[mapserver-commits] r8436 - trunk/docs/mapfile

[svn at osgeo.org]

Author: tbonfort
Date: 2009-01-16 13:43:47 -0500 (Fri, 16 Jan 2009)
New Revision: 8436

Modified:
   trunk/docs/mapfile/outputformat.txt
Log:
add AGG and formatoptions


Modified: trunk/docs/mapfile/outputformat.txt
===================================================================
--- trunk/docs/mapfile/outputformat.txt	2009-01-16 18:35:22 UTC (rev 8435)
+++ trunk/docs/mapfile/outputformat.txt	2009-01-16 18:43:47 UTC (rev 8436)
@@ -16,6 +16,13 @@
   
 .. code-block:: mapfile
 
+    OUTPUTFOTMAT
+      NAME aggpng24
+      DRIVER AGG/PNG
+      MIMETYPE "image/png"
+      IMAGEMODE RGB
+      EXTENSION "png"
+    END
     OUTPUTFORMAT
       NAME gif
       DRIVER "GD/GIF"
@@ -67,7 +74,8 @@
 DRIVER [name]
     The name of the driver to use to generate this output format. Some driver 
     names include the definition of the format if the driver supports multiple 
-    formats. For GD the possible driver names are "GD/Gif", "GD/PNG", 
+    formats. For AGG, the possbile driver names are "AGG/PNG" and "AGG/JPEG".
+    For GD the possible driver names are "GD/Gif", "GD/PNG", 
     "GD/WBMP" and "GD/JPEG". For flash the driver is just called "SWF". For 
     output through GDAL the GDAL shortname for the format is appended, such 
     as "GDAL/GTiff". Note that PNG, JPEG and GIF output can be generated with 
@@ -83,7 +91,8 @@
     - RGB: Render in 24bit Red/Green/Blue mode. Supports all colors 
       but does not support transparency.
     - RGBA: Render in 32bit Red/Green/Blue/Alpha mode. Supports all 
-      colors, and alpha based transparency.
+      colors, and alpha based transparency. All features are rendered
+      against an initially transparent background.
     - BYTE:  Render raw 8bit pixel values (no presentation).  Only 
       works for RASTER layers (through GDAL) and WMS layers currently.
     - INT16: Render raw 16bit signed pixel values (no presentation). 
@@ -127,4 +136,31 @@
     - GDAL/\*: All FORMATOPTIONs are passed onto the GDAL create function. 
       Options supported by GDAL are described in the detailed documentation 
       for each GDAL format
-           
+    - GD/PNG and AGG/PNG both support quantizing from 24/32 bits to 8bits,
+      in order to reduce the final image size (and therefore save bandwidth) (see
+      also http://trac.osgeo.org/mapserver/ticket/2436#comment:4 for strategies
+      when applying these options):
+
+        - "QUANTIZE_FORCE=on" used to reduce an RGB or RGBA image into an 8bit 
+          (or less) paletted images. The colors used in the palette are selected
+          to best fit the actual colors in the RGB image (RGBA quantization was
+          introduced in mapserver 5.2, and requires the --enable-experimental-png
+          flag to be used when configuring)
+        - "QUANTIZE_COLORS=256" used to specify the number of colors to be used
+          when applying quantization. Maximum value is 256. Specifying anything
+          between 17 and 255 is probably a waste of quality as each pixel is still
+          encoded with a full byte. Specifying a value under 16 will produce tiny
+          images, but severly degraded.
+        - "QUANTIZE_NEW=on" used to force using a slower but higher quality 
+          quantization algorithm for RGB images. The default behavior is to use
+          a fast but crude quantization algorithm, that usually results in duplicate
+          entries in the palette, and visible artifacts when tiling.
+        - "PALETTE_FORCE=on" is used to reduce image depth with a predefined palette.
+          This option is incompatible with the previous quantization options.
+        - "PALETTE=/path/to/palette.txt" is used to define the absolute path where 
+          palette colors can be found. This file must contain 256 entries of r,g,b
+          triplets for RGB imagemodes, or r,g,b,a quadruplets for
+          RGBA imagemodes. The expected format is one triplet (or quadruplet) per
+          line, each value separated by commas, and each triplet/quadruplet on a
+          single line. If less than 256 triplets are found
+          in the file, 256-n remaining colors will be computed by quantization.