Wednesday, 7 April 2010

Assigning and Converting Colour Spaces

All this colour management is difficult to understand.  I’ve found the following helpful from:

Assigning Profiles

The main goal of assigning an ICC profile to an image is to tell Photoshop what the RGB (or CMYK) values mean, so that the colours and tones in your image can be displayed as accurately as possible. When you assign a profile you are not changing the red, green, and blue code values; rather, you’re changing how those values are interpreted by Photoshop, and how the image will appear.

To tell Photoshop what the RGB values mean, launch the “Assign Profile” window from “Edit” in the menu bar in Photoshop CS2 or from “Image” then ”Mode” in Photoshop 6, 7 and CS.  As you select different profiles in this pull-down interface with the Preview checked, you will notice that the image visibly changes with each profile selected. Each profile tells Photoshop that the RGB values in the image correspond to different tones and colours or Lab values.

The goal of assigning a profile is to tell Photoshop what the RGB values mean so that the colours and tones of an image can be displayed as accurately as possible.

Practical reasons to assign a profile include:

  1. Missing Color Space Profiles from Digital Camera or Other Files: If a digital camera does not embed a working color space to its jpeg files or the file you are opening doesn’t have an embed a profile, but you know the images are processed to sRGB, AdobeRGB(1998), or a generic RGB, just assign the sRGB, AdobeRGB(1998), or generic RGB to get the color right.
  2. Improved Accuracy from Input Devices: By assigning custom scanner profiles to their corresponding images, you get improved color matches to the original prints or transparencies you are scanning. By assigning custom digital camera/light source profiles to their corresponding images, you can get colors that better match the colors of the original scene.

However, you should NOT assign a printer or output profile from a lab or a printer or paper manufacturer. This is when you use the “Convert to Profile” tool.

Converting to Profiles

As opposed to assigning a profile, when you perform a conversion you go from one colour space into another. Here, the goal is to change the RGB values, so that the colours are kept the same.

To convert an image in Photoshop we launch the “Convert to Profile” window from Edit in the menu bar. 

In performing a conversion, we have three options.

  1. The first is the colour engine or colour management module (CMM). This does the math under the hood when a conversion is performed. At the moment this selection does not have a big impact on the conversion and Adobe (ACE) is a good general choice.
  2. The second option is rendering intent. The rendering intent controls gamut compression or how in-gamut colours, that can be reproduced, are handled during a conversion. Out-of-gamut colours, that can’t be reproduced, are always brought into the gamut of our destination space, but with photographic images we have two choices for handling the in-gamut colours:
    1. We can shift the in-gamut colours to maintain the overall relationship of colours, called Perceptual, or
    2. We can keep the in-gamut colours unchanged or as close as possible, which is called Relative Colorimetric. Using the Preview check box, you can decide which intent you prefer for the image(s).
  3. The final options are two check boxes. “Use Black Point Compensation” is checked to help maintain shadow detail. It is especially important to have this checked with colorimetric rendering intents. “Use Dither” is checked to help maintain smoothness in gradations, like the sky of this image.

Practical reasons to convert to a profile could include:

  1. Putting Images into Working Color Spaces: Before working on images in Photoshop, it is better to have them in standard, independent, neutral colour spaces, like sRGB, ColorMatchRGB, or Adobe(RGB). One reason is so that the Info Palate’s RGB values will have more meaning. When the red, green, and blue values are equal in one of these spaces the colour is a true neutral. If you are in a device space, such as scanner or monitor space, this is, most likely, not the case and values can’t be used.
  2. Preparing Images for Lab Printing: As I discussed in my May article, you can convert using your lab’s printer and paper profile before sending images. But remember it’s important to tell them not to make any adjustments or conversions to these images once you have already converted them.
  3. Preparing Images for the Web: If you are saving files for the Web, it’s best to convert the images to sRGB, which is currently the standard colour space for this industry. Make sure to embed the profile when saving the image.

Another place Photoshop performs conversions is in the “Print with Preview” window. You select the option “Let Photoshop Determine Colours” under the Colour Management tab and select the printer paper profile. This could be a generic profile or, preferably, a custom profile built by you or a professional. The conversion to the printer and paper colour space happens on the fly as you send the image to the printer. This will help you get the best match from monitor to your printer and paper combination.