Perception of light and colors by the human eye

To the problem of having an objective reference to define the colors is added the fact that the definition of the color is intrinsically related to the perception that one has, and this perception obviously differs from one population to another, from one individual to another, and it’s also related to the environment in which one perceives the color.

Hint

In ancient history, Democritus even suggested that colors are only imagination.

Comprehending this subjective interpretation of color is important for understanding how color management in digital image processing was designed and how it is evolving.

The Eye

As soon as light rays are received by the eye and the retina, light is broken down and interpreted. Two types of photoreceptors make up the retina, cones and rods.

Rods are the useful and active cells in half-light; they reach total saturation as soon as light becomes too strong, around 500 photons per second.

Cones are the effective cells for reception of more intense lights; they activate only above an intensity of 10 photons per second. Divided into three different types, they’re able to capture rays over a wider range of frequencies; this division into three types also allows the hue of the light to be interpreted: the nervous system can compare the intensities on the three different types of cones and deduce a hue.

The nervous system

The information received by the cells of the retina are converted into nerve impulses and interpreted by the brain. It should also be noted that the response of the retina isn’t linear: a proportional variation in light intensity doesn’t translate into a proportional variation in nerve impulses.

This process explains in particular the fact that objectively different blends of monochromatic sources can be perceived in exactly the same way and subjectively identical; two white colors perceived in the same way and indistinguishable by an individual can therefore in reality be composed of different blends of monochromatic lights1.

Conversely, colors that are objectively identical can be perceived in completely different ways depending on their environment. This is particularly true for whites and grays; what we perceive as white color greatly varies depending on the types of light sources and the general hue of the environment. This is the equivalent of white balance in photography, but it’s true for all colors.

The nervous system is therefore responsible for an interpretation of the colors which isn’t directly related to the objective signal received by the eye, but which could be seen as a subjective ” adjustment ” in post-production: it adapts the image according to the environment, exactly as the photographer does by adjusting the contrasts and the white balance (the tints), to “normalize” the picture, and increase the quantity of perceivable details in the reconstructed image.

All these different steps change the perception of a physical and objective signal: the same light ray will not be “seen” in the same way in two different places, and according to the quantity and the nature of the other rays which accompany it.

Consequences

The consequences of this subjective perception are important to understand the historical choices that were made on the techniques allowing the artificial reproduction and processing of images (analog and digital).

On brightness and contrast

The presence of the rods on the retina and their performance in weaker lights, to the detriment of the perception of the colors in these lights, makes that the eye is globally more efficient to discern the details in the half-light than in strong lights, and that it’s in general more efficient to distinguish contrasts of intensity than variations of tint (the decomposition in precise wavelengths on the spectrum)

On hues

The division into three types of cones, each performing on specific wavelengths, means that three primary colors2 are sufficient to reproduce all the colors the human being can perceive, as long as these primaries correspond approximately to the performance range of these three types of cones3.

The three “primary” colors closest perception ranges of the cone cells are the combination of Red, Green, and Blue, even if in reality the cones rather perceive yellow, green and blue.

Note

Trichromy isn’t the only model that can represent color vision. For example, a model where the vision distinguishes the white-black, blue-yellow and red-green oppositions can be conceived4.

Hence this way of perceiving the colors influences the choices of systems to reproduce them artificially, as we’ll see later, in particular with the gamma, and the RGB or YUV system.

Sources & References

  1. There are different “qualities” of white for seemingly identical lights, depending on whether they’re composed of a more or less varied ranges of rays made of different wavelengths (or different tints). ↩

  2. There is a vocabulary confusion, in particular in the audio-visual production, between primary color (which cannot be obtained by blending other colors), and complementary colors (whose blending gives black, gray or white). We use here the word primary in the sense of digital color spaces: whose mixing gives white. ↩

  3. Isaac Newton had discovered that it’s quite possible to reproduce a white light with only two monochromatic colors but a third primary is necessary to depart from the simple gradation between the first two. ↩

  4. Ewald Hering, a physiologist, was vigorously opposed to trichromy. Based on the psychological study of perception, his model, which Erwin Schrödinger showed the mathematical equivalence with trichromacy, has since been confirmed by neuroscience studies. His evidence stemmed from color-adaptation experiments and the linguistic observation that certain color names cannot be combined into one, like “yellowish blue” or “reddish green”. In this model, colors are perceived through mechanisms sensitive to three pairs of opponent colors: white-black; blue-yellow; green-red. ↩