Easy.. It's very unscientific
The preamp out is connected directly to an 8" active monitor (bi-amp). I have to be careful that my distance to the monitor is the same for both tests. I'm usually sitting down, and I always look at the same reference point (say, a picture frame, or spot on the wall, or the tweeter cone). That way, my head position is always the same, and therefore my ears are at the same spot too (or roughly the same spot).
I have the stock pre, and the modified pre (or prototype) on my desk. First, use the 1st pre. I'll loop the same music material continuously, until I'm used to the sound. Then I'll switch jacks and run the 2nd pre... same material, same distance, same head position. I'll take note of specific musical passages, and keep looping it. Then I switch again and see if what I think I heard is really there, or just my imagination. I use orchestral music, with very fine piano passages, and loud percussive sound in other parts, mixed with some strings, french horns, etc in other passages. I'll take note of the crash cymbals or the taiko drum, or the violin strings. How do they sound.
I remember on an early preamp prototype last year, my pre would "trip up" when one specific section came up (very loud, very transient percussive sound) on one specific music (It was a Yanni CD
) It didn't happen on any other loud music material or CD... just this one particular song. So hook up the oscilloscope, run the audio generator and sweep up and down the frequency until I replicated the problem. So that got fixed.
In another case, I noticed the output of the 2nd pre was lower in volume than the 1st pre. It seems just a tad bit lower, but hardly noticeable listening to it. So I hooked up an oscilloscope and measured it, and found the output of the other pre was 0.4V less. Reviewing the board, I noticed a 2K67 resistor instead of a 3K32 in it's place. Replaced it, hooked up the oscilloscope and now they're on the same volume, and listening tests confirm it too.
And I'll do the same listening tests several times, a few days apart also.
I also have a software spectrum analyzer just to see if anything changed... But it's really hard to tell because the graph doesn't tell the whole story. (Not unless the change is really obvious.)
Swapping chips with another model/brand can also reveal the weakness of the older chip models. The high frequencies just get smeared and harsh to the ear (sounds like a fissssssh instead of a pissssss) , and the music just sound lifeless and dead. It's like the difference between listening to a speaker covered with a blanket, and listening to one without. That's best how I can describe it, like a veil being lifted.
I also try different mics, dynamics, condenser and ribbons.
Sometimes, I'll record the same performance on 2 different signal paths, and then chop the resulting audio and butt them together alternately, and listen to it to see if there are differences. Here's an example:
http://www.fivefishstudios.com/audio/SC-1ActiveDI-vs-RadialActiveDI.mp3Everytime you hear a click or tick, that's audio being swapped between the 2 sources. This is a comparison between my Active DI and a commercial $200 Active DI unit. The $200 DI has a tiny bit more low end than mine, but nobody would even notice it if not pointed out.
Anyways, I don't know if that answered your question but that's how I do things. Whether it's right or wrong, who knows? Each person will have their own ears, their own personal biases, etc... See the June 2008 issue of AES Journal about how personal bias affects listening tests. (factors like price, look, brand, equipment model#, and also mood, music preference, nationality, etc.)