I suppose we should thank Mr Austin for saying the quiet part out loud: his statement confirms that he accepts that there are psychological as well as electro-acoustic aspects to musical reproduction and listening. He should go one step further and recognise that his own chosen role is far more in the former area than the latter. He doesn't though: he would rather attribute the transcendent, emotional feelings which music triggers in the wonderful, living, breathing human mind to "unmeasurable" qualities of a box of cold, dead electronics. That's "gear fetishism"!
I would never criticise someone for how they gain pleasure (I'm a smoker, how could I?) but examining the mechanisms by which pleasure is gained, and looking into which can be generalised and which must remain purely personal, is not criticism or attack.
We are right to point to the placebo effect when judging the results of sighted listening tests: it is a far simpler explanation than inventing stories about measurements not yet possible (or even about measurements which will never be possible!). It is also something that can be tested through blind testing.
But what if, having failed to distinguish between two components in a blind test, an individual still consistently prefers one of them in "normal" listening? In such a case, there is no shame at all in recognising that non-auditory psychological factors are at work, and yes, the music *actually does* sound better in the mind of the individual concerned. I was going to say this is no problem at all, but it can be - when snake oil salesmen and other fraudsters use the power of suggestion to take advantage of people.
Then there are the things that are measurably "lower fi" than others, but are nonetheless preferred in subjective review. If such a preference survives a blind test, are the measurements irrelevant or wrong? Of course not. What has been demonstrated in this case is a *personal preference* for certain distortions of the musical signal, which is fine, until such a subjective preference is generalised into statements that the component in question is objectively *better*.
@GXAlan mentions this above, that certain objectively inferior measurements can be preferred. Again, no problem, but these days it is likely to be far cheaper and more reliable to use DSP rather than struggle to put together a collection of differently-distorting components which even if successful would only ever really work for one person in a one particular room. I've never understood people who obsess over keeping things bit perfect in the digital domain and then "tune" the analogue signal with analogue components chosen for their ("euphonic") distortions.
