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 MAX_FILTER_SIZE 256
47 #define ALT32_CORR (-1)
64 typedef enum SwsDither {
72 typedef int (*SwsFunc)(struct SwsContext *context, const uint8_t *src[],
73 int srcStride[], int srcSliceY, int srcSliceH,
74 uint8_t *dst[], int dstStride[]);
77 * Write one line of horizontally scaled data to planar output
78 * without any additional vertical scaling (or point-scaling).
80 * @param src scaled source data, 15bit for 8-10bit output,
81 * 19-bit for 16bit output (in int32_t)
82 * @param dest pointer to the output plane. For >8bit
83 * output, this is in uint16_t
84 * @param dstW width of destination in pixels
85 * @param dither ordered dither array of type int16_t and size 8
86 * @param offset Dither offset
88 typedef void (*yuv2planar1_fn)(const int16_t *src, uint8_t *dest, int dstW,
89 const uint8_t *dither, int offset);
92 * Write one line of horizontally scaled data to planar output
93 * with multi-point vertical scaling between input pixels.
95 * @param filter vertical luma/alpha scaling coefficients, 12bit [0,4096]
96 * @param src scaled luma (Y) or alpha (A) source data, 15bit for 8-10bit output,
97 * 19-bit for 16bit output (in int32_t)
98 * @param filterSize number of vertical input lines to scale
99 * @param dest pointer to output plane. For >8bit
100 * output, this is in uint16_t
101 * @param dstW width of destination pixels
102 * @param offset Dither offset
104 typedef void (*yuv2planarX_fn)(const int16_t *filter, int filterSize,
105 const int16_t **src, uint8_t *dest, int dstW,
106 const uint8_t *dither, int offset);
109 * Write one line of horizontally scaled chroma to interleaved output
110 * with multi-point vertical scaling between input pixels.
112 * @param c SWS scaling context
113 * @param chrFilter vertical chroma scaling coefficients, 12bit [0,4096]
114 * @param chrUSrc scaled chroma (U) source data, 15bit for 8-10bit output,
115 * 19-bit for 16bit output (in int32_t)
116 * @param chrVSrc scaled chroma (V) source data, 15bit for 8-10bit output,
117 * 19-bit for 16bit output (in int32_t)
118 * @param chrFilterSize number of vertical chroma input lines to scale
119 * @param dest pointer to the output plane. For >8bit
120 * output, this is in uint16_t
121 * @param dstW width of chroma planes
123 typedef void (*yuv2interleavedX_fn)(struct SwsContext *c,
124 const int16_t *chrFilter,
126 const int16_t **chrUSrc,
127 const int16_t **chrVSrc,
128 uint8_t *dest, int dstW);
131 * Write one line of horizontally scaled Y/U/V/A to packed-pixel YUV/RGB
132 * output without any additional vertical scaling (or point-scaling). Note
133 * that this function may do chroma scaling, see the "uvalpha" argument.
135 * @param c SWS scaling context
136 * @param lumSrc scaled luma (Y) source data, 15bit for 8-10bit output,
137 * 19-bit for 16bit output (in int32_t)
138 * @param chrUSrc scaled chroma (U) source data, 15bit for 8-10bit output,
139 * 19-bit for 16bit output (in int32_t)
140 * @param chrVSrc scaled chroma (V) source data, 15bit for 8-10bit output,
141 * 19-bit for 16bit output (in int32_t)
142 * @param alpSrc scaled alpha (A) source data, 15bit for 8-10bit output,
143 * 19-bit for 16bit output (in int32_t)
144 * @param dest pointer to the output plane. For 16bit output, this is
146 * @param dstW width of lumSrc and alpSrc in pixels, number of pixels
147 * to write into dest[]
148 * @param uvalpha chroma scaling coefficient for the second line of chroma
149 * pixels, either 2048 or 0. If 0, one chroma input is used
150 * for 2 output pixels (or if the SWS_FLAG_FULL_CHR_INT flag
151 * is set, it generates 1 output pixel). If 2048, two chroma
152 * input pixels should be averaged for 2 output pixels (this
153 * only happens if SWS_FLAG_FULL_CHR_INT is not set)
154 * @param y vertical line number for this output. This does not need
155 * to be used to calculate the offset in the destination,
156 * but can be used to generate comfort noise using dithering
157 * for some output formats.
159 typedef void (*yuv2packed1_fn)(struct SwsContext *c, const int16_t *lumSrc,
160 const int16_t *chrUSrc[2],
161 const int16_t *chrVSrc[2],
162 const int16_t *alpSrc, uint8_t *dest,
163 int dstW, int uvalpha, int y);
165 * Write one line of horizontally scaled Y/U/V/A to packed-pixel YUV/RGB
166 * output by doing bilinear scaling between two input lines.
