No subject

is where the problem lies - the higher level display programs are deliberately
restricted to setting a small number of rgb values so that 8-bit colour is
used.  The way it does this is to generate a (possibly immutable) 216 colour
colour-cube.

My gripe is that in many cases I am only using a corner of it,
so I only get effectively 50 or so colours - either the corners (?)
or (probably) the quantization of the cube are not decided by looking
at histograms of the data, so the displayed range on the 8 bit display
is not optimised for your dataset.  It is probably possible to work around
this by judicious insertion of histogram stretches, etc, but I think this
would be hard work and it all gets you away from the underlying data!

This diagnosis was arrived at by displaying the present colortable using
one of the x display tools, while mucking around with the colour tables for
some rasters in grass, as well as the abovementioned inspection of the code
("looking under the hood" as someone once said on this list).  I may be barking
up completely the wrong tree (stalk?) so I would be glad to hear comments.

Simon Cox

----
___________________________________________________
      __  L     Dr Simon Cox			
   ,~'  L_|\    VIEPS Department of Earth Sciences, 
,-'         \   Monash University, Clayton Vic 3168        
(            \  Australia
\    ___     /          Phone +61 3 565 5762
 L,~'   "\_x/           Fax   +61 3 565 5062
           u    simon at cerberus.earth.monash.edu.au
___________________________________________________