[Bf-committers] Particle experiments - looking for feedback

Leon Turner bf-committers@blender.org
Sat, 1 May 2004 21:06:44 +0900


yup, it's entirely possible that my earlier attempts had problems even
with completely planar quads. There's still *some* leaking: mainly
when the particles have pretty much "settled" on a plane due to
damping, and then hit a seam with another face (one which isn't planar
with the first). I'm not entirely sure how this can be easily fixed
(any thoughts appreciated!), but on the whole it shouldn't be too bad.

Also, there might still be problems in this build (although it seemed
fine when I tried a subdivided plane just now). If you have windows
and could try and reproduce using the build I linked to, that would be
really helpful.

I'll have a think about treating quads as two triangles in all cases:
at the moment, I first calculate whether the particle has crossed the
plane of the face. For this, I just use the first three vertices, and
this is a very quick check. If it has, I then go on to calculate the
intersection with the triangles of the face - for this, I do treat all
quads as two triangles.



On Sat, 1 May 2004 13:30:12 +0200 (CEST), Alexander Ewering
<blender@instinctive.de> wrote:

> I'm not sure if it only happened for non-planar quads. I remember one
> of your comments in tuhopuu's featurelist.txt where you assumed that
> "lost collisions" happened due to excessively high particle or object
> speed...
> However, a lot more leakage happens due to quads, and I *think*
> I tried it with a simple subdivided plane, in which all faces are
> definitely planar. And there too, there was a huge amount of particle
> leakage, and converting all quads to triangles with ctrl-t completely
> fixed it.
> Maybe I remember wrong or you have fixed it in the meantime. Anyway,
> I would indeed propose that you treat all quads as two triangles (you
> can use the fact that the two "correct" triangles in a quad always have
> the same order (123, 134 I think), or at least provide a button "perfect
> collisions". :)
> Otherwise, your extensions really are great. And people here know I don't
> say that easily :-)
> | alexander ewering               instinctive new media
> | ae[@]instinctive[.]de   http://www[.]instinctive[.]de
> _______________________________________________
> Bf-committers mailing list
> Bf-committers@blender.org
> http://www.blender.org/mailman/listinfo/bf-committers