Gilles Tran wrote:
> 
> > That looks very promising.  Are those also suited for large size renders?
> The goal is to make the technique suitable in quality and render times for
> large size renders. I've tested them at 1600*600 and 1600*1200 (6 hours each
> on my P4) which is already very acceptable considering that these are pretty
> complex scenes with radiosity, media and isosurfaces, but I'm doing further
> tests. It won't be 100% perfect and will have some limitations but I hope it
> will be usable.

6 hours sounds good.  Do they interact with the light (i.e. cast shadows
etc.)?

> > I also like the terrain.  Looks like made with isosurfaces...
> The terrain is a function-generated heigh-field and the mountain is a big
> hill of isosurfaces. Both render quite fast too.

Using a heightfield for the background terrain is surely an efficient
solution if you can afford it memory-wise.  With a large render this will
always be tricky though - the higher the render resolution the stronger
the differences between foreground and background will get in the result. 
In case of your sample this won't look too artificial since the step from
foreground to background is quite large but this is a real problem when it
is more fluent.

In my experience the most serious disadvantage of using isosurfaces in
such situations is the strong slowdown you get when using radiosity.  It
could be worth trying variable radiosity settings here (i have made tests
with that some time ago but not with landscape scenes).

Christoph

-- 
POV-Ray tutorials, include files, Sim-POV,
HCR-Edit and more: http://www.tu-bs.de/~y0013390/
Last updated 28 Feb. 2003 _____./\/^>_*_<^\/\.______

Post a reply to this message