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
|
#include "cacher.h"
#include "project/clip.h"
#include "project/sequence.h"
#include "project/transition.h"
#include "project/footage.h"
#include "playback/audio.h"
#include "playback/playback.h"
#include "project/effect.h"
#include "panels/timeline.h"
#include "panels/project.h"
#include "playback/audio.h"
#include "panels/panels.h"
#include "panels/viewer.h"
#include "project/media.h"
#include "debug.h"
extern "C" {
#include <libavformat/avformat.h>
#include <libavcodec/avcodec.h>
#include <libswscale/swscale.h>
#include <libswresample/swresample.h>
#include <libavfilter/avfilter.h>
#include <libavfilter/buffersrc.h>
#include <libavfilter/buffersink.h>
#include <libavutil/opt.h>
#include <libavutil/pixdesc.h>
}
#include <QOpenGLFramebufferObject>
#include <QtMath>
#include <QAudioOutput>
#include <math.h>
// temp debug shit
//#define AUDIOWARNINGS
//int dest_format = AV_PIX_FMT_RGBA;
double bytes_to_seconds(int nb_bytes, int nb_channels, int sample_rate) {
return ((double) (nb_bytes >> 1) / nb_channels / sample_rate);
}
void apply_audio_effects(Clip* c, double timecode_start, AVFrame* frame, int nb_bytes, QVector<Clip*> nests) {
// perform all audio effects
double timecode_end;
timecode_end = timecode_start + bytes_to_seconds(nb_bytes, frame->channels, frame->sample_rate);
for (int j=0;j<c->effects.size();j++) {
Effect* e = c->effects.at(j);
if (e->is_enabled()) e->process_audio(timecode_start, timecode_end, frame->data[0], nb_bytes, 2);
}
if (c->get_opening_transition() != NULL) {
if (c->media != NULL && c->media->get_type() == MEDIA_TYPE_FOOTAGE) {
double transition_start = (c->get_clip_in_with_transition() / c->sequence->frame_rate);
double transition_end = (c->get_clip_in_with_transition() + c->get_opening_transition()->get_length()) / c->sequence->frame_rate;
if (timecode_end < transition_end) {
double adjustment = transition_end - transition_start;
double adjusted_range_start = (timecode_start - transition_start) / adjustment;
double adjusted_range_end = (timecode_end - transition_start) / adjustment;
c->get_opening_transition()->process_audio(adjusted_range_start, adjusted_range_end, frame->data[0], nb_bytes, TA_OPENING_TRANSITION);
}
}
}
if (c->get_closing_transition() != NULL) {
if (c->media != NULL && c->media->get_type() == MEDIA_TYPE_FOOTAGE) {
long length_with_transitions = c->get_timeline_out_with_transition() - c->get_timeline_in_with_transition();
double transition_start = (c->get_clip_in_with_transition() + length_with_transitions - c->get_closing_transition()->get_length()) / c->sequence->frame_rate;
double transition_end = (c->get_clip_in_with_transition() + length_with_transitions) / c->sequence->frame_rate;
if (timecode_start > transition_start) {
double adjustment = transition_end - transition_start;
double adjusted_range_start = (timecode_start - transition_start) / adjustment;
double adjusted_range_end = (timecode_end - transition_start) / adjustment;
c->get_closing_transition()->process_audio(adjusted_range_start, adjusted_range_end, frame->data[0], nb_bytes, TA_CLOSING_TRANSITION);
}
}
}
if (!nests.isEmpty()) {
Clip* next_nest = nests.last();
nests.removeLast();
apply_audio_effects(next_nest, timecode_start + (((double)c->get_timeline_in_with_transition()-c->get_clip_in_with_transition())/c->sequence->frame_rate), frame, nb_bytes, nests);
}
}
#define AUDIO_BUFFER_PADDING 2048
void cache_audio_worker(Clip* c, bool scrubbing, QVector<Clip*>& nests) {
long timeline_in = c->get_timeline_in_with_transition();
long timeline_out = c->get_timeline_out_with_transition();
long target_frame = c->audio_target_frame;
long frame_skip = 0;
double last_fr = c->sequence->frame_rate;
if (!nests.isEmpty()) {
for (int i=nests.size()-1;i>=0;i--) {
