00001
00002
00003
00004
00005
00006
00007
00008
00009
00010
00011
00012
00013
00014
00015
00016
00017
00018
00019
00020
00021
00028 #include "avcodec.h"
00029 #include "mpegvideo.h"
00030 #include "h264pred.h"
00031
00032 static void pred4x4_vertical_c(uint8_t *src, uint8_t *topright, int stride){
00033 const uint32_t a= ((uint32_t*)(src-stride))[0];
00034 ((uint32_t*)(src+0*stride))[0]= a;
00035 ((uint32_t*)(src+1*stride))[0]= a;
00036 ((uint32_t*)(src+2*stride))[0]= a;
00037 ((uint32_t*)(src+3*stride))[0]= a;
00038 }
00039
00040 static void pred4x4_horizontal_c(uint8_t *src, uint8_t *topright, int stride){
00041 ((uint32_t*)(src+0*stride))[0]= src[-1+0*stride]*0x01010101;
00042 ((uint32_t*)(src+1*stride))[0]= src[-1+1*stride]*0x01010101;
00043 ((uint32_t*)(src+2*stride))[0]= src[-1+2*stride]*0x01010101;
00044 ((uint32_t*)(src+3*stride))[0]= src[-1+3*stride]*0x01010101;
00045 }
00046
00047 static void pred4x4_dc_c(uint8_t *src, uint8_t *topright, int stride){
00048 const int dc= ( src[-stride] + src[1-stride] + src[2-stride] + src[3-stride]
00049 + src[-1+0*stride] + src[-1+1*stride] + src[-1+2*stride] + src[-1+3*stride] + 4) >>3;
00050
00051 ((uint32_t*)(src+0*stride))[0]=
00052 ((uint32_t*)(src+1*stride))[0]=
00053 ((uint32_t*)(src+2*stride))[0]=
00054 ((uint32_t*)(src+3*stride))[0]= dc* 0x01010101;
00055 }
00056
00057 static void pred4x4_left_dc_c(uint8_t *src, uint8_t *topright, int stride){
00058 const int dc= ( src[-1+0*stride] + src[-1+1*stride] + src[-1+2*stride] + src[-1+3*stride] + 2) >>2;
00059
00060 ((uint32_t*)(src+0*stride))[0]=
00061 ((uint32_t*)(src+1*stride))[0]=
00062 ((uint32_t*)(src+2*stride))[0]=
00063 ((uint32_t*)(src+3*stride))[0]= dc* 0x01010101;
00064 }
00065
00066 static void pred4x4_top_dc_c(uint8_t *src, uint8_t *topright, int stride){
00067 const int dc= ( src[-stride] + src[1-stride] + src[2-stride] + src[3-stride] + 2) >>2;
00068
00069 ((uint32_t*)(src+0*stride))[0]=
00070 ((uint32_t*)(src+1*stride))[0]=
00071 ((uint32_t*)(src+2*stride))[0]=
00072 ((uint32_t*)(src+3*stride))[0]= dc* 0x01010101;
00073 }
00074
00075 static void pred4x4_128_dc_c(uint8_t *src, uint8_t *topright, int stride){
00076 ((uint32_t*)(src+0*stride))[0]=
00077 ((uint32_t*)(src+1*stride))[0]=
00078 ((uint32_t*)(src+2*stride))[0]=
00079 ((uint32_t*)(src+3*stride))[0]= 128U*0x01010101U;
00080 }
00081
00082
00083 #define LOAD_TOP_RIGHT_EDGE\
00084 const int av_unused t4= topright[0];\
00085 const int av_unused t5= topright[1];\
00086 const int av_unused t6= topright[2];\
00087 const int av_unused t7= topright[3];\
00088
00089 #define LOAD_DOWN_LEFT_EDGE\
00090 const int av_unused l4= src[-1+4*stride];\
00091 const int av_unused l5= src[-1+5*stride];\
00092 const int av_unused l6= src[-1+6*stride];\
00093 const int av_unused l7= src[-1+7*stride];\
00094
00095 #define LOAD_LEFT_EDGE\
00096 const int av_unused l0= src[-1+0*stride];\
00097 const int av_unused l1= src[-1+1*stride];\
00098 const int av_unused l2= src[-1+2*stride];\
00099 const int av_unused l3= src[-1+3*stride];\
00100
00101 #define LOAD_TOP_EDGE\
00102 const int av_unused t0= src[ 0-1*stride];\
00103 const int av_unused t1= src[ 1-1*stride];\
00104 const int av_unused t2= src[ 2-1*stride];\
00105 const int av_unused t3= src[ 3-1*stride];\
00106
00107 static void pred4x4_down_right_c(uint8_t *src, uint8_t *topright, int stride){
00108 const int lt= src[-1-1*stride];
00109 LOAD_TOP_EDGE
00110 LOAD_LEFT_EDGE
00111
00112 src[0+3*stride]=(l3 + 2*l2 + l1 + 2)>>2;
00113 src[0+2*stride]=
00114 src[1+3*stride]=(l2 + 2*l1 + l0 + 2)>>2;
00115 src[0+1*stride]=
00116 src[1+2*stride]=
00117 src[2+3*stride]=(l1 + 2*l0 + lt + 2)>>2;
00118 src[0+0*stride]=
00119 src[1+1*stride]=
00120 src[2+2*stride]=
00121 src[3+3*stride]=(l0 + 2*lt + t0 + 2)>>2;
00122 src[1+0*stride]=
00123 src[2+1*stride]=
00124 src[3+2*stride]=(lt + 2*t0 + t1 + 2)>>2;
00125 src[2+0*stride]=
00126 src[3+1*stride]=(t0 + 2*t1 + t2 + 2)>>2;
00127 src[3+0*stride]=(t1 + 2*t2 + t3 + 2)>>2;
00128 }
00129
00130 static void pred4x4_down_left_c(uint8_t *src, uint8_t *topright, int stride){
00131 LOAD_TOP_EDGE
00132 LOAD_TOP_RIGHT_EDGE
00133
00134
00135 src[0+0*stride]=(t0 + t2 + 2*t1 + 2)>>2;
00136 src[1+0*stride]=
00137 src[0+1*stride]=(t1 + t3 + 2*t2 + 2)>>2;
00138 src[2+0*stride]=
00139 src[1+1*stride]=
00140 src[0+2*stride]=(t2 + t4 + 2*t3 + 2)>>2;
00141 src[3+0*stride]=
00142 src[2+1*stride]=
00143 src[1+2*stride]=
00144 src[0+3*stride]=(t3 + t5 + 2*t4 + 2)>>2;
00145 src[3+1*stride]=
00146 src[2+2*stride]=
00147 src[1+3*stride]=(t4 + t6 + 2*t5 + 2)>>2;
00148 src[3+2*stride]=
00149 src[2+3*stride]=(t5 + t7 + 2*t6 + 2)>>2;
00150 src[3+3*stride]=(t6 + 3*t7 + 2)>>2;
00151 }
00152
00153 static void pred4x4_down_left_svq3_c(uint8_t *src, uint8_t *topright, int stride){
00154 LOAD_TOP_EDGE
00155 LOAD_LEFT_EDGE
00156 const av_unused int unu0= t0;
00157 const av_unused int unu1= l0;
00158
00159 src[0+0*stride]=(l1 + t1)>>1;
00160 src[1+0*stride]=
00161 src[0+1*stride]=(l2 + t2)>>1;
00162 src[2+0*stride]=
00163 src[1+1*stride]=
00164 src[0+2*stride]=
00165 src[3+0*stride]=
00166 src[2+1*stride]=
00167 src[1+2*stride]=
00168 src[0+3*stride]=
00169 src[3+1*stride]=
00170 src[2+2*stride]=
00171 src[1+3*stride]=
00172 src[3+2*stride]=
00173 src[2+3*stride]=
00174 src[3+3*stride]=(l3 + t3)>>1;
00175 }
00176
00177 static void pred4x4_down_left_rv40_c(uint8_t *src, uint8_t *topright, int stride){
00178 LOAD_TOP_EDGE
00179 LOAD_TOP_RIGHT_EDGE
00180 LOAD_LEFT_EDGE
00181 LOAD_DOWN_LEFT_EDGE
00182
00183 src[0+0*stride]=(t0 + t2 + 2*t1 + 2 + l0 + l2 + 2*l1 + 2)>>3;
00184 src[1+0*stride]=
00185 src[0+1*stride]=(t1 + t3 + 2*t2 + 2 + l1 + l3 + 2*l2 + 2)>>3;
00186 src[2+0*stride]=
00187 src[1+1*stride]=
00188 src[0+2*stride]=(t2 + t4 + 2*t3 + 2 + l2 + l4 + 2*l3 + 2)>>3;
00189 src[3+0*stride]=
00190 src[2+1*stride]=
00191 src[1+2*stride]=
00192 src[0+3*stride]=(t3 + t5 + 2*t4 + 2 + l3 + l5 + 2*l4 + 2)>>3;
00193 src[3+1*stride]=
00194 src[2+2*stride]=
00195 src[1+3*stride]=(t4 + t6 + 2*t5 + 2 + l4 + l6 + 2*l5 + 2)>>3;
00196 src[3+2*stride]=
00197 src[2+3*stride]=(t5 + t7 + 2*t6 + 2 + l5 + l7 + 2*l6 + 2)>>3;
00198 src[3+3*stride]=(t6 + t7 + 1 + l6 + l7 + 1)>>2;
00199 }
00200
00201 static void pred4x4_down_left_rv40_notop_c(uint8_t *src, uint8_t *topright, int stride){
00202 LOAD_LEFT_EDGE
00203 LOAD_DOWN_LEFT_EDGE
00204
00205 src[0+0*stride]=(l0 + l2 + 2*l1 + 2)>>2;
00206 src[1+0*stride]=
00207 src[0+1*stride]=(l1 + l3 + 2*l2 + 2)>>2;
00208 src[2+0*stride]=
00209 src[1+1*stride]=
00210 src[0+2*stride]=(l2 + l4 + 2*l3 + 2)>>2;
00211 src[3+0*stride]=
00212 src[2+1*stride]=
00213 src[1+2*stride]=
00214 src[0+3*stride]=(l3 + l5 + 2*l4 + 2)>>2;
00215 src[3+1*stride]=
00216 src[2+2*stride]=
00217 src[1+3*stride]=(l4 + l6 + 2*l5 + 2)>>2;
00218 src[3+2*stride]=
00219 src[2+3*stride]=(l5 + l7 + 2*l6 + 2)>>2;
00220 src[3+3*stride]=(l6 + l7 + 1)>>1;
00221 }
00222
00223 static void pred4x4_down_left_rv40_nodown_c(uint8_t *src, uint8_t *topright, int stride){
00224 LOAD_TOP_EDGE
00225 LOAD_TOP_RIGHT_EDGE
00226 LOAD_LEFT_EDGE
00227
00228 src[0+0*stride]=(t0 + t2 + 2*t1 + 2 + l0 + l2 + 2*l1 + 2)>>3;
00229 src[1+0*stride]=
00230 src[0+1*stride]=(t1 + t3 + 2*t2 + 2 + l1 + l3 + 2*l2 + 2)>>3;
00231 src[2+0*stride]=
00232 src[1+1*stride]=
00233 src[0+2*stride]=(t2 + t4 + 2*t3 + 2 + l2 + 3*l3 + 2)>>3;
00234 src[3+0*stride]=
00235 src[2+1*stride]=
00236 src[1+2*stride]=
00237 src[0+3*stride]=(t3 + t5 + 2*t4 + 2 + l3*4 + 2)>>3;
00238 src[3+1*stride]=
00239 src[2+2*stride]=
00240 src[1+3*stride]=(t4 + t6 + 2*t5 + 2 + l3*4 + 2)>>3;
00241 src[3+2*stride]=
00242 src[2+3*stride]=(t5 + t7 + 2*t6 + 2 + l3*4 + 2)>>3;
00243 src[3+3*stride]=(t6 + t7 + 1 + 2*l3 + 1)>>2;
00244 }
00245
00246 static void pred4x4_vertical_right_c(uint8_t *src, uint8_t *topright, int stride){
00247 const int lt= src[-1-1*stride];
00248 LOAD_TOP_EDGE
00249 LOAD_LEFT_EDGE
00250
00251 src[0+0*stride]=
00252 src[1+2*stride]=(lt + t0 + 1)>>1;
00253 src[1+0*stride]=
00254 src[2+2*stride]=(t0 + t1 + 1)>>1;
00255 src[2+0*stride]=
00256 src[3+2*stride]=(t1 + t2 + 1)>>1;
00257 src[3+0*stride]=(t2 + t3 + 1)>>1;
00258 src[0+1*stride]=
00259 src[1+3*stride]=(l0 + 2*lt + t0 + 2)>>2;
00260 src[1+1*stride]=
00261 src[2+3*stride]=(lt + 2*t0 + t1 + 2)>>2;
00262 src[2+1*stride]=
00263 src[3+3*stride]=(t0 + 2*t1 + t2 + 2)>>2;
00264 src[3+1*stride]=(t1 + 2*t2 + t3 + 2)>>2;
00265 src[0+2*stride]=(lt + 2*l0 + l1 + 2)>>2;
00266 src[0+3*stride]=(l0 + 2*l1 + l2 + 2)>>2;
00267 }
00268
00269 static void pred4x4_vertical_left_c(uint8_t *src, uint8_t *topright, int stride){
00270 LOAD_TOP_EDGE
00271 LOAD_TOP_RIGHT_EDGE
00272
00273 src[0+0*stride]=(t0 + t1 + 1)>>1;
00274 src[1+0*stride]=
00275 src[0+2*stride]=(t1 + t2 + 1)>>1;
00276 src[2+0*stride]=
00277 src[1+2*stride]=(t2 + t3 + 1)>>1;
00278 src[3+0*stride]=
00279 src[2+2*stride]=(t3 + t4+ 1)>>1;
00280 src[3+2*stride]=(t4 + t5+ 1)>>1;
00281 src[0+1*stride]=(t0 + 2*t1 + t2 + 2)>>2;
00282 src[1+1*stride]=
00283 src[0+3*stride]=(t1 + 2*t2 + t3 + 2)>>2;
00284 src[2+1*stride]=
00285 src[1+3*stride]=(t2 + 2*t3 + t4 + 2)>>2;
00286 src[3+1*stride]=
00287 src[2+3*stride]=(t3 + 2*t4 + t5 + 2)>>2;
00288 src[3+3*stride]=(t4 + 2*t5 + t6 + 2)>>2;
00289 }
00290
00291 static void pred4x4_vertical_left_rv40_c(uint8_t *src, uint8_t *topright, int stride){
00292 LOAD_TOP_EDGE
00293 LOAD_TOP_RIGHT_EDGE
00294 LOAD_LEFT_EDGE
00295
00296 src[0+0*stride]=(2*t0 + 2*t1 + l1 + 2*l2 + l3 + 4)>>3;
00297 src[1+0*stride]=
00298 src[0+2*stride]=(t1 + t2 + 1)>>1;
00299 src[2+0*stride]=
00300 src[1+2*stride]=(t2 + t3 + 1)>>1;
00301 src[3+0*stride]=
00302 src[2+2*stride]=(t3 + t4+ 1)>>1;
00303 src[3+2*stride]=(t4 + t5+ 1)>>1;
00304 src[0+1*stride]=(t0 + 2*t1 + t2 + l2 + 2*l3 + l3 + 4)>>3;
00305 src[1+1*stride]=
00306 src[0+3*stride]=(t1 + 2*t2 + t3 + 2)>>2;
00307 src[2+1*stride]=
00308 src[1+3*stride]=(t2 + 2*t3 + t4 + 2)>>2;
00309 src[3+1*stride]=
00310 src[2+3*stride]=(t3 + 2*t4 + t5 + 2)>>2;
00311 src[3+3*stride]=(t4 + 2*t5 + t6 + 2)>>2;
00312 }
00313
00314 static void pred4x4_horizontal_up_c(uint8_t *src, uint8_t *topright, int stride){
00315 LOAD_LEFT_EDGE
00316
00317 src[0+0*stride]=(l0 + l1 + 1)>>1;
00318 src[1+0*stride]=(l0 + 2*l1 + l2 + 2)>>2;
00319 src[2+0*stride]=
00320 src[0+1*stride]=(l1 + l2 + 1)>>1;
00321 src[3+0*stride]=
00322 src[1+1*stride]=(l1 + 2*l2 + l3 + 2)>>2;
00323 src[2+1*stride]=
00324 src[0+2*stride]=(l2 + l3 + 1)>>1;
00325 src[3+1*stride]=
00326 src[1+2*stride]=(l2 + 2*l3 + l3 + 2)>>2;
00327 src[3+2*stride]=
00328 src[1+3*stride]=
00329 src[0+3*stride]=
00330 src[2+2*stride]=
00331 src[2+3*stride]=
00332 src[3+3*stride]=l3;
00333 }
00334
00335 static void pred4x4_horizontal_up_rv40_c(uint8_t *src, uint8_t *topright, int stride){
00336 LOAD_LEFT_EDGE
00337 LOAD_DOWN_LEFT_EDGE
00338 LOAD_TOP_EDGE
00339 LOAD_TOP_RIGHT_EDGE
00340
00341 src[0+0*stride]=(t1 + 2*t2 + t3 + 2*l0 + 2*l1 + 4)>>3;
00342 src[1+0*stride]=(t2 + 2*t3 + t4 + l0 + 2*l1 + l2 + 4)>>3;
00343 src[2+0*stride]=
00344 src[0+1*stride]=(t3 + 2*t4 + t5 + 2*l1 + 2*l2 + 4)>>3;
00345 src[3+0*stride]=
00346 src[1+1*stride]=(t4 + 2*t5 + t6 + l1 + 2*l2 + l3 + 4)>>3;
00347 src[2+1*stride]=
00348 src[0+2*stride]=(t5 + 2*t6 + t7 + 2*l2 + 2*l3 + 4)>>3;
00349 src[3+1*stride]=
00350 src[1+2*stride]=(t6 + 3*t7 + l2 + 3*l3 + 4)>>3;
00351 src[3+2*stride]=
00352 src[1+3*stride]=(l3 + 2*l4 + l5 + 2)>>2;
00353 src[0+3*stride]=
00354 src[2+2*stride]=(t6 + t7 + l3 + l4 + 2)>>2;
00355 src[2+3*stride]=(l4 + l5 + 1)>>1;
00356 src[3+3*stride]=(l4 + 2*l5 + l6 + 2)>>2;
00357 }
00358
00359 static void pred4x4_horizontal_up_rv40_nodown_c(uint8_t *src, uint8_t *topright, int stride){
00360 LOAD_LEFT_EDGE
00361 LOAD_TOP_EDGE
00362 LOAD_TOP_RIGHT_EDGE
00363
00364 src[0+0*stride]=(t1 + 2*t2 + t3 + 2*l0 + 2*l1 + 4)>>3;
00365 src[1+0*stride]=(t2 + 2*t3 + t4 + l0 + 2*l1 + l2 + 4)>>3;
00366 src[2+0*stride]=
00367 src[0+1*stride]=(t3 + 2*t4 + t5 + 2*l1 + 2*l2 + 4)>>3;
00368 src[3+0*stride]=
00369 src[1+1*stride]=(t4 + 2*t5 + t6 + l1 + 2*l2 + l3 + 4)>>3;
00370 src[2+1*stride]=
00371 src[0+2*stride]=(t5 + 2*t6 + t7 + 2*l2 + 2*l3 + 4)>>3;
00372 src[3+1*stride]=
00373 src[1+2*stride]=(t6 + 3*t7 + l2 + 3*l3 + 4)>>3;
00374 src[3+2*stride]=
00375 src[1+3*stride]=l3;
00376 src[0+3*stride]=
00377 src[2+2*stride]=(t6 + t7 + 2*l3 + 2)>>2;
00378 src[2+3*stride]=
00379 src[3+3*stride]=l3;
00380 }
00381
00382 static void pred4x4_horizontal_down_c(uint8_t *src, uint8_t *topright, int stride){
00383 const int lt= src[-1-1*stride];
00384 LOAD_TOP_EDGE
00385 LOAD_LEFT_EDGE
00386
00387 src[0+0*stride]=
00388 src[2+1*stride]=(lt + l0 + 1)>>1;
00389 src[1+0*stride]=
00390 src[3+1*stride]=(l0 + 2*lt + t0 + 2)>>2;
00391 src[2+0*stride]=(lt + 2*t0 + t1 + 2)>>2;
00392 src[3+0*stride]=(t0 + 2*t1 + t2 + 2)>>2;
00393 src[0+1*stride]=
00394 src[2+2*stride]=(l0 + l1 + 1)>>1;
00395 src[1+1*stride]=
00396 src[3+2*stride]=(lt + 2*l0 + l1 + 2)>>2;
00397 src[0+2*stride]=
00398 src[2+3*stride]=(l1 + l2+ 1)>>1;
00399 src[1+2*stride]=
00400 src[3+3*stride]=(l0 + 2*l1 + l2 + 2)>>2;
00401 src[0+3*stride]=(l2 + l3 + 1)>>1;
00402 src[1+3*stride]=(l1 + 2*l2 + l3 + 2)>>2;
00403 }
00404
00405 static void pred16x16_vertical_c(uint8_t *src, int stride){
00406 int i;
00407 const uint32_t a= ((uint32_t*)(src-stride))[0];
00408 const uint32_t b= ((uint32_t*)(src-stride))[1];
00409 const uint32_t c= ((uint32_t*)(src-stride))[2];
00410 const uint32_t d= ((uint32_t*)(src-stride))[3];
00411
00412 for(i=0; i<16; i++){
00413 ((uint32_t*)(src+i*stride))[0]= a;
00414 ((uint32_t*)(src+i*stride))[1]= b;
00415 ((uint32_t*)(src+i*stride))[2]= c;
00416 ((uint32_t*)(src+i*stride))[3]= d;
00417 }
00418 }
00419
00420 static void pred16x16_horizontal_c(uint8_t *src, int stride){
00421 int i;
00422
00423 for(i=0; i<16; i++){
00424 ((uint32_t*)(src+i*stride))[0]=
00425 ((uint32_t*)(src+i*stride))[1]=
00426 ((uint32_t*)(src+i*stride))[2]=
00427 ((uint32_t*)(src+i*stride))[3]= src[-1+i*stride]*0x01010101;
00428 }
00429 }
00430
00431 static void pred16x16_dc_c(uint8_t *src, int stride){
00432 int i, dc=0;
00433
00434 for(i=0;i<16; i++){
00435 dc+= src[-1+i*stride];
00436 }
00437
00438 for(i=0;i<16; i++){
00439 dc+= src[i-stride];
00440 }
00441
00442 dc= 0x01010101*((dc + 16)>>5);
00443
00444 for(i=0; i<16; i++){
00445 ((uint32_t*)(src+i*stride))[0]=
00446 ((uint32_t*)(src+i*stride))[1]=
00447 ((uint32_t*)(src+i*stride))[2]=
