Traditional Malaysian Musical Genres Classification Based on the Analysis of Beat Feature in Audio

Abstract

Interest on automated genre classification systems is growing following the increase in the number of musical digital data collections. Many of these systems have been researched and developed to classify Western musical genres such as pop, rock or classical. However, adapting these systems for the classification of Traditional Malay Musical (TMM) genres which includes Gamelan, Inang and Zapin, is difficult due to the differences in musical structures and modes. This study investigates the effects of various factors and audio feature set combinations towards the classification of TMM genres. Results from experiments conducted in several phases show that factors such as dataset size, track length and location¸ together with various combinations of audio feature sets comprising Short Time Fourier Transform (STFT), Mel-Frequency Cepstral Coefficients (MFCCs) and Beat Features affect classification. Based on parameters optimized for TMM genres, classification performances were evaluated against three groups of human subjects: experts, trained and untrained. Performances of both machine and human were shown to be comparable.

Genre Classification of Traditional Malay Music Using Spectrogram Correlation

“…The best overall average accuracy rate of 61.8% obtained here is comparable to the spectrogram technique used for Latin music [11] with the advantage of not having to extract features from the spectrogram. The result obtained using this proposed technique is however lower than those reported by others for Malaysian music genre classification of which traditional Malay music is a subset at 88.6% [7] and 69.1% [8], where conventional techniques of extracting several features directly from audio signals were used. It may be argued that the feature used in this paper, which is the whole spectrogram image, is simpler.…”

“…Automatic music genre classification is important for information retrieval systems, sprouting numerous research in this field, especially for Western music genres [3][4][5][6]. Other genres have also been studied including Malaysian music genres, of which traditional Malay music is a subset [7][8]. However, as pointed out by these researchers, the main drawback for studying local genres is that there is no large dataset available such as those used for Western genres research.…”

“…Once features are extracted from a music tract, a classifier will be employed to determine the genre. Supervised learning approaches such as K-Nearest Neighbors, Hidden Markov Models, Neural Networks, Support Vector Machines and many others similar to those used for speech recognition have been reported [3][4][5][6][7][8]. A number of studies tried to use different classifiers to try to improve performance.…”