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
32 #include "libavutil/avassert.h"
33 #include "libavutil/avutil.h"
34 #include "libavutil/common.h"
35 #include "libavutil/intreadwrite.h"
36 #include "libavutil/log.h"
37 #include "libavutil/pixfmt.h"
38 #include "libavutil/pixdesc.h"
40 #define STR(s) AV_TOSTRING(s) // AV_STRINGIFY is too long
42 #define YUVRGB_TABLE_HEADROOM 256
44 #define MAX_FILTER_SIZE SWS_MAX_FILTER_SIZE
49 #define ALT32_CORR (-1)
64 #define RETCODE_USE_CASCADE -12345
68 typedef enum SwsDither {
78 typedef enum SwsAlphaBlend {
79 SWS_ALPHA_BLEND_NONE = 0,
80 SWS_ALPHA_BLEND_UNIFORM,
81 SWS_ALPHA_BLEND_CHECKERBOARD,
85 typedef int (*SwsFunc)(struct SwsContext *context, const uint8_t *src[],
86 int srcStride[], int srcSliceY, int srcSliceH,
87 uint8_t *dst[], int dstStride[]);
90 * Write one line of horizontally scaled data to planar output
91 * without any additional vertical scaling (or point-scaling).
93 * @param src scaled source data, 15bit for 8-10bit output,
94 * 19-bit for 16bit output (in int32_t)
95 * @param dest pointer to the output plane. For >8bit
96 * output, this is in uint16_t
97 * @param dstW width of destination in pixels
98 * @param dither ordered dither array of type int16_t and size 8
99 * @param offset Dither offset
101 typedef void (*yuv2planar1_fn)(const int16_t *src, uint8_t *dest, int dstW,
102 const uint8_t *dither, int offset);
105 * Write one line of horizontally scaled data to planar output
106 * with multi-point vertical scaling between input pixels.
108 * @param filter vertical luma/alpha scaling coefficients, 12bit [0,4096]
109 * @param src scaled luma (Y) or alpha (A) source data, 15bit for 8-10bit output,
110 * 19-bit for 16bit output (in int32_t)
111 * @param filterSize number of vertical input lines to scale
112 * @param dest pointer to output plane. For >8bit
113 * output, this is in uint16_t
114 * @param dstW width of destination pixels
115 * @param offset Dither offset
117 typedef void (*yuv2planarX_fn)(const int16_t *filter, int filterSize,
118 const int16_t **src, uint8_t *dest, int dstW,
119 const uint8_t *dither, int offset);
122 * Write one line of horizontally scaled chroma to interleaved output
123 * with multi-point vertical scaling between input pixels.
125 * @param c SWS scaling context
126 * @param chrFilter vertical chroma scaling coefficients, 12bit [0,4096]
127 * @param chrUSrc scaled chroma (U) source data, 15bit for 8-10bit output,
128 * 19-bit for 16bit output (in int32_t)
129 * @param chrVSrc scaled chroma (V) source data, 15bit for 8-10bit output,
130 * 19-bit for 16bit output (in int32_t)
131 * @param chrFilterSize number of vertical chroma input lines to scale
132 * @param dest pointer to the output plane. For >8bit
133 * output, this is in uint16_t
134 * @param dstW width of chroma planes
136 typedef void (*yuv2interleavedX_fn)(struct SwsContext *c,
137 const int16_t *chrFilter,
139 const int16_t **chrUSrc,
140 const int16_t **chrVSrc,
141 uint8_t *dest, int dstW);
144 * Write one line of horizontally scaled Y/U/V/A to packed-pixel YUV/RGB
145 * output without any additional vertical scaling (or point-scaling). Note
146 * that this function may do chroma scaling, see the "uvalpha" argument.
148 * @param c SWS scaling context
149 * @param lumSrc scaled luma (Y) source data, 15bit for 8-10bit output,
150 * 19-bit for 16bit output (in int32_t)
151 * @param chrUSrc scaled chroma (U) source data, 15bit for 8-10bit output,
152 * 19-bit for 16bit output (in int32_t)
153 * @param chrVSrc scaled chroma (V) source data, 15bit for 8-10bit output,
154 * 19-bit for 16bit output (in int32_t)
155 * @param alpSrc scaled alpha (A) source data, 15bit for 8-10bit output,
156 * 19-bit for 16bit output (in int32_t)
157 * @param dest pointer to the output plane. For 16bit output, this is
159 * @param dstW width of lumSrc and alpSrc in pixels, number of pixels
160 * to write into dest[]
161 * @param uvalpha chroma scaling coefficient for the second line of chroma
162 * pixels, either 2048 or 0. If 0, one chroma input is used
163 * for 2 output pixels (or if the SWS_FLAG_FULL_CHR_INT flag
164 * is set, it generates 1 output pixel). If 2048, two chroma
165 * input pixels should be averaged for 2 output pixels (this
166 * only happens if SWS_FLAG_FULL_CHR_INT is not set)
167 * @param y vertical line number for this output. This does not need
168 * to be used to calculate the offset in the destination,
169 * but can be used to generate comfort noise using dithering
170 * for some output formats.
172 typedef void (*yuv2packed1_fn)(struct SwsContext *c, const int16_t *lumSrc,
173 const int16_t *chrUSrc[2],
174 const int16_t *chrVSrc[2],
175 const int16_t *alpSrc, uint8_t *dest,
176 int dstW, int uvalpha, int y);
178 * Write one line of horizontally scaled Y/U/V/A to packed-pixel YUV/RGB
179 * output by doing bilinear scaling between two input lines.
181 * @param c SWS scaling context
182 * @param lumSrc scaled luma (Y) source data, 15bit for 8-10bit output,
183 * 19-bit for 16bit output (in int32_t)
184 * @param chrUSrc scaled chroma (U) source data, 15bit for 8-10bit output,
185 * 19-bit for 16bit output (in int32_t)
186 * @param chrVSrc scaled chroma (V) source data, 15bit for 8-10bit output,
187 * 19-bit for 16bit output (in int32_t)
188 * @param alpSrc scaled alpha (A) source data, 15bit for 8-10bit output,
189 * 19-bit for 16bit output (in int32_t)
190 * @param dest pointer to the output plane. For 16bit output, this is
192 * @param dstW width of lumSrc and alpSrc in pixels, number of pixels
193 * to write into dest[]
194 * @param yalpha luma/alpha scaling coefficients for the second input line.
