• Home
  • Line#
  • Scopes#
  • Navigate#
  • Raw
  • Download
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 #include "version.h"
26 
27 #include "libavutil/avassert.h"
28 #include "libavutil/avutil.h"
29 #include "libavutil/common.h"
30 #include "libavutil/intreadwrite.h"
31 #include "libavutil/log.h"
32 #include "libavutil/mem_internal.h"
33 #include "libavutil/pixfmt.h"
34 #include "libavutil/pixdesc.h"
35 #include "libavutil/ppc/util_altivec.h"
36 
37 #define STR(s) AV_TOSTRING(s) // AV_STRINGIFY is too long
38 
39 #define YUVRGB_TABLE_HEADROOM 512
40 #define YUVRGB_TABLE_LUMA_HEADROOM 512
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 #define RETCODE_USE_CASCADE -12345
63 
64 struct SwsContext;
65 
66 typedef enum SwsDither {
67     SWS_DITHER_NONE = 0,
68     SWS_DITHER_AUTO,
69     SWS_DITHER_BAYER,
70     SWS_DITHER_ED,
71     SWS_DITHER_A_DITHER,
72     SWS_DITHER_X_DITHER,
73     NB_SWS_DITHER,
74 } SwsDither;
75 
76 typedef enum SwsAlphaBlend {
77     SWS_ALPHA_BLEND_NONE  = 0,
78     SWS_ALPHA_BLEND_UNIFORM,
79     SWS_ALPHA_BLEND_CHECKERBOARD,
80     SWS_ALPHA_BLEND_NB,
81 } SwsAlphaBlend;
82 
83 typedef int (*SwsFunc)(struct SwsContext *context, const uint8_t *src[],
84                        int srcStride[], int srcSliceY, int srcSliceH,
85                        uint8_t *dst[], int dstStride[]);
86 
87 /**
88  * Write one line of horizontally scaled data to planar output
89  * without any additional vertical scaling (or point-scaling).
90  *
91  * @param src     scaled source data, 15 bits for 8-10-bit output,
92  *                19 bits for 16-bit output (in int32_t)
93  * @param dest    pointer to the output plane. For >8-bit
94  *                output, this is in uint16_t
95  * @param dstW    width of destination in pixels
96  * @param dither  ordered dither array of type int16_t and size 8
97  * @param offset  Dither offset
98  */
99 typedef void (*yuv2planar1_fn)(const int16_t *src, uint8_t *dest, int dstW,
100                                const uint8_t *dither, int offset);
101 
102 /**
103  * Write one line of horizontally scaled data to planar output
104  * with multi-point vertical scaling between input pixels.
105  *
106  * @param filter        vertical luma/alpha scaling coefficients, 12 bits [0,4096]
107  * @param src           scaled luma (Y) or alpha (A) source data, 15 bits for
108  *                      8-10-bit output, 19 bits for 16-bit output (in int32_t)
109  * @param filterSize    number of vertical input lines to scale
110  * @param dest          pointer to output plane. For >8-bit
111  *                      output, this is in uint16_t
112  * @param dstW          width of destination pixels
113  * @param offset        Dither offset
114  */
115 typedef void (*yuv2planarX_fn)(const int16_t *filter, int filterSize,
116                                const int16_t **src, uint8_t *dest, int dstW,
117                                const uint8_t *dither, int offset);
118 
119 /**
120  * Write one line of horizontally scaled chroma to interleaved output
121  * with multi-point vertical scaling between input pixels.
122  *
123  * @param dstFormat     destination pixel format
124  * @param chrDither     ordered dither array of type uint8_t and size 8
125  * @param chrFilter     vertical chroma scaling coefficients, 12 bits [0,4096]
126  * @param chrUSrc       scaled chroma (U) source data, 15 bits for 8-10-bit
127  *                      output, 19 bits for 16-bit output (in int32_t)
128  * @param chrVSrc       scaled chroma (V) source data, 15 bits for 8-10-bit
129  *                      output, 19 bits for 16-bit output (in int32_t)
130  * @param chrFilterSize number of vertical chroma input lines to scale
131  * @param dest          pointer to the output plane. For >8-bit
132  *                      output, this is in uint16_t
133  * @param dstW          width of chroma planes
134  */
135 typedef void (*yuv2interleavedX_fn)(enum AVPixelFormat dstFormat,
136                                     const uint8_t *chrDither,
137                                     const int16_t *chrFilter,
138                                     int chrFilterSize,
139                                     const int16_t **chrUSrc,
140                                     const int16_t **chrVSrc,
141                                     uint8_t *dest, int dstW);
142 
143 /**
144  * Write one line of horizontally scaled Y/U/V/A to packed-pixel YUV/RGB
145  * output without any additional vertical scaling (or point-scaling). Note
146  * that this function may do chroma scaling, see the "uvalpha" argument.
147  *
148  * @param c       SWS scaling context
149  * @param lumSrc  scaled luma (Y) source data, 15 bits for 8-10-bit output,
150  *                19 bits for 16-bit output (in int32_t)
151  * @param chrUSrc scaled chroma (U) source data, 15 bits for 8-10-bit output,
152  *                19 bits for 16-bit output (in int32_t)
153  * @param chrVSrc scaled chroma (V) source data, 15 bits for 8-10-bit output,
154  *                19 bits for 16-bit output (in int32_t)
155  * @param alpSrc  scaled alpha (A) source data, 15 bits for 8-10-bit output,
156  *                19 bits for 16-bit output (in int32_t)
157  * @param dest    pointer to the output plane. For 16-bit output, this is
158  *                uint16_t
159  * @param dstW    width of lumSrc and alpSrc in pixels, number of pixels
160  *                to write into dest[]
161  * @param uvalpha chroma scaling coefficient for the second line of chroma
162  *                pixels, either 2048 or 0. If 0, one chroma input is used
163  *                for 2 output pixels (or if the SWS_FLAG_FULL_CHR_INT flag
164  *                is set, it generates 1 output pixel). If 2048, two chroma
165  *                input pixels should be averaged for 2 output pixels (this
166  *                only happens if SWS_FLAG_FULL_CHR_INT is not set)
167  * @param y       vertical line number for this output. This does not need
168  *                to be used to calculate the offset in the destination,
169  *                but can be used to generate comfort noise using dithering
170  *                for some output formats.
171  */
172 typedef void (*yuv2packed1_fn)(struct SwsContext *c, const int16_t *lumSrc,
173                                const int16_t *chrUSrc[2],
174                                const int16_t *chrVSrc[2],
175                                const int16_t *alpSrc, uint8_t *dest,
176                                int dstW, int uvalpha, int y);
177 /**
178  * Write one line of horizontally scaled Y/U/V/A to packed-pixel YUV/RGB
179  * output by doing bilinear scaling between two input lines.
