[vtkusers] The best way to represent a multi-dimensional dataset?

[Jesse Neri]

> -----Original Message-----
> From: vtkusers-admin at public.kitware.com
> [mailto:vtkusers-admin at public.kitware.com]On Behalf Of Maurice van de
> Rijzen
> Sent: Wednesday, September 04, 2002 8:24 AM
> To: vtkusers
> Subject: [vtkusers] The best way to represent a multi-dimensional
> dataset?
> 
> 
> Dear all,
> 
> I want to visualize a 5D-seismic dataset which is on a 
> uniform sampled grid.
> In order to do this I want  to create a 3D subset which I visualize 
> using slices.
> To import the dataset I want to use vtkImageReader2.
> But vtkImageReader2 can hold 2D or 3D data.
> Can someone tell me how I can import a 5D dataset.
> Can someone also give me a hint on  how to use the 5D dataset 
> to obtain 
> a 3D subset( By taking for 2 dimensions a constan, user-given value).
> 
> Thanks,
> Maurice
>  
This mya not be the best or only way, but here is how I do it:

Read all the data into a vtkFieldArray.  The preferred method to do 
this is is with a vtkProgrammableDataObjectSource.

Then, construct the structured grid data (spatial coordinates only, 
not data values!) using the vtkDataObjectToDataSetFilter.

Then, I choose which data is mapped to the coordinates, using the
vtkFieldDataToAttributeDataFilter. I use repeated calls to this
to update the data values, and reuse the spatial cooridates.
It cna be used for scalar and vecotr data.  

A useful example of this process is the FinancialField.tcl
example, in the DataManiuplation category.

Jess

--
Jess Neri   neri at hobbes.nrl.navy.mil   (202) 404-4365 
Pulsed Power Physics Branch, Plasma Physics Division  
Naval Research Laboratory, Washington DC   20375