8d03cb629d3d50d799452a54958af81a6de5e2ae
[libav.git] / libswscale / swscale_internal.h
1 /*
2 * Copyright (C) 2001-2003 Michael Niedermayer <michaelni@gmx.at>
3 *
4 * This file is part of Libav.
5 *
6 * Libav is free software; you can redistribute it and/or
7 * modify it under the terms of the GNU Lesser General Public
8 * License as published by the Free Software Foundation; either
9 * version 2.1 of the License, or (at your option) any later version.
10 *
11 * Libav is distributed in the hope that it will be useful,
12 * but WITHOUT ANY WARRANTY; without even the implied warranty of
13 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
14 * Lesser General Public License for more details.
15 *
16 * You should have received a copy of the GNU Lesser General Public
17 * License along with Libav; if not, write to the Free Software
18 * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
19 */
20
21 #ifndef SWSCALE_SWSCALE_INTERNAL_H
22 #define SWSCALE_SWSCALE_INTERNAL_H
23
24 #include "config.h"
25
26 #if HAVE_ALTIVEC_H
27 #include <altivec.h>
28 #endif
29
30 #include "libavutil/avutil.h"
31 #include "libavutil/pixfmt.h"
32
33 #define STR(s) AV_TOSTRING(s) //AV_STRINGIFY is too long
34
35 #define FAST_BGR2YV12 //use 7-bit instead of 15-bit coefficients
36
37 #define MAX_FILTER_SIZE 256
38
39 #if HAVE_BIGENDIAN
40 #define ALT32_CORR (-1)
41 #else
42 #define ALT32_CORR 1
43 #endif
44
45 #if ARCH_X86_64
46 # define APCK_PTR2 8
47 # define APCK_COEF 16
48 # define APCK_SIZE 24
49 #else
50 # define APCK_PTR2 4
51 # define APCK_COEF 8
52 # define APCK_SIZE 16
53 #endif
54
55 struct SwsContext;
56
57 typedef int (*SwsFunc)(struct SwsContext *context, const uint8_t* src[],
58 int srcStride[], int srcSliceY, int srcSliceH,
59 uint8_t* dst[], int dstStride[]);
60
61 /**
62 * Write one line of horizontally scaled Y/U/V/A to planar output
63 * without any additional vertical scaling (or point-scaling).
64 *
65 * @param c SWS scaling context
66 * @param lumSrc scaled luma (Y) source data, 15bit for 8bit output
67 * @param chrUSrc scaled chroma (U) source data, 15bit for 8bit output
68 * @param chrVSrc scaled chroma (V) source data, 15bit for 8bit output
69 * @param alpSrc scaled alpha (A) source data, 15bit for 8bit output
70 * @param dest pointer to the 4 output planes (Y/U/V/A)
71 * @param dstW width of dest[0], dest[3], lumSrc and alpSrc in pixels
72 * @param chrDstW width of dest[1], dest[2], chrUSrc and chrVSrc
73 */
74 typedef void (*yuv2planar1_fn) (struct SwsContext *c,
75 const int16_t *lumSrc, const int16_t *chrUSrc,
76 const int16_t *chrVSrc, const int16_t *alpSrc,
77 uint8_t *dest[4], int dstW, int chrDstW);
78 /**
79 * Write one line of horizontally scaled Y/U/V/A to planar output
80 * with multi-point vertical scaling between input pixels.
81 *
82 * @param c SWS scaling context
83 * @param lumFilter vertical luma/alpha scaling coefficients, 12bit [0,4096]
84 * @param lumSrc scaled luma (Y) source data, 15bit for 8bit output
85 * @param lumFilterSize number of vertical luma/alpha input lines to scale
86 * @param chrFilter vertical chroma scaling coefficients, 12bit [0,4096]
87 * @param chrUSrc scaled chroma (U) source data, 15bit for 8bit output
88 * @param chrVSrc scaled chroma (V) source data, 15bit for 8bit output
89 * @param chrFilterSize number of vertical chroma input lines to scale
90 * @param alpSrc scaled alpha (A) source data, 15bit for 8bit output
91 * @param dest pointer to the 4 output planes (Y/U/V/A)
92 * @param dstW width of dest[0], dest[3], lumSrc and alpSrc in pixels
93 * @param chrDstW width of dest[1], dest[2], chrUSrc and chrVSrc
94 */
95 typedef void (*yuv2planarX_fn) (struct SwsContext *c, const int16_t *lumFilter,
96 const int16_t **lumSrc, int lumFilterSize,
97 const int16_t *chrFilter, const int16_t **chrUSrc,
98 const int16_t **chrVSrc, int chrFilterSize,
99 const int16_t **alpSrc, uint8_t *dest[4],
100 int dstW, int chrDstW);
101 /**
102 * Write one line of horizontally scaled Y/U/V/A to packed-pixel YUV/RGB
103 * output without any additional vertical scaling (or point-scaling). Note
104 * that this function may do chroma scaling, see the "uvalpha" argument.
