Change of hue near shadows My brother is an artist and he told me his paintings look more natural when he slightly change the hue (color) of the surface near the transition between shadow and light. Here's three examples: (1) without this effect, looks artificial, (2) with this effect, looks natural, (3) with overuse of this effect, looks strange. Assume pure white light (this hue change is only caused by surface properties).

What would be the physical explanation for this effect? The hypothesis I got is that when an object is painted (semi-transparent surface over pure reflective surface), if the rays reach the surface in a near parallel direction, then we get some refraction. If the reflectance of the surface is not a pure wavelength, then we can see some hue drift due to that refraction. Does this make sense? Is there a simpler explanation?
 A: Refraction has nothing to do with the effect.
There is a very straightforward explanation of this effect in terms of simple geometric optics, and the simple observation that the size of physical sources of light is always larger than zero, compared to mathematically ideal point-source.
A diagram is probably worth a 1,000 words of explanation here:


Edit in reply to comment: Physically, a yellow object in shadow will be perceived as the same color because a shadow is the absence of light as opposed to black light. So you merely see less yellow instead of a mixture of yellow and black. To reproduce the effect of shadows in paintings, you have to mix black pigment with the color pigment of the object you're shading. But by adding black to yellow you've inadvertently shifted the yellow hue in the greenish-blue direction (see the first image you provided in your question). Painters solve this problem by mixing in pigments which shift the hue in the complementary direction (in this case purplish-red), and the idea is that two shifts of cancel out to yellow again.
This is why painting is hard. The optics is easy though, and it should be emphasized that this part of your question belongs to color theory. The subtlety here is of paints, not optics.
A: Color perception is rather complicated. An object's perceived color doesn't just depend on the light frequencies it emits/reflects. The light frequencies entering the eye from the area  surrounding the object also play a role. (See "color constancy") This is the basis for some color illusions.
So the effect might be due to how our brains process images rather than to enhanced or diminished reflection of certain wavelengths off of a surface.
