Nicolas Farcet (farcet++at++thomson-lcr.fr)
Tue, 26 Mar 1996 15:51:55 -0600
I would like to test a special culling architecture
with Performer 2.0. It's a "meta-culling" compared
to typical Performer hierarchical Bounding Volume culling.
In order to implement this culling, I use a special
structure that references objects, and an algorithm tells
me which ones to draw.
I wonder how this specific culling architecture would
fit in Performer 2.0.
The first approach I considered was to use culling callbacks
on nodes that I could control from the outside with my culling algorithm.
But it is a bit awkward and would produce an overhead
because my culling architecture has nothing to do with
Performer scene graph, and therefore, using Performer
depth-first scene graph evaluation for culling seems
inadequate.
The second approach would be to plug a general culling
callback, and discard pfCull(). This would allow me to
select the nodes I want to draw. But it implies that I
can do myself what pfCull() does - that is, generate
pfDispList and so on.
The second method seems cleaner and more adequate to me.
So my questions are :
- Isn't it technically insane to replace pfCull() by my own function ?
- Are there hard problems I seem to be missing ?
- Where could I find informations on what pfCull exactly does ?
Thanks very much.
-- _______________________________________________________________________________Nicolas Farcet
Tel : (+ 33 1) 69 33 08 07 Thomson-CSF LCR Fax : (+ 33 1) 69 33 08 65 Domaine de Corbeville Email : farcet++at++thomson-lcr.fr F-91404 Orsay cedex - FRANCE _______________________________________________________________________________
This archive was generated by hypermail 2.0b2 on Mon Aug 10 1998 - 17:52:35 PDT