[Paraview] Question about volume viz algorithm(s)

[Utkarsh Ayachit]

Comer,

If your data is structured points, it is indeed a vtkImageData and
hence you should be able to volume render it directly as well as use
vtkImageShiftScale filter.

Utkarsh

On Thu, Mar 18, 2010 at 2:10 PM, Comer Duncan <comer.duncan at gmail.com> wrote:
> Utkarsh,
>
> Thanks very much for your reply.  I appreciate it.
>
>
>> > 1. I would like to know more about the algorithm that is used in
>> > Paraview to
>> > do volume viz.  I have used Paraview to viz a cartesian grid data set
>> > with
>> > reaonable success.  I can not seem to locate in the documentation much
>> > at
>> > all about just which algorithm is being used.  And how come there is a
>> > need
>> > to stuff the data through tetrahedralize first?
>>
>> What is your data type? If it's a vtkImageData then you shouldn't need
>> to tetrahedralize. If it's vtkStructuredGrid then you'll need to
>> tetrahedralize since ParaView currently only supports volume rendering
>> for regular cartesian grids i.e image data and unstructured grids.
>
> My data is a 31x31x31 array of doubles in a 3d cartesian grid.  The vtk file
> header plus a couple of numbers is:
>
> # vtk DataFile Version 2.0
> brillset_a6_n64_dx0p15625_gxx.dat
> ASCII
> DATASET STRUCTURED_POINTS
> DIMENSIONS 31 31 31
> ORIGIN 0 0 0
> SPACING 0.412500 0.412500 0.412500
> POINT_DATA 29791
>  SCALARS gxx float
> LOOKUP_TABLE default
> 1.00006867
> 1.00006955
> ...
> ...
> ...
>
> so, it's structured. The min is .953... and the max is 1.030... so it covers
> a pretty modest range.  I have viewed this data with Paraview and thought I
> had to stuff it through tetrahedralize first...was wondering how come?
> The volume viz of it shows most of the action is in the middle of the
> volume. Have played around with diverging and other maps. Have run
> isosurface on it with success.
>
> I am just trying to understand what methods are being used in Paraview in
> order to better learn how to do better viz.
>
>
>>
>> For image data, ParaView uses fixed-point-ray-cast mapper (software)
>> or a gpu-based raycaster. For details for the two algorithms used look
>> at
>> *
>> http://www.vtk.org/doc/nightly/html/classvtkFixedPointVolumeRayCastMapper.html#_details
>>
>> For unstructured grid, we have:
>>
>> * ZSweep -
>> http://www.vtk.org/doc/nightly/html/classvtkUnstructuredGridVolumeZSweepMapper.html#_details
>> * Projected Tetrahedra  -
>>
>> http://www.vtk.org/doc/nightly/html/classvtkProjectedTetrahedraMapper.html#_details
>> * Sofware Raycaster (Bunyk) -
>>
>> http://www.vtk.org/doc/nightly/html/classvtkUnstructuredGridVolumeRayCastMapper.html#_details
>> * HAVS Mapper -
>> http://www.vtk.org/doc/nightly/html/classvtkHAVSVolumeMapper.html#_details
>>
>
> Thanks very much for the pointers. Will take a look.
>
>>
>> > 2. I have recently taken a look at the vtk examples for volume viz.
>> > They
>> > are nice.  However, I decided to take a look at modifying
>> > SimpleRayCast.py
>> > to viz my 31^3 data set i doubles but seem to find that the routine
>> > which
>> > does the heavy lifting does not take doubles.  What do I need to do
>> > first to
>> > make use of this example?
>>
>> You may want to use something like vtkImageShiftScale
>> (http://www.vtk.org/doc/nightly/html/classvtkImageShiftScale.html) to
>> transform your data to the appropriate type.
>>
>
> I guess not, as my data isn't an image.
>
>
> Thanks again for taking the time to answer my questions.
>
> Comer
>
>