falkTX
/
FFmpeg
mirror of https://github.com/falkTX/FFmpeg.git
1
0
Fork 0
Code Issues 0 Releases 338 Wiki Activity
You can not select more than 25 topics Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.
91973 Commits
32 Branches
283MB
Tree: bd10c1e9a8
Branches Tags
${ item.name }
Create branch ${ searchTerm }
from 'bd10c1e9a8'
${ noResults }
FFmpeg/libswscale/utils.c

2426 lines
86KB
Raw Blame History 

  1. /*

  2.  * Copyright (C) 2001-2003 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. #include "config.h"

  22. 

  23. #define _DEFAULT_SOURCE

  24. #define _SVID_SOURCE // needed for MAP_ANONYMOUS

  25. #define _DARWIN_C_SOURCE // needed for MAP_ANON

  26. #include <inttypes.h>

  27. #include <math.h>

  28. #include <stdio.h>

  29. #include <string.h>

  30. #if HAVE_MMAP

  31. #include <sys/mman.h>

  32. #if defined(MAP_ANON) && !defined(MAP_ANONYMOUS)

  33. #define MAP_ANONYMOUS MAP_ANON

  34. #endif

  35. #endif

  36. #if HAVE_VIRTUALALLOC

  37. #define WIN32_LEAN_AND_MEAN

  38. #include <windows.h>

  39. #endif

  40. 

  41. #include "libavutil/attributes.h"

  42. #include "libavutil/avassert.h"

  43. #include "libavutil/avutil.h"

  44. #include "libavutil/bswap.h"

  45. #include "libavutil/cpu.h"

  46. #include "libavutil/imgutils.h"

  47. #include "libavutil/intreadwrite.h"

  48. #include "libavutil/libm.h"

  49. #include "libavutil/mathematics.h"

  50. #include "libavutil/opt.h"

  51. #include "libavutil/pixdesc.h"

  52. #include "libavutil/aarch64/cpu.h"

  53. #include "libavutil/ppc/cpu.h"

  54. #include "libavutil/x86/asm.h"

  55. #include "libavutil/x86/cpu.h"

  56. 

  57. // We have to implement deprecated functions until they are removed, this is the

  58. // simplest way to prevent warnings

  59. #undef attribute_deprecated

  60. #define attribute_deprecated

  61. 

  62. #include "rgb2rgb.h"

  63. #include "swscale.h"

  64. #include "swscale_internal.h"

  65. 

  66. #if !FF_API_SWS_VECTOR

  67. static SwsVector *sws_getIdentityVec(void);

  68. static void sws_addVec(SwsVector *a, SwsVector *b);

  69. static void sws_shiftVec(SwsVector *a, int shift);

  70. static void sws_printVec2(SwsVector *a, AVClass *log_ctx, int log_level);

  71. #endif

  72. 

  73. static void handle_formats(SwsContext *c);

  74. 

  75. unsigned swscale_version(void)

  76. {

  77.     av_assert0(LIBSWSCALE_VERSION_MICRO >= 100);

  78.     return LIBSWSCALE_VERSION_INT;

  79. }

  80. 

  81. const char *swscale_configuration(void)

  82. {

  83.     return FFMPEG_CONFIGURATION;

  84. }

  85. 

  86. const char *swscale_license(void)

  87. {

  88. #define LICENSE_PREFIX "libswscale license: "

  89.     return LICENSE_PREFIX FFMPEG_LICENSE + sizeof(LICENSE_PREFIX) - 1;

  90. }

  91. 

  92. typedef struct FormatEntry {

  93.     uint8_t is_supported_in         :1;

  94.     uint8_t is_supported_out        :1;

  95.     uint8_t is_supported_endianness :1;

  96. } FormatEntry;

  97. 

  98. static const FormatEntry format_entries[AV_PIX_FMT_NB] = {

  99.     [AV_PIX_FMT_YUV420P]     = { 1, 1 },

  100.     [AV_PIX_FMT_YUYV422]     = { 1, 1 },

  101.     [AV_PIX_FMT_RGB24]       = { 1, 1 },

  102.     [AV_PIX_FMT_BGR24]       = { 1, 1 },

  103.     [AV_PIX_FMT_YUV422P]     = { 1, 1 },

  104.     [AV_PIX_FMT_YUV444P]     = { 1, 1 },

  105.     [AV_PIX_FMT_YUV410P]     = { 1, 1 },

  106.     [AV_PIX_FMT_YUV411P]     = { 1, 1 },

  107.     [AV_PIX_FMT_GRAY8]       = { 1, 1 },

  108.     [AV_PIX_FMT_MONOWHITE]   = { 1, 1 },

  109.     [AV_PIX_FMT_MONOBLACK]   = { 1, 1 },

  110.     [AV_PIX_FMT_PAL8]        = { 1, 0 },

  111.     [AV_PIX_FMT_YUVJ420P]    = { 1, 1 },

  112.     [AV_PIX_FMT_YUVJ411P]    = { 1, 1 },

  113.     [AV_PIX_FMT_YUVJ422P]    = { 1, 1 },

  114.     [AV_PIX_FMT_YUVJ444P]    = { 1, 1 },

  115.     [AV_PIX_FMT_YVYU422]     = { 1, 1 },

  116.     [AV_PIX_FMT_UYVY422]     = { 1, 1 },

  117.     [AV_PIX_FMT_UYYVYY411]   = { 0, 0 },

  118.     [AV_PIX_FMT_BGR8]        = { 1, 1 },

  119.     [AV_PIX_FMT_BGR4]        = { 0, 1 },

  120.     [AV_PIX_FMT_BGR4_BYTE]   = { 1, 1 },

  121.     [AV_PIX_FMT_RGB8]        = { 1, 1 },

  122.     [AV_PIX_FMT_RGB4]        = { 0, 1 },

  123.     [AV_PIX_FMT_RGB4_BYTE]   = { 1, 1 },

  124.     [AV_PIX_FMT_NV12]        = { 1, 1 },

  125.     [AV_PIX_FMT_NV21]        = { 1, 1 },

  126.     [AV_PIX_FMT_ARGB]        = { 1, 1 },

  127.     [AV_PIX_FMT_RGBA]        = { 1, 1 },

  128.     [AV_PIX_FMT_ABGR]        = { 1, 1 },

  129.     [AV_PIX_FMT_BGRA]        = { 1, 1 },

  130.     [AV_PIX_FMT_0RGB]        = { 1, 1 },

  131.     [AV_PIX_FMT_RGB0]        = { 1, 1 },

  132.     [AV_PIX_FMT_0BGR]        = { 1, 1 },

  133.     [AV_PIX_FMT_BGR0]        = { 1, 1 },

  134.     [AV_PIX_FMT_GRAY9BE]     = { 1, 1 },

  135.     [AV_PIX_FMT_GRAY9LE]     = { 1, 1 },

  136.     [AV_PIX_FMT_GRAY10BE]    = { 1, 1 },

  137.     [AV_PIX_FMT_GRAY10LE]    = { 1, 1 },

  138.     [AV_PIX_FMT_GRAY12BE]    = { 1, 1 },

  139.     [AV_PIX_FMT_GRAY12LE]    = { 1, 1 },

  140.     [AV_PIX_FMT_GRAY14BE]    = { 1, 1 },

  141.     [AV_PIX_FMT_GRAY14LE]    = { 1, 1 },

  142.     [AV_PIX_FMT_GRAY16BE]    = { 1, 1 },

  143.     [AV_PIX_FMT_GRAY16LE]    = { 1, 1 },

  144.     [AV_PIX_FMT_YUV440P]     = { 1, 1 },

  145.     [AV_PIX_FMT_YUVJ440P]    = { 1, 1 },

  146.     [AV_PIX_FMT_YUV440P10LE] = { 1, 1 },

  147.     [AV_PIX_FMT_YUV440P10BE] = { 1, 1 },

  148.     [AV_PIX_FMT_YUV440P12LE] = { 1, 1 },

  149.     [AV_PIX_FMT_YUV440P12BE] = { 1, 1 },

  150.     [AV_PIX_FMT_YUVA420P]    = { 1, 1 },

  151.     [AV_PIX_FMT_YUVA422P]    = { 1, 1 },

  152.     [AV_PIX_FMT_YUVA444P]    = { 1, 1 },

  153.     [AV_PIX_FMT_YUVA420P9BE] = { 1, 1 },

  154.     [AV_PIX_FMT_YUVA420P9LE] = { 1, 1 },

  155.     [AV_PIX_FMT_YUVA422P9BE] = { 1, 1 },

  156.     [AV_PIX_FMT_YUVA422P9LE] = { 1, 1 },

  157.     [AV_PIX_FMT_YUVA444P9BE] = { 1, 1 },

  158.     [AV_PIX_FMT_YUVA444P9LE] = { 1, 1 },

  159.     [AV_PIX_FMT_YUVA420P10BE]= { 1, 1 },

  160.     [AV_PIX_FMT_YUVA420P10LE]= { 1, 1 },

  161.     [AV_PIX_FMT_YUVA422P10BE]= { 1, 1 },

  162.     [AV_PIX_FMT_YUVA422P10LE]= { 1, 1 },

  163.     [AV_PIX_FMT_YUVA444P10BE]= { 1, 1 },

  164.     [AV_PIX_FMT_YUVA444P10LE]= { 1, 1 },

  165.     [AV_PIX_FMT_YUVA420P16BE]= { 1, 1 },

  166.     [AV_PIX_FMT_YUVA420P16LE]= { 1, 1 },

  167.     [AV_PIX_FMT_YUVA422P16BE]= { 1, 1 },

  168.     [AV_PIX_FMT_YUVA422P16LE]= { 1, 1 },

  169.     [AV_PIX_FMT_YUVA444P16BE]= { 1, 1 },

  170.     [AV_PIX_FMT_YUVA444P16LE]= { 1, 1 },

  171.     [AV_PIX_FMT_RGB48BE]     = { 1, 1 },

  172.     [AV_PIX_FMT_RGB48LE]     = { 1, 1 },

  173.     [AV_PIX_FMT_RGBA64BE]    = { 1, 1, 1 },

  174.     [AV_PIX_FMT_RGBA64LE]    = { 1, 1, 1 },

  175.     [AV_PIX_FMT_RGB565BE]    = { 1, 1 },

  176.     [AV_PIX_FMT_RGB565LE]    = { 1, 1 },

  177.     [AV_PIX_FMT_RGB555BE]    = { 1, 1 },

  178.     [AV_PIX_FMT_RGB555LE]    = { 1, 1 },

  179.     [AV_PIX_FMT_BGR565BE]    = { 1, 1 },

  180.     [AV_PIX_FMT_BGR565LE]    = { 1, 1 },

  181.     [AV_PIX_FMT_BGR555BE]    = { 1, 1 },

  182.     [AV_PIX_FMT_BGR555LE]    = { 1, 1 },

  183.     [AV_PIX_FMT_YUV420P16LE] = { 1, 1 },

  184.     [AV_PIX_FMT_YUV420P16BE] = { 1, 1 },

  185.     [AV_PIX_FMT_YUV422P16LE] = { 1, 1 },

  186.     [AV_PIX_FMT_YUV422P16BE] = { 1, 1 },

  187.     [AV_PIX_FMT_YUV444P16LE] = { 1, 1 },

  188.     [AV_PIX_FMT_YUV444P16BE] = { 1, 1 },

  189.     [AV_PIX_FMT_RGB444LE]    = { 1, 1 },

  190.     [AV_PIX_FMT_RGB444BE]    = { 1, 1 },

  191.     [AV_PIX_FMT_BGR444LE]    = { 1, 1 },

  192.     [AV_PIX_FMT_BGR444BE]    = { 1, 1 },

  193.     [AV_PIX_FMT_YA8]         = { 1, 1 },

  194.     [AV_PIX_FMT_YA16BE]      = { 1, 0 },

  195.     [AV_PIX_FMT_YA16LE]      = { 1, 0 },

  196.     [AV_PIX_FMT_BGR48BE]     = { 1, 1 },

  197.     [AV_PIX_FMT_BGR48LE]     = { 1, 1 },

  198.     [AV_PIX_FMT_BGRA64BE]    = { 1, 1, 1 },

  199.     [AV_PIX_FMT_BGRA64LE]    = { 1, 1, 1 },

  200.     [AV_PIX_FMT_YUV420P9BE]  = { 1, 1 },

  201.     [AV_PIX_FMT_YUV420P9LE]  = { 1, 1 },

  202.     [AV_PIX_FMT_YUV420P10BE] = { 1, 1 },

  203.     [AV_PIX_FMT_YUV420P10LE] = { 1, 1 },

  204.     [AV_PIX_FMT_YUV420P12BE] = { 1, 1 },

  205.     [AV_PIX_FMT_YUV420P12LE] = { 1, 1 },

  206.     [AV_PIX_FMT_YUV420P14BE] = { 1, 1 },

  207.     [AV_PIX_FMT_YUV420P14LE] = { 1, 1 },

  208.     [AV_PIX_FMT_YUV422P9BE]  = { 1, 1 },

  209.     [AV_PIX_FMT_YUV422P9LE]  = { 1, 1 },

  210.     [AV_PIX_FMT_YUV422P10BE] = { 1, 1 },

  211.     [AV_PIX_FMT_YUV422P10LE] = { 1, 1 },

  212.     [AV_PIX_FMT_YUV422P12BE] = { 1, 1 },

  213.     [AV_PIX_FMT_YUV422P12LE] = { 1, 1 },

  214.     [AV_PIX_FMT_YUV422P14BE] = { 1, 1 },

  215.     [AV_PIX_FMT_YUV422P14LE] = { 1, 1 },

  216.     [AV_PIX_FMT_YUV444P9BE]  = { 1, 1 },

  217.     [AV_PIX_FMT_YUV444P9LE]  = { 1, 1 },

  218.     [AV_PIX_FMT_YUV444P10BE] = { 1, 1 },

  219.     [AV_PIX_FMT_YUV444P10LE] = { 1, 1 },

  220.     [AV_PIX_FMT_YUV444P12BE] = { 1, 1 },

  221.     [AV_PIX_FMT_YUV444P12LE] = { 1, 1 },

  222.     [AV_PIX_FMT_YUV444P14BE] = { 1, 1 },

  223.     [AV_PIX_FMT_YUV444P14LE] = { 1, 1 },

  224.     [AV_PIX_FMT_GBRP]        = { 1, 1 },

  225.     [AV_PIX_FMT_GBRP9LE]     = { 1, 1 },

  226.     [AV_PIX_FMT_GBRP9BE]     = { 1, 1 },

  227.     [AV_PIX_FMT_GBRP10LE]    = { 1, 1 },

  228.     [AV_PIX_FMT_GBRP10BE]    = { 1, 1 },

  229.     [AV_PIX_FMT_GBRAP10LE]   = { 1, 1 },

  230.     [AV_PIX_FMT_GBRAP10BE]   = { 1, 1 },

  231.     [AV_PIX_FMT_GBRP12LE]    = { 1, 1 },

  232.     [AV_PIX_FMT_GBRP12BE]    = { 1, 1 },

  233.     [AV_PIX_FMT_GBRAP12LE]   = { 1, 1 },

  234.     [AV_PIX_FMT_GBRAP12BE]   = { 1, 1 },

  235.     [AV_PIX_FMT_GBRP14LE]    = { 1, 1 },

  236.     [AV_PIX_FMT_GBRP14BE]    = { 1, 1 },

  237.     [AV_PIX_FMT_GBRP16LE]    = { 1, 1 },

  238.     [AV_PIX_FMT_GBRP16BE]    = { 1, 1 },

  239.     [AV_PIX_FMT_GBRAP]       = { 1, 1 },

  240.     [AV_PIX_FMT_GBRAP16LE]   = { 1, 1 },

  241.     [AV_PIX_FMT_GBRAP16BE]   = { 1, 1 },

  242.     [AV_PIX_FMT_BAYER_BGGR8] = { 1, 0 },

  243.     [AV_PIX_FMT_BAYER_RGGB8] = { 1, 0 },

  244.     [AV_PIX_FMT_BAYER_GBRG8] = { 1, 0 },

  245.     [AV_PIX_FMT_BAYER_GRBG8] = { 1, 0 },

  246.     [AV_PIX_FMT_BAYER_BGGR16LE] = { 1, 0 },

  247.     [AV_PIX_FMT_BAYER_BGGR16BE] = { 1, 0 },

  248.     [AV_PIX_FMT_BAYER_RGGB16LE] = { 1, 0 },

  249.     [AV_PIX_FMT_BAYER_RGGB16BE] = { 1, 0 },

  250.     [AV_PIX_FMT_BAYER_GBRG16LE] = { 1, 0 },

  251.     [AV_PIX_FMT_BAYER_GBRG16BE] = { 1, 0 },

  252.     [AV_PIX_FMT_BAYER_GRBG16LE] = { 1, 0 },

  253.     [AV_PIX_FMT_BAYER_GRBG16BE] = { 1, 0 },

  254.     [AV_PIX_FMT_XYZ12BE]     = { 1, 1, 1 },

  255.     [AV_PIX_FMT_XYZ12LE]     = { 1, 1, 1 },

  256.     [AV_PIX_FMT_AYUV64LE]    = { 1, 1},

  257.     [AV_PIX_FMT_P010LE]      = { 1, 1 },

  258.     [AV_PIX_FMT_P010BE]      = { 1, 1 },

  259.     [AV_PIX_FMT_P016LE]      = { 1, 1 },

  260.     [AV_PIX_FMT_P016BE]      = { 1, 1 },

  261.     [AV_PIX_FMT_GRAYF32LE]   = { 1, 1 },

  262.     [AV_PIX_FMT_GRAYF32BE]   = { 1, 1 },

  263. };

  264. 

