2
* H.26L/H.264/AVC/JVT/14496-10/... encoder/decoder
3
* Copyright (c) 2003 Michael Niedermayer <michaelni@gmx.at>
5
* This file is part of FFmpeg.
7
* FFmpeg is free software; you can redistribute it and/or
8
* modify it under the terms of the GNU Lesser General Public
9
* License as published by the Free Software Foundation; either
10
* version 2.1 of the License, or (at your option) any later version.
12
* FFmpeg is distributed in the hope that it will be useful,
13
* but WITHOUT ANY WARRANTY; without even the implied warranty of
14
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
15
* Lesser General Public License for more details.
17
* You should have received a copy of the GNU Lesser General Public
18
* License along with FFmpeg; if not, write to the Free Software
19
* Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
24
* H.264 / AVC / MPEG4 part10 prediction functions.
25
* @author Michael Niedermayer <michaelni@gmx.at>
29
#include "mpegvideo.h"
32
static void pred4x4_vertical_c(uint8_t *src, uint8_t *topright, int stride){
33
const uint32_t a= ((uint32_t*)(src-stride))[0];
34
((uint32_t*)(src+0*stride))[0]= a;
35
((uint32_t*)(src+1*stride))[0]= a;
36
((uint32_t*)(src+2*stride))[0]= a;
37
((uint32_t*)(src+3*stride))[0]= a;
40
static void pred4x4_horizontal_c(uint8_t *src, uint8_t *topright, int stride){
41
((uint32_t*)(src+0*stride))[0]= src[-1+0*stride]*0x01010101;
42
((uint32_t*)(src+1*stride))[0]= src[-1+1*stride]*0x01010101;
43
((uint32_t*)(src+2*stride))[0]= src[-1+2*stride]*0x01010101;
44
((uint32_t*)(src+3*stride))[0]= src[-1+3*stride]*0x01010101;
47
static void pred4x4_dc_c(uint8_t *src, uint8_t *topright, int stride){
48
const int dc= ( src[-stride] + src[1-stride] + src[2-stride] + src[3-stride]
49
+ src[-1+0*stride] + src[-1+1*stride] + src[-1+2*stride] + src[-1+3*stride] + 4) >>3;
51
((uint32_t*)(src+0*stride))[0]=
52
((uint32_t*)(src+1*stride))[0]=
53
((uint32_t*)(src+2*stride))[0]=
54
((uint32_t*)(src+3*stride))[0]= dc* 0x01010101;
57
static void pred4x4_left_dc_c(uint8_t *src, uint8_t *topright, int stride){
58
const int dc= ( src[-1+0*stride] + src[-1+1*stride] + src[-1+2*stride] + src[-1+3*stride] + 2) >>2;
60
((uint32_t*)(src+0*stride))[0]=
61
((uint32_t*)(src+1*stride))[0]=
62
((uint32_t*)(src+2*stride))[0]=
63
((uint32_t*)(src+3*stride))[0]= dc* 0x01010101;
66
static void pred4x4_top_dc_c(uint8_t *src, uint8_t *topright, int stride){
67
const int dc= ( src[-stride] + src[1-stride] + src[2-stride] + src[3-stride] + 2) >>2;
69
((uint32_t*)(src+0*stride))[0]=
70
((uint32_t*)(src+1*stride))[0]=
71
((uint32_t*)(src+2*stride))[0]=
72
((uint32_t*)(src+3*stride))[0]= dc* 0x01010101;
75
static void pred4x4_128_dc_c(uint8_t *src, uint8_t *topright, int stride){
76
((uint32_t*)(src+0*stride))[0]=
77
((uint32_t*)(src+1*stride))[0]=
78
((uint32_t*)(src+2*stride))[0]=
79
((uint32_t*)(src+3*stride))[0]= 128U*0x01010101U;
83
#define LOAD_TOP_RIGHT_EDGE\
84
const int av_unused t4= topright[0];\
85
const int av_unused t5= topright[1];\
86
const int av_unused t6= topright[2];\
87
const int av_unused t7= topright[3];\
89
#define LOAD_DOWN_LEFT_EDGE\
90
const int av_unused l4= src[-1+4*stride];\
91
const int av_unused l5= src[-1+5*stride];\
92
const int av_unused l6= src[-1+6*stride];\
93
const int av_unused l7= src[-1+7*stride];\
95
#define LOAD_LEFT_EDGE\
96
const int av_unused l0= src[-1+0*stride];\
97
const int av_unused l1= src[-1+1*stride];\
98
const int av_unused l2= src[-1+2*stride];\
99
const int av_unused l3= src[-1+3*stride];\
101
#define LOAD_TOP_EDGE\
102
const int av_unused t0= src[ 0-1*stride];\
103
const int av_unused t1= src[ 1-1*stride];\
104
const int av_unused t2= src[ 2-1*stride];\
105
const int av_unused t3= src[ 3-1*stride];\
107
static void pred4x4_down_right_c(uint8_t *src, uint8_t *topright, int stride){
108
const int lt= src[-1-1*stride];
112
src[0+3*stride]=(l3 + 2*l2 + l1 + 2)>>2;
114
src[1+3*stride]=(l2 + 2*l1 + l0 + 2)>>2;
117
src[2+3*stride]=(l1 + 2*l0 + lt + 2)>>2;
121
src[3+3*stride]=(l0 + 2*lt + t0 + 2)>>2;
124
src[3+2*stride]=(lt + 2*t0 + t1 + 2)>>2;
126
src[3+1*stride]=(t0 + 2*t1 + t2 + 2)>>2;
127
src[3+0*stride]=(t1 + 2*t2 + t3 + 2)>>2;
130
static void pred4x4_down_left_c(uint8_t *src, uint8_t *topright, int stride){
135
src[0+0*stride]=(t0 + t2 + 2*t1 + 2)>>2;
137
src[0+1*stride]=(t1 + t3 + 2*t2 + 2)>>2;
140
src[0+2*stride]=(t2 + t4 + 2*t3 + 2)>>2;
144
src[0+3*stride]=(t3 + t5 + 2*t4 + 2)>>2;
147
src[1+3*stride]=(t4 + t6 + 2*t5 + 2)>>2;
149
src[2+3*stride]=(t5 + t7 + 2*t6 + 2)>>2;
150
src[3+3*stride]=(t6 + 3*t7 + 2)>>2;
153
static void pred4x4_down_left_svq3_c(uint8_t *src, uint8_t *topright, int stride){
156
const av_unused int unu0= t0;
157
const av_unused int unu1= l0;
159
src[0+0*stride]=(l1 + t1)>>1;
161
src[0+1*stride]=(l2 + t2)>>1;
174
src[3+3*stride]=(l3 + t3)>>1;
177
static void pred4x4_down_left_rv40_c(uint8_t *src, uint8_t *topright, int stride){
183
src[0+0*stride]=(t0 + t2 + 2*t1 + 2 + l0 + l2 + 2*l1 + 2)>>3;
185
