The Film Was Calibrated for Somebody

Colour film has to be calibrated. A negative contains no fixed instruction about how it should be printed, so a laboratory needs a reference: an image whose correct appearance is agreed, against which the printing settings can be judged.

From the mid-1950s, Kodak supplied those references. They were called Shirley cards in the trade, after the Kodak employee who modelled for the first one, and they showed a woman against a standard background with colour patches. Labs printed against her. For decades, every Shirley was white, and the cards were often labelled with the word normal.

This is not a story about a secret instruction. Nobody needed to issue one. The bias operated through the entire chain at once.

The emulsion's dynamic range was developed and optimised against the reference. The printing settings technicians used were derived from the reference. The technician's own eye, trained on thousands of prints judged against the reference, was calibrated to it. And the whole arrangement was presented not as a choice but as measurement, which is the part that made it durable.

The result was consistent and well documented by the people it affected. Dark skin rendered badly: features lost in shadow, no separation between tones, a mask rather than a face. In a group shot with mixed skin tones, one exposure could not serve both, and the exposure that was standard served one of them.

Now the sequence of events that fixed it, which is worth getting in the right order.

The emulsion improvement came first, in the 1980s, and the pressure for it did not come from film studios or civil rights campaigners. It came from commercial clients. Furniture manufacturers could not photograph dark wood so that the grain was visible. Confectionery companies could not photograph dark chocolate so that it looked like chocolate. Kodak developed film with a better response in the dark end of the range, and the marketing for it talked about photographing a dark horse in low light.

The multiracial reference card, showing women of several skin tones, arrived in 1995.

So: the chemistry changed when furniture and chocolate needed it, and the reference image changed a decade later.

Meanwhile cinematographers were solving it in the room, because they had to. More light on dark-skinned actors, different exposure, specific stock choices, sometimes different lighting setups for different performers in the same scene. Several admired looks in the history of the medium are that compensation, done well, by people who were doing unpaid extra work to get a result the equipment made easy for somebody else.

Which is the general point to carry out of this, and it is not confined to film.

A technical standard encodes a decision about the normal case. Whoever is not the normal case pays for it, in extra labour, worse results, or both. That is true of film emulsion, of automatic exposure algorithms, of pulse oximeters, of voice recognition, of a great deal that presents itself as a measurement rather than as a choice. The useful question about any standard is not whether it is accurate. It is: accurate for whom, and who was in the room when that was decided.