Artists Don't Know Better Than the Rest of Us How Things Look
It's almost a ritual, in discussions of the phenomenology of vision, to praise "artists" -- meaning those in the visual arts -- for having an appreciation of visual phenomenology that most of the rest of us lack. However I believe that the truth is the reverse.
Thomas Reid is typical:
I cannot therefore entertain the hope of being intelligible to those readers who have not, by pains and practice, acquired the habit of distinguishing the appearance of objects to the eye, from the judgment which we form by sight of their colour, distance, magnitude, and figure. The only profession in life wherein it is necessary to make this distinction, is that of painting. The painter hath occasion for an abstraction, with regard to visible objects, somewhat similar to that which we here require: and this indeed is the most difficult part of his art. For it is evident, that if he could fix in his imagination the visible appearance of objects, without confounding it with the things signified by that appearance, it would be as easy for him to paint from life, and to give every figure its proper shading and relief, and its perspective proportions, as it is to paint from a copy (An Inquiry into the Human Mind, 1764/1997, p. 82-83).
Now this much I'll grant Reid and others who share his view: Traditional, representational painters have the difficult skill of rendering on a two-dimensional canvas an arrangement of paint such that it produces for the eye an arrangement of light importantly similar to what would be produced by the actual three-dimensional scene they are rendering; and it takes much practice to see outward things in terms of how they can be presented on a canvas, for example foreshortened and rendered in the right two-dimensional shapes. But that skill is not the skill of appreciating real visual appearances.
For one thing, the view makes no geometric sense. Three dimensional scenes cannot be rendered in two-dimensions without geometric distortion in size and/or angle -- distortion that becomes more evident the greater the visual angle encompassed. This is why there is always something a little wrong with panoramic photographs. This geometrical difficulty could be avoided if artists drew on concave semispheres instead of flat rectangles. But they don't; and they'd have to relearn the rules of perspective to do so.
Even setting that issue aside: We should not infer from the fact that to create a sense of realism in the viewer an artist must color shadows in such-and-such a way that we really visually experience shadows as colored in that way. We should not infer from the fact that light, and water, and distance, and motion, can be rendered a certain way on canvas to the fact that our visual experience light and water and distance and motion matches such renditions (e.g., motion as either a series of freeze-frames or as blur). The painter learns the skill of seeing the world in a certain way for the purpose of a certain technique, not the skill of apprehending our visual experience as it is in itself.
Since most visual artists don't seem to appreciate this fact, their reports about their visual experience are likely to be less accurate than the reports of non-artists -- distorted by the false assumption that the world as seen for painting is the world as seen for life.