KatsBits Community

Make a normal map from a window texture (time-lapse)

kat · 1 · 10665

0 Members and 1 Guest are viewing this topic.

Offline kat

  • Administrator
  • Hero Member
  • *
    • Posts: 2692
    • KatsBits

The above video is a 30 minute time-lapse of the 3+ hours taken to make a grey-scale template image, using an RtCW window texture from "village", that is then passed through a normal map filter - nJob in this instance - to produce a normal map with a reasonable 'normalised' effect under dynamic lighting conditions.

Scenario
A set a diffuse textures has already been produced which need to be converted for use with Normal Maps. Ideally these images should be used in some way to speed the production of multiple assets.

Problem
Typically a textures can be desaturated and passed through a normal map filter but in doing so errors and normalisation issues occur because the filter process is not able to distinguish between dark shadows and depth; the two are equal procedurally. This is further exacerbated by the complex nature of certain textures or types of surfaces described by a texture, organic materials in particular cause additional issues due to their random nature.

Solution
A relatively quick approach to making normal maps is to use a diffuse texture as a guide only, to aid production of a simplified image specifically tasked for conversion to a normal map. The reason for this relates to the way grey-scale tones are interpreted by the conversion process such that shadows and other 'darkened' areas are read by the process as recessed. Making a 'proper' grey-scale template where dark tones represent depth, light, height, allows for the production of normal maps that exhibit a better normalisation effect under dynamic lighting conditions (as shown in Blender using a Lamp object in 'Blender Game' and 'GLSL' modes).

Additional Resources
How not to make normal maps from images,  photos or 2D artwork
Make normal maps from 2D image artwork (not photos)
What to look out for and how to make normal maps from photographs