Hi Eric,
[EMAIL PROTECTED] wrote:
> Hello all,
> I've been working with the SplitGraphOp and have a few questions on
> it's intended use and expected behavior.
>
> Questions:
> 1) does RenderAction::setVolumeDrawing(true) ignore the bounding volume
> of geometries, if so, how do I visualize the bounding volumes of an
> object after it has run through SplitGraphOp.
Hm, it should draw them.
> 2) are there known limitations/problems/errors with using SplitGraphOp
> on very large models? With OBJ files? VRML files? I get different
> results depending on whether the model was loaded as VRML or OBJ (at
> least in terms of what setVolumeDrawing() shows).
It does eat up quite a bit of memory, but it should be repeatable. Concerning
OBJ vs VRML: the scene graphs of OBJ and VRML files looks different
(MaterialGroups etc), so I wouldn't be surprised to see some differences.
> 3) when I turn on volume drawing I get a brown/yellow box moving around
> my scene wildly and seems to be attached to the camera but not identical
> to it. what does this color represent? It looks like some form of
> frustum. (see attached)
That's supposed to be the frustum, yes, but apparently it's not rendered in the
right position. Not sure what's wrong there, but I don't want to spend time
fixing it right now (too much going on).
> Background:
>
> I have several large models from Pascall Mueller's site
> (http://www.vision.ee.ethz.ch/~pmueller/wiki/CityEngine/DownloadModelsPompeii).
>
> I would like to break up these models using SplitGraphOp to make use of
> frustum and occlusion culling. when I run SplitGraphOp on them, the
> number of occlusion tests never go above 1, and the number of nodes
> drawn is a little higher (7nodes vs. 28nodes). Sadly, the # of nodes
> doesn't represent my "max_polygon" value nor does the # of nodes drawn
> change during runtime as I add/remove things from view.
>
> Using the OpenSG Traversal example, it states that the # of nodes
> before/after is the same for this model. So, I tried a simpler model, a
> tessellated sphere loaded from file, and received slightly different
> results. It ocassionally broke up the object into many parts, but the
> drawn volumes didn't update. I still had only one bounding box drawn.
>
> How have others approached such a scenario? Any insight would be greatly
> appreciated.
That's what the SplitGraphOp is meant for, but sadly it has not gotten a lot of
exercise since it's initial writing many years ago, so I would not be too
surprised if it didn't quite work the way it's expected. I tried it on the
Pompeii models and I found the problem there: the SGO didn't do anything if the
root node was a leaf. I fixed that (committed).
Hope it helps
Dirk
-------------------------------------------------------------------------
This SF.net email is sponsored by DB2 Express
Download DB2 Express C - the FREE version of DB2 express and take
control of your XML. No limits. Just data. Click to get it now.
http://sourceforge.net/powerbar/db2/
_______________________________________________
Opensg-users mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/opensg-users
