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/log.h"
34 #include "libavutil/pixfmt.h"
35 #include "libavutil/pixdesc.h"
37 #define STR(s) AV_TOSTRING(s) // AV_STRINGIFY is too long
39 #define YUVRGB_TABLE_HEADROOM 128
41 #define FAST_BGR2YV12 // use 7-bit instead of 15-bit coefficients
43 #define MAX_FILTER_SIZE 256
48 #define ALT32_CORR (-1)
65 typedef int (*SwsFunc)(struct SwsContext *context, const uint8_t *src[],
66 int srcStride[], int srcSliceY, int srcSliceH,
67 uint8_t *dst[], int dstStride[]);
70 * Write one line of horizontally scaled data to planar output
71 * without any additional vertical scaling (or point-scaling).
73 * @param src scaled source data, 15bit for 8-10bit output,
74 * 19-bit for 16bit output (in int32_t)
75 * @param dest pointer to the output plane. For >8bit
76 * output, this is in uint16_t
77 * @param dstW width of destination in pixels
78 * @param dither ordered dither array of type int16_t and size 8
79 * @param offset Dither offset
81 typedef void (*yuv2planar1_fn)(const int16_t *src, uint8_t *dest, int dstW,
82 const uint8_t *dither, int offset);
85 * Write one line of horizontally scaled data to planar output
86 * with multi-point vertical scaling between input pixels.
88 * @param filter vertical luma/alpha scaling coefficients, 12bit [0,4096]
89 * @param src scaled luma (Y) or alpha (A) source data, 15bit for 8-10bit output,
90 * 19-bit for 16bit output (in int32_t)
91 * @param filterSize number of vertical input lines to scale
92 * @param dest pointer to output plane. For >8bit
93 * output, this is in uint16_t
94 * @param dstW width of destination pixels
95 * @param offset Dither offset
97 typedef void (*yuv2planarX_fn)(const int16_t *filter, int filterSize,
98 const int16_t **src, uint8_t *dest, int dstW,
99 const uint8_t *dither, int offset);
102 * Write one line of horizontally scaled chroma to interleaved output
103 * with multi-point vertical scaling between input pixels.
105 * @param c SWS scaling context
106 * @param chrFilter vertical chroma scaling coefficients, 12bit [0,4096]
107 * @param chrUSrc scaled chroma (U) source data, 15bit for 8-10bit output,
108 * 19-bit for 16bit output (in int32_t)
109 * @param chrVSrc scaled chroma (V) source data, 15bit for 8-10bit output,
110 * 19-bit for 16bit output (in int32_t)
111 * @param chrFilterSize number of vertical chroma input lines to scale
112 * @param dest pointer to the output plane. For >8bit
113 * output, this is in uint16_t
114 * @param dstW width of chroma planes
116 typedef void (*yuv2interleavedX_fn)(struct SwsContext *c,
117 const int16_t *chrFilter,
119 const int16_t **chrUSrc,
120 const int16_t **chrVSrc,
121 uint8_t *dest, int dstW);
124 * Write one line of horizontally scaled Y/U/V/A to packed-pixel YUV/RGB
125 * output without any additional vertical scaling (or point-scaling). Note
126 * that this function may do chroma scaling, see the "uvalpha" argument.
128 * @param c SWS scaling context
129 * @param lumSrc scaled luma (Y) source data, 15bit for 8-10bit output,
130 * 19-bit for 16bit output (in int32_t)
131 * @param chrUSrc scaled chroma (U) source data, 15bit for 8-10bit output,
132 * 19-bit for 16bit output (in int32_t)
133 * @param chrVSrc scaled chroma (V) source data, 15bit for 8-10bit output,
134 * 19-bit for 16bit output (in int32_t)
135 * @param alpSrc scaled alpha (A) source data, 15bit for 8-10bit output,
136 * 19-bit for 16bit output (in int32_t)
137 * @param dest pointer to the output plane. For 16bit output, this is
139 * @param dstW width of lumSrc and alpSrc in pixels, number of pixels
140 * to write into dest[]
141 * @param uvalpha chroma scaling coefficient for the second line of chroma
142 * pixels, either 2048 or 0. If 0, one chroma input is used
143 * for 2 output pixels (or if the SWS_FLAG_FULL_CHR_INT flag
144 * is set, it generates 1 output pixel). If 2048, two chroma
145 * input pixels should be averaged for 2 output pixels (this
146 * only happens if SWS_FLAG_FULL_CHR_INT is not set)
147 * @param y vertical line number for this output. This does not need
148 * to be used to calculate the offset in the destination,
149 * but can be used to generate comfort noise using dithering
150 * for some output formats.
152 typedef void (*yuv2packed1_fn)(struct SwsContext *c, const int16_t *lumSrc,
153 const int16_t *chrUSrc[2],
154 const int16_t *chrVSrc[2],
155 const int16_t *alpSrc, uint8_t *dest,
156 int dstW, int uvalpha, int y);
158 * Write one line of horizontally scaled Y/U/V/A to packed-pixel YUV/RGB
159 * output by doing bilinear scaling between two input lines.
161 * @param c SWS scaling context
162 * @param lumSrc scaled luma (Y) source data, 15bit for 8-10bit output,
163 * 19-bit for 16bit output (in int32_t)
164 * @param chrUSrc scaled chroma (U) source data, 15bit for 8-10bit output,
165 * 19-bit for 16bit output (in int32_t)
166 * @param chrVSrc scaled chroma (V) source data, 15bit for 8-10bit output,
167 * 19-bit for 16bit output (in int32_t)
168 * @param alpSrc scaled alpha (A) source data, 15bit for 8-10bit output,
169 * 19-bit for 16bit output (in int32_t)
170 * @param dest pointer to the output plane. For 16bit output, this is
172 * @param dstW width of lumSrc and alpSrc in pixels, number of pixels
173 * to write into dest[]
174 * @param yalpha luma/alpha scaling coefficients for the second input line.
175 * The first line's coefficients can be calculated by using
177 * @param uvalpha chroma scaling coefficient for the second input line. The
178 * first line's coefficients can be calculated by using
180 * @param y vertical line number for this output. This does not need
181 * to be used to calculate the offset in the destination,
182 * but can be used to generate comfort noise using dithering
183 * for some output formats.
185 typedef void (*yuv2packed2_fn)(struct SwsContext *c, const int16_t *lumSrc[2],
186 const int16_t *chrUSrc[2],
187 const int16_t *chrVSrc[2],
188 const int16_t *alpSrc[2],
190 int dstW, int yalpha, int uvalpha, int y);
192 * Write one line of horizontally scaled Y/U/V/A to packed-pixel YUV/RGB
193 * output by doing multi-point vertical scaling between input pixels.
195 * @param c SWS scaling context
196 * @param lumFilter vertical luma/alpha scaling coefficients, 12bit [0,4096]
197 * @param lumSrc scaled luma (Y) source data, 15bit for 8-10bit output,
198 * 19-bit for 16bit output (in int32_t)
199 * @param lumFilterSize number of vertical luma/alpha input lines to scale
200 * @param chrFilter vertical chroma scaling coefficients, 12bit [0,4096]
201 * @param chrUSrc scaled chroma (U) source data, 15bit for 8-10bit output,
202 * 19-bit for 16bit output (in int32_t)
203 * @param chrVSrc scaled chroma (V) source data, 15bit for 8-10bit output,
204 * 19-bit for 16bit output (in int32_t)
205 * @param chrFilterSize number of vertical chroma input lines to scale
206 * @param alpSrc scaled alpha (A) source data, 15bit for 8-10bit output,
207 * 19-bit for 16bit output (in int32_t)
208 * @param dest pointer to the output plane. For 16bit output, this is
210 * @param dstW width of lumSrc and alpSrc in pixels, number of pixels
211 * to write into dest[]
212 * @param y vertical line number for this output. This does not need
213 * to be used to calculate the offset in the destination,
214 * but can be used to generate comfort noise using dithering
215 * or some output formats.
217 typedef void (*yuv2packedX_fn)(struct SwsContext *c, const int16_t *lumFilter,
218 const int16_t **lumSrc, int lumFilterSize,
219 const int16_t *chrFilter,
220 const int16_t **chrUSrc,
221 const int16_t **chrVSrc, int chrFilterSize,
222 const int16_t **alpSrc, uint8_t *dest,
225 /* This struct should be aligned on at least a 32-byte boundary. */
226 typedef struct SwsContext {
228 * info on struct for av_log
230 const AVClass *av_class;
233 * Note that src, dst, srcStride, dstStride will be copied in the
234 * sws_scale() wrapper so they can be freely modified here.
237 int srcW; ///< Width of source luma/alpha planes.
238 int srcH; ///< Height of source luma/alpha planes.
239 int dstH; ///< Height of destination luma/alpha planes.
240 int chrSrcW; ///< Width of source chroma planes.
241 int chrSrcH; ///< Height of source chroma planes.
242 int chrDstW; ///< Width of destination chroma planes.
243 int chrDstH; ///< Height of destination chroma planes.
244 int lumXInc, chrXInc;
245 int lumYInc, chrYInc;
246 enum AVPixelFormat dstFormat; ///< Destination pixel format.
247 enum AVPixelFormat srcFormat; ///< Source pixel format.
248 int dstFormatBpp; ///< Number of bits per pixel of the destination pixel format.
249 int srcFormatBpp; ///< Number of bits per pixel of the source pixel format.
251 int chrSrcHSubSample; ///< Binary logarithm of horizontal subsampling factor between luma/alpha and chroma planes in source image.
252 int chrSrcVSubSample; ///< Binary logarithm of vertical subsampling factor between luma/alpha and chroma planes in source image.
253 int chrDstHSubSample; ///< Binary logarithm of horizontal subsampling factor between luma/alpha and chroma planes in destination image.
254 int chrDstVSubSample; ///< Binary logarithm of vertical subsampling factor between luma/alpha and chroma planes in destination image.
255 int vChrDrop; ///< Binary logarithm of extra vertical subsampling factor in source image chroma planes specified by user.
256 int sliceDir; ///< Direction that slices are fed to the scaler (1 = top-to-bottom, -1 = bottom-to-top).
257 double param[2]; ///< Input parameters for scaling algorithms that need them.
259 uint32_t pal_yuv[256];
260 uint32_t pal_rgb[256];
263 * @name Scaled horizontal lines ring buffer.
264 * The horizontal scaler keeps just enough scaled lines in a ring buffer
265 * so they may be passed to the vertical scaler. The pointers to the
266 * allocated buffers for each line are duplicated in sequence in the ring
267 * buffer to simplify indexing and avoid wrapping around between lines
268 * inside the vertical scaler code. The wrapping is done before the
269 * vertical scaler is called.
272 int16_t **lumPixBuf; ///< Ring buffer for scaled horizontal luma plane lines to be fed to the vertical scaler.
273 int16_t **chrUPixBuf; ///< Ring buffer for scaled horizontal chroma plane lines to be fed to the vertical scaler.
274 int16_t **chrVPixBuf; ///< Ring buffer for scaled horizontal chroma plane lines to be fed to the vertical scaler.
275 int16_t **alpPixBuf; ///< Ring buffer for scaled horizontal alpha plane lines to be fed to the vertical scaler.
276 int vLumBufSize; ///< Number of vertical luma/alpha lines allocated in the ring buffer.
277 int vChrBufSize; ///< Number of vertical chroma lines allocated in the ring buffer.
278 int lastInLumBuf; ///< Last scaled horizontal luma/alpha line from source in the ring buffer.
279 int lastInChrBuf; ///< Last scaled horizontal chroma line from source in the ring buffer.
280 int lumBufIndex; ///< Index in ring buffer of the last scaled horizontal luma/alpha line from source.
281 int chrBufIndex; ///< Index in ring buffer of the last scaled horizontal chroma line from source.
284 uint8_t *formatConvBuffer;
287 * @name Horizontal and vertical filters.
288 * To better understand the following fields, here is a pseudo-code of
289 * their usage in filtering a horizontal line:
291 * for (i = 0; i < width; i++) {
293 * for (j = 0; j < filterSize; j++)
294 * dst[i] += src[ filterPos[i] + j ] * filter[ filterSize * i + j ];
295 * dst[i] >>= FRAC_BITS; // The actual implementation is fixed-point.
300 int16_t *hLumFilter; ///< Array of horizontal filter coefficients for luma/alpha planes.
301 int16_t *hChrFilter; ///< Array of horizontal filter coefficients for chroma planes.
302 int16_t *vLumFilter; ///< Array of vertical filter coefficients for luma/alpha planes.
303 int16_t *vChrFilter; ///< Array of vertical filter coefficients for chroma planes.
304 int32_t *hLumFilterPos; ///< Array of horizontal filter starting positions for each dst[i] for luma/alpha planes.
305 int32_t *hChrFilterPos; ///< Array of horizontal filter starting positions for each dst[i] for chroma planes.
306 int32_t *vLumFilterPos; ///< Array of vertical filter starting positions for each dst[i] for luma/alpha planes.
307 int32_t *vChrFilterPos; ///< Array of vertical filter starting positions for each dst[i] for chroma planes.
308 int hLumFilterSize; ///< Horizontal filter size for luma/alpha pixels.
309 int hChrFilterSize; ///< Horizontal filter size for chroma pixels.
310 int vLumFilterSize; ///< Vertical filter size for luma/alpha pixels.
311 int vChrFilterSize; ///< Vertical filter size for chroma pixels.
314 int lumMmx2FilterCodeSize; ///< Runtime-generated MMX2 horizontal fast bilinear scaler code size for luma/alpha planes.
315 int chrMmx2FilterCodeSize; ///< Runtime-generated MMX2 horizontal fast bilinear scaler code size for chroma planes.
316 uint8_t *lumMmx2FilterCode; ///< Runtime-generated MMX2 horizontal fast bilinear scaler code for luma/alpha planes.
317 uint8_t *chrMmx2FilterCode; ///< Runtime-generated MMX2 horizontal fast bilinear scaler code for chroma planes.
321 int dstY; ///< Last destination vertical line output from last slice.
322 int flags; ///< Flags passed by the user to select scaler algorithm, optimizations, subsampling, etc...
323 void *yuvTable; // pointer to the yuv->rgb table start so it can be freed()
324 uint8_t *table_rV[256 + 2*YUVRGB_TABLE_HEADROOM];
325 uint8_t *table_gU[256 + 2*YUVRGB_TABLE_HEADROOM];
326 int table_gV[256 + 2*YUVRGB_TABLE_HEADROOM];
327 uint8_t *table_bU[256 + 2*YUVRGB_TABLE_HEADROOM];
330 int contrast, brightness, saturation; // for sws_getColorspaceDetails
331 int srcColorspaceTable[4];
332 int dstColorspaceTable[4];
333 int srcRange; ///< 0 = MPG YUV range, 1 = JPG YUV range (source image).
334 int dstRange; ///< 0 = MPG YUV range, 1 = JPG YUV range (destination image).
337 int yuv2rgb_y_offset;
339 int yuv2rgb_v2r_coeff;
340 int yuv2rgb_v2g_coeff;
341 int yuv2rgb_u2g_coeff;
342 int yuv2rgb_u2b_coeff;
344 #define RED_DITHER "0*8"
345 #define GREEN_DITHER "1*8"
346 #define BLUE_DITHER "2*8"
347 #define Y_COEFF "3*8"
348 #define VR_COEFF "4*8"
349 #define UB_COEFF "5*8"
350 #define VG_COEFF "6*8"
351 #define UG_COEFF "7*8"
352 #define Y_OFFSET "8*8"
353 #define U_OFFSET "9*8"
354 #define V_OFFSET "10*8"
355 #define LUM_MMX_FILTER_OFFSET "11*8"
356 #define CHR_MMX_FILTER_OFFSET "11*8+4*4*256"
357 #define DSTW_OFFSET "11*8+4*4*256*2" //do not change, it is hardcoded in the ASM
358 #define ESP_OFFSET "11*8+4*4*256*2+8"
359 #define VROUNDER_OFFSET "11*8+4*4*256*2+16"
360 #define U_TEMP "11*8+4*4*256*2+24"
361 #define V_TEMP "11*8+4*4*256*2+32"
362 #define Y_TEMP "11*8+4*4*256*2+40"
363 #define ALP_MMX_FILTER_OFFSET "11*8+4*4*256*2+48"
364 #define UV_OFF_PX "11*8+4*4*256*3+48"
365 #define UV_OFF_BYTE "11*8+4*4*256*3+56"
366 #define DITHER16 "11*8+4*4*256*3+64"
367 #define DITHER32 "11*8+4*4*256*3+80"
369 DECLARE_ALIGNED(8, uint64_t, redDither);
370 DECLARE_ALIGNED(8, uint64_t, greenDither);
371 DECLARE_ALIGNED(8, uint64_t, blueDither);
373 DECLARE_ALIGNED(8, uint64_t, yCoeff);
374 DECLARE_ALIGNED(8, uint64_t, vrCoeff);
375 DECLARE_ALIGNED(8, uint64_t, ubCoeff);
376 DECLARE_ALIGNED(8, uint64_t, vgCoeff);
377 DECLARE_ALIGNED(8, uint64_t, ugCoeff);
378 DECLARE_ALIGNED(8, uint64_t, yOffset);
379 DECLARE_ALIGNED(8, uint64_t, uOffset);
380 DECLARE_ALIGNED(8, uint64_t, vOffset);
381 int32_t lumMmxFilter[4 * MAX_FILTER_SIZE];
382 int32_t chrMmxFilter[4 * MAX_FILTER_SIZE];
383 int dstW; ///< Width of destination luma/alpha planes.
384 DECLARE_ALIGNED(8, uint64_t, esp);
385 DECLARE_ALIGNED(8, uint64_t, vRounder);
386 DECLARE_ALIGNED(8, uint64_t, u_temp);
387 DECLARE_ALIGNED(8, uint64_t, v_temp);
388 DECLARE_ALIGNED(8, uint64_t, y_temp);
389 int32_t alpMmxFilter[4 * MAX_FILTER_SIZE];
390 // alignment of these values is not necessary, but merely here
391 // to maintain the same offset across x8632 and x86-64. Once we
392 // use proper offset macros in the asm, they can be removed.
393 DECLARE_ALIGNED(8, ptrdiff_t, uv_off); ///< offset (in pixels) between u and v planes
394 DECLARE_ALIGNED(8, ptrdiff_t, uv_offx2); ///< offset (in bytes) between u and v planes
395 DECLARE_ALIGNED(8, uint16_t, dither16)[8];
396 DECLARE_ALIGNED(8, uint32_t, dither32)[8];
398 const uint8_t *chrDither8, *lumDither8;
401 vector signed short CY;
402 vector signed short CRV;
403 vector signed short CBU;
404 vector signed short CGU;
405 vector signed short CGV;
406 vector signed short OY;
407 vector unsigned short CSHIFT;
408 vector signed short *vYCoeffsBank, *vCCoeffsBank;
412 DECLARE_ALIGNED(4, uint32_t, oy);
413 DECLARE_ALIGNED(4, uint32_t, oc);
414 DECLARE_ALIGNED(4, uint32_t, zero);
415 DECLARE_ALIGNED(4, uint32_t, cy);
416 DECLARE_ALIGNED(4, uint32_t, crv);
417 DECLARE_ALIGNED(4, uint32_t, rmask);
418 DECLARE_ALIGNED(4, uint32_t, cbu);
419 DECLARE_ALIGNED(4, uint32_t, bmask);
420 DECLARE_ALIGNED(4, uint32_t, cgu);
421 DECLARE_ALIGNED(4, uint32_t, cgv);
422 DECLARE_ALIGNED(4, uint32_t, gmask);
426 DECLARE_ALIGNED(8, uint64_t, sparc_coeffs)[10];
430 /* function pointers for swScale() */
431 yuv2planar1_fn yuv2plane1;
432 yuv2planarX_fn yuv2planeX;
433 yuv2interleavedX_fn yuv2nv12cX;
434 yuv2packed1_fn yuv2packed1;
435 yuv2packed2_fn yuv2packed2;
436 yuv2packedX_fn yuv2packedX;
438 /// Unscaled conversion of luma plane to YV12 for horizontal scaler.
439 void (*lumToYV12)(uint8_t *dst, const uint8_t *src, const uint8_t *src2, const uint8_t *src3,
440 int width, uint32_t *pal);
441 /// Unscaled conversion of alpha plane to YV12 for horizontal scaler.
442 void (*alpToYV12)(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 chroma planes to YV12 for horizontal scaler.
445 void (*chrToYV12)(uint8_t *dstU, uint8_t *dstV,
446 const uint8_t *src1, const uint8_t *src2, const uint8_t *src3,
447 int width, uint32_t *pal);
450 * Functions to read planar input, such as planar RGB, and convert
451 * internally to Y/UV.
454 void (*readLumPlanar)(uint8_t *dst, const uint8_t *src[4], int width);
455 void (*readChrPlanar)(uint8_t *dstU, uint8_t *dstV, const uint8_t *src[4],
460 * Scale one horizontal line of input data using a bilinear filter
461 * to produce one line of output data. Compared to SwsContext->hScale(),
462 * please take note of the following caveats when using these:
463 * - Scaling is done using only 7bit instead of 14bit coefficients.
464 * - You can use no more than 5 input pixels to produce 4 output
465 * pixels. Therefore, this filter should not be used for downscaling
466 * by more than ~20% in width (because that equals more than 5/4th
467 * downscaling and thus more than 5 pixels input per 4 pixels output).
468 * - In general, bilinear filters create artifacts during downscaling
469 * (even when <20%), because one output pixel will span more than one
470 * input pixel, and thus some pixels will need edges of both neighbor
471 * pixels to interpolate the output pixel. Since you can use at most
472 * two input pixels per output pixel in bilinear scaling, this is
473 * impossible and thus downscaling by any size will create artifacts.
474 * To enable this type of scaling, set SWS_FLAG_FAST_BILINEAR
475 * in SwsContext->flags.
478 void (*hyscale_fast)(struct SwsContext *c,
479 int16_t *dst, int dstWidth,
480 const uint8_t *src, int srcW, int xInc);
481 void (*hcscale_fast)(struct SwsContext *c,
482 int16_t *dst1, int16_t *dst2, int dstWidth,
483 const uint8_t *src1, const uint8_t *src2,
488 * Scale one horizontal line of input data using a filter over the input
489 * lines, to produce one (differently sized) line of output data.
491 * @param dst pointer to destination buffer for horizontally scaled
492 * data. If the number of bits per component of one
493 * destination pixel (SwsContext->dstBpc) is <= 10, data
494 * will be 15bpc in 16bits (int16_t) width. Else (i.e.
495 * SwsContext->dstBpc == 16), data will be 19bpc in
496 * 32bits (int32_t) width.
497 * @param dstW width of destination image
498 * @param src pointer to source data to be scaled. If the number of
499 * bits per component of a source pixel (SwsContext->srcBpc)
500 * is 8, this is 8bpc in 8bits (uint8_t) width. Else
501 * (i.e. SwsContext->dstBpc > 8), this is native depth
502 * in 16bits (uint16_t) width. In other words, for 9-bit
503 * YUV input, this is 9bpc, for 10-bit YUV input, this is
504 * 10bpc, and for 16-bit RGB or YUV, this is 16bpc.
505 * @param filter filter coefficients to be used per output pixel for
506 * scaling. This contains 14bpp filtering coefficients.
507 * Guaranteed to contain dstW * filterSize entries.
508 * @param filterPos position of the first input pixel to be used for
509 * each output pixel during scaling. Guaranteed to
510 * contain dstW entries.
511 * @param filterSize the number of input coefficients to be used (and
512 * thus the number of input pixels to be used) for
513 * creating a single output pixel. Is aligned to 4
514 * (and input coefficients thus padded with zeroes)
515 * to simplify creating SIMD code.
518 void (*hyScale)(struct SwsContext *c, int16_t *dst, int dstW,
519 const uint8_t *src, const int16_t *filter,
520 const int32_t *filterPos, int filterSize);
521 void (*hcScale)(struct SwsContext *c, int16_t *dst, int dstW,
522 const uint8_t *src, const int16_t *filter,
523 const int32_t *filterPos, int filterSize);
526 /// Color range conversion function for luma plane if needed.
527 void (*lumConvertRange)(int16_t *dst, int width);
528 /// Color range conversion function for chroma planes if needed.
529 void (*chrConvertRange)(int16_t *dst1, int16_t *dst2, int width);
531 int needs_hcscale; ///< Set if there are chroma planes to be converted.
533 //FIXME check init (where 0)
535 SwsFunc ff_yuv2rgb_get_func_ptr(SwsContext *c);
536 int ff_yuv2rgb_c_init_tables(SwsContext *c, const int inv_table[4],
537 int fullRange, int brightness,
538 int contrast, int saturation);
540 void ff_yuv2rgb_init_tables_altivec(SwsContext *c, const int inv_table[4],
541 int brightness, int contrast, int saturation);
542 void updateMMXDitherTables(SwsContext *c, int dstY, int lumBufIndex, int chrBufIndex,
543 int lastInLumBuf, int lastInChrBuf);
545 SwsFunc ff_yuv2rgb_init_mmx(SwsContext *c);
546 SwsFunc ff_yuv2rgb_init_vis(SwsContext *c);
547 SwsFunc ff_yuv2rgb_init_altivec(SwsContext *c);
548 SwsFunc ff_yuv2rgb_get_func_ptr_bfin(SwsContext *c);
549 void ff_bfin_get_unscaled_swscale(SwsContext *c);
551 #if FF_API_SWS_FORMAT_NAME
553 * @deprecated Use av_get_pix_fmt_name() instead.