105 *
106 * @param c SWS scaling context
107 * @param lumSrc scaled luma (Y) source data, 15bit for 8bit output
108 * @param chrUSrc scaled chroma (U) source data, 15bit for 8bit output
109 * @param chrVSrc scaled chroma (V) source data, 15bit for 8bit output
110 * @param alpSrc scaled alpha (A) source data, 15bit for 8bit output
111 * @param dest pointer to the output plane
112 * @param dstW width of lumSrc and alpSrc in pixels, number of pixels
113 * to write into dest[]
114 * @param uvalpha chroma scaling coefficient for the second line of chroma
115 * pixels, either 2048 or 0. If 0, one chroma input is used
116 * for 2 output pixels (or if the SWS_FLAG_FULL_CHR_INT flag
117 * is set, it generates 1 output pixel). If 2048, two chroma
118 * input pixels should be averaged for 2 output pixels (this
119 * only happens if SWS_FLAG_FULL_CHR_INT is not set)
120 * @param y vertical line number for this output. This does not need
121 * to be used to calculate the offset in the destination,
122 * but can be used to generate comfort noise using dithering
123 * for some output formats.
124 */
125 typedef void (*yuv2packed1_fn) (struct SwsContext *c, const int16_t *lumSrc,
126 const int16_t *chrUSrc[2], const int16_t *chrVSrc[2],
127 const int16_t *alpSrc, uint8_t *dest,
128 int dstW, int uvalpha, int y);
129 /**
130 * Write one line of horizontally scaled Y/U/V/A to packed-pixel YUV/RGB
131 * output by doing bilinear scaling between two input lines.
132 *
133 * @param c SWS scaling context
134 * @param lumSrc scaled luma (Y) source data, 15bit for 8bit output
135 * @param chrUSrc scaled chroma (U) source data, 15bit for 8bit output
136 * @param chrVSrc scaled chroma (V) source data, 15bit for 8bit output
137 * @param alpSrc scaled alpha (A) source data, 15bit for 8bit output
138 * @param dest pointer to the output plane
139 * @param dstW width of lumSrc and alpSrc in pixels, number of pixels
140 * to write into dest[]
141 * @param yalpha luma/alpha scaling coefficients for the second input line.
142 * The first line's coefficients can be calculated by using
143 * 4096 - yalpha
144 * @param uvalpha chroma scaling coefficient for the second input line. The
145 * first line's coefficients can be calculated by using
146 * 4096 - uvalpha
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.
151 */
152 typedef void (*yuv2packed2_fn) (struct SwsContext *c, const int16_t *lumSrc[2],
153 const int16_t *chrUSrc[2], const int16_t *chrVSrc[2],
154 const int16_t *alpSrc[2], uint8_t *dest,
155 int dstW, int yalpha, int uvalpha, int y);
156 /**
157 * Write one line of horizontally scaled Y/U/V/A to packed-pixel YUV/RGB
158 * output by doing multi-point vertical scaling between input pixels.
159 *
160 * @param c SWS scaling context
161 * @param lumFilter vertical luma/alpha scaling coefficients, 12bit [0,4096]
162 * @param lumSrc scaled luma (Y) source data, 15bit for 8bit output
163 * @param lumFilterSize number of vertical luma/alpha input lines to scale
164 * @param chrFilter vertical chroma scaling coefficients, 12bit [0,4096]
165 * @param chrUSrc scaled chroma (U) source data, 15bit for 8bit output
166 * @param chrVSrc scaled chroma (V) source data, 15bit for 8bit output
167 * @param chrFilterSize number of vertical chroma input lines to scale
168 * @param alpSrc scaled alpha (A) source data, 15bit for 8bit output
169 * @param dest pointer to the output plane
170 * @param dstW width of lumSrc and alpSrc in pixels, number of pixels
171 * to write into dest[]
172 * @param y vertical line number for this output. This does not need
173 * to be used to calculate the offset in the destination,
174 * but can be used to generate comfort noise using dithering
175 * or some output formats.
