UNCERTAINTY OF INPUT DATA FOR WAVE-BASED ROOM ACOUSTIC SIMULATIONS IN LARGE NON-TRIVIAL ENVIRONMENTS

In the last decades, wave-based simulation methods have been applied to an increasing number of 3D virtual rooms thanks to the scientific and computational advances in numerical models. However, there is still a lack of adequate material properties required for those simulations, in terms of accessible lists of various frequency-dependent boundary conditions. Such input parameters can be retrieved from sound absorption coefficients, exploiting the availability of several consolidated datasets typically employed in ray-tracing simulations. The present work aims at quantifying and assessing the degree of uncertainty underlying this critical step in non-trivial environments. With this purpose, parallel calibrations have been carried out on distinct case studies based on experimental data using finite-difference time-domain (FDTD) and geometrical acoustics (GA) approaches. The outcomes highlight significant discrepancies at low frequencies between the different input data for various materials, suggesting a potential decrease (up to 45%) in sound absorption coefficient before the conversion to specific acoustic impedances.