2 * Copyright (C) 2001-2011 Michael Niedermayer <michaelni@gmx.at>
4 * This file is part of FFmpeg.
6 * FFmpeg is free software; you can redistribute it and/or
7 * modify it under the terms of the GNU Lesser General Public
8 * License as published by the Free Software Foundation; either
9 * version 2.1 of the License, or (at your option) any later version.
11 * FFmpeg is distributed in the hope that it will be useful,
12 * but WITHOUT ANY WARRANTY; without even the implied warranty of
13 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
14 * Lesser General Public License for more details.
16 * You should have received a copy of the GNU Lesser General Public
17 * License along with FFmpeg; if not, write to the Free Software
18 * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
21 #ifndef SWSCALE_SWSCALE_INTERNAL_H
22 #define SWSCALE_SWSCALE_INTERNAL_H
30 #include "libavutil/avassert.h"
31 #include "libavutil/avutil.h"
32 #include "libavutil/common.h"
33 #include "libavutil/intreadwrite.h"
34 #include "libavutil/log.h"
35 #include "libavutil/pixfmt.h"
36 #include "libavutil/pixdesc.h"
38 #define STR(s) AV_TOSTRING(s) // AV_STRINGIFY is too long
40 #define YUVRGB_TABLE_HEADROOM 128
42 #define FAST_BGR2YV12 // use 7-bit instead of 15-bit coefficients
44 #define MAX_FILTER_SIZE 256
49 #define ALT32_CORR (-1)
66 typedef int (*SwsFunc)(struct SwsContext *context, const uint8_t *src[],
67 int srcStride[], int srcSliceY, int srcSliceH,
68 uint8_t *dst[], int dstStride[]);
71 * Write one line of horizontally scaled data to planar output
72 * without any additional vertical scaling (or point-scaling).
74 * @param src scaled source data, 15bit for 8-10bit output,
75 * 19-bit for 16bit output (in int32_t)
76 * @param dest pointer to the output plane. For >8bit
77 * output, this is in uint16_t
78 * @param dstW width of destination in pixels
79 * @param dither ordered dither array of type int16_t and size 8
80 * @param offset Dither offset
82 typedef void (*yuv2planar1_fn)(const int16_t *src, uint8_t *dest, int dstW,
83 const uint8_t *dither, int offset);
86 * Write one line of horizontally scaled data to planar output
87 * with multi-point vertical scaling between input pixels.
89 * @param filter vertical luma/alpha scaling coefficients, 12bit [0,4096]
90 * @param src scaled luma (Y) or alpha (A) source data, 15bit for 8-10bit output,
91 * 19-bit for 16bit output (in int32_t)
92 * @param filterSize number of vertical input lines to scale
93 * @param dest pointer to output plane. For >8bit
94 * output, this is in uint16_t
95 * @param dstW width of destination pixels
96 * @param offset Dither offset
98 typedef void (*yuv2planarX_fn)(const int16_t *filter, int filterSize,
99 const int16_t **src, uint8_t *dest, int dstW,
100 const uint8_t *dither, int offset);
103 * Write one line of horizontally scaled chroma to interleaved output
104 * with multi-point vertical scaling between input pixels.
106 * @param c SWS scaling context
107 * @param chrFilter vertical chroma scaling coefficients, 12bit [0,4096]
108 * @param chrUSrc scaled chroma (U) source data, 15bit for 8-10bit output,
109 * 19-bit for 16bit output (in int32_t)
110 * @param chrVSrc scaled chroma (V) source data, 15bit for 8-10bit output,
111 * 19-bit for 16bit output (in int32_t)
112 * @param chrFilterSize number of vertical chroma input lines to scale
113 * @param dest pointer to the output plane. For >8bit
114 * output, this is in uint16_t
115 * @param dstW width of chroma planes
117 typedef void (*yuv2interleavedX_fn)(struct SwsContext *c,
118 const int16_t *chrFilter,
120 const int16_t **chrUSrc,
121 const int16_t **chrVSrc,
122 uint8_t *dest, int dstW);
125 * Write one line of horizontally scaled Y/U/V/A to packed-pixel YUV/RGB
126 * output without any additional vertical scaling (or point-scaling). Note
127 * that this function may do chroma scaling, see the "uvalpha" argument.
