Probably a bad idea to congest the limited bandwidth of Tor with voice chat.
Plain speech can be compressed pretty well. I’m not an expert by any means, but I suspect latency would be the bigger issue.

Latency is a huge issue, but it goes away with the PTT model. I tried full duplex on initial prototyping but it was trash.

PTT forces solves this by simply forcing the listen, digest, then respond. You can expect about 2-3 seconds of delay from when you release the ptt, to when the other side hears it.

That makes sense. Thanks, I was still half asleep and didn’t register the PTT.