src[0+1*stride]=(t1 + t3 + 2*t2 + 2 + l1 + l3 + 2*l2 + 2)>>3;
188
src[0+2*stride]=(t2 + t4 + 2*t3 + 2 + l2 + l4 + 2*l3 + 2)>>3;
192
src[0+3*stride]=(t3 + t5 + 2*t4 + 2 + l3 + l5 + 2*l4 + 2)>>3;
195
src[1+3*stride]=(t4 + t6 + 2*t5 + 2 + l4 + l6 + 2*l5 + 2)>>3;
197
src[2+3*stride]=(t5 + t7 + 2*t6 + 2 + l5 + l7 + 2*l6 + 2)>>3;
198
src[3+3*stride]=(t6 + t7 + 1 + l6 + l7 + 1)>>2;
201
static void pred4x4_down_left_rv40_nodown_c(uint8_t *src, uint8_t *topright, int stride){
206
src[0+0*stride]=(t0 + t2 + 2*t1 + 2 + l0 + l2 + 2*l1 + 2)>>3;
208
src[0+1*stride]=(t1 + t3 + 2*t2 + 2 + l1 + l3 + 2*l2 + 2)>>3;
211
src[0+2*stride]=(t2 + t4 + 2*t3 + 2 + l2 + 3*l3 + 2)>>3;
215
src[0+3*stride]=(t3 + t5 + 2*t4 + 2 + l3*4 + 2)>>3;
218
src[1+3*stride]=(t4 + t6 + 2*t5 + 2 + l3*4 + 2)>>3;
220
src[2+3*stride]=(t5 + t7 + 2*t6 + 2 + l3*4 + 2)>>3;
221
src[3+3*stride]=(t6 + t7 + 1 + 2*l3 + 1)>>2;
224
static void pred4x4_vertical_right_c(uint8_t *src, uint8_t *topright, int stride){
225
const int lt= src[-1-1*stride];
230
src[1+2*stride]=(lt + t0 + 1)>>1;
232
src[2+2*stride]=(t0 + t1 + 1)>>1;
234
src[3+2*stride]=(t1 + t2 + 1)>>1;
235
src[3+0*stride]=(t2 + t3 + 1)>>1;
237
src[1+3*stride]=(l0 + 2*lt + t0 + 2)>>2;
239
src[2+3*stride]=(lt + 2*t0 + t1 + 2)>>2;
241
src[3+3*stride]=(t0 + 2*t1 + t2 + 2)>>2;
242
src[3+1*stride]=(t1 + 2*t2 + t3 + 2)>>2;
243
src[0+2*stride]=(lt + 2*l0 + l1 + 2)>>2;
244
src[0+3*stride]=(l0 + 2*l1 + l2 + 2)>>2;
247
static void pred4x4_vertical_left_c(uint8_t *src, uint8_t *topright, int stride){
251
src[0+0*stride]=(t0 + t1 + 1)>>1;
253
src[0+2*stride]=(t1 + t2 + 1)>>1;
255
src[1+2*stride]=(t2 + t3 + 1)>>1;
257
src[2+2*stride]=(t3 + t4+ 1)>>1;
258
src[3+2*stride]=(t4 + t5+ 1)>>1;
259
src[0+1*stride]=(t0 + 2*t1 + t2 + 2)>>2;
261
src[0+3*stride]=(t1 + 2*t2 + t3 + 2)>>2;
263
src[1+3*stride]=(t2 + 2*t3 + t4 + 2)>>2;
265
src[2+3*stride]=(t3 + 2*t4 + t5 + 2)>>2;
266
src[3+3*stride]=(t4 + 2*t5 + t6 + 2)>>2;
269
static void pred4x4_vertical_left_rv40(uint8_t *src, uint8_t *topright, int stride,
270
const int l0, const int l1, const int l2, const int l3, const int l4){
274
src[0+0*stride]=(2*t0 + 2*t1 + l1 + 2*l2 + l3 + 4)>>3;
276
src[0+2*stride]=(t1 + t2 + 1)>>1;
278
src[1+2*stride]=(t2 + t3 + 1)>>1;
280
src[2+2*stride]=(t3 + t4+ 1)>>1;
281
src[3+2*stride]=(t4 + t5+ 1)>>1;
282
src[0+1*stride]=(t0 + 2*t1 + t2 + l2 + 2*l3 + l4 + 4)>>3;
284
src[0+3*stride]=(t1 + 2*t2 + t3 + 2)>>2;
286
src[1+3*stride]=(t2 + 2*t3 + t4 + 2)>>2;
288
src[2+3*stride]=(t3 + 2*t4 + t5 + 2)>>2;
289
src[3+3*stride]=(t4 + 2*t5 + t6 + 2)>>2;
292
static void pred4x4_vertical_left_rv40_c(uint8_t *src, uint8_t *topright, int stride){
296
pred4x4_vertical_left_rv40(src, topright, stride, l0, l1, l2, l3, l4);
299
static void pred4x4_vertical_left_rv40_nodown_c(uint8_t *src, uint8_t *topright, int stride){
302
pred4x4_vertical_left_rv40(src, topright, stride, l0, l1, l2, l3, l3);
305
static void pred4x4_horizontal_up_c(uint8_t *src, uint8_t *topright, int stride){
308
src[0+0*stride]=(l0 + l1 + 1)>>1;
309
src[1+0*stride]=(l0 + 2*l1 + l2 + 2)>>2;
311
src[0+1*stride]=(l1 + l2 + 1)>>1;
313
src[1+1*stride]=(l1 + 2*l2 + l3 + 2)>>2;
315
src[0+2*stride]=(l2 + l3 + 1)>>1;
317
src[1+2*stride]=(l2 + 2*l3 + l3 + 2)>>2;
326
static void pred4x4_horizontal_up_rv40_c(uint8_t *src, uint8_t *topright, int stride){
332
src[0+0*stride]=(t1 + 2*t2 + t3 + 2*l0 + 2*l1 + 4)>>3;
333
src[1+0*stride]=(t2 + 2*t3 + t4 + l0 + 2*l1 + l2 + 4)>>3;
335
src[0+1*stride]=(t3 + 2*t4 + t5 + 2*l1 + 2*l2 + 4)>>3;
337
src[1+1*stride]=(t4 + 2*t5 + t6 + l1 + 2*l2 + l3 + 4)>>3;
339
src[0+2*stride]=(t5 + 2*t6 + t7 + 2*l2 + 2*l3 + 4)>>3;
341
src[1+2*stride]=(t6 + 3*t7 + l2 + 3*l3 + 4)>>3;
343
src[1+3*stride]=(l3 + 2*l4 + l5 + 2)>>2;
345
src[2+2*stride]=(t6 + t7 + l3 + l4 + 2)>>2;
346
src[2+3*stride]=(l4 + l5 + 1)>>1;
347
src[3+3*stride]=(l4 + 2*l5 + l6 + 2)>>2;
350
static void pred4x4_horizontal_up_rv40_nodown_c(uint8_t *src, uint8_t *topright, int stride){
355
src[0+0*stride]=(t1 + 2*t2 + t3 + 2*l0 + 2*l1 + 4)>>3;
356
src[1+0*stride]=(t2 + 2*t3 + t4 + l0 + 2*l1 + l2 + 4)>>3;
358
src[0+1*stride]=(t3 + 2*t4 + t5 + 2*l1 + 2*l2 + 4)>>3;
360
src[1+1*stride]=(t4 + 2*t5 + t6 + l1 + 2*l2 + l3 + 4)>>3;
362
src[0+2*stride]=(t5 + 2*t6 + t7 + 2*l2 + 2*l3 + 4)>>3;
364
src[1+2*stride]=(t6 + 3*t7 + l2 + 3*l3 + 4)>>3;
368
src[2+2*stride]=(t6 + t7 + 2*l3 + 2)>>2;
373
static void pred4x4_horizontal_down_c(uint8_t *src, uint8_t *topright, int stride){
374
const int lt= src[-1-1*stride];
379
src[2+1*stride]=(lt + l0 + 1)>>1;
381
src[3+1*stride]=(l0 + 2*lt + t0 + 2)>>2;
382
src[2+0*stride]=(lt + 2*t0 + t1 + 2)>>2;
383
src[3+0*stride]=(t0 + 2*t1 + t2 + 2)>>2;
385
src[2+2*stride]=(l0 + l1 + 1)>>1;
387
src[3+2*stride]=(lt + 2*l0 + l1 + 2)>>2;
389
src[2+3*stride]=(l1 + l2+ 1)>>1;
391
src[3+3*stride]=(l0 + 2*l1 + l2 + 2)>>2;
392
src[0+3*stride]=(l2 + l3 + 1)>>1;
393
src[1+3*stride]=(l1 + 2*l2 + l3 + 2)>>2;
396
static void pred16x16_vertical_c(uint8_t *src, int stride){
398
const uint32_t a= ((uint32_t*)(src-stride))[0];
399
const uint32_t b= ((uint32_t*)(src-stride))[1];
400
const uint32_t c= ((uint32_t*)(src-stride))[2];
401
const uint32_t d= ((uint32_t*)(src-stride))[3];
404
((uint32_t*)(src+i*stride))[0]= a;
405
((uint32_t*)(src+i*stride))[1]= b;
406
((uint32_t*)(src+i*stride))[2]= c;
407
((uint32_t*)(src+i*stride))[3]= d;
411
static void pred16x16_horizontal_c(uint8_t *src, int stride){
415
((uint32_t*)(src+i*stride))[0]=
416
((uint32_t*)(src+i*stride))[1]=
417
((uint32_t*)(src+i*stride))[2]=
418
((uint32_t*)(src+i*stride))[3]= src[-1+i*stride]*0x01010101;
422
static void pred16x16_dc_c(uint8_t *src, int stride){
426
dc+= src[-1+i*stride];
433
dc= 0x01010101*((dc + 16)>>5);
436
((uint32_t*)(src+i*stride))[0]=
437
((uint32_t*)(src+i*stride))[1]=
438
((uint32_t*)(src+i*stride))[2]=
439
((uint32_t*)(src+i*stride))[3]= dc;
443
static void pred16x16_left_dc_c(uint8_t *src, int stride){
447
dc+= src[-1+i*stride];
450
dc= 0x01010101*((dc + 8)>>4);
453
((uint32_t*)(src+i*stride))[0]=
454
((uint32_t*)(src+i*stride))[1]=
455
((uint32_t*)(src+i*stride))[2]=
456
((uint32_t*)(src+i*stride))[3]= dc;
460
static void pred16x16_top_dc_c(uint8_t *src, int stride){
466
dc= 0x01010101*((dc + 8)>>4);
469
((uint32_t*)(src+i*stride))[0]=
470
((uint32_t*)(src+i*stride))[1]=
471
((uint32_t*)(src+i*stride))[2]=
472
((uint32_t*)(src+i*stride))[3]= dc;
476
static void pred16x16_128_dc_c(uint8_t *src, int stride){
480
((uint32_t*)(src+i*stride))[0]=
481
((uint32_t*)(src+i*stride))[1]=
482
((uint32_t*)(src+i*stride))[2]=
483
((uint32_t*)(src+i*stride))[3]= 0x01010101U*128U;
487
static inline void pred16x16_plane_compat_c(uint8_t *src, int stride, const int svq3, const int rv40){
490
uint8_t *cm = ff_cropTbl + MAX_NEG_CROP;
491
const uint8_t * const src0 = src+7-stride;
492
const uint8_t *src1 = src+8*stride-1;
493
const uint8_t *src2 = src1-2*stride; // == src+6*stride-1;
494
int H = src0[1] - src0[-1];
495
int V = src1[0] - src2[ 0];
496
for(k=2; k<=8; ++k) {
497
src1 += stride; src2 -= stride;
498
H += k*(src0[k] - src0[-k]);
499
V += k*(src1[0] - src2[ 0]);
502
H = ( 5*(H/4) ) / 16;
503
V = ( 5*(V/4) ) / 16;
505
/* required for 100% accuracy */
508
H = ( H + (H>>2) ) >> 4;
509
V = ( V + (V>>2) ) >> 4;
515
a = 16*(src1[0] + src2[16] + 1) - 7*(V+H);
516
for(j=16; j>0; --j) {
519
for(i=-16; i<0; i+=4) {
520
src[16+i] = cm[ (b ) >> 5 ];
521
src[17+i] = cm[ (b+ H) >> 5 ];
522
src[18+i] = cm[ (b+2*H) >> 5 ];
523
src[19+i] = cm[ (b+3*H) >> 5 ];
530
static void pred16x16_plane_c(uint8_t *src, int stride){
531
pred16x16_plane_compat_c(src, stride, 0, 0);
534
static void pred16x16_plane_svq3_c(uint8_t *src, int stride){
535
pred16x16_plane_compat_c(src, stride, 1, 0);
538
static void pred16x16_plane_rv40_c(uint8_t *src, int stride){
539
pred16x16_plane_compat_c(src, stride, 0, 1);
542
static void pred8x8_vertical_c(uint8_t *src, int stride){
544
const uint32_t a= ((uint32_t*)(src-stride))[0];
545
const uint32_t b= ((uint32_t*)(src-stride))[1];
548
((uint32_t*)(src+i*stride))[0]= a;
549
((uint32_t*)(src+i*stride))[1]= b;
553
static void pred8x8_horizontal_c(uint8_t *src, int stride){
557
((uint32_t*)(src+i*stride))[0]=
558
((uint32_t*)(src+i*stride))[1]= src[-1+i*stride]*0x01010101;
562
static void pred8x8_128_dc_c(uint8_t *src, int stride){
566
((uint32_t*)(src+i*stride))[0]=
567
((uint32_t*)(src+i*stride))[1]= 0x01010101U*128U;
571
static void pred8x8_left_dc_c(uint8_t *src, int stride){
577
dc0+= src[-1+i*stride];
578
dc2+= src[-1+(i+4)*stride];
580
dc0= 0x01010101*((dc0 + 2)>>2);
581
dc2= 0x01010101*((dc2 + 2)>>2);
584
((uint32_t*)(src+i*stride))[0]=
585
((uint32_t*)(src+i*stride))[1]= dc0;
588
((uint32_t*)(src+i*stride))[0]=
589
((uint32_t*)(src+i*stride))[1]= dc2;
593
static void pred8x8_left_dc_rv40_c(uint8_t *src, int stride){
599
dc0+= src[-1+i*stride];
600
dc0= 0x01010101*((dc0 + 4)>>3);
603
((uint32_t*)(src+i*stride))[0]=
604
((uint32_t*)(src+i*stride))[1]= dc0;
608
static void pred8x8_top_dc_c(uint8_t *src, int stride){
615
dc1+= src[4+i-stride];
617
dc0= 0x01010101*((dc0 + 2)>>2);
618
dc1= 0x01010101*((dc1 + 2)>>2);
621
((uint32_t*)(src+i*stride))[0]= dc0;
622
((uint32_t*)(src+i*stride))[1]= dc1;
625
((uint32_t*)(src+i*stride))[0]= dc0;
626
((uint32_t*)(src+i*stride))[1]= dc1;
630
static void pred8x8_top_dc_rv40_c(uint8_t *src, int stride){
637
dc0= 0x01010101*((dc0 + 4)>>3);
640
((uint32_t*)(src+i*stride))[0]=
641
((uint32_t*)(src+i*stride))[1]= dc0;
646
static void pred8x8_dc_c(uint8_t *src, int stride){
648
int dc0, dc1, dc2, dc3;
652
dc0+= src[-1+i*stride] + src[i-stride];
653
dc1+= src[4+i-stride];
654
dc2+= src[-1+(i+4)*stride];
656
dc3= 0x01010101*((dc1 + dc2 + 4)>>3);
657
dc0= 0x01010101*((dc0 + 4)>>3);
658
dc1= 0x01010101*((dc1 + 2)>>2);
659
dc2= 0x01010101*((dc2 + 2)>>2);
662
((uint32_t*)(src+i*stride))[0]= dc0;
663
((uint32_t*)(src+i*stride))[1]= dc1;
666
((uint32_t*)(src+i*stride))[0]= dc2;
667
((uint32_t*)(src+i*stride))[1]= dc3;
671
//the following 4 function should not be optimized!
672
static void pred8x8_mad_cow_dc_l0t(uint8_t *src, int stride){
673
pred8x8_top_dc_c(src, stride);
674
pred4x4_dc_c(src, NULL, stride);
677
static void pred8x8_mad_cow_dc_0lt(uint8_t *src, int stride){
678
pred8x8_dc_c(src, stride);
679
pred4x4_top_dc_c(src, NULL, stride);
682
static void pred8x8_mad_cow_dc_l00(uint8_t *src, int stride){
683
pred8x8_left_dc_c(src, stride);
684
pred4x4_128_dc_c(src + 4*stride , NULL, stride);
685
pred4x4_128_dc_c(src + 4*stride + 4, NULL, stride);
688
static void pred8x8_mad_cow_dc_0l0(uint8_t *src, int stride){
689
pred8x8_left_dc_c(src, stride);
690
pred4x4_128_dc_c(src , NULL, stride);
691
pred4x4_128_dc_c(src + 4, NULL, stride);
694
static void pred8x8_dc_rv40_c(uint8_t *src, int stride){
699
dc0+= src[-1+i*stride] + src[i-stride];
700
dc0+= src[4+i-stride];
701
dc0+= src[-1+(i+4)*stride];
703
dc0= 0x01010101*((dc0 + 8)>>4);
706
((uint32_t*)(src+i*stride))[0]= dc0;
707
((uint32_t*)(src+i*stride))[1]= dc0;
710
((uint32_t*)(src+i*stride))[0]= dc0;
711
((uint32_t*)(src+i*stride))[1]= dc0;
715
static void pred8x8_plane_c(uint8_t *src, int stride){
718
uint8_t *cm = ff_cropTbl + MAX_NEG_CROP;
719
const uint8_t * const src0 = src+3-stride;
720
const uint8_t *src1 = src+4*stride-1;
721
const uint8_t *src2 = src1-2*stride; // == src+2*stride-1;
722
int H = src0[1] - src0[-1];
723
int V = src1[0] - src2[ 0];
724
for(k=2; k<=4; ++k) {
725
src1 += stride; src2 -= stride;
726
H += k*(src0[k] - src0[-k]);
727
V += k*(src1[0] - src2[ 0]);
729
H = ( 17*H+16 ) >> 5;
730
V = ( 17*V+16 ) >> 5;
732
a = 16*(src1[0] + src2[8]+1) - 3*(V+H);
736
src[0] = cm[ (b ) >> 5 ];
737
src[1] = cm[ (b+ H) >> 5 ];
738
src[2] = cm[ (b+2*H) >> 5 ];
739
src[3] = cm[ (b+3*H) >> 5 ];
740
src[4] = cm[ (b+4*H) >> 5 ];
741
src[5] = cm[ (b+5*H) >> 5 ];
742
src[6] = cm[ (b+6*H) >> 5 ];
743
src[7] = cm[ (b+7*H) >> 5 ];
748
#define SRC(x,y) src[(x)+(y)*stride]
750
const int l##y = (SRC(-1,y-1) + 2*SRC(-1,y) + SRC(-1,y+1) + 2) >> 2;
751
#define PREDICT_8x8_LOAD_LEFT \
752
const int l0 = ((has_topleft ? SRC(-1,-1) : SRC(-1,0)) \
753
+ 2*SRC(-1,0) + SRC(-1,1) + 2) >> 2; \
754
PL(1) PL(2) PL(3) PL(4) PL(5) PL(6) \
755
const int l7 av_unused = (SRC(-1,6) + 3*SRC(-1,7) + 2) >> 2
758
const int t##x = (SRC(x-1,-1) + 2*SRC(x,-1) + SRC(x+1,-1) + 2) >> 2;
759
#define PREDICT_8x8_LOAD_TOP \
760
const int t0 = ((has_topleft ? SRC(-1,-1) : SRC(0,-1)) \
761
+ 2*SRC(0,-1) + SRC(1,-1) + 2) >> 2; \
762
PT(1) PT(2) PT(3) PT(4) PT(5) PT(6) \
763
const int t7 av_unused = ((has_topright ? SRC(8,-1) : SRC(7,-1)) \
764
+ 2*SRC(7,-1) + SRC(6,-1) + 2) >> 2
767
t##x = (SRC(x-1,-1) + 2*SRC(x,-1) + SRC(x+1,-1) + 2) >> 2;
768
#define PREDICT_8x8_LOAD_TOPRIGHT \
769
int t8, t9, t10, t11, t12, t13, t14, t15; \
771
PTR(8) PTR(9) PTR(10) PTR(11) PTR(12) PTR(13) PTR(14) \
772
t15 = (SRC(14,-1) + 3*SRC(15,-1) + 2) >> 2; \
773
} else t8=t9=t10=t11=t12=t13=t14=t15= SRC(7,-1);
775
#define PREDICT_8x8_LOAD_TOPLEFT \
776
const int lt = (SRC(-1,0) + 2*SRC(-1,-1) + SRC(0,-1) + 2) >> 2
778
#define PREDICT_8x8_DC(v) \
780
for( y = 0; y < 8; y++ ) { \
781
((uint32_t*)src)[0] = \
782
((uint32_t*)src)[1] = v; \
786
static void pred8x8l_128_dc_c(uint8_t *src, int has_topleft, int has_topright, int stride)
788
PREDICT_8x8_DC(0x80808080);
790
static void pred8x8l_left_dc_c(uint8_t *src, int has_topleft, int has_topright, int stride)
792
PREDICT_8x8_LOAD_LEFT;
793
const uint32_t dc = ((l0+l1+l2+l3+l4+l5+l6+l7+4) >> 3) * 0x01010101;
796
static void pred8x8l_top_dc_c(uint8_t *src, int has_topleft, int has_topright, int stride)
798
PREDICT_8x8_LOAD_TOP;
799
const uint32_t dc = ((t0+t1+t2+t3+t4+t5+t6+t7+4) >> 3) * 0x01010101;
802
static void pred8x8l_dc_c(uint8_t *src, int has_topleft, int has_topright, int stride)
804
PREDICT_8x8_LOAD_LEFT;
805
PREDICT_8x8_LOAD_TOP;
806
const uint32_t dc = ((l0+l1+l2+l3+l4+l5+l6+l7
807
+t0+t1+t2+t3+t4+t5+t6+t7+8) >> 4) * 0x01010101;
810
static void pred8x8l_horizontal_c(uint8_t *src, int has_topleft, int has_topright, int stride)
812
PREDICT_8x8_LOAD_LEFT;
813
#define ROW(y) ((uint32_t*)(src+y*stride))[0] =\
814
((uint32_t*)(src+y*stride))[1] = 0x01010101 * l##y
815
ROW(0); ROW(1); ROW(2); ROW(3); ROW(4); ROW(5); ROW(6); ROW(7);
818
static void pred8x8l_vertical_c(uint8_t *src, int has_topleft, int has_topright, int stride)
821
PREDICT_8x8_LOAD_TOP;
830
for( y = 1; y < 8; y++ )
831
*(uint64_t*)(src+y*stride) = *(uint64_t*)src;
833
static void pred8x8l_down_left_c(uint8_t *src, int has_topleft, int has_topright, int stride)
835
PREDICT_8x8_LOAD_TOP;
836
PREDICT_8x8_LOAD_TOPRIGHT;
837
SRC(0,0)= (t0 + 2*t1 + t2 + 2) >> 2;
838
SRC(0,1)=SRC(1,0)= (t1 + 2*t2 + t3 + 2) >> 2;
839
SRC(0,2)=SRC(1,1)=SRC(2,0)= (t2 + 2*t3 + t4 + 2) >> 2;
840
SRC(0,3)=SRC(1,2)=SRC(2,1)=SRC(3,0)= (t3 + 2*t4 + t5 + 2) >> 2;
841
SRC(0,4)=SRC(1,3)=SRC(2,2)=SRC(3,1)=SRC(4,0)= (t4 + 2*t5 + t6 + 2) >> 2;
842
SRC(0,5)=SRC(1,4)=SRC(2,3)=SRC(3,2)=SRC(4,1)=SRC(5,0)= (t5 + 2*t6 + t7 + 2) >> 2;
843
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;
844
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;
845
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;
846
SRC(2,7)=SRC(3,6)=SRC(4,5)=SRC(5,4)=SRC(6,3)=SRC(7,2)= (t9 + 2*t10 + t11 + 2) >> 2;
847
SRC(3,7)=SRC(4,6)=SRC(5,5)=SRC(6,4)=SRC(7,3)= (t10 + 2*t11 + t12 + 2) >> 2;
848
SRC(4,7)=SRC(5,6)=SRC(6,5)=SRC(7,4)= (t11 + 2*t12 + t13 + 2) >> 2;
849
SRC(5,7)=SRC(6,6)=SRC(7,5)= (t12 + 2*t13 + t14 + 2) >> 2;
850
SRC(6,7)=SRC(7,6)= (t13 + 2*t14 + t15 + 2) >> 2;
851
SRC(7,7)= (t14 + 3*t15 + 2) >> 2;
853
static void pred8x8l_down_right_c(uint8_t *src, int has_topleft, int has_topright, int stride)
855
PREDICT_8x8_LOAD_TOP;
856
PREDICT_8x8_LOAD_LEFT;
857
PREDICT_8x8_LOAD_TOPLEFT;
858
SRC(0,7)= (l7 + 2*l6 + l5 + 2) >> 2;
859
SRC(0,6)=SRC(1,7)= (l6 + 2*l5 + l4 + 2) >> 2;
860
SRC(0,5)=SRC(1,6)=SRC(2,7)= (l5 + 2*l4 + l3 + 2) >> 2;
861
SRC(0,4)=SRC(1,5)=SRC(2,6)=SRC(3,7)= (l4 + 2*l3 + l2 + 2) >> 2;
862
SRC(0,3)=SRC(1,4)=SRC(2,5)=SRC(3,6)=SRC(4,7)= (l3 + 2*l2 + l1 + 2) >> 2;
863
SRC(0,2)=SRC(1,3)=SRC(2,4)=SRC(3,5)=SRC(4,6)=SRC(5,7)= (l2 + 2*l1 + l0 + 2) >> 2;
864
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;
865
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;
866
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;
867
SRC(2,0)=SRC(3,1)=SRC(4,2)=SRC(5,3)=SRC(6,4)=SRC(7,5)= (t0 + 2*t1 + t2 + 2) >> 2;
868
SRC(3,0)=SRC(4,1)=SRC(5,2)=SRC(6,3)=SRC(7,4)= (t1 + 2*t2 + t3 + 2) >> 2;
869
SRC(4,0)=SRC(5,1)=SRC(6,2)=SRC(7,3)= (t2 + 2*t3 + t4 + 2) >> 2;
870
SRC(5,0)=SRC(6,1)=SRC(7,2)= (t3 + 2*t4 + t5 + 2) >> 2;
871
SRC(6,0)=SRC(7,1)= (t4 + 2*t5 + t6 + 2) >> 2;
872
SRC(7,0)= (t5 + 2*t6 + t7 + 2) >> 2;
875
static void pred8x8l_vertical_right_c(uint8_t *src, int has_topleft, int has_topright, int stride)
877
PREDICT_8x8_LOAD_TOP;
878
PREDICT_8x8_LOAD_LEFT;
879
PREDICT_8x8_LOAD_TOPLEFT;
880
SRC(0,6)= (l5 + 2*l4 + l3 + 2) >> 2;
881
SRC(0,7)= (l6 + 2*l5 + l4 + 2) >> 2;
882
SRC(0,4)=SRC(1,6)= (l3 + 2*l2 + l1 + 2) >> 2;
883
SRC(0,5)=SRC(1,7)= (l4 + 2*l3 + l2 + 2) >> 2;
884
SRC(0,2)=SRC(1,4)=SRC(2,6)= (l1 + 2*l0 + lt + 2) >> 2;
885
SRC(0,3)=SRC(1,5)=SRC(2,7)= (l2 + 2*l1 + l0 + 2) >> 2;
886
SRC(0,1)=SRC(1,3)=SRC(2,5)=SRC(3,7)= (l0 + 2*lt + t0 + 2) >> 2;
887
SRC(0,0)=SRC(1,2)=SRC(2,4)=SRC(3,6)= (lt + t0 + 1) >> 1;
888
SRC(1,1)=SRC(2,3)=SRC(3,5)=SRC(4,7)= (lt + 2*t0 + t1 + 2) >> 2;
889
SRC(1,0)=SRC(2,2)=SRC(3,4)=SRC(4,6)= (t0 + t1 + 1) >> 1;
890
SRC(2,1)=SRC(3,3)=SRC(4,5)=SRC(5,7)= (t0 + 2*t1 + t2 + 2) >> 2;
891
SRC(2,0)=SRC(3,2)=SRC(4,4)=SRC(5,6)= (t1 + t2 + 1) >> 1;
892
SRC(3,1)=SRC(4,3)=SRC(5,5)=SRC(6,7)= (t1 + 2*t2 + t3 + 2) >> 2;
893
SRC(3,0)=SRC(4,2)=SRC(5,4)=SRC(6,6)= (t2 + t3 + 1) >> 1;
894
SRC(4,1)=SRC(5,3)=SRC(6,5)=SRC(7,7)= (t2 + 2*t3 + t4 + 2) >> 2;
895
SRC(4,0)=SRC(5,2)=SRC(6,4)=SRC(7,6)= (t3 + t4 + 1) >> 1;
896
SRC(5,1)=SRC(6,3)=SRC(7,5)= (t3 + 2*t4 + t5 + 2) >> 2;
897
SRC(5,0)=SRC(6,2)=SRC(7,4)= (t4 + t5 + 1) >> 1;
898
SRC(6,1)=SRC(7,3)= (t4 + 2*t5 + t6 + 2) >> 2;
899
SRC(6,0)=SRC(7,2)= (t5 + t6 + 1) >> 1;
900
SRC(7,1)= (t5 + 2*t6 + t7 + 2) >> 2;
901
SRC(7,0)= (t6 + t7 + 1) >> 1;
903
static void pred8x8l_horizontal_down_c(uint8_t *src, int has_topleft, int has_topright, int stride)
905
PREDICT_8x8_LOAD_TOP;
906
PREDICT_8x8_LOAD_LEFT;
907
PREDICT_8x8_LOAD_TOPLEFT;
908
SRC(0,7)= (l6 + l7 + 1) >> 1;
909
SRC(1,7)= (l5 + 2*l6 + l7 + 2) >> 2;
910
SRC(0,6)=SRC(2,7)= (l5 + l6 + 1) >> 1;
911
SRC(1,6)=SRC(3,7)= (l4 + 2*l5 + l6 + 2) >> 2;
912
SRC(0,5)=SRC(2,6)=SRC(4,7)= (l4 + l5 + 1) >> 1;
913
SRC(1,5)=SRC(3,6)=SRC(5,7)= (l3 + 2*l4 + l5 + 2) >> 2;
914
SRC(0,4)=SRC(2,5)=SRC(4,6)=SRC(6,7)= (l3 + l4 + 1) >> 1;
915
SRC(1,4)=SRC(3,5)=SRC(5,6)=SRC(7,7)= (l2 + 2*l3 + l4 + 2) >> 2;
916
SRC(0,3)=SRC(2,4)=SRC(4,5)=SRC(6,6)= (l2 + l3 + 1) >> 1;
917
SRC(1,3)=SRC(3,4)=SRC(5,5)=SRC(7,6)= (l1 + 2*l2 + l3 + 2) >> 2;
918
SRC(0,2)=SRC(2,3)=SRC(4,4)=SRC(6,5)= (l1 + l2 + 1) >> 1;
919
SRC(1,2)=SRC(3,3)=SRC(5,4)=SRC(7,5)= (l0 + 2*l1 + l2 + 2) >> 2;
920
SRC(0,1)=SRC(2,2)=SRC(4,3)=SRC(6,4)= (l0 + l1 + 1) >> 1;
921
SRC(1,1)=SRC(3,2)=SRC(5,3)=SRC(7,4)= (lt + 2*l0 + l1 + 2) >> 2;
922
SRC(0,0)=SRC(2,1)=SRC(4,2)=SRC(6,3)= (lt + l0 + 1) >> 1;
923
SRC(1,0)=SRC(3,1)=SRC(5,2)=SRC(7,3)= (l0 + 2*lt + t0 + 2) >> 2;
924
SRC(2,0)=SRC(4,1)=SRC(6,2)= (t1 + 2*t0 + lt + 2) >> 2;
925
SRC(3,0)=SRC(5,1)=SRC(7,2)= (t2 + 2*t1 + t0 + 2) >> 2;
926
SRC(4,0)=SRC(6,1)= (t3 + 2*t2 + t1 + 2) >> 2;
927
SRC(5,0)=SRC(7,1)= (t4 + 2*t3 + t2 + 2) >> 2;
928
SRC(6,0)= (t5 + 2*t4 + t3 + 2) >> 2;
929
SRC(7,0)= (t6 + 2*t5 + t4 + 2) >> 2;
931
static void pred8x8l_vertical_left_c(uint8_t *src, int has_topleft, int has_topright, int stride)
933
PREDICT_8x8_LOAD_TOP;
934
PREDICT_8x8_LOAD_TOPRIGHT;
935
SRC(0,0)= (t0 + t1 + 1) >> 1;
936
SRC(0,1)= (t0 + 2*t1 + t2 + 2) >> 2;
937
SRC(0,2)=SRC(1,0)= (t1 + t2 + 1) >> 1;
938
SRC(0,3)=SRC(1,1)= (t1 + 2*t2 + t3 + 2) >> 2;
939
SRC(0,4)=SRC(1,2)=SRC(2,0)= (t2 + t3 + 1) >> 1;
940
SRC(0,5)=SRC(1,3)=SRC(2,1)= (t2 + 2*t3 + t4 + 2) >> 2;
941
SRC(0,6)=SRC(1,4)=SRC(2,2)=SRC(3,0)= (t3 + t4 + 1) >> 1;
942
SRC(0,7)=SRC(1,5)=SRC(2,3)=SRC(3,1)= (t3 + 2*t4 + t5 + 2) >> 2;
943
SRC(1,6)=SRC(2,4)=SRC(3,2)=SRC(4,0)= (t4 + t5 + 1) >> 1;
944
SRC(1,7)=SRC(2,5)=SRC(3,3)=SRC(4,1)= (t4 + 2*t5 + t6 + 2) >> 2;
945
SRC(2,6)=SRC(3,4)=SRC(4,2)=SRC(5,0)= (t5 + t6 + 1) >> 1;
946
SRC(2,7)=SRC(3,5)=SRC(4,3)=SRC(5,1)= (t5 + 2*t6 + t7 + 2) >> 2;
947
SRC(3,6)=SRC(4,4)=SRC(5,2)=SRC(6,0)= (t6 + t7 + 1) >> 1;
948
SRC(3,7)=SRC(4,5)=SRC(5,3)=SRC(6,1)= (t6 + 2*t7 + t8 + 2) >> 2;
949
SRC(4,6)=SRC(5,4)=SRC(6,2)=SRC(7,0)= (t7 + t8 + 1) >> 1;
950
SRC(4,7)=SRC(5,5)=SRC(6,3)=SRC(7,1)= (t7 + 2*t8 + t9 + 2) >> 2;
951
SRC(5,6)=SRC(6,4)=SRC(7,2)= (t8 + t9 + 1) >> 1;
952
SRC(5,7)=SRC(6,5)=SRC(7,3)= (t8 + 2*t9 + t10 + 2) >> 2;
953
SRC(6,6)=SRC(7,4)= (t9 + t10 + 1) >> 1;
954
SRC(6,7)=SRC(7,5)= (t9 + 2*t10 + t11 + 2) >> 2;
955
SRC(7,6)= (t10 + t11 + 1) >> 1;
956
SRC(7,7)= (t10 + 2*t11 + t12 + 2) >> 2;
958
static void pred8x8l_horizontal_up_c(uint8_t *src, int has_topleft, int has_topright, int stride)
960
PREDICT_8x8_LOAD_LEFT;
961
SRC(0,0)= (l0 + l1 + 1) >> 1;
962
SRC(1,0)= (l0 + 2*l1 + l2 + 2) >> 2;
963
SRC(0,1)=SRC(2,0)= (l1 + l2 + 1) >> 1;
964
SRC(1,1)=SRC(3,0)= (l1 + 2*l2 + l3 + 2) >> 2;
965
SRC(0,2)=SRC(2,1)=SRC(4,0)= (l2 + l3 + 1) >> 1;
966
SRC(1,2)=SRC(3,1)=SRC(5,0)= (l2 + 2*l3 + l4 + 2) >> 2;
967
SRC(0,3)=SRC(2,2)=SRC(4,1)=SRC(6,0)= (l3 + l4 + 1) >> 1;
968
SRC(1,3)=SRC(3,2)=SRC(5,1)=SRC(7,0)= (l3 + 2*l4 + l5 + 2) >> 2;
969
SRC(0,4)=SRC(2,3)=SRC(4,2)=SRC(6,1)= (l4 + l5 + 1) >> 1;
970
SRC(1,4)=SRC(3,3)=SRC(5,2)=SRC(7,1)= (l4 + 2*l5 + l6 + 2) >> 2;
971
SRC(0,5)=SRC(2,4)=SRC(4,3)=SRC(6,2)= (l5 + l6 + 1) >> 1;
972
SRC(1,5)=SRC(3,4)=SRC(5,3)=SRC(7,2)= (l5 + 2*l6 + l7 + 2) >> 2;
973
SRC(0,6)=SRC(2,5)=SRC(4,4)=SRC(6,3)= (l6 + l7 + 1) >> 1;
974
SRC(1,6)=SRC(3,5)=SRC(5,4)=SRC(7,3)= (l6 + 3*l7 + 2) >> 2;
975
SRC(0,7)=SRC(1,7)=SRC(2,6)=SRC(2,7)=SRC(3,6)=
976
SRC(3,7)=SRC(4,5)=SRC(4,6)=SRC(4,7)=SRC(5,5)=
977
SRC(5,6)=SRC(5,7)=SRC(6,4)=SRC(6,5)=SRC(6,6)=
978
SRC(6,7)=SRC(7,4)=SRC(7,5)=SRC(7,6)=SRC(7,7)= l7;
980
#undef PREDICT_8x8_LOAD_LEFT
981
#undef PREDICT_8x8_LOAD_TOP
982
#undef PREDICT_8x8_LOAD_TOPLEFT
983
#undef PREDICT_8x8_LOAD_TOPRIGHT
984
#undef PREDICT_8x8_DC
990
static void pred4x4_vertical_add_c(uint8_t *pix, const DCTELEM *block, int stride){
995
pix[1*stride]= v += block[0];
996
pix[2*stride]= v += block[4];
997
pix[3*stride]= v += block[8];
998
pix[4*stride]= v + block[12];
1004
static void pred4x4_horizontal_add_c(uint8_t *pix, const DCTELEM *block, int stride){
1007
uint8_t v = pix[-1];
1008
pix[0]= v += block[0];
1009
pix[1]= v += block[1];
1010
pix[2]= v += block[2];
1011
pix[3]= v + block[3];
1017
static void pred8x8l_vertical_add_c(uint8_t *pix, const DCTELEM *block, int stride){
1022
pix[1*stride]= v += block[0];
1023
pix[2*stride]= v += block[8];
1024
pix[3*stride]= v += block[16];
1025
pix[4*stride]= v += block[24];
1026
pix[5*stride]= v += block[32];
1027
pix[6*stride]= v += block[40];
1028
pix[7*stride]= v += block[48];
1029
pix[8*stride]= v + block[56];
1035
static void pred8x8l_horizontal_add_c(uint8_t *pix, const DCTELEM *block, int stride){
1038
uint8_t v = pix[-1];
1039
pix[0]= v += block[0];
1040
pix[1]= v += block[1];
1041
pix[2]= v += block[2];
1042
pix[3]= v += block[3];
1043
pix[4]= v += block[4];
1044
pix[5]= v += block[5];
1045
pix[6]= v += block[6];
1046
pix[7]= v + block[7];
1052
static void pred16x16_vertical_add_c(uint8_t *pix, const int *block_offset, const DCTELEM *block, int stride){
1055
pred4x4_vertical_add_c(pix + block_offset[i], block + i*16, stride);
1058
static void pred16x16_horizontal_add_c(uint8_t *pix, const int *block_offset, const DCTELEM *block, int stride){
1061
pred4x4_horizontal_add_c(pix + block_offset[i], block + i*16, stride);
1064
static void pred8x8_vertical_add_c(uint8_t *pix, const int *block_offset, const DCTELEM *block, int stride){
1067
pred4x4_vertical_add_c(pix + block_offset[i], block + i*16, stride);
1070
static void pred8x8_horizontal_add_c(uint8_t *pix, const int *block_offset, const DCTELEM *block, int stride){
1073
pred4x4_horizontal_add_c(pix + block_offset[i], block + i*16, stride);
1078
* Sets the intra prediction function pointers.
