Forum: Poser - OFFICIAL


Subject: Conforming clothes->dynamic clothing = Extreme slowdown. =(

seanlo opened this issue on Jan 21, 2008 · 12 posts


Conniekat8 posted Mon, 21 January 2008 at 5:31 PM

Quote - seanlo:

I've read often about how Poser doesn't work well with triangles, but I have to say that in my experience it just isn't true.  I've converted conforming clothes to .obj's and deliberately "split" the rectangles into triangles to get a smoother look.  Also, there are dynamic clothes--made as dynamic clothes--consisting entirely of triangles that work like a charm and run very fast in simulation.  So I would say to ignore the warnings and try it yourself.

 

The issue is not with triangles themselves. If you take a quad based surfacem and densify it so that each quad becomes two triangles, there's no problems.

It's when you try and simplify a mesh, most mesh decimating routines will make triangles spanning 1, 2 , 3, 5 or even more old quads or triangles, resulting in looong skinny tiangles wherever the object surface is pretty uniform.

Mesh decimating routines don't take small triangles or quads, and make them into uniformly spread larger triangles. (imagine inverse tessalation. they don't do this.)
When you decimate a mesh you get a very irreguilar mix of triangle sizes and placement. The problem isn't with triangles, the problem is with irregilarity of a decimated mesh.
Same would happen if there was a lot of irregularly sized and placed quads, but I'm yet to se a decimator that leaves you with quads.

Hi, my namez: "NO, Bad Kitteh, NO!"  Whaz yurs?
BadKittehCo Store  BadKittehCo Freebies and product support