ffmpeg/libswscale/swscale.c

/*
 * Copyright (C) 2001-2003 Michael Niedermayer <michaelni@gmx.at>
 *
 * This file is part of FFmpeg.
 *
 * FFmpeg is free software; you can redistribute it and/or
 * modify it under the terms of the GNU Lesser General Public
 * License as published by the Free Software Foundation; either
 * version 2.1 of the License, or (at your option) any later version.
 *
 * FFmpeg is distributed in the hope that it will be useful,
 * but WITHOUT ANY WARRANTY; without even the implied warranty of
 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
 * Lesser General Public License for more details.
 *
 * You should have received a copy of the GNU Lesser General Public
 * License along with FFmpeg; if not, write to the Free Software
 * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
 */

/*
  supported Input formats: YV12, I420/IYUV, YUY2, UYVY, BGR32, BGR32_1, BGR24, BGR16, BGR15, RGB32, RGB32_1, RGB24, Y8/Y800, YVU9/IF09, PAL8
  supported output formats: YV12, I420/IYUV, YUY2, UYVY, {BGR,RGB}{1,4,8,15,16,24,32}, Y8/Y800, YVU9/IF09
  {BGR,RGB}{1,4,8,15,16} support dithering

  unscaled special converters (YV12=I420=IYUV, Y800=Y8)
  YV12 -> {BGR,RGB}{1,4,8,12,15,16,24,32}
  x -> x
  YUV9 -> YV12
  YUV9/YV12 -> Y800
  Y800 -> YUV9/YV12
  BGR24 -> BGR32 & RGB24 -> RGB32
  BGR32 -> BGR24 & RGB32 -> RGB24
  BGR15 -> BGR16
*/

/*
tested special converters (most are tested actually, but I did not write it down ...)
 YV12 -> BGR12/BGR16
 YV12 -> YV12
 BGR15 -> BGR16
 BGR16 -> BGR16
 YVU9 -> YV12

untested special converters
  YV12/I420 -> BGR15/BGR24/BGR32 (it is the yuv2rgb stuff, so it should be OK)
  YV12/I420 -> YV12/I420
  YUY2/BGR15/BGR24/BGR32/RGB24/RGB32 -> same format
  BGR24 -> BGR32 & RGB24 -> RGB32
  BGR32 -> BGR24 & RGB32 -> RGB24
  BGR24 -> YV12
*/

#include <inttypes.h>
#include <string.h>
#include <math.h>
#include <stdio.h>
#include "config.h"
#include <assert.h>
#include "swscale.h"
#include "swscale_internal.h"
#include "rgb2rgb.h"
#include "libavutil/avassert.h"
#include "libavutil/intreadwrite.h"
#include "libavutil/x86_cpu.h"
#include "libavutil/cpu.h"
#include "libavutil/avutil.h"
#include "libavutil/mathematics.h"
#include "libavutil/bswap.h"
#include "libavutil/pixdesc.h"

#undef MOVNTQ
#undef PAVGB

#define DITHER1XBPP

#define isPacked(x)         (       \
           (x)==PIX_FMT_PAL8        \
        || (x)==PIX_FMT_YUYV422     \
        || (x)==PIX_FMT_UYVY422     \
        || (x)==PIX_FMT_GRAY8A       \
        || isAnyRGB(x)              \
    )

#define RGB2YUV_SHIFT 15
#define BY ( (int)(0.114*219/255*(1<<RGB2YUV_SHIFT)+0.5))
#define BV (-(int)(0.081*224/255*(1<<RGB2YUV_SHIFT)+0.5))
#define BU ( (int)(0.500*224/255*(1<<RGB2YUV_SHIFT)+0.5))
#define GY ( (int)(0.587*219/255*(1<<RGB2YUV_SHIFT)+0.5))
#define GV (-(int)(0.419*224/255*(1<<RGB2YUV_SHIFT)+0.5))
#define GU (-(int)(0.331*224/255*(1<<RGB2YUV_SHIFT)+0.5))
#define RY ( (int)(0.299*219/255*(1<<RGB2YUV_SHIFT)+0.5))
#define RV ( (int)(0.500*224/255*(1<<RGB2YUV_SHIFT)+0.5))
#define RU (-(int)(0.169*224/255*(1<<RGB2YUV_SHIFT)+0.5))

static const double rgb2yuv_table[8][9]={
    {0.7152, 0.0722, 0.2126, -0.386, 0.5, -0.115, -0.454, -0.046, 0.5}, //ITU709
    {0.7152, 0.0722, 0.2126, -0.386, 0.5, -0.115, -0.454, -0.046, 0.5}, //ITU709
    {0.587 , 0.114 , 0.299 , -0.331, 0.5, -0.169, -0.419, -0.081, 0.5}, //DEFAULT / ITU601 / ITU624 / SMPTE 170M
    {0.587 , 0.114 , 0.299 , -0.331, 0.5, -0.169, -0.419, -0.081, 0.5}, //DEFAULT / ITU601 / ITU624 / SMPTE 170M
    {0.59  , 0.11  , 0.30  , -0.331, 0.5, -0.169, -0.421, -0.079, 0.5}, //FCC
    {0.587 , 0.114 , 0.299 , -0.331, 0.5, -0.169, -0.419, -0.081, 0.5}, //DEFAULT / ITU601 / ITU624 / SMPTE 170M
    {0.587 , 0.114 , 0.299 , -0.331, 0.5, -0.169, -0.419, -0.081, 0.5}, //DEFAULT / ITU601 / ITU624 / SMPTE 170M
    {0.701 , 0.087 , 0.212 , -0.384, 0.5, -0.116, -0.445, -0.055, 0.5}, //SMPTE 240M
};

/*
NOTES
Special versions: fast Y 1:1 scaling (no interpolation in y direction)

TODO
more intelligent misalignment avoidance for the horizontal scaler
write special vertical cubic upscale version
optimize C code (YV12 / minmax)
add support for packed pixel YUV input & output
add support for Y8 output
optimize BGR24 & BGR32
add BGR4 output support
write special BGR->BGR scaler
*/

DECLARE_ALIGNED(8, static const uint8_t, dither_2x2_4)[2][8]={
{  1,   3,   1,   3,   1,   3,   1,   3, },
{  2,   0,   2,   0,   2,   0,   2,   0, },
};

DECLARE_ALIGNED(8, static const uint8_t, dither_2x2_8)[2][8]={
{  6,   2,   6,   2,   6,   2,   6,   2, },
{  0,   4,   0,   4,   0,   4,   0,   4, },
};

DECLARE_ALIGNED(8, const uint8_t, dither_4x4_16)[4][8]={
{  8,   4,  11,   7,   8,   4,  11,   7, },
{  2,  14,   1,  13,   2,  14,   1,  13, },
{ 10,   6,   9,   5,  10,   6,   9,   5, },
{  0,  12,   3,  15,   0,  12,   3,  15, },
};

DECLARE_ALIGNED(8, const uint8_t, dither_8x8_32)[8][8]={
{ 17,   9,  23,  15,  16,   8,  22,  14, },
{  5,  29,   3,  27,   4,  28,   2,  26, },
{ 21,  13,  19,  11,  20,  12,  18,  10, },
{  0,  24,   6,  30,   1,  25,   7,  31, },
{ 16,   8,  22,  14,  17,   9,  23,  15, },
{  4,  28,   2,  26,   5,  29,   3,  27, },
{ 20,  12,  18,  10,  21,  13,  19,  11, },
{  1,  25,   7,  31,   0,  24,   6,  30, },
};

DECLARE_ALIGNED(8, const uint8_t, dither_8x8_73)[8][8]={
{  0,  55,  14,  68,   3,  58,  17,  72, },
{ 37,  18,  50,  32,  40,  22,  54,  35, },
{  9,  64,   5,  59,  13,  67,   8,  63, },
{ 46,  27,  41,  23,  49,  31,  44,  26, },
{  2,  57,  16,  71,   1,  56,  15,  70, },
{ 39,  21,  52,  34,  38,  19,  51,  33, },
{ 11,  66,   7,  62,  10,  65,   6,  60, },
{ 48,  30,  43,  25,  47,  29,  42,  24, },
};

#if 1
DECLARE_ALIGNED(8, const uint8_t, dither_8x8_220)[8][8]={
{117,  62, 158, 103, 113,  58, 155, 100, },
{ 34, 199,  21, 186,  31, 196,  17, 182, },
{144,  89, 131,  76, 141,  86, 127,  72, },
{  0, 165,  41, 206,  10, 175,  52, 217, },
{110,  55, 151,  96, 120,  65, 162, 107, },
{ 28, 193,  14, 179,  38, 203,  24, 189, },
{138,  83, 124,  69, 148,  93, 134,  79, },
{  7, 172,  48, 213,   3, 168,  45, 210, },
};
#elif 1
// tries to correct a gamma of 1.5
DECLARE_ALIGNED(8, const uint8_t, dither_8x8_220)[8][8]={
{  0, 143,  18, 200,   2, 156,  25, 215, },
{ 78,  28, 125,  64,  89,  36, 138,  74, },
{ 10, 180,   3, 161,  16, 195,   8, 175, },
{109,  51,  93,  38, 121,  60, 105,  47, },
{  1, 152,  23, 210,   0, 147,  20, 205, },
{ 85,  33, 134,  71,  81,  30, 130,  67, },
{ 14, 190,   6, 171,  12, 185,   5, 166, },
{117,  57, 101,  44, 113,  54,  97,  41, },
};
#elif 1
// tries to correct a gamma of 2.0
DECLARE_ALIGNED(8, const uint8_t, dither_8x8_220)[8][8]={
{  0, 124,   8, 193,   0, 140,  12, 213, },
{ 55,  14, 104,  42,  66,  19, 119,  52, },
{  3, 168,   1, 145,   6, 187,   3, 162, },
{ 86,  31,  70,  21,  99,  39,  82,  28, },
{  0, 134,  11, 206,   0, 129,   9, 200, },
{ 62,  17, 114,  48,  58,  16, 109,  45, },
{  5, 181,   2, 157,   4, 175,   1, 151, },
{ 95,  36,  78,  26,  90,  34,  74,  24, },
};
#else
// tries to correct a gamma of 2.5
DECLARE_ALIGNED(8, const uint8_t, dither_8x8_220)[8][8]={
{  0, 107,   3, 187,   0, 125,   6, 212, },
{ 39,   7,  86,  28,  49,  11, 102,  36, },
{  1, 158,   0, 131,   3, 180,   1, 151, },
{ 68,  19,  52,  12,  81,  25,  64,  17, },
{  0, 119,   5, 203,   0, 113,   4, 195, },
{ 45,   9,  96,  33,  42,   8,  91,  30, },
{  2, 172,   1, 144,   2, 165,   0, 137, },
{ 77,  23,  60,  15,  72,  21,  56,  14, },
};
#endif

DECLARE_ALIGNED(8, const uint8_t, dithers)[8][8][8]={
{
  {   0,  1,  0,  1,  0,  1,  0,  1,},
  {   1,  0,  1,  0,  1,  0,  1,  0,},
  {   0,  1,  0,  1,  0,  1,  0,  1,},
  {   1,  0,  1,  0,  1,  0,  1,  0,},
  {   0,  1,  0,  1,  0,  1,  0,  1,},
  {   1,  0,  1,  0,  1,  0,  1,  0,},
  {   0,  1,  0,  1,  0,  1,  0,  1,},
  {   1,  0,  1,  0,  1,  0,  1,  0,},
},{
  {   1,  2,  1,  2,  1,  2,  1,  2,},
  {   3,  0,  3,  0,  3,  0,  3,  0,},
  {   1,  2,  1,  2,  1,  2,  1,  2,},
  {   3,  0,  3,  0,  3,  0,  3,  0,},
  {   1,  2,  1,  2,  1,  2,  1,  2,},
  {   3,  0,  3,  0,  3,  0,  3,  0,},
  {   1,  2,  1,  2,  1,  2,  1,  2,},
  {   3,  0,  3,  0,  3,  0,  3,  0,},
},{
  {   2,  4,  3,  5,  2,  4,  3,  5,},
  {   6,  0,  7,  1,  6,  0,  7,  1,},
  {   3,  5,  2,  4,  3,  5,  2,  4,},
  {   7,  1,  6,  0,  7,  1,  6,  0,},
  {   2,  4,  3,  5,  2,  4,  3,  5,},
  {   6,  0,  7,  1,  6,  0,  7,  1,},
  {   3,  5,  2,  4,  3,  5,  2,  4,},
  {   7,  1,  6,  0,  7,  1,  6,  0,},
},{
  {   4,  8,  7, 11,  4,  8,  7, 11,},
  {  12,  0, 15,  3, 12,  0, 15,  3,},
  {   6, 10,  5,  9,  6, 10,  5,  9,},
  {  14,  2, 13,  1, 14,  2, 13,  1,},
  {   4,  8,  7, 11,  4,  8,  7, 11,},
  {  12,  0, 15,  3, 12,  0, 15,  3,},
  {   6, 10,  5,  9,  6, 10,  5,  9,},
  {  14,  2, 13,  1, 14,  2, 13,  1,},
},{
  {   9, 17, 15, 23,  8, 16, 14, 22,},
  {  25,  1, 31,  7, 24,  0, 30,  6,},
  {  13, 21, 11, 19, 12, 20, 10, 18,},
  {  29,  5, 27,  3, 28,  4, 26,  2,},
  {   8, 16, 14, 22,  9, 17, 15, 23,},
  {  24,  0, 30,  6, 25,  1, 31,  7,},
  {  12, 20, 10, 18, 13, 21, 11, 19,},
  {  28,  4, 26,  2, 29,  5, 27,  3,},
},{
  {  18, 34, 30, 46, 17, 33, 29, 45,},
  {  50,  2, 62, 14, 49,  1, 61, 13,},
  {  26, 42, 22, 38, 25, 41, 21, 37,},
  {  58, 10, 54,  6, 57,  9, 53,  5,},
  {  16, 32, 28, 44, 19, 35, 31, 47,},
  {  48,  0, 60, 12, 51,  3, 63, 15,},
  {  24, 40, 20, 36, 27, 43, 23, 39,},
  {  56,  8, 52,  4, 59, 11, 55,  7,},
},{
  {  18, 34, 30, 46, 17, 33, 29, 45,},
  {  50,  2, 62, 14, 49,  1, 61, 13,},
  {  26, 42, 22, 38, 25, 41, 21, 37,},
  {  58, 10, 54,  6, 57,  9, 53,  5,},
  {  16, 32, 28, 44, 19, 35, 31, 47,},
  {  48,  0, 60, 12, 51,  3, 63, 15,},
  {  24, 40, 20, 36, 27, 43, 23, 39,},
  {  56,  8, 52,  4, 59, 11, 55,  7,},
},{
  {  36, 68, 60, 92, 34, 66, 58, 90,},
  { 100,  4,124, 28, 98,  2,122, 26,},
  {  52, 84, 44, 76, 50, 82, 42, 74,},
  { 116, 20,108, 12,114, 18,106, 10,},
  {  32, 64, 56, 88, 38, 70, 62, 94,},
  {  96,  0,120, 24,102,  6,126, 30,},
  {  48, 80, 40, 72, 54, 86, 46, 78,},
  { 112, 16,104,  8,118, 22,110, 14,},
}};

static const uint8_t flat64[8]={64,64,64,64,64,64,64,64};

uint16_t dither_scale[15][16]={
{    2,    3,    3,    5,    5,    5,    5,    5,    5,    5,    5,    5,    5,    5,    5,    5,},
{    2,    3,    7,    7,   13,   13,   25,   25,   25,   25,   25,   25,   25,   25,   25,   25,},
{    3,    3,    4,   15,   15,   29,   57,   57,   57,  113,  113,  113,  113,  113,  113,  113,},
{    3,    4,    4,    5,   31,   31,   61,  121,  241,  241,  241,  241,  481,  481,  481,  481,},
{    3,    4,    5,    5,    6,   63,   63,  125,  249,  497,  993,  993,  993,  993,  993, 1985,},
{    3,    5,    6,    6,    6,    7,  127,  127,  253,  505, 1009, 2017, 4033, 4033, 4033, 4033,},
{    3,    5,    6,    7,    7,    7,    8,  255,  255,  509, 1017, 2033, 4065, 8129,16257,16257,},
{    3,    5,    6,    8,    8,    8,    8,    9,  511,  511, 1021, 2041, 4081, 8161,16321,32641,},
{    3,    5,    7,    8,    9,    9,    9,    9,   10, 1023, 1023, 2045, 4089, 8177,16353,32705,},
{    3,    5,    7,    8,   10,   10,   10,   10,   10,   11, 2047, 2047, 4093, 8185,16369,32737,},
{    3,    5,    7,    8,   10,   11,   11,   11,   11,   11,   12, 4095, 4095, 8189,16377,32753,},
{    3,    5,    7,    9,   10,   12,   12,   12,   12,   12,   12,   13, 8191, 8191,16381,32761,},
{    3,    5,    7,    9,   10,   12,   13,   13,   13,   13,   13,   13,   14,16383,16383,32765,},
{    3,    5,    7,    9,   10,   12,   14,   14,   14,   14,   14,   14,   14,   15,32767,32767,},
{    3,    5,    7,    9,   11,   12,   14,   15,   15,   15,   15,   15,   15,   15,   16,65535,},
};

static av_always_inline void yuv2yuvX16inC_template(const int16_t *lumFilter, const int16_t **lumSrc, int lumFilterSize,
                                                    const int16_t *chrFilter, const int16_t **chrUSrc,
                                                    const int16_t **chrVSrc, int chrFilterSize,
                                                    const int16_t **alpSrc, uint16_t *dest, uint16_t *uDest, uint16_t *vDest, uint16_t *aDest,
                                                    int dstW, int chrDstW, int big_endian, int output_bits)
{
    //FIXME Optimize (just quickly written not optimized..)
    int i;
    int shift = 11 + 16 - output_bits;

#define output_pixel(pos, val) \
    if (big_endian) { \
        if (output_bits == 16) { \
            AV_WB16(pos, av_clip_uint16(val >> shift)); \
        } else { \
            AV_WB16(pos, av_clip_uintp2(val >> shift, output_bits)); \
        } \
    } else { \
        if (output_bits == 16) { \
            AV_WL16(pos, av_clip_uint16(val >> shift)); \
        } else { \
            AV_WL16(pos, av_clip_uintp2(val >> shift, output_bits)); \
        } \
    }
    for (i = 0; i < dstW; i++) {
        int val = 1 << (26-output_bits);
        int j;

        for (j = 0; j < lumFilterSize; j++)
            val += lumSrc[j][i] * lumFilter[j];

        output_pixel(&dest[i], val);
    }

    if (uDest) {
        for (i = 0; i < chrDstW; i++) {
            int u = 1 << (26-output_bits);
            int v = 1 << (26-output_bits);
            int j;

            for (j = 0; j < chrFilterSize; j++) {
                u += chrUSrc[j][i] * chrFilter[j];
                v += chrVSrc[j][i] * chrFilter[j];
            }

            output_pixel(&uDest[i], u);
            output_pixel(&vDest[i], v);
        }
    }

    if (CONFIG_SWSCALE_ALPHA && aDest) {
        for (i = 0; i < dstW; i++) {
            int val = 1 << (26-output_bits);
            int j;

            for (j = 0; j < lumFilterSize; j++)
                val += alpSrc[j][i] * lumFilter[j];

            output_pixel(&aDest[i], val);
        }
    }
}

#define yuv2NBPS(bits, BE_LE, is_be) \
static void yuv2yuvX ## bits ## BE_LE ## _c(const int16_t *lumFilter, \
                              const int16_t **lumSrc, int lumFilterSize, \
                              const int16_t *chrFilter, const int16_t **chrUSrc, \
                              const int16_t **chrVSrc, \
                              int chrFilterSize, const int16_t **alpSrc, \
                              uint16_t *dest, uint16_t *uDest, uint16_t *vDest, \
                              uint16_t *aDest, int dstW, int chrDstW) \
{ \
    yuv2yuvX16inC_template(lumFilter, lumSrc, lumFilterSize, \
                           chrFilter, chrUSrc, chrVSrc, chrFilterSize, \
                           alpSrc, \
                           dest, uDest, vDest, aDest, \
                           dstW, chrDstW, is_be, bits); \
}
yuv2NBPS( 9, BE, 1);
yuv2NBPS( 9, LE, 0);
yuv2NBPS(10, BE, 1);
yuv2NBPS(10, LE, 0);
yuv2NBPS(16, BE, 1);
yuv2NBPS(16, LE, 0);

static inline void yuv2yuvX16inC(const int16_t *lumFilter, const int16_t **lumSrc, int lumFilterSize,
                                 const int16_t *chrFilter, const int16_t **chrUSrc, const int16_t **chrVSrc, int chrFilterSize,
                                 const int16_t **alpSrc, uint16_t *dest, uint16_t *uDest, uint16_t *vDest, uint16_t *aDest, int dstW, int chrDstW,
                                 enum PixelFormat dstFormat)
{
    if (isNBPS(dstFormat)) {
        const int depth = av_pix_fmt_descriptors[dstFormat].comp[0].depth_minus1+1;
        yuv2yuvX16inC_template(lumFilter, lumSrc, lumFilterSize,
                              chrFilter, chrUSrc, chrVSrc, chrFilterSize,
                              alpSrc,
                              dest, uDest, vDest, aDest,
                              dstW, chrDstW, isBE(dstFormat), depth);
    } else {
        if (isBE(dstFormat)) {
            yuv2yuvX16inC_template(lumFilter, lumSrc, lumFilterSize,
                                   chrFilter, chrUSrc, chrVSrc, chrFilterSize,
                                   alpSrc,
                                   dest, uDest, vDest, aDest,
                                   dstW, chrDstW, 1, 16);
        } else {
            yuv2yuvX16inC_template(lumFilter, lumSrc, lumFilterSize,
                                   chrFilter, chrUSrc, chrVSrc, chrFilterSize,
                                   alpSrc,
                                   dest, uDest, vDest, aDest,
                                   dstW, chrDstW, 0, 16);
        }
    }
}

static inline void yuv2yuvXinC(const int16_t *lumFilter, const int16_t **lumSrc, int lumFilterSize,
                               const int16_t *chrFilter, const int16_t **chrUSrc,
                               const int16_t **chrVSrc, int chrFilterSize,
                               const int16_t **alpSrc, uint8_t *dest, uint8_t *uDest, uint8_t *vDest, uint8_t *aDest, int dstW, int chrDstW,
                               const uint8_t *lumDither, const uint8_t *chrDither)
{
    //FIXME Optimize (just quickly written not optimized..)
    int i;
    for (i=0; i<dstW; i++) {
        int val = lumDither[i&7] << 12;
        int j;
        for (j=0; j<lumFilterSize; j++)
            val += lumSrc[j][i] * lumFilter[j];

        dest[i]= av_clip_uint8(val>>19);
    }

    if (uDest)
        for (i=0; i<chrDstW; i++) {
            int u = chrDither[i&7] << 12;
            int v = chrDither[(i+3)&7] << 12;
            int j;
            for (j=0; j<chrFilterSize; j++) {
                u += chrUSrc[j][i] * chrFilter[j];
                v += chrVSrc[j][i] * chrFilter[j];
            }

            uDest[i]= av_clip_uint8(u>>19);
            vDest[i]= av_clip_uint8(v>>19);
        }

    if (CONFIG_SWSCALE_ALPHA && aDest)
        for (i=0; i<dstW; i++) {
            int val = lumDither[i&7] << 12;
            int j;
            for (j=0; j<lumFilterSize; j++)
                val += alpSrc[j][i] * lumFilter[j];

            aDest[i]= av_clip_uint8(val>>19);
        }

}

static inline void yuv2nv12XinC(const int16_t *lumFilter, const int16_t **lumSrc, int lumFilterSize,
                                const int16_t *chrFilter, const int16_t **chrUSrc,
                                const int16_t **chrVSrc, int chrFilterSize,
                                uint8_t *dest, uint8_t *uDest, int dstW, int chrDstW, int dstFormat,
                                const uint8_t *lumDither, const uint8_t *chrDither)
{
    //FIXME Optimize (just quickly written not optimized..)
    int i;
    for (i=0; i<dstW; i++) {
        int val = lumDither[i&7]<<12;
        int j;
        for (j=0; j<lumFilterSize; j++)
            val += lumSrc[j][i] * lumFilter[j];

        dest[i]= av_clip_uint8(val>>19);
    }

    if (!uDest)
        return;

    if (dstFormat == PIX_FMT_NV12)
        for (i=0; i<chrDstW; i++) {
            int u = chrDither[i&7]<<12;
            int v = chrDither[(i+3)&7]<<12;
            int j;
            for (j=0; j<chrFilterSize; j++) {
                u += chrUSrc[j][i] * chrFilter[j];
                v += chrVSrc[j][i] * chrFilter[j];
            }

            uDest[2*i]= av_clip_uint8(u>>19);
            uDest[2*i+1]= av_clip_uint8(v>>19);
        }
    else
        for (i=0; i<chrDstW; i++) {
            int u = chrDither[i&7]<<12;
            int v = chrDither[(i+3)&7]<<12;
            int j;
            for (j=0; j<chrFilterSize; j++) {
                u += chrUSrc[j][i] * chrFilter[j];
                v += chrVSrc[j][i] * chrFilter[j];
            }

            uDest[2*i]= av_clip_uint8(v>>19);
            uDest[2*i+1]= av_clip_uint8(u>>19);
        }
}

