Like many people, you may find yourself laboring under the misapprehension that the world you live in is colored. It (the world), in fact, is not (inherently colored). Terrible, no? What we normally think of as color is our brain’s response to a specific physical stimulus, electromagnetic wavelength, much like sound is our brain’s response to pressure waves traveling through air. Big deal, that’s just a technical version by a different name, you may think. This is more than just a small syntactic distinction, because our conscious experience of the colored world doesn’t smoothly correspond with those pesky wavelengths of light. What we see as purple, for example, is actually a combination of the shortest and the longest wavelengths. If seeing color were like hearing sound, then looking at purple would be like hitting the highest and lowest notes on a piano and hearing the same pitch even though one string is much shorter. No one knows why.
That is what color is, but how does it work? The truth is that although the visual system is the best understood part of the brain, we know very little about how color works once it leaves the eye. The eye is probably only the tip of the iceberg, but it is better than nothing. [while we do know a bit about color in the brain, a) that’s a later post, and b) even the location of color in the brain is contested].
color in the eye
Light, once it enters your eye, eventually hits the photoreceptors on your retina. The chromatic or color sensing photoreceptors can be sensitive to either short, middle or long wavelengths, or what we might call red, green and blue. The signals from the cones are then added together and rearranged so instead of sensing absolute amounts of red, green and blue light, color is represented by either a red vs. green or blue vs. yellow opponent process. Yellow is just green and red at the same time (mixing light is quite unlike mixing paint). This is why you can’t have a reddish green, or a yellow with a hint of blue- these color are polar opposites. Green, red, blue and yellow are (ironically) known as the Hering colors, named after the person who first proposed them.
An interesting effect of having only three cones is that various combinations of different wavelengths look the same to us- these are known as metamers. Now, everyone has the same cones in their eyes, so people argue that the red one person sees must be the same as what another person sees, because the same causes must produce the same effects. While this is true in principle, it turns out that the distribution of different types of cones varies wildly from person to person. Scientists have no idea what this means. The larger issue is that cones aren’t responsible for our experience of color- some region in the brain is, and that could be different from person to person. It can’t be too different (I will explain this later), but it is possible that my red is a little different from yours.
Also, some women have four different types of cones in their eyes, as opposed to three- they are known as tetrachromats, and see things very different from the rest of us. They probably don’t see any ‘alien’ colors, instead they just have less metamers. According to one tetrachromat, what looks like a bright blue sky to us is instead dappled with pinks and oranges to her. (source: http://www.radiolab.org/2012/may/21/ , a recent radiolab episode about color). Isn’t that weird?
Finally, the retina can detect more colors than it currently sees- the low wavelength colors are absorbed by the eye’s lens. If this lens is removed, people gain the ability to see what is traditionally known as ultra-violet light. This happened to Monet, who had the lens of one of his eyes removed due to cataracts clouding his vision (source: http://www.downloadtheuniverse.com/dtu/2012/04/monets-ultraviolet-eye.html). Without a lens, he was able to see light well into the 300 nm range, whereas the rest of us can only see electromagnetic waves from 400 to 700 (roughly) nm. In fact, this paper (http://neuronresearch.net/vision/files/tetrachromat.htm) argues that humans are more accurately labeled as blocked tetrachromats instead of tri-chromats.
Hue is generally the term used to describe the thing that is different between one color and another. Or, the amount by which a color is similar to either red or green, or blue or yellow.
Saturation/Chroma generally refers to vividness, or a lack of dustiness to a color. In other words, this means how ‘pure’ a color is perceived to be. However, chroma is sometimes confused with lightness. I like the word chroma better than saturation, so that is the one I will be using.
Lightness refers to the amount of white light added to the hue/chroma mixture, i.e. the difference between light blue and dark blue.
The difference between lightness and chroma isnot always obvious, and many people confuse the two. Try it- can you think of a vivid brown or a vivid very light blue? I would bet you will never describe colors like that. In fact, we don’t quite know the exact psychological dimensions of color- instead we just try and describe the physical ones. People make up new qualities of color all the time (tint/tone/shade/brightness/illumination etc.), and there are nuances to each one, but the qualities described above shall do for now.
It is estimated that the average human eye can register around 10 million distinct colors- finding a way to organize the colors we can see is non-trivial. In fact, there are at least 10 ways to do it, none of them perfect. This stems from the fact that we don’t really have any idea as to how color actually works in the brain, so we are left to guess and infer.
Originally, colors were organized in color wheels. Goethe and Newton debated as to the nature of color, and both had their own ideas.
For more, check out this excellent post: http://imprint.printmag.com/color/the-wondrous-color-wheel-part-1/
[I will be doing a history of color and related research sequence later on].
However, the only real way to make sense of color is to use three dimensions and make a color space, instead of using two. One of the (seemingly) simplest ways to organize color is to use varying amounts of red, green and blue light. Due to metamers, all one needs are three sufficiently spaced out color primaries to make most of the colors that the human eye can detect. This is the way your computer does it, with red green and blue each represented at one of 256 different levels. If you were to model this in three dimensions, you would get a 256 by 256 by 256 cube. Which roughly looks like:
If you squish the cube into a cone or two and go from rectangular coordinates to cylindrical coordinates, you get the Hue, Saturation and Lightness model. Which looks like:
HSL has the advantage of trying to separate out the different physical properties of colors, but as mentioned above, this isn’t exactly how color works in the brain. Both RGB and HSL are rather non-intuitive and not perceptually grounded. For example, one cannot figure out what color would be ‘in the middle’ or the average of two colors using an RGB/HSL representation. In other words, Euclidian distance (aka ‘distance) doesn’t match up well with perceptual distance.
To overcome these problems, the Commission internationale de l’éclairage or the International Commission on Illumination (world authority on color spaces) came up with a number of color spaces. The most widely used is called CIELa*b*, where physical distance does match up with perceptual distance, at least for short distances. The coordinates of La*b* are in terms of lightness (l) and then the red v. green opponent process (a*) and the blue vs. yellow one (b*). Another great thing about La*b* space is that it is device independent, so that, in theory, you can actually be certain that the colors I see on my screen match the colors you see on yours. Due to these properties, it is widely used in the scientific community. It is however, complicated, to try and use this color space. Accurately converting from RGB to LA*B* space requires two different conversions and a whole lot of matrix multiplication. Currently you can look online to do the conversion, but no place that I’ve seen offers a way of taking care of all the variables. The folks over at easyrgb.com go most of the way, but without a method to specify primaries and gamma, the whole point of doing a conversion is pointless. I will have up on another page and implemented sometime in the next month.
Finally, there is one more color space worth mentioning. (I saved the best for last, you know). The Munsell color space is composed of a series of color chips that are perceptually even and uniform. In theory, the difference between one chip and an adjacent chip is supposed to be the same in any direction. The outer skin of the Munsell color solid has been used in a wide variety of color research, including the World Color Survey, the Mesoamerican Color Survey and some of the stuff I work on. (there will be much more about everything in the previous sentence in future posts). Here is a computer representation of the color chips:
Note that the colors are neither exactly accurate nor especially vivid, which isn’t how they look in real life.
The interesting thing is that the Munsell color solid is not a perfect shape, like a sphere or a cube or two cones. Instead, there are interesting bulges – in particular, there is a large bulge out at yellow, and a smaller one at red. I made this plot of the Munsell color chips plotted in CieLa*b* space (i’m still trying to work out how to animate it):
So, there you have some color spaces- If you want to know more, check out the following helpful sites:
And as always, feel free to reply with questions or corrections.