176 */
177 typedef void (*yuv2packedX_fn) (struct SwsContext *c, const int16_t *lumFilter,
178 const int16_t **lumSrc, int lumFilterSize,
179 const int16_t *chrFilter, const int16_t **chrUSrc,
180 const int16_t **chrVSrc, int chrFilterSize,
181 const int16_t **alpSrc, uint8_t *dest,
182 int dstW, int y);
183
184 /* This struct should be aligned on at least a 32-byte boundary. */
185 typedef struct SwsContext {
186 /**
187 * info on struct for av_log
188 */
189 const AVClass *av_class;
190
191 /**
192 * Note that src, dst, srcStride, dstStride will be copied in the
193 * sws_scale() wrapper so they can be freely modified here.
194 */
195 SwsFunc swScale;
196 int srcW; ///< Width of source luma/alpha planes.
197 int srcH; ///< Height of source luma/alpha planes.
198 int dstH; ///< Height of destination luma/alpha planes.
199 int chrSrcW; ///< Width of source chroma planes.
200 int chrSrcH; ///< Height of source chroma planes.
201 int chrDstW; ///< Width of destination chroma planes.
202 int chrDstH; ///< Height of destination chroma planes.
203 int lumXInc, chrXInc;
204 int lumYInc, chrYInc;
205 enum PixelFormat dstFormat; ///< Destination pixel format.
206 enum PixelFormat srcFormat; ///< Source pixel format.
207 int dstFormatBpp; ///< Number of bits per pixel of the destination pixel format.
208 int srcFormatBpp; ///< Number of bits per pixel of the source pixel format.
209 int scalingBpp;
210 int chrSrcHSubSample; ///< Binary logarithm of horizontal subsampling factor between luma/alpha and chroma planes in source image.
211 int chrSrcVSubSample; ///< Binary logarithm of vertical subsampling factor between luma/alpha and chroma planes in source image.
212 int chrDstHSubSample; ///< Binary logarithm of horizontal subsampling factor between luma/alpha and chroma planes in destination image.
213 int chrDstVSubSample; ///< Binary logarithm of vertical subsampling factor between luma/alpha and chroma planes in destination image.
214 int vChrDrop; ///< Binary logarithm of extra vertical subsampling factor in source image chroma planes specified by user.
215 int sliceDir; ///< Direction that slices are fed to the scaler (1 = top-to-bottom, -1 = bottom-to-top).
216 double param[2]; ///< Input parameters for scaling algorithms that need them.
217
218 uint32_t pal_yuv[256];
219 uint32_t pal_rgb[256];
220
221 /**
222 * @name Scaled horizontal lines ring buffer.
223 * The horizontal scaler keeps just enough scaled lines in a ring buffer
224 * so they may be passed to the vertical scaler. The pointers to the
225 * allocated buffers for each line are duplicated in sequence in the ring
226 * buffer to simplify indexing and avoid wrapping around between lines
227 * inside the vertical scaler code. The wrapping is done before the
228 * vertical scaler is called.
229 */
230 //@{
231 int16_t **lumPixBuf; ///< Ring buffer for scaled horizontal luma plane lines to be fed to the vertical scaler.
232 int16_t **chrUPixBuf; ///< Ring buffer for scaled horizontal chroma plane lines to be fed to the vertical scaler.
233 int16_t **chrVPixBuf; ///< Ring buffer for scaled horizontal chroma plane lines to be fed to the vertical scaler.
234 int16_t **alpPixBuf; ///< Ring buffer for scaled horizontal alpha plane lines to be fed to the vertical scaler.
235 int vLumBufSize; ///< Number of vertical luma/alpha lines allocated in the ring buffer.
236 int vChrBufSize; ///< Number of vertical chroma lines allocated in the ring buffer.
237 int lastInLumBuf; ///< Last scaled horizontal luma/alpha line from source in the ring buffer.
238 int lastInChrBuf; ///< Last scaled horizontal chroma line from source in the ring buffer.
239 int lumBufIndex; ///< Index in ring buffer of the last scaled horizontal luma/alpha line from source.
240 int chrBufIndex; ///< Index in ring buffer of the last scaled horizontal chroma line from source.
241 //@}
242
243 uint8_t *formatConvBuffer;
244
245 /**
246 * @name Horizontal and vertical filters.
247 * To better understand the following fields, here is a pseudo-code of
248 * their usage in filtering a horizontal line:
249 * @code
250 * for (i = 0; i < width; i++) {
251 * dst[i] = 0;
252 * for (j = 0; j < filterSize; j++)
253 * dst[i] += src[ filterPos[i] + j ] * filter[ filterSize * i + j ];
254 * dst[i] >>= FRAC_BITS; // The actual implementation is fixed-point.
255 * }
256 * @endcode
257 */
258 //@{
259 int16_t *hLumFilter; ///< Array of horizontal filter coefficients for luma/alpha planes.
260 int16_t *hChrFilter; ///< Array of horizontal filter coefficients for chroma planes.
261 int16_t *vLumFilter; ///< Array of vertical filter coefficients for luma/alpha planes.
262 int16_t *vChrFilter; ///< Array of vertical filter coefficients for chroma planes.
263 int16_t *hLumFilterPos; ///< Array of horizontal filter starting positions for each dst[i] for luma/alpha planes.
264 int16_t *hChrFilterPos; ///< Array of horizontal filter starting positions for each dst[i] for chroma planes.
265 int16_t *vLumFilterPos; ///< Array of vertical filter starting positions for each dst[i] for luma/alpha planes.
266 int16_t *vChrFilterPos; ///< Array of vertical filter starting positions for each dst[i] for chroma planes.
267 int hLumFilterSize; ///< Horizontal filter size for luma/alpha pixels.
268 int hChrFilterSize; ///< Horizontal filter size for chroma pixels.
269 int vLumFilterSize; ///< Vertical filter size for luma/alpha pixels.
270 int vChrFilterSize; ///< Vertical filter size for chroma pixels.
271 //@}
272
273 int lumMmx2FilterCodeSize; ///< Runtime-generated MMX2 horizontal fast bilinear scaler code size for luma/alpha planes.
274 int chrMmx2FilterCodeSize; ///< Runtime-generated MMX2 horizontal fast bilinear scaler code size for chroma planes.
275 uint8_t *lumMmx2FilterCode; ///< Runtime-generated MMX2 horizontal fast bilinear scaler code for luma/alpha planes.
276 uint8_t *chrMmx2FilterCode; ///< Runtime-generated MMX2 horizontal fast bilinear scaler code for chroma planes.
277
278 int canMMX2BeUsed;
279
280 int dstY; ///< Last destination vertical line output from last slice.
281 int flags; ///< Flags passed by the user to select scaler algorithm, optimizations, subsampling, etc...
282 void * yuvTable; // pointer to the yuv->rgb table start so it can be freed()
283 uint8_t * table_rV[256];
284 uint8_t * table_gU[256];
285 int table_gV[256];
286 uint8_t * table_bU[256];
287
288 //Colorspace stuff
289 int contrast, brightness, saturation; // for sws_getColorspaceDetails
290 int srcColorspaceTable[4];
291 int dstColorspaceTable[4];
292 int srcRange; ///< 0 = MPG YUV range, 1 = JPG YUV range (source image).
293 int dstRange; ///< 0 = MPG YUV range, 1 = JPG YUV range (destination image).
294 int yuv2rgb_y_offset;
295 int yuv2rgb_y_coeff;
296 int yuv2rgb_v2r_coeff;
297 int yuv2rgb_v2g_coeff;
298 int yuv2rgb_u2g_coeff;
299 int yuv2rgb_u2b_coeff;
300
301 #define RED_DITHER "0*8"
302 #define GREEN_DITHER "1*8"
303 #define BLUE_DITHER "2*8"
304 #define Y_COEFF "3*8"
305 #define VR_COEFF "4*8"
306 #define UB_COEFF "5*8"
307 #define VG_COEFF "6*8"
308 #define UG_COEFF "7*8"
309 #define Y_OFFSET "8*8"
310 #define U_OFFSET "9*8"
311 #define V_OFFSET "10*8"
312 #define LUM_MMX_FILTER_OFFSET "11*8"
313 #define CHR_MMX_FILTER_OFFSET "11*8+4*4*256"
314 #define DSTW_OFFSET "11*8+4*4*256*2" //do not change, it is hardcoded in the ASM
315 #define ESP_OFFSET "11*8+4*4*256*2+8"
316 #define VROUNDER_OFFSET "11*8+4*4*256*2+16"
317 #define U_TEMP "11*8+4*4*256*2+24"
318 #define V_TEMP "11*8+4*4*256*2+32"
319 #define Y_TEMP "11*8+4*4*256*2+40"
320 #define ALP_MMX_FILTER_OFFSET "11*8+4*4*256*2+48"
321 #define UV_OFF "11*8+4*4*256*3+48"
322 #define UV_OFFx2 "11*8+4*4*256*3+56"
323
324 DECLARE_ALIGNED(8, uint64_t, redDither);
325 DECLARE_ALIGNED(8, uint64_t, greenDither);
326 DECLARE_ALIGNED(8, uint64_t, blueDither);
327
328 DECLARE_ALIGNED(8, uint64_t, yCoeff);
329 DECLARE_ALIGNED(8, uint64_t, vrCoeff);
330 DECLARE_ALIGNED(8, uint64_t, ubCoeff);
331 DECLARE_ALIGNED(8, uint64_t, vgCoeff);
332 DECLARE_ALIGNED(8, uint64_t, ugCoeff);
333 DECLARE_ALIGNED(8, uint64_t, yOffset);
334 DECLARE_ALIGNED(8, uint64_t, uOffset);
335 DECLARE_ALIGNED(8, uint64_t, vOffset);
336 int32_t lumMmxFilter[4*MAX_FILTER_SIZE];
337 int32_t chrMmxFilter[4*MAX_FILTER_SIZE];
338 int dstW; ///< Width of destination luma/alpha planes.
339 DECLARE_ALIGNED(8, uint64_t, esp);
340 DECLARE_ALIGNED(8, uint64_t, vRounder);
341 DECLARE_ALIGNED(8, uint64_t, u_temp);
342 DECLARE_ALIGNED(8, uint64_t, v_temp);
343 DECLARE_ALIGNED(8, uint64_t, y_temp);
344 int32_t alpMmxFilter[4*MAX_FILTER_SIZE];
345 DECLARE_ALIGNED(8, ptrdiff_t, uv_off); ///< offset (in pixels) between u and v planes
346 DECLARE_ALIGNED(8, ptrdiff_t, uv_offx2); ///< offset (in bytes) between u and v planes
347
348 #if HAVE_ALTIVEC
349 vector signed short CY;
350 vector signed short CRV;
351 vector signed short CBU;
352 vector signed short CGU;
353 vector signed short CGV;
354 vector signed short OY;
355 vector unsigned short CSHIFT;
356 vector signed short *vYCoeffsBank, *vCCoeffsBank;
357 #endif
358
359 #if ARCH_BFIN
360 DECLARE_ALIGNED(4, uint32_t, oy);
361 DECLARE_ALIGNED(4, uint32_t, oc);
362 DECLARE_ALIGNED(4, uint32_t, zero);
363 DECLARE_ALIGNED(4, uint32_t, cy);
364 DECLARE_ALIGNED(4, uint32_t, crv);
365 DECLARE_ALIGNED(4, uint32_t, rmask);
366 DECLARE_ALIGNED(4, uint32_t, cbu);
367 DECLARE_ALIGNED(4, uint32_t, bmask);
368 DECLARE_ALIGNED(4, uint32_t, cgu);
369 DECLARE_ALIGNED(4, uint32_t, cgv);
370 DECLARE_ALIGNED(4, uint32_t, gmask);
371 #endif
372
373 #if HAVE_VIS
374 DECLARE_ALIGNED(8, uint64_t, sparc_coeffs)[10];
375 #endif
376
377 /* function pointers for swScale() */
378 yuv2planar1_fn yuv2yuv1;
379 yuv2planarX_fn yuv2yuvX;
380 yuv2packed1_fn yuv2packed1;
381 yuv2packed2_fn yuv2packed2;
382 yuv2packedX_fn yuv2packedX;
383
384 void (*lumToYV12)(uint8_t *dst, const uint8_t *src,
385 int width, uint32_t *pal); ///< Unscaled conversion of luma plane to YV12 for horizontal scaler.
386 void (*alpToYV12)(uint8_t *dst, const uint8_t *src,
387 int width, uint32_t *pal); ///< Unscaled conversion of alpha plane to YV12 for horizontal scaler.
388 void (*chrToYV12)(uint8_t *dstU, uint8_t *dstV,
389 const uint8_t *src1, const uint8_t *src2,
390 int width, uint32_t *pal); ///< Unscaled conversion of chroma planes to YV12 for horizontal scaler.