129 * @param c SWS scaling context
130 * @param lumSrc scaled luma (Y) source data, 15bit for 8-10bit output,
131 * 19-bit for 16bit output (in int32_t)
132 * @param chrUSrc scaled chroma (U) source data, 15bit for 8-10bit output,
133 * 19-bit for 16bit output (in int32_t)
134 * @param chrVSrc scaled chroma (V) source data, 15bit for 8-10bit output,
135 * 19-bit for 16bit output (in int32_t)
136 * @param alpSrc scaled alpha (A) source data, 15bit for 8-10bit output,
137 * 19-bit for 16bit output (in int32_t)
138 * @param dest pointer to the output plane. For 16bit output, this is
140 * @param dstW width of lumSrc and alpSrc in pixels, number of pixels
141 * to write into dest[]
142 * @param uvalpha chroma scaling coefficient for the second line of chroma
143 * pixels, either 2048 or 0. If 0, one chroma input is used
144 * for 2 output pixels (or if the SWS_FLAG_FULL_CHR_INT flag
145 * is set, it generates 1 output pixel). If 2048, two chroma
146 * input pixels should be averaged for 2 output pixels (this
147 * only happens if SWS_FLAG_FULL_CHR_INT is not set)
148 * @param y vertical line number for this output. This does not need
149 * to be used to calculate the offset in the destination,
150 * but can be used to generate comfort noise using dithering
151 * for some output formats.
153 typedef void (*yuv2packed1_fn)(struct SwsContext *c, const int16_t *lumSrc,
154 const int16_t *chrUSrc[2],
155 const int16_t *chrVSrc[2],
156 const int16_t *alpSrc, uint8_t *dest,
157 int dstW, int uvalpha, int y);
159 * Write one line of horizontally scaled Y/U/V/A to packed-pixel YUV/RGB
160 * output by doing bilinear scaling between two input lines.
162 * @param c SWS scaling context
163 * @param lumSrc scaled luma (Y) source data, 15bit for 8-10bit output,
164 * 19-bit for 16bit output (in int32_t)
165 * @param chrUSrc scaled chroma (U) source data, 15bit for 8-10bit output,
166 * 19-bit for 16bit output (in int32_t)
167 * @param chrVSrc scaled chroma (V) source data, 15bit for 8-10bit output,
168 * 19-bit for 16bit output (in int32_t)
169 * @param alpSrc scaled alpha (A) source data, 15bit for 8-10bit output,
170 * 19-bit for 16bit output (in int32_t)
171 * @param dest pointer to the output plane. For 16bit output, this is
173 * @param dstW width of lumSrc and alpSrc in pixels, number of pixels
174 * to write into dest[]
175 * @param yalpha luma/alpha scaling coefficients for the second input line.
176 * The first line's coefficients can be calculated by using
178 * @param uvalpha chroma scaling coefficient for the second input line. The
179 * first line's coefficients can be calculated by using
181 * @param y vertical line number for this output. This does not need
182 * to be used to calculate the offset in the destination,
183 * but can be used to generate comfort noise using dithering
184 * for some output formats.
186 typedef void (*yuv2packed2_fn)(struct SwsContext *c, const int16_t *lumSrc[2],
187 const int16_t *chrUSrc[2],
188 const int16_t *chrVSrc[2],
189 const int16_t *alpSrc[2],
191 int dstW, int yalpha, int uvalpha, int y);
193 * Write one line of horizontally scaled Y/U/V/A to packed-pixel YUV/RGB
194 * output by doing multi-point vertical scaling between input pixels.
196 * @param c SWS scaling context
197 * @param lumFilter vertical luma/alpha scaling coefficients, 12bit [0,4096]
198 * @param lumSrc scaled luma (Y) source data, 15bit for 8-10bit output,
199 * 19-bit for 16bit output (in int32_t)
200 * @param lumFilterSize number of vertical luma/alpha input lines to scale
201 * @param chrFilter vertical chroma scaling coefficients, 12bit [0,4096]
202 * @param chrUSrc scaled chroma (U) source data, 15bit for 8-10bit output,
203 * 19-bit for 16bit output (in int32_t)
204 * @param chrVSrc scaled chroma (V) source data, 15bit for 8-10bit output,
205 * 19-bit for 16bit output (in int32_t)
206 * @param chrFilterSize number of vertical chroma input lines to scale
207 * @param alpSrc scaled alpha (A) source data, 15bit for 8-10bit output,
208 * 19-bit for 16bit output (in int32_t)
209 * @param dest pointer to the output plane. For 16bit output, this is
211 * @param dstW width of lumSrc and alpSrc in pixels, number of pixels
212 * to write into dest[]
213 * @param y vertical line number for this output. This does not need
214 * to be used to calculate the offset in the destination,
215 * but can be used to generate comfort noise using dithering
216 * or some output formats.
