Indeed Womba & LTig! It's been explained on ASR many times:
- Only a properly conducted listening test is worth anything (double blind and level-matched), and they are annoying and difficult to undertake!
- Cognitive biases cause HAVOCK with sighted casual listening tests, for everyone (me included) because we are human and ALL subject to these biases, so they have zero value (ie results cannot be replicated).
- We can measure everything that can be heard, this has been supported here many times, there is no magic, there is nothing that we can hear that cannot be measured, so measurements capture everything that we can hear (even more). Properly conducted listening tests have verified this.
Scott Wurcer designed the 797 opamp (sorry for outing you Scott) and knows more about them than anyone else on this forum, probably more than most people on the planet. Ask him about opamp rolling. Yes pma there are differences between opamps, but a competent designer today will select the optimum opamp for a particular part of a circuit and substituting another will, 99% of the time, offer no value but it will sometimes give worse performance.
A whole lot of the performance of a piece of modern equipment will depend on the circuit layout and the details of the components chosen. Modern engineers are often very very good. Unlike 40 years ago when substituting a high-performance opamp could probably improve the performance, that is no longer true. Bruno Putzey's balanced preamp/volume controller that was in Linear Audio (and online in EDN I think) is a case in point, substituting other opamps will almost certainly degrade the performance.
If you have something 40 years old and swap the opamps you might, might, improve the performance, but a higher performance opamp may also break into oscillation in the MHz region etc etc (which could present in the audio band as noise) and actually degrade the performance.
If we practiced medicine the same way as subjective audio reviews we'd still be selling snake-oil to cure cancer...