/* * Copyright (C) 2001-2003 Michael Niedermayer * * 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 #include #include #include #include "config.h" #include #include "swscale.h" #include "swscale_internal.h" #include "rgb2rgb.h" #include "libavutil/avassert.h" #include "libavutil/intreadwrite.h" #include "libavutil/cpu.h" #include "libavutil/avutil.h" #include "libavutil/mathematics.h" #include "libavutil/bswap.h" #include "libavutil/pixdesc.h" #define RGB2YUV_SHIFT 15 #define BY ( (int)(0.114*219/255*(1<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 yuv2yuvX16_c_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) \ { \ yuv2yuvX16_c_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 yuv2yuvX16_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, enum PixelFormat dstFormat) { if (isNBPS(dstFormat)) { const int depth = av_pix_fmt_descriptors[dstFormat].comp[0].depth_minus1+1; yuv2yuvX16_c_template(lumFilter, lumSrc, lumFilterSize, chrFilter, chrUSrc, chrVSrc, chrFilterSize, alpSrc, dest, uDest, vDest, aDest, dstW, chrDstW, isBE(dstFormat), depth); } else { if (isBE(dstFormat)) { yuv2yuvX16_c_template(lumFilter, lumSrc, lumFilterSize, chrFilter, chrUSrc, chrVSrc, chrFilterSize, alpSrc, dest, uDest, vDest, aDest, dstW, chrDstW, 1, 16); } else { yuv2yuvX16_c_template(lumFilter, lumSrc, lumFilterSize, chrFilter, chrUSrc, chrVSrc, chrFilterSize, alpSrc, dest, uDest, vDest, aDest, dstW, chrDstW, 0, 16); } } } static void yuv2yuvX_c(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, 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>19); } if (uDest) for (i=0; i>19); vDest[i]= av_clip_uint8(v>>19); } if (CONFIG_SWSCALE_ALPHA && aDest) for (i=0; i>19); } } static inline void yuv2nv12X_c(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, 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>19); } if (!uDest) return; if (dstFormat == PIX_FMT_NV12) for (i=0; i>19); uDest[2*i+1]= av_clip_uint8(v>>19); } else for (i=0; i>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>=19;\ Y2>>=19;\ U >>=19;\ V >>=19;\ if (alpha) {\ A1 = 1<<18;\ A2 = 1<<18;\ for (j=0; j>=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>=10;\ U >>=10;\ V >>=10;\ if (alpha) {\ A = rnd;\ for (j=0; j>=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>=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>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>=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 void yuv2packedX_c(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 yuv2rgbX_c_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> (RGB2YUV_SHIFT-6+8);\ }\ }\ \ static inline void name ## ToUV_c(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_c(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 void name ## _c(int16_t *dst, const uint8_t *src, int width, uint32_t *unused)\ {\ int i;\ for (i=0; 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 void abgrToA_c(int16_t *dst, const uint8_t *src, int width, uint32_t *unused) { int i; for (i=0; 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 void name ## _half_c(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>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 void palToA_c(int16_t *dst, const uint8_t *src, int width, uint32_t *pal) { int i; for (i=0; i> 24)<<6; } } static void palToY_c(int16_t *dst, const uint8_t *src, long width, uint32_t *pal) { int i; for (i=0; i> 8)<<6; dstV[i]= (uint8_t)(p>>16)<<6; } } static void monowhite2Y_c(int16_t *dst, const uint8_t *src, int width, uint32_t *unused) { int i, j; for (i=0; i>(7-j))&1)*16383; } } static void monoblack2Y_c(int16_t *dst, const uint8_t *src, int width, uint32_t *unused) { int i, j; for (i=0; i>(7-j))&1)*16383; } } static void yuv2yuv1_c(SwsContext *c, const int16_t *lumSrc, const int16_t *chrUSrc, const int16_t *chrVSrc, 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) { int i; for (i=0; i>7; dest[i]= av_clip_uint8(val); } if (uDest) for (i=0; i>7; int v=(chrVSrc[i]+chrDither[(i+3)&7])>>7; uDest[i]= av_clip_uint8(u); vDest[i]= av_clip_uint8(v); } if (CONFIG_SWSCALE_ALPHA && aDest) for (i=0; i>7; aDest[i]= av_clip_uint8(val); } } /** * vertical bilinear scale YV12 to RGB */ static void yuv2packed2_c(SwsContext *c, const uint16_t *buf0, const uint16_t *buf1, const uint16_t *ubuf0, const uint16_t *ubuf1, const uint16_t *vbuf0, const uint16_t *vbuf1, const uint16_t *abuf0, const uint16_t *abuf1, uint8_t *dest, int dstW, int yalpha, int uvalpha, int y) { int yalpha1=4095- yalpha; int uvalpha1=4095-uvalpha; int i; YSCALE_YUV_2_ANYRGB_C(YSCALE_YUV_2_RGB2_C, YSCALE_YUV_2_PACKED2_C(void,0), YSCALE_YUV_2_GRAY16_2_C, YSCALE_YUV_2_MONO2_C) } /** * YV12 to RGB without scaling or interpolating */ static void yuv2packed1_c(SwsContext *c, const uint16_t *buf0, const uint16_t *ubuf0, const uint16_t *ubuf1, const uint16_t *vbuf0, const uint16_t *vbuf1, const uint16_t *abuf0, uint8_t *dest, int dstW, int uvalpha, enum PixelFormat dstFormat, int flags, int y) { const int yalpha1=0; int i; const uint16_t *buf1= buf0; //FIXME needed for RGB1/BGR1 const int yalpha= 4096; //FIXME ... if (uvalpha < 2048) { YSCALE_YUV_2_ANYRGB_C(YSCALE_YUV_2_RGB1_C, YSCALE_YUV_2_PACKED1_C(void,0), YSCALE_YUV_2_GRAY16_1_C, YSCALE_YUV_2_MONO2_C) } else { YSCALE_YUV_2_ANYRGB_C(YSCALE_YUV_2_RGB1B_C, YSCALE_YUV_2_PACKED1B_C(void,0), YSCALE_YUV_2_GRAY16_1_C, YSCALE_YUV_2_MONO2_C) } } //FIXME yuy2* can read up to 7 samples too much static void yuy2ToY_c(uint8_t *dst, const uint8_t *src, int width, uint32_t *unused) { int i; for (i=0; i>(depth-8); \ dstV[i] = rfunc(&srcV[i])>>(depth-8); \ } \ } \ \ static void endianness ## depth ## ToY_c(uint8_t *dstY, const uint16_t *srcY, int width, uint32_t *unused) \ { \ int i; \ for (i = 0; i < width; i++) \ dstY[i] = rfunc(&srcY[i])>>(depth-8); \ } \ YUV_NBPS( 9, LE, AV_RL16) YUV_NBPS( 9, BE, AV_RB16) YUV_NBPS(10, LE, AV_RL16) YUV_NBPS(10, BE, AV_RB16) #endif // YUV_NBPS static inline void nv12ToUV_c(uint8_t *dstU, uint8_t *dstV, const uint8_t *src1, const uint8_t *src2, int width, uint32_t *unused) { nvXXtoUV_c(dstU, dstV, src1, width); } static inline void nv21ToUV_c(uint8_t *dstU, uint8_t *dstV, const uint8_t *src1, const uint8_t *src2, int width, uint32_t *unused) { nvXXtoUV_c(dstV, dstU, src1, width); } static void bgr24ToY_c(int16_t *dst, const uint8_t *src, int width, uint32_t *unused) { int i; for (i=0; i>(RGB2YUV_SHIFT-6)); } } static void bgr24ToUV_c(int16_t *dstU, int16_t *dstV, const uint8_t *src1, const uint8_t *src2, int width, uint32_t *unused) { int i; for (i=0; i>(RGB2YUV_SHIFT-6); dstV[i]= (RV*r + GV*g + BV*b + (256<<(RGB2YUV_SHIFT-1)) + (1<<(RGB2YUV_SHIFT-7)))>>(RGB2YUV_SHIFT-6); } assert(src1 == src2); } static void bgr24ToUV_half_c(int16_t *dstU, int16_t *dstV, const uint8_t *src1, const uint8_t *src2, int width, uint32_t *unused) { int i; for (i=0; i>(RGB2YUV_SHIFT-5); dstV[i]= (RV*r + GV*g + BV*b + (256<>(RGB2YUV_SHIFT-5); } assert(src1 == src2); } static void rgb24ToY_c(int16_t *dst, const uint8_t *src, int width, uint32_t *unused) { int i; for (i=0; i>(RGB2YUV_SHIFT-6)); } } static void rgb24ToUV_c(int16_t *dstU, int16_t *dstV, const uint8_t *src1, const uint8_t *src2, int width, uint32_t *unused) { int i; assert(src1==src2); for (i=0; i>(RGB2YUV_SHIFT-6); dstV[i]= (RV*r + GV*g + BV*b + (256<<(RGB2YUV_SHIFT-1)) + (1<<(RGB2YUV_SHIFT-7)))>>(RGB2YUV_SHIFT-6); } } static void rgb24ToUV_half_c(int16_t *dstU, int16_t *dstV, const uint8_t *src1, const uint8_t *src2, int width, uint32_t *unused) { int i; assert(src1==src2); for (i=0; i>(RGB2YUV_SHIFT-5); dstV[i]= (RV*r + GV*g + BV*b + (256<>(RGB2YUV_SHIFT-5); } } // bilinear / bicubic scaling static void hScale_c(int16_t *dst, int dstW, const uint8_t *src, int srcW, int xInc, const int16_t *filter, const int16_t *filterPos, int filterSize) { int i; for (i=0; i>7, (1<<15)-1); // the cubic equation does overflow ... //dst[i] = val>>7; } } static inline void