218 typedef void (*yuv2packedX_fn)(struct SwsContext *c, const int16_t *lumFilter,
219 const int16_t **lumSrc, int lumFilterSize,
220 const int16_t *chrFilter,
221 const int16_t **chrUSrc,
222 const int16_t **chrVSrc, int chrFilterSize,
223 const int16_t **alpSrc, uint8_t *dest,
226 /* This struct should be aligned on at least a 32-byte boundary. */
227 typedef struct SwsContext {
229 * info on struct for av_log
231 const AVClass *av_class;
234 * Note that src, dst, srcStride, dstStride will be copied in the
235 * sws_scale() wrapper so they can be freely modified here.
238 int srcW; ///< Width of source luma/alpha planes.
239 int srcH; ///< Height of source luma/alpha planes.
240 int dstH; ///< Height of destination luma/alpha planes.
241 int chrSrcW; ///< Width of source chroma planes.
242 int chrSrcH; ///< Height of source chroma planes.
243 int chrDstW; ///< Width of destination chroma planes.
244 int chrDstH; ///< Height of destination chroma planes.
245 int lumXInc, chrXInc;
246 int lumYInc, chrYInc;
247 enum AVPixelFormat dstFormat; ///< Destination pixel format.
248 enum AVPixelFormat srcFormat; ///< Source pixel format.
249 int dstFormatBpp; ///< Number of bits per pixel of the destination pixel format.
250 int srcFormatBpp; ///< Number of bits per pixel of the source pixel format.
252 int chrSrcHSubSample; ///< Binary logarithm of horizontal subsampling factor between luma/alpha and chroma planes in source image.
253 int chrSrcVSubSample; ///< Binary logarithm of vertical subsampling factor between luma/alpha and chroma planes in source image.
254 int chrDstHSubSample; ///< Binary logarithm of horizontal subsampling factor between luma/alpha and chroma planes in destination image.
255 int chrDstVSubSample; ///< Binary logarithm of vertical subsampling factor between luma/alpha and chroma planes in destination image.
256 int vChrDrop; ///< Binary logarithm of extra vertical subsampling factor in source image chroma planes specified by user.
257 int sliceDir; ///< Direction that slices are fed to the scaler (1 = top-to-bottom, -1 = bottom-to-top).
258 double param[2]; ///< Input parameters for scaling algorithms that need them.
260 uint32_t pal_yuv[256];
261 uint32_t pal_rgb[256];
264 * @name Scaled horizontal lines ring buffer.
265 * The horizontal scaler keeps just enough scaled lines in a ring buffer
266 * so they may be passed to the vertical scaler. The pointers to the
267 * allocated buffers for each line are duplicated in sequence in the ring
268 * buffer to simplify indexing and avoid wrapping around between lines
269 * inside the vertical scaler code. The wrapping is done before the
270 * vertical scaler is called.
273 int16_t **lumPixBuf; ///< Ring buffer for scaled horizontal luma plane lines to be fed to the vertical scaler.
274 int16_t **chrUPixBuf; ///< Ring buffer for scaled horizontal chroma plane lines to be fed to the vertical scaler.
275 int16_t **chrVPixBuf; ///< Ring buffer for scaled horizontal chroma plane lines to be fed to the vertical scaler.
276 int16_t **alpPixBuf; ///< Ring buffer for scaled horizontal alpha plane lines to be fed to the vertical scaler.
277 int vLumBufSize; ///< Number of vertical luma/alpha lines allocated in the ring buffer.
278 int vChrBufSize; ///< Number of vertical chroma lines allocated in the ring buffer.
279 int lastInLumBuf; ///< Last scaled horizontal luma/alpha line from source in the ring buffer.
280 int lastInChrBuf; ///< Last scaled horizontal chroma line from source in the ring buffer.
281 int lumBufIndex; ///< Index in ring buffer of the last scaled horizontal luma/alpha line from source.
282 int chrBufIndex; ///< Index in ring buffer of the last scaled horizontal chroma line from source.
285 uint8_t *formatConvBuffer;
288 * @name Horizontal and vertical filters.
289 * To better understand the following fields, here is a pseudo-code of
290 * their usage in filtering a horizontal line:
292 * for (i = 0; i < width; i++) {
294 * for (j = 0; j < filterSize; j++)
295 * dst[i] += src[ filterPos[i] + j ] * filter[ filterSize * i + j ];
296 * dst[i] >>= FRAC_BITS; // The actual implementation is fixed-point.
301 int16_t *hLumFilter; ///< Array of horizontal filter coefficients for luma/alpha planes.
302 int16_t *hChrFilter; ///< Array of horizontal filter coefficients for chroma planes.
303 int16_t *vLumFilter; ///< Array of vertical filter coefficients for luma/alpha planes.
304 int16_t *vChrFilter; ///< Array of vertical filter coefficients for chroma planes.
305 int32_t *hLumFilterPos; ///< Array of horizontal filter starting positions for each dst[i] for luma/alpha planes.
306 int32_t *hChrFilterPos; ///< Array of horizontal filter starting positions for each dst[i] for chroma planes.
307 int32_t *vLumFilterPos; ///< Array of vertical filter starting positions for each dst[i] for luma/alpha planes.
308 int32_t *vChrFilterPos; ///< Array of vertical filter starting positions for each dst[i] for chroma planes.
309 int hLumFilterSize; ///< Horizontal filter size for luma/alpha pixels.
310 int hChrFilterSize; ///< Horizontal filter size for chroma pixels.
311 int vLumFilterSize; ///< Vertical filter size for luma/alpha pixels.
312 int vChrFilterSize; ///< Vertical filter size for chroma pixels.
315 int lumMmxextFilterCodeSize; ///< Runtime-generated MMXEXT horizontal fast bilinear scaler code size for luma/alpha planes.
316 int chrMmxextFilterCodeSize; ///< Runtime-generated MMXEXT horizontal fast bilinear scaler code size for chroma planes.
317 uint8_t *lumMmxextFilterCode; ///< Runtime-generated MMXEXT horizontal fast bilinear scaler code for luma/alpha planes.
318 uint8_t *chrMmxextFilterCode; ///< Runtime-generated MMXEXT horizontal fast bilinear scaler code for chroma planes.
322 int dstY; ///< Last destination vertical line output from last slice.
323 int flags; ///< Flags passed by the user to select scaler algorithm, optimizations, subsampling, etc...
324 void *yuvTable; // pointer to the yuv->rgb table start so it can be freed()
325 uint8_t *table_rV[256 + 2*YUVRGB_TABLE_HEADROOM];
326 uint8_t *table_gU[256 + 2*YUVRGB_TABLE_HEADROOM];
327 int table_gV[256 + 2*YUVRGB_TABLE_HEADROOM];
328 uint8_t *table_bU[256 + 2*YUVRGB_TABLE_HEADROOM];
330 int *dither_error[4];
333 int contrast, brightness, saturation; // for sws_getColorspaceDetails
334 int srcColorspaceTable[4];
335 int dstColorspaceTable[4];
336 int srcRange; ///< 0 = MPG YUV range, 1 = JPG YUV range (source image).
337 int dstRange; ///< 0 = MPG YUV range, 1 = JPG YUV range (destination image).
340 int yuv2rgb_y_offset;
342 int yuv2rgb_v2r_coeff;
343 int yuv2rgb_v2g_coeff;
344 int yuv2rgb_u2g_coeff;
345 int yuv2rgb_u2b_coeff;
347 #define RED_DITHER "0*8"
348 #define GREEN_DITHER "1*8"
349 #define BLUE_DITHER "2*8"
350 #define Y_COEFF "3*8"
351 #define VR_COEFF "4*8"
352 #define UB_COEFF "5*8"
353 #define VG_COEFF "6*8"
354 #define UG_COEFF "7*8"
355 #define Y_OFFSET "8*8"
356 #define U_OFFSET "9*8"
357 #define V_OFFSET "10*8"
358 #define LUM_MMX_FILTER_OFFSET "11*8"
359 #define CHR_MMX_FILTER_OFFSET "11*8+4*4*256"
360 #define DSTW_OFFSET "11*8+4*4*256*2" //do not change, it is hardcoded in the ASM
361 #define ESP_OFFSET "11*8+4*4*256*2+8"
362 #define VROUNDER_OFFSET "11*8+4*4*256*2+16"
363 #define U_TEMP "11*8+4*4*256*2+24"
364 #define V_TEMP "11*8+4*4*256*2+32"
365 #define Y_TEMP "11*8+4*4*256*2+40"
366 #define ALP_MMX_FILTER_OFFSET "11*8+4*4*256*2+48"
367 #define UV_OFF_PX "11*8+4*4*256*3+48"
368 #define UV_OFF_BYTE "11*8+4*4*256*3+56"
369 #define DITHER16 "11*8+4*4*256*3+64"
370 #define DITHER32 "11*8+4*4*256*3+80"
372 DECLARE_ALIGNED(8, uint64_t, redDither);
373 DECLARE_ALIGNED(8, uint64_t, greenDither);
374 DECLARE_ALIGNED(8, uint64_t, blueDither);
376 DECLARE_ALIGNED(8, uint64_t, yCoeff);
377 DECLARE_ALIGNED(8, uint64_t, vrCoeff);
378 DECLARE_ALIGNED(8, uint64_t, ubCoeff);
379 DECLARE_ALIGNED(8, uint64_t, vgCoeff);
380 DECLARE_ALIGNED(8, uint64_t, ugCoeff);
381 DECLARE_ALIGNED(8, uint64_t, yOffset);
382 DECLARE_ALIGNED(8, uint64_t, uOffset);
383 DECLARE_ALIGNED(8, uint64_t, vOffset);
384 int32_t lumMmxFilter[4 * MAX_FILTER_SIZE];
385 int32_t chrMmxFilter[4 * MAX_FILTER_SIZE];
386 int dstW; ///< Width of destination luma/alpha planes.
