I'd guess every app that uses real-time live audio does this. It's far simpler to constantly encode and selectively transmit than it is to start the encoding e

I'd guess every app that uses real-time live audio does this. It's far simpler to constantly encode and selectively transmit than it is to start the encoding each time you unmute. It would probably also be missing the first few hundred milliseconds (and if it was buffered, it would be out of sync).