Choosing Colours - the book

How colour works - extracted from the introduction to Choosing Colours

The following four pages contain various colour models - structures that are used to explain how colours relate to each other and how we perceive them - that are all useful in helping to understand how the palettes in this book work.

Colour models have been in use for millenia; Aristotle developed the first western scale of colours, which in turn influenced Newton's decision to settle on seven colours for the rainbow (he had at various points thought there to be five or eleven). Even modern systems still refer to colour in essentially Newtonian terms.

To an extent all colour models are valid in some way. Even the most sophisticated 3-D colour models in use by scientists still cannot account for some colours, so no one model necessarily reigns supreme; they are all fallible. That's because human vision itself is not consistent and it is not neat. It's not consistent because we each of us are blessed with minutely different configurations of optical receptors in our eyes (the rods and cones). It's not neat because vision is not a logical, but a biological mechanism that has developed as an important tool for a myriad number of human activities: it is analogue, not digital, and its sensitivity to the colours of the rainbow varies (see colour model 3 on page 10).

Colour model 1 is the one that we learnt at school. The primary paint colours - which cannot be produced by mixing other colours - are yellow, red and blue and their secondaries are made by intermixing. The result is that when the model is arranged as a circle, the complementary of red is green (being opposite). Complementaries appear to 'fight' when placed together and when mixed together make brown of some kind. Many of the subtler and more complex colours in the palettes involve the dilution of one colour with a small quantity of its complementary plus the addition of white and/or black. Other palettes pair complementary colours together to deliberate effect. This model is a pragmatic one because it works for pigments and chemical colorants, which in turn are often imperfect. It is also known as the 'subtractive' model because as the colours are mixed, they cancel out a degree of their own luminosity and become darker. So purple and green are less bright than their combined constituent primaries. A case of the sum being less than the parts.

Colour model 2 is a four-colour primary model, unusual at first sight but actually formulated as long ago as 1878 by the german physiologist Ewald Hering, who thought that yellow, red, blue and green, together with black and white, formed a palette of six 'natural' colours. This is partly because green is perceived as a colour independent of its component subtractive colours, blue and yellow. (The eye's physiology sets up three signal channels to the brain, each one of opposing colours: black-white, red-green, blue-yellow. Green has its name up there with the others.)

This four-colour palette creates an interesting set of four secondary colours: orange, violet, turquoise and lime green. One result of this, for example, is that the complementary of turquoise becomes orange and you can see a muted example of this complementary relationship in palette 58. The four-colour model was refined in the twentieth century, published and republished and then technically perfected by the Swedish Colour Centre Foundation, who issued it as the Standard Color Atlas of the Natural Color System (NCS), in 1979. It's now adopted as several national standards and by paint and coatings manufacturers worldwide. Palette 62 contains all the primary, secondary and tertiary (between the secondaries) colours of this four-colour system.

Colour model 3 reverts to three primaries and they're different again from colour model 1. The reason is that these are the primary and secondary colours of light: green, warm red and purple-blue light when mixed together produce white light (just as red, yellow and blue paint are supposed to make black but, because of pigment imperfection, make a murky brown). Their secondary colours are also interesting: cyan blue (similar to turquoise), magenta and yellow, which most bizarrely of all is produced by mixing red and green light. Note that because the secondaries result from the addition of two other light colours, they are also more luminous. Not surprisingly, this is called an additive palette.

There is no magic behind the choice of colours and no alchemy in their mixing. The choice of red, green and blue primaries from among the colours of the visible spectrum (i.e. the rainbow) is all down to the way we're built. We perceive a very narrow band of electromagnetic radiation (which includes radio and gamma waves) and we call that visible portion light. The three types of sensor on our retinas that respond to different colours within the visible spectrum have peak sensitivity in different areas: one peaks in the blue part of the spectrum, one in the green and one in the red. If we had sensors that were sensitive to infra-red or ultra-violet (as some animals do) we might see more colours.

As it is because we have only three types (plus one for monochrome vision), our entire colour world is dependent on them: all perceivable colours are made up from various combinations of their activities. That gives us a measly, separately identifiable 16 million colours to play with.

Colour model 4 takes the secondary colours of light - cyan, magenta and yellow - and works them backwards. The theory is that if you take three sheets of transparent plastic in these three colours, you should be able to produce the light primaries by subtraction. Thus a sheet of cyan plastic film held over a sheet of magenta plastic film should give a less luminous, but blue light. It works a bit like mixing paint and these theoretical subtractive primaries, cyan, magenta and yellow, ought (in opposition to colour model 3) to make black when overlaid. Of course they don't because coloured plastic, like coloured anything (other than light), is never 100 percent totally purely coloured.

It's the same story with printing inks. This model lies behind modern printing methods, which use cyan, magenta and yellow transparent inks over white paper (a source of reflected light) to reproduce full-colour photographs in books and magazines. In practice printers also use black to bolster the performance of the three colours and the resulting printing method (which is part-additive, part-subtractive) is known as CMYK process printing, now used worldwide.

But CMYK has always had its weaknesses, notably poor greens and especially oranges, which always look dull and muddy, mainly due, again, to the technical limitations of the inks. Pantone®, one of the largest colour authorities in the world, and certainly the single most persuasive voice in the graphics and printing industries, has solved these problems to a large extent with the introduction of two further ink colours, green and orange, as shown on the left. Together with CMYK they make a six-colour process method called Hexachrome®. This book is one of very few consumer volumes to be printed in Hexachrome®; it was an essential choice for the accurate colour rendering of the palettes and an added advantage in that it reproduces vibrant photographs well.

Paint matches >>

^ Back to the top ^