UPDATE: September 2, 2013 A month or so ago I did a Powerpoint presentation for our photo meetup group on the Essentials of Digital Photography as a starting point for a talk on Lightroom, and I created a video of that presentation. This is complementary to but different from the post below.
This started out as a quick response to a friend from a local photography meetup group on how to profile her scanner, but some 5000 characters later I thought maybe I should simply post it here instead. As I mention (several times) below, I’m far from being a colour expert, so at best this is a layman’s explanation. Any errors are wholly mine, and if you want to add corrections in the comments below, feel free to do so. Just remember that this is targeted toward the average reader!Hi There:
At last night’s meeting I asked if you’d calibrated your scanner, and you replied that you’re still new to the world of digital photography so I thought I’d provide a bit of information for you. Proper colour management is really at the heart of a good digital workflow, but it all comes down to what you’re planning to do with your images. For example, someone who shoots images only to put them on the web really doesn’t need to concern themselves with colour management because the ‘net is largely uncoordinated. There are actually a couple of web browsers that have colour management options – Safari and Firefox for example – but it’s highly unlikely that most of the people viewing these images online will have any idea what colour management is or how to use it.
However, if you want to print your images, then it’s a different story. First of all, colour management is about getting consistent colour, not accurate colour. Accurate colour really has no meaning because to be accurate something has to be measured against some predetermined standard. Is this red, or is this red? That sort of thing. Ideally one would have consistent colour from input (camera or scanner for example) to monitor to output (printer or press).
Now, it’s possible to get REALLY deep into colour management, and even in the wading pool you’ll come across terms like LAB, gamut, CIE (International Commission on Illumination (usually abbreviated CIE for its French name, Commission internationale de l’éclairage)), ICC (International Colour Consortium), HSL (hue, saturation and luminance) and so on. I haven’t gotten much beyond the wading pool myself Actually, LAB and HSL are very similar in some respects, so let’s start there. Hue refers to the shifting of colour. For example, if we had yellow–orange–red paints side by side and we added more yellow to the orange or more red to the orange we’d be changing the hue. Saturation refers to the ‘amount’ of colour. Something that’s black and white has a saturation value of ‘0’ – no colour. Something with 100% saturation would look, well, gaudy. Luminance refers to how ‘light’ or how ‘dark’ a colour is. Imagine having a can of paint and adding white paint to it or adding black paint to it. Doing so changes the luminance of that colour. All clear as mud now?
Before I go any further I should mention about the values 0-255 that you see in terms of colour values and where it comes from. You’ve probably heard that .jpg files are 8-bit colour, and .tif files are 16-bit colour. What does that mean exactly? Well it goes back to binary, because binary is all computers understand. For example, in 8-bit colour each pixel is represented by 2 bytes. 20 is 0, 21 is 2, 22 is 4, etc. so if we had exponents and values lined up they’d look like this:
So 28 is 256 and ‘8-bit’ colour has 256 values, from 0-255. 0 is black, and 255 is white. Since we’re dealing with exponents, ’16-bit’colour is 216, which is 65366. 8-bit colour has 256 different separations, and 16-bit colour has 65366 different separations – for each colour (see next paragraph). A VGA monitor was one capable of displaying 8-bit colour. How many people (beside me) remember a time before colour monitors? So-called ‘true colour’ is 24-bit (28 x 3 colours), and beyond that you can even get into 36-bit or 48-bit colour, which have 12-bits or 16-bits per channel. Images can have 64-bit pixels with 48-bit color and a 16-bit alpha channel.. For more on ‘Color Depth’ I’d suggest visiting this site (with thanks to Wikipedia!).
The next thing to consider is RGB vs. CMYK, or what’s known as additive colour vs. subtractive colour. Computer monitors are emissive; they give off light. RGB stands for Red, Green, Blue. If we start with nothing and add red light, we get red, if we add green light we get green, and if we add blue we get blue. If we add equal parts red and green we get yellow, if we add green and blue we get cyan, and if we add red and blue we get magenta. By varying the ratios of the red, green and blue we create other colours. By the way, 256 x 256 x 256 works out to 16,777,216 possible different colour combinations of RGB values possible on a computer monitor..
Printers use subtractive colour. If we start with a white sheet of paper (reflects all colours) and add a colour to it, the paper begins to absorb some of the colour instead. CMYK stands for Cyan, Magenta, Yellow and Black (they used the K (‘key’) for black because the cyan, magenta and yellow plates are aligned with the key or black plate. If we mix cyan, magenta and yellow together in different increments (0-255 again) we can create different colours, and by adding 0-255 units of black we can change the luminance.
Because cameras use additive colour and printers use subtractive colour we need some way to get the colour from one type to approximate the colour from the other type.
A given range of colours is selectively called a colour space, and there are certain standards. Consider a box of crayons – the really big Crayola box with all of the colours. This imaginary box of crayons represents every colour possible. Now let’s say we know we’re working on the ‘net and we know that the average computer monitor can’t display all of those colours, so there’s no point in having the whole box available. From that big box of crayons we select only those colours that can be displayed on a computer monitor. The range and number of colours we choose is called the gamut. For the internet and monitors in general, we use a standard called sRGB (standard RGB). sRGB is a relatively small gamut (8-bit) and it works for this case.
BUT. You knew there was going to be a but there. The camera’s sensor can record more colours than can be displayed on the ‘net and if you were to assign every image you have an sRGB colour space you’d be taking your big box of crayons and throwing away all of those other colours. There are other gamuts as well; CMYK is the smallest gamut and Lightroom for example uses something proprietary (Melissa RGB) but that is very close to what’s known as ProPhoto RGB. ProPhoto RGB is a much larger gamut and can hold all of the colours generated by a digital camera. Adobe RGB is a median standard used by many photographers. For more on gamuts, see the links at the bottom of this post.
Incidentally, this is one of the differences between shooting RAW and shooting .jpg – .jpg files use sRGB, which is 8-bit colour. RAW files use 16-bit colour. So imagine taking the 65366 possible colour separations/channel available in 16-bit colour and throwing away all of them but the 256 colour separations used in sRGB. The flip side is that an image with 16-bit colour has a lot more information available and so yields larger files. This requires more computer power to work with them and more storage space on the hard drive. Again, it depends on what you want to do with your images.
Now because different equipment uses different colour types (additive vs. subtractive) and has a different colour space, sometimes we can have colours from one piece of equipment (say a camera) that aren’t available in the gamut of the other piece of equipment (say a printer). These colours are said to be ‘out of gamut’. Software like Photoshop uses what is known as ‘rendering intents’ to try to squeeze those ‘out of gamut’ colours into something as close as possible to what it CAN display. Again, see the links at the bottom for more information.
That’s a very quick overview of colour management. When you calibrate your equipment – whether it be camera, scanner, printer or offset press, what you’re doing is using hardware and software to create ‘profiles’ so that a colour of say R-66, G-127, B-239 will look approximately the same on every device – as close as possible. A profile reads what’s being displayed for a range of colours and then creates a map that says essentially, ‘Okay, the monitor wants to display R-59, G-135, B-222, so we need to make these adjustments: R+7, G-8, B+17 to make it look right’.
Now, to calibrate different devices you need both hardware and software. Your eyes and your brain are simply too accommodating. No matter the light, we’ll see something that’s supposed to be ‘white’ as white, even though it might be yellowish or bluish or whatever. This again is a whole world unto itself, and it depends on how far down the rabbit hole you want to go.
To calibrate your camera for example, you can use something like the X-Rite Colour Checker Passport or even a spectrally neutral gray card – in the light you’re shooting (because every light has its own colour or ‘temperature’) – and then use your image-processing software to adjust the colour balance of the image(s) to make the colours neutral.
For a monitor you need a colorimeter or spectrophotometer that sits on your screen and measures the light being emitted by the screen and software that creates a profile for your monitor. NB: When profiling your monitor you need to be aware of two things. One, your monitor will change colours as it warms up, so let it run for 30 minutes or more before profiling. Two, your monitor will change over time, so depending on how diligent you want to be you should re-do the profile weekly or monthly.
For your printer, the profile will depend on the printer, the inks and the paper, so again you’ll need a spectrophotometer and software. You’ll print out of a series of colour patches and have the device read those colours then create a profile from them. There are all kinds of variables over whether to create a 64-colour patch or a thousand-colour patch, depending on whether you’re printing out pictures of your kids, creating fine art or somewhere in between, but I’m trying to keep this somewhat simplified. How often you profile your printer depends on paper batches, worn out/ replaced print heads, etc. but at the very least you will need a different profile for each printer/ paper combination. If you lack the necessary hardware and software you can also hire someone to make profiles for you. The other option for a printer is to use what’s known as a ‘canned’ profile. These are usually available online from the paper manufacturers and depending on how picky you are they may work just fine.
If you’re displaying your images using a projector, you can also use similar hardware/software to create a profile for your projector.
Now for a scanner, you have two choices. Basically to calibrate a scanner you need what’s known as a ‘target’ – an image showing the different colour patches and grayscale bars from white to black. If you’re doing this for yourself and you’ve properly calibrated your monitor and your printer and have achieved printouts you’re happy with, you can print out a target to use with your scanner. You can find target images online.
The other choice is to buy what’s known as an ‘IT-8’ target. Again, these are available online and from some camera stores. You scan the target and then use the resulting image with some software to create a profile for your scanner. There is a freeware program you can use called CoCa that will read your target scan and create a profile for you. See the links at the bottom.
So, that’s a very quick overview of colour management – as I understand it, and I’m certainly no expert. One person who is an expert is Andrew Rodney; you can find him at the Digital Dog site. A few beginner-level articles on colour management for photographers are:
Update, December 21, 2011: I see a lot of questions from people about their images and the Internet. “Why don’t my images look good on Facebook?” for example. There are a couple of reasons for this. One is that the ‘net uses an sRGB colour space, so if you export your images in AdobeRGB for example, they won’t look right because too many colours are being crammed into too few. The other reason is that not all web browsers are colour-managed. There’s an excellent article on this, here: WEB BROWSER COLOR MANAGEMENT TUTORIAL.
Update, May 2, 2013: Laura Shoe has an excellent tutorial on bit depth over on her site: 8 bit, 12 bit, 14 bit, 16 bit — What Does It Really Mean to Digital Photographers? She also has an article on Chromatic Aberration that explains it well.
P.S. There’s also:
Why Does “Brightness” Wash Colors to White?: an article from Jeffrey Friedl on why adjusting exposure/brightness in programs like Lightroom and Photoshop can reduce saturation. There’s also a brief description of ‘colour’ vs. ‘chromaticity’
Colour Management for Photographers: a series of nine articles from Digital Outback Photo
Introduction to Colour Management: a series of articles on colour management from Phil Cruse of Graphic Quality Consultancy
From Camera to Print and Screen: a series of video tutorials (some 13 hours) by Michael Reichmann and Jeff Schewe, available for purchase from the Luminous Landscape
CoCa Color Camera Calibrator – freeware colour calibration software from Andrew Stawowczyk Long, National Library of Australia. Follow the link for the DOHM Sitemap at the bottom of the page, then click on the CoCa link.
Correcting White Balance: a tutorial on white balance correction, primarily using the X-Rite Colour Checker Passport