2 * Copyright (C) 2001-2003 Michael Niedermayer <michaelni@gmx.at>
4 * This file is part of FFmpeg.
6 * FFmpeg 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 * FFmpeg 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 FFmpeg; if not, write to the Free Software
18 * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
23 #define _SVID_SOURCE // needed for MAP_ANONYMOUS
24 #define _DARWIN_C_SOURCE // needed for MAP_ANON
31 #if defined(MAP_ANON) && !defined(MAP_ANONYMOUS)
32 #define MAP_ANONYMOUS MAP_ANON
36 #define WIN32_LEAN_AND_MEAN
40 #include "libavutil/attributes.h"
41 #include "libavutil/avassert.h"
42 #include "libavutil/avutil.h"
43 #include "libavutil/bswap.h"
44 #include "libavutil/cpu.h"
45 #include "libavutil/imgutils.h"
46 #include "libavutil/intreadwrite.h"
47 #include "libavutil/mathematics.h"
48 #include "libavutil/opt.h"
49 #include "libavutil/pixdesc.h"
50 #include "libavutil/ppc/cpu.h"
51 #include "libavutil/x86/asm.h"
52 #include "libavutil/x86/cpu.h"
55 #include "swscale_internal.h"
57 static void handle_formats(SwsContext *c);
59 unsigned swscale_version(void)
61 av_assert0(LIBSWSCALE_VERSION_MICRO >= 100);
62 return LIBSWSCALE_VERSION_INT;
65 const char *swscale_configuration(void)
67 return FFMPEG_CONFIGURATION;
70 const char *swscale_license(void)
72 #define LICENSE_PREFIX "libswscale license: "
73 return LICENSE_PREFIX FFMPEG_LICENSE + sizeof(LICENSE_PREFIX) - 1;
76 typedef struct FormatEntry {
77 uint8_t is_supported_in :1;
78 uint8_t is_supported_out :1;
79 uint8_t is_supported_endianness :1;
82 static const FormatEntry format_entries[AV_PIX_FMT_NB] = {
83 [AV_PIX_FMT_YUV420P] = { 1, 1 },
84 [AV_PIX_FMT_YUYV422] = { 1, 1 },
85 [AV_PIX_FMT_RGB24] = { 1, 1 },
86 [AV_PIX_FMT_BGR24] = { 1, 1 },
87 [AV_PIX_FMT_YUV422P] = { 1, 1 },
88 [AV_PIX_FMT_YUV444P] = { 1, 1 },
89 [AV_PIX_FMT_YUV410P] = { 1, 1 },
90 [AV_PIX_FMT_YUV411P] = { 1, 1 },
91 [AV_PIX_FMT_GRAY8] = { 1, 1 },
92 [AV_PIX_FMT_MONOWHITE] = { 1, 1 },
93 [AV_PIX_FMT_MONOBLACK] = { 1, 1 },
94 [AV_PIX_FMT_PAL8] = { 1, 0 },
95 [AV_PIX_FMT_YUVJ420P] = { 1, 1 },
96 [AV_PIX_FMT_YUVJ411P] = { 1, 1 },
97 [AV_PIX_FMT_YUVJ422P] = { 1, 1 },
98 [AV_PIX_FMT_YUVJ444P] = { 1, 1 },
99 [AV_PIX_FMT_YVYU422] = { 1, 1 },
100 [AV_PIX_FMT_UYVY422] = { 1, 1 },
101 [AV_PIX_FMT_UYYVYY411] = { 0, 0 },
102 [AV_PIX_FMT_BGR8] = { 1, 1 },
103 [AV_PIX_FMT_BGR4] = { 0, 1 },
104 [AV_PIX_FMT_BGR4_BYTE] = { 1, 1 },
105 [AV_PIX_FMT_RGB8] = { 1, 1 },
106 [AV_PIX_FMT_RGB4] = { 0, 1 },
107 [AV_PIX_FMT_RGB4_BYTE] = { 1, 1 },
108 [AV_PIX_FMT_NV12] = { 1, 1 },
109 [AV_PIX_FMT_NV21] = { 1, 1 },
110 [AV_PIX_FMT_ARGB] = { 1, 1 },
111 [AV_PIX_FMT_RGBA] = { 1, 1 },
112 [AV_PIX_FMT_ABGR] = { 1, 1 },
113 [AV_PIX_FMT_BGRA] = { 1, 1 },
114 [AV_PIX_FMT_0RGB] = { 1, 1 },
115 [AV_PIX_FMT_RGB0] = { 1, 1 },
116 [AV_PIX_FMT_0BGR] = { 1, 1 },
117 [AV_PIX_FMT_BGR0] = { 1, 1 },
118 [AV_PIX_FMT_GRAY16BE] = { 1, 1 },
119 [AV_PIX_FMT_GRAY16LE] = { 1, 1 },
120 [AV_PIX_FMT_YUV440P] = { 1, 1 },
121 [AV_PIX_FMT_YUVJ440P] = { 1, 1 },
122 [AV_PIX_FMT_YUVA420P] = { 1, 1 },
123 [AV_PIX_FMT_YUVA422P] = { 1, 1 },
124 [AV_PIX_FMT_YUVA444P] = { 1, 1 },
125 [AV_PIX_FMT_YUVA420P9BE] = { 1, 1 },
126 [AV_PIX_FMT_YUVA420P9LE] = { 1, 1 },
127 [AV_PIX_FMT_YUVA422P9BE] = { 1, 1 },
128 [AV_PIX_FMT_YUVA422P9LE] = { 1, 1 },
129 [AV_PIX_FMT_YUVA444P9BE] = { 1, 1 },
130 [AV_PIX_FMT_YUVA444P9LE] = { 1, 1 },
131 [AV_PIX_FMT_YUVA420P10BE]= { 1, 1 },
132 [AV_PIX_FMT_YUVA420P10LE]= { 1, 1 },
133 [AV_PIX_FMT_YUVA422P10BE]= { 1, 1 },
134 [AV_PIX_FMT_YUVA422P10LE]= { 1, 1 },
135 [AV_PIX_FMT_YUVA444P10BE]= { 1, 1 },
136 [AV_PIX_FMT_YUVA444P10LE]= { 1, 1 },
137 [AV_PIX_FMT_YUVA420P16BE]= { 1, 1 },
138 [AV_PIX_FMT_YUVA420P16LE]= { 1, 1 },
139 [AV_PIX_FMT_YUVA422P16BE]= { 1, 1 },
140 [AV_PIX_FMT_YUVA422P16LE]= { 1, 1 },
141 [AV_PIX_FMT_YUVA444P16BE]= { 1, 1 },
142 [AV_PIX_FMT_YUVA444P16LE]= { 1, 1 },
143 [AV_PIX_FMT_RGB48BE] = { 1, 1 },
144 [AV_PIX_FMT_RGB48LE] = { 1, 1 },
145 [AV_PIX_FMT_RGBA64BE] = { 1, 1, 1 },
146 [AV_PIX_FMT_RGBA64LE] = { 1, 1, 1 },
147 [AV_PIX_FMT_RGB565BE] = { 1, 1 },
148 [AV_PIX_FMT_RGB565LE] = { 1, 1 },
149 [AV_PIX_FMT_RGB555BE] = { 1, 1 },
150 [AV_PIX_FMT_RGB555LE] = { 1, 1 },
151 [AV_PIX_FMT_BGR565BE] = { 1, 1 },
152 [AV_PIX_FMT_BGR565LE] = { 1, 1 },
153 [AV_PIX_FMT_BGR555BE] = { 1, 1 },
154 [AV_PIX_FMT_BGR555LE] = { 1, 1 },
155 [AV_PIX_FMT_YUV420P16LE] = { 1, 1 },
156 [AV_PIX_FMT_YUV420P16BE] = { 1, 1 },
157 [AV_PIX_FMT_YUV422P16LE] = { 1, 1 },
158 [AV_PIX_FMT_YUV422P16BE] = { 1, 1 },
159 [AV_PIX_FMT_YUV444P16LE] = { 1, 1 },
160 [AV_PIX_FMT_YUV444P16BE] = { 1, 1 },
161 [AV_PIX_FMT_RGB444LE] = { 1, 1 },
162 [AV_PIX_FMT_RGB444BE] = { 1, 1 },
163 [AV_PIX_FMT_BGR444LE] = { 1, 1 },
164 [AV_PIX_FMT_BGR444BE] = { 1, 1 },
165 [AV_PIX_FMT_YA8] = { 1, 0 },
166 [AV_PIX_FMT_YA16BE] = { 1, 0 },
167 [AV_PIX_FMT_YA16LE] = { 1, 0 },
168 [AV_PIX_FMT_BGR48BE] = { 1, 1 },
169 [AV_PIX_FMT_BGR48LE] = { 1, 1 },
170 [AV_PIX_FMT_BGRA64BE] = { 1, 1, 1 },
171 [AV_PIX_FMT_BGRA64LE] = { 1, 1, 1 },
172 [AV_PIX_FMT_YUV420P9BE] = { 1, 1 },
173 [AV_PIX_FMT_YUV420P9LE] = { 1, 1 },
174 [AV_PIX_FMT_YUV420P10BE] = { 1, 1 },
175 [AV_PIX_FMT_YUV420P10LE] = { 1, 1 },
176 [AV_PIX_FMT_YUV420P12BE] = { 1, 1 },
177 [AV_PIX_FMT_YUV420P12LE] = { 1, 1 },
178 [AV_PIX_FMT_YUV420P14BE] = { 1, 1 },
179 [AV_PIX_FMT_YUV420P14LE] = { 1, 1 },
180 [AV_PIX_FMT_YUV422P9BE] = { 1, 1 },
181 [AV_PIX_FMT_YUV422P9LE] = { 1, 1 },
182 [AV_PIX_FMT_YUV422P10BE] = { 1, 1 },
183 [AV_PIX_FMT_YUV422P10LE] = { 1, 1 },
184 [AV_PIX_FMT_YUV422P12BE] = { 1, 1 },
185 [AV_PIX_FMT_YUV422P12LE] = { 1, 1 },
186 [AV_PIX_FMT_YUV422P14BE] = { 1, 1 },
187 [AV_PIX_FMT_YUV422P14LE] = { 1, 1 },
188 [AV_PIX_FMT_YUV444P9BE] = { 1, 1 },
189 [AV_PIX_FMT_YUV444P9LE] = { 1, 1 },
190 [AV_PIX_FMT_YUV444P10BE] = { 1, 1 },
191 [AV_PIX_FMT_YUV444P10LE] = { 1, 1 },
192 [AV_PIX_FMT_YUV444P12BE] = { 1, 1 },
193 [AV_PIX_FMT_YUV444P12LE] = { 1, 1 },
194 [AV_PIX_FMT_YUV444P14BE] = { 1, 1 },
195 [AV_PIX_FMT_YUV444P14LE] = { 1, 1 },
196 [AV_PIX_FMT_GBRP] = { 1, 1 },
197 [AV_PIX_FMT_GBRP9LE] = { 1, 1 },
198 [AV_PIX_FMT_GBRP9BE] = { 1, 1 },
199 [AV_PIX_FMT_GBRP10LE] = { 1, 1 },
200 [AV_PIX_FMT_GBRP10BE] = { 1, 1 },
201 [AV_PIX_FMT_GBRP12LE] = { 1, 1 },
202 [AV_PIX_FMT_GBRP12BE] = { 1, 1 },
203 [AV_PIX_FMT_GBRP14LE] = { 1, 1 },
204 [AV_PIX_FMT_GBRP14BE] = { 1, 1 },
205 [AV_PIX_FMT_GBRP16LE] = { 1, 0 },
206 [AV_PIX_FMT_GBRP16BE] = { 1, 0 },
207 [AV_PIX_FMT_GBRAP] = { 1, 1 },
208 [AV_PIX_FMT_GBRAP16LE] = { 1, 0 },
209 [AV_PIX_FMT_GBRAP16BE] = { 1, 0 },
210 [AV_PIX_FMT_BAYER_BGGR8] = { 1, 0 },
211 [AV_PIX_FMT_BAYER_RGGB8] = { 1, 0 },
212 [AV_PIX_FMT_BAYER_GBRG8] = { 1, 0 },
213 [AV_PIX_FMT_BAYER_GRBG8] = { 1, 0 },
214 [AV_PIX_FMT_BAYER_BGGR16LE] = { 1, 0 },
215 [AV_PIX_FMT_BAYER_BGGR16BE] = { 1, 0 },
216 [AV_PIX_FMT_BAYER_RGGB16LE] = { 1, 0 },
217 [AV_PIX_FMT_BAYER_RGGB16BE] = { 1, 0 },
218 [AV_PIX_FMT_BAYER_GBRG16LE] = { 1, 0 },
219 [AV_PIX_FMT_BAYER_GBRG16BE] = { 1, 0 },
220 [AV_PIX_FMT_BAYER_GRBG16LE] = { 1, 0 },
221 [AV_PIX_FMT_BAYER_GRBG16BE] = { 1, 0 },
222 [AV_PIX_FMT_XYZ12BE] = { 1, 1, 1 },
223 [AV_PIX_FMT_XYZ12LE] = { 1, 1, 1 },
226 int sws_isSupportedInput(enum AVPixelFormat pix_fmt)
228 return (unsigned)pix_fmt < AV_PIX_FMT_NB ?
