specify and connect colorbar ranges

Hello,

I'm having troubles making several colorplots and 'connect' the color ranges between them. For example, the plotted values are between 0.1 and 0.25 for the first and 0.25 to 0.5 for the second. What I want is plot the first data set using colors that range from blue to green, and then the second with colors ranging from green to red. Or something like this but at least making sure that the max color of plot 1 is the min color of plot 2. As they are in different scripts, I need an easy way to manually 'set' the color ranges. I have been trying to get this to work with caxis/ clim, but these seem to only change the bar limits, rather than the actual colors... I can't get it to work.

Here is my code for the first graph:

for j = 1: length(stat)
file=char(stat(j,1));               %metric to be plotted
temp=eval(file);                    %get data  
figure(j)
h=pcolor(X/10,Y/10,temp); hold on   %start pseudoplot
set(h, 'edgecolor','none');         %get rid of edgelines
title(titles(j),'fontsize',11, 'fontweight', 'bold'); % add title

caxis([0.1 0.25])                   % specify colorbar limits
colorbar;
end

Please help me to accomplish this! Thanks very much for any help!

BW

Jasper

Answer 1

[Adam]

You may want/need to create your own colourmap, but just use

caxis([0 1])

or whatever the full range across all your plots would be for every plot rather than setting the caxis range to cover the data in the individual plots. If you do this then that range will always map to the full colourmap.

If you use [0 1] range for every plot then e.g. yellow will always be between 0 and 0.1 irrespective of whether a given plot has data there and your boundary of e.g. 0.25 will map to the same colour in both plots.

I have never used pcolor so I don't know if it has any special properties that make this not work, but for things like imagesc which I usually use it should be fine so I assume for pcolor too.

[Jasper]

Thanks for your answer, and it was originally what I did as well. Issue is that the 'gradient' in the first plot is smaller than the second, so when using general colorscale (e.g. 0.1-0.5), the first one is basically one color without clear gradients, although they are important. Thats why I want to look into 'stretching' one of the colormaps once I get this working. Hope this makes sense. Jasper

[Adam]

Well, if you create your own colourmap you can define more colours or faster varying colours in some parts than others.

If the data to which the colourmap will be applied is always predictable (e.g. 0 to 1) then you can define that e.g. colour changes quickly up to 0.25 then more gradually afterwards. You can also define the level of granularity of the colourmap. I'm not sure what the highest is, but with builtin colourmaps you can create them with a large number of sample points - e.g.

jet( 1024 );

[Jasper]

Allright, I will look into that, thanks for the advice.

Anyone with experience with that, feel free to provide me with some basics. Especially changing color gradients within the map.

Jasper

[Adam]

c1 = [1 0 0];
c2 = [0 1 0];
c3 = [0 0 1];

for i = 1:3
   c(1:200,i) = linspace( c1(i), c2(i), 200 );
   c(200:256,i) = linspace( c2(i), c3(i), 57 );
end

for example, would create a colourmap that graduates from red to green over 200 samples, then green to red over the remaining 56 samples of the colourmap.

Obviously you could extend it to as many colour points as you wish. You also don't have to restrict yourself to 256-sized colourmap either.