LCOV - code coverage report
Current view: top level - libswscale - utils.c (source / functions) Hit Total Coverage
Test: coverage.info Lines: 608 1193 51.0 %
Date: 2017-10-24 00:14:43 Functions: 19 47 40.4 %

          Line data    Source code
       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_SYS_MMAN_H
      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        5431 : unsigned swscale_version(void)
      76             : {
      77             :     av_assert0(LIBSWSCALE_VERSION_MICRO >= 100);
      78        5431 :     return LIBSWSCALE_VERSION_INT;
      79             : }
      80             : 
      81        5431 : const char *swscale_configuration(void)
      82             : {
      83        5431 :     return FFMPEG_CONFIGURATION;
      84             : }
      85             : 
      86           0 : const char *swscale_license(void)
      87             : {
      88             : #define LICENSE_PREFIX "libswscale license: "
      89           0 :     return LICENSE_PREFIX FFMPEG_LICENSE + sizeof(LICENSE_PREFIX) - 1;
      90             : }
      91             : 
      92             : typedef struct FormatEntry {
      93             :     uint8_t is_supported_in         :1;
      94             :     uint8_t is_supported_out        :1;
      95             :     uint8_t is_supported_endianness :1;
      96             : } FormatEntry;
      97             : 
      98             : static const FormatEntry format_entries[AV_PIX_FMT_NB] = {
      99             :     [AV_PIX_FMT_YUV420P]     = { 1, 1 },
     100             :     [AV_PIX_FMT_YUYV422]     = { 1, 1 },
     101             :     [AV_PIX_FMT_RGB24]       = { 1, 1 },
     102             :     [AV_PIX_FMT_BGR24]       = { 1, 1 },
     103             :     [AV_PIX_FMT_YUV422P]     = { 1, 1 },
     104             :     [AV_PIX_FMT_YUV444P]     = { 1, 1 },
     105             :     [AV_PIX_FMT_YUV410P]     = { 1, 1 },
     106             :     [AV_PIX_FMT_YUV411P]     = { 1, 1 },
     107             :     [AV_PIX_FMT_GRAY8]       = { 1, 1 },
     108             :     [AV_PIX_FMT_MONOWHITE]   = { 1, 1 },
     109             :     [AV_PIX_FMT_MONOBLACK]   = { 1, 1 },
     110             :     [AV_PIX_FMT_PAL8]        = { 1, 0 },
     111             :     [AV_PIX_FMT_YUVJ420P]    = { 1, 1 },
     112             :     [AV_PIX_FMT_YUVJ411P]    = { 1, 1 },
     113             :     [AV_PIX_FMT_YUVJ422P]    = { 1, 1 },
     114             :     [AV_PIX_FMT_YUVJ444P]    = { 1, 1 },
     115             :     [AV_PIX_FMT_YVYU422]     = { 1, 1 },
     116             :     [AV_PIX_FMT_UYVY422]     = { 1, 1 },
     117             :     [AV_PIX_FMT_UYYVYY411]   = { 0, 0 },
     118             :     [AV_PIX_FMT_BGR8]        = { 1, 1 },
     119             :     [AV_PIX_FMT_BGR4]        = { 0, 1 },
     120             :     [AV_PIX_FMT_BGR4_BYTE]   = { 1, 1 },
     121             :     [AV_PIX_FMT_RGB8]        = { 1, 1 },
     122             :     [AV_PIX_FMT_RGB4]        = { 0, 1 },
     123             :     [AV_PIX_FMT_RGB4_BYTE]   = { 1, 1 },
     124             :     [AV_PIX_FMT_NV12]        = { 1, 1 },
     125             :     [AV_PIX_FMT_NV21]        = { 1, 1 },
     126             :     [AV_PIX_FMT_ARGB]        = { 1, 1 },
     127             :     [AV_PIX_FMT_RGBA]        = { 1, 1 },
     128             :     [AV_PIX_FMT_ABGR]        = { 1, 1 },
     129             :     [AV_PIX_FMT_BGRA]        = { 1, 1 },
     130             :     [AV_PIX_FMT_0RGB]        = { 1, 1 },
     131             :     [AV_PIX_FMT_RGB0]        = { 1, 1 },
     132             :     [AV_PIX_FMT_0BGR]        = { 1, 1 },
     133             :     [AV_PIX_FMT_BGR0]        = { 1, 1 },
     134             :     [AV_PIX_FMT_GRAY9BE]     = { 1, 1 },
     135             :     [AV_PIX_FMT_GRAY9LE]     = { 1, 1 },
     136             :     [AV_PIX_FMT_GRAY10BE]    = { 1, 1 },
     137             :     [AV_PIX_FMT_GRAY10LE]    = { 1, 1 },
     138             :     [AV_PIX_FMT_GRAY12BE]    = { 1, 1 },
     139             :     [AV_PIX_FMT_GRAY12LE]    = { 1, 1 },
     140             :     [AV_PIX_FMT_GRAY16BE]    = { 1, 1 },
     141             :     [AV_PIX_FMT_GRAY16LE]    = { 1, 1 },
     142             :     [AV_PIX_FMT_YUV440P]     = { 1, 1 },
     143             :     [AV_PIX_FMT_YUVJ440P]    = { 1, 1 },
     144             :     [AV_PIX_FMT_YUV440P10LE] = { 1, 1 },
     145             :     [AV_PIX_FMT_YUV440P10BE] = { 1, 1 },
     146             :     [AV_PIX_FMT_YUV440P12LE] = { 1, 1 },
     147             :     [AV_PIX_FMT_YUV440P12BE] = { 1, 1 },
     148             :     [AV_PIX_FMT_YUVA420P]    = { 1, 1 },
     149             :     [AV_PIX_FMT_YUVA422P]    = { 1, 1 },
     150             :     [AV_PIX_FMT_YUVA444P]    = { 1, 1 },
     151             :     [AV_PIX_FMT_YUVA420P9BE] = { 1, 1 },
     152             :     [AV_PIX_FMT_YUVA420P9LE] = { 1, 1 },
     153             :     [AV_PIX_FMT_YUVA422P9BE] = { 1, 1 },
     154             :     [AV_PIX_FMT_YUVA422P9LE] = { 1, 1 },
     155             :     [AV_PIX_FMT_YUVA444P9BE] = { 1, 1 },
     156             :     [AV_PIX_FMT_YUVA444P9LE] = { 1, 1 },
     157             :     [AV_PIX_FMT_YUVA420P10BE]= { 1, 1 },
     158             :     [AV_PIX_FMT_YUVA420P10LE]= { 1, 1 },
     159             :     [AV_PIX_FMT_YUVA422P10BE]= { 1, 1 },
     160             :     [AV_PIX_FMT_YUVA422P10LE]= { 1, 1 },
     161             :     [AV_PIX_FMT_YUVA444P10BE]= { 1, 1 },
     162             :     [AV_PIX_FMT_YUVA444P10LE]= { 1, 1 },
     163             :     [AV_PIX_FMT_YUVA420P16BE]= { 1, 1 },
     164             :     [AV_PIX_FMT_YUVA420P16LE]= { 1, 1 },
     165             :     [AV_PIX_FMT_YUVA422P16BE]= { 1, 1 },
     166             :     [AV_PIX_FMT_YUVA422P16LE]= { 1, 1 },
     167             :     [AV_PIX_FMT_YUVA444P16BE]= { 1, 1 },
     168             :     [AV_PIX_FMT_YUVA444P16LE]= { 1, 1 },
     169             :     [AV_PIX_FMT_RGB48BE]     = { 1, 1 },
     170             :     [AV_PIX_FMT_RGB48LE]     = { 1, 1 },
     171             :     [AV_PIX_FMT_RGBA64BE]    = { 1, 1, 1 },
     172             :     [AV_PIX_FMT_RGBA64LE]    = { 1, 1, 1 },
     173             :     [AV_PIX_FMT_RGB565BE]    = { 1, 1 },
     174             :     [AV_PIX_FMT_RGB565LE]    = { 1, 1 },
     175             :     [AV_PIX_FMT_RGB555BE]    = { 1, 1 },
     176             :     [AV_PIX_FMT_RGB555LE]    = { 1, 1 },
     177             :     [AV_PIX_FMT_BGR565BE]    = { 1, 1 },
     178             :     [AV_PIX_FMT_BGR565LE]    = { 1, 1 },
     179             :     [AV_PIX_FMT_BGR555BE]    = { 1, 1 },
     180             :     [AV_PIX_FMT_BGR555LE]    = { 1, 1 },
     181             :     [AV_PIX_FMT_YUV420P16LE] = { 1, 1 },
     182             :     [AV_PIX_FMT_YUV420P16BE] = { 1, 1 },
     183             :     [AV_PIX_FMT_YUV422P16LE] = { 1, 1 },
     184             :     [AV_PIX_FMT_YUV422P16BE] = { 1, 1 },
     185             :     [AV_PIX_FMT_YUV444P16LE] = { 1, 1 },
     186             :     [AV_PIX_FMT_YUV444P16BE] = { 1, 1 },
     187             :     [AV_PIX_FMT_RGB444LE]    = { 1, 1 },
     188             :     [AV_PIX_FMT_RGB444BE]    = { 1, 1 },
     189             :     [AV_PIX_FMT_BGR444LE]    = { 1, 1 },
     190             :     [AV_PIX_FMT_BGR444BE]    = { 1, 1 },
     191             :     [AV_PIX_FMT_YA8]         = { 1, 1 },
     192             :     [AV_PIX_FMT_YA16BE]      = { 1, 0 },
     193             :     [AV_PIX_FMT_YA16LE]      = { 1, 0 },
     194             :     [AV_PIX_FMT_BGR48BE]     = { 1, 1 },
     195             :     [AV_PIX_FMT_BGR48LE]     = { 1, 1 },
     196             :     [AV_PIX_FMT_BGRA64BE]    = { 1, 1, 1 },
     197             :     [AV_PIX_FMT_BGRA64LE]    = { 1, 1, 1 },
     198             :     [AV_PIX_FMT_YUV420P9BE]  = { 1, 1 },
     199             :     [AV_PIX_FMT_YUV420P9LE]  = { 1, 1 },
     200             :     [AV_PIX_FMT_YUV420P10BE] = { 1, 1 },
     201             :     [AV_PIX_FMT_YUV420P10LE] = { 1, 1 },
     202             :     [AV_PIX_FMT_YUV420P12BE] = { 1, 1 },
     203             :     [AV_PIX_FMT_YUV420P12LE] = { 1, 1 },
     204             :     [AV_PIX_FMT_YUV420P14BE] = { 1, 1 },
     205             :     [AV_PIX_FMT_YUV420P14LE] = { 1, 1 },
     206             :     [AV_PIX_FMT_YUV422P9BE]  = { 1, 1 },
     207             :     [AV_PIX_FMT_YUV422P9LE]  = { 1, 1 },
     208             :     [AV_PIX_FMT_YUV422P10BE] = { 1, 1 },
     209             :     [AV_PIX_FMT_YUV422P10LE] = { 1, 1 },
     210             :     [AV_PIX_FMT_YUV422P12BE] = { 1, 1 },
     211             :     [AV_PIX_FMT_YUV422P12LE] = { 1, 1 },
     212             :     [AV_PIX_FMT_YUV422P14BE] = { 1, 1 },
     213             :     [AV_PIX_FMT_YUV422P14LE] = { 1, 1 },
     214             :     [AV_PIX_FMT_YUV444P9BE]  = { 1, 1 },
     215             :     [AV_PIX_FMT_YUV444P9LE]  = { 1, 1 },
     216             :     [AV_PIX_FMT_YUV444P10BE] = { 1, 1 },
     217             :     [AV_PIX_FMT_YUV444P10LE] = { 1, 1 },
     218             :     [AV_PIX_FMT_YUV444P12BE] = { 1, 1 },
     219             :     [AV_PIX_FMT_YUV444P12LE] = { 1, 1 },
     220             :     [AV_PIX_FMT_YUV444P14BE] = { 1, 1 },
     221             :     [AV_PIX_FMT_YUV444P14LE] = { 1, 1 },
     222             :     [AV_PIX_FMT_GBRP]        = { 1, 1 },
     223             :     [AV_PIX_FMT_GBRP9LE]     = { 1, 1 },
     224             :     [AV_PIX_FMT_GBRP9BE]     = { 1, 1 },
     225             :     [AV_PIX_FMT_GBRP10LE]    = { 1, 1 },
     226             :     [AV_PIX_FMT_GBRP10BE]    = { 1, 1 },
     227             :     [AV_PIX_FMT_GBRAP10LE]   = { 1, 1 },
     228             :     [AV_PIX_FMT_GBRAP10BE]   = { 1, 1 },
     229             :     [AV_PIX_FMT_GBRP12LE]    = { 1, 1 },
     230             :     [AV_PIX_FMT_GBRP12BE]    = { 1, 1 },
     231             :     [AV_PIX_FMT_GBRAP12LE]   = { 1, 1 },
     232             :     [AV_PIX_FMT_GBRAP12BE]   = { 1, 1 },
     233             :     [AV_PIX_FMT_GBRP14LE]    = { 1, 1 },
     234             :     [AV_PIX_FMT_GBRP14BE]    = { 1, 1 },
     235             :     [AV_PIX_FMT_GBRP16LE]    = { 1, 1 },
     236             :     [AV_PIX_FMT_GBRP16BE]    = { 1, 1 },
     237             :     [AV_PIX_FMT_GBRAP]       = { 1, 1 },
     238             :     [AV_PIX_FMT_GBRAP16LE]   = { 1, 1 },
     239             :     [AV_PIX_FMT_GBRAP16BE]   = { 1, 1 },
     240             :     [AV_PIX_FMT_BAYER_BGGR8] = { 1, 0 },
     241             :     [AV_PIX_FMT_BAYER_RGGB8] = { 1, 0 },
     242             :     [AV_PIX_FMT_BAYER_GBRG8] = { 1, 0 },
     243             :     [AV_PIX_FMT_BAYER_GRBG8] = { 1, 0 },
     244             :     [AV_PIX_FMT_BAYER_BGGR16LE] = { 1, 0 },
     245             :     [AV_PIX_FMT_BAYER_BGGR16BE] = { 1, 0 },
     246             :     [AV_PIX_FMT_BAYER_RGGB16LE] = { 1, 0 },
     247             :     [AV_PIX_FMT_BAYER_RGGB16BE] = { 1, 0 },
     248             :     [AV_PIX_FMT_BAYER_GBRG16LE] = { 1, 0 },
     249             :     [AV_PIX_FMT_BAYER_GBRG16BE] = { 1, 0 },
     250             :     [AV_PIX_FMT_BAYER_GRBG16LE] = { 1, 0 },
     251             :     [AV_PIX_FMT_BAYER_GRBG16BE] = { 1, 0 },
     252             :     [AV_PIX_FMT_XYZ12BE]     = { 1, 1, 1 },
     253             :     [AV_PIX_FMT_XYZ12LE]     = { 1, 1, 1 },
     254             :     [AV_PIX_FMT_AYUV64LE]    = { 1, 1},
     255             :     [AV_PIX_FMT_P010LE]      = { 1, 1 },
     256             :     [AV_PIX_FMT_P010BE]      = { 1, 1 },
     257             :     [AV_PIX_FMT_P016LE]      = { 1, 0 },
     258             :     [AV_PIX_FMT_P016BE]      = { 1, 0 },
     259             : };
     260             : 
     261      554828 : int sws_isSupportedInput(enum AVPixelFormat pix_fmt)
     262             : {
     263      554828 :     return (unsigned)pix_fmt < AV_PIX_FMT_NB ?
     264      554828 :            format_entries[pix_fmt].is_supported_in : 0;
     265             : }
     266             : 
     267      554828 : int sws_isSupportedOutput(enum AVPixelFormat pix_fmt)
     268             : {
     269      554828 :     return (unsigned)pix_fmt < AV_PIX_FMT_NB ?
     270      554828 :            format_entries[pix_fmt].is_supported_out : 0;
     271             : }
     272             : 
     273      196502 : int sws_isSupportedEndiannessConversion(enum AVPixelFormat pix_fmt)
     274             : {
     275      196502 :     return (unsigned)pix_fmt < AV_PIX_FMT_NB ?
     276      196502 :            format_entries[pix_fmt].is_supported_endianness : 0;
     277             : }
     278             : 
     279           0 : static double getSplineCoeff(double a, double b, double c, double d,
     280             :                              double dist)
     281             : {
     282           0 :     if (dist <= 1.0)
     283           0 :         return ((d * dist + c) * dist + b) * dist + a;
     284             :     else
     285           0 :         return getSplineCoeff(0.0,
     286           0 :                                b + 2.0 * c + 3.0 * d,
     287           0 :                                c + 3.0 * d,
     288           0 :                               -b - 3.0 * c - 6.0 * d,
     289             :                               dist - 1.0);
     290             : }
     291             : 
     292       22576 : static av_cold int get_local_pos(SwsContext *s, int chr_subsample, int pos, int dir)
     293             : {
     294       22576 :     if (pos == -1 || pos <= -513) {
     295        8779 :         pos = (128 << chr_subsample) - 128;
     296             :     }
     297       22576 :     pos += 128; // relative to ideal left edge
     298       22576 :     return pos >> chr_subsample;
     299             : }
     300             : 
     301             : typedef struct {
     302             :     int flag;                   ///< flag associated to the algorithm
     303             :     const char *description;    ///< human-readable description
     304             :     int size_factor;            ///< size factor used when initing the filters
     305             : } ScaleAlgorithm;
     306             : 
     307             : static const ScaleAlgorithm scale_algorithms[] = {
     308             :     { SWS_AREA,          "area averaging",                  1 /* downscale only, for upscale it is bilinear */ },
     309             :     { SWS_BICUBIC,       "bicubic",                         4 },
     310             :     { SWS_BICUBLIN,      "luma bicubic / chroma bilinear", -1 },
     311             :     { SWS_BILINEAR,      "bilinear",                        2 },
     312             :     { SWS_FAST_BILINEAR, "fast bilinear",                  -1 },
     313             :     { SWS_GAUSS,         "Gaussian",                        8 /* infinite ;) */ },
     314             :     { SWS_LANCZOS,       "Lanczos",                        -1 /* custom */ },
     315             :     { SWS_POINT,         "nearest neighbor / point",       -1 },
     316             :     { SWS_SINC,          "sinc",                           20 /* infinite ;) */ },
     317             :     { SWS_SPLINE,        "bicubic spline",                 20 /* infinite :)*/ },
     318             :     { SWS_X,             "experimental",                    8 },
     319             : };
     320             : 
     321       11288 : static av_cold int initFilter(int16_t **outFilter, int32_t **filterPos,
     322             :                               int *outFilterSize, int xInc, int srcW,
     323             :                               int dstW, int filterAlign, int one,
     324             :                               int flags, int cpu_flags,
     325             :                               SwsVector *srcFilter, SwsVector *dstFilter,
     326             :                               double param[2], int srcPos, int dstPos)
     327             : {
     328             :     int i;
     329             :     int filterSize;
     330             :     int filter2Size;
     331             :     int minFilterSize;
     332       11288 :     int64_t *filter    = NULL;
     333       11288 :     int64_t *filter2   = NULL;
     334       11288 :     const int64_t fone = 1LL << (54 - FFMIN(av_log2(srcW/dstW), 8));
     335       11288 :     int ret            = -1;
     336             : 
     337       11288 :     emms_c(); // FIXME should not be required but IS (even for non-MMX versions)
     338             : 
     339             :     // NOTE: the +3 is for the MMX(+1) / SSE(+3) scaler which reads over the end
     340       11288 :     FF_ALLOC_ARRAY_OR_GOTO(NULL, *filterPos, (dstW + 3), sizeof(**filterPos), fail);
     341             : 
     342       18402 :     if (FFABS(xInc - 0x10000) < 10 && srcPos == dstPos) { // unscaled
     343             :         int i;
     344        7114 :         filterSize = 1;
     345        7114 :         FF_ALLOCZ_ARRAY_OR_GOTO(NULL, filter,
     346             :                                 dstW, sizeof(*filter) * filterSize, fail);
     347             : 
     348     1921151 :         for (i = 0; i < dstW; i++) {
     349     1914037 :             filter[i * filterSize] = fone;
     350     1914037 :             (*filterPos)[i]        = i;
     351             :         }
     352        4174 :     } else if (flags & SWS_POINT) { // lame looking point sampling mode
     353             :         int i;
     354             :         int64_t xDstInSrc;
     355         188 :         filterSize = 1;
     356         188 :         FF_ALLOC_ARRAY_OR_GOTO(NULL, filter,
     357             :                                dstW, sizeof(*filter) * filterSize, fail);
     358             : 
     359         188 :         xDstInSrc = ((dstPos*(int64_t)xInc)>>8) - ((srcPos*0x8000LL)>>7);
     360       43549 :         for (i = 0; i < dstW; i++) {
     361       43361 :             int xx = (xDstInSrc - ((filterSize - 1) << 15) + (1 << 15)) >> 16;
     362             : 
     363       43361 :             (*filterPos)[i] = xx;
     364       43361 :             filter[i]       = fone;
     365       43361 :             xDstInSrc      += xInc;
     366             :         }
     367        7920 :     } else if ((xInc <= (1 << 16) && (flags & SWS_AREA)) ||
     368        3986 :                (flags & SWS_FAST_BILINEAR)) { // bilinear upscale
     369             :         int i;
     370             :         int64_t xDstInSrc;
     371          52 :         filterSize = 2;
     372          52 :         FF_ALLOC_ARRAY_OR_GOTO(NULL, filter,
     373             :                                dstW, sizeof(*filter) * filterSize, fail);
     374             : 
     375          52 :         xDstInSrc = ((dstPos*(int64_t)xInc)>>8) - ((srcPos*0x8000LL)>>7);
     376       29092 :         for (i = 0; i < dstW; i++) {
     377       29040 :             int xx = (xDstInSrc - ((filterSize - 1) << 15) + (1 << 15)) >> 16;
     378             :             int j;
     379             : 
     380       29040 :             (*filterPos)[i] = xx;
     381             :             // bilinear upscale / linear interpolate / area averaging
     382       87120 :             for (j = 0; j < filterSize; j++) {
     383       58080 :                 int64_t coeff= fone - FFABS(((int64_t)xx<<16) - xDstInSrc)*(fone>>16);
     384       58080 :                 if (coeff < 0)
     385           0 :                     coeff = 0;
     386       58080 :                 filter[i * filterSize + j] = coeff;
     387       58080 :                 xx++;
     388             :             }
     389       29040 :             xDstInSrc += xInc;
     390             :         }
     391             :     } else {
     392             :         int64_t xDstInSrc;
     393        3934 :         int sizeFactor = -1;
     394             : 
     395        7770 :         for (i = 0; i < FF_ARRAY_ELEMS(scale_algorithms); i++) {
     396        7770 :             if (flags & scale_algorithms[i].flag && scale_algorithms[i].size_factor > 0) {
     397        3934 :                 sizeFactor = scale_algorithms[i].size_factor;
     398        3934 :                 break;
     399             :             }
     400             :         }
     401        3934 :         if (flags & SWS_LANCZOS)
     402           0 :             sizeFactor = param[0] != SWS_PARAM_DEFAULT ? ceil(2 * param[0]) : 6;
     403        3934 :         av_assert0(sizeFactor > 0);
     404             : 
     405        3934 :         if (xInc <= 1 << 16)
     406        2706 :             filterSize = 1 + sizeFactor;    // upscale
     407             :         else
     408        1228 :             filterSize = 1 + (sizeFactor * srcW + dstW - 1) / dstW;
     409             : 
     410        3934 :         filterSize = FFMIN(filterSize, srcW - 2);
     411        3934 :         filterSize = FFMAX(filterSize, 1);
     412             : 
     413        3934 :         FF_ALLOC_ARRAY_OR_GOTO(NULL, filter,
     414             :                                dstW, sizeof(*filter) * filterSize, fail);
     415             : 
     416        3934 :         xDstInSrc = ((dstPos*(int64_t)xInc)>>7) - ((srcPos*0x10000LL)>>7);
     417     1237637 :         for (i = 0; i < dstW; i++) {
     418     1233703 :             int xx = (xDstInSrc - (filterSize - 2) * (1LL<<16)) / (1 << 17);
     419             :             int j;
     420     1233703 :             (*filterPos)[i] = xx;
     421     8298734 :             for (j = 0; j < filterSize; j++) {
     422     7065031 :                 int64_t d = (FFABS(((int64_t)xx * (1 << 17)) - xDstInSrc)) << 13;
     423             :                 double floatd;
     424             :                 int64_t coeff;
     425             : 
     426     7065031 :                 if (xInc > 1 << 16)
     427     1752576 :                     d = d * dstW / srcW;
     428     7065031 :                 floatd = d * (1.0 / (1 << 30));
     429             : 
     430     7065031 :                 if (flags & SWS_BICUBIC) {
     431     6734054 :                     int64_t B = (param[0] != SWS_PARAM_DEFAULT ? param[0] :   0) * (1 << 24);
     432     6734054 :                     int64_t C = (param[1] != SWS_PARAM_DEFAULT ? param[1] : 0.6) * (1 << 24);
     433             : 
     434     6734054 :                     if (d >= 1LL << 31) {
     435     1278556 :                         coeff = 0.0;
     436             :                     } else {
     437     5455498 :                         int64_t dd  = (d  * d) >> 30;
     438     5455498 :                         int64_t ddd = (dd * d) >> 30;
     439             : 
     440     5455498 :                         if (d < 1LL << 30)
     441     8194260 :                             coeff =  (12 * (1 << 24) -  9 * B - 6 * C) * ddd +
     442     2731420 :                                     (-18 * (1 << 24) + 12 * B + 6 * C) *  dd +
     443     2731420 :                                       (6 * (1 << 24) -  2 * B)         * (1 << 30);
     444             :                         else
     445     8172234 :                             coeff =      (-B -  6 * C) * ddd +
     446     5448156 :                                       (6 * B + 30 * C) * dd  +
     447     2724078 :                                     (-12 * B - 48 * C) * d   +
     448     2724078 :                                       (8 * B + 24 * C) * (1 << 30);
     449             :                     }
     450     6734054 :                     coeff /= (1LL<<54)/fone;
     451      330977 :                 } else if (flags & SWS_X) {
     452           0 :                     double A = param[0] != SWS_PARAM_DEFAULT ? param[0] : 1.0;
     453             :                     double c;
     454             : 
     455           0 :                     if (floatd < 1.0)
     456           0 :                         c = cos(floatd * M_PI);
     457             :                     else
     458           0 :                         c = -1.0;
     459           0 :                     if (c < 0.0)
     460           0 :                         c = -pow(-c, A);
     461             :                     else
     462           0 :                         c = pow(c, A);
     463           0 :                     coeff = (c * 0.5 + 0.5) * fone;
     464      330977 :                 } else if (flags & SWS_AREA) {
     465      288927 :                     int64_t d2 = d - (1 << 29);
     466      288927 :                     if (d2 * xInc < -(1LL << (29 + 16)))
     467      218924 :                         coeff = 1.0 * (1LL << (30 + 16));
     468       70003 :                     else if (d2 * xInc < (1LL << (29 + 16)))
     469       60908 :                         coeff = -d2 * xInc + (1LL << (29 + 16));
     470             :                     else
     471        9095 :                         coeff = 0.0;
     472      288927 :                     coeff *= fone >> (30 + 16);
     473       42050 :                 } else if (flags & SWS_GAUSS) {
     474           0 :                     double p = param[0] != SWS_PARAM_DEFAULT ? param[0] : 3.0;
     475           0 :                     coeff = exp2(-p * floatd * floatd) * fone;
     476       42050 :                 } else if (flags & SWS_SINC) {
     477           0 :                     coeff = (d ? sin(floatd * M_PI) / (floatd * M_PI) : 1.0) * fone;
     478       42050 :                 } else if (flags & SWS_LANCZOS) {
     479           0 :                     double p = param[0] != SWS_PARAM_DEFAULT ? param[0] : 3.0;
     480           0 :                     coeff = (d ? sin(floatd * M_PI) * sin(floatd * M_PI / p) /
     481           0 :                              (floatd * floatd * M_PI * M_PI / p) : 1.0) * fone;
     482           0 :                     if (floatd > p)
     483           0 :                         coeff = 0;
     484       42050 :                 } else if (flags & SWS_BILINEAR) {
     485       42050 :                     coeff = (1 << 30) - d;
     486       42050 :                     if (coeff < 0)
     487        5238 :                         coeff = 0;
     488       42050 :                     coeff *= fone >> 30;
     489           0 :                 } else if (flags & SWS_SPLINE) {
     490           0 :                     double p = -2.196152422706632;
     491           0 :                     coeff = getSplineCoeff(1.0, 0.0, p, -p - 1.0, floatd) * fone;
     492             :                 } else {
     493           0 :                     av_assert0(0);
     494             :                 }
     495             : 
     496     7065031 :                 filter[i * filterSize + j] = coeff;
     497     7065031 :                 xx++;
     498             :             }
     499     1233703 :             xDstInSrc += 2 * xInc;
     500             :         }
     501             :     }
     502             : 
     503             :     /* apply src & dst Filter to filter -> filter2
     504             :      * av_free(filter);
     505             :      */
     506       11288 :     av_assert0(filterSize > 0);
     507       11288 :     filter2Size = filterSize;
     508       11288 :     if (srcFilter)
     509           0 :         filter2Size += srcFilter->length - 1;
     510       11288 :     if (dstFilter)
     511           0 :         filter2Size += dstFilter->length - 1;
     512       11288 :     av_assert0(filter2Size > 0);
     513       11288 :     FF_ALLOCZ_ARRAY_OR_GOTO(NULL, filter2, dstW, filter2Size * sizeof(*filter2), fail);
     514             : 
     515     3231429 :     for (i = 0; i < dstW; i++) {
     516             :         int j, k;
     517             : 
     518     3220141 :         if (srcFilter) {
     519           0 :             for (k = 0; k < srcFilter->length; k++) {
     520           0 :                 for (j = 0; j < filterSize; j++)
     521           0 :                     filter2[i * filter2Size + k + j] +=
     522           0 :                         srcFilter->coeff[k] * filter[i * filterSize + j];
     523             :             }
     524             :         } else {
     525    12300650 :             for (j = 0; j < filterSize; j++)
     526     9080509 :                 filter2[i * filter2Size + j] = filter[i * filterSize + j];
     527             :         }
     528             :         // FIXME dstFilter
     529             : 
     530     3220141 :         (*filterPos)[i] += (filterSize - 1) / 2 - (filter2Size - 1) / 2;
     531             :     }
     532       11288 :     av_freep(&filter);
     533             : 
     534             :     /* try to reduce the filter-size (step1 find size and shift left) */
     535             :     // Assume it is near normalized (*0.5 or *2.0 is OK but * 0.001 is not).
     536       11288 :     minFilterSize = 0;
     537     3231429 :     for (i = dstW - 1; i >= 0; i--) {
     538     3220141 :         int min = filter2Size;
     539             :         int j;
     540     3220141 :         int64_t cutOff = 0.0;
     541             : 
     542             :         /* get rid of near zero elements on the left by shifting left */
     543     3853275 :         for (j = 0; j < filter2Size; j++) {
     544             :             int k;
     545     3853275 :             cutOff += FFABS(filter2[i * filter2Size]);
     546             : 
     547     3853275 :             if (cutOff > SWS_MAX_REDUCE_CUTOFF * fone)
     548     3219040 :                 break;
     549             : 
     550             :             /* preserve monotonicity because the core can't handle the
     551             :              * filter otherwise */
     552      634235 :             if (i < dstW - 1 && (*filterPos)[i] >= (*filterPos)[i + 1])
     553        1101 :                 break;
     554             : 
     555             :             // move filter coefficients left
     556     3707722 :             for (k = 1; k < filter2Size; k++)
     557     3074588 :                 filter2[i * filter2Size + k - 1] = filter2[i * filter2Size + k];
     558      633134 :             filter2[i * filter2Size + k - 1] = 0;
     559      633134 :             (*filterPos)[i]++;
     560             :         }
     561             : 
     562     3220141 :         cutOff = 0;
     563             :         /* count near zeros on the right */
     564     4549216 :         for (j = filter2Size - 1; j > 0; j--) {
     565     2591715 :             cutOff += FFABS(filter2[i * filter2Size + j]);
     566             : 
     567     2591715 :             if (cutOff > SWS_MAX_REDUCE_CUTOFF * fone)
     568     1262640 :                 break;
     569     1329075 :             min--;
     570             :         }
     571             : 
     572     3220141 :         if (min > minFilterSize)
     573       11746 :             minFilterSize = min;
     574             :     }
     575             : 
     576             :     if (PPC_ALTIVEC(cpu_flags)) {
     577             :         // we can handle the special case 4, so we don't want to go the full 8
     578             :         if (minFilterSize < 5)
     579             :             filterAlign = 4;
     580             : 
     581             :         /* We really don't want to waste our time doing useless computation, so
     582             :          * fall back on the scalar C code for very small filters.
