I don't know if you are still having a problem with this, but there is a similar issue here with a new solution:
http://bulletphysics.org/Bullet/phpBB3/ ... f=9&t=9100
Search found 24 matches
- Fri Aug 14, 2015 3:58 pm
- Forum: General Bullet Physics Support and Feedback
- Topic: Rolling box "sucked" through mesh terrain; video example!
- Replies: 2
- Views: 4537
- Fri Aug 14, 2015 3:32 pm
- Forum: General Bullet Physics Support and Feedback
- Topic: btBvhTriangleMeshShape strange bug
- Replies: 2
- Views: 5649
Re: btBvhTriangleMeshShape strange bug
Sorry this reply is so slow in coming, but thanks for your post with such a clean example of this problem. I just updated from 2.76 to 2.82 and had the same issue, but noticed that it didn't happen if I ported your example to 2.83. Anyway, the problem is in btGjkPairDetector.cpp. If you set m_fixCon...
- Thu Dec 09, 2010 8:45 pm
- Forum: General Bullet Physics Support and Feedback
- Topic: btQuantizedBvh serialization
- Replies: 2
- Views: 2551
Re: btQuantizedBvh serialization
The original allocator was resized 1 larger than it needed to be, and the iteration/recursion on the tree was limited by m_curNodeIndex rather than by the array size, so the extra slot was not necessary in the serialized data.
- Thu Oct 28, 2010 7:23 pm
- Forum: General Bullet Physics Support and Feedback
- Topic: Aabb test
- Replies: 2
- Views: 3124
Re: Aabb test
m_clientObject is the btCollisionObject.
- Thu Oct 28, 2010 7:06 pm
- Forum: General Bullet Physics Support and Feedback
- Topic: Animating a Collision Mesh?
- Replies: 1
- Views: 4303
Re: Animating a Collision Mesh?
You can use a btCompoundShape to do this. Your model likely has an internal hierarchy that you may not need to preserve in the collision, however, so you can flatten it to only have a 1 level deep hierarchy in your compound. Use convex primitives, such as boxes and cylinders/capsules for the body pa...
- Thu Oct 28, 2010 6:56 pm
- Forum: General Bullet Physics Support and Feedback
- Topic: Sharing btCollisionShape with different scale
- Replies: 2
- Views: 4970
Re: Sharing btCollisionShape with different scale
You can use the btUniformScalingShape to accomplish this.
- Fri Mar 19, 2010 2:47 pm
- Forum: General Bullet Physics Support and Feedback
- Topic: Capsule vs Multisphere
- Replies: 1
- Views: 2686
Re: Capsule vs Multisphere
Multisphere is fine. It is slightly slower than Capsule, but not enough to worry about.
- Tue Apr 21, 2009 3:01 pm
- Forum: General Bullet Physics Support and Feedback
- Topic: Sphere / Box collision generates multiple contact points
- Replies: 1
- Views: 4720
Re: Sphere / Box collision generates multiple contact points
Bullet allows some penetration, and because of this, you can have multiple contact points. Bullet keeps up to 4 points, 3 to specify a contact triangle, and 1 that represents the deepest point of penetration. In your case, if you only want to consider 1 point, you can most likely just use the point ...
- Tue Apr 21, 2009 2:51 pm
- Forum: General Bullet Physics Support and Feedback
- Topic: Simplest possible C/D between two objects?
- Replies: 7
- Views: 6450
Re: Simplest possible C/D between two objects?
In addition, adding a simple AABB check before the processCollision call can make Dominik's code fairly optimal.
- Fri Jan 02, 2009 11:51 pm
- Forum: General Bullet Physics Support and Feedback
- Topic: Use only narrow phase collision detection
- Replies: 2
- Views: 3766
Re: Use only narrow phase collision detection
Yes, this can be done. Setup a btDefaultCollisionConfiguration, btCollisionDispatcher, and btDispatcherInfo. If you want easy object management, you can use a btCollisionWorld, but if you manage your own objects and pairings like it sounds like you are doing (using only Bullet's narrow phase), this ...
- Tue Aug 14, 2007 7:27 pm
- Forum: General Bullet Physics Support and Feedback
- Topic: Bullet Ray Cast
- Replies: 10
- Views: 12974
Intuitively it seemed like the correct point should have been available during the last step of the subsimplex cast, but I couldn't convince myself whether the correct point was there or not when stepping through the algorithm. I will probably switch to using the GJK cast instead as Erwin suggests, ...
- Tue Aug 14, 2007 5:42 pm
- Forum: General Bullet Physics Support and Feedback
- Topic: Bullet Ray Cast
- Replies: 10
- Views: 12974
The support in the motion direction would not be the correct point in general. Imagine a triangle moving in the direction of one of its vertices, with a glancing collision just at another vertex. The support would pick the wrong vertex in this case. A support that is along the returned collision nor...
- Tue Aug 14, 2007 5:04 pm
- Forum: General Bullet Physics Support and Feedback
- Topic: Bullet Ray Cast
- Replies: 10
- Views: 12974
- Wed Aug 01, 2007 4:41 pm
- Forum: General Bullet Physics Support and Feedback
- Topic: Box (instead of ray) tracing?
- Replies: 6
- Views: 9465
I have been using this code extensively, and it works as it is supposed to with the exception of the normal direction. Other than this issue, I have tested this code against many different convex shape types and have had no issues once I fed my data to Bullet correctly. Note that this last point is ...
- Tue Jul 10, 2007 3:24 pm
- Forum: General Bullet Physics Support and Feedback
- Topic: Suggestion for higher precision velocity integration
- Replies: 5
- Views: 5733
Have you tried Runge-Kutta? It is usually considered the best general case numerical differential equation solver, and may work for your situation as well. Second order Runge-Kutta (RK2) is sometimes adequate, but RK4 seems to be the most often used method. Here are some links: http://mathworld.wolf...