June 21, 2021

I, Science

The science magazine of Imperial College

Dave Warrell explains how our eyes are able to decipher colour

Dave Warrell
5th March 2021

I’m currently having a debate with my girlfriend about whether or not the chair she wants to buy is more blue or more green. We both agree it’s a sort of teal, but it seems teal is something ever so slightly different to us. And this gets me thinking: is she wrong, or do colours look different to different people?

In some senses, it’s a question that is borderline impossible to answer – just try describing a colour without comparing it to another colour. I can never know whether my blue is the exact same as your blue, just in the same way I can never understand whether your feeling of happiness is the same as mine. It’s a question that wanders into the territory of philosophy, away from the firm ground that I, Science is familiar with.

But, clawing back onto ground that I feel a bit sturdier on, there are some things that physics, biology and visual neuroscience can tell us about the answer to this question.

What is colour?

An important thing to cover off first is what determines a colour from a physical perspective. Put simply, the colour that we perceive something to be is determined by the light which comes from it.

More specifically, the colours that we see correspond to the amount of energy belonging to the photon – the tiny particle that makes up every type of radiation on the electromagnetic spectrum, including light – travelling from an object and hitting our eye.

The range of photons that our eyes are able to detect is actually vanishingly small compared to all of the photons that are flying around the universe. Photons with between 1.65-3.29 electron volts (eV) of energy can be detected by our eyes, so our eyes can manage a range of roughly 2eV. For reference, the range of energies that photons can have throughout the entire electromagnetic spectrum is roughly 1020eV. Written out in full, that is a range of 100,000,000,000,000,000,000eV compared to our piddly 2eV.

The entire visible spectrum and all the colours that we see exist within this 2eV window. Red is the bottom end of that window (1.65eV), while violet is what we see at the upper end (3.29eV).

So, in any given situation, the same scene is firing photons with the same energies into our eyes. This physical process is the same for all of us.

But photon energy isn’t all that defines colour. Without a visual system, colour is just a small segment of light, itself a tiny sliver of the electromagnetic spectrum. We all know what we mean when we say red (at least in relation to other colours), but that does not mean that red is a physical thing. Not like, say, a chair is.

Colour is really just a layer of information added by our brains to the physical scene in front of us. It essentially allows us to differentiate between objects based on which photons bounce off them, to helps us navigate our environment.

To answer whether we see the same colours, then, we have to understand our visual systems and how these could differ between people.

How is colour made?

Light is focussed onto your retina, the part of your eye with all the processing power, which houses your cone cells. Cone cells are able to react to light, a property known as photosensitivity. Cone cells absorb a range of photons, and each has a ‘favourite’ energy of photon.

When cone cells are hit by the photons in their range, they switch on, which sends a signal to the brain. Importantly, cone cells don’t just have two modes: ‘on’ or ‘off’. Instead, each cone cell can respond with 100 different intensities depending on how close the photon is to having the cone cell’s preferred energy. The most intense response comes from their favourite photons, and the intensity of the cone cell’s response then peters out either side.

I find it helpful to think about them like a dimmer switch, or a volume knob, with a twist: while a normal dimmer switch would be completely off if turned all the way to the left and be at maximum output when turned all the way to the right, this dimmer switch is completely off when turned to all the way to the left or right, but at maximum output when bang in the middle.

Because of this property, cone cells alone are unable to decipher colour on their own. Think of it this way: the light bulb controlled by our dimmer switch reaches peak brightness when the switch is right in the middle. If you saw a lightbulb at 40% of its maximum brightness, you could never tell which side of the dimmer switch the knob had been turned to if you couldn’t see the switch. The brain has the exact same problem: it doesn’t know how much energy the photon has, so can’t translate what the intensity of a cone cell’s response means.

graph depicting a normal distribution of photon energy vs cone response intensity
The problem with one cone. Either side of the peak two different photons will elicit the same response, which the brain can’t decipher.
(Graph: I, Science. Colour Gradient: Wikimedia Commons)

That is why monochromats, animals (or very rarely, people) with one type of cone cell, see in greyscale – there is no way of working out what quantity of energy a photon has, so it is impossible to add colour to a visual scene.

Can dogs see colour?

Adding another cone cell allows the brain to add some colour. This type of colour vision is what most mammals (including dogs) have. One type of cone prefers photons from one end of the spectrum (blue) and the other type prefers those from the other end (yellow). Because the signals from the two types of cone cells can now be compared, the brain is able to add a layer of information to the visual scene. It knows whether the photons have more or less energy and fills that in as colour; things appear either more blue or more yellow. The ability of the brain to compare the information coming from cone cells that it ‘knows’ prefer different photons is what allows us to see colour.

Graph showing both cones' relative cone response intensity
Two cones are better than one.
(Graph: I, Science. Colour Gradient: Wikimedia Commons)

Where monochromats can see around 100 colours (which are just shades of grey), dichromats can see something in the order of 10,000. That’s because for any one of the 100 responses a ‘blue’ cone can have, a ‘yellow’ cone can have 100 responses – meaning we can multiply the number of possibilities together. For any animal with an advanced visual system, the number of colours they can see is roughly 100x, where x is the number of different types of cone cells they have. (Incidentally, the lack of higher processing power is why the dodecachromatic mantis shrimp – who have twelve different types of cones – cannot see 10012 colours.)

What about humans?

That brings us back to human. We have three sets of cones: S-cones favour roughly what we see as blue light (which corresponds to photons with around 2.6eV of energy), for M-cones it’s green (2.2eV), and for L-cones it’s yellow (2.1eV).

Together, they allow our brain to compare three different inputs. As a result, we can see something in the order of 1003, or one million, colours. Let’s use orange light as an example to explore how this works exactly.

An orange light is shone into our eyes (please don’t try this at home) and hits each different type of cone cell. The S-cone is not interested at all and remains off. The M-cone reacts a little, while the L-cone gets quite excited. Because all of the cones have reacted differently compared to one another (as they would for any wavelength of light), the brain can compare the relative response of the cones. A 40% response from the M-cones and a 60% response from the L-cones tell it that we must be looking either at something green or orange. But a 0% response from the S-cones means it can’t be green, so it must be orange. A similar process happens for every wavelength of light.

graph showing three cones' response intensities
Humans have three sets of cones.
(Graph: I, Science. Colour Gradient: Wikimedia Commons)

Do we all see the same colours?

Applying this, we can say fairly confidently what people lacking a certain type of cone see in relation to us. They see fewer colours. And if we know which type of cone they are missing, we can tell what colours they see – either their vision skews towards blue and yellow or cyan and red, depending on which cone is missing. Below is an example of the same colour spectrum from a monochromatic, dichromatic and trichromatic perspective.

four different colour gradients
Top to bottom: monochromatic, dichromatic (red-green colour blind), dichromatic (blue-yellow colour blindness), trichromatic.
Made using Coblis. Original image sourced from Wikimedia Commons.

But trichromats can also see slight variations in colour compared to one another. For example, some people may have a greater density of one type of cone cell than others, and so might see certain colours slightly more intensely than others. It’s also possible that mutations in genes important to cone cells can slightly alter the photon energy that cone cells prefer, and as a result some people might see some colours slightly differently to others. Although this is a bit trickier to get your head around, let’s try: think about teal. It’s bluey green. If someone had a mutation to an M-cone, making it favour photons with slightly higher energy (‘bluer’) than normal, it’s sort of possible to see how people with that mutation might see teal as more blue than you.

What is impossible to grasp is what people with an additional cone, tetrachromats, can see. Tetrachromats are, in theory, able to see 100 million colours. Many animals are tetrachromats, including goldfish and chickens. We are as colour blind to them as dogs are to us. And it is impossible for us to imagine what they see – because our visual systems lack the capability to add that layer of information, we simply can never see the richness of colours that many animals can.

That’s not strictly true for all of us, because there are tetrachromatic people, too. It is incredibly rare, and something that only affects women. The crucial genes for M-and L-cones are located on the X-chromosome, so women have 4 copies of these genes. In a very small percentage of them, a mutation exists in one of those 4 which gives cone cells a unique preference for photons.

Tetrachromats, we have learned, are able to see vibrant colours in what we see as dull monotones, similar to how dichromats see a dull yellow where we see orange. For every colour we see, at least in theory, they can see an array of 100 colours. It certainly sounds like we’re missing out on a lot, but it’s not all bad news – tetrachromats can sometimes find things a bit overwhelming. One, who is an art teacher of all things, said that her favourite colour is white because it’s a break for her eyes.

So, to return to the question at hand: our visual systems are all based on the same principles. The photons are the same, and we all have the genes to be able to see the full spectrum of visible light. About 8% of males have a mutant gene that’s faulty, making them dichromats. And a very small percentage of females have a different mutant gene that makes them tetrachromats. Beyond that, tiny variations in our visual systems mean that, although the colours we see are likely to be very similar, they are unlikely to be the exact same.

So, the odds are, as a result of minor variations in our visual systems, that my girlfriend and I do see the chair as a slightly different colour.

I still think it’s more blue.


Dave is a science communicator with a background in biomedical sciences. He loves making the everyday interesting and thinking about how science can work for everyone.