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
41 #define VOFW 2048 // faster on PPC and not tested on others
47 #define ALT32_CORR (-1)
64 typedef int (*SwsFunc)(struct SwsContext *context, const uint8_t* src[],
65 int srcStride[], int srcSliceY, int srcSliceH,
66 uint8_t* dst[], int dstStride[]);
68 /* This struct should be aligned on at least a 32-byte boundary. */
69 typedef struct SwsContext {
71 * info on struct for av_log
73 const AVClass *av_class;
76 * Note that src, dst, srcStride, dstStride will be copied in the
77 * sws_scale() wrapper so they can be freely modified here.
80 int srcW; ///< Width of source luma/alpha planes.
81 int srcH; ///< Height of source luma/alpha planes.
82 int dstH; ///< Height of destination luma/alpha planes.
83 int chrSrcW; ///< Width of source chroma planes.
84 int chrSrcH; ///< Height of source chroma planes.
85 int chrDstW; ///< Width of destination chroma planes.
86 int chrDstH; ///< Height of destination chroma planes.
89 enum PixelFormat dstFormat; ///< Destination pixel format.
90 enum PixelFormat srcFormat; ///< Source pixel format.
91 int dstFormatBpp; ///< Number of bits per pixel of the destination pixel format.
92 int srcFormatBpp; ///< Number of bits per pixel of the source pixel format.
93 int chrSrcHSubSample; ///< Binary logarithm of horizontal subsampling factor between luma/alpha and chroma planes in source image.
94 int chrSrcVSubSample; ///< Binary logarithm of vertical subsampling factor between luma/alpha and chroma planes in source image.
95 int chrDstHSubSample; ///< Binary logarithm of horizontal subsampling factor between luma/alpha and chroma planes in destination image.
96 int chrDstVSubSample; ///< Binary logarithm of vertical subsampling factor between luma/alpha and chroma planes in destination image.
97 int vChrDrop; ///< Binary logarithm of extra vertical subsampling factor in source image chroma planes specified by user.
98 int sliceDir; ///< Direction that slices are fed to the scaler (1 = top-to-bottom, -1 = bottom-to-top).
99 double param[2]; ///< Input parameters for scaling algorithms that need them.
101 uint32_t pal_yuv[256];
102 uint32_t pal_rgb[256];
105 * @name Scaled horizontal lines ring buffer.
106 * The horizontal scaler keeps just enough scaled lines in a ring buffer
107 * so they may be passed to the vertical scaler. The pointers to the
108 * allocated buffers for each line are duplicated in sequence in the ring
109 * buffer to simplify indexing and avoid wrapping around between lines
110 * inside the vertical scaler code. The wrapping is done before the
111 * vertical scaler is called.
114 int16_t **lumPixBuf; ///< Ring buffer for scaled horizontal luma plane lines to be fed to the vertical scaler.
115 int16_t **chrUPixBuf; ///< Ring buffer for scaled horizontal chroma plane lines to be fed to the vertical scaler.
116 int16_t **chrVPixBuf; ///< Ring buffer for scaled horizontal chroma plane lines to be fed to the vertical scaler.
117 int16_t **alpPixBuf; ///< Ring buffer for scaled horizontal alpha plane lines to be fed to the vertical scaler.
118 int vLumBufSize; ///< Number of vertical luma/alpha lines allocated in the ring buffer.
119 int vChrBufSize; ///< Number of vertical chroma lines allocated in the ring buffer.
120 int lastInLumBuf; ///< Last scaled horizontal luma/alpha line from source in the ring buffer.
121 int lastInChrBuf; ///< Last scaled horizontal chroma line from source in the ring buffer.
122 int lumBufIndex; ///< Index in ring buffer of the last scaled horizontal luma/alpha line from source.
123 int chrBufIndex; ///< Index in ring buffer of the last scaled horizontal chroma line from source.
126 uint8_t *formatConvBuffer;
129 * @name Horizontal and vertical filters.
130 * To better understand the following fields, here is a pseudo-code of
131 * their usage in filtering a horizontal line:
133 * for (i = 0; i < width; i++) {
135 * for (j = 0; j < filterSize; j++)
136 * dst[i] += src[ filterPos[i] + j ] * filter[ filterSize * i + j ];
137 * dst[i] >>= FRAC_BITS; // The actual implementation is fixed-point.
142 int16_t *hLumFilter; ///< Array of horizontal filter coefficients for luma/alpha planes.
143 int16_t *hChrFilter; ///< Array of horizontal filter coefficients for chroma planes.
