William F Pokorny <ano### [at] anonymousorg> wrote:
> Hi Louis,
> Answering blind in not really understanding what you are trying. It
> helps to see the code.
>
> If you have not yet found it, check out Mike William's tutorial page at
> http://www.econym.demon.co.uk/isotut/index.htm. It's got lots of useful
> information on isosurfaces plus working code you can download for all
> the pages.
>
> When you get into max gradients as large as yours without the resultant
> object you want, it very likely means the isosurface
> "function/pigment-image_map-function" you are rendering has - or
> effectively has - discontinuities. There are changes in value which are
> too rapid to accurately find the surface.
>
> In other words, something is wrong(1). I've never gone over 1000 for the
> max gradient setting in anything I've done and I normally run <<200. If
> nicely behaved, slowly changing, functions <<20 is typical for a max
> gradient.
>
> I didn't follow what you did with your bit depth change. You too say
> heightfield, which is not an isosurface, but another kind of povray object.
>
> Taking you to be working with a base isosurface shape where you are
> adding a surface displacement via an input image's grey, red, green or
> blue value:
>
> - Are you using interpolation to smooth your input image? If not, do.
>
> - Are you using the right image_map map_type for the base iso shape? The
> default is 0 or planar.
>
> - Larger input images for the pigment image_map function often help.
>
> Bill P.
>
> (1) - I'm strictly lying as it is OK to have max gradient warnings so
> long as you are getting the iso object you want.

Thanks for your reply. I'm mapping one of the "blue marble" heightfields to a
sphere (http://visibleearth.nasa.gov/view.php?id=73934) with interpolate 3.

After removing the "precision" declaration the problem was allot less, I'm now
at 1200 and it renders ok. Not sure what value I had there but it must have
been excessive.

Cheers,
Louis

Post a reply to this message