Hello dear Blender community,

I'm working since a few days a good way to create heightmaps from meshes. Everything works fine, but when rendering (eevee) the colors are always wrong. White should be black and vice versa. If I now add an invert node in the compositing tab, the image is totally blended and you can no longer see any structures.

Picture 1: Without invert node, you can clearly see the buildings.

Picture 2: With invert node, hardly anything can be seen of the buildings.

What am I doing wrong? Shouldn't it actually be totally simple?

Without invert node, you can clearly see the buildings.

With invert node, hardly anything can be seen of the buildings.

  • $\begingroup$ First of all, heightmaps don't need to rely on what is visible to your eye. As long as the values are correct, you can use them as heightmap. The other things Gamma correction and other things related to how colors are displayed on your screen. I guess you have the default Color Management settings enabled in your Render Properties, which means View Transform > Filmic and Look > None. You can change the contrast by changing the Look or the Exposure or Gamma, but for a heightmap it might be best to not use Filmic at all but set it to Standard and None. $\endgroup$ Jan 10 at 14:52
  • $\begingroup$ The other thing is, you are using the Mist pass. This is highly dependent on what distances you have set for Start and Depth, because the 0 to 1 greyscale range is distributed over the this dimension. Everything closer will all be 0 and everything further away will be 1. The Mist Pass doesn't output values below 0 or above 1, no matter how close or far objects are to the camera. So simply inverting the colors could mean that values outside this range might get lost because the values are clamped between 0 and 1. $\endgroup$ Jan 10 at 14:56

1 Answer 1


Colors in Blender are in linear space. This means the color values are proportional to the actual amount/intensity of light, eg. 0.8 is twice as much light as 0.4. But the human eye has a non-linear perception of light, it is much better at distinguishing colors near 0 than near 1.

For example, the colors 0.1 and 0.08 are easy to distinguish, but the difference between their inverses, 0.9 and 0.92, is nearly invisible.

So if what you care about is the numerical values, there isn't necessarily anything wrong, it's just a matter of visualization.

  • $\begingroup$ Oh, okay! I understand. Thank's a lot! :) $\endgroup$
    – Knudsen
    Jan 10 at 17:53

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .