FFmpeg
 All Data Structures Namespaces Files Functions Variables Typedefs Enumerations Enumerator Macros Groups Pages
swscale_internal.h
Go to the documentation of this file.
1 /*
2  * Copyright (C) 2001-2011 Michael Niedermayer <michaelni@gmx.at>
3  *
4  * This file is part of FFmpeg.
5  *
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.
10  *
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.
15  *
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
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/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"
37 
38 #define STR(s) AV_TOSTRING(s) // AV_STRINGIFY is too long
39 
40 #define YUVRGB_TABLE_HEADROOM 128
41 
42 #define MAX_FILTER_SIZE 256
43 
44 #define DITHER1XBPP
45 
46 #if HAVE_BIGENDIAN
47 #define ALT32_CORR (-1)
48 #else
49 #define ALT32_CORR 1
50 #endif
51 
52 #if ARCH_X86_64
53 # define APCK_PTR2 8
54 # define APCK_COEF 16
55 # define APCK_SIZE 24
56 #else
57 # define APCK_PTR2 4
58 # define APCK_COEF 8
59 # define APCK_SIZE 16
60 #endif
61 
62 struct SwsContext;
63 
64 typedef int (*SwsFunc)(struct SwsContext *context, const uint8_t *src[],
65  int srcStride[], int srcSliceY, int srcSliceH,
66  uint8_t *dst[], int dstStride[]);
67 
68 /**
69  * Write one line of horizontally scaled data to planar output
70  * without any additional vertical scaling (or point-scaling).
71  *
72  * @param src scaled source data, 15bit for 8-10bit output,
73  * 19-bit for 16bit output (in int32_t)
74  * @param dest pointer to the output plane. For >8bit
75  * output, this is in uint16_t
76  * @param dstW width of destination in pixels
77  * @param dither ordered dither array of type int16_t and size 8
78  * @param offset Dither offset
79  */
80 typedef void (*yuv2planar1_fn)(const int16_t *src, uint8_t *dest, int dstW,
81  const uint8_t *dither, int offset);
82 
83 /**
84  * Write one line of horizontally scaled data to planar output
85  * with multi-point vertical scaling between input pixels.
86  *
87  * @param filter vertical luma/alpha scaling coefficients, 12bit [0,4096]
88  * @param src scaled luma (Y) or alpha (A) source data, 15bit for 8-10bit output,
89  * 19-bit for 16bit output (in int32_t)
90  * @param filterSize number of vertical input lines to scale
91  * @param dest pointer to output plane. For >8bit
92  * output, this is in uint16_t
93  * @param dstW width of destination pixels
94  * @param offset Dither offset
95  */
96 typedef void (*yuv2planarX_fn)(const int16_t *filter, int filterSize,
97  const int16_t **src, uint8_t *dest, int dstW,
98  const uint8_t *dither, int offset);
99 
100 /**
101  * Write one line of horizontally scaled chroma to interleaved output
102  * with multi-point vertical scaling between input pixels.
103  *
104  * @param c SWS scaling context
105  * @param chrFilter vertical chroma scaling coefficients, 12bit [0,4096]
106  * @param chrUSrc scaled chroma (U) source data, 15bit for 8-10bit output,
107  * 19-bit for 16bit output (in int32_t)
108  * @param chrVSrc scaled chroma (V) source data, 15bit for 8-10bit output,
109  * 19-bit for 16bit output (in int32_t)
110  * @param chrFilterSize number of vertical chroma input lines to scale
111  * @param dest pointer to the output plane. For >8bit
112  * output, this is in uint16_t
113  * @param dstW width of chroma planes
114  */
116  const int16_t *chrFilter,
117  int chrFilterSize,
118  const int16_t **chrUSrc,
119  const int16_t **chrVSrc,
120  uint8_t *dest, int dstW);
121 
122 /**
123  * Write one line of horizontally scaled Y/U/V/A to packed-pixel YUV/RGB
124  * output without any additional vertical scaling (or point-scaling). Note
125  * that this function may do chroma scaling, see the "uvalpha" argument.
126  *
127  * @param c SWS scaling context
128  * @param lumSrc scaled luma (Y) source data, 15bit for 8-10bit output,
129  * 19-bit for 16bit output (in int32_t)
130  * @param chrUSrc scaled chroma (U) source data, 15bit for 8-10bit output,
131  * 19-bit for 16bit output (in int32_t)
132  * @param chrVSrc scaled chroma (V) source data, 15bit for 8-10bit output,
133  * 19-bit for 16bit output (in int32_t)
134  * @param alpSrc scaled alpha (A) source data, 15bit for 8-10bit output,
135  * 19-bit for 16bit output (in int32_t)
136  * @param dest pointer to the output plane. For 16bit output, this is
137  * uint16_t
138  * @param dstW width of lumSrc and alpSrc in pixels, number of pixels
139  * to write into dest[]
140  * @param uvalpha chroma scaling coefficient for the second line of chroma
141  * pixels, either 2048 or 0. If 0, one chroma input is used
142  * for 2 output pixels (or if the SWS_FLAG_FULL_CHR_INT flag
143  * is set, it generates 1 output pixel). If 2048, two chroma
144  * input pixels should be averaged for 2 output pixels (this
145  * only happens if SWS_FLAG_FULL_CHR_INT is not set)
146  * @param y vertical line number for this output. This does not need
147  * to be used to calculate the offset in the destination,
148  * but can be used to generate comfort noise using dithering
149  * for some output formats.
150  */
151 typedef void (*yuv2packed1_fn)(struct SwsContext *c, const int16_t *lumSrc,
152  const int16_t *chrUSrc[2],
153  const int16_t *chrVSrc[2],
154  const int16_t *alpSrc, uint8_t *dest,
155  int dstW, 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 bilinear scaling between two input lines.
159  *
160  * @param c SWS scaling context
161  * @param lumSrc scaled luma (Y) source data, 15bit for 8-10bit output,
162  * 19-bit for 16bit output (in int32_t)
163  * @param chrUSrc scaled chroma (U) source data, 15bit for 8-10bit output,
164  * 19-bit for 16bit output (in int32_t)
165  * @param chrVSrc scaled chroma (V) source data, 15bit for 8-10bit output,
166  * 19-bit for 16bit output (in int32_t)
167  * @param alpSrc scaled alpha (A) source data, 15bit for 8-10bit output,
168  * 19-bit for 16bit output (in int32_t)
169  * @param dest pointer to the output plane. For 16bit output, this is
170  * uint16_t
171  * @param dstW width of lumSrc and alpSrc in pixels, number of pixels
172  * to write into dest[]
173  * @param yalpha luma/alpha scaling coefficients for the second input line.
174  * The first line's coefficients can be calculated by using
175  * 4096 - yalpha
176  * @param uvalpha chroma scaling coefficient for the second input line. The
177  * first line's coefficients can be calculated by using
178  * 4096 - uvalpha
179  * @param y vertical line number for this output. This does not need
180  * to be used to calculate the offset in the destination,
181  * but can be used to generate comfort noise using dithering
182  * for some output formats.
183  */
184 typedef void (*yuv2packed2_fn)(struct SwsContext *c, const int16_t *lumSrc[2],
185  const int16_t *chrUSrc[2],
186  const int16_t *chrVSrc[2],
187  const int16_t *alpSrc[2],
188  uint8_t *dest,
189  int dstW, int yalpha, int uvalpha, int y);
190 /**
191  * Write one line of horizontally scaled Y/U/V/A to packed-pixel YUV/RGB
192  * output by doing multi-point vertical scaling between input pixels.
193  *
194  * @param c SWS scaling context
195  * @param lumFilter vertical luma/alpha scaling coefficients, 12bit [0,4096]
196  * @param lumSrc scaled luma (Y) source data, 15bit for 8-10bit output,
197  * 19-bit for 16bit output (in int32_t)
198  * @param lumFilterSize number of vertical luma/alpha input lines to scale
199  * @param chrFilter vertical chroma scaling coefficients, 12bit [0,4096]
200  * @param chrUSrc scaled chroma (U) source data, 15bit for 8-10bit output,
201  * 19-bit for 16bit output (in int32_t)
202  * @param chrVSrc scaled chroma (V) source data, 15bit for 8-10bit output,
203  * 19-bit for 16bit output (in int32_t)
204  * @param chrFilterSize number of vertical chroma input lines to scale
205  * @param alpSrc scaled alpha (A) source data, 15bit for 8-10bit output,
206  * 19-bit for 16bit output (in int32_t)
207  * @param dest pointer to the output plane. For 16bit output, this is
208  * uint16_t
209  * @param dstW width of lumSrc and alpSrc in pixels, number of pixels
210  * to write into dest[]
211  * @param y vertical line number for this output. This does not need
212  * to be used to calculate the offset in the destination,
213  * but can be used to generate comfort noise using dithering
214  * or some output formats.
215  */
216 typedef void (*yuv2packedX_fn)(struct SwsContext *c, const int16_t *lumFilter,
217  const int16_t **lumSrc, int lumFilterSize,
218  const int16_t *chrFilter,
219  const int16_t **chrUSrc,
220  const int16_t **chrVSrc, int chrFilterSize,
221  const int16_t **alpSrc, uint8_t *dest,
222  int dstW, int y);
223 
224 /**
225  * Write one line of horizontally scaled Y/U/V/A to YUV/RGB
226  * output by doing multi-point vertical scaling between input pixels.
