Auditory scene analysis as the basis for designing auditory widgets

Abstract

This paper presents a methodology for the design of fast-rate auditory presentations based on the Auditory Scene Analysis of Bregman. The auditory scene is hierarchically organized and based on the concept of auditory streams described in terms of two types of structures, one across streams (at an instant) and one within each stream (over time). Each stream constitutes a perceptual entity on which attention can be focused. Integrating information across streams requires effort and practice because the auditory system is inclined to derive properties such as temporal order or rhythm on a within-stream basis. The methodology distinguishes between the structure of the scene and the physical dimensions of sound used to produce this structure. Task performance may be critically affected by the structure of the scene. For this reason the structure is designed with respect to the tasks to be supported. When the structural aspects of the auditory scene have been defined, the physical dimensions of sounds are selected so as to provide the desired structure.