195 * The first line's coefficients can be calculated by using
197 * @param uvalpha chroma scaling coefficient for the second input line. The
198 * first line's coefficients can be calculated by using
200 * @param y vertical line number for this output. This does not need
201 * to be used to calculate the offset in the destination,
202 * but can be used to generate comfort noise using dithering
203 * for some output formats.
205 typedef void (*yuv2packed2_fn)(struct SwsContext *c, const int16_t *lumSrc[2],
206 const int16_t *chrUSrc[2],
207 const int16_t *chrVSrc[2],
208 const int16_t *alpSrc[2],
210 int dstW, int yalpha, int uvalpha, int y);
212 * Write one line of horizontally scaled Y/U/V/A to packed-pixel YUV/RGB
213 * output by doing multi-point vertical scaling between input pixels.
215 * @param c SWS scaling context
216 * @param lumFilter vertical luma/alpha scaling coefficients, 12bit [0,4096]
217 * @param lumSrc scaled luma (Y) source data, 15bit for 8-10bit output,
218 * 19-bit for 16bit output (in int32_t)
219 * @param lumFilterSize number of vertical luma/alpha input lines to scale
220 * @param chrFilter vertical chroma scaling coefficients, 12bit [0,4096]
221 * @param chrUSrc scaled chroma (U) source data, 15bit for 8-10bit output,
222 * 19-bit for 16bit output (in int32_t)
223 * @param chrVSrc scaled chroma (V) source data, 15bit for 8-10bit output,
224 * 19-bit for 16bit output (in int32_t)
225 * @param chrFilterSize number of vertical chroma input lines to scale
226 * @param alpSrc scaled alpha (A) source data, 15bit for 8-10bit output,
227 * 19-bit for 16bit output (in int32_t)
228 * @param dest pointer to the output plane. For 16bit output, this is
230 * @param dstW width of lumSrc and alpSrc in pixels, number of pixels
231 * to write into dest[]
232 * @param y vertical line number for this output. This does not need
233 * to be used to calculate the offset in the destination,
234 * but can be used to generate comfort noise using dithering
235 * or some output formats.
237 typedef void (*yuv2packedX_fn)(struct SwsContext *c, const int16_t *lumFilter,
238 const int16_t **lumSrc, int lumFilterSize,
239 const int16_t *chrFilter,
240 const int16_t **chrUSrc,
241 const int16_t **chrVSrc, int chrFilterSize,
242 const int16_t **alpSrc, uint8_t *dest,
246 * Write one line of horizontally scaled Y/U/V/A to YUV/RGB
247 * output by doing multi-point vertical scaling between input pixels.
249 * @param c SWS scaling context
250 * @param lumFilter vertical luma/alpha scaling coefficients, 12bit [0,4096]
251 * @param lumSrc scaled luma (Y) source data, 15bit for 8-10bit output,
252 * 19-bit for 16bit output (in int32_t)
253 * @param lumFilterSize number of vertical luma/alpha input lines to scale
254 * @param chrFilter vertical chroma scaling coefficients, 12bit [0,4096]
255 * @param chrUSrc scaled chroma (U) source data, 15bit for 8-10bit output,
256 * 19-bit for 16bit output (in int32_t)
257 * @param chrVSrc scaled chroma (V) source data, 15bit for 8-10bit output,
258 * 19-bit for 16bit output (in int32_t)
259 * @param chrFilterSize number of vertical chroma input lines to scale
260 * @param alpSrc scaled alpha (A) source data, 15bit for 8-10bit output,
261 * 19-bit for 16bit output (in int32_t)
262 * @param dest pointer to the output planes. For 16bit output, this is
264 * @param dstW width of lumSrc and alpSrc in pixels, number of pixels
265 * to write into dest[]
266 * @param y vertical line number for this output. This does not need
267 * to be used to calculate the offset in the destination,
268 * but can be used to generate comfort noise using dithering
269 * or some output formats.
271 typedef void (*yuv2anyX_fn)(struct SwsContext *c, const int16_t *lumFilter,
272 const int16_t **lumSrc, int lumFilterSize,
273 const int16_t *chrFilter,
274 const int16_t **chrUSrc,
275 const int16_t **chrVSrc, int chrFilterSize,
276 const int16_t **alpSrc, uint8_t **dest,
280 struct SwsFilterDescriptor;
282 /* This struct should be aligned on at least a 32-byte boundary. */
283 typedef struct SwsContext {
285 * info on struct for av_log
287 const AVClass *av_class;
290 * Note that src, dst, srcStride, dstStride will be copied in the
291 * sws_scale() wrapper so they can be freely modified here.
294 int srcW; ///< Width of source luma/alpha planes.
295 int srcH; ///< Height of source luma/alpha planes.
296 int dstH; ///< Height of destination luma/alpha planes.
297 int chrSrcW; ///< Width of source chroma planes.
298 int chrSrcH; ///< Height of source chroma planes.
299 int chrDstW; ///< Width of destination chroma planes.
300 int chrDstH; ///< Height of destination chroma planes.
301 int lumXInc, chrXInc;
302 int lumYInc, chrYInc;
303 enum AVPixelFormat dstFormat; ///< Destination pixel format.
304 enum AVPixelFormat srcFormat; ///< Source pixel format.
305 int dstFormatBpp; ///< Number of bits per pixel of the destination pixel format.
306 int srcFormatBpp; ///< Number of bits per pixel of the source pixel format.
308 int chrSrcHSubSample; ///< Binary logarithm of horizontal subsampling factor between luma/alpha and chroma planes in source image.
309 int chrSrcVSubSample; ///< Binary logarithm of vertical subsampling factor between luma/alpha and chroma planes in source image.
310 int chrDstHSubSample; ///< Binary logarithm of horizontal subsampling factor between luma/alpha and chroma planes in destination image.
311 int chrDstVSubSample; ///< Binary logarithm of vertical subsampling factor between luma/alpha and chroma planes in destination image.
312 int vChrDrop; ///< Binary logarithm of extra vertical subsampling factor in source image chroma planes specified by user.
313 int sliceDir; ///< Direction that slices are fed to the scaler (1 = top-to-bottom, -1 = bottom-to-top).
314 double param[2]; ///< Input parameters for scaling algorithms that need them.
316 /* The cascaded_* fields allow spliting a scaler task into multiple
317 * sequential steps, this is for example used to limit the maximum
318 * downscaling factor that needs to be supported in one scaler.
320 struct SwsContext *cascaded_context[3];
321 int cascaded_tmpStride[4];
322 uint8_t *cascaded_tmp[4];
323 int cascaded1_tmpStride[4];
324 uint8_t *cascaded1_tmp[4];
325 int cascaded_mainindex;
329 int is_internal_gamma;
336 struct SwsSlice *slice;
337 struct SwsFilterDescriptor *desc;
339 uint32_t pal_yuv[256];
340 uint32_t pal_rgb[256];
343 * @name Scaled horizontal lines ring buffer.
344 * The horizontal scaler keeps just enough scaled lines in a ring buffer
345 * so they may be passed to the vertical scaler. The pointers to the
346 * allocated buffers for each line are duplicated in sequence in the ring
347 * buffer to simplify indexing and avoid wrapping around between lines
348 * inside the vertical scaler code. The wrapping is done before the
349 * vertical scaler is called.
352 int16_t **lumPixBuf; ///< Ring buffer for scaled horizontal luma plane lines to be fed to the vertical scaler.
353 int16_t **chrUPixBuf; ///< Ring buffer for scaled horizontal chroma plane lines to be fed to the vertical scaler.
354 int16_t **chrVPixBuf; ///< Ring buffer for scaled horizontal chroma plane lines to be fed to the vertical scaler.
355 int16_t **alpPixBuf; ///< Ring buffer for scaled horizontal alpha plane lines to be fed to the vertical scaler.
356 int vLumBufSize; ///< Number of vertical luma/alpha lines allocated in the ring buffer.
357 int vChrBufSize; ///< Number of vertical chroma lines allocated in the ring buffer.
358 int lastInLumBuf; ///< Last scaled horizontal luma/alpha line from source in the ring buffer.
359 int lastInChrBuf; ///< Last scaled horizontal chroma line from source in the ring buffer.
360 int lumBufIndex; ///< Index in ring buffer of the last scaled horizontal luma/alpha line from source.
361 int chrBufIndex; ///< Index in ring buffer of the last scaled horizontal chroma line from source.
364 uint8_t *formatConvBuffer;
367 * @name Horizontal and vertical filters.
368 * To better understand the following fields, here is a pseudo-code of
369 * their usage in filtering a horizontal line:
371 * for (i = 0; i < width; i++) {
373 * for (j = 0; j < filterSize; j++)
374 * dst[i] += src[ filterPos[i] + j ] * filter[ filterSize * i + j ];
375 * dst[i] >>= FRAC_BITS; // The actual implementation is fixed-point.
380 int16_t *hLumFilter; ///< Array of horizontal filter coefficients for luma/alpha planes.
381 int16_t *hChrFilter; ///< Array of horizontal filter coefficients for chroma planes.
382 int16_t *vLumFilter; ///< Array of vertical filter coefficients for luma/alpha planes.
383 int16_t *vChrFilter; ///< Array of vertical filter coefficients for chroma planes.
384 int32_t *hLumFilterPos; ///< Array of horizontal filter starting positions for each dst[i] for luma/alpha planes.
385 int32_t *hChrFilterPos; ///< Array of horizontal filter starting positions for each dst[i] for chroma planes.
386 int32_t *vLumFilterPos; ///< Array of vertical filter starting positions for each dst[i] for luma/alpha planes.
387 int32_t *vChrFilterPos; ///< Array of vertical filter starting positions for each dst[i] for chroma planes.
388 int hLumFilterSize; ///< Horizontal filter size for luma/alpha pixels.
389 int hChrFilterSize; ///< Horizontal filter size for chroma pixels.
390 int vLumFilterSize; ///< Vertical filter size for luma/alpha pixels.
391 int vChrFilterSize; ///< Vertical filter size for chroma pixels.
394 int lumMmxextFilterCodeSize; ///< Runtime-generated MMXEXT horizontal fast bilinear scaler code size for luma/alpha planes.
395 int chrMmxextFilterCodeSize; ///< Runtime-generated MMXEXT horizontal fast bilinear scaler code size for chroma planes.
396 uint8_t *lumMmxextFilterCode; ///< Runtime-generated MMXEXT horizontal fast bilinear scaler code for luma/alpha planes.
397 uint8_t *chrMmxextFilterCode; ///< Runtime-generated MMXEXT horizontal fast bilinear scaler code for chroma planes.
401 int dstY; ///< Last destination vertical line output from last slice.
402 int flags; ///< Flags passed by the user to select scaler algorithm, optimizations, subsampling, etc...
403 void *yuvTable; // pointer to the yuv->rgb table start so it can be freed()
404 // alignment ensures the offset can be added in a single
405 // instruction on e.g. ARM
406 DECLARE_ALIGNED(16, int, table_gV)[256 + 2*YUVRGB_TABLE_HEADROOM];
407 uint8_t *table_rV[256 + 2*YUVRGB_TABLE_HEADROOM];
408 uint8_t *table_gU[256 + 2*YUVRGB_TABLE_HEADROOM];
409 uint8_t *table_bU[256 + 2*YUVRGB_TABLE_HEADROOM];
410 DECLARE_ALIGNED(16, int32_t, input_rgb2yuv_table)[16+40*4]; // This table can contain both C and SIMD formatted values, the C vales are always at the XY_IDX points
420 #define RGB2YUV_SHIFT 15
422 int *dither_error[4];
425 int contrast, brightness, saturation; // for sws_getColorspaceDetails
426 int srcColorspaceTable[4];
427 int dstColorspaceTable[4];
428 int srcRange; ///< 0 = MPG YUV range, 1 = JPG YUV range (source image).
