Color Management Myths 35-39

From ColorWiki

(Difference between revisions)
Jump to: navigation, search
m
m (Myth #37: Lab is perceptually uniform)
Line 25: Line 25:
When you open up the ColorThink Grapher and view a gamut of a profile, you're viewing it in a representation of CIE L*a*b space.  Lab is supposed to be perceptually uniform, so that a distance between any two colors in one part of the color space, gives the same basic difference as the same distance in another part of the color space.    For most purposes, it works well for what it is, but there are times when it is good to keep in mind that this Lab space has a subtle warp to it.  You'll see this in action any time you bring a printer profile into the ColorThink Grapher and observe how the blue end of the spectrum twists around in a slight counter-clockwise direction.  In most cases, it's not that your printer's behavior is skewed, it's that a good profile will have this "twist" in the shape in order to give you good color in a twisted Lab space.  This is sometimes the reason that older profile building software would give you blues that turned slightly purple, yellows that were green and reds that were orange.
When you open up the ColorThink Grapher and view a gamut of a profile, you're viewing it in a representation of CIE L*a*b space.  Lab is supposed to be perceptually uniform, so that a distance between any two colors in one part of the color space, gives the same basic difference as the same distance in another part of the color space.    For most purposes, it works well for what it is, but there are times when it is good to keep in mind that this Lab space has a subtle warp to it.  You'll see this in action any time you bring a printer profile into the ColorThink Grapher and observe how the blue end of the spectrum twists around in a slight counter-clockwise direction.  In most cases, it's not that your printer's behavior is skewed, it's that a good profile will have this "twist" in the shape in order to give you good color in a twisted Lab space.  This is sometimes the reason that older profile building software would give you blues that turned slightly purple, yellows that were green and reds that were orange.
-
 
==Myth #38:  Delta E 2000 is not a good equation to choose; I've heard that xxx is better.==
==Myth #38:  Delta E 2000 is not a good equation to choose; I've heard that xxx is better.==

Revision as of 20:36, 14 August 2013

Contents

Color Management Myths #35 - 39

It's been a while since we added to our collection of color management myths. Here we take on some of the common misperceptions that we hear from various sources & explain what the truth is and why. A little advice: Having some foreknowledge from the mistakes that other people make is a particularly easy way to learn valuable information in a short amount of time. Find an extensive list of our previous myths gleaned from our newsletters here.

Myth #35: Wide gamut monitors are the best!  Everybody should have one.

Modern backlighting and LCD panel technology has made it possible to make displays that can produce much more saturated colors, that have a larger gamut than the typical displays we're used to. Obviously it's great to have a display that is capable of producing more colors in order to view what you're sending to a printer that also has an enlarged gamut.

So they really are marvels of modern science, but they are not for everyone. If you work only in web design, then you will want to keep to the sRGB gamut which is the standard for internet images. You'll have no use for all that added saturation. It could naturally even give you a false sense of how colorful your images are. Most of these wide gamut displays can also be "dumbed down, " can be calibrated to emulate sRGB, but then the question is why get one if you're not going to take advantage of the benefits of a wide gamut?

Wide gamut displays also carry the possibility for banding because with the common 8-bit workflow most of us are working with, there's only barely enough steps to smoothly define all the colors in a normal display. So that adding a whole extra range of colors to define makes it more likely that some banding might occur. This is generally not a big problem for displays with built-in graphics capabilities like Eizo and NEC, but the possibility is something to be aware of. More information on banding is here.

Myth #36: A wide gamut monitor will show me all the colors my printer can print.

I know what you're thinking…. Finally! Now that I have a wide gamut display, something that can display 98% of the AdobeRGB gamut, I can finally truly soft-proof all the colors that my printer can print!

Hey, I don't blame you for falling for this myth. The size of these gamuts are breathtaking. AdobeRGB is a large color space and these displays get you almost all of it. But printers have been gaining ground in developing new ink sets and they have been producing more saturation in recent years too.

Keep in mind also the fundamental nature of these two coloring systems. A printers's gamut is shaped something like a round hole and a monitor's gamut is shaped like a 3 cornered peg. And you know what they say about putting an odd-shapped peg in a round hole.

Like most things in the world of color management, it all depends… In this case, your printer gamut will depend on your printer, your inks and especially your paper. And we often find that a modern inkjet printing onto good quality glossy paper can produce an abundance of cyans and yellows that are well outside the gamut of your AdobeRGB display.

Myth #37: Lab is perceptually uniform

When you open up the ColorThink Grapher and view a gamut of a profile, you're viewing it in a representation of CIE L*a*b space. Lab is supposed to be perceptually uniform, so that a distance between any two colors in one part of the color space, gives the same basic difference as the same distance in another part of the color space. For most purposes, it works well for what it is, but there are times when it is good to keep in mind that this Lab space has a subtle warp to it. You'll see this in action any time you bring a printer profile into the ColorThink Grapher and observe how the blue end of the spectrum twists around in a slight counter-clockwise direction. In most cases, it's not that your printer's behavior is skewed, it's that a good profile will have this "twist" in the shape in order to give you good color in a twisted Lab space. This is sometimes the reason that older profile building software would give you blues that turned slightly purple, yellows that were green and reds that were orange.

Myth #38: Delta E 2000 is not a good equation to choose; I've heard that xxx is better.

Delta E is a unit of measure that indicates how far apart two colors are. A delta E of 1 is the smallest change between two colors that the trained human eye can see. Delta E values are thrown around a lot in the industry when you're talking about how far off a colors is. We have a little history of the development of delta E in a previous ColorNews article. DeltaE 76 was the first method of calculation developed soon after the CIE created the L*a*b* space in 1976.

It was a simple formula and it turns out that it was not consistent in all areas of Lab space. (See the myth "Lab is perceptually uniform" above.) A solution to this was published soon after 1994 called delta E 94. Many people still swear by dE 94, but it has the drawback of not being symmetric in its calculations. It does not work the same way both ways. Using dE 94 you can get a different results when you compare color A with color B - than you get when you compare color B with color A. Occasionally we hear from people who have connections with the garment industry that they like to use a calculation called CMC.

With delta E 2000 (developed around the year 2000), the general structure of dE 94 was taken and improved. It is the latest thing in color distance calculation. This is a well-proven formula and the official word from the CIE (the official body that make rules about these things) is that CIE2000 is officially the recommended color-difference equation. It has been shown to be more accurate than dE94 and CMC.

http://onlinelibrary.wiley.com/doi/10.1002/col.1049/abstract
http://www.ncbi.nlm.nih.gov/pubmed/15603061

Myth #39: Curve3 will make my printer print GRACoL

We have touched on this in our article on G7 Myths, but this issue comes up often enough that it is worth clarifying. Curve3 (as well as Curve2 and the original IDEAlink Curve) has some very useful functions besides just giving you G7 output curves. The "Analyze" tab will allow you to compare your press with the actual ink requirements for specifications like GRACoL and SWOP.

You can look at the "spidergraph" and see how your ink colors lay out. The shape of the "arms" of the graph give you an idea of how smooth your ink curves are, as they stretch out in the direction of the bullseye circles that represent where your color should end up according to the specification that is selected.

Curve3 is fundamentally a linearization tool. It helps you define your curves from 0 to 100. But note that you always start at 0 and end up at 100.

People get into trouble when they look at the 100% target bullseye in the spider graph and somehow expect the Curve software to yank the solid ink color over to where the bullseye specification is. There is no amount of curve adjustment that can ever make an ink solid change its color. (That would literally be like a leopard changing its spots.) The color of an ink solid is determined primarily by the color of the ink itself, as well as your choice of ink limits, with a little influence by the color and nature of the paper or substrate and how it reacts with the ink.

So Curve3 will do a fabulous G7 linearization for you, and it will tell you how close you are to GRACoL, but you have to take care of some of the basics if you are aiming for a specification like GRACoL.

Thanks for reading,

Patrick Herold CHROMiX

Personal tools
Namespaces
Variants
Actions
Navigation
Toolbox