Hi Justin,
>X-Accept-Language: en
>MIME-Version: 1.0
>Content-Transfer-Encoding: 7bit
>Date: Fri, 2 Mar 2001 20:14:13 +1100
>From: Justin Couch <[EMAIL PROTECTED]>
>Subject: Re: [JAVA3D] Oh No! More on Collision Detection!
>To: [EMAIL PROTECTED]
>
>Kelvin Chung wrote:
>
>>is return for USE_BOUNDS. Otherwise, for USE_GEOMETRY, current
>>implemetation do a brute force O(n^2) check for each triangle.
>>So the first pass is pretty quick but the second pass is very
>>slow.
>
>Ok, for the article - clarification point - even if you are have
>USE_GEOMETRY, it will do a fast bounds cull first and then do polygonal
>tests on the remaining contentious bounding boxes.
>
Yes, thanks for the clarification.
>Now second question: Are you doing anything like checking for vectors of
>the polygons that are facing in the direction of the contact point? I'm
>wondering if the collisions are assuming twosided polygons and testing
>every single polygon in the object or if you are doing a visibility
>style backface cull of polygons that can't be seen from the current
>"position" before doing tests on actual polygon intersections? I suppose
>you can't do this on object to object collisions, but viewpoint to
>object collisions it should give a significant performance increase as
>you only have a simple dot product to calculate rather than the more
>complex polygonal intersection tests.
No, current implementation did not do back face culling for collision
either. But this is a good suggestion for optimization in the
future. Thanks for your effort in writing the collsion
document.
- Kelvin
--------------
Java 3D Team
Sun Microsystems Inc.
===========================================================================
To unsubscribe, send email to [EMAIL PROTECTED] and include in the body
of the message "signoff JAVA3D-INTEREST". For general help, send email to
[EMAIL PROTECTED] and include in the body of the message "help".
