Results! FIR for the win!

Plotted results from Mathematica

No, above is not the result of an FFT, this is the result of running an audio file through a C# program that was running a large set of FIR filters in various octaves that resulted in a mountain of data that I pre-processed with Perl, then plotted with Mathematica.

The difference you get with this technique is added resolution in both the frequency domain (lower frequencies on top), as well as temporal resolution. Win-win.

I still need to tune the filtering scheme, I thing the filters are still a bit too wide for my liking, but that's perfectly OK for now. Results are results. Once you get to this point, you can go on to tune these things. Until you have a baseline, you have no way to improve.