How Colours Relate
A colour on its own carries no feeling. Colour acquires meaning by comparison with another colour.
The previous chapter dealt with the three dimensions of a single colour. But a photograph almost never contains a single colour — what you see is always one colour with another colour beside it.
So the question on location is this one:
How many colours lead here, and on the wheel are they neighbours or opposites?
Lay the hues around a circle: red, orange, yellow, green, cyan, blue, violet, back to red. The relationship between two colours is their distance on that circle. Close is analogous, furthest apart is complementary, zero is monochromatic. This chapter is about what each of those three distances can do and how each of them fails.
Why opposite colours reinforce each other
The usual explanation is that complementary colours “go together”. That explains nothing, and it has the causality backwards. They are not opposite because they suit each other; they suit each other because the visual system encodes them on one shared channel, which is what puts them opposite.
Three cone types turn light into three numbers, as the last chapter described. Those numbers do not travel to the brain untouched. In the retina they are immediately re-encoded into opponent signals: red minus green, blue minus yellow, light minus dark. What the nerve carries is not “how much red” but “how much more red than green”.
A single line can only point one way at a time, so one location cannot be red and green at once. You can picture a green that leans blue, or a red that leans yellow; nobody can picture a greenish red. That is not a gap in vocabulary — the signal does not physiologically exist.
Two consequences that photographers use directly:
Simultaneous contrast. A colour pushes its surroundings towards its own opposite. One grey patch beside orange looks bluish; the same grey beside blue looks warm. So when complementary colours meet, each shoves the other further away — the contrast at the boundary is amplified by the visual system, not recorded by the camera.
Afterimage. Stare at a saturated red for twenty seconds and then at a white wall, and a cyan patch floats up. That is the red-green channel rebounding after fatigue, and it is the easiest way to verify for yourself exactly which square on the wheel counts as the opposite of a given colour.
Simultaneous contrast has a consequence for how you edit, not just how you shoot. If you grade a photograph against a coloured desktop background, that surround is pushing every judgement you make in the opposite direction, and you will correct against a bias that lives outside the picture. It is the reason editing applications default to neutral grey chrome, and the reason gallery walls are so often off-white — the frame around a picture is part of the picture’s colour.
One clarification while we are here: there are two colour wheels. The pigment wheel of the school art room pairs red with green and blue with orange. The light wheel pairs red with cyan, green with magenta, blue with yellow. Photography, screens and white balance all run on the second one, which is the one this site means. It is also why the shift axes in your camera are labelled red↔cyan and blue↔yellow.
Complementary: opposite sides
Orange and blue is the most used pairing, for three reasons rather than one:
- Opposite hues, so maximum hue contrast.
- Different luminance too — orange sits near 0.48 in relative luminance and blue near 0.07, a factor of seven — so the contrast survives conversion to black and white.
- It maps onto the commonest real situation: warm people and lamps, cool sky and shadow.
The third reason is what separates it from red-and-green. Red and green are nearly as far apart on the wheel, but they correspond to no common narrative, and their luminances are too close, so the pairing reads as noisy rather than strong.
Area matters more than hue
The most useful sentence in this chapter: two complementary colours at equal area will fight.
The cause is the opponent channel again. At equal area, both sides push equally hard along one axis, nothing wins, and the boundary acquires an unstable shimmer — which is precisely why fluorescent signage and cheap advertising feel abrasive. Make one large and one small, and the small patch is driven to maximum intensity by the field around it, becoming the place the whole photograph lands.
Roughly:
| Ratio of areas | How it reads | What it suits |
|---|---|---|
| About 1:1 | Confrontation, tension, sometimes glare | Deliberate conflict; graphic design more than photography |
| About 1:3 | Balanced contrast | Two subjects, a person and their setting |
| About 1:9 or wider | One landing point, everything else is field | Street work, a figure in a landscape |
Those two are deliberately paired because together they rule out an easy wrong conclusion — that the device needs a dark background. The variable is the gap in saturation and area, not the brightness of the ground.
In practice this is a street photography staple: find a grey wall, a dark alley or a bank of fog, and wait for someone in red. Note that the order of operations is the reverse of what most people do — choose the field first, then wait for the point. Raising the camera because you spotted a red coat leaves the background entirely to luck.
Analogous: neighbours on the wheel
Combinations like yellow–orange–red or blue–cyan–green produce harmony, because they share underlying pigment or spectral content. The whole photograph is unified inside one family and reads calm and coherent.
The pair is here to dismantle a common misreading: analogous is not “one colour at different depths”. The canyon runs from red to gold, the moss from yellow-green to brown, and both cross a substantial arc. The family is wider than most people assume.
The failure mode of an analogous scene is flatness — everything in one key with nothing to land on. The cure is usually not colour, since you have already committed to one stretch of the wheel. It is luminance. The canyon relies on the shaft of light pouring in from above; the moss forest relies on bright green behind and dark green in front.
Monochromatic: one hue only
A single hue, carrying its layering through saturation and luminance alone. This is the most extreme and most easily unified approach — the yellow wall in the previous chapter — and it rarely fails outright, because there is no second colour available to argue with it.
One warning about how it fails, since it fails quietly rather than loudly: a monochrome frame that is not working does not look wrong, it looks empty, and the usual instinct is to blame the subject rather than the treatment.
Fog, snow, night, underwater and dust all push a scene towards monochrome without help. This is part of why bad weather is worth going out in: it collapses the colour variable on your behalf.
The test for a monochrome frame differs from the other two. Complementary work leans on tension and analogous work on family resemblance; monochrome has neither, so it stands entirely on luminance and texture. Use the chapter ten method: shrink the picture to thumbnail size, and if subject and layering still read, it holds. If they do not, saturation will not save it, because the fault is in the skeleton.
Which leads to an honest test worth applying: if a monochrome photograph loses nothing when converted to black and white, the colour was never working. It was a tint, not a subject.
One more axis: warm and cool
Distance on the wheel is not the only relationship available. The other one — as important and much less often stated — is warm against cool, and what it governs is depth.
Warm colours advance in a frame; cool colours recede. There are two mechanisms. The physical one is that the atmosphere scatters short wavelengths more strongly, so distant hills and distant buildings are natively bluer and paler. That is aerial perspective, one of the oldest depth cues the brain uses, and painters have exploited it since Leonardo wrote it down. The optical one is that the eye refracts wavelengths differently, focusing red slightly behind blue, so red and blue on one flat plane acquire a faint stereo separation.
The practical consequence is direct: to separate a subject from its background, make the subject warm and the background cool — it is cheaper than any compositional device. It also explains why all three accents above — the orange sari, the yellow raincoat, the warm walls of the mountain town — are warm marks in cool or neutral fields rather than the reverse.
The inversion works too. A cool subject in a warm field reads as receding, distant, isolated. When that is the feeling you are after, it is the only thing that produces it.
Six common symptoms and their causes
Always: the picture shows X → the usual cause is Y → change Z.
- Both colours shout and the frame has no centre → complementary colours at similar area with no hierarchy → move to change the ratio until one is below a tenth of the frame, or step closer so the larger one fills the background.
- A red coat against green trees looks noisy rather than handsome → pure red and pure green actually separate in luminance, but a real coat and real leaves are not pure; once each is mixed with grey their luminances land close together, leaving hue to do all the work alone → change angle to drop the background into shade and let luminance share the load.
- An analogous landscape photographs flat → no luminance landing point, everything in one key → wait for raking or side light, or find a dark foreground; do not try to fix it with saturation.
- You wanted a small patch of red and got a large one → you stood too close, or used too long a lens → step back, go wider, let the surroundings regain their area. The device works by proportion, not by the colour itself.
- A colour pairing looks strong on screen and weak in print → simultaneous contrast is added by the visual system, and print gamut compression removes the saturated end first → widen the area ratio at capture rather than relying on saturation to carry it.
- The whole frame is warm and looks like a filter was applied → the analogous run has no cool reference anywhere, so the viewer cannot tell scene from post-production → keep one small neutral or cool thing in frame, even a corner of sky, a patch of shadow, a white wall. Its job is not beauty; its job is to be a witness.
The rule was discovered in a tapestry works
Michel Eugène Chevreul (1786–1889) was a French chemist who in 1824 took charge of the dye house at the Gobelins royal tapestry works. Complaints arrived steadily: certain black wools, once woven into a tapestry, looked insufficiently black — greyish, weak.
He assumed a dye problem and had batch after batch analysed. The dyes were fine. The same black thread was pure black in isolation, looked brownish woven beside blue, and greenish woven beside red.
The fault was not in the pigment. It was in the neighbours.
In 1839 he published the law of simultaneous contrast of colours: every colour pushes the region beside it towards its own complement. He had no access to the opponent-channel mechanism — that was a century off — yet from a stream of customer complaints he described its behaviour exactly.
The book escaped the tapestry works entirely. Delacroix read it, the Impressionists read it, and Seurat copied out its passages on painting — he also met the same ideas through Charles Blanc and Ogden Rood — then pushed them to their conclusion: not mixing blue and yellow into green on the palette, but placing pure blue and pure yellow dots side by side and letting the mixing happen in the viewer’s eye. That is Chevreul’s mechanism run in reverse.
A photographer has no palette. What a photographer has is position. The only thing you get to decide is which two colours end up adjacent in the frame — which is what this entire chapter has been about.
Why orange and blue became the default
There is a much more recent history as well. Around 2000, film post-production moved from cutting negative to the digital intermediate, and for the first time every frame of a feature could have its hues pushed independently. The cheapest move available to a colourist was to push skin towards orange and everything else towards teal.
The reasoning is the same as everything above: skin already sits in the orange region, so flooding the background with its complement lifts people off it without touching a single light. Between roughly 2003 and 2015 the move was used to exhaustion, which is why “teal and orange” now sounds like an accusation.
It became the default not through fashion but because the three reasons above could, for the first time, be manufactured in post — there was no longer any need for the scene to actually contain warm light and cool shade. This is the one place in this chapter where a colour relationship can be decided after the fact, and the cost is legible: a relationship that did not grow out of the scene eventually reads as a technique rather than a place. Knowing why it works is what lets you decide when not to use it.
A table for the field
| What you see | Which relationship it is | What to do |
|---|---|---|
| Two colours, opposite, similar area | Complementary but fighting | Move until one drops below a tenth of the frame |
| A large pale field, one small saturated patch | Landing-point structure | Change nothing; wait for the patch to reach the right place |
| One hue family throughout | Analogous | Find a luminance landing point: side light, dark foreground, a shaft |
| One hue family and still flat | Analogous, failed | Come back at another hour, or consider black and white (chapter ten) |
| Only one colour | Monochromatic | Carry it on texture and luminance; move closer and let it fill |
| More colours than you can count | No relationship at all | Back to the previous chapter: reduce the number before discussing relationships |
| Subject glued to background | Warm and cool not separated | Drop the background into shade — shadow is natively blue — or wait for the subject to walk into warm light |
| You want isolation and distance | Warm and cool inverted | Cool subject, warm surroundings, and the subject recedes on its own |