ODE in simple VR simulation

[Peter...@btinternet.com]

I want to move some fairly simple objects around in a VR type 3D environment.

ODE looks suitable, but I can't find any examples of trimesh-trimesh collision handling code.  I've looked in the demos but noting is obvious,

PeterO

[Alexandre Vaillancourt]

Hello Peter,

Often, we'll use sets of boxes and spheres to model objects, instead of making those objets using trimesh. Those have proven to be working well in our situation. We'll use trimesh only for the terrain.

-V

--
You received this message because you are subscribed to the Google Groups "ode-users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to ode-users+...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/ode-users/80f4e424-cef4-4859-abed-3151747e9bb6n%40googlegroups.com.

[Peter...@btinternet.com]

That's not very helpful when I clearly asked about  trimesh-trimesh collisions. 

I already have low-poly-count bounding shapes of my objects, and if ODE can't do trimesh-trimesh collisoins then please update the table at

http://ode.org/wiki/index.php?title=Manual#Collision_tests_supported that says it does !

PeterO

[Peter...@btinternet.com]

Somehow I had missed "demo_moving_trimesh"  !  Seems to have the code I need :-)

PeterO

[Danny Chapman]

On Wed, 7 Apr 2021 at 14:36, 'Peter...@btinternet.com' via ode-users <ode-...@googlegroups.com> wrote:

Somehow I had missed "demo_moving_trimesh"  !  Seems to have the code I need :-)

ODE's (opcode) trimesh->trimesh is actually pretty good. 

Anything involving convex and cylinder shapes seems to be completely broken (at least with the version I have, which is a couple of years old now, but I'm not aware of any relevant changes since then.

The collision detection part of ODE is a lot weaker than the dynamics, in my opinion. I thought a while ago that it would be quite nice to take the collision detection from PhysX (which is really good, and I believe pretty stand-alone) and combine it with ODE's solver... but never got around to it!

--

http://www.rowlhouse.co.uk/main.html

[Rodrigo Hernandez]

> Anything involving convex and cylinder shapes seems to be completely
> broken (at least with the version I have, which is a couple of years
> old now, but I'm not aware of any relevant changes since then.

Depending on how many years are in a "couple", there was this fix on
convex-ray
collision:

https://bitbucket.org/odedevs/ode/commits/8298b271c37f8a0be207d33321481bb574850f7c

ray collision being the basis for cylinder collision, it is likely it
was fixed,
then again, its been a while for me too since I last ran the demos.
That said, if you find errors with the convex code,
you should probably report them on this list making sure the subject
contains the word convex,
otherwise I tend to ignore or just take a glance at the messages since
most of the times
I have no ready available answers. If there is a report on this, I
probably missed it.

As for adding code to use a proprietary library for collision detection,
I think it is fine as long as it is made optional and the default is
disabled.

-R

> --
> You received this message because you are subscribed to the Google
> Groups "ode-users" group.
> To unsubscribe from this group and stop receiving emails from it, send
> an email to ode-users+...@googlegroups.com.
> To view this discussion on the web visit

> https://groups.google.com/d/msgid/ode-users/CAFf13ptnjtVSnnMm%2BX0ixoxTMwU4RnPjhxcrNLwANJSGpxcLDw%40mail.gmail.com
> [1].
>
>
> Links:
> ------
> [1]
> https://groups.google.com/d/msgid/ode-users/CAFf13ptnjtVSnnMm%2BX0ixoxTMwU4RnPjhxcrNLwANJSGpxcLDw%40mail.gmail.com?utm_medium=email&utm_source=footer