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.
97341 Commits
32 Branches
940MB
Tree: 40a9363033
Branches Tags
${ item.name }
Create branch ${ searchTerm }
from '40a9363033'
${ noResults }
FFmpeg/libswscale/utils.c

2433 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[] = {

  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, 1 },

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

  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.     [AV_PIX_FMT_YUVA422P12BE] = { 1, 1 },

  264.     [AV_PIX_FMT_YUVA422P12LE] = { 1, 1 },

  265.     [AV_PIX_FMT_YUVA444P12BE] = { 1, 1 },

  266.     [AV_PIX_FMT_YUVA444P12LE] = { 1, 1 },

  267.     [AV_PIX_FMT_NV24]        = { 1, 1 },

  268.     [AV_PIX_FMT_NV42]        = { 1, 1 },

  269.     [AV_PIX_FMT_Y210LE]      = { 1, 0 },

  270. };

  271. 

  272. int sws_isSupportedInput(enum AVPixelFormat pix_fmt)

  273. {

  274.     return (unsigned)pix_fmt < FF_ARRAY_ELEMS(format_entries) ?

  275.            format_entries[pix_fmt].is_supported_in : 0;

  276. }

  277. 

  278. int sws_isSupportedOutput(enum AVPixelFormat pix_fmt)

  279. {

  280.     return (unsigned)pix_fmt < FF_ARRAY_ELEMS(format_entries) ?

  281.            format_entries[pix_fmt].is_supported_out : 0;

  282. }

  283. 

  284. int sws_isSupportedEndiannessConversion(enum AVPixelFormat pix_fmt)

  285. {

  286.     return (unsigned)pix_fmt < FF_ARRAY_ELEMS(format_entries) ?

  287.            format_entries[pix_fmt].is_supported_endianness : 0;

  288. }

  289. 

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

  291.                              double dist)

  292. {

  293.     if (dist <= 1.0)

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

  295.     else

  296.         return getSplineCoeff(0.0,

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

  298.                                c + 3.0 * d,

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

  300.                               dist - 1.0);

  301. }

  302. 

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

  304. {

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

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

  307.     }

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

  309.     return pos >> chr_subsample;

  310. }

  311. 

  312. typedef struct {

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

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

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

  316. } ScaleAlgorithm;

  317. 

  318. static const ScaleAlgorithm scale_algorithms[] = {

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

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

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

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

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

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

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

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

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

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

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

  330. };

  331. 

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

  333.                               int *outFilterSize, int xInc, int srcW,

  334.                               int dstW, int filterAlign, int one,

  335.                               int flags, int cpu_flags,

  336.                               SwsVector *srcFilter, SwsVector *dstFilter,

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

  338. {

  339.     int i;

  340.     int filterSize;

  341.     int filter2Size;

  342.     int minFilterSize;

  343.     int64_t *filter    = NULL;

  344.     int64_t *filter2   = NULL;

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

  346.     int ret            = -1;

  347. 

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

  349. 

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

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

  352. 

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

  354.         int i;

  355.         filterSize = 1;

  356.         FF_ALLOCZ_ARRAY_OR_GOTO(NULL, filter,

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

  358. 

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

  360.             filter[i * filterSize] = fone;

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

  362.         }

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

  364.         int i;

  365.         int64_t xDstInSrc;

  366.         filterSize = 1;

  367.         FF_ALLOC_ARRAY_OR_GOTO(NULL, filter,

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

  369. 

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

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

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

  373. 

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

  375.             filter[i]       = fone;

  376.             xDstInSrc      += xInc;

  377.         }

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

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

  380.         int i;

  381.         int64_t xDstInSrc;

  382.         filterSize = 2;

  383.         FF_ALLOC_ARRAY_OR_GOTO(NULL, filter,

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

  385. 

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

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

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

  389.             int j;

  390. 

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

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

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

  394.                 int64_t coeff = fone - FFABS((int64_t)xx * (1 << 16) - xDstInSrc) * (fone >> 16);

  395.                 if (coeff < 0)

  396.                     coeff = 0;

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

  398.                 xx++;

  399.             }

  400.             xDstInSrc += xInc;

  401.         }

  402.     } else {

  403.         int64_t xDstInSrc;

  404.         int sizeFactor = -1;

  405. 

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

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

  408.                 sizeFactor = scale_algorithms[i].size_factor;

  409.                 break;

  410.             }

  411.         }

  412.         if (flags & SWS_LANCZOS)

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

  414.         av_assert0(sizeFactor > 0);

  415. 

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

  417.             filterSize = 1 + sizeFactor;    // upscale

  418.         else

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

  420. 

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

  422.         filterSize = FFMAX(filterSize, 1);

  423. 

  424.         FF_ALLOC_ARRAY_OR_GOTO(NULL, filter,

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

  426. 

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

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

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

  430.             int j;

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

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

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

  434.                 double floatd;

  435.                 int64_t coeff;

  436. 

  437.                 if (xInc > 1 << 16)

  438.                     d = d * dstW / srcW;

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

  440. 

  441.                 if (flags & SWS_BICUBIC) {

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

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

  444. 

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

  446.                         coeff = 0.0;

  447.                     } else {

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

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

  450. 

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

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

  453.                                     (-18 * (1 << 24) + 12 * B + 6 * C) *  dd +

  454.                                       (6 * (1 << 24) -  2 * B)         * (1 << 30);

  455.                         else

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

  457.                                       (6 * B + 30 * C) * dd  +

  458.                                     (-12 * B - 48 * C) * d   +

  459.                                       (8 * B + 24 * C) * (1 << 30);

  460.                     }

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

  462.                 } else if (flags & SWS_X) {

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

  464.                     double c;

  465. 

  466.                     if (floatd < 1.0)

  467.                         c = cos(floatd * M_PI);

  468.                     else

  469.                         c = -1.0;

  470.                     if (c < 0.0)

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

  472.                     else

  473.                         c = pow(c, A);

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

  475.                 } else if (flags & SWS_AREA) {

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

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

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

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

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

  481.                     else

  482.                         coeff = 0.0;

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

  484.                 } else if (flags & SWS_GAUSS) {

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

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

  487.                 } else if (flags & SWS_SINC) {

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

  489.                 } else if (flags & SWS_LANCZOS) {

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

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

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

  493.                     if (floatd > p)

  494.                         coeff = 0;

  495.                 } else if (flags & SWS_BILINEAR) {

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

  497.                     if (coeff < 0)

  498.                         coeff = 0;

  499.                     coeff *= fone >> 30;

  500.                 } else if (flags & SWS_SPLINE) {

  501.                     double p = -2.196152422706632;

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

  503.                 } else {

  504.                     av_assert0(0);

  505.                 }

  506. 

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

  508.                 xx++;

  509.             }

  510.             xDstInSrc += 2 * xInc;

  511.         }

  512.     }

  513. 

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

  515.      * av_free(filter);

  516.      */

  517.     av_assert0(filterSize > 0);

  518.     filter2Size = filterSize;

  519.     if (srcFilter)

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

  521.     if (dstFilter)

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

  523.     av_assert0(filter2Size > 0);

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

  525. 

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

  527.         int j, k;

  528. 

  529.         if (srcFilter) {

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

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

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

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

  534.             }

  535.         } else {

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

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

  538.         }

  539.         // FIXME dstFilter

  540. 

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

  542.     }

  543.     av_freep(&filter);

  544. 

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

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

  547.     minFilterSize = 0;

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

  549.         int min = filter2Size;

  550.         int j;

  551.         int64_t cutOff = 0.0;

  552. 

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

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

  555.             int k;

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

  557. 

  558.             if (cutOff > SWS_MAX_REDUCE_CUTOFF * fone)

  559.                 break;

  560. 

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

  562.              * filter otherwise */

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

  564.                 break;

  565. 

  566.             // move filter coefficients left

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

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

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

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

  571.         }

  572. 

  573.         cutOff = 0;

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

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

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

  577. 

  578.             if (cutOff > SWS_MAX_REDUCE_CUTOFF * fone)

  579.                 break;

  580.             min--;

  581.         }

  582. 

  583.         if (min > minFilterSize)

  584.             minFilterSize = min;

  585.     }

  586. 

  587.     if (PPC_ALTIVEC(cpu_flags)) {

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

  589.         if (minFilterSize < 5)

  590.             filterAlign = 4;

  591. 

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

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

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

  595.         if (minFilterSize < 3)

  596.             filterAlign = 1;

  597.     }

  598. 

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

  600.         // special case for unscaled vertical filtering

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

  602.             filterAlign = 1;

  603.     }

  604. 

  605.     av_assert0(minFilterSize > 0);

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

  607.     av_assert0(filterSize > 0);

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

  609.     if (!filter)

  610.         goto fail;

  611.     if (filterSize >= MAX_FILTER_SIZE * 16 /

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

  613.         ret = RETCODE_USE_CASCADE;

  614.         goto fail;

  615.     }

  616.     *outFilterSize = filterSize;

  617. 

  618.     if (flags & SWS_PRINT_INFO)

  619.         av_log(NULL, AV_LOG_VERBOSE,

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

  621.                filter2Size, filterSize);

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

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

  624.         int j;

  625. 

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

  627.             if (j >= filter2Size)

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

  629.             else

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

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

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

  633.         }

  634.     }

  635. 

  636.     // FIXME try to align filterPos if possible

  637. 

  638.     // fix borders

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

  640.         int j;

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

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

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

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

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

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

  647.             }

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

  649.         }

  650. 

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

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

  653.             int64_t acc = 0;

  654. 

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

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

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

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

  659.                 }

  660.             }

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

  662.                 if (j < shift) {

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

  664.                 } else {

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

  666.                 }

  667.             }

  668. 

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

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

  671.         }

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

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

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

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

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

  677.             }

  678.         }

  679.     }

  680. 

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

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

  683.     FF_ALLOCZ_ARRAY_OR_GOTO(NULL, *outFilter,

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

  685. 

  686.     /* normalize & store in outFilter */

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

  688.         int j;

  689.         int64_t error = 0;

  690.         int64_t sum   = 0;

  691. 

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

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

  694.         }

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

  696.         if (!sum) {

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

  698.             sum = 1;

  699.         }

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

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

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

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

  704.             error                                  = v - intV * sum;

  705.         }

  706.     }

  707. 

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

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

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

  711.                                                       * read over the end */

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

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

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

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

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

  717.     }

  718. 

  719.     ret = 0;

  720. 

  721. fail:

  722.     if(ret < 0)

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

  724.     av_free(filter);

  725.     av_free(filter2);

  726.     return ret;

  727. }

  728. 

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

  730. {

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

  732.     int64_t vr =  table[0];

  733.     int64_t ub =  table[1];

  734.     int64_t ug = -table[2];

  735.     int64_t vg = -table[3];

  736.     int64_t ONE = 65536;

  737.     int64_t cy = ONE;

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

  739.     int i;

  740.     static const int8_t map[] = {

  741.     BY_IDX, GY_IDX, -1    , BY_IDX, BY_IDX, GY_IDX, -1    , BY_IDX,

  742.     RY_IDX, -1    , GY_IDX, RY_IDX, RY_IDX, -1    , GY_IDX, RY_IDX,

  743.     RY_IDX, GY_IDX, -1    , RY_IDX, RY_IDX, GY_IDX, -1    , RY_IDX,

  744.     BY_IDX, -1    , GY_IDX, BY_IDX, BY_IDX, -1    , GY_IDX, BY_IDX,

  745.     BU_IDX, GU_IDX, -1    , BU_IDX, BU_IDX, GU_IDX, -1    , BU_IDX,

  746.     RU_IDX, -1    , GU_IDX, RU_IDX, RU_IDX, -1    , GU_IDX, RU_IDX,

  747.     RU_IDX, GU_IDX, -1    , RU_IDX, RU_IDX, GU_IDX, -1    , RU_IDX,

  748.     BU_IDX, -1    , GU_IDX, BU_IDX, BU_IDX, -1    , GU_IDX, BU_IDX,

  749.     BV_IDX, GV_IDX, -1    , BV_IDX, BV_IDX, GV_IDX, -1    , BV_IDX,

  750.     RV_IDX, -1    , GV_IDX, RV_IDX, RV_IDX, -1    , GV_IDX, RV_IDX,

  751.     RV_IDX, GV_IDX, -1    , RV_IDX, RV_IDX, GV_IDX, -1    , RV_IDX,

  752.     BV_IDX, -1    , GV_IDX, BV_IDX, BV_IDX, -1    , GV_IDX, BV_IDX,

  753.     RY_IDX, BY_IDX, RY_IDX, BY_IDX, RY_IDX, BY_IDX, RY_IDX, BY_IDX,

  754.     BY_IDX, RY_IDX, BY_IDX, RY_IDX, BY_IDX, RY_IDX, BY_IDX, RY_IDX,

  755.     GY_IDX, -1    , GY_IDX, -1    , GY_IDX, -1    , GY_IDX, -1    ,

  756.     -1    , GY_IDX, -1    , GY_IDX, -1    , GY_IDX, -1    , GY_IDX,

  757.     RU_IDX, BU_IDX, RU_IDX, BU_IDX, RU_IDX, BU_IDX, RU_IDX, BU_IDX,

  758.     BU_IDX, RU_IDX, BU_IDX, RU_IDX, BU_IDX, RU_IDX, BU_IDX, RU_IDX,

  759.     GU_IDX, -1    , GU_IDX, -1    , GU_IDX, -1    , GU_IDX, -1    ,

  760.     -1    , GU_IDX, -1    , GU_IDX, -1    , GU_IDX, -1    , GU_IDX,

  761.     RV_IDX, BV_IDX, RV_IDX, BV_IDX, RV_IDX, BV_IDX, RV_IDX, BV_IDX,

  762.     BV_IDX, RV_IDX, BV_IDX, RV_IDX, BV_IDX, RV_IDX, BV_IDX, RV_IDX,

  763.     GV_IDX, -1    , GV_IDX, -1    , GV_IDX, -1    , GV_IDX, -1    ,

  764.     -1    , GV_IDX, -1    , GV_IDX, -1    , GV_IDX, -1    , GV_IDX, //23

  765.     -1    , -1    , -1    , -1    , -1    , -1    , -1    , -1    , //24

  766.     -1    , -1    , -1    , -1    , -1    , -1    , -1    , -1    , //25

  767.     -1    , -1    , -1    , -1    , -1    , -1    , -1    , -1    , //26

  768.     -1    , -1    , -1    , -1    , -1    , -1    , -1    , -1    , //27

  769.     -1    , -1    , -1    , -1    , -1    , -1    , -1    , -1    , //28

  770.     -1    , -1    , -1    , -1    , -1    , -1    , -1    , -1    , //29

  771.     -1    , -1    , -1    , -1    , -1    , -1    , -1    , -1    , //30

  772.     -1    , -1    , -1    , -1    , -1    , -1    , -1    , -1    , //31

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

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

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

  776.     };

  777. 

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

  779. 

  780.     if (!dstRange) {

  781.         cy = cy * 255 / 219;

  782.     } else {

  783.         vr = vr * 224 / 255;

  784.         ub = ub * 224 / 255;

  785.         ug = ug * 224 / 255;

  786.         vg = vg * 224 / 255;

  787.     }

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

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

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

  791. 

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

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

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

  795. 

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

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

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

  799. 

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

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

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

  803. 

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

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

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

  807. 

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

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

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

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

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

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

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

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

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

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

  818.     }

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

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

  821. }

  822. 

  823. static void fill_xyztables(struct SwsContext *c)

  824. {

  825.     int i;

  826.     double xyzgamma = XYZ_GAMMA;

  827.     double rgbgamma = 1.0 / RGB_GAMMA;

  828.     double xyzgammainv = 1.0 / XYZ_GAMMA;

  829.     double rgbgammainv = RGB_GAMMA;

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

  831.         {13270, -6295, -2041},

  832.         {-3969,  7682,   170},

  833.         {  228,  -835,  4329} };

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

  835.         {1689, 1464,  739},

  836.         { 871, 2929,  296},

  837.         {  79,  488, 3891} };

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

  839. 

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

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

  842.     c->xyzgamma = xyzgamma_tab;

  843.     c->rgbgamma = rgbgamma_tab;

  844.     c->xyzgammainv = xyzgammainv_tab;

  845.     c->rgbgammainv = rgbgammainv_tab;

  846. 

  847.     if (rgbgamma_tab[4095])

  848.         return;

  849. 

  850.     /* set gamma vectors */

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

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

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

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

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

  856.     }

  857. }

  858. 

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

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

  861.                              int brightness, int contrast, int saturation)

  862. {

  863.     const AVPixFmtDescriptor *desc_dst;

  864.     const AVPixFmtDescriptor *desc_src;

  865.     int need_reinit = 0;

  866. 

  867.     handle_formats(c);

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

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

  870. 

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

  872.         dstRange = 0;

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

  874.         srcRange = 0;

  875. 

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

  877.         c->dstRange != dstRange ||

  878.         c->brightness != brightness ||

  879.         c->contrast   != contrast ||

  880.         c->saturation != saturation ||

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

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

  883.     )

  884.         need_reinit = 1;

  885. 

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

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

  888. 

  889. 

  890. 

  891.     c->brightness = brightness;

  892.     c->contrast   = contrast;

  893.     c->saturation = saturation;

  894.     c->srcRange   = srcRange;

  895.     c->dstRange   = dstRange;

  896. 

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

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

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

  900.         ff_sws_init_range_convert(c);

  901. 

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

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

  904. 

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

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

  907. 

  908.     if (!need_reinit)

  909.         return 0;

  910. 

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

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

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

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

  915.             enum AVPixelFormat tmp_format;

  916.             int tmp_width, tmp_height;

  917.             int srcW = c->srcW;

  918.             int srcH = c->srcH;

  919.             int dstW = c->dstW;

  920.             int dstH = c->dstH;

  921.             int ret;

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

  923. 

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

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

  926.                     tmp_format = AV_PIX_FMT_BGRA64;

  927.                 } else {

  928.                     tmp_format = AV_PIX_FMT_BGR48;

  929.                 }

  930.             } else {

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

  932.                     tmp_format = AV_PIX_FMT_BGRA;

  933.                 } else {

  934.                     tmp_format = AV_PIX_FMT_BGR24;

  935.                 }

  936.             }

  937. 

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

  939.                 tmp_width  = dstW;

  940.                 tmp_height = dstH;

  941.             } else {

  942.                 tmp_width  = srcW;

  943.                 tmp_height = srcH;

  944.             }

  945. 

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

  947.                                 tmp_width, tmp_height, tmp_format, 64);

  948.             if (ret < 0)

  949.                 return ret;

  950. 

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

  952.                                                         tmp_width, tmp_height, tmp_format,

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

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

  955.                 return -1;

  956. 

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

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

  959.             if (ret < 0)

  960.                 return ret;

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

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

  963.                                      srcRange, table, dstRange,

  964.                                      brightness, contrast, saturation);

  965. 

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

  967.                                                     dstW, dstH, c->dstFormat,

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

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

  970.                 return -1;

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

  972.                                      srcRange, table, dstRange,

  973.                                      0, 1 << 16, 1 << 16);

  974.             return 0;

  975.         }

  976.         return -1;

  977.     }

  978. 

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

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

  981.                                  contrast, saturation);

  982.         // FIXME factorize

  983. 

  984.         if (ARCH_PPC)

  985.             ff_yuv2rgb_init_tables_ppc(c, inv_table, brightness,

  986.                                        contrast, saturation);

  987.     }

  988. 

  989.     fill_rgb2yuv_table(c, table, dstRange);

  990. 

  991.     return 0;

  992. }

  993. 

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

  995.                              int *srcRange, int **table, int *dstRange,

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

  997. {

  998.     if (!c )

  999.         return -1;

  1000. 

  1001.     *inv_table  = c->srcColorspaceTable;

  1002.     *table      = c->dstColorspaceTable;

  1003.     *srcRange   = c->srcRange;

  1004.     *dstRange   = c->dstRange;

  1005.     *brightness = c->brightness;

  1006.     *contrast   = c->contrast;

  1007.     *saturation = c->saturation;

  1008. 

  1009.     return 0;

  1010. }

  1011. 

  1012. static int handle_jpeg(enum AVPixelFormat *format)

  1013. {

  1014.     switch (*format) {

  1015.     case AV_PIX_FMT_YUVJ420P:

  1016.         *format = AV_PIX_FMT_YUV420P;

  1017.         return 1;

  1018.     case AV_PIX_FMT_YUVJ411P:

  1019.         *format = AV_PIX_FMT_YUV411P;

  1020.         return 1;

  1021.     case AV_PIX_FMT_YUVJ422P:

  1022.         *format = AV_PIX_FMT_YUV422P;

  1023.         return 1;

  1024.     case AV_PIX_FMT_YUVJ444P:

  1025.         *format = AV_PIX_FMT_YUV444P;

  1026.         return 1;

  1027.     case AV_PIX_FMT_YUVJ440P:

  1028.         *format = AV_PIX_FMT_YUV440P;

  1029.         return 1;

  1030.     case AV_PIX_FMT_GRAY8:

  1031.     case AV_PIX_FMT_YA8:

  1032.     case AV_PIX_FMT_GRAY9LE:

  1033.     case AV_PIX_FMT_GRAY9BE:

  1034.     case AV_PIX_FMT_GRAY10LE:

  1035.     case AV_PIX_FMT_GRAY10BE:

  1036.     case AV_PIX_FMT_GRAY12LE:

  1037.     case AV_PIX_FMT_GRAY12BE:

  1038.     case AV_PIX_FMT_GRAY14LE:

  1039.     case AV_PIX_FMT_GRAY14BE:

  1040.     case AV_PIX_FMT_GRAY16LE:

  1041.     case AV_PIX_FMT_GRAY16BE:

  1042.     case AV_PIX_FMT_YA16BE:

  1043.     case AV_PIX_FMT_YA16LE:

  1044.         return 1;

  1045.     default:

  1046.         return 0;

  1047.     }

  1048. }

  1049. 

  1050. static int handle_0alpha(enum AVPixelFormat *format)

  1051. {

  1052.     switch (*format) {

  1053.     case AV_PIX_FMT_0BGR    : *format = AV_PIX_FMT_ABGR   ; return 1;

  1054.     case AV_PIX_FMT_BGR0    : *format = AV_PIX_FMT_BGRA   ; return 4;

  1055.     case AV_PIX_FMT_0RGB    : *format = AV_PIX_FMT_ARGB   ; return 1;

  1056.     case AV_PIX_FMT_RGB0    : *format = AV_PIX_FMT_RGBA   ; return 4;

  1057.     default:                                          return 0;

  1058.     }

  1059. }

  1060. 

  1061. static int handle_xyz(enum AVPixelFormat *format)

  1062. {

  1063.     switch (*format) {

  1064.     case AV_PIX_FMT_XYZ12BE : *format = AV_PIX_FMT_RGB48BE; return 1;

  1065.     case AV_PIX_FMT_XYZ12LE : *format = AV_PIX_FMT_RGB48LE; return 1;

  1066.     default:                                                return 0;

  1067.     }

  1068. }

  1069. 

  1070. static void handle_formats(SwsContext *c)

  1071. {

  1072.     c->src0Alpha |= handle_0alpha(&c->srcFormat);

  1073.     c->dst0Alpha |= handle_0alpha(&c->dstFormat);

  1074.     c->srcXYZ    |= handle_xyz(&c->srcFormat);

  1075.     c->dstXYZ    |= handle_xyz(&c->dstFormat);

  1076.     if (c->srcXYZ || c->dstXYZ)

  1077.         fill_xyztables(c);

  1078. }

  1079. 

