How would you put a physics equation to music?

willum070's icon

I am totally in awe here, and probably out of my depth, but I'm listening to Luke Dubois "Year in MP3s". There is a track (89) that he says he generated by using physics equations. I wouldn't be so presumptuous as to ask for his source code, but I would like to prod the community as to what sort of approaches one would take in order to do this. Firstly, how does one select a satisfactory equation? Secondly, once you've got your equation into a patcher, how would you make it do stuff?

By the way you should check out the music: http://lukedubois.com/year. It's a great listen, and it's totally inspiring that someone actually made a track a day for an entire year. I particularly like his quote in this op-ed piece (http://opinionator.blogs.nytimes.com/2010/09/09/a-year-in-mp3s/): "...part of the challenge was avoiding my comfort zone of repetitive practice; with 365 pieces to make and no real expected listenership outside of my friends in my various creative communities, I felt, for the first time in years, that I could experiment and fail." Man, that's music to my ears. Thanks Luke for the inspiration!!!!

Brad Garton's icon

Luke's year is an *amazing* project. Don't feel it's too presumptuous to ask for his code, his plan is to post a lot of the "recipes" he used. In the meantime, check out a bunch of our classes (and on-line code/patches):

Luke's work is scattered throughout.

PS: people asking about my various objects, sorry for non-response. Life has been a little overwhelming lately. Soon... (I promise!)

willum070's icon

Hey thanks Brad, I appreciate it!

Michael Sperone's icon

I've based a lot of my music off of physics equations and theories. Though I wouldn't assume that my process of doing so is the same as Luke DuBois'.

But if it helps to hear my thoughts, here it goes: I think it's more of a compositional question than a patching question. How do you translate something like F=ma or E=mc^2 into music? There has to be some sort of compositional choices that must be made as to how this translates. Once you decide that, then you make the patch to carry out those compositional choices, which can as be simple or complex as you choose.

I have based my music not so much on the mathematical equations but more on the abstract theories behind them. Making connections from a scientific idea to a musical idea...

I guess I didn't really say too much that would be helpful :) But I guess you can see my point of view on the subject.

seejayjames's icon

You might do better by choosing some functions which are periodic, having some sin/cos elements to them. They could make some better tones at least. You could use other equations for things like amplitude over time, pitch shifting, or modulation, that might be cool.

Try making a simple interactive equation with [expr] [uzi] and [scale] which generates a bunch of points (y-values) across a given domain (x-values) that makes sense---a full period, or a half, or two, etc. Then you can mess with coefficients and exponents to see the fun wavelets form... add offsets... average two together... multiply two together...

you could do it with [vexpr] too, using a pre-made list for your domain. plenty of jit.expr for all this too.

Also check out [jit.bfg] for some cool ready-made functions, they could provide a great starting point. With a 512 x 8 matrix jit.bfg, you could grab each of the 8 rows with jit.spill and put them into 8 individual buffers~ which each play the "sample". With something like noise.voronoi you might get some cool cool results... hmm maybe I'll have to try that out...

Steven Miller's icon

A quick websearch for 'sonification' will give you some ideas on where to start, perhaps.

Roman Thilenius's icon

i do not think E=mc^2 can so easily be translated into (good) music, and E=mc^2 is not really releated to physics either.

two things come to mind: physical modeling synthesis, and physical movement data for composing.

you might have heard about the "falling balls" effect and similar constructs where you use gravity, velocity,
reflection and such data to create for example, a pairs of numbers for delta time of event and gain of event.

such data combined with simple FM synthesis makes you quickly something quite "physical".

last but bot least, there is a lot of "physics" already in basic classes - a tapping delay or a
resonator max object is more or less a physical formula or process itself.

-110