556 const char *sws_format_name(enum AVPixelFormat format);
559 static av_always_inline int is16BPS(enum AVPixelFormat pix_fmt)
561 const AVPixFmtDescriptor *desc = av_pix_fmt_desc_get(pix_fmt);
563 return desc->comp[0].depth_minus1 == 15;
566 static av_always_inline int is9_OR_10BPS(enum AVPixelFormat pix_fmt)
568 const AVPixFmtDescriptor *desc = av_pix_fmt_desc_get(pix_fmt);
570 return desc->comp[0].depth_minus1 >= 8 && desc->comp[0].depth_minus1 <= 13;
573 #define isNBPS(x) is9_OR_10BPS(x)
575 static av_always_inline int isBE(enum AVPixelFormat pix_fmt)
577 const AVPixFmtDescriptor *desc = av_pix_fmt_desc_get(pix_fmt);
579 return desc->flags & PIX_FMT_BE;
582 static av_always_inline int isYUV(enum AVPixelFormat pix_fmt)
584 const AVPixFmtDescriptor *desc = av_pix_fmt_desc_get(pix_fmt);
586 return !(desc->flags & PIX_FMT_RGB) && desc->nb_components >= 2;
589 static av_always_inline int isPlanarYUV(enum AVPixelFormat pix_fmt)
591 const AVPixFmtDescriptor *desc = av_pix_fmt_desc_get(pix_fmt);
593 return ((desc->flags & PIX_FMT_PLANAR) && isYUV(pix_fmt));
596 static av_always_inline int isRGB(enum AVPixelFormat pix_fmt)
598 const AVPixFmtDescriptor *desc = av_pix_fmt_desc_get(pix_fmt);
600 return (desc->flags & PIX_FMT_RGB);
605 (!(av_pix_fmt_desc_get(x)->flags & PIX_FMT_PAL) && \
606 av_pix_fmt_desc_get(x)->nb_components <= 2)
609 ((x) == AV_PIX_FMT_GRAY8 || \
610 (x) == AV_PIX_FMT_Y400A || \
611 (x) == AV_PIX_FMT_GRAY16BE || \
612 (x) == AV_PIX_FMT_GRAY16LE)
615 #define isRGBinInt(x) \
617 (x) == AV_PIX_FMT_RGB48BE || \
618 (x) == AV_PIX_FMT_RGB48LE || \
619 (x) == AV_PIX_FMT_RGBA64BE || \
620 (x) == AV_PIX_FMT_RGBA64LE || \
621 (x) == AV_PIX_FMT_RGB32 || \
622 (x) == AV_PIX_FMT_RGB32_1 || \
623 (x) == AV_PIX_FMT_RGB24 || \
624 (x) == AV_PIX_FMT_RGB565BE || \
625 (x) == AV_PIX_FMT_RGB565LE || \
626 (x) == AV_PIX_FMT_RGB555BE || \
627 (x) == AV_PIX_FMT_RGB555LE || \
628 (x) == AV_PIX_FMT_RGB444BE || \
629 (x) == AV_PIX_FMT_RGB444LE || \
630 (x) == AV_PIX_FMT_RGB8 || \
631 (x) == AV_PIX_FMT_RGB4 || \
632 (x) == AV_PIX_FMT_RGB4_BYTE || \
633 (x) == AV_PIX_FMT_MONOBLACK || \
634 (x) == AV_PIX_FMT_MONOWHITE \
636 #define isBGRinInt(x) \
638 (x) == AV_PIX_FMT_BGR48BE || \
639 (x) == AV_PIX_FMT_BGR48LE || \
640 (x) == AV_PIX_FMT_BGRA64BE || \
641 (x) == AV_PIX_FMT_BGRA64LE || \
642 (x) == AV_PIX_FMT_BGR32 || \
643 (x) == AV_PIX_FMT_BGR32_1 || \
644 (x) == AV_PIX_FMT_BGR24 || \
645 (x) == AV_PIX_FMT_BGR565BE || \
646 (x) == AV_PIX_FMT_BGR565LE || \
647 (x) == AV_PIX_FMT_BGR555BE || \
648 (x) == AV_PIX_FMT_BGR555LE || \
649 (x) == AV_PIX_FMT_BGR444BE || \
650 (x) == AV_PIX_FMT_BGR444LE || \
651 (x) == AV_PIX_FMT_BGR8 || \
652 (x) == AV_PIX_FMT_BGR4 || \
653 (x) == AV_PIX_FMT_BGR4_BYTE || \
654 (x) == AV_PIX_FMT_MONOBLACK || \
655 (x) == AV_PIX_FMT_MONOWHITE \
658 #define isRGBinBytes(x) ( \
659 (x) == AV_PIX_FMT_RGB48BE \
660 || (x) == AV_PIX_FMT_RGB48LE \
661 || (x) == AV_PIX_FMT_RGBA64BE \
662 || (x) == AV_PIX_FMT_RGBA64LE \
