Perceived suitability of reverberation in large coupled volume concert halls.

Abstract

Large acoustically coupled volumes can be found in a variety of places such as churches, factories, train stations, and concert halls, where sound reverberation occurs. In recent decades, coupled volume concert halls have garnered increasing attention (Beranek, 2004) as the use of reverberation chambers can provide both sound clarity and reverberance due to the nonexponential temporal sound energy decay. Furthermore, variations of clarity and reverberance are essentially controlled by modifying the coupling area between the adjacent volumes. The present study focuses on the appreciation of changes in reverberation in terms of perceived suitability for different musical instruments or ensembles. Reverberation is often described in terms of reverberation time (reaction time [RT]), defined as the time required for the energy of a stopped sound to decay by 60 dB. However this metric does not provide a unique quantification of the decay profile and is therefore not sufficient to characterize nonexponential decays. To focus the study on physically possible variations of a given architectural space, the present study relies on variations of an architectural element, the coupling area, which embeds acoustical parameter variations, estimated from room impulse responses. Participants were asked to rate the suitability of reverberation settings for different instruments and ensembles. The sound stimuli were synthesized with an auralization method, providing control over the coupling area.Previous ResearchSingle volume concert halls have been widely studied while in comparison little research has been conducted on perceptual aspects of coupled volume concert halls. Perceptual thresholds, or just noticeable differences (JND), based on reverberation time have been estimated in large single volume rooms (Billon & Embrechts, 2012; Katz, 2004; Seraphim, 1958) and were found to lie between 5% and 10% of a given RT value (for RT 1.0 sec). In coupled volume concert halls, a recent study (Luizard et al., 2013; Luizard et al., 2015) based on an architectural parameter whose variations affect the room reverberance found a perceptual threshold of changes of 10% of a given open coupling area.Previous research has determined that although discrimination performance does not seem to depend on sound material (Frissen et al., 2010), reverberation preference, or evaluation, depends on the sound material (Frissen et al., 2010; Kuhl, 1954) proposed to the listeners. Preference for reverberation in the context of concert hall design has been previously studied (Ando et al., 1997) with the definition of four subjective criteria: sound level, initial time delay gap, reverberance, and interaural cross correlation; these acoustical parameters being defined in the ISO standard (ISO 3382-1, 2009). Another study (Ando et al., 1982) in single volume concert halls has shown that long reverberation times (around 2.7 sec.) were preferred for baroque music played with an orchestra, slow classical music required lower RTs (around 1.2 sec.) and preferred RTs for speech were 0.3 sec. However, another research group (Schroeder et al., 1974) has tested one of these classical music pieces using the characteristics of 11 concert halls and listeners preferred the stimuli corresponding to an RT 2.0 sec. This discrepancy can be explained by variations in stimuli parameters other than RT, such as initial time delay gap or interaural cross correlation, which are linked to spatial aspects and depend on the room geometry. Preference of reverberation in coupled volumes as compared to single-slope reverberation has been studied (Ermann, 2007) by means of listening tests with sound stimuli generated by ray-tracing software with no significant differences being observed between single-slope and double-slope reverberation.Preference of reverberation within coupled spaces has also been studied (Bradley & Wang, 2010) by means of geometrical acoustics simulations with participants preferring low to medium levels of decay curvature, as compared to high levels of curvature and classical linear slope exponential reverberation. …