1080
void ff_h264_pred_init(H264PredContext *h, int codec_id){
1081
// MpegEncContext * const s = &h->s;
1083
if(codec_id != CODEC_ID_RV40){
1084
h->pred4x4[VERT_PRED ]= pred4x4_vertical_c;
1085
h->pred4x4[HOR_PRED ]= pred4x4_horizontal_c;
1086
h->pred4x4[DC_PRED ]= pred4x4_dc_c;
1087
if(codec_id == CODEC_ID_SVQ3)
1088
h->pred4x4[DIAG_DOWN_LEFT_PRED ]= pred4x4_down_left_svq3_c;
1090
h->pred4x4[DIAG_DOWN_LEFT_PRED ]= pred4x4_down_left_c;
1091
h->pred4x4[DIAG_DOWN_RIGHT_PRED]= pred4x4_down_right_c;
1092
h->pred4x4[VERT_RIGHT_PRED ]= pred4x4_vertical_right_c;
1093
h->pred4x4[HOR_DOWN_PRED ]= pred4x4_horizontal_down_c;
1094
h->pred4x4[VERT_LEFT_PRED ]= pred4x4_vertical_left_c;
1095
h->pred4x4[HOR_UP_PRED ]= pred4x4_horizontal_up_c;
1096
h->pred4x4[LEFT_DC_PRED ]= pred4x4_left_dc_c;
1097
h->pred4x4[TOP_DC_PRED ]= pred4x4_top_dc_c;
1098
h->pred4x4[DC_128_PRED ]= pred4x4_128_dc_c;
1100
h->pred4x4[VERT_PRED ]= pred4x4_vertical_c;
1101
h->pred4x4[HOR_PRED ]= pred4x4_horizontal_c;
1102
h->pred4x4[DC_PRED ]= pred4x4_dc_c;
1103
h->pred4x4[DIAG_DOWN_LEFT_PRED ]= pred4x4_down_left_rv40_c;
1104
h->pred4x4[DIAG_DOWN_RIGHT_PRED]= pred4x4_down_right_c;
1105
h->pred4x4[VERT_RIGHT_PRED ]= pred4x4_vertical_right_c;
1106
h->pred4x4[HOR_DOWN_PRED ]= pred4x4_horizontal_down_c;
1107
h->pred4x4[VERT_LEFT_PRED ]= pred4x4_vertical_left_rv40_c;
1108
h->pred4x4[HOR_UP_PRED ]= pred4x4_horizontal_up_rv40_c;
1109
h->pred4x4[LEFT_DC_PRED ]= pred4x4_left_dc_c;
1110
h->pred4x4[TOP_DC_PRED ]= pred4x4_top_dc_c;
1111
h->pred4x4[DC_128_PRED ]= pred4x4_128_dc_c;
1112
h->pred4x4[DIAG_DOWN_LEFT_PRED_RV40_NODOWN]= pred4x4_down_left_rv40_nodown_c;
1113
h->pred4x4[HOR_UP_PRED_RV40_NODOWN]= pred4x4_horizontal_up_rv40_nodown_c;
1114
h->pred4x4[VERT_LEFT_PRED_RV40_NODOWN]= pred4x4_vertical_left_rv40_nodown_c;
1117
h->pred8x8l[VERT_PRED ]= pred8x8l_vertical_c;
1118
h->pred8x8l[HOR_PRED ]= pred8x8l_horizontal_c;
1119
h->pred8x8l[DC_PRED ]= pred8x8l_dc_c;
1120
h->pred8x8l[DIAG_DOWN_LEFT_PRED ]= pred8x8l_down_left_c;
1121
h->pred8x8l[DIAG_DOWN_RIGHT_PRED]= pred8x8l_down_right_c;
1122
h->pred8x8l[VERT_RIGHT_PRED ]= pred8x8l_vertical_right_c;
1123
h->pred8x8l[HOR_DOWN_PRED ]= pred8x8l_horizontal_down_c;
1124
h->pred8x8l[VERT_LEFT_PRED ]= pred8x8l_vertical_left_c;
1125
h->pred8x8l[HOR_UP_PRED ]= pred8x8l_horizontal_up_c;
1126
h->pred8x8l[LEFT_DC_PRED ]= pred8x8l_left_dc_c;
1127
h->pred8x8l[TOP_DC_PRED ]= pred8x8l_top_dc_c;
1128
h->pred8x8l[DC_128_PRED ]= pred8x8l_128_dc_c;
1130
h->pred8x8[VERT_PRED8x8 ]= pred8x8_vertical_c;
1131
h->pred8x8[HOR_PRED8x8 ]= pred8x8_horizontal_c;
1132
h->pred8x8[PLANE_PRED8x8 ]= pred8x8_plane_c;
1133
if(codec_id != CODEC_ID_RV40){
1134
h->pred8x8[DC_PRED8x8 ]= pred8x8_dc_c;
1135
h->pred8x8[LEFT_DC_PRED8x8]= pred8x8_left_dc_c;
1136
h->pred8x8[TOP_DC_PRED8x8 ]= pred8x8_top_dc_c;
1137
h->pred8x8[ALZHEIMER_DC_L0T_PRED8x8 ]= pred8x8_mad_cow_dc_l0t;
1138
h->pred8x8[ALZHEIMER_DC_0LT_PRED8x8 ]= pred8x8_mad_cow_dc_0lt;
1139
h->pred8x8[ALZHEIMER_DC_L00_PRED8x8 ]= pred8x8_mad_cow_dc_l00;
1140
h->pred8x8[ALZHEIMER_DC_0L0_PRED8x8 ]= pred8x8_mad_cow_dc_0l0;
1142
h->pred8x8[DC_PRED8x8 ]= pred8x8_dc_rv40_c;
1143
h->pred8x8[LEFT_DC_PRED8x8]= pred8x8_left_dc_rv40_c;
1144
h->pred8x8[TOP_DC_PRED8x8 ]= pred8x8_top_dc_rv40_c;
1146
h->pred8x8[DC_128_PRED8x8 ]= pred8x8_128_dc_c;
1148
h->pred16x16[DC_PRED8x8 ]= pred16x16_dc_c;
1149
h->pred16x16[VERT_PRED8x8 ]= pred16x16_vertical_c;
1150
h->pred16x16[HOR_PRED8x8 ]= pred16x16_horizontal_c;
1151
h->pred16x16[PLANE_PRED8x8 ]= pred16x16_plane_c;
1154
h->pred16x16[PLANE_PRED8x8 ]= pred16x16_plane_svq3_c;
1157
h->pred16x16[PLANE_PRED8x8 ]= pred16x16_plane_rv40_c;
1160
h->pred16x16[PLANE_PRED8x8 ]= pred16x16_plane_c;
1162
h->pred16x16[LEFT_DC_PRED8x8]= pred16x16_left_dc_c;
1163
h->pred16x16[TOP_DC_PRED8x8 ]= pred16x16_top_dc_c;
1164
h->pred16x16[DC_128_PRED8x8 ]= pred16x16_128_dc_c;
1166
//special lossless h/v prediction for h264
1167
h->pred4x4_add [VERT_PRED ]= pred4x4_vertical_add_c;
1168
h->pred4x4_add [ HOR_PRED ]= pred4x4_horizontal_add_c;
1169
h->pred8x8l_add [VERT_PRED ]= pred8x8l_vertical_add_c;
1170
h->pred8x8l_add [ HOR_PRED ]= pred8x8l_horizontal_add_c;
1171
h->pred8x8_add [VERT_PRED8x8]= pred8x8_vertical_add_c;
1172
h->pred8x8_add [ HOR_PRED8x8]= pred8x8_horizontal_add_c;
1173
h->pred16x16_add[VERT_PRED8x8]= pred16x16_vertical_add_c;
1174
h->pred16x16_add[ HOR_PRED8x8]= pred16x16_horizontal_add_c;
1176
if (ARCH_ARM) ff_h264_pred_init_arm(h, codec_id);