One small sound, several of which may make up a spoken word. Used to analyse voice input to recognise words or to produce speech by playing back a sequence of phonemes