Jean Daigle (jaydee++at++thor.ats.qc.ca)
Tue, 2 May 1995 09:57:46 -0400
On May 2, 5:39pm, deepa++at++tcsernet.tcs.ernet.in wrote:
...
} I noticed a problem with perfly and am wondering if anybody else
} has noticed it too. This occurs when one reads in a model of house for example
} and moves through it using perfly. After colliding with a wall, if one tries to move at a very small angle to this wall or along it, the wall
} starts disappearing and one can see the other side.
} But at this stage collision is still working
} because though one can see the other side one cannot cross over (atleast most
} of the time).
}
} Does anyone have any idea why this happens?
}
...
}-- End of excerpt from deepa++at++tcsernet.tcs.ernet.in
It sounds as though your near clipping plane is far enough from
the viewpoint that the wall (or other object you are colliding with)
is not being drawn. The wall is outside the view volume and is culled
from the scene, allowing an unobstructed view of objects and/or
terrain on the other side.
Since the intersection traversal is separate from the cull traversal
of your scene graph, the collision detection works as you expect.
This isn't, as far as I can tell, a bug in Performer so much as
an unanticipated interaction of some of its features.
Perhaps you could try intersecting a set of segments defining the
frustum from the eyepoint to the near clipping plane against
scene geometry. The man pages for pfHit and pfSegsIsectNode might
be helpful.
Regards,
Jean Daigle.
-----------------------------------------------------------------
| Jean Daigle ATS Aerospace Inc. |
| Software Designer 1250 Boul Marie-Victorin |
| St. Bruno, QC J3V 6B8 |
| jaydee++at++ats.qc.ca Tel: (514) 441-9000 Fax: (514) 441-6789 |
-----------------------------------------------------------------
This archive was generated by hypermail 2.0b2 on Mon Aug 10 1998 - 17:51:28 PDT