2 * Copyright (C) 2001-2003 Michael Niedermayer <michaelni@gmx.at>
4 * This file is part of Libav.
6 * Libav is free software; you can redistribute it and/or
7 * modify it under the terms of the GNU Lesser General Public
8 * License as published by the Free Software Foundation; either
9 * version 2.1 of the License, or (at your option) any later version.
11 * Libav is distributed in the hope that it will be useful,
12 * but WITHOUT ANY WARRANTY; without even the implied warranty of
13 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
14 * Lesser General Public License for more details.
16 * You should have received a copy of the GNU Lesser General Public
17 * License along with Libav; if not, write to the Free Software
18 * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
27 #include "libavutil/avutil.h"
28 #include "libavutil/bswap.h"
29 #include "libavutil/cpu.h"
30 #include "libavutil/intreadwrite.h"
31 #include "libavutil/mathematics.h"
32 #include "libavutil/pixdesc.h"
36 #include "swscale_internal.h"
38 #define RGB2YUV_SHIFT 15
39 #define BY ((int)(0.114 * 219 / 255 * (1 << RGB2YUV_SHIFT) + 0.5))
40 #define BV (-(int)(0.081 * 224 / 255 * (1 << RGB2YUV_SHIFT) + 0.5))
41 #define BU ((int)(0.500 * 224 / 255 * (1 << RGB2YUV_SHIFT) + 0.5))
42 #define GY ((int)(0.587 * 219 / 255 * (1 << RGB2YUV_SHIFT) + 0.5))
43 #define GV (-(int)(0.419 * 224 / 255 * (1 << RGB2YUV_SHIFT) + 0.5))
44 #define GU (-(int)(0.331 * 224 / 255 * (1 << RGB2YUV_SHIFT) + 0.5))
45 #define RY ((int)(0.299 * 219 / 255 * (1 << RGB2YUV_SHIFT) + 0.5))
46 #define RV ((int)(0.500 * 224 / 255 * (1 << RGB2YUV_SHIFT) + 0.5))
47 #define RU (-(int)(0.169 * 224 / 255 * (1 << RGB2YUV_SHIFT) + 0.5))
49 #define input_pixel(pos) (isBE(origin) ? AV_RB16(pos) : AV_RL16(pos))
51 #define r ((origin == AV_PIX_FMT_BGR48BE || origin == AV_PIX_FMT_BGR48LE) ? b_r : r_b)
52 #define b ((origin == AV_PIX_FMT_BGR48BE || origin == AV_PIX_FMT_BGR48LE) ? r_b : b_r)
54 static av_always_inline void rgb48ToY_c_template(uint16_t *dst,
55 const uint16_t *src, int width,
56 enum AVPixelFormat origin)
59 for (i = 0; i < width; i++) {
60 unsigned int r_b = input_pixel(&src[i * 3 + 0]);
61 unsigned int g = input_pixel(&src[i * 3 + 1]);
62 unsigned int b_r = input_pixel(&src[i * 3 + 2]);
64 dst[i] = (RY * r + GY * g + BY * b + (0x2001 << (RGB2YUV_SHIFT - 1))) >> RGB2YUV_SHIFT;
68 static av_always_inline void rgb48ToUV_c_template(uint16_t *dstU,
73 enum AVPixelFormat origin)
77 for (i = 0; i < width; i++) {
78 int r_b = input_pixel(&src1[i * 3 + 0]);
79 int g = input_pixel(&src1[i * 3 + 1]);
80 int b_r = input_pixel(&src1[i * 3 + 2]);
82 dstU[i] = (RU * r + GU * g + BU * b + (0x10001 << (RGB2YUV_SHIFT - 1))) >> RGB2YUV_SHIFT;
83 dstV[i] = (RV * r + GV * g + BV * b + (0x10001 << (RGB2YUV_SHIFT - 1))) >> RGB2YUV_SHIFT;
87 static av_always_inline void rgb48ToUV_half_c_template(uint16_t *dstU,
92 enum AVPixelFormat origin)
96 for (i = 0; i < width; i++) {
97 int r_b = (input_pixel(&src1[6 * i + 0]) +
98 input_pixel(&src1[6 * i + 3]) + 1) >> 1;
99 int g = (input_pixel(&src1[6 * i + 1]) +
100 input_pixel(&src1[6 * i + 4]) + 1) >> 1;
101 int b_r = (input_pixel(&src1[6 * i + 2]) +
102 input_pixel(&src1[6 * i + 5]) + 1) >> 1;
104 dstU[i] = (RU * r + GU * g + BU * b + (0x10001 << (RGB2YUV_SHIFT - 1))) >> RGB2YUV_SHIFT;
105 dstV[i] = (RV * r + GV * g + BV * b + (0x10001 << (RGB2YUV_SHIFT - 1))) >> RGB2YUV_SHIFT;
113 #define rgb48funcs(pattern, BE_LE, origin) \
114 static void pattern ## 48 ## BE_LE ## ToY_c(uint8_t *_dst, \
115 const uint8_t *_src, \
119 const uint16_t *src = (const uint16_t *)_src; \
120 uint16_t *dst = (uint16_t *)_dst; \
121 rgb48ToY_c_template(dst, src, width, origin); \
124 static void pattern ## 48 ## BE_LE ## ToUV_c(uint8_t *_dstU, \
126 const uint8_t *_src1, \
127 const uint8_t *_src2, \
131 const uint16_t *src1 = (const uint16_t *)_src1, \
132 *src2 = (const uint16_t *)_src2; \
133 uint16_t *dstU = (uint16_t *)_dstU, \
134 *dstV = (uint16_t *)_dstV; \
135 rgb48ToUV_c_template(dstU, dstV, src1, src2, width, origin); \
138 static void pattern ## 48 ## BE_LE ## ToUV_half_c(uint8_t *_dstU, \
140 const uint8_t *_src1, \
141 const uint8_t *_src2, \
145 const uint16_t *src1 = (const uint16_t *)_src1, \
146 *src2 = (const uint16_t *)_src2; \
147 uint16_t *dstU = (uint16_t *)_dstU, \
148 *dstV = (uint16_t *)_dstV; \
149 rgb48ToUV_half_c_template(dstU, dstV, src1, src2, width, origin); \
152 rgb48funcs(rgb, LE, AV_PIX_FMT_RGB48LE)
153 rgb48funcs(rgb, BE, AV_PIX_FMT_RGB48BE)
154 rgb48funcs(bgr, LE, AV_PIX_FMT_BGR48LE)
155 rgb48funcs(bgr, BE, AV_PIX_FMT_BGR48BE)
157 #define input_pixel(i) ((origin == AV_PIX_FMT_RGBA || \
158 origin == AV_PIX_FMT_BGRA || \
159 origin == AV_PIX_FMT_ARGB || \
160 origin == AV_PIX_FMT_ABGR) \
161 ? AV_RN32A(&src[(i) * 4]) \
162 : (isBE(origin) ? AV_RB16(&src[(i) * 2]) \
163 : AV_RL16(&src[(i) * 2])))
165 static av_always_inline void rgb16_32ToY_c_template(uint8_t *dst,
168 enum AVPixelFormat origin,
171 int maskr, int maskg,
173 int gsh, int bsh, int S)
175 const int ry = RY << rsh, gy = GY << gsh, by = BY << bsh;
176 const unsigned rnd = 33u << (S - 1);
179 for (i = 0; i < width; i++) {
180 int px = input_pixel(i) >> shp;
181 int b = (px & maskb) >> shb;
182 int g = (px & maskg) >> shg;
183 int r = (px & maskr) >> shr;
185 dst[i] = (ry * r + gy * g + by * b + rnd) >> S;
189 static av_always_inline void rgb16_32ToUV_c_template(uint8_t *dstU,
193 enum AVPixelFormat origin,
196 int maskr, int maskg,
198 int gsh, int bsh, int S)
200 const int ru = RU << rsh, gu = GU << gsh, bu = BU << bsh,
201 rv = RV << rsh, gv = GV << gsh, bv = BV << bsh;
202 const unsigned rnd = 257u << (S - 1);
205 for (i = 0; i < width; i++) {
206 int px = input_pixel(i) >> shp;
207 int b = (px & maskb) >> shb;
208 int g = (px & maskg) >> shg;
209 int r = (px & maskr) >> shr;
211 dstU[i] = (ru * r + gu * g + bu * b + rnd) >> S;
212 dstV[i] = (rv * r + gv * g + bv * b + rnd) >> S;
216 static av_always_inline void rgb16_32ToUV_half_c_template(uint8_t *dstU,
220 enum AVPixelFormat origin,
223 int maskr, int maskg,
225 int gsh, int bsh, int S)
227 const int ru = RU << rsh, gu = GU << gsh, bu = BU << bsh,
228 rv = RV << rsh, gv = GV << gsh, bv = BV << bsh,
229 maskgx = ~(maskr | maskb);
230 const unsigned rnd = 257u << S;
236 for (i = 0; i < width; i++) {
237 int px0 = input_pixel(2 * i + 0) >> shp;
238 int px1 = input_pixel(2 * i + 1) >> shp;
239 int b, r, g = (px0 & maskgx) + (px1 & maskgx);
240 int rb = px0 + px1 - g;
242 b = (rb & maskb) >> shb;
244 origin == AV_PIX_FMT_BGR565LE || origin == AV_PIX_FMT_BGR565BE ||
245 origin == AV_PIX_FMT_RGB565LE || origin == AV_PIX_FMT_RGB565BE) {
248 g = (g & maskg) >> shg;
250 r = (rb & maskr) >> shr;
252 dstU[i] = (ru * r + gu * g + bu * b + rnd) >> (S + 1);
253 dstV[i] = (rv * r + gv * g + bv * b + rnd) >> (S + 1);
259 #define rgb16_32_wrapper(fmt, name, shr, shg, shb, shp, maskr, \
260 maskg, maskb, rsh, gsh, bsh, S) \
261 static void name ## ToY_c(uint8_t *dst, const uint8_t *src, \
262 int width, uint32_t *unused) \
264 rgb16_32ToY_c_template(dst, src, width, fmt, shr, shg, shb, shp, \
265 maskr, maskg, maskb, rsh, gsh, bsh, S); \
268 static void name ## ToUV_c(uint8_t *dstU, uint8_t *dstV, \
269 const uint8_t *src, const uint8_t *dummy, \
270 int width, uint32_t *unused) \
272 rgb16_32ToUV_c_template(dstU, dstV, src, width, fmt, \
273 shr, shg, shb, shp, \
274 maskr, maskg, maskb, rsh, gsh, bsh, S); \
277 static void name ## ToUV_half_c(uint8_t *dstU, uint8_t *dstV, \
278 const uint8_t *src, \
279 const uint8_t *dummy, \
280 int width, uint32_t *unused) \
282 rgb16_32ToUV_half_c_template(dstU, dstV, src, width, fmt, \
283 shr, shg, shb, shp, \
284 maskr, maskg, maskb, \
288 rgb16_32_wrapper(AV_PIX_FMT_BGR32, bgr32, 16, 0, 0, 0, 0xFF0000, 0xFF00, 0x00FF, 8, 0, 8, RGB2YUV_SHIFT + 8)
289 rgb16_32_wrapper(AV_PIX_FMT_BGR32_1, bgr321, 16, 0, 0, 8, 0xFF0000, 0xFF00, 0x00FF, 8, 0, 8, RGB2YUV_SHIFT + 8)
290 rgb16_32_wrapper(AV_PIX_FMT_RGB32, rgb32, 0, 0, 16, 0, 0x00FF, 0xFF00, 0xFF0000, 8, 0, 8, RGB2YUV_SHIFT + 8)
291 rgb16_32_wrapper(AV_PIX_FMT_RGB32_1, rgb321, 0, 0, 16, 8, 0x00FF, 0xFF00, 0xFF0000, 8, 0, 8, RGB2YUV_SHIFT + 8)
292 rgb16_32_wrapper(AV_PIX_FMT_BGR565LE, bgr16le, 0, 0, 0, 0, 0x001F, 0x07E0, 0xF800, 11, 5, 0, RGB2YUV_SHIFT + 8)
293 rgb16_32_wrapper(AV_PIX_FMT_BGR555LE, bgr15le, 0, 0, 0, 0, 0x001F, 0x03E0, 0x7C00, 10, 5, 0, RGB2YUV_SHIFT + 7)
294 rgb16_32_wrapper(AV_PIX_FMT_BGR444LE, bgr12le, 0, 0, 0, 0, 0x000F, 0x00F0, 0x0F00, 8, 4, 0, RGB2YUV_SHIFT + 4)
295 rgb16_32_wrapper(AV_PIX_FMT_RGB565LE, rgb16le, 0, 0, 0, 0, 0xF800, 0x07E0, 0x001F, 0, 5, 11, RGB2YUV_SHIFT + 8)
296 rgb16_32_wrapper(AV_PIX_FMT_RGB555LE, rgb15le, 0, 0, 0, 0, 0x7C00, 0x03E0, 0x001F, 0, 5, 10, RGB2YUV_SHIFT + 7)
297 rgb16_32_wrapper(AV_PIX_FMT_RGB444LE, rgb12le, 0, 0, 0, 0, 0x0F00, 0x00F0, 0x000F, 0, 4, 8, RGB2YUV_SHIFT + 4)
298 rgb16_32_wrapper(AV_PIX_FMT_BGR565BE, bgr16be, 0, 0, 0, 0, 0x001F, 0x07E0, 0xF800, 11, 5, 0, RGB2YUV_SHIFT + 8)
299 rgb16_32_wrapper(AV_PIX_FMT_BGR555BE, bgr15be, 0, 0, 0, 0, 0x001F, 0x03E0, 0x7C00, 10, 5, 0, RGB2YUV_SHIFT + 7)
300 rgb16_32_wrapper(AV_PIX_FMT_BGR444BE, bgr12be, 0, 0, 0, 0, 0x000F, 0x00F0, 0x0F00, 8, 4, 0, RGB2YUV_SHIFT + 4)
301 rgb16_32_wrapper(AV_PIX_FMT_RGB565BE, rgb16be, 0, 0, 0, 0, 0xF800, 0x07E0, 0x001F, 0, 5, 11, RGB2YUV_SHIFT + 8)
302 rgb16_32_wrapper(AV_PIX_FMT_RGB555BE, rgb15be, 0, 0, 0, 0, 0x7C00, 0x03E0, 0x001F, 0, 5, 10, RGB2YUV_SHIFT + 7)
303 rgb16_32_wrapper(AV_PIX_FMT_RGB444BE, rgb12be, 0, 0, 0, 0, 0x0F00, 0x00F0, 0x000F, 0, 4, 8, RGB2YUV_SHIFT + 4)
305 static void abgrToA_c(uint8_t *dst, const uint8_t *src, int width,
309 for (i = 0; i < width; i++)
313 static void rgbaToA_c(uint8_t *dst, const uint8_t *src, int width,
317 for (i = 0; i < width; i++)
318 dst[i] = src[4 * i + 3];
321 static void palToY_c(uint8_t *dst, const uint8_t *src, int width, uint32_t *pal)
324 for (i = 0; i < width; i++) {
327 dst[i] = pal[d] & 0xFF;
331 static void palToUV_c(uint8_t *dstU, uint8_t *dstV,
332 const uint8_t *src1, const uint8_t *src2,
333 int width, uint32_t *pal)
336 assert(src1 == src2);
337 for (i = 0; i < width; i++) {
338 int p = pal[src1[i]];
345 static void monowhite2Y_c(uint8_t *dst, const uint8_t *src,
346 int width, uint32_t *unused)
349 width = (width + 7) >> 3;
350 for (i = 0; i < width; i++) {
352 for (j = 0; j < 8; j++)
353 dst[8 * i + j] = ((d >> (7 - j)) & 1) * 255;
