PDA

View Full Version : Motion Designer Gravity Simulation?

roberthurt
05-05-2005, 07:15 PM
So my problem is that I want to try to simulate the gravitational interaction between 3 objects, all mutually attracting one another. In principle it seems like Motion Designer should handle this, but in practice the results are badly erratic. In trying to set this up I've found some odd behaviors that I wonder whether anyone can illuminate. I'm using 8.2 on a Mac (OSX 10.3.9).

My initial setup for a single object is this:

0) Set the end frame to 150 (we want more than the default of 60.

2) Under Motion Options/IK and Modifiers add an FX_Motion plug. I set its resistance low (0.1) and give it a little initial velocity vector (+0.5 in the X direction). Depending on your gridsize/camera distance you need to increase the vector to see the motion.

3) Hit the play arrow to see your motion (not sure if you need to hit

FIRST WEIRDNESS: Notice that, if you still have the object selected, there is a little cross that follows behind the object as if it's slowly accelerating, which is odd since there are as yet no accelerations, and the object has a constant initial velocity.

SECOND WEIRDNESS: On my playback at frame 60 suddenly this cross jumps and begins racing towards the moving object. However if I scrub the timeline, after frame 60, the cross seems to be jumping back and forth rapidly between two positions.

These both are good reason to expect there will be worse problems downstream, and there are...

4) Go to Object Properties/Dynamics and add FX Gravity to the object. Change its type to Point, it's power to -20%

5) Hit the play arrow

THIRD WEIRDNESS: At this point my object begins oscillating wildly back and forth along the X axis, and that odd cross from the last stage just jumps madly back and forth. OK, so I deduce that maybe the object is interacting with it's own gravity, and the effect is large since the object center is close (but not exactly on) the gravity center. We continue...

6) Create two new groups, Obj1 and Obj2. Assign the object's FX_Motion/Mode group to Obj1, and assign it's Object Properties/Dynamics/FX Gravity to Obj 2. This should eliminate self-gravity affecting the object

7) Hit play. OK no weirdness... it just moves along as expected.

8) Clone the object. Offset the two a little bit on the Z axis so they're not on top of one another

9) Go to the 2nd object. For its FX_Motion, change its group to Obj2 and reverse its initial velocity. For it's gravity, change its group to Obj1. This should leave us with two objects whose gravity only affect each other and are starting off in a kind of orbit

10) Hit play. You may need to adjust the gravity strength and radius to see an effect

FOURTH WEIRDNESS: From my experiments, it looks like the objects act as if the gravity is staying put at the initial positions and that the objects are otherwise ignoring one another.

I'd love any insights on how to make the gravity stick to the dynamically moving objects!

Dodgy
05-06-2005, 06:45 AM
This isn't really going to work. At least not in this setup. In this case you have a feed back loop, which this type of simulation isn't going to be able to do. You might want to make a feature request for something like this... Basically as one moves it'll move the other and so on back and forth, which I don't think LW's calculations can resolve.