Dan Margulis Applied Color Theory

Creating a 4/c Black and White

   Date: Mon, 3 Jan 2005 07:36:07 EST
   From:Hal Silverman
Subject: Converting to B/W

Hello and Happy New Year to all.

I just completed shooting a job, where many(24) of the color images will be reproduced in a brochure as 4 color b/w. Through the "Channel Mixer," I have converted the images to a monochromatic state while keeping them RGB. They look great!

Now, how should I go about the CMYK conversion? What should I look for in keeping the neutral color in the images? I noticed that in some of the images I converted over while testing, the cmy values shifted a percentage here and there. I might mention that the images were shot under the same conditions, so they are very consistent in their appearance.

I contacted the printer for their recommendation; they were of no assistance.

Best regards,
Hal

Hal Silverman Studio, Inc.
6511B Basile Rowe
E. Syracuse, NY  13057

p. 315.434.9121
w. www.silvermanstudio.com
___________________________________________________________________________

   Date: Mon, 03 Jan 2005 12:59:21 +0000
   From: Shangara Singh
Subject: Re: Converting to B/W

It was 3/1/05 12:36 pm, whenHal Silverman wrote:

I just completed shooting a job, where many(24) of the color images will be
reproduced in a brochure as 4 color b/w. Through the "Channel Mixer," I have
converted the images to a monochromatic state while keeping them RGB. They
look great!
Now, how should I go about the CMYK conversion? What should I look for in
keeping the neutral color in the images?

Stephen Marsh has some good info regarding basic CMYK separations on his Website. It's in three parts. The info you seek is in part three, I think. It's worth reading all three parts.

http://members.ozemail.com.au/~binaryfx/index.html
 
Shangara Singh.
___________________________________________________________________________

   Date: Mon, 3 Jan 2005 12:48:32 -0500
   From: Terry Wyse
Subject: Re: Converting to B/W

I'd create a "Custom CMYK" profile (I can't believe I just recommended using the "Custom CMYK" engine) and set it up for either "Heavy GCR" or "Maximum GCR" with a Black limit at or near 100% with a total ink of not more than about 300%. The idea is you want a separation made up of almost exclusively black ink with only enough CMY to extend the dynamic/density range out to the limits of what the press can achieve. The other option would be to create a "traditional" double-black duotone or gray+black duotone.

As a side note, I recently did a bit of press profiling where I included a total ink chart in the press run. Interestingly (or not), the highest Dmax's were achieved where K=100% but the total ink was no more than about 260-280%. At K<100%, the total ink amount could be much higher (340%+) but in no case did the Dmax exceed the K=100%, TAC= 260-280% combination. FYI, this was on a 6/color 40" Heidelberg sheetfed press.

Cheers,
Terry
_____________________________
WyseConsul
Color Management Consulting
704.843.0858
http://www.colormanagementgroup.com
http://www.wyseconsul.com (coming soon)
________________________________________________________________________

   Date: Tue, 04 Jan 2005 01:05:22 -0000
   From:Stephen Marsh
Subject: Re: Converting to B/W

Hal Silverman wrote:

Now, how should I go about the CMYK conversion? What should I look for in
keeping the neutral color in the images?

Two basic choices - grey balance is your main concern here, even though using a heavy GCR this is still a concern for lighter tones.

i) Using a modified default Custom CMYK using Heavy GCR and whatever dot gain and max K and TIL etc.

This will result in one type of neutral grey balance, that many printers who are used to working with Photoshop separations over the years will be comfortable with.

ii) Using an EDITED Custom CMYK setting that has ALTERED LAB INK VALUES from the default, to produce the grey balance that you would like to hit, rather than the Photoshop legacy aimpoint. For example, I have made an ICC profile from Custom CMYK that is very close to TR001 grey balance (what most mean when they say SWOP). But I would only use this hacked profile for four colour greyscales.

