summaryrefslogtreecommitdiffstats
path: root/libhb/sync.c
blob: c95f9ad26a068990768a157742466cb3cd1dd73a (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
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
1001
1002
1003
1004
1005
1006
1007
1008
1009
1010
1011
1012
1013
1014
1015
1016
1017
1018
1019
1020
1021
1022
1023
1024
1025
1026
1027
1028
1029
1030
1031
1032
1033
1034
1035
1036
1037
1038
1039
1040
1041
1042
1043
1044
1045
1046
1047
1048
1049
1050
1051
1052
1053
1054
1055
1056
1057
1058
1059
1060
1061
1062
1063
1064
1065
1066
1067
1068
1069
1070
1071
1072
1073
1074
1075
1076
1077
1078
1079
1080
1081
1082
1083
1084
1085
1086
1087
1088
1089
1090
1091
1092
1093
1094
1095
1096
1097
1098
1099
1100
1101
1102
1103
1104
1105
1106
1107
1108
1109
1110
1111
1112
1113
1114
1115
1116
1117
1118
1119
1120
1121
1122
1123
1124
1125
1126
1127
1128
1129
1130
1131
1132
1133
1134
1135
1136
1137
1138
1139
1140
1141
1142
1143
1144
1145
1146
1147
1148
1149
1150
1151
1152
1153
1154
1155
1156
1157
1158
1159
1160
1161
1162
1163
1164
1165
1166
1167
1168
1169
1170
1171
1172
1173
1174
1175
1176
1177
1178
1179
1180
1181
1182
1183
1184
1185
1186
1187
1188
1189
1190
1191
1192
1193
1194
1195
1196
1197
1198
1199
1200
1201
1202
1203
1204
1205
1206
1207
1208
1209
1210
1211
1212
1213
1214
1215
1216
1217
1218
1219
1220
1221
1222
1223
1224
1225
1226
1227
1228
1229
1230
1231
1232
1233
1234
1235
1236
1237
1238
1239
1240
1241
1242
1243
1244
1245
1246
1247
1248
1249
1250
1251
1252
1253
1254
1255
1256
1257
1258
1259
1260
1261
1262
1263
1264
1265
1266
1267
1268
1269
1270
1271
1272
1273
1274
1275
1276
1277
1278
1279
1280
1281
1282
1283
1284
1285
1286
1287
1288
1289
1290
1291
1292
1293
1294
1295
1296
1297
1298
1299
1300
1301
1302
1303
1304
1305
1306
1307
1308
1309
1310
1311
1312
1313
1314
1315
1316
1317
1318
1319
1320
1321
1322
1323
1324
1325
1326
1327
1328
1329
1330
1331
1332
1333
1334
1335
1336
1337
1338
1339
1340
1341
1342
1343
1344
1345
1346
1347
1348
1349
1350
1351
1352
1353
1354
1355
1356
1357
1358
1359
1360
1361
1362
1363
1364
1365
1366
1367
1368
1369
1370
1371
1372
1373
1374
1375
1376
1377
1378
1379
1380
1381
1382
1383
1384
1385
1386
1387
1388
1389
1390
1391
1392
1393
1394
1395
1396
1397
1398
1399
1400
1401
1402
1403
1404
1405
1406
1407
1408
1409
1410
1411
1412
1413
1414
1415
1416
1417
1418
1419
1420
1421
1422
1423
1424
1425
1426
1427
1428
1429
1430
1431
1432
1433
1434
1435
1436
1437
1438
1439
1440
1441
1442
1443
1444
1445
1446
1447
1448
1449
1450
1451
1452
1453
1454
1455
1456
1457
1458
1459
1460
1461
1462
1463
1464
1465
1466
1467
1468
1469
1470
1471
1472
1473
1474
1475
1476
1477
1478
1479
1480
1481
1482
1483
1484
1485
1486
1487
1488
1489
1490
1491
1492
1493
1494
1495
1496
1497
1498
1499
1500
1501
1502
1503
1504
1505
1506
1507
1508
1509
1510
1511
1512
1513
1514
1515
1516
1517
1518
1519
1520
1521
1522
1523
1524
1525
1526
1527
1528
1529
1530
1531
1532
1533
1534
1535
1536
1537
1538
1539
1540
1541
1542
1543
1544
1545
1546
1547
1548
1549
1550
1551
1552
1553
1554
1555
1556
1557
1558
1559
1560
1561
1562
1563
1564
1565
1566
1567
1568
1569
1570
1571
1572
1573
1574
1575
1576
1577
1578
1579
1580
1581
1582
1583
1584
1585
1586
1587
1588
1589
1590
1591
1592
1593
1594
1595
1596
1597
1598
1599
1600
1601
1602
1603
1604
1605
1606
1607
1608
1609
1610
1611
1612
1613
1614
1615
1616
1617
1618
1619
1620
1621
1622
1623
1624
1625
1626
1627
1628
1629
1630
1631
1632
/* sync.c

   Copyright (c) 2003-2013 HandBrake Team
   This file is part of the HandBrake source code
   Homepage: <http://handbrake.fr/>.
   It may be used under the terms of the GNU General Public License v2.
   For full terms see the file COPYING file or visit http://www.gnu.org/licenses/gpl-2.0.html
 */

#include "hb.h"
#include "hbffmpeg.h"
#include <stdio.h>
#include "samplerate.h"

#ifdef INT64_MIN
#undef INT64_MIN /* Because it isn't defined correctly in Zeta */
#endif
#define INT64_MIN (-9223372036854775807LL-1)

typedef struct
{
    hb_lock_t * mutex;
    int         ref;        /* Reference count to tell us when it's unused */
    int         count_frames;
    int64_t     audio_pts_slip;
    int64_t     video_pts_slip;
    int64_t     pts_offset;

    /* Frame based point-to-point support */
    int64_t     audio_pts_thresh;
    int         start_found;
    hb_cond_t * next_frame;
    int         pts_count;
    int64_t   * first_pts;
} hb_sync_common_t;

typedef struct
{
    int          index;
    double       next_start;   /* start time of next output frame */
    int64_t      first_drop;   /* PTS of first 'went backwards' frame dropped */
    int          drop_count;   /* count of 'time went backwards' drops */

    /* Raw */
    SRC_STATE  * state;
    SRC_DATA     data;

    int          silence_size;
    uint8_t    * silence_buf;

    int          drop_video_to_sync;

    double       gain_factor;
} hb_sync_audio_t;

typedef struct
{
    int             link;
    int             merge;
    hb_buffer_t   * list_current;
    hb_buffer_t   * last;
} subtitle_sanitizer_t;

typedef struct
{
    /* Video */
    int        first_frame;
    int64_t    pts_skip;
    int64_t    next_start;    /* start time of next output frame */
    int64_t    first_drop;    /* PTS of first 'went backwards' frame dropped */
    int        drop_count;    /* count of 'time went backwards' drops */
    int        drops;         /* frames dropped to make a cbr video stream */
    int        dups;          /* frames duplicated to make a cbr video stream */
    int        video_sequence;
    int        count_frames_max;
    int        chap_mark;     /* to propagate chapter mark across a drop */
    hb_buffer_t * cur;        /* The next picture to process */

    subtitle_sanitizer_t *subtitle_sanitizer;

    /* Statistics */
    uint64_t   st_counts[4];
    uint64_t   st_dates[4];
    uint64_t   st_first;
} hb_sync_video_t;

struct hb_work_private_s
{
    hb_job_t * job;
    hb_sync_common_t * common;
    union
    {
        hb_sync_video_t video;
        hb_sync_audio_t audio;
    } type;
};

/***********************************************************************
 * Local prototypes
 **********************************************************************/
static void getPtsOffset( hb_work_object_t * w );
static int  checkPtsOffset( hb_work_object_t * w );
static void InitAudio( hb_job_t * job, hb_sync_common_t * common, int i );
static void InitSubtitle( hb_job_t * job, hb_sync_video_t * sync, int i );
static void InsertSilence( hb_work_object_t * w, int64_t d );
static void UpdateState( hb_work_object_t * w );
static void UpdateSearchState( hb_work_object_t * w, int64_t start );
static hb_buffer_t * OutputAudioFrame( hb_audio_t *audio, hb_buffer_t *buf,
                                       hb_sync_audio_t *sync );

/***********************************************************************
 * hb_work_sync_init
 ***********************************************************************
 * Initialize the work object
 **********************************************************************/
hb_work_object_t * hb_sync_init( hb_job_t * job )
{
    hb_title_t        * title = job->title;
    hb_chapter_t      * chapter;
    int                 i;
    uint64_t            duration;
    hb_work_private_t * pv;
    hb_sync_video_t   * sync;
    hb_work_object_t  * w;
    hb_work_object_t  * ret = NULL;

    pv = calloc( 1, sizeof( hb_work_private_t ) );
    sync = &pv->type.video;
    pv->common = calloc( 1, sizeof( hb_sync_common_t ) );
    pv->common->ref++;
    pv->common->mutex = hb_lock_init();
    pv->common->audio_pts_thresh = AV_NOPTS_VALUE;
    pv->common->next_frame = hb_cond_init();
    pv->common->pts_count = 1;
    if ( job->frame_to_start || job->pts_to_start )
    {
        pv->common->start_found = 0;
    }
    else
    {
        pv->common->start_found = 1;
    }

    ret = w = hb_get_work( WORK_SYNC_VIDEO );
    w->private_data = pv;
    w->fifo_in = job->fifo_raw;

    // When doing subtitle indepth scan, the pipeline ends at sync
    if ( !job->indepth_scan )
        w->fifo_out = job->fifo_sync;
    else
        w->fifo_out = NULL;

    pv->job            = job;
    pv->common->pts_offset   = INT64_MIN;
    sync->first_frame = 1;

    if( job->pass == 2 )
    {
        /* We already have an accurate frame count from pass 1 */
        hb_interjob_t * interjob = hb_interjob_get( job->h );
        sync->count_frames_max = interjob->frame_count;
    }
    else
    {
        /* Calculate how many video frames we are expecting */
        if ( job->pts_to_stop )
        {
            duration = job->pts_to_stop + 90000;
        }
        else if( job->frame_to_stop )
        {
            /* Set the duration to a rough estimate */
            duration = ( job->frame_to_stop / ( title->rate / title->rate_base ) ) * 90000;
        }
        else
        {
            duration = 0;
            for( i = job->chapter_start; i <= job->chapter_end; i++ )
            {
                chapter   = hb_list_item( job->list_chapter, i - 1 );
                duration += chapter->duration;
            }
        }
        sync->count_frames_max = duration * title->rate / title->rate_base / 90000;
    }

    hb_log( "sync: expecting %d video frames", sync->count_frames_max );

    /* Initialize libsamplerate for every audio track we have */
    if ( ! job->indepth_scan )
    {
        for( i = 0; i < hb_list_count( job->list_audio ); i++ )
        {
            InitAudio( job, pv->common, i );
        }
    }
    pv->common->first_pts = malloc( sizeof(int64_t) * pv->common->pts_count );
    for ( i = 0; i < pv->common->pts_count; i++ )
        pv->common->first_pts[i] = INT64_MAX;

    int count = hb_list_count(job->list_subtitle);
    sync->subtitle_sanitizer = calloc(count, sizeof(subtitle_sanitizer_t));
    for( i = 0; i < count; i++ )
    {
        InitSubtitle(job, sync, i);
    }
    return ret;
}

static void InitSubtitle( hb_job_t * job, hb_sync_video_t * sync, int i )
{
    hb_subtitle_t * subtitle;

    subtitle = hb_list_item( job->list_subtitle, i );
    if (subtitle->format == TEXTSUB &&
        subtitle->config.dest == PASSTHRUSUB &&
        (job->mux & HB_MUX_MASK_MP4))
    {
        // Merge overlapping subtitles since mpv tx3g does not support them
        sync->subtitle_sanitizer[i].merge = 1;
    }
    if (subtitle->config.dest == PASSTHRUSUB)
    {
        // Fill in stop time when it is missing
        sync->subtitle_sanitizer[i].link = 1;
    }
}

static void CloseSubtitle(hb_sync_video_t * sync, int ii)
{
    hb_buffer_close(&sync->subtitle_sanitizer[ii].list_current);
}

/***********************************************************************
 * Close Video
 ***********************************************************************
 *
 **********************************************************************/
void syncVideoClose( hb_work_object_t * w )
{
    hb_work_private_t * pv = w->private_data;
    hb_job_t          * job   = pv->job;
    hb_sync_video_t   * sync = &pv->type.video;
    int ii;

    // Wake up audio sync if it's still waiting on condition.
    pv->common->pts_offset = 0;
    pv->common->start_found = 1;
    hb_cond_broadcast( pv->common->next_frame );

    if( sync->cur )
    {
        hb_buffer_close( &sync->cur );
    }

    hb_log( "sync: got %d frames, %d expected",
            pv->common->count_frames, sync->count_frames_max );

    /* save data for second pass */
    if( job->pass == 1 )
    {
        /* Preserve frame count for better accuracy in pass 2 */
        hb_interjob_t * interjob = hb_interjob_get( job->h );
        interjob->frame_count = pv->common->count_frames;
        interjob->last_job = job->sequence_id;
    }

    if (sync->drops || sync->dups )
    {
        hb_log( "sync: %d frames dropped, %d duplicated", 
                sync->drops, sync->dups );
    }

    int count = hb_list_count(job->list_subtitle);
    for( ii = 0; ii < count; ii++ )
    {
        CloseSubtitle(sync, ii);
    }
    free(sync->subtitle_sanitizer);

    hb_lock( pv->common->mutex );
    if ( --pv->common->ref == 0 )
    {
        hb_unlock( pv->common->mutex );
        hb_cond_close( &pv->common->next_frame );
        hb_lock_close( &pv->common->mutex );
        free( pv->common->first_pts );
        free( pv->common );
    }
    else
    {
        hb_unlock( pv->common->mutex );
    }

    free( pv );
    w->private_data = NULL;
}

#define ABS(a)  ((a) < 0 ? -(a) : (a))

static hb_buffer_t * mergeSubtitles(subtitle_sanitizer_t *sanitizer, int end)
{
    hb_buffer_t *a, *b, *buf, *out = NULL, *last = NULL;

    do
    {
        a = sanitizer->list_current;