387 DECLARE_ALIGNED(8, uint64_t, esp);
388 DECLARE_ALIGNED(8, uint64_t, vRounder);
389 DECLARE_ALIGNED(8, uint64_t, u_temp);
390 DECLARE_ALIGNED(8, uint64_t, v_temp);
391 DECLARE_ALIGNED(8, uint64_t, y_temp);
392 int32_t alpMmxFilter[4 * MAX_FILTER_SIZE];
393 // alignment of these values is not necessary, but merely here
394 // to maintain the same offset across x8632 and x86-64. Once we
395 // use proper offset macros in the asm, they can be removed.
396 DECLARE_ALIGNED(8, ptrdiff_t, uv_off); ///< offset (in pixels) between u and v planes
397 DECLARE_ALIGNED(8, ptrdiff_t, uv_offx2); ///< offset (in bytes) between u and v planes
398 DECLARE_ALIGNED(8, uint16_t, dither16)[8];
399 DECLARE_ALIGNED(8, uint32_t, dither32)[8];
401 const uint8_t *chrDither8, *lumDither8;
404 vector signed short CY;
405 vector signed short CRV;
406 vector signed short CBU;
407 vector signed short CGU;
408 vector signed short CGV;
409 vector signed short OY;
410 vector unsigned short CSHIFT;
411 vector signed short *vYCoeffsBank, *vCCoeffsBank;
415 DECLARE_ALIGNED(4, uint32_t, oy);
416 DECLARE_ALIGNED(4, uint32_t, oc);
417 DECLARE_ALIGNED(4, uint32_t, zero);
418 DECLARE_ALIGNED(4, uint32_t, cy);
419 DECLARE_ALIGNED(4, uint32_t, crv);
420 DECLARE_ALIGNED(4, uint32_t, rmask);
421 DECLARE_ALIGNED(4, uint32_t, cbu);
422 DECLARE_ALIGNED(4, uint32_t, bmask);
423 DECLARE_ALIGNED(4, uint32_t, cgu);
424 DECLARE_ALIGNED(4, uint32_t, cgv);
425 DECLARE_ALIGNED(4, uint32_t, gmask);
429 DECLARE_ALIGNED(8, uint64_t, sparc_coeffs)[10];
433 /* function pointers for swScale() */
434 yuv2planar1_fn yuv2plane1;
435 yuv2planarX_fn yuv2planeX;
436 yuv2interleavedX_fn yuv2nv12cX;
437 yuv2packed1_fn yuv2packed1;
438 yuv2packed2_fn yuv2packed2;
439 yuv2packedX_fn yuv2packedX;
441 /// Unscaled conversion of luma plane to YV12 for horizontal scaler.
442 void (*lumToYV12)(uint8_t *dst, const uint8_t *src, const uint8_t *src2, const uint8_t *src3,
443 int width, uint32_t *pal);
444 /// Unscaled conversion of alpha plane to YV12 for horizontal scaler.
445 void (*alpToYV12)(uint8_t *dst, const uint8_t *src, const uint8_t *src2, const uint8_t *src3,
446 int width, uint32_t *pal);
447 /// Unscaled conversion of chroma planes to YV12 for horizontal scaler.
448 void (*chrToYV12)(uint8_t *dstU, uint8_t *dstV,
449 const uint8_t *src1, const uint8_t *src2, const uint8_t *src3,
450 int width, uint32_t *pal);
453 * Functions to read planar input, such as planar RGB, and convert
454 * internally to Y/UV.
457 void (*readLumPlanar)(uint8_t *dst, const uint8_t *src[4], int width);
458 void (*readChrPlanar)(uint8_t *dstU, uint8_t *dstV, const uint8_t *src[4],
463 * Scale one horizontal line of input data using a bilinear filter
464 * to produce one line of output data. Compared to SwsContext->hScale(),
465 * please take note of the following caveats when using these:
466 * - Scaling is done using only 7bit instead of 14bit coefficients.
467 * - You can use no more than 5 input pixels to produce 4 output
468 * pixels. Therefore, this filter should not be used for downscaling
469 * by more than ~20% in width (because that equals more than 5/4th
470 * downscaling and thus more than 5 pixels input per 4 pixels output).
471 * - In general, bilinear filters create artifacts during downscaling
472 * (even when <20%), because one output pixel will span more than one
473 * input pixel, and thus some pixels will need edges of both neighbor
474 * pixels to interpolate the output pixel. Since you can use at most
475 * two input pixels per output pixel in bilinear scaling, this is
476 * impossible and thus downscaling by any size will create artifacts.
477 * To enable this type of scaling, set SWS_FLAG_FAST_BILINEAR
478 * in SwsContext->flags.
481 void (*hyscale_fast)(struct SwsContext *c,
482 int16_t *dst, int dstWidth,
483 const uint8_t *src, int srcW, int xInc);
484 void (*hcscale_fast)(struct SwsContext *c,
485 int16_t *dst1, int16_t *dst2, int dstWidth,
486 const uint8_t *src1, const uint8_t *src2,
491 * Scale one horizontal line of input data using a filter over the input
492 * lines, to produce one (differently sized) line of output data.
494 * @param dst pointer to destination buffer for horizontally scaled
495 * data. If the number of bits per component of one
496 * destination pixel (SwsContext->dstBpc) is <= 10, data
497 * will be 15bpc in 16bits (int16_t) width. Else (i.e.
498 * SwsContext->dstBpc == 16), data will be 19bpc in
499 * 32bits (int32_t) width.
500 * @param dstW width of destination image
501 * @param src pointer to source data to be scaled. If the number of
502 * bits per component of a source pixel (SwsContext->srcBpc)
503 * is 8, this is 8bpc in 8bits (uint8_t) width. Else
504 * (i.e. SwsContext->dstBpc > 8), this is native depth
505 * in 16bits (uint16_t) width. In other words, for 9-bit
506 * YUV input, this is 9bpc, for 10-bit YUV input, this is
507 * 10bpc, and for 16-bit RGB or YUV, this is 16bpc.
508 * @param filter filter coefficients to be used per output pixel for
509 * scaling. This contains 14bpp filtering coefficients.
510 * Guaranteed to contain dstW * filterSize entries.
511 * @param filterPos position of the first input pixel to be used for
512 * each output pixel during scaling. Guaranteed to
513 * contain dstW entries.
514 * @param filterSize the number of input coefficients to be used (and
515 * thus the number of input pixels to be used) for
516 * creating a single output pixel. Is aligned to 4
517 * (and input coefficients thus padded with zeroes)
518 * to simplify creating SIMD code.
521 void (*hyScale)(struct SwsContext *c, int16_t *dst, int dstW,
522 const uint8_t *src, const int16_t *filter,
523 const int32_t *filterPos, int filterSize);
524 void (*hcScale)(struct SwsContext *c, int16_t *dst, int dstW,
525 const uint8_t *src, const int16_t *filter,
526 const int32_t *filterPos, int filterSize);
529 /// Color range conversion function for luma plane if needed.
530 void (*lumConvertRange)(int16_t *dst, int width);
531 /// Color range conversion function for chroma planes if needed.
532 void (*chrConvertRange)(int16_t *dst1, int16_t *dst2, int width);
534 int needs_hcscale; ///< Set if there are chroma planes to be converted.
536 //FIXME check init (where 0)
538 SwsFunc ff_yuv2rgb_get_func_ptr(SwsContext *c);
539 int ff_yuv2rgb_c_init_tables(SwsContext *c, const int inv_table[4],
540 int fullRange, int brightness,
541 int contrast, int saturation);
543 void ff_yuv2rgb_init_tables_altivec(SwsContext *c, const int inv_table[4],
544 int brightness, int contrast, int saturation);
545 void updateMMXDitherTables(SwsContext *c, int dstY, int lumBufIndex, int chrBufIndex,
546 int lastInLumBuf, int lastInChrBuf);
548 SwsFunc ff_yuv2rgb_init_mmx(SwsContext *c);
549 SwsFunc ff_yuv2rgb_init_vis(SwsContext *c);
550 SwsFunc ff_yuv2rgb_init_altivec(SwsContext *c);
551 SwsFunc ff_yuv2rgb_get_func_ptr_bfin(SwsContext *c);
552 void ff_bfin_get_unscaled_swscale(SwsContext *c);
554 #if FF_API_SWS_FORMAT_NAME
556 * @deprecated Use av_get_pix_fmt_name() instead.
