thickening complex surfaces

 From:  Michael Gibson
6131.5 In reply to 6131.1 
Also this surface as well may be probematic due to it's corner shape too:



The surface is highly compressed in that small area and if the 2 surface edges are exactly anti-tangent there (maybe they're not quite, not quite sure), that's also a type of degenerate corner situation as well which again is bad for offset operations.

The way surface normals of a surface are defined is by a direction that is a shared perpendicular between the U and V parametric directions at that point. When the directions form a straight line with one another, either by pointing in the same direction or exactly opposite directions this calculation becomes indistinct.

So because of that areas that are shaped like that usually can't be a direct part of the base underlying surface and instead need to be trim curves that are on a broader more regularly formed surface instead.

Anyway, these 2 particular surfaces are going to be the most problematic things here.

- Michael
Attachments: