[Paraview] paraview3beta: isosurface coloring and transparency

[Berk Geveci]

Depth-peeling is off in 2.6 so that parameter would have no effect.
However, at the same time, transparency is plain wrong in 2.6.
Depth-peeling addresses this problem. To see this in action, compare
how two sphere with transparency are rendered between 2.6 and 3.0.

-berk


On 5/8/07, W. Bryan Smith <bryan at ncmir.ucsd.edu> wrote:
>
>
> On 5/8/07, Berk Geveci <berk.geveci at kitware.com> wrote:
> > Ouch! That is some seriously complicated geometry!
>
> yes, i am working with biologists... complex is the name of the game!
>
> > I think you are
> > running into the limitations of the depth-peeling algorithm that is
> > used to render transparent geometry. You can try to increase some of
> > these limits by recompiling ParaView. In the source code, you have to
> > edit VTK/Rendering/vtkRenderer.cxx and change MaximumNumberOfPeels on
> > line 119 to something higher, maybe 10-20. However, this will reduce
> > rendering speed because each peel roughly corresponds to 1 render.
>
>
>
> i am not sure this is the problem... when i look at the source for
> vtkRenderer.cxx in ParaView2.6 and ParaView3, these fields are
> exactly the same:
>
>  this->MaximumNumberOfPeels=4;
>
>  and the exact same dataset on the exact same machine appears
>  as expected in ParaView2.6 (i.e. without these raster artifacts).  or
> is there something i am missing here?  i looked at the diff between
> the two files, and there is a bunch of different stuff (version number
> is 1.231 versus 1.233) including a bunch of calls where, in 1.231
> there are references to "Render translucent geometry" that in ver
> 1.233 it says "Render translucent polygonal geometry".  these sorts
> of things appear to make up a majority of the diff b/t the versions.  i
> can post the diff if it will be useful (though i am sure everyone here
> can easily generate it themselves).
>
> thanks again,
> bryan
>
>


-- 
 Berk Geveci
 Kitware Inc.
 28 Corporate Drive
 Clifton Park, NY, 12065