PDA

View Full Version : Motion Designer Tests


Mike Pauza
07-12-2003, 05:49 PM
I'm researching how to use structural engineering type meshes (ie not regular geometry) to accurately simulate various scenarios using Motion Designer. Below are some animated examples. I'll try to post a tutorial in the near future.




http://mywebpages.comcast.net/robpauza/Cup_Soft.avi (0.46 MB)

http://mywebpages.comcast.net/robpauza/Cup_Hard.avi (0.43 MB)

http://mywebpages.comcast.net/robpauza/BridgeEarthquake2.avi (4.5 MB)

http://mywebpages.comcast.net/robpauza/CubeArray_Twist.avi (1.8 MB)

http://mywebpages.comcast.net/robpauza/CubeArray_Stretch.avi (1.3 MB)

http://mywebpages.comcast.net/robpauza/Truss_Shake.avi (0.26 MB)

http://mywebpages.comcast.net/robpauza/Truss_Snake.avi (0.89 MB)




I plan to keep creating these little sims, so please comment. :)

-Mike Pauza

Kvaalen
07-12-2003, 07:06 PM
Looks very cool! A tutorial would be great! :thumbsup:
You got some really great ideas there!

I think the cup could have a little more gravity, also it seems to slide a bit afterwards. :shrug:

artista3d
07-12-2003, 07:13 PM
Pretty nice sims you got there, pal! :thumbsup:

By the way, is there an easy way of creating this "X" substructures on objects from quads?

Keep posting new sims!!!

Triple G
07-12-2003, 07:30 PM
Very nice stuff...I really like the earthquake one. Very impressive. The snake one looks like it could be very useful also. Keep up the good work! :beer:

BoulYaBase
07-12-2003, 07:49 PM
Hey Mike,

Great stuff! Guess you got some server space. Hehe... I was wondering why you hadn't taken me up on my offer. :)

Seeing these samples makes me wonder.... (pardon me if this is outright impossible, I don't know much about writing plug-ins or LScripts... :hmm: )

If you built something out of truss-like components (as you've done above), how hard would it be to use a Weight Map to create fracture/break points in the mesh. When an object collides with another object or is stressed via torsional or rotational forces, the magnitude of the strain on the vertices could be compared to the Weight Map. If the strain exceeds the Weight Map value, the points could be Unwelded and the object could begin to come apart.

Is that sort of thing possible via a plug-in, or would the Unwelding process not work when called by Layout?

Mike Pauza
07-12-2003, 07:49 PM
Kvaalen:

Thanks.

With the cup sim I was having trouble getting MD to handle collision and sliding friction properly...I don't think those are it's strong porints, but it can handle spring based deformation fairly well. I hope to post a tutorial soon.




GoodieBR:

Thanks dude. I wish there was an easy way of getting all those "X"'s in there. It's really not that hard but quite tedious...a plug-in that takes an all-quad model and automatically converts it into a "simulation mesh" would be really sweet.




Triple G:

Thanks man...the quake one is my favorite too. I hope I can create some little characters/machines using these techniques in the next few days.




-Mike Pauza

Mike Pauza
07-12-2003, 08:00 PM
BoulYaBase:

Hehe...I did try to send those vids to you, guess it didn't get through...email servers don't seem to like large attachments much. :)

I've been told that Layout cannot create geometry on the fly, so fracture would be hard at this point. That would be very nice though.

I'm writing some sim code that's working pretty well (mostly an illustration of MD's weaknesses / potential fixes...see the "Biggest problem with SOFT BODY DYNAMICS???" thread), and I know someone who is going to try and convert it into a LW plug during the next month or so.




-Mike Pauza

NanoGator
07-12-2003, 08:07 PM
I can't wait until body dynamics (soft or hard) gets better with LW, including a better interface. There's lots of things I'd like to do with it.

Cool stuff. :)

Celshader
07-12-2003, 08:44 PM
Cool, Mike -- I dig the bridge!

Originally posted by Mike Pauza
With the cup sim I was having trouble getting MD to handle collision and sliding friction properly...I don't think those are it's strong porints, but it can handle spring based deformation fairly well. I hope to post a tutorial soon.

For what it's worth, in Motion Designer friction is a polygon property -- not a point property.

Also, the FX_Collision planes in LW do not pass along any friction information to MD, no matter how high you crank up the Friction on the FX_Collision object. You're going to need to make a collision object if you want friction. Load the floor object into MD as a Collision object. Set the Friction on the floor object to 100 and you'll see more friction happen in the calculation. Make sure the Skin Thickness of the floor is set high enough to keep the cup from falling through the floor. :D

:shrug:

Mike Pauza
07-12-2003, 09:14 PM
Celshader:

Thanks!!! I have a strong background in numerical dynamics simulation (so I know exactly what I want to do), but I'm still almost a total MD Newbie...hehe. I really appreciate your suggestions...I'll try them out tonight. :)




Everybody Else:

Check out Celshader's thread if you haven't already, she has some really sweet character animation with MD clothing, MD hair, and MD ears:

http://www.cgtalk.com/showthread.php?s=&threadid=75158

This is probably the most impressive use of MD I've seen.




-Mike Pauza

HowardM
07-12-2003, 09:51 PM
Originally posted by Mike Pauza
Thanks dude. I wish there was an easy way of getting all those "X"'s in there. It's really not that hard but quite tedious...a plug-in that takes an all-quad model and automatically converts it into a "simulation mesh" would be really sweet.
-Mike Pauza [/B]

basically your Xs are like Substructure right?
hopefully someone would be able to code the new md so it could automatically make X structures work with all quad objects...

great tests and movies! keep em coming!
:)

..and Jen, you RULE!...you truly are the queen of celshading! If DJ hadnt nabbed you I surely would of!
:p

Celshader
07-12-2003, 10:13 PM
Heh... :D Thanks, guys.

Originally posted by Mike Pauza
Thanks dude. I wish there was an easy way of getting all those "X"'s in there. It's really not that hard but quite tedious...a plug-in that takes an all-quad model and automatically converts it into a "simulation mesh" would be really sweet.

Hey, I think I just found one for you -- the Spikey tool!

Load your mesh into Modeler. Select the polygons that you want to "X," or deselect everything so that all polygons will be affected.

Turn on Multiply->Extend->Spikey tool.

Type n to activate the numeric panel for the Spikey tool. Leave the Spike Factor at 0%.

This will "X" all of your selected quads.

:beer:

HowardM
07-12-2003, 10:22 PM
STOP! STOP making us love you! great tip! :)

you need to start a school, a celshading school for women!
we needs mo' dames like yous, ya know!? :p

..well mike, will spikey Xs work the same way?

Joviex
07-13-2003, 02:30 AM
Originally posted by Mike Pauza
BoulYaBase:

Hehe...I did try to send those vids to you, guess it didn't get through...email servers don't seem to like large attachments much. :)

I've been told that Layout cannot create geometry on the fly, so fracture would be hard at this point. That would be very nice though.

I'm writing some sim code that's working pretty well (mostly an illustration of MD's weaknesses / potential fixes...see the "Biggest problem with SOFT BODY DYNAMICS???" thread), and I know someone who is going to try and convert it into a LW plug during the next month or so.




-Mike Pauza


Hey Mike, still here in vacation land, get back to you soon.

Great stuff you have going on, can't wait to get back and dig in.

Kvaalen
07-13-2003, 09:16 AM
Originally posted by HowardM
basically your Xs are like Substructure right?
hopefully someone would be able to code the new md so it could automatically make X structures work with all quad objects...


It is possible to do this with David Ikedas Create Pole plugin.
Select all the quads you want and then activate it. :thumbsup:

Brett H.
07-13-2003, 05:25 PM
Spikey won't do the substructure he's using. See attached. You need to connect the corners of a cube for example, through the middle of the cube to gain that substructure that makes this work. (the pic is of an uncompleted cube with just the substructure created to show what I mean).