  265. int sws_isSupportedInput(enum AVPixelFormat pix_fmt)

  266. {

  267.     return (unsigned)pix_fmt < AV_PIX_FMT_NB ?

  268.            format_entries[pix_fmt].is_supported_in : 0;

  269. }

  270. 

  271. int sws_isSupportedOutput(enum AVPixelFormat pix_fmt)

  272. {

  273.     return (unsigned)pix_fmt < AV_PIX_FMT_NB ?

  274.            format_entries[pix_fmt].is_supported_out : 0;

  275. }

  276. 

  277. int sws_isSupportedEndiannessConversion(enum AVPixelFormat pix_fmt)

  278. {

  279.     return (unsigned)pix_fmt < AV_PIX_FMT_NB ?

  280.            format_entries[pix_fmt].is_supported_endianness : 0;

  281. }

  282. 

  283. static double getSplineCoeff(double a, double b, double c, double d,

  284.                              double dist)

  285. {

  286.     if (dist <= 1.0)

  287.         return ((d * dist + c) * dist + b) * dist + a;

  288.     else

  289.         return getSplineCoeff(0.0,

  290.                                b + 2.0 * c + 3.0 * d,

  291.                                c + 3.0 * d,

  292.                               -b - 3.0 * c - 6.0 * d,

  293.                               dist - 1.0);

  294. }

  295. 

  296. static av_cold int get_local_pos(SwsContext *s, int chr_subsample, int pos, int dir)

  297. {

  298.     if (pos == -1 || pos <= -513) {

  299.         pos = (128 << chr_subsample) - 128;

  300.     }

  301.     pos += 128; // relative to ideal left edge

  302.     return pos >> chr_subsample;

  303. }

  304. 

  305. typedef struct {

  306.     int flag;                   ///< flag associated to the algorithm

  307.     const char *description;    ///< human-readable description

  308.     int size_factor;            ///< size factor used when initing the filters

  309. } ScaleAlgorithm;

  310. 

  311. static const ScaleAlgorithm scale_algorithms[] = {

  312.     { SWS_AREA,          "area averaging",                  1 /* downscale only, for upscale it is bilinear */ },

  313.     { SWS_BICUBIC,       "bicubic",                         4 },

  314.     { SWS_BICUBLIN,      "luma bicubic / chroma bilinear", -1 },

  315.     { SWS_BILINEAR,      "bilinear",                        2 },

  316.     { SWS_FAST_BILINEAR, "fast bilinear",                  -1 },

  317.     { SWS_GAUSS,         "Gaussian",                        8 /* infinite ;) */ },

  318.     { SWS_LANCZOS,       "Lanczos",                        -1 /* custom */ },

  319.     { SWS_POINT,         "nearest neighbor / point",       -1 },

  320.     { SWS_SINC,          "sinc",                           20 /* infinite ;) */ },

  321.     { SWS_SPLINE,        "bicubic spline",                 20 /* infinite :)*/ },

  322.     { SWS_X,             "experimental",                    8 },

  323. };

  324. 

  325. static av_cold int initFilter(int16_t **outFilter, int32_t **filterPos,

  326.                               int *outFilterSize, int xInc, int srcW,

  327.                               int dstW, int filterAlign, int one,

  328.                               int flags, int cpu_flags,

  329.                               SwsVector *srcFilter, SwsVector *dstFilter,

  330.                               double param[2], int srcPos, int dstPos)

  331. {

  332.     int i;

  333.     int filterSize;

  334.     int filter2Size;

  335.     int minFilterSize;

  336.     int64_t *filter    = NULL;

  337.     int64_t *filter2   = NULL;

  338.     const int64_t fone = 1LL << (54 - FFMIN(av_log2(srcW/dstW), 8));

  339.     int ret            = -1;

  340. 

  341.     emms_c(); // FIXME should not be required but IS (even for non-MMX versions)

  342. 

  343.     // NOTE: the +3 is for the MMX(+1) / SSE(+3) scaler which reads over the end

  344.     FF_ALLOC_ARRAY_OR_GOTO(NULL, *filterPos, (dstW + 3), sizeof(**filterPos), fail);

  345. 

  346.     if (FFABS(xInc - 0x10000) < 10 && srcPos == dstPos) { // unscaled

  347.         int i;

  348.         filterSize = 1;

  349.         FF_ALLOCZ_ARRAY_OR_GOTO(NULL, filter,

  350.                                 dstW, sizeof(*filter) * filterSize, fail);

  351. 

  352.         for (i = 0; i < dstW; i++) {

  353.             filter[i * filterSize] = fone;

  354.             (*filterPos)[i]        = i;

  355.         }

  356.     } else if (flags & SWS_POINT) { // lame looking point sampling mode

  357.         int i;

  358.         int64_t xDstInSrc;

  359.         filterSize = 1;

  360.         FF_ALLOC_ARRAY_OR_GOTO(NULL, filter,

  361.                                dstW, sizeof(*filter) * filterSize, fail);

  362. 

  363.         xDstInSrc = ((dstPos*(int64_t)xInc)>>8) - ((srcPos*0x8000LL)>>7);

  364.         for (i = 0; i < dstW; i++) {

  365.             int xx = (xDstInSrc - ((filterSize - 1) << 15) + (1 << 15)) >> 16;

  366. 

  367.             (*filterPos)[i] = xx;

  368.             filter[i]       = fone;

  369.             xDstInSrc      += xInc;

  370.         }

  371.     } else if ((xInc <= (1 << 16) && (flags & SWS_AREA)) ||

  372.                (flags & SWS_FAST_BILINEAR)) { // bilinear upscale

  373.         int i;

  374.         int64_t xDstInSrc;

  375.         filterSize = 2;

  376.         FF_ALLOC_ARRAY_OR_GOTO(NULL, filter,

  377.                                dstW, sizeof(*filter) * filterSize, fail);

  378. 

  379.         xDstInSrc = ((dstPos*(int64_t)xInc)>>8) - ((srcPos*0x8000LL)>>7);

  380.         for (i = 0; i < dstW; i++) {

  381.             int xx = (xDstInSrc - ((filterSize - 1) << 15) + (1 << 15)) >> 16;

  382.             int j;

  383. 

  384.             (*filterPos)[i] = xx;

  385.             // bilinear upscale / linear interpolate / area averaging

  386.             for (j = 0; j < filterSize; j++) {

  387.                 int64_t coeff= fone - FFABS(((int64_t)xx<<16) - xDstInSrc)*(fone>>16);

  388.                 if (coeff < 0)

  389.                     coeff = 0;

  390.                 filter[i * filterSize + j] = coeff;

  391.                 xx++;

  392.             }

  393.             xDstInSrc += xInc;

  394.         }

  395.     } else {

  396.         int64_t xDstInSrc;

  397.         int sizeFactor = -1;

  398. 

  399.         for (i = 0; i < FF_ARRAY_ELEMS(scale_algorithms); i++) {

  400.             if (flags & scale_algorithms[i].flag && scale_algorithms[i].size_factor > 0) {

  401.                 sizeFactor = scale_algorithms[i].size_factor;

  402.                 break;

  403.             }

  404.         }

  405.         if (flags & SWS_LANCZOS)

  406.             sizeFactor = param[0] != SWS_PARAM_DEFAULT ? ceil(2 * param[0]) : 6;

  407.         av_assert0(sizeFactor > 0);

  408. 

  409.         if (xInc <= 1 << 16)

  410.             filterSize = 1 + sizeFactor;    // upscale

  411.         else

  412.             filterSize = 1 + (sizeFactor * srcW + dstW - 1) / dstW;

  413. 

  414.         filterSize = FFMIN(filterSize, srcW - 2);

  415.         filterSize = FFMAX(filterSize, 1);

  416. 

  417.         FF_ALLOC_ARRAY_OR_GOTO(NULL, filter,

  418.                                dstW, sizeof(*filter) * filterSize, fail);

  419. 

  420.         xDstInSrc = ((dstPos*(int64_t)xInc)>>7) - ((srcPos*0x10000LL)>>7);

  421.         for (i = 0; i < dstW; i++) {

  422.             int xx = (xDstInSrc - (filterSize - 2) * (1LL<<16)) / (1 << 17);

  423.             int j;

  424.             (*filterPos)[i] = xx;

  425.             for (j = 0; j < filterSize; j++) {

  426.                 int64_t d = (FFABS(((int64_t)xx * (1 << 17)) - xDstInSrc)) << 13;

  427.                 double floatd;

  428.                 int64_t coeff;

  429. 

  430.                 if (xInc > 1 << 16)

  431.                     d = d * dstW / srcW;

  432.                 floatd = d * (1.0 / (1 << 30));

  433. 

  434.                 if (flags & SWS_BICUBIC) {

  435.                     int64_t B = (param[0] != SWS_PARAM_DEFAULT ? param[0] :   0) * (1 << 24);

  436.                     int64_t C = (param[1] != SWS_PARAM_DEFAULT ? param[1] : 0.6) * (1 << 24);

  437. 

  438.                     if (d >= 1LL << 31) {

  439.                         coeff = 0.0;

  440.                     } else {

  441.                         int64_t dd  = (d  * d) >> 30;

  442.                         int64_t ddd = (dd * d) >> 30;

  443. 

  444.                         if (d < 1LL << 30)

  445.                             coeff =  (12 * (1 << 24) -  9 * B - 6 * C) * ddd +

  446.                                     (-18 * (1 << 24) + 12 * B + 6 * C) *  dd +

  447.                                       (6 * (1 << 24) -  2 * B)         * (1 << 30);

  448.                         else

  449.                             coeff =      (-B -  6 * C) * ddd +

  450.                                       (6 * B + 30 * C) * dd  +

  451.                                     (-12 * B - 48 * C) * d   +

  452.                                       (8 * B + 24 * C) * (1 << 30);

  453.                     }

  454.                     coeff /= (1LL<<54)/fone;

  455.                 } else if (flags & SWS_X) {

  456.                     double A = param[0] != SWS_PARAM_DEFAULT ? param[0] : 1.0;

  457.                     double c;

  458. 

  459.                     if (floatd < 1.0)

  460.                         c = cos(floatd * M_PI);

  461.                     else

  462.                         c = -1.0;

  463.                     if (c < 0.0)

  464.                         c = -pow(-c, A);

  465.                     else

  466.                         c = pow(c, A);

  467.                     coeff = (c * 0.5 + 0.5) * fone;

  468.                 } else if (flags & SWS_AREA) {

  469.                     int64_t d2 = d - (1 << 29);

  470.                     if (d2 * xInc < -(1LL << (29 + 16)))

  471.                         coeff = 1.0 * (1LL << (30 + 16));

  472.                     else if (d2 * xInc < (1LL << (29 + 16)))

  473.                         coeff = -d2 * xInc + (1LL << (29 + 16));

  474.                     else

  475.                         coeff = 0.0;

  476.                     coeff *= fone >> (30 + 16);

  477.                 } else if (flags & SWS_GAUSS) {

  478.                     double p = param[0] != SWS_PARAM_DEFAULT ? param[0] : 3.0;

  479.                     coeff = exp2(-p * floatd * floatd) * fone;

  480.                 } else if (flags & SWS_SINC) {

  481.                     coeff = (d ? sin(floatd * M_PI) / (floatd * M_PI) : 1.0) * fone;

  482.                 } else if (flags & SWS_LANCZOS) {

  483.                     double p = param[0] != SWS_PARAM_DEFAULT ? param[0] : 3.0;

  484.                     coeff = (d ? sin(floatd * M_PI) * sin(floatd * M_PI / p) /

  485.                              (floatd * floatd * M_PI * M_PI / p) : 1.0) * fone;

  486.                     if (floatd > p)

  487.                         coeff = 0;

  488.                 } else if (flags & SWS_BILINEAR) {

  489.                     coeff = (1 << 30) - d;

  490.                     if (coeff < 0)

  491.                         coeff = 0;

  492.                     coeff *= fone >> 30;

  493.                 } else if (flags & SWS_SPLINE) {

  494.                     double p = -2.196152422706632;

  495.                     coeff = getSplineCoeff(1.0, 0.0, p, -p - 1.0, floatd) * fone;

  496.                 } else {

  497.                     av_assert0(0);

  498.                 }

  499. 

  500.                 filter[i * filterSize + j] = coeff;

  501.                 xx++;

  502.             }

  503.             xDstInSrc += 2 * xInc;

  504.         }

  505.     }

  506. 

  507.     /* apply src & dst Filter to filter -> filter2

  508.      * av_free(filter);

  509.      */

  510.     av_assert0(filterSize > 0);

  511.     filter2Size = filterSize;

  512.     if (srcFilter)

  513.         filter2Size += srcFilter->length - 1;

  514.     if (dstFilter)

  515.         filter2Size += dstFilter->length - 1;

  516.     av_assert0(filter2Size > 0);

  517.     FF_ALLOCZ_ARRAY_OR_GOTO(NULL, filter2, dstW, filter2Size * sizeof(*filter2), fail);

  518. 

  519.     for (i = 0; i < dstW; i++) {

  520.         int j, k;

  521. 

  522.         if (srcFilter) {

  523.             for (k = 0; k < srcFilter->length; k++) {

  524.                 for (j = 0; j < filterSize; j++)

  525.                     filter2[i * filter2Size + k + j] +=

  526.                         srcFilter->coeff[k] * filter[i * filterSize + j];

  527.             }

  528.         } else {

  529.             for (j = 0; j < filterSize; j++)

  530.                 filter2[i * filter2Size + j] = filter[i * filterSize + j];

  531.         }

  532.         // FIXME dstFilter

  533. 

  534.         (*filterPos)[i] += (filterSize - 1) / 2 - (filter2Size - 1) / 2;

  535.     }

  536.     av_freep(&filter);

  537. 

  538.     /* try to reduce the filter-size (step1 find size and shift left) */

  539.     // Assume it is near normalized (*0.5 or *2.0 is OK but * 0.001 is not).

  540.     minFilterSize = 0;

  541.     for (i = dstW - 1; i >= 0; i--) {

  542.         int min = filter2Size;

  543.         int j;

  544.         int64_t cutOff = 0.0;

  545. 

  546.         /* get rid of near zero elements on the left by shifting left */

  547.         for (j = 0; j < filter2Size; j++) {

  548.             int k;

  549.             cutOff += FFABS(filter2[i * filter2Size]);

  550. 

  551.             if (cutOff > SWS_MAX_REDUCE_CUTOFF * fone)

  552.                 break;

  553. 

  554.             /* preserve monotonicity because the core can't handle the

  555.              * filter otherwise */

  556.             if (i < dstW - 1 && (*filterPos)[i] >= (*filterPos)[i + 1])

  557.                 break;

  558. 

  559.             // move filter coefficients left

  560.             for (k = 1; k < filter2Size; k++)

  561.                 filter2[i * filter2Size + k - 1] = filter2[i * filter2Size + k];

  562.             filter2[i * filter2Size + k - 1] = 0;

  563.             (*filterPos)[i]++;

  564.         }

  565. 

  566.         cutOff = 0;

  567.         /* count near zeros on the right */

  568.         for (j = filter2Size - 1; j > 0; j--) {

  569.             cutOff += FFABS(filter2[i * filter2Size + j]);

  570. 

  571.             if (cutOff > SWS_MAX_REDUCE_CUTOFF * fone)

  572.                 break;

  573.             min--;

  574.         }

  575. 

