Constrained least-squares and maximum-likelihood calibration of absorption coefficients in reverberation time equations

Abstract

Calibration of material properties in room acoustics simulations is essential to producing realistic results. However, the number of available algorithms for calibrating room acoustics models remain limited. This work presents two methods for calibrating absorption coefficient values of reverberation time equations. The first uses a constrained least-squares estimate, while the second uses a constrained maximum-likelihood estimate. In addition to matching measured and estimated reverberation times, stochastic simulations demonstrate that the proposed methods decrease deviations in estimated absorption coefficient values relative to the initial estimate. A case study application of the calibration method to a scattering-delay-network reverberator and geometrical acoustics models suggests the portability of the approach in calibrating more complex room acoustic numerical simulations.

Keywords

• room acoustics
• geometric acoustics
• model calibration
• absorption coefficient
• reverberation time