forked from webmproject/libvpx
-
Notifications
You must be signed in to change notification settings - Fork 2
/
y4minput.c
1127 lines (1094 loc) · 38.6 KB
/
y4minput.c
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
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
/*
* Copyright (c) 2010 The WebM project authors. All Rights Reserved.
*
* Use of this source code is governed by a BSD-style license
* that can be found in the LICENSE file in the root of the source
* tree. An additional intellectual property rights grant can be found
* in the file PATENTS. All contributing project authors may
* be found in the AUTHORS file in the root of the source tree.
*
* Based on code from the OggTheora software codec source code,
* Copyright (C) 2002-2010 The Xiph.Org Foundation and contributors.
*/
#include <errno.h>
#include <stdlib.h>
#include <string.h>
#include "vpx/vpx_integer.h"
#include "y4minput.h"
// Reads 'size' bytes from 'file' into 'buf' with some fault tolerance.
// Returns true on success.
static int file_read(void *buf, size_t size, FILE *file) {
const int kMaxRetries = 5;
int retry_count = 0;
int file_error;
size_t len = 0;
do {
const size_t n = fread((uint8_t *)buf + len, 1, size - len, file);
len += n;
file_error = ferror(file);
if (file_error) {
if (errno == EINTR || errno == EAGAIN) {
clearerr(file);
continue;
} else {
fprintf(stderr, "Error reading file: %u of %u bytes read, %d: %s\n",
(uint32_t)len, (uint32_t)size, errno, strerror(errno));
return 0;
}
}
} while (!feof(file) && len < size && ++retry_count < kMaxRetries);
if (!feof(file) && len != size) {
fprintf(stderr,
"Error reading file: %u of %u bytes read,"
" error: %d, retries: %d, %d: %s\n",
(uint32_t)len, (uint32_t)size, file_error, retry_count, errno,
strerror(errno));
}
return len == size;
}
static int y4m_parse_tags(y4m_input *_y4m, char *_tags) {
int got_w;
int got_h;
int got_fps;
int got_interlace;
int got_par;
int got_chroma;
char *p;
char *q;
got_w = got_h = got_fps = got_interlace = got_par = got_chroma = 0;
for (p = _tags;; p = q) {
/*Skip any leading spaces.*/
while (*p == ' ') p++;
/*If that's all we have, stop.*/
if (p[0] == '\0') break;
/*Find the end of this tag.*/
for (q = p + 1; *q != '\0' && *q != ' '; q++) {
}
/*Process the tag.*/
switch (p[0]) {
case 'W': {
if (sscanf(p + 1, "%d", &_y4m->pic_w) != 1) return -1;
got_w = 1;
break;
}
case 'H': {
if (sscanf(p + 1, "%d", &_y4m->pic_h) != 1) return -1;
got_h = 1;
break;
}
case 'F': {
if (sscanf(p + 1, "%d:%d", &_y4m->fps_n, &_y4m->fps_d) != 2) {
return -1;
}
got_fps = 1;
break;
}
case 'I': {
_y4m->interlace = p[1];
got_interlace = 1;
break;
}
case 'A': {
if (sscanf(p + 1, "%d:%d", &_y4m->par_n, &_y4m->par_d) != 2) {
return -1;
}
got_par = 1;
break;
}
case 'C': {
if (q - p > 16) return -1;
memcpy(_y4m->chroma_type, p + 1, q - p - 1);
_y4m->chroma_type[q - p - 1] = '\0';
got_chroma = 1;
break;
}
/*Ignore unknown tags.*/
}
}
if (!got_w || !got_h || !got_fps) return -1;
if (!got_interlace) _y4m->interlace = '?';
if (!got_par) _y4m->par_n = _y4m->par_d = 0;
/*Chroma-type is not specified in older files, e.g., those generated by
mplayer.*/
if (!got_chroma) strcpy(_y4m->chroma_type, "420");
return 0;
}
/*All anti-aliasing filters in the following conversion functions are based on
one of two window functions:
The 6-tap Lanczos window (for down-sampling and shifts):
sinc(\pi*t)*sinc(\pi*t/3), |t|<3 (sinc(t)==sin(t)/t)
0, |t|>=3
The 4-tap Mitchell window (for up-sampling):
7|t|^3-12|t|^2+16/3, |t|<1
-(7/3)|x|^3+12|x|^2-20|x|+32/3, |t|<2
0, |t|>=2
The number of taps is intentionally kept small to reduce computational
overhead and limit ringing.
The taps from these filters are scaled so that their sum is 1, and the
result is scaled by 128 and rounded to integers to create a filter whose
intermediate values fit inside 16 bits.
Coefficients are rounded in such a way as to ensure their sum is still 128,
which is usually equivalent to normal rounding.
