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(uint8_t *src, uint8_t *topright, int stride,
00292 const int l0, const int l1, const int l2, const int l3, const int l4){
00293 LOAD_TOP_EDGE
00294 LOAD_TOP_RIGHT_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 + l4 + 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_vertical_left_rv40_c(uint8_t *src, uint8_t *topright, int stride){
00315 LOAD_LEFT_EDGE
00316 LOAD_DOWN_LEFT_EDGE
00317
00318 pred4x4_vertical_left_rv40(src, topright, stride, l0, l1, l2, l3, l4);
00319 }
00320
00321 static void pred4x4_vertical_left_rv40_nodown_c(uint8_t *src, uint8_t *topright, int stride){
00322 LOAD_LEFT_EDGE
00323
00324 pred4x4_vertical_left_rv40(src, topright, stride, l0, l1, l2, l3, l3);
00325 }
00326
00327 static void pred4x4_horizontal_up_c(uint8_t *src, uint8_t *topright, int stride){
00328 LOAD_LEFT_EDGE
00329
00330 src[0+0*stride]=(l0 + l1 + 1)>>1;
00331 src[1+0*stride]=(l0 + 2*l1 + l2 + 2)>>2;
00332 src[2+0*stride]=
00333 src[0+1*stride]=(l1 + l2 + 1)>>1;
00334 src[3+0*stride]=
00335 src[1+1*stride]=(l1 + 2*l2 + l3 + 2)>>2;
00336 src[2+1*stride]=
00337 src[0+2*stride]=(l2 + l3 + 1)>>1;
00338 src[3+1*stride]=
00339 src[1+2*stride]=(l2 + 2*l3 + l3 + 2)>>2;
00340 src[3+2*stride]=
00341 src[1+3*stride]=
00342 src[0+3*stride]=
00343 src[2+2*stride]=
00344 src[2+3*stride]=
00345 src[3+3*stride]=l3;
00346 }
00347
00348 static void pred4x4_horizontal_up_rv40_c(uint8_t *src, uint8_t *topright, int stride){
00349 LOAD_LEFT_EDGE
00350 LOAD_DOWN_LEFT_EDGE
00351 LOAD_TOP_EDGE
00352 LOAD_TOP_RIGHT_EDGE
00353
00354 src[0+0*stride]=(t1 + 2*t2 + t3 + 2*l0 + 2*l1 + 4)>>3;
00355 src[1+0*stride]=(t2 + 2*t3 + t4 + l0 + 2*l1 + l2 + 4)>>3;
00356 src[2+0*stride]=
00357 src[0+1*stride]=(t3 + 2*t4 + t5 + 2*l1 + 2*l2 + 4)>>3;
00358 src[3+0*stride]=
00359 src[1+1*stride]=(t4 + 2*t5 + t6 + l1 + 2*l2 + l3 + 4)>>3;
00360 src[2+1*stride]=
00361 src[0+2*stride]=(t5 + 2*t6 + t7 + 2*l2 + 2*l3 + 4)>>3;
00362 src[3+1*stride]=
00363 src[1+2*stride]=(t6 + 3*t7 + l2 + 3*l3 + 4)>>3;
00364 src[3+2*stride]=
00365 src[1+3*stride]=(l3 + 2*l4 + l5 + 2)>>2;
00366 src[0+3*stride]=
00367 src[2+2*stride]=(t6 + t7 + l3 + l4 + 2)>>2;
00368 src[2+3*stride]=(l4 + l5 + 1)>>1;
00369 src[3+3*stride]=(l4 + 2*l5 + l6 + 2)>>2;
00370 }
00371
00372 static void pred4x4_horizontal_up_rv40_nodown_c(uint8_t *src, uint8_t *topright, int stride){
00373 LOAD_LEFT_EDGE
00374 LOAD_TOP_EDGE
00375 LOAD_TOP_RIGHT_EDGE
00376
00377 src[0+0*stride]=(t1 + 2*t2 + t3 + 2*l0 + 2*l1 + 4)>>3;
00378 src[1+0*stride]=(t2 + 2*t3 + t4 + l0 + 2*l1 + l2 + 4)>>3;
00379 src[2+0*stride]=
00380 src[0+1*stride]=(t3 + 2*t4 + t5 + 2*l1 + 2*l2 + 4)>>3;
00381 src[3+0*stride]=
00382 src[1+1*stride]=(t4 + 2*t5 + t6 + l1 + 2*l2 + l3 + 4)>>3;
00383 src[2+1*stride]=
00384 src[0+2*stride]=(t5 + 2*t6 + t7 + 2*l2 + 2*l3 + 4)>>3;
00385 src[3+1*stride]=
00386 src[1+2*stride]=(t6 + 3*t7 + l2 + 3*l3 + 4)>>3;
00387 src[3+2*stride]=
00388 src[1+3*stride]=l3;
00389 src[0+3*stride]=
00390 src[2+2*stride]=(t6 + t7 + 2*l3 + 2)>>2;
00391 src[2+3*stride]=
00392 src[3+3*stride]=l3;
00393 }
00394
00395 static void pred4x4_horizontal_down_c(uint8_t *src, uint8_t *topright, int stride){
00396 const int lt= src[-1-1*stride];
00397 LOAD_TOP_EDGE
00398 LOAD_LEFT_EDGE
00399
00400 src[0+0*stride]=
00401 src[2+1*stride]=(lt + l0 + 1)>>1;
00402 src[1+0*stride]=
00403 src[3+1*stride]=(l0 + 2*lt + t0 + 2)>>2;
00404 src[2+0*stride]=(lt + 2*t0 + t1 + 2)>>2;
