Learning about hall acoustics from multi-directional audio recordings

Abstract

Multi-channel audio recording, for playback in ambisonic format, is common when documenting a newly completed auditorium or an existing auditorium to be renovated. These recordings provide opportunities for comparative listening of different auditoria, and also between an acoustical model and the built auditorium. To allow meaningful comparison, recordings have been made using the same musicians playing the same musical excerpts on the same instruments. Anechoic recordings made similarly are used for convolution with synthetic (modelled) impulse responses. Comparative listening experience has yielded a number of lessons for acoustical design of auditoria and for acoustical modeling in design, including (a) the importance of instrument directivity to the sound of music in an auditorium, and perceptual responses to variations thereof, (b) the importance and challenge of loudness calibration during comparative listening, (c) inter-dependency of room size, perceived intensity and distribution of reverberation, and source directivity, (d) limitations in typical geometric acoustical modeling techniques, and (e) usefulness of directive loudspeakers when acquiring measured impulse responses for the purpose of convolving with anechoic audio. The process and results of these efforts will be presented and discussed. Selected listening segments will be available following the presentation.