Brett

HowardM
07-13-2003, 05:44 PM
ok, how about a triple>mirror triple?
:)
if we can triple polys instantly, theres gotta be an EASY way to triple them twice, once reversed....
!!!!!
:thumbsup:

Mike RB
07-13-2003, 05:56 PM
? Just use the spikey tool with no offset.

Mike Pauza
07-13-2003, 08:35 PM
Celshader, HowardM, Kvaalen, Mike RB, Brett H:

Brett's correct that Spikey is kinda the right idea, but it won't work how I want it to on internal structure. See Brett's pic...a 10x10x10 cube array would require 1000 of these things. This sort of setup is great for solids. For shells/cloth the concept of angular springs really needs to be utilized...I still need to explore this idea in detail.

Modeling truss objects (like a bridge) is a little more staightforeward.




amorano:

Good to hear from you. Have a great vacation man.



BTW. I started looking at cloth and air resistence last night with decent results. Tonight I'm going to try and build a glider that actually flys.

Thanks for the interest. -Mike Pauza

HowardM
07-13-2003, 09:04 PM
oooooh ahhhh! that would be killer, let it loose on random MD winds!!! I WANT THAT SCENE!
:p

is there any way to combine the 3D techniques of volumetrics with MotionD to create structures that have thickness?
bet it would take forever, but maybe it could be REALLY accurate....but Im no math wiz like you guys, what do I know?!
:)

Mike Pauza
07-13-2003, 09:46 PM
HowardM:

I'm very interested in "density field" simulation (like water bubbles, air heat ripples, sci-fi force fields, etc) rendered w "HV-like" technology, and yes the whole MD/HV connection in general seems very promising, but IMO you won't see that sort of thing in LightWave for a while. That seems to be kinda the direction Maya is going though.

If I can can get the glider scene working I'll send it to you. :)

-Mike

HowardM
07-13-2003, 10:27 PM
cool! thanks!
but could volumetric MD be used for solid objects like a bridge or a cup?...youre talking more about particles affecting MD type force fields...
I was just thinking out loud before , take an object, figure out its 3D space volumetricly (sp?!) and use that instead of X structures...is that what Maya is doing?

Kvaalen
07-13-2003, 10:49 PM
Originally posted by Mike Pauza
Celshader, HowardM, Kvaalen, Mike RB, Brett H:

Brett's correct that Spikey is kinda the right idea, but it won't work how I want it to on internal structure. See Brett's pic...a 10x10x10 cube array would require 1000 of these things. This sort of setup is great for solids. For shells/cloth the concept of angular springs really needs to be utilized...I still need to explore this idea in detail.


I didn't mention the spiky tool, I mentioned David Ikedas Create Pole. Here is an example. It just takes one click:

http://mapage.noos.fr/samuelLK/CreatePoles.jpg

Et voila! :thumbsup:

HowardM
07-13-2003, 11:17 PM
but see, Mikes X structures make solid cubes, where those are just X structures on edges, so does spikey, tipled polys, etc...so we need a way to compute solid structures to get good results.
it needs to be like Mike has shown us, built with solid building blocks...so whats the easiest way to build em?
that would be cool if lets say you made a nice realisitic bridge model, then either made or took points off it, and a plugin creates the X structures from those points, that the new MD4000 would use as a cage for the real model...or can this actually be done with MD_plug now if someone writes a plugin to make the cages?
:applause:

Mike Pauza
07-14-2003, 09:10 PM
I got an MD Glider working. Yeah baby yeah!

It's not very pretty, and it took me way too long, but I'm excited. Hopefully I'll have video links up tonight or tomorrow. I can't really "fly it" interactively because there's no way for me to control flaps or give it any power (am I wrong about that?). Anyway, this sort of sim shows the power of MD's parallel air resistence and sets the stage for more complex dynamic machines...like ILM's pod racing simulations for example.