391 /**
392 * Scale one horizontal line of input data using a bilinear filter
393 * to produce one line of output data. Compared to SwsContext->hScale(),
394 * please take note of the following caveats when using these:
395 * - Scaling is done using only 7bit instead of 14bit coefficients.
396 * - You can use no more than 5 input pixels to produce 4 output
397 * pixels. Therefore, this filter should not be used for downscaling
398 * by more than ~20% in width (because that equals more than 5/4th
399 * downscaling and thus more than 5 pixels input per 4 pixels output).
400 * - In general, bilinear filters create artifacts during downscaling
401 * (even when <20%), because one output pixel will span more than one
402 * input pixel, and thus some pixels will need edges of both neighbor
403 * pixels to interpolate the output pixel. Since you can use at most
404 * two input pixels per output pixel in bilinear scaling, this is
405 * impossible and thus downscaling by any size will create artifacts.
406 * To enable this type of scaling, set SWS_FLAG_FAST_BILINEAR
407 * in SwsContext->flags.
408 */
409 /** @{ */
410 void (*hyscale_fast)(struct SwsContext *c,
411 int16_t *dst, int dstWidth,
412 const uint8_t *src, int srcW, int xInc);
413 void (*hcscale_fast)(struct SwsContext *c,
414 int16_t *dst1, int16_t *dst2, int dstWidth,
415 const uint8_t *src1, const uint8_t *src2,
416 int srcW, int xInc);
417 /** @} */
418
419 /**
420 * Scale one horizontal line of input data using a filter over the input
421 * lines, to produce one (differently sized) line of output data.
422 *
423 * @param dst pointer to destination buffer for horizontally scaled
424 * data. If the scaling depth (SwsContext->scalingBpp) is
425 * 8, data will be 15bpp in 16bits (int16_t) width. If
426 * scaling depth is 16, data will be 19bpp in 32bpp
427 * (int32_t) width.
428 * @param dstW width of destination image
429 * @param src pointer to source data to be scaled. If scaling depth
430 * is 8, this is 8bpp in 8bpp (uint8_t) width. If scaling
431 * depth is 16, this is 16bpp in 16bpp (uint16_t) depth.
432 * @param filter filter coefficients to be used per output pixel for
433 * scaling. This contains 14bpp filtering coefficients.
434 * Guaranteed to contain dstW * filterSize entries.
435 * @param filterPos position of the first input pixel to be used for
436 * each output pixel during scaling. Guaranteed to
437 * contain dstW entries.
438 * @param filterSize the number of input coefficients to be used (and
439 * thus the number of input pixels to be used) for
440 * creating a single output pixel. Is aligned to 4
441 * (and input coefficients thus padded with zeroes)
442 * to simplify creating SIMD code.
443 */
444 void (*hScale)(struct SwsContext *c, int16_t *dst, int dstW, const uint8_t *src,
445 const int16_t *filter, const int16_t *filterPos,
446 int filterSize);
447
448 void (*lumConvertRange)(int16_t *dst, int width); ///< Color range conversion function for luma plane if needed.
449 void (*chrConvertRange)(int16_t *dst1, int16_t *dst2, int width); ///< Color range conversion function for chroma planes if needed.
450
451 /**
452 * dst[..] = (src[..] << 8) | src[..];
453 */
454 void (*scale8To16Rv)(uint16_t *dst, const uint8_t *src, int len);
455 /**
456 * dst[..] = src[..] >> 4;
457 */
458 void (*scale19To15Fw)(int16_t *dst, const int32_t *src, int len);
459
460 int needs_hcscale; ///< Set if there are chroma planes to be converted.