00405 src[3+0*stride]=(t0 + 2*t1 + t2 + 2)>>2;
00406 src[0+1*stride]=
00407 src[2+2*stride]=(l0 + l1 + 1)>>1;
00408 src[1+1*stride]=
00409 src[3+2*stride]=(lt + 2*l0 + l1 + 2)>>2;
00410 src[0+2*stride]=
00411 src[2+3*stride]=(l1 + l2+ 1)>>1;
00412 src[1+2*stride]=
00413 src[3+3*stride]=(l0 + 2*l1 + l2 + 2)>>2;
00414 src[0+3*stride]=(l2 + l3 + 1)>>1;
00415 src[1+3*stride]=(l1 + 2*l2 + l3 + 2)>>2;
00416 }
00417
00418 static void pred16x16_vertical_c(uint8_t *src, int stride){
00419 int i;
00420 const uint32_t a= ((uint32_t*)(src-stride))[0];
00421 const uint32_t b= ((uint32_t*)(src-stride))[1];
00422 const uint32_t c= ((uint32_t*)(src-stride))[2];
00423 const uint32_t d= ((uint32_t*)(src-stride))[3];
00424
00425 for(i=0; i<16; i++){
00426 ((uint32_t*)(src+i*stride))[0]= a;
00427 ((uint32_t*)(src+i*stride))[1]= b;
00428 ((uint32_t*)(src+i*stride))[2]= c;
00429 ((uint32_t*)(src+i*stride))[3]= d;
00430 }
00431 }
00432
00433 static void pred16x16_horizontal_c(uint8_t *src, int stride){
00434 int i;
00435
00436 for(i=0; i<16; i++){
00437 ((uint32_t*)(src+i*stride))[0]=
00438 ((uint32_t*)(src+i*stride))[1]=
00439 ((uint32_t*)(src+i*stride))[2]=
00440 ((uint32_t*)(src+i*stride))[3]= src[-1+i*stride]*0x01010101;
00441 }
00442 }
00443
00444 static void pred16x16_dc_c(uint8_t *src, int stride){
00445 int i, dc=0;
00446
00447 for(i=0;i<16; i++){
00448 dc+= src[-1+i*stride];
00449 }
00450
00451 for(i=0;i<16; i++){
00452 dc+= src[i-stride];
00453 }
00454
00455 dc= 0x01010101*((dc + 16)>>5);
00456
00457 for(i=0; i<16; i++){
00458 ((uint32_t*)(src+i*stride))[0]=
00459 ((uint32_t*)(src+i*stride))[1]=
00460 ((uint32_t*)(src+i*stride))[2]=
00461 ((uint32_t*)(src+i*stride))[3]= dc;
00462 }
00463 }
00464
00465 static void pred16x16_left_dc_c(uint8_t *src, int stride){
00466 int i, dc=0;
00467
00468 for(i=0;i<16; i++){
00469 dc+= src[-1+i*stride];
00470 }
00471
00472 dc= 0x01010101*((dc + 8)>>4);
00473
00474 for(i=0; i<16; i++){
00475 ((uint32_t*)(src+i*stride))[0]=
00476 ((uint32_t*)(src+i*stride))[1]=
00477 ((uint32_t*)(src+i*stride))[2]=
00478 ((uint32_t*)(src+i*stride))[3]= dc;
00479 }
00480 }
00481
00482 static void pred16x16_top_dc_c(uint8_t *src, int stride){
00483 int i, dc=0;
00484
00485 for(i=0;i<16; i++){
00486 dc+= src[i-stride];
00487 }
00488 dc= 0x01010101*((dc + 8)>>4);
00489
00490 for(i=0; i<16; i++){
00491 ((uint32_t*)(src+i*stride))[0]=
00492 ((uint32_t*)(src+i*stride))[1]=
00493 ((uint32_t*)(src+i*stride))[2]=
00494 ((uint32_t*)(src+i*stride))[3]= dc;
00495 }
00496 }
00497
00498 static void pred16x16_128_dc_c(uint8_t *src, int stride){
00499 int i;
00500
00501 for(i=0; i<16; i++){
00502 ((uint32_t*)(src+i*stride))[0]=
00503 ((uint32_t*)(src+i*stride))[1]=
00504 ((uint32_t*)(src+i*stride))[2]=
00505 ((uint32_t*)(src+i*stride))[3]= 0x01010101U*128U;
00506 }
00507 }
00508
00509 static inline void pred16x16_plane_compat_c(uint8_t *src, int stride, const int svq3, const int rv40){
00510 int i, j, k;
00511 int a;
00512 uint8_t *cm = ff_cropTbl + MAX_NEG_CROP;
00513 const uint8_t * const src0 = src+7-stride;
00514 const uint8_t *src1 = src+8*stride-1;
00515 const uint8_t *src2 = src1-2*stride;
00516 int H = src0[1] - src0[-1];
00517 int V = src1[0] - src2[ 0];
00518 for(k=2; k<=8; ++k) {
00519 src1 += stride; src2 -= stride;
00520 H += k*(src0[k] - src0[-k]);
00521 V += k*(src1[0] - src2[ 0]);
00522 }
00523 if(svq3){
00524 H = ( 5*(H/4) ) / 16;
00525 V = ( 5*(V/4) ) / 16;
00526
00527
00528 i = H; H = V; V = i;
00529 }else if(rv40){
