *src, uint8_t *dst, int src_size)
230
{
231
    int i, num_pixels = src_size >> 1;
232

233 234 235
    for (i = 0; i < num_pixels; i++) {
        unsigned rgb         = ((const uint16_t *)src)[i];
        ((uint16_t *)dst)[i] = (rgb >> 11) | (rgb & 0x7E0) | (rgb << 11);
236
    }
237 238
}

239
void rgb16tobgr15(const uint8_t *src, uint8_t *dst, int src_size)
240
{
241
    int i, num_pixels = src_size >> 1;
242

243 244 245
    for (i = 0; i < num_pixels; i++) {
        unsigned rgb         = ((const uint16_t *)src)[i];
        ((uint16_t *)dst)[i] = (rgb >> 11) | ((rgb & 0x7C0) >> 1) | ((rgb & 0x1F) << 10);
246
    }
247 248
}

249
void rgb15tobgr32(const uint8_t *src, uint8_t *dst, int src_size)
250
{
251 252
    uint8_t *d          = dst;
    const uint16_t *s   = (const uint16_t *)src;
253 254
    const uint16_t *end = s + src_size / 2;

Ramiro Polla's avatar
Ramiro Polla committed
255
    while (s < end) {
256
        register uint16_t bgr = *s++;
Ramiro Polla's avatar
Ramiro Polla committed
257 258
#if HAVE_BIGENDIAN
        *d++ = 255;
259 260
        *d++ = ((bgr&0x001F)<<3) | ((bgr&0x001F)>> 2);
        *d++ = ((bgr&0x03E0)>>2) | ((bgr&0x03E0)>> 7);
261
        *d++ = ((bgr&0x7C00)>>7) | ((bgr&0x7C00)>>12);
Ramiro Polla's avatar
Ramiro Polla committed
262
#else
263
        *d++ = ((bgr&0x7C00)>>7) | ((bgr&0x7C00)>>12);
264 265
        *d++ = ((bgr&0x03E0)>>2) | ((bgr&0x03E0)>> 7);
        *d++ = ((bgr&0x001F)<<3) | ((bgr&0x001F)>> 2);
Ramiro Polla's avatar
Ramiro Polla committed
266 267
        *d++ = 255;
#endif
268
    }
269 270
}

271
void rgb15to24(const uint8_t *src, uint8_t *dst, int src_size)
272
{
273 274
    uint8_t *d          = dst;
    const uint16_t *s   = (const uint16_t *)src;
275 276
    const uint16_t *end = s + src_size / 2;

Ramiro Polla's avatar
Ramiro Polla committed
277
    while (s < end) {
278
        register uint16_t bgr = *s++;
279
        *d++ = ((bgr&0x7C00)>>7) | ((bgr&0x7C00)>>12);
280 281
        *d++ = ((bgr&0x03E0)>>2) | ((bgr&0x03E0)>> 7);
        *d++ = ((bgr&0x001F)<<3) | ((bgr&0x001F)>> 2);
282
    }
283 284
}

285
void rgb15tobgr16(const uint8_t *src, uint8_t *dst, int src_size)
286
{
287
    int i, num_pixels = src_size >> 1;
288

289 290 291
    for (i = 0; i < num_pixels; i++) {
        unsigned rgb         = ((const uint16_t *)src)[i];
        ((uint16_t *)dst)[i] = ((rgb & 0x7C00) >> 10) | ((rgb & 0x3E0) << 1) | (rgb << 11);
292
    }
293 294
}

295
void rgb15tobgr15(const uint8_t *src, uint8_t *dst, int src_size)
296
{
297
    int i, num_pixels = src_size >> 1;
298

299 300 301 302
    for (i = 0; i < num_pixels; i++) {
        unsigned rgb         = ((const uint16_t *)src)[i];
        unsigned br          = rgb & 0x7C1F;
        ((uint16_t *)dst)[i] = (br >> 10) | (rgb & 0x3E0) | (br << 10);
303
    }
304 305
}

Paul B Mahol's avatar
Paul B Mahol committed
306 307
void rgb12tobgr12(const uint8_t *src, uint8_t *dst, int src_size)
{
308 309
    uint16_t *d = (uint16_t *)dst;
    uint16_t *s = (uint16_t *)src;
310
    int i, num_pixels = src_size >> 1;
Paul B Mahol's avatar
Paul B Mahol committed
311 312

    for (i = 0; i < num_pixels; i++) {
Paul B Mahol's avatar
Paul B Mahol committed
313
        unsigned rgb = s[i];
314
        d[i]         = (rgb << 8 | rgb & 0xF0 | rgb >> 8) & 0xFFF;
Paul B Mahol's avatar
Paul B Mahol committed
315 316 317
    }
}

318 319

#define DEFINE_SHUFFLE_BYTES(a, b, c, d)                                \
320 321
void shuffle_bytes_ ## a ## b ## c ## d(const uint8_t *src,             \
                                        uint8_t *dst, int src_size)     \
322
{                                                                       \
323
    int i;                                                              \
324
                                                                        \
325
    for (i = 0; i < src_size; i += 4) {                                 \
326 327 328 329 330 331 332
        dst[i + 0] = src[i + a];                                        \
        dst[i + 1] = src[i + b];                                        \
        dst[i + 2] = src[i + c];                                        \
        dst[i + 3] = src[i + d];                                        \
    }                                                                   \
}

333 334 335 336
DEFINE_SHUFFLE_BYTES(0, 3, 2, 1)
DEFINE_SHUFFLE_BYTES(1, 2, 3, 0)
DEFINE_SHUFFLE_BYTES(3, 0, 1, 2)
DEFINE_SHUFFLE_BYTES(3, 2, 1, 0)
337

338 339 340
#define DEFINE_RGB48TOBGR48(need_bswap, swap)                           \
void rgb48tobgr48_ ## need_bswap(const uint8_t *src,                    \
                                 uint8_t *dst, int src_size)            \
341 342 343 344 345 346
{                                                                       \
    uint16_t *d = (uint16_t *)dst;                                      \
    uint16_t *s = (uint16_t *)src;                                      \
    int i, num_pixels = src_size >> 1;                                  \
                                                                        \
    for (i = 0; i < num_pixels; i += 3) {                               \
347 348 349
        d[i    ] = swap ? av_bswap16(s[i + 2]) : s[i + 2];              \
        d[i + 1] = swap ? av_bswap16(s[i + 1]) : s[i + 1];              \
        d[i + 2] = swap ? av_bswap16(s[i    ]) : s[i    ];              \
350 351 352
    }                                                                   \
}

353 354
DEFINE_RGB48TOBGR48(nobswap, 0)
DEFINE_RGB48TOBGR48(bswap, 1)
355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372

#define DEFINE_RGB64TOBGR48(need_bswap, swap)                           \
void rgb64tobgr48_ ## need_bswap(const uint8_t *src,                    \
                                 uint8_t *dst, int src_size)            \
{                                                                       \
    uint16_t *d = (uint16_t *)dst;                                      \
    uint16_t *s = (uint16_t *)src;                                      \
    int i, num_pixels = src_size >> 3;                                  \
                                                                        \
    for (i = 0; i < num_pixels; i++) {                                  \
        d[3 * i    ] = swap ? av_bswap16(s[4 * i + 2]) : s[4 * i + 2];  \
        d[3 * i + 1] = swap ? av_bswap16(s[4 * i + 1]) : s[4 * i + 1];  \
        d[3 * i + 2] = swap ? av_bswap16(s[4 * i    ]) : s[4 * i    ];  \
    }                                                                   \
}

DEFINE_RGB64TOBGR48(nobswap, 0)
DEFINE_RGB64TOBGR48(bswap, 1)
373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390

#define DEFINE_RGB64TO48(need_bswap, swap)                              \
void rgb64to48_ ## need_bswap(const uint8_t *src,                       \
                              uint8_t *dst, int src_size)               \
{                                                                       \
    uint16_t *d = (uint16_t *)dst;                                      \
    uint16_t *s = (uint16_t *)src;                                      \
    int i, num_pixels = src_size >> 3;                                  \
                                                                        \
    for (i = 0; i < num_pixels; i++) {                                  \
        d[3 * i    ] = swap ? av_bswap16(s[4 * i    ]) : s[4 * i    ];  \
        d[3 * i + 1] = swap ? av_bswap16(s[4 * i + 1]) : s[4 * i + 1];  \
        d[3 * i + 2] = swap ? av_bswap16(s[4 * i + 2]) : s[4 * i + 2];  \
    }                                                                   \
}

DEFINE_RGB64TO48(nobswap, 0)
DEFINE_RGB64TO48(bswap, 1)