Colour film, and whose skin it was calibrated for

Colour photography poses a problem that black and white does not. A monochrome image can be described as correct if the tonal range is pleasing. A colour image has to be correct relative to something, and that something has to be chosen. Every colour reproduction system in history, in photography, film, printing and television, contains a decision about which colours matter most, and that decision is usually invisible to the people affected by it.

The card

In colour film processing, laboratories used reference images to set printing exposure and colour balance. The standard reference for consumer colour negative printing in North America and much of the world was a card showing a woman in a high contrast outfit against a controlled background, together with colour patches and grey scales. The cards became known generically as Shirley cards, after the name of an early model.

The point of the card was consistency. An operator adjusted the printer until the reference rendered correctly, and then ran the day's work. That is a sound engineering practice. The difficulty is what correctly meant. For decades the reference showed a light skinned woman, and the chemistry, the printing latitude and the operator's judgement were all tuned to render that particular set of tones well.

The research that established this as a documented history rather than an anecdote is principally that of the communications scholar Lorna Roth, whose article on the Shirley cards, published in the Canadian Journal of Communication in 2009, set out the practice, its consequences and the industry's eventual response. It is the standard citation, and anyone writing about this subject is drawing on it whether or not they say so.

What actually changed it

The improvement in film emulsions' rendering of darker tones came, according to Roth's account, under commercial pressure from unexpected directions. Manufacturers of dark wood furniture complained that grain and finish were not being reproduced. Chocolate manufacturers complained about their product photography. Those complaints came from industrial customers with purchasing power and were acted on.

That detail is often reported with a note of outrage, and the outrage is understandable, but the useful reading is structural rather than moral. Photographic manufacturing responded to the customers who complained in a language it recognised. Whose skin looked right was not, in the internal logic of the business, a specification. Whose furniture looked right was.

Multiracial reference cards were introduced from the 1990s, and emulsion improvements over the same period widened the range of tones rendered well. Roth's work also documents earlier efforts by individual photographers and laboratories working outside the mainstream, who developed their own exposure and printing practices precisely because the standard ones failed them.

Somebody always decides what correct looks like. The question worth asking is who was in the room, and what they were being asked to optimise.

Lighting, and the other half of the problem

Film stock is only part of it. Cinematography is a system of stock, exposure, lighting and processing, and the conventions of studio lighting described in the close up essay were themselves developed against particular assumptions.

Classical Hollywood lighting practice was codified around a specific set of reflectance assumptions, and the standard three point setup with strong key and controlled fill was designed with a particular expected subject in mind. Cinematographers working with darker skinned performers have described, in interviews and in the technical literature, the additional work required to obtain a good result within a system that was not built for it: different key placement, more separation from the background, attention to specular highlight rather than to diffuse reflectance.

None of this is mysterious once stated. It is the ordinary consequence of a technical standard being set against a narrow reference and then treated as neutral, which is the general shape of the problem and not specific to photography.

The digital inheritance

Digital imaging did not start from a clean sheet. Sensors, colour matrices, automatic exposure, automatic white balance and, later, face detection all had to be tuned against sample data, and the composition of that data determines performance.

The most rigorously documented modern case is face recognition. The United States National Institute of Standards and Technology has published large scale evaluations of face recognition algorithms including analysis of demographic differentials, and its 2019 report on demographic effects found measurable variation in false match rates across demographic groups for many of the algorithms tested. That is a formal, adversarial, published measurement rather than an assertion, which makes it unusually solid ground in a field full of claims.

Camera automatic exposure and face detection are less well documented publicly, because the algorithms are proprietary, but the mechanism is the same: a system tuned on a sample performs best on subjects resembling the sample.

Why this belongs in a magazine about beauty

Because the image is where beauty standards are transmitted, and the image is manufactured. A generation of people learned what a good face looks like from photographs, films, magazines and television, all produced through a chain of technical decisions that were made by engineers optimising against a reference nobody outside the industry ever saw.

The cosmetic consequence is direct and it is well known within the industry: for decades, colour cosmetics ranges were built with very few shades at the darker end, and the standard commercial justification was demand. Demand is measured through sales, sales require stock, stock requires shelf space, and the whole loop is self confirming. The recent and rapid expansion of shade ranges in the mass market is one of the clearer cases of a long standing constraint turning out to have been a commercial assumption rather than a fact about customers, and it belongs alongside the story of the beauty counter, where the assumptions were physically built into the fixtures.

Where to look

Readers who want to go further should start with the holdings themselves rather than with summaries of them. The National Science and Media Museum and the Science Museum Group both publish catalogue and research material relevant to this article. Both links are given for reference only. Neither institution has been assessed by us, neither has any connection to this magazine, and nothing here is an endorsement of any organisation.