Re: [Algorithms] Dynamic Plane Shifting BSP Traversal Problem

[Damian C. <dam...@gm...>]

Jay Stelly: You're right, it's not going to nicely map to a sphere's
surface, it's just an approximation.  It's a fast approximation, which is
the real selling point - you get reasonable collisions for not much CPU.
Since my last email to this list I have been successful with this technique
and have the so-called 'zero volume cells' algorithm working.  It kicks
butt.  I have suffered no noticeable performance hit and I'm colliding
spheres and cylinders nicely.

Jon Watte: This technique solves more than just spheres. Yes, for just
spheres, you don't need to jump through these hoops.  With this technique
you can implement any convex shape for which you can calculate a point
tangential to a plane.  So an AABB, a cylinder, and yes a sphere too.  I
suppose if you had a translating, rotating box you collide that too.

~Damian();


On Wed, Feb 25, 2009 at 6:19 PM, Gregory Junker <gj...@da...> wrote:

> @ Jon -- the OP is at the bottom; the person you are replying to is not the
> OP.
>
> Greg
>
> > -----Original Message-----
> > From: Jon Watte [mailto:jw...@gm...]
> > Sent: Tuesday, February 24, 2009 3:48 PM
> > To: Game Development Algorithms
> > Subject: Re: [Algorithms] Dynamic Plane Shifting BSP Traversal Problem
> >
> > Jay Stelly wrote:
> > > The bevelling doesn't completely solve the problem though.  In order
> > > to shrink a sphere down to a point you need to spherically extend
> > > (i.e. convolve with a sphere) the BSP geometry.  Or to put it another
> > > way - you need to add bevel planes to the BSP for every potential
> > > separating axis.
> >
> > The word for that kind of expansion of geometry is "Minkowski sum". You
> > get some Google hits on that, but most of them are not quite real-time
> > or implementation-oriented enough for games.
> >
> > Color me confused, though: I have previously collided spheres against
> > geometry in a BSP tree, and it works just fine without the plane
> > shifting -- without any artifacts. Is there some other benefit you are
> > trying to get out of the plane shifting than avoiding having to do
> > sphere/polygon tests (i e, live with only the slightly faster
> > sphere/plane tests)?
> > It seems to me that the cost of sphere/poly tests, once culled based on
> > a BSP, is such a small cost as to be negligible, and certainly not worth
> > all the hours that have been spent on it already even on this list :-)
> >
> > Sincerely,
> >
> > jw
> >
> >
> >
> --------------------------------------------------------------------------
> > ----
> > Open Source Business Conference (OSBC), March 24-25, 2009, San Francisco,
> > CA
> > -OSBC tackles the biggest issue in open source: Open Sourcing the
> > Enterprise
> > -Strategies to boost innovation and cut costs with open source
> > participation
> > -Receive a $600 discount off the registration fee with the source code:
> > SFAD
> > http://p.sf.net/sfu/XcvMzF8H
> > _______________________________________________
> > GDAlgorithms-list mailing list
> > GDA...@li...
> > https://lists.sourceforge.net/lists/listinfo/gdalgorithms-list
> > Archives:
> >
> http://sourceforge.net/mailarchive/forum.php?forum_name=gdalgorithms-list
>
>
>
> ------------------------------------------------------------------------------
> Open Source Business Conference (OSBC), March 24-25, 2009, San Francisco,
> CA
> -OSBC tackles the biggest issue in open source: Open Sourcing the
> Enterprise
> -Strategies to boost innovation and cut costs with open source
> participation
> -Receive a $600 discount off the registration fee with the source code:
> SFAD
> http://p.sf.net/sfu/XcvMzF8H
> _______________________________________________
> GDAlgorithms-list mailing list
> GDA...@li...
> https://lists.sourceforge.net/lists/listinfo/gdalgorithms-list
> Archives:
> http://sourceforge.net/mailarchive/forum.php?forum_name=gdalgorithms-list
>