posted by [identity profile] whc.livejournal.com at 03:24pm on 2007-06-15
Instead of comparing waveforms, how about running them through an FFT algorithm and comparing the spectra?
 
posted by [identity profile] dglenn.livejournal.com at 11:58am on 2007-06-16
Something along those lines did eventually occur to me, in a vauge, half-formed way, when I started wondering which overtones added up to each of those two waveforms (and not expecting it to be as easy to find my trial-and-error (and guessing) as the y = sin(θ) + sin(2×θ+π/2) for the two-notes-on-one-string analysis from a couple months ago (http://dglenn.livejournal.com/919887.html) -- maybe FFT would have come to mind more quickly if I'd worked with Fourier transforms before instead of just reading about them once in a while. First step, I suppose, is 'man -k fourier ; man -k fft' to see whether I've already got tools to do the magic for me without my having to bother reading up on the math. :-)
 
posted by [identity profile] dglenn.livejournal.com at 02:34am on 2007-06-18
Did you mean something like this?
Image
It turns out that Audacity does this, but I couldn't convince it to use the same vertical scaling for both plots, making it difficult to compose them together in GIMP. But there's an "export" button that generates a text file of values, and I've got GnuPlot, so ...

There were a bunch of options on the Audacity "analyze spectrum" dialogue that I'm really not sure I understand right, so I'm not sure what I got (pretty much the default) is the most useful version.)

Links

January

SunMonTueWedThuFriSat
1 2 3 4 5 6 7
8 9 10 11 12 13 14
15 16 17 18 19 20 21
22 23 24
 
25
 
26
 
27
 
28
 
29
 
30
 
31