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_YUV440P10LE] = { 1, 1 },
123 [AV_PIX_FMT_YUV440P10BE] = { 1, 1 },
124 [AV_PIX_FMT_YUV440P12LE] = { 1, 1 },
125 [AV_PIX_FMT_YUV440P12BE] = { 1, 1 },
126 [AV_PIX_FMT_YUVA420P] = { 1, 1 },
127 [AV_PIX_FMT_YUVA422P] = { 1, 1 },
128 [AV_PIX_FMT_YUVA444P] = { 1, 1 },
129 [AV_PIX_FMT_YUVA420P9BE] = { 1, 1 },
130 [AV_PIX_FMT_YUVA420P9LE] = { 1, 1 },
131 [AV_PIX_FMT_YUVA422P9BE] = { 1, 1 },
132 [AV_PIX_FMT_YUVA422P9LE] = { 1, 1 },
133 [AV_PIX_FMT_YUVA444P9BE] = { 1, 1 },
134 [AV_PIX_FMT_YUVA444P9LE] = { 1, 1 },
135 [AV_PIX_FMT_YUVA420P10BE]= { 1, 1 },
136 [AV_PIX_FMT_YUVA420P10LE]= { 1, 1 },
137 [AV_PIX_FMT_YUVA422P10BE]= { 1, 1 },
138 [AV_PIX_FMT_YUVA422P10LE]= { 1, 1 },
139 [AV_PIX_FMT_YUVA444P10BE]= { 1, 1 },
140 [AV_PIX_FMT_YUVA444P10LE]= { 1, 1 },
141 [AV_PIX_FMT_YUVA420P16BE]= { 1, 1 },
142 [AV_PIX_FMT_YUVA420P16LE]= { 1, 1 },
143 [AV_PIX_FMT_YUVA422P16BE]= { 1, 1 },
144 [AV_PIX_FMT_YUVA422P16LE]= { 1, 1 },
145 [AV_PIX_FMT_YUVA444P16BE]= { 1, 1 },
146 [AV_PIX_FMT_YUVA444P16LE]= { 1, 1 },
147 [AV_PIX_FMT_RGB48BE] = { 1, 1 },
148 [AV_PIX_FMT_RGB48LE] = { 1, 1 },
149 [AV_PIX_FMT_RGBA64BE] = { 1, 1, 1 },
150 [AV_PIX_FMT_RGBA64LE] = { 1, 1, 1 },
151 [AV_PIX_FMT_RGB565BE] = { 1, 1 },
152 [AV_PIX_FMT_RGB565LE] = { 1, 1 },
153 [AV_PIX_FMT_RGB555BE] = { 1, 1 },
154 [AV_PIX_FMT_RGB555LE] = { 1, 1 },
155 [AV_PIX_FMT_BGR565BE] = { 1, 1 },
156 [AV_PIX_FMT_BGR565LE] = { 1, 1 },
157 [AV_PIX_FMT_BGR555BE] = { 1, 1 },
158 [AV_PIX_FMT_BGR555LE] = { 1, 1 },
159 [AV_PIX_FMT_YUV420P16LE] = { 1, 1 },
160 [AV_PIX_FMT_YUV420P16BE] = { 1, 1 },
161 [AV_PIX_FMT_YUV422P16LE] = { 1, 1 },
162 [AV_PIX_FMT_YUV422P16BE] = { 1, 1 },
163 [AV_PIX_FMT_YUV444P16LE] = { 1, 1 },
164 [AV_PIX_FMT_YUV444P16BE] = { 1, 1 },
165 [AV_PIX_FMT_RGB444LE] = { 1, 1 },
166 [AV_PIX_FMT_RGB444BE] = { 1, 1 },
167 [AV_PIX_FMT_BGR444LE] = { 1, 1 },
168 [AV_PIX_FMT_BGR444BE] = { 1, 1 },
169 [AV_PIX_FMT_YA8] = { 1, 1 },
170 [AV_PIX_FMT_YA16BE] = { 1, 0 },
171 [AV_PIX_FMT_YA16LE] = { 1, 0 },
172 [AV_PIX_FMT_BGR48BE] = { 1, 1 },
173 [AV_PIX_FMT_BGR48LE] = { 1, 1 },
174 [AV_PIX_FMT_BGRA64BE] = { 1, 1, 1 },
175 [AV_PIX_FMT_BGRA64LE] = { 1, 1, 1 },
176 [AV_PIX_FMT_YUV420P9BE] = { 1, 1 },
177 [AV_PIX_FMT_YUV420P9LE] = { 1, 1 },
178 [AV_PIX_FMT_YUV420P10BE] = { 1, 1 },
179 [AV_PIX_FMT_YUV420P10LE] = { 1, 1 },
180 [AV_PIX_FMT_YUV420P12BE] = { 1, 1 },
181 [AV_PIX_FMT_YUV420P12LE] = { 1, 1 },
182 [AV_PIX_FMT_YUV420P14BE] = { 1, 1 },
183 [AV_PIX_FMT_YUV420P14LE] = { 1, 1 },
184 [AV_PIX_FMT_YUV422P9BE] = { 1, 1 },
185 [AV_PIX_FMT_YUV422P9LE] = { 1, 1 },
186 [AV_PIX_FMT_YUV422P10BE] = { 1, 1 },
187 [AV_PIX_FMT_YUV422P10LE] = { 1, 1 },
188 [AV_PIX_FMT_YUV422P12BE] = { 1, 1 },
189 [AV_PIX_FMT_YUV422P12LE] = { 1, 1 },
190 [AV_PIX_FMT_YUV422P14BE] = { 1, 1 },
191 [AV_PIX_FMT_YUV422P14LE] = { 1, 1 },
192 [AV_PIX_FMT_YUV444P9BE] = { 1, 1 },
193 [AV_PIX_FMT_YUV444P9LE] = { 1, 1 },
194 [AV_PIX_FMT_YUV444P10BE] = { 1, 1 },
195 [AV_PIX_FMT_YUV444P10LE] = { 1, 1 },
196 [AV_PIX_FMT_YUV444P12BE] = { 1, 1 },
197 [AV_PIX_FMT_YUV444P12LE] = { 1, 1 },
198 [AV_PIX_FMT_YUV444P14BE] = { 1, 1 },
199 [AV_PIX_FMT_YUV444P14LE] = { 1, 1 },
200 [AV_PIX_FMT_GBRP] = { 1, 1 },
201 [AV_PIX_FMT_GBRP9LE] = { 1, 1 },
202 [AV_PIX_FMT_GBRP9BE] = { 1, 1 },
203 [AV_PIX_FMT_GBRP10LE] = { 1, 1 },
204 [AV_PIX_FMT_GBRP10BE] = { 1, 1 },
205 [AV_PIX_FMT_GBRP12LE] = { 1, 1 },
206 [AV_PIX_FMT_GBRP12BE] = { 1, 1 },
207 [AV_PIX_FMT_GBRP14LE] = { 1, 1 },
208 [AV_PIX_FMT_GBRP14BE] = { 1, 1 },
209 [AV_PIX_FMT_GBRP16LE] = { 1, 0 },
210 [AV_PIX_FMT_GBRP16BE] = { 1, 0 },
211 [AV_PIX_FMT_GBRAP] = { 1, 1 },
212 [AV_PIX_FMT_GBRAP16LE] = { 1, 0 },
213 [AV_PIX_FMT_GBRAP16BE] = { 1, 0 },
214 [AV_PIX_FMT_BAYER_BGGR8] = { 1, 0 },
215 [AV_PIX_FMT_BAYER_RGGB8] = { 1, 0 },
216 [AV_PIX_FMT_BAYER_GBRG8] = { 1, 0 },
217 [AV_PIX_FMT_BAYER_GRBG8] = { 1, 0 },
218 [AV_PIX_FMT_BAYER_BGGR16LE] = { 1, 0 },
219 [AV_PIX_FMT_BAYER_BGGR16BE] = { 1, 0 },
220 [AV_PIX_FMT_BAYER_RGGB16LE] = { 1, 0 },
221 [AV_PIX_FMT_BAYER_RGGB16BE] = { 1, 0 },
222 [AV_PIX_FMT_BAYER_GBRG16LE] = { 1, 0 },
223 [AV_PIX_FMT_BAYER_GBRG16BE] = { 1, 0 },
224 [AV_PIX_FMT_BAYER_GRBG16LE] = { 1, 0 },
225 [AV_PIX_FMT_BAYER_GRBG16BE] = { 1, 0 },
226 [AV_PIX_FMT_XYZ12BE] = { 1, 1, 1 },
227 [AV_PIX_FMT_XYZ12LE] = { 1, 1, 1 },
228 [AV_PIX_FMT_AYUV64LE] = { 1, 1},
231 int sws_isSupportedInput(enum AVPixelFormat pix_fmt)
233 return (unsigned)pix_fmt < AV_PIX_FMT_NB ?
234 format_entries[pix_fmt].is_supported_in : 0;
237 int sws_isSupportedOutput(enum AVPixelFormat pix_fmt)
239 return (unsigned)pix_fmt < AV_PIX_FMT_NB ?
240 format_entries[pix_fmt].is_supported_out : 0;
243 int sws_isSupportedEndiannessConversion(enum AVPixelFormat pix_fmt)
245 return (unsigned)pix_fmt < AV_PIX_FMT_NB ?