663 || (x) == AV_PIX_FMT_RGBA \
664 || (x) == AV_PIX_FMT_ARGB \
665 || (x) == AV_PIX_FMT_RGB24 \
667 #define isBGRinBytes(x) ( \
668 (x) == AV_PIX_FMT_BGR48BE \
669 || (x) == AV_PIX_FMT_BGR48LE \
670 || (x) == AV_PIX_FMT_BGRA64BE \
671 || (x) == AV_PIX_FMT_BGRA64LE \
672 || (x) == AV_PIX_FMT_BGRA \
673 || (x) == AV_PIX_FMT_ABGR \
674 || (x) == AV_PIX_FMT_BGR24 \
677 #define isAnyRGB(x) \
681 (x)==AV_PIX_FMT_GBR24P \
684 static av_always_inline int isALPHA(enum AVPixelFormat pix_fmt)
686 const AVPixFmtDescriptor *desc = av_pix_fmt_desc_get(pix_fmt);
688 return desc->nb_components == 2 || desc->nb_components == 4;
692 #define isPacked(x) ( \
693 (x)==AV_PIX_FMT_PAL8 \
694 || (x)==AV_PIX_FMT_YUYV422 \
695 || (x)==AV_PIX_FMT_UYVY422 \
696 || (x)==AV_PIX_FMT_Y400A \
701 static av_always_inline int isPacked(enum AVPixelFormat pix_fmt)
703 const AVPixFmtDescriptor *desc = av_pix_fmt_desc_get(pix_fmt);
705 return ((desc->nb_components >= 2 && !(desc->flags & PIX_FMT_PLANAR)) ||
706 pix_fmt == AV_PIX_FMT_PAL8);
710 static av_always_inline int isPlanar(enum AVPixelFormat pix_fmt)
712 const AVPixFmtDescriptor *desc = av_pix_fmt_desc_get(pix_fmt);
714 return (desc->nb_components >= 2 && (desc->flags & PIX_FMT_PLANAR));
717 static av_always_inline int isPackedRGB(enum AVPixelFormat pix_fmt)
719 const AVPixFmtDescriptor *desc = av_pix_fmt_desc_get(pix_fmt);
721 return ((desc->flags & (PIX_FMT_PLANAR | PIX_FMT_RGB)) == PIX_FMT_RGB);
724 static av_always_inline int isPlanarRGB(enum AVPixelFormat pix_fmt)
726 const AVPixFmtDescriptor *desc = av_pix_fmt_desc_get(pix_fmt);
728 return ((desc->flags & (PIX_FMT_PLANAR | PIX_FMT_RGB)) ==
729 (PIX_FMT_PLANAR | PIX_FMT_RGB));
732 static av_always_inline int usePal(enum AVPixelFormat pix_fmt)
734 const AVPixFmtDescriptor *desc = av_pix_fmt_desc_get(pix_fmt);
736 return (desc->flags & PIX_FMT_PAL) || (desc->flags & PIX_FMT_PSEUDOPAL);
739 extern const uint64_t ff_dither4[2];
740 extern const uint64_t ff_dither8[2];
741 extern const uint8_t dithers[8][8][8];
742 extern const uint16_t dither_scale[15][16];
745 extern const AVClass sws_context_class;
748 * Set c->swScale to an unscaled converter if one exists for the specific
749 * source and destination formats, bit depths, flags, etc.
751 void ff_get_unscaled_swscale(SwsContext *c);
753 void ff_swscale_get_unscaled_altivec(SwsContext *c);
756 * Return function pointer to fastest main scaler path function depending
757 * on architecture and available optimizations.
759 SwsFunc ff_getSwsFunc(SwsContext *c);
761 void ff_sws_init_input_funcs(SwsContext *c);
762 void ff_sws_init_output_funcs(SwsContext *c,
763 yuv2planar1_fn *yuv2plane1,
764 yuv2planarX_fn *yuv2planeX,
765 yuv2interleavedX_fn *yuv2nv12cX,
766 yuv2packed1_fn *yuv2packed1,
767 yuv2packed2_fn *yuv2packed2,
768 yuv2packedX_fn *yuv2packedX);
769 void ff_sws_init_swScale_altivec(SwsContext *c);
770 void ff_sws_init_swScale_mmx(SwsContext *c);
772 #endif /* SWSCALE_SWSCALE_INTERNAL_H */