Evaluation of Reverberation Time Models with Variable Acoustics

Karolina Prawda, Sebastian J. Schlecht, Vesa Välimäki

Companion page for a paper in the The 17th Sound and Music Computing Conference, 24-26 June, 2020

Abstract

Reverberation time of a room is the most prominent parameter considered when designing the acoustics of physical spaces. Techniques for predicting reverberation of enclosed spaces started emerging over one hundred years ago. Since then, several formulas to estimate the reverberation time in different room types were proposed. Although validations of those models were conducted in the past, they lack testing in a space with a high granularity of controllable absorptive and reflective conditions. The present study discusses the reverberation time estimation techniques by comparing various formulas. Moreover, the reverberation time measurements in a variable acoustic laboratory for different combinations of reflective and absorptive panels are shown. The values calculated with presented models are compared with the ones obtained via measurements. The results show that all formulas predict reverberation time values inaccurately, with an average error of 16% or larger. Among the analyzed models, Fitzroy's formula performs the best, giving the smallest error.

Room layout and variable acoustic panels

Layout of the variable acoustics laboratory Arni.

View and cross section of the variable acoustic panels.

Impulse response examples

All panels open (source 1, microphone 1)

All panels closed (source 1, microphone 3)

All panels on the ceiling closed (source 2, microphone 5)

First wall closed (source 2, microphone 4)

First and second wall closed (source 1, microphone 2)

First, second and third wall closed (source 2, microphone 1)

All walls closed, ceiling open (source 2, microphone 5)

Results of calculations with noise added to the absorption coefficient values.
The case of all panels closing one by one

The case of panels on the ceiling closing one by one

---