11.0

GRAIN SAMPLE MANIPULATOR207
A few words about granular synthesis
Grain utilizes “granular synthesis” to generate sounds. This synthesis method results in playback of a series of snip-
pets of audio data - grains - “extracted” from an audio sample. The grains could be of a selectable length and spac-
ing, and could be from anywhere in the original sample. The grains could also be played back in a number different
ways - with or without crossfades between the grains.
The picture below shows the basic principle of granular synthesis:
An example of a signal generated from 5 grains of a sample.
Here is what happens in the example above:
• The original sample at the top is used as base for the granular synthesis.
• 5 grains (of the same lengths and the same distances between them) are “extracted” from the original sample.
The distance between the grains is determined by the current sample playback speed. The grains could contain
common audio data in some parts (like in the beginnings and ends in the example above).
• The 5 grains are then placed after one another, partly overlapping each other.
The distance between the grains is determined by the playback rate.
• When the grains are played back, big parts of the grains are played back together (since they are overlapping).
In the example above, there are also crossfades between the grains to make the overlaps smoother.
Note that the picture above only describes one basic example of granular synthesis - the “Long Grains” playback
algorithm in Grain. Grain uses a number of different granular synthesis and spectral synthesis techniques, with
different functionality and characteristics.
Time
Level
Time
Level
Original sample
5 “extracted” grains
The resulting signal is generated by
appending and crossfading the grains.