If you find this work useful please cite
Peter Kovesi. Good Colour Maps: How to Design Them.
arXiv:1509.03700 [cs.GR] 2015
This website presents a collection of colour maps that have been designed to have uniform perceptual contrast over their whole range.
Many colour maps provided by vendors have highly uneven perceptual contrast over their range. Colour maps may have points of locally high colour contrast leading to the perception of false anomalies in your data when there is none. Conversely colour maps may also have 'flat spots' of low perceptual contrast that prevent you from seeing features in the data.
To illustrate this the colour maps shown below are rendered on a test image consisting of a sine wave superimposed on a ramp function. The amplitude of the sine wave is modulated from its full value at the top of the image to 0 at the bottom.
What we are hoping to see is the sine wave uniformly visible across the image from left to right. We also want the contrast level, the distance down the image, at which the sine wave remains discernible to be uniform across the image. At the very bottom of the image, where the sine wave amplitude is 0, we just have a linear ramp which simply reproduces the colour map. Given that the underlying data is a featureless ramp we should not perceive any identifiable features across the bottom of the image.
At the top row of the test image the sine wave amplitude from peak to trough is 10% of the total data range. It is not unusual for the sine wave pattern to completely disappear in parts of some vendor colour maps. On the other hand the perceptually uniform colour maps exhibit no false features and the sine wave pattern is uniformly visible across the full width of the test image.
Note the colour maps presented here are intended for the display of data that varies over a continuous range. For data sets containing a limited set of categorical values it is suggested that you refer to the work by Cynthia Brewer at www.colorbrewer2.org
These colour maps are released under the Creative Commons BY License. A summary of the conditions can be found here. Basically, you are free to use these colour maps in anyway you wish as long as you give appropriate credit.
If you find this work useful please cite:
Peter Kovesi. Good Colour Maps: How to Design Them.
arXiv:1509.03700 [cs.GR] 2015
|Adobe Colour Table|| CETperceptual_act.zip|
Adobe's .act format for use within Photoshop.
|ArcGIS style file|| CETperceptual_ArcGIS.style|
The maps are approximated using multi-part ramps with 64 segments. Please note that I have had to apply an adjustment to the colour ramps in an attempt to compensate for some strangeness in the way ArcGIS converts CIELAB values in the style file to RGB. I have been unable to reconcile the ArcGIS conversion results with those I obtain from any other published conversion routines. I am not entirely satisfied with the result but I believe the final ramps, as rendered, are reasonably close to their original design.
|CSV floating point RGB values in the range 0-1.|| CETperceptual_csv_0_1.zip |
Can be used with the Madagascar package.
|CSV integer RGB values in the range 0-255.||CETperceptual_csv_0_255.zip|
|ER Mapper, Intrepid||CETperceptual_ERMapper.zip |
ER Mapper's .lut format. Used by Intrepid and I think these can also be used with MapInfo.
|Geosoft Oasis montaj||
Geosoft's .tbl format.
GeoGraphix binary .PAL format.
Dynamic color palette tables (.cpt) files for GMT, The Generic Mapping Tools.
If you use gmt makecpt to construct a static CPT file from these files you should avoid using a large z interval when using the -T option otherwise the properties of some colour maps may be compromised. Choose a z interval such that the total range is divided into at least 64 intervals.
Paradigm GOCAD .cmap format.
ImageJ's .lut format.
For those working in Julia this package allows you to geneate all these colour maps for use with various Julia plotting packages.
|Landmark DecisionSpace Geophysics|| CETperceptual_cl2.zip |
Landmark's .cl2 format.
|MagicPlot||These colour maps are available in MagicPlot's format at magicplot.com/wiki/palettes|
A stand-alone function that contains pre-generated arrays of the perceptually uniform colour maps.
|MicroImages TNTmips|| No need to download |
These colour maps are incorporated in their latest release.
|Micromine|| No need to download |
These colour maps are included in Micromine 2018.
NCL, NCAR Command Language .rgb files.
Load using: Color map editor -> Choose preset -> Import
|Petrel, DUG Insight||CETperceptual_alut.zip |
Schlumberger Petrel .alut format. These can also be used with DownUnder GeoSolutions Insight.
Petrosys .pal format. These colour maps are now also available directly from Petrosys.
This package gives you access to these colour maps for use with Python plotting programs such as Bokeh, Matplotlib, HoloViews GeoViews, and Datashader. Thanks to James Bednar.
The pals package gathers together a number of colour maps including the CET perceptually uniform colour maps and also includes R code for generating the colour map test image.
|QGIS style file||CETperceptual_QGIS.xml|
The maps are approximated using multi-part ramps with 64 segments. Load the style file using the QGIS menu sequence:
Settings -> Style Manager -> Share -> Import
Golden Software's Surfer .clr format.
Let me know of any other formats I should generate.
Please note that I have not been able to test all these file formats as I do not have access to all these packages. If you encounter any problems please let me know. Any other feedback you might have would also be appreciated.
This work was supported by the Centre for Exploration Targeting, School of Earth and Environment at The University of Western Australia.
I am indebted to Tom Horrocks for his help in generating the ArcGIS style file.
The colour maps are organised according to the attributes: Linear, Diverging, Rainbow, Cyclic, and Isoluminant.
Linear colour maps are intended for general use and have colour lightness values that increase or decrease linearly over the colour map's range.
Diverging colour maps are suitable where the data being displayed has a well defined reference value and we are interested in differentiating values that lie above, or below, the reference value. The centre point of the colour map will be white, black or grey. It should be noted that, in general, diverging colour maps have a small perceptual flat spot at the centre. The exception being linear-diverging maps which avoid this problem.
Rainbow colour maps are widely used but often misused. It is suggested that they be avoided because they have reversals in the lightness gradient at yellow and red which can upset a viewer's perceptual ordering of the colours in the colour map. However, they are attractive and perhaps can have a legitimate use where the main aim is to differentiate data values rather than communicate a data ordering. I believe the two rainbow colour maps presented here have minimal badness though they do have localised perceptual flat spots at yellow and red.
Cyclic colour maps have colours that are matched at each end. They are intended for the presentation of data that is cyclic such as orientation values or angular phase data. They require particular care in their design (the standard colour circle is not a good map).
Isoluminant colour maps are constructed from colours of equal perceptual lightness. These colour maps are designed for use with relief shading. On their own these colour maps are not very useful because features in the data are very hard to discern. However, when used in conjunction with relief shading their constant lightness means that the colour map does not induce an independent shading pattern that will interfere with, or even hide, the structures induced by the relief shading. The relief shading provides the structural information and the colours provide the data classification information.
Colour Blind colour maps. These are not designed to be merely 'colour blind safe'. These maps have been constructed to lie within either the 2D model of protanopic/deuteranopic colour space, or the 2D model of tritanopic colour space. Hopefully by working within these colour spaces people who are colour blind will be able to share a common perceptual interpretation of data with those who have normal colour vision. It also ensures maximal use of the available colour spaces, and allows chroma and lightness to be properly used in the design of colour maps. I would value any feedback on the usefulness, or otherwise, of these maps.
I devised the following naming convention to allow semi-automatic naming of colour maps in a way that described their key characteristics. However the resulting names are too complex, too long to type, impossible to remember, and I do not use them!
In practice I simply refer to the colour maps by a number prefixed by one or two characters to indicate whether the map is linear (l), diverging (d), rainbow (r), cyclic(c), isoluminant (i) or colour blind (cb). Thus, for example, the linear greyscale is 'l1' or 'cet-l1', the heat colour map is 'l3' or 'cet-l3' and the blue-white-red diverging map is 'd1' or 'cet-d1' etc, etc. In the visual catalogue shown further below you will see each colour map with both its simple and complex names.
Anyway, here is the description of the complex naming scheme.
|r - red||g - green||b - blue|
|c - cyan||m - magenta||y - yellow|
|o - orange||v - violet|
|k - black||w - white||j - grey|