diff options
author | John Stebbins <[email protected]> | 2016-01-21 09:11:09 -0800 |
---|---|---|
committer | John Stebbins <[email protected]> | 2016-01-21 09:11:09 -0800 |
commit | 254b1fac8bb35bb29688560dbc8a3811be652f36 (patch) | |
tree | 8232f55f36175db115ee08ebd5f5e41c29c53949 | |
parent | 56925edbfa59cda34d3fa45f4c61f48970cee5fb (diff) | |
parent | fcb78d5c24387bf2dad2e3c38b4417431b2836ae (diff) |
Merge pull request #34 from jstebbins/sync
sync: gut and rewrite
-rw-r--r-- | libhb/common.h | 2 | ||||
-rw-r--r-- | libhb/decavcodec.c | 1 | ||||
-rw-r--r-- | libhb/enc_qsv.c | 3 | ||||
-rw-r--r-- | libhb/fifo.c | 26 | ||||
-rw-r--r-- | libhb/hb.c | 1 | ||||
-rw-r--r-- | libhb/internal.h | 11 | ||||
-rw-r--r-- | libhb/sync.c | 2816 | ||||
-rw-r--r-- | libhb/vfr.c | 2 | ||||
-rw-r--r-- | libhb/work.c | 7 |
9 files changed, 1657 insertions, 1212 deletions
diff --git a/libhb/common.h b/libhb/common.h index 2dc1c9002..15f847095 100644 --- a/libhb/common.h +++ b/libhb/common.h @@ -1157,7 +1157,6 @@ struct hb_work_object_s hb_work_private_t * private_data; hb_thread_t * thread; - int yield; volatile int * done; volatile int * die; int status; @@ -1172,6 +1171,7 @@ struct hb_work_object_s extern hb_work_object_t hb_sync_video; extern hb_work_object_t hb_sync_audio; +extern hb_work_object_t hb_sync_subtitle; extern hb_work_object_t hb_decvobsub; extern hb_work_object_t hb_encvobsub; extern hb_work_object_t hb_deccc608; diff --git a/libhb/decavcodec.c b/libhb/decavcodec.c index b21b19de4..c7ed9494b 100644 --- a/libhb/decavcodec.c +++ b/libhb/decavcodec.c @@ -918,6 +918,7 @@ static hb_buffer_t *copy_frame( hb_work_private_t *pv ) pv->qsv.config.io_pattern == MFX_IOPATTERN_OUT_OPAQUE_MEMORY) { buf->qsv_details.qsv_atom = pv->frame->data[2]; + buf->qsv_details.ctx = pv->job->qsv.ctx; return buf; } #endif diff --git a/libhb/enc_qsv.c b/libhb/enc_qsv.c index 360db65b5..87a180629 100644 --- a/libhb/enc_qsv.c +++ b/libhb/enc_qsv.c @@ -1966,6 +1966,9 @@ int encqsvWork(hb_work_object_t *w, hb_buffer_t **buf_in, hb_buffer_t **buf_out) { qsv_atom = in->qsv_details.qsv_atom; surface = av_qsv_get_last_stage(qsv_atom)->out.p_surface; + // At this point, enc_qsv takes ownership of the QSV resources + // in the 'in' buffer. + in->qsv_details.qsv_atom = NULL; /* * QSV decoding fills the QSV context's dts_seq list, we need to diff --git a/libhb/fifo.c b/libhb/fifo.c index 634d0a336..62a34b8b3 100644 --- a/libhb/fifo.c +++ b/libhb/fifo.c @@ -9,6 +9,9 @@ #include "hb.h" #include "openclwrapper.h" +#ifdef USE_QSV +#include "libavcodec/qsv.h" +#endif #ifndef SYS_DARWIN #include <malloc.h> @@ -676,6 +679,29 @@ void hb_buffer_close( hb_buffer_t ** _b ) while( b ) { +#ifdef USE_QSV + // Reclaim QSV resources before dropping the buffer. + // when decoding without QSV, the QSV atom will be NULL. + if (b->qsv_details.qsv_atom != NULL && b->qsv_details.ctx != NULL) + { + av_qsv_stage *stage = av_qsv_get_last_stage(b->qsv_details.qsv_atom); + if (stage != NULL) + { + av_qsv_wait_on_sync(b->qsv_details.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(b->qsv_details.ctx->pipes, + &b->qsv_details.qsv_atom); + } +#endif + hb_buffer_t * next = b->next; hb_fifo_t *buffer_pool = size_to_pool( b->alloc ); diff --git a/libhb/hb.c b/libhb/hb.c index 42ac83a9d..41347f613 100644 --- a/libhb/hb.c +++ b/libhb/hb.c @@ -1746,6 +1746,7 @@ int hb_global_init() hb_register(&hb_reader); hb_register(&hb_sync_video); hb_register(&hb_sync_audio); + hb_register(&hb_sync_subtitle); hb_register(&hb_decavcodecv); hb_register(&hb_decavcodeca); hb_register(&hb_declpcm); diff --git a/libhb/internal.h b/libhb/internal.h index 59510961a..8fe461dac 100644 --- a/libhb/internal.h +++ b/libhb/internal.h @@ -9,6 +9,9 @@ #include "hbffmpeg.h" #include "extras/cl.h" +#ifdef USE_QSV +#include "libavcodec/qsv.h" +#endif /*********************************************************************** * Hardware Decode Context @@ -145,11 +148,14 @@ struct hb_buffer_s int size; } plane[4]; // 3 Color components + alpha +#ifdef USE_QSV struct qsv { - void *qsv_atom; - void *filter_details; + void * qsv_atom; + void * filter_details; + av_qsv_context * ctx; } qsv_details; +#endif /* OpenCL */ struct cl_data @@ -426,6 +432,7 @@ enum WORK_NONE = 0, WORK_SYNC_VIDEO, WORK_SYNC_AUDIO, + WORK_SYNC_SUBTITLE, WORK_DECCC608, WORK_DECVOBSUB, WORK_DECSRTSUB, diff --git a/libhb/sync.c b/libhb/sync.c index 5520603bb..9272d30d3 100644 --- a/libhb/sync.c +++ b/libhb/sync.c @@ -12,50 +12,29 @@ #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) +#define SYNC_MAX_VIDEO_QUEUE_LEN 20 +#define SYNC_MIN_VIDEO_QUEUE_LEN 10 -typedef struct -{ - /* Audio/Video sync thread synchronization */ - hb_lock_t * mutex; - hb_cond_t * next_frame; - int64_t volatile * last_pts; - int pts_count; +// Audio is small, buffer a lot. It helps to ensure that we see +// the initial PTS from all input streams before setting the 'zero' point. +#define SYNC_MAX_AUDIO_QUEUE_LEN 60 +#define SYNC_MIN_AUDIO_QUEUE_LEN 30 - /* PTS synchronization */ - int64_t audio_pts_slip; - int64_t video_pts_slip; - int64_t pts_offset; +#define SYNC_MAX_SUBTITLE_QUEUE_LEN INT_MAX +#define SYNC_MIN_SUBTITLE_QUEUE_LEN 0 - /* point-to-point support */ - int start_found; - int count_frames; - - /* sync audio work objects */ - hb_list_t * list_work; -} hb_sync_common_t; +typedef enum +{ + SYNC_TYPE_VIDEO, + SYNC_TYPE_AUDIO, + SYNC_TYPE_SUBTITLE, +} sync_type_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; + int64_t pts; + int64_t delta; +} sync_delta_t; typedef struct { @@ -64,1320 +43,1770 @@ typedef struct hb_buffer_list_t list_current; } subtitle_sanitizer_t; +typedef struct sync_common_s sync_common_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 */ + sync_common_t * common; + + // Stream I/O control + hb_list_t * in_queue; + int max_len; + int min_len; + hb_cond_t * cond_full; + hb_buffer_list_t out_queue; + int eof; + + // PTS synchronization + hb_list_t * delta_list; + int64_t pts_slip; + double next_pts; + + // frame statistics + int64_t first_pts; + int64_t min_frame_duration; + int64_t max_frame_duration; + int64_t current_duration; + int frame_count; + + // Error reporting stats + int64_t drop_duration; + int drop; + int64_t drop_pts; + int64_t gap_duration; + int64_t gap_pts; + + int first_frame; + + // stream type specific context + sync_type_t type; + union + { + struct + { + double dejitter_pts; + } video; + + // Audio stream context + struct + { + hb_audio_t * audio; + + // Audio filter settings + // Samplerate conversion + struct + { + SRC_STATE * ctx; + SRC_DATA pkt; + } src; + double gain_factor; + } audio; + + // Subtitle stream context + struct + { + hb_subtitle_t * subtitle; + subtitle_sanitizer_t sanitizer; + } subtitle; + }; +} sync_stream_t; + +struct sync_common_s +{ + /* Audio/Video sync thread synchronization */ + hb_job_t * job; + hb_lock_t * mutex; + int stream_count; + sync_stream_t * streams; + int found_first_pts; + int done; + + // point-to-point support + int start_found; + + // sync audio work objects + hb_list_t * list_work; + + // UpdateState Statistics + int est_frame_count; 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; + sync_common_t * common; + sync_stream_t * stream; }; /*********************************************************************** * Local prototypes **********************************************************************/ -static void setSyncPTS(hb_work_private_t * pv, int64_t pts, int index); -static void getPtsOffset( hb_work_private_t * pv ); -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 ); +static void UpdateState( sync_common_t * common, int frame_count ); +static void UpdateSearchState( sync_common_t * common, int64_t start, + int frame_count ); +static hb_buffer_t * FilterAudioFrame( sync_stream_t * stream, + hb_buffer_t *buf ); -/*********************************************************************** - * hb_work_sync_init - *********************************************************************** - * Initialize the work object - **********************************************************************/ -int syncVideoInit( hb_work_object_t * w, hb_job_t * job) +static int fillQueues( sync_common_t * common ) { - 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; + int ii, wait = 0, abort = 0; - pv = calloc( 1, sizeof( hb_work_private_t ) ); - sync = &pv->type.video; - pv->common = calloc( 1, sizeof( hb_sync_common_t ) ); - pv->common->list_work = hb_list_init(); - pv->common->mutex = hb_lock_init(); - 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 + for (ii = 0; ii < common->stream_count; ii++) { - pv->common->start_found = 1; - } + sync_stream_t *stream = &common->streams[ii]; - w->private_data = pv; - w->fifo_in = job->fifo_raw; - // Register condition with fifo to wake us up immediately if - // the fifo becomes full - hb_fifo_register_full_cond(w->fifo_in, pv->common->next_frame); - - // 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_id == HB_PASS_ENCODE_2ND ) - { - /* 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 ) + // Don't let the queues grow indefinitely + // abort when too large + if (hb_list_count(stream->in_queue) > stream->max_len) { - duration = job->pts_to_stop + 90000; + abort = 1; } - else if( job->frame_to_stop ) - { - /* Set the duration to a rough estimate */ - duration = (int64_t)(job->frame_to_stop + 1) * - title->vrate.den * 90000 / title->vrate.num; - } - 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->vrate.num / title->vrate.den / 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++ ) + if (hb_list_count(stream->in_queue) < stream->min_len) { - InitAudio( job, pv->common, i ); + wait = 1; } } - pv->common->last_pts = malloc( sizeof(int64_t) * pv->common->pts_count ); - for ( i = 0; i < pv->common->pts_count; i++ ) - { - pv->common->last_pts[i] = AV_NOPTS_VALUE; - } - - 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); - } - - /* Launch audio processing threads */ - for (i = 0; i < hb_list_count(pv->common->list_work); i++) - { - hb_work_object_t * audio_work; - audio_work = hb_list_item(pv->common->list_work, i); - audio_work->done = w->done; - audio_work->thread = hb_thread_init(audio_work->name, hb_work_loop, - audio_work, HB_LOW_PRIORITY); - } - return 0; + return !wait || abort; } -static void InitSubtitle( hb_job_t * job, hb_sync_video_t * sync, int ii ) +static void signalBuffer( sync_stream_t * stream ) { - hb_subtitle_t * subtitle; - - subtitle = hb_list_item( job->list_subtitle, ii ); - 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[ii].merge = 1; - } - // PGS subtitles don't need to be linked because there are explicit - // "clear" subtitle packets that indicate the end time of the - // previous subtitle - if (subtitle->config.dest == PASSTHRUSUB && - subtitle->source != PGSSUB) + if (hb_list_count(stream->in_queue) < stream->max_len) { - // Fill in stop time when it is missing - sync->subtitle_sanitizer[ii].link = 1; + hb_cond_signal(stream->cond_full); } - hb_buffer_list_clear(&sync->subtitle_sanitizer[ii].list_current); -} - -static void CloseSubtitle(hb_sync_video_t * sync, int ii) -{ - hb_buffer_list_close(&sync->subtitle_sanitizer[ii].list_current); } -/*********************************************************************** - * Close Video - *********************************************************************** - * - **********************************************************************/ -void syncVideoClose( hb_work_object_t * w ) +static void allSlip( sync_common_t * common, int64_t delta ) { - 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; - // Unblock anybody waiting on this threads last PTS - setSyncPTS(pv, INT64_MAX, 0); - - if( sync->cur ) + for (ii = 0; ii < common->stream_count; ii++) { - hb_buffer_close( &sync->cur ); + common->streams[ii].pts_slip += delta; + if (common->streams[ii].next_pts != AV_NOPTS_VALUE) + { + common->streams[ii].next_pts -= delta; + } } +} - hb_log( "sync: got %d frames, %d expected", - pv->common->count_frames, sync->count_frames_max ); +static void shiftTS( sync_common_t * common, int64_t delta ) +{ + int ii, jj; - /* save data for second pass */ - if( job->pass_id == HB_PASS_ENCODE_1ST ) + allSlip(common, delta); + for (ii = 0; ii < common->stream_count; ii++) { - /* 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; + sync_stream_t * stream = &common->streams[ii]; + int count = hb_list_count(stream->in_queue); + for (jj = 0; jj < count; jj++) + { + hb_buffer_t * buf = hb_list_item(stream->in_queue, jj); + buf->s.start -= delta; + if (buf->s.stop != AV_NOPTS_VALUE) + { + buf->s.stop -= delta; + } + } } +} - if (sync->drops || sync->dups ) - { - hb_log( "sync: %d frames dropped, %d duplicated", - sync->drops, sync->dups ); - } +static void checkFirstPts( sync_common_t * common ) +{ + int ii; + int64_t first_pts = INT64_MAX; - int count = hb_list_count(job->list_subtitle); - for( ii = 0; ii < count; ii++ ) + for (ii = 0; ii < common->stream_count; ii++) { - CloseSubtitle(sync, ii); - } - free(sync->subtitle_sanitizer); + sync_stream_t *stream = &common->streams[ii]; + if (stream->type == SYNC_TYPE_SUBTITLE) + { + // only wait for audio and video + continue; + } - // Close audio work threads - hb_work_object_t * audio_work; - while ((audio_work = hb_list_item(pv->common->list_work, 0))) - { - hb_list_rem(pv->common->list_work, audio_work); - if (audio_work->thread != NULL) + // If buffers are queued, find the lowest initial PTS + if (hb_list_count(stream->in_queue) > 0) { - hb_thread_close(&audio_work->thread); + hb_buffer_t * buf = hb_list_item(stream->in_queue, 0); + if (first_pts > buf->s.start) + { + first_pts = buf->s.start; + } } - audio_work->close(audio_work); - free(audio_work); } - hb_list_close(&pv->common->list_work); - - hb_cond_close( &pv->common->next_frame ); - hb_lock_close( &pv->common->mutex ); - free((void*)pv->common->last_pts); - free(pv->common); - free( pv ); - w->private_data = NULL; + if (first_pts != INT64_MAX) + { + // Add a fudge factor to first pts to prevent negative + // timestamps from leaking through. The pipeline can + // handle a positive offset, but some things choke on + // negative offsets + //first_pts -= 500000; + shiftTS(common, first_pts); + } + common->found_first_pts = 1; } -static hb_buffer_t * merge_ssa(hb_buffer_t *a, hb_buffer_t *b) +static void addDelta( sync_common_t * common, int64_t start, int64_t delta) { - int len, ii; - char *text; - hb_buffer_t *buf = hb_buffer_init(a->size + b->size); - buf->s = a->s; + int ii; - // Find the text in the second SSA sub - text = (char*)b->data; - for (ii = 0; ii < 8; ii++) - { - text = strchr(text, ','); - if (text == NULL) - break; - text++; - } - if (text != NULL) + for (ii = 0; ii < common->stream_count; ii++) { - len = sprintf((char*)buf->data, "%s\n%s", a->data, text); - if (len >= 0) - buf->size = len + 1; + sync_delta_t * delta_item = malloc(sizeof(sync_delta_t)); + delta_item->pts = start; + delta_item->delta = delta; + hb_list_add(common->streams[ii].delta_list, delta_item); } - else - { - memcpy(buf->data, a->data, a->size); - buf->size = a->size; - } - - return buf; } -static hb_buffer_t * mergeSubtitles(subtitle_sanitizer_t *sanitizer, int end) +static void applyDeltas( sync_common_t * common ) { - hb_buffer_t *a, *b, *buf; - hb_buffer_list_t list; + int ii; - hb_buffer_list_clear(&list); - - do + // Apply delta to any applicable buffers in the queue + for (ii = 0; ii < common->stream_count; ii++) { - a = hb_buffer_list_head(&sanitizer->list_current); - if (a == NULL) - { - break; - } - b = a->next; + sync_stream_t * stream = &common->streams[ii]; - buf = NULL; - if (b == NULL && end) - { - buf = hb_buffer_list_rem_head(&sanitizer->list_current); - } - else if (a->s.stop != AV_NOPTS_VALUE) + // Make adjustments for deltas found in other streams + sync_delta_t * delta = hb_list_item(stream->delta_list, 0); + if (delta != NULL) { - if (!sanitizer->merge) - { - buf = hb_buffer_list_rem_head(&sanitizer->list_current); - } - else if (b != NULL && a->s.stop > b->s.start) + int jj, index = -1; + int64_t prev_start, max = 0; + hb_buffer_t * buf; + + prev_start = stream->next_pts; + for (jj = 0; jj < hb_list_count(stream->in_queue); jj++) { - // Overlap - if (ABS(a->s.start - b->s.start) <= 18000) + buf = hb_list_item(stream->in_queue, jj); + if (stream->type == SYNC_TYPE_SUBTITLE) { - if (b->s.stop == AV_NOPTS_VALUE && !end) + if (buf->s.start > delta->pts) { - // To evaluate overlaps, we need the stop times - // for a and b + // absorb gaps in subtitles as soon as possible + index = jj; break; } - a = hb_buffer_list_rem_head(&sanitizer->list_current); - - // subtitles start within 1/5 second of eachother, merge - if (a->s.stop > b->s.stop && b->s.stop != AV_NOPTS_VALUE) + } + else if (buf->s.start > delta->pts) + { + // absorb gap in largest gap found in this stream. + if (buf->s.start - prev_start > max) { - // a continues after b, reorder the list and swap - b = a; - a = hb_buffer_list_rem_head(&sanitizer->list_current); - hb_buffer_list_prepend(&sanitizer->list_current, b); + max = buf->s.start - prev_start; + index = jj; } + if (stream->type == SYNC_TYPE_AUDIO && max >= delta->delta) + { + // absorb gaps in audio as soon as possible + // when there is a gap that will absorb it. + break; + } + } + prev_start = buf->s.start; + } - b->s.start = a->s.stop; - - buf = merge_ssa(a, b); - hb_buffer_close(&a); - a = buf; - - if (b->s.stop != AV_NOPTS_VALUE && - ABS(b->s.stop - b->s.start) <= 18000) + if (index >= 0) + { + for (jj = index; jj < hb_list_count(stream->in_queue); jj++) + { + buf = hb_list_item(stream->in_queue, jj); + buf->s.start -= delta->delta; + if (buf->s.stop != AV_NOPTS_VALUE) { - // b and a completely overlap, remove b - b = hb_buffer_list_rem_head(&sanitizer->list_current); - hb_buffer_close(&b); + buf->s.stop -= delta->delta; } } - else + // Correct the duration of the video buffer before + // the affected timestamp correction. + if (stream->type == SYNC_TYPE_VIDEO && index > 0) { - // a starts before b, output copy of a and update a start - buf = hb_buffer_dup(a); - buf->s.stop = b->s.start; - a->s.start = b->s.start; + buf = hb_list_item(stream->in_queue, index - 1); + if (buf->s.duration > delta->delta) + { + buf->s.duration -= delta->delta; + buf->s.stop -= delta->delta; + } + else + { + buf->s.duration = 0; + buf->s.stop = buf->s.start; + } } - } - else if (b != NULL && a->s.stop <= b->s.start) - { - // a starts and ends before b - buf = hb_buffer_list_rem_head(&sanitizer->list_current); + stream->pts_slip += delta->delta; + hb_list_rem(stream->delta_list, delta); + free(delta); } } + } +} - 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; - hb_buffer_list_append(&list, buf); - } - } while (hb_buffer_list_count(&sanitizer->list_current) >= 2 || end); +static void removeVideoJitter( sync_stream_t * stream, int stop ) +{ + int ii; + hb_buffer_t * buf; + double frame_duration, next_pts; - return hb_buffer_list_clear(&list); + frame_duration = 90000. * stream->common->job->title->vrate.den / + stream->common->job->title->vrate.num; + + buf = hb_list_item(stream->in_queue, 0); + buf->s.start = stream->next_pts; + next_pts = stream->next_pts + frame_duration; + for (ii = 1; ii <= stop; ii++) + { + buf->s.duration = frame_duration; + buf->s.stop = next_pts; + buf = hb_list_item(stream->in_queue, ii); + buf->s.start = next_pts; + next_pts += frame_duration; + } } -static hb_buffer_t * sanitizeSubtitle( - hb_work_private_t * pv, - int i, - hb_buffer_t * sub) +// Look for a sequence of packets whose duration as measure by +// vrate closely matches the duration as measured by timestamp +// differences. When close matches are found, smooth the timestamps. +// +// Most often, video dejitter is applied when there is jitter due to +// soft telecine. I also have a couple sample files that have very +// bad video jitter that this corrects. +static void dejitterVideo( sync_stream_t * stream ) { - hb_sync_video_t * sync; - subtitle_sanitizer_t * sanitizer; + int ii, count, jitter_stop; + double frame_duration, duration; + hb_buffer_t * buf; - sync = &pv->type.video; - sanitizer = &sync->subtitle_sanitizer[i]; + count = hb_list_count(stream->in_queue); + if (count < 2) + { + return; + } - if (!sanitizer->link && !sanitizer->merge) + frame_duration = 90000. * stream->common->job->title->vrate.den / + stream->common->job->title->vrate.num; + + // Look for start of jittered sequence + buf = hb_list_item(stream->in_queue, 1); + duration = buf->s.start - stream->next_pts; + if (ABS(duration - frame_duration) < 1.1) { - if (sub != NULL) + // Ignore small jitter + return; + } + + // Look for end of jittered sequence + jitter_stop = 0; + for (ii = 1; ii < count; ii++) + { + buf = hb_list_item(stream->in_queue, ii); + duration = buf->s.start - stream->next_pts; + + // Only dejitter video that aligns periodically + // with the frame durations. + if (ABS(duration - ii * frame_duration) < 0.1) { - if (sub->s.stop != AV_NOPTS_VALUE) - sub->s.duration = sub->s.stop - sub->s.start; - else - sub->s.duration = 0; - 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; + jitter_stop = ii; } - return sub; } - if (sub == NULL) + if (jitter_stop > 0) { - return mergeSubtitles(sanitizer, 1); + removeVideoJitter(stream, jitter_stop); } +} - 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; +// Fix video overlaps that could not be corrected by dejitter +static void fixVideoOverlap( sync_stream_t * stream ) +{ + int drop = 0, new_chap = 0; + int64_t overlap; + hb_buffer_t * buf; - hb_buffer_t *last = hb_buffer_list_tail(&sanitizer->list_current); - if (last != NULL && last->s.stop == AV_NOPTS_VALUE) + if (!stream->first_frame) { - last->s.stop = sub->s.start; + // There are no overlaps if we haven't seen the first frame yet. + return; } - if (sub->s.start == sub->s.stop) + // If time goes backwards drop the frame. + // Check if subsequent buffers also overlap. + while ((buf = hb_list_item(stream->in_queue, 0)) != NULL) { - // Used to indicate "clear" subtitles when the duration - // of subtitles is not encoded in the stream - hb_buffer_close(&sub); + // For video, an overlap is where the entire frame is + // in the past. + overlap = stream->next_pts - buf->s.stop; + if (overlap >= 0) + { + if (stream->drop == 0) + { + stream->drop_pts = buf->s.start; + } + // Preserve chapter marks + if (buf->s.new_chap > 0) + { + new_chap = buf->s.new_chap; + } + hb_list_rem(stream->in_queue, buf); + signalBuffer(stream); + stream->drop_duration += buf->s.duration; + stream->drop++; + drop++; + hb_buffer_close(&buf); + } + else + { + if (new_chap > 0) + { + buf->s.new_chap = new_chap; + } + break; + } } - hb_buffer_list_append(&sanitizer->list_current, sub); - return mergeSubtitles(sanitizer, 0); + if (drop <= 0 && stream->drop > 0) + { + hb_log("sync: video time went backwards %d ms, dropped %d frames. " + "PTS %"PRId64"", + (int)stream->drop_duration / 90, stream->drop, stream->drop_pts); + stream->drop_duration = 0; + stream->drop = 0; + } } -static void setSyncPTS(hb_work_private_t * pv, int64_t pts, int index) +static void removeAudioJitter(sync_stream_t * stream, int stop) { - hb_lock(pv->common->mutex); - pv->common->last_pts[index] = pts; - hb_unlock(pv->common->mutex); - hb_cond_broadcast(pv->common->next_frame); + int ii; + hb_buffer_t * buf; + double next_pts; + + // If duration of sum of packet durations is close to duration + // as measured by timestamps, align timestamps to packet durations. + // The packet durations are computed based on samplerate and + // number of samples and are therefore a reliable measure + // of the actual duration of an audio frame. + buf = hb_list_item(stream->in_queue, 0); + buf->s.start = stream->next_pts; + next_pts = stream->next_pts + buf->s.duration; + for (ii = 1; ii <= stop; ii++) + { + // Duration can be fractional, so track fractional PTS + buf->s.stop = next_pts; + buf = hb_list_item(stream->in_queue, ii); + buf->s.start = next_pts; + next_pts += buf->s.duration; + } } -static void resetSync(hb_work_private_t * pv) +// Look for a sequence of packets whose duration as measure by packet +// durations closely matches the duration as measured by timestamp +// differences. When close matches are found, smooth the timestamps. +// +// This fixes issues where there are false overlaps and gaps in audio +// timestamps. libav creates this type of problem sometimes with it's +// timestamp guessing code. +static void dejitterAudio( sync_stream_t * stream ) { - int ii; + int ii, count, jitter_stop; + double duration; + hb_buffer_t * buf, * buf0, * buf1; - hb_lock(pv->common->mutex); - for (ii = 0; ii < pv->common->pts_count; ii++) + count = hb_list_count(stream->in_queue); + if (count < 4) { - // Unblock any sync thread that are waiting for a PTS - pv->common->last_pts[ii] = INT64_MAX; + return; } - hb_unlock(pv->common->mutex); - hb_cond_broadcast(pv->common->next_frame); - hb_yield(); -} - -// Keeps sync tasks "in sync". I.e. the lowest pts will always be -// output first. -static int waitForSync(hb_work_object_t * w, int64_t pts, int index, - hb_fifo_t *fifo) -{ - hb_work_private_t * pv = w->private_data; - setSyncPTS(pv, pts, index); + // Look for start of jitter sequence + jitter_stop = 0; + buf0 = hb_list_item(stream->in_queue, 0); + buf1 = hb_list_item(stream->in_queue, 1); + if (ABS(buf0->s.duration - (buf1->s.start - stream->next_pts)) < 1.1) + { + // Ignore very small jitter + return; + } + buf = hb_list_item(stream->in_queue, 0); + duration = buf->s.duration; - int ii; - hb_lock(pv->common->mutex); - for (ii = 0; ii < pv->common->pts_count; ii++) + // Look for end of jitter sequence + for (ii = 1; ii < count; ii++) { - while (pts > pv->common->last_pts[ii]) + buf = hb_list_item(stream->in_queue, ii); + if (ABS(duration - (buf->s.start - stream->next_pts)) < (90 * 40)) { - // wait for other streams to catch up - // since fifos can become full and clog up the works, - // check if our fifo is full when waking. - // Also check if encoding was canceled. - hb_cond_timedwait(pv->common->next_frame, pv->common->mutex, 200); - if (*w->done || - (pts > pv->common->last_pts[ii] && hb_fifo_is_full(fifo))) - { - hb_unlock(pv->common->mutex); - getPtsOffset(pv); - return 0; - } + // Finds the largest span that has low jitter + jitter_stop = ii; } + duration += buf->s.duration; + } + if (jitter_stop >= 4) + { + removeAudioJitter(stream, jitter_stop); } - hb_unlock(pv->common->mutex); - getPtsOffset(pv); - return 1; } -static void flushSubtitles(hb_work_private_t *pv) +// Fix audio gaps that could not be corrected with dejitter +static void fixAudioGap( sync_stream_t * stream ) { - hb_job_t * job = pv->job; - hb_subtitle_t * subtitle; - int ii; + int64_t gap; + hb_buffer_t * buf; - /* - * Push through any subtitle EOFs in case they were not synced through. - */ - for (ii = 0; ii < hb_list_count(job->list_subtitle); ii++) + if (hb_list_count(stream->in_queue) < 1 || !stream->first_frame) { - subtitle = hb_list_item(job->list_subtitle, ii); - // flush out any pending subtitle buffers in the sanitizer - hb_buffer_t *out = sanitizeSubtitle(pv, ii, NULL); - if (out != NULL) + // Can't find gaps with < 1 buffers + return; + } + + buf = hb_list_item(stream->in_queue, 0); + gap = buf->s.start - stream->next_pts; + + // there's a gap of more than a minute between the last + // frame and this. assume we got a corrupted timestamp + if (gap > 90 * 20 && gap < 90000LL * 60) + { + if (stream->gap_duration <= 0) { - hb_fifo_push(subtitle->fifo_out, out); + stream->gap_pts = buf->s.start; } - if (subtitle->config.dest == PASSTHRUSUB) + addDelta(stream->common, stream->next_pts, gap); + applyDeltas(stream->common); + stream->gap_duration += gap; + } + else + { + // If not fixing the gap, carry to the next frame + // so they do not accumulate. Do not carry negative gaps. + // Those are overlaps and are handled by fixAudioOverlap. + if (gap >= 90000LL * 60) { - hb_fifo_push(subtitle->fifo_out, hb_buffer_eof_init()); + // Fix "corrupted" timestamp + buf->s.start = stream->next_pts; + } + if (stream->gap_duration > 0) + { + hb_deep_log(3, "sync: audio 0x%x time gap %d ms. PTS %"PRId64"", + stream->audio.audio->id, (int)stream->gap_duration / 90, + stream->gap_pts); + stream->gap_duration = 0; } } } -/*********************************************************************** - * syncVideoWork - *********************************************************************** - * - **********************************************************************/ -int syncVideoWork( hb_work_object_t * w, hb_buffer_t ** buf_in, - hb_buffer_t ** buf_out ) +// Fix audio overlaps that could not be corrected with dejitter +static void fixAudioOverlap( sync_stream_t * stream ) { - 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; + int drop = 0; + int64_t overlap; + hb_buffer_t * buf; - next = *buf_in; - *buf_in = NULL; + if (!stream->first_frame) + { + // There are no overlaps if we haven't seen the first frame yet. + return; + } - if (next->s.flags & HB_BUF_FLAG_EOF) + // If time goes backwards drop the frame. + // Check if subsequent buffers also overlap. + while ((buf = hb_list_item(stream->in_queue, 0)) != NULL) { - if (sync->cur != NULL) + overlap = stream->next_pts - buf->s.start; + if (overlap > 90 * 20) { - cur = sync->cur; - cur->s.start = sync->next_start; - cur->s.stop = cur->s.start + 90000LL * - job->vrate.den / job->vrate.num; - - /* 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 = next; + if (stream->drop == 0) + { + stream->drop_pts = buf->s.start; + } + // This is likely to generate a gap in audio timestamps + // This will be subsequently handled by the call to + // fix AudioGap in Synchronize(). Small gaps will be handled + // by just shifting the timestamps and carrying the gap + // along. + hb_list_rem(stream->in_queue, buf); + signalBuffer(stream); + stream->drop_duration += buf->s.duration; + stream->drop++; + drop++; + hb_buffer_close(&buf); } else { - *buf_out = next; - } - flushSubtitles(pv); - pv->common->start_found = 1; - // Unblock anybody waiting on this threads last PTS - setSyncPTS(pv, INT64_MAX, 0); - if (job->indepth_scan) - { - // During subtitle scan, sync is the end of the pipeline. - // Terminate job when EOF reached. - *w->done = 1; + break; } - return HB_WORK_DONE; } - next_start = next->s.start - pv->common->video_pts_slip; - if (pv->common->pts_offset == INT64_MIN || !pv->common->start_found || - job->frame_to_stop > 0) + if (drop <= 0 && stream->drop > 0) { - waitForSync(w, next_start, 0, w->fifo_in); - // video_pts_slip may change in during waitForSync - next_start = next->s.start - pv->common->video_pts_slip; + hb_log("sync: audio 0x%x time went backwards %d ms, dropped %d frames. " + "PTS %"PRId64"", + stream->audio.audio->id, (int)stream->drop_duration / 90, + stream->drop, stream->drop_pts); + stream->drop_duration = 0; + stream->drop = 0; } - else +} + +static void fixStreamTimestamps( sync_stream_t * stream ) +{ + // Fix gaps and overlaps in queue + if (stream->type == SYNC_TYPE_AUDIO) { - setSyncPTS(pv, next_start, 0); + dejitterAudio(stream); + fixAudioOverlap(stream); + fixAudioGap(stream); } + else if (stream->type == SYNC_TYPE_VIDEO) + { + dejitterVideo(stream); + fixVideoOverlap(stream); + } +} + +static void sendEof( sync_common_t * common ) +{ + int ii; - /* Wait for start of point-to-point encoding */ - if (!pv->common->start_found) + for (ii = 0; ii < common->stream_count; ii++) { - if (pv->common->count_frames < job->frame_to_start || - next->s.start < job->pts_to_start) + hb_buffer_list_append(&common->streams[ii].out_queue, + hb_buffer_eof_init()); + } +} + +static void streamFlush( sync_stream_t * stream ) +{ + while (hb_list_count(stream->in_queue) > 0) + { + fixStreamTimestamps(stream); + hb_buffer_t * buf = hb_list_item(stream->in_queue, 0); + if (buf != NULL) { - UpdateSearchState( w, next_start ); + hb_list_rem(stream->in_queue, buf); + if ((buf->s.start < 0) || + (stream->type == SYNC_TYPE_VIDEO && buf->s.duration < 256)) + { + // The pipeline can't handle negative timestamps + // and it is sometimes not possible to avoid one + // at the start of the video. There can be a + // significant delay before we see the first buffers + // from all streams. We can't buffer indefinitely + // until we have seen the first PTS for all streams + // so sometimes we may start before we have seen + // the earliest PTS + // + // Also, encx264.c can't handle timestamps that are spaced + // less than 256 ticks apart. + hb_buffer_close(&buf); + continue; + } - // Flush any subtitles that have pts prior to the - // current frame - for( i = 0; i < hb_list_count( job->list_subtitle ); i++) + if (!stream->first_frame) { - subtitle = hb_list_item( job->list_subtitle, i ); - while( ( sub = hb_fifo_see( subtitle->fifo_raw ) ) ) + switch (stream->type) { - if ( sub->s.start > next->s.start ) + case SYNC_TYPE_VIDEO: + hb_log("sync: first pts video is %"PRId64, + buf->s.start); + break; + case SYNC_TYPE_AUDIO: + hb_log("sync: first pts audio 0x%x is %"PRId64, + stream->audio.audio->id, buf->s.start); + break; + case SYNC_TYPE_SUBTITLE: + hb_log("sync: first pts subtitle 0x%x is %"PRId64, + stream->subtitle.subtitle->id, buf->s.start); + break; + default: break; - sub = hb_fifo_get( subtitle->fifo_raw ); - hb_buffer_close( &sub ); } + stream->first_frame = 1; + stream->first_pts = buf->s.start; + stream->next_pts = buf->s.start; + stream->min_frame_duration = buf->s.duration; } -#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) + if (stream->type == SYNC_TYPE_AUDIO) { - 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); + buf = FilterAudioFrame(stream, buf); } -#endif - hb_buffer_close( &next ); + if (stream->type == SYNC_TYPE_AUDIO || + stream->type == SYNC_TYPE_VIDEO) + { + buf->s.start = stream->next_pts; + buf->s.stop = stream->next_pts + buf->s.duration; + } + + stream->next_pts += buf->s.duration; - return HB_WORK_OK; + if (buf->s.stop > 0) + { + stream->current_duration = buf->s.stop - stream->first_pts; + } + stream->frame_count++; + if (buf->s.duration > 0 && + stream->min_frame_duration > buf->s.duration) + { + stream->min_frame_duration = buf->s.duration; + } + if (stream->max_frame_duration < buf->s.duration) + { + stream->max_frame_duration = buf->s.duration; + } + hb_buffer_list_append(&stream->out_queue, buf); } - hb_lock( pv->common->mutex ); - pv->common->audio_pts_slip += next_start; - pv->common->video_pts_slip += next_start; - pv->common->start_found = 1; - pv->common->count_frames = 0; - hb_unlock( pv->common->mutex ); - next_start = 0; - sync->st_first = 0; - resetSync(pv); - } - - /* 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_log("sync: reached %d frames, exiting early", - pv->common->count_frames); - hb_buffer_close(&sync->cur); - hb_buffer_close(&next); - *buf_out = hb_buffer_eof_init(); - flushSubtitles(pv); - // Unblock anybody waiting on this threads last PTS - setSyncPTS(pv, INT64_MAX, 0); - return HB_WORK_DONE; } + hb_buffer_list_append(&stream->out_queue, hb_buffer_eof_init()); +} - /* 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", - sync->cur->s.start); - hb_buffer_close(&sync->cur); - hb_buffer_close( &next ); - *buf_out = hb_buffer_eof_init(); - flushSubtitles(pv); - // Unblock anybody waiting on this threads last PTS - setSyncPTS(pv, INT64_MAX, 0); - return HB_WORK_DONE; - } +// OutputBuffer pulls buffers from the internal sync buffer queues in +// lowest PTS first order. It then processes the queue the buffer is +// pulled from for frame overlaps and gaps. +// +// If a queue reaches MAX depth, it is possible another queue is too low +// to achieve both goals of pulling lowest PTS first *and* perform +// timestamp correction. In this scenario, we forego lowest PTS and pull +// the next lowest PTS that has enough buffers in the queue to perform +// timestamp correction. +static void OutputBuffer( sync_common_t * common ) +{ + int ii, full; + int64_t pts; + sync_stream_t * out_stream; + hb_buffer_t * buf; - if (sync->cur == NULL) + if (common->done) { - sync->cur = next; - return HB_WORK_OK; + // It is possible to get here when one stream triggers + // end of output (i.e. pts_to_stop or frame_to_stop) while + // another stream is waiting on the mutex. + return; } - - // At this point, we have 2 video frames wich allows us to set the - // duration of the first and output it. - cur = sync->cur; - if (sync->first_frame) + do { - /* This is our first frame */ - if (cur->s.start > 0) + full = 0; + out_stream = NULL; + pts = INT64_MAX; + + // Find lowest PTS and output that buffer + for (ii = 0; ii < common->stream_count; ii++) { - /* - * 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_stream_t * stream = &common->streams[ii]; + // We need at least 2 buffers in the queue in order to fix + // frame overlaps and inter-frame gaps. So if a queue is + // low, do not do normal PTS interleaving with this queue. + // Except for subtitles which are not processed for gaps + // and overlaps. + if (hb_list_count(stream->in_queue) > stream->min_len) + { + buf = hb_list_item(stream->in_queue, 0); + if (buf->s.start < pts) + { + pts = buf->s.start; + out_stream = stream; + } + } + // But continue output of buffers as long as one of the queues + // is above the maximum queue level. + if(hb_list_count(stream->in_queue) > stream->max_len) + { + full = 1; + } } - 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 ) + if (out_stream == NULL) { - sync->first_drop = next_start; + // This should only happen if all queues are below the + // minimum queue level + break; } - ++sync->drop_count; - if ( next->s.new_chap ) + + if (out_stream->next_pts == AV_NOPTS_VALUE) { - // don't drop a chapter mark when we drop the buffer - sync->chap_mark = next->s.new_chap; + // Initialize next_pts, it is used to make timestamp corrections + buf = hb_list_item(out_stream->in_queue, 0); + out_stream->next_pts = buf->s.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) + // Make timestamp adjustments to eliminate jitter, gaps, and overlaps + fixStreamTimestamps(out_stream); + + buf = hb_list_item(out_stream->in_queue, 0); + if (buf == NULL) + { + // In case some timestamp sanitization causes the one and + // only buffer in the queue to be deleted... + // This really shouldn't happen. + continue; + } + if (!common->start_found) { - av_qsv_stage *stage = av_qsv_get_last_stage(next->qsv_details.qsv_atom); - if (stage != NULL) + // pts_to_start or frame_to_start were specified. + // Wait for the appropriate start point. + if (out_stream->type == SYNC_TYPE_VIDEO && + common->job->frame_to_start > 0 && + out_stream->frame_count >= common->job->frame_to_start) { - av_qsv_wait_on_sync(job->qsv.ctx, stage); - if (stage->out.sync->in_use > 0) + common->start_found = 1; + out_stream->frame_count = 0; + } + else if (common->job->pts_to_start > 0 && + buf->s.start >= common->job->pts_to_start) + { + common->start_found = 1; + common->streams[0].frame_count = 0; + } + else + { + if (out_stream->type == SYNC_TYPE_VIDEO) { - ff_qsv_atomic_dec(&stage->out.sync->in_use); + UpdateSearchState(common, buf->s.start, + out_stream->frame_count); + out_stream->frame_count++; } - if (stage->out.p_surface->Data.Locked > 0) + hb_list_rem(out_stream->in_queue, buf); + signalBuffer(out_stream); + hb_buffer_close(&buf); + continue; + } + // reset frame count to track number of frames after + // the start position till the end of encode. + shiftTS(common, buf->s.start); + } + + // If pts_to_stop or frame_to_stop were specified, stop output + if (common->job->pts_to_stop && + buf->s.start >= common->job->pts_to_stop ) + { + hb_log("sync: reached pts %"PRId64", exiting early", buf->s.start); + common->done = 1; + sendEof(common); + return; + } + if (out_stream->type == SYNC_TYPE_VIDEO && + common->job->frame_to_stop && + out_stream->frame_count >= common->job->frame_to_stop) + { + hb_log("sync: reached %d frames, exiting early", + out_stream->frame_count); + common->done = 1; + sendEof(common); + return; + } + + // Out the buffer goes... + hb_list_rem(out_stream->in_queue, buf); + signalBuffer(out_stream); + if ((buf->s.start < 0) || + (out_stream->type == SYNC_TYPE_VIDEO && buf->s.duration < 256)) + { + // The pipeline can't handle negative timestamps + // and it is sometimes not possible to avoid one + // at the start of the video. There can be a + // significant delay before we see the first buffers + // from all streams. We can't buffer indefinitely + // until we have seen the first PTS for all streams + // so sometimes we may start before we have seen + // the earliest PTS + // + // Also, encx264.c can't handle timestamps that are spaced + // less than 256 ticks apart. + hb_buffer_close(&buf); + } + else + { + if (out_stream->type == SYNC_TYPE_VIDEO) + { + UpdateState(common, out_stream->frame_count); + } + if (!out_stream->first_frame) + { + switch (out_stream->type) { - ff_qsv_atomic_dec(&stage->out.p_surface->Data.Locked); + case SYNC_TYPE_VIDEO: + hb_log("sync: first pts video is %"PRId64, + buf->s.start); + break; + case SYNC_TYPE_AUDIO: + hb_log("sync: first pts audio 0x%x is %"PRId64, + out_stream->audio.audio->id, buf->s.start); + break; + case SYNC_TYPE_SUBTITLE: + hb_log("sync: first pts subtitle 0x%x is %"PRId64, + out_stream->subtitle.subtitle->id, buf->s.start); + break; + default: + break; } + out_stream->first_frame = 1; + out_stream->first_pts = buf->s.start; + out_stream->next_pts = buf->s.start; + out_stream->min_frame_duration = buf->s.duration; + } + + if (out_stream->type == SYNC_TYPE_AUDIO) + { + buf = FilterAudioFrame(out_stream, buf); + } + if (out_stream->type == SYNC_TYPE_AUDIO || + out_stream->type == SYNC_TYPE_VIDEO) + { + buf->s.start = out_stream->next_pts; + buf->s.stop = out_stream->next_pts + buf->s.duration; + } + + out_stream->next_pts += buf->s.duration; + + if (buf->s.stop > 0) + { + out_stream->current_duration = buf->s.stop - + out_stream->first_pts; + } + out_stream->frame_count++; + if (buf->s.duration > 0 && + out_stream->min_frame_duration > buf->s.duration) + { + out_stream->min_frame_duration = buf->s.duration; + } + if (out_stream->max_frame_duration < buf->s.duration) + { + out_stream->max_frame_duration = buf->s.duration; } - av_qsv_flush_stages(job->qsv.ctx->pipes, - &next->qsv_details.qsv_atom); + hb_buffer_list_append(&out_stream->out_queue, buf); } -#endif + } while (full); +} + +static void Synchronize( sync_common_t * common ) +{ + hb_lock(common->mutex); - hb_buffer_close( &next ); - return HB_WORK_OK; + if (!fillQueues(common)) + { + hb_unlock(common->mutex); + return; } - if ( sync->first_drop ) + if (!common->found_first_pts) { - 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; + checkFirstPts(common); } + OutputBuffer(common); - /* - * 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. + hb_unlock(common->mutex); +} - for( i = 0; i < hb_list_count( job->list_subtitle ); i++) +static void updateDuration( sync_stream_t * stream, int64_t start ) +{ + // The video decoder does not set an initial duration for frames. + // So set it here. + if (stream->type == SYNC_TYPE_VIDEO) { - 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 ) + int count = hb_list_count(stream->in_queue); + if (count > 0) { - if (!(next->s.flags & HB_BUF_FLAG_EOF)) + hb_buffer_t * buf = hb_list_item(stream->in_queue, count - 1); + double duration = start - buf->s.start; + if (duration > 0) { - out = sanitizeSubtitle(pv, i, sub); - if (out != NULL) - hb_fifo_push( subtitle->fifo_out, out ); + buf->s.duration = duration; + buf->s.stop = start; } else { - // Push the end of stream marker - hb_fifo_push( subtitle->fifo_out, sub ); + buf->s.duration = 0.; + buf->s.stop = buf->s.start; } } } +} - /* - * 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->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 ); +static void QueueBuffer( sync_stream_t * stream, hb_buffer_t * buf ) +{ + hb_lock(stream->common->mutex); - return HB_WORK_OK; + // We do not place a limit on the number of subtitle frames + // that are buffered becuase there are cases where we will + // receive all the subtitles for a file all at once (SSA subs). + // If we did not buffer these subs here, the following deadlock + // condition would occur: + // 1. Subtitle decoder blocks trying to generate more subtitle + // lines than will fit in sync input buffers. + // 2. This blocks the reader. Reader doesn't read any more + // audio or video, so sync won't receive buffers it needs + // to unblock subtitles. + while (hb_list_count(stream->in_queue) > stream->max_len) + { + hb_cond_wait(stream->cond_full, stream->common->mutex); + } + + buf->s.start -= stream->pts_slip; + if (buf->s.stop != AV_NOPTS_VALUE) + { + buf->s.stop -= stream->pts_slip; + } + // Render offset is only useful for decoders, which are all + // upstream of sync. Squash it. + buf->s.renderOffset = AV_NOPTS_VALUE; + updateDuration(stream, buf->s.start); + hb_list_add(stream->in_queue, buf); + + // Make adjustments for gaps found in other streams + applyDeltas(stream->common); + + hb_unlock(stream->common->mutex); } -hb_work_object_t hb_sync_video = +static int InitAudio( sync_common_t * common, int index ) { - WORK_SYNC_VIDEO, - "Video Synchronization", - syncVideoInit, - syncVideoWork, - syncVideoClose -}; + hb_work_object_t * w = NULL; + hb_work_private_t * pv; + hb_audio_t * audio; + + audio = hb_list_item(common->job->list_audio, index); + if (audio->priv.fifo_raw == NULL) + { + // No input fifo, not configured + return 0; + } + pv = calloc(1, sizeof(hb_work_private_t)); + if (pv == NULL) goto fail; + + pv->common = common; + pv->stream = &common->streams[1 + index]; + pv->stream->common = common; + pv->stream->cond_full = hb_cond_init(); + if (pv->stream->cond_full == NULL) goto fail; + pv->stream->in_queue = hb_list_init(); + pv->stream->max_len = SYNC_MAX_AUDIO_QUEUE_LEN; + pv->stream->min_len = SYNC_MIN_AUDIO_QUEUE_LEN; + if (pv->stream->in_queue == NULL) goto fail; + pv->stream->delta_list = hb_list_init(); + if (pv->stream->delta_list == NULL) goto fail; + pv->stream->type = SYNC_TYPE_AUDIO; + pv->stream->first_pts = AV_NOPTS_VALUE; + pv->stream->next_pts = AV_NOPTS_VALUE; + pv->stream->audio.audio = audio; + + w = hb_get_work(common->job->h, WORK_SYNC_AUDIO); + w->private_data = pv; + w->audio = audio; + w->fifo_in = audio->priv.fifo_raw; + if (audio->config.out.codec & HB_ACODEC_PASS_FLAG) + { + w->fifo_out = audio->priv.fifo_out; + } + else + { + w->fifo_out = audio->priv.fifo_sync; + } -/*********************************************************************** - * Close Audio - *********************************************************************** - * - **********************************************************************/ -void syncAudioClose( hb_work_object_t * w ) + if (!(audio->config.out.codec & HB_ACODEC_PASS_FLAG) && + audio->config.in.samplerate != audio->config.out.samplerate) + { + /* Initialize libsamplerate */ + int error; + pv->stream->audio.src.ctx = src_new(SRC_SINC_MEDIUM_QUALITY, + hb_mixdown_get_discrete_channel_count(audio->config.out.mixdown), + &error); + if (pv->stream->audio.src.ctx == NULL) goto fail; + pv->stream->audio.src.pkt.end_of_input = 0; + } + + pv->stream->audio.gain_factor = pow(10, audio->config.out.gain / 20); + + hb_list_add(common->list_work, w); + + return 0; + +fail: + if (pv != NULL) + { + if (pv->stream != NULL) + { + if (pv->stream->audio.src.ctx) + { + src_delete(pv->stream->audio.src.ctx); + } + hb_list_close(&pv->stream->delta_list); + hb_list_close(&pv->stream->in_queue); + hb_cond_close(&pv->stream->cond_full); + } + } + free(pv); + free(w); + + return 1; +} + +static int InitSubtitle( sync_common_t * common, int index ) { - hb_work_private_t * pv = w->private_data; - hb_sync_audio_t * sync = &pv->type.audio; + hb_work_object_t * w = NULL; + hb_work_private_t * pv; + hb_subtitle_t * subtitle; - // Unblock anybody waiting on this threads last PTS - setSyncPTS(pv, INT64_MAX, sync->index+1); + subtitle = hb_list_item(common->job->list_subtitle, index); + if (subtitle->fifo_raw == NULL) + { + // No input fifo, not configured + return 0; + } + pv = calloc(1, sizeof(hb_work_private_t)); + if (pv == NULL) goto fail; + + pv->common = common; + pv->stream = + &common->streams[1 + hb_list_count(common->job->list_audio) + index]; + pv->stream->common = common; + pv->stream->cond_full = hb_cond_init(); + if (pv->stream->cond_full == NULL) goto fail; + pv->stream->in_queue = hb_list_init(); + pv->stream->max_len = SYNC_MAX_SUBTITLE_QUEUE_LEN; + pv->stream->min_len = SYNC_MIN_SUBTITLE_QUEUE_LEN; + if (pv->stream->in_queue == NULL) goto fail; + pv->stream->delta_list = hb_list_init(); + if (pv->stream->delta_list == NULL) goto fail; + pv->stream->type = SYNC_TYPE_SUBTITLE; + pv->stream->first_pts = AV_NOPTS_VALUE; + pv->stream->next_pts = AV_NOPTS_VALUE; + pv->stream->subtitle.subtitle = subtitle; + + w = hb_get_work(common->job->h, WORK_SYNC_SUBTITLE); + w->private_data = pv; + w->subtitle = subtitle; + w->fifo_in = subtitle->fifo_raw; + w->fifo_out = subtitle->fifo_out; - if( sync->silence_buf ) + memset(&pv->stream->subtitle.sanitizer, 0, + sizeof(pv->stream->subtitle.sanitizer)); + if (subtitle->format == TEXTSUB && subtitle->config.dest == PASSTHRUSUB && + (common->job->mux & HB_MUX_MASK_MP4)) { - free( sync->silence_buf ); + // Merge overlapping subtitles since mpv tx3g does not support them + pv->stream->subtitle.sanitizer.merge = 1; } - if ( sync->state ) + // PGS subtitles don't need to be linked because there are explicit + // "clear" subtitle packets that indicate the end time of the + // previous subtitle + if (subtitle->config.dest == PASSTHRUSUB && + subtitle->source != PGSSUB) { - src_delete( sync->state ); + // Fill in stop time when it is missing + pv->stream->subtitle.sanitizer.link = 1; } + hb_buffer_list_clear(&pv->stream->subtitle.sanitizer.list_current); - free( pv ); - w->private_data = NULL; -} + hb_list_add(common->list_work, w); -int syncAudioInit( hb_work_object_t * w, hb_job_t * job) -{ return 0; + +fail: + if (pv != NULL) + { + if (pv->stream != NULL) + { + hb_list_close(&pv->stream->delta_list); + hb_list_close(&pv->stream->in_queue); + hb_cond_close(&pv->stream->cond_full); + } + } + free(pv); + free(w); + + return 1; } /*********************************************************************** - * SyncAudio - *********************************************************************** - * + * Initialize the work object **********************************************************************/ -static int syncAudioWork( hb_work_object_t * w, hb_buffer_t ** buf_in, - hb_buffer_t ** buf_out ) +static int syncVideoInit( hb_work_object_t * w, hb_job_t * job) { - 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; + hb_work_private_t * pv; + int ii; - buf = *buf_in; - *buf_in = NULL; + pv = calloc(1, sizeof(hb_work_private_t)); + if (pv == NULL) goto fail; + w->private_data = pv; + pv->common = calloc(1, sizeof(sync_common_t)); + if (pv->common == NULL) goto fail; + pv->common->job = job; - /* if the next buffer is an eof send it downstream */ - if (buf->s.flags & HB_BUF_FLAG_EOF) - { - *buf_out = buf; - // Unblock anybody waiting on this threads last PTS - setSyncPTS(pv, INT64_MAX, sync->index+1); - return HB_WORK_DONE; + // count number of streams we need + pv->common->stream_count = 1; + pv->common->stream_count += hb_list_count(job->list_audio); + pv->common->stream_count += hb_list_count(job->list_subtitle); + pv->common->streams = calloc(pv->common->stream_count, + sizeof(sync_stream_t)); + + // Allocate streams + if (pv->common->streams == NULL) goto fail; + + // create audio and subtitle work list + pv->common->list_work = hb_list_init(); + if (pv->common->list_work == NULL) goto fail; + + // mutex to mediate access to pv->common + pv->common->mutex = hb_lock_init(); + if (pv->common->mutex == NULL) goto fail; + + // Set up video sync work object + pv->stream = &pv->common->streams[0]; + pv->stream->common = pv->common; + pv->stream->cond_full = hb_cond_init(); + if (pv->stream->cond_full == NULL) goto fail; + pv->stream->in_queue = hb_list_init(); + pv->stream->max_len = SYNC_MAX_VIDEO_QUEUE_LEN; + pv->stream->min_len = SYNC_MIN_VIDEO_QUEUE_LEN; + if (pv->stream->in_queue == NULL) goto fail; + pv->stream->delta_list = hb_list_init(); + if (pv->stream->delta_list == NULL) goto fail; + pv->stream->type = SYNC_TYPE_VIDEO; + pv->stream->first_pts = AV_NOPTS_VALUE; + pv->stream->next_pts = AV_NOPTS_VALUE; + + w->fifo_in = job->fifo_raw; + // sync performs direct output to fifos + w->fifo_out = job->fifo_sync; + if (job->indepth_scan) + { + // When doing subtitle indepth scan, the pipeline ends at sync + w->fifo_out = NULL; } - start = buf->s.start - pv->common->audio_pts_slip; - if (pv->common->pts_offset == INT64_MIN || !pv->common->start_found || - job->frame_to_stop > 0) + if (job->pass_id == HB_PASS_ENCODE_2ND) { - waitForSync(w, start, sync->index+1, w->fifo_in); - // audio_pts_slip may change in waitForSync() - start = buf->s.start - pv->common->audio_pts_slip; + /* We already have an accurate frame count from pass 1 */ + hb_interjob_t * interjob = hb_interjob_get(job->h); + pv->common->est_frame_count = interjob->frame_count; } else { - setSyncPTS(pv, start, sync->index+1); - } - - // Wait for start frame if doing point-to-point - // - // When doing frame p-to-p, video leads the way. The video thead will set - // start_found when we have reached the start point. - if (!pv->common->start_found) - { - if (job->pts_to_start > 0 && buf->s.start >= job->pts_to_start) + /* Calculate how many video frames we are expecting */ + if (job->frame_to_stop) { - hb_lock( pv->common->mutex ); - pv->common->start_found = 1; - pv->common->audio_pts_slip += start; - pv->common->video_pts_slip += start; - pv->common->count_frames = 0; - hb_unlock( pv->common->mutex ); - resetSync(pv); - start = 0; + pv->common->est_frame_count = job->frame_to_stop; } else { - // For frame-to-frame encoding, the video sync thread will - // tell us when it is ok to start - hb_buffer_close(&buf); - return HB_WORK_OK; + int64_t duration; + if (job->pts_to_stop) + { + duration = job->pts_to_stop + 90000; + } + else + { + duration = 0; + for (ii = job->chapter_start; ii <= job->chapter_end; ii++) + { + hb_chapter_t * chapter; + chapter = hb_list_item(job->list_chapter, ii - 1); + if (chapter != NULL) + { + duration += chapter->duration; + } + } + } + pv->common->est_frame_count = duration * job->title->vrate.num / + job->title->vrate.den / 90000; } } + hb_log("sync: expecting %d video frames", pv->common->est_frame_count); - // Check for p-to-p end time - if (job->frame_to_stop && pv->common->count_frames >= job->frame_to_stop) + // Initialize audio sync work objects + for (ii = 0; ii < hb_list_count(job->list_audio); ii++ ) { - hb_buffer_close( &buf ); - *buf_out = hb_buffer_eof_init(); - // Unblock anybody waiting on this threads last PTS - setSyncPTS(pv, INT64_MAX, sync->index+1); - return HB_WORK_DONE; + if (InitAudio(pv->common, ii)) goto fail; } - if (job->pts_to_stop && sync->next_start >= job->pts_to_stop) + + // Initialize subtitle sync work objects + for (ii = 0; ii < hb_list_count(job->list_subtitle); ii++ ) { - hb_buffer_close( &buf ); - *buf_out = hb_buffer_eof_init(); - // Unblock anybody waiting on this threads last PTS - setSyncPTS(pv, INT64_MAX, sync->index+1); - return HB_WORK_DONE; + if (InitSubtitle(pv->common, ii)) goto fail; } - // 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 ) + /* Launch work processing threads */ + for (ii = 0; ii < hb_list_count(pv->common->list_work); ii++) { - // 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; + hb_work_object_t * work; + work = hb_list_item(pv->common->list_work, ii); + work->done = w->done; + work->thread = hb_thread_init(work->name, hb_work_loop, + work, HB_LOW_PRIORITY); } - if ( sync->first_drop ) + + if (job->frame_to_start || job->pts_to_start) { - // 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; + pv->common->start_found = 0; } - if ( start - sync->next_start >= (90 * 70) ) + else { - 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 ) + pv->common->start_found = 1; + } + + return 0; + +fail: + if (pv != NULL) + { + if (pv->common != NULL) { - 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; + for (ii = 0; ii < hb_list_count(pv->common->list_work); ii++) + { + hb_work_object_t * work; + work = hb_list_item(pv->common->list_work, ii); + if (work->close) work->close(work); + } + hb_list_close(&pv->common->list_work); + hb_lock_close(&pv->common->mutex); + if (pv->stream != NULL) + { + hb_list_close(&pv->stream->delta_list); + hb_list_close(&pv->stream->in_queue); + hb_cond_close(&pv->stream->cond_full); + } + free(pv->common->streams); + free(pv->common); } - 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; + } + free(pv); + w->private_data = NULL; + + return 1; } -hb_work_object_t hb_sync_audio = +/*********************************************************************** + * Close Video + *********************************************************************** + * + **********************************************************************/ +static void syncVideoClose( hb_work_object_t * w ) { - WORK_SYNC_AUDIO, - "AudioSynchronization", - syncAudioInit, - syncAudioWork, - syncAudioClose -}; + hb_work_private_t * pv = w->private_data; + hb_job_t * job = pv->common->job; -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; + if (pv == NULL) + { + return; + } + + hb_log("sync: got %d frames, %d expected", + pv->stream->frame_count, pv->common->est_frame_count ); + if (pv->stream->min_frame_duration > 0 && + pv->stream->max_frame_duration > 0 && + pv->stream->current_duration > 0) + { + hb_log("sync: framerate min %.3f fps, max %.3f fps, avg %.3f fps", + 90000. / pv->stream->max_frame_duration, + 90000. / pv->stream->min_frame_duration, + (pv->stream->frame_count * 90000.) / + pv->stream->current_duration); + } - pv = calloc( 1, sizeof( hb_work_private_t ) ); - sync = &pv->type.audio; - sync->index = i; - pv->job = job; - pv->common = common; - pv->common->pts_count++; + /* save data for second pass */ + if( job->pass_id == HB_PASS_ENCODE_1ST ) + { + /* Preserve frame count for better accuracy in pass 2 */ + hb_interjob_t * interjob = hb_interjob_get( job->h ); + interjob->frame_count = pv->stream->frame_count; + } + sync_delta_t * delta; + while ((delta = hb_list_item(pv->stream->delta_list, 0)) != NULL) + { + hb_list_rem(pv->stream->delta_list, delta); + free(delta); + } + hb_list_close(&pv->stream->delta_list); + hb_buffer_list_close(&pv->stream->out_queue); + hb_list_empty(&pv->stream->in_queue); + hb_cond_close(&pv->stream->cond_full); - w = hb_get_work( job->h, WORK_SYNC_AUDIO ); - hb_list_add(common->list_work, w); + // Close work threads + hb_work_object_t * work; + while ((work = hb_list_item(pv->common->list_work, 0))) + { + hb_list_rem(pv->common->list_work, work); + if (work->thread != NULL) + { + hb_thread_close(&work->thread); + } + if (work->close) work->close(work); + free(work); + } + hb_list_close(&pv->common->list_work); - w->private_data = pv; - w->audio = hb_list_item( job->list_audio, i ); - w->fifo_in = w->audio->priv.fifo_raw; - // Register condition with fifo to wake us up immediately if - // the fifo becomes full - hb_fifo_register_full_cond(w->fifo_in, pv->common->next_frame); + hb_lock_close(&pv->common->mutex); + free(pv->common->streams); + free(pv->common); + free(pv); + w->private_data = NULL; +} + +static hb_buffer_t * merge_ssa(hb_buffer_t *a, hb_buffer_t *b) +{ + int len, ii; + char *text; + hb_buffer_t *buf = hb_buffer_init(a->size + b->size); + buf->s = a->s; - if ( w->audio->config.out.codec & HB_ACODEC_PASS_FLAG ) + // Find the text in the second SSA sub + text = (char*)b->data; + for (ii = 0; ii < 8; ii++) { - w->fifo_out = w->audio->priv.fifo_out; + text = strchr(text, ','); + if (text == NULL) + break; + text++; + } + if (text != NULL) + { + len = sprintf((char*)buf->data, "%s\n%s", a->data, text); + if (len >= 0) + buf->size = len + 1; } else { - w->fifo_out = w->audio->priv.fifo_sync; + memcpy(buf->data, a->data, a->size); + buf->size = a->size; } - if( w->audio->config.out.codec == HB_ACODEC_AC3_PASS || - w->audio->config.out.codec == HB_ACODEC_AAC_PASS || - w->audio->config.out.codec == HB_ACODEC_EAC3_PASS || - w->audio->config.out.codec == HB_ACODEC_FLAC_PASS ) - { - /* Have a silent frame ready in case we have to fill a gap */ - AVDictionary *av_opts = NULL; - AVCodec *codec; - AVCodecContext *c; + return buf; +} + +static void setSubDuration(hb_buffer_t * buf) +{ + if (buf->s.stop != AV_NOPTS_VALUE) + buf->s.duration = buf->s.stop - buf->s.start; + else + buf->s.duration = AV_NOPTS_VALUE; +} + +static hb_buffer_t * mergeSubtitles(subtitle_sanitizer_t *sanitizer) +{ + hb_buffer_t *a, *b, *buf; + hb_buffer_list_t list; - switch ( w->audio->config.out.codec ) + hb_buffer_list_clear(&list); + + if (!sanitizer->merge) + { + // Handle all but the last buffer + while (hb_buffer_list_count(&sanitizer->list_current) > 1) { - case HB_ACODEC_AC3_PASS: - { - codec = avcodec_find_encoder( AV_CODEC_ID_AC3 ); - } break; - case HB_ACODEC_EAC3_PASS: - { - codec = avcodec_find_encoder( AV_CODEC_ID_EAC3 ); - } break; - case HB_ACODEC_AAC_PASS: - { - codec = avcodec_find_encoder_by_name("aac"); - } break; - case HB_ACODEC_FLAC_PASS: - { - codec = avcodec_find_encoder( AV_CODEC_ID_FLAC ); - } break; - default: - { - // Never gets here - codec = NULL; // Silence compiler warning - } break; + buf = hb_buffer_list_rem_head(&sanitizer->list_current); + setSubDuration(buf); + hb_buffer_list_append(&list, buf); } - - 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); - - /* - * lossless codecs may encode differently depending on the bit depth, so - * we need to set it correctly if we want the bitstream to be as close - * as possible to what we're passing through - */ - if (w->audio->config.out.codec == HB_ACODEC_FLAC_PASS) + // Check last buffer + buf = hb_buffer_list_head(&sanitizer->list_current); + if (buf != NULL) { - if (w->audio->config.in.sample_bit_depth <= 16) + if (buf->s.flags & HB_BUF_FLAG_EOF) { - hb_ff_set_sample_fmt(c, codec, AV_SAMPLE_FMT_S16); + buf = hb_buffer_list_rem_head(&sanitizer->list_current); + hb_buffer_list_append(&list, buf); } - else + else if (buf->s.stop != AV_NOPTS_VALUE) { - hb_ff_set_sample_fmt(c, codec, AV_SAMPLE_FMT_S32); + buf = hb_buffer_list_rem_head(&sanitizer->list_current); + setSubDuration(buf); + hb_buffer_list_append(&list, buf); } - c->bits_per_raw_sample = w->audio->config.in.sample_bit_depth; - } - else - { - hb_ff_set_sample_fmt(c, codec, AV_SAMPLE_FMT_FLTP); } + return hb_buffer_list_clear(&list); + } - /* - * the default frame size selected by the encoder may not match - * that of the input stream, but the FLAC encoder will honor whatever - * frame size we set as long as it's a valid FLAC block size. - * - * for AC-3, the frame size is the same for all streams. - * - * for E-AC-3, using the same bitrate and sample rate as the input - * should result in the frame size being the same as the source's. - */ - if (w->audio->config.out.codec == HB_ACODEC_FLAC_PASS) + // We only reach here if we are merging subtitles + while (hb_buffer_list_count(&sanitizer->list_current) > 1) + { + a = hb_buffer_list_head(&sanitizer->list_current); + if (a->s.flags & HB_BUF_FLAG_EOF) { - c->frame_size = w->audio->config.in.samples_per_frame; + buf = hb_buffer_list_rem_head(&sanitizer->list_current); + hb_buffer_list_append(&list, buf); + break; } - - /* - * we want the output to be as close as possible to what we're passing - * through, and we do have access to the source's matrix encoding mode. - */ - if (w->audio->config.out.codec == HB_ACODEC_AC3_PASS || - w->audio->config.out.codec == HB_ACODEC_EAC3_PASS) + b = a->next; + if (b->s.flags & HB_BUF_FLAG_EOF) { - switch (w->audio->config.in.matrix_encoding) - { - case AV_MATRIX_ENCODING_DOLBY: - case AV_MATRIX_ENCODING_DPLII: - av_dict_set(&av_opts, "dsur_mode", "on", 0); - break; - case AV_MATRIX_ENCODING_DPLIIX: - case AV_MATRIX_ENCODING_DOLBYEX: - av_dict_set(&av_opts, "dsurex_mode", "on", 0); - break; - case AV_MATRIX_ENCODING_DPLIIZ: - av_dict_set(&av_opts, "dsurex_mode", "dpliiz", 0); - break; - case AV_MATRIX_ENCODING_DOLBYHEADPHONE: - av_dict_set(&av_opts, "dheadphone_mode", "on", 0); - break; - default: - break; - } + buf = hb_buffer_list_rem_head(&sanitizer->list_current); + setSubDuration(buf); + hb_buffer_list_append(&list, buf); + buf = hb_buffer_list_rem_head(&sanitizer->list_current); + hb_buffer_list_append(&list, buf); + break; } - if (w->audio->config.in.channel_layout == AV_CH_LAYOUT_STEREO_DOWNMIX) + if (a->s.stop == AV_NOPTS_VALUE) { - c->channel_layout = AV_CH_LAYOUT_STEREO; - } - else - { - c->channel_layout = w->audio->config.in.channel_layout; + // To evaluate overlaps, we need the stop times + // This should really never happen... + break; } - if (hb_avcodec_open(c, codec, &av_opts, 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 + buf = NULL; + if (a->s.stop > b->s.start) { - // 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++ ) + // Overlap + if (ABS(a->s.start - b->s.start) <= 18000) { - // 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 ) + if (b->s.stop == AV_NOPTS_VALUE) { - hb_log("sync: track %d, avcodec_encode_audio() failed, dropping video to sync", - w->audio->config.out.track); - sync->drop_video_to_sync = 1; + // To evaluate overlaps, we need the stop times + // for a and b break; } + a = hb_buffer_list_rem_head(&sanitizer->list_current); - if ( got_packet ) + // subtitles start within 1/5 second of eachother, merge + if (a->s.stop > b->s.stop && b->s.stop != AV_NOPTS_VALUE) { - sync->silence_size = pkt.size; - break; + // a continues after b, reorder the list and swap + b = a; + a = hb_buffer_list_rem_head(&sanitizer->list_current); + hb_buffer_list_prepend(&sanitizer->list_current, b); } - else if (ii + 1 == 10) + + b->s.start = a->s.stop; + + buf = merge_ssa(a, b); + hb_buffer_close(&a); + a = buf; + + if (b->s.stop != AV_NOPTS_VALUE && + ABS(b->s.stop - b->s.start) <= 18000) { - 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; + // b and a completely overlap, remove b + b = hb_buffer_list_rem_head(&sanitizer->list_current); + hb_buffer_close(&b); } } - free( zeros ); - hb_avcodec_close( c ); + else + { + // a starts before b, output copy of a and update a start + buf = hb_buffer_dup(a); + buf->s.stop = b->s.start; + a->s.start = b->s.start; + } } - AVDictionaryEntry *opt = NULL; - while ((opt = av_dict_get(av_opts, "", opt, AV_DICT_IGNORE_SUFFIX)) != NULL) + else if (a->s.stop <= b->s.start) { - hb_log("InitAudio: unknown option '%s'", opt->key); + // a starts and ends before b + buf = hb_buffer_list_rem_head(&sanitizer->list_current); + } + + if (buf != NULL) + { + setSubDuration(buf); + hb_buffer_list_append(&list, buf); } - av_dict_free( &av_opts ); - av_free( c ); } - else + + return hb_buffer_list_clear(&list); +} + +static hb_buffer_t * sanitizeSubtitle( + subtitle_sanitizer_t * sanitizer, + hb_buffer_t * sub) +{ + hb_buffer_list_t list; + + if (sub == NULL) { - if( w->audio->config.out.codec & HB_ACODEC_PASS_FLAG ) + return NULL; + } + + hb_buffer_list_set(&list, sub); + if (!sanitizer->link && !sanitizer->merge) + { + sub = hb_buffer_list_head(&list); + while (sub != NULL) { - sync->drop_video_to_sync = 1; + if (sub->s.stop != AV_NOPTS_VALUE) + sub->s.duration = sub->s.stop - sub->s.start; + else + sub->s.duration = AV_NOPTS_VALUE; + sub = sub->next; + } + return hb_buffer_list_clear(&list); + } + + sub = hb_buffer_list_rem_head(&list); + while (sub != NULL) + { + if (sub->s.flags & HB_BUF_FLAG_EOF) + { + hb_buffer_list_append(&sanitizer->list_current, sub); + break; + } + hb_buffer_t *last = hb_buffer_list_tail(&sanitizer->list_current); + if (last != NULL && last->s.stop == AV_NOPTS_VALUE) + { + last->s.stop = sub->s.start; + } + + if (sub->s.start == sub->s.stop) + { + // Used to indicate "clear" subtitles when the duration + // of subtitles is not encoded in the stream + hb_buffer_close(&sub); } 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; + hb_buffer_list_append(&sanitizer->list_current, sub); } + sub = hb_buffer_list_rem_head(&list); + } + + return mergeSubtitles(sanitizer); +} + +/*********************************************************************** + * syncVideoWork + *********************************************************************** + * + **********************************************************************/ +static int syncVideoWork( hb_work_object_t * w, hb_buffer_t ** buf_in, + hb_buffer_t ** buf_out ) +{ + hb_work_private_t * pv = w->private_data; + hb_buffer_t * in = *buf_in; + + if (pv->common->done) + { + *buf_out = hb_buffer_list_clear(&pv->stream->out_queue); + return HB_WORK_DONE; + } + if (in->s.flags & HB_BUF_FLAG_EOF) + { + streamFlush(pv->stream); + *buf_out = hb_buffer_list_clear(&pv->stream->out_queue); + return HB_WORK_DONE; + } + + *buf_in = NULL; + QueueBuffer(pv->stream, in); + Synchronize(pv->common); + *buf_out = hb_buffer_list_clear(&pv->stream->out_queue); + + return HB_WORK_OK; +} + +hb_work_object_t hb_sync_video = +{ + WORK_SYNC_VIDEO, + "Video Synchronization", + syncVideoInit, + syncVideoWork, + syncVideoClose +}; + +/*********************************************************************** + * Close Audio + *********************************************************************** + * + **********************************************************************/ +static void syncAudioClose( hb_work_object_t * w ) +{ + hb_work_private_t * pv = w->private_data; + + if (pv == NULL) + { + return; + } + + // Free samplerate conversion context + if (pv->stream->audio.src.ctx) + { + src_delete(pv->stream->audio.src.ctx); + } + + sync_delta_t * delta; + while ((delta = hb_list_item(pv->stream->delta_list, 0)) != NULL) + { + hb_list_rem(pv->stream->delta_list, delta); + free(delta); + } + hb_list_close(&pv->stream->delta_list); + hb_buffer_list_close(&pv->stream->out_queue); + hb_list_empty(&pv->stream->in_queue); + hb_cond_close(&pv->stream->cond_full); + free(pv); + w->private_data = NULL; +} + +static 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_buffer_t * in = *buf_in; + + if (pv->common->done) + { + *buf_out = hb_buffer_list_clear(&pv->stream->out_queue); + return HB_WORK_DONE; + } + if (in->s.flags & HB_BUF_FLAG_EOF) + { + streamFlush(pv->stream); + *buf_out = hb_buffer_list_clear(&pv->stream->out_queue); + return HB_WORK_DONE; } - sync->gain_factor = pow(10, w->audio->config.out.gain / 20); + *buf_in = NULL; + QueueBuffer(pv->stream, in); + Synchronize(pv->common); + *buf_out = hb_buffer_list_clear(&pv->stream->out_queue); + + return HB_WORK_OK; } -static hb_buffer_t * OutputAudioFrame( hb_audio_t *audio, hb_buffer_t *buf, - hb_sync_audio_t *sync ) +hb_work_object_t hb_sync_audio = +{ + WORK_SYNC_AUDIO, + "Audio Synchronization", + syncAudioInit, + syncAudioWork, + syncAudioClose +}; + +// FilterAudioFrame is called after audio timestamp discontinuities +// have all been corrected. So we expect smooth continuous audio +// here. +static hb_buffer_t * FilterAudioFrame( sync_stream_t * stream, + hb_buffer_t *buf ) { - int64_t start = (int64_t)sync->next_start; + hb_audio_t * audio = stream->audio.audio; // 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 (buf->s.duration <= 0) + { + buf->s.duration = buf->s.stop - buf->s.start; + } - if ( !( audio->config.out.codec & HB_ACODEC_PASS_FLAG ) ) + if (!