Of course you definitely should take your time ;) As we were waiting for ages for DEM, it doesn't matter :D It seems that 4008 compared to 3997 has the issue that several tiles are not compiled and the not enough room exception is written in the logs. I remember you mentioned it before, but couldn't find it. Im usually using 1,200,000 for splitter, reduced now to 1,000,000 but still have problems. Please find the tiles as osm.pbf file. The areas tagged as water are not compiled with 4008, the meadow-tile is fine. It seems the limit is not depending on real area size, as the measdow tile is much larger as several of the water-tiles. But there are not so many hills in, compared to the smaller ones. As said before, 3997 is all fine. The additional time looks pretty good already. I attached you also my mkgmap logs. If I should log it more detailed or send you some more data, please let me know. Henning On 21.12.2017 16:51, Gerd Petermann wrote:
Hi Henning,
yes, that matches my findings. So far I found a small difference in the encoding of plateau lenghts but even when I changed my code to match that of BuildDEMFile I did not see an improvement. I am now trying to find out why BuildDEMFile sometimes interpolates a different height value. The values differ by a few cm and thus sometimes mkgmap rounds down while BuildDEMFile rounds up (or vice versa). I am pretty sure that this is not a problem but it makes it nearly impossible to compare the resulting bit streams. Just a matter of time ...
Gerd
________________________________________ Von: mkgmap-dev <mkgmap-dev-bounces@lists.mkgmap.org.uk> im Auftrag von Henning Scholland <osm@hscholland.de> Gesendet: Donnerstag, 21. Dezember 2017 00:59:38 An: mkgmap-dev@lists.mkgmap.org.uk Betreff: Re: [mkgmap-dev] r4006: 1st alpha version to write DEM data
Hi Gerd, just a brief feedback: Overall it seems to be fine. In MapSource and Basecamp most of the DEM-Data is displayed. But there a many squares completely empty (I guess not read at all by Garmin). Of these tiles a lot are along the coast and nearby lake-area. Maybe this is related to large areas with no changing elevation. But also there are some coast-tiles working well. If you need additional information, please let me know.
Bye Henning
On 21.12.2017 00:36, Gerd Petermann wrote:
Hi Henning,
please try with r4008 which implements that and also interpolation of hgt values. There seems to be a major bug somewhere in the encoder, sometimes MapSource displays only parts of the map and relief looks wrong. I did not yet find out where exactly the problem is. I tried to compare the results from mkgmp with those from BuildDEMFile. It seems that mkgmap sometimes calculates different positions because of rounding differences. I tried to do most calculations with integers while BuildDEMFile uses doubles. Will try again tomorrow.
Gerd
________________________________________ Von: mkgmap-dev <mkgmap-dev-bounces@lists.mkgmap.org.uk> im Auftrag von Henning Scholland <osm@hscholland.de> Gesendet: Mittwoch, 20. Dezember 2017 17:06:25 An: mkgmap-dev@lists.mkgmap.org.uk Betreff: Re: [mkgmap-dev] r4006: 1st alpha version to write DEM data
Hi Gerd, it's a great news. I wanted to try it on my map of China. But unfortunately failed because of missing files in the sea area between Korea and Japan. Maybe it's a good 1st approach to interpret a missing file as an area with 0 elevation.
Anyway, thanks for all the research work on DEM!!!
Bye, Henning _______________________________________________ mkgmap-dev mailing list mkgmap-dev@lists.mkgmap.org.uk http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev _______________________________________________ mkgmap-dev mailing list mkgmap-dev@lists.mkgmap.org.uk http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev
_______________________________________________ mkgmap-dev mailing list mkgmap-dev@lists.mkgmap.org.uk http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev _______________________________________________ mkgmap-dev mailing list mkgmap-dev@lists.mkgmap.org.uk http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev