Converting Images to Grayscale for Print

[michaelsgoldman]

Sorry, need help ASAP. I designed a book with 20 digital pencil illustrations. I have a hardware calibrated BenQ monitor. Sheridan (the Printer) specifies that for black only printing, all images should be converted to grayscale. Makes sense. They also specify the expected Dot Gain for uncoated stock on a Web Press is 25-27%. Ok cool. However, when I ask about grayscale profiles for 1) converting images to grayscale and 2) soft proofing the images on screen, they just keep answering that no profiles are needed for grayscale. To my knowledge, you cannot convert an image or even view an image without a profile. If a custom one is not selected, Photoshop will use the default profile. Which in my case is Dot Gain 20%. I don't know if I'm thinking about it all wrong, but here are my questions. PS, if you didn't guess yet, I have never printed a book interior in single color black only:

1. What is the best way to convert sRGB illustrations to grayscale for offset web press printing on uncoated stock? I want to prevent the images from darkening too much and losing the fine tone/texture detail in my digital pencil drawings.
2. I tried converting with Dot Gain 20%. It makes the images appear lighter on screen. I tried Dot Gain 25% and it make the images appear darker on screen than the 20% profile. Shouldn't the 25% profile appear lighter on screen than the 20% profile because it is compensating for more dot gain? The on screen effect is the opposite of what I expected.
3. Are Dot Gain profiles even still used in a modern printing workflow? I read on a post here that using just the K channel in CMYK is better for modern workflows.
4. For cover printing, when I asked about color printer profiles Sheridan recommended Coated FOGRA 39. On one of their documents they recommend GRACoL2006_Coated1v2.icc. And on another document for exporting to PDF they recommend U.S. WEB COAT (SWOP) v.2 if the document is not already setup for the correct color space. Are they trying to be confusing or are these profile so similar that it wouldn't matter a whole lot?
5. In the end, I just need to know how to ensure that my illustrations will print nicely on uncoated paper. How can I do this. Is it possible to soft proof for this on screen.

[michaelsgoldman]

Also forgot to mention that I am editing the images in Photoshop. Book layout is in InDesign.

[D Fosse · Accepted solution]

Open Color Settings, and in the Working Gray rolldown, click Load Gray and navigate to the CMYK profile that will be used for print:

If you start with an RGB file, converting to grayscale will now convert to the K component in the CMYK profile, which is what you want. It will also appear as a Convert to Profile option if you have grayscale files in other gray spaces.

InDesign does not have proper grayscale color management support. A grayscale file will just be sent directly and unmanaged to the K channel in CMYK and output on the black plate.

So this is the way to get the correct result. Effectively, since InDesign does not color manage grayscale, you're doing it in Photoshop prior to sending it to InDesign.

[michaelsgoldman]

Loading a CMYK profile as the gray space is a great tip. Didn't know you could do that.

I keep asking for a profile though and they have said multiple times that no profile is neede for grayscale. For the cover they mentioned the two in my original post (Fogra 39 and Gracol 2006). Then they have a document that says to output from InDesing to U.S. WEB COAT (SWOP) v.2 if the document is not already setup for the correct color space. But that is for coated paper.

How do I account for Dot Gain if they won't provide a profile for uncoated paper? Also, I would really like to know why the Dot Gain 20% profile looks lighter on screen than the Dot Gain 25% profile.

Thanks!

[D Fosse · Accepted solution]

Dot gain (ink spread in the paper) is built into all CMYK profiles. The dot gain profiles are outdated and not really used anymore. They're not accurate and don't correspond to modern printing processes.

But you need to know which CMYK profile. Keep nagging! You do need it for grayscale, as per above.

[michaelsgoldman]

Awesome, I will thanks! If I convert the illustrations to grayscale with a CMYK profile in Photoshop, and then place those illustraitons into InDesign, technically the file would already be in the correct color space and I would not need to convert on export again from InDesign? ie I could select "No Color Conversion" in InDesign?

And I'm still just curious. I don't understand why the Dot Gain 20% profile would display lighter on screen than the 25% profile. Isn't is supposed to be the opposite?

[D Fosse]

It depends on what else is in the ID file. There may be other RGB images that need to be converted to press CMYK. That does no harm, your grayscale images already are in that color space, so nothing happens to them.

How grayscale displays depends on whether the application supports grayscale color management. Photoshop does, so any grayscale profile will display correctly (and thus identically) as long as there is an embedded grayscale profile. The different tone curves are remapped into the monitor profile's tone curve.

But InDesign does not have grayscale color management. Any grayscale image will be interpreted as the K component in the ID working CMYK. And if it already is, that displays and prints correctly.

[michaelsgoldman]

You've been super helpful. Thank you so much! That all is making a lot of sense.

Since it seems important to included color profiles when moving files between applications and hardware, why would the printer specify that if the document is in the correct color space already to select "No Color Conversion" and "Don't Include Profiles" from the InDesign export settings?

[michaelsgoldman]

[D Fosse]

What I've always been told is that some RIPs - that's what produces the halftone screen - don't work well with profiles. I suspect that's older processes and it's not a real problem anymore, but I don't know.

It doesn't really matter in practice. A press-ready PDF will already be converted into final press CMYK, so the profile is redundant at this point and is not needed. No further conversions will be done.

[michaelsgoldman]

Ahh, I see. Then here's my final question:

If the printers prepress team is viewing different files on their monitors, converted to different profiles based on the clients paper specs etc, wouldn't they need the profile to view it accurately? Or perhaps they don't care how it looks on their screen, and they just assume that the client chose all the right settings for whatever printing equipment it's set to run on?

[D Fosse]

This is where the profile needs to be embedded. Then standard display color management applies: the document profile is converted into each system's custom monitor profile. Those corrected numbers are sent to screen, again representing the file correctly. If all these operators have properly profiled monitors, they will all see the same thing.

[Stephen Marsh]

Also, I would really like to know why the Dot Gain 20% profile looks lighter on screen than the Dot Gain 25% profile.

Thanks!

By @michaelsgoldman

Looks lighter where?

Photoshop or InDesign? And with what exact colour settings?

[michaelsgoldman]

The whole image onscreen in Photoshop looks lighter with Dot Gain 20% vs Dot Gain 25%. Shouldn't it be the opposite if Dot Gain 25% is compensating for more dot gain by adjusting the tone curve so that the midtones had lower tonal values in that range?

[D Fosse]

1 - higher dot gain requires a lighter image to compensate for the higher ink bleed.

2 - There should not be a difference in Photoshop as long as you convert between them so that the numbers are recalculated. If you assign there will be a difference because the numbers aren't changed, but the meaning of the numbers are. 

The whole point of color management is that color spaces are remapped from source (document) to destination (screen/print), so that the visual appearance is maintained. Or you can initiate the process manually if you want to change the color space of the document, again, preserving visual appearance.

InDesign doesn't color manage grayscale, so there you will always see the difference.

[Stephen Marsh]

The whole image onscreen in Photoshop looks lighter with Dot Gain 20% vs Dot Gain 25%. Shouldn't it be the opposite if Dot Gain 25% is compensating for more dot gain by adjusting the tone curve so that the midtones had lower tonal values in that range?

By @michaelsgoldman

There should be no difference in appearance or in Lab values between the 20% and 25% dot gain profile conversions in Photoshop. The grayscale tonal values will be different as expected from the different compensation curve.

[michaelsgoldman]

How can the appearance in Photoshop be the same when the grayscale tonal values are different?

Here are some screenshots. What looks like is happening is that if I convert to a grayscale profile and merge the layers with a white background, the appearance in Photoshop does not change. If I merge just the transparent layers (no white background) the appearance does change.  Converting to Dot gain 20% makes it look lighter. Dot gain 30% makes it look slightly darker than the original, but almost the same. What is happening?

Do I want the appearance to change or not? If it does not change, is it still compensating for dot gain if it doesn't visually look different on screen? Sorry that I'm not understanding.

[Stephen Marsh]

@michaelsgoldman 

Colour mode conversions are as a general rule best flattened during conversion.

[michaelsgoldman]

That's the weird part, I am flattening. If I have just transparent layers and flatten before or during conversion, the appearance changes on screen. Lighter than the original RGB for Dot Gain 20% and darker than the original for Dot Gain 30%. If I add a white background and flatten, the appearan does not change during conversion.

Do you want to see your image lighten for Dot Gain compensation or do you want it to look the same on screen? If it looks the same, will the dot gain during printing still be compensated for?

[Stephen Marsh]

But what about the actual tonal value %?

Look at the values using the info panel. The value is compensated for.

[Stephen Marsh]

But what about the actual tonal value %?

Look at the values using the info panel. The value is compensated for.

By @Stephen Marsh

To illustrate my previous comment:

When evaluated with their respective profiles - the sRGB, 20% gray dot gain and 25% gray dot gain all have the same Lab value... However you can see that the gray value for the 25% conversion is 5% lighter than the 20% conversion.

[michaelsgoldman]

That example image really helped too, thanks!

[D Fosse]

@michaelsgoldman 

I don't want to interrupt this prepress discussion (my expertise is as a working photographer) - but what you still have trouble with is the basic principle of color management, whether applied to RGB or grayscale, or document, display or print.

Yes, numbers will change with different profiles, even when they look the same. RGB or K%, same principle.

The key is that there is a reference to it all, the profile connection space, PCS. All conversions go through there, whether it's from document to document, or from document to your display, or from document to print. The PCS is either Lab or CIE XYZ, but people will usually be more familiar with Lab.

Since I'm a visual guy, here's how it works:

[michaelsgoldman]

I want to thank you all for taking the time to help me understand color management better. I really appreciate all the info provided! I think this is where I am getting hung up:

When I merge transparent layers and convert to Grayscale there seems to be a glitch. I do not get the same results when the background is transparent vs having a white background.

Predictable Results = Converting to grayscale with white background

• Image appearance stays the same in photoshop
• Dot Gain 25% K values are lower than Dot Gain 20% k values (as expected)

  

Unpredictable Results = Converting to grayscale with transparent background

• Image appearance lightens in photoshop and K% values are lower than if converted with white background
• Dot Gain 25% K values are higher than Dot Gain 20% K values (unexpected)
• (Note: a White background was added after merging and converting, otherwise all K% values show as 100%…)

Am I doing something wrong or is this a glitch? I don't want a white background, but I'm not sure which method is more accurate so that the illustrations do not print out too dark for web offset on uncoated natural stock.

[michaelsgoldman]

Argh, sorry, uploaded the wrong images for the second set…

Unpredictable Results = Converting to grayscale with transparent background

• Image appearance lightens in photoshop and K% values are lower than if converted with white background
• Dot Gain 20 and 25% K values  are the same (unexpected)
• (Note: a White background was added after merging and converting, otherwise all K% values show as 100%…)