     583             :          * Vectorizing is worth it only if you have a decent-sized vector. */
     584             :         if (minFilterSize < 3)
     585             :             filterAlign = 1;
     586             :     }
     587             : 
     588       11288 :     if (HAVE_MMX && cpu_flags & AV_CPU_FLAG_MMX) {
     589             :         // special case for unscaled vertical filtering
     590          28 :         if (minFilterSize == 1 && filterAlign == 2)
     591          11 :             filterAlign = 1;
     592             :     }
     593             : 
     594       11288 :     av_assert0(minFilterSize > 0);
     595       11288 :     filterSize = (minFilterSize + (filterAlign - 1)) & (~(filterAlign - 1));
     596       11288 :     av_assert0(filterSize > 0);
     597       11288 :     filter = av_malloc_array(dstW, filterSize * sizeof(*filter));
     598       11288 :     if (!filter)
     599           0 :         goto fail;
     600       11288 :     if (filterSize >= MAX_FILTER_SIZE * 16 /
     601       11288 :                       ((flags & SWS_ACCURATE_RND) ? APCK_SIZE : 16)) {
     602           0 :         ret = RETCODE_USE_CASCADE;
     603           0 :         goto fail;
     604             :     }
     605       11288 :     *outFilterSize = filterSize;
     606             : 
     607       11288 :     if (flags & SWS_PRINT_INFO)
     608           0 :         av_log(NULL, AV_LOG_VERBOSE,
     609             :                "SwScaler: reducing / aligning filtersize %d -> %d\n",
     610             :                filter2Size, filterSize);
     611             :     /* try to reduce the filter-size (step2 reduce it) */
     612     3231429 :     for (i = 0; i < dstW; i++) {
     613             :         int j;
     614             : 
     615    11041867 :         for (j = 0; j < filterSize; j++) {
     616     7821726 :             if (j >= filter2Size)
     617        7056 :                 filter[i * filterSize + j] = 0;
     618             :             else
     619     7814670 :                 filter[i * filterSize + j] = filter2[i * filter2Size + j];
     620     7821726 :             if ((flags & SWS_BITEXACT) && j >= minFilterSize)
     621        4320 :                 filter[i * filterSize + j] = 0;
     622             :         }
     623             :     }
     624             : 
     625             :     // FIXME try to align filterPos if possible
     626             : 
     627             :     // fix borders
     628     3231429 :     for (i = 0; i < dstW; i++) {
     629             :         int j;
     630     3220141 :         if ((*filterPos)[i] < 0) {
     631             :             // move filter coefficients left to compensate for filterPos
     632       53739 :             for (j = 1; j < filterSize; j++) {
     633       43089 :                 int left = FFMAX(j + (*filterPos)[i], 0);
     634       43089 :                 filter[i * filterSize + left] += filter[i * filterSize + j];
     635       43089 :                 filter[i * filterSize + j]     = 0;
     636             :             }
     637       10650 :             (*filterPos)[i]= 0;
     638             :         }
     639             : 
     640     3220141 :         if ((*filterPos)[i] + filterSize > srcW) {
     641       15186 :             int shift = (*filterPos)[i] + FFMIN(filterSize - srcW, 0);
     642       15186 :             int64_t acc = 0;
     643             : 
     644       89397 :             for (j = filterSize - 1; j >= 0; j--) {
     645       74211 :                 if ((*filterPos)[i] + j >= srcW) {
     646       22928 :                     acc += filter[i * filterSize + j];
     647       22928 :                     filter[i * filterSize + j] = 0;
     648             :                 }
     649             :             }
     650       89397 :             for (j = filterSize - 1; j >= 0; j--) {
     651       74211 :                 if (j < shift) {
     652       22928 :                     filter[i * filterSize + j] = 0;
     653             :                 } else {
     654       51283 :                     filter[i * filterSize + j] = filter[i * filterSize + j - shift];
     655             :                 }
     656             :             }
     657             : 
     658       15186 :             (*filterPos)[i]-= shift;
     659       15186 :             filter[i * filterSize + srcW - 1 - (*filterPos)[i]] += acc;
     660             :         }
     661     3220141 :         av_assert0((*filterPos)[i] >= 0);
     662     3220141 :         av_assert0((*filterPos)[i] < srcW);
     663     3220141 :         if ((*filterPos)[i] + filterSize > srcW) {
     664           0 :             for (j = 0; j < filterSize; j++) {
     665           0 :                 av_assert0((*filterPos)[i] + j < srcW || !filter[i * filterSize + j]);
     666             :             }
     667             :         }
     668             :     }
     669             : 
     670             :     // Note the +1 is for the MMX scaler which reads over the end
     671             :     /* align at 16 for AltiVec (needed by hScale_altivec_real) */
     672       11288 :     FF_ALLOCZ_ARRAY_OR_GOTO(NULL, *outFilter,
     673             :                             (dstW + 3), *outFilterSize * sizeof(int16_t), fail);
     674             : 
     675             :     /* normalize & store in outFilter */
     676     3231429 :     for (i = 0; i < dstW; i++) {
     677             :         int j;
     678     3220141 :         int64_t error = 0;
     679     3220141 :         int64_t sum   = 0;
     680             : 
     681    11041867 :         for (j = 0; j < filterSize; j++) {
     682     7821726 :             sum += filter[i * filterSize + j];
     683             :         }
     684     3220141 :         sum = (sum + one / 2) / one;
     685     3220141 :         if (!sum) {
     686           0 :             av_log(NULL, AV_LOG_WARNING, "SwScaler: zero vector in scaling\n");
     687           0 :             sum = 1;
     688             :         }
     689    11041867 :         for (j = 0; j < *outFilterSize; j++) {
     690     7821726 :             int64_t v = filter[i * filterSize + j] + error;
     691     7821726 :             int intV  = ROUNDED_DIV(v, sum);
     692     7821726 :             (*outFilter)[i * (*outFilterSize) + j] = intV;
     693     7821726 :             error                                  = v - intV * sum;
     694             :         }
     695             :     }
     696             : 
     697       22576 :     (*filterPos)[dstW + 0] =
     698       33864 :     (*filterPos)[dstW + 1] =
     699       22576 :     (*filterPos)[dstW + 2] = (*filterPos)[dstW - 1]; /* the MMX/SSE scaler will
     700             :                                                       * read over the end */
     701       43070 :     for (i = 0; i < *outFilterSize; i++) {
     702       31782 :         int k = (dstW - 1) * (*outFilterSize) + i;
     703       63564 :         (*outFilter)[k + 1 * (*outFilterSize)] =
     704       95346 :         (*outFilter)[k + 2 * (*outFilterSize)] =
     705       63564 :         (*outFilter)[k + 3 * (*outFilterSize)] = (*outFilter)[k];
     706             :     }
     707             : 
     708       11288 :     ret = 0;
     709             : 
     710       11288 : fail:
     711       11288 :     if(ret < 0)
     712           0 :         av_log(NULL, ret == RETCODE_USE_CASCADE ? AV_LOG_DEBUG : AV_LOG_ERROR, "sws: initFilter failed\n");
     713       11288 :     av_free(filter);
     714       11288 :     av_free(filter2);
     715       11288 :     return ret;
     716             : }
     717             : 
     718        1145 : static void fill_rgb2yuv_table(SwsContext *c, const int table[4], int dstRange)
     719             : {
     720             :     int64_t W, V, Z, Cy, Cu, Cv;
     721        1145 :     int64_t vr =  table[0];
     722        1145 :     int64_t ub =  table[1];
     723        1145 :     int64_t ug = -table[2];
     724        1145 :     int64_t vg = -table[3];
     725        1145 :     int64_t ONE = 65536;
     726        1145 :     int64_t cy = ONE;
     727        1145 :     uint8_t *p = (uint8_t*)c->input_rgb2yuv_table;
     728             :     int i;
     729             :     static const int8_t map[] = {
     730             :     BY_IDX, GY_IDX, -1    , BY_IDX, BY_IDX, GY_IDX, -1    , BY_IDX,
     731             :     RY_IDX, -1    , GY_IDX, RY_IDX, RY_IDX, -1    , GY_IDX, RY_IDX,
     732             :     RY_IDX, GY_IDX, -1    , RY_IDX, RY_IDX, GY_IDX, -1    , RY_IDX,
     733             :     BY_IDX, -1    , GY_IDX, BY_IDX, BY_IDX, -1    , GY_IDX, BY_IDX,
     734             :     BU_IDX, GU_IDX, -1    , BU_IDX, BU_IDX, GU_IDX, -1    , BU_IDX,
     735             :     RU_IDX, -1    , GU_IDX, RU_IDX, RU_IDX, -1    , GU_IDX, RU_IDX,
     736             :     RU_IDX, GU_IDX, -1    , RU_IDX, RU_IDX, GU_IDX, -1    , RU_IDX,
     737             :     BU_IDX, -1    , GU_IDX, BU_IDX, BU_IDX, -1    , GU_IDX, BU_IDX,
     738             :     BV_IDX, GV_IDX, -1    , BV_IDX, BV_IDX, GV_IDX, -1    , BV_IDX,
     739             :     RV_IDX, -1    , GV_IDX, RV_IDX, RV_IDX, -1    , GV_IDX, RV_IDX,
     740             :     RV_IDX, GV_IDX, -1    , RV_IDX, RV_IDX, GV_IDX, -1    , RV_IDX,
     741             :     BV_IDX, -1    , GV_IDX, BV_IDX, BV_IDX, -1    , GV_IDX, BV_IDX,
     742             :     RY_IDX, BY_IDX, RY_IDX, BY_IDX, RY_IDX, BY_IDX, RY_IDX, BY_IDX,
     743             :     BY_IDX, RY_IDX, BY_IDX, RY_IDX, BY_IDX, RY_IDX, BY_IDX, RY_IDX,
     744             :     GY_IDX, -1    , GY_IDX, -1    , GY_IDX, -1    , GY_IDX, -1    ,
     745             :     -1    , GY_IDX, -1    , GY_IDX, -1    , GY_IDX, -1    , GY_IDX,
     746             :     RU_IDX, BU_IDX, RU_IDX, BU_IDX, RU_IDX, BU_IDX, RU_IDX, BU_IDX,
     747             :     BU_IDX, RU_IDX, BU_IDX, RU_IDX, BU_IDX, RU_IDX, BU_IDX, RU_IDX,
     748             :     GU_IDX, -1    , GU_IDX, -1    , GU_IDX, -1    , GU_IDX, -1    ,
     749             :     -1    , GU_IDX, -1    , GU_IDX, -1    , GU_IDX, -1    , GU_IDX,
     750             :     RV_IDX, BV_IDX, RV_IDX, BV_IDX, RV_IDX, BV_IDX, RV_IDX, BV_IDX,
     751             :     BV_IDX, RV_IDX, BV_IDX, RV_IDX, BV_IDX, RV_IDX, BV_IDX, RV_IDX,
     752             :     GV_IDX, -1    , GV_IDX, -1    , GV_IDX, -1    , GV_IDX, -1    ,
     753             :     -1    , GV_IDX, -1    , GV_IDX, -1    , GV_IDX, -1    , GV_IDX, //23
     754             :     -1    , -1    , -1    , -1    , -1    , -1    , -1    , -1    , //24
     755             :     -1    , -1    , -1    , -1    , -1    , -1    , -1    , -1    , //25
     756             :     -1    , -1    , -1    , -1    , -1    , -1    , -1    , -1    , //26
     757             :     -1    , -1    , -1    , -1    , -1    , -1    , -1    , -1    , //27
     758             :     -1    , -1    , -1    , -1    , -1    , -1    , -1    , -1    , //28
     759             :     -1    , -1    , -1    , -1    , -1    , -1    , -1    , -1    , //29
     760             :     -1    , -1    , -1    , -1    , -1    , -1    , -1    , -1    , //30
     761             :     -1    , -1    , -1    , -1    , -1    , -1    , -1    , -1    , //31
     762             :     BY_IDX, GY_IDX, RY_IDX, -1    , -1    , -1    , -1    , -1    , //32
     763             :     BU_IDX, GU_IDX, RU_IDX, -1    , -1    , -1    , -1    , -1    , //33
     764             :     BV_IDX, GV_IDX, RV_IDX, -1    , -1    , -1    , -1    , -1    , //34
     765             :     };
     766             : 
     767        1145 :     dstRange = 0; //FIXME range = 1 is handled elsewhere
     768             : 
     769        1145 :     if (!dstRange) {
     770        1145 :         cy = cy * 255 / 219;
     771             :     } else {
     772           0 :         vr = vr * 224 / 255;
     773           0 :         ub = ub * 224 / 255;
     774           0 :         ug = ug * 224 / 255;
     775           0 :         vg = vg * 224 / 255;
     776             :     }
     777        1145 :     W = ROUNDED_DIV(ONE*ONE*ug, ub);
     778        1145 :     V = ROUNDED_DIV(ONE*ONE*vg, vr);
     779        1145 :     Z = ONE*ONE-W-V;
     780             : 
     781        1145 :     Cy = ROUNDED_DIV(cy*Z, ONE);
     782        1145 :     Cu = ROUNDED_DIV(ub*Z, ONE);
     783        1145 :     Cv = ROUNDED_DIV(vr*Z, ONE);
     784             : 
     785        1145 :     c->input_rgb2yuv_table[RY_IDX] = -ROUNDED_DIV((1 << RGB2YUV_SHIFT)*V        , Cy);
     786        1145 :     c->input_rgb2yuv_table[GY_IDX] =  ROUNDED_DIV((1 << RGB2YUV_SHIFT)*ONE*ONE  , Cy);
     787        1145 :     c->input_rgb2yuv_table[BY_IDX] = -ROUNDED_DIV((1 << RGB2YUV_SHIFT)*W        , Cy);
     788             : 
     789        1145 :     c->input_rgb2yuv_table[RU_IDX] =  ROUNDED_DIV((1 << RGB2YUV_SHIFT)*V        , Cu);
     790        1145 :     c->input_rgb2yuv_table[GU_IDX] = -ROUNDED_DIV((1 << RGB2YUV_SHIFT)*ONE*ONE  , Cu);
     791        1145 :     c->input_rgb2yuv_table[BU_IDX] =  ROUNDED_DIV((1 << RGB2YUV_SHIFT)*(Z+W)    , Cu);
     792             : 
     793        1145 :     c->input_rgb2yuv_table[RV_IDX] =  ROUNDED_DIV((1 << RGB2YUV_SHIFT)*(V+Z)    , Cv);
     794        1145 :     c->input_rgb2yuv_table[GV_IDX] = -ROUNDED_DIV((1 << RGB2YUV_SHIFT)*ONE*ONE  , Cv);
     795        1145 :     c->input_rgb2yuv_table[BV_IDX] =  ROUNDED_DIV((1 << RGB2YUV_SHIFT)*W        , Cv);
     796             : 
     797        1145 :     if(/*!dstRange && */!memcmp(table, ff_yuv2rgb_coeffs[SWS_CS_DEFAULT], sizeof(ff_yuv2rgb_coeffs[SWS_CS_DEFAULT]))) {
     798        1145 :         c->input_rgb2yuv_table[BY_IDX] =  ((int)(0.114 * 219 / 255 * (1 << RGB2YUV_SHIFT) + 0.5));
     799        1145 :         c->input_rgb2yuv_table[BV_IDX] = (-(int)(0.081 * 224 / 255 * (1 << RGB2YUV_SHIFT) + 0.5));
     800        1145 :         c->input_rgb2yuv_table[BU_IDX] =  ((int)(0.500 * 224 / 255 * (1 << RGB2YUV_SHIFT) + 0.5));
     801        1145 :         c->input_rgb2yuv_table[GY_IDX] =  ((int)(0.587 * 219 / 255 * (1 << RGB2YUV_SHIFT) + 0.5));
     802        1145 :         c->input_rgb2yuv_table[GV_IDX] = (-(int)(0.419 * 224 / 255 * (1 << RGB2YUV_SHIFT) + 0.5));
     803        1145 :         c->input_rgb2yuv_table[GU_IDX] = (-(int)(0.331 * 224 / 255 * (1 << RGB2YUV_SHIFT) + 0.5));
     804        1145 :         c->input_rgb2yuv_table[RY_IDX] =  ((int)(0.299 * 219 / 255 * (1 << RGB2YUV_SHIFT) + 0.5));
     805        1145 :         c->input_rgb2yuv_table[RV_IDX] =  ((int)(0.500 * 224 / 255 * (1 << RGB2YUV_SHIFT) + 0.5));
     806        1145 :         c->input_rgb2yuv_table[RU_IDX] = (-(int)(0.169 * 224 / 255 * (1 << RGB2YUV_SHIFT) + 0.5));
     807             :     }
     808      321745 :     for(i=0; i<FF_ARRAY_ELEMS(map); i++)
     809      320600 :         AV_WL16(p + 16*4 + 2*i, map[i] >= 0 ? c->input_rgb2yuv_table[map[i]] : 0);
     810        1145 : }
     811             : 
     812          77 : static void fill_xyztables(struct SwsContext *c)
     813             : {
     814             :     int i;
     815          77 :     double xyzgamma = XYZ_GAMMA;
     816          77 :     double rgbgamma = 1.0 / RGB_GAMMA;
     817          77 :     double xyzgammainv = 1.0 / XYZ_GAMMA;
     818          77 :     double rgbgammainv = RGB_GAMMA;
     819             :     static const int16_t xyz2rgb_matrix[3][4] = {
     820             :         {13270, -6295, -2041},
     821             :         {-3969,  7682,   170},
     822             :         {  228,  -835,  4329} };
     823             :     static const int16_t rgb2xyz_matrix[3][4] = {
     824             :         {1689, 1464,  739},
     825             :         { 871, 2929,  296},
     826             :         {  79,  488, 3891} };
     827             :     static int16_t xyzgamma_tab[4096], rgbgamma_tab[4096], xyzgammainv_tab[4096], rgbgammainv_tab[4096];
     828             : 
     829          77 :     memcpy(c->xyz2rgb_matrix, xyz2rgb_matrix, sizeof(c->xyz2rgb_matrix));
     830          77 :     memcpy(c->rgb2xyz_matrix, rgb2xyz_matrix, sizeof(c->rgb2xyz_matrix));
     831          77 :     c->xyzgamma = xyzgamma_tab;
     832          77 :     c->rgbgamma = rgbgamma_tab;
     833          77 :     c->xyzgammainv = xyzgammainv_tab;
     834          77 :     c->rgbgammainv = rgbgammainv_tab;
     835             : 
     836          77 :     if (rgbgamma_tab[4095])
     837          56 :         return;
     838             : 
     839             :     /* set gamma vectors */
     840       86037 :     for (i = 0; i < 4096; i++) {
     841       86016 :         xyzgamma_tab[i] = lrint(pow(i / 4095.0, xyzgamma) * 4095.0);
     842       86016 :         rgbgamma_tab[i] = lrint(pow(i / 4095.0, rgbgamma) * 4095.0);
