Sound energy decay analysis in multiple coupled volume systems.

Abstract

There has been an increasing need in analyzing sound energy decays consisting of more than one decay slope, so‐called non‐single‐exponential decays. It has been considered very challenging to estimate parameters associated with double‐slope decay characteristics, even more challenging when the coupled‐volume systems contain more than two decay processes. To meet the need of characterizing energy decays of multiple decay processes, this work reports investigations using both acoustical scale‐models and numerical models of three coupled volumes. Two levels of Bayesian inference: model selection and parameter estimation, have been applied to detection and characterization of sound energy decays. Experimental results measured in acoustical scale models and numerical results from diffusion‐equation‐based models are used to validate the Bayesian analysis methods. The analysis methods are then applied to geometric‐acoustics simulations of a conceptual concert hall. This work demonstrates that the analysis method within the Bayesian framework is capable of determining more than two decay slopes and estimating the corresponding decay parameters and parameter uncertainties.