POV-Ray : Newsgroups : povray.advanced-users : milky glass problem : Re: milky glass problem Server Time
29 Jul 2024 22:20:36 EDT (-0400)
  Re: milky glass problem  
From:
Date: 26 Feb 2002 09:31:55
Message: <3c7b9c5b@news.povray.org>
> > doesn't work on triangles in meshes). Will tracing speed decrease and
memory
> > usage increase??
>
> Yes, you will lose this.  You can compensate, a little, by breaking the
big
> union up into smaller unions and bounding them separately.

*lol* this is the point where I started from. Then I made it a single mesh
to speed it up (I had 30000 objects which made bounding test very slow). Now
I will have to flip back again ....

> In addition, if you specify different interiors for different triangles,
you
> will have other problems.  One of the other problems you will have is that
> this:
>
>   union {
>    triangle {x,y,x+y interior {...whatever...}}
>    triangle {x+z,y+z,1 interior {...whatever...}}
>   }
>
> won't get rendered the way you expect it to, because the two interiors
> there are utterly independent of each other.  This:

Yes I am aware of this. But since the interior has no structure (just a
foggy -something-) this wouldn't be a problem.

>   union {
>    triangle {x,y,x+y}
>    triangle {x+z,y+z,1}
>    interior {...whatever...}
>   }
>
> will be closer to what you want.

Yep. And this would be also necessary for bounding reasons.

> In addition, it's still possible for a ray to sneak between the two
triangles
> and give you funny borders around your triangle.  That might be the cause
of
> the black lines you mentioned before.

The current thing doesn't use interior at all. The borders seem to be a
"numeric resolution" problem. It is alwys a single pixel regardless of the
rendering resolution.

regards
SY


Post a reply to this message

Copyright 2003-2023 Persistence of Vision Raytracer Pty. Ltd.