summaryrefslogtreecommitdiff
path: root/libavcodec/snow.c
blob: 101c8f83e5c19e0cb98304ba00a48ff05b41db05 (plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
/*
 * Copyright (C) 2004 Michael Niedermayer <michaelni@gmx.at>
 *
 * This file is part of FFmpeg.
 *
 * FFmpeg is free software; you can redistribute it and/or
 * modify it under the terms of the GNU Lesser General Public
 * License as published by the Free Software Foundation; either
 * version 2.1 of the License, or (at your option) any later version.
 *
 * FFmpeg is distributed in the hope that it will be useful,
 * but WITHOUT ANY WARRANTY; without even the implied warranty of
 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
 * Lesser General Public License for more details.
 *
 * You should have received a copy of the GNU Lesser General Public
 * License along with FFmpeg; if not, write to the Free Software
 * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
 */

#include "libavutil/intmath.h"
#include "libavutil/log.h"
#include "libavutil/opt.h"
#include "avcodec.h"
#include "me_cmp.h"
#include "snow_dwt.h"
#include "internal.h"
#include "snow.h"
#include "snowdata.h"

#include "rangecoder.h"
#include "mathops.h"
#include "h263.h"


void ff_snow_inner_add_yblock(const uint8_t *obmc, const int obmc_stride, uint8_t * * block, int b_w, int b_h,
                              int src_x, int src_y, int src_stride, slice_buffer * sb, int add, uint8_t * dst8){
    int y, x;
    IDWTELEM * dst;
    for(y=0; y<b_h; y++){
        //FIXME ugly misuse of obmc_stride
        const uint8_t *obmc1= obmc + y*obmc_stride;
        const uint8_t *obmc2= obmc1+ (obmc_stride>>1);
        const uint8_t *obmc3= obmc1+ obmc_stride*(obmc_stride>>1);
        const uint8_t *obmc4= obmc3+ (obmc_stride>>1);
        dst = slice_buffer_get_line(sb, src_y + y);
        for(x=0; x<b_w; x++){
            int v=   obmc1[x] * block[3][x + y*src_stride]
                    +obmc2[x] * block[2][x + y*src_stride]
                    +obmc3[x] * block[1][x + y*src_stride]
                    +obmc4[x] * block[0][x + y*src_stride];

            v <<= 8 - LOG2_OBMC_MAX;
            if(FRAC_BITS != 8){
                v >>= 8 - FRAC_BITS;
            }
            if(add){
                v += dst[x + src_x];
                v = (v + (1<<(FRAC_BITS-1))) >> FRAC_BITS;
                if(v&(~255)) v= ~(v>>31);
                dst8[x + y*src_stride] = v;
            }else{
                dst[x + src_x] -= v;
            }
        }
    }
}

int ff_snow_get_buffer(SnowContext *s, AVFrame *frame)
{
    int ret, i;
    int edges_needed = av_codec_is_encoder(s->avctx->codec);

    frame->width  = s->avctx->width ;
    frame->height = s->avctx->height;
    if (edges_needed) {
        frame->width  += 2 * EDGE_WIDTH;
        frame->height += 2 * EDGE_WIDTH;
    }
    if ((ret = ff_get_buffer(s->avctx, frame, AV_GET_BUFFER_FLAG_REF)) < 0)
        return ret;
    if (edges_needed) {
        for (i = 0; frame->data[i]; i++) {
            int offset = (EDGE_WIDTH >> (i ? s->chroma_v_shift : 0)) *
                            frame->linesize[i] +
                            (EDGE_WIDTH >> (i ? s->chroma_h_shift : 0));
            frame->data[i] += offset;
        }
        frame->width  = s->avctx->width;
        frame->height = s->avctx->height;
    }

    return 0;
}

void ff_snow_reset_contexts(SnowContext *s){ //FIXME better initial contexts
    int plane_index, level, orientation;

    for(plane_index=0; plane_index<3; plane_index++){
        for(level=0; level<MAX_DECOMPOSITIONS; level++){
            for(orientation=level ? 1:0; orientation<4; orientation++){
                memset(s->plane[plane_index].band[level][orientation].state, MID_STATE, sizeof(s->plane[plane_index].band[level][orientation].state));
            }
        }
    }
    memset(s->header_state, MID_STATE, sizeof(s->header_state));
    memset(s->block_state, MID_STATE, sizeof(s->block_state));
}

int ff_snow_alloc_blocks(SnowContext *s){
    int w= FF_CEIL_RSHIFT(s->avctx->width,  LOG2_MB_SIZE);
    int h= FF_CEIL_RSHIFT(s->avctx->height, LOG2_MB_SIZE);

    s->b_width = w;
    s->b_height= h;

    av_free(s->block);
    s->block= av_mallocz_array(w * h,  sizeof(BlockNode) << (s->block_max_depth*2));
    if (!s->block)
        return AVERROR(ENOMEM);

    return 0;
}

static av_cold void init_qexp(void){
    int i;
    double v=128;

    for(i=0; i<QROOT; i++){
        ff_qexp[i]= lrintf(v);
        v *= pow(2, 1.0 / QROOT);
    }
}
static void mc_block(Plane *p, uint8_t *dst, const uint8_t *src, int stride, int b_w, int b_h, int dx, int dy){
    static const uint8_t weight[64]={
    8,7,6,5,4,3,2,1,
    7,7,0,0,0,0,0,1,
    6,0,6,0,0,0,2,0,
    5,0,0,5,0,3,0,0,
    4,0,0,0,4,0,0,0,
    3,0,0,5,0,3,0,0,
    2,0,6,0,0,0,2,0,
    1,7,0,0,0,0,0,1,
    };

    static const uint8_t brane[256]={
    0x00,0x01,0x01,0x01,0x01,0x01,0x01,0x01,0x11,0x12,0x12,0x12,0x12,0x12,0x12,0x12,
    0x04,0x05,0xcc,0xcc,0xcc,0xcc,0xcc,0x41,0x15,0x16,0xcc,0xcc,0xcc,0xcc,0xcc,0x52,
    0x04,0xcc,0x05,0xcc,0xcc,0xcc,0x41,0xcc,0x15,0xcc,0x16,0xcc,0xcc,0xcc,0x52,0xcc,
    0x04,0xcc,0xcc,0x05,0xcc,0x41,0xcc,0xcc,0x15,0xcc,0xcc,0x16,0xcc,0x52,0xcc,0xcc,
    0x04,0xcc,0xcc,0xcc,0x41,0xcc,0xcc,0xcc,0x15,0xcc,0xcc,0xcc,0x16,0xcc,0xcc,0xcc,
    0x04,0xcc,0xcc,0x41,0xcc,0x05,0xcc,0xcc,0x15,0xcc,0xcc,0x52,0xcc,0x16,0xcc,0xcc,
    0x04,0xcc,0x41,0xcc,0xcc,0xcc,0x05,0xcc,0x15,0xcc,0x52,0xcc,0xcc,0xcc,0x16,0xcc,
    0x04,0x41,0xcc,0xcc,0xcc,0xcc,0xcc,0x05,0x15,0x52,0xcc,0xcc,0xcc,0xcc,0xcc,0x16,
    0x44,0x45,0x45,0x45,0x45,0x45,0x45,0x45,0x55,0x56,0x56,0x56,0x56,0x56,0x56,0x56,
    0x48,0x49,0xcc,0xcc,0xcc,0xcc,0xcc,0x85,0x59,0x5A,0xcc,0xcc,0xcc,0xcc,0xcc,0x96,
    0x48,0xcc,0x49,0xcc,0xcc,0xcc,0x85,0xcc,0x59,0xcc,0x5A,0xcc,0xcc,0xcc,0x96,0xcc,
    0x48,0xcc,0xcc,0x49,0xcc,0x85,0xcc,0xcc,0x59,0xcc,0xcc,0x5A,0xcc,0x96,0xcc,0xcc,
    0x48,0xcc,0xcc,0xcc,0x49,0xcc,0xcc,0xcc,0x59,0xcc,0xcc,0xcc,0x96,0xcc,0xcc,0xcc,
    0x48,0xcc,0xcc,0x85,0xcc,0x49,0xcc,0xcc,0x59,0xcc,0xcc,0x96,0xcc,0x5A,0xcc,0xcc,
    0x48,0xcc,0x85,0xcc,0xcc,0xcc,0x49,0xcc,0x59,0xcc,0x96,0xcc,0xcc,0xcc,0x5A,0xcc,
    0x48,0x85,0xcc,0xcc,0xcc,0xcc,0xcc,0x49,0x59,0x96,0xcc,0xcc,0xcc,0xcc,0xcc,0x5A,
    };

    static const uint8_t needs[16]={
    0,1,0,0,
    2,4,2,0,
    0,1,0,0,
    15
    };

    int x, y, b, r, l;
    int16_t tmpIt   [64*(32+HTAPS_MAX)];
    uint8_t tmp2t[3][64*(32+HTAPS_MAX)];
    int16_t *tmpI= tmpIt;
    uint8_t *tmp2= tmp2t[0];
    const uint8_t *hpel[11];
    av_assert2(dx<16 && dy<16);
    r= brane[dx + 16*dy]&15;
    l= brane[dx + 16*dy]>>4;

    b= needs[l] | needs[r];
    if(p && !p->diag_mc)
        b= 15;

    if(b&5){
        for(y=0; y < b_h+HTAPS_MAX-1; y++){
            for(x=0; x < b_w; x++){
                int a_1=src[x + HTAPS_MAX/2-4];
                int a0= src[x + HTAPS_MAX/2-3];
                int a1= src[x + HTAPS_MAX/2-2];
                int a2= src[x + HTAPS_MAX/2-1];
                int a3= src[x + HTAPS_MAX/2+0];
                int a4= src[x + HTAPS_MAX/2+1];
                int a5= src[x + HTAPS_MAX/2+2];
                int a6= src[x + HTAPS_MAX/2+3];
                int am=0;
                if(!p || p->fast_mc){
                    am= 20*(a2+a3) - 5*(a1+a4) + (a0+a5);
                    tmpI[x]= am;
                    am= (am+16)>>5;
                }else{
                    am= p->hcoeff[0]*(a2+a3) + p->hcoeff[1]*(a1+a4) + p->hcoeff[2]*(a0+a5) + p->hcoeff[3]*(a_1+a6);
                    tmpI[x]= am;
                    am= (am+32)>>6;
                }

                if(am&(~255)) am= ~(am>>31);
                tmp2[x]= am;
            }
            tmpI+= 64;
            tmp2+= 64;
            src += stride;
        }
        src -= stride*y;
    }
    src += HTAPS_MAX/2 - 1;
    tmp2= tmp2t[1];

    if(b&2){
        for(y=0; y < b_h; y++){
            for(x=0; x < b_w+1; x++){
                int a_1=src[x + (HTAPS_MAX/2-4)*stride];
                int a0= src[x + (HTAPS_MAX/2-3)*stride];
                int a1= src[x + (HTAPS_MAX/2-2)*stride];
                int a2= src[x + (HTAPS_MAX/2-1)*stride];
                int a3= src[x + (HTAPS_MAX/2+0)*stride];
                int a4= src[x + (HTAPS_MAX/2+1)*stride];
                int a5= src[x + (HTAPS_MAX/2+2)*stride];
                int a6= src[x + (HTAPS_MAX/2+3)*stride];
                int am=0;
                if(!p || p->fast_mc)
                    am= (20*(a2+a3) - 5*(a1+a4) + (a0+a5) + 16)>>5;
                else
                    am= (p->hcoeff[0]*(a2+a3) + p->hcoeff[1]*(a1+a4) + p->hcoeff[2]*(a0+a5) + p->hcoeff[3]*(a_1+a6) + 32)>>6;

                if(am&(~255)) am= ~(am>>31);
                tmp2[x]= am;
            }
            src += stride;
            tmp2+= 64;
        }
        src -= stride*y;
    }
    src += stride*(HTAPS_MAX/2 - 1);
    tmp2= tmp2t[2];
    tmpI= tmpIt;
    if(b&4){
        for(y=0; y < b_h; y++){
            for(x=0; x < b_w; x++){
                int a_1=tmpI[x + (HTAPS_MAX/2-4)*64];
                int a0= tmpI[x + (HTAPS_MAX/2-3)*64];
                int a1= tmpI[x + (HTAPS_MAX/2-2)*64];
                int a2= tmpI[x + (HTAPS_MAX/2-1)*64];
                int a3= tmpI[x + (HTAPS_MAX/2+0)*64];
                int a4= tmpI[x + (HTAPS_MAX/2+1)*64];
                int a5= tmpI[x + (HTAPS_MAX/2+2)*64];
                int a6= tmpI[x + (HTAPS_MAX/2+3)*64];
                int am=0;
                if(!p || p->fast_mc)
                    am= (20*(a2+a3) - 5*(a1+a4) + (a0+a5) + 512)>>10;
                else
                    am= (p->hcoeff[0]*(a2+a3) + p->hcoeff[1]*(a1+a4) + p->hcoeff[2]*(a0+a5) + p->hcoeff[3]*(a_1+a6) + 2048)>>12;
                if(am&(~255)) am= ~(am>>31);
                tmp2[x]= am;
            }
            tmpI+= 64;
            tmp2+= 64;
        }
    }

