Hi, I am new to Jitsi but wondering if this exists now and if not whether your community is building anything like it. The idea is that as a permanent feature of a conference room where the screen displays people as life size, and walking up to the wall of screens is the same as walking up to the camera, not a single-camera fishtank view. Would take a lot of processing in real time to knit the live video feeds together into one big image but maybe you’re already doing that? Thanks, John!
2 Likes
after more research does look like the researchers of software algorithms for this have been awarded a patent: https://patents.google.com/patent/US20160028950A1/
would love this
what I was thinking is to be able to walk right up to the screen instead of seeing a “fishbowl” view from on POV on a large screen:
I think the stitching of live video fees is very possible using OpenCV:
and another person who has used something else back in 2016 shows a demo here:
1 Like
this is the kind of topology I think it would need, with the live video stitching/knitting done by some capable graphics processor
and I see the following software services needed to get it done, lots of this has to already exist in jitsi codebase? 