No announcement yet.

Static Mesh & Boolean Results

  • Filter
  • Time
  • Show
Clear All
new posts

    Static Mesh & Boolean Results

    When you take a static mesh you have created in a 3D modeler and perform Boolean operations on it (intersect/union/difference) it creates extra lines and faces on it. I can merge vertices/join faces to eliminate most of these to get to the smallest amount of quad faces, but is it worth the effort? I mean I'm going to have to triagulate everything before I export to ASE. Should I worry about this or does it have much impact on performance in a level other than looking messy in wireframe mode?

    This is a difficult question to answer, because it depends....

    Personally I don't usually bool anything because it often creates a mess that requires a lot of manual fixup, and the usual resulting thin sliver triangles are a pain to get proper smoothing groups and often are a pain for unwrapped skins and lightmaps.

    I assume that by "triangulate" you are referring to collapsing/converting the mesh to a set of base triangle polygons as opposed to "tesselation"...

    For single staticmeshes you are better off creating as few faces as possible to get the final mesh shape, and have them as clean as possible. This facilitates the creation of better UVUnwraps for lightmap coordinates.
    As often as is possible I try to create all staticmeshes with as few triangles as possible to get the shape well, and design it so that it uses a single unwrapped skin so that AmbientOcclusion can be baked and folded into the diffuse, so that the NormalMap is easier to create, so that the lightmap unwrap is easier to create or can be derived directly from the skin unwrap channel, and so that there are as few triangles as possible for faster rendering. For smaller low-poly objects this also can mean that you don't have to spend time creating a simplified collision.