Originally published in 2009 at http://www.math.ucla.edu/~rcompton/musical_sorting_algorithms/musical_sorting_algorithms.html

Suppose you drop a set of drums and they land randomly ordered in a row on the floor. You want to put the drums back in order but can only pick up and swap two at a time. A good strategy to minimize the number of swaps you must make is to follow the Quicksort algorithim http://en.wikipedia.org/wiki/Quicksort.

If you order the drums by their general MIDI number and simultaneously strike any two which you swap then you will produce a sound similar to this:

https://www.youtube.com/watch?v=g2IWUd3p30I

To hear what’s going on with quick sort a little better consider the case where you have dropped 12 guitar strings whose frequencies vary expoentially, ie

frequency of string \(i = 220.0*( 2.0^{\frac{i}{12.0}} )\)

for \(i=0,...,11\) for more information http://en.wikipedia.org/wiki/Pitch_(music)

https://www.youtube.com/watch?v=YR6VAZUGAMo

I could not hear what’s going on all that well either. An aesthetically pleasing quadratic running time strategy you may follow is insertion sort http://en.wikipedia.org/wiki/Insertion_sort. Here’s how insertion sort sounds in a scenario involving many more strings on the floor:

https://www.youtube.com/watch?v=DNAmWDmIAZQ

And if you were to sort drums with insertion sort you would get something like this:

https://www.youtube.com/watch?v=zXTlRc6QM-M

The sounds were made using STK http://ccrma.stanford.edu/software/stk/ the videos were made using the python gnuplot module http://gnuplot-py.sourceforge.net and mencoder. The source is at https://code.google.com/p/musical-sorting-algorithms/

Update 2009-08-11: This got written up in Make