Output-based objective speech quality measure using self-organizing map

This paper proposes a new output-based method for assessing speech quality and evaluates its performance. The measure is based on comparing the output speech to an artificial reference signal representing the closest match from an appropriately formulated codebook. The codebook holds a number of optimally clustered speech parameter vectors, extracted from an undistorted clean speech database, and provides a reference for computing objective auditory distance measures for distorted speech. The median minimum distance is used as a measure of the objective auditory distance. The required clustering and matching processes are achieved by using an efficient data mining technique known as the self-organising map. Speech parameters derived from perceptual linear prediction (PLP) and bark spectrum analysis are used to provide speaker independent information as required by an output-based objective approach for speech quality measure.