GitHub - alextongue/markov-coltrane: MIDI melody prediction using a second-order Markov models. Includes a repoirtoire of 20 melodies from John Coltrane.

alextongue / markov-coltrane Public

Notifications You must be signed in to change notification settings
Fork 0
Star 0

MIDI melody prediction using a second-order Markov models. Includes a repoirtoire of 20 melodies from John Coltrane.

Notifications

Name		Name	Last commit message	Last commit date
Latest commit History 2 Commits
images		images
repertoire		repertoire
.gitignore		.gitignore
Manifest.toml		Manifest.toml
Project.toml		Project.toml
README.txt		README.txt
markovtrane.jl		markovtrane.jl
out.mid		out.mid

Repository files navigation

Second Order Markov Model Chain

[INSTRUCTIONS]
1. Open a terminal to this directory and start Julia.
2. In the Julia REPL, type ']' and enter "activate ." to activate the environment. This uses a few packages that are included in the "general registries".
3. Press backspace to return to the REPL and enter: include("markovtrane.jl")
4. A file named "out.mid" will be written to this directory. Figures should also come up, but I'm not sure if your shell runs the same as the my computer.

[NOTE 1] If you'd like to play with some parameters, they are included in the first code block of the file. I have only tested this up to second order (predicting a note based on previous two notes), but you are welcome to change the first two notes that prime this framework. It takes inMIDI note values.

[NOTE 2] There is an option to construct a separate train/inference Markov model for velocities. I find that it's a little more sensitive to being primed right, but there are instructions to set that up in the beginning of the script as well.