pixfmt: Add yuv420p12 pixel format
[libav.git] / libswscale / output.c
CommitLineData
21449410
RB
1/*
2 * Copyright (C) 2001-2003 Michael Niedermayer <michaelni@gmx.at>
3 *
4 * This file is part of Libav.
5 *
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.
10 *
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.
15 *
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
19 */
20
21449410
RB
21#include <math.h>
22#include <stdint.h>
23#include <stdio.h>
24#include <string.h>
25
5a6e3c03 26#include "libavutil/attributes.h"
21449410
RB
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"
33#include "config.h"
34#include "rgb2rgb.h"
35#include "swscale.h"
36#include "swscale_internal.h"
37
38DECLARE_ALIGNED(8, static const uint8_t, dither_2x2_4)[2][8]={
39{ 1, 3, 1, 3, 1, 3, 1, 3, },
40{ 2, 0, 2, 0, 2, 0, 2, 0, },
41};
42
43DECLARE_ALIGNED(8, static const uint8_t, dither_2x2_8)[2][8]={
44{ 6, 2, 6, 2, 6, 2, 6, 2, },
45{ 0, 4, 0, 4, 0, 4, 0, 4, },
46};
47
aa2ba8c9 48DECLARE_ALIGNED(8, const uint8_t, ff_dither_4x4_16)[4][8] = {
21449410
RB
49{ 8, 4, 11, 7, 8, 4, 11, 7, },
50{ 2, 14, 1, 13, 2, 14, 1, 13, },
51{ 10, 6, 9, 5, 10, 6, 9, 5, },
52{ 0, 12, 3, 15, 0, 12, 3, 15, },
53};
54
aa2ba8c9 55DECLARE_ALIGNED(8, const uint8_t, ff_dither_8x8_32)[8][8] = {
21449410
RB
56{ 17, 9, 23, 15, 16, 8, 22, 14, },
57{ 5, 29, 3, 27, 4, 28, 2, 26, },
58{ 21, 13, 19, 11, 20, 12, 18, 10, },
59{ 0, 24, 6, 30, 1, 25, 7, 31, },
60{ 16, 8, 22, 14, 17, 9, 23, 15, },
61{ 4, 28, 2, 26, 5, 29, 3, 27, },
62{ 20, 12, 18, 10, 21, 13, 19, 11, },
63{ 1, 25, 7, 31, 0, 24, 6, 30, },
64};
65
aa2ba8c9 66DECLARE_ALIGNED(8, const uint8_t, ff_dither_8x8_73)[8][8] = {
21449410
RB
67{ 0, 55, 14, 68, 3, 58, 17, 72, },
68{ 37, 18, 50, 32, 40, 22, 54, 35, },
69{ 9, 64, 5, 59, 13, 67, 8, 63, },
70{ 46, 27, 41, 23, 49, 31, 44, 26, },
71{ 2, 57, 16, 71, 1, 56, 15, 70, },
72{ 39, 21, 52, 34, 38, 19, 51, 33, },
73{ 11, 66, 7, 62, 10, 65, 6, 60, },
74{ 48, 30, 43, 25, 47, 29, 42, 24, },
75};
76
77#if 1
aa2ba8c9 78DECLARE_ALIGNED(8, const uint8_t, ff_dither_8x8_220)[8][8] = {
21449410
RB
79{117, 62, 158, 103, 113, 58, 155, 100, },
80{ 34, 199, 21, 186, 31, 196, 17, 182, },
81{144, 89, 131, 76, 141, 86, 127, 72, },
82{ 0, 165, 41, 206, 10, 175, 52, 217, },
83{110, 55, 151, 96, 120, 65, 162, 107, },
84{ 28, 193, 14, 179, 38, 203, 24, 189, },
85{138, 83, 124, 69, 148, 93, 134, 79, },
86{ 7, 172, 48, 213, 3, 168, 45, 210, },
87};
88#elif 1
89// tries to correct a gamma of 1.5
aa2ba8c9 90DECLARE_ALIGNED(8, const uint8_t, ff_dither_8x8_220)[8][8] = {
21449410
RB
91{ 0, 143, 18, 200, 2, 156, 25, 215, },
92{ 78, 28, 125, 64, 89, 36, 138, 74, },
93{ 10, 180, 3, 161, 16, 195, 8, 175, },
94{109, 51, 93, 38, 121, 60, 105, 47, },
95{ 1, 152, 23, 210, 0, 147, 20, 205, },
96{ 85, 33, 134, 71, 81, 30, 130, 67, },
97{ 14, 190, 6, 171, 12, 185, 5, 166, },
98{117, 57, 101, 44, 113, 54, 97, 41, },
99};
100#elif 1
101// tries to correct a gamma of 2.0
aa2ba8c9 102DECLARE_ALIGNED(8, const uint8_t, ff_dither_8x8_220)[8][8] = {
21449410
RB
103{ 0, 124, 8, 193, 0, 140, 12, 213, },
104{ 55, 14, 104, 42, 66, 19, 119, 52, },
105{ 3, 168, 1, 145, 6, 187, 3, 162, },
106{ 86, 31, 70, 21, 99, 39, 82, 28, },
107{ 0, 134, 11, 206, 0, 129, 9, 200, },
108{ 62, 17, 114, 48, 58, 16, 109, 45, },
109{ 5, 181, 2, 157, 4, 175, 1, 151, },
110{ 95, 36, 78, 26, 90, 34, 74, 24, },
111};
112#else
113// tries to correct a gamma of 2.5
aa2ba8c9 114DECLARE_ALIGNED(8, const uint8_t, ff_dither_8x8_220)[8][8] = {
21449410
RB
115{ 0, 107, 3, 187, 0, 125, 6, 212, },
116{ 39, 7, 86, 28, 49, 11, 102, 36, },
117{ 1, 158, 0, 131, 3, 180, 1, 151, },
118{ 68, 19, 52, 12, 81, 25, 64, 17, },
119{ 0, 119, 5, 203, 0, 113, 4, 195, },
120{ 45, 9, 96, 33, 42, 8, 91, 30, },
121{ 2, 172, 1, 144, 2, 165, 0, 137, },
122{ 77, 23, 60, 15, 72, 21, 56, 14, },
123};
124#endif
125
126#define output_pixel(pos, val, bias, signedness) \
127 if (big_endian) { \
128 AV_WB16(pos, bias + av_clip_ ## signedness ## 16(val >> shift)); \
129 } else { \
130 AV_WL16(pos, bias + av_clip_ ## signedness ## 16(val >> shift)); \
131 }
132
133static av_always_inline void
134yuv2plane1_16_c_template(const int32_t *src, uint16_t *dest, int dstW,
135 int big_endian, int output_bits)
136{
137 int i;
138 int shift = 19 - output_bits;
139
140 for (i = 0; i < dstW; i++) {
141 int val = src[i] + (1 << (shift - 1));
142 output_pixel(&dest[i], val, 0, uint);
143 }
144}
145
146static av_always_inline void
147yuv2planeX_16_c_template(const int16_t *filter, int filterSize,
148 const int32_t **src, uint16_t *dest, int dstW,
149 int big_endian, int output_bits)
150{
151 int i;
152 int shift = 15 + 16 - output_bits;
153
154 for (i = 0; i < dstW; i++) {
155 int val = 1 << (30-output_bits);
156 int j;
157
158 /* range of val is [0,0x7FFFFFFF], so 31 bits, but with lanczos/spline
159 * filters (or anything with negative coeffs, the range can be slightly
160 * wider in both directions. To account for this overflow, we subtract
161 * a constant so it always fits in the signed range (assuming a
162 * reasonable filterSize), and re-add that at the end. */
163 val -= 0x40000000;
164 for (j = 0; j < filterSize; j++)
165 val += src[j][i] * filter[j];
166
167 output_pixel(&dest[i], val, 0x8000, int);
168 }
169}
170
171#undef output_pixel
172
173#define output_pixel(pos, val) \
174 if (big_endian) { \
175 AV_WB16(pos, av_clip_uintp2(val >> shift, output_bits)); \
176 } else { \
177 AV_WL16(pos, av_clip_uintp2(val >> shift, output_bits)); \
178 }
179
180static av_always_inline void
181yuv2plane1_10_c_template(const int16_t *src, uint16_t *dest, int dstW,
182 int big_endian, int output_bits)
183{
184 int i;
185 int shift = 15 - output_bits;
186
187 for (i = 0; i < dstW; i++) {
188 int val = src[i] + (1 << (shift - 1));
189 output_pixel(&dest[i], val);
190 }
191}
192
193static av_always_inline void
194yuv2planeX_10_c_template(const int16_t *filter, int filterSize,
195 const int16_t **src, uint16_t *dest, int dstW,
196 int big_endian, int output_bits)
197{
198 int i;
199 int shift = 11 + 16 - output_bits;
200
201 for (i = 0; i < dstW; i++) {
202 int val = 1 << (26-output_bits);
203 int j;
204
205 for (j = 0; j < filterSize; j++)
206 val += src[j][i] * filter[j];
207
208 output_pixel(&dest[i], val);
209 }
210}
211
212#undef output_pixel
213
214#define yuv2NBPS(bits, BE_LE, is_be, template_size, typeX_t) \
215static void yuv2plane1_ ## bits ## BE_LE ## _c(const int16_t *src, \
216 uint8_t *dest, int dstW, \
217 const uint8_t *dither, int offset)\
218{ \
219 yuv2plane1_ ## template_size ## _c_template((const typeX_t *) src, \
220 (uint16_t *) dest, dstW, is_be, bits); \
221}\
222static void yuv2planeX_ ## bits ## BE_LE ## _c(const int16_t *filter, int filterSize, \
223 const int16_t **src, uint8_t *dest, int dstW, \
224 const uint8_t *dither, int offset)\
225{ \
226 yuv2planeX_## template_size ## _c_template(filter, \
