We have recently started developing a platform that uses Jitsi, and we’re running into some problems with recording.
So for context: this is a chess tournament platform and the videos are used for fair play purposes (i.e. we check the players aren’t looking down at their phone etc). As such, we need to record every video stream independently, rather than the ‘active speaker’ view that Jibri seems to record.
However, even if we could use gallery view, it wouldn’t be sufficient for us since Jibri is so resource intensive and we’re going to have hundreds or even thousands of games at once, so we can’t be running a chromium instance for every one.
I’m sure we’re not the first people who have wanted to do something like this, but I can’t find any docs or any relevant threads. So does anyone have a good idea of how to do this?
Even if we had to ‘join’ the meetings from another server that handles recording, that would be fine.
In that case, you might want to consider a screen recording application. As you likely know already, Jibri records one conference at a time so for mltiple conferences, you’ll need an equivalent umber of Jibri instances.
Thanks for the reply
Well my point is that we don’t want to do the ‘screen recording’ thing that Jibri does. I’m wondering if there’s a way to just subscribe to the individual video streams somehow - clearly they exist since on the frontend we can move individual videos around
Given that Jibri sometimes takes actual full gigabytes of RAM, we can’t really run one instance per conference
sure, it’s asked about one time per month as you could have found by searching the forum’s archive. The answer is always the same; this solution has been tried and found wanting, and so Jibri was born. It’s extremely unlikely that the Jitsi project will resurrect the old way.
Let me clarify: I have been through hundreds of threads on this forum and didn’t find any reasonable conclusions, everyone is just talking about Jibri and we don’t want to use Jibri because there is no way we can scale it cost effectively.
What I want is to be able to get the raw video streams in a conference and write my own software to record them, that doesn’t use render them in chromium - we need recordings of videos, not conferences
Jitsi-meet is a tool for meetings. To manage meetings efficiently and at the lowest cost, what is displayed and then saved is *interestiing data’, that is, people currently speaking. To record a meeting, it’s not useful to save the video of people scratching their noses. That’s why Jibri exists. Its aim is to save the interesting data from a meeting without wasting bandwidth and Cpu at every level.
What you want is video surveillance, it’s a different kind of use case. Jitsi-meet is not a good fit for that.