429 int dstRange; ///< 0 = MPG YUV range, 1 = JPG YUV range (destination image).
438 int yuv2rgb_y_offset;
440 int yuv2rgb_v2r_coeff;
441 int yuv2rgb_v2g_coeff;
442 int yuv2rgb_u2g_coeff;
443 int yuv2rgb_u2b_coeff;
445 #define RED_DITHER "0*8"
446 #define GREEN_DITHER "1*8"
447 #define BLUE_DITHER "2*8"
448 #define Y_COEFF "3*8"
449 #define VR_COEFF "4*8"
450 #define UB_COEFF "5*8"
451 #define VG_COEFF "6*8"
452 #define UG_COEFF "7*8"
453 #define Y_OFFSET "8*8"
454 #define U_OFFSET "9*8"
455 #define V_OFFSET "10*8"
456 #define LUM_MMX_FILTER_OFFSET "11*8"
457 #define CHR_MMX_FILTER_OFFSET "11*8+4*4*"AV_STRINGIFY(MAX_FILTER_SIZE)
458 #define DSTW_OFFSET "11*8+4*4*"AV_STRINGIFY(MAX_FILTER_SIZE)"*2"
459 #define ESP_OFFSET "11*8+4*4*"AV_STRINGIFY(MAX_FILTER_SIZE)"*2+8"
460 #define VROUNDER_OFFSET "11*8+4*4*"AV_STRINGIFY(MAX_FILTER_SIZE)"*2+16"
461 #define U_TEMP "11*8+4*4*"AV_STRINGIFY(MAX_FILTER_SIZE)"*2+24"
462 #define V_TEMP "11*8+4*4*"AV_STRINGIFY(MAX_FILTER_SIZE)"*2+32"
463 #define Y_TEMP "11*8+4*4*"AV_STRINGIFY(MAX_FILTER_SIZE)"*2+40"
464 #define ALP_MMX_FILTER_OFFSET "11*8+4*4*"AV_STRINGIFY(MAX_FILTER_SIZE)"*2+48"
465 #define UV_OFF_PX "11*8+4*4*"AV_STRINGIFY(MAX_FILTER_SIZE)"*3+48"
466 #define UV_OFF_BYTE "11*8+4*4*"AV_STRINGIFY(MAX_FILTER_SIZE)"*3+56"
467 #define DITHER16 "11*8+4*4*"AV_STRINGIFY(MAX_FILTER_SIZE)"*3+64"
468 #define DITHER32 "11*8+4*4*"AV_STRINGIFY(MAX_FILTER_SIZE)"*3+80"
469 #define DITHER32_INT (11*8+4*4*MAX_FILTER_SIZE*3+80) // value equal to above, used for checking that the struct hasn't been changed by mistake
471 DECLARE_ALIGNED(8, uint64_t, redDither);
472 DECLARE_ALIGNED(8, uint64_t, greenDither);
473 DECLARE_ALIGNED(8, uint64_t, blueDither);
475 DECLARE_ALIGNED(8, uint64_t, yCoeff);
476 DECLARE_ALIGNED(8, uint64_t, vrCoeff);
477 DECLARE_ALIGNED(8, uint64_t, ubCoeff);
478 DECLARE_ALIGNED(8, uint64_t, vgCoeff);
479 DECLARE_ALIGNED(8, uint64_t, ugCoeff);
480 DECLARE_ALIGNED(8, uint64_t, yOffset);
481 DECLARE_ALIGNED(8, uint64_t, uOffset);
482 DECLARE_ALIGNED(8, uint64_t, vOffset);
483 int32_t lumMmxFilter[4 * MAX_FILTER_SIZE];
484 int32_t chrMmxFilter[4 * MAX_FILTER_SIZE];
485 int dstW; ///< Width of destination luma/alpha planes.
486 DECLARE_ALIGNED(8, uint64_t, esp);
487 DECLARE_ALIGNED(8, uint64_t, vRounder);
488 DECLARE_ALIGNED(8, uint64_t, u_temp);
489 DECLARE_ALIGNED(8, uint64_t, v_temp);
490 DECLARE_ALIGNED(8, uint64_t, y_temp);
491 int32_t alpMmxFilter[4 * MAX_FILTER_SIZE];
492 // alignment of these values is not necessary, but merely here
493 // to maintain the same offset across x8632 and x86-64. Once we
494 // use proper offset macros in the asm, they can be removed.
495 DECLARE_ALIGNED(8, ptrdiff_t, uv_off); ///< offset (in pixels) between u and v planes
496 DECLARE_ALIGNED(8, ptrdiff_t, uv_offx2); ///< offset (in bytes) between u and v planes
497 DECLARE_ALIGNED(8, uint16_t, dither16)[8];
498 DECLARE_ALIGNED(8, uint32_t, dither32)[8];
500 const uint8_t *chrDither8, *lumDither8;
503 vector signed short CY;
504 vector signed short CRV;
505 vector signed short CBU;
506 vector signed short CGU;
507 vector signed short CGV;
508 vector signed short OY;
509 vector unsigned short CSHIFT;
510 vector signed short *vYCoeffsBank, *vCCoeffsBank;
515 /* pre defined color-spaces gamma */
516 #define XYZ_GAMMA (2.6f)
517 #define RGB_GAMMA (2.2f)
520 int16_t *xyzgammainv;
521 int16_t *rgbgammainv;
522 int16_t xyz2rgb_matrix[3][4];
523 int16_t rgb2xyz_matrix[3][4];
525 /* function pointers for swscale() */
526 yuv2planar1_fn yuv2plane1;
527 yuv2planarX_fn yuv2planeX;
528 yuv2interleavedX_fn yuv2nv12cX;
529 yuv2packed1_fn yuv2packed1;
530 yuv2packed2_fn yuv2packed2;
531 yuv2packedX_fn yuv2packedX;
532 yuv2anyX_fn yuv2anyX;
534 /// Unscaled conversion of luma plane to YV12 for horizontal scaler.
