"Jan Walzer" <nos### [at] informatikuni-hallede> writes:

> > > Can't we use it as a kind of Media ???
> >
> > You're not the first one to come up with this idea. It has already be
> > phrased by Kajiya and Kay in "Rendering Fur with Three Dimensional
> > Textures", Computer Graphics, Volume 23, Number 3, July 1989.
> Where to get this paper ?

First try the library of the university in Halle. If it's not available
there, you can get it using <insert what "Fernleihe" means in English>.
Alternatively, it may be downloadable with a credit card at www.acm.org,
because it's in a magazine that is published by the ACM (which I forgot to
mention). 
 
> [...]
> > In my opinion, this approach has a lot of potential. Apart from fur,
> > I think that grass and forests can be renderered this way. Of course,
> > someone must first find an adequate lighting model for these objects.
> > However, you should be aware that this method will only work for
> > distant views.
> hmmm ... but am I right, about the rendertime ...
> I've thougt about the last night, and found that therefore the "steps in the
> media" (or how was it called) have to be massivly increased, to get all the
> branches of a high-detailed tree, don't they?
> The standard for this is AFAIK 10 steps, and how can this make a good tree ?

I cannot speak for trees but for fur. Using media makes mainly sense, when
you don't care about details. Instead of a higly detailed image, you want
to look your fur (tree, grass, forest) good from a distance. So, you 
approximate the reflections instead of generating a lot of cylinders.

As far as I know, the number of steps is also calculated adaptively.
That's why I suspect the number of steps to be higher by default in
such applications. However, I didn't care about them too much so far.
I have tried the new sampling methods in the megapatch. They didn't
work very well in case of my furry torus.

Thomas

-- 
http://thomas.willhalm.de/ (includes pgp key)

Post a reply to this message