• Main Page
  • Related Pages
  • Modules
  • Data Structures
  • Files
  • File List
  • Globals

libavfilter/libmpcodecs/vf_spp.c

Go to the documentation of this file.
00001 /*
00002  * Copyright (C) 2003 Michael Niedermayer <michaelni@gmx.at>
00003  *
00004  * This file is part of MPlayer.
00005  *
00006  * MPlayer is free software; you can redistribute it and/or modify
00007  * it under the terms of the GNU General Public License as published by
00008  * the Free Software Foundation; either version 2 of the License, or
00009  * (at your option) any later version.
00010  *
00011  * MPlayer 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
00014  * GNU General Public License for more details.
00015  *
00016  * You should have received a copy of the GNU General Public License along
00017  * with MPlayer; if not, write to the Free Software Foundation, Inc.,
00018  * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
00019  */
00020 
00021 /*
00022  * This implementation is based on an algorithm described in
00023  * "Aria Nosratinia Embedded Post-Processing for
00024  * Enhancement of Compressed Images (1999)"
00025  * (http://citeseer.nj.nec.com/nosratinia99embedded.html)
00026  */
00027 
00028 
00029 #include <stdio.h>
00030 #include <stdlib.h>
00031 #include <string.h>
00032 #include <inttypes.h>
00033 #include <math.h>
00034 
00035 #include "config.h"
00036 
00037 #include "mp_msg.h"
00038 #include "cpudetect.h"
00039 
00040 #include "libavutil/internal.h"
00041 #include "libavutil/intreadwrite.h"
00042 #include "libavcodec/avcodec.h"
00043 #include "libavcodec/dsputil.h"
00044 
00045 #undef fprintf
00046 #undef free
00047 #undef malloc
00048 
00049 #include "img_format.h"
00050 #include "mp_image.h"
00051 #include "vf.h"
00052 #include "vd_ffmpeg.h"
00053 #include "libvo/fastmemcpy.h"
00054 
00055 #define XMIN(a,b) ((a) < (b) ? (a) : (b))
00056 
00057 //===========================================================================//
00058 static const uint8_t  __attribute__((aligned(8))) dither[8][8]={
00059 {  0,  48,  12,  60,   3,  51,  15,  63, },
00060 { 32,  16,  44,  28,  35,  19,  47,  31, },
00061 {  8,  56,   4,  52,  11,  59,   7,  55, },
00062 { 40,  24,  36,  20,  43,  27,  39,  23, },
00063 {  2,  50,  14,  62,   1,  49,  13,  61, },
00064 { 34,  18,  46,  30,  33,  17,  45,  29, },
00065 { 10,  58,   6,  54,   9,  57,   5,  53, },
00066 { 42,  26,  38,  22,  41,  25,  37,  21, },
00067 };
00068 
00069 static const uint8_t offset[127][2]= {
00070 {0,0},
00071 {0,0}, {4,4},
00072 {0,0}, {2,2}, {6,4}, {4,6},
00073 {0,0}, {5,1}, {2,2}, {7,3}, {4,4}, {1,5}, {6,6}, {3,7},
00074 
00075 {0,0}, {4,0}, {1,1}, {5,1}, {3,2}, {7,2}, {2,3}, {6,3},
00076 {0,4}, {4,4}, {1,5}, {5,5}, {3,6}, {7,6}, {2,7}, {6,7},
00077 
00078 {0,0}, {0,2}, {0,4}, {0,6}, {1,1}, {1,3}, {1,5}, {1,7},
00079 {2,0}, {2,2}, {2,4}, {2,6}, {3,1}, {3,3}, {3,5}, {3,7},
00080 {4,0}, {4,2}, {4,4}, {4,6}, {5,1}, {5,3}, {5,5}, {5,7},
00081 {6,0}, {6,2}, {6,4}, {6,6}, {7,1}, {7,3}, {7,5}, {7,7},
00082 
00083 {0,0}, {4,4}, {0,4}, {4,0}, {2,2}, {6,6}, {2,6}, {6,2},