168 * @param c SWS scaling context
169 * @param lumSrc scaled luma (Y) source data, 15bit for 8-10bit output,
170 * 19-bit for 16bit output (in int32_t)
171 * @param chrUSrc scaled chroma (U) source data, 15bit for 8-10bit output,
172 * 19-bit for 16bit output (in int32_t)
173 * @param chrVSrc scaled chroma (V) source data, 15bit for 8-10bit output,
174 * 19-bit for 16bit output (in int32_t)
175 * @param alpSrc scaled alpha (A) source data, 15bit for 8-10bit output,
176 * 19-bit for 16bit output (in int32_t)
177 * @param dest pointer to the output plane. For 16bit output, this is
179 * @param dstW width of lumSrc and alpSrc in pixels, number of pixels
180 * to write into dest[]
181 * @param yalpha luma/alpha scaling coefficients for the second input line.
182 * The first line's coefficients can be calculated by using
184 * @param uvalpha chroma scaling coefficient for the second input line. The
185 * first line's coefficients can be calculated by using
187 * @param y vertical line number for this output. This does not need
188 * to be used to calculate the offset in the destination,
189 * but can be used to generate comfort noise using dithering
190 * for some output formats.
192 typedef void (*yuv2packed2_fn)(struct SwsContext *c, const int16_t *lumSrc[2],
193 const int16_t *chrUSrc[2],
194 const int16_t *chrVSrc[2],
195 const int16_t *alpSrc[2],
197 int dstW, int yalpha, int uvalpha, int y);
199 * Write one line of horizontally scaled Y/U/V/A to packed-pixel YUV/RGB
200 * output by doing multi-point vertical scaling between input pixels.
202 * @param c SWS scaling context
203 * @param lumFilter vertical luma/alpha scaling coefficients, 12bit [0,4096]
204 * @param lumSrc scaled luma (Y) source data, 15bit for 8-10bit output,
205 * 19-bit for 16bit output (in int32_t)
206 * @param lumFilterSize number of vertical luma/alpha input lines to scale
207 * @param chrFilter vertical chroma scaling coefficients, 12bit [0,4096]
208 * @param chrUSrc scaled chroma (U) source data, 15bit for 8-10bit output,
209 * 19-bit for 16bit output (in int32_t)
210 * @param chrVSrc scaled chroma (V) source data, 15bit for 8-10bit output,
211 * 19-bit for 16bit output (in int32_t)
212 * @param chrFilterSize number of vertical chroma input lines to scale
213 * @param alpSrc scaled alpha (A) source data, 15bit for 8-10bit output,
214 * 19-bit for 16bit output (in int32_t)
215 * @param dest pointer to the output plane. For 16bit output, this is
217 * @param dstW width of lumSrc and alpSrc in pixels, number of pixels
218 * to write into dest[]
219 * @param y vertical line number for this output. This does not need
220 * to be used to calculate the offset in the destination,
221 * but can be used to generate comfort noise using dithering
222 * or some output formats.
224 typedef void (*yuv2packedX_fn)(struct SwsContext *c, const int16_t *lumFilter,
225 const int16_t **lumSrc, int lumFilterSize,
226 const int16_t *chrFilter,
227 const int16_t **chrUSrc,
228 const int16_t **chrVSrc, int chrFilterSize,
229 const int16_t **alpSrc, uint8_t *dest,
233 * Write one line of horizontally scaled Y/U/V/A to YUV/RGB
234 * output by doing multi-point vertical scaling between input pixels.
236 * @param c SWS scaling context
237 * @param lumFilter vertical luma/alpha scaling coefficients, 12bit [0,4096]
238 * @param lumSrc scaled luma (Y) source data, 15bit for 8-10bit output,
239 * 19-bit for 16bit output (in int32_t)
240 * @param lumFilterSize number of vertical luma/alpha input lines to scale
241 * @param chrFilter vertical chroma scaling coefficients, 12bit [0,4096]
242 * @param chrUSrc scaled chroma (U) source data, 15bit for 8-10bit output,
243 * 19-bit for 16bit output (in int32_t)
244 * @param chrVSrc scaled chroma (V) source data, 15bit for 8-10bit output,
245 * 19-bit for 16bit output (in int32_t)
246 * @param chrFilterSize number of vertical chroma input lines to scale
247 * @param alpSrc scaled alpha (A) source data, 15bit for 8-10bit output,
248 * 19-bit for 16bit output (in int32_t)
249 * @param dest pointer to the output planes. For 16bit output, this is
251 * @param dstW width of lumSrc and alpSrc in pixels, number of pixels
252 * to write into dest[]
253 * @param y vertical line number for this output. This does not need
254 * to be used to calculate the offset in the destination,
255 * but can be used to generate comfort noise using dithering
256 * or some output formats.