timeline_in = refactor_frame_number(timeline_in, last_fr, nests.at(i)->sequence->frame_rate) + nests.at(i)->get_timeline_in_with_transition() - nests.at(i)->get_clip_in_with_transition();
timeline_out = refactor_frame_number(timeline_out, last_fr, nests.at(i)->sequence->frame_rate) + nests.at(i)->get_timeline_in_with_transition() - nests.at(i)->get_clip_in_with_transition();
target_frame = refactor_frame_number(target_frame, last_fr, nests.at(i)->sequence->frame_rate) + nests.at(i)->get_timeline_in_with_transition() - nests.at(i)->get_clip_in_with_transition();
timeline_out = qMin(timeline_out, nests.at(i)->get_timeline_out_with_transition());
frame_skip = refactor_frame_number(frame_skip, last_fr, nests.at(i)->sequence->frame_rate);
long validator = nests.at(i)->get_timeline_in_with_transition() - timeline_in;
if (validator > 0) {
frame_skip += validator;
//timeline_in = nests.at(i)->get_timeline_in_with_transition();
}
last_fr = nests.at(i)->sequence->frame_rate;
}
}
while (true) {
AVFrame* frame;
int nb_bytes = INT_MAX;
if (c->media == NULL) {
frame = c->frame;
nb_bytes = frame->nb_samples * av_get_bytes_per_sample(static_cast<AVSampleFormat>(frame->format)) * frame->channels;
while ((c->frame_sample_index == -1 || c->frame_sample_index >= nb_bytes) && nb_bytes > 0) {
// create "new frame"
memset(c->frame->data[0], 0, nb_bytes);
apply_audio_effects(c, bytes_to_seconds(frame->pts, frame->channels, frame->sample_rate), frame, nb_bytes, nests);
c->frame->pts += nb_bytes;
c->frame_sample_index = 0;
if (c->audio_buffer_write == 0) {
c->audio_buffer_write = get_buffer_offset_from_frame(last_fr, qMax(timeline_in, target_frame));
}
int offset = audio_ibuffer_read - c->audio_buffer_write;
if (offset > 0) {
c->audio_buffer_write += offset;
c->frame_sample_index += offset;
}
}
} else if (c->media->get_type() == MEDIA_TYPE_FOOTAGE) {
double timebase = av_q2d(c->stream->time_base);
frame = c->queue.at(0);
// retrieve frame
bool new_frame = false;
while ((c->frame_sample_index == -1 || c->frame_sample_index >= nb_bytes) && nb_bytes > 0) {
// no more audio left in frame, get a new one
if (!c->reached_end) {
int loop = 0;
if (c->reverse && !c->audio_just_reset) {
avcodec_flush_buffers(c->codecCtx);
c->reached_end = false;
int64_t backtrack_seek = qMax(c->reverse_target - static_cast<int64_t>(av_q2d(av_inv_q(c->stream->time_base))), static_cast<int64_t>(0));
av_seek_frame(c->formatCtx, c->stream->index, backtrack_seek, AVSEEK_FLAG_BACKWARD);
#ifdef AUDIOWARNINGS
if (backtrack_seek == 0) {
dout << "backtracked to 0";
}
#endif
}
do {
av_frame_unref(frame);
int ret;
while ((ret = av_buffersink_get_frame(c->buffersink_ctx, frame)) == AVERROR(EAGAIN)) {
ret = retrieve_next_frame(c, c->frame);
if (ret >= 0) {
if ((ret = av_buffersrc_add_frame_flags(c->buffersrc_ctx, c->frame, AV_BUFFERSRC_FLAG_KEEP_REF)) < 0) {
dout << "[ERROR] Could not feed filtergraph -" << ret;
break;
}
} else {
if (ret == AVERROR_EOF) {
#ifdef AUDIOWARNINGS
dout << "reached EOF while reading";
#endif
// TODO revise usage of reached_end in audio
if (!c->reverse) {
c->reached_end = true;
} else {
}
} else {
dout << "[WARNING] Raw audio frame data could not be retrieved." << ret;
c->reached_end = true;
}
break;
}
}
if (ret < 0) {
if (ret != AVERROR_EOF) {
dout << "[ERROR] Could not pull from filtergraph";
c->reached_end = true;
break;
} else {
#ifdef AUDIOWARNINGS
dout << "reached EOF while pulling from filtergraph";
#endif
if (!c->reverse) break;
}
}
if (c->reverse) {
if (loop > 1) {
AVFrame* rev_frame = c->queue.at(1);
if (ret != AVERROR_EOF) {
if (loop == 2) {
#ifdef AUDIOWARNINGS
dout << "starting rev_frame";
#endif
rev_frame->nb_samples = 0;
rev_frame->pts = c->frame->pkt_pts;
}
int offset = rev_frame->nb_samples * av_get_bytes_per_sample(static_cast<AVSampleFormat>(rev_frame->format)) * rev_frame->channels;
#ifdef AUDIOWARNINGS
dout << "offset 1:" << offset;
dout << "retrieved samples:" << frame->nb_samples << "size:" << (frame->nb_samples * av_get_bytes_per_sample(static_cast<AVSampleFormat>(frame->format)) * frame->channels);
#endif
memcpy(
rev_frame->data[0]+offset,
frame->data[0],
(frame->nb_samples * av_get_bytes_per_sample(static_cast<AVSampleFormat>(frame->format)) * frame->channels)
);
#ifdef AUDIOWARNINGS
dout << "pts:" << c->frame->pts << "dur:" << c->frame->pkt_duration << "rev_target:" << c->reverse_target << "offset:" << offset << "limit:" << rev_frame->linesize[0];
#endif
}
rev_frame->nb_samples += frame->nb_samples;
if ((c->frame->pts >= c->reverse_target) || (ret == AVERROR_EOF)) {
/*
#ifdef AUDIOWARNINGS
dout << "time for the end of rev cache" << rev_frame->nb_samples << c->rev_target << c->frame->pts << c->frame->pkt_duration << c->frame->nb_samples;
dout << "diff:" << (c->frame->pkt_pts + c->frame->pkt_duration) - c->rev_target;
#endif
int cutoff = qRound64((((c->frame->pkt_pts + c->frame->pkt_duration) - c->reverse_target) * timebase) * audio_output->format().sampleRate());
if (cutoff > 0) {
#ifdef AUDIOWARNINGS
dout << "cut off" << cutoff << "samples (rate:" << audio_output->format().sampleRate() << ")";
#endif
rev_frame->nb_samples -= cutoff;
}
*/
#ifdef AUDIOWARNINGS
dout << "pre cutoff deets::: rev_frame.pts:" << rev_frame->pts << "rev_frame.nb_samples" << rev_frame->nb_samples << "rev_target:" << c->reverse_target;
#endif
rev_frame->nb_samples = qRound64(static_cast<double>(c->reverse_target - rev_frame->pts) / c->stream->codecpar->sample_rate * (audio_output->format().sampleRate() / c->speed));
#ifdef AUDIOWARNINGS
dout << "post cutoff deets::" << rev_frame->nb_samples;
#endif
int frame_size = rev_frame->nb_samples * rev_frame->channels * av_get_bytes_per_sample(static_cast<AVSampleFormat>(rev_frame->format));
int half_frame_size = frame_size >> 1;
int sample_size = rev_frame->channels*av_get_bytes_per_sample(static_cast<AVSampleFormat>(rev_frame->format));
char* temp_chars = new char[sample_size];
for (int i=0;i<half_frame_size;i+=sample_size) {
for (int j=0;j<sample_size;j++) {
temp_chars[j] = rev_frame->data[0][i+j];
}
for (int j=0;j<sample_size;j++) {
rev_frame->data[0][i+j] = rev_frame->data[0][frame_size-i-sample_size+j];
}
for (int j=0;j<sample_size;j++) {
rev_frame->data[0][frame_size-i-sample_size+j] = temp_chars[j];
}
}
delete [] temp_chars;
c->reverse_target = rev_frame->pts;
frame = rev_frame;
break;
}
}
loop++;
#ifdef AUDIOWARNINGS
dout << "loop" << loop;
#endif
} else {
frame->pts = c->frame->pts;
break;
}
} while (true);
} else {
// if there is no more data in the file, we flush the remainder out of swresample
break;
}
new_frame = true;
if (c->frame_sample_index < 0) {
c->frame_sample_index = 0;
} else {
c->frame_sample_index -= nb_bytes;
}
nb_bytes = frame->nb_samples * av_get_bytes_per_sample(static_cast<AVSampleFormat>(frame->format)) * frame->channels;
if (c->audio_just_reset) {
// get precise sample offset for the elected clip_in from this audio frame
double target_sts = playhead_to_clip_seconds(c, c->audio_target_frame);
double frame_sts = ((frame->pts - c->stream->start_time) * timebase);
int nb_samples = qRound64((target_sts - frame_sts)*audio_output->format().sampleRate());
c->frame_sample_index = nb_samples * 4;
#ifdef AUDIOWARNINGS
dout << "fsts:" << frame_sts << "tsts:" << target_sts << "nbs:" << nb_samples << "nbb:" << nb_bytes << "rev_targetToSec:" << (c->reverse_target * timebase);
dout << "fsi-calc:" << c->frame_sample_index;
#endif
if (c->reverse) c->frame_sample_index = nb_bytes - c->frame_sample_index;
c->audio_just_reset = false;
}
#ifdef AUDIOWARNINGS
dout << "fsi-post-post:" << c->frame_sample_index;
#endif
if (c->audio_buffer_write == 0) {
