Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

Rework connect_pad_added closure #19

Closed
fengalin opened this issue Aug 15, 2017 · 4 comments
Closed

Rework connect_pad_added closure #19

fengalin opened this issue Aug 15, 2017 · 4 comments
Assignees
@fengalin
Labels
enhancement

Comments

@fengalin
Copy link
Owner

There are several possible enhancements:

  • Audio processing: a tee can be added to split the audio processing in two queues (which also implies 2 threads). One will handle audio playback just as right now, the other will be in charge of pre-processing the samples to be ready for waveform rendering.
  • Video sink construction: it seems that the way the sink is constructed right now is not really thread safe and may not work on other platforms. See this discussion.
  • Caps might change during playback, so this should be monitored. See this discussion.
@fengalin fengalin self-assigned this Aug 15, 2017
@fengalin
Copy link
Owner Author

Audio processing: see this tutorial and this conversation.

@fengalin
Copy link
Owner Author

Configure buffering to make sure we always have more than 1s worse of samples ahead.

@fengalin
Copy link
Owner Author

For the video widget potential issue, maybe send-cell would be applicable to guarantee thread safety.

fengalin added a commit that referenced this issue Aug 18, 2017
@fengalin
Audio: use a tee and an appsink (#19)
35f4666
@fengalin
Copy link
Owner Author

Possible enhancements in audio buffers management. See this comment. The only problem is the buffer drain that requires knowing samples/pixel.

fengalin added a commit that referenced this issue Aug 19, 2017
@fengalin
Fill the normalized sample AudioBuffer in the appsink callbacks (#19)
5e37e85 
This allows taking advantage of the viualization queue thread to perform all the required computation as soon as the samples are received. There is no need to send the gst::Buffers through the channel: the AudioBuffer is now shared between the AudioController and the appsink.

Using a similar approach, the video widget box is now passed at Context creation. The video sink is constructed before the pipeline creation which ensures it is created in the GTK thread and which allows adding the widget to the video widget box without using channels.

Since channels are no longer required for realtime communication, the listener can be invoked less frequently.

This also fixes:
- Handling of media with multiple audio or video streams.
- Waveform flickering with some files (#18)
fengalin added a commit that referenced this issue Aug 19, 2017
@fengalin
README: remove waveform rendering adaptation mention (#19)
edd028d 
No longer needed as there is enough data buffered to guarantee rendering on time.
fengalin added a commit that referenced this issue Aug 19, 2017
@fengalin
Remove magic constant from AudioBuffer drain (#19)
3d0a77a
fengalin added a commit that referenced this issue Aug 19, 2017
@fengalin
Better handling for AudioBuffer duration smaller than window duration (…
5b049f7 
…#19)
fengalin added a commit that referenced this issue Aug 19, 2017
@fengalin
Update dependencies (#19)
f63dfe6
fengalin added a commit that referenced this issue Aug 20, 2017
@fengalin
Read position upon waveform renderding (#19)
a7f5af9 
The waveform rendering used to rely on a position that was read before in the main controller. In order to limit the drifts between the waveform rendered and the actual position, the position is now read during renderding. This requires the Context to be visible in AudioController::draw. So the Context is now a Weak Ref of AudioController.
fengalin added a commit that referenced this issue Aug 20, 2017
@fengalin
Code simplifications (#19)
1c86195
fengalin added a commit that referenced this issue Aug 22, 2017
@fengalin
Double buffering for waveform rendering (#19)
5b7fba4 
This modifications introduce a double buffering mechanism in order to smoothen the waveform rendering. However, another cause of locks was found: context.get_position()
fengalin added a commit that referenced this issue Aug 28, 2017
@fengalin
Replace context.get_position with a clock.get_time (#19)
1497066 
It doesn't solve the stuttering in waveform rendering, but seems to reduce latency getting the position
fengalin added a commit that referenced this issue Aug 29, 2017
@fengalin
Waveform buffer: incremental updates (#19)
16813d4
fengalin added a commit that referenced this issue Aug 30, 2017
@fengalin
Register / unregister tracker when switch playing state (#19)
f1cbae4
fengalin added a commit that referenced this issue Aug 31, 2017
@fengalin
Fix rendering for positions highers than half window and remove buffe…
2dfe924 
…r offset (#19)
@fengalin fengalin closed this as completed Sep 1, 2017
@fengalin fengalin reopened this Sep 1, 2017
fengalin added a commit that referenced this issue Sep 1, 2017
@fengalin
Fix rendering when end of stream has been reached (#19)
5145760
fengalin added a commit that referenced this issue Sep 1, 2017
@fengalin
Another fix for rendering in specific conditions (#19)
587e310
fengalin added a commit that referenced this issue Sep 2, 2017
@fengalin
Waveform: use an iterator and draw the visible part only (#19)
caedcf4
fengalin added a commit that referenced this issue Sep 2, 2017
@fengalin
Restore AudioBuffer draining (#19)
98eca08
@fengalin fengalin closed this as completed Sep 2, 2017
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Assignees

@fengalin fengalin

Labels
enhancement
Projects
None yet
Milestone
No milestone
Development

No branches or pull requests
1 participant
@fengalin