     843       86016 :         xyzgammainv_tab[i] = lrint(pow(i / 4095.0, xyzgammainv) * 4095.0);
     844       86016 :         rgbgammainv_tab[i] = lrint(pow(i / 4095.0, rgbgammainv) * 4095.0);
     845             :     }
     846             : }
     847             : 
     848       40377 : int sws_setColorspaceDetails(struct SwsContext *c, const int inv_table[4],
     849             :                              int srcRange, const int table[4], int dstRange,
     850             :                              int brightness, int contrast, int saturation)
     851             : {
     852             :     const AVPixFmtDescriptor *desc_dst;
     853             :     const AVPixFmtDescriptor *desc_src;
     854       40377 :     int need_reinit = 0;
     855             : 
     856       40377 :     handle_formats(c);
     857       40377 :     desc_dst = av_pix_fmt_desc_get(c->dstFormat);
     858       40377 :     desc_src = av_pix_fmt_desc_get(c->srcFormat);
     859             : 
     860       40377 :     if(!isYUV(c->dstFormat) && !isGray(c->dstFormat))
     861       15861 :         dstRange = 0;
     862       40377 :     if(!isYUV(c->srcFormat) && !isGray(c->srcFormat))
     863       12375 :         srcRange = 0;
     864             : 
     865       80754 :     if (c->srcRange != srcRange ||
     866       80754 :         c->dstRange != dstRange ||
     867       80754 :         c->brightness != brightness ||
     868       77932 :         c->contrast   != contrast ||
     869       75110 :         c->saturation != saturation ||
     870       75025 :         memcmp(c->srcColorspaceTable, inv_table, sizeof(int) * 4) ||
     871       37470 :         memcmp(c->dstColorspaceTable,     table, sizeof(int) * 4)
     872             :     )
     873        2907 :         need_reinit = 1;
     874             : 
     875       40377 :     memmove(c->srcColorspaceTable, inv_table, sizeof(int) * 4);
     876       40377 :     memmove(c->dstColorspaceTable, table, sizeof(int) * 4);
     877             : 
     878             : 
     879             : 
     880       40377 :     c->brightness = brightness;
     881       40377 :     c->contrast   = contrast;
     882       40377 :     c->saturation = saturation;
     883       40377 :     c->srcRange   = srcRange;
     884       40377 :     c->dstRange   = dstRange;
     885             : 
     886             :     //The srcBpc check is possibly wrong but we seem to lack a definitive reference to test this
     887             :     //and what we have in ticket 2939 looks better with this check
     888       40377 :     if (need_reinit && (c->srcBpc == 8 || !isYUV(c->srcFormat)))
     889         416 :         ff_sws_init_range_convert(c);
     890             : 
     891       40377 :     c->dstFormatBpp = av_get_bits_per_pixel(desc_dst);
     892       40377 :     c->srcFormatBpp = av_get_bits_per_pixel(desc_src);
     893             : 
     894       40377 :     if (c->cascaded_context[c->cascaded_mainindex])
     895           0 :         return sws_setColorspaceDetails(c->cascaded_context[c->cascaded_mainindex],inv_table, srcRange,table, dstRange, brightness,  contrast, saturation);
     896             : 
     897       40377 :     if (!need_reinit)
     898       37470 :         return 0;
     899             : 
     900        2907 :     if ((isYUV(c->dstFormat) || isGray(c->dstFormat)) && (isYUV(c->srcFormat) || isGray(c->srcFormat))) {
     901        3524 :         if (!c->cascaded_context[0] &&
     902        1762 :             memcmp(c->dstColorspaceTable, c->srcColorspaceTable, sizeof(int) * 4) &&
     903           0 :             c->srcW && c->srcH && c->dstW && c->dstH) {
     904             :             enum AVPixelFormat tmp_format;
     905             :             int tmp_width, tmp_height;
     906           0 :             int srcW = c->srcW;
     907           0 :             int srcH = c->srcH;
     908           0 :             int dstW = c->dstW;
     909           0 :             int dstH = c->dstH;
     910             :             int ret;
     911           0 :             av_log(c, AV_LOG_VERBOSE, "YUV color matrix differs for YUV->YUV, using intermediate RGB to convert\n");
     912             : 
     913           0 :             if (isNBPS(c->dstFormat) || is16BPS(c->dstFormat)) {
     914           0 :                 if (isALPHA(c->srcFormat) && isALPHA(c->dstFormat)) {
     915           0 :                     tmp_format = AV_PIX_FMT_BGRA64;
     916             :                 } else {
     917           0 :                     tmp_format = AV_PIX_FMT_BGR48;
     918             :                 }
     919             :             } else {
     920           0 :                 if (isALPHA(c->srcFormat) && isALPHA(c->dstFormat)) {
     921           0 :                     tmp_format = AV_PIX_FMT_BGRA;
     922             :                 } else {
     923           0 :                     tmp_format = AV_PIX_FMT_BGR24;
     924             :                 }
     925             :             }
     926             : 
     927           0 :             if (srcW*srcH > dstW*dstH) {
     928           0 :                 tmp_width  = dstW;
     929           0 :                 tmp_height = dstH;
     930             :             } else {
     931           0 :                 tmp_width  = srcW;
     932           0 :                 tmp_height = srcH;
     933             :             }
     934             : 
     935           0 :             ret = av_image_alloc(c->cascaded_tmp, c->cascaded_tmpStride,
     936             :                                 tmp_width, tmp_height, tmp_format, 64);
     937           0 :             if (ret < 0)
     938           0 :                 return ret;
     939             : 
     940           0 :             c->cascaded_context[0] = sws_alloc_set_opts(srcW, srcH, c->srcFormat,
     941             :                                                         tmp_width, tmp_height, tmp_format,
     942           0 :                                                         c->flags, c->param);
     943           0 :             if (!c->cascaded_context[0])
     944           0 :                 return -1;
     945             : 
     946           0 :             c->cascaded_context[0]->alphablend = c->alphablend;
     947           0 :             ret = sws_init_context(c->cascaded_context[0], NULL , NULL);
     948           0 :             if (ret < 0)
     949           0 :                 return ret;
     950             :             //we set both src and dst depending on that the RGB side will be ignored
     951           0 :             sws_setColorspaceDetails(c->cascaded_context[0], inv_table,
     952             :                                      srcRange, table, dstRange,
     953             :                                      brightness, contrast, saturation);
     954             : 
     955           0 :             c->cascaded_context[1] = sws_getContext(tmp_width, tmp_height, tmp_format,
     956             :                                                     dstW, dstH, c->dstFormat,
     957           0 :                                                     c->flags, NULL, NULL, c->param);
     958           0 :             if (!c->cascaded_context[1])
     959           0 :                 return -1;
     960           0 :             sws_setColorspaceDetails(c->cascaded_context[1], inv_table,
     961             :                                      srcRange, table, dstRange,
     962             :                                      0, 1 << 16, 1 << 16);
     963           0 :             return 0;
     964             :         }
     965        1762 :         return -1;
     966             :     }
     967             : 
     968        1145 :     if (!isYUV(c->dstFormat) && !isGray(c->dstFormat)) {
     969        1029 :         ff_yuv2rgb_c_init_tables(c, inv_table, srcRange, brightness,
     970             :                                  contrast, saturation);
     971             :         // FIXME factorize
     972             : 
     973             :         if (ARCH_PPC)
     974             :             ff_yuv2rgb_init_tables_ppc(c, inv_table, brightness,
     975             :                                        contrast, saturation);
     976             :     }
     977             : 
     978        1145 :     fill_rgb2yuv_table(c, table, dstRange);
     979             : 
     980        1145 :     return 0;
     981             : }
     982             : 
     983       37555 : int sws_getColorspaceDetails(struct SwsContext *c, int **inv_table,
     984             :                              int *srcRange, int **table, int *dstRange,
     985             :                              int *brightness, int *contrast, int *saturation)
     986             : {
     987       37555 :     if (!c )
     988           0 :         return -1;
     989             : 
     990       37555 :     *inv_table  = c->srcColorspaceTable;
     991       37555 :     *table      = c->dstColorspaceTable;
     992       37555 :     *srcRange   = c->srcRange;
     993       37555 :     *dstRange   = c->dstRange;
     994       37555 :     *brightness = c->brightness;
     995       37555 :     *contrast   = c->contrast;
     996       37555 :     *saturation = c->saturation;
     997             : 
     998       37555 :     return 0;
     999             : }
    1000             : 
    1001        5644 : static int handle_jpeg(enum AVPixelFormat *format)
    1002             : {
    1003        5644 :     switch (*format) {
    1004          67 :     case AV_PIX_FMT_YUVJ420P:
    1005          67 :         *format = AV_PIX_FMT_YUV420P;
    1006          67 :         return 1;
    1007          15 :     case AV_PIX_FMT_YUVJ411P:
    1008          15 :         *format = AV_PIX_FMT_YUV411P;
    1009          15 :         return 1;
    1010          30 :     case AV_PIX_FMT_YUVJ422P:
    1011          30 :         *format = AV_PIX_FMT_YUV422P;
    1012          30 :         return 1;
    1013          38 :     case AV_PIX_FMT_YUVJ444P:
    1014          38 :         *format = AV_PIX_FMT_YUV444P;
    1015          38 :         return 1;
    1016          21 :     case AV_PIX_FMT_YUVJ440P:
    1017          21 :         *format = AV_PIX_FMT_YUV440P;
    1018          21 :         return 1;
    1019         164 :     case AV_PIX_FMT_GRAY8:
    1020             :     case AV_PIX_FMT_YA8:
    1021             :     case AV_PIX_FMT_GRAY9LE:
    1022             :     case AV_PIX_FMT_GRAY9BE:
    1023             :     case AV_PIX_FMT_GRAY10LE:
    1024             :     case AV_PIX_FMT_GRAY10BE:
    1025             :     case AV_PIX_FMT_GRAY12LE:
    1026             :     case AV_PIX_FMT_GRAY12BE:
    1027             :     case AV_PIX_FMT_GRAY16LE:
    1028             :     case AV_PIX_FMT_GRAY16BE:
    1029             :     case AV_PIX_FMT_YA16BE:
    1030             :     case AV_PIX_FMT_YA16LE:
    1031         164 :         return 1;
    1032        5309 :     default:
    1033        5309 :         return 0;
    1034             :     }
    1035             : }
    1036             : 
    1037       86398 : static int handle_0alpha(enum AVPixelFormat *format)
    1038             : {
    1039       86398 :     switch (*format) {
    1040          15 :     case AV_PIX_FMT_0BGR    : *format = AV_PIX_FMT_ABGR   ; return 1;
    1041          46 :     case AV_PIX_FMT_BGR0    : *format = AV_PIX_FMT_BGRA   ; return 4;
    1042          15 :     case AV_PIX_FMT_0RGB    : *format = AV_PIX_FMT_ARGB   ; return 1;
    1043          15 :     case AV_PIX_FMT_RGB0    : *format = AV_PIX_FMT_RGBA   ; return 4;
    1044       86307 :     default:                                          return 0;
    1045             :     }
    1046             : }
    1047             : 
    1048       86398 : static int handle_xyz(enum AVPixelFormat *format)
    1049             : {
    1050       86398 :     switch (*format) {
    1051          12 :     case AV_PIX_FMT_XYZ12BE : *format = AV_PIX_FMT_RGB48BE; return 1;
    1052          13 :     case AV_PIX_FMT_XYZ12LE : *format = AV_PIX_FMT_RGB48LE; return 1;
    1053       86373 :     default:                                                return 0;
    1054             :     }
    1055             : }
    1056             : 
    1057       43199 : static void handle_formats(SwsContext *c)
    1058             : {
    1059       43199 :     c->src0Alpha |= handle_0alpha(&c->srcFormat);
    1060       43199 :     c->dst0Alpha |= handle_0alpha(&c->dstFormat);
    1061       43199 :     c->srcXYZ    |= handle_xyz(&c->srcFormat);
    1062       43199 :     c->dstXYZ    |= handle_xyz(&c->dstFormat);
    1063       43199 :     if (c->srcXYZ || c->dstXYZ)
    1064          77 :         fill_xyztables(c);
    1065       43199 : }
    1066             : 
    1067        9078 : SwsContext *sws_alloc_context(void)
    1068             : {
    1069        9078 :     SwsContext *c = av_mallocz(sizeof(SwsContext));
    1070             : 
    1071             :     av_assert0(offsetof(SwsContext, redDither) + DITHER32_INT == offsetof(SwsContext, dither32));
    1072             : 
    1073        9078 :     if (c) {
    1074        9078 :         c->av_class = &ff_sws_context_class;
    1075        9078 :         av_opt_set_defaults(c);
    1076             :     }
    1077             : 
    1078        9078 :     return c;
    1079             : }
    1080             : 
    1081           0 : static uint16_t * alloc_gamma_tbl(double e)
    1082             : {
    1083           0 :     int i = 0;
    1084             :     uint16_t * tbl;
    1085           0 :     tbl = (uint16_t*)av_malloc(sizeof(uint16_t) * 1 << 16);
    1086           0 :     if (!tbl)
    1087           0 :         return NULL;
    1088             : 
    1089           0 :     for (i = 0; i < 65536; ++i) {
    1090           0 :         tbl[i] = pow(i / 65535.0, e) * 65535.0;
    1091             :     }
    1092           0 :     return tbl;
    1093             : }
    1094             : 
    1095         131 : static enum AVPixelFormat alphaless_fmt(enum AVPixelFormat fmt)
    1096             : {
    1097         131 :     switch(fmt) {
    1098           1 :     case AV_PIX_FMT_ARGB:       return AV_PIX_FMT_RGB24;
    1099           1 :     case AV_PIX_FMT_RGBA:       return AV_PIX_FMT_RGB24;
    1100           0 :     case AV_PIX_FMT_ABGR:       return AV_PIX_FMT_BGR24;
    1101          29 :     case AV_PIX_FMT_BGRA:       return AV_PIX_FMT_BGR24;
    1102           2 :     case AV_PIX_FMT_YA8:        return AV_PIX_FMT_GRAY8;
    1103             : 
    1104           0 :     case AV_PIX_FMT_YUVA420P:   return AV_PIX_FMT_YUV420P;
    1105           3 :     case AV_PIX_FMT_YUVA422P:   return AV_PIX_FMT_YUV422P;
    1106           4 :     case AV_PIX_FMT_YUVA444P:           return AV_PIX_FMT_YUV444P;
    1107             : 
    1108           5 :     case AV_PIX_FMT_GBRAP:              return AV_PIX_FMT_GBRP;
    1109             : 
    1110           0 :     case AV_PIX_FMT_GBRAP10LE:          return AV_PIX_FMT_GBRP10;
    1111           0 :     case AV_PIX_FMT_GBRAP10BE:          return AV_PIX_FMT_GBRP10;
    1112             : 
    1113           0 :     case AV_PIX_FMT_GBRAP12LE:          return AV_PIX_FMT_GBRP12;
    1114           0 :     case AV_PIX_FMT_GBRAP12BE:          return AV_PIX_FMT_GBRP12;
    1115             : 
    1116           0 :     case AV_PIX_FMT_GBRAP16LE:          return AV_PIX_FMT_GBRP16;
    1117           1 :     case AV_PIX_FMT_GBRAP16BE:          return AV_PIX_FMT_GBRP16;
    1118             : 
    1119           0 :     case AV_PIX_FMT_RGBA64LE:   return AV_PIX_FMT_RGB48;
    1120           1 :     case AV_PIX_FMT_RGBA64BE:   return AV_PIX_FMT_RGB48;
    1121           0 :     case AV_PIX_FMT_BGRA64LE:   return AV_PIX_FMT_BGR48;
    1122           0 :     case AV_PIX_FMT_BGRA64BE:   return AV_PIX_FMT_BGR48;
    1123             : 
    1124           2 :     case AV_PIX_FMT_YA16BE:             return AV_PIX_FMT_GRAY16;
    1125           0 :     case AV_PIX_FMT_YA16LE:             return AV_PIX_FMT_GRAY16;
    1126             : 
    1127           0 :     case AV_PIX_FMT_YUVA420P9BE:        return AV_PIX_FMT_YUV420P9;
    1128           0 :     case AV_PIX_FMT_YUVA422P9BE:        return AV_PIX_FMT_YUV422P9;
    1129           0 :     case AV_PIX_FMT_YUVA444P9BE:        return AV_PIX_FMT_YUV444P9;
    1130           0 :     case AV_PIX_FMT_YUVA420P9LE:        return AV_PIX_FMT_YUV420P9;
    1131           0 :     case AV_PIX_FMT_YUVA422P9LE:        return AV_PIX_FMT_YUV422P9;
    1132           0 :     case AV_PIX_FMT_YUVA444P9LE:        return AV_PIX_FMT_YUV444P9;