180  *
181  * @param c       SWS scaling context
182  * @param lumSrc  scaled luma (Y) source data, 15 bits for 8-10-bit output,
183  *                19 bits for 16-bit output (in int32_t)
184  * @param chrUSrc scaled chroma (U) source data, 15 bits for 8-10-bit output,
185  *                19 bits for 16-bit output (in int32_t)
186  * @param chrVSrc scaled chroma (V) source data, 15 bits for 8-10-bit output,
187  *                19 bits for 16-bit output (in int32_t)
188  * @param alpSrc  scaled alpha (A) source data, 15 bits for 8-10-bit output,
189  *                19 bits for 16-bit output (in int32_t)
190  * @param dest    pointer to the output plane. For 16-bit output, this is
191  *                uint16_t
192  * @param dstW    width of lumSrc and alpSrc in pixels, number of pixels
193  *                to write into dest[]
194  * @param yalpha  luma/alpha scaling coefficients for the second input line.
195  *                The first line's coefficients can be calculated by using
196  *                4096 - yalpha
197  * @param uvalpha chroma scaling coefficient for the second input line. The
198  *                first line's coefficients can be calculated by using
199  *                4096 - uvalpha
200  * @param y       vertical line number for this output. This does not need
201  *                to be used to calculate the offset in the destination,
202  *                but can be used to generate comfort noise using dithering
203  *                for some output formats.
204  */
205 typedef void (*yuv2packed2_fn)(struct SwsContext *c, const int16_t *lumSrc[2],
206                                const int16_t *chrUSrc[2],
207                                const int16_t *chrVSrc[2],
208                                const int16_t *alpSrc[2],
209                                uint8_t *dest,
210                                int dstW, int yalpha, int uvalpha, int y);
211 /**
212  * Write one line of horizontally scaled Y/U/V/A to packed-pixel YUV/RGB
213  * output by doing multi-point vertical scaling between input pixels.
214  *
215  * @param c             SWS scaling context
216  * @param lumFilter     vertical luma/alpha scaling coefficients, 12 bits [0,4096]
217  * @param lumSrc        scaled luma (Y) source data, 15 bits for 8-10-bit output,
218  *                      19 bits for 16-bit output (in int32_t)
219  * @param lumFilterSize number of vertical luma/alpha input lines to scale
220  * @param chrFilter     vertical chroma scaling coefficients, 12 bits [0,4096]
221  * @param chrUSrc       scaled chroma (U) source data, 15 bits for 8-10-bit output,
222  *                      19 bits for 16-bit output (in int32_t)
223  * @param chrVSrc       scaled chroma (V) source data, 15 bits for 8-10-bit output,
224  *                      19 bits for 16-bit output (in int32_t)
225  * @param chrFilterSize number of vertical chroma input lines to scale
226  * @param alpSrc        scaled alpha (A) source data, 15 bits for 8-10-bit output,
227  *                      19 bits for 16-bit output (in int32_t)
228  * @param dest          pointer to the output plane. For 16-bit output, this is
229  *                      uint16_t
230  * @param dstW          width of lumSrc and alpSrc in pixels, number of pixels
231  *                      to write into dest[]
232  * @param y             vertical line number for this output. This does not need
233  *                      to be used to calculate the offset in the destination,
234  *                      but can be used to generate comfort noise using dithering
235  *                      or some output formats.
236  */
237 typedef void (*yuv2packedX_fn)(struct SwsContext *c, const int16_t *lumFilter,
238                                const int16_t **lumSrc, int lumFilterSize,
239                                const int16_t *chrFilter,
240                                const int16_t **chrUSrc,
241                                const int16_t **chrVSrc, int chrFilterSize,
242                                const int16_t **alpSrc, uint8_t *dest,
243                                int dstW, int y);
244 
245 /**
246  * Write one line of horizontally scaled Y/U/V/A to YUV/RGB
247  * output by doing multi-point vertical scaling between input pixels.
248  *
249  * @param c             SWS scaling context
250  * @param lumFilter     vertical luma/alpha scaling coefficients, 12 bits [0,4096]
251  * @param lumSrc        scaled luma (Y) source data, 15 bits for 8-10-bit output,
252  *                      19 bits for 16-bit output (in int32_t)
253  * @param lumFilterSize number of vertical luma/alpha input lines to scale
254  * @param chrFilter     vertical chroma scaling coefficients, 12 bits [0,4096]
255  * @param chrUSrc       scaled chroma (U) source data, 15 bits for 8-10-bit output,
256  *                      19 bits for 16-bit output (in int32_t)
257  * @param chrVSrc       scaled chroma (V) source data, 15 bits for 8-10-bit output,
258  *                      19 bits for 16-bit output (in int32_t)
259  * @param chrFilterSize number of vertical chroma input lines to scale
260  * @param alpSrc        scaled alpha (A) source data, 15 bits for 8-10-bit output,
261  *                      19 bits for 16-bit output (in int32_t)
262  * @param dest          pointer to the output planes. For 16-bit output, this is
263  *                      uint16_t
264  * @param dstW          width of lumSrc and alpSrc in pixels, number of pixels
265  *                      to write into dest[]
266  * @param y             vertical line number for this output. This does not need
267  *                      to be used to calculate the offset in the destination,
268  *                      but can be used to generate comfort noise using dithering
269  *                      or some output formats.
270  */
271 typedef void (*yuv2anyX_fn)(struct SwsContext *c, const int16_t *lumFilter,
272                             const int16_t **lumSrc, int lumFilterSize,
273                             const int16_t *chrFilter,
274                             const int16_t **chrUSrc,
275                             const int16_t **chrVSrc, int chrFilterSize,
276                             const int16_t **alpSrc, uint8_t **dest,
277                             int dstW, int y);
278 
279 struct SwsSlice;
280 struct SwsFilterDescriptor;
281 
282 /* This struct should be aligned on at least a 32-byte boundary. */
283 typedef struct SwsContext {
284     /**
285      * info on struct for av_log
286      */
287     const AVClass *av_class;
288 
289     /**
290      * Note that src, dst, srcStride, dstStride will be copied in the
291      * sws_scale() wrapper so they can be freely modified here.
292      */
293     SwsFunc swscale;
294     int srcW;                     ///< Width  of source      luma/alpha planes.
295     int srcH;                     ///< Height of source      luma/alpha planes.
296     int dstH;                     ///< Height of destination luma/alpha planes.
297     int chrSrcW;                  ///< Width  of source      chroma     planes.
298     int chrSrcH;                  ///< Height of source      chroma     planes.
299     int chrDstW;                  ///< Width  of destination chroma     planes.
300     int chrDstH;                  ///< Height of destination chroma     planes.
301     int lumXInc, chrXInc;
302     int lumYInc, chrYInc;
303     enum AVPixelFormat dstFormat; ///< Destination pixel format.
304     enum AVPixelFormat srcFormat; ///< Source      pixel format.
305     int dstFormatBpp;             ///< Number of bits per pixel of the destination pixel format.
306     int srcFormatBpp;             ///< Number of bits per pixel of the source      pixel format.
307     int dstBpc, srcBpc;
308     int chrSrcHSubSample;         ///< Binary logarithm of horizontal subsampling factor between luma/alpha and chroma planes in source      image.
309     int chrSrcVSubSample;         ///< Binary logarithm of vertical   subsampling factor between luma/alpha and chroma planes in source      image.
310     int chrDstHSubSample;         ///< Binary logarithm of horizontal subsampling factor between luma/alpha and chroma planes in destination image.
311     int chrDstVSubSample;         ///< Binary logarithm of vertical   subsampling factor between luma/alpha and chroma planes in destination image.
312     int vChrDrop;                 ///< Binary logarithm of extra vertical subsampling factor in source image chroma planes specified by user.
313     int sliceDir;                 ///< Direction that slices are fed to the scaler (1 = top-to-bottom, -1 = bottom-to-top).
314     double param[2];              ///< Input parameters for scaling algorithms that need them.
315 
316     /* The cascaded_* fields allow spliting a scaler task into multiple
317      * sequential steps, this is for example used to limit the maximum
318      * downscaling factor that needs to be supported in one scaler.
319      */
320     struct SwsContext *cascaded_context[3];
321     int cascaded_tmpStride[4];
322     uint8_t *cascaded_tmp[4];
323     int cascaded1_tmpStride[4];
324     uint8_t *cascaded1_tmp[4];
325     int cascaded_mainindex;
326 
327     double gamma_value;
328     int gamma_flag;
329     int is_internal_gamma;
330     uint16_t *gamma;
331     uint16_t *inv_gamma;
332 
333     int numDesc;
334     int descIndex[2];
335     int numSlice;
336     struct SwsSlice *slice;
337     struct SwsFilterDescriptor *desc;
338 
339     uint32_t pal_yuv[256];
340     uint32_t pal_rgb[256];
341 
342     float uint2float_lut[256];
343 
344     /**
345      * @name Scaled horizontal lines ring buffer.
346      * The horizontal scaler keeps just enough scaled lines in a ring buffer
347      * so they may be passed to the vertical scaler. The pointers to the
348      * allocated buffers for each line are duplicated in sequence in the ring
349      * buffer to simplify indexing and avoid wrapping around between lines
350      * inside the vertical scaler code. The wrapping is done before the
351      * vertical scaler is called.
352      */
353     //@{
354     int lastInLumBuf;             ///< Last scaled horizontal luma/alpha line from source in the ring buffer.
355     int lastInChrBuf;             ///< Last scaled horizontal chroma     line from source in the ring buffer.
356     //@}
357 
358     uint8_t *formatConvBuffer;
359     int needAlpha;
360 
361     /**
362      * @name Horizontal and vertical filters.
363      * To better understand the following fields, here is a pseudo-code of
364      * their usage in filtering a horizontal line:
365      * @code
366      * for (i = 0; i < width; i++) {
367      *     dst[i] = 0;
368      *     for (j = 0; j < filterSize; j++)
369      *         dst[i] += src[ filterPos[i] + j ] * filter[ filterSize * i + j ];
370      *     dst[i] >>= FRAC_BITS; // The actual implementation is fixed-point.
371      * }
372      * @endcode
373      */
374     //@{
375     int16_t *hLumFilter;          ///< Array of horizontal filter coefficients for luma/alpha planes.
376     int16_t *hChrFilter;          ///< Array of horizontal filter coefficients for chroma     planes.
377     int16_t *vLumFilter;          ///< Array of vertical   filter coefficients for luma/alpha planes.
378     int16_t *vChrFilter;          ///< Array of vertical   filter coefficients for chroma     planes.
379     int32_t *hLumFilterPos;       ///< Array of horizontal filter starting positions for each dst[i] for luma/alpha planes.
380     int32_t *hChrFilterPos;       ///< Array of horizontal filter starting positions for each dst[i] for chroma     planes.
381     int32_t *vLumFilterPos;       ///< Array of vertical   filter starting positions for each dst[i] for luma/alpha planes.
382     int32_t *vChrFilterPos;       ///< Array of vertical   filter starting positions for each dst[i] for chroma     planes.
383     int hLumFilterSize;           ///< Horizontal filter size for luma/alpha pixels.
384     int hChrFilterSize;           ///< Horizontal filter size for chroma     pixels.
385     int vLumFilterSize;           ///< Vertical   filter size for luma/alpha pixels.
386     int vChrFilterSize;           ///< Vertical   filter size for chroma     pixels.
387     //@}
388 
389     int lumMmxextFilterCodeSize;  ///< Runtime-generated MMXEXT horizontal fast bilinear scaler code size for luma/alpha planes.
390     int chrMmxextFilterCodeSize;  ///< Runtime-generated MMXEXT horizontal fast bilinear scaler code size for chroma planes.
391     uint8_t *lumMmxextFilterCode; ///< Runtime-generated MMXEXT horizontal fast bilinear scaler code for luma/alpha planes.
392     uint8_t *chrMmxextFilterCode; ///< Runtime-generated MMXEXT horizontal fast bilinear scaler code for chroma planes.
393 
394     int canMMXEXTBeUsed;
395     int warned_unuseable_bilinear;
396 
397     int dstY;                     ///< Last destination vertical line output from last slice.
398     int flags;                    ///< Flags passed by the user to select scaler algorithm, optimizations, subsampling, etc...
399     void *yuvTable;             // pointer to the yuv->rgb table start so it can be freed()
400     // alignment ensures the offset can be added in a single
401     // instruction on e.g. ARM
402     DECLARE_ALIGNED(16, int, table_gV)[256 + 2*YUVRGB_TABLE_HEADROOM];
403     uint8_t *table_rV[256 + 2*YUVRGB_TABLE_HEADROOM];
404     uint8_t *table_gU[256 + 2*YUVRGB_TABLE_HEADROOM];
405     uint8_t *table_bU[256 + 2*YUVRGB_TABLE_HEADROOM];
406     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
407 #define RY_IDX 0
408 #define GY_IDX 1
409 #define BY_IDX 2
410 #define RU_IDX 3
411 #define GU_IDX 4
412 #define BU_IDX 5
413 #define RV_IDX 6
414 #define GV_IDX 7
415 #define BV_IDX 8
416 #define RGB2YUV_SHIFT 15
417 
418     int *dither_error[4];
419 
420     //Colorspace stuff
421     int contrast, brightness, saturation;    // for sws_getColorspaceDetails
422     int srcColorspaceTable[4];
423     int dstColorspaceTable[4];
424     int srcRange;                 ///< 0 = MPG YUV range, 1 = JPG YUV range (source      image).
425     int dstRange;                 ///< 0 = MPG YUV range, 1 = JPG YUV range (destination image).