227  *
228  * @param c SWS scaling context
229  * @param lumFilter vertical luma/alpha scaling coefficients, 12bit [0,4096]
230  * @param lumSrc scaled luma (Y) source data, 15bit for 8-10bit output,
231  * 19-bit for 16bit output (in int32_t)
232  * @param lumFilterSize number of vertical luma/alpha input lines to scale
233  * @param chrFilter vertical chroma scaling coefficients, 12bit [0,4096]
234  * @param chrUSrc scaled chroma (U) source data, 15bit for 8-10bit output,
235  * 19-bit for 16bit output (in int32_t)
236  * @param chrVSrc scaled chroma (V) source data, 15bit for 8-10bit output,
237  * 19-bit for 16bit output (in int32_t)
238  * @param chrFilterSize number of vertical chroma input lines to scale
239  * @param alpSrc scaled alpha (A) source data, 15bit for 8-10bit output,
240  * 19-bit for 16bit output (in int32_t)
241  * @param dest pointer to the output planes. For 16bit output, this is
242  * uint16_t
243  * @param dstW width of lumSrc and alpSrc in pixels, number of pixels
244  * to write into dest[]
245  * @param y vertical line number for this output. This does not need
246  * to be used to calculate the offset in the destination,
247  * but can be used to generate comfort noise using dithering
248  * or some output formats.
249  */
250 typedef void (*yuv2anyX_fn)(struct SwsContext *c, const int16_t *lumFilter,
251  const int16_t **lumSrc, int lumFilterSize,
252  const int16_t *chrFilter,
253  const int16_t **chrUSrc,
254  const int16_t **chrVSrc, int chrFilterSize,
255  const int16_t **alpSrc, uint8_t **dest,
256  int dstW, int y);
257 
258 /* This struct should be aligned on at least a 32-byte boundary. */
259 typedef struct SwsContext {
260  /**
261  * info on struct for av_log
262  */
264 
265  /**
266  * Note that src, dst, srcStride, dstStride will be copied in the
267  * sws_scale() wrapper so they can be freely modified here.
268  */
270  int srcW; ///< Width of source luma/alpha planes.
271  int srcH; ///< Height of source luma/alpha planes.
272  int dstH; ///< Height of destination luma/alpha planes.
273  int chrSrcW; ///< Width of source chroma planes.
274  int chrSrcH; ///< Height of source chroma planes.
275  int chrDstW; ///< Width of destination chroma planes.
276  int chrDstH; ///< Height of destination chroma planes.
279  enum AVPixelFormat dstFormat; ///< Destination pixel format.
280  enum AVPixelFormat srcFormat; ///< Source pixel format.
281  int dstFormatBpp; ///< Number of bits per pixel of the destination pixel format.
282  int srcFormatBpp; ///< Number of bits per pixel of the source pixel format.
284  int chrSrcHSubSample; ///< Binary logarithm of horizontal subsampling factor between luma/alpha and chroma planes in source image.
285  int chrSrcVSubSample; ///< Binary logarithm of vertical subsampling factor between luma/alpha and chroma planes in source image.
286  int chrDstHSubSample; ///< Binary logarithm of horizontal subsampling factor between luma/alpha and chroma planes in destination image.
287  int chrDstVSubSample; ///< Binary logarithm of vertical subsampling factor between luma/alpha and chroma planes in destination image.
288  int vChrDrop; ///< Binary logarithm of extra vertical subsampling factor in source image chroma planes specified by user.
289  int sliceDir; ///< Direction that slices are fed to the scaler (1 = top-to-bottom, -1 = bottom-to-top).
290  double param[2]; ///< Input parameters for scaling algorithms that need them.
291 
292  uint32_t pal_yuv[256];
293  uint32_t pal_rgb[256];
294 
295  /**
296  * @name Scaled horizontal lines ring buffer.
297  * The horizontal scaler keeps just enough scaled lines in a ring buffer
298  * so they may be passed to the vertical scaler. The pointers to the
299  * allocated buffers for each line are duplicated in sequence in the ring
300  * buffer to simplify indexing and avoid wrapping around between lines
301  * inside the vertical scaler code. The wrapping is done before the
302  * vertical scaler is called.
303  */
304  //@{
305  int16_t **lumPixBuf; ///< Ring buffer for scaled horizontal luma plane lines to be fed to the vertical scaler.
306  int16_t **chrUPixBuf; ///< Ring buffer for scaled horizontal chroma plane lines to be fed to the vertical scaler.
307  int16_t **chrVPixBuf; ///< Ring buffer for scaled horizontal chroma plane lines to be fed to the vertical scaler.
308  int16_t **alpPixBuf; ///< Ring buffer for scaled horizontal alpha plane lines to be fed to the vertical scaler.
309  int vLumBufSize; ///< Number of vertical luma/alpha lines allocated in the ring buffer.
310  int vChrBufSize; ///< Number of vertical chroma lines allocated in the ring buffer.
311  int lastInLumBuf; ///< Last scaled horizontal luma/alpha line from source in the ring buffer.
312  int lastInChrBuf; ///< Last scaled horizontal chroma line from source in the ring buffer.
313  int lumBufIndex; ///< Index in ring buffer of the last scaled horizontal luma/alpha line from source.
314  int chrBufIndex; ///< Index in ring buffer of the last scaled horizontal chroma line from source.
315  //@}
316 
318 
319  /**
320  * @name Horizontal and vertical filters.
321  * To better understand the following fields, here is a pseudo-code of
322  * their usage in filtering a horizontal line:
323  * @code
324  * for (i = 0; i < width; i++) {
325  * dst[i] = 0;
326  * for (j = 0; j < filterSize; j++)
327  * dst[i] += src[ filterPos[i] + j ] * filter[ filterSize * i + j ];
328  * dst[i] >>= FRAC_BITS; // The actual implementation is fixed-point.
329  * }
330  * @endcode
331  */
332  //@{
333  int16_t *hLumFilter; ///< Array of horizontal filter coefficients for luma/alpha planes.
334  int16_t *hChrFilter; ///< Array of horizontal filter coefficients for chroma planes.
335  int16_t *vLumFilter; ///< Array of vertical filter coefficients for luma/alpha planes.
336  int16_t *vChrFilter; ///< Array of vertical filter coefficients for chroma planes.
337  int32_t *hLumFilterPos; ///< Array of horizontal filter starting positions for each dst[i] for luma/alpha planes.
338  int32_t *hChrFilterPos; ///< Array of horizontal filter starting positions for each dst[i] for chroma planes.
339  int32_t *vLumFilterPos; ///< Array of vertical filter starting positions for each dst[i] for luma/alpha planes.
340  int32_t *vChrFilterPos; ///< Array of vertical filter starting positions for each dst[i] for chroma planes.
341  int hLumFilterSize; ///< Horizontal filter size for luma/alpha pixels.
342  int hChrFilterSize; ///< Horizontal filter size for chroma pixels.
343  int vLumFilterSize; ///< Vertical filter size for luma/alpha pixels.
344  int vChrFilterSize; ///< Vertical filter size for chroma pixels.
345  //@}
346 
347  int lumMmxextFilterCodeSize; ///< Runtime-generated MMXEXT horizontal fast bilinear scaler code size for luma/alpha planes.
348  int chrMmxextFilterCodeSize; ///< Runtime-generated MMXEXT horizontal fast bilinear scaler code size for chroma planes.
349  uint8_t *lumMmxextFilterCode; ///< Runtime-generated MMXEXT horizontal fast bilinear scaler code for luma/alpha planes.
350  uint8_t *chrMmxextFilterCode; ///< Runtime-generated MMXEXT horizontal fast bilinear scaler code for chroma planes.
351 
353 
354  int dstY; ///< Last destination vertical line output from last slice.
355  int flags; ///< Flags passed by the user to select scaler algorithm, optimizations, subsampling, etc...
356  void *yuvTable; // pointer to the yuv->rgb table start so it can be freed()
361  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
362 #define RY_IDX 0
363 #define GY_IDX 1
364 #define BY_IDX 2
365 #define RU_IDX 3
366 #define GU_IDX 4
367 #define BU_IDX 5
368 #define RV_IDX 6
369 #define GV_IDX 7
370 #define BV_IDX 8
371 #define RGB2YUV_SHIFT 15
372 
373  int *dither_error[4];
374 
375  //Colorspace stuff
376  int contrast, brightness, saturation; // for sws_getColorspaceDetails
379  int srcRange; ///< 0 = MPG YUV range, 1 = JPG YUV range (source image).
380  int dstRange; ///< 0 = MPG YUV range, 1 = JPG YUV range (destination image).
