Yer foolin' yourself if you think you can reliably calibrate sound levels without a sound meter. Can't remember the source, but I read an article several years ago that demonstrated the need for this to be accurate.
They took a room, with two 5.1 setups. One was a high end speaker system (can't remember the brand). The other was a set of cheapo Rat-Shack small speakers. The RS speakers were calibrated to within 1 db. The others were left mismatched, but still within a small range (can't remember, something like 2-3 dB). Both sets were hidden from view by acoustically-transparent screens.
The participants consistently picked the RS setup as having the more convincing soundfield.

Note, this doesn't mean the best, most detailed sound. When watching a movie, you can have the most revealing setup in the world, but if pans change in volume or character as an object moves, say form left to center to right, the illusion is broken, and you hear the speakers instead of the movie sound.

So, if we can't even reliably calibrate speakers by ear to within 1 db, how can anyone be trusted to know that Amp A sounds 'more detailed' than Amp B, when they heard them on different days, different speakers, etc??

Today, any quality differences between high quality amps are largely neglible. The interaction between a given amp and a given set of speakers is far more important. All speakers have a widely varying impedance across the frequency spectrum. For example, my Martin Logan reQuests are 'nominally' 4 ohms. But, this impedance drops to as low as 1.2 ohms at some freqs -- damn near a dead short! This drastic change is enough to alter the performance of some amps, and even make some give up completely. So, even though an amp meets certain specs, it may fail when hooked up to systems like this.
So, while I might prefer Amp A over Amp B in my system, I don't think this applies to all other systems. Likewise, of the set of Amps that can handle difficult loads like MLs, I think the sound difference is very marginal, if audible at all.


[This message has been edited by bigmac (edited August 04, 2002).]