144 int16_t *vLumFilter; ///< Array of vertical filter coefficients for luma/alpha planes.
145 int16_t *vChrFilter; ///< Array of vertical filter coefficients for chroma planes.
146 int16_t *hLumFilterPos; ///< Array of horizontal filter starting positions for each dst[i] for luma/alpha planes.
147 int16_t *hChrFilterPos; ///< Array of horizontal filter starting positions for each dst[i] for chroma planes.
148 int16_t *vLumFilterPos; ///< Array of vertical filter starting positions for each dst[i] for luma/alpha planes.
149 int16_t *vChrFilterPos; ///< Array of vertical filter starting positions for each dst[i] for chroma planes.
150 int hLumFilterSize; ///< Horizontal filter size for luma/alpha pixels.
151 int hChrFilterSize; ///< Horizontal filter size for chroma pixels.
152 int vLumFilterSize; ///< Vertical filter size for luma/alpha pixels.
153 int vChrFilterSize; ///< Vertical filter size for chroma pixels.
156 int lumMmx2FilterCodeSize; ///< Runtime-generated MMX2 horizontal fast bilinear scaler code size for luma/alpha planes.
157 int chrMmx2FilterCodeSize; ///< Runtime-generated MMX2 horizontal fast bilinear scaler code size for chroma planes.
158 uint8_t *lumMmx2FilterCode; ///< Runtime-generated MMX2 horizontal fast bilinear scaler code for luma/alpha planes.
159 uint8_t *chrMmx2FilterCode; ///< Runtime-generated MMX2 horizontal fast bilinear scaler code for chroma planes.
163 int dstY; ///< Last destination vertical line output from last slice.
164 int flags; ///< Flags passed by the user to select scaler algorithm, optimizations, subsampling, etc...
165 void * yuvTable; // pointer to the yuv->rgb table start so it can be freed()
166 uint8_t * table_rV[256];
167 uint8_t * table_gU[256];
169 uint8_t * table_bU[256];
172 int contrast, brightness, saturation; // for sws_getColorspaceDetails
173 int srcColorspaceTable[4];
174 int dstColorspaceTable[4];
175 int srcRange; ///< 0 = MPG YUV range, 1 = JPG YUV range (source image).
176 int dstRange; ///< 0 = MPG YUV range, 1 = JPG YUV range (destination image).
177 int yuv2rgb_y_offset;
179 int yuv2rgb_v2r_coeff;
180 int yuv2rgb_v2g_coeff;
181 int yuv2rgb_u2g_coeff;
182 int yuv2rgb_u2b_coeff;
184 #define RED_DITHER "0*8"
185 #define GREEN_DITHER "1*8"
186 #define BLUE_DITHER "2*8"
187 #define Y_COEFF "3*8"
188 #define VR_COEFF "4*8"
189 #define UB_COEFF "5*8"
190 #define VG_COEFF "6*8"
191 #define UG_COEFF "7*8"
192 #define Y_OFFSET "8*8"
193 #define U_OFFSET "9*8"
194 #define V_OFFSET "10*8"
195 #define LUM_MMX_FILTER_OFFSET "11*8"
196 #define CHR_MMX_FILTER_OFFSET "11*8+4*4*256"
197 #define DSTW_OFFSET "11*8+4*4*256*2" //do not change, it is hardcoded in the ASM
198 #define ESP_OFFSET "11*8+4*4*256*2+8"
199 #define VROUNDER_OFFSET "11*8+4*4*256*2+16"
200 #define U_TEMP "11*8+4*4*256*2+24"
201 #define V_TEMP "11*8+4*4*256*2+32"
202 #define Y_TEMP "11*8+4*4*256*2+40"
203 #define ALP_MMX_FILTER_OFFSET "11*8+4*4*256*2+48"
204 #define UV_OFF "11*8+4*4*256*3+48"
206 DECLARE_ALIGNED(8, uint64_t, redDither);
207 DECLARE_ALIGNED(8, uint64_t, greenDither);
208 DECLARE_ALIGNED(8, uint64_t, blueDither);
210 DECLARE_ALIGNED(8, uint64_t, yCoeff);
211 DECLARE_ALIGNED(8, uint64_t, vrCoeff);
212 DECLARE_ALIGNED(8, uint64_t, ubCoeff);
213 DECLARE_ALIGNED(8, uint64_t, vgCoeff);
214 DECLARE_ALIGNED(8, uint64_t, ugCoeff);
215 DECLARE_ALIGNED(8, uint64_t, yOffset);
216 DECLARE_ALIGNED(8, uint64_t, uOffset);
217 DECLARE_ALIGNED(8, uint64_t, vOffset);
218 int32_t lumMmxFilter[4*MAX_FILTER_SIZE];
219 int32_t chrMmxFilter[4*MAX_FILTER_SIZE];
220 int dstW; ///< Width of destination luma/alpha planes.
221 DECLARE_ALIGNED(8, uint64_t, esp);
222 DECLARE_ALIGNED(8, uint64_t, vRounder);
223 DECLARE_ALIGNED(8, uint64_t, u_temp);
224 DECLARE_ALIGNED(8, uint64_t, v_temp);
225 DECLARE_ALIGNED(8, uint64_t, y_temp);
226 int32_t alpMmxFilter[4*MAX_FILTER_SIZE];
227 DECLARE_ALIGNED(8, ptrdiff_t, uv_off); ///< offset (in pixels) between u and v planes
230 vector signed short CY;
231 vector signed short CRV;
232 vector signed short CBU;
233 vector signed short CGU;
234 vector signed short CGV;
235 vector signed short OY;
236 vector unsigned short CSHIFT;
237 vector signed short *vYCoeffsBank, *vCCoeffsBank;
241 DECLARE_ALIGNED(4, uint32_t, oy);
242 DECLARE_ALIGNED(4, uint32_t, oc);
243 DECLARE_ALIGNED(4, uint32_t, zero);
244 DECLARE_ALIGNED(4, uint32_t, cy);
245 DECLARE_ALIGNED(4, uint32_t, crv);
246 DECLARE_ALIGNED(4, uint32_t, rmask);
247 DECLARE_ALIGNED(4, uint32_t, cbu);
248 DECLARE_ALIGNED(4, uint32_t, bmask);
249 DECLARE_ALIGNED(4, uint32_t, cgu);
250 DECLARE_ALIGNED(4, uint32_t, cgv);
251 DECLARE_ALIGNED(4, uint32_t, gmask);
255 DECLARE_ALIGNED(8, uint64_t, sparc_coeffs)[10];
258 /* function pointers for swScale() */
259 void (*yuv2nv12X )(struct SwsContext *c,
260 const int16_t *lumFilter, const int16_t **lumSrc, int lumFilterSize,
261 const int16_t *chrFilter, const int16_t **chrUSrc,
262 const int16_t **chrVSrc, int chrFilterSize,
263 uint8_t *dest, uint8_t *uDest,
264 int dstW, int chrDstW, int dstFormat);
265 void (*yuv2yuv1 )(struct SwsContext *c,
266 const int16_t *lumSrc, const int16_t *chrUSrc,
267 const int16_t *chrVSrc, const int16_t *alpSrc,
269 uint8_t *uDest, uint8_t *vDest, uint8_t *aDest,
270 long dstW, long chrDstW);
271 void (*yuv2yuvX )(struct SwsContext *c,
272 const int16_t *lumFilter, const int16_t **lumSrc, int lumFilterSize,
273 const int16_t *chrFilter, const int16_t **chrUSrc,
274 const int16_t **chrVSrc, int chrFilterSize,
275 const int16_t **alpSrc,
277 uint8_t *uDest, uint8_t *vDest, uint8_t *aDest,
278 long dstW, long chrDstW);
279 void (*yuv2packed1)(struct SwsContext *c,
280 const uint16_t *buf0,
281 const uint16_t *ubuf0, const uint16_t *ubuf1,
282 const uint16_t *vbuf0, const uint16_t *vbuf1,
283 const uint16_t *abuf0,
285 int dstW, int uvalpha, int dstFormat, int flags, int y);
286 void (*yuv2packed2)(struct SwsContext *c,
287 const uint16_t *buf0, const uint16_t *buf1,
288 const uint16_t *ubuf0, const uint16_t *ubuf1,
289 const uint16_t *vbuf0, const uint16_t *vbuf1,
290 const uint16_t *abuf0, const uint16_t *abuf1,
292 int dstW, int yalpha, int uvalpha, int y);
293 void (*yuv2packedX)(struct SwsContext *c,
294 const int16_t *lumFilter, const int16_t **lumSrc, int lumFilterSize,
295 const int16_t *chrFilter, const int16_t **chrUSrc,
296 const int16_t **chrVSrc, int chrFilterSize,
297 const int16_t **alpSrc, uint8_t *dest,
298 long dstW, long dstY);
300 void (*lumToYV12)(uint8_t *dst, const uint8_t *src,
301 long width, uint32_t *pal); ///< Unscaled conversion of luma plane to YV12 for horizontal scaler.
302 void (*alpToYV12)(uint8_t *dst, const uint8_t *src,
303 long width, uint32_t *pal); ///< Unscaled conversion of alpha plane to YV12 for horizontal scaler.
304 void (*chrToYV12)(uint8_t *dstU, uint8_t *dstV,
305 const uint8_t *src1, const uint8_t *src2,
306 long width, uint32_t *pal); ///< Unscaled conversion of chroma planes to YV12 for horizontal scaler.
307 void (*hyscale_fast)(struct SwsContext *c,
308 int16_t *dst, long dstWidth,
309 const uint8_t *src, int srcW, int xInc);
310 void (*hcscale_fast)(struct SwsContext *c,
311 int16_t *dst1, int16_t *dst2, long dstWidth,
312 const uint8_t *src1, const uint8_t *src2,
315 void (*hScale)(int16_t *dst, int dstW, const uint8_t *src, int srcW,
316 int xInc, const int16_t *filter, const int16_t *filterPos,
319 void (*lumConvertRange)(uint16_t *dst, int width); ///< Color range conversion function for luma plane if needed.
320 void (*chrConvertRange)(uint16_t *dst1, uint16_t *dst2, int width); ///< Color range conversion function for chroma planes if needed.
322 int lumSrcOffset; ///< Offset given to luma src pointers passed to horizontal input functions.
323 int chrSrcOffset; ///< Offset given to chroma src pointers passed to horizontal input functions.
324 int alpSrcOffset; ///< Offset given to alpha src pointers passed to horizontal input functions.
326 int needs_hcscale; ///< Set if there are chroma planes to be converted.
329 //FIXME check init (where 0)
331 SwsFunc ff_yuv2rgb_get_func_ptr(SwsContext *c);
332 int ff_yuv2rgb_c_init_tables(SwsContext *c, const int inv_table[4],
333 int fullRange, int brightness,
334 int contrast, int saturation);
336 void ff_yuv2rgb_init_tables_altivec(SwsContext *c, const int inv_table[4],
337 int brightness, int contrast, int saturation);
338 SwsFunc ff_yuv2rgb_init_mmx(SwsContext *c);
339 SwsFunc ff_yuv2rgb_init_vis(SwsContext *c);
340 SwsFunc ff_yuv2rgb_init_mlib(SwsContext *c);
341 SwsFunc ff_yuv2rgb_init_altivec(SwsContext *c);
342 SwsFunc ff_yuv2rgb_get_func_ptr_bfin(SwsContext *c);
343 void ff_bfin_get_unscaled_swscale(SwsContext *c);
344 void ff_yuv2packedX_altivec(SwsContext *c,
345 const int16_t *lumFilter, const int16_t **lumSrc, int lumFilterSize,
346 const int16_t *chrFilter, const int16_t **chrSrc, int chrFilterSize,
347 uint8_t *dest, int dstW, int dstY);
349 const char *sws_format_name(enum PixelFormat format);
351 //FIXME replace this with something faster
352 #define is16BPS(x) ( \
353 (x)==PIX_FMT_GRAY16BE \
354 || (x)==PIX_FMT_GRAY16LE \
355 || (x)==PIX_FMT_BGR48BE \
356 || (x)==PIX_FMT_BGR48LE \
357 || (x)==PIX_FMT_RGB48BE \
358 || (x)==PIX_FMT_RGB48LE \
359 || (x)==PIX_FMT_YUV420P16LE \
360 || (x)==PIX_FMT_YUV422P16LE \
361 || (x)==PIX_FMT_YUV444P16LE \
362 || (x)==PIX_FMT_YUV420P16BE \
363 || (x)==PIX_FMT_YUV422P16BE \
364 || (x)==PIX_FMT_YUV444P16BE \
366 #define is9_OR_10BPS(x) ( \
367 (x)==PIX_FMT_YUV420P9LE \
