Thanks @holmz for the various very good points and suggestions. Just a clarification on one point, as this is one of the things I’m trying to understand here:
I doubt that the amps are any “faster”, as most amps play 20kHz just fine, and nothing is going to arrive faster from a 2kHz signal, than a 20kHz signal.
With fast I mean that the transients, the typical example being the hit on a snare, are quick and powerful. So in a graph measuring the sound pressure, the peak would be higher and narrower on a "fast" amplifier and lower and flatter on a "slow" amp. The sound is more compressed in a "slow" amplifier and thus the hit of a snare or the initial transient of a piano note will be felt as softer by the ear.
I understand that there are other factors discussed that heavily influence the perception, but I do believe that this is part of the equation. Low powered amplifiers would in my understanding be therefore more suitable to avoid the problems I’ve had. It is not so much sustained notes that cause issues for me, but the percussive sounds and initial attacks on many instruments.
I get the description of the hypothesis.
If the speakers were hard to drive then more instantaneous current would help if, and only if, one did not have enough current to keep up.
It would be like speaker compression, but I suppose for the amp.
And I suppose one could (maybe) see if it shows up in the impulse response of the system.
But it might be easier to have the same snare drum hit recorded from the 4 amps (3 good and one bad one), and compare them in the time domain?
If the bad speaker was high amplitude than the other three, then it would show as you describe.
My hypothesis is that the bad sounding amp likely has more harmonics or IMD.
So I am thinking exactly along the lines of what @helomech mentioned:
I recommend looking into separates. Pre-owned Parasound A21 amps can be found for ~$1200. Their distortion profile is primarily 2nd order, which is very rare among amps below $5K, and output enough power to provide plenty of headroom for nearly any speaker. Pair an A21 with a low distortion preamp like the Topping Pre90 and you’ll have a very smooth and low-fatigue setup that can takes up little more real estate than an integrated.
(And user atmasphere talks about in various threads.)
… So it might be better to take the snare drum recording and do the comparison in the frequency domain.
Of course going to that effort one would want to do both.
I am not sure there is any way to reach an objective conclusion as to what the causal mechanism is for what you are hearing, without doing some measurements, But I also know that I also usually have problems with the SS gear showing up only in those sorts of sounds. And also I find sibilance to be distressing to me ears.
If it is as you say, then I would think that it would show up more in music with has a high dynamic range, where the peak level would be much higher than the RMS level. And then on music that’s more compressed, we would not have the snare drum attacks at the higher amplitude to begin with.
Have you noticed something like ^that^ which correlates with the bad sounds?