PDA

View Full Version : LW 8 dynamics questions...Proton?



agrippa
08-12-2003, 12:03 PM
Hey gang...

It was great to see the cool new stuff that's finding a home in LW8. The new dynamics engine looks very cool. Question for Proton: In the dynamics sims where you blew up robots (die, bad robots), it looked as if the robots, though blowing up nicely, retained the actual "parts" that they were constructed from. If the object being blown up was one continuous mesh (i.e. aircraft fuselage, asteroid, hot air baloon, whatever), how would the dynamics engine handle the sim? Would the single mesh be blasted apart into random chunks? Or do the hard body dynamics only calc the individual components that make up an object?

Thanks in advance...

problemchild
08-12-2003, 01:41 PM
My understanding after talking to Proton was that if the polys are connected, it's the same chunk. So everything can be in one layer and still be scattered to the wind with dynamics. so... will have to slice and dice any contiguous meshes and perhaps object replace just before the annihilation... It's not going to disassemble a subpatch mesh, but a little bit of timely unwelding and I think we'll have some fun.

cheers

-roger

jb_gfx
08-12-2003, 03:04 PM
You could just use a free LScript called CrackIt to cut your object. Or do it by hand.. or maybe we will have such function in LW8's modeler, that would be pretty cool.

agrippa
08-12-2003, 04:18 PM
Just tried crack it! Seems to work pretty well for what it is...though it would be nice to have some sort of "gizmo" (a la Max) to set the core impact point on an object, rather than such a random break up. Or, for example you have an object (could be a bottle or something) where only PART of the object gets broken away...it would be cool to have a bit of control over the effect of the blast. I know, I know...it can be done by hand, but that would be contrary to LW8's new workflow, workflow, workflow mantra (right Deuce?)... :)

hrgiger
08-12-2003, 10:30 PM
I'm still waiting to hear whether the new softbodies will respond to weightmaps. I'm hopin.

SplineGod
08-12-2003, 10:55 PM
The rigid bodies can see an object in one layer as separate parts.
Anything you can click on and hit selected connected will be seen as one piece. An example is one layer with 10 bowling pins. Technically its a single object but all the pieces dont share vertices.
Steve,
If I remember correctly (and I might not) I think I recall that they do see weight maps in some capacity. :)

Emmanuel
08-13-2003, 05:30 AM
Yeah, but lets say I have the bowling pins all in one layer, how does
the simulation know where the pivot/center of gravities are if all the pins share one pivot point ??

hrgiger
08-13-2003, 05:46 AM
Originally posted by SplineGod

Steve,
If I remember correctly (and I might not) I think I recall that they do see weight maps in some capacity. :)

In some capacity? I thought it would either see recognize weight mpas or not recognize weight maps. I didn't know there was an inbetween option??

SplineGod
08-13-2003, 08:41 AM
Emmanuel,
Withoug getting too technical, it just does. :)
I wasnt able to ask that but I did see it in action. You can also grab the pins and adjust their motions paths.
Steve,
Im pretty sure I saw something about weight maps in there.
Its hard to remember it all. :)

prospector
08-13-2003, 09:29 AM
It looks like it takes all Connected polys (even in single layer) and calculates the center of mass for those connected polys and uses those as individual objects.

But that's speculation on my part.

Mike Pauza
08-13-2003, 09:43 AM
Yeah, but does it calculate the actual center of mass or just the center of mass of the vertices? That's a big difference. Most people don't care how the algorithms work, but power users need to know this stuff.

Emmanuel
08-13-2003, 10:38 AM
To me, the fact alone that You can alter the motion path after the calculation is great.
Thats a good way to have "character" parts for example in an explosion where You want a certain part to have a certain behavior.

BTW, I hope the problem with FX_Link has been solved, where You could only set the number of objects at the beginning.
I would love to be able to handle that dynamically, based on the number of particles generated.
Like, when it rains frogs, I would like to let rain as many frogs as there are particles, not a fixed number like it is now :(

Mike Pauza
08-13-2003, 10:39 AM
For all you technical folks, here's an example of what I mean:

For the object below, the center of mass should be about at the origin...and so the object should balance and rotate about its geometric center. If MD calculates the CM based on vertices alone, then the object's CM would incorrectly be positioned at the "X" for this example. I put extra geometry on the "left end" on purpose to illustrate the effect (there are 4 vertices on the right end, and 52 on the left). If NewTek uses a goemetric based CM calculation algorithm then all will be well.

I don't know what algorithms they are using in LW8 so this is all academic, but my real point is that when LightWave8 does ship, NewTek needs to be quite clear in the docs how everything works, so we can figure out how to use it.

-Mike Pauza

taO_js
08-13-2003, 10:49 AM
learning new things everyday!

=]

I luv this stuff