246 format_entries[pix_fmt].is_supported_endianness : 0;
249 static double getSplineCoeff(double a, double b, double c, double d,
253 return ((d * dist + c) * dist + b) * dist + a;
255 return getSplineCoeff(0.0,
256 b + 2.0 * c + 3.0 * d,
258 -b - 3.0 * c - 6.0 * d,
262 static av_cold int get_local_pos(SwsContext *s, int chr_subsample, int pos, int dir)
264 if (pos == -1 || pos <= -513) {
265 pos = (128 << chr_subsample) - 128;
267 pos += 128; // relative to ideal left edge
268 return pos >> chr_subsample;
272 int flag; ///< flag associated to the algorithm
273 const char *description; ///< human-readable description
274 int size_factor; ///< size factor used when initing the filters
277 static const ScaleAlgorithm scale_algorithms[] = {
278 { SWS_AREA, "area averaging", 1 /* downscale only, for upscale it is bilinear */ },
279 { SWS_BICUBIC, "bicubic", 4 },
280 { SWS_BICUBLIN, "luma bicubic / chroma bilinear", -1 },
281 { SWS_BILINEAR, "bilinear", 2 },
282 { SWS_FAST_BILINEAR, "fast bilinear", -1 },
283 { SWS_GAUSS, "Gaussian", 8 /* infinite ;) */ },
284 { SWS_LANCZOS, "Lanczos", -1 /* custom */ },
285 { SWS_POINT, "nearest neighbor / point", -1 },
286 { SWS_SINC, "sinc", 20 /* infinite ;) */ },
287 { SWS_SPLINE, "bicubic spline", 20 /* infinite :)*/ },
288 { SWS_X, "experimental", 8 },
291 static av_cold int initFilter(int16_t **outFilter, int32_t **filterPos,
292 int *outFilterSize, int xInc, int srcW,
293 int dstW, int filterAlign, int one,
294 int flags, int cpu_flags,
295 SwsVector *srcFilter, SwsVector *dstFilter,
296 double param[2], int srcPos, int dstPos)
302 int64_t *filter = NULL;
303 int64_t *filter2 = NULL;
304 const int64_t fone = 1LL << (54 - FFMIN(av_log2(srcW/dstW), 8));
307 emms_c(); // FIXME should not be required but IS (even for non-MMX versions)
309 // NOTE: the +3 is for the MMX(+1) / SSE(+3) scaler which reads over the end
310 FF_ALLOC_ARRAY_OR_GOTO(NULL, *filterPos, (dstW + 3), sizeof(**filterPos), fail);
312 if (FFABS(xInc - 0x10000) < 10 && srcPos == dstPos) { // unscaled
315 FF_ALLOCZ_ARRAY_OR_GOTO(NULL, filter,
316 dstW, sizeof(*filter) * filterSize, fail);
318 for (i = 0; i < dstW; i++) {
319 filter[i * filterSize] = fone;
322 } else if (flags & SWS_POINT) { // lame looking point sampling mode
326 FF_ALLOC_ARRAY_OR_GOTO(NULL, filter,
327 dstW, sizeof(*filter) * filterSize, fail);
329 xDstInSrc = ((dstPos*(int64_t)xInc)>>8) - ((srcPos*0x8000LL)>>7);
330 for (i = 0; i < dstW; i++) {
331 int xx = (xDstInSrc - ((filterSize - 1) << 15) + (1 << 15)) >> 16;
333 (*filterPos)[i] = xx;
337 } else if ((xInc <= (1 << 16) && (flags & SWS_AREA)) ||
338 (flags & SWS_FAST_BILINEAR)) { // bilinear upscale
342 FF_ALLOC_ARRAY_OR_GOTO(NULL, filter,
343 dstW, sizeof(*filter) * filterSize, fail);
345 xDstInSrc = ((dstPos*(int64_t)xInc)>>8) - ((srcPos*0x8000LL)>>7);
346 for (i = 0; i < dstW; i++) {
347 int xx = (xDstInSrc - ((filterSize - 1) << 15) + (1 << 15)) >> 16;
350 (*filterPos)[i] = xx;
351 // bilinear upscale / linear interpolate / area averaging
352 for (j = 0; j < filterSize; j++) {
353 int64_t coeff= fone - FFABS(((int64_t)xx<<16) - xDstInSrc)*(fone>>16);
356 filter[i * filterSize + j] = coeff;
365 for (i = 0; i < FF_ARRAY_ELEMS(scale_algorithms); i++) {
366 if (flags & scale_algorithms[i].flag && scale_algorithms[i].size_factor > 0) {
367 sizeFactor = scale_algorithms[i].size_factor;
371 if (flags & SWS_LANCZOS)
372 sizeFactor = param[0] != SWS_PARAM_DEFAULT ? ceil(2 * param[0]) : 6;
373 av_assert0(sizeFactor > 0);
376 filterSize = 1 + sizeFactor; // upscale
378 filterSize = 1 + (sizeFactor * srcW + dstW - 1) / dstW;
380 filterSize = FFMIN(filterSize, srcW - 2);
381 filterSize = FFMAX(filterSize, 1);
383 FF_ALLOC_ARRAY_OR_GOTO(NULL, filter,
384 dstW, sizeof(*filter) * filterSize, fail);
386 xDstInSrc = ((dstPos*(int64_t)xInc)>>7) - ((srcPos*0x10000LL)>>7);
387 for (i = 0; i < dstW; i++) {
388 int xx = (xDstInSrc - ((int64_t)(filterSize - 2) << 16)) / (1 << 17);
390 (*filterPos)[i] = xx;
391 for (j = 0; j < filterSize; j++) {
392 int64_t d = (FFABS(((int64_t)xx * (1 << 17)) - xDstInSrc)) << 13;
398 floatd = d * (1.0 / (1 << 30));
400 if (flags & SWS_BICUBIC) {
401 int64_t B = (param[0] != SWS_PARAM_DEFAULT ? param[0] : 0) * (1 << 24);
402 int64_t C = (param[1] != SWS_PARAM_DEFAULT ? param[1] : 0.6) * (1 << 24);
404 if (d >= 1LL << 31) {
407 int64_t dd = (d * d) >> 30;
408 int64_t ddd = (dd * d) >> 30;
411 coeff = (12 * (1 << 24) - 9 * B - 6 * C) * ddd +
412 (-18 * (1 << 24) + 12 * B + 6 * C) * dd +
413 (6 * (1 << 24) - 2 * B) * (1 << 30);
415 coeff = (-B - 6 * C) * ddd +
416 (6 * B + 30 * C) * dd +
417 (-12 * B - 48 * C) * d +
418 (8 * B + 24 * C) * (1 << 30);
420 coeff /= (1LL<<54)/fone;
423 else if (flags & SWS_X) {
424 double p = param ? param * 0.01 : 0.3;
425 coeff = d ? sin(d * M_PI) / (d * M_PI) : 1.0;
426 coeff *= pow(2.0, -p * d * d);
429 else if (flags & SWS_X) {
430 double A = param[0] != SWS_PARAM_DEFAULT ? param[0] : 1.0;
434 c = cos(floatd * M_PI);
441 coeff = (c * 0.5 + 0.5) * fone;
442 } else if (flags & SWS_AREA) {
443 int64_t d2 = d - (1 << 29);
444 if (d2 * xInc < -(1LL << (29 + 16)))
445 coeff = 1.0 * (1LL << (30 + 16));
446 else if (d2 * xInc < (1LL << (29 + 16)))
447 coeff = -d2 * xInc + (1LL << (29 + 16));
450 coeff *= fone >> (30 + 16);
451 } else if (flags & SWS_GAUSS) {
452 double p = param[0] != SWS_PARAM_DEFAULT ? param[0] : 3.0;
453 coeff = (pow(2.0, -p * floatd * floatd)) * fone;
454 } else if (flags & SWS_SINC) {
455 coeff = (d ? sin(floatd * M_PI) / (floatd * M_PI) : 1.0) * fone;
456 } else if (flags & SWS_LANCZOS) {
457 double p = param[0] != SWS_PARAM_DEFAULT ? param[0] : 3.0;
458 coeff = (d ? sin(floatd * M_PI) * sin(floatd * M_PI / p) /
459 (floatd * floatd * M_PI * M_PI / p) : 1.0) * fone;
462 } else if (flags & SWS_BILINEAR) {
463 coeff = (1 << 30) - d;
467 } else if (flags & SWS_SPLINE) {
468 double p = -2.196152422706632;
469 coeff = getSplineCoeff(1.0, 0.0, p, -p - 1.0, floatd) * fone;
474 filter[i * filterSize + j] = coeff;
477 xDstInSrc += 2 * xInc;
481 /* apply src & dst Filter to filter -> filter2
484 av_assert0(filterSize > 0);
485 filter2Size = filterSize;
487 filter2Size += srcFilter->length - 1;
489 filter2Size += dstFilter->length - 1;
490 av_assert0(filter2Size > 0);
491 FF_ALLOCZ_ARRAY_OR_GOTO(NULL, filter2, dstW, filter2Size * sizeof(*filter2), fail);
493 for (i = 0; i < dstW; i++) {
497 for (k = 0; k < srcFilter->length; k++) {
498 for (j = 0; j < filterSize; j++)
499 filter2[i * filter2Size + k + j] +=
500 srcFilter->coeff[k] * filter[i * filterSize + j];
503 for (j = 0; j < filterSize; j++)
504 filter2[i * filter2Size + j] = filter[i * filterSize + j];
508 (*filterPos)[i] += (filterSize - 1) / 2 - (filter2Size - 1) / 2;
512 /* try to reduce the filter-size (step1 find size and shift left) */
513 // Assume it is near normalized (*0.5 or *2.0 is OK but * 0.001 is not).
515 for (i = dstW - 1; i >= 0; i--) {
516 int min = filter2Size;
518 int64_t cutOff = 0.0;
520 /* get rid of near zero elements on the left by shifting left */
521 for (j = 0; j < filter2Size; j++) {
523 cutOff += FFABS(filter2[i * filter2Size]);
525 if (cutOff > SWS_MAX_REDUCE_CUTOFF * fone)
528 /* preserve monotonicity because the core can't handle the
529 * filter otherwise */
530 if (i < dstW - 1 && (*filterPos)[i] >= (*filterPos)[i + 1])
533 // move filter coefficients left
534 for (k = 1; k < filter2Size; k++)
535 filter2[i * filter2Size + k - 1] = filter2[i * filter2Size + k];
536 filter2[i * filter2Size + k - 1] = 0;
541 /* count near zeros on the right */
542 for (j = filter2Size - 1; j > 0; j--) {
543 cutOff += FFABS(filter2[i * filter2Size + j]);
545 if (cutOff > SWS_MAX_REDUCE_CUTOFF * fone)
550 if (min > minFilterSize)
554 if (PPC_ALTIVEC(cpu_flags)) {
555 // we can handle the special case 4, so we don't want to go the full 8
556 if (minFilterSize < 5)
559 /* We really don't want to waste our time doing useless computation, so
560 * fall back on the scalar C code for very small filters.
