| 1 | /* |
| 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) |
| 8 | * |
| 9 | * This file is part of FFmpeg. |
| 10 | * |
| 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. |
| 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 | * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU |
| 19 | * Lesser General Public License for more details. |
| 20 | * |
| 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 |
| 24 | */ |
| 25 | |
| 26 | #include <inttypes.h> |
| 27 | |
| 28 | #include "libavutil/attributes.h" |
| 29 | #include "libavutil/bswap.h" |
| 30 | #include "config.h" |
| 31 | #include "rgb2rgb.h" |
| 32 | #include "swscale.h" |
| 33 | #include "swscale_internal.h" |
| 34 | |
| 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); |
| 44 | |
| 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); |
| 53 | |
| 54 | void (*shuffle_bytes_2103)(const uint8_t *src, uint8_t *dst, int src_size); |
| 55 | |
| 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, |
| 80 | int32_t *rgb2yuv); |
| 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 (*deinterleaveBytes)(const uint8_t *src, uint8_t *dst1, uint8_t *dst2, |
| 87 | int width, int height, int srcStride, |
| 88 | int dst1Stride, int dst2Stride); |
| 89 | void (*vu9_to_vu12)(const uint8_t *src1, const uint8_t *src2, |
| 90 | uint8_t *dst1, uint8_t *dst2, |
| 91 | int width, int height, |
| 92 | int srcStride1, int srcStride2, |
| 93 | int dstStride1, int dstStride2); |
| 94 | void (*yvu9_to_yuy2)(const uint8_t *src1, const uint8_t *src2, |
| 95 | const uint8_t *src3, uint8_t *dst, |
| 96 | int width, int height, |
| 97 | int srcStride1, int srcStride2, |
| 98 | int srcStride3, int dstStride); |
| 99 | void (*uyvytoyuv420)(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 (*uyvytoyuv422)(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 (*yuyvtoyuv420)(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); |
| 108 | void (*yuyvtoyuv422)(uint8_t *ydst, uint8_t *udst, uint8_t *vdst, |
| 109 | const uint8_t *src, int width, int height, |
| 110 | int lumStride, int chromStride, int srcStride); |
| 111 | |
| 112 | #define BY ((int)( 0.098 * (1 << RGB2YUV_SHIFT) + 0.5)) |
| 113 | #define BV ((int)(-0.071 * (1 << RGB2YUV_SHIFT) + 0.5)) |
| 114 | #define BU ((int)( 0.439 * (1 << RGB2YUV_SHIFT) + 0.5)) |
| 115 | #define GY ((int)( 0.504 * (1 << RGB2YUV_SHIFT) + 0.5)) |
| 116 | #define GV ((int)(-0.368 * (1 << RGB2YUV_SHIFT) + 0.5)) |
| 117 | #define GU ((int)(-0.291 * (1 << RGB2YUV_SHIFT) + 0.5)) |
| 118 | #define RY ((int)( 0.257 * (1 << RGB2YUV_SHIFT) + 0.5)) |
| 119 | #define RV ((int)( 0.439 * (1 << RGB2YUV_SHIFT) + 0.5)) |
| 120 | #define RU ((int)(-0.148 * (1 << RGB2YUV_SHIFT) + 0.5)) |
| 121 | |
| 122 | //plain C versions |
| 123 | #include "rgb2rgb_template.c" |
| 124 | |
| 125 | /* |
| 126 | * RGB15->RGB16 original by Strepto/Astral |
| 127 | * ported to gcc & bugfixed : A'rpi |
| 128 | * MMXEXT, 3DNOW optimization by Nick Kurshev |
| 129 | * 32-bit C version, and and&add trick by Michael Niedermayer |
| 130 | */ |
| 131 | |
| 132 | av_cold void sws_rgb2rgb_init(void) |
| 133 | { |
| 134 | rgb2rgb_init_c(); |
| 135 | if (ARCH_X86) |
| 136 | rgb2rgb_init_x86(); |
| 137 | } |
| 138 | |
| 139 | void rgb32to24(const uint8_t *src, uint8_t *dst, int src_size) |
| 140 | { |
| 141 | int i, num_pixels = src_size >> 2; |
| 142 | |
| 143 | for (i = 0; i < num_pixels; i++) { |
| 144 | #if HAVE_BIGENDIAN |
| 145 | /* RGB32 (= A,B,G,R) -> BGR24 (= B,G,R) */ |
| 146 | dst[3 * i + 0] = src[4 * i + 1]; |
| 147 | dst[3 * i + 1] = src[4 * i + 2]; |
| 148 | dst[3 * i + 2] = src[4 * i + 3]; |
| 149 | #else |
| 150 | dst[3 * i + 0] = src[4 * i + 2]; |
| 151 | dst[3 * i + 1] = src[4 * i + 1]; |
| 152 | dst[3 * i + 2] = src[4 * i + 0]; |
| 153 | #endif |
| 154 | } |
| 155 | } |
| 156 | |
| 157 | void rgb24to32(const uint8_t *src, uint8_t *dst, int src_size) |
| 158 | { |
| 159 | int i; |
| 160 | |
| 161 | for (i = 0; 3 * i < src_size; i++) { |
| 162 | #if HAVE_BIGENDIAN |
| 163 | /* RGB24 (= R, G, B) -> BGR32 (= A, R, G, B) */ |
| 164 | dst[4 * i + 0] = 255; |
| 165 | dst[4 * i + 1] = src[3 * i + 0]; |
| 166 | dst[4 * i + 2] = src[3 * i + 1]; |
| 167 | dst[4 * i + 3] = src[3 * i + 2]; |
| 168 | #else |
| 169 | dst[4 * i + 0] = src[3 * i + 2]; |
| 170 | dst[4 * i + 1] = src[3 * i + 1]; |
| 171 | dst[4 * i + 2] = src[3 * i + 0]; |
| 172 | dst[4 * i + 3] = 255; |
| 173 | #endif |
| 174 | } |
| 175 | } |
| 176 | |
| 177 | void rgb16tobgr32(const uint8_t *src, uint8_t *dst, int src_size) |
| 178 | { |
| 179 | uint8_t *d = dst; |
| 180 | const uint16_t *s = (const uint16_t *)src; |
| 181 | const uint16_t *end = s + src_size / 2; |
| 182 | |
| 183 | while (s < end) { |
| 184 | register uint16_t bgr = *s++; |
| 185 | #if HAVE_BIGENDIAN |
| 186 | *d++ = 255; |
| 187 | *d++ = ((bgr&0x001F)<<3) | ((bgr&0x001F)>> 2); |
| 188 | *d++ = ((bgr&0x07E0)>>3) | ((bgr&0x07E0)>> 9); |
| 189 | *d++ = ((bgr&0xF800)>>8) | ((bgr&0xF800)>>13); |
| 190 | #else |
| 191 | *d++ = ((bgr&0xF800)>>8) | ((bgr&0xF800)>>13); |
| 192 | *d++ = ((bgr&0x07E0)>>3) | ((bgr&0x07E0)>> 9); |
| 193 | *d++ = ((bgr&0x001F)<<3) | ((bgr&0x001F)>> 2); |
| 194 | *d++ = 255; |
| 195 | #endif |
| 196 | } |
| 197 | } |
| 198 | |
| 199 | void rgb12to15(const uint8_t *src, uint8_t *dst, int src_size) |
| 200 | { |
| 201 | uint16_t rgb, r, g, b; |
| 202 | uint16_t *d = (uint16_t *)dst; |
| 203 | const uint16_t *s = (const uint16_t *)src; |
| 204 | const uint16_t *end = s + src_size / 2; |
| 205 | |
| 206 | while (s < end) { |
| 207 | rgb = *s++; |
| 208 | r = rgb & 0xF00; |
| 209 | g = rgb & 0x0F0; |
| 210 | b = rgb & 0x00F; |
| 211 | r = (r << 3) | ((r & 0x800) >> 1); |
| 212 | g = (g << 2) | ((g & 0x080) >> 2); |
| 213 | b = (b << 1) | ( b >> 3); |
| 214 | *d++ = r | g | b; |
| 215 | } |
