Archive for the ‘cielab color space’ Category

Universal SmoothnessTest.tif (2,3 MB)

This test image contains 16,7 million colors of a
24-bit RGB file. Convert this test image with the designated ICC profile. So you can evaluate very quick and easily the color reproduction and smoothness quality of the output profile.


I recently told a friend I was writing an article about painless RGB to CMYK conversion. “A whole article?” he frowned. “Can’t you cover that in one sentence?”

Now, my friend is a designer with years of prepress experience, so I was somewhat taken aback by his question. Nonetheless, I thought I knew what he meant. “Is ‘Use Photoshop to change the color mode from RGB to CMYK’ the sentence you had in mind?” I asked.

“Well, yes,” he said. “I mean, that’s pretty easy.” The key word here is “painless.” For our purposes, let’s define painless RGB to CMYK conversion as “changing the color space of an image in a way that honors the original creator’s intent (within the limits of the printing process) and doing so in as automated a fashion as possible.”

My friend confessed that his customers were sometimes so unhappy with the color resulting from his “easy” conversion method that he often had to correct the images. As for process automation, he admitted his company had none—other than allowing RGB images to convert to CMYK in its RIP. Because the RIP isn’t set up for color management, this method frequently results in a poor color conversion, and, in some instances, a free second press run. Many printers have a better method for dealing with supplied RGB color, but my friend’s technique is not unique and I would certainly not call it painless!

Digital cameras spawn more RGB files
Dealing with supplied RGB images is a relatively recent phenomenon for commercial printers, coming on the heels of digital photography’s growing acceptance. Digital cameras and scanners capture images through red, green and blue (RGB) filters, producing RGB images. Of course, high-end CMYK drum scanners (now mostly a thing of the past) capture color in the RGB color space as well, but the the scanner’s computer handles the conversion to CMYK automatically. Consequently, longtime time prepress experts are used to working with CMYK images, and many printing companies still prefer to receive CMYK files from their clients, perhaps believing this absolves them from responsibility for the potentially “bad” color that might result from performing the conversion in-house.

As the growing use of images captured from digital cameras or designer-driven scanners has all but eliminated scanning as a prepress function, printers will be getting more and more incoming RGB image files. It is critical to determine how to successfully convert these RGB files into the cyan, magenta, yellow and black colors that will be used to print them.

Color management is absolutely essential. While color reproduction has always been “managed” for print production, color management today is almost purely electronic. Simply put, every device in the workflow that touches color—computers, monitors, digital cameras, scanners, proofers and printers—captures, displays or outputs color a little differently. Color management takes into account the different capabilities of image capture devices, display devices and output devices, and helps ensure the color displayed on one computer monitor can closely match another, or that a proof from an inkjet printer approximates what will be produced on a press.

Profiles—descriptions of each device in the workflow’s color capabilities—are the key to good color management. Profiles can be proprietary, but acceptance of ICC profiles, based on the specifications put forth by the International Color Consortium, is nearly universal. By combining source profiles with destination profiles, input color can be transformed via rendering intents to the proper color on output. The mapping of these profiles is performed by a color management/matching module (CMM). CMMs exist in many places in the production workflow, including image editing and page layout software, the computer operating system, printer drivers and RIPs.

RGB ‘s big three
In a print production workflow, there are three primary places where RGB data can be converted to CMYK:

  1. At an individual desktop computer.
  2. With a server-based system.
  3. In the RIP.

There also is at least one color conversion option available online at, an ASP-style service that offers RGB to CMYK conversion of individual images, accomplished by color experts on a pay-per-conversion basis. The system includes a unique compression method that allows very large images to be color corrected via a Web connection without a long upload/download time.At the personal computer level, Adobe Photoshop reigns as the supreme image-editing tool used universally by photographers, designers and prepress providers. Like my friend, the majority of these users employ Photoshop to make color space conversions. Using consistent color settings and the right profiles, Photoshop can convert images from RGB to CMYK successfully. (The process can even be automated through batch processing and scripting options.) Photoshop, however, is not the only option for desktop-level color conversion. You can use an image editing application, such as the Windows-based Picture Window Pro offered by Digital Light & Color (Cambridge, MA) or Binuscan’s (Hartsdale, NY) CMYK+. Scanner interface software (Binuscan ColorPro, LaserSoft Imaging’s SilverFast, Creo oXYgen, or Heidelberg’s LinoColor) naturally allow for conversion to CMYK, much like the drum scanners of old. The latest version of Adobe’s Acrobat Distiller includes an option to convert color to CMYK. And the Mac operating system’s “Save as PDF” allows savvy users to create ColorSync filters to be used to convert images, text or vector artwork to CMYK using ICC profiles in the process of creating a PDF file.

Doing it properly with desktop publishing applications
Desktop publishing applications such as QuarkXPress or Adobe InDesign also offer a way to convert color between color modes from the print or export options. Sometimes RGB to CMYK conversion is done incidentally from a desktop publishing application, as, for example, when RGB images are placed in QuarkXPress layouts, then output with the “Print Colors” option set to “Composite CMYK.” If Quark Color Management has not been enabled, this will convert any RGB image in the document to the CMYK color space using a built-in algorithm rather than ICC profiles. The resulting conversions often aren’t very pretty.

But when color management is set up properly in desktop publishing applications, placed RGB images can be converted quite successfully to CMYK. Adobe, in particular, has made consistent color settings between its Creative Suite tools a priority. All CS tools (including InDesign, Illustrator, Photoshop, Acrobat and Distiller) can use the same “color settings file” (.csf). This allows users to color manage content within all the CS applications in the same way. And with the latest version of the Adobe Creative Suite (CS2), all applications that can export PDF files will share common “Adobe PDF Settings,” so a PDF file can be created from Illustrator in the same way as Distiller or InDesign. It is possible, then, to set up master color and PDF setting files and use them with all instances of the Adobe applications within a workgroup. A printing company might share these settings with its clients, so it won’t matter where color is being converted with an Adobe-based tool—it will be done consistently and to a master specification. Now we’re on the road to painless.

There is an ever-growing number of color management aficionados who recommend storing, manipulating and color correcting images in the RGB color space for CMYK conversion just prior to final output. This works even for stalwart CMYK prepress professionals, because imaged editing tools like Photoshop allow the user to accurately preview and manipulate CMYK data while working within an RGB image mode. The benefits of an RGB workflow are many. It is far easier to maintain an accurate gray balance in RGB images, because equal amounts of RGB indicate a balanced gray. Edits to RGB images won’t result in out-of-range shifts, either, as maximum blacks and minimum whites will be set to the right level automatically when CMYK conversion is accomplished using a good profile. RGB files are smaller and easier to store, and retouching or color edits to RGB images need only be done once, even if the image will be separated later for any number of different printing processes.

Server-based options
Several server-based image/workflow management tools make color conversion a highly automated and consistent process. Adobe now offers an array of server solutions, including the Adobe Graphics Server. The Adobe Graphics Server integrates database-driven digital asset management and color management, automating the process of managing and repurposing images. Users of the Graphics Server can create scripts to automatically convert color mode and attach ICC profiles as required, relieving the desktop computer user from that time-consuming task.

Binuscan offers the IPM Workflow Server, which promises to automate many aspects of image management from color mode conversion (based on ICC profiles or conventional separation tables) to image correction such as unsharp masking or gamma adjustments, and geometrical adjustments such as rotations, crops and resizing. Like most of these servers, IPM Workflow Server is a client/server solution and is compatible with Mac and Windows platforms.

Helios (Sacremento, CA), a company with long-time name recognition in the prepress industry, offers EtherShare OPI , a server-based image replacement product, which includes ColorSync/ICC-based color management. With EtherShare OPI, the high-resolution master image can be in RGB, CMYK or Lab color mode, while the corresponding low-resolution proxy image can be in CMYK mode. When the file ultimately is separated, the properly converted CMYK image is swapped in on the fly. (See

Alwan Color (Lyon, France) has positioned its CMYK Optimizer Server as both a preflight and image manipulation tool. CMYK Optimizer Server is based on hot folders, each of which can be associated with a series of tasks, including color inspection as well as mode conversion. Image files can even be directly imported into the server from an FTP location. As a preflight tool, CMYK Optimizer Server analyzes images placed in hot folders, generating a report of potential problems, including over-limit total area coverage or black ink issues and images requiring mode transformation. (See

Enfocus (San Mateo, CA) Pitstop Server provides a similar service for PDF workflows. Pitstop Server is hot folder-based as well, and lets users preflight and perform any action that can be accomplished with Pitstop Pro to PDF files. For example, a folder can be set up to automatically detect RGB images or text in a PDF file and convert it to RGB and Black based on a specific ICC profile. (See

RGB to CMYK conversion also can be handled at the end of the production chain. While RGB to CMYK conversion is part of all Adobe CPSI RIPs, the default settings are not optimal for most printing conditions. Most prepress workflow solutions, such as Creo Prinergy, Esko-Graphics’ Scope BackStage server or Heidelberg Prinect MetaDimension (to name just three of dozens), integrate enhanced color management tools for both proofing and final output to film or plate. These systems can be set up to apply the right profile for each specific output device in the workflow. So deciding to image a job originally intended for one press to another at the last moment won’t make any difference in terms of color, when imaging properly tagged RGB images, because the RGB data will be converted using the right profile for that press.

RGB to CMYK conversion can be accomplished at the computer desktop, with a server-based solution or in the RIP. Any of these methods can work “painlessly,” but all work best when some form of color management is used.

Julie Shaffer is the director of PIA/GATF’s Center for Imaging Excellence


This is a very basic introduction to two related colour models which are becoming increasingly important in the world of colour reproduction. A colour model is merely a way of describing colour. These are among the tristimulus (three-dimensional) colour models (‘spaces’) developed by the C.I.E.

What is the CIE? C.I.E. is short for ‘Commission Internationale de l’Eclairage’, which in English is the ‘International Commission on Illumination’.  A professional scientific organisation founded over 90 years ago to exchange information on ‘all matters relating to the science and art of lighting’. The standards for colour spaces representing the visible spectrum were established in 1931, but have been revised more recently.
For those of us involved in creating colour which will be reproduced on a printed page, it is easy to forget that there are other industries which need to accurately describe colour! RGB or CMYK descriptions won’t be of any use to paint or textile manufacturers! Terms such as ‘maroon’ or ‘navy blue’ won’t be precise enough.
There are many CIE colour spaces, more correctly known as models, which serve different purposes. They are all device independent, unlike RGB or CMYK colour spaces which are related to a specific device (camera, scanner, or press, etc.) and/or material type (paper, ink set, film emulsion or lighting, etc.). These RGB and CMYK spaces usually do not cover the entire visible colour spectrum or gamut. The CIE also specify lighting conditions.

The CIE LCH Colour Space or Colour Model.

This is possibly a little easier to comprehend than the Lab colour space, with which it shares several features. It is more correctly known as  L*C*H*.  Essentially it is in the form of a sphere. There are three axes; L* , C* and H°.

The L* axis represents Lightness. This is vertical; from 0, which has no lightness (i.e. absolute black), at the bottom; through 50 in the middle, to 100 which is maximum lightness (i.e. absolute white) at the top.

The C* axis represents Chroma or ‘saturation’. This ranges from 0 at the centre of the circle, which is completely unsaturated (i.e. a neutral grey, black or white) to 100 or more at the edge of the circle for very high Chroma (saturation) or ‘colour purity’.

If we take a horizontal slice through the centre, we see a coloured circle. Around the edge of the circle we see every possible saturated colour, or Hue. This circular axis is known as H° for Hue. The units are in the form of degrees° (or angles), ranging from 0° (red) through 90° (yellow), 180° (green), 270° (blue) and back to  0°.

The LCH colour model is very useful for retouching images in a colour managed workflow, using high-end editing applications. LCH is device-independent. A similar colour model is HSB or HSL, for Hue, Saturation and Brightness(Lightness), which can be used in Adobe Photoshop and other applications. Technically this is ‘device-dependent’, however it is particularly useful for editing RGB images. For example to edit a green:  Adjust the Hue angle by increasing it to make it ‘bluish’ or by reducing it to make it ‘yellowish’; Increase the Saturation (Chroma) to make it ‘cleaner’;  increase the Brightness or Lightness to make it lighter. Go on give it a try!

The CIE Lab Colour Space or Colour Model

CIE Lab colour space used in ICC Colour ManagementThis is more correctly known as  L*a*b*.
Just as in LCH, the vertical L* axis represents Lightness, ranging from 0-100.  The other (horizontal) axes are now represented by a* and b*.  These are at right angles to each other and cross each other in the centre, which is neutral (grey, black or white). They are based on the principal that a colour cannot be both red and green, or blue and yellow.
The a* axis is green at one extremity (represented by -a), and red at the other (+a).
The b* axis has blue at one end (-b), and yellow (+b) at the other.
The centre of each axis is 0. A value of 0 or very low numbers of both a* and b* will describe a neutral or near neutral. In theory there are no maximum values of a* and b*, but in practice they are usually numbered from -128 to +127 (256 levels).

CIE Lab is extensively used in many industries apart from printing and photography. It’s uses include providing exact colour specifications for paint (including automotive, household, etc.), dyes (including textiles, plastics, etc.), printing ink and paper. Nowadays it is becoming of increasing importance in specifying printing standards such as in ISO-12647, where it is usually used instead of densitometry.
For example Paper Type 1 (115gsm gloss coated white, wood-free) has ‘Paper Shade’ described as ‘L* 95, a* 0, b* -2’. So the L*95 is very light, the a*0 neutral, and the b*-2 very slightly ‘blueish’.
Paper Type 5 (115gsm uncoated yellowish offset) is described as ‘L* 90, a* 0, b* 9’. So it is a darker, more ‘yellow’ paper. If you compare the different Lab values for Type 1 & 5 you will understand the descriptions.

Lab measurements can be used to control printing, typically by monitoring a 3-colour neutral grey mid-tone patch. It is also very useful for specifying a spot colour, perhaps an important “house” or “corporate” colour such as “Coca-Cola Red”. The same colour definition could be used for printed matter, vehicles, clothing, buildings, and of course tin cans.

To obtain CIE Lab measurements from an RGB image in Photoshop etc., you will need to have assigned the correct ICC profile to that image.

In ICC Colour Management  CIE Lab is often used as the Profile Connection Space (PCS) where it provides a link between two colour profiles, such as Input RGB (scanner or camera) and Output (CMYK or RGB press or inkjet). All ICC profiles contain a PCS. In an input  profile the tables will convert the scanner’s or camera’s RGB space to the PCS (Lab). An output profile will convert the  PCS (Lab) to the digital printer or printing press colour space (CMYK). The other PCS colour space is CIE XYZ, which is often also used by spectrophotometers to report colour, see the nextarticle.

Delta E Differences and Tolerances.

The difference between two colour samples is often expressed as Delta E, also called  DE, or ΔE. ‘Δ’ is the Greek letter for ‘D’. This can be used in quality control to show whether a printed sample, such as a colour swatch or proof, is in tolerance with a reference sample or industry standard. The difference between the L*, a* and b* values of the reference and sample will be shown as Delta E (ΔE). The resulting Delta E number will show how far apart visually the two samples are in the colour ‘sphere’.

Customers may specify that their contract proofs must have tolerances within ΔE 2.0 for example. Different tolerances may be specified for greys and primary colours. A value of less than 2 is common for greys, and less than 5 for primary CMYK and overprints. This is somewhat contentious however. Proofing RIPs sometimes have verification software to check a proof against  a standard scale, such as a Ugra/Fogra Media Wedge, using a spectrophotometer. Various software applications are available to check colour swatches and spot colours, proofs, and printed sheets.

Delta E displays the difference as a single value for colour and lightness. ΔE values of  4  and over will normally be visible to the average person, while those of 2 and over may be visible to an experienced observer. Values for neutrals are lower. Note that there are several subtly different variations of Delta E: CIE 1976, 1994, 2000, cmc delta e. When no particular ‘flavour’ of DE is mentioned, it is usually DE76 (as in 1976).  The later DE2000 is more perceptually uniform, and becoming more common.

Information source:

Download SWOP ICC  profiles

SWOP2006_Coated3_GCR_bas.icc (1,9 MB) 
This ICC profile is based on the characterization data „SWOP2006_Coated3.txt“. The maximum ink coverage for the separation is 300% and the maximum black is 98%. In addition a very strong GCR in the neutral axis and a medium GCR in the more saturated colors has been used for the profile creation.


ICC profile

SWOP2006_Coated5_GCR_bas.icc (1,9 MB) 
This ICC profile is based on the characterization data „SWOP2006_Coated5.txt“. The maximum ink coverage for the separation is 260% and the maximum black is 97%. In addition a very strong GCR in the neutral axis and a medium GCR in the more saturated colors has been used for the profile creation.


ICC profile