Hi Gino,

This is rather a data specific issue and really not something that is
generally applicable.  If you do wish to change the way the
transparent objects in the scene are handled then it would most
appropriate to modify the StateSet attached to the prototype
"DepthSortedBin" by doing something like:

osg::StateSet* startset =
osgUtil::RenderBin::getRenderBinPrototype("DepthSortedBin")->getStateSet();
if (!stateset)
{
    stateset = new osgUtil::RenderBin;
    
osgUtil::RenderBin::getRenderBinPrototype("DepthSortedBin")->setStateSet(stateset);
 }

  stateset->setAttribute(... stuff you want to set...);

Robert.


On Tue, Jun 16, 2009 at 12:43 PM, Gino van den
Bergen<[email protected]> wrote:
> Hi Guys,
>
> I've added the disabled depth write. I've been using VRML models that are
> exported by VMD (http://www.ks.uiuc.edu/Research/vmd/). Here atoms in
> molecules may show a transparent sphere to denote the van derWaals fields.
> Without disabling the depth write, the overlapping transparent spheres would
> show popping when a molecule is rotated due to swaps in the depth sort. I
> know that strictly speaking the draw order may be violated for overlapping
> transparent spheres, but this is less disturbing than a popping depth cull
> on part of a transparent sphere. So my suggestion would be to keep the depth
> draw disabled on transparent objects. In the best case you do not see any
> difference since transparent objects are already drawn back-to-front. In the
> worst case a  non-commutative blend operation is performed in the wrong
> order.
>
> My 2 cents,
>
> Gino
>
>
>
>
> Robert Osfield wrote:
>>
>> Thanks for the explanation, makes sense now.  I've commented out this
>> line and have checked all the changes into svn/trunk.  Cheers, Robert.
>>
>> On Mon, Jun 8, 2009 at 10:31 AM, Johan
>> Nouvel<[email protected]> wrote:
>>
>>>
>>> Hello,
>>>
>>> You're right Jan. I doesn't known why this code is here, but as I doesn't
>>> want to change usual behaviour of the ReaderWriterVRML I have
>>> copied/pasted
>>> this code from the readerWriterVRML itself. It is coherent with previous
>>> behaviour.
>>>
>>> For me, we could completely remove this code, in my part (line 452) and
>>> when
>>> there is a material with transparency too (line 388).
>>>
>>> Regards,
>>> Johan.
>>>
>>>
>>> Jan Ciger a écrit :
>>>
>>>>
>>>> -----BEGIN PGP SIGNED MESSAGE-----
>>>> Hash: SHA1
>>>>
>>>> Hi Robert,
>>>>
>>>> Robert Osfield wrote:
>>>>
>>>>>
>>>>> HI Johan,
>>>>>
>>>>> I'm just doing a review of your changes relative to svn/trunk and
>>>>> concerned about the comment:
>>>>>
>>>>> // GvdB: transparent objects do not write depth
>>>>>
>>>>>
>>>>
>>>> ...
>>>>
>>>>>
>>>>> Normally transparent objects do not disable depth write.  Could you
>>>>> please explain why you've done this as it would seem like something
>>>>> that is
>>>>> very model specific not a general feature that you'd enable.
>>>>>
>>>>
>>>> I think that these changes weren't made by him - I recall seeing this
>>>> when I was porting the plugin to new OpenVRML. I think it was added by
>>>> Gino van den Bergen (gino at dtecta.com) last time. I am not sure why he
>>>> did this change neither.
>>>>
>>>> Regards,
>>>>
>>>> Jan
>>>>
>>>> -----BEGIN PGP SIGNATURE-----
>>>> Version: GnuPG v1.4.9 (GNU/Linux)
>>>> Comment: Using GnuPG with Mandriva - http://enigmail.mozdev.org
>>>>
>>>> iD8DBQFKKdFbn11XseNj94gRAobvAKCGE6CT7TRUGt762PszB6b/9scGaACgsBev
>>>> yUDt/KRaaT9bAH3pbE6mIUs=
>>>> =oMmR
>>>> -----END PGP SIGNATURE-----
>>>> _______________________________________________
>>>> osg-submissions mailing list
>>>> [email protected]
>>>>
>>>>
>>>> http://lists.openscenegraph.org/listinfo.cgi/osg-submissions-openscenegraph.org
>>>>
>>>
>>> --
>>> Johan Nouvel, R&D Engineer
>>> http://www.archivideo.com
>>> tel : +33 (0)2 99 86 30 20
>>> ARCHIVIDEO, 40 rue des Veyettes, 35000 RENNES, FRANCE
>>> _______________________________________________
>>> osg-submissions mailing list
>>> [email protected]
>>>
>>> http://lists.openscenegraph.org/listinfo.cgi/osg-submissions-openscenegraph.org
>>>
>>>
>>
>> _______________________________________________
>> osg-submissions mailing list
>> [email protected]
>>
>> http://lists.openscenegraph.org/listinfo.cgi/osg-submissions-openscenegraph.org
>>
>
> _______________________________________________
> osg-submissions mailing list
> [email protected]
> http://lists.openscenegraph.org/listinfo.cgi/osg-submissions-openscenegraph.org
>
_______________________________________________
osg-submissions mailing list
[email protected]
http://lists.openscenegraph.org/listinfo.cgi/osg-submissions-openscenegraph.org

Reply via email to