Abstract. This project was primarily aimed to create an automated system for classification model for music genres. The first step included finding good features that demarcated genre boundaries clearly. A total of five features, namely MFCC vector, chroma frequencies, spectral rolloff, spectral centroid, zero-crossing rate were used for obtaining feature vectors for the classifiers from the GTZAN genre dataset [5]. Many different classifiers were trained and used to classify, each yielding varying degrees of accuracy in prediction. An ensemble classifier based on majority voting was then created to incorporate all of the classifiers into one.
Key words: music, genre, classification, MFCC, GTZAN genre dataset
Data Sets:
We have used the GTZAN dataset from the MARYSAS website. This is the dataset used in [5]. It contains 9 music genres, each genre has 100 audio clips in .au format. The genres are - blues, classical, country, disco, pop, jazz, reggae, rock, metal. Each audio clips has a length 30 seconds, are 22050Hz Mono 16-bit files. The dataset incorporates samples from variety of sources like CDs, radios, microphone recordings etc. We split the datset in 0.9 : 0.1 ratio and used 5-fold cross validation for reporting the results.
Music genre classification is widely discussed in the MIR (Music Information Retrieval) Society and has been of interest for a variety of reasons, including management of large music collections. The main purpose of this project was to implement few classification algorithms and compare their performance when applied to a practical problem. Specifically, we performed music genre classification of songs on a dataset containing 100 songs from 10 different genres.