Auralization of concert hall acoustics using finite difference time domain methods and wave field synthesis

Abstract

A finite difference time domain (FDTD) simulation on a three-dimensional face-centered cubic grid, combined at a crossover frequency of 800 Hz with a CATT-AcousticTM simulation, was found to have a reverberation time, direct to reverberant sound energy ratio, and early reflection pattern that more closely matched measured data from the hall compared to a CATT-AcousticTM simulation and other hybrid simulations. In the CRAIVE lab, nine experienced listeners found all hybrid auralizations (with varying source location, grid resolution, crossover frequency, and number of loudspeakers) to be more perceptually plausible than the CATT-AcousticTM auralization. The FDTD simulation required two days to compute, while the CATT-AcousticTM simulation required three separate TUCTTM computations, each taking four hours, to accommodate the large number of receivers. Given the perceptual advantages realized with WFS for auralization of a large, inhomogeneous sound field, it is recommended that hybrid simulations be used in the future to achieve more accurate and plausible auralizations. Predictions are made for a parallelized version of the simulation code that could achieve such auralizations in less than one hour, making the tool practical for everyday application.

Auralization methods have been used for a long time to simulate the acoustics of a concert hall for different seat positions. The goal of this thesis was to apply the concept of auralization to a larger audience area that the listener could walk through to compare differences in acoustics for a wide range of seat positions. For this purpose, the acoustics of Rensselaer’s Experimental Media and Performing Arts Center (EMPAC) Concert Hall were simulated to create signals for a 136 channel wave field synthesis (WFS) system located at Rensselaer’s Collaborative Research Augmented Immersive Virtual Environment (CRAIVE) Laboratory. By allowing multiple people to dynamically experience the concert hall’s acoustics at the same time, this research gained perspective on what is important for achieving objective accuracy and subjective plausibility in an auralization.