461
462 } SwsContext;
463 //FIXME check init (where 0)
464
465 SwsFunc ff_yuv2rgb_get_func_ptr(SwsContext *c);
466 int ff_yuv2rgb_c_init_tables(SwsContext *c, const int inv_table[4],
467 int fullRange, int brightness,
468 int contrast, int saturation);
469
470 void ff_yuv2rgb_init_tables_altivec(SwsContext *c, const int inv_table[4],
471 int brightness, int contrast, int saturation);
472 void updateMMXDitherTables(SwsContext *c, int dstY, int lumBufIndex, int chrBufIndex,
473 int lastInLumBuf, int lastInChrBuf);
474
475 SwsFunc ff_yuv2rgb_init_mmx(SwsContext *c);
476 SwsFunc ff_yuv2rgb_init_vis(SwsContext *c);
477 SwsFunc ff_yuv2rgb_init_mlib(SwsContext *c);
478 SwsFunc ff_yuv2rgb_init_altivec(SwsContext *c);
479 SwsFunc ff_yuv2rgb_get_func_ptr_bfin(SwsContext *c);
480 void ff_bfin_get_unscaled_swscale(SwsContext *c);
481
482 const char *sws_format_name(enum PixelFormat format);
483
484 //FIXME replace this with something faster
485 #define is16BPS(x) ( \
486 (x)==PIX_FMT_GRAY16BE \
487 || (x)==PIX_FMT_GRAY16LE \
488 || (x)==PIX_FMT_BGR48BE \
489 || (x)==PIX_FMT_BGR48LE \
490 || (x)==PIX_FMT_RGB48BE \
491 || (x)==PIX_FMT_RGB48LE \
492 || (x)==PIX_FMT_YUV420P16LE \
493 || (x)==PIX_FMT_YUV422P16LE \
494 || (x)==PIX_FMT_YUV444P16LE \
495 || (x)==PIX_FMT_YUV420P16BE \
496 || (x)==PIX_FMT_YUV422P16BE \
497 || (x)==PIX_FMT_YUV444P16BE \
498 )
499 #define is9_OR_10BPS(x) ( \
500 (x)==PIX_FMT_YUV420P9LE \
501 || (x)==PIX_FMT_YUV420P9BE \
502 || (x)==PIX_FMT_YUV444P9BE \
503 || (x)==PIX_FMT_YUV444P9LE \
504 || (x)==PIX_FMT_YUV422P10BE \
505 || (x)==PIX_FMT_YUV422P10LE \
506 || (x)==PIX_FMT_YUV444P10BE \
507 || (x)==PIX_FMT_YUV444P10LE \
508 || (x)==PIX_FMT_YUV420P10LE \
509 || (x)==PIX_FMT_YUV420P10BE \
510 )
511 #define isBE(x) ((x)&1)
512 #define isPlanar8YUV(x) ( \
513 (x)==PIX_FMT_YUV410P \
514 || (x)==PIX_FMT_YUV420P \
515 || (x)==PIX_FMT_YUVA420P \
516 || (x)==PIX_FMT_YUV411P \
517 || (x)==PIX_FMT_YUV422P \
518 || (x)==PIX_FMT_YUV444P \
519 || (x)==PIX_FMT_YUV440P \
520 || (x)==PIX_FMT_NV12 \
521 || (x)==PIX_FMT_NV21 \
522 )
523 #define isPlanarYUV(x) ( \
524 isPlanar8YUV(x) \
525 || (x)==PIX_FMT_YUV420P9LE \
526 || (x)==PIX_FMT_YUV444P9LE \
527 || (x)==PIX_FMT_YUV420P10LE \
528 || (x)==PIX_FMT_YUV422P10LE \
529 || (x)==PIX_FMT_YUV444P10LE \
530 || (x)==PIX_FMT_YUV420P16LE \
531 || (x)==PIX_FMT_YUV422P16LE \
532 || (x)==PIX_FMT_YUV444P16LE \
533 || (x)==PIX_FMT_YUV420P9BE \
534 || (x)==PIX_FMT_YUV444P9BE \
535 || (x)==PIX_FMT_YUV420P10BE \
536 || (x)==PIX_FMT_YUV422P10BE \
537 || (x)==PIX_FMT_YUV444P10BE \
538 || (x)==PIX_FMT_YUV420P16BE \
539 || (x)==PIX_FMT_YUV422P16BE \
540 || (x)==PIX_FMT_YUV444P16BE \
541 )
542 #define isYUV(x) ( \
543 (x)==PIX_FMT_UYVY422 \
544 || (x)==PIX_FMT_YUYV422 \
545 || isPlanarYUV(x) \
546 )
547 #define isGray(x) ( \
548 (x)==PIX_FMT_GRAY8 \
549 || (x)==PIX_FMT_Y400A \