hScale16_c(int16_t *dst, int dstW, const uint16_t *src, int srcW, int xInc, const int16_t *filter, const int16_t *filterPos, long filterSize, int shift) { int i, j; for (i=0; i>shift, (1<<15)-1); // the cubic equation does overflow ... } } static inline void hScale16X_c(int16_t *dst, int dstW, const uint16_t *src, int srcW, int xInc, const int16_t *filter, const int16_t *filterPos, long filterSize, int shift) { int i, j; for (i=0; i>shift, (1<<15)-1); // the cubic equation does overflow ... } } //FIXME all pal and rgb srcFormats could do this convertion as well //FIXME all scalers more complex than bilinear could do half of this transform static void chrRangeToJpeg_c(int16_t *dstU, int16_t *dstV, int width) { int i; for (i = 0; i < width; i++) { dstU[i] = (FFMIN(dstU[i],30775)*4663 - 9289992)>>12; //-264 dstV[i] = (FFMIN(dstV[i],30775)*4663 - 9289992)>>12; //-264 } } static void chrRangeFromJpeg_c(int16_t *dstU, int16_t *dstV, int width) { int i; for (i = 0; i < width; i++) { dstU[i] = (dstU[i]*1799 + 4081085)>>11; //1469 dstV[i] = (dstV[i]*1799 + 4081085)>>11; //1469 } } static void lumRangeToJpeg_c(int16_t *dst, int width) { int i; for (i = 0; i < width; i++) dst[i] = (FFMIN(dst[i],30189)*19077 - 39057361)>>14; } static void lumRangeFromJpeg_c(int16_t *dst, int width) { int i; for (i = 0; i < width; i++) dst[i] = (dst[i]*14071 + 33561947)>>14; } static void hyscale_fast_c(SwsContext *c, int16_t *dst, int dstWidth, const uint8_t *src, int srcW, int xInc) { int i; unsigned int xpos=0; for (i=0;i>16; register unsigned int xalpha=(xpos&0xFFFF)>>9; dst[i]= (src[xx]<<7) + (src[xx+1] - src[xx])*xalpha; xpos+=xInc; } for (i=dstWidth-1; (i*xInc)>>16 >=srcW-1; i--) dst[i] = src[srcW-1]*128; } // *** horizontal scale Y line to temp buffer static inline void hyscale(SwsContext *c, uint16_t *dst, int dstWidth, const uint8_t *src, int srcW, int xInc, const int16_t *hLumFilter, const int16_t *hLumFilterPos, int hLumFilterSize, uint8_t *formatConvBuffer, uint32_t *pal, int isAlpha) { void (*toYV12)(uint8_t *, const uint8_t *, int, uint32_t *) = isAlpha ? c->alpToYV12 : c->lumToYV12; void (*convertRange)(int16_t *, int) = isAlpha ? NULL : c->lumConvertRange; src += isAlpha ? c->alpSrcOffset : c->lumSrcOffset; if (toYV12) { toYV12(formatConvBuffer, src, srcW, pal); src= formatConvBuffer; } if (c->hScale16) { int shift= isAnyRGB(c->srcFormat) || c->srcFormat==PIX_FMT_PAL8 ? 13 : av_pix_fmt_descriptors[c->srcFormat].comp[0].depth_minus1; c->hScale16(dst, dstWidth, (const uint16_t*)src, srcW, xInc, hLumFilter, hLumFilterPos, hLumFilterSize, shift); } else if (!c->hyscale_fast) { c->hScale(dst, dstWidth, src, srcW, xInc, hLumFilter, hLumFilterPos, hLumFilterSize); } else { // fast bilinear upscale / crap downscale c->hyscale_fast(c, dst, dstWidth, src, srcW, xInc); } if (convertRange) convertRange(dst, dstWidth); } static void hcscale_fast_c(SwsContext *c, int16_t *dst1, int16_t *dst2, int dstWidth, const uint8_t *src1, const uint8_t *src2, int srcW, int xInc) { int i; unsigned int xpos=0; for (i=0;i>16; register unsigned int xalpha=(xpos&0xFFFF)>>9; dst1[i]=(src1[xx]*(xalpha^127)+src1[xx+1]*xalpha); dst2[i]=(src2[xx]*(xalpha^127)+src2[xx+1]*xalpha); xpos+=xInc; } for (i=dstWidth-1; (i*xInc)>>16 >=srcW-1; i--) { dst1[i] = src1[srcW-1]*128; dst2[i] = src2[srcW-1]*128; } } static inline void hcscale(SwsContext *c, uint16_t *dst1, uint16_t *dst2, int dstWidth, const uint8_t *src1, const uint8_t *src2, int srcW, int xInc, const int16_t *hChrFilter, const int16_t *hChrFilterPos, int hChrFilterSize, uint8_t *formatConvBuffer, uint32_t *pal) { src1 += c->chrSrcOffset; src2 += c->chrSrcOffset; if (c->chrToYV12) { uint8_t *buf2 = formatConvBuffer + FFALIGN(srcW*2+78, 16); c->chrToYV12(formatConvBuffer, buf2, src1, src2, srcW, pal); src1= formatConvBuffer; src2= buf2; } if (c->hScale16) { int shift= isAnyRGB(c->srcFormat) || c->srcFormat==PIX_FMT_PAL8 ? 13 : av_pix_fmt_descriptors[c->srcFormat].comp[0].depth_minus1; c->hScale16(dst1, dstWidth, (const uint16_t*)src1, srcW, xInc, hChrFilter, hChrFilterPos, hChrFilterSize, shift); c->hScale16(dst2, dstWidth, (const uint16_t*)src2, srcW, xInc, hChrFilter, hChrFilterPos, hChrFilterSize, shift); } else if (!c->hcscale_fast) { c->hScale(dst1, dstWidth, src1, srcW, xInc, hChrFilter, hChrFilterPos, hChrFilterSize); c->hScale(dst2, dstWidth, src2, srcW, xInc, hChrFilter, hChrFilterPos, hChrFilterSize); } else { // fast bilinear upscale / crap downscale c->hcscale_fast(c, dst1, dst2, dstWidth, src1, src2, srcW, xInc); } if (c->chrConvertRange) c->chrConvertRange(dst1, dst2, dstWidth); } #define DEBUG_SWSCALE_BUFFERS 0 #define DEBUG_BUFFERS(...) if (DEBUG_SWSCALE_BUFFERS) av_log(c, AV_LOG_DEBUG, __VA_ARGS__) static int swScale(SwsContext *c, const uint8_t* src[], int srcStride[], int srcSliceY, int srcSliceH, uint8_t* dst[], int dstStride[]) { /* load a few things into local vars to make the code more readable? and faster */ const int srcW= c->srcW; const int dstW= c->dstW; const int dstH= c->dstH; const int chrDstW= c->chrDstW; const int chrSrcW= c->chrSrcW; const int lumXInc= c->lumXInc; const int chrXInc= c->chrXInc; const enum PixelFormat dstFormat= c->dstFormat; const int flags= c->flags; int16_t *vLumFilterPos= c->vLumFilterPos; int16_t *vChrFilterPos= c->vChrFilterPos; int16_t *hLumFilterPos= c->hLumFilterPos; int16_t *hChrFilterPos= c->hChrFilterPos; int16_t *vLumFilter= c->vLumFilter; int16_t *vChrFilter= c->vChrFilter; int16_t *hLumFilter= c->hLumFilter; int16_t *hChrFilter= c->hChrFilter; int32_t *lumMmxFilter= c->lumMmxFilter; int32_t *chrMmxFilter= c->chrMmxFilter; int32_t av_unused *alpMmxFilter= c->alpMmxFilter; const int vLumFilterSize= c->vLumFilterSize; const int vChrFilterSize= c->vChrFilterSize; const int hLumFilterSize= c->hLumFilterSize; const int hChrFilterSize= c->hChrFilterSize; int16_t **lumPixBuf= c->lumPixBuf; int16_t **chrUPixBuf= c->chrUPixBuf; int16_t **chrVPixBuf= c->chrVPixBuf; int16_t **alpPixBuf= c->alpPixBuf; const int vLumBufSize= c->vLumBufSize; const int vChrBufSize= c->vChrBufSize; uint8_t *formatConvBuffer= c->formatConvBuffer; const int chrSrcSliceY= srcSliceY >> c->chrSrcVSubSample; const int chrSrcSliceH= -((-srcSliceH) >> c->chrSrcVSubSample); int lastDstY; uint32_t *pal=c->pal_yuv; int should_dither= isNBPS(c->srcFormat) || is16BPS(c->srcFormat); /* vars which will change and which we need to store back in the context */ int dstY= c->dstY; int lumBufIndex= c->lumBufIndex; int chrBufIndex= c->chrBufIndex; int lastInLumBuf= c->lastInLumBuf; int lastInChrBuf= c->lastInChrBuf; if (isPacked(c->srcFormat)) { src[0]= src[1]= src[2]= src[3]= src[0]; srcStride[0]= srcStride[1]= srcStride[2]= srcStride[3]= srcStride[0]; } srcStride[1]<<= c->vChrDrop; srcStride[2]<<= c->vChrDrop; DEBUG_BUFFERS("swScale() %p[%d] %p[%d] %p[%d] %p[%d] -> %p[%d] %p[%d] %p[%d] %p[%d]\n", src[0], srcStride[0], src[1], srcStride[1], src[2], srcStride[2], src[3], srcStride[3], dst[0], dstStride[0], dst[1], dstStride[1], dst[2], dstStride[2], dst[3], dstStride[3]); DEBUG_BUFFERS("srcSliceY: %d srcSliceH: %d dstY: %d dstH: %d\n", srcSliceY, srcSliceH, dstY, dstH); DEBUG_BUFFERS("vLumFilterSize: %d vLumBufSize: %d vChrFilterSize: %d vChrBufSize: %d\n", vLumFilterSize, vLumBufSize, vChrFilterSize, vChrBufSize); if (dstStride[0]%8 !=0 || dstStride[1]%8 !=0 || dstStride[2]%8 !=0 || dstStride[3]%8 != 0) { static int warnedAlready=0; //FIXME move this into the context perhaps if (flags & SWS_PRINT_INFO && !warnedAlready) { av_log(c, AV_LOG_WARNING, "Warning: dstStride is not aligned!\n" " ->cannot do aligned memory accesses anymore\n"); warnedAlready=1; } } /* Note the user might start scaling the picture in the middle so this will not get executed. This is not really intended but works currently, so people might do it. */ if (srcSliceY ==0) { lumBufIndex=-1; chrBufIndex=-1; dstY=0; lastInLumBuf= -1; lastInChrBuf= -1; } lastDstY= dstY; for (;dstY < dstH; dstY++) { unsigned char *dest =dst[0]+dstStride[0]*dstY; const int chrDstY= dstY>>c->chrDstVSubSample; unsigned char *uDest=dst[1]+dstStride[1]*chrDstY; unsigned char *vDest=dst[2]+dstStride[2]*chrDstY; unsigned char *aDest=(CONFIG_SWSCALE_ALPHA && alpPixBuf) ? dst[3]+dstStride[3]*dstY : NULL; const uint8_t *lumDither= should_dither ? dithers[7][dstY &7] : flat64; const uint8_t *chrDither= should_dither ? dithers[7][chrDstY&7] : flat64; const int firstLumSrcY= vLumFilterPos[dstY]; //First line needed as input const int firstLumSrcY2= vLumFilterPos[FFMIN(dstY | ((1<chrDstVSubSample) - 1), dstH-1)]; const int firstChrSrcY= vChrFilterPos[chrDstY]; //First line needed as input int lastLumSrcY= firstLumSrcY + vLumFilterSize -1; // Last line needed as input int lastLumSrcY2=firstLumSrcY2+ vLumFilterSize -1; // Last line needed as input int lastChrSrcY= firstChrSrcY + vChrFilterSize -1; // Last line needed as input int enough_lines; //handle holes (FAST_BILINEAR & weird filters) if (firstLumSrcY > lastInLumBuf) lastInLumBuf= firstLumSrcY-1; if (firstChrSrcY > lastInChrBuf) lastInChrBuf= firstChrSrcY-1; assert(firstLumSrcY >= lastInLumBuf - vLumBufSize + 1); assert(firstChrSrcY >= lastInChrBuf - vChrBufSize + 1); DEBUG_BUFFERS("dstY: %d\n", dstY); DEBUG_BUFFERS("\tfirstLumSrcY: %d lastLumSrcY: %d lastInLumBuf: %d\n", firstLumSrcY, lastLumSrcY, lastInLumBuf); DEBUG_BUFFERS("\tfirstChrSrcY: %d lastChrSrcY: %d lastInChrBuf: %d\n", firstChrSrcY, lastChrSrcY, lastInChrBuf); // Do we have enough lines in this slice to output the dstY line enough_lines = lastLumSrcY2 < srcSliceY + srcSliceH && lastChrSrcY < -((-srcSliceY - srcSliceH)>>c->chrSrcVSubSample); if (!enough_lines) { lastLumSrcY = srcSliceY + srcSliceH - 1; lastChrSrcY = chrSrcSliceY + chrSrcSliceH - 1; DEBUG_BUFFERS("buffering slice: lastLumSrcY %d lastChrSrcY %d\n", lastLumSrcY, lastChrSrcY); } //Do horizontal scaling while(lastInLumBuf < lastLumSrcY) { const uint8_t *src1= src[0]+(lastInLumBuf + 1 - srcSliceY)*srcStride[0]; const uint8_t *src2= src[3]+(lastInLumBuf + 1 - srcSliceY)*srcStride[3]; lumBufIndex++; assert(lumBufIndex < 2*vLumBufSize); assert(lastInLumBuf + 1 - srcSliceY < srcSliceH); assert(lastInLumBuf + 1 - srcSliceY >= 0); hyscale(c, lumPixBuf[ lumBufIndex ], dstW, src1, srcW, lumXInc, hLumFilter, hLumFilterPos, hLumFilterSize, formatConvBuffer, pal, 0); if (CONFIG_SWSCALE_ALPHA && alpPixBuf) hyscale(c, alpPixBuf[ lumBufIndex ], dstW, src2, srcW, lumXInc, hLumFilter, hLumFilterPos, hLumFilterSize, formatConvBuffer, pal, 1); lastInLumBuf++; DEBUG_BUFFERS("\t\tlumBufIndex %d: lastInLumBuf: %d\n", lumBufIndex, lastInLumBuf); } while(lastInChrBuf < lastChrSrcY) { const uint8_t *src1= src[1]+(lastInChrBuf + 1 - chrSrcSliceY)*srcStride[1]; const uint8_t *src2= src[2]+(lastInChrBuf + 1 - chrSrcSliceY)*srcStride[2]; chrBufIndex++; assert(chrBufIndex < 2*vChrBufSize); assert(lastInChrBuf + 1 - chrSrcSliceY < (chrSrcSliceH)); assert(lastInChrBuf + 1 - chrSrcSliceY >= 0); //FIXME replace parameters through context struct (some at least) if (c->needs_hcscale) hcscale(c, chrUPixBuf[chrBufIndex], chrVPixBuf[chrBufIndex], chrDstW, src1, src2, chrSrcW, chrXInc, hChrFilter, hChrFilterPos, hChrFilterSize, formatConvBuffer, pal); lastInChrBuf++; DEBUG_BUFFERS("\t\tchrBufIndex %d: lastInChrBuf: %d\n", chrBufIndex, lastInChrBuf); } //wrap buf index around to stay inside the ring buffer if (lumBufIndex >= vLumBufSize) lumBufIndex-= vLumBufSize; if (chrBufIndex >= vChrBufSize) chrBufIndex-= vChrBufSize; if (!enough_lines) break; //we can't output a dstY line so let's try with the next slice #if HAVE_MMX updateMMXDitherTables(c, dstY, lumBufIndex, chrBufIndex, lastInLumBuf, lastInChrBuf); #endif if (dstY < dstH-2) { const int16_t **lumSrcPtr= (const int16_t **) lumPixBuf + lumBufIndex + firstLumSrcY - lastInLumBuf + vLumBufSize; const int16_t **chrUSrcPtr= (const int16_t **) chrUPixBuf + chrBufIndex + firstChrSrcY - lastInChrBuf + vChrBufSize; const int16_t **chrVSrcPtr= (const int16_t **) chrVPixBuf + chrBufIndex + firstChrSrcY - lastInChrBuf + vChrBufSize; const int16_t **alpSrcPtr= (CONFIG_SWSCALE_ALPHA && alpPixBuf) ? (const int16_t **) alpPixBuf + lumBufIndex + firstLumSrcY - lastInLumBuf + vLumBufSize : NULL; if (dstFormat == PIX_FMT_NV12 || dstFormat == PIX_FMT_NV21) { const int chrSkipMask= (1<chrDstVSubSample)-1; if (dstY&chrSkipMask) uDest= NULL; //FIXME split functions in lumi / chromi c->yuv2nv12X(c, vLumFilter+dstY*vLumFilterSize , lumSrcPtr, vLumFilterSize, vChrFilter+chrDstY*vChrFilterSize, chrUSrcPtr, chrVSrcPtr, vChrFilterSize, dest, uDest, dstW, chrDstW, dstFormat, lumDither, chrDither); } else if (isPlanarYUV(dstFormat) || dstFormat==PIX_FMT_GRAY8) { //YV12 like const int chrSkipMask= (1<chrDstVSubSample)-1; if ((dstY&chrSkipMask) || isGray(dstFormat)) uDest=vDest= NULL; //FIXME split functions in lumi / chromi if (is16BPS(dstFormat) || isNBPS(dstFormat)) { yuv2yuvX16_c(vLumFilter+dstY*vLumFilterSize , lumSrcPtr, vLumFilterSize, vChrFilter+chrDstY*vChrFilterSize, chrUSrcPtr, chrVSrcPtr, vChrFilterSize, alpSrcPtr, (uint16_t *) dest, (uint16_t *) uDest, (uint16_t *) vDest, (uint16_t *) aDest, dstW, chrDstW, dstFormat); } else if (vLumFilterSize == 1 && vChrFilterSize == 1) { // unscaled YV12 const int16_t *lumBuf = lumSrcPtr[0]; const int16_t *chrUBuf= chrUSrcPtr[0]; const int16_t *chrVBuf= chrVSrcPtr[0]; const int16_t *alpBuf= (CONFIG_SWSCALE_ALPHA && alpPixBuf) ? alpSrcPtr[0] : NULL; c->yuv2yuv1(c, lumBuf, chrUBuf, chrVBuf, alpBuf, dest, uDest, vDest, aDest, dstW, chrDstW, lumDither, chrDither); } else { //General YV12 c->yuv2yuvX(c, vLumFilter+dstY*vLumFilterSize , lumSrcPtr, vLumFilterSize, vChrFilter+chrDstY*vChrFilterSize, chrUSrcPtr, chrVSrcPtr, vChrFilterSize, alpSrcPtr, dest, uDest, vDest, aDest, dstW, chrDstW, lumDither, chrDither); } } else { assert(lumSrcPtr + vLumFilterSize - 1 < lumPixBuf + vLumBufSize*2); assert(chrUSrcPtr + vChrFilterSize - 1 < chrUPixBuf + vChrBufSize*2); if (vLumFilterSize == 1 && vChrFilterSize == 2) { //unscaled RGB int chrAlpha= vChrFilter[2*dstY+1]; if(flags & SWS_FULL_CHR_H_INT) { yuv2rgbX_c_full(c, //FIXME write a packed1_full function vLumFilter+dstY*vLumFilterSize, lumSrcPtr, vLumFilterSize, vChrFilter+dstY*vChrFilterSize, chrUSrcPtr, chrVSrcPtr, vChrFilterSize, alpSrcPtr, dest, dstW, dstY); } else { c->yuv2packed1(c, *lumSrcPtr, *chrUSrcPtr, *(chrUSrcPtr+1), *chrVSrcPtr, *(chrVSrcPtr+1), alpPixBuf ? *alpSrcPtr : NULL, dest, dstW, chrAlpha, dstFormat, flags, dstY); } } else if (vLumFilterSize == 2 && vChrFilterSize == 2) { //bilinear upscale RGB int lumAlpha= vLumFilter[2*dstY+1]; int chrAlpha= vChrFilter[2*dstY+1]; lumMmxFilter[2]= lumMmxFilter[3]= vLumFilter[2*dstY ]*0x10001; chrMmxFilter[2]= chrMmxFilter[3]= vChrFilter[2*chrDstY]*0x10001; if(flags & SWS_FULL_CHR_H_INT) { yuv2rgbX_c_full(c, //FIXME write a packed2_full function vLumFilter+dstY*vLumFilterSize, lumSrcPtr, vLumFilterSize, vChrFilter+dstY*vChrFilterSize, chrUSrcPtr, chrVSrcPtr, vChrFilterSize, alpSrcPtr, dest, dstW, dstY); } else { c->yuv2packed2(c, *lumSrcPtr, *(lumSrcPtr+1), *chrUSrcPtr, *(chrUSrcPtr+1), *chrVSrcPtr, *(chrVSrcPtr+1), alpPixBuf ? *alpSrcPtr : NULL, alpPixBuf ? *(alpSrcPtr+1) : NULL, dest, dstW, lumAlpha, chrAlpha, dstY); } } else { //general RGB if(flags & SWS_FULL_CHR_H_INT) { yuv2rgbX_c_full(c, vLumFilter+dstY*vLumFilterSize, lumSrcPtr, vLumFilterSize, vChrFilter+dstY*vChrFilterSize, chrUSrcPtr, chrVSrcPtr, vChrFilterSize, alpSrcPtr, dest, dstW, dstY); } else { c->yuv2packedX(c, vLumFilter+dstY*vLumFilterSize, lumSrcPtr, vLumFilterSize, vChrFilter+dstY*vChrFilterSize, chrUSrcPtr, chrVSrcPtr, vChrFilterSize, alpSrcPtr, dest, dstW, dstY); } } } } else { // hmm looks like we can't use MMX here without overwriting this array's tail const int16_t **lumSrcPtr= (const int16_t **)lumPixBuf + lumBufIndex + firstLumSrcY - lastInLumBuf + vLumBufSize; const int16_t **chrUSrcPtr= (const int16_t **)chrUPixBuf + chrBufIndex + firstChrSrcY - lastInChrBuf + vChrBufSize; const int16_t **chrVSrcPtr= (const int16_t **)chrVPixBuf + chrBufIndex + firstChrSrcY - lastInChrBuf + vChrBufSize; const int16_t **alpSrcPtr= (CONFIG_SWSCALE_ALPHA && alpPixBuf) ? (const int16_t **)alpPixBuf + lumBufIndex + firstLumSrcY - lastInLumBuf + vLumBufSize : NULL; if (dstFormat == PIX_FMT_NV12 || dstFormat == PIX_FMT_NV21) { const int chrSkipMask= (1<chrDstVSubSample)-1; if (dstY&chrSkipMask) uDest= NULL; //FIXME split functions in lumi / chromi yuv2nv12X_c(c, vLumFilter+dstY*vLumFilterSize, lumSrcPtr, vLumFilterSize, vChrFilter+chrDstY*vChrFilterSize, chrUSrcPtr, chrVSrcPtr, vChrFilterSize, dest, uDest, dstW, chrDstW, dstFormat, lumDither, chrDither); } else if (isPlanarYUV(dstFormat) || dstFormat==PIX_FMT_GRAY8) { //YV12 const int chrSkipMask= (1<chrDstVSubSample)-1; if ((dstY&chrSkipMask) || isGray(dstFormat)) uDest=vDest= NULL; //FIXME split functions in lumi / chromi if (is16BPS(dstFormat) || is9_OR_10BPS(dstFormat)) { yuv2yuvX16_c(vLumFilter+dstY*vLumFilterSize , lumSrcPtr, vLumFilterSize, vChrFilter+chrDstY*vChrFilterSize, chrUSrcPtr, chrVSrcPtr, vChrFilterSize, alpSrcPtr, (uint16_t *) dest, (uint16_t *) uDest, (uint16_t *) vDest, (uint16_t *) aDest, dstW, chrDstW, dstFormat); } else { yuv2yuvX_c(c, vLumFilter+dstY*vLumFilterSize, lumSrcPtr, vLumFilterSize, vChrFilter+chrDstY*vChrFilterSize, chrUSrcPtr, chrVSrcPtr, vChrFilterSize, alpSrcPtr, dest, uDest, vDest, aDest, dstW, chrDstW, lumDither, chrDither); } } else { assert(lumSrcPtr + vLumFilterSize - 1 < lumPixBuf + vLumBufSize*2); assert(chrUSrcPtr + vChrFilterSize - 1 < chrUPixBuf + vChrBufSize*2); if(flags & SWS_FULL_CHR_H_INT) { yuv2rgbX_c_full(c, vLumFilter+dstY*vLumFilterSize, lumSrcPtr, vLumFilterSize, vChrFilter+dstY*vChrFilterSize, chrUSrcPtr, chrVSrcPtr, vChrFilterSize, alpSrcPtr, dest, dstW, dstY); } else { yuv2packedX_c(c, vLumFilter+dstY*vLumFilterSize, lumSrcPtr, vLumFilterSize, vChrFilter+dstY*vChrFilterSize, chrUSrcPtr, chrVSrcPtr, vChrFilterSize, alpSrcPtr, dest, dstW, dstY); } } } } if ((dstFormat == PIX_FMT_YUVA420P) && !alpPixBuf) fillPlane(dst[3], dstStride[3], dstW, dstY-lastDstY, lastDstY, 255); #if HAVE_MMX2 if (av_get_cpu_flags() & AV_CPU_FLAG_MMX2) __asm__ volatile("sfence":::"memory"); #endif emms_c(); /* store changed local vars back in the context */ c->dstY= dstY; c->lumBufIndex= lumBufIndex; c->chrBufIndex= chrBufIndex; c->lastInLumBuf= lastInLumBuf; c->lastInChrBuf= lastInChrBuf; return dstY - lastDstY; } static void sws_init_swScale_c(SwsContext *c) { enum PixelFormat srcFormat = c->srcFormat; c->yuv2nv12X = yuv2nv12X_c; c->yuv2yuv1 = yuv2yuv1_c; c->yuv2yuvX = yuv2yuvX_c; c->yuv2packed1 = yuv2packed1_c; c->yuv2packed2 = yuv2packed2_c; c->yuv2packedX = yuv2packedX_c; c->hScale = hScale_c; if (c->flags & SWS_FAST_BILINEAR) { c->hyscale_fast = hyscale_fast_c; c->hcscale_fast = hcscale_fast_c; } c->chrToYV12 = NULL; switch(srcFormat) { case PIX_FMT_YUYV422 : c->chrToYV12 = yuy2ToUV_c; break; case PIX_FMT_UYVY422 : c->chrToYV12 = uyvyToUV_c; break; case PIX_FMT_NV12 : c->chrToYV12 = nv12ToUV_c; break; case PIX_FMT_NV21 : c->chrToYV12 = nv21ToUV_c; break; case PIX_FMT_RGB8 : case PIX_FMT_BGR8 : case PIX_FMT_PAL8 : case PIX_FMT_BGR4_BYTE: case PIX_FMT_RGB4_BYTE: c->chrToYV12 = palToUV_c; break; case PIX_FMT_GRAY16BE : case PIX_FMT_YUV420P9BE: case PIX_FMT_YUV422P10BE: case PIX_FMT_YUV420P10BE: case PIX_FMT_YUV420P16BE: case PIX_FMT_YUV422P16BE: case PIX_FMT_YUV444P16BE: c->hScale16= HAVE_BIGENDIAN ? hScale16_c : hScale16X_c; break; case PIX_FMT_GRAY16LE : case PIX_FMT_YUV420P9LE: case PIX_FMT_YUV422P10LE: case PIX_FMT_YUV420P10LE: case PIX_FMT_YUV420P16LE: case PIX_FMT_YUV422P16LE: case PIX_FMT_YUV444P16LE: c->hScale16= HAVE_BIGENDIAN ? hScale16X_c : hScale16_c; break; } if (c->chrSrcHSubSample) { switch(srcFormat) { case PIX_FMT_RGB48BE: c->chrToYV12 = rgb48BEToUV_half_c; break; case