On the simulation of seat-dip effect using geometrical acoustics software

Abstract

A group of Italian churches was surveyed in order to measure the most important acoustic parameters according to ISO 3382 Standard. Computer models of the same churches were made using acoustic simulation software. Absorption coefficients found in the literature were used and later calibrated to match predicted and measured T30 values. The results of the simulations were compared with the observed values, showing some discrepancies at mid-low frequencies. This discrepancy appeared to be due to a lack of direct sound in the measured responses, particularly at the 250-Hz and 500-Hz octave bands, indicating the probable presence of a seat-dip effect caused by the wooden pews. Since the acoustic simulation software provided the possibility to use acoustically semitransparent planes, this feature was used to simulate the effect of selective absorption of the direct sound due to the seat-dip effect. The comparison between measured acoustic parameters and those predicted including the simulation of the seat-dip effect showed that an improvement in the prediction accuracy can be achieved. Different configurations were tested in order to define the optimal placing of the semitransparent plane, and a criterion to choose the transparency coefficients is finally proposed.