HowardM:

Without getting into a long winded discussion about different types of simulation, Maya's roots are in scientific visualization so they have a number of fairly interesting tools, but they seem to be focusing on grid based (as opposed to particle based) fluids right now...that's where they shine right now...that and rigid bodies.

About your "HV Idea" I think you are talking about something like the marching cubes algorithm...if you constuct an array around geometry, can can reduce it to a cuberile approximation and possibly "shave" the cubes that are partially in and partially out of the original geometry to try to get a decent fit.

A tool like like would be very useful in some cases and worthless in others. A tire for example would be a bad object to approximate that way. But yeah, varients of tools like this would make regridding a lot easier.




-Mike Pauza

Facial Deluxe
07-14-2003, 10:30 PM
That kills....:applause:

Mike Pauza
07-15-2003, 04:31 AM
Here's the glider video (BTW, does anybody know if this is this the first MD flying machine???):
http://mywebpages.comcast.net/robpauza/!Glider2.avi


Here's a falling sheet video (what about a hang glider made out of fabric like this?):
http://mywebpages.comcast.net/robpauza/FallingSheet.avi


Still far from sold on MD, but the parallel air resistence is super cool I have to admit. Hopefully I can play with some glider steering and wind before I move on to more machines. Again, all this is basically just my attempt to test what MD's strengths and weaknesses are. Please comment if you can. Oh, if you want the scene just email me at sharonpauza@aol.com or mike@3dpiranha.com.

-Mike Pauza




Facial:

Thanks a bunch Oliver. How are you?

Steve Warner
07-15-2003, 04:34 AM
Hi Mike,

Can you check the link to the glider video? Seems to be down.

The sheet looks cool! You rock! :)

Steve

Mike Pauza
07-15-2003, 05:20 AM
Thanks Steve.


@#$#@%%@#$#$!!$!$!$!$!%^^$%^&$&$ I'll see if I can get it on-line tomorrow night. If anybody wants me to email you the scene and or video just let me know.

-Mike

ViperGT
07-15-2003, 07:14 AM
Hey Mike/anyone trying to watch the glider vid:
Just get rid of that "!" before the filename and it works. ;)

Love what you've been doing with md Mike. Just starting to get into it lately because of the projects I've been working on. This stuff gives me some hope that it'll pull through for me :rolleyes:


-Viper

Mike Pauza
07-15-2003, 02:29 PM
They should work now, but here they are again:

http://mywebpages.comcast.net/robpauza/!Glider2.avi
http://mywebpages.comcast.net/robpauza/FallingSheet.avi




ViperGT:

Thanks. Best of luck with your MD projects.




-Mike

anieves
07-15-2003, 02:49 PM
Looking good Mike! that bridge looks awesome. Reminds me of a tv ad with a bridge displaced like that by a radio blasting in a car... cool

Mike Pauza
07-15-2003, 05:49 PM
anieves:

Thanks man. That commercial sounds cool.




BTW, The glider can probably be controled by morphing the steering polys. For those interested, take a look at (I'll be attempting some of this tonight.):

http://www.imadjinn-cgi.com/TUT/curtain_tute/curtain_tute.html




-Mike

Gnome
08-13-2003, 07:20 PM
I was wondering with the bridge simulation, how you managed to move the different "fixed" points in relation to one another.
As I understand it, the entire bridge has to be one welded set of geometry. You then label the fixed portions with one surface and the rest with another. To vibrate the model you grab the whole thing and give it a wiggle. But you have managed to get different regions on the bridge to animate differently.

Are the coloured sections in the foreground fixed, or are they part of the dynamic bridge?

If so, is there a way you can think of to animate the base of the two main stands in different ways so the connecting trusses are pulled in different directions?

CGTalk Moderation
01-15-2006, 03:00 PM
This thread has been automatically closed as it remained inactive for 12 months. If you wish to continue the discussion, please create a new thread in the appropriate forum.