looking for a training course or collaboration

Engines, Libraries and Tools Xith3D Forums
#1

Hi folks,

I’m developping the GUI (in java) of a software which simulates reactions in a cell.
One part of the GUI should display the geometry model (meaning the shape of the cell he’s doing the simulation) in 3D. And we would like to add some kind of interactions
with this 3D view.
I contact you because I dont know anything about java 3d or xith 3d and I guess it will take me long time to make something nice (or may be you will tell me that it’s easy like hell and I dont need any help! :D)
Actually my boss proposed me to go to a training course specialised in java 3d stuff or to do a collaboration.
Would you know such a course or a team/someone with who we could collaborate?
I’m working in a Scientific European Institute.

thanks to answer me

mathilde

#2

What kind of simulation is this ?
Reactive flow simulations using FVM or FEM ?
Molecular Dynamics in a box ?
Simple animation or streaming datasets ?
Simple mouse interactions with your 3D view ?

Some more info. would help anyone who may be interested …

Anyways, you should seriously consider taking a 3D Java course if you get to find one !

#3

Best of luck finding a course. There’s nobody offering one that I know of, anywhere. If you are familiar with 3D graphics programming in general, and somewhat with any scene graph technology, you’re 80% of the way there. The rest just is adapting to the Java3D/Xith3D-specific way of doing things. More general 3D training courses such as OpenGL primers would not be much use to you though as you don’t have the level of control they do.

Edit: Hit Post too early - If you are doin sci-viz, you may also need more than J3D/Xith3D can provide as you’ll probably need proper volume rendering capabilities. For that, you’ll really need to look into one of the C++ alternatives such as OpenSceneGraph or OpenSG. My project, Aviatrix3D is about to get some volume rendering capabilities too, but not the really advanced versions, and will rely on using OpenGL programmable shaders rather than fixed functionality, so will require some very modern hardware to run on.

#4

That must be an interesting project. You may probably want to check out the work done at Uni Stuttgart and Erlangen. Sometime back I worked on unstructured volume rendering with Java3D and using volume slicing - requires only alpha blending support but generates wagonload of polys. There are possibly more hardware oriented algorithms now with the advent of shaders.

#5

Not really. The shader stuff is still using view-aligned polys slicing through the 3D texture with alpha blending after that. Where the interesting stuff comes in is in the use of shaders for more interesting rendering effects with the transfer functions and segmented data sets. There’s also some new techniques floating around dealing with preloading the z-buffer to make use of early z-culling. Our intention is to only initially go for simple 1D and 2D transfer functions and then get paid projects to do the more interesting stuff. We have some interested parties sniffing around now, so we need to get the proof-of-concept code running first.