oak-team/oak-editor
3
1
Fork 0
Code Issues Pull Requests Actions 1 Packages Projects Releases Wiki Activity

nested sequence audio is still garbage

Browse Source
This commit is contained in:
itsmattkc
2018-11-10 13:34:25 +11:00
parent 586b24beed
commit 2461cd20cd
7 changed files with 106 additions and 93 deletions
Download Patch File Download Diff File Expand all files Collapse all files
playback/cacher.cpp
+63 -47
View File
@@ -74,13 +74,33 @@ int get_read_buffer_index(Clip* nest) {
return (nest == NULL) ? audio_ibuffer_read : nest->frame->pts;
}
void cache_audio_worker(Clip* c, bool scrubbing, Clip* nest) {
int get_real_buffer_write(Clip* c, Clip* nest, long frame) {
if (nest == NULL) {
return get_buffer_offset_from_frame(c->sequence, frame);
} else {
return qMax(0, qFloor((((double) frame / c->sequence->frame_rate) * sequence->audio_frequency * 4)/* - nest->frame->pkt_dts*/));
dout << "ti:" << c->timeline_in << "atf:" << c->audio_target_frame << "abw:" << c->audio_buffer_write << "nest pts:" << nest->frame->pts;
}
}
void cache_audio_worker(Clip* c, bool scrubbing, QVector<Clip*> nests) {
long timeline_in = c->timeline_in;
long timeline_out = c->timeline_out;
if (nest != NULL) {
timeline_in = refactor_frame_number(timeline_in, c->sequence->frame_rate, nest->sequence->frame_rate) + nest->timeline_in - nest->clip_in;
timeline_out = refactor_frame_number(timeline_out, c->sequence->frame_rate, nest->sequence->frame_rate) + nest->timeline_in - nest->clip_in;
}
Clip* nest = NULL;
long audio_target_frame = c->audio_target_frame;
if (nests.size() > 0) {
/*double last_fr = c->sequence->frame_rate;
for (int i=nests.size()-1;i>=0;i--) {
timeline_in = refactor_frame_number(timeline_in, last_fr, nests.at(i)->sequence->frame_rate) + nests.at(i)->timeline_in - nests.at(i)->clip_in;
timeline_out = refactor_frame_number(timeline_out, last_fr, nests.at(i)->sequence->frame_rate) + nests.at(i)->timeline_in - nests.at(i)->clip_in;
audio_target_frame = refactor_frame_number(audio_target_frame, last_fr, nests.at(i)->sequence->frame_rate) + nests.at(i)->timeline_in - nests.at(i)->clip_in;
last_fr = nests.at(i)->sequence->frame_rate;
}*/
nest = nests.last();
}
while (true) {
AVFrame* frame;
@@ -257,7 +277,7 @@ void cache_audio_worker(Clip* c, bool scrubbing, Clip* nest) {
if (c->audio_just_reset) {
// get precise sample offset for the elected clip_in from this audio frame
double target_sts = playhead_to_clip_seconds(c, c->audio_target_frame);
double target_sts = playhead_to_clip_seconds(c, audio_target_frame);
double frame_sts = ((frame->pts - c->stream->start_time) * timebase);
int nb_samples = qRound64((target_sts - frame_sts)*c->sequence->audio_frequency);
c->frame_sample_index = nb_samples * 4;
@@ -272,7 +292,10 @@ void cache_audio_worker(Clip* c, bool scrubbing, Clip* nest) {
#ifdef AUDIOWARNINGS
dout << "fsi-post-post:" << c->frame_sample_index;
#endif
if (c->audio_buffer_write == 0) c->audio_buffer_write = get_buffer_offset_from_frame(c->sequence, qMax(timeline_in, c->audio_target_frame));
if (c->audio_buffer_write == 0) {
c->audio_buffer_write = get_real_buffer_write(c, nest, qMax(c->timeline_in, c->audio_target_frame));
//if (nest != NULL) c->audio_buffer_write -= nest->frame->pkt_dts;
}
int offset = get_read_buffer_index(nest) + AUDIO_BUFFER_PADDING - c->audio_buffer_write;
if (offset > 0) {
@@ -309,7 +332,7 @@ void cache_audio_worker(Clip* c, bool scrubbing, Clip* nest) {
apply_audio_effects(c, bytes_to_seconds(frame->pts, frame->channels, frame->sample_rate), frame, nb_bytes);
c->frame->pts += nb_bytes;
c->frame_sample_index = 0;
if (c->audio_buffer_write == 0) c->audio_buffer_write = get_buffer_offset_from_frame(c->sequence, qMax(timeline_in, c->audio_target_frame));