    1133           0 :     case AV_PIX_FMT_YUVA420P10BE:       return AV_PIX_FMT_YUV420P10;
    1134           0 :     case AV_PIX_FMT_YUVA422P10BE:       return AV_PIX_FMT_YUV422P10;
    1135           0 :     case AV_PIX_FMT_YUVA444P10BE:       return AV_PIX_FMT_YUV444P10;
    1136           0 :     case AV_PIX_FMT_YUVA420P10LE:       return AV_PIX_FMT_YUV420P10;
    1137           0 :     case AV_PIX_FMT_YUVA422P10LE:       return AV_PIX_FMT_YUV422P10;
    1138           0 :     case AV_PIX_FMT_YUVA444P10LE:       return AV_PIX_FMT_YUV444P10;
    1139           0 :     case AV_PIX_FMT_YUVA420P16BE:       return AV_PIX_FMT_YUV420P16;
    1140           0 :     case AV_PIX_FMT_YUVA422P16BE:       return AV_PIX_FMT_YUV422P16;
    1141           0 :     case AV_PIX_FMT_YUVA444P16BE:       return AV_PIX_FMT_YUV444P16;
    1142           0 :     case AV_PIX_FMT_YUVA420P16LE:       return AV_PIX_FMT_YUV420P16;
    1143           1 :     case AV_PIX_FMT_YUVA422P16LE:       return AV_PIX_FMT_YUV422P16;
    1144           0 :     case AV_PIX_FMT_YUVA444P16LE:       return AV_PIX_FMT_YUV444P16;
    1145             : 
    1146             : //     case AV_PIX_FMT_AYUV64LE:
    1147             : //     case AV_PIX_FMT_AYUV64BE:
    1148             : //     case AV_PIX_FMT_PAL8:
    1149          81 :     default: return AV_PIX_FMT_NONE;
    1150             :     }
    1151             : }
    1152             : 
    1153        2822 : av_cold int sws_init_context(SwsContext *c, SwsFilter *srcFilter,
    1154             :                              SwsFilter *dstFilter)
    1155             : {
    1156             :     int i;
    1157             :     int usesVFilter, usesHFilter;
    1158             :     int unscaled;
    1159        2822 :     SwsFilter dummyFilter = { NULL, NULL, NULL, NULL };
    1160        2822 :     int srcW              = c->srcW;
    1161        2822 :     int srcH              = c->srcH;
    1162        2822 :     int dstW              = c->dstW;
    1163        2822 :     int dstH              = c->dstH;
    1164        2822 :     int dst_stride        = FFALIGN(dstW * sizeof(int16_t) + 66, 16);
    1165             :     int flags, cpu_flags;
    1166        2822 :     enum AVPixelFormat srcFormat = c->srcFormat;
    1167        2822 :     enum AVPixelFormat dstFormat = c->dstFormat;
    1168             :     const AVPixFmtDescriptor *desc_src;
    1169             :     const AVPixFmtDescriptor *desc_dst;
    1170        2822 :     int ret = 0;
    1171             :     enum AVPixelFormat tmpFmt;
    1172             : 
    1173        2822 :     cpu_flags = av_get_cpu_flags();
    1174        2822 :     flags     = c->flags;
    1175        2822 :     emms_c();
    1176        2822 :     if (!rgb15to16)
    1177        2620 :         ff_sws_rgb2rgb_init();
    1178             : 
    1179        2822 :     unscaled = (srcW == dstW && srcH == dstH);
    1180             : 
    1181        2822 :     c->srcRange |= handle_jpeg(&c->srcFormat);
    1182        2822 :     c->dstRange |= handle_jpeg(&c->dstFormat);
    1183             : 
    1184        2822 :     if(srcFormat!=c->srcFormat || dstFormat!=c->dstFormat)
    1185         166 :         av_log(c, AV_LOG_WARNING, "deprecated pixel format used, make sure you did set range correctly\n");
    1186             : 
    1187        2822 :     if (!c->contrast && !c->saturation && !c->dstFormatBpp)
    1188        2822 :         sws_setColorspaceDetails(c, ff_yuv2rgb_coeffs[SWS_CS_DEFAULT], c->srcRange,
    1189             :                                  ff_yuv2rgb_coeffs[SWS_CS_DEFAULT],
    1190             :                                  c->dstRange, 0, 1 << 16, 1 << 16);
    1191             : 
    1192        2822 :     handle_formats(c);
    1193        2822 :     srcFormat = c->srcFormat;
    1194        2822 :     dstFormat = c->dstFormat;
    1195        2822 :     desc_src = av_pix_fmt_desc_get(srcFormat);
    1196        2822 :     desc_dst = av_pix_fmt_desc_get(dstFormat);
    1197             : 
    1198             :     // If the source has no alpha then disable alpha blendaway
    1199        2822 :     if (c->src0Alpha)
    1200          25 :         c->alphablend = SWS_ALPHA_BLEND_NONE;
    1201             : 
    1202        2824 :     if (!(unscaled && sws_isSupportedEndiannessConversion(srcFormat) &&
    1203           2 :           av_pix_fmt_swap_endianness(srcFormat) == dstFormat)) {
    1204        2822 :     if (!sws_isSupportedInput(srcFormat)) {
    1205           0 :         av_log(c, AV_LOG_ERROR, "%s is not supported as input pixel format\n",
    1206             :                av_get_pix_fmt_name(srcFormat));
    1207           0 :         return AVERROR(EINVAL);
    1208             :     }
    1209        2822 :     if (!sws_isSupportedOutput(dstFormat)) {
    1210           0 :         av_log(c, AV_LOG_ERROR, "%s is not supported as output pixel format\n",
    1211             :                av_get_pix_fmt_name(dstFormat));
    1212           0 :         return AVERROR(EINVAL);
    1213             :     }
    1214             :     }
    1215             :     av_assert2(desc_src && desc_dst);
    1216             : 
    1217        2822 :     i = flags & (SWS_POINT         |
    1218             :                  SWS_AREA          |
    1219             :                  SWS_BILINEAR      |
    1220             :                  SWS_FAST_BILINEAR |
    1221             :                  SWS_BICUBIC       |
    1222             :                  SWS_X             |
    1223             :                  SWS_GAUSS         |
    1224             :                  SWS_LANCZOS       |
    1225             :                  SWS_SINC          |
    1226             :                  SWS_SPLINE        |
    1227             :                  SWS_BICUBLIN);
    1228             : 
    1229             :     /* provide a default scaler if not set by caller */
    1230        2822 :     if (!i) {
    1231          30 :         if (dstW < srcW && dstH < srcH)
    1232           8 :             flags |= SWS_BICUBIC;
    1233          22 :         else if (dstW > srcW && dstH > srcH)
    1234           1 :             flags |= SWS_BICUBIC;
    1235             :         else
    1236          21 :             flags |= SWS_BICUBIC;
    1237          30 :         c->flags = flags;
    1238        2792 :     } else if (i & (i - 1)) {
    1239           0 :         av_log(c, AV_LOG_ERROR,
    1240             :                "Exactly one scaler algorithm must be chosen, got %X\n", i);
    1241           0 :         return AVERROR(EINVAL);
    1242             :     }
    1243             :     /* sanity check */
    1244        2822 :     if (srcW < 1 || srcH < 1 || dstW < 1 || dstH < 1) {
    1245             :         /* FIXME check if these are enough and try to lower them after
    1246             :          * fixing the relevant parts of the code */
    1247           0 :         av_log(c, AV_LOG_ERROR, "%dx%d -> %dx%d is invalid scaling dimension\n",
    1248             :                srcW, srcH, dstW, dstH);
    1249           0 :         return AVERROR(EINVAL);
    1250             :     }
    1251        2822 :     if (flags & SWS_FAST_BILINEAR) {
    1252           0 :         if (srcW < 8 || dstW < 8) {
    1253           0 :             flags ^= SWS_FAST_BILINEAR | SWS_BILINEAR;
    1254           0 :             c->flags = flags;
    1255             :         }
    1256             :     }
    1257             : 
    1258        2822 :     if (!dstFilter)
    1259        2822 :         dstFilter = &dummyFilter;
    1260        2822 :     if (!srcFilter)
    1261        2822 :         srcFilter = &dummyFilter;
    1262             : 
    1263        2822 :     c->lumXInc      = (((int64_t)srcW << 16) + (dstW >> 1)) / dstW;
    1264        2822 :     c->lumYInc      = (((int64_t)srcH << 16) + (dstH >> 1)) / dstH;
    1265        2822 :     c->dstFormatBpp = av_get_bits_per_pixel(desc_dst);
    1266        2822 :     c->srcFormatBpp = av_get_bits_per_pixel(desc_src);
    1267        2822 :     c->vRounder     = 4 * 0x0001000100010001ULL;
    1268             : 
    1269        8466 :     usesVFilter = (srcFilter->lumV && srcFilter->lumV->length > 1) ||
    1270        5644 :                   (srcFilter->chrV && srcFilter->chrV->length > 1) ||
    1271        8466 :                   (dstFilter->lumV && dstFilter->lumV->length > 1) ||
    1272        2822 :                   (dstFilter->chrV && dstFilter->chrV->length > 1);
    1273        8466 :     usesHFilter = (srcFilter->lumH && srcFilter->lumH->length > 1) ||
    1274        5644 :                   (srcFilter->chrH && srcFilter->chrH->length > 1) ||
    1275        8466 :                   (dstFilter->lumH && dstFilter->lumH->length > 1) ||
    1276        2822 :                   (dstFilter->chrH && dstFilter->chrH->length > 1);
    1277             : 
    1278        2822 :     av_pix_fmt_get_chroma_sub_sample(srcFormat, &c->chrSrcHSubSample, &c->chrSrcVSubSample);
    1279        2822 :     av_pix_fmt_get_chroma_sub_sample(dstFormat, &c->chrDstHSubSample, &c->chrDstVSubSample);
    1280             : 
    1281        2822 :     if (isAnyRGB(dstFormat) && !(flags&SWS_FULL_CHR_H_INT)) {
    1282        1017 :         if (dstW&1) {
    1283          28 :             av_log(c, AV_LOG_DEBUG, "Forcing full internal H chroma due to odd output size\n");
    1284          28 :             flags |= SWS_FULL_CHR_H_INT;
    1285          28 :             c->flags = flags;
    1286             :         }
    1287             : 
    1288        1017 :         if (   c->chrSrcHSubSample == 0
    1289         237 :             && c->chrSrcVSubSample == 0
    1290         237 :             && c->dither != SWS_DITHER_BAYER //SWS_FULL_CHR_H_INT is currently not supported with SWS_DITHER_BAYER
    1291         237 :             && !(c->flags & SWS_FAST_BILINEAR)
    1292             :         ) {
    1293         237 :             av_log(c, AV_LOG_DEBUG, "Forcing full internal H chroma due to input having non subsampled chroma\n");
    1294         237 :             flags |= SWS_FULL_CHR_H_INT;
    1295         237 :             c->flags = flags;
    1296             :         }
    1297             :     }
    1298             : 
    1299        2822 :     if (c->dither == SWS_DITHER_AUTO) {
    1300        2822 :         if (flags & SWS_ERROR_DIFFUSION)
    1301           0 :             c->dither = SWS_DITHER_ED;
    1302             :     }
    1303             : 
    1304        2822 :     if(dstFormat == AV_PIX_FMT_BGR4_BYTE ||
    1305        2797 :        dstFormat == AV_PIX_FMT_RGB4_BYTE ||
    1306        2776 :        dstFormat == AV_PIX_FMT_BGR8 ||
    1307             :        dstFormat == AV_PIX_FMT_RGB8) {
    1308          60 :         if (c->dither == SWS_DITHER_AUTO)
    1309          60 :             c->dither = (flags & SWS_FULL_CHR_H_INT) ? SWS_DITHER_ED : SWS_DITHER_BAYER;
    1310          60 :         if (!(flags & SWS_FULL_CHR_H_INT)) {
    1311          50 :             if (c->dither == SWS_DITHER_ED || c->dither == SWS_DITHER_A_DITHER || c->dither == SWS_DITHER_X_DITHER) {
    1312           0 :                 av_log(c, AV_LOG_DEBUG,
    1313             :                     "Desired dithering only supported in full chroma interpolation for destination format '%s'\n",
    1314             :                     av_get_pix_fmt_name(dstFormat));
    1315           0 :                 flags   |= SWS_FULL_CHR_H_INT;
    1316           0 :                 c->flags = flags;
    1317             :             }
    1318             :         }
    1319          60 :         if (flags & SWS_FULL_CHR_H_INT) {
    1320          10 :             if (c->dither == SWS_DITHER_BAYER) {
    1321           0 :                 av_log(c, AV_LOG_DEBUG,
    1322             :                     "Ordered dither is not supported in full chroma interpolation for destination format '%s'\n",
    1323             :                     av_get_pix_fmt_name(dstFormat));
    1324           0 :                 c->dither = SWS_DITHER_ED;
    1325             :             }
    1326             :         }
    1327             :     }
    1328        2822 :     if (isPlanarRGB(dstFormat)) {
    1329         242 :         if (!(flags & SWS_FULL_CHR_H_INT)) {
    1330         209 :             av_log(c, AV_LOG_DEBUG,
    1331             :                    "%s output is not supported with half chroma resolution, switching to full\n",
    1332             :                    av_get_pix_fmt_name(dstFormat));
    1333         209 :             flags   |= SWS_FULL_CHR_H_INT;
    1334         209 :             c->flags = flags;
    1335             :         }
    1336             :     }
    1337             : 
    1338             :     /* reuse chroma for 2 pixels RGB/BGR unless user wants full
    1339             :      * chroma interpolation */
    1340        3280 :     if (flags & SWS_FULL_CHR_H_INT &&
    1341         916 :         isAnyRGB(dstFormat)        &&
    1342         674 :         !isPlanarRGB(dstFormat)    &&
    1343         215 :         dstFormat != AV_PIX_FMT_RGBA64LE &&
    1344         214 :         dstFormat != AV_PIX_FMT_RGBA64BE &&
    1345         213 :         dstFormat != AV_PIX_FMT_BGRA64LE &&
    1346         212 :         dstFormat != AV_PIX_FMT_BGRA64BE &&
    1347         209 :         dstFormat != AV_PIX_FMT_RGB48LE &&
    1348         206 :         dstFormat != AV_PIX_FMT_RGB48BE &&
    1349         205 :         dstFormat != AV_PIX_FMT_BGR48LE &&
    1350         204 :         dstFormat != AV_PIX_FMT_BGR48BE &&
    1351         197 :         dstFormat != AV_PIX_FMT_RGBA  &&
    1352         194 :         dstFormat != AV_PIX_FMT_ARGB  &&
    1353         160 :         dstFormat != AV_PIX_FMT_BGRA  &&
    1354         158 :         dstFormat != AV_PIX_FMT_ABGR  &&
    1355          35 :         dstFormat != AV_PIX_FMT_RGB24 &&
    1356          24 :         dstFormat != AV_PIX_FMT_BGR24 &&
    1357          22 :         dstFormat != AV_PIX_FMT_BGR4_BYTE &&
    1358          20 :         dstFormat != AV_PIX_FMT_RGB4_BYTE &&
    1359          16 :         dstFormat != AV_PIX_FMT_BGR8 &&
    1360             :         dstFormat != AV_PIX_FMT_RGB8
    1361             :     ) {
    1362          14 :         av_log(c, AV_LOG_WARNING,
    1363             :                "full chroma interpolation for destination format '%s' not yet implemented\n",
    1364             :                av_get_pix_fmt_name(dstFormat));
    1365          14 :         flags   &= ~SWS_FULL_CHR_H_INT;
    1366          14 :         c->flags = flags;
    1367             :     }
    1368        2822 :     if (isAnyRGB(dstFormat) && !(flags & SWS_FULL_CHR_H_INT))
    1369         585 :         c->chrDstHSubSample = 1;
    1370             : 
    1371             :     // drop some chroma lines if the user wants it
    1372        2822 :     c->vChrDrop          = (flags & SWS_SRC_V_CHR_DROP_MASK) >>
    1373             :                            SWS_SRC_V_CHR_DROP_SHIFT;
    1374        2822 :     c->chrSrcVSubSample += c->vChrDrop;
    1375             : 
    1376             :     /* drop every other pixel for chroma calculation unless user
    1377             :      * wants full chroma */
    1378        2822 :     if (isAnyRGB(srcFormat) && !(flags & SWS_FULL_CHR_H_INP)   &&
    1379         250 :         srcFormat != AV_PIX_FMT_RGB8 && srcFormat != AV_PIX_FMT_BGR8 &&
    1380         249 :         srcFormat != AV_PIX_FMT_RGB4 && srcFormat != AV_PIX_FMT_BGR4 &&
    1381         248 :         srcFormat != AV_PIX_FMT_RGB4_BYTE && srcFormat != AV_PIX_FMT_BGR4_BYTE &&
    1382         246 :         srcFormat != AV_PIX_FMT_GBRP9BE   && srcFormat != AV_PIX_FMT_GBRP9LE  &&
    1383         244 :         srcFormat != AV_PIX_FMT_GBRP10BE  && srcFormat != AV_PIX_FMT_GBRP10LE &&
    1384         241 :         srcFormat != AV_PIX_FMT_GBRAP10BE && srcFormat != AV_PIX_FMT_GBRAP10LE &&
    1385         239 :         srcFormat != AV_PIX_FMT_GBRP12BE  && srcFormat != AV_PIX_FMT_GBRP12LE &&
    1386         237 :         srcFormat != AV_PIX_FMT_GBRAP12BE && srcFormat != AV_PIX_FMT_GBRAP12LE &&
    1387         235 :         srcFormat != AV_PIX_FMT_GBRP14BE  && srcFormat != AV_PIX_FMT_GBRP14LE &&
    1388         232 :         srcFormat != AV_PIX_FMT_GBRP16BE  && srcFormat != AV_PIX_FMT_GBRP16LE &&
    1389         446 :         srcFormat != AV_PIX_FMT_GBRAP16BE  && srcFormat != AV_PIX_FMT_GBRAP16LE &&
    1390         341 :         ((dstW >> c->chrDstHSubSample) <= (srcW >> 1) ||
    1391         119 :          (flags & SWS_FAST_BILINEAR)))
    1392         103 :         c->chrSrcHSubSample = 1;
    1393             : 
    1394             :     // Note the AV_CEIL_RSHIFT is so that we always round toward +inf.