00084 {0,2}, {4,6}, {0,6}, {4,2}, {2,0}, {6,4}, {2,4}, {6,0},
00085 {1,1}, {5,5}, {1,5}, {5,1}, {3,3}, {7,7}, {3,7}, {7,3},
00086 {1,3}, {5,7}, {1,7}, {5,3}, {3,1}, {7,5}, {3,5}, {7,1},
00087 {0,1}, {4,5}, {0,5}, {4,1}, {2,3}, {6,7}, {2,7}, {6,3},
00088 {0,3}, {4,7}, {0,7}, {4,3}, {2,1}, {6,5}, {2,5}, {6,1},
00089 {1,0}, {5,4}, {1,4}, {5,0}, {3,2}, {7,6}, {3,6}, {7,2},
00090 {1,2}, {5,6}, {1,6}, {5,2}, {3,0}, {7,4}, {3,4}, {7,0},
00091 };
00092 
00093 struct vf_priv_s {
00094         int log2_count;
00095         int qp;
00096         int mode;
00097         int mpeg2;
00098         int temp_stride;
00099         uint8_t *src;
00100         int16_t *temp;
00101         AVCodecContext *avctx;
00102         DSPContext dsp;
00103         char *non_b_qp;
00104 };
00105 
00106 #define SHIFT 22
00107 
00108 static void hardthresh_c(DCTELEM dst[64], DCTELEM src[64], int qp, uint8_t *permutation){
00109         int i;
00110         int bias= 0; //FIXME
00111         unsigned int threshold1, threshold2;
00112 
00113         threshold1= qp*((1<<4) - bias) - 1;
00114         threshold2= (threshold1<<1);
00115 
00116         memset(dst, 0, 64*sizeof(DCTELEM));
00117         dst[0]= (src[0] + 4)>>3;
00118 
00119         for(i=1; i<64; i++){
00120                 int level= src[i];
00121                 if(((unsigned)(level+threshold1))>threshold2){
00122                         const int j= permutation[i];
00123                         dst[j]= (level + 4)>>3;
00124                 }
00125         }
00126 }
00127 
00128 static void softthresh_c(DCTELEM dst[64], DCTELEM src[64], int qp, uint8_t *permutation){
00129         int i;
00130         int bias= 0; //FIXME
00131         unsigned int threshold1, threshold2;
00132 
00133         threshold1= qp*((1<<4) - bias) - 1;
00134         threshold2= (threshold1<<1);
00135 
00136         memset(dst, 0, 64*sizeof(DCTELEM));
00137         dst[0]= (src[0] + 4)>>3;
00138 
00139         for(i=1; i<64; i++){
00140                 int level= src[i];
00141                 if(((unsigned)(level+threshold1))>threshold2){
00142                         const int j= permutation[i];
00143                         if(level>0)
00144                                 dst[j]= (level - threshold1 + 4)>>3;
00145                         else
00146                                 dst[j]= (level + threshold1 + 4)>>3;
00147                 }
00148         }
00149 }
00150 
00151 #if HAVE_MMX
00152 static void hardthresh_mmx(DCTELEM dst[64], DCTELEM src[64], int qp, uint8_t *permutation){
00153         int bias= 0; //FIXME
00154         unsigned int threshold1;
00155 
00156         threshold1= qp*((1<<4) - bias) - 1;
00157 
00158         __asm__ volatile(
00159 #define REQUANT_CORE(dst0, dst1, dst2, dst3, src0, src1, src2, src3) \
00160                 "movq " #src0 ", %%mm0        \n\t"\
00161                 "movq " #src1 ", %%mm1        \n\t"\
00162                 "movq " #src2 ", %%mm2        \n\t"\
00163                 "movq " #src3 ", %%mm3        \n\t"\
00164                 "psubw %%mm4, %%mm0        \n\t"\
