What is Colour Management? Managing colour is something that we all do when we create and print images. When you assess a scan on your monitor and make adjustments so that you get a good print from your inkjet printer you are using colour management. Your own experience may tell you that your inkjet printer tends to produce slightly cool results compared to what you see on screen. To get an acceptable proof you need to make a small adjustment to the colour balance to warm the image prior to printing. Some people see this process as a failure of their equipment – it isn’t. The problems arise from a shortcoming in mathematics and colour descriptions! All colour devices have their own character; they excel in some colours and don’t do so well with others. You've probably heard the term *gamut* being used as a description of the number of colours that a printer is capable of reproducing. When a manufacturer creates new inks and/or papers they measure the colour gamut and then assign numerical values to the colours. The purest red, green and blue values are all set to the maximum allowed in RGB colour space – 255. Sometime later a new dye is discovered that produces a much more vivid red and the manufacturer launches a new range of inks. Once again the colour gamut is measured. The red is incredibly rich and vibrant, so what colour value do they assign it? 255 of course - as they haven’t got any other option! So if you send the same RGB value to printers using old and new inks you get different colours! The RGB colour model is called *device dependent* because the colour reproduced is totally dependent on the device. This concept is really important to understand so I’ll list the device dependent colour spaces that you are likely to meet: RGB
CMY
CMYK
HSB
HSV And to reiterate that if you are given an image in any of these colour spaces they *will* reproduce *differently* on any device that you choose to print or proof on. So how do you ensure that you get the correct (or best matching) colours from different devices? Next we'll discuss ICC Profiles. ICC Profiles In theory colour management is very simple: 1. You invent a colour model that can assign a numerical value to *any colour* that it is possible to create on *any device* – in other words a *device independent* colour model. 2. You build a profile of every device that you intend to use. This allows you to assign a numeric value to all the colours that the device can produce. It also gives you formula for translating those *device values* to our new *device independent* colour model. IMPORTANT: The profile does NOT change the device colours. A profiled printer will not produce more vivid colours. A profiled scanner will not record different colours. The profile works only when the colours are converted for output on another device. So they will change the colours that you see on your monitor and will change the colours when you convert an RGB scan to CMYK. 3. You attach this profile to all images associated with your device so that when the images are sent to different devices the colours can be accurately translated: existing device values => device independent values => new device values. And everybody gets accurate colour results without any bother at all You, the user, don’t have to bother with learning about any of the maths involved with the trans- formations. All you need to do is ensure that you are using the appropriate profiles in your workflow. The profiles are commonly called ICC profiles because they conform to standards that have been devised and published by the International Colour Consortium (http://www.color.org). While I’m mentioning acronyms, I may as well throw in a couple more that you might be faced with when checking your settings: CMM = Colour Management Module. Software code that manages the mathematical transformations required to move your image from one colour space to another. Different manufacturers have their own varieties. ACE = Adobe Color Engine. This is Adobe’s CMM, you’ll find it lurking in all of their graphic software. PCS = Profile Connection Space. This is the name given to the device independent colour model used by your profiles. ICC profiles may use either CIE-XYZ or CIELAB colour models. It all sounds good in theory - we'll look at some practical applications next. Practical Applications for ICC Profiles Scanners are RGB devices because they record the colours that they *see* as RGB values. The values are *device dependent* so we know that many of the colours will be inaccurate when viewed on another RGB device - including your monitor. So, we haven’t even done a scan yet and already our colours are wrong! First job is to give the RGB scanner values some real meaning that other devices can understand; we need to build a profile. This is done by scanning an image with known colour values and comparing the scan of the image to those known values. Some scanner software has this function built in and you may have been supplied an IT8 colour target with your scanning software. Comparison of the scanned results with the known values allows your computer to build a translation table and a mathematical formula for converting scanned RGB values to Lab (or XYZ) values. This information is the scanner profile and should be attached to all of your scanned images. Please remember that the scanner profile does *not* alter or manipulate the RGB values recorded by your scanner. Scans carried out before and after profiling will look no different if you check the RGB data values. But, once profiled, your RGB scanner values now have independent verification of what it is they are trying to describe. This working method can be used to your advantage because if you profile your scanner tomorrow, you can still apply the new profile to the scans that you did yesterday! Some words of warning: scanner profiles should not be used as *working spaces*. If you need to retouch your images they should be converted to an appropriate working space (Adobe1998, DonRGB, BestRGB etc). We'll discuss the reasons for this next week when we look at working space profiles and monitor profiles. This colour management seems pretty easy doesn’t it? It doesn’t mess around with your scanning process, you may already have all the tools you need to build your own profiles and you get to improve existing images just by applying your new profile! If you don’t have the software or targets then you’ll need to purchase some software that builds scanner profiles. Scanner profiling software is relatively cheap (compared to the stuff you need to build printer profiles) because they aren’t that difficult to produce. However, you should be prepared to pay a premium for good scanner targets because “the better the target, the better your profile”. Next we'll look at monitor and working space profiles. Working Spaces & Monitor Spaces Viewing an accurate image of the RGB scan in an image editing programme like Adobe Photoshop requires two further profiles: working space and monitor. RGB Working Space Profile Your working space is an RGB space that gives meaning, colour and depth to the RGB values that you intend to view and manipulate. Working space profiles are selected using two criteria: size and gamma. The working space has to be big enough to describe all of the colours you wish to manipulate and reproduce without clipping. Ideally this means that your working space should have a larger gamut than both your scanner and your printer. However, it shouldn’t be too big, or you will run into problems caused by stretching 255 possible values in each channel too far – leading to quantization (banding or stepping). Most people associate the term gamma with screen brightness and therefore don't realise its importance to image manipulation. Gamma can be more usefully thought of as a description of how much space is allocated to the highlight and shadow details in your images. Gamma values of 1.8 allocate more space to the highlight end of your image. Gamma values of 2.2 allocate more space to the shadow end. Working spaces like sRGB and Adobe(1998) have gamma values of 2.2, whereas AppleRGB and Colormatch use 1.8. How do you use the working space profile? Scanner spaces are no good as working spaces. They are too distorted and uneven and likely to produce some odd results during heavy editing. The working space is an *ideal* space where colour values are evenly spaced and predictable. For example: an RGB value of 100:100:100 will give a perfectly neutral tone in sRGB or any other RGB working space. We can pretty much guarantee that this isn't the case in a scanner profile where some kind of cast correction will be going on. Open the image that you want to edit. If it is tagged with a scanner profile then you should *convert* the image by selecting: Mode=>Convert to Profile Working Space profiles also differ from Scanner Profiles because they have information embedded in them that enable the RGB/Lab conversions to work both ways so that you may convert *to* your working space and convert *from* your working space. The Monitor Profile Like any other RGB device, your computer screen has its own colour character. It can be profiled in a similar manner to other RGB devices – send known values to the screen, use a device to measure the emitted colour values, compare the real colours to the known values and bingo, you’ve got yourself a profile. Unlike the other profiles in Photoshop, the monitor profile is loaded into the Operating System so that almost all applications are affected by the monitor profile. Building a good monitor profile is essential if you want really accurate soft proofing on your monitor. However, you can get pretty good results without using expensive hardware and software. The most common problems with un-calibrated systems are: Incorrect Gamma
Incorrect White Point Which will produce images that are too dark and too warm on press. Setting your monitor gamma to 2.2 and lowering the White Point from 9300K to 6500K will instantly improve the quality of your images. And don't forget to set your Desktop to a neutral gray! Making choices about colour bias is impossible against any other background. Next we'll look at printer profiles. Printing profiles are more complex and more expensive to produce than scanner or monitor profiles because they are more complex to produce and require more expensive measuring devices. RGB Print Profiles The simplest devices to profile are those that use RGB input - including desktop inkjets and the Fuji Pictrograph. Note that we're talking about "RGB input" not the "output" of the device. This is an important distinction when calibrating ink jet printers. Many desktop models use up to 7 output colours (CcMmYKk) to create a proof, yet need nothing more than a simple RGB profile to get them colour correct. The printer drivers that are created and supplied by the manufacturer do some really complex stuff like providing the correct amount of ink to the heads and working out where the dark cyan stops and the light cyan begins. It is possible to bypass the manufacturers drivers and to take complete control of the output of all 7 colours, but doing so is way beyond the software and hardware available to the average user. So, the best method of ensuring accurate results is to use a profile to slightly modify the RGB input data and let the printer driver do the rest of the work. CMYK Print Profiles CMYK devices like printing presses are a bit more difficult to profile. The first problem is maintaining consistent output. Your average ink jet printer is an incredibly consistent device - the output matches from day to day, week to week etc. Printing presses are nothing like so consistent. They not only vary over time, they also vary across the page depending on how much ink is being used in different areas of the sheet! The profiles themselves are also more complex because they need to take account of: 1. Dot Gain 2. Maximum ink limits 3. Black length - where should black start and finish 4. Black width - how much black should be generated Dot gain and ink limits are usually imposed by the type of press so are fairly easy to work out. But calculation of black generation includes *personal* preferences and may vary with the types of images that are being printed - not an easy task. Other Profiles The profiles mentioned above are all known as device profiles because they relate to specific devices (or classes of device). There are other profile types that you might meet: 1. Device Link Profiles. Many profile-building packages have the ability to combine profiles together so that you can use one profile to handle two or more processes. These profiles are commonly found in proofing devices where you might want to combine a paper profile with the printing profile of a device that you are simulating. 2. Abstract Profiles. Profiles can be used to create specific effects or colour enhancements. 3. Named Colour Profiles. Specific colours, for example Pantones, may be specified with corresponding colour values. |
