-- Mark J. Stevens Climate and Global Dynamics Division NCAR Mesa Laboratory, Room 600 ph: 303-497-1707 stevens@ucar.edu http://www.cgd.ucar.edu/~stevens
attached mail follows:
Jonathan,
This is a problem I looked into before. The problem isn't in the postscript plot, it arises when viewing the postscript plot or converting it to another format. You can see this by using gv. Start gv with the postscript file. At the top of gv there is a menu button labeled "State". Click on the item labeled "Antialias" to turn it on or off. With the antialias turned on the problem will be there, with the antialias turned off the problem will go away. If you use gs I think by default the antialiasing is off so the problem doesn't appear, but there are options to turn it on or off.
When you use "convert" the antialiasing is by default used so the problem is visible. You can turn it off with the +antialias option.
The real problem is that when the antialiasing is turned off the plot area will be fine but the fonts will look really bad. Also sometimes this problem appears and other times it doesn't. When I run scripts sometimes I get this problem in one plot, but not in another.
Mark
-------------------------------------------------------------------------------------------------------------------------
On Tuesday 23 March 2004 02:57 pm, Jonathan Vigh wrote: > Greetings All, > > I am trying to make some publication-quality plots in NCL, but am > experiencing an odd strippling/streaking pattern when I output to > postscript or eps. I've reduced my plot down to a very simple > configuration (I plot a map using gsn_csm_map and resources for color > fill and boundary outlines). The strippling/streaking often appears in a > north-south oriented box, and sometimes other faint streaks appear going > from north to south. In the example plot attached, the streaking is > visible in a region to the north of Cuba. If I change the domain of the > map to a different region like India, the strippling also appears, so > perhaps the problem is related to the low-level area routines? If I just > plot over the open ocean, with no land in the domain, I think it goes > away. > > Maybe this is a bug in the NCL postscript drivers? My NCL version is > a031. (I was making similar plots previously using version a024 and I > don't think I had this problem, but I'm not sure.) > > Interestingly, the strippling isn't visible using the Ghostscript viewer > (gs), but it is visible in Ghostview (gv). It's not just a quirk of > Ghostview, because when I convert (using convert) to another format like > png, the streaks are still visible. I've attached my script (test.ncl). > The output file is too large for this list, but I can e-mail it to any > of you off-list if you can't reproduce it with the script. > > Has anyone experienced this before, and if so, do you know what the > problem is, or how to prevent this strippling? > > Thanks, > > Jonathan
-- Mark J. Stevens Climate and Global Dynamics Division NCAR Mesa Laboratory, Room 600 ph: 303-497-1707 stevens@ucar.edu http://www.cgd.ucar.edu/~stevens
_______________________________________________ ncl-talk mailing list ncl-talk@ucar.edu http://mailman.ucar.edu/mailman/listinfo/ncl-talk
This archive was generated by hypermail 2b29 : Tue Mar 23 2004 - 15:29:34 MST