00530 H = ( H + (H>>2) ) >> 4;
00531 V = ( V + (V>>2) ) >> 4;
00532 }else{
00533 H = ( 5*H+32 ) >> 6;
00534 V = ( 5*V+32 ) >> 6;
00535 }
00536
00537 a = 16*(src1[0] + src2[16] + 1) - 7*(V+H);
00538 for(j=16; j>0; --j) {
00539 int b = a;
00540 a += V;
00541 for(i=-16; i<0; i+=4) {
00542 src[16+i] = cm[ (b ) >> 5 ];
00543 src[17+i] = cm[ (b+ H) >> 5 ];
00544 src[18+i] = cm[ (b+2*H) >> 5 ];
00545 src[19+i] = cm[ (b+3*H) >> 5 ];
00546 b += 4*H;
00547 }
00548 src += stride;
00549 }
00550 }
00551
00552 static void pred16x16_plane_c(uint8_t *src, int stride){
00553 pred16x16_plane_compat_c(src, stride, 0, 0);
00554 }
00555
00556 static void pred16x16_plane_svq3_c(uint8_t *src, int stride){
00557 pred16x16_plane_compat_c(src, stride, 1, 0);
00558 }
00559
00560 static void pred16x16_plane_rv40_c(uint8_t *src, int stride){
00561 pred16x16_plane_compat_c(src, stride, 0, 1);
00562 }
00563
00564 static void pred8x8_vertical_c(uint8_t *src, int stride){
00565 int i;
00566 const uint32_t a= ((uint32_t*)(src-stride))[0];
00567 const uint32_t b= ((uint32_t*)(src-stride))[1];
00568
00569 for(i=0; i<8; i++){
00570 ((uint32_t*)(src+i*stride))[0]= a;
00571 ((uint32_t*)(src+i*stride))[1]= b;
00572 }
00573 }
00574
00575 static void pred8x8_horizontal_c(uint8_t *src, int stride){
00576 int i;
00577
00578 for(i=0; i<8; i++){
00579 ((uint32_t*)(src+i*stride))[0]=
00580 ((uint32_t*)(src+i*stride))[1]= src[-1+i*stride]*0x01010101;
00581 }
00582 }
00583
00584 static void pred8x8_128_dc_c(uint8_t *src, int stride){
00585 int i;
00586
00587 for(i=0; i<8; i++){
00588 ((uint32_t*)(src+i*stride))[0]=
00589 ((uint32_t*)(src+i*stride))[1]= 0x01010101U*128U;
00590 }
00591 }
00592
00593 static void pred8x8_left_dc_c(uint8_t *src, int stride){
00594 int i;
00595 int dc0, dc2;
00596
00597 dc0=dc2=0;
00598 for(i=0;i<4; i++){
00599 dc0+= src[-1+i*stride];
00600 dc2+= src[-1+(i+4)*stride];
00601 }
00602 dc0= 0x01010101*((dc0 + 2)>>2);
00603 dc2= 0x01010101*((dc2 + 2)>>2);
00604
00605 for(i=0; i<4; i++){
00606 ((uint32_t*)(src+i*stride))[0]=
00607 ((uint32_t*)(src+i*stride))[1]= dc0;
00608 }
00609 for(i=4; i<8; i++){
00610 ((uint32_t*)(src+i*stride))[0]=
00611 ((uint32_t*)(src+i*stride))[1]= dc2;
00612 }
00613 }
00614
00615 static void pred8x8_left_dc_rv40_c(uint8_t *src, int stride){
00616 int i;
00617 int dc0;
00618
00619 dc0=0;
00620 for(i=0;i<8; i++)
00621 dc0+= src[-1+i*stride];
00622 dc0= 0x01010101*((dc0 + 4)>>3);
00623
00624 for(i=0; i<8; i++){
00625 ((uint32_t*)(src+i*stride))[0]=
00626 ((uint32_t*)(src+i*stride))[1]= dc0;
00627 }
00628 }
00629
00630 static void pred8x8_top_dc_c(uint8_t *src, int stride){
00631 int i;
00632 int dc0, dc1;
00633
00634 dc0=dc1=0;
00635 for(i=0;i<4; i++){
00636 dc0+= src[i-stride];
00637 dc1+= src[4+i-stride];
00638 }
00639 dc0= 0x01010101*((dc0 + 2)>>2);
00640 dc1= 0x01010101*((dc1 + 2)>>2);
00641
00642 for(i=0; i<4; i++){
00643 ((uint32_t*)(src+i*stride))[0]= dc0;
00644 ((uint32_t*)(src+i*stride))[1]= dc1;
00645 }
00646 for(i=4; i<8; i++){
00647 ((uint32_t*)(src+i*stride))[0]= dc0;
00648 ((uint32_t*)(src+i*stride))[1]= dc1;
00649 }
00650 }
00651
00652 static void pred8x8_top_dc_rv40_c(uint8_t *src, int stride){
00653 int i;
00654 int dc0;
00655
00656 dc0=0;
00657 for(i=0;i<8; i++)
00658 dc0+= src[i-stride];
00659 dc0= 0x01010101*((dc0 + 4)>>3);
00660
00661 for(i=0; i<8; i++){
00662 ((uint32_t*)(src+i*stride))[0]=
00663 ((uint32_t*)(src+i*stride))[1]= dc0;
00664 }
00665 }
00666
00667
00668 static void pred8x8_dc_c(uint8_t *src, int stride){
00669 int i;
00670 int dc0, dc1, dc2, dc3;
00671
00672 dc0=dc1=dc2=0;
00673 for(i=0;i<4; i++){
00674 dc0+= src[-1+i*stride] + src[i-stride];
00675 dc1+= src[4+i-stride];
00676 dc2+= src[-1+(i+4)*stride];
00677 }
00678 dc3= 0x01010101*((dc1 + dc2 + 4)>>3);
00679 dc0= 0x01010101*((dc0 + 4)>>3);
00680 dc1= 0x01010101*((dc1 + 2)>>2);
00681 dc2= 0x01010101*((dc2 + 2)>>2);
00682
00683 for(i=0; i<4; i++){
00684 ((uint32_t*)(src+i*stride))[0]= dc0;
00685 ((uint32_t*)(src+i*stride))[1]= dc1;
00686 }
00687 for(i=4; i<8; i++){
00688 ((uint32_t*)(src+i*stride))[0]= dc2;
00689 ((uint32_t*)(src+i*stride))[1]= dc3;
00690 }
00691 }
00692
00693 static void pred8x8_dc_rv40_c(uint8_t *src, int stride){
00694 int i;
00695 int dc0=0;
00696
00697 for(i=0;i<4; i++){
00698 dc0+= src[-1+i*stride] + src[i-stride];
00699 dc0+= src[4+i-stride];
00700 dc0+= src[-1+(i+4)*stride];
00701 }
00702 dc0= 0x01010101*((dc0 + 8)>>4);
00703
00704 for(i=0; i<4; i++){
00705 ((uint32_t*)(src+i*stride))[0]= dc0;
00706 ((uint32_t*)(src+i*stride))[1]= dc0;
00707 }
00708 for(i=4; i<8; i++){
00709 ((uint32_t*)(src+i*stride))[0]= dc0;
00710 ((uint32_t*)(src+i*stride))[1]= dc0;
00711 }
00712 }
00713
00714 static void pred8x8_plane_c(uint8_t *src, int stride){
00715 int j, k;
00716 int a;
00717 uint8_t *cm = ff_cropTbl + MAX_NEG_CROP;
00718 const uint8_t * const src0 = src+3-stride;
00719 const uint8_t *src1 = src+4*stride-1;
00720 const uint8_t *src2 = src1-2*stride;
00721 int H = src0[1] - src0[-1];
00722 int V = src1[0] - src2[ 0];
00723 for(k=2; k<=4; ++k) {
00724 src1 += stride; src2 -= stride;
00725 H += k*(src0[k] - src0[-k]);
00726 V += k*(src1[0] - src2[ 0]);
00727 }
00728 H = ( 17*H+16 ) >> 5;
00729 V = ( 17*V+16 ) >> 5;
00730
00731 a = 16*(src1[0] + src2[8]+1) - 3*(V+H);
00732 for(j=8; j>0; --j) {
00733 int b = a;
00734 a += V;
00735 src[0] = cm[ (b ) >> 5 ];
00736 src[1] = cm[ (b+ H) >> 5 ];
00737 src[2] = cm[ (b+2*H) >> 5 ];
00738 src[3] = cm[ (b+3*H) >> 5 ];
00739 src[4] = cm[ (b+4*H) >> 5 ];
00740 src[5] = cm[ (b+5*H) >> 5 ];
00741 src[6] = cm[ (b+6*H) >> 5 ];
00742 src[7] = cm[ (b+7*H) >> 5 ];
00743 src += stride;
00744 }
00745 }
00746
00747 #define SRC(x,y) src[(x)+(y)*stride]
00748 #define PL(y) \
00749 const int l##y = (SRC(-1,y-1) + 2*SRC(-1,y) + SRC(-1,y+1) + 2) >> 2;
00750 #define PREDICT_8x8_LOAD_LEFT \
00751 const int l0 = ((has_topleft ? SRC(-1,-1) : SRC(-1,0)) \
00752 + 2*SRC(-1,0) + SRC(-1,1) + 2) >> 2; \
00753 PL(1) PL(2) PL(3) PL(4) PL(5) PL(6) \
00754 const int l7 av_unused = (SRC(-1,6) + 3*SRC(-1,7) + 2) >> 2
00755
00756 #define PT(x) \
00757 const int t##x = (SRC(x-1,-1) + 2*SRC(x,-1) + SRC(x+1,-1) + 2) >> 2;
00758 #define PREDICT_8x8_LOAD_TOP \
00759 const int t0 = ((has_topleft ? SRC(-1,-1) : SRC(0,-1)) \
00760 + 2*SRC(0,-1) + SRC(1,-1) + 2) >> 2; \
00761 PT(1) PT(2) PT(3) PT(4) PT(5) PT(6) \
00762 const int t7 av_unused = ((has_topright ? SRC(8,-1) : SRC(7,-1)) \
00763 + 2*SRC(7,-1) + SRC(6,-1) + 2) >> 2
00764
00765 #define PTR(x) \
00766 t##x = (SRC(x-1,-1) + 2*SRC(x,-1) + SRC(x+1,-1) + 2) >> 2;
00767 #define PREDICT_8x8_LOAD_TOPRIGHT \
00768 int t8, t9, t10, t11, t12, t13, t14, t15; \
00769 if(has_topright) { \
00770 PTR(8) PTR(9) PTR(10) PTR(11) PTR(12) PTR(13) PTR(14) \
00771 t15 = (SRC(14,-1) + 3*SRC(15,-1) + 2) >> 2; \
00772 } else t8=t9=t10=t11=t12=t13=t14=t15= SRC(7,-1);
00773
00774 #define PREDICT_8x8_LOAD_TOPLEFT \
00775 const int lt = (SRC(-1,0) + 2*SRC(-1,-1) + SRC(0,-1) + 2) >> 2
00776
00777 #define PREDICT_8x8_DC(v) \
00778 int y; \
00779 for( y = 0; y < 8; y++ ) { \
00780 ((uint32_t*)src)[0] = \
00781 ((uint32_t*)src)[1] = v; \
00782 src += stride; \
00783 }
00784
00785 static void pred8x8l_128_dc_c(uint8_t *src, int has_topleft, int has_topright, int stride)
00786 {
00787 PREDICT_8x8_DC(0x80808080);
00788 }
00789 static void pred8x8l_left_dc_c(uint8_t *src, int has_topleft, int has_topright, int stride)
00790 {
00791 PREDICT_8x8_LOAD_LEFT;
00792 const uint32_t dc = ((l0+l1+l2+l3+l4+l5+l6+l7+4) >> 3) * 0x01010101;
00793 PREDICT_8x8_DC(dc);
00794 }
00795 static void pred8x8l_top_dc_c(uint8_t *src, int has_topleft, int has_topright, int stride)
00796 {
00797 PREDICT_8x8_LOAD_TOP;
00798 const uint32_t dc = ((t0+t1+t2+t3+t4+t5+t6+t7+4) >> 3) * 0x01010101;
00799 PREDICT_8x8_DC(dc);
00800 }
00801 static void pred8x8l_dc_c(uint8_t *src, int has_topleft, int has_topright, int stride)
00802 {
00803 PREDICT_8x8_LOAD_LEFT;
00804 PREDICT_8x8_LOAD_TOP;
00805 const uint32_t dc = ((l0+l1+l2+l3+l4+l5+l6+l7
00806 +t0+t1+t2+t3+t4+t5+t6+t7+8) >> 4) * 0x01010101;
00807 PREDICT_8x8_DC(dc);
00808 }
00809 static void pred8x8l_horizontal_c(uint8_t *src, int has_topleft, int has_topright, int stride)
00810 {
00811 PREDICT_8x8_LOAD_LEFT;
00812 #define ROW(y) ((uint32_t*)(src+y*stride))[0] =\
00813 ((uint32_t*)(src+y*stride))[1] = 0x01010101 * l##y
00814 ROW(0); ROW(1); ROW(2); ROW(3); ROW(4); ROW(5); ROW(6); ROW(7);
00815 #undef ROW
00816 }
00817 static void pred8x8l_vertical_c(uint8_t *src, int has_topleft, int has_topright, int stride)
00818 {
00819 int y;
00820 PREDICT_8x8_LOAD_TOP;
00821 src[0] = t0;
00822 src[1] = t1;
00823 src[2] = t2;
00824 src[3] = t3;
00825 src[4] = t4;
00826 src[5] = t5;
00827 src[6] = t6;
00828 src[7] = t7;
00829 for( y = 1; y < 8; y++ )
00830 *(uint64_t*)(src+y*stride) = *(uint64_t*)src;
00831 }
00832 static void pred8x8l_down_left_c(uint8_t *src, int has_topleft, int has_topright, int stride)
00833 {
00834 PREDICT_8x8_LOAD_TOP;
00835 PREDICT_8x8_LOAD_TOPRIGHT;
00836 SRC(0,0)= (t0 + 2*t1 + t2 + 2) >> 2;
00837 SRC(0,1)=SRC(1,0)= (t1 + 2*t2 + t3 + 2) >> 2;
00838 SRC(0,2)=SRC(1,1)=SRC(2,0)= (t2 + 2*t3 + t4 + 2) >> 2;
00839 SRC(0,3)=SRC(1,2)=SRC(2,1)=SRC(3,0)= (t3 + 2*t4 + t5 + 2) >> 2;
00840 SRC(0,4)=SRC(1,3)=SRC(2,2)=SRC(3,1)=SRC(4,0)= (t4 + 2*t5 + t6 + 2) >> 2;
00841 SRC(0,5)=SRC(1,4)=SRC(2,3)=SRC(3,2)=SRC(4,1)=SRC(5,0)= (t5 + 2*t6 + t7 + 2) >> 2;
00842 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;
00843 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;
00844 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;
00845 SRC(2,7)=SRC(3,6)=SRC(4,5)=SRC(5,4)=SRC(6,3)=SRC(7,2)= (t9 + 2*t10 + t11 + 2) >> 2;
00846 SRC(3,7)=SRC(4,6)=SRC(5,5)=SRC(6,4)=SRC(7,3)= (t10 + 2*t11 + t12 + 2) >> 2;
00847 SRC(4,7)=SRC(5,6)=SRC(6,5)=SRC(7,4)= (t11 + 2*t12 + t13 + 2) >> 2;
00848 SRC(5,7)=SRC(6,6)=SRC(7,5)= (t12 + 2*t13 + t14 + 2) >> 2;
00849 SRC(6,7)=SRC(7,6)= (t13 + 2*t14 + t15 + 2) >> 2;
00850 SRC(7,7)= (t14 + 3*t15 + 2) >> 2;
00851 }
00852 static void pred8x8l_down_right_c(uint8_t *src, int has_topleft, int has_topright, int stride)
00853 {
00854 PREDICT_8x8_LOAD_TOP;
00855 PREDICT_8x8_LOAD_LEFT;
00856 PREDICT_8x8_LOAD_TOPLEFT;
00857 SRC(0,7)= (l7 + 2*l6 + l5 + 2) >> 2;
00858 SRC(0,6)=SRC(1,7)= (l6 + 2*l5 + l4 + 2) >> 2;
00859 SRC(0,5)=SRC(1,6)=SRC(2,7)= (l5 + 2*l4 + l3 + 2) >> 2;
00860 SRC(0,4)=SRC(1,5)=SRC(2,6)=SRC(3,7)= (l4 + 2*l3 + l2 + 2) >> 2;
00861 SRC(0,3)=SRC(1,4)=SRC(2,5)=SRC(3,6)=SRC(4,7)= (l3 + 2*l2 + l1 + 2) >> 2;
00862 SRC(0,2)=SRC(1,3)=SRC(2,4)=SRC(3,5)=SRC(4,6)=SRC(5,7)= (l2 + 2*l1 + l0 + 2) >> 2;
00863 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;
00864 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;
00865 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;
00866 SRC(2,0)=SRC(3,1)=SRC(4,2)=SRC(5,3)=SRC(6,4)=SRC(7,5)= (t0 + 2*t1 + t2 + 2) >> 2;
00867 SRC(3,0)=SRC(4,1)=SRC(5,2)=SRC(6,3)=SRC(7,4)= (t1 + 2*t2 + t3 + 2) >> 2;
00868 SRC(4,0)=SRC(5,1)=SRC(6,2)=SRC(7,3)= (t2 + 2*t3 + t4 + 2) >> 2;
00869 SRC(5,0)=SRC(6,1)=SRC(7,2)= (t3 + 2*t4 + t5 + 2) >> 2;
00870 SRC(6,0)=SRC(7,1)= (t4 + 2*t5 + t6 + 2) >> 2;
00871 SRC(7,0)= (t5 + 2*t6 + t7 + 2) >> 2;
00872
00873 }
00874 static void pred8x8l_vertical_right_c(uint8_t *src, int has_topleft, int has_topright, int stride)
00875 {
00876 PREDICT_8x8_LOAD_TOP;
00877 PREDICT_8x8_LOAD_LEFT;
00878 PREDICT_8x8_LOAD_TOPLEFT;
00879 SRC(0,6)= (l5 + 2*l4 + l3 + 2) >> 2;
00880 SRC(0,7)= (l6 + 2*l5 + l4 + 2) >> 2;
00881 SRC(0,4)=SRC(1,6)= (l3 + 2*l2 + l1 + 2) >> 2;
00882 SRC(0,5)=SRC(1,7)= (l4 + 2*l3 + l2 + 2) >> 2;
00883 SRC(0,2)=SRC(1,4)=SRC(2,6)= (l1 + 2*l0 + lt + 2) >> 2;
00884 SRC(0,3)=SRC(1,5)=SRC(2,7)= (l2 + 2*l1 + l0 + 2) >> 2;
00885 SRC(0,1)=SRC(1,3)=SRC(2,5)=SRC(3,7)= (l0 + 2*lt + t0 + 2) >> 2;
00886 SRC(0,0)=SRC(1,2)=SRC(2,4)=SRC(3,6)= (lt + t0 + 1) >> 1;
00887 SRC(1,1)=SRC(2,3)=SRC(3,5)=SRC(4,7)= (lt + 2*t0 + t1 + 2) >> 2;
00888 SRC(1,0)=SRC(2,2)=SRC(3,4)=SRC(4,6)= (t0 + t1 + 1) >> 1;
00889 SRC(2,1)=SRC(3,3)=SRC(4,5)=SRC(5,7)= (t0 + 2*t1 + t2 + 2) >> 2;
00890 SRC(2,0)=SRC(3,2)=SRC(4,4)=SRC(5,6)= (t1 + t2 + 1) >> 1;
00891 SRC(3,1)=SRC(4,3)=SRC(5,5)=SRC(6,7)= (t1 + 2*t2 + t3 + 2) >> 2;
00892 SRC(3,0)=SRC(4,2)=SRC(5,4)=SRC(6,6)= (t2 + t3 + 1) >> 1;
00893 SRC(4,1)=SRC(5,3)=SRC(6,5)=SRC(7,7)= (t2 + 2*t3 + t4 + 2) >> 2;
00894 SRC(4,0)=SRC(5,2)=SRC(6,4)=SRC(7,6)= (t3 + t4 + 1) >> 1;
00895 SRC(5,1)=SRC(6,3)=SRC(7,5)= (t3 + 2*t4 + t5 + 2) >> 2;
00896 SRC(5,0)=SRC(6,2)=SRC(7,4)= (t4 + t5 + 1) >> 1;
00897 SRC(6,1)=SRC(7,3)= (t4 + 2*t5 + t6 + 2) >> 2;
00898 SRC(6,0)=SRC(7,2)= (t5 + t6 + 1) >> 1;
00899 SRC(7,1)= (t5 + 2*t6 + t7 + 2) >> 2;
00900 SRC(7,0)= (t6 + t7 + 1) >> 1;
00901 }
00902 static void pred8x8l_horizontal_down_c(uint8_t *src, int has_topleft, int has_topright, int stride)
00903 {
00904 PREDICT_8x8_LOAD_TOP;
00905 PREDICT_8x8_LOAD_LEFT;
00906 PREDICT_8x8_LOAD_TOPLEFT;
00907 SRC(0,7)= (l6 + l7 + 1) >> 1;
00908 SRC(1,7)= (l5 + 2*l6 + l7 + 2) >> 2;
00909 SRC(0,6)=SRC(2,7)= (l5 + l6 + 1) >> 1;
00910 SRC(1,6)=SRC(3,7)= (l4 + 2*l5 + l6 + 2) >> 2;
00911 SRC(0,5)=SRC(2,6)=SRC(4,7)= (l4 + l5 + 1) >> 1;
00912 SRC(1,5)=SRC(3,6)=SRC(5,7)= (l3 + 2*l4 + l5 + 2) >> 2;
00913 SRC(0,4)=SRC(2,5)=SRC(4,6)=SRC(6,7)= (l3 + l4 + 1) >> 1;
00914 SRC(1,4)=SRC(3,5)=SRC(5,6)=SRC(7,7)= (l2 + 2*l3 + l4 + 2) >> 2;
00915 SRC(0,3)=SRC(2,4)=SRC(4,5)=SRC(6,6)= (l2 + l3 + 1) >> 1;
00916 SRC(1,3)=SRC(3,4)=SRC(5,5)=SRC(7,6)= (l1 + 2*l2 + l3 + 2) >> 2;
00917 SRC(0,2)=SRC(2,3)=SRC(4,4)=SRC(6,5)= (l1 + l2 + 1) >> 1;
00918 SRC(1,2)=SRC(3,3)=SRC(5,4)=SRC(7,5)= (l0 + 2*l1 + l2 + 2) >> 2;
00919 SRC(0,1)=SRC(2,2)=SRC(4,3)=SRC(6,4)= (l0 + l1 + 1) >> 1;
00920 SRC(1,1)=SRC(3,2)=SRC(5,3)=SRC(7,4)= (lt + 2*l0 + l1 + 2) >> 2;
00921 SRC(0,0)=SRC(2,1)=SRC(4,2)=SRC(6,3)= (lt + l0 + 1) >> 1;
00922 SRC(1,0)=SRC(3,1)=SRC(5,2)=SRC(7,3)= (l0 + 2*lt + t0 + 2) >> 2;
00923 SRC(2,0)=SRC(4,1)=SRC(6,2)= (t1 + 2*t0 + lt + 2) >> 2;
