Indeed. Science has demonstrated that domestic dogs have evolved to talk to our emotions and get us to feed them. I’m happy about that, I like dogs.
As regards lenses, Zeiss once told me that in the end it works as a lens/sensor/processing combo, or it fails. It’s not numbers, it’s psychophysics, an understanding of human perception. In hifi or in photography the human is the ultimate judge, not the graphs.
I have no problem with the idea that when it comes to matters of sensory perception, humans matter more than measurements and numbers. But only in a very caveated way.
The job of the measurer is to get at the heart of what factors influence perceptions and measure those ie model the human perception accurately. If they can pull that off, the measurements become more reliable than human perception itself because perception is variable - it can be influenced by external factors, such as mood or environment.
That is the problem to me with subjectivists: they are correct in demanding that a measurement that doesn't correlate with human perception is not useful for judging the factors a device needs to have high fidelity. Working with such measurements is barking up the wrong tree. But they are dead wrong to reject
all measurements just because
one measurement isn't appropriate. If we have the
right measurements that
correlate closely with human perception, the machine is usually much more accurate and consistent than subjective opinion. And on average you get better results for more people if you design based on the measurements, than if you design around one person's subjective view.
Take the CIE colour model thingy (I forget what it is called). It is a colour model for average human colour perception. It is based on measurements, so is consistent and reliable. But it is also based on actual testing of a representative sample of humans, so it correlates very closely with what the average person perceives. It is, therefore, a better basis for designing the colour response of a camera system that needs to appeal to wide number of people than, say, asking me my opinion on what nice colour is.
The hifi world is full of problems because so many people believe good sound can't be measured, only experienced. Unscrupulous manufacturers take advantage of this naivety by designing bad sounding gear that measures poorly on well correlated metrics, then using marketing lies to appeal to customer prejudices (previously established by earlier marketing). It often turns out that people can fool themselves (temporarily at least) into preferring gear that sounds worse, because it sounds
different ie bad, and stands out against well designed neutral gear that all sounds the same. It stands out so must be better, is the thinking. They then defend that choice (confirmation bias) against facts because they are committed psychologically to their opinion. The lie to all this is exposed in double blind testing where broken gear no longer sounds good.
Blind testing is very important when it comes to identifying and characterising human perceptions. It removes additional cues so the only information available is the factor under test. The subjectivity tends to vanish when there are no additional cues to bias opinion, and under these tight constraints human perception tends to become more accurate and more consistent and reliable. That is why blind tests are so good.
Hifi nuts (of course) and homeopathy fans (of course) hate blind testing alike, because it keeps exposing that their cherished beliefs don't stand up in a neutral cueless environment, where the only information available to test subjects are the facts under consideration. Human biases taint everything, which is why objective measurements are so useful. It's easy to have an opinion, much harder to prove it is correct...
This is an interesting demonstration of how subjective impressions (that most people rely on for their beliefs) are not completely immune (as some people like to claim) to being broken down into numerical components, combined and modelled in ways that actually explain those perceptions quite precisely (assuming a good model).
Lens character isn't some kind of mystical characteristic that can't be explained and just has to be accepted as magic; it's actually a combination of the effects of separate modelable aberrations and other lens parameters. There is no magic if you are able to measure and think in the right way; everything really is explicable with science and numbers.
As Feynman famously argued, science doesn't detract from the beauty of nature, but rather adds to it by revealing deeper layers of complexity and wonder. Understanding the scientific principles behind natural phenomena enhances, rather than diminishes, the appreciation of their beauty. Yet so many people are willing to reject understanding, reject the maths, because they believe that, somehow, intuition is more beautiful.
I think that the reason for this is because proper understanding requires thinking very hard, and thinking very hard is hard work. You mentioned something about people not being willing to think hard enough about things. I think that is dead right, but it is understandable: anything that is hard work uses energy and people evolved in an energy-poor environment, so careful hoarding of energy and the reluctance to expend it is an utterly natural and built-in part of human nature. Of course it is, laziness is a survival characteristic.
Watch a pet dog zippering back and forth sniffing out their environment and compare that to the dead straight path of a coyote. One is permanently fat, the other permanently starved. Of course we are all lazy by nature...
