Re: [mkgmap-dev] [PATCH] patch (V8) for simplifying ways and reducing img size
Patch V8 updated to r1404. - do not merge lines at resolution 24, as the dp filter is disabled. With this change there should be no longer routing problems. ------------------------------------------------------------------- Patch V7 is now based on the R1041 (and the SizeFilter.patch) - removing of the -- suppress-dead-end-nodes option, as it causes problems with routing. - code cleanup, introduction of new classes. ------------------------------------------------------------------- Patch V4 now supports three options. Its build on top of r972. With all three options used, it will reduce the four bavaria tiles from 58MB to 46MB. Drawing at low zoom levels gets speed up significantly at my etrex. This will not be the last version of the patch. I'm working further on it, but have no idea, how fast work will progress. The options are as follows: +--reduce-point-density=num + Simplifies the ways with the douglas peucker algorithm. + num is the maximal allowed error distance, by which the resulting + way may differ from the original one. + This distance gets shifted with lower zoom levels. + Recommended setting is 10, this should lead to only small differences + (Default is 2.6, which should lead to invisible changes) + +-- suppress-dead-end-nodes + Do not generate nodes at the end of an dead end way. + Decrease file size and should speed up routing slightly. + May possibly cause errors, but not seen any one at the moment. + +--merge-lines + Try to merge lines. This helps the simplify filter to straighten out + longer chunks at lower zoom levels. Decreases file size more. + Increases paint speed at low zoom levels .
I wanted to apply it but I get "The chunk size did not match the number of added/removed lines!" Is there possibly something wrong with Patch V8? Johann Gail wrote:
Patch V8 updated to r1404. - do not merge lines at resolution 24, as the dp filter is disabled. With this change there should be no longer routing problems.
-------------------------------------------------------------------
Patch V7 is now based on the R1041 (and the SizeFilter.patch) - removing of the -- suppress-dead-end-nodes option, as it causes problems with routing. - code cleanup, introduction of new classes.
-------------------------------------------------------------------
Patch V4 now supports three options. Its build on top of r972. With all three options used, it will reduce the four bavaria tiles from 58MB to 46MB. Drawing at low zoom levels gets speed up significantly at my etrex.
This will not be the last version of the patch. I'm working further on it, but have no idea, how fast work will progress.
The options are as follows:
+--reduce-point-density=num + Simplifies the ways with the douglas peucker algorithm. + num is the maximal allowed error distance, by which the resulting + way may differ from the original one. + This distance gets shifted with lower zoom levels. + Recommended setting is 10, this should lead to only small differences + (Default is 2.6, which should lead to invisible changes) + +-- suppress-dead-end-nodes + Do not generate nodes at the end of an dead end way. + Decrease file size and should speed up routing slightly. + May possibly cause errors, but not seen any one at the moment. + +--merge-lines + Try to merge lines. This helps the simplify filter to straighten out + longer chunks at lower zoom levels. Decreases file size more. + Increases paint speed at low zoom levels . ------------------------------------------------------------------------
_______________________________________________ mkgmap-dev mailing list mkgmap-dev@lists.mkgmap.org.uk http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev
Hi Felix, are you a night worker or do you live in another timezone? :-)
I wanted to apply it but I get "The chunk size did not match the number of added/removed lines!"
Is there possibly something wrong with Patch V8?
Yes, I have seen some minor changes in the patch, which should not be included, so I modified it by hand. Seems to be no good idea. Attached the corrected patch. Please note: The patch should work as far it is implemented now. But I have some more ideas for the dp filter, which would need some rework of the base classes, for which I have not the time at the moment. It should be clearly an improvement, but shows not its full potential at the moment.
participants (2)
-
Felix Hartmann -
Johann Gail