  1080. SwsContext *sws_alloc_context(void)

  1081. {

  1082.     SwsContext *c = av_mallocz(sizeof(SwsContext));

  1083. 

  1084.     av_assert0(offsetof(SwsContext, redDither) + DITHER32_INT == offsetof(SwsContext, dither32));

  1085. 

  1086.     if (c) {

  1087.         c->av_class = &ff_sws_context_class;

  1088.         av_opt_set_defaults(c);

  1089.     }

  1090. 

  1091.     return c;

  1092. }

  1093. 

  1094. static uint16_t * alloc_gamma_tbl(double e)

  1095. {

  1096.     int i = 0;

  1097.     uint16_t * tbl;

  1098.     tbl = (uint16_t*)av_malloc(sizeof(uint16_t) * 1 << 16);

  1099.     if (!tbl)

  1100.         return NULL;

  1101. 

  1102.     for (i = 0; i < 65536; ++i) {

  1103.         tbl[i] = pow(i / 65535.0, e) * 65535.0;

  1104.     }

  1105.     return tbl;

  1106. }

  1107. 

  1108. static enum AVPixelFormat alphaless_fmt(enum AVPixelFormat fmt)

  1109. {

  1110.     switch(fmt) {

  1111.     case AV_PIX_FMT_ARGB:       return AV_PIX_FMT_RGB24;

  1112.     case AV_PIX_FMT_RGBA:       return AV_PIX_FMT_RGB24;

  1113.     case AV_PIX_FMT_ABGR:       return AV_PIX_FMT_BGR24;

  1114.     case AV_PIX_FMT_BGRA:       return AV_PIX_FMT_BGR24;

  1115.     case AV_PIX_FMT_YA8:        return AV_PIX_FMT_GRAY8;

  1116. 

  1117.     case AV_PIX_FMT_YUVA420P:   return AV_PIX_FMT_YUV420P;

  1118.     case AV_PIX_FMT_YUVA422P:   return AV_PIX_FMT_YUV422P;

  1119.     case AV_PIX_FMT_YUVA444P:           return AV_PIX_FMT_YUV444P;

  1120. 

  1121.     case AV_PIX_FMT_GBRAP:              return AV_PIX_FMT_GBRP;

  1122. 

  1123.     case AV_PIX_FMT_GBRAP10LE:          return AV_PIX_FMT_GBRP10;

  1124.     case AV_PIX_FMT_GBRAP10BE:          return AV_PIX_FMT_GBRP10;

  1125. 

  1126.     case AV_PIX_FMT_GBRAP12LE:          return AV_PIX_FMT_GBRP12;

  1127.     case AV_PIX_FMT_GBRAP12BE:          return AV_PIX_FMT_GBRP12;

  1128. 

  1129.     case AV_PIX_FMT_GBRAP16LE:          return AV_PIX_FMT_GBRP16;

  1130.     case AV_PIX_FMT_GBRAP16BE:          return AV_PIX_FMT_GBRP16;

  1131. 

  1132.     case AV_PIX_FMT_RGBA64LE:   return AV_PIX_FMT_RGB48;

  1133.     case AV_PIX_FMT_RGBA64BE:   return AV_PIX_FMT_RGB48;

  1134.     case AV_PIX_FMT_BGRA64LE:   return AV_PIX_FMT_BGR48;

  1135.     case AV_PIX_FMT_BGRA64BE:   return AV_PIX_FMT_BGR48;

  1136. 

  1137.     case AV_PIX_FMT_YA16BE:             return AV_PIX_FMT_GRAY16;

  1138.     case AV_PIX_FMT_YA16LE:             return AV_PIX_FMT_GRAY16;

  1139. 

  1140.     case AV_PIX_FMT_YUVA420P9BE:        return AV_PIX_FMT_YUV420P9;

  1141.     case AV_PIX_FMT_YUVA422P9BE:        return AV_PIX_FMT_YUV422P9;

  1142.     case AV_PIX_FMT_YUVA444P9BE:        return AV_PIX_FMT_YUV444P9;

  1143.     case AV_PIX_FMT_YUVA420P9LE:        return AV_PIX_FMT_YUV420P9;

  1144.     case AV_PIX_FMT_YUVA422P9LE:        return AV_PIX_FMT_YUV422P9;

  1145.     case AV_PIX_FMT_YUVA444P9LE:        return AV_PIX_FMT_YUV444P9;

  1146.     case AV_PIX_FMT_YUVA420P10BE:       return AV_PIX_FMT_YUV420P10;

  1147.     case AV_PIX_FMT_YUVA422P10BE:       return AV_PIX_FMT_YUV422P10;

  1148.     case AV_PIX_FMT_YUVA444P10BE:       return AV_PIX_FMT_YUV444P10;

  1149.     case AV_PIX_FMT_YUVA420P10LE:       return AV_PIX_FMT_YUV420P10;

  1150.     case AV_PIX_FMT_YUVA422P10LE:       return AV_PIX_FMT_YUV422P10;

  1151.     case AV_PIX_FMT_YUVA444P10LE:       return AV_PIX_FMT_YUV444P10;

  1152.     case AV_PIX_FMT_YUVA420P16BE:       return AV_PIX_FMT_YUV420P16;

  1153.     case AV_PIX_FMT_YUVA422P16BE:       return AV_PIX_FMT_YUV422P16;

  1154.     case AV_PIX_FMT_YUVA444P16BE:       return AV_PIX_FMT_YUV444P16;

  1155.     case AV_PIX_FMT_YUVA420P16LE:       return AV_PIX_FMT_YUV420P16;

  1156.     case AV_PIX_FMT_YUVA422P16LE:       return AV_PIX_FMT_YUV422P16;

  1157.     case AV_PIX_FMT_YUVA444P16LE:       return AV_PIX_FMT_YUV444P16;

  1158. 

  1159. //     case AV_PIX_FMT_AYUV64LE:

  1160. //     case AV_PIX_FMT_AYUV64BE:

  1161. //     case AV_PIX_FMT_PAL8:

  1162.     default: return AV_PIX_FMT_NONE;

  1163.     }

  1164. }

  1165. 

  1166. av_cold int sws_init_context(SwsContext *c, SwsFilter *srcFilter,

  1167.                              SwsFilter *dstFilter)

  1168. {

  1169.     int i;

  1170.     int usesVFilter, usesHFilter;

  1171.     int unscaled;

  1172.     SwsFilter dummyFilter = { NULL, NULL, NULL, NULL };

  1173.     int srcW              = c->srcW;

  1174.     int srcH              = c->srcH;

  1175.     int dstW              = c->dstW;

  1176.     int dstH              = c->dstH;

  1177.     int dst_stride        = FFALIGN(dstW * sizeof(int16_t) + 66, 16);

  1178.     int flags, cpu_flags;

  1179.     enum AVPixelFormat srcFormat = c->srcFormat;

  1180.     enum AVPixelFormat dstFormat = c->dstFormat;

  1181.     const AVPixFmtDescriptor *desc_src;

  1182.     const AVPixFmtDescriptor *desc_dst;

  1183.     int ret = 0;

  1184.     enum AVPixelFormat tmpFmt;

  1185.     static const float float_mult = 1.0f / 255.0f;

  1186. 

  1187.     cpu_flags = av_get_cpu_flags();

  1188.     flags     = c->flags;

  1189.     emms_c();

  1190.     if (!rgb15to16)

  1191.         ff_sws_rgb2rgb_init();

  1192. 

  1193.     unscaled = (srcW == dstW && srcH == dstH);

  1194. 

  1195.     c->srcRange |= handle_jpeg(&c->srcFormat);

  1196.     c->dstRange |= handle_jpeg(&c->dstFormat);

  1197. 

  1198.     if(srcFormat!=c->srcFormat || dstFormat!=c->dstFormat)

  1199.         av_log(c, AV_LOG_WARNING, "deprecated pixel format used, make sure you did set range correctly\n");

  1200. 

  1201.     if (!c->contrast && !c->saturation && !c->dstFormatBpp)

  1202.         sws_setColorspaceDetails(c, ff_yuv2rgb_coeffs[SWS_CS_DEFAULT], c->srcRange,

  1203.                                  ff_yuv2rgb_coeffs[SWS_CS_DEFAULT],

  1204.                                  c->dstRange, 0, 1 << 16, 1 << 16);

  1205. 

  1206.     handle_formats(c);

  1207.     srcFormat = c->srcFormat;

  1208.     dstFormat = c->dstFormat;

  1209.     desc_src = av_pix_fmt_desc_get(srcFormat);

  1210.     desc_dst = av_pix_fmt_desc_get(dstFormat);

  1211. 

  1212.     // If the source has no alpha then disable alpha blendaway

  1213.     if (c->src0Alpha)

  1214.         c->alphablend = SWS_ALPHA_BLEND_NONE;

  1215. 

  1216.     if (!(unscaled && sws_isSupportedEndiannessConversion(srcFormat) &&

  1217.           av_pix_fmt_swap_endianness(srcFormat) == dstFormat)) {

  1218.     if (!sws_isSupportedInput(srcFormat)) {

  1219.         av_log(c, AV_LOG_ERROR, "%s is not supported as input pixel format\n",

  1220.                av_get_pix_fmt_name(srcFormat));

  1221.         return AVERROR(EINVAL);

  1222.     }

  1223.     if (!sws_isSupportedOutput(dstFormat)) {

  1224.         av_log(c, AV_LOG_ERROR, "%s is not supported as output pixel format\n",

  1225.                av_get_pix_fmt_name(dstFormat));

  1226.         return AVERROR(EINVAL);

  1227.     }

  1228.     }

  1229.     av_assert2(desc_src && desc_dst);

  1230. 

  1231.     i = flags & (SWS_POINT         |

  1232.                  SWS_AREA          |

  1233.                  SWS_BILINEAR      |

  1234.                  SWS_FAST_BILINEAR |

  1235.                  SWS_BICUBIC       |

  1236.                  SWS_X             |

  1237.                  SWS_GAUSS         |

  1238.                  SWS_LANCZOS       |

  1239.                  SWS_SINC          |

  1240.                  SWS_SPLINE        |

  1241.                  SWS_BICUBLIN);

  1242. 

  1243.     /* provide a default scaler if not set by caller */

  1244.     if (!i) {

  1245.         if (dstW < srcW && dstH < srcH)

  1246.             flags |= SWS_BICUBIC;

  1247.         else if (dstW > srcW && dstH > srcH)

  1248.             flags |= SWS_BICUBIC;

  1249.         else

  1250.             flags |= SWS_BICUBIC;

  1251.         c->flags = flags;

  1252.     } else if (i & (i - 1)) {

  1253.         av_log(c, AV_LOG_ERROR,

  1254.                "Exactly one scaler algorithm must be chosen, got %X\n", i);

  1255.         return AVERROR(EINVAL);

  1256.     }

  1257.     /* sanity check */

  1258.     if (srcW < 1 || srcH < 1 || dstW < 1 || dstH < 1) {

  1259.         /* FIXME check if these are enough and try to lower them after

  1260.          * fixing the relevant parts of the code */

  1261.         av_log(c, AV_LOG_ERROR, "%dx%d -> %dx%d is invalid scaling dimension\n",

  1262.                srcW, srcH, dstW, dstH);

  1263.         return AVERROR(EINVAL);

  1264.     }

  1265.     if (flags & SWS_FAST_BILINEAR) {

  1266.         if (srcW < 8 || dstW < 8) {

  1267.             flags ^= SWS_FAST_BILINEAR | SWS_BILINEAR;

  1268.             c->flags = flags;

  1269.         }

  1270.     }

  1271. 

  1272.     if (!dstFilter)

  1273.         dstFilter = &dummyFilter;

  1274.     if (!srcFilter)

  1275.         srcFilter = &dummyFilter;

  1276. 

  1277.     c->lumXInc      = (((int64_t)srcW << 16) + (dstW >> 1)) / dstW;

  1278.     c->lumYInc      = (((int64_t)srcH << 16) + (dstH >> 1)) / dstH;

  1279.     c->dstFormatBpp = av_get_bits_per_pixel(desc_dst);

  1280.     c->srcFormatBpp = av_get_bits_per_pixel(desc_src);

  1281.     c->vRounder     = 4 * 0x0001000100010001ULL;

  1282. 

  1283.     usesVFilter = (srcFilter->lumV && srcFilter->lumV->length > 1) ||

  1284.                   (srcFilter->chrV && srcFilter->chrV->length > 1) ||

  1285.                   (dstFilter->lumV && dstFilter->lumV->length > 1) ||

  1286.                   (dstFilter->chrV && dstFilter->chrV->length > 1);

  1287.     usesHFilter = (srcFilter->lumH && srcFilter->lumH->length > 1) ||

  1288.                   (srcFilter->chrH && srcFilter->chrH->length > 1) ||

  1289.                   (dstFilter->lumH && dstFilter->lumH->length > 1) ||

  1290.                   (dstFilter->chrH && dstFilter->chrH->length > 1);

  1291. 

  1292.     av_pix_fmt_get_chroma_sub_sample(srcFormat, &c->chrSrcHSubSample, &c->chrSrcVSubSample);

  1293.     av_pix_fmt_get_chroma_sub_sample(dstFormat, &c->chrDstHSubSample, &c->chrDstVSubSample);

  1294. 

  1295.     if (isAnyRGB(dstFormat) && !(flags&SWS_FULL_CHR_H_INT)) {

  1296.         if (dstW&1) {

  1297.             av_log(c, AV_LOG_DEBUG, "Forcing full internal H chroma due to odd output size\n");

  1298.             flags |= SWS_FULL_CHR_H_INT;

  1299.             c->flags = flags;

  1300.         }

  1301. 

  1302.         if (   c->chrSrcHSubSample == 0

  1303.             && c->chrSrcVSubSample == 0

  1304.             && c->dither != SWS_DITHER_BAYER //SWS_FULL_CHR_H_INT is currently not supported with SWS_DITHER_BAYER

  1305.             && !(c->flags & SWS_FAST_BILINEAR)

  1306.         ) {

  1307.             av_log(c, AV_LOG_DEBUG, "Forcing full internal H chroma due to input having non subsampled chroma\n");

  1308.             flags |= SWS_FULL_CHR_H_INT;

  1309.             c->flags = flags;

  1310.         }

  1311.     }

  1312. 

  1313.     if (c->dither == SWS_DITHER_AUTO) {

  1314.         if (flags & SWS_ERROR_DIFFUSION)

  1315.             c->dither = SWS_DITHER_ED;

  1316.     }

  1317. 

  1318.     if(dstFormat == AV_PIX_FMT_BGR4_BYTE ||

  1319.        dstFormat == AV_PIX_FMT_RGB4_BYTE ||

  1320.        dstFormat == AV_PIX_FMT_BGR8 ||

  1321.        dstFormat == AV_PIX_FMT_RGB8) {

  1322.         if (c->dither == SWS_DITHER_AUTO)

  1323.             c->dither = (flags & SWS_FULL_CHR_H_INT) ? SWS_DITHER_ED : SWS_DITHER_BAYER;

  1324.         if (!(flags & SWS_FULL_CHR_H_INT)) {

  1325.             if (c->dither == SWS_DITHER_ED || c->dither == SWS_DITHER_A_DITHER || c->dither == SWS_DITHER_X_DITHER) {

  1326.                 av_log(c, AV_LOG_DEBUG,

  1327.                     "Desired dithering only supported in full chroma interpolation for destination format '%s'\n",

  1328.                     av_get_pix_fmt_name(dstFormat));

  1329.                 flags   |= SWS_FULL_CHR_H_INT;

  1330.                 c->flags = flags;

  1331.             }

  1332.         }

  1333.         if (flags & SWS_FULL_CHR_H_INT) {

  1334.             if (c->dither == SWS_DITHER_BAYER) {

  1335.                 av_log(c, AV_LOG_DEBUG,

  1336.                     "Ordered dither is not supported in full chroma interpolation for destination format '%s'\n",

  1337.                     av_get_pix_fmt_name(dstFormat));

  1338.                 c->dither = SWS_DITHER_ED;

  1339.             }

  1340.         }

  1341.     }

  1342.     if (isPlanarRGB(dstFormat)) {

  1343.         if (!(flags & SWS_FULL_CHR_H_INT)) {

  1344.             av_log(c, AV_LOG_DEBUG,

  1345.                    "%s output is not supported with half chroma resolution, switching to full\n",

  1346.                    av_get_pix_fmt_name(dstFormat));

  1347.             flags   |= SWS_FULL_CHR_H_INT;

  1348.             c->flags = flags;

  1349.         }

  1350.     }

  1351. 

  1352.     /* reuse chroma for 2 pixels RGB/BGR unless user wants full

  1353.      * chroma interpolation */

  1354.     if (flags & SWS_FULL_CHR_H_INT &&

  1355.         isAnyRGB(dstFormat)        &&

  1356.         !isPlanarRGB(dstFormat)    &&

  1357.         dstFormat != AV_PIX_FMT_RGBA64LE &&

  1358.         dstFormat != AV_PIX_FMT_RGBA64BE &&

  1359.         dstFormat != AV_PIX_FMT_BGRA64LE &&

  1360.         dstFormat != AV_PIX_FMT_BGRA64BE &&

  1361.         dstFormat != AV_PIX_FMT_RGB48LE &&

  1362.         dstFormat != AV_PIX_FMT_RGB48BE &&

  1363.         dstFormat != AV_PIX_FMT_BGR48LE &&

  1364.         dstFormat != AV_PIX_FMT_BGR48BE &&

  1365.         dstFormat != AV_PIX_FMT_RGBA  &&

  1366.         dstFormat != AV_PIX_FMT_ARGB  &&

  1367.         dstFormat != AV_PIX_FMT_BGRA  &&

  1368.         dstFormat != AV_PIX_FMT_ABGR  &&

  1369.         dstFormat != AV_PIX_FMT_RGB24 &&

  1370.         dstFormat != AV_PIX_FMT_BGR24 &&

  1371.         dstFormat != AV_PIX_FMT_BGR4_BYTE &&

  1372.         dstFormat != AV_PIX_FMT_RGB4_BYTE &&

  1373.         dstFormat != AV_PIX_FMT_BGR8 &&

  1374.         dstFormat != AV_PIX_FMT_RGB8

  1375.     ) {

  1376.         av_log(c, AV_LOG_WARNING,

  1377.                "full chroma interpolation for destination format '%s' not yet implemented\n",

  1378.                av_get_pix_fmt_name(dstFormat));

  1379.         flags   &= ~SWS_FULL_CHR_H_INT;

  1380.         c->flags = flags;

  1381.     }

  1382.     if (isAnyRGB(dstFormat) && !(flags & SWS_FULL_CHR_H_INT))

  1383.         c->chrDstHSubSample = 1;

  1384. 

  1385.     // drop some chroma lines if the user wants it

  1386.     c->vChrDrop          = (flags & SWS_SRC_V_CHR_DROP_MASK) >>

  1387.                            SWS_SRC_V_CHR_DROP_SHIFT;

  1388.     c->chrSrcVSubSample += c->vChrDrop;

  1389. 

  1390.     /* drop every other pixel for chroma calculation unless user

  1391.      * wants full chroma */

  1392.     if (isAnyRGB(srcFormat) && !(flags & SWS_FULL_CHR_H_INP)   &&

  1393.         srcFormat != AV_PIX_FMT_RGB8 && srcFormat != AV_PIX_FMT_BGR8 &&

  1394.         srcFormat != AV_PIX_FMT_RGB4 && srcFormat != AV_PIX_FMT_BGR4 &&

  1395.         srcFormat != AV_PIX_FMT_RGB4_BYTE && srcFormat != AV_PIX_FMT_BGR4_BYTE &&

  1396.         srcFormat != AV_PIX_FMT_GBRP9BE   && srcFormat != AV_PIX_FMT_GBRP9LE  &&

  1397.         srcFormat != AV_PIX_FMT_GBRP10BE  && srcFormat != AV_PIX_FMT_GBRP10LE &&

  1398.         srcFormat != AV_PIX_FMT_GBRAP10BE && srcFormat != AV_PIX_FMT_GBRAP10LE &&

  1399.         srcFormat != AV_PIX_FMT_GBRP12BE  && srcFormat != AV_PIX_FMT_GBRP12LE &&

  1400.         srcFormat != AV_PIX_FMT_GBRAP12BE && srcFormat != AV_PIX_FMT_GBRAP12LE &&

  1401.         srcFormat != AV_PIX_FMT_GBRP14BE  && srcFormat != AV_PIX_FMT_GBRP14LE &&

  1402.         srcFormat != AV_PIX_FMT_GBRP16BE  && srcFormat != AV_PIX_FMT_GBRP16LE &&

  1403.         srcFormat != AV_PIX_FMT_GBRAP16BE  && srcFormat != AV_PIX_FMT_GBRAP16LE &&

  1404.         ((dstW >> c->chrDstHSubSample) <= (srcW >> 1) ||

  1405.          (flags & SWS_FAST_BILINEAR)))

  1406.         c->chrSrcHSubSample = 1;

  1407. 

  1408.     // Note the AV_CEIL_RSHIFT is so that we always round toward +inf.

  1409.     c->chrSrcW = AV_CEIL_RSHIFT(srcW, c->chrSrcHSubSample);

  1410.     c->chrSrcH = AV_CEIL_RSHIFT(srcH, c->chrSrcVSubSample);

  1411.     c->chrDstW = AV_CEIL_RSHIFT(dstW, c->chrDstHSubSample);

  1412.     c->chrDstH = AV_CEIL_RSHIFT(dstH, c->chrDstVSubSample);

  1413. 

  1414.     FF_ALLOCZ_OR_GOTO(c, c->formatConvBuffer, FFALIGN(srcW*2+78, 16) * 2, fail);

  1415. 

  1416.     c->srcBpc = desc_src->comp[0].depth;

  1417.     if (c->srcBpc < 8)

  1418.         c->srcBpc = 8;

  1419.     c->dstBpc = desc_dst->comp[0].depth;

  1420.     if (c->dstBpc < 8)

  1421.         c->dstBpc = 8;

  1422.     if (isAnyRGB(srcFormat) || srcFormat == AV_PIX_FMT_PAL8)

  1423.         c->srcBpc = 16;

  1424.     if (c->dstBpc == 16)

  1425.         dst_stride <<= 1;

  1426. 

  1427.     if (INLINE_MMXEXT(cpu_flags) && c->srcBpc == 8 && c->dstBpc <= 14) {

  1428.         c->canMMXEXTBeUsed = dstW >= srcW && (dstW & 31) == 0 &&

  1429.                              c->chrDstW >= c->chrSrcW &&

  1430.                              (srcW & 15) == 0;

  1431.         if (!c->canMMXEXTBeUsed && dstW >= srcW && c->chrDstW >= c->chrSrcW && (srcW & 15) == 0

  1432. 

  1433.             && (flags & SWS_FAST_BILINEAR)) {

  1434.             if (flags & SWS_PRINT_INFO)

  1435.                 av_log(c, AV_LOG_INFO,

  1436.                        "output width is not a multiple of 32 -> no MMXEXT scaler\n");

  1437.         }

  1438.         if (usesHFilter || isNBPS(c->srcFormat) || is16BPS(c->srcFormat) || isAnyRGB(c->srcFormat))

  1439.             c->canMMXEXTBeUsed = 0;

  1440.     } else

  1441.         c->canMMXEXTBeUsed = 0;

  1442. 

  1443.     c->chrXInc = (((int64_t)c->chrSrcW << 16) + (c->chrDstW >> 1)) / c->chrDstW;

  1444.     c->chrYInc = (((int64_t)c->chrSrcH << 16) + (c->chrDstH >> 1)) / c->chrDstH;

  1445. 

  1446.     /* Match pixel 0 of the src to pixel 0 of dst and match pixel n-2 of src

  1447.      * to pixel n-2 of dst, but only for the FAST_BILINEAR mode otherwise do

  1448.      * correct scaling.

  1449.      * n-2 is the last chrominance sample available.

  1450.      * This is not perfect, but no one should notice the difference, the more

  1451.      * correct variant would be like the vertical one, but that would require

  1452.      * some special code for the first and last pixel */

  1453.     if (flags & SWS_FAST_BILINEAR) {

  1454.         if (c->canMMXEXTBeUsed) {

  1455.             c->lumXInc += 20;

  1456.             c->chrXInc += 20;

  1457.         }

  1458.         // we don't use the x86 asm scaler if MMX is available

  1459.         else if (INLINE_MMX(cpu_flags) && c->dstBpc <= 14) {

  1460.             c->lumXInc = ((int64_t)(srcW       - 2) << 16) / (dstW       - 2) - 20;

  1461.             c->chrXInc = ((int64_t)(c->chrSrcW - 2) << 16) / (c->chrDstW - 2) - 20;

  1462.         }

  1463.     }

  1464. 

  1465.     // hardcoded for now

  1466.     c->gamma_value = 2.2;

  1467.     tmpFmt = AV_PIX_FMT_RGBA64LE;

  1468. 

  1469. 

  1470.     if (!unscaled && c->gamma_flag && (srcFormat != tmpFmt || dstFormat != tmpFmt)) {

  1471.         SwsContext *c2;

  1472.         c->cascaded_context[0] = NULL;

  1473. 

  1474.         ret = av_image_alloc(c->cascaded_tmp, c->cascaded_tmpStride,

  1475.                             srcW, srcH, tmpFmt, 64);

  1476.         if (ret < 0)

  1477.             return ret;

  1478. 

  1479.         c->cascaded_context[0] = sws_getContext(srcW, srcH, srcFormat,

  1480.                                                 srcW, srcH, tmpFmt,

  1481.                                                 flags, NULL, NULL, c->param);

  1482.         if (!c->cascaded_context[0]) {

  1483.             return -1;

  1484.         }

  1485. 

  1486.         c->cascaded_context[1] = sws_getContext(srcW, srcH, tmpFmt,

  1487.                                                 dstW, dstH, tmpFmt,

  1488.                                                 flags, srcFilter, dstFilter, c->param);

  1489. 

  1490.         if (!c->cascaded_context[1])

  1491.             return -1;

  1492. 

  1493.         c2 = c->cascaded_context[1];

  1494.         c2->is_internal_gamma = 1;

  1495.         c2->gamma     = alloc_gamma_tbl(    c->gamma_value);

  1496.         c2->inv_gamma = alloc_gamma_tbl(1.f/c->gamma_value);

  1497.         if (!c2->gamma || !c2->inv_gamma)

  1498.             return AVERROR(ENOMEM);

  1499. 

  1500.         // is_internal_flag is set after creating the context

  1501.         // to properly create the gamma convert FilterDescriptor

  1502.         // we have to re-initialize it

  1503.         ff_free_filters(c2);

  1504.         if (ff_init_filters(c2) < 0) {

  1505.             sws_freeContext(c2);

  1506.             c->cascaded_context[1] = NULL;

  1507.             return -1;

  1508.         }

  1509. 

  1510.         c->cascaded_context[2] = NULL;

  1511.         if (dstFormat != tmpFmt) {

  1512.             ret = av_image_alloc(c->cascaded1_tmp, c->cascaded1_tmpStride,

  1513.                                 dstW, dstH, tmpFmt, 64);

  1514.             if (ret < 0)

  1515.                 return ret;

  1516. 

  1517.             c->cascaded_context[2] = sws_getContext(dstW, dstH, tmpFmt,

  1518.                                                 dstW, dstH, dstFormat,

  1519.                                                 flags, NULL, NULL, c->param);

  1520.             if (!c->cascaded_context[2])

  1521.                 return -1;

  1522.         }

  1523.         return 0;

  1524.     }

  1525. 

  1526.     if (isBayer(srcFormat)) {

  1527.         if (!unscaled ||

  1528.             (dstFormat != AV_PIX_FMT_RGB24 && dstFormat != AV_PIX_FMT_YUV420P)) {

  1529.             enum AVPixelFormat tmpFormat = AV_PIX_FMT_RGB24;

  1530. 

  1531.             ret = av_image_alloc(c->cascaded_tmp, c->cascaded_tmpStride,

  1532.                                 srcW, srcH, tmpFormat, 64);

  1533.             if (ret < 0)

  1534.                 return ret;

  1535. 

  1536.             c->cascaded_context[0] = sws_getContext(srcW, srcH, srcFormat,

  1537.                                                     srcW, srcH, tmpFormat,

  1538.                                                     flags, srcFilter, NULL, c->param);

  1539.             if (!c->cascaded_context[0])

  1540.                 return -1;

  1541. 

  1542.             c->cascaded_context[1] = sws_getContext(srcW, srcH, tmpFormat,

  1543.                                                     dstW, dstH, dstFormat,

  1544.                                                     flags, NULL, dstFilter, c->param);

  1545.             if (!c->cascaded_context[1])

  1546.                 return -1;

  1547.             return 0;

  1548.         }

  1549.     }

  1550. 

  1551.     if (unscaled && c->srcBpc == 8 && dstFormat == AV_PIX_FMT_GRAYF32){

  1552.         for (i = 0; i < 256; ++i){

  1553.             c->uint2float_lut[i] = (float)i * float_mult;

  1554.         }

  1555.     }

  1556. 

  1557.     // float will be converted to uint16_t

  1558.     if ((srcFormat == AV_PIX_FMT_GRAYF32BE || srcFormat == AV_PIX_FMT_GRAYF32LE) &&

  1559.         (!unscaled || unscaled && dstFormat != srcFormat && (srcFormat != AV_PIX_FMT_GRAYF32 ||

  1560.         dstFormat != AV_PIX_FMT_GRAY8))){

  1561.         c->srcBpc = 16;

  1562.     }

  1563. 

  1564.     if (CONFIG_SWSCALE_ALPHA && isALPHA(srcFormat) && !isALPHA(dstFormat)) {

  1565.         enum AVPixelFormat tmpFormat = alphaless_fmt(srcFormat);

  1566. 

  1567.         if (tmpFormat != AV_PIX_FMT_NONE && c->alphablend != SWS_ALPHA_BLEND_NONE)

  1568.         if (!unscaled ||

  1569.             dstFormat != tmpFormat ||

  1570.             usesHFilter || usesVFilter ||

  1571.             c->srcRange != c->dstRange

  1572.         ) {

  1573.             c->cascaded_mainindex = 1;

  1574.             ret = av_image_alloc(c->cascaded_tmp, c->cascaded_tmpStride,

  1575.                                 srcW, srcH, tmpFormat, 64);

  1576.             if (ret < 0)

  1577.                 return ret;

  1578. 

  1579.             c->cascaded_context[0] = sws_alloc_set_opts(srcW, srcH, srcFormat,

  1580.                                                         srcW, srcH, tmpFormat,

  1581.                                                         flags, c->param);

  1582.             if (!c->cascaded_context[0])

  1583.                 return -1;

  1584.             c->cascaded_context[0]->alphablend = c->alphablend;

  1585.             ret = sws_init_context(c->cascaded_context[0], NULL , NULL);

  1586.             if (ret < 0)

  1587.                 return ret;

  1588. 

  1589.             c->cascaded_context[1] = sws_alloc_set_opts(srcW, srcH, tmpFormat,

  1590.                                                         dstW, dstH, dstFormat,

  1591.                                                         flags, c->param);

  1592.             if (!c->cascaded_context[1])

  1593.                 return -1;

  1594. 

  1595.             c->cascaded_context[1]->srcRange = c->srcRange;

  1596.             c->cascaded_context[1]->dstRange = c->dstRange;

  1597.             ret = sws_init_context(c->cascaded_context[1], srcFilter , dstFilter);

  1598.             if (ret < 0)

  1599.                 return ret;

  1600. 

  1601.             return 0;

  1602.         }

  1603.     }

  1604. 

  1605. #if HAVE_MMAP && HAVE_MPROTECT && defined(MAP_ANONYMOUS)

  1606. #define USE_MMAP 1

  1607. #else

  1608. #define USE_MMAP 0

  1609. #endif

  1610. 

  1611.     /* precalculate horizontal scaler filter coefficients */

  1612.     {

  1613. #if HAVE_MMXEXT_INLINE

  1614. // can't downscale !!!

  1615.         if (c->canMMXEXTBeUsed && (flags & SWS_FAST_BILINEAR)) {

  1616.             c->lumMmxextFilterCodeSize = ff_init_hscaler_mmxext(dstW, c->lumXInc, NULL,

  1617.                                                              NULL, NULL, 8);

  1618.             c->chrMmxextFilterCodeSize = ff_init_hscaler_mmxext(c->chrDstW, c->chrXInc,

  1619.                                                              NULL, NULL, NULL, 4);

  1620. 

  1621. #if USE_MMAP

  1622.             c->lumMmxextFilterCode = mmap(NULL, c->lumMmxextFilterCodeSize,

  1623.                                           PROT_READ | PROT_WRITE,

  1624.                                           MAP_PRIVATE | MAP_ANONYMOUS,

  1625.                                           -1, 0);

  1626.             c->chrMmxextFilterCode = mmap(NULL, c->chrMmxextFilterCodeSize,

  1627.                                           PROT_READ | PROT_WRITE,

  1628.                                           MAP_PRIVATE | MAP_ANONYMOUS,

  1629.                                           -1, 0);

  1630. #elif HAVE_VIRTUALALLOC

  1631.             c->lumMmxextFilterCode = VirtualAlloc(NULL,

  1632.                                                   c->lumMmxextFilterCodeSize,

  1633.                                                   MEM_COMMIT,

  1634.                                                   PAGE_EXECUTE_READWRITE);

  1635.             c->chrMmxextFilterCode = VirtualAlloc(NULL,

  1636.                                                   c->chrMmxextFilterCodeSize,

  1637.                                                   MEM_COMMIT,

  1638.                                                   PAGE_EXECUTE_READWRITE);

  1639. #else

  1640.             c->lumMmxextFilterCode = av_malloc(c->lumMmxextFilterCodeSize);

  1641.             c->chrMmxextFilterCode = av_malloc(c->chrMmxextFilterCodeSize);

  1642. #endif

  1643. 

  1644. #ifdef MAP_ANONYMOUS

  1645.             if (c->lumMmxextFilterCode == MAP_FAILED || c->chrMmxextFilterCode == MAP_FAILED)

  1646. #else

  1647.             if (!c->lumMmxextFilterCode || !c->chrMmxextFilterCode)

  1648. #endif

  1649.             {

  1650.                 av_log(c, AV_LOG_ERROR, "Failed to allocate MMX2FilterCode\n");

  1651.                 return AVERROR(ENOMEM);

  1652.             }

  1653. 

  1654.             FF_ALLOCZ_OR_GOTO(c, c->hLumFilter,    (dstW           / 8 + 8) * sizeof(int16_t), fail);

  1655.             FF_ALLOCZ_OR_GOTO(c, c->hChrFilter,    (c->chrDstW     / 4 + 8) * sizeof(int16_t), fail);

  1656.             FF_ALLOCZ_OR_GOTO(c, c->hLumFilterPos, (dstW       / 2 / 8 + 8) * sizeof(int32_t), fail);

  1657.             FF_ALLOCZ_OR_GOTO(c, c->hChrFilterPos, (c->chrDstW / 2 / 4 + 8) * sizeof(int32_t), fail);

  1658. 

  1659.             ff_init_hscaler_mmxext(      dstW, c->lumXInc, c->lumMmxextFilterCode,

  1660.                                 c->hLumFilter, (uint32_t*)c->hLumFilterPos, 8);

  1661.             ff_init_hscaler_mmxext(c->chrDstW, c->chrXInc, c->chrMmxextFilterCode,

  1662.                                 c->hChrFilter, (uint32_t*)c->hChrFilterPos, 4);

  1663. 

  1664. #if USE_MMAP

  1665.             if (   mprotect(c->lumMmxextFilterCode, c->lumMmxextFilterCodeSize, PROT_EXEC | PROT_READ) == -1

  1666.                 || mprotect(c->chrMmxextFilterCode, c->chrMmxextFilterCodeSize, PROT_EXEC | PROT_READ) == -1) {

  1667.                 av_log(c, AV_LOG_ERROR, "mprotect failed, cannot use fast bilinear scaler\n");

  1668.                 goto fail;

  1669.             }

  1670. #endif

  1671.         } else

  1672. #endif /* HAVE_MMXEXT_INLINE */

  1673.         {

  1674.             const int filterAlign = X86_MMX(cpu_flags)     ? 4 :

  1675.                                     PPC_ALTIVEC(cpu_flags) ? 8 :

  1676.                                     have_neon(cpu_flags)   ? 8 : 1;

  1677. 

  1678.             if ((ret = initFilter(&c->hLumFilter, &c->hLumFilterPos,

  1679.                            &c->hLumFilterSize, c->lumXInc,

  1680.                            srcW, dstW, filterAlign, 1 << 14,

  1681.                            (flags & SWS_BICUBLIN) ? (flags | SWS_BICUBIC) : flags,

  1682.                            cpu_flags, srcFilter->lumH, dstFilter->lumH,

  1683.                            c->param,

  1684.                            get_local_pos(c, 0, 0, 0),

  1685.                            get_local_pos(c, 0, 0, 0))) < 0)

  1686.                 goto fail;

  1687.             if ((ret = initFilter(&c->hChrFilter, &c->hChrFilterPos,

  1688.                            &c->hChrFilterSize, c->chrXInc,

  1689.                            c->chrSrcW, c->chrDstW, filterAlign, 1 << 14,

  1690.                            (flags & SWS_BICUBLIN) ? (flags | SWS_BILINEAR) : flags,

  1691.                            cpu_flags, srcFilter->chrH, dstFilter->chrH,

  1692.                            c->param,

  1693.                            get_local_pos(c, c->chrSrcHSubSample, c->src_h_chr_pos, 0),

  1694.                            get_local_pos(c, c->chrDstHSubSample, c->dst_h_chr_pos, 0))) < 0)

  1695.                 goto fail;

  1696.         }

  1697.     } // initialize horizontal stuff

  1698. 

  1699.     /* precalculate vertical scaler filter coefficients */

  1700.     {

  1701.         const int filterAlign = X86_MMX(cpu_flags)     ? 2 :

  1702.                                 PPC_ALTIVEC(cpu_flags) ? 8 :

  1703.                                 have_neon(cpu_flags)   ? 2 : 1;

  1704. 

  1705.         if ((ret = initFilter(&c->vLumFilter, &c->vLumFilterPos, &c->vLumFilterSize,

  1706.                        c->lumYInc, srcH, dstH, filterAlign, (1 << 12),

  1707.                        (flags & SWS_BICUBLIN) ? (flags | SWS_BICUBIC) : flags,

  1708.                        cpu_flags, srcFilter->lumV, dstFilter->lumV,

  1709.                        c->param,

  1710.                        get_local_pos(c, 0, 0, 1),

  1711.                        get_local_pos(c, 0, 0, 1))) < 0)

  1712.             goto fail;

  1713.         if ((ret = initFilter(&c->vChrFilter, &c->vChrFilterPos, &c->vChrFilterSize,

  1714.                        c->chrYInc, c->chrSrcH, c->chrDstH,

  1715.                        filterAlign, (1 << 12),

  1716.                        (flags & SWS_BICUBLIN) ? (flags | SWS_BILINEAR) : flags,

  1717.                        cpu_flags, srcFilter->chrV, dstFilter->chrV,

  1718.                        c->param,

  1719.                        get_local_pos(c, c->chrSrcVSubSample, c->src_v_chr_pos, 1),

  1720.                        get_local_pos(c, c->chrDstVSubSample, c->dst_v_chr_pos, 1))) < 0)

  1721. 

  1722.             goto fail;

  1723. 

  1724. #if HAVE_ALTIVEC

  1725.         FF_ALLOC_OR_GOTO(c, c->vYCoeffsBank, sizeof(vector signed short) * c->vLumFilterSize * c->dstH,    fail);

  1726.         FF_ALLOC_OR_GOTO(c, c->vCCoeffsBank, sizeof(vector signed short) * c->vChrFilterSize * c->chrDstH, fail);

  1727. 

  1728.         for (i = 0; i < c->vLumFilterSize * c->dstH; i++) {

  1729.             int j;

  1730.             short *p = (short *)&c->vYCoeffsBank[i];

  1731.             for (j = 0; j < 8; j++)

  1732.                 p[j] = c->vLumFilter[i];

  1733.         }

  1734. 

  1735.         for (i = 0; i < c->vChrFilterSize * c->chrDstH; i++) {

  1736.             int j;

  1737.             short *p = (short *)&c->vCCoeffsBank[i];

  1738.             for (j = 0; j < 8; j++)

  1739.                 p[j] = c->vChrFilter[i];

  1740.         }

  1741. #endif

  1742.     }

  1743. 

  1744.     for (i = 0; i < 4; i++)

  1745.         FF_ALLOCZ_OR_GOTO(c, c->dither_error[i], (c->dstW+2) * sizeof(int), fail);

  1746. 

  1747.     c->needAlpha = (CONFIG_SWSCALE_ALPHA && isALPHA(c->srcFormat) && isALPHA(c->dstFormat)) ? 1 : 0;

  1748. 

  1749.     // 64 / c->scalingBpp is the same as 16 / sizeof(scaling_intermediate)

  1750.     c->uv_off   = (dst_stride>>1) + 64 / (c->dstBpc &~ 7);

  1751.     c->uv_offx2 = dst_stride + 16;

  1752. 

  1753.     av_assert0(c->chrDstH <= dstH);

  1754. 

  1755.     if (flags & SWS_PRINT_INFO) {

  1756.         const char *scaler = NULL, *cpucaps;

  1757. 

  1758.         for (i = 0; i < FF_ARRAY_ELEMS(scale_algorithms); i++) {

  1759.             if (flags & scale_algorithms[i].flag) {

  1760.                 scaler = scale_algorithms[i].description;

  1761.                 break;

  1762.             }

  1763.         }

  1764.         if (!scaler)

  1765.             scaler =  "ehh flags invalid?!";

  1766.         av_log(c, AV_LOG_INFO, "%s scaler, from %s to %s%s ",

  1767.                scaler,

  1768.                av_get_pix_fmt_name(srcFormat),

  1769. #ifdef DITHER1XBPP

  1770.                dstFormat == AV_PIX_FMT_BGR555   || dstFormat == AV_PIX_FMT_BGR565   ||

  1771.                dstFormat == AV_PIX_FMT_RGB444BE || dstFormat == AV_PIX_FMT_RGB444LE ||

  1772.                dstFormat == AV_PIX_FMT_BGR444BE || dstFormat == AV_PIX_FMT_BGR444LE ?

  1773.                                                              "dithered " : "",

  1774. #else

  1775.                "",

  1776. #endif

  1777.                av_get_pix_fmt_name(dstFormat));

  1778. 

  1779.         if (INLINE_MMXEXT(cpu_flags))

  1780.             cpucaps = "MMXEXT";

  1781.         else if (INLINE_AMD3DNOW(cpu_flags))

  1782.             cpucaps = "3DNOW";

  1783.         else if (INLINE_MMX(cpu_flags))

  1784.             cpucaps = "MMX";

  1785.         else if (PPC_ALTIVEC(cpu_flags))

  1786.             cpucaps = "AltiVec";

  1787.         else

  1788.             cpucaps = "C";

  1789. 

  1790.         av_log(c, AV_LOG_INFO, "using %s\n", cpucaps);

  1791. 

  1792.         av_log(c, AV_LOG_VERBOSE, "%dx%d -> %dx%d\n", srcW, srcH, dstW, dstH);

  1793.         av_log(c, AV_LOG_DEBUG,

  1794.                "lum srcW=%d srcH=%d dstW=%d dstH=%d xInc=%d yInc=%d\n",

  1795.                c->srcW, c->srcH, c->dstW, c->dstH, c->lumXInc, c->lumYInc);

  1796.         av_log(c, AV_LOG_DEBUG,

  1797.                "chr srcW=%d srcH=%d dstW=%d dstH=%d xInc=%d yInc=%d\n",

  1798.                c->chrSrcW, c->chrSrcH, c->chrDstW, c->chrDstH,

  1799.                c->chrXInc, c->chrYInc);

  1800.     }

  1801. 

  1802.     /* alpha blend special case, note this has been split via cascaded contexts if its scaled */

  1803.     if (unscaled && !usesHFilter && !usesVFilter &&

  1804.         c->alphablend != SWS_ALPHA_BLEND_NONE &&

  1805.         isALPHA(srcFormat) &&

  1806.         (c->srcRange == c->dstRange || isAnyRGB(dstFormat)) &&

  1807.         alphaless_fmt(srcFormat) == dstFormat

  1808.     ) {

  1809.         c->swscale = ff_sws_alphablendaway;

  1810. 

  1811.         if (flags & SWS_PRINT_INFO)

  1812.             av_log(c, AV_LOG_INFO,

  1813.                     "using alpha blendaway %s -> %s special converter\n",

  1814.                     av_get_pix_fmt_name(srcFormat), av_get_pix_fmt_name(dstFormat));

  1815.         return 0;

  1816.     }

  1817. 

  1818.     /* unscaled special cases */

  1819.     if (unscaled && !usesHFilter && !usesVFilter &&

  1820.         (c->srcRange == c->dstRange || isAnyRGB(dstFormat) ||

  1821.          isFloat(srcFormat) || isFloat(dstFormat))){

  1822.         ff_get_unscaled_swscale(c);

  1823. 

  1824.         if (c->swscale) {

  1825.             if (flags & SWS_PRINT_INFO)

  1826.                 av_log(c, AV_LOG_INFO,

  1827.                        "using unscaled %s -> %s special converter\n",

  1828.                        av_get_pix_fmt_name(srcFormat), av_get_pix_fmt_name(dstFormat));

  1829.             return 0;

  1830.         }

  1831.     }

  1832. 

  1833.     c->swscale = ff_getSwsFunc(c);

  1834.     return ff_init_filters(c);

  1835. fail: // FIXME replace things by appropriate error codes

  1836.     if (ret == RETCODE_USE_CASCADE)  {

  1837.         int tmpW = sqrt(srcW * (int64_t)dstW);

  1838.         int tmpH = sqrt(srcH * (int64_t)dstH);

  1839.         enum AVPixelFormat tmpFormat = AV_PIX_FMT_YUV420P;

  1840. 

  1841.         if (isALPHA(srcFormat))

  1842.             tmpFormat = AV_PIX_FMT_YUVA420P;

  1843. 

