Input Profiles and Working Spaces

From ColorWiki

Jump to: navigation, search
Reserved Article

This page is a
Reserved Article.
For more details see
Reserved ColorWiki Articles


This reserved article originally appeared in CHROMiX ColorNews Issue 8 on July 15, 2003.

Click here to see the original in its original context.
colornews(at) to subscribe to the ColorNews newsletter.

by Steve Upton


Input Profiles and Color Space Conversion Guidelines

This issue's article is on the subtle but important issue of bringing files into your workflow.

Let's say you have a great scanner profile that you took time to create with an accurate target and it is delivering results that you love AND it saves you tons of time.

When you bring files from your scanner into Photoshop you diligently assign the scanner profile. Once Photoshop opens them, your scanner profile links up with your display profile and the image you see on your display is bright, accurate and wonderful.

Convert to Working Space?

Should you convert it into your working space or not? or when? or... why?

First, the similarities... let's compare an original scan file and its embedded scanner profile to the same file after you have converted it to a working space, say Adobe RGB (1998). Both files will have their embedded profiles used when displaying on your screen, creating soft proofs and printing. Even though the files contain different RGB values for each respective pixel, the included profiles convert them to approximately the same color so they both look good. This is all fine and dandy and the way things should work normally.

So, you are asking... I'm several paragraphs into this article and I still don't understand why he brought it up...

Fair enough, now with the differences.

"Scanner-space" Advantages

First, the advantages of leaving your file in "scanner-space". When ever you convert a file, you lose accuracy. So the fewer times you convert on the way to print, the better. It's easy to see that the scanner->print path is shorter than the scanner-> working space->print path. Also, scanner gamuts are typically larger than working space gamuts. If your primary destination is print, then the important question is how many scanner colors will print well on the printer and not be "clipped" or otherwise altered when printed. If you move your file through a working space you run the risk of clipping even more colors - those that were in the scan and can print but don't fit through the "keyhole" of your working space. This may be a good time to graph your scanner, working space, and print profile in ColorThink to see if you should choose a different working space...

Now that you have your head around scanner-space advantages and it seems like a great idea, I'm afraid I'm going to turn it all around....

"Working-space" Advantages

Converting your scan file to your working space has a bunch of advantages as well. First, scanner profiles are filled with 3D look-up-tables (LUTs) that are fairly large, so embedding a scanner profile can add 300-600 KB to each file. Working space profiles are typically matrix-based so they consist of only a few bytes of color information and are often 3-4 K in length - negligible. On large files this may not be a big difference but if have many files in your workflow, adding 400 KB to each file can plug things up.

Second, scanner profiles are made to convert from scanner RGB to Lab. Only. They are a one-way trip out of scanner-land and there are no tables in the profile to convert back (at least according to the ICC spec). If you had a scanner RGB file in Photoshop and you wanted to paste an image from another color space into your file, you would be unable to. If you wanted to specify a color using Lab in the color picker you could not.

Third, scanner space is non-uniform. For example. The RGB setting 100,100,100 coming from your scanner is probably not gray. That's OK as your scanner profile has already figured out that your scanner is actually seeing gray when it scans RGB = 100,105,92. When your scanner profile does its job all is fine. But lets say you want to put a gray border around the edge of your scanner RGB-based image. You click into the color picker and specify a nice middle gray such as 125,125,125. But as soon as you start applying this gray in your image it looks strange - and decidedly NOT gray. This is the same problem in reverse. 125,125,125 in scanner RGB is NOT gray and trying to figure out what is gray is not worth the effort. If you convert your file to Adobe RGB then 125,125,125 is gray as expected and all is fine. This gray example is only one case of this problem. This issue exists throughout the color space and is even tougher to figure out in other colors.

The process of converting a file from a non-uniform device space to a uniform working space is often called "normalizing" the file. It's a bit like running a comb through your hair in the morning, a good idea. So it is fair to say that a file in scanner RGB is not really in a working space at all as much of the work you might want to do is impossible.

The final benefit of converting to a working space is when people mess up your files. If you send a scanner RGB file to someone who views in in their web browser or (gasp!) strips the profile off the file when opening in Photoshop it is going to look bad. Scanner RGB mistreated as monitor RGB is not pretty (see gray problems above). A file in Adobe RGB will not look right as monitor RGB or sRGB but it won't look as bad. I know that's little comfort but sometimes that's what happens to our carefully created files. Now that I have fully confused you I should give you some tasty bit of wisdom to help you venture out in the world with this new-found knowledge. Unfortunately this issue is like many other issues in which answers like "it depends" rule the land.

If you are concerned about the greatest fidelity possible and will perform minimal editing on your file, leaving your file in scanner RGB is probably the best choice.

If, like most of us, you want to hand the file off downstream in RGB digital form, are going to perform more than the most cursory edits, are planning on compositing multiple images together, or just want gray to be predictable, then conversion to an appropriate working space is a good idea. For more information about whether or not you are using the best working space for your workflow, look into ColorThink. 2D and 3D graphs comparing scanner, working space and print gamuts will quickly show you what you need to know.

Personal tools