Audio Classification to Speech and Music using SVM and MLP

In this paper, the performance of some features based on wavelet transform are evaluated through classification of audio to speech and music using both the SVM and the MLP classifiers. The wavelet features compared to typical MFCC features as input into an audio classifier. Classification results show the wavelet features are quite successful in speech/music classification. Experimental comparisons using different wavelets are presented and discussed. By using some wavelet features, extracted from 1-second segments of the signal, we achieved 97.19% accuracy in the audio classification.

Keywords:

component , wavelet , Support vector machine , multi layer perceptron , audio classification