I'm trying to figure out the various image settings and how they relate to a bump map texture. Creating the image with "32-bit float" certainly helps give smoother results, but it seems like enabling "float buffer" does the same thing. Is there a difference between these settings? If so, what is it? Which is better suited for use in bump and height maps?

EDIT: The "float buffer" setting is found in the UV image editor's image settings (right under the resolution settings) after creating an image, before saving it. It seems to be automatically enabled when the image is created with "32-bit float" checked, and disabled if it's created with "32-bit float" left unchecked. I think it's just a way to alter the same setting after the image is created, but I'm not certain.

  • $\begingroup$ I think OpenEXR will give better result (and of course larger images) but it looks like painting a height doesn't give any good result for the moment, or I'm missing something, I hope someone will be able to answer $\endgroup$
    – moonboots
    Commented Apr 5, 2022 at 11:14
  • $\begingroup$ The "float buffer" probably keeps track of the data as 32-bit floating point, which is then downsampled to 8 bit integer color for typical saving. Saving the image shouldn't change the data-- but quitting and restarting Blender should clear the buffer, leading to reloading from just saved image color, rather than the baked/rendered color. $\endgroup$
    – Nathan
    Commented Apr 12, 2022 at 4:48

1 Answer 1


I have always used 32-bit float for anything normal-map-baking-related as well as bump map mixing. The "buffer" you mention, I cannot find it. With The "Compositor on viewport" patch, I can think that the "float buffer" is offering a way to view it on the viewport. Ultimately, you need this into a file, which is why I recommend using the 32-bit float from the image node directly. (i.e: there's a buffer for Bloom, but ultimately you want it to affect the alpha to view it on your viewport). I covered buffers briefly a while back:

32 bit float

enter image description here

  • $\begingroup$ Added an update to the question specifying the location of the "float buffer" option. $\endgroup$
    – Legoman
    Commented Apr 7, 2022 at 19:19

You must log in to answer this question.

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