On 04/24/2016 11:36 AM, Le_Forgeron wrote:
> Le 24/04/2016 17:19, William F Pokorny a écrit :
>
>> I don't believe the function bounding being suggested is going to help with such
isosurfaces. It is at least not apparent to me how I could employ it. Perhaps I could
though if this new method could associate many containers - as in say a spline of
spheres/ranges - with
>> one function?
>
> Now thinking about it: what is the gain of one function being the sum of bounded
boxed functions,
> compared to a multitude of distinct functions which are bounded by their own box ?
>
> Wouldn't it duplicate the union{} for isosurfaces ?
>

I think I did a poor job of describing what I was thinking while 
thinking aloud...

With the sphere sweep iso, the one function is based upon a spline 
function along which an f_sphere is swept(1).

What I'd like - or think I'd like - in cases like this is a way to 
better bound this one function. To that end I was thinking about being 
able to specify set of spheres/ranges to bound the function roughly 
following the same spline, but larger as a collection of bounding 
regions which I presumed would fit nicely in any new search scheme.

If the point in space being tested is inside - any - of the spheres in 
this collection of bounding spheres the one function would be evaluated. 
It does need a many bounding shapes to on function mechanism for this. 
The gain or aim would be evaluating the one function a good deal less 
than happens today - or would happen if I had access to only one 
container shape for the function.

I think what you are suggesting is a set up multiple f_sphere/blobbing 
functions already placed following the spline each with there own 
bounding shape? This would probably work too I guess with some work on 
how to distribute the sphere/blobs ahead of time along the spline so as 
to look right. This approach would have the significant advantage the 
iso sphere sweep could fold every which way with no issues. Maybe this 
is the better way to go given internal bounding...

Aside: Some more thinking aloud. Not seen it stated out right in this 
thread, but to be most general we too would like to be able to distort - 
with passed coordinates - both the functions and containers/ranges 
together. Or, is the thinking we'd set the sub function containers up to 
handle any distortion as is the case with the outer contained_by shape 
today ?

Bill P.

(1)- Lying a bit as there were derivative adjustments too, to keep the 
sphere rounder as it was swept.

Post a reply to this message