[mkgmap-dev] DEM Resolution and size savings

Sat Mar 31 12:28:00 BST 2018

Observation based on Austria:
Well if I create a map with DEM layer with
resolution=0=24,1=23,2=22,3=21,4=20,5=19,6=18 and set dem-dists=6624  the
resulting quality will be pretty low.
If I create that map instead with resolution=0=23,1=22,2:21.... and
dem-dists=3312 - the filesize will be similar or actually smaller, the
level of detail/quality of the DEM however much higher. Compared to the
much bigger (in MB/filesize) option of  0=24,1=23,2=22... and
dem-dists=3312 there is virtually no visual difference in Basecamp.
Even with dem-dists=1656 there is no visual difference if you create the
DEM layer starting with resolution 0=24 or 0=23. For dem-dists=3312 there
is really no reason to go higher than resolution 23. Even resolution 22
would be fine. For dem-dists=1656 resolution 0=23 is good enough.

So actually if mkgmap could somehow make use of this to optimize  the
quality/size ratio of DEM layer that would be pretty good.
Even though dem-dists=1652 looks pretty neat in Basecamp - I'm not sure if
it is not sometimes creating detail that is not there. For sure if you
create a route and look at the altitude profile the overall climb/descent
will be overstated - but that's of course also due to ways usually
following more the possibility of lowest climb/descent vs shortest
distance. In general the resolution of both OSM and DEM I guess is then not
good enough and climb/descent will be overestimated a bit.

For Viewfinderpanormas 1" DEM files - the DEM produced with resolution 0=23
and dem-dists=3312 seems to be a good compromise if size is not a big
factor. If size is a factor resolution 0=22 and dem-dists=3312 will be the
optimum. For best visual quality resolution 0=23 and dem-dists=1656 will be
best (though I'm not sure if we go for fake accuracy here. Resolution 0=24
and dem-dists=1562 is really not worth it. It maybe however that the 1"DEM
is not up to actually improving quality for dem-dists=1656 in the Alps so
best quality default would actually be resolution 0=23 dem-dists=3312 and
best quality/size 0=22 and dem-dists=3312.

On 31 March 2018 at 13:00, Gerd Petermann <gpetermann_muenchen at hotmail.com>
wrote:

> Hi Felix,
>
> Sorry, I don't understand how you connect resolution and DEM. Can you
> explain this more detailed?
>
> Gerd
>
> ________________________________________
> Von: mkgmap-dev <mkgmap-dev-bounces at lists.mkgmap.org.uk> im Auftrag von
> Felix Hartmann <extremecarver at gmail.com>
> Gesendet: Samstag, 31. März 2018 12:33:44
> An: Development list for mkgmap
> Betreff: Re: [mkgmap-dev] DEM Resolution and size savings
>
> Yes I know there are still improvements to be done. It was just a
> suggestion because the result is much better than saving space/data by
> decreasing the dem-dist value. Even resolution 22 as highest value is still
> pretty good - but with 22 on 3312 you start to see some very small changes
> already. Still way better than 6624 at resolution 24.
> Actually with resolution 22 it just looks a little bit flatter but level
> of detail still seems to be the same (similar to decreasing the elevation
> exageration by 20% in Basecamp). Only at resolution 21 you really start to
> miss detail (in general it seems to me that the DEM detail is not that good
> in Basecamp - but that also applies to original garmin maps).
>
> Maybe to save size (because right now DEM at resolution 24 can get quite
> huge) - there could be an option to have the DEM always saved like this -
> so same as 3312 on resolution 22 but at 24....
>
> On 31 March 2018 at 08:40, Gerd Petermann <gpetermann_muenchen at hotmail.com
> <mailto:gpetermann_muenchen at hotmail.com>> wrote:
> Hi Felix,
>
> yes, the DEM format is not yet fully understood. I assume what you have is
> a map that uses a shrink factor <> 1.
> The shrink factor is used like this:
> The height deltas are devided by this value before encoding and multiplied
> when extracting. The effect is that the deltas
> are smaller and therefore the size is also smaller, but of course you also
> lose a bit of information, because only the integer
> part is stored.
> The problem is that Garmin also uses slightly different rules for the
> encoder, and we did not yet find out all details.
> Frank Stinners program BuildDEMFile allows to use this but sometimes
> produces invalid data.
> The tool DemDisplay shows my current knowledge.
>
> Gerd
>
> ________________________________________
> Von: mkgmap-dev <mkgmap-dev-bounces at lists.mkgmap.org.uk<mailto:mkgmap-
> dev-bounces at lists.mkgmap.org.uk>> im Auftrag von Felix Hartmann <
> extremecarver at gmail.com<mailto:extremecarver at gmail.com>>
> Gesendet: Freitag, 30. März 2018 23:07:10
> An: Development list for mkgmap
> Betreff: [mkgmap-dev] DEM Resolution and size savings
>
> Hi everyone,
>
> I noticed that the DEM layer if created for resolution 23 only (with a map
> that has not 24 resolution) will only be half the size of the DEM in
> resolution 24 (dem-dist=3312) - however in Basecamp/Mapsource the detail is
> virtually identical - I cannot see any difference in quality.
>
>
> So I think there must be some way to still save a lot of data/space - but
> it's not by going for dem-dits=6624 - that will result in much worse DEM
> detail.
>
> (I still really haven't found a good solution for DEM on GPS devices
> though. Need more time trying out different values and possibilities. Right
> now I think best is probably a separate transparent but except for DEM
> empty DEM only gmapsupp.img).
>
> --
> Felix Hartman - Openmtbmap.org & VeloMap.org
> Schusterbergweg 32/8
> 6020 Innsbruck
> Austria - Österreich
> _______________________________________________
> mkgmap-dev mailing list
> mkgmap-dev at lists.mkgmap.org.uk<mailto:mkgmap-dev at lists.mkgmap.org.uk>
> http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev
>
>
>
> --
> Felix Hartman - Openmtbmap.org & VeloMap.org
> Schusterbergweg 32/8
> 6020 Innsbruck
> Austria - Österreich
> _______________________________________________
> mkgmap-dev mailing list
> mkgmap-dev at lists.mkgmap.org.uk
> http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev
>

-- 
Felix Hartman - Openmtbmap.org & VeloMap.org
Schusterbergweg 32/8
6020 Innsbruck
Austria - Österreich
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.mkgmap.org.uk/pipermail/mkgmap-dev/attachments/20180331/d8584d62/attachment.html>