258 typedef void (*yuv2anyX_fn)(struct SwsContext *c, const int16_t *lumFilter,
259 const int16_t **lumSrc, int lumFilterSize,
260 const int16_t *chrFilter,
261 const int16_t **chrUSrc,
262 const int16_t **chrVSrc, int chrFilterSize,
263 const int16_t **alpSrc, uint8_t **dest,
266 /* This struct should be aligned on at least a 32-byte boundary. */
267 typedef struct SwsContext {
269 * info on struct for av_log
271 const AVClass *av_class;
274 * Note that src, dst, srcStride, dstStride will be copied in the
275 * sws_scale() wrapper so they can be freely modified here.
278 int srcW; ///< Width of source luma/alpha planes.
279 int srcH; ///< Height of source luma/alpha planes.
280 int dstH; ///< Height of destination luma/alpha planes.
281 int chrSrcW; ///< Width of source chroma planes.
282 int chrSrcH; ///< Height of source chroma planes.
283 int chrDstW; ///< Width of destination chroma planes.
284 int chrDstH; ///< Height of destination chroma planes.
285 int lumXInc, chrXInc;
286 int lumYInc, chrYInc;
287 enum AVPixelFormat dstFormat; ///< Destination pixel format.
288 enum AVPixelFormat srcFormat; ///< Source pixel format.
289 int dstFormatBpp; ///< Number of bits per pixel of the destination pixel format.
290 int srcFormatBpp; ///< Number of bits per pixel of the source pixel format.
292 int chrSrcHSubSample; ///< Binary logarithm of horizontal subsampling factor between luma/alpha and chroma planes in source image.
293 int chrSrcVSubSample; ///< Binary logarithm of vertical subsampling factor between luma/alpha and chroma planes in source image.
294 int chrDstHSubSample; ///< Binary logarithm of horizontal subsampling factor between luma/alpha and chroma planes in destination image.
295 int chrDstVSubSample; ///< Binary logarithm of vertical subsampling factor between luma/alpha and chroma planes in destination image.
296 int vChrDrop; ///< Binary logarithm of extra vertical subsampling factor in source image chroma planes specified by user.
297 int sliceDir; ///< Direction that slices are fed to the scaler (1 = top-to-bottom, -1 = bottom-to-top).
298 double param[2]; ///< Input parameters for scaling algorithms that need them.
300 uint32_t pal_yuv[256];
301 uint32_t pal_rgb[256];
304 * @name Scaled horizontal lines ring buffer.
305 * The horizontal scaler keeps just enough scaled lines in a ring buffer
306 * so they may be passed to the vertical scaler. The pointers to the
307 * allocated buffers for each line are duplicated in sequence in the ring
308 * buffer to simplify indexing and avoid wrapping around between lines
309 * inside the vertical scaler code. The wrapping is done before the
310 * vertical scaler is called.
313 int16_t **lumPixBuf; ///< Ring buffer for scaled horizontal luma plane lines to be fed to the vertical scaler.
314 int16_t **chrUPixBuf; ///< Ring buffer for scaled horizontal chroma plane lines to be fed to the vertical scaler.
315 int16_t **chrVPixBuf; ///< Ring buffer for scaled horizontal chroma plane lines to be fed to the vertical scaler.
316 int16_t **alpPixBuf; ///< Ring buffer for scaled horizontal alpha plane lines to be fed to the vertical scaler.
317 int vLumBufSize; ///< Number of vertical luma/alpha lines allocated in the ring buffer.
318 int vChrBufSize; ///< Number of vertical chroma lines allocated in the ring buffer.
319 int lastInLumBuf; ///< Last scaled horizontal luma/alpha line from source in the ring buffer.
320 int lastInChrBuf; ///< Last scaled horizontal chroma line from source in the ring buffer.
321 int lumBufIndex; ///< Index in ring buffer of the last scaled horizontal luma/alpha line from source.
322 int chrBufIndex; ///< Index in ring buffer of the last scaled horizontal chroma line from source.
325 uint8_t *formatConvBuffer;
328 * @name Horizontal and vertical filters.
329 * To better understand the following fields, here is a pseudo-code of
330 * their usage in filtering a horizontal line:
332 * for (i = 0; i < width; i++) {
334 * for (j = 0; j < filterSize; j++)
335 * dst[i] += src[ filterPos[i] + j ] * filter[ filterSize * i + j ];
336 * dst[i] >>= FRAC_BITS; // The actual implementation is fixed-point.
341 int16_t *hLumFilter; ///< Array of horizontal filter coefficients for luma/alpha planes.
342 int16_t *hChrFilter; ///< Array of horizontal filter coefficients for chroma planes.
343 int16_t *vLumFilter; ///< Array of vertical filter coefficients for luma/alpha planes.
344 int16_t *vChrFilter; ///< Array of vertical filter coefficients for chroma planes.
345 int32_t *hLumFilterPos; ///< Array of horizontal filter starting positions for each dst[i] for luma/alpha planes.
346 int32_t *hChrFilterPos; ///< Array of horizontal filter starting positions for each dst[i] for chroma planes.
347 int32_t *vLumFilterPos; ///< Array of vertical filter starting positions for each dst[i] for luma/alpha planes.
