Re: Questions regarding VisAD from Brookhaven National Lab


VisAD supports linear, gridded and irregular sets, where a linear set is
an N-D grid that is evenly spaced, a gridded set is an N-D grid with a
convex (but not necessarily evenly spaced) topology, and an irregular set
is a collection of M points in N-D with no particular topology.

VisAD supports volume rendering of data with a linear set. Although it
does not do volume rendering of non-linear gridded or irregular data,
VisAD has a powerful resampling feature that allows you to resample your
data from a non-linear gridded or irregular set into a linear set, which
can then be volume rendered.

If your data is distributed in 3-D in such a way that it cannot be
ordered, use Irregular3DSet (resampling to a Linear3DSet to do volume
rendering). If your data can be ordered in 3-D somehow, consider using a
Gridded3DSet, Linear3DSet or Integer3DSet, since they are more efficient
(no need to perform an implicit Delaunay triangulation).

To give you a headstart, I've written two quick examples. The first
(LinearRenderTest) is a volume rendering example that uses Linear3DSet.
Hopefully it is demonstrates how to do volume rendering in VisAD
straightforwardly. The second (IrregularRenderTest) is an example that
uses Irregular3DSet, resampling to a Linear3DSet to do the volume
rendering. It is very similar to the first example, with a few key
changes. You can find the examples on my "Additional VisAD examples"
web page at:

I'm CC'ing the VisAD mailing list, so that anyone else with similar
questions can find this response in the archives. I'd be happy to answer
any further questions you might have, but please CC the VisAD list,
since there are other VisAD developers who can also assist you.


At 02:46 PM 6/16/2004, you wrote:
>Hello Curtis Rueden,
>We are a research team at Brookhaven National Lab implementing visualization
>for the Atmospheric Science Division.
>We have been charged to implement the visualization of a 4D volume rendering
>of chemical species in the atmosphere, e.g. sulfur dioxide distribution in
>the northern hemisphere.  Due to real topography, the altitude is not a
>regular grid.  On top of this issue, we are also rendering the chemical
>The initial study from a year ago suggested either a Vis5D+ or VisAD
>approach.  However, after research was conducted during that time, it was
>decided that VisAD was not capable of displaying our data in an irregular
>After some current research, we are inclined to VisAD thanks to an article
>from 4DWX on "Why VisAD" for a particular project, as well as documentation
>on IrregularSets from the "VisAD Java Class Library Developers Guide".
>However, we need concrete evidence about VisAD supporting "irregular grids"
>in volume rendering.
>We would like to borrow some wisdom regarding any experience in the use of
>"irregular grids".  Any insight and/or suggestions would be greatly
>Matthew Weber