  576.         if (min > minFilterSize)

  577.             minFilterSize = min;

  578.     }

  579. 

  580.     if (PPC_ALTIVEC(cpu_flags)) {

  581.         // we can handle the special case 4, so we don't want to go the full 8

  582.         if (minFilterSize < 5)

  583.             filterAlign = 4;

  584. 

  585.         /* We really don't want to waste our time doing useless computation, so

  586.          * fall back on the scalar C code for very small filters.

  587.          * Vectorizing is worth it only if you have a decent-sized vector. */

  588.         if (minFilterSize < 3)

  589.             filterAlign = 1;

  590.     }

  591. 

  592.     if (HAVE_MMX && cpu_flags & AV_CPU_FLAG_MMX) {

  593.         // special case for unscaled vertical filtering

  594.         if (minFilterSize == 1 && filterAlign == 2)

  595.             filterAlign = 1;

  596.     }

  597. 

  598.     av_assert0(minFilterSize > 0);

  599.     filterSize = (minFilterSize + (filterAlign - 1)) & (~(filterAlign - 1));

  600.     av_assert0(filterSize > 0);

  601.     filter = av_malloc_array(dstW, filterSize * sizeof(*filter));

  602.     if (!filter)

  603.         goto fail;

  604.     if (filterSize >= MAX_FILTER_SIZE * 16 /

  605.                       ((flags & SWS_ACCURATE_RND) ? APCK_SIZE : 16)) {

  606.         ret = RETCODE_USE_CASCADE;

  607.         goto fail;

  608.     }

  609.     *outFilterSize = filterSize;

  610. 

  611.     if (flags & SWS_PRINT_INFO)

  612.         av_log(NULL, AV_LOG_VERBOSE,

  613.                "SwScaler: reducing / aligning filtersize %d -> %d\n",

  614.                filter2Size, filterSize);

  615.     /* try to reduce the filter-size (step2 reduce it) */

  616.     for (i = 0; i < dstW; i++) {

  617.         int j;

  618. 

  619.         for (j = 0; j < filterSize; j++) {

  620.             if (j >= filter2Size)

  621.                 filter[i * filterSize + j] = 0;

  622.             else

  623.                 filter[i * filterSize + j] = filter2[i * filter2Size + j];

  624.             if ((flags & SWS_BITEXACT) && j >= minFilterSize)

  625.                 filter[i * filterSize + j] = 0;

  626.         }

  627.     }

  628. 

  629.     // FIXME try to align filterPos if possible

  630. 

  631.     // fix borders

  632.     for (i = 0; i < dstW; i++) {

  633.         int j;

  634.         if ((*filterPos)[i] < 0) {

  635.             // move filter coefficients left to compensate for filterPos

  636.             for (j = 1; j < filterSize; j++) {

  637.                 int left = FFMAX(j + (*filterPos)[i], 0);

  638.                 filter[i * filterSize + left] += filter[i * filterSize + j];

  639.                 filter[i * filterSize + j]     = 0;

  640.             }

  641.             (*filterPos)[i]= 0;

  642.         }

  643. 

  644.         if ((*filterPos)[i] + filterSize > srcW) {

  645.             int shift = (*filterPos)[i] + FFMIN(filterSize - srcW, 0);

  646.             int64_t acc = 0;

  647. 

  648.             for (j = filterSize - 1; j >= 0; j--) {

  649.                 if ((*filterPos)[i] + j >= srcW) {

  650.                     acc += filter[i * filterSize + j];

  651.                     filter[i * filterSize + j] = 0;

  652.                 }

  653.             }

  654.             for (j = filterSize - 1; j >= 0; j--) {

  655.                 if (j < shift) {

  656.                     filter[i * filterSize + j] = 0;

  657.                 } else {

  658.                     filter[i * filterSize + j] = filter[i * filterSize + j - shift];

  659.                 }

  660.             }

  661. 

  662.             (*filterPos)[i]-= shift;

  663.             filter[i * filterSize + srcW - 1 - (*filterPos)[i]] += acc;

  664.         }

  665.         av_assert0((*filterPos)[i] >= 0);

  666.         av_assert0((*filterPos)[i] < srcW);

  667.         if ((*filterPos)[i] + filterSize > srcW) {

  668.             for (j = 0; j < filterSize; j++) {

  669.                 av_assert0((*filterPos)[i] + j < srcW || !filter[i * filterSize + j]);

  670.             }

  671.         }

  672.     }

  673. 

  674.     // Note the +1 is for the MMX scaler which reads over the end

  675.     /* align at 16 for AltiVec (needed by hScale_altivec_real) */

  676.     FF_ALLOCZ_ARRAY_OR_GOTO(NULL, *outFilter,

  677.                             (dstW + 3), *outFilterSize * sizeof(int16_t), fail);

  678. 

  679.     /* normalize & store in outFilter */

  680.     for (i = 0; i < dstW; i++) {

  681.         int j;

  682.         int64_t error = 0;

  683.         int64_t sum   = 0;

  684. 

  685.         for (j = 0; j < filterSize; j++) {

  686.             sum += filter[i * filterSize + j];

  687.         }

  688.         sum = (sum + one / 2) / one;

  689.         if (!sum) {

  690.             av_log(NULL, AV_LOG_WARNING, "SwScaler: zero vector in scaling\n");

  691.             sum = 1;

  692.         }

  693.         for (j = 0; j < *outFilterSize; j++) {

  694.             int64_t v = filter[i * filterSize + j] + error;

  695.             int intV  = ROUNDED_DIV(v, sum);

  696.             (*outFilter)[i * (*outFilterSize) + j] = intV;

  697.             error                                  = v - intV * sum;

  698.         }

  699.     }

  700. 

  701.     (*filterPos)[dstW + 0] =

  702.     (*filterPos)[dstW + 1] =

  703.     (*filterPos)[dstW + 2] = (*filterPos)[dstW - 1]; /* the MMX/SSE scaler will

  704.                                                       * read over the end */

  705.     for (i = 0; i < *outFilterSize; i++) {

  706.         int k = (dstW - 1) * (*outFilterSize) + i;

  707.         (*outFilter)[k + 1 * (*outFilterSize)] =

  708.         (*outFilter)[k + 2 * (*outFilterSize)] =

  709.         (*outFilter)[k + 3 * (*outFilterSize)] = (*outFilter)[k];

  710.     }

  711. 

  712.     ret = 0;

  713. 

  714. fail:

  715.     if(ret < 0)

  716.         av_log(NULL, ret == RETCODE_USE_CASCADE ? AV_LOG_DEBUG : AV_LOG_ERROR, "sws: initFilter failed\n");

  717.     av_free(filter);

  718.     av_free(filter2);

  719.     return ret;

  720. }

  721. 

  722. static void fill_rgb2yuv_table(SwsContext *c, const int table[4], int dstRange)

  723. {

  724.     int64_t W, V, Z, Cy, Cu, Cv;

  725.     int64_t vr =  table[0];

  726.     int64_t ub =  table[1];

  727.     int64_t ug = -table[2];

  728.     int64_t vg = -table[3];

  729.     int64_t ONE = 65536;

  730.     int64_t cy = ONE;

  731.     uint8_t *p = (uint8_t*)c->input_rgb2yuv_table;

  732.     int i;

  733.     static const int8_t map[] = {

  734.     BY_IDX, GY_IDX, -1    , BY_IDX, BY_IDX, GY_IDX, -1    , BY_IDX,

  735.     RY_IDX, -1    , GY_IDX, RY_IDX, RY_IDX, -1    , GY_IDX, RY_IDX,

  736.     RY_IDX, GY_IDX, -1    , RY_IDX, RY_IDX, GY_IDX, -1    , RY_IDX,

  737.     BY_IDX, -1    , GY_IDX, BY_IDX, BY_IDX, -1    , GY_IDX, BY_IDX,

  738.     BU_IDX, GU_IDX, -1    , BU_IDX, BU_IDX, GU_IDX, -1    , BU_IDX,

  739.     RU_IDX, -1    , GU_IDX, RU_IDX, RU_IDX, -1    , GU_IDX, RU_IDX,

  740.     RU_IDX, GU_IDX, -1    , RU_IDX, RU_IDX, GU_IDX, -1    , RU_IDX,

  741.     BU_IDX, -1    , GU_IDX, BU_IDX, BU_IDX, -1    , GU_IDX, BU_IDX,

  742.     BV_IDX, GV_IDX, -1    , BV_IDX, BV_IDX, GV_IDX, -1    , BV_IDX,

  743.     RV_IDX, -1    , GV_IDX, RV_IDX, RV_IDX, -1    , GV_IDX, RV_IDX,

  744.     RV_IDX, GV_IDX, -1    , RV_IDX, RV_IDX, GV_IDX, -1    , RV_IDX,

  745.     BV_IDX, -1    , GV_IDX, BV_IDX, BV_IDX, -1    , GV_IDX, BV_IDX,

  746.     RY_IDX, BY_IDX, RY_IDX, BY_IDX, RY_IDX, BY_IDX, RY_IDX, BY_IDX,

  747.     BY_IDX, RY_IDX, BY_IDX, RY_IDX, BY_IDX, RY_IDX, BY_IDX, RY_IDX,

  748.     GY_IDX, -1    , GY_IDX, -1    , GY_IDX, -1    , GY_IDX, -1    ,

  749.     -1    , GY_IDX, -1    , GY_IDX, -1    , GY_IDX, -1    , GY_IDX,

  750.     RU_IDX, BU_IDX, RU_IDX, BU_IDX, RU_IDX, BU_IDX, RU_IDX, BU_IDX,

  751.     BU_IDX, RU_IDX, BU_IDX, RU_IDX, BU_IDX, RU_IDX, BU_IDX, RU_IDX,

  752.     GU_IDX, -1    , GU_IDX, -1    , GU_IDX, -1    , GU_IDX, -1    ,

  753.     -1    , GU_IDX, -1    , GU_IDX, -1    , GU_IDX, -1    , GU_IDX,

  754.     RV_IDX, BV_IDX, RV_IDX, BV_IDX, RV_IDX, BV_IDX, RV_IDX, BV_IDX,

  755.     BV_IDX, RV_IDX, BV_IDX, RV_IDX, BV_IDX, RV_IDX, BV_IDX, RV_IDX,

  756.     GV_IDX, -1    , GV_IDX, -1    , GV_IDX, -1    , GV_IDX, -1    ,

  757.     -1    , GV_IDX, -1    , GV_IDX, -1    , GV_IDX, -1    , GV_IDX, //23

  758.     -1    , -1    , -1    , -1    , -1    , -1    , -1    , -1    , //24

  759.     -1    , -1    , -1    , -1    , -1    , -1    , -1    , -1    , //25

  760.     -1    , -1    , -1    , -1    , -1    , -1    , -1    , -1    , //26

  761.     -1    , -1    , -1    , -1    , -1    , -1    , -1    , -1    , //27

  762.     -1    , -1    , -1    , -1    , -1    , -1    , -1    , -1    , //28

  763.     -1    , -1    , -1    , -1    , -1    , -1    , -1    , -1    , //29

  764.     -1    , -1    , -1    , -1    , -1    , -1    , -1    , -1    , //30

  765.     -1    , -1    , -1    , -1    , -1    , -1    , -1    , -1    , //31

  766.     BY_IDX, GY_IDX, RY_IDX, -1    , -1    , -1    , -1    , -1    , //32

  767.     BU_IDX, GU_IDX, RU_IDX, -1    , -1    , -1    , -1    , -1    , //33

  768.     BV_IDX, GV_IDX, RV_IDX, -1    , -1    , -1    , -1    , -1    , //34

  769.     };

  770. 

  771.     dstRange = 0; //FIXME range = 1 is handled elsewhere

  772. 

  773.     if (!dstRange) {

  774.         cy = cy * 255 / 219;

  775.     } else {

  776.         vr = vr * 224 / 255;

  777.         ub = ub * 224 / 255;

  778.         ug = ug * 224 / 255;

  779.         vg = vg * 224 / 255;

  780.     }

  781.     W = ROUNDED_DIV(ONE*ONE*ug, ub);

  782.     V = ROUNDED_DIV(ONE*ONE*vg, vr);

  783.     Z = ONE*ONE-W-V;

  784. 

  785.     Cy = ROUNDED_DIV(cy*Z, ONE);

  786.     Cu = ROUNDED_DIV(ub*Z, ONE);

  787.     Cv = ROUNDED_DIV(vr*Z, ONE);

  788. 

  789.     c->input_rgb2yuv_table[RY_IDX] = -ROUNDED_DIV((1 << RGB2YUV_SHIFT)*V        , Cy);

  790.     c->input_rgb2yuv_table[GY_IDX] =  ROUNDED_DIV((1 << RGB2YUV_SHIFT)*ONE*ONE  , Cy);

  791.     c->input_rgb2yuv_table[BY_IDX] = -ROUNDED_DIV((1 << RGB2YUV_SHIFT)*W        , Cy);

  792. 

  793.     c->input_rgb2yuv_table[RU_IDX] =  ROUNDED_DIV((1 << RGB2YUV_SHIFT)*V        , Cu);

  794.     c->input_rgb2yuv_table[GU_IDX] = -ROUNDED_DIV((1 << RGB2YUV_SHIFT)*ONE*ONE  , Cu);

  795.     c->input_rgb2yuv_table[BU_IDX] =  ROUNDED_DIV((1 << RGB2YUV_SHIFT)*(Z+W)    , Cu);

  796. 

  797.     c->input_rgb2yuv_table[RV_IDX] =  ROUNDED_DIV((1 << RGB2YUV_SHIFT)*(V+Z)    , Cv);

  798.     c->input_rgb2yuv_table[GV_IDX] = -ROUNDED_DIV((1 << RGB2YUV_SHIFT)*ONE*ONE  , Cv);

  799.     c->input_rgb2yuv_table[BV_IDX] =  ROUNDED_DIV((1 << RGB2YUV_SHIFT)*W        , Cv);

  800. 

  801.     if(/*!dstRange && */!memcmp(table, ff_yuv2rgb_coeffs[SWS_CS_DEFAULT], sizeof(ff_yuv2rgb_coeffs[SWS_CS_DEFAULT]))) {

  802.         c->input_rgb2yuv_table[BY_IDX] =  ((int)(0.114 * 219 / 255 * (1 << RGB2YUV_SHIFT) + 0.5));

  803.         c->input_rgb2yuv_table[BV_IDX] = (-(int)(0.081 * 224 / 255 * (1 << RGB2YUV_SHIFT) + 0.5));

  804.         c->input_rgb2yuv_table[BU_IDX] =  ((int)(0.500 * 224 / 255 * (1 << RGB2YUV_SHIFT) + 0.5));

  805.         c->input_rgb2yuv_table[GY_IDX] =  ((int)(0.587 * 219 / 255 * (1 << RGB2YUV_SHIFT) + 0.5));

  806.         c->input_rgb2yuv_table[GV_IDX] = (-(int)(0.419 * 224 / 255 * (1 << RGB2YUV_SHIFT) + 0.5));

  807.         c->input_rgb2yuv_table[GU_IDX] = (-(int)(0.331 * 224 / 255 * (1 << RGB2YUV_SHIFT) + 0.5));

  808.         c->input_rgb2yuv_table[RY_IDX] =  ((int)(0.299 * 219 / 255 * (1 << RGB2YUV_SHIFT) + 0.5));

  809.         c->input_rgb2yuv_table[RV_IDX] =  ((int)(0.500 * 224 / 255 * (1 << RGB2YUV_SHIFT) + 0.5));

  810.         c->input_rgb2yuv_table[RU_IDX] = (-(int)(0.169 * 224 / 255 * (1 << RGB2YUV_SHIFT) + 0.5));

  811.     }

  812.     for(i=0; i<FF_ARRAY_ELEMS(map); i++)

  813.         AV_WL16(p + 16*4 + 2*i, map[i] >= 0 ? c->input_rgb2yuv_table[map[i]] : 0);

  814. }

  815. 

  816. static void fill_xyztables(struct SwsContext *c)

  817. {

  818.     int i;

  819.     double xyzgamma = XYZ_GAMMA;

  820.     double rgbgamma = 1.0 / RGB_GAMMA;

  821.     double xyzgammainv = 1.0 / XYZ_GAMMA;

  822.     double rgbgammainv = RGB_GAMMA;

  823.     static const int16_t xyz2rgb_matrix[3][4] = {

  824.         {13270, -6295, -2041},

  825.         {-3969,  7682,   170},

  826.         {  228,  -835,  4329} };

  827.     static const int16_t rgb2xyz_matrix[3][4] = {

  828.         {1689, 1464,  739},

  829.         { 871, 2929,  296},

  830.         {  79,  488, 3891} };

  831.     static int16_t xyzgamma_tab[4096], rgbgamma_tab[4096], xyzgammainv_tab[4096], rgbgammainv_tab[4096];

  832. 

