This paper and accompanying Max/MSP patch explore corpus-based concatenative synthesis (CBCS) as a compositional and sound design tool, with a focus on the relationship between audio descriptors, corpus construction, and timbral navigation.

The patch implements a real-time mosaicing system built with FluCoMa, designed to analyse a live audio input, such as an acoustic instrument or voice, and reconstruct it in real-time using fragments drawn from a user-defined corpus.