Wyn: I see you added a note on "the positions of Nelson Pass"; if you keep it, you probably should elaborate which ones. He's got a few.
Not a critique of the paper, but more an attempt to sort my own thoughts on the distortion spectrum stuff:
Everything in the mechanical realm we live in has overtone characteristics with an overall decaying applitude of the overtone series; of purely mechanical reasons, you rarely find stuff in the nature with audible overtones above about 5th or so. Some man made materials manipulate the limits of this, and have often been used for creation of specific sounds - mostly as metallophone instruments.
The basis for the "modern" take on distortions is that the ear itself has same resonance mechanisms, producing its own overtone series from the received input – by definition a harmonic distortion. These series of course follow the the same average distribution as every other mechanical overtone series. But the interesting thing is that the auditory system – the ear and the brain – has a built in compensation for this internal distortion, effectively masking it so it won't distort our perception of our surroundings. And since there is no "distorted" sound in nature (only real sounds with their actual overtone series), the same mechanism also works for masking harmonic distortion from reproduced sound - to a limit.
When reproducing sound in audio gear, transducers give the same kind of overtone distribution; the overtones generated in loudspeakers, cartridges and microphones (and transformers?) all follow the same natural distribution. But this distribution becomes upended when passing active electronics, most markedly in the sort Wyn describes as having about 15 dB of feedback. This gives an "unnatural" distribution of the generated overtones, with a higher level of the higher order overtones. Since the ear isn't built to handle this kind of sound spectrum, it sounds wrong, in a way we don't even have an established vocabulary to describe.
One effect of this is that the concept of total harmonic distortion becomes more or less meaningless, since it hides away the spectral distribution of the distortion. For instance, this explains why the objectively high distortion measurements from turntables and loudspeakers are so much easier to accept than the miniscule distortion from many transistor amplifiers. And also why distortion specs say so little of how good or bad audio gadget really sound.
A couple of decades ago these distortion effects were barely measurable, being hidden behind the total harmonic distortion specs, or perhaps "visible" in the deformation of an oscilloscope trace. Indentifying them instead required a spectral analysis.
The various metrics Wyn touches on, the cheevers, gedlees an such, are all methods for creating a simplified "weighted" spec, with a higher correlation between specs and subjective perception. The resulting specs are based on a combination of several different kinds of criteria, including spectral analysis and psycho acoustic theory on how diffent spectral distributions are perceived. This makes the proposed metrics very abstract, and detached from both the actual measurements and any normally relatable sensual experience, which of course doesn't make them attractive to anybody without first reading up on a lot of theory.
As for the next step, how to use this knowledge for actually designing good gear, it all becomes even vaguer. The most common use has been in giving more slack to the acceptable levels of lower overtone distortions, since those don't sound as bad as the higher ones. I think a diluted version of the perspective is most popular amongst the valve, non-feedback crowd, and of course has been popularised by Nelson Pass, cleverly using it for spicing up the sound from some of his gear.
I think Wyn may be the first designer instead combining this perspective with a high feedback design, giving the possibility to not just create a more benign distortion spectrum, but to push down the entire spectrum under an extremely low noise floor.
Of course I'd love to see a power amp design based on the same ideas.