00165                 "psubw %%mm4, %%mm1        \n\t"\
00166                 "psubw %%mm4, %%mm2        \n\t"\
00167                 "psubw %%mm4, %%mm3        \n\t"\
00168                 "paddusw %%mm5, %%mm0        \n\t"\
00169                 "paddusw %%mm5, %%mm1        \n\t"\
00170                 "paddusw %%mm5, %%mm2        \n\t"\
00171                 "paddusw %%mm5, %%mm3        \n\t"\
00172                 "paddw %%mm6, %%mm0        \n\t"\
00173                 "paddw %%mm6, %%mm1        \n\t"\
00174                 "paddw %%mm6, %%mm2        \n\t"\
00175                 "paddw %%mm6, %%mm3        \n\t"\
00176                 "psubusw %%mm6, %%mm0        \n\t"\
00177                 "psubusw %%mm6, %%mm1        \n\t"\
00178                 "psubusw %%mm6, %%mm2        \n\t"\
00179                 "psubusw %%mm6, %%mm3        \n\t"\
00180                 "psraw $3, %%mm0        \n\t"\
00181                 "psraw $3, %%mm1        \n\t"\
00182                 "psraw $3, %%mm2        \n\t"\
00183                 "psraw $3, %%mm3        \n\t"\
00184 \
00185                 "movq %%mm0, %%mm7        \n\t"\
00186                 "punpcklwd %%mm2, %%mm0        \n\t" /*A*/\
00187                 "punpckhwd %%mm2, %%mm7        \n\t" /*C*/\
00188                 "movq %%mm1, %%mm2        \n\t"\
00189                 "punpcklwd %%mm3, %%mm1        \n\t" /*B*/\
00190                 "punpckhwd %%mm3, %%mm2        \n\t" /*D*/\
00191                 "movq %%mm0, %%mm3        \n\t"\
00192                 "punpcklwd %%mm1, %%mm0        \n\t" /*A*/\
00193                 "punpckhwd %%mm7, %%mm3        \n\t" /*C*/\
00194                 "punpcklwd %%mm2, %%mm7        \n\t" /*B*/\
00195                 "punpckhwd %%mm2, %%mm1        \n\t" /*D*/\
00196 \
00197                 "movq %%mm0, " #dst0 "        \n\t"\
00198                 "movq %%mm7, " #dst1 "        \n\t"\
00199                 "movq %%mm3, " #dst2 "        \n\t"\
00200                 "movq %%mm1, " #dst3 "        \n\t"
00201 
00202                 "movd %2, %%mm4                \n\t"
00203                 "movd %3, %%mm5                \n\t"
00204                 "movd %4, %%mm6                \n\t"
00205                 "packssdw %%mm4, %%mm4        \n\t"
00206                 "packssdw %%mm5, %%mm5        \n\t"
00207                 "packssdw %%mm6, %%mm6        \n\t"
00208                 "packssdw %%mm4, %%mm4        \n\t"
00209                 "packssdw %%mm5, %%mm5        \n\t"
00210                 "packssdw %%mm6, %%mm6        \n\t"
00211                 REQUANT_CORE(  (%1),  8(%1), 16(%1), 24(%1),  (%0), 8(%0), 64(%0), 72(%0))
00212                 REQUANT_CORE(32(%1), 40(%1), 48(%1), 56(%1),16(%0),24(%0), 48(%0), 56(%0))
00213                 REQUANT_CORE(64(%1), 72(%1), 80(%1), 88(%1),32(%0),40(%0), 96(%0),104(%0))
00214                 REQUANT_CORE(96(%1),104(%1),112(%1),120(%1),80(%0),88(%0),112(%0),120(%0))
00215                 : : "r" (src), "r" (dst), "g" (threshold1+1), "g" (threshold1+5), "g" (threshold1-4) //FIXME maybe more accurate then needed?
