All sounds are calculated in "real time", no samples are used.