        buf = NULL;
        if (a != NULL && end)
        {
            sanitizer->list_current = a->next;
            if (sanitizer->list_current == NULL)
                sanitizer->last = NULL;
            a->next = NULL;
            buf = a;
        }
        else if (a != NULL && a->s.stop != AV_NOPTS_VALUE)
        {
            b = a->next;

            if (b != NULL && !sanitizer->merge)
            {
                sanitizer->list_current = a->next;
                if (sanitizer->list_current == NULL)
                    sanitizer->last = NULL;
                a->next = NULL;
                buf = a;
            }
            else if (b != NULL && a->s.stop > b->s.start)
            {
                // Overlap
                if (ABS(a->s.start - b->s.start) <= 18000)
                {
                    // subtitles start within 1/5 second of eachother, merge
                    sanitizer->list_current = a->next;
                    if (sanitizer->list_current == NULL)
                        sanitizer->last = NULL;
                    a->next = NULL;
                    b->s.start = a->s.stop;

                    buf = hb_buffer_init(a->size + b->size);
                    buf->s = a->s;
                    sprintf((char*)buf->data, "%s\n%s", a->data, b->data);
                    hb_buffer_close(&a);

                    if (b->s.stop != AV_NOPTS_VALUE && ABS(b->s.stop - b->s.start) <= 18000)
                    {
                        // b and a completely overlap, remove b
                        sanitizer->list_current = b->next;
                        if (sanitizer->list_current == NULL)
                            sanitizer->last = NULL;
                        hb_buffer_close(&b);
                    }
                }
                else
                {
                    // a starts before b, output copy of a and
                    buf = hb_buffer_dup(a);
                    buf->s.stop = b->s.start;
                    a->s.start = b->s.start;
                }
            }
            else if (b != NULL && a->s.stop <= b->s.start)
            {
                sanitizer->list_current = a->next;
                if (sanitizer->list_current == NULL)
                    sanitizer->last = NULL;
                a->next = NULL;
                buf = a;
            }
        }

        if (buf != NULL)
        {
            if (buf->s.stop != AV_NOPTS_VALUE)
                buf->s.duration = buf->s.stop - buf->s.start;
            else
                buf->s.duration = AV_NOPTS_VALUE;
            if (last == NULL)
            {
                out = last = buf;
            }
            else
            {
                last->next = buf;
                last = buf;
            }
        }
    } while (buf != NULL);

    return out;
}

static hb_buffer_t * sanitizeSubtitle(
    hb_work_private_t * pv,
    int                 i,
    hb_buffer_t       * sub)
{
    hb_sync_video_t       * sync;
    subtitle_sanitizer_t  * sanitizer;

    sync = &pv->type.video;
    sanitizer = &sync->subtitle_sanitizer[i];

    if (!sanitizer->link && !sanitizer->merge)
    {
        return sub;
    }

    if (sub == NULL)
    {
        return mergeSubtitles(sanitizer, 1);
    }

    hb_lock( pv->common->mutex );
    sub->s.start -= pv->common->video_pts_slip;
    if (sub->s.stop != AV_NOPTS_VALUE)
        sub->s.stop -= pv->common->video_pts_slip;
    if (sub->s.renderOffset != AV_NOPTS_VALUE)
        sub->s.renderOffset -= pv->common->video_pts_slip;
    hb_unlock( pv->common->mutex );

    if (sanitizer->last != NULL && sanitizer->last->s.stop == AV_NOPTS_VALUE)
    {
        sanitizer->last->s.stop = sub->s.start;
    }

    if (sanitizer->last == NULL)
    {
        sanitizer->list_current = sanitizer->last = sub;
    }
    else
    {
        sanitizer->last->next = sub;
        sanitizer->last = sub;
    }
    return mergeSubtitles(sanitizer, 0);
}

/***********************************************************************
 * syncVideoWork
 ***********************************************************************
 *
 **********************************************************************/
int syncVideoWork( hb_work_object_t * w, hb_buffer_t ** buf_in,
              hb_buffer_t ** buf_out )
{
    hb_buffer_t * cur, * next, * sub = NULL;
    hb_work_private_t * pv = w->private_data;
    hb_job_t          * job = pv->job;
    hb_subtitle_t     * subtitle;
    hb_sync_video_t   * sync = &pv->type.video;
    int i;
    int64_t next_start;

    *buf_out = NULL;
    next = *buf_in;
    *buf_in = NULL;

    /* Wait till we can determine the initial pts of all streams */
    if( next->size != 0 && pv->common->pts_offset == INT64_MIN )
    {
        pv->common->first_pts[0] = next->s.start;
        hb_lock( pv->common->mutex );
        while( pv->common->pts_offset == INT64_MIN && !*w->done )
        {
            // Full fifos will make us wait forever, so get the
            // pts offset from the available streams if full
            if ( hb_fifo_is_full( job->fifo_raw ) )
            {
                getPtsOffset( w );
                hb_cond_broadcast( pv->common->next_frame );
            }
            else if ( checkPtsOffset( w ) )
                hb_cond_broadcast( pv->common->next_frame );
            else
                hb_cond_timedwait( pv->common->next_frame, pv->common->mutex, 200 );
        }
        hb_unlock( pv->common->mutex );
    }

    hb_lock( pv->common->mutex );
    next_start = next->s.start - pv->common->video_pts_slip;
    hb_unlock( pv->common->mutex );

    /* Wait for start of point-to-point encoding */
    if( !pv->common->start_found )
    {
        hb_sync_video_t   * sync = &pv->type.video;

        if( next->size == 0 )
        {
            *buf_out = next;
            pv->common->start_found = 1;
            pv->common->first_pts[0] = INT64_MAX - 1;
            hb_cond_broadcast( pv->common->next_frame );

            /*
             * Push through any subtitle EOFs in case they 
             * were not synced through.
             */
            for( i = 0; i < hb_list_count( job->list_subtitle ); i++)
            {
                subtitle = hb_list_item( job->list_subtitle, i );
                if( subtitle->config.dest == PASSTHRUSUB )
                {
                    hb_fifo_push( subtitle->fifo_out, hb_buffer_init( 0 ) );
                }
            }
            return HB_WORK_DONE;
        }
        if ( pv->common->count_frames < job->frame_to_start ||
             next->s.start < job->pts_to_start )
        {
            // Flush any subtitles that have pts prior to the
            // current frame
            for( i = 0; i < hb_list_count( job->list_subtitle ); i++)
            {
                subtitle = hb_list_item( job->list_subtitle, i );
                while( ( sub = hb_fifo_see( subtitle->fifo_raw ) ) )
                {
                    if ( sub->s.start > next->s.start )
                        break;
                    sub = hb_fifo_get( subtitle->fifo_raw );
                    hb_buffer_close( &sub );
                }
            }
            hb_lock( pv->common->mutex );
            if (job->frame_to_start > 0)
            {
                // When doing frame based p-to-p we must update the audio
                // start point with each frame skipped.
                //
                // Tell the audio threads what must be dropped
                pv->common->audio_pts_thresh = next->s.start;
            }
            hb_cond_broadcast( pv->common->next_frame );
            hb_unlock( pv->common->mutex );

            UpdateSearchState( w, next_start );
#ifdef USE_QSV
            // reclaim QSV resources before dropping the buffer
            // when decoding without QSV, the QSV atom will be NULL
            if (job != NULL && job->qsv.ctx != NULL &&
                next->qsv_details.qsv_atom  != NULL)
            {
                av_qsv_stage *stage = av_qsv_get_last_stage(next->qsv_details.qsv_atom);
                if (stage != NULL)
                {
                    av_qsv_wait_on_sync(job->qsv.ctx, stage);
                    if (stage->out.sync->in_use > 0)
                    {
                        ff_qsv_atomic_dec(&stage->out.sync->in_use);
                    }
                    if (stage->out.p_surface->Data.Locked > 0)
                    {
                        ff_qsv_atomic_dec(&stage->out.p_surface->Data.Locked);
                    }
                }
                av_qsv_flush_stages(job->qsv.ctx->pipes,
                                    &next->qsv_details.qsv_atom);
            }
#endif
            hb_buffer_close( &next );

            return HB_WORK_OK;
        }
        hb_lock( pv->common->mutex );
        pv->common->audio_pts_thresh = 0;
        pv->common->audio_pts_slip += next_start;
        pv->common->video_pts_slip += next_start;
        next_start = 0;
        pv->common->start_found = 1;
        pv->common->count_frames = 0;
        hb_cond_broadcast( pv->common->next_frame );
        hb_unlock( pv->common->mutex );
        sync->st_first = 0;
    }

    if( !sync->cur )
    {
        sync->cur = next;
        if (next->size == 0)
        {
            /* we got an end-of-stream as our first video packet? 
             * Feed it downstream & signal that we're done. 
             */
            *buf_out = next;
            sync->cur = NULL;

            pv->common->start_found = 1;
            pv->common->first_pts[0] = INT64_MAX - 1;
            hb_cond_broadcast( pv->common->next_frame );

            /*
             * Push through any subtitle EOFs in case they 
             * were not synced through.
             */
            for( i = 0; i < hb_list_count( job->list_subtitle ); i++)
            {
                subtitle = hb_list_item( job->list_subtitle, i );
                if( subtitle->config.dest == PASSTHRUSUB )
                {
                    hb_fifo_push( subtitle->fifo_out, hb_buffer_init( 0 ) );
                }
            }
            return HB_WORK_DONE;
        }
        return HB_WORK_OK;
    }
    cur = sync->cur;
    /* At this point we have a frame to process. Let's check
        1) if we will be able to push into the fifo ahead
        2) if the next frame is there already, since we need it to
           compute the duration of the current frame*/
    if( next->size == 0 )
    {
        hb_buffer_close( &next );

        pv->common->first_pts[0] = INT64_MAX - 1;
        cur->s.start = sync->next_start;
        cur->s.stop = cur->s.start + 90000. / ((double)job->vrate / (double)job->vrate_base);
        sync->next_start += cur->s.stop - cur->s.start;;

        /* Make sure last frame is reflected in frame count */
        pv->common->count_frames++;

        /* Push the frame to the renderer */
        *buf_out = cur;
        sync->cur = NULL;

        /* we got an end-of-stream. Feed it downstream & signal that
         * we're done. Note that this means we drop the final frame of
         * video (we don't know its duration). On DVDs the final frame
         * is often strange and dropping it seems to be a good idea. */
        (*buf_out)->next = hb_buffer_init( 0 );

        /*
         * Push through any subtitle EOFs in case they were not synced through.
         */
        for( i = 0; i < hb_list_count( job->list_subtitle ); i++)
        {
            subtitle = hb_list_item( job->list_subtitle, i );
            // flush out any pending subtitle buffers in the sanitizer
            hb_buffer_t *out = sanitizeSubtitle(pv, i, NULL);
            if (out != NULL)
                hb_fifo_push( subtitle->fifo_out, out );
            if( subtitle->config.dest == PASSTHRUSUB )
            {
                hb_fifo_push( subtitle->fifo_out, hb_buffer_init( 0 ) );
            }
        }
        pv->common->start_found = 1;
        hb_cond_broadcast( pv->common->next_frame );
        return HB_WORK_DONE;
    }

    /* Check for end of point-to-point frame encoding */
    if( job->frame_to_stop && pv->common->count_frames > job->frame_to_stop )
    {
        // Drop an empty buffer into our output to ensure that things
        // get flushed all the way out.
        hb_buffer_close( &sync->cur );
        hb_buffer_close( &next );
        *buf_out = hb_buffer_init( 0 );
        hb_log( "sync: reached %d frames, exiting early",
                pv->common->count_frames );

        /*
         * Push through any subtitle EOFs in case they were not synced through.
         */
        for( i = 0; i < hb_list_count( job->list_subtitle ); i++)
        {
            subtitle = hb_list_item( job->list_subtitle, i );
            // flush out any pending subtitle buffers in the sanitizer
            hb_buffer_t *out = sanitizeSubtitle(pv, i, NULL);
            if (out != NULL)
                hb_fifo_push( subtitle->fifo_out, out );
            if( subtitle->config.dest == PASSTHRUSUB )
            {
                hb_fifo_push( subtitle->fifo_out, hb_buffer_init( 0 ) );
            }
        }
        return HB_WORK_DONE;
    }

    /* Check for end of point-to-point pts encoding */
    if( job->pts_to_stop && sync->next_start >= job->pts_to_stop )
    {
        // Drop an empty buffer into our output to ensure that things
        // get flushed all the way out.
        hb_log( "sync: reached pts %"PRId64", exiting early", cur->s.start );
        hb_buffer_close( &sync->cur );
        hb_buffer_close( &next );
        *buf_out = hb_buffer_init( 0 );

        /*
         * Push through any subtitle EOFs in case they were not synced through.
         */
        for( i = 0; i < hb_list_count( job->list_subtitle ); i++)
        {
            subtitle = hb_list_item( job->list_subtitle, i );
            // flush out any pending subtitle buffers in the sanitizer
            hb_buffer_t *out = sanitizeSubtitle(pv, i, NULL);
            if (out != NULL)
                hb_fifo_push( subtitle->fifo_out, out );
            if( subtitle->config.dest == PASSTHRUSUB )
            {
                hb_fifo_push( subtitle->fifo_out, hb_buffer_init( 0 ) );
            }
        }
        return HB_WORK_DONE;
    }

    if( sync->first_frame )
    {
        /* This is our first frame */
        if ( cur->s.start > 0 )
        {
            /*
             * The first pts from a dvd should always be zero but
             * can be non-zero with a transport or program stream since
             * we're not guaranteed to start on an IDR frame. If we get
             * a non-zero initial PTS extend its duration so it behaves
             * as if it started at zero so that our audio timing will
             * be in sync.
             */
            hb_log( "sync: first pts is %"PRId64, cur->s.start );
            cur->s.start = 0;
        }
        sync->first_frame = 0;
    }

    /*
     * since the first frame is always 0 and the upstream reader code
     * is taking care of adjusting for pts discontinuities, we just have
     * to deal with the next frame's start being in the past. This can
     * happen when the PTS is adjusted after data loss but video frame
     * reordering causes some frames with the old clock to appear after
     * the clock change. This creates frames that overlap in time which
     * looks to us like time going backward. The downstream muxing code
     * can deal with overlaps of up to a frame time but anything larger
     * we handle by dropping frames here.
     */
    if ( next_start - cur->s.start <= 0 )
    {
        if ( sync->first_drop == 0 )
        {
            sync->first_drop = next_start;
        }
        ++sync->drop_count;
        if ( next->s.new_chap )
        {
            // don't drop a chapter mark when we drop the buffer
            sync->chap_mark = next->s.new_chap;
        }