  833.     memcpy(c->xyz2rgb_matrix, xyz2rgb_matrix, sizeof(c->xyz2rgb_matrix));

  834.     memcpy(c->rgb2xyz_matrix, rgb2xyz_matrix, sizeof(c->rgb2xyz_matrix));

  835.     c->xyzgamma = xyzgamma_tab;

  836.     c->rgbgamma = rgbgamma_tab;

  837.     c->xyzgammainv = xyzgammainv_tab;

  838.     c->rgbgammainv = rgbgammainv_tab;

  839. 

  840.     if (rgbgamma_tab[4095])

  841.         return;

  842. 

  843.     /* set gamma vectors */

  844.     for (i = 0; i < 4096; i++) {

  845.         xyzgamma_tab[i] = lrint(pow(i / 4095.0, xyzgamma) * 4095.0);

  846.         rgbgamma_tab[i] = lrint(pow(i / 4095.0, rgbgamma) * 4095.0);

  847.         xyzgammainv_tab[i] = lrint(pow(i / 4095.0, xyzgammainv) * 4095.0);

  848.         rgbgammainv_tab[i] = lrint(pow(i / 4095.0, rgbgammainv) * 4095.0);

  849.     }

  850. }

  851. 

  852. int sws_setColorspaceDetails(struct SwsContext *c, const int inv_table[4],

  853.                              int srcRange, const int table[4], int dstRange,

  854.                              int brightness, int contrast, int saturation)

  855. {

  856.     const AVPixFmtDescriptor *desc_dst;

  857.     const AVPixFmtDescriptor *desc_src;

  858.     int need_reinit = 0;

  859. 

  860.     handle_formats(c);

  861.     desc_dst = av_pix_fmt_desc_get(c->dstFormat);

  862.     desc_src = av_pix_fmt_desc_get(c->srcFormat);

  863. 

  864.     if(!isYUV(c->dstFormat) && !isGray(c->dstFormat))

  865.         dstRange = 0;

  866.     if(!isYUV(c->srcFormat) && !isGray(c->srcFormat))

  867.         srcRange = 0;

  868. 

  869.     if (c->srcRange != srcRange ||

  870.         c->dstRange != dstRange ||

  871.         c->brightness != brightness ||

  872.         c->contrast   != contrast ||

  873.         c->saturation != saturation ||

  874.         memcmp(c->srcColorspaceTable, inv_table, sizeof(int) * 4) ||

  875.         memcmp(c->dstColorspaceTable,     table, sizeof(int) * 4)

  876.     )

  877.         need_reinit = 1;

  878. 

  879.     memmove(c->srcColorspaceTable, inv_table, sizeof(int) * 4);

  880.     memmove(c->dstColorspaceTable, table, sizeof(int) * 4);

  881. 

  882. 

  883. 

  884.     c->brightness = brightness;

  885.     c->contrast   = contrast;

  886.     c->saturation = saturation;

  887.     c->srcRange   = srcRange;

  888.     c->dstRange   = dstRange;

  889. 

  890.     //The srcBpc check is possibly wrong but we seem to lack a definitive reference to test this

  891.     //and what we have in ticket 2939 looks better with this check

  892.     if (need_reinit && (c->srcBpc == 8 || !isYUV(c->srcFormat)))

  893.         ff_sws_init_range_convert(c);

  894. 

  895.     c->dstFormatBpp = av_get_bits_per_pixel(desc_dst);

  896.     c->srcFormatBpp = av_get_bits_per_pixel(desc_src);

  897. 

  898.     if (c->cascaded_context[c->cascaded_mainindex])

  899.         return sws_setColorspaceDetails(c->cascaded_context[c->cascaded_mainindex],inv_table, srcRange,table, dstRange, brightness,  contrast, saturation);

  900. 

  901.     if (!need_reinit)

  902.         return 0;

  903. 

  904.     if ((isYUV(c->dstFormat) || isGray(c->dstFormat)) && (isYUV(c->srcFormat) || isGray(c->srcFormat))) {

  905.         if (!c->cascaded_context[0] &&

  906.             memcmp(c->dstColorspaceTable, c->srcColorspaceTable, sizeof(int) * 4) &&

  907.             c->srcW && c->srcH && c->dstW && c->dstH) {

  908.             enum AVPixelFormat tmp_format;

  909.             int tmp_width, tmp_height;

  910.             int srcW = c->srcW;

  911.             int srcH = c->srcH;

  912.             int dstW = c->dstW;

  913.             int dstH = c->dstH;

  914.             int ret;

  915.             av_log(c, AV_LOG_VERBOSE, "YUV color matrix differs for YUV->YUV, using intermediate RGB to convert\n");

  916. 

  917.             if (isNBPS(c->dstFormat) || is16BPS(c->dstFormat)) {

  918.                 if (isALPHA(c->srcFormat) && isALPHA(c->dstFormat)) {

  919.                     tmp_format = AV_PIX_FMT_BGRA64;

  920.                 } else {

  921.                     tmp_format = AV_PIX_FMT_BGR48;

  922.                 }

  923.             } else {

  924.                 if (isALPHA(c->srcFormat) && isALPHA(c->dstFormat)) {

  925.                     tmp_format = AV_PIX_FMT_BGRA;

  926.                 } else {

  927.                     tmp_format = AV_PIX_FMT_BGR24;

  928.                 }

  929.             }

  930. 

  931.             if (srcW*srcH > dstW*dstH) {

  932.                 tmp_width  = dstW;

  933.                 tmp_height = dstH;

  934.             } else {

  935.                 tmp_width  = srcW;

  936.                 tmp_height = srcH;

  937.             }

  938. 

  939.             ret = av_image_alloc(c->cascaded_tmp, c->cascaded_tmpStride,

  940.                                 tmp_width, tmp_height, tmp_format, 64);

  941.             if (ret < 0)

  942.                 return ret;

  943. 

  944.             c->cascaded_context[0] = sws_alloc_set_opts(srcW, srcH, c->srcFormat,

  945.                                                         tmp_width, tmp_height, tmp_format,

  946.                                                         c->flags, c->param);

  947.             if (!c->cascaded_context[0])

  948.                 return -1;

  949. 

  950.             c->cascaded_context[0]->alphablend = c->alphablend;

  951.             ret = sws_init_context(c->cascaded_context[0], NULL , NULL);

  952.             if (ret < 0)

  953.                 return ret;

  954.             //we set both src and dst depending on that the RGB side will be ignored

  955.             sws_setColorspaceDetails(c->cascaded_context[0], inv_table,

  956.                                      srcRange, table, dstRange,

  957.                                      brightness, contrast, saturation);

  958. 

  959.             c->cascaded_context[1] = sws_getContext(tmp_width, tmp_height, tmp_format,

  960.                                                     dstW, dstH, c->dstFormat,

  961.                                                     c->flags, NULL, NULL, c->param);

  962.             if (!c->cascaded_context[1])

  963.                 return -1;

  964.             sws_setColorspaceDetails(c->cascaded_context[1], inv_table,

  965.                                      srcRange, table, dstRange,

  966.                                      0, 1 << 16, 1 << 16);

  967.             return 0;

  968.         }

  969.         return -1;

  970.     }

  971. 

  972.     if (!isYUV(c->dstFormat) && !isGray(c->dstFormat)) {

  973.         ff_yuv2rgb_c_init_tables(c, inv_table, srcRange, brightness,

  974.                                  contrast, saturation);

  975.         // FIXME factorize

  976. 

  977.         if (ARCH_PPC)

  978.             ff_yuv2rgb_init_tables_ppc(c, inv_table, brightness,

  979.                                        contrast, saturation);

  980.     }

  981. 

  982.     fill_rgb2yuv_table(c, table, dstRange);

  983. 

  984.     return 0;

  985. }

  986. 

  987. int sws_getColorspaceDetails(struct SwsContext *c, int **inv_table,

  988.                              int *srcRange, int **table, int *dstRange,

  989.                              int *brightness, int *contrast, int *saturation)

  990. {

  991.     if (!c )

  992.         return -1;

  993. 

  994.     *inv_table  = c->srcColorspaceTable;

  995.     *table      = c->dstColorspaceTable;

  996.     *srcRange   = c->srcRange;

  997.     *dstRange   = c->dstRange;

  998.     *brightness = c->brightness;

  999.     *contrast   = c->contrast;

  1000.     *saturation = c->saturation;

  1001. 

  1002.     return 0;

  1003. }

  1004. 

  1005. static int handle_jpeg(enum AVPixelFormat *format)

  1006. {

  1007.     switch (*format) {

  1008.     case AV_PIX_FMT_YUVJ420P:

  1009.         *format = AV_PIX_FMT_YUV420P;

  1010.         return 1;

  1011.     case AV_PIX_FMT_YUVJ411P:

  1012.         *format = AV_PIX_FMT_YUV411P;

  1013.         return 1;

  1014.     case AV_PIX_FMT_YUVJ422P:

  1015.         *format = AV_PIX_FMT_YUV422P;

  1016.         return 1;

  1017.     case AV_PIX_FMT_YUVJ444P:

  1018.         *format = AV_PIX_FMT_YUV444P;

  1019.         return 1;

  1020.     case AV_PIX_FMT_YUVJ440P:

  1021.         *format = AV_PIX_FMT_YUV440P;

  1022.         return 1;

  1023.     case AV_PIX_FMT_GRAY8:

  1024.     case AV_PIX_FMT_YA8:

  1025.     case AV_PIX_FMT_GRAY9LE:

  1026.     case AV_PIX_FMT_GRAY9BE:

  1027.     case AV_PIX_FMT_GRAY10LE:

  1028.     case AV_PIX_FMT_GRAY10BE:

  1029.     case AV_PIX_FMT_GRAY12LE:

  1030.     case AV_PIX_FMT_GRAY12BE:

  1031.     case AV_PIX_FMT_GRAY14LE:

  1032.     case AV_PIX_FMT_GRAY14BE:

  1033.     case AV_PIX_FMT_GRAY16LE:

  1034.     case AV_PIX_FMT_GRAY16BE:

  1035.     case AV_PIX_FMT_YA16BE:

  1036.     case AV_PIX_FMT_YA16LE:

  1037.         return 1;

  1038.     default:

  1039.         return 0;

  1040.     }

  1041. }

  1042. 

  1043. static int handle_0alpha(enum AVPixelFormat *format)

  1044. {

  1045.     switch (*format) {

  1046.     case AV_PIX_FMT_0BGR    : *format = AV_PIX_FMT_ABGR   ; return 1;

  1047.     case AV_PIX_FMT_BGR0    : *format = AV_PIX_FMT_BGRA   ; return 4;

  1048.     case AV_PIX_FMT_0RGB    : *format = AV_PIX_FMT_ARGB   ; return 1;

  1049.     case AV_PIX_FMT_RGB0    : *format = AV_PIX_FMT_RGBA   ; return 4;

  1050.     default:                                          return 0;

  1051.     }

  1052. }

  1053. 

  1054. static int handle_xyz(enum AVPixelFormat *format)

  1055. {

  1056.     switch (*format) {

  1057.     case AV_PIX_FMT_XYZ12BE : *format = AV_PIX_FMT_RGB48BE; return 1;

  1058.     case AV_PIX_FMT_XYZ12LE : *format = AV_PIX_FMT_RGB48LE; return 1;

  1059.     default:                                                return 0;

  1060.     }

  1061. }

  1062. 

  1063. static void handle_formats(SwsContext *c)

  1064. {

  1065.     c->src0Alpha |= handle_0alpha(&c->srcFormat);

  1066.     c->dst0Alpha |= handle_0alpha(&c->dstFormat);

  1067.     c->srcXYZ    |= handle_xyz(&c->srcFormat);

  1068.     c->dstXYZ    |= handle_xyz(&c->dstFormat);

  1069.     if (c->srcXYZ || c->dstXYZ)

  1070.         fill_xyztables(c);

  1071. }

  1072. 

  1073. SwsContext *sws_alloc_context(void)

  1074. {

  1075.     SwsContext *c = av_mallocz(sizeof(SwsContext));

  1076. 

  1077.     av_assert0(offsetof(SwsContext, redDither) + DITHER32_INT == offsetof(SwsContext, dither32));

  1078. 

  1079.     if (c) {

  1080.         c->av_class = &ff_sws_context_class;

  1081.         av_opt_set_defaults(c);

  1082.     }

  1083. 

  1084.     return c;

  1085. }

  1086. 

  1087. static uint16_t * alloc_gamma_tbl(double e)

  1088. {

  1089.     int i = 0;

  1090.     uint16_t * tbl;

  1091.     tbl = (uint16_t*)av_malloc(sizeof(uint16_t) * 1 << 16);

  1092.     if (!tbl)

  1093.         return NULL;

  1094. 

  1095.     for (i = 0; i < 65536; ++i) {

  1096.         tbl[i] = pow(i / 65535.0, e) * 65535.0;

  1097.     }

  1098.     return tbl;

  1099. }

  1100. 

  1101. static enum AVPixelFormat alphaless_fmt(enum AVPixelFormat fmt)

  1102. {

  1103.     switch(fmt) {

  1104.     case AV_PIX_FMT_ARGB:       return AV_PIX_FMT_RGB24;

  1105.     case AV_PIX_FMT_RGBA:       return AV_PIX_FMT_RGB24;

  1106.     case AV_PIX_FMT_ABGR:       return AV_PIX_FMT_BGR24;

  1107.     case AV_PIX_FMT_BGRA:       return AV_PIX_FMT_BGR24;

  1108.     case AV_PIX_FMT_YA8:        return AV_PIX_FMT_GRAY8;

  1109. 

  1110.     case AV_PIX_FMT_YUVA420P:   return AV_PIX_FMT_YUV420P;

  1111.     case AV_PIX_FMT_YUVA422P:   return AV_PIX_FMT_YUV422P;

  1112.     case AV_PIX_FMT_YUVA444P:           return AV_PIX_FMT_YUV444P;

  1113. 

  1114.     case AV_PIX_FMT_GBRAP:              return AV_PIX_FMT_GBRP;

  1115. 

  1116.     case AV_PIX_FMT_GBRAP10LE:          return AV_PIX_FMT_GBRP10;

  1117.     case AV_PIX_FMT_GBRAP10BE:          return AV_PIX_FMT_GBRP10;

  1118. 

  1119.     case AV_PIX_FMT_GBRAP12LE:          return AV_PIX_FMT_GBRP12;

  1120.     case AV_PIX_FMT_GBRAP12BE:          return AV_PIX_FMT_GBRP12;

  1121. 

  1122.     case AV_PIX_FMT_GBRAP16LE:          return AV_PIX_FMT_GBRP16;

  1123.     case AV_PIX_FMT_GBRAP16BE:          return AV_PIX_FMT_GBRP16;

  1124. 

  1125.     case AV_PIX_FMT_RGBA64LE:   return AV_PIX_FMT_RGB48;

  1126.     case AV_PIX_FMT_RGBA64BE:   return AV_PIX_FMT_RGB48;

  1127.     case AV_PIX_FMT_BGRA64LE:   return AV_PIX_FMT_BGR48;

  1128.     case AV_PIX_FMT_BGRA64BE:   return AV_PIX_FMT_BGR48;

  1129. 

  1130.     case AV_PIX_FMT_YA16BE:             return AV_PIX_FMT_GRAY16;

  1131.     case AV_PIX_FMT_YA16LE:             return AV_PIX_FMT_GRAY16;

  1132. 

  1133.     case AV_PIX_FMT_YUVA420P9BE:        return AV_PIX_FMT_YUV420P9;

  1134.     case AV_PIX_FMT_YUVA422P9BE:        return AV_PIX_FMT_YUV422P9;

  1135.     case AV_PIX_FMT_YUVA444P9BE:        return AV_PIX_FMT_YUV444P9;

  1136.     case AV_PIX_FMT_YUVA420P9LE:        return AV_PIX_FMT_YUV420P9;

  1137.     case AV_PIX_FMT_YUVA422P9LE:        return AV_PIX_FMT_YUV422P9;

  1138.     case AV_PIX_FMT_YUVA444P9LE:        return AV_PIX_FMT_YUV444P9;

  1139.     case AV_PIX_FMT_YUVA420P10BE:       return AV_PIX_FMT_YUV420P10;

  1140.     case AV_PIX_FMT_YUVA422P10BE:       return AV_PIX_FMT_YUV422P10;

  1141.     case AV_PIX_FMT_YUVA444P10BE:       return AV_PIX_FMT_YUV444P10;

  1142.     case AV_PIX_FMT_YUVA420P10LE:       return AV_PIX_FMT_YUV420P10;

  1143.     case AV_PIX_FMT_YUVA422P10LE:       return AV_PIX_FMT_YUV422P10;

  1144.     case AV_PIX_FMT_YUVA444P10LE:       return AV_PIX_FMT_YUV444P10;

  1145.     case AV_PIX_FMT_YUVA420P16BE:       return AV_PIX_FMT_YUV420P16;

  1146.     case AV_PIX_FMT_YUVA422P16BE:       return AV_PIX_FMT_YUV422P16;

  1147.     case AV_PIX_FMT_YUVA444P16BE:       return AV_PIX_FMT_YUV444P16;

  1148.     case AV_PIX_FMT_YUVA420P16LE:       return AV_PIX_FMT_YUV420P16;

  1149.     case AV_PIX_FMT_YUVA422P16LE:       return AV_PIX_FMT_YUV422P16;

  1150.     case AV_PIX_FMT_YUVA444P16LE:       return AV_PIX_FMT_YUV444P16;

  1151. 

  1152. //     case AV_PIX_FMT_AYUV64LE:

  1153. //     case AV_PIX_FMT_AYUV64BE:

  1154. //     case AV_PIX_FMT_PAL8:

  1155.     default: return AV_PIX_FMT_NONE;

  1156.     }

  1157. }

  1158. 

  1159. av_cold int sws_init_context(SwsContext *c, SwsFilter *srcFilter,

  1160.                              SwsFilter *dstFilter)

  1161. {

  1162.     int i;

  1163.     int usesVFilter, usesHFilter;

  1164.     int unscaled;

  1165.     SwsFilter dummyFilter = { NULL, NULL, NULL, NULL };

  1166.     int srcW              = c->srcW;

  1167.     int srcH              = c->srcH;

  1168.     int dstW              = c->dstW;

  1169.     int dstH              = c->dstH;

  1170.     int dst_stride        = FFALIGN(dstW * sizeof(int16_t) + 66, 16);

  1171.     int flags, cpu_flags;

  1172.     enum AVPixelFormat srcFormat = c->srcFormat;

  1173.     enum AVPixelFormat dstFormat = c->dstFormat;

  1174.     const AVPixFmtDescriptor *desc_src;

  1175.     const AVPixFmtDescriptor *desc_dst;

  1176.     int ret = 0;

  1177.     enum AVPixelFormat tmpFmt;

  1178.     static const float float_mult = 1.0f / 255.0f;

  1179. 

  1180.     cpu_flags = av_get_cpu_flags();

  1181.     flags     = c->flags;

  1182.     emms_c();

  1183.     if (!rgb15to16)

  1184.         ff_sws_rgb2rgb_init();

  1185. 

  1186.     unscaled = (srcW == dstW && srcH == dstH);

  1187. 

  1188.     c->srcRange |= handle_jpeg(&c->srcFormat);

  1189.     c->dstRange |= handle_jpeg(&c->dstFormat);

  1190. 

  1191.     if(srcFormat!=c->srcFormat || dstFormat!=c->dstFormat)

  1192.         av_log(c, AV_LOG_WARNING, "deprecated pixel format used, make sure you did set range correctly\n");

  1193. 

  1194.     if (!c->contrast && !c->saturation && !c->dstFormatBpp)

  1195.         sws_setColorspaceDetails(c, ff_yuv2rgb_coeffs[SWS_CS_DEFAULT], c->srcRange,

  1196.                                  ff_yuv2rgb_coeffs[SWS_CS_DEFAULT],

  1197.                                  c->dstRange, 0, 1 << 16, 1 << 16);

  1198. 

  1199.     handle_formats(c);

  1200.     srcFormat = c->srcFormat;

  1201.     dstFormat = c->dstFormat;

  1202.     desc_src = av_pix_fmt_desc_get(srcFormat);

  1203.     desc_dst = av_pix_fmt_desc_get(dstFormat);

  1204. 

  1205.     // If the source has no alpha then disable alpha blendaway

  1206.     if (c->src0Alpha)

  1207.         c->alphablend = SWS_ALPHA_BLEND_NONE;

  1208. 

  1209.     if (!(unscaled && sws_isSupportedEndiannessConversion(srcFormat) &&

  1210.           av_pix_fmt_swap_endianness(srcFormat) == dstFormat)) {

  1211.     if (!sws_isSupportedInput(srcFormat)) {

  1212.         av_log(c, AV_LOG_ERROR, "%s is not supported as input pixel format\n",

  1213.                av_get_pix_fmt_name(srcFormat));

  1214.         return AVERROR(EINVAL);

  1215.     }

  1216.     if (!sws_isSupportedOutput(dstFormat)) {

  1217.         av_log(c, AV_LOG_ERROR, "%s is not supported as output pixel format\n",

  1218.                av_get_pix_fmt_name(dstFormat));

  1219.         return AVERROR(EINVAL);

  1220.     }

  1221.     }

  1222.     av_assert2(desc_src && desc_dst);

  1223. 

  1224.     i = flags & (SWS_POINT         |

  1225.                  SWS_AREA          |

  1226.                  SWS_BILINEAR      |

  1227.                  SWS_FAST_BILINEAR |

  1228.                  SWS_BICUBIC       |

  1229.                  SWS_X             |

  1230.                  SWS_GAUSS         |

  1231.                  SWS_LANCZOS       |

  1232.                  SWS_SINC          |

  1233.                  SWS_SPLINE        |

  1234.                  SWS_BICUBLIN);

  1235. 

  1236.     /* provide a default scaler if not set by caller */

  1237.     if (!i) {

  1238.         if (dstW < srcW && dstH < srcH)

  1239.             flags |= SWS_BICUBIC;

  1240.         else if (dstW > srcW && dstH > srcH)

  1241.             flags |= SWS_BICUBIC;

  1242.         else

  1243.             flags |= SWS_BICUBIC;

  1244.         c->flags = flags;

  1245.     } else if (i & (i - 1)) {

  1246.         av_log(c, AV_LOG_ERROR,

  1247.                "Exactly one scaler algorithm must be chosen, got %X\n", i);

  1248.         return AVERROR(EINVAL);

  1249.     }

  1250.     /* sanity check */

  1251.     if (srcW < 1 || srcH < 1 || dstW < 1 || dstH < 1) {

  1252.         /* FIXME check if these are enough and try to lower them after

  1253.          * fixing the relevant parts of the code */

  1254.         av_log(c, AV_LOG_ERROR, "%dx%d -> %dx%d is invalid scaling dimension\n",

  1255.                srcW, srcH, dstW, dstH);

  1256.         return AVERROR(EINVAL);

  1257.     }

  1258.     if (flags & SWS_FAST_BILINEAR) {

  1259.         if (srcW < 8 || dstW < 8) {

  1260.             flags ^= SWS_FAST_BILINEAR | SWS_BILINEAR;

  1261.             c->flags = flags;

  1262.         }

  1263.     }

  1264. 

  1265.     if (!dstFilter)

  1266.         dstFilter = &dummyFilter;

  1267.     if (!srcFilter)

  1268.         srcFilter = &dummyFilter;

  1269. 

  1270.     c->lumXInc      = (((int64_t)srcW << 16) + (dstW >> 1)) / dstW;

  1271.     c->lumYInc      = (((int64_t)srcH << 16) + (dstH >> 1)) / dstH;

  1272.     c->dstFormatBpp = av_get_bits_per_pixel(desc_dst);

  1273.     c->srcFormatBpp = av_get_bits_per_pixel(desc_src);

  1274.     c->vRounder     = 4 * 0x0001000100010001ULL;

  1275. 

  1276.     usesVFilter = (srcFilter->lumV && srcFilter->lumV->length > 1) ||

  1277.                   (srcFilter->chrV && srcFilter->chrV->length > 1) ||

  1278.                   (dstFilter->lumV && dstFilter->lumV->length > 1) ||

  1279.                   (dstFilter->chrV && dstFilter->chrV->length > 1);

  1280.     usesHFilter = (srcFilter->lumH && srcFilter->lumH->length > 1) ||

  1281.                   (srcFilter->chrH && srcFilter->chrH->length > 1) ||

  1282.                   (dstFilter->lumH && dstFilter->lumH->length > 1) ||

  1283.                   (dstFilter->chrH && dstFilter->chrH->length > 1);

  1284. 

  1285.     av_pix_fmt_get_chroma_sub_sample(srcFormat, &c->chrSrcHSubSample, &c->chrSrcVSubSample);

  1286.     av_pix_fmt_get_chroma_sub_sample(dstFormat, &c->chrDstHSubSample, &c->chrDstVSubSample);

  1287. 

  1288.     if (isAnyRGB(dstFormat) && !(flags&SWS_FULL_CHR_H_INT)) {

  1289.         if (dstW&1) {

  1290.             av_log(c, AV_LOG_DEBUG, "Forcing full internal H chroma due to odd output size\n");

  1291.             flags |= SWS_FULL_CHR_H_INT;

  1292.             c->flags = flags;

  1293.         }

  1294. 

  1295.         if (   c->chrSrcHSubSample == 0

  1296.             && c->chrSrcVSubSample == 0

  1297.             && c->dither != SWS_DITHER_BAYER //SWS_FULL_CHR_H_INT is currently not supported with SWS_DITHER_BAYER

  1298.             && !(c->flags & SWS_FAST_BILINEAR)

  1299.         ) {

  1300.             av_log(c, AV_LOG_DEBUG, "Forcing full internal H chroma due to input having non subsampled chroma\n");

  1301.             flags |= SWS_FULL_CHR_H_INT;

  1302.             c->flags = flags;

  1303.         }

  1304.     }

  1305. 

  1306.     if (c->dither == SWS_DITHER_AUTO) {

  1307.         if (flags & SWS_ERROR_DIFFUSION)

  1308.             c->dither = SWS_DITHER_ED;

  1309.     }

  1310. 

  1311.     if(dstFormat == AV_PIX_FMT_BGR4_BYTE ||

  1312.        dstFormat == AV_PIX_FMT_RGB4_BYTE ||

  1313.        dstFormat == AV_PIX_FMT_BGR8 ||

  1314.        dstFormat == AV_PIX_FMT_RGB8) {

  1315.         if (c->dither == SWS_DITHER_AUTO)

  1316.             c->dither = (flags & SWS_FULL_CHR_H_INT) ? SWS_DITHER_ED : SWS_DITHER_BAYER;

  1317.         if (!(flags & SWS_FULL_CHR_H_INT)) {

  1318.             if (c->dither == SWS_DITHER_ED || c->dither == SWS_DITHER_A_DITHER || c->dither == SWS_DITHER_X_DITHER) {

  1319.                 av_log(c, AV_LOG_DEBUG,

  1320.                     "Desired dithering only supported in full chroma interpolation for destination format '%s'\n",

  1321.                     av_get_pix_fmt_name(dstFormat));

  1322.                 flags   |= SWS_FULL_CHR_H_INT;

  1323.                 c->flags = flags;

  1324.             }

  1325.         }

  1326.         if (flags & SWS_FULL_CHR_H_INT) {

  1327.             if (c->dither == SWS_DITHER_BAYER) {

  1328.                 av_log(c, AV_LOG_DEBUG,

  1329.                     "Ordered dither is not supported in full chroma interpolation for destination format '%s'\n",

  1330.                     av_get_pix_fmt_name(dstFormat));

  1331.                 c->dither = SWS_DITHER_ED;

  1332.             }

  1333.         }

  1334.     }

  1335.     if (isPlanarRGB(dstFormat)) {

  1336.         if (!(flags & SWS_FULL_CHR_H_INT)) {

  1337.             av_log(c, AV_LOG_DEBUG,

  1338.                    "%s output is not supported with half chroma resolution, switching to full\n",

  1339.                    av_get_pix_fmt_name(dstFormat));

  1340.             flags   |= SWS_FULL_CHR_H_INT;

  1341.             c->flags = flags;

  1342.         }

  1343.     }

  1344. 

  1345.     /* reuse chroma for 2 pixels RGB/BGR unless user wants full

  1346.      * chroma interpolation */

  1347.     if (flags & SWS_FULL_CHR_H_INT &&

  1348.         isAnyRGB(dstFormat)        &&

  1349.         !isPlanarRGB(dstFormat)    &&

  1350.         dstFormat != AV_PIX_FMT_RGBA64LE &&

  1351.         dstFormat != AV_PIX_FMT_RGBA64BE &&

  1352.         dstFormat != AV_PIX_FMT_BGRA64LE &&

  1353.         dstFormat != AV_PIX_FMT_BGRA64BE &&

  1354.         dstFormat != AV_PIX_FMT_RGB48LE &&

  1355.         dstFormat != AV_PIX_FMT_RGB48BE &&

  1356.         dstFormat != AV_PIX_FMT_BGR48LE &&

  1357.         dstFormat != AV_PIX_FMT_BGR48BE &&

  1358.         dstFormat != AV_PIX_FMT_RGBA  &&

  1359.         dstFormat != AV_PIX_FMT_ARGB  &&

  1360.         dstFormat != AV_PIX_FMT_BGRA  &&

  1361.         dstFormat != AV_PIX_FMT_ABGR  &&

  1362.         dstFormat != AV_PIX_FMT_RGB24 &&

  1363.         dstFormat != AV_PIX_FMT_BGR24 &&

  1364.         dstFormat != AV_PIX_FMT_BGR4_BYTE &&

  1365.         dstFormat != AV_PIX_FMT_RGB4_BYTE &&

  1366.         dstFormat != AV_PIX_FMT_BGR8 &&

  1367.         dstFormat != AV_PIX_FMT_RGB8

  1368.     ) {

  1369.         av_log(c, AV_LOG_WARNING,

  1370.                "full chroma interpolation for destination format '%s' not yet implemented\n",

  1371.                av_get_pix_fmt_name(dstFormat));

  1372.         flags   &= ~SWS_FULL_CHR_H_INT;

  1373.         c->flags = flags;

  1374.     }

  1375.     if (isAnyRGB(dstFormat) && !(flags & SWS_FULL_CHR_H_INT))

  1376.         c->chrDstHSubSample = 1;

  1377. 

  1378.     // drop some chroma lines if the user wants it

  1379.     c->vChrDrop          = (flags & SWS_SRC_V_CHR_DROP_MASK) >>

  1380.                            SWS_SRC_V_CHR_DROP_SHIFT;

  1381.     c->chrSrcVSubSample += c->vChrDrop;

  1382. 

  1383.     /* drop every other pixel for chroma calculation unless user

  1384.      * wants full chroma */

  1385.     if (isAnyRGB(srcFormat) && !(flags & SWS_FULL_CHR_H_INP)   &&

  1386.         srcFormat != AV_PIX_FMT_RGB8 && srcFormat != AV_PIX_FMT_BGR8 &&

  1387.         srcFormat != AV_PIX_FMT_RGB4 && srcFormat != AV_PIX_FMT_BGR4 &&

  1388.         srcFormat != AV_PIX_FMT_RGB4_BYTE && srcFormat != AV_PIX_FMT_BGR4_BYTE &&

  1389.         srcFormat != AV_PIX_FMT_GBRP9BE   && srcFormat != AV_PIX_FMT_GBRP9LE  &&

  1390.         srcFormat != AV_PIX_FMT_GBRP10BE  && srcFormat != AV_PIX_FMT_GBRP10LE &&

  1391.         srcFormat != AV_PIX_FMT_GBRAP10BE && srcFormat != AV_PIX_FMT_GBRAP10LE &&

  1392.         srcFormat != AV_PIX_FMT_GBRP12BE  && srcFormat != AV_PIX_FMT_GBRP12LE &&

  1393.         srcFormat != AV_PIX_FMT_GBRAP12BE && srcFormat != AV_PIX_FMT_GBRAP12LE &&

  1394.         srcFormat != AV_PIX_FMT_GBRP14BE  && srcFormat != AV_PIX_FMT_GBRP14LE &&

  1395.         srcFormat != AV_PIX_FMT_GBRP16BE  && srcFormat != AV_PIX_FMT_GBRP16LE &&

  1396.         srcFormat != AV_PIX_FMT_GBRAP16BE  && srcFormat != AV_PIX_FMT_GBRAP16LE &&

  1397.         ((dstW >> c->chrDstHSubSample) <= (srcW >> 1) ||

  1398.          (flags & SWS_FAST_BILINEAR)))

  1399.         c->chrSrcHSubSample = 1;

  1400. 

