Finite-difference diffusion-equation modeling of reverberation chambers in time-domain

Abstract

Conventional chamber-based methods of random incidence absorbing coefficients overlook the non-diffused sound field in the room acoustics, which decreases accuracy and may even lead to conflicting results. This work applies the diffusion equation model to room-acoustic simulations of standardized reverberation chambers. The simulations can more efficiently capture the chamber's non-diffuse sound field and energy flows than wave-based simulation models. The diffusion equation represents the governing equation of reverberant sound energy densities within the reverberation chamber which is solved using the finite-difference time-domain method. Its computational efficiency of diffusion equation-based modeling lies in a highly sparse domain meshing condition dictated by the mean-free path length rather than wavelengths and still derives a wideband simulation result. This work also dedicates the effort to reexamine the meshing condition, particularly for standardized finite sizes of sound absorbers for measurements of random incidence absorption coefficients. By comparing the outcomes of simulations with measurement data, an a posteriori absorption coefficient is inversely estimated involving Bayesian parameter estimation.