In my previous post on colour, we saw a need for colour management. This post will provide an overview of how colour management works.
Colour management aims to provide a translation between each device so that a particular colour captured by a given device can be accurately displayed or printed by another device.
The are four key parts which make up colour management; the profile connection space (PCS), device profiles, a colour management module, and rendering intents.
Profile Connection Space
Rather than needing a formula or look-up table to translate colour information from each input device that we may use to each output device we use (this would result in MANY possible combinations), a profile connection space (PCS) is used. This means that a profile is used to translate the information provided by the input device in to the PCS.
In most photography workflows, the profile connection space is CIE LAB. In this PCS, L*a*b* represents three axis, where ‘a’ is an axis with green at one end and red at the other, ‘b’ is an axis (perpendicular to ‘a’) with blue at one end and yellow at the other, and L is an axis (perpendicular to ‘a’ and ‘b’), with black at one end and white at the other.
Unlike device profiles, which describe the way a particular device responds to or produces colour, the relationship between colours and numbers in the PCS is arbitrary.
To devices, colours are just numbers. Because each device has a unique range of colours that it can capture, display, or print, the numbers used to represent particular colours vary from device to device.
A device profile describes the colour behaviour of a particular device (for a particular light source/illuminant).
Our cameras capture raw red, green and blue values. In order for the colours that were captured by camera to be accurately translated into the PCS, we need understand how the camera behaves with regards to colour.
There are profiles that are provided by camera manufacturers and Adobe, and these do a pretty good job of describing the way that our cameras behave under most circumstances.
A profile is created by having known colours and either capturing them, or attempting to output them. For example, the image at the beginning of this post is of an x-rite colour rendition chart. By photographing this chart, which has known colour values, a profile can be built that describes how far off each known colour the capture was. This profile can be used by the colour management module to translate the raw colour numbers into numbers within the PCS which correspond with the colours on the chart. The result is that the colour described in in the PCS matches the colour in the scene captured by the camera.
Colour Management Module
A colour management module (CMM) is a piece of software that uses device profiles to translate colour values from an input device into the PCS, or from the PCS to an output device.
A CMM is needed, because a device profile doesn’t contain every possible colour value. A CMM takes the values defined in the profile, and interpolates to come up with a suitable translated output for a particular input.
It does the thinking.
The namesake of this blog (although the blog is so named because of our aim – having our prints rendered the way we intend them to be)!
Each output device can produce a particular range of colours. This is known as the colour gamut of the device. Take a monitor. Once it is producing the most saturated blue that it can, ‘asking’ it to produce a more saturated blue will not result in a more saturated blue being displayed. The same is true of a printer-paper combination.
When printing, you may have noticed that you can select a rendering intent in software. The rendering intent that you select tells the CMM what to do when it encounters colours that are out of gamut.
For printing photographs, there are two types of rendering intent we should consider. These are perceptual and relative colorimetric.
Perceptual aims to maintain colour relationships – it reduces the saturation of all colours in the in the image so that they fit within the gamut of the output device. Our visual systems are very good at perceiving the relationships between colours, but not very good at perceiving absolute colours. The perceptual rending intent works on this basis. When there are a lot of out of gamut colours, it prevents the out of gamut colours being clipped to the closest colour within the gamut.
Relative colorimetric maps the white value in the source to the white of the output (paper in the case of a print). It accurately reproduces all in gamut colours and clips out of gamut colours to the nearest colour that can be produced by the output device. Unless there are a large number of out of gamut colours, and in situations where accurate colour is critical, it is a good choice, as the colours in the output more accurately reflect original colours.
You can use soft proofing to determine which colours will be out of gamut, and assess the impact of selecting different rendering intents. Soft proofing will be covered in a future post.
Next time, we’ll look at working spaces.