Self organizing feature map network for musical instrument sounds

Abstract

Self Organizing Feature Map Network is used as a classifier for classification of features extracted of the sound database of Musical Instruments. However, the database of Indian Classical Musical Instruments is prepared for 15 instruments. The different types of features such as Temporal, Spectral and Cepstral features are available out of which we have considered Spectral features i.e. Roll off, Centroid, RMS Energy, Zero Crossing Rate, Spectral Irregularity & Spectral Brightness.