348 int32_t *vChrFilterPos; ///< Array of vertical filter starting positions for each dst[i] for chroma planes.
349 int hLumFilterSize; ///< Horizontal filter size for luma/alpha pixels.
350 int hChrFilterSize; ///< Horizontal filter size for chroma pixels.
351 int vLumFilterSize; ///< Vertical filter size for luma/alpha pixels.
352 int vChrFilterSize; ///< Vertical filter size for chroma pixels.
355 int lumMmxextFilterCodeSize; ///< Runtime-generated MMXEXT horizontal fast bilinear scaler code size for luma/alpha planes.
356 int chrMmxextFilterCodeSize; ///< Runtime-generated MMXEXT horizontal fast bilinear scaler code size for chroma planes.
357 uint8_t *lumMmxextFilterCode; ///< Runtime-generated MMXEXT horizontal fast bilinear scaler code for luma/alpha planes.
358 uint8_t *chrMmxextFilterCode; ///< Runtime-generated MMXEXT horizontal fast bilinear scaler code for chroma planes.
362 int dstY; ///< Last destination vertical line output from last slice.
363 int flags; ///< Flags passed by the user to select scaler algorithm, optimizations, subsampling, etc...
364 void *yuvTable; // pointer to the yuv->rgb table start so it can be freed()
365 uint8_t *table_rV[256 + 2*YUVRGB_TABLE_HEADROOM];
366 uint8_t *table_gU[256 + 2*YUVRGB_TABLE_HEADROOM];
367 int table_gV[256 + 2*YUVRGB_TABLE_HEADROOM];
368 uint8_t *table_bU[256 + 2*YUVRGB_TABLE_HEADROOM];
369 DECLARE_ALIGNED(16, int32_t, input_rgb2yuv_table)[16+40*4]; // This table can contain both C and SIMD formatted values, teh C vales are always at the XY_IDX points
379 #define RGB2YUV_SHIFT 15
381 int *dither_error[4];
384 int contrast, brightness, saturation; // for sws_getColorspaceDetails
385 int srcColorspaceTable[4];
386 int dstColorspaceTable[4];
387 int srcRange; ///< 0 = MPG YUV range, 1 = JPG YUV range (source image).
388 int dstRange; ///< 0 = MPG YUV range, 1 = JPG YUV range (destination image).
397 int yuv2rgb_y_offset;
399 int yuv2rgb_v2r_coeff;
400 int yuv2rgb_v2g_coeff;
401 int yuv2rgb_u2g_coeff;
402 int yuv2rgb_u2b_coeff;
404 #define RED_DITHER "0*8"
405 #define GREEN_DITHER "1*8"
406 #define BLUE_DITHER "2*8"
407 #define Y_COEFF "3*8"
408 #define VR_COEFF "4*8"
409 #define UB_COEFF "5*8"
410 #define VG_COEFF "6*8"
411 #define UG_COEFF "7*8"
412 #define Y_OFFSET "8*8"
413 #define U_OFFSET "9*8"
414 #define V_OFFSET "10*8"
415 #define LUM_MMX_FILTER_OFFSET "11*8"
416 #define CHR_MMX_FILTER_OFFSET "11*8+4*4*256"
417 #define DSTW_OFFSET "11*8+4*4*256*2" //do not change, it is hardcoded in the ASM
418 #define ESP_OFFSET "11*8+4*4*256*2+8"
419 #define VROUNDER_OFFSET "11*8+4*4*256*2+16"
420 #define U_TEMP "11*8+4*4*256*2+24"
421 #define V_TEMP "11*8+4*4*256*2+32"
422 #define Y_TEMP "11*8+4*4*256*2+40"
423 #define ALP_MMX_FILTER_OFFSET "11*8+4*4*256*2+48"
424 #define UV_OFF_PX "11*8+4*4*256*3+48"
425 #define UV_OFF_BYTE "11*8+4*4*256*3+56"
426 #define DITHER16 "11*8+4*4*256*3+64"
427 #define DITHER32 "11*8+4*4*256*3+80"
429 DECLARE_ALIGNED(8, uint64_t, redDither);
430 DECLARE_ALIGNED(8, uint64_t, greenDither);
431 DECLARE_ALIGNED(8, uint64_t, blueDither);
433 DECLARE_ALIGNED(8, uint64_t, yCoeff);
434 DECLARE_ALIGNED(8, uint64_t, vrCoeff);
435 DECLARE_ALIGNED(8, uint64_t, ubCoeff);
436 DECLARE_ALIGNED(8, uint64_t, vgCoeff);
437 DECLARE_ALIGNED(8, uint64_t, ugCoeff);
438 DECLARE_ALIGNED(8, uint64_t, yOffset);
439 DECLARE_ALIGNED(8, uint64_t, uOffset);
440 DECLARE_ALIGNED(8, uint64_t, vOffset);
441 int32_t lumMmxFilter[4 * MAX_FILTER_SIZE];
442 int32_t chrMmxFilter[4 * MAX_FILTER_SIZE];
443 int dstW; ///< Width of destination luma/alpha planes.
444 DECLARE_ALIGNED(8, uint64_t, esp);
445 DECLARE_ALIGNED(8, uint64_t, vRounder);
446 DECLARE_ALIGNED(8, uint64_t, u_temp);
447 DECLARE_ALIGNED(8, uint64_t, v_temp);
448 DECLARE_ALIGNED(8, uint64_t, y_temp);
449 int32_t alpMmxFilter[4 * MAX_FILTER_SIZE];
450 // alignment of these values is not necessary, but merely here
451 // to maintain the same offset across x8632 and x86-64. Once we
452 // use proper offset macros in the asm, they can be removed.
453 DECLARE_ALIGNED(8, ptrdiff_t, uv_off); ///< offset (in pixels) between u and v planes
454 DECLARE_ALIGNED(8, ptrdiff_t, uv_offx2); ///< offset (in bytes) between u and v planes
455 DECLARE_ALIGNED(8, uint16_t, dither16)[8];
456 DECLARE_ALIGNED(8, uint32_t, dither32)[8];
458 const uint8_t *chrDither8, *lumDither8;
461 vector signed short CY;
462 vector signed short CRV;
463 vector signed short CBU;
464 vector signed short CGU;
465 vector signed short CGV;
466 vector signed short OY;
467 vector unsigned short CSHIFT;
468 vector signed short *vYCoeffsBank, *vCCoeffsBank;
472 DECLARE_ALIGNED(4, uint32_t, oy);
473 DECLARE_ALIGNED(4, uint32_t, oc);
474 DECLARE_ALIGNED(4, uint32_t, zero);
475 DECLARE_ALIGNED(4, uint32_t, cy);
476 DECLARE_ALIGNED(4, uint32_t, crv);
477 DECLARE_ALIGNED(4, uint32_t, rmask);
478 DECLARE_ALIGNED(4, uint32_t, cbu);
479 DECLARE_ALIGNED(4, uint32_t, bmask);
480 DECLARE_ALIGNED(4, uint32_t, cgu);
481 DECLARE_ALIGNED(4, uint32_t, cgv);
482 DECLARE_ALIGNED(4, uint32_t, gmask);
486 DECLARE_ALIGNED(8, uint64_t, sparc_coeffs)[10];
490 /* pre defined color-spaces gamma */
491 #define XYZ_GAMMA (2.6f)
492 #define RGB_GAMMA (2.2f)
495 int16_t *xyzgammainv;
496 int16_t *rgbgammainv;
497 int16_t xyz2rgb_matrix[3][4];
498 int16_t rgb2xyz_matrix[3][4];
500 /* function pointers for swScale() */
501 yuv2planar1_fn yuv2plane1;
502 yuv2planarX_fn yuv2planeX;
503 yuv2interleavedX_fn yuv2nv12cX;
504 yuv2packed1_fn yuv2packed1;
505 yuv2packed2_fn yuv2packed2;
506 yuv2packedX_fn yuv2packedX;
507 yuv2anyX_fn yuv2anyX;
509 /// Unscaled conversion of luma plane to YV12 for horizontal scaler.
510 void (*lumToYV12)(uint8_t *dst, const uint8_t *src, const uint8_t *src2, const uint8_t *src3,
511 int width, uint32_t *pal);
512 /// Unscaled conversion of alpha plane to YV12 for horizontal scaler.
513 void (*alpToYV12)(uint8_t *dst, const uint8_t *src, const uint8_t *src2, const uint8_t *src3,
514 int width, uint32_t *pal);
515 /// Unscaled conversion of chroma planes to YV12 for horizontal scaler.
516 void (*chrToYV12)(uint8_t *dstU, uint8_t *dstV,
517 const uint8_t *src1, const uint8_t *src2, const uint8_t *src3,
518 int width, uint32_t *pal);
521 * Functions to read planar input, such as planar RGB, and convert
522 * internally to Y/UV/A.
525 void (*readLumPlanar)(uint8_t *dst, const uint8_t *src[4], int width, int32_t *rgb2yuv);
526 void (*readChrPlanar)(uint8_t *dstU, uint8_t *dstV, const uint8_t *src[4],
527 int width, int32_t *rgb2yuv);
528 void (*readAlpPlanar)(uint8_t *dst, const uint8_t *src[4], int width, int32_t *rgb2yuv);
532 * Scale one horizontal line of input data using a bilinear filter
533 * to produce one line of output data. Compared to SwsContext->hScale(),
534 * please take note of the following caveats when using these:
535 * - Scaling is done using only 7bit instead of 14bit coefficients.
536 * - You can use no more than 5 input pixels to produce 4 output
537 * pixels. Therefore, this filter should not be used for downscaling
538 * by more than ~20% in width (because that equals more than 5/4th
539 * downscaling and thus more than 5 pixels input per 4 pixels output).
