PDA

View Full Version : NewTek - HIRE THIS GUY NOW!



Mike Pauza
08-20-2004, 04:22 PM
NewTek - HIRE THIS GUY NOW!

--------------------------------------------------------------------------------

This is maybe the sharpest guy in Dynamics Simulation right now, and he's looking for a job:
http://www.cc.gatech.edu/%7Ecarlson/

Our existing dynamics system (fundamentally based on points) most needs robust collision detection, dynamics hyerarchies, and smart particles (like RealFlow).

The guy mentioned above is not a particle guy (so he's not the man to spruce up our existing dynamics), but he's on the verge of comming up with the absolute best "Grid based simulation system" (think hypervoxels) around. This is the type of stuff we're seeing from Maya Fluids...only much better. His existing code will quickly simulate incompressible fluids and solids together at the same time. It won't do soft body's or compressible fluids...YET, but he's thinking about it and there's no reason that can't be done as well to get ONE PIECE OF SIMULATION CODE THAT CAN SIMULATE VIRTUALLY ANYTHING.

Check out his animations, then pick your jaw up off the floor.

-Mike

lesterfoster
08-20-2004, 05:28 PM
Boy. This stuff is quite good. Is it all LightWAVE?

CB_3D
08-20-2004, 06:25 PM
wow, impressive stuff, indeed.

wacom
08-20-2004, 09:46 PM
I say put some of my upgrade money on this guy!

js33
08-20-2004, 10:15 PM
It looks like Real Flow has some competition. :D

Cheers,
JS

CB_3D
08-20-2004, 11:20 PM
From my limited understanding this is so much more than what real flow (which is excellent) does. This is simply another league, and it´s amazing seeing these kind of results in a student work.

pure genius.

Hervé
08-21-2004, 01:13 AM
I believe he was already contacted by XSI....:D :D

If NT would cut down.... large parties events, tee shirt distribution, crispy creams.... they 'll have themoney to hire him...$$$.:D :D :D

I am joking....

revolution is on the way..... we are mathematics !:D

tokyo drifter
08-21-2004, 01:24 AM
Originally posted by CB_3D
From my limited understanding this is so much more than what real flow (which is excellent) does. This is simply another league, and it´s amazing seeing these kind of results in a student work.

pure genius. What makes this different then Real Flow?

Mylenium
08-21-2004, 03:24 AM
I really don't understand your excitement, people... This is nothing really new and it's not even competition for RealFlow. What you all are forgetting is the simple fact, that no algorithm or formula, no matter how fast or good, is of any use unless it is implemented in a usable way. You see, Fluid simulation such as in Maya and Real Flow has been around several years before, but did anybody ever use them except for geeky scientific stuff? No! Only after those programs implemented some user interface with understandable paradigms, everybody began to use their functionality. Under those circumstances it is wrong to assume that we would get a good and production prove tool just if NT hired this guy.

Mylenium

pauland
08-21-2004, 04:36 AM
Originally posted by Mylenium
I really don't understand your excitement, people...

Mylenium, you are a spoilsport.

We all have customers that want animations of their logos dropped into water and coloured balls.

Anyway, that's what I'd tell my wife if the LW plugin came on the market.. ;-)

Mike Pauza
08-21-2004, 11:37 AM
Carlson's simulation is quite different from RealFlow. Realflow is particle based and therefore much slower, but RF is
"better" at small volumes of water and "not so great" at larger volumes. Carlson's code has not yet been integrated
into an animation system, but it's similiar (but with more potential) to Maya's Fluid Effects which is a proven and highly
vaunted production tool.

Hervé said Avid may have alread contacted Carlson. That would be a shame.



-Mike

Mylenium
08-21-2004, 01:43 PM
Carlson's simulation is quite different from RealFlow. Realflow is particle based and therefore much slower,

Well, his sims are, too. RealFlow also uses a simple Lagrange integration method (admittedly much simpler) to calculate energy feedback between "particles" in the grid, Maya uses Navier-Stokes, which is totally different and actually not comparable to this.

Sure it would make some nice addition to LW's toolset, but once more, just having some code does not make a good app or plugin. There is so much that hinges on a good interface at the user end...

Mylenium

Mike Pauza
08-21-2004, 02:44 PM
Originally posted by Mylenium
Well, his sims are, too. RealFlow also uses a simple Lagrange integration method (admittedly much simpler) to calculate energy feedback between "particles" in the grid, Maya uses Navier-Stokes, which is totally different and actually not comparable to this.

Sure it would make some nice addition to LW's toolset, but once more, just having some code does not make a good app or plugin. There is so much that hinges on a good interface at the user end...

Mylenium


I don't see your point. Simulations are either corpuscular (particle based) or continuous (field equations).

RealFlow is corpuscular. They may use a very old "energy grid" technique (See "Simulation using particles" by Hockney & Eastwood) to approximate or even throw away distant particle pair calculations (otherwise it's an N^2 process which is very computationally prohibitive), but the inclusion of a grid for calulation simplification doesn't make a particle system a field equation simulator!

Yes, Carlson and Maya (Jos Stam) both use finite difference methods to solve the Navier Stokes field equations...If your solving field equations for incompressible fluids then your using the Navier Stokes eqations, or at least your own varient of them.

What Carlson is doing differently from other computational fluid dynamics guys is trying to come up with a method that can solve for incompressible materials "in general"...fluids, solids, iamorphous solids, multiple materials, etc. He hasn't tackled compressible materials yet, but that's his next goal. The whole point of having one simulator that can handle "anything" is that you get to stop struggling with how to come up with an effect, you can simply do anything you want with simple controls. For example...If you want to melt a teapot, you simply assign it a material, set a hot plate under it, and press go!

I realize you have to integrate the thing into LightWave, but it's very doable!

If NewTek doesn't get into finite difference simulation, then at the very least we need a particle simulator with user definable interparticle interaction (that's easy to code). I would rather have both!

-Mike Pauza

Librarian
08-21-2004, 04:04 PM
Mike, don´t wait for NewTek.
I think they have to implement tons of other basic stuff before carry on with such advanced stuff.

My suggestion is take action by yourself.
I know you hold a degree in computational physics, so it´s your playing-field :).

Dodgy
08-21-2004, 06:30 PM
I agree, the particle interaction stuff could use some work. Somebody code this up already :)

Mylenium
08-22-2004, 03:53 AM
Ah, I see. Sorry, I'm not that terribly into the theoretical side of things, it was more like quoting the little bits I know and some marketing hubub I heard every now and then... Your explanation makes a lot of sense. Still, as Dodgy said, there are more pressing issues in LW and at some point I too begin to wonder, if you are such a fit guy in the theoretical stuff, why you don't develop something yourself? You see, every now and then you, Lightwolf and a few others pop up spreading your wisdom and once the truth is out, it's even more frustrating to not see such things coming to LW but knowing that some able guys already have pinned down the problem.

Mylenium

Hervé
08-22-2004, 09:51 AM
I'd say forget about this guy, look here, we now have Norm with a real new tool...Fantastic !

Watch the demo he posted !

http://vbulletin.newtek.com/showthread.php?s=&postid=206725#post206725

:D :D ;)

Big potential !

hunter
08-22-2004, 12:16 PM
Originally posted by Hervé
I'd say forget about this guy, look here, we now have Norm with a real new tool...Fantastic !

Watch the demo he posted !

http://vbulletin.newtek.com/showthread.php?s=&postid=206725#post206725

:D :D ;)

Big potential !

Sure that's cool, but what does it have to do with dynamics? That's captured from modeler.

Hervé
08-22-2004, 12:26 PM
well not much in fact, but since you were talking about new stuff, I just thought tat wouldinterest you, and let's also say it looks more " right now"....:D ;)

hunter
08-22-2004, 07:38 PM
Ah, I see, Maybe in 8.0b?

Kurtis
08-23-2004, 06:52 AM
We have a request list, although people tend to post in General Discussion and Community (like this thread). Members of the Marketing and Engineering teams do keep an eye on it. It is LW - Feature Requests (http://vbulletin.newtek.com/forumdisplay.php?s=&forumid=27).

Although you are welcome to post threads like this in other forums, this kind of thing is what LW - Feature Requests (http://vbulletin.newtek.com/forumdisplay.php?s=&forumid=27) is for, and will get the best effect there. If you would all like, I would be happy to relocate this thread there for you.

And of course, there is also the e-mail address [email protected] which goes directly to the Engineering department.

Mike Pauza
08-23-2004, 09:36 AM
Mylenium:

I want to develop dynamics scripts (I have no desire to use the SDK), but LScript just isn't powerful enough yet...there's no point/vertice control in Layout for example.



Kurtis:

Thanks for the feature requests link. I just posted one.



-Mike

nerdyguy227
08-23-2004, 12:50 PM
It would be a good investment to hire someone with advanced skill in what many intermeadiat users would like to do with more simple, esier tools.