#define YSCALE_YUV_2_PACKEDX_NOCLIP_C(type,alpha) \
    for (i=0; i<(dstW>>1); i++) {\
        int j;\
        int Y1 = 1<<18;\
        int Y2 = 1<<18;\
        int U  = 1<<18;\
        int V  = 1<<18;\
        int av_unused A1, A2;\
        type av_unused *r, *b, *g;\
        const int i2= 2*i;\
        \
        for (j=0; j<lumFilterSize; j++) {\
            Y1 += lumSrc[j][i2] * lumFilter[j];\
            Y2 += lumSrc[j][i2+1] * lumFilter[j];\
        }\
        for (j=0; j<chrFilterSize; j++) {\
            U += chrUSrc[j][i] * chrFilter[j];\
            V += chrVSrc[j][i] * chrFilter[j];\
        }\
        Y1>>=19;\
        Y2>>=19;\
        U >>=19;\
        V >>=19;\
        if (alpha) {\
            A1 = 1<<18;\
            A2 = 1<<18;\
            for (j=0; j<lumFilterSize; j++) {\
                A1 += alpSrc[j][i2  ] * lumFilter[j];\
                A2 += alpSrc[j][i2+1] * lumFilter[j];\
            }\
            A1>>=19;\
            A2>>=19;\
        }

#define YSCALE_YUV_2_PACKEDX_C(type,alpha) \
        YSCALE_YUV_2_PACKEDX_NOCLIP_C(type,alpha)\
        if ((Y1|Y2|U|V)&256) {\
            if (Y1>255)   Y1=255; \
            else if (Y1<0)Y1=0;   \
            if (Y2>255)   Y2=255; \
            else if (Y2<0)Y2=0;   \
            if (U>255)    U=255;  \
            else if (U<0) U=0;    \
            if (V>255)    V=255;  \
            else if (V<0) V=0;    \
        }\
        if (alpha && ((A1|A2)&256)) {\
            A1=av_clip_uint8(A1);\
            A2=av_clip_uint8(A2);\
        }

#define YSCALE_YUV_2_PACKEDX_FULL_C(rnd,alpha) \
    for (i=0; i<dstW; i++) {\
        int j;\
        int Y = 1<<9;\
        int U = (1<<9)-(128<<19);\
        int V = (1<<9)-(128<<19);\
        int av_unused A;\
        int R,G,B;\
        \
        for (j=0; j<lumFilterSize; j++) {\
            Y += lumSrc[j][i     ] * lumFilter[j];\
        }\
        for (j=0; j<chrFilterSize; j++) {\
            U += chrUSrc[j][i] * chrFilter[j];\
            V += chrVSrc[j][i] * chrFilter[j];\
        }\
        Y >>=10;\
        U >>=10;\
        V >>=10;\
        if (alpha) {\
            A = rnd;\
            for (j=0; j<lumFilterSize; j++)\
                A += alpSrc[j][i     ] * lumFilter[j];\
            A >>=19;\
            if (A&256)\
                A = av_clip_uint8(A);\
        }

#define YSCALE_YUV_2_RGBX_FULL_C(rnd,alpha) \
    YSCALE_YUV_2_PACKEDX_FULL_C(rnd>>3,alpha)\
        Y-= c->yuv2rgb_y_offset;\
        Y*= c->yuv2rgb_y_coeff;\
        Y+= rnd;\
        R= Y + V*c->yuv2rgb_v2r_coeff;\
        G= Y + V*c->yuv2rgb_v2g_coeff + U*c->yuv2rgb_u2g_coeff;\
        B= Y +                          U*c->yuv2rgb_u2b_coeff;\
        if ((R|G|B)&(0xC0000000)) {\
            if (R>=(256<<22))   R=(256<<22)-1; \
            else if (R<0)R=0;   \
            if (G>=(256<<22))   G=(256<<22)-1; \
            else if (G<0)G=0;   \
            if (B>=(256<<22))   B=(256<<22)-1; \
            else if (B<0)B=0;   \
        }

#define YSCALE_YUV_2_GRAY16_C \
    for (i=0; i<(dstW>>1); i++) {\
        int j;\
        int Y1 = 1<<18;\
        int Y2 = 1<<18;\
        int U  = 1<<18;\
        int V  = 1<<18;\
        \
        const int i2= 2*i;\
        \
        for (j=0; j<lumFilterSize; j++) {\
            Y1 += lumSrc[j][i2] * lumFilter[j];\
            Y2 += lumSrc[j][i2+1] * lumFilter[j];\
        }\
        Y1>>=11;\
        Y2>>=11;\
        if ((Y1|Y2|U|V)&65536) {\
            if (Y1>65535)   Y1=65535; \
            else if (Y1<0)Y1=0;   \
            if (Y2>65535)   Y2=65535; \
            else if (Y2<0)Y2=0;   \
        }

#define YSCALE_YUV_2_RGBX_C(type,alpha) \
    YSCALE_YUV_2_PACKEDX_C(type,alpha)  /* FIXME fix tables so that clipping is not needed and then use _NOCLIP*/\
    r = (type *)c->table_rV[V];   \
    g = (type *)(c->table_gU[U] + c->table_gV[V]); \
    b = (type *)c->table_bU[U];

#define YSCALE_YUV_2_PACKED2_C(type,alpha)   \
    for (i=0; i<(dstW>>1); i++) { \
        const int i2= 2*i;       \
        int Y1= (buf0[i2  ]*yalpha1+buf1[i2  ]*yalpha)>>19;           \
        int Y2= (buf0[i2+1]*yalpha1+buf1[i2+1]*yalpha)>>19;           \
        int U= (ubuf0[i]*uvalpha1+ubuf1[i]*uvalpha)>>19;              \
        int V= (vbuf0[i]*uvalpha1+vbuf1[i]*uvalpha)>>19;              \
        type av_unused *r, *b, *g;                                    \
        int av_unused A1, A2;                                         \
        if (alpha) {\
            A1= (abuf0[i2  ]*yalpha1+abuf1[i2  ]*yalpha)>>19;         \
            A2= (abuf0[i2+1]*yalpha1+abuf1[i2+1]*yalpha)>>19;         \
        }

#define YSCALE_YUV_2_GRAY16_2_C   \
    for (i=0; i<(dstW>>1); i++) { \
        const int i2= 2*i;       \
        int Y1= (buf0[i2  ]*yalpha1+buf1[i2  ]*yalpha)>>11;           \
        int Y2= (buf0[i2+1]*yalpha1+buf1[i2+1]*yalpha)>>11;

#define YSCALE_YUV_2_RGB2_C(type,alpha) \
    YSCALE_YUV_2_PACKED2_C(type,alpha)\
    r = (type *)c->table_rV[V];\
    g = (type *)(c->table_gU[U] + c->table_gV[V]);\
    b = (type *)c->table_bU[U];

#define YSCALE_YUV_2_PACKED1_C(type,alpha) \
    for (i=0; i<(dstW>>1); i++) {\
        const int i2= 2*i;\
        int Y1= buf0[i2  ]>>7;\
        int Y2= buf0[i2+1]>>7;\
        int U= (ubuf1[i])>>7;\
        int V= (vbuf1[i])>>7;\
        type av_unused *r, *b, *g;\
        int av_unused A1, A2;\
        if (alpha) {\
            A1= abuf0[i2  ]>>7;\
            A2= abuf0[i2+1]>>7;\
        }

#define YSCALE_YUV_2_GRAY16_1_C \
    for (i=0; i<(dstW>>1); i++) {\
        const int i2= 2*i;\
        int Y1= buf0[i2  ]<<1;\
        int Y2= buf0[i2+1]<<1;

#define YSCALE_YUV_2_RGB1_C(type,alpha) \
    YSCALE_YUV_2_PACKED1_C(type,alpha)\
    r = (type *)c->table_rV[V];\
    g = (type *)(c->table_gU[U] + c->table_gV[V]);\
    b = (type *)c->table_bU[U];

#define YSCALE_YUV_2_PACKED1B_C(type,alpha) \
    for (i=0; i<(dstW>>1); i++) {\
        const int i2= 2*i;\
        int Y1= buf0[i2  ]>>7;\
        int Y2= buf0[i2+1]>>7;\
        int U= (ubuf0[i] + ubuf1[i])>>8;\
        int V= (vbuf0[i] + vbuf1[i])>>8;\
        type av_unused *r, *b, *g;\
        int av_unused A1, A2;\
        if (alpha) {\
            A1= abuf0[i2  ]>>7;\
            A2= abuf0[i2+1]>>7;\
        }

#define YSCALE_YUV_2_RGB1B_C(type,alpha) \
    YSCALE_YUV_2_PACKED1B_C(type,alpha)\
    r = (type *)c->table_rV[V];\
    g = (type *)(c->table_gU[U] + c->table_gV[V]);\
    b = (type *)c->table_bU[U];

#define YSCALE_YUV_2_MONO2_C \
    const uint8_t * const d128=dither_8x8_220[y&7];\
    uint8_t *g= c->table_gU[128] + c->table_gV[128];\
    for (i=0; i<dstW-7; i+=8) {\
        int acc;\
        acc =       g[((buf0[i  ]*yalpha1+buf1[i  ]*yalpha)>>19) + d128[0]];\
        acc+= acc + g[((buf0[i+1]*yalpha1+buf1[i+1]*yalpha)>>19) + d128[1]];\
        acc+= acc + g[((buf0[i+2]*yalpha1+buf1[i+2]*yalpha)>>19) + d128[2]];\
        acc+= acc + g[((buf0[i+3]*yalpha1+buf1[i+3]*yalpha)>>19) + d128[3]];\
        acc+= acc + g[((buf0[i+4]*yalpha1+buf1[i+4]*yalpha)>>19) + d128[4]];\
        acc+= acc + g[((buf0[i+5]*yalpha1+buf1[i+5]*yalpha)>>19) + d128[5]];\
        acc+= acc + g[((buf0[i+6]*yalpha1+buf1[i+6]*yalpha)>>19) + d128[6]];\
        acc+= acc + g[((buf0[i+7]*yalpha1+buf1[i+7]*yalpha)>>19) + d128[7]];\
        ((uint8_t*)dest)[0]= c->dstFormat == PIX_FMT_MONOBLACK ? acc : ~acc;\
        dest++;\
    }

#define YSCALE_YUV_2_MONOX_C \
    const uint8_t * const d128=dither_8x8_220[y&7];\
    uint8_t *g= c->table_gU[128] + c->table_gV[128];\
    int acc=0;\
    for (i=0; i<dstW-1; i+=2) {\
        int j;\
        int Y1=1<<18;\
        int Y2=1<<18;\
\
        for (j=0; j<lumFilterSize; j++) {\
            Y1 += lumSrc[j][i] * lumFilter[j];\
            Y2 += lumSrc[j][i+1] * lumFilter[j];\
        }\
        Y1>>=19;\
        Y2>>=19;\
        if ((Y1|Y2)&256) {\
            if (Y1>255)   Y1=255;\
            else if (Y1<0)Y1=0;\
            if (Y2>255)   Y2=255;\
            else if (Y2<0)Y2=0;\
        }\
        acc+= acc + g[Y1+d128[(i+0)&7]];\
        acc+= acc + g[Y2+d128[(i+1)&7]];\
        if ((i&7)==6) {\
            ((uint8_t*)dest)[0]= c->dstFormat == PIX_FMT_MONOBLACK ? acc : ~acc;\
            dest++;\
        }\
    }

