Ideal Flow Network (IFN) can be used to generate music. Unlike random music generation, IFN music is learned from whatever music you give. For instance, if you give Mozart song as the training example, the music generated would be similar to Mozart song. If you give other song such as Kakatua song, the music generated would be similar to Kakatua song. That is exactly how Machine Learning in Artificial Intelligence (AI) works. Using the same and blackbox generic model of IFN, you can create new things simply by changing your training sample. When you train the AI to be Mozart, it would behave like Mozart. If you train the AI to be Kakatua, it would behave like Kakatua.
The purpose of this page is not to claim that IFN is better than many music algorithms but simply to demonstrate that IFN can also be used for music generation.
1. Select Sample Music Notes
2. Play / Stop the Music
Click "Play" button to play your sample music. Don't forget to turn ON your volume
3. Let IFN learns from your training notes and Generate New Music Notes for You
Select how many bars do you want to generate and click "Generate New Music Notes" and then "Play" buttons. Don't forget to turn on your volume
Total bars:
Optional Setting your Training Example of Music Notes and Music Parameters
Music Notes
Speed:
(higher is faster)
Time Signature: /
(total note values/beat) per measure
Optional Notes Manipulation
Operation:
Alternatively, you can also code your own Music Notes
How to write your own music notes?
Each measure (or bar) is separated by a comma, except the last line
In one measure, the total tempo must be equaled to the time signature (total notes/beat)
A musical note consist of 3 to 4 characters:
The first character tonic is either A,B,C,D,E,F,G or rest
The second character is chromatic, either empty, b, x, or #
The third character is octave, an integer number to represent the high or low pitch
The fourth character is tempo, an integer number to represent the high or low division by time