c->audio_buffer_write = get_buffer_offset_from_frame(last_fr, qMax(timeline_in, target_frame));
if (frame_skip > 0) {
int target = get_buffer_offset_from_frame(last_fr, qMax(timeline_in + frame_skip, target_frame));
c->frame_sample_index += (target - c->audio_buffer_write);
c->audio_buffer_write = target;
}
}
int offset = audio_ibuffer_read - c->audio_buffer_write;
if (offset > 0) {
c->audio_buffer_write += offset;
c->frame_sample_index += offset;
}
// try to correct negative fsi
if (c->frame_sample_index < 0) {
c->audio_buffer_write -= c->frame_sample_index;
c->frame_sample_index = 0;
}
}
if (c->reverse) frame = c->queue.at(1);
#ifdef AUDIOWARNINGS
dout << "j" << c->frame_sample_index << nb_bytes;
#endif
// apply any audio effects to the data
if (nb_bytes == INT_MAX) nb_bytes = frame->nb_samples * av_get_bytes_per_sample(static_cast<AVSampleFormat>(frame->format)) * frame->channels;
if (new_frame) {
apply_audio_effects(c, bytes_to_seconds(c->audio_buffer_write, 2, audio_output->format().sampleRate()) + audio_ibuffer_timecode + ((double)c->get_clip_in_with_transition()/c->sequence->frame_rate) - ((double)timeline_in/last_fr), frame, nb_bytes, nests);
}
} else {
// shouldn't ever get here
dout << "[ERROR] Tried to cache a non-footage/tone clip";
return;
}
// mix audio into internal buffer
if (frame->nb_samples == 0) {
break;
} else {
long buffer_timeline_out = get_buffer_offset_from_frame(c->sequence->frame_rate, timeline_out);
audio_write_lock.lock();
while (c->frame_sample_index < nb_bytes
&& c->audio_buffer_write < audio_ibuffer_read+(audio_ibuffer_size>>1)
&& c->audio_buffer_write < buffer_timeline_out) {
int upper_byte_index = (c->audio_buffer_write+1)%audio_ibuffer_size;
int lower_byte_index = (c->audio_buffer_write)%audio_ibuffer_size;
qint16 old_sample = static_cast<qint16>((audio_ibuffer[upper_byte_index] & 0xFF) << 8 | (audio_ibuffer[lower_byte_index] & 0xFF));
qint16 new_sample = static_cast<qint16>((frame->data[0][c->frame_sample_index+1] & 0xFF) << 8 | (frame->data[0][c->frame_sample_index] & 0xFF));
qint16 mixed_sample = mix_audio_sample(old_sample, new_sample);
audio_ibuffer[upper_byte_index] = static_cast<quint8>((mixed_sample >> 8) & 0xFF);
audio_ibuffer[lower_byte_index] = static_cast<quint8>(mixed_sample & 0xFF);
c->audio_buffer_write+=2;
c->frame_sample_index+=2;
}
#ifdef AUDIOWARNINGS
if (c->audio_buffer_write >= buffer_timeline_out) dout << "timeline out at fsi" << c->frame_sample_index << "of frame ts" << c->frame->pts;
#endif
audio_write_lock.unlock();
if (scrubbing) {
if (audio_thread != NULL) audio_thread->notifyReceiver();
}
if (c->frame_sample_index == nb_bytes) {
c->frame_sample_index = -1;
} else {
// assume we have no more data to send
break;
}
// dout << "ended" << c->frame_sample_index << nb_bytes;
}
if (c->reached_end) {
frame->nb_samples = 0;
}
if (scrubbing) {
break;
}
}
QMetaObject::invokeMethod(panel_footage_viewer, "play_wake", Qt::QueuedConnection);
QMetaObject::invokeMethod(panel_sequence_viewer, "play_wake", Qt::QueuedConnection);
}
void cache_video_worker(Clip* c, long playhead) {
int read_ret, send_ret, retr_ret;
int64_t target_pts = seconds_to_timestamp(c, playhead_to_clip_seconds(c, playhead));
int limit = c->max_queue_size;
if (c->ignore_reverse) {
// waiting for one frame
limit = c->queue.size() + 1;
} else if (c->reverse) {
limit *= 2;
}
if (c->queue.size() < limit) {
bool reverse = (c->reverse && !c->ignore_reverse);
c->ignore_reverse = false;
int64_t smallest_pts = INT64_MAX;
if (reverse && c->queue.size() > 0) {
int64_t quarter_sec = qRound64(av_q2d(av_inv_q(c->stream->time_base))) >> 2;
for (int i=0;i<c->queue.size();i++) {
smallest_pts = qMin(smallest_pts, c->queue.at(i)->pts);
}
avcodec_flush_buffers(c->codecCtx);
c->reached_end = false;
int64_t seek_ts = qMax(static_cast<int64_t>(0), smallest_pts - quarter_sec);
av_seek_frame(c->formatCtx, c->stream->index, seek_ts, AVSEEK_FLAG_BACKWARD);
} else {
smallest_pts = target_pts;
}
if (c->multithreaded && c->cacher->interrupt) { // ignore interrupts for now
c->cacher->interrupt = false;
}
while (true) {
AVFrame* frame = av_frame_alloc();
Footage* media = c->media->to_footage();
FootageStream* ms = media->get_stream_from_file_index(true, c->media_stream);
while ((retr_ret = av_buffersink_get_frame(c->buffersink_ctx, frame)) == AVERROR(EAGAIN)) {
if (c->multithreaded && c->cacher->interrupt) return; // abort
AVFrame* send_frame = c->frame;
read_ret = (c->use_existing_frame) ? 0 : retrieve_next_frame(c, send_frame);
c->use_existing_frame = false;
if (read_ret >= 0) {
bool send_it = true;
/*if (reverse) {
send_it = true;
} else if (send_frame->pts > target_pts - eighth_second) {
send_it = true;
} else if (media->get_stream_from_file_index(true, c->media_stream)->infinite_length) {
send_it = true;
} else {
dout << "skipped adding a frame to the queue - fpts:" << send_frame->pts << "target:" << target_pts;
}*/
if (send_it) {
if ((send_ret = av_buffersrc_add_frame_flags(c->buffersrc_ctx, send_frame, AV_BUFFERSRC_FLAG_KEEP_REF)) < 0) {
dout << "[ERROR] Failed to add frame to buffer source." << send_ret;
break;
}
}
av_frame_unref(c->frame);
} else {
if (read_ret == AVERROR_EOF) {
c->reached_end = true;
} else {
dout << "[ERROR] Failed to read frame." << read_ret;
}
break;
}
}
if (retr_ret < 0) {
if (retr_ret == AVERROR_EOF) {
c->reached_end = true;
} else {
dout << "[ERROR] Failed to retrieve frame from buffersink." << retr_ret;
}
av_frame_free(&frame);
break;
} else {
if (reverse && ((smallest_pts == target_pts && frame->pts >= smallest_pts) || (smallest_pts != target_pts && frame->pts > smallest_pts))) {
av_frame_free(&frame);
break;
} else {
// thread-safety while adding frame to the queue
c->queue_lock.lock();
c->queue.append(frame);
if (!ms->infinite_length && !reverse && c->queue.size() == limit) {
// see if we got the frame we needed (used for speed ups primarily)
bool found = false;
for (int i=0;i<c->queue.size();i++) {
if (c->queue.at(i)->pts >= target_pts) {
found = true;
break;
}
}
if (found) {
c->queue_lock.unlock();
break;
} else {
// remove earliest frame and loop to store another
c->queue_remove_earliest();
}
}
c->queue_lock.unlock();
}
}
if (c->multithreaded && c->cacher->interrupt) { // abort
return;
}
}
}
}
void reset_cache(Clip* c, long target_frame) {
// if we seek to a whole other place in the timeline, we'll need to reset the cache with new values
if (c->media == NULL) {
if (c->track >= 0) {
// tone clip
c->reached_end = false;
c->audio_target_frame = target_frame;
c->frame_sample_index = -1;
c->frame->pts = 0;
}
} else {
FootageStream* ms = c->media->to_footage()->get_stream_from_file_index(c->track < 0, c->media_stream);
if (ms->infinite_length) {
/*avcodec_flush_buffers(c->codecCtx);
av_seek_frame(c->formatCtx, ms->file_index, 0, AVSEEK_FLAG_BACKWARD);*/
c->use_existing_frame = false;
} else {
if (c->stream->codecpar->codec_type == AVMEDIA_TYPE_VIDEO) {
// clear current queue
c->queue_clear();
// seeks to nearest keyframe (target_frame represents internal clip frame)
int64_t target_ts = seconds_to_timestamp(c, playhead_to_clip_seconds(c, target_frame));
int64_t seek_ts = target_ts;
int64_t timebase_half_second = qRound64(av_q2d(av_inv_q(c->stream->time_base)));
if (c->reverse) seek_ts -= timebase_half_second;
while (true) {
// flush ffmpeg codecs
avcodec_flush_buffers(c->codecCtx);
c->reached_end = false;
if (seek_ts > 0) {
av_seek_frame(c->formatCtx, ms->file_index, seek_ts, AVSEEK_FLAG_BACKWARD);
av_frame_unref(c->frame);
int ret = retrieve_next_frame(c, c->frame);
if (ret < 0) {
dout << "[WARNING] Seeking terminated prematurely";