#define YSCALE_YUV_2_ANYRGB_C(func, func2, func_g16, func_monoblack)\
    switch(c->dstFormat) {\
    case PIX_FMT_RGB48BE:\
    case PIX_FMT_RGB48LE:\
        func(uint8_t,0)\
            ((uint8_t*)dest)[ 0]= r[Y1];\
            ((uint8_t*)dest)[ 1]= r[Y1];\
            ((uint8_t*)dest)[ 2]= g[Y1];\
            ((uint8_t*)dest)[ 3]= g[Y1];\
            ((uint8_t*)dest)[ 4]= b[Y1];\
            ((uint8_t*)dest)[ 5]= b[Y1];\
            ((uint8_t*)dest)[ 6]= r[Y2];\
            ((uint8_t*)dest)[ 7]= r[Y2];\
            ((uint8_t*)dest)[ 8]= g[Y2];\
            ((uint8_t*)dest)[ 9]= g[Y2];\
            ((uint8_t*)dest)[10]= b[Y2];\
            ((uint8_t*)dest)[11]= b[Y2];\
            dest+=12;\
        }\
        break;\
    case PIX_FMT_BGR48BE:\
    case PIX_FMT_BGR48LE:\
        func(uint8_t,0)\
            ((uint8_t*)dest)[ 0] = ((uint8_t*)dest)[ 1] = b[Y1];\
            ((uint8_t*)dest)[ 2] = ((uint8_t*)dest)[ 3] = g[Y1];\
            ((uint8_t*)dest)[ 4] = ((uint8_t*)dest)[ 5] = r[Y1];\
            ((uint8_t*)dest)[ 6] = ((uint8_t*)dest)[ 7] = b[Y2];\
            ((uint8_t*)dest)[ 8] = ((uint8_t*)dest)[ 9] = g[Y2];\
            ((uint8_t*)dest)[10] = ((uint8_t*)dest)[11] = r[Y2];\
            dest+=12;\
        }\
        break;\
    case PIX_FMT_RGBA:\
    case PIX_FMT_BGRA:\
        if (CONFIG_SMALL) {\
            int needAlpha = CONFIG_SWSCALE_ALPHA && c->alpPixBuf;\
            func(uint32_t,needAlpha)\
                ((uint32_t*)dest)[i2+0]= r[Y1] + g[Y1] + b[Y1] + (needAlpha ? (A1<<24) : 0);\
                ((uint32_t*)dest)[i2+1]= r[Y2] + g[Y2] + b[Y2] + (needAlpha ? (A2<<24) : 0);\
            }\
        } else {\
            if (CONFIG_SWSCALE_ALPHA && c->alpPixBuf) {\
                func(uint32_t,1)\
                    ((uint32_t*)dest)[i2+0]= r[Y1] + g[Y1] + b[Y1] + (A1<<24);\
                    ((uint32_t*)dest)[i2+1]= r[Y2] + g[Y2] + b[Y2] + (A2<<24);\
                }\
            } else {\
                func(uint32_t,0)\
                    ((uint32_t*)dest)[i2+0]= r[Y1] + g[Y1] + b[Y1];\
                    ((uint32_t*)dest)[i2+1]= r[Y2] + g[Y2] + b[Y2];\
                }\
            }\
        }\
        break;\
    case PIX_FMT_ARGB:\
    case PIX_FMT_ABGR:\
        if (CONFIG_SMALL) {\
            int needAlpha = CONFIG_SWSCALE_ALPHA && c->alpPixBuf;\
            func(uint32_t,needAlpha)\
                ((uint32_t*)dest)[i2+0]= r[Y1] + g[Y1] + b[Y1] + (needAlpha ? A1 : 0);\
                ((uint32_t*)dest)[i2+1]= r[Y2] + g[Y2] + b[Y2] + (needAlpha ? A2 : 0);\
            }\
        } else {\
            if (CONFIG_SWSCALE_ALPHA && c->alpPixBuf) {\
                func(uint32_t,1)\
                    ((uint32_t*)dest)[i2+0]= r[Y1] + g[Y1] + b[Y1] + A1;\
                    ((uint32_t*)dest)[i2+1]= r[Y2] + g[Y2] + b[Y2] + A2;\
                }\
            } else {\
                func(uint32_t,0)\
                    ((uint32_t*)dest)[i2+0]= r[Y1] + g[Y1] + b[Y1];\
                    ((uint32_t*)dest)[i2+1]= r[Y2] + g[Y2] + b[Y2];\
                }\
            }\
        }                \
        break;\
    case PIX_FMT_RGB24:\
        func(uint8_t,0)\
            ((uint8_t*)dest)[0]= r[Y1];\
            ((uint8_t*)dest)[1]= g[Y1];\
            ((uint8_t*)dest)[2]= b[Y1];\
            ((uint8_t*)dest)[3]= r[Y2];\
            ((uint8_t*)dest)[4]= g[Y2];\
            ((uint8_t*)dest)[5]= b[Y2];\
            dest+=6;\
        }\
        break;\
    case PIX_FMT_BGR24:\
        func(uint8_t,0)\
            ((uint8_t*)dest)[0]= b[Y1];\
            ((uint8_t*)dest)[1]= g[Y1];\
            ((uint8_t*)dest)[2]= r[Y1];\
            ((uint8_t*)dest)[3]= b[Y2];\
            ((uint8_t*)dest)[4]= g[Y2];\
            ((uint8_t*)dest)[5]= r[Y2];\
            dest+=6;\
        }\
        break;\
    case PIX_FMT_RGB565BE:\
    case PIX_FMT_RGB565LE:\
    case PIX_FMT_BGR565BE:\
    case PIX_FMT_BGR565LE:\
        {\
            const int dr1= dither_2x2_8[y&1    ][0];\
            const int dg1= dither_2x2_4[y&1    ][0];\
            const int db1= dither_2x2_8[(y&1)^1][0];\
            const int dr2= dither_2x2_8[y&1    ][1];\
            const int dg2= dither_2x2_4[y&1    ][1];\
            const int db2= dither_2x2_8[(y&1)^1][1];\
            func(uint16_t,0)\
                ((uint16_t*)dest)[i2+0]= r[Y1+dr1] + g[Y1+dg1] + b[Y1+db1];\
                ((uint16_t*)dest)[i2+1]= r[Y2+dr2] + g[Y2+dg2] + b[Y2+db2];\
            }\
        }\
        break;\
    case PIX_FMT_RGB555BE:\
    case PIX_FMT_RGB555LE:\
    case PIX_FMT_BGR555BE:\
    case PIX_FMT_BGR555LE:\
        {\
            const int dr1= dither_2x2_8[y&1    ][0];\
            const int dg1= dither_2x2_8[y&1    ][1];\
            const int db1= dither_2x2_8[(y&1)^1][0];\
            const int dr2= dither_2x2_8[y&1    ][1];\
            const int dg2= dither_2x2_8[y&1    ][0];\
            const int db2= dither_2x2_8[(y&1)^1][1];\
            func(uint16_t,0)\
                ((uint16_t*)dest)[i2+0]= r[Y1+dr1] + g[Y1+dg1] + b[Y1+db1];\
                ((uint16_t*)dest)[i2+1]= r[Y2+dr2] + g[Y2+dg2] + b[Y2+db2];\
            }\
        }\
        break;\
    case PIX_FMT_RGB444BE:\
    case PIX_FMT_RGB444LE:\
    case PIX_FMT_BGR444BE:\
    case PIX_FMT_BGR444LE:\
        {\
            const int dr1= dither_4x4_16[y&3    ][0];\
            const int dg1= dither_4x4_16[y&3    ][1];\
            const int db1= dither_4x4_16[(y&3)^3][0];\
            const int dr2= dither_4x4_16[y&3    ][1];\
            const int dg2= dither_4x4_16[y&3    ][0];\
            const int db2= dither_4x4_16[(y&3)^3][1];\
            func(uint16_t,0)\
                ((uint16_t*)dest)[i2+0]= r[Y1+dr1] + g[Y1+dg1] + b[Y1+db1];\
                ((uint16_t*)dest)[i2+1]= r[Y2+dr2] + g[Y2+dg2] + b[Y2+db2];\
            }\
        }\
        break;\
    case PIX_FMT_RGB8:\
    case PIX_FMT_BGR8:\
        {\
            const uint8_t * const d64= dither_8x8_73[y&7];\
            const uint8_t * const d32= dither_8x8_32[y&7];\
            func(uint8_t,0)\
                ((uint8_t*)dest)[i2+0]= r[Y1+d32[(i2+0)&7]] + g[Y1+d32[(i2+0)&7]] + b[Y1+d64[(i2+0)&7]];\
                ((uint8_t*)dest)[i2+1]= r[Y2+d32[(i2+1)&7]] + g[Y2+d32[(i2+1)&7]] + b[Y2+d64[(i2+1)&7]];\
            }\
        }\
        break;\
    case PIX_FMT_RGB4:\
    case PIX_FMT_BGR4:\
        {\
            const uint8_t * const d64= dither_8x8_73 [y&7];\
            const uint8_t * const d128=dither_8x8_220[y&7];\
            func(uint8_t,0)\
                ((uint8_t*)dest)[i]= r[Y1+d128[(i2+0)&7]] + g[Y1+d64[(i2+0)&7]] + b[Y1+d128[(i2+0)&7]]\
                                 + ((r[Y2+d128[(i2+1)&7]] + g[Y2+d64[(i2+1)&7]] + b[Y2+d128[(i2+1)&7]])<<4);\
            }\
        }\
        break;\
    case PIX_FMT_RGB4_BYTE:\
    case PIX_FMT_BGR4_BYTE:\
        {\
            const uint8_t * const d64= dither_8x8_73 [y&7];\
            const uint8_t * const d128=dither_8x8_220[y&7];\
            func(uint8_t,0)\
                ((uint8_t*)dest)[i2+0]= r[Y1+d128[(i2+0)&7]] + g[Y1+d64[(i2+0)&7]] + b[Y1+d128[(i2+0)&7]];\
                ((uint8_t*)dest)[i2+1]= r[Y2+d128[(i2+1)&7]] + g[Y2+d64[(i2+1)&7]] + b[Y2+d128[(i2+1)&7]];\
            }\
        }\
        break;\
    case PIX_FMT_MONOBLACK:\
    case PIX_FMT_MONOWHITE:\
        {\
            func_monoblack\
        }\
        break;\
    case PIX_FMT_YUYV422:\
        func2\
            ((uint8_t*)dest)[2*i2+0]= Y1;\
            ((uint8_t*)dest)[2*i2+1]= U;\
            ((uint8_t*)dest)[2*i2+2]= Y2;\
            ((uint8_t*)dest)[2*i2+3]= V;\
        }                \
        break;\
    case PIX_FMT_UYVY422:\
        func2\
            ((uint8_t*)dest)[2*i2+0]= U;\
            ((uint8_t*)dest)[2*i2+1]= Y1;\
            ((uint8_t*)dest)[2*i2+2]= V;\
            ((uint8_t*)dest)[2*i2+3]= Y2;\
        }                \
        break;\
    case PIX_FMT_GRAY16BE:\
        func_g16\
            ((uint8_t*)dest)[2*i2+0]= Y1>>8;\
            ((uint8_t*)dest)[2*i2+1]= Y1;\
            ((uint8_t*)dest)[2*i2+2]= Y2>>8;\
            ((uint8_t*)dest)[2*i2+3]= Y2;\
        }                \
        break;\
    case PIX_FMT_GRAY16LE:\
        func_g16\
            ((uint8_t*)dest)[2*i2+0]= Y1;\
            ((uint8_t*)dest)[2*i2+1]= Y1>>8;\
            ((uint8_t*)dest)[2*i2+2]= Y2;\
            ((uint8_t*)dest)[2*i2+3]= Y2>>8;\
        }                \
        break;\
    }

static inline void yuv2packedXinC(SwsContext *c, const int16_t *lumFilter,
                                  const int16_t **lumSrc, int lumFilterSize,
                                  const int16_t *chrFilter, const int16_t **chrUSrc,
                                  const int16_t **chrVSrc, int chrFilterSize,
                                  const int16_t **alpSrc, uint8_t *dest, int dstW, int y)
{
    int i;
    YSCALE_YUV_2_ANYRGB_C(YSCALE_YUV_2_RGBX_C, YSCALE_YUV_2_PACKEDX_C(void,0), YSCALE_YUV_2_GRAY16_C, YSCALE_YUV_2_MONOX_C)
}

static inline void yuv2rgbXinC_full(SwsContext *c, const int16_t *lumFilter,
                                    const int16_t **lumSrc, int lumFilterSize,
                                    const int16_t *chrFilter, const int16_t **chrUSrc,
                                    const int16_t **chrVSrc, int chrFilterSize,
                                    const int16_t **alpSrc, uint8_t *dest, int dstW, int y)
{
    int i;
    int step= c->dstFormatBpp/8;
    int aidx= 3;

    switch(c->dstFormat) {
    case PIX_FMT_ARGB:
        dest++;
        aidx= 0;
    case PIX_FMT_RGB24:
        aidx--;
    case PIX_FMT_RGBA:
        if (CONFIG_SMALL) {
            int needAlpha = CONFIG_SWSCALE_ALPHA && c->alpPixBuf;
            YSCALE_YUV_2_RGBX_FULL_C(1<<21, needAlpha)
                dest[aidx]= needAlpha ? A : 255;
                dest[0]= R>>22;
                dest[1]= G>>22;
                dest[2]= B>>22;
                dest+= step;
            }
        } else {
            if (CONFIG_SWSCALE_ALPHA && c->alpPixBuf) {
                YSCALE_YUV_2_RGBX_FULL_C(1<<21, 1)
                    dest[aidx]= A;
                    dest[0]= R>>22;
                    dest[1]= G>>22;
                    dest[2]= B>>22;
                    dest+= step;
                }
            } else {
                YSCALE_YUV_2_RGBX_FULL_C(1<<21, 0)
                    dest[aidx]= 255;
                    dest[0]= R>>22;
                    dest[1]= G>>22;
                    dest[2]= B>>22;
                    dest+= step;
                }
            }
        }
        break;
    case PIX_FMT_ABGR:
        dest++;
        aidx= 0;
    case PIX_FMT_BGR24:
        aidx--;
    case PIX_FMT_BGRA:
        if (CONFIG_SMALL) {
            int needAlpha = CONFIG_SWSCALE_ALPHA && c->alpPixBuf;
            YSCALE_YUV_2_RGBX_FULL_C(1<<21, needAlpha)
                dest[aidx]= needAlpha ? A : 255;
                dest[0]= B>>22;
                dest[1]= G>>22;
                dest[2]= R>>22;
                dest+= step;
            }
        } else {
            if (CONFIG_SWSCALE_ALPHA && c->alpPixBuf) {
                YSCALE_YUV_2_RGBX_FULL_C(1<<21, 1)
                    dest[aidx]= A;
                    dest[0]= B>>22;
                    dest[1]= G>>22;
                    dest[2]= R>>22;
                    dest+= step;
                }
            } else {
                YSCALE_YUV_2_RGBX_FULL_C(1<<21, 0)
                    dest[aidx]= 255;
                    dest[0]= B>>22;
                    dest[1]= G>>22;
                    dest[2]= R>>22;
                    dest+= step;
                }
            }
        }
        break;
    default:
        assert(0);
    }
}

static void fillPlane(uint8_t* plane, int stride, int width, int height, int y, uint8_t val)
{
    int i;
    uint8_t *ptr = plane + stride*y;
    for (i=0; i<height; i++) {
        memset(ptr, val, width);
        ptr += stride;
    }
}

#define RGB48(name, R, B, READ)\
static inline void name ## ToY(int16_t *dst, const uint16_t *src, int width, uint32_t *unused)\
{\
    int i;\
    for (i = 0; i < width; i++) {\
        int r = READ(&src[i*3+R]);\
        int g = READ(&src[i*3+1]);\
        int b = READ(&src[i*3+B]);\
\
        dst[i] = (RY*r + GY*g + BY*b + (32<<(RGB2YUV_SHIFT-1+8)) + (1<<(RGB2YUV_SHIFT-7+8))) >> (RGB2YUV_SHIFT-6+8);\
    }\
}\
\
static inline void name ## ToUV(int16_t *dstU, int16_t *dstV,\
                             const uint16_t *src1, const uint16_t *src2,\
                             int width, uint32_t *unused)\
{\
    int i;\
    assert(src1==src2);\
    for (i = 0; i < width; i++) {\
        int r = READ(&src1[3*i + R]);\
        int g = READ(&src1[3*i + 1]);\
        int b = READ(&src1[3*i + B]);\
\
        dstU[i] = (RU*r + GU*g + BU*b + (256<<(RGB2YUV_SHIFT-1+8)) + (1<<(RGB2YUV_SHIFT-7+8))) >> (RGB2YUV_SHIFT-6+8);\
        dstV[i] = (RV*r + GV*g + BV*b + (256<<(RGB2YUV_SHIFT-1+8)) + (1<<(RGB2YUV_SHIFT-7+8))) >> (RGB2YUV_SHIFT-6+8);\
    }\
}\
\
static inline void name ## ToUV_half(int16_t *dstU, int16_t *dstV,\
                                  const uint16_t *src1, const uint16_t *src2,\
                                  int width, uint32_t *unused)\
{\
    int i;\
    assert(src1==src2);\
    for (i = 0; i < width; i++) {\
        int r= READ(&src1[6*i + R]) + READ(&src1[6*i + 3+R]);\
        int g= READ(&src1[6*i + 1]) + READ(&src1[6*i + 4]);\
        int b= READ(&src1[6*i + B]) + READ(&src1[6*i + 3+B]);\
\
        dstU[i]= (RU*r + GU*g + BU*b + (256U<<(RGB2YUV_SHIFT+8)) + (1<<(RGB2YUV_SHIFT-6+8))) >> (RGB2YUV_SHIFT-5+8);\
        dstV[i]= (RV*r + GV*g + BV*b + (256U<<(RGB2YUV_SHIFT+8)) + (1<<(RGB2YUV_SHIFT-6+8))) >> (RGB2YUV_SHIFT-5+8);\
    }\
}

RGB48(rgb48LE, 0, 2, AV_RL16)
RGB48(rgb48BE, 0, 2, AV_RB16)
RGB48(bgr48LE, 2, 0, AV_RL16)
RGB48(bgr48BE, 2, 0, AV_RB16)

#define BGR2Y(type, name, shr, shg, shb, maskr, maskg, maskb, RY, GY, BY, S)\
static inline void name(int16_t *dst, const uint8_t *src, int width, uint32_t *unused)\
{\
    int i;\
    for (i=0; i<width; i++) {\
        int b= (((const type*)src)[i]>>shb)&maskb;\
        int g= (((const type*)src)[i]>>shg)&maskg;\
        int r= (((const type*)src)[i]>>shr)&maskr;\
\
        dst[i]= (((RY)*r + (GY)*g + (BY)*b + (32<<((S)-1)) + (1<<(S-7)))>>((S)-6));\
    }\
}

BGR2Y(uint32_t, bgr32ToY,16, 0, 0, 0x00FF, 0xFF00, 0x00FF, RY<< 8, GY   , BY<< 8, RGB2YUV_SHIFT+8)
BGR2Y(uint32_t,bgr321ToY,16,16, 0, 0xFF00, 0x00FF, 0xFF00, RY    , GY<<8, BY    , RGB2YUV_SHIFT+8)
BGR2Y(uint32_t, rgb32ToY, 0, 0,16, 0x00FF, 0xFF00, 0x00FF, RY<< 8, GY   , BY<< 8, RGB2YUV_SHIFT+8)
BGR2Y(uint32_t,rgb321ToY, 0,16,16, 0xFF00, 0x00FF, 0xFF00, RY    , GY<<8, BY    , RGB2YUV_SHIFT+8)
BGR2Y(uint16_t, bgr16ToY, 0, 0, 0, 0x001F, 0x07E0, 0xF800, RY<<11, GY<<5, BY    , RGB2YUV_SHIFT+8)
BGR2Y(uint16_t, bgr15ToY, 0, 0, 0, 0x001F, 0x03E0, 0x7C00, RY<<10, GY<<5, BY    , RGB2YUV_SHIFT+7)
BGR2Y(uint16_t, rgb16ToY, 0, 0, 0, 0xF800, 0x07E0, 0x001F, RY    , GY<<5, BY<<11, RGB2YUV_SHIFT+8)
BGR2Y(uint16_t, rgb15ToY, 0, 0, 0, 0x7C00, 0x03E0, 0x001F, RY    , GY<<5, BY<<10, RGB2YUV_SHIFT+7)

static inline void abgrToA(int16_t *dst, const uint8_t *src, int width, uint32_t *unused)
{
    int i;
    for (i=0; i<width; i++) {
        dst[i]= src[4*i]<<6;
    }
}

#define BGR2UV(type, name, shr, shg, shb, shp, maskr, maskg, maskb, RU, GU, BU, RV, GV, BV, S) \
static inline void name(int16_t *dstU, int16_t *dstV, const uint8_t *src, const uint8_t *dummy, int width, uint32_t *unused)\
{\
    int i;\
    for (i=0; i<width; i++) {\
        int b= ((((const type*)src)[i]>>shp)&maskb)>>shb;\
        int g= ((((const type*)src)[i]>>shp)&maskg)>>shg;\
        int r= ((((const type*)src)[i]>>shp)&maskr)>>shr;\
\
        dstU[i]= ((RU)*r + (GU)*g + (BU)*b + (256<<((S)-1)) + (1<<(S-7)))>>((S)-6);\
        dstV[i]= ((RV)*r + (GV)*g + (BV)*b + (256<<((S)-1)) + (1<<(S-7)))>>((S)-6);\
    }\
}\
static inline void name ## _half(int16_t *dstU, int16_t *dstV, const uint8_t *src, const uint8_t *dummy, int width, uint32_t *unused)\
{\
    int i;\
    for (i=0; i<width; i++) {\
        int pix0= ((const type*)src)[2*i+0]>>shp;\
        int pix1= ((const type*)src)[2*i+1]>>shp;\
        int g= (pix0&~(maskr|maskb))+(pix1&~(maskr|maskb));\
        int b= ((pix0+pix1-g)&(maskb|(2*maskb)))>>shb;\
        int r= ((pix0+pix1-g)&(maskr|(2*maskr)))>>shr;\
        g&= maskg|(2*maskg);\
\
        g>>=shg;\
\
        dstU[i]= ((RU)*r + (GU)*g + (BU)*b + (256U<<(S)) + (1<<(S-6)))>>((S)-6+1);\
        dstV[i]= ((RV)*r + (GV)*g + (BV)*b + (256U<<(S)) + (1<<(S-6)))>>((S)-6+1);\
    }\
}

BGR2UV(uint32_t, bgr32ToUV,16, 0, 0, 0, 0xFF0000, 0xFF00,   0x00FF, RU<< 8, GU   , BU<< 8, RV<< 8, GV   , BV<< 8, RGB2YUV_SHIFT+8)
BGR2UV(uint32_t,bgr321ToUV,16, 0, 0, 8, 0xFF0000, 0xFF00,   0x00FF, RU<< 8, GU   , BU<< 8, RV<< 8, GV   , BV<< 8, RGB2YUV_SHIFT+8)
BGR2UV(uint32_t, rgb32ToUV, 0, 0,16, 0,   0x00FF, 0xFF00, 0xFF0000, RU<< 8, GU   , BU<< 8, RV<< 8, GV   , BV<< 8, RGB2YUV_SHIFT+8)
BGR2UV(uint32_t,rgb321ToUV, 0, 0,16, 8,   0x00FF, 0xFF00, 0xFF0000, RU<< 8, GU   , BU<< 8, RV<< 8, GV   , BV<< 8, RGB2YUV_SHIFT+8)
BGR2UV(uint16_t, bgr16ToUV, 0, 0, 0, 0,   0x001F, 0x07E0,   0xF800, RU<<11, GU<<5, BU    , RV<<11, GV<<5, BV    , RGB2YUV_SHIFT+8)
BGR2UV(uint16_t, bgr15ToUV, 0, 0, 0, 0,   0x001F, 0x03E0,   0x7C00, RU<<10, GU<<5, BU    , RV<<10, GV<<5, BV    , RGB2YUV_SHIFT+7)
BGR2UV(uint16_t, rgb16ToUV, 0, 0, 0, 0,   0xF800, 0x07E0,   0x001F, RU    , GU<<5, BU<<11, RV    , GV<<5, BV<<11, RGB2YUV_SHIFT+8)
BGR2UV(uint16_t, rgb15ToUV, 0, 0, 0, 0,   0x7C00, 0x03E0,   0x001F, RU    , GU<<5, BU<<10, RV    , GV<<5, BV<<10, RGB2YUV_SHIFT+7)

static inline void palToA(int16_t *dst, const uint8_t *src, int width, uint32_t *pal)
{
    int i;
    for (i=0; i<width; i++) {
        int d= src[i];

        dst[i]= (pal[d] >> 24)<<6;
    }
}

static inline void palToY(int16_t *dst, const uint8_t *src, long width, uint32_t *pal)
{
    int i;
    for (i=0; i<width; i++) {
        int d= src[i];

        dst[i]= (pal[d] & 0xFF)<<6;
    }
}

static inline void palToUV(uint16_t *dstU, int16_t *dstV,
                           const uint8_t *src1, const uint8_t *src2,
                           int width, uint32_t *pal)
{
    int i;
    assert(src1 == src2);
    for (i=0; i<width; i++) {
        int p= pal[src1[i]];

        dstU[i]= (uint8_t)(p>> 8)<<6;
        dstV[i]= (uint8_t)(p>>16)<<6;
    }
}

static inline void monowhite2Y(int16_t *dst, const uint8_t *src, int width, uint32_t *unused)
{
    int i, j;
    for (i=0; i<width/8; i++) {
        int d= ~src[i];
        for(j=0; j<8; j++)
            dst[8*i+j]= ((d>>(7-j))&1)*16383;
    }
}

static inline void monoblack2Y(int16_t *dst, const uint8_t *src, int width, uint32_t *unused)
{
    int i, j;
    for (i=0; i<width/8; i++) {
        int d= src[i];
        for(j=0; j<8; j++)
            dst[8*i+j]= ((d>>(7-j))&1)*16383;
    }
}

//Note: we have C, MMX, MMX2, 3DNOW versions, there is no 3DNOW+MMX2 one
//Plain C versions

#define COMPILE_TEMPLATE_MMX2 0
#define COMPILE_TEMPLATE_ALTIVEC 0

#include "swscale_template.c"

#if HAVE_ALTIVEC
#undef RENAME
#undef COMPILE_TEMPLATE_ALTIVEC
#define COMPILE_TEMPLATE_ALTIVEC 1
#define RENAME(a) a ## _altivec
#include "ppc/swscale_template.c"
#endif

//MMX versions
#if HAVE_MMX
#undef RENAME
#undef COMPILE_TEMPLATE_MMX2
#define COMPILE_TEMPLATE_MMX2 0
#define RENAME(a) a ## _MMX
#include "x86/swscale_template.c"
#endif

//MMX2 versions
#if HAVE_MMX2
#undef RENAME
#undef COMPILE_TEMPLATE_MMX2
#define COMPILE_TEMPLATE_MMX2 1
#define RENAME(a) a ## _MMX2
#include "x86/swscale_template.c"
#endif

SwsFunc ff_getSwsFunc(SwsContext *c)
{
    int cpu_flags = av_get_cpu_flags();

    sws_init_swScale_c(c);