00216         );
00217         dst[0]= (src[0] + 4)>>3;
00218 }
00219 
00220 static void softthresh_mmx(DCTELEM dst[64], DCTELEM src[64], int qp, uint8_t *permutation){
00221         int bias= 0; //FIXME
00222         unsigned int threshold1;
00223 
00224         threshold1= qp*((1<<4) - bias) - 1;
00225 
00226         __asm__ volatile(
00227 #undef REQUANT_CORE
00228 #define REQUANT_CORE(dst0, dst1, dst2, dst3, src0, src1, src2, src3) \
00229                 "movq " #src0 ", %%mm0        \n\t"\
00230                 "movq " #src1 ", %%mm1        \n\t"\
00231                 "pxor %%mm6, %%mm6        \n\t"\
00232                 "pxor %%mm7, %%mm7        \n\t"\
00233                 "pcmpgtw %%mm0, %%mm6        \n\t"\
00234                 "pcmpgtw %%mm1, %%mm7        \n\t"\
00235                 "pxor %%mm6, %%mm0        \n\t"\
00236                 "pxor %%mm7, %%mm1        \n\t"\
00237                 "psubusw %%mm4, %%mm0        \n\t"\
00238                 "psubusw %%mm4, %%mm1        \n\t"\
00239                 "pxor %%mm6, %%mm0        \n\t"\
00240                 "pxor %%mm7, %%mm1        \n\t"\
00241                 "movq " #src2 ", %%mm2        \n\t"\
00242                 "movq " #src3 ", %%mm3        \n\t"\
00243                 "pxor %%mm6, %%mm6        \n\t"\
00244                 "pxor %%mm7, %%mm7        \n\t"\
00245                 "pcmpgtw %%mm2, %%mm6        \n\t"\
00246                 "pcmpgtw %%mm3, %%mm7        \n\t"\
00247                 "pxor %%mm6, %%mm2        \n\t"\
00248                 "pxor %%mm7, %%mm3        \n\t"\
00249                 "psubusw %%mm4, %%mm2        \n\t"\
00250                 "psubusw %%mm4, %%mm3        \n\t"\
00251                 "pxor %%mm6, %%mm2        \n\t"\
00252                 "pxor %%mm7, %%mm3        \n\t"\
00253 \
00254                 "paddsw %%mm5, %%mm0        \n\t"\
00255                 "paddsw %%mm5, %%mm1        \n\t"\
00256                 "paddsw %%mm5, %%mm2        \n\t"\
00257                 "paddsw %%mm5, %%mm3        \n\t"\
00258                 "psraw $3, %%mm0        \n\t"\
00259                 "psraw $3, %%mm1        \n\t"\
00260                 "psraw $3, %%mm2        \n\t"\
00261                 "psraw $3, %%mm3        \n\t"\
00262 \
00263                 "movq %%mm0, %%mm7        \n\t"\
00264                 "punpcklwd %%mm2, %%mm0        \n\t" /*A*/\
00265                 "punpckhwd %%mm2, %%mm7        \n\t" /*C*/\
00266                 "movq %%mm1, %%mm2        \n\t"\
00267                 "punpcklwd %%mm3, %%mm1        \n\t" /*B*/\
00268                 "punpckhwd %%mm3, %%mm2        \n\t" /*D*/\
00269                 "movq %%mm0, %%mm3        \n\t"\
00270                 "punpcklwd %%mm1, %%mm0        \n\t" /*A*/\
00271                 "punpckhwd %%mm7, %%mm3        \n\t" /*C*/\
00272                 "punpcklwd %%mm2, %%mm7        \n\t" /*B*/\
00273                 "punpckhwd %%mm2, %%mm1        \n\t" /*D*/\
00274 \
00275                 "movq %%mm0, " #dst0 "        \n\t"\
00276                 "movq %%mm7, " #dst1 "        \n\t"\
00277                 "movq %%mm3, " #dst2 "        \n\t"\
00278                 "movq %%mm1, " #dst3 "        \n\t"
00279 
00280                 "movd %2, %%mm4                \n\t"
00281                 "movd %3, %%mm5                \n\t"
00282                 "packssdw %%mm4, %%mm4        \n\t"
00283                 "packssdw %%mm5, %%mm5        \n\t"
00284                 "packssdw %%mm4, %%mm4        \n\t"
00285                 "packssdw %%mm5, %%mm5        \n\t"
00286                 REQUANT_CORE(  (%1),  8(%1), 16(%1), 24(%1),  (%0), 8(%0), 64(%0), 72(%0))
00287                 REQUANT_CORE(32(%1), 40(%1), 48(%1), 56(%1),16(%0),24(%0), 48(%0), 56(%0))
00288                 REQUANT_CORE(64(%1), 72(%1), 80(%1), 88(%1),32(%0),40(%0), 96(%0),104(%0))
00289                 REQUANT_CORE(96(%1),104(%1),112(%1),120(%1),80(%0),88(%0),112(%0),120(%0))
00290                 : : "r" (src), "r" (dst), "g" (threshold1), "rm" (4) //FIXME maybe more accurate then needed?
