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 _DEFAULT_SOURCE
24 #define _SVID_SOURCE // needed for MAP_ANONYMOUS
25 #define _DARWIN_C_SOURCE // needed for MAP_ANON
32 #if defined(MAP_ANON) && !defined(MAP_ANONYMOUS)
33 #define MAP_ANONYMOUS MAP_ANON
37 #define WIN32_LEAN_AND_MEAN
41 #include "libavutil/attributes.h"
42 #include "libavutil/avassert.h"
43 #include "libavutil/avutil.h"
44 #include "libavutil/bswap.h"
45 #include "libavutil/cpu.h"
46 #include "libavutil/imgutils.h"
47 #include "libavutil/intreadwrite.h"
48 #include "libavutil/libm.h"
49 #include "libavutil/mathematics.h"
50 #include "libavutil/opt.h"
51 #include "libavutil/pixdesc.h"
52 #include "libavutil/aarch64/cpu.h"
53 #include "libavutil/ppc/cpu.h"
54 #include "libavutil/x86/asm.h"
55 #include "libavutil/x86/cpu.h"
57 // We have to implement deprecated functions until they are removed, this is the
58 // simplest way to prevent warnings
59 #undef attribute_deprecated
60 #define attribute_deprecated
64 #include "swscale_internal.h"
66 #if !FF_API_SWS_VECTOR
67 static SwsVector *sws_getIdentityVec(void);
68 static void sws_addVec(SwsVector *a, SwsVector *b);
69 static void sws_shiftVec(SwsVector *a, int shift);
70 static void sws_printVec2(SwsVector *a, AVClass *log_ctx, int log_level);
73 static void handle_formats(SwsContext *c);
75 unsigned swscale_version(void)
77 av_assert0(LIBSWSCALE_VERSION_MICRO >= 100);
78 return LIBSWSCALE_VERSION_INT;
81 const char *swscale_configuration(void)
83 return FFMPEG_CONFIGURATION;
86 const char *swscale_license(void)
88 #define LICENSE_PREFIX "libswscale license: "
89 return &LICENSE_PREFIX FFMPEG_LICENSE[sizeof(LICENSE_PREFIX) - 1];
92 typedef struct FormatEntry {
93 uint8_t is_supported_in :1;
94 uint8_t is_supported_out :1;
95 uint8_t is_supported_endianness :1;
98 static const FormatEntry format_entries[] = {
99 [AV_PIX_FMT_YUV420P] = { 1, 1 },
100 [AV_PIX_FMT_YUYV422] = { 1, 1 },
101 [AV_PIX_FMT_RGB24] = { 1, 1 },
102 [AV_PIX_FMT_BGR24] = { 1, 1 },
103 [AV_PIX_FMT_YUV422P] = { 1, 1 },
104 [AV_PIX_FMT_YUV444P] = { 1, 1 },
105 [AV_PIX_FMT_YUV410P] = { 1, 1 },
106 [AV_PIX_FMT_YUV411P] = { 1, 1 },
107 [AV_PIX_FMT_GRAY8] = { 1, 1 },
108 [AV_PIX_FMT_MONOWHITE] = { 1, 1 },
109 [AV_PIX_FMT_MONOBLACK] = { 1, 1 },
110 [AV_PIX_FMT_PAL8] = { 1, 0 },
111 [AV_PIX_FMT_YUVJ420P] = { 1, 1 },
112 [AV_PIX_FMT_YUVJ411P] = { 1, 1 },
113 [AV_PIX_FMT_YUVJ422P] = { 1, 1 },
114 [AV_PIX_FMT_YUVJ444P] = { 1, 1 },
115 [AV_PIX_FMT_YVYU422] = { 1, 1 },
116 [AV_PIX_FMT_UYVY422] = { 1, 1 },
117 [AV_PIX_FMT_UYYVYY411] = { 0, 0 },
118 [AV_PIX_FMT_BGR8] = { 1, 1 },
119 [AV_PIX_FMT_BGR4] = { 0, 1 },
120 [AV_PIX_FMT_BGR4_BYTE] = { 1, 1 },
121 [AV_PIX_FMT_RGB8] = { 1, 1 },
122 [AV_PIX_FMT_RGB4] = { 0, 1 },
123 [AV_PIX_FMT_RGB4_BYTE] = { 1, 1 },
124 [AV_PIX_FMT_NV12] = { 1, 1 },
125 [AV_PIX_FMT_NV21] = { 1, 1 },
126 [AV_PIX_FMT_ARGB] = { 1, 1 },
127 [AV_PIX_FMT_RGBA] = { 1, 1 },
128 [AV_PIX_FMT_ABGR] = { 1, 1 },
129 [AV_PIX_FMT_BGRA] = { 1, 1 },
130 [AV_PIX_FMT_0RGB] = { 1, 1 },
131 [AV_PIX_FMT_RGB0] = { 1, 1 },
132 [AV_PIX_FMT_0BGR] = { 1, 1 },
133 [AV_PIX_FMT_BGR0] = { 1, 1 },
134 [AV_PIX_FMT_GRAY9BE] = { 1, 1 },
135 [AV_PIX_FMT_GRAY9LE] = { 1, 1 },
136 [AV_PIX_FMT_GRAY10BE] = { 1, 1 },
137 [AV_PIX_FMT_GRAY10LE] = { 1, 1 },
138 [AV_PIX_FMT_GRAY12BE] = { 1, 1 },
139 [AV_PIX_FMT_GRAY12LE] = { 1, 1 },
140 [AV_PIX_FMT_GRAY14BE] = { 1, 1 },
141 [AV_PIX_FMT_GRAY14LE] = { 1, 1 },
142 [AV_PIX_FMT_GRAY16BE] = { 1, 1 },
143 [AV_PIX_FMT_GRAY16LE] = { 1, 1 },
144 [AV_PIX_FMT_YUV440P] = { 1, 1 },
145 [AV_PIX_FMT_YUVJ440P] = { 1, 1 },
146 [AV_PIX_FMT_YUV440P10LE] = { 1, 1 },
147 [AV_PIX_FMT_YUV440P10BE] = { 1, 1 },
148 [AV_PIX_FMT_YUV440P12LE] = { 1, 1 },
149 [AV_PIX_FMT_YUV440P12BE] = { 1, 1 },
150 [AV_PIX_FMT_YUVA420P] = { 1, 1 },
151 [AV_PIX_FMT_YUVA422P] = { 1, 1 },
152 [AV_PIX_FMT_YUVA444P] = { 1, 1 },
153 [AV_PIX_FMT_YUVA420P9BE] = { 1, 1 },
154 [AV_PIX_FMT_YUVA420P9LE] = { 1, 1 },
155 [AV_PIX_FMT_YUVA422P9BE] = { 1, 1 },
156 [AV_PIX_FMT_YUVA422P9LE] = { 1, 1 },
157 [AV_PIX_FMT_YUVA444P9BE] = { 1, 1 },
158 [AV_PIX_FMT_YUVA444P9LE] = { 1, 1 },
159 [AV_PIX_FMT_YUVA420P10BE]= { 1, 1 },
160 [AV_PIX_FMT_YUVA420P10LE]= { 1, 1 },
161 [AV_PIX_FMT_YUVA422P10BE]= { 1, 1 },
162 [AV_PIX_FMT_YUVA422P10LE]= { 1, 1 },
163 [AV_PIX_FMT_YUVA444P10BE]= { 1, 1 },
164 [AV_PIX_FMT_YUVA444P10LE]= { 1, 1 },
165 [AV_PIX_FMT_YUVA420P16BE]= { 1, 1 },
166 [AV_PIX_FMT_YUVA420P16LE]= { 1, 1 },
167 [AV_PIX_FMT_YUVA422P16BE]= { 1, 1 },
168 [AV_PIX_FMT_YUVA422P16LE]= { 1, 1 },
169 [AV_PIX_FMT_YUVA444P16BE]= { 1, 1 },
170 [AV_PIX_FMT_YUVA444P16LE]= { 1, 1 },
171 [AV_PIX_FMT_RGB48BE] = { 1, 1 },
172 [AV_PIX_FMT_RGB48LE] = { 1, 1 },
173 [AV_PIX_FMT_RGBA64BE] = { 1, 1, 1 },
174 [AV_PIX_FMT_RGBA64LE] = { 1, 1, 1 },
175 [AV_PIX_FMT_RGB565BE] = { 1, 1 },
176 [AV_PIX_FMT_RGB565LE] = { 1, 1 },
177 [AV_PIX_FMT_RGB555BE] = { 1, 1 },
178 [AV_PIX_FMT_RGB555LE] = { 1, 1 },
179 [AV_PIX_FMT_BGR565BE] = { 1, 1 },
180 [AV_PIX_FMT_BGR565LE] = { 1, 1 },
181 [AV_PIX_FMT_BGR555BE] = { 1, 1 },
182 [AV_PIX_FMT_BGR555LE] = { 1, 1 },
183 [AV_PIX_FMT_YUV420P16LE] = { 1, 1 },
184 [AV_PIX_FMT_YUV420P16BE] = { 1, 1 },
185 [AV_PIX_FMT_YUV422P16LE] = { 1, 1 },
186 [AV_PIX_FMT_YUV422P16BE] = { 1, 1 },
187 [AV_PIX_FMT_YUV444P16LE] = { 1, 1 },
188 [AV_PIX_FMT_YUV444P16BE] = { 1, 1 },
189 [AV_PIX_FMT_RGB444LE] = { 1, 1 },
190 [AV_PIX_FMT_RGB444BE] = { 1, 1 },
191 [AV_PIX_FMT_BGR444LE] = { 1, 1 },
192 [AV_PIX_FMT_BGR444BE] = { 1, 1 },
193 [AV_PIX_FMT_YA8] = { 1, 1 },
194 [AV_PIX_FMT_YA16BE] = { 1, 1 },
195 [AV_PIX_FMT_YA16LE] = { 1, 1 },
196 [AV_PIX_FMT_BGR48BE] = { 1, 1 },
197 [AV_PIX_FMT_BGR48LE] = { 1, 1 },
198 [AV_PIX_FMT_BGRA64BE] = { 1, 1, 1 },
199 [AV_PIX_FMT_BGRA64LE] = { 1, 1, 1 },
200 [AV_PIX_FMT_YUV420P9BE] = { 1, 1 },
201 [AV_PIX_FMT_YUV420P9LE] = { 1, 1 },
202 [AV_PIX_FMT_YUV420P10BE] = { 1, 1 },
203 [AV_PIX_FMT_YUV420P10LE] = { 1, 1 },
204 [AV_PIX_FMT_YUV420P12BE] = { 1, 1 },
205 [AV_PIX_FMT_YUV420P12LE] = { 1, 1 },
206 [AV_PIX_FMT_YUV420P14BE] = { 1, 1 },
207 [AV_PIX_FMT_YUV420P14LE] = { 1, 1 },
208 [AV_PIX_FMT_YUV422P9BE] = { 1, 1 },
209 [AV_PIX_FMT_YUV422P9LE] = { 1, 1 },
210 [AV_PIX_FMT_YUV422P10BE] = { 1, 1 },
211 [AV_PIX_FMT_YUV422P10LE] = { 1, 1 },
212 [AV_PIX_FMT_YUV422P12BE] = { 1, 1 },
213 [AV_PIX_FMT_YUV422P12LE] = { 1, 1 },
214 [AV_PIX_FMT_YUV422P14BE] = { 1, 1 },
215 [AV_PIX_FMT_YUV422P14LE] = { 1, 1 },
216 [AV_PIX_FMT_YUV444P9BE] = { 1, 1 },
217 [AV_PIX_FMT_YUV444P9LE] = { 1, 1 },
218 [AV_PIX_FMT_YUV444P10BE] = { 1, 1 },
219 [AV_PIX_FMT_YUV444P10LE] = { 1, 1 },
220 [AV_PIX_FMT_YUV444P12BE] = { 1, 1 },
221 [AV_PIX_FMT_YUV444P12LE] = { 1, 1 },
222 [AV_PIX_FMT_YUV444P14BE] = { 1, 1 },
223 [AV_PIX_FMT_YUV444P14LE] = { 1, 1 },
224 [AV_PIX_FMT_GBRP] = { 1, 1 },
225 [AV_PIX_FMT_GBRP9LE] = { 1, 1 },
226 [AV_PIX_FMT_GBRP9BE] = { 1, 1 },
227 [AV_PIX_FMT_GBRP10LE] = { 1, 1 },
228 [AV_PIX_FMT_GBRP10BE] = { 1, 1 },
229 [AV_PIX_FMT_GBRAP10LE] = { 1, 1 },
230 [AV_PIX_FMT_GBRAP10BE] = { 1, 1 },
231 [AV_PIX_FMT_GBRP12LE] = { 1, 1 },
232 [AV_PIX_FMT_GBRP12BE] = { 1, 1 },
233 [AV_PIX_FMT_GBRAP12LE] = { 1, 1 },
234 [AV_PIX_FMT_GBRAP12BE] = { 1, 1 },
235 [AV_PIX_FMT_GBRP14LE] = { 1, 1 },
236 [AV_PIX_FMT_GBRP14BE] = { 1, 1 },
237 [AV_PIX_FMT_GBRP16LE] = { 1, 1 },
238 [AV_PIX_FMT_GBRP16BE] = { 1, 1 },
239 [AV_PIX_FMT_GBRPF32LE] = { 1, 1 },
240 [AV_PIX_FMT_GBRPF32BE] = { 1, 1 },
241 [AV_PIX_FMT_GBRAPF32LE] = { 1, 1 },
242 [AV_PIX_FMT_GBRAPF32BE] = { 1, 1 },
243 [AV_PIX_FMT_GBRAP] = { 1, 1 },
244 [AV_PIX_FMT_GBRAP16LE] = { 1, 1 },
245 [AV_PIX_FMT_GBRAP16BE] = { 1, 1 },
246 [AV_PIX_FMT_BAYER_BGGR8] = { 1, 0 },
247 [AV_PIX_FMT_BAYER_RGGB8] = { 1, 0 },
248 [AV_PIX_FMT_BAYER_GBRG8] = { 1, 0 },
249 [AV_PIX_FMT_BAYER_GRBG8] = { 1, 0 },
250 [AV_PIX_FMT_BAYER_BGGR16LE] = { 1, 0 },
251 [AV_PIX_FMT_BAYER_BGGR16BE] = { 1, 0 },
252 [AV_PIX_FMT_BAYER_RGGB16LE] = { 1, 0 },
253 [AV_PIX_FMT_BAYER_RGGB16BE] = { 1, 0 },
254 [AV_PIX_FMT_BAYER_GBRG16LE] = { 1, 0 },
255 [AV_PIX_FMT_BAYER_GBRG16BE] = { 1, 0 },
256 [AV_PIX_FMT_BAYER_GRBG16LE] = { 1, 0 },
257 [AV_PIX_FMT_BAYER_GRBG16BE] = { 1, 0 },
258 [AV_PIX_FMT_XYZ12BE] = { 1, 1, 1 },
259 [AV_PIX_FMT_XYZ12LE] = { 1, 1, 1 },
260 [AV_PIX_FMT_AYUV64LE] = { 1, 1},
261 [AV_PIX_FMT_P010LE] = { 1, 1 },
262 [AV_PIX_FMT_P010BE] = { 1, 1 },
263 [AV_PIX_FMT_P016LE] = { 1, 1 },
264 [AV_PIX_FMT_P016BE] = { 1, 1 },
265 [AV_PIX_FMT_GRAYF32LE] = { 1, 1 },
266 [AV_PIX_FMT_GRAYF32BE] = { 1, 1 },
267 [AV_PIX_FMT_YUVA422P12BE] = { 1, 1 },
268 [AV_PIX_FMT_YUVA422P12LE] = { 1, 1 },
269 [AV_PIX_FMT_YUVA444P12BE] = { 1, 1 },
270 [AV_PIX_FMT_YUVA444P12LE] = { 1, 1 },
271 [AV_PIX_FMT_NV24] = { 1, 1 },
272 [AV_PIX_FMT_NV42] = { 1, 1 },
273 [AV_PIX_FMT_Y210LE] = { 1, 0 },
274 [AV_PIX_FMT_X2RGB10LE] = { 1, 1 },
277 int sws_isSupportedInput(enum AVPixelFormat pix_fmt)
279 return (unsigned)pix_fmt < FF_ARRAY_ELEMS(format_entries) ?
280 format_entries[pix_fmt].is_supported_in : 0;
283 int sws_isSupportedOutput(enum AVPixelFormat pix_fmt)
285 return (unsigned)pix_fmt < FF_ARRAY_ELEMS(format_entries) ?
286 format_entries[pix_fmt].is_supported_out : 0;
289 int sws_isSupportedEndiannessConversion(enum AVPixelFormat pix_fmt)
291 return (unsigned)pix_fmt < FF_ARRAY_ELEMS(format_entries) ?
292 format_entries[pix_fmt].is_supported_endianness : 0;
295 static double getSplineCoeff(double a, double b, double c, double d,
299 return ((d * dist + c) * dist + b) * dist + a;
301 return getSplineCoeff(0.0,
302 b + 2.0 * c + 3.0 * d,
304 -b - 3.0 * c - 6.0 * d,
308 static av_cold int get_local_pos(SwsContext *s, int chr_subsample, int pos, int dir)
310 if (pos == -1 || pos <= -513) {
311 pos = (128 << chr_subsample) - 128;
313 pos += 128; // relative to ideal left edge
314 return pos >> chr_subsample;
318 int flag; ///< flag associated to the algorithm
319 const char *description; ///< human-readable description
320 int size_factor; ///< size factor used when initing the filters
323 static const ScaleAlgorithm scale_algorithms[] = {
324 { SWS_AREA, "area averaging", 1 /* downscale only, for upscale it is bilinear */ },
325 { SWS_BICUBIC, "bicubic", 4 },
326 { SWS_BICUBLIN, "luma bicubic / chroma bilinear", -1 },
327 { SWS_BILINEAR, "bilinear", 2 },
328 { SWS_FAST_BILINEAR, "fast bilinear", -1 },
329 { SWS_GAUSS, "Gaussian", 8 /* infinite ;) */ },
330 { SWS_LANCZOS, "Lanczos", -1 /* custom */ },
331 { SWS_POINT, "nearest neighbor / point", -1 },
332 { SWS_SINC, "sinc", 20 /* infinite ;) */ },
333 { SWS_SPLINE, "bicubic spline", 20 /* infinite :)*/ },
334 { SWS_X, "experimental", 8 },
337 static av_cold int initFilter(int16_t **outFilter, int32_t **filterPos,
338 int *outFilterSize, int xInc, int srcW,
339 int dstW, int filterAlign, int one,
340 int flags, int cpu_flags,
341 SwsVector *srcFilter, SwsVector *dstFilter,
342 double param[2], int srcPos, int dstPos)
348 int64_t *filter = NULL;
349 int64_t *filter2 = NULL;
350 const int64_t fone = 1LL << (54 - FFMIN(av_log2(srcW/dstW), 8));
353 emms_c(); // FIXME should not be required but IS (even for non-MMX versions)
355 // NOTE: the +3 is for the MMX(+1) / SSE(+3) scaler which reads over the end
356 if (!FF_ALLOC_TYPED_ARRAY(*filterPos, dstW + 3))
359 if (FFABS(xInc - 0x10000) < 10 && srcPos == dstPos) { // unscaled
362 if (!FF_ALLOCZ_TYPED_ARRAY(filter, dstW * filterSize))
365 for (i = 0; i < dstW; i++) {
366 filter[i * filterSize] = fone;
369 } else if (flags & SWS_POINT) { // lame looking point sampling mode
373 if (!FF_ALLOC_TYPED_ARRAY(filter, dstW * filterSize))
376 xDstInSrc = ((dstPos*(int64_t)xInc)>>8) - ((srcPos*0x8000LL)>>7);
377 for (i = 0; i < dstW; i++) {
378 int xx = (xDstInSrc - ((filterSize - 1) << 15) + (1 << 15)) >> 16;
380 (*filterPos)[i] = xx;
384 } else if ((xInc <= (1 << 16) && (flags & SWS_AREA)) ||
385 (flags & SWS_FAST_BILINEAR)) { // bilinear upscale
389 if (!FF_ALLOC_TYPED_ARRAY(filter, dstW * filterSize))
392 xDstInSrc = ((dstPos*(int64_t)xInc)>>8) - ((srcPos*0x8000LL)>>7);
393 for (i = 0; i < dstW; i++) {
394 int xx = (xDstInSrc - ((filterSize - 1) << 15) + (1 << 15)) >> 16;
397 (*filterPos)[i] = xx;
398 // bilinear upscale / linear interpolate / area averaging
399 for (j = 0; j < filterSize; j++) {
400 int64_t coeff = fone - FFABS((int64_t)xx * (1 << 16) - xDstInSrc) * (fone >> 16);
403 filter[i * filterSize + j] = coeff;
412 for (i = 0; i < FF_ARRAY_ELEMS(scale_algorithms); i++) {
413 if (flags & scale_algorithms[i].flag && scale_algorithms[i].size_factor > 0) {
414 sizeFactor = scale_algorithms[i].size_factor;
418 if (flags & SWS_LANCZOS)
419 sizeFactor = param[0] != SWS_PARAM_DEFAULT ? ceil(2 * param[0]) : 6;
420 av_assert0(sizeFactor > 0);
423 filterSize = 1 + sizeFactor; // upscale
425 filterSize = 1 + (sizeFactor * srcW + dstW - 1) / dstW;
427 filterSize = FFMIN(filterSize, srcW - 2);
428 filterSize = FFMAX(filterSize, 1);
430 if (!FF_ALLOC_TYPED_ARRAY(filter, dstW * filterSize))
432 xDstInSrc = ((dstPos*(int64_t)xInc)>>7) - ((srcPos*0x10000LL)>>7);
433 for (i = 0; i < dstW; i++) {
434 int xx = (xDstInSrc - (filterSize - 2) * (1LL<<16)) / (1 << 17);
436 (*filterPos)[i] = xx;
437 for (j = 0; j < filterSize; j++) {
438 int64_t d = (FFABS(((int64_t)xx * (1 << 17)) - xDstInSrc)) << 13;
444 floatd = d * (1.0 / (1 << 30));
446 if (flags & SWS_BICUBIC) {
447 int64_t B = (param[0] != SWS_PARAM_DEFAULT ? param[0] : 0) * (1 << 24);
448 int64_t C = (param[1] != SWS_PARAM_DEFAULT ? param[1] : 0.6) * (1 << 24);
450 if (d >= 1LL << 31) {
453 int64_t dd = (d * d) >> 30;
454 int64_t ddd = (dd * d) >> 30;
457 coeff = (12 * (1 << 24) - 9 * B - 6 * C) * ddd +
458 (-18 * (1 << 24) + 12 * B + 6 * C) * dd +
459 (6 * (1 << 24) - 2 * B) * (1 << 30);
461 coeff = (-B - 6 * C) * ddd +
462 (6 * B + 30 * C) * dd +
463 (-12 * B - 48 * C) * d +
464 (8 * B + 24 * C) * (1 << 30);
466 coeff /= (1LL<<54)/fone;
467 } else if (flags & SWS_X) {
468 double A = param[0] != SWS_PARAM_DEFAULT ? param[0] : 1.0;
472 c = cos(floatd * M_PI);
479 coeff = (c * 0.5 + 0.5) * fone;
480 } else if (flags & SWS_AREA) {
481 int64_t d2 = d - (1 << 29);
482 if (d2 * xInc < -(1LL << (29 + 16)))
483 coeff = 1.0 * (1LL << (30 + 16));
484 else if (d2 * xInc < (1LL << (29 + 16)))
485 coeff = -d2 * xInc + (1LL << (29 + 16));
488 coeff *= fone >> (30 + 16);
489 } else if (flags & SWS_GAUSS) {
490 double p = param[0] != SWS_PARAM_DEFAULT ? param[0] : 3.0;
491 coeff = exp2(-p * floatd * floatd) * fone;
492 } else if (flags & SWS_SINC) {
493 coeff = (d ? sin(floatd * M_PI) / (floatd * M_PI) : 1.0) * fone;
494 } else if (flags & SWS_LANCZOS) {
495 double p = param[0] != SWS_PARAM_DEFAULT ? param[0] : 3.0;
496 coeff = (d ? sin(floatd * M_PI) * sin(floatd * M_PI / p) /
497 (floatd * floatd * M_PI * M_PI / p) : 1.0) * fone;
500 } else if (flags & SWS_BILINEAR) {
501 coeff = (1 << 30) - d;
505 } else if (flags & SWS_SPLINE) {
506 double p = -2.196152422706632;
507 coeff = getSplineCoeff(1.0, 0.0, p, -p - 1.0, floatd) * fone;
512 filter[i * filterSize + j] = coeff;
515 xDstInSrc += 2 * xInc;
519 /* apply src & dst Filter to filter -> filter2
522 av_assert0(filterSize > 0);
523 filter2Size = filterSize;
525 filter2Size += srcFilter->length - 1;
527 filter2Size += dstFilter->length - 1;
528 av_assert0(filter2Size > 0);
529 if (!FF_ALLOCZ_TYPED_ARRAY(filter2, dstW * filter2Size))
531 for (i = 0; i < dstW; i++) {
535 for (k = 0; k < srcFilter->length; k++) {
536 for (j = 0; j < filterSize; j++)
537 filter2[i * filter2Size + k + j] +=
538 srcFilter->coeff[k] * filter[i * filterSize + j];
541 for (j = 0; j < filterSize; j++)
542 filter2[i * filter2Size + j] = filter[i * filterSize + j];
546 (*filterPos)[i] += (filterSize - 1) / 2 - (filter2Size - 1) / 2;
550 /* try to reduce the filter-size (step1 find size and shift left) */
551 // Assume it is near normalized (*0.5 or *2.0 is OK but * 0.001 is not).
