good points, all, lawrie but we can't limit our examination to a 2 dimensional waveform on an o-scope. if we considered each characteristic as a vector and analyze it in 3 dimensions (assuming that is made possible) then we would be able to extract a lot more information from the waveform, orders of magnitude more. perhaps the o-scope analogy is wrong. we need a multi-dimensional means of measuring perhaps 2 or 3 dozen factors simultaneously then running the results through a computer program capable of isolating the instruments. a friend, roger cota, designed a hand-held lie detector than detected vocal micro-tremors to indicate when the speaker was lying. in effect, he was using a a little known aspect of sound that had escaped other engineers' attention. that is the sort of thing i am alluding to. another circuit that fascinated me is the 'timbre gate' found in synthesizers. what if it were reversed and thereby used to identify various instruments rather than to clone them?
just some random thoughts.

Donny

SoundClick: http://www.soundclick.com/bands/default.cfm?bandID=1209254&content=music
SoundCloud: https://soundcloud.com/donald-gaynor
BIAB 2021 w/current updates.