How to tell if a CMYK build printed as requested

[HarveyDunn]

Newbie question here: lets say I asked a print shop to give me a passage that was Cyan 30 - Yellow 100, GRACoL profile. I do not have a physical swatch to show them, and I do not have a LAB value to give them. I just know that I want a passage that is Cyan 30 - Yellow 100, GRACoL profile. Is there an instrument or technique that can analyze the press sheet to see if that is what is really being produced?

 

[jonteo]

You can use a spectrophotometer to measure the screen values, allowing for dot gain on the press for that stock. With access to the colour bar on the press sheet, you can also verify if the job was printed to Gracol specs (G7 - solid ink Lab values, pre G7 - TVI, PC etc).

 

[HarveyDunn]

You can use a spectrophotometer to measure the screen values, allowing for dot gain on the press for that stock.

How does that work? I though all you could get from a spectrophotometer was the LAB value of the sample - and as I said, in my hypothetical, I don't know what the LAB value of the swatch is supposed to be.

 

[Colorblind]

I will let the color management experts correct me if I'm wrong but my understanding is: Cyan 30 - Yellow 100, GRACoL profile translates to L:77 a:-20 b:72 absolute colorimetric (according to Adobe CMS engine) so if your print shop already prints to Gracol specifications, then using a loupe on a print sheet, you would see only solid yellow and a screened value of Cyan and reading the printed patch with a spectro should result in something relatively close to L:77 a:-20 b:72. Since what you really want to achieve is L:77 a:-20 b:72, a printer that's not printing to Gracol specs or that would want to simulate that color on another device (inkjet proofer, laser color copier) would need to convert your Cyan 30 - Yellow 100 to their printing device color space (what we call color management) in order to achieve the same end color result of L:77 a:-20 b:72. You would most likely see a different mix of ink on the sheet but get the same close spectro reading. The key here IMO is to communicate clearly that you want L:77 a:-20 b:72 and not only Cyan 30 - Yellow 100

 

[HarveyDunn]

I see. So even though I did not supply a LAB value, the print shop would, on its own initiative, look up what the LAB value should be for the requested build, and use that for their analysis. Correct?

 

[jonteo]

On my Beta Spectrodensitometer (not spectophotometer), there is a mode to measure printed screen values.

What's spec'ed as Cyan 30 - Yellow 100 will most likely not be those values on the printed sheet. If the printer is printing to some sort of specification, he would have applied a plate curve to correct for ink (different ink Lab colours from different ink manufacturers), plate, paper characteristics and other press variables to achieve the correct Lab colours.

 

[HarveyDunn]

I've never heard of that instrument type before. Interesting.

"What's spec'ed as Cyan 30 - Yellow 100 will most likely not be those values on the printed sheet. "
Also very interesting. Which leads me to another question. I have often wondered why we spec colors as CMYK builds instead of LAB values. Other than the danger of typing in a color that is out of gamut, that is. I can easily see how that is a danger. But other than that - why not?

 

[jonteo]

I see. So even though I did not supply a LAB value, the print shop would, on its own initiative, look up what the LAB value should be for the requested build, and use that for their analysis. Correct?

Most of this will be automated through a workflow. They will not convert CMYK to Lab in your digital file. They calibrate the press and build a colour profile based on press variables (ink, paper, plate etc) and use the colour bar on the press sheet to verify and adjust.

 

[jonteo]

I've never heard of that instrument type before. Interesting.

Here's a link to their website: Color Densitometers for Printers
Most of them will measure dot area.

I have often wondered why we spec colors as CMYK builds instead of LAB values. Other than the danger of typing in a color that is out of gamut, that is. I can easily see how that is a danger. But other than that - why not?

Hmm... you can spec colours as Lab but they get converted to CMYK at some point in the workflow. Sorry, I don't have an intelligent answer.

 

[gordo]

I honestly don't believe that measuring a color patch on a press sheet will give you a Lab value the same as the Lab value derived from the GRACoL profile according to Adobe CMS or any other engine. It might, or it might not. There are just too many variables in production printing.

Newbie question here: lets say I asked a print shop to give me a passage that was Cyan 30 - Yellow 100, GRACoL profile. I do not have a physical swatch to show them, and I do not have a LAB value to give them. I just know that I want a passage that is Cyan 30 - Yellow 100, GRACoL profile. Is there an instrument or technique that can analyze the press sheet to see if that is what is really being produced?

Not sure what you mean by "passage" but if you're looking at a press sheet then you would cut the press sheet through the area of colour that you are concerned about and, in the press operator's light booth, lay one half over the proof that you signed off on and examine for colour alignment ("match").
If your specifying the color and need a guide then you might use a color atlas either provide by the printshop (rare) or a commercially available one like Heidelberg's (http://the-print-guide.blogspot.ca/2009/12/heidelbergs-color-atlas.html).
But no matter how you go about it - there will be some degree of color mismatch.

They calibrate the press and build a colour profile based on press variables (ink, paper, plate etc) and use the colour bar on the press sheet to verify and adjust.

I don't believe that any production printer builds a colour profile based on their press variables nor uses the colour bar on the press sheet to verify and adjust. But maybe I'm misunderstanding you?

I've never heard of that instrument type before. Interesting.

"What's spec'ed as Cyan 30 - Yellow 100 will most likely not be those values on the printed sheet. "
Also very interesting. Which leads me to another question. I have often wondered why we spec colors as CMYK builds instead of LAB values. Other than the danger of typing in a color that is out of gamut, that is. I can easily see how that is a danger. But other than that - why not?

The reason is probably partly historical - since back during the steam-powered days, areas of specified screen tint values would be assembled from pieces of pre-halftoned film.
You can specify color using Lab values - however I don't know of a practical way of guaranteeing that you end up with that Lab value in the presswork.

best, gordo

 

[Schnitzel]

Working with Lab is not intuitive. When I tell you to add 5% cyan or subtract 10% yellow, you can have a basic concept of the impact on the graphic file or print. But if I tell you to add 8 to the a* and subtract 5 from the b*, that's harder to imagine.
Also, working with Lab will prevent you from getting "pure" colors on press (e.g., 50% cyan only), and from getting colors consisting of black only (all conversions between Lab and CMYK produce a rich black. The only workaround is a device-link ICC profile). Sometimes people are not interested in an exact colorimetric match, but rather in specific CMYK numbers, whatever the visual result may be.

 

[jonteo]

I don't believe that any production printer builds a colour profile based on their press variables nor uses the colour bar on the press sheet to verify and adjust. But maybe I'm misunderstanding you?

What I meant was the printer would profile their press to a standard and when your job comes through, it goes through their workflow software which applies corrections (is this the right terminology? I call it a "curve" to adjust for dot gains) based on that printer's equipment to visually (measured by tools) match the output from the press to that standard (Gracol, SWOP etc). The colour bar is used to verify they are printing to the specifications.

 

[gordo]

What I meant was the printer would profile their press to a standard and when your job comes through, it goes through their workflow software which applies corrections (is this the right terminology? I call it a "curve" to adjust for dot gains) based on that printer's equipment to visually (measured by tools) match the output from the press to that standard (Gracol, SWOP etc). The colour bar is used to verify they are printing to the specifications.

The terminology is wobbly but your meaning is clearer. I'll PM you.

best, gordo

 

[Lukas Engqvist]

Do you want the colour or do you want the values? Very rarely can you have both. Going only by numbers is like boiling an egg by a clock… as long as you are exactly at the same altitude it will be boiled the way you want… but go up a mountain and it will be too soft put it in a pressure cooker and it will be too hard. Exact same time but different conditions.
With ink it's more complicated… but with a correct colour management there is a fair chance to predict what to expect, and to get there (within reason). You have great advice in the other posts. i hope I have not confused you but I guess what I'm trying to convey is the real world is complex, and we need to submit to how it behaves.

 

[HarveyDunn]

This is all very interesting and helpful, thanks to all of you.

As we know, there are many different CMYK builds that can produce the same color. Lets say that I've got some very light (L value higher than 80) but somewhat dull colors in my design. I don't want any black in my build, because sometimes it makes light colors look a bit dirty (at least, that has been my experience - as always, please correct me if I am wrong). Therefore, I make a point of choosing builds with no black. For example, C6 M6 Y12 K0 instead of C3 M12 Y9 K3.

Is that actually what I'm going to get in my output? Or is the software at the printshop going to convert my carefully selected build into LAB and then convert it back to CMYK using its own color engine which may in fact put back that little bit of black that I was trying to take out?

 

[HarveyDunn]

PS: Sheetfed job, using GRACoL profile.

 

[gordo]

Is that actually what I'm going to get in my output? Or is the software at the printshop going to convert my carefully selected build into LAB and then convert it back to CMYK using its own color engine which may in fact put back that little bit of black that I was trying to take out?

Unless your files are specifically being reseparated (typically for newspaper application, less so for magazines, and rarely for general offset) your input cmyk screen tint builds are not changed. However they will not be the same on plate or presswork.

Gordo

 

[Stephen Marsh]

Or is the software at the printshop going to convert my carefully selected build into LAB and then convert it back to CMYK using its own color engine which may in fact put back that little bit of black that I was trying to take out?

Is this a theoretical question, or do you see K highlight dots in the print when using a loupe or microscope?

The source CMYK numbers are translated through the Lab colour space into the destination colour space. This is not the same thing as actually converting a file into the Lab colour space - however it does cause the loss of the original CMYK data and black could be introduced when the job is converted to the destination space.

This could happen if a print shop decided to run a ICC device profile conversion on all incoming CMYK to convert it to another type of CMYK. I would hope that this is not so common.

What may be "more common", is that a print shop may have "ink optimizing" or "normalizing" (Max GCR) iCC DeviceLink profile conversion software that is changing source CMYK values direct to destination CMYK values (DeviceLink profiles do not use Lab, there is no PCS or profile connection space, two CMYK profiles are linked directly together with a rendering intent and other options above and beyond normal device profiles).


Stephen Marsh

 

[HarveyDunn]

Unless your files are specifically being reseparated (typically for newspaper application, less so for magazines, and rarely for general offset) your input cmyk screen tint builds are not changed. However they will not be the same on plate or presswork.

Gordo

Can you explain how it is both true that they are not change but they are not the same?

 

[HarveyDunn]

Is this a theoretical question, or do you see K highlight dots in the print when using a loupe or microscope?

I'm very new at this. I've seen people peering through loupes, but I have no idea what they are hoping to learn. If you could fill me, I would appreciate it.

It is not an entirely theoretical question: the last job I did had some light "colored greys" in it. By which I mean colors that were above L80 in value, and dull in color, but not completely neutral - pale pinks and oranges and yellows. They really didn't look very good. I was told that is just what I should expect when using a small amount of black in a light color - it makes the color look dirty.