The International Color Consortium Profile Format

ColorSync supports the profile format defined by the International Color Consortium (ICC). The ICC format provides a single cross-platform standard for translating color data across devices. The ICC defines several types of profiles, including input, output, and display profiles. Each of these types specifies a different required set of information, but all follow the same format.

The founding members of the ICC include Adobe Systems Inc.; Agfa-Gevaert N.V.; Apple Computer, Inc.; Eastman Kodak Company; FOGRA (Honorary); Microsoft Corporation; Silicon Graphics, Inc.; and Sun Microsystems, Inc. These companies have committed to full support of this specification in their operating systems, platforms, and applications.

To obtain a copy of the International Color Consortium Profile Format Specification, or to get other information about the ICC, visit the ICC Web site at <http://www.color.org/>.

ColorSync and ICC Profile Format Version Numbers

The first version of ColorSync used a 1.0 profile format that preceded the ICC profile format definition. Starting with version 2.0 of the ColorSync Manager, ColorSync uses a 2.x profile format that supports all current ICC profile format versions. As of this writing, that includes ICC versions 2.0 and 2.1. The ICC defines the profile format version as part of the profile header. For more information on the differences between these profile format versions, see ColorSync 1.0 Profiles and Version 2.x Profiles.

Source and Destination Profiles

When a ColorSync-supportive scanning application creates a scanned image, it embeds a profile for the scanner in the image. The profile that is associated with the image and describes the characteristics of the device on which the image was created is called the source profile. If the colors in the image are subsequently converted to another color space by the scanning application or by another ColorSync-supportive application, ColorSync can use that source profile to identify the original colors and to match them to colors expressed in the new color space.

Displaying the image requires using another profile, which is associated with the output device, such as a display. The profile for that device is called the destination profile. If the image is destined for a display, ColorSync can use the display’s profile (the destination profile) along with the image’s source profile to match the image’s colors to the display’s gamut. If the image is printed, ColorSync can use the printer’s profile to match the image’s colors to the printer, including generating black and removing excessive color densities (known as undercolor removal, or UCR) where appropriate.

Profile Classes

The ColorSync Manager supports seven classes, or types, of profiles. These classes are defined below. Three of the profile classes define device profiles for different types of devices: input, output, and display devices. The other four profile classes include definitions for an abstract profile, a color space profile, a named color space profile, and a device link profile. The constants used to specify these classes are described in ColorSync Manager Reference.

A device profile characterizes a particular device: that is, it describes the characteristics of a color space for a physical device in a particular state. A display, for example, might have a single profile, or it might have several, based on differences in gamma value and white point. A printer might have a different profile for each paper type or ink type it uses because each paper type and ink type constitutes a different printer state. When an application calls a ColorSync Manager color-matching function to match colors between devices, such as a display and a printer, it specifies the profile for each device.

Device profiles are divided into three broad classifications:

• input devices, such as scanners and digital cameras

• display devices, such as monitors and flat-panel screens

• output devices, such as printers, film recorders, and printing presses.

Each device profile class has its own signature. The ColorSync constants for these signatures are described in ColorSync Manager Reference. For related information, see Devices and Their Profiles.

A profile connection space (PCS) is a device-independent color space used as an intermediate when converting from one device-dependent color space to another. Profile connection spaces are typically based on spaces derived from the CIE color space, which is described in Device-Independent Color Spaces. ColorSync supports two of these spaces, XYZ and L*a*b.

A color space profile contains the data necessary to convert color values between a PCS and a non-device color space (such as L*a*b to or from L*u*v, or XYZ to or from Yxy), for color matching. The ColorSync Manager uses color space profiles when mapping colors between different color spaces. Color space profiles also provide a convenient means for CMMs to convert between different non-device profiles.

L*a*b, L*u*v, XYZ, and Yxy color spaces are described in Color Spaces.

Abstract profiles allow applications to perform special color effects independent of the devices on which the effects are rendered. For example, an application may choose to implement an abstract profile that increases yellow hue on all devices. Abstract profiles allow users of the application to make subjective color changes to images or graphics objects.

A device link profile represents a one-way link or connection between devices. It can be created from a set of multiple profiles, such as various device profiles associated with the creation and editing of an image. It does not represent any device model, nor can it be embedded into images.

For more information on device link profiles, see CWNewLinkProfile and CMConcatProfileSet.

A named color space profile contains data for a list of named colors. The profile specifies a device color value and the corresponding CIE value for each color in the list. Profiles are typically stored as individual files in the ColorSync Profiles folder. For example, device-specific profiles provided by hardware vendors should be stored in the ColorSync Profiles folder. The location and use of the ColorSync Profiles folder has changed beginning in version 2.5. For a description of these changes, see Profile Search Locations.

Profiles can also be embedded within images. For example, profiles can be embedded in PICT, EPS, and TIFF files and in the private file formats used by applications. Embedded profiles allow for the automatic interpretation of color information as the color image is transferred from one device to another.

Embedding a profile in an image guarantees that the image can be rendered correctly on a different system. However, profiles can be large—the largest can be several hundred KB or even larger. A profile identifier is an abbreviated data structure that uniquely identifies, and possibly modifies, a profile in memory or on disk, but takes up much less space than a large profile. For example, an application might embed a profile identifier to change just the rendering intent in an image without having to embed an entire new profile. Rendering intents are described in Rendering Intents. For more information on embedding profile information, see Embedding Profiles and Profile Identifiers.

Profile Properties

Profiles can contain different kinds of information. For example, a scanner profile and a printer profile have different sets of minimum required tags and element data. However, all profiles have at least a header followed by a required element tag table. The required tags may represent lookup tables, for example. The required tags for various profile classes are described in the International Color Consortium Profile Format Specification.

Profiles contain additional information, such as a specification for how to apply matching. For more information, see Color Management Modules. Profiles may also have a series of optional and private tagged elements. These private tagged elements may contain custom information used by particular color management modules.

Profile Location

In most cases, a ColorSync version 2.x profile is stored in a disk file. However, to support special requirements, a profile can also be located in memory or in an arbitrary location that is accessed by a procedure you specify. See ColorSync Manager Reference for a description of ColorSync Manager structures for working with profiles that are stored in each of these locations. See Opening a Profile and Obtaining a Reference to It for information on working with profile locations in your application.

Setting Default Profiles

Prior to version 2.5, the default system profile (or simply the system profile) served as the default display profile; it also served as the default profile for color operations for which no profile was specified. The system profile had to be an RGB profile. A user could specify the system profile through the ColorSync control panel (formerly called the ColorSync<Superscript>™ System Profile control panel). If a user did not specify a system profile, then by default ColorSync used the Apple 13-inch color display profile.

Because the system profile was used for two dissimilar functions (default display profile and default profile for some RGB operations), there were several limitations:

• A user could not specify default profiles for color spaces other than RGB.

• A user could not specify separate profiles for more than one monitor.

• When matching an image without an embedded monitor profile, no matching occurred because the source and destination profiles were the same (system) profile.

Starting with ColorSync version 2.5, a user can set default profiles for RGB and CMYK color spaces, as well as for the system profile, using the ColorSync control panel shown in Figure 3-1. In addition, your application can call routines to get and set default profiles for the RGB, CMYK, Lab, and XYZ color spaces. As in previous versions of ColorSync, you can also call routines to get and set the current system profile.

Also starting with ColorSync version 2.5, a user can specify a separate profile for each monitor using the Monitors & Sound control panel. In addition, your application can call routines to get and set the profile for each display.

Because ColorSync version 2.5 provides capabilities for getting and setting default profiles for color spaces and for assigning a profile to each monitor, your application and anyone using it can more precisely specify source and destination profiles. For example, your application can set the destination profile for an operation to be the profile for a specific monitor or the source profile to be the default CMYK profile.

For information on getting and setting profiles in code, see ColorSync Manager Reference. Monitor Calibration and Profiles, describes how a user can specify a separate profile for each available monitor.

Where ColorSync Searches for Profiles

Prior to ColorSync 2.5, profile search routines such as CMNewProfileSearch looked for profiles only in the profiles folder (within the Preferences folder). Starting with version 2.5, the search routines look in the following locations:

• in the ColorSync Profiles folder (within the System folder)

• in first-level subfolders of the ColorSync Profiles folder

• in locations specified by aliases in the ColorSync Profiles folder (whether the aliases are to single profiles or to folders containing profiles)

With this searching support, you can group profiles in subfolders within the profiles folder (one level of subfolders is currently allowed). For example, you might store all scanner profiles in one folder and a variety of monitor profiles for your primary monitor in another. You can also store aliases to other profiles and profile folders within the ColorSync Profiles folder. ColorSync search routines will find all profiles in the specified locations.

The Profile Cache and Optimized Searching describes how your application can perform optimized profile searching with ColorSync 2.5.

Where ColorSync Does Not Look for Profiles

Because profile searching in ColorSync 2.5 can only go two levels deep, ColorSync search routines will not find a profile in the following cases:

• The profile is located in a folder that is within a folder in the profiles folder (requires more than two levels of searching).

• The profile is located in a folder that is within a folder specified by an alias in the profiles folder (again, requires more than two levels of searching).

• The profile is in a folder whose name starts with a parenthesis.

• The profile is specified by an alias to an unmounted volume (only mounted volumes are searched).

Temporarily Hiding a Profile Folder

To temporarily hide a folder from ColorSync’s search path, put parentheses around the name of the folder or the alias to the folder.

Setting a Preferred CMM

Starting with version 2.5, the ColorSync control panel, shown in Figure 3-1, lets you choose a preferred CMM from any CMMs that are present (that is, registered with the Component Manager—see Creating a Component Resource for a CMM for a description of how to create a CMM the ColorSync Manager can use).

If you choose a preferred CMM with the ColorSync control panel, and if that CMM is available, ColorSync will attempt to use that CMM for all color conversion and matching operations. If you specify Automatic instead, or if the specified CMM is no longer present or cannot provide the required matching service, or for versions prior to version 2.5, ColorSync follows an algorithm described in How the ColorSync Manager Selects a CMM to determine which available CMM to use for matching.

Note that if you want color conversion and matching operations to use the same CMM-selection algorithm they did in versions prior to ColorSync 2.5, specify Automatic in the ColorSync control panel. <8bat>u

Starting with ColorSync 2.5, your application can determine the preferred CMM by calling the function CMGetPreferredCMM.

Rendering Intents

Rendering intent refers to the approach taken when a CMM maps or translates the colors of an image to the color gamut of a destination device—that is, a rendering intent specifies a gamut-matching strategy. The ICC specification defines a profile tag for each of four rendering intents: perceptual matching, relative colorimetric matching, saturation matching, and absolute colorimetric matching. These rendering intents are described in Table 3-1.

Color professionals and technically-sophisticated users are likely to be familiar with the ICC terms for rendering intent and the gamut-matching strategies they represent. If your application is aimed at novice users, however, you may prefer to add a simplified terminology based on the typical image content associated with a rendering intent (for example, perceptual matching is commonly used for photographic images). Table 3-1 lists the ICC-specified rendering intents and the corresponding image content term, where applicable. Note that there is no simplified terminology for describing absolute colorimetric matching. However, novice users are not likely to need this rendering intent.

For information about the actual rendering intent tags, you can obtain the latest ICC profile specification by visiting either the ICC Web site at <http://www.color.org/> or the ColorSync Web site at <http://colorsync.apple.com/>.

General Purpose Color-Matching Functions

A general purpose color-matching function is one that uses a color world to characterize how to perform color-matching. General purpose functions depend on the information contained in the profiles that you supply when you set up the color world. You can define a color world for color transformations between a source profile and a destination profile, or define a color world for color transformations between a series of concatenated profiles.

Creating a Color World to Use With the General Purpose Functions provides a code sample for working with general purpose functions. Matching Colors Using the General Purpose Functions lists the general purpose functions and provides a description of each function.

In contrast to the general purpose color-matching functions, the QuickDraw-Specific Color-Matching Functions are tailored for color-matching with QuickDraw. Note, however, that you can also use the general purpose functions when working with QuickDraw—for example, if you need the greater level of control the general purpose functions provide.

QuickDraw-Specific Color-Matching Functions

A QuickDraw-specific color-matching function is one that uses QuickDraw to provide images showing consistent colors across displays. The ColorSync Manager provides two QuickDraw-specific functions that your application can call to draw a color picture to the current display:

• NCMBeginMatching uses the source and destination profiles you specify to match the colors of the source image to the colors of the device for which it is destined.

• NCMDrawMatchedPicture matches a QuickDraw picture’s colors to a destination device’s color gamut as the picture is drawn, using the specified destination profile. Uses the system profile as the initial source profile but switches to any embedded profiles as they are encountered.

Matching to Displays Using QuickDraw-Specific Operations provides a code sample for working with QuickDraw-specific functions. ColorSync Manager Reference lists the QuickDraw-specific functions and provides a description of each function. Note that the QuickDraw-specific functions call upon the general purpose functions to perform their operations, as shown in Figure 3-2.

Setting a Profile for Each Monitor

Because Monitors & Sound displays a panel for each available monitor, a user can also select, for each monitor, a separate profile from the list of available profiles. When a user sets a profile for a monitor in the Monitors & Sound control panel, ColorSync makes that profile the current system profile, as described in Setting Default Profiles. When your application sets a profile for a monitor, it may also wish to make that profile the system profile. If so, it must set the system profile explicitly by calling CMSetSystemProfile. For information on how to set monitor profiles in your code, see ColorSync Manager Reference.

Calibration

Calibration and characterization are related terms, but with important differences. Calibration is the process of setting a device’s parameters according to its factory standards. This is also known as linearizing or linearization. Characterization is the process of learning the color character of a monitor so that a profile can be created to describe it. Unlike input and output devices, whose calibration and characterization steps are not the same, a monitor is calibrated and characterized in one step.

The Calibrate button on the Monitors & Sound control panel provides the launching point for monitor calibration. A user can calibrate each monitor and create one or more color profiles for each, based on variations in gamma, white point, and so on. For related information on profiles, see Profiles and Devices and Their Profiles.

AppleVision and Apple ColorSync monitors are self-calibrating, so you will not see a Calibrate button for these monitors, unless there is a third-party calibrator installed in your Extensions folder.

Calibrating a monitor can be a challenging task for a naive user, but Apple Computer supplies a default calibrator that leads the user through a series of calibration steps. Using the default calibrator, even a novice should have a reasonable chance for success.

The calibration framework uses a plug-in architecture that is fully accessible to third-party calibration plug-ins. When a user clicks on the Calibrate button, the Monitors & Sound control panel provides a list of all available calibrator plug-ins. To appear in the list, a plug-in must meet the following criteria:

• It must be stored in the Extensions folder.

• It must be a shared library (file type 'shlb').

• Its shared library must export the symbols CanCalibrate and Calibrate.

• It should have a unique creator type (registered with Apple).

• The name of the library’s code fragment (specified in the 'cfrg' resource) must be unique (among all currently loaded shared libraries) and begin with 'Cali'. For example, you might want to name the library by appending your creator type to 'Cali'.

You can find source code for sample monitor calibration plug-ins in the ColorSync 2.5.1 SDK. If you plan to create a monitor calibration plug-in, you should read the next section Video Card Gamma.

Video Card Gamma

Starting with version 2.5, ColorSync supports an optional profile tag for video card gamma, which you specify with the cmVideoCardGammaTag constant. The tag specifies gamma information, stored either as a formula or in table format, to be loaded into the video card when the profile containing the tag is put into use. When you call the function CMSetProfileByAVID and specify a profile that contains a video card gamma tag, ColorSync will extract the tag from the profile and set the video card based on the tag.

If you provide monitor calibration software, you should include the video card gamma tag in the profiles you create. For information on the constants and data types you use to work with video card gamma, see ColorSync Manager Reference. You can get more information about AVID values from the Display Manager SDK.

Scriptable Properties

ColorSync provides scriptable support for getting and setting the following properties:

• system profile (the default system profile)

• default profiles for RGB, CMYK, Lab, and XYZ color spaces

• quit delay (the time in seconds for auto-quit, where 0 = never)

• profile location (a file specification)

For the following, you can only get, not set, the property:

• profile folder (the ColorSync profile folder)

Location is the only property currently supported for profiles, but future support is planned for additional profile properties.

Scriptable Operations

ColorSync supports the following scriptable operations:

• Matching an image.

• Matching an image with a device link profile.

• Proofing an image or a series of images.

• Embedding a profile in an image (very helpful for converting archives of legacy film).

Scriptable image operations currently work only on TIFF files, but support for other formats is planned.

Extending the Scripting Framework

The scripting framework uses a plug-in architecture that is fully accessible to third-party scripting plug-ins. When a user invokes a script to perform a ColorSync operation on an image, ColorSync (operating as a faceless background application) automatically builds a list of all available scripting plug-ins. It then attempts to call each of the plug-ins in the list until one of them successfully executes the desired operation. To appear in the list, a plug-in must meet the following criteria:

• It must be stored in the Extensions folder.

• It must be a shared library (file type 'shlb').

• It should have a unique creator type (registered with Apple).

• The name of the library’s code fragment (specified in the 'cfrg' resource) must be unique (among all currently loaded shared libraries) and begin with 'CSSP'. For example, you might want to name the library by appending your creator type to 'CSSP'.

Sample Scripts

The ColorSync SDK includes several sample scripts that demonstrate how to perform common operations. You can use the scripts as is, or borrow from them for your own custom scripts. For more information, see the detailed Read Me files that accompany the sample scripts.

When ColorSync Uses Multiple Processors

The default CMM takes advantage of multiprocessor support only if the following conditions are satisfied:

1. The MPLibrary was successfully loaded at boot time.

2. The CMM successfully links against the MPLibrary at runtime.

3. The number of processors available is greater than one.

4. The number of rows in the image is greater than the number of processors.

5. The source and destination buffers have the same number of bytes per row or have different locations in memory.

Unless all of these conditions are met, matching will proceed without acceleration. Multiprocessor support is currently supplied only for the following component request codes:

• kCMMMatchBitMap

• kCMMMatchPixMap

As a result, the default CMM invokes multiprocessor support only in response to the general purpose CWMatchPixMap and CWMatchBitmap functions, or when those calls are invoked as a result of a call to the QuickDraw-specific matching routines, such as NCMBeginMatching.

Efficiency of ColorSync’s Multiprocessor Support

Depending on the image and other factors, ColorSync’s matching algorithms take advantage of multiple processors with up to 95% efficiency (your mileage may vary). If you have two processors, for example, ColorSync can complete a matching operation in as little 53% of the time required by one processor. Additional processors should provide proportional improvement.