PIX_FMT_RGB48LE: c->chrToYV12 = rgb48LEToUV_half_c; break; case PIX_FMT_BGR48BE: c->chrToYV12 = bgr48BEToUV_half_c; break; case PIX_FMT_BGR48LE: c->chrToYV12 = bgr48LEToUV_half_c; break; case PIX_FMT_RGB32 : c->chrToYV12 = bgr32ToUV_half_c; break; case PIX_FMT_RGB32_1: c->chrToYV12 = bgr321ToUV_half_c; break; case PIX_FMT_BGR24 : c->chrToYV12 = bgr24ToUV_half_c; break; case PIX_FMT_BGR565 : c->chrToYV12 = bgr16ToUV_half_c; break; case PIX_FMT_BGR555 : c->chrToYV12 = bgr15ToUV_half_c; break; case PIX_FMT_BGR32 : c->chrToYV12 = rgb32ToUV_half_c; break; case PIX_FMT_BGR32_1: c->chrToYV12 = rgb321ToUV_half_c; break; case PIX_FMT_RGB24 : c->chrToYV12 = rgb24ToUV_half_c; break; case PIX_FMT_RGB565 : c->chrToYV12 = rgb16ToUV_half_c; break; case PIX_FMT_RGB555 : c->chrToYV12 = rgb15ToUV_half_c; break; } } else { switch(srcFormat) { case PIX_FMT_RGB48BE: c->chrToYV12 = rgb48BEToUV_c; break; case PIX_FMT_RGB48LE: c->chrToYV12 = rgb48LEToUV_c; break; case PIX_FMT_BGR48BE: c->chrToYV12 = bgr48BEToUV_c; break; case PIX_FMT_BGR48LE: c->chrToYV12 = bgr48LEToUV_c; break; case PIX_FMT_RGB32 : c->chrToYV12 = bgr32ToUV_c; break; case PIX_FMT_RGB32_1: c->chrToYV12 = bgr321ToUV_c; break; case PIX_FMT_BGR24 : c->chrToYV12 = bgr24ToUV_c; break; case PIX_FMT_BGR565 : c->chrToYV12 = bgr16ToUV_c; break; case PIX_FMT_BGR555 : c->chrToYV12 = bgr15ToUV_c; break; case PIX_FMT_BGR32 : c->chrToYV12 = rgb32ToUV_c; break; case PIX_FMT_BGR32_1: c->chrToYV12 = rgb321ToUV_c; break; case PIX_FMT_RGB24 : c->chrToYV12 = rgb24ToUV_c; break; case PIX_FMT_RGB565 : c->chrToYV12 = rgb16ToUV_c; break; case PIX_FMT_RGB555 : c->chrToYV12 = rgb15ToUV_c; break; } } c->lumToYV12 = NULL; c->alpToYV12 = NULL; switch (srcFormat) { case PIX_FMT_YUYV422 : case PIX_FMT_GRAY8A : c->lumToYV12 = yuy2ToY_c; break; case PIX_FMT_UYVY422 : c->lumToYV12 = uyvyToY_c; break; case PIX_FMT_BGR24 : c->lumToYV12 = bgr24ToY_c; break; case PIX_FMT_BGR565 : c->lumToYV12 = bgr16ToY_c; break; case PIX_FMT_BGR555 : c->lumToYV12 = bgr15ToY_c; break; case PIX_FMT_RGB24 : c->lumToYV12 = rgb24ToY_c; break; case PIX_FMT_RGB565 : c->lumToYV12 = rgb16ToY_c; break; case PIX_FMT_RGB555 : c->lumToYV12 = rgb15ToY_c; break; case PIX_FMT_RGB8 : case PIX_FMT_BGR8 : case PIX_FMT_PAL8 : case PIX_FMT_BGR4_BYTE: case PIX_FMT_RGB4_BYTE: c->lumToYV12 = palToY_c; break; case PIX_FMT_MONOBLACK: c->lumToYV12 = monoblack2Y_c; break; case PIX_FMT_MONOWHITE: c->lumToYV12 = monowhite2Y_c; break; case PIX_FMT_RGB32 : c->lumToYV12 = bgr32ToY_c; break; case PIX_FMT_RGB32_1: c->lumToYV12 = bgr321ToY_c; break; case PIX_FMT_BGR32 : c->lumToYV12 = rgb32ToY_c; break; case PIX_FMT_BGR32_1: c->lumToYV12 = rgb321ToY_c; break; case PIX_FMT_RGB48BE: c->lumToYV12 = rgb48BEToY_c; break; case PIX_FMT_RGB48LE: c->lumToYV12 = rgb48LEToY_c; break; case PIX_FMT_BGR48BE: c->lumToYV12 = bgr48BEToY_c; break; case PIX_FMT_BGR48LE: c->lumToYV12 = bgr48LEToY_c; break; } if (c->alpPixBuf) { switch (srcFormat) { case PIX_FMT_RGB32 : case PIX_FMT_RGB32_1: case PIX_FMT_BGR32 : case PIX_FMT_BGR32_1: c->alpToYV12 = abgrToA_c; break; case PIX_FMT_GRAY8A : c->alpToYV12 = yuy2ToY_c; break; case PIX_FMT_PAL8 : c->alpToYV12 = palToA_c; break; } } if(isAnyRGB(c->srcFormat) || c->srcFormat == PIX_FMT_PAL8) c->hScale16= hScale16_c; switch (srcFormat) { case PIX_FMT_GRAY8A : c->alpSrcOffset = 1; break; case PIX_FMT_RGB32 : case PIX_FMT_BGR32 : c->alpSrcOffset = 3; break; } if (c->srcRange != c->dstRange && !isAnyRGB(c->dstFormat)) { if (c->srcRange) { c->lumConvertRange = lumRangeFromJpeg_c; c->chrConvertRange = chrRangeFromJpeg_c; } else { c->lumConvertRange = lumRangeToJpeg_c; c->chrConvertRange = chrRangeToJpeg_c; } } if (!(isGray(srcFormat) || isGray(c->dstFormat) || srcFormat == PIX_FMT_MONOBLACK || srcFormat == PIX_FMT_MONOWHITE)) c->needs_hcscale = 1; } SwsFunc ff_getSwsFunc(SwsContext *c) { sws_init_swScale_c(c); if (HAVE_MMX) ff_sws_init_swScale_mmx(c); if (HAVE_ALTIVEC) ff_sws_init_swScale_altivec(c); return swScale; }