553 for (i = dstW - 1; i >= 0; i--) {
554 int min = filter2Size;
556 int64_t cutOff = 0.0;
558 /* get rid of near zero elements on the left by shifting left */
559 for (j = 0; j < filter2Size; j++) {
561 cutOff += FFABS(filter2[i * filter2Size]);
563 if (cutOff > SWS_MAX_REDUCE_CUTOFF * fone)
566 /* preserve monotonicity because the core can't handle the
567 * filter otherwise */
568 if (i < dstW - 1 && (*filterPos)[i] >= (*filterPos)[i + 1])
571 // move filter coefficients left
572 for (k = 1; k < filter2Size; k++)
573 filter2[i * filter2Size + k - 1] = filter2[i * filter2Size + k];
574 filter2[i * filter2Size + k - 1] = 0;
579 /* count near zeros on the right */
580 for (j = filter2Size - 1; j > 0; j--) {
581 cutOff += FFABS(filter2[i * filter2Size + j]);
583 if (cutOff > SWS_MAX_REDUCE_CUTOFF * fone)
588 if (min > minFilterSize)
592 if (PPC_ALTIVEC(cpu_flags)) {
593 // we can handle the special case 4, so we don't want to go the full 8
594 if (minFilterSize < 5)
597 /* We really don't want to waste our time doing useless computation, so
598 * fall back on the scalar C code for very small filters.
599 * Vectorizing is worth it only if you have a decent-sized vector. */
600 if (minFilterSize < 3)
604 if (HAVE_MMX && cpu_flags & AV_CPU_FLAG_MMX) {
605 // special case for unscaled vertical filtering
606 if (minFilterSize == 1 && filterAlign == 2)
610 av_assert0(minFilterSize > 0);
611 filterSize = (minFilterSize + (filterAlign - 1)) & (~(filterAlign - 1));
612 av_assert0(filterSize > 0);
613 filter = av_malloc_array(dstW, filterSize * sizeof(*filter));
616 if (filterSize >= MAX_FILTER_SIZE * 16 /
617 ((flags & SWS_ACCURATE_RND) ? APCK_SIZE : 16)) {
618 ret = RETCODE_USE_CASCADE;
621 *outFilterSize = filterSize;
623 if (flags & SWS_PRINT_INFO)
624 av_log(NULL, AV_LOG_VERBOSE,
625 "SwScaler: reducing / aligning filtersize %d -> %d\n",
626 filter2Size, filterSize);
627 /* try to reduce the filter-size (step2 reduce it) */
628 for (i = 0; i < dstW; i++) {
631 for (j = 0; j < filterSize; j++) {
632 if (j >= filter2Size)
633 filter[i * filterSize + j] = 0;
635 filter[i * filterSize + j] = filter2[i * filter2Size + j];
636 if ((flags & SWS_BITEXACT) && j >= minFilterSize)
637 filter[i * filterSize + j] = 0;
641 // FIXME try to align filterPos if possible
644 for (i = 0; i < dstW; i++) {
646 if ((*filterPos)[i] < 0) {
647 // move filter coefficients left to compensate for filterPos
648 for (j = 1; j < filterSize; j++) {
649 int left = FFMAX(j + (*filterPos)[i], 0);
650 filter[i * filterSize + left] += filter[i * filterSize + j];
651 filter[i * filterSize + j] = 0;
656 if ((*filterPos)[i] + filterSize > srcW) {
657 int shift = (*filterPos)[i] + FFMIN(filterSize - srcW, 0);
660 for (j = filterSize - 1; j >= 0; j--) {
661 if ((*filterPos)[i] + j >= srcW) {
662 acc += filter[i * filterSize + j];
663 filter[i * filterSize + j] = 0;
666 for (j = filterSize - 1; j >= 0; j--) {
668 filter[i * filterSize + j] = 0;
670 filter[i * filterSize + j] = filter[i * filterSize + j - shift];
674 (*filterPos)[i]-= shift;
675 filter[i * filterSize + srcW - 1 - (*filterPos)[i]] += acc;
677 av_assert0((*filterPos)[i] >= 0);
678 av_assert0((*filterPos)[i] < srcW);
679 if ((*filterPos)[i] + filterSize > srcW) {
680 for (j = 0; j < filterSize; j++) {
681 av_assert0((*filterPos)[i] + j < srcW || !filter[i * filterSize + j]);
686 // Note the +1 is for the MMX scaler which reads over the end
687 /* align at 16 for AltiVec (needed by hScale_altivec_real) */
688 if (!FF_ALLOCZ_TYPED_ARRAY(*outFilter, *outFilterSize * (dstW + 3)))
691 /* normalize & store in outFilter */
692 for (i = 0; i < dstW; i++) {
697 for (j = 0; j < filterSize; j++) {
698 sum += filter[i * filterSize + j];
700 sum = (sum + one / 2) / one;
702 av_log(NULL, AV_LOG_WARNING, "SwScaler: zero vector in scaling\n");
705 for (j = 0; j < *outFilterSize; j++) {
706 int64_t v = filter[i * filterSize + j] + error;
707 int intV = ROUNDED_DIV(v, sum);
708 (*outFilter)[i * (*outFilterSize) + j] = intV;
709 error = v - intV * sum;
713 (*filterPos)[dstW + 0] =
714 (*filterPos)[dstW + 1] =
715 (*filterPos)[dstW + 2] = (*filterPos)[dstW - 1]; /* the MMX/SSE scaler will
716 * read over the end */
717 for (i = 0; i < *outFilterSize; i++) {
718 int k = (dstW - 1) * (*outFilterSize) + i;
719 (*outFilter)[k + 1 * (*outFilterSize)] =
720 (*outFilter)[k + 2 * (*outFilterSize)] =
721 (*outFilter)[k + 3 * (*outFilterSize)] = (*outFilter)[k];
727 ret = AVERROR(ENOMEM);
730 av_log(NULL, ret == RETCODE_USE_CASCADE ? AV_LOG_DEBUG : AV_LOG_ERROR, "sws: initFilter failed\n");
737 static void fill_rgb2yuv_table(SwsContext *c, const int table[4], int dstRange)
739 int64_t W, V, Z, Cy, Cu, Cv;
740 int64_t vr = table[0];
741 int64_t ub = table[1];
742 int64_t ug = -table[2];
743 int64_t vg = -table[3];
746 uint8_t *p = (uint8_t*)c->input_rgb2yuv_table;
748 static const int8_t map[] = {
749 BY_IDX, GY_IDX, -1 , BY_IDX, BY_IDX, GY_IDX, -1 , BY_IDX,
750 RY_IDX, -1 , GY_IDX, RY_IDX, RY_IDX, -1 , GY_IDX, RY_IDX,
751 RY_IDX, GY_IDX, -1 , RY_IDX, RY_IDX, GY_IDX, -1 , RY_IDX,
752 BY_IDX, -1 , GY_IDX, BY_IDX, BY_IDX, -1 , GY_IDX, BY_IDX,
753 BU_IDX, GU_IDX, -1 , BU_IDX, BU_IDX, GU_IDX, -1 , BU_IDX,
754 RU_IDX, -1 , GU_IDX, RU_IDX, RU_IDX, -1 , GU_IDX, RU_IDX,
755 RU_IDX, GU_IDX, -1 , RU_IDX, RU_IDX, GU_IDX, -1 , RU_IDX,
756 BU_IDX, -1 , GU_IDX, BU_IDX, BU_IDX, -1 , GU_IDX, BU_IDX,
757 BV_IDX, GV_IDX, -1 , BV_IDX, BV_IDX, GV_IDX, -1 , BV_IDX,
758 RV_IDX, -1 , GV_IDX, RV_IDX, RV_IDX, -1 , GV_IDX, RV_IDX,
759 RV_IDX, GV_IDX, -1 , RV_IDX, RV_IDX, GV_IDX, -1 , RV_IDX,
760 BV_IDX, -1 , GV_IDX, BV_IDX, BV_IDX, -1 , GV_IDX, BV_IDX,
761 RY_IDX, BY_IDX, RY_IDX, BY_IDX, RY_IDX, BY_IDX, RY_IDX, BY_IDX,
762 BY_IDX, RY_IDX, BY_IDX, RY_IDX, BY_IDX, RY_IDX, BY_IDX, RY_IDX,
763 GY_IDX, -1 , GY_IDX, -1 , GY_IDX, -1 , GY_IDX, -1 ,
764 -1 , GY_IDX, -1 , GY_IDX, -1 , GY_IDX, -1 , GY_IDX,
765 RU_IDX, BU_IDX, RU_IDX, BU_IDX, RU_IDX, BU_IDX, RU_IDX, BU_IDX,
766 BU_IDX, RU_IDX, BU_IDX, RU_IDX, BU_IDX, RU_IDX, BU_IDX, RU_IDX,
767 GU_IDX, -1 , GU_IDX, -1 , GU_IDX, -1 , GU_IDX, -1 ,
768 -1 , GU_IDX, -1 , GU_IDX, -1 , GU_IDX, -1 , GU_IDX,
769 RV_IDX, BV_IDX, RV_IDX, BV_IDX, RV_IDX, BV_IDX, RV_IDX, BV_IDX,
770 BV_IDX, RV_IDX, BV_IDX, RV_IDX, BV_IDX, RV_IDX, BV_IDX, RV_IDX,
771 GV_IDX, -1 , GV_IDX, -1 , GV_IDX, -1 , GV_IDX, -1 ,
772 -1 , GV_IDX, -1 , GV_IDX, -1 , GV_IDX, -1 , GV_IDX, //23
773 -1 , -1 , -1 , -1 , -1 , -1 , -1 , -1 , //24
774 -1 , -1 , -1 , -1 , -1 , -1 , -1 , -1 , //25
775 -1 , -1 , -1 , -1 , -1 , -1 , -1 , -1 , //26
776 -1 , -1 , -1 , -1 , -1 , -1 , -1 , -1 , //27
777 -1 , -1 , -1 , -1 , -1 , -1 , -1 , -1 , //28
778 -1 , -1 , -1 , -1 , -1 , -1 , -1 , -1 , //29
779 -1 , -1 , -1 , -1 , -1 , -1 , -1 , -1 , //30
780 -1 , -1 , -1 , -1 , -1 , -1 , -1 , -1 , //31
781 BY_IDX, GY_IDX, RY_IDX, -1 , -1 , -1 , -1 , -1 , //32
782 BU_IDX, GU_IDX, RU_IDX, -1 , -1 , -1 , -1 , -1 , //33
783 BV_IDX, GV_IDX, RV_IDX, -1 , -1 , -1 , -1 , -1 , //34
786 dstRange = 0; //FIXME range = 1 is handled elsewhere
796 W = ROUNDED_DIV(ONE*ONE*ug, ub);
797 V = ROUNDED_DIV(ONE*ONE*vg, vr);
800 Cy = ROUNDED_DIV(cy*Z, ONE);
801 Cu = ROUNDED_DIV(ub*Z, ONE);
802 Cv = ROUNDED_DIV(vr*Z, ONE);
804 c->input_rgb2yuv_table[RY_IDX] = -ROUNDED_DIV((1 << RGB2YUV_SHIFT)*V , Cy);
805 c->input_rgb2yuv_table[GY_IDX] = ROUNDED_DIV((1 << RGB2YUV_SHIFT)*ONE*ONE , Cy);
806 c->input_rgb2yuv_table[BY_IDX] = -ROUNDED_DIV((1 << RGB2YUV_SHIFT)*W , Cy);
808 c->input_rgb2yuv_table[RU_IDX] = ROUNDED_DIV((1 << RGB2YUV_SHIFT)*V , Cu);
809 c->input_rgb2yuv_table[GU_IDX] = -ROUNDED_DIV((1 << RGB2YUV_SHIFT)*ONE*ONE , Cu);
810 c->input_rgb2yuv_table[BU_IDX] = ROUNDED_DIV((1 << RGB2YUV_SHIFT)*(Z+W) , Cu);
812 c->input_rgb2yuv_table[RV_IDX] = ROUNDED_DIV((1 << RGB2YUV_SHIFT)*(V+Z) , Cv);
813 c->input_rgb2yuv_table[GV_IDX] = -ROUNDED_DIV((1 << RGB2YUV_SHIFT)*ONE*ONE , Cv);
814 c->input_rgb2yuv_table[BV_IDX] = ROUNDED_DIV((1 << RGB2YUV_SHIFT)*W , Cv);
816 if(/*!dstRange && */!memcmp(table, ff_yuv2rgb_coeffs[SWS_CS_DEFAULT], sizeof(ff_yuv2rgb_coeffs[SWS_CS_DEFAULT]))) {
817 c->input_rgb2yuv_table[BY_IDX] = ((int)(0.114 * 219 / 255 * (1 << RGB2YUV_SHIFT) + 0.5));
818 c->input_rgb2yuv_table[BV_IDX] = (-(int)(0.081 * 224 / 255 * (1 << RGB2YUV_SHIFT) + 0.5));
819 c->input_rgb2yuv_table[BU_IDX] = ((int)(0.500 * 224 / 255 * (1 << RGB2YUV_SHIFT) + 0.5));
820 c->input_rgb2yuv_table[GY_IDX] = ((int)(0.587 * 219 / 255 * (1 << RGB2YUV_SHIFT) + 0.5));
821 c->input_rgb2yuv_table[GV_IDX] = (-(int)(0.419 * 224 / 255 * (1 << RGB2YUV_SHIFT) + 0.5));
822 c->input_rgb2yuv_table[GU_IDX] = (-(int)(0.331 * 224 / 255 * (1 << RGB2YUV_SHIFT) + 0.5));
823 c->input_rgb2yuv_table[RY_IDX] = ((int)(0.299 * 219 / 255 * (1 << RGB2YUV_SHIFT) + 0.5));
824 c->input_rgb2yuv_table[RV_IDX] = ((int)(0.500 * 224 / 255 * (1 << RGB2YUV_SHIFT) + 0.5));
825 c->input_rgb2yuv_table[RU_IDX] = (-(int)(0.169 * 224 / 255 * (1 << RGB2YUV_SHIFT) + 0.5));
827 for(i=0; i<FF_ARRAY_ELEMS(map); i++)
828 AV_WL16(p + 16*4 + 2*i, map[i] >= 0 ? c->input_rgb2yuv_table[map[i]] : 0);
831 static void fill_xyztables(struct SwsContext *c)
834 double xyzgamma = XYZ_GAMMA;
835 double rgbgamma = 1.0 / RGB_GAMMA;
836 double xyzgammainv = 1.0 / XYZ_GAMMA;
837 double rgbgammainv = RGB_GAMMA;
838 static const int16_t xyz2rgb_matrix[3][4] = {
839 {13270, -6295, -2041},
841 { 228, -835, 4329} };
842 static const int16_t rgb2xyz_matrix[3][4] = {
846 static int16_t xyzgamma_tab[4096], rgbgamma_tab[4096], xyzgammainv_tab[4096], rgbgammainv_tab[4096];
848 memcpy(c->xyz2rgb_matrix, xyz2rgb_matrix, sizeof(c->xyz2rgb_matrix));
849 memcpy(c->rgb2xyz_matrix, rgb2xyz_matrix, sizeof(c->rgb2xyz_matrix));
850 c->xyzgamma = xyzgamma_tab;
851 c->rgbgamma = rgbgamma_tab;
852 c->xyzgammainv = xyzgammainv_tab;
853 c->rgbgammainv = rgbgammainv_tab;
855 if (rgbgamma_tab[4095])
858 /* set gamma vectors */
859 for (i = 0; i < 4096; i++) {
860 xyzgamma_tab[i] = lrint(pow(i / 4095.0, xyzgamma) * 4095.0);
861 rgbgamma_tab[i] = lrint(pow(i / 4095.0, rgbgamma) * 4095.0);
862 xyzgammainv_tab[i] = lrint(pow(i / 4095.0, xyzgammainv) * 4095.0);
863 rgbgammainv_tab[i] = lrint(pow(i / 4095.0, rgbgammainv) * 4095.0);
867 static int range_override_needed(enum AVPixelFormat format)
869 return !isYUV(format) && !isGray(format);
872 int sws_setColorspaceDetails(struct SwsContext *c, const int inv_table[4],
873 int srcRange, const int table[4], int dstRange,
874 int brightness, int contrast, int saturation)
876 const AVPixFmtDescriptor *desc_dst;
877 const AVPixFmtDescriptor *desc_src;
881 desc_dst = av_pix_fmt_desc_get(c->dstFormat);
882 desc_src = av_pix_fmt_desc_get(c->srcFormat);
884 if(range_override_needed(c->dstFormat))
886 if(range_override_needed(c->srcFormat))
889 if (c->srcRange != srcRange ||
890 c->dstRange != dstRange ||
891 c->brightness != brightness ||
892 c->contrast != contrast ||
893 c->saturation != saturation ||
894 memcmp(c->srcColorspaceTable, inv_table, sizeof(int) * 4) ||
895 memcmp(c->dstColorspaceTable, table, sizeof(int) * 4)
899 memmove(c->srcColorspaceTable, inv_table, sizeof(int) * 4);
900 memmove(c->dstColorspaceTable, table, sizeof(int) * 4);
904 c->brightness = brightness;
905 c->contrast = contrast;
906 c->saturation = saturation;
907 c->srcRange = srcRange;
908 c->dstRange = dstRange;
910 //The srcBpc check is possibly wrong but we seem to lack a definitive reference to test this
911 //and what we have in ticket 2939 looks better with this check
912 if (need_reinit && (c->srcBpc == 8 || !isYUV(c->srcFormat)))
913 ff_sws_init_range_convert(c);
915 c->dstFormatBpp = av_get_bits_per_pixel(desc_dst);
916 c->srcFormatBpp = av_get_bits_per_pixel(desc_src);
918 if (c->cascaded_context[c->cascaded_mainindex])
919 return sws_setColorspaceDetails(c->cascaded_context[c->cascaded_mainindex],inv_table, srcRange,table, dstRange, brightness, contrast, saturation);
924 if ((isYUV(c->dstFormat) || isGray(c->dstFormat)) && (isYUV(c->srcFormat) || isGray(c->srcFormat))) {
925 if (!c->cascaded_context[0] &&
926 memcmp(c->dstColorspaceTable, c->srcColorspaceTable, sizeof(int) * 4) &&
927 c->srcW && c->srcH && c->dstW && c->dstH) {
928 enum AVPixelFormat tmp_format;
929 int tmp_width, tmp_height;
935 av_log(c, AV_LOG_VERBOSE, "YUV color matrix differs for YUV->YUV, using intermediate RGB to convert\n");
937 if (isNBPS(c->dstFormat) || is16BPS(c->dstFormat)) {
938 if (isALPHA(c->srcFormat) && isALPHA(c->dstFormat)) {
939 tmp_format = AV_PIX_FMT_BGRA64;
941 tmp_format = AV_PIX_FMT_BGR48;
944 if (isALPHA(c->srcFormat) && isALPHA(c->dstFormat)) {
945 tmp_format = AV_PIX_FMT_BGRA;
947 tmp_format = AV_PIX_FMT_BGR24;
951 if (srcW*srcH > dstW*dstH) {
959 ret = av_image_alloc(c->cascaded_tmp, c->cascaded_tmpStride,
960 tmp_width, tmp_height, tmp_format, 64);
964 c->cascaded_context[0] = sws_alloc_set_opts(srcW, srcH, c->srcFormat,
965 tmp_width, tmp_height, tmp_format,
967 if (!c->cascaded_context[0])
970 c->cascaded_context[0]->alphablend = c->alphablend;
971 ret = sws_init_context(c->cascaded_context[0], NULL , NULL);
974 //we set both src and dst depending on that the RGB side will be ignored
975 sws_setColorspaceDetails(c->cascaded_context[0], inv_table,
976 srcRange, table, dstRange,
977 brightness, contrast, saturation);
979 c->cascaded_context[1] = sws_getContext(tmp_width, tmp_height, tmp_format,
980 dstW, dstH, c->dstFormat,
981 c->flags, NULL, NULL, c->param);
982 if (!c->cascaded_context[1])
984 sws_setColorspaceDetails(c->cascaded_context[1], inv_table,
985 srcRange, table, dstRange,
986 0, 1 << 16, 1 << 16);
992 if (!isYUV(c->dstFormat) && !isGray(c->dstFormat)) {
993 ff_yuv2rgb_c_init_tables(c, inv_table, srcRange, brightness,
994 contrast, saturation);
998 ff_yuv2rgb_init_tables_ppc(c, inv_table, brightness,
999 contrast, saturation);
1002 fill_rgb2yuv_table(c, table, dstRange);
1007 int sws_getColorspaceDetails(struct SwsContext *c, int **inv_table,
1008 int *srcRange, int **table, int *dstRange,
1009 int *brightness, int *contrast, int *saturation)
1014 *inv_table = c->srcColorspaceTable;
1015 *table = c->dstColorspaceTable;
1016 *srcRange = range_override_needed(c->srcFormat) ? 1 : c->srcRange;
1017 *dstRange = range_override_needed(c->dstFormat) ? 1 : c->dstRange;
1018 *brightness = c->brightness;
1019 *contrast = c->contrast;
1020 *saturation = c->saturation;
1025 static int handle_jpeg(enum AVPixelFormat *format)
1028 case AV_PIX_FMT_YUVJ420P:
1029 *format = AV_PIX_FMT_YUV420P;
1031 case AV_PIX_FMT_YUVJ411P:
1032 *format = AV_PIX_FMT_YUV411P;
1034 case AV_PIX_FMT_YUVJ422P:
1035 *format = AV_PIX_FMT_YUV422P;
1037 case AV_PIX_FMT_YUVJ444P:
1038 *format = AV_PIX_FMT_YUV444P;
1040 case AV_PIX_FMT_YUVJ440P:
1041 *format = AV_PIX_FMT_YUV440P;
1043 case AV_PIX_FMT_GRAY8:
1044 case AV_PIX_FMT_YA8:
1045 case AV_PIX_FMT_GRAY9LE:
1046 case AV_PIX_FMT_GRAY9BE:
1047 case AV_PIX_FMT_GRAY10LE:
1048 case AV_PIX_FMT_GRAY10BE:
1049 case AV_PIX_FMT_GRAY12LE:
1050 case AV_PIX_FMT_GRAY12BE:
1051 case AV_PIX_FMT_GRAY14LE:
1052 case AV_PIX_FMT_GRAY14BE:
1053 case AV_PIX_FMT_GRAY16LE:
1054 case AV_PIX_FMT_GRAY16BE:
1055 case AV_PIX_FMT_YA16BE:
1056 case AV_PIX_FMT_YA16LE:
1063 static int handle_0alpha(enum AVPixelFormat *format)
1066 case AV_PIX_FMT_0BGR : *format = AV_PIX_FMT_ABGR ; return 1;
1067 case AV_PIX_FMT_BGR0 : *format = AV_PIX_FMT_BGRA ; return 4;
1068 case AV_PIX_FMT_0RGB : *format = AV_PIX_FMT_ARGB ; return 1;
1069 case AV_PIX_FMT_RGB0 : *format = AV_PIX_FMT_RGBA ; return 4;
1074 static int handle_xyz(enum AVPixelFormat *format)
1077 case AV_PIX_FMT_XYZ12BE : *format = AV_PIX_FMT_RGB48BE; return 1;
1078 case AV_PIX_FMT_XYZ12LE : *format = AV_PIX_FMT_RGB48LE; return 1;
1083 static void handle_formats(SwsContext *c)
1085 c->src0Alpha |= handle_0alpha(&c->srcFormat);
1086 c->dst0Alpha |= handle_0alpha(&c->dstFormat);
1087 c->srcXYZ |= handle_xyz(&c->srcFormat);
1088 c->dstXYZ |= handle_xyz(&c->dstFormat);
1089 if (c->srcXYZ || c->dstXYZ)
1093 SwsContext *sws_alloc_context(void)
1095 SwsContext *c = av_mallocz(sizeof(SwsContext));
1097 av_assert0(offsetof(SwsContext, redDither) + DITHER32_INT == offsetof(SwsContext, dither32));
1100 c->av_class = &ff_sws_context_class;
1101 av_opt_set_defaults(c);
1107 static uint16_t * alloc_gamma_tbl(double e)
1111 tbl = (uint16_t*)av_malloc(sizeof(uint16_t) * 1 << 16);
1115 for (i = 0; i < 65536; ++i) {
1116 tbl[i] = pow(i / 65535.0, e) * 65535.0;
1121 static enum AVPixelFormat alphaless_fmt(enum AVPixelFormat fmt)
1124 case AV_PIX_FMT_ARGB: return AV_PIX_FMT_RGB24;
1125 case AV_PIX_FMT_RGBA: return AV_PIX_FMT_RGB24;
1126 case AV_PIX_FMT_ABGR: return AV_PIX_FMT_BGR24;
1127 case AV_PIX_FMT_BGRA: return AV_PIX_FMT_BGR24;
1128 case AV_PIX_FMT_YA8: return AV_PIX_FMT_GRAY8;
1130 case AV_PIX_FMT_YUVA420P: return AV_PIX_FMT_YUV420P;
1131 case AV_PIX_FMT_YUVA422P: return AV_PIX_FMT_YUV422P;
1132 case AV_PIX_FMT_YUVA444P: return AV_PIX_FMT_YUV444P;
1134 case AV_PIX_FMT_GBRAP: return AV_PIX_FMT_GBRP;
1136 case AV_PIX_FMT_GBRAP10LE: return AV_PIX_FMT_GBRP10;
1137 case AV_PIX_FMT_GBRAP10BE: return AV_PIX_FMT_GBRP10;
1139 case AV_PIX_FMT_GBRAP12LE: return AV_PIX_FMT_GBRP12;
1140 case AV_PIX_FMT_GBRAP12BE: return AV_PIX_FMT_GBRP12;
1142 case AV_PIX_FMT_GBRAP16LE: return AV_PIX_FMT_GBRP16;
1143 case AV_PIX_FMT_GBRAP16BE: return AV_PIX_FMT_GBRP16;
1145 case AV_PIX_FMT_RGBA64LE: return AV_PIX_FMT_RGB48;
1146 case AV_PIX_FMT_RGBA64BE: return AV_PIX_FMT_RGB48;
1147 case AV_PIX_FMT_BGRA64LE: return AV_PIX_FMT_BGR48;
1148 case AV_PIX_FMT_BGRA64BE: return AV_PIX_FMT_BGR48;
1150 case AV_PIX_FMT_YA16BE: return AV_PIX_FMT_GRAY16;
1151 case AV_PIX_FMT_YA16LE: return AV_PIX_FMT_GRAY16;
1153 case AV_PIX_FMT_YUVA420P9BE: return AV_PIX_FMT_YUV420P9;
1154 case AV_PIX_FMT_YUVA422P9BE: return AV_PIX_FMT_YUV422P9;
1155 case AV_PIX_FMT_YUVA444P9BE: return AV_PIX_FMT_YUV444P9;
1156 case AV_PIX_FMT_YUVA420P9LE: return AV_PIX_FMT_YUV420P9;
1157 case AV_PIX_FMT_YUVA422P9LE: return AV_PIX_FMT_YUV422P9;
1158 case AV_PIX_FMT_YUVA444P9LE: return AV_PIX_FMT_YUV444P9;
1159 case AV_PIX_FMT_YUVA420P10BE: return AV_PIX_FMT_YUV420P10;
1160 case AV_PIX_FMT_YUVA422P10BE: return AV_PIX_FMT_YUV422P10;
1161 case AV_PIX_FMT_YUVA444P10BE: return AV_PIX_FMT_YUV444P10;
