Procedural generation is a potent tool for creating highly replay-able video game content.

Given a sufficiently advanced procedure, and a fresh seed, it is capable of making unique landscapes, characters, equipment, stories, and other features to the player, on every playthrough.

Following from this, music proves a uniquely suitable seed, due to its near infinite compositions, and how players can provide a diverse selection of their own tracks that will appeal to them.

Thus, this work presents MLoDI; a game created to test what data can be gathered from music, how it can be applied to a game, and to study the methods to do so.