libswscale/swscale_internal.h
Go to the documentation of this file.
00001 /*
00002  * Copyright (C) 2001-2003 Michael Niedermayer <michaelni@gmx.at>
00003  *
00004  * This file is part of Libav.
00005  *
00006  * Libav is free software; you can redistribute it and/or
00007  * modify it under the terms of the GNU Lesser General Public
00008  * License as published by the Free Software Foundation; either
00009  * version 2.1 of the License, or (at your option) any later version.
00010  *
00011  * Libav is distributed in the hope that it will be useful,
00012  * but WITHOUT ANY WARRANTY; without even the implied warranty of
00013  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
00014  * Lesser General Public License for more details.
00015  *
00016  * You should have received a copy of the GNU Lesser General Public
00017  * License along with Libav; if not, write to the Free Software
00018  * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
00019  */
00020 
00021 #ifndef SWSCALE_SWSCALE_INTERNAL_H
00022 #define SWSCALE_SWSCALE_INTERNAL_H
00023 
00024 #include "config.h"
00025 
00026 #if HAVE_ALTIVEC_H
00027 #include <altivec.h>
00028 #endif
00029 
00030 #include "libavutil/avutil.h"
00031 #include "libavutil/log.h"
00032 #include "libavutil/pixfmt.h"
00033 #include "libavutil/pixdesc.h"
00034 
00035 #define STR(s) AV_TOSTRING(s) // AV_STRINGIFY is too long
00036 
00037 #define FAST_BGR2YV12 // use 7-bit instead of 15-bit coefficients
00038 
00039 #define MAX_FILTER_SIZE 256
00040 
00041 #if HAVE_BIGENDIAN
00042 #define ALT32_CORR (-1)
00043 #else
00044 #define ALT32_CORR   1
00045 #endif
00046 
00047 #if ARCH_X86_64
00048 #   define APCK_PTR2  8
00049 #   define APCK_COEF 16
00050 #   define APCK_SIZE 24
00051 #else
00052 #   define APCK_PTR2  4
00053 #   define APCK_COEF  8
00054 #   define APCK_SIZE 16
00055 #endif
00056 
00057 struct SwsContext;
00058 
00059 typedef int (*SwsFunc)(struct SwsContext *context, const uint8_t *src[],
00060                        int srcStride[], int srcSliceY, int srcSliceH,
00061                        uint8_t *dst[], int dstStride[]);
00062 
00075 typedef void (*yuv2planar1_fn)(const int16_t *src, uint8_t *dest, int dstW,
00076                                const uint8_t *dither, int offset);
00077 
00091 typedef void (*yuv2planarX_fn)(const int16_t *filter, int filterSize,
00092                                const int16_t **src, uint8_t *dest, int dstW,
00093                                const uint8_t *dither, int offset);
00094 
00110 typedef void (*yuv2interleavedX_fn)(struct SwsContext *c,
00111                                     const int16_t *chrFilter,
00112                                     int chrFilterSize,
00113                                     const int16_t **chrUSrc,
00114                                     const int16_t **chrVSrc,
00115                                     uint8_t *dest, int dstW);
00116 
00146 typedef void (*yuv2packed1_fn)(struct SwsContext *c, const int16_t *lumSrc,
00147                                const int16_t *chrUSrc[2],
00148                                const int16_t *chrVSrc[2],
00149                                const int16_t *alpSrc, uint8_t *dest,
00150                                int dstW, int uvalpha, int y);
00179 typedef void (*yuv2packed2_fn)(struct SwsContext *c, const int16_t *lumSrc[2],
00180                                const int16_t *chrUSrc[2],
00181                                const int16_t *chrVSrc[2],
00182                                const int16_t *alpSrc[2],
00183                                uint8_t *dest,
