I like most others here, don't have access to such equipment. If this data is so readily available, why isn't it routinely being published?
I might not completely understand what your suggesting, with re: to speakers and square waves, but to my knowledge the only readily available commercial speakers that can accurately reproduce square waves would not be considered "high efficiency".
Unsound, To answer the immediate question, IMO its because this information is not very convenient. From the rest of this quote I feel that you *might* be missing my point... I mention this test because it can be done with almost any speaker, as the 0VU value is not particularly demanding. It is just a very simple demonstration of a principle of human perceptual rules, and one that is easily duplicated.
IOW, its a proof. The point of it is that we use odd ordered harmonics to tell how loud a sound is. That's how our ear/brain system is wired and is true of all humans. So from my point of view, if our ability to tell how loud a sound is has to do with odd ordered harmonics (already present in the signal) then it is simple logic that we do everything we can to avoid distorting them since the ear is more sensitive to them then just about anything else. That of course is my conclusion and stand, but it is only simple logic.
Reducing odd ordered harmonics is also why I mess around with tubes, as its a lot harder to do the same thing with transistors. That is a preference on my part. However I do not like many of the colorations of tubes (read: lush midrange associated with 2nd harmonic); I prefer things to be neutral.
BTW its Norman Crowhurst, not Cromwell... If you look at his writings you find that he does not eschew negative feedback, but he also does point out its weaknesses. At one point he talks about how when feedback is added, the noise floor of the amplifier circuit then contains harmonics up to the 81st, plus inharmonic information related to intermodulations at the feedback node. In fact he points out that this *is* the noise floor of the amplifier.
(BTW, Chaos Theory also predicts this very same phenomena. If you read his books, you saw the formula for feedback in an amplifier too...)
Now I am not sure exactly when the perceptual rule known as 'masking' became understood, but this is the rule that made mp3s possible, and is the idea that louder sounds mask the presence of quieter sounds. The masking rule is why we cannot hear sounds below the harmonic noise floor created by feedback.
Now it happens that if you have a noise floor consisting of hiss alone, that you can hear about 15-20 db into that noise floor (I suspect that this peculiar exception to the masking rule is the result of necessities of evolution, as wind and hiss are very similar). Whatever the reason, this does seem to be one explanation of why an amp without feedback can seem to have more low level detail than an amp with that employs feedback.
BTW, I think we are getting a bit off-topic- is this a subject for a different thread?