Do CMYK builds have to be in 5% increments?

[HarveyDunn]

Hello, I'm a designer working on a color critical project. I'm using the GRACoL2006 profile for color management.

Today's question: do CMYK builds have to be in 5% increments? My understanding is that modern technology has moved on, and that builds can be in 1% steps, although it is best if you avoid 1%, 2%, 98% and 99% dots. The former will drop out, the latter will fill in. But numbers between 3% and 97% will be respected. However, I was at a service bureau today, and the guy there, who knows a lot about color, questioned whether I've got my fact right. The fact that the Pantone guides still move in 5% steps was brought up as a datapoint suggestive of the possibility that while you can ASK for 63%, 18%, 32%, etc., you should not be surprised if what you actually GET is 65%, 20%, 30%. In other words, always round to the nearest number divisible by 5.

 

[dabob]

To the best of my knowledge the simple answer is no - they can be any percentage including numbers like 12.45. That said, it will mostly be determined by the accuracy of your output device, if your going to film it won't be as good as direct to plate and the calibration of the output device will be critical. I remember years ago when an area printer went to a new ctp device they were so excited by being able to control their dot to a .5% increment - they said that they had almost stopped using tint varnishes and were just using cmyk builds. We have a heidelberg prosetter and can control our dot at least to a 1% accuracy.

I think that halftones would really look ugly if they could only reproduce dots in 5% steps . . . . .

 

[HarveyDunn]

"they can be any percentage including numbers like 12.45."

Is that why Adobe products like InDesign show LAB- or RGB-to-CMYK conversions with two decimal points?

 

[gordo]

Today's question: do CMYK builds have to be in 5% increments?

No. You can specify whatever percent you want.

while you can ASK for 63%, 18%, 32%, etc., you should not be surprised if what you actually GET is 65%, 20%, 30%. In other words, always round to the nearest number divisible by 5.

Here's the thing....what you ask for in your file you probably won't get on film or plate because most print shops don't output to linear (film or plate). The service bureau might out put linear film though.

Best, gordo

 

[HarveyDunn]

Does that mean the print shop is going to give me the color I ask for in my file, just perhaps via a different route? I realize that for any one color there can be many ways to build it.

 

[Al Ferrari]

Hi Gordon,

Please explain what you mean by alluding to "other than linear" output.

Thanks,

Al

 

[gordo]

Does that mean the print shop is going to give me the color I ask for in my file, just perhaps via a different route? I realize that for any one color there can be many ways to build it.

Well, dot values are tone, not color, values. So, the final color you get will depend on, among other things the ink hues being used and the tone reproduction your printer has targeted.

Gordo

 

[gordo]

Hi Gordon,

Please explain what you mean by alluding to "other than linear" output.

Well, when it comes to CtP plates, the tones on the plate are not the target. The tones in the presswork that the dot values on the plates deliver are what is important. So CtP plates may or (more likely) may not be linear. I.e. if you ask for 50% in your file you probably won't get 50% on the plate. You'll get whatever tone on the plate is needed to deliver the desired tone in the presswork for a 50% file request. Probably aabout a measured 68% in the presswork.

Gordo

 

[HarveyDunn]

Can you explain what the word "tone" means in this context?

 

[Al Ferrari]

The concepts in your response make perfect sense, but the last sentence ruins it. Perhaps you mean that a 50% on the plate produces ~68% in the presswork. But it would have served us better if in your last sentence you had guessed at the plate value that would produce a 50% in the presswork.

Thanks,

Al

 

[Lukas Engqvist]

% is relative, So 50% of a colour has no meaning if you don't know the colour (including the intensity/luminosity of black). 50% of a black on one paper may be the same appearance as 60% of another black in another printing device/substrate, this is the reason for colour management. Defining the colour as precise as possible helps reach that goal.

As to the decimals, think of it as exchanging currency, in the calculation you may have decimals, but you may find you loose the decimals when you move to cash, but the decimals help to correctly round up or down. How smooth the below 5% and above 95% are varies greatly depending process.

 

[gordo]

The concepts in your response make perfect sense, but the last sentence ruins it. Perhaps you mean that a 50% on the plate produces ~68% in the presswork. But it would have served us better if in your last sentence you had guessed at the plate value that would produce a 50% in the presswork.

I was trying to explain the basic concept. Once understood - guessing is not needed.

Gordo

 

[gordo]

Can you explain what the word "tone" means in this context?

Tone or dot area - the final percent achieved when a percent value of an ink is specified by an authoring application.

Gordo

 

[rich apollo]

do CMYK builds have to be in 5% increments?

Actually, all your file values have a resolution of 8 bits per channel. This means that each channel (C,M,Y,K or RGB) has 256 possible levels (0-255). Your mid-point falls somewhere between levels 127 and 128. The 50% you specified isn't even a fifty.

The output devices (imagesetters and platesetters) operate on the same basis, with the exception that some objects, like smooth shades (vector-based gradients) can be output at 12 bits per channel on some systems.

I have personally printed on press what measured as a 0.5% dot on the plate. Once you get on press, you probably won't be able to distinguish a 1% from a 2%, and from 95% on up you run the chance of the screens plugging, depending on how much ink is being laid down.

The pressroom is not a digital environment. You're talking about applying a viscous, oily liquid to a supple substrate of .003" thick under tons of pressure at a rate of about 10,000 times per hour. And the only reason that works at all is because oil and water don't mix.

 

[Bill W]

In 1991 when we purchased our first imagesetter, I thought good, now I can make a 50% in the file print 50% on the press. Through fingerprinting, flexo, we discovered that a 33% printed a 50%. So I created a curve to pull 50 to 33 and printed my 50% in the file as a 50% on the label - big mistake. When I did that my images looked too light. A bit of research showed that, as Gordon points out, the world is used to seeing a 50% in the file print as approximately a 68% on press.

Photoshop's default is to show the approximate 20% gain the the world expects, so your images ought to look close, on a calibrated monitor, to what is going on a press that honors the 20% gain.

It is possible that the individual at your service was referencing what I have seen in most printed tint books - values in 5% increments. It would be way to time, and material, consuming to print a book that has printed CMYK patches in 1% increments. Most books I have seen, and the one we printed years ago only shown CMY patches as adding K to the entire book was cost / time prohibitive. We did run 1 - 10% tints because of the higher dot gain in flexo, but then ran 5% increments up to 100%.

 

[Stephen Marsh]

Bill, might be on to something here, it is common for printed colour reference "look up table" guides to use a 5% or 10% step. An example of such a guide can be found here:

http://www.fho-emden.de/~hoffmann/swatch22112002.pdf


Stephen Marsh