You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
The live-streaming latency, even when in low-latency mode, leaves a little to be desired when compared to proprietary solutions
A project I'm heading is working on a possible solution in our fork, and wanted to gauge if this should be upstreamed once complete
Describe the solution you would like
Implement a dedicated low latency media loader using VHS (as P2P is disabled in low latency mode anyways)
This would also open the door for future improvements, such as implementing LL-HLS once support is available
Utilize the latency compensation technique from owncast to keep the player as close as possible to the live edge
Add a "live segment transmission delay" admin configuration option (to handle cases where a CDN has been configured), and use that to set the targets used in compensation
As this may be non-conformant (iirc), pushing closer than the HLS spec minimum buffer could be opt-in for the user (as is done on twitch)
The text was updated successfully, but these errors were encountered:
I think it would be simpler to wait LL-HLS support in hls.js https://github.com/video-dev/hls.js/projects/7
Unfortunately (fortunately?) we're happy with hls.js and don't want to switch on VHS
Understood! Nothing should theoretically prevent hls.js from doing the same (it even has a similar latency compensation mechanism via maxLiveSyncPlaybackRate), however we were never able to get it to function properly with liveSyncDurationCount under 2, while VHS has little problem with this.
If we can find out how to get this working in hls.js, we'll switch over to that to open a path to upstream
Describe the problem to be solved
The live-streaming latency, even when in low-latency mode, leaves a little to be desired when compared to proprietary solutions
A project I'm heading is working on a possible solution in our fork, and wanted to gauge if this should be upstreamed once complete
Describe the solution you would like
The text was updated successfully, but these errors were encountered: