What is computer generated music?
During the spring of 2017, a few software design students teamed up to generate music on the computer. The program uses Markov analysis of musical notes to suggest various new patterns of notes and therefore making new music! In addition to using Markov analysis, we also implemented a few additional pattern identifiers in order to produce smooth sounding music using techniques in music theory. The program lets a user choose from a variety of songs to base their new music creation off of.
Our minimal viable product for this project was a fifteen second tune generated from the command line that isn’t necessarily congruent. Intermediate goals included harmony generation, using multiple instruments, creating a full three minute song, having a web user interface, and having multiple song genres. Our stretch goal was be to implement Music Theory to produce a professional sounding song and to add lyrics. We reached a majority of our goals. We chose to not pursue adding lyrics and instead to focus on applying music theory.
Some members of our team have had extensive musical experience. The remaining members were curious about music theory. Our mixed backgrounds gave us insight to make a webapp that people without musical experience could use and understand. Our goal was to implement an algorithm that take a song, analyses it, and produces a new melody that sounds just as good. This leads to the question of, well what sounds ‘good.’ This is where music theory comes into play.
Please explore the rest of this site to see the ins and outs of our program.
Here is a diagram of what why we decided to create computational music.