00448 ((uint32_t*)(src+i*stride))[3]= dc;
00449 }
00450 }
00451
00452 static void pred16x16_left_dc_c(uint8_t *src, int stride){
00453 int i, dc=0;
00454
00455 for(i=0;i<16; i++){
00456 dc+= src[-1+i*stride];
00457 }
00458
00459 dc= 0x01010101*((dc + 8)>>4);
00460
00461 for(i=0; i<16; i++){
00462 ((uint32_t*)(src+i*stride))[0]=
00463 ((uint32_t*)(src+i*stride))[1]=
00464 ((uint32_t*)(src+i*stride))[2]=
00465 ((uint32_t*)(src+i*stride))[3]= dc;
00466 }
00467 }
00468
00469 static void pred16x16_top_dc_c(uint8_t *src, int stride){
00470 int i, dc=0;
00471
00472 for(i=0;i<16; i++){
00473 dc+= src[i-stride];
00474 }
00475 dc= 0x01010101*((dc + 8)>>4);
00476
00477 for(i=0; i<16; i++){
00478 ((uint32_t*)(src+i*stride))[0]=
00479 ((uint32_t*)(src+i*stride))[1]=
00480 ((uint32_t*)(src+i*stride))[2]=
00481 ((uint32_t*)(src+i*stride))[3]= dc;
00482 }
00483 }
00484
00485 static void pred16x16_128_dc_c(uint8_t *src, int stride){
00486 int i;
00487
00488 for(i=0; i<16; i++){
00489 ((uint32_t*)(src+i*stride))[0]=
00490 ((uint32_t*)(src+i*stride))[1]=
00491 ((uint32_t*)(src+i*stride))[2]=
00492 ((uint32_t*)(src+i*stride))[3]= 0x01010101U*128U;
00493 }
00494 }
00495
00496 static inline void pred16x16_plane_compat_c(uint8_t *src, int stride, const int svq3, const int rv40){
00497 int i, j, k;
00498 int a;
00499 uint8_t *cm = ff_cropTbl + MAX_NEG_CROP;
00500 const uint8_t * const src0 = src+7-stride;
00501 const uint8_t *src1 = src+8*stride-1;
00502 const uint8_t *src2 = src1-2*stride;
00503 int H = src0[1] - src0[-1];
00504 int V = src1[0] - src2[ 0];
00505 for(k=2; k<=8; ++k) {
00506 src1 += stride; src2 -= stride;
00507 H += k*(src0[k] - src0[-k]);
00508 V += k*(src1[0] - src2[ 0]);
00509 }
00510 if(svq3){
00511 H = ( 5*(H/4) ) / 16;
00512 V = ( 5*(V/4) ) / 16;
00513
00514
00515 i = H; H = V; V = i;
00516 }else if(rv40){
00517 H = ( H + (H>>2) ) >> 4;
00518 V = ( V + (V>>2) ) >> 4;
00519 }else{
00520 H = ( 5*H+32 ) >> 6;
00521 V = ( 5*V+32 ) >> 6;
00522 }
00523
00524 a = 16*(src1[0] + src2[16] + 1) - 7*(V+H);
00525 for(j=16; j>0; --j) {
00526 int b = a;
00527 a += V;
00528 for(i=-16; i<0; i+=4) {
00529 src[16+i] = cm[ (b ) >> 5 ];
00530 src[17+i] = cm[ (b+ H) >> 5 ];
00531 src[18+i] = cm[ (b+2*H) >> 5 ];
00532 src[19+i] = cm[ (b+3*H) >> 5 ];
00533 b += 4*H;
00534 }
00535 src += stride;
00536 }
00537 }
00538
00539 static void pred16x16_plane_c(uint8_t *src, int stride){
00540 pred16x16_plane_compat_c(src, stride, 0, 0);
00541 }
00542
00543 static void pred16x16_plane_svq3_c(uint8_t *src, int stride){
00544 pred16x16_plane_compat_c(src, stride, 1, 0);
00545 }
00546
00547 static void pred16x16_plane_rv40_c(uint8_t *src, int stride){
00548 pred16x16_plane_compat_c(src, stride, 0, 1);
00549 }
00550
00551 static void pred8x8_vertical_c(uint8_t *src, int stride){
00552 int i;
00553 const uint32_t a= ((uint32_t*)(src-stride))[0];
00554 const uint32_t b= ((uint32_t*)(src-stride))[1];
00555
00556 for(i=0; i<8; i++){
00557 ((uint32_t*)(src+i*stride))[0]= a;
00558 ((uint32_t*)(src+i*stride))[1]= b;
00559 }
00560 }
00561
00562 static void pred8x8_horizontal_c(uint8_t *src, int stride){
00563 int i;
00564
00565 for(i=0; i<8; i++){
00566 ((uint32_t*)(src+i*stride))[0]=
00567 ((uint32_t*)(src+i*stride))[1]= src[-1+i*stride]*0x01010101;
00568 }
00569 }
00570
00571 static void pred8x8_128_dc_c(uint8_t *src, int stride){
00572 int i;
00573
00574 for(i=0; i<8; i++){
00575 ((uint32_t*)(src+i*stride))[0]=
00576 ((uint32_t*)(src+i*stride))[1]= 0x01010101U*128U;
00577 }
00578 }
00579
00580 static void pred8x8_left_dc_c(uint8_t *src, int stride){
00581 int i;
00582 int dc0, dc2;
00583
00584 dc0=dc2=0;
00585 for(i=0;i<4; i++){
00586 dc0+= src[-1+i*stride];
00587 dc2+= src[-1+(i+4)*stride];
00588 }
00589 dc0= 0x01010101*((dc0 + 2)>>2);
00590 dc2= 0x01010101*((dc2 + 2)>>2);
00591
00592 for(i=0; i<4; i++){
00593 ((uint32_t*)(src+i*stride))[0]=
00594 ((uint32_t*)(src+i*stride))[1]= dc0;
00595 }
00596 for(i=4; i<8; i++){
00597 ((uint32_t*)(src+i*stride))[0]=
00598 ((uint32_t*)(src+i*stride))[1]= dc2;
00599 }
00600 }
00601
00602 static void pred8x8_left_dc_rv40_c(uint8_t *src, int stride){
00603 int i;
00604 int dc0;
00605
00606 dc0=0;
00607 for(i=0;i<8; i++)
00608 dc0+= src[-1+i*stride];
00609 dc0= 0x01010101*((dc0 + 4)>>3);
00610
00611 for(i=0; i<8; i++){
00612 ((uint32_t*)(src+i*stride))[0]=
00613 ((uint32_t*)(src+i*stride))[1]= dc0;
00614 }
00615 }
00616
00617 static void pred8x8_top_dc_c(uint8_t *src, int stride){
00618 int i;
00619 int dc0, dc1;
00620
00621 dc0=dc1=0;
00622 for(i=0;i<4; i++){
00623 dc0+= src[i-stride];
00624 dc1+= src[4+i-stride];
00625 }
00626 dc0= 0x01010101*((dc0 + 2)>>2);
00627 dc1= 0x01010101*((dc1 + 2)>>2);
00628
00629 for(i=0; i<4; i++){
00630 ((uint32_t*)(src+i*stride))[0]= dc0;