  1401.     // Note the AV_CEIL_RSHIFT is so that we always round toward +inf.

  1402.     c->chrSrcW = AV_CEIL_RSHIFT(srcW, c->chrSrcHSubSample);

  1403.     c->chrSrcH = AV_CEIL_RSHIFT(srcH, c->chrSrcVSubSample);

  1404.     c->chrDstW = AV_CEIL_RSHIFT(dstW, c->chrDstHSubSample);

  1405.     c->chrDstH = AV_CEIL_RSHIFT(dstH, c->chrDstVSubSample);

  1406. 

  1407.     FF_ALLOCZ_OR_GOTO(c, c->formatConvBuffer, FFALIGN(srcW*2+78, 16) * 2, fail);

  1408. 

  1409.     c->srcBpc = desc_src->comp[0].depth;

  1410.     if (c->srcBpc < 8)

  1411.         c->srcBpc = 8;

  1412.     c->dstBpc = desc_dst->comp[0].depth;

  1413.     if (c->dstBpc < 8)

  1414.         c->dstBpc = 8;

  1415.     if (isAnyRGB(srcFormat) || srcFormat == AV_PIX_FMT_PAL8)

  1416.         c->srcBpc = 16;

  1417.     if (c->dstBpc == 16)

  1418.         dst_stride <<= 1;

  1419. 

  1420.     if (INLINE_MMXEXT(cpu_flags) && c->srcBpc == 8 && c->dstBpc <= 14) {

  1421.         c->canMMXEXTBeUsed = dstW >= srcW && (dstW & 31) == 0 &&

  1422.                              c->chrDstW >= c->chrSrcW &&

  1423.                              (srcW & 15) == 0;

  1424.         if (!c->canMMXEXTBeUsed && dstW >= srcW && c->chrDstW >= c->chrSrcW && (srcW & 15) == 0

  1425. 

  1426.             && (flags & SWS_FAST_BILINEAR)) {

  1427.             if (flags & SWS_PRINT_INFO)

  1428.                 av_log(c, AV_LOG_INFO,

  1429.                        "output width is not a multiple of 32 -> no MMXEXT scaler\n");

  1430.         }

  1431.         if (usesHFilter || isNBPS(c->srcFormat) || is16BPS(c->srcFormat) || isAnyRGB(c->srcFormat))

  1432.             c->canMMXEXTBeUsed = 0;

  1433.     } else

  1434.         c->canMMXEXTBeUsed = 0;

  1435. 

  1436.     c->chrXInc = (((int64_t)c->chrSrcW << 16) + (c->chrDstW >> 1)) / c->chrDstW;

  1437.     c->chrYInc = (((int64_t)c->chrSrcH << 16) + (c->chrDstH >> 1)) / c->chrDstH;

  1438. 

  1439.     /* Match pixel 0 of the src to pixel 0 of dst and match pixel n-2 of src

  1440.      * to pixel n-2 of dst, but only for the FAST_BILINEAR mode otherwise do

  1441.      * correct scaling.

  1442.      * n-2 is the last chrominance sample available.

  1443.      * This is not perfect, but no one should notice the difference, the more

  1444.      * correct variant would be like the vertical one, but that would require

  1445.      * some special code for the first and last pixel */

  1446.     if (flags & SWS_FAST_BILINEAR) {

  1447.         if (c->canMMXEXTBeUsed) {

  1448.             c->lumXInc += 20;

  1449.             c->chrXInc += 20;

  1450.         }

  1451.         // we don't use the x86 asm scaler if MMX is available

  1452.         else if (INLINE_MMX(cpu_flags) && c->dstBpc <= 14) {

  1453.             c->lumXInc = ((int64_t)(srcW       - 2) << 16) / (dstW       - 2) - 20;

  1454.             c->chrXInc = ((int64_t)(c->chrSrcW - 2) << 16) / (c->chrDstW - 2) - 20;

  1455.         }

  1456.     }

  1457. 

  1458.     // hardcoded for now

  1459.     c->gamma_value = 2.2;

  1460.     tmpFmt = AV_PIX_FMT_RGBA64LE;

  1461. 

  1462. 

  1463.     if (!unscaled && c->gamma_flag && (srcFormat != tmpFmt || dstFormat != tmpFmt)) {

  1464.         SwsContext *c2;

  1465.         c->cascaded_context[0] = NULL;

  1466. 

  1467.         ret = av_image_alloc(c->cascaded_tmp, c->cascaded_tmpStride,

  1468.                             srcW, srcH, tmpFmt, 64);

  1469.         if (ret < 0)

  1470.             return ret;

  1471. 

  1472.         c->cascaded_context[0] = sws_getContext(srcW, srcH, srcFormat,

  1473.                                                 srcW, srcH, tmpFmt,

  1474.                                                 flags, NULL, NULL, c->param);

  1475.         if (!c->cascaded_context[0]) {

  1476.             return -1;

  1477.         }

  1478. 

  1479.         c->cascaded_context[1] = sws_getContext(srcW, srcH, tmpFmt,

  1480.                                                 dstW, dstH, tmpFmt,

  1481.                                                 flags, srcFilter, dstFilter, c->param);

  1482. 

  1483.         if (!c->cascaded_context[1])

  1484.             return -1;

  1485. 

  1486.         c2 = c->cascaded_context[1];

  1487.         c2->is_internal_gamma = 1;

  1488.         c2->gamma     = alloc_gamma_tbl(    c->gamma_value);

  1489.         c2->inv_gamma = alloc_gamma_tbl(1.f/c->gamma_value);

  1490.         if (!c2->gamma || !c2->inv_gamma)

  1491.             return AVERROR(ENOMEM);

  1492. 

  1493.         // is_internal_flag is set after creating the context

  1494.         // to properly create the gamma convert FilterDescriptor

  1495.         // we have to re-initialize it

  1496.         ff_free_filters(c2);

  1497.         if (ff_init_filters(c2) < 0) {

  1498.             sws_freeContext(c2);

  1499.             return -1;

  1500.         }

  1501. 

  1502.         c->cascaded_context[2] = NULL;

  1503.         if (dstFormat != tmpFmt) {

  1504.             ret = av_image_alloc(c->cascaded1_tmp, c->cascaded1_tmpStride,

  1505.                                 dstW, dstH, tmpFmt, 64);

  1506.             if (ret < 0)

  1507.                 return ret;

  1508. 

  1509.             c->cascaded_context[2] = sws_getContext(dstW, dstH, tmpFmt,

  1510.                                                 dstW, dstH, dstFormat,

  1511.                                                 flags, NULL, NULL, c->param);

  1512.             if (!c->cascaded_context[2])

  1513.                 return -1;

  1514.         }

  1515.         return 0;

  1516.     }

  1517. 

  1518.     if (isBayer(srcFormat)) {

  1519.         if (!unscaled ||

  1520.             (dstFormat != AV_PIX_FMT_RGB24 && dstFormat != AV_PIX_FMT_YUV420P)) {

  1521.             enum AVPixelFormat tmpFormat = AV_PIX_FMT_RGB24;

  1522. 

  1523.             ret = av_image_alloc(c->cascaded_tmp, c->cascaded_tmpStride,

  1524.                                 srcW, srcH, tmpFormat, 64);

  1525.             if (ret < 0)

  1526.                 return ret;

  1527. 

  1528.             c->cascaded_context[0] = sws_getContext(srcW, srcH, srcFormat,

  1529.                                                     srcW, srcH, tmpFormat,

  1530.                                                     flags, srcFilter, NULL, c->param);

  1531.             if (!c->cascaded_context[0])

  1532.                 return -1;

  1533. 

  1534.             c->cascaded_context[1] = sws_getContext(srcW, srcH, tmpFormat,

  1535.                                                     dstW, dstH, dstFormat,

  1536.                                                     flags, NULL, dstFilter, c->param);

  1537.             if (!c->cascaded_context[1])

  1538.                 return -1;

  1539.             return 0;

  1540.         }

  1541.     }

  1542. 

  1543.     if (unscaled && c->srcBpc == 8 && dstFormat == AV_PIX_FMT_GRAYF32){

  1544.         for (i = 0; i < 256; ++i){

  1545.             c->uint2float_lut[i] = (float)i * float_mult;

  1546.         }

  1547.     }

  1548. 

  1549.     // float will be converted to uint16_t

  1550.     if ((srcFormat == AV_PIX_FMT_GRAYF32BE || srcFormat == AV_PIX_FMT_GRAYF32LE) &&

  1551.         (!unscaled || unscaled && dstFormat != srcFormat && (srcFormat != AV_PIX_FMT_GRAYF32 ||

  1552.         dstFormat != AV_PIX_FMT_GRAY8))){

  1553.         c->srcBpc = 16;

  1554.     }

  1555. 

  1556.     if (CONFIG_SWSCALE_ALPHA && isALPHA(srcFormat) && !isALPHA(dstFormat)) {

  1557.         enum AVPixelFormat tmpFormat = alphaless_fmt(srcFormat);

  1558. 

  1559.         if (tmpFormat != AV_PIX_FMT_NONE && c->alphablend != SWS_ALPHA_BLEND_NONE)

  1560.         if (!unscaled ||

  1561.             dstFormat != tmpFormat ||

  1562.             usesHFilter || usesVFilter ||

  1563.             c->srcRange != c->dstRange

  1564.         ) {

  1565.             c->cascaded_mainindex = 1;

  1566.             ret = av_image_alloc(c->cascaded_tmp, c->cascaded_tmpStride,

  1567.                                 srcW, srcH, tmpFormat, 64);

  1568.             if (ret < 0)

  1569.                 return ret;

  1570. 

  1571.             c->cascaded_context[0] = sws_alloc_set_opts(srcW, srcH, srcFormat,

  1572.                                                         srcW, srcH, tmpFormat,

  1573.                                                         flags, c->param);

  1574.             if (!c->cascaded_context[0])

  1575.                 return -1;

  1576.             c->cascaded_context[0]->alphablend = c->alphablend;

  1577.             ret = sws_init_context(c->cascaded_context[0], NULL , NULL);

  1578.             if (ret < 0)

  1579.                 return ret;

  1580. 

  1581.             c->cascaded_context[1] = sws_alloc_set_opts(srcW, srcH, tmpFormat,

  1582.                                                         dstW, dstH, dstFormat,

  1583.                                                         flags, c->param);

  1584.             if (!c->cascaded_context[1])

  1585.                 return -1;

  1586. 

  1587.             c->cascaded_context[1]->srcRange = c->srcRange;

  1588.             c->cascaded_context[1]->dstRange = c->dstRange;

  1589.             ret = sws_init_context(c->cascaded_context[1], srcFilter , dstFilter);

  1590.             if (ret < 0)

  1591.                 return ret;

  1592. 

  1593.             return 0;

  1594.         }

  1595.     }

  1596. 

  1597. #if HAVE_MMAP && HAVE_MPROTECT && defined(MAP_ANONYMOUS)

  1598. #define USE_MMAP 1

  1599. #else

  1600. #define USE_MMAP 0

  1601. #endif

  1602. 

  1603.     /* precalculate horizontal scaler filter coefficients */

  1604.     {

  1605. #if HAVE_MMXEXT_INLINE

  1606. // can't downscale !!!

  1607.         if (c->canMMXEXTBeUsed && (flags & SWS_FAST_BILINEAR)) {

  1608.             c->lumMmxextFilterCodeSize = ff_init_hscaler_mmxext(dstW, c->lumXInc, NULL,

  1609.                                                              NULL, NULL, 8);

  1610.             c->chrMmxextFilterCodeSize = ff_init_hscaler_mmxext(c->chrDstW, c->chrXInc,

  1611.                                                              NULL, NULL, NULL, 4);

  1612. 

  1613. #if USE_MMAP

  1614.             c->lumMmxextFilterCode = mmap(NULL, c->lumMmxextFilterCodeSize,

  1615.                                           PROT_READ | PROT_WRITE,

  1616.                                           MAP_PRIVATE | MAP_ANONYMOUS,

  1617.                                           -1, 0);

  1618.             c->chrMmxextFilterCode = mmap(NULL, c->chrMmxextFilterCodeSize,

  1619.                                           PROT_READ | PROT_WRITE,

  1620.                                           MAP_PRIVATE | MAP_ANONYMOUS,

  1621.                                           -1, 0);

  1622. #elif HAVE_VIRTUALALLOC

  1623.             c->lumMmxextFilterCode = VirtualAlloc(NULL,

  1624.                                                   c->lumMmxextFilterCodeSize,

  1625.                                                   MEM_COMMIT,

  1626.                                                   PAGE_EXECUTE_READWRITE);

  1627.             c->chrMmxextFilterCode = VirtualAlloc(NULL,

  1628.                                                   c->chrMmxextFilterCodeSize,

  1629.                                                   MEM_COMMIT,

  1630.                                                   PAGE_EXECUTE_READWRITE);

  1631. #else

  1632.             c->lumMmxextFilterCode = av_malloc(c->lumMmxextFilterCodeSize);

  1633.             c->chrMmxextFilterCode = av_malloc(c->chrMmxextFilterCodeSize);

  1634. #endif

  1635. 

  1636. #ifdef MAP_ANONYMOUS

  1637.             if (c->lumMmxextFilterCode == MAP_FAILED || c->chrMmxextFilterCode == MAP_FAILED)

  1638. #else

  1639.             if (!c->lumMmxextFilterCode || !c->chrMmxextFilterCode)

  1640. #endif

  1641.             {

  1642.                 av_log(c, AV_LOG_ERROR, "Failed to allocate MMX2FilterCode\n");

  1643.                 return AVERROR(ENOMEM);

  1644.             }

  1645. 

  1646.             FF_ALLOCZ_OR_GOTO(c, c->hLumFilter,    (dstW           / 8 + 8) * sizeof(int16_t), fail);

  1647.             FF_ALLOCZ_OR_GOTO(c, c->hChrFilter,    (c->chrDstW     / 4 + 8) * sizeof(int16_t), fail);

  1648.             FF_ALLOCZ_OR_GOTO(c, c->hLumFilterPos, (dstW       / 2 / 8 + 8) * sizeof(int32_t), fail);

  1649.             FF_ALLOCZ_OR_GOTO(c, c->hChrFilterPos, (c->chrDstW / 2 / 4 + 8) * sizeof(int32_t), fail);

  1650. 

  1651.             ff_init_hscaler_mmxext(      dstW, c->lumXInc, c->lumMmxextFilterCode,

  1652.                                 c->hLumFilter, (uint32_t*)c->hLumFilterPos, 8);

  1653.             ff_init_hscaler_mmxext(c->chrDstW, c->chrXInc, c->chrMmxextFilterCode,

  1654.                                 c->hChrFilter, (uint32_t*)c->hChrFilterPos, 4);

  1655. 

  1656. #if USE_MMAP

  1657.             if (   mprotect(c->lumMmxextFilterCode, c->lumMmxextFilterCodeSize, PROT_EXEC | PROT_READ) == -1

  1658.                 || mprotect(c->chrMmxextFilterCode, c->chrMmxextFilterCodeSize, PROT_EXEC | PROT_READ) == -1) {

  1659.                 av_log(c, AV_LOG_ERROR, "mprotect failed, cannot use fast bilinear scaler\n");

  1660.                 goto fail;

  1661.             }

  1662. #endif

  1663.         } else

  1664. #endif /* HAVE_MMXEXT_INLINE */

  1665.         {

  1666.             const int filterAlign = X86_MMX(cpu_flags)     ? 4 :

  1667.                                     PPC_ALTIVEC(cpu_flags) ? 8 :

  1668.                                     have_neon(cpu_flags)   ? 8 : 1;

  1669. 