Conversions which require both horizontal and vertical filtering could
have these steps pipelined, for less memory consumption and better cache
performance, but we do them separately for simplicity.*/
#define OC_MINI(_a, _b) ((_a) > (_b) ? (_b) : (_a))
#define OC_MAXI(_a, _b) ((_a) < (_b) ? (_b) : (_a))
#define OC_CLAMPI(_a, _b, _c) (OC_MAXI(_a, OC_MINI(_b, _c)))
/*420jpeg chroma samples are sited like:
Y-------Y-------Y-------Y-------
| | | |
| BR | | BR |
| | | |
Y-------Y-------Y-------Y-------
| | | |
| | | |
| | | |
Y-------Y-------Y-------Y-------
| | | |
| BR | | BR |
| | | |
Y-------Y-------Y-------Y-------
| | | |
| | | |
| | | |
420mpeg2 chroma samples are sited like:
Y-------Y-------Y-------Y-------
| | | |
BR | BR |
| | | |
Y-------Y-------Y-------Y-------
| | | |
| | | |
| | | |
Y-------Y-------Y-------Y-------
| | | |
BR | BR |
| | | |
Y-------Y-------Y-------Y-------
| | | |
| | | |
| | | |
We use a resampling filter to shift the site locations one quarter pixel (at
the chroma plane's resolution) to the right.
The 4:2:2 modes look exactly the same, except there are twice as many chroma
lines, and they are vertically co-sited with the luma samples in both the
mpeg2 and jpeg cases (thus requiring no vertical resampling).*/
static void y4m_42xmpeg2_42xjpeg_helper(unsigned char *_dst,
const unsigned char *_src, int _c_w,
int _c_h) {
int y;
int x;
for (y = 0; y < _c_h; y++) {
/*Filter: [4 -17 114 35 -9 1]/128, derived from a 6-tap Lanczos
window.*/
for (x = 0; x < OC_MINI(_c_w, 2); x++) {
_dst[x] = (unsigned char)OC_CLAMPI(
0,
(4 * _src[0] - 17 * _src[OC_MAXI(x - 1, 0)] + 114 * _src[x] +
35 * _src[OC_MINI(x + 1, _c_w - 1)] -
9 * _src[OC_MINI(x + 2, _c_w - 1)] + _src[OC_MINI(x + 3, _c_w - 1)] +
64) >>
7,
255);
}
for (; x < _c_w - 3; x++) {
_dst[x] = (unsigned char)OC_CLAMPI(
0,
(4 * _src[x - 2] - 17 * _src[x - 1] + 114 * _src[x] +
35 * _src[x + 1] - 9 * _src[x + 2] + _src[x + 3] + 64) >>
7,
255);
}
for (; x < _c_w; x++) {
_dst[x] = (unsigned char)OC_CLAMPI(
0,
(4 * _src[x - 2] - 17 * _src[x - 1] + 114 * _src[x] +
35 * _src[OC_MINI(x + 1, _c_w - 1)] -
9 * _src[OC_MINI(x + 2, _c_w - 1)] + _src[_c_w - 1] + 64) >>
7,
255);
}
_dst += _c_w;
_src += _c_w;
}
}
/*Handles both 422 and 420mpeg2 to 422jpeg and 420jpeg, respectively.*/
static void y4m_convert_42xmpeg2_42xjpeg(y4m_input *_y4m, unsigned char *_dst,
unsigned char *_aux) {
int c_w;
int c_h;
int c_sz;
int pli;
/*Skip past the luma data.*/
_dst += _y4m->pic_w * _y4m->pic_h;
/*Compute the size of each chroma plane.*/
c_w = (_y4m->pic_w + _y4m->dst_c_dec_h - 1) / _y4m->dst_c_dec_h;
c_h = (_y4m->pic_h + _y4m->dst_c_dec_v - 1) / _y4m->dst_c_dec_v;
c_sz = c_w * c_h;
for (pli = 1; pli < 3; pli++) {
y4m_42xmpeg2_42xjpeg_helper(_dst, _aux, c_w, c_h);
_dst += c_sz;
_aux += c_sz;
}
}
/*This format is only used for interlaced content, but is included for
completeness.
420jpeg chroma samples are sited like:
Y-------Y-------Y-------Y-------
| | | |
| BR | | BR |
| | | |
Y-------Y-------Y-------Y-------
| | | |
| | | |
| | | |
Y-------Y-------Y-------Y-------
| | | |
| BR | | BR |
| | | |
Y-------Y-------Y-------Y-------
| | | |
| | | |
| | | |
420paldv chroma samples are sited like:
YR------Y-------YR------Y-------
| | | |
| | | |
| | | |
YB------Y-------YB------Y-------
| | | |
| | | |
| | | |
YR------Y-------YR------Y-------
| | | |
| | | |
| | | |
YB------Y-------YB------Y-------
| | | |
| | | |
| | | |
We use a resampling filter to shift the site locations one quarter pixel (at
the chroma plane's resolution) to the right.
Then we use another filter to move the C_r location down one quarter pixel,
and the C_b location up one quarter pixel.*/
static void y4m_convert_42xpaldv_42xjpeg(y4m_input *_y4m, unsigned char *_dst,
unsigned char *_aux) {
unsigned char *tmp;
int c_w;
int c_h;
int c_sz;
int pli;
int y;
int x;
/*Skip past the luma data.*/
_dst += _y4m->pic_w * _y4m->pic_h;
/*Compute the size of each chroma plane.*/
c_w = (_y4m->pic_w + 1) / 2;
c_h = (_y4m->pic_h + _y4m->dst_c_dec_h - 1) / _y4m->dst_c_dec_h;
c_sz = c_w * c_h;
tmp = _aux + 2 * c_sz;
for (pli = 1; pli < 3; pli++) {
/*First do the horizontal re-sampling.
This is the same as the mpeg2 case, except that after the horizontal
case, we need to apply a second vertical filter.*/
y4m_42xmpeg2_42xjpeg_helper(tmp, _aux, c_w, c_h);
_aux += c_sz;
switch (pli) {
case 1: {
/*Slide C_b up a quarter-pel.
This is the same filter used above, but in the other order.*/
for (x = 0; x < c_w; x++) {
for (y = 0; y < OC_MINI(c_h, 3); y++) {
_dst[y * c_w] = (unsigned char)OC_CLAMPI(
0,
(tmp[0] - 9 * tmp[OC_MAXI(y - 2, 0) * c_w] +
35 * tmp[OC_MAXI(y - 1, 0) * c_w] + 114 * tmp[y * c_w] -
17 * tmp[OC_MINI(y + 1, c_h - 1) * c_w] +
4 * tmp[OC_MINI(y + 2, c_h - 1) * c_w] + 64) >>
7,
255);
}
for (; y < c_h - 2; y++) {
_dst[y * c_w] = (unsigned char)OC_CLAMPI(
0,
(tmp[(y - 3) * c_w] - 9 * tmp[(y - 2) * c_w] +
35 * tmp[(y - 1) * c_w] + 114 * tmp[y * c_w] -
17 * tmp[(y + 1) * c_w] + 4 * tmp[(y + 2) * c_w] + 64) >>
7,
255);
}
for (; y < c_h; y++) {
_dst[y * c_w] = (unsigned char)OC_CLAMPI(
0,
(tmp[(y - 3) * c_w] - 9 * tmp[(y - 2) * c_w] +
35 * tmp[(y - 1) * c_w] + 114 * tmp[y * c_w] -
17 * tmp[OC_MINI(y + 1, c_h - 1) * c_w] +
4 * tmp[(c_h - 1) * c_w] + 64) >>
7,
255);
}
_dst++;
tmp++;
}
_dst += c_sz - c_w;
tmp -= c_w;
break;
}
case 2: {
/*Slide C_r down a quarter-pel.
This is the same as the horizontal filter.*/
for (x = 0; x < c_w; x++) {
for (y = 0; y < OC_MINI(c_h, 2); y++) {
_dst[y * c_w] = (unsigned char)OC_CLAMPI(
0,
(4 * tmp[0] - 17 * tmp[OC_MAXI(y - 1, 0) * c_w] +
114 * tmp[y * c_w] + 35 * tmp[OC_MINI(y + 1, c_h - 1) * c_w] -
9 * tmp[OC_MINI(y + 2, c_h - 1) * c_w] +
tmp[OC_MINI(y + 3, c_h - 1) * c_w] + 64) >>
7,
255);
}
for (; y < c_h - 3; y++) {
_dst[y * c_w] = (unsigned char)OC_CLAMPI(
0,
(4 * tmp[(y - 2) * c_w] - 17 * tmp[(y - 1) * c_w] +
114 * tmp[y * c_w] + 35 * tmp[(y + 1) * c_w] -
9 * tmp[(y + 2) * c_w] + tmp[(y + 3) * c_w] + 64) >>
7,
255);
}
for (; y < c_h; y++) {
_dst[y * c_w] = (unsigned char)OC_CLAMPI(
0,
(4 * tmp[(y - 2) * c_w] - 17 * tmp[(y - 1) * c_w] +
114 * tmp[y * c_w] + 35 * tmp[OC_MINI(y + 1, c_h - 1) * c_w] -
9 * tmp[OC_MINI(y + 2, c_h - 1) * c_w] + tmp[(c_h - 1) * c_w] +
64) >>
7,
255);
}
_dst++;
tmp++;
}
break;
}
}
/*For actual interlaced material, this would have to be done separately on
each field, and the shift amounts would be different.
C_r moves down 1/8, C_b up 3/8 in the top field, and C_r moves down 3/8,
C_b up 1/8 in the bottom field.
The corresponding filters would be:
Down 1/8 (reverse order for up): [3 -11 125 15 -4 0]/128
Down 3/8 (reverse order for up): [4 -19 98 56 -13 2]/128*/
}
}
/*Perform vertical filtering to reduce a single plane from 4:2:2 to 4:2:0.
