[jitsi-dev] [jitsi-videobridge] Payload-type consistency enforced across endpoints? (#133)


#1

Should a specific payload-type be enforced for the conference and all associated endpoints since JVB doesn't transcode?

···

---
Reply to this email directly or view it on GitHub:
https://github.com/jitsi/jitsi-videobridge/issues/133


#2

We don't want to explicitly make the assumption that payload-type mappings are the same for all endpoints, and try to work without it whenever possible, but in our environment (jitsi-meet + jicofo) it holds. So problems might creep in if you don't enforce the same PT mappings.

···

---
Reply to this email directly or view it on GitHub:
https://github.com/jitsi/jitsi-videobridge/issues/133#issuecomment-182963547


#3

ditto what @bgrozev said, we unfortunately have a few constants here and there that assume 116 for RED, 100 for VP8.

···

---
Reply to this email directly or view it on GitHub:
https://github.com/jitsi/jitsi-videobridge/issues/133#issuecomment-182965152


#4

JVB *does* support transcoding in mixer mode. Feel free to put in more checks for the translator mode, but please don't break the transcoding in mixer mode.

···

---
Reply to this email directly or view it on GitHub:
https://github.com/jitsi/jitsi-videobridge/issues/133#issuecomment-183218559


#5

Right, since JVB is first and foremost an SFU, clients must be vetted for codec compatibility before endpoint creation as far as I understand. e.g. Chrome/Firefox SDP would need to be checked

···

---
Reply to this email directly or view it on GitHub:
https://github.com/jitsi/jitsi-videobridge/issues/133#issuecomment-183515255