Texture Artists: Diffuse and Normals
Author of this post: Adam Capone | About Notes on Game Dev Authors »>
Last week I talked about assets, mapping, and light bake in Know Your Assets. Currently in industry, meshes are far more complicated and require various other maps, which in some cases require different tools from one another, so I’ve put together a breakdown of how I generally work. This week, I’ll talk about Diffuse Map and Normal Map.
Diffuse Map
The most important map in which the main purpose is to provide the colour information. During the early stages of 3D video game era most if not all games had just a diffuse map used on meshes.
I’ll start off by blocking out simple shapes, maybe throw in a few unedited texture photos and then keep checking within a 3D application to make sure everything is appearing in the right places with no obvious mapping errors. Once everything is in place I then gradually build up the quality of the texture map.
Normal Map
I like to break normal maps into two types. Firstly there is what I call ‘base shape normals’ which are what give the areas within the meshes shape, indent and embossed areas such as a decorative carving on a cabinet. Then there are ’surface texture normal maps’ that would contain elements such as the grain surface texture of the cabinet, when both are combined you have a complete normal map. Usually texture artists will be provided with a ‘base shape normal’ map from the modeler. However, one task texture artists are expected to do is the ’surface texture normal map’ that is made from gray scale painting of which a lot can simply be pulled from your diffuse layer. There are currently 3 main ways of actually making normals:
1. Making high poly meshes in 3D app and then render out a map based on the mapping of the original low poly mesh.
2. Creating gray scale maps with depth in Photoshop and then using a normal map filter that will convert that info into a normal map.
3. Using a high end modeling app such as Zbrush or Mudbox. Then export normals from those.
All have their own advantages and which one you use usually depends on the type of asset you’re working on or time. Method 1 can get you better results for mechanical style modelling. Method 2 is the fastest and easiest method for texture artists though you can not get as much depth out of your normal map. However, in many cases it is perfectly fine for ’surface normals’ such as wood grain. Method 3 fits very well with high end character modeling since it’s very organic and feels like molding clay. It also allows you see in real time where you are placing details instead of switching again and again from Photoshop to 3D application to make sure.
Normal maps are now expected and will be for some time yet. It is more or less expected now that all potential texture artists looking for a job know at least the basics. The more you know the better your chances and if you are able to add Zbrush or Mudbox skills to your CV that will really push you closer to the front.
Next week’s entry will conclude Adam’s breakdown, so stay linked!












July 17th, 2008 at 12:16 am
[…] Adam Capone’s last two entries, Know Your Assets and Diffuse and Normals, he covered the basics any texture artist should know about the process of making textures and the […]