RECOGNITION OF MUSICAL INSTRUMENT FROM AUDIO FILE

Abstract

The software involves a formal method of automatic ontology generation for web based audio signal
processing application. Ontology is a set of concepts and category in audio file that shows their property and features. In
this system various musical instrument from wind to string families are classified using timber features extracted from
audio. To obtain models of the analyze instrument recordings k-means clustering is used to determine and optimize code
book of line spectral frequencies (LSF) or met frequency spectral coefficient (MFCCs. To classification technique based
on multi-layer perception(MLP) neural network and support vector machine (SVM)were tested. The input of software is
music audio collection and output name of instrument with it’s image.