Monday, April 13, 2009

Selection Sets and Physics

I have now revamped part of the GUI work the why I would have always liked it to work. To make a long story short you can now set up parameters on a selection of objects or even a selection Set. This can make larger scenes much easier to work with.

There are however a few things you should know. If you selected a group of objects, then every selection set that is a subset of the current selection will also be selected. Also if you selected a selection set --all its selection set subsets are also selected.

Now the tricky part is when the editor is disabled. This means that all the objects that are selected do NOT have the same parameters or type. Or alternatively the type selected from the drop down is different. At this stage you can choose to set all of the current selection to what the editor has via the set to selection button.

So anyway a little playing around should make things clear.

The other thing that has changed is the way other objects are handled. In particular soft-bodies. They now have there own iterative solver. Its not different at this stage, but uses a smaller time step from the rest of the simulation, and hence the whole simulation can proceed much faster (ie with larger time steps). You can control this with the internal reps parameters for soft-bodies. Higher means more internal iterations. However this won't solve all stability issues so smaller global time steps will still be generally required for fluid soft-body interactions.

The other nice change is that I now check to see if the soft body simulation is trying to use all the ram. At this point it throws an exception rather than just consuming every last Meg on your machine. In this case either increase the internal reps or reduce the time step.

Don't forget to delete the physics object on older scenes and recreate a new one.

Here it is


Anonymous said...

Too late for easter.

Just to say thank you Delt0r.
It´s so much fun.


Anonymous said...

After reading your description over and over I test softbodies now with a reps setting of 2.
And this seem pretty stable with my scenes.

Good work! Would be even better with more doks - but it looks and feels already a lot easier for starters!


TroY said...


sorry for being late but I'm quite busy at the moment. :(

The new GUI works very well and the integration with the selections plugin is of great help. That's exactly what we need when it comes to "massive" simulations. :)

I got a little problem: Maybe I'm just doing something wrong but it seems like all water is just pouring straight downwards. If I rotate an emitter, that doesn't have any effect. "Normal factor" doesn't have any effect, either. I always use TriMeshes as emitters. Any hints?

By the way, are there any future plans concerning scripting? I could think of turning an emitter on/off, changing a soft body's paramters, ... Scripting would be the most powerful way to do so.

Last but not least, thanks again for your work! And thank you for sharing it. :)