Contrast: The Volume of a Picture — The Grammar of Light

Contrast is difference. The further apart two things in a frame are, the louder the relationship between them.

So the question on location is not whether the picture is bright enough. It is:

What is my subject different from its surroundings in — and is that difference big enough?

When a photograph feels like it has no point, the cause is usually not the wrong subject. It is that the subject is not different enough from what is around it. And the most effective fix for that is rarely a slider at home. It is two steps sideways at the time.

Contrast is a relationship, so it is about where you stand

An expensive misconception first: that “add contrast” means “push the contrast slider right”.

Those are not the same operation. The slider works on the tone curve of the whole picture — brights pushed brighter, darks pushed darker. The frame gets more emphatic and the price is detail crushed at both ends. But if your subject was already the same brightness as its background, the slider moves them together. The relationship does not change and the subject still will not separate.

What actually decides whether a subject stands out is how far it differs from what is immediately behind it, and that is almost entirely a function of where you are standing:

Crouch half a metre and the dark hedge behind your subject is replaced by bright sky.
Move two steps left and the white wall behind a face becomes a shaded arcade.
Wait thirty seconds for the cloud to move and the subject is lit while the background is still in shadow.

None of those touches a setting, and all three change the relationship. A slider cannot change a relationship. It can only amplify or reduce the one you already photographed.

The eye measures ratios

There is physiology under this, and you can watch it happen.

The visual system does not measure absolute brightness. It measures the ratio between adjacent regions. Put the same mid-grey patch on a black ground and on a white ground: the two patches are numerically identical in the file, and you will plainly see the one on black as lighter. That is simultaneous contrast, and the lateral inhibition from the previous part is its mechanism — cells on either side of a boundary suppress one another, so differences get exaggerated.

Edward Adelson’s checkerboard is the most ruthless demonstration. Two squares on a chequered surface, one in a cast shadow, are printed in exactly the same grey — and they remain impossible to see as the same, because your visual system is not reporting the light arriving from them, it is reporting its best guess at what colour the surfaces are. Perception is already an interpretation before you get to make one.

There is a second layer: the eye responds to brightness roughly logarithmically, not linearly, an observation formalised as the Weber–Fechner relationship. Going from one unit of light to two feels like the same step as going from a hundred to two hundred — both are “a stop brighter”. That is why photography counts exposure in stops (doublings) rather than in absolute numbers, and why lifting deep shadows by a small amount produces a far larger visible change than lifting highlights by the same amount.

Together they give a very practical corollary: next to something dark, a little brightness is plenty. That is what lets a low-key photograph carry the whole frame on one small bright area. You do not need the subject lit hard. You need its surroundings dark enough.

Contrast is not only tonal

Most people think only about light and dark, but several kinds are available and they stack. The right-hand column is the point of the table — each one has a corresponding action on location.

Kind of contrast	Example	How to make it happen
Tonal	A bright face on a dark background	Change the background by moving yourself, or wait for the light
Colour	An orange boat on a blue sea	Find a complementary backdrop, or wait for the subject to walk into light of another colour temperature
Sharpness	A sharp person in a blurred crowd	Open the aperture, use a longer lens, or move closer
Scale	A huge mountain, a tiny figure	Step back and put a person in the frame, or go wide and close to a foreground
Texture	Smooth glass against rough brick	Stand where two materials meet
Density	An empty left half, a crowded right half	Wait for the subject to reach the empty side

The tonal contrast here is actually weak — the orange hull and the blue water are close in brightness. What holds the picture up is colour: orange and blue sit opposite each other on the wheel, the strongest colour pairing there is. Convert this to black and white and it collapses, because both become much the same grey. Richard West, 2023 · CC BY-SA 2.0 · Source

Which carries a warning: colour contrast does not automatically become tonal contrast. If you plan to convert to black and white later, a pairing like orange and blue disappears entirely. Chapter ten returns to this.

Stacking multiplies, conflicting splits

Several kinds of contrast can run at once, but they have to point at the same subject.

Pointed the same way the effect multiplies. Something bright, sharp and large beside things dark, blurred and small, and a viewer knows where to look within a second. That is the standard portrait recipe — a wide aperture for sharpness contrast, a dark background for tonal contrast, moving closer for scale contrast, all three pushing the same person.

The trouble starts when they point at different things. A small bright object on the left, a large dark mass filling the right: each wins on one axis, the viewer’s eye shuttles between them and then gives up. It is the competing-fixations problem from the previous part seen from another side: contrast is not something you want more of, it is something you want all of in one place.

The check on location: before firing, ask whether the brightest thing, the sharpest thing and the largest thing are the same thing. If they are not, the photograph has not yet decided who it is about.

The photograph that proved the camera is not the eye

1927, Adams below Half Dome. His eye saw a bright sky and his first negative recorded a grey one, which was not what he wanted. So he fitted a deep red filter: red absorbs blue, and on black-and-white film a blue sky goes nearly black. He later described it as the first time he had seen the finished print before releasing the shutter. Ansel Adams, 1927 · Public domain · Source

This is among the most quoted photographs ever made, because it states one thing very plainly: a photograph is not what the eye saw, it is what you decided it would be.

And what it demonstrates is exactly this chapter’s opening claim. Adams did not make the whole picture more contrasty. He changed the relationship between sky and rock by pushing one of them down on its own. The digital equivalents are the filter options in a monochrome mode (Fujifilm’s ACROS+R and Monochrome+R), channel mixing, and per-hue luminance in HSL. The principle is identical: pick one colour, move only that.

When the contrast is too large to record

A seascape from the 1850s. No emulsion of the period could hold a bright sky and a dark sea at once — so Le Gray used two negatives, one exposed for each, and combined them on a single sheet of paper. The earliest HDR, by a century and a half. Gustave Le Gray, 1855 · Public domain · Source

Some numbers first, because they explain why you so often have to choose. A sunny scene containing both direct sun and shadow routinely spans more than fourteen stops — a ratio above twenty thousand to one. A modern sensor holds roughly twelve to fourteen in raw, and considerably less in a straight JPEG. Your eye copes only by cheating continuously: the pupil opens and closes, and the retina adapts locally to whatever small patch you are fixating, so you never actually see all of it at one instant. A camera gets one exposure and has no such trick.

That is why the trade-off cannot be avoided. When a scene exceeds what the sensor can record, the options are:

Give one end up — usually protect the highlights and let the shadows go down. The most common answer and frequently the best-looking one.
Compress the highlights — which is precisely what Fujifilm’s DR200 and DR400 do; the Fujifilm part covers it properly.
Add light — a reflector or a flash lifts the shadows and narrows the gap.
Wait — for cloud, or for a lower sun. Contrast is not a fixed property of a place.
Combine — bracket and merge. Le Gray’s method, modernised.

The commonest beginner error is forcing the shadows up, which turns the whole frame grey and drags every bit of noise into view. Letting dark things stay dark is a decision, not a failure. A blown highlight is pure white with no data to recover; a blocked shadow is at least black, and black reads as deliberate.

Global contrast versus local contrast

These get conflated, but they operate on different parts of the picture:

Global contrast is the distance between the brightest and darkest points. Increasing it makes the frame more emphatic at the cost of detail at both ends.
Local contrast is the difference between neighbouring regions — what a clarity control does. Increasing it brings mid-tone detail forward and makes things look more solid and more “sharp”.

The global contrast here is enormous — cloud near white, forest near black — and the mid-tones are nonetheless full of gradation. High contrast and rich tonality are not opposites. The difficult part is never raising contrast; it is raising it without losing the middle. Ansel Adams, 1937 · Public domain · Source

Local contrast has a different cost from global, and a harder one to notice: it leaves halos on either side of a strong edge.

A frame filled with bare twigs against an even sky is where clarity gives itself away fastest — thousands of hard boundaries packed together. Raise it and every twig picks up a rim brighter than the sky behind it, so the whole canopy appears to glow from a distance. Frames like this want clarity pulled back, not pushed up. Yitzilitt, 2021 · CC BY-SA 4.0 · Source

The test is therefore simple: look for large numbers of fine, hard edges. Branches against sky, backlit hair, a building edge against clean cloud — all of them want restraint. Large, low-frequency subjects (rock faces, cloud, leather, old plaster) will absorb a great deal more.

Low contrast is not a deficiency

No black anywhere and no white anywhere; the entire range is squeezed into a short stretch of the upper middle. It is not short of contrast at all — fog has converted distance into tone, near trees darker and far ones paler, and that narrow band of greys is carrying every layer. The narrowness is the content. Dietmar Rabich, 2020 · CC BY-SA 4.0 · Source

Deliberately low contrast — Fujifilm’s ETERNA, a faded curve in processing — produces something quiet, hazy, memory-like. Fog, rain and early morning are low contrast to begin with, and forcing them up loses precisely the quality that made you stop.

Worth noticing that a low-contrast photograph is not one without separation. It is one that achieves separation with very small differences. The brightness gap between successive ranks of trees above may be a few per cent, and because the visual system measures ratios and amplifies boundaries anyway, a few per cent is enough. It also explains why these frames are so fragile in processing: push a little too far and those fine steps blow apart into hard slabs.

Distance lowers contrast by itself

Low contrast is not only weather. Distance does it too. The air between you and a subject scatters light and lays it over the subject, so the further away something is, the lower its contrast, the paler its colour and the bluer its cast. Painters have used this aerial perspective since the Renaissance.

Three consequences follow directly:

Contrast is a depth cue. Within one picture, high-contrast things read as near and low-contrast things as far, even at the same size. So lowering the contrast of a background pushes the subject forward.
Raising the contrast of distant things makes a photograph look false. Pull a far ridge until it is as crisp as the foreground and the brain receives contradictory depth cues; that is most of what people mean by the HDR look. Distant material is safer desaturated and lightened than hardened.
It is also what fog is. The frame above is aerial perspective taken to its limit — enough scattering to wash out trees a few metres away.

The question stays the same one: do you want this photograph to speak loudly or quietly?

Six symptoms and their causes

The format throughout: the picture shows X → the usual cause is Y → change Z.

The subject will not separate however much contrast you add. → It is too close in brightness to what is immediately behind it, and the slider moves both together. → Change the background: crouch, step sideways, or wait for the subject to move onto a different tone.
After adding contrast the picture looks dirty and the shadows go blotchy. → Global contrast has crushed shadow gradation that was thin to begin with. → Add less, or use local contrast to lift only the mid-tones.
Twigs, hair or building edges have picked up a bright rim. → Clarity too high, producing halos at hard boundaries. → Pull it back; subjects like these have a low ceiling.
An overcast frame stays flat whatever you do. → The scene’s range genuinely is narrow, and you are treating a quiet picture by pulling its ends apart. → Accept that it is low contrast and build the difference from colour, shape or density instead.
The sky is dead white while the ground is still dark. → The scene exceeds the sensor and metering favoured the ground. → Protect the highlights with negative compensation and lift the shadows, or wait, add light, or bracket and merge.
A black-and-white conversion turns into a slab of mid grey. → The contrast was built from hue and hue does not survive the conversion. → Find the difference in luminance instead, or use a filter (+R, +Ye) to push one colour down on its own.

In the darkroom, contrast was a kind of paper

Contrast arrives today as a slider, which makes it easy to assume it is a modern adjustable. It has always been adjustable. It used to be physical.

Black-and-white enlarging paper came in grades, 0 to 5. Grade 0 is soft and compresses the negative’s range into a gentle band of greys; grade 5 is hard, and the same negative prints as emphatic black and white. The stack of paper beside the enlarger was the contrast slider. Variable-contrast paper made it more direct still — one paper, and a filter from yellow-green to magenta to choose the grade.

That is worth remembering because it exposes an order of operations people often miss: the negative records the range, the paper decides the contrast. The negative’s job is to get as much of the scene’s span inside it as possible; deciding how far to spread that span is the print’s job. Digital inherited the division exactly — the raw file is the negative and the curve in your processor is the stack of paper. So “protect both ends when shooting and decide contrast later” is not a modern technique. It has been the darkroom’s default workflow for a century.

There is a further reason the vocabulary still fits. Hurter and Driffield began measuring, in the 1890s, exactly how much density a film returns for a given exposure, and plotted it — the characteristic curve, with its toe in the shadows, its straight section through the mid-tones and its shoulder in the highlights. The slope of that straight section is contrast, and they called it gamma. Every tone curve you have dragged in software is a redrawing of their graph.

Adams’s Zone System pushes the whole idea to its conclusion: decide before the exposure which print value each brightness in the scene should land on, then work backwards to the exposure, the development and the paper grade. The next chapter picks that up.

The decision table

What you see	The contrast problem is	What to do
Subject the same brightness as its background	Insufficient relationship, not insufficient strength	Move yourself to change the background; leave the slider alone
Blown sky over a black foreground	The scene exceeds the camera	Protect highlights; add light, wait, or bracket
Everything flat and grey	The scene’s range is genuinely narrow	Decide whether you want it quiet; if so, build difference from colour or shape
Flat mid-tones, ends are fine	What is missing is local contrast	Add some clarity, after checking for fine hard edges
Rimmed edges, looks artificial	Local contrast overdone	Pull back; this subject has a low ceiling
Strong colours that die in monochrome	The contrast lives in hue	Filter one colour down, or rebuild the difference in luminance

Further watching

Street Photography: Light & Dark Contrast
The Candid Frame · 9 min
Why this one The whole video is two photographs, and what it says about them is this chapter's opening line. At 1:28: "what's key here is the area immediately behind her." At 2:15 he goes further — had the cyclist pedalled a few feet on, her shape would have fallen against a door just as dark, and the picture would have been gone. At 5:58 he runs the same argument on a second photograph. He never touches a slider. From 7:15 it is a plug for his podcast.
Multigrade Filters - ILFORD Photo Darkroom Guides
ILFORD Photo · 4 min
Why this one At 1:24 one negative is printed at grade 0 and grade 5 and shown side by side, captioned on screen: "Nothing else has been changed apart from the actual filter itself." That is this chapter's point that the negative records the range and the paper decides the contrast — and it is the only place on this site where you can see the thing itself. Commissioned by ILFORD, which the presenter says at the top; the product happens to be this section's subject.