Towards enabling measurement of similarity of acoustical environments using mobile devices

Abstract

There is a similarity between algorithm for automatic speaker recognition, speech or emotion recognition in speech, and in general automatic sound recognition. Algorithm based on use of mel-frequency cepstral coefficients and covariance matrices, previously used for automatic speaker recognition, is extended to automatic sound recognition. One example of using application for automatic sound recognition is presented. This application can show the similarity between different acoustical environments.