561 * Vectorizing is worth it only if you have a decent-sized vector. */
562 if (minFilterSize < 3)
566 if (HAVE_MMX && cpu_flags & AV_CPU_FLAG_MMX) {
567 // special case for unscaled vertical filtering
568 if (minFilterSize == 1 && filterAlign == 2)
572 av_assert0(minFilterSize > 0);
573 filterSize = (minFilterSize + (filterAlign - 1)) & (~(filterAlign - 1));
574 av_assert0(filterSize > 0);
575 filter = av_malloc_array(dstW, filterSize * sizeof(*filter));
578 if (filterSize >= MAX_FILTER_SIZE * 16 /
579 ((flags & SWS_ACCURATE_RND) ? APCK_SIZE : 16)) {
580 ret = RETCODE_USE_CASCADE;
583 *outFilterSize = filterSize;
585 if (flags & SWS_PRINT_INFO)
586 av_log(NULL, AV_LOG_VERBOSE,
587 "SwScaler: reducing / aligning filtersize %d -> %d\n",
588 filter2Size, filterSize);
589 /* try to reduce the filter-size (step2 reduce it) */
590 for (i = 0; i < dstW; i++) {
593 for (j = 0; j < filterSize; j++) {
594 if (j >= filter2Size)
595 filter[i * filterSize + j] = 0;
597 filter[i * filterSize + j] = filter2[i * filter2Size + j];
598 if ((flags & SWS_BITEXACT) && j >= minFilterSize)
599 filter[i * filterSize + j] = 0;
603 // FIXME try to align filterPos if possible
606 for (i = 0; i < dstW; i++) {
608 if ((*filterPos)[i] < 0) {
609 // move filter coefficients left to compensate for filterPos
610 for (j = 1; j < filterSize; j++) {
611 int left = FFMAX(j + (*filterPos)[i], 0);
612 filter[i * filterSize + left] += filter[i * filterSize + j];
613 filter[i * filterSize + j] = 0;
618 if ((*filterPos)[i] + filterSize > srcW) {
619 int shift = (*filterPos)[i] + FFMIN(filterSize - srcW, 0);
622 for (j = filterSize - 1; j >= 0; j--) {
623 if ((*filterPos)[i] + j >= srcW) {
624 acc += filter[i * filterSize + j];
625 filter[i * filterSize + j] = 0;
628 for (j = filterSize - 1; j >= 0; j--) {
630 filter[i * filterSize + j] = 0;
632 filter[i * filterSize + j] = filter[i * filterSize + j - shift];
636 (*filterPos)[i]-= shift;
637 filter[i * filterSize + srcW - 1 - (*filterPos)[i]] += acc;
639 av_assert0((*filterPos)[i] >= 0);
640 av_assert0((*filterPos)[i] < srcW);
641 if ((*filterPos)[i] + filterSize > srcW) {
642 for (j = 0; j < filterSize; j++) {
643 av_assert0((*filterPos)[i] + j < srcW || !filter[i * filterSize + j]);
648 // Note the +1 is for the MMX scaler which reads over the end
649 /* align at 16 for AltiVec (needed by hScale_altivec_real) */
650 FF_ALLOCZ_ARRAY_OR_GOTO(NULL, *outFilter,
651 (dstW + 3), *outFilterSize * sizeof(int16_t), fail);
653 /* normalize & store in outFilter */
654 for (i = 0; i < dstW; i++) {
659 for (j = 0; j < filterSize; j++) {
660 sum += filter[i * filterSize + j];
662 sum = (sum + one / 2) / one;
664 av_log(NULL, AV_LOG_WARNING, "SwScaler: zero vector in scaling\n");
667 for (j = 0; j < *outFilterSize; j++) {
668 int64_t v = filter[i * filterSize + j] + error;
669 int intV = ROUNDED_DIV(v, sum);
670 (*outFilter)[i * (*outFilterSize) + j] = intV;
671 error = v - intV * sum;
675 (*filterPos)[dstW + 0] =
676 (*filterPos)[dstW + 1] =
677 (*filterPos)[dstW + 2] = (*filterPos)[dstW - 1]; /* the MMX/SSE scaler will
678 * read over the end */
679 for (i = 0; i < *outFilterSize; i++) {
680 int k = (dstW - 1) * (*outFilterSize) + i;
681 (*outFilter)[k + 1 * (*outFilterSize)] =
682 (*outFilter)[k + 2 * (*outFilterSize)] =
683 (*outFilter)[k + 3 * (*outFilterSize)] = (*outFilter)[k];
690 av_log(NULL, ret == RETCODE_USE_CASCADE ? AV_LOG_DEBUG : AV_LOG_ERROR, "sws: initFilter failed\n");
696 static void fill_rgb2yuv_table(SwsContext *c, const int table[4], int dstRange)
698 int64_t W, V, Z, Cy, Cu, Cv;
699 int64_t vr = table[0];
700 int64_t ub = table[1];
701 int64_t ug = -table[2];
702 int64_t vg = -table[3];
705 uint8_t *p = (uint8_t*)c->input_rgb2yuv_table;
707 static const int8_t map[] = {
708 BY_IDX, GY_IDX, -1 , BY_IDX, BY_IDX, GY_IDX, -1 , BY_IDX,
709 RY_IDX, -1 , GY_IDX, RY_IDX, RY_IDX, -1 , GY_IDX, RY_IDX,
710 RY_IDX, GY_IDX, -1 , RY_IDX, RY_IDX, GY_IDX, -1 , RY_IDX,
711 BY_IDX, -1 , GY_IDX, BY_IDX, BY_IDX, -1 , GY_IDX, BY_IDX,
712 BU_IDX, GU_IDX, -1 , BU_IDX, BU_IDX, GU_IDX, -1 , BU_IDX,
713 RU_IDX, -1 , GU_IDX, RU_IDX, RU_IDX, -1 , GU_IDX, RU_IDX,
714 RU_IDX, GU_IDX, -1 , RU_IDX, RU_IDX, GU_IDX, -1 , RU_IDX,
715 BU_IDX, -1 , GU_IDX, BU_IDX, BU_IDX, -1 , GU_IDX, BU_IDX,
716 BV_IDX, GV_IDX, -1 , BV_IDX, BV_IDX, GV_IDX, -1 , BV_IDX,
717 RV_IDX, -1 , GV_IDX, RV_IDX, RV_IDX, -1 , GV_IDX, RV_IDX,
718 RV_IDX, GV_IDX, -1 , RV_IDX, RV_IDX, GV_IDX, -1 , RV_IDX,
719 BV_IDX, -1 , GV_IDX, BV_IDX, BV_IDX, -1 , GV_IDX, BV_IDX,
720 RY_IDX, BY_IDX, RY_IDX, BY_IDX, RY_IDX, BY_IDX, RY_IDX, BY_IDX,
721 BY_IDX, RY_IDX, BY_IDX, RY_IDX, BY_IDX, RY_IDX, BY_IDX, RY_IDX,
722 GY_IDX, -1 , GY_IDX, -1 , GY_IDX, -1 , GY_IDX, -1 ,
723 -1 , GY_IDX, -1 , GY_IDX, -1 , GY_IDX, -1 , GY_IDX,
724 RU_IDX, BU_IDX, RU_IDX, BU_IDX, RU_IDX, BU_IDX, RU_IDX, BU_IDX,
725 BU_IDX, RU_IDX, BU_IDX, RU_IDX, BU_IDX, RU_IDX, BU_IDX, RU_IDX,
726 GU_IDX, -1 , GU_IDX, -1 , GU_IDX, -1 , GU_IDX, -1 ,
727 -1 , GU_IDX, -1 , GU_IDX, -1 , GU_IDX, -1 , GU_IDX,
728 RV_IDX, BV_IDX, RV_IDX, BV_IDX, RV_IDX, BV_IDX, RV_IDX, BV_IDX,
729 BV_IDX, RV_IDX, BV_IDX, RV_IDX, BV_IDX, RV_IDX, BV_IDX, RV_IDX,
730 GV_IDX, -1 , GV_IDX, -1 , GV_IDX, -1 , GV_IDX, -1 ,
731 -1 , GV_IDX, -1 , GV_IDX, -1 , GV_IDX, -1 , GV_IDX, //23
732 -1 , -1 , -1 , -1 , -1 , -1 , -1 , -1 , //24
733 -1 , -1 , -1 , -1 , -1 , -1 , -1 , -1 , //25
734 -1 , -1 , -1 , -1 , -1 , -1 , -1 , -1 , //26
735 -1 , -1 , -1 , -1 , -1 , -1 , -1 , -1 , //27
736 -1 , -1 , -1 , -1 , -1 , -1 , -1 , -1 , //28
737 -1 , -1 , -1 , -1 , -1 , -1 , -1 , -1 , //29
738 -1 , -1 , -1 , -1 , -1 , -1 , -1 , -1 , //30
739 -1 , -1 , -1 , -1 , -1 , -1 , -1 , -1 , //31
740 BY_IDX, GY_IDX, RY_IDX, -1 , -1 , -1 , -1 , -1 , //32
741 BU_IDX, GU_IDX, RU_IDX, -1 , -1 , -1 , -1 , -1 , //33
742 BV_IDX, GV_IDX, RV_IDX, -1 , -1 , -1 , -1 , -1 , //34
745 dstRange = 0; //FIXME range = 1 is handled elsewhere
755 W = ROUNDED_DIV(ONE*ONE*ug, ub);
756 V = ROUNDED_DIV(ONE*ONE*vg, vr);
759 Cy = ROUNDED_DIV(cy*Z, ONE);
760 Cu = ROUNDED_DIV(ub*Z, ONE);
761 Cv = ROUNDED_DIV(vr*Z, ONE);
763 c->input_rgb2yuv_table[RY_IDX] = -ROUNDED_DIV((1 << RGB2YUV_SHIFT)*V , Cy);
764 c->input_rgb2yuv_table[GY_IDX] = ROUNDED_DIV((1 << RGB2YUV_SHIFT)*ONE*ONE , Cy);
765 c->input_rgb2yuv_table[BY_IDX] = -ROUNDED_DIV((1 << RGB2YUV_SHIFT)*W , Cy);
767 c->input_rgb2yuv_table[RU_IDX] = ROUNDED_DIV((1 << RGB2YUV_SHIFT)*V , Cu);
768 c->input_rgb2yuv_table[GU_IDX] = -ROUNDED_DIV((1 << RGB2YUV_SHIFT)*ONE*ONE , Cu);
769 c->input_rgb2yuv_table[BU_IDX] = ROUNDED_DIV((1 << RGB2YUV_SHIFT)*(Z+W) , Cu);
771 c->input_rgb2yuv_table[RV_IDX] = ROUNDED_DIV((1 << RGB2YUV_SHIFT)*(V+Z) , Cv);
772 c->input_rgb2yuv_table[GV_IDX] = -ROUNDED_DIV((1 << RGB2YUV_SHIFT)*ONE*ONE , Cv);
773 c->input_rgb2yuv_table[BV_IDX] = ROUNDED_DIV((1 << RGB2YUV_SHIFT)*W , Cv);
775 if(/*!dstRange && */!memcmp(table, ff_yuv2rgb_coeffs[SWS_CS_DEFAULT], sizeof(ff_yuv2rgb_coeffs[SWS_CS_DEFAULT]))) {
776 c->input_rgb2yuv_table[BY_IDX] = ((int)(0.114 * 219 / 255 * (1 << RGB2YUV_SHIFT) + 0.5));
777 c->input_rgb2yuv_table[BV_IDX] = (-(int)(0.081 * 224 / 255 * (1 << RGB2YUV_SHIFT) + 0.5));
778 c->input_rgb2yuv_table[BU_IDX] = ((int)(0.500 * 224 / 255 * (1 << RGB2YUV_SHIFT) + 0.5));
779 c->input_rgb2yuv_table[GY_IDX] = ((int)(0.587 * 219 / 255 * (1 << RGB2YUV_SHIFT) + 0.5));
780 c->input_rgb2yuv_table[GV_IDX] = (-(int)(0.419 * 224 / 255 * (1 << RGB2YUV_SHIFT) + 0.5));
781 c->input_rgb2yuv_table[GU_IDX] = (-(int)(0.331 * 224 / 255 * (1 << RGB2YUV_SHIFT) + 0.5));
782 c->input_rgb2yuv_table[RY_IDX] = ((int)(0.299 * 219 / 255 * (1 << RGB2YUV_SHIFT) + 0.5));
783 c->input_rgb2yuv_table[RV_IDX] = ((int)(0.500 * 224 / 255 * (1 << RGB2YUV_SHIFT) + 0.5));
784 c->input_rgb2yuv_table[RU_IDX] = (-(int)(0.169 * 224 / 255 * (1 << RGB2YUV_SHIFT) + 0.5));
786 for(i=0; i<FF_ARRAY_ELEMS(map); i++)
787 AV_WL16(p + 16*4 + 2*i, map[i] >= 0 ? c->input_rgb2yuv_table[map[i]] : 0);
790 static void fill_xyztables(struct SwsContext *c)
793 double xyzgamma = XYZ_GAMMA;
794 double rgbgamma = 1.0 / RGB_GAMMA;
795 double xyzgammainv = 1.0 / XYZ_GAMMA;
796 double rgbgammainv = RGB_GAMMA;
797 static const int16_t xyz2rgb_matrix[3][4] = {
798 {13270, -6295, -2041},
800 { 228, -835, 4329} };
801 static const int16_t rgb2xyz_matrix[3][4] = {
805 static int16_t xyzgamma_tab[4096], rgbgamma_tab[4096], xyzgammainv_tab[4096], rgbgammainv_tab[4096];
807 memcpy(c->xyz2rgb_matrix, xyz2rgb_matrix, sizeof(c->xyz2rgb_matrix));
808 memcpy(c->rgb2xyz_matrix, rgb2xyz_matrix, sizeof(c->rgb2xyz_matrix));
809 c->xyzgamma = xyzgamma_tab;
810 c->rgbgamma = rgbgamma_tab;
811 c->xyzgammainv = xyzgammainv_tab;
