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 SWS_MAX_FILTER_SIZE
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 enum SwsDither {
72 } SwsDither;
73 
74 typedef int (*SwsFunc)(struct SwsContext *context, const uint8_t *src[],
75  int srcStride[], int srcSliceY, int srcSliceH,
76  uint8_t *dst[], int dstStride[]);
77 
78 /**
79  * Write one line of horizontally scaled data to planar output
80  * without any additional vertical scaling (or point-scaling).
81  *
82  * @param src scaled source data, 15bit for 8-10bit output,
83  * 19-bit for 16bit output (in int32_t)
84  * @param dest pointer to the output plane. For >8bit
85  * output, this is in uint16_t
86  * @param dstW width of destination in pixels
87  * @param dither ordered dither array of type int16_t and size 8
88  * @param offset Dither offset
89  */
90 typedef void (*yuv2planar1_fn)(const int16_t *src, uint8_t *dest, int dstW,
91  const uint8_t *dither, int offset);
92 
93 /**
94  * Write one line of horizontally scaled data to planar output
95  * with multi-point vertical scaling between input pixels.
96  *
97  * @param filter vertical luma/alpha scaling coefficients, 12bit [0,4096]
98  * @param src scaled luma (Y) or alpha (A) source data, 15bit for 8-10bit output,
99  * 19-bit for 16bit output (in int32_t)
100  * @param filterSize number of vertical input lines to scale
101  * @param dest pointer to output plane. For >8bit
102  * output, this is in uint16_t
103  * @param dstW width of destination pixels
104  * @param offset Dither offset
105  */
106 typedef void (*yuv2planarX_fn)(const int16_t *filter, int filterSize,
107  const int16_t **src, uint8_t *dest, int dstW,
108  const uint8_t *dither, int offset);
109 
110 /**
111  * Write one line of horizontally scaled chroma to interleaved output
112  * with multi-point vertical scaling between input pixels.
113  *
114  * @param c SWS scaling context
115  * @param chrFilter vertical chroma scaling coefficients, 12bit [0,4096]
116  * @param chrUSrc scaled chroma (U) source data, 15bit for 8-10bit output,
117  * 19-bit for 16bit output (in int32_t)
118  * @param chrVSrc scaled chroma (V) source data, 15bit for 8-10bit output,
119  * 19-bit for 16bit output (in int32_t)
120  * @param chrFilterSize number of vertical chroma input lines to scale
121  * @param dest pointer to the output plane. For >8bit
122  * output, this is in uint16_t
123  * @param dstW width of chroma planes
124  */
126  const int16_t *chrFilter,
127  int chrFilterSize,
128  const int16_t **chrUSrc,
129  const int16_t **chrVSrc,
130  uint8_t *dest, int dstW);
131 
132 /**
133  * Write one line of horizontally scaled Y/U/V/A to packed-pixel YUV/RGB
134  * output without any additional vertical scaling (or point-scaling). Note
135  * that this function may do chroma scaling, see the "uvalpha" argument.
136  *
137  * @param c SWS scaling context
138  * @param lumSrc scaled luma (Y) source data, 15bit for 8-10bit output,
139  * 19-bit for 16bit output (in int32_t)
140  * @param chrUSrc scaled chroma (U) source data, 15bit for 8-10bit output,
141  * 19-bit for 16bit output (in int32_t)
142  * @param chrVSrc scaled chroma (V) source data, 15bit for 8-10bit output,
143  * 19-bit for 16bit output (in int32_t)
144  * @param alpSrc scaled alpha (A) source data, 15bit for 8-10bit output,
145  * 19-bit for 16bit output (in int32_t)
146  * @param dest pointer to the output plane. For 16bit output, this is
147  * uint16_t
148  * @param dstW width of lumSrc and alpSrc in pixels, number of pixels
149  * to write into dest[]
150  * @param uvalpha chroma scaling coefficient for the second line of chroma
151  * pixels, either 2048 or 0. If 0, one chroma input is used
152  * for 2 output pixels (or if the SWS_FLAG_FULL_CHR_INT flag
153  * is set, it generates 1 output pixel). If 2048, two chroma
154  * input pixels should be averaged for 2 output pixels (this
155  * only happens if SWS_FLAG_FULL_CHR_INT is not set)
156  * @param y vertical line number for this output. This does not need
157  * to be used to calculate the offset in the destination,
158  * but can be used to generate comfort noise using dithering
159  * for some output formats.
160  */
161 typedef void (*yuv2packed1_fn)(struct SwsContext *c, const int16_t *lumSrc,
162  const int16_t *chrUSrc[2],
163  const int16_t *chrVSrc[2],
164  const int16_t *alpSrc, uint8_t *dest,
165  int dstW, int uvalpha, int y);
166 /**
167  * Write one line of horizontally scaled Y/U/V/A to packed-pixel YUV/RGB
168  * output by doing bilinear scaling between two input lines.