383  int srcXYZ;
384  int dstXYZ;
391 
392 #define RED_DITHER "0*8"
393 #define GREEN_DITHER "1*8"
394 #define BLUE_DITHER "2*8"
395 #define Y_COEFF "3*8"
396 #define VR_COEFF "4*8"
397 #define UB_COEFF "5*8"
398 #define VG_COEFF "6*8"
399 #define UG_COEFF "7*8"
400 #define Y_OFFSET "8*8"
401 #define U_OFFSET "9*8"
402 #define V_OFFSET "10*8"
403 #define LUM_MMX_FILTER_OFFSET "11*8"
404 #define CHR_MMX_FILTER_OFFSET "11*8+4*4*256"
405 #define DSTW_OFFSET "11*8+4*4*256*2" //do not change, it is hardcoded in the ASM
406 #define ESP_OFFSET "11*8+4*4*256*2+8"
407 #define VROUNDER_OFFSET "11*8+4*4*256*2+16"
408 #define U_TEMP "11*8+4*4*256*2+24"
409 #define V_TEMP "11*8+4*4*256*2+32"
410 #define Y_TEMP "11*8+4*4*256*2+40"
411 #define ALP_MMX_FILTER_OFFSET "11*8+4*4*256*2+48"
412 #define UV_OFF_PX "11*8+4*4*256*3+48"
413 #define UV_OFF_BYTE "11*8+4*4*256*3+56"
414 #define DITHER16 "11*8+4*4*256*3+64"
415 #define DITHER32 "11*8+4*4*256*3+80"
416 
417  DECLARE_ALIGNED(8, uint64_t, redDither);
420 
421  DECLARE_ALIGNED(8, uint64_t, yCoeff);
422  DECLARE_ALIGNED(8, uint64_t, vrCoeff);
423  DECLARE_ALIGNED(8, uint64_t, ubCoeff);
424  DECLARE_ALIGNED(8, uint64_t, vgCoeff);
425  DECLARE_ALIGNED(8, uint64_t, ugCoeff);
426  DECLARE_ALIGNED(8, uint64_t, yOffset);
427  DECLARE_ALIGNED(8, uint64_t, uOffset);
428  DECLARE_ALIGNED(8, uint64_t, vOffset);
431  int dstW; ///< Width of destination luma/alpha planes.
432  DECLARE_ALIGNED(8, uint64_t, esp);
433  DECLARE_ALIGNED(8, uint64_t, vRounder);
434  DECLARE_ALIGNED(8, uint64_t, u_temp);
435  DECLARE_ALIGNED(8, uint64_t, v_temp);
436  DECLARE_ALIGNED(8, uint64_t, y_temp);
438  // alignment of these values is not necessary, but merely here
439  // to maintain the same offset across x8632 and x86-64. Once we
440  // use proper offset macros in the asm, they can be removed.
441  DECLARE_ALIGNED(8, ptrdiff_t, uv_off); ///< offset (in pixels) between u and v planes
442  DECLARE_ALIGNED(8, ptrdiff_t, uv_offx2); ///< offset (in bytes) between u and v planes
443  DECLARE_ALIGNED(8, uint16_t, dither16)[8];
444  DECLARE_ALIGNED(8, uint32_t, dither32)[8];
445 
447 
448 #if HAVE_ALTIVEC
449  vector signed short CY;
450  vector signed short CRV;
451  vector signed short CBU;
452  vector signed short CGU;
453  vector signed short CGV;
454  vector signed short OY;
455  vector unsigned short CSHIFT;
456  vector signed short *vYCoeffsBank, *vCCoeffsBank;
457 #endif
458 
459 #if ARCH_BFIN
460  DECLARE_ALIGNED(4, uint32_t, oy);
461  DECLARE_ALIGNED(4, uint32_t, oc);
462  DECLARE_ALIGNED(4, uint32_t, zero);
463  DECLARE_ALIGNED(4, uint32_t, cy);
464  DECLARE_ALIGNED(4, uint32_t, crv);
465  DECLARE_ALIGNED(4, uint32_t, rmask);
466  DECLARE_ALIGNED(4, uint32_t, cbu);
467  DECLARE_ALIGNED(4, uint32_t, bmask);
468  DECLARE_ALIGNED(4, uint32_t, cgu);
469  DECLARE_ALIGNED(4, uint32_t, cgv);
470  DECLARE_ALIGNED(4, uint32_t, gmask);
471 #endif
472 
473 #if HAVE_VIS
474  DECLARE_ALIGNED(8, uint64_t, sparc_coeffs)[10];
475 #endif
477 
478 /* pre defined color-spaces gamma */
479 #define XYZ_GAMMA (2.6f)
480 #define RGB_GAMMA (2.2f)
481  int16_t *xyzgamma;
482  int16_t *rgbgamma;
483  int16_t xyz2rgb_matrix[3][4];
484 
485  /* function pointers for swScale() */
493 
494  /// Unscaled conversion of luma plane to YV12 for horizontal scaler.
495  void (*lumToYV12)(uint8_t *dst, const uint8_t *src, const uint8_t *src2, const uint8_t *src3,
496  int width, uint32_t *pal);
497  /// Unscaled conversion of alpha plane to YV12 for horizontal scaler.
498  void (*alpToYV12)(uint8_t *dst, const uint8_t *src, const uint8_t *src2, const uint8_t *src3,
499  int width, uint32_t *pal);
500  /// Unscaled conversion of chroma planes to YV12 for horizontal scaler.
501  void (*chrToYV12)(uint8_t *dstU, uint8_t *dstV,
502  const uint8_t *src1, const uint8_t *src2, const uint8_t *src3,
503  int width, uint32_t *pal);
504 
505  /**
506  * Functions to read planar input, such as planar RGB, and convert
507  * internally to Y/UV/A.
508  */
509  /** @{ */
510  void (*readLumPlanar)(uint8_t *dst, const uint8_t *src[4], int width, int32_t *rgb2yuv);
511  void (*readChrPlanar)(uint8_t *dstU, uint8_t *dstV, const uint8_t *src[4],
512  int width, int32_t *rgb2yuv);
513  void (*readAlpPlanar)(uint8_t *dst, const uint8_t *src[4], int width, int32_t *rgb2yuv);
514  /** @} */
515 
516  /**
517  * Scale one horizontal line of input data using a bilinear filter
518  * to produce one line of output data. Compared to SwsContext->hScale(),
519  * please take note of the following caveats when using these:
520  * - Scaling is done using only 7bit instead of 14bit coefficients.
521  * - You can use no more than 5 input pixels to produce 4 output
522  * pixels. Therefore, this filter should not be used for downscaling
523  * by more than ~20% in width (because that equals more than 5/4th
524  * downscaling and thus more than 5 pixels input per 4 pixels output).
525  * - In general, bilinear filters create artifacts during downscaling
526  * (even when <20%), because one output pixel will span more than one
527  * input pixel, and thus some pixels will need edges of both neighbor
528  * pixels to interpolate the output pixel. Since you can use at most
529  * two input pixels per output pixel in bilinear scaling, this is
530  * impossible and thus downscaling by any size will create artifacts.
531  * To enable this type of scaling, set SWS_FLAG_FAST_BILINEAR
532  * in SwsContext->flags.
533  */
534  /** @{ */
536  int16_t *dst, int dstWidth,
537  const uint8_t *src, int srcW, int xInc);
539  int16_t *dst1, int16_t *dst2, int dstWidth,
540  const uint8_t *src1, const uint8_t *src2,
541  int srcW, int xInc);
542  /** @} */
543 
544  /**
545  * Scale one horizontal line of input data using a filter over the input
546  * lines, to produce one (differently sized) line of output data.
547  *
548  * @param dst pointer to destination buffer for horizontally scaled
549  * data. If the number of bits per component of one
550  * destination pixel (SwsContext->dstBpc) is <= 10, data
551  * will be 15bpc in 16bits (int16_t) width. Else (i.e.
552  * SwsContext->dstBpc == 16), data will be 19bpc in
553  * 32bits (int32_t) width.
554  * @param dstW width of destination image
555  * @param src pointer to source data to be scaled. If the number of
556  * bits per component of a source pixel (SwsContext->srcBpc)
557  * is 8, this is 8bpc in 8bits (uint8_t) width. Else
558  * (i.e. SwsContext->dstBpc > 8), this is native depth
559  * in 16bits (uint16_t) width. In other words, for 9-bit
560  * YUV input, this is 9bpc, for 10-bit YUV input, this is
561  * 10bpc, and for 16-bit RGB or YUV, this is 16bpc.
562  * @param filter filter coefficients to be used per output pixel for
563  * scaling. This contains 14bpp filtering coefficients.
564  * Guaranteed to contain dstW * filterSize entries.
565  * @param filterPos position of the first input pixel to be used for
566  * each output pixel during scaling. Guaranteed to
567  * contain dstW entries.
568  * @param filterSize the number of input coefficients to be used (and
569  * thus the number of input pixels to be used) for
570  * creating a single output pixel. Is aligned to 4
571  * (and input coefficients thus padded with zeroes)
572  * to simplify creating SIMD code.
573  */
574  /** @{ */
575  void (*hyScale)(struct SwsContext *c, int16_t *dst, int dstW,
576  const uint8_t *src, const int16_t *filter,
577  const int32_t *filterPos, int filterSize);
578  void (*hcScale)(struct SwsContext *c, int16_t *dst, int dstW,
579  const uint8_t *src, const int16_t *filter,
580  const int32_t *filterPos, int filterSize);
581  /** @} */
582 
583  /// Color range conversion function for luma plane if needed.
584  void (*lumConvertRange)(int16_t *dst, int width);
585  /// Color range conversion function for chroma planes if needed.
586  void (*chrConvertRange)(int16_t *dst1, int16_t *dst2, int width);
587 
588  int needs_hcscale; ///< Set if there are chroma planes to be converted.
589 } SwsContext;
590 //FIXME check init (where 0)
591 
593 int ff_yuv2rgb_c_init_tables(SwsContext *c, const int inv_table[4],
594  int fullRange, int brightness,
595  int contrast, int saturation);
596 
597 void ff_yuv2rgb_init_tables_altivec(SwsContext *c, const int inv_table[4],
598  int brightness, int contrast, int saturation);
600  int lastInLumBuf, int lastInChrBuf);
601 
607 
608 #if FF_API_SWS_FORMAT_NAME
609 /**
610  * @deprecated Use av_get_pix_fmt_name() instead.
611  */
613 const char *sws_format_name(enum AVPixelFormat format);
614 #endif
615 
617 {
618  const AVPixFmtDescriptor *desc = av_pix_fmt_desc_get(pix_fmt);
619  av_assert0(desc);
620  return desc->comp[0].depth_minus1 == 15;
621 }
622 
624 {
625  const AVPixFmtDescriptor *desc = av_pix_fmt_desc_get(pix_fmt);
626  av_assert0(desc);
627  return desc->comp[0].depth_minus1 >= 8 && desc->comp[0].depth_minus1 <= 13;
628 }
629 
630 #define isNBPS(x) is9_OR_10BPS(x)
631 
633 {
634  const AVPixFmtDescriptor *desc = av_pix_fmt_desc_get(pix_fmt);
635  av_assert0(desc);
636  return desc->flags & AV_PIX_FMT_FLAG_BE;
637 }
638 
640 {
641  const AVPixFmtDescriptor *desc = av_pix_fmt_desc_get(pix_fmt);
642  av_assert0(desc);
643  return !(desc->flags & AV_PIX_FMT_FLAG_RGB) && desc->nb_components >= 2;
644 }
645 
647 {
648  const AVPixFmtDescriptor *desc = av_pix_fmt_desc_get(pix_fmt);
649  av_assert0(desc);
650  return ((desc->flags & AV_PIX_FMT_FLAG_PLANAR) && isYUV(pix_fmt));
651 }
652 
654 {
655  const AVPixFmtDescriptor *desc = av_pix_fmt_desc_get(pix_fmt);
656  av_assert0(desc);
657  return (desc->flags & AV_PIX_FMT_FLAG_RGB);
658 }
659 
660 #if 0 // FIXME
661 #define isGray(x) \
662  (!(av_pix_fmt_desc_get(x)->flags & AV_PIX_FMT_FLAG_PAL) && \
663  av_pix_fmt_desc_get(x)->nb_components <= 2)
664 #else
665 #define isGray(x) \
666  ((x) == AV_PIX_FMT_GRAY8 || \
667  (x) == AV_PIX_FMT_Y400A || \
668  (x) == AV_PIX_FMT_GRAY16BE || \
669  (x) == AV_PIX_FMT_GRAY16LE)
670 #endif
671 
672 #define isRGBinInt(x) \
673  ( \
674  (x) == AV_PIX_FMT_RGB48BE || \
675  (x) == AV_PIX_FMT_RGB48LE || \
676  (x) == AV_PIX_FMT_RGBA64BE || \
677  (x) == AV_PIX_FMT_RGBA64LE || \
678  (x) == AV_PIX_FMT_RGB32 || \
679  (x) == AV_PIX_FMT_RGB32_1 || \
680  (x) == AV_PIX_FMT_RGB24 || \
681  (x) == AV_PIX_FMT_RGB565BE || \
682  (x) == AV_PIX_FMT_RGB565LE || \
683  (x) == AV_PIX_FMT_RGB555BE || \
684  (x) == AV_PIX_FMT_RGB555LE || \
685  (x) == AV_PIX_FMT_RGB444BE || \
686  (x) == AV_PIX_FMT_RGB444LE || \
687  (x) == AV_PIX_FMT_RGB8 || \
688  (x) == AV_PIX_FMT_RGB4 || \
689  (x) == AV_PIX_FMT_RGB4_BYTE || \
690  (x) == AV_PIX_FMT_MONOBLACK || \
691  (x) == AV_PIX_FMT_MONOWHITE \
692  )
693 #define isBGRinInt(x) \
694  ( \
695  (x) == AV_PIX_FMT_BGR48BE || \
696  (x) == AV_PIX_FMT_BGR48LE || \
697  (x) == AV_PIX_FMT_BGRA64BE || \
698  (x) == AV_PIX_FMT_BGRA64LE || \
699  (x) == AV_PIX_FMT_BGR32 || \
700  (x) == AV_PIX_FMT_BGR32_1 || \
701  (x) == AV_PIX_FMT_BGR24 || \
702  (x) == AV_PIX_FMT_BGR565BE || \
703  (x) == AV_PIX_FMT_BGR565LE || \
704  (x) == AV_PIX_FMT_BGR555BE || \
705  (x) == AV_PIX_FMT_BGR555LE || \
706  (x) == AV_PIX_FMT_BGR444BE || \
707  (x) == AV_PIX_FMT_BGR444LE || \
708  (x) == AV_PIX_FMT_BGR8 || \
709  (x) == AV_PIX_FMT_BGR4 || \
710  (x) == AV_PIX_FMT_BGR4_BYTE || \
711  (x) == AV_PIX_FMT_MONOBLACK || \
712  (x) == AV_PIX_FMT_MONOWHITE \
713  )
714 
715 #define isRGBinBytes(x) ( \
716  (x) == AV_PIX_FMT_RGB48BE \
717  || (x) == AV_PIX_FMT_RGB48LE \
718  || (x) == AV_PIX_FMT_RGBA64BE \