00291         );
00292 
00293         dst[0]= (src[0] + 4)>>3;
00294 }
00295 #endif
00296 
00297 static inline void add_block(int16_t *dst, int stride, DCTELEM block[64]){
00298         int y;
00299 
00300         for(y=0; y<8; y++){
00301                 *(uint32_t*)&dst[0 + y*stride]+= *(uint32_t*)&block[0 + y*8];
00302                 *(uint32_t*)&dst[2 + y*stride]+= *(uint32_t*)&block[2 + y*8];
00303                 *(uint32_t*)&dst[4 + y*stride]+= *(uint32_t*)&block[4 + y*8];
00304                 *(uint32_t*)&dst[6 + y*stride]+= *(uint32_t*)&block[6 + y*8];
00305         }
00306 }
00307 
00308 static void store_slice_c(uint8_t *dst, int16_t *src, int dst_stride, int src_stride, int width, int height, int log2_scale){
00309         int y, x;
00310 
00311 #define STORE(pos) \
00312         temp= ((src[x + y*src_stride + pos]<<log2_scale) + d[pos])>>6;\
00313         if(temp & 0x100) temp= ~(temp>>31);\
00314         dst[x + y*dst_stride + pos]= temp;
00315 
00316         for(y=0; y<height; y++){
00317                 const uint8_t *d= dither[y];
00318                 for(x=0; x<width; x+=8){
00319                         int temp;
00320                         STORE(0);
00321                         STORE(1);
00322                         STORE(2);
00323                         STORE(3);
00324                         STORE(4);
00325                         STORE(5);
00326                         STORE(6);
00327                         STORE(7);
00328                 }
00329         }
00330 }
00331 
00332 #if HAVE_MMX
00333 static void store_slice_mmx(uint8_t *dst, int16_t *src, int dst_stride, int src_stride, int width, int height, int log2_scale){
00334         int y;
00335 
00336         for(y=0; y<height; y++){
00337                 uint8_t *dst1= dst;
00338                 int16_t *src1= src;
00339                 __asm__ volatile(
00340                         "movq (%3), %%mm3        \n\t"
00341                         "movq (%3), %%mm4        \n\t"
00342                         "movd %4, %%mm2                \n\t"
00343                         "pxor %%mm0, %%mm0        \n\t"
00344                         "punpcklbw %%mm0, %%mm3        \n\t"
00345                         "punpckhbw %%mm0, %%mm4        \n\t"
00346                         "psraw %%mm2, %%mm3        \n\t"
00347                         "psraw %%mm2, %%mm4        \n\t"
00348                         "movd %5, %%mm2                \n\t"
00349                         "1:                        \n\t"
00350                         "movq (%0), %%mm0        \n\t"
00351                         "movq 8(%0), %%mm1        \n\t"
00352                         "paddw %%mm3, %%mm0        \n\t"
00353                         "paddw %%mm4, %%mm1        \n\t"
00354                         "psraw %%mm2, %%mm0        \n\t"
00355                         "psraw %%mm2, %%mm1        \n\t"
00356                         "packuswb %%mm1, %%mm0        \n\t"
00357                         "movq %%mm0, (%1)         \n\t"
00358                         "add $16, %0                \n\t"
00359                         "add $8, %1                \n\t"
00360                         "cmp %2, %1                \n\t"
00361                         " jb 1b                        \n\t"
00362                         : "+r" (src1), "+r"(dst1)
00363                         : "r"(dst + width), "r"(dither[y]), "g"(log2_scale), "g"(6-log2_scale)
00364                 );
00365                 src += src_stride;
00366                 dst += dst_stride;
00367         }
00368 //        if(width != mmxw)