812 c->rgbgammainv = rgbgammainv_tab;
814 if (rgbgamma_tab[4095])
817 /* set gamma vectors */
818 for (i = 0; i < 4096; i++) {
819 xyzgamma_tab[i] = lrint(pow(i / 4095.0, xyzgamma) * 4095.0);
820 rgbgamma_tab[i] = lrint(pow(i / 4095.0, rgbgamma) * 4095.0);
821 xyzgammainv_tab[i] = lrint(pow(i / 4095.0, xyzgammainv) * 4095.0);
822 rgbgammainv_tab[i] = lrint(pow(i / 4095.0, rgbgammainv) * 4095.0);
826 int sws_setColorspaceDetails(struct SwsContext *c, const int inv_table[4],
827 int srcRange, const int table[4], int dstRange,
828 int brightness, int contrast, int saturation)
830 const AVPixFmtDescriptor *desc_dst;
831 const AVPixFmtDescriptor *desc_src;
833 memmove(c->srcColorspaceTable, inv_table, sizeof(int) * 4);
834 memmove(c->dstColorspaceTable, table, sizeof(int) * 4);
837 desc_dst = av_pix_fmt_desc_get(c->dstFormat);
838 desc_src = av_pix_fmt_desc_get(c->srcFormat);
840 if(!isYUV(c->dstFormat) && !isGray(c->dstFormat))
842 if(!isYUV(c->srcFormat) && !isGray(c->srcFormat))
845 c->brightness = brightness;
846 c->contrast = contrast;
847 c->saturation = saturation;
848 if (c->srcRange != srcRange || c->dstRange != dstRange)
850 c->srcRange = srcRange;
851 c->dstRange = dstRange;
853 //The srcBpc check is possibly wrong but we seem to lack a definitive reference to test this
854 //and what we have in ticket 2939 looks better with this check
855 if (need_reinit && (c->srcBpc == 8 || !isYUV(c->srcFormat)))
856 ff_sws_init_range_convert(c);
858 c->dstFormatBpp = av_get_bits_per_pixel(desc_dst);
859 c->srcFormatBpp = av_get_bits_per_pixel(desc_src);
861 if (c->cascaded_context[0])
862 return sws_setColorspaceDetails(c->cascaded_context[0],inv_table, srcRange,table, dstRange, brightness, contrast, saturation);
864 if ((isYUV(c->dstFormat) || isGray(c->dstFormat)) && (isYUV(c->srcFormat) || isGray(c->srcFormat))) {
865 if (!c->cascaded_context[0] &&
866 memcmp(c->dstColorspaceTable, c->srcColorspaceTable, sizeof(int) * 4) &&
867 c->srcW && c->srcH && c->dstW && c->dstH) {
868 enum AVPixelFormat tmp_format;
869 int tmp_width, tmp_height;
875 av_log(c, AV_LOG_VERBOSE, "YUV color matrix differs for YUV->YUV, using intermediate RGB to convert\n");
877 if (isNBPS(c->dstFormat) || is16BPS(c->dstFormat)) {
878 if (isALPHA(c->srcFormat) && isALPHA(c->dstFormat)) {
879 tmp_format = AV_PIX_FMT_BGRA64;
881 tmp_format = AV_PIX_FMT_BGR48;
884 if (isALPHA(c->srcFormat) && isALPHA(c->dstFormat)) {
885 tmp_format = AV_PIX_FMT_BGRA;
887 tmp_format = AV_PIX_FMT_BGR24;
891 if (srcW*srcH > dstW*dstH) {
899 ret = av_image_alloc(c->cascaded_tmp, c->cascaded_tmpStride,
900 tmp_width, tmp_height, tmp_format, 64);
904 c->cascaded_context[0] = sws_alloc_set_opts(srcW, srcH, c->srcFormat,
905 tmp_width, tmp_height, tmp_format,
907 if (!c->cascaded_context[0])
910 c->cascaded_context[0]->alphablend = c->alphablend;
911 ret = sws_init_context(c->cascaded_context[0], NULL , NULL);
914 //we set both src and dst depending on that the RGB side will be ignored
915 sws_setColorspaceDetails(c->cascaded_context[0], inv_table,
916 srcRange, table, dstRange,
917 brightness, contrast, saturation);
919 c->cascaded_context[1] = sws_getContext(tmp_width, tmp_height, tmp_format,
920 dstW, dstH, c->dstFormat,
921 c->flags, NULL, NULL, c->param);
922 if (!c->cascaded_context[1])
924 sws_setColorspaceDetails(c->cascaded_context[1], inv_table,
925 srcRange, table, dstRange,
926 0, 1 << 16, 1 << 16);
932 if (!isYUV(c->dstFormat) && !isGray(c->dstFormat)) {
933 ff_yuv2rgb_c_init_tables(c, inv_table, srcRange, brightness,
934 contrast, saturation);
938 ff_yuv2rgb_init_tables_ppc(c, inv_table, brightness,
939 contrast, saturation);
942 fill_rgb2yuv_table(c, table, dstRange);
947 int sws_getColorspaceDetails(struct SwsContext *c, int **inv_table,
948 int *srcRange, int **table, int *dstRange,
949 int *brightness, int *contrast, int *saturation)
954 *inv_table = c->srcColorspaceTable;
955 *table = c->dstColorspaceTable;
956 *srcRange = c->srcRange;
957 *dstRange = c->dstRange;
958 *brightness = c->brightness;
959 *contrast = c->contrast;
960 *saturation = c->saturation;
965 static int handle_jpeg(enum AVPixelFormat *format)
968 case AV_PIX_FMT_YUVJ420P:
969 *format = AV_PIX_FMT_YUV420P;
971 case AV_PIX_FMT_YUVJ411P:
972 *format = AV_PIX_FMT_YUV411P;
974 case AV_PIX_FMT_YUVJ422P:
975 *format = AV_PIX_FMT_YUV422P;
977 case AV_PIX_FMT_YUVJ444P:
978 *format = AV_PIX_FMT_YUV444P;
980 case AV_PIX_FMT_YUVJ440P:
981 *format = AV_PIX_FMT_YUV440P;
983 case AV_PIX_FMT_GRAY8:
984 case AV_PIX_FMT_GRAY16LE:
985 case AV_PIX_FMT_GRAY16BE:
992 static int handle_0alpha(enum AVPixelFormat *format)
995 case AV_PIX_FMT_0BGR : *format = AV_PIX_FMT_ABGR ; return 1;
996 case AV_PIX_FMT_BGR0 : *format = AV_PIX_FMT_BGRA ; return 4;
997 case AV_PIX_FMT_0RGB : *format = AV_PIX_FMT_ARGB ; return 1;
998 case AV_PIX_FMT_RGB0 : *format = AV_PIX_FMT_RGBA ; return 4;
1003 static int handle_xyz(enum AVPixelFormat *format)
1006 case AV_PIX_FMT_XYZ12BE : *format = AV_PIX_FMT_RGB48BE; return 1;
1007 case AV_PIX_FMT_XYZ12LE : *format = AV_PIX_FMT_RGB48LE; return 1;
1012 static void handle_formats(SwsContext *c)
1014 c->src0Alpha |= handle_0alpha(&c->srcFormat);
1015 c->dst0Alpha |= handle_0alpha(&c->dstFormat);
1016 c->srcXYZ |= handle_xyz(&c->srcFormat);
1017 c->dstXYZ |= handle_xyz(&c->dstFormat);
1018 if (c->srcXYZ || c->dstXYZ)
1022 SwsContext *sws_alloc_context(void)
1024 SwsContext *c = av_mallocz(sizeof(SwsContext));
1026 av_assert0(offsetof(SwsContext, redDither) + DITHER32_INT == offsetof(SwsContext, dither32));
1029 c->av_class = &sws_context_class;
1030 av_opt_set_defaults(c);
1036 static uint16_t * alloc_gamma_tbl(double e)
1040 tbl = (uint16_t*)av_malloc(sizeof(uint16_t) * 1 << 16);
1044 for (i = 0; i < 65536; ++i) {
1045 tbl[i] = pow(i / 65535.0, e) * 65535.0;
1050 static enum AVPixelFormat alphaless_fmt(enum AVPixelFormat fmt)
1053 case AV_PIX_FMT_ARGB: return AV_PIX_FMT_RGB24;
1054 case AV_PIX_FMT_RGBA: return AV_PIX_FMT_RGB24;
1055 case AV_PIX_FMT_ABGR: return AV_PIX_FMT_BGR24;
1056 case AV_PIX_FMT_BGRA: return AV_PIX_FMT_BGR24;
1057 case AV_PIX_FMT_YA8: return AV_PIX_FMT_GRAY8;
1059 case AV_PIX_FMT_YUVA420P: return AV_PIX_FMT_YUV420P;
1060 case AV_PIX_FMT_YUVA422P: return AV_PIX_FMT_YUV422P;
1061 case AV_PIX_FMT_YUVA444P: return AV_PIX_FMT_YUV444P;
1063 case AV_PIX_FMT_GBRAP: return AV_PIX_FMT_GBRP;
1065 case AV_PIX_FMT_GBRAP16LE: return AV_PIX_FMT_GBRP16;
1066 case AV_PIX_FMT_GBRAP16BE: return AV_PIX_FMT_GBRP16;
1068 case AV_PIX_FMT_RGBA64LE: return AV_PIX_FMT_RGB48;
1069 case AV_PIX_FMT_RGBA64BE: return AV_PIX_FMT_RGB48;
1070 case AV_PIX_FMT_BGRA64LE: return AV_PIX_FMT_BGR48;
1071 case AV_PIX_FMT_BGRA64BE: return AV_PIX_FMT_BGR48;
1073 case AV_PIX_FMT_YA16BE: return AV_PIX_FMT_GRAY16;
1074 case AV_PIX_FMT_YA16LE: return AV_PIX_FMT_GRAY16;
1076 case AV_PIX_FMT_YUVA420P9BE: return AV_PIX_FMT_YUV420P9;
1077 case AV_PIX_FMT_YUVA422P9BE: return AV_PIX_FMT_YUV422P9;
1078 case AV_PIX_FMT_YUVA444P9BE: return AV_PIX_FMT_YUV444P9;
1079 case AV_PIX_FMT_YUVA420P9LE: return AV_PIX_FMT_YUV420P9;
1080 case AV_PIX_FMT_YUVA422P9LE: return AV_PIX_FMT_YUV422P9;
1081 case AV_PIX_FMT_YUVA444P9LE: return AV_PIX_FMT_YUV444P9;
1082 case AV_PIX_FMT_YUVA420P10BE: return AV_PIX_FMT_YUV420P10;
1083 case AV_PIX_FMT_YUVA422P10BE: return AV_PIX_FMT_YUV422P10;
1084 case AV_PIX_FMT_YUVA444P10BE: return AV_PIX_FMT_YUV444P10;
1085 case AV_PIX_FMT_YUVA420P10LE: return AV_PIX_FMT_YUV420P10;
1086 case AV_PIX_FMT_YUVA422P10LE: return AV_PIX_FMT_YUV422P10;
1087 case AV_PIX_FMT_YUVA444P10LE: return AV_PIX_FMT_YUV444P10;
1088 case AV_PIX_FMT_YUVA420P16BE: return AV_PIX_FMT_YUV420P16;
1089 case AV_PIX_FMT_YUVA422P16BE: return AV_PIX_FMT_YUV422P16;
1090 case AV_PIX_FMT_YUVA444P16BE: return AV_PIX_FMT_YUV444P16;
1091 case AV_PIX_FMT_YUVA420P16LE: return AV_PIX_FMT_YUV420P16;
1092 case AV_PIX_FMT_YUVA422P16LE: return AV_PIX_FMT_YUV422P16;
1093 case AV_PIX_FMT_YUVA444P16LE: return AV_PIX_FMT_YUV444P16;
1095 // case AV_PIX_FMT_AYUV64LE:
1096 // case AV_PIX_FMT_AYUV64BE:
1097 // case AV_PIX_FMT_PAL8:
1098 default: return AV_PIX_FMT_NONE;
1102 av_cold int sws_init_context(SwsContext *c, SwsFilter *srcFilter,
1103 SwsFilter *dstFilter)
1106 int usesVFilter, usesHFilter;
1108 SwsFilter dummyFilter = { NULL, NULL, NULL, NULL };
1113 int dst_stride = FFALIGN(dstW * sizeof(int16_t) + 66, 16);
1114 int flags, cpu_flags;
1115 enum AVPixelFormat srcFormat = c->srcFormat;
1116 enum AVPixelFormat dstFormat = c->dstFormat;
1117 const AVPixFmtDescriptor *desc_src;
1118 const AVPixFmtDescriptor *desc_dst;
1120 enum AVPixelFormat tmpFmt;
1122 cpu_flags = av_get_cpu_flags();
