Boundary conditions for hybrid simulations in a rectangular environment with sound-absorbing ceiling

Abstract

Acoustic numerical models facilitate sound field prediction in challenging real-world scenarios, such as environments with non-uniform sound absorption distribution. The accuracy of their results strongly depends on the reliability of boundary conditions required as input data. Research has largely covered analytical models of pressure-based boundary conditions for wave-based simulation techniques. However, accessible lists of frequency-dependent acoustic impedances remain limited compared to the energy-based datasets widely available in the literature. Consequently, random-incidence absorption coefficients are often converted into complex surface impedances through non-unique processes. This work explores the potential discrepancies between the input data of a wave-based finite-element model (hybridized with ray-tracing), and the energy-based coefficients employed in analytical predictions and geometrical acoustics simulations. The 3D model of an existing rectangular space with a highly sound-absorbing surface (the ceiling) is a suitable test environment for this investigation. Room criteria obtained with in-field acoustic measurements, i.e., reverberation time and speech clarity, are the experimental reference data throughout the work. Focusing on the air-backed sound-absorbing tiles at the suspended ceiling, results reveal gaps in input data, suggesting a potential percentage of discrepancies between the analytical formula and numerical models' input data up to 25% at low-mid frequencies (125 Hz - 250 Hz - 500 Hz).