Extend Jitsi to Wowza, increase room seat no

Hi,

jvb2 was supposed to remove the barrier of 35 members in a room and a hard limitation of 75. How is it now? Where are these things defined, please?

Also, we have a very powerful server with Wowza Streaming Engine running on it just fine for online television. We needed a chat feature, too, with 2 way audio video features and not really in the business of reinventing the wheel. Jitsi seamed a very good candidate but requires some serious adjustments to support large amount of people in a room (we have 2k+ concurrent viewer peaks):

  • room global settings => mic, camera, chat, share screen etc all disabled till granted to the room or individuals
  • different member organizer - to easy manage large no of people
  • link Jitsi to Wowza -> both broadcast to wowza and playback from wowza - or any other video server.

If anyone can point us to the right directions, as I said, not reinventing the wheel, maybe you guys already made some adjustments like this, please let me know.

Best!

2 Likes

Hi. This would be something I would like to implement too. If there is any direction on this I would be happy to test a solution

I’m trying to implement this as well. I’m messed around with the webRTC streaming/ingest features of the latest Wowza server, but I’m not seeing an obvious way to link the two systems together.

I’m very new to WebRTC in general though… does anybody know if it’s possible to include Wowza as a hidden meeting participant or something?

It’s possible to stream to Wowza using Jibri. There are many posts in the forum about how to stream a custom RTMP server.

I was initially hoping to use Wowza’s WebRTC support to bypass Jibri and potentially dozens of VMs…one for each concurrent meeting. However, now that I’ve read up more on WebRTC and spent hours looking for browserless solutions, I see the issue.

The WebRTC component of Jitsi is just an unorganized collection of video/audio streams. The Jitsi Meet web app is responsible for organizing them into a coherent presentation, with participant name labels, highlighting current speaker, etc. Jibri needs to actually run the web app in Chrome to do all of that, then use ffmpeg to screen capture and save to a nice mp4 video.

So, theoretically, one could bypass Chrome and use use something like node-webrtc to capture the individual webrtc streams, but then you’d need to basically re-implement all of jitsi’s presentation logic, format it onto node-canvas, and then wind up encoding the view with ffmpeg anyway…it might be more efficient than screen-grabbing Chrome in the long run…but it’s a long way to go to find out out.

Sorry if this is all obvious to most of you here, but I figured I’d share for the other WebRTC noobs looking for a “better” solution.