[ODE] Trimesh vs Convex Hull

Nguyen Binh ngbinh at gmail.com
Fri Mar 31 14:33:48 MST 2006


I think there are rooms for improvement in Bullet + ODE pair.
So I think it should be good if we have 2 versions:
1) Current ODE
2) ODE + Bullet

There are some bugs in my implementations of cylinder (now it's capsule) but
I think they are still there.
Anyone should fix it before 1.0 I'd guess.


On 3/31/06, Andy O'Neil <aoneil at austin.rr.com> wrote:
>
> Hi,
>
> There already is Bullet support in ODE in the main (old) branch, Erwin
> Coumans added it quite a while ago, but it never made it over to any other
> branches.
>
> http://cvs.sourceforge.net/viewcvs.py/opende/ode/ode/src/
>
> I've used ODE with Bullet and I can confirm that it works. I ran into a
> couple of issues, such as the collision pair cache being a static global
> in
> the BulletODECollide lib. If you link this against the ODE .dll, then link
> it as a lib against your project, you'll get some issues when invalidating
> Bullet geoms. I dealt with this by building Bullet / BulletODECollide /
> ODE
> all as libraries.
>
> Also, I believe that Bullet is more focused on techniques than
> performance,
> so that might be an issue for some users. Things might have changed
> recently
> on this front however, so take that statement with a pinch of salt.
>
> Andy
>
> ________________________________
>
> From: ode-bounces at q12.org [mailto:ode-bounces at q12.org] On Behalf Of Nguyen
> Binh
> Sent: Friday, March 31, 2006 6:58 AM
> To: Mohsin Hasan
> Cc: ode at q12.org
> Subject: Re: [ODE] Trimesh vs Convex Hull
>
>
> I 'd doubt it.
> Anyway, Bullet has a good implelementation of GJK algorithm so it supports
> convex shape.
> It would be nice if we couple Bullet in ODE.
>
>
>
> On 3/31/06, Mohsin Hasan <mohsin.hasan at trivor.com> wrote:
>
>         I'm not sure if this has been discussed earlier but wouldn't it be
> great if ODE natively supported a ConvexHull primitive. It would give
> better
> collision representation of a mesh compared to simple primitives besides
> being more efficient and accurate in terms of detecting collision compared
> to Trimesh. Does anyone else also think that they are more important than
> Trimesh/Trimesh.
>
>
>
>         If supporting ConvexHull is not possible for whatever reason, is
> it
> possible to do something similar using Trimesh? I mean if we can generate
> a
> convex hull of our geometry through whatever way and pass it as Trimesh to
> ODE, is Trimesh-Trimesh collisions for convex hull meshes accurate?
>
>
>
>         Mohsin
>
>         ------
>
>         See tech demos of our engine supporting ODE based Physics at
> http://www.itrango.com/demo.html
>
>
>
>
>         _______________________________________________
>         ODE mailing list
>         ODE at q12.org
>         http://q12.org/mailman/listinfo/ode
> <http://q12.org/mailman/listinfo/ode>
>
>
>
>
>
>
>
> --
> --------------------------------------------------
> Binh Nguyen
> Computer Science Department
> Rensselaer Polytechnic Institute
> Troy, NY, 12180
> --------------------------------------------------
>
>


--
--------------------------------------------------
Binh Nguyen
Computer Science Department
Rensselaer Polytechnic Institute
Troy, NY, 12180
--------------------------------------------------
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://q12.org/pipermail/ode/attachments/20060331/7ff18834/attachment.htm


More information about the ODE mailing list