#if HAVE_MMX
    if (cpu_flags & AV_CPU_FLAG_MMX)
        sws_init_swScale_MMX(c);
#endif
#if HAVE_MMX2
    if (cpu_flags & AV_CPU_FLAG_MMX2)
        sws_init_swScale_MMX2(c);
#endif
#if HAVE_ALTIVEC
    if (cpu_flags & AV_CPU_FLAG_ALTIVEC)
        sws_init_swScale_altivec(c);
#endif

    return swScale_c;
}

static void copyPlane(const uint8_t *src, int srcStride,
                      int srcSliceY, int srcSliceH, int width,
                      uint8_t *dst, int dstStride)
{
    dst += dstStride * srcSliceY;
    if (dstStride == srcStride && srcStride > 0) {
        memcpy(dst, src, srcSliceH * dstStride);
    } else {
        int i;
        for (i=0; i<srcSliceH; i++) {
            memcpy(dst, src, width);
            src += srcStride;
            dst += dstStride;
        }
    }
}

static int planarToNv12Wrapper(SwsContext *c, const uint8_t* src[], int srcStride[], int srcSliceY,
                               int srcSliceH, uint8_t* dstParam[], int dstStride[])
{
    uint8_t *dst = dstParam[1] + dstStride[1]*srcSliceY/2;

    copyPlane(src[0], srcStride[0], srcSliceY, srcSliceH, c->srcW,
              dstParam[0], dstStride[0]);

    if (c->dstFormat == PIX_FMT_NV12)
        interleaveBytes(src[1], src[2], dst, c->srcW/2, srcSliceH/2, srcStride[1], srcStride[2], dstStride[0]);
    else
        interleaveBytes(src[2], src[1], dst, c->srcW/2, srcSliceH/2, srcStride[2], srcStride[1], dstStride[0]);

    return srcSliceH;
}

static int planarToYuy2Wrapper(SwsContext *c, const uint8_t* src[], int srcStride[], int srcSliceY,
                               int srcSliceH, uint8_t* dstParam[], int dstStride[])
{
    uint8_t *dst=dstParam[0] + dstStride[0]*srcSliceY;

    yv12toyuy2(src[0], src[1], src[2], dst, c->srcW, srcSliceH, srcStride[0], srcStride[1], dstStride[0]);

    return srcSliceH;
}

static int planarToUyvyWrapper(SwsContext *c, const uint8_t* src[], int srcStride[], int srcSliceY,
                               int srcSliceH, uint8_t* dstParam[], int dstStride[])
{
    uint8_t *dst=dstParam[0] + dstStride[0]*srcSliceY;

    yv12touyvy(src[0], src[1], src[2], dst, c->srcW, srcSliceH, srcStride[0], srcStride[1], dstStride[0]);

    return srcSliceH;
}

static int yuv422pToYuy2Wrapper(SwsContext *c, const uint8_t* src[], int srcStride[], int srcSliceY,
                                int srcSliceH, uint8_t* dstParam[], int dstStride[])
{
    uint8_t *dst=dstParam[0] + dstStride[0]*srcSliceY;

    yuv422ptoyuy2(src[0],src[1],src[2],dst,c->srcW,srcSliceH,srcStride[0],srcStride[1],dstStride[0]);

    return srcSliceH;
}

static int yuv422pToUyvyWrapper(SwsContext *c, const uint8_t* src[], int srcStride[], int srcSliceY,
                                int srcSliceH, uint8_t* dstParam[], int dstStride[])
{
    uint8_t *dst=dstParam[0] + dstStride[0]*srcSliceY;

    yuv422ptouyvy(src[0],src[1],src[2],dst,c->srcW,srcSliceH,srcStride[0],srcStride[1],dstStride[0]);

    return srcSliceH;
}

static int yuyvToYuv420Wrapper(SwsContext *c, const uint8_t* src[], int srcStride[], int srcSliceY,
                               int srcSliceH, uint8_t* dstParam[], int dstStride[])
{
    uint8_t *ydst=dstParam[0] + dstStride[0]*srcSliceY;
    uint8_t *udst=dstParam[1] + dstStride[1]*srcSliceY/2;
    uint8_t *vdst=dstParam[2] + dstStride[2]*srcSliceY/2;

    yuyvtoyuv420(ydst, udst, vdst, src[0], c->srcW, srcSliceH, dstStride[0], dstStride[1], srcStride[0]);

    if (dstParam[3])
        fillPlane(dstParam[3], dstStride[3], c->srcW, srcSliceH, srcSliceY, 255);

    return srcSliceH;
}

static int yuyvToYuv422Wrapper(SwsContext *c, const uint8_t* src[], int srcStride[], int srcSliceY,
                               int srcSliceH, uint8_t* dstParam[], int dstStride[])
{
    uint8_t *ydst=dstParam[0] + dstStride[0]*srcSliceY;
    uint8_t *udst=dstParam[1] + dstStride[1]*srcSliceY;
    uint8_t *vdst=dstParam[2] + dstStride[2]*srcSliceY;

    yuyvtoyuv422(ydst, udst, vdst, src[0], c->srcW, srcSliceH, dstStride[0], dstStride[1], srcStride[0]);

    return srcSliceH;
}

static int uyvyToYuv420Wrapper(SwsContext *c, const uint8_t* src[], int srcStride[], int srcSliceY,
                               int srcSliceH, uint8_t* dstParam[], int dstStride[])
{
    uint8_t *ydst=dstParam[0] + dstStride[0]*srcSliceY;
    uint8_t *udst=dstParam[1] + dstStride[1]*srcSliceY/2;
    uint8_t *vdst=dstParam[2] + dstStride[2]*srcSliceY/2;

    uyvytoyuv420(ydst, udst, vdst, src[0], c->srcW, srcSliceH, dstStride[0], dstStride[1], srcStride[0]);

    if (dstParam[3])
        fillPlane(dstParam[3], dstStride[3], c->srcW, srcSliceH, srcSliceY, 255);

    return srcSliceH;
}

static int uyvyToYuv422Wrapper(SwsContext *c, const uint8_t* src[], int srcStride[], int srcSliceY,
                               int srcSliceH, uint8_t* dstParam[], int dstStride[])
{
    uint8_t *ydst=dstParam[0] + dstStride[0]*srcSliceY;
    uint8_t *udst=dstParam[1] + dstStride[1]*srcSliceY;
    uint8_t *vdst=dstParam[2] + dstStride[2]*srcSliceY;

    uyvytoyuv422(ydst, udst, vdst, src[0], c->srcW, srcSliceH, dstStride[0], dstStride[1], srcStride[0]);

    return srcSliceH;
}

static void gray8aToPacked32(const uint8_t *src, uint8_t *dst, int num_pixels, const uint8_t *palette)
{
    int i;
    for (i=0; i<num_pixels; i++)
        ((uint32_t *) dst)[i] = ((const uint32_t *)palette)[src[i<<1]] | (src[(i<<1)+1] << 24);
}

static void gray8aToPacked32_1(const uint8_t *src, uint8_t *dst, int num_pixels, const uint8_t *palette)
{
    int i;

    for (i=0; i<num_pixels; i++)
        ((uint32_t *) dst)[i] = ((const uint32_t *)palette)[src[i<<1]] | src[(i<<1)+1];
}

static void gray8aToPacked24(const uint8_t *src, uint8_t *dst, int num_pixels, const uint8_t *palette)
{
    int i;

    for (i=0; i<num_pixels; i++) {
        //FIXME slow?
        dst[0]= palette[src[i<<1]*4+0];
        dst[1]= palette[src[i<<1]*4+1];
        dst[2]= palette[src[i<<1]*4+2];
        dst+= 3;
    }
}

static int palToRgbWrapper(SwsContext *c, const uint8_t* src[], int srcStride[], int srcSliceY,
                           int srcSliceH, uint8_t* dst[], int dstStride[])
{
    const enum PixelFormat srcFormat= c->srcFormat;
    const enum PixelFormat dstFormat= c->dstFormat;
    void (*conv)(const uint8_t *src, uint8_t *dst, int num_pixels,
                 const uint8_t *palette)=NULL;
    int i;
    uint8_t *dstPtr= dst[0] + dstStride[0]*srcSliceY;
    const uint8_t *srcPtr= src[0];

    if (srcFormat == PIX_FMT_GRAY8A) {
        switch (dstFormat) {
        case PIX_FMT_RGB32  : conv = gray8aToPacked32; break;
        case PIX_FMT_BGR32  : conv = gray8aToPacked32; break;
        case PIX_FMT_BGR32_1: conv = gray8aToPacked32_1; break;
        case PIX_FMT_RGB32_1: conv = gray8aToPacked32_1; break;
        case PIX_FMT_RGB24  : conv = gray8aToPacked24; break;
        case PIX_FMT_BGR24  : conv = gray8aToPacked24; break;
        }
    } else if (usePal(srcFormat)) {
        switch (dstFormat) {
        case PIX_FMT_RGB32  : conv = sws_convertPalette8ToPacked32; break;
        case PIX_FMT_BGR32  : conv = sws_convertPalette8ToPacked32; break;
        case PIX_FMT_BGR32_1: conv = sws_convertPalette8ToPacked32; break;
        case PIX_FMT_RGB32_1: conv = sws_convertPalette8ToPacked32; break;
        case PIX_FMT_RGB24  : conv = sws_convertPalette8ToPacked24; break;
        case PIX_FMT_BGR24  : conv = sws_convertPalette8ToPacked24; break;
        }
    }

    if (!conv)
        av_log(c, AV_LOG_ERROR, "internal error %s -> %s converter\n",
               av_get_pix_fmt_name(srcFormat), av_get_pix_fmt_name(dstFormat));
    else {
        for (i=0; i<srcSliceH; i++) {
            conv(srcPtr, dstPtr, c->srcW, (uint8_t *) c->pal_rgb);
            srcPtr+= srcStride[0];
            dstPtr+= dstStride[0];
        }
    }

    return srcSliceH;
}

#define isRGBA32(x) (            \
           (x) == PIX_FMT_ARGB   \
        || (x) == PIX_FMT_RGBA   \
        || (x) == PIX_FMT_BGRA   \
        || (x) == PIX_FMT_ABGR   \
        )

/* {RGB,BGR}{15,16,24,32,32_1} -> {RGB,BGR}{15,16,24,32} */
static int rgbToRgbWrapper(SwsContext *c, const uint8_t* src[], int srcStride[], int srcSliceY,
                           int srcSliceH, uint8_t* dst[], int dstStride[])
{
    const enum PixelFormat srcFormat= c->srcFormat;
    const enum PixelFormat dstFormat= c->dstFormat;
    const int srcBpp= (c->srcFormatBpp + 7) >> 3;
    const int dstBpp= (c->dstFormatBpp + 7) >> 3;
    const int srcId= c->srcFormatBpp >> 2; /* 1:0, 4:1, 8:2, 15:3, 16:4, 24:6, 32:8 */
    const int dstId= c->dstFormatBpp >> 2;
    void (*conv)(const uint8_t *src, uint8_t *dst, int src_size)=NULL;

#define CONV_IS(src, dst) (srcFormat == PIX_FMT_##src && dstFormat == PIX_FMT_##dst)

    if (isRGBA32(srcFormat) && isRGBA32(dstFormat)) {
        if (     CONV_IS(ABGR, RGBA)
              || CONV_IS(ARGB, BGRA)
              || CONV_IS(BGRA, ARGB)
              || CONV_IS(RGBA, ABGR)) conv = shuffle_bytes_3210;
        else if (CONV_IS(ABGR, ARGB)
              || CONV_IS(ARGB, ABGR)) conv = shuffle_bytes_0321;
        else if (CONV_IS(ABGR, BGRA)
              || CONV_IS(ARGB, RGBA)) conv = shuffle_bytes_1230;
        else if (CONV_IS(BGRA, RGBA)
              || CONV_IS(RGBA, BGRA)) conv = shuffle_bytes_2103;
        else if (CONV_IS(BGRA, ABGR)
              || CONV_IS(RGBA, ARGB)) conv = shuffle_bytes_3012;
    } else
    /* BGR -> BGR */
    if (  (isBGRinInt(srcFormat) && isBGRinInt(dstFormat))
       || (isRGBinInt(srcFormat) && isRGBinInt(dstFormat))) {
        switch(srcId | (dstId<<4)) {
        case 0x34: conv= rgb16to15; break;
        case 0x36: conv= rgb24to15; break;
        case 0x38: conv= rgb32to15; break;
        case 0x43: conv= rgb15to16; break;
        case 0x46: conv= rgb24to16; break;
        case 0x48: conv= rgb32to16; break;
        case 0x63: conv= rgb15to24; break;
        case 0x64: conv= rgb16to24; break;
        case 0x68: conv= rgb32to24; break;
        case 0x83: conv= rgb15to32; break;
        case 0x84: conv= rgb16to32; break;
        case 0x86: conv= rgb24to32; break;
        }
    } else if (  (isBGRinInt(srcFormat) && isRGBinInt(dstFormat))
             || (isRGBinInt(srcFormat) && isBGRinInt(dstFormat))) {
        switch(srcId | (dstId<<4)) {
        case 0x33: conv= rgb15tobgr15; break;
        case 0x34: conv= rgb16tobgr15; break;
        case 0x36: conv= rgb24tobgr15; break;
        case 0x38: conv= rgb32tobgr15; break;
        case 0x43: conv= rgb15tobgr16; break;
        case 0x44: conv= rgb16tobgr16; break;
        case 0x46: conv= rgb24tobgr16; break;
        case 0x48: conv= rgb32tobgr16; break;
        case 0x63: conv= rgb15tobgr24; break;
        case 0x64: conv= rgb16tobgr24; break;
        case 0x66: conv= rgb24tobgr24; break;
        case 0x68: conv= rgb32tobgr24; break;
        case 0x83: conv= rgb15tobgr32; break;
        case 0x84: conv= rgb16tobgr32; break;
        case 0x86: conv= rgb24tobgr32; break;
        }
    }

