repositories / deb_ffmpeg.git / blame
? search:
summary | shortlog | log | commit | commitdiff | tree
blob | blame (incremental) | history | HEAD
Imported Debian version 2.5.0~trusty1.1
[deb_ffmpeg.git] / ffmpeg / libswscale / utils.c
CommitLineData
2ba45a60
DM		 1/*
2 * Copyright (C) 2001-2003 Michael Niedermayer <michaelni@gmx.at>
3 *
4 * This file is part of FFmpeg.
5 *
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.
10 *
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.
15 *
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
19 */
20
21#include "config.h"
22
23#define _SVID_SOURCE // needed for MAP_ANONYMOUS
24#define _DARWIN_C_SOURCE // needed for MAP_ANON
25#include <inttypes.h>
26#include <math.h>
27#include <stdio.h>
28#include <string.h>
29#if HAVE_SYS_MMAN_H
30#include <sys/mman.h>
31#if defined(MAP_ANON) && !defined(MAP_ANONYMOUS)
32#define MAP_ANONYMOUS MAP_ANON
33#endif
34#endif
35#if HAVE_VIRTUALALLOC
36#define WIN32_LEAN_AND_MEAN
37#include <windows.h>
38#endif
39
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"
f6fa7814 45#include "libavutil/imgutils.h"
2ba45a60
DM		 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"
53#include "rgb2rgb.h"
54#include "swscale.h"
55#include "swscale_internal.h"
56
57static void handle_formats(SwsContext *c);
58
59unsigned swscale_version(void)
60{
61 av_assert0(LIBSWSCALE_VERSION_MICRO >= 100);
62 return LIBSWSCALE_VERSION_INT;
63}
64
65const char *swscale_configuration(void)
66{
67 return FFMPEG_CONFIGURATION;
68}
69
70const char *swscale_license(void)
71{
72#define LICENSE_PREFIX "libswscale license: "
73 return LICENSE_PREFIX FFMPEG_LICENSE + sizeof(LICENSE_PREFIX) - 1;
74}
75
76typedef struct FormatEntry {
77 uint8_t is_supported_in :1;
78 uint8_t is_supported_out :1;
79 uint8_t is_supported_endianness :1;
80} FormatEntry;
81
82static const FormatEntry format_entries[AV_PIX_FMT_NB] = {
83 [AV_PIX_FMT_YUV420P] = { 1, 1 },
84 [AV_PIX_FMT_YUYV422] = { 1, 1 },
85 [AV_PIX_FMT_RGB24] = { 1, 1 },
86 [AV_PIX_FMT_BGR24] = { 1, 1 },
87 [AV_PIX_FMT_YUV422P] = { 1, 1 },
88 [AV_PIX_FMT_YUV444P] = { 1, 1 },
89 [AV_PIX_FMT_YUV410P] = { 1, 1 },
90 [AV_PIX_FMT_YUV411P] = { 1, 1 },
91 [AV_PIX_FMT_GRAY8] = { 1, 1 },
92 [AV_PIX_FMT_MONOWHITE] = { 1, 1 },
93 [AV_PIX_FMT_MONOBLACK] = { 1, 1 },
94 [AV_PIX_FMT_PAL8] = { 1, 0 },
95 [AV_PIX_FMT_YUVJ420P] = { 1, 1 },
96 [AV_PIX_FMT_YUVJ411P] = { 1, 1 },
97 [AV_PIX_FMT_YUVJ422P] = { 1, 1 },
98 [AV_PIX_FMT_YUVJ444P] = { 1, 1 },
99 [AV_PIX_FMT_YVYU422] = { 1, 1 },
100 [AV_PIX_FMT_UYVY422] = { 1, 1 },
101 [AV_PIX_FMT_UYYVYY411] = { 0, 0 },
102 [AV_PIX_FMT_BGR8] = { 1, 1 },
103 [AV_PIX_FMT_BGR4] = { 0, 1 },
104 [AV_PIX_FMT_BGR4_BYTE] = { 1, 1 },
105 [AV_PIX_FMT_RGB8] = { 1, 1 },
106 [AV_PIX_FMT_RGB4] = { 0, 1 },
107 [AV_PIX_FMT_RGB4_BYTE] = { 1, 1 },
108 [AV_PIX_FMT_NV12] = { 1, 1 },
109 [AV_PIX_FMT_NV21] = { 1, 1 },
110 [AV_PIX_FMT_ARGB] = { 1, 1 },
111 [AV_PIX_FMT_RGBA] = { 1, 1 },
112 [AV_PIX_FMT_ABGR] = { 1, 1 },
113 [AV_PIX_FMT_BGRA] = { 1, 1 },
114 [AV_PIX_FMT_0RGB] = { 1, 1 },
115 [AV_PIX_FMT_RGB0] = { 1, 1 },
116 [AV_PIX_FMT_0BGR] = { 1, 1 },
117 [AV_PIX_FMT_BGR0] = { 1, 1 },
118 [AV_PIX_FMT_GRAY16BE] = { 1, 1 },
119 [AV_PIX_FMT_GRAY16LE] = { 1, 1 },
120 [AV_PIX_FMT_YUV440P] = { 1, 1 },
121 [AV_PIX_FMT_YUVJ440P] = { 1, 1 },
122 [AV_PIX_FMT_YUVA420P] = { 1, 1 },
123 [AV_PIX_FMT_YUVA422P] = { 1, 1 },
124 [AV_PIX_FMT_YUVA444P] = { 1, 1 },
125 [AV_PIX_FMT_YUVA420P9BE] = { 1, 1 },
126 [AV_PIX_FMT_YUVA420P9LE] = { 1, 1 },
127 [AV_PIX_FMT_YUVA422P9BE] = { 1, 1 },
128 [AV_PIX_FMT_YUVA422P9LE] = { 1, 1 },
129 [AV_PIX_FMT_YUVA444P9BE] = { 1, 1 },
130 [AV_PIX_FMT_YUVA444P9LE] = { 1, 1 },
131 [AV_PIX_FMT_YUVA420P10BE]= { 1, 1 },
132 [AV_PIX_FMT_YUVA420P10LE]= { 1, 1 },
133 [AV_PIX_FMT_YUVA422P10BE]= { 1, 1 },
134 [AV_PIX_FMT_YUVA422P10LE]= { 1, 1 },
135 [AV_PIX_FMT_YUVA444P10BE]= { 1, 1 },
136 [AV_PIX_FMT_YUVA444P10LE]= { 1, 1 },
137 [AV_PIX_FMT_YUVA420P16BE]= { 1, 1 },
138 [AV_PIX_FMT_YUVA420P16LE]= { 1, 1 },
139 [AV_PIX_FMT_YUVA422P16BE]= { 1, 1 },
140 [AV_PIX_FMT_YUVA422P16LE]= { 1, 1 },
141 [AV_PIX_FMT_YUVA444P16BE]= { 1, 1 },
142 [AV_PIX_FMT_YUVA444P16LE]= { 1, 1 },
143 [AV_PIX_FMT_RGB48BE] = { 1, 1 },
144 [AV_PIX_FMT_RGB48LE] = { 1, 1 },
145 [AV_PIX_FMT_RGBA64BE] = { 1, 1, 1 },
146 [AV_PIX_FMT_RGBA64LE] = { 1, 1, 1 },
147 [AV_PIX_FMT_RGB565BE] = { 1, 1 },
148 [AV_PIX_FMT_RGB565LE] = { 1, 1 },
149 [AV_PIX_FMT_RGB555BE] = { 1, 1 },
150 [AV_PIX_FMT_RGB555LE] = { 1, 1 },
151 [AV_PIX_FMT_BGR565BE] = { 1, 1 },
152 [AV_PIX_FMT_BGR565LE] = { 1, 1 },
153 [AV_PIX_FMT_BGR555BE] = { 1, 1 },
154 [AV_PIX_FMT_BGR555LE] = { 1, 1 },
155 [AV_PIX_FMT_YUV420P16LE] = { 1, 1 },
156 [AV_PIX_FMT_YUV420P16BE] = { 1, 1 },
157 [AV_PIX_FMT_YUV422P16LE] = { 1, 1 },
158 [AV_PIX_FMT_YUV422P16BE] = { 1, 1 },
159 [AV_PIX_FMT_YUV444P16LE] = { 1, 1 },
160 [AV_PIX_FMT_YUV444P16BE] = { 1, 1 },
161 [AV_PIX_FMT_RGB444LE] = { 1, 1 },
162 [AV_PIX_FMT_RGB444BE] = { 1, 1 },
163 [AV_PIX_FMT_BGR444LE] = { 1, 1 },
164 [AV_PIX_FMT_BGR444BE] = { 1, 1 },
165 [AV_PIX_FMT_YA8] = { 1, 0 },
166 [AV_PIX_FMT_YA16BE] = { 1, 0 },
167 [AV_PIX_FMT_YA16LE] = { 1, 0 },
168 [AV_PIX_FMT_BGR48BE] = { 1, 1 },
169 [AV_PIX_FMT_BGR48LE] = { 1, 1 },
170 [AV_PIX_FMT_BGRA64BE] = { 1, 1, 1 },
171 [AV_PIX_FMT_BGRA64LE] = { 1, 1, 1 },
172 [AV_PIX_FMT_YUV420P9BE] = { 1, 1 },
173 [AV_PIX_FMT_YUV420P9LE] = { 1, 1 },
174 [AV_PIX_FMT_YUV420P10BE] = { 1, 1 },
175 [AV_PIX_FMT_YUV420P10LE] = { 1, 1 },
176 [AV_PIX_FMT_YUV420P12BE] = { 1, 1 },
177 [AV_PIX_FMT_YUV420P12LE] = { 1, 1 },
178 [AV_PIX_FMT_YUV420P14BE] = { 1, 1 },
179 [AV_PIX_FMT_YUV420P14LE] = { 1, 1 },
180 [AV_PIX_FMT_YUV422P9BE] = { 1, 1 },
181 [AV_PIX_FMT_YUV422P9LE] = { 1, 1 },
182 [AV_PIX_FMT_YUV422P10BE] = { 1, 1 },
183 [AV_PIX_FMT_YUV422P10LE] = { 1, 1 },
184 [AV_PIX_FMT_YUV422P12BE] = { 1, 1 },
185 [AV_PIX_FMT_YUV422P12LE] = { 1, 1 },
186 [AV_PIX_FMT_YUV422P14BE] = { 1, 1 },
187 [AV_PIX_FMT_YUV422P14LE] = { 1, 1 },
188 [AV_PIX_FMT_YUV444P9BE] = { 1, 1 },
189 [AV_PIX_FMT_YUV444P9LE] = { 1, 1 },
190 [AV_PIX_FMT_YUV444P10BE] = { 1, 1 },
191 [AV_PIX_FMT_YUV444P10LE] = { 1, 1 },
192 [AV_PIX_FMT_YUV444P12BE] = { 1, 1 },
193 [AV_PIX_FMT_YUV444P12LE] = { 1, 1 },
194 [AV_PIX_FMT_YUV444P14BE] = { 1, 1 },
195 [AV_PIX_FMT_YUV444P14LE] = { 1, 1 },
196 [AV_PIX_FMT_GBRP] = { 1, 1 },
197 [AV_PIX_FMT_GBRP9LE] = { 1, 1 },
198 [AV_PIX_FMT_GBRP9BE] = { 1, 1 },
199 [AV_PIX_FMT_GBRP10LE] = { 1, 1 },
200 [AV_PIX_FMT_GBRP10BE] = { 1, 1 },
201 [AV_PIX_FMT_GBRP12LE] = { 1, 1 },
202 [AV_PIX_FMT_GBRP12BE] = { 1, 1 },
203 [AV_PIX_FMT_GBRP14LE] = { 1, 1 },
204 [AV_PIX_FMT_GBRP14BE] = { 1, 1 },
205 [AV_PIX_FMT_GBRP16LE] = { 1, 0 },
206 [AV_PIX_FMT_GBRP16BE] = { 1, 0 },
207 [AV_PIX_FMT_XYZ12BE] = { 1, 1, 1 },
208 [AV_PIX_FMT_XYZ12LE] = { 1, 1, 1 },
209 [AV_PIX_FMT_GBRAP] = { 1, 1 },
210 [AV_PIX_FMT_GBRAP16LE] = { 1, 0 },
211 [AV_PIX_FMT_GBRAP16BE] = { 1, 0 },
212 [AV_PIX_FMT_BAYER_BGGR8] = { 1, 0 },
213 [AV_PIX_FMT_BAYER_RGGB8] = { 1, 0 },
214 [AV_PIX_FMT_BAYER_GBRG8] = { 1, 0 },
215 [AV_PIX_FMT_BAYER_GRBG8] = { 1, 0 },
216 [AV_PIX_FMT_BAYER_BGGR16LE] = { 1, 0 },
217 [AV_PIX_FMT_BAYER_BGGR16BE] = { 1, 0 },
218 [AV_PIX_FMT_BAYER_RGGB16LE] = { 1, 0 },
219 [AV_PIX_FMT_BAYER_RGGB16BE] = { 1, 0 },
220 [AV_PIX_FMT_BAYER_GBRG16LE] = { 1, 0 },
221 [AV_PIX_FMT_BAYER_GBRG16BE] = { 1, 0 },
222 [AV_PIX_FMT_BAYER_GRBG16LE] = { 1, 0 },
223 [AV_PIX_FMT_BAYER_GRBG16BE] = { 1, 0 },
224};
225
226int sws_isSupportedInput(enum AVPixelFormat pix_fmt)
227{
228 return (unsigned)pix_fmt < AV_PIX_FMT_NB ?
