Dan Margulis Applied Color Theory - How Harsh a Black?

From: "PEarle@gpconline.com", INTERNET:PEarle@gpconline.com
Date: Mon, Feb 19, 2001, 4:33 PM
RE: [colortheory] Problems with black printer

It's as if a poor printer doesn't have enough problems. Today's "challenge" was a job we're printing from furnished CMYK files in Quark 4.05. We ripped the furnished page files to our Brisque and proofed them on Iris page proofs. Our customer OKed the page proofs. We imposed and output film, and the customer came in to check the job on press.

On press, one image had noticeable dark splotches in a "sky" area. The image was an artsy, blurred image of a marsh with a very golden cast overall. The sky area is in the range of 22c18m30y0k, except there are several areas (sort of dark islands in the sky) with 1% black. These areas were not noticeable on the Iris proof, but really stood out on the press-sheet. The CMYK image file reads 1% black in these areas. The black printer has very noticeable areas, corresponding to the dark splotches, that measure 1%, surrounded by areas with a slight dot but not measurable on our transmission densitometer. The balance of the area is solid on the negatives--0% dot.

We "solved" the problem by opaquing the most noticible of the dark splotches, and our customer accepted the others as part of the picture. The new black plate solved satisfied our customer.

To study the situation further, we converted the original CMYK file to RGB and then back to CMYK using our standard CMYK set-up. The 22c18m30y0K area converted to 196r190g163b and then back to 22c18m30y0k. However, the original file must have been been converted to CMYK with much stronger black-generation. It had significant amounts of black printing in highlight and mid-tone areas where our our converted CMYK image had no black in it. One area that was originally 25c29m46y15k converted to 36c38m53y0k.

From this, we think there are a couple of lessons: the Iris proofs were very close "color" proofs, but there were some areas which looked different on the press-sheet from the proofs. We are assuming that analog proofs (Fuji ColorArt) might have shown that 1% black dot, but we haven't checked that out yet. Further, using more than "light" black generation can lead to black being added to highlight areas in which the transition from 0 to 1 to 2 to 3 to 4 pixel-values in the black printer (out of the 256 range) becomes a very noticeable step from 0% to 1% on the film, plate, and press sheet. Before this experience, I might have argued that there is no noticeable difference between 0% and 1% on a press sheet, but in the black printer, there certainly was. (Could dot-gain have contributed to this difference?)

Have any of you had similar experiences with small but very noticeable black dots? Is this a good reason not to use much black generation in any images that contain large, light-colored areas? Any other thoughts or experiences?

Preston Earle
PEarle@gpcmail.com


From: INTERNET:sdelvecchio@ambrosi.com, INTERNET:sdelvecchio@ambrosi.com
Date: Tue, Feb 20, 2001, 9:06 AM
RE: [Fwd: [colortheory] Problems with black printer]

You also have to watch if you have done retouching and add noise to a CMYK image. Adding noise will add 1% in all color including Black in the highlight areas. I will usually use curves and -1% in the start point to avoid the problem below.

Steven Del Vecchio
Ambrosi & Assoc. Inc.
Chicago, Ill.


From: Dan Margulis, INTERNET:76270.1033@compuserve.com
Date: Tue, Feb 20, 2001, 8:39 PM
RE: [colortheory] Problems with black printer

Preston writes,

>>On press, one image had noticeable dark splotches in a "sky" area. The image was an artsy, blurred image of a marsh with a very golden cast overall. The sky area is in the range of 22c18m30y0k, except there are several areas (sort of dark islands in the sky) with 1% black.>>

That's a pretty weird sky. 22c18m30y is a beige.

>>These areas were not noticeable on the Iris proof, but really stood out on the press-sheet. The CMYK image file reads 1% black in these areas. The black printer has very noticeable areas, corresponding to the dark splotches, that measure 1%, surrounded by areas with a slight dot but not measurable on our transmission densitometer. The balance of the area is solid on the negatives--0% dot.>>

The issue isn't what the dot read on the film, but on the plate. A number of printers overexpose their plates just to be able to hold small dots, and 1% dots can sometimes become 2% dots.

>>Is this a good reason not to use much black generation in any images that contain large, light-colored areas?>>

Not if the black generation is sensible. Any such system introduces the black very gradually at first, to avoid just such problems as you are having. That is, if you are getting a jump of 1k in your island you should be getting a jump of 2c2m2y (or more) as well. So, given a sensible system, the island is more in the cmy than it is in the black.

The problem is, your client may not have been using a sensible system. Photoshop's black generation, even in its heavier incarnations, starts gradually, as it should. Third parties, however, occasionally don't understand the realities of presswork, and assume that they can just blithely start throwing black in as though it were any other ink.

Generally, for presswork one stays away from heavier black generation, except in certain classes of image. But the reasons for this don't have much to do with your particular problem, which is probably due either to an unexpected gain in the small dots during the platemaking process, or to a faulty black generation method on the part of your client.

Dan Margulis

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