2 * Copyright (C) 2001-2003 Michael Niedermayer <michaelni@gmx.at>
4 * This file is part of Libav.
6 * Libav is free software; you can redistribute it and/or
7 * modify it under the terms of the GNU Lesser General Public
8 * License as published by the Free Software Foundation; either
9 * version 2.1 of the License, or (at your option) any later version.
11 * Libav is distributed in the hope that it will be useful,
12 * but WITHOUT ANY WARRANTY; without even the implied warranty of
13 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
14 * Lesser General Public License for more details.
16 * You should have received a copy of the GNU Lesser General Public
17 * License along with Libav; if not, write to the Free Software
18 * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
21 #ifndef SWSCALE_SWSCALE_INTERNAL_H
22 #define SWSCALE_SWSCALE_INTERNAL_H
30 #include "libavutil/avutil.h"
32 #define STR(s) AV_TOSTRING(s) //AV_STRINGIFY is too long
34 #define FAST_BGR2YV12 //use 7-bit instead of 15-bit coefficients
36 #define MAX_FILTER_SIZE 256
39 #define ALT32_CORR (-1)
56 typedef int (*SwsFunc)(struct SwsContext *context, const uint8_t* src[],
57 int srcStride[], int srcSliceY, int srcSliceH,
58 uint8_t* dst[], int dstStride[]);
60 /* This struct should be aligned on at least a 32-byte boundary. */
61 typedef struct SwsContext {
63 * info on struct for av_log
65 const AVClass *av_class;
68 * Note that src, dst, srcStride, dstStride will be copied in the
69 * sws_scale() wrapper so they can be freely modified here.
72 int srcW; ///< Width of source luma/alpha planes.
73 int srcH; ///< Height of source luma/alpha planes.
74 int dstH; ///< Height of destination luma/alpha planes.
75 int chrSrcW; ///< Width of source chroma planes.
76 int chrSrcH; ///< Height of source chroma planes.
77 int chrDstW; ///< Width of destination chroma planes.
78 int chrDstH; ///< Height of destination chroma planes.
81 enum PixelFormat dstFormat; ///< Destination pixel format.
82 enum PixelFormat srcFormat; ///< Source pixel format.
83 int dstFormatBpp; ///< Number of bits per pixel of the destination pixel format.
84 int srcFormatBpp; ///< Number of bits per pixel of the source pixel format.
85 int chrSrcHSubSample; ///< Binary logarithm of horizontal subsampling factor between luma/alpha and chroma planes in source image.
86 int chrSrcVSubSample; ///< Binary logarithm of vertical subsampling factor between luma/alpha and chroma planes in source image.
87 int chrDstHSubSample; ///< Binary logarithm of horizontal subsampling factor between luma/alpha and chroma planes in destination image.
88 int chrDstVSubSample; ///< Binary logarithm of vertical subsampling factor between luma/alpha and chroma planes in destination image.
89 int vChrDrop; ///< Binary logarithm of extra vertical subsampling factor in source image chroma planes specified by user.
90 int sliceDir; ///< Direction that slices are fed to the scaler (1 = top-to-bottom, -1 = bottom-to-top).
91 double param[2]; ///< Input parameters for scaling algorithms that need them.
93 uint32_t pal_yuv[256];
94 uint32_t pal_rgb[256];
97 * @name Scaled horizontal lines ring buffer.
98 * The horizontal scaler keeps just enough scaled lines in a ring buffer
99 * so they may be passed to the vertical scaler. The pointers to the
100 * allocated buffers for each line are duplicated in sequence in the ring
101 * buffer to simplify indexing and avoid wrapping around between lines
102 * inside the vertical scaler code. The wrapping is done before the
103 * vertical scaler is called.
106 int16_t **lumPixBuf; ///< Ring buffer for scaled horizontal luma plane lines to be fed to the vertical scaler.
107 int16_t **chrUPixBuf; ///< Ring buffer for scaled horizontal chroma plane lines to be fed to the vertical scaler.
108 int16_t **chrVPixBuf; ///< Ring buffer for scaled horizontal chroma plane lines to be fed to the vertical scaler.
109 int16_t **alpPixBuf; ///< Ring buffer for scaled horizontal alpha plane lines to be fed to the vertical scaler.
110 int vLumBufSize; ///< Number of vertical luma/alpha lines allocated in the ring buffer.
111 int vChrBufSize; ///< Number of vertical chroma lines allocated in the ring buffer.
112 int lastInLumBuf; ///< Last scaled horizontal luma/alpha line from source in the ring buffer.
113 int lastInChrBuf; ///< Last scaled horizontal chroma line from source in the ring buffer.
114 int lumBufIndex; ///< Index in ring buffer of the last scaled horizontal luma/alpha line from source.
115 int chrBufIndex; ///< Index in ring buffer of the last scaled horizontal chroma line from source.
118 uint8_t *formatConvBuffer;
121 * @name Horizontal and vertical filters.
122 * To better understand the following fields, here is a pseudo-code of
123 * their usage in filtering a horizontal line:
125 * for (i = 0; i < width; i++) {
127 * for (j = 0; j < filterSize; j++)
128 * dst[i] += src[ filterPos[i] + j ] * filter[ filterSize * i + j ];
129 * dst[i] >>= FRAC_BITS; // The actual implementation is fixed-point.
134 int16_t *hLumFilter; ///< Array of horizontal filter coefficients for luma/alpha planes.
135 int16_t *hChrFilter; ///< Array of horizontal filter coefficients for chroma planes.
136 int16_t *vLumFilter; ///< Array of vertical filter coefficients for luma/alpha planes.
137 int16_t *vChrFilter; ///< Array of vertical filter coefficients for chroma planes.
138 int16_t *hLumFilterPos; ///< Array of horizontal filter starting positions for each dst[i] for luma/alpha planes.
139 int16_t *hChrFilterPos; ///< Array of horizontal filter starting positions for each dst[i] for chroma planes.
140 int16_t *vLumFilterPos; ///< Array of vertical filter starting positions for each dst[i] for luma/alpha planes.
141 int16_t *vChrFilterPos; ///< Array of vertical filter starting positions for each dst[i] for chroma planes.
142 int hLumFilterSize; ///< Horizontal filter size for luma/alpha pixels.
143 int hChrFilterSize; ///< Horizontal filter size for chroma pixels.
144 int vLumFilterSize; ///< Vertical filter size for luma/alpha pixels.
145 int vChrFilterSize; ///< Vertical filter size for chroma pixels.
148 int lumMmx2FilterCodeSize; ///< Runtime-generated MMX2 horizontal fast bilinear scaler code size for luma/alpha planes.
149 int chrMmx2FilterCodeSize; ///< Runtime-generated MMX2 horizontal fast bilinear scaler code size for chroma planes.
150 uint8_t *lumMmx2FilterCode; ///< Runtime-generated MMX2 horizontal fast bilinear scaler code for luma/alpha planes.
151 uint8_t *chrMmx2FilterCode; ///< Runtime-generated MMX2 horizontal fast bilinear scaler code for chroma planes.
155 int dstY; ///< Last destination vertical line output from last slice.
156 int flags; ///< Flags passed by the user to select scaler algorithm, optimizations, subsampling, etc...
157 void * yuvTable; // pointer to the yuv->rgb table start so it can be freed()
158 uint8_t * table_rV[256];
159 uint8_t * table_gU[256];
161 uint8_t * table_bU[256];
164 int contrast, brightness, saturation; // for sws_getColorspaceDetails
165 int srcColorspaceTable[4];
166 int dstColorspaceTable[4];
167 int srcRange; ///< 0 = MPG YUV range, 1 = JPG YUV range (source image).
168 int dstRange; ///< 0 = MPG YUV range, 1 = JPG YUV range (destination image).
169 int yuv2rgb_y_offset;
171 int yuv2rgb_v2r_coeff;
172 int yuv2rgb_v2g_coeff;
173 int yuv2rgb_u2g_coeff;
174 int yuv2rgb_u2b_coeff;
176 #define RED_DITHER "0*8"
177 #define GREEN_DITHER "1*8"
178 #define BLUE_DITHER "2*8"
179 #define Y_COEFF "3*8"
180 #define VR_COEFF "4*8"
181 #define UB_COEFF "5*8"
182 #define VG_COEFF "6*8"
183 #define UG_COEFF "7*8"
184 #define Y_OFFSET "8*8"
185 #define U_OFFSET "9*8"
186 #define V_OFFSET "10*8"
187 #define LUM_MMX_FILTER_OFFSET "11*8"
188 #define CHR_MMX_FILTER_OFFSET "11*8+4*4*256"
189 #define DSTW_OFFSET "11*8+4*4*256*2" //do not change, it is hardcoded in the ASM
190 #define ESP_OFFSET "11*8+4*4*256*2+8"
191 #define VROUNDER_OFFSET "11*8+4*4*256*2+16"
192 #define U_TEMP "11*8+4*4*256*2+24"
193 #define V_TEMP "11*8+4*4*256*2+32"
194 #define Y_TEMP "11*8+4*4*256*2+40"
195 #define ALP_MMX_FILTER_OFFSET "11*8+4*4*256*2+48"
196 #define UV_OFF "11*8+4*4*256*3+48"
197 #define UV_OFFx2 "11*8+4*4*256*3+56"
199 DECLARE_ALIGNED(8, uint64_t, redDither);
200 DECLARE_ALIGNED(8, uint64_t, greenDither);
201 DECLARE_ALIGNED(8, uint64_t, blueDither);
203 DECLARE_ALIGNED(8, uint64_t, yCoeff);
204 DECLARE_ALIGNED(8, uint64_t, vrCoeff);
205 DECLARE_ALIGNED(8, uint64_t, ubCoeff);
206 DECLARE_ALIGNED(8, uint64_t, vgCoeff);
207 DECLARE_ALIGNED(8, uint64_t, ugCoeff);
208 DECLARE_ALIGNED(8, uint64_t, yOffset);
209 DECLARE_ALIGNED(8, uint64_t, uOffset);
210 DECLARE_ALIGNED(8, uint64_t, vOffset);
211 int32_t lumMmxFilter[4*MAX_FILTER_SIZE];
212 int32_t chrMmxFilter[4*MAX_FILTER_SIZE];
213 int dstW; ///< Width of destination luma/alpha planes.
214 DECLARE_ALIGNED(8, uint64_t, esp);
215 DECLARE_ALIGNED(8, uint64_t, vRounder);
216 DECLARE_ALIGNED(8, uint64_t, u_temp);
217 DECLARE_ALIGNED(8, uint64_t, v_temp);
218 DECLARE_ALIGNED(8, uint64_t, y_temp);
219 int32_t alpMmxFilter[4*MAX_FILTER_SIZE];
220 DECLARE_ALIGNED(8, ptrdiff_t, uv_off); ///< offset (in pixels) between u and v planes
221 DECLARE_ALIGNED(8, ptrdiff_t, uv_offx2); ///< offset (in bytes) between u and v planes
224 vector signed short CY;
225 vector signed short CRV;
226 vector signed short CBU;
227 vector signed short CGU;
228 vector signed short CGV;
229 vector signed short OY;
230 vector unsigned short CSHIFT;
231 vector signed short *vYCoeffsBank, *vCCoeffsBank;
235 DECLARE_ALIGNED(4, uint32_t, oy);
236 DECLARE_ALIGNED(4, uint32_t, oc);
237 DECLARE_ALIGNED(4, uint32_t, zero);
238 DECLARE_ALIGNED(4, uint32_t, cy);
239 DECLARE_ALIGNED(4, uint32_t, crv);
240 DECLARE_ALIGNED(4, uint32_t, rmask);
241 DECLARE_ALIGNED(4, uint32_t, cbu);
242 DECLARE_ALIGNED(4, uint32_t, bmask);
243 DECLARE_ALIGNED(4, uint32_t, cgu);
244 DECLARE_ALIGNED(4, uint32_t, cgv);
245 DECLARE_ALIGNED(4, uint32_t, gmask);
249 DECLARE_ALIGNED(8, uint64_t, sparc_coeffs)[10];
252 /* function pointers for swScale() */
253 void (*yuv2nv12X )(struct SwsContext *c,
254 const int16_t *lumFilter, const int16_t **lumSrc, int lumFilterSize,
255 const int16_t *chrFilter, const int16_t **chrUSrc,
256 const int16_t **chrVSrc, int chrFilterSize,
257 uint8_t *dest, uint8_t *uDest,
258 int dstW, int chrDstW, int dstFormat);
259 void (*yuv2yuv1 )(struct SwsContext *c,
260 const int16_t *lumSrc, const int16_t *chrUSrc,
261 const int16_t *chrVSrc, const int16_t *alpSrc,
263 uint8_t *uDest, uint8_t *vDest, uint8_t *aDest,
264 int dstW, int chrDstW);
265 void (*yuv2yuvX )(struct SwsContext *c,
266 const int16_t *lumFilter, const int16_t **lumSrc, int lumFilterSize,
267 const int16_t *chrFilter, const int16_t **chrUSrc,
268 const int16_t **chrVSrc, int chrFilterSize,
269 const int16_t **alpSrc,
271 uint8_t *uDest, uint8_t *vDest, uint8_t *aDest,
272 int dstW, int chrDstW);
273 void (*yuv2packed1)(struct SwsContext *c,
274 const uint16_t *buf0,
275 const uint16_t *ubuf0, const uint16_t *ubuf1,
276 const uint16_t *vbuf0, const uint16_t *vbuf1,
277 const uint16_t *abuf0,
279 int dstW, int uvalpha, int dstFormat, int flags, int y);
280 void (*yuv2packed2)(struct SwsContext *c,
281 const uint16_t *buf0, const uint16_t *buf1,
282 const uint16_t *ubuf0, const uint16_t *ubuf1,
283 const uint16_t *vbuf0, const uint16_t *vbuf1,
284 const uint16_t *abuf0, const uint16_t *abuf1,
286 int dstW, int yalpha, int uvalpha, int y);
287 void (*yuv2packedX)(struct SwsContext *c,
288 const int16_t *lumFilter, const int16_t **lumSrc, int lumFilterSize,
289 const int16_t *chrFilter, const int16_t **chrUSrc,
290 const int16_t **chrVSrc, int chrFilterSize,
291 const int16_t **alpSrc, uint8_t *dest,
294 void (*lumToYV12)(uint8_t *dst, const uint8_t *src,
295 int width, uint32_t *pal); ///< Unscaled conversion of luma plane to YV12 for horizontal scaler.
296 void (*alpToYV12)(uint8_t *dst, const uint8_t *src,
297 int width, uint32_t *pal); ///< Unscaled conversion of alpha plane to YV12 for horizontal scaler.
298 void (*chrToYV12)(uint8_t *dstU, uint8_t *dstV,
299 const uint8_t *src1, const uint8_t *src2,
300 int width, uint32_t *pal); ///< Unscaled conversion of chroma planes to YV12 for horizontal scaler.
301 void (*hyscale_fast)(struct SwsContext *c,
302 int16_t *dst, int dstWidth,
303 const uint8_t *src, int srcW, int xInc);
304 void (*hcscale_fast)(struct SwsContext *c,
305 int16_t *dst1, int16_t *dst2, int dstWidth,
306 const uint8_t *src1, const uint8_t *src2,
309 void (*hScale)(int16_t *dst, int dstW, const uint8_t *src, int srcW,
310 int xInc, const int16_t *filter, const int16_t *filterPos,
313 void (*lumConvertRange)(uint16_t *dst, int width); ///< Color range conversion function for luma plane if needed.
314 void (*chrConvertRange)(uint16_t *dst1, uint16_t *dst2, int width); ///< Color range conversion function for chroma planes if needed.
316 int lumSrcOffset; ///< Offset given to luma src pointers passed to horizontal input functions.
317 int chrSrcOffset; ///< Offset given to chroma src pointers passed to horizontal input functions.
318 int alpSrcOffset; ///< Offset given to alpha src pointers passed to horizontal input functions.
320 int needs_hcscale; ///< Set if there are chroma planes to be converted.
323 //FIXME check init (where 0)
325 SwsFunc ff_yuv2rgb_get_func_ptr(SwsContext *c);
326 int ff_yuv2rgb_c_init_tables(SwsContext *c, const int inv_table[4],
327 int fullRange, int brightness,
328 int contrast, int saturation);
330 void ff_yuv2rgb_init_tables_altivec(SwsContext *c, const int inv_table[4],
331 int brightness, int contrast, int saturation);
332 void updateMMXDitherTables(SwsContext *c, int dstY, int lumBufIndex, int chrBufIndex,
333 int lastInLumBuf, int lastInChrBuf);
335 SwsFunc ff_yuv2rgb_init_mmx(SwsContext *c);
336 SwsFunc ff_yuv2rgb_init_vis(SwsContext *c);
337 SwsFunc ff_yuv2rgb_init_mlib(SwsContext *c);
338 SwsFunc ff_yuv2rgb_init_altivec(SwsContext *c);
339 SwsFunc ff_yuv2rgb_get_func_ptr_bfin(SwsContext *c);
340 void ff_bfin_get_unscaled_swscale(SwsContext *c);
342 const char *sws_format_name(enum PixelFormat format);
344 //FIXME replace this with something faster
345 #define is16BPS(x) ( \
346 (x)==PIX_FMT_GRAY16BE \
347 || (x)==PIX_FMT_GRAY16LE \
348 || (x)==PIX_FMT_BGR48BE \
349 || (x)==PIX_FMT_BGR48LE \
350 || (x)==PIX_FMT_RGB48BE \
351 || (x)==PIX_FMT_RGB48LE \
352 || (x)==PIX_FMT_YUV420P16LE \
353 || (x)==PIX_FMT_YUV422P16LE \
354 || (x)==PIX_FMT_YUV444P16LE \
355 || (x)==PIX_FMT_YUV420P16BE \
356 || (x)==PIX_FMT_YUV422P16BE \
357 || (x)==PIX_FMT_YUV444P16BE \
359 #define is9_OR_10BPS(x) ( \
360 (x)==PIX_FMT_YUV420P9LE \
361 || (x)==PIX_FMT_YUV420P9BE \
362 || (x)==PIX_FMT_YUV420P10LE \
363 || (x)==PIX_FMT_YUV420P10BE \
365 #define isBE(x) ((x)&1)
366 #define isPlanar8YUV(x) ( \
367 (x)==PIX_FMT_YUV410P \
368 || (x)==PIX_FMT_YUV420P \
369 || (x)==PIX_FMT_YUVA420P \
370 || (x)==PIX_FMT_YUV411P \
371 || (x)==PIX_FMT_YUV422P \
372 || (x)==PIX_FMT_YUV444P \
373 || (x)==PIX_FMT_YUV440P \
374 || (x)==PIX_FMT_NV12 \
375 || (x)==PIX_FMT_NV21 \
377 #define isPlanarYUV(x) ( \
379 || (x)==PIX_FMT_YUV420P9LE \
380 || (x)==PIX_FMT_YUV420P10LE \
381 || (x)==PIX_FMT_YUV420P16LE \
382 || (x)==PIX_FMT_YUV422P16LE \
383 || (x)==PIX_FMT_YUV444P16LE \
384 || (x)==PIX_FMT_YUV420P9BE \
385 || (x)==PIX_FMT_YUV420P10BE \
386 || (x)==PIX_FMT_YUV420P16BE \
387 || (x)==PIX_FMT_YUV422P16BE \
388 || (x)==PIX_FMT_YUV444P16BE \
391 (x)==PIX_FMT_UYVY422 \
392 || (x)==PIX_FMT_YUYV422 \
395 #define isGray(x) ( \
397 || (x)==PIX_FMT_Y400A \
398 || (x)==PIX_FMT_GRAY16BE \
399 || (x)==PIX_FMT_GRAY16LE \
401 #define isGray16(x) ( \
402 (x)==PIX_FMT_GRAY16BE \
403 || (x)==PIX_FMT_GRAY16LE \
405 #define isRGBinInt(x) ( \
406 (x)==PIX_FMT_RGB48BE \
407 || (x)==PIX_FMT_RGB48LE \
408 || (x)==PIX_FMT_RGB32 \
409 || (x)==PIX_FMT_RGB32_1 \
410 || (x)==PIX_FMT_RGB24 \
411 || (x)==PIX_FMT_RGB565BE \
412 || (x)==PIX_FMT_RGB565LE \
413 || (x)==PIX_FMT_RGB555BE \
414 || (x)==PIX_FMT_RGB555LE \
415 || (x)==PIX_FMT_RGB444BE \
416 || (x)==PIX_FMT_RGB444LE \
417 || (x)==PIX_FMT_RGB8 \
418 || (x)==PIX_FMT_RGB4 \
419 || (x)==PIX_FMT_RGB4_BYTE \
420 || (x)==PIX_FMT_MONOBLACK \
421 || (x)==PIX_FMT_MONOWHITE \
423 #define isBGRinInt(x) ( \
424 (x)==PIX_FMT_BGR48BE \
425 || (x)==PIX_FMT_BGR48LE \
426 || (x)==PIX_FMT_BGR32 \
427 || (x)==PIX_FMT_BGR32_1 \
428 || (x)==PIX_FMT_BGR24 \
429 || (x)==PIX_FMT_BGR565BE \
430 || (x)==PIX_FMT_BGR565LE \
431 || (x)==PIX_FMT_BGR555BE \
432 || (x)==PIX_FMT_BGR555LE \
433 || (x)==PIX_FMT_BGR444BE \
434 || (x)==PIX_FMT_BGR444LE \
435 || (x)==PIX_FMT_BGR8 \
436 || (x)==PIX_FMT_BGR4 \
437 || (x)==PIX_FMT_BGR4_BYTE \
438 || (x)==PIX_FMT_MONOBLACK \
439 || (x)==PIX_FMT_MONOWHITE \
441 #define isRGBinBytes(x) ( \
442 (x)==PIX_FMT_RGB48BE \
443 || (x)==PIX_FMT_RGB48LE \
444 || (x)==PIX_FMT_RGBA \
445 || (x)==PIX_FMT_ARGB \
446 || (x)==PIX_FMT_RGB24 \
448 #define isBGRinBytes(x) ( \
449 (x)==PIX_FMT_BGR48BE \
450 || (x)==PIX_FMT_BGR48LE \
451 || (x)==PIX_FMT_BGRA \
452 || (x)==PIX_FMT_ABGR \
453 || (x)==PIX_FMT_BGR24 \
455 #define isAnyRGB(x) ( \
459 #define isALPHA(x) ( \
461 || (x)==PIX_FMT_BGR32_1 \
462 || (x)==PIX_FMT_RGB32 \
463 || (x)==PIX_FMT_RGB32_1 \
464 || (x)==PIX_FMT_Y400A \
465 || (x)==PIX_FMT_YUVA420P \
467 #define isPacked(x) ( \
469 || (x)==PIX_FMT_YUYV422 \
470 || (x)==PIX_FMT_UYVY422 \
471 || (x)==PIX_FMT_Y400A \
474 #define usePal(x) ((av_pix_fmt_descriptors[x].flags & PIX_FMT_PAL) || (x) == PIX_FMT_Y400A)
476 extern const uint64_t ff_dither4[2];
477 extern const uint64_t ff_dither8[2];
479 extern const AVClass sws_context_class;
482 * Sets c->swScale to an unscaled converter if one exists for the specific
483 * source and destination formats, bit depths, flags, etc.
485 void ff_get_unscaled_swscale(SwsContext *c);
487 void ff_swscale_get_unscaled_altivec(SwsContext *c);
490 * Returns function pointer to fastest main scaler path function depending
491 * on architecture and available optimizations.
493 SwsFunc ff_getSwsFunc(SwsContext *c);
495 void ff_sws_init_swScale_altivec(SwsContext *c);
496 void ff_sws_init_swScale_mmx(SwsContext *c);
498 #endif /* SWSCALE_SWSCALE_INTERNAL_H */