Hi René! Thanks for the words of encouragement and your continued support!
For questions like this, the best medium at the moment is github discussions. For your application, `input.harbor` is not what I would recommend. It uses a TCP connection and requires buffering so you will never be able to get below a certain threshold for latency. The best I would suggest for latency is: * Use a udp-based input. `input.srt` is probably the best for us. * Reduce the frame size. In `2.2.x` you can cheat a little bit by setting the video rate to 0 and then the frame size can be whatever you want. See for instance: https://www.liquidsoap.info/doc-2.2.4/cookbook.html#alsa-unbuffered-output In `2.3.x`, the current `main` this is now a plain feature but this is also the experimental version. If you wanted to try with this one, though, I would personally help fixing whatever bug you encounter. Lastly, you would also need a suitable output. Again, `output.srt` would be great but it's not for general public consumption so you would then need another broadcast server to receive it and make it available to your listeners. If you give more details we can see about this further. Also, happy to move this to github discussion where you might be more feedback from other users. -- Romain Le lun. 22 avr. 2024 à 09:10, René Jaun <lis...@jaunscorner.com> a écrit : > Hello, > We've been using Liquidsoap for more than six years now. It's very > valuable for us. So thanks for doing such a great job. > Currently, I am trying to set up a new system for our realtime broadcasts > of soccer games. > We would like to be able to provide an audio stream for people listening > inside stadiums. This stream should of course have no or very little > latency (ideally under 0.5 seconds). > I read somewhere that Liquidsoap may not be suited for this, having > somebuffers built-in which can't be changed. > At the same time, I'd really like to keep using Liquidsoap to capture the > audio from our hosts at the stadium, as it really works well with their > broadcasting app, Backpack Studio. > Thus my question is something like: Is there a way to have Liquidsoap > relay / output the audio signal it gets via input.harbor without any > buffering / delay to another program which could serve the realtime-output > to our audience in the stadium? (I am looking at a Janus-based option for > this). > This should happen in addition to our classical Icecast Output method, > which would provide our complete program with more latency. delay. > > Thanks for any help! > Best regards > René > > > > _______________________________________________ > Savonet-users mailing list > Savonet-users@lists.sourceforge.net > https://lists.sourceforge.net/lists/listinfo/savonet-users >
_______________________________________________ Savonet-users mailing list Savonet-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/savonet-users
