Personalized audio systems - a Bayesian approach

Nielsen, J. B., Sand Jensen, B. , Hansen, T. J. and Larsen, J. (2013) Personalized audio systems - a Bayesian approach. In: 135th International AES Convention, New York City, NY, USA, 17-20 Oct 2013,

Full text not currently available from Enlighten.

Publisher's URL:


Modern audio systems are typically equipped with several user-adjustable parameters unfamiliar to most users listening to the system. To obtain the best possible setting, the user is forced into multi-parameter optimization with respect to the users's own objective and preference. To address this, the present paper presents a general inter-active framework for personalization of such audio systems. The framework builds on Bayesian Gaussian process regression in which a model of the users's objective function is updated sequentially. The parameter setting to be evaluated in a given trial is selected by model-based sequential experimental design. A Gaussian process model is proposed which incorporates correlation among particular parameters providing better modeling capabilities compared to a standard model. A ve-band equalizer is considered for demonstration purposes, in which the parameters are optimized using the proposed framework. Twelve test subjects obtain a personalized setting with the framework, and these settings are signicantly preferred to those obtained with random experimentation.

Item Type:Conference Proceedings
Glasgow Author(s) Enlighten ID:Jensen, Dr Bjorn
Authors: Nielsen, J. B., Sand Jensen, B., Hansen, T. J., and Larsen, J.
College/School:College of Science and Engineering > School of Computing Science
Related URLs:

University Staff: Request a correction | Enlighten Editors: Update this record