Yes, we are not making ANY scientific claims regarding human hearing capacity. This is not our specialty. The ony claims we make are related to our units conventional specifications, which as Amir noted, are reasonably consistent (i.e., accurate and repeatable).
The Cyril Bateman articles do analyze passive R and C and show differences at those levels.
Our listening tests confirm the preference for a ranking similar to what his results showed - i.e, in general, the resistor and capacitor technology that he showed as having "better" performance, isolated down at those levels, was directionally similar to what we perceived in our testing.
Therefore, our only sincere hypothesis is that the human ear must be able to perceive distortion well down into the noise floor (this is generally known but probably conventionally accepted as 10 to 20+dB into the noise floor. We hypothesize that it may be greater than this.
D.E.L. Shorter's 1949 article is available at:
https://www.bbc.co.uk/rd/publications/rdreport_1949_30
Attached are our original calculations based on Shorter's formula. To our knowledge we have not seen them posted elsewhere with an actual derivation by harmonic. They have been peer reviewed by an ASR member and the calcuations shown have small insignificant rounding errors.
Essentially they propose a weighting formula that increases the upper harmonics (above the 2nd) in total contribution to a new "weighted THD" metric.
The top of page 611 of RHD4 shows an analysis of 5 BBC broadcast transmission plants, some AM, some FM, whereby the ear's assessment of the "impared-ness" of the plant was ranked in correct order with the "measured result" when the "measured result" was via this new weighting. It is the "more drastic weighting" as mentioned by the editor of RDH4 on page 610.
We don't refute Shannon, et. al; we just question the applicability to this discussion.
Keith Richardson
Director of Engineering
Darlington Labs LLC
www.darlingtonlabs.com