so this model can only ever understand music which is classified, described, labelled, standardized. and recombine those. sounds boring, sounds like the opposite of what (I would like to believe) people listen to music for, outside of a corporate stock audio context.