What’s in a colour?

By Marcus Wilson 23/07/2013 4


When I was young (about six-ish)  I had a variety of ambitions. Some of them I shared with a lot of other boys of my age, such as being a train driver and playing cricket for England. Some were more particular to me, such as becoming a biologist and discovering a new colour. 

Needless to say I failed on all accounts. One I got close to – being a physicist is not so far away from being a biologist.  I’ve at least watched England play cricket (including an England v India match at Lord’s – in the members’ guests area – that was rather neat) and stood on the footplate of a steam engine. Discovering a new colour, however, is something I was not likely to achieve from the outset.

I had a vague idea that if I mixed enough paints together I’d hit on a combination that no-one had tried before (maybe purple and green with just a hint of orange) and, hey-presto, they’d mix together to some entirely colour previously unknown to science. The colour would naturally be named after me, and become an instant hit with home decorators. Out would go ‘Magnolia’, in would come ‘Wilurple’. 

I gave up on the ambition long before I found out why it was unlikely to work. The CIE colour chart encapsulates the situation neatly. There are only three different colour receptors (‘cones’)  in the human eye. By having the ‘red’, ‘green’ and ‘blue’ cones stimulated differently, one sees different colours. The CIE chart puts all possible colours onto a 2d grid. One defines the variable ‘x’ as being the fraction of the total stimulation that is accounted for by the red cones; the variable ‘y’ as the fraction of the total that is accounted for by the green cones. (One could define ‘z’ in a similar way for the blue cones, but it is redundant since x plus y plus z must equal 1.) Then ‘x’ and ‘y’ defines a colour. The chart shows it. 

All possible colours are shown on this chart. The outside of the curved space shows the colours of the spectrum – those stimulated by a pure wavelength of light. The others are due to combinations of wavelengths. At x=1/3, y=1/3 (and so z=1/3) there is white. It isn’t possible to go outside this chart, and therefore it contains all possible colours. D’oh.

But, there is hope. The response of the green cones of the eye is entirely overlapped by those of the red and the blue. This means it isn’t possible to find a wavelength of light that stimulates JUST the green cones. If, somehow, one could stimulate cells artificially, one might be able to trigger green cones to fire without any response from red and blue. And then the person would be seeing a colour they’ve never experienced before. 

 


4 Responses to “What’s in a colour?”

    • The chart doesn’t show the intensity of the colour. The mapping to x and y coordinates normalizes by dividing by the total intensity, so anything that stimulates the R, G, and B by the same amount will have x=1/3, y=1/3 (and z=1/3). Grey and black are just like white in this sense – at 1/3, 1/3, 1/3. I’m not sure about brown. I have a vague memory of attending a talk about colour in which ‘brown’ was labelled as a bit of an oddity colour-wise.

  • Brown is a dark yellow. Try taking two yellow cards. Cut a hole in one. Lay one behind the other, then lower the back one so it is shaded. Voila! Brown! If the back of the front card is white and you stand towards a light source, the white reflects the source onto the shaded “back” card which enhances the effect.

    It has the same CIE coordinates as yellow, The luminance value is different.

  • “Brown” is basically dull yellow (different shades of brown head further towards the green or red ends of yellow). Go into MS Paint (or any other program that lets you pick specific colours) and mess with the RGB (or HSV) sliders. The fun with colours really starts when you look at non-spectral colours (that’s the straight line bit at the lower edge of the graph), tetrachromats (people with 4 different cones), and stuff like that!