Open pmarini-nc opened 11 months ago
The audio stream will be the one that gets to the starter. So there are situations in which a participant other than the starter is talking but there is no visible change in the recording to show that. That is, in the recording we don't know who was speaking.
That should only happen when Janus 1.0 is used instead of 0.13 (which is recommended).
The current intended behaviour is that always the video of the "speaker" is shown. Not the starter. The starter is only considered special when uploading the recording to put it into their directory.
Thanks for the feedback.
In my tests I'm using Janus 1.2.0. Is there some configuration switch to be enabled on nextcloud-talk-recording side?
And in my case it is not the speaker that is shown, but always the starter. Only when the speaker shares the screen, his/her screen is shown in the recording, but in any case not his / her video.
In my tests I'm using Janus 1.2.0
That version is currently not supported. The "Is speaking" information is transmitted via data channels and Janus renamed the default data channel iirc. So you don't receive the "is speaking" information and that is exactly the problem you are facing.
Please downgrade to Janus 0.13.x for now
Thanks. I have verified that with 0.13.4 "the video goes with the audio".
It would be great to have a comment on the feasibility of this too:
webinars or seminars live streaming: starter point of view, with only the speaker(s) screen(s) as main screen area and the speaker(s) camera(s) shown aside, below or above (or not shown at all if there are privacy concerns, but the option should be given)
As it is now, if we define the participant who starts the recording as "the starter", the recording will be showing:
The audio stream will be the one that gets to the starter. So there are situations in which a participant other than the starter is talking but there is no visible change in the recording to show that. That is, in the recording we don't know who was speaking.
In short, starter point of view is as expected for the audio, but not for the video.
In my opinion, the expected video recording layout depends on the use case:
private calls, group or 1-to-1: starter point of view, in speaker mode (or grid mode, maybe configurable?)
webinars or seminars live streaming: starter point of view, with only the speaker(s) screen(s) as main screen area and the speaker(s) camera(s) shown aside, below or above (or not shown at all if there are privacy concerns, but the option should be given)
I think this is quite common in other session recording streaming.