#ifdef USE_QSV
        // reclaim QSV resources before dropping the buffer
        // when decoding without QSV, the QSV atom will be NULL
        if (job != NULL && job->qsv.ctx != NULL &&
            next->qsv_details.qsv_atom  != NULL)
        {
            av_qsv_stage *stage = av_qsv_get_last_stage(next->qsv_details.qsv_atom);
            if (stage != NULL)
            {
                av_qsv_wait_on_sync(job->qsv.ctx, stage);
                if (stage->out.sync->in_use > 0)
                {
                    ff_qsv_atomic_dec(&stage->out.sync->in_use);
                }
                if (stage->out.p_surface->Data.Locked > 0)
                {
                    ff_qsv_atomic_dec(&stage->out.p_surface->Data.Locked);
                }
            }
            av_qsv_flush_stages(job->qsv.ctx->pipes,
                                &next->qsv_details.qsv_atom);
        }
#endif

        hb_buffer_close( &next );
        return HB_WORK_OK;
    }
    if ( sync->first_drop )
    {
        hb_log( "sync: video time didn't advance - dropped %d frames "
                "(delta %d ms, current %"PRId64", next %"PRId64", dur %d)",
                sync->drop_count, (int)( cur->s.start - sync->first_drop ) / 90,
                cur->s.start, next_start, (int)( next_start - cur->s.start ) );
        sync->first_drop = 0;
        sync->drop_count = 0;
    }

    /*
     * Track the video sequence number locally so that we can sync the audio
     * to it using the sequence number as well as the PTS.
     */
    sync->video_sequence = cur->sequence;
    
    /* Process subtitles that apply to this video frame */
    // NOTE: There is no logic in either subtitle-sync algorithm that waits
    // for the subtitle-decoder if it is lagging behind the video-decoder.
    //       
    // Therefore there is the implicit assumption that the subtitle-decoder 
    // is always faster than the video-decoder. This assumption is definitely 
    // incorrect in some cases where the SSA subtitle decoder is used.

    for( i = 0; i < hb_list_count( job->list_subtitle ); i++)
    {
        hb_buffer_t *out;

        subtitle = hb_list_item( job->list_subtitle, i );
        
        // Sanitize subtitle start and stop times, then pass to 
        // muxer or renderer filter.
        while ( ( sub = hb_fifo_get( subtitle->fifo_raw ) ) != NULL )
        {
            if (sub->size > 0)
            {
                out = sanitizeSubtitle(pv, i, sub);
                if (out != NULL)
                    hb_fifo_push( subtitle->fifo_out, out );
            }
        }
    }

    /*
     * Adjust the pts of the current frame so that it's contiguous
     * with the previous frame. The start time of the current frame
     * has to be the end time of the previous frame and the stop
     * time has to be the start of the next frame.  We don't
     * make any adjustments to the source timestamps other than removing
     * the clock offsets (which also removes pts discontinuities).
     * This means we automatically encode at the source's frame rate.
     * MP2 uses an implicit duration (frames end when the next frame
     * starts) but more advanced containers like MP4 use an explicit
     * duration. Since we're looking ahead one frame we set the
     * explicit stop time from the start time of the next frame.
     */
    *buf_out = cur;
    int64_t duration = next_start - cur->s.start;
    sync->cur = cur = next;
    cur->sub = NULL;
    cur->s.start -= pv->common->video_pts_slip;
    if (cur->s.renderOffset != AV_NOPTS_VALUE)
        cur->s.renderOffset -= pv->common->video_pts_slip;
    cur->s.stop -= pv->common->video_pts_slip;
    sync->pts_skip = 0;
    if ( duration <= 0 )
    {
        hb_log( "sync: invalid video duration %"PRId64", start %"PRId64", next %"PRId64"",
                duration, cur->s.start, next_start );
    }

    (*buf_out)->s.start = sync->next_start;
    sync->next_start += duration;
    (*buf_out)->s.stop = sync->next_start;

    if ( sync->chap_mark )
    {
        // we have a pending chapter mark from a recent drop - put it on this
        // buffer (this may make it one frame late but we can't do any better).
        (*buf_out)->s.new_chap = sync->chap_mark;
        sync->chap_mark = 0;
    }

    /* Update UI */
    UpdateState( w );

    return HB_WORK_OK;
}

// sync*Init does nothing because sync has a special initializer
// that takes care of initializing video and all audio tracks
int syncVideoInit( hb_work_object_t * w, hb_job_t * job)
{
    return 0;
}

hb_work_object_t hb_sync_video =
{
    WORK_SYNC_VIDEO,
    "Video Synchronization",
    syncVideoInit,
    syncVideoWork,
    syncVideoClose
};

/***********************************************************************
 * Close Audio
 ***********************************************************************
 *
 **********************************************************************/
void syncAudioClose( hb_work_object_t * w )
{
    hb_work_private_t * pv    = w->private_data;
    hb_sync_audio_t   * sync  = &pv->type.audio;

    if( sync->silence_buf )
    {
        free( sync->silence_buf );
    }
    if ( sync->state )
    {
        src_delete( sync->state );
    }

    hb_lock( pv->common->mutex );
    if ( --pv->common->ref == 0 )
    {
        hb_unlock( pv->common->mutex );
        hb_cond_close( &pv->common->next_frame );
        hb_lock_close( &pv->common->mutex );
        free( pv->common->first_pts );
        free( pv->common );
    }
    else
    {
        hb_unlock( pv->common->mutex );
    }

    free( pv );
    w->private_data = NULL;
}

int syncAudioInit( hb_work_object_t * w, hb_job_t * job)
{
    return 0;
}

/***********************************************************************
 * SyncAudio
 ***********************************************************************
 *
 **********************************************************************/
static int syncAudioWork( hb_work_object_t * w, hb_buffer_t ** buf_in,
                       hb_buffer_t ** buf_out )
{
    hb_work_private_t * pv = w->private_data;
    hb_job_t        * job = pv->job;
    hb_sync_audio_t * sync = &pv->type.audio;
    hb_buffer_t     * buf;
    int64_t start;

    *buf_out = NULL;
    buf = *buf_in;
    *buf_in = NULL;
    /* if the next buffer is an eof send it downstream */
    if ( buf->size <= 0 )
    {
        hb_buffer_close( &buf );
        *buf_out = hb_buffer_init( 0 );
        pv->common->first_pts[sync->index+1] = INT64_MAX - 1;
        return HB_WORK_DONE;
    }

    /* Wait till we can determine the initial pts of all streams */
    if( pv->common->pts_offset == INT64_MIN )
    {
        pv->common->first_pts[sync->index+1] = buf->s.start;
        hb_lock( pv->common->mutex );
        while( pv->common->pts_offset == INT64_MIN && !*w->done)
        {
            // Full fifos will make us wait forever, so get the
            // pts offset from the available streams if full
            if (hb_fifo_is_full(w->fifo_in))
            {
                getPtsOffset( w );
                hb_cond_broadcast( pv->common->next_frame );
            }
            else if ( checkPtsOffset( w ) )
                hb_cond_broadcast( pv->common->next_frame );
            else
                hb_cond_timedwait( pv->common->next_frame, pv->common->mutex, 200 );
        }
        hb_unlock( pv->common->mutex );
    }