00924 SRC(3,0)=SRC(5,1)=SRC(7,2)= (t2 + 2*t1 + t0 + 2) >> 2;
00925 SRC(4,0)=SRC(6,1)= (t3 + 2*t2 + t1 + 2) >> 2;
00926 SRC(5,0)=SRC(7,1)= (t4 + 2*t3 + t2 + 2) >> 2;
00927 SRC(6,0)= (t5 + 2*t4 + t3 + 2) >> 2;
00928 SRC(7,0)= (t6 + 2*t5 + t4 + 2) >> 2;
00929 }
00930 static void pred8x8l_vertical_left_c(uint8_t *src, int has_topleft, int has_topright, int stride)
00931 {
00932 PREDICT_8x8_LOAD_TOP;
00933 PREDICT_8x8_LOAD_TOPRIGHT;
00934 SRC(0,0)= (t0 + t1 + 1) >> 1;
00935 SRC(0,1)= (t0 + 2*t1 + t2 + 2) >> 2;
00936 SRC(0,2)=SRC(1,0)= (t1 + t2 + 1) >> 1;
00937 SRC(0,3)=SRC(1,1)= (t1 + 2*t2 + t3 + 2) >> 2;
00938 SRC(0,4)=SRC(1,2)=SRC(2,0)= (t2 + t3 + 1) >> 1;
00939 SRC(0,5)=SRC(1,3)=SRC(2,1)= (t2 + 2*t3 + t4 + 2) >> 2;
00940 SRC(0,6)=SRC(1,4)=SRC(2,2)=SRC(3,0)= (t3 + t4 + 1) >> 1;
00941 SRC(0,7)=SRC(1,5)=SRC(2,3)=SRC(3,1)= (t3 + 2*t4 + t5 + 2) >> 2;
00942 SRC(1,6)=SRC(2,4)=SRC(3,2)=SRC(4,0)= (t4 + t5 + 1) >> 1;
00943 SRC(1,7)=SRC(2,5)=SRC(3,3)=SRC(4,1)= (t4 + 2*t5 + t6 + 2) >> 2;
00944 SRC(2,6)=SRC(3,4)=SRC(4,2)=SRC(5,0)= (t5 + t6 + 1) >> 1;
00945 SRC(2,7)=SRC(3,5)=SRC(4,3)=SRC(5,1)= (t5 + 2*t6 + t7 + 2) >> 2;
00946 SRC(3,6)=SRC(4,4)=SRC(5,2)=SRC(6,0)= (t6 + t7 + 1) >> 1;
00947 SRC(3,7)=SRC(4,5)=SRC(5,3)=SRC(6,1)= (t6 + 2*t7 + t8 + 2) >> 2;
00948 SRC(4,6)=SRC(5,4)=SRC(6,2)=SRC(7,0)= (t7 + t8 + 1) >> 1;
00949 SRC(4,7)=SRC(5,5)=SRC(6,3)=SRC(7,1)= (t7 + 2*t8 + t9 + 2) >> 2;
00950 SRC(5,6)=SRC(6,4)=SRC(7,2)= (t8 + t9 + 1) >> 1;
00951 SRC(5,7)=SRC(6,5)=SRC(7,3)= (t8 + 2*t9 + t10 + 2) >> 2;
00952 SRC(6,6)=SRC(7,4)= (t9 + t10 + 1) >> 1;
00953 SRC(6,7)=SRC(7,5)= (t9 + 2*t10 + t11 + 2) >> 2;
00954 SRC(7,6)= (t10 + t11 + 1) >> 1;
00955 SRC(7,7)= (t10 + 2*t11 + t12 + 2) >> 2;
00956 }
00957 static void pred8x8l_horizontal_up_c(uint8_t *src, int has_topleft, int has_topright, int stride)
00958 {
00959 PREDICT_8x8_LOAD_LEFT;
00960 SRC(0,0)= (l0 + l1 + 1) >> 1;
00961 SRC(1,0)= (l0 + 2*l1 + l2 + 2) >> 2;
00962 SRC(0,1)=SRC(2,0)= (l1 + l2 + 1) >> 1;
00963 SRC(1,1)=SRC(3,0)= (l1 + 2*l2 + l3 + 2) >> 2;
00964 SRC(0,2)=SRC(2,1)=SRC(4,0)= (l2 + l3 + 1) >> 1;
00965 SRC(1,2)=SRC(3,1)=SRC(5,0)= (l2 + 2*l3 + l4 + 2) >> 2;
00966 SRC(0,3)=SRC(2,2)=SRC(4,1)=SRC(6,0)= (l3 + l4 + 1) >> 1;
00967 SRC(1,3)=SRC(3,2)=SRC(5,1)=SRC(7,0)= (l3 + 2*l4 + l5 + 2) >> 2;
00968 SRC(0,4)=SRC(2,3)=SRC(4,2)=SRC(6,1)= (l4 + l5 + 1) >> 1;
00969 SRC(1,4)=SRC(3,3)=SRC(5,2)=SRC(7,1)= (l4 + 2*l5 + l6 + 2) >> 2;
00970 SRC(0,5)=SRC(2,4)=SRC(4,3)=SRC(6,2)= (l5 + l6 + 1) >> 1;
00971 SRC(1,5)=SRC(3,4)=SRC(5,3)=SRC(7,2)= (l5 + 2*l6 + l7 + 2) >> 2;
00972 SRC(0,6)=SRC(2,5)=SRC(4,4)=SRC(6,3)= (l6 + l7 + 1) >> 1;
00973 SRC(1,6)=SRC(3,5)=SRC(5,4)=SRC(7,3)= (l6 + 3*l7 + 2) >> 2;
00974 SRC(0,7)=SRC(1,7)=SRC(2,6)=SRC(2,7)=SRC(3,6)=
00975 SRC(3,7)=SRC(4,5)=SRC(4,6)=SRC(4,7)=SRC(5,5)=
00976 SRC(5,6)=SRC(5,7)=SRC(6,4)=SRC(6,5)=SRC(6,6)=
00977 SRC(6,7)=SRC(7,4)=SRC(7,5)=SRC(7,6)=SRC(7,7)= l7;
00978 }
00979 #undef PREDICT_8x8_LOAD_LEFT
00980 #undef PREDICT_8x8_LOAD_TOP
00981 #undef PREDICT_8x8_LOAD_TOPLEFT
00982 #undef PREDICT_8x8_LOAD_TOPRIGHT
00983 #undef PREDICT_8x8_DC
00984 #undef PTR
00985 #undef PT
00986 #undef PL
00987 #undef SRC
00988
00992 void ff_h264_pred_init(H264PredContext *h, int codec_id){
00993
00994
00995 if(codec_id != CODEC_ID_RV40){
00996 h->pred4x4[VERT_PRED ]= pred4x4_vertical_c;
00997 h->pred4x4[HOR_PRED ]= pred4x4_horizontal_c;
00998 h->pred4x4[DC_PRED ]= pred4x4_dc_c;
00999 if(codec_id == CODEC_ID_SVQ3)
01000 h->pred4x4[DIAG_DOWN_LEFT_PRED ]= pred4x4_down_left_svq3_c;
01001 else
01002 h->pred4x4[DIAG_DOWN_LEFT_PRED ]= pred4x4_down_left_c;
01003 h->pred4x4[DIAG_DOWN_RIGHT_PRED]= pred4x4_down_right_c;
01004 h->pred4x4[VERT_RIGHT_PRED ]= pred4x4_vertical_right_c;
01005 h->pred4x4[HOR_DOWN_PRED ]= pred4x4_horizontal_down_c;
01006 h->pred4x4[VERT_LEFT_PRED ]= pred4x4_vertical_left_c;
01007 h->pred4x4[HOR_UP_PRED ]= pred4x4_horizontal_up_c;
01008 h->pred4x4[LEFT_DC_PRED ]= pred4x4_left_dc_c;
01009 h->pred4x4[TOP_DC_PRED ]= pred4x4_top_dc_c;
01010 h->pred4x4[DC_128_PRED ]= pred4x4_128_dc_c;
01011 }else{
01012 h->pred4x4[VERT_PRED ]= pred4x4_vertical_c;
01013 h->pred4x4[HOR_PRED ]= pred4x4_horizontal_c;
01014 h->pred4x4[DC_PRED ]= pred4x4_dc_c;
01015 h->pred4x4[DIAG_DOWN_LEFT_PRED ]= pred4x4_down_left_rv40_c;
01016 h->pred4x4[DIAG_DOWN_RIGHT_PRED]= pred4x4_down_right_c;
01017 h->pred4x4[VERT_RIGHT_PRED ]= pred4x4_vertical_right_c;
01018 h->pred4x4[HOR_DOWN_PRED ]= pred4x4_horizontal_down_c;
01019 h->pred4x4[VERT_LEFT_PRED ]= pred4x4_vertical_left_rv40_c;
01020 h->pred4x4[HOR_UP_PRED ]= pred4x4_horizontal_up_rv40_c;
01021 h->pred4x4[LEFT_DC_PRED ]= pred4x4_left_dc_c;
01022 h->pred4x4[TOP_DC_PRED ]= pred4x4_top_dc_c;
01023 h->pred4x4[DC_128_PRED ]= pred4x4_128_dc_c;
01024 h->pred4x4[DIAG_DOWN_LEFT_PRED_RV40_NODOWN]= pred4x4_down_left_rv40_nodown_c;
01025 h->pred4x4[HOR_UP_PRED_RV40_NODOWN]= pred4x4_horizontal_up_rv40_nodown_c;
01026 h->pred4x4[VERT_LEFT_PRED_RV40_NODOWN]= pred4x4_vertical_left_rv40_nodown_c;
01027 }
01028
01029 h->pred8x8l[VERT_PRED ]= pred8x8l_vertical_c;
01030 h->pred8x8l[HOR_PRED ]= pred8x8l_horizontal_c;
01031 h->pred8x8l[DC_PRED ]= pred8x8l_dc_c;
01032 h->pred8x8l[DIAG_DOWN_LEFT_PRED ]= pred8x8l_down_left_c;
01033 h->pred8x8l[DIAG_DOWN_RIGHT_PRED]= pred8x8l_down_right_c;
01034 h->pred8x8l[VERT_RIGHT_PRED ]= pred8x8l_vertical_right_c;
01035 h->pred8x8l[HOR_DOWN_PRED ]= pred8x8l_horizontal_down_c;
01036 h->pred8x8l[VERT_LEFT_PRED ]= pred8x8l_vertical_left_c;
01037 h->pred8x8l[HOR_UP_PRED ]= pred8x8l_horizontal_up_c;
01038 h->pred8x8l[LEFT_DC_PRED ]= pred8x8l_left_dc_c;
01039 h->pred8x8l[TOP_DC_PRED ]= pred8x8l_top_dc_c;
01040 h->pred8x8l[DC_128_PRED ]= pred8x8l_128_dc_c;
01041
01042 h->pred8x8[VERT_PRED8x8 ]= pred8x8_vertical_c;
01043 h->pred8x8[HOR_PRED8x8 ]= pred8x8_horizontal_c;
01044 h->pred8x8[PLANE_PRED8x8 ]= pred8x8_plane_c;
01045 if(codec_id != CODEC_ID_RV40){
01046 h->pred8x8[DC_PRED8x8 ]= pred8x8_dc_c;
01047 h->pred8x8[LEFT_DC_PRED8x8]= pred8x8_left_dc_c;
01048 h->pred8x8[TOP_DC_PRED8x8 ]= pred8x8_top_dc_c;
01049 }else{
01050 h->pred8x8[DC_PRED8x8 ]= pred8x8_dc_rv40_c;
01051 h->pred8x8[LEFT_DC_PRED8x8]= pred8x8_left_dc_rv40_c;
01052 h->pred8x8[TOP_DC_PRED8x8 ]= pred8x8_top_dc_rv40_c;
01053 }
01054 h->pred8x8[DC_128_PRED8x8 ]= pred8x8_128_dc_c;
01055
01056 h->pred16x16[DC_PRED8x8 ]= pred16x16_dc_c;
01057 h->pred16x16[VERT_PRED8x8 ]= pred16x16_vertical_c;
01058 h->pred16x16[HOR_PRED8x8 ]= pred16x16_horizontal_c;
01059 h->pred16x16[PLANE_PRED8x8 ]= pred16x16_plane_c;
01060 switch(codec_id){
01061 case CODEC_ID_SVQ3:
01062 h->pred16x16[PLANE_PRED8x8 ]= pred16x16_plane_svq3_c;
01063 break;
01064 case CODEC_ID_RV40:
01065 h->pred16x16[PLANE_PRED8x8 ]= pred16x16_plane_rv40_c;
01066 break;
01067 default:
01068 h->pred16x16[PLANE_PRED8x8 ]= pred16x16_plane_c;
01069 }
01070 h->pred16x16[LEFT_DC_PRED8x8]= pred16x16_left_dc_c;
01071 h->pred16x16[TOP_DC_PRED8x8 ]= pred16x16_top_dc_c;
01072 h->pred16x16[DC_128_PRED8x8 ]= pred16x16_128_dc_c;
01073 }