Music Generation for Novices Using Recurrent Neural Network (RNN)

Abstract

Listening to music is a pastime most people enjoy. We're all fascinated with music and resort to listening to it in times when we're in a good mood and also while in distress. While a variety of applications and softwares exist that let musicians make music, there is not much development in the field for novices who do not understand music. This paper aims to change that. Not everyone should need to be an expert in the field to be able to create melodious pieces of music. This paper gives an approach to be able to do the same using Recurrent Neural Networks. The idea is to build a model that trains using existing melodies or instrumentals and generate new music based on the training. The approach will not only be helpful to people who do not know the field well but also to musicians to be able to generate fine quality music that can be developed further to make decent length songs. We aim to create music without having a need to play musical instruments physically.