227 filterSize, (const typeX_t **) src, \
228 (uint16_t *) dest, dstW, is_be, bits); \
229}
230yuv2NBPS( 9, BE, 1, 10, int16_t)
231yuv2NBPS( 9, LE, 0, 10, int16_t)
232yuv2NBPS(10, BE, 1, 10, int16_t)
233yuv2NBPS(10, LE, 0, 10, int16_t)
85406e7a
LB
234yuv2NBPS(12, BE, 1, 10, int16_t)
235yuv2NBPS(12, LE, 0, 10, int16_t)
21449410
RB
236yuv2NBPS(16, BE, 1, 16, int32_t)
237yuv2NBPS(16, LE, 0, 16, int32_t)
238
239static void yuv2planeX_8_c(const int16_t *filter, int filterSize,
240 const int16_t **src, uint8_t *dest, int dstW,
241 const uint8_t *dither, int offset)
242{
243 int i;
244 for (i=0; i<dstW; i++) {
245 int val = dither[(i + offset) & 7] << 12;
246 int j;
247 for (j=0; j<filterSize; j++)
248 val += src[j][i] * filter[j];
249
250 dest[i]= av_clip_uint8(val>>19);
251 }
252}
253
254static void yuv2plane1_8_c(const int16_t *src, uint8_t *dest, int dstW,
255 const uint8_t *dither, int offset)
256{
257 int i;
258 for (i=0; i<dstW; i++) {
259 int val = (src[i] + dither[(i + offset) & 7]) >> 7;
260 dest[i]= av_clip_uint8(val);
261 }
262}
263
264static void yuv2nv12cX_c(SwsContext *c, const int16_t *chrFilter, int chrFilterSize,
265 const int16_t **chrUSrc, const int16_t **chrVSrc,
266 uint8_t *dest, int chrDstW)
267{
716d413c 268 enum AVPixelFormat dstFormat = c->dstFormat;
21449410
RB
269 const uint8_t *chrDither = c->chrDither8;
270 int i;
271
716d413c 272 if (dstFormat == AV_PIX_FMT_NV12)
21449410
RB
273 for (i=0; i<chrDstW; i++) {
274 int u = chrDither[i & 7] << 12;
275 int v = chrDither[(i + 3) & 7] << 12;
276 int j;
277 for (j=0; j<chrFilterSize; j++) {
278 u += chrUSrc[j][i] * chrFilter[j];
279 v += chrVSrc[j][i] * chrFilter[j];
280 }
281
282 dest[2*i]= av_clip_uint8(u>>19);
283 dest[2*i+1]= av_clip_uint8(v>>19);
284 }
285 else
286 for (i=0; i<chrDstW; i++) {
287 int u = chrDither[i & 7] << 12;
288 int v = chrDither[(i + 3) & 7] << 12;
289 int j;
290 for (j=0; j<chrFilterSize; j++) {
291 u += chrUSrc[j][i] * chrFilter[j];
292 v += chrVSrc[j][i] * chrFilter[j];
293 }
294
295 dest[2*i]= av_clip_uint8(v>>19);
296 dest[2*i+1]= av_clip_uint8(u>>19);
297 }
298}
299
25c42234
RB
300#define accumulate_bit(acc, val) \
301 acc <<= 1; \
302 acc |= (val) >= (128 + 110)
21449410 303#define output_pixel(pos, acc) \
716d413c 304 if (target == AV_PIX_FMT_MONOBLACK) { \
21449410
RB
305 pos = acc; \
306 } else { \
307 pos = ~acc; \
308 }
309
310static av_always_inline void
311yuv2mono_X_c_template(SwsContext *c, const int16_t *lumFilter,
312 const int16_t **lumSrc, int lumFilterSize,
313 const int16_t *chrFilter, const int16_t **chrUSrc,
314 const int16_t **chrVSrc, int chrFilterSize,
315 const int16_t **alpSrc, uint8_t *dest, int dstW,
716d413c 316 int y, enum AVPixelFormat target)
21449410 317{
aa2ba8c9 318 const uint8_t * const d128 = ff_dither_8x8_220[y&7];
21449410
RB
319 int i;
320 unsigned acc = 0;
321
bc0bdda7 322 for (i = 0; i < dstW; i += 2) {
21449410
RB
323 int j;
324 int Y1 = 1 << 18;
325 int Y2 = 1 << 18;
326
327 for (j = 0; j < lumFilterSize; j++) {
328 Y1 += lumSrc[j][i] * lumFilter[j];
329 Y2 += lumSrc[j][i+1] * lumFilter[j];
330 }
331 Y1 >>= 19;
332 Y2 >>= 19;
333 if ((Y1 | Y2) & 0x100) {
334 Y1 = av_clip_uint8(Y1);
335 Y2 = av_clip_uint8(Y2);
336 }
25c42234
RB
337 accumulate_bit(acc, Y1 + d128[(i + 0) & 7]);
338 accumulate_bit(acc, Y2 + d128[(i + 1) & 7]);
21449410
RB
339 if ((i & 7) == 6) {
340 output_pixel(*dest++, acc);
341 }
342 }
bc0bdda7
RB
343
344 if (i & 6) {
345 output_pixel(*dest, acc);
346 }
21449410
RB
347}
348
349static av_always_inline void
350yuv2mono_2_c_template(SwsContext *c, const int16_t *buf[2],
351 const int16_t *ubuf[2], const int16_t *vbuf[2],
352 const int16_t *abuf[2], uint8_t *dest, int dstW,
353 int yalpha, int uvalpha, int y,
716d413c 354 enum AVPixelFormat target)
21449410
RB
355{
356 const int16_t *buf0 = buf[0], *buf1 = buf[1];
aa2ba8c9 357 const uint8_t * const d128 = ff_dither_8x8_220[y & 7];
48606252 358 int yalpha1 = 4096 - yalpha;
21449410
RB
359 int i;
360
bc0bdda7 361 for (i = 0; i < dstW; i += 8) {
25c42234
RB
362 int Y, acc = 0;
363
364 Y = (buf0[i + 0] * yalpha1 + buf1[i + 0] * yalpha) >> 19;
365 accumulate_bit(acc, Y + d128[0]);
366 Y = (buf0[i + 1] * yalpha1 + buf1[i + 1] * yalpha) >> 19;
367 accumulate_bit(acc, Y + d128[1]);
368 Y = (buf0[i + 2] * yalpha1 + buf1[i + 2] * yalpha) >> 19;
369 accumulate_bit(acc, Y + d128[2]);
370 Y = (buf0[i + 3] * yalpha1 + buf1[i + 3] * yalpha) >> 19;
371 accumulate_bit(acc, Y + d128[3]);
372 Y = (buf0[i + 4] * yalpha1 + buf1[i + 4] * yalpha) >> 19;
373 accumulate_bit(acc, Y + d128[4]);
374 Y = (buf0[i + 5] * yalpha1 + buf1[i + 5] * yalpha) >> 19;
375 accumulate_bit(acc, Y + d128[5]);
376 Y = (buf0[i + 6] * yalpha1 + buf1[i + 6] * yalpha) >> 19;
377 accumulate_bit(acc, Y + d128[6]);
378 Y = (buf0[i + 7] * yalpha1 + buf1[i + 7] * yalpha) >> 19;
379 accumulate_bit(acc, Y + d128[7]);
380
21449410
RB
381 output_pixel(*dest++, acc);
382 }
383}
384
385static av_always_inline void
386yuv2mono_1_c_template(SwsContext *c, const int16_t *buf0,
387 const int16_t *ubuf[2], const int16_t *vbuf[2],
388 const int16_t *abuf0, uint8_t *dest, int dstW,
716d413c 389 int uvalpha, int y, enum AVPixelFormat target)
21449410 390{
aa2ba8c9 391 const uint8_t * const d128 = ff_dither_8x8_220[y & 7];
21449410
RB
392 int i;
393
bc0bdda7 394 for (i = 0; i < dstW; i += 8) {
25c42234
RB
395 int acc = 0;
396
397 accumulate_bit(acc, (buf0[i + 0] >> 7) + d128[0]);
398 accumulate_bit(acc, (buf0[i + 1] >> 7) + d128[1]);
399 accumulate_bit(acc, (buf0[i + 2] >> 7) + d128[2]);
400 accumulate_bit(acc, (buf0[i + 3] >> 7) + d128[3]);
401 accumulate_bit(acc, (buf0[i + 4] >> 7) + d128[4]);
402 accumulate_bit(acc, (buf0[i + 5] >> 7) + d128[5]);
403 accumulate_bit(acc, (buf0[i + 6] >> 7) + d128[6]);
404 accumulate_bit(acc, (buf0[i + 7] >> 7) + d128[7]);
405
21449410
RB
406 output_pixel(*dest++, acc);
407 }
408}
409
410#undef output_pixel
25c42234 411#undef accumulate_bit
21449410
RB
412
413#define YUV2PACKEDWRAPPER(name, base, ext, fmt) \
414static void name ## ext ## _X_c(SwsContext *c, const int16_t *lumFilter, \
415 const int16_t **lumSrc, int lumFilterSize, \
416 const int16_t *chrFilter, const int16_t **chrUSrc, \
417 const int16_t **chrVSrc, int chrFilterSize, \
418 const int16_t **alpSrc, uint8_t *dest, int dstW, \
419 int y) \
420{ \
421 name ## base ## _X_c_template(c, lumFilter, lumSrc, lumFilterSize, \
422 chrFilter, chrUSrc, chrVSrc, chrFilterSize, \
423 alpSrc, dest, dstW, y, fmt); \
424} \
425 \
426static void name ## ext ## _2_c(SwsContext *c, const int16_t *buf[2], \
427 const int16_t *ubuf[2], const int16_t *vbuf[2], \
428 const int16_t *abuf[2], uint8_t *dest, int dstW, \
429 int yalpha, int uvalpha, int y) \
430{ \
431 name ## base ## _2_c_template(c, buf, ubuf, vbuf, abuf, \
432 dest, dstW, yalpha, uvalpha, y, fmt); \
433} \
434 \
