[vtkusers] vtkOpenGLHAVSVolumeMapper a problem or not ?

Moreland, Kenneth kmorel at sandia.gov
Wed Jul 18 18:27:31 EDT 2007


Gourbish,

In general, you do NOT want to convert your structured points to an
unstructured grid.  The structured arrangement provides implicit
connectivity that many algorithms, including volume rendering, can
exploit to provide much more efficient processing.  You will lose that
if you convert to an unstructured grid.

So, in the case of volume rendering, you will be much better off keeping
your data in a vtkStructuredPoints class and using one of the subclasses
of vtkVolumeMapper to do the volume rendering.

-Ken

> -----Original Message-----
> From: Alfredo Gourbish [mailto:the.answer3 at libertysurf.fr]
> Sent: Wednesday, July 18, 2007 1:05 PM
> To: Moreland, Kenneth; vtkusers
> Subject: Re: [vtkusers] vtkOpenGLHAVSVolumeMapper a problem or not ?
> 
> Okay, thanks you Ken for these explanations.  It's true that I have a
> structured data too. But now, I have another question : how can read
Dicom
> images in order to write an unstructured grid ? For the moment, I read
> these
> images with ITK and write a VTK volume which is a "structuredpoints"
data.
> 
> Gourbish
> 
> ----- Original Message -----
> From: "Moreland, Kenneth" <kmorel at sandia.gov>
> To: <the.answer3 at libertysurf.fr>; "vtkusers" <vtkusers at vtk.org>
> Sent: Wednesday, July 18, 2007 8:41 PM
> Subject: RE: [vtkusers] vtkOpenGLHAVSVolumeMapper a problem or not ?
> 
> 
> The HAVS volume mapper was really designed for unstructured grids.  I
> believe the test, for convenience, converts a structured grid to an
> unstructured grid.  But this conversion causes the number of cells to
grow
> significantly and virtually guarantees that the rendering rate per
cell
> will
> be slower.  Neither of things matter for a regression test, but they
make
> a
> big deal in practice.
> 
> In practice, you would really only use the HAVS volume mapper in
> situations
> where you start with an unstructured grid.  In this case, the volume
ray
> cast mapper is not really an option.
> 
> -Ken
> 
> > -----Original Message-----
> > From: vtkusers-bounces+kmorel=sandia.gov at vtk.org [mailto:vtkusers-
> > bounces+kmorel=sandia.gov at vtk.org] On Behalf Of
> the.answer3 at libertysurf.fr
> > Sent: Wednesday, July 18, 2007 6:46 AM
> > To: vtkusers
> > Subject: [vtkusers] vtkOpenGLHAVSVolumeMapper a problem or not ?
> >
> > Hi all,
> >
> > I tried to use the vtkOpenGLHAVSVolumeMapper to render a big vtk
volume
> > (14 081 Ko). I watched the test TestHAVSVolumeMapper.cxx to see how
this
> > mapper works. So I used the same filters, etc... in my program than
in
> the
> > test.
> > Finally, I have bad results with this mapper since the result is
very
> > slower and less detailled than when I use vtkVolumeRayCastMapper. I
can
> > understand that the RayCast is more detailled but not faster than
the
> > HAVS.
> >
> > Here is the config of my working PC :
> > Intel Core 2 6600 - 2Go RAM - NVIDIA GeForce 8800 GTS
> >
> > That's why I wanted to use the GPU in order to accelerate the
render...
> >
> > Does somebody have an idea on why I have so bad results ?
> >
> > Gourbish
> > _______________________________________________
> > This is the private VTK discussion list.
> > Please keep messages on-topic. Check the FAQ at:
> > http://www.vtk.org/Wiki/VTK_FAQ
> > Follow this link to subscribe/unsubscribe:
> > http://www.vtk.org/mailman/listinfo/vtkusers
> 
> 
> 
> 





More information about the vtkusers mailing list