oh, i forgot one other thing.  it is better to have your object defined
as a unit sixed object and then to scale it.  media works better that
way.  ex. i had a unit sphere where my media was spherical in density. 
i then scaled my sphere and the media accordingly with no problem.  one
other thing i do when scaling is to scale in small steps doing a test at
each step to make sure that each is what i expect.

Bob Hughes wrote:
> 
> "How Camp" <hca### [at] ksuedu> wrote in message news:396a412c@news.povray.org...
> |
> | Unfortunately, I tried this.  I assumed the same thing.  My object (a simple
> | cone) is scaled <0.04, 0.04, 0.1>.  So, I figured I could apply a
> | transformation to the texture of <25, 25, 10> just as you suggested.  The
> | results are the same as without the transformation.
> 
> Don't scale the media overall or you'll get a different pattern size,  so
> scale either the emission (absorption too) or density or scattering.
> Also, if you start with a container size larger or smaller than unit-size the
> media will not fill it in the usual way.  It will become thinned out from
> either a lack of distance or too much distance for the samples to be checked
> through.  You should use scaling of the container+media at the end of the
> objects statement block instead.
> Scaling of the densities might prove difficult if there are many indexed or
> are in separate density statements used in a density map.
> I think you should be able to get a reasonable media through scaling in that
> way.
> 
> Bob

Post a reply to this message