I was just watching the talk on Protege at zkm 2006 and a bit of a crazy idea occured to me. In addition to storing qualities of each instrument like lead, string, percussive etc, a spectral profile could be generated for each instrument and stored in the database. Then, you could take each channel of your recording and pipe them (individually) into an app which would analyse their spectral content, and then suggest alternative patches to make the 'whole' sound like some idealised mix (there would be a selection of these, based on Genre perhaps). It would be like a magic button which would make a track sound immediately better. Any thoughts ? Bruce.