1128 unscaled = (srcW == dstW && srcH == dstH);
1130 c->srcRange |= handle_jpeg(&c->srcFormat);
1131 c->dstRange |= handle_jpeg(&c->dstFormat);
1133 if(srcFormat!=c->srcFormat || dstFormat!=c->dstFormat)
1134 av_log(c, AV_LOG_WARNING, "deprecated pixel format used, make sure you did set range correctly\n");
1136 if (!c->contrast && !c->saturation && !c->dstFormatBpp)
1137 sws_setColorspaceDetails(c, ff_yuv2rgb_coeffs[SWS_CS_DEFAULT], c->srcRange,
1138 ff_yuv2rgb_coeffs[SWS_CS_DEFAULT],
1139 c->dstRange, 0, 1 << 16, 1 << 16);
1142 srcFormat = c->srcFormat;
1143 dstFormat = c->dstFormat;
1144 desc_src = av_pix_fmt_desc_get(srcFormat);
1145 desc_dst = av_pix_fmt_desc_get(dstFormat);
1147 if (!(unscaled && sws_isSupportedEndiannessConversion(srcFormat) &&
1148 av_pix_fmt_swap_endianness(srcFormat) == dstFormat)) {
1149 if (!sws_isSupportedInput(srcFormat)) {
1150 av_log(c, AV_LOG_ERROR, "%s is not supported as input pixel format\n",
1151 av_get_pix_fmt_name(srcFormat));
1152 return AVERROR(EINVAL);
1154 if (!sws_isSupportedOutput(dstFormat)) {
1155 av_log(c, AV_LOG_ERROR, "%s is not supported as output pixel format\n",
1156 av_get_pix_fmt_name(dstFormat));
1157 return AVERROR(EINVAL);
1160 av_assert2(desc_src && desc_dst);
1162 i = flags & (SWS_POINT |
1174 /* provide a default scaler if not set by caller */
1176 if (dstW < srcW && dstH < srcH)
1177 flags |= SWS_BICUBIC;
1178 else if (dstW > srcW && dstH > srcH)
1179 flags |= SWS_BICUBIC;
1181 flags |= SWS_BICUBIC;
1183 } else if (i & (i - 1)) {
1184 av_log(c, AV_LOG_ERROR,
1185 "Exactly one scaler algorithm must be chosen, got %X\n", i);
1186 return AVERROR(EINVAL);
1189 if (srcW < 1 || srcH < 1 || dstW < 1 || dstH < 1) {
1190 /* FIXME check if these are enough and try to lower them after
1191 * fixing the relevant parts of the code */
1192 av_log(c, AV_LOG_ERROR, "%dx%d -> %dx%d is invalid scaling dimension\n",
1193 srcW, srcH, dstW, dstH);
1194 return AVERROR(EINVAL);
1198 dstFilter = &dummyFilter;
1200 srcFilter = &dummyFilter;
1202 c->lumXInc = (((int64_t)srcW << 16) + (dstW >> 1)) / dstW;
1203 c->lumYInc = (((int64_t)srcH << 16) + (dstH >> 1)) / dstH;
1204 c->dstFormatBpp = av_get_bits_per_pixel(desc_dst);
1205 c->srcFormatBpp = av_get_bits_per_pixel(desc_src);
1206 c->vRounder = 4 * 0x0001000100010001ULL;
1208 usesVFilter = (srcFilter->lumV && srcFilter->lumV->length > 1) ||
1209 (srcFilter->chrV && srcFilter->chrV->length > 1) ||
1210 (dstFilter->lumV && dstFilter->lumV->length > 1) ||
1211 (dstFilter->chrV && dstFilter->chrV->length > 1);
1212 usesHFilter = (srcFilter->lumH && srcFilter->lumH->length > 1) ||
1213 (srcFilter->chrH && srcFilter->chrH->length > 1) ||
1214 (dstFilter->lumH && dstFilter->lumH->length > 1) ||
1215 (dstFilter->chrH && dstFilter->chrH->length > 1);
1217 av_pix_fmt_get_chroma_sub_sample(srcFormat, &c->chrSrcHSubSample, &c->chrSrcVSubSample);
1218 av_pix_fmt_get_chroma_sub_sample(dstFormat, &c->chrDstHSubSample, &c->chrDstVSubSample);
1220 if (isAnyRGB(dstFormat) && !(flags&SWS_FULL_CHR_H_INT)) {
1222 av_log(c, AV_LOG_DEBUG, "Forcing full internal H chroma due to odd output size\n");
1223 flags |= SWS_FULL_CHR_H_INT;
1227 if ( c->chrSrcHSubSample == 0
1228 && c->chrSrcVSubSample == 0
1229 && c->dither != SWS_DITHER_BAYER //SWS_FULL_CHR_H_INT is currently not supported with SWS_DITHER_BAYER
1230 && !(c->flags & SWS_FAST_BILINEAR)
1232 av_log(c, AV_LOG_DEBUG, "Forcing full internal H chroma due to input having non subsampled chroma\n");
1233 flags |= SWS_FULL_CHR_H_INT;
1238 if (c->dither == SWS_DITHER_AUTO) {
1239 if (flags & SWS_ERROR_DIFFUSION)
1240 c->dither = SWS_DITHER_ED;
1243 if(dstFormat == AV_PIX_FMT_BGR4_BYTE ||
1244 dstFormat == AV_PIX_FMT_RGB4_BYTE ||
1245 dstFormat == AV_PIX_FMT_BGR8 ||
1246 dstFormat == AV_PIX_FMT_RGB8) {
1247 if (c->dither == SWS_DITHER_AUTO)
1248 c->dither = (flags & SWS_FULL_CHR_H_INT) ? SWS_DITHER_ED : SWS_DITHER_BAYER;
1249 if (!(flags & SWS_FULL_CHR_H_INT)) {
1250 if (c->dither == SWS_DITHER_ED || c->dither == SWS_DITHER_A_DITHER || c->dither == SWS_DITHER_X_DITHER) {
1251 av_log(c, AV_LOG_DEBUG,
1252 "Desired dithering only supported in full chroma interpolation for destination format '%s'\n",
1253 av_get_pix_fmt_name(dstFormat));
1254 flags |= SWS_FULL_CHR_H_INT;
1258 if (flags & SWS_FULL_CHR_H_INT) {
1259 if (c->dither == SWS_DITHER_BAYER) {
1260 av_log(c, AV_LOG_DEBUG,
1261 "Ordered dither is not supported in full chroma interpolation for destination format '%s'\n",
1262 av_get_pix_fmt_name(dstFormat));
1263 c->dither = SWS_DITHER_ED;
1267 if (isPlanarRGB(dstFormat)) {
1268 if (!(flags & SWS_FULL_CHR_H_INT)) {
1269 av_log(c, AV_LOG_DEBUG,
1270 "%s output is not supported with half chroma resolution, switching to full\n",
1271 av_get_pix_fmt_name(dstFormat));
1272 flags |= SWS_FULL_CHR_H_INT;
1277 /* reuse chroma for 2 pixels RGB/BGR unless user wants full
1278 * chroma interpolation */
1279 if (flags & SWS_FULL_CHR_H_INT &&
1280 isAnyRGB(dstFormat) &&
1281 !isPlanarRGB(dstFormat) &&
1282 dstFormat != AV_PIX_FMT_RGBA64LE &&
1283 dstFormat != AV_PIX_FMT_RGBA64BE &&
1284 dstFormat != AV_PIX_FMT_BGRA64LE &&
1285 dstFormat != AV_PIX_FMT_BGRA64BE &&
1286 dstFormat != AV_PIX_FMT_RGB48LE &&
1287 dstFormat != AV_PIX_FMT_RGB48BE &&
1288 dstFormat != AV_PIX_FMT_BGR48LE &&
1289 dstFormat != AV_PIX_FMT_BGR48BE &&
1290 dstFormat != AV_PIX_FMT_RGBA &&
1291 dstFormat != AV_PIX_FMT_ARGB &&
1292 dstFormat != AV_PIX_FMT_BGRA &&
1293 dstFormat != AV_PIX_FMT_ABGR &&
1294 dstFormat != AV_PIX_FMT_RGB24 &&
1295 dstFormat != AV_PIX_FMT_BGR24 &&
1296 dstFormat != AV_PIX_FMT_BGR4_BYTE &&
1297 dstFormat != AV_PIX_FMT_RGB4_BYTE &&
1298 dstFormat != AV_PIX_FMT_BGR8 &&
1299 dstFormat != AV_PIX_FMT_RGB8
1301 av_log(c, AV_LOG_WARNING,
1302 "full chroma interpolation for destination format '%s' not yet implemented\n",
1303 av_get_pix_fmt_name(dstFormat));
1304 flags &= ~SWS_FULL_CHR_H_INT;
1307 if (isAnyRGB(dstFormat) && !(flags & SWS_FULL_CHR_H_INT))
1308 c->chrDstHSubSample = 1;
1310 // drop some chroma lines if the user wants it
1311 c->vChrDrop = (flags & SWS_SRC_V_CHR_DROP_MASK) >>
1312 SWS_SRC_V_CHR_DROP_SHIFT;
1313 c->chrSrcVSubSample += c->vChrDrop;
1315 /* drop every other pixel for chroma calculation unless user
1316 * wants full chroma */
1317 if (isAnyRGB(srcFormat) && !(flags & SWS_FULL_CHR_H_INP) &&
1318 srcFormat != AV_PIX_FMT_RGB8 && srcFormat != AV_PIX_FMT_BGR8 &&
1319 srcFormat != AV_PIX_FMT_RGB4 && srcFormat != AV_PIX_FMT_BGR4 &&
1320 srcFormat != AV_PIX_FMT_RGB4_BYTE && srcFormat != AV_PIX_FMT_BGR4_BYTE &&
1321 srcFormat != AV_PIX_FMT_GBRP9BE && srcFormat != AV_PIX_FMT_GBRP9LE &&
1322 srcFormat != AV_PIX_FMT_GBRP10BE && srcFormat != AV_PIX_FMT_GBRP10LE &&
1323 srcFormat != AV_PIX_FMT_GBRP12BE && srcFormat != AV_PIX_FMT_GBRP12LE &&
1324 srcFormat != AV_PIX_FMT_GBRP14BE && srcFormat != AV_PIX_FMT_GBRP14LE &&
1325 srcFormat != AV_PIX_FMT_GBRP16BE && srcFormat != AV_PIX_FMT_GBRP16LE &&
1326 ((dstW >> c->chrDstHSubSample) <= (srcW >> 1) ||
1327 (flags & SWS_FAST_BILINEAR)))
1328 c->chrSrcHSubSample = 1;
1330 // Note the FF_CEIL_RSHIFT is so that we always round toward +inf.
1331 c->chrSrcW = FF_CEIL_RSHIFT(srcW, c->chrSrcHSubSample);
1332 c->chrSrcH = FF_CEIL_RSHIFT(srcH, c->chrSrcVSubSample);
1333 c->chrDstW = FF_CEIL_RSHIFT(dstW, c->chrDstHSubSample);
1334 c->chrDstH = FF_CEIL_RSHIFT(dstH, c->chrDstVSubSample);
1336 FF_ALLOCZ_OR_GOTO(c, c->formatConvBuffer, FFALIGN(srcW*2+78, 16) * 2, fail);
1338 c->srcBpc = 1 + desc_src->comp[0].depth_minus1;
1341 c->dstBpc = 1 + desc_dst->comp[0].depth_minus1;
1344 if (isAnyRGB(srcFormat) || srcFormat == AV_PIX_FMT_PAL8)
1346 if (c->dstBpc == 16)
1349 if (INLINE_MMXEXT(cpu_flags) && c->srcBpc == 8 && c->dstBpc <= 14) {
1350 c->canMMXEXTBeUsed = dstW >= srcW && (dstW & 31) == 0 &&
1351 c->chrDstW >= c->chrSrcW &&
1353 if (!c->canMMXEXTBeUsed && dstW >= srcW && c->chrDstW >= c->chrSrcW && (srcW & 15) == 0
1355 && (flags & SWS_FAST_BILINEAR)) {
1356 if (flags & SWS_PRINT_INFO)
1357 av_log(c, AV_LOG_INFO,
1358 "output width is not a multiple of 32 -> no MMXEXT scaler\n");
1360 if (usesHFilter || isNBPS(c->srcFormat) || is16BPS(c->srcFormat) || isAnyRGB(c->srcFormat))
1361 c->canMMXEXTBeUsed = 0;
1363 c->canMMXEXTBeUsed = 0;
1365 c->chrXInc = (((int64_t)c->chrSrcW << 16) + (c->chrDstW >> 1)) / c->chrDstW;
1366 c->chrYInc = (((int64_t)c->chrSrcH << 16) + (c->chrDstH >> 1)) / c->chrDstH;
1368 /* Match pixel 0 of the src to pixel 0 of dst and match pixel n-2 of src
1369 * to pixel n-2 of dst, but only for the FAST_BILINEAR mode otherwise do
1371 * n-2 is the last chrominance sample available.