    hpel[ 0]= src;
    hpel[ 1]= tmp2t[0] + 64*(HTAPS_MAX/2-1);
    hpel[ 2]= src + 1;

    hpel[ 4]= tmp2t[1];
    hpel[ 5]= tmp2t[2];
    hpel[ 6]= tmp2t[1] + 1;

    hpel[ 8]= src + stride;
    hpel[ 9]= hpel[1] + 64;
    hpel[10]= hpel[8] + 1;

#define MC_STRIDE(x) (needs[x] ? 64 : stride)

    if(b==15){
        int dxy = dx / 8 + dy / 8 * 4;
        const uint8_t *src1 = hpel[dxy    ];
        const uint8_t *src2 = hpel[dxy + 1];
        const uint8_t *src3 = hpel[dxy + 4];
        const uint8_t *src4 = hpel[dxy + 5];
        int stride1 = MC_STRIDE(dxy);
        int stride2 = MC_STRIDE(dxy + 1);
        int stride3 = MC_STRIDE(dxy + 4);
        int stride4 = MC_STRIDE(dxy + 5);
        dx&=7;
        dy&=7;
        for(y=0; y < b_h; y++){
            for(x=0; x < b_w; x++){
                dst[x]= ((8-dx)*(8-dy)*src1[x] + dx*(8-dy)*src2[x]+
                         (8-dx)*   dy *src3[x] + dx*   dy *src4[x]+32)>>6;
            }
            src1+=stride1;
            src2+=stride2;
            src3+=stride3;
            src4+=stride4;
            dst +=stride;
        }
    }else{
        const uint8_t *src1= hpel[l];
        const uint8_t *src2= hpel[r];
        int stride1 = MC_STRIDE(l);
        int stride2 = MC_STRIDE(r);
        int a= weight[((dx&7) + (8*(dy&7)))];
        int b= 8-a;
        for(y=0; y < b_h; y++){
            for(x=0; x < b_w; x++){
                dst[x]= (a*src1[x] + b*src2[x] + 4)>>3;
            }
            src1+=stride1;
            src2+=stride2;
            dst +=stride;
        }
    }
}

