NOTICE: This version of the NSF Unidata web site (archive.unidata.ucar.edu) is no longer being updated.
Current content can be found at unidata.ucar.edu.
To learn about what's going on, see About the Archive Site.
NOTICE: This version of the NSF Unidata web site (archive.unidata.ucar.edu) is no longer being updated.
Current content can be found at unidata.ucar.edu.
To learn about what's going on, see About the Archive Site.
Hi Ian, We will need a test case for debugging this. But first a couple questions: 1. Are you generating your own topology passed to DelaunayCustom, or are you letting the system compute its own? 2. When you offset X and y values, do you do that before you or the system computes the topology, or do you compute the topology and then offset the x and y values? In any case, we will need a program, any necessary data files, and instructions for compiling and running the program. Cheers, Bill ---------------------------------------------------------- Bill Hibbard, SSEC, 1225 W. Dayton St., Madison, WI 53706 hibbard@xxxxxxxxxxxxxxxxx 608-263-4427 fax: 608-263-6738 http://www.ssec.wisc.edu/~billh/vis.html On Mon, 28 Apr 2003, Ian Graham wrote: > I'm trying to resample a field for which the domain is an x/y tuple defined > by an Irregular2DSet, and the range is a simple RealType. Basically I want > to make an interpolated image out of a scatter plot, so I'm resampling my > field using the points of a Linear2DSet. > > Any simple examples I create seem to work fine, so I'm unable to give a code > example that demonstrates the problem, but my symptom is that large portions > of my resampled field are filled with NaN. If I offset my x values so they > are all positive, then missing horizontal strips are resampled properly. If > I offset my y values so they are all positive, then the missing top > portion(which is sometimes everything) of the data gets displayed properly. > > I find the resampling code pretty intimidating, but I'm hoping a developer > can look through it and pick out the problem. If not, then I'll have to > send my whole set of data(2900 points), or the Irregular2DSet created by > them. I'm pretty sure the triangulation is OK, because a plot of the field > whose domain is the irregular set itself looks reasonable, and I can easily > evaluate the field for any pair of coordinates OK. > > I've even tried to create an equally complex irregular set from random > points, and I still can't recreate the problem except with my specific set > of points. > > Anyway, I put some quick screen-dumps illustrating the problems at > http://gritstone.ca/personal/igraham/VisADProblems/VisADResamplingProblem.ht > ml > > Thanks for any help resolving this! > > Ian > >
visad archives: