[ncl-talk] Plot looks fuzzy when transparency is used

[Kyle Griffin]

Hi all,

Curious case I've been looking at for the past few weeks, and I think I
finally have enough to share with you. The two images below are made by the
same set of resources applied (via setvalues) directly to relevant map and
field objects in NCL - the only difference is from some changes in the fill
opacities and the draw order. The image ending in 000 (the first link)
appears, to me, to be fuzzy - especially with the vector edges, especially
when compared to the 012 image (second link).

With partially opaque fills:
http://aos.wisc.edu/dynamaps/15011512/jet_pv_irrwind/jet_pv_irrwind_africa_15011512_000.jpg
Without partially opaque fills:
http://aos.wisc.edu/dynamaps/15011512/jet_pv_irrwind/jet_pv_irrwind_africa_15011512_012.jpg

Your first thought, like mine, may jump to the image type (JPEG) which is
not designed to handle transparencies. Since NCL doesn't produce JPG files,
it's obvious these have been run through 'convert', but I don't believe
that to be the source of the problem either. The curious thing, however, is
that this occurs regardless of the original source file. I've done this by
directly producing EPS and PNG files and the final result, with
transparencies, is the same.

Direct product of NCL (png):
http://aos.wisc.edu/dynamaps/15011512/jet_pv_irrwind/ncl_transp_orig.png

Is there something inherent to my use of the object-oriented applications
that degrades image quality with the use of opacity settings? I've made
similar/nearly identical images via gsn_xxxx calls with perfect success.
Any thoughts here would be appreciated. I can provide some code, but I can
assure you it's not something that will be easily abbreviated into a
typical ncl-talk example (sorry).


Kyle

----------------------------------------
Kyle S. Griffin
Department of Atmospheric and Oceanic Sciences
University of Wisconsin - Madison
Room 1421
1225 W Dayton St, Madison, WI 53706
Email: ksgriffin2 at wisc.edu
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mailman.ucar.edu/pipermail/ncl-talk/attachments/20150115/f0c8ce8e/attachment.html