AUDIO CLASSIFICATION IN SPEECH AND MUSIC: A COMPARISON OF DIFFERENT APPROACHES

This paper presents a comparison between different techniques for audio classification into homogeneous segments of speech and music. The first method is based on Zero Crossing Rate and Bayesian Classification (ZB), and it is very simple from a computational point of view. The second approach uses a Multi Layer Perceptron network (MLP) and requires therefore more computations. The performance of the proposed algorithms has been evaluated in terms of misclassification errors and precision in music-speech change detection. Both the proposed algorithms give good results, even if the MLP shows the best performance.