On 09/18/2014 11:03 PM, Jukka Rahkonen wrote:
Peter Hopfgartner <peter.hopfgartner <at> r3-gis.com> writes:
Salut Evan,
thanks for you recipe. Unfortunately, what I am really interested in is to
use this from MapServer. Could
you give me a pointer in the source where the resampling takes place?
Hi,
Hi Jukka
Why don't you test with Mapserver then before starting to code?
You are certainly right that starting with MapServer is a more
reasonable approach. Anyway, sometimes a look at the code might help to
better understand the problem. (black box vs. white box)
If you use
Mapserver as WMS GDAL can request correct resolution directly and there
should be no need at all for resampling.
This is true for WMS layers. Unfortunately, with tiles this is
different. I guess (and I think to have seen this in the code), that
GDAL is looking for the best matching resolution for this particular
tile server, fetches the necessary tiles and then resamples the tiles to
the final pixmap, which, if not by incredible chance, has a lower
resolution then the cpomplete set of fetched tiles. My guess is, that in
this resampling the labels get unreadable to to nearest point
interpolation, which is really fast but delivers a lower quality then
e.g. bilinear interpolation.
Most Mapservers are also configured
to use png as one of the outputformats which should lead to better quality
than the jpeg compressed tiles you have been playing with.
Even here the, I have to follow the rules set by the service provider,
as specified in http://www.basemap.at/wmts/1.0.0/WMTSCapabilities.xml ,
which limits me to JPEG. I have to take what I get.
-Jukka Rahkonen-
Regards,
Peter
--
Peter Hopfgartner
R3 GIS
web : www.r3-gis.com
_______________________________________________
gdal-dev mailing list
gdal-dev@lists.osgeo.org
http://lists.osgeo.org/mailman/listinfo/gdal-dev
