ABSTRACT
This chapter describes a model of the primary stage of auditory scene analysis. This stage of analysis consists of two parts: segmenting the auditory scene to a collection of elementary auditory units and estimation of features such as onset, offset, and frequency and amplitude dynamics for each elementary unit. The formation of elementary units in the presented model is psychophysically and physiologically motivated. Both the segmentation and the feature estimation algorithms were tested on variety of auditory scenes and found to be a very good basis for computational auditory scene analysis grouping models.