169  *
170  * @param c SWS scaling context
171  * @param lumSrc scaled luma (Y) source data, 15bit for 8-10bit output,
172  * 19-bit for 16bit output (in int32_t)
173  * @param chrUSrc scaled chroma (U) source data, 15bit for 8-10bit output,
174  * 19-bit for 16bit output (in int32_t)
175  * @param chrVSrc scaled chroma (V) source data, 15bit for 8-10bit output,
176  * 19-bit for 16bit output (in int32_t)
177  * @param alpSrc scaled alpha (A) source data, 15bit for 8-10bit output,
178  * 19-bit for 16bit output (in int32_t)
179  * @param dest pointer to the output plane. For 16bit output, this is
180  * uint16_t
181  * @param dstW width of lumSrc and alpSrc in pixels, number of pixels
182  * to write into dest[]
183  * @param yalpha luma/alpha scaling coefficients for the second input line.
184  * The first line's coefficients can be calculated by using
185  * 4096 - yalpha
186  * @param uvalpha chroma scaling coefficient for the second input line. The
187  * first line's coefficients can be calculated by using
188  * 4096 - uvalpha
189  * @param y vertical line number for this output. This does not need
190  * to be used to calculate the offset in the destination,
191  * but can be used to generate comfort noise using dithering
192  * for some output formats.
193  */
194 typedef void (*yuv2packed2_fn)(struct SwsContext *c, const int16_t *lumSrc[2],
195  const int16_t *chrUSrc[2],
196  const int16_t *chrVSrc[2],
197  const int16_t *alpSrc[2],
198  uint8_t *dest,
199  int dstW, int yalpha, int uvalpha, int y);
200 /**
201  * Write one line of horizontally scaled Y/U/V/A to packed-pixel YUV/RGB
202  * output by doing multi-point vertical scaling between input pixels.
203  *
204  * @param c SWS scaling context
205  * @param lumFilter vertical luma/alpha scaling coefficients, 12bit [0,4096]
206  * @param lumSrc scaled luma (Y) source data, 15bit for 8-10bit output,
207  * 19-bit for 16bit output (in int32_t)
208  * @param lumFilterSize number of vertical luma/alpha input lines to scale
209  * @param chrFilter vertical chroma scaling coefficients, 12bit [0,4096]
210  * @param chrUSrc scaled chroma (U) source data, 15bit for 8-10bit output,
211  * 19-bit for 16bit output (in int32_t)
212  * @param chrVSrc scaled chroma (V) source data, 15bit for 8-10bit output,
213  * 19-bit for 16bit output (in int32_t)
214  * @param chrFilterSize number of vertical chroma input lines to scale
215  * @param alpSrc scaled alpha (A) source data, 15bit for 8-10bit output,
216  * 19-bit for 16bit output (in int32_t)
217  * @param dest pointer to the output plane. For 16bit output, this is
218  * uint16_t
219  * @param dstW width of lumSrc and alpSrc in pixels, number of pixels
220  * to write into dest[]
221  * @param y vertical line number for this output. This does not need
222  * to be used to calculate the offset in the destination,
223  * but can be used to generate comfort noise using dithering
224  * or some output formats.
225  */
226 typedef void (*yuv2packedX_fn)(struct SwsContext *c, const int16_t *lumFilter,
227  const int16_t **lumSrc, int lumFilterSize,
228  const int16_t *chrFilter,
229  const int16_t **chrUSrc,
230  const int16_t **chrVSrc, int chrFilterSize,
231  const int16_t **alpSrc, uint8_t *dest,
232  int dstW, int y);
233 
234 /**
235  * Write one line of horizontally scaled Y/U/V/A to YUV/RGB
236  * output by doing multi-point vertical scaling between input pixels.
237  *
238  * @param c SWS scaling context
239  * @param lumFilter vertical luma/alpha scaling coefficients, 12bit [0,4096]
240  * @param lumSrc scaled luma (Y) source data, 15bit for 8-10bit output,
241  * 19-bit for 16bit output (in int32_t)
242  * @param lumFilterSize number of vertical luma/alpha input lines to scale
243  * @param chrFilter vertical chroma scaling coefficients, 12bit [0,4096]
244  * @param chrUSrc scaled chroma (U) source data, 15bit for 8-10bit output,
245  * 19-bit for 16bit output (in int32_t)
246  * @param chrVSrc scaled chroma (V) source data, 15bit for 8-10bit output,
247  * 19-bit for 16bit output (in int32_t)
248  * @param chrFilterSize number of vertical chroma input lines to scale
249  * @param alpSrc scaled alpha (A) source data, 15bit for 8-10bit output,
250  * 19-bit for 16bit output (in int32_t)
251  * @param dest pointer to the output planes. For 16bit output, this is
252  * uint16_t
253  * @param dstW width of lumSrc and alpSrc in pixels, number of pixels
254  * to write into dest[]
255  * @param y vertical line number for this output. This does not need
256  * to be used to calculate the offset in the destination,
257  * but can be used to generate comfort noise using dithering
258  * or some output formats.
