Jim Wilson <[EMAIL PROTECTED]> said: > Andy Ross <[EMAIL PROTECTED]> said: > > > Anyway, I think I've finally got this thing working. Grab a new copy > > at http://www.plausible.org/vertsplit/vertsplit2.tar.gz, dump it into > > your ssg directory and try it out. I've run it on a bunch of > > FlightGear aircraft and am quite happy with the results. No more > > "seam" along the texture boundaries on the 747, and the (ahem) > > creatively tesselated bo105 model displays correctly* now. I can't > > find anything that breaks it or looks wrong. > > > > Still something going on here. This is a wip model (that's why the tex > mapping is screwed up) so let me look at it more before you get too upset :-) > I can tell you there are no dups/degenerates there. And the surfaces on that > edge are not triangulated in the ac file. > > http://www.spiderbark.com/fgfs/leadingedge.png > > Haven't checked yet to see if the cvs 747 does it, but there's a good chance > it does.
Ok the problem shows up with objects triangulated as follows: Picture a diamond shape. Split it vertically so that you now have two triangles, one on the left and the other on the right. Split the one on the right horizontally. Now you have three triangles. The two on the right share a vertex the one on the left does not contain. These are problematic arrangements anyway so it probably doesn't indicate a serious bug. Also I'm not 100% sure of the precise condition, but this represents the general idea of the issue. The possibility that the problem in the model could lead to degenerates during stripification is certainly there. BTW it appears that the "46" threshold indicates something other than the inside angle between surface normals. Is this the intent? So far the code looks very efficient. Nice job! I'm wondering now if we should be locating this work in simgear rather than plib. It would be easy enough to do so, and it would and allow us a great deal of flexibility in the implementation different algorithms without having to intefere with other plib based projects. All we'd have to do is use ssgLoadAC() and call our own optimizers based on model wrapper configuration properties. Best, Jim _______________________________________________ Flightgear-devel mailing list [EMAIL PROTECTED] http://mail.flightgear.org/mailman/listinfo/flightgear-devel
