In a real world environment, in which I have freedome of movement and the ability to see the light souce, I've no problems correctly interpreting lighting and shadows of of surface features. However, those are known features, e.g. I put them there, or have a lot of experience of seeing them.
The kind of "stubborn perception" of light source position is only something I experience when seeing 2D images of features that I have no direct experience of, so don't know for certian where the light souce is, or that nothing other than areas of light and shadow, to inform my perception (e.g. no lense flare or glow form a light source).
My theory that Photoshop layer effect defaults might have bent my mind, seems like I was wrong. I just checked and top left isn't the default in the current version I used (which has it at center top). Possibly it was the default in the older Photoshop 5, that I used for about 2 decades (I make no apologies), that had a top left lighting angle default. However, I'm not even sure of that now (and can't summon the energy to install it right now to find out), as it may have just been my default behaviour to change it from top center, to top left.
I did a little looking around, to see if there are any conventions regarding lighting in art, that I might have picked up on over the years and built into my own perception and there seems to be one, in Western art at least, for light being from the top left.
designsynopsis.wordpress.com/2021/07/08/1070-top-left-lighting-conventions/
frontiersin.org/journals/psychology/articles/10.3389/fpsyg.2018.00454/full
This is the kind of thing I meant when I said in my previous post about "cultural assumptions of "light is from the left"
To complicate things just a little more, I did a little experimentation, after horiziontally flipping the pair of images at the top of your post my ability to flip my perception of lightng direction (so seeing a crater or dome), varies with the apparent lighting directions, available.
Here's the image I made, with all 4 flips in it.
View attachment 87266
Image 1 - At a glance this is a dome, and after a lot of effort I can see it as a crater.
Image 2 - At a glance this is a crater, I can flip this at will, roughly about once per second.
Image 3 - At a glance this is a dome and it takes huge effort (hardest of all 4 images) to see it as a crater. It takes a long time to see it as a dome and as soon as I do, it flips back to a dome, a moment later.
Image 4 - At a glance this is a crater and after staring at it for a few seconds I can start flipping between dome and crater, but not as quickly or as easily as Image 2.
As an extra tidbit of information, when I manage to change my perception of the hardest to flips ones (1 and 3), my right eye (the one with nearest to normal, uncorrected vision) gives me a "stop doing that" sensation... so I stop doing it. For the others, no sensation like that.
One thing that I noticed, is that to flip my perception of any of these, it's much easier if I "soft focus" at the center of each image, as if trying to take in the structure as a whole and not to stare at specific features, as doing that makes it much harder to flip.
It's all quite annoying really... dammit brain, do what I tell you