559 const char *sws_format_name(enum AVPixelFormat format);
562 static av_always_inline int is16BPS(enum AVPixelFormat pix_fmt)
564 const AVPixFmtDescriptor *desc = av_pix_fmt_desc_get(pix_fmt);
566 return desc->comp[0].depth_minus1 == 15;
569 static av_always_inline int is9_OR_10BPS(enum AVPixelFormat pix_fmt)
571 const AVPixFmtDescriptor *desc = av_pix_fmt_desc_get(pix_fmt);
573 return desc->comp[0].depth_minus1 >= 8 && desc->comp[0].depth_minus1 <= 13;
576 #define isNBPS(x) is9_OR_10BPS(x)
578 static av_always_inline int isBE(enum AVPixelFormat pix_fmt)
580 const AVPixFmtDescriptor *desc = av_pix_fmt_desc_get(pix_fmt);
582 return desc->flags & PIX_FMT_BE;
585 static av_always_inline int isYUV(enum AVPixelFormat pix_fmt)
587 const AVPixFmtDescriptor *desc = av_pix_fmt_desc_get(pix_fmt);
589 return !(desc->flags & PIX_FMT_RGB) && desc->nb_components >= 2;
592 static av_always_inline int isPlanarYUV(enum AVPixelFormat pix_fmt)
594 const AVPixFmtDescriptor *desc = av_pix_fmt_desc_get(pix_fmt);
596 return ((desc->flags & PIX_FMT_PLANAR) && isYUV(pix_fmt));
599 static av_always_inline int isRGB(enum AVPixelFormat pix_fmt)
601 const AVPixFmtDescriptor *desc = av_pix_fmt_desc_get(pix_fmt);
603 return (desc->flags & PIX_FMT_RGB);
608 (!(av_pix_fmt_desc_get(x)->flags & PIX_FMT_PAL) && \
609 av_pix_fmt_desc_get(x)->nb_components <= 2)
612 ((x) == AV_PIX_FMT_GRAY8 || \
613 (x) == AV_PIX_FMT_Y400A || \
614 (x) == AV_PIX_FMT_GRAY16BE || \
615 (x) == AV_PIX_FMT_GRAY16LE)
618 #define isRGBinInt(x) \
620 (x) == AV_PIX_FMT_RGB48BE || \
621 (x) == AV_PIX_FMT_RGB48LE || \
622 (x) == AV_PIX_FMT_RGBA64BE || \
623 (x) == AV_PIX_FMT_RGBA64LE || \
624 (x) == AV_PIX_FMT_RGB32 || \
625 (x) == AV_PIX_FMT_RGB32_1 || \
626 (x) == AV_PIX_FMT_RGB24 || \
627 (x) == AV_PIX_FMT_RGB565BE || \
628 (x) == AV_PIX_FMT_RGB565LE || \
629 (x) == AV_PIX_FMT_RGB555BE || \
630 (x) == AV_PIX_FMT_RGB555LE || \
631 (x) == AV_PIX_FMT_RGB444BE || \
632 (x) == AV_PIX_FMT_RGB444LE || \
633 (x) == AV_PIX_FMT_RGB8 || \
634 (x) == AV_PIX_FMT_RGB4 || \
635 (x) == AV_PIX_FMT_RGB4_BYTE || \
636 (x) == AV_PIX_FMT_MONOBLACK || \
637 (x) == AV_PIX_FMT_MONOWHITE \
639 #define isBGRinInt(x) \
641 (x) == AV_PIX_FMT_BGR48BE || \
642 (x) == AV_PIX_FMT_BGR48LE || \
643 (x) == AV_PIX_FMT_BGRA64BE || \
644 (x) == AV_PIX_FMT_BGRA64LE || \
645 (x) == AV_PIX_FMT_BGR32 || \
646 (x) == AV_PIX_FMT_BGR32_1 || \
647 (x) == AV_PIX_FMT_BGR24 || \
648 (x) == AV_PIX_FMT_BGR565BE || \
649 (x) == AV_PIX_FMT_BGR565LE || \
650 (x) == AV_PIX_FMT_BGR555BE || \
651 (x) == AV_PIX_FMT_BGR555LE || \
652 (x) == AV_PIX_FMT_BGR444BE || \
653 (x) == AV_PIX_FMT_BGR444LE || \
654 (x) == AV_PIX_FMT_BGR8 || \
655 (x) == AV_PIX_FMT_BGR4 || \
656 (x) == AV_PIX_FMT_BGR4_BYTE || \
657 (x) == AV_PIX_FMT_MONOBLACK || \
658 (x) == AV_PIX_FMT_MONOWHITE \
661 #define isRGBinBytes(x) ( \
662 (x) == AV_PIX_FMT_RGB48BE \
663 || (x) == AV_PIX_FMT_RGB48LE \
664 || (x) == AV_PIX_FMT_RGBA64BE \
665 || (x) == AV_PIX_FMT_RGBA64LE \
666 || (x) == AV_PIX_FMT_RGBA \
667 || (x) == AV_PIX_FMT_ARGB \
668 || (x) == AV_PIX_FMT_RGB24 \
670 #define isBGRinBytes(x) ( \
671 (x) == AV_PIX_FMT_BGR48BE \
