Re: [OSM-dev] Out of copyright WMS offering - looking for assistance
Andy Robinson (blackadder-lists) wrote: Oh, and I forgot the most important issue. The plug-in is looking for png's when the ooc map tiles are all jpgs. Cheers Andy Must they be .jpgs? You'll get a lot of sharp edge artefacts using the .jpg format on a map. In theory at least (and in practice whenever I've done it) .png will give better results on material such as that. Steve ___ dev mailing list dev@openstreetmap.org http://lists.openstreetmap.org/listinfo/dev
Re: [OSM-dev] Out of copyright WMS offering - looking for assistance
Steve Hosgood wrote: Sent: 19 October 2009 9:39 AM To: dev@openstreetmap.org Cc: 'josm-dev' Subject: Re: [OSM-dev] Out of copyright WMS offering - looking for assistance Andy Robinson (blackadder-lists) wrote: Oh, and I forgot the most important issue. The plug-in is looking for png's when the ooc map tiles are all jpgs. Cheers Andy Must they be .jpgs? You'll get a lot of sharp edge artefacts using the .jpg format on a map. In theory at least (and in practice whenever I've done it) .png will give better results on material such as that. Bear in mind these are printed maps, not rasterised vector maps. We tried with png but you can get much lower file size without loosing too much definition with the jpg's. Now, if someone wants to spend the time proving me wrong they are welcome to a an original scan to try it. Cheers Andy Steve ___ dev mailing list dev@openstreetmap.org http://lists.openstreetmap.org/listinfo/dev ___ dev mailing list dev@openstreetmap.org http://lists.openstreetmap.org/listinfo/dev
Re: [OSM-dev] Out of copyright WMS offering - looking for assistance
Andy Robinson (blackadder-lists) wrote: Steve Hosgood wrote: Sent: 19 October 2009 9:39 AM To: dev@openstreetmap.org Cc: 'josm-dev' Subject: Re: [OSM-dev] Out of copyright WMS offering - looking for assistance Andy Robinson (blackadder-lists) wrote: Oh, and I forgot the most important issue. The plug-in is looking for png's when the ooc map tiles are all jpgs. Cheers Andy Must they be .jpgs? You'll get a lot of sharp edge artefacts using the .jpg format on a map. In theory at least (and in practice whenever I've done it) .png will give better results on material such as that. Bear in mind these are printed maps, not rasterised vector maps. We tried with png but you can get much lower file size without loosing too much definition with the jpg's. That's OK then. You've done the tests, you've got your results. Now, if someone wants to spend the time proving me wrong they are welcome to a an original scan to try it. :-) No - you've done the tests and in this case evidently .jpg wins. I'm surprised, but I've seen that happen before. I suppose it depends on just how sharp the sharp edges were on your images. Steve. ___ dev mailing list dev@openstreetmap.org http://lists.openstreetmap.org/listinfo/dev
Re: [OSM-dev] Out of copyright WMS offering - looking for assistance
2009/10/19 Andy Robinson (blackadder-lists) ajrli...@googlemail.com: Steve Hosgood wrote: Sent: 19 October 2009 9:39 AM To: dev@openstreetmap.org Cc: 'josm-dev' Subject: Re: [OSM-dev] Out of copyright WMS offering - looking for assistance Must they be .jpgs? You'll get a lot of sharp edge artefacts using the .jpg format on a map. In theory at least (and in practice whenever I've done it) .png will give better results on material such as that. Bear in mind these are printed maps, not rasterised vector maps. We tried with png but you can get much lower file size without loosing too much definition with the jpg's. Now, if someone wants to spend the time proving me wrong they are welcome to a an original scan to try it. I assume you've run something like optipng [1] over the images. I've found with complex images you can get quite a size decrease with 'optipng -o7 *.png'. However, if the size of the images at present is ok, there's no need to change anything. The quality looks fine as it is. On another note, in my tests I managed to reduce the size of the default Mapnik tiles by about 5% but I guess that reduction is small since they've already been slightly optimised. -- Matt Williams http://milliams.com [1] http://optipng.sourceforge.net ___ dev mailing list dev@openstreetmap.org http://lists.openstreetmap.org/listinfo/dev
[OSM-dev] Code for OSTN02?
Hi all, Does anyone know a linux compatible C or C++ implementation of OSTN02? (that is the fancy OS grid to ETRS89 which is a local version of WGS84). I need it to run under linux which rules out the OS implementation. I am currently using Geo-Coordinates-OSGB-2.04 in perl and although it works fine, it would be a nightmare to deploy to other machines. Open source solutions are strongly preferred. Zero monetary cost is a must! Also, the Helmert approximation seems to be not accurate enough. Ideas anyone? Tim ___ dev mailing list dev@openstreetmap.org http://lists.openstreetmap.org/listinfo/dev
Re: [OSM-dev] Code for OSTN02?
Good luck, this is the one projection that the OS actually sell for professional uses. I thought that OSTN02 wouldn't actually be needed unless for the very high precision stuff. The standard OSGB36 transform should be good enough otherwise? On Mon, 2009-10-19 at 20:33 +0100, TimSC wrote: Hi all, Does anyone know a linux compatible C or C++ implementation of OSTN02? (that is the fancy OS grid to ETRS89 which is a local version of WGS84). I need it to run under linux which rules out the OS implementation. I am currently using Geo-Coordinates-OSGB-2.04 in perl and although it works fine, it would be a nightmare to deploy to other machines. Open source solutions are strongly preferred. Zero monetary cost is a must! Also, the Helmert approximation seems to be not accurate enough. Ideas anyone? Tim ___ dev mailing list dev@openstreetmap.org http://lists.openstreetmap.org/listinfo/dev ___ dev mailing list dev@openstreetmap.org http://lists.openstreetmap.org/listinfo/dev
Re: [OSM-dev] Code for OSTN02?
Thomas Wood wrote: Good luck, this is the one projection that the OS actually sell for professional uses. I thought that OSTN02 wouldn't actually be needed unless for the very high precision stuff. The standard OSGB36 transform should be good enough otherwise? I don't think they are that restrictive. They seem to have a freeware library and API for windows (closed source) here http://www.ordnancesurvey.co.uk/oswebsite/gps/osnetfreeservices/furtherinfo/questdeveloper.html On accuracy, converting NR730100 to ETRS89, OSTN02: 55.33149630, -5.58040715 Helmert (based on OS spreadsheet[1], ported to C++): 55.331258, -5.580440849 Nearby.org.uk: 55.331495, -5.580444 http://www.rutter.uklinux.net/ostowiki.html: 55.33149480 -5.58044426 Streetmap.co.uk: 55.331967, -5.579682 ganzc (by Chuck Gantz): 55.3314950512,-5.58040512526 [1] http://www.ordnancesurvey.co.uk/oswebsite/gps/docs/ProjectionandTransformationCalculations.xls First observation: there is a range of accuracies from various sources! ganzc was better than I remember. I will investigate that one further. Tim ___ dev mailing list dev@openstreetmap.org http://lists.openstreetmap.org/listinfo/dev