break;
}
if (c->frame->pts <= target_ts) {
c->use_existing_frame = true;
break;
} else {
seek_ts -= timebase_half_second;
}
} else {
av_frame_unref(c->frame);
av_seek_frame(c->formatCtx, ms->file_index, 0, AVSEEK_FLAG_BACKWARD);
c->use_existing_frame = false;
break;
}
}
} else if (c->stream->codecpar->codec_type == AVMEDIA_TYPE_AUDIO) {
// flush ffmpeg codecs
avcodec_flush_buffers(c->codecCtx);
c->reached_end = false;
// seek (target_frame represents timeline timecode in frames, not clip timecode)
int64_t timestamp = seconds_to_timestamp(c, playhead_to_clip_seconds(c, target_frame));
if (c->reverse) {
c->reverse_target = timestamp;
timestamp -= av_q2d(av_inv_q(c->stream->time_base));
#ifdef AUDIOWARNINGS
dout << "seeking to" << timestamp << "(originally" << c->reverse_target << ")";
} else {
dout << "reset called; seeking to" << timestamp;
#endif
}
av_seek_frame(c->formatCtx, ms->file_index, timestamp, AVSEEK_FLAG_BACKWARD);
c->audio_target_frame = target_frame;
c->frame_sample_index = -1;
c->audio_just_reset = true;
}
}
}
}
Cacher::Cacher(Clip* c) : clip(c) {}
AVSampleFormat sample_format = AV_SAMPLE_FMT_S16;
void open_clip_worker(Clip* clip) {
if (clip->media == NULL) {
if (clip->track >= 0) {
clip->frame = av_frame_alloc();
clip->frame->format = sample_format;
clip->frame->channel_layout = clip->sequence->audio_layout;
clip->frame->channels = av_get_channel_layout_nb_channels(clip->frame->channel_layout);
clip->frame->sample_rate = audio_output->format().sampleRate();
clip->frame->nb_samples = 2048;
av_frame_make_writable(clip->frame);
if (av_frame_get_buffer(clip->frame, 0)) {
dout << "[ERROR] Could not allocate buffer for tone clip";
}
clip->audio_reset = true;
}
} else if (clip->media->get_type() == MEDIA_TYPE_FOOTAGE) {
// opens file resource for FFmpeg and prepares Clip struct for playback
Footage* m = clip->media->to_footage();
QByteArray ba = m->url.toUtf8();
const char* filename = ba.constData();
FootageStream* ms = m->get_stream_from_file_index(clip->track < 0, clip->media_stream);
int errCode = avformat_open_input(
&clip->formatCtx,
filename,
NULL,
NULL
);
if (errCode != 0) {
char err[1024];
av_strerror(errCode, err, 1024);
dout << "[ERROR] Could not open" << filename << "-" << err;
return;
}
errCode = avformat_find_stream_info(clip->formatCtx, NULL);
if (errCode < 0) {
char err[1024];
av_strerror(errCode, err, 1024);
dout << "[ERROR] Could not open" << filename << "-" << err;
return;
}
av_dump_format(clip->formatCtx, 0, filename, 0);
clip->stream = clip->formatCtx->streams[ms->file_index];
clip->codec = avcodec_find_decoder(clip->stream->codecpar->codec_id);
clip->codecCtx = avcodec_alloc_context3(clip->codec);
avcodec_parameters_to_context(clip->codecCtx, clip->stream->codecpar);
clip->max_queue_size = (ms->infinite_length) ? 1 : qCeil(ms->video_frame_rate*0.5);
if (ms->video_interlacing != VIDEO_PROGRESSIVE) clip->max_queue_size *= 2;
clip->opts = NULL;
// optimized decoding settings
if (clip->stream->codecpar->codec_id != AV_CODEC_ID_PNG &&
clip->stream->codecpar->codec_id != AV_CODEC_ID_APNG &&
clip->stream->codecpar->codec_id != AV_CODEC_ID_TIFF &&
clip->stream->codecpar->codec_id != AV_CODEC_ID_PSD) {
av_dict_set(&clip->opts, "threads", "auto", 0);
}
if (clip->stream->codecpar->codec_id == AV_CODEC_ID_H264) {
av_dict_set(&clip->opts, "tune", "fastdecode", 0);
av_dict_set(&clip->opts, "tune", "zerolatency", 0);
}
// Open codec
if (avcodec_open2(clip->codecCtx, clip->codec, &clip->opts) < 0) {
dout << "[ERROR] Could not open codec";
}
// allocate filtergraph
clip->filter_graph = avfilter_graph_alloc();
if (clip->filter_graph == NULL) {
dout << "[ERROR] Could not create filtergraph";
}
char filter_args[512];
if (clip->stream->codecpar->codec_type == AVMEDIA_TYPE_VIDEO) {
