Tomas,
Thanks to some debugging from Ghansham Sangar (SAC/ISRO) and plots
from Chris Pudney, this indeed looks like a precision problem manifested
in the surface normal generation. The double-precision Set classes won't
help with this. To help verifiy this is the case, you might try
scaling up the spatial dimensions of your data while we consider some
possible solutions.
Tom
On 5/22/12 9:27 PM, Pudney Chris (ext) GBJH wrote:
G'day,
Here's another example of the issue Tomas is talking about (from the
MZmine software).
The artefacts appear when the sampling resolution along the m/z axis
is high (attachment 1).
They're not present at lower resolutions (attachment 2).
Floating point precision could well be the issue.
Regards,
Chris.
On 22/05/12 08:11, Tomas Pluskal wrote:
Hi Tom,
Thank you for the tip.
I would like to try that, but what class should I use instead of the
Linear2DSet?
There seems to be only Gridded2DDoubleSet, but no equivalent for the
linear set.
By the way, please check the other batch of screenshots I sent:
http://www.unidata.ucar.edu/mailing_lists/archives/visad/2012/msg00057.html
There the same issue appears although the spacing between the data
points is not so close (around 1).
Tomas
On May 21, 2012, at 10:41 PM, Tom Rink wrote:
Hi Tomas,
With the Linear2DSet you've defined, the points are pretty close, maybe
too close for single precision. You could try the double-precision
version
of the Set classes.
Tom
===============================================
Tomás( Pluskal
G0 Cell Unit, Okinawa Institute of Science and Technology Graduate
University
1919-1 Tancha, Onna-son, Okinawa 904-0495, Japan
TEL: +81-98-966-8684
Fax: +81-98-966-2890
This message may contain confidential information. If you are not the
designated recipient, please notify the sender immediately, and delete
the original and any copies. Any use of the message by you is prohibited.
_______________________________________________
visad mailing list
visad@xxxxxxxxxxxxxxxx
For list information, to unsubscribe, visit:
http://www.unidata.ucar.edu/mailing_lists/
