Effect of reference value and ensemble size on perception of reverberation in simulated listening environments

Abstract

Reverberation time is an important metric characterizing one of the key differences between performance venues caused by various furnishing, materials, and geometry within such spaces. To justify alterations to a venue, architects and acousticians refer to the just-noticeable-difference (JND) of reverberation time, cited in the ISO 3382-1 room acoustics standard as 5% of the reference value. This JND was determined experimentally using impulsive noise with varying reverberation. In real performance venues, overlapping and simultaneous sounds are more common than isolated impulses. As such, the JND noted in literature may not provide a complete picture of perceived reverberation for persons listening to live performed music. For this study, multiple test methods were evaluated by using different reference reverberation times as well as using a variety of instrumentation and genres for musical signals. Results were generated using test participants who were asked to listen to multiple virtual soundfields and provide subjective evaluation. Reverberant soundfields were recreated in an anechoic listening environment and auralizations were produced using ODEON room acoustics software. Results indicated statistically significant mean differences in perceived reverberation across instrumentation subsets, genre subsets, as well as through different testing configurations. [Work supported by The Paul S. Veneklasen Research Foundation.]