We see several types of interactions already:
- ornamenting: expands or embroiders the other sound, creating features/events (in time) that was not there before
- transplanting: transfers its own timbral character to the other sound
- inhibiting: when one sound plays, the other is damped/muted (via amplitude, eq, spectrum, other), this is also used in traditional sidechainingto make room for a sound in the mix.
- enhancing: the opposite of inhibiting, e.g. the other sound is only hear when the first one plays.
Some types (e.g. inhibiting and enhancing) will be hard to apply to a live performance due to acoustic sound bleed. It is probable that it will also be hard to apply for any kind of realtime interactive situation with instruments that have a significant acoustic sound level. The other sound source can be isolated, but my own sound will always be heard directly (e.g. for a singer or a wind instrument), so the other musician can not mute me in a way that I will experience in the same way as the audience (or other player) will