We are trying to implement the following:
- 1 stream (at least 1920x1080) will be watched by many subscribers (possibly over 1000)
- every subscriber should be able to “open the microphone” and talk back to the host
- all the audio-streams should be mixed serverside (which can be over 1000) and send back to the host
- roundtrip latency should be around 1 second or less
As far as I understand scalability does not yet support that many subscribers. Maybe we can use a scalable media-server (like Ant) to stream the main media to all subscribers and put around 500 subscribers into one openvidu call, where the audio streams get mixed and sent to the host. Like a cascade of “submixes”.
Is it possible to achieve the “audio-mixing” part serverside with openvidu?