This is used as a helper by several converation routines.*/
static void y4m_422jpeg_420jpeg_helper(unsigned char *_dst,
const unsigned char *_src, int _c_w,
int _c_h) {
int y;
int x;
/*Filter: [3 -17 78 78 -17 3]/128, derived from a 6-tap Lanczos window.*/
for (x = 0; x < _c_w; x++) {
for (y = 0; y < OC_MINI(_c_h, 2); y += 2) {
_dst[(y >> 1) * _c_w] =
OC_CLAMPI(0,
(64 * _src[0] + 78 * _src[OC_MINI(1, _c_h - 1) * _c_w] -
17 * _src[OC_MINI(2, _c_h - 1) * _c_w] +
3 * _src[OC_MINI(3, _c_h - 1) * _c_w] + 64) >>
7,
255);
}
for (; y < _c_h - 3; y += 2) {
_dst[(y >> 1) * _c_w] =
OC_CLAMPI(0,
(3 * (_src[(y - 2) * _c_w] + _src[(y + 3) * _c_w]) -
17 * (_src[(y - 1) * _c_w] + _src[(y + 2) * _c_w]) +
78 * (_src[y * _c_w] + _src[(y + 1) * _c_w]) + 64) >>
7,
255);
}
for (; y < _c_h; y += 2) {
_dst[(y >> 1) * _c_w] = OC_CLAMPI(
0,
(3 * (_src[(y - 2) * _c_w] + _src[(_c_h - 1) * _c_w]) -
17 * (_src[(y - 1) * _c_w] + _src[OC_MINI(y + 2, _c_h - 1) * _c_w]) +
78 * (_src[y * _c_w] + _src[OC_MINI(y + 1, _c_h - 1) * _c_w]) +
64) >>
7,
255);
}
_src++;
_dst++;
}
}
/*420jpeg chroma samples are sited like:
Y-------Y-------Y-------Y-------
| | | |
| BR | | BR |
| | | |
Y-------Y-------Y-------Y-------
| | | |
| | | |
| | | |
Y-------Y-------Y-------Y-------
| | | |
| BR | | BR |
| | | |
Y-------Y-------Y-------Y-------
| | | |
| | | |
| | | |
422jpeg chroma samples are sited like:
Y---BR--Y-------Y---BR--Y-------
| | | |
| | | |
| | | |
Y---BR--Y-------Y---BR--Y-------
| | | |
| | | |
| | | |
Y---BR--Y-------Y---BR--Y-------
| | | |
| | | |
| | | |
Y---BR--Y-------Y---BR--Y-------
| | | |
| | | |
| | | |
We use a resampling filter to decimate the chroma planes by two in the
vertical direction.*/
static void y4m_convert_422jpeg_420jpeg(y4m_input *_y4m, unsigned char *_dst,
unsigned char *_aux) {
int c_w;
int c_h;
int c_sz;
int dst_c_w;
int dst_c_h;
int dst_c_sz;
int pli;
/*Skip past the luma data.*/
_dst += _y4m->pic_w * _y4m->pic_h;
/*Compute the size of each chroma plane.*/
c_w = (_y4m->pic_w + _y4m->src_c_dec_h - 1) / _y4m->src_c_dec_h;
c_h = _y4m->pic_h;
dst_c_w = (_y4m->pic_w + _y4m->dst_c_dec_h - 1) / _y4m->dst_c_dec_h;
dst_c_h = (_y4m->pic_h + _y4m->dst_c_dec_v - 1) / _y4m->dst_c_dec_v;
c_sz = c_w * c_h;
dst_c_sz = dst_c_w * dst_c_h;
for (pli = 1; pli < 3; pli++) {
y4m_422jpeg_420jpeg_helper(_dst, _aux, c_w, c_h);
_aux += c_sz;
_dst += dst_c_sz;
}
}
/*420jpeg chroma samples are sited like:
Y-------Y-------Y-------Y-------
| | | |
| BR | | BR |
| | | |
Y-------Y-------Y-------Y-------
| | | |
| | | |
| | | |
Y-------Y-------Y-------Y-------
| | | |
| BR | | BR |
| | | |
Y-------Y-------Y-------Y-------
| | | |
| | | |
| | | |
422 chroma samples are sited like:
YBR-----Y-------YBR-----Y-------
| | | |
| | | |
| | | |
YBR-----Y-------YBR-----Y-------
| | | |
| | | |
| | | |
YBR-----Y-------YBR-----Y-------
| | | |
| | | |
| | | |
YBR-----Y-------YBR-----Y-------
| | | |
| | | |
| | | |
We use a resampling filter to shift the original site locations one quarter
pixel (at the original chroma resolution) to the right.
Then we use a second resampling filter to decimate the chroma planes by two
in the vertical direction.*/
static void y4m_convert_422_420jpeg(y4m_input *_y4m, unsigned char *_dst,
unsigned char *_aux) {
unsigned char *tmp;
int c_w;
int c_h;
int c_sz;
int dst_c_h;
int dst_c_sz;
int pli;
/*Skip past the luma data.*/
_dst += _y4m->pic_w * _y4m->pic_h;
/*Compute the size of each chroma plane.*/
c_w = (_y4m->pic_w + _y4m->src_c_dec_h - 1) / _y4m->src_c_dec_h;
c_h = _y4m->pic_h;
dst_c_h = (_y4m->pic_h + _y4m->dst_c_dec_v - 1) / _y4m->dst_c_dec_v;
c_sz = c_w * c_h;
dst_c_sz = c_w * dst_c_h;
tmp = _aux + 2 * c_sz;
for (pli = 1; pli < 3; pli++) {
/*In reality, the horizontal and vertical steps could be pipelined, for
less memory consumption and better cache performance, but we do them
separately for simplicity.*/
/*First do horizontal filtering (convert to 422jpeg)*/
y4m_42xmpeg2_42xjpeg_helper(tmp, _aux, c_w, c_h);
/*Now do the vertical filtering.*/
y4m_422jpeg_420jpeg_helper(_dst, tmp, c_w, c_h);
_aux += c_sz;
_dst += dst_c_sz;
}
}
/*420jpeg chroma samples are sited like:
Y-------Y-------Y-------Y-------
| | | |
| BR | | BR |
| | | |
Y-------Y-------Y-------Y-------
| | | |
| | | |
| | | |
Y-------Y-------Y-------Y-------
| | | |
| BR | | BR |
| | | |
Y-------Y-------Y-------Y-------
| | | |
| | | |
| | | |
411 chroma samples are sited like:
YBR-----Y-------Y-------Y-------
| | | |
| | | |
| | | |
YBR-----Y-------Y-------Y-------
| | | |
| | | |
| | | |
YBR-----Y-------Y-------Y-------
| | | |
| | | |
| | | |
YBR-----Y-------Y-------Y-------
| | | |
| | | |
| | | |
We use a filter to resample at site locations one eighth pixel (at the source
chroma plane's horizontal resolution) and five eighths of a pixel to the
right.
Then we use another filter to decimate the planes by 2 in the vertical
direction.*/
static void y4m_convert_411_420jpeg(y4m_input *_y4m, unsigned char *_dst,
unsigned char *_aux) {
unsigned char *tmp;
int c_w;
int c_h;
int c_sz;
int dst_c_w;
int dst_c_h;
int dst_c_sz;
int tmp_sz;
int pli;
int y;
int x;
/*Skip past the luma data.*/
_dst += _y4m->pic_w * _y4m->pic_h;
/*Compute the size of each chroma plane.*/
c_w = (_y4m->pic_w + _y4m->src_c_dec_h - 1) / _y4m->src_c_dec_h;
c_h = _y4m->pic_h;
dst_c_w = (_y4m->pic_w + _y4m->dst_c_dec_h - 1) / _y4m->dst_c_dec_h;
dst_c_h = (_y4m->pic_h + _y4m->dst_c_dec_v - 1) / _y4m->dst_c_dec_v;
c_sz = c_w * c_h;
dst_c_sz = dst_c_w * dst_c_h;
tmp_sz = dst_c_w * c_h;
tmp = _aux + 2 * c_sz;
for (pli = 1; pli < 3; pli++) {
/*In reality, the horizontal and vertical steps could be pipelined, for
less memory consumption and better cache performance, but we do them
separately for simplicity.*/
/*First do horizontal filtering (convert to 422jpeg)*/
for (y = 0; y < c_h; y++) {
/*Filters: [1 110 18 -1]/128 and [-3 50 86 -5]/128, both derived from a
4-tap Mitchell window.*/
for (x = 0; x < OC_MINI(c_w, 1); x++) {
tmp[x << 1] = (unsigned char)OC_CLAMPI(
0,
(111 * _aux[0] + 18 * _aux[OC_MINI(1, c_w - 1)] -
_aux[OC_MINI(2, c_w - 1)] + 64) >>
7,
255);
tmp[x << 1 | 1] = (unsigned char)OC_CLAMPI(
0,
(47 * _aux[0] + 86 * _aux[OC_MINI(1, c_w - 1)] -
5 * _aux[OC_MINI(2, c_w - 1)] + 64) >>
7,
255);
}
for (; x < c_w - 2; x++) {
tmp[x << 1] =
(unsigned char)OC_CLAMPI(0,
(_aux[x - 1] + 110 * _aux[x] +
18 * _aux[x + 1] - _aux[x + 2] + 64) >>
7,
255);
tmp[x << 1 | 1] = (unsigned char)OC_CLAMPI(
0,
(-3 * _aux[x - 1] + 50 * _aux[x] + 86 * _aux[x + 1] -
5 * _aux[x + 2] + 64) >>
7,
255);