void ff_snow_pred_block(SnowContext *s, uint8_t *dst, uint8_t *tmp, ptrdiff_t stride, int sx, int sy, int b_w, int b_h, const BlockNode *block, int plane_index, int w, int h){
    if(block->type & BLOCK_INTRA){
        int x, y;
        const unsigned color  = block->color[plane_index];
        const unsigned color4 = color*0x01010101;
        if(b_w==32){
            for(y=0; y < b_h; y++){
                *(uint32_t*)&dst[0 + y*stride]= color4;
                *(uint32_t*)&dst[4 + y*stride]= color4;
                *(uint32_t*)&dst[8 + y*stride]= color4;
                *(uint32_t*)&dst[12+ y*stride]= color4;
                *(uint32_t*)&dst[16+ y*stride]= color4;
                *(uint32_t*)&dst[20+ y*stride]= color4;
                *(uint32_t*)&dst[24+ y*stride]= color4;
                *(uint32_t*)&dst[28+ y*stride]= color4;
            }
        }else if(b_w==16){
            for(y=0; y < b_h; y++){
                *(uint32_t*)&dst[0 + y*stride]= color4;
                *(uint32_t*)&dst[4 + y*stride]= color4;
                *(uint32_t*)&dst[8 + y*stride]= color4;
                *(uint32_t*)&dst[12+ y*stride]= color4;
            }
        }else if(b_w==8){
            for(y=0; y < b_h; y++){
                *(uint32_t*)&dst[0 + y*stride]= color4;
                *(uint32_t*)&dst[4 + y*stride]= color4;
            }
        }else if(b_w==4){
            for(y=0; y < b_h; y++){
                *(uint32_t*)&dst[0 + y*stride]= color4;
            }
        }else{
            for(y=0; y < b_h; y++){
                for(x=0; x < b_w; x++){
                    dst[x + y*stride]= color;
                }
            }
        }
    }else{
        uint8_t *src= s->last_picture[block->ref]->data[plane_index];
        const int scale= plane_index ?  (2*s->mv_scale)>>s->chroma_h_shift : 2*s->mv_scale;
        int mx= block->mx*scale;
        int my= block->my*scale;
        const int dx= mx&15;
        const int dy= my&15;
        const int tab_index= 3 - (b_w>>2) + (b_w>>4);
        sx += (mx>>4) - (HTAPS_MAX/2-1);
        sy += (my>>4) - (HTAPS_MAX/2-1);
        src += sx + sy*stride;
        if(   (unsigned)sx >= FFMAX(w - b_w - (HTAPS_MAX-2), 0)
           || (unsigned)sy >= FFMAX(h - b_h - (HTAPS_MAX-2), 0)){
            s->vdsp.emulated_edge_mc(tmp + MB_SIZE, src,
                                     stride, stride,
                                     b_w+HTAPS_MAX-1, b_h+HTAPS_MAX-1,
                                     sx, sy, w, h);
            src= tmp + MB_SIZE;
        }

        av_assert2(s->chroma_h_shift == s->chroma_v_shift); // only one mv_scale

        av_assert2((tab_index>=0 && tab_index<4) || b_w==32);
        if(    (dx&3) || (dy&3)
            || !(b_w == b_h || 2*b_w == b_h || b_w == 2*b_h)
            || (b_w&(b_w-1))
            || b_w == 1
            || b_h == 1
            || !s->plane[plane_index].fast_mc )
            mc_block(&s->plane[plane_index], dst, src, stride, b_w, b_h, dx, dy);
        else if(b_w==32){
            int y;
            for(y=0; y<b_h; y+=16){
                s->h264qpel.put_h264_qpel_pixels_tab[0][dy+(dx>>2)](dst + y*stride, src + 3 + (y+3)*stride,stride);
                s->h264qpel.put_h264_qpel_pixels_tab[0][dy+(dx>>2)](dst + 16 + y*stride, src + 19 + (y+3)*stride,stride);
            }
        }else if(b_w==b_h)
            s->h264qpel.put_h264_qpel_pixels_tab[tab_index  ][dy+(dx>>2)](dst,src + 3 + 3*stride,stride);
        else if(b_w==2*b_h){
            s->h264qpel.put_h264_qpel_pixels_tab[tab_index+1][dy+(dx>>2)](dst    ,src + 3       + 3*stride,stride);
            s->h264qpel.put_h264_qpel_pixels_tab[tab_index+1][dy+(dx>>2)](dst+b_h,src + 3 + b_h + 3*stride,stride);
        }else{
            av_assert2(2*b_w==b_h);
            s->h264qpel.put_h264_qpel_pixels_tab[tab_index  ][dy+(dx>>2)](dst           ,src + 3 + 3*stride           ,stride);
            s->h264qpel.put_h264_qpel_pixels_tab[tab_index  ][dy+(dx>>2)](dst+b_w*stride,src + 3 + 3*stride+b_w*stride,stride);
        }
    }
}

#define mca(dx,dy,b_w)\
static void mc_block_hpel ## dx ## dy ## b_w(uint8_t *dst, const uint8_t *src, ptrdiff_t stride, int h){\
    av_assert2(h==b_w);\
    mc_block(NULL, dst, src-(HTAPS_MAX/2-1)-(HTAPS_MAX/2-1)*stride, stride, b_w, b_w, dx, dy);\
}

mca( 0, 0,16)
mca( 8, 0,16)
mca( 0, 8,16)
mca( 8, 8,16)
mca( 0, 0,8)
mca( 8, 0,8)
mca( 0, 8,8)
mca( 8, 8,8)

av_cold int ff_snow_common_init(AVCodecContext *avctx){
    SnowContext *s = avctx->priv_data;
    int width, height;
    int i, j;

    s->avctx= avctx;
    s->max_ref_frames=1; //just make sure it's not an invalid value in case of no initial keyframe
    s->spatial_decomposition_count = 1;

    ff_me_cmp_init(&s->mecc, avctx);
    ff_hpeldsp_init(&s->hdsp, avctx->flags);
    ff_videodsp_init(&s->vdsp, 8);
    ff_dwt_init(&s->dwt);
    ff_h264qpel_init(&s->h264qpel, 8);

#define mcf(dx,dy)\
    s->qdsp.put_qpel_pixels_tab       [0][dy+dx/4]=\
    s->qdsp.put_no_rnd_qpel_pixels_tab[0][dy+dx/4]=\
        s->h264qpel.put_h264_qpel_pixels_tab[0][dy+dx/4];\
    s->qdsp.put_qpel_pixels_tab       [1][dy+dx/4]=\
    s->qdsp.put_no_rnd_qpel_pixels_tab[1][dy+dx/4]=\
        s->h264qpel.put_h264_qpel_pixels_tab[1][dy+dx/4];

    mcf( 0, 0)
    mcf( 4, 0)
    mcf( 8, 0)
    mcf(12, 0)
    mcf( 0, 4)
    mcf( 4, 4)
    mcf( 8, 4)
    mcf(12, 4)
    mcf( 0, 8)
    mcf( 4, 8)
    mcf( 8, 8)
    mcf(12, 8)
    mcf( 0,12)
    mcf( 4,12)
    mcf( 8,12)
    mcf(12,12)

#define mcfh(dx,dy)\
    s->hdsp.put_pixels_tab       [0][dy/4+dx/8]=\
    s->hdsp.put_no_rnd_pixels_tab[0][dy/4+dx/8]=\
        mc_block_hpel ## dx ## dy ## 16;\
    s->hdsp.put_pixels_tab       [1][dy/4+dx/8]=\
    s->hdsp.put_no_rnd_pixels_tab[1][dy/4+dx/8]=\
        mc_block_hpel ## dx ## dy ## 8;

    mcfh(0, 0)
    mcfh(8, 0)
    mcfh(0, 8)
    mcfh(8, 8)

    init_qexp();

//    dec += FFMAX(s->chroma_h_shift, s->chroma_v_shift);

    width= s->avctx->width;
    height= s->avctx->height;

    FF_ALLOCZ_ARRAY_OR_GOTO(avctx, s->spatial_idwt_buffer, width, height * sizeof(IDWTELEM), fail);
    FF_ALLOCZ_ARRAY_OR_GOTO(avctx, s->spatial_dwt_buffer,  width, height * sizeof(DWTELEM),  fail); //FIXME this does not belong here
    FF_ALLOCZ_ARRAY_OR_GOTO(avctx, s->temp_dwt_buffer,     width, sizeof(DWTELEM),  fail);
    FF_ALLOCZ_ARRAY_OR_GOTO(avctx, s->temp_idwt_buffer,    width, sizeof(IDWTELEM), fail);
    FF_ALLOC_ARRAY_OR_GOTO(avctx,  s->run_buffer,          ((width + 1) >> 1), ((height + 1) >> 1) * sizeof(*s->run_buffer), fail);

    for(i=0; i<MAX_REF_FRAMES; i++) {
        for(j=0; j<MAX_REF_FRAMES; j++)
            ff_scale_mv_ref[i][j] = 256*(i+1)/(j+1);
        s->last_picture[i] = av_frame_alloc();
        if (!s->last_picture[i])
            goto fail;
    }

    s->mconly_picture = av_frame_alloc();
    s->current_picture = av_frame_alloc();
    if (!s->mconly_picture || !s->current_picture)
        goto fail;

    return 0;
fail:
    return AVERROR(ENOMEM);
}

int ff_snow_common_init_after_header(AVCodecContext *avctx) {
    SnowContext *s = avctx->priv_data;
    int plane_index, level, orientation;
    int ret, emu_buf_size;

