Color Munki use

You are not going to have good workflow without display calibration and printer profiles are not going to work. Junk in junk out.
To be clear, the printer profiles will absolutely work, the are not connected in any way to the display or it's profile. What will not work is a visual match. It would make deciding what rendering intent to select via a soft proof, far more difficult. That's a rather important selection to make depending on the image! I agree, it's not a good workflow not to have a calibrated and profiled display! But as I wrote earlier, if all you care about is the print and you can waste as much time, ink and paper, money as you wish to get that print, you don't have to have a profiled display. But I highly recommend everyone do so.
 
Well it seems there is at least one point of agreement - profiling the monitor does not affect the profiling of printer-paper-ink.

But for matching print to screen you need both outputs to be profiled, So much for "Color Munki use"

BTW, thank you for those instructive videos!
 
There is also the relation between the source of the image and the monitor. When a camera gives you a JPG in sRGB, you will not see what it is intended to be if your monitor does not display sRGB values accurately. Or working from the raw file, if you have your processing program output an image in Adobe RGB color space, you will not see it accurately if the monitor does not do a reasonably good job of displaying Adobe RGB.

Monitors vary. Some hold steady, while others drift over time, that is, they display an R,G,B trio today much differently than they did last month. In addition, most middle- and low-priced monitors do not reach a significant portion of the colors in Adobe RGB color space. It really makes a difference when you see your photo in Adobe RGB on a NEC PA model monitor, for example. But it also makes a difference to your credit card balance.

As for printer profiles, my experience is with those who report that profiles from the paper manufacturer deliver very good results. And for black and white, letting the printer driver do its thing is as good as or better than giving it an ICC profile.
 
There is also the relation between the source of the image and the monitor. When a camera gives you a JPG in sRGB, you will not see what it is intended to be if your monitor does not display sRGB values accurately.
Here we go again with the accuracy term without a metric. Chris didn't seem to get that either when he wrote and was called out by stating: drycreek profiles are 'accurate'.

Here's how you do it correctly without any speculation. You take a number of device values that HAVE to fall within the gamut of the device you're trying to define as accurate (so in this case, sRGB). The more the better to a point. Let's use 806!

Now you need to convert the values through each profile (sRGB and the display profile) to get Lab values the profile predicts. Then you get this report below. This is from my sRGB-like MacBook Pro display:

--------------------------------------------------

dE Report

Number of Samples: 806

Delta-E Formula dE2000

Overall - (806 colors)

--------------------------------------------------

Average dE: 0.62

Max dE: 3.19


Min dE: 0.00

StdDev dE: 0.49

Best 90% - (725 colors)

--------------------------------------------------

Average dE: 0.49

Max dE: 1.18

Min dE: 0.00

StdDev dE: 0.28

Worst 10% - (81 colors)

--------------------------------------------------

Average dE: 1.74

Max dE: 3.19

Min dE: 1.19

StdDev dE: 0.50

--------------------------------------------------

Is this accurate? Well with an average dE of 0.62 for 806 patches, yes! Under dE of 1 which is just a visible difference. But wait, there's more. The max dE, the one worst color match is 3.19 which for a display isn't bad but visible.

Easy to sort by worst dE per patch and we see orange and blues are the worst offenders:

Worst dE patches in accuracy between sRGB and an sRGB like display. Oranges and blues!

Worst dE patches in accuracy between sRGB and an sRGB like display. Oranges and blues!

BEWARE of anyone who talks of color accuracy without providing a metric of what that means! It's just color speculation and your mileage and numbers may (will very likely) vary.

The fact deniers in another forum didn't even address this video but for new folks lurking here:

Delta-E and color accuracy

In this 7 minute video I'll cover: What is Delta-E and how we use it to evaluate color differences. Color Accuracy: what it really means, how we measure it using ColorThink Pro and BableColor CT&A. This is an edited subset of a video covering RGB working spaces from raw data (sRGB urban legend Part 1).

Low Rez:

High Rez: http://digitaldog.net/files/Delta-E and Color Accuracy Video.mp4


Accuracy of color isn't subjective, it's a metric based on (in this discussion) colorimetry and the above testing is what's required to accuracy talk about color accuracy!