00369 //                store_slice_c(dst + mmxw, src + mmxw, dst_stride, src_stride, width - mmxw, log2_scale);
00370 }
00371 #endif
00372 
00373 static void (*store_slice)(uint8_t *dst, int16_t *src, int dst_stride, int src_stride, int width, int height, int log2_scale)= store_slice_c;
00374 
00375 static void (*requantize)(DCTELEM dst[64], DCTELEM src[64], int qp, uint8_t *permutation)= hardthresh_c;
00376 
00377 static void filter(struct vf_priv_s *p, uint8_t *dst, uint8_t *src, int dst_stride, int src_stride, int width, int height, uint8_t *qp_store, int qp_stride, int is_luma){
00378         int x, y, i;
00379         const int count= 1<<p->log2_count;
00380         const int stride= is_luma ? p->temp_stride : ((width+16+15)&(~15));
00381         uint64_t __attribute__((aligned(16))) block_align[32];
00382         DCTELEM *block = (DCTELEM *)block_align;
00383         DCTELEM *block2= (DCTELEM *)(block_align+16);
00384 
00385         if (!src || !dst) return; // HACK avoid crash for Y8 colourspace
00386         for(y=0; y<height; y++){
00387                 int index= 8 + 8*stride + y*stride;
00388                 fast_memcpy(p->src + index, src + y*src_stride, width);
00389                 for(x=0; x<8; x++){
00390                         p->src[index         - x - 1]= p->src[index +         x    ];
00391                         p->src[index + width + x    ]= p->src[index + width - x - 1];
00392                 }
00393         }
00394         for(y=0; y<8; y++){
00395                 fast_memcpy(p->src + (      7-y)*stride, p->src + (      y+8)*stride, stride);
00396                 fast_memcpy(p->src + (height+8+y)*stride, p->src + (height-y+7)*stride, stride);
00397         }
00398         //FIXME (try edge emu)
00399 
00400         for(y=0; y<height+8; y+=8){
00401                 memset(p->temp + (8+y)*stride, 0, 8*stride*sizeof(int16_t));
00402                 for(x=0; x<width+8; x+=8){
00403                         const int qps= 3 + is_luma;
00404                         int qp;
00405 
00406                         if(p->qp)
00407                                 qp= p->qp;
00408                         else{
00409                                 qp= qp_store[ (XMIN(x, width-1)>>qps) + (XMIN(y, height-1)>>qps) * qp_stride];
00410                                 qp = FFMAX(1, norm_qscale(qp, p->mpeg2));
00411                         }
00412                         for(i=0; i<count; i++){
00413                                 const int x1= x + offset[i+count-1][0];
00414                                 const int y1= y + offset[i+count-1][1];
00415                                 const int index= x1 + y1*stride;
00416                                 p->dsp.get_pixels(block, p->src + index, stride);
00417                                 p->dsp.fdct(block);
00418                                 requantize(block2, block, qp, p->dsp.idct_permutation);
00419                                 p->dsp.idct(block2);
00420                                 add_block(p->temp + index, stride, block2);
00421                         }
00422                 }
00423                 if(y)
00424                         store_slice(dst + (y-8)*dst_stride, p->temp + 8 + y*stride, dst_stride, stride, width, XMIN(8, height+8-y), 6-p->log2_count);
00425         }
00426 #if 0
00427         for(y=0; y<height; y++){
00428                 for(x=0; x<width; x++){
00429                         if((((x>>6) ^ (y>>6)) & 1) == 0)
00430                                 dst[x + y*dst_stride]= p->src[8 + 8*stride  + x + y*stride];
00431                         if((x&63) == 0 || (y&63)==0)
00432                                 dst[x + y*dst_stride] += 128;
00433                 }
00434         }
00435 #endif
00436         //FIXME reorder for better caching
00437 }
00438 
00439 static int config(struct vf_instance *vf,
00440         int width, int height, int d_width, int d_height,
00441         unsigned int flags, unsigned int outfmt){
00442         int h= (height+16+15)&(~15);
00443 
00444         vf->priv->temp_stride= (width+16+15)&(~15);
00445         vf->priv->temp= malloc(vf->priv->temp_stride*h*sizeof(int16_t));
00446         vf->priv->src = malloc(vf->priv->temp_stride*h*sizeof(uint8_t));
00447 
00448         return vf_next_config(vf,width,height,d_width,d_height,flags,outfmt);
00449 }
00450 
00451 static void get_image(struct vf_instance *vf, mp_image_t *mpi){
00452     if(mpi->flags&MP_IMGFLAG_PRESERVE) return; // don't change
00453     // ok, we can do pp in-place (or pp disabled):
00454     vf->dmpi=vf_get_image(vf->next,mpi->imgfmt,
00455         mpi->type, mpi->flags | MP_IMGFLAG_READABLE, mpi->width, mpi->height);
00456     mpi->planes[0]=vf->dmpi->planes[0];
00457     mpi->stride[0]=vf->dmpi->stride[0];
00458     mpi->width=vf->dmpi->width;
00459     if(mpi->flags&MP_IMGFLAG_PLANAR){
00460         mpi->planes[1]=vf->dmpi->planes[1];
00461         mpi->planes[2]=vf->dmpi->planes[2];
00462         mpi->stride[1]=vf->dmpi->stride[1];
00463         mpi->stride[2]=vf->dmpi->stride[2];
00464     }
00465     mpi->flags|=MP_IMGFLAG_DIRECT;
00466 }
00467 
00468 static int put_image(struct vf_instance *vf, mp_image_t *mpi, double pts){
00469         mp_image_t *dmpi;
00470 
00471         if(!(mpi->flags&MP_IMGFLAG_DIRECT)){
00472                 // no DR, so get a new image! hope we'll get DR buffer:
00473                 dmpi=vf_get_image(vf->next,mpi->imgfmt,
00474                     MP_IMGTYPE_TEMP,
00475                     MP_IMGFLAG_ACCEPT_STRIDE|MP_IMGFLAG_PREFER_ALIGNED_STRIDE,
00476                     mpi->width,mpi->height);
00477                 vf_clone_mpi_attributes(dmpi, mpi);
00478         }else{
00479            dmpi=vf->dmpi;
00480         }
00481 
00482         vf->priv->mpeg2= mpi->qscale_type;
00483         if(mpi->pict_type != 3 && mpi->qscale && !vf->priv->qp){
00484             int w = mpi->qstride;
00485             int h = (mpi->h + 15) >> 4;
00486             if (!w) {
00487                 w = (mpi->w + 15) >> 4;
00488                 h = 1;
00489             }
00490             if(!vf->priv->non_b_qp)
00491                 vf->priv->non_b_qp= malloc(w*h);
00492             fast_memcpy(vf->priv->non_b_qp, mpi->qscale, w*h);
00493         }
00494         if(vf->priv->log2_count || !(mpi->flags&MP_IMGFLAG_DIRECT)){
00495             char *qp_tab= vf->priv->non_b_qp;
00496             if((vf->priv->mode&4) || !qp_tab)
00497                 qp_tab= mpi->qscale;
00498 
00499             if(qp_tab || vf->priv->qp){
00500                 filter(vf->priv, dmpi->planes[0], mpi->planes[0], dmpi->stride[0], mpi->stride[0], mpi->w, mpi->h, qp_tab, mpi->qstride, 1);
00501                 filter(vf->priv, dmpi->planes[1], mpi->planes[1], dmpi->stride[1], mpi->stride[1], mpi->w>>mpi->chroma_x_shift, mpi->h>>mpi->chroma_y_shift, qp_tab, mpi->qstride, 0);
00502                 filter(vf->priv, dmpi->planes[2], mpi->planes[2], dmpi->stride[2], mpi->stride[2], mpi->w>>mpi->chroma_x_shift, mpi->h>>mpi->chroma_y_shift, qp_tab, mpi->qstride, 0);
00503             }else{
00504                 memcpy_pic(dmpi->planes[0], mpi->planes[0], mpi->w, mpi->h, dmpi->stride[0], mpi->stride[0]);
00505                 memcpy_pic(dmpi->planes[1], mpi->planes[1], mpi->w>>mpi->chroma_x_shift, mpi->h>>mpi->chroma_y_shift, dmpi->stride[1], mpi->stride[1]);
