Automatic genre classification of Indian Tamil and western music using fractional MFCC

Abstract

This paper presents the automatic genre classification of Indian Tamil music and western music using timbral features and fractional Fourier transform (FrFT) based Mel frequency cepstral coefficient (MFCC) features. The classifier model for the proposed system has been built using K-nearest neighbours and support vector machine (SVM) classifiers. In this work, the performance of various features extracted from music excerpts have been analyzed, to identify the appropriate feature descriptors for the two major genres of Indian Tamil music, namely classical music (Carnatic based devotional hymn compositions) and folk music. The results have shown that the feature combination of spectral roll off, spectral flux, spectral skewness and spectral kurtosis, combined with fractional MFCC features, outperforms all other feature combinations, to yield a higher classification accuracy of 96.05 %, as compared to the accuracy of 84.21 % with conventional MFCC. It has also been observed, that the FrFT based MFCC, with timbral features and SVM, efficiently classifies the two western genres of rock and classical music, from the GTZAN dataset, with fewer features and a higher classification accuracy of 96.25 %, as compared to the classification accuracy of 80 % with conventional MFCC.

BMNet-5: A Novel Approach of Neural Network to Classify the Genre of Bengali Music Based on Audio Features

“…An accuracy of 91.25% was found which took place due to the feature combination although it was evaluated comparatively on a smaller dataset with only 500 songs. The right combination of features can make a positive impact on the accuracy of genre selection as seen in the research of Betsy et al [24]. Compared to the conventional MFCC feature giving an accuracy of 84.21% combination of spectral roll-off, spectral flux, spectral skewness, and spectral kurtosis, combined with fractional MFCC features, outperforms all other feature combinations showing an accuracy of 96.05%.…”

Some New Results on Automatic Identification of Vietnamese Folk Songs Cheo and Quanho

“…In 2016, Rajesh, Betsy, and DG Bhalke [36] conducted the classification Tamil folk songs (Southern India) on a dataset of 216 songs (103 traditional songs + 113 folk songs) with 30 seconds duration for each song. The data for training in each type is 70 songs and the data for testing is 33 songs and 43 songs for each type.…”

Mô Hình GMM Định Danh Tự Động Một Số Làn Điệu Dân Ca Quan Họ Bắc Ninh

“…Đối với bộ dữ liệu âm nhạc Indian Tamil [19] gồm có 216 trích đoạn bài hát với độ dài mỗi đoạn là 30s và tập tham số đặc trưng là MFCC + Skewness + Kurtosis + Flux + Spectral Roll-off, độ chính xác đạt được với bộ phân lớp KNN là 66,23%. Với bộ phân lớp SVM, độ chính xác đạt được là 84,21%.…”