[vtkusers] nightly build (8.1.0?) where did vtkVolumeTextureMapper3D go

[gohlinchia]

Hi Andras & Aashish,

Thanks for explaining the issue. I have a better understanding of the
problems you guys are facing now. Thanks also for fixing the two previous
issues I mentioned. I downloaded the latest master from GitHub and verified
that the bugs are fixed.

Speaking of problems, I hope I can tag on to this thread to sound out two of
the problems we are facing now with the GPU ray cast volume renderer:

1. The first is the incorrect stereo rendering as mentioned before. The two
viewpoints generated are
the same except that they are slightly offset to give the illusion of
stereoscopic rendering. This same problem also appears in ParaView 5.4.1.
This problem only seem to affect ray cast volume rendering and not the
polygonal meshes. The horizontal left-right split viewports shows this very
clearly:

<http://vtk.1045678.n5.nabble.com/file/t342340/Split_Viewports.png> 

We suspect that while rendering the ray cast volumes, the same
viewpoint/perspective matrix was used in both the left and right views.


2. The second problem we encounter is that ray cast volume renderer doesn't
seem to work well when volume depth peeling is activated. I can't explain it
very well, but it seems to us that the depth peeling of ray cast volumes
doesn't seem to be depth correct. The rendering looks better than it is
without depth peeling but at certain angles, it feels as if the depth
peeling has no effect at all. I have attached the following images to help
explain what we are seeing:

<http://vtk.1045678.n5.nabble.com/file/t342340/Without_Depth_Peeling_0.png> 
This image shows two volumes (MRI is grey and CT is yellowish) rendered
without depth peeling. The CT is almost fully occluded.

<http://vtk.1045678.n5.nabble.com/file/t342340/With_Depth_Peeling_1.png> 
The next image shows the same two volumes rendered with depth peeling
activated. As you can see, the CT (yellowish) is smaller than the MRI (grey)
and the CT is almost fully enclosed within the MRI.

The following series of images shows the rendering effect with depth peeling
activated, as viewed from different angles. The CT should be
partially/mostly occluded by the MRI because it is a physically smaller
volume that is located/enclosed within the MRI itself. But as the series of
images shows, it looks as if the CT is floating outside of the MRI or that
the depth peeling is deactivated (it is activated, we checked).

<http://vtk.1045678.n5.nabble.com/file/t342340/With_Depth_Peeling_2.png> 
<http://vtk.1045678.n5.nabble.com/file/t342340/With_Depth_Peeling_3.png> 
<http://vtk.1045678.n5.nabble.com/file/t342340/With_Depth_Peeling_4.png> 
<http://vtk.1045678.n5.nabble.com/file/t342340/With_Depth_Peeling_5.png> 
<http://vtk.1045678.n5.nabble.com/file/t342340/With_Depth_Peeling_6.png> 
<http://vtk.1045678.n5.nabble.com/file/t342340/With_Depth_Peeling_7.png> 
<http://vtk.1045678.n5.nabble.com/file/t342340/With_Depth_Peeling_8.png> 
<http://vtk.1045678.n5.nabble.com/file/t342340/With_Depth_Peeling_9.png> 
<http://vtk.1045678.n5.nabble.com/file/t342340/With_Depth_Peeling_10.png> 
<http://vtk.1045678.n5.nabble.com/file/t342340/With_Depth_Peeling_11.png> 

We wonder what could be the problem.

Thank you so much!



-----
Chris
--
Sent from: http://vtk.1045678.n5.nabble.com/VTK-Users-f1224199.html