229 format_entries[pix_fmt].is_supported_in : 0;
232 int sws_isSupportedOutput(enum AVPixelFormat pix_fmt)
234 return (unsigned)pix_fmt < AV_PIX_FMT_NB ?
235 format_entries[pix_fmt].is_supported_out : 0;
238 int sws_isSupportedEndiannessConversion(enum AVPixelFormat pix_fmt)
240 return (unsigned)pix_fmt < AV_PIX_FMT_NB ?
241 format_entries[pix_fmt].is_supported_endianness : 0;
244 static double getSplineCoeff(double a, double b, double c, double d,
248 return ((d * dist + c) * dist + b) * dist + a;
250 return getSplineCoeff(0.0,
251 b + 2.0 * c + 3.0 * d,
253 -b - 3.0 * c - 6.0 * d,
257 static av_cold int get_local_pos(SwsContext *s, int chr_subsample, int pos, int dir)
259 if (pos == -1 || pos <= -513) {
260 pos = (128 << chr_subsample) - 128;
262 pos += 128; // relative to ideal left edge
263 return pos >> chr_subsample;
267 int flag; ///< flag associated to the algorithm
268 const char *description; ///< human-readable description
269 int size_factor; ///< size factor used when initing the filters
272 static const ScaleAlgorithm scale_algorithms[] = {
273 { SWS_AREA, "area averaging", 1 /* downscale only, for upscale it is bilinear */ },
274 { SWS_BICUBIC, "bicubic", 4 },
275 { SWS_BICUBLIN, "luma bicubic / chroma bilinear", -1 },
276 { SWS_BILINEAR, "bilinear", 2 },
277 { SWS_FAST_BILINEAR, "fast bilinear", -1 },
278 { SWS_GAUSS, "Gaussian", 8 /* infinite ;) */ },
279 { SWS_LANCZOS, "Lanczos", -1 /* custom */ },
280 { SWS_POINT, "nearest neighbor / point", -1 },
281 { SWS_SINC, "sinc", 20 /* infinite ;) */ },
282 { SWS_SPLINE, "bicubic spline", 20 /* infinite :)*/ },
283 { SWS_X, "experimental", 8 },
286 static av_cold int initFilter(int16_t **outFilter, int32_t **filterPos,
287 int *outFilterSize, int xInc, int srcW,
288 int dstW, int filterAlign, int one,
289 int flags, int cpu_flags,
290 SwsVector *srcFilter, SwsVector *dstFilter,
291 double param[2], int srcPos, int dstPos)
297 int64_t *filter = NULL;
298 int64_t *filter2 = NULL;
299 const int64_t fone = 1LL << (54 - FFMIN(av_log2(srcW/dstW), 8));
302 emms_c(); // FIXME should not be required but IS (even for non-MMX versions)
304 // NOTE: the +3 is for the MMX(+1) / SSE(+3) scaler which reads over the end
305 FF_ALLOC_ARRAY_OR_GOTO(NULL, *filterPos, (dstW + 3), sizeof(**filterPos), fail);
307 if (FFABS(xInc - 0x10000) < 10 && srcPos == dstPos) { // unscaled
310 FF_ALLOCZ_ARRAY_OR_GOTO(NULL, filter,
311 dstW, sizeof(*filter) * filterSize, fail);
313 for (i = 0; i < dstW; i++) {
314 filter[i * filterSize] = fone;
317 } else if (flags & SWS_POINT) { // lame looking point sampling mode
321 FF_ALLOC_ARRAY_OR_GOTO(NULL, filter,
322 dstW, sizeof(*filter) * filterSize, fail);
324 xDstInSrc = ((dstPos*(int64_t)xInc)>>8) - ((srcPos*0x8000LL)>>7);
325 for (i = 0; i < dstW; i++) {
326 int xx = (xDstInSrc - ((filterSize - 1) << 15) + (1 << 15)) >> 16;
328 (*filterPos)[i] = xx;
332 } else if ((xInc <= (1 << 16) && (flags & SWS_AREA)) ||
333 (flags & SWS_FAST_BILINEAR)) { // bilinear upscale
337 FF_ALLOC_ARRAY_OR_GOTO(NULL, filter,
338 dstW, sizeof(*filter) * filterSize, fail);
340 xDstInSrc = ((dstPos*(int64_t)xInc)>>8) - ((srcPos*0x8000LL)>>7);
341 for (i = 0; i < dstW; i++) {
342 int xx = (xDstInSrc - ((filterSize - 1) << 15) + (1 << 15)) >> 16;
345 (*filterPos)[i] = xx;
346 // bilinear upscale / linear interpolate / area averaging
347 for (j = 0; j < filterSize; j++) {
348 int64_t coeff= fone - FFABS(((int64_t)xx<<16) - xDstInSrc)*(fone>>16);
351 filter[i * filterSize + j] = coeff;
360 for (i = 0; i < FF_ARRAY_ELEMS(scale_algorithms); i++) {
361 if (flags & scale_algorithms[i].flag && scale_algorithms[i].size_factor > 0) {
362 sizeFactor = scale_algorithms[i].size_factor;
366 if (flags & SWS_LANCZOS)
367 sizeFactor = param[0] != SWS_PARAM_DEFAULT ? ceil(2 * param[0]) : 6;
368 av_assert0(sizeFactor > 0);
371 filterSize = 1 + sizeFactor; // upscale
373 filterSize = 1 + (sizeFactor * srcW + dstW - 1) / dstW;
375 filterSize = FFMIN(filterSize, srcW - 2);
376 filterSize = FFMAX(filterSize, 1);
378 FF_ALLOC_ARRAY_OR_GOTO(NULL, filter,
379 dstW, sizeof(*filter) * filterSize, fail);
381 xDstInSrc = ((dstPos*(int64_t)xInc)>>7) - ((srcPos*0x10000LL)>>7);
382 for (i = 0; i < dstW; i++) {
383 int xx = (xDstInSrc - ((int64_t)(filterSize - 2) << 16)) / (1 << 17);
385 (*filterPos)[i] = xx;
386 for (j = 0; j < filterSize; j++) {
387 int64_t d = (FFABS(((int64_t)xx * (1 << 17)) - xDstInSrc)) << 13;
393 floatd = d * (1.0 / (1 << 30));
395 if (flags & SWS_BICUBIC) {
396 int64_t B = (param[0] != SWS_PARAM_DEFAULT ? param[0] : 0) * (1 << 24);
397 int64_t C = (param[1] != SWS_PARAM_DEFAULT ? param[1] : 0.6) * (1 << 24);
399 if (d >= 1LL << 31) {
402 int64_t dd = (d * d) >> 30;
403 int64_t ddd = (dd * d) >> 30;
406 coeff = (12 * (1 << 24) - 9 * B - 6 * C) * ddd +
407 (-18 * (1 << 24) + 12 * B + 6 * C) * dd +
408 (6 * (1 << 24) - 2 * B) * (1 << 30);
410 coeff = (-B - 6 * C) * ddd +
411 (6 * B + 30 * C) * dd +
412 (-12 * B - 48 * C) * d +
413 (8 * B + 24 * C) * (1 << 30);
415 coeff /= (1LL<<54)/fone;
418 else if (flags & SWS_X) {
419 double p = param ? param * 0.01 : 0.3;
420 coeff = d ? sin(d * M_PI) / (d * M_PI) : 1.0;
421 coeff *= pow(2.0, -p * d * d);
424 else if (flags & SWS_X) {
425 double A = param[0] != SWS_PARAM_DEFAULT ? param[0] : 1.0;
429 c = cos(floatd * M_PI);
436 coeff = (c * 0.5 + 0.5) * fone;
437 } else if (flags & SWS_AREA) {
438 int64_t d2 = d - (1 << 29);
439 if (d2 * xInc < -(1LL << (29 + 16)))
440 coeff = 1.0 * (1LL << (30 + 16));
441 else if (d2 * xInc < (1LL << (29 + 16)))
442 coeff = -d2 * xInc + (1LL << (29 + 16));
445 coeff *= fone >> (30 + 16);
446 } else if (flags & SWS_GAUSS) {
447 double p = param[0] != SWS_PARAM_DEFAULT ? param[0] : 3.0;
448 coeff = (pow(2.0, -p * floatd * floatd)) * fone;
449 } else if (flags & SWS_SINC) {
450 coeff = (d ? sin(floatd * M_PI) / (floatd * M_PI) : 1.0) * fone;
451 } else if (flags & SWS_LANCZOS) {
452 double p = param[0] != SWS_PARAM_DEFAULT ? param[0] : 3.0;
453 coeff = (d ? sin(floatd * M_PI) * sin(floatd * M_PI / p) /
454 (floatd * floatd * M_PI * M_PI / p) : 1.0) * fone;
457 } else if (flags & SWS_BILINEAR) {
458 coeff = (1 << 30) - d;
462 } else if (flags & SWS_SPLINE) {
463 double p = -2.196152422706632;
464 coeff = getSplineCoeff(1.0, 0.0, p, -p - 1.0, floatd) * fone;
469 filter[i * filterSize + j] = coeff;
472 xDstInSrc += 2 * xInc;
476 /* apply src & dst Filter to filter -> filter2
479 av_assert0(filterSize > 0);
480 filter2Size = filterSize;
482 filter2Size += srcFilter->length - 1;
484 filter2Size += dstFilter->length - 1;
485 av_assert0(filter2Size > 0);
486 FF_ALLOCZ_ARRAY_OR_GOTO(NULL, filter2, dstW, filter2Size * sizeof(*filter2), fail);
488 for (i = 0; i < dstW; i++) {
492 for (k = 0; k < srcFilter->length; k++) {
493 for (j = 0; j < filterSize; j++)
494 filter2[i * filter2Size + k + j] +=
495 srcFilter->coeff[k] * filter[i * filterSize + j];
498 for (j = 0; j < filterSize; j++)
499 filter2[i * filter2Size + j] = filter[i * filterSize + j];
503 (*filterPos)[i] += (filterSize - 1) / 2 - (filter2Size - 1) / 2;
507 /* try to reduce the filter-size (step1 find size and shift left) */
508 // Assume it is near normalized (*0.5 or *2.0 is OK but * 0.001 is not).