435static void name ## ext ## _1_c(SwsContext *c, const int16_t *buf0, \
436 const int16_t *ubuf[2], const int16_t *vbuf[2], \
437 const int16_t *abuf0, uint8_t *dest, int dstW, \
438 int uvalpha, int y) \
439{ \
440 name ## base ## _1_c_template(c, buf0, ubuf, vbuf, \
441 abuf0, dest, dstW, uvalpha, \
442 y, fmt); \
443}
444
716d413c
AK
445YUV2PACKEDWRAPPER(yuv2mono,, white, AV_PIX_FMT_MONOWHITE)
446YUV2PACKEDWRAPPER(yuv2mono,, black, AV_PIX_FMT_MONOBLACK)
21449410
RB
447
448#define output_pixels(pos, Y1, U, Y2, V) \
716d413c 449 if (target == AV_PIX_FMT_YUYV422) { \
21449410
RB
450 dest[pos + 0] = Y1; \
451 dest[pos + 1] = U; \
452 dest[pos + 2] = Y2; \
453 dest[pos + 3] = V; \
92b099da
VG
454 } else if (target == AV_PIX_FMT_YVYU422) { \
455 dest[pos + 0] = Y1; \
456 dest[pos + 1] = V; \
457 dest[pos + 2] = Y2; \
458 dest[pos + 3] = U; \
459 } else { /* AV_PIX_FMT_UYVY422 */ \
21449410
RB
460 dest[pos + 0] = U; \
461 dest[pos + 1] = Y1; \
462 dest[pos + 2] = V; \
463 dest[pos + 3] = Y2; \
464 }
465
466static av_always_inline void
467yuv2422_X_c_template(SwsContext *c, const int16_t *lumFilter,
468 const int16_t **lumSrc, int lumFilterSize,
469 const int16_t *chrFilter, const int16_t **chrUSrc,
470 const int16_t **chrVSrc, int chrFilterSize,
471 const int16_t **alpSrc, uint8_t *dest, int dstW,
716d413c 472 int y, enum AVPixelFormat target)
21449410
RB
473{
474 int i;
475
9206ac80 476 for (i = 0; i < ((dstW + 1) >> 1); i++) {
21449410
RB
477 int j;
478 int Y1 = 1 << 18;
479 int Y2 = 1 << 18;
480 int U = 1 << 18;
481 int V = 1 << 18;
482
483 for (j = 0; j < lumFilterSize; j++) {
484 Y1 += lumSrc[j][i * 2] * lumFilter[j];
485 Y2 += lumSrc[j][i * 2 + 1] * lumFilter[j];
486 }
487 for (j = 0; j < chrFilterSize; j++) {
488 U += chrUSrc[j][i] * chrFilter[j];
489 V += chrVSrc[j][i] * chrFilter[j];
490 }
491 Y1 >>= 19;
492 Y2 >>= 19;
493 U >>= 19;
494 V >>= 19;
495 if ((Y1 | Y2 | U | V) & 0x100) {
496 Y1 = av_clip_uint8(Y1);
497 Y2 = av_clip_uint8(Y2);
498 U = av_clip_uint8(U);
499 V = av_clip_uint8(V);
500 }
501 output_pixels(4*i, Y1, U, Y2, V);
502 }
503}
504
505static av_always_inline void
506yuv2422_2_c_template(SwsContext *c, const int16_t *buf[2],
507 const int16_t *ubuf[2], const int16_t *vbuf[2],
508 const int16_t *abuf[2], uint8_t *dest, int dstW,
509 int yalpha, int uvalpha, int y,
716d413c 510 enum AVPixelFormat target)
21449410
RB
511{
512 const int16_t *buf0 = buf[0], *buf1 = buf[1],
513 *ubuf0 = ubuf[0], *ubuf1 = ubuf[1],
514 *vbuf0 = vbuf[0], *vbuf1 = vbuf[1];
48606252
RB
515 int yalpha1 = 4096 - yalpha;
516 int uvalpha1 = 4096 - uvalpha;
21449410
RB
517 int i;
518
9206ac80 519 for (i = 0; i < ((dstW + 1) >> 1); i++) {
21449410
RB
520 int Y1 = (buf0[i * 2] * yalpha1 + buf1[i * 2] * yalpha) >> 19;
521 int Y2 = (buf0[i * 2 + 1] * yalpha1 + buf1[i * 2 + 1] * yalpha) >> 19;
522 int U = (ubuf0[i] * uvalpha1 + ubuf1[i] * uvalpha) >> 19;
523 int V = (vbuf0[i] * uvalpha1 + vbuf1[i] * uvalpha) >> 19;
524
9487fb4d
RB
525 Y1 = av_clip_uint8(Y1);
526 Y2 = av_clip_uint8(Y2);
527 U = av_clip_uint8(U);
528 V = av_clip_uint8(V);
529
21449410
RB
530 output_pixels(i * 4, Y1, U, Y2, V);
531 }
532}
533
534static av_always_inline void
535yuv2422_1_c_template(SwsContext *c, const int16_t *buf0,
536 const int16_t *ubuf[2], const int16_t *vbuf[2],
537 const int16_t *abuf0, uint8_t *dest, int dstW,
716d413c 538 int uvalpha, int y, enum AVPixelFormat target)
21449410 539{
1ca7dc60 540 const int16_t *ubuf0 = ubuf[0], *vbuf0 = vbuf[0];
21449410
RB
541 int i;
542
543 if (uvalpha < 2048) {
9206ac80 544 for (i = 0; i < ((dstW + 1) >> 1); i++) {
21449410
RB
545 int Y1 = buf0[i * 2] >> 7;
546 int Y2 = buf0[i * 2 + 1] >> 7;
1ca7dc60
RB
547 int U = ubuf0[i] >> 7;
548 int V = vbuf0[i] >> 7;
21449410 549
9487fb4d
RB
550 Y1 = av_clip_uint8(Y1);
551 Y2 = av_clip_uint8(Y2);
552 U = av_clip_uint8(U);
553 V = av_clip_uint8(V);
554
21449410
RB
555 output_pixels(i * 4, Y1, U, Y2, V);
556 }
557 } else {
1ca7dc60 558 const int16_t *ubuf1 = ubuf[1], *vbuf1 = vbuf[1];
9206ac80 559 for (i = 0; i < ((dstW + 1) >> 1); i++) {
21449410
RB
560 int Y1 = buf0[i * 2] >> 7;
561 int Y2 = buf0[i * 2 + 1] >> 7;
562 int U = (ubuf0[i] + ubuf1[i]) >> 8;
563 int V = (vbuf0[i] + vbuf1[i]) >> 8;
564
9487fb4d
RB
565 Y1 = av_clip_uint8(Y1);
566 Y2 = av_clip_uint8(Y2);
567 U = av_clip_uint8(U);
568 V = av_clip_uint8(V);
569
21449410
RB
570 output_pixels(i * 4, Y1, U, Y2, V);
571 }
572 }
573}
574
575#undef output_pixels
576
716d413c 577YUV2PACKEDWRAPPER(yuv2, 422, yuyv422, AV_PIX_FMT_YUYV422)
92b099da 578YUV2PACKEDWRAPPER(yuv2, 422, yvyu422, AV_PIX_FMT_YVYU422)
716d413c 579YUV2PACKEDWRAPPER(yuv2, 422, uyvy422, AV_PIX_FMT_UYVY422)
21449410 580
716d413c
AK
581#define R_B ((target == AV_PIX_FMT_RGB48LE || target == AV_PIX_FMT_RGB48BE) ? R : B)
582#define B_R ((target == AV_PIX_FMT_RGB48LE || target == AV_PIX_FMT_RGB48BE) ? B : R)
21449410
RB
583#define output_pixel(pos, val) \
584 if (isBE(target)) { \
585 AV_WB16(pos, val); \
586 } else { \
587 AV_WL16(pos, val); \
588 }
589
590static av_always_inline void
591yuv2rgb48_X_c_template(SwsContext *c, const int16_t *lumFilter,
592 const int32_t **lumSrc, int lumFilterSize,
593 const int16_t *chrFilter, const int32_t **chrUSrc,
594 const int32_t **chrVSrc, int chrFilterSize,
595 const int32_t **alpSrc, uint16_t *dest, int dstW,
716d413c 596 int y, enum AVPixelFormat target)
21449410
RB
597{
598 int i;
599
9206ac80 600 for (i = 0; i < ((dstW + 1) >> 1); i++) {
21449410
RB
601 int j;
602 int Y1 = -0x40000000;
603 int Y2 = -0x40000000;
604 int U = -128 << 23; // 19
605 int V = -128 << 23;
606 int R, G, B;
607
608 for (j = 0; j < lumFilterSize; j++) {
609 Y1 += lumSrc[j][i * 2] * lumFilter[j];
610 Y2 += lumSrc[j][i * 2 + 1] * lumFilter[j];
611 }
612 for (j = 0; j < chrFilterSize; j++) {
613 U += chrUSrc[j][i] * chrFilter[j];
614 V += chrVSrc[j][i] * chrFilter[j];
615 }
616
41ed7ab4 617 // 8 bits: 12+15=27; 16 bits: 12+19=31
21449410
RB
618 Y1 >>= 14; // 10
619 Y1 += 0x10000;
620 Y2 >>= 14;
621 Y2 += 0x10000;
622 U >>= 14;
623 V >>= 14;
624
41ed7ab4 625 // 8 bits: 27 -> 17 bits, 16 bits: 31 - 14 = 17 bits
21449410
RB
626 Y1 -= c->yuv2rgb_y_offset;
627 Y2 -= c->yuv2rgb_y_offset;
628 Y1 *= c->yuv2rgb_y_coeff;
629 Y2 *= c->yuv2rgb_y_coeff;
630 Y1 += 1 << 13; // 21
631 Y2 += 1 << 13;
41ed7ab4 632 // 8 bits: 17 + 13 bits = 30 bits, 16 bits: 17 + 13 bits = 30 bits
21449410
RB
633
634 R = V * c->yuv2rgb_v2r_coeff;
635 G = V * c->yuv2rgb_v2g_coeff + U * c->yuv2rgb_u2g_coeff;
636 B = U * c->yuv2rgb_u2b_coeff;
637
41ed7ab4 638 // 8 bits: 30 - 22 = 8 bits, 16 bits: 30 bits - 14 = 16 bits
21449410
RB
639 output_pixel(&dest[0], av_clip_uintp2(R_B + Y1, 30) >> 14);
640 output_pixel(&dest[1], av_clip_uintp2( G + Y1, 30) >> 14);
641 output_pixel(&dest[2], av_clip_uintp2(B_R + Y1, 30) >> 14);
642 output_pixel(&dest[3], av_clip_uintp2(R_B + Y2, 30) >> 14);
643 output_pixel(&dest[4], av_clip_uintp2( G + Y2, 30) >> 14);
