"Alain" <ele### [at] netscapenet> schreef in bericht 
news:454089a3$1@news.povray.org...
>
> When you have very detailled trees, you only use 1 or 2, maybe 3. For the 
> background trees, it's usualy beter to use mesh trees.
> Not all isosurfaces are very slow. It depend on it's complexity and 
> gradiant.
>
Thanks Alain!
Yes, I agree about the background trees. It wouldn't make really sense to 
use isosurfaces for that, so my question there was a little unrealistic. 
Still, I am curious about render time of Kenneth's isosurface tree in 
conjunction with (as you mentioned) its complexity, and the levels of detail 
(bark in particular).

Thomas

Post a reply to this message

"Thomas de Groot" <t.d### [at] internlnet> wrote:
 soon.
> >
> I am curious about this: Isn't your render time very slow for an isosurface
> tree? And what about planting a forest with them?

Yes, my current tree renders quite slowly. Interestingly, it gets slower the
more functions I add to it (regardless of the max_gradient value.)  I
suppose there's a good mathematical reason for that(?)

I haven't yet tried #declare-ing the tree and using it multiple times--to
see if the render time stays the same, or if it increases by the number of
trees. I don't think it would increase by a great deal. I have a vague
impression (based on different experiments that I've done over time) that
an isosurface is *somehow* converted to a triangle mesh during
rendering--but that's just a wild conjecture.  Anybody know the answer?

Ken

Post a reply to this message

Kenneth nous apporta ses lumieres en ce 27/10/2006 15:54:
> "Thomas de Groot" <t.d### [at] internlnet> wrote:
>  soon.
>> I am curious about this: Isn't your render time very slow for an isosurface
>> tree? And what about planting a forest with them?
> 
> Yes, my current tree renders quite slowly. Interestingly, it gets slower the
> more functions I add to it (regardless of the max_gradient value.)  I
> suppose there's a good mathematical reason for that(?)
> 
> I haven't yet tried #declare-ing the tree and using it multiple times--to
> see if the render time stays the same, or if it increases by the number of
> trees. I don't think it would increase by a great deal. I have a vague
> impression (based on different experiments that I've done over time) that
> an isosurface is *somehow* converted to a triangle mesh during
> rendering--but that's just a wild conjecture.  Anybody know the answer?
> 
> Ken
> 
An isosurface is not converted into anything remotely resembling a mesh.

-- 
Alain
-------------------------------------------------
Islam: If shit happens, it is the will of Allah.

Post a reply to this message