535 void (*lumToYV12)(uint8_t *dst, const uint8_t *src, const uint8_t *src2, const uint8_t *src3,
536 int width, uint32_t *pal);
537 /// Unscaled conversion of alpha plane to YV12 for horizontal scaler.
538 void (*alpToYV12)(uint8_t *dst, const uint8_t *src, const uint8_t *src2, const uint8_t *src3,
539 int width, uint32_t *pal);
540 /// Unscaled conversion of chroma planes to YV12 for horizontal scaler.
541 void (*chrToYV12)(uint8_t *dstU, uint8_t *dstV,
542 const uint8_t *src1, const uint8_t *src2, const uint8_t *src3,
543 int width, uint32_t *pal);
546 * Functions to read planar input, such as planar RGB, and convert
547 * internally to Y/UV/A.
550 void (*readLumPlanar)(uint8_t *dst, const uint8_t *src[4], int width, int32_t *rgb2yuv);
551 void (*readChrPlanar)(uint8_t *dstU, uint8_t *dstV, const uint8_t *src[4],
552 int width, int32_t *rgb2yuv);
553 void (*readAlpPlanar)(uint8_t *dst, const uint8_t *src[4], int width, int32_t *rgb2yuv);
557 * Scale one horizontal line of input data using a bilinear filter
558 * to produce one line of output data. Compared to SwsContext->hScale(),
559 * please take note of the following caveats when using these:
560 * - Scaling is done using only 7bit instead of 14bit coefficients.
561 * - You can use no more than 5 input pixels to produce 4 output
562 * pixels. Therefore, this filter should not be used for downscaling
563 * by more than ~20% in width (because that equals more than 5/4th
564 * downscaling and thus more than 5 pixels input per 4 pixels output).
565 * - In general, bilinear filters create artifacts during downscaling
566 * (even when <20%), because one output pixel will span more than one
567 * input pixel, and thus some pixels will need edges of both neighbor
568 * pixels to interpolate the output pixel. Since you can use at most
569 * two input pixels per output pixel in bilinear scaling, this is
570 * impossible and thus downscaling by any size will create artifacts.
571 * To enable this type of scaling, set SWS_FLAG_FAST_BILINEAR
572 * in SwsContext->flags.
575 void (*hyscale_fast)(struct SwsContext *c,
576 int16_t *dst, int dstWidth,
577 const uint8_t *src, int srcW, int xInc);
578 void (*hcscale_fast)(struct SwsContext *c,
579 int16_t *dst1, int16_t *dst2, int dstWidth,
580 const uint8_t *src1, const uint8_t *src2,
585 * Scale one horizontal line of input data using a filter over the input
586 * lines, to produce one (differently sized) line of output data.
588 * @param dst pointer to destination buffer for horizontally scaled
589 * data. If the number of bits per component of one
590 * destination pixel (SwsContext->dstBpc) is <= 10, data
591 * will be 15bpc in 16bits (int16_t) width. Else (i.e.
592 * SwsContext->dstBpc == 16), data will be 19bpc in
593 * 32bits (int32_t) width.
594 * @param dstW width of destination image
595 * @param src pointer to source data to be scaled. If the number of
596 * bits per component of a source pixel (SwsContext->srcBpc)
597 * is 8, this is 8bpc in 8bits (uint8_t) width. Else
598 * (i.e. SwsContext->dstBpc > 8), this is native depth
599 * in 16bits (uint16_t) width. In other words, for 9-bit
600 * YUV input, this is 9bpc, for 10-bit YUV input, this is
601 * 10bpc, and for 16-bit RGB or YUV, this is 16bpc.
602 * @param filter filter coefficients to be used per output pixel for
603 * scaling. This contains 14bpp filtering coefficients.
604 * Guaranteed to contain dstW * filterSize entries.
605 * @param filterPos position of the first input pixel to be used for
606 * each output pixel during scaling. Guaranteed to
607 * contain dstW entries.
608 * @param filterSize the number of input coefficients to be used (and
609 * thus the number of input pixels to be used) for
610 * creating a single output pixel. Is aligned to 4
611 * (and input coefficients thus padded with zeroes)
612 * to simplify creating SIMD code.
615 void (*hyScale)(struct SwsContext *c, int16_t *dst, int dstW,
616 const uint8_t *src, const int16_t *filter,
617 const int32_t *filterPos, int filterSize);
618 void (*hcScale)(struct SwsContext *c, int16_t *dst, int dstW,
619 const uint8_t *src, const int16_t *filter,
620 const int32_t *filterPos, int filterSize);
623 /// Color range conversion function for luma plane if needed.
624 void (*lumConvertRange)(int16_t *dst, int width);
625 /// Color range conversion function for chroma planes if needed.
626 void (*chrConvertRange)(int16_t *dst1, int16_t *dst2, int width);
628 int needs_hcscale; ///< Set if there are chroma planes to be converted.
632 SwsAlphaBlend alphablend;
634 //FIXME check init (where 0)
636 SwsFunc ff_yuv2rgb_get_func_ptr(SwsContext *c);
637 int ff_yuv2rgb_c_init_tables(SwsContext *c, const int inv_table[4],
638 int fullRange, int brightness,
639 int contrast, int saturation);
640 void ff_yuv2rgb_init_tables_ppc(SwsContext *c, const int inv_table[4],
641 int brightness, int contrast, int saturation);
643 void ff_updateMMXDitherTables(SwsContext *c, int dstY, int lumBufIndex, int chrBufIndex,
644 int lastInLumBuf, int lastInChrBuf);
646 av_cold void ff_sws_init_range_convert(SwsContext *c);
648 SwsFunc ff_yuv2rgb_init_x86(SwsContext *c);
649 SwsFunc ff_yuv2rgb_init_ppc(SwsContext *c);
651 static av_always_inline int is16BPS(enum AVPixelFormat pix_fmt)
653 const AVPixFmtDescriptor *desc = av_pix_fmt_desc_get(pix_fmt);
655 return desc->comp[0].depth == 16;
658 static av_always_inline int is9_OR_10BPS(enum AVPixelFormat pix_fmt)
660 const AVPixFmtDescriptor *desc = av_pix_fmt_desc_get(pix_fmt);
662 return desc->comp[0].depth >= 9 && desc->comp[0].depth <= 14;
665 #define isNBPS(x) is9_OR_10BPS(x)
667 static av_always_inline int isBE(enum AVPixelFormat pix_fmt)
669 const AVPixFmtDescriptor *desc = av_pix_fmt_desc_get(pix_fmt);
671 return desc->flags & AV_PIX_FMT_FLAG_BE;
674 static av_always_inline int isYUV(enum AVPixelFormat pix_fmt)
676 const AVPixFmtDescriptor *desc = av_pix_fmt_desc_get(pix_fmt);
678 return !(desc->flags & AV_PIX_FMT_FLAG_RGB) && desc->nb_components >= 2;
681 static av_always_inline int isPlanarYUV(enum AVPixelFormat pix_fmt)
683 const AVPixFmtDescriptor *desc = av_pix_fmt_desc_get(pix_fmt);
685 return ((desc->flags & AV_PIX_FMT_FLAG_PLANAR) && isYUV(pix_fmt));
688 static av_always_inline int isRGB(enum AVPixelFormat pix_fmt)
690 const AVPixFmtDescriptor *desc = av_pix_fmt_desc_get(pix_fmt);
692 return (desc->flags & AV_PIX_FMT_FLAG_RGB);
697 (!(av_pix_fmt_desc_get(x)->flags & AV_PIX_FMT_FLAG_PAL) && \
698 av_pix_fmt_desc_get(x)->nb_components <= 2)
701 ((x) == AV_PIX_FMT_GRAY8 || \
702 (x) == AV_PIX_FMT_YA8 || \
703 (x) == AV_PIX_FMT_GRAY16BE || \
704 (x) == AV_PIX_FMT_GRAY16LE || \
705 (x) == AV_PIX_FMT_YA16BE || \
706 (x) == AV_PIX_FMT_YA16LE)
709 #define isRGBinInt(x) \
711 (x) == AV_PIX_FMT_RGB48BE || \
712 (x) == AV_PIX_FMT_RGB48LE || \
713 (x) == AV_PIX_FMT_RGB32 || \
714 (x) == AV_PIX_FMT_RGB32_1 || \
715 (x) == AV_PIX_FMT_RGB24 || \
716 (x) == AV_PIX_FMT_RGB565BE || \
717 (x) == AV_PIX_FMT_RGB565LE || \
718 (x) == AV_PIX_FMT_RGB555BE || \
719 (x) == AV_PIX_FMT_RGB555LE || \
720 (x) == AV_PIX_FMT_RGB444BE || \
721 (x) == AV_PIX_FMT_RGB444LE || \
722 (x) == AV_PIX_FMT_RGB8 || \
723 (x) == AV_PIX_FMT_RGB4 || \
724 (x) == AV_PIX_FMT_RGB4_BYTE || \
725 (x) == AV_PIX_FMT_RGBA64BE || \
726 (x) == AV_PIX_FMT_RGBA64LE || \
727 (x) == AV_PIX_FMT_MONOBLACK || \
728 (x) == AV_PIX_FMT_MONOWHITE \
730 #define isBGRinInt(x) \
732 (x) == AV_PIX_FMT_BGR48BE || \
733 (x) == AV_PIX_FMT_BGR48LE || \
734 (x) == AV_PIX_FMT_BGR32 || \
735 (x) == AV_PIX_FMT_BGR32_1 || \
736 (x) == AV_PIX_FMT_BGR24 || \
737 (x) == AV_PIX_FMT_BGR565BE || \
738 (x) == AV_PIX_FMT_BGR565LE || \
739 (x) == AV_PIX_FMT_BGR555BE || \
740 (x) == AV_PIX_FMT_BGR555LE || \
741 (x) == AV_PIX_FMT_BGR444BE || \
742 (x) == AV_PIX_FMT_BGR444LE || \
743 (x) == AV_PIX_FMT_BGR8 || \
744 (x) == AV_PIX_FMT_BGR4 || \
745 (x) == AV_PIX_FMT_BGR4_BYTE || \
746 (x) == AV_PIX_FMT_BGRA64BE || \
747 (x) == AV_PIX_FMT_BGRA64LE || \
748 (x) == AV_PIX_FMT_MONOBLACK || \
749 (x) == AV_PIX_FMT_MONOWHITE \
752 #define isRGBinBytes(x) ( \
753 (x) == AV_PIX_FMT_RGB48BE \
754 || (x) == AV_PIX_FMT_RGB48LE \
755 || (x) == AV_PIX_FMT_RGBA64BE \
756 || (x) == AV_PIX_FMT_RGBA64LE \
757 || (x) == AV_PIX_FMT_RGBA \
758 || (x) == AV_PIX_FMT_ARGB \
759 || (x) == AV_PIX_FMT_RGB24 \
761 #define isBGRinBytes(x) ( \
762 (x) == AV_PIX_FMT_BGR48BE \
763 || (x) == AV_PIX_FMT_BGR48LE \
764 || (x) == AV_PIX_FMT_BGRA64BE \
765 || (x) == AV_PIX_FMT_BGRA64LE \
766 || (x) == AV_PIX_FMT_BGRA \
767 || (x) == AV_PIX_FMT_ABGR \
768 || (x) == AV_PIX_FMT_BGR24 \
771 #define isBayer(x) ( \
772 (x)==AV_PIX_FMT_BAYER_BGGR8 \
773 || (x)==AV_PIX_FMT_BAYER_BGGR16LE \
774 || (x)==AV_PIX_FMT_BAYER_BGGR16BE \
