Classification of Recorded Classical Music using Neural Networks

Authors

Abstract

As a result of recent technological innovations, there has been a tremendous growth in the Electronic Music Distribution industry. Factors like the widespread access to the Internet, bandwidth increasing in domestic accesses or the generalized use of compact audio formats with CD or near CD quality, such as mp3, have given a great contribution to that boom. It opens new and exciting challenges, such as automatic music genre classification.
Automatic music genre recognition involves issues like feature extraction and development of classifiers using the obtained features. As for feature extraction, we use features such as the number of zero crossings, loudness, spectral centroid, bandwidth and uniformity. These are statistically manipulated, making a total of 40 features. As for the task of genre modeling, we train a feedforward neural network (FFNN) with the Levenberg?Marquardt algorithm. We use a taxonomy of subgenres of classical music. We consider three classification problems. In the first one, we try to discriminate between music for flute, piano and violin. In the second problem, we distinguish choral music from opera. Finally, in the third one, we try to discriminate between all the abovementioned five genres together.
We obtained 85% classification accuracy in the three-class problem, 90% in the two-class problem and 76% in the five-class problem. These results are encouraging and show that the presented methodology may be a good starting point for addressing more challenging tasks, such as using a broader range of musical categories.

Keywords

neural networks, music information retrieval, music classification, music signal analysis

Subject

Music Classification

Conference

EIS'2004, February 2004

PDF File