Wasn't it Simen Kvaal who wrote:
>You seem to know these built-in-functions well!

It didn't start out that way. It took a lot of hard work to obtain all
that information.

>I've played with them, but never got anything to work. So I decided to give
>your source-files a try. I fired up algbr015.pov (Duphin_Cyclid-thing), and
>it worked fine! However, when I tried to make my own scene, I found that it
>seems the behaviour of these bulit-in functions are somewhat erratic. For
>example, when _removing_ the last four lines of algbr015.pov, the ones
>placing the transparent container, the Isosurface only shows something like
>a solid ball! And without affecting the code for the iso at all.
>
>Du you know why such things happen (stuff like this happen to my isos all
>the time), and mayby how to avoid/predict them?

What's happening is that several built in functions produce surfaces
that are actually inside out, and the Duphin_Cyclid is one of them.
What's happening is that the contained_by sphere is intersecting the
"inside" of the cyclid at all points, so what actually gets rendered is
the contained_by sphere, and the cyclid itself is a hollow space inside
this sphere where you can't see it. I don't know an easy way to predict
inside-out-ness, but the solution is easy, use "sign -1".

A subtle, and not quite intentional, side effect of the presence of my
container surfaces is that they clobber the parts where the surface runs
round the contained_by sphere. Because my transparent container surfaces
are exactly coincident with the contained_by surface, only one of them
gets rendered, and that happens to be the transparent surface.

An alternative solution is to add the keyword "open" just after the
contained_by statement.

I'll go back through my tutorial to check for inside-out surfaces and
turn them right side out.

-- 
Mike Williams
Gentleman of Leisure

Post a reply to this message