Preserving Greyscale Dot Gain in CMYK mode

[LAOPeroxide]

I have a question about managing dot gain in greyscale images in a CMYK workflow.

I have an RGB image that's superficially greyscale (as in, it looks like grey tones, even though it's made of rgb values). It will be placed in InD in a document that has some colour pages/images, and some black and white. The document will be exported as a PDF/X-1a with a cmyk output intent. The image in question (plus the rest of the b&w content) has to print in k black only, so I convert it to Greyscale mode in PS, adjusting levels as desired, then place it in InD, then export to PDF- in Acrobat's Output Preview, the separations read as k-values only for the image. So far so good.

The question:

1) My printer says to compensate for 25% dot gain on all images. In general, can mixed b&w and colour content in the same pdf be adequately compensated for with a single destination profile/output intent? (In this case, everything is being treated like cmyk, so will the cmyk output profile/intent be sufficient for the dot gain in the greyscale images that only have k values? Or do images using just k dots versus all four colour dots require different compensation curves, even if they're in the same "mode"?)

2) If not, where in the process can I apply a dot gain compensation to the greyscale images? I assume embedding a profile would be useless, as the PDF/X-1a conversion will strip it.

3) It's worth mentioning that, for some reason, my printer did not supply me with a profile to use, even when directly asked. I'm assuming that without a specific printing profile, the 25% dot gain instruction is actually useless? (I've just picked US Web Coated SWOP v2 as my workflow profile, since its Adobe's the standard working space and stuff does indeed come off the press looking good when we use it. But that still brings me back to 1)- is US Web Coated SWOP v2, as a cmyk output intent, sufficient for both cmyk and k only dot gain?)

Any clarification greatly appreciated!

[Bob_Hallam]

Sounds like this printer is very low tech.  26% dot gain is high as well, and it's rare (Like I have never seen one) where a printer will have equal dot gains in all colors.  Also of note is it's the printers' responsibility to adjust their process to a standard.  So that begs the questions, What print standard do they use and What print process is this? (ie Offset Flexo, Gravure etc) 

So go through those questions and try to give us a better picture of what you're dealing with if you can.  

The reason for those questions is there are 2 ways now to get to your solution.  

One is to convert to CMYK with a standard profile that compensates for a "normal" amount of dot gain and lighten the files afterward using curves to go the rest of the way your printer needs.  I suggest this but I'd go so far as to say that your printer must, if they have been in business for very long received file that are at print standards and uses them.  They compensate with plate curves and life is good.  That would be the expectation of anyone supplying files to a printer.  

The other is doing a custom CMYK profile in PS where your responsibility is total for every little bit of how that conversion and print process will happen. I don't recommend this because so far at least the information received here doesn't tell us if you have any proofing capabilities.  Without proofing capabilities this is a bad road to take.  

[NB, colourmanagement net]

Hi

Basically you need an accurate ICC CMYK profile from the printer. That way no discussion on dotgain will be needed. 

Many printers have adopted the sensible approach of standards based printing, that’s what ISO 12647 and GRACoL are all about. 

These guys seem rather seriously out of step with current working practices 

I hope this helps

if so, please "like" my reply

thanks

neil barstow, colourmanagement.net :: adobe forum volunteer

[please do not use the reply button on a message in the thread, only use the one at the top of the page, to maintain chronological order]

[LAOPeroxide]

In reply to both, this is a cold web offset press and we're printing on newsprint. Their specs are at least partially based on SNAP. The printer told me 25% dot gain, and PDF/X-1a, but didn't give a profile. (They're a large company too- it's strange)

My manager told me that in the past we'd had problems with images coming out too dark. I don't know what the previous workflow was, but I do know that when we work in US Web Coated (SWOP) v2, manually edit greyscale images in Photoshop before placing (bring up midtones, heighten contrast), and place/export them in the workflow I described, they generally come out looking good. So I've been doing something similar to what you suggested, @Bob_Hallam1- just not in any fixed, quantified way.

So would I be best off taking those images that did turn out well, and basing all future curve/levels edits on those ones?
(We aren't set up for reliable soft-proofing. The output happens to match our screens decently, but a) we can't calibrate them and b) without a profile it guess wouldn't be much use anyways.)

[rob day]

If you use a profile that is designed for the destination press, you shouldn’t have to manually compensate because the assigned profile affects the preview. You can see that here where I’m comparing Black Ink-US Newsprint with Photoshop’s Dot Gain 25% curve.

The DotGain 25% preview shows more contrast than will likely happen in the print—newsprint flattens contrast because of the absorbent paper, and the Black Ink US Newsprint profile is displaying the expected lower contrast.

