ICC Profile Evaluation Print ???

[[email protected]]

OK- I have built a profile for ColorBurst RIP (think its a good one). What do I use as an evaluation print and how do I set it up to insure I don't mess up the evaluation print numbers (and resulting output.)

I downloaded the G7 Digital Press Form 1- CMYK - to use as a test form. It has a section that uses CMY, K, and CMYK in to reproduce the same image in order to eveluate grey balance (all black, just CMY or cmy plus black.)

But I am perplexed about something- The G7 eval form is an untagged raw CMYK file- (without an icc profile.) The way ColorBurst is set-up- it always assumes an ICC profile for CMYK files if they aren't tagged with one. In my case - ColorBurst assummes that untagged CMYK files are Gracol 2006 coated #1.

If I drop the untagged file in to the RIP - it will take the raw CMYK numbers and process them assuming they are Gracol Coated #1. Is that correct?

I tried assigning the Gracol Coated #1 to the CMYK file before sending it to ColorBurst - but that had undesirable consequences- The idnetical images made from CMY, K and CMYK that were to be used for greyscale evaluation were all converted into cmyk - all with equal values all with a overload of magenta.

I gather I am missing some fundamantal understanding here- can someone please hekp me out?

Thanks: Kirk

 

[TerryWyse]

OK- I have built a profile for ColorBurst RIP (think its a good one). What do I use as an evaluation print and how do I set it up to insure I don't mess up the evaluation print numbers (and resulting output.)

I downloaded the G7 Digital Press Form 1- CMYK - to use as a test form. It has a section that uses CMY, K, and CMYK in to reproduce the same image in order to eveluate grey balance (all black, just CMY or cmy plus black.)

The G7 Digital Press Form....or any of the smaller G7 test forms....are good for evaluation, assuming G7 and GRACoL/SWOP colorimetry is your goal.

But I am perplexed about something- The G7 eval form is an untagged raw CMYK file- (without an icc profile.) The way ColorBurst is set-up- it always assumes an ICC profile for CMYK files if they aren't tagged with one. In my case - ColorBurst assummes that untagged CMYK files are Gracol 2006 coated #1.

If I drop the untagged file in to the RIP - it will take the raw CMYK numbers and process them assuming they are Gracol Coated #1. Is that correct?

That's correct. Untagged CMYK files will always get assigned the CMYK Input profile (should be called "Source" profile as ColorBurst continues to use incorrect terminology). If you have "Use Embedded Profiles" checked, untagged will still get assigned the source profile but tagged files will use their embedded profile as the source.

*Generally speaking*, when the application is CMYK press proofing, you do NOT want to use the "Use Embedded Profiles" checkbox as this could result in a completely different proof print than what you expect. This is NOT the case for RGB prints as you generally want ColorBurst to ALWAYS honor the embedded profile in this situation. The ideal workflow in my opinion would be to have a "Use Embedded Profiles" checkbox for each color space/model that way you could force the assignment of the CMYK source profile for proofing but honor embedded profiles for RGB printing.

I tried assigning the Gracol Coated #1 to the CMYK file before sending it to ColorBurst - but that had undesirable consequences- The idnetical images made from CMY, K and CMYK that were to be used for greyscale evaluation were all converted into cmyk - all with equal values all with a overload of magenta.

I gather I am missing some fundamantal understanding here- can someone please hekp me out?

My guess is Kirk that you didn't *assign* the GRACoL profile but instead *converted* to it prior to dropping it in ColorBurst. In any case, those CMY, K and CMYK elements will ALL get converted to CMYK when passed through ColorBurst no matter if you assigned/embedded the GRACoL profile or not.....converting everything to the output/printer profile (in this case a CMYK printer profile) is how color-managed proofing works.

As far as the "overload of magenta", I'm not sure what you mean. Are you simply saying the gray balance appeared magenta-casted? If so, it could be a result of poor printer/output profile in your ColorBurst environment or it could be poor/incorrect viewing conditions. Inkjet proofs will often appear to have a color cast, especially in the neutrals, when viewed under incorrect viewing conditions.

Regards,
Terry

 

[Stephen Marsh]

OK- I have built a profile for ColorBurst RIP (think its a good one).

Kirk, I think that it would be good to go back to basics, before the water is muddied with custom profiles...

I think that it is helpful to have a "known state" to compare to. Your custom profile is unknown.

Load in standard Epson media, that is supported for your ColorBurst Environment (printer profile for the target resolution on the target printer with Epson ink and media). Run the proof using this default/canned environment setting.

Use this "default/canned" print as a visual reference.

Then move on and compare the results to your custom profile with the same RIP colour settings using the custom profile print settings.


Hope this helps,

Stephen Marsh

 

[[email protected]]

OK- I have built a profile for ColorBurst RIP

OK- I have built a profile for ColorBurst RIP

Terry- I want to be sure about your response to one of my burning questions. You wrote:

"My guess is Kirk that you didn't *assign* the GRACoL profile but instead *converted* to it prior to dropping it in ColorBurst. In any case, those CMY, K and CMYK elements will ALL get converted to CMYK when passed through ColorBurst no matter if you assigned/embedded the GRACoL profile or not.....converting everything to the output/printer profile (in this case a CMYK printer profile) is how color-managed proofing works."​

I understand that when the LAB values go from the Profile Connection Space through the printer profile to the destination (printer) space they will be converted to CMYK values- However- what happens at the stage of assigning a source profile to the untagged cmyk values in the original file? I would think here that a value like 20 K (only black) would not be converted durring an assignment to a mix of cmyk. (Now that I have succinctly defined my question- it's pretty easy for me to check this out.)

 

[rich apollo]

what happens at the stage of assigning a source profile to the untagged cmyk values in the original file?

Nothing. The CMYK values are unchanged when an ICC profile is assigned. The ICC profile provides a context in which those values can be better understood.

The same CMYK values going down on cold-set newsprint, #1 grade coated paper (sheet-fed), and an Epson 9900 will yield different colors. The ICC profile assigned to a file tries to communicate the color associated with those CMYK values.

 

[[email protected]]

evaluation print comments

evaluation print comments

Thanks for the replies from everyone.

Rich- that us what I thought- assigning a profile will not change the cmyk values of the original file. However, the file will yield different LAB values in the PCS (Profile Connection Space) during the AtoB transform depending on the profile you assign it.

Terry- I am pretty sure I converted and not assigned the GRACoL #1 the first time I set up the test file for printing. I think that changed the cmyk values. I was working in Photoshop when I converted and although the original had no profile - I believe Photoshop assumed a profile and converted to GRACoL - remapping K values to cmyk values and more. I am going to check this again by test. If that is the case- sending these values in the AtoB transform would definetly chnage the LAB values.

 

[TerryWyse]

Thanks for the replies from everyone.

Rich- that us what I thought- assigning a profile will not change the cmyk values of the original file. However, the file will yield different LAB values in the PCS (Profile Connection Space) during the AtoB transform depending on the profile you assign it.

Exactly correct......at the moment you assign the profile, nothing at all happens but the moment you start talking about the PCS, things change....different profile assignments will not alter the CMYK values one bit but the conversion to the PCS is something else entirely....that depends on the profile that's being assigned.

Terry- I am pretty sure I converted and not assigned the GRACoL #1 the first time I set up the test file for printing. I think that changed the cmyk values. I was working in Photoshop when I converted and although the original had no profile - I believe Photoshop assumed a profile and converted to GRACoL - remapping K values to cmyk values and more. I am going to check this again by test. If that is the case- sending these values in the AtoB transform would definetly chnage the LAB values.

NOPE.....CMYK values will absolutely NOT change with a new or different profile *assignment*....if they changed then a *conversion* took place, whether it was intended or not. Again, we're just talking about assigning a profile at this point, not about the A2B/PCS conversion or anything else that might be happening downstream.

Terry

 

[[email protected]]

Evaluation Print setup topic

Evaluation Print setup topic

One thing I found out during testing yesterday is that Photoshop will always assign the working CMYK space to an untagged PDF file you are opening. Even if the Photoshop color settings are set to keep the existing color space when opening and ask if there is a mismatched or missing color profile.

When I opened the Digital Print Form PDF (non tagged)- Photoshop did not warn of a missing profile and did not ask if I wanted to apply a profile- it simply applied the working profile and opened the file. I confirmed with several tests. That was a surprise.

Terry- now the conversion instead of assignement becomes an issue. For example- my working color space in photoshop was set to SWOP coated when I opened the file. My erroneous conversion took the cmyk numbers assigned SWOP to the PCS and then out to GRACoL. The cmyk values of the file were altered during the conversion- and in fact represent different LAB values than if I had assigned GRACoL as the working space when opening.

Thanks for everyone's help- another piece of the puzzle and an educational process.

 

[meddington]

When I opened the Digital Print Form PDF (non tagged)- Photoshop did not warn of a missing profile and did not ask if I wanted to apply a profile- it simply applied the working profile and opened the file. I confirmed with several tests. That was a surprise.

if you're opening a PDf in Photoshop, your essentially rasterizing the file. you should have gotten an "Import PDF" dialog, and yes, if your mode was set to CMYKcolor, the working space is used by default and will be assigned.

 

[[email protected]]

Evaluation Print setup topic

Evaluation Print setup topic

reply to opening pdf- yes precisely- you get the "import pdf" dialog box and if you select CMYK as the colorspace for an untagged PDF you are opening, Photoshop assigns the working CMYK color space.

 

[Correct Color]

Kirk,

In answer to your original question: If I had to use only one file to evaluate a printer profile, it would be the Granger Rainbow.

Here's a little info regarding how I use it, including a link to one you can download:

Correct Color Commentary

I'll also point out the what-should-be obvious here that if you're really serious about learning color management, you'll be money ahead in the long run if you hire someone to come teach you.

Thanks for everyone's help- another piece of the puzzle and an educational process.

You're getting some good advice here, but keep in mind that you're getting it from people who have spent a lifetime learning this stuff. Trying to learn color management on message boards is about as complete an educational process as, say, attempting to learn to become a doctor by mail order.

Mike Adams
Correct Color

 

[[email protected]]

Evaluating ICC Profiles

Evaluating ICC Profiles

Hi Mike:

Grainger rainbow has been widely used since it was lauded in "Real World Color Management" in 2003. The authors; Fraser, Murphy and Bunting made the file available and did a pretty good job of explaining its use. I do appreciate getting your take on it.

I regularly use it on the RGB Profile side. It is an untagged RGB file- it has a wide color gamut especially compared to the CMYK gamut I work with on the prepress and proofing side. None the less it is still a useful visual reference although to not as great an extent as with an RGB profile because the gamut is reduced and areas posterized simply because of the cmyk conversion.

For CMYK prepress and proofing the Idealliance Digital Print Forms contain a lot of good imagery and targets for visual profile evaluation. They are very well designed- I especially like forms 3 and 4. Additionally measured evaluation of printed Idealliance characterization data sets can provide hard Delta E info regarding the quality of a profile's output.

On the matter of education- it is an ongoing process and even old dogs can learn new tricks. I started selling, supporting and training in Color Management as a dealer for Epson, ColorBlind, Kodak Colorflow and others in 1994 and haven't stopped since. I am not embarassed that I do not know everything, and always appreciate the process of learning more and gaining an in depth understanding. I love stepping behind the curtain- it help makes all the pieces of the puzzle fit together.

My Best
[email protected]

 

[Correct Color]

Kirk,

Grainger rainbow has been widely used since it was lauded in "Real World Color Management" in 2003. The authors; Fraser, Murphy and Bunting made the file available and did a pretty good job of explaining its use. I do appreciate getting your take on it.

