[Paraview] Opacity Issues

[David Lonie]

On Fri, Jun 24, 2016 at 2:15 PM, David Lonie <david.lonie at kitware.com>
wrote:

> On Fri, Jun 24, 2016 at 1:10 PM, Paluszek, Lukasz <
> lukasz.paluszek at airbus.com> wrote:
>
>> It doesn’t happen all the time, I suppose you need a fairly large surface
>> dataset. I managed to reproduce the problem on a wavelet source (attached
>> state file), try increasing the number of elements if you do not see any
>> artifacts. I have bult Paraview from source on OpenSuse 13.1 but also
>> cross-checked with the Linux binaries from Paraview.org and had same
>> issues. There was no problem with the osmesa-llvm flavor though.
>>
>
> I couldn't reproduce this here on arch linux with a newer nVidia driver
> (367.27), so I'm suspicious that this is a driver bug that's affecting the
> new depth peeling implementation.
>
> Since you're building from source, can you try editing
> VTK/Rendering/OpenGL2/vtkOpenGLRenderer.cxx, and change line 322
>
> if (dualDepthPeelingSupported)
>
> to
>
> if (false)
>
> in DeviceRenderTranslucentPolygonalGeometry, recompile using the OpenGL2
> backend, and let me know if the issue goes away? I suspect it will since
> both the OpenGL1 backend and OpenGL2 mesa/llvm do not use the new code.
>
> I'll put together a patch to force the older peeling implementation via an
> environment variable to make the work-around easier to apply.
>

Patch is here: https://gitlab.kitware.com/vtk/vtk/merge_requests/1589

By the way, are you running paraview with remote rendering via pvserver? I
was just made aware of a similar issue with remote rendering and depth
peeling.

Dave
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://public.kitware.com/pipermail/paraview/attachments/20160624/cfe6efeb/attachment.html>