540 * - In general, bilinear filters create artifacts during downscaling
541 * (even when <20%), because one output pixel will span more than one
542 * input pixel, and thus some pixels will need edges of both neighbor
543 * pixels to interpolate the output pixel. Since you can use at most
544 * two input pixels per output pixel in bilinear scaling, this is
545 * impossible and thus downscaling by any size will create artifacts.
546 * To enable this type of scaling, set SWS_FLAG_FAST_BILINEAR
547 * in SwsContext->flags.
550 void (*hyscale_fast)(struct SwsContext *c,
551 int16_t *dst, int dstWidth,
552 const uint8_t *src, int srcW, int xInc);
553 void (*hcscale_fast)(struct SwsContext *c,
554 int16_t *dst1, int16_t *dst2, int dstWidth,
555 const uint8_t *src1, const uint8_t *src2,
560 * Scale one horizontal line of input data using a filter over the input
561 * lines, to produce one (differently sized) line of output data.
563 * @param dst pointer to destination buffer for horizontally scaled
564 * data. If the number of bits per component of one
565 * destination pixel (SwsContext->dstBpc) is <= 10, data
566 * will be 15bpc in 16bits (int16_t) width. Else (i.e.
567 * SwsContext->dstBpc == 16), data will be 19bpc in
568 * 32bits (int32_t) width.
569 * @param dstW width of destination image
570 * @param src pointer to source data to be scaled. If the number of
571 * bits per component of a source pixel (SwsContext->srcBpc)
572 * is 8, this is 8bpc in 8bits (uint8_t) width. Else
573 * (i.e. SwsContext->dstBpc > 8), this is native depth
574 * in 16bits (uint16_t) width. In other words, for 9-bit
575 * YUV input, this is 9bpc, for 10-bit YUV input, this is
576 * 10bpc, and for 16-bit RGB or YUV, this is 16bpc.
577 * @param filter filter coefficients to be used per output pixel for
578 * scaling. This contains 14bpp filtering coefficients.
579 * Guaranteed to contain dstW * filterSize entries.
580 * @param filterPos position of the first input pixel to be used for
581 * each output pixel during scaling. Guaranteed to
582 * contain dstW entries.
583 * @param filterSize the number of input coefficients to be used (and
584 * thus the number of input pixels to be used) for
585 * creating a single output pixel. Is aligned to 4
586 * (and input coefficients thus padded with zeroes)
587 * to simplify creating SIMD code.
590 void (*hyScale)(struct SwsContext *c, int16_t *dst, int dstW,
591 const uint8_t *src, const int16_t *filter,
592 const int32_t *filterPos, int filterSize);
593 void (*hcScale)(struct SwsContext *c, int16_t *dst, int dstW,
594 const uint8_t *src, const int16_t *filter,
595 const int32_t *filterPos, int filterSize);
598 /// Color range conversion function for luma plane if needed.
599 void (*lumConvertRange)(int16_t *dst, int width);
600 /// Color range conversion function for chroma planes if needed.
601 void (*chrConvertRange)(int16_t *dst1, int16_t *dst2, int width);
603 int needs_hcscale; ///< Set if there are chroma planes to be converted.
607 //FIXME check init (where 0)
609 SwsFunc ff_yuv2rgb_get_func_ptr(SwsContext *c);
610 int ff_yuv2rgb_c_init_tables(SwsContext *c, const int inv_table[4],
611 int fullRange, int brightness,
612 int contrast, int saturation);
614 void ff_yuv2rgb_init_tables_altivec(SwsContext *c, const int inv_table[4],
615 int brightness, int contrast, int saturation);
616 void updateMMXDitherTables(SwsContext *c, int dstY, int lumBufIndex, int chrBufIndex,
617 int lastInLumBuf, int lastInChrBuf);
619 SwsFunc ff_yuv2rgb_init_mmx(SwsContext *c);
620 SwsFunc ff_yuv2rgb_init_vis(SwsContext *c);
621 SwsFunc ff_yuv2rgb_init_altivec(SwsContext *c);
622 SwsFunc ff_yuv2rgb_get_func_ptr_bfin(SwsContext *c);
623 void ff_bfin_get_unscaled_swscale(SwsContext *c);
625 #if FF_API_SWS_FORMAT_NAME
627 * @deprecated Use av_get_pix_fmt_name() instead.
630 const char *sws_format_name(enum AVPixelFormat format);
633 static av_always_inline int is16BPS(enum AVPixelFormat pix_fmt)
635 const AVPixFmtDescriptor *desc = av_pix_fmt_desc_get(pix_fmt);
637 return desc->comp[0].depth_minus1 == 15;
640 static av_always_inline int is9_OR_10BPS(enum AVPixelFormat pix_fmt)
642 const AVPixFmtDescriptor *desc = av_pix_fmt_desc_get(pix_fmt);
644 return desc->comp[0].depth_minus1 >= 8 && desc->comp[0].depth_minus1 <= 13;
647 #define isNBPS(x) is9_OR_10BPS(x)
649 static av_always_inline int isBE(enum AVPixelFormat pix_fmt)
651 const AVPixFmtDescriptor *desc = av_pix_fmt_desc_get(pix_fmt);
653 return desc->flags & AV_PIX_FMT_FLAG_BE;
656 static av_always_inline int isYUV(enum AVPixelFormat pix_fmt)
658 const AVPixFmtDescriptor *desc = av_pix_fmt_desc_get(pix_fmt);
660 return !(desc->flags & AV_PIX_FMT_FLAG_RGB) && desc->nb_components >= 2;
663 static av_always_inline int isPlanarYUV(enum AVPixelFormat pix_fmt)
665 const AVPixFmtDescriptor *desc = av_pix_fmt_desc_get(pix_fmt);
667 return ((desc->flags & AV_PIX_FMT_FLAG_PLANAR) && isYUV(pix_fmt));
670 static av_always_inline int isRGB(enum AVPixelFormat pix_fmt)
672 const AVPixFmtDescriptor *desc = av_pix_fmt_desc_get(pix_fmt);
674 return (desc->flags & AV_PIX_FMT_FLAG_RGB);
679 (!(av_pix_fmt_desc_get(x)->flags & AV_PIX_FMT_FLAG_PAL) && \
680 av_pix_fmt_desc_get(x)->nb_components <= 2)
683 ((x) == AV_PIX_FMT_GRAY8 || \
684 (x) == AV_PIX_FMT_Y400A || \
685 (x) == AV_PIX_FMT_GRAY16BE || \
686 (x) == AV_PIX_FMT_GRAY16LE)
689 #define isRGBinInt(x) \
691 (x) == AV_PIX_FMT_RGB48BE || \
692 (x) == AV_PIX_FMT_RGB48LE || \
693 (x) == AV_PIX_FMT_RGBA64BE || \
694 (x) == AV_PIX_FMT_RGBA64LE || \
695 (x) == AV_PIX_FMT_RGB32 || \
696 (x) == AV_PIX_FMT_RGB32_1 || \
697 (x) == AV_PIX_FMT_RGB24 || \
698 (x) == AV_PIX_FMT_RGB565BE || \
699 (x) == AV_PIX_FMT_RGB565LE || \
700 (x) == AV_PIX_FMT_RGB555BE || \
701 (x) == AV_PIX_FMT_RGB555LE || \
702 (x) == AV_PIX_FMT_RGB444BE || \
703 (x) == AV_PIX_FMT_RGB444LE || \
704 (x) == AV_PIX_FMT_RGB8 || \
705 (x) == AV_PIX_FMT_RGB4 || \
706 (x) == AV_PIX_FMT_RGB4_BYTE || \
707 (x) == AV_PIX_FMT_MONOBLACK || \
708 (x) == AV_PIX_FMT_MONOWHITE \
710 #define isBGRinInt(x) \
712 (x) == AV_PIX_FMT_BGR48BE || \
713 (x) == AV_PIX_FMT_BGR48LE || \
714 (x) == AV_PIX_FMT_BGRA64BE || \
715 (x) == AV_PIX_FMT_BGRA64LE || \
716 (x) == AV_PIX_FMT_BGR32 || \
717 (x) == AV_PIX_FMT_BGR32_1 || \
718 (x) == AV_PIX_FMT_BGR24 || \
719 (x) == AV_PIX_FMT_BGR565BE || \
720 (x) == AV_PIX_FMT_BGR565LE || \
721 (x) == AV_PIX_FMT_BGR555BE || \
722 (x) == AV_PIX_FMT_BGR555LE || \
723 (x) == AV_PIX_FMT_BGR444BE || \
724 (x) == AV_PIX_FMT_BGR444LE || \
725 (x) == AV_PIX_FMT_BGR8 || \
726 (x) == AV_PIX_FMT_BGR4 || \
727 (x) == AV_PIX_FMT_BGR4_BYTE || \
728 (x) == AV_PIX_FMT_MONOBLACK || \
729 (x) == AV_PIX_FMT_MONOWHITE \
732 #define isRGBinBytes(x) ( \
733 (x) == AV_PIX_FMT_RGB48BE \
734 || (x) == AV_PIX_FMT_RGB48LE \
735 || (x) == AV_PIX_FMT_RGBA64BE \
736 || (x) == AV_PIX_FMT_RGBA64LE \
737 || (x) == AV_PIX_FMT_RGBA \
738 || (x) == AV_PIX_FMT_ARGB \
739 || (x) == AV_PIX_FMT_RGB24 \
741 #define isBGRinBytes(x) ( \
742 (x) == AV_PIX_FMT_BGR48BE \
743 || (x) == AV_PIX_FMT_BGR48LE \
744 || (x) == AV_PIX_FMT_BGRA64BE \
745 || (x) == AV_PIX_FMT_BGRA64LE \
746 || (x) == AV_PIX_FMT_BGRA \
747 || (x) == AV_PIX_FMT_ABGR \
