I think a lot of the complaints being voiced can be worked around with enough thought.
1. Mic spamming: most other places that offer a voice feature also offer the ability to mute on a player-to-player or global basis as well. I don't see why this would need to be any different.
2. Song muting: there would only be a reason to mute the song for the person speaking; their voice could be added over the song (at a selectable volume) in the lobby or during breaks like the drum and cymbal sound effects found on the beatmap itself-- no need to stop the music unless YOU feel the need to talk (presumably at an opportune moment).
3. THERE'S NOTHING TO SAY!: Well yeah, if you're voice chatting DURING the song you're probably doing something wrong, but it would be nice in the lobby where a lot of conversation happens anyway, and sometimes we DO want to gripe about that last spinner during breaks but don't have time to type and hide the chat window again.
4.Not many people have a mic: I don't know about that, they're common enough playing other games. Maybe some people would even go out and pick up a cheap ten dollar headset if they could use it here.
5. It'll slow down the server: I said in the original post that most of this could (I think, at least) be handled on the player end to leave Bancho out of the chat thing entirely. We could even have a dedicated player, an "operator" if you would, with a fast connection that handles the voice chat separate from the ever-changing host. And of course if you don't have a mic or you're running Australian dial-up you can always just disable the whole thing and be done with it-- and if you want to talk there's always text; no need to ruin everyone's fun because you can't talk.