00184                                int dstW, int yalpha, int uvalpha, int y);
00211 typedef void (*yuv2packedX_fn)(struct SwsContext *c, const int16_t *lumFilter,
00212                                const int16_t **lumSrc, int lumFilterSize,
00213                                const int16_t *chrFilter,
00214                                const int16_t **chrUSrc,
00215                                const int16_t **chrVSrc, int chrFilterSize,
00216                                const int16_t **alpSrc, uint8_t *dest,
00217                                int dstW, int y);
00218 
00219 /* This struct should be aligned on at least a 32-byte boundary. */
00220 typedef struct SwsContext {
00224     const AVClass *av_class;
00225 
00230     SwsFunc swScale;
00231     int srcW;                     
00232     int srcH;                     
00233     int dstH;                     
00234     int chrSrcW;                  
00235     int chrSrcH;                  
00236     int chrDstW;                  
00237     int chrDstH;                  
00238     int lumXInc, chrXInc;
00239     int lumYInc, chrYInc;
00240     enum PixelFormat dstFormat;   
00241     enum PixelFormat srcFormat;   
00242     int dstFormatBpp;             
00243     int srcFormatBpp;             
00244     int dstBpc, srcBpc;
00245     int chrSrcHSubSample;         
00246     int chrSrcVSubSample;         
00247     int chrDstHSubSample;         
00248     int chrDstVSubSample;         
00249     int vChrDrop;                 
00250     int sliceDir;                 
00251     double param[2];              
00252 
00253     uint32_t pal_yuv[256];
00254     uint32_t pal_rgb[256];
00255 
00266     int16_t **lumPixBuf;          
00267     int16_t **chrUPixBuf;         
00268     int16_t **chrVPixBuf;         
00269     int16_t **alpPixBuf;          
00270     int vLumBufSize;              
00271     int vChrBufSize;              
00272     int lastInLumBuf;             
00273     int lastInChrBuf;             
00274     int lumBufIndex;              
00275     int chrBufIndex;              
00276 
00277 
00278     uint8_t *formatConvBuffer;
00279 
00294     int16_t *hLumFilter;          
00295     int16_t *hChrFilter;          
00296     int16_t *vLumFilter;          
00297     int16_t *vChrFilter;          
00298     int32_t *hLumFilterPos;       
00299     int32_t *hChrFilterPos;       
00300     int32_t *vLumFilterPos;       
00301     int32_t *vChrFilterPos;       
00302     int hLumFilterSize;           
00303     int hChrFilterSize;           
00304     int vLumFilterSize;           
00305     int vChrFilterSize;           
00306 
00307 
00308     int lumMmx2FilterCodeSize;    
00309     int chrMmx2FilterCodeSize;    
00310     uint8_t *lumMmx2FilterCode;   
00311     uint8_t *chrMmx2FilterCode;   
00312 
00313     int canMMX2BeUsed;
00314 
00315     int dstY;                     
00316     int flags;                    
00317     void *yuvTable;             // pointer to the yuv->rgb table start so it can be freed()
00318     uint8_t *table_rV[256];
00319     uint8_t *table_gU[256];
00320     int table_gV[256];
00321     uint8_t *table_bU[256];
00322 
00323     //Colorspace stuff
00324     int contrast, brightness, saturation;    // for sws_getColorspaceDetails
00325     int srcColorspaceTable[4];
00326     int dstColorspaceTable[4];
00327     int srcRange;                 
00328     int dstRange;                 
00329     int yuv2rgb_y_offset;
00330     int yuv2rgb_y_coeff;
00331     int yuv2rgb_v2r_coeff;
00332     int yuv2rgb_v2g_coeff;
00333     int yuv2rgb_u2g_coeff;
00334     int yuv2rgb_u2b_coeff;
00335 
00336 #define RED_DITHER            "0*8"
00337 #define GREEN_DITHER          "1*8"
00338 #define BLUE_DITHER           "2*8"
00339 #define Y_COEFF               "3*8"
00340 #define VR_COEFF              "4*8"
00341 #define UB_COEFF              "5*8"
00342 #define VG_COEFF              "6*8"
00343 #define UG_COEFF              "7*8"
00344 #define Y_OFFSET              "8*8"
00345 #define U_OFFSET              "9*8"
00346 #define V_OFFSET              "10*8"
00347 #define LUM_MMX_FILTER_OFFSET "11*8"
00348 #define CHR_MMX_FILTER_OFFSET "11*8+4*4*256"
00349 #define DSTW_OFFSET           "11*8+4*4*256*2" //do not change, it is hardcoded in the ASM
00350 #define ESP_OFFSET            "11*8+4*4*256*2+8"
00351 #define VROUNDER_OFFSET       "11*8+4*4*256*2+16"
00352 #define U_TEMP                "11*8+4*4*256*2+24"
00353 #define V_TEMP                "11*8+4*4*256*2+32"
00354 #define Y_TEMP                "11*8+4*4*256*2+40"
00355 #define ALP_MMX_FILTER_OFFSET "11*8+4*4*256*2+48"
00356 #define UV_OFF_PX             "11*8+4*4*256*3+48"
00357 #define UV_OFF_BYTE           "11*8+4*4*256*3+56"
00358 #define DITHER16              "11*8+4*4*256*3+64"
00359 #define DITHER32              "11*8+4*4*256*3+80"
00360 
00361     DECLARE_ALIGNED(8, uint64_t, redDither);
00362     DECLARE_ALIGNED(8, uint64_t, greenDither);
00363     DECLARE_ALIGNED(8, uint64_t, blueDither);
00364 
00365     DECLARE_ALIGNED(8, uint64_t, yCoeff);
00366     DECLARE_ALIGNED(8, uint64_t, vrCoeff);
00367     DECLARE_ALIGNED(8, uint64_t, ubCoeff);
00368     DECLARE_ALIGNED(8, uint64_t, vgCoeff);
00369     DECLARE_ALIGNED(8, uint64_t, ugCoeff);
00370     DECLARE_ALIGNED(8, uint64_t, yOffset);
00371     DECLARE_ALIGNED(8, uint64_t, uOffset);
00372     DECLARE_ALIGNED(8, uint64_t, vOffset);
00373     int32_t lumMmxFilter[4 * MAX_FILTER_SIZE];
00374     int32_t chrMmxFilter[4 * MAX_FILTER_SIZE];
00375     int dstW;                     
00376     DECLARE_ALIGNED(8, uint64_t, esp);
00377     DECLARE_ALIGNED(8, uint64_t, vRounder);
00378     DECLARE_ALIGNED(8, uint64_t, u_temp);
00379     DECLARE_ALIGNED(8, uint64_t, v_temp);
00380     DECLARE_ALIGNED(8, uint64_t, y_temp);
00381     int32_t alpMmxFilter[4 * MAX_FILTER_SIZE];
00382     // alignment of these values is not necessary, but merely here
00383     // to maintain the same offset across x8632 and x86-64. Once we
00384     // use proper offset macros in the asm, they can be removed.
00385     DECLARE_ALIGNED(8, ptrdiff_t, uv_off_px);   
00386     DECLARE_ALIGNED(8, ptrdiff_t, uv_off_byte); 
00387     DECLARE_ALIGNED(8, uint16_t, dither16)[8];
00388     DECLARE_ALIGNED(8, uint32_t, dither32)[8];
00389 
00390     const uint8_t *chrDither8, *lumDither8;
00391 
00392 #if HAVE_ALTIVEC
00393     vector signed short   CY;
00394     vector signed short   CRV;
00395     vector signed short   CBU;
00396     vector signed short   CGU;
00397     vector signed short   CGV;
00398     vector signed short   OY;
00399     vector unsigned short CSHIFT;
00400     vector signed short  *vYCoeffsBank, *vCCoeffsBank;
00401 #endif
00402 
00403 #if ARCH_BFIN
00404     DECLARE_ALIGNED(4, uint32_t, oy);
00405     DECLARE_ALIGNED(4, uint32_t, oc);
00406     DECLARE_ALIGNED(4, uint32_t, zero);
00407     DECLARE_ALIGNED(4, uint32_t, cy);
00408     DECLARE_ALIGNED(4, uint32_t, crv);
00409     DECLARE_ALIGNED(4, uint32_t, rmask);
00410     DECLARE_ALIGNED(4, uint32_t, cbu);
00411     DECLARE_ALIGNED(4, uint32_t, bmask);
00412     DECLARE_ALIGNED(4, uint32_t, cgu);
00413     DECLARE_ALIGNED(4, uint32_t, cgv);
00414     DECLARE_ALIGNED(4, uint32_t, gmask);
00415 #endif
00416 
00417 #if HAVE_VIS
00418     DECLARE_ALIGNED(8, uint64_t, sparc_coeffs)[10];
00419 #endif
00420 
00421     /* function pointers for swScale() */
00422     yuv2planar1_fn yuv2plane1;
00423     yuv2planarX_fn yuv2planeX;
00424     yuv2interleavedX_fn yuv2nv12cX;
00425     yuv2packed1_fn yuv2packed1;
00426     yuv2packed2_fn yuv2packed2;
00427     yuv2packedX_fn yuv2packedX;
00428 
00430     void (*lumToYV12)(uint8_t *dst, const uint8_t *src,
00431                       int width, uint32_t *pal);
00433     void (*alpToYV12)(uint8_t *dst, const uint8_t *src,
00434                       int width, uint32_t *pal);
00436     void (*chrToYV12)(uint8_t *dstU, uint8_t *dstV,
00437                       const uint8_t *src1, const uint8_t *src2,
00438                       int width, uint32_t *pal);
00439 
00445     void (*readLumPlanar)(uint8_t *dst, const uint8_t *src[4], int width);
00446     void (*readChrPlanar)(uint8_t *dstU, uint8_t *dstV, const uint8_t *src[4],
00447                           int width);
00469     void (*hyscale_fast)(struct SwsContext *c,
00470                          int16_t *dst, int dstWidth,
00471                          const uint8_t *src, int srcW, int xInc);
00472     void (*hcscale_fast)(struct SwsContext *c,
00473                          int16_t *dst1, int16_t *dst2, int dstWidth,
00474                          const uint8_t *src1, const uint8_t *src2,
00475                          int srcW, int xInc);
00509     void (*hyScale)(struct SwsContext *c, int16_t *dst, int dstW,
00510                     const uint8_t *src, const int16_t *filter,
00511                     const int32_t *filterPos, int filterSize);
00512     void (*hcScale)(struct SwsContext *c, int16_t *dst, int dstW,
00513                     const uint8_t *src, const int16_t *filter,
00514                     const int32_t *filterPos, int filterSize);
00517 
00518     void (*lumConvertRange)(int16_t *dst, int width);
00520     void (*chrConvertRange)(int16_t *dst1, int16_t *dst2, int width);
00521 
00522     int needs_hcscale; 
00523 } SwsContext;
00524 //FIXME check init (where 0)
00525 
00526 SwsFunc ff_yuv2rgb_get_func_ptr(SwsContext *c);
00527 int ff_yuv2rgb_c_init_tables(SwsContext *c, const int inv_table[4],
00528                              int fullRange, int brightness,
00529                              int contrast, int saturation);
00530 
00531 void ff_yuv2rgb_init_tables_altivec(SwsContext *c, const int inv_table[4],
00532                                     int brightness, int contrast, int saturation);
00533 void updateMMXDitherTables(SwsContext *c, int dstY, int lumBufIndex, int chrBufIndex,
00534                            int lastInLumBuf, int lastInChrBuf);
00535 
00536 SwsFunc ff_yuv2rgb_init_mmx(SwsContext *c);
00537 SwsFunc ff_yuv2rgb_init_vis(SwsContext *c);
00538 SwsFunc ff_yuv2rgb_init_mlib(SwsContext *c);
00539 SwsFunc ff_yuv2rgb_init_altivec(SwsContext *c);
00540 SwsFunc ff_yuv2rgb_get_func_ptr_bfin(SwsContext *c);
00541 void ff_bfin_get_unscaled_swscale(SwsContext *c);
00542 
00543 const char *sws_format_name(enum PixelFormat format);
00544 
00545 #define is16BPS(x) \
00546     (av_pix_fmt_descriptors[x].comp[0].depth_minus1 == 15)
00547 
00548 #define is9_OR_10BPS(x) \
00549     (av_pix_fmt_descriptors[x].comp[0].depth_minus1 == 8 || \
00550      av_pix_fmt_descriptors[x].comp[0].depth_minus1 == 9)
00551 
00552 #define isBE(x) \
00553     (av_pix_fmt_descriptors[x].flags & PIX_FMT_BE)
00554 
00555 #define isYUV(x) \
00556     (!(av_pix_fmt_descriptors[x].flags & PIX_FMT_RGB) && \
00557      av_pix_fmt_descriptors[x].nb_components >= 2)
00558 
00559 #define isPlanarYUV(x) \
00560     ((av_pix_fmt_descriptors[x].flags & PIX_FMT_PLANAR) && \
00561      isYUV(x))
00562 
00563 #define isRGB(x) \
00564     (av_pix_fmt_descriptors[x].flags & PIX_FMT_RGB)
00565 
00566 #if 0 // FIXME
00567 #define isGray(x) \
00568     (!(av_pix_fmt_descriptors[x].flags & PIX_FMT_PAL) && \
00569      av_pix_fmt_descriptors[x].nb_components <= 2)
00570 #else
00571 #define isGray(x)                      \
00572     ((x) == PIX_FMT_GRAY8       ||     \
00573      (x) == PIX_FMT_Y400A       ||     \
00574      (x) == PIX_FMT_GRAY16BE    ||     \
00575      (x) == PIX_FMT_GRAY16LE)
00576 #endif
00577 
00578 #define isRGBinInt(x)                  \
00579     ((x) == PIX_FMT_RGB48BE     ||     \
00580      (x) == PIX_FMT_RGB48LE     ||     \
00581      (x) == PIX_FMT_RGB32       ||     \
00582      (x) == PIX_FMT_RGB32_1     ||     \
00583      (x) == PIX_FMT_RGB24       ||     \
00584      (x) == PIX_FMT_RGB565BE    ||     \
00585      (x) == PIX_FMT_RGB565LE    ||     \
00586      (x) == PIX_FMT_RGB555BE    ||     \
00587      (x) == PIX_FMT_RGB555LE    ||     \
00588      (x) == PIX_FMT_RGB444BE    ||     \
00589      (x) == PIX_FMT_RGB444LE    ||     \
00590      (x) == PIX_FMT_RGB8        ||     \
00591      (x) == PIX_FMT_RGB4        ||     \
00592      (x) == PIX_FMT_RGB4_BYTE   ||     \
00593      (x) == PIX_FMT_MONOBLACK   ||     \
00594      (x) == PIX_FMT_MONOWHITE)
00595 
00596 #define isBGRinInt(x)                  \
00597     ((x) == PIX_FMT_BGR48BE     ||     \
00598      (x) == PIX_FMT_BGR48LE     ||     \
00599      (x) == PIX_FMT_BGR32       ||     \
00600      (x) == PIX_FMT_BGR32_1     ||     \
00601      (x) == PIX_FMT_BGR24       ||     \
00602      (x) == PIX_FMT_BGR565BE    ||     \
00603      (x) == PIX_FMT_BGR565LE    ||     \
00604      (x) == PIX_FMT_BGR555BE    ||     \
00605      (x) == PIX_FMT_BGR555LE    ||     \
00606      (x) == PIX_FMT_BGR444BE    ||     \
00607      (x) == PIX_FMT_BGR444LE    ||     \
00608      (x) == PIX_FMT_BGR8        ||     \
00609      (x) == PIX_FMT_BGR4        ||     \
00610      (x) == PIX_FMT_BGR4_BYTE   ||     \
00611      (x) == PIX_FMT_MONOBLACK   ||     \
00612      (x) == PIX_FMT_MONOWHITE)
00613 
00614 #define isAnyRGB(x)                    \
00615     (isRGBinInt(x)              ||     \
00616      isBGRinInt(x))
00617 
00618 #define isALPHA(x)                                            \
00619     (av_pix_fmt_descriptors[x].nb_components == 2          || \
00620      av_pix_fmt_descriptors[x].nb_components == 4)
00621 
00622 #define isPacked(x)                                            \
00623     ((av_pix_fmt_descriptors[x].nb_components >= 2         &&  \
00624       !(av_pix_fmt_descriptors[x].flags & PIX_FMT_PLANAR)) ||  \
00625      (x) == PIX_FMT_PAL8)
00626 
00627 #define isPlanar(x) \
00628     (av_pix_fmt_descriptors[x].nb_components >= 2          &&  \
00629      (av_pix_fmt_descriptors[x].flags & PIX_FMT_PLANAR))
00630 
00631 #define usePal(x) ((av_pix_fmt_descriptors[x].flags & PIX_FMT_PAL) || (x) == PIX_FMT_Y400A)
00632 
00633 extern const uint64_t ff_dither4[2];
00634 extern const uint64_t ff_dither8[2];
00635 
00636 extern const AVClass sws_context_class;
00637 
00642 void ff_get_unscaled_swscale(SwsContext *c);
00643 
00644 void ff_swscale_get_unscaled_altivec(SwsContext *c);
00645 
00650 SwsFunc ff_getSwsFunc(SwsContext *c);
00651 
00652 void ff_sws_init_swScale_altivec(SwsContext *c);
00653 void ff_sws_init_swScale_mmx(SwsContext *c);
00654 
00655 #endif /* SWSCALE_SWSCALE_INTERNAL_H */