Grayscale PDFs for print, dot gain and profiles

[ukzembla]

Hi, I work on a lot of books for print that just are K plate only. So far I've generally been asked by various printers to manually prevent TAC going over 80–95% depending on the images and the press, and I've  always done this using curves in Photoshop on a grayscale image with max + min highlight / shadow. Is there a better way than doing this manually?!… I haven't read a lot on here about people working in black and white so much but there must be lots in the same situation as me. I believe that InDesign just ignores whatever profiles are attached to a grayscale image when it's imported, but you can still export to greyscale with whatever dot gain you want – is that going to be better / worse / more / less work than tweaking things in Photoshop? Thanks! I would ask the printer but … they are quite rarely clear or helpful on profiles.

[rob day]

So far I've generally been asked by various printers to manually prevent TAC going over 80–95%

Limiting the grayscale black point percentage tops my list of printer’s advice to ignore. That conventional wisdom originates from the 90s when there was no reliable way to soft proof profiled grayscale images. The idea is you can prevent images from going too dark from dot gain or "filling in", but all you are really doing is limiting the already limited dynamic range of a black only image printing on an offset press.

Dot gain is most noticeable in the mid range values, and a Gray profile will soft proof the gain through out the scale. Consider this scan of a black only image printed on coated stock from a 4-color press running to the Coated GRACoL CMYK profile:

The patches at the bottom are in 5% increments. There’s plenty of range between 80 and 100%, and there is a clear difference between the 95% and 100% patch. Limiting the black point of this image to 85% would only wash it out

Which gets to the other printer problem:

I would ask the printer but … they are quite rarely clear or helpful on profiles.

In fairness to the printers, most probably don’t know what the profile for a black only press run is—the profile could change depending on the amount of ink density the press person decides to run. But the Gray profile you assign to a Grayscale image in photoshop soft proofs dot gain so there’s no need to put a limit on the black point. Black Ink profiles tend to be more accrate through out the scale:

Black Ink - US Sheetfed Coated compensates for more gain than Black Ink GRACoL Coated, or Dot Gain 20%

More on Black Ink profiles here

https://community.adobe.com/t5/indesign-discussions/exporting-rgb-cmyk-and-grayscale-to-print-pdf/m-...

[ukzembla]

Thanks for such a detailed and thought-out reply, Rob! Does it make a material difference that it's books we're working with, and so things move though the press pretty fast, and are pressed + bound not long after? I should perhaps have said too that they're typically made on uncoated paper. There's also the problem of ink transfer onto the facing page (I've temporarily forgotten the term for this, sorry), which I've also been told can be caused by too high an ink density. Do you have experience working with book printers too?

[rob day · Accepted solution]

which I've also been told can be caused by too high an ink density.

Total Ink drying problems are only a problem with 4-color printing—there’s no way black only at 100% should cause drying problems—if it did you wouldn’t be able to print 4-colors.

CMYK profiles limit ink on a color conversion—typically coated sheetfed profiles limit total ink to between 300% and 350%. US Sheetfed is 350%, US Web Coated SWOP is 300%. Uncoated profiles are usually less, Uncoated GRACoL is 280%. The limit is only enforced on a color conversion—you can color correct a CMYK image and force the total ink over the limit.

Also, InDesign does ignore embedded Gray profiles, and with Overprint Preview turned on, InDesign displays the grayscale as it would print on the CMYK black plate—the preview of a grayscale would change depending the ID document’s CMYK profile, and how the profile is soft proofing the black plate, but the gray output values do not change.

If you are color correcting to Gray Gamma 2.2 in Photoshop, the soft proof of the grayscale is going to change when it is placed in a document with the default US Web Coated SWOP and Overprint turned on. Gray Gamma is a Display class profile, not a press output profile. It might coincidentally profile the press, but it’s not an Output profile for any press.

[rob day]

Also, if you are having problems with the shadows printing too dark I think this is the curve correction to make:

Not this, which kills the dynamic range:

[reproo2773183]

Rob's replys are excellent as usual.

If I could add that the reason the Printer may be telling you to max out at say 90% in an image rather than 100% is to do with on press adjustments. Generally book printing on uncoated you would want to run the Black Density up so that the text gets nice and black, when you do this you are moving the step wedge away from the standards and you won't see that difference between 95 and 100, in some cases/paper stocks you'd lose that difference even at 90%.

My experience says that with images that show a great deal of shadow detail on screen and a coated proof you just can't reproduce it in print without applying the sort of curve above – but after applying the curve you can no longer proof that image. Half the problem is that most of the time the black is showing maximum black on screen, half the problem is that the standard uncoated profiles are somewhere in the middle of a very large range.

If I was retouching an image that sat on a solid black background I'd use similar to Rob's top curve (maybe even more severe* between 85 and 100, if it was sat on white maybe the lower one but add maybe 5% in the midtones.

If any of this is 4colour you can get the contrast you see on screen because you have extra ink to strengthen the blacks so you don't need to worry.

If any of this is printed digitally then you are not dealing with Ink Density so you can usually adjust to the point where you do get a difference in the shadow detail.

*more severe

 

[rob day]

the reason the Printer may be telling you to max out at say 90% in an image rather than 100%...

If I was retouching an image that sat on a solid black background I'd use similar to Rob's top curve (maybe even more severe* between 85 and 100

Hi @reproo2773183 , Just to clarify, both of our curve adjustments are allowing 100% gray values, so with those adjustments the full grayscale range is still output. 

Also, it’s possible to build a custom Gray profile that would soft proof the "filling in" of the 85-100% shadow range. You can build a custom dot gain curve via visual comparison to a printed press sheet—Color Settings>Gray>Custom Dot Gain:

This Grayscale has no Gray profile assigned, so the default Dot Gain 20% soft proofs tonal separation between 85 and 100%:

I  can build a custom dot gain curve that replicates the shadow filling in a soft proof without changing the output values

If I assign my custom dot gain curve, the soft proof of my shadow curve adjustment is more accurate—here I’m adjusting 90 to 78% and 95 to 87%:

[ukzembla]

Thanks so much everyone, again. As a rule of thumb, I usually know I'm printing on uncoated paper, and that's about it, but as an approximation, is it reasonable to soft proof in InDesign / Photoshop for greyscale using dot gain of 20–30%, rather than the uncoated CMYK equivalent (if it's not going to be printed on CMY plates)? Results are quite drastically different in overprint preview when I'm looking at greyscale images using dot gain vs e.g. PSO Uncoated v3, which I hadn't realised – is the dot gain preview going to be the more realistic of the two, and the dot gain using a curve like the one @rob day suggested even more accurate for my type of setup (with all the usual provisos of course)? I'm still unsure why printers are asking me not to print at 100% K, since you're both using that in your profiles, but @reproo2773183 is I think right that they may be increasing density for print here. 

[reproo2773183]

The series of images I had to use this severe curve on, had large areas of detail between 85 and 100  similar to Rob's bottom left corner of the Car Radiator Grille. All these images were solid black at the edges and at on solid Black background objects so I had to leave my 100% point at 100%. If I hadn't had to hide the edge of the image against the background I'd probably of gone with a curves adjustment similar to below.

At this point I'm ignoring what I'm seeing on screen and trying to make the print on press look like it did before the curves adjustment. I'm using the eyedropper to try and get those high 90s to high 80s.

Maybe Rob's custom Dot Gain curve is the way to go for simulating print on screen but how would you go about fine tuning it for the print process? And how would it help outside of Photoshop or Downstream?

 

[rob day]

Maybe Rob's custom Dot Gain curve is the way to go for simulating print on screen but how would you go about fine tuning it for the print process?

You would need a grayscale target printed from the destination press that you can use for a visual comparison as you adjust the dot gain curve— ideally an image with full tonal range. Dot Gain curves can be saved out as .icc Gray profiles via Color Settings>Gray>Save Gray... 

InDesign will ignore the embedded Gray profile and use the document’s CMYK black ink for its Overprint Preview, but if you export to PDF/X-4 the grascales will export as DeviceGray (no profile) and in that case AcrobatPro will preview DeviceGray

objects using the Gray profile set in the Color Management Preferences, which can be your saved custom gray profile:

[ukzembla]

Hmmm, and does Photoshop work in the same way in terms of softproofing? ie it'd need to be done in Acrobat? Thank you so much for all the invaluable advice! If you have any pointers to doing this successfully online in a Youtube video or anything please let me know, otherwise I'll have a forage around!

[rob day]

In both Photoshop and AcrobatPro the appearance of a grayscale depends on the images’s Gray profile assignment—for example in Photoshop assign the Dot Gain 10% profile (Edit>Assign profiles) and then change the profile assignment to Dot Gain 30% and you will get a very different appearance without the Info panel’s output numbers changing.

If the PS document has no profile assignment, the soft proof falls back to the current Color Settings Gray Working space.

Acrobat works the same way, if you export an InDesign file with a placed grayscale using the default PDF/X-4 preset, the document grayscales will export with no profile embedded (DeviceGray). AcrobatPro will preview DeviceGray in the current Gray Working Space—Preferences>Color Management.

[reproo2773183]

Hi Rob,

I don't disagree with anything you've posted. Fundamentally I just don't like the way Photoshop displays the shadow end of grayscales, the blacks are too contrasty against the white of the sceen and you can see clear differences between 98% and solids whereas in Print the Black would be dark grey and there would be nowhere near the contrast you are seeing on screen.

Your screenshot with CoatedGRACol 2013 as the Simulation Profile is a much more reasonable expectation of 1 colour Print than anything you've screengrabbed from photoshop.

A Custom Dot Gain curve doesn't fix the problem of the Black point being viewed as black as RGB can go and AFAIK the white point too.

So you end up needing to paste the Grayscale into the black channel of a profiled cmyk.

And whilst that should work nicely for most Coated stocks with a similar white/yellow point, for that to be meaningful with Uncoated you'd need to have a profile for the specific stock and process that you are printing on.

And even if you go to all this effort and expense there is nothing to stop you falling out with Printer A and Printer B giving you a different result because their RIP is set up to handle cmyk Profiles differently.

Not sure about any of it any more. Is there anyway to set Photoshop's White and Black points but still use a greyscale?

[rob day]

Your screenshot with CoatedGRACol 2013 as the Simulation Profile is a much more reasonable expectation of 1 colour Print than anything you've screengrabbed from photoshop.

Sorry, I posted the wrong capture. You can save out a custom Dot Gain curve as an .icc profile and use it in Acrobat as the Simulation Profile—for Acrobat to see the profile it needs to be saved in:

⁨Library⁩ ▸ ⁨Application Support⁩ ▸ ⁨Adobe⁩ ▸ ⁨Color⁩ ▸ ⁨Profiles⁩ ▸ ⁨Recommended⁩

Here the Simulation profile is my Custom Dot Gain profile. The bottom image has no correction—95% is outputting as 95% and its soft proof is a similar value to 100%:

The top image has my correction applied 95% is corrected to 87%:

The Photoshop Grayscales are displaying the same with my Custom Dot Gain profile assigned:

Fundamentally I just don't like the way Photoshop displays the shadow end of grayscales

Gray .icc profiles do have a limitation in that they can’t specifiy the "color" of black ink the way a CMYK profile can—100% gray is always displayed as absolute black, which wouldn’t happen on an uncoated sheet, so it’s a problem with Gray profiles not Photoshop.

As you suggest we can overcome the gray profile 100% display problem by working in the Black channel of the CMYK file, but for 1-color only printing where there’s no CMY on the sheet, I think it can be dangerous to assume the press is going to run exactly to something like Uncoated GRACoL where the profile might assume a lower black density. With black only I think the press operator might push the black in order to overcome the "grayness" of one color on an uncoated sheet. For color work where there’s CMY supporting the black, the tendency might be to hold back on the black density because the blacks are going to be richer.

[reproo2773183]

With black only I think the press operator might push the black in order to overcome the "grayness" of one color on an uncoated sheet. For color work where there’s CMY supporting the black, the tendency might be to hold back on the black density because the blacks are going to be richer.

Yes. and Yes again. I think you you are explaining my initial post so much better with these words.