748 || (x) == AV_PIX_FMT_BGR24 \
751 #define isAnyRGB(x) \
758 static av_always_inline int isALPHA(enum AVPixelFormat pix_fmt)
760 const AVPixFmtDescriptor *desc = av_pix_fmt_desc_get(pix_fmt);
762 if (pix_fmt == AV_PIX_FMT_PAL8)
764 return desc->flags & AV_PIX_FMT_FLAG_ALPHA;
768 #define isPacked(x) ( \
769 (x)==AV_PIX_FMT_PAL8 \
770 || (x)==AV_PIX_FMT_YUYV422 \
771 || (x)==AV_PIX_FMT_UYVY422 \
772 || (x)==AV_PIX_FMT_Y400A \
777 static av_always_inline int isPacked(enum AVPixelFormat pix_fmt)
779 const AVPixFmtDescriptor *desc = av_pix_fmt_desc_get(pix_fmt);
781 return ((desc->nb_components >= 2 && !(desc->flags & AV_PIX_FMT_FLAG_PLANAR)) ||
782 pix_fmt == AV_PIX_FMT_PAL8);
786 static av_always_inline int isPlanar(enum AVPixelFormat pix_fmt)
788 const AVPixFmtDescriptor *desc = av_pix_fmt_desc_get(pix_fmt);
790 return (desc->nb_components >= 2 && (desc->flags & AV_PIX_FMT_FLAG_PLANAR));
793 static av_always_inline int isPackedRGB(enum AVPixelFormat pix_fmt)
795 const AVPixFmtDescriptor *desc = av_pix_fmt_desc_get(pix_fmt);
797 return ((desc->flags & (AV_PIX_FMT_FLAG_PLANAR | AV_PIX_FMT_FLAG_RGB)) == AV_PIX_FMT_FLAG_RGB);
800 static av_always_inline int isPlanarRGB(enum AVPixelFormat pix_fmt)
802 const AVPixFmtDescriptor *desc = av_pix_fmt_desc_get(pix_fmt);
804 return ((desc->flags & (AV_PIX_FMT_FLAG_PLANAR | AV_PIX_FMT_FLAG_RGB)) ==
805 (AV_PIX_FMT_FLAG_PLANAR | AV_PIX_FMT_FLAG_RGB));
808 static av_always_inline int usePal(enum AVPixelFormat pix_fmt)
810 const AVPixFmtDescriptor *desc = av_pix_fmt_desc_get(pix_fmt);
812 return (desc->flags & AV_PIX_FMT_FLAG_PAL) || (desc->flags & AV_PIX_FMT_FLAG_PSEUDOPAL);
815 extern const uint64_t ff_dither4[2];
816 extern const uint64_t ff_dither8[2];
818 extern const uint8_t ff_dither_2x2_4[3][8];
819 extern const uint8_t ff_dither_2x2_8[3][8];
820 extern const uint8_t ff_dither_4x4_16[5][8];
821 extern const uint8_t ff_dither_8x8_32[9][8];
822 extern const uint8_t ff_dither_8x8_73[9][8];
823 extern const uint8_t ff_dither_8x8_128[9][8];
824 extern const uint8_t ff_dither_8x8_220[9][8];
826 extern const int32_t ff_yuv2rgb_coeffs[8][4];
828 extern const AVClass sws_context_class;
831 * Set c->swScale to an unscaled converter if one exists for the specific
832 * source and destination formats, bit depths, flags, etc.
834 void ff_get_unscaled_swscale(SwsContext *c);
836 void ff_swscale_get_unscaled_altivec(SwsContext *c);
839 * Return function pointer to fastest main scaler path function depending
840 * on architecture and available optimizations.
842 SwsFunc ff_getSwsFunc(SwsContext *c);
844 void ff_sws_init_input_funcs(SwsContext *c);
845 void ff_sws_init_output_funcs(SwsContext *c,
846 yuv2planar1_fn *yuv2plane1,
847 yuv2planarX_fn *yuv2planeX,
848 yuv2interleavedX_fn *yuv2nv12cX,
849 yuv2packed1_fn *yuv2packed1,
850 yuv2packed2_fn *yuv2packed2,
851 yuv2packedX_fn *yuv2packedX,
852 yuv2anyX_fn *yuv2anyX);
853 void ff_sws_init_swScale_altivec(SwsContext *c);
854 void ff_sws_init_swScale_mmx(SwsContext *c);
856 static inline void fillPlane16(uint8_t *plane, int stride, int width, int height, int y,
857 int alpha, int bits, const int big_endian)
860 uint8_t *ptr = plane + stride * y;
861 int v = alpha ? 0xFFFF>>(15-bits) : (1<<bits);
862 for (i = 0; i < height; i++) {
863 #define FILL(wfunc) \
864 for (j = 0; j < width; j++) {\
876 #endif /* SWSCALE_SWSCALE_INTERNAL_H */