A Cognitive Model to Mimic an Aspect of Low Level Perception of Sound: Modelling Reverberation Perception by Statistical Signal Analysis

Abstract

Sound reproduced and perceived in different environments, or transmitted via diverse transmission channels shows distinctive acoustic characteristics, sometimes quoted simply as acoustics. Acoustics of an auditorium can be described by the transfer function from the source  to the receiver, or the impulse response in the time domain, which can be measured by instrumentation using a number of methods. On the other hand human listeners, especiallytrained musicians, sound engineers and acousticians canaccurately differentiate acoustics of auditoria by  listening tothe sound effects, indicating acoustics can be viewed as a lowlevel human perception of sounds. This paper presents acomputing model and algorithms to mimic human perceptionof reverberation, arguably a most significant aspect of acousticperception. This is done by statistical signal analysis usingmaximum likelihood estimation with a purposely chosenenergy decay model.