510 for (i = dstW - 1; i >= 0; i--) {
511 int min = filter2Size;
513 int64_t cutOff = 0.0;
515 /* get rid of near zero elements on the left by shifting left */
516 for (j = 0; j < filter2Size; j++) {
518 cutOff += FFABS(filter2[i * filter2Size]);
520 if (cutOff > SWS_MAX_REDUCE_CUTOFF * fone)
523 /* preserve monotonicity because the core can't handle the
524 * filter otherwise */
525 if (i < dstW - 1 && (*filterPos)[i] >= (*filterPos)[i + 1])
528 // move filter coefficients left
529 for (k = 1; k < filter2Size; k++)
530 filter2[i * filter2Size + k - 1] = filter2[i * filter2Size + k];
531 filter2[i * filter2Size + k - 1] = 0;
536 /* count near zeros on the right */
537 for (j = filter2Size - 1; j > 0; j--) {
538 cutOff += FFABS(filter2[i * filter2Size + j]);
540 if (cutOff > SWS_MAX_REDUCE_CUTOFF * fone)
545 if (min > minFilterSize)
549 if (PPC_ALTIVEC(cpu_flags)) {
550 // we can handle the special case 4, so we don't want to go the full 8
551 if (minFilterSize < 5)
554 /* We really don't want to waste our time doing useless computation, so
555 * fall back on the scalar C code for very small filters.
556 * Vectorizing is worth it only if you have a decent-sized vector. */
557 if (minFilterSize < 3)
561 if (HAVE_MMX && cpu_flags & AV_CPU_FLAG_MMX) {
562 // special case for unscaled vertical filtering
563 if (minFilterSize == 1 && filterAlign == 2)
567 av_assert0(minFilterSize > 0);
568 filterSize = (minFilterSize + (filterAlign - 1)) & (~(filterAlign - 1));
569 av_assert0(filterSize > 0);
570 filter = av_malloc_array(dstW, filterSize * sizeof(*filter));
573 if (filterSize >= MAX_FILTER_SIZE * 16 /
574 ((flags & SWS_ACCURATE_RND) ? APCK_SIZE : 16)) {
575 ret = RETCODE_USE_CASCADE;
578 *outFilterSize = filterSize;
580 if (flags & SWS_PRINT_INFO)
581 av_log(NULL, AV_LOG_VERBOSE,
582 "SwScaler: reducing / aligning filtersize %d -> %d\n",
583 filter2Size, filterSize);
584 /* try to reduce the filter-size (step2 reduce it) */
585 for (i = 0; i < dstW; i++) {
588 for (j = 0; j < filterSize; j++) {
589 if (j >= filter2Size)
590 filter[i * filterSize + j] = 0;
592 filter[i * filterSize + j] = filter2[i * filter2Size + j];
593 if ((flags & SWS_BITEXACT) && j >= minFilterSize)
594 filter[i * filterSize + j] = 0;
598 // FIXME try to align filterPos if possible
601 for (i = 0; i < dstW; i++) {
603 if ((*filterPos)[i] < 0) {
604 // move filter coefficients left to compensate for filterPos
605 for (j = 1; j < filterSize; j++) {
606 int left = FFMAX(j + (*filterPos)[i], 0);
607 filter[i * filterSize + left] += filter[i * filterSize + j];
608 filter[i * filterSize + j] = 0;
613 if ((*filterPos)[i] + filterSize > srcW) {
614 int shift = (*filterPos)[i] + FFMIN(filterSize - srcW, 0);
617 for (j = filterSize - 1; j >= 0; j--) {
618 if ((*filterPos)[i] + j >= srcW) {
619 acc += filter[i * filterSize + j];
620 filter[i * filterSize + j] = 0;
623 for (j = filterSize - 1; j >= 0; j--) {
625 filter[i * filterSize + j] = 0;
627 filter[i * filterSize + j] = filter[i * filterSize + j - shift];
631 (*filterPos)[i]-= shift;
632 filter[i * filterSize + srcW - 1 - (*filterPos)[i]] += acc;
634 av_assert0((*filterPos)[i] >= 0);
635 av_assert0((*filterPos)[i] < srcW);
636 if ((*filterPos)[i] + filterSize > srcW) {
637 for (j = 0; j < filterSize; j++) {
638 av_assert0((*filterPos)[i] + j < srcW || !filter[i * filterSize + j]);
643 // Note the +1 is for the MMX scaler which reads over the end
644 /* align at 16 for AltiVec (needed by hScale_altivec_real) */
645 FF_ALLOCZ_ARRAY_OR_GOTO(NULL, *outFilter,
646 (dstW + 3), *outFilterSize * sizeof(int16_t), fail);
648 /* normalize & store in outFilter */
649 for (i = 0; i < dstW; i++) {
654 for (j = 0; j < filterSize; j++) {
655 sum += filter[i * filterSize + j];
657 sum = (sum + one / 2) / one;
659 av_log(NULL, AV_LOG_WARNING, "SwScaler: zero vector in scaling\n");
662 for (j = 0; j < *outFilterSize; j++) {
663 int64_t v = filter[i * filterSize + j] + error;
664 int intV = ROUNDED_DIV(v, sum);
665 (*outFilter)[i * (*outFilterSize) + j] = intV;
666 error = v - intV * sum;
670 (*filterPos)[dstW + 0] =
671 (*filterPos)[dstW + 1] =
672 (*filterPos)[dstW + 2] = (*filterPos)[dstW - 1]; /* the MMX/SSE scaler will
673 * read over the end */
674 for (i = 0; i < *outFilterSize; i++) {
675 int k = (dstW - 1) * (*outFilterSize) + i;
676 (*outFilter)[k + 1 * (*outFilterSize)] =
677 (*outFilter)[k + 2 * (*outFilterSize)] =
678 (*outFilter)[k + 3 * (*outFilterSize)] = (*outFilter)[k];
685 av_log(NULL, ret == RETCODE_USE_CASCADE ? AV_LOG_DEBUG : AV_LOG_ERROR, "sws: initFilter failed\n");
691 static void fill_rgb2yuv_table(SwsContext *c, const int table[4], int dstRange)
693 int64_t W, V, Z, Cy, Cu, Cv;
694 int64_t vr = table[0];
695 int64_t ub = table[1];
696 int64_t ug = -table[2];
697 int64_t vg = -table[3];
700 uint8_t *p = (uint8_t*)c->input_rgb2yuv_table;
702 static const int8_t map[] = {
703 BY_IDX, GY_IDX, -1 , BY_IDX, BY_IDX, GY_IDX, -1 , BY_IDX,
704 RY_IDX, -1 , GY_IDX, RY_IDX, RY_IDX, -1 , GY_IDX, RY_IDX,
705 RY_IDX, GY_IDX, -1 , RY_IDX, RY_IDX, GY_IDX, -1 , RY_IDX,
706 BY_IDX, -1 , GY_IDX, BY_IDX, BY_IDX, -1 , GY_IDX, BY_IDX,
707 BU_IDX, GU_IDX, -1 , BU_IDX, BU_IDX, GU_IDX, -1 , BU_IDX,
708 RU_IDX, -1 , GU_IDX, RU_IDX, RU_IDX, -1 , GU_IDX, RU_IDX,
709 RU_IDX, GU_IDX, -1 , RU_IDX, RU_IDX, GU_IDX, -1 , RU_IDX,
710 BU_IDX, -1 , GU_IDX, BU_IDX, BU_IDX, -1 , GU_IDX, BU_IDX,
711 BV_IDX, GV_IDX, -1 , BV_IDX, BV_IDX, GV_IDX, -1 , BV_IDX,
712 RV_IDX, -1 , GV_IDX, RV_IDX, RV_IDX, -1 , GV_IDX, RV_IDX,
713 RV_IDX, GV_IDX, -1 , RV_IDX, RV_IDX, GV_IDX, -1 , RV_IDX,
714 BV_IDX, -1 , GV_IDX, BV_IDX, BV_IDX, -1 , GV_IDX, BV_IDX,
715 RY_IDX, BY_IDX, RY_IDX, BY_IDX, RY_IDX, BY_IDX, RY_IDX, BY_IDX,
716 BY_IDX, RY_IDX, BY_IDX, RY_IDX, BY_IDX, RY_IDX, BY_IDX, RY_IDX,
717 GY_IDX, -1 , GY_IDX, -1 , GY_IDX, -1 , GY_IDX, -1 ,
718 -1 , GY_IDX, -1 , GY_IDX, -1 , GY_IDX, -1 , GY_IDX,
719 RU_IDX, BU_IDX, RU_IDX, BU_IDX, RU_IDX, BU_IDX, RU_IDX, BU_IDX,
720 BU_IDX, RU_IDX, BU_IDX, RU_IDX, BU_IDX, RU_IDX, BU_IDX, RU_IDX,
721 GU_IDX, -1 , GU_IDX, -1 , GU_IDX, -1 , GU_IDX, -1 ,
722 -1 , GU_IDX, -1 , GU_IDX, -1 , GU_IDX, -1 , GU_IDX,
723 RV_IDX, BV_IDX, RV_IDX, BV_IDX, RV_IDX, BV_IDX, RV_IDX, BV_IDX,
724 BV_IDX, RV_IDX, BV_IDX, RV_IDX, BV_IDX, RV_IDX, BV_IDX, RV_IDX,
725 GV_IDX, -1 , GV_IDX, -1 , GV_IDX, -1 , GV_IDX, -1 ,
726 -1 , GV_IDX, -1 , GV_IDX, -1 , GV_IDX, -1 , GV_IDX, //23
727 -1 , -1 , -1 , -1 , -1 , -1 , -1 , -1 , //24
728 -1 , -1 , -1 , -1 , -1 , -1 , -1 , -1 , //25
729 -1 , -1 , -1 , -1 , -1 , -1 , -1 , -1 , //26
730 -1 , -1 , -1 , -1 , -1 , -1 , -1 , -1 , //27
731 -1 , -1 , -1 , -1 , -1 , -1 , -1 , -1 , //28
732 -1 , -1 , -1 , -1 , -1 , -1 , -1 , -1 , //29
733 -1 , -1 , -1 , -1 , -1 , -1 , -1 , -1 , //30
734 -1 , -1 , -1 , -1 , -1 , -1 , -1 , -1 , //31
735 BY_IDX, GY_IDX, RY_IDX, -1 , -1 , -1 , -1 , -1 , //32
736 BU_IDX, GU_IDX, RU_IDX, -1 , -1 , -1 , -1 , -1 , //33
737 BV_IDX, GV_IDX, RV_IDX, -1 , -1 , -1 , -1 , -1 , //34
740 dstRange = 0; //FIXME range = 1 is handled elsewhere
750 W = ROUNDED_DIV(ONE*ONE*ug, ub);
751 V = ROUNDED_DIV(ONE*ONE*vg, vr);
754 Cy = ROUNDED_DIV(cy*Z, ONE);
755 Cu = ROUNDED_DIV(ub*Z, ONE);
756 Cv = ROUNDED_DIV(vr*Z, ONE);
758 c->input_rgb2yuv_table[RY_IDX] = -ROUNDED_DIV((1 << RGB2YUV_SHIFT)*V , Cy);
759 c->input_rgb2yuv_table[GY_IDX] = ROUNDED_DIV((1 << RGB2YUV_SHIFT)*ONE*ONE , Cy);
760 c->input_rgb2yuv_table[BY_IDX] = -ROUNDED_DIV((1 << RGB2YUV_SHIFT)*W , Cy);
762 c->input_rgb2yuv_table[RU_IDX] = ROUNDED_DIV((1 << RGB2YUV_SHIFT)*V , Cu);
763 c->input_rgb2yuv_table[GU_IDX] = -ROUNDED_DIV((1 << RGB2YUV_SHIFT)*ONE*ONE , Cu);
764 c->input_rgb2yuv_table[BU_IDX] = ROUNDED_DIV((1 << RGB2YUV_SHIFT)*(Z+W) , Cu);
766 c->input_rgb2yuv_table[RV_IDX] = ROUNDED_DIV((1 << RGB2YUV_SHIFT)*(V+Z) , Cv);
767 c->input_rgb2yuv_table[GV_IDX] = -ROUNDED_DIV((1 << RGB2YUV_SHIFT)*ONE*ONE , Cv);
768 c->input_rgb2yuv_table[BV_IDX] = ROUNDED_DIV((1 << RGB2YUV_SHIFT)*W , Cv);
770 if(/*!dstRange && */!memcmp(table, ff_yuv2rgb_coeffs[SWS_CS_DEFAULT], sizeof(ff_yuv2rgb_coeffs[SWS_CS_DEFAULT]))) {
771 c->input_rgb2yuv_table[BY_IDX] = ((int)(0.114 * 219 / 255 * (1 << RGB2YUV_SHIFT) + 0.5));
772 c->input_rgb2yuv_table[BV_IDX] = (-(int)(0.081 * 224 / 255 * (1 << RGB2YUV_SHIFT) + 0.5));
773 c->input_rgb2yuv_table[BU_IDX] = ((int)(0.500 * 224 / 255 * (1 << RGB2YUV_SHIFT) + 0.5));
774 c->input_rgb2yuv_table[GY_IDX] = ((int)(0.587 * 219 / 255 * (1 << RGB2YUV_SHIFT) + 0.5));
775 c->input_rgb2yuv_table[GV_IDX] = (-(int)(0.419 * 224 / 255 * (1 << RGB2YUV_SHIFT) + 0.5));
776 c->input_rgb2yuv_table[GU_IDX] = (-(int)(0.331 * 224 / 255 * (1 << RGB2YUV_SHIFT) + 0.5));
777 c->input_rgb2yuv_table[RY_IDX] = ((int)(0.299 * 219 / 255 * (1 << RGB2YUV_SHIFT) + 0.5));
778 c->input_rgb2yuv_table[RV_IDX] = ((int)(0.500 * 224 / 255 * (1 << RGB2YUV_SHIFT) + 0.5));
779 c->input_rgb2yuv_table[RU_IDX] = (-(int)(0.169 * 224 / 255 * (1 << RGB2YUV_SHIFT) + 0.5));
781 for(i=0; i<FF_ARRAY_ELEMS(map); i++)
782 AV_WL16(p + 16*4 + 2*i, map[i] >= 0 ? c->input_rgb2yuv_table[map[i]] : 0);
785 static void fill_xyztables(struct SwsContext *c)
788 double xyzgamma = XYZ_GAMMA;
789 double rgbgamma = 1.0 / RGB_GAMMA;
790 double xyzgammainv = 1.0 / XYZ_GAMMA;
791 double rgbgammainv = RGB_GAMMA;
792 static const int16_t xyz2rgb_matrix[3][4] = {
793 {13270, -6295, -2041},
795 { 228, -835, 4329} };
796 static const int16_t rgb2xyz_matrix[3][4] = {
800 static int16_t xyzgamma_tab[4096], rgbgamma_tab[4096], xyzgammainv_tab[4096], rgbgammainv_tab[4096];
802 memcpy(c->xyz2rgb_matrix, xyz2rgb_matrix, sizeof(c->xyz2rgb_matrix));
803 memcpy(c->rgb2xyz_matrix, rgb2xyz_matrix, sizeof(c->rgb2xyz_matrix));
804 c->xyzgamma = xyzgamma_tab;
805 c->rgbgamma = rgbgamma_tab;
806 c->xyzgammainv = xyzgammainv_tab;
807 c->rgbgammainv = rgbgammainv_tab;
809 if (rgbgamma_tab[4095])
812 /* set gamma vectors */
813 for (i = 0; i < 4096; i++) {
814 xyzgamma_tab[i] = lrint(pow(i / 4095.0, xyzgamma) * 4095.0);
815 rgbgamma_tab[i] = lrint(pow(i / 4095.0, rgbgamma) * 4095.0);
816 xyzgammainv_tab[i] = lrint(pow(i / 4095.0, xyzgammainv) * 4095.0);
817 rgbgammainv_tab[i] = lrint(pow(i / 4095.0, rgbgammainv) * 4095.0);
821 int sws_setColorspaceDetails(struct SwsContext *c, const int inv_table[4],
822 int srcRange, const int table[4], int dstRange,
823 int brightness, int contrast, int saturation)
825 const AVPixFmtDescriptor *desc_dst;
826 const AVPixFmtDescriptor *desc_src;
828 memmove(c->srcColorspaceTable, inv_table, sizeof(int) * 4);
829 memmove(c->dstColorspaceTable, table, sizeof(int) * 4);
832 desc_dst = av_pix_fmt_desc_get(c->dstFormat);
833 desc_src = av_pix_fmt_desc_get(c->srcFormat);
835 if(!isYUV(c->dstFormat) && !isGray(c->dstFormat))
837 if(!isYUV(c->srcFormat) && !isGray(c->srcFormat))
840 c->brightness = brightness;
841 c->contrast = contrast;
842 c->saturation = saturation;
843 if (c->srcRange != srcRange || c->dstRange != dstRange)
845 c->srcRange = srcRange;
846 c->dstRange = dstRange;
848 //The srcBpc check is possibly wrong but we seem to lack a definitive reference to test this
849 //and what we have in ticket 2939 looks better with this check
850 if (need_reinit && (c->srcBpc == 8 || !isYUV(c->srcFormat)))
851 ff_sws_init_range_convert(c);
853 if ((isYUV(c->dstFormat) || isGray(c->dstFormat)) && (isYUV(c->srcFormat) || isGray(c->srcFormat)))
856 c->dstFormatBpp = av_get_bits_per_pixel(desc_dst);
857 c->srcFormatBpp = av_get_bits_per_pixel(desc_src);
859 if (!isYUV(c->dstFormat) && !isGray(c->dstFormat)) {
860 ff_yuv2rgb_c_init_tables(c, inv_table, srcRange, brightness,
861 contrast, saturation);
865 ff_yuv2rgb_init_tables_ppc(c, inv_table, brightness,
866 contrast, saturation);
869 fill_rgb2yuv_table(c, table, dstRange);
874 int sws_getColorspaceDetails(struct SwsContext *c, int **inv_table,
875 int *srcRange, int **table, int *dstRange,
876 int *brightness, int *contrast, int *saturation)
881 *inv_table = c->srcColorspaceTable;
882 *table = c->dstColorspaceTable;
883 *srcRange = c->srcRange;
884 *dstRange = c->dstRange;
885 *brightness = c->brightness;
886 *contrast = c->contrast;
887 *saturation = c->saturation;
892 static int handle_jpeg(enum AVPixelFormat *format)
895 case AV_PIX_FMT_YUVJ420P:
896 *format = AV_PIX_FMT_YUV420P;
898 case AV_PIX_FMT_YUVJ411P:
899 *format = AV_PIX_FMT_YUV411P;
901 case AV_PIX_FMT_YUVJ422P:
902 *format = AV_PIX_FMT_YUV422P;
904 case AV_PIX_FMT_YUVJ444P:
905 *format = AV_PIX_FMT_YUV444P;
907 case AV_PIX_FMT_YUVJ440P:
908 *format = AV_PIX_FMT_YUV440P;
910 case AV_PIX_FMT_GRAY8:
911 case AV_PIX_FMT_GRAY16LE:
912 case AV_PIX_FMT_GRAY16BE:
919 static int handle_0alpha(enum AVPixelFormat *format)
922 case AV_PIX_FMT_0BGR : *format = AV_PIX_FMT_ABGR ; return 1;
923 case AV_PIX_FMT_BGR0 : *format = AV_PIX_FMT_BGRA ; return 4;
924 case AV_PIX_FMT_0RGB : *format = AV_PIX_FMT_ARGB ; return 1;
925 case AV_PIX_FMT_RGB0 : *format = AV_PIX_FMT_RGBA ; return 4;
930 static int handle_xyz(enum AVPixelFormat *format)
933 case AV_PIX_FMT_XYZ12BE : *format = AV_PIX_FMT_RGB48BE; return 1;
934 case AV_PIX_FMT_XYZ12LE : *format = AV_PIX_FMT_RGB48LE; return 1;
939 static void handle_formats(SwsContext *c)
941 c->src0Alpha |= handle_0alpha(&c->srcFormat);
942 c->dst0Alpha |= handle_0alpha(&c->dstFormat);
943 c->srcXYZ |= handle_xyz(&c->srcFormat);
944 c->dstXYZ |= handle_xyz(&c->dstFormat);
945 if (c->srcXYZ || c->dstXYZ)
949 SwsContext *sws_alloc_context(void)
951 SwsContext *c = av_mallocz(sizeof(SwsContext));
953 av_assert0(offsetof(SwsContext, redDither) + DITHER32_INT == offsetof(SwsContext, dither32));
956 c->av_class = &sws_context_class;
957 av_opt_set_defaults(c);
963 static uint16_t * alloc_gamma_tbl(double e)
967 tbl = (uint16_t*)av_malloc(sizeof(uint16_t) * 1 << 16);
971 for (i = 0; i < 65536; ++i) {
972 tbl[i] = pow(i / 65535.0, e) * 65535.0;
977 av_cold int sws_init_context(SwsContext *c, SwsFilter *srcFilter,
978 SwsFilter *dstFilter)
981 int usesVFilter, usesHFilter;
983 SwsFilter dummyFilter = { NULL, NULL, NULL, NULL };
988 int dst_stride = FFALIGN(dstW * sizeof(int16_t) + 66, 16);
989 int flags, cpu_flags;
990 enum AVPixelFormat srcFormat = c->srcFormat;
991 enum AVPixelFormat dstFormat = c->dstFormat;
992 const AVPixFmtDescriptor *desc_src;
993 const AVPixFmtDescriptor *desc_dst;
995 enum AVPixelFormat tmpFmt;
997 cpu_flags = av_get_cpu_flags();
1003 unscaled = (srcW == dstW && srcH == dstH);
1005 c->srcRange |= handle_jpeg(&c->srcFormat);
1006 c->dstRange |= handle_jpeg(&c->dstFormat);
1008 if(srcFormat!=c->srcFormat || dstFormat!=c->dstFormat)
1009 av_log(c, AV_LOG_WARNING, "deprecated pixel format used, make sure you did set range correctly\n");
1011 if (!c->contrast && !c->saturation && !c->dstFormatBpp)
1012 sws_setColorspaceDetails(c, ff_yuv2rgb_coeffs[SWS_CS_DEFAULT], c->srcRange,
1013 ff_yuv2rgb_coeffs[SWS_CS_DEFAULT],
1014 c->dstRange, 0, 1 << 16, 1 << 16);
1017 srcFormat = c->srcFormat;
1018 dstFormat = c->dstFormat;
1019 desc_src = av_pix_fmt_desc_get(srcFormat);
1020 desc_dst = av_pix_fmt_desc_get(dstFormat);
1022 if (!(unscaled && sws_isSupportedEndiannessConversion(srcFormat) &&
1023 av_pix_fmt_swap_endianness(srcFormat) == dstFormat)) {
1024 if (!sws_isSupportedInput(srcFormat)) {
1025 av_log(c, AV_LOG_ERROR, "%s is not supported as input pixel format\n",
1026 av_get_pix_fmt_name(srcFormat));
1027 return AVERROR(EINVAL);
1029 if (!sws_isSupportedOutput(dstFormat)) {
1030 av_log(c, AV_LOG_ERROR, "%s is not supported as output pixel format\n",
1031 av_get_pix_fmt_name(dstFormat));
1032 return AVERROR(EINVAL);
1036 i = flags & (SWS_POINT |
1048 /* provide a default scaler if not set by caller */
1050 if (dstW < srcW && dstH < srcH)
1051 flags |= SWS_BICUBIC;
1052 else if (dstW > srcW && dstH > srcH)
1053 flags |= SWS_BICUBIC;
1055 flags |= SWS_BICUBIC;
1057 } else if (i & (i - 1)) {
1058 av_log(c, AV_LOG_ERROR,
1059 "Exactly one scaler algorithm must be chosen, got %X\n", i);
1060 return AVERROR(EINVAL);
1063 if (srcW < 1 || srcH < 1 || dstW < 1 || dstH < 1) {
1064 /* FIXME check if these are enough and try to lower them after
1065 * fixing the relevant parts of the code */
1066 av_log(c, AV_LOG_ERROR, "%dx%d -> %dx%d is invalid scaling dimension\n",
1067 srcW, srcH, dstW, dstH);
1068 return AVERROR(EINVAL);
1072 dstFilter = &dummyFilter;
1074 srcFilter = &dummyFilter;
1076 c->lumXInc = (((int64_t)srcW << 16) + (dstW >> 1)) / dstW;
1077 c->lumYInc = (((int64_t)srcH << 16) + (dstH >> 1)) / dstH;
1078 c->dstFormatBpp = av_get_bits_per_pixel(desc_dst);
1079 c->srcFormatBpp = av_get_bits_per_pixel(desc_src);
1080 c->vRounder = 4 * 0x0001000100010001ULL;
1082 usesVFilter = (srcFilter->lumV && srcFilter->lumV->length > 1) ||
1083 (srcFilter->chrV && srcFilter->chrV->length > 1) ||
1084 (dstFilter->lumV && dstFilter->lumV->length > 1) ||
1085 (dstFilter->chrV && dstFilter->chrV->length > 1);
1086 usesHFilter = (srcFilter->lumH && srcFilter->lumH->length > 1) ||
1087 (srcFilter->chrH && srcFilter->chrH->length > 1) ||
1088 (dstFilter->lumH && dstFilter->lumH->length > 1) ||
1089 (dstFilter->chrH && dstFilter->chrH->length > 1);
1091 av_pix_fmt_get_chroma_sub_sample(srcFormat, &c->chrSrcHSubSample, &c->chrSrcVSubSample);
1092 av_pix_fmt_get_chroma_sub_sample(dstFormat, &c->chrDstHSubSample, &c->chrDstVSubSample);
1094 if (isAnyRGB(dstFormat) && !(flags&SWS_FULL_CHR_H_INT)) {
1096 av_log(c, AV_LOG_DEBUG, "Forcing full internal H chroma due to odd output size\n");
1097 flags |= SWS_FULL_CHR_H_INT;
1101 if ( c->chrSrcHSubSample == 0
1102 && c->chrSrcVSubSample == 0
1103 && c->dither != SWS_DITHER_BAYER //SWS_FULL_CHR_H_INT is currently not supported with SWS_DITHER_BAYER
1104 && !(c->flags & SWS_FAST_BILINEAR)
1106 av_log(c, AV_LOG_DEBUG, "Forcing full internal H chroma due to input having non subsampled chroma\n");
1107 flags |= SWS_FULL_CHR_H_INT;
1112 if (c->dither == SWS_DITHER_AUTO) {
1113 if (flags & SWS_ERROR_DIFFUSION)
1114 c->dither = SWS_DITHER_ED;
1117 if(dstFormat == AV_PIX_FMT_BGR4_BYTE ||
1118 dstFormat == AV_PIX_FMT_RGB4_BYTE ||
1119 dstFormat == AV_PIX_FMT_BGR8 ||
1120 dstFormat == AV_PIX_FMT_RGB8) {
1121 if (c->dither == SWS_DITHER_AUTO)
1122 c->dither = (flags & SWS_FULL_CHR_H_INT) ? SWS_DITHER_ED : SWS_DITHER_BAYER;
1123 if (!(flags & SWS_FULL_CHR_H_INT)) {
1124 if (c->dither == SWS_DITHER_ED || c->dither == SWS_DITHER_A_DITHER || c->dither == SWS_DITHER_X_DITHER) {
1125 av_log(c, AV_LOG_DEBUG,
1126 "Desired dithering only supported in full chroma interpolation for destination format '%s'\n",
1127 av_get_pix_fmt_name(dstFormat));
1128 flags |= SWS_FULL_CHR_H_INT;
1132 if (flags & SWS_FULL_CHR_H_INT) {
1133 if (c->dither == SWS_DITHER_BAYER) {
1134 av_log(c, AV_LOG_DEBUG,
1135 "Ordered dither is not supported in full chroma interpolation for destination format '%s'\n",
1136 av_get_pix_fmt_name(dstFormat));
1137 c->dither = SWS_DITHER_ED;
1141 if (isPlanarRGB(dstFormat)) {
1142 if (!(flags & SWS_FULL_CHR_H_INT)) {
1143 av_log(c, AV_LOG_DEBUG,
1144 "%s output is not supported with half chroma resolution, switching to full\n",
1145 av_get_pix_fmt_name(dstFormat));
1146 flags |= SWS_FULL_CHR_H_INT;
1151 /* reuse chroma for 2 pixels RGB/BGR unless user wants full
1152 * chroma interpolation */
1153 if (flags & SWS_FULL_CHR_H_INT &&
1154 isAnyRGB(dstFormat) &&
1155 !isPlanarRGB(dstFormat) &&
1156 dstFormat != AV_PIX_FMT_RGBA &&
1157 dstFormat != AV_PIX_FMT_ARGB &&
1158 dstFormat != AV_PIX_FMT_BGRA &&
1159 dstFormat != AV_PIX_FMT_ABGR &&
1160 dstFormat != AV_PIX_FMT_RGB24 &&
1161 dstFormat != AV_PIX_FMT_BGR24 &&
1162 dstFormat != AV_PIX_FMT_BGR4_BYTE &&
1163 dstFormat != AV_PIX_FMT_RGB4_BYTE &&
1164 dstFormat != AV_PIX_FMT_BGR8 &&
1165 dstFormat != AV_PIX_FMT_RGB8
1167 av_log(c, AV_LOG_WARNING,
1168 "full chroma interpolation for destination format '%s' not yet implemented\n",
1169 av_get_pix_fmt_name(dstFormat));
1170 flags &= ~SWS_FULL_CHR_H_INT;
1173 if (isAnyRGB(dstFormat) && !(flags & SWS_FULL_CHR_H_INT))
1174 c->chrDstHSubSample = 1;
1176 // drop some chroma lines if the user wants it
1177 c->vChrDrop = (flags & SWS_SRC_V_CHR_DROP_MASK) >>
1178 SWS_SRC_V_CHR_DROP_SHIFT;
1179 c->chrSrcVSubSample += c->vChrDrop;
1181 /* drop every other pixel for chroma calculation unless user
1182 * wants full chroma */
1183 if (isAnyRGB(srcFormat) && !(flags & SWS_FULL_CHR_H_INP) &&
1184 srcFormat != AV_PIX_FMT_RGB8 && srcFormat != AV_PIX_FMT_BGR8 &&
1185 srcFormat != AV_PIX_FMT_RGB4 && srcFormat != AV_PIX_FMT_BGR4 &&
1186 srcFormat != AV_PIX_FMT_RGB4_BYTE && srcFormat != AV_PIX_FMT_BGR4_BYTE &&
1187 srcFormat != AV_PIX_FMT_GBRP9BE && srcFormat != AV_PIX_FMT_GBRP9LE &&
1188 srcFormat != AV_PIX_FMT_GBRP10BE && srcFormat != AV_PIX_FMT_GBRP10LE &&
1189 srcFormat != AV_PIX_FMT_GBRP12BE && srcFormat != AV_PIX_FMT_GBRP12LE &&
1190 srcFormat != AV_PIX_FMT_GBRP14BE && srcFormat != AV_PIX_FMT_GBRP14LE &&
1191 srcFormat != AV_PIX_FMT_GBRP16BE && srcFormat != AV_PIX_FMT_GBRP16LE &&
1192 ((dstW >> c->chrDstHSubSample) <= (srcW >> 1) ||
1193 (flags & SWS_FAST_BILINEAR)))
1194 c->chrSrcHSubSample = 1;
1196 // Note the FF_CEIL_RSHIFT is so that we always round toward +inf.
