DATA-DRIVEN RECOMPOSITION USING THE HIERARCHICAL DIRICHLET PROCESS HIDDEN MARKOV MODEL

Hidden Markov Models (HMMs) have been widely used in various audio analysis tasks such as speech recogni- tion and genre classification. In this paper we show how HMMs can be used tosynthesize new audio clips of unlim- ited length inspired by the temporal structure and percep- tual content of a training recording or set of such record- ings. We use Markov chain techniques similar to those that have long been used to generate symbolic data such as text and musical scores to instead generate sequences of continuous audio feature data that can then be transformed into audio using feature-based and concatenative synthesis techniques. Additionally, we explore the use of the Hier- archical Dirichlet Process HMM (HDP-HMM) for music, which sidesteps some difficulties with traditional HMMs, and extend the HDP-HMM to allow multiple song mod- els to be trained simultaneously in a way that allows the blending of different models to produce output that is a hybrid of multiple input recordings.