--
Andrew Rodney
Author: Color Management for Photographers
The Digital Dog
 
Not sure I understand. That product will calibrate and profile a display too.
That's why I profiled both. To improve "print to screen matching"

I just want to make clear that a profile of a printer-paper-ink ONLY can be as useful as a profile of the monitor ONLY. And to my eyes, with my equipment (2 rather outdated standard office monitors and a variety of different printers and ink technologies) printer profiling maybe more important than for most other users...
 
There is also the relation between the source of the image and the monitor. When a camera gives you a JPG in sRGB, you will not see what it is intended to be if your monitor does not display sRGB values accurately.
Here we go again with the accuracy term without a metric.
--
Andrew Rodney
(commercial link omitted)
To repeat: you will not see what an sRGB JPG from your camera is intended to be if your monitor does not display sRGB values accurately. This is true, and in the context of the thread, it supports the point that a monitor should be calibrated. That is all.

Can accuracy be examined quantitatively? Of course. Those who like to measure delta-E's as one metric are welcome to do so.

That is no excuse for Mr. Rodney's post to scream, "BEWARE of anyone who talks of color accuracy without providing a metric!" and to hurl insults about "fact deniers" in response to a post that is short, practical, and denies no facts.

For once, could this man participate in a thread in a civil manner?

--
The comment above probably has an example in my photostream at http://www.flickr.com/photos/41790885@N08/
 
There is also the relation between the source of the image and the monitor. When a camera gives you a JPG in sRGB, you will not see what it is intended to be if your monitor does not display sRGB values accurately.
Here we go again with the accuracy term without a metric.
--
Andrew Rodney
(commercial link omitted)
To repeat: you will not see what an sRGB JPG from your camera is intended to be if your monitor does not display sRGB values accurately.
Speculation. Further, if your goal is to soft proof to an output device, something some of us are talking about, the display calibration has nothing to do with sRGB!
This is true, and in the context of the thread, it supports the point that a monitor should be calibrated. That is all.
I'm the last guy who would suggest otherwise and I'm simply addressing the unproven discussion by some, talking about color accuracy. There's not an ounce of proof what is or isn't accurate or by how much.
Can accuracy be examined quantitatively? Of course.
Why can't you do that?
Those who like to measure delta-E's as one metric are welcome to do so.
Because we can and because it proves with science and measurements what we are talking about. Whose who can't simply speculate which is just unsubstantiated opinions devoid of facts or measurements.
That is no excuse for Mr. Rodney's post to scream, "BEWARE of anyone who talks of color accuracy without providing a metric!" and to hurl insults about "fact deniers" in response to a post that is short, practical, and denies no facts.
Because it's mostly nonsense. It's an assumption. It's not an accurate! It's not factual.

When you measure the foundation of a wall with your foot, because you assume your foot is 12 inches (and it's really 11), you'll find that proper measurement data is kind of useful.
For once, could this man participate in a thread in a civil manner?
What isn't civil about expecting a fact based answer or fact based text on the subject of color management?

IF I went into any photo forum here and stated that F22 is a wider aperture than F8, I'd be called out, corrected. As I should be for stating an opinion that has zero basis in fact. There's nothing uncivil about correcting misinformation!

You ARE entitled to your own opinion. You're not entitled to your own made up facts.

ac·cu·ra·cyˈakyərəsē/noun
  1. the quality or state of being correct or precise."we have confidence in the accuracy of the statistics"
    • technicalthe degree to which the result of a measurement, calculation, or specification conforms to the correct value or a standard.plural noun: accuracies"the accuracy of radiocarbon dating"
 
Last edited:
digidog wrote: Now you need to convert the values through each profile (sRGB and the display profile) to get Lab values the profile predicts. Then you get this report below.
Note: this was an exercise to illustrate how to define color accuracy and it does tell us the accuracy of the profile (what it predicts) vs. sRGB. It does not tell us anything about the accuracy of the display! We need to measure it with a spectroradiometer, ideally using the same number of color samples and do the comparison report again. Wanted to point that out! This was simply a quick and dirty illustration of the accuracy of the MacBook's ICC Profile, not the MacBook's current behavior.

You can do the same with printer profiles. You output a target and measure it. Then you compare the measurements of what was printed to what was predicted by the profile (and you can check how differing tables within the profile work). But again, the point is, any discussion of color accuracy needs a reference (what should the values be?) and some measurement of what was produced. Then you can give average, max, dE values and correctly define color accuracy without a lick of speculation!

My foot is 12 inches. It's accurate.

No, I just compared it to a ruler from Home Depot, it's 11.5 inches. Using my foot to assume what I measure is one foot is inaccurate. And we now know it's inaccurate by 1/2 an inch, a number of a specific, understood value of distance.

--
Andrew Rodney
Author: Color Management for Photographers
The Digital Dog
http://www.digitaldog.net
 
Last edited:
+1

I did see in his first post that he had made a hundred paper profiles.

Maybe he can hit the 1000 mark.
 
You are not going to have good workflow without display calibration and printer profiles are not going to work. Junk in junk out.
To be clear, the printer profiles will absolutely work, the are not connected in any way to the display or it's profile. What will not work is a visual match. It would make deciding what rendering intent to select via a soft proof, far more difficult. That's a rather important selection to make depending on the image! I agree, it's not a good workflow not to have a calibrated and profiled display! But as I wrote earlier, if all you care about is the print and you can waste as much time, ink and paper, money as you wish to get that print, you don't have to have a profiled display. But I highly recommend everyone do so.
 
I never said that I don't want to profile my monitor, I said that a monitor profile BY ITSELF does not affect the printer output and vice versa. It's the person in front of the monitor who does...

So even a profiled monitor doesn't prevent you from editing color cast.

Profile a certain printer-paper-ink-combination and tell me at which step the monitor would come into play!? Many users seem to believe this, but it is as wrong as the idea that while profiling a monitor the printer output would come into play (well, nobody believes the second, why do some believe the first?)

BTW, (@aikenmooney) with let's say 5 printers, 5 sorts of inks and at least 10 sorts of paper, how many profiles you would need?

--
Printers: Canon IP4000, MP810, Epson R285, L300 and pro3880;
Refill Ink from PC, octoinkjet and recently out of Fujifilm Drylab carts
Forum for Refillers: http://www.printerknowledge.com/
...please excuse my german english, schooldays are long ago...
 
Last edited:
I never said that I don't want to profile my monitor, I said that a monitor profile BY ITSELF does not affect the printer output and vice versa.
What display is that? And calibration and profiling are of course different.
So even a profiled monitor doesn't prevent you from editing color cast.
Correct. But many prefer having a display that visually matches their output so they don't do that.
Profile a certain printer-paper-ink-combination and tell me at which step the monitor would come into play!?
Each profile has two tables. One affects the output of course. The other affects the soft proof. That soft proof you see is based on this profile and the profile used to define the calibration of the display! Again, calibration is the step of 'adjusting' the display to a desired condition. Profiling defines or describes this to the color management system. So the printer profile and the display profile work hand in hand.
Many users seem to believe this, but it is as wrong as the idea that while profiling a monitor the printer output would come into play (well, nobody believes the second, why do some believe the first?)
You really should do both. Without, it's like one hand clapping. Again, if you don't care at all how the image appears on screen, you don't have to calibrate and profile it, you don't have to invoke a soft proof. And you don't have to pick an important consideration that the profile provides; what rendering intent to use when printing. Just print using RelCol, Perceptual and Saturation and pick the one you like. Time consuming and expensive.
BTW, (@aikenmooney) with let's say 5 printers, 5 sorts of inks and at least 10 sorts of paper, how many profiles you would need?
10 printer profiles, one display profile at a minimum but possibly 10 display profiles. Again, if you view the video I referenced on dark prints, you'll see there are products like my NEC SpectraView that allows me to calibrate to each paper and of course that means a profile for each calibration!

A matt paper and a glossy paper need differing calibration to produce a visual match on-screen. Assuming you care to see that match. Many of us do.

--
Andrew Rodney
Author: Color Management for Photographers
The Digital Dog
http://www.digitaldog.net
 
Last edited:
I never said that I don't want to profile my monitor, I said that a monitor profile BY ITSELF does not affect the printer output and vice versa.
What display is that? And calibration and profiling are of course different.
What display..?? I don't understand the question... said nothing about calibration... think I'm misunderstood
So even a profiled monitor doesn't prevent you from editing color cast.
Correct. But many prefer having a display that visually matches their output so they don't do that.
Profile a certain printer-paper-ink-combination and tell me at which step the monitor would come into play!?
Each profile has two tables. One affects the output of course. The other affects the soft proof. That soft proof you see is based on this profile and the profile used to define the calibration of the display! Again, calibration is the step of 'adjusting' the display to a desired condition. Profiling defines or describes this to the color management system. So the printer profile and the display profile work hand in hand.
Again you seem to misunderstand my Question, which is meant rethorical: The monitor does not come into play at all while profiling a printer-paper-ink.

How else could a service make a profile for you - with chartprints from your printer - using a complete different computer and monitor than yours?

Printer and monitor profiles indeed should work hand in hand, if correctly installed - BUT the profiling PROCESS is independend from each other, and so are the profiles: correct or not, good or bad, but never "knowing of each other". Because they are both output profiles, I have to repeat, and you won't do wrong when profiling a paper using an unknown even B&W monitor..
Many users seem to believe this, but it is as wrong as the idea that while profiling a monitor the printer output would come into play (well, nobody believes the second, why do some believe the first?)
You really should do both. Without, it's like one hand clapping. Again, if you don't care at all how the image appears on screen, you don't have to calibrate and profile it, you don't have to invoke a soft proof. And you don't have to pick an important consideration that the profile provides; what rendering intent to use when printing. Just print using RelCol, Perceptual and Saturation and pick the one you like. Time consuming and expensive.
Agreed (but again I spoke of the profiling process - you speak of how they should be used
BTW, (@aikenmooney) with let's say 5 printers, 5 sorts of inks and at least 10 sorts of paper, how many profiles you would need?
10 printer profiles, one display profile at a minimum but possibly 10 display profiles. Again, if you view the video I referenced on dark prints, you'll see there are products like my NEC SpectraView that allows me to calibrate to each paper and of course that means a profile for each calibration!

A matt paper and a glossy paper need differing calibration to produce a visual match on-screen. Assuming you care to see that match. Many of us do.
Now we seem to really talking past each other.. so I answer my question: 10 paper sorts, makes 10 profiles for 1 printer and 1 ink. would make 50 profiles for 5 printers, and 250 profiles for 250 printer-paper-ink-combinations. I'm less accurate, made only around 100 for my most used media (much experimenting with aquarell art papers)

Most printers I know don't have any option for "hardware calibration". You would need a special RIP, which cost more than a color munki. My only printer which is capable to do so, is a HP B8850. The "calibration" of a printer is mainly adjusting the ink limits for each color channel... When processing a profile with the color munki, you use to switch off any color correction - BUT you can not switch off the choice of one paper out of the printer driver's list: This is equivalent to "calibration", because the printer wiil use different ink limits. This is done before color correction, just as the settings for gloss optimizers or matte or photo black. (BTW what the colormunki photo does with printers is "profiling")
 
What display..?? I don't understand the question... said nothing about calibration... think I'm misunderstood
OK, I think I misunderstood the text, let it go.
Again you seem to misunderstand my Question, which is meant rethorical: The monitor does not come into play at all while profiling a printer-paper-ink.
Absolutely agree and I think I said so.
How else could a service make a profile for you - with chartprints from your printer - using a complete different computer and monitor than yours?
The process of making an output profile is separate from making a display profile. But as discussed, both play a role (indirectly).
Printer and monitor profiles indeed should work hand in hand, if correctly installed - BUT the profiling PROCESS is independend from each other, and so are the profiles: correct or not, good or bad, but never "knowing of each other". Because they are both output profiles, I have to repeat, and you won't do wrong when profiling a paper using an unknown even B&W monitor..
Yes, agreed, independent processes.
Now we seem to really talking past each other.. so I answer my question: 10 paper sorts, makes 10 profiles for 1 printer and 1 ink. would make 50 profiles for 5 printers, and 250 profiles for 250 printer-paper-ink-combinations. I'm less accurate, made only around 100 for my most used media (much experimenting with aquarell art papers)
IF you have 10 paper profiles, you might have 10 calibration settings for the display which also means 10 display profiles. If you had 10 papers that were very similar in their paper white, dynamic range and using the same inkset, you might only need 1 display calibration and thus 1 display profile.