426     int src0Alpha;
427     int dst0Alpha;
428     int srcXYZ;
429     int dstXYZ;
430     int src_h_chr_pos;
431     int dst_h_chr_pos;
432     int src_v_chr_pos;
433     int dst_v_chr_pos;
434     int yuv2rgb_y_offset;
435     int yuv2rgb_y_coeff;
436     int yuv2rgb_v2r_coeff;
437     int yuv2rgb_v2g_coeff;
438     int yuv2rgb_u2g_coeff;
439     int yuv2rgb_u2b_coeff;
440 
441 #define RED_DITHER            "0*8"
442 #define GREEN_DITHER          "1*8"
443 #define BLUE_DITHER           "2*8"
444 #define Y_COEFF               "3*8"
445 #define VR_COEFF              "4*8"
446 #define UB_COEFF              "5*8"
447 #define VG_COEFF              "6*8"
448 #define UG_COEFF              "7*8"
449 #define Y_OFFSET              "8*8"
450 #define U_OFFSET              "9*8"
451 #define V_OFFSET              "10*8"
452 #define LUM_MMX_FILTER_OFFSET "11*8"
453 #define CHR_MMX_FILTER_OFFSET "11*8+4*4*"AV_STRINGIFY(MAX_FILTER_SIZE)
454 #define DSTW_OFFSET           "11*8+4*4*"AV_STRINGIFY(MAX_FILTER_SIZE)"*2"
455 #define ESP_OFFSET            "11*8+4*4*"AV_STRINGIFY(MAX_FILTER_SIZE)"*2+8"
456 #define VROUNDER_OFFSET       "11*8+4*4*"AV_STRINGIFY(MAX_FILTER_SIZE)"*2+16"
457 #define U_TEMP                "11*8+4*4*"AV_STRINGIFY(MAX_FILTER_SIZE)"*2+24"
458 #define V_TEMP                "11*8+4*4*"AV_STRINGIFY(MAX_FILTER_SIZE)"*2+32"
459 #define Y_TEMP                "11*8+4*4*"AV_STRINGIFY(MAX_FILTER_SIZE)"*2+40"
460 #define ALP_MMX_FILTER_OFFSET "11*8+4*4*"AV_STRINGIFY(MAX_FILTER_SIZE)"*2+48"
461 #define UV_OFF_PX             "11*8+4*4*"AV_STRINGIFY(MAX_FILTER_SIZE)"*3+48"
462 #define UV_OFF_BYTE           "11*8+4*4*"AV_STRINGIFY(MAX_FILTER_SIZE)"*3+56"
463 #define DITHER16              "11*8+4*4*"AV_STRINGIFY(MAX_FILTER_SIZE)"*3+64"
464 #define DITHER32              "11*8+4*4*"AV_STRINGIFY(MAX_FILTER_SIZE)"*3+80"
465 #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
466 
467     DECLARE_ALIGNED(8, uint64_t, redDither);
468     DECLARE_ALIGNED(8, uint64_t, greenDither);
469     DECLARE_ALIGNED(8, uint64_t, blueDither);
470 
471     DECLARE_ALIGNED(8, uint64_t, yCoeff);
472     DECLARE_ALIGNED(8, uint64_t, vrCoeff);
473     DECLARE_ALIGNED(8, uint64_t, ubCoeff);
474     DECLARE_ALIGNED(8, uint64_t, vgCoeff);
475     DECLARE_ALIGNED(8, uint64_t, ugCoeff);
476     DECLARE_ALIGNED(8, uint64_t, yOffset);
477     DECLARE_ALIGNED(8, uint64_t, uOffset);
478     DECLARE_ALIGNED(8, uint64_t, vOffset);
479     int32_t lumMmxFilter[4 * MAX_FILTER_SIZE];
480     int32_t chrMmxFilter[4 * MAX_FILTER_SIZE];
481     int dstW;                     ///< Width  of destination luma/alpha planes.
482     DECLARE_ALIGNED(8, uint64_t, esp);
483     DECLARE_ALIGNED(8, uint64_t, vRounder);
484     DECLARE_ALIGNED(8, uint64_t, u_temp);
485     DECLARE_ALIGNED(8, uint64_t, v_temp);
486     DECLARE_ALIGNED(8, uint64_t, y_temp);
487     int32_t alpMmxFilter[4 * MAX_FILTER_SIZE];
488     // alignment of these values is not necessary, but merely here
489     // to maintain the same offset across x8632 and x86-64. Once we
490     // use proper offset macros in the asm, they can be removed.
491     DECLARE_ALIGNED(8, ptrdiff_t, uv_off); ///< offset (in pixels) between u and v planes
492     DECLARE_ALIGNED(8, ptrdiff_t, uv_offx2); ///< offset (in bytes) between u and v planes
493     DECLARE_ALIGNED(8, uint16_t, dither16)[8];
494     DECLARE_ALIGNED(8, uint32_t, dither32)[8];
495 
496     const uint8_t *chrDither8, *lumDither8;
497 
498 #if HAVE_ALTIVEC
499     vector signed short   CY;
500     vector signed short   CRV;
501     vector signed short   CBU;
502     vector signed short   CGU;
503     vector signed short   CGV;
504     vector signed short   OY;
505     vector unsigned short CSHIFT;
506     vector signed short  *vYCoeffsBank, *vCCoeffsBank;
507 #endif
508 
509     int use_mmx_vfilter;
510 
511 /* pre defined color-spaces gamma */
512 #define XYZ_GAMMA (2.6f)
513 #define RGB_GAMMA (2.2f)
514     int16_t *xyzgamma;
515     int16_t *rgbgamma;
516     int16_t *xyzgammainv;
517     int16_t *rgbgammainv;
518     int16_t xyz2rgb_matrix[3][4];
519     int16_t rgb2xyz_matrix[3][4];
520 
521     /* function pointers for swscale() */
522     yuv2planar1_fn yuv2plane1;
523     yuv2planarX_fn yuv2planeX;
524     yuv2interleavedX_fn yuv2nv12cX;
525     yuv2packed1_fn yuv2packed1;
526     yuv2packed2_fn yuv2packed2;
527     yuv2packedX_fn yuv2packedX;
528     yuv2anyX_fn yuv2anyX;
529 
530     /// Unscaled conversion of luma plane to YV12 for horizontal scaler.
531     void (*lumToYV12)(uint8_t *dst, const uint8_t *src, const uint8_t *src2, const uint8_t *src3,
532                       int width, uint32_t *pal);
533     /// Unscaled conversion of alpha plane to YV12 for horizontal scaler.
534     void (*alpToYV12)(uint8_t *dst, const uint8_t *src, const uint8_t *src2, const uint8_t *src3,
535                       int width, uint32_t *pal);
536     /// Unscaled conversion of chroma planes to YV12 for horizontal scaler.
537     void (*chrToYV12)(uint8_t *dstU, uint8_t *dstV,
538                       const uint8_t *src1, const uint8_t *src2, const uint8_t *src3,
539                       int width, uint32_t *pal);
540 
541     /**
542      * Functions to read planar input, such as planar RGB, and convert
543      * internally to Y/UV/A.
544      */
545     /** @{ */
546     void (*readLumPlanar)(uint8_t *dst, const uint8_t *src[4], int width, int32_t *rgb2yuv);
547     void (*readChrPlanar)(uint8_t *dstU, uint8_t *dstV, const uint8_t *src[4],
548                           int width, int32_t *rgb2yuv);
549     void (*readAlpPlanar)(uint8_t *dst, const uint8_t *src[4], int width, int32_t *rgb2yuv);
550     /** @} */
551 
552     /**
553      * Scale one horizontal line of input data using a bilinear filter
554      * to produce one line of output data. Compared to SwsContext->hScale(),
555      * please take note of the following caveats when using these:
556      * - Scaling is done using only 7 bits instead of 14-bit coefficients.
557      * - You can use no more than 5 input pixels to produce 4 output
558      *   pixels. Therefore, this filter should not be used for downscaling
559      *   by more than ~20% in width (because that equals more than 5/4th
560      *   downscaling and thus more than 5 pixels input per 4 pixels output).
561      * - In general, bilinear filters create artifacts during downscaling
562      *   (even when <20%), because one output pixel will span more than one
563      *   input pixel, and thus some pixels will need edges of both neighbor
564      *   pixels to interpolate the output pixel. Since you can use at most
565      *   two input pixels per output pixel in bilinear scaling, this is
566      *   impossible and thus downscaling by any size will create artifacts.
567      * To enable this type of scaling, set SWS_FLAG_FAST_BILINEAR
568      * in SwsContext->flags.
569      */
570     /** @{ */
571     void (*hyscale_fast)(struct SwsContext *c,
572                          int16_t *dst, int dstWidth,
573                          const uint8_t *src, int srcW, int xInc);
574     void (*hcscale_fast)(struct SwsContext *c,
575                          int16_t *dst1, int16_t *dst2, int dstWidth,
576                          const uint8_t *src1, const uint8_t *src2,
577                          int srcW, int xInc);
578     /** @} */
579 
580     /**
581      * Scale one horizontal line of input data using a filter over the input
582      * lines, to produce one (differently sized) line of output data.
583      *
584      * @param dst        pointer to destination buffer for horizontally scaled
585      *                   data. If the number of bits per component of one
586      *                   destination pixel (SwsContext->dstBpc) is <= 10, data
587      *                   will be 15 bpc in 16 bits (int16_t) width. Else (i.e.
588      *                   SwsContext->dstBpc == 16), data will be 19bpc in
589      *                   32 bits (int32_t) width.
590      * @param dstW       width of destination image
591      * @param src        pointer to source data to be scaled. If the number of
592      *                   bits per component of a source pixel (SwsContext->srcBpc)
593      *                   is 8, this is 8bpc in 8 bits (uint8_t) width. Else
594      *                   (i.e. SwsContext->dstBpc > 8), this is native depth
595      *                   in 16 bits (uint16_t) width. In other words, for 9-bit
596      *                   YUV input, this is 9bpc, for 10-bit YUV input, this is
597      *                   10bpc, and for 16-bit RGB or YUV, this is 16bpc.
598      * @param filter     filter coefficients to be used per output pixel for
599      *                   scaling. This contains 14bpp filtering coefficients.
600      *                   Guaranteed to contain dstW * filterSize entries.
601      * @param filterPos  position of the first input pixel to be used for
602      *                   each output pixel during scaling. Guaranteed to
603      *                   contain dstW entries.
604      * @param filterSize the number of input coefficients to be used (and
605      *                   thus the number of input pixels to be used) for
606      *                   creating a single output pixel. Is aligned to 4
607      *                   (and input coefficients thus padded with zeroes)
608      *                   to simplify creating SIMD code.
609      */
610     /** @{ */
611     void (*hyScale)(struct SwsContext *c, int16_t *dst, int dstW,
612                     const uint8_t *src, const int16_t *filter,
613                     const int32_t *filterPos, int filterSize);
614     void (*hcScale)(struct SwsContext *c, int16_t *dst, int dstW,
615                     const uint8_t *src, const int16_t *filter,
616                     const int32_t *filterPos, int filterSize);
617     /** @} */
618 
619     /// Color range conversion function for luma plane if needed.
620     void (*lumConvertRange)(int16_t *dst, int width);
621     /// Color range conversion function for chroma planes if needed.
622     void (*chrConvertRange)(int16_t *dst1, int16_t *dst2, int width);
623 
624     int needs_hcscale; ///< Set if there are chroma planes to be converted.
625 
626     SwsDither dither;
627 
628     SwsAlphaBlend alphablend;
629 } SwsContext;
630 //FIXME check init (where 0)
631 
632 SwsFunc ff_yuv2rgb_get_func_ptr(SwsContext *c);
633 int ff_yuv2rgb_c_init_tables(SwsContext *c, const int inv_table[4],
634                              int fullRange, int brightness,
635                              int contrast, int saturation);
636 void ff_yuv2rgb_init_tables_ppc(SwsContext *c, const int inv_table[4],
637                                 int brightness, int contrast, int saturation);
638 
639 void ff_updateMMXDitherTables(SwsContext *c, int dstY);
640 
641 av_cold void ff_sws_init_range_convert(SwsContext *c);
642 
643 SwsFunc ff_yuv2rgb_init_x86(SwsContext *c);
644 SwsFunc ff_yuv2rgb_init_ppc(SwsContext *c);
645 
is16BPS(enum AVPixelFormat pix_fmt)646 static av_always_inline int is16BPS(enum AVPixelFormat pix_fmt)
647 {
648     const AVPixFmtDescriptor *desc = av_pix_fmt_desc_get(pix_fmt);
649     av_assert0(desc);
650     return desc->comp[0].depth == 16;
651 }
652 
is32BPS(enum AVPixelFormat pix_fmt)653 static av_always_inline int is32BPS(enum AVPixelFormat pix_fmt)
654 {
655     const AVPixFmtDescriptor *desc = av_pix_fmt_desc_get(pix_fmt);
656     av_assert0(desc);
657     return desc->comp[0].depth == 32;
658 }
659 
isNBPS(enum AVPixelFormat pix_fmt)660 static av_always_inline int isNBPS(enum AVPixelFormat pix_fmt)
661 {
662     const AVPixFmtDescriptor *desc = av_pix_fmt_desc_get(pix_fmt);
663     av_assert0(desc);
664     return desc->comp[0].depth >= 9 && desc->comp[0].depth <= 14;
665 }
666 
isBE(enum AVPixelFormat pix_fmt)667 static av_always_inline int isBE(enum AVPixelFormat pix_fmt)
668 {
669     const AVPixFmtDescriptor *desc = av_pix_fmt_desc_get(pix_fmt);
670     av_assert0(desc);
671     return desc->flags & AV_PIX_FMT_FLAG_BE;
672 }
673 
isYUV(enum AVPixelFormat pix_fmt)674 static av_always_inline int isYUV(enum AVPixelFormat pix_fmt)
675 {
676     const AVPixFmtDescriptor *desc = av_pix_fmt_desc_get(pix_fmt);
677     av_assert0(desc);
678     return !(desc->flags & AV_PIX_FMT_FLAG_RGB) && desc->nb_components >= 2;
679 }
680 
isPlanarYUV(enum AVPixelFormat pix_fmt)681 static av_always_inline int isPlanarYUV(enum AVPixelFormat pix_fmt)
682 {
683     const AVPixFmtDescriptor *desc = av_pix_fmt_desc_get(pix_fmt);
684     av_assert0(desc);
685     return ((desc->flags & AV_PIX_FMT_FLAG_PLANAR) && isYUV(pix_fmt));
686 }
687 
688 /*
689  * Identity semi-planar YUV formats. Specifically, those are YUV formats
690  * where the second and third components (U & V) are on the same plane.
