What if Zoom had spatial audio?

This seems very interesting, and definitely something that I could see getting integrated into a product like Zoom eventually.

Is the Spatial Audio function fast-enough to be done client-side, so it can deal with the way different clients might be rendering the talking heads in different orders? If I make my client really wide, so that all heads are in a line, it should sound different than if all the heads are stacked vertically. (Also, people on "page 2" of a big meeting could sound like they are further away -- indeed, that latter feature is something I've always thought should exist for a teacher to be able to listen to the soft hubub of breakout rooms).

If the transfer function is fast, I could also see it being eventually paired with the camera that is already watching your face, so that it could tweak it subtly as you turn your head. I've noticed that people already naturally cock their heads towards someone they're straining to hear, even though they are wearing headphones so it makes no difference (except as a subtle hint to the person talking).

I was just speculating a week or so ago that it might be much better if they would use stereo output for multiple speakers. That seems to be the thing that we've had the most problem with: two people start talking and the combined audio is harder to understand over videoconferencing than in person. But stereo output is pretty easy and nearly universally supported (one co-worker has a little puck-mic/speaker that's designed for conferencing and works pretty well, but has no stereo separation, for example).

This seems to take it to the next level.

I remember sometimes making stereo phone calls in the 1980s with two phones.

FYSA, High Fidelity is a virtual world startup / standards project created by the Philip Rosedale, the guy that founded Linden Labs (makers of Second Life).