    if(!s->scratchbuf) {
        if ((ret = ff_get_buffer(s->avctx, s->mconly_picture,
                                 AV_GET_BUFFER_FLAG_REF)) < 0)
            return ret;
        FF_ALLOCZ_ARRAY_OR_GOTO(avctx, s->scratchbuf, FFMAX(s->mconly_picture->linesize[0], 2*avctx->width+256), 7*MB_SIZE, fail);
        emu_buf_size = FFMAX(s->mconly_picture->linesize[0], 2*avctx->width+256) * (2 * MB_SIZE + HTAPS_MAX - 1);
        FF_ALLOC_OR_GOTO(avctx, s->emu_edge_buffer, emu_buf_size, fail);
    }

    if(s->mconly_picture->format != avctx->pix_fmt) {
        av_log(avctx, AV_LOG_ERROR, "pixel format changed\n");
        return AVERROR_INVALIDDATA;
    }

    for(plane_index=0; plane_index < s->nb_planes; plane_index++){
        int w= s->avctx->width;
        int h= s->avctx->height;

        if(plane_index){
            w>>= s->chroma_h_shift;
            h>>= s->chroma_v_shift;
        }
        s->plane[plane_index].width = w;
        s->plane[plane_index].height= h;

        for(level=s->spatial_decomposition_count-1; level>=0; level--){
            for(orientation=level ? 1 : 0; orientation<4; orientation++){
                SubBand *b= &s->plane[plane_index].band[level][orientation];

                b->buf= s->spatial_dwt_buffer;
                b->level= level;
                b->stride= s->plane[plane_index].width << (s->spatial_decomposition_count - level);
                b->width = (w + !(orientation&1))>>1;
                b->height= (h + !(orientation>1))>>1;

                b->stride_line = 1 << (s->spatial_decomposition_count - level);
                b->buf_x_offset = 0;
                b->buf_y_offset = 0;

                if(orientation&1){
                    b->buf += (w+1)>>1;
                    b->buf_x_offset = (w+1)>>1;
                }
                if(orientation>1){
                    b->buf += b->stride>>1;
                    b->buf_y_offset = b->stride_line >> 1;
                }
                b->ibuf= s->spatial_idwt_buffer + (b->buf - s->spatial_dwt_buffer);

                if(level)
                    b->parent= &s->plane[plane_index].band[level-1][orientation];
                //FIXME avoid this realloc
                av_freep(&b->x_coeff);
                b->x_coeff=av_mallocz_array(((b->width+1) * b->height+1), sizeof(x_and_coeff));
                if (!b->x_coeff)
                    goto fail;
            }
            w= (w+1)>>1;
            h= (h+1)>>1;
        }
    }

    return 0;
fail:
    return AVERROR(ENOMEM);
}

#define USE_HALFPEL_PLANE 0

static int halfpel_interpol(SnowContext *s, uint8_t *halfpel[4][4], AVFrame *frame){
    int p,x,y;

    for(p=0; p < s->nb_planes; p++){
        int is_chroma= !!p;
        int w= is_chroma ? s->avctx->width >>s->chroma_h_shift : s->avctx->width;
        int h= is_chroma ? s->avctx->height>>s->chroma_v_shift : s->avctx->height;
        int ls= frame->linesize[p];
        uint8_t *src= frame->data[p];

        halfpel[1][p] = av_malloc_array(ls, (h + 2 * EDGE_WIDTH));
        halfpel[2][p] = av_malloc_array(ls, (h + 2 * EDGE_WIDTH));
        halfpel[3][p] = av_malloc_array(ls, (h + 2 * EDGE_WIDTH));
        if (!halfpel[1][p] || !halfpel[2][p] || !halfpel[3][p]) {
            av_freep(&halfpel[1][p]);
            av_freep(&halfpel[2][p]);
            av_freep(&halfpel[3][p]);
            return AVERROR(ENOMEM);
        }
        halfpel[1][p] += EDGE_WIDTH * (1 + ls);
        halfpel[2][p] += EDGE_WIDTH * (1 + ls);
        halfpel[3][p] += EDGE_WIDTH * (1 + ls);

        halfpel[0][p]= src;
        for(y=0; y<h; y++){
            for(x=0; x<w; x++){
                int i= y*ls + x;

                halfpel[1][p][i]= (20*(src[i] + src[i+1]) - 5*(src[i-1] + src[i+2]) + (src[i-2] + src[i+3]) + 16 )>>5;
            }
        }
        for(y=0; y<h; y++){
            for(x=0; x<w; x++){
                int i= y*ls + x;

                halfpel[2][p][i]= (20*(src[i] + src[i+ls]) - 5*(src[i-ls] + src[i+2*ls]) + (src[i-2*ls] + src[i+3*ls]) + 16 )>>5;
            }
        }
        src= halfpel[1][p];
        for(y=0; y<h; y++){
            for(x=0; x<w; x++){
                int i= y*ls + x;

                halfpel[3][p][i]= (20*(src[i] + src[i+ls]) - 5*(src[i-ls] + src[i+2*ls]) + (src[i-2*ls] + src[i+3*ls]) + 16 )>>5;
            }
        }

//FIXME border!
    }
    return 0;
}

void ff_snow_release_buffer(AVCodecContext *avctx)
{
    SnowContext *s = avctx->priv_data;
    int i;

    if(s->last_picture[s->max_ref_frames-1]->data[0]){
        av_frame_unref(s->last_picture[s->max_ref_frames-1]);
        for(i=0; i<9; i++)
            if(s->halfpel_plane[s->max_ref_frames-1][1+i/3][i%3]) {
                av_free(s->halfpel_plane[s->max_ref_frames-1][1+i/3][i%3] - EDGE_WIDTH*(1+s->current_picture->linesize[i%3]));
                s->halfpel_plane[s->max_ref_frames-1][1+i/3][i%3] = NULL;
            }
    }
}

int ff_snow_frame_start(SnowContext *s){
   AVFrame *tmp;
   int i, ret;

    ff_snow_release_buffer(s->avctx);

    tmp= s->last_picture[s->max_ref_frames-1];
    for(i=s->max_ref_frames-1; i>0; i--)
        s->last_picture[i] = s->last_picture[i-1];
    memmove(s->halfpel_plane+1, s->halfpel_plane, (s->max_ref_frames-1)*sizeof(void*)*4*4);
    if(USE_HALFPEL_PLANE && s->current_picture->data[0]) {
        if((ret = halfpel_interpol(s, s->halfpel_plane[0], s->current_picture)) < 0)
            return ret;
    }
    s->last_picture[0] = s->current_picture;
    s->current_picture = tmp;

    if(s->keyframe){
        s->ref_frames= 0;
    }else{
        int i;
        for(i=0; i<s->max_ref_frames && s->last_picture[i]->data[0]; i++)
            if(i && s->last_picture[i-1]->key_frame)
                break;
        s->ref_frames= i;
        if(s->ref_frames==0){
            av_log(s->avctx,AV_LOG_ERROR, "No reference frames\n");
            return -1;
        }
    }
    if ((ret = ff_snow_get_buffer(s, s->current_picture)) < 0)
        return ret;

    s->current_picture->key_frame= s->keyframe;

    return 0;
}

av_cold void ff_snow_common_end(SnowContext *s)
{
    int plane_index, level, orientation, i;

    av_freep(&s->spatial_dwt_buffer);
    av_freep(&s->temp_dwt_buffer);
    av_freep(&s->spatial_idwt_buffer);
    av_freep(&s->temp_idwt_buffer);
    av_freep(&s->run_buffer);

    s->m.me.temp= NULL;
    av_freep(&s->m.me.scratchpad);
    av_freep(&s->m.me.map);
    av_freep(&s->m.me.score_map);
    av_freep(&s->m.obmc_scratchpad);

    av_freep(&s->block);
    av_freep(&s->scratchbuf);
    av_freep(&s->emu_edge_buffer);

    for(i=0; i<MAX_REF_FRAMES; i++){
        av_freep(&s->ref_mvs[i]);
        av_freep(&s->ref_scores[i]);
        if(s->last_picture[i] && s->last_picture[i]->data[0]) {
            av_assert0(s->last_picture[i]->data[0] != s->current_picture->data[0]);
        }
        av_frame_free(&s->last_picture[i]);
    }

    for(plane_index=0; plane_index < MAX_PLANES; plane_index++){
        for(level=MAX_DECOMPOSITIONS-1; level>=0; level--){
            for(orientation=level ? 1 : 0; orientation<4; orientation++){
                SubBand *b= &s->plane[plane_index].band[level][orientation];

                av_freep(&b->x_coeff);
            }
        }
    }
    av_frame_free(&s->mconly_picture);
    av_frame_free(&s->current_picture);
}