691  */
isSemiPlanarYUV(enum AVPixelFormat pix_fmt)692 static av_always_inline int isSemiPlanarYUV(enum AVPixelFormat pix_fmt)
693 {
694     const AVPixFmtDescriptor *desc = av_pix_fmt_desc_get(pix_fmt);
695     av_assert0(desc);
696     return (isPlanarYUV(pix_fmt) && desc->comp[1].plane == desc->comp[2].plane);
697 }
698 
isRGB(enum AVPixelFormat pix_fmt)699 static av_always_inline int isRGB(enum AVPixelFormat pix_fmt)
700 {
701     const AVPixFmtDescriptor *desc = av_pix_fmt_desc_get(pix_fmt);
702     av_assert0(desc);
703     return (desc->flags & AV_PIX_FMT_FLAG_RGB);
704 }
705 
isGray(enum AVPixelFormat pix_fmt)706 static av_always_inline int isGray(enum AVPixelFormat pix_fmt)
707 {
708     const AVPixFmtDescriptor *desc = av_pix_fmt_desc_get(pix_fmt);
709     av_assert0(desc);
710     return !(desc->flags & AV_PIX_FMT_FLAG_PAL) &&
711            !(desc->flags & AV_PIX_FMT_FLAG_HWACCEL) &&
712            desc->nb_components <= 2 &&
713            pix_fmt != AV_PIX_FMT_MONOBLACK &&
714            pix_fmt != AV_PIX_FMT_MONOWHITE;
715 }
716 
isRGBinInt(enum AVPixelFormat pix_fmt)717 static av_always_inline int isRGBinInt(enum AVPixelFormat pix_fmt)
718 {
719     return pix_fmt == AV_PIX_FMT_RGB48BE     ||
720            pix_fmt == AV_PIX_FMT_RGB48LE     ||
721            pix_fmt == AV_PIX_FMT_RGB32       ||
722            pix_fmt == AV_PIX_FMT_RGB32_1     ||
723            pix_fmt == AV_PIX_FMT_RGB24       ||
724            pix_fmt == AV_PIX_FMT_RGB565BE    ||
725            pix_fmt == AV_PIX_FMT_RGB565LE    ||
726            pix_fmt == AV_PIX_FMT_RGB555BE    ||
727            pix_fmt == AV_PIX_FMT_RGB555LE    ||
728            pix_fmt == AV_PIX_FMT_RGB444BE    ||
729            pix_fmt == AV_PIX_FMT_RGB444LE    ||
730            pix_fmt == AV_PIX_FMT_RGB8        ||
731            pix_fmt == AV_PIX_FMT_RGB4        ||
732            pix_fmt == AV_PIX_FMT_RGB4_BYTE   ||
733            pix_fmt == AV_PIX_FMT_RGBA64BE    ||
734            pix_fmt == AV_PIX_FMT_RGBA64LE    ||
735            pix_fmt == AV_PIX_FMT_MONOBLACK   ||
736            pix_fmt == AV_PIX_FMT_MONOWHITE;
737 }
738 
isBGRinInt(enum AVPixelFormat pix_fmt)739 static av_always_inline int isBGRinInt(enum AVPixelFormat pix_fmt)
740 {
741     return pix_fmt == AV_PIX_FMT_BGR48BE     ||
742            pix_fmt == AV_PIX_FMT_BGR48LE     ||
743            pix_fmt == AV_PIX_FMT_BGR32       ||
744            pix_fmt == AV_PIX_FMT_BGR32_1     ||
745            pix_fmt == AV_PIX_FMT_BGR24       ||
746            pix_fmt == AV_PIX_FMT_BGR565BE    ||
747            pix_fmt == AV_PIX_FMT_BGR565LE    ||
748            pix_fmt == AV_PIX_FMT_BGR555BE    ||
749            pix_fmt == AV_PIX_FMT_BGR555LE    ||
750            pix_fmt == AV_PIX_FMT_BGR444BE    ||
751            pix_fmt == AV_PIX_FMT_BGR444LE    ||
752            pix_fmt == AV_PIX_FMT_BGR8        ||
753            pix_fmt == AV_PIX_FMT_BGR4        ||
754            pix_fmt == AV_PIX_FMT_BGR4_BYTE   ||
755            pix_fmt == AV_PIX_FMT_BGRA64BE    ||
756            pix_fmt == AV_PIX_FMT_BGRA64LE    ||
757            pix_fmt == AV_PIX_FMT_MONOBLACK   ||
758            pix_fmt == AV_PIX_FMT_MONOWHITE;
759 }
760 
isBayer(enum AVPixelFormat pix_fmt)761 static av_always_inline int isBayer(enum AVPixelFormat pix_fmt)
762 {
763     const AVPixFmtDescriptor *desc = av_pix_fmt_desc_get(pix_fmt);
764     av_assert0(desc);
765     return !!(desc->flags & AV_PIX_FMT_FLAG_BAYER);
766 }
767 
isBayer16BPS(enum AVPixelFormat pix_fmt)768 static av_always_inline int isBayer16BPS(enum AVPixelFormat pix_fmt)
769 {
770     const AVPixFmtDescriptor *desc = av_pix_fmt_desc_get(pix_fmt);
771     av_assert0(desc);
772     return desc->comp[1].depth == 8;
773 }
774 
isAnyRGB(enum AVPixelFormat pix_fmt)775 static av_always_inline int isAnyRGB(enum AVPixelFormat pix_fmt)
776 {
777     const AVPixFmtDescriptor *desc = av_pix_fmt_desc_get(pix_fmt);
778     av_assert0(desc);
779     return (desc->flags & AV_PIX_FMT_FLAG_RGB) ||
780             pix_fmt == AV_PIX_FMT_MONOBLACK || pix_fmt == AV_PIX_FMT_MONOWHITE;
781 }
782 
isFloat(enum AVPixelFormat pix_fmt)783 static av_always_inline int isFloat(enum AVPixelFormat pix_fmt)
784 {
785     const AVPixFmtDescriptor *desc = av_pix_fmt_desc_get(pix_fmt);
786     av_assert0(desc);
787     return desc->flags & AV_PIX_FMT_FLAG_FLOAT;
788 }
789 
isALPHA(enum AVPixelFormat pix_fmt)790 static av_always_inline int isALPHA(enum AVPixelFormat pix_fmt)
791 {
792     const AVPixFmtDescriptor *desc = av_pix_fmt_desc_get(pix_fmt);
793     av_assert0(desc);
794     if (pix_fmt == AV_PIX_FMT_PAL8)
795         return 1;
796     return desc->flags & AV_PIX_FMT_FLAG_ALPHA;
797 }
798 
isPacked(enum AVPixelFormat pix_fmt)799 static av_always_inline int isPacked(enum AVPixelFormat pix_fmt)