  1670.             if ((ret = initFilter(&c->hLumFilter, &c->hLumFilterPos,

  1671.                            &c->hLumFilterSize, c->lumXInc,

  1672.                            srcW, dstW, filterAlign, 1 << 14,

  1673.                            (flags & SWS_BICUBLIN) ? (flags | SWS_BICUBIC) : flags,

  1674.                            cpu_flags, srcFilter->lumH, dstFilter->lumH,

  1675.                            c->param,

  1676.                            get_local_pos(c, 0, 0, 0),

  1677.                            get_local_pos(c, 0, 0, 0))) < 0)

  1678.                 goto fail;

  1679.             if ((ret = initFilter(&c->hChrFilter, &c->hChrFilterPos,

  1680.                            &c->hChrFilterSize, c->chrXInc,

  1681.                            c->chrSrcW, c->chrDstW, filterAlign, 1 << 14,

  1682.                            (flags & SWS_BICUBLIN) ? (flags | SWS_BILINEAR) : flags,

  1683.                            cpu_flags, srcFilter->chrH, dstFilter->chrH,

  1684.                            c->param,

  1685.                            get_local_pos(c, c->chrSrcHSubSample, c->src_h_chr_pos, 0),

  1686.                            get_local_pos(c, c->chrDstHSubSample, c->dst_h_chr_pos, 0))) < 0)

  1687.                 goto fail;

  1688.         }

  1689.     } // initialize horizontal stuff

  1690. 

  1691.     /* precalculate vertical scaler filter coefficients */

  1692.     {

  1693.         const int filterAlign = X86_MMX(cpu_flags)     ? 2 :

  1694.                                 PPC_ALTIVEC(cpu_flags) ? 8 :

  1695.                                 have_neon(cpu_flags)   ? 2 : 1;

  1696. 

  1697.         if ((ret = initFilter(&c->vLumFilter, &c->vLumFilterPos, &c->vLumFilterSize,

  1698.                        c->lumYInc, srcH, dstH, filterAlign, (1 << 12),

  1699.                        (flags & SWS_BICUBLIN) ? (flags | SWS_BICUBIC) : flags,

  1700.                        cpu_flags, srcFilter->lumV, dstFilter->lumV,

  1701.                        c->param,

  1702.                        get_local_pos(c, 0, 0, 1),

  1703.                        get_local_pos(c, 0, 0, 1))) < 0)

  1704.             goto fail;

  1705.         if ((ret = initFilter(&c->vChrFilter, &c->vChrFilterPos, &c->vChrFilterSize,

  1706.                        c->chrYInc, c->chrSrcH, c->chrDstH,

  1707.                        filterAlign, (1 << 12),

  1708.                        (flags & SWS_BICUBLIN) ? (flags | SWS_BILINEAR) : flags,

  1709.                        cpu_flags, srcFilter->chrV, dstFilter->chrV,

  1710.                        c->param,

  1711.                        get_local_pos(c, c->chrSrcVSubSample, c->src_v_chr_pos, 1),

  1712.                        get_local_pos(c, c->chrDstVSubSample, c->dst_v_chr_pos, 1))) < 0)

  1713. 

  1714.             goto fail;

  1715. 

  1716. #if HAVE_ALTIVEC

  1717.         FF_ALLOC_OR_GOTO(c, c->vYCoeffsBank, sizeof(vector signed short) * c->vLumFilterSize * c->dstH,    fail);

  1718.         FF_ALLOC_OR_GOTO(c, c->vCCoeffsBank, sizeof(vector signed short) * c->vChrFilterSize * c->chrDstH, fail);

  1719. 

  1720.         for (i = 0; i < c->vLumFilterSize * c->dstH; i++) {

  1721.             int j;

  1722.             short *p = (short *)&c->vYCoeffsBank[i];

  1723.             for (j = 0; j < 8; j++)

  1724.                 p[j] = c->vLumFilter[i];

  1725.         }

  1726. 

  1727.         for (i = 0; i < c->vChrFilterSize * c->chrDstH; i++) {

  1728.             int j;

  1729.             short *p = (short *)&c->vCCoeffsBank[i];

  1730.             for (j = 0; j < 8; j++)

  1731.                 p[j] = c->vChrFilter[i];

  1732.         }

  1733. #endif

  1734.     }

  1735. 

  1736.     for (i = 0; i < 4; i++)

  1737.         FF_ALLOCZ_OR_GOTO(c, c->dither_error[i], (c->dstW+2) * sizeof(int), fail);

  1738. 

  1739.     c->needAlpha = (CONFIG_SWSCALE_ALPHA && isALPHA(c->srcFormat) && isALPHA(c->dstFormat)) ? 1 : 0;

  1740. 

  1741.     // 64 / c->scalingBpp is the same as 16 / sizeof(scaling_intermediate)

  1742.     c->uv_off   = (dst_stride>>1) + 64 / (c->dstBpc &~ 7);

  1743.     c->uv_offx2 = dst_stride + 16;

  1744. 

  1745.     av_assert0(c->chrDstH <= dstH);

  1746. 

  1747.     if (flags & SWS_PRINT_INFO) {

  1748.         const char *scaler = NULL, *cpucaps;

  1749. 

  1750.         for (i = 0; i < FF_ARRAY_ELEMS(scale_algorithms); i++) {

  1751.             if (flags & scale_algorithms[i].flag) {

  1752.                 scaler = scale_algorithms[i].description;

  1753.                 break;

  1754.             }

  1755.         }

  1756.         if (!scaler)

  1757.             scaler =  "ehh flags invalid?!";

  1758.         av_log(c, AV_LOG_INFO, "%s scaler, from %s to %s%s ",

  1759.                scaler,

  1760.                av_get_pix_fmt_name(srcFormat),

  1761. #ifdef DITHER1XBPP

  1762.                dstFormat == AV_PIX_FMT_BGR555   || dstFormat == AV_PIX_FMT_BGR565   ||

  1763.                dstFormat == AV_PIX_FMT_RGB444BE || dstFormat == AV_PIX_FMT_RGB444LE ||

  1764.                dstFormat == AV_PIX_FMT_BGR444BE || dstFormat == AV_PIX_FMT_BGR444LE ?

  1765.                                                              "dithered " : "",

  1766. #else

  1767.                "",

  1768. #endif

  1769.                av_get_pix_fmt_name(dstFormat));

  1770. 

  1771.         if (INLINE_MMXEXT(cpu_flags))

  1772.             cpucaps = "MMXEXT";

  1773.         else if (INLINE_AMD3DNOW(cpu_flags))

  1774.             cpucaps = "3DNOW";

  1775.         else if (INLINE_MMX(cpu_flags))

  1776.             cpucaps = "MMX";

  1777.         else if (PPC_ALTIVEC(cpu_flags))

  1778.             cpucaps = "AltiVec";

  1779.         else

  1780.             cpucaps = "C";

  1781. 

  1782.         av_log(c, AV_LOG_INFO, "using %s\n", cpucaps);

  1783. 

  1784.         av_log(c, AV_LOG_VERBOSE, "%dx%d -> %dx%d\n", srcW, srcH, dstW, dstH);

  1785.         av_log(c, AV_LOG_DEBUG,

  1786.                "lum srcW=%d srcH=%d dstW=%d dstH=%d xInc=%d yInc=%d\n",

  1787.                c->srcW, c->srcH, c->dstW, c->dstH, c->lumXInc, c->lumYInc);

  1788.         av_log(c, AV_LOG_DEBUG,

  1789.                "chr srcW=%d srcH=%d dstW=%d dstH=%d xInc=%d yInc=%d\n",

  1790.                c->chrSrcW, c->chrSrcH, c->chrDstW, c->chrDstH,

  1791.                c->chrXInc, c->chrYInc);

  1792.     }

  1793. 

  1794.     /* alpha blend special case, note this has been split via cascaded contexts if its scaled */

  1795.     if (unscaled && !usesHFilter && !usesVFilter &&

  1796.         c->alphablend != SWS_ALPHA_BLEND_NONE &&

  1797.         isALPHA(srcFormat) &&

  1798.         (c->srcRange == c->dstRange || isAnyRGB(dstFormat)) &&

  1799.         alphaless_fmt(srcFormat) == dstFormat

  1800.     ) {

  1801.         c->swscale = ff_sws_alphablendaway;

  1802. 

  1803.         if (flags & SWS_PRINT_INFO)

  1804.             av_log(c, AV_LOG_INFO,

  1805.                     "using alpha blendaway %s -> %s special converter\n",

  1806.                     av_get_pix_fmt_name(srcFormat), av_get_pix_fmt_name(dstFormat));

  1807.         return 0;

  1808.     }

  1809. 

  1810.     /* unscaled special cases */

  1811.     if (unscaled && !usesHFilter && !usesVFilter &&

  1812.         (c->srcRange == c->dstRange || isAnyRGB(dstFormat) ||

  1813.          srcFormat == AV_PIX_FMT_GRAYF32 && dstFormat == AV_PIX_FMT_GRAY8 ||

  1814.          srcFormat == AV_PIX_FMT_GRAY8 && dstFormat == AV_PIX_FMT_GRAYF32)) {

  1815.         ff_get_unscaled_swscale(c);

  1816. 

  1817.         if (c->swscale) {

  1818.             if (flags & SWS_PRINT_INFO)

  1819.                 av_log(c, AV_LOG_INFO,

  1820.                        "using unscaled %s -> %s special converter\n",

  1821.                        av_get_pix_fmt_name(srcFormat), av_get_pix_fmt_name(dstFormat));

  1822.             return 0;

  1823.         }

  1824.     }

  1825. 

  1826.     c->swscale = ff_getSwsFunc(c);

  1827.     return ff_init_filters(c);

  1828. fail: // FIXME replace things by appropriate error codes

  1829.     if (ret == RETCODE_USE_CASCADE)  {

  1830.         int tmpW = sqrt(srcW * (int64_t)dstW);

  1831.         int tmpH = sqrt(srcH * (int64_t)dstH);

  1832.         enum AVPixelFormat tmpFormat = AV_PIX_FMT_YUV420P;

  1833. 

  1834.         if (isALPHA(srcFormat))

  1835.             tmpFormat = AV_PIX_FMT_YUVA420P;

  1836. 

  1837.         if (srcW*(int64_t)srcH <= 4LL*dstW*dstH)

  1838.             return AVERROR(EINVAL);

  1839. 

  1840.         ret = av_image_alloc(c->cascaded_tmp, c->cascaded_tmpStride,

  1841.                              tmpW, tmpH, tmpFormat, 64);

  1842.         if (ret < 0)

  1843.             return ret;

  1844. 

  1845.         c->cascaded_context[0] = sws_getContext(srcW, srcH, srcFormat,

  1846.                                                 tmpW, tmpH, tmpFormat,

  1847.                                                 flags, srcFilter, NULL, c->param);

  1848.         if (!c->cascaded_context[0])

  1849.             return -1;

  1850. 

  1851.         c->cascaded_context[1] = sws_getContext(tmpW, tmpH, tmpFormat,

  1852.                                                 dstW, dstH, dstFormat,

  1853.                                                 flags, NULL, dstFilter, c->param);

  1854.         if (!c->cascaded_context[1])

  1855.             return -1;

  1856.         return 0;

  1857.     }

  1858.     return -1;

  1859. }

  1860. 

  1861. SwsContext *sws_alloc_set_opts(int srcW, int srcH, enum AVPixelFormat srcFormat,

  1862.                                int dstW, int dstH, enum AVPixelFormat dstFormat,

  1863.                                int flags, const double *param)

  1864. {

  1865.     SwsContext *c;

  1866. 

  1867.     if (!(c = sws_alloc_context()))

  1868.         return NULL;

  1869. 

  1870.     c->flags     = flags;

  1871.     c->srcW      = srcW;

  1872.     c->srcH      = srcH;

  1873.     c->dstW      = dstW;

  1874.     c->dstH      = dstH;

  1875.     c->srcFormat = srcFormat;

  1876.     c->dstFormat = dstFormat;

  1877. 

  1878.     if (param) {

  1879.         c->param[0] = param[0];

  1880.         c->param[1] = param[1];

  1881.     }

  1882. 

  1883.     return c;

  1884. }

  1885. 

  1886. SwsContext *sws_getContext(int srcW, int srcH, enum AVPixelFormat srcFormat,

  1887.                            int dstW, int dstH, enum AVPixelFormat dstFormat,

  1888.                            int flags, SwsFilter *srcFilter,

  1889.                            SwsFilter *dstFilter, const double *param)

  1890. {

  1891.     SwsContext *c;

  1892. 

  1893.     c = sws_alloc_set_opts(srcW, srcH, srcFormat,

  1894.                            dstW, dstH, dstFormat,

  1895.                            flags, param);

  1896.     if (!c)

  1897.         return NULL;

  1898. 

  1899.     if (sws_init_context(c, srcFilter, dstFilter) < 0) {

  1900.         sws_freeContext(c);

  1901.         return NULL;

  1902.     }

  1903. 

  1904.     return c;

  1905. }

  1906. 

  1907. static int isnan_vec(SwsVector *a)

  1908. {

  1909.     int i;

  1910.     for (i=0; i<a->length; i++)

  1911.         if (isnan(a->coeff[i]))

  1912.             return 1;

  1913.     return 0;

  1914. }

  1915. 

  1916. static void makenan_vec(SwsVector *a)

  1917. {

  1918.     int i;

  1919.     for (i=0; i<a->length; i++)

  1920.         a->coeff[i] = NAN;

  1921. }

  1922. 

  1923. SwsFilter *sws_getDefaultFilter(float lumaGBlur, float chromaGBlur,

  1924.                                 float lumaSharpen, float chromaSharpen,

  1925.                                 float chromaHShift, float chromaVShift,

  1926.                                 int verbose)

  1927. {

  1928.     SwsFilter *filter = av_malloc(sizeof(SwsFilter));

  1929.     if (!filter)

  1930.         return NULL;

  1931. 

  1932.     if (lumaGBlur != 0.0) {

  1933.         filter->lumH = sws_getGaussianVec(lumaGBlur, 3.0);

  1934.         filter->lumV = sws_getGaussianVec(lumaGBlur, 3.0);

  1935.     } else {

  1936.         filter->lumH = sws_getIdentityVec();

  1937.         filter->lumV = sws_getIdentityVec();

  1938.     }

  1939. 

  1940.     if (chromaGBlur != 0.0) {

  1941.         filter->chrH = sws_getGaussianVec(chromaGBlur, 3.0);

  1942.         filter->chrV = sws_getGaussianVec(chromaGBlur, 3.0);

  1943.     } else {

  1944.         filter->chrH = sws_getIdentityVec();

  1945.         filter->chrV = sws_getIdentityVec();

  1946.     }

  1947. 

  1948.     if (!filter->lumH || !filter->lumV || !filter->chrH || !filter->chrV)

  1949.         goto fail;

  1950. 

  1951.     if (chromaSharpen != 0.0) {

  1952.         SwsVector *id = sws_getIdentityVec();

  1953.         if (!id)

  1954.             goto fail;

  1955.         sws_scaleVec(filter->chrH, -chromaSharpen);

  1956.         sws_scaleVec(filter->chrV, -chromaSharpen);

  1957.         sws_addVec(filter->chrH, id);

  1958.         sws_addVec(filter->chrV, id);

  1959.         sws_freeVec(id);

  1960.     }

  1961. 

  1962.     if (lumaSharpen != 0.0) {

  1963.         SwsVector *id = sws_getIdentityVec();

  1964.         if (!id)

  1965.             goto fail;

  1966.         sws_scaleVec(filter->lumH, -lumaSharpen);

  1967.         sws_scaleVec(filter->lumV, -lumaSharpen);

  1968.         sws_addVec(filter->lumH, id);

  1969.         sws_addVec(filter->lumV, id);

  1970.         sws_freeVec(id);

  1971.     }

  1972. 

  1973.     if (chromaHShift != 0.0)

  1974.         sws_shiftVec(filter->chrH, (int)(chromaHShift + 0.5));

  1975. 

  1976.     if (chromaVShift != 0.0)

  1977.         sws_shiftVec(filter->chrV, (int)(chromaVShift + 0.5));

  1978. 

  1979.     sws_normalizeVec(filter->chrH, 1.0);

  1980.     sws_normalizeVec(filter->chrV, 1.0);

  1981.     sws_normalizeVec(filter->lumH, 1.0);

  1982.     sws_normalizeVec(filter->lumV, 1.0);

  1983. 

  1984.     if (isnan_vec(filter->chrH) ||

  1985.         isnan_vec(filter->chrV) ||

  1986.         isnan_vec(filter->lumH) ||

  1987.         isnan_vec(filter->lumV))

  1988.         goto fail;

  1989. 

  1990.     if (verbose)

  1991.         sws_printVec2(filter->chrH, NULL, AV_LOG_DEBUG);

  1992.     if (verbose)

  1993.         sws_printVec2(filter->lumH, NULL, AV_LOG_DEBUG);

  1994. 

  1995.     return filter;

  1996. 