    if (!conv) {
        av_log(c, AV_LOG_ERROR, "internal error %s -> %s converter\n",
               av_get_pix_fmt_name(srcFormat), av_get_pix_fmt_name(dstFormat));
    } else {
        const uint8_t *srcPtr= src[0];
              uint8_t *dstPtr= dst[0];
        if ((srcFormat == PIX_FMT_RGB32_1 || srcFormat == PIX_FMT_BGR32_1) && !isRGBA32(dstFormat))
            srcPtr += ALT32_CORR;

        if ((dstFormat == PIX_FMT_RGB32_1 || dstFormat == PIX_FMT_BGR32_1) && !isRGBA32(srcFormat))
            dstPtr += ALT32_CORR;

        if (dstStride[0]*srcBpp == srcStride[0]*dstBpp && srcStride[0] > 0 && !(srcStride[0]%srcBpp))
            conv(srcPtr, dstPtr + dstStride[0]*srcSliceY, srcSliceH*srcStride[0]);
        else {
            int i;
            dstPtr += dstStride[0]*srcSliceY;

            for (i=0; i<srcSliceH; i++) {
                conv(srcPtr, dstPtr, c->srcW*srcBpp);
                srcPtr+= srcStride[0];
                dstPtr+= dstStride[0];
            }
        }
    }
    return srcSliceH;
}

static int bgr24ToYv12Wrapper(SwsContext *c, const uint8_t* src[], int srcStride[], int srcSliceY,
                              int srcSliceH, uint8_t* dst[], int dstStride[])
{
    rgb24toyv12(
        src[0],
        dst[0]+ srcSliceY    *dstStride[0],
        dst[1]+(srcSliceY>>1)*dstStride[1],
        dst[2]+(srcSliceY>>1)*dstStride[2],
        c->srcW, srcSliceH,
        dstStride[0], dstStride[1], srcStride[0]);
    if (dst[3])
        fillPlane(dst[3], dstStride[3], c->srcW, srcSliceH, srcSliceY, 255);
    return srcSliceH;
}

static int yvu9ToYv12Wrapper(SwsContext *c, const uint8_t* src[], int srcStride[], int srcSliceY,
                             int srcSliceH, uint8_t* dst[], int dstStride[])
{
    copyPlane(src[0], srcStride[0], srcSliceY, srcSliceH, c->srcW,
              dst[0], dstStride[0]);

    planar2x(src[1], dst[1] + dstStride[1]*(srcSliceY >> 1), c->chrSrcW,
             srcSliceH >> 2, srcStride[1], dstStride[1]);
    planar2x(src[2], dst[2] + dstStride[2]*(srcSliceY >> 1), c->chrSrcW,
             srcSliceH >> 2, srcStride[2], dstStride[2]);
    if (dst[3])
        fillPlane(dst[3], dstStride[3], c->srcW, srcSliceH, srcSliceY, 255);
    return srcSliceH;
}

/* unscaled copy like stuff (assumes nearly identical formats) */
static int packedCopyWrapper(SwsContext *c, const uint8_t* src[], int srcStride[], int srcSliceY,
                             int srcSliceH, uint8_t* dst[], int dstStride[])
{
    if (dstStride[0]==srcStride[0] && srcStride[0] > 0)
        memcpy(dst[0] + dstStride[0]*srcSliceY, src[0], srcSliceH*dstStride[0]);
    else {
        int i;
        const uint8_t *srcPtr= src[0];
        uint8_t *dstPtr= dst[0] + dstStride[0]*srcSliceY;
        int length=0;

        /* universal length finder */
        while(length+c->srcW <= FFABS(dstStride[0])
           && length+c->srcW <= FFABS(srcStride[0])) length+= c->srcW;
        assert(length!=0);

        for (i=0; i<srcSliceH; i++) {
            memcpy(dstPtr, srcPtr, length);
            srcPtr+= srcStride[0];
            dstPtr+= dstStride[0];
        }
    }
    return srcSliceH;
}

#define DITHER_COPY(dst, dstStride, src, srcStride, bswap, dbswap)\
    uint16_t scale= dither_scale[dst_depth-1][src_depth-1];\
    int shift= src_depth-dst_depth + dither_scale[src_depth-2][dst_depth-1];\
    for (i = 0; i < height; i++) {\
        const uint8_t *dither= dithers[src_depth-9][i&7];\
        for (j = 0; j < length-7; j+=8){\
            dst[j+0] = dbswap((bswap(src[j+0]) + dither[0])*scale>>shift);\
            dst[j+1] = dbswap((bswap(src[j+1]) + dither[1])*scale>>shift);\
            dst[j+2] = dbswap((bswap(src[j+2]) + dither[2])*scale>>shift);\
            dst[j+3] = dbswap((bswap(src[j+3]) + dither[3])*scale>>shift);\
            dst[j+4] = dbswap((bswap(src[j+4]) + dither[4])*scale>>shift);\
            dst[j+5] = dbswap((bswap(src[j+5]) + dither[5])*scale>>shift);\
            dst[j+6] = dbswap((bswap(src[j+6]) + dither[6])*scale>>shift);\
            dst[j+7] = dbswap((bswap(src[j+7]) + dither[7])*scale>>shift);\
        }\
        for (; j < length; j++)\
            dst[j] = dbswap((bswap(src[j]) + dither[j&7])*scale>>shift);\
        dst += dstStride;\
        src += srcStride;\
    }


static int planarCopyWrapper(SwsContext *c, const uint8_t* src[], int srcStride[], int srcSliceY,
                             int srcSliceH, uint8_t* dst[], int dstStride[])
{
    int plane, i, j;
    for (plane=0; plane<4; plane++) {
        int length= (plane==0 || plane==3) ? c->srcW  : -((-c->srcW  )>>c->chrDstHSubSample);
        int y=      (plane==0 || plane==3) ? srcSliceY: -((-srcSliceY)>>c->chrDstVSubSample);
        int height= (plane==0 || plane==3) ? srcSliceH: -((-srcSliceH)>>c->chrDstVSubSample);
        const uint8_t *srcPtr= src[plane];
        uint8_t *dstPtr= dst[plane] + dstStride[plane]*y;

        if (!dst[plane]) continue;
        // ignore palette for GRAY8
        if (plane == 1 && !dst[2]) continue;
        if (!src[plane] || (plane == 1 && !src[2])) {
            if(is16BPS(c->dstFormat))
                length*=2;
            fillPlane(dst[plane], dstStride[plane], length, height, y, (plane==3) ? 255 : 128);
        } else {
            if(isNBPS(c->srcFormat) || isNBPS(c->dstFormat)
               || (is16BPS(c->srcFormat) != is16BPS(c->dstFormat))
            ) {
                const int src_depth = av_pix_fmt_descriptors[c->srcFormat].comp[plane].depth_minus1+1;
                const int dst_depth = av_pix_fmt_descriptors[c->dstFormat].comp[plane].depth_minus1+1;
                const uint16_t *srcPtr2 = (const uint16_t*)srcPtr;
                uint16_t *dstPtr2 = (uint16_t*)dstPtr;

                if (dst_depth == 8) {
                    if(isBE(c->srcFormat) == HAVE_BIGENDIAN){
                        DITHER_COPY(dstPtr, dstStride[plane], srcPtr2, srcStride[plane]/2, , )
                    } else {
                        DITHER_COPY(dstPtr, dstStride[plane], srcPtr2, srcStride[plane]/2, av_bswap16, )
                    }
                } else if (src_depth == 8) {
                    for (i = 0; i < height; i++) {
                        if(isBE(c->dstFormat)){
                            for (j = 0; j < length; j++)
                                AV_WB16(&dstPtr2[j], (srcPtr[j]<<(dst_depth-8)) |
                                                     (srcPtr[j]>>(2*8-dst_depth)));
                        } else {
                            for (j = 0; j < length; j++)
                                AV_WL16(&dstPtr2[j], (srcPtr[j]<<(dst_depth-8)) |
                                                     (srcPtr[j]>>(2*8-dst_depth)));
                        }
                        dstPtr2 += dstStride[plane]/2;
                        srcPtr  += srcStride[plane];
                    }
                } else if (src_depth <= dst_depth) {
                    for (i = 0; i < height; i++) {
#define COPY_UP(r,w) \
    for (j = 0; j < length; j++){ \
        unsigned int v= r(&srcPtr2[j]);\
        w(&dstPtr2[j], (v<<(dst_depth-src_depth)) | \
                       (v>>(2*src_depth-dst_depth)));\
    }
                        if(isBE(c->srcFormat)){
                            if(isBE(c->dstFormat)){
                                COPY_UP(AV_RB16, AV_WB16)
                            } else {
                                COPY_UP(AV_RB16, AV_WL16)
                            }
                        } else {
                            if(isBE(c->dstFormat)){
                                COPY_UP(AV_RL16, AV_WB16)
                            } else {
                                COPY_UP(AV_RL16, AV_WL16)
                            }
                        }
                        dstPtr2 += dstStride[plane]/2;
                        srcPtr2 += srcStride[plane]/2;
                    }
                } else {
                    if(isBE(c->srcFormat) == HAVE_BIGENDIAN){
                        if(isBE(c->dstFormat) == HAVE_BIGENDIAN){
                            DITHER_COPY(dstPtr2, dstStride[plane]/2, srcPtr2, srcStride[plane]/2, , )
                        } else {
                            DITHER_COPY(dstPtr2, dstStride[plane]/2, srcPtr2, srcStride[plane]/2, , av_bswap16)
                        }
                    }else{
                        if(isBE(c->dstFormat) == HAVE_BIGENDIAN){
                            DITHER_COPY(dstPtr2, dstStride[plane]/2, srcPtr2, srcStride[plane]/2, av_bswap16, )
                        } else {
                            DITHER_COPY(dstPtr2, dstStride[plane]/2, srcPtr2, srcStride[plane]/2, av_bswap16, av_bswap16)
                        }
                    }
                }
            } else if(is16BPS(c->srcFormat) && is16BPS(c->dstFormat)
                  && isBE(c->srcFormat) != isBE(c->dstFormat)) {

                for (i=0; i<height; i++) {
                    for (j=0; j<length; j++)
                        ((uint16_t*)dstPtr)[j] = av_bswap16(((const uint16_t*)srcPtr)[j]);
                    srcPtr+= srcStride[plane];
                    dstPtr+= dstStride[plane];
                }
            } else if (dstStride[plane] == srcStride[plane] &&
                       srcStride[plane] > 0 && srcStride[plane] == length) {
                memcpy(dst[plane] + dstStride[plane]*y, src[plane],
                       height*dstStride[plane]);
            } else {
                if(is16BPS(c->srcFormat) && is16BPS(c->dstFormat))
                    length*=2;
                for (i=0; i<height; i++) {
                    memcpy(dstPtr, srcPtr, length);
                    srcPtr+= srcStride[plane];
                    dstPtr+= dstStride[plane];
                }
            }
        }
    }
    return srcSliceH;
}

void ff_get_unscaled_swscale(SwsContext *c)
{
    const enum PixelFormat srcFormat = c->srcFormat;
    const enum PixelFormat dstFormat = c->dstFormat;
    const int flags = c->flags;
    const int dstH = c->dstH;
    int needsDither;

    needsDither= isAnyRGB(dstFormat)
        &&  c->dstFormatBpp < 24
        && (c->dstFormatBpp < c->srcFormatBpp || (!isAnyRGB(srcFormat)));

    /* yv12_to_nv12 */
    if ((srcFormat == PIX_FMT_YUV420P || srcFormat == PIX_FMT_YUVA420P) && (dstFormat == PIX_FMT_NV12 || dstFormat == PIX_FMT_NV21)) {
        c->swScale= planarToNv12Wrapper;
    }
    /* yuv2bgr */
    if ((srcFormat==PIX_FMT_YUV420P || srcFormat==PIX_FMT_YUV422P || srcFormat==PIX_FMT_YUVA420P) && isAnyRGB(dstFormat)
        && !(flags & SWS_ACCURATE_RND) && !(dstH&1)) {
        c->swScale= ff_yuv2rgb_get_func_ptr(c);
    }

    if (srcFormat==PIX_FMT_YUV410P && (dstFormat==PIX_FMT_YUV420P || dstFormat==PIX_FMT_YUVA420P) && !(flags & SWS_BITEXACT)) {
        c->swScale= yvu9ToYv12Wrapper;
    }

    /* bgr24toYV12 */
    if (srcFormat==PIX_FMT_BGR24 && (dstFormat==PIX_FMT_YUV420P || dstFormat==PIX_FMT_YUVA420P) && !(flags & SWS_ACCURATE_RND))
        c->swScale= bgr24ToYv12Wrapper;

