3 ms·

The way I would describe things (being a programmer) is that my actual eye and my "mind's eye" are using different display devices. The actual eye comes out as

The way I would describe things (being a programmer) is that my actual eye and my "mind's eye" are using different display devices. The actual eye comes out as a PNG in its full vividness, but the "mind's eye" only "sees"... something like an SVG? It definitely comes across that the "mind's eye" is encoding visual information differently, and words like "wireframe" or "storyboard" (for entire scenes) seem to be a better way of describing it than anything else, although I wouldn't call them great descriptions [1]. The most notable difference is that there's no "visual" encoding of color in my "mind's eye"--all that's there is "this thing is filled with color 3df76346cdcad147dcf0efc07e347bd0"--and when I look at a color palette with my physical eye, I can tell you if that shade is "color 3df76346cdcad147dcf0efc07e347bd0" or not, but if it isn't, I'd struggle to tell you what needs to be done to match the color. It really does seem to be some sort of one-way hash function in encoding color; I can "visualize" a red square or a green square, but something like a color gradient just makes my "mind's eye" go ¯\_(ツ)_/¯. [1] I think the general idea to convey is that details can be very flexible. For example, as I wrote this, someone turned on their car (remote start, I presume), got into it, backed out of their parking space, and drove away, and I only know about this through hearing it. I can visualize all that in my mind's eye without having to decide if the car is a sedan, an SUV, a minivan, a convertible--knowing that it's a car is sufficient detail, and the mental imagery doesn't give it any more detail. Then I can decide that it's an SUV, one of those that has some piping in front of the metal grill, and make a pretty detailed image of that front of the car. But the mind's eye usually doesn't bother doing anything beyond a very low level of detail unless prompted.

Twisol 5y ago
Just to add my own anecdatum here -- my mind's eye definitely encodes color, and I can relatively easily match a color in the real world to one in my mind's eye. Color gradients are instant for me, although I summon the incorrect red-green gradient caused by incorrect gamma correction to mind by default rather than the more pleasing one with correct gamma correction. In the case of the red-green gradient, I'm not exactly computing the colors intentionally; I've just seen quads with interpolated coloring often enough that those memories feed my mind's eye. I can fairly easily visualize a blue-purple or orange-teal gradient, too, but I can't attest to their accuracy. > But the mind's eye usually doesn't bother doing anything beyond a very low level of detail unless prompted. That's my experience, too. Reading your description of somebody driving away in their car, I was able to easily reconstruct an imagined scene in my mind's eye, but I naturally filled in the missing details with a kind of "default" -- camera position and angle, whether the parking is covered, the car is like the one Dexter's dad drives in Dexter's Laboratory (but not a cartoon), etc. The details are... there-ish, if I care to query them, but for a simple replay of the scene, my imagination doesn't bother rendering them to any significant fidelity unless I intend to.

Twisol 5y ago

Just to add my own anecdatum here -- my mind's eye definitely encodes color, and I can relatively easily match a color in the real world to one in my mind's eye. Color gradients are instant for me, although I summon the incorrect red-green gradient caused by incorrect gamma correction to mind by default rather than the more pleasing one with correct gamma correction. In the case of the red-green gradient, I'm not exactly computing the colors intentionally; I've just seen quads with interpolated coloring often enough that those memories feed my mind's eye. I can fairly easily visualize a blue-purple or orange-teal gradient, too, but I can't attest to their accuracy. > But the mind's eye usually doesn't bother doing anything beyond a very low level of detail unless prompted. That's my experience, too. Reading your description of somebody driving away in their car, I was able to easily reconstruct an imagined scene in my mind's eye, but I naturally filled in the missing details with a kind of "default" -- camera position and angle, whether the parking is covered, the car is like the one Dexter's dad drives in Dexter's Laboratory (but not a cartoon), etc. The details are... there-ish, if I care to query them, but for a simple replay of the scene, my imagination doesn't bother rendering them to any significant fidelity unless I intend to.