You might be interested in Calla, which provides exactly that kind of spatialized audio for meeting participants (via avatars) on top of Jitsi conferencing:
No probs! If I remember correctly, I heard about it on HN around the start of COVID-19 restriction time in this great side-projects thread: https://news.ycombinator.com/item?id=23170881 . Glad to see it gaining more attention.
https://github.com/capnmidnight/Calla