259  */
260 typedef void (*yuv2anyX_fn)(struct SwsContext *c, const int16_t *lumFilter,
261  const int16_t **lumSrc, int lumFilterSize,
262  const int16_t *chrFilter,
263  const int16_t **chrUSrc,
264  const int16_t **chrVSrc, int chrFilterSize,
265  const int16_t **alpSrc, uint8_t **dest,
266  int dstW, int y);
267 
268 /* This struct should be aligned on at least a 32-byte boundary. */
269 typedef struct SwsContext {
270  /**
271  * info on struct for av_log
272  */
274 
275  /**
276  * Note that src, dst, srcStride, dstStride will be copied in the
277  * sws_scale() wrapper so they can be freely modified here.
278  */
280  int srcW; ///< Width of source luma/alpha planes.
281  int srcH; ///< Height of source luma/alpha planes.
282  int dstH; ///< Height of destination luma/alpha planes.
283  int chrSrcW; ///< Width of source chroma planes.
284  int chrSrcH; ///< Height of source chroma planes.
285  int chrDstW; ///< Width of destination chroma planes.
286  int chrDstH; ///< Height of destination chroma planes.
289  enum AVPixelFormat dstFormat; ///< Destination pixel format.
290  enum AVPixelFormat srcFormat; ///< Source pixel format.
291  int dstFormatBpp; ///< Number of bits per pixel of the destination pixel format.
292  int srcFormatBpp; ///< Number of bits per pixel of the source pixel format.
294  int chrSrcHSubSample; ///< Binary logarithm of horizontal subsampling factor between luma/alpha and chroma planes in source image.
295  int chrSrcVSubSample; ///< Binary logarithm of vertical subsampling factor between luma/alpha and chroma planes in source image.
296  int chrDstHSubSample; ///< Binary logarithm of horizontal subsampling factor between luma/alpha and chroma planes in destination image.
297  int chrDstVSubSample; ///< Binary logarithm of vertical subsampling factor between luma/alpha and chroma planes in destination image.
298  int vChrDrop; ///< Binary logarithm of extra vertical subsampling factor in source image chroma planes specified by user.
299  int sliceDir; ///< Direction that slices are fed to the scaler (1 = top-to-bottom, -1 = bottom-to-top).
300  double param[2]; ///< Input parameters for scaling algorithms that need them.
301 
302  uint32_t pal_yuv[256];
303  uint32_t pal_rgb[256];
304 
305  /**
306  * @name Scaled horizontal lines ring buffer.
307  * The horizontal scaler keeps just enough scaled lines in a ring buffer
308  * so they may be passed to the vertical scaler. The pointers to the
309  * allocated buffers for each line are duplicated in sequence in the ring
310  * buffer to simplify indexing and avoid wrapping around between lines
311  * inside the vertical scaler code. The wrapping is done before the
312  * vertical scaler is called.
313  */
314  //@{
315  int16_t **lumPixBuf; ///< Ring buffer for scaled horizontal luma plane lines to be fed to the vertical scaler.
316  int16_t **chrUPixBuf; ///< Ring buffer for scaled horizontal chroma plane lines to be fed to the vertical scaler.
317  int16_t **chrVPixBuf; ///< Ring buffer for scaled horizontal chroma plane lines to be fed to the vertical scaler.
318  int16_t **alpPixBuf; ///< Ring buffer for scaled horizontal alpha plane lines to be fed to the vertical scaler.
319  int vLumBufSize; ///< Number of vertical luma/alpha lines allocated in the ring buffer.
320  int vChrBufSize; ///< Number of vertical chroma lines allocated in the ring buffer.
321  int lastInLumBuf; ///< Last scaled horizontal luma/alpha line from source in the ring buffer.
322  int lastInChrBuf; ///< Last scaled horizontal chroma line from source in the ring buffer.
323  int lumBufIndex; ///< Index in ring buffer of the last scaled horizontal luma/alpha line from source.
324  int chrBufIndex; ///< Index in ring buffer of the last scaled horizontal chroma line from source.
325  //@}
326 
328 
329  /**
330  * @name Horizontal and vertical filters.
331  * To better understand the following fields, here is a pseudo-code of
332  * their usage in filtering a horizontal line:
333  * @code
334  * for (i = 0; i < width; i++) {
335  * dst[i] = 0;
336  * for (j = 0; j < filterSize; j++)
337  * dst[i] += src[ filterPos[i] + j ] * filter[ filterSize * i + j ];
338  * dst[i] >>= FRAC_BITS; // The actual implementation is fixed-point.