1197 c->chrSrcW = FF_CEIL_RSHIFT(srcW, c->chrSrcHSubSample);
1198 c->chrSrcH = FF_CEIL_RSHIFT(srcH, c->chrSrcVSubSample);
1199 c->chrDstW = FF_CEIL_RSHIFT(dstW, c->chrDstHSubSample);
1200 c->chrDstH = FF_CEIL_RSHIFT(dstH, c->chrDstVSubSample);
1202 FF_ALLOCZ_OR_GOTO(c, c->formatConvBuffer, FFALIGN(srcW*2+78, 16) * 2, fail);
1204 c->srcBpc = 1 + desc_src->comp[0].depth_minus1;
1207 c->dstBpc = 1 + desc_dst->comp[0].depth_minus1;
1210 if (isAnyRGB(srcFormat) || srcFormat == AV_PIX_FMT_PAL8)
1212 if (c->dstBpc == 16)
1215 if (INLINE_MMXEXT(cpu_flags) && c->srcBpc == 8 && c->dstBpc <= 14) {
1216 c->canMMXEXTBeUsed = dstW >= srcW && (dstW & 31) == 0 &&
1217 c->chrDstW >= c->chrSrcW &&
1219 if (!c->canMMXEXTBeUsed && dstW >= srcW && c->chrDstW >= c->chrSrcW && (srcW & 15) == 0
1221 && (flags & SWS_FAST_BILINEAR)) {
1222 if (flags & SWS_PRINT_INFO)
1223 av_log(c, AV_LOG_INFO,
1224 "output width is not a multiple of 32 -> no MMXEXT scaler\n");
1226 if (usesHFilter || isNBPS(c->srcFormat) || is16BPS(c->srcFormat) || isAnyRGB(c->srcFormat))
1227 c->canMMXEXTBeUsed = 0;
1229 c->canMMXEXTBeUsed = 0;
1231 c->chrXInc = (((int64_t)c->chrSrcW << 16) + (c->chrDstW >> 1)) / c->chrDstW;
1232 c->chrYInc = (((int64_t)c->chrSrcH << 16) + (c->chrDstH >> 1)) / c->chrDstH;
1234 /* Match pixel 0 of the src to pixel 0 of dst and match pixel n-2 of src
1235 * to pixel n-2 of dst, but only for the FAST_BILINEAR mode otherwise do
1237 * n-2 is the last chrominance sample available.
1238 * This is not perfect, but no one should notice the difference, the more
1239 * correct variant would be like the vertical one, but that would require
1240 * some special code for the first and last pixel */
1241 if (flags & SWS_FAST_BILINEAR) {
1242 if (c->canMMXEXTBeUsed) {
1246 // we don't use the x86 asm scaler if MMX is available
1247 else if (INLINE_MMX(cpu_flags) && c->dstBpc <= 14) {
1248 c->lumXInc = ((int64_t)(srcW - 2) << 16) / (dstW - 2) - 20;
1249 c->chrXInc = ((int64_t)(c->chrSrcW - 2) << 16) / (c->chrDstW - 2) - 20;
1253 // hardcoded for now
1254 c->gamma_value = 2.2;
1255 tmpFmt = AV_PIX_FMT_RGBA64LE;
1258 if (!unscaled && c->gamma_flag && (srcFormat != tmpFmt || dstFormat != tmpFmt)) {
1260 c->cascaded_context[0] = NULL;
1262 ret = av_image_alloc(c->cascaded_tmp, c->cascaded_tmpStride,
1263 srcW, srcH, tmpFmt, 64);
1267 c->cascaded_context[0] = sws_getContext(srcW, srcH, srcFormat,
1269 flags, NULL, NULL, c->param);
1270 if (!c->cascaded_context[0]) {
1274 c->cascaded_context[1] = sws_getContext(srcW, srcH, tmpFmt,
1276 flags, srcFilter, dstFilter, c->param);
1278 if (!c->cascaded_context[1])
1281 c2 = c->cascaded_context[1];
1282 c2->is_internal_gamma = 1;
1283 c2->gamma = alloc_gamma_tbl( c->gamma_value);
1284 c2->inv_gamma = alloc_gamma_tbl(1.f/c->gamma_value);
1285 if (!c2->gamma || !c2->inv_gamma)
1286 return AVERROR(ENOMEM);
1288 c->cascaded_context[2] = NULL;
1289 if (dstFormat != tmpFmt) {
1290 ret = av_image_alloc(c->cascaded1_tmp, c->cascaded1_tmpStride,
1291 dstW, dstH, tmpFmt, 64);
1295 c->cascaded_context[2] = sws_getContext(dstW, dstH, tmpFmt,
1296 dstW, dstH, dstFormat,
1297 flags, NULL, NULL, c->param);
1298 if (!c->cascaded_context[2])
1304 if (isBayer(srcFormat)) {
1306 (dstFormat != AV_PIX_FMT_RGB24 && dstFormat != AV_PIX_FMT_YUV420P)) {
1307 enum AVPixelFormat tmpFormat = AV_PIX_FMT_RGB24;
1309 ret = av_image_alloc(c->cascaded_tmp, c->cascaded_tmpStride,
1310 srcW, srcH, tmpFormat, 64);
1314 c->cascaded_context[0] = sws_getContext(srcW, srcH, srcFormat,
1315 srcW, srcH, tmpFormat,
1316 flags, srcFilter, NULL, c->param);
1317 if (!c->cascaded_context[0])
1320 c->cascaded_context[1] = sws_getContext(srcW, srcH, tmpFormat,
1321 dstW, dstH, dstFormat,
1322 flags, NULL, dstFilter, c->param);
1323 if (!c->cascaded_context[1])
1329 #define USE_MMAP (HAVE_MMAP && HAVE_MPROTECT && defined MAP_ANONYMOUS)
1331 /* precalculate horizontal scaler filter coefficients */
1333 #if HAVE_MMXEXT_INLINE
1334 // can't downscale !!!
1335 if (c->canMMXEXTBeUsed && (flags & SWS_FAST_BILINEAR)) {
1336 c->lumMmxextFilterCodeSize = ff_init_hscaler_mmxext(dstW, c->lumXInc, NULL,
1338 c->chrMmxextFilterCodeSize = ff_init_hscaler_mmxext(c->chrDstW, c->chrXInc,
1339 NULL, NULL, NULL, 4);
1342 c->lumMmxextFilterCode = mmap(NULL, c->lumMmxextFilterCodeSize,
1343 PROT_READ | PROT_WRITE,
1344 MAP_PRIVATE | MAP_ANONYMOUS,
1346 c->chrMmxextFilterCode = mmap(NULL, c->chrMmxextFilterCodeSize,
1347 PROT_READ | PROT_WRITE,
1348 MAP_PRIVATE | MAP_ANONYMOUS,
1350 #elif HAVE_VIRTUALALLOC
1351 c->lumMmxextFilterCode = VirtualAlloc(NULL,
1352 c->lumMmxextFilterCodeSize,
1354 PAGE_EXECUTE_READWRITE);
1355 c->chrMmxextFilterCode = VirtualAlloc(NULL,
1356 c->chrMmxextFilterCodeSize,
1358 PAGE_EXECUTE_READWRITE);
1360 c->lumMmxextFilterCode = av_malloc(c->lumMmxextFilterCodeSize);
1361 c->chrMmxextFilterCode = av_malloc(c->chrMmxextFilterCodeSize);
1364 #ifdef MAP_ANONYMOUS
1365 if (c->lumMmxextFilterCode == MAP_FAILED || c->chrMmxextFilterCode == MAP_FAILED)
1367 if (!c->lumMmxextFilterCode || !c->chrMmxextFilterCode)
1370 av_log(c, AV_LOG_ERROR, "Failed to allocate MMX2FilterCode\n");
1371 return AVERROR(ENOMEM);
1374 FF_ALLOCZ_OR_GOTO(c, c->hLumFilter, (dstW / 8 + 8) * sizeof(int16_t), fail);
1375 FF_ALLOCZ_OR_GOTO(c, c->hChrFilter, (c->chrDstW / 4 + 8) * sizeof(int16_t), fail);
1376 FF_ALLOCZ_OR_GOTO(c, c->hLumFilterPos, (dstW / 2 / 8 + 8) * sizeof(int32_t), fail);
1377 FF_ALLOCZ_OR_GOTO(c, c->hChrFilterPos, (c->chrDstW / 2 / 4 + 8) * sizeof(int32_t), fail);
1379 ff_init_hscaler_mmxext( dstW, c->lumXInc, c->lumMmxextFilterCode,
1380 c->hLumFilter, (uint32_t*)c->hLumFilterPos, 8);
1381 ff_init_hscaler_mmxext(c->chrDstW, c->chrXInc, c->chrMmxextFilterCode,
1382 c->hChrFilter, (uint32_t*)c->hChrFilterPos, 4);
1385 if ( mprotect(c->lumMmxextFilterCode, c->lumMmxextFilterCodeSize, PROT_EXEC | PROT_READ) == -1
1386 || mprotect(c->chrMmxextFilterCode, c->chrMmxextFilterCodeSize, PROT_EXEC | PROT_READ) == -1) {
1387 av_log(c, AV_LOG_ERROR, "mprotect failed, cannot use fast bilinear scaler\n");
1392 #endif /* HAVE_MMXEXT_INLINE */
1394 const int filterAlign = X86_MMX(cpu_flags) ? 4 :
1395 PPC_ALTIVEC(cpu_flags) ? 8 : 1;
1397 if ((ret = initFilter(&c->hLumFilter, &c->hLumFilterPos,
1398 &c->hLumFilterSize, c->lumXInc,
1399 srcW, dstW, filterAlign, 1 << 14,
1400 (flags & SWS_BICUBLIN) ? (flags | SWS_BICUBIC) : flags,
1401 cpu_flags, srcFilter->lumH, dstFilter->lumH,
1403 get_local_pos(c, 0, 0, 0),
1404 get_local_pos(c, 0, 0, 0))) < 0)
1406 if ((ret = initFilter(&c->hChrFilter, &c->hChrFilterPos,
1407 &c->hChrFilterSize, c->chrXInc,
1408 c->chrSrcW, c->chrDstW, filterAlign, 1 << 14,
1409 (flags & SWS_BICUBLIN) ? (flags | SWS_BILINEAR) : flags,
1410 cpu_flags, srcFilter->chrH, dstFilter->chrH,
1412 get_local_pos(c, c->chrSrcHSubSample, c->src_h_chr_pos, 0),
1413 get_local_pos(c, c->chrDstHSubSample, c->dst_h_chr_pos, 0))) < 0)
1416 } // initialize horizontal stuff
1418 /* precalculate vertical scaler filter coefficients */
1420 const int filterAlign = X86_MMX(cpu_flags) ? 2 :
1421 PPC_ALTIVEC(cpu_flags) ? 8 : 1;
1423 if ((ret = initFilter(&c->vLumFilter, &c->vLumFilterPos, &c->vLumFilterSize,
1424 c->lumYInc, srcH, dstH, filterAlign, (1 << 12),
1425 (flags & SWS_BICUBLIN) ? (flags | SWS_BICUBIC) : flags,
1426 cpu_flags, srcFilter->lumV, dstFilter->lumV,
1428 get_local_pos(c, 0, 0, 1),
1429 get_local_pos(c, 0, 0, 1))) < 0)
1431 if ((ret = initFilter(&c->vChrFilter, &c->vChrFilterPos, &c->vChrFilterSize,
1432 c->chrYInc, c->chrSrcH, c->chrDstH,
1433 filterAlign, (1 << 12),
1434 (flags & SWS_BICUBLIN) ? (flags | SWS_BILINEAR) : flags,
1435 cpu_flags, srcFilter->chrV, dstFilter->chrV,
1437 get_local_pos(c, c->chrSrcVSubSample, c->src_v_chr_pos, 1),
1438 get_local_pos(c, c->chrDstVSubSample, c->dst_v_chr_pos, 1))) < 0)
1443 FF_ALLOC_OR_GOTO(c, c->vYCoeffsBank, sizeof(vector signed short) * c->vLumFilterSize * c->dstH, fail);