}
for (; x < c_w; x++) {
tmp[x << 1] = (unsigned char)OC_CLAMPI(
0,
(_aux[x - 1] + 110 * _aux[x] + 18 * _aux[OC_MINI(x + 1, c_w - 1)] -
_aux[c_w - 1] + 64) >>
7,
255);
if ((x << 1 | 1) < dst_c_w) {
tmp[x << 1 | 1] = (unsigned char)OC_CLAMPI(
0,
(-3 * _aux[x - 1] + 50 * _aux[x] +
86 * _aux[OC_MINI(x + 1, c_w - 1)] - 5 * _aux[c_w - 1] + 64) >>
7,
255);
}
}
tmp += dst_c_w;
_aux += c_w;
}
tmp -= tmp_sz;
/*Now do the vertical filtering.*/
y4m_422jpeg_420jpeg_helper(_dst, tmp, dst_c_w, c_h);
_dst += dst_c_sz;
}
}
/*Convert 444 to 420jpeg.*/
static void y4m_convert_444_420jpeg(y4m_input *_y4m, unsigned char *_dst,
unsigned char *_aux) {
unsigned char *tmp;
int c_w;
int c_h;
int c_sz;
int dst_c_w;
int dst_c_h;
int dst_c_sz;
int tmp_sz;
int pli;
int y;
int x;
/*Skip past the luma data.*/
_dst += _y4m->pic_w * _y4m->pic_h;
/*Compute the size of each chroma plane.*/
c_w = (_y4m->pic_w + _y4m->src_c_dec_h - 1) / _y4m->src_c_dec_h;
c_h = _y4m->pic_h;
dst_c_w = (_y4m->pic_w + _y4m->dst_c_dec_h - 1) / _y4m->dst_c_dec_h;
dst_c_h = (_y4m->pic_h + _y4m->dst_c_dec_v - 1) / _y4m->dst_c_dec_v;
c_sz = c_w * c_h;
dst_c_sz = dst_c_w * dst_c_h;
tmp_sz = dst_c_w * c_h;
tmp = _aux + 2 * c_sz;
for (pli = 1; pli < 3; pli++) {
/*Filter: [3 -17 78 78 -17 3]/128, derived from a 6-tap Lanczos window.*/
for (y = 0; y < c_h; y++) {
for (x = 0; x < OC_MINI(c_w, 2); x += 2) {
tmp[x >> 1] = OC_CLAMPI(0,
(64 * _aux[0] + 78 * _aux[OC_MINI(1, c_w - 1)] -
17 * _aux[OC_MINI(2, c_w - 1)] +
3 * _aux[OC_MINI(3, c_w - 1)] + 64) >>
7,
255);
}
for (; x < c_w - 3; x += 2) {
tmp[x >> 1] = OC_CLAMPI(0,
(3 * (_aux[x - 2] + _aux[x + 3]) -
17 * (_aux[x - 1] + _aux[x + 2]) +
78 * (_aux[x] + _aux[x + 1]) + 64) >>
7,
255);
}
for (; x < c_w; x += 2) {
tmp[x >> 1] =
OC_CLAMPI(0,
(3 * (_aux[x - 2] + _aux[c_w - 1]) -
17 * (_aux[x - 1] + _aux[OC_MINI(x + 2, c_w - 1)]) +
78 * (_aux[x] + _aux[OC_MINI(x + 1, c_w - 1)]) + 64) >>
7,
255);
}
tmp += dst_c_w;
_aux += c_w;
}
tmp -= tmp_sz;
/*Now do the vertical filtering.*/
y4m_422jpeg_420jpeg_helper(_dst, tmp, dst_c_w, c_h);
_dst += dst_c_sz;
}
}
/*The image is padded with empty chroma components at 4:2:0.*/
static void y4m_convert_mono_420jpeg(y4m_input *_y4m, unsigned char *_dst,
unsigned char *_aux) {
int c_sz;
(void)_aux;
_dst += _y4m->pic_w * _y4m->pic_h;
c_sz = ((_y4m->pic_w + _y4m->dst_c_dec_h - 1) / _y4m->dst_c_dec_h) *
((_y4m->pic_h + _y4m->dst_c_dec_v - 1) / _y4m->dst_c_dec_v);
memset(_dst, 128, c_sz * 2);
}
/*No conversion function needed.*/
static void y4m_convert_null(y4m_input *_y4m, unsigned char *_dst,
unsigned char *_aux) {
(void)_y4m;
(void)_dst;
(void)_aux;
}
int y4m_input_open(y4m_input *_y4m, FILE *_fin, char *_skip, int _nskip,
int only_420) {
char buffer[80] = { 0 };
int ret;
int i;
/*Read until newline, or 80 cols, whichever happens first.*/
for (i = 0; i < 79; i++) {
if (_nskip > 0) {
buffer[i] = *_skip++;
_nskip--;
} else {
if (!file_read(buffer + i, 1, _fin)) return -1;
}
if (buffer[i] == '\n') break;
}
/*We skipped too much header data.*/
if (_nskip > 0) return -1;
if (i == 79) {
fprintf(stderr, "Error parsing header; not a YUV2MPEG2 file?\n");
return -1;
}
buffer[i] = '\0';
if (memcmp(buffer, "YUV4MPEG", 8)) {
fprintf(stderr, "Incomplete magic for YUV4MPEG file.\n");
return -1;
}
if (buffer[8] != '2') {
fprintf(stderr, "Incorrect YUV input file version; YUV4MPEG2 required.\n");
}
ret = y4m_parse_tags(_y4m, buffer + 5);
if (ret < 0) {
fprintf(stderr, "Error parsing YUV4MPEG2 header.\n");