719  || (x) == AV_PIX_FMT_RGBA64LE \
720  || (x) == AV_PIX_FMT_RGBA \
721  || (x) == AV_PIX_FMT_ARGB \
722  || (x) == AV_PIX_FMT_RGB24 \
723  )
724 #define isBGRinBytes(x) ( \
725  (x) == AV_PIX_FMT_BGR48BE \
726  || (x) == AV_PIX_FMT_BGR48LE \
727  || (x) == AV_PIX_FMT_BGRA64BE \
728  || (x) == AV_PIX_FMT_BGRA64LE \
729  || (x) == AV_PIX_FMT_BGRA \
730  || (x) == AV_PIX_FMT_ABGR \
731  || (x) == AV_PIX_FMT_BGR24 \
732  )
733 
734 #define isAnyRGB(x) \
735  ( \
736  isRGBinInt(x) || \
737  isBGRinInt(x) || \
738  isRGB(x) \
739  )
740 
742 {
743  const AVPixFmtDescriptor *desc = av_pix_fmt_desc_get(pix_fmt);
744  av_assert0(desc);
745  if (pix_fmt == AV_PIX_FMT_PAL8)
746  return 1;
747  return desc->flags & AV_PIX_FMT_FLAG_ALPHA;
748 }
749 
750 #if 1
751 #define isPacked(x) ( \
752  (x)==AV_PIX_FMT_PAL8 \
753  || (x)==AV_PIX_FMT_YUYV422 \
754  || (x)==AV_PIX_FMT_UYVY422 \
755  || (x)==AV_PIX_FMT_Y400A \
756  || isRGBinInt(x) \
757  || isBGRinInt(x) \
758  )
759 #else
761 {
762  const AVPixFmtDescriptor *desc = av_pix_fmt_desc_get(pix_fmt);
763  av_assert0(desc);
764  return ((desc->nb_components >= 2 && !(desc->flags & AV_PIX_FMT_FLAG_PLANAR)) ||
765  pix_fmt == AV_PIX_FMT_PAL8);
766 }
767 
768 #endif
770 {
771  const AVPixFmtDescriptor *desc = av_pix_fmt_desc_get(pix_fmt);
772  av_assert0(desc);
773  return (desc->nb_components >= 2 && (desc->flags & AV_PIX_FMT_FLAG_PLANAR));
774 }
775 
777 {
778  const AVPixFmtDescriptor *desc = av_pix_fmt_desc_get(pix_fmt);
779  av_assert0(desc);
781 }
782 
784 {
785  const AVPixFmtDescriptor *desc = av_pix_fmt_desc_get(pix_fmt);
786  av_assert0(desc);
787  return ((desc->flags & (AV_PIX_FMT_FLAG_PLANAR | AV_PIX_FMT_FLAG_RGB)) ==
789 }
790 
792 {
793  const AVPixFmtDescriptor *desc = av_pix_fmt_desc_get(pix_fmt);
794  av_assert0(desc);
795  return (desc->flags & AV_PIX_FMT_FLAG_PAL) || (desc->flags & AV_PIX_FMT_FLAG_PSEUDOPAL);
796 }
797 
798 extern const uint64_t ff_dither4[2];
799 extern const uint64_t ff_dither8[2];
800 extern const uint8_t dithers[8][8][8];
801 extern const uint16_t dither_scale[15][16];
802 
803 
804 extern const AVClass sws_context_class;
805 
806 /**
807  * Set c->swScale to an unscaled converter if one exists for the specific
808  * source and destination formats, bit depths, flags, etc.
809  */
811 
813 
814 /**
815  * Return function pointer to fastest main scaler path function depending
816  * on architecture and available optimizations.
817  */
819 
831 
832 static inline void fillPlane16(uint8_t *plane, int stride, int width, int height, int y,
833  int alpha, int bits, const int big_endian)
834 {
835  int i, j;
836  uint8_t *ptr = plane + stride * y;
837  int v = alpha ? 0xFFFF>>(15-bits) : (1<<bits);
838  for (i = 0; i < height; i++) {
839 #define FILL(wfunc) \
840  for (j = 0; j < width; j++) {\
841  wfunc(ptr+2*j, v);\
842  }
843  if (big_endian) {
844  FILL(AV_WB16);
845  } else {
846  FILL(AV_WL16);
847  }
848  ptr += stride;
849  }
850 }
851 
852 #endif /* SWSCALE_SWSCALE_INTERNAL_H */