Is it possible to generate a musical instrument's sounds using only algorithms? or can it only be done via pre-recorded sound samples?