Multi-Genre Symbolic Music Generation using Deep Convolutional Generative Adversarial Network

Music is an art that uses sound to convey emotions and ideas. It a universal language transcends cultural boundaries can move inspire individuals of all ages cultures. As with every form, the generation music complex challenging task. Despite challenges, has made considerable strides in recent years, owing application artificial intelligence machine learning. However, most research was focused on only one genre music, i.e., classical, jazz, etc., while there are more than 40 genres each sub-genres. This paper proposes model for multi-genre using Generative Adversarial Networks (GAN). Considering symbolic MIDI tracks were converted into piano-roll form after extraction musical information. Subsequently, GAN based trained learn distribution training data, it generates new data learned parameters. Generated evaluated survey musicians professionals. The results validate GAN’s ability generate multiple genres.