Using Jitsi for a networked music performance. Stereo, server based

Hello everybody,

I work for the Milan Conservatoire of Music Giuseppe Verdi as consultant on networked music learning and performance.

I am currently having an online workshop on networked music with students from three different academies in Europe, our objective is to create music online using latency as a parameter for music making.

At the moment we are in the process of choosing the software we will use for our performance, organized on April 22nd 2020. We have NREN networks (GARR, Nordunet, 1 full gbps, symmetric) and very good hardware and other software. I am very intrested in Jitsi because open source will help people later replicate what we are doing in other universities and academies.

I have an understanding of how WebRTC and networks work, but I am still trying to figure out a couple of key aspects, listed below:

  • Is Stereo input working? I read a couple topics saying it was there but sometimes it breaks, can I have a confirmation that stereo inputs is working properly?
  • Do I need a jitsi server to enable client/server based communication?
  • If yes, is there a way to host a server on a Windows 10 based machine?
  • Does the server include STUN and TURN servers?
  • Is there a way to force OPUS audio quality to a fixed (maximum) bitrate?
  • Is there a way to disable all noise cancelling and input processing so that we have clean full range input? (I know we’re still compressing and how OPUS works, but I’ll call its maximum quality setting “full range”).

Thank you for any help you could give me and I can’t wait to try more of jitsi myself

2 Likes

What do you mean? Jitsi-meet is the web client which uses a backend. You need to either deploy it yourself, or you can directly use the meet.jit.si deployment.

Although it is possible to do it, you will spend a lot of time configuring and installing it, and this is not recommended. Nobody is actively using it like that … or at least I don’t know of such a case.

The latest unstable packages install a turn server by default. It is needed only in p2p cases when the direct connection between the 2 participants cannot be established.

Yes, there is a url param you can use to disable that and test it on meet.jit.si. https://github.com/jitsi/lib-jitsi-meet/blob/master/modules/RTC/RTCUtils.js#L79
So you can use https://meet.jit.si/mytestroom#config.disableAP=true to try it out.

3 Likes

What do you mean? Jitsi-meet is the web client which uses a backend. You need to either deploy it yourself, or you can directly use the meet.jit.si deployment.

I asked because I don’t understand wether jitsi can act as a server (like Polycom hardware does) as well or if I need another machine for that. We’ll try simply using the meet.jit.si deployment anyways, I’m mainly interested in the possibility to use the highest opus quality settings and as long as the deployment allows that I’m perfectly fine with it!

Any info about stereo audio?

Thank you very much for your help

Hi Enrico,

I tried to confirm that stereo works, but I don’t have a stereo mic to test with. You can test on meet.jit.si by appending #config.stereo=true to the URL. You’ll need to have 3 participants, as in peer-to-peer mode it definitely does not work.

You can also test using https://appr.tc/?stereo=true to make sure that your environment and browser support it.

Boris

1 Like

Hi,

This is great, just a couple of questions:

1- What would be the format of the URL if we want to disable multiple settings for instance both AP and AGC
2- All participents should use the URL as you mentioned or only the first person who is creating the meet?