The Rhythm Method: Crowd-sourcing Drum Science

It’s always fun making Sixty Symbols/Numberphile videos with Brady Haran but the most recent filming brought together quite a number of my core enthusiasms — Rush, physics, drums, and noise (in all senses of the term) — and so was even more enjoyable than usual.

Brady uploaded the videos this morning (here and here). The first discusses a fascinating recent paper by Esa Räsänen and colleagues which focuses on the fluctuations in timing in the virtuoso drum pattern played by Jeff Porcaro in Michael McDonald’s “I Keep Forgettin’ “. I’m not a huge fan of the song but Porcaro’s groove is certainly infectious. He’s also responsible for the fantastic drumming in this classic, among so many other others. (Please excuse the excruciatingly high cheese factor in that Toto video. It was the Eighties.)

Fortuitously, I read the paper by Esa, Holger and their colleagues at around about the time an e-mail arrived asking for suggestions for undergraduate projects. As Esa et al. state in the conclusions of their paper, there is particularly exciting scope to extend their analysis to other songs, drummers, and styles. So I proposed an analysis of fluctuations in drum beats as an undergraduate project and was delighted when two 3rd year Physics students at Nottingham, Easel Kandola-McNicholas and Adeel Bokhari, selected the project.

What we want to do is analyse the fluctuations in timing/rhythm for not just one drummer — as Esa, Holger et al. did — but for as many drummers as possible. Enter Sixty Symbols. While we could have stuck with an analysis of the Porcaro pattern — and, indeed, if you’re a drummer, please feel free to send us your version of “I Keep Forgettin’ ” to the address below — there’s another single-handed 16th note pattern which is very famous among drummers: Rush’s “Tom Sawyer”. Neil Peart is widely recognised as among the very best rock drummers in the world, and there are very interesting parallels with the Porcaro track in terms of the way the 16th note pattern is played, so, in many ways, “Tom Sawyer” is a natural choice. (My love of Rush is, of course, entirely coincidental…)

I put together this short video (using the wonder that is Aerodrums — see below) to show some examples of the 16th note patterns in “Tom Sawyer” and to explain what we need for the analysis.

Our aim is to publish the analysis and include the names of all those who contributed their version of “Tom Sawyer” (and/or “I Keep Forgettin’ “) in the paper. If you’re a drummer and you’d like to contribute please e-mail your WAV, MP3, or MIDI file (or any other appropriate file type) to There’s no deadline — we’ll accept drum tracks for as long as it takes to get good statistics for the analysis. The more, the merrier.

By the way, Aerodrums are available here. I enthusiastically recommend them! I’m not a drummer and so Aerodrums are ideal for learning to play and for putting down rhythms when song-writing or demoing tracks. But they’re much more than this — in very many ways, Aerodrums are just as good as a real kit, as this impressive example of virtuoso aerodrumming shows…

[A huge thank you to the University of Nottingham BandSoc and, in particular, Jedd Bellamy-Carter for providing the practice room for the video and for all their help with equipment].

Author: Philip Moriarty

Physicist. Rush fan. Father of three. (Not Rush fans. Yet.) Rants not restricted to the key of E minor...

22 thoughts on “The Rhythm Method: Crowd-sourcing Drum Science”

  1. Are you planning on including live performances by Neil Peart/ RUSH and covers by other bands and what not, or just submissions?

    Also is it only fluctuations in hi-hats that you’re studying? Or also things like kick dynamics, ghost notes, fills, looseness of hi-hats etc.?

    I can’t personally contribute at the moment living in an apartment without a kit, but as a Canadian drummer I do feel obliged…

    ps another Canadian song with 16th hi-hats…

    It’s completely different from Tom Sawyer in that it’s simple, two-handed, and for the most part the accents are all in the same places, but I suppose that leaves it less prone to drummer error and more open to drummer interpretation… I don’t know, I’m just really excited about all this


    1. Hi, Jon.

      Thanks for those great comments. At the moment we’ll focus on the hi-hat but all of the other aspects are of interest for future work. Ghost notes are of particular interest in terms of the subtlety of the timing and the extent to which they add to the “groove” or overall appeal of a drum track.