800 {
801     const AVPixFmtDescriptor *desc = av_pix_fmt_desc_get(pix_fmt);
802     av_assert0(desc);
803     return (desc->nb_components >= 2 && !(desc->flags & AV_PIX_FMT_FLAG_PLANAR)) ||
804             pix_fmt == AV_PIX_FMT_PAL8 ||
805             pix_fmt == AV_PIX_FMT_MONOBLACK || pix_fmt == AV_PIX_FMT_MONOWHITE;
806 }
807 
isPlanar(enum AVPixelFormat pix_fmt)808 static av_always_inline int isPlanar(enum AVPixelFormat pix_fmt)
809 {
810     const AVPixFmtDescriptor *desc = av_pix_fmt_desc_get(pix_fmt);
811     av_assert0(desc);
812     return (desc->nb_components >= 2 && (desc->flags & AV_PIX_FMT_FLAG_PLANAR));
813 }
814 
isPackedRGB(enum AVPixelFormat pix_fmt)815 static av_always_inline int isPackedRGB(enum AVPixelFormat pix_fmt)
816 {
817     const AVPixFmtDescriptor *desc = av_pix_fmt_desc_get(pix_fmt);
818     av_assert0(desc);
819     return ((desc->flags & (AV_PIX_FMT_FLAG_PLANAR | AV_PIX_FMT_FLAG_RGB)) == AV_PIX_FMT_FLAG_RGB);
820 }
821 
isPlanarRGB(enum AVPixelFormat pix_fmt)822 static av_always_inline int isPlanarRGB(enum AVPixelFormat pix_fmt)
823 {
824     const AVPixFmtDescriptor *desc = av_pix_fmt_desc_get(pix_fmt);
825     av_assert0(desc);
826     return ((desc->flags & (AV_PIX_FMT_FLAG_PLANAR | AV_PIX_FMT_FLAG_RGB)) ==
827             (AV_PIX_FMT_FLAG_PLANAR | AV_PIX_FMT_FLAG_RGB));
828 }
829 
usePal(enum AVPixelFormat pix_fmt)830 static av_always_inline int usePal(enum AVPixelFormat pix_fmt)
831 {
832     switch (pix_fmt) {
833     case AV_PIX_FMT_PAL8:
834     case AV_PIX_FMT_BGR4_BYTE:
835     case AV_PIX_FMT_BGR8:
836     case AV_PIX_FMT_GRAY8:
837     case AV_PIX_FMT_RGB4_BYTE:
838     case AV_PIX_FMT_RGB8:
839         return 1;
840     default:
841         return 0;
842     }
843 }
844 
845 extern const uint64_t ff_dither4[2];
846 extern const uint64_t ff_dither8[2];
847 
848 extern const uint8_t ff_dither_2x2_4[3][8];
849 extern const uint8_t ff_dither_2x2_8[3][8];
850 extern const uint8_t ff_dither_4x4_16[5][8];
851 extern const uint8_t ff_dither_8x8_32[9][8];
852 extern const uint8_t ff_dither_8x8_73[9][8];
853 extern const uint8_t ff_dither_8x8_128[9][8];
854 extern const uint8_t ff_dither_8x8_220[9][8];
855 
856 extern const int32_t ff_yuv2rgb_coeffs[11][4];
857 
858 extern const AVClass ff_sws_context_class;
859 
860 /**
861  * Set c->swscale to an unscaled converter if one exists for the specific
862  * source and destination formats, bit depths, flags, etc.
863  */
864 void ff_get_unscaled_swscale(SwsContext *c);
865 void ff_get_unscaled_swscale_ppc(SwsContext *c);
866 void ff_get_unscaled_swscale_arm(SwsContext *c);
867 void ff_get_unscaled_swscale_aarch64(SwsContext *c);
868 
869 /**
870  * Return function pointer to fastest main scaler path function depending
871  * on architecture and available optimizations.
872  */
873 SwsFunc ff_getSwsFunc(SwsContext *c);
874 
875 void ff_sws_init_input_funcs(SwsContext *c);
876 void ff_sws_init_output_funcs(SwsContext *c,
877                               yuv2planar1_fn *yuv2plane1,
878                               yuv2planarX_fn *yuv2planeX,
879                               yuv2interleavedX_fn *yuv2nv12cX,
880                               yuv2packed1_fn *yuv2packed1,
881                               yuv2packed2_fn *yuv2packed2,
882                               yuv2packedX_fn *yuv2packedX,
883                               yuv2anyX_fn *yuv2anyX);
884 void ff_sws_init_swscale_ppc(SwsContext *c);
885 void ff_sws_init_swscale_vsx(SwsContext *c);
886 void ff_sws_init_swscale_x86(SwsContext *c);
887 void ff_sws_init_swscale_aarch64(SwsContext *c);
888 void ff_sws_init_swscale_arm(SwsContext *c);
889 
890 void ff_hyscale_fast_c(SwsContext *c, int16_t *dst, int dstWidth,
891                        const uint8_t *src, int srcW, int xInc);
892 void ff_hcscale_fast_c(SwsContext *c, int16_t *dst1, int16_t *dst2,
893                        int dstWidth, const uint8_t *src1,
894                        const uint8_t *src2, int srcW, int xInc);
895 int ff_init_hscaler_mmxext(int dstW, int xInc, uint8_t *filterCode,
896                            int16_t *filter, int32_t *filterPos,
897                            int numSplits);
898 void ff_hyscale_fast_mmxext(SwsContext *c, int16_t *dst,
899                             int dstWidth, const uint8_t *src,
900                             int srcW, int xInc);
901 void ff_hcscale_fast_mmxext(SwsContext *c, int16_t *dst1, int16_t *dst2,
902                             int dstWidth, const uint8_t *src1,
903                             const uint8_t *src2, int srcW, int xInc);
904 
905 /**
906  * Allocate and return an SwsContext.
907  * This is like sws_getContext() but does not perform the init step, allowing
908  * the user to set additional AVOptions.