1162 case AV_PIX_FMT_YUVA420P10LE: return AV_PIX_FMT_YUV420P10;
1163 case AV_PIX_FMT_YUVA422P10LE: return AV_PIX_FMT_YUV422P10;
1164 case AV_PIX_FMT_YUVA444P10LE: return AV_PIX_FMT_YUV444P10;
1165 case AV_PIX_FMT_YUVA420P16BE: return AV_PIX_FMT_YUV420P16;
1166 case AV_PIX_FMT_YUVA422P16BE: return AV_PIX_FMT_YUV422P16;
1167 case AV_PIX_FMT_YUVA444P16BE: return AV_PIX_FMT_YUV444P16;
1168 case AV_PIX_FMT_YUVA420P16LE: return AV_PIX_FMT_YUV420P16;
1169 case AV_PIX_FMT_YUVA422P16LE: return AV_PIX_FMT_YUV422P16;
1170 case AV_PIX_FMT_YUVA444P16LE: return AV_PIX_FMT_YUV444P16;
1172 // case AV_PIX_FMT_AYUV64LE:
1173 // case AV_PIX_FMT_AYUV64BE:
1174 // case AV_PIX_FMT_PAL8:
1175 default: return AV_PIX_FMT_NONE;
1179 av_cold int sws_init_context(SwsContext *c, SwsFilter *srcFilter,
1180 SwsFilter *dstFilter)
1183 int usesVFilter, usesHFilter;
1185 SwsFilter dummyFilter = { NULL, NULL, NULL, NULL };
1190 int dst_stride = FFALIGN(dstW * sizeof(int16_t) + 66, 16);
1191 int flags, cpu_flags;
1192 enum AVPixelFormat srcFormat = c->srcFormat;
1193 enum AVPixelFormat dstFormat = c->dstFormat;
1194 const AVPixFmtDescriptor *desc_src;
1195 const AVPixFmtDescriptor *desc_dst;
1197 enum AVPixelFormat tmpFmt;
1198 static const float float_mult = 1.0f / 255.0f;
1200 cpu_flags = av_get_cpu_flags();
1204 ff_sws_rgb2rgb_init();
1206 unscaled = (srcW == dstW && srcH == dstH);
1208 c->srcRange |= handle_jpeg(&c->srcFormat);
1209 c->dstRange |= handle_jpeg(&c->dstFormat);
1211 if(srcFormat!=c->srcFormat || dstFormat!=c->dstFormat)
1212 av_log(c, AV_LOG_WARNING, "deprecated pixel format used, make sure you did set range correctly\n");
1214 if (!c->contrast && !c->saturation && !c->dstFormatBpp)
1215 sws_setColorspaceDetails(c, ff_yuv2rgb_coeffs[SWS_CS_DEFAULT], c->srcRange,
1216 ff_yuv2rgb_coeffs[SWS_CS_DEFAULT],
1217 c->dstRange, 0, 1 << 16, 1 << 16);
1220 srcFormat = c->srcFormat;
1221 dstFormat = c->dstFormat;
1222 desc_src = av_pix_fmt_desc_get(srcFormat);
1223 desc_dst = av_pix_fmt_desc_get(dstFormat);
1225 // If the source has no alpha then disable alpha blendaway
1227 c->alphablend = SWS_ALPHA_BLEND_NONE;
1229 if (!(unscaled && sws_isSupportedEndiannessConversion(srcFormat) &&
1230 av_pix_fmt_swap_endianness(srcFormat) == dstFormat)) {
1231 if (!sws_isSupportedInput(srcFormat)) {
1232 av_log(c, AV_LOG_ERROR, "%s is not supported as input pixel format\n",
1233 av_get_pix_fmt_name(srcFormat));
1234 return AVERROR(EINVAL);
1236 if (!sws_isSupportedOutput(dstFormat)) {
1237 av_log(c, AV_LOG_ERROR, "%s is not supported as output pixel format\n",
1238 av_get_pix_fmt_name(dstFormat));
1239 return AVERROR(EINVAL);
1242 av_assert2(desc_src && desc_dst);
1244 i = flags & (SWS_POINT |
1256 /* provide a default scaler if not set by caller */
1258 if (dstW < srcW && dstH < srcH)
1259 flags |= SWS_BICUBIC;
1260 else if (dstW > srcW && dstH > srcH)
1261 flags |= SWS_BICUBIC;
1263 flags |= SWS_BICUBIC;
1265 } else if (i & (i - 1)) {
1266 av_log(c, AV_LOG_ERROR,
1267 "Exactly one scaler algorithm must be chosen, got %X\n", i);
1268 return AVERROR(EINVAL);
1271 if (srcW < 1 || srcH < 1 || dstW < 1 || dstH < 1) {
1272 /* FIXME check if these are enough and try to lower them after
1273 * fixing the relevant parts of the code */
1274 av_log(c, AV_LOG_ERROR, "%dx%d -> %dx%d is invalid scaling dimension\n",
1275 srcW, srcH, dstW, dstH);
1276 return AVERROR(EINVAL);
1278 if (flags & SWS_FAST_BILINEAR) {
1279 if (srcW < 8 || dstW < 8) {
1280 flags ^= SWS_FAST_BILINEAR | SWS_BILINEAR;
1286 dstFilter = &dummyFilter;
1288 srcFilter = &dummyFilter;
1290 c->lumXInc = (((int64_t)srcW << 16) + (dstW >> 1)) / dstW;
1291 c->lumYInc = (((int64_t)srcH << 16) + (dstH >> 1)) / dstH;
1292 c->dstFormatBpp = av_get_bits_per_pixel(desc_dst);
1293 c->srcFormatBpp = av_get_bits_per_pixel(desc_src);
1294 c->vRounder = 4 * 0x0001000100010001ULL;
1296 usesVFilter = (srcFilter->lumV && srcFilter->lumV->length > 1) ||
1297 (srcFilter->chrV && srcFilter->chrV->length > 1) ||
1298 (dstFilter->lumV && dstFilter->lumV->length > 1) ||
1299 (dstFilter->chrV && dstFilter->chrV->length > 1);
1300 usesHFilter = (srcFilter->lumH && srcFilter->lumH->length > 1) ||
1301 (srcFilter->chrH && srcFilter->chrH->length > 1) ||
1302 (dstFilter->lumH && dstFilter->lumH->length > 1) ||
1303 (dstFilter->chrH && dstFilter->chrH->length > 1);
1305 av_pix_fmt_get_chroma_sub_sample(srcFormat, &c->chrSrcHSubSample, &c->chrSrcVSubSample);
1306 av_pix_fmt_get_chroma_sub_sample(dstFormat, &c->chrDstHSubSample, &c->chrDstVSubSample);
1308 if (isAnyRGB(dstFormat) && !(flags&SWS_FULL_CHR_H_INT)) {
1310 av_log(c, AV_LOG_DEBUG, "Forcing full internal H chroma due to odd output size\n");
1311 flags |= SWS_FULL_CHR_H_INT;
1315 if ( c->chrSrcHSubSample == 0
1316 && c->chrSrcVSubSample == 0
1317 && c->dither != SWS_DITHER_BAYER //SWS_FULL_CHR_H_INT is currently not supported with SWS_DITHER_BAYER
1318 && !(c->flags & SWS_FAST_BILINEAR)
1320 av_log(c, AV_LOG_DEBUG, "Forcing full internal H chroma due to input having non subsampled chroma\n");
1321 flags |= SWS_FULL_CHR_H_INT;
1326 if (c->dither == SWS_DITHER_AUTO) {
1327 if (flags & SWS_ERROR_DIFFUSION)
1328 c->dither = SWS_DITHER_ED;
1331 if(dstFormat == AV_PIX_FMT_BGR4_BYTE ||
1332 dstFormat == AV_PIX_FMT_RGB4_BYTE ||
1333 dstFormat == AV_PIX_FMT_BGR8 ||
1334 dstFormat == AV_PIX_FMT_RGB8) {
1335 if (c->dither == SWS_DITHER_AUTO)
1336 c->dither = (flags & SWS_FULL_CHR_H_INT) ? SWS_DITHER_ED : SWS_DITHER_BAYER;
1337 if (!(flags & SWS_FULL_CHR_H_INT)) {
1338 if (c->dither == SWS_DITHER_ED || c->dither == SWS_DITHER_A_DITHER || c->dither == SWS_DITHER_X_DITHER) {
1339 av_log(c, AV_LOG_DEBUG,
1340 "Desired dithering only supported in full chroma interpolation for destination format '%s'\n",
1341 av_get_pix_fmt_name(dstFormat));
1342 flags |= SWS_FULL_CHR_H_INT;
1346 if (flags & SWS_FULL_CHR_H_INT) {
1347 if (c->dither == SWS_DITHER_BAYER) {
1348 av_log(c, AV_LOG_DEBUG,
1349 "Ordered dither is not supported in full chroma interpolation for destination format '%s'\n",
1350 av_get_pix_fmt_name(dstFormat));
1351 c->dither = SWS_DITHER_ED;
1355 if (isPlanarRGB(dstFormat)) {
1356 if (!(flags & SWS_FULL_CHR_H_INT)) {
1357 av_log(c, AV_LOG_DEBUG,
1358 "%s output is not supported with half chroma resolution, switching to full\n",
1359 av_get_pix_fmt_name(dstFormat));
1360 flags |= SWS_FULL_CHR_H_INT;
1365 /* reuse chroma for 2 pixels RGB/BGR unless user wants full
1366 * chroma interpolation */
1367 if (flags & SWS_FULL_CHR_H_INT &&
1368 isAnyRGB(dstFormat) &&
1369 !isPlanarRGB(dstFormat) &&
1370 dstFormat != AV_PIX_FMT_RGBA64LE &&
1371 dstFormat != AV_PIX_FMT_RGBA64BE &&
1372 dstFormat != AV_PIX_FMT_BGRA64LE &&
1373 dstFormat != AV_PIX_FMT_BGRA64BE &&
1374 dstFormat != AV_PIX_FMT_RGB48LE &&
1375 dstFormat != AV_PIX_FMT_RGB48BE &&
1376 dstFormat != AV_PIX_FMT_BGR48LE &&
1377 dstFormat != AV_PIX_FMT_BGR48BE &&
1378 dstFormat != AV_PIX_FMT_RGBA &&
1379 dstFormat != AV_PIX_FMT_ARGB &&
1380 dstFormat != AV_PIX_FMT_BGRA &&
1381 dstFormat != AV_PIX_FMT_ABGR &&
1382 dstFormat != AV_PIX_FMT_RGB24 &&
1383 dstFormat != AV_PIX_FMT_BGR24 &&
1384 dstFormat != AV_PIX_FMT_BGR4_BYTE &&
1385 dstFormat != AV_PIX_FMT_RGB4_BYTE &&
1386 dstFormat != AV_PIX_FMT_BGR8 &&
1387 dstFormat != AV_PIX_FMT_RGB8
1389 av_log(c, AV_LOG_WARNING,
1390 "full chroma interpolation for destination format '%s' not yet implemented\n",
1391 av_get_pix_fmt_name(dstFormat));
1392 flags &= ~SWS_FULL_CHR_H_INT;
1395 if (isAnyRGB(dstFormat) && !(flags & SWS_FULL_CHR_H_INT))
1396 c->chrDstHSubSample = 1;
1398 // drop some chroma lines if the user wants it
1399 c->vChrDrop = (flags & SWS_SRC_V_CHR_DROP_MASK) >>
1400 SWS_SRC_V_CHR_DROP_SHIFT;
1401 c->chrSrcVSubSample += c->vChrDrop;
1403 /* drop every other pixel for chroma calculation unless user
1404 * wants full chroma */
1405 if (isAnyRGB(srcFormat) && !(flags & SWS_FULL_CHR_H_INP) &&
1406 srcFormat != AV_PIX_FMT_RGB8 && srcFormat != AV_PIX_FMT_BGR8 &&
1407 srcFormat != AV_PIX_FMT_RGB4 && srcFormat != AV_PIX_FMT_BGR4 &&
1408 srcFormat != AV_PIX_FMT_RGB4_BYTE && srcFormat != AV_PIX_FMT_BGR4_BYTE &&
1409 srcFormat != AV_PIX_FMT_GBRP9BE && srcFormat != AV_PIX_FMT_GBRP9LE &&
1410 srcFormat != AV_PIX_FMT_GBRP10BE && srcFormat != AV_PIX_FMT_GBRP10LE &&
1411 srcFormat != AV_PIX_FMT_GBRAP10BE && srcFormat != AV_PIX_FMT_GBRAP10LE &&
1412 srcFormat != AV_PIX_FMT_GBRP12BE && srcFormat != AV_PIX_FMT_GBRP12LE &&
1413 srcFormat != AV_PIX_FMT_GBRAP12BE && srcFormat != AV_PIX_FMT_GBRAP12LE &&
1414 srcFormat != AV_PIX_FMT_GBRP14BE && srcFormat != AV_PIX_FMT_GBRP14LE &&
1415 srcFormat != AV_PIX_FMT_GBRP16BE && srcFormat != AV_PIX_FMT_GBRP16LE &&
1416 srcFormat != AV_PIX_FMT_GBRAP16BE && srcFormat != AV_PIX_FMT_GBRAP16LE &&
1417 srcFormat != AV_PIX_FMT_GBRPF32BE && srcFormat != AV_PIX_FMT_GBRPF32LE &&
1418 srcFormat != AV_PIX_FMT_GBRAPF32BE && srcFormat != AV_PIX_FMT_GBRAPF32LE &&
1419 ((dstW >> c->chrDstHSubSample) <= (srcW >> 1) ||
1420 (flags & SWS_FAST_BILINEAR)))
1421 c->chrSrcHSubSample = 1;
1423 // Note the AV_CEIL_RSHIFT is so that we always round toward +inf.
1424 c->chrSrcW = AV_CEIL_RSHIFT(srcW, c->chrSrcHSubSample);
1425 c->chrSrcH = AV_CEIL_RSHIFT(srcH, c->chrSrcVSubSample);
1426 c->chrDstW = AV_CEIL_RSHIFT(dstW, c->chrDstHSubSample);
1427 c->chrDstH = AV_CEIL_RSHIFT(dstH, c->chrDstVSubSample);
1429 if (!FF_ALLOCZ_TYPED_ARRAY(c->formatConvBuffer, FFALIGN(srcW * 2 + 78, 16) * 2))
1432 c->srcBpc = desc_src->comp[0].depth;
1435 c->dstBpc = desc_dst->comp[0].depth;
1438 if (isAnyRGB(srcFormat) || srcFormat == AV_PIX_FMT_PAL8)
1440 if (c->dstBpc == 16)
1443 if (INLINE_MMXEXT(cpu_flags) && c->srcBpc == 8 && c->dstBpc <= 14) {
1444 c->canMMXEXTBeUsed = dstW >= srcW && (dstW & 31) == 0 &&
1445 c->chrDstW >= c->chrSrcW &&
1447 if (!c->canMMXEXTBeUsed && dstW >= srcW && c->chrDstW >= c->chrSrcW && (srcW & 15) == 0
1449 && (flags & SWS_FAST_BILINEAR)) {
1450 if (flags & SWS_PRINT_INFO)
1451 av_log(c, AV_LOG_INFO,
1452 "output width is not a multiple of 32 -> no MMXEXT scaler\n");
1454 if (usesHFilter || isNBPS(c->srcFormat) || is16BPS(c->srcFormat) || isAnyRGB(c->srcFormat))
1455 c->canMMXEXTBeUsed = 0;
1457 c->canMMXEXTBeUsed = 0;
1459 c->chrXInc = (((int64_t)c->chrSrcW << 16) + (c->chrDstW >> 1)) / c->chrDstW;
1460 c->chrYInc = (((int64_t)c->chrSrcH << 16) + (c->chrDstH >> 1)) / c->chrDstH;
1462 /* Match pixel 0 of the src to pixel 0 of dst and match pixel n-2 of src
1463 * to pixel n-2 of dst, but only for the FAST_BILINEAR mode otherwise do
1465 * n-2 is the last chrominance sample available.
1466 * This is not perfect, but no one should notice the difference, the more
1467 * correct variant would be like the vertical one, but that would require
1468 * some special code for the first and last pixel */
1469 if (flags & SWS_FAST_BILINEAR) {
1470 if (c->canMMXEXTBeUsed) {
1474 // we don't use the x86 asm scaler if MMX is available
1475 else if (INLINE_MMX(cpu_flags) && c->dstBpc <= 14) {
1476 c->lumXInc = ((int64_t)(srcW - 2) << 16) / (dstW - 2) - 20;
1477 c->chrXInc = ((int64_t)(c->chrSrcW - 2) << 16) / (c->chrDstW - 2) - 20;
1481 // hardcoded for now
1482 c->gamma_value = 2.2;
1483 tmpFmt = AV_PIX_FMT_RGBA64LE;
1486 if (!unscaled && c->gamma_flag && (srcFormat != tmpFmt || dstFormat != tmpFmt)) {
1488 c->cascaded_context[0] = NULL;
1490 ret = av_image_alloc(c->cascaded_tmp, c->cascaded_tmpStride,
1491 srcW, srcH, tmpFmt, 64);
1495 c->cascaded_context[0] = sws_getContext(srcW, srcH, srcFormat,
1497 flags, NULL, NULL, c->param);
1498 if (!c->cascaded_context[0]) {
1499 return AVERROR(ENOMEM);
1502 c->cascaded_context[1] = sws_getContext(srcW, srcH, tmpFmt,
1504 flags, srcFilter, dstFilter, c->param);
1506 if (!c->cascaded_context[1])
1507 return AVERROR(ENOMEM);
1509 c2 = c->cascaded_context[1];
1510 c2->is_internal_gamma = 1;
1511 c2->gamma = alloc_gamma_tbl( c->gamma_value);
1512 c2->inv_gamma = alloc_gamma_tbl(1.f/c->gamma_value);
1513 if (!c2->gamma || !c2->inv_gamma)
1514 return AVERROR(ENOMEM);
1516 // is_internal_flag is set after creating the context
1517 // to properly create the gamma convert FilterDescriptor
1518 // we have to re-initialize it
1519 ff_free_filters(c2);
1520 if ((ret = ff_init_filters(c2)) < 0) {
1521 sws_freeContext(c2);
1522 c->cascaded_context[1] = NULL;
1526 c->cascaded_context[2] = NULL;
1527 if (dstFormat != tmpFmt) {
1528 ret = av_image_alloc(c->cascaded1_tmp, c->cascaded1_tmpStride,
1529 dstW, dstH, tmpFmt, 64);
1533 c->cascaded_context[2] = sws_getContext(dstW, dstH, tmpFmt,
1534 dstW, dstH, dstFormat,
1535 flags, NULL, NULL, c->param);
1536 if (!c->cascaded_context[2])
1537 return AVERROR(ENOMEM);
1542 if (isBayer(srcFormat)) {
1544 (dstFormat != AV_PIX_FMT_RGB24 && dstFormat != AV_PIX_FMT_YUV420P &&
1545 dstFormat != AV_PIX_FMT_RGB48)) {
1546 enum AVPixelFormat tmpFormat = isBayer16BPS(srcFormat) ? AV_PIX_FMT_RGB48 : AV_PIX_FMT_RGB24;
1548 ret = av_image_alloc(c->cascaded_tmp, c->cascaded_tmpStride,
1549 srcW, srcH, tmpFormat, 64);
1553 c->cascaded_context[0] = sws_getContext(srcW, srcH, srcFormat,
1554 srcW, srcH, tmpFormat,
1555 flags, srcFilter, NULL, c->param);
1556 if (!c->cascaded_context[0])
1557 return AVERROR(ENOMEM);
1559 c->cascaded_context[1] = sws_getContext(srcW, srcH, tmpFormat,
1560 dstW, dstH, dstFormat,
1561 flags, NULL, dstFilter, c->param);
1562 if (!c->cascaded_context[1])
1563 return AVERROR(ENOMEM);
1568 if (unscaled && c->srcBpc == 8 && dstFormat == AV_PIX_FMT_GRAYF32){
1569 for (i = 0; i < 256; ++i){
1570 c->uint2float_lut[i] = (float)i * float_mult;
1574 // float will be converted to uint16_t
1575 if ((srcFormat == AV_PIX_FMT_GRAYF32BE || srcFormat == AV_PIX_FMT_GRAYF32LE) &&
1576 (!unscaled || unscaled && dstFormat != srcFormat && (srcFormat != AV_PIX_FMT_GRAYF32 ||
1577 dstFormat != AV_PIX_FMT_GRAY8))){
1581 if (CONFIG_SWSCALE_ALPHA && isALPHA(srcFormat) && !isALPHA(dstFormat)) {
1582 enum AVPixelFormat tmpFormat = alphaless_fmt(srcFormat);
1584 if (tmpFormat != AV_PIX_FMT_NONE && c->alphablend != SWS_ALPHA_BLEND_NONE) {
1586 dstFormat != tmpFormat ||
1587 usesHFilter || usesVFilter ||
1588 c->srcRange != c->dstRange
1590 c->cascaded_mainindex = 1;
1591 ret = av_image_alloc(c->cascaded_tmp, c->cascaded_tmpStride,
1592 srcW, srcH, tmpFormat, 64);
1596 c->cascaded_context[0] = sws_alloc_set_opts(srcW, srcH, srcFormat,
1597 srcW, srcH, tmpFormat,
1599 if (!c->cascaded_context[0])
1600 return AVERROR(EINVAL);
1601 c->cascaded_context[0]->alphablend = c->alphablend;
1602 ret = sws_init_context(c->cascaded_context[0], NULL , NULL);
1606 c->cascaded_context[1] = sws_alloc_set_opts(srcW, srcH, tmpFormat,
1607 dstW, dstH, dstFormat,
1609 if (!c->cascaded_context[1])
1610 return AVERROR(EINVAL);
1612 c->cascaded_context[1]->srcRange = c->srcRange;
1613 c->cascaded_context[1]->dstRange = c->dstRange;
1614 ret = sws_init_context(c->cascaded_context[1], srcFilter , dstFilter);
1623 #if HAVE_MMAP && HAVE_MPROTECT && defined(MAP_ANONYMOUS)
1629 /* precalculate horizontal scaler filter coefficients */
1631 #if HAVE_MMXEXT_INLINE
1632 // can't downscale !!!