550 || (x)==PIX_FMT_GRAY16BE \
551 || (x)==PIX_FMT_GRAY16LE \
552 )
553 #define isGray16(x) ( \
554 (x)==PIX_FMT_GRAY16BE \
555 || (x)==PIX_FMT_GRAY16LE \
556 )
557 #define isRGBinInt(x) ( \
558 (x)==PIX_FMT_RGB48BE \
559 || (x)==PIX_FMT_RGB48LE \
560 || (x)==PIX_FMT_RGB32 \
561 || (x)==PIX_FMT_RGB32_1 \
562 || (x)==PIX_FMT_RGB24 \
563 || (x)==PIX_FMT_RGB565BE \
564 || (x)==PIX_FMT_RGB565LE \
565 || (x)==PIX_FMT_RGB555BE \
566 || (x)==PIX_FMT_RGB555LE \
567 || (x)==PIX_FMT_RGB444BE \
568 || (x)==PIX_FMT_RGB444LE \
569 || (x)==PIX_FMT_RGB8 \
570 || (x)==PIX_FMT_RGB4 \
571 || (x)==PIX_FMT_RGB4_BYTE \
572 || (x)==PIX_FMT_MONOBLACK \
573 || (x)==PIX_FMT_MONOWHITE \
574 )
575 #define isBGRinInt(x) ( \
576 (x)==PIX_FMT_BGR48BE \
577 || (x)==PIX_FMT_BGR48LE \
578 || (x)==PIX_FMT_BGR32 \
579 || (x)==PIX_FMT_BGR32_1 \
580 || (x)==PIX_FMT_BGR24 \
581 || (x)==PIX_FMT_BGR565BE \
582 || (x)==PIX_FMT_BGR565LE \
583 || (x)==PIX_FMT_BGR555BE \
584 || (x)==PIX_FMT_BGR555LE \
585 || (x)==PIX_FMT_BGR444BE \
586 || (x)==PIX_FMT_BGR444LE \
587 || (x)==PIX_FMT_BGR8 \
588 || (x)==PIX_FMT_BGR4 \
589 || (x)==PIX_FMT_BGR4_BYTE \
590 || (x)==PIX_FMT_MONOBLACK \
591 || (x)==PIX_FMT_MONOWHITE \
592 )
593 #define isRGBinBytes(x) ( \
594 (x)==PIX_FMT_RGB48BE \
595 || (x)==PIX_FMT_RGB48LE \
596 || (x)==PIX_FMT_RGBA \
597 || (x)==PIX_FMT_ARGB \
598 || (x)==PIX_FMT_RGB24 \
599 )
600 #define isBGRinBytes(x) ( \
601 (x)==PIX_FMT_BGR48BE \
602 || (x)==PIX_FMT_BGR48LE \
603 || (x)==PIX_FMT_BGRA \
604 || (x)==PIX_FMT_ABGR \
605 || (x)==PIX_FMT_BGR24 \
606 )
607 #define isAnyRGB(x) ( \
608 isRGBinInt(x) \
609 || isBGRinInt(x) \
610 )
611 #define isALPHA(x) ( \
612 (x)==PIX_FMT_BGR32 \
613 || (x)==PIX_FMT_BGR32_1 \
614 || (x)==PIX_FMT_RGB32 \
615 || (x)==PIX_FMT_RGB32_1 \
616 || (x)==PIX_FMT_Y400A \
617 || (x)==PIX_FMT_YUVA420P \
618 )
619 #define isPacked(x) ( \
620 (x)==PIX_FMT_PAL8 \
621 || (x)==PIX_FMT_YUYV422 \
622 || (x)==PIX_FMT_UYVY422 \
623 || (x)==PIX_FMT_Y400A \
624 || isAnyRGB(x) \
625 )
626 #define usePal(x) ((av_pix_fmt_descriptors[x].flags & PIX_FMT_PAL) || (x) == PIX_FMT_Y400A)
627
628 extern const uint64_t ff_dither4[2];
629 extern const uint64_t ff_dither8[2];
630
631 extern const AVClass sws_context_class;
632
633 /**
634 * Sets c->swScale to an unscaled converter if one exists for the specific
635 * source and destination formats, bit depths, flags, etc.
636 */
637 void ff_get_unscaled_swscale(SwsContext *c);
638
639 void ff_swscale_get_unscaled_altivec(SwsContext *c);
640
641 /**
642 * Returns function pointer to fastest main scaler path function depending
643 * on architecture and available optimizations.
644 */
645 SwsFunc ff_getSwsFunc(SwsContext *c);
646
647 void ff_sws_init_swScale_altivec(SwsContext *c);
648 void ff_sws_init_swScale_mmx(SwsContext *c);
649
650 #endif /* SWSCALE_SWSCALE_INTERNAL_H */