Perception-based simplification for binaural room auralisation

Abstract

Room auralisation refers to the process by which the acoustic response of a room is rendered audible using signal processing techniques. The major part of the computational complexity in a binaural room auralisation system arises from the processing of the early reflections. However, most of the early reflections in a room are suppressed by the auditory system in a variety of psychoacoustical processes such as temporal masking and the precedence effect. This paper presents a perception-based data reduction method based on a mathematical model of the precedence effect. Results of a subjective evaluation in the form of a virtual sourceidentification experiment are presented. It is shown that it is possible to reduce the total number of reflections by 70% without significantly affecting the localisation acuity.12.