I recently completed my first sculpted game character including rigging and animation. I am still desperately trying to find reasons to do PBR texture paint in blender 2.8 rather than substance but I am still on the fence. For the time being I used only blender cycles.
The steps I took:
- Sculpted a very high poly character.
- Created a retopologised (low-poly) version of this character.
- Baked a normal-map from the high-poly sculpt onto the low-poly mesh.
- Added the normal-map to the low-poly mesh and then painted that in texture painter.
My question is "Is it better in general to paint the high-poly sculpt and then bake all the textures (diffuse/albedo, roughness & normal-map) afterwards?"
The upside I can see to that approach might be that if I need to adjust the resulting tri-count of the low poly model I don't need to repaint - I can just make now low-poly mesh and then re-bake accordingly.
I see some quite conflicting information online.
One thing I noticed in painting the low -poly mesh with normal-map applied is that occasionally the paint landed slightly off the target I was aiming for because of the real topology underneath.
The efficacy of workflows can be subjective but I am interested to know if my suggested workflow is better than the one I used or not.
Thanks in advance.