    // Wait for start frame if doing point-to-point
    //
    // When doing p-to-p, video leads the way. The video thead will set
    // start_found when we have reached the start point.
    //
    // When doing frame based p-to-p, as each video frame is processed
    // it advances audio_pts_thresh which informs us how much audio must
    // be dropped.
    hb_lock( pv->common->mutex );
    while ( !pv->common->start_found && !*w->done )
    {
        if ( pv->common->audio_pts_thresh < 0 )
        {
            // I would initialize this in hb_sync_init, but 
            // job->pts_to_start can be modified by reader 
            // after hb_sync_init is called.
            pv->common->audio_pts_thresh = job->pts_to_start;
        }
        if ( buf->s.start < pv->common->audio_pts_thresh )
        {
            hb_buffer_close( &buf );
            hb_unlock( pv->common->mutex );
            return HB_WORK_OK;
        }

        // We should only get here when doing frame based p-to-p.
        // In frame based p-to-p, the video sync thread updates
        // audio_pts_thresh as it discards frames.  So wait here
        // until the current audio frame needs to be discarded
        // or start point is found.
        while ( !pv->common->start_found && 
                buf->s.start >= pv->common->audio_pts_thresh && !*w->done )
        {
            hb_cond_timedwait( pv->common->next_frame, pv->common->mutex, 10 );
            // There is an unfortunate unavoidable deadlock that can occur.
            // Since we need to wait for a specific frame in syncVideoWork,
            // syncAudioWork can be stalled indefinitely.  The video decoder
            // often drops multiple of the initial frames after starting
            // because they require references that have not been decoded yet.
            // This allows a lot of audio to be queued in the fifo and the
            // audio fifo fills before we get a single video frame.  So we
            // must drop some audio to unplug the pipeline and allow the first
            // video frame to be decoded.
            if ( hb_fifo_is_full(w->fifo_in) )
            {
                hb_buffer_t *tmp;
                tmp = buf = hb_fifo_get( w->fifo_in );
                while ( tmp )
                {
                    tmp = hb_fifo_get( w->fifo_in );
                    if ( tmp )
                    {
                        hb_buffer_close( &buf );
                        buf = tmp;
                    }
                }
            }
        }
    }
    start = buf->s.start - pv->common->audio_pts_slip;
    hb_unlock( pv->common->mutex );

    // When doing p-to-p, video determines when the start point has been
    // found.  Since audio and video are processed asynchronously, there
    // may yet be audio packets in the pipe that are before the start point.
    // These packets will have negative start times after applying
    // audio_pts_slip.
    if (start < 0)
    {
        hb_buffer_close(&buf);
        return HB_WORK_OK;
    }

    if( job->frame_to_stop && pv->common->count_frames >= job->frame_to_stop )
    {
        hb_buffer_close( &buf );
        *buf_out = hb_buffer_init( 0 );
        return HB_WORK_DONE;
    }

    if( job->pts_to_stop && sync->next_start >= job->pts_to_stop )
    {
        hb_buffer_close( &buf );
        *buf_out = hb_buffer_init( 0 );
        return HB_WORK_DONE;
    }

    // audio time went backwards.
    // If our output clock is more than a half frame ahead of the
    // input clock drop this frame to move closer to sync.
    // Otherwise drop frames until the input clock matches the output clock.
    if ( sync->next_start - start > 90*15 )
    {
        // Discard data that's in the past.
        if ( sync->first_drop == 0 )
        {
            sync->first_drop = start;
        }
        ++sync->drop_count;
        hb_buffer_close( &buf );
        return HB_WORK_OK;
    }
    if ( sync->first_drop )
    {
        // we were dropping old data but input buf time is now current
        hb_log( "sync: audio 0x%x time went backwards %d ms, dropped %d frames "
                "(start %"PRId64", next %"PRId64")", w->audio->id,
                (int)( sync->next_start - sync->first_drop ) / 90,
                sync->drop_count, sync->first_drop, (int64_t)sync->next_start );
        sync->first_drop = 0;
        sync->drop_count = 0;
    }
    if ( start - sync->next_start >= (90 * 70) )
    {
        if ( start - sync->next_start > (90000LL * 60) )
        {
            // there's a gap of more than a minute between the last
            // frame and this. assume we got a corrupted timestamp
            // and just drop the next buf.
            hb_log( "sync: %d minute time gap in audio 0x%x - dropping buf"
                    "  start %"PRId64", next %"PRId64,
                    (int)((start - sync->next_start) / (90000*60)),
                    w->audio->id, start, (int64_t)sync->next_start );
            hb_buffer_close( &buf );
            return HB_WORK_OK;
        }
        /*
         * there's a gap of at least 70ms between the last
         * frame we processed & the next. Fill it with silence.
         * Or in the case of DCA, skip some frames from the
         * other streams.
         */
        if ( sync->drop_video_to_sync )
        {
            hb_log( "sync: audio gap %d ms. Skipping frames. Audio 0x%x"
                    "  start %"PRId64", next %"PRId64,
                    (int)((start - sync->next_start) / 90),
                    w->audio->id, start, (int64_t)sync->next_start );
            hb_lock( pv->common->mutex );
            pv->common->audio_pts_slip += (start - sync->next_start);
            pv->common->video_pts_slip += (start - sync->next_start);
            hb_unlock( pv->common->mutex );
            *buf_out = OutputAudioFrame( w->audio, buf, sync );
            return HB_WORK_OK;
        }
        hb_log( "sync: adding %d ms of silence to audio 0x%x"
                "  start %"PRId64", next %"PRId64,
                (int)((start - sync->next_start) / 90),
                w->audio->id, start, (int64_t)sync->next_start );
        InsertSilence( w, start - sync->next_start );
    }

    /*
     * When we get here we've taken care of all the dups and gaps in the
     * audio stream and are ready to inject the next input frame into
     * the output stream.
     */
    *buf_out = OutputAudioFrame( w->audio, buf, sync );
    return HB_WORK_OK;
}

hb_work_object_t hb_sync_audio =
{
    WORK_SYNC_AUDIO,
    "AudioSynchronization",
    syncAudioInit,
    syncAudioWork,
    syncAudioClose
};

static void InitAudio( hb_job_t * job, hb_sync_common_t * common, int i )
{
    hb_work_object_t  * w;
    hb_work_private_t * pv;
    hb_sync_audio_t   * sync;

    pv = calloc( 1, sizeof( hb_work_private_t ) );
    sync = &pv->type.audio;
    sync->index = i;
    pv->job    = job;
    pv->common = common;
    pv->common->ref++;
    pv->common->pts_count++;

    w = hb_get_work( WORK_SYNC_AUDIO );
    w->private_data = pv;
    w->audio = hb_list_item( job->list_audio, i );
    w->fifo_in = w->audio->priv.fifo_raw;

    if ( w->audio->config.out.codec & HB_ACODEC_PASS_FLAG )
    {
        w->fifo_out = w->audio->priv.fifo_out;
    }
    else
    {
        w->fifo_out = w->audio->priv.fifo_sync;
    }