      Given that you don’t have a kit at the moment, I thoroughly recommend Aerodrums. They’re inexpensive (~£100), exceptionally easy to set up, feel much like playing a real kit, and they’re entirely noise free. I’m not being paid commission by Aerodrums to advertise their product — promise! It’s just that I really am blown away by how good Aerodrums are.

      All the very best,



  2. When is the closing date for submissions for drum samples? (It may be a few days until I’d be able to submit to this.)


    1. Hi, Ryan.

      There’s no closing date any time soon! We’ll keep taking drum tracks for the next few months at least. Thanks for submitting your track — really appreciate this.


  3. Great project, but I can see three different sources of data that need to be differentiated. The first is the ‘solo drummer’, keeping track of the song in their head with nothing to respond to as it were. Perhaps they are playing to a click track. The second is solo drummer playing to backing track. The drummer responds to the song in perhaps a different way than if there were no backing. Finally, drummer in a band. Here, the responses are more complex, as each band member is responding to every other. I am an ex-physicist and musician, and am very aware that my playing very much depends on whom I am playing with. Granted, my music is largely improvised, and I would never be able to faithfully recreate a particular tune without imparting a bit of the moment, but surely these variables are still there and need to be isolated. Would it be appropriate to find out the conditions and setup when each recording was made? By the way, great project


    1. Hi, Andy.

      Agree entirely with your points! Indeed, the group to whose paper I refer in the blog post have previously looked at the role of playing to a metronome/click-track in some detail. In the PLOS ONE paper they state that their analysis of Porcaro’s playing would lead them to propose that Porcaro didn’t play with a click-track (which certainly sounds more than plausible to me!)

      We plan to send out a very short questionnaire to all participants to attempt to address some of the points you raise.

      Thanks for your perceptive and helpful comments,

      All the best,



  4. Hi Phil-

    I’m going to try to get a recording of my playing this weekend. A few questions:

    1. Is there a minimum sound quality threshold you need? In other words, do you think recording it on my phone will not give a high enough sound quality? I don’t think I have easy access to better recording equipment right now.

    2. Does the song have to be played on the original tempo? So, should I play to a track, instead of just freestyling it? I’m sure my internal sense of tempo is flawed enough that there would be a difference

    3. Is there any benefit to sending a video? While watching the Sixty Symbols video I couldn’t help but thing about single-hand playing technique, which I know has an impact on the sound (as you mention yourself in the video).



    1. Hi, Sam.

      Thanks so much for submitting a track!

      1. If you can isolate the hi-hat track (or have it as high as possible in the ‘mix’) that’d be very helpful.If not, don’t worry — we’ll do our best to extract it.

      2. Yes, we need the song at the original tempo, I’m afraid.

      3. A video would be great but not essential.

      Thanks again for helping us with this,



  5. Hey!!

    i just found your video on the hi-hat experiment(sixty symbols) and i got very exited when you said you wanted tom sawyer as its probably my favorite song to play to.
    my only concern is that i’m too late to contribute as the last few posts were posted in October, in fact i’m almost 3 months late but id still be more than happy to send in a recording if you guys want,

    if i’m too late i can understand that, commend you guys for studying such an epic song and rock on!!


    1. Hi there.

      Thanks for leaving a comment. It’s definitely not too late. We’ve had about 80 submissions thus far but the more the merrier! Please send your track to If you can isolate the hi-hat track that is ideal. If you can’t then if you can keep the hi-hat as loud as possible in the “mix” that makes the analysis a little more straight-forward.

      Thanks again.



    1. Hi, Phillip.

      Thanks for your interest in the project and for volunteering to get involved. The students (Easel and Adeel) who’ve been doing the analysis handed in their reports last week. We’ll write up and submit a paper on the data over the summer break.

      One issue that we encountered is that many of the tracks submitted were not sufficiently long to enable a robust study of long-range correlations. (This was entirely my fault for not realising just how much data is required for a reliable analysis). If you could submit a track of 16th notes on the hi-hat at 88 bpm which lasts for at least three minutes that would be fantastic. We can then compare that against a number of 3 minute (and longer) tracks we have.

      If you could submit a track (to which is 3 mins of just the hi-hat (or where the hi-hat is very high in the “mix”) that’d be fantastic.

      Thanks again,



Comments are closed.