Once upon a time when we were young many of us had our introduction to painting by using the paint by the number system. It’s a fairly easy concept to grasp. We start with a legend for instance that says one equals red so where we see the number one we paint in red. If done properly then we should end up with a pretty picture of something colourful. Digital photography at its core works about the same except in reverse. When we expose a pixel in a camera sensor to light the camera will measure the red green and blue light and assign a value to the pixel in the form of an RGB value.
Later when we import the photo into our computer, the computer reads the RGB value and uses the RGB value to cause a pixel in a monitor to light up the same way. Capture and display enough of these pixels and we can make a picture (hopefully pretty). The paint by number system we used as children is no different then the way computer monitors render our photographs.
Now going back to the colour by number example, what happens if say the number one corresponded to the colour blue so when we paint the picture we end up with a blue stop sign? In digital photography it would be the equivalent to perhaps shooting with the wrong white balance setting. In the paint by number example the solution would be to repaint the same scene again and use the colour red for when we see the number one. Correcting colour in digital photography is really no different.
Understanding the Three Properties of Digital Colour
In order to properly understand digital colour we need to understand a few things. The first thing is that digital colour actually is composed of three properties: hue, saturation and luminosity. Hue is basically the electromagnetic spectrum we see when light is reflected from an object. We typically refer to hue in terms such as red, blue, green etc. The luminosity of the object is the amount of light that is reflected from the the object. The saturation of the object refers to the intensity of the hue.
In digital photography the exposure represents the luminosity, the hue is the white balance and saturation is a combination of the luminosity and the hue. Consider a bright red exotic sports car. At high noon in full daylight the car will be bright red. Put the same car in a light tight box and its colour is black because there is no light being reflected.
Digital Colour Spaces
Knowing that we have these three properties of digital colour we need a way to represent them digitally (digitally just means as a number). The digital representation of digital colour is referred to as the colour space (or image mode in Photoshop). The two we concern ourselves with the most are RGB and CMYK. RGB stands for Red, Green, Blue and is typically used by by devices like cameras, scanners and monitors. CMYK stands for Cyan, Magenta, Yellow, blacK and is typically used by printers to print our pictures.
One of the interesting things to note about the red, green and blue values in a RGB pixel is one colour value contains the hue, saturation and luminosity for that colour. The same applies to CMYK. Despite their differences one of the most important traits both colour spaces have in common is they are device dependent (more on this later).
Another colour space that you should be aware of is the LAB colour space. LAB consists of three channels the Luminosity, A and B channels. In LAB the A channel contains both magenta (red) and green while the B channel contains blue and yellow. In LAB the luminosity is stored in the L channel while the hue and saturation is stored in both the A and B channels. One of the most important aspects of LAB is unlike RGB and CMYK it is a device independent colour space.
The Problems with Device Dependent Colour Spaces
If you made it this far hopefully you have an understanding about what digital colour is and how it is represented in digital form. Now I am going to illustrate why its all important. As mentioned earlier RGB is a device dependent colour space. That means what one device defines as red will depend on the device capturing or displaying it. Every camera and every monitor for example has a slightly different idea about what red is.
The same also applies to devices that use the CMYK colour space. The best example of this would be colour printers. We capture our digital images in camera and then we display it and edit it on the monitor in RGB knowing that the monitor has its own interpretation about what RGB should look like. Next we send our picture to the printer where it translates from RGB to CMYK and prints our picture. More often then not what we get from the printer looks nothing like what we wanted to. What our digital workflow just did was something like the other game we played as children called telephone (you know where a group of people have to pass a long message from one to person to the next verbally and usually the message at the end looks nothing like the beginning).
Chromatic Adaptation aka Why we are the weak link
Usually when we talk about digital workflow and colour management we restrict our discussions to the computer side of things. However we as photographers do play a part in this process as well. Something we need to understand is the concept of chromatic adaptation. It essentially means we see what we need to see and not what is actually there. If we look at white paper under florescent lighting we see white paper. Shoot white paper under florescent lighting with daylight white balance and there will be a green colour cast in the white. This usually baffles new photographers but its only because our brain is programmed to see white because we know it should be white. Digital cameras do not have this bias.
What this means from the colour management perspective is we need to be aware that when we look at colours on screen we are going to see the colours we need to see not what is really there. Given that no two RGB devices will render the same RGB colours exactly the same its one of the reasons the same picture can look different on two different monitors and even more different on two different printers.
Understanding the Mechanics of Colour Calibration
One of the popular methods for handing this situation is to colour calibrate one’s monitor to get the colours right. One of the problems with profiling RGB display devices is they are device dependent so colour calibration software will work in the LAB colour space because its device independent. As one example, colour calibration software will render a green LAB target (a known green object in the LAB colour space) and then adjust the monitor accordingly so it properly displays the green LAB target as a green LAB target. The adjustments it records form the profile so you can see green properly.
The best part about colour calibration is it sets up your work environment perfectly for LAB colour. The downside is the majority of us do not work in LAB colour and even if we did eventually our work will either need to go back to RGB for the Internet or CMYK for printing. Contrary to what many choose to believe a colour corrected monitor does not solve everything (or does it really come close).