Aside from any OS supported colour management such as the ICM in Microsoft Windows XP or WCS in Vista, at the heart of any user controlled colour management system is some form of image editing software. Most frequently, for Microsoft Windows users this will be Adobe Photoshop (now at version CS4). A good and rather cheaper alternative is available from Corel. Linux users have the strangely named GIMP (GNU Image Manipulation Program) which does now include some colour management functions.
This article takes an in-depth look at colour management in both Adobe and Corel products. Without relating these colour settings to colour management principles, it can be difficult to understand what they do and the choice, number and combination of user settings can be totally bewildering. Comparing the approaches of two competing image editing products is also very helpful in understanding how colour management works in practice.
Adobe And Corel Products
Digital still images are either bitmap or vector based, with bitmaps being used for photographic images and vectors for design graphics and text. Adobe are perhaps best known for their bitmap based photo editing software — Photoshop, while Corel are known for their vector based graphics design program — CorelDRAW. Both products are powerful image manipulation tools and because of the number of supported features have quite a steep learning curve. Many graphics design professionals use both programs and move image files between the two. A clear knowledge of the colour management in each product is required to be able to do this successfully.
In addition to their flagship products both companies market a range of image and graphics design bitmap and vector based software that also incorporates colour management. For example Illustrator is Adobe’s vector graphics design and editing program. Although Corel include a bitmap photo editor, Photo-Paint X4, with the CorelDRAW Graphics Suite X4, they also sell Paint Shop Pro, a stand-alone bitmap photo editor. Because some of these products started life in development by other companies and were bought up by Adobe and Corel, the way that colour management was implemented was not entirely consistent across the product range. For example Paint Shop Pro does not use the excellent flow chart diagram found in CorelDRAW to represent the colour work flow, but rather three menu branches — Color Management, Color Working Space and Monitor Calibration all of which access their own control panel windows. This menu branch is reached from the top line application menu through clicking File > Color Mangement (or using the [Alt] [F] [G] control keys).
Adobe have made more effort to integrate colour management across their product range. The Adobe Creative Suites include Adobe Bridge, a program that automatically synchronises colour settings across applications.
Unfortunately, in a perhaps misguided attempt to simplify colour management, Corel have chosen to use non-standard names for their Rendering Intent settings in Paint Shop Pro. These are referred to as Pictures, Proof, Graphics and Match.
Adobe Photoshop and Illustrator both run on the Apple Mac OS as well as on Windows while the Corel products are Windows only.
The Reference Colour Space
So that accurate image colour relationships can be maintained as image data moves from capture (input) to an edited image ready for reproduction (output), image editing software uses an internal colour space as a reference for colour space translation calculations. This intermediate reference space is used because this minimises the number of device profiles and transforms required to translate from any input to any output.
Gradually over the years reference colour spaces and colour conversion transforms have moved towards better representations of perceptual linearity and towards greater accuracy. Perceptual linearity means that the individual coordinates that represent each of the measured colour shades the average human can discriminate are equi-distantly spaced within the colour space volume. As with all digital representations colour volumes are quantised, in this case into a number of points on a 3D grid. The spacing of the grid is determined by the bit depth used to represent it. For example if eight bit values are used for the height width and depth of a cube the cube will have 256 divisions on each axis and contain 16,777,216 equi-distant defined points. If the colour coordinates of all of the discernible shades coincided exactly with these grid points, 8 bit coordinates would be sufficiently accurate. However depending on the colour space used they may not be.
As described in a previous feature the earliest colour model in use is the CIE XYZ (1931) and this is based directly on measurements. The original measurements were made using a sample group of only seventeen people, although the results have since been confirmed using larger sample groups. Within the XYZ colour space some areas of colour discrimination of the human visual system are packed closely together and other areas are widely spaced. For example the greens occupy a large portion of the space while yellows occupy only a small portion. Based on the original XYZ measurements, CIE L*a*b* (1976) is one attempt to create a more perceptually linear colour space. Adobe products use CIE L*a*b* as their reference space while CorelDRAW Graphics Suite X4 uses sRGB.
The embedded colour management in Microsoft Windows from Windows 95 up to XP also uses the sRGB colour space for reference. This colour space was originally created by Microsoft and HP as a reference space for colour images on the world wide web. It is based on the typical device colour space for a CRT monitor. In this sense it is an idealised device dependant colour space and has quite a small gamut.
The colour engine or Colour Management Module (CMM — Corel define this as Color Matching Module) is the piece of software that makes the calculations required to convert a colour image from one colour device and its associated colour space to another. These conversions are performed via the reference colour space. To make the conversion the colour engine uses the reference space, the source and destination colour space, the associated input and output device profiles and the users choice of rendering intent. Because of the complexity of the calculations involved and a variety of methods for exactly how the transformations are performed, there are a number of different colour engines in use.
Some image editing software provides a choice of which colour engine or CMM to use. For example by left clicking on the central colour icon in the colour management control panel in CorelDRAW (X3 or X4) the user has the choice of either the Kodak Digital Science or Microsoft ICM 2.0 CMMs. With Microsoft Vista there is also the additional choice of — Microsoft WCS. In Adobe Photoshop, opening the Color Settings control panel from the Edit menu ([Shift] [Ctrl] [K]) with the More Options button selected, under Conversion Options — Engine, allows the user to choose either Adobe (ACE) or Microsoft ICM. Differences in the algorithms used in different colour engines, coupled with the use of different reference colour spaces can result in visible differences in edited images when the same source file is processed by different image editing software.
Moving Colour Images Between Adobe And Corel
There some limitations on file format support between the competing products. The native file formats are: PSD for Photoshop, CDR for CorelDRAW and PSP for Paint Shop Pro. Paint Shop Pro X2 can save files in PSD format and CorelDRAW can save in AI Adobe Illustrator format. Adobe Photoshop formats are more restricted, although it can save in one of the standard image formats such as JPG or TIF which Corel programs can open or import.
CorelDRAW cannot open JPG files, these must be imported using the File > Import menu choices. For an image to have the same colour appearance in CorelDRAW that it has in Photoshop the colour work flows must be set up with the same colour spaces and profiles.
Microsoft And Colour Management
Starting with Windows 95, in versions of Windows preceding Vista, Microsoft introduced the Image Color Management (ICM) and later ICM2, colour models, which are ICC based.
With Windows Vista, Microsoft, in cooperation with Canon, introduced a new colour model called the Windows Colour System (WCS). Microsoft say the ICC approach (i.e. the system used by just about every company except Microsoft, including Adobe and Corel) is fundamentally flawed, that problems with previous colour management systems are due to not using a perceptually accurate colour model, and that Vista with WCS provides this necessary perceptual accuracy (see Article References 3). Although by default WCS still uses sRGB (IEC 61966-2-1) as its reference colour space it does allow the use of other colour spaces, such as scRGB (IEC 61966-2-2 ). Colour accuracy is greater in scRGB because, although it uses identical color primaries and white/black points as the sRGB color space, it uses 16-bit floating point (half precision) linear values instead of the gamma compressed 8-bit integers used in sRGB. Microsoft call the colour engine used by WCS the Color Infrastructure and Translation Engine (CITE). It uses improved transform algorithms based on the CIECAM02 perceptual model to perform the colour calculations (see Article References 4).
Setting The Colour Controls
In Adobe Photoshop the the colour management controls are accessed from Edit > Color Settings (or [Shift] [Ctrl] [K]).
Adobe Photoshop and Corel Photo Shop have a setting for Working Space. This sets the default colour space that any un-tagged image files will be converted into when opened and also will be saved as. There are separate settings for RGB and CMYK files. Depending on the Policies settings, opening image files that are tagged with a colour space that is not the same as the working space can trigger an options warning.
Below the CMYK setting are choices for Dot Gain. Dot gain relates to the absorbency of the paper used for printing. Dots represented as a certain size in the image file tend to grow in size when printed as the ink soaks in to the paper. This affects the appearance of printed colours.
Engine and Intent settings are as previously mentioned in the section on Colour Engines.
Rather than use a control panel with a series of drop down menus and entry boxes Corel have adopted a much clearer flow chart diagram to illustrate the colour work flow. In CorelDRAW X3, in X4 and in the associated Photo-Paint, the menu selection Tools > Color Management, launches a control panel displaying a graphical representation of the colour work flow (pressing the Alt o, Alt c key combinations also opens this window). The internal colour space is shown at the centre with five icons representing scanner or camera input, printing press output, display screen, desktop colour print output and file input/output surrounding it in an equally spaced ring. Drop down menus below each icon allow selection of profiles, while arrows running between the internal colour space and the device icons allow selection of different colour paths. Clicking on the arrows toggles the appropriate profiles on and off. Further arrows running between the two print icons and the display icon allow settings for soft proofing. Left clicking on all of the icons except the scanner/camera opens an Advanced settings control panel associated with that icon.
Display Calibration And Profiling
To be able to maintain any control over colour it is absolutely essential that the computer display is calibrated and correctly profiled. The display is the users visual window and reference on what’s happening to colour at various stages in the colour editing process.
Calibration involves setting the display to known levels of brightness, contrast and colour balance. This is achieved through using a suitable calibration device such as a colorimeter and adjusting the physical controls on the display until the measurements reach the required levels. This sets the display to a known and repeatable state so that successive profiles are always created from the same starting point.
The input to output transfer characteristics of displays vary, two examples of the same model LCD or CRT monitor may not produce exactly the same brightness or colour when driven by identical video signals. Most computer monitors do not have individual colour channel linearity controls. Uncorrected monitors not only deviate from the ideal desired transfer curve, but are also likely to have different transfer characteristics for each colour channel. If the Red, Green and Blue channels do not track together the display will exhibit unwanted, shade dependant hue variations. Poor RGB tracking is most apparent where obvious colour tints appear in monochrome images with a wide range of greys.
Software Display Calibrators
Although the approach has been abandoned in Adobe Photoshop CS4 and Adobe now specifically recommend using a third party hardware based display calibrator, earlier versions of Photoshop included a software only display calibrator and profiler applet called Adobe Gamma. The Photoshop install places this in the Windows Control Panel and also installs a short cut in the Start folder to the Adobe profile loader.
Adobe Gamma attempts to calibrate and profile a display by eye. First a graduated brightness scale is used to calibrate the brightness and contrast settings then a series of solid colour / alternate line patterns is displayed. The center solid patch is adjusted in brightness by eye until it matches the brightness of the alternate lines. Required corrections are then derived from these relative settings and used to construct a profile. Corel Paint Shop Pro Photo X2 has a similar utility, but improves accuracy by providing adjustment at 5 different brightness levels. Adobe Gamma only corrects at a single brightness level. These ‘eyeball’ calibrators are perhaps better than nothing, but are not as accurate as the current hardware calibrators.
Recent display drivers include controls which can be used to modify display characteristic, such as gamma controls. These should be turned off, or left at default values, before (and after) a display is calibrated and profiled.
Boot Time Display Profile Loaders
The display colour response is not corrected by making adjustments to the display itself but by applying corrections to the graphics card. Usually these corrections are loaded by a profile loader applet loading the correction values from the display profile into the colour Look Up Table (LUT) on the graphics card at boot time. These corrections are therefore global, affecting everything displayed, following boot up.
Most colour profiling applications are supplied with a profile loader applet which is installed during installation of the main program. A short cut to the profile loader is placed in the Start folder and is therefore called during operating system boot. Under Windows Vista these can be found via Start > All Programs and opening the Startup folder. Any profile loaders should be listed here.
If several applications with colour management that support profile boot loaders are installed on the same PC there can be some confusion. The LUT will eventually be loaded by whichever loader and profile is loaded last in the boot sequence and this may not be the desired profile. Users should manually edit the Start folder and remove all loader short cuts except to the desired loader. Ideally this should be for a hardware display calibrator and its associated software.
Hard Copy Colour Proofing
In theory the ideal colour proof work flow is to print proofs on the printer that will be used for production runs. In the case of commercial press printing this is impractical because the presses involved are not designed to print just the few pages required for colour proofing. In the past hard copy press print proofs were produced using a (DuPont) Chromalin, (Kodak) Matchprint or similar proofing system. Unfortunately the word ‘Chromalin’ is often used like the word ‘Hoover’ is used as a generic term for vacuum cleaners and if you ask a printer for a Chromalin proof that may not literally be what you get.
By the time pages reach the printing press the edited page layouts have been split into the component C, M, Y and K (known as ‘separations’) used to print the finished pages. To make a Chromalin proof, the separations, either in digital file form, or as page spread size photographic negatives are used to produce positive colour exposures of the four colours on four sheets of plastic. These sheets are then aligned and bonded together to form the final colour proof. Other photographic proofing systems use the separation negatives to produce a colour photographic print on photo print paper. The digital files or negatives used to make the proof are also used to make the printing plates for the press, so in principle there is an exact correspondence between the proof and the press print.
The problem with all these colour press proofing systems is that they do not use the inks and paper that will be used in the final print run. They will show errors such as registration problems or obvious colour errors introduced at the separation stage, but will not necessarily produce exactly the same colour appearance as the final printed output. This is particularly obvious with proofs printed onto photographic paper, as the white of the photo print paper normally won’t be the same as the white of the paper used in the press.
Hard copy colour proofs are relatively expensive and time consuming to produce, commercial press operators or print service bureaus apply a charge for each colour proof (in the UK, £45 to 50 each for A4 proofs). So publications will normally only request hard copy colour proofs for images of particular importance.
Soft proofing attempts to use either a computer display or a desktop colour inkjet or laser printer to emulate the performance of the final output media, i.e. a printing press.
It is cheap, convenient and fast. However the term ‘soft proofing’ is also used to refer to colour proofs, such as PDF files sent by email, that do not attempt accurate colour reproduction, but serve merely as general proofs for checking layout, text and overall appearance.
True soft proofing relies on the gamut of the display or desktop printer being larger than the gamut of the printing press.
This allows the colour engine in the colour management software to shift and constrain the image colour values inside the confines of the display or printer gamut so that the image appears as it will when printed on the press.
Out Of Gamut Colours
As mentioned in a previous feature because the gamut of most printers is smaller than that of capture devices such as cameras and scanners, some colours may be lost when converting from the captured image into the print device colour space (see also Intents). Colour editing software usually has a setting which will highlight out of gamut colours.
Comparing Soft Proofs With Prints
In practice, when final printed output is compared side by side with the screen, soft proofing to a display often appears to be only partially successful. This is because, assuming the display has been calibrated and profiled correctly, the image will be displayed with the chosen white point (for example 6,500K) at a certain brightness (often 80 cd/m2) while also being affected to some extent by the ambient illumination reflected from the screen. In contrast the hard copy proof will be illuminated by the available light and this will almost certainly not be of the same colour temperature as the screen, nor will the proof reflect at the same brightness.
For the best correspondence between display soft proofs and final output the display should be set up alongside a viewing booth, with both the display and the booth illuminated by the same fairly low level of neutral ambient light. The colour temperature, and brightness of the display must be adjusted to be as close as possible. If the booth has a brightness control this may be the easiest way to balance the brightness to that of the display.
The display colour temperature should be set to match that of the lamps in the booth; 5,000K for D50 lamps and 6,500K for D65. Unfortunately the stated CT of viewing booth lamps isn’t particularly accurate. It may be necessary to set the monitor CT to something other than one of the standard settings to match the booth exactly and to re-calibrate and re-profile the display for this setting.
Desktop Printers And Colour Management
Desktop colour printers, particularly inkjet are designed as ‘RGB’ printers. The operating system sends print data as RGB and the printer driver and printer perform the necessary conversions to print the image correctly using CMYK inks. When images are printed from an image editing application that provides user controls for colour management a choice must be made over whether either the colour management in the application, or in the printer driver should be used. It is a common mistake to use both, which results in unpredictable colours in the final print. Only very recently have printer drivers started to include warnings indicating that both management systems may be in use.
The best approach is to select the correct printer profile in the colour controls for the image editor and to turn off the colour management in the printer driver. Unfortunately in many desktop printer drivers it is often difficult to determine exactly how to disable their internal colour and image controls. For example in the current Epson driver it is necessary to select Printer Preferences, click the Advanced button, select ICM and tick the Off (No color adjustment) box.
Conclusion — Defining Your Work Flows
Perhaps the best way to approach colour management is to first define the work flows you will use.
For example a photographer might only ever work with bitmap images and will mainly work in RGB up until the printing stage. If these are to be only small run prints rather than printed as part of a publication on a commercial press then is no need to prepare separations, most of the work flow can be in RGB and the CMYK conversions either left to the desktop printer and its driver. A hobbyist photographer might choose to set their camera for sRGB output. Depending on the camera sRGB might be the only choice. A professional photographer is more likely to choose RAW mode, a mode which outputs the image data as registered by the CCD sensor, without any intermediate processing by the camera electronics. Using this mode often requires the use of a dedicated RAW mode driver supplied by the camera manufacturer.
A graphics designer will often import digital photographs, but will manipulate and combine these with vector graphic elements. Graphic design work is usually intended for publication via a commercial printing press so it is sensible to choose to work with the effects in CMYK (this is CorelDRAWs default mode) and to soft proof using the commercial press profile.