Just had a quick question for modeling concepts. I have a, for example, building that is getting somewhat complex. There are many different pieces, some smaller than others. I wanted to take my finished mesh into mudbox to sculpt out the stone walls, the wood textures etc. However when I import the mesh into mudbox some parts become very small, so small that it doesn’t really register any sculpting(though they can be clear seen in game, and the details would look a lot better). Would it be better to break the mesh up into pieces (stone walls, roof, wood pieces etc) and have separate normal/uv maps for each piece? Or would there be a better way to go about this. Thanks!
To get enough detail you would probably want to use tiling textures instead of sculpting stuff in Mudbox, unless there’s a lot of unique detail and the surfaces aren’t just flat it would be a huge amount of time to do things that way. With tiling textures you can get more detail with smaller resolutions and save a bunch of time. You will have to consider how to break up the mesh, at the very least you have to consider lightmapping, each mesh only gets one lightmap so if there’s too much geometry then it can’t get enough detail in the lightmap even at a high resolution. So for like a building, you might have the walls as a single mesh and the floor as another mesh. That can also help with materials, the walls with one material and the floors with another. A single mesh can have multiple materials, but you have to keep that low as well, each material is an additional draw call.
Awesome, thanks for the quick response! I’m still learning all this stuff, so what should I be looking for (in terms of tiling textures) for just a quick example/tutorial etc?
Or do you simply just mean apply tileable textures to the mesh and generate some normals from that?
You can find tileable diffuse maps and normal maps, or there’s various tools out there to create a normal map from an image.