Thanks Mart68. These questions are not easy to answer, which is one reason I find them so intriguing.
Any conclusion for "what it is reasonable to believe of our sense perception" will have to account for two seemingly diverging facts:
1. Our perception can be unreliable. (See: optical illusions, and any number of bias effects)
and
2. Our perception can be reliable. (See: our ability to successfully navigate the world through sight and sound, drive cars, identify people objects reliably etc)
It doesn't matter what domain we are in - one that deals with visual perception, or one that deals with sound perception, we are not going to arrive at a coherent
picture by over-emphasizing one or the other of those two facts. If you over-emphasize the reliability of our perception, miss all the ways it is unreliable, you end up with an untenable confidence in your subjective experience. You end up believing all the audio snake oil. If you over-emphasize our perception being unreliable, basing your view on a hyper-skepticism, you will likewise end up in an untenable position because you haven't accounted for the countless instances in which our perception proves reliable. You end up throwing the baby out with the bathwater by disavowing useful information based on sighted perception.
So any account for the worth...or worthlessness...of sighted listening to speakers will have to ultimately be some synthesis of the above two facts, not an emphasis of one at the expense of another. It seems to force at least the general conclusion that our perception isn't fully reliable, but neither is it fully unreliable. Working out the details for justifying belief in any conclusion from our perception - even synthesizing the results of blind testing - is the tough part.
I've given my own pragmatic "solution" before (the heuristic "extraordinary claims require extraordinary evidence").
But I think we are on the same page in wanting to synthesize the above two facts, to make the case for the use of blind testing. A synthesis of #1 and #2 would generally entail: blind testing results are
more reliable for determining the sound people prefer, though that does not leave sighted listening as
wholly unreliable for the same purposes.
but since the variations introduced by sighted listening are a construct of the mind they are not permanent. One day you may really like the speaker, the next day, not so much.
At least if you selected the speaker blind then you know you have a speaker that you genuinely like, even if in normal use your perception of its sound varies due to sighted bias, your mood, etc. That still seems to me like a better baseline to work from.
Yes that does sound, at least at first glance, like a reasonable conclusion, when trying to work out the usefulness of blind testing for speaker preferences.
I've thought through that before. And one reason I've found it somewhat compelling is based on my own experience blind testing. For instance I've mentioned before on the forum that when I switched music servers it *seemed* at first like the new one was a touch brighter sounding. That didn't make technical sense to me so I blind tested them, heard no difference, and from then on didn't worry about it and didn't perceive this "slight brightness" to the new server. So on first glance, that type of experience seems to justify your line of reasoning: That the "errors" in our perception are unstable, but the actual SOUND is more stable, so the actual sound (as can be determined in the blind testing) is ultimately of greater relevance and influence.
But..note what that necessarily implies for our listening pleasure: The Sound comes to us only through our perception. To say that there is something about The Actual Sound that will ultimately prove more stable is to necessarily say that we will find some stability in OUR PERCEPTION of the sound. And, presumably, this is because over time our perception will be accurate enough to perceive How It Really Sounds - that would be the explanation for why the blind tests could result in ultimate listener satisfaction. So that STILL ultimately results in verifying our sighted perception!
So back to my music server issue. The blind testing did help me
quickly resolve my perception of added brightness (I did this within a couple of days of "hearing" the issue). But I think it most likely that this would have faded pretty quickly to my baseline, since the sound was in fact not different, I'm sure I would have ceased thinking it was "bright" pretty soon. In other words: even in sighted conditions, my perception of the real sound would tend to win out over time.
Ok, so should we then conclude "INITIAL impressions of sound under sighted conditions are unreliable?" (E.g. auditions at an audio showroom). Wish it was that easy. Except that too becomes problematic if pushed too far, because there are so many examples of initial impressions of sound being correct/reliable.
Also: what do we do with the countless examples of persistent preferences and perception? Over the years tons of audiophiles have found great satisfaction using speakers that would not pass the ASR/blind testing criteria - avid Klipsch, Maggie, Horn, Zu owners and on and on. Blind testing suggests they shouldn't really be so enthusiastic about those brands (and many audiophiles have heard other more neutral speakers in their time, yet still found happiness with those speakers). And yet, there's plenty of evidence their perception of the sound persists year after year. If we go with the conclusion that blind testing is useful because The Real Sound will ultimately be more stably perceived over time than our fleeting biased distortions, it suggests these speaker owners are reacting to the
actual sound of their speakers and truly do love the sound. That it's not all pure, unreliable delusion.
So...life's messy. We can take one pragmatic position or another to get us through the messiness, but we shouldn't pretend it's all be tied up in a nice bow when it hasn't been.
So I acknowledge your view for justifying the usefulness of blind speaker testing. I don't see it as unreasonable.
As I've said before, my approach to synthesizing the two facts (perception can be fallible/can be reliable) is pragmatic: scale my confidence levels to the available evidence and plausibility, employing heuristic tools like "extraordinary claims require extraordinary evidence." If my pal came back from auditioning two different speakers and described the sound, and why he preferred A to B - e.g. "speaker A was too bright/too lean in the bass etc" it is entirely plausible he really heard those differences, as speakers can differ in those audible parameters. In that sense it is an "ordinary claim" that does not contradict known science or engineering.
If I wanted HIGHER confidence levels in his perception of the sound, I'd want to employ controls for possible sighted bias. But on a pragmatic level, it's fine to provisionally accept his claims, while acknowledging the lower confidence levels and room for the influence of bias. If he was making technically implausible claims about the sound of different USB cables, I'd demand more rigorous evidence before accepting what is an extraordinary claim.
If this type of epistemic pragmatism wasn't reasonable, I don't see how we couldn't justify our daily inferences from our perception (which aren't done under scientific controls), let alone communicate with other human beings.
Cheers!