[Paraview] [EXTERNAL] fft of selection over time filter

[Moreland, Kenneth]

Alan,

I assume you are talking about the FFT of Selection Over Time filter. I'm not sure I understand all of your questions but I'll try to answer them.


  *   I got the filter to work on point data. Two caveats though. First, I think the data needs to be of a scalar floating point type. It doesn't work on vectors and it doesn't seem to work on ids. You also need to make a point selection (vs. a cell selection). I also noticed some funniness with how the selection got populated in the filter. For some reason when I create the filter it does not seem to populate the active selection correctly. I have to hit the "Copy Active Selection" button for it to get my selection properly.
  *   [Shrug] Most filter's don't automatically create magnitudes of vectors. I guess if auto convert feature is on maybe it should, but that could also be expensive since the filter takes the FFT of all fields, not just one selected field.
  *   You mean pick some range of time to apply the FFT over? That should have more to do with clipping the data values than picking some number of values to apply the FFT to. Is there a use case for that feature?
  *   I recall implementing the filter a while ago. It's a little flaky, but it seems to work on the can dataset.
  *   Open can.ex2. Select a cell at the bottom of the can. Add the FFT of Selection Over Time filter. Click "Copy Active Selection". Apply. You get a plot that is in kind of a U shape.

-Ken

From: <Scott>, Walter Scott <wascott at sandia.gov<mailto:wascott at sandia.gov>>
Date: Friday, September 12, 2014 6:52 PM
To: "paraview at paraview.org<mailto:paraview at paraview.org>" <paraview at paraview.org<mailto:paraview at paraview.org>>
Subject: [EXTERNAL] [Paraview] fft of selection over time filter

I am trying to figure out the FFT of selection over time filter.  My user is trying to use it on a velocity point vector field (in 2d, no less).  So, a few questions:

  *   This filter doesn't work on point data, but only cell data.  Shouldn't it work on point data?
  *   This filter doesn't create magnitudes of vectors.  Shouldn't it if needed?
  *   Since an FFT spans some array of data, shouldn't this span be changeable?  For instance, 16 or 32 or 128?
  *   Does this filter even work correctly?  Has anyone ever used it?
  *   Can anyone create a simple example using this filter?


Thanks,


Alan



-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://public.kitware.com/pipermail/paraview/attachments/20140915/e36fdb33/attachment-0001.html>