[rob day]

The printer told me 25% dot gain, and PDF/X-1a, but didn't give a profile. (They're a large company too- it's strange)

Newsprint is challenging because the profile of the press is constantly changing because the ink density is harder to control, but you would not want to convert to the default US Web Coated SWOP, and try to compensate by color correcting the numbers. At the very least you would have a total ink problem—SWOP’s 300% vs SNAP’s 220%

[rob day]

InDesign doesn’t have a grayscale space—placed grayscale objects output on the document’s CMYK black plate with their output values unchanged. When Overprint Preview is turned on, the document CMYK profile handles the preview of placed grayscale objects—an embedded Gray profile will show in the Links panel but it is ignored. With Overprint Preview turned off grayscale objects soft proof as sGray (2.2 Gamma), which would be useful if you are designing for web or screen display and not print.

Photoshop does have a Gray space, and if you want the soft proof of gray values to match between Photoshop and InDesign you have to use a Black Ink profile that matches the InDesign document’s CMYK profile.

For example, if your InDesign document has US Sheetfed Coated as its CMYK assignment you would want to use Black Ink -US Sheetfed Coated as the assigned Photoshop grayscale profile. Photoshop’s Info panel Actual Color shows the gray values that will output to the plates, and the Gray profile adjusts the soft proof to show the affect of dotgain:

Black Ink-US Web Coated assumes less dot gain, so the same output values display as lighter values:

To use a Black Ink profile in Photoshop open Color Settings>Gray Working Space>Load Gray... and choose the desired CMYK profile. The Gray profile will be listed as the CMYK profile with a Black Ink - prefix:

If the printer will not provide a CMYK profile, but is expecting you to make CMYK conversions (PDF/X-1a) you are shooting in the dark for both CMYK and Gray color management.

[LAOPeroxide]

Black Ink-US Web Coated assumes less dot gain, so the same output values display as lighter values:

Am I correct then that Photoshop's soft-proof is approximating what the absolute k values will look like after they've printed and dot gain has occured? (As opposed to visualizing the degree of lightening/compensation built into the profile before printing.)

If the printer will not provide a CMYK profile, but is expecting you to make CMYK conversions (PDF/X-1a) you are shooting in the dark for both CMYK and Gray color management.

Unfortunately this is the case. We've got limited resources to begin with, and without a profile I've just had to guess. We've been lucky in that the first things sent to print turned out well, so I've stuck with that workflow since.

...you would not want to convert to the default US Web Coated SWOP, and try to compensate by color correcting the numbers. At the very least you would have a total ink problem—SWOP’s 300% vs SNAP’s 220%

They did give me a total area coverage, which is 240% for 4 colour content and 140% for 2 colour content. (Is this kind of split TAC normal?) I have been manually editing colour images in Photoshop, using levels, curves, and channel mixers to get colours under this limit and checking them with Acrobat Output Preview/ InDesign's Separations Preview (I don't know of an equivalent photoshop feature that displays full-image ink levels) It's time-consuming, but seems to work, even though I'm working in US Web Coated (SWOP) v2.

In this situation, what cmyk profile would you recommend working in? Would it still be better to choose a profile with a higher-than-specified ink limit and correct heavy colours, rather than choose a lighter profile and make all your darks lighter than necessary?

[LAOPeroxide]

"if you use a profile that is designed for the destination press, you shouldn’t have to manually compensate because the assigned profile affects the preview."

"Black Ink-US Web Coated assumes less dot gain, so the same output values display as lighter values"

Just to clarify, I wasn't compensating to create an accurate preview- my intent was to lighten midtones and improve contrast so that the image might print better regardless of dot gain. ie. With no profile I don't know how much dot gain there'll be. But regardless, I figured that with any amount of dot gain, lightened midtones and heightened contrast will print more readable than a flat image with fewer tonal values.

But with regards to grey soft-proofing, am I correct then that Photoshop's proof is showing what the absolute k values will look like after print, in dot gain conditions simulated by the profile?

If the printer will not provide a CMYK profile, but is expecting you to make CMYK conversions (PDF/X-1a) you are shooting in the dark for both CMYK and Gray color management.

Unfortunatley this is the case. I've had to guess- but I've been lucky in that our first print job turned out well. So I've stuck with the workflow for that first project ever since.

...you would not want to convert to the default US Web Coated SWOP, and try to compensate by color correcting the numbers. At the very least you would have a total ink problem—SWOP’s 300% vs SNAP’s 220%

They did give me an ink limit- 240% for 4-colour content and 140% for 2-colour content. (Is this kind of split limit normal?) To get colour images under this limit, I have been editing them in Photoshop with levels/curves/channel mixers + Info Panel, and then checking TAC for the whole image in Acrobat or InDesign. It's time consuming, but seems to work. I'm working in US Web Coated (SWOP) v2 but never hit 300% simply because I've edited those colours out. (Unless I'm missing the point- are you saying that a given ink coverage in two different profiles entails a different, absolute volume of ink in each process being simulated? ie. is 220% in SWOP a heavier ink volume than 220% in SNAP?)

What cmyk profile would you recommend using in this situation? With this target TAC, is it better to use a profile that's over or under?

[rob day]

Am I correct then that Photoshop's soft-proof is approximating what the absolute k values will look like after they've printed and dot gain has occured?

Yes, the K values listed in the Info panel are the values output to the press plate and not the printed sheet after the gain has occurred. If you place a grayscale in an InDesign document and export to PDF/X-1a with document CMYK as the destination, the values you read in AcrobatPro’s Output Preview would be unchanged from Photoshop. The soft proof (the conversion from the gray profile to your monitor’s profile for display) shows the expected affect of dot gain. You can see that via Edit>Assign Profile... and choosing 10% Dot Gain vs. 30% Dot Gain—the output numbers don’t change but the preview does:

A curve based profile like 25% Dot Gain is somewhat problematic in that it is simply a midtone adjustment, and when you print on a paper like newsprint the shadows appear lighter and the highlights appear darker (less contrast), and a Black Ink newsprint profile would soft proof that affect assuming you have an accurate monitor profile.

It's time-consuming, but seems to work, even though I'm working in US Web Coated (SWOP) v2.

It may work, but you are fighting against the built in color management system where output profiles automatically handle all of the press variables on the conversion from RGB to CMYK, which you are trying to blindly compensate for after the conversion via CMYK color corrections.

A newsprint profile compensates for more than dot gain, it handles the black generation, the CMYK ink colors (which are different than SWOP because of the paper), gray balance, and total ink.

In this situation, what cmyk profile would you recommend working in?

I don’t know why you wouldn’t use a newsprint profile. Here you can see the affect off the destination CMYK profile on the output numbers while maintaining the color appearance using Convert to Profile on an AdobeRGB image.

The SNAP 0|0|0 RGB Black (#4) converts to 217%. You can see the profile’s gray balance, and black generation are affecting the conversion of mid gray 128|128|128 (#3). Also the saturated colors are a bit more saturated relative to the SWOP output again compensating for the paper. There wouldn’t be much reason to make any post CMYK color corrections to the SNAP conversion. You could even leave the color as RGB and make the conversion to the SNAP destination on the PDF export from InDesign—both ID and PS use the same color management system.

SNAP:

SWOP

[LAOPeroxide]

So if I understand- even in the absence of a specific printer profile, it's better to be editing images in RGB, then converting to a cmyk newsprint profile that at least has similar characteristics to my output device and given specs, rather than use the default working space. Greyscale images can be converted to greyscale using the black ink of my chosen newsprint profile, and then converted to cmyk in the final PDF/X-1a

(I guess what gets me is the gamut differences between rgb and cmyk- I've always edited images in my chosen cmyk profile because it's the lowest common denominator, colour-wise. I've never wanted to fall in the trap of relying on bright, saturated rgb colours that can't be repoduced by my printer, so I never bothered with rgb for printed images and edited in cmyk from the start, to see what I have to work with. Yet I hear everyone saying it's best practise to work in rgb mode and convert afterwards. Can someone speak to this?)

Tangent aside- I'm going to see if it's feasible for my workplace get some test prints done, to compare this workflow to what I've been using. (Not to mention keep hounding the printer for a profile. I imagine that trying to calibrate monitors for good soft-proofing will still be guesswork without this.)

[Bob_Hallam]

Hearing the recommendation to use a newspaper profile because the dot gain is at 26% would work, but o Ky if you're printing on newspaper.  If you're not then don't use a newsprint profile.  This are more reasons than dot gain alone that make a profil work for a print process.  Not the least is gamut. Newspaper profiles have a very small gamut and high dot gain because of the stock and print process.  It is a low quality high speed process.  So if your work isn't being printed that way don't go there.  

if you're printing on coated stock use a coated profile that's going to get you close to the gamut of the print process your printer is using.   If they don't have web presses don't use the SWOP.  For coated sheet Fed work use GraCol 2013 or CRPC-6 as the destination profile and lighten the file by 5% in the mid tones using curves.   If you tried using a newspaper profile on coated sheet Fed printing the results would be very poor.   

most of today's high quality work is edited in RGB.  I prefer Adobe 1998 RGB for most work and supply color corrected RGB to printing companies.  Those printing companies convert to the appropriate press profile and there is no disappointment.  

[rob day]

Yet I hear everyone saying it's best practise to work in rgb mode and convert afterwards. Can someone speak to this?)