If you had a paper with a vastly different paper white and contrast ratio from another, you'd ideally have two display calibrations, one for each, and therefore you'd have 2 display profiles too. So again, with 10 papers, the number of display profiles depends.

The huge advantage of a smart display system like SpectraView is:
  1. There's a tremendous amount of control over calibration such when soft proofing, you get a very good display to print match. The soft proof match isn't based solely on the output profile's preview table. Here's where the output profile and the display profile (and calibration) work hand in hand.
  2. You can build as many calibrations and thus profiles as you need for as many papers as you're using.
  3. You can switch on the fly from calibration and associated profile in the software to any other you've created.
The calibration takes place in the panel itself and controlled by the software. As you update the calibration of the display from paper A to paper B, the calibration for each is produced and the software loads the correct ICC profile that reflects the calibration. That's what a good reference grade, smart display system provides among other attributes.

You can calibrate to one aim point and hope the preview table in each ICC Profile will produce a soft proof that produces a good match to the print. But again, if the papers are vastly different, of you're working with differing printers, ONE calibration and profile doesn't fit all needs. That results in a mismatch between print and display. That's not ideal.
Most printers I know don't have any option for "hardware calibration".
It's not the printer per se, it's what drives that printer. The driver itself. There's virtually nothing to calibrate on most Epson's with the Epson driver. You can profile it of course.
You would need a special RIP, which cost more than a color munki.
Exactly. The so called RIP is the driver in this case.
My only printer which is capable to do so, is a HP B8850. The "calibration" of a printer is mainly adjusting the ink limits for each color channel... When processing a profile with the color munki, you use to switch off any color correction - BUT you can not switch off the choice of one paper out of the printer driver's list: This is equivalent to "calibration", because the printer wiil use different ink limits.
Not really IMHO. Calibration and optionally a different process, linearization, attempts to put the device in a desired behavior after which you profile that behavior. That's exactly what SpectraView is doing with it's multiple calibrations! Producing a behavior that results in a better visual match of print and display based on the paper and printer.

On my various Epson printers, I have no option with their driver to do anything but profile it's behavior. No calibration, no linearization. I could use a different driver (which may or may not be a RIP).
 
I have a color munki photo since one year. And made more than 100 printer profiles - one for the monitor. Which did not change very much. Other than the profiles for different papers, ink and printers... it will adjust the output of some unknown cheap but decent photo paper in a non-pro canon or epson a4 printer as well as in a 3880.

Printer profiles are much more important to me personally, they are my "must have" while a custom monitor profile turns out as "nice to have". Perhaps I will change my mind if I buy a better monitor.. :-O But one thing I want to underline: a profiled monitor by itself does not affect the printer output, neither does a printer profile change the monitor (except you use it for soft proof) - because both are output profiles, which are made independent from the other.
I am intrigued by a workflow which doesn't use a calibrated and profiled monitor. When you edit your image do you print a copy out after every edit so as to see what you've done?
 
I never said that I don't want to profile my monitor, I said that a monitor profile BY ITSELF does not affect the printer output and vice versa. It's the person in front of the monitor who does...

So even a profiled monitor doesn't prevent you from editing color cast.

Profile a certain printer-paper-ink-combination and tell me at which step the monitor would come into play!? Many users seem to believe this, but it is as wrong as the idea that while profiling a monitor the printer output would come into play (well, nobody believes the second, why do some believe the first?)

BTW, (@aikenmooney) with let's say 5 printers, 5 sorts of inks and at least 10 sorts of paper, how many profiles you would need?
 
Yes the monitor output and calibration will affect the printer output IF you use them to guide your efforts to adjust the appearance of the image. The adjusted image will print out differently than if it were not adjusted. Since that is the whole point of having an accurate monitor, I just don't understand what you are trying to say.
I'm saying nothing else:
I never said that I don't want to profile my monitor, I said that a monitor profile
does not affect the printer output and vice versa. It's the person in front of the monitor who does...

So even a profiled monitor doesn't prevent you from editing color cast.
So profiling your monitor should improve your editing - but it will not change the printer output of images which you leave as they are

And regarding the main topic "color munki use" I wanted to point out, that you CAN PROFILE a printer BEFORE or even WITHOUT profiling the monitor. Anyway I said it all...
 

Keyboard shortcuts

Back
Top