644 output_pixel(&dest[5], av_clip_uintp2(B_R + Y2, 30) >> 14);
645 dest += 6;
646 }
647}
648
649static av_always_inline void
650yuv2rgb48_2_c_template(SwsContext *c, const int32_t *buf[2],
651 const int32_t *ubuf[2], const int32_t *vbuf[2],
652 const int32_t *abuf[2], uint16_t *dest, int dstW,
653 int yalpha, int uvalpha, int y,
716d413c 654 enum AVPixelFormat target)
21449410
RB
655{
656 const int32_t *buf0 = buf[0], *buf1 = buf[1],
657 *ubuf0 = ubuf[0], *ubuf1 = ubuf[1],
658 *vbuf0 = vbuf[0], *vbuf1 = vbuf[1];
48606252
RB
659 int yalpha1 = 4096 - yalpha;
660 int uvalpha1 = 4096 - uvalpha;
21449410
RB
661 int i;
662
9206ac80 663 for (i = 0; i < ((dstW + 1) >> 1); i++) {
21449410
RB
664 int Y1 = (buf0[i * 2] * yalpha1 + buf1[i * 2] * yalpha) >> 14;
665 int Y2 = (buf0[i * 2 + 1] * yalpha1 + buf1[i * 2 + 1] * yalpha) >> 14;
666 int U = (ubuf0[i] * uvalpha1 + ubuf1[i] * uvalpha + (-128 << 23)) >> 14;
667 int V = (vbuf0[i] * uvalpha1 + vbuf1[i] * uvalpha + (-128 << 23)) >> 14;
668 int R, G, B;
669
670 Y1 -= c->yuv2rgb_y_offset;
671 Y2 -= c->yuv2rgb_y_offset;
672 Y1 *= c->yuv2rgb_y_coeff;
673 Y2 *= c->yuv2rgb_y_coeff;
674 Y1 += 1 << 13;
675 Y2 += 1 << 13;
676
677 R = V * c->yuv2rgb_v2r_coeff;
678 G = V * c->yuv2rgb_v2g_coeff + U * c->yuv2rgb_u2g_coeff;
679 B = U * c->yuv2rgb_u2b_coeff;
680
681 output_pixel(&dest[0], av_clip_uintp2(R_B + Y1, 30) >> 14);
682 output_pixel(&dest[1], av_clip_uintp2( G + Y1, 30) >> 14);
683 output_pixel(&dest[2], av_clip_uintp2(B_R + Y1, 30) >> 14);
684 output_pixel(&dest[3], av_clip_uintp2(R_B + Y2, 30) >> 14);
685 output_pixel(&dest[4], av_clip_uintp2( G + Y2, 30) >> 14);
686 output_pixel(&dest[5], av_clip_uintp2(B_R + Y2, 30) >> 14);
687 dest += 6;
688 }
689}
690
691static av_always_inline void
692yuv2rgb48_1_c_template(SwsContext *c, const int32_t *buf0,
693 const int32_t *ubuf[2], const int32_t *vbuf[2],
694 const int32_t *abuf0, uint16_t *dest, int dstW,
716d413c 695 int uvalpha, int y, enum AVPixelFormat target)
21449410 696{
1ca7dc60 697 const int32_t *ubuf0 = ubuf[0], *vbuf0 = vbuf[0];
21449410
RB
698 int i;
699
700 if (uvalpha < 2048) {
9206ac80 701 for (i = 0; i < ((dstW + 1) >> 1); i++) {
21449410
RB
702 int Y1 = (buf0[i * 2] ) >> 2;
703 int Y2 = (buf0[i * 2 + 1]) >> 2;
704 int U = (ubuf0[i] + (-128 << 11)) >> 2;
705 int V = (vbuf0[i] + (-128 << 11)) >> 2;
706 int R, G, B;
707
708 Y1 -= c->yuv2rgb_y_offset;
709 Y2 -= c->yuv2rgb_y_offset;
710 Y1 *= c->yuv2rgb_y_coeff;
711 Y2 *= c->yuv2rgb_y_coeff;
712 Y1 += 1 << 13;
713 Y2 += 1 << 13;
714
715 R = V * c->yuv2rgb_v2r_coeff;
716 G = V * c->yuv2rgb_v2g_coeff + U * c->yuv2rgb_u2g_coeff;
717 B = U * c->yuv2rgb_u2b_coeff;
718
719 output_pixel(&dest[0], av_clip_uintp2(R_B + Y1, 30) >> 14);
720 output_pixel(&dest[1], av_clip_uintp2( G + Y1, 30) >> 14);
721 output_pixel(&dest[2], av_clip_uintp2(B_R + Y1, 30) >> 14);
722 output_pixel(&dest[3], av_clip_uintp2(R_B + Y2, 30) >> 14);
723 output_pixel(&dest[4], av_clip_uintp2( G + Y2, 30) >> 14);
724 output_pixel(&dest[5], av_clip_uintp2(B_R + Y2, 30) >> 14);
725 dest += 6;
726 }
727 } else {
1ca7dc60 728 const int32_t *ubuf1 = ubuf[1], *vbuf1 = vbuf[1];
9206ac80 729 for (i = 0; i < ((dstW + 1) >> 1); i++) {
21449410
RB
730 int Y1 = (buf0[i * 2] ) >> 2;
731 int Y2 = (buf0[i * 2 + 1]) >> 2;
d63b7d8c
RB
732 int U = (ubuf0[i] + ubuf1[i] + (-128 << 12)) >> 3;
733 int V = (vbuf0[i] + vbuf1[i] + (-128 << 12)) >> 3;
21449410
RB
734 int R, G, B;
735
736 Y1 -= c->yuv2rgb_y_offset;
737 Y2 -= c->yuv2rgb_y_offset;
738 Y1 *= c->yuv2rgb_y_coeff;
739 Y2 *= c->yuv2rgb_y_coeff;
740 Y1 += 1 << 13;
741 Y2 += 1 << 13;
742
743 R = V * c->yuv2rgb_v2r_coeff;
744 G = V * c->yuv2rgb_v2g_coeff + U * c->yuv2rgb_u2g_coeff;
745 B = U * c->yuv2rgb_u2b_coeff;
746
747 output_pixel(&dest[0], av_clip_uintp2(R_B + Y1, 30) >> 14);
748 output_pixel(&dest[1], av_clip_uintp2( G + Y1, 30) >> 14);
749 output_pixel(&dest[2], av_clip_uintp2(B_R + Y1, 30) >> 14);
750 output_pixel(&dest[3], av_clip_uintp2(R_B + Y2, 30) >> 14);
751 output_pixel(&dest[4], av_clip_uintp2( G + Y2, 30) >> 14);
752 output_pixel(&dest[5], av_clip_uintp2(B_R + Y2, 30) >> 14);
753 dest += 6;
754 }
755 }
756}
757
758#undef output_pixel
759#undef r_b
760#undef b_r
761
3cac4756
RB
762#define YUV2PACKED16WRAPPER(name, base, ext, fmt) \
763static void name ## ext ## _X_c(SwsContext *c, const int16_t *lumFilter, \
764 const int16_t **_lumSrc, int lumFilterSize, \
765 const int16_t *chrFilter, const int16_t **_chrUSrc, \
766 const int16_t **_chrVSrc, int chrFilterSize, \
767 const int16_t **_alpSrc, uint8_t *_dest, int dstW, \
768 int y) \
769{ \
770 const int32_t **lumSrc = (const int32_t **) _lumSrc, \
771 **chrUSrc = (const int32_t **) _chrUSrc, \
772 **chrVSrc = (const int32_t **) _chrVSrc, \
773 **alpSrc = (const int32_t **) _alpSrc; \
774 uint16_t *dest = (uint16_t *) _dest; \
775 name ## base ## _X_c_template(c, lumFilter, lumSrc, lumFilterSize, \
776 chrFilter, chrUSrc, chrVSrc, chrFilterSize, \
777 alpSrc, dest, dstW, y, fmt); \
778} \
779 \
780static void name ## ext ## _2_c(SwsContext *c, const int16_t *_buf[2], \
781 const int16_t *_ubuf[2], const int16_t *_vbuf[2], \
782 const int16_t *_abuf[2], uint8_t *_dest, int dstW, \
783 int yalpha, int uvalpha, int y) \
784{ \
785 const int32_t **buf = (const int32_t **) _buf, \
786 **ubuf = (const int32_t **) _ubuf, \
787 **vbuf = (const int32_t **) _vbuf, \
788 **abuf = (const int32_t **) _abuf; \
789 uint16_t *dest = (uint16_t *) _dest; \
790 name ## base ## _2_c_template(c, buf, ubuf, vbuf, abuf, \
791 dest, dstW, yalpha, uvalpha, y, fmt); \
792} \
793 \
794static void name ## ext ## _1_c(SwsContext *c, const int16_t *_buf0, \
795 const int16_t *_ubuf[2], const int16_t *_vbuf[2], \
796 const int16_t *_abuf0, uint8_t *_dest, int dstW, \
797 int uvalpha, int y) \
798{ \
799 const int32_t *buf0 = (const int32_t *) _buf0, \
800 **ubuf = (const int32_t **) _ubuf, \
801 **vbuf = (const int32_t **) _vbuf, \
802 *abuf0 = (const int32_t *) _abuf0; \
803 uint16_t *dest = (uint16_t *) _dest; \
804 name ## base ## _1_c_template(c, buf0, ubuf, vbuf, abuf0, dest, \
805 dstW, uvalpha, y, fmt); \
806}
807
716d413c
AK
808YUV2PACKED16WRAPPER(yuv2, rgb48, rgb48be, AV_PIX_FMT_RGB48BE)
809YUV2PACKED16WRAPPER(yuv2, rgb48, rgb48le, AV_PIX_FMT_RGB48LE)
810YUV2PACKED16WRAPPER(yuv2, rgb48, bgr48be, AV_PIX_FMT_BGR48BE)
811YUV2PACKED16WRAPPER(yuv2, rgb48, bgr48le, AV_PIX_FMT_BGR48LE)
21449410
RB
812
813/*
814 * Write out 2 RGB pixels in the target pixel format. This function takes a
815 * R/G/B LUT as generated by ff_yuv2rgb_c_init_tables(), which takes care of
816 * things like endianness conversion and shifting. The caller takes care of
817 * setting the correct offset in these tables from the chroma (U/V) values.
818 * This function then uses the luminance (Y1/Y2) values to write out the
819 * correct RGB values into the destination buffer.