In Photoshop you don’t have to actually make the conversion to CMYK in order to soft proof how the color will convert to the destination CMYK profile, or if you prefer, color correct by the output numbers.

If you set your Color Settings Working CMYK space to the destination profile, set your Proofing Setup to Working CMYK, and turn on Proof Colors, the soft proof will show the conversion while you work in your RGB editing space. The same will happen in InDesign if you turn on Overprint Preview.

Here I have an RGB image with some patches that are clearly out-of-gamut. With Proof Colors turned off I see the RGB gamut—the Info panel set to CMYK (or Proof Color) shows the CMYK output numbers for the Working CMYK profile. If the color is out-of-gamut to the proofing space there will be an ! with the output numbers. The Color Settings’ Conversion Options also affect the output numbers.

Proof Colors off:

Proof Colors on:

[LAOPeroxide]

Ah, okay. That all makes sense- I'd been avoiding relying on soft proofs because of our uncalibrated system, but I think going forward I'm just going to push for this to be a priority.

Thank you, everyone who's replied with their advice- it's been tremendously helpful! 😉

[rob day]

I'd been avoiding relying on soft proofs because of our uncalibrated system

Keep in mind the soft proof accuracy depends on both the output profile and your monitor profile. If you are soft proofing CMYK output there’s a conversion from profiled CMYK to your Monitor profile for the display (CMYK>Lab>MonitorRGB).

If you are using OSX, it has a software calibrator that will generate a monitor profile—probably not as good as a hardware calibrator, but better than nothing.

[NB, colourmanagement net]

Hi

I believe that you need decent display calibration and profiling, period. Otherwise the screen cannot show you what the file's data really means. I believe you need a hardware calibration sensor to do that properly.

A messed up job could cost a reputation, or a client's business - X-Rites i1 Display Pro is a good entry level display calibration / profiling setup and the cost is insignificant in comparison to the value of a hard earned reputation.

I hope this helps

if so, please "like" my reply

thanks

neil barstow, colourmanagement.net :: adobe forum volunteer

[rob day]

you need decent display calibration and profiling, period.

On a conversion to CMYK, the ouput profile is equally important—it also affects the soft proof as well as the output numbers. If the destination is newsprint and you convert to the default US Web Coated SWOP, the soft proof and output numbers would be wrong even with a perfect monitor profile.

[NB, colourmanagement net]

Rob

absolutely right, thanks for adding that tip about the output profile in proof setup

neil barstow, colourmanagement.net :: adobe forum volunteer

[edy54443724]

Hi Rob. I read this procedure for using the black plate of the CMYK profile in use here and in another post, and got me curious.

What then is the advantage/benefit of doing this vs Image > Mode > Grayscale (after editing in Adjustments > Black & White, of course)?

Is it about dot gain? Contrast? I tried both conversions and, with Overprint Preview turned on, saw the grayscale turn a bit lighter.

[rob day]

What then is the advantage/benefit of doing this vs Image > Mode > Grayscale

When you convert via Mode>Grayscale, the conversion is to whatever you have set as the Working Gray space in Color Settings, that could be the default Dot Gain 20% or any other Gray Profile including Black Ink gray profiles.

If you are placing in InDesign, where there is no Gray working space, using a matching Black Ink profile is important in order to get a matching appearance in InDesign when Overprint Preview is on.

[edy54443724]

Thanks, mucho thanks. That there is one good piece of advice I haven't encountered in Ps and press books. 

Alternatively, would you say another good way to convert to grayscale (to place InDesign) is to covert to 4-color grayscale? Perhaps appying a light GCR?

[rob day]

covert to 4-color grayscale? Perhaps appying a light GCR?

That would be a conversion to CMYK not Grayscale. With a CMYK conversion there is the advantage of a richer black point, but the risk of a color cast if the press is offset. You would also have to control of the CMYK values all the way to output and be able to ensure there are no additional CMYK-to-CMYK conversions downstream. In InDesign the document's CMYK Policy should be Preserve Numbers (Ignore embedded Profiles).

[D Fosse]

@edy54443724 

In Photoshop color settings, click "load gray" from the working gray rolldown, and navigate to the appropriate CMYK profile.

Then the working gray changes to Black Ink - <CMYK profile> like this:

The crucial part is that this now also becomes available as a grayscale profile in "convert to profile". Now you can convert your Dot Gain document, and place that grayscale file in InDesign. This will now output correctly on the black plate only.

[Bob_Hallam]

The Method that DFosse mentions above will work well for most commercial print applications.  Adobe includes different options in the pulldown menue for Gray depending on the region your in.  In the US, we have the choices below: But even those can accomidate most print conditions .  

[edy54443724]

My thanks to @rob day and @D Fosse (and everybody else who chimed in) for your comments. Very informative indeed.