Now if Adobe and other profile suppliers would be kind enough to offer freely available Heavy GCR TR001 profiles to the public - then that would be great. But nobody does. At least Chromix offered UCR and LT and MED GCR variants of TR001. But nobody seems to care about four colour greyscale work.

Another 'radical' option is to make your own greyscale conversion where the image is K only, then to fudge support screens in the CMY in the deep shadows/three quarter tones. Easy enough to automate.

Stephen Marsh.
___________________________________________________________________________   

Date: Wed, 5 Jan 2005 19:03:48 EST
   From: Dan Margulis
Subject: Re: Converting to B/W

Terry Wyse writes,

I'd create a "Custom CMYK" profile (I can't believe I just recommended
using the "Custom CMYK" engine)

I can't believe that in the year of our Lord 2005, that Custom CMYK is the only option Adobe chooses to give us for dealing with drop shadows, 4/c B/W, and the like.

and set it up for either "Heavy GCR" or "Maximum GCR" with a Black limit at
or near 100% with a total ink of not more than about 300%.

Not Maximum GCR. That would give you K only, no CMY at all. Also, an important note: By default, Custom CMYK underestimates black dot gain. That's not a big deal when using a skeleton black, but in this case it *is* a big deal.  Ideally one goes in and edits the black dot gain curve, but for these limited purposes Hal could just increase the overall dot gain setting by four points before separating.

As a side note, I recently did a bit of press profiling where I included a
total ink chart in the press run. Interestingly (or not), the highest Dmax's
were achieved where K=100% but the total ink was no more than about 260-280%.
At K<100%, the total ink amount could be much higher (340%+) but in no case did
the Dmax exceed the K=100%, TAC=260-280% combination.

Yup. If you're using a relatively heavy black, trying to force extra CMY into the shadow is counterproductive, as we discussed some months ago when various parties found it illogical that the black generation setting in Custom CMYK supersedes a contrary ink limit.

Dan Margulis
___________________________________________________________________________

   Date: Wed, 5 Jan 2005 20:46:20 -0500
   From: Terry Wyse
Subject: Re: Converting to B/W
 
I guess I was thinking more in terms of what "MaxK" or maximum GCR does in profiling applications. It will give you mostly black (linear curve shape) but it will add enough CMY to give you, in essence, a "Rich Black" style of profile. You'll get enough CMY to add up to your requested total ink limit in most cases. And of course a profiling app will compensate correctly for the K dot gain as this is known from the actual press characteristics/measurement data.

Cheers,
Terry

___________________________________________________________________________

   Date: Thu, 6 Jan 2005 08:50:30 EST
   From: Dan Margulis
Subject: Re: Converting to B/W

Terry Wyse writes,

I guess I was thinking more in terms of what "MaxK" or maximum GCR does
in profiling applications. It will give you mostly black (linear curve
shape) but it will add enough CMY to give you, in essence, a "Rich
Black" style of profile.

What people call that kind of profile is unimportant; the point is that any method of conversion that professionals are apt to adopt needs to have the capability of making that kind of separation as an option.

Also, I hope the profiling app has the capability of making the type of sep that Photoshop designates as "Maximum" GCR, i.e., no CMY ink at all where possible. That type of sep is also necessary from time to time, although not to make 4/c B/Ws. But, if somebody gives you an image of a comic strip in RGB and you have to separate it, you want those fine black lines to come out as 0c0m0y100k.

Dan Margulis
___________________________________________________________________________

   Date: Thu, 6 Jan 2005 14:57:11 -0500
   From: todie
Subject: Re: Converting to B/W

On Jan 6, 2005, at 8:50 AM, Dan Margulis wrote:

...if somebody gives you an image of a comic strip in RGB and
you have to separate it, you want those fine black lines to come out as
0c0m0y100k.

… so you'd start by telling somebody that the comic strip was inked wrong.

The line work should've been scanned at at least 600dpi (spi/ppi?:) saved as a bitmap TIFF, downsampled to 300ppi (or less), choked (the oposite of "spread" in old timer speak : ), inked (filled with color?), saved as another TIFF, than placed in Quark with the line scan in a color "none" picture box (set to overprint) on top.

Laurentiu Todie
m
___________________________________________________________________________

   Date: Thu, 6 Jan 2005 21:59:04 EST
   From:Hal Silverman
Subject: Re: Converting to B/W...Follow-up

Hello everyone!

I want to thank you for your input in this recent thread.

Initially, I had contacted the Prepress dept. of the printer for their input. It amazes me, on daily basis, what lack of knowledge there is out there. They were lost as how I should approach the issue.

I ended up creating a custom profile, using Heavy GCR, 90% black Ink Limit, and keeping the Dot Gain at Standard 17%. This profile was used after I did the initial b/w conversion using the Channel Mixer, and keeping the file RGB. I did the initial conversion to CMYK, and found a bit of Y in the files (1-2% higher than M). The images looked a tad flat too. I created a curve in the Y channel to neutralize, then another one in the K to bump the contrast. I did some proofs on the Epson, and was amazed at the tonal range of the prints. My client was thrilled too. I am eager to see the proofs from the printer, hoping my approach proves positive.

 I must pat myself on the back here, as I used to avoid CMYK like the plague! But the recent confidence I have gained is all due to the knowledge from this list, and Dan's book.

Best regards,
Hal

Hal Silverman Studio, Inc.
6511B Basile Rowe
E. Syracuse, NY  13057
___________________________________________________________________________

   Date: Fri, 07 Jan 2005 04:32:15 -0000
   From:Stephen Marsh
Subject: Re: Converting to B/W...Follow-up

Hal Silverman wrote:

I did the initial conversion to CMYK, and found a bit of Y in the files (1-2%
higher than M). The images looked a tad flat too. I created a curve in the Y
channel to neutralize, then another one in the K to bump the contrast.

Hal, you have made the decision that there is too much yellow.

WHY???

What are you making this decision off?

If the file was neutral to start with in RGB, then you convert to CMYK then the CMYK will be neutral, if you mess around with the CMY curves you can mess things up if not careful.

So, you converted to the legacy gray balance found in the old SWOP Coated settings - which is NOT TR001, but it is something that many folk are used to printing for historical reasons.

If you do know the final output condition, then you could assign this profile to the Custom CMYK Heavy GCR separation, then view the files numbers as LAB numbers and not CMYK, then you could adjust things with a more known outcome. But if you don't know the final conditions and you just decide to pull a yellow curve or whatever - how do you know what you are really doing?

The other way is to try to hit known neutral CMY aimpoints over the tonal scale, this way the 'profile' is in your head and not in Photoshop, and you adjust the CMY to match these target values, is this what you were doing when you pulled the Y curve?

With the legacy custom CMYK profile assigned to this image with the yellow curve, what LAB values do you get? I would presume that Photoshop now thinks this grey is no longer neutral gray. If you had to convert to RGB or to another CMYK and expected a true neutral for the new conversion - IT WOULD NOT BE NEUTRAL.

Stephen Marsh.
___________________________________________________________________________

   Date: Fri, 7 Jan 2005 08:35:44 EST
   From: Hal Silverman
Subject: Re: Re: Converting to B/W...Follow-up

Stephen,

I sort of understand what you are writing.

I was going strictly by the numbers. I did not check in LAB, but only in RGB and CMYK. You are right though....

The RGB values were identical after the conversion to CMYK, but I thought the Y and M needed to be identical as well. That's what threw me.  The Y was 1-2% higher.
I guess I took the wrong road!

Regards,
Hal


Hal Silverman Studio, Inc.
6511B Basile Rowe
E. Syracuse, NY  13057

p. 315.434.9121

Adobe Photoshop training classes are taught in the US by Sterling Ledet & Associates, Inc.