909  *
910  * @see sws_getContext()
911  */
912 struct SwsContext *sws_alloc_set_opts(int srcW, int srcH, enum AVPixelFormat srcFormat,
913                                       int dstW, int dstH, enum AVPixelFormat dstFormat,
914                                       int flags, const double *param);
915 
916 int ff_sws_alphablendaway(SwsContext *c, const uint8_t *src[],
917                           int srcStride[], int srcSliceY, int srcSliceH,
918                           uint8_t *dst[], int dstStride[]);
919 
fillPlane16(uint8_t * plane,int stride,int width,int height,int y,int alpha,int bits,const int big_endian)920 static inline void fillPlane16(uint8_t *plane, int stride, int width, int height, int y,
921                                int alpha, int bits, const int big_endian)
922 {
923     int i, j;
924     uint8_t *ptr = plane + stride * y;
925     int v = alpha ? 0xFFFF>>(16-bits) : (1<<(bits-1));
926     for (i = 0; i < height; i++) {
927 #define FILL(wfunc) \
928         for (j = 0; j < width; j++) {\
929             wfunc(ptr+2*j, v);\
930         }
931         if (big_endian) {
932             FILL(AV_WB16);
933         } else {
934             FILL(AV_WL16);
935         }
936         ptr += stride;
937     }
938 #undef FILL
939 }
940 
fillPlane32(uint8_t * plane,int stride,int width,int height,int y,int alpha,int bits,const int big_endian,int is_float)941 static inline void fillPlane32(uint8_t *plane, int stride, int width, int height, int y,
942                                int alpha, int bits, const int big_endian, int is_float)
943 {
944     int i, j;
945     uint8_t *ptr = plane + stride * y;
946     uint32_t v;
947     uint32_t onef32 = 0x3f800000;
948     if (is_float)
949         v = alpha ? onef32 : 0;
950     else
951         v = alpha ? 0xFFFFFFFF>>(32-bits) : (1<<(bits-1));
952 
953     for (i = 0; i < height; i++) {
954 #define FILL(wfunc) \
955         for (j = 0; j < width; j++) {\
956             wfunc(ptr+4*j, v);\
957         }
958         if (big_endian) {
959             FILL(AV_WB32);
960         } else {
961             FILL(AV_WL32);
962         }
963         ptr += stride;
964     }
965 #undef FILL
966 }
967 
968 
969 #define MAX_SLICE_PLANES 4
970 
971 /// Slice plane
972 typedef struct SwsPlane
973 {
974     int available_lines;    ///< max number of lines that can be hold by this plane
975     int sliceY;             ///< index of first line
976     int sliceH;             ///< number of lines
977     uint8_t **line;         ///< line buffer
978     uint8_t **tmp;          ///< Tmp line buffer used by mmx code
979 } SwsPlane;
980 
981 /**
982  * Struct which defines a slice of an image to be scaled or an output for
983  * a scaled slice.
984  * A slice can also be used as intermediate ring buffer for scaling steps.
985  */
986 typedef struct SwsSlice
987 {
988     int width;              ///< Slice line width
989     int h_chr_sub_sample;   ///< horizontal chroma subsampling factor
990     int v_chr_sub_sample;   ///< vertical chroma subsampling factor
991     int is_ring;            ///< flag to identify if this slice is a ring buffer
992     int should_free_lines;  ///< flag to identify if there are dynamic allocated lines
993     enum AVPixelFormat fmt; ///< planes pixel format
994     SwsPlane plane[MAX_SLICE_PLANES];   ///< color planes
995 } SwsSlice;
996 
997 /**
998  * Struct which holds all necessary data for processing a slice.
999  * A processing step can be a color conversion or horizontal/vertical scaling.
1000  */
1001 typedef struct SwsFilterDescriptor
1002 {
1003     SwsSlice *src;  ///< Source slice
1004     SwsSlice *dst;  ///< Output slice
1005 
1006     int alpha;      ///< Flag for processing alpha channel
1007     void *instance; ///< Filter instance data
1008 
1009     /// Function for processing input slice sliceH lines starting from line sliceY
1010     int (*process)(SwsContext *c, struct SwsFilterDescriptor *desc, int sliceY, int sliceH);
1011 } SwsFilterDescriptor;
1012 
1013 // warp input lines in the form (src + width*i + j) to slice format (line[i][j])
1014 // relative=true means first line src[x][0] otherwise first line is src[x][lum/crh Y]
1015 int ff_init_slice_from_src(SwsSlice * s, uint8_t *src[4], int stride[4], int srcW, int lumY, int lumH, int chrY, int chrH, int relative);
1016 
1017 // Initialize scaler filter descriptor chain
1018 int ff_init_filters(SwsContext *c);
1019 
1020 // Free all filter data
1021 int ff_free_filters(SwsContext *c);
1022 
1023 /*
1024  function for applying ring buffer logic into slice s
1025  It checks if the slice can hold more @lum lines, if yes
1026  do nothing otherwise remove @lum least used lines.
1027  It applies the same procedure for @chr lines.
1028 */
1029 int ff_rotate_slice(SwsSlice *s, int lum, int chr);
1030 
1031 /// initializes gamma conversion descriptor
1032 int ff_init_gamma_convert(SwsFilterDescriptor *desc, SwsSlice * src, uint16_t *table);
1033 
1034 /// initializes lum pixel format conversion descriptor
1035 int ff_init_desc_fmt_convert(SwsFilterDescriptor *desc, SwsSlice * src, SwsSlice *dst, uint32_t *pal);
1036 
1037 /// initializes lum horizontal scaling descriptor
1038 int ff_init_desc_hscale(SwsFilterDescriptor *desc, SwsSlice *src, SwsSlice *dst, uint16_t *filter, int * filter_pos, int filter_size, int xInc);
1039 
1040 /// initializes chr pixel format conversion descriptor
1041 int ff_init_desc_cfmt_convert(SwsFilterDescriptor *desc, SwsSlice * src, SwsSlice *dst, uint32_t *pal);
1042 
1043 /// initializes chr horizontal scaling descriptor
1044 int ff_init_desc_chscale(SwsFilterDescriptor *desc, SwsSlice *src, SwsSlice *dst, uint16_t *filter, int * filter_pos, int filter_size, int xInc);
1045 
1046 int ff_init_desc_no_chr(SwsFilterDescriptor *desc, SwsSlice * src, SwsSlice *dst);
1047 
1048 /// initializes vertical scaling descriptors
1049 int ff_init_vscale(SwsContext *c, SwsFilterDescriptor *desc, SwsSlice *src, SwsSlice *dst);
1050 
1051 /// setup vertical scaler functions
1052 void ff_init_vscale_pfn(SwsContext *c, yuv2planar1_fn yuv2plane1, yuv2planarX_fn yuv2planeX,
1053     yuv2interleavedX_fn yuv2nv12cX, yuv2packed1_fn yuv2packed1, yuv2packed2_fn yuv2packed2,
1054     yuv2packedX_fn yuv2packedX, yuv2anyX_fn yuv2anyX, int use_mmx);
1055 
1056 //number of extra lines to process
1057 #define MAX_LINES_AHEAD 4
1058 
1059 #endif /* SWSCALE_SWSCALE_INTERNAL_H */
1060