1372 * This is not perfect, but no one should notice the difference, the more
1373 * correct variant would be like the vertical one, but that would require
1374 * some special code for the first and last pixel */
1375 if (flags & SWS_FAST_BILINEAR) {
1376 if (c->canMMXEXTBeUsed) {
1380 // we don't use the x86 asm scaler if MMX is available
1381 else if (INLINE_MMX(cpu_flags) && c->dstBpc <= 14) {
1382 c->lumXInc = ((int64_t)(srcW - 2) << 16) / (dstW - 2) - 20;
1383 c->chrXInc = ((int64_t)(c->chrSrcW - 2) << 16) / (c->chrDstW - 2) - 20;
1387 // hardcoded for now
1388 c->gamma_value = 2.2;
1389 tmpFmt = AV_PIX_FMT_RGBA64LE;
1392 if (!unscaled && c->gamma_flag && (srcFormat != tmpFmt || dstFormat != tmpFmt)) {
1394 c->cascaded_context[0] = NULL;
1396 ret = av_image_alloc(c->cascaded_tmp, c->cascaded_tmpStride,
1397 srcW, srcH, tmpFmt, 64);
1401 c->cascaded_context[0] = sws_getContext(srcW, srcH, srcFormat,
1403 flags, NULL, NULL, c->param);
1404 if (!c->cascaded_context[0]) {
1408 c->cascaded_context[1] = sws_getContext(srcW, srcH, tmpFmt,
1410 flags, srcFilter, dstFilter, c->param);
1412 if (!c->cascaded_context[1])
1415 c2 = c->cascaded_context[1];
1416 c2->is_internal_gamma = 1;
1417 c2->gamma = alloc_gamma_tbl( c->gamma_value);
1418 c2->inv_gamma = alloc_gamma_tbl(1.f/c->gamma_value);
1419 if (!c2->gamma || !c2->inv_gamma)
1420 return AVERROR(ENOMEM);
1422 c->cascaded_context[2] = NULL;
1423 if (dstFormat != tmpFmt) {
1424 ret = av_image_alloc(c->cascaded1_tmp, c->cascaded1_tmpStride,
1425 dstW, dstH, tmpFmt, 64);
1429 c->cascaded_context[2] = sws_getContext(dstW, dstH, tmpFmt,
1430 dstW, dstH, dstFormat,
1431 flags, NULL, NULL, c->param);
1432 if (!c->cascaded_context[2])
1438 if (isBayer(srcFormat)) {
1440 (dstFormat != AV_PIX_FMT_RGB24 && dstFormat != AV_PIX_FMT_YUV420P)) {
1441 enum AVPixelFormat tmpFormat = AV_PIX_FMT_RGB24;
1443 ret = av_image_alloc(c->cascaded_tmp, c->cascaded_tmpStride,
1444 srcW, srcH, tmpFormat, 64);
1448 c->cascaded_context[0] = sws_getContext(srcW, srcH, srcFormat,
1449 srcW, srcH, tmpFormat,
1450 flags, srcFilter, NULL, c->param);
1451 if (!c->cascaded_context[0])
1454 c->cascaded_context[1] = sws_getContext(srcW, srcH, tmpFormat,
1455 dstW, dstH, dstFormat,
1456 flags, NULL, dstFilter, c->param);
1457 if (!c->cascaded_context[1])
1463 if (CONFIG_SWSCALE_ALPHA && isALPHA(srcFormat) && !isALPHA(dstFormat)) {
1464 enum AVPixelFormat tmpFormat = alphaless_fmt(srcFormat);
1466 if (tmpFormat != AV_PIX_FMT_NONE && c->alphablend != SWS_ALPHA_BLEND_NONE)
1468 dstFormat != tmpFormat ||
1469 usesHFilter || usesVFilter ||
1470 c->srcRange != c->dstRange
1472 ret = av_image_alloc(c->cascaded_tmp, c->cascaded_tmpStride,
1473 srcW, srcH, tmpFormat, 64);
1477 c->cascaded_context[0] = sws_alloc_set_opts(srcW, srcH, srcFormat,
1478 srcW, srcH, tmpFormat,
1480 if (!c->cascaded_context[0])
1482 c->cascaded_context[0]->alphablend = c->alphablend;
1483 ret = sws_init_context(c->cascaded_context[0], NULL , NULL);
1487 c->cascaded_context[1] = sws_getContext(srcW, srcH, tmpFormat,
1488 dstW, dstH, dstFormat,
1489 flags, srcFilter, dstFilter, c->param);
1490 if (!c->cascaded_context[1])
1496 #define USE_MMAP (HAVE_MMAP && HAVE_MPROTECT && defined MAP_ANONYMOUS)
1498 /* precalculate horizontal scaler filter coefficients */
1500 #if HAVE_MMXEXT_INLINE
1501 // can't downscale !!!
1502 if (c->canMMXEXTBeUsed && (flags & SWS_FAST_BILINEAR)) {
1503 c->lumMmxextFilterCodeSize = ff_init_hscaler_mmxext(dstW, c->lumXInc, NULL,
1505 c->chrMmxextFilterCodeSize = ff_init_hscaler_mmxext(c->chrDstW, c->chrXInc,
1506 NULL, NULL, NULL, 4);
1509 c->lumMmxextFilterCode = mmap(NULL, c->lumMmxextFilterCodeSize,
1510 PROT_READ | PROT_WRITE,
1511 MAP_PRIVATE | MAP_ANONYMOUS,
1513 c->chrMmxextFilterCode = mmap(NULL, c->chrMmxextFilterCodeSize,
1514 PROT_READ | PROT_WRITE,
1515 MAP_PRIVATE | MAP_ANONYMOUS,
1517 #elif HAVE_VIRTUALALLOC
1518 c->lumMmxextFilterCode = VirtualAlloc(NULL,
1519 c->lumMmxextFilterCodeSize,
1521 PAGE_EXECUTE_READWRITE);
1522 c->chrMmxextFilterCode = VirtualAlloc(NULL,
1523 c->chrMmxextFilterCodeSize,
1525 PAGE_EXECUTE_READWRITE);
1527 c->lumMmxextFilterCode = av_malloc(c->lumMmxextFilterCodeSize);
1528 c->chrMmxextFilterCode = av_malloc(c->chrMmxextFilterCodeSize);
1531 #ifdef MAP_ANONYMOUS
1532 if (c->lumMmxextFilterCode == MAP_FAILED || c->chrMmxextFilterCode == MAP_FAILED)
1534 if (!c->lumMmxextFilterCode || !c->chrMmxextFilterCode)
1537 av_log(c, AV_LOG_ERROR, "Failed to allocate MMX2FilterCode\n");
1538 return AVERROR(ENOMEM);
1541 FF_ALLOCZ_OR_GOTO(c, c->hLumFilter, (dstW / 8 + 8) * sizeof(int16_t), fail);
1542 FF_ALLOCZ_OR_GOTO(c, c->hChrFilter, (c->chrDstW / 4 + 8) * sizeof(int16_t), fail);
1543 FF_ALLOCZ_OR_GOTO(c, c->hLumFilterPos, (dstW / 2 / 8 + 8) * sizeof(int32_t), fail);
1544 FF_ALLOCZ_OR_GOTO(c, c->hChrFilterPos, (c->chrDstW / 2 / 4 + 8) * sizeof(int32_t), fail);
1546 ff_init_hscaler_mmxext( dstW, c->lumXInc, c->lumMmxextFilterCode,
1547 c->hLumFilter, (uint32_t*)c->hLumFilterPos, 8);
1548 ff_init_hscaler_mmxext(c->chrDstW, c->chrXInc, c->chrMmxextFilterCode,
1549 c->hChrFilter, (uint32_t*)c->hChrFilterPos, 4);
1552 if ( mprotect(c->lumMmxextFilterCode, c->lumMmxextFilterCodeSize, PROT_EXEC | PROT_READ) == -1
1553 || mprotect(c->chrMmxextFilterCode, c->chrMmxextFilterCodeSize, PROT_EXEC | PROT_READ) == -1) {
1554 av_log(c, AV_LOG_ERROR, "mprotect failed, cannot use fast bilinear scaler\n");
1559 #endif /* HAVE_MMXEXT_INLINE */
1561 const int filterAlign = X86_MMX(cpu_flags) ? 4 :
1562 PPC_ALTIVEC(cpu_flags) ? 8 : 1;
1564 if ((ret = initFilter(&c->hLumFilter, &c->hLumFilterPos,
1565 &c->hLumFilterSize, c->lumXInc,
1566 srcW, dstW, filterAlign, 1 << 14,
1567 (flags & SWS_BICUBLIN) ? (flags | SWS_BICUBIC) : flags,
1568 cpu_flags, srcFilter->lumH, dstFilter->lumH,
1570 get_local_pos(c, 0, 0, 0),
1571 get_local_pos(c, 0, 0, 0))) < 0)
1573 if ((ret = initFilter(&c->hChrFilter, &c->hChrFilterPos,
1574 &c->hChrFilterSize, c->chrXInc,
1575 c->chrSrcW, c->chrDstW, filterAlign, 1 << 14,
1576 (flags & SWS_BICUBLIN) ? (flags | SWS_BILINEAR) : flags,
1577 cpu_flags, srcFilter->chrH, dstFilter->chrH,
1579 get_local_pos(c, c->chrSrcHSubSample, c->src_h_chr_pos, 0),
1580 get_local_pos(c, c->chrDstHSubSample, c->dst_h_chr_pos, 0))) < 0)
1583 } // initialize horizontal stuff
1585 /* precalculate vertical scaler filter coefficients */
1587 const int filterAlign = X86_MMX(cpu_flags) ? 2 :
1588 PPC_ALTIVEC(cpu_flags) ? 8 : 1;
1590 if ((ret = initFilter(&c->vLumFilter, &c->vLumFilterPos, &c->vLumFilterSize,
1591 c->lumYInc, srcH, dstH, filterAlign, (1 << 12),
1592 (flags & SWS_BICUBLIN) ? (flags | SWS_BICUBIC) : flags,
1593 cpu_flags, srcFilter->lumV, dstFilter->lumV,
1595 get_local_pos(c, 0, 0, 1),
1596 get_local_pos(c, 0, 0, 1))) < 0)
1598 if ((ret = initFilter(&c->vChrFilter, &c->vChrFilterPos, &c->vChrFilterSize,
1599 c->chrYInc, c->chrSrcH, c->chrDstH,
1600 filterAlign, (1 << 12),
1601 (flags & SWS_BICUBLIN) ? (flags | SWS_BILINEAR) : flags,
1602 cpu_flags, srcFilter->chrV, dstFilter->chrV,
1604 get_local_pos(c, c->chrSrcVSubSample, c->src_v_chr_pos, 1),
1605 get_local_pos(c, c->chrDstVSubSample, c->dst_v_chr_pos, 1))) < 0)
1610 FF_ALLOC_OR_GOTO(c, c->vYCoeffsBank, sizeof(vector signed short) * c->vLumFilterSize * c->dstH, fail);
1611 FF_ALLOC_OR_GOTO(c, c->vCCoeffsBank, sizeof(vector signed short) * c->vChrFilterSize * c->chrDstH, fail);
1613 for (i = 0; i < c->vLumFilterSize * c->dstH; i++) {
1615 short *p = (short *)&c->vYCoeffsBank[i];
1616 for (j = 0; j < 8; j++)
1617 p[j] = c->vLumFilter[i];
1620 for (i = 0; i < c->vChrFilterSize * c->chrDstH; i++) {
1622 short *p = (short *)&c->vCCoeffsBank[i];
1623 for (j = 0; j < 8; j++)
1624 p[j] = c->vChrFilter[i];
1629 // calculate buffer sizes so that they won't run out while handling these damn slices
1630 c->vLumBufSize = c->vLumFilterSize;
1631 c->vChrBufSize = c->vChrFilterSize;
1632 for (i = 0; i < dstH; i++) {
1633 int chrI = (int64_t)i * c->chrDstH / dstH;
1634 int nextSlice = FFMAX(c->vLumFilterPos[i] + c->vLumFilterSize - 1,
