|
 |
Hi all,
I have a scene with about 30.000 hexagonal-Z prisms which in their turn
generate one large hexagonal-Z prism (i.e. a hexagonal-Z lattice from a
numerical calculation). Each small prism has its own color based on the
results from the calculation. If I just render this scene with the large
prism a union{ prism ... prism... prism ..} of the small prisms, it takes
about 10 seconds in total (parsing+rendering). However, when I use a
clipped_by {plane {-z,0}} to cut the large prism in half to see the inside,
rendering time becomes 10 hours. Is there a faster way (clipping each small
prism seperately, specifying the bounding box by hand,...) to do this?
thanks,
gert
Post a reply to this message
|
|
|
|
Wasn't it Gert Van den Eynde who wrote:
>Hi all,
>
>I have a scene with about 30.000 hexagonal-Z prisms which in their turn
>generate one large hexagonal-Z prism (i.e. a hexagonal-Z lattice from a
>numerical calculation). Each small prism has its own color based on the
>results from the calculation. If I just render this scene with the large
>prism a union{ prism ... prism... prism ..} of the small prisms, it takes
>about 10 seconds in total (parsing+rendering). However, when I use a
>clipped_by {plane {-z,0}} to cut the large prism in half to see the inside,
>rendering time becomes 10 hours. Is there a faster way (clipping each small
>prism seperately, specifying the bounding box by hand,...) to do this?
I guess that you're getting the same problem with clipped_by as is
generally encountered with difference. In a union, POV can apply
sensible bounds around each component, but when dealing with differences
and clipped_by it can only apply a bound around the whole object, which
means that each ray ends up being tested against every component.
The usual solution is to clip each component separately.
See <http://www.econym.demon.co.uk/holetut/index.htm> for full details
of the difference problem and examples of how to observe what's
happening by observing the stats.
--
Mike Williams
Gentleman of Leisure
Post a reply to this message
|
|