339  * }
340  * @endcode
341  */
342  //@{
343  int16_t *hLumFilter; ///< Array of horizontal filter coefficients for luma/alpha planes.
344  int16_t *hChrFilter; ///< Array of horizontal filter coefficients for chroma planes.
345  int16_t *vLumFilter; ///< Array of vertical filter coefficients for luma/alpha planes.
346  int16_t *vChrFilter; ///< Array of vertical filter coefficients for chroma planes.
347  int32_t *hLumFilterPos; ///< Array of horizontal filter starting positions for each dst[i] for luma/alpha planes.
348  int32_t *hChrFilterPos; ///< Array of horizontal filter starting positions for each dst[i] for chroma planes.
349  int32_t *vLumFilterPos; ///< Array of vertical filter starting positions for each dst[i] for luma/alpha planes.
350  int32_t *vChrFilterPos; ///< Array of vertical filter starting positions for each dst[i] for chroma planes.
351  int hLumFilterSize; ///< Horizontal filter size for luma/alpha pixels.
352  int hChrFilterSize; ///< Horizontal filter size for chroma pixels.
353  int vLumFilterSize; ///< Vertical filter size for luma/alpha pixels.
354  int vChrFilterSize; ///< Vertical filter size for chroma pixels.
355  //@}
356 
357  int lumMmxextFilterCodeSize; ///< Runtime-generated MMXEXT horizontal fast bilinear scaler code size for luma/alpha planes.
358  int chrMmxextFilterCodeSize; ///< Runtime-generated MMXEXT horizontal fast bilinear scaler code size for chroma planes.
359  uint8_t *lumMmxextFilterCode; ///< Runtime-generated MMXEXT horizontal fast bilinear scaler code for luma/alpha planes.
360  uint8_t *chrMmxextFilterCode; ///< Runtime-generated MMXEXT horizontal fast bilinear scaler code for chroma planes.
361 
363 
364  int dstY; ///< Last destination vertical line output from last slice.
365  int flags; ///< Flags passed by the user to select scaler algorithm, optimizations, subsampling, etc...
366  void *yuvTable; // pointer to the yuv->rgb table start so it can be freed()
367  // alignment ensures the offset can be added in a single
368  // instruction on e.g. ARM
373  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
374 #define RY_IDX 0
375 #define GY_IDX 1
376 #define BY_IDX 2
377 #define RU_IDX 3
378 #define GU_IDX 4
379 #define BU_IDX 5
380 #define RV_IDX 6
381 #define GV_IDX 7
382 #define BV_IDX 8
383 #define RGB2YUV_SHIFT 15
384 
385  int *dither_error[4];
386 
387  //Colorspace stuff
388  int contrast, brightness, saturation; // for sws_getColorspaceDetails
391  int srcRange; ///< 0 = MPG YUV range, 1 = JPG YUV range (source image).
392  int dstRange; ///< 0 = MPG YUV range, 1 = JPG YUV range (destination image).
395  int srcXYZ;
396  int dstXYZ;
407 
408 #define RED_DITHER "0*8"
409 #define GREEN_DITHER "1*8"
410 #define BLUE_DITHER "2*8"
411 #define Y_COEFF "3*8"
412 #define VR_COEFF "4*8"
413 #define UB_COEFF "5*8"
414 #define VG_COEFF "6*8"
415 #define UG_COEFF "7*8"
416 #define Y_OFFSET "8*8"
417 #define U_OFFSET "9*8"
418 #define V_OFFSET "10*8"
419 #define LUM_MMX_FILTER_OFFSET "11*8"
420 #define CHR_MMX_FILTER_OFFSET "11*8+4*4*"AV_STRINGIFY(MAX_FILTER_SIZE)
421 #define DSTW_OFFSET "11*8+4*4*"AV_STRINGIFY(MAX_FILTER_SIZE)"*2"
422 #define ESP_OFFSET "11*8+4*4*"AV_STRINGIFY(MAX_FILTER_SIZE)"*2+8"
423 #define VROUNDER_OFFSET "11*8+4*4*"AV_STRINGIFY(MAX_FILTER_SIZE)"*2+16"
424 #define U_TEMP "11*8+4*4*"AV_STRINGIFY(MAX_FILTER_SIZE)"*2+24"
425 #define V_TEMP "11*8+4*4*"AV_STRINGIFY(MAX_FILTER_SIZE)"*2+32"
426 #define Y_TEMP "11*8+4*4*"AV_STRINGIFY(MAX_FILTER_SIZE)"*2+40"
427 #define ALP_MMX_FILTER_OFFSET "11*8+4*4*"AV_STRINGIFY(MAX_FILTER_SIZE)"*2+48"
428 #define UV_OFF_PX "11*8+4*4*"AV_STRINGIFY(MAX_FILTER_SIZE)"*3+48"
429 #define UV_OFF_BYTE "11*8+4*4*"AV_STRINGIFY(MAX_FILTER_SIZE)"*3+56"
430 #define DITHER16 "11*8+4*4*"AV_STRINGIFY(MAX_FILTER_SIZE)"*3+64"
431 #define DITHER32 "11*8+4*4*"AV_STRINGIFY(MAX_FILTER_SIZE)"*3+80"
432 #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
433 
434  DECLARE_ALIGNED(8, uint64_t, redDither);
437 
438  DECLARE_ALIGNED(8, uint64_t, yCoeff);
439  DECLARE_ALIGNED(8, uint64_t, vrCoeff);
440  DECLARE_ALIGNED(8, uint64_t, ubCoeff);
441  DECLARE_ALIGNED(8, uint64_t, vgCoeff);
442  DECLARE_ALIGNED(8, uint64_t, ugCoeff);
443  DECLARE_ALIGNED(8, uint64_t, yOffset);
444  DECLARE_ALIGNED(8, uint64_t, uOffset);
445  DECLARE_ALIGNED(8, uint64_t, vOffset);
448  int dstW; ///< Width of destination luma/alpha planes.
449  DECLARE_ALIGNED(8, uint64_t, esp);
450  DECLARE_ALIGNED(8, uint64_t, vRounder);
451  DECLARE_ALIGNED(8, uint64_t, u_temp);
452  DECLARE_ALIGNED(8, uint64_t, v_temp);
453  DECLARE_ALIGNED(8, uint64_t, y_temp);
455  // alignment of these values is not necessary, but merely here
456  // to maintain the same offset across x8632 and x86-64. Once we
457  // use proper offset macros in the asm, they can be removed.
458  DECLARE_ALIGNED(8, ptrdiff_t, uv_off); ///< offset (in pixels) between u and v planes
459  DECLARE_ALIGNED(8, ptrdiff_t, uv_offx2); ///< offset (in bytes) between u and v planes
460  DECLARE_ALIGNED(8, uint16_t, dither16)[8];
461  DECLARE_ALIGNED(8, uint32_t, dither32)[8];
462 
464 
465 #if HAVE_ALTIVEC
466  vector signed short CY;
467  vector signed short CRV;
468  vector signed short CBU;
469  vector signed short CGU;
470  vector signed short CGV;
471  vector signed short OY;
472  vector unsigned short CSHIFT;
473  vector signed short *vYCoeffsBank, *vCCoeffsBank;
474 #endif
475 
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 *xyzgammainv;
484  int16_t *rgbgammainv;
485  int16_t xyz2rgb_matrix[3][4];
486  int16_t rgb2xyz_matrix[3][4];
487 
488  /* function pointers for swscale() */
496 
497  /// Unscaled conversion of luma plane to YV12 for horizontal scaler.
498  void (*lumToYV12)(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 alpha plane to YV12 for horizontal scaler.
501  void (*alpToYV12)(uint8_t *dst, const uint8_t *src, const uint8_t *src2, const uint8_t *src3,
502  int width, uint32_t *pal);
503  /// Unscaled conversion of chroma planes to YV12 for horizontal scaler.
504  void (*chrToYV12)(uint8_t *dstU, uint8_t *dstV,
505  const uint8_t *src1, const uint8_t *src2, const uint8_t *src3,
506  int width, uint32_t *pal);
507 
508  /**
509  * Functions to read planar input, such as planar RGB, and convert
510  * internally to Y/UV/A.
511  */
512  /** @{ */
513  void (*readLumPlanar)(uint8_t *dst, const uint8_t *src[4], int width, int32_t *rgb2yuv);
514  void (*readChrPlanar)(uint8_t *dstU, uint8_t *dstV, const uint8_t *src[4],
515  int width, int32_t *rgb2yuv);
516  void (*readAlpPlanar)(uint8_t *dst, const uint8_t *src[4], int width, int32_t *rgb2yuv);
517  /** @} */
518 
519  /**
520  * Scale one horizontal line of input data using a bilinear filter
521  * to produce one line of output data. Compared to SwsContext->hScale(),
522  * please take note of the following caveats when using these:
523  * - Scaling is done using only 7bit instead of 14bit coefficients.
524  * - You can use no more than 5 input pixels to produce 4 output
525  * pixels. Therefore, this filter should not be used for downscaling
526  * by more than ~20% in width (because that equals more than 5/4th
527  * downscaling and thus more than 5 pixels input per 4 pixels output).
528  * - In general, bilinear filters create artifacts during downscaling
529  * (even when <20%), because one output pixel will span more than one
530  * input pixel, and thus some pixels will need edges of both neighbor
531  * pixels to interpolate the output pixel. Since you can use at most
532  * two input pixels per output pixel in bilinear scaling, this is
533  * impossible and thus downscaling by any size will create artifacts.
