Re: pfiPick and FCS:es

New Message Reply Date view Thread view Subject view Author view

MLM Veraart (Veraart++at++fel.tno.nl)
Wed, 17 Mar 1999 14:19:17 +0100


Yair Kurzion wrote:
>
> Hi Jan !
>
> > In my app I'm trying to pick an object using pfiPick. The object is
> > placed below a switch node and the switch node below an FCS. My problem
> > is that picking allways fails when using FCS:es. If I use DCS:es there's
> > no problem picking at all.
>
> Are you using STATIC bounding volumes on the FCS node and on the nodes under
> it ?
>
> The main difference between a DCS and an FCS is that when an FCS matrix
> changes, performer does NOT re-calculate the bounding volumes of the nodes
> under and above the FCS. When using a DCS node, any matrix change causes a
> re-calculation of all bounding volumes under the DCS and on the path from the
> root to the DCS.
I'm not so very sure about updating all the child node bounding volumes.
In case the child is shared under many different DCS's you get an error
in the culling phase.

Mario
>
> The intersection code uses this bounding volume in order to minimize the number
> of geometric primitives that it intersects. If the bounding volume is incorrect,
> the intersection will sometimes miss the objects.
>
> The way around this (when using FCS) is to define a STATIC bounding volume for
> all the geometry under the FCS node. Find a bounding volume that bounds all the
> possible matrices that your application will store in the FCS node. Set this
> bounding volume as the static bounding volume of the nodes under the FCS.
>
> -yair


New Message Reply Date view Thread view Subject view Author view

This archive was generated by hypermail 2.0b2 on Wed Mar 17 1999 - 05:18:37 PST

This message has been cleansed for anti-spam protection. Replace '++at++' in any mail addresses with the '@' symbol.