357 static void monoblack2Y_c(uint8_t *dst, const uint8_t *src,
358 int width, uint32_t *unused)
361 width = (width + 7) >> 3;
362 for (i = 0; i < width; i++) {
364 for (j = 0; j < 8; j++)
365 dst[8 * i + j] = ((d >> (7 - j)) & 1) * 255;
369 static void yuy2ToY_c(uint8_t *dst, const uint8_t *src, int width,
373 for (i = 0; i < width; i++)
377 static void yuy2ToUV_c(uint8_t *dstU, uint8_t *dstV, const uint8_t *src1,
378 const uint8_t *src2, int width, uint32_t *unused)
381 for (i = 0; i < width; i++) {
382 dstU[i] = src1[4 * i + 1];
383 dstV[i] = src1[4 * i + 3];
385 assert(src1 == src2);
388 static void bswap16Y_c(uint8_t *_dst, const uint8_t *_src, int width,
392 const uint16_t *src = (const uint16_t *)_src;
393 uint16_t *dst = (uint16_t *)_dst;
394 for (i = 0; i < width; i++)
395 dst[i] = av_bswap16(src[i]);
398 static void bswap16UV_c(uint8_t *_dstU, uint8_t *_dstV, const uint8_t *_src1,
399 const uint8_t *_src2, int width, uint32_t *unused)
402 const uint16_t *src1 = (const uint16_t *)_src1,
403 *src2 = (const uint16_t *)_src2;
404 uint16_t *dstU = (uint16_t *)_dstU, *dstV = (uint16_t *)_dstV;
405 for (i = 0; i < width; i++) {
406 dstU[i] = av_bswap16(src1[i]);
407 dstV[i] = av_bswap16(src2[i]);
411 /* This is almost identical to the previous, end exists only because
412 * yuy2ToY/UV)(dst, src + 1, ...) would have 100% unaligned accesses. */
413 static void uyvyToY_c(uint8_t *dst, const uint8_t *src, int width,
417 for (i = 0; i < width; i++)
418 dst[i] = src[2 * i + 1];
421 static void uyvyToUV_c(uint8_t *dstU, uint8_t *dstV, const uint8_t *src1,
422 const uint8_t *src2, int width, uint32_t *unused)
425 for (i = 0; i < width; i++) {
426 dstU[i] = src1[4 * i + 0];
427 dstV[i] = src1[4 * i + 2];
429 assert(src1 == src2);
432 static av_always_inline void nvXXtoUV_c(uint8_t *dst1, uint8_t *dst2,
433 const uint8_t *src, int width)
436 for (i = 0; i < width; i++) {
437 dst1[i] = src[2 * i + 0];
438 dst2[i] = src[2 * i + 1];
442 static void nv12ToUV_c(uint8_t *dstU, uint8_t *dstV,
443 const uint8_t *src1, const uint8_t *src2,
444 int width, uint32_t *unused)
446 nvXXtoUV_c(dstU, dstV, src1, width);
449 static void nv21ToUV_c(uint8_t *dstU, uint8_t *dstV,
450 const uint8_t *src1, const uint8_t *src2,
451 int width, uint32_t *unused)
453 nvXXtoUV_c(dstV, dstU, src1, width);
456 #define input_pixel(pos) (isBE(origin) ? AV_RB16(pos) : AV_RL16(pos))
458 static void bgr24ToY_c(uint8_t *dst, const uint8_t *src,
459 int width, uint32_t *unused)
462 for (i = 0; i < width; i++) {
463 int b = src[i * 3 + 0];
464 int g = src[i * 3 + 1];
465 int r = src[i * 3 + 2];
467 dst[i] = ((RY * r + GY * g + BY * b + (33 << (RGB2YUV_SHIFT - 1))) >> RGB2YUV_SHIFT);
471 static void bgr24ToUV_c(uint8_t *dstU, uint8_t *dstV, const uint8_t *src1,
472 const uint8_t *src2, int width, uint32_t *unused)
475 for (i = 0; i < width; i++) {
476 int b = src1[3 * i + 0];
477 int g = src1[3 * i + 1];
478 int r = src1[3 * i + 2];
480 dstU[i] = (RU * r + GU * g + BU * b + (257 << (RGB2YUV_SHIFT - 1))) >> RGB2YUV_SHIFT;
481 dstV[i] = (RV * r + GV * g + BV * b + (257 << (RGB2YUV_SHIFT - 1))) >> RGB2YUV_SHIFT;
483 assert(src1 == src2);
486 static void bgr24ToUV_half_c(uint8_t *dstU, uint8_t *dstV, const uint8_t *src1,
487 const uint8_t *src2, int width, uint32_t *unused)
490 for (i = 0; i < width; i++) {
491 int b = src1[6 * i + 0] + src1[6 * i + 3];
492 int g = src1[6 * i + 1] + src1[6 * i + 4];
493 int r = src1[6 * i + 2] + src1[6 * i + 5];
