Only a GOD could answer this question in detail and not theory but that is why I am posting this question because its never truly been solved. What is the best cpu friendly way to decompose any audio signal into individual signals for each instrument regardless of polyphony and subharmonic instruments like cymbals and drums? Their have been a dozen attempts and approaches but most that I have read are so blasphamus I can tell the writers have never actually tested them. If you have any thoughts on this question but more importantly if you would like to share or look at some of my source code and theories on the subject than post. I only ask because as we speak I intend to solve this program and design software that does just what I have been talking about. If you know anything about digital audio and especially dsp it is your duty to respond. I am serious about this and my determination to solve this problem would make Dr. Frankestein pull his hair out.

Please don't crosspost on these forums.