775 || (x)==AV_PIX_FMT_BAYER_RGGB8 \
776 || (x)==AV_PIX_FMT_BAYER_RGGB16LE \
777 || (x)==AV_PIX_FMT_BAYER_RGGB16BE \
778 || (x)==AV_PIX_FMT_BAYER_GBRG8 \
779 || (x)==AV_PIX_FMT_BAYER_GBRG16LE \
780 || (x)==AV_PIX_FMT_BAYER_GBRG16BE \
781 || (x)==AV_PIX_FMT_BAYER_GRBG8 \
782 || (x)==AV_PIX_FMT_BAYER_GRBG16LE \
783 || (x)==AV_PIX_FMT_BAYER_GRBG16BE \
786 #define isAnyRGB(x) \
794 static av_always_inline int isALPHA(enum AVPixelFormat pix_fmt)
796 const AVPixFmtDescriptor *desc = av_pix_fmt_desc_get(pix_fmt);
798 if (pix_fmt == AV_PIX_FMT_PAL8)
800 return desc->flags & AV_PIX_FMT_FLAG_ALPHA;
804 #define isPacked(x) ( \
805 (x)==AV_PIX_FMT_PAL8 \
806 || (x)==AV_PIX_FMT_YUYV422 \
807 || (x)==AV_PIX_FMT_YVYU422 \
808 || (x)==AV_PIX_FMT_UYVY422 \
809 || (x)==AV_PIX_FMT_YA8 \
810 || (x)==AV_PIX_FMT_YA16LE \
811 || (x)==AV_PIX_FMT_YA16BE \
812 || (x)==AV_PIX_FMT_AYUV64LE \
813 || (x)==AV_PIX_FMT_AYUV64BE \
818 static av_always_inline int isPacked(enum AVPixelFormat pix_fmt)
820 const AVPixFmtDescriptor *desc = av_pix_fmt_desc_get(pix_fmt);
822 return ((desc->nb_components >= 2 && !(desc->flags & AV_PIX_FMT_FLAG_PLANAR)) ||
823 pix_fmt == AV_PIX_FMT_PAL8);
827 static av_always_inline int isPlanar(enum AVPixelFormat pix_fmt)
829 const AVPixFmtDescriptor *desc = av_pix_fmt_desc_get(pix_fmt);
831 return (desc->nb_components >= 2 && (desc->flags & AV_PIX_FMT_FLAG_PLANAR));
834 static av_always_inline int isPackedRGB(enum AVPixelFormat pix_fmt)
836 const AVPixFmtDescriptor *desc = av_pix_fmt_desc_get(pix_fmt);
838 return ((desc->flags & (AV_PIX_FMT_FLAG_PLANAR | AV_PIX_FMT_FLAG_RGB)) == AV_PIX_FMT_FLAG_RGB);
841 static av_always_inline int isPlanarRGB(enum AVPixelFormat pix_fmt)
843 const AVPixFmtDescriptor *desc = av_pix_fmt_desc_get(pix_fmt);
845 return ((desc->flags & (AV_PIX_FMT_FLAG_PLANAR | AV_PIX_FMT_FLAG_RGB)) ==
846 (AV_PIX_FMT_FLAG_PLANAR | AV_PIX_FMT_FLAG_RGB));
849 static av_always_inline int usePal(enum AVPixelFormat pix_fmt)
851 const AVPixFmtDescriptor *desc = av_pix_fmt_desc_get(pix_fmt);
853 return (desc->flags & AV_PIX_FMT_FLAG_PAL) || (desc->flags & AV_PIX_FMT_FLAG_PSEUDOPAL);
856 extern const uint64_t ff_dither4[2];
857 extern const uint64_t ff_dither8[2];
859 extern const uint8_t ff_dither_2x2_4[3][8];
860 extern const uint8_t ff_dither_2x2_8[3][8];
861 extern const uint8_t ff_dither_4x4_16[5][8];
862 extern const uint8_t ff_dither_8x8_32[9][8];
863 extern const uint8_t ff_dither_8x8_73[9][8];
864 extern const uint8_t ff_dither_8x8_128[9][8];
865 extern const uint8_t ff_dither_8x8_220[9][8];
867 extern const int32_t ff_yuv2rgb_coeffs[8][4];
869 extern const AVClass sws_context_class;
872 * Set c->swscale to an unscaled converter if one exists for the specific
873 * source and destination formats, bit depths, flags, etc.
875 void ff_get_unscaled_swscale(SwsContext *c);
876 void ff_get_unscaled_swscale_ppc(SwsContext *c);
877 void ff_get_unscaled_swscale_arm(SwsContext *c);
880 * Return function pointer to fastest main scaler path function depending
881 * on architecture and available optimizations.
883 SwsFunc ff_getSwsFunc(SwsContext *c);
885 void ff_sws_init_input_funcs(SwsContext *c);
886 void ff_sws_init_output_funcs(SwsContext *c,
887 yuv2planar1_fn *yuv2plane1,
888 yuv2planarX_fn *yuv2planeX,
889 yuv2interleavedX_fn *yuv2nv12cX,
890 yuv2packed1_fn *yuv2packed1,
891 yuv2packed2_fn *yuv2packed2,
892 yuv2packedX_fn *yuv2packedX,
893 yuv2anyX_fn *yuv2anyX);
894 void ff_sws_init_swscale_ppc(SwsContext *c);
895 void ff_sws_init_swscale_x86(SwsContext *c);
897 void ff_hyscale_fast_c(SwsContext *c, int16_t *dst, int dstWidth,
898 const uint8_t *src, int srcW, int xInc);
899 void ff_hcscale_fast_c(SwsContext *c, int16_t *dst1, int16_t *dst2,
900 int dstWidth, const uint8_t *src1,
901 const uint8_t *src2, int srcW, int xInc);
902 int ff_init_hscaler_mmxext(int dstW, int xInc, uint8_t *filterCode,
903 int16_t *filter, int32_t *filterPos,
905 void ff_hyscale_fast_mmxext(SwsContext *c, int16_t *dst,
906 int dstWidth, const uint8_t *src,
908 void ff_hcscale_fast_mmxext(SwsContext *c, int16_t *dst1, int16_t *dst2,
909 int dstWidth, const uint8_t *src1,
910 const uint8_t *src2, int srcW, int xInc);
913 * Allocate and return an SwsContext.
