08-17-2010 10:46 AM - edited 08-17-2010 10:48 AM
When faceting curved faces with holes I am not getting good results. The problem seems to be that after faceting, the hole edge is no longer flush with the face (now faces). The hole portrudes out to the distance of the original smooth curve, while the edge of the face is approximated with secants. This causes a "starburst" of unnecessary facets from the hole curve to the outer edge segments. I'm attaching visual examples of a curved wall with a hole, the current R21 faceting results, and what we would like to get instead.
08-17-2010 11:16 AM
I am sorry I don't know quite what you dislike about our faceting. The mesh you showed as the desired result (the second picture, right?) did not contain all triangles. Our faceter output is a triangular mesh. If you have a face with two loops like that, in order to triangulate it, you need to connect it something like we did.
Perhaps it would help to elaborate about what you are trying to do with the mesh and why having the star of facets connecting the inner hole with the outer hole is causing problems for you.
08-17-2010 01:53 PM - edited 08-18-2010 09:15 AM
Yes, I'm not making myself clear. It's not the segments needed for triangulation themselves that are the problem. It's the fact that the hole edge is not flush with the quad face, and so the segments do not lie on the plane of the quad. It causes a kind of fileted pyramid bulging out, with the hole at the "top" of the pyramid. Without the hole, the geometry of the mesh face would be a planar quad (using the "no triangles" facet option). The existence of a relatively tiny hole actually changes the shape of the convex hull of the mesh significantly.
We are using the mesh geometry to do a hidden line approximation. If the triangular faces around the hole were coplanar, we could easily eliminate them.
08-21-2010 02:00 PM
Sorry it has been a couple of days since I could respond. If I understand correctly, you have a model with a nearly planar face with a hole in it Because the face is nearly flat, it has no grid lines. This gives a pyramid which looks bad in HLR.
Would it help if there were more grid lines? If so, this could be achieved in a lot of different ways using the faceter controls (e.g., set max edge length in facet _options_precise to the face box diagonal size over 4 or 5, tighten the surface tolerance, tighten the normal tolerance ...) You may find more details regarding how to control the faceter at http://doc.spatial.com/index.php/Using_Facet_Optio
If the geometry actually has a slight bulge in it, I think it is correct for the faceter to give a bulge in the output. If you believe the faceter is giving results which don't match the actual geometry within the requested tolerances, please file a bug for the issue.
Please feel free to clarify the issue further if I am not giving useful help.