A place to discuss everything related to Newton Dynamics.
Moderators: Sascha Willems, walaber
by neimod » Thu Aug 07, 2008 11:46 am
Hi all,
I've got a problem with the ball physics in my game.
The game has a Tree collision which represents the level world, and it has alot of ramps.
When the ball is going up a ramp, it rolls up a bit, but then it hits the ramp and takes a tiny flight into the air, and then quickly hits the ground again, and then it rolls up a bit, takes another tiny flight into the air, and then hits the ground again, etc.
Is this normal with a physics engine?
This is a little bit annoying because it makes the velocity go up and down constantly with tiny amounts, and I need this velocity to be smooth because I am amplifying it and using it for force feedback.
For moving the ball I am using a force which is always in the XZ plane.
Thanks in advance.
-
neimod
-
- Posts: 8
- Joined: Wed Apr 23, 2008 9:34 pm
by agi_shi » Thu Aug 07, 2008 11:50 am
When you NewtonTreeCollisionEndBuild(), did you set the last parameter (optimize) to 1? How is your tree collision modeled, are there too many unneeded triangles that might cause edge issues? Implement the debug display to see what the physics engine sees in terms of collision geometry.
-
agi_shi
-
- Posts: 263
- Joined: Fri Aug 17, 2007 6:54 pm
by neimod » Thu Aug 07, 2008 12:04 pm
agi_shi wrote:When you NewtonTreeCollisionEndBuild(), did you set the last parameter (optimize) to 1? How is your tree collision modeled, are there too many unneeded triangles that might cause edge issues? Implement the debug display to see what the physics engine sees in terms of collision geometry.
Yes I am setting the optimize parameter to 1. The model is made using blender with bezier curves, and then converted to a poly mesh.
Here's a screenshot of the collision tree:
[url]dead[/url]
Last edited by
neimod on Thu Aug 28, 2008 2:16 am, edited 1 time in total.
-
neimod
-
- Posts: 8
- Joined: Wed Apr 23, 2008 9:34 pm
by Julio Jerez » Thu Aug 07, 2008 12:26 pm
is this with 1.53?
-
Julio Jerez
- Moderator
-
- Posts: 12249
- Joined: Sun Sep 14, 2003 2:18 pm
- Location: Los Angeles
-
by neimod » Thu Aug 07, 2008 12:38 pm
This is with 1.53 yes. I can try it with the 2.0 beta shortly.
-
neimod
-
- Posts: 8
- Joined: Wed Apr 23, 2008 9:34 pm
by Julio Jerez » Thu Aug 07, 2008 1:24 pm
we are doing some work to smooth contact normals across edged for round objects, but it is only for archemeadia.
would you consider, migrating to 2.0?
-
Julio Jerez
- Moderator
-
- Posts: 12249
- Joined: Sun Sep 14, 2003 2:18 pm
- Location: Los Angeles
-
by neimod » Fri Aug 08, 2008 9:53 am
I have migrated to SDK 2.0 beta 17, but sadly the problem still is there.
Infact, it does not seem much different at all
.
-
neimod
-
- Posts: 8
- Joined: Wed Apr 23, 2008 9:34 pm
by Julio Jerez » Fri Aug 08, 2008 10:02 am
I did add code to solve that problem
if you migrated to to 2.0, do you have a demo that I can debug?
-
Julio Jerez
- Moderator
-
- Posts: 12249
- Joined: Sun Sep 14, 2003 2:18 pm
- Location: Los Angeles
-
by neimod » Fri Aug 08, 2008 2:13 pm
I can't give you a demo at the moment.
The whole problem seems that the ball is too easily starting to bounce...
Is it not possible to make the ball "stick" more to the ground, and it will only come loose if somekind of force threshold is exceeded?
-
neimod
-
- Posts: 8
- Joined: Wed Apr 23, 2008 9:34 pm
by Julio Jerez » Fri Aug 08, 2008 2:22 pm
as far as I know the problem is fixed, if you cannot product a demo to reproduce the error I cna not help.
There is not "stick more to the ground" physics law in newton, I do not know what that is.
I you cannot trust us, I suggest you try another physics solution all of the comercials one are free now.
-
Julio Jerez
- Moderator
-
- Posts: 12249
- Joined: Sun Sep 14, 2003 2:18 pm
- Location: Los Angeles
-
Return to General Discussion
Who is online
Users browsing this forum: No registered users and 15 guests