1635 ((c->vChrFilterPos[chrI] + c->vChrFilterSize - 1)
1636 << c->chrSrcVSubSample));
1638 nextSlice >>= c->chrSrcVSubSample;
1639 nextSlice <<= c->chrSrcVSubSample;
1640 if (c->vLumFilterPos[i] + c->vLumBufSize < nextSlice)
1641 c->vLumBufSize = nextSlice - c->vLumFilterPos[i];
1642 if (c->vChrFilterPos[chrI] + c->vChrBufSize <
1643 (nextSlice >> c->chrSrcVSubSample))
1644 c->vChrBufSize = (nextSlice >> c->chrSrcVSubSample) -
1645 c->vChrFilterPos[chrI];
1648 for (i = 0; i < 4; i++)
1649 FF_ALLOCZ_OR_GOTO(c, c->dither_error[i], (c->dstW+2) * sizeof(int), fail);
1651 /* Allocate pixbufs (we use dynamic allocation because otherwise we would
1652 * need to allocate several megabytes to handle all possible cases) */
1653 FF_ALLOCZ_OR_GOTO(c, c->lumPixBuf, c->vLumBufSize * 3 * sizeof(int16_t *), fail);
1654 FF_ALLOCZ_OR_GOTO(c, c->chrUPixBuf, c->vChrBufSize * 3 * sizeof(int16_t *), fail);
1655 FF_ALLOCZ_OR_GOTO(c, c->chrVPixBuf, c->vChrBufSize * 3 * sizeof(int16_t *), fail);
1656 if (CONFIG_SWSCALE_ALPHA && isALPHA(c->srcFormat) && isALPHA(c->dstFormat))
1657 FF_ALLOCZ_OR_GOTO(c, c->alpPixBuf, c->vLumBufSize * 3 * sizeof(int16_t *), fail);
1658 /* Note we need at least one pixel more at the end because of the MMX code
1659 * (just in case someone wants to replace the 4000/8000). */
1660 /* align at 16 bytes for AltiVec */
1661 for (i = 0; i < c->vLumBufSize; i++) {
1662 FF_ALLOCZ_OR_GOTO(c, c->lumPixBuf[i + c->vLumBufSize],
1663 dst_stride + 16, fail);
1664 c->lumPixBuf[i] = c->lumPixBuf[i + c->vLumBufSize];
1666 // 64 / c->scalingBpp is the same as 16 / sizeof(scaling_intermediate)
1667 c->uv_off = (dst_stride>>1) + 64 / (c->dstBpc &~ 7);
1668 c->uv_offx2 = dst_stride + 16;
1669 for (i = 0; i < c->vChrBufSize; i++) {
1670 FF_ALLOC_OR_GOTO(c, c->chrUPixBuf[i + c->vChrBufSize],
1671 dst_stride * 2 + 32, fail);
1672 c->chrUPixBuf[i] = c->chrUPixBuf[i + c->vChrBufSize];
1673 c->chrVPixBuf[i] = c->chrVPixBuf[i + c->vChrBufSize]
1674 = c->chrUPixBuf[i] + (dst_stride >> 1) + 8;
1676 if (CONFIG_SWSCALE_ALPHA && c->alpPixBuf)
1677 for (i = 0; i < c->vLumBufSize; i++) {
1678 FF_ALLOCZ_OR_GOTO(c, c->alpPixBuf[i + c->vLumBufSize],
1679 dst_stride + 16, fail);
1680 c->alpPixBuf[i] = c->alpPixBuf[i + c->vLumBufSize];
1683 // try to avoid drawing green stuff between the right end and the stride end
1684 for (i = 0; i < c->vChrBufSize; i++)
1685 if(desc_dst->comp[0].depth_minus1 == 15){
1686 av_assert0(c->dstBpc > 14);
1687 for(j=0; j<dst_stride/2+1; j++)
1688 ((int32_t*)(c->chrUPixBuf[i]))[j] = 1<<18;
1690 for(j=0; j<dst_stride+1; j++)
1691 ((int16_t*)(c->chrUPixBuf[i]))[j] = 1<<14;
1693 av_assert0(c->chrDstH <= dstH);
1695 if (flags & SWS_PRINT_INFO) {
1696 const char *scaler = NULL, *cpucaps;
1698 for (i = 0; i < FF_ARRAY_ELEMS(scale_algorithms); i++) {
1699 if (flags & scale_algorithms[i].flag) {
1700 scaler = scale_algorithms[i].description;
1705 scaler = "ehh flags invalid?!";
1706 av_log(c, AV_LOG_INFO, "%s scaler, from %s to %s%s ",
1708 av_get_pix_fmt_name(srcFormat),
1710 dstFormat == AV_PIX_FMT_BGR555 || dstFormat == AV_PIX_FMT_BGR565 ||
1711 dstFormat == AV_PIX_FMT_RGB444BE || dstFormat == AV_PIX_FMT_RGB444LE ||
1712 dstFormat == AV_PIX_FMT_BGR444BE || dstFormat == AV_PIX_FMT_BGR444LE ?
1717 av_get_pix_fmt_name(dstFormat));
1719 if (INLINE_MMXEXT(cpu_flags))
1721 else if (INLINE_AMD3DNOW(cpu_flags))
1723 else if (INLINE_MMX(cpu_flags))
1725 else if (PPC_ALTIVEC(cpu_flags))
1726 cpucaps = "AltiVec";
1730 av_log(c, AV_LOG_INFO, "using %s\n", cpucaps);
1732 av_log(c, AV_LOG_VERBOSE, "%dx%d -> %dx%d\n", srcW, srcH, dstW, dstH);
1733 av_log(c, AV_LOG_DEBUG,
1734 "lum srcW=%d srcH=%d dstW=%d dstH=%d xInc=%d yInc=%d\n",
1735 c->srcW, c->srcH, c->dstW, c->dstH, c->lumXInc, c->lumYInc);
1736 av_log(c, AV_LOG_DEBUG,
1737 "chr srcW=%d srcH=%d dstW=%d dstH=%d xInc=%d yInc=%d\n",
1738 c->chrSrcW, c->chrSrcH, c->chrDstW, c->chrDstH,
1739 c->chrXInc, c->chrYInc);
1742 /* alpha blend special case, note this has been split via cascaded contexts if its scaled */
1743 if (unscaled && !usesHFilter && !usesVFilter &&
1744 c->alphablend != SWS_ALPHA_BLEND_NONE &&
1745 isALPHA(srcFormat) &&
1746 (c->srcRange == c->dstRange || isAnyRGB(dstFormat)) &&
1747 alphaless_fmt(srcFormat) == dstFormat
1749 c->swscale = ff_sws_alphablendaway;
1751 if (flags & SWS_PRINT_INFO)
1752 av_log(c, AV_LOG_INFO,
1753 "using alpha blendaway %s -> %s special converter\n",
1754 av_get_pix_fmt_name(srcFormat), av_get_pix_fmt_name(dstFormat));
1758 /* unscaled special cases */
1759 if (unscaled && !usesHFilter && !usesVFilter &&
1760 (c->srcRange == c->dstRange || isAnyRGB(dstFormat))) {
1761 ff_get_unscaled_swscale(c);
1764 if (flags & SWS_PRINT_INFO)
1765 av_log(c, AV_LOG_INFO,
1766 "using unscaled %s -> %s special converter\n",
1767 av_get_pix_fmt_name(srcFormat), av_get_pix_fmt_name(dstFormat));
1772 c->swscale = ff_getSwsFunc(c);
1773 return ff_init_filters(c);
1774 fail: // FIXME replace things by appropriate error codes
1775 if (ret == RETCODE_USE_CASCADE) {
1776 int tmpW = sqrt(srcW * (int64_t)dstW);
1777 int tmpH = sqrt(srcH * (int64_t)dstH);
1778 enum AVPixelFormat tmpFormat = AV_PIX_FMT_YUV420P;
1780 if (srcW*(int64_t)srcH <= 4LL*dstW*dstH)
1781 return AVERROR(EINVAL);
1783 ret = av_image_alloc(c->cascaded_tmp, c->cascaded_tmpStride,
1784 tmpW, tmpH, tmpFormat, 64);
1788 c->cascaded_context[0] = sws_getContext(srcW, srcH, srcFormat,
1789 tmpW, tmpH, tmpFormat,
1790 flags, srcFilter, NULL, c->param);
1791 if (!c->cascaded_context[0])
1794 c->cascaded_context[1] = sws_getContext(tmpW, tmpH, tmpFormat,
1795 dstW, dstH, dstFormat,
1796 flags, NULL, dstFilter, c->param);
1797 if (!c->cascaded_context[1])
1804 SwsContext *sws_alloc_set_opts(int srcW, int srcH, enum AVPixelFormat srcFormat,
1805 int dstW, int dstH, enum AVPixelFormat dstFormat,
1806 int flags, const double *param)
1810 if (!(c = sws_alloc_context()))
1818 c->srcFormat = srcFormat;
1819 c->dstFormat = dstFormat;
1822 c->param[0] = param[0];
1823 c->param[1] = param[1];
1829 SwsContext *sws_getContext(int srcW, int srcH, enum AVPixelFormat srcFormat,
1830 int dstW, int dstH, enum AVPixelFormat dstFormat,
1831 int flags, SwsFilter *srcFilter,
1832 SwsFilter *dstFilter, const double *param)
1836 c = sws_alloc_set_opts(srcW, srcH, srcFormat,
1837 dstW, dstH, dstFormat,
1842 if (sws_init_context(c, srcFilter, dstFilter) < 0) {
1850 static int isnan_vec(SwsVector *a)
1853 for (i=0; i<a->length; i++)
1854 if (isnan(a->coeff[i]))
1859 static void makenan_vec(SwsVector *a)
1862 for (i=0; i<a->length; i++)
1866 SwsFilter *sws_getDefaultFilter(float lumaGBlur, float chromaGBlur,
1867 float lumaSharpen, float chromaSharpen,
1868 float chromaHShift, float chromaVShift,
1871 SwsFilter *filter = av_malloc(sizeof(SwsFilter));
1875 if (lumaGBlur != 0.0) {
1876 filter->lumH = sws_getGaussianVec(lumaGBlur, 3.0);
1877 filter->lumV = sws_getGaussianVec(lumaGBlur, 3.0);
1879 filter->lumH = sws_getIdentityVec();
1880 filter->lumV = sws_getIdentityVec();
1883 if (chromaGBlur != 0.0) {
1884 filter->chrH = sws_getGaussianVec(chromaGBlur, 3.0);
1885 filter->chrV = sws_getGaussianVec(chromaGBlur, 3.0);
1887 filter->chrH = sws_getIdentityVec();
1888 filter->chrV = sws_getIdentityVec();
1891 if (!filter->lumH || !filter->lumV || !filter->chrH || !filter->chrV)
1894 if (chromaSharpen != 0.0) {
1895 SwsVector *id = sws_getIdentityVec();
1898 sws_scaleVec(filter->chrH, -chromaSharpen);
1899 sws_scaleVec(filter->chrV, -chromaSharpen);
1900 sws_addVec(filter->chrH, id);
1901 sws_addVec(filter->chrV, id);
1905 if (lumaSharpen != 0.0) {
1906 SwsVector *id = sws_getIdentityVec();
1909 sws_scaleVec(filter->lumH, -lumaSharpen);
1910 sws_scaleVec(filter->lumV, -lumaSharpen);
1911 sws_addVec(filter->lumH, id);
1912 sws_addVec(filter->lumV, id);
1916 if (chromaHShift != 0.0)
1917 sws_shiftVec(filter->chrH, (int)(chromaHShift + 0.5));
1919 if (chromaVShift != 0.0)
1920 sws_shiftVec(filter->chrV, (int)(chromaVShift + 0.5));
1922 sws_normalizeVec(filter->chrH, 1.0);
1923 sws_normalizeVec(filter->chrV, 1.0);
1924 sws_normalizeVec(filter->lumH, 1.0);