00631 ((uint32_t*)(src+i*stride))[1]= dc1;
00632 }
00633 for(i=4; i<8; i++){
00634 ((uint32_t*)(src+i*stride))[0]= dc0;
00635 ((uint32_t*)(src+i*stride))[1]= dc1;
00636 }
00637 }
00638
00639 static void pred8x8_top_dc_rv40_c(uint8_t *src, int stride){
00640 int i;
00641 int dc0;
00642
00643 dc0=0;
00644 for(i=0;i<8; i++)
00645 dc0+= src[i-stride];
00646 dc0= 0x01010101*((dc0 + 4)>>3);
00647
00648 for(i=0; i<8; i++){
00649 ((uint32_t*)(src+i*stride))[0]=
00650 ((uint32_t*)(src+i*stride))[1]= dc0;
00651 }
00652 }
00653
00654
00655 static void pred8x8_dc_c(uint8_t *src, int stride){
00656 int i;
00657 int dc0, dc1, dc2, dc3;
00658
00659 dc0=dc1=dc2=0;
00660 for(i=0;i<4; i++){
00661 dc0+= src[-1+i*stride] + src[i-stride];
00662 dc1+= src[4+i-stride];
00663 dc2+= src[-1+(i+4)*stride];
00664 }
00665 dc3= 0x01010101*((dc1 + dc2 + 4)>>3);
00666 dc0= 0x01010101*((dc0 + 4)>>3);
00667 dc1= 0x01010101*((dc1 + 2)>>2);
00668 dc2= 0x01010101*((dc2 + 2)>>2);
00669
00670 for(i=0; i<4; i++){
00671 ((uint32_t*)(src+i*stride))[0]= dc0;
00672 ((uint32_t*)(src+i*stride))[1]= dc1;
00673 }
00674 for(i=4; i<8; i++){
00675 ((uint32_t*)(src+i*stride))[0]= dc2;
00676 ((uint32_t*)(src+i*stride))[1]= dc3;
00677 }
00678 }
00679
00680
00681 static void pred8x8_mad_cow_dc_l0t(uint8_t *src, int stride){
00682 pred8x8_top_dc_c(src, stride);
00683 pred4x4_dc_c(src, NULL, stride);
00684 }
00685
00686 static void pred8x8_mad_cow_dc_0lt(uint8_t *src, int stride){
00687 pred8x8_dc_c(src, stride);
00688 pred4x4_top_dc_c(src, NULL, stride);
00689 }
00690
00691 static void pred8x8_mad_cow_dc_l00(uint8_t *src, int stride){
00692 pred8x8_left_dc_c(src, stride);
00693 pred4x4_128_dc_c(src + 4*stride , NULL, stride);
00694 pred4x4_128_dc_c(src + 4*stride + 4, NULL, stride);
00695 }
00696
00697 static void pred8x8_mad_cow_dc_0l0(uint8_t *src, int stride){
00698 pred8x8_left_dc_c(src, stride);
00699 pred4x4_128_dc_c(src , NULL, stride);
00700 pred4x4_128_dc_c(src + 4, NULL, stride);
00701 }
00702
00703 static void pred8x8_dc_rv40_c(uint8_t *src, int stride){
00704 int i;
00705 int dc0=0;
00706
00707 for(i=0;i<4; i++){
00708 dc0+= src[-1+i*stride] + src[i-stride];
00709 dc0+= src[4+i-stride];
00710 dc0+= src[-1+(i+4)*stride];
00711 }
00712 dc0= 0x01010101*((dc0 + 8)>>4);
00713
00714 for(i=0; i<4; i++){
00715 ((uint32_t*)(src+i*stride))[0]= dc0;
00716 ((uint32_t*)(src+i*stride))[1]= dc0;
00717 }
00718 for(i=4; i<8; i++){
00719 ((uint32_t*)(src+i*stride))[0]= dc0;
00720 ((uint32_t*)(src+i*stride))[1]= dc0;
00721 }
00722 }
00723
00724 static void pred8x8_plane_c(uint8_t *src, int stride){
00725 int j, k;
00726 int a;
00727 uint8_t *cm = ff_cropTbl + MAX_NEG_CROP;
00728 const uint8_t * const src0 = src+3-stride;
00729 const uint8_t *src1 = src+4*stride-1;
00730 const uint8_t *src2 = src1-2*stride;
00731 int H = src0[1] - src0[-1];
00732 int V = src1[0] - src2[ 0];
00733 for(k=2; k<=4; ++k) {
00734 src1 += stride; src2 -= stride;
00735 H += k*(src0[k] - src0[-k]);
00736 V += k*(src1[0] - src2[ 0]);
00737 }
00738 H = ( 17*H+16 ) >> 5;
00739 V = ( 17*V+16 ) >> 5;
00740
00741 a = 16*(src1[0] + src2[8]+1) - 3*(V+H);
00742 for(j=8; j>0; --j) {
00743 int b = a;
00744 a += V;
00745 src[0] = cm[ (b ) >> 5 ];
00746 src[1] = cm[ (b+ H) >> 5 ];
00747 src[2] = cm[ (b+2*H) >> 5 ];
00748 src[3] = cm[ (b+3*H) >> 5 ];
00749 src[4] = cm[ (b+4*H) >> 5 ];
00750 src[5] = cm[ (b+5*H) >> 5 ];
00751 src[6] = cm[ (b+6*H) >> 5 ];
00752 src[7] = cm[ (b+7*H) >> 5 ];
00753 src += stride;
00754 }
00755 }
00756
00757 #define SRC(x,y) src[(x)+(y)*stride]
00758 #define PL(y) \
00759 const int l##y = (SRC(-1,y-1) + 2*SRC(-1,y) + SRC(-1,y+1) + 2) >> 2;
00760 #define PREDICT_8x8_LOAD_LEFT \
00761 const int l0 = ((has_topleft ? SRC(-1,-1) : SRC(-1,0)) \
00762 + 2*SRC(-1,0) + SRC(-1,1) + 2) >> 2; \
00763 PL(1) PL(2) PL(3) PL(4) PL(5) PL(6) \
00764 const int l7 av_unused = (SRC(-1,6) + 3*SRC(-1,7) + 2) >> 2
00765
00766 #define PT(x) \
00767 const int t##x = (SRC(x-1,-1) + 2*SRC(x,-1) + SRC(x+1,-1) + 2) >> 2;
00768 #define PREDICT_8x8_LOAD_TOP \
00769 const int t0 = ((has_topleft ? SRC(-1,-1) : SRC(0,-1)) \
00770 + 2*SRC(0,-1) + SRC(1,-1) + 2) >> 2; \
00771 PT(1) PT(2) PT(3) PT(4) PT(5) PT(6) \
00772 const int t7 av_unused = ((has_topright ? SRC(8,-1) : SRC(7,-1)) \
00773 + 2*SRC(7,-1) + SRC(6,-1) + 2) >> 2
00774
00775 #define PTR(x) \
00776 t##x = (SRC(x-1,-1) + 2*SRC(x,-1) + SRC(x+1,-1) + 2) >> 2;
00777 #define PREDICT_8x8_LOAD_TOPRIGHT \
00778 int t8, t9, t10, t11, t12, t13, t14, t15; \
00779 if(has_topright) { \
00780 PTR(8) PTR(9) PTR(10) PTR(11) PTR(12) PTR(13) PTR(14) \
00781 t15 = (SRC(14,-1) + 3*SRC(15,-1) + 2) >> 2; \
00782 } else t8=t9=t10=t11=t12=t13=t14=t15= SRC(7,-1);
00783
00784 #define PREDICT_8x8_LOAD_TOPLEFT \
00785 const int lt = (SRC(-1,0) + 2*SRC(-1,-1) + SRC(0,-1) + 2) >> 2
00786
00787 #define PREDICT_8x8_DC(v) \
00788 int y; \
00789 for( y = 0; y < 8; y++ ) { \
00790 ((uint32_t*)src)[0] = \
00791 ((uint32_t*)src)[1] = v; \
00792 src += stride; \
00793 }
00794
00795 static void pred8x8l_128_dc_c(uint8_t *src, int has_topleft, int has_topright, int stride)
00796 {
00797 PREDICT_8x8_DC(0x80808080);
00798 }
00799 static void pred8x8l_left_dc_c(uint8_t *src, int has_topleft, int has_topright, int stride)
00800 {
00801 PREDICT_8x8_LOAD_LEFT;
00802 const uint32_t dc = ((l0+l1+l2+l3+l4+l5+l6+l7+4) >> 3) * 0x01010101;
00803 PREDICT_8x8_DC(dc);
00804 }
00805 static void pred8x8l_top_dc_c(uint8_t *src, int has_topleft, int has_topright, int stride)
00806 {
00807 PREDICT_8x8_LOAD_TOP;
00808 const uint32_t dc = ((t0+t1+t2+t3+t4+t5+t6+t7+4) >> 3) * 0x01010101;
00809 PREDICT_8x8_DC(dc);
00810 }
00811 static void pred8x8l_dc_c(uint8_t *src, int has_topleft, int has_topright, int stride)
00812 {
00813 PREDICT_8x8_LOAD_LEFT;
00814 PREDICT_8x8_LOAD_TOP;
00815 const uint32_t dc = ((l0+l1+l2+l3+l4+l5+l6+l7
00816 +t0+t1+t2+t3+t4+t5+t6+t7+8) >> 4) * 0x01010101;
00817 PREDICT_8x8_DC(dc);
00818 }
00819 static void pred8x8l_horizontal_c(uint8_t *src, int has_topleft, int has_topright, int stride)
00820 {
00821 PREDICT_8x8_LOAD_LEFT;
00822 #define ROW(y) ((uint32_t*)(src+y*stride))[0] =\
00823 ((uint32_t*)(src+y*stride))[1] = 0x01010101 * l##y
00824 ROW(0); ROW(1); ROW(2); ROW(3); ROW(4); ROW(5); ROW(6); ROW(7);
00825 #undef ROW
00826 }
00827 static void pred8x8l_vertical_c(uint8_t *src, int has_topleft, int has_topright, int stride)
00828 {
00829 int y;
00830 PREDICT_8x8_LOAD_TOP;
00831 src[0] = t0;
00832 src[1] = t1;
00833 src[2] = t2;
00834 src[3] = t3;
00835 src[4] = t4;
00836 src[5] = t5;
00837 src[6] = t6;
00838 src[7] = t7;
00839 for( y = 1; y < 8; y++ )
00840 *(uint64_t*)(src+y*stride) = *(uint64_t*)src;
00841 }
00842 static void pred8x8l_down_left_c(uint8_t *src, int has_topleft, int has_topright, int stride)
00843 {
00844 PREDICT_8x8_LOAD_TOP;
00845 PREDICT_8x8_LOAD_TOPRIGHT;
00846 SRC(0,0)= (t0 + 2*t1 + t2 + 2) >> 2;
00847 SRC(0,1)=SRC(1,0)= (t1 + 2*t2 + t3 + 2) >> 2;
00848 SRC(0,2)=SRC(1,1)=SRC(2,0)= (t2 + 2*t3 + t4 + 2) >> 2;
00849 SRC(0,3)=SRC(1,2)=SRC(2,1)=SRC(3,0)= (t3 + 2*t4 + t5 + 2) >> 2;
00850 SRC(0,4)=SRC(1,3)=SRC(2,2)=SRC(3,1)=SRC(4,0)= (t4 + 2*t5 + t6 + 2) >> 2;
00851 SRC(0,5)=SRC(1,4)=SRC(2,3)=SRC(3,2)=SRC(4,1)=SRC(5,0)= (t5 + 2*t6 + t7 + 2) >> 2;
00852 SRC(0,6)=SRC(1,5)=SRC(2,4)=SRC(3,3)=SRC(4,2)=SRC(5,1)=SRC(6,0)= (t6 + 2*t7 + t8 + 2) >> 2;
00853 SRC(0,7)=SRC(1,6)=SRC(2,5)=SRC(3,4)=SRC(4,3)=SRC(5,2)=SRC(6,1)=SRC(7,0)= (t7 + 2*t8 + t9 + 2) >> 2;
00854 SRC(1,7)=SRC(2,6)=SRC(3,5)=SRC(4,4)=SRC(5,3)=SRC(6,2)=SRC(7,1)= (t8 + 2*t9 + t10 + 2) >> 2;
00855 SRC(2,7)=SRC(3,6)=SRC(4,5)=SRC(5,4)=SRC(6,3)=SRC(7,2)= (t9 + 2*t10 + t11 + 2) >> 2;
00856 SRC(3,7)=SRC(4,6)=SRC(5,5)=SRC(6,4)=SRC(7,3)= (t10 + 2*t11 + t12 + 2) >> 2;
00857 SRC(4,7)=SRC(5,6)=SRC(6,5)=SRC(7,4)= (t11 + 2*t12 + t13 + 2) >> 2;
00858 SRC(5,7)=SRC(6,6)=SRC(7,5)= (t12 + 2*t13 + t14 + 2) >> 2;
00859 SRC(6,7)=SRC(7,6)= (t13 + 2*t14 + t15 + 2) >> 2;
00860 SRC(7,7)= (t14 + 3*t15 + 2) >> 2;
00861 }
00862 static void pred8x8l_down_right_c(uint8_t *src, int has_topleft, int has_topright, int stride)
00863 {
00864 PREDICT_8x8_LOAD_TOP;
00865 PREDICT_8x8_LOAD_LEFT;
00866 PREDICT_8x8_LOAD_TOPLEFT;
00867 SRC(0,7)= (l7 + 2*l6 + l5 + 2) >> 2;
00868 SRC(0,6)=SRC(1,7)= (l6 + 2*l5 + l4 + 2) >> 2;
00869 SRC(0,5)=SRC(1,6)=SRC(2,7)= (l5 + 2*l4 + l3 + 2) >> 2;
00870 SRC(0,4)=SRC(1,5)=SRC(2,6)=SRC(3,7)= (l4 + 2*l3 + l2 + 2) >> 2;
00871 SRC(0,3)=SRC(1,4)=SRC(2,5)=SRC(3,6)=SRC(4,7)= (l3 + 2*l2 + l1 + 2) >> 2;
00872 SRC(0,2)=SRC(1,3)=SRC(2,4)=SRC(3,5)=SRC(4,6)=SRC(5,7)= (l2 + 2*l1 + l0 + 2) >> 2;
00873 SRC(0,1)=SRC(1,2)=SRC(2,3)=SRC(3,4)=SRC(4,5)=SRC(5,6)=SRC(6,7)= (l1 + 2*l0 + lt + 2) >> 2;
00874 SRC(0,0)=SRC(1,1)=SRC(2,2)=SRC(3,3)=SRC(4,4)=SRC(5,5)=SRC(6,6)=SRC(7,7)= (l0 + 2*lt + t0 + 2) >> 2;
00875 SRC(1,0)=SRC(2,1)=SRC(3,2)=SRC(4,3)=SRC(5,4)=SRC(6,5)=SRC(7,6)= (lt + 2*t0 + t1 + 2) >> 2;
00876 SRC(2,0)=SRC(3,1)=SRC(4,2)=SRC(5,3)=SRC(6,4)=SRC(7,5)= (t0 + 2*t1 + t2 + 2) >> 2;
00877 SRC(3,0)=SRC(4,1)=SRC(5,2)=SRC(6,3)=SRC(7,4)= (t1 + 2*t2 + t3 + 2) >> 2;
00878 SRC(4,0)=SRC(5,1)=SRC(6,2)=SRC(7,3)= (t2 + 2*t3 + t4 + 2) >> 2;
00879 SRC(5,0)=SRC(6,1)=SRC(7,2)= (t3 + 2*t4 + t5 + 2) >> 2;
00880 SRC(6,0)=SRC(7,1)= (t4 + 2*t5 + t6 + 2) >> 2;
00881 SRC(7,0)= (t5 + 2*t6 + t7 + 2) >> 2;
00882
00883 }
00884 static void pred8x8l_vertical_right_c(uint8_t *src, int has_topleft, int has_topright, int stride)
00885 {
00886 PREDICT_8x8_LOAD_TOP;
00887 PREDICT_8x8_LOAD_LEFT;
00888 PREDICT_8x8_LOAD_TOPLEFT;
00889 SRC(0,6)= (l5 + 2*l4 + l3 + 2) >> 2;
00890 SRC(0,7)= (l6 + 2*l5 + l4 + 2) >> 2;
