If your mapserver is compiled with --enable-experimental-png, you can change QUANTIZE_FORCE to QUANTIZE_NEW to use a slower but higher quality quantization algorithm that usually fixes these problems.
best regards, thomas www.camptocamp.com +33 4 79 26 57 97 On Wed, May 6, 2009 at 05:52, David Nugent <dav...@datalinktech.com.au> wrote: > Greetings all, > Once I completed upgrading to mapserver 5.4, I decided I'd give the new AGG > renderer a try. > To date I have been using the GD renderer in 8 bit color mode. > One problem I had with it the renderer, and my main motivation to use the > AGG renderer (apart from what appears to be an overall improvement in > appearance) was that any layer containing "ANGLE FOLLOW" within a class > label definition caused php/mapscript to crash rendering meta tiles (for > ka-map, if that matters). The AGG renderer solved that problem immediately > and it works flawlessly in that respect. > However, as a by-product of changing renderer, the metatiles produced by > mapserver differ slightly in colour. Some agree exactly to the RBG colours I > am providing, others are rendered with slight variations, and when viewed on > the web it becomes fairly obvious where the meta tile borders are. > I have tried both > OUTPUTFORMAT > NAME 'AGG_Q' > DRIVER AGG/PNG > IMAGEMODE RGB > FORMATOPTION "QUANTIZE_FORCE=ON" > FORMATOPTION "QUANTIZE_DITHER=OFF" > FORMATOPTION "QUANTIZE_COLORS=256" > END > and > OUTPUTFORMAT > NAME 'AGG' > DRIVER AGG/PNG > IMAGEMODE RGB > END > but both produce the same results. I have even tried fiddling with the > FORMATOPTIONs in the AGG_Q block to no effect. > I've double-checked the mapscript used to render the tiles to make sure that > the correct output format is being selected. > Is there a known solution to this? How can I force all tiles produced to use > the exact same colour settings? I would have thought the quantize option > would resove it, although obviously this would have a slight detrimental > side-effect for anti-aliasing. > My map file does include various "ANTIALIAS" options (which the document > states are ignored in AGG mode) so I can fall back to GD rendering should I > ever get "ANGLE FOLLOW" working again. > Regards, > -- > David Nugent (dav...@datalinktech.com.au) > Software Engineer > datalink technologies pty ltd > Mobile GPS/GPRS business process solutions > http://www.datalinktech.com.au/ > > _______________________________________________ > mapserver-users mailing list > mapserver-users@lists.osgeo.org > http://lists.osgeo.org/mailman/listinfo/mapserver-users > > _______________________________________________ mapserver-users mailing list mapserver-users@lists.osgeo.org http://lists.osgeo.org/mailman/listinfo/mapserver-users