Wasn't it Nicolas Alvarez who wrote:

>I think it would be a lot faster to use 6 isosurface "planes" than an 
>actual box (intersection of planes in a single isosurface).

If you actually try it, it turns out to be slightly faster to use
   function { max((y*y-1),(x*x-1),(z*z-1)) +wobble(x,y,z) }
than to use the intersection of 6 isosurface "planes" that use the same 
contained_by (even though the "planes" use a lower max_gradient). 
There's not a huge difference.

I guess that if you were to create a tighter contained_by for each 
"plane", then you might get that to run faster, but I suspect that the 
gain would not be significant.

-- 
Mike Williams
Gentleman of Leisure

Post a reply to this message