672 || (x) == AV_PIX_FMT_BGR48LE \
673 || (x) == AV_PIX_FMT_BGRA64BE \
674 || (x) == AV_PIX_FMT_BGRA64LE \
675 || (x) == AV_PIX_FMT_BGRA \
676 || (x) == AV_PIX_FMT_ABGR \
677 || (x) == AV_PIX_FMT_BGR24 \
680 #define isAnyRGB(x) \
684 (x)==AV_PIX_FMT_GBR24P \
687 static av_always_inline int isALPHA(enum AVPixelFormat pix_fmt)
689 const AVPixFmtDescriptor *desc = av_pix_fmt_desc_get(pix_fmt);
691 return desc->flags & PIX_FMT_ALPHA;
695 #define isPacked(x) ( \
696 (x)==AV_PIX_FMT_PAL8 \
697 || (x)==AV_PIX_FMT_YUYV422 \
698 || (x)==AV_PIX_FMT_UYVY422 \
699 || (x)==AV_PIX_FMT_Y400A \
704 static av_always_inline int isPacked(enum AVPixelFormat pix_fmt)
706 const AVPixFmtDescriptor *desc = av_pix_fmt_desc_get(pix_fmt);
708 return ((desc->nb_components >= 2 && !(desc->flags & PIX_FMT_PLANAR)) ||
709 pix_fmt == AV_PIX_FMT_PAL8);
713 static av_always_inline int isPlanar(enum AVPixelFormat pix_fmt)
715 const AVPixFmtDescriptor *desc = av_pix_fmt_desc_get(pix_fmt);
717 return (desc->nb_components >= 2 && (desc->flags & PIX_FMT_PLANAR));
720 static av_always_inline int isPackedRGB(enum AVPixelFormat pix_fmt)
722 const AVPixFmtDescriptor *desc = av_pix_fmt_desc_get(pix_fmt);
724 return ((desc->flags & (PIX_FMT_PLANAR | PIX_FMT_RGB)) == PIX_FMT_RGB);
727 static av_always_inline int isPlanarRGB(enum AVPixelFormat pix_fmt)
729 const AVPixFmtDescriptor *desc = av_pix_fmt_desc_get(pix_fmt);
731 return ((desc->flags & (PIX_FMT_PLANAR | PIX_FMT_RGB)) ==
732 (PIX_FMT_PLANAR | PIX_FMT_RGB));
735 static av_always_inline int usePal(enum AVPixelFormat pix_fmt)
737 const AVPixFmtDescriptor *desc = av_pix_fmt_desc_get(pix_fmt);
739 return (desc->flags & PIX_FMT_PAL) || (desc->flags & PIX_FMT_PSEUDOPAL);
742 extern const uint64_t ff_dither4[2];
743 extern const uint64_t ff_dither8[2];
744 extern const uint8_t dithers[8][8][8];
745 extern const uint16_t dither_scale[15][16];
748 extern const AVClass sws_context_class;
751 * Set c->swScale to an unscaled converter if one exists for the specific
752 * source and destination formats, bit depths, flags, etc.
754 void ff_get_unscaled_swscale(SwsContext *c);
756 void ff_swscale_get_unscaled_altivec(SwsContext *c);
759 * Return function pointer to fastest main scaler path function depending
760 * on architecture and available optimizations.
762 SwsFunc ff_getSwsFunc(SwsContext *c);
764 void ff_sws_init_input_funcs(SwsContext *c);
765 void ff_sws_init_output_funcs(SwsContext *c,
766 yuv2planar1_fn *yuv2plane1,
767 yuv2planarX_fn *yuv2planeX,
768 yuv2interleavedX_fn *yuv2nv12cX,
769 yuv2packed1_fn *yuv2packed1,
770 yuv2packed2_fn *yuv2packed2,
771 yuv2packedX_fn *yuv2packedX);
772 void ff_sws_init_swScale_altivec(SwsContext *c);
773 void ff_sws_init_swScale_mmx(SwsContext *c);
775 static inline void fillPlane16(uint8_t *plane, int stride, int width, int height, int y,
776 int alpha, int bits, const int big_endian)
779 uint8_t *ptr = plane + stride * y;
780 int v = alpha ? 0xFFFF>>(15-bits) : (1<<bits);
781 for (i = 0; i < height; i++) {
782 #define FILL(wfunc) \
783 for (j = 0; j < width; j++) {\
795 #endif /* SWSCALE_SWSCALE_INTERNAL_H */