Procedural music, taught with data from real artists. Live demo
Check requirements.txt for python dependencies
To convert to mp3, run the following:
brew install libsndfile lame brew install --with-libsndfile fluidsynth
Neural Network structure
The following program uses a Feed Forward Neural network that is trained with midi data to generate a song. There two hidden layers present and 25 nodes on the first hidden layer and 10 on the second.
The following files have the corresponding functions.
|tomidi.py||Converts an intermediate text file with midi chord data to a midi file to be played|
|learn.py||Creates an intermediate file from the all the trained data.|
Basic music theory
Major: 1 2 3 4 5 6 7 M m m M M m o Natural minor: 1 2 3 4 5 6 7 m o M m m M M