Matching digital colours: the hard part

Red is red is red, isn't it?

Consider the situation where we’re trying to photograph, and eventually print, a red object.

The most saturated ‘pure red’ in the scene would be recorded as 255,0,0 by the camera (assuming it were an 8-bit device). To get the most saturated pure red on the monitor we should send the triple 255,0,0 to emitters telling it to emit maximum red, no other colour. To get maximum pure red on the printer we should send 255,0,0 to the printer driver telling it to squirt inks that will reflect maximum red light, and absorb all green and blue.

Easy, right?

Unfortunately, in the real world it is just a little more complicated!

The first challenge relates to the different colours of filters, emitters and dyes.

Filters, monitors, and dyes are different physically and electrically. Because of their physical differences what is ‘red’ to a camera, is not the same colour as ‘red’ on monitor, and nor same colour as ‘red’ on printer.

A human observer can easily see that the same signal sent to the three devices produces different colours of red.

However, imagine trying to describe the three different 'reds' here to someone over the telephone, or in an email, what a particular shade of red you are seeing looks like?

It’s not easy, so we shouldn’t be too hard on mere machines!

255,0,0 means fully saturated red, nothing else, in camera, monitor and printer. To each it means the most saturated red possible. But the most saturated red on a camera will not be the same red as the most saturated red on a monitor, or the most saturated red reflected from a print.

The above applies equally to ‘pure green’ and ‘pure blue’ and any other colour for that matter.

The colour spaces of the camera, printer and monitor are all different. They are device-dependent. The same colours on each device are associated with different number triples.

A real photographic image is made up of millions of points of different colours.

Each model of camera produces a unique digital signal to represent the colour and intensity of each spot in a particular scene.

For example, your digital camera may measure a reddish-hued spot in your picture and record it as 221, 23, 98. The same triple sent to your monitor may produce a dot with a reddish hue but, because the monitor's colour space is different, it will almost certainly not be the same colour that the camera captured. Each monitor requires its own digital signal to reproduce its version of the same colour and intensity of the corresponding spot.

Because of the way the printer works it too will almost certainly produce a very different colour if sent the same triple. Each model of printer needs its own unique digital signal to lay down the amounts of ink needed to reproduce the same colour and intensity of that spot in the scene.

So the first challenge of colour management is to try and ensure that every individual colour that is recorded by the camera is translated into numbers that will be displayed as the same colour on the monitor and printer.

next page: colour gamut