if (c->audio_buffer_write == 0) c->audio_buffer_write = get_buffer_offset_from_frame(c->sequence, qMax(timeline_in, audio_target_frame));
int offset = (get_read_buffer_index(nest) + AUDIO_BUFFER_PADDING) - c->audio_buffer_write;
if (offset > 0) {
c->audio_buffer_write += offset;
@@ -319,12 +342,14 @@ void cache_audio_worker(Clip* c, bool scrubbing, Clip* nest) {
break;
case MEDIA_TYPE_SEQUENCE:
frame = c->frame;
nb_bytes = frame->nb_samples * av_get_bytes_per_sample(static_cast<AVSampleFormat>(frame->format)) * frame->channels;
nb_bytes = c->frame->linesize[0];
if (c->frame_sample_index == -1 || c->frame_sample_index >= nb_bytes) {
c->frame_sample_index = 0;
if (c->audio_buffer_write == 0) c->audio_buffer_write = get_buffer_offset_from_frame(c->sequence, qMax(timeline_in, c->audio_target_frame));
if (c->audio_buffer_write == 0) {
c->audio_buffer_write = get_buffer_offset_from_frame(c->sequence, qMax(timeline_in, audio_target_frame));
}
int offset = (get_read_buffer_index(nest) + AUDIO_BUFFER_PADDING) - c->audio_buffer_write;
if (offset > 0) {
@@ -333,19 +358,6 @@ void cache_audio_worker(Clip* c, bool scrubbing, Clip* nest) {
c->frame->pts += offset;
}
}
//c->audio_buffer_write = 0;
/*while ((c->frame_sample_index == -1 || c->frame_sample_index >= nb_bytes) && nb_bytes > 0) {
// create "new frame"
int offset = (get_read_buffer_index(nest) + AUDIO_BUFFER_PADDING) - c->audio_buffer_write;
if (offset > 0) {
c->audio_buffer_write += offset;
c->frame_sample_index += offset;
}
}*/
break;
}
@@ -353,29 +365,29 @@ void cache_audio_worker(Clip* c, bool scrubbing, Clip* nest) {
if (frame->nb_samples == 0) {
break;
} else {
long buffer_timeline_out = get_buffer_offset_from_frame(c->sequence, timeline_out);
long buffer_timeline_out = get_real_buffer_write(c, nest, timeline_out);
int max_write = (nest == NULL) ? audio_ibuffer_size : nest->frame->nb_samples * av_get_bytes_per_sample(static_cast<AVSampleFormat>(nest->frame->format)) * nest->frame->channels;
if (nest == NULL) {
int max_write = (nests.isEmpty()) ? audio_ibuffer_size : nest->frame->nb_samples * av_get_bytes_per_sample(static_cast<AVSampleFormat>(nest->frame->format)) * nest->frame->channels;
if (nests.size() == 0) {
audio_write_lock.lock();
} else {
c->queue_lock.lock();
}
uint8_t* buffer = (nest == NULL) ? audio_ibuffer : nest->frame->data[0];
dout << c->name << "|" <<
c->audio_buffer_write << "<" << get_read_buffer_index(nest)+(max_write>>1) << "|" <<
c->audio_buffer_write << "<" << buffer_timeline_out << "|" <<
c->frame_sample_index << "<" << nb_bytes << "|";
uint8_t* buffer = (nests.isEmpty()) ? audio_ibuffer : nest->frame->data[0];
int written = 0;
dout << c->name << "reports" << c->frame_sample_index << "(" << nb_bytes << ")" << c->audio_buffer_write << "(" << get_read_buffer_index(nest)+(max_write>>1) << "/" << buffer_timeline_out << ")";
while (c->frame_sample_index < nb_bytes
&& c->audio_buffer_write < get_read_buffer_index(nest)+(max_write>>1)
&& c->audio_buffer_write < buffer_timeline_out) {
int upper_byte_index = (c->audio_buffer_write+1)%max_write;
int lower_byte_index = (c->audio_buffer_write)%max_write;
qint16 old_sample = static_cast<int16_t>((buffer[upper_byte_index] & 0xFF) << 8 | (buffer[lower_byte_index] & 0xFF));
qint16 new_sample = static_cast<int16_t>((frame->data[0][c->frame_sample_index+1] & 0xFF) << 8 | (frame->data[0][c->frame_sample_index] & 0xFF));
qint16 mixed_sample = mix_audio_sample(old_sample, new_sample);
int16_t old_sample = static_cast<int16_t>((buffer[upper_byte_index] & 0xFF) << 8 | (buffer[lower_byte_index] & 0xFF));
int16_t new_sample = static_cast<int16_t>((frame->data[0][c->frame_sample_index+1] & 0xFF) << 8 | (frame->data[0][c->frame_sample_index] & 0xFF));