    /* RGB/BGR -> RGB/BGR (no dither needed forms) */
    if (   isAnyRGB(srcFormat)
        && isAnyRGB(dstFormat)
        && srcFormat != PIX_FMT_BGR8      && dstFormat != PIX_FMT_BGR8
        && srcFormat != PIX_FMT_RGB8      && dstFormat != PIX_FMT_RGB8
        && srcFormat != PIX_FMT_BGR4      && dstFormat != PIX_FMT_BGR4
        && srcFormat != PIX_FMT_RGB4      && dstFormat != PIX_FMT_RGB4
        && srcFormat != PIX_FMT_BGR4_BYTE && dstFormat != PIX_FMT_BGR4_BYTE
        && srcFormat != PIX_FMT_RGB4_BYTE && dstFormat != PIX_FMT_RGB4_BYTE
        && srcFormat != PIX_FMT_MONOBLACK && dstFormat != PIX_FMT_MONOBLACK
        && srcFormat != PIX_FMT_MONOWHITE && dstFormat != PIX_FMT_MONOWHITE
        && srcFormat != PIX_FMT_RGB48LE   && dstFormat != PIX_FMT_RGB48LE
        && srcFormat != PIX_FMT_RGB48BE   && dstFormat != PIX_FMT_RGB48BE
        && srcFormat != PIX_FMT_BGR48LE   && dstFormat != PIX_FMT_BGR48LE
        && srcFormat != PIX_FMT_BGR48BE   && dstFormat != PIX_FMT_BGR48BE
        && (!needsDither || (c->flags&(SWS_FAST_BILINEAR|SWS_POINT))))
        c->swScale= rgbToRgbWrapper;

    if ((usePal(srcFormat) && (
        dstFormat == PIX_FMT_RGB32   ||
        dstFormat == PIX_FMT_RGB32_1 ||
        dstFormat == PIX_FMT_RGB24   ||
        dstFormat == PIX_FMT_BGR32   ||
        dstFormat == PIX_FMT_BGR32_1 ||
        dstFormat == PIX_FMT_BGR24)))
        c->swScale= palToRgbWrapper;

    if (srcFormat == PIX_FMT_YUV422P) {
        if (dstFormat == PIX_FMT_YUYV422)
            c->swScale= yuv422pToYuy2Wrapper;
        else if (dstFormat == PIX_FMT_UYVY422)
            c->swScale= yuv422pToUyvyWrapper;
    }

    /* LQ converters if -sws 0 or -sws 4*/
    if (c->flags&(SWS_FAST_BILINEAR|SWS_POINT)) {
        /* yv12_to_yuy2 */
        if (srcFormat == PIX_FMT_YUV420P || srcFormat == PIX_FMT_YUVA420P) {
            if (dstFormat == PIX_FMT_YUYV422)
                c->swScale= planarToYuy2Wrapper;
            else if (dstFormat == PIX_FMT_UYVY422)
                c->swScale= planarToUyvyWrapper;
        }
    }
    if(srcFormat == PIX_FMT_YUYV422 && (dstFormat == PIX_FMT_YUV420P || dstFormat == PIX_FMT_YUVA420P))
        c->swScale= yuyvToYuv420Wrapper;
    if(srcFormat == PIX_FMT_UYVY422 && (dstFormat == PIX_FMT_YUV420P || dstFormat == PIX_FMT_YUVA420P))
        c->swScale= uyvyToYuv420Wrapper;
    if(srcFormat == PIX_FMT_YUYV422 && dstFormat == PIX_FMT_YUV422P)
        c->swScale= yuyvToYuv422Wrapper;
    if(srcFormat == PIX_FMT_UYVY422 && dstFormat == PIX_FMT_YUV422P)
        c->swScale= uyvyToYuv422Wrapper;

#if HAVE_ALTIVEC
    if ((av_get_cpu_flags() & AV_CPU_FLAG_ALTIVEC) &&
        !(c->flags & SWS_BITEXACT) &&
        srcFormat == PIX_FMT_YUV420P) {
        // unscaled YV12 -> packed YUV, we want speed
        if (dstFormat == PIX_FMT_YUYV422)
            c->swScale= yv12toyuy2_unscaled_altivec;
        else if (dstFormat == PIX_FMT_UYVY422)
            c->swScale= yv12touyvy_unscaled_altivec;
    }
#endif

    /* simple copy */
    if (  srcFormat == dstFormat
        || (srcFormat == PIX_FMT_YUVA420P && dstFormat == PIX_FMT_YUV420P)
        || (srcFormat == PIX_FMT_YUV420P && dstFormat == PIX_FMT_YUVA420P)
        || (isPlanarYUV(srcFormat) && isGray(dstFormat))
        || (isPlanarYUV(dstFormat) && isGray(srcFormat))
        || (isGray(dstFormat) && isGray(srcFormat))
        || (isPlanarYUV(srcFormat) && isPlanarYUV(dstFormat)
            && c->chrDstHSubSample == c->chrSrcHSubSample
            && c->chrDstVSubSample == c->chrSrcVSubSample
            && dstFormat != PIX_FMT_NV12 && dstFormat != PIX_FMT_NV21
            && srcFormat != PIX_FMT_NV12 && srcFormat != PIX_FMT_NV21))
    {
        if (isPacked(c->srcFormat))
            c->swScale= packedCopyWrapper;
        else /* Planar YUV or gray */
            c->swScale= planarCopyWrapper;
    }
#if ARCH_BFIN
    ff_bfin_get_unscaled_swscale (c);
#endif
}

static void reset_ptr(const uint8_t* src[], int format)
{
    if(!isALPHA(format))
        src[3]=NULL;
    if(!isPlanarYUV(format)) {
        src[3]=src[2]=NULL;

        if (!usePal(format))
            src[1]= NULL;
    }
}

static int check_image_pointers(uint8_t *data[4], enum PixelFormat pix_fmt,
                                const int linesizes[4])
{
    const AVPixFmtDescriptor *desc = &av_pix_fmt_descriptors[pix_fmt];
    int i;

    for (i = 0; i < 4; i++) {
        int plane = desc->comp[i].plane;
        if (!data[plane] || !linesizes[plane])
            return 0;
    }

    return 1;
}

/**
 * swscale wrapper, so we don't need to export the SwsContext.
 * Assumes planar YUV to be in YUV order instead of YVU.
 */
int sws_scale(SwsContext *c, const uint8_t* const src[], const int srcStride[], int srcSliceY,
              int srcSliceH, uint8_t* const dst[], const int dstStride[])
{
    int i;
    const uint8_t* src2[4]= {src[0], src[1], src[2], src[3]};
    uint8_t* dst2[4]= {dst[0], dst[1], dst[2], dst[3]};

    // do not mess up sliceDir if we have a "trailing" 0-size slice
    if (srcSliceH == 0)
        return 0;

    if (!check_image_pointers(src, c->srcFormat, srcStride)) {
        av_log(c, AV_LOG_ERROR, "bad src image pointers\n");
        return 0;
    }
    if (!check_image_pointers(dst, c->dstFormat, dstStride)) {
        av_log(c, AV_LOG_ERROR, "bad dst image pointers\n");
        return 0;
    }

    if (c->sliceDir == 0 && srcSliceY != 0 && srcSliceY + srcSliceH != c->srcH) {
        av_log(c, AV_LOG_ERROR, "Slices start in the middle!\n");
        return 0;
    }
    if (c->sliceDir == 0) {
        if (srcSliceY == 0) c->sliceDir = 1; else c->sliceDir = -1;
    }

    if (usePal(c->srcFormat)) {
        for (i=0; i<256; i++) {
            int p, r, g, b, y, u, v, a = 0xff;
            if(c->srcFormat == PIX_FMT_PAL8) {
                p=((const uint32_t*)(src[1]))[i];
                a= (p>>24)&0xFF;
                r= (p>>16)&0xFF;
                g= (p>> 8)&0xFF;
                b=  p     &0xFF;
            } else if(c->srcFormat == PIX_FMT_RGB8) {
                r= (i>>5    )*36;
                g= ((i>>2)&7)*36;
                b= (i&3     )*85;
            } else if(c->srcFormat == PIX_FMT_BGR8) {
                b= (i>>6    )*85;
                g= ((i>>3)&7)*36;
                r= (i&7     )*36;
            } else if(c->srcFormat == PIX_FMT_RGB4_BYTE) {
                r= (i>>3    )*255;
                g= ((i>>1)&3)*85;
                b= (i&1     )*255;
            } else if(c->srcFormat == PIX_FMT_GRAY8 || c->srcFormat == PIX_FMT_GRAY8A) {
                r = g = b = i;
            } else {
                assert(c->srcFormat == PIX_FMT_BGR4_BYTE);
                b= (i>>3    )*255;
                g= ((i>>1)&3)*85;
                r= (i&1     )*255;
            }
            y= av_clip_uint8((RY*r + GY*g + BY*b + ( 33<<(RGB2YUV_SHIFT-1)))>>RGB2YUV_SHIFT);
            u= av_clip_uint8((RU*r + GU*g + BU*b + (257<<(RGB2YUV_SHIFT-1)))>>RGB2YUV_SHIFT);
            v= av_clip_uint8((RV*r + GV*g + BV*b + (257<<(RGB2YUV_SHIFT-1)))>>RGB2YUV_SHIFT);
            c->pal_yuv[i]= y + (u<<8) + (v<<16) + (a<<24);

            switch(c->dstFormat) {
            case PIX_FMT_BGR32:
#if !HAVE_BIGENDIAN
            case PIX_FMT_RGB24:
#endif
                c->pal_rgb[i]=  r + (g<<8) + (b<<16) + (a<<24);
                break;
            case PIX_FMT_BGR32_1:
#if HAVE_BIGENDIAN
            case PIX_FMT_BGR24:
#endif
                c->pal_rgb[i]= a + (r<<8) + (g<<16) + (b<<24);
                break;
            case PIX_FMT_RGB32_1:
#if HAVE_BIGENDIAN
            case PIX_FMT_RGB24:
#endif
                c->pal_rgb[i]= a + (b<<8) + (g<<16) + (r<<24);
                break;
            case PIX_FMT_RGB32:
#if !HAVE_BIGENDIAN
            case PIX_FMT_BGR24:
#endif
            default:
                c->pal_rgb[i]=  b + (g<<8) + (r<<16) + (a<<24);
            }
        }
    }

    // copy strides, so they can safely be modified
    if (c->sliceDir == 1) {
        // slices go from top to bottom
        int srcStride2[4]= {srcStride[0], srcStride[1], srcStride[2], srcStride[3]};
        int dstStride2[4]= {dstStride[0], dstStride[1], dstStride[2], dstStride[3]};

        reset_ptr(src2, c->srcFormat);
        reset_ptr((const uint8_t**)dst2, c->dstFormat);

        /* reset slice direction at end of frame */
        if (srcSliceY + srcSliceH == c->srcH)
            c->sliceDir = 0;

        return c->swScale(c, src2, srcStride2, srcSliceY, srcSliceH, dst2, dstStride2);
    } else {
        // slices go from bottom to top => we flip the image internally
        int srcStride2[4]= {-srcStride[0], -srcStride[1], -srcStride[2], -srcStride[3]};
        int dstStride2[4]= {-dstStride[0], -dstStride[1], -dstStride[2], -dstStride[3]};

        src2[0] += (srcSliceH-1)*srcStride[0];
        if (!usePal(c->srcFormat))
            src2[1] += ((srcSliceH>>c->chrSrcVSubSample)-1)*srcStride[1];
        src2[2] += ((srcSliceH>>c->chrSrcVSubSample)-1)*srcStride[2];
        src2[3] += (srcSliceH-1)*srcStride[3];
        dst2[0] += ( c->dstH                      -1)*dstStride[0];
        dst2[1] += ((c->dstH>>c->chrDstVSubSample)-1)*dstStride[1];
        dst2[2] += ((c->dstH>>c->chrDstVSubSample)-1)*dstStride[2];
        dst2[3] += ( c->dstH                      -1)*dstStride[3];

        reset_ptr(src2, c->srcFormat);
        reset_ptr((const uint8_t**)dst2, c->dstFormat);

        /* reset slice direction at end of frame */
        if (!srcSliceY)
            c->sliceDir = 0;

        return c->swScale(c, src2, srcStride2, c->srcH-srcSliceY-srcSliceH, srcSliceH, dst2, dstStride2);
    }
}

#if LIBSWSCALE_VERSION_MAJOR < 1
int sws_scale_ordered(SwsContext *c, const uint8_t* const src[], int srcStride[], int srcSliceY,
                      int srcSliceH, uint8_t* dst[], int dstStride[])
{
    return sws_scale(c, src, srcStride, srcSliceY, srcSliceH, dst, dstStride);
}
#endif

/* Convert the palette to the same packed 32-bit format as the palette */
void sws_convertPalette8ToPacked32(const uint8_t *src, uint8_t *dst, int num_pixels, const uint8_t *palette)
{
    int i;

    for (i=0; i<num_pixels; i++)
        ((uint32_t *) dst)[i] = ((const uint32_t *) palette)[src[i]];
}

/* Palette format: ABCD -> dst format: ABC */
void sws_convertPalette8ToPacked24(const uint8_t *src, uint8_t *dst, int num_pixels, const uint8_t *palette)
{
    int i;

    for (i=0; i<num_pixels; i++) {
        //FIXME slow?
        dst[0]= palette[src[i]*4+0];
        dst[1]= palette[src[i]*4+1];
        dst[2]= palette[src[i]*4+2];
        dst+= 3;
    }
}