Within a week, Billions of melody is copyright free

Alfred Wang
4 min readDec 8, 2020


Thanks to Riehl and Rubin

As Technology continually improves, making music have become easier , a lot of songs we listen now often feels like a mimic of the other, these kind of situation often let music producers facing copyright issues

But sometimes, someone on this planet, some where, might’ve wrote a song similar to yours already.

Katy Perry’s Dark Horse vs Joyful Noise

A federal jury of six people found Katy Perry guilty of copying a Christian rap song to create 2013’s “Dark Horse,” one of the biggest hits of her career.

The case even though Katy Perry claimed never heard of it before, however, in the end , she has won the ongoing plagiarism dispute over her 2013 hit “Dark Horse.”

“We’ve created an application to generate by brute force all mathematically possible melodies and write them to MIDI files. The application accepts various parameters (e.g., pitch, rhythm, length) to mathematically exhaust all melodies that have ever been — and are mathematically possible.”

Riehl and Rubin uses programming and algorithms solved this issue.

How does the program works?

Using piano as an example

First, on all the notes we give them a number, such as “Do” = 1, “Re” = 2 until “Si”, all 7th of them but also with the five black keys, now that’s an octave.

Total of 12 notes right ?

Most pop melodies runs fewer than 12 notes.

So the program basically just goes







Remember the story where the king wanted to reward the inventor of the chess?

“one grain of wheat for the first square, two for the second, four for the third…”

However, if you generate every possible melody with just one of the C scale, that’d be 8¹² melodies, which is 68,719,476,736.

And by doing this, all the melodies in the earth within one octave is now collected, there’s no such thing as rhythm, beats, this is just math. However, this only puts all the possible order, so it might not sound right to your music taste, it also makes some of them really awkward.

But that doesn't matter, the purpose of this program is to have it all. They built a system to make those melodies into MIDI files. (a file for computers to talk to keyboards) They originally started with Python but switched to Rust, since it gives them the speed they need. They have made all the melodies in here

“When MIDI files were written to disk, they were copyrighted automatically and then dedicated all newly created files (to which we have legal rights) to the Creative Commons Zero (CC0).”

Melodies are math: Finite combinations of notes/pitches

The reason for Riehl and Rubin is to protect the current music producers with copyright issues. All melodic combinations are mere math.

What is the status of those machine-generated works: copyrightable or uncopyrightable?

  • IF copyrightable,
    THEN All the Music has dedicated all works to which it has copyright into Creative Commons Zero (CC0).
  • IF uncopyrightable because Melodies = Math = Unoriginal/Facts = Uncopyrightable (Or Thin Copyright)
    THEN all ATM(All The Music)melodies are uncopyrightable facts, available for others to use.

Under either scenario, we think that most (if not all) of the ATM dataset might be available for all songwriters to use.

What about rhythms ?

Mashup: “Twinkle / ABCs”

Consider this example: Many people have been surprised to realize that these three songs share the same melody


  • SONG 1. Twinkle, Twinkle Little Star
  • SONG 2. ABCs

Even though under current copy right laws, rhythm doesn't matter, it is still included.The application accepts various parameters (e.g., pitch, rhythm, length) to mathematically exhaust all melodies that have ever been — and are mathematically possible.

Axis of Awesome — 4 Chord Song

There are still debates where or not these machine generated are considered copyrights or not because of the way it is generated, however, the purpose of this program is to protect a innocent writer to be able reference to this open space. For more about legal questions, the website has a very detailed FAQ.



Alfred Wang

Learning one step at a time