534  * To enable this type of scaling, set SWS_FLAG_FAST_BILINEAR
535  * in SwsContext->flags.
536  */
537  /** @{ */
539  int16_t *dst, int dstWidth,
540  const uint8_t *src, int srcW, int xInc);
542  int16_t *dst1, int16_t *dst2, int dstWidth,
543  const uint8_t *src1, const uint8_t *src2,
544  int srcW, int xInc);
545  /** @} */
546 
547  /**
548  * Scale one horizontal line of input data using a filter over the input
549  * lines, to produce one (differently sized) line of output data.
550  *
551  * @param dst pointer to destination buffer for horizontally scaled
552  * data. If the number of bits per component of one
553  * destination pixel (SwsContext->dstBpc) is <= 10, data
554  * will be 15bpc in 16bits (int16_t) width. Else (i.e.
555  * SwsContext->dstBpc == 16), data will be 19bpc in
556  * 32bits (int32_t) width.
557  * @param dstW width of destination image
558  * @param src pointer to source data to be scaled. If the number of
559  * bits per component of a source pixel (SwsContext->srcBpc)
560  * is 8, this is 8bpc in 8bits (uint8_t) width. Else
561  * (i.e. SwsContext->dstBpc > 8), this is native depth
562  * in 16bits (uint16_t) width. In other words, for 9-bit
563  * YUV input, this is 9bpc, for 10-bit YUV input, this is
564  * 10bpc, and for 16-bit RGB or YUV, this is 16bpc.
565  * @param filter filter coefficients to be used per output pixel for
566  * scaling. This contains 14bpp filtering coefficients.
567  * Guaranteed to contain dstW * filterSize entries.
568  * @param filterPos position of the first input pixel to be used for
569  * each output pixel during scaling. Guaranteed to
570  * contain dstW entries.
571  * @param filterSize the number of input coefficients to be used (and
572  * thus the number of input pixels to be used) for
573  * creating a single output pixel. Is aligned to 4
574  * (and input coefficients thus padded with zeroes)
575  * to simplify creating SIMD code.
576  */
577  /** @{ */
578  void (*hyScale)(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  void (*hcScale)(struct SwsContext *c, int16_t *dst, int dstW,
582  const uint8_t *src, const int16_t *filter,
583  const int32_t *filterPos, int filterSize);
584  /** @} */
585 
586  /// Color range conversion function for luma plane if needed.
587  void (*lumConvertRange)(int16_t *dst, int width);
588  /// Color range conversion function for chroma planes if needed.
589  void (*chrConvertRange)(int16_t *dst1, int16_t *dst2, int width);
590 
591  int needs_hcscale; ///< Set if there are chroma planes to be converted.
592 
594 } SwsContext;
595 //FIXME check init (where 0)
596 
598 int ff_yuv2rgb_c_init_tables(SwsContext *c, const int inv_table[4],
599  int fullRange, int brightness,
600  int contrast, int saturation);
601 void ff_yuv2rgb_init_tables_ppc(SwsContext *c, const int inv_table[4],
602  int brightness, int contrast, int saturation);
603 
605  int lastInLumBuf, int lastInChrBuf);
606 
608 
611 
612 #if FF_API_SWS_FORMAT_NAME
613 /**
614  * @deprecated Use av_get_pix_fmt_name() instead.
615  */
617 const char *sws_format_name(enum AVPixelFormat format);
618 #endif
619 
621 {
622  const AVPixFmtDescriptor *desc = av_pix_fmt_desc_get(pix_fmt);
623  av_assert0(desc);
624  return desc->comp[0].depth_minus1 == 15;
625 }
626 
628 {
629  const AVPixFmtDescriptor *desc = av_pix_fmt_desc_get(pix_fmt);
630  av_assert0(desc);
631  return desc->comp[0].depth_minus1 >= 8 && desc->comp[0].depth_minus1 <= 13;
632 }
633 
634 #define isNBPS(x) is9_OR_10BPS(x)
635 
637 {
638  const AVPixFmtDescriptor *desc = av_pix_fmt_desc_get(pix_fmt);
639  av_assert0(desc);
640  return desc->flags & AV_PIX_FMT_FLAG_BE;
641 }
642 
644 {
645  const AVPixFmtDescriptor *desc = av_pix_fmt_desc_get(pix_fmt);
646  av_assert0(desc);
647  return !(desc->flags & AV_PIX_FMT_FLAG_RGB) && desc->nb_components >= 2;
648 }
649 
651 {
652  const AVPixFmtDescriptor *desc = av_pix_fmt_desc_get(pix_fmt);
653  av_assert0(desc);
654  return ((desc->flags & AV_PIX_FMT_FLAG_PLANAR) && isYUV(pix_fmt));
655 }
656 
658 {
659  const AVPixFmtDescriptor *desc = av_pix_fmt_desc_get(pix_fmt);
660  av_assert0(desc);
661  return (desc->flags & AV_PIX_FMT_FLAG_RGB);
662 }
663 
664 #if 0 // FIXME
665 #define isGray(x) \
666  (!(av_pix_fmt_desc_get(x)->flags & AV_PIX_FMT_FLAG_PAL) && \
667  av_pix_fmt_desc_get(x)->nb_components <= 2)
668 #else
669 #define isGray(x) \
670  ((x) == AV_PIX_FMT_GRAY8 || \
671  (x) == AV_PIX_FMT_Y400A || \
672  (x) == AV_PIX_FMT_GRAY16BE || \
673  (x) == AV_PIX_FMT_GRAY16LE)
674 #endif
675 
676 #define isRGBinInt(x) \
677  ( \
678  (x) == AV_PIX_FMT_RGB48BE || \
679  (x) == AV_PIX_FMT_RGB48LE || \
680  (x) == AV_PIX_FMT_RGB32 || \
681  (x) == AV_PIX_FMT_RGB32_1 || \
682  (x) == AV_PIX_FMT_RGB24 || \
683  (x) == AV_PIX_FMT_RGB565BE || \
684  (x) == AV_PIX_FMT_RGB565LE || \
685  (x) == AV_PIX_FMT_RGB555BE || \
686  (x) == AV_PIX_FMT_RGB555LE || \
687  (x) == AV_PIX_FMT_RGB444BE || \
688  (x) == AV_PIX_FMT_RGB444LE || \
689  (x) == AV_PIX_FMT_RGB8 || \
690  (x) == AV_PIX_FMT_RGB4 || \
691  (x) == AV_PIX_FMT_RGB4_BYTE || \
692  (x) == AV_PIX_FMT_RGBA64BE || \
693  (x) == AV_PIX_FMT_RGBA64LE || \
694  (x) == AV_PIX_FMT_MONOBLACK || \
695  (x) == AV_PIX_FMT_MONOWHITE \
696  )
697 #define isBGRinInt(x) \
698  ( \
699  (x) == AV_PIX_FMT_BGR48BE || \
700  (x) == AV_PIX_FMT_BGR48LE || \
701  (x) == AV_PIX_FMT_BGR32 || \
702  (x) == AV_PIX_FMT_BGR32_1 || \
703  (x) == AV_PIX_FMT_BGR24 || \
704  (x) == AV_PIX_FMT_BGR565BE || \
705  (x) == AV_PIX_FMT_BGR565LE || \
706  (x) == AV_PIX_FMT_BGR555BE || \
707  (x) == AV_PIX_FMT_BGR555LE || \
708  (x) == AV_PIX_FMT_BGR444BE || \
709  (x) == AV_PIX_FMT_BGR444LE || \
710  (x) == AV_PIX_FMT_BGR8 || \
711  (x) == AV_PIX_FMT_BGR4 || \
712  (x) == AV_PIX_FMT_BGR4_BYTE || \
713  (x) == AV_PIX_FMT_BGRA64BE || \
714  (x) == AV_PIX_FMT_BGRA64LE || \
715  (x) == AV_PIX_FMT_MONOBLACK || \
716  (x) == AV_PIX_FMT_MONOWHITE \
717  )
718 
719 #define isRGBinBytes(x) ( \
720  (x) == AV_PIX_FMT_RGB48BE \
721  || (x) == AV_PIX_FMT_RGB48LE \
722  || (x) == AV_PIX_FMT_RGBA64BE \
723  || (x) == AV_PIX_FMT_RGBA64LE \
724  || (x) == AV_PIX_FMT_RGBA \
725  || (x) == AV_PIX_FMT_ARGB \
726  || (x) == AV_PIX_FMT_RGB24 \
727  )
728 #define isBGRinBytes(x) ( \
729  (x) == AV_PIX_FMT_BGR48BE \
730  || (x) == AV_PIX_FMT_BGR48LE \
731  || (x) == AV_PIX_FMT_BGRA64BE \
732  || (x) == AV_PIX_FMT_BGRA64LE \
733  || (x) == AV_PIX_FMT_BGRA \
734  || (x) == AV_PIX_FMT_ABGR \
735  || (x) == AV_PIX_FMT_BGR24 \
736  )
737 
