Unfortunately, there doesn't seem to be a good way for two geographically-separated people to sing together in real-time. As far as I can tell, Zoom and probably most others simply try to identify the person speaking, and mute everybody else. The problem is the delay -- Bob can sing along with Alice, but what everyone else -- including Alice! -- is going to hear is Alice with Bob singing along several milliseconds later. It's worse if Bob is listening to Alice over speakers, because then people are going to hear Alice and delayed Alice mixed together, which leads to some weird interference effects. It's called flanging, if it's done deliberately. You sometimes hear flanging with just one person singing with a guitar; I've heard noise reduction blamed for it, but my personal guess is that it's at least partly the result of using speakers rather than headphones.

There _are_ a few programs that purport to let people sing together, but I've seen mixed reviews, and they're unlikely to work at all if people aren't located pretty close to one another.

There are a couple of ways to make it _sound_ like people are singing together, but that usually involves singing against a click track, singing with a conductor (both of which require recording everyone and mixing the result later -- I don't think there's any software out there that lets you mix multiple video streams in realtime) or adding layers -- Alice sings her part, Bob sings along and uses a second computer to mix his part with Alice's, and so on. So for N people you need 2N-1 computers.

Is that making sense? It got a little long.

There are two