Automatically adjust each emitter video quality to the age of her/his last speaker status

Currently, the number of users with video isn’t scalable. If you increase the number of videos, you will reach the bandwidth limit even with simulcast. If you have a huge bandwidth you will just put back this number of user limit.
Besides this problem, I’m not sure you need a high quality for each video when you have 100 video users. You actually need a better quality for the speaker and the persons who have the higher probability to get the speaker status into the coming minutes.
I propose that the JVB handles 1 queue of users ordered by latest speaker status date. When the video-bridge evaluates the bandwidth quality of each user, it will propose to each web client that it emits with the video quality which matches the age of its latest speaker status. The more recent speakers will emit video with the higher quality.
The number of users with high, middle or low quality will be chosen by the JVB in order the fit the bandwidth limitation. This evaluation could be performed each 10 minutes for example.

1 Like