From hendrik.hoeth at cern.ch  Sat Oct 11 18:16:57 2008
From: hendrik.hoeth at cern.ch (Hendrik Hoeth)
Date: Sat, 11 Oct 2008 18:16:57 +0200
Subject: [Kflog-devel] New terrain files for KFLog/Cumulus
Message-ID: <20081011161657.GG2866@mail.physik.uni-wuppertal.de>

Hi,

after doing a bit of code-assisted reverse engineering I've now managed
to generate new terrain files for KFLog and cumulus, using void filled
data from the Shuttle Radar Topography Mission.

At full resolution of 3 arc seconds (as compared to 30 arc seconds of
our current data base) there is a great amount of detail, but the files
are large and processing time is an issue. So I downsampled the SRTM
data to 9 and 15 arc seconds.

You can find a comparison, including some screenshots, at
http://www.linta.de/~hoeth/kflog/srtm/

The web page also has links to terrain files for Germany (46-54N, 6-14E)
at different resolutions, so you can put them into your mapdata
directory and compare them to the original files. There are many places
where the difference in detail is much more apparent than in the
screenshots on my webpage -- whole river valleys disappear in GTOPO30 --
but I selected the screenshot area to also show the tile edge artefacts.

I would be glad if someone with a Nokia device could give me feedback
about the speed when these terrain files are used in cumulus.

What do you think about this effort? Which resolution do you prefer?
I'm willing to generate new terrain and ground files for the whole
earth, but I'd like to hear your opinion first.

Cheers,

    Hendrik

-- 
Those who give up essential liberties for temporary safety
deserve neither liberty nor safety.   -- Benjamin Franklin