820 */
821static av_always_inline void
822yuv2rgb_write(uint8_t *_dest, int i, unsigned Y1, unsigned Y2,
823 unsigned A1, unsigned A2,
824 const void *_r, const void *_g, const void *_b, int y,
716d413c 825 enum AVPixelFormat target, int hasAlpha)
21449410 826{
716d413c
AK
827 if (target == AV_PIX_FMT_ARGB || target == AV_PIX_FMT_RGBA ||
828 target == AV_PIX_FMT_ABGR || target == AV_PIX_FMT_BGRA) {
21449410
RB
829 uint32_t *dest = (uint32_t *) _dest;
830 const uint32_t *r = (const uint32_t *) _r;
831 const uint32_t *g = (const uint32_t *) _g;
832 const uint32_t *b = (const uint32_t *) _b;
833
834#if CONFIG_SMALL
716d413c 835 int sh = hasAlpha ? ((target == AV_PIX_FMT_RGB32_1 || target == AV_PIX_FMT_BGR32_1) ? 0 : 24) : 0;
21449410
RB
836
837 dest[i * 2 + 0] = r[Y1] + g[Y1] + b[Y1] + (hasAlpha ? A1 << sh : 0);
838 dest[i * 2 + 1] = r[Y2] + g[Y2] + b[Y2] + (hasAlpha ? A2 << sh : 0);
839#else
840 if (hasAlpha) {
716d413c 841 int sh = (target == AV_PIX_FMT_RGB32_1 || target == AV_PIX_FMT_BGR32_1) ? 0 : 24;
21449410
RB
842
843 dest[i * 2 + 0] = r[Y1] + g[Y1] + b[Y1] + (A1 << sh);
844 dest[i * 2 + 1] = r[Y2] + g[Y2] + b[Y2] + (A2 << sh);
845 } else {
846 dest[i * 2 + 0] = r[Y1] + g[Y1] + b[Y1];
847 dest[i * 2 + 1] = r[Y2] + g[Y2] + b[Y2];
848 }
849#endif
716d413c 850 } else if (target == AV_PIX_FMT_RGB24 || target == AV_PIX_FMT_BGR24) {
21449410
RB
851 uint8_t *dest = (uint8_t *) _dest;
852 const uint8_t *r = (const uint8_t *) _r;
853 const uint8_t *g = (const uint8_t *) _g;
854 const uint8_t *b = (const uint8_t *) _b;
855
716d413c
AK
856#define r_b ((target == AV_PIX_FMT_RGB24) ? r : b)
857#define b_r ((target == AV_PIX_FMT_RGB24) ? b : r)
21449410
RB
858 dest[i * 6 + 0] = r_b[Y1];
859 dest[i * 6 + 1] = g[Y1];
860 dest[i * 6 + 2] = b_r[Y1];
861 dest[i * 6 + 3] = r_b[Y2];
862 dest[i * 6 + 4] = g[Y2];
863 dest[i * 6 + 5] = b_r[Y2];
864#undef r_b
865#undef b_r
716d413c
AK
866 } else if (target == AV_PIX_FMT_RGB565 || target == AV_PIX_FMT_BGR565 ||
867 target == AV_PIX_FMT_RGB555 || target == AV_PIX_FMT_BGR555 ||
868 target == AV_PIX_FMT_RGB444 || target == AV_PIX_FMT_BGR444) {
21449410
RB
869 uint16_t *dest = (uint16_t *) _dest;
870 const uint16_t *r = (const uint16_t *) _r;
871 const uint16_t *g = (const uint16_t *) _g;
872 const uint16_t *b = (const uint16_t *) _b;
873 int dr1, dg1, db1, dr2, dg2, db2;
874
716d413c 875 if (target == AV_PIX_FMT_RGB565 || target == AV_PIX_FMT_BGR565) {
21449410
RB
876 dr1 = dither_2x2_8[ y & 1 ][0];
877 dg1 = dither_2x2_4[ y & 1 ][0];
878 db1 = dither_2x2_8[(y & 1) ^ 1][0];
879 dr2 = dither_2x2_8[ y & 1 ][1];
880 dg2 = dither_2x2_4[ y & 1 ][1];
881 db2 = dither_2x2_8[(y & 1) ^ 1][1];
716d413c 882 } else if (target == AV_PIX_FMT_RGB555 || target == AV_PIX_FMT_BGR555) {
21449410
RB
883 dr1 = dither_2x2_8[ y & 1 ][0];
884 dg1 = dither_2x2_8[ y & 1 ][1];
885 db1 = dither_2x2_8[(y & 1) ^ 1][0];
886 dr2 = dither_2x2_8[ y & 1 ][1];
887 dg2 = dither_2x2_8[ y & 1 ][0];
888 db2 = dither_2x2_8[(y & 1) ^ 1][1];
889 } else {
aa2ba8c9
DB
890 dr1 = ff_dither_4x4_16[ y & 3 ][0];
891 dg1 = ff_dither_4x4_16[ y & 3 ][1];
892 db1 = ff_dither_4x4_16[(y & 3) ^ 3][0];
893 dr2 = ff_dither_4x4_16[ y & 3 ][1];
894 dg2 = ff_dither_4x4_16[ y & 3 ][0];
895 db2 = ff_dither_4x4_16[(y & 3) ^ 3][1];
21449410
RB
896 }
897
898 dest[i * 2 + 0] = r[Y1 + dr1] + g[Y1 + dg1] + b[Y1 + db1];
899 dest[i * 2 + 1] = r[Y2 + dr2] + g[Y2 + dg2] + b[Y2 + db2];
41ed7ab4 900 } else /* 8/4 bits */ {
21449410
RB
901 uint8_t *dest = (uint8_t *) _dest;
902 const uint8_t *r = (const uint8_t *) _r;
903 const uint8_t *g = (const uint8_t *) _g;
904 const uint8_t *b = (const uint8_t *) _b;
905 int dr1, dg1, db1, dr2, dg2, db2;
906
716d413c 907 if (target == AV_PIX_FMT_RGB8 || target == AV_PIX_FMT_BGR8) {
aa2ba8c9
DB
908 const uint8_t * const d64 = ff_dither_8x8_73[y & 7];
909 const uint8_t * const d32 = ff_dither_8x8_32[y & 7];
21449410
RB
910 dr1 = dg1 = d32[(i * 2 + 0) & 7];
911 db1 = d64[(i * 2 + 0) & 7];
912 dr2 = dg2 = d32[(i * 2 + 1) & 7];
913 db2 = d64[(i * 2 + 1) & 7];
914 } else {
aa2ba8c9
DB
915 const uint8_t * const d64 = ff_dither_8x8_73 [y & 7];
916 const uint8_t * const d128 = ff_dither_8x8_220[y & 7];
21449410
RB
917 dr1 = db1 = d128[(i * 2 + 0) & 7];
918 dg1 = d64[(i * 2 + 0) & 7];
919 dr2 = db2 = d128[(i * 2 + 1) & 7];
920 dg2 = d64[(i * 2 + 1) & 7];
921 }
922
716d413c 923 if (target == AV_PIX_FMT_RGB4 || target == AV_PIX_FMT_BGR4) {
21449410
RB
924 dest[i] = r[Y1 + dr1] + g[Y1 + dg1] + b[Y1 + db1] +
925 ((r[Y2 + dr2] + g[Y2 + dg2] + b[Y2 + db2]) << 4);
926 } else {
927 dest[i * 2 + 0] = r[Y1 + dr1] + g[Y1 + dg1] + b[Y1 + db1];
928 dest[i * 2 + 1] = r[Y2 + dr2] + g[Y2 + dg2] + b[Y2 + db2];
929 }
930 }
931}
932
933static av_always_inline void
934yuv2rgb_X_c_template(SwsContext *c, const int16_t *lumFilter,
935 const int16_t **lumSrc, int lumFilterSize,
936 const int16_t *chrFilter, const int16_t **chrUSrc,
937 const int16_t **chrVSrc, int chrFilterSize,
938 const int16_t **alpSrc, uint8_t *dest, int dstW,
716d413c 939 int y, enum AVPixelFormat target, int hasAlpha)
21449410
RB
940{
941 int i;
942
9206ac80 943 for (i = 0; i < ((dstW + 1) >> 1); i++) {
7331b6e7 944 int j, A1, A2;
21449410
RB
945 int Y1 = 1 << 18;
946 int Y2 = 1 << 18;
947 int U = 1 << 18;
948 int V = 1 << 18;
21449410
RB
949 const void *r, *g, *b;
950
951 for (j = 0; j < lumFilterSize; j++) {
952 Y1 += lumSrc[j][i * 2] * lumFilter[j];
953 Y2 += lumSrc[j][i * 2 + 1] * lumFilter[j];
954 }
955 for (j = 0; j < chrFilterSize; j++) {
956 U += chrUSrc[j][i] * chrFilter[j];
957 V += chrVSrc[j][i] * chrFilter[j];
958 }
959 Y1 >>= 19;
960 Y2 >>= 19;
961 U >>= 19;
962 V >>= 19;
963 if ((Y1 | Y2 | U | V) & 0x100) {
964 Y1 = av_clip_uint8(Y1);
965 Y2 = av_clip_uint8(Y2);
966 U = av_clip_uint8(U);
967 V = av_clip_uint8(V);
968 }
969 if (hasAlpha) {
970 A1 = 1 << 18;
971 A2 = 1 << 18;
972 for (j = 0; j < lumFilterSize; j++) {
973 A1 += alpSrc[j][i * 2 ] * lumFilter[j];
974 A2 += alpSrc[j][i * 2 + 1] * lumFilter[j];
975 }
976 A1 >>= 19;
977 A2 >>= 19;
978 if ((A1 | A2) & 0x100) {
979 A1 = av_clip_uint8(A1);
980 A2 = av_clip_uint8(A2);
981 }
982 }
983
984 /* FIXME fix tables so that clipping is not needed and then use _NOCLIP*/
985 r = c->table_rV[V];
986 g = (c->table_gU[U] + c->table_gV[V]);
987 b = c->table_bU[U];
988
989 yuv2rgb_write(dest, i, Y1, Y2, hasAlpha ? A1 : 0, hasAlpha ? A2 : 0,
990 r, g, b, y, target, hasAlpha);
991 }
992}
993
994static av_always_inline void
995yuv2rgb_2_c_template(SwsContext *c, const int16_t *buf[2],
996 const int16_t *ubuf[2], const int16_t *vbuf[2],
997 const int16_t *abuf[2], uint8_t *dest, int dstW,
998 int yalpha, int uvalpha, int y,
716d413c 999 enum AVPixelFormat target, int hasAlpha)
21449410
RB
1000{
1001 const int16_t *buf0 = buf[0], *buf1 = buf[1],
1002 *ubuf0 = ubuf[0], *ubuf1 = ubuf[1],
1003 *vbuf0 = vbuf[0], *vbuf1 = vbuf[1],
1004 *abuf0 = hasAlpha ? abuf[0] : NULL,
1005 *abuf1 = hasAlpha ? abuf[1] : NULL;
48606252
RB
1006 int yalpha1 = 4096 - yalpha;
1007 int uvalpha1 = 4096 - uvalpha;
21449410
RB
1008 int i;
1009
9206ac80 1010 for (i = 0; i < ((dstW + 1) >> 1); i++) {
21449410
RB
1011 int Y1 = (buf0[i * 2] * yalpha1 + buf1[i * 2] * yalpha) >> 19;
1012 int Y2 = (buf0[i * 2 + 1] * yalpha1 + buf1[i * 2 + 1] * yalpha) >> 19;
1013 int U = (ubuf0[i] * uvalpha1 + ubuf1[i] * uvalpha) >> 19;
1014 int V = (vbuf0[i] * uvalpha1 + vbuf1[i] * uvalpha) >> 19;
1015 int A1, A2;
7f77e904 1016 const void *r, *g, *b;
21449410 1017
9487fb4d
RB
1018 Y1 = av_clip_uint8(Y1);
1019 Y2 = av_clip_uint8(Y2);
1020 U = av_clip_uint8(U);
1021 V = av_clip_uint8(V);
1022
7f77e904
RB
1023 r = c->table_rV[V];
1024 g = (c->table_gU[U] + c->table_gV[V]);
1025 b = c->table_bU[U];
1026
21449410
RB
1027 if (hasAlpha) {
1028 A1 = (abuf0[i * 2 ] * yalpha1 + abuf1[i * 2 ] * yalpha) >> 19;
1029 A2 = (abuf0[i * 2 + 1] * yalpha1 + abuf1[i * 2 + 1] * yalpha) >> 19;
9487fb4d
RB
1030 A1 = av_clip_uint8(A1);
1031 A2 = av_clip_uint8(A2);
21449410
RB
1032 }
1033
1034 yuv2rgb_write(dest, i, Y1, Y2, hasAlpha ? A1 : 0, hasAlpha ? A2 : 0,
1035 r, g, b, y, target, hasAlpha);
1036 }
1037}
1038
1039static av_always_inline void
1040yuv2rgb_1_c_template(SwsContext *c, const int16_t *buf0,
1041 const int16_t *ubuf[2], const int16_t *vbuf[2],
1042 const int16_t *abuf0, uint8_t *dest, int dstW,
716d413c 1043 int uvalpha, int y, enum AVPixelFormat target,
21449410
RB
1044 int hasAlpha)
1045{
1ca7dc60 1046 const int16_t *ubuf0 = ubuf[0], *vbuf0 = vbuf[0];
21449410
RB
1047 int i;
1048
1049 if (uvalpha < 2048) {
9206ac80 1050 for (i = 0; i < ((dstW + 1) >> 1); i++) {
21449410
RB
1051 int Y1 = buf0[i * 2] >> 7;
1052 int Y2 = buf0[i * 2 + 1] >> 7;
1ca7dc60
RB
1053 int U = ubuf0[i] >> 7;
1054 int V = vbuf0[i] >> 7;
21449410 1055 int A1, A2;
7f77e904 1056 const void *r, *g, *b;
21449410 1057
9487fb4d
RB
1058 Y1 = av_clip_uint8(Y1);
1059 Y2 = av_clip_uint8(Y2);
1060 U = av_clip_uint8(U);
1061 V = av_clip_uint8(V);
1062
7f77e904
RB
1063 r = c->table_rV[V];
1064 g = (c->table_gU[U] + c->table_gV[V]);
1065 b = c->table_bU[U];
1066
21449410
RB
1067 if (hasAlpha) {
1068 A1 = abuf0[i * 2 ] >> 7;
1069 A2 = abuf0[i * 2 + 1] >> 7;
9487fb4d
RB
1070 A1 = av_clip_uint8(A1);
1071 A2 = av_clip_uint8(A2);
21449410
RB
1072 }
1073
1074 yuv2rgb_write(dest, i, Y1, Y2, hasAlpha ? A1 : 0, hasAlpha ? A2 : 0,
1075 r, g, b, y, target, hasAlpha);
1076 }
1077 } else {
1ca7dc60 1078 const int16_t *ubuf1 = ubuf[1], *vbuf1 = vbuf[1];
9206ac80 1079 for (i = 0; i < ((dstW + 1) >> 1); i++) {
21449410
RB
1080 int Y1 = buf0[i * 2] >> 7;
1081 int Y2 = buf0[i * 2 + 1] >> 7;
1082 int U = (ubuf0[i] + ubuf1[i]) >> 8;
1083 int V = (vbuf0[i] + vbuf1[i]) >> 8;
1084 int A1, A2;
7f77e904 1085 const void *r, *g, *b;
21449410 1086
9487fb4d
RB
1087 Y1 = av_clip_uint8(Y1);
1088 Y2 = av_clip_uint8(Y2);
1089 U = av_clip_uint8(U);
1090 V = av_clip_uint8(V);
1091
7f77e904
RB
1092 r = c->table_rV[V];
1093 g = (c->table_gU[U] + c->table_gV[V]);
1094 b = c->table_bU[U];
1095
21449410
RB
1096 if (hasAlpha) {
1097 A1 = abuf0[i * 2 ] >> 7;
1098 A2 = abuf0[i * 2 + 1] >> 7;
9487fb4d
RB
1099 A1 = av_clip_uint8(A1);
1100 A2 = av_clip_uint8(A2);
21449410
RB
1101 }
1102
1103 yuv2rgb_write(dest, i, Y1, Y2, hasAlpha ? A1 : 0, hasAlpha ? A2 : 0,
1104 r, g, b, y, target, hasAlpha);
1105 }
1106 }
1107}
1108
1109#define YUV2RGBWRAPPERX(name, base, ext, fmt, hasAlpha) \
1110static void name ## ext ## _X_c(SwsContext *c, const int16_t *lumFilter, \
1111 const int16_t **lumSrc, int lumFilterSize, \
1112 const int16_t *chrFilter, const int16_t **chrUSrc, \
1113 const int16_t **chrVSrc, int chrFilterSize, \
1114 const int16_t **alpSrc, uint8_t *dest, int dstW, \
1115 int y) \
1116{ \
1117 name ## base ## _X_c_template(c, lumFilter, lumSrc, lumFilterSize, \
1118 chrFilter, chrUSrc, chrVSrc, chrFilterSize, \
1119 alpSrc, dest, dstW, y, fmt, hasAlpha); \
1120}
1121#define YUV2RGBWRAPPER(name, base, ext, fmt, hasAlpha) \
1122YUV2RGBWRAPPERX(name, base, ext, fmt, hasAlpha) \
1123static void name ## ext ## _2_c(SwsContext *c, const int16_t *buf[2], \
1124 const int16_t *ubuf[2], const int16_t *vbuf[2], \
1125 const int16_t *abuf[2], uint8_t *dest, int dstW, \
1126 int yalpha, int uvalpha, int y) \
1127{ \
1128 name ## base ## _2_c_template(c, buf, ubuf, vbuf, abuf, \
1129 dest, dstW, yalpha, uvalpha, y, fmt, hasAlpha); \
1130} \
1131 \
1132static void name ## ext ## _1_c(SwsContext *c, const int16_t *buf0, \
1133 const int16_t *ubuf[2], const int16_t *vbuf[2], \
1134 const int16_t *abuf0, uint8_t *dest, int dstW, \
1135 int uvalpha, int y) \
1136{ \
1137 name ## base ## _1_c_template(c, buf0, ubuf, vbuf, abuf0, dest, \
1138 dstW, uvalpha, y, fmt, hasAlpha); \
1139}
1140
1141#if CONFIG_SMALL
716d413c
AK
1142YUV2RGBWRAPPER(yuv2rgb,, 32_1, AV_PIX_FMT_RGB32_1, CONFIG_SWSCALE_ALPHA && c->alpPixBuf)
1143YUV2RGBWRAPPER(yuv2rgb,, 32, AV_PIX_FMT_RGB32, CONFIG_SWSCALE_ALPHA && c->alpPixBuf)
21449410
RB
1144#else
1145#if CONFIG_SWSCALE_ALPHA
716d413c
AK
1146YUV2RGBWRAPPER(yuv2rgb,, a32_1, AV_PIX_FMT_RGB32_1, 1)
1147YUV2RGBWRAPPER(yuv2rgb,, a32, AV_PIX_FMT_RGB32, 1)
21449410 1148#endif
716d413c
AK
1149YUV2RGBWRAPPER(yuv2rgb,, x32_1, AV_PIX_FMT_RGB32_1, 0)
1150YUV2RGBWRAPPER(yuv2rgb,, x32, AV_PIX_FMT_RGB32, 0)
21449410 1151#endif
716d413c
AK
1152YUV2RGBWRAPPER(yuv2, rgb, rgb24, AV_PIX_FMT_RGB24, 0)
1153YUV2RGBWRAPPER(yuv2, rgb, bgr24, AV_PIX_FMT_BGR24, 0)
1154YUV2RGBWRAPPER(yuv2rgb,, 16, AV_PIX_FMT_RGB565, 0)
1155YUV2RGBWRAPPER(yuv2rgb,, 15, AV_PIX_FMT_RGB555, 0)
1156YUV2RGBWRAPPER(yuv2rgb,, 12, AV_PIX_FMT_RGB444, 0)
1157YUV2RGBWRAPPER(yuv2rgb,, 8, AV_PIX_FMT_RGB8, 0)
1158YUV2RGBWRAPPER(yuv2rgb,, 4, AV_PIX_FMT_RGB4, 0)
1159YUV2RGBWRAPPER(yuv2rgb,, 4b, AV_PIX_FMT_RGB4_BYTE, 0)
21449410
RB
1160
1161static av_always_inline void
1162yuv2rgb_full_X_c_template(SwsContext *c, const int16_t *lumFilter,
1163 const int16_t **lumSrc, int lumFilterSize,
1164 const int16_t *chrFilter, const int16_t **chrUSrc,
1165 const int16_t **chrVSrc, int chrFilterSize,
1166 const int16_t **alpSrc, uint8_t *dest,
716d413c 1167 int dstW, int y, enum AVPixelFormat target, int hasAlpha)
21449410
RB
1168{
1169 int i;
716d413c 1170 int step = (target == AV_PIX_FMT_RGB24 || target == AV_PIX_FMT_BGR24) ? 3 : 4;
21449410
RB
1171
1172 for (i = 0; i < dstW; i++) {
1173 int j;
1174 int Y = 0;
1175 int U = -128 << 19;
1176 int V = -128 << 19;
7331b6e7 1177 int R, G, B, A;
21449410
RB
1178
1179 for (j = 0; j < lumFilterSize; j++) {
1180 Y += lumSrc[j][i] * lumFilter[j];
1181 }
1182 for (j = 0; j < chrFilterSize; j++) {
1183 U += chrUSrc[j][i] * chrFilter[j];
1184 V += chrVSrc[j][i] * chrFilter[j];
1185 }
