Psychoacoustic so, why should we care.
If you are meaning we fool ourselves while listening, I would not care. Someone recently had the opportunity to ask the "experts" at least 4 times this question concerning dbt/abx testing. (This was on another forum.) There is more information though out this post concerning "scientific" dbt/abx testing.
From another forum:
How can one obtain 90% confidence when 50% of a group are positioned in bass increasing areas, and 50% are in bass decreasing areas?
Every single "expert" dodged the question, time after time. Never did get a response.
Of course, one has to also address variable confounds such as cochlea fatigue (tired ear), habituation to stimuli (after 2 or 3 ABs, will tend to sound the same), spl, type, quality of recordings, bandwidth, echoes etc.
If one checks out books, websites, only "sight" is listed as a variable confound, so one will never perform a correct test procedure, so the conclusion will virtually always be no sonic difference. Just a typical marketing tatic.
Measurements: which ones really matter?
I suppose power output, frequency response because some are down a db at 20khz etc. Harmonic distortion since a few might be high at 1 watt output. Also depends upon which harmonics are involved.
Is flat response desire able or even musical?
I would say flat, but just my opinion. I have a caveat, that +/- 0.1db means virtually nothing. Varying the FR by 0,2db is like varying the frequency response only around 54db down from the fundamental.
Personally, I am tweaking the frequency response (FR) 1 part in ~200,000 on my speaker xovers (which equates to about ~105 - 110db down), and less than 0,2% on cap values in my mono block amps.
I can easily see why components "sound" different.
Speaker design?
I have found 2nd order xover is about as far as I wish to go, no 3rd order. Just too difficult to design 3rd order to sound right.
Are shielded interconnects necessary or even matter?
I do not use shielded ICs even for my turn table to phono stage. At full volume just a slight hum at the driver cone. And I live ~6 miles from megawatt tv stations. But I suppose it would not hurt to use shielded, but I prefer both to twisted.
I am sure most are probably not concerned with such minute sonic changes, but it is good to understand how sensitive our ears are. I believe that our ears are more reliable than dbt/abx tests, which are inherently flawed virtually all the time.
Cheers
Steve