Music Genre Classification

Abstract

Music genre classification has taken on a new significance in recent years as a result of the significant expansion in the accessibility of music data. We need to correctly index them in order to have better access to them. When working with a large collection of music, using an automatic music genre classification is essential. Researchers have favored machine learning methods for the majority of contemporary music genre classification methods. We used two distinct genre-specific datasets in this investigation. The system is trained and categorized with the help of a Deep Learning technique. Classification and training are carried out with a convolution neural network. The feature extraction step is the most important part of speech analysis. The primary method for extracting audio features is the Mel Frequency Cepstral Coefficient (MFCC). The proposed method divides music into several genres by extracting the feature vector. Based on our findings, our system has an accuracy level of 80%, which will make music genre classification much easier and significantly improve with additional training. The term "music" is a common way to classify the various kinds of music. It must be distinguished from other types of music. There are many different ways that music can be divided into genres. Pop, classical, rock, and other genres make up the music. In the field of music information retrieval, the most difficult task is to classify music by genre. In order to obtain music from a large collection, automatic music genre classification is essential. It has applications in the real world, such as automatically tagging unknown music (which is useful for apps like savaan and wynk etc.)