    1395        2822 :     c->chrSrcW = AV_CEIL_RSHIFT(srcW, c->chrSrcHSubSample);
    1396        2822 :     c->chrSrcH = AV_CEIL_RSHIFT(srcH, c->chrSrcVSubSample);
    1397        2822 :     c->chrDstW = AV_CEIL_RSHIFT(dstW, c->chrDstHSubSample);
    1398        2822 :     c->chrDstH = AV_CEIL_RSHIFT(dstH, c->chrDstVSubSample);
    1399             : 
    1400        2822 :     FF_ALLOCZ_OR_GOTO(c, c->formatConvBuffer, FFALIGN(srcW*2+78, 16) * 2, fail);
    1401             : 
    1402        2822 :     c->srcBpc = desc_src->comp[0].depth;
    1403        2822 :     if (c->srcBpc < 8)
    1404          56 :         c->srcBpc = 8;
    1405        2822 :     c->dstBpc = desc_dst->comp[0].depth;
    1406        2822 :     if (c->dstBpc < 8)
    1407         249 :         c->dstBpc = 8;
    1408        2822 :     if (isAnyRGB(srcFormat) || srcFormat == AV_PIX_FMT_PAL8)
    1409         340 :         c->srcBpc = 16;
    1410        2822 :     if (c->dstBpc == 16)
    1411         392 :         dst_stride <<= 1;
    1412             : 
    1413        2822 :     if (INLINE_MMXEXT(cpu_flags) && c->srcBpc == 8 && c->dstBpc <= 14) {
    1414           6 :         c->canMMXEXTBeUsed = dstW >= srcW && (dstW & 31) == 0 &&
    1415           6 :                              c->chrDstW >= c->chrSrcW &&
    1416           2 :                              (srcW & 15) == 0;
    1417           2 :         if (!c->canMMXEXTBeUsed && dstW >= srcW && c->chrDstW >= c->chrSrcW && (srcW & 15) == 0
    1418             : 
    1419           0 :             && (flags & SWS_FAST_BILINEAR)) {
    1420           0 :             if (flags & SWS_PRINT_INFO)
    1421           0 :                 av_log(c, AV_LOG_INFO,
    1422             :                        "output width is not a multiple of 32 -> no MMXEXT scaler\n");
    1423             :         }
    1424           4 :         if (usesHFilter || isNBPS(c->srcFormat) || is16BPS(c->srcFormat) || isAnyRGB(c->srcFormat))
    1425           0 :             c->canMMXEXTBeUsed = 0;
    1426             :     } else
    1427        2820 :         c->canMMXEXTBeUsed = 0;
    1428             : 
    1429        2822 :     c->chrXInc = (((int64_t)c->chrSrcW << 16) + (c->chrDstW >> 1)) / c->chrDstW;
    1430        2822 :     c->chrYInc = (((int64_t)c->chrSrcH << 16) + (c->chrDstH >> 1)) / c->chrDstH;
    1431             : 
    1432             :     /* Match pixel 0 of the src to pixel 0 of dst and match pixel n-2 of src
    1433             :      * to pixel n-2 of dst, but only for the FAST_BILINEAR mode otherwise do
    1434             :      * correct scaling.
    1435             :      * n-2 is the last chrominance sample available.
    1436             :      * This is not perfect, but no one should notice the difference, the more
    1437             :      * correct variant would be like the vertical one, but that would require
    1438             :      * some special code for the first and last pixel */
    1439        2822 :     if (flags & SWS_FAST_BILINEAR) {
    1440           0 :         if (c->canMMXEXTBeUsed) {
    1441           0 :             c->lumXInc += 20;
    1442           0 :             c->chrXInc += 20;
    1443             :         }
    1444             :         // we don't use the x86 asm scaler if MMX is available
    1445           0 :         else if (INLINE_MMX(cpu_flags) && c->dstBpc <= 14) {
    1446           0 :             c->lumXInc = ((int64_t)(srcW       - 2) << 16) / (dstW       - 2) - 20;
    1447           0 :             c->chrXInc = ((int64_t)(c->chrSrcW - 2) << 16) / (c->chrDstW - 2) - 20;
    1448             :         }
    1449             :     }
    1450             : 
    1451             :     // hardcoded for now
    1452        2822 :     c->gamma_value = 2.2;
    1453        2822 :     tmpFmt = AV_PIX_FMT_RGBA64LE;
    1454             : 
    1455             : 
    1456        2822 :     if (!unscaled && c->gamma_flag && (srcFormat != tmpFmt || dstFormat != tmpFmt)) {
    1457             :         SwsContext *c2;
    1458           0 :         c->cascaded_context[0] = NULL;
    1459             : 
    1460           0 :         ret = av_image_alloc(c->cascaded_tmp, c->cascaded_tmpStride,
    1461             :                             srcW, srcH, tmpFmt, 64);
    1462           0 :         if (ret < 0)
    1463           0 :             return ret;
    1464             : 
    1465           0 :         c->cascaded_context[0] = sws_getContext(srcW, srcH, srcFormat,
    1466             :                                                 srcW, srcH, tmpFmt,
    1467           0 :                                                 flags, NULL, NULL, c->param);
    1468           0 :         if (!c->cascaded_context[0]) {
    1469           0 :             return -1;
    1470             :         }
    1471             : 
    1472           0 :         c->cascaded_context[1] = sws_getContext(srcW, srcH, tmpFmt,
    1473             :                                                 dstW, dstH, tmpFmt,
    1474           0 :                                                 flags, srcFilter, dstFilter, c->param);
    1475             : 
    1476           0 :         if (!c->cascaded_context[1])
    1477           0 :             return -1;
    1478             : 
    1479           0 :         c2 = c->cascaded_context[1];
    1480           0 :         c2->is_internal_gamma = 1;
    1481           0 :         c2->gamma     = alloc_gamma_tbl(    c->gamma_value);
    1482           0 :         c2->inv_gamma = alloc_gamma_tbl(1.f/c->gamma_value);
    1483           0 :         if (!c2->gamma || !c2->inv_gamma)
    1484           0 :             return AVERROR(ENOMEM);
    1485             : 
    1486             :         // is_internal_flag is set after creating the context
    1487             :         // to properly create the gamma convert FilterDescriptor
    1488             :         // we have to re-initialize it
    1489           0 :         ff_free_filters(c2);
    1490           0 :         if (ff_init_filters(c2) < 0) {
    1491           0 :             sws_freeContext(c2);
    1492           0 :             return -1;
    1493             :         }
    1494             : 
    1495           0 :         c->cascaded_context[2] = NULL;
    1496           0 :         if (dstFormat != tmpFmt) {
    1497           0 :             ret = av_image_alloc(c->cascaded1_tmp, c->cascaded1_tmpStride,
    1498             :                                 dstW, dstH, tmpFmt, 64);
    1499           0 :             if (ret < 0)
    1500           0 :                 return ret;
    1501             : 
    1502           0 :             c->cascaded_context[2] = sws_getContext(dstW, dstH, tmpFmt,
    1503             :                                                 dstW, dstH, dstFormat,
    1504           0 :                                                 flags, NULL, NULL, c->param);
    1505           0 :             if (!c->cascaded_context[2])
    1506           0 :                 return -1;
    1507             :         }
    1508           0 :         return 0;
    1509             :     }
    1510             : 
    1511        2822 :     if (isBayer(srcFormat)) {
    1512           0 :         if (!unscaled ||
    1513           0 :             (dstFormat != AV_PIX_FMT_RGB24 && dstFormat != AV_PIX_FMT_YUV420P)) {
    1514           0 :             enum AVPixelFormat tmpFormat = AV_PIX_FMT_RGB24;
    1515             : 
    1516           0 :             ret = av_image_alloc(c->cascaded_tmp, c->cascaded_tmpStride,
    1517             :                                 srcW, srcH, tmpFormat, 64);
    1518           0 :             if (ret < 0)
    1519           0 :                 return ret;
    1520             : 
    1521           0 :             c->cascaded_context[0] = sws_getContext(srcW, srcH, srcFormat,
    1522             :                                                     srcW, srcH, tmpFormat,
    1523           0 :                                                     flags, srcFilter, NULL, c->param);
    1524           0 :             if (!c->cascaded_context[0])
    1525           0 :                 return -1;
    1526             : 
    1527           0 :             c->cascaded_context[1] = sws_getContext(srcW, srcH, tmpFormat,
    1528             :                                                     dstW, dstH, dstFormat,
    1529           0 :                                                     flags, NULL, dstFilter, c->param);
    1530           0 :             if (!c->cascaded_context[1])
    1531           0 :                 return -1;
    1532           0 :             return 0;
    1533             :         }
    1534             :     }
    1535             : 
    1536        2822 :     if (CONFIG_SWSCALE_ALPHA && isALPHA(srcFormat) && !isALPHA(dstFormat)) {
    1537         131 :         enum AVPixelFormat tmpFormat = alphaless_fmt(srcFormat);
    1538             : 
    1539         131 :         if (tmpFormat != AV_PIX_FMT_NONE && c->alphablend != SWS_ALPHA_BLEND_NONE)
    1540           0 :         if (!unscaled ||
    1541           0 :             dstFormat != tmpFormat ||
    1542           0 :             usesHFilter || usesVFilter ||
    1543           0 :             c->srcRange != c->dstRange
    1544             :         ) {
    1545           0 :             c->cascaded_mainindex = 1;
    1546           0 :             ret = av_image_alloc(c->cascaded_tmp, c->cascaded_tmpStride,
    1547             :                                 srcW, srcH, tmpFormat, 64);
    1548           0 :             if (ret < 0)
    1549           0 :                 return ret;
    1550             : 
    1551           0 :             c->cascaded_context[0] = sws_alloc_set_opts(srcW, srcH, srcFormat,
    1552             :                                                         srcW, srcH, tmpFormat,
    1553           0 :                                                         flags, c->param);
    1554           0 :             if (!c->cascaded_context[0])
    1555           0 :                 return -1;
    1556           0 :             c->cascaded_context[0]->alphablend = c->alphablend;
    1557           0 :             ret = sws_init_context(c->cascaded_context[0], NULL , NULL);
    1558           0 :             if (ret < 0)
    1559           0 :                 return ret;
    1560             : 
    1561           0 :             c->cascaded_context[1] = sws_alloc_set_opts(srcW, srcH, tmpFormat,
    1562             :                                                         dstW, dstH, dstFormat,
    1563           0 :                                                         flags, c->param);
    1564           0 :             if (!c->cascaded_context[1])
    1565           0 :                 return -1;
    1566             : 
    1567           0 :             c->cascaded_context[1]->srcRange = c->srcRange;
    1568           0 :             c->cascaded_context[1]->dstRange = c->dstRange;
    1569           0 :             ret = sws_init_context(c->cascaded_context[1], srcFilter , dstFilter);
    1570           0 :             if (ret < 0)
    1571           0 :                 return ret;
    1572             : 
    1573           0 :             return 0;
    1574             :         }
    1575             :     }
    1576             : 
    1577             : #if HAVE_MMAP && HAVE_MPROTECT && defined(MAP_ANONYMOUS)
    1578             : #define USE_MMAP 1
    1579             : #else
    1580             : #define USE_MMAP 0
    1581             : #endif
    1582             : 
    1583             :     /* precalculate horizontal scaler filter coefficients */
    1584             :     {
    1585             : #if HAVE_MMXEXT_INLINE
    1586             : // can't downscale !!!
    1587        2822 :         if (c->canMMXEXTBeUsed && (flags & SWS_FAST_BILINEAR)) {
    1588           0 :             c->lumMmxextFilterCodeSize = ff_init_hscaler_mmxext(dstW, c->lumXInc, NULL,
    1589             :                                                              NULL, NULL, 8);
    1590           0 :             c->chrMmxextFilterCodeSize = ff_init_hscaler_mmxext(c->chrDstW, c->chrXInc,
    1591             :                                                              NULL, NULL, NULL, 4);
    1592             : 
    1593             : #if USE_MMAP
    1594           0 :             c->lumMmxextFilterCode = mmap(NULL, c->lumMmxextFilterCodeSize,
    1595             :                                           PROT_READ | PROT_WRITE,
    1596             :                                           MAP_PRIVATE | MAP_ANONYMOUS,
    1597             :                                           -1, 0);
    1598           0 :             c->chrMmxextFilterCode = mmap(NULL, c->chrMmxextFilterCodeSize,
    1599             :                                           PROT_READ | PROT_WRITE,
    1600             :                                           MAP_PRIVATE | MAP_ANONYMOUS,
    1601             :                                           -1, 0);
    1602             : #elif HAVE_VIRTUALALLOC
    1603             :             c->lumMmxextFilterCode = VirtualAlloc(NULL,
    1604             :                                                   c->lumMmxextFilterCodeSize,
    1605             :                                                   MEM_COMMIT,
    1606             :                                                   PAGE_EXECUTE_READWRITE);
    1607             :             c->chrMmxextFilterCode = VirtualAlloc(NULL,
    1608             :                                                   c->chrMmxextFilterCodeSize,
    1609             :                                                   MEM_COMMIT,
    1610             :                                                   PAGE_EXECUTE_READWRITE);
    1611             : #else
    1612             :             c->lumMmxextFilterCode = av_malloc(c->lumMmxextFilterCodeSize);
    1613             :             c->chrMmxextFilterCode = av_malloc(c->chrMmxextFilterCodeSize);
    1614             : #endif
    1615             : 
    1616             : #ifdef MAP_ANONYMOUS
    1617           0 :             if (c->lumMmxextFilterCode == MAP_FAILED || c->chrMmxextFilterCode == MAP_FAILED)
    1618             : #else
    1619             :             if (!c->lumMmxextFilterCode || !c->chrMmxextFilterCode)
    1620             : #endif
    1621             :             {
    1622           0 :                 av_log(c, AV_LOG_ERROR, "Failed to allocate MMX2FilterCode\n");
    1623           0 :                 return AVERROR(ENOMEM);
    1624             :             }
    1625             : 
    1626           0 :             FF_ALLOCZ_OR_GOTO(c, c->hLumFilter,    (dstW           / 8 + 8) * sizeof(int16_t), fail);
    1627           0 :             FF_ALLOCZ_OR_GOTO(c, c->hChrFilter,    (c->chrDstW     / 4 + 8) * sizeof(int16_t), fail);
    1628           0 :             FF_ALLOCZ_OR_GOTO(c, c->hLumFilterPos, (dstW       / 2 / 8 + 8) * sizeof(int32_t), fail);
    1629           0 :             FF_ALLOCZ_OR_GOTO(c, c->hChrFilterPos, (c->chrDstW / 2 / 4 + 8) * sizeof(int32_t), fail);
    1630             : 
    1631           0 :             ff_init_hscaler_mmxext(      dstW, c->lumXInc, c->lumMmxextFilterCode,
    1632             :                                 c->hLumFilter, (uint32_t*)c->hLumFilterPos, 8);
    1633           0 :             ff_init_hscaler_mmxext(c->chrDstW, c->chrXInc, c->chrMmxextFilterCode,
    1634             :                                 c->hChrFilter, (uint32_t*)c->hChrFilterPos, 4);
    1635             : 
    1636             : #if USE_MMAP
    1637           0 :             if (   mprotect(c->lumMmxextFilterCode, c->lumMmxextFilterCodeSize, PROT_EXEC | PROT_READ) == -1
    1638           0 :                 || mprotect(c->chrMmxextFilterCode, c->chrMmxextFilterCodeSize, PROT_EXEC | PROT_READ) == -1) {
    1639           0 :                 av_log(c, AV_LOG_ERROR, "mprotect failed, cannot use fast bilinear scaler\n");
    1640           0 :                 goto fail;
    1641             :             }
    1642             : #endif
    1643             :         } else
    1644             : #endif /* HAVE_MMXEXT_INLINE */
    1645             :         {
    1646        2822 :             const int filterAlign = X86_MMX(cpu_flags)     ? 4 :
    1647             :                                     PPC_ALTIVEC(cpu_flags) ? 8 :
    1648             :                                     have_neon(cpu_flags)   ? 8 : 1;
    1649             : 
    1650        8466 :             if ((ret = initFilter(&c->hLumFilter, &c->hLumFilterPos,
    1651             :                            &c->hLumFilterSize, c->lumXInc,
    1652             :                            srcW, dstW, filterAlign, 1 << 14,
    1653        2822 :                            (flags & SWS_BICUBLIN) ? (flags | SWS_BICUBIC) : flags,
    1654             :                            cpu_flags, srcFilter->lumH, dstFilter->lumH,
    1655        2822 :                            c->param,
    1656             :                            get_local_pos(c, 0, 0, 0),
    1657             :                            get_local_pos(c, 0, 0, 0))) < 0)
    1658           0 :                 goto fail;
    1659        8466 :             if ((ret = initFilter(&c->hChrFilter, &c->hChrFilterPos,
    1660             :                            &c->hChrFilterSize, c->chrXInc,
    1661             :                            c->chrSrcW, c->chrDstW, filterAlign, 1 << 14,
    1662        2822 :                            (flags & SWS_BICUBLIN) ? (flags | SWS_BILINEAR) : flags,
    1663             :                            cpu_flags, srcFilter->chrH, dstFilter->chrH,
    1664        2822 :                            c->param,
    1665             :                            get_local_pos(c, c->chrSrcHSubSample, c->src_h_chr_pos, 0),
    1666             :                            get_local_pos(c, c->chrDstHSubSample, c->dst_h_chr_pos, 0))) < 0)
    1667           0 :                 goto fail;
    1668             :         }
    1669             :     } // initialize horizontal stuff
    1670             : 
    1671             :     /* precalculate vertical scaler filter coefficients */
    1672             :     {
    1673        2822 :         const int filterAlign = X86_MMX(cpu_flags)     ? 2 :
    1674             :                                 PPC_ALTIVEC(cpu_flags) ? 8 :
    1675             :                                 have_neon(cpu_flags)   ? 2 : 1;
    1676             : 
    1677        8466 :         if ((ret = initFilter(&c->vLumFilter, &c->vLumFilterPos, &c->vLumFilterSize,
    1678             :                        c->lumYInc, srcH, dstH, filterAlign, (1 << 12),
    1679        2822 :                        (flags & SWS_BICUBLIN) ? (flags | SWS_BICUBIC) : flags,
    1680             :                        cpu_flags, srcFilter->lumV, dstFilter->lumV,
    1681        2822 :                        c->param,
    1682             :                        get_local_pos(c, 0, 0, 1),
    1683             :                        get_local_pos(c, 0, 0, 1))) < 0)
    1684           0 :             goto fail;
    1685        8466 :         if ((ret = initFilter(&c->vChrFilter, &c->vChrFilterPos, &c->vChrFilterSize,
    1686             :                        c->chrYInc, c->chrSrcH, c->chrDstH,
    1687             :                        filterAlign, (1 << 12),
    1688        2822 :                        (flags & SWS_BICUBLIN) ? (flags | SWS_BILINEAR) : flags,
    1689             :                        cpu_flags, srcFilter->chrV, dstFilter->chrV,
    1690        2822 :                        c->param,
    1691             :                        get_local_pos(c, c->chrSrcVSubSample, c->src_v_chr_pos, 1),
    1692             :                        get_local_pos(c, c->chrDstVSubSample, c->dst_v_chr_pos, 1))) < 0)
    1693             : 
    1694           0 :             goto fail;
    1695             : 
    1696             : #if HAVE_ALTIVEC
    1697             :         FF_ALLOC_OR_GOTO(c, c->vYCoeffsBank, sizeof(vector signed short) * c->vLumFilterSize * c->dstH,    fail);
    1698             :         FF_ALLOC_OR_GOTO(c, c->vCCoeffsBank, sizeof(vector signed short) * c->vChrFilterSize * c->chrDstH, fail);
    1699             : 
    1700             :         for (i = 0; i < c->vLumFilterSize * c->dstH; i++) {
    1701             :             int j;
    1702             :             short *p = (short *)&c->vYCoeffsBank[i];
    1703             :             for (j = 0; j < 8; j++)
    1704             :                 p[j] = c->vLumFilter[i];
    1705             :         }
    1706             : 
    1707             :         for (i = 0; i < c->vChrFilterSize * c->chrDstH; i++) {
    1708             :             int j;
    1709             :             short *p = (short *)&c->vCCoeffsBank[i];
    1710             :             for (j = 0; j < 8; j++)
    1711             :                 p[j] = c->vChrFilter[i];
    1712             :         }
    1713             : #endif
    1714             :     }
    1715             : 
    1716       14110 :     for (i = 0; i < 4; i++)
    1717       11288 :         FF_ALLOCZ_OR_GOTO(c, c->dither_error[i], (c->dstW+2) * sizeof(int), fail);
    1718             : 
    1719        2822 :     c->needAlpha = (CONFIG_SWSCALE_ALPHA && isALPHA(c->srcFormat) && isALPHA(c->dstFormat)) ? 1 : 0;
    1720             : 
    1721             :     // 64 / c->scalingBpp is the same as 16 / sizeof(scaling_intermediate)
    1722        2822 :     c->uv_off   = (dst_stride>>1) + 64 / (c->dstBpc &~ 7);
    1723        2822 :     c->uv_offx2 = dst_stride + 16;
    1724             : 
    1725        2822 :     av_assert0(c->chrDstH <= dstH);
    1726             : 
    1727        2822 :     if (flags & SWS_PRINT_INFO) {
    1728           0 :         const char *scaler = NULL, *cpucaps;
    1729             : 
    1730           0 :         for (i = 0; i < FF_ARRAY_ELEMS(scale_algorithms); i++) {
    1731           0 :             if (flags & scale_algorithms[i].flag) {
    1732           0 :                 scaler = scale_algorithms[i].description;
    1733           0 :                 break;
    1734             :             }
    1735             :         }
    1736           0 :         if (!scaler)
    1737           0 :             scaler =  "ehh flags invalid?!";
    1738           0 :         av_log(c, AV_LOG_INFO, "%s scaler, from %s to %s%s ",
    1739             :                scaler,
    1740             :                av_get_pix_fmt_name(srcFormat),
    1741             : #ifdef DITHER1XBPP
    1742           0 :                dstFormat == AV_PIX_FMT_BGR555   || dstFormat == AV_PIX_FMT_BGR565   ||
    1743           0 :                dstFormat == AV_PIX_FMT_RGB444BE || dstFormat == AV_PIX_FMT_RGB444LE ||
    1744           0 :                dstFormat == AV_PIX_FMT_BGR444BE || dstFormat == AV_PIX_FMT_BGR444LE ?
