Black and White: Removing Colour as a Variable

Beginners tend to assume black and white is more artistic, convert an unsatisfying colour photograph, and hope it improves. It usually does not.

The question worth asking — on location, or at the computer — is a different one:

Is colour helping this photograph, or getting in its way?

Converting only addresses the second case. If the trouble is loose composition, an unclear subject or flat light, removing the colour leaves all of it in place and makes it more obvious, because there is now nothing else to look at.

Converting is dimensional reduction, so it always costs something

Worth being precise about this, because it decides when the trade is worth making.

Every pixel of a colour image carries three independent numbers. Think of them as three dimensions: hue (which colour), saturation (how strongly) and lightness (how bright). After conversion each pixel has one number.

Three dimensions compressed into one must lose information, and nothing can prevent it. So whether to convert is not a matter of taste. It is a specific trade:

• If the two discarded dimensions were helping — colour is the subject, colour is separating things, colour is carrying the mood — converting is a pure loss.
• If they were interfering — colour is noisy, colour is flat, colour temperatures are mixed — converting deletes two dimensions that were generating noise, and the one that remains gets cleaner.

Which also explains why converting a bad photograph rarely helps: composition and lighting problems live in the lightness dimension, and that is the dimension you kept.

When to convert

From that principle, five situations where colour genuinely is in the way:

• Colour is noisy. Signs, umbrellas and cars in every hue, pulling attention everywhere at once.
• Colour is flat. Overcast days, grey walls, winter branches. Colour contributes nothing while still occupying two dimensions.
• Colour temperatures are mixed. Tungsten, fluorescent and window light in one room, where no white balance makes all of it right. Black and white deletes the problem outright.
• Shape and texture are the subject. Architectural geometry, wrinkles, bark, shadow patterns — all of which already live in lightness.
• You want to remove period cues. Black and white strips a photograph of its decade.

Conversely there are two cases where you almost certainly should not: colour is the subject (a sunset, a market, flowers), and subject and background are separated by colour alone — which is the next section.

Where the colour goes

Conversion combines each pixel’s RGB values into one grey value. The weights are neither arbitrary nor fixed — and the fact that they are not fixed is the source of all your control.

The default weights are roughly 30% red, 59% green, 11% blue. Why so much green? Because human daylight spectral sensitivity peaks near green, around 555 nanometres, and the cone type sensitive to middle wavelengths is the most numerous. In other words this is not an aesthetic choice, it is an attempt to model how bright a colour looks.

But weights can be changed:

Emphasise this channel	Effect on blue sky	Effect on skin
Red	Sky goes dark, clouds separate strongly	Skin lightens, blemishes fade
Yellow	Sky darkens moderately — the gentle version of red	Natural skin; the usual choice
Green	Sky lightens	Skin darkens and texture is emphasised; foliage lightens

On film this was done with a coloured filter on the lens. A filter passes its own colour and absorbs the complementary one, so a red filter lightens red objects and darkens a blue sky. Digitally you choose it in the camera or in processing. Fujifilm’s ACROS+Ye / +R / +G and Monochrome+Ye / +R / +G are exactly this.

Two colours, one grey

The cruel part of black and white: if subject and background are close in brightness, they merge.

A red coat on green grass looks strongly contrasted in colour. Converted, both may land on much the same mid-grey, and the person disappears into the field.

There are only two ways to handle this while shooting:

1. Look for the difference in brightness instead. Ask how many stops apart two things are, not whether their colours go together.
2. Use a filter to move one of them. In the orange-and-blue case, a red or yellow filter lightens the hull and darkens the water, and they separate immediately. It is precisely what the Half Dome frame is doing.

In practice there is a third and faster route: set the camera to black and white and shoot that way. The electronic viewfinders on the X-T5 and X-E4 show a monochrome image directly, so you see the tonal relationships on location. This is one of digital’s largest advantages over film — imagining which grey a colour would become used to take years of experience, and now you simply look.

(Shoot raw alongside and the colour information is still fully preserved, so you can change your mind at home. The approach costs nothing.)

What standing on tone alone looks like

This uses the same device as the Cameron portrait: establish one large difference in brightness, then let everything else fall in behind it. Cameron had white hair against a dark ground; Vachon had a wet street duplicating the highlights.

The four most reliable black-and-white subjects, then: something with a large tonal difference (backlight, white hair, a dark object against a pale wall), texture, shape and pattern, and scenes that had little colour to start with (fog, snow, overcast, rain). What they share is that their information already lived in lightness, so reducing the dimensions costs almost nothing.

Black and white is stricter about composition

One indirect consequence matters more than it sounds: a black-and-white frame has to be cleaner than a colour one.

The reason goes back to the reduction. In colour, a piece of clutter in a corner at much the same brightness as the subject may still read as a separate thing because its hue differs. Black and white has no such insurance — things of similar brightness fuse. The “check the four corners first” habit from the second part costs more here if you skip it.

There is a test for this needing no equipment: shrink the photograph to thumbnail size. A colour picture still leaves blocks of colour at that scale; a monochrome one leaves only the skeleton of its light and dark masses. If that skeleton still reads at thumbnail size — where the subject is, where the eye travels — the conversion works. If it does not, no amount of contrast or filtering will rescue it, because the fault is in the skeleton.

Exposure works differently in black and white

With two dimensions gone, more is being asked of the one left. Three things differ from shooting in colour:

One: you need a real black. A colour photograph can sit entirely in the mid-tones and survive on hue. A monochrome one cannot; without a true black, or something close to it, it will almost always look soft and washed. The previous chapter’s histogram gains an extra use here — confirm something actually reaches the left end. Not a large mass of it. A small patch is enough to give the picture a reference point.

Two: highlight texture is worth more than in colour. A small clipped patch of sky often goes unnoticed in a colour frame. In monochrome that patch of pure white is glaring, because it is the single brightest point and there is no colour to distract from it. Be more conservative with highlights.

Three: per-channel clipping stops being visible, but does not stop mattering. The red-channel clipping from the previous chapter will not show up as a colour shift in a monochrome output, but it still destroys detail — the texture in that red area was flattened at capture, and no channel mixing afterwards recovers it. Watch the RGB histogram even when the output will be black and white.

Grain is an asset, not a defect

Here is another place where monochrome inverts the colour rules. Colour noise is ugly — red, green and blue speckle corresponding to nothing in the world. Monochrome grain reads as texture, because it resembles the silver halide crystals of film, and viewers long ago learned to read it as “this is a photograph” rather than “this is a flaw”.

Two practical consequences: you can work a stop or two higher in ISO without feeling the loss; and a monochrome frame that looks too clean and too digital often becomes more convincing with a little grain added. Fujifilm’s Grain Effect and the grain modelled into ACROS exist for this, and the Fujifilm part covers the detail.

Black and white is not necessarily neutral

One more overlooked dimension: the white in a black-and-white print does not have to be neutral.

In the darkroom this was toning. Bathe a finished print in various solutions and some of the silver is replaced by other compounds, tinting the whole image — selenium towards purplish-red while deepening the blacks, sulphide towards brown (the sepia of old photographs), iron towards a cold blue. It was never only a style: selenium and sulphide toning also extend the life of the print, so the brownness of old museum prints is partly a survival strategy.

The digital equivalent is monochrome adjustment: on Fujifilm bodies, a Warm–Cool control under ACROS or MONOCHROME (older bodies called it Monochromatic Color; newer ones add a second axis), and split toning or channel curves in software. The principle is as simple as it is on location:

• Warm — brown or yellow — reads as memory, as an old photograph, as temperature.
• Cool — towards blue — reads as distance, as winter, as clinical.
• A little is enough. Toning is background music, not the subject. Once a viewer’s first observation is that the picture has a colour in it, it has gone too far.

Six symptoms and their causes

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

1. After conversion the subject disappears into the background. → Different hues, similar brightness, merged into one grey by the reduction. → Move one of them with a filter (red or yellow to drop blue, green to drop red), or reshoot looking for the difference in brightness.
2. A converted portrait looks dirty, with every blemish showing. → A green filter or a green-weighted mix, darkening skin and emphasising texture. → Use yellow or red, which lighten skin and soften blemishes.
3. A converted landscape has a blank white sky with no clouds. → Under default weights a blue sky is already bright, much like the white cloud in it. → Add red or yellow to push the sky down and the clouds forward.
4. A bad photograph is still bad in black and white, more obviously. → The problems are in composition or light, both of which live in the dimension you kept. → Monochrome cannot fix it; go back and fix those.
5. The monochrome frame looks grey and weak. → It has no true black or true white, with the whole range bunched in the middle. → Back to the previous chapter — check that the histogram reaches an end. Monochrome needs a definite black more than colour does.
6. The camera’s monochrome preview looked right and the computer’s conversion does not. → The camera applied a film simulation and filter setting while the software used default weights. → Match the channel mix by hand, or use in-camera raw conversion to reapply the same settings.

Why every nineteenth-century sky is white

Black and white began as a limitation rather than a choice. What makes it interesting is that the limitation was not uniform — early emulsions responded wildly differently to different colours.

The earliest ones reacted to blue and ultraviolet, barely to green, and not at all to red. That material is called orthochromatic, and its consequences are visible in any nineteenth-century photograph:

• Blue skies are always white. Blue light exposed the plate hard and saturated it immediately, so early landscape skies are blank paper with no cloud recorded at all. It is why Le Gray needed two negatives for the seascape in chapter eight, and the origin of the problem Adams later solved with a red filter.
• Red is black. Red lips, red fabric and red brick all render deep black. The unusually dark mouths in Victorian portraits are not makeup; they are the emulsion.
• Freckles and sun damage stand out, being reddish. A good share of a portrait studio’s retouching labour went on exactly that.

Hermann Vogel found in 1873 that certain dyes made emulsions sensitive to green — hence “orthochromatic” — and panchromatic stock, sensitive to red as well, only became general in the early twentieth century. And the entire practice of filtering only becomes meaningful after panchromatic film arrives: once the material finally responds to every colour, you have something to selectively withhold.

The change was disruptive enough to reach the actors. When Hollywood moved from orthochromatic to panchromatic stock in the mid-1920s, existing screen makeup — formulated for a film that could not see red — rendered completely differently, and had to be reformulated from scratch; Max Factor’s panchromatic makeup of 1928 was the answer, and it won an Academy Award. An emulsion’s spectral response changed what faces were painted with.

So the filter table earlier in this chapter is the residue of a technical history: first the film could not see red, then it could see everything, and then photographers began deliberately making it unsee particular things.

The decision table

What you see	What black and white does to it	What to do
Noisy, multicoloured signage	Deletes the noisy dimensions; the frame gets cleaner	Convert, and make it your first thought
Subject and background differing in hue, close in brightness	The two merge into one grey	Do not simply convert; separate with a filter, or recompose on brightness
Blue sky with white cloud	Sky and cloud become equally white	Add red or yellow so the sky sinks
A portrait where skin matters	Green darkens and coarsens skin	Use yellow or red; avoid green
Texture, geometry, shadow patterns	Almost nothing lost, usually improved	Convert, and use side light to strengthen texture
Fog, snow, rain, overcast	Little colour there to lose	Convert, and make sure you keep a definite black