Nested sampling for room-acoustics energy decay analysis.

Abstract

Room-acoustic energy decays often exhibit single-rate or multiple-rate behavior in a wide variety of enclosures. There has been an ever increasing need to investigate where and under what conditions the single-rate and or multiple-rate energy decays occur and to quantify/characterize the energy decay process by estimating a set of decay parameters. In this architectural acoustics application, both energy decay model selection and decay parameter estimation are of practical significance. This paper discusses a model-based sound energy decay analysis within a Bayesian framework and applies the most recent Bayesian sampling method (nested sampling) to sound energy decay analysis. The nested sampling reverses the historical Bayesian computation approaches, yielding the Bayesian evidence as the prime target with representative posterior samples available as optional by-products. Bayesian evidence is crucial for model selection, while the representative posterior samples are of central importance for parameter estimation. Taking the energy decay analysis in architectural acoustics as an example, this paper demonstrates that two different levels of inference, decay model-selection and decay parameter estimation, can be cohesively accomplished by nested sampling.