Audiophiles should learn from people who created audio

Having been involved in work on video projection engines(how individual circuits I gear affects final outcome), both analog and digital, the optical intricacies (glass coatings, lens design, etc)..the screens, their placement, their design at the (photonic/atomic level)..the interlacing and associated scaling algorithms, and so on....basically the entire pathway..even the entire package of the image capture devices (including their fundamental physics) ..the entire package, in all it’s minute and interconnected overarching intricacies -and many aspects that folks don’t generally understand yet....and then going back to how people see...how they see differently...

....I can say that visual people suffer from the same complex issues as do the audio people.

I've done probably...1000-1500 'single cause analysis' tests in the video work in these mentioned areas. At least double that number in the world of audio. This is why I'm (now) in my 50's and getting into finally publishing works as product. The apprenticeship was fairly long it seems.... I like pure research, and selling product, building product - kills the fun of pure research. It's a grist mill, business is..it can ruin a mind and a life with ...well..what we see here in these threads. I avoided it as long as I could.

@willemj  

I agree that intrumentation is far more resolving and reliable than human hearing.

However, it is still astonishing that we can hear up to 10 dB below the noise floor. Our hearing is very much like a spectrum analyzer. I dont subscribe to the idea we can hear the shape of wavefronts - this is nonsense. However, given enough audio signal it is astonishing how well we can work out the frequency content. 

In theory, and according to the designer, there shoukd not be an audible difference from Benchmark DAc 2 to DAc 3 but I hear something. Extremely subtle and I admit that I would not detect this in a blind test but with rapid A to B I hear something. Instrumentation of course sees quite a difference in THD+N performance but at a level that should be inaudible.

This is, of course, interesting because we know from the measurements that there is a difference. Methodologicaly the problem is that you do not believe it is big enough to show up in a double blind listening test, and it is beyond the level where by common consent it is believed the limits are of human hearing acuity. So it could just be expectation bias. How do we decide that what you hear is real?
My real concern is when measurements show that there is nothing there, or even, that the audiophile marvel measures badly and is still praised by golden ears.

Eek!