229 format_entries[pix_fmt].is_supported_in : 0;
230}
231
232int sws_isSupportedOutput(enum AVPixelFormat pix_fmt)
233{
234 return (unsigned)pix_fmt < AV_PIX_FMT_NB ?
235 format_entries[pix_fmt].is_supported_out : 0;
236}
237
238int sws_isSupportedEndiannessConversion(enum AVPixelFormat pix_fmt)
239{
240 return (unsigned)pix_fmt < AV_PIX_FMT_NB ?
241 format_entries[pix_fmt].is_supported_endianness : 0;
242}
243
2ba45a60
DM		 244static double getSplineCoeff(double a, double b, double c, double d,
245 double dist)
246{
247 if (dist <= 1.0)
248 return ((d * dist + c) * dist + b) * dist + a;
249 else
250 return getSplineCoeff(0.0,
251 b + 2.0 * c + 3.0 * d,
252 c + 3.0 * d,
253 -b - 3.0 * c - 6.0 * d,
254 dist - 1.0);
255}
256
257static av_cold int get_local_pos(SwsContext *s, int chr_subsample, int pos, int dir)
258{
259 if (pos == -1 || pos <= -513) {
260 pos = (128 << chr_subsample) - 128;
261 }
262 pos += 128; // relative to ideal left edge
263 return pos >> chr_subsample;
264}
265
266typedef struct {
267 int flag; ///< flag associated to the algorithm
268 const char *description; ///< human-readable description
269 int size_factor; ///< size factor used when initing the filters
270} ScaleAlgorithm;
271
272static const ScaleAlgorithm scale_algorithms[] = {
273 { SWS_AREA, "area averaging", 1 /* downscale only, for upscale it is bilinear */ },
274 { SWS_BICUBIC, "bicubic", 4 },
275 { SWS_BICUBLIN, "luma bicubic / chroma bilinear", -1 },
276 { SWS_BILINEAR, "bilinear", 2 },
277 { SWS_FAST_BILINEAR, "fast bilinear", -1 },
278 { SWS_GAUSS, "Gaussian", 8 /* infinite ;) */ },
279 { SWS_LANCZOS, "Lanczos", -1 /* custom */ },
280 { SWS_POINT, "nearest neighbor / point", -1 },
281 { SWS_SINC, "sinc", 20 /* infinite ;) */ },
282 { SWS_SPLINE, "bicubic spline", 20 /* infinite :)*/ },
283 { SWS_X, "experimental", 8 },
284};
285
286static av_cold int initFilter(int16_t **outFilter, int32_t **filterPos,
287 int *outFilterSize, int xInc, int srcW,
288 int dstW, int filterAlign, int one,
289 int flags, int cpu_flags,
290 SwsVector *srcFilter, SwsVector *dstFilter,
291 double param[2], int srcPos, int dstPos)
292{
293 int i;
294 int filterSize;
295 int filter2Size;
296 int minFilterSize;
297 int64_t *filter = NULL;
298 int64_t *filter2 = NULL;
299 const int64_t fone = 1LL << (54 - FFMIN(av_log2(srcW/dstW), 8));
300 int ret = -1;
301
302 emms_c(); // FIXME should not be required but IS (even for non-MMX versions)
303
304 // NOTE: the +3 is for the MMX(+1) / SSE(+3) scaler which reads over the end
305 FF_ALLOC_ARRAY_OR_GOTO(NULL, *filterPos, (dstW + 3), sizeof(**filterPos), fail);
306
307 if (FFABS(xInc - 0x10000) < 10 && srcPos == dstPos) { // unscaled
308 int i;
309 filterSize = 1;
310 FF_ALLOCZ_ARRAY_OR_GOTO(NULL, filter,
311 dstW, sizeof(*filter) * filterSize, fail);
312
313 for (i = 0; i < dstW; i++) {
314 filter[i * filterSize] = fone;
315 (*filterPos)[i] = i;
316 }
317 } else if (flags & SWS_POINT) { // lame looking point sampling mode
318 int i;
319 int64_t xDstInSrc;
320 filterSize = 1;
321 FF_ALLOC_ARRAY_OR_GOTO(NULL, filter,
322 dstW, sizeof(*filter) * filterSize, fail);
323
324 xDstInSrc = ((dstPos*(int64_t)xInc)>>8) - ((srcPos*0x8000LL)>>7);
325 for (i = 0; i < dstW; i++) {
326 int xx = (xDstInSrc - ((filterSize - 1) << 15) + (1 << 15)) >> 16;
327
328 (*filterPos)[i] = xx;
329 filter[i] = fone;
330 xDstInSrc += xInc;
331 }
332 } else if ((xInc <= (1 << 16) && (flags & SWS_AREA)) ||
333 (flags & SWS_FAST_BILINEAR)) { // bilinear upscale
334 int i;
335 int64_t xDstInSrc;
336 filterSize = 2;
337 FF_ALLOC_ARRAY_OR_GOTO(NULL, filter,
338 dstW, sizeof(*filter) * filterSize, fail);
339
340 xDstInSrc = ((dstPos*(int64_t)xInc)>>8) - ((srcPos*0x8000LL)>>7);
341 for (i = 0; i < dstW; i++) {
342 int xx = (xDstInSrc - ((filterSize - 1) << 15) + (1 << 15)) >> 16;
343 int j;
344
345 (*filterPos)[i] = xx;
346 // bilinear upscale / linear interpolate / area averaging
347 for (j = 0; j < filterSize; j++) {
348 int64_t coeff= fone - FFABS(((int64_t)xx<<16) - xDstInSrc)*(fone>>16);
349 if (coeff < 0)
350 coeff = 0;
351 filter[i * filterSize + j] = coeff;
352 xx++;
353 }
354 xDstInSrc += xInc;
355 }
356 } else {
357 int64_t xDstInSrc;
358 int sizeFactor = -1;
359
360 for (i = 0; i < FF_ARRAY_ELEMS(scale_algorithms); i++) {
361 if (flags & scale_algorithms[i].flag && scale_algorithms[i].size_factor > 0) {
362 sizeFactor = scale_algorithms[i].size_factor;
363 break;
364 }
365 }
366 if (flags & SWS_LANCZOS)
367 sizeFactor = param[0] != SWS_PARAM_DEFAULT ? ceil(2 * param[0]) : 6;
368 av_assert0(sizeFactor > 0);
369
370 if (xInc <= 1 << 16)
371 filterSize = 1 + sizeFactor; // upscale
372 else
373 filterSize = 1 + (sizeFactor * srcW + dstW - 1) / dstW;
374
375 filterSize = FFMIN(filterSize, srcW - 2);
376 filterSize = FFMAX(filterSize, 1);
377
378 FF_ALLOC_ARRAY_OR_GOTO(NULL, filter,
379 dstW, sizeof(*filter) * filterSize, fail);
380
381 xDstInSrc = ((dstPos*(int64_t)xInc)>>7) - ((srcPos*0x10000LL)>>7);
382 for (i = 0; i < dstW; i++) {
383 int xx = (xDstInSrc - ((int64_t)(filterSize - 2) << 16)) / (1 << 17);
384 int j;
385 (*filterPos)[i] = xx;
386 for (j = 0; j < filterSize; j++) {
387 int64_t d = (FFABS(((int64_t)xx << 17) - xDstInSrc)) << 13;
388 double floatd;
389 int64_t coeff;
390
391 if (xInc > 1 << 16)
392 d = d * dstW / srcW;
393 floatd = d * (1.0 / (1 << 30));
394
395 if (flags & SWS_BICUBIC) {
396 int64_t B = (param[0] != SWS_PARAM_DEFAULT ? param[0] : 0) * (1 << 24);
397 int64_t C = (param[1] != SWS_PARAM_DEFAULT ? param[1] : 0.6) * (1 << 24);
398
399 if (d >= 1LL << 31) {
400 coeff = 0.0;
401 } else {
402 int64_t dd = (d * d) >> 30;
403 int64_t ddd = (dd * d) >> 30;
404
405 if (d < 1LL << 30)
406 coeff = (12 * (1 << 24) - 9 * B - 6 * C) * ddd +
407 (-18 * (1 << 24) + 12 * B + 6 * C) * dd +
408 (6 * (1 << 24) - 2 * B) * (1 << 30);
409 else
410 coeff = (-B - 6 * C) * ddd +
411 (6 * B + 30 * C) * dd +
412 (-12 * B - 48 * C) * d +
413 (8 * B + 24 * C) * (1 << 30);
414 }
415 coeff /= (1LL<<54)/fone;
416 }
417#if 0
418 else if (flags & SWS_X) {
419 double p = param ? param * 0.01 : 0.3;
420 coeff = d ? sin(d * M_PI) / (d * M_PI) : 1.0;
421 coeff *= pow(2.0, -p * d * d);
422 }
423#endif
424 else if (flags & SWS_X) {
425 double A = param[0] != SWS_PARAM_DEFAULT ? param[0] : 1.0;
426 double c;
427
428 if (floatd < 1.0)
429 c = cos(floatd * M_PI);
430 else
431 c = -1.0;
432 if (c < 0.0)
433 c = -pow(-c, A);
434 else
435 c = pow(c, A);
436 coeff = (c * 0.5 + 0.5) * fone;
437 } else if (flags & SWS_AREA) {
438 int64_t d2 = d - (1 << 29);
439 if (d2 * xInc < -(1LL << (29 + 16)))
440 coeff = 1.0 * (1LL << (30 + 16));
441 else if (d2 * xInc < (1LL << (29 + 16)))
442 coeff = -d2 * xInc + (1LL << (29 + 16));
443 else
444 coeff = 0.0;
445 coeff *= fone >> (30 + 16);
446 } else if (flags & SWS_GAUSS) {
447 double p = param[0] != SWS_PARAM_DEFAULT ? param[0] : 3.0;
448 coeff = (pow(2.0, -p * floatd * floatd)) * fone;
449 } else if (flags & SWS_SINC) {
450 coeff = (d ? sin(floatd * M_PI) / (floatd * M_PI) : 1.0) * fone;
451 } else if (flags & SWS_LANCZOS) {
452 double p = param[0] != SWS_PARAM_DEFAULT ? param[0] : 3.0;
453 coeff = (d ? sin(floatd * M_PI) * sin(floatd * M_PI / p) /
454 (floatd * floatd * M_PI * M_PI / p) : 1.0) * fone;
455 if (floatd > p)
456 coeff = 0;
457 } else if (flags & SWS_BILINEAR) {
458 coeff = (1 << 30) - d;
459 if (coeff < 0)
460 coeff = 0;
461 coeff *= fone >> 30;
462 } else if (flags & SWS_SPLINE) {
463 double p = -2.196152422706632;
464 coeff = getSplineCoeff(1.0, 0.0, p, -p - 1.0, floatd) * fone;
465 } else {
466 av_assert0(0);
467 }
468
469 filter[i * filterSize + j] = coeff;
470 xx++;
471 }
472 xDstInSrc += 2 * xInc;
473 }
474 }
475
476 /* apply src & dst Filter to filter -> filter2
477 * av_free(filter);
478 */
479 av_assert0(filterSize > 0);
480 filter2Size = filterSize;
481 if (srcFilter)
482 filter2Size += srcFilter->length - 1;
483 if (dstFilter)
484 filter2Size += dstFilter->length - 1;
485 av_assert0(filter2Size > 0);
486 FF_ALLOCZ_ARRAY_OR_GOTO(NULL, filter2, dstW, filter2Size * sizeof(*filter2), fail);
487
488 for (i = 0; i < dstW; i++) {
489 int j, k;
490
491 if (srcFilter) {
492 for (k = 0; k < srcFilter->length; k++) {
493 for (j = 0; j < filterSize; j++)
494 filter2[i * filter2Size + k + j] +=
495 srcFilter->coeff[k] * filter[i * filterSize + j];
496 }
497 } else {
498 for (j = 0; j < filterSize; j++)
499 filter2[i * filter2Size + j] = filter[i * filterSize + j];
500 }
501 // FIXME dstFilter
502
503 (*filterPos)[i] += (filterSize - 1) / 2 - (filter2Size - 1) / 2;
504 }
505 av_freep(&filter);
506
507 /* try to reduce the filter-size (step1 find size and shift left) */
508 // Assume it is near normalized (*0.5 or *2.0 is OK but * 0.001 is not).
509 minFilterSize = 0;
510 for (i = dstW - 1; i >= 0; i--) {
511 int min = filter2Size;
512 int j;
513 int64_t cutOff = 0.0;
514
515 /* get rid of near zero elements on the left by shifting left */
516 for (j = 0; j < filter2Size; j++) {
517 int k;
518 cutOff += FFABS(filter2[i * filter2Size]);
519
520 if (cutOff > SWS_MAX_REDUCE_CUTOFF * fone)
521 break;
522
523 /* preserve monotonicity because the core can't handle the
524 * filter otherwise */
525 if (i < dstW - 1 && (*filterPos)[i] >= (*filterPos)[i + 1])
526 break;
527
528 // move filter coefficients left
529 for (k = 1; k < filter2Size; k++)
530 filter2[i * filter2Size + k - 1] = filter2[i * filter2Size + k];
531 filter2[i * filter2Size + k - 1] = 0;
532 (*filterPos)[i]++;
533 }
534
535 cutOff = 0;
536 /* count near zeros on the right */
537 for (j = filter2Size - 1; j > 0; j--) {
538 cutOff += FFABS(filter2[i * filter2Size + j]);
539
540 if (cutOff > SWS_MAX_REDUCE_CUTOFF * fone)
541 break;
542 min--;
543 }
544
545 if (min > minFilterSize)
546 minFilterSize = min;
547 }
548
549 if (PPC_ALTIVEC(cpu_flags)) {
550 // we can handle the special case 4, so we don't want to go the full 8
551 if (minFilterSize < 5)
552 filterAlign = 4;
553
554 /* We really don't want to waste our time doing useless computation, so
555 * fall back on the scalar C code for very small filters.
556 * Vectorizing is worth it only if you have a decent-sized vector. */
557 if (minFilterSize < 3)
558 filterAlign = 1;
559 }
560
561 if (HAVE_MMX && cpu_flags & AV_CPU_FLAG_MMX) {
562 // special case for unscaled vertical filtering
563 if (minFilterSize == 1 && filterAlign == 2)
564 filterAlign = 1;
565 }
566
567 av_assert0(minFilterSize > 0);
568 filterSize = (minFilterSize + (filterAlign - 1)) & (~(filterAlign - 1));
569 av_assert0(filterSize > 0);
570 filter = av_malloc_array(dstW, filterSize * sizeof(*filter));
571 if (!filter)
572 goto fail;
573 if (filterSize >= MAX_FILTER_SIZE * 16 /
574 ((flags & SWS_ACCURATE_RND) ? APCK_SIZE : 16)) {
f6fa7814 575 ret = RETCODE_USE_CASCADE;
2ba45a60
DM		 576 goto fail;
577 }
578 *outFilterSize = filterSize;
579
580 if (flags & SWS_PRINT_INFO)
581 av_log(NULL, AV_LOG_VERBOSE,
582 "SwScaler: reducing / aligning filtersize %d -> %d\n",
583 filter2Size, filterSize);
584 /* try to reduce the filter-size (step2 reduce it) */
585 for (i = 0; i < dstW; i++) {
586 int j;
587
588 for (j = 0; j < filterSize; j++) {
589 if (j >= filter2Size)
590 filter[i * filterSize + j] = 0;
591 else
592 filter[i * filterSize + j] = filter2[i * filter2Size + j];
593 if ((flags & SWS_BITEXACT) && j >= minFilterSize)
594 filter[i * filterSize + j] = 0;
595 }
596 }
597
598 // FIXME try to align filterPos if possible
599
600 // fix borders
601 for (i = 0; i < dstW; i++) {
602 int j;
603 if ((*filterPos)[i] < 0) {
604 // move filter coefficients left to compensate for filterPos
605 for (j = 1; j < filterSize; j++) {
606 int left = FFMAX(j + (*filterPos)[i], 0);
607 filter[i * filterSize + left] += filter[i * filterSize + j];
608 filter[i * filterSize + j] = 0;
609 }
610 (*filterPos)[i]= 0;
611 }
612
613 if ((*filterPos)[i] + filterSize > srcW) {
614 int shift = (*filterPos)[i] + filterSize - srcW;
615 // move filter coefficients right to compensate for filterPos
616 for (j = filterSize - 2; j >= 0; j--) {
617 int right = FFMIN(j + shift, filterSize - 1);
618 filter[i * filterSize + right] += filter[i * filterSize + j];
619 filter[i * filterSize + j] = 0;
620 }
621 (*filterPos)[i]= srcW - filterSize;
622 }
623 }
624
625 // Note the +1 is for the MMX scaler which reads over the end
626 /* align at 16 for AltiVec (needed by hScale_altivec_real) */
627 FF_ALLOCZ_ARRAY_OR_GOTO(NULL, *outFilter,
628 (dstW + 3), *outFilterSize * sizeof(int16_t), fail);
629
630 /* normalize & store in outFilter */
631 for (i = 0; i < dstW; i++) {
632 int j;
633 int64_t error = 0;
634 int64_t sum = 0;
635
636 for (j = 0; j < filterSize; j++) {
637 sum += filter[i * filterSize + j];
638 }
639 sum = (sum + one / 2) / one;
640 if (!sum) {
641 av_log(NULL, AV_LOG_WARNING, "SwScaler: zero vector in scaling\n");
642 sum = 1;
643 }
644 for (j = 0; j < *outFilterSize; j++) {
645 int64_t v = filter[i * filterSize + j] + error;
646 int intV = ROUNDED_DIV(v, sum);
647 (*outFilter)[i * (*outFilterSize) + j] = intV;
648 error = v - intV * sum;
649 }
650 }
651
652 (*filterPos)[dstW + 0] =
653 (*filterPos)[dstW + 1] =
654 (*filterPos)[dstW + 2] = (*filterPos)[dstW - 1]; /* the MMX/SSE scaler will
655 * read over the end */
656 for (i = 0; i < *outFilterSize; i++) {
657 int k = (dstW - 1) * (*outFilterSize) + i;
658 (*outFilter)[k + 1 * (*outFilterSize)] =
659 (*outFilter)[k + 2 * (*outFilterSize)] =
660 (*outFilter)[k + 3 * (*outFilterSize)] = (*outFilter)[k];
661 }
662
663 ret = 0;
664
665fail:
666 if(ret < 0)
f6fa7814 667 av_log(NULL, ret == RETCODE_USE_CASCADE ? AV_LOG_DEBUG : AV_LOG_ERROR, "sws: initFilter failed\n");
2ba45a60
DM		 668 av_free(filter);
669 av_free(filter2);
670 return ret;
671}
672
673static void fill_rgb2yuv_table(SwsContext *c, const int table[4], int dstRange)
674{
675 int64_t W, V, Z, Cy, Cu, Cv;
676 int64_t vr = table[0];
677 int64_t ub = table[1];
678 int64_t ug = -table[2];
679 int64_t vg = -table[3];
680 int64_t ONE = 65536;
681 int64_t cy = ONE;
682 uint8_t *p = (uint8_t*)c->input_rgb2yuv_table;
683 int i;
684 static const int8_t map[] = {
685 BY_IDX, GY_IDX, -1 , BY_IDX, BY_IDX, GY_IDX, -1 , BY_IDX,
686 RY_IDX, -1 , GY_IDX, RY_IDX, RY_IDX, -1 , GY_IDX, RY_IDX,
687 RY_IDX, GY_IDX, -1 , RY_IDX, RY_IDX, GY_IDX, -1 , RY_IDX,
688 BY_IDX, -1 , GY_IDX, BY_IDX, BY_IDX, -1 , GY_IDX, BY_IDX,
689 BU_IDX, GU_IDX, -1 , BU_IDX, BU_IDX, GU_IDX, -1 , BU_IDX,
690 RU_IDX, -1 , GU_IDX, RU_IDX, RU_IDX, -1 , GU_IDX, RU_IDX,
691 RU_IDX, GU_IDX, -1 , RU_IDX, RU_IDX, GU_IDX, -1 , RU_IDX,
692 BU_IDX, -1 , GU_IDX, BU_IDX, BU_IDX, -1 , GU_IDX, BU_IDX,
693 BV_IDX, GV_IDX, -1 , BV_IDX, BV_IDX, GV_IDX, -1 , BV_IDX,
694 RV_IDX, -1 , GV_IDX, RV_IDX, RV_IDX, -1 , GV_IDX, RV_IDX,
695 RV_IDX, GV_IDX, -1 , RV_IDX, RV_IDX, GV_IDX, -1 , RV_IDX,
696 BV_IDX, -1 , GV_IDX, BV_IDX, BV_IDX, -1 , GV_IDX, BV_IDX,
697 RY_IDX, BY_IDX, RY_IDX, BY_IDX, RY_IDX, BY_IDX, RY_IDX, BY_IDX,
698 BY_IDX, RY_IDX, BY_IDX, RY_IDX, BY_IDX, RY_IDX, BY_IDX, RY_IDX,
699 GY_IDX, -1 , GY_IDX, -1 , GY_IDX, -1 , GY_IDX, -1 ,
700 -1 , GY_IDX, -1 , GY_IDX, -1 , GY_IDX, -1 , GY_IDX,
701 RU_IDX, BU_IDX, RU_IDX, BU_IDX, RU_IDX, BU_IDX, RU_IDX, BU_IDX,
702 BU_IDX, RU_IDX, BU_IDX, RU_IDX, BU_IDX, RU_IDX, BU_IDX, RU_IDX,
703 GU_IDX, -1 , GU_IDX, -1 , GU_IDX, -1 , GU_IDX, -1 ,
704 -1 , GU_IDX, -1 , GU_IDX, -1 , GU_IDX, -1 , GU_IDX,
705 RV_IDX, BV_IDX, RV_IDX, BV_IDX, RV_IDX, BV_IDX, RV_IDX, BV_IDX,
706 BV_IDX, RV_IDX, BV_IDX, RV_IDX, BV_IDX, RV_IDX, BV_IDX, RV_IDX,
707 GV_IDX, -1 , GV_IDX, -1 , GV_IDX, -1 , GV_IDX, -1 ,
708 -1 , GV_IDX, -1 , GV_IDX, -1 , GV_IDX, -1 , GV_IDX, //23
709 -1 , -1 , -1 , -1 , -1 , -1 , -1 , -1 , //24
710 -1 , -1 , -1 , -1 , -1 , -1 , -1 , -1 , //25
711 -1 , -1 , -1 , -1 , -1 , -1 , -1 , -1 , //26
712 -1 , -1 , -1 , -1 , -1 , -1 , -1 , -1 , //27
713 -1 , -1 , -1 , -1 , -1 , -1 , -1 , -1 , //28
714 -1 , -1 , -1 , -1 , -1 , -1 , -1 , -1 , //29
715 -1 , -1 , -1 , -1 , -1 , -1 , -1 , -1 , //30
716 -1 , -1 , -1 , -1 , -1 , -1 , -1 , -1 , //31
717 BY_IDX, GY_IDX, RY_IDX, -1 , -1 , -1 , -1 , -1 , //32
718 BU_IDX, GU_IDX, RU_IDX, -1 , -1 , -1 , -1 , -1 , //33
719 BV_IDX, GV_IDX, RV_IDX, -1 , -1 , -1 , -1 , -1 , //34
720 };
721
722 dstRange = 0; //FIXME range = 1 is handled elsewhere
723
724 if (!dstRange) {
725 cy = cy * 255 / 219;
726 } else {
727 vr = vr * 224 / 255;
728 ub = ub * 224 / 255;
729 ug = ug * 224 / 255;
730 vg = vg * 224 / 255;
731 }
732 W = ROUNDED_DIV(ONE*ONE*ug, ub);
733 V = ROUNDED_DIV(ONE*ONE*vg, vr);
734 Z = ONE*ONE-W-V;
735
736 Cy = ROUNDED_DIV(cy*Z, ONE);
737 Cu = ROUNDED_DIV(ub*Z, ONE);
738 Cv = ROUNDED_DIV(vr*Z, ONE);
739
740 c->input_rgb2yuv_table[RY_IDX] = -ROUNDED_DIV((1 << RGB2YUV_SHIFT)*V , Cy);
741 c->input_rgb2yuv_table[GY_IDX] = ROUNDED_DIV((1 << RGB2YUV_SHIFT)*ONE*ONE , Cy);
742 c->input_rgb2yuv_table[BY_IDX] = -ROUNDED_DIV((1 << RGB2YUV_SHIFT)*W , Cy);
743
744 c->input_rgb2yuv_table[RU_IDX] = ROUNDED_DIV((1 << RGB2YUV_SHIFT)*V , Cu);
745 c->input_rgb2yuv_table[GU_IDX] = -ROUNDED_DIV((1 << RGB2YUV_SHIFT)*ONE*ONE , Cu);
746 c->input_rgb2yuv_table[BU_IDX] = ROUNDED_DIV((1 << RGB2YUV_SHIFT)*(Z+W) , Cu);
747
748 c->input_rgb2yuv_table[RV_IDX] = ROUNDED_DIV((1 << RGB2YUV_SHIFT)*(V+Z) , Cv);
749 c->input_rgb2yuv_table[GV_IDX] = -ROUNDED_DIV((1 << RGB2YUV_SHIFT)*ONE*ONE , Cv);
750 c->input_rgb2yuv_table[BV_IDX] = ROUNDED_DIV((1 << RGB2YUV_SHIFT)*W , Cv);
751
752 if(/*!dstRange && */!memcmp(table, ff_yuv2rgb_coeffs[SWS_CS_DEFAULT], sizeof(ff_yuv2rgb_coeffs[SWS_CS_DEFAULT]))) {
753 c->input_rgb2yuv_table[BY_IDX] = ((int)(0.114 * 219 / 255 * (1 << RGB2YUV_SHIFT) + 0.5));
754 c->input_rgb2yuv_table[BV_IDX] = (-(int)(0.081 * 224 / 255 * (1 << RGB2YUV_SHIFT) + 0.5));
755 c->input_rgb2yuv_table[BU_IDX] = ((int)(0.500 * 224 / 255 * (1 << RGB2YUV_SHIFT) + 0.5));
756 c->input_rgb2yuv_table[GY_IDX] = ((int)(0.587 * 219 / 255 * (1 << RGB2YUV_SHIFT) + 0.5));
757 c->input_rgb2yuv_table[GV_IDX] = (-(int)(0.419 * 224 / 255 * (1 << RGB2YUV_SHIFT) + 0.5));
758 c->input_rgb2yuv_table[GU_IDX] = (-(int)(0.331 * 224 / 255 * (1 << RGB2YUV_SHIFT) + 0.5));
759 c->input_rgb2yuv_table[RY_IDX] = ((int)(0.299 * 219 / 255 * (1 << RGB2YUV_SHIFT) + 0.5));
760 c->input_rgb2yuv_table[RV_IDX] = ((int)(0.500 * 224 / 255 * (1 << RGB2YUV_SHIFT) + 0.5));
761 c->input_rgb2yuv_table[RU_IDX] = (-(int)(0.169 * 224 / 255 * (1 << RGB2YUV_SHIFT) + 0.5));
762 }
763 for(i=0; i<FF_ARRAY_ELEMS(map); i++)
764 AV_WL16(p + 16*4 + 2*i, map[i] >= 0 ? c->input_rgb2yuv_table[map[i]] : 0);
765}
766
767static void fill_xyztables(struct SwsContext *c)
768{
769 int i;
770 double xyzgamma = XYZ_GAMMA;
771 double rgbgamma = 1.0 / RGB_GAMMA;
772 double xyzgammainv = 1.0 / XYZ_GAMMA;
773 double rgbgammainv = RGB_GAMMA;
774 static const int16_t xyz2rgb_matrix[3][4] = {
775 {13270, -6295, -2041},
776 {-3969, 7682, 170},
777 { 228, -835, 4329} };
778 static const int16_t rgb2xyz_matrix[3][4] = {
779 {1689, 1464, 739},
780 { 871, 2929, 296},
781 { 79, 488, 3891} };
782 static int16_t xyzgamma_tab[4096], rgbgamma_tab[4096], xyzgammainv_tab[4096], rgbgammainv_tab[4096];
783
784 memcpy(c->xyz2rgb_matrix, xyz2rgb_matrix, sizeof(c->xyz2rgb_matrix));
785 memcpy(c->rgb2xyz_matrix, rgb2xyz_matrix, sizeof(c->rgb2xyz_matrix));
786 c->xyzgamma = xyzgamma_tab;
787 c->rgbgamma = rgbgamma_tab;
788 c->xyzgammainv = xyzgammainv_tab;
789 c->rgbgammainv = rgbgammainv_tab;
790
791 if (rgbgamma_tab[4095])
792 return;
793
794 /* set gamma vectors */
795 for (i = 0; i < 4096; i++) {
796 xyzgamma_tab[i] = lrint(pow(i / 4095.0, xyzgamma) * 4095.0);
797 rgbgamma_tab[i] = lrint(pow(i / 4095.0, rgbgamma) * 4095.0);
798 xyzgammainv_tab[i] = lrint(pow(i / 4095.0, xyzgammainv) * 4095.0);
799 rgbgammainv_tab[i] = lrint(pow(i / 4095.0, rgbgammainv) * 4095.0);
800 }
801}
802
803int sws_setColorspaceDetails(struct SwsContext *c, const int inv_table[4],
804 int srcRange, const int table[4], int dstRange,
805 int brightness, int contrast, int saturation)
806{
807 const AVPixFmtDescriptor *desc_dst;
808 const AVPixFmtDescriptor *desc_src;
809 int need_reinit = 0;
810 memmove(c->srcColorspaceTable, inv_table, sizeof(int) * 4);
811 memmove(c->dstColorspaceTable, table, sizeof(int) * 4);
812
813 handle_formats(c);
814 desc_dst = av_pix_fmt_desc_get(c->dstFormat);
815 desc_src = av_pix_fmt_desc_get(c->srcFormat);
816
817 if(!isYUV(c->dstFormat) && !isGray(c->dstFormat))
818 dstRange = 0;
819 if(!isYUV(c->srcFormat) && !isGray(c->srcFormat))
820 srcRange = 0;
821
822 c->brightness = brightness;
823 c->contrast = contrast;
824 c->saturation = saturation;
825 if (c->srcRange != srcRange || c->dstRange != dstRange)
826 need_reinit = 1;
827 c->srcRange = srcRange;
828 c->dstRange = dstRange;
829
830 //The srcBpc check is possibly wrong but we seem to lack a definitive reference to test this
831 //and what we have in ticket 2939 looks better with this check
832 if (need_reinit && (c->srcBpc == 8 || !isYUV(c->srcFormat)))
833 ff_sws_init_range_convert(c);
834
835 if ((isYUV(c->dstFormat) || isGray(c->dstFormat)) && (isYUV(c->srcFormat) || isGray(c->srcFormat)))
836 return -1;
837
838 c->dstFormatBpp = av_get_bits_per_pixel(desc_dst);
839 c->srcFormatBpp = av_get_bits_per_pixel(desc_src);
840
841 if (!isYUV(c->dstFormat) && !isGray(c->dstFormat)) {
842 ff_yuv2rgb_c_init_tables(c, inv_table, srcRange, brightness,
843 contrast, saturation);
844 // FIXME factorize
845
846 if (ARCH_PPC)
847 ff_yuv2rgb_init_tables_ppc(c, inv_table, brightness,
848 contrast, saturation);
849 }
850
851 fill_rgb2yuv_table(c, table, dstRange);
852
853 return 0;
854}
855
856int sws_getColorspaceDetails(struct SwsContext *c, int **inv_table,
857 int *srcRange, int **table, int *dstRange,
858 int *brightness, int *contrast, int *saturation)
859{
860 if (!c )
861 return -1;
862
863 *inv_table = c->srcColorspaceTable;
864 *table = c->dstColorspaceTable;
865 *srcRange = c->srcRange;
866 *dstRange = c->dstRange;
867 *brightness = c->brightness;
868 *contrast = c->contrast;
869 *saturation = c->saturation;
870
871 return 0;
872}
873
874static int handle_jpeg(enum AVPixelFormat *format)
875{
876 switch (*format) {
877 case AV_PIX_FMT_YUVJ420P:
878 *format = AV_PIX_FMT_YUV420P;
879 return 1;
880 case AV_PIX_FMT_YUVJ411P:
881 *format = AV_PIX_FMT_YUV411P;
882 return 1;
883 case AV_PIX_FMT_YUVJ422P:
884 *format = AV_PIX_FMT_YUV422P;
885 return 1;
886 case AV_PIX_FMT_YUVJ444P:
887 *format = AV_PIX_FMT_YUV444P;
888 return 1;
889 case AV_PIX_FMT_YUVJ440P:
890 *format = AV_PIX_FMT_YUV440P;
891 return 1;
892 case AV_PIX_FMT_GRAY8:
893 case AV_PIX_FMT_GRAY16LE:
894 case AV_PIX_FMT_GRAY16BE:
895 return 1;
896 default:
897 return 0;
898 }
899}
900
901static int handle_0alpha(enum AVPixelFormat *format)
902{
903 switch (*format) {
904 case AV_PIX_FMT_0BGR : *format = AV_PIX_FMT_ABGR ; return 1;
905 case AV_PIX_FMT_BGR0 : *format = AV_PIX_FMT_BGRA ; return 4;
906 case AV_PIX_FMT_0RGB : *format = AV_PIX_FMT_ARGB ; return 1;
907 case AV_PIX_FMT_RGB0 : *format = AV_PIX_FMT_RGBA ; return 4;
908 default: return 0;
909 }
910}
911
912static int handle_xyz(enum AVPixelFormat *format)
913{
914 switch (*format) {
915 case AV_PIX_FMT_XYZ12BE : *format = AV_PIX_FMT_RGB48BE; return 1;
916 case AV_PIX_FMT_XYZ12LE : *format = AV_PIX_FMT_RGB48LE; return 1;
917 default: return 0;
918 }
919}
920
921static void handle_formats(SwsContext *c)
922{
923 c->src0Alpha |= handle_0alpha(&c->srcFormat);
924 c->dst0Alpha |= handle_0alpha(&c->dstFormat);
925 c->srcXYZ |= handle_xyz(&c->srcFormat);
926 c->dstXYZ |= handle_xyz(&c->dstFormat);
927 if (c->srcXYZ || c->dstXYZ)
928 fill_xyztables(c);
929}
930
931SwsContext *sws_alloc_context(void)
932{
933 SwsContext *c = av_mallocz(sizeof(SwsContext));
934
935 av_assert0(offsetof(SwsContext, redDither) + DITHER32_INT == offsetof(SwsContext, dither32));
936
937 if (c) {
938 c->av_class = &sws_context_class;
939 av_opt_set_defaults(c);
940 }
941
942 return c;
943}
944
945av_cold int sws_init_context(SwsContext *c, SwsFilter *srcFilter,
946 SwsFilter *dstFilter)
947{
948 int i, j;
949 int usesVFilter, usesHFilter;
950 int unscaled;
951 SwsFilter dummyFilter = { NULL, NULL, NULL, NULL };
952 int srcW = c->srcW;
953 int srcH = c->srcH;
954 int dstW = c->dstW;
955 int dstH = c->dstH;
956 int dst_stride = FFALIGN(dstW * sizeof(int16_t) + 66, 16);
957 int flags, cpu_flags;
958 enum AVPixelFormat srcFormat = c->srcFormat;
959 enum AVPixelFormat dstFormat = c->dstFormat;
960 const AVPixFmtDescriptor *desc_src;
961 const AVPixFmtDescriptor *desc_dst;
f6fa7814 962 int ret = 0;
2ba45a60
DM		 963
964 cpu_flags = av_get_cpu_flags();
965 flags = c->flags;
966 emms_c();
967 if (!rgb15to16)
968 sws_rgb2rgb_init();
969
970 unscaled = (srcW == dstW && srcH == dstH);
971
972 c->srcRange |= handle_jpeg(&c->srcFormat);
973 c->dstRange |= handle_jpeg(&c->dstFormat);
974
975 if(srcFormat!=c->srcFormat || dstFormat!=c->dstFormat)
976 av_log(c, AV_LOG_WARNING, "deprecated pixel format used, make sure you did set range correctly\n");
977
978 if (!c->contrast && !c->saturation && !c->dstFormatBpp)
979 sws_setColorspaceDetails(c, ff_yuv2rgb_coeffs[SWS_CS_DEFAULT], c->srcRange,
980 ff_yuv2rgb_coeffs[SWS_CS_DEFAULT],
981 c->dstRange, 0, 1 << 16, 1 << 16);
982
983 handle_formats(c);
984 srcFormat = c->srcFormat;
985 dstFormat = c->dstFormat;
986 desc_src = av_pix_fmt_desc_get(srcFormat);
987 desc_dst = av_pix_fmt_desc_get(dstFormat);
988
989 if (!(unscaled && sws_isSupportedEndiannessConversion(srcFormat) &&
990 av_pix_fmt_swap_endianness(srcFormat) == dstFormat)) {
991 if (!sws_isSupportedInput(srcFormat)) {
992 av_log(c, AV_LOG_ERROR, "%s is not supported as input pixel format\n",
993 av_get_pix_fmt_name(srcFormat));
994 return AVERROR(EINVAL);
995 }
996 if (!sws_isSupportedOutput(dstFormat)) {
997 av_log(c, AV_LOG_ERROR, "%s is not supported as output pixel format\n",
998 av_get_pix_fmt_name(dstFormat));
999 return AVERROR(EINVAL);
1000 }
1001 }
1002
1003 i = flags & (SWS_POINT |
1004 SWS_AREA |
1005 SWS_BILINEAR |
1006 SWS_FAST_BILINEAR |
1007 SWS_BICUBIC |
1008 SWS_X |
1009 SWS_GAUSS |
1010 SWS_LANCZOS |
1011 SWS_SINC |
1012 SWS_SPLINE |
1013 SWS_BICUBLIN);
1014
1015 /* provide a default scaler if not set by caller */
1016 if (!i) {
1017 if (dstW < srcW && dstH < srcH)
1018 flags |= SWS_BICUBIC;
1019 else if (dstW > srcW && dstH > srcH)
1020 flags |= SWS_BICUBIC;
1021 else
1022 flags |= SWS_BICUBIC;
1023 c->flags = flags;
1024 } else if (i & (i - 1)) {
1025 av_log(c, AV_LOG_ERROR,
1026 "Exactly one scaler algorithm must be chosen, got %X\n", i);
1027 return AVERROR(EINVAL);
1028 }
1029 /* sanity check */
1030 if (srcW < 1 || srcH < 1 || dstW < 1 || dstH < 1) {
1031 /* FIXME check if these are enough and try to lower them after
1032 * fixing the relevant parts of the code */
1033 av_log(c, AV_LOG_ERROR, "%dx%d -> %dx%d is invalid scaling dimension\n",
1034 srcW, srcH, dstW, dstH);
1035 return AVERROR(EINVAL);
1036 }
1037
1038 if (!dstFilter)
1039 dstFilter = &dummyFilter;
1040 if (!srcFilter)
1041 srcFilter = &dummyFilter;
1042
1043 c->lumXInc = (((int64_t)srcW << 16) + (dstW >> 1)) / dstW;
1044 c->lumYInc = (((int64_t)srcH << 16) + (dstH >> 1)) / dstH;
1045 c->dstFormatBpp = av_get_bits_per_pixel(desc_dst);
1046 c->srcFormatBpp = av_get_bits_per_pixel(desc_src);
1047 c->vRounder = 4 * 0x0001000100010001ULL;
1048
1049 usesVFilter = (srcFilter->lumV && srcFilter->lumV->length > 1) ||
1050 (srcFilter->chrV && srcFilter->chrV->length > 1) ||
1051 (dstFilter->lumV && dstFilter->lumV->length > 1) ||
1052 (dstFilter->chrV && dstFilter->chrV->length > 1);
1053 usesHFilter = (srcFilter->lumH && srcFilter->lumH->length > 1) ||
1054 (srcFilter->chrH && srcFilter->chrH->length > 1) ||
1055 (dstFilter->lumH && dstFilter->lumH->length > 1) ||
1056 (dstFilter->chrH && dstFilter->chrH->length > 1);
1057
1058 av_pix_fmt_get_chroma_sub_sample(srcFormat, &c->chrSrcHSubSample, &c->chrSrcVSubSample);
1059 av_pix_fmt_get_chroma_sub_sample(dstFormat, &c->chrDstHSubSample, &c->chrDstVSubSample);
1060
1061 if (isAnyRGB(dstFormat) && !(flags&SWS_FULL_CHR_H_INT)) {
1062 if (dstW&1) {
1063 av_log(c, AV_LOG_DEBUG, "Forcing full internal H chroma due to odd output size\n");
1064 flags |= SWS_FULL_CHR_H_INT;
1065 c->flags = flags;
1066 }
1067
1068 if ( c->chrSrcHSubSample == 0
1069 && c->chrSrcVSubSample == 0
1070 && c->dither != SWS_DITHER_BAYER //SWS_FULL_CHR_H_INT is currently not supported with SWS_DITHER_BAYER
1071 && !(c->flags & SWS_FAST_BILINEAR)
1072 ) {
1073 av_log(c, AV_LOG_DEBUG, "Forcing full internal H chroma due to input having non subsampled chroma\n");
1074 flags |= SWS_FULL_CHR_H_INT;
1075 c->flags = flags;
1076 }
1077 }
1078
1079 if (c->dither == SWS_DITHER_AUTO) {
1080 if (flags & SWS_ERROR_DIFFUSION)
1081 c->dither = SWS_DITHER_ED;
1082 }
1083
1084 if(dstFormat == AV_PIX_FMT_BGR4_BYTE ||
1085 dstFormat == AV_PIX_FMT_RGB4_BYTE ||
1086 dstFormat == AV_PIX_FMT_BGR8 ||
1087 dstFormat == AV_PIX_FMT_RGB8) {
1088 if (c->dither == SWS_DITHER_AUTO)
1089 c->dither = (flags & SWS_FULL_CHR_H_INT) ? SWS_DITHER_ED : SWS_DITHER_BAYER;
1090 if (!(flags & SWS_FULL_CHR_H_INT)) {
1091 if (c->dither == SWS_DITHER_ED || c->dither == SWS_DITHER_A_DITHER || c->dither == SWS_DITHER_X_DITHER) {
1092 av_log(c, AV_LOG_DEBUG,
1093 "Desired dithering only supported in full chroma interpolation for destination format '%s'\n",
1094 av_get_pix_fmt_name(dstFormat));
1095 flags |= SWS_FULL_CHR_H_INT;
1096 c->flags = flags;
1097 }
1098 }
1099 if (flags & SWS_FULL_CHR_H_INT) {
1100 if (c->dither == SWS_DITHER_BAYER) {
1101 av_log(c, AV_LOG_DEBUG,
1102 "Ordered dither is not supported in full chroma interpolation for destination format '%s'\n",
1103 av_get_pix_fmt_name(dstFormat));
1104 c->dither = SWS_DITHER_ED;
1105 }
1106 }
1107 }
1108 if (isPlanarRGB(dstFormat)) {
1109 if (!(flags & SWS_FULL_CHR_H_INT)) {
1110 av_log(c, AV_LOG_DEBUG,
1111 "%s output is not supported with half chroma resolution, switching to full\n",
1112 av_get_pix_fmt_name(dstFormat));
1113 flags |= SWS_FULL_CHR_H_INT;
1114 c->flags = flags;
1115 }
1116 }
1117
1118 /* reuse chroma for 2 pixels RGB/BGR unless user wants full
1119 * chroma interpolation */
1120 if (flags & SWS_FULL_CHR_H_INT &&
1121 isAnyRGB(dstFormat) &&
1122 !isPlanarRGB(dstFormat) &&
1123 dstFormat != AV_PIX_FMT_RGBA &&
1124 dstFormat != AV_PIX_FMT_ARGB &&
1125 dstFormat != AV_PIX_FMT_BGRA &&
1126 dstFormat != AV_PIX_FMT_ABGR &&
1127 dstFormat != AV_PIX_FMT_RGB24 &&
1128 dstFormat != AV_PIX_FMT_BGR24 &&
1129 dstFormat != AV_PIX_FMT_BGR4_BYTE &&
1130 dstFormat != AV_PIX_FMT_RGB4_BYTE &&
1131 dstFormat != AV_PIX_FMT_BGR8 &&
1132 dstFormat != AV_PIX_FMT_RGB8
1133 ) {
1134 av_log(c, AV_LOG_WARNING,
1135 "full chroma interpolation for destination format '%s' not yet implemented\n",
1136 av_get_pix_fmt_name(dstFormat));
1137 flags &= ~SWS_FULL_CHR_H_INT;
1138 c->flags = flags;
1139 }
1140 if (isAnyRGB(dstFormat) && !(flags & SWS_FULL_CHR_H_INT))
1141 c->chrDstHSubSample = 1;
1142
1143 // drop some chroma lines if the user wants it
1144 c->vChrDrop = (flags & SWS_SRC_V_CHR_DROP_MASK) >>
1145 SWS_SRC_V_CHR_DROP_SHIFT;
1146 c->chrSrcVSubSample += c->vChrDrop;
1147
1148 /* drop every other pixel for chroma calculation unless user
1149 * wants full chroma */
1150 if (isAnyRGB(srcFormat) && !(flags & SWS_FULL_CHR_H_INP) &&
1151 srcFormat != AV_PIX_FMT_RGB8 && srcFormat != AV_PIX_FMT_BGR8 &&
1152 srcFormat != AV_PIX_FMT_RGB4 && srcFormat != AV_PIX_FMT_BGR4 &&
1153 srcFormat != AV_PIX_FMT_RGB4_BYTE && srcFormat != AV_PIX_FMT_BGR4_BYTE &&
1154 srcFormat != AV_PIX_FMT_GBRP9BE && srcFormat != AV_PIX_FMT_GBRP9LE &&
1155 srcFormat != AV_PIX_FMT_GBRP10BE && srcFormat != AV_PIX_FMT_GBRP10LE &&
1156 srcFormat != AV_PIX_FMT_GBRP12BE && srcFormat != AV_PIX_FMT_GBRP12LE &&
1157 srcFormat != AV_PIX_FMT_GBRP14BE && srcFormat != AV_PIX_FMT_GBRP14LE &&
1158 srcFormat != AV_PIX_FMT_GBRP16BE && srcFormat != AV_PIX_FMT_GBRP16LE &&
1159 ((dstW >> c->chrDstHSubSample) <= (srcW >> 1) ||
1160 (flags & SWS_FAST_BILINEAR)))
1161 c->chrSrcHSubSample = 1;
1162
1163 // Note the FF_CEIL_RSHIFT is so that we always round toward +inf.
1164 c->chrSrcW = FF_CEIL_RSHIFT(srcW, c->chrSrcHSubSample);
1165 c->chrSrcH = FF_CEIL_RSHIFT(srcH, c->chrSrcVSubSample);
1166 c->chrDstW = FF_CEIL_RSHIFT(dstW, c->chrDstHSubSample);
1167 c->chrDstH = FF_CEIL_RSHIFT(dstH, c->chrDstVSubSample);
1168
1169 FF_ALLOC_OR_GOTO(c, c->formatConvBuffer, FFALIGN(srcW*2+78, 16) * 2, fail);
1170
1171 c->srcBpc = 1 + desc_src->comp[0].depth_minus1;
1172 if (c->srcBpc < 8)
1173 c->srcBpc = 8;
1174 c->dstBpc = 1 + desc_dst->comp[0].depth_minus1;
1175 if (c->dstBpc < 8)
1176 c->dstBpc = 8;
1177 if (isAnyRGB(srcFormat) || srcFormat == AV_PIX_FMT_PAL8)
1178 c->srcBpc = 16;
1179 if (c->dstBpc == 16)
1180 dst_stride <<= 1;
1181
1182 if (INLINE_MMXEXT(cpu_flags) && c->srcBpc == 8 && c->dstBpc <= 14) {
1183 c->canMMXEXTBeUsed = dstW >= srcW && (dstW & 31) == 0 &&
1184 c->chrDstW >= c->chrSrcW &&
1185 (srcW & 15) == 0;
1186 if (!c->canMMXEXTBeUsed && dstW >= srcW && c->chrDstW >= c->chrSrcW && (srcW & 15) == 0
1187
1188 && (flags & SWS_FAST_BILINEAR)) {
1189 if (flags & SWS_PRINT_INFO)
1190 av_log(c, AV_LOG_INFO,
1191 "output width is not a multiple of 32 -> no MMXEXT scaler\n");
1192 }
1193 if (usesHFilter || isNBPS(c->srcFormat) || is16BPS(c->srcFormat) || isAnyRGB(c->srcFormat))
1194 c->canMMXEXTBeUsed = 0;
1195 } else
1196 c->canMMXEXTBeUsed = 0;
1197
1198 c->chrXInc = (((int64_t)c->chrSrcW << 16) + (c->chrDstW >> 1)) / c->chrDstW;
1199 c->chrYInc = (((int64_t)c->chrSrcH << 16) + (c->chrDstH >> 1)) / c->chrDstH;
1200
1201 /* Match pixel 0 of the src to pixel 0 of dst and match pixel n-2 of src
1202 * to pixel n-2 of dst, but only for the FAST_BILINEAR mode otherwise do
1203 * correct scaling.
1204 * n-2 is the last chrominance sample available.
1205 * This is not perfect, but no one should notice the difference, the more
1206 * correct variant would be like the vertical one, but that would require
1207 * some special code for the first and last pixel */
1208 if (flags & SWS_FAST_BILINEAR) {
1209 if (c->canMMXEXTBeUsed) {
1210 c->lumXInc += 20;
1211 c->chrXInc += 20;
1212 }
1213 // we don't use the x86 asm scaler if MMX is available
1214 else if (INLINE_MMX(cpu_flags) && c->dstBpc <= 14) {
1215 c->lumXInc = ((int64_t)(srcW - 2) << 16) / (dstW - 2) - 20;
1216 c->chrXInc = ((int64_t)(c->chrSrcW - 2) << 16) / (c->chrDstW - 2) - 20;
1217 }
1218 }
1219
f6fa7814
DM		 1220 if (isBayer(srcFormat)) {
1221 if (!unscaled ||
1222 (dstFormat != AV_PIX_FMT_RGB24 && dstFormat != AV_PIX_FMT_YUV420P)) {
1223 enum AVPixelFormat tmpFormat = AV_PIX_FMT_RGB24;
1224
1225 ret = av_image_alloc(c->cascaded_tmp, c->cascaded_tmpStride,
1226 srcW, srcH, tmpFormat, 64);
1227 if (ret < 0)
1228 return ret;
1229
1230 c->cascaded_context[0] = sws_getContext(srcW, srcH, srcFormat,
1231 srcW, srcH, tmpFormat,
1232 flags, srcFilter, NULL, c->param);
1233 if (!c->cascaded_context[0])
1234 return -1;
1235
1236 c->cascaded_context[1] = sws_getContext(srcW, srcH, tmpFormat,
1237 dstW, dstH, dstFormat,
1238 flags, NULL, dstFilter, c->param);
1239 if (!c->cascaded_context[1])
1240 return -1;
1241 return 0;
1242 }
1243 }
1244
2ba45a60
DM		 1245#define USE_MMAP (HAVE_MMAP && HAVE_MPROTECT && defined MAP_ANONYMOUS)
1246
1247 /* precalculate horizontal scaler filter coefficients */
1248 {
1249#if HAVE_MMXEXT_INLINE
1250// can't downscale !!!
1251 if (c->canMMXEXTBeUsed && (flags & SWS_FAST_BILINEAR)) {
1252 c->lumMmxextFilterCodeSize = ff_init_hscaler_mmxext(dstW, c->lumXInc, NULL,
1253 NULL, NULL, 8);
1254 c->chrMmxextFilterCodeSize = ff_init_hscaler_mmxext(c->chrDstW, c->chrXInc,
1255 NULL, NULL, NULL, 4);
1256
1257#if USE_MMAP
1258 c->lumMmxextFilterCode = mmap(NULL, c->lumMmxextFilterCodeSize,
1259 PROT_READ | PROT_WRITE,
1260 MAP_PRIVATE | MAP_ANONYMOUS,
1261 -1, 0);
1262 c->chrMmxextFilterCode = mmap(NULL, c->chrMmxextFilterCodeSize,
1263 PROT_READ | PROT_WRITE,
1264 MAP_PRIVATE | MAP_ANONYMOUS,
1265 -1, 0);
1266#elif HAVE_VIRTUALALLOC
1267 c->lumMmxextFilterCode = VirtualAlloc(NULL,
1268 c->lumMmxextFilterCodeSize,
1269 MEM_COMMIT,
1270 PAGE_EXECUTE_READWRITE);
1271 c->chrMmxextFilterCode = VirtualAlloc(NULL,
1272 c->chrMmxextFilterCodeSize,
1273 MEM_COMMIT,
1274 PAGE_EXECUTE_READWRITE);
1275#else
1276 c->lumMmxextFilterCode = av_malloc(c->lumMmxextFilterCodeSize);
1277 c->chrMmxextFilterCode = av_malloc(c->chrMmxextFilterCodeSize);
1278#endif
1279
1280#ifdef MAP_ANONYMOUS
1281 if (c->lumMmxextFilterCode == MAP_FAILED || c->chrMmxextFilterCode == MAP_FAILED)
1282#else
1283 if (!c->lumMmxextFilterCode || !c->chrMmxextFilterCode)
1284#endif
1285 {
1286 av_log(c, AV_LOG_ERROR, "Failed to allocate MMX2FilterCode\n");
1287 return AVERROR(ENOMEM);
1288 }
1289
1290 FF_ALLOCZ_OR_GOTO(c, c->hLumFilter, (dstW / 8 + 8) * sizeof(int16_t), fail);
1291 FF_ALLOCZ_OR_GOTO(c, c->hChrFilter, (c->chrDstW / 4 + 8) * sizeof(int16_t), fail);
1292 FF_ALLOCZ_OR_GOTO(c, c->hLumFilterPos, (dstW / 2 / 8 + 8) * sizeof(int32_t), fail);
1293 FF_ALLOCZ_OR_GOTO(c, c->hChrFilterPos, (c->chrDstW / 2 / 4 + 8) * sizeof(int32_t), fail);
1294
1295 ff_init_hscaler_mmxext( dstW, c->lumXInc, c->lumMmxextFilterCode,
1296 c->hLumFilter, (uint32_t*)c->hLumFilterPos, 8);
1297 ff_init_hscaler_mmxext(c->chrDstW, c->chrXInc, c->chrMmxextFilterCode,
1298 c->hChrFilter, (uint32_t*)c->hChrFilterPos, 4);
1299
1300#if USE_MMAP
1301 if ( mprotect(c->lumMmxextFilterCode, c->lumMmxextFilterCodeSize, PROT_EXEC | PROT_READ) == -1
1302 || mprotect(c->chrMmxextFilterCode, c->chrMmxextFilterCodeSize, PROT_EXEC | PROT_READ) == -1) {
1303 av_log(c, AV_LOG_ERROR, "mprotect failed, cannot use fast bilinear scaler\n");
1304 goto fail;
1305 }
1306#endif
1307 } else
1308#endif /* HAVE_MMXEXT_INLINE */
1309 {
1310 const int filterAlign = X86_MMX(cpu_flags) ? 4 :
1311 PPC_ALTIVEC(cpu_flags) ? 8 : 1;
1312
f6fa7814 1313 if ((ret = initFilter(&c->hLumFilter, &c->hLumFilterPos,
2ba45a60
DM		 1314 &c->hLumFilterSize, c->lumXInc,
1315 srcW, dstW, filterAlign, 1 << 14,
1316 (flags & SWS_BICUBLIN) ? (flags | SWS_BICUBIC) : flags,
1317 cpu_flags, srcFilter->lumH, dstFilter->lumH,
1318 c->param,
1319 get_local_pos(c, 0, 0, 0),
f6fa7814 1320 get_local_pos(c, 0, 0, 0))) < 0)
2ba45a60 1321 goto fail;
f6fa7814 1322 if ((ret = initFilter(&c->hChrFilter, &c->hChrFilterPos,
2ba45a60
DM		 1323 &c->hChrFilterSize, c->chrXInc,
1324 c->chrSrcW, c->chrDstW, filterAlign, 1 << 14,
1325 (flags & SWS_BICUBLIN) ? (flags | SWS_BILINEAR) : flags,
1326 cpu_flags, srcFilter->chrH, dstFilter->chrH,
1327 c->param,
1328 get_local_pos(c, c->chrSrcHSubSample, c->src_h_chr_pos, 0),
f6fa7814 1329 get_local_pos(c, c->chrDstHSubSample, c->dst_h_chr_pos, 0))) < 0)
2ba45a60
DM		 1330 goto fail;
1331 }
1332 } // initialize horizontal stuff
1333
1334 /* precalculate vertical scaler filter coefficients */
1335 {
1336 const int filterAlign = X86_MMX(cpu_flags) ? 2 :
1337 PPC_ALTIVEC(cpu_flags) ? 8 : 1;
1338
f6fa7814 1339 if ((ret = initFilter(&c->vLumFilter, &c->vLumFilterPos, &c->vLumFilterSize,
2ba45a60
DM		 1340 c->lumYInc, srcH, dstH, filterAlign, (1 << 12),
1341 (flags & SWS_BICUBLIN) ? (flags | SWS_BICUBIC) : flags,
1342 cpu_flags, srcFilter->lumV, dstFilter->lumV,
1343 c->param,
1344 get_local_pos(c, 0, 0, 1),
f6fa7814 1345 get_local_pos(c, 0, 0, 1))) < 0)
2ba45a60 1346 goto fail;
f6fa7814 1347 if ((ret = initFilter(&c->vChrFilter, &c->vChrFilterPos, &c->vChrFilterSize,
2ba45a60
DM		 1348 c->chrYInc, c->chrSrcH, c->chrDstH,
1349 filterAlign, (1 << 12),
1350 (flags & SWS_BICUBLIN) ? (flags | SWS_BILINEAR) : flags,
1351 cpu_flags, srcFilter->chrV, dstFilter->chrV,
1352 c->param,
1353 get_local_pos(c, c->chrSrcVSubSample, c->src_v_chr_pos, 1),
f6fa7814 1354 get_local_pos(c, c->chrDstVSubSample, c->dst_v_chr_pos, 1))) < 0)
2ba45a60
DM		 1355
1356 goto fail;
1357
1358#if HAVE_ALTIVEC
1359 FF_ALLOC_OR_GOTO(c, c->vYCoeffsBank, sizeof(vector signed short) * c->vLumFilterSize * c->dstH, fail);
1360 FF_ALLOC_OR_GOTO(c, c->vCCoeffsBank, sizeof(vector signed short) * c->vChrFilterSize * c->chrDstH, fail);
1361
1362 for (i = 0; i < c->vLumFilterSize * c->dstH; i++) {
1363 int j;
1364 short *p = (short *)&c->vYCoeffsBank[i];
1365 for (j = 0; j < 8; j++)
1366 p[j] = c->vLumFilter[i];
1367 }
1368
1369 for (i = 0; i < c->vChrFilterSize * c->chrDstH; i++) {
1370 int j;
1371 short *p = (short *)&c->vCCoeffsBank[i];
1372 for (j = 0; j < 8; j++)
1373 p[j] = c->vChrFilter[i];
1374 }
1375#endif
1376 }
1377
1378 // calculate buffer sizes so that they won't run out while handling these damn slices
1379 c->vLumBufSize = c->vLumFilterSize;
1380 c->vChrBufSize = c->vChrFilterSize;
1381 for (i = 0; i < dstH; i++) {
1382 int chrI = (int64_t)i * c->chrDstH / dstH;
1383 int nextSlice = FFMAX(c->vLumFilterPos[i] + c->vLumFilterSize - 1,
1384 ((c->vChrFilterPos[chrI] + c->vChrFilterSize - 1)
1385 << c->chrSrcVSubSample));
1386
1387 nextSlice >>= c->chrSrcVSubSample;
1388 nextSlice <<= c->chrSrcVSubSample;
1389 if (c->vLumFilterPos[i] + c->vLumBufSize < nextSlice)
1390 c->vLumBufSize = nextSlice - c->vLumFilterPos[i];
1391 if (c->vChrFilterPos[chrI] + c->vChrBufSize <
1392 (nextSlice >> c->chrSrcVSubSample))
1393 c->vChrBufSize = (nextSlice >> c->chrSrcVSubSample) -
1394 c->vChrFilterPos[chrI];
1395 }
1396
1397 for (i = 0; i < 4; i++)
1398 FF_ALLOCZ_OR_GOTO(c, c->dither_error[i], (c->dstW+2) * sizeof(int), fail);
1399
1400 /* Allocate pixbufs (we use dynamic allocation because otherwise we would
1401 * need to allocate several megabytes to handle all possible cases) */
1402 FF_ALLOC_OR_GOTO(c, c->lumPixBuf, c->vLumBufSize * 3 * sizeof(int16_t *), fail);
1403 FF_ALLOC_OR_GOTO(c, c->chrUPixBuf, c->vChrBufSize * 3 * sizeof(int16_t *), fail);
1404 FF_ALLOC_OR_GOTO(c, c->chrVPixBuf, c->vChrBufSize * 3 * sizeof(int16_t *), fail);
1405 if (CONFIG_SWSCALE_ALPHA && isALPHA(c->srcFormat) && isALPHA(c->dstFormat))
1406 FF_ALLOCZ_OR_GOTO(c, c->alpPixBuf, c->vLumBufSize * 3 * sizeof(int16_t *), fail);
1407 /* Note we need at least one pixel more at the end because of the MMX code
1408 * (just in case someone wants to replace the 4000/8000). */
1409 /* align at 16 bytes for AltiVec */
1410 for (i = 0; i < c->vLumBufSize; i++) {
1411 FF_ALLOCZ_OR_GOTO(c, c->lumPixBuf[i + c->vLumBufSize],
1412 dst_stride + 16, fail);
1413 c->lumPixBuf[i] = c->lumPixBuf[i + c->vLumBufSize];
1414 }
1415 // 64 / c->scalingBpp is the same as 16 / sizeof(scaling_intermediate)
1416 c->uv_off = (dst_stride>>1) + 64 / (c->dstBpc &~ 7);
1417 c->uv_offx2 = dst_stride + 16;
1418 for (i = 0; i < c->vChrBufSize; i++) {
1419 FF_ALLOC_OR_GOTO(c, c->chrUPixBuf[i + c->vChrBufSize],
1420 dst_stride * 2 + 32, fail);
1421 c->chrUPixBuf[i] = c->chrUPixBuf[i + c->vChrBufSize];
1422 c->chrVPixBuf[i] = c->chrVPixBuf[i + c->vChrBufSize]
1423 = c->chrUPixBuf[i] + (dst_stride >> 1) + 8;
1424 }
1425 if (CONFIG_SWSCALE_ALPHA && c->alpPixBuf)
1426 for (i = 0; i < c->vLumBufSize; i++) {
1427 FF_ALLOCZ_OR_GOTO(c, c->alpPixBuf[i + c->vLumBufSize],
1428 dst_stride + 16, fail);
1429 c->alpPixBuf[i] = c->alpPixBuf[i + c->vLumBufSize];
1430 }
1431
1432 // try to avoid drawing green stuff between the right end and the stride end
1433 for (i = 0; i < c->vChrBufSize; i++)
1434 if(desc_dst->comp[0].depth_minus1 == 15){
1435 av_assert0(c->dstBpc > 14);
1436 for(j=0; j<dst_stride/2+1; j++)
1437 ((int32_t*)(c->chrUPixBuf[i]))[j] = 1<<18;
1438 } else
1439 for(j=0; j<dst_stride+1; j++)
1440 ((int16_t*)(c->chrUPixBuf[i]))[j] = 1<<14;
1441
1442 av_assert0(c->chrDstH <= dstH);
1443
1444 if (flags & SWS_PRINT_INFO) {
1445 const char *scaler = NULL, *cpucaps;
1446
1447 for (i = 0; i < FF_ARRAY_ELEMS(scale_algorithms); i++) {
1448 if (flags & scale_algorithms[i].flag) {
1449 scaler = scale_algorithms[i].description;
1450 break;
1451 }
1452 }
1453 if (!scaler)
1454 scaler = "ehh flags invalid?!";
1455 av_log(c, AV_LOG_INFO, "%s scaler, from %s to %s%s ",
1456 scaler,
1457 av_get_pix_fmt_name(srcFormat),
1458#ifdef DITHER1XBPP
1459 dstFormat == AV_PIX_FMT_BGR555 || dstFormat == AV_PIX_FMT_BGR565 ||
1460 dstFormat == AV_PIX_FMT_RGB444BE || dstFormat == AV_PIX_FMT_RGB444LE ||
1461 dstFormat == AV_PIX_FMT_BGR444BE || dstFormat == AV_PIX_FMT_BGR444LE ?
1462 "dithered " : "",
1463#else
1464 "",
1465#endif
1466 av_get_pix_fmt_name(dstFormat));
1467
1468 if (INLINE_MMXEXT(cpu_flags))
1469 cpucaps = "MMXEXT";
1470 else if (INLINE_AMD3DNOW(cpu_flags))
1471 cpucaps = "3DNOW";
1472 else if (INLINE_MMX(cpu_flags))
1473 cpucaps = "MMX";
1474 else if (PPC_ALTIVEC(cpu_flags))
1475 cpucaps = "AltiVec";
1476 else
1477 cpucaps = "C";
1478
1479 av_log(c, AV_LOG_INFO, "using %s\n", cpucaps);
1480
1481 av_log(c, AV_LOG_VERBOSE, "%dx%d -> %dx%d\n", srcW, srcH, dstW, dstH);
1482 av_log(c, AV_LOG_DEBUG,
1483 "lum srcW=%d srcH=%d dstW=%d dstH=%d xInc=%d yInc=%d\n",
1484 c->srcW, c->srcH, c->dstW, c->dstH, c->lumXInc, c->lumYInc);
1485 av_log(c, AV_LOG_DEBUG,
1486 "chr srcW=%d srcH=%d dstW=%d dstH=%d xInc=%d yInc=%d\n",
1487 c->chrSrcW, c->chrSrcH, c->chrDstW, c->chrDstH,
1488 c->chrXInc, c->chrYInc);
1489 }
1490
1491 /* unscaled special cases */
1492 if (unscaled && !usesHFilter && !usesVFilter &&
1493 (c->srcRange == c->dstRange || isAnyRGB(dstFormat))) {
1494 ff_get_unscaled_swscale(c);
1495
1496 if (c->swscale) {
1497 if (flags & SWS_PRINT_INFO)
1498 av_log(c, AV_LOG_INFO,
1499 "using unscaled %s -> %s special converter\n",
1500 av_get_pix_fmt_name(srcFormat), av_get_pix_fmt_name(dstFormat));
1501 return 0;
1502 }
1503 }
1504
1505 c->swscale = ff_getSwsFunc(c);
1506 return 0;
1507fail: // FIXME replace things by appropriate error codes
f6fa7814
DM		 1508 if (ret == RETCODE_USE_CASCADE) {
1509 int tmpW = sqrt(srcW * (int64_t)dstW);
1510 int tmpH = sqrt(srcH * (int64_t)dstH);
1511 enum AVPixelFormat tmpFormat = AV_PIX_FMT_YUV420P;
1512
1513 if (srcW*(int64_t)srcH <= 4LL*dstW*dstH)
1514 return AVERROR(EINVAL);
1515
1516 ret = av_image_alloc(c->cascaded_tmp, c->cascaded_tmpStride,
1517 tmpW, tmpH, tmpFormat, 64);
1518 if (ret < 0)
1519 return ret;
1520
1521 c->cascaded_context[0] = sws_getContext(srcW, srcH, srcFormat,
1522 tmpW, tmpH, tmpFormat,
1523 flags, srcFilter, NULL, c->param);
1524 if (!c->cascaded_context[0])
1525 return -1;
1526
1527 c->cascaded_context[1] = sws_getContext(tmpW, tmpH, tmpFormat,
1528 dstW, dstH, dstFormat,
1529 flags, NULL, dstFilter, c->param);
1530 if (!c->cascaded_context[1])
1531 return -1;
1532 return 0;
1533 }
2ba45a60
DM		 1534 return -1;
1535}
1536
1537SwsContext *sws_getContext(int srcW, int srcH, enum AVPixelFormat srcFormat,
1538 int dstW, int dstH, enum AVPixelFormat dstFormat,
1539 int flags, SwsFilter *srcFilter,
1540 SwsFilter *dstFilter, const double *param)
1541{
1542 SwsContext *c;
1543
1544 if (!(c = sws_alloc_context()))
1545 return NULL;
1546
1547 c->flags = flags;
1548 c->srcW = srcW;
1549 c->srcH = srcH;
1550 c->dstW = dstW;
1551 c->dstH = dstH;
1552 c->srcFormat = srcFormat;
1553 c->dstFormat = dstFormat;
1554
1555 if (param) {
1556 c->param[0] = param[0];
1557 c->param[1] = param[1];
1558 }
1559
1560 if (sws_init_context(c, srcFilter, dstFilter) < 0) {
1561 sws_freeContext(c);
1562 return NULL;
1563 }
1564
1565 return c;
1566}
1567
1568SwsFilter *sws_getDefaultFilter(float lumaGBlur, float chromaGBlur,
1569 float lumaSharpen, float chromaSharpen,
1570 float chromaHShift, float chromaVShift,
1571 int verbose)
1572{
1573 SwsFilter *filter = av_malloc(sizeof(SwsFilter));
1574 if (!filter)
1575 return NULL;
1576
1577 if (lumaGBlur != 0.0) {
1578 filter->lumH = sws_getGaussianVec(lumaGBlur, 3.0);
1579 filter->lumV = sws_getGaussianVec(lumaGBlur, 3.0);
1580 } else {
1581 filter->lumH = sws_getIdentityVec();
1582 filter->lumV = sws_getIdentityVec();
1583 }
1584
1585 if (chromaGBlur != 0.0) {
1586 filter->chrH = sws_getGaussianVec(chromaGBlur, 3.0);
1587 filter->chrV = sws_getGaussianVec(chromaGBlur, 3.0);
1588 } else {
1589 filter->chrH = sws_getIdentityVec();
1590 filter->chrV = sws_getIdentityVec();
1591 }
1592
1593 if (chromaSharpen != 0.0) {
1594 SwsVector *id = sws_getIdentityVec();
1595 sws_scaleVec(filter->chrH, -chromaSharpen);
1596 sws_scaleVec(filter->chrV, -chromaSharpen);
1597 sws_addVec(filter->chrH, id);
1598 sws_addVec(filter->chrV, id);
1599 sws_freeVec(id);
1600 }
1601
1602 if (lumaSharpen != 0.0) {
1603 SwsVector *id = sws_getIdentityVec();
1604 sws_scaleVec(filter->lumH, -lumaSharpen);
1605 sws_scaleVec(filter->lumV, -lumaSharpen);
1606 sws_addVec(filter->lumH, id);
1607 sws_addVec(filter->lumV, id);
1608 sws_freeVec(id);
1609 }
1610
1611 if (chromaHShift != 0.0)
1612 sws_shiftVec(filter->chrH, (int)(chromaHShift + 0.5));
1613
1614 if (chromaVShift != 0.0)
1615 sws_shiftVec(filter->chrV, (int)(chromaVShift + 0.5));
1616
1617 sws_normalizeVec(filter->chrH, 1.0);
1618 sws_normalizeVec(filter->chrV, 1.0);
1619 sws_normalizeVec(filter->lumH, 1.0);
1620 sws_normalizeVec(filter->lumV, 1.0);
1621
1622 if (verbose)
1623 sws_printVec2(filter->chrH, NULL, AV_LOG_DEBUG);
1624 if (verbose)
1625 sws_printVec2(filter->lumH, NULL, AV_LOG_DEBUG);
1626
1627 return filter;
1628}
1629
1630SwsVector *sws_allocVec(int length)
1631{
1632 SwsVector *vec;
1633
1634 if(length <= 0 || length > INT_MAX/ sizeof(double))
1635 return NULL;
1636
1637 vec = av_malloc(sizeof(SwsVector));
1638 if (!vec)
1639 return NULL;
1640 vec->length = length;
1641 vec->coeff = av_malloc(sizeof(double) * length);
1642 if (!vec->coeff)
1643 av_freep(&vec);
1644 return vec;
1645}
1646
1647SwsVector *sws_getGaussianVec(double variance, double quality)