| 216 | } |
| 217 | |
| 218 | void rgb16to24(const uint8_t *src, uint8_t *dst, int src_size) |
| 219 | { |
| 220 | uint8_t *d = dst; |
| 221 | const uint16_t *s = (const uint16_t *)src; |
| 222 | const uint16_t *end = s + src_size / 2; |
| 223 | |
| 224 | while (s < end) { |
| 225 | register uint16_t bgr = *s++; |
| 226 | *d++ = ((bgr&0xF800)>>8) | ((bgr&0xF800)>>13); |
| 227 | *d++ = ((bgr&0x07E0)>>3) | ((bgr&0x07E0)>> 9); |
| 228 | *d++ = ((bgr&0x001F)<<3) | ((bgr&0x001F)>> 2); |
| 229 | } |
| 230 | } |
| 231 | |
| 232 | void rgb16tobgr16(const uint8_t *src, uint8_t *dst, int src_size) |
| 233 | { |
| 234 | int i, num_pixels = src_size >> 1; |
| 235 | |
| 236 | for (i = 0; i < num_pixels; i++) { |
| 237 | unsigned rgb = ((const uint16_t *)src)[i]; |
| 238 | ((uint16_t *)dst)[i] = (rgb >> 11) | (rgb & 0x7E0) | (rgb << 11); |
| 239 | } |
| 240 | } |
| 241 | |
| 242 | void rgb16tobgr15(const uint8_t *src, uint8_t *dst, int src_size) |
| 243 | { |
| 244 | int i, num_pixels = src_size >> 1; |
| 245 | |
| 246 | for (i = 0; i < num_pixels; i++) { |
| 247 | unsigned rgb = ((const uint16_t *)src)[i]; |
| 248 | ((uint16_t *)dst)[i] = (rgb >> 11) | ((rgb & 0x7C0) >> 1) | ((rgb & 0x1F) << 10); |
| 249 | } |
| 250 | } |
| 251 | |
| 252 | void rgb15tobgr32(const uint8_t *src, uint8_t *dst, int src_size) |
| 253 | { |
| 254 | uint8_t *d = dst; |
| 255 | const uint16_t *s = (const uint16_t *)src; |
| 256 | const uint16_t *end = s + src_size / 2; |
| 257 | |
| 258 | while (s < end) { |
| 259 | register uint16_t bgr = *s++; |
| 260 | #if HAVE_BIGENDIAN |
| 261 | *d++ = 255; |
| 262 | *d++ = ((bgr&0x001F)<<3) | ((bgr&0x001F)>> 2); |
| 263 | *d++ = ((bgr&0x03E0)>>2) | ((bgr&0x03E0)>> 7); |
| 264 | *d++ = ((bgr&0x7C00)>>7) | ((bgr&0x7C00)>>12); |
| 265 | #else |
| 266 | *d++ = ((bgr&0x7C00)>>7) | ((bgr&0x7C00)>>12); |
| 267 | *d++ = ((bgr&0x03E0)>>2) | ((bgr&0x03E0)>> 7); |
| 268 | *d++ = ((bgr&0x001F)<<3) | ((bgr&0x001F)>> 2); |
| 269 | *d++ = 255; |
| 270 | #endif |
| 271 | } |
| 272 | } |
| 273 | |
| 274 | void rgb15to24(const uint8_t *src, uint8_t *dst, int src_size) |
| 275 | { |
| 276 | uint8_t *d = dst; |
| 277 | const uint16_t *s = (const uint16_t *)src; |
| 278 | const uint16_t *end = s + src_size / 2; |
| 279 | |
| 280 | while (s < end) { |
| 281 | register uint16_t bgr = *s++; |
| 282 | *d++ = ((bgr&0x7C00)>>7) | ((bgr&0x7C00)>>12); |
| 283 | *d++ = ((bgr&0x03E0)>>2) | ((bgr&0x03E0)>> 7); |
| 284 | *d++ = ((bgr&0x001F)<<3) | ((bgr&0x001F)>> 2); |
| 285 | } |
| 286 | } |
| 287 | |
| 288 | void rgb15tobgr16(const uint8_t *src, uint8_t *dst, int src_size) |
| 289 | { |
| 290 | int i, num_pixels = src_size >> 1; |
| 291 | |
| 292 | for (i = 0; i < num_pixels; i++) { |
| 293 | unsigned rgb = ((const uint16_t *)src)[i]; |
| 294 | ((uint16_t *)dst)[i] = ((rgb & 0x7C00) >> 10) | ((rgb & 0x3E0) << 1) | (rgb << 11); |
| 295 | } |
| 296 | } |
| 297 | |
| 298 | void rgb15tobgr15(const uint8_t *src, uint8_t *dst, int src_size) |
| 299 | { |
| 300 | int i, num_pixels = src_size >> 1; |
| 301 | |
| 302 | for (i = 0; i < num_pixels; i++) { |
| 303 | unsigned rgb = ((const uint16_t *)src)[i]; |
| 304 | unsigned br = rgb & 0x7C1F; |
| 305 | ((uint16_t *)dst)[i] = (br >> 10) | (rgb & 0x3E0) | (br << 10); |
| 306 | } |
| 307 | } |
| 308 | |
| 309 | void rgb12tobgr12(const uint8_t *src, uint8_t *dst, int src_size) |
| 310 | { |
| 311 | uint16_t *d = (uint16_t *)dst; |
| 312 | uint16_t *s = (uint16_t *)src; |
| 313 | int i, num_pixels = src_size >> 1; |
| 314 | |
| 315 | for (i = 0; i < num_pixels; i++) { |
| 316 | unsigned rgb = s[i]; |
| 317 | d[i] = (rgb << 8 | rgb & 0xF0 | rgb >> 8) & 0xFFF; |
| 318 | } |
| 319 | } |
| 320 | |
| 321 | |
| 322 | #define DEFINE_SHUFFLE_BYTES(a, b, c, d) \ |
| 323 | void shuffle_bytes_ ## a ## b ## c ## d(const uint8_t *src, \ |
| 324 | uint8_t *dst, int src_size) \ |
| 325 | { \ |
| 326 | int i; \ |
| 327 | \ |
| 328 | for (i = 0; i < src_size; i += 4) { \ |
| 329 | dst[i + 0] = src[i + a]; \ |
| 330 | dst[i + 1] = src[i + b]; \ |
| 331 | dst[i + 2] = src[i + c]; \ |
| 332 | dst[i + 3] = src[i + d]; \ |
| 333 | } \ |
| 334 | } |
| 335 | |
| 336 | DEFINE_SHUFFLE_BYTES(0, 3, 2, 1) |
| 337 | DEFINE_SHUFFLE_BYTES(1, 2, 3, 0) |
| 338 | DEFINE_SHUFFLE_BYTES(3, 0, 1, 2) |
| 339 | DEFINE_SHUFFLE_BYTES(3, 2, 1, 0) |
| 340 | |
| 341 | #define DEFINE_RGB48TOBGR48(need_bswap, swap) \ |
| 342 | void rgb48tobgr48_ ## need_bswap(const uint8_t *src, \ |
| 343 | uint8_t *dst, int src_size) \ |
| 344 | { \ |
| 345 | uint16_t *d = (uint16_t *)dst; \ |
| 346 | uint16_t *s = (uint16_t *)src; \ |
| 347 | int i, num_pixels = src_size >> 1; \ |
| 348 | \ |
| 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 ]; \ |
| 353 | } \ |
| 354 | } |
| 355 | |
| 356 | DEFINE_RGB48TOBGR48(nobswap, 0) |
| 357 | DEFINE_RGB48TOBGR48(bswap, 1) |
| 358 | |
| 359 | #define DEFINE_RGB64TOBGR48(need_bswap, swap) \ |
| 360 | void rgb64tobgr48_ ## need_bswap(const uint8_t *src, \ |
| 361 | uint8_t *dst, int src_size) \ |
| 362 | { \ |
| 363 | uint16_t *d = (uint16_t *)dst; \ |
| 364 | uint16_t *s = (uint16_t *)src; \ |
| 365 | int i, num_pixels = src_size >> 3; \ |
| 366 | \ |
| 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 ]; \ |
| 371 | } \ |
| 372 | } |
| 373 | |
| 374 | DEFINE_RGB64TOBGR48(nobswap, 0) |
| 375 | DEFINE_RGB64TOBGR48(bswap, 1) |
| 376 | |
| 377 | #define DEFINE_RGB64TO48(need_bswap, swap) \ |
| 378 | void rgb64to48_ ## need_bswap(const uint8_t *src, \ |
| 379 | uint8_t *dst, int src_size) \ |
| 380 | { \ |
| 381 | uint16_t *d = (uint16_t *)dst; \ |
| 382 | uint16_t *s = (uint16_t *)src; \ |
| 383 | int i, num_pixels = src_size >> 3; \ |
| 384 | \ |
| 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]; \ |
| 389 | } \ |
| 390 | } |
| 391 | |
| 392 | DEFINE_RGB64TO48(nobswap, 0) |
| 393 | DEFINE_RGB64TO48(bswap, 1) |