I now remembered why I know about DOF issues: years ago, there was too much of it! I used to have an Epson flatbed scanner that had a slide adapter, i.e., a replacement cover with its own light source from above. First I noticed a strange, faint, repetitive pattern on all my slides I scanned. I could not figure out where that comes from, as the light source was covered with a thick frosted glass. The texture of the glass did not look like the patterns I was getting on the scanned images, so I dismantled the cover in search of the cause. I found it had two or three fluorescent tubes as source of light AND a metal mesh between the tubes and the glass, the purpose of which I guess was to defuse the light even before it hits the frosted glass.
Now, those who use dedicated film scanners would know that they have extremely limited DOF... so much so, that if you do not use a glass carrier for the film strip or slide, the curvature of the film would result in unsharp edges. Imagine then my surprise that my flatbed scanner would go beyond the surface of the film, beyond the frosted glass, and actually pick the pattern of the metal mesh behind. Then I did a different experiment and scanned a software box, standing upright: you could not only read the bottom of the box that was in direct contact with the scanner glass, but you could read the letters on the box sides all the way to the top!