1633 if (c->canMMXEXTBeUsed && (flags & SWS_FAST_BILINEAR)) {
1634 c->lumMmxextFilterCodeSize = ff_init_hscaler_mmxext(dstW, c->lumXInc, NULL,
1636 c->chrMmxextFilterCodeSize = ff_init_hscaler_mmxext(c->chrDstW, c->chrXInc,
1637 NULL, NULL, NULL, 4);
1640 c->lumMmxextFilterCode = mmap(NULL, c->lumMmxextFilterCodeSize,
1641 PROT_READ | PROT_WRITE,
1642 MAP_PRIVATE | MAP_ANONYMOUS,
1644 c->chrMmxextFilterCode = mmap(NULL, c->chrMmxextFilterCodeSize,
1645 PROT_READ | PROT_WRITE,
1646 MAP_PRIVATE | MAP_ANONYMOUS,
1648 #elif HAVE_VIRTUALALLOC
1649 c->lumMmxextFilterCode = VirtualAlloc(NULL,
1650 c->lumMmxextFilterCodeSize,
1652 PAGE_EXECUTE_READWRITE);
1653 c->chrMmxextFilterCode = VirtualAlloc(NULL,
1654 c->chrMmxextFilterCodeSize,
1656 PAGE_EXECUTE_READWRITE);
1658 c->lumMmxextFilterCode = av_malloc(c->lumMmxextFilterCodeSize);
1659 c->chrMmxextFilterCode = av_malloc(c->chrMmxextFilterCodeSize);
1662 #ifdef MAP_ANONYMOUS
1663 if (c->lumMmxextFilterCode == MAP_FAILED || c->chrMmxextFilterCode == MAP_FAILED)
1665 if (!c->lumMmxextFilterCode || !c->chrMmxextFilterCode)
1668 av_log(c, AV_LOG_ERROR, "Failed to allocate MMX2FilterCode\n");
1669 return AVERROR(ENOMEM);
1672 if (!FF_ALLOCZ_TYPED_ARRAY(c->hLumFilter, dstW / 8 + 8) ||
1673 !FF_ALLOCZ_TYPED_ARRAY(c->hChrFilter, c->chrDstW / 4 + 8) ||
1674 !FF_ALLOCZ_TYPED_ARRAY(c->hLumFilterPos, dstW / 2 / 8 + 8) ||
1675 !FF_ALLOCZ_TYPED_ARRAY(c->hChrFilterPos, c->chrDstW / 2 / 4 + 8))
1678 ff_init_hscaler_mmxext( dstW, c->lumXInc, c->lumMmxextFilterCode,
1679 c->hLumFilter, (uint32_t*)c->hLumFilterPos, 8);
1680 ff_init_hscaler_mmxext(c->chrDstW, c->chrXInc, c->chrMmxextFilterCode,
1681 c->hChrFilter, (uint32_t*)c->hChrFilterPos, 4);
1684 if ( mprotect(c->lumMmxextFilterCode, c->lumMmxextFilterCodeSize, PROT_EXEC | PROT_READ) == -1
1685 || mprotect(c->chrMmxextFilterCode, c->chrMmxextFilterCodeSize, PROT_EXEC | PROT_READ) == -1) {
1686 av_log(c, AV_LOG_ERROR, "mprotect failed, cannot use fast bilinear scaler\n");
1687 ret = AVERROR(EINVAL);
1692 #endif /* HAVE_MMXEXT_INLINE */
1694 const int filterAlign = X86_MMX(cpu_flags) ? 4 :
1695 PPC_ALTIVEC(cpu_flags) ? 8 :
1696 have_neon(cpu_flags) ? 8 : 1;
1698 if ((ret = initFilter(&c->hLumFilter, &c->hLumFilterPos,
1699 &c->hLumFilterSize, c->lumXInc,
1700 srcW, dstW, filterAlign, 1 << 14,
1701 (flags & SWS_BICUBLIN) ? (flags | SWS_BICUBIC) : flags,
1702 cpu_flags, srcFilter->lumH, dstFilter->lumH,
1704 get_local_pos(c, 0, 0, 0),
1705 get_local_pos(c, 0, 0, 0))) < 0)
1707 if ((ret = initFilter(&c->hChrFilter, &c->hChrFilterPos,
1708 &c->hChrFilterSize, c->chrXInc,
1709 c->chrSrcW, c->chrDstW, filterAlign, 1 << 14,
1710 (flags & SWS_BICUBLIN) ? (flags | SWS_BILINEAR) : flags,
1711 cpu_flags, srcFilter->chrH, dstFilter->chrH,
1713 get_local_pos(c, c->chrSrcHSubSample, c->src_h_chr_pos, 0),
1714 get_local_pos(c, c->chrDstHSubSample, c->dst_h_chr_pos, 0))) < 0)
1717 } // initialize horizontal stuff
1719 /* precalculate vertical scaler filter coefficients */
1721 const int filterAlign = X86_MMX(cpu_flags) ? 2 :
1722 PPC_ALTIVEC(cpu_flags) ? 8 :
1723 have_neon(cpu_flags) ? 2 : 1;
1725 if ((ret = initFilter(&c->vLumFilter, &c->vLumFilterPos, &c->vLumFilterSize,
1726 c->lumYInc, srcH, dstH, filterAlign, (1 << 12),
1727 (flags & SWS_BICUBLIN) ? (flags | SWS_BICUBIC) : flags,
1728 cpu_flags, srcFilter->lumV, dstFilter->lumV,
1730 get_local_pos(c, 0, 0, 1),
1731 get_local_pos(c, 0, 0, 1))) < 0)
1733 if ((ret = initFilter(&c->vChrFilter, &c->vChrFilterPos, &c->vChrFilterSize,
1734 c->chrYInc, c->chrSrcH, c->chrDstH,
1735 filterAlign, (1 << 12),
1736 (flags & SWS_BICUBLIN) ? (flags | SWS_BILINEAR) : flags,
1737 cpu_flags, srcFilter->chrV, dstFilter->chrV,
1739 get_local_pos(c, c->chrSrcVSubSample, c->src_v_chr_pos, 1),
1740 get_local_pos(c, c->chrDstVSubSample, c->dst_v_chr_pos, 1))) < 0)
1745 if (!FF_ALLOC_TYPED_ARRAY(c->vYCoeffsBank, c->vLumFilterSize * c->dstH) ||
1746 !FF_ALLOC_TYPED_ARRAY(c->vCCoeffsBank, c->vChrFilterSize * c->chrDstH))
1749 for (i = 0; i < c->vLumFilterSize * c->dstH; i++) {
1751 short *p = (short *)&c->vYCoeffsBank[i];
1752 for (j = 0; j < 8; j++)
1753 p[j] = c->vLumFilter[i];
1756 for (i = 0; i < c->vChrFilterSize * c->chrDstH; i++) {
1758 short *p = (short *)&c->vCCoeffsBank[i];
1759 for (j = 0; j < 8; j++)
1760 p[j] = c->vChrFilter[i];
1765 for (i = 0; i < 4; i++)
1766 if (!FF_ALLOCZ_TYPED_ARRAY(c->dither_error[i], c->dstW + 2))
1769 c->needAlpha = (CONFIG_SWSCALE_ALPHA && isALPHA(c->srcFormat) && isALPHA(c->dstFormat)) ? 1 : 0;
1771 // 64 / c->scalingBpp is the same as 16 / sizeof(scaling_intermediate)
1772 c->uv_off = (dst_stride>>1) + 64 / (c->dstBpc &~ 7);
1773 c->uv_offx2 = dst_stride + 16;
1775 av_assert0(c->chrDstH <= dstH);
1777 if (flags & SWS_PRINT_INFO) {
1778 const char *scaler = NULL, *cpucaps;
1780 for (i = 0; i < FF_ARRAY_ELEMS(scale_algorithms); i++) {
1781 if (flags & scale_algorithms[i].flag) {
1782 scaler = scale_algorithms[i].description;
1787 scaler = "ehh flags invalid?!";
1788 av_log(c, AV_LOG_INFO, "%s scaler, from %s to %s%s ",
1790 av_get_pix_fmt_name(srcFormat),
1792 dstFormat == AV_PIX_FMT_BGR555 || dstFormat == AV_PIX_FMT_BGR565 ||
1793 dstFormat == AV_PIX_FMT_RGB444BE || dstFormat == AV_PIX_FMT_RGB444LE ||
1794 dstFormat == AV_PIX_FMT_BGR444BE || dstFormat == AV_PIX_FMT_BGR444LE ?
1799 av_get_pix_fmt_name(dstFormat));
1801 if (INLINE_MMXEXT(cpu_flags))
1803 else if (INLINE_AMD3DNOW(cpu_flags))
1805 else if (INLINE_MMX(cpu_flags))
1807 else if (PPC_ALTIVEC(cpu_flags))
1808 cpucaps = "AltiVec";
1812 av_log(c, AV_LOG_INFO, "using %s\n", cpucaps);
1814 av_log(c, AV_LOG_VERBOSE, "%dx%d -> %dx%d\n", srcW, srcH, dstW, dstH);
1815 av_log(c, AV_LOG_DEBUG,
1816 "lum srcW=%d srcH=%d dstW=%d dstH=%d xInc=%d yInc=%d\n",
1817 c->srcW, c->srcH, c->dstW, c->dstH, c->lumXInc, c->lumYInc);
1818 av_log(c, AV_LOG_DEBUG,
1819 "chr srcW=%d srcH=%d dstW=%d dstH=%d xInc=%d yInc=%d\n",
1820 c->chrSrcW, c->chrSrcH, c->chrDstW, c->chrDstH,
1821 c->chrXInc, c->chrYInc);
1824 /* alpha blend special case, note this has been split via cascaded contexts if its scaled */
1825 if (unscaled && !usesHFilter && !usesVFilter &&
1826 c->alphablend != SWS_ALPHA_BLEND_NONE &&
1827 isALPHA(srcFormat) &&
1828 (c->srcRange == c->dstRange || isAnyRGB(dstFormat)) &&
1829 alphaless_fmt(srcFormat) == dstFormat
1831 c->swscale = ff_sws_alphablendaway;
1833 if (flags & SWS_PRINT_INFO)
1834 av_log(c, AV_LOG_INFO,
1835 "using alpha blendaway %s -> %s special converter\n",
1836 av_get_pix_fmt_name(srcFormat), av_get_pix_fmt_name(dstFormat));
1840 /* unscaled special cases */
1841 if (unscaled && !usesHFilter && !usesVFilter &&
1842 (c->srcRange == c->dstRange || isAnyRGB(dstFormat) ||
1843 isFloat(srcFormat) || isFloat(dstFormat))){
1844 ff_get_unscaled_swscale(c);
1847 if (flags & SWS_PRINT_INFO)
1848 av_log(c, AV_LOG_INFO,
1849 "using unscaled %s -> %s special converter\n",
1850 av_get_pix_fmt_name(srcFormat), av_get_pix_fmt_name(dstFormat));
1855 c->swscale = ff_getSwsFunc(c);
1856 return ff_init_filters(c);
1858 ret = AVERROR(ENOMEM);
1859 fail: // FIXME replace things by appropriate error codes
1860 if (ret == RETCODE_USE_CASCADE) {
1861 int tmpW = sqrt(srcW * (int64_t)dstW);
1862 int tmpH = sqrt(srcH * (int64_t)dstH);
1863 enum AVPixelFormat tmpFormat = AV_PIX_FMT_YUV420P;
1865 if (isALPHA(srcFormat))
1866 tmpFormat = AV_PIX_FMT_YUVA420P;
1868 if (srcW*(int64_t)srcH <= 4LL*dstW*dstH)
1869 return AVERROR(EINVAL);
1871 ret = av_image_alloc(c->cascaded_tmp, c->cascaded_tmpStride,
1872 tmpW, tmpH, tmpFormat, 64);
1876 c->cascaded_context[0] = sws_getContext(srcW, srcH, srcFormat,
1877 tmpW, tmpH, tmpFormat,
1878 flags, srcFilter, NULL, c->param);
1879 if (!c->cascaded_context[0])
1880 return AVERROR(ENOMEM);
1882 c->cascaded_context[1] = sws_getContext(tmpW, tmpH, tmpFormat,
1883 dstW, dstH, dstFormat,
1884 flags, NULL, dstFilter, c->param);
1885 if (!c->cascaded_context[1])
1886 return AVERROR(ENOMEM);
1892 SwsContext *sws_alloc_set_opts(int srcW, int srcH, enum AVPixelFormat srcFormat,
1893 int dstW, int dstH, enum AVPixelFormat dstFormat,
1894 int flags, const double *param)
1898 if (!(c = sws_alloc_context()))
1906 c->srcFormat = srcFormat;
1907 c->dstFormat = dstFormat;
1910 c->param[0] = param[0];
1911 c->param[1] = param[1];
1917 SwsContext *sws_getContext(int srcW, int srcH, enum AVPixelFormat srcFormat,
1918 int dstW, int dstH, enum AVPixelFormat dstFormat,
1919 int flags, SwsFilter *srcFilter,
1920 SwsFilter *dstFilter, const double *param)
1924 c = sws_alloc_set_opts(srcW, srcH, srcFormat,
1925 dstW, dstH, dstFormat,
1930 if (sws_init_context(c, srcFilter, dstFilter) < 0) {
1938 static int isnan_vec(SwsVector *a)
1941 for (i=0; i<a->length; i++)
1942 if (isnan(a->coeff[i]))
1947 static void makenan_vec(SwsVector *a)
1950 for (i=0; i<a->length; i++)
1954 SwsFilter *sws_getDefaultFilter(float lumaGBlur, float chromaGBlur,
1955 float lumaSharpen, float chromaSharpen,
1956 float chromaHShift, float chromaVShift,
1959 SwsFilter *filter = av_malloc(sizeof(SwsFilter));
1963 if (lumaGBlur != 0.0) {
1964 filter->lumH = sws_getGaussianVec(lumaGBlur, 3.0);
1965 filter->lumV = sws_getGaussianVec(lumaGBlur, 3.0);
1967 filter->lumH = sws_getIdentityVec();
1968 filter->lumV = sws_getIdentityVec();
1971 if (chromaGBlur != 0.0) {
1972 filter->chrH = sws_getGaussianVec(chromaGBlur, 3.0);
1973 filter->chrV = sws_getGaussianVec(chromaGBlur, 3.0);
1975 filter->chrH = sws_getIdentityVec();
1976 filter->chrV = sws_getIdentityVec();
1979 if (!filter->lumH || !filter->lumV || !filter->chrH || !filter->chrV)
1982 if (chromaSharpen != 0.0) {
1983 SwsVector *id = sws_getIdentityVec();
1986 sws_scaleVec(filter->chrH, -chromaSharpen);
1987 sws_scaleVec(filter->chrV, -chromaSharpen);
1988 sws_addVec(filter->chrH, id);
1989 sws_addVec(filter->chrV, id);
1993 if (lumaSharpen != 0.0) {
1994 SwsVector *id = sws_getIdentityVec();
1997 sws_scaleVec(filter->lumH, -lumaSharpen);
1998 sws_scaleVec(filter->lumV, -lumaSharpen);
1999 sws_addVec(filter->lumH, id);
2000 sws_addVec(filter->lumV, id);
2004 if (chromaHShift != 0.0)
2005 sws_shiftVec(filter->chrH, (int)(chromaHShift + 0.5));
2007 if (chromaVShift != 0.0)
2008 sws_shiftVec(filter->chrV, (int)(chromaVShift + 0.5));
2010 sws_normalizeVec(filter->chrH, 1.0);
2011 sws_normalizeVec(filter->chrV, 1.0);
2012 sws_normalizeVec(filter->lumH, 1.0);
2013 sws_normalizeVec(filter->lumV, 1.0);
2015 if (isnan_vec(filter->chrH) ||
2016 isnan_vec(filter->chrV) ||
2017 isnan_vec(filter->lumH) ||
2018 isnan_vec(filter->lumV))
2022 sws_printVec2(filter->chrH, NULL, AV_LOG_DEBUG);
2024 sws_printVec2(filter->lumH, NULL, AV_LOG_DEBUG);
2029 sws_freeVec(filter->lumH);
2030 sws_freeVec(filter->lumV);
2031 sws_freeVec(filter->chrH);
2032 sws_freeVec(filter->chrV);
2037 SwsVector *sws_allocVec(int length)
2041 if(length <= 0 || length > INT_MAX/ sizeof(double))
2044 vec = av_malloc(sizeof(SwsVector));
2047 vec->length = length;
2048 vec->coeff = av_malloc(sizeof(double) * length);
2054 SwsVector *sws_getGaussianVec(double variance, double quality)
2056 const int length = (int)(variance * quality + 0.5) | 1;
2058 double middle = (length - 1) * 0.5;
2061 if(variance < 0 || quality < 0)
2064 vec = sws_allocVec(length);
2069 for (i = 0; i < length; i++) {
2070 double dist = i - middle;
2071 vec->coeff[i] = exp(-dist * dist / (2 * variance * variance)) /
2072 sqrt(2 * variance * M_PI);
2075 sws_normalizeVec(vec, 1.0);
2081 * Allocate and return a vector with length coefficients, all
2082 * with the same value c.
2084 #if !FF_API_SWS_VECTOR
2087 SwsVector *sws_getConstVec(double c, int length)
2090 SwsVector *vec = sws_allocVec(length);
2095 for (i = 0; i < length; i++)
2102 * Allocate and return a vector with just one coefficient, with
2105 #if !FF_API_SWS_VECTOR
2108 SwsVector *sws_getIdentityVec(void)
2110 return sws_getConstVec(1.0, 1);
2113 static double sws_dcVec(SwsVector *a)
2118 for (i = 0; i < a->length; i++)
2124 void sws_scaleVec(SwsVector *a, double scalar)
2128 for (i = 0; i < a->length; i++)
2129 a->coeff[i] *= scalar;
2132 void sws_normalizeVec(SwsVector *a, double height)
2134 sws_scaleVec(a, height / sws_dcVec(a));
2137 #if FF_API_SWS_VECTOR
2138 static SwsVector *sws_getConvVec(SwsVector *a, SwsVector *b)
2140 int length = a->length + b->length - 1;
2142 SwsVector *vec = sws_getConstVec(0.0, length);
2147 for (i = 0; i < a->length; i++) {
2148 for (j = 0; j < b->length; j++) {
2149 vec->coeff[i + j] += a->coeff[i] * b->coeff[j];
2157 static SwsVector *sws_sumVec(SwsVector *a, SwsVector *b)
2159 int length = FFMAX(a->length, b->length);
2161 SwsVector *vec = sws_getConstVec(0.0, length);
2166 for (i = 0; i < a->length; i++)
2167 vec->coeff[i + (length - 1) / 2 - (a->length - 1) / 2] += a->coeff[i];
2168 for (i = 0; i < b->length; i++)
2169 vec->coeff[i + (length - 1) / 2 - (b->length - 1) / 2] += b->coeff[i];
2174 #if FF_API_SWS_VECTOR
2175 static SwsVector *sws_diffVec(SwsVector *a, SwsVector *b)
2177 int length = FFMAX(a->length, b->length);
2179 SwsVector *vec = sws_getConstVec(0.0, length);
2184 for (i = 0; i < a->length; i++)
2185 vec->coeff[i + (length - 1) / 2 - (a->length - 1) / 2] += a->coeff[i];
2186 for (i = 0; i < b->length; i++)
2187 vec->coeff[i + (length - 1) / 2 - (b->length - 1) / 2] -= b->coeff[i];
2193 /* shift left / or right if "shift" is negative */
2194 static SwsVector *sws_getShiftedVec(SwsVector *a, int shift)
2196 int length = a->length + FFABS(shift) * 2;
2198 SwsVector *vec = sws_getConstVec(0.0, length);
2203 for (i = 0; i < a->length; i++) {
2204 vec->coeff[i + (length - 1) / 2 -
2205 (a->length - 1) / 2 - shift] = a->coeff[i];
2211 #if !FF_API_SWS_VECTOR
2214 void sws_shiftVec(SwsVector *a, int shift)
2216 SwsVector *shifted = sws_getShiftedVec(a, shift);
2222 a->coeff = shifted->coeff;
2223 a->length = shifted->length;
2227 #if !FF_API_SWS_VECTOR
2230 void sws_addVec(SwsVector *a, SwsVector *b)
2232 SwsVector *sum = sws_sumVec(a, b);
2238 a->coeff = sum->coeff;
2239 a->length = sum->length;
2243 #if FF_API_SWS_VECTOR
2244 void sws_subVec(SwsVector *a, SwsVector *b)
2246 SwsVector *diff = sws_diffVec(a, b);
2252 a->coeff = diff->coeff;
2253 a->length = diff->length;
2257 void sws_convVec(SwsVector *a, SwsVector *b)
2259 SwsVector *conv = sws_getConvVec(a, b);
2265 a->coeff = conv->coeff;
2266 a->length = conv->length;
2270 SwsVector *sws_cloneVec(SwsVector *a)
2272 SwsVector *vec = sws_allocVec(a->length);
2277 memcpy(vec->coeff, a->coeff, a->length * sizeof(*a->coeff));
2284 * Print with av_log() a textual representation of the vector a
2285 * if log_level <= av_log_level.
2287 #if !FF_API_SWS_VECTOR
2290 void sws_printVec2(SwsVector *a, AVClass *log_ctx, int log_level)
2297 for (i = 0; i < a->length; i++)
2298 if (a->coeff[i] > max)
2301 for (i = 0; i < a->length; i++)
2302 if (a->coeff[i] < min)
2307 for (i = 0; i < a->length; i++) {
2308 int x = (int)((a->coeff[i] - min) * 60.0 / range + 0.5);
2309 av_log(log_ctx, log_level, "%1.3f ", a->coeff[i]);
2311 av_log(log_ctx, log_level, " ");
2312 av_log(log_ctx, log_level, "|\n");
2316 void sws_freeVec(SwsVector *a)
2320 av_freep(&a->coeff);
2325 void sws_freeFilter(SwsFilter *filter)
2330 sws_freeVec(filter->lumH);
2331 sws_freeVec(filter->lumV);
2332 sws_freeVec(filter->chrH);
2333 sws_freeVec(filter->chrV);
2337 void sws_freeContext(SwsContext *c)
2343 for (i = 0; i < 4; i++)
2344 av_freep(&c->dither_error[i]);
2346 av_freep(&c->vLumFilter);
2347 av_freep(&c->vChrFilter);
2348 av_freep(&c->hLumFilter);
2349 av_freep(&c->hChrFilter);
2351 av_freep(&c->vYCoeffsBank);
2352 av_freep(&c->vCCoeffsBank);
2355 av_freep(&c->vLumFilterPos);
2356 av_freep(&c->vChrFilterPos);
2357 av_freep(&c->hLumFilterPos);
2358 av_freep(&c->hChrFilterPos);
2362 if (c->lumMmxextFilterCode)
2363 munmap(c->lumMmxextFilterCode, c->lumMmxextFilterCodeSize);
2364 if (c->chrMmxextFilterCode)
2365 munmap(c->chrMmxextFilterCode, c->chrMmxextFilterCodeSize);
2366 #elif HAVE_VIRTUALALLOC
2367 if (c->lumMmxextFilterCode)
2368 VirtualFree(c->lumMmxextFilterCode, 0, MEM_RELEASE);
2369 if (c->chrMmxextFilterCode)
2370 VirtualFree(c->chrMmxextFilterCode, 0, MEM_RELEASE);
2372 av_free(c->lumMmxextFilterCode);
2373 av_free(c->chrMmxextFilterCode);
2375 c->lumMmxextFilterCode = NULL;
2376 c->chrMmxextFilterCode = NULL;
2377 #endif /* HAVE_MMX_INLINE */
2379 av_freep(&c->yuvTable);
2380 av_freep(&c->formatConvBuffer);
2382 sws_freeContext(c->cascaded_context[0]);
2383 sws_freeContext(c->cascaded_context[1]);
2384 sws_freeContext(c->cascaded_context[2]);
2385 memset(c->cascaded_context, 0, sizeof(c->cascaded_context));
2386 av_freep(&c->cascaded_tmp[0]);
2387 av_freep(&c->cascaded1_tmp[0]);
2389 av_freep(&c->gamma);
2390 av_freep(&c->inv_gamma);
2397 struct SwsContext *sws_getCachedContext(struct SwsContext *context, int srcW,
2398 int srcH, enum AVPixelFormat srcFormat,
2400 enum AVPixelFormat dstFormat, int flags,
2401 SwsFilter *srcFilter,
2402 SwsFilter *dstFilter,
2403 const double *param)
2405 static const double default_param[2] = { SWS_PARAM_DEFAULT,
2406 SWS_PARAM_DEFAULT };
2407 int64_t src_h_chr_pos = -513, dst_h_chr_pos = -513,
2408 src_v_chr_pos = -513, dst_v_chr_pos = -513;
2411 param = default_param;
2414 (context->srcW != srcW ||
2415 context->srcH != srcH ||
2416 context->srcFormat != srcFormat ||
2417 context->dstW != dstW ||
2418 context->dstH != dstH ||
2419 context->dstFormat != dstFormat ||
2420 context->flags != flags ||
2421 context->param[0] != param[0] ||
2422 context->param[1] != param[1])) {
2424 av_opt_get_int(context, "src_h_chr_pos", 0, &src_h_chr_pos);
2425 av_opt_get_int(context, "src_v_chr_pos", 0, &src_v_chr_pos);
2426 av_opt_get_int(context, "dst_h_chr_pos", 0, &dst_h_chr_pos);
2427 av_opt_get_int(context, "dst_v_chr_pos", 0, &dst_v_chr_pos);
2428 sws_freeContext(context);
2433 if (!(context = sws_alloc_context()))
2435 context->srcW = srcW;
2436 context->srcH = srcH;
2437 context->srcFormat = srcFormat;
2438 context->dstW = dstW;
2439 context->dstH = dstH;
2440 context->dstFormat = dstFormat;
2441 context->flags = flags;
2442 context->param[0] = param[0];
2443 context->param[1] = param[1];
2445 av_opt_set_int(context, "src_h_chr_pos", src_h_chr_pos, 0);
2446 av_opt_set_int(context, "src_v_chr_pos", src_v_chr_pos, 0);
2447 av_opt_set_int(context, "dst_h_chr_pos", dst_h_chr_pos, 0);
2448 av_opt_set_int(context, "dst_v_chr_pos", dst_v_chr_pos, 0);
2450 if (sws_init_context(context, srcFilter, dstFilter) < 0) {
2451 sws_freeContext(context);