00891 SRC(0,4)=SRC(1,6)= (l3 + 2*l2 + l1 + 2) >> 2;
00892 SRC(0,5)=SRC(1,7)= (l4 + 2*l3 + l2 + 2) >> 2;
00893 SRC(0,2)=SRC(1,4)=SRC(2,6)= (l1 + 2*l0 + lt + 2) >> 2;
00894 SRC(0,3)=SRC(1,5)=SRC(2,7)= (l2 + 2*l1 + l0 + 2) >> 2;
00895 SRC(0,1)=SRC(1,3)=SRC(2,5)=SRC(3,7)= (l0 + 2*lt + t0 + 2) >> 2;
00896 SRC(0,0)=SRC(1,2)=SRC(2,4)=SRC(3,6)= (lt + t0 + 1) >> 1;
00897 SRC(1,1)=SRC(2,3)=SRC(3,5)=SRC(4,7)= (lt + 2*t0 + t1 + 2) >> 2;
00898 SRC(1,0)=SRC(2,2)=SRC(3,4)=SRC(4,6)= (t0 + t1 + 1) >> 1;
00899 SRC(2,1)=SRC(3,3)=SRC(4,5)=SRC(5,7)= (t0 + 2*t1 + t2 + 2) >> 2;
00900 SRC(2,0)=SRC(3,2)=SRC(4,4)=SRC(5,6)= (t1 + t2 + 1) >> 1;
00901 SRC(3,1)=SRC(4,3)=SRC(5,5)=SRC(6,7)= (t1 + 2*t2 + t3 + 2) >> 2;
00902 SRC(3,0)=SRC(4,2)=SRC(5,4)=SRC(6,6)= (t2 + t3 + 1) >> 1;
00903 SRC(4,1)=SRC(5,3)=SRC(6,5)=SRC(7,7)= (t2 + 2*t3 + t4 + 2) >> 2;
00904 SRC(4,0)=SRC(5,2)=SRC(6,4)=SRC(7,6)= (t3 + t4 + 1) >> 1;
00905 SRC(5,1)=SRC(6,3)=SRC(7,5)= (t3 + 2*t4 + t5 + 2) >> 2;
00906 SRC(5,0)=SRC(6,2)=SRC(7,4)= (t4 + t5 + 1) >> 1;
00907 SRC(6,1)=SRC(7,3)= (t4 + 2*t5 + t6 + 2) >> 2;
00908 SRC(6,0)=SRC(7,2)= (t5 + t6 + 1) >> 1;
00909 SRC(7,1)= (t5 + 2*t6 + t7 + 2) >> 2;
00910 SRC(7,0)= (t6 + t7 + 1) >> 1;
00911 }
00912 static void pred8x8l_horizontal_down_c(uint8_t *src, int has_topleft, int has_topright, int stride)
00913 {
00914 PREDICT_8x8_LOAD_TOP;
00915 PREDICT_8x8_LOAD_LEFT;
00916 PREDICT_8x8_LOAD_TOPLEFT;
00917 SRC(0,7)= (l6 + l7 + 1) >> 1;
00918 SRC(1,7)= (l5 + 2*l6 + l7 + 2) >> 2;
00919 SRC(0,6)=SRC(2,7)= (l5 + l6 + 1) >> 1;
00920 SRC(1,6)=SRC(3,7)= (l4 + 2*l5 + l6 + 2) >> 2;
00921 SRC(0,5)=SRC(2,6)=SRC(4,7)= (l4 + l5 + 1) >> 1;
00922 SRC(1,5)=SRC(3,6)=SRC(5,7)= (l3 + 2*l4 + l5 + 2) >> 2;
00923 SRC(0,4)=SRC(2,5)=SRC(4,6)=SRC(6,7)= (l3 + l4 + 1) >> 1;
00924 SRC(1,4)=SRC(3,5)=SRC(5,6)=SRC(7,7)= (l2 + 2*l3 + l4 + 2) >> 2;
00925 SRC(0,3)=SRC(2,4)=SRC(4,5)=SRC(6,6)= (l2 + l3 + 1) >> 1;
00926 SRC(1,3)=SRC(3,4)=SRC(5,5)=SRC(7,6)= (l1 + 2*l2 + l3 + 2) >> 2;
00927 SRC(0,2)=SRC(2,3)=SRC(4,4)=SRC(6,5)= (l1 + l2 + 1) >> 1;
00928 SRC(1,2)=SRC(3,3)=SRC(5,4)=SRC(7,5)= (l0 + 2*l1 + l2 + 2) >> 2;
00929 SRC(0,1)=SRC(2,2)=SRC(4,3)=SRC(6,4)= (l0 + l1 + 1) >> 1;
00930 SRC(1,1)=SRC(3,2)=SRC(5,3)=SRC(7,4)= (lt + 2*l0 + l1 + 2) >> 2;
00931 SRC(0,0)=SRC(2,1)=SRC(4,2)=SRC(6,3)= (lt + l0 + 1) >> 1;
00932 SRC(1,0)=SRC(3,1)=SRC(5,2)=SRC(7,3)= (l0 + 2*lt + t0 + 2) >> 2;
00933 SRC(2,0)=SRC(4,1)=SRC(6,2)= (t1 + 2*t0 + lt + 2) >> 2;
00934 SRC(3,0)=SRC(5,1)=SRC(7,2)= (t2 + 2*t1 + t0 + 2) >> 2;
00935 SRC(4,0)=SRC(6,1)= (t3 + 2*t2 + t1 + 2) >> 2;
00936 SRC(5,0)=SRC(7,1)= (t4 + 2*t3 + t2 + 2) >> 2;
00937 SRC(6,0)= (t5 + 2*t4 + t3 + 2) >> 2;
00938 SRC(7,0)= (t6 + 2*t5 + t4 + 2) >> 2;
00939 }
00940 static void pred8x8l_vertical_left_c(uint8_t *src, int has_topleft, int has_topright, int stride)
00941 {
00942 PREDICT_8x8_LOAD_TOP;
00943 PREDICT_8x8_LOAD_TOPRIGHT;
00944 SRC(0,0)= (t0 + t1 + 1) >> 1;
00945 SRC(0,1)= (t0 + 2*t1 + t2 + 2) >> 2;
00946 SRC(0,2)=SRC(1,0)= (t1 + t2 + 1) >> 1;
00947 SRC(0,3)=SRC(1,1)= (t1 + 2*t2 + t3 + 2) >> 2;
00948 SRC(0,4)=SRC(1,2)=SRC(2,0)= (t2 + t3 + 1) >> 1;
00949 SRC(0,5)=SRC(1,3)=SRC(2,1)= (t2 + 2*t3 + t4 + 2) >> 2;
00950 SRC(0,6)=SRC(1,4)=SRC(2,2)=SRC(3,0)= (t3 + t4 + 1) >> 1;
00951 SRC(0,7)=SRC(1,5)=SRC(2,3)=SRC(3,1)= (t3 + 2*t4 + t5 + 2) >> 2;
00952 SRC(1,6)=SRC(2,4)=SRC(3,2)=SRC(4,0)= (t4 + t5 + 1) >> 1;
00953 SRC(1,7)=SRC(2,5)=SRC(3,3)=SRC(4,1)= (t4 + 2*t5 + t6 + 2) >> 2;
00954 SRC(2,6)=SRC(3,4)=SRC(4,2)=SRC(5,0)= (t5 + t6 + 1) >> 1;
00955 SRC(2,7)=SRC(3,5)=SRC(4,3)=SRC(5,1)= (t5 + 2*t6 + t7 + 2) >> 2;
00956 SRC(3,6)=SRC(4,4)=SRC(5,2)=SRC(6,0)= (t6 + t7 + 1) >> 1;
00957 SRC(3,7)=SRC(4,5)=SRC(5,3)=SRC(6,1)= (t6 + 2*t7 + t8 + 2) >> 2;
00958 SRC(4,6)=SRC(5,4)=SRC(6,2)=SRC(7,0)= (t7 + t8 + 1) >> 1;
00959 SRC(4,7)=SRC(5,5)=SRC(6,3)=SRC(7,1)= (t7 + 2*t8 + t9 + 2) >> 2;
00960 SRC(5,6)=SRC(6,4)=SRC(7,2)= (t8 + t9 + 1) >> 1;
00961 SRC(5,7)=SRC(6,5)=SRC(7,3)= (t8 + 2*t9 + t10 + 2) >> 2;
00962 SRC(6,6)=SRC(7,4)= (t9 + t10 + 1) >> 1;
00963 SRC(6,7)=SRC(7,5)= (t9 + 2*t10 + t11 + 2) >> 2;
00964 SRC(7,6)= (t10 + t11 + 1) >> 1;
00965 SRC(7,7)= (t10 + 2*t11 + t12 + 2) >> 2;
00966 }
00967 static void pred8x8l_horizontal_up_c(uint8_t *src, int has_topleft, int has_topright, int stride)
00968 {
00969 PREDICT_8x8_LOAD_LEFT;
00970 SRC(0,0)= (l0 + l1 + 1) >> 1;
00971 SRC(1,0)= (l0 + 2*l1 + l2 + 2) >> 2;
00972 SRC(0,1)=SRC(2,0)= (l1 + l2 + 1) >> 1;
00973 SRC(1,1)=SRC(3,0)= (l1 + 2*l2 + l3 + 2) >> 2;
00974 SRC(0,2)=SRC(2,1)=SRC(4,0)= (l2 + l3 + 1) >> 1;
00975 SRC(1,2)=SRC(3,1)=SRC(5,0)= (l2 + 2*l3 + l4 + 2) >> 2;
00976 SRC(0,3)=SRC(2,2)=SRC(4,1)=SRC(6,0)= (l3 + l4 + 1) >> 1;
00977 SRC(1,3)=SRC(3,2)=SRC(5,1)=SRC(7,0)= (l3 + 2*l4 + l5 + 2) >> 2;
00978 SRC(0,4)=SRC(2,3)=SRC(4,2)=SRC(6,1)= (l4 + l5 + 1) >> 1;
00979 SRC(1,4)=SRC(3,3)=SRC(5,2)=SRC(7,1)= (l4 + 2*l5 + l6 + 2) >> 2;
00980 SRC(0,5)=SRC(2,4)=SRC(4,3)=SRC(6,2)= (l5 + l6 + 1) >> 1;
00981 SRC(1,5)=SRC(3,4)=SRC(5,3)=SRC(7,2)= (l5 + 2*l6 + l7 + 2) >> 2;
00982 SRC(0,6)=SRC(2,5)=SRC(4,4)=SRC(6,3)= (l6 + l7 + 1) >> 1;
00983 SRC(1,6)=SRC(3,5)=SRC(5,4)=SRC(7,3)= (l6 + 3*l7 + 2) >> 2;
00984 SRC(0,7)=SRC(1,7)=SRC(2,6)=SRC(2,7)=SRC(3,6)=
00985 SRC(3,7)=SRC(4,5)=SRC(4,6)=SRC(4,7)=SRC(5,5)=
00986 SRC(5,6)=SRC(5,7)=SRC(6,4)=SRC(6,5)=SRC(6,6)=
00987 SRC(6,7)=SRC(7,4)=SRC(7,5)=SRC(7,6)=SRC(7,7)= l7;
00988 }
00989 #undef PREDICT_8x8_LOAD_LEFT
00990 #undef PREDICT_8x8_LOAD_TOP
00991 #undef PREDICT_8x8_LOAD_TOPLEFT
00992 #undef PREDICT_8x8_LOAD_TOPRIGHT
00993 #undef PREDICT_8x8_DC
00994 #undef PTR
00995 #undef PT
00996 #undef PL
00997 #undef SRC
00998
01002 void ff_h264_pred_init(H264PredContext *h, int codec_id){
01003
01004
01005 if(codec_id != CODEC_ID_RV40){
01006 h->pred4x4[VERT_PRED ]= pred4x4_vertical_c;
01007 h->pred4x4[HOR_PRED ]= pred4x4_horizontal_c;
01008 h->pred4x4[DC_PRED ]= pred4x4_dc_c;
01009 if(codec_id == CODEC_ID_SVQ3)
01010 h->pred4x4[DIAG_DOWN_LEFT_PRED ]= pred4x4_down_left_svq3_c;
01011 else
01012 h->pred4x4[DIAG_DOWN_LEFT_PRED ]= pred4x4_down_left_c;
01013 h->pred4x4[DIAG_DOWN_RIGHT_PRED]= pred4x4_down_right_c;
01014 h->pred4x4[VERT_RIGHT_PRED ]= pred4x4_vertical_right_c;
01015 h->pred4x4[HOR_DOWN_PRED ]= pred4x4_horizontal_down_c;
01016 h->pred4x4[VERT_LEFT_PRED ]= pred4x4_vertical_left_c;
01017 h->pred4x4[HOR_UP_PRED ]= pred4x4_horizontal_up_c;
01018 h->pred4x4[LEFT_DC_PRED ]= pred4x4_left_dc_c;
01019 h->pred4x4[TOP_DC_PRED ]= pred4x4_top_dc_c;
01020 h->pred4x4[DC_128_PRED ]= pred4x4_128_dc_c;
01021 }else{
01022 h->pred4x4[VERT_PRED ]= pred4x4_vertical_c;
01023 h->pred4x4[HOR_PRED ]= pred4x4_horizontal_c;
01024 h->pred4x4[DC_PRED ]= pred4x4_dc_c;
01025 h->pred4x4[DIAG_DOWN_LEFT_PRED ]= pred4x4_down_left_rv40_c;
01026 h->pred4x4[DIAG_DOWN_RIGHT_PRED]= pred4x4_down_right_c;
01027 h->pred4x4[VERT_RIGHT_PRED ]= pred4x4_vertical_right_c;
01028 h->pred4x4[HOR_DOWN_PRED ]= pred4x4_horizontal_down_c;
01029 h->pred4x4[VERT_LEFT_PRED ]= pred4x4_vertical_left_rv40_c;
01030 h->pred4x4[HOR_UP_PRED ]= pred4x4_horizontal_up_rv40_c;
01031 h->pred4x4[LEFT_DC_PRED ]= pred4x4_left_dc_c;
01032 h->pred4x4[TOP_DC_PRED ]= pred4x4_top_dc_c;
01033 h->pred4x4[DC_128_PRED ]= pred4x4_128_dc_c;
01034 h->pred4x4[DIAG_DOWN_LEFT_PRED_RV40_NODOWN]= pred4x4_down_left_rv40_nodown_c;
01035 h->pred4x4[HOR_UP_PRED_RV40_NODOWN]= pred4x4_horizontal_up_rv40_nodown_c;
01036 }
01037
01038 h->pred8x8l[VERT_PRED ]= pred8x8l_vertical_c;
01039 h->pred8x8l[HOR_PRED ]= pred8x8l_horizontal_c;
01040 h->pred8x8l[DC_PRED ]= pred8x8l_dc_c;
01041 h->pred8x8l[DIAG_DOWN_LEFT_PRED ]= pred8x8l_down_left_c;
01042 h->pred8x8l[DIAG_DOWN_RIGHT_PRED]= pred8x8l_down_right_c;
01043 h->pred8x8l[VERT_RIGHT_PRED ]= pred8x8l_vertical_right_c;
01044 h->pred8x8l[HOR_DOWN_PRED ]= pred8x8l_horizontal_down_c;
01045 h->pred8x8l[VERT_LEFT_PRED ]= pred8x8l_vertical_left_c;
01046 h->pred8x8l[HOR_UP_PRED ]= pred8x8l_horizontal_up_c;
01047 h->pred8x8l[LEFT_DC_PRED ]= pred8x8l_left_dc_c;
01048 h->pred8x8l[TOP_DC_PRED ]= pred8x8l_top_dc_c;
01049 h->pred8x8l[DC_128_PRED ]= pred8x8l_128_dc_c;
01050
01051 h->pred8x8[VERT_PRED8x8 ]= pred8x8_vertical_c;
01052 h->pred8x8[HOR_PRED8x8 ]= pred8x8_horizontal_c;
01053 h->pred8x8[PLANE_PRED8x8 ]= pred8x8_plane_c;
01054 if(codec_id != CODEC_ID_RV40){
01055 h->pred8x8[DC_PRED8x8 ]= pred8x8_dc_c;
01056 h->pred8x8[LEFT_DC_PRED8x8]= pred8x8_left_dc_c;
01057 h->pred8x8[TOP_DC_PRED8x8 ]= pred8x8_top_dc_c;
01058 h->pred8x8[ALZHEIMER_DC_L0T_PRED8x8 ]= pred8x8_mad_cow_dc_l0t;
01059 h->pred8x8[ALZHEIMER_DC_0LT_PRED8x8 ]= pred8x8_mad_cow_dc_0lt;
01060 h->pred8x8[ALZHEIMER_DC_L00_PRED8x8 ]= pred8x8_mad_cow_dc_l00;
01061 h->pred8x8[ALZHEIMER_DC_0L0_PRED8x8 ]= pred8x8_mad_cow_dc_0l0;
01062 }else{
01063 h->pred8x8[DC_PRED8x8 ]= pred8x8_dc_rv40_c;
01064 h->pred8x8[LEFT_DC_PRED8x8]= pred8x8_left_dc_rv40_c;
01065 h->pred8x8[TOP_DC_PRED8x8 ]= pred8x8_top_dc_rv40_c;
01066 }
01067 h->pred8x8[DC_128_PRED8x8 ]= pred8x8_128_dc_c;
01068
01069 h->pred16x16[DC_PRED8x8 ]= pred16x16_dc_c;
01070 h->pred16x16[VERT_PRED8x8 ]= pred16x16_vertical_c;
01071 h->pred16x16[HOR_PRED8x8 ]= pred16x16_horizontal_c;
01072 h->pred16x16[PLANE_PRED8x8 ]= pred16x16_plane_c;
01073 switch(codec_id){
01074 case CODEC_ID_SVQ3:
01075 h->pred16x16[PLANE_PRED8x8 ]= pred16x16_plane_svq3_c;
01076 break;
01077 case CODEC_ID_RV40:
01078 h->pred16x16[PLANE_PRED8x8 ]= pred16x16_plane_rv40_c;
01079 break;
01080 default:
01081 h->pred16x16[PLANE_PRED8x8 ]= pred16x16_plane_c;
01082 }
01083 h->pred16x16[LEFT_DC_PRED8x8]= pred16x16_left_dc_c;
01084 h->pred16x16[TOP_DC_PRED8x8 ]= pred16x16_top_dc_c;
01085 h->pred16x16[DC_128_PRED8x8 ]= pred16x16_128_dc_c;
01086 }
