[jitsi-dev] [jitsi-videobridge] Focus specific video stream overriding last n (#117)


#1

I'm using last-n on the videobridge via colibri/focus controller, but I would like the ability for users to "override" the last-n setting for a single video stream. There are two specific use cases for this:

1. **Implicit override:** lastN = 2; Users A and B are speaking; User C is sharing screen. I would like User C's screen stream to always be visible to other users, regardless of the fact that Users A and B are speaking.

2. **Explicit override:** User D wants to view User B's video stream regardless of who is speaking, so User D focuses User B in the user interface, triggering some signal to always forward user B's video stream, overriding the Last N endpoints.

Are either of these possible?

···

---
Reply to this email directly or view it on GitHub:
https://github.com/jitsi/jitsi-videobridge/issues/117


#2

We have the concept of a "pinned" endpoint (which we plan to extend to a set of "pinned" endpoints in the near future) which does pretty much what your case "2" describes. It is configured with a message like this from a client to the bridge over the data channel:
{colibriClass: "PinnedEndpointChangedEvent", pinnedEndpoint: "endpoint-id"}

This can also be used to implement case "1", although not in the cleanest possible way. It seems that for a conference-wide "pin", signalling from the focus (via colibri) is more appropriate, but we don't have any of that implemented.

···

---
Reply to this email directly or view it on GitHub:
https://github.com/jitsi/jitsi-videobridge/issues/117#issuecomment-167809192


#3

@bgrozev Oh neat, that sounds promising. Is there a difference between a selected endpoint and a pinned endpoint?

···

---
Reply to this email directly or view it on GitHub:
https://github.com/jitsi/jitsi-videobridge/issues/117#issuecomment-167850110


#4

"Pinned" has to do with Last-N: a pinned endpoint's video will not be dropped, even it it isn't the active speaker.

"Selected" has to do with simulcast: video from a selected endpoint will always be forwarded in high quality.

···

---
Reply to this email directly or view it on GitHub:
https://github.com/jitsi/jitsi-videobridge/issues/117#issuecomment-167862054


#5

Perfect, thanks.

···

---
Reply to this email directly or view it on GitHub:
https://github.com/jitsi/jitsi-videobridge/issues/117#issuecomment-167863013


#6

Closed #117.

···

---
You are receiving this because you commented.
Reply to this email directly or view it on GitHub:
https://github.com/jitsi/jitsi-videobridge/issues/117#event-705263390


#7

Now included here: https://github.com/jitsi/jitsi-videobridge/blob/master/doc/last-n.md

···

---
You are receiving this because you commented.
Reply to this email directly or view it on GitHub:
https://github.com/jitsi/jitsi-videobridge/issues/117#issuecomment-228815654


#8

What Will happen if we set last n=1 ?

···

Le 29 déc. 2015 à 20:41, bgrozev <notifications@github.com> a écrit :

"Pinned" has to do with Last-N: a pinned endpoint's video will not be dropped, even it it isn't the active speaker.

"Selected" has to do with simulcast: video from a selected endpoint will always be forwarded in high quality.


Reply to this email directly or view it on GitHub.

_______________________________________________
dev mailing list
dev@jitsi.org
Unsubscribe instructions and other list options:
http://lists.jitsi.org/mailman/listinfo/dev


#9

Just what you might expect: video will be forwarded only for the current dominant speaker (and the current dominant speaker will receive the previous dominant speaker's video).

Regards,
Boris

···

On 29/12/15 23:23, Frederic wrote:

What Will happen if we set last n=1 ?