Abstract
A nonlinear, binaural auditory model was developed which is able to act as an artificial listener for assessing features related to the quality of room acoustics. The model is able to derive objective parameters for reverberance, clarity, apparent source width (ASW) and listener envelopment (LEV). In order to derive these parameters, the model splits the input audio into two streams: one related to the source (direct sound) and one related to the environment (room). In order to derive these two streams, the splitting algorithm makes use of the nonlinear properties of the auditory model. The whole algorithm works on arbitrary binaural recordings and therefore the parameters can for example be determined in a concert situation using an artificial head. This way the signal type is automatically taken into account, meaning that the acoustics of a room can be tested for multiple types of stimuli. Listening tests show that the resulting objective parameters correlate highly with perceptual results, even in cases where conventional objective parameters show low correlation coefficients.
Get full access to this article
View all access options for this article.