1925 sws_normalizeVec(filter->lumV, 1.0);
1927 if (isnan_vec(filter->chrH) ||
1928 isnan_vec(filter->chrV) ||
1929 isnan_vec(filter->lumH) ||
1930 isnan_vec(filter->lumV))
1934 sws_printVec2(filter->chrH, NULL, AV_LOG_DEBUG);
1936 sws_printVec2(filter->lumH, NULL, AV_LOG_DEBUG);
1941 sws_freeVec(filter->lumH);
1942 sws_freeVec(filter->lumV);
1943 sws_freeVec(filter->chrH);
1944 sws_freeVec(filter->chrV);
1949 SwsVector *sws_allocVec(int length)
1953 if(length <= 0 || length > INT_MAX/ sizeof(double))
1956 vec = av_malloc(sizeof(SwsVector));
1959 vec->length = length;
1960 vec->coeff = av_malloc(sizeof(double) * length);
1966 SwsVector *sws_getGaussianVec(double variance, double quality)
1968 const int length = (int)(variance * quality + 0.5) | 1;
1970 double middle = (length - 1) * 0.5;
1973 if(variance < 0 || quality < 0)
1976 vec = sws_allocVec(length);
1981 for (i = 0; i < length; i++) {
1982 double dist = i - middle;
1983 vec->coeff[i] = exp(-dist * dist / (2 * variance * variance)) /
1984 sqrt(2 * variance * M_PI);
1987 sws_normalizeVec(vec, 1.0);
1992 SwsVector *sws_getConstVec(double c, int length)
1995 SwsVector *vec = sws_allocVec(length);
2000 for (i = 0; i < length; i++)
2006 SwsVector *sws_getIdentityVec(void)
2008 return sws_getConstVec(1.0, 1);
2011 static double sws_dcVec(SwsVector *a)
2016 for (i = 0; i < a->length; i++)
2022 void sws_scaleVec(SwsVector *a, double scalar)
2026 for (i = 0; i < a->length; i++)
2027 a->coeff[i] *= scalar;
2030 void sws_normalizeVec(SwsVector *a, double height)
2032 sws_scaleVec(a, height / sws_dcVec(a));
2035 static SwsVector *sws_getConvVec(SwsVector *a, SwsVector *b)
2037 int length = a->length + b->length - 1;
2039 SwsVector *vec = sws_getConstVec(0.0, length);
2044 for (i = 0; i < a->length; i++) {
2045 for (j = 0; j < b->length; j++) {
2046 vec->coeff[i + j] += a->coeff[i] * b->coeff[j];
2053 static SwsVector *sws_sumVec(SwsVector *a, SwsVector *b)
2055 int length = FFMAX(a->length, b->length);
2057 SwsVector *vec = sws_getConstVec(0.0, length);
2062 for (i = 0; i < a->length; i++)
2063 vec->coeff[i + (length - 1) / 2 - (a->length - 1) / 2] += a->coeff[i];
2064 for (i = 0; i < b->length; i++)
2065 vec->coeff[i + (length - 1) / 2 - (b->length - 1) / 2] += b->coeff[i];
2070 static SwsVector *sws_diffVec(SwsVector *a, SwsVector *b)
2072 int length = FFMAX(a->length, b->length);
2074 SwsVector *vec = sws_getConstVec(0.0, length);
2079 for (i = 0; i < a->length; i++)
2080 vec->coeff[i + (length - 1) / 2 - (a->length - 1) / 2] += a->coeff[i];
2081 for (i = 0; i < b->length; i++)
2082 vec->coeff[i + (length - 1) / 2 - (b->length - 1) / 2] -= b->coeff[i];
2087 /* shift left / or right if "shift" is negative */
2088 static SwsVector *sws_getShiftedVec(SwsVector *a, int shift)
2090 int length = a->length + FFABS(shift) * 2;
2092 SwsVector *vec = sws_getConstVec(0.0, length);
2097 for (i = 0; i < a->length; i++) {
2098 vec->coeff[i + (length - 1) / 2 -
2099 (a->length - 1) / 2 - shift] = a->coeff[i];
2105 void sws_shiftVec(SwsVector *a, int shift)
2107 SwsVector *shifted = sws_getShiftedVec(a, shift);
2113 a->coeff = shifted->coeff;
2114 a->length = shifted->length;
2118 void sws_addVec(SwsVector *a, SwsVector *b)
2120 SwsVector *sum = sws_sumVec(a, b);
2126 a->coeff = sum->coeff;
2127 a->length = sum->length;
2131 void sws_subVec(SwsVector *a, SwsVector *b)
2133 SwsVector *diff = sws_diffVec(a, b);
2139 a->coeff = diff->coeff;
2140 a->length = diff->length;
2144 void sws_convVec(SwsVector *a, SwsVector *b)
2146 SwsVector *conv = sws_getConvVec(a, b);
2152 a->coeff = conv->coeff;
2153 a->length = conv->length;
2157 SwsVector *sws_cloneVec(SwsVector *a)
2159 SwsVector *vec = sws_allocVec(a->length);
2164 memcpy(vec->coeff, a->coeff, a->length * sizeof(*a->coeff));
2169 void sws_printVec2(SwsVector *a, AVClass *log_ctx, int log_level)
2176 for (i = 0; i < a->length; i++)
2177 if (a->coeff[i] > max)
2180 for (i = 0; i < a->length; i++)
2181 if (a->coeff[i] < min)
2186 for (i = 0; i < a->length; i++) {
2187 int x = (int)((a->coeff[i] - min) * 60.0 / range + 0.5);
2188 av_log(log_ctx, log_level, "%1.3f ", a->coeff[i]);
2190 av_log(log_ctx, log_level, " ");
2191 av_log(log_ctx, log_level, "|\n");
2195 void sws_freeVec(SwsVector *a)
2199 av_freep(&a->coeff);
2204 void sws_freeFilter(SwsFilter *filter)
2209 sws_freeVec(filter->lumH);
2210 sws_freeVec(filter->lumV);
2211 sws_freeVec(filter->chrH);
2212 sws_freeVec(filter->chrV);
2216 void sws_freeContext(SwsContext *c)
2223 for (i = 0; i < c->vLumBufSize; i++)
2224 av_freep(&c->lumPixBuf[i]);
2225 av_freep(&c->lumPixBuf);
2228 if (c->chrUPixBuf) {
2229 for (i = 0; i < c->vChrBufSize; i++)
2230 av_freep(&c->chrUPixBuf[i]);
2231 av_freep(&c->chrUPixBuf);
2232 av_freep(&c->chrVPixBuf);
2235 if (CONFIG_SWSCALE_ALPHA && c->alpPixBuf) {
2236 for (i = 0; i < c->vLumBufSize; i++)
2237 av_freep(&c->alpPixBuf[i]);
2238 av_freep(&c->alpPixBuf);
2241 for (i = 0; i < 4; i++)
2242 av_freep(&c->dither_error[i]);
2244 av_freep(&c->vLumFilter);
2245 av_freep(&c->vChrFilter);
2246 av_freep(&c->hLumFilter);
2247 av_freep(&c->hChrFilter);
2249 av_freep(&c->vYCoeffsBank);
2250 av_freep(&c->vCCoeffsBank);
2253 av_freep(&c->vLumFilterPos);
2254 av_freep(&c->vChrFilterPos);
2255 av_freep(&c->hLumFilterPos);
2256 av_freep(&c->hChrFilterPos);
2260 if (c->lumMmxextFilterCode)
2261 munmap(c->lumMmxextFilterCode, c->lumMmxextFilterCodeSize);
2262 if (c->chrMmxextFilterCode)
2263 munmap(c->chrMmxextFilterCode, c->chrMmxextFilterCodeSize);
2264 #elif HAVE_VIRTUALALLOC
2265 if (c->lumMmxextFilterCode)
2266 VirtualFree(c->lumMmxextFilterCode, 0, MEM_RELEASE);
2267 if (c->chrMmxextFilterCode)
2268 VirtualFree(c->chrMmxextFilterCode, 0, MEM_RELEASE);
2270 av_free(c->lumMmxextFilterCode);
2271 av_free(c->chrMmxextFilterCode);
2273 c->lumMmxextFilterCode = NULL;
2274 c->chrMmxextFilterCode = NULL;
2275 #endif /* HAVE_MMX_INLINE */
2277 av_freep(&c->yuvTable);
2278 av_freep(&c->formatConvBuffer);
2280 sws_freeContext(c->cascaded_context[0]);
2281 sws_freeContext(c->cascaded_context[1]);
2282 sws_freeContext(c->cascaded_context[2]);
2283 memset(c->cascaded_context, 0, sizeof(c->cascaded_context));
2284 av_freep(&c->cascaded_tmp[0]);
2285 av_freep(&c->cascaded1_tmp[0]);
2287 av_freep(&c->gamma);
2288 av_freep(&c->inv_gamma);
2295 struct SwsContext *sws_getCachedContext(struct SwsContext *context, int srcW,
2296 int srcH, enum AVPixelFormat srcFormat,
2298 enum AVPixelFormat dstFormat, int flags,
2299 SwsFilter *srcFilter,
2300 SwsFilter *dstFilter,
2301 const double *param)
2303 static const double default_param[2] = { SWS_PARAM_DEFAULT,
2304 SWS_PARAM_DEFAULT };
2305 int64_t src_h_chr_pos = -513, dst_h_chr_pos = -513,
2306 src_v_chr_pos = -513, dst_v_chr_pos = -513;
2309 param = default_param;
2312 (context->srcW != srcW ||
2313 context->srcH != srcH ||
2314 context->srcFormat != srcFormat ||
2315 context->dstW != dstW ||
2316 context->dstH != dstH ||
2317 context->dstFormat != dstFormat ||
2318 context->flags != flags ||
2319 context->param[0] != param[0] ||
2320 context->param[1] != param[1])) {
2322 av_opt_get_int(context, "src_h_chr_pos", 0, &src_h_chr_pos);
2323 av_opt_get_int(context, "src_v_chr_pos", 0, &src_v_chr_pos);
2324 av_opt_get_int(context, "dst_h_chr_pos", 0, &dst_h_chr_pos);
2325 av_opt_get_int(context, "dst_v_chr_pos", 0, &dst_v_chr_pos);
2326 sws_freeContext(context);
2331 if (!(context = sws_alloc_context()))
2333 context->srcW = srcW;
2334 context->srcH = srcH;
2335 context->srcFormat = srcFormat;
2336 context->dstW = dstW;
2337 context->dstH = dstH;
2338 context->dstFormat = dstFormat;
2339 context->flags = flags;
2340 context->param[0] = param[0];
2341 context->param[1] = param[1];
2343 av_opt_set_int(context, "src_h_chr_pos", src_h_chr_pos, 0);
2344 av_opt_set_int(context, "src_v_chr_pos", src_v_chr_pos, 0);
2345 av_opt_set_int(context, "dst_h_chr_pos", dst_h_chr_pos, 0);
2346 av_opt_set_int(context, "dst_v_chr_pos", dst_v_chr_pos, 0);
2348 if (sws_init_context(context, srcFilter, dstFilter) < 0) {
2349 sws_freeContext(context);