No anti-aliasing for cartoline symbol?

[Ken Lord]

Hi James,

The documentation states that 'antialias true' within the symbol
definition is just for truetype fonts, 'antialias true' in the style
affects both truetype fonts and cartolines.

Try your map with no antialias statement in the symbol or in the
style, then compare the results with 'antialias true' just within the
style.

I just started using the cartoline symbol for roads myself, and I can
see a good difference ... still a little jaggy, but a good difference.

Perhaps the 72dpi resolution of the output image is keeping the jaggies visible.

Cheers,
Ken Lord
Vancouver BC



On Apr 1, 2005 7:11 AM, James Sohr <jms4 at po.cwru.edu> wrote:
> Like so many others, I've used the circle symbol with various  sizes to
> draw my lines for the past few years.  The results have been pretty good,
> but the aliasing on the roads has always bothered me, especially when the
> true-type labels are beautfully anti-aliased right beside them.
>
> I recently read in the archives that the relatively new cartoline symbol
> supports anti-aliasing.  However, when I use the cartoline symbol and turn
> antialiasing on, the lines drawn are still aliased.
>
> I'm using Mapserver 4.5 on Windows XP with the latest GD library.  Is
> there something in my setup that is preventing the cartoline from being
> anti-aliased?  Does someone have an online example that shows the
> cartoline symbol being anti-aliased?
>
> The pertinent sections in my symbols file are:
>
>      SYMBOL
>       NAME 'road'
>       TYPE CARTOLINE
>       ANTIALIAS TRUE
>      END
>
>      SYMBOL
>        NAME 'roadov'
>        TYPE CARTOLINE
>        ANTIALIAS TRUE
>      END
>
> And in my mapfile:
>
>        CLASS
>
>            STYLE
>              SYMBOL 'road'
>              SIZE 5
>              COLOR 255 0 0
>              ANTIALIAS TRUE
>           END
>
>           STYLE
>              SYMBOL 'roadov'
>              SIZE 3
>              COLOR 255 255 255
>              ANTIALIAS TRUE
>           END
>      ....
>
> Any help is appreciated.
>
> Thanks,
>  James
>