495 dstU[i] = (RU * r + GU * g + BU * b + (257 << RGB2YUV_SHIFT)) >> (RGB2YUV_SHIFT + 1);
496 dstV[i] = (RV * r + GV * g + BV * b + (257 << RGB2YUV_SHIFT)) >> (RGB2YUV_SHIFT + 1);
498 assert(src1 == src2);
501 static void rgb24ToY_c(uint8_t *dst, const uint8_t *src, int width,
505 for (i = 0; i < width; i++) {
506 int r = src[i * 3 + 0];
507 int g = src[i * 3 + 1];
508 int b = src[i * 3 + 2];
510 dst[i] = ((RY * r + GY * g + BY * b + (33 << (RGB2YUV_SHIFT - 1))) >> RGB2YUV_SHIFT);
514 static void rgb24ToUV_c(uint8_t *dstU, uint8_t *dstV, const uint8_t *src1,
515 const uint8_t *src2, int width, uint32_t *unused)
518 assert(src1 == src2);
519 for (i = 0; i < width; i++) {
520 int r = src1[3 * i + 0];
521 int g = src1[3 * i + 1];
522 int b = src1[3 * i + 2];
524 dstU[i] = (RU * r + GU * g + BU * b + (257 << (RGB2YUV_SHIFT - 1))) >> RGB2YUV_SHIFT;
525 dstV[i] = (RV * r + GV * g + BV * b + (257 << (RGB2YUV_SHIFT - 1))) >> RGB2YUV_SHIFT;
529 static void rgb24ToUV_half_c(uint8_t *dstU, uint8_t *dstV, const uint8_t *src1,
530 const uint8_t *src2, int width, uint32_t *unused)
533 assert(src1 == src2);
534 for (i = 0; i < width; i++) {
535 int r = src1[6 * i + 0] + src1[6 * i + 3];
536 int g = src1[6 * i + 1] + src1[6 * i + 4];
537 int b = src1[6 * i + 2] + src1[6 * i + 5];
539 dstU[i] = (RU * r + GU * g + BU * b + (257 << RGB2YUV_SHIFT)) >> (RGB2YUV_SHIFT + 1);
540 dstV[i] = (RV * r + GV * g + BV * b + (257 << RGB2YUV_SHIFT)) >> (RGB2YUV_SHIFT + 1);
544 static void planar_rgb_to_y(uint8_t *dst, const uint8_t *src[4], int width)
547 for (i = 0; i < width; i++) {
552 dst[i] = ((RY * r + GY * g + BY * b + (33 << (RGB2YUV_SHIFT - 1))) >> RGB2YUV_SHIFT);
556 static void planar_rgb_to_uv(uint8_t *dstU, uint8_t *dstV, const uint8_t *src[4], int width)
559 for (i = 0; i < width; i++) {
564 dstU[i] = (RU * r + GU * g + BU * b + (257 << (RGB2YUV_SHIFT - 1))) >> RGB2YUV_SHIFT;
565 dstV[i] = (RV * r + GV * g + BV * b + (257 << (RGB2YUV_SHIFT - 1))) >> RGB2YUV_SHIFT;
570 is_be ? AV_RB16(src) : AV_RL16(src)
571 static av_always_inline void planar_rgb16_to_y(uint8_t *_dst, const uint8_t *_src[4],
572 int width, int bpc, int is_be)
575 const uint16_t **src = (const uint16_t **)_src;
576 uint16_t *dst = (uint16_t *)_dst;
577 for (i = 0; i < width; i++) {
578 int g = rdpx(src[0] + i);
579 int b = rdpx(src[1] + i);
580 int r = rdpx(src[2] + i);
582 dst[i] = ((RY * r + GY * g + BY * b + (33 << (RGB2YUV_SHIFT + bpc - 9))) >> RGB2YUV_SHIFT);
586 static void planar_rgb9le_to_y(uint8_t *dst, const uint8_t *src[4], int w)
588 planar_rgb16_to_y(dst, src, w, 9, 0);
591 static void planar_rgb9be_to_y(uint8_t *dst, const uint8_t *src[4], int w)
593 planar_rgb16_to_y(dst, src, w, 9, 1);
596 static void planar_rgb10le_to_y(uint8_t *dst, const uint8_t *src[4], int w)
598 planar_rgb16_to_y(dst, src, w, 10, 0);
601 static void planar_rgb10be_to_y(uint8_t *dst, const uint8_t *src[4], int w)
603 planar_rgb16_to_y(dst, src, w, 10, 1);
606 static void planar_rgb16le_to_y(uint8_t *dst, const uint8_t *src[4], int w)
608 planar_rgb16_to_y(dst, src, w, 16, 0);
611 static void planar_rgb16be_to_y(uint8_t *dst, const uint8_t *src[4], int w)
613 planar_rgb16_to_y(dst, src, w, 16, 1);
616 static av_always_inline void planar_rgb16_to_uv(uint8_t *_dstU, uint8_t *_dstV,
617 const uint8_t *_src[4], int width,
621 const uint16_t **src = (const uint16_t **)_src;
622 uint16_t *dstU = (uint16_t *)_dstU;
623 uint16_t *dstV = (uint16_t *)_dstV;
624 for (i = 0; i < width; i++) {
625 int g = rdpx(src[0] + i);
626 int b = rdpx(src[1] + i);
627 int r = rdpx(src[2] + i);
629 dstU[i] = (RU * r + GU * g + BU * b + (257 << (RGB2YUV_SHIFT + bpc - 9))) >> RGB2YUV_SHIFT;
630 dstV[i] = (RV * r + GV * g + BV * b + (257 << (RGB2YUV_SHIFT + bpc - 9))) >> RGB2YUV_SHIFT;
635 static void planar_rgb9le_to_uv(uint8_t *dstU, uint8_t *dstV,
636 const uint8_t *src[4], int w)
638 planar_rgb16_to_uv(dstU, dstV, src, w, 9, 0);
641 static void planar_rgb9be_to_uv(uint8_t *dstU, uint8_t *dstV,
642 const uint8_t *src[4], int w)
644 planar_rgb16_to_uv(dstU, dstV, src, w, 9, 1);
647 static void planar_rgb10le_to_uv(uint8_t *dstU, uint8_t *dstV,
648 const uint8_t *src[4], int w)
650 planar_rgb16_to_uv(dstU, dstV, src, w, 10, 0);
653 static void planar_rgb10be_to_uv(uint8_t *dstU, uint8_t *dstV,
654 const uint8_t *src[4], int w)
656 planar_rgb16_to_uv(dstU, dstV, src, w, 10, 1);
659 static void planar_rgb16le_to_uv(uint8_t *dstU, uint8_t *dstV,
660 const uint8_t *src[4], int w)
662 planar_rgb16_to_uv(dstU, dstV, src, w, 16, 0);
665 static void planar_rgb16be_to_uv(uint8_t *dstU, uint8_t *dstV,
666 const uint8_t *src[4], int w)
668 planar_rgb16_to_uv(dstU, dstV, src, w, 16, 1);
671 av_cold void ff_sws_init_input_funcs(SwsContext *c)
673 enum AVPixelFormat srcFormat = c->srcFormat;
677 case AV_PIX_FMT_YUYV422:
678 c->chrToYV12 = yuy2ToUV_c;
680 case AV_PIX_FMT_UYVY422:
681 c->chrToYV12 = uyvyToUV_c;
683 case AV_PIX_FMT_NV12:
684 c->chrToYV12 = nv12ToUV_c;
686 case AV_PIX_FMT_NV21:
687 c->chrToYV12 = nv21ToUV_c;
689 case AV_PIX_FMT_RGB8:
690 case AV_PIX_FMT_BGR8:
691 case AV_PIX_FMT_PAL8:
692 case AV_PIX_FMT_BGR4_BYTE:
693 case AV_PIX_FMT_RGB4_BYTE:
694 c->chrToYV12 = palToUV_c;
696 case AV_PIX_FMT_GBRP9LE:
697 c->readChrPlanar = planar_rgb9le_to_uv;
699 case AV_PIX_FMT_GBRP10LE:
700 c->readChrPlanar = planar_rgb10le_to_uv;
702 case AV_PIX_FMT_GBRP16LE:
703 c->readChrPlanar = planar_rgb16le_to_uv;
705 case AV_PIX_FMT_GBRP9BE:
706 c->readChrPlanar = planar_rgb9be_to_uv;
708 case AV_PIX_FMT_GBRP10BE:
709 c->readChrPlanar = planar_rgb10be_to_uv;
711 case AV_PIX_FMT_GBRP16BE:
712 c->readChrPlanar = planar_rgb16be_to_uv;
714 case AV_PIX_FMT_GBRP:
715 c->readChrPlanar = planar_rgb_to_uv;
718 case AV_PIX_FMT_YUV444P9LE:
719 case AV_PIX_FMT_YUV422P9LE:
720 case AV_PIX_FMT_YUV420P9LE:
721 case AV_PIX_FMT_YUV422P10LE:
722 case AV_PIX_FMT_YUV444P10LE:
723 case AV_PIX_FMT_YUV420P10LE:
724 case AV_PIX_FMT_YUV420P16LE:
725 case AV_PIX_FMT_YUV422P16LE:
726 case AV_PIX_FMT_YUV444P16LE:
727 case AV_PIX_FMT_YUVA444P9LE:
728 case AV_PIX_FMT_YUVA422P9LE:
729 case AV_PIX_FMT_YUVA420P9LE:
730 case AV_PIX_FMT_YUVA422P10LE:
731 case AV_PIX_FMT_YUVA444P10LE:
732 case AV_PIX_FMT_YUVA420P10LE:
733 case AV_PIX_FMT_YUVA420P16LE:
734 case AV_PIX_FMT_YUVA422P16LE:
735 case AV_PIX_FMT_YUVA444P16LE:
736 c->chrToYV12 = bswap16UV_c;
739 case AV_PIX_FMT_YUV444P9BE:
740 case AV_PIX_FMT_YUV422P9BE:
741 case AV_PIX_FMT_YUV420P9BE:
742 case AV_PIX_FMT_YUV444P10BE:
743 case AV_PIX_FMT_YUV422P10BE:
744 case AV_PIX_FMT_YUV420P10BE:
745 case AV_PIX_FMT_YUV420P16BE:
746 case AV_PIX_FMT_YUV422P16BE:
747 case AV_PIX_FMT_YUV444P16BE:
748 case AV_PIX_FMT_YUVA444P9BE:
749 case AV_PIX_FMT_YUVA422P9BE:
750 case AV_PIX_FMT_YUVA420P9BE:
751 case AV_PIX_FMT_YUVA422P10BE:
752 case AV_PIX_FMT_YUVA444P10BE:
753 case AV_PIX_FMT_YUVA420P10BE:
754 case AV_PIX_FMT_YUVA420P16BE:
755 case AV_PIX_FMT_YUVA422P16BE:
756 case AV_PIX_FMT_YUVA444P16BE:
757 c->chrToYV12 = bswap16UV_c;
761 if (c->chrSrcHSubSample) {
763 case AV_PIX_FMT_RGB48BE:
764 c->chrToYV12 = rgb48BEToUV_half_c;
766 case AV_PIX_FMT_RGB48LE:
767 c->chrToYV12 = rgb48LEToUV_half_c;
769 case AV_PIX_FMT_BGR48BE:
770 c->chrToYV12 = bgr48BEToUV_half_c;
772 case AV_PIX_FMT_BGR48LE:
773 c->chrToYV12 = bgr48LEToUV_half_c;
775 case AV_PIX_FMT_RGB32:
776 c->chrToYV12 = bgr32ToUV_half_c;
778 case AV_PIX_FMT_RGB32_1:
779 c->chrToYV12 = bgr321ToUV_half_c;
781 case AV_PIX_FMT_BGR24:
782 c->chrToYV12 = bgr24ToUV_half_c;
784 case AV_PIX_FMT_BGR565LE:
785 c->chrToYV12 = bgr16leToUV_half_c;
787 case AV_PIX_FMT_BGR565BE:
788 c->chrToYV12 = bgr16beToUV_half_c;
790 case AV_PIX_FMT_BGR555LE:
791 c->chrToYV12 = bgr15leToUV_half_c;
793 case AV_PIX_FMT_BGR555BE:
794 c->chrToYV12 = bgr15beToUV_half_c;
796 case AV_PIX_FMT_BGR444LE:
797 c->chrToYV12 = bgr12leToUV_half_c;
799 case AV_PIX_FMT_BGR444BE:
800 c->chrToYV12 = bgr12beToUV_half_c;
802 case AV_PIX_FMT_BGR32:
803 c->chrToYV12 = rgb32ToUV_half_c;
805 case AV_PIX_FMT_BGR32_1:
806 c->chrToYV12 = rgb321ToUV_half_c;
808 case AV_PIX_FMT_RGB24:
809 c->chrToYV12 = rgb24ToUV_half_c;
811 case AV_PIX_FMT_RGB565LE:
812 c->chrToYV12 = rgb16leToUV_half_c;
814 case AV_PIX_FMT_RGB565BE:
815 c->chrToYV12 = rgb16beToUV_half_c;
817 case AV_PIX_FMT_RGB555LE:
818 c->chrToYV12 = rgb15leToUV_half_c;
820 case AV_PIX_FMT_RGB555BE:
821 c->chrToYV12 = rgb15beToUV_half_c;
823 case AV_PIX_FMT_RGB444LE:
824 c->chrToYV12 = rgb12leToUV_half_c;
826 case AV_PIX_FMT_RGB444BE:
827 c->chrToYV12 = rgb12beToUV_half_c;
832 case AV_PIX_FMT_RGB48BE:
833 c->chrToYV12 = rgb48BEToUV_c;
835 case AV_PIX_FMT_RGB48LE:
836 c->chrToYV12 = rgb48LEToUV_c;
838 case AV_PIX_FMT_BGR48BE:
839 c->chrToYV12 = bgr48BEToUV_c;
841 case AV_PIX_FMT_BGR48LE:
842 c->chrToYV12 = bgr48LEToUV_c;
844 case AV_PIX_FMT_RGB32:
845 c->chrToYV12 = bgr32ToUV_c;
847 case AV_PIX_FMT_RGB32_1:
848 c->chrToYV12 = bgr321ToUV_c;
850 case AV_PIX_FMT_BGR24:
851 c->chrToYV12 = bgr24ToUV_c;
853 case AV_PIX_FMT_BGR565LE:
854 c->chrToYV12 = bgr16leToUV_c;
856 case AV_PIX_FMT_BGR565BE:
857 c->chrToYV12 = bgr16beToUV_c;
859 case AV_PIX_FMT_BGR555LE:
860 c->chrToYV12 = bgr15leToUV_c;
862 case AV_PIX_FMT_BGR555BE:
863 c->chrToYV12 = bgr15beToUV_c;
865 case AV_PIX_FMT_BGR444LE:
866 c->chrToYV12 = bgr12leToUV_c;
868 case AV_PIX_FMT_BGR444BE:
869 c->chrToYV12 = bgr12beToUV_c;
871 case AV_PIX_FMT_BGR32:
872 c->chrToYV12 = rgb32ToUV_c;
874 case AV_PIX_FMT_BGR32_1:
875 c->chrToYV12 = rgb321ToUV_c;
877 case AV_PIX_FMT_RGB24:
878 c->chrToYV12 = rgb24ToUV_c;
880 case AV_PIX_FMT_RGB565LE:
881 c->chrToYV12 = rgb16leToUV_c;
883 case AV_PIX_FMT_RGB565BE:
884 c->chrToYV12 = rgb16beToUV_c;
886 case AV_PIX_FMT_RGB555LE:
887 c->chrToYV12 = rgb15leToUV_c;
889 case AV_PIX_FMT_RGB555BE:
890 c->chrToYV12 = rgb15beToUV_c;
892 case AV_PIX_FMT_RGB444LE:
893 c->chrToYV12 = rgb12leToUV_c;
895 case AV_PIX_FMT_RGB444BE:
896 c->chrToYV12 = rgb12beToUV_c;
904 case AV_PIX_FMT_GBRP9LE:
905 c->readLumPlanar = planar_rgb9le_to_y;
907 case AV_PIX_FMT_GBRP10LE:
908 c->readLumPlanar = planar_rgb10le_to_y;
910 case AV_PIX_FMT_GBRP16LE:
911 c->readLumPlanar = planar_rgb16le_to_y;
913 case AV_PIX_FMT_GBRP9BE:
914 c->readLumPlanar = planar_rgb9be_to_y;
916 case AV_PIX_FMT_GBRP10BE:
917 c->readLumPlanar = planar_rgb10be_to_y;
919 case AV_PIX_FMT_GBRP16BE:
920 c->readLumPlanar = planar_rgb16be_to_y;
922 case AV_PIX_FMT_GBRP:
923 c->readLumPlanar = planar_rgb_to_y;
926 case AV_PIX_FMT_YUV444P9LE:
927 case AV_PIX_FMT_YUV422P9LE:
928 case AV_PIX_FMT_YUV420P9LE:
929 case AV_PIX_FMT_YUV444P10LE:
930 case AV_PIX_FMT_YUV422P10LE:
931 case AV_PIX_FMT_YUV420P10LE:
932 case AV_PIX_FMT_YUV420P16LE:
933 case AV_PIX_FMT_YUV422P16LE:
934 case AV_PIX_FMT_YUV444P16LE:
935 case AV_PIX_FMT_GRAY16LE:
936 c->lumToYV12 = bswap16Y_c;
938 case AV_PIX_FMT_YUVA444P9LE:
939 case AV_PIX_FMT_YUVA422P9LE:
940 case AV_PIX_FMT_YUVA420P9LE:
941 case AV_PIX_FMT_YUVA444P10LE:
942 case AV_PIX_FMT_YUVA422P10LE:
943 case AV_PIX_FMT_YUVA420P10LE:
944 case AV_PIX_FMT_YUVA420P16LE:
945 case AV_PIX_FMT_YUVA422P16LE:
946 case AV_PIX_FMT_YUVA444P16LE:
947 c->lumToYV12 = bswap16Y_c;
948 c->alpToYV12 = bswap16Y_c;
951 case AV_PIX_FMT_YUV444P9BE:
952 case AV_PIX_FMT_YUV422P9BE:
953 case AV_PIX_FMT_YUV420P9BE:
954 case AV_PIX_FMT_YUV444P10BE:
955 case AV_PIX_FMT_YUV422P10BE:
956 case AV_PIX_FMT_YUV420P10BE:
957 case AV_PIX_FMT_YUV420P16BE:
958 case AV_PIX_FMT_YUV422P16BE:
959 case AV_PIX_FMT_YUV444P16BE:
960 case AV_PIX_FMT_GRAY16BE:
961 c->lumToYV12 = bswap16Y_c;
963 case AV_PIX_FMT_YUVA444P9BE:
964 case AV_PIX_FMT_YUVA422P9BE:
965 case AV_PIX_FMT_YUVA420P9BE:
966 case AV_PIX_FMT_YUVA444P10BE:
967 case AV_PIX_FMT_YUVA422P10BE:
968 case AV_PIX_FMT_YUVA420P10BE:
969 case AV_PIX_FMT_YUVA420P16BE:
970 case AV_PIX_FMT_YUVA422P16BE:
971 case AV_PIX_FMT_YUVA444P16BE:
972 c->lumToYV12 = bswap16Y_c;
973 c->alpToYV12 = bswap16Y_c;
976 case AV_PIX_FMT_YUYV422:
977 case AV_PIX_FMT_Y400A:
978 c->lumToYV12 = yuy2ToY_c;
980 case AV_PIX_FMT_UYVY422:
981 c->lumToYV12 = uyvyToY_c;
983 case AV_PIX_FMT_BGR24:
984 c->lumToYV12 = bgr24ToY_c;
986 case AV_PIX_FMT_BGR565LE:
987 c->lumToYV12 = bgr16leToY_c;
989 case AV_PIX_FMT_BGR565BE:
990 c->lumToYV12 = bgr16beToY_c;
992 case AV_PIX_FMT_BGR555LE:
993 c->lumToYV12 = bgr15leToY_c;
995 case AV_PIX_FMT_BGR555BE:
996 c->lumToYV12 = bgr15beToY_c;
998 case AV_PIX_FMT_BGR444LE:
999 c->lumToYV12 = bgr12leToY_c;
1001 case AV_PIX_FMT_BGR444BE:
1002 c->lumToYV12 = bgr12beToY_c;
1004 case AV_PIX_FMT_RGB24:
1005 c->lumToYV12 = rgb24ToY_c;
1007 case AV_PIX_FMT_RGB565LE:
1008 c->lumToYV12 = rgb16leToY_c;
1010 case AV_PIX_FMT_RGB565BE:
1011 c->lumToYV12 = rgb16beToY_c;
1013 case AV_PIX_FMT_RGB555LE:
1014 c->lumToYV12 = rgb15leToY_c;
1016 case AV_PIX_FMT_RGB555BE:
1017 c->lumToYV12 = rgb15beToY_c;
1019 case AV_PIX_FMT_RGB444LE:
1020 c->lumToYV12 = rgb12leToY_c;
1022 case AV_PIX_FMT_RGB444BE:
1023 c->lumToYV12 = rgb12beToY_c;
1025 case AV_PIX_FMT_RGB8:
1026 case AV_PIX_FMT_BGR8:
1027 case AV_PIX_FMT_PAL8:
1028 case AV_PIX_FMT_BGR4_BYTE:
1029 case AV_PIX_FMT_RGB4_BYTE:
1030 c->lumToYV12 = palToY_c;
1032 case AV_PIX_FMT_MONOBLACK:
1033 c->lumToYV12 = monoblack2Y_c;
1035 case AV_PIX_FMT_MONOWHITE:
1036 c->lumToYV12 = monowhite2Y_c;
1038 case AV_PIX_FMT_RGB32:
1039 c->lumToYV12 = bgr32ToY_c;
1041 case AV_PIX_FMT_RGB32_1:
1042 c->lumToYV12 = bgr321ToY_c;
1044 case AV_PIX_FMT_BGR32:
1045 c->lumToYV12 = rgb32ToY_c;
1047 case AV_PIX_FMT_BGR32_1:
1048 c->lumToYV12 = rgb321ToY_c;
1050 case AV_PIX_FMT_RGB48BE:
1051 c->lumToYV12 = rgb48BEToY_c;
1053 case AV_PIX_FMT_RGB48LE:
1054 c->lumToYV12 = rgb48LEToY_c;
1056 case AV_PIX_FMT_BGR48BE:
1057 c->lumToYV12 = bgr48BEToY_c;
1059 case AV_PIX_FMT_BGR48LE:
1060 c->lumToYV12 = bgr48LEToY_c;
1064 switch (srcFormat) {
1065 case AV_PIX_FMT_BGRA:
1066 case AV_PIX_FMT_RGBA:
1067 c->alpToYV12 = rgbaToA_c;
1069 case AV_PIX_FMT_ABGR:
1070 case AV_PIX_FMT_ARGB:
1071 c->alpToYV12 = abgrToA_c;
1073 case AV_PIX_FMT_Y400A:
1074 c->alpToYV12 = uyvyToY_c;