  1997. fail:

  1998.     sws_freeVec(filter->lumH);

  1999.     sws_freeVec(filter->lumV);

  2000.     sws_freeVec(filter->chrH);

  2001.     sws_freeVec(filter->chrV);

  2002.     av_freep(&filter);

  2003.     return NULL;

  2004. }

  2005. 

  2006. SwsVector *sws_allocVec(int length)

  2007. {

  2008.     SwsVector *vec;

  2009. 

  2010.     if(length <= 0 || length > INT_MAX/ sizeof(double))

  2011.         return NULL;

  2012. 

  2013.     vec = av_malloc(sizeof(SwsVector));

  2014.     if (!vec)

  2015.         return NULL;

  2016.     vec->length = length;

  2017.     vec->coeff  = av_malloc(sizeof(double) * length);

  2018.     if (!vec->coeff)

  2019.         av_freep(&vec);

  2020.     return vec;

  2021. }

  2022. 

  2023. SwsVector *sws_getGaussianVec(double variance, double quality)

  2024. {

  2025.     const int length = (int)(variance * quality + 0.5) | 1;

  2026.     int i;

  2027.     double middle  = (length - 1) * 0.5;

  2028.     SwsVector *vec;

  2029. 

  2030.     if(variance < 0 || quality < 0)

  2031.         return NULL;

  2032. 

  2033.     vec = sws_allocVec(length);

  2034. 

  2035.     if (!vec)

  2036.         return NULL;

  2037. 

  2038.     for (i = 0; i < length; i++) {

  2039.         double dist = i - middle;

  2040.         vec->coeff[i] = exp(-dist * dist / (2 * variance * variance)) /

  2041.                         sqrt(2 * variance * M_PI);

  2042.     }

  2043. 

  2044.     sws_normalizeVec(vec, 1.0);

  2045. 

  2046.     return vec;

  2047. }

  2048. 

  2049. /**

  2050.  * Allocate and return a vector with length coefficients, all

  2051.  * with the same value c.

  2052.  */

  2053. #if !FF_API_SWS_VECTOR

  2054. static

  2055. #endif

  2056. SwsVector *sws_getConstVec(double c, int length)

  2057. {

  2058.     int i;

  2059.     SwsVector *vec = sws_allocVec(length);

  2060. 

  2061.     if (!vec)

  2062.         return NULL;

  2063. 

  2064.     for (i = 0; i < length; i++)

  2065.         vec->coeff[i] = c;

  2066. 

  2067.     return vec;

  2068. }

  2069. 

  2070. /**

  2071.  * Allocate and return a vector with just one coefficient, with

  2072.  * value 1.0.

  2073.  */

  2074. #if !FF_API_SWS_VECTOR

  2075. static

  2076. #endif

  2077. SwsVector *sws_getIdentityVec(void)

  2078. {

  2079.     return sws_getConstVec(1.0, 1);

  2080. }

  2081. 

  2082. static double sws_dcVec(SwsVector *a)

  2083. {

  2084.     int i;

  2085.     double sum = 0;

  2086. 

  2087.     for (i = 0; i < a->length; i++)

  2088.         sum += a->coeff[i];

  2089. 

  2090.     return sum;

  2091. }

  2092. 

  2093. void sws_scaleVec(SwsVector *a, double scalar)

  2094. {

  2095.     int i;

  2096. 

  2097.     for (i = 0; i < a->length; i++)

  2098.         a->coeff[i] *= scalar;

  2099. }

  2100. 

  2101. void sws_normalizeVec(SwsVector *a, double height)

  2102. {

  2103.     sws_scaleVec(a, height / sws_dcVec(a));

  2104. }

  2105. 

  2106. #if FF_API_SWS_VECTOR

  2107. static SwsVector *sws_getConvVec(SwsVector *a, SwsVector *b)

  2108. {

  2109.     int length = a->length + b->length - 1;

  2110.     int i, j;

  2111.     SwsVector *vec = sws_getConstVec(0.0, length);

  2112. 

  2113.     if (!vec)

  2114.         return NULL;

  2115. 

  2116.     for (i = 0; i < a->length; i++) {

  2117.         for (j = 0; j < b->length; j++) {

  2118.             vec->coeff[i + j] += a->coeff[i] * b->coeff[j];

  2119.         }

  2120.     }

  2121. 

  2122.     return vec;

  2123. }

  2124. #endif

  2125. 

  2126. static SwsVector *sws_sumVec(SwsVector *a, SwsVector *b)

  2127. {

  2128.     int length = FFMAX(a->length, b->length);

  2129.     int i;

  2130.     SwsVector *vec = sws_getConstVec(0.0, length);

  2131. 

  2132.     if (!vec)

  2133.         return NULL;

  2134. 

  2135.     for (i = 0; i < a->length; i++)

  2136.         vec->coeff[i + (length - 1) / 2 - (a->length - 1) / 2] += a->coeff[i];

  2137.     for (i = 0; i < b->length; i++)

  2138.         vec->coeff[i + (length - 1) / 2 - (b->length - 1) / 2] += b->coeff[i];

  2139. 

  2140.     return vec;

  2141. }

  2142. 

  2143. #if FF_API_SWS_VECTOR

  2144. static SwsVector *sws_diffVec(SwsVector *a, SwsVector *b)

  2145. {

  2146.     int length = FFMAX(a->length, b->length);

  2147.     int i;

  2148.     SwsVector *vec = sws_getConstVec(0.0, length);

  2149. 

  2150.     if (!vec)

  2151.         return NULL;

  2152. 

  2153.     for (i = 0; i < a->length; i++)

  2154.         vec->coeff[i + (length - 1) / 2 - (a->length - 1) / 2] += a->coeff[i];

  2155.     for (i = 0; i < b->length; i++)

  2156.         vec->coeff[i + (length - 1) / 2 - (b->length - 1) / 2] -= b->coeff[i];

  2157. 

  2158.     return vec;

  2159. }

  2160. #endif

  2161. 

  2162. /* shift left / or right if "shift" is negative */

  2163. static SwsVector *sws_getShiftedVec(SwsVector *a, int shift)

  2164. {

  2165.     int length = a->length + FFABS(shift) * 2;

  2166.     int i;

  2167.     SwsVector *vec = sws_getConstVec(0.0, length);

  2168. 

  2169.     if (!vec)

  2170.         return NULL;

  2171. 

  2172.     for (i = 0; i < a->length; i++) {

  2173.         vec->coeff[i + (length    - 1) / 2 -

  2174.                        (a->length - 1) / 2 - shift] = a->coeff[i];

  2175.     }

  2176. 

  2177.     return vec;

  2178. }

  2179. 

  2180. #if !FF_API_SWS_VECTOR

  2181. static

  2182. #endif

  2183. void sws_shiftVec(SwsVector *a, int shift)

  2184. {

  2185.     SwsVector *shifted = sws_getShiftedVec(a, shift);

  2186.     if (!shifted) {

  2187.         makenan_vec(a);

  2188.         return;

  2189.     }

  2190.     av_free(a->coeff);

  2191.     a->coeff  = shifted->coeff;

  2192.     a->length = shifted->length;

  2193.     av_free(shifted);

  2194. }

  2195. 

  2196. #if !FF_API_SWS_VECTOR

  2197. static

  2198. #endif

  2199. void sws_addVec(SwsVector *a, SwsVector *b)

  2200. {

  2201.     SwsVector *sum = sws_sumVec(a, b);

  2202.     if (!sum) {

  2203.         makenan_vec(a);

  2204.         return;

  2205.     }

  2206.     av_free(a->coeff);

  2207.     a->coeff  = sum->coeff;

  2208.     a->length = sum->length;

  2209.     av_free(sum);

  2210. }

  2211. 

  2212. #if FF_API_SWS_VECTOR

  2213. void sws_subVec(SwsVector *a, SwsVector *b)

  2214. {

  2215.     SwsVector *diff = sws_diffVec(a, b);

  2216.     if (!diff) {

  2217.         makenan_vec(a);

  2218.         return;

  2219.     }

  2220.     av_free(a->coeff);

  2221.     a->coeff  = diff->coeff;

  2222.     a->length = diff->length;

  2223.     av_free(diff);

  2224. }

  2225. 

  2226. void sws_convVec(SwsVector *a, SwsVector *b)

  2227. {

  2228.     SwsVector *conv = sws_getConvVec(a, b);

  2229.     if (!conv) {

  2230.         makenan_vec(a);

  2231.         return;

  2232.     }

  2233.     av_free(a->coeff);

  2234.     a->coeff  = conv->coeff;

  2235.     a->length = conv->length;

  2236.     av_free(conv);

  2237. }

  2238. 

  2239. SwsVector *sws_cloneVec(SwsVector *a)

  2240. {

  2241.     SwsVector *vec = sws_allocVec(a->length);

  2242. 

  2243.     if (!vec)

  2244.         return NULL;

  2245. 

  2246.     memcpy(vec->coeff, a->coeff, a->length * sizeof(*a->coeff));

  2247. 

  2248.     return vec;

  2249. }

  2250. #endif

  2251. 

  2252. /**

  2253.  * Print with av_log() a textual representation of the vector a

  2254.  * if log_level <= av_log_level.

  2255.  */

  2256. #if !FF_API_SWS_VECTOR

  2257. static

  2258. #endif

  2259. void sws_printVec2(SwsVector *a, AVClass *log_ctx, int log_level)

  2260. {

  2261.     int i;

  2262.     double max = 0;

  2263.     double min = 0;

  2264.     double range;

  2265. 

  2266.     for (i = 0; i < a->length; i++)

  2267.         if (a->coeff[i] > max)

  2268.             max = a->coeff[i];

  2269. 

  2270.     for (i = 0; i < a->length; i++)

  2271.         if (a->coeff[i] < min)

  2272.             min = a->coeff[i];

  2273. 

  2274.     range = max - min;

  2275. 

  2276.     for (i = 0; i < a->length; i++) {

  2277.         int x = (int)((a->coeff[i] - min) * 60.0 / range + 0.5);

  2278.         av_log(log_ctx, log_level, "%1.3f ", a->coeff[i]);

  2279.         for (; x > 0; x--)

  2280.             av_log(log_ctx, log_level, " ");

  2281.         av_log(log_ctx, log_level, "|\n");

  2282.     }

  2283. }

  2284. 

  2285. void sws_freeVec(SwsVector *a)

  2286. {

  2287.     if (!a)

  2288.         return;

  2289.     av_freep(&a->coeff);

  2290.     a->length = 0;

  2291.     av_free(a);

  2292. }

  2293. 

  2294. void sws_freeFilter(SwsFilter *filter)

  2295. {

  2296.     if (!filter)

  2297.         return;

  2298. 

  2299.     sws_freeVec(filter->lumH);

  2300.     sws_freeVec(filter->lumV);

  2301.     sws_freeVec(filter->chrH);

  2302.     sws_freeVec(filter->chrV);

  2303.     av_free(filter);

  2304. }

  2305. 

  2306. void sws_freeContext(SwsContext *c)

  2307. {

  2308.     int i;

  2309.     if (!c)

  2310.         return;

  2311. 

  2312.     for (i = 0; i < 4; i++)

  2313.         av_freep(&c->dither_error[i]);

  2314. 

  2315.     av_freep(&c->vLumFilter);

  2316.     av_freep(&c->vChrFilter);

  2317.     av_freep(&c->hLumFilter);

  2318.     av_freep(&c->hChrFilter);

  2319. #if HAVE_ALTIVEC

  2320.     av_freep(&c->vYCoeffsBank);

  2321.     av_freep(&c->vCCoeffsBank);

  2322. #endif

  2323. 

  2324.     av_freep(&c->vLumFilterPos);

  2325.     av_freep(&c->vChrFilterPos);

  2326.     av_freep(&c->hLumFilterPos);

  2327.     av_freep(&c->hChrFilterPos);

  2328. 

  2329. #if HAVE_MMX_INLINE

  2330. #if USE_MMAP

  2331.     if (c->lumMmxextFilterCode)

  2332.         munmap(c->lumMmxextFilterCode, c->lumMmxextFilterCodeSize);

  2333.     if (c->chrMmxextFilterCode)

  2334.         munmap(c->chrMmxextFilterCode, c->chrMmxextFilterCodeSize);

  2335. #elif HAVE_VIRTUALALLOC

  2336.     if (c->lumMmxextFilterCode)

  2337.         VirtualFree(c->lumMmxextFilterCode, 0, MEM_RELEASE);

  2338.     if (c->chrMmxextFilterCode)

  2339.         VirtualFree(c->chrMmxextFilterCode, 0, MEM_RELEASE);

  2340. #else

  2341.     av_free(c->lumMmxextFilterCode);

  2342.     av_free(c->chrMmxextFilterCode);

  2343. #endif

  2344.     c->lumMmxextFilterCode = NULL;

  2345.     c->chrMmxextFilterCode = NULL;

  2346. #endif /* HAVE_MMX_INLINE */

  2347. 

  2348.     av_freep(&c->yuvTable);

  2349.     av_freep(&c->formatConvBuffer);

  2350. 

  2351.     sws_freeContext(c->cascaded_context[0]);

  2352.     sws_freeContext(c->cascaded_context[1]);

  2353.     sws_freeContext(c->cascaded_context[2]);

  2354.     memset(c->cascaded_context, 0, sizeof(c->cascaded_context));

  2355.     av_freep(&c->cascaded_tmp[0]);

  2356.     av_freep(&c->cascaded1_tmp[0]);

  2357. 

  2358.     av_freep(&c->gamma);

  2359.     av_freep(&c->inv_gamma);

  2360. 

  2361.     ff_free_filters(c);

  2362. 

  2363.     av_free(c);

  2364. }

  2365. 

  2366. struct SwsContext *sws_getCachedContext(struct SwsContext *context, int srcW,

  2367.                                         int srcH, enum AVPixelFormat srcFormat,

  2368.                                         int dstW, int dstH,

  2369.                                         enum AVPixelFormat dstFormat, int flags,

  2370.                                         SwsFilter *srcFilter,

  2371.                                         SwsFilter *dstFilter,

  2372.                                         const double *param)

  2373. {

  2374.     static const double default_param[2] = { SWS_PARAM_DEFAULT,

  2375.                                              SWS_PARAM_DEFAULT };

  2376.     int64_t src_h_chr_pos = -513, dst_h_chr_pos = -513,

  2377.             src_v_chr_pos = -513, dst_v_chr_pos = -513;

  2378. 

  2379.     if (!param)

  2380.         param = default_param;

  2381. 

  2382.     if (context &&

  2383.         (context->srcW      != srcW      ||

  2384.          context->srcH      != srcH      ||

  2385.          context->srcFormat != srcFormat ||

  2386.          context->dstW      != dstW      ||

  2387.          context->dstH      != dstH      ||

  2388.          context->dstFormat != dstFormat ||

  2389.          context->flags     != flags     ||

  2390.          context->param[0]  != param[0]  ||

  2391.          context->param[1]  != param[1])) {

  2392. 

  2393.         av_opt_get_int(context, "src_h_chr_pos", 0, &src_h_chr_pos);

  2394.         av_opt_get_int(context, "src_v_chr_pos", 0, &src_v_chr_pos);

  2395.         av_opt_get_int(context, "dst_h_chr_pos", 0, &dst_h_chr_pos);

  2396.         av_opt_get_int(context, "dst_v_chr_pos", 0, &dst_v_chr_pos);

  2397.         sws_freeContext(context);

  2398.         context = NULL;

  2399.     }

  2400. 

  2401.     if (!context) {

  2402.         if (!(context = sws_alloc_context()))

  2403.             return NULL;

  2404.         context->srcW      = srcW;

  2405.         context->srcH      = srcH;

  2406.         context->srcFormat = srcFormat;

  2407.         context->dstW      = dstW;

  2408.         context->dstH      = dstH;

  2409.         context->dstFormat = dstFormat;

  2410.         context->flags     = flags;

  2411.         context->param[0]  = param[0];

  2412.         context->param[1]  = param[1];

  2413. 

  2414.         av_opt_set_int(context, "src_h_chr_pos", src_h_chr_pos, 0);

  2415.         av_opt_set_int(context, "src_v_chr_pos", src_v_chr_pos, 0);

  2416.         av_opt_set_int(context, "dst_h_chr_pos", dst_h_chr_pos, 0);

  2417.         av_opt_set_int(context, "dst_v_chr_pos", dst_v_chr_pos, 0);

  2418. 

  2419.         if (sws_init_context(context, srcFilter, dstFilter) < 0) {

  2420.             sws_freeContext(context);

  2421.             return NULL;

  2422.         }

  2423.     }

  2424.     return context;

  2425. }