[ODE] Should trimeshes react to physics?

Nguyen Binh ngbinh at gmail.com
Mon Nov 21 16:37:25 MST 2005


The mass will make alot of different when you try some mechanic simulation
scene like the part feeder (Put a shape in part feeder, the heavier end will
come out). But also, ODE precision isn't enough to actually "simulate" such
scenes, especially the friction model.


On 11/21/05, Megan Fox <shalinor at gmail.com> wrote:
>
> Well, it would of course depend on your setup, but "usually," the big
> terrain trimesh has no body. If it has no body, it can not experience
> a force... math-wise, it is an infinitely massive object that is
> unaffected by gravity.
>
> If, instead, you were to assign a body to your huge trimesh? It would
> experience gravity, and it would recieve forces, and you would have to
> set a mass to that body. This scenario isn't terribly useful, though.
>
>
> The concept of mass is unrelated to the geometry you assign to a
> shape, incidentally. The mass tensor matrix you specify can be
> calculated via a particular shape (the dMassSetBox and similar
> functions), but this doesn't have to match the shape of the body in
> question...
>
> For instance: you could create a body, attach a single sphere geom,
> but set a mass calculated via dMassSetBox. This would create a shape
> which had the inertial properties of a box, but the collision and
> friction properties of a sphere. (which isn't sensical or useful, and
> you probably wouldn't notice the difference, but you can do it)
>
> In my scenes, everything just uses mass properties calculated via
> dMassSetBox. I never notice the difference, and it's simple and
> general (easy to determine from the shape's visual AABB, and anything
> physical is visible in my sim). This beats trying to calculate a
> to-the-frame accurate intertial representation of a multi-limbed human
> entity that is constantly deforming (animating).
>
> On 11/21/05, Keith Wiley <kwiley at cs.unm.edu> wrote:
> > I'm trying to understand trimesh roles in ODE. I realize that,
> > according to the docs, they are new and not fully developed. I am
> > using one massive trimesh for a nonoverlapping extruded terrain right
> > now. One thing I have noticed is that the trimesh is unaffect by
> > gravity. If I create the entire terrain at some height up in the
> > air, it doesn't fall down, it just floats permanently in space, even
> > though other objects are rolling and crashing around on the top
> > surface of the terrain, so it must clearly be enabled due to all the
> > collisions it is experiencing.
> >
> > Likewise, since a trimesh is infinitely thin, I'm curious how mass
> > works with respect to trimeshes.
> >
> > There is an option to precalculate the face normals for a trimesh and
> > give them to the trimesh building routine. When should this be used
> > and when shouldn't it be used?
> >
> > And lastly, there is a brief mention of temporal coherence, which as
> > I understand it, reduces overall collision checks by assuming that
> > objects don't move very far between adjacent timesteps. So, it
> > sounds like a good idea. When would it *not* be a good idea to use?
> >
> > Thanks.
> >
> > ________________________________________________________________________
> > Keith Wiley kwiley at cs.unm.edu http://www.unm.edu/~keithw
> >
> > "Luminous beings are we, not this crude matter."
> > -- Yoda
> > ________________________________________________________________________
> >
> > _______________________________________________
> > ODE mailing list
> > ODE at q12.org
> > http://q12.org/mailman/listinfo/ode
> >
>
>
> --
> -Megan Fox
> Lead Developer, Elium Project
> http://www.elium.tk
>
> _______________________________________________
> ODE mailing list
> ODE at q12.org
> http://q12.org/mailman/listinfo/ode
>



--
--------------------------------------------------
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/20051121/f002f882/attachment-0001.htm


More information about the ODE mailing list