AJ, my own experiments with DBTs have shown me that either the testing apparatus is not sufficient competent, I'm thinking the Foobar extension here, or that the use with music content is doomed to give poor results - an example of the latter instance is when I did experiments with reversing dynamic decompression some time ago, I ended up horribly confused as to which version was the 'best', audible solution. And I note that other people have similar difficulties.
DBTs work well with sound snippets with no meaning beyond the set of frequencies used for the test - once they move into the realm of time meaningful content, like music, then all bets are off. That's what this interesting field of ASA makes very clear - the mind quite happily manufactures the content that's not really there, so that the model of what the brain expects to hear next, what's called a schema, is satisfied. In simple terms, the auditory system we all have is constantly deluding us, all the time - we never actually 'hear' the sound waves that are impinging, these are always rearranged internally to suit, that give the best fit. Which makes DBT on music problematic, at the very least ...