1648{
1649 const int length = (int)(variance * quality + 0.5) | 1;
1650 int i;
1651 double middle = (length - 1) * 0.5;
1652 SwsVector *vec;
1653
1654 if(variance < 0 || quality < 0)
1655 return NULL;
1656
1657 vec = sws_allocVec(length);
1658
1659 if (!vec)
1660 return NULL;
1661
1662 for (i = 0; i < length; i++) {
1663 double dist = i - middle;
1664 vec->coeff[i] = exp(-dist * dist / (2 * variance * variance)) /
1665 sqrt(2 * variance * M_PI);
1666 }
1667
1668 sws_normalizeVec(vec, 1.0);
1669
1670 return vec;
1671}
1672
1673SwsVector *sws_getConstVec(double c, int length)
1674{
1675 int i;
1676 SwsVector *vec = sws_allocVec(length);
1677
1678 if (!vec)
1679 return NULL;
1680
1681 for (i = 0; i < length; i++)
1682 vec->coeff[i] = c;
1683
1684 return vec;
1685}
1686
1687SwsVector *sws_getIdentityVec(void)
1688{
1689 return sws_getConstVec(1.0, 1);
1690}
1691
1692static double sws_dcVec(SwsVector *a)
1693{
1694 int i;
1695 double sum = 0;
1696
1697 for (i = 0; i < a->length; i++)
1698 sum += a->coeff[i];
1699
1700 return sum;
1701}
1702
1703void sws_scaleVec(SwsVector *a, double scalar)
1704{
1705 int i;
1706
1707 for (i = 0; i < a->length; i++)
1708 a->coeff[i] *= scalar;
1709}
1710
1711void sws_normalizeVec(SwsVector *a, double height)
1712{
1713 sws_scaleVec(a, height / sws_dcVec(a));
1714}
1715
1716static SwsVector *sws_getConvVec(SwsVector *a, SwsVector *b)
1717{
1718 int length = a->length + b->length - 1;
1719 int i, j;
1720 SwsVector *vec = sws_getConstVec(0.0, length);
1721
1722 if (!vec)
1723 return NULL;
1724
1725 for (i = 0; i < a->length; i++) {
1726 for (j = 0; j < b->length; j++) {
1727 vec->coeff[i + j] += a->coeff[i] * b->coeff[j];
1728 }
1729 }
1730
1731 return vec;
1732}
1733
1734static SwsVector *sws_sumVec(SwsVector *a, SwsVector *b)
1735{
1736 int length = FFMAX(a->length, b->length);
1737 int i;
1738 SwsVector *vec = sws_getConstVec(0.0, length);
1739
1740 if (!vec)
1741 return NULL;
1742
1743 for (i = 0; i < a->length; i++)
1744 vec->coeff[i + (length - 1) / 2 - (a->length - 1) / 2] += a->coeff[i];
1745 for (i = 0; i < b->length; i++)
1746 vec->coeff[i + (length - 1) / 2 - (b->length - 1) / 2] += b->coeff[i];
1747
1748 return vec;
1749}
1750
1751static SwsVector *sws_diffVec(SwsVector *a, SwsVector *b)
1752{
1753 int length = FFMAX(a->length, b->length);
1754 int i;
1755 SwsVector *vec = sws_getConstVec(0.0, length);
1756
1757 if (!vec)
1758 return NULL;
1759
1760 for (i = 0; i < a->length; i++)
1761 vec->coeff[i + (length - 1) / 2 - (a->length - 1) / 2] += a->coeff[i];
1762 for (i = 0; i < b->length; i++)
1763 vec->coeff[i + (length - 1) / 2 - (b->length - 1) / 2] -= b->coeff[i];
1764
1765 return vec;
1766}
1767
1768/* shift left / or right if "shift" is negative */
1769static SwsVector *sws_getShiftedVec(SwsVector *a, int shift)
1770{
1771 int length = a->length + FFABS(shift) * 2;
1772 int i;
1773 SwsVector *vec = sws_getConstVec(0.0, length);
1774
1775 if (!vec)
1776 return NULL;
1777
1778 for (i = 0; i < a->length; i++) {
1779 vec->coeff[i + (length - 1) / 2 -
1780 (a->length - 1) / 2 - shift] = a->coeff[i];
1781 }
1782
1783 return vec;
1784}
1785
1786void sws_shiftVec(SwsVector *a, int shift)
1787{
1788 SwsVector *shifted = sws_getShiftedVec(a, shift);
1789 av_free(a->coeff);
1790 a->coeff = shifted->coeff;
1791 a->length = shifted->length;
1792 av_free(shifted);
1793}
1794
1795void sws_addVec(SwsVector *a, SwsVector *b)
1796{
1797 SwsVector *sum = sws_sumVec(a, b);
1798 av_free(a->coeff);
1799 a->coeff = sum->coeff;
1800 a->length = sum->length;
1801 av_free(sum);
1802}
1803
1804void sws_subVec(SwsVector *a, SwsVector *b)
1805{
1806 SwsVector *diff = sws_diffVec(a, b);
1807 av_free(a->coeff);
1808 a->coeff = diff->coeff;
1809 a->length = diff->length;
1810 av_free(diff);
1811}
1812
1813void sws_convVec(SwsVector *a, SwsVector *b)
1814{
1815 SwsVector *conv = sws_getConvVec(a, b);
1816 av_free(a->coeff);
1817 a->coeff = conv->coeff;
1818 a->length = conv->length;
1819 av_free(conv);
1820}
1821
1822SwsVector *sws_cloneVec(SwsVector *a)
1823{
1824 SwsVector *vec = sws_allocVec(a->length);
1825
1826 if (!vec)
1827 return NULL;
1828
1829 memcpy(vec->coeff, a->coeff, a->length * sizeof(*a->coeff));
1830
1831 return vec;
1832}
1833
1834void sws_printVec2(SwsVector *a, AVClass *log_ctx, int log_level)
1835{
1836 int i;
1837 double max = 0;
1838 double min = 0;
1839 double range;
1840
1841 for (i = 0; i < a->length; i++)
1842 if (a->coeff[i] > max)
1843 max = a->coeff[i];
1844
1845 for (i = 0; i < a->length; i++)
1846 if (a->coeff[i] < min)
1847 min = a->coeff[i];
1848
1849 range = max - min;
1850
1851 for (i = 0; i < a->length; i++) {
1852 int x = (int)((a->coeff[i] - min) * 60.0 / range + 0.5);
1853 av_log(log_ctx, log_level, "%1.3f ", a->coeff[i]);
1854 for (; x > 0; x--)
1855 av_log(log_ctx, log_level, " ");
1856 av_log(log_ctx, log_level, "|\n");
1857 }
1858}
1859
1860void sws_freeVec(SwsVector *a)
1861{
1862 if (!a)
1863 return;
1864 av_freep(&a->coeff);
1865 a->length = 0;
1866 av_free(a);
1867}
1868
1869void sws_freeFilter(SwsFilter *filter)
1870{
1871 if (!filter)
1872 return;
1873
1874 sws_freeVec(filter->lumH);
1875 sws_freeVec(filter->lumV);
1876 sws_freeVec(filter->chrH);
1877 sws_freeVec(filter->chrV);
1878 av_free(filter);
1879}
1880
1881void sws_freeContext(SwsContext *c)
1882{
1883 int i;
1884 if (!c)
1885 return;
1886
1887 if (c->lumPixBuf) {
1888 for (i = 0; i < c->vLumBufSize; i++)
1889 av_freep(&c->lumPixBuf[i]);
1890 av_freep(&c->lumPixBuf);
1891 }
1892
1893 if (c->chrUPixBuf) {
1894 for (i = 0; i < c->vChrBufSize; i++)
1895 av_freep(&c->chrUPixBuf[i]);
1896 av_freep(&c->chrUPixBuf);
1897 av_freep(&c->chrVPixBuf);
1898 }
1899
1900 if (CONFIG_SWSCALE_ALPHA && c->alpPixBuf) {
1901 for (i = 0; i < c->vLumBufSize; i++)
1902 av_freep(&c->alpPixBuf[i]);
1903 av_freep(&c->alpPixBuf);
1904 }
1905
1906 for (i = 0; i < 4; i++)
1907 av_freep(&c->dither_error[i]);
1908
1909 av_freep(&c->vLumFilter);
1910 av_freep(&c->vChrFilter);
1911 av_freep(&c->hLumFilter);
1912 av_freep(&c->hChrFilter);
1913#if HAVE_ALTIVEC
1914 av_freep(&c->vYCoeffsBank);
1915 av_freep(&c->vCCoeffsBank);
1916#endif
1917
1918 av_freep(&c->vLumFilterPos);
1919 av_freep(&c->vChrFilterPos);
1920 av_freep(&c->hLumFilterPos);
1921 av_freep(&c->hChrFilterPos);
1922
1923#if HAVE_MMX_INLINE
1924#if USE_MMAP
1925 if (c->lumMmxextFilterCode)
1926 munmap(c->lumMmxextFilterCode, c->lumMmxextFilterCodeSize);
1927 if (c->chrMmxextFilterCode)
1928 munmap(c->chrMmxextFilterCode, c->chrMmxextFilterCodeSize);
1929#elif HAVE_VIRTUALALLOC
1930 if (c->lumMmxextFilterCode)
1931 VirtualFree(c->lumMmxextFilterCode, 0, MEM_RELEASE);
1932 if (c->chrMmxextFilterCode)
1933 VirtualFree(c->chrMmxextFilterCode, 0, MEM_RELEASE);
1934#else
1935 av_free(c->lumMmxextFilterCode);
1936 av_free(c->chrMmxextFilterCode);
1937#endif
1938 c->lumMmxextFilterCode = NULL;
1939 c->chrMmxextFilterCode = NULL;
1940#endif /* HAVE_MMX_INLINE */
1941
1942 av_freep(&c->yuvTable);
1943 av_freep(&c->formatConvBuffer);
1944
f6fa7814
DM		 1945 sws_freeContext(c->cascaded_context[0]);
1946 sws_freeContext(c->cascaded_context[1]);
1947 memset(c->cascaded_context, 0, sizeof(c->cascaded_context));
1948 av_freep(&c->cascaded_tmp[0]);
1949
2ba45a60
DM		 1950 av_free(c);
1951}
1952
1953struct SwsContext *sws_getCachedContext(struct SwsContext *context, int srcW,
1954 int srcH, enum AVPixelFormat srcFormat,
1955 int dstW, int dstH,
1956 enum AVPixelFormat dstFormat, int flags,
1957 SwsFilter *srcFilter,
1958 SwsFilter *dstFilter,
1959 const double *param)
1960{
1961 static const double default_param[2] = { SWS_PARAM_DEFAULT,
1962 SWS_PARAM_DEFAULT };
1963
1964 if (!param)
1965 param = default_param;
1966
1967 if (context &&
1968 (context->srcW != srcW ||
1969 context->srcH != srcH ||
1970 context->srcFormat != srcFormat ||
1971 context->dstW != dstW ||
1972 context->dstH != dstH ||
1973 context->dstFormat != dstFormat ||
1974 context->flags != flags ||
1975 context->param[0] != param[0] ||
1976 context->param[1] != param[1])) {
1977 sws_freeContext(context);
1978 context = NULL;
1979 }
1980
1981 if (!context) {
1982 if (!(context = sws_alloc_context()))
1983 return NULL;
1984 context->srcW = srcW;
1985 context->srcH = srcH;
1986 context->srcFormat = srcFormat;
1987 context->dstW = dstW;
1988 context->dstH = dstH;
1989 context->dstFormat = dstFormat;
1990 context->flags = flags;
1991 context->param[0] = param[0];
1992 context->param[1] = param[1];
1993 if (sws_init_context(context, srcFilter, dstFilter) < 0) {
1994 sws_freeContext(context);
1995 return NULL;
1996 }
1997 }
1998 return context;
1999}
ffmpeg Piment Noir Debian package repository