Choosing audio channels for translation

For webinars or such, it would be very helpful to be able to choose what channel to listen to. You can currently mute people, but you have no choice with a single speaker. What if a speaker can have multiple channels?

Describe the solution you'd like
Below a speaker avatar you can choose a sound channel, only when there are multiple channels of course.
Speakers must be able to send multiple streams - that probably makes most sense again for a webinar, where you have a single speaker.

Describe alternatives you've considered
You can have multiple speakers now, each sending a stream, and rely on users to mute the ones they don't want to hear. But the user experience of that is sub-optimal: when you join you hear them all, it's a mess! Plus, which one should be shown big?

Start by tracing how audio streams and speaker controls are handled in the project, then inspect the UI around each speaker avatar. Define how multiple streams from one speaker are represented, how a listener selects one, and how the default selection behaves; the feature is done when users can choose among a speaker’s channels without hearing all streams by default.