For me, rather, the question is whether your claims actually result in audible effects, since it would appear that with properly designed amps the distortion products are all extremely low; the feedback doesn't taper off as soon as your example states; and the higher order harmonics of distortion where feedback is lower are above 20kHz.
What is meant by 'properly designed amps'?
To prevent feedback from decreasing, you need a really high value of Gain Bandwidth Product (and for those that don't know, this is simply the frequency at which the circuit's gain has fallen to a value of 1, or 'unity gain'). To give you an understanding, let's say the amp has 30dB of gain, which is pretty common in higher powered solid state amps. That's a gain of 1000; if the GBP is 1 MHz, then we divide that by 1000 and we get 1KHz which will be the turnover frequency at which distortion will increase. Quite a lot of solid state amps have values like this that might be considered 'properly designed'.
You really need a GBP much higher than that! 10MHz might be a good start. Then with the same gain, the turnover frequency is 10KHz, so harmonics are far less likely to be annoying- except to a cat or dog (if they leave the room when you turn the stereo on...).
Put another way, feedback has been poorly applied in most amplifiers made, tube or solid state. This is for two reasons, the first just pointed out, the second being that usually the feedback is applied to a non-linear node in the circuit which distorts the feedback before it can do its job. FWIW opamps don't do this since the feedback is mixed with the incoming signal outside the actual opamp circuit. The result is correction that is far more accurate, and no surprise, its fairly easy to design opamp audio circuits that are very neutral, especially with modern opamps.
You can imagine that if distortion were rising with frequency in a preamp circuit, and then that was driving an amplifier with the same problem, the result would be quite a lot more distortion than the numbers (THD) would have suggested, and it would be mostly higher orders!
Going out on a limb here (mostly on account of the response this comment is likely to engender), blind trust in the numbers is no better and no different than blind trust in your ears- its still blind trust. Both can fool you (and quite often spec sheets are devised to do exactly that; thus might be actually be easier to fool those who rely solely on the numbers). That different amps might sound different isn't weird or religious, its a simple fact. You have to look at the distortion spectra, if distortion rises with frequency and if so, what frequency, does it rise faster after an octave or two and so on. Most of these measurements are simply never done, and worse there is a paucity of understanding of their significance. If two very different amps that have different cost and appearance have similar distortion spectra and the like, they are probably going to sound the same, since the distortion of the amp is also its 'sonic signature' (and in some cases, so is the output impedance) and if those amps happen to be the ones being compared, you might fool yourself into thinking that all amps sound the same.
Put another way, there is a direct link between what we can hear and what can be measured. Both the subjectivist guys and the measurement guys seem to hate that so I get a lot of pushback, but I've found that if you place belief above science you're likely going to make up a story that isn't true and then live your life as if it were true. It may well be that I am mistaken, that there is no link at all between what we can hear and what we measure; Occam's Razor and my experience in the audio biz of the last 45 years says otherwise.
I don't think this issue of subjectivist/objectivist is going away any time soon due to the lack of education and the lack of in-depth measurements. Its pretty safe to assume that threads like this will still be around 10 years from now. Enjoy the ride- its a long one.
