This module takes the original selection and creates new channels with the granulated sounds. Grain lengths between 0.01 and 0.06 seconds are best, and a variance of 0.002 helps to remove particular spectral characteristics.
The density parameter refers to the percentage of grains that are copied across, so that you could have a sound 'appearing' or 'disappearing' (although in practice I've never managed to get this sounding convincing).