1444 FF_ALLOC_OR_GOTO(c, c->vCCoeffsBank, sizeof(vector signed short) * c->vChrFilterSize * c->chrDstH, fail);
1446 for (i = 0; i < c->vLumFilterSize * c->dstH; i++) {
1448 short *p = (short *)&c->vYCoeffsBank[i];
1449 for (j = 0; j < 8; j++)
1450 p[j] = c->vLumFilter[i];
1453 for (i = 0; i < c->vChrFilterSize * c->chrDstH; i++) {
1455 short *p = (short *)&c->vCCoeffsBank[i];
1456 for (j = 0; j < 8; j++)
1457 p[j] = c->vChrFilter[i];
1462 // calculate buffer sizes so that they won't run out while handling these damn slices
1463 c->vLumBufSize = c->vLumFilterSize;
1464 c->vChrBufSize = c->vChrFilterSize;
1465 for (i = 0; i < dstH; i++) {
1466 int chrI = (int64_t)i * c->chrDstH / dstH;
1467 int nextSlice = FFMAX(c->vLumFilterPos[i] + c->vLumFilterSize - 1,
1468 ((c->vChrFilterPos[chrI] + c->vChrFilterSize - 1)
1469 << c->chrSrcVSubSample));
1471 nextSlice >>= c->chrSrcVSubSample;
1472 nextSlice <<= c->chrSrcVSubSample;
1473 if (c->vLumFilterPos[i] + c->vLumBufSize < nextSlice)
1474 c->vLumBufSize = nextSlice - c->vLumFilterPos[i];
1475 if (c->vChrFilterPos[chrI] + c->vChrBufSize <
1476 (nextSlice >> c->chrSrcVSubSample))
1477 c->vChrBufSize = (nextSlice >> c->chrSrcVSubSample) -
1478 c->vChrFilterPos[chrI];
1481 for (i = 0; i < 4; i++)
1482 FF_ALLOCZ_OR_GOTO(c, c->dither_error[i], (c->dstW+2) * sizeof(int), fail);
1484 /* Allocate pixbufs (we use dynamic allocation because otherwise we would
1485 * need to allocate several megabytes to handle all possible cases) */
1486 FF_ALLOC_OR_GOTO(c, c->lumPixBuf, c->vLumBufSize * 3 * sizeof(int16_t *), fail);
1487 FF_ALLOC_OR_GOTO(c, c->chrUPixBuf, c->vChrBufSize * 3 * sizeof(int16_t *), fail);
1488 FF_ALLOC_OR_GOTO(c, c->chrVPixBuf, c->vChrBufSize * 3 * sizeof(int16_t *), fail);
1489 if (CONFIG_SWSCALE_ALPHA && isALPHA(c->srcFormat) && isALPHA(c->dstFormat))
1490 FF_ALLOCZ_OR_GOTO(c, c->alpPixBuf, c->vLumBufSize * 3 * sizeof(int16_t *), fail);
1491 /* Note we need at least one pixel more at the end because of the MMX code
1492 * (just in case someone wants to replace the 4000/8000). */
1493 /* align at 16 bytes for AltiVec */
1494 for (i = 0; i < c->vLumBufSize; i++) {
1495 FF_ALLOCZ_OR_GOTO(c, c->lumPixBuf[i + c->vLumBufSize],
1496 dst_stride + 16, fail);
1497 c->lumPixBuf[i] = c->lumPixBuf[i + c->vLumBufSize];
1499 // 64 / c->scalingBpp is the same as 16 / sizeof(scaling_intermediate)
1500 c->uv_off = (dst_stride>>1) + 64 / (c->dstBpc &~ 7);
1501 c->uv_offx2 = dst_stride + 16;
1502 for (i = 0; i < c->vChrBufSize; i++) {
1503 FF_ALLOC_OR_GOTO(c, c->chrUPixBuf[i + c->vChrBufSize],
1504 dst_stride * 2 + 32, fail);
1505 c->chrUPixBuf[i] = c->chrUPixBuf[i + c->vChrBufSize];
1506 c->chrVPixBuf[i] = c->chrVPixBuf[i + c->vChrBufSize]
1507 = c->chrUPixBuf[i] + (dst_stride >> 1) + 8;
1509 if (CONFIG_SWSCALE_ALPHA && c->alpPixBuf)
1510 for (i = 0; i < c->vLumBufSize; i++) {
1511 FF_ALLOCZ_OR_GOTO(c, c->alpPixBuf[i + c->vLumBufSize],
1512 dst_stride + 16, fail);
1513 c->alpPixBuf[i] = c->alpPixBuf[i + c->vLumBufSize];
1516 // try to avoid drawing green stuff between the right end and the stride end
1517 for (i = 0; i < c->vChrBufSize; i++)
1518 if(desc_dst->comp[0].depth_minus1 == 15){
1519 av_assert0(c->dstBpc > 14);
1520 for(j=0; j<dst_stride/2+1; j++)
1521 ((int32_t*)(c->chrUPixBuf[i]))[j] = 1<<18;
1523 for(j=0; j<dst_stride+1; j++)
1524 ((int16_t*)(c->chrUPixBuf[i]))[j] = 1<<14;
1526 av_assert0(c->chrDstH <= dstH);
1528 if (flags & SWS_PRINT_INFO) {
1529 const char *scaler = NULL, *cpucaps;
1531 for (i = 0; i < FF_ARRAY_ELEMS(scale_algorithms); i++) {
1532 if (flags & scale_algorithms[i].flag) {
1533 scaler = scale_algorithms[i].description;
1538 scaler = "ehh flags invalid?!";
1539 av_log(c, AV_LOG_INFO, "%s scaler, from %s to %s%s ",
1541 av_get_pix_fmt_name(srcFormat),
1543 dstFormat == AV_PIX_FMT_BGR555 || dstFormat == AV_PIX_FMT_BGR565 ||
1544 dstFormat == AV_PIX_FMT_RGB444BE || dstFormat == AV_PIX_FMT_RGB444LE ||
1545 dstFormat == AV_PIX_FMT_BGR444BE || dstFormat == AV_PIX_FMT_BGR444LE ?
1550 av_get_pix_fmt_name(dstFormat));
1552 if (INLINE_MMXEXT(cpu_flags))
1554 else if (INLINE_AMD3DNOW(cpu_flags))
1556 else if (INLINE_MMX(cpu_flags))
1558 else if (PPC_ALTIVEC(cpu_flags))
1559 cpucaps = "AltiVec";
1563 av_log(c, AV_LOG_INFO, "using %s\n", cpucaps);
1565 av_log(c, AV_LOG_VERBOSE, "%dx%d -> %dx%d\n", srcW, srcH, dstW, dstH);
1566 av_log(c, AV_LOG_DEBUG,
1567 "lum srcW=%d srcH=%d dstW=%d dstH=%d xInc=%d yInc=%d\n",
1568 c->srcW, c->srcH, c->dstW, c->dstH, c->lumXInc, c->lumYInc);
1569 av_log(c, AV_LOG_DEBUG,
1570 "chr srcW=%d srcH=%d dstW=%d dstH=%d xInc=%d yInc=%d\n",
1571 c->chrSrcW, c->chrSrcH, c->chrDstW, c->chrDstH,
1572 c->chrXInc, c->chrYInc);
1575 /* unscaled special cases */
1576 if (unscaled && !usesHFilter && !usesVFilter &&
1577 (c->srcRange == c->dstRange || isAnyRGB(dstFormat))) {
1578 ff_get_unscaled_swscale(c);
1581 if (flags & SWS_PRINT_INFO)
1582 av_log(c, AV_LOG_INFO,
1583 "using unscaled %s -> %s special converter\n",
1584 av_get_pix_fmt_name(srcFormat), av_get_pix_fmt_name(dstFormat));
1589 c->swscale = ff_getSwsFunc(c);
1591 fail: // FIXME replace things by appropriate error codes
1592 if (ret == RETCODE_USE_CASCADE) {
1593 int tmpW = sqrt(srcW * (int64_t)dstW);
1594 int tmpH = sqrt(srcH * (int64_t)dstH);
1595 enum AVPixelFormat tmpFormat = AV_PIX_FMT_YUV420P;
1597 if (srcW*(int64_t)srcH <= 4LL*dstW*dstH)
1598 return AVERROR(EINVAL);
1600 ret = av_image_alloc(c->cascaded_tmp, c->cascaded_tmpStride,
1601 tmpW, tmpH, tmpFormat, 64);
1605 c->cascaded_context[0] = sws_getContext(srcW, srcH, srcFormat,
1606 tmpW, tmpH, tmpFormat,
1607 flags, srcFilter, NULL, c->param);
1608 if (!c->cascaded_context[0])
1611 c->cascaded_context[1] = sws_getContext(tmpW, tmpH, tmpFormat,
1612 dstW, dstH, dstFormat,
1613 flags, NULL, dstFilter, c->param);
1614 if (!c->cascaded_context[1])
1621 SwsContext *sws_getContext(int srcW, int srcH, enum AVPixelFormat srcFormat,
1622 int dstW, int dstH, enum AVPixelFormat dstFormat,
1623 int flags, SwsFilter *srcFilter,
1624 SwsFilter *dstFilter, const double *param)
1628 if (!(c = sws_alloc_context()))
1636 c->srcFormat = srcFormat;
1637 c->dstFormat = dstFormat;
1640 c->param[0] = param[0];
1641 c->param[1] = param[1];
1644 if (sws_init_context(c, srcFilter, dstFilter) < 0) {
1652 SwsFilter *sws_getDefaultFilter(float lumaGBlur, float chromaGBlur,
1653 float lumaSharpen, float chromaSharpen,
1654 float chromaHShift, float chromaVShift,
1657 SwsFilter *filter = av_malloc(sizeof(SwsFilter));
1661 if (lumaGBlur != 0.0) {
1662 filter->lumH = sws_getGaussianVec(lumaGBlur, 3.0);
1663 filter->lumV = sws_getGaussianVec(lumaGBlur, 3.0);
1665 filter->lumH = sws_getIdentityVec();
1666 filter->lumV = sws_getIdentityVec();
1669 if (chromaGBlur != 0.0) {
1670 filter->chrH = sws_getGaussianVec(chromaGBlur, 3.0);
1671 filter->chrV = sws_getGaussianVec(chromaGBlur, 3.0);
1673 filter->chrH = sws_getIdentityVec();
1674 filter->chrV = sws_getIdentityVec();
1677 if (!filter->lumH || !filter->lumV || !filter->chrH || !filter->chrV)
1680 if (chromaSharpen != 0.0) {
1681 SwsVector *id = sws_getIdentityVec();
1684 sws_scaleVec(filter->chrH, -chromaSharpen);
1685 sws_scaleVec(filter->chrV, -chromaSharpen);
1686 sws_addVec(filter->chrH, id);
1687 sws_addVec(filter->chrV, id);
1691 if (lumaSharpen != 0.0) {
1692 SwsVector *id = sws_getIdentityVec();
1695 sws_scaleVec(filter->lumH, -lumaSharpen);
1696 sws_scaleVec(filter->lumV, -lumaSharpen);
1697 sws_addVec(filter->lumH, id);
1698 sws_addVec(filter->lumV, id);
1702 if (chromaHShift != 0.0)
1703 sws_shiftVec(filter->chrH, (int)(chromaHShift + 0.5));
1705 if (chromaVShift != 0.0)
1706 sws_shiftVec(filter->chrV, (int)(chromaVShift + 0.5));
1708 sws_normalizeVec(filter->chrH, 1.0);
1709 sws_normalizeVec(filter->chrV, 1.0);
1710 sws_normalizeVec(filter->lumH, 1.0);
1711 sws_normalizeVec(filter->lumV, 1.0);
1714 sws_printVec2(filter->chrH, NULL, AV_LOG_DEBUG);
1716 sws_printVec2(filter->lumH, NULL, AV_LOG_DEBUG);
1721 sws_freeVec(filter->lumH);
1722 sws_freeVec(filter->lumV);
1723 sws_freeVec(filter->chrH);
1724 sws_freeVec(filter->chrV);
1729 SwsVector *sws_allocVec(int length)
1733 if(length <= 0 || length > INT_MAX/ sizeof(double))
1736 vec = av_malloc(sizeof(SwsVector));
1739 vec->length = length;
1740 vec->coeff = av_malloc(sizeof(double) * length);
1746 SwsVector *sws_getGaussianVec(double variance, double quality)
1748 const int length = (int)(variance * quality + 0.5) | 1;
1750 double middle = (length - 1) * 0.5;
1753 if(variance < 0 || quality < 0)
1756 vec = sws_allocVec(length);
1761 for (i = 0; i < length; i++) {
1762 double dist = i - middle;
1763 vec->coeff[i] = exp(-dist * dist / (2 * variance * variance)) /
1764 sqrt(2 * variance * M_PI);
1767 sws_normalizeVec(vec, 1.0);
1772 SwsVector *sws_getConstVec(double c, int length)
1775 SwsVector *vec = sws_allocVec(length);
1780 for (i = 0; i < length; i++)
1786 SwsVector *sws_getIdentityVec(void)
1788 return sws_getConstVec(1.0, 1);
1791 static double sws_dcVec(SwsVector *a)
1796 for (i = 0; i < a->length; i++)
1802 void sws_scaleVec(SwsVector *a, double scalar)
1806 for (i = 0; i < a->length; i++)
1807 a->coeff[i] *= scalar;
1810 void sws_normalizeVec(SwsVector *a, double height)
1812 sws_scaleVec(a, height / sws_dcVec(a));
1815 static SwsVector *sws_getConvVec(SwsVector *a, SwsVector *b)
1817 int length = a->length + b->length - 1;
1819 SwsVector *vec = sws_getConstVec(0.0, length);
1824 for (i = 0; i < a->length; i++) {
1825 for (j = 0; j < b->length; j++) {
1826 vec->coeff[i + j] += a->coeff[i] * b->coeff[j];
1833 static SwsVector *sws_sumVec(SwsVector *a, SwsVector *b)
1835 int length = FFMAX(a->length, b->length);
1837 SwsVector *vec = sws_getConstVec(0.0, length);
1842 for (i = 0; i < a->length; i++)
1843 vec->coeff[i + (length - 1) / 2 - (a->length - 1) / 2] += a->coeff[i];
1844 for (i = 0; i < b->length; i++)
1845 vec->coeff[i + (length - 1) / 2 - (b->length - 1) / 2] += b->coeff[i];
1850 static SwsVector *sws_diffVec(SwsVector *a, SwsVector *b)
1852 int length = FFMAX(a->length, b->length);
1854 SwsVector *vec = sws_getConstVec(0.0, length);
1859 for (i = 0; i < a->length; i++)
1860 vec->coeff[i + (length - 1) / 2 - (a->length - 1) / 2] += a->coeff[i];
1861 for (i = 0; i < b->length; i++)
1862 vec->coeff[i + (length - 1) / 2 - (b->length - 1) / 2] -= b->coeff[i];
1867 /* shift left / or right if "shift" is negative */
1868 static SwsVector *sws_getShiftedVec(SwsVector *a, int shift)
1870 int length = a->length + FFABS(shift) * 2;
1872 SwsVector *vec = sws_getConstVec(0.0, length);
1877 for (i = 0; i < a->length; i++) {
1878 vec->coeff[i + (length - 1) / 2 -
1879 (a->length - 1) / 2 - shift] = a->coeff[i];
1885 void sws_shiftVec(SwsVector *a, int shift)
1887 SwsVector *shifted = sws_getShiftedVec(a, shift);
1889 a->coeff = shifted->coeff;
1890 a->length = shifted->length;
1894 void sws_addVec(SwsVector *a, SwsVector *b)
1896 SwsVector *sum = sws_sumVec(a, b);
1898 a->coeff = sum->coeff;
1899 a->length = sum->length;
1903 void sws_subVec(SwsVector *a, SwsVector *b)
1905 SwsVector *diff = sws_diffVec(a, b);
1907 a->coeff = diff->coeff;
1908 a->length = diff->length;
1912 void sws_convVec(SwsVector *a, SwsVector *b)
1914 SwsVector *conv = sws_getConvVec(a, b);
1916 a->coeff = conv->coeff;
1917 a->length = conv->length;
1921 SwsVector *sws_cloneVec(SwsVector *a)
1923 SwsVector *vec = sws_allocVec(a->length);
1928 memcpy(vec->coeff, a->coeff, a->length * sizeof(*a->coeff));
1933 void sws_printVec2(SwsVector *a, AVClass *log_ctx, int log_level)
1940 for (i = 0; i < a->length; i++)
1941 if (a->coeff[i] > max)
1944 for (i = 0; i < a->length; i++)
1945 if (a->coeff[i] < min)
1950 for (i = 0; i < a->length; i++) {
1951 int x = (int)((a->coeff[i] - min) * 60.0 / range + 0.5);
1952 av_log(log_ctx, log_level, "%1.3f ", a->coeff[i]);
1954 av_log(log_ctx, log_level, " ");
1955 av_log(log_ctx, log_level, "|\n");
1959 void sws_freeVec(SwsVector *a)
1963 av_freep(&a->coeff);
1968 void sws_freeFilter(SwsFilter *filter)
1973 sws_freeVec(filter->lumH);
1974 sws_freeVec(filter->lumV);
1975 sws_freeVec(filter->chrH);
1976 sws_freeVec(filter->chrV);
1980 void sws_freeContext(SwsContext *c)
1987 for (i = 0; i < c->vLumBufSize; i++)
1988 av_freep(&c->lumPixBuf[i]);
1989 av_freep(&c->lumPixBuf);
1992 if (c->chrUPixBuf) {
1993 for (i = 0; i < c->vChrBufSize; i++)
1994 av_freep(&c->chrUPixBuf[i]);
1995 av_freep(&c->chrUPixBuf);
1996 av_freep(&c->chrVPixBuf);
1999 if (CONFIG_SWSCALE_ALPHA && c->alpPixBuf) {
2000 for (i = 0; i < c->vLumBufSize; i++)
2001 av_freep(&c->alpPixBuf[i]);
2002 av_freep(&c->alpPixBuf);
2005 for (i = 0; i < 4; i++)
2006 av_freep(&c->dither_error[i]);
2008 av_freep(&c->vLumFilter);
2009 av_freep(&c->vChrFilter);
2010 av_freep(&c->hLumFilter);
2011 av_freep(&c->hChrFilter);
2013 av_freep(&c->vYCoeffsBank);
2014 av_freep(&c->vCCoeffsBank);
2017 av_freep(&c->vLumFilterPos);
2018 av_freep(&c->vChrFilterPos);
2019 av_freep(&c->hLumFilterPos);
2020 av_freep(&c->hChrFilterPos);
2024 if (c->lumMmxextFilterCode)
2025 munmap(c->lumMmxextFilterCode, c->lumMmxextFilterCodeSize);
2026 if (c->chrMmxextFilterCode)
2027 munmap(c->chrMmxextFilterCode, c->chrMmxextFilterCodeSize);
2028 #elif HAVE_VIRTUALALLOC
2029 if (c->lumMmxextFilterCode)
2030 VirtualFree(c->lumMmxextFilterCode, 0, MEM_RELEASE);
2031 if (c->chrMmxextFilterCode)
2032 VirtualFree(c->chrMmxextFilterCode, 0, MEM_RELEASE);
2034 av_free(c->lumMmxextFilterCode);
2035 av_free(c->chrMmxextFilterCode);
2037 c->lumMmxextFilterCode = NULL;
2038 c->chrMmxextFilterCode = NULL;
2039 #endif /* HAVE_MMX_INLINE */
2041 av_freep(&c->yuvTable);
2042 av_freep(&c->formatConvBuffer);
2044 sws_freeContext(c->cascaded_context[0]);
2045 sws_freeContext(c->cascaded_context[1]);
2046 sws_freeContext(c->cascaded_context[2]);
2047 memset(c->cascaded_context, 0, sizeof(c->cascaded_context));
2048 av_freep(&c->cascaded_tmp[0]);
2049 av_freep(&c->cascaded1_tmp[0]);
2051 av_freep(&c->gamma);
2052 av_freep(&c->inv_gamma);
2058 struct SwsContext *sws_getCachedContext(struct SwsContext *context, int srcW,
2059 int srcH, enum AVPixelFormat srcFormat,
2061 enum AVPixelFormat dstFormat, int flags,
2062 SwsFilter *srcFilter,
2063 SwsFilter *dstFilter,
2064 const double *param)
2066 static const double default_param[2] = { SWS_PARAM_DEFAULT,
2067 SWS_PARAM_DEFAULT };
2068 int64_t src_h_chr_pos = -513, dst_h_chr_pos = -513,
2069 src_v_chr_pos = -513, dst_v_chr_pos = -513;
2072 param = default_param;
2075 (context->srcW != srcW ||
2076 context->srcH != srcH ||
2077 context->srcFormat != srcFormat ||
2078 context->dstW != dstW ||
2079 context->dstH != dstH ||
2080 context->dstFormat != dstFormat ||
2081 context->flags != flags ||
2082 context->param[0] != param[0] ||
2083 context->param[1] != param[1])) {
2085 av_opt_get_int(context, "src_h_chr_pos", 0, &src_h_chr_pos);
2086 av_opt_get_int(context, "src_v_chr_pos", 0, &src_v_chr_pos);
2087 av_opt_get_int(context, "dst_h_chr_pos", 0, &dst_h_chr_pos);
2088 av_opt_get_int(context, "dst_v_chr_pos", 0, &dst_v_chr_pos);
2089 sws_freeContext(context);
2094 if (!(context = sws_alloc_context()))
2096 context->srcW = srcW;
2097 context->srcH = srcH;
2098 context->srcFormat = srcFormat;
2099 context->dstW = dstW;
2100 context->dstH = dstH;
2101 context->dstFormat = dstFormat;
2102 context->flags = flags;
2103 context->param[0] = param[0];
2104 context->param[1] = param[1];
2106 av_opt_set_int(context, "src_h_chr_pos", src_h_chr_pos, 0);
2107 av_opt_set_int(context, "src_v_chr_pos", src_v_chr_pos, 0);
2108 av_opt_set_int(context, "dst_h_chr_pos", dst_h_chr_pos, 0);
2109 av_opt_set_int(context, "dst_v_chr_pos", dst_v_chr_pos, 0);
2111 if (sws_init_context(context, srcFilter, dstFilter) < 0) {
2112 sws_freeContext(context);