Elevation colour bands in Aven

[Olly Betts]

On Mon, Oct 22, 2007 at 12:05:21PM +0100, Paul Footleg Fretwell wrote:
> I am trying to plot all the caves over a wide area (entire Matienzo 
> expedition area) in Aven, so that I can generate overlays for Googlemaps. 
> One thing that is a problem with this is that elevation detail is 
> difficult to determine by colour due to the wide elevation range of the 
> caves (615m to 40m). What appears to happen is most of the caves are 
> pushed into the blue banding. Looking at the colour bar in Aven, it 
> appears that the blue zone takes up a lot of the elevation range, as does 
> the red zone. In the middle the bands for Orange, yellow, green are fairly 
> narrow.

It does depend a bit on your display, but the current colour gradient
could probably be improved.  Here's the current one numerically (blue at the
top, red at the bottom, some lines are duplicated for some reason):

/* color_map file created by the GIMP */
/* http://www.gimp.org/               */
color_map {
        [0.000000 color rgbt <0.236267, 0.265919, 1.000000, 0.000000>]
        [0.164434 color rgbt <0.236267, 0.476027, 0.853163, 0.000000>]
        [0.377295 color rgbt <0.236267, 0.686135, 0.512695, 0.000000>]
        [0.377295 color rgbt <0.236267, 0.686135, 0.512695, 0.000000>]
        [0.470785 color rgbt <0.557528, 0.713981, 0.256348, 0.000000>]
        [0.558222 color rgbt <0.878788, 0.741828, 0.000000, 0.000000>]
        [0.558222 color rgbt <0.878788, 0.741828, 0.000000, 0.000000>]
        [0.651085 color rgbt <0.840094, 0.633394, 0.090664, 0.000000>]
        [0.729549 color rgbt <0.801401, 0.524960, 0.181328, 0.000000>]
        [0.729549 color rgbt <0.801401, 0.524960, 0.181328, 0.000000>]
        [0.875000 color rgbt <0.859034, 0.275473, 0.123558, 0.000000>]
        [1.000000 color rgbt <0.916667, 0.025986, 0.065789, 0.000000>]
} /* color_map */

Unfortunately, changing this currently requires a recompile, but it wouldn't
be too hard to make it a configuration file.

Martin Green tried making a "spectroscopically correct" gradient, but I think
we decided it actually looked less good.  It's crossed my mind since that
perhaps he didn't factor in the human eye's different responsiveness to
light at different frequencies.  Can you remember more about this Martin?

> Would it be possible to extend the colour banding to include 
> purple at the bottom, and lessen the bias of the red and blue bands to 
> give better visual differentiation of passages at elevations in the middle 
> of the range?

I'll try to make it possible to specify the gradient, then people can try
out different ones and we can try to find an improved default.

As Erin says, it would also be useful to be able to vary which heights map
to where (and similarly for dates, and errors), but that's a separate issue.

Cheers,
    Olly