return ret;
}
if (_y4m->interlace == '?') {
fprintf(stderr,
"Warning: Input video interlacing format unknown; "
"assuming progressive scan.\n");
} else if (_y4m->interlace != 'p') {
fprintf(stderr,
"Input video is interlaced; "
"Only progressive scan handled.\n");
return -1;
}
_y4m->vpx_fmt = VPX_IMG_FMT_I420;
_y4m->bps = 12;
_y4m->bit_depth = 8;
if (strcmp(_y4m->chroma_type, "420") == 0 ||
strcmp(_y4m->chroma_type, "420jpeg") == 0) {
_y4m->src_c_dec_h = _y4m->dst_c_dec_h = _y4m->src_c_dec_v =
_y4m->dst_c_dec_v = 2;
_y4m->dst_buf_read_sz =
_y4m->pic_w * _y4m->pic_h +
2 * ((_y4m->pic_w + 1) / 2) * ((_y4m->pic_h + 1) / 2);
/* Natively supported: no conversion required. */
_y4m->aux_buf_sz = _y4m->aux_buf_read_sz = 0;
_y4m->convert = y4m_convert_null;
} else if (strcmp(_y4m->chroma_type, "420p10") == 0) {
_y4m->src_c_dec_h = 2;
_y4m->dst_c_dec_h = 2;
_y4m->src_c_dec_v = 2;
_y4m->dst_c_dec_v = 2;
_y4m->dst_buf_read_sz =
2 * (_y4m->pic_w * _y4m->pic_h +
2 * ((_y4m->pic_w + 1) / 2) * ((_y4m->pic_h + 1) / 2));
/* Natively supported: no conversion required. */
_y4m->aux_buf_sz = _y4m->aux_buf_read_sz = 0;
_y4m->convert = y4m_convert_null;
_y4m->bit_depth = 10;
_y4m->bps = 15;
_y4m->vpx_fmt = VPX_IMG_FMT_I42016;
if (only_420) {
fprintf(stderr, "Unsupported conversion from 420p10 to 420jpeg\n");
return -1;
}
} else if (strcmp(_y4m->chroma_type, "420p12") == 0) {
_y4m->src_c_dec_h = 2;
_y4m->dst_c_dec_h = 2;
_y4m->src_c_dec_v = 2;
_y4m->dst_c_dec_v = 2;
_y4m->dst_buf_read_sz =
2 * (_y4m->pic_w * _y4m->pic_h +
2 * ((_y4m->pic_w + 1) / 2) * ((_y4m->pic_h + 1) / 2));
/* Natively supported: no conversion required. */
_y4m->aux_buf_sz = _y4m->aux_buf_read_sz = 0;
_y4m->convert = y4m_convert_null;
_y4m->bit_depth = 12;
_y4m->bps = 18;
_y4m->vpx_fmt = VPX_IMG_FMT_I42016;
if (only_420) {
fprintf(stderr, "Unsupported conversion from 420p12 to 420jpeg\n");
return -1;
}
} else if (strcmp(_y4m->chroma_type, "420mpeg2") == 0) {
_y4m->src_c_dec_h = _y4m->dst_c_dec_h = _y4m->src_c_dec_v =
_y4m->dst_c_dec_v = 2;
_y4m->dst_buf_read_sz = _y4m->pic_w * _y4m->pic_h;
/*Chroma filter required: read into the aux buf first.*/
_y4m->aux_buf_sz = _y4m->aux_buf_read_sz =
2 * ((_y4m->pic_w + 1) / 2) * ((_y4m->pic_h + 1) / 2);
_y4m->convert = y4m_convert_42xmpeg2_42xjpeg;
} else if (strcmp(_y4m->chroma_type, "420paldv") == 0) {
_y4m->src_c_dec_h = _y4m->dst_c_dec_h = _y4m->src_c_dec_v =
_y4m->dst_c_dec_v = 2;
_y4m->dst_buf_read_sz = _y4m->pic_w * _y4m->pic_h;
/*Chroma filter required: read into the aux buf first.
We need to make two filter passes, so we need some extra space in the
aux buffer.*/
_y4m->aux_buf_sz = 3 * ((_y4m->pic_w + 1) / 2) * ((_y4m->pic_h + 1) / 2);
_y4m->aux_buf_read_sz =
2 * ((_y4m->pic_w + 1) / 2) * ((_y4m->pic_h + 1) / 2);
_y4m->convert = y4m_convert_42xpaldv_42xjpeg;
} else if (strcmp(_y4m->chroma_type, "422jpeg") == 0) {
_y4m->src_c_dec_h = _y4m->dst_c_dec_h = 2;
_y4m->src_c_dec_v = 1;
_y4m->dst_c_dec_v = 2;
_y4m->dst_buf_read_sz = _y4m->pic_w * _y4m->pic_h;
/*Chroma filter required: read into the aux buf first.*/
_y4m->aux_buf_sz = _y4m->aux_buf_read_sz =
2 * ((_y4m->pic_w + 1) / 2) * _y4m->pic_h;
_y4m->convert = y4m_convert_422jpeg_420jpeg;
} else if (strcmp(_y4m->chroma_type, "422") == 0) {
_y4m->src_c_dec_h = 2;
_y4m->src_c_dec_v = 1;
if (only_420) {
_y4m->dst_c_dec_h = 2;
_y4m->dst_c_dec_v = 2;
_y4m->dst_buf_read_sz = _y4m->pic_w * _y4m->pic_h;
/*Chroma filter required: read into the aux buf first.
We need to make two filter passes, so we need some extra space in the
aux buffer.*/
_y4m->aux_buf_read_sz = 2 * ((_y4m->pic_w + 1) / 2) * _y4m->pic_h;
_y4m->aux_buf_sz =
_y4m->aux_buf_read_sz + ((_y4m->pic_w + 1) / 2) * _y4m->pic_h;
_y4m->convert = y4m_convert_422_420jpeg;
} else {
_y4m->vpx_fmt = VPX_IMG_FMT_I422;
_y4m->bps = 16;
_y4m->dst_c_dec_h = _y4m->src_c_dec_h;
_y4m->dst_c_dec_v = _y4m->src_c_dec_v;
_y4m->dst_buf_read_sz =
_y4m->pic_w * _y4m->pic_h + 2 * ((_y4m->pic_w + 1) / 2) * _y4m->pic_h;
/*Natively supported: no conversion required.*/
_y4m->aux_buf_sz = _y4m->aux_buf_read_sz = 0;
_y4m->convert = y4m_convert_null;
}
} else if (strcmp(_y4m->chroma_type, "422p10") == 0) {
_y4m->src_c_dec_h = 2;
_y4m->src_c_dec_v = 1;
_y4m->vpx_fmt = VPX_IMG_FMT_I42216;
_y4m->bps = 20;
_y4m->bit_depth = 10;
_y4m->dst_c_dec_h = _y4m->src_c_dec_h;
_y4m->dst_c_dec_v = _y4m->src_c_dec_v;
_y4m->dst_buf_read_sz = 2 * (_y4m->pic_w * _y4m->pic_h +
2 * ((_y4m->pic_w + 1) / 2) * _y4m->pic_h);
_y4m->aux_buf_sz = _y4m->aux_buf_read_sz = 0;
_y4m->convert = y4m_convert_null;
if (only_420) {
fprintf(stderr, "Unsupported conversion from 422p10 to 420jpeg\n");
return -1;
}
} else if (strcmp(_y4m->chroma_type, "422p12") == 0) {
_y4m->src_c_dec_h = 2;
_y4m->src_c_dec_v = 1;
_y4m->vpx_fmt = VPX_IMG_FMT_I42216;
_y4m->bps = 24;
_y4m->bit_depth = 12;
_y4m->dst_c_dec_h = _y4m->src_c_dec_h;
_y4m->dst_c_dec_v = _y4m->src_c_dec_v;
_y4m->dst_buf_read_sz = 2 * (_y4m->pic_w * _y4m->pic_h +
2 * ((_y4m->pic_w + 1) / 2) * _y4m->pic_h);
_y4m->aux_buf_sz = _y4m->aux_buf_read_sz = 0;
_y4m->convert = y4m_convert_null;
if (only_420) {
fprintf(stderr, "Unsupported conversion from 422p12 to 420jpeg\n");
return -1;
}
} else if (strcmp(_y4m->chroma_type, "411") == 0) {
_y4m->src_c_dec_h = 4;
_y4m->dst_c_dec_h = 2;
_y4m->src_c_dec_v = 1;
_y4m->dst_c_dec_v = 2;
_y4m->dst_buf_read_sz = _y4m->pic_w * _y4m->pic_h;
/*Chroma filter required: read into the aux buf first.
We need to make two filter passes, so we need some extra space in the
aux buffer.*/
_y4m->aux_buf_read_sz = 2 * ((_y4m->pic_w + 3) / 4) * _y4m->pic_h;
_y4m->aux_buf_sz =
_y4m->aux_buf_read_sz + ((_y4m->pic_w + 1) / 2) * _y4m->pic_h;
_y4m->convert = y4m_convert_411_420jpeg;
fprintf(stderr, "Unsupported conversion from yuv 411\n");
return -1;
} else if (strcmp(_y4m->chroma_type, "444") == 0) {
_y4m->src_c_dec_h = 1;
_y4m->src_c_dec_v = 1;
if (only_420) {
_y4m->dst_c_dec_h = 2;
_y4m->dst_c_dec_v = 2;
_y4m->dst_buf_read_sz = _y4m->pic_w * _y4m->pic_h;
/*Chroma filter required: read into the aux buf first.
We need to make two filter passes, so we need some extra space in the
aux buffer.*/
_y4m->aux_buf_read_sz = 2 * _y4m->pic_w * _y4m->pic_h;
_y4m->aux_buf_sz =
_y4m->aux_buf_read_sz + ((_y4m->pic_w + 1) / 2) * _y4m->pic_h;
_y4m->convert = y4m_convert_444_420jpeg;
} else {
_y4m->vpx_fmt = VPX_IMG_FMT_I444;
_y4m->bps = 24;
_y4m->dst_c_dec_h = _y4m->src_c_dec_h;
_y4m->dst_c_dec_v = _y4m->src_c_dec_v;
_y4m->dst_buf_read_sz = 3 * _y4m->pic_w * _y4m->pic_h;
/*Natively supported: no conversion required.*/