Well, they did a reasonable job explaining how they used it. But that's not the way I use it.

I regularly use it on the RGB Profile side. It is an untagged RGB file- it has a wide color gamut especially compared to the CMYK gamut I work with on the prepress and proofing side. None the less it is still a useful visual reference although to not as great an extent as with an RGB profile because the gamut is reduced and areas posterized simply because of the cmyk conversion.

See, that's why I suggested if you're serious about learning color management, you might want to consider hiring a professional to come out and teach you. It's unfortunately possible to learn a very lot about color management online, and yet still not fundamentally understand it at all.

Yes, the Granger Rainbow Colorremedies has available for download--it's not the one I use but no matter--is untagged. The whole point is that it has no gamut until you assign a profile to it. Since it extends from 0 to 255 in R, G, and B, of course what it will do when you assign a profile to it is extend to the outer edges of the gamut of whatever color space you assign to it. That may be a very large gamut, or it may be a very small one.

And that's fine for proofing RGB devices by assigning their profiles to it as described in RWCM. However, that's not how I use it. And I'm not attempting to take anything away from RWCM by saying that either. It's a good book and just about everyone who's at all serious about color management has read it. However, it's getting a little old now, and it's actually pretty light on useful information regarding large-format inkjet. And I did make the assumption that if you're profiling in ColorBurst, you're profiling probably either an Epson or a VuTek, but an inkjet, regardless.

Believe me, there's a huge misunderstanding on the part or many who think they understand color management that the secret to properly profiling printing devices is can be found in the making of ICC profiles.

It can not.

Yes, there are better devices than others; better software than others; better charts and better methodologies to making good profiles and they're not inconsequential. But they're also just "wrench turning" so to speak.

Making a good profile involves creating the state of a machine, and then characterizing the machine in that state. It's the creating the state of the machine that determines how much of that machine's capabilities you capture with the profile, and how well you capture them. And that boils down to single channel ink limits, linearization, and multi-channel ink limits. As simple and as complex as that.

And while the Granger Rainbow as I describe using it doesn't tell a whole lot about linearization, it will instantly tell you if you've got any issues in the balance of your single-channel and multi-channel ink limits.

And believe me, I've been in many situations and seen many profile evaluation files over the years. And unfortunately, many of them can print out fine even though the profile itself might have some huge issues in this very area.

So, and again as I said, after visually inspecting the gamut of the profile, and after checking the curves in CTP, my next step is a soft proof of my own eval sheet which includes a Granger Rainbow. Any imperfections in its transition to black, and it's time to start over. And if there was only one file I could use for profile evaluation, the Granger Rainbow would be it.


Mike Adams
Correct Color

 

[[email protected]]

You "Fundamentaly Understand All of Color Management."- Impressive

You "Fundamentaly Understand All of Color Management."- Impressive

RE: Grainger Rainbow-

Well, they did a reasonable job explaining how they used it. But that's not the way I use it.

What you say on your site regarding evaluating with the Granger Rainbow is very similar to what they say in RWC. Regarding the color gamut- I misstated in technical terms regarding color gamut of the Grainger Rainbow being large (it is not defined until a profile is assigned- referenced by me correctly in this same topic on 07-18-2011 12:38 PM if you read it). However in terms of what you consider on your site as “GOOD” when evaluating the quality of a profile with a soft proof of the Grainger Rainbow – namely “no posterization, no untoward artifacts; and in particular very good transitions- these qualities are much more difficult to achieve in CMYK prepress spaces (my point.)​

In spite of your addressing me in what I consider a disrespectful manner (you have no idea of who I am and what I know) - I will experiment with evaluating CMYK profiles with the GR to see what I think- for my own benefit.

To Mike Adams- Color Correct- something to consider:

It is often those that proclaim the loudest how much they know, that have no idea how much they don’t.