738 #define isBayer(x) ( \
739  (x)==AV_PIX_FMT_BAYER_BGGR8 \
740  || (x)==AV_PIX_FMT_BAYER_BGGR16LE \
741  || (x)==AV_PIX_FMT_BAYER_BGGR16BE \
742  || (x)==AV_PIX_FMT_BAYER_RGGB8 \
743  || (x)==AV_PIX_FMT_BAYER_RGGB16LE \
744  || (x)==AV_PIX_FMT_BAYER_RGGB16BE \
745  || (x)==AV_PIX_FMT_BAYER_GBRG8 \
746  || (x)==AV_PIX_FMT_BAYER_GBRG16LE \
747  || (x)==AV_PIX_FMT_BAYER_GBRG16BE \
748  || (x)==AV_PIX_FMT_BAYER_GRBG8 \
749  || (x)==AV_PIX_FMT_BAYER_GRBG16LE \
750  || (x)==AV_PIX_FMT_BAYER_GRBG16BE \
751  )
752 
753 #define isAnyRGB(x) \
754  ( \
755  isBayer(x) || \
756  isRGBinInt(x) || \
757  isBGRinInt(x) || \
758  isRGB(x) \
759  )
760 
762 {
763  const AVPixFmtDescriptor *desc = av_pix_fmt_desc_get(pix_fmt);
764  av_assert0(desc);
765  if (pix_fmt == AV_PIX_FMT_PAL8)
766  return 1;
767  return desc->flags & AV_PIX_FMT_FLAG_ALPHA;
768 }
769 
770 #if 1
771 #define isPacked(x) ( \
772  (x)==AV_PIX_FMT_PAL8 \
773  || (x)==AV_PIX_FMT_YUYV422 \
774  || (x)==AV_PIX_FMT_YVYU422 \
775  || (x)==AV_PIX_FMT_UYVY422 \
776  || (x)==AV_PIX_FMT_Y400A \
777  || isRGBinInt(x) \
778  || isBGRinInt(x) \
779  )
780 #else
782 {
783  const AVPixFmtDescriptor *desc = av_pix_fmt_desc_get(pix_fmt);
784  av_assert0(desc);
785  return ((desc->nb_components >= 2 && !(desc->flags & AV_PIX_FMT_FLAG_PLANAR)) ||
786  pix_fmt == AV_PIX_FMT_PAL8);
787 }
788 
789 #endif
791 {
792  const AVPixFmtDescriptor *desc = av_pix_fmt_desc_get(pix_fmt);
793  av_assert0(desc);
794  return (desc->nb_components >= 2 && (desc->flags & AV_PIX_FMT_FLAG_PLANAR));
795 }
796 
798 {
799  const AVPixFmtDescriptor *desc = av_pix_fmt_desc_get(pix_fmt);
800  av_assert0(desc);
802 }
803 
805 {
806  const AVPixFmtDescriptor *desc = av_pix_fmt_desc_get(pix_fmt);
807  av_assert0(desc);
808  return ((desc->flags & (AV_PIX_FMT_FLAG_PLANAR | AV_PIX_FMT_FLAG_RGB)) ==
810 }
811 
813 {
814  const AVPixFmtDescriptor *desc = av_pix_fmt_desc_get(pix_fmt);
815  av_assert0(desc);
816  return (desc->flags & AV_PIX_FMT_FLAG_PAL) || (desc->flags & AV_PIX_FMT_FLAG_PSEUDOPAL);
817 }
818 
819 extern const uint64_t ff_dither4[2];
820 extern const uint64_t ff_dither8[2];
821 
822 extern const uint8_t ff_dither_2x2_4[3][8];
823 extern const uint8_t ff_dither_2x2_8[3][8];
824 extern const uint8_t ff_dither_4x4_16[5][8];
825 extern const uint8_t ff_dither_8x8_32[9][8];
826 extern const uint8_t ff_dither_8x8_73[9][8];
827 extern const uint8_t ff_dither_8x8_128[9][8];
828 extern const uint8_t ff_dither_8x8_220[9][8];
829 
830 extern const int32_t ff_yuv2rgb_coeffs[8][4];
831 
832 extern const AVClass sws_context_class;
833 
834 /**
835  * Set c->swscale to an unscaled converter if one exists for the specific
836  * source and destination formats, bit depths, flags, etc.
837  */
841 
842 /**
843  * Return function pointer to fastest main scaler path function depending
844  * on architecture and available optimizations.
845  */
847 
859 
860 static inline void fillPlane16(uint8_t *plane, int stride, int width, int height, int y,
861  int alpha, int bits, const int big_endian)
862 {
863  int i, j;
864  uint8_t *ptr = plane + stride * y;
865  int v = alpha ? 0xFFFF>>(15-bits) : (1<<bits);
866  for (i = 0; i < height; i++) {
867 #define FILL(wfunc) \
868  for (j = 0; j < width; j++) {\
869  wfunc(ptr+2*j, v);\
870  }
871  if (big_endian) {
872  FILL(AV_WB16);
873  } else {
874  FILL(AV_WL16);
875  }
876  ptr += stride;
877  }
878 }
879 
880 #endif /* SWSCALE_SWSCALE_INTERNAL_H */