1186 Y >>= 10;
1187 U >>= 10;
1188 V >>= 10;
1189 if (hasAlpha) {
1190 A = 1 << 21;
1191 for (j = 0; j < lumFilterSize; j++) {
1192 A += alpSrc[j][i] * lumFilter[j];
1193 }
1194 A >>= 19;
1195 if (A & 0x100)
1196 A = av_clip_uint8(A);
1197 }
1198 Y -= c->yuv2rgb_y_offset;
1199 Y *= c->yuv2rgb_y_coeff;
1200 Y += 1 << 21;
1201 R = Y + V*c->yuv2rgb_v2r_coeff;
1202 G = Y + V*c->yuv2rgb_v2g_coeff + U*c->yuv2rgb_u2g_coeff;
1203 B = Y + U*c->yuv2rgb_u2b_coeff;
1204 if ((R | G | B) & 0xC0000000) {
1205 R = av_clip_uintp2(R, 30);
1206 G = av_clip_uintp2(G, 30);
1207 B = av_clip_uintp2(B, 30);
1208 }
1209
1210 switch(target) {
716d413c 1211 case AV_PIX_FMT_ARGB:
21449410
RB
1212 dest[0] = hasAlpha ? A : 255;
1213 dest[1] = R >> 22;
1214 dest[2] = G >> 22;
1215 dest[3] = B >> 22;
1216 break;
716d413c 1217 case AV_PIX_FMT_RGB24:
21449410
RB
1218 dest[0] = R >> 22;
1219 dest[1] = G >> 22;
1220 dest[2] = B >> 22;
1221 break;
716d413c 1222 case AV_PIX_FMT_RGBA:
21449410
RB
1223 dest[0] = R >> 22;
1224 dest[1] = G >> 22;
1225 dest[2] = B >> 22;
1226 dest[3] = hasAlpha ? A : 255;
1227 break;
716d413c 1228 case AV_PIX_FMT_ABGR:
21449410
RB
1229 dest[0] = hasAlpha ? A : 255;
1230 dest[1] = B >> 22;
1231 dest[2] = G >> 22;
1232 dest[3] = R >> 22;
1233 dest += 4;
1234 break;
716d413c 1235 case AV_PIX_FMT_BGR24:
21449410
RB
1236 dest[0] = B >> 22;
1237 dest[1] = G >> 22;
1238 dest[2] = R >> 22;
1239 break;
716d413c 1240 case AV_PIX_FMT_BGRA:
21449410
RB
1241 dest[0] = B >> 22;
1242 dest[1] = G >> 22;
1243 dest[2] = R >> 22;
1244 dest[3] = hasAlpha ? A : 255;
1245 break;
1246 }
1247 dest += step;
1248 }
1249}
1250
1251#if CONFIG_SMALL
716d413c
AK
1252YUV2RGBWRAPPERX(yuv2, rgb_full, bgra32_full, AV_PIX_FMT_BGRA, CONFIG_SWSCALE_ALPHA && c->alpPixBuf)
1253YUV2RGBWRAPPERX(yuv2, rgb_full, abgr32_full, AV_PIX_FMT_ABGR, CONFIG_SWSCALE_ALPHA && c->alpPixBuf)
1254YUV2RGBWRAPPERX(yuv2, rgb_full, rgba32_full, AV_PIX_FMT_RGBA, CONFIG_SWSCALE_ALPHA && c->alpPixBuf)
1255YUV2RGBWRAPPERX(yuv2, rgb_full, argb32_full, AV_PIX_FMT_ARGB, CONFIG_SWSCALE_ALPHA && c->alpPixBuf)
21449410
RB
1256#else
1257#if CONFIG_SWSCALE_ALPHA
716d413c
AK
1258YUV2RGBWRAPPERX(yuv2, rgb_full, bgra32_full, AV_PIX_FMT_BGRA, 1)
1259YUV2RGBWRAPPERX(yuv2, rgb_full, abgr32_full, AV_PIX_FMT_ABGR, 1)
1260YUV2RGBWRAPPERX(yuv2, rgb_full, rgba32_full, AV_PIX_FMT_RGBA, 1)
1261YUV2RGBWRAPPERX(yuv2, rgb_full, argb32_full, AV_PIX_FMT_ARGB, 1)
21449410 1262#endif
716d413c
AK
1263YUV2RGBWRAPPERX(yuv2, rgb_full, bgrx32_full, AV_PIX_FMT_BGRA, 0)
1264YUV2RGBWRAPPERX(yuv2, rgb_full, xbgr32_full, AV_PIX_FMT_ABGR, 0)
1265YUV2RGBWRAPPERX(yuv2, rgb_full, rgbx32_full, AV_PIX_FMT_RGBA, 0)
1266YUV2RGBWRAPPERX(yuv2, rgb_full, xrgb32_full, AV_PIX_FMT_ARGB, 0)
21449410 1267#endif
716d413c
AK
1268YUV2RGBWRAPPERX(yuv2, rgb_full, bgr24_full, AV_PIX_FMT_BGR24, 0)
1269YUV2RGBWRAPPERX(yuv2, rgb_full, rgb24_full, AV_PIX_FMT_RGB24, 0)
21449410 1270
4eb93bed
MN
1271static void
1272yuv2gbrp_full_X_c(SwsContext *c, const int16_t *lumFilter,
1273 const int16_t **lumSrc, int lumFilterSize,
1274 const int16_t *chrFilter, const int16_t **chrUSrc,
1275 const int16_t **chrVSrc, int chrFilterSize,
1276 const int16_t **alpSrc, uint8_t **dest,
1277 int dstW, int y)
1278{
d6d5ef55 1279 const AVPixFmtDescriptor *desc = av_pix_fmt_desc_get(c->dstFormat);
4eb93bed 1280 int i;
92806167 1281 int hasAlpha = (desc->flags & AV_PIX_FMT_FLAG_ALPHA) && alpSrc;
d6d5ef55 1282 uint16_t **dest16 = (uint16_t**)dest;
2268db2c 1283 int SH = 22 + 8 - desc->comp[0].depth;
4eb93bed
MN
1284
1285 for (i = 0; i < dstW; i++) {
1286 int j;
1287 int Y = 1 << 9;
1288 int U = (1 << 9) - (128 << 19);
1289 int V = (1 << 9) - (128 << 19);
1290 int R, G, B, A;
1291
1292 for (j = 0; j < lumFilterSize; j++)
1293 Y += lumSrc[j][i] * lumFilter[j];
1294
1295 for (j = 0; j < chrFilterSize; j++) {
1296 U += chrUSrc[j][i] * chrFilter[j];
1297 V += chrVSrc[j][i] * chrFilter[j];
1298 }
1299
1300 Y >>= 10;
1301 U >>= 10;
1302 V >>= 10;
1303
1304 if (hasAlpha) {
1305 A = 1 << 18;
1306
1307 for (j = 0; j < lumFilterSize; j++)
1308 A += alpSrc[j][i] * lumFilter[j];
1309
1310 A >>= 19;
1311
1312 if (A & 0x100)
1313 A = av_clip_uint8(A);
1314 }
1315
1316 Y -= c->yuv2rgb_y_offset;
1317 Y *= c->yuv2rgb_y_coeff;
1318 Y += 1 << 21;
1319 R = Y + V * c->yuv2rgb_v2r_coeff;
1320 G = Y + V * c->yuv2rgb_v2g_coeff + U * c->yuv2rgb_u2g_coeff;
1321 B = Y + U * c->yuv2rgb_u2b_coeff;
1322
1323 if ((R | G | B) & 0xC0000000) {
1324 R = av_clip_uintp2(R, 30);
1325 G = av_clip_uintp2(G, 30);
1326 B = av_clip_uintp2(B, 30);
1327 }
1328
d6d5ef55
MN
1329 if (SH != 22) {
1330 dest16[0][i] = G >> SH;
1331 dest16[1][i] = B >> SH;
1332 dest16[2][i] = R >> SH;
92806167
PM
1333 if (hasAlpha)
1334 dest16[3][i] = A;
d6d5ef55
MN
1335 } else {
1336 dest[0][i] = G >> 22;
1337 dest[1][i] = B >> 22;
1338 dest[2][i] = R >> 22;
92806167
PM
1339 if (hasAlpha)
1340 dest[3][i] = A;
d6d5ef55
MN
1341 }
1342 }
1343 if (SH != 22 && (!isBE(c->dstFormat)) != (!HAVE_BIGENDIAN)) {
1344 for (i = 0; i < dstW; i++) {
1345 dest16[0][i] = av_bswap16(dest16[0][i]);
1346 dest16[1][i] = av_bswap16(dest16[1][i]);
1347 dest16[2][i] = av_bswap16(dest16[2][i]);
92806167
PM
1348 if (hasAlpha)
1349 dest16[3][i] = av_bswap16(dest16[3][i]);
d6d5ef55 1350 }
4eb93bed
MN
1351 }
1352}
1353
5a6e3c03
DB
1354av_cold void ff_sws_init_output_funcs(SwsContext *c,
1355 yuv2planar1_fn *yuv2plane1,
1356 yuv2planarX_fn *yuv2planeX,
1357 yuv2interleavedX_fn *yuv2nv12cX,
1358 yuv2packed1_fn *yuv2packed1,
1359 yuv2packed2_fn *yuv2packed2,
4eb93bed
MN
1360 yuv2packedX_fn *yuv2packedX,
1361 yuv2anyX_fn *yuv2anyX)
21449410 1362{
716d413c 1363 enum AVPixelFormat dstFormat = c->dstFormat;
0a7068fa 1364 const AVPixFmtDescriptor *desc = av_pix_fmt_desc_get(dstFormat);
21449410
RB
1365
1366 if (is16BPS(dstFormat)) {
1367 *yuv2planeX = isBE(dstFormat) ? yuv2planeX_16BE_c : yuv2planeX_16LE_c;
1368 *yuv2plane1 = isBE(dstFormat) ? yuv2plane1_16BE_c : yuv2plane1_16LE_c;
2b5b1e1e 1369 } else if (is9_15BPS(dstFormat)) {
2268db2c 1370 if (desc->comp[0].depth == 9) {
21449410
RB
1371 *yuv2planeX = isBE(dstFormat) ? yuv2planeX_9BE_c : yuv2planeX_9LE_c;
1372 *yuv2plane1 = isBE(dstFormat) ? yuv2plane1_9BE_c : yuv2plane1_9LE_c;
85406e7a 1373 } else if (desc->comp[0].depth == 10) {
21449410
RB
1374 *yuv2planeX = isBE(dstFormat) ? yuv2planeX_10BE_c : yuv2planeX_10LE_c;
1375 *yuv2plane1 = isBE(dstFormat) ? yuv2plane1_10BE_c : yuv2plane1_10LE_c;
85406e7a
LB
1376 } else if (desc->comp[0].depth == 12) {
1377 *yuv2planeX = isBE(dstFormat) ? yuv2planeX_12BE_c : yuv2planeX_12LE_c;
1378 *yuv2plane1 = isBE(dstFormat) ? yuv2plane1_12BE_c : yuv2plane1_12LE_c;
1379 } else {
1380 assert(0);
21449410
RB
1381 }
1382 } else {
1383 *yuv2plane1 = yuv2plane1_8_c;
1384 *yuv2planeX = yuv2planeX_8_c;