914 * This is like sws_getContext() but does not perform the init step, allowing
915 * the user to set additional AVOptions.
917 * @see sws_getContext()
919 struct SwsContext *sws_alloc_set_opts(int srcW, int srcH, enum AVPixelFormat srcFormat,
920 int dstW, int dstH, enum AVPixelFormat dstFormat,
921 int flags, const double *param);
923 int ff_sws_alphablendaway(SwsContext *c, const uint8_t *src[],
924 int srcStride[], int srcSliceY, int srcSliceH,
925 uint8_t *dst[], int dstStride[]);
927 static inline void fillPlane16(uint8_t *plane, int stride, int width, int height, int y,
928 int alpha, int bits, const int big_endian)
931 uint8_t *ptr = plane + stride * y;
932 int v = alpha ? 0xFFFF>>(16-bits) : (1<<(bits-1));
933 for (i = 0; i < height; i++) {
934 #define FILL(wfunc) \
935 for (j = 0; j < width; j++) {\
947 #define MAX_SLICE_PLANES 4
950 typedef struct SwsPlane
952 int available_lines; ///< max number of lines that can be hold by this plane
953 int sliceY; ///< index of first line
954 int sliceH; ///< number of lines
955 uint8_t **line; ///< line buffer
956 uint8_t **tmp; ///< Tmp line buffer used by mmx code
960 * Struct which defines a slice of an image to be scaled or a output for
962 * A slice can also be used as intermediate ring buffer for scaling steps.
964 typedef struct SwsSlice
966 int width; ///< Slice line width
967 int h_chr_sub_sample; ///< horizontal chroma subsampling factor
968 int v_chr_sub_sample; ///< vertical chroma subsampling factor
969 int is_ring; ///< flag to identify if this slice is a ring buffer
970 int should_free_lines; ///< flag to identify if there are dynamic allocated lines
971 enum AVPixelFormat fmt; ///< planes pixel format
972 SwsPlane plane[MAX_SLICE_PLANES]; ///< color planes
976 * Struct which holds all necessary data for processing a slice.
977 * A processing step can be a color conversion or horizontal/vertical scaling.
979 typedef struct SwsFilterDescriptor
981 SwsSlice *src; ///< Source slice
982 SwsSlice *dst; ///< Output slice
984 int alpha; ///< Flag for processing alpha channel
985 void *instance; ///< Filter instance data
987 /// Function for processing input slice sliceH lines starting from line sliceY
988 int (*process)(SwsContext *c, struct SwsFilterDescriptor *desc, int sliceY, int sliceH);
989 } SwsFilterDescriptor;
991 /// Color conversion instance data
992 typedef struct ColorContext
997 /// Scaler instance data
998 typedef struct FilterContext
1006 typedef struct VScalerContext
1008 uint16_t *filter[2];
1009 int32_t *filter_pos;
1015 // warp input lines in the form (src + width*i + j) to slice format (line[i][j])
1016 int ff_init_slice_from_src(SwsSlice * s, uint8_t *src[4], int stride[4], int srcW, int lumY, int lumH, int chrY, int chrH);
1018 // Initialize scaler filter descriptor chain
1019 int ff_init_filters(SwsContext *c);
1021 // Free all filter data
1022 int ff_free_filters(SwsContext *c);
1025 function for applying ring buffer logic into slice s
1026 It checks if the slice can hold more @lum lines, if yes
1027 do nothing otherwise remove @lum least used lines.
1028 It applies the same procedure for @chr lines.
1030 int ff_rotate_slice(SwsSlice *s, int lum, int chr);
1032 /// initializes gamma conversion descriptor
1033 int ff_init_gamma_convert(SwsFilterDescriptor *desc, SwsSlice * src, uint16_t *table);
1035 /// initializes lum pixel format conversion descriptor
1036 int ff_init_desc_fmt_convert(SwsFilterDescriptor *desc, SwsSlice * src, SwsSlice *dst, uint32_t *pal);
1038 /// initializes lum horizontal scaling descriptor
1039 int ff_init_desc_hscale(SwsFilterDescriptor *desc, SwsSlice *src, SwsSlice *dst, uint16_t *filter, int * filter_pos, int filter_size, int xInc);
1041 /// initializes chr pixel format conversion descriptor
1042 int ff_init_desc_cfmt_convert(SwsFilterDescriptor *desc, SwsSlice * src, SwsSlice *dst, uint32_t *pal);
1044 /// initializes chr horizontal scaling descriptor
1045 int ff_init_desc_chscale(SwsFilterDescriptor *desc, SwsSlice *src, SwsSlice *dst, uint16_t *filter, int * filter_pos, int filter_size, int xInc);
1047 int ff_init_desc_no_chr(SwsFilterDescriptor *desc, SwsSlice * src, SwsSlice *dst);
1049 /// initializes vertical scaling descriptors
1050 int ff_init_vscale(SwsContext *c, SwsFilterDescriptor *desc, SwsSlice *src, SwsSlice *dst);
1052 /// setup vertical scaler functions
1053 void ff_init_vscale_pfn(SwsContext *c, yuv2planar1_fn yuv2plane1, yuv2planarX_fn yuv2planeX,
1054 yuv2interleavedX_fn yuv2nv12cX, yuv2packed1_fn yuv2packed1, yuv2packed2_fn yuv2packed2,
1055 yuv2packedX_fn yuv2packedX, yuv2anyX_fn yuv2anyX, int use_mmx);
1057 //number of extra lines to process
1058 #define MAX_LINES_AHEAD 4
1060 // enable use of refactored scaler code
1063 #endif /* SWSCALE_SWSCALE_INTERNAL_H */