I am not sure if this kind of question is on-topic or not.

I just heard the soundtrack used in Scott Manley's recent video SpaceX Radarsat Launch Highlights (fairly quiet sound level) with the notes

Twitter sized ( less than 140 seconds) with music by Test Shot Starfish.

It reminds me very much of Jamie xx's Gosh (loud sound level).

I've set the two time cues for the videos to roughly analogous points.

Are these in fact two versions of the same music, or is the resemblance just a coincidence? Are there objective, unambiguous ways that the level of similarity can be measured? If so, how might they be applied in this case?

This one has a fairly quiet sound level:

This one has a loud sound level:


3 Answers 3


No, there's no objective mathematical distance metric for music excerpts. Although there's immense marketing funding to make one, it would depend on:

  • listeners' cultural background: if you grew up listening to ragas, Beethoven and the Beatles are indistinguishable.

  • human psychoacoustic messiness: listening conditions enormously affect which aspects of the music you can attend to, nevermind recall in an A-B-A test.

  • personal musical history: within the Western tradition, there's only so many chord progressions. Air on the G String, Pachelbel's Canon, Wachet Auf, Warm and Tender Love all are sisters harmonically to Whiter Shade of Pale. When one considers its lyrics, its sisters are even more numerous.

My recent peer-reviewed paper about this shows how far we still fall short of what you desire.


The answer is mostly "it depends." Some examples:

"Cottontail" is a jazz number which uses the exact chord sequence of "I Got Rhythm."

Many classical composers wrote "Variations on a Theme by {other composer}."

Several pieces by Mahler, Bartok, Dvorak, to name a few, incorporate folk-music themes.

What I'm suggesting is that any given person can generate his own "figure of merit score" based on things like chord sequence, rhythms, lead tune and decide how much alike two pieces are.

  • In fact, rhythm changes are dirt common in jazz. I've read anecdotes that jazz music students are sometimes assigned to write music using the rhythm changes.
    – Dekkadeci
    Commented Jul 1, 2019 at 13:35
  • 2
    "Cherokee" does not use rhythm changes.
    – PiedPiper
    Commented Jul 1, 2019 at 14:48
  • @PiedPiper you are correct; I was ambiguous in my statements Commented Jul 1, 2019 at 15:24
  • 1
    Double checking; the answer to "Are there objective, unambiguous ways..." is it depends? Suggesting that any given person can generate his own "figure of merit score" doesn't seem to count as an "objective, unambiguous way..." which is what I have asked about. To my question as written, perhaps then your answer is no?
    – uhoh
    Commented Jul 3, 2019 at 0:52
  • A distinctive feature of Rhythm changes is the middle 8's 4 chained V-I progressions. Cherokee has a chain of 8.
    – Rosie F
    Commented Oct 28, 2019 at 8:42

If you could reduce the two examples of music to some kind of code then you have an objective measure; two codebases can always be compared with tools.

So maybe some kind of frequency or signal curve analysis? It's not my field.

  • This is essentially what tools like Shazam do to identify not only the piece but the specific performance. I don't know what the Shazam's algorithms' error rate is for two performances or songs that are almost identical. Commented Jul 1, 2019 at 13:21
  • @CarlWitthoft I use SoundHound rather than Shazam but there the error rate for different recordings of the same piece of classical music is extremely high. It mostly recognizes the piece, but which version it claims to identify is pure chance.
    – PiedPiper
    Commented Jul 1, 2019 at 15:43

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.