(audio->config.out.codec & HB_ACODEC_PASS_FLAG)) { + // TODO: this should all be replaced by an audio filter chain. + // Audio is not passthru. Check if we need to modify the audio // in any way. - if( audio->config.in.samplerate != audio->config.out.samplerate ) + if (stream->audio.src.ctx != NULL) { /* 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 ); + int sample_size = hb_mixdown_get_discrete_channel_count( + audio->config.out.mixdown ) * sizeof( float ); count_in = buf_raw->size / sample_size; /* @@ -1389,35 +1818,35 @@ static hb_buffer_t * OutputAudioFrame( hb_audio_t *audio, hb_buffer_t *buf, * 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; + count_out = (buf->s.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; + stream->audio.src.pkt.input_frames = count_in; + stream->audio.src.pkt.output_frames = count_out; + stream->audio.src.pkt.src_ratio = + (double)audio->config.out.samplerate / + 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 ) ) + buf->s = buf_raw->s; + stream->audio.src.pkt.data_in = (float *) buf_raw->data; + stream->audio.src.pkt.data_out = (float *) buf->data; + if (src_process(stream->audio.src.ctx, &stream->audio.src.pkt)) { /* XXX If this happens, we're screwed */ - hb_log( "sync: audio 0x%x src_process failed", audio->id ); + hb_error("sync: audio 0x%x src_process failed", audio->id); } - hb_buffer_close( &buf_raw ); + hb_buffer_close(&buf_raw); - if (sync->data.output_frames_gen <= 0) + if (stream->audio.src.pkt.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; + buf->s.duration = 90000. * stream->audio.src.pkt.output_frames_gen / + audio->config.out.samplerate; } - if( audio->config.out.gain > 0.0 ) + if (audio->config.out.gain > 0.0) { int count, ii; @@ -1427,7 +1856,7 @@ static hb_buffer_t * OutputAudioFrame( hb_audio_t *audio, hb_buffer_t *buf, double sample; sample = (double)*(((float*)buf->data)+ii); - sample *= sync->gain_factor; + sample *= stream->audio.gain_factor; if (sample > 0) sample = MIN(sample, 1.0); else @@ -1445,7 +1874,7 @@ static hb_buffer_t * OutputAudioFrame( hb_audio_t *audio, hb_buffer_t *buf, double sample; sample = (double)*(((float*)buf->data)+ii); - sample *= sync->gain_factor; + sample *= stream->audio.gain_factor; *(((float*)buf->data)+ii) = sample; } } @@ -1454,127 +1883,57 @@ static hb_buffer_t * OutputAudioFrame( hb_audio_t *audio, hb_buffer_t *buf, 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 ) +static void UpdateState( sync_common_t * common, int frame_count ) { - 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_job_t * job = common->job; hb_state_t state; - hb_get_state2( pv->job->h, &state ); - if ( !pv->common->count_frames ) + hb_get_state2(job->h, &state); + if (frame_count == 0) { - sync->st_first = hb_get_date(); - pv->job->st_pause_date = -1; - pv->job->st_paused = 0; + common->st_first = hb_get_date(); + job->st_pause_date = -1; + job->st_paused = 0; } - pv->common->count_frames++; - if (pv->job->indepth_scan) + if (job->indepth_scan) { // Progress for indept scan is handled by reader - // pv->common->count_frames is used during indepth_scan + // frame_count is used during indepth_scan // to find start & end points. return; } - if( hb_get_date() > sync->st_dates[3] + 1000 ) + if (hb_get_date() > common->st_dates[3] + 1000) { - memmove( &sync->st_dates[0], &sync->st_dates[1], + memmove( &common->st_dates[0], &common->st_dates[1], 3 * sizeof( uint64_t ) ); - memmove( &sync->st_counts[0], &sync->st_counts[1], + memmove( &common->st_counts[0], &common->st_counts[1], 3 * sizeof( uint64_t ) ); - sync->st_dates[3] = hb_get_date(); - sync->st_counts[3] = pv->common->count_frames; + common->st_dates[3] = hb_get_date(); + common->st_counts[3] = frame_count; } #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 = (float)frame_count / common->est_frame_count; + 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 ) + p.rate_cur = 1000.0 * (common->st_counts[3] - common->st_counts[0]) / + (common->st_dates[3] - common->st_dates[0]); + if (hb_get_date() > common->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.rate_avg = 1000.0 * common->st_counts[3] / + (common->st_dates[3] - common->st_first - job->st_paused); + eta = (common->est_frame_count - common->st_counts[3]) / p.rate_avg; p.hours = eta / 3600; - p.minutes = ( eta % 3600 ) / 60; + p.minutes = (eta % 3600) / 60; p.seconds = eta % 60; } else @@ -1586,65 +1945,63 @@ static void UpdateState( hb_work_object_t * w ) } #undef p - hb_set_state( pv->job->h, &state ); + hb_set_state(job->h, &state); } -static void UpdateSearchState( hb_work_object_t * w, int64_t start ) +static void UpdateSearchState( sync_common_t * common, int64_t start, + int frame_count ) { - 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; + hb_job_t * job = common->job; + hb_state_t state; + uint64_t now; + double avg; now = hb_get_date(); - if( !pv->common->count_frames ) + if (frame_count == 0) { - sync->st_first = now; - pv->job->st_pause_date = -1; - pv->job->st_paused = 0; + common->st_first = now; + job->st_pause_date = -1; + job->st_paused = 0; } - pv->common->count_frames++; - if (pv->job->indepth_scan) + if (job->indepth_scan) { // Progress for indept scan is handled by reader - // pv->common->count_frames is used during indepth_scan + // frame_count is used during indepth_scan // to find start & end points. return; } - hb_get_state2(pv->job->h, &state); + hb_get_state2(job->h, &state); #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; + if (job->frame_to_start) + p.progress = (float)frame_count / job->frame_to_start; + else if (job->pts_to_start) + p.progress = (float) start / job->pts_to_start; else p.progress = 0; - if( p.progress > 1.0 ) + if (p.progress > 1.0) { p.progress = 1.0; } - if (now > sync->st_first) + if (now > common->st_first) { int eta = 0; - if ( pv->job->frame_to_start ) + if (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; + avg = 1000.0 * frame_count / (now - common->st_first); + eta = (job->frame_to_start - frame_count ) / avg; } - else if ( pv->job->pts_to_start ) + else if (job->pts_to_start) { - avg = 1000.0 * (double)start / (now - sync->st_first); - eta = ( pv->job->pts_to_start - start ) / avg; + avg = 1000.0 * start / (now - common->st_first); + eta = (job->pts_to_start - start) / avg; } p.hours = eta / 3600; - p.minutes = ( eta % 3600 ) / 60; + p.minutes = (eta % 3600) / 60; p.seconds = eta % 60; } else @@ -1656,25 +2013,82 @@ static void UpdateSearchState( hb_work_object_t * w, int64_t start ) } #undef p - hb_set_state( pv->job->h, &state ); + hb_set_state(job->h, &state); } -static void getPtsOffset(hb_work_private_t * pv) +static int syncSubtitleInit( hb_work_object_t * w, hb_job_t * job ) { - if (pv->common->pts_offset != INT64_MIN) + return 0; +} + +static void syncSubtitleClose( hb_work_object_t * w ) +{ + hb_work_private_t * pv = w->private_data; + + if (pv == NULL) + { return; + } - int64_t first_pts = INT64_MAX; - int ii; - for (ii = 0; ii < pv->common->pts_count; ii++) + sync_delta_t * delta; + while ((delta = hb_list_item(pv->stream->delta_list, 0)) != NULL) + { + hb_list_rem(pv->stream->delta_list, delta); + free(delta); + } + hb_list_close(&pv->stream->delta_list); + hb_buffer_list_close(&pv->stream->out_queue); + hb_list_empty(&pv->stream->in_queue); + hb_cond_close(&pv->stream->cond_full); + hb_buffer_list_close(&pv->stream->subtitle.sanitizer.list_current); + free(pv); + w->private_data = NULL; +} + +static int syncSubtitleWork( hb_work_object_t * w, hb_buffer_t ** buf_in, + hb_buffer_t ** buf_out ) +{ + hb_work_private_t * pv = w->private_data; + hb_buffer_t * in = *buf_in; + + if (pv->common->done) + { + hb_buffer_list_append(&pv->stream->out_queue, hb_buffer_eof_init()); + in = hb_buffer_list_clear(&pv->stream->out_queue); + *buf_out = sanitizeSubtitle(&pv->stream->subtitle.sanitizer, in); + return HB_WORK_DONE; + } + if (in->s.flags & HB_BUF_FLAG_EOF) { - if (pv->common->last_pts[ii] != AV_NOPTS_VALUE && - pv->common->last_pts[ii] < first_pts) + streamFlush(pv->stream); + in = hb_buffer_list_clear(&pv->stream->out_queue); + *buf_out = sanitizeSubtitle(&pv->stream->subtitle.sanitizer, in); + if (pv->common->job->indepth_scan) { - first_pts = pv->common->last_pts[ii]; + // When doing subtitle indepth scan, the pipeline ends at sync. + // Terminate job when EOF reached. + *w->done = 1; } + return HB_WORK_DONE; + } + + *buf_in = NULL; + QueueBuffer(pv->stream, in); + Synchronize(pv->common); + in = hb_buffer_list_clear(&pv->stream->out_queue); + if (in != NULL) + { + *buf_out = sanitizeSubtitle(&pv->stream->subtitle.sanitizer, in); } - pv->common->video_pts_slip = pv->common->audio_pts_slip = - pv->common->pts_offset = first_pts; - return; + + return HB_WORK_OK; } + +hb_work_object_t hb_sync_subtitle = +{ + WORK_SYNC_SUBTITLE, + "Subitle Synchronization", + syncSubtitleInit, + syncSubtitleWork, + syncSubtitleClose +}; diff --git a/libhb/vfr.c b/libhb/vfr.c index 50eddd0e7..b1ba3196a 100644 --- a/libhb/vfr.c +++ b/libhb/vfr.c @@ -501,7 +501,7 @@ static int hb_vfr_work( hb_filter_object_t * filter, // If there is a gap between the last stop and the current start // then frame(s) were dropped. - if ( in->s.start > pv->last_stop[0] ) + if (hb_fifo_size(pv->delay_queue) > 0 && in->s.start > pv->last_stop[0]) { /* We need to compensate for the time lost by dropping frame(s). Spread its duration out in quarters, because usually dropped frames diff --git a/libhb/work.c b/libhb/work.c index 31e4cad96..1ed356dd1 100644 --- a/libhb/work.c +++ b/libhb/work.c @@ -169,13 +169,11 @@ hb_work_object_t* hb_audio_decoder(hb_handle_t *h, int codec) if (codec & HB_ACODEC_FF_MASK) { w = hb_get_work(h, WORK_DECAVCODEC); - w->yield = 1; // decoders yield to keep sync fifos more even } switch (codec) { case HB_ACODEC_LPCM: w = hb_get_work(h, WORK_DECLPCM); - w->yield = 1; // decoders yield to keep sync fifos more even break; default: break; @@ -1570,7 +1568,6 @@ static void do_job(hb_job_t *job) *job->die = 1; goto cleanup; } - w->yield = 1; // decoders yield to keep sync fifos more even w->fifo_in = job->fifo_mpeg2; w->fifo_out = job->fifo_raw; hb_list_add(job->list_work, w); @@ -1868,10 +1865,6 @@ void hb_work_loop( void * _w ) } } } - if (w->yield) - { - hb_yield(); - } } if ( buf_out ) { |