    1745             :                                                              "dithered " : "",
    1746             : #else
    1747             :                "",
    1748             : #endif
    1749             :                av_get_pix_fmt_name(dstFormat));
    1750             : 
    1751           0 :         if (INLINE_MMXEXT(cpu_flags))
    1752           0 :             cpucaps = "MMXEXT";
    1753           0 :         else if (INLINE_AMD3DNOW(cpu_flags))
    1754           0 :             cpucaps = "3DNOW";
    1755           0 :         else if (INLINE_MMX(cpu_flags))
    1756           0 :             cpucaps = "MMX";
    1757             :         else if (PPC_ALTIVEC(cpu_flags))
    1758             :             cpucaps = "AltiVec";
    1759             :         else
    1760           0 :             cpucaps = "C";
    1761             : 
    1762           0 :         av_log(c, AV_LOG_INFO, "using %s\n", cpucaps);
    1763             : 
    1764           0 :         av_log(c, AV_LOG_VERBOSE, "%dx%d -> %dx%d\n", srcW, srcH, dstW, dstH);
    1765           0 :         av_log(c, AV_LOG_DEBUG,
    1766             :                "lum srcW=%d srcH=%d dstW=%d dstH=%d xInc=%d yInc=%d\n",
    1767             :                c->srcW, c->srcH, c->dstW, c->dstH, c->lumXInc, c->lumYInc);
    1768           0 :         av_log(c, AV_LOG_DEBUG,
    1769             :                "chr srcW=%d srcH=%d dstW=%d dstH=%d xInc=%d yInc=%d\n",
    1770             :                c->chrSrcW, c->chrSrcH, c->chrDstW, c->chrDstH,
    1771             :                c->chrXInc, c->chrYInc);
    1772             :     }
    1773             : 
    1774             :     /* alpha blend special case, note this has been split via cascaded contexts if its scaled */
    1775        5276 :     if (unscaled && !usesHFilter && !usesVFilter &&
    1776        2454 :         c->alphablend != SWS_ALPHA_BLEND_NONE &&
    1777           0 :         isALPHA(srcFormat) &&
    1778           0 :         (c->srcRange == c->dstRange || isAnyRGB(dstFormat)) &&
    1779           0 :         alphaless_fmt(srcFormat) == dstFormat
    1780             :     ) {
    1781           0 :         c->swscale = ff_sws_alphablendaway;
    1782             : 
    1783           0 :         if (flags & SWS_PRINT_INFO)
    1784           0 :             av_log(c, AV_LOG_INFO,
    1785             :                     "using alpha blendaway %s -> %s special converter\n",
    1786             :                     av_get_pix_fmt_name(srcFormat), av_get_pix_fmt_name(dstFormat));
    1787           0 :         return 0;
    1788             :     }
    1789             : 
    1790             :     /* unscaled special cases */
    1791        5276 :     if (unscaled && !usesHFilter && !usesVFilter &&
    1792        2755 :         (c->srcRange == c->dstRange || isAnyRGB(dstFormat))) {
    1793        2163 :         ff_get_unscaled_swscale(c);
    1794             : 
    1795        2163 :         if (c->swscale) {
    1796         416 :             if (flags & SWS_PRINT_INFO)
    1797           0 :                 av_log(c, AV_LOG_INFO,
    1798             :                        "using unscaled %s -> %s special converter\n",
    1799             :                        av_get_pix_fmt_name(srcFormat), av_get_pix_fmt_name(dstFormat));
    1800         416 :             return 0;
    1801             :         }
    1802             :     }
    1803             : 
    1804        2406 :     c->swscale = ff_getSwsFunc(c);
    1805        2406 :     return ff_init_filters(c);
    1806           0 : fail: // FIXME replace things by appropriate error codes
    1807           0 :     if (ret == RETCODE_USE_CASCADE)  {
    1808           0 :         int tmpW = sqrt(srcW * (int64_t)dstW);
    1809           0 :         int tmpH = sqrt(srcH * (int64_t)dstH);
    1810           0 :         enum AVPixelFormat tmpFormat = AV_PIX_FMT_YUV420P;
    1811             : 
    1812           0 :         if (isALPHA(srcFormat))
    1813           0 :             tmpFormat = AV_PIX_FMT_YUVA420P;
    1814             : 
    1815           0 :         if (srcW*(int64_t)srcH <= 4LL*dstW*dstH)
    1816           0 :             return AVERROR(EINVAL);
    1817             : 
    1818           0 :         ret = av_image_alloc(c->cascaded_tmp, c->cascaded_tmpStride,
    1819             :                              tmpW, tmpH, tmpFormat, 64);
    1820           0 :         if (ret < 0)
    1821           0 :             return ret;
    1822             : 
    1823           0 :         c->cascaded_context[0] = sws_getContext(srcW, srcH, srcFormat,
    1824             :                                                 tmpW, tmpH, tmpFormat,
    1825           0 :                                                 flags, srcFilter, NULL, c->param);
    1826           0 :         if (!c->cascaded_context[0])
    1827           0 :             return -1;
    1828             : 
    1829           0 :         c->cascaded_context[1] = sws_getContext(tmpW, tmpH, tmpFormat,
    1830             :                                                 dstW, dstH, dstFormat,
    1831           0 :                                                 flags, NULL, dstFilter, c->param);
    1832           0 :         if (!c->cascaded_context[1])
    1833           0 :             return -1;
    1834           0 :         return 0;
    1835             :     }
    1836           0 :     return -1;
    1837             : }
    1838             : 
    1839           0 : SwsContext *sws_alloc_set_opts(int srcW, int srcH, enum AVPixelFormat srcFormat,
    1840             :                                int dstW, int dstH, enum AVPixelFormat dstFormat,
    1841             :                                int flags, const double *param)
    1842             : {
    1843             :     SwsContext *c;
    1844             : 
    1845           0 :     if (!(c = sws_alloc_context()))
    1846           0 :         return NULL;
    1847             : 
    1848           0 :     c->flags     = flags;
    1849           0 :     c->srcW      = srcW;
    1850           0 :     c->srcH      = srcH;
    1851           0 :     c->dstW      = dstW;
    1852           0 :     c->dstH      = dstH;
    1853           0 :     c->srcFormat = srcFormat;
    1854           0 :     c->dstFormat = dstFormat;
    1855             : 
    1856           0 :     if (param) {
    1857           0 :         c->param[0] = param[0];
    1858           0 :         c->param[1] = param[1];
    1859             :     }
    1860             : 
    1861           0 :     return c;
    1862             : }
    1863             : 
    1864           0 : SwsContext *sws_getContext(int srcW, int srcH, enum AVPixelFormat srcFormat,
    1865             :                            int dstW, int dstH, enum AVPixelFormat dstFormat,
    1866             :                            int flags, SwsFilter *srcFilter,
    1867             :                            SwsFilter *dstFilter, const double *param)
    1868             : {
    1869             :     SwsContext *c;
    1870             : 
    1871           0 :     c = sws_alloc_set_opts(srcW, srcH, srcFormat,
    1872             :                            dstW, dstH, dstFormat,
    1873             :                            flags, param);
    1874           0 :     if (!c)
    1875           0 :         return NULL;
    1876             : 
    1877           0 :     if (sws_init_context(c, srcFilter, dstFilter) < 0) {
    1878           0 :         sws_freeContext(c);
    1879           0 :         return NULL;
    1880             :     }
    1881             : 
    1882           0 :     return c;
    1883             : }
    1884             : 
    1885           0 : static int isnan_vec(SwsVector *a)
    1886             : {
    1887             :     int i;
    1888           0 :     for (i=0; i<a->length; i++)
    1889           0 :         if (isnan(a->coeff[i]))
    1890           0 :             return 1;
    1891           0 :     return 0;
    1892             : }
    1893             : 
    1894           0 : static void makenan_vec(SwsVector *a)
    1895             : {
    1896             :     int i;
    1897           0 :     for (i=0; i<a->length; i++)
    1898           0 :         a->coeff[i] = NAN;
    1899           0 : }
    1900             : 
    1901           0 : SwsFilter *sws_getDefaultFilter(float lumaGBlur, float chromaGBlur,
    1902             :                                 float lumaSharpen, float chromaSharpen,
    1903             :                                 float chromaHShift, float chromaVShift,
    1904             :                                 int verbose)
    1905             : {
    1906           0 :     SwsFilter *filter = av_malloc(sizeof(SwsFilter));
    1907           0 :     if (!filter)
    1908           0 :         return NULL;
    1909             : 
    1910           0 :     if (lumaGBlur != 0.0) {
    1911           0 :         filter->lumH = sws_getGaussianVec(lumaGBlur, 3.0);
    1912           0 :         filter->lumV = sws_getGaussianVec(lumaGBlur, 3.0);
    1913             :     } else {
    1914           0 :         filter->lumH = sws_getIdentityVec();
    1915           0 :         filter->lumV = sws_getIdentityVec();
    1916             :     }
    1917             : 
    1918           0 :     if (chromaGBlur != 0.0) {
    1919           0 :         filter->chrH = sws_getGaussianVec(chromaGBlur, 3.0);
    1920           0 :         filter->chrV = sws_getGaussianVec(chromaGBlur, 3.0);
    1921             :     } else {
    1922           0 :         filter->chrH = sws_getIdentityVec();
    1923           0 :         filter->chrV = sws_getIdentityVec();
    1924             :     }
    1925             : 
    1926           0 :     if (!filter->lumH || !filter->lumV || !filter->chrH || !filter->chrV)
    1927             :         goto fail;
    1928             : 
    1929           0 :     if (chromaSharpen != 0.0) {
    1930           0 :         SwsVector *id = sws_getIdentityVec();
    1931           0 :         if (!id)
    1932           0 :             goto fail;
    1933           0 :         sws_scaleVec(filter->chrH, -chromaSharpen);
    1934           0 :         sws_scaleVec(filter->chrV, -chromaSharpen);
    1935           0 :         sws_addVec(filter->chrH, id);
    1936           0 :         sws_addVec(filter->chrV, id);
    1937           0 :         sws_freeVec(id);
    1938             :     }
    1939             : 
    1940           0 :     if (lumaSharpen != 0.0) {
    1941           0 :         SwsVector *id = sws_getIdentityVec();
    1942           0 :         if (!id)
    1943           0 :             goto fail;
    1944           0 :         sws_scaleVec(filter->lumH, -lumaSharpen);
    1945           0 :         sws_scaleVec(filter->lumV, -lumaSharpen);
    1946           0 :         sws_addVec(filter->lumH, id);
    1947           0 :         sws_addVec(filter->lumV, id);
    1948           0 :         sws_freeVec(id);
    1949             :     }
    1950             : 
    1951           0 :     if (chromaHShift != 0.0)
    1952           0 :         sws_shiftVec(filter->chrH, (int)(chromaHShift + 0.5));
    1953             : 
    1954           0 :     if (chromaVShift != 0.0)
    1955           0 :         sws_shiftVec(filter->chrV, (int)(chromaVShift + 0.5));
    1956             : 
    1957           0 :     sws_normalizeVec(filter->chrH, 1.0);
    1958           0 :     sws_normalizeVec(filter->chrV, 1.0);
    1959           0 :     sws_normalizeVec(filter->lumH, 1.0);
    1960           0 :     sws_normalizeVec(filter->lumV, 1.0);
    1961             : 
    1962           0 :     if (isnan_vec(filter->chrH) ||
    1963           0 :         isnan_vec(filter->chrV) ||
    1964           0 :         isnan_vec(filter->lumH) ||
    1965           0 :         isnan_vec(filter->lumV))
    1966             :         goto fail;
    1967             : 
    1968           0 :     if (verbose)
    1969           0 :         sws_printVec2(filter->chrH, NULL, AV_LOG_DEBUG);
    1970           0 :     if (verbose)
    1971           0 :         sws_printVec2(filter->lumH, NULL, AV_LOG_DEBUG);
    1972             : 
    1973           0 :     return filter;
    1974             : 
    1975           0 : fail:
    1976           0 :     sws_freeVec(filter->lumH);
    1977           0 :     sws_freeVec(filter->lumV);
    1978           0 :     sws_freeVec(filter->chrH);
    1979           0 :     sws_freeVec(filter->chrV);
    1980           0 :     av_freep(&filter);
    1981           0 :     return NULL;
    1982             : }
    1983             : 
    1984           0 : SwsVector *sws_allocVec(int length)
    1985             : {
    1986             :     SwsVector *vec;
    1987             : 
    1988           0 :     if(length <= 0 || length > INT_MAX/ sizeof(double))
    1989           0 :         return NULL;
    1990             : 
    1991           0 :     vec = av_malloc(sizeof(SwsVector));
    1992           0 :     if (!vec)
    1993           0 :         return NULL;
    1994           0 :     vec->length = length;
    1995           0 :     vec->coeff  = av_malloc(sizeof(double) * length);
    1996           0 :     if (!vec->coeff)
    1997           0 :         av_freep(&vec);
    1998           0 :     return vec;
    1999             : }
    2000             : 
    2001           0 : SwsVector *sws_getGaussianVec(double variance, double quality)
    2002             : {
    2003           0 :     const int length = (int)(variance * quality + 0.5) | 1;
    2004             :     int i;
    2005           0 :     double middle  = (length - 1) * 0.5;
    2006             :     SwsVector *vec;
    2007             : 
    2008           0 :     if(variance < 0 || quality < 0)
    2009           0 :         return NULL;
    2010             : 
    2011           0 :     vec = sws_allocVec(length);
    2012             : 
    2013           0 :     if (!vec)
    2014           0 :         return NULL;
    2015             : 
    2016           0 :     for (i = 0; i < length; i++) {
    2017           0 :         double dist = i - middle;
    2018           0 :         vec->coeff[i] = exp(-dist * dist / (2 * variance * variance)) /
    2019           0 :                         sqrt(2 * variance * M_PI);
    2020             :     }
    2021             : 
    2022           0 :     sws_normalizeVec(vec, 1.0);
    2023             : 
    2024           0 :     return vec;
    2025             : }
    2026             : 
    2027             : /**
    2028             :  * Allocate and return a vector with length coefficients, all
    2029             :  * with the same value c.
    2030             :  */
    2031             : #if !FF_API_SWS_VECTOR
    2032             : static
    2033             : #endif
    2034           0 : SwsVector *sws_getConstVec(double c, int length)
    2035             : {
    2036             :     int i;
    2037           0 :     SwsVector *vec = sws_allocVec(length);
    2038             : 
    2039           0 :     if (!vec)
    2040           0 :         return NULL;
    2041             : 
    2042           0 :     for (i = 0; i < length; i++)
    2043           0 :         vec->coeff[i] = c;
    2044             : 
    2045           0 :     return vec;
    2046             : }
    2047             : 
    2048             : /**
    2049             :  * Allocate and return a vector with just one coefficient, with
    2050             :  * value 1.0.
    2051             :  */
    2052             : #if !FF_API_SWS_VECTOR
    2053             : static
    2054             : #endif
    2055           0 : SwsVector *sws_getIdentityVec(void)
    2056             : {
    2057           0 :     return sws_getConstVec(1.0, 1);
    2058             : }
    2059             : 
    2060           0 : static double sws_dcVec(SwsVector *a)
    2061             : {
    2062             :     int i;
    2063           0 :     double sum = 0;
    2064             : 
    2065           0 :     for (i = 0; i < a->length; i++)
    2066           0 :         sum += a->coeff[i];
    2067             : 
    2068           0 :     return sum;
    2069             : }
    2070             : 
    2071           0 : void sws_scaleVec(SwsVector *a, double scalar)
    2072             : {
    2073             :     int i;
    2074             : 
    2075           0 :     for (i = 0; i < a->length; i++)
    2076           0 :         a->coeff[i] *= scalar;
    2077           0 : }
    2078             : 
    2079           0 : void sws_normalizeVec(SwsVector *a, double height)
    2080             : {
    2081           0 :     sws_scaleVec(a, height / sws_dcVec(a));
    2082           0 : }
    2083             : 
    2084             : #if FF_API_SWS_VECTOR
    2085           0 : static SwsVector *sws_getConvVec(SwsVector *a, SwsVector *b)
    2086             : {
    2087           0 :     int length = a->length + b->length - 1;
    2088             :     int i, j;
    2089           0 :     SwsVector *vec = sws_getConstVec(0.0, length);
    2090             : 
    2091           0 :     if (!vec)
    2092           0 :         return NULL;
    2093             : 
    2094           0 :     for (i = 0; i < a->length; i++) {
    2095           0 :         for (j = 0; j < b->length; j++) {
    2096           0 :             vec->coeff[i + j] += a->coeff[i] * b->coeff[j];
    2097             :         }
    2098             :     }
    2099             : 
    2100           0 :     return vec;
    2101             : }
    2102             : #endif
    2103             : 
    2104           0 : static SwsVector *sws_sumVec(SwsVector *a, SwsVector *b)
    2105             : {
    2106           0 :     int length = FFMAX(a->length, b->length);
    2107             :     int i;
    2108           0 :     SwsVector *vec = sws_getConstVec(0.0, length);
    2109             : 
    2110           0 :     if (!vec)
    2111           0 :         return NULL;
    2112             : 
    2113           0 :     for (i = 0; i < a->length; i++)
    2114           0 :         vec->coeff[i + (length - 1) / 2 - (a->length - 1) / 2] += a->coeff[i];
    2115           0 :     for (i = 0; i < b->length; i++)
    2116           0 :         vec->coeff[i + (length - 1) / 2 - (b->length - 1) / 2] += b->coeff[i];
    2117             : 
    2118           0 :     return vec;
    2119             : }
    2120             : 
    2121             : #if FF_API_SWS_VECTOR
    2122           0 : static SwsVector *sws_diffVec(SwsVector *a, SwsVector *b)
    2123             : {
    2124           0 :     int length = FFMAX(a->length, b->length);
    2125             :     int i;
    2126           0 :     SwsVector *vec = sws_getConstVec(0.0, length);
    2127             : 
    2128           0 :     if (!vec)
    2129           0 :         return NULL;
    2130             : 
    2131           0 :     for (i = 0; i < a->length; i++)
    2132           0 :         vec->coeff[i + (length - 1) / 2 - (a->length - 1) / 2] += a->coeff[i];
    2133           0 :     for (i = 0; i < b->length; i++)
    2134           0 :         vec->coeff[i + (length - 1) / 2 - (b->length - 1) / 2] -= b->coeff[i];
    2135             : 
    2136           0 :     return vec;
    2137             : }
    2138             : #endif
    2139             : 
    2140             : /* shift left / or right if "shift" is negative */
    2141           0 : static SwsVector *sws_getShiftedVec(SwsVector *a, int shift)
    2142             : {
    2143           0 :     int length = a->length + FFABS(shift) * 2;
    2144             :     int i;
    2145           0 :     SwsVector *vec = sws_getConstVec(0.0, length);
    2146             : 
    2147           0 :     if (!vec)
    2148           0 :         return NULL;
    2149             : 
    2150           0 :     for (i = 0; i < a->length; i++) {
    2151           0 :         vec->coeff[i + (length    - 1) / 2 -
    2152           0 :                        (a->length - 1) / 2 - shift] = a->coeff[i];
    2153             :     }
    2154             : 
    2155           0 :     return vec;
    2156             : }
    2157             : 
    2158             : #if !FF_API_SWS_VECTOR
    2159             : static
    2160             : #endif
    2161           0 : void sws_shiftVec(SwsVector *a, int shift)
    2162             : {
    2163           0 :     SwsVector *shifted = sws_getShiftedVec(a, shift);
    2164           0 :     if (!shifted) {
    2165           0 :         makenan_vec(a);
    2166           0 :         return;
    2167             :     }
    2168           0 :     av_free(a->coeff);
    2169           0 :     a->coeff  = shifted->coeff;
    2170           0 :     a->length = shifted->length;
    2171           0 :     av_free(shifted);
    2172             : }
    2173             : 
    2174             : #if !FF_API_SWS_VECTOR
    2175             : static
    2176             : #endif
    2177           0 : void sws_addVec(SwsVector *a, SwsVector *b)
    2178             : {
    2179           0 :     SwsVector *sum = sws_sumVec(a, b);
    2180           0 :     if (!sum) {
    2181           0 :         makenan_vec(a);
    2182           0 :         return;
    2183             :     }
    2184           0 :     av_free(a->coeff);
    2185           0 :     a->coeff  = sum->coeff;
    2186           0 :     a->length = sum->length;
    2187           0 :     av_free(sum);
    2188             : }
    2189             : 
    2190             : #if FF_API_SWS_VECTOR
    2191           0 : void sws_subVec(SwsVector *a, SwsVector *b)
    2192             : {
    2193           0 :     SwsVector *diff = sws_diffVec(a, b);
    2194           0 :     if (!diff) {
    2195           0 :         makenan_vec(a);
    2196           0 :         return;
    2197             :     }
    2198           0 :     av_free(a->coeff);
    2199           0 :     a->coeff  = diff->coeff;
    2200           0 :     a->length = diff->length;
    2201           0 :     av_free(diff);
    2202             : }
    2203             : 
    2204           0 : void sws_convVec(SwsVector *a, SwsVector *b)
    2205             : {
    2206           0 :     SwsVector *conv = sws_getConvVec(a, b);
    2207           0 :     if (!conv) {
    2208           0 :         makenan_vec(a);
    2209           0 :         return;
    2210             :     }
    2211           0 :     av_free(a->coeff);
    2212           0 :     a->coeff  = conv->coeff;
    2213           0 :     a->length = conv->length;
    2214           0 :     av_free(conv);
    2215             : }
    2216             : 
    2217           0 : SwsVector *sws_cloneVec(SwsVector *a)
    2218             : {
    2219           0 :     SwsVector *vec = sws_allocVec(a->length);
    2220             : 
    2221           0 :     if (!vec)
    2222           0 :         return NULL;
    2223             : 
    2224           0 :     memcpy(vec->coeff, a->coeff, a->length * sizeof(*a->coeff));
    2225             : 
    2226           0 :     return vec;
    2227             : }
    2228             : #endif
    2229             : 
    2230             : /**
    2231             :  * Print with av_log() a textual representation of the vector a
    2232             :  * if log_level <= av_log_level.
    2233             :  */
    2234             : #if !FF_API_SWS_VECTOR
    2235             : static
    2236             : #endif
    2237           0 : void sws_printVec2(SwsVector *a, AVClass *log_ctx, int log_level)
    2238             : {
    2239             :     int i;
    2240           0 :     double max = 0;
    2241           0 :     double min = 0;
    2242             :     double range;
    2243             : 
    2244           0 :     for (i = 0; i < a->length; i++)
    2245           0 :         if (a->coeff[i] > max)
    2246           0 :             max = a->coeff[i];
    2247             : 
    2248           0 :     for (i = 0; i < a->length; i++)
    2249           0 :         if (a->coeff[i] < min)
    2250           0 :             min = a->coeff[i];
    2251             : 
    2252           0 :     range = max - min;
    2253             : 
    2254           0 :     for (i = 0; i < a->length; i++) {
    2255           0 :         int x = (int)((a->coeff[i] - min) * 60.0 / range + 0.5);
    2256           0 :         av_log(log_ctx, log_level, "%1.3f ", a->coeff[i]);
    2257           0 :         for (; x > 0; x--)
    2258           0 :             av_log(log_ctx, log_level, " ");
    2259           0 :         av_log(log_ctx, log_level, "|\n");
    2260             :     }
    2261           0 : }
    2262             : 
    2263           0 : void sws_freeVec(SwsVector *a)
    2264             : {
    2265           0 :     if (!a)
    2266           0 :         return;
    2267           0 :     av_freep(&a->coeff);
    2268           0 :     a->length = 0;
    2269           0 :     av_free(a);
    2270             : }
    2271             : 
    2272           0 : void sws_freeFilter(SwsFilter *filter)
    2273             : {
    2274           0 :     if (!filter)
    2275           0 :         return;
    2276             : 
    2277           0 :     sws_freeVec(filter->lumH);
    2278           0 :     sws_freeVec(filter->lumV);
    2279           0 :     sws_freeVec(filter->chrH);
    2280           0 :     sws_freeVec(filter->chrV);
    2281           0 :     av_free(filter);
    2282             : }
    2283             : 
    2284       42614 : void sws_freeContext(SwsContext *c)
    2285             : {
    2286             :     int i;
    2287       42614 :     if (!c)
    2288       33536 :         return;
    2289             : 
    2290       45390 :     for (i = 0; i < 4; i++)
    2291       36312 :         av_freep(&c->dither_error[i]);
    2292             : 
    2293        9078 :     av_freep(&c->vLumFilter);
    2294        9078 :     av_freep(&c->vChrFilter);
    2295        9078 :     av_freep(&c->hLumFilter);
    2296        9078 :     av_freep(&c->hChrFilter);
    2297             : #if HAVE_ALTIVEC
    2298             :     av_freep(&c->vYCoeffsBank);
    2299             :     av_freep(&c->vCCoeffsBank);
    2300             : #endif
    2301             : 
    2302        9078 :     av_freep(&c->vLumFilterPos);
    2303        9078 :     av_freep(&c->vChrFilterPos);
    2304        9078 :     av_freep(&c->hLumFilterPos);
    2305        9078 :     av_freep(&c->hChrFilterPos);
    2306             : 
    2307             : #if HAVE_MMX_INLINE
    2308             : #if USE_MMAP
    2309        9078 :     if (c->lumMmxextFilterCode)
    2310           0 :         munmap(c->lumMmxextFilterCode, c->lumMmxextFilterCodeSize);
    2311        9078 :     if (c->chrMmxextFilterCode)
    2312           0 :         munmap(c->chrMmxextFilterCode, c->chrMmxextFilterCodeSize);
    2313             : #elif HAVE_VIRTUALALLOC
    2314             :     if (c->lumMmxextFilterCode)
    2315             :         VirtualFree(c->lumMmxextFilterCode, 0, MEM_RELEASE);
    2316             :     if (c->chrMmxextFilterCode)
    2317             :         VirtualFree(c->chrMmxextFilterCode, 0, MEM_RELEASE);
    2318             : #else
    2319             :     av_free(c->lumMmxextFilterCode);
    2320             :     av_free(c->chrMmxextFilterCode);
    2321             : #endif
    2322        9078 :     c->lumMmxextFilterCode = NULL;
    2323        9078 :     c->chrMmxextFilterCode = NULL;
    2324             : #endif /* HAVE_MMX_INLINE */
    2325             : 
    2326        9078 :     av_freep(&c->yuvTable);
    2327        9078 :     av_freep(&c->formatConvBuffer);
    2328             : 
    2329        9078 :     sws_freeContext(c->cascaded_context[0]);
    2330        9078 :     sws_freeContext(c->cascaded_context[1]);
    2331        9078 :     sws_freeContext(c->cascaded_context[2]);
    2332        9078 :     memset(c->cascaded_context, 0, sizeof(c->cascaded_context));
    2333        9078 :     av_freep(&c->cascaded_tmp[0]);
    2334        9078 :     av_freep(&c->cascaded1_tmp[0]);
    2335             : 
    2336        9078 :     av_freep(&c->gamma);
    2337        9078 :     av_freep(&c->inv_gamma);
    2338             : 
    2339        9078 :     ff_free_filters(c);
    2340             : 
    2341        9078 :     av_free(c);
    2342             : }
    2343             : 
    2344           0 : struct SwsContext *sws_getCachedContext(struct SwsContext *context, int srcW,
    2345             :                                         int srcH, enum AVPixelFormat srcFormat,
    2346             :                                         int dstW, int dstH,
    2347             :                                         enum AVPixelFormat dstFormat, int flags,
    2348             :                                         SwsFilter *srcFilter,
    2349             :                                         SwsFilter *dstFilter,
    2350             :                                         const double *param)
    2351             : {
    2352             :     static const double default_param[2] = { SWS_PARAM_DEFAULT,
    2353             :                                              SWS_PARAM_DEFAULT };
    2354           0 :     int64_t src_h_chr_pos = -513, dst_h_chr_pos = -513,
    2355           0 :             src_v_chr_pos = -513, dst_v_chr_pos = -513;
    2356             : 
    2357           0 :     if (!param)
    2358           0 :         param = default_param;
    2359             : 
    2360           0 :     if (context &&
    2361           0 :         (context->srcW      != srcW      ||
    2362           0 :          context->srcH      != srcH      ||
    2363           0 :          context->srcFormat != srcFormat ||
    2364           0 :          context->dstW      != dstW      ||
    2365           0 :          context->dstH      != dstH      ||
    2366           0 :          context->dstFormat != dstFormat ||
    2367           0 :          context->flags     != flags     ||
    2368           0 :          context->param[0]  != param[0]  ||
    2369           0 :          context->param[1]  != param[1])) {
    2370             : 
    2371           0 :         av_opt_get_int(context, "src_h_chr_pos", 0, &src_h_chr_pos);
    2372           0 :         av_opt_get_int(context, "src_v_chr_pos", 0, &src_v_chr_pos);
    2373           0 :         av_opt_get_int(context, "dst_h_chr_pos", 0, &dst_h_chr_pos);
    2374           0 :         av_opt_get_int(context, "dst_v_chr_pos", 0, &dst_v_chr_pos);
    2375           0 :         sws_freeContext(context);
    2376           0 :         context = NULL;
    2377             :     }
    2378             : 
    2379           0 :     if (!context) {
    2380           0 :         if (!(context = sws_alloc_context()))
    2381           0 :             return NULL;
    2382           0 :         context->srcW      = srcW;
    2383           0 :         context->srcH      = srcH;
    2384           0 :         context->srcFormat = srcFormat;
    2385           0 :         context->dstW      = dstW;
    2386           0 :         context->dstH      = dstH;
    2387           0 :         context->dstFormat = dstFormat;
    2388           0 :         context->flags     = flags;
    2389           0 :         context->param[0]  = param[0];
    2390           0 :         context->param[1]  = param[1];
    2391             : 
    2392           0 :         av_opt_set_int(context, "src_h_chr_pos", src_h_chr_pos, 0);
    2393           0 :         av_opt_set_int(context, "src_v_chr_pos", src_v_chr_pos, 0);
    2394           0 :         av_opt_set_int(context, "dst_h_chr_pos", dst_h_chr_pos, 0);
    2395           0 :         av_opt_set_int(context, "dst_v_chr_pos", dst_v_chr_pos, 0);
    2396             : 
    2397           0 :         if (sws_init_context(context, srcFilter, dstFilter) < 0) {
    2398           0 :             sws_freeContext(context);
    2399           0 :             return NULL;
    2400             :         }
    2401             :     }
    2402           0 :     return context;
    2403             : }

Generated by: LCOV version 1.13