00506                 memcpy_pic(dmpi->planes[2], mpi->planes[2], mpi->w>>mpi->chroma_x_shift, mpi->h>>mpi->chroma_y_shift, dmpi->stride[2], mpi->stride[2]);
00507             }
00508         }
00509 
00510 #if HAVE_MMX
00511         if(gCpuCaps.hasMMX) __asm__ volatile ("emms\n\t");
00512 #endif
00513 #if HAVE_MMX2
00514         if(gCpuCaps.hasMMX2) __asm__ volatile ("sfence\n\t");
00515 #endif
00516 
00517         return vf_next_put_image(vf,dmpi, pts);
00518 }
00519 
00520 static void uninit(struct vf_instance *vf){
00521         if(!vf->priv) return;
00522 
00523         free(vf->priv->temp);
00524         vf->priv->temp= NULL;
00525         free(vf->priv->src);
00526         vf->priv->src= NULL;
00527         free(vf->priv->avctx);
00528         vf->priv->avctx= NULL;
00529         free(vf->priv->non_b_qp);
00530         vf->priv->non_b_qp= NULL;
00531 
00532         free(vf->priv);
00533         vf->priv=NULL;
00534 }
00535 
00536 //===========================================================================//
00537 static int query_format(struct vf_instance *vf, unsigned int fmt){
00538     switch(fmt){
00539         case IMGFMT_YVU9:
00540         case IMGFMT_IF09:
00541         case IMGFMT_YV12:
00542         case IMGFMT_I420:
00543         case IMGFMT_IYUV:
00544         case IMGFMT_CLPL:
00545         case IMGFMT_Y800:
00546         case IMGFMT_Y8:
00547         case IMGFMT_444P:
00548         case IMGFMT_422P:
00549         case IMGFMT_411P:
00550             return vf_next_query_format(vf,fmt);
00551     }
00552     return 0;
00553 }
00554 
00555 static int control(struct vf_instance *vf, int request, void* data){
00556     switch(request){
00557     case VFCTRL_QUERY_MAX_PP_LEVEL:
00558         return 6;
00559     case VFCTRL_SET_PP_LEVEL:
00560         vf->priv->log2_count= *((unsigned int*)data);
00561         return CONTROL_TRUE;
00562     }
00563     return vf_next_control(vf,request,data);
00564 }
00565 
00566 static int vf_open(vf_instance_t *vf, char *args){
00567 
00568     int log2c=-1;
00569 
00570     vf->config=config;
00571     vf->put_image=put_image;
00572     vf->get_image=get_image;
00573     vf->query_format=query_format;
00574     vf->uninit=uninit;
00575     vf->control= control;
00576     vf->priv=malloc(sizeof(struct vf_priv_s));
00577     memset(vf->priv, 0, sizeof(struct vf_priv_s));
00578 
00579     init_avcodec();
00580 
00581     vf->priv->avctx= avcodec_alloc_context();
00582     dsputil_init(&vf->priv->dsp, vf->priv->avctx);
00583 
00584     vf->priv->log2_count= 3;
00585 
00586     if (args) sscanf(args, "%d:%d:%d", &log2c, &vf->priv->qp, &vf->priv->mode);
00587 
00588     if( log2c >=0 && log2c <=6 )
00589         vf->priv->log2_count = log2c;
00590 
00591     if(vf->priv->qp < 0)
00592         vf->priv->qp = 0;
00593 
00594     switch(vf->priv->mode&3){
00595         default:
00596         case 0: requantize= hardthresh_c; break;
00597         case 1: requantize= softthresh_c; break;
00598     }
00599 
00600 #if HAVE_MMX
00601     if(gCpuCaps.hasMMX){
00602         store_slice= store_slice_mmx;
00603         switch(vf->priv->mode&3){
00604             case 0: requantize= hardthresh_mmx; break;
00605             case 1: requantize= softthresh_mmx; break;
00606         }
00607     }
00608 #endif
00609 
00610     return 1;
00611 }
00612 
00613 const vf_info_t vf_info_spp = {
00614     "simple postprocess",
00615     "spp",
00616     "Michael Niedermayer",
00617     "",
00618     vf_open,
00619     NULL
00620 };
Generated on Fri Feb 1 2013 14:34:49 for FFmpeg by doxygen 1.7.1