I was playing around with some interesting recording techniques last night and discovered some expected and unexpected behavior in the waveforms.
Tonight I'll be experimenting with an oscilloscope and other tools to analyze the waveform produced by my amp/speaker/cab/tonestack with scientific rigor and thoroughness.
All but the last image are the same waveform over different periods of time. Each is a note played through my guitar and amp, recorded by a mic in front of the speaker cabinet. The prominent "tornado" shape in the first waveform is a factor of the way the tubes are biased and a "ripple" in the power supply caused by the current draw of the tubes, the load of the speaker/output transformer, and the properties of the power transformer itself. The last waveform shows a note transitioning from hard, gritty, clipping distortion to a smooth, pleasing, compression(blues) distortion, and eventually fading into a clean(mostly) tone. The asymmetry is caused by asymmetric biasing on the plate voltage to the tubes and the larger(200 volt+) voltage swings. The curious thing is that the asymmetric clipping is on the wrong harmonics. Either my tubes are biased too hot, or the signal is inverted somewhere along the way. The odd harmonics(the ones on top) are being clipped too much, and the even harmonics(the ones that sound a little better when clipped/compressed) are not being clipped/compressed enough. Either that, or the signal is inverted in (A) the cathode follower, (B) the power output stage, (C) speaker is connected with wrong polarity, (D) the microphone inverts the signals it recieves, or (E) the preamp on the microphone inverts. This is important because the polarity of the signal determines whether its a compression or decompression in the air. Odd harmonics drive the speaker forward, and thus compress the air, and even harmonics pull it back, decompressing. 100% fidelity in odd harmonics and 95% fidelity in even harmonics will produce a tone that is actually more pleasing than the 100%/100% fidelity tone. As a refference, when I post all of my work/findings, I'll include some transisor-based amplifiers and their results run through essentially the same tests, as well as the results of software/signal processors designed to emulate the sound of tubes.