I have a suspicion, but would like to confirm it. Both
normal-texture:RGB -> Normal Map -> shader-Normal
and
height-texture:grayscale -> Bump Map -> shader-Normal
seem to be able to produce a similar rendered appearance. I assume that similar calculations happen in both cases? I further assume that normal-texture:RGB produces better quality, because more bits of information per pixel?
I am asking because it’s more practical in a few cases (i.e. text) to create a height-texture with any image editor, rather than hi-poly modeling and baking normals.
(Interestingly, after applying a bump map, the result can be baked to get a normal texture.)