|Abstract||A system is introduced that learns the structure of an audio recording of a rhythmical percussion fragment in an unsupervised manner and synthesizes musical variations from it. The procedure consists of 1) segmentation, 2) symbolization (feature extraction, clustering, sequence structure analysis, temporal alignment), and 3) synthesis. The symbolization step yields a sequence of event classes. Simultaneously, representations are maintained that cluster the events into few or many classes. Moreover, a tempo estimation procedure is used to preserve the metrical structure in the generated sequence. Employing variable length Markov chains, the final synthesis is performed recombining the audio material derived from the sample itself. In particular, the level of refinement of the clustering procedure is selected, choosing a representation that displays maximal regularity. Examples synthesized from percussion patterns such as the amen break and beat boxing are available on the web. For a broad variety of musical styles the musical characteristics of the original are preserved. At the same time, considerable variability is introduced.