  1844.         if (srcW*(int64_t)srcH <= 4LL*dstW*dstH)

  1845.             return AVERROR(EINVAL);

  1846. 

  1847.         ret = av_image_alloc(c->cascaded_tmp, c->cascaded_tmpStride,

  1848.                              tmpW, tmpH, tmpFormat, 64);

  1849.         if (ret < 0)

  1850.             return ret;

  1851. 

  1852.         c->cascaded_context[0] = sws_getContext(srcW, srcH, srcFormat,

  1853.                                                 tmpW, tmpH, tmpFormat,

  1854.                                                 flags, srcFilter, NULL, c->param);

  1855.         if (!c->cascaded_context[0])

  1856.             return -1;

  1857. 

  1858.         c->cascaded_context[1] = sws_getContext(tmpW, tmpH, tmpFormat,

  1859.                                                 dstW, dstH, dstFormat,

  1860.                                                 flags, NULL, dstFilter, c->param);

  1861.         if (!c->cascaded_context[1])

  1862.             return -1;

  1863.         return 0;

  1864.     }

  1865.     return -1;

  1866. }

  1867. 

  1868. SwsContext *sws_alloc_set_opts(int srcW, int srcH, enum AVPixelFormat srcFormat,

  1869.                                int dstW, int dstH, enum AVPixelFormat dstFormat,

  1870.                                int flags, const double *param)

  1871. {

  1872.     SwsContext *c;

  1873. 

  1874.     if (!(c = sws_alloc_context()))

  1875.         return NULL;

  1876. 

  1877.     c->flags     = flags;

  1878.     c->srcW      = srcW;

  1879.     c->srcH      = srcH;

  1880.     c->dstW      = dstW;

  1881.     c->dstH      = dstH;

  1882.     c->srcFormat = srcFormat;

  1883.     c->dstFormat = dstFormat;

  1884. 

  1885.     if (param) {

  1886.         c->param[0] = param[0];

  1887.         c->param[1] = param[1];

  1888.     }

  1889. 

  1890.     return c;

  1891. }

  1892. 

  1893. SwsContext *sws_getContext(int srcW, int srcH, enum AVPixelFormat srcFormat,

  1894.                            int dstW, int dstH, enum AVPixelFormat dstFormat,

  1895.                            int flags, SwsFilter *srcFilter,

  1896.                            SwsFilter *dstFilter, const double *param)

  1897. {

  1898.     SwsContext *c;

  1899. 

  1900.     c = sws_alloc_set_opts(srcW, srcH, srcFormat,

  1901.                            dstW, dstH, dstFormat,

  1902.                            flags, param);

  1903.     if (!c)

  1904.         return NULL;

  1905. 

  1906.     if (sws_init_context(c, srcFilter, dstFilter) < 0) {

  1907.         sws_freeContext(c);

  1908.         return NULL;

  1909.     }

  1910. 

  1911.     return c;

  1912. }

  1913. 

  1914. static int isnan_vec(SwsVector *a)

  1915. {

  1916.     int i;

  1917.     for (i=0; i<a->length; i++)

  1918.         if (isnan(a->coeff[i]))

  1919.             return 1;

  1920.     return 0;

  1921. }

  1922. 

  1923. static void makenan_vec(SwsVector *a)

  1924. {

  1925.     int i;

  1926.     for (i=0; i<a->length; i++)

  1927.         a->coeff[i] = NAN;

  1928. }

  1929. 

  1930. SwsFilter *sws_getDefaultFilter(float lumaGBlur, float chromaGBlur,

  1931.                                 float lumaSharpen, float chromaSharpen,

  1932.                                 float chromaHShift, float chromaVShift,

  1933.                                 int verbose)

  1934. {

  1935.     SwsFilter *filter = av_malloc(sizeof(SwsFilter));

  1936.     if (!filter)

  1937.         return NULL;

  1938. 

  1939.     if (lumaGBlur != 0.0) {

  1940.         filter->lumH = sws_getGaussianVec(lumaGBlur, 3.0);

  1941.         filter->lumV = sws_getGaussianVec(lumaGBlur, 3.0);

  1942.     } else {

  1943.         filter->lumH = sws_getIdentityVec();

  1944.         filter->lumV = sws_getIdentityVec();

  1945.     }

  1946. 

  1947.     if (chromaGBlur != 0.0) {

  1948.         filter->chrH = sws_getGaussianVec(chromaGBlur, 3.0);

  1949.         filter->chrV = sws_getGaussianVec(chromaGBlur, 3.0);

  1950.     } else {

  1951.         filter->chrH = sws_getIdentityVec();

  1952.         filter->chrV = sws_getIdentityVec();

  1953.     }

  1954. 

  1955.     if (!filter->lumH || !filter->lumV || !filter->chrH || !filter->chrV)

  1956.         goto fail;

  1957. 

  1958.     if (chromaSharpen != 0.0) {

  1959.         SwsVector *id = sws_getIdentityVec();

  1960.         if (!id)

  1961.             goto fail;

  1962.         sws_scaleVec(filter->chrH, -chromaSharpen);

  1963.         sws_scaleVec(filter->chrV, -chromaSharpen);

  1964.         sws_addVec(filter->chrH, id);

  1965.         sws_addVec(filter->chrV, id);

  1966.         sws_freeVec(id);

  1967.     }

  1968. 

  1969.     if (lumaSharpen != 0.0) {

  1970.         SwsVector *id = sws_getIdentityVec();

  1971.         if (!id)

  1972.             goto fail;

  1973.         sws_scaleVec(filter->lumH, -lumaSharpen);

  1974.         sws_scaleVec(filter->lumV, -lumaSharpen);

  1975.         sws_addVec(filter->lumH, id);

  1976.         sws_addVec(filter->lumV, id);

  1977.         sws_freeVec(id);

  1978.     }

  1979. 

  1980.     if (chromaHShift != 0.0)

  1981.         sws_shiftVec(filter->chrH, (int)(chromaHShift + 0.5));

  1982. 

  1983.     if (chromaVShift != 0.0)

  1984.         sws_shiftVec(filter->chrV, (int)(chromaVShift + 0.5));

  1985. 

  1986.     sws_normalizeVec(filter->chrH, 1.0);

  1987.     sws_normalizeVec(filter->chrV, 1.0);

  1988.     sws_normalizeVec(filter->lumH, 1.0);

  1989.     sws_normalizeVec(filter->lumV, 1.0);

  1990. 

  1991.     if (isnan_vec(filter->chrH) ||

  1992.         isnan_vec(filter->chrV) ||

  1993.         isnan_vec(filter->lumH) ||

  1994.         isnan_vec(filter->lumV))

  1995.         goto fail;

  1996. 

  1997.     if (verbose)

  1998.         sws_printVec2(filter->chrH, NULL, AV_LOG_DEBUG);

  1999.     if (verbose)

  2000.         sws_printVec2(filter->lumH, NULL, AV_LOG_DEBUG);

  2001. 

  2002.     return filter;

  2003. 

  2004. fail:

  2005.     sws_freeVec(filter->lumH);

  2006.     sws_freeVec(filter->lumV);

  2007.     sws_freeVec(filter->chrH);

  2008.     sws_freeVec(filter->chrV);

  2009.     av_freep(&filter);

  2010.     return NULL;

  2011. }

  2012. 

  2013. SwsVector *sws_allocVec(int length)

  2014. {

  2015.     SwsVector *vec;

  2016. 

  2017.     if(length <= 0 || length > INT_MAX/ sizeof(double))

  2018.         return NULL;

  2019. 

  2020.     vec = av_malloc(sizeof(SwsVector));

  2021.     if (!vec)

  2022.         return NULL;

  2023.     vec->length = length;

  2024.     vec->coeff  = av_malloc(sizeof(double) * length);

  2025.     if (!vec->coeff)

  2026.         av_freep(&vec);

  2027.     return vec;

  2028. }

  2029. 

  2030. SwsVector *sws_getGaussianVec(double variance, double quality)

  2031. {

  2032.     const int length = (int)(variance * quality + 0.5) | 1;

  2033.     int i;

  2034.     double middle  = (length - 1) * 0.5;

  2035.     SwsVector *vec;

  2036. 

  2037.     if(variance < 0 || quality < 0)

  2038.         return NULL;

  2039. 

  2040.     vec = sws_allocVec(length);

  2041. 

  2042.     if (!vec)

  2043.         return NULL;

  2044. 

  2045.     for (i = 0; i < length; i++) {

  2046.         double dist = i - middle;

  2047.         vec->coeff[i] = exp(-dist * dist / (2 * variance * variance)) /

  2048.                         sqrt(2 * variance * M_PI);

  2049.     }

  2050. 

  2051.     sws_normalizeVec(vec, 1.0);

  2052. 

  2053.     return vec;

  2054. }

  2055. 

  2056. /**

  2057.  * Allocate and return a vector with length coefficients, all

  2058.  * with the same value c.

  2059.  */

  2060. #if !FF_API_SWS_VECTOR

  2061. static

  2062. #endif

  2063. SwsVector *sws_getConstVec(double c, int length)

  2064. {

  2065.     int i;

  2066.     SwsVector *vec = sws_allocVec(length);

  2067. 

  2068.     if (!vec)

  2069.         return NULL;

  2070. 

  2071.     for (i = 0; i < length; i++)

  2072.         vec->coeff[i] = c;

  2073. 

  2074.     return vec;

  2075. }

  2076. 

  2077. /**

  2078.  * Allocate and return a vector with just one coefficient, with

  2079.  * value 1.0.

  2080.  */

  2081. #if !FF_API_SWS_VECTOR

  2082. static

  2083. #endif

  2084. SwsVector *sws_getIdentityVec(void)

  2085. {

  2086.     return sws_getConstVec(1.0, 1);

  2087. }

  2088. 

  2089. static double sws_dcVec(SwsVector *a)

  2090. {

  2091.     int i;

  2092.     double sum = 0;

  2093. 

  2094.     for (i = 0; i < a->length; i++)

  2095.         sum += a->coeff[i];

  2096. 

  2097.     return sum;

  2098. }

  2099. 

  2100. void sws_scaleVec(SwsVector *a, double scalar)

  2101. {

  2102.     int i;

  2103. 

  2104.     for (i = 0; i < a->length; i++)

  2105.         a->coeff[i] *= scalar;

  2106. }

  2107. 

  2108. void sws_normalizeVec(SwsVector *a, double height)

  2109. {

  2110.     sws_scaleVec(a, height / sws_dcVec(a));

  2111. }

  2112. 

  2113. #if FF_API_SWS_VECTOR

  2114. static SwsVector *sws_getConvVec(SwsVector *a, SwsVector *b)

  2115. {

  2116.     int length = a->length + b->length - 1;

  2117.     int i, j;

  2118.     SwsVector *vec = sws_getConstVec(0.0, length);

  2119. 

  2120.     if (!vec)

  2121.         return NULL;

  2122. 

  2123.     for (i = 0; i < a->length; i++) {

  2124.         for (j = 0; j < b->length; j++) {

  2125.             vec->coeff[i + j] += a->coeff[i] * b->coeff[j];

  2126.         }

  2127.     }

  2128. 

  2129.     return vec;

  2130. }

  2131. #endif

  2132. 

  2133. static SwsVector *sws_sumVec(SwsVector *a, SwsVector *b)

  2134. {

  2135.     int length = FFMAX(a->length, b->length);

  2136.     int i;

  2137.     SwsVector *vec = sws_getConstVec(0.0, length);

  2138. 

  2139.     if (!vec)

  2140.         return NULL;

  2141. 

  2142.     for (i = 0; i < a->length; i++)

  2143.         vec->coeff[i + (length - 1) / 2 - (a->length - 1) / 2] += a->coeff[i];

  2144.     for (i = 0; i < b->length; i++)

  2145.         vec->coeff[i + (length - 1) / 2 - (b->length - 1) / 2] += b->coeff[i];

  2146. 

