Undergraduate YPEC 2022

UG13 – Apply machine learning to generating sheet music

Modern pop music attracts young people and inspires them to learn how to play the music with instruments. However, most pop music is released without sheet music, which requires people with the capability of absolute pitch and the knowledge of music theory to edit sheet music on the notation paper manually or through music composition software. Many beginners have to search online and download existing sheet music, which is usually not free in the current market. In addition, a large number of songs have not been converted into sheet music. According to the market research, almost no application provides automatic generation of sheet music. Therefore, our team proposes a machine learning model to fill this gap. Piano works as one of the most common instruments for playing popular music. Every key on the piano corresponds to a constant frequency and a specific pitch. It is possible to generate sheet music automatically with the development of machine learning technology. The machine learning model will be trained with the input of frequency and the output of various combinations of pitch, which can decode a piece of music and generate piano sheet music.