Imagine being able to make music just by making faces? This interface (hah!) allows a 'synthesizer' (not really - more a pre- prepared jukebox of riffs, beats and licks) to be able to respond as if you were composing by using your face. Certain common actions would map to certain common musical dynamics, such as transposition, sequence shifting, and so on. Smaller granularities of facial variation would map to smaller variations of musical expression, such as playing an actual melody, making it loud, soft, etc. Intermediate granularities would affect style and patch selection, and the bigger granularities would affect total arrangement. That sort of stuff.