I've decided that I will at some point get an SPL meter and test some things. I seriously doubt it will change the obvious tonal differences audible between, say, OPAMPs. I might try hooking my amp back into the computer using the pre-amp outputs to see if I can measure some tonal differences though. If only I could justify an oscilloscope.
Run the analog output back via an ADC to a PC and record it in something like Audacity, repeat 10 times to get over random variations, trim and align all 10 samples, run the recorded samples through the spectrum analyzer, export the analyses to spreadsheet average the results and compare against the reference.
repeat as necessary for chages in variables...
It is a touch tedious but the results will be pretty solid.