snprintf(filter_args, sizeof(filter_args), "video_size=%dx%d:pix_fmt=%d:time_base=%d/%d:pixel_aspect=%d/%d",
clip->stream->codecpar->width,
clip->stream->codecpar->height,
clip->stream->codecpar->format,
clip->stream->time_base.num,
clip->stream->time_base.den,
clip->stream->codecpar->sample_aspect_ratio.num,
clip->stream->codecpar->sample_aspect_ratio.den
);
avfilter_graph_create_filter(&clip->buffersrc_ctx, avfilter_get_by_name("buffer"), "in", filter_args, NULL, clip->filter_graph);
avfilter_graph_create_filter(&clip->buffersink_ctx, avfilter_get_by_name("buffersink"), "out", NULL, NULL, clip->filter_graph);
AVFilterContext* last_filter = clip->buffersrc_ctx;
if (ms->video_interlacing != VIDEO_PROGRESSIVE) {
AVFilterContext* yadif_filter;
char yadif_args[100];
snprintf(yadif_args, sizeof(yadif_args), "mode=3:parity=%d", ((ms->video_interlacing == VIDEO_TOP_FIELD_FIRST) ? 0 : 1)); // there's a CUDA version if we start using nvdec/nvenc
avfilter_graph_create_filter(&yadif_filter, avfilter_get_by_name("yadif"), "yadif", yadif_args, NULL, clip->filter_graph);
avfilter_link(last_filter, 0, yadif_filter, 0);
last_filter = yadif_filter;
}
/* stabilization code */
bool stabilize = false;
if (stabilize) {
AVFilterContext* stab_filter;
int stab_ret = avfilter_graph_create_filter(&stab_filter, avfilter_get_by_name("vidstabtransform"), "vidstab", "input=/media/matt/Home/samples/transforms.trf", NULL, clip->filter_graph);
if (stab_ret < 0) {
char err[100];
av_strerror(stab_ret, err, sizeof(err));
} else {
avfilter_link(last_filter, 0, stab_filter, 0);
last_filter = stab_filter;
}
}
enum AVPixelFormat valid_pix_fmts[] = {
AV_PIX_FMT_RGB24,
AV_PIX_FMT_RGBA,
AV_PIX_FMT_NONE
};
clip->pix_fmt = avcodec_find_best_pix_fmt_of_list(valid_pix_fmts, static_cast<enum AVPixelFormat>(clip->stream->codecpar->format), 1, NULL);
const char* chosen_format = av_get_pix_fmt_name(static_cast<enum AVPixelFormat>(clip->pix_fmt));
char format_args[100];
snprintf(format_args, sizeof(format_args), "pix_fmts=%s", chosen_format);
AVFilterContext* format_conv;
avfilter_graph_create_filter(&format_conv, avfilter_get_by_name("format"), "fmt", format_args, NULL, clip->filter_graph);
avfilter_link(last_filter, 0, format_conv, 0);
avfilter_link(format_conv, 0, clip->buffersink_ctx, 0);
avfilter_graph_config(clip->filter_graph, NULL);
} else if (clip->stream->codecpar->codec_type == AVMEDIA_TYPE_AUDIO) {
if (clip->codecCtx->channel_layout == 0) clip->codecCtx->channel_layout = av_get_default_channel_layout(clip->stream->codecpar->channels);
// set up cache
clip->queue.append(av_frame_alloc());
if (clip->reverse) {
AVFrame* reverse_frame = av_frame_alloc();
reverse_frame->format = sample_format;
reverse_frame->nb_samples = audio_output->format().sampleRate()*2;
reverse_frame->channel_layout = clip->sequence->audio_layout;
reverse_frame->channels = av_get_channel_layout_nb_channels(clip->sequence->audio_layout);
av_frame_get_buffer(reverse_frame, 0);
clip->queue.append(reverse_frame);
}
snprintf(filter_args, sizeof(filter_args), "time_base=%d/%d:sample_rate=%d:sample_fmt=%s:channel_layout=0x%" PRIx64,
clip->stream->time_base.num,
clip->stream->time_base.den,
clip->stream->codecpar->sample_rate,
av_get_sample_fmt_name(clip->codecCtx->sample_fmt),
clip->codecCtx->channel_layout
);
avfilter_graph_create_filter(&clip->buffersrc_ctx, avfilter_get_by_name("abuffer"), "in", filter_args, NULL, clip->filter_graph);
avfilter_graph_create_filter(&clip->buffersink_ctx, avfilter_get_by_name("abuffersink"), "out", NULL, NULL, clip->filter_graph);
enum AVSampleFormat sample_fmts[] = { sample_format, static_cast<AVSampleFormat>(-1) };
if (av_opt_set_int_list(clip->buffersink_ctx, "sample_fmts", sample_fmts, -1, AV_OPT_SEARCH_CHILDREN) < 0) {
dout << "[ERROR] Could not set output sample format";
}
int64_t channel_layouts[] = { AV_CH_LAYOUT_STEREO, static_cast<AVSampleFormat>(-1) };
if (av_opt_set_int_list(clip->buffersink_ctx, "channel_layouts", channel_layouts, -1, AV_OPT_SEARCH_CHILDREN) < 0) {
dout << "[ERROR] Could not set output sample format";
}
int target_sample_rate = audio_output->format().sampleRate();
if (qFuzzyCompare(clip->speed, 1.0)) {
avfilter_link(clip->buffersrc_ctx, 0, clip->buffersink_ctx, 0);
} else if (clip->maintain_audio_pitch) {
AVFilterContext* previous_filter = clip->buffersrc_ctx;
AVFilterContext* last_filter = clip->buffersrc_ctx;
char speed_param[10];
if (clip->speed != 1.0) {
double base = (clip->speed > 1.0) ? 2.0 : 0.5;
double speedlog = log(clip->speed) / log(base);
int whole2 = qFloor(speedlog);
speedlog -= whole2;
if (whole2 > 0) {
snprintf(speed_param, sizeof(speed_param), "%f", base);
for (int i=0;i<whole2;i++) {
AVFilterContext* tempo_filter = NULL;
avfilter_graph_create_filter(&tempo_filter, avfilter_get_by_name("atempo"), "atempo", speed_param, NULL, clip->filter_graph);
avfilter_link(previous_filter, 0, tempo_filter, 0);
previous_filter = tempo_filter;
}
}
snprintf(speed_param, sizeof(speed_param), "%f", qPow(base, speedlog));
last_filter = NULL;
avfilter_graph_create_filter(&last_filter, avfilter_get_by_name("atempo"), "atempo", speed_param, NULL, clip->filter_graph);
avfilter_link(previous_filter, 0, last_filter, 0);
}
avfilter_link(last_filter, 0, clip->buffersink_ctx, 0);
} else {
target_sample_rate = qRound64(target_sample_rate / clip->speed);
avfilter_link(clip->buffersrc_ctx, 0, clip->buffersink_ctx, 0);
}
int sample_rates[] = { target_sample_rate, 0 };
if (av_opt_set_int_list(clip->buffersink_ctx, "sample_rates", sample_rates, 0, AV_OPT_SEARCH_CHILDREN) < 0) {
dout << "[ERROR] Could not set output sample rates";
}
avfilter_graph_config(clip->filter_graph, NULL);
clip->audio_reset = true;
}
clip->frame = av_frame_alloc();
}
for (int i=0;i<clip->effects.size();i++) {
clip->effects.at(i)->open();
}
clip->finished_opening = true;
dout << "[INFO] Clip opened on track" << clip->track;
}
void cache_clip_worker(Clip* clip, long playhead, bool reset, bool scrubbing, QVector<Clip*> nests) {
if (reset) {
// note: for video, playhead is in "internal clip" frames - for audio, it's the timeline playhead
reset_cache(clip, playhead);
clip->audio_reset = false;
}
if (clip->media == NULL) {
if (clip->track >= 0) {
cache_audio_worker(clip, scrubbing, nests);
}
} else if (clip->media->get_type() == MEDIA_TYPE_FOOTAGE) {
if (clip->stream->codecpar->codec_type == AVMEDIA_TYPE_VIDEO) {
cache_video_worker(clip, playhead);
} else if (clip->stream->codecpar->codec_type == AVMEDIA_TYPE_AUDIO) {
cache_audio_worker(clip, scrubbing, nests);
}
}
}
void close_clip_worker(Clip* clip) {
clip->finished_opening = false;
if (clip->media != NULL && clip->media->get_type() == MEDIA_TYPE_FOOTAGE) {
clip->queue_clear();
avfilter_graph_free(&clip->filter_graph);
avcodec_close(clip->codecCtx);
avcodec_free_context(&clip->codecCtx);
av_dict_free(&clip->opts);
avformat_close_input(&clip->formatCtx);
}
av_frame_free(&clip->frame);
clip->reset();
dout << "[INFO] Clip closed on track" << clip->track;
}
void Cacher::run() {
// open_lock is used to prevent the clip from being destroyed before the cacher has closed it properly
clip->lock.lock();
clip->finished_opening = false;
clip->open = true;
caching = true;
interrupt = false;
open_clip_worker(clip);
while (caching) {
clip->can_cache.wait(&clip->lock);
if (!caching) {
break;
} else {
while (true) {
cache_clip_worker(clip, playhead, reset, scrubbing, nests);
if (clip->multithreaded && clip->cacher->interrupt && clip->track < 0) {
clip->cacher->interrupt = false;
} else {
break;
}
}
}
}
close_clip_worker(clip);
clip->lock.unlock();
clip->open_lock.unlock();
}
|
