Bayesian inference for boundary admittance estimation using a multipole model for room-acoustic simulation

Abstract

Acoustic surface admittance/impedance at room boundaries is essential for wave-based room-acoustic simulations. In this work, two levels of Bayesian inference are applied to estimate the surface admittance based on a multipole admittance model. This work estimates the order of the multipole admittance model through the high level of inference, Bayesian model selection. The first (low) level of inference, Bayesian parameter estimation, is applied to estimate the parameter values of the surface admittance model once model order is selected. This work approximates the frequency-dependent admittance from experimentally measured a set of acoustic surface admittance data. Analysis results demonstrate that multipole model-based Bayesian inference is well suited in estimating the frequency-dependent boundary condition within wave-based simulation framework. Numerical simulations verify the estimation results of Bayesian inference.