    if( w->audio->config.out.codec == HB_ACODEC_AC3_PASS ||
        w->audio->config.out.codec == HB_ACODEC_AAC_PASS )
    {
        /* Have a silent AC-3/AAC frame ready in case we have to fill a
           gap */
        AVCodec        * codec;
        AVCodecContext * c;

        switch ( w->audio->config.out.codec )
        {
            case HB_ACODEC_AC3_PASS:
            {
                codec = avcodec_find_encoder( AV_CODEC_ID_AC3 );
            } break;
            case HB_ACODEC_AAC_PASS:
            {
                codec = avcodec_find_encoder_by_name("aac");
            } break;
            default:
            {
                // Never gets here
                codec = NULL; // Silence compiler warning
            } break;
        }

        c              = avcodec_alloc_context3(codec);
        c->bit_rate    = w->audio->config.in.bitrate;
        c->sample_rate = w->audio->config.in.samplerate;
        c->channels    =
            av_get_channel_layout_nb_channels(w->audio->config.in.channel_layout);
        hb_ff_set_sample_fmt(c, codec, AV_SAMPLE_FMT_FLT);

        if (w->audio->config.in.channel_layout == AV_CH_LAYOUT_STEREO_DOWNMIX)
        {
            c->channel_layout = AV_CH_LAYOUT_STEREO;
        }
        else
        {
            c->channel_layout = w->audio->config.in.channel_layout;
        }

        if (hb_avcodec_open(c, codec, NULL, 0) < 0)
        {
            hb_log("sync: track %d, hb_avcodec_open() failed, dropping video to sync",
                     w->audio->config.out.track);
            sync->drop_video_to_sync = 1;
        }
        else
        {
            // Prepare input frame
            AVFrame frame = { .nb_samples = c->frame_size, .pts = 0, };
            int input_size = av_samples_get_buffer_size(NULL, c->channels,
                                                        frame.nb_samples,
                                                        c->sample_fmt, 1);
            uint8_t *zeros = calloc(1, input_size);
            avcodec_fill_audio_frame(&frame, c->channels, c->sample_fmt, zeros,
                                     input_size, 1);

            // Allocate enough space for the encoded silence
            // The output should be < the input
            sync->silence_buf  = malloc( input_size );

            // There is some delay in getting output from some audio encoders.
            // So encode a few packets till we get output.
            int ii;
            for ( ii = 0; ii < 10; ii++ )
            {
                // Prepare output packet
                AVPacket pkt;
                int got_packet;
                av_init_packet(&pkt);
                pkt.data = sync->silence_buf;
                pkt.size = input_size;

                int ret = avcodec_encode_audio2( c, &pkt, &frame, &got_packet);
                if ( ret < 0 )
                {
                    hb_log("sync: track %d, avcodec_encode_audio() failed, dropping video to sync",
                           w->audio->config.out.track);
                    sync->drop_video_to_sync = 1;
                    break;
                }

                if ( got_packet )
                {
                    sync->silence_size = pkt.size;
                    break;
                }
                else if (ii + 1 == 10)
                {
                    hb_log("sync: track %d, failed to get output packet, dropping video to sync",
                           w->audio->config.out.track);
                    sync->drop_video_to_sync = 1;
                }
            }
            free( zeros );
            hb_avcodec_close( c );
        }
        av_free( c );
    }
    else
    {
        if( w->audio->config.out.codec & HB_ACODEC_PASS_FLAG )
        {
            sync->drop_video_to_sync = 1;
        }
        else
        {
            /* Not passthru, initialize libsamplerate */
            int error;
            sync->state = src_new( SRC_SINC_MEDIUM_QUALITY,
                                   hb_mixdown_get_discrete_channel_count( w->audio->config.out.mixdown ),
                                   &error );
            sync->data.end_of_input = 0;
        }
    }

    sync->gain_factor = pow(10, w->audio->config.out.gain / 20);

    hb_list_add( job->list_work, w );
}

static hb_buffer_t * OutputAudioFrame( hb_audio_t *audio, hb_buffer_t *buf,
                                       hb_sync_audio_t *sync )
{
    int64_t start = (int64_t)sync->next_start;

    // Can't count of buf->s.stop - buf->s.start for accurate duration
    // due to integer rounding, so use buf->s.duration when it is set
    // (which should be always if I didn't miss anything)
    double duration;
    if ( buf->s.duration > 0 )
        duration = buf->s.duration;
    else
        duration = buf->s.stop - buf->s.start;

    if ( !( audio->config.out.codec & HB_ACODEC_PASS_FLAG ) )
    {
        // Audio is not passthru.  Check if we need to modify the audio
        // in any way.
        if( audio->config.in.samplerate != audio->config.out.samplerate )
        {
            /* do sample rate conversion */
            int count_in, count_out;
            hb_buffer_t * buf_raw = buf;
            int sample_size = hb_mixdown_get_discrete_channel_count( audio->config.out.mixdown ) *
                              sizeof( float );

            count_in  = buf_raw->size / sample_size;
            /*
             * When using stupid rates like 44.1 there will always be some
             * truncation error. E.g., a 1536 sample AC3 frame will turn into a
             * 1536*44.1/48.0 = 1411.2 sample frame. If we just truncate the .2
             * the error will build up over time and eventually the audio will
             * substantially lag the video. libsamplerate will keep track of the
             * fractional sample & give it to us when appropriate if we give it
             * an extra sample of space in the output buffer.
             */
            count_out = ( duration * audio->config.out.samplerate ) / 90000 + 1;

            sync->data.input_frames = count_in;
            sync->data.output_frames = count_out;
            sync->data.src_ratio = (double)audio->config.out.samplerate /
                                   (double)audio->config.in.samplerate;

            buf = hb_buffer_init( count_out * sample_size );
            sync->data.data_in  = (float *) buf_raw->data;
            sync->data.data_out = (float *) buf->data;
            if( src_process( sync->state, &sync->data ) )
            {
                /* XXX If this happens, we're screwed */
                hb_log( "sync: audio 0x%x src_process failed", audio->id );
            }
            hb_buffer_close( &buf_raw );

            if (sync->data.output_frames_gen <= 0)
            {
                // XXX: don't send empty buffers downstream (EOF)
                // possibly out-of-sync audio is better than no audio at all
                hb_buffer_close(&buf);
                return NULL;
            }
            buf->size = sync->data.output_frames_gen * sample_size;
            duration = (double)( sync->data.output_frames_gen * 90000 ) /
                       audio->config.out.samplerate;
        }
        if( audio->config.out.gain > 0.0 )
        {
            int count, ii;

            count  = buf->size / sizeof(float);
            for ( ii = 0; ii < count; ii++ )
            {
                double sample;

                sample = (double)*(((float*)buf->data)+ii);
                sample *= sync->gain_factor;
                if (sample > 0)
                    sample = MIN(sample, 1.0);
                else
                    sample = MAX(sample, -1.0);
                *(((float*)buf->data)+ii) = sample;
            }
        }
        else if( audio->config.out.gain < 0.0 )
        {
            int count, ii;

            count  = buf->size / sizeof(float);
            for ( ii = 0; ii < count; ii++ )
            {
                double sample;

                sample = (double)*(((float*)buf->data)+ii);
                sample *= sync->gain_factor;
                *(((float*)buf->data)+ii) = sample;
            }
        }
    }

    buf->s.type = AUDIO_BUF;
    buf->s.frametype = HB_FRAME_AUDIO;

    buf->s.start = start;
    sync->next_start += duration;
    buf->s.stop  = (int64_t)sync->next_start;
    return buf;
}

static void InsertSilence( hb_work_object_t * w, int64_t duration )
{
    hb_work_private_t * pv = w->private_data;
    hb_sync_audio_t *sync = &pv->type.audio;
    hb_buffer_t     *buf;
    hb_fifo_t       *fifo;
    int frame_dur;

    // to keep pass-thru and regular audio in sync we generate silence in
    // frame-sized units. If the silence duration isn't an integer multiple
    // of the frame duration we will truncate or round up depending on
    // which minimizes the timing error.
    if( w->audio->config.out.codec & HB_ACODEC_PASS_FLAG )
    {
        frame_dur = ( 90000 * w->audio->config.in.samples_per_frame ) /
                                            w->audio->config.in.samplerate;
    }
    else
    {
        frame_dur = ( 90000 * w->audio->config.out.samples_per_frame ) /
                                            w->audio->config.in.samplerate;
    }

    while (duration >= frame_dur >> 2)
    {
        if( w->audio->config.out.codec & HB_ACODEC_PASS_FLAG )
        {
            buf        = hb_buffer_init( sync->silence_size );
            buf->s.start = sync->next_start;
            buf->s.stop  = buf->s.start + frame_dur;
            memcpy( buf->data, sync->silence_buf, buf->size );
            fifo = w->audio->priv.fifo_out;
            duration -= frame_dur;
        }
        else
        {
            int channel_count = hb_mixdown_get_discrete_channel_count( w->audio->config.out.mixdown );
            int size = sizeof( float ) *
                       w->audio->config.out.samples_per_frame *
                       channel_count;
            if (frame_dur > duration)
            {
                int samples = duration * w->audio->config.in.samplerate / 90000;
                if (samples == 0)
                {
                    break;
                }
                size = sizeof(float) * samples * channel_count;
                frame_dur = (90000 * samples) / w->audio->config.in.samplerate;
            }
            buf = hb_buffer_init(size);
            buf->s.start = sync->next_start;
            buf->s.duration = frame_dur;
            buf->s.stop  = buf->s.start + frame_dur;
            memset( buf->data, 0, buf->size );
            fifo = w->audio->priv.fifo_sync;
            duration -= frame_dur;
        }
        buf = OutputAudioFrame( w->audio, buf, sync );
        hb_fifo_push( fifo, buf );
    }
}

static void UpdateState( hb_work_object_t * w )
{
    hb_work_private_t * pv = w->private_data;
    hb_sync_video_t   * sync = &pv->type.video;
    hb_state_t state;

    if( !pv->common->count_frames )
    {
        sync->st_first = hb_get_date();
        pv->job->st_pause_date = -1;
        pv->job->st_paused = 0;
    }
    pv->common->count_frames++;

    if (pv->job->indepth_scan)
    {
        // Progress for indept scan is handled by reader
        // pv->common->count_frames is used during indepth_scan
        // to find start & end points.
        return;
    }

    if( hb_get_date() > sync->st_dates[3] + 1000 )
    {
        memmove( &sync->st_dates[0], &sync->st_dates[1],
                 3 * sizeof( uint64_t ) );
        memmove( &sync->st_counts[0], &sync->st_counts[1],
                 3 * sizeof( uint64_t ) );
        sync->st_dates[3]  = hb_get_date();
        sync->st_counts[3] = pv->common->count_frames;
    }

#define p state.param.working
    state.state = HB_STATE_WORKING;
    p.progress  = (float) pv->common->count_frames / (float) sync->count_frames_max;
    if( p.progress > 1.0 )
    {
        p.progress = 1.0;
    }
    p.rate_cur   = 1000.0 *
        (float) ( sync->st_counts[3] - sync->st_counts[0] ) /
        (float) ( sync->st_dates[3] - sync->st_dates[0] );
    if( hb_get_date() > sync->st_first + 4000 )
    {
        int eta;
        p.rate_avg = 1000.0 * (float) sync->st_counts[3] /
            (float) ( sync->st_dates[3] - sync->st_first - pv->job->st_paused);
        eta = (float) ( sync->count_frames_max - sync->st_counts[3] ) /
            p.rate_avg;
        p.hours   = eta / 3600;
        p.minutes = ( eta % 3600 ) / 60;
        p.seconds = eta % 60;
    }
    else
    {
        p.rate_avg = 0.0;
        p.hours    = -1;
        p.minutes  = -1;
        p.seconds  = -1;
    }
#undef p

    hb_set_state( pv->job->h, &state );
}

static void UpdateSearchState( hb_work_object_t * w, int64_t start )
{
    hb_work_private_t * pv = w->private_data;
    hb_sync_video_t   * sync = &pv->type.video;
    hb_state_t state;
    uint64_t now;
    double avg;

    now = hb_get_date();
    if( !pv->common->count_frames )
    {
        sync->st_first = now;
        pv->job->st_pause_date = -1;
        pv->job->st_paused = 0;
    }
    pv->common->count_frames++;

    if (pv->job->indepth_scan)
    {
        // Progress for indept scan is handled by reader
        // pv->common->count_frames is used during indepth_scan
        // to find start & end points.
        return;
    }

#define p state.param.working
    state.state = HB_STATE_SEARCHING;
    if ( pv->job->frame_to_start )
        p.progress  = (float) pv->common->count_frames / 
                      (float) pv->job->frame_to_start;
    else if ( pv->job->pts_to_start )
        p.progress  = (float) start / (float) pv->job->pts_to_start;
    else
        p.progress = 0;
    if( p.progress > 1.0 )
    {
        p.progress = 1.0;
    }
    if (now > sync->st_first)
    {
        int eta = 0;

        if ( pv->job->frame_to_start )
        {
            avg = 1000.0 * (double)pv->common->count_frames / (now - sync->st_first);
            eta = ( pv->job->frame_to_start - pv->common->count_frames ) / avg;
        }
        else if ( pv->job->pts_to_start )
        {
            avg = 1000.0 * (double)start / (now - sync->st_first);
            eta = ( pv->job->pts_to_start - start ) / avg;
        }
        p.hours   = eta / 3600;
        p.minutes = ( eta % 3600 ) / 60;
        p.seconds = eta % 60;
    }
    else
    {
        p.rate_avg = 0.0;
        p.hours    = -1;
        p.minutes  = -1;
        p.seconds  = -1;
    }
#undef p

    hb_set_state( pv->job->h, &state );
}

static void getPtsOffset( hb_work_object_t * w )
{
    hb_work_private_t * pv = w->private_data;
    int           i ;
    int64_t       first_pts = INT64_MAX;

    for( i = 0; i < pv->common->pts_count; i++ )
    {
        if ( pv->common->first_pts[i] < first_pts )
            first_pts = pv->common->first_pts[i];
    }
    pv->common->video_pts_slip = pv->common->audio_pts_slip = pv->common->pts_offset = first_pts;
    return;
}

static int checkPtsOffset( hb_work_object_t * w )
{
    hb_work_private_t * pv = w->private_data;
    int           i ;

    for( i = 0; i < pv->common->pts_count; i++ )
    {
        if ( pv->common->first_pts[i] == INT64_MAX )
            return 0;
    }
    getPtsOffset( w );
    return 1;
}