Abstract:

This presentation will describe the state of the art in model-based approaches to audio restoration, with an emphasis on results and an explanation of theory suitable for the nonspecialist. An overview of work conducted over the past several years by members of the Cambridge Signal Processing Group will be presented, including techniques for the detection, correction, and interpolation of degraded or missing audio data. The latest advances in a fully Bayesian statistical methodology for modeling of music and speech signals will also be detailed, including a common unifying framework for the treatment of both global and localized audio degradations.