Classification of punjabi folk musical instruments based on acoustic features

Abstract

Automatic musical instrument classification can be achieved using various features extracted such as pitch, skewness, energy, etc., from extensive number of musical database. Various feature extractionmethods have already been employed to represent data set. The crucial step in the feature extraction process is to find the best features that represent the appropriate characteristics of data set suitable for classification. This paper focuses on classification of Punjabi folk musical instruments from their audio segments. Five Punjabi folk musical instruments are considered for study. Twelve acoustic features such as entropy, kurtosis, brightness, event density, etc., including pitch are used to characterize eachmusical instrument from 150 songs. J48 classifier is used for the classification. Using the acoustic features, recognition accuracy of 91% is achieved. � Springer Science+Business Media Singapore 2017.