Patrick Ikedi Egbuchulam
MIT EECS | CS+HASS Undergraduate Research and Innovation Scholar
Dynamic Background Music for Action Adventure Video Games
- Artificial Intelligence and Machine Learning
Michael Scott Cuthbert
Imagine playing through your favorite video game for the 20th time an exciting and tough adventure game. While the dynamic gameplay adapts to you as the player, the soundtrack hasn’t changed since your very first play-through. A game that responds to your controls should also have responsive music. We propose a system that will fit the music to your current experience (be it triumphant victory, inevitable defeat, or something in between). This project will solve three challenges to creating such a system. First, we’ll use our music theory knowledge to take a background theme input and output a variation to fit the player’s state. Second, we’ll use Machine Learning techniques to predict the game’s state from the engine. Finally, we’ll use audio engineering to replace the game’s audio engine with our own.
I’ve had the opportunity to take many music classes while at MIT, including Professor Eran Egozy’s Interactive Music Systems (21M.385), after which I realized that I want to continue doing substantial work that combines music composition with my computer science major. I’m very excited to be able to work on building a new music system with Professor Michael Cuthbert through the SuperUROP program.