Critique of certain concert‐hall design criteria

Abstract

Current concert‐hall designs are often seriously unsatisfactory for performers and musically experienced listeners, who tend to pay more attention to the music itself than to the “ambiance” of a hall. Formal experiments dovetail with practical experience to support the idea that the auditory system uses early reflections to compile information on tone, pitch, spatial, and temporal location (as well as loudness) via an extension of the precedence effect. This suggests that the “early reflection” criteria with 60 < t < 80 ms, which were originally proposed before Haas, should be based on t < 30 ms for at least a handful of early reflections to assure a good sampling of (among other things) the radiation patterns of the instruments. Data abound showing poor instrumental recognition when stimuli are recorded under anechoic conditions that give the listener only what the instrument happens to radiate in the direction of the microphone. However, source/listener motions and/or a few close‐in reflectors (with or without instrumental onsets and decays) restore recognizability to recorded sounds. This is consistent with experiments that compare perception of carphone‐presented stimuli with those arising from the statistical sound field of a room. Outlines of such experiments and of practical examples will illustrate and support the views described, and suggest the insecure basis of hall designs guided by undefined “listener preference” data obtained with variously processed versions of music recorded in an anechoic chamber. [Work supported by NSF.]