368 || (x)==PIX_FMT_YUV420P9BE \
369 || (x)==PIX_FMT_YUV420P10LE \
370 || (x)==PIX_FMT_YUV420P10BE \
372 #define isBE(x) ((x)&1)
373 #define isPlanar8YUV(x) ( \
374 (x)==PIX_FMT_YUV410P \
375 || (x)==PIX_FMT_YUV420P \
376 || (x)==PIX_FMT_YUVA420P \
377 || (x)==PIX_FMT_YUV411P \
378 || (x)==PIX_FMT_YUV422P \
379 || (x)==PIX_FMT_YUV444P \
380 || (x)==PIX_FMT_YUV440P \
381 || (x)==PIX_FMT_NV12 \
382 || (x)==PIX_FMT_NV21 \
384 #define isPlanarYUV(x) ( \
386 || (x)==PIX_FMT_YUV420P9LE \
387 || (x)==PIX_FMT_YUV420P10LE \
388 || (x)==PIX_FMT_YUV420P16LE \
389 || (x)==PIX_FMT_YUV422P16LE \
390 || (x)==PIX_FMT_YUV444P16LE \
391 || (x)==PIX_FMT_YUV420P9BE \
392 || (x)==PIX_FMT_YUV420P10BE \
393 || (x)==PIX_FMT_YUV420P16BE \
394 || (x)==PIX_FMT_YUV422P16BE \
395 || (x)==PIX_FMT_YUV444P16BE \
398 (x)==PIX_FMT_UYVY422 \
399 || (x)==PIX_FMT_YUYV422 \
402 #define isGray(x) ( \
404 || (x)==PIX_FMT_Y400A \
405 || (x)==PIX_FMT_GRAY16BE \
406 || (x)==PIX_FMT_GRAY16LE \
408 #define isGray16(x) ( \
409 (x)==PIX_FMT_GRAY16BE \
410 || (x)==PIX_FMT_GRAY16LE \
412 #define isRGBinInt(x) ( \
413 (x)==PIX_FMT_RGB48BE \
414 || (x)==PIX_FMT_RGB48LE \
415 || (x)==PIX_FMT_RGB32 \
416 || (x)==PIX_FMT_RGB32_1 \
417 || (x)==PIX_FMT_RGB24 \
418 || (x)==PIX_FMT_RGB565BE \
419 || (x)==PIX_FMT_RGB565LE \
420 || (x)==PIX_FMT_RGB555BE \
421 || (x)==PIX_FMT_RGB555LE \
422 || (x)==PIX_FMT_RGB444BE \
423 || (x)==PIX_FMT_RGB444LE \
424 || (x)==PIX_FMT_RGB8 \
425 || (x)==PIX_FMT_RGB4 \
426 || (x)==PIX_FMT_RGB4_BYTE \
427 || (x)==PIX_FMT_MONOBLACK \
428 || (x)==PIX_FMT_MONOWHITE \
430 #define isBGRinInt(x) ( \
431 (x)==PIX_FMT_BGR48BE \
432 || (x)==PIX_FMT_BGR48LE \
433 || (x)==PIX_FMT_BGR32 \
434 || (x)==PIX_FMT_BGR32_1 \
435 || (x)==PIX_FMT_BGR24 \
436 || (x)==PIX_FMT_BGR565BE \
437 || (x)==PIX_FMT_BGR565LE \
438 || (x)==PIX_FMT_BGR555BE \
439 || (x)==PIX_FMT_BGR555LE \
440 || (x)==PIX_FMT_BGR444BE \
441 || (x)==PIX_FMT_BGR444LE \
442 || (x)==PIX_FMT_BGR8 \
443 || (x)==PIX_FMT_BGR4 \
444 || (x)==PIX_FMT_BGR4_BYTE \
445 || (x)==PIX_FMT_MONOBLACK \
446 || (x)==PIX_FMT_MONOWHITE \
448 #define isRGBinBytes(x) ( \
449 (x)==PIX_FMT_RGB48BE \
450 || (x)==PIX_FMT_RGB48LE \
451 || (x)==PIX_FMT_RGBA \
452 || (x)==PIX_FMT_ARGB \
453 || (x)==PIX_FMT_RGB24 \
455 #define isBGRinBytes(x) ( \
456 (x)==PIX_FMT_BGR48BE \
457 || (x)==PIX_FMT_BGR48LE \
458 || (x)==PIX_FMT_BGRA \
459 || (x)==PIX_FMT_ABGR \
460 || (x)==PIX_FMT_BGR24 \
462 #define isAnyRGB(x) ( \
466 #define isALPHA(x) ( \
468 || (x)==PIX_FMT_BGR32_1 \
469 || (x)==PIX_FMT_RGB32 \
470 || (x)==PIX_FMT_RGB32_1 \
471 || (x)==PIX_FMT_Y400A \
472 || (x)==PIX_FMT_YUVA420P \
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);
488 * Returns function pointer to fastest main scaler path function depending
489 * on architecture and available optimizations.
491 SwsFunc ff_getSwsFunc(SwsContext *c);
493 #endif /* SWSCALE_SWSCALE_INTERNAL_H */