Granular Synthesis

Cut sound into millisecond grains and rearrange them into a cloud.

Related

TimbreEqualizerThe software synthesizerContemporary

The question

How does two seconds become a ten-minute cloud, or slow without changing pitch? Granular synthesis cuts audio into short pieces and decides when each appears and its pitch.

How it works

A grain lasts a few milliseconds to about a hundred, with a window to avoid clicks. The read position drifts through the source, and density sets how many layers overlap. At low density you hear grains; at high density they overlap into a cloud. Pitch comes from each grain's playback rate, separate from position, so time stretches independently. When grain phases disagree, high frequencies cancel and the result softens or turns metallic.

Settings and what you hear

Grain length sets roughness; short grains sound metallic, long ones keep the original tone. Density decides dotted or clouded, and spray sets how far grains stray from the centre. Slow position moves stretch time, fast ones scan. Pitch offset shifts the whole cloud, and the window shape softens or hardens edges.

What to listen for

Hear the seashore as a bed and ask whether it is continuous water or repeated fragments. Then hear the metallic pad and judge whether the top survives or is smoothed away. Finally hear the cave scene and notice fine grains floating across the field rather than a reverb tail.

Common misconceptions

That granular is a reverb, when reverb reflects and grains rearrange; that smaller grains are smoother, when too small adds sidebands; that it belongs only to experimental music, when film and pop textures often use it.

Sources

• Curtis Roads, Microsound — grain windows and clouds checked, not quoted
• Udo Zölzer, DAFX — Digital Audio Effects, granular processing chapter — background checked, not quoted

Used for fact-checking only; all prose is original.