716d413c 1385 if (dstFormat == AV_PIX_FMT_NV12 || dstFormat == AV_PIX_FMT_NV21)
21449410
RB
1386 *yuv2nv12cX = yuv2nv12cX_c;
1387 }
1388
1389 if(c->flags & SWS_FULL_CHR_H_INT) {
1390 switch (dstFormat) {
716d413c 1391 case AV_PIX_FMT_RGBA:
21449410
RB
1392#if CONFIG_SMALL
1393 *yuv2packedX = yuv2rgba32_full_X_c;
1394#else
1395#if CONFIG_SWSCALE_ALPHA
1396 if (c->alpPixBuf) {
1397 *yuv2packedX = yuv2rgba32_full_X_c;
1398 } else
1399#endif /* CONFIG_SWSCALE_ALPHA */
1400 {
1401 *yuv2packedX = yuv2rgbx32_full_X_c;
1402 }
1403#endif /* !CONFIG_SMALL */
1404 break;
716d413c 1405 case AV_PIX_FMT_ARGB:
21449410
RB
1406#if CONFIG_SMALL
1407 *yuv2packedX = yuv2argb32_full_X_c;
1408#else
1409#if CONFIG_SWSCALE_ALPHA
1410 if (c->alpPixBuf) {
1411 *yuv2packedX = yuv2argb32_full_X_c;
1412 } else
1413#endif /* CONFIG_SWSCALE_ALPHA */
1414 {
1415 *yuv2packedX = yuv2xrgb32_full_X_c;
1416 }
1417#endif /* !CONFIG_SMALL */
1418 break;
716d413c 1419 case AV_PIX_FMT_BGRA:
21449410
RB
1420#if CONFIG_SMALL
1421 *yuv2packedX = yuv2bgra32_full_X_c;
1422#else
1423#if CONFIG_SWSCALE_ALPHA
1424 if (c->alpPixBuf) {
1425 *yuv2packedX = yuv2bgra32_full_X_c;
1426 } else
1427#endif /* CONFIG_SWSCALE_ALPHA */
1428 {
1429 *yuv2packedX = yuv2bgrx32_full_X_c;
1430 }
1431#endif /* !CONFIG_SMALL */
1432 break;
716d413c 1433 case AV_PIX_FMT_ABGR:
21449410
RB
1434#if CONFIG_SMALL
1435 *yuv2packedX = yuv2abgr32_full_X_c;
1436#else
1437#if CONFIG_SWSCALE_ALPHA
1438 if (c->alpPixBuf) {
1439 *yuv2packedX = yuv2abgr32_full_X_c;
1440 } else
1441#endif /* CONFIG_SWSCALE_ALPHA */
1442 {
1443 *yuv2packedX = yuv2xbgr32_full_X_c;
1444 }
1445#endif /* !CONFIG_SMALL */
1446 break;
716d413c 1447 case AV_PIX_FMT_RGB24:
21449410
RB
1448 *yuv2packedX = yuv2rgb24_full_X_c;
1449 break;
716d413c 1450 case AV_PIX_FMT_BGR24:
21449410
RB
1451 *yuv2packedX = yuv2bgr24_full_X_c;
1452 break;
4eb93bed 1453 case AV_PIX_FMT_GBRP:
d6d5ef55
MN
1454 case AV_PIX_FMT_GBRP9BE:
1455 case AV_PIX_FMT_GBRP9LE:
1456 case AV_PIX_FMT_GBRP10BE:
1457 case AV_PIX_FMT_GBRP10LE:
1458 case AV_PIX_FMT_GBRP16BE:
1459 case AV_PIX_FMT_GBRP16LE:
92806167 1460 case AV_PIX_FMT_GBRAP:
4eb93bed
MN
1461 *yuv2anyX = yuv2gbrp_full_X_c;
1462 break;
21449410
RB
1463 }
1464 } else {
1465 switch (dstFormat) {
716d413c 1466 case AV_PIX_FMT_RGB48LE:
21449410
RB
1467 *yuv2packed1 = yuv2rgb48le_1_c;
1468 *yuv2packed2 = yuv2rgb48le_2_c;
1469 *yuv2packedX = yuv2rgb48le_X_c;
1470 break;
716d413c 1471 case AV_PIX_FMT_RGB48BE:
21449410
RB
1472 *yuv2packed1 = yuv2rgb48be_1_c;
1473 *yuv2packed2 = yuv2rgb48be_2_c;
1474 *yuv2packedX = yuv2rgb48be_X_c;
1475 break;
716d413c 1476 case AV_PIX_FMT_BGR48LE:
21449410
RB
1477 *yuv2packed1 = yuv2bgr48le_1_c;
1478 *yuv2packed2 = yuv2bgr48le_2_c;
1479 *yuv2packedX = yuv2bgr48le_X_c;
1480 break;
716d413c 1481 case AV_PIX_FMT_BGR48BE:
21449410
RB
1482 *yuv2packed1 = yuv2bgr48be_1_c;
1483 *yuv2packed2 = yuv2bgr48be_2_c;
1484 *yuv2packedX = yuv2bgr48be_X_c;
1485 break;
716d413c
AK
1486 case AV_PIX_FMT_RGB32:
1487 case AV_PIX_FMT_BGR32:
21449410
RB
1488#if CONFIG_SMALL
1489 *yuv2packed1 = yuv2rgb32_1_c;
1490 *yuv2packed2 = yuv2rgb32_2_c;
1491 *yuv2packedX = yuv2rgb32_X_c;
1492#else
1493#if CONFIG_SWSCALE_ALPHA
1494 if (c->alpPixBuf) {
1495 *yuv2packed1 = yuv2rgba32_1_c;
1496 *yuv2packed2 = yuv2rgba32_2_c;
1497 *yuv2packedX = yuv2rgba32_X_c;
1498 } else
1499#endif /* CONFIG_SWSCALE_ALPHA */
1500 {
1501 *yuv2packed1 = yuv2rgbx32_1_c;
1502 *yuv2packed2 = yuv2rgbx32_2_c;
1503 *yuv2packedX = yuv2rgbx32_X_c;
1504 }
1505#endif /* !CONFIG_SMALL */
1506 break;
716d413c
AK
1507 case AV_PIX_FMT_RGB32_1:
1508 case AV_PIX_FMT_BGR32_1:
21449410
RB
1509#if CONFIG_SMALL
1510 *yuv2packed1 = yuv2rgb32_1_1_c;
1511 *yuv2packed2 = yuv2rgb32_1_2_c;
1512 *yuv2packedX = yuv2rgb32_1_X_c;
1513#else
1514#if CONFIG_SWSCALE_ALPHA
1515 if (c->alpPixBuf) {
1516 *yuv2packed1 = yuv2rgba32_1_1_c;
1517 *yuv2packed2 = yuv2rgba32_1_2_c;
1518 *yuv2packedX = yuv2rgba32_1_X_c;
1519 } else
1520#endif /* CONFIG_SWSCALE_ALPHA */
1521 {
1522 *yuv2packed1 = yuv2rgbx32_1_1_c;
1523 *yuv2packed2 = yuv2rgbx32_1_2_c;
1524 *yuv2packedX = yuv2rgbx32_1_X_c;
1525 }
1526#endif /* !CONFIG_SMALL */
1527 break;
716d413c 1528 case AV_PIX_FMT_RGB24:
21449410
RB
1529 *yuv2packed1 = yuv2rgb24_1_c;
1530 *yuv2packed2 = yuv2rgb24_2_c;
1531 *yuv2packedX = yuv2rgb24_X_c;
1532 break;
716d413c 1533 case AV_PIX_FMT_BGR24:
21449410
RB
1534 *yuv2packed1 = yuv2bgr24_1_c;
1535 *yuv2packed2 = yuv2bgr24_2_c;
1536 *yuv2packedX = yuv2bgr24_X_c;
1537 break;
716d413c
AK
1538 case AV_PIX_FMT_RGB565LE:
1539 case AV_PIX_FMT_RGB565BE:
1540 case AV_PIX_FMT_BGR565LE:
1541 case AV_PIX_FMT_BGR565BE:
21449410
RB
1542 *yuv2packed1 = yuv2rgb16_1_c;
1543 *yuv2packed2 = yuv2rgb16_2_c;
1544 *yuv2packedX = yuv2rgb16_X_c;
1545 break;
716d413c
AK
1546 case AV_PIX_FMT_RGB555LE:
1547 case AV_PIX_FMT_RGB555BE:
1548 case AV_PIX_FMT_BGR555LE:
1549 case AV_PIX_FMT_BGR555BE:
21449410
RB
1550 *yuv2packed1 = yuv2rgb15_1_c;
1551 *yuv2packed2 = yuv2rgb15_2_c;
1552 *yuv2packedX = yuv2rgb15_X_c;
1553 break;
716d413c
AK
1554 case AV_PIX_FMT_RGB444LE:
1555 case AV_PIX_FMT_RGB444BE:
1556 case AV_PIX_FMT_BGR444LE:
1557 case AV_PIX_FMT_BGR444BE:
21449410
RB
1558 *yuv2packed1 = yuv2rgb12_1_c;
1559 *yuv2packed2 = yuv2rgb12_2_c;
1560 *yuv2packedX = yuv2rgb12_X_c;
1561 break;
716d413c
AK
1562 case AV_PIX_FMT_RGB8:
1563 case AV_PIX_FMT_BGR8:
21449410
RB
1564 *yuv2packed1 = yuv2rgb8_1_c;
1565 *yuv2packed2 = yuv2rgb8_2_c;
1566 *yuv2packedX = yuv2rgb8_X_c;
1567 break;
716d413c
AK
1568 case AV_PIX_FMT_RGB4:
1569 case AV_PIX_FMT_BGR4:
21449410
RB
1570 *yuv2packed1 = yuv2rgb4_1_c;
1571 *yuv2packed2 = yuv2rgb4_2_c;
1572 *yuv2packedX = yuv2rgb4_X_c;
1573 break;
716d413c
AK
1574 case AV_PIX_FMT_RGB4_BYTE:
1575 case AV_PIX_FMT_BGR4_BYTE:
21449410
RB
1576 *yuv2packed1 = yuv2rgb4b_1_c;
1577 *yuv2packed2 = yuv2rgb4b_2_c;
1578 *yuv2packedX = yuv2rgb4b_X_c;
1579 break;
1580 }
1581 }
1582 switch (dstFormat) {
716d413c 1583 case AV_PIX_FMT_MONOWHITE:
21449410
RB
1584 *yuv2packed1 = yuv2monowhite_1_c;
1585 *yuv2packed2 = yuv2monowhite_2_c;
1586 *yuv2packedX = yuv2monowhite_X_c;
1587 break;
716d413c 1588 case AV_PIX_FMT_MONOBLACK:
21449410
RB
1589 *yuv2packed1 = yuv2monoblack_1_c;
1590 *yuv2packed2 = yuv2monoblack_2_c;
1591 *yuv2packedX = yuv2monoblack_X_c;
1592 break;
716d413c 1593 case AV_PIX_FMT_YUYV422:
21449410
RB
1594 *yuv2packed1 = yuv2yuyv422_1_c;
1595 *yuv2packed2 = yuv2yuyv422_2_c;
1596 *yuv2packedX = yuv2yuyv422_X_c;
1597 break;
92b099da
VG
1598 case AV_PIX_FMT_YVYU422:
1599 *yuv2packed1 = yuv2yvyu422_1_c;
1600 *yuv2packed2 = yuv2yvyu422_2_c;
1601 *yuv2packedX = yuv2yvyu422_X_c;
1602 break;
716d413c 1603 case AV_PIX_FMT_UYVY422:
21449410
RB
1604 *yuv2packed1 = yuv2uyvy422_1_c;
1605 *yuv2packed2 = yuv2uyvy422_2_c;
1606 *yuv2packedX = yuv2uyvy422_X_c;
1607 break;
1608 }
1609}