As is often suggested on these type of threads, measurements don't tell the whole story. This should be obvious, as if they did, every designer would base their designs on the best possible measurements, yet many don't.
That's not to say that they ignore distortion, etc., but rather that the best measurements don't necessarily equate to the most compelling listening experiences.
There are many examples, but Nelson Pass has shared more than a few interesting, related anecdotes and opinions. Here's a quote from an interview:
I like measurements, and I use them all the time, but they don’t get the last word. Generally, there isn’t much conflict between what we measure and what we like to hear. I’ve spent decades working to correlate good sound with measurements, and we have a reasonable picture of what works, and you see that in our products. These represent our own listening tastes, and appeal to a large enough portion of customers to keep us in business.
With low-order harmonics, there is agreement that low-order distortion is much less offensive than high-order harmonics and IM (Inter modulated) sidebands, and this drove the development of the Threshold 800A back in 1975, where the nature of the distortion was considered as important as the size of the “single number.”
By 1991 when I started Pass Labs, I began to focus more on the specific character of second and third harmonics, and slowly settled into a character where negative-phase second harmonic dominates at low levels, segueing into symmetric third harmonic at higher power. The second fosters an illusion of expanded space and localization, and the third seems to improve dynamics. The distortion of these amplifiers is still quite low, but they are not sterile.