  2147.     return vec;

  2148. }

  2149. 

  2150. #if FF_API_SWS_VECTOR

  2151. static SwsVector *sws_diffVec(SwsVector *a, SwsVector *b)

  2152. {

  2153.     int length = FFMAX(a->length, b->length);

  2154.     int i;

  2155.     SwsVector *vec = sws_getConstVec(0.0, length);

  2156. 

  2157.     if (!vec)

  2158.         return NULL;

  2159. 

  2160.     for (i = 0; i < a->length; i++)

  2161.         vec->coeff[i + (length - 1) / 2 - (a->length - 1) / 2] += a->coeff[i];

  2162.     for (i = 0; i < b->length; i++)

  2163.         vec->coeff[i + (length - 1) / 2 - (b->length - 1) / 2] -= b->coeff[i];

  2164. 

  2165.     return vec;

  2166. }

  2167. #endif

  2168. 

  2169. /* shift left / or right if "shift" is negative */

  2170. static SwsVector *sws_getShiftedVec(SwsVector *a, int shift)

  2171. {

  2172.     int length = a->length + FFABS(shift) * 2;

  2173.     int i;

  2174.     SwsVector *vec = sws_getConstVec(0.0, length);

  2175. 

  2176.     if (!vec)

  2177.         return NULL;

  2178. 

  2179.     for (i = 0; i < a->length; i++) {

  2180.         vec->coeff[i + (length    - 1) / 2 -

  2181.                        (a->length - 1) / 2 - shift] = a->coeff[i];

  2182.     }

  2183. 

  2184.     return vec;

  2185. }

  2186. 

  2187. #if !FF_API_SWS_VECTOR

  2188. static

  2189. #endif

  2190. void sws_shiftVec(SwsVector *a, int shift)

  2191. {

  2192.     SwsVector *shifted = sws_getShiftedVec(a, shift);

  2193.     if (!shifted) {

  2194.         makenan_vec(a);

  2195.         return;

  2196.     }

  2197.     av_free(a->coeff);

  2198.     a->coeff  = shifted->coeff;

  2199.     a->length = shifted->length;

  2200.     av_free(shifted);

  2201. }

  2202. 

  2203. #if !FF_API_SWS_VECTOR

  2204. static

  2205. #endif

  2206. void sws_addVec(SwsVector *a, SwsVector *b)

  2207. {

  2208.     SwsVector *sum = sws_sumVec(a, b);

  2209.     if (!sum) {

  2210.         makenan_vec(a);

  2211.         return;

  2212.     }

  2213.     av_free(a->coeff);

  2214.     a->coeff  = sum->coeff;

  2215.     a->length = sum->length;

  2216.     av_free(sum);

  2217. }

  2218. 

  2219. #if FF_API_SWS_VECTOR

  2220. void sws_subVec(SwsVector *a, SwsVector *b)

  2221. {

  2222.     SwsVector *diff = sws_diffVec(a, b);

  2223.     if (!diff) {

  2224.         makenan_vec(a);

  2225.         return;

  2226.     }

  2227.     av_free(a->coeff);

  2228.     a->coeff  = diff->coeff;

  2229.     a->length = diff->length;

  2230.     av_free(diff);

  2231. }

  2232. 

  2233. void sws_convVec(SwsVector *a, SwsVector *b)

  2234. {

  2235.     SwsVector *conv = sws_getConvVec(a, b);

  2236.     if (!conv) {

  2237.         makenan_vec(a);

  2238.         return;

  2239.     }

  2240.     av_free(a->coeff);

  2241.     a->coeff  = conv->coeff;

  2242.     a->length = conv->length;

  2243.     av_free(conv);

  2244. }

  2245. 

  2246. SwsVector *sws_cloneVec(SwsVector *a)

  2247. {

  2248.     SwsVector *vec = sws_allocVec(a->length);

  2249. 

  2250.     if (!vec)

  2251.         return NULL;

  2252. 

  2253.     memcpy(vec->coeff, a->coeff, a->length * sizeof(*a->coeff));

  2254. 

  2255.     return vec;

  2256. }

  2257. #endif

  2258. 

  2259. /**

  2260.  * Print with av_log() a textual representation of the vector a

  2261.  * if log_level <= av_log_level.

  2262.  */

  2263. #if !FF_API_SWS_VECTOR

  2264. static

  2265. #endif

  2266. void sws_printVec2(SwsVector *a, AVClass *log_ctx, int log_level)

  2267. {

  2268.     int i;

  2269.     double max = 0;

  2270.     double min = 0;

  2271.     double range;

  2272. 

  2273.     for (i = 0; i < a->length; i++)

  2274.         if (a->coeff[i] > max)

  2275.             max = a->coeff[i];

  2276. 

  2277.     for (i = 0; i < a->length; i++)

  2278.         if (a->coeff[i] < min)

  2279.             min = a->coeff[i];

  2280. 

  2281.     range = max - min;

  2282. 

  2283.     for (i = 0; i < a->length; i++) {

  2284.         int x = (int)((a->coeff[i] - min) * 60.0 / range + 0.5);

  2285.         av_log(log_ctx, log_level, "%1.3f ", a->coeff[i]);

  2286.         for (; x > 0; x--)

  2287.             av_log(log_ctx, log_level, " ");

  2288.         av_log(log_ctx, log_level, "|\n");

  2289.     }

  2290. }

  2291. 

  2292. void sws_freeVec(SwsVector *a)

  2293. {

  2294.     if (!a)

  2295.         return;

  2296.     av_freep(&a->coeff);

  2297.     a->length = 0;

  2298.     av_free(a);

  2299. }

  2300. 

  2301. void sws_freeFilter(SwsFilter *filter)

  2302. {

  2303.     if (!filter)

  2304.         return;

  2305. 

  2306.     sws_freeVec(filter->lumH);

  2307.     sws_freeVec(filter->lumV);

  2308.     sws_freeVec(filter->chrH);

  2309.     sws_freeVec(filter->chrV);

  2310.     av_free(filter);

  2311. }

  2312. 

  2313. void sws_freeContext(SwsContext *c)

  2314. {

  2315.     int i;

  2316.     if (!c)

  2317.         return;

  2318. 

  2319.     for (i = 0; i < 4; i++)

  2320.         av_freep(&c->dither_error[i]);

  2321. 

  2322.     av_freep(&c->vLumFilter);

  2323.     av_freep(&c->vChrFilter);

  2324.     av_freep(&c->hLumFilter);

  2325.     av_freep(&c->hChrFilter);

  2326. #if HAVE_ALTIVEC

  2327.     av_freep(&c->vYCoeffsBank);

  2328.     av_freep(&c->vCCoeffsBank);

  2329. #endif

  2330. 

  2331.     av_freep(&c->vLumFilterPos);

  2332.     av_freep(&c->vChrFilterPos);

  2333.     av_freep(&c->hLumFilterPos);

  2334.     av_freep(&c->hChrFilterPos);

  2335. 

  2336. #if HAVE_MMX_INLINE

  2337. #if USE_MMAP

  2338.     if (c->lumMmxextFilterCode)

  2339.         munmap(c->lumMmxextFilterCode, c->lumMmxextFilterCodeSize);

  2340.     if (c->chrMmxextFilterCode)

  2341.         munmap(c->chrMmxextFilterCode, c->chrMmxextFilterCodeSize);

  2342. #elif HAVE_VIRTUALALLOC

  2343.     if (c->lumMmxextFilterCode)

  2344.         VirtualFree(c->lumMmxextFilterCode, 0, MEM_RELEASE);

  2345.     if (c->chrMmxextFilterCode)

  2346.         VirtualFree(c->chrMmxextFilterCode, 0, MEM_RELEASE);

  2347. #else

  2348.     av_free(c->lumMmxextFilterCode);

  2349.     av_free(c->chrMmxextFilterCode);

  2350. #endif

  2351.     c->lumMmxextFilterCode = NULL;

  2352.     c->chrMmxextFilterCode = NULL;

  2353. #endif /* HAVE_MMX_INLINE */

  2354. 

  2355.     av_freep(&c->yuvTable);

  2356.     av_freep(&c->formatConvBuffer);

  2357. 

  2358.     sws_freeContext(c->cascaded_context[0]);

  2359.     sws_freeContext(c->cascaded_context[1]);

  2360.     sws_freeContext(c->cascaded_context[2]);

  2361.     memset(c->cascaded_context, 0, sizeof(c->cascaded_context));

  2362.     av_freep(&c->cascaded_tmp[0]);

  2363.     av_freep(&c->cascaded1_tmp[0]);

  2364. 

  2365.     av_freep(&c->gamma);

  2366.     av_freep(&c->inv_gamma);

  2367. 

  2368.     ff_free_filters(c);

  2369. 

  2370.     av_free(c);

  2371. }

  2372. 

  2373. struct SwsContext *sws_getCachedContext(struct SwsContext *context, int srcW,

  2374.                                         int srcH, enum AVPixelFormat srcFormat,

  2375.                                         int dstW, int dstH,

  2376.                                         enum AVPixelFormat dstFormat, int flags,

  2377.                                         SwsFilter *srcFilter,

  2378.                                         SwsFilter *dstFilter,

  2379.                                         const double *param)

  2380. {

  2381.     static const double default_param[2] = { SWS_PARAM_DEFAULT,

  2382.                                              SWS_PARAM_DEFAULT };

  2383.     int64_t src_h_chr_pos = -513, dst_h_chr_pos = -513,

  2384.             src_v_chr_pos = -513, dst_v_chr_pos = -513;

  2385. 

  2386.     if (!param)

  2387.         param = default_param;

  2388. 

  2389.     if (context &&

  2390.         (context->srcW      != srcW      ||

  2391.          context->srcH      != srcH      ||

  2392.          context->srcFormat != srcFormat ||

  2393.          context->dstW      != dstW      ||

  2394.          context->dstH      != dstH      ||

  2395.          context->dstFormat != dstFormat ||

  2396.          context->flags     != flags     ||

  2397.          context->param[0]  != param[0]  ||

  2398.          context->param[1]  != param[1])) {

  2399. 

  2400.         av_opt_get_int(context, "src_h_chr_pos", 0, &src_h_chr_pos);

  2401.         av_opt_get_int(context, "src_v_chr_pos", 0, &src_v_chr_pos);

  2402.         av_opt_get_int(context, "dst_h_chr_pos", 0, &dst_h_chr_pos);

  2403.         av_opt_get_int(context, "dst_v_chr_pos", 0, &dst_v_chr_pos);

  2404.         sws_freeContext(context);

  2405.         context = NULL;

  2406.     }

  2407. 

  2408.     if (!context) {

  2409.         if (!(context = sws_alloc_context()))

  2410.             return NULL;

  2411.         context->srcW      = srcW;

  2412.         context->srcH      = srcH;

  2413.         context->srcFormat = srcFormat;

  2414.         context->dstW      = dstW;

  2415.         context->dstH      = dstH;

  2416.         context->dstFormat = dstFormat;

  2417.         context->flags     = flags;

  2418.         context->param[0]  = param[0];

  2419.         context->param[1]  = param[1];

  2420. 

  2421.         av_opt_set_int(context, "src_h_chr_pos", src_h_chr_pos, 0);

  2422.         av_opt_set_int(context, "src_v_chr_pos", src_v_chr_pos, 0);

  2423.         av_opt_set_int(context, "dst_h_chr_pos", dst_h_chr_pos, 0);

  2424.         av_opt_set_int(context, "dst_v_chr_pos", dst_v_chr_pos, 0);

  2425. 

  2426.         if (sws_init_context(context, srcFilter, dstFilter) < 0) {

  2427.             sws_freeContext(context);

  2428.             return NULL;

  2429.         }

  2430.     }

  2431.     return context;

  2432. }