int16_t mixed_sample = mix_audio_sample(old_sample, new_sample);
buffer[upper_byte_index] = static_cast<uint8_t>((mixed_sample >> 8) & 0xFF);
buffer[lower_byte_index] = static_cast<uint8_t>(mixed_sample & 0xFF);
@@ -388,15 +400,17 @@ void cache_audio_worker(Clip* c, bool scrubbing, Clip* nest) {
c->audio_buffer_write+=2;
c->frame_sample_index+=2;
written += 2;
}
dout << c->name << "|" << c->frame_sample_index << "<" << nb_bytes << "LELELEL";
dout << c->name << "wrote" << written;
#ifdef AUDIOWARNINGS
if (c->audio_buffer_write >= buffer_timeline_out) dout << "timeline out at fsi" << c->frame_sample_index << "of frame ts" << c->frame->pts;
#endif
if (nest == NULL) {
if (nests.isEmpty()) {
audio_write_lock.unlock();
} else {
c->queue_lock.unlock();
@@ -416,8 +430,6 @@ void cache_audio_worker(Clip* c, bool scrubbing, Clip* nest) {
if (c->media_type == MEDIA_TYPE_SEQUENCE) {
break;
}
// dout << "ended" << c->frame_sample_index << nb_bytes;
}
if (c->reached_end) {
frame->nb_samples = 0;
@@ -479,8 +491,6 @@ void cache_video_worker(Clip* c, long playhead) {
send_it = true;
} else if (media->get_stream_from_file_index(true, c->media_stream)->infinite_length) {
send_it = true;
} else {
dout << "skipped adding a frame to the queue - fpts:" << send_frame->pts << "target:" << target_pts;
}
if (send_it) {
@@ -522,7 +532,6 @@ void cache_video_worker(Clip* c, long playhead) {
// see if we got the frame we needed (used for speed ups primarily)
bool found = false;
for (int i=0;i<c->queue.size();i++) {
// TODO/NOTE: this will not work on clips that are sped up AND reversed
if (c->queue.at(i)->pts >= target_pts) {
found = true;
break;
@@ -618,9 +627,13 @@ void reset_cache(Clip* c, long target_frame) {
}
break;
case MEDIA_TYPE_SEQUENCE:
c->frame->pts = 0;
c->audio_just_reset = true;
c->audio_buffer_write = 0;
c->frame_sample_index = -1;
c->audio_target_frame = target_frame;
c->frame->pts = qFloor((double) c->clip_in / sequence->audio_frequency);
memset(c->frame->data[0], 0, c->frame->linesize[0]);
c->frame->pkt_dts = qFloor((((double) (c->timeline_in - c->clip_in)/c->sequence->frame_rate) - audio_ibuffer_timecode)*sequence->audio_frequency*4);
break;
case MEDIA_TYPE_TONE:
c->reached_end = false;
@@ -872,7 +885,10 @@ void open_clip_worker(Clip* clip) {
if (av_frame_get_buffer(clip->frame, 0)) {
dout << "[ERROR] Could not allocate buffer for tone clip";
}
clip->audio_reset = true;
clip->reset_audio();
break;
case MEDIA_TYPE_SEQUENCE:
break;
}
@@ -885,7 +901,7 @@ void open_clip_worker(Clip* clip) {
dout << "[INFO] Clip opened on track" << clip->track;
}
void cache_clip_worker(Clip* clip, long playhead, bool reset, bool scrubbing, Clip* nest) {
void cache_clip_worker(Clip* clip, long playhead, bool reset, bool scrubbing, QVector<Clip*> nests) {
if (reset) {
// note: for video, playhead is in "internal clip" frames - for audio, it's the timeline playhead
reset_cache(clip, playhead);
@@ -897,12 +913,12 @@ void cache_clip_worker(Clip* clip, long playhead, bool reset, bool scrubbing, Cl
if (clip->stream->codecpar->codec_type == AVMEDIA_TYPE_VIDEO) {
cache_video_worker(clip, playhead);
} else if (clip->stream->codecpar->codec_type == AVMEDIA_TYPE_AUDIO) {
cache_audio_worker(clip, scrubbing, nest);
cache_audio_worker(clip, scrubbing, nests);
}
break;
case MEDIA_TYPE_TONE:
case MEDIA_TYPE_SEQUENCE:
cache_audio_worker(clip, scrubbing, nest);
cache_audio_worker(clip, scrubbing, nests);
break;
}
}
@@ -941,7 +957,7 @@ void Cacher::run() {
if (!caching) {
break;
} else {
cache_clip_worker(clip, playhead, reset, scrubbing, nest);
cache_clip_worker(clip, playhead, reset, scrubbing, nests);
}
}