Hi Felix,

I think one reason for the zig-zagging is that we don't merge roads at lower 
resolutions and maybe DouglasPeucker should ignore "preseved" points at level 1 
and higher.

Gerd

________________________________________
Von: mkgmap-dev <[email protected]> im Auftrag von Felix 
Hartmann <[email protected]>
Gesendet: Samstag, 1. Mai 2021 11:17
An: Development list for mkgmap
Betreff: Re: [mkgmap-dev] Request - set --reduce-point-density= for each        
resolution

haha, no worries. I could have only said, I never noticed that problem. The 
main problem is the zig-zagging at lower resolution.

On Sat, 1 May 2021 at 16:54, Gerd Petermann 
<[email protected]<mailto:[email protected]>> wrote:
Hi all,

sorry for the noise. Got confused by the special case with mkgmap:fast_road=yes 
in the default style.

Gerd

________________________________________
Von: Gerd Petermann 
<[email protected]<mailto:[email protected]>>
Gesendet: Samstag, 1. Mai 2021 10:46
An: Development list for mkgmap
Betreff: AW: [mkgmap-dev] Request - set --reduce-point-density= for each        
resolution

Hi Felix,

reg. low resolutions:
I think there is also an error in mkgmap which filters e.g. large parts (almost 
all) of motorways at resolution 14.

Gerd

________________________________________
Von: mkgmap-dev 
<[email protected]<mailto:[email protected]>>
 im Auftrag von Felix Hartmann 
<[email protected]<mailto:[email protected]>>
Gesendet: Freitag, 30. April 2021 16:55
An: Development list for mkgmap
Betreff: Re: [mkgmap-dev] Request - set --reduce-point-density= for each        
resolution

sorry - patch has some typos as i was still playing around with values. here is 
a working version. These values I feel are good with 1.3 at commandline.

On Fri, 30 Apr 2021 at 22:44, Felix Hartmann 
<[email protected]<mailto:[email protected]><mailto:[email protected]<mailto:[email protected]>>>
 wrote:
I hardcoded a patch - as I was not sure someone is interested in it. This is 
doing the job - but I do feel others could profit as well. And yes I now 
understood how I could change this for a map starting at resolution 23 as well. 
I'm still not sure why do we not run the DP filter on resolution 24? I would 
image a very small value would give some improvement without visually changing 
anything.

I haven't really looked at polygons yet - but my patch looks pretty good to me 
if you use a value of 1.3. So just as recommendation for default values. I feel 
the current default of 2.6 was exactly that low - because at resolution 23 a 
high value is really bad, while at lower resolutions there is a need for 
simplification. it's a bad compromise.

On Fri, 30 Apr 2021 at 20:38, Gerd Petermann 
<[email protected]<mailto:[email protected]><mailto:[email protected]<mailto:[email protected]>>>
 wrote:
Hi Felix,

you are probably right, the DP filter was not adapted to work with a map with 
resolution 23 as highest resolution. A quick change would be to disable the 
filter at level 0, similar to a map with max resolution 24.
The --reduce-point-density sets the error distance for the Douglas-Peucker (DP) 
algorithm.
If I got you right that's not enough for a map with levels = 0:24, 1:23, ... 
because you would want to use a rather small error distance for the first two 
layers and a higher value for the rest.  Sounds reasonable to me. Probably also 
for --reduce-point-density-polygon.

I'll try to implement both.

Gerd

________________________________________
Von: mkgmap-dev 
<[email protected]<mailto:[email protected]><mailto:[email protected]<mailto:[email protected]>>>
 im Auftrag von Felix Hartmann 
<[email protected]<mailto:[email protected]><mailto:[email protected]<mailto:[email protected]>>>
Gesendet: Mittwoch, 28. April 2021 05:03
An: Development list for mkgmap
Betreff: [mkgmap-dev] Request - set --reduce-point-density= for each    
resolution

As I'm not sure everyone is following the other topic, I would like to ask to 
introduce the possibility to set different values for the DP filter based on 
resolution.

Why is this important?

a) maps which are produced with resolution 23 as highest resolution need to 
have a much lower level than subsequent resolutions.
b) For resolution 23 and 22 a lower level makes sense - it is not needed as 
much here to speed up the gps device map drawing speed, while for lower 
resolutions like 18-21 a high level is really needed else map panning becomes 
painfully slow (usually basemap kicks in at 17).

If I'm correct right now 2.6 at resolution 23 would be 5.4 at resolution 22, 
10.8 at resolution 21 and so on. It's okay to keep that behaviour - but the 
value doubled just needs to be possible to be set for each resolution. It 
doesn't matter to me if we keep it like that or if we change the numbers than 
ourselves by resolution. Important thing is that it's possible to be set for 
each resolution

c) maybe also introduce the possibility to set DP for resolution 24. I think a 
value of 0.3-0.6 could be worth it. I'm not sure if there were some problems 
preventing DP to be used at 24. However as by now I think mkgmap is able to 
produce routable maps starting at 23 instead of 24, those problems should be 
fixed...
In general I think a map made for automotive use (car, motorcycle) would be 
fine using resolution 23 as a start instead of 24. Like Garmin City Navigator 
maps do. But then yes it's really important to not have a high DP value on 
resolution 23. Same as contourlines only maps which make sense at resolution 23 
instead of 24.

--
Felix Hartman - Openmtbmap.org & VeloMap.org

_______________________________________________
mkgmap-dev mailing list
[email protected]<mailto:[email protected]><mailto:[email protected]<mailto:[email protected]>>
https://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev


--
Felix Hartman - Openmtbmap.org & VeloMap.org



--
Felix Hartman - Openmtbmap.org & VeloMap.org

_______________________________________________
mkgmap-dev mailing list
[email protected]<mailto:[email protected]>
https://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev


--
Felix Hartman - Openmtbmap.org & VeloMap.org

_______________________________________________
mkgmap-dev mailing list
[email protected]
https://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev

Reply via email to