10 replies to this topic

[arnaud]

Description
Hi all,

Here is a small engine I worked a bit. Today everybody is doing rigid body, so I wanted to play with "non rigid" ones :-)

There is several stuff to improve, especially the collision routine which is done with sphere instead of meshes (that's why collision are not accurate, this is not the solver fault).

You can download a working version here:

http://leonard.oxg.f...ny/NonRigid.zip

The solver itself does integration using Verlet algorithm for stabilty. The collisions are a bit twicked to get speed boost, so nothing is very stable when adding a lots of balls. You must have DirectX 9.0 runtime on your machine.

You can rotate the world with the mouse, add objects or revert gravity to get an exiting experience ! (sounds like marketing stuff, huh ?)

The demo is only 25Kb long , so enjoy small toys ! :-)

[Nautilus]

What do you mean by "rigid" and "non-rigid"?

TIA :)

[roel]

rigid body physics is the name for physics simulations where one assumes that the used bodies (the meshes) do not deform as reaction on the forces applied to them.

[Alex]

Is it spring based? I'd love to see different materials..like glass, playdoh, wood behaving nicely...

Alex

[Nodlehs]

Thats pretty cool! Repost when you have mesh collisions working, keep up the good stuff!

[roel]

Alex said:

Is it spring based? I'd love to see different materials..like glass, playdoh, wood behaving nicely...

Alex

I guess it isn't, as it is rigid.

[swordphish]

Iirc, there is such thing as a 'rigid spring' :yes:

[Razor]

roel said:

I guess it isn't, as it is rigid.

Lol, no it's not. It even says in the title: "Small Non Rigid body engine".

EDIT: I couldn't get it to work. The exe just failed to do anything at all, I even tried from the command line.

[corey]

Ran great even with a lot of objects. Rotational limits on the controls were a little odd :)

A couple clipping issues when objects were colliding/coming to stop.

corey

[roel]

swordphish said:

Iirc, there is such thing as a 'rigid spring' :yes:

You can make a spring with a stiffness of infinity and call it rigid, but I guess that it isn't a very stable thing in a physics simulation.

[PoiNT]

Très joli !

Félicitations Arnaud pour cette démo. Est-ce que tu vas rendre les sources dispo ou pas ?

:worthy:

ps: sorry for this message in French, just asking Arnaud if he will put sourcecode available