New to color management

[jfryw]

First of all our work flow was set up by someone who is no longer here. Now I am questioning it and am trying to understand all the pieces. The thought that I am going over is wondering if the files that are sent to the CTP and proofer correctly processed.

Since I am fairly new to color management and I wanted to put this out there for comment. And this is what my basic understanding is on rendering intents.

Relative Colormetric - produce colors as accurate as possible. Colors are remapped so white point paper white will be paper white.

Absolute Colormetric - all colors as accurate to original as possible. This produces a scum dot on proof and press at the white point.

Perceptual - I have heard to things with this and am unclear about which direction to lean. 1. should be used when making separations. or 2. is used with film to keep image tonality

Thanks for the input

 

[kyle]

jfryw:

Your descriptions of relative and absolute colorimetric rendering intents sound pretty good. Relative colorimetric and absolute colorimetric both use the same look-up table to transform colors. When an absolute conversion is done, the color management process performs the same steps as it would with a relative conversion, but it just nudges all the values to compensate for the white point (which is encoded in the profile as a single color value). It can add ink to white, and it can also remove it, depending upon wether the destination white point is lighter or darker than the source. The colors are remapped to make the white point pure when the profile is originally created, then the real white point is recorded in the profile. When you choose an absolute intent, the colors get shifted back using this information.

The perceptual intent is assigned the task of preserving color relationships when going to a smaller gamut, and to take advantage of larger gamuts by expanding into them. Because the other color space is variable and cannot be known when a profile is created, the perceptual intent of many profiles is not very useful. It is therefore very appropriately named "perceptual intent" instead of "perceptual result." Because perceptual sort of means "color preservation be damned, make it look as good as possible," it is of no use for proofing purposes, but may yield preferable results when color-managing upstream from the proofing.

Relative colorimetric is generally the best for proofing, unless you want to use different profiles with accurate white points for different stocks, have more reflective proofing paper than most of your stocks, and want to print dots where there are none to try to match the color of the paper. There are advantages and disadvantages to each.

In regard to making separations, and tone values on film: rendering intents are irrelevant and not applicable. The standard rendering intents come from the ICC specification, and therefore apply only within the scope of color profiles. Color management for proofing happens before film/plates are output to represent the color that will result if you print what you have ready with your film/plates, using your curves, on your press, with your ink, and your stock using the same processes and under the same conditions as the press run that the profile was based upon. On the first monday after a new moon in November. This kind of color management only affects how accurately your proofs match your press sheets. The other kind of color management happens upstream, and determines what colors you actually get on press by altering the color of images and other elements.

The only related concept applicable to separations and film is curves, also known as plate curves or press curves. You could think of these as single-color color profiles, because they compensate for device-specific and process-specific distortions of a subjective ideal "device-independent" response.

 

[Grnofslt]

I wish I could help you with the color management directly, but the company I had worked for didn't really need color management for the work we produced, All of our artwork was black and white. Also, I didn't know that much when I went into the pre-press department there, most of what I know now, was through tutorial videos, taking what was being taught on those videos and applying it to the work flow that we had at work.

That being said, my advice to you would be for you to keep accurate and detailed notes on all the steps of your work flow processes, step by step, hints and tips as side bars. Create a manual for your work flow and lay it out as a book using the software tools you have within your work flow. Many might scoff and say that it's unnecessary, but in the long run, you might solve a problem once, and not see it again for a long time and have to struggle to remember what it was you did when you had that problem in the past.

Also something like what I've mentioned above may be a very useful tool when you have to train someone as a co-worker or as something for a new person who comes into your work flow to learn from if you find a new and better job.

Also, if you create something like this, it really is a portfolio piece for prospective employers to look at, showing your expertise with the software, and how you handled problems and laid out a template for fixing problems as they repeat.

It's just a suggestion, one that I am following for myself

Bill J

 

[MickyJ]

jfryw:

Color management for proofing happens before film/plates are output to represent the color that will result if you print what you have ready with your film/plates, using your curves, on your press, with your ink, and your stock using the same processes and under the same conditions as the press run that the profile was based upon. On the first monday after a new moon in November. This kind of color management only affects how accurately your proofs match your press sheets.

Very good description. I'm going to use this quote from now on whenever someone complains that my proofs aren't matching the press. But what can you do when the press guys tell you they don't need to calibrate their measuring devices.

 

[telenow2002]

jfryw the rendering intents are well documented all over the web and in so many books, what they rarely do is help you understand when you should be using one or the other and that really depends on the the type of work you are producing.
Since you are questioning the workflow as well you should provide us with some more detail so we can help you with your questioning. As you are probably realizing, colour management is a mammoth task and the understanding of workflow based CMM and it's application is something many claim to be experts in, but very few really are when they need to apply it across multiple disciplines (sorry all you CMM experts!)

And MickyJ if your press guys think they need no calibration then I would be questioning their process control procedures and quality assurance processes as without these they are in no position to question the quality of a proof when they are not measuring the quality of their print process!

 

[telenow2002]

jfryw the rendering intents are well documented all over the web and in so many books, what they rarely do is help you understand when you should be using one or the other and that really depends on the the type of work you are producing.
Since you are questioning the workflow as well you should provide us with some more detail so we can help you with your questioning. As you are probably realizing, colour management is a mammoth task and the understanding of workflow based CMM and it's application is something many claim to be experts in, but very few really are when they need to apply it across multiple disciplines (sorry all you CMM experts!)

And MickyJ if your press guys think they need no calibration then I would be questioning their process control procedures and quality assurance processes as without these they are in no position to question the quality of a proof when they are not measuring the quality of their print process!

 

[David Ingram]

With regard to final output, are you outputting all plates/ negs as linear, or do you you use a curve to cutback the tone when printing on uncoated? Do you have different setups for different stocks? I found the perceptual worked well when going to a copier device, otherwise relative worked best for final output. What I've found is MOST important, coming from the photoshop 3-7 days was JUST DON'T CHANGE THE PROFILE! whatever the scan comes in as, KEEP it! The customer/agency / photographer KNOWS what they HAVE seen, changing profiles within photoshop is usually a no-no.

 

[jfryw]

Since you are questioning the workflow as well you should provide us with some more detail so we can help you with your questioning. As you are probably realizing, colour management is a mammoth task

I would very much agree that this is a mammoth task!

I began questioning our workflow because I read that Perceptual should be used for film output. (sorry I dont remember where I read that). Currently in our Prepress anything that comes through the door is converted to CMYK using Perceptual. Then to the proofer or to CTP.

I decided to test this in PS and duplicate the conversions that take place in our workflow. I went through the conversions from common profiles that we receive to our press profile. I was a little surprised to see how much change took place with the conversion using Perceptual compared to Relative Colormetric.

Then I scratched my head and asked myself don't I want to a conversion to produce the most accurate colors possible? And thats when I began to question if Perceptual was the correct choice because Relative appears to be more accurate.

Like I said before I am new to color management and I just hope that I havnt taken to big of a bite before I really understand this. Thanks for the input.

 

[Lukas Engqvist]

The amount of change that takes place in relativa and perceptual is very different depending on the application that made the ICC. We use Relative and blackpoint compensation for coated papers. But when going to uncoated papers there are benefits of going with perceptual.
One vital step is to realise that what colours are in the risk zone. What you will find is that out of gamut colours are thos that change most when changing profile or intent.

One could say that perceptual is the way to go for those that don't proof colours, it is the "lazy" compromise to make printable any image. Here I would argue with the comment not to change profiles in Photoshop. You cannot assign profiles without changing colours, but converting before editing is not allways a bad idea, especially since some RGB profiles have weaknesses (eg sRGB has weakness in Cyan, Adobe RGB has a different whitepoint than printning standard).
We find that often when a customer complains that they want more from a picture convertinig them from sRGB to ECI RGB and tweaking the image (pushing it to the edge of the printable gamut, (working in Preview ofcourse) letts us maximise the images.

One of the most common errore though is that images with too much noise will topple your workflow because there is a dither of unprintable colours even thoug the average colour is printable.

You should split your process in two:
1
Prepress making good repro. Printable files with a good GCR that will give you a stable print run without too much ink coerage per paper. Use standard measure data (for your region) to generate your ICC profiles.

2
Havinig a press workflow, and quality management that means you have a solid repetetive environment that produces consistent results that you continually push to get closer to the regional standard for each paper.

 

[J_Lindgren]

Jfryw, you are definitely on the right track with your reasoning.

To help you clear this out in your mind, let me add my 2 cents worth on practical application of these intents.

As a rule-of-thumb, Perceptual intent is aimed for such colour conversions, where the difference in source and destination gamuts (the space of reproducible colours) is large. In practice, this usually means going from RGB to CMYK. In this conversion, the Perceptual intent will endeavor to keep the *visual* relationship between different colours of the original image, and *not* trying to reproduce colours 1:1. If it did, any RGB colours that were out of the gamut of your destination CMYK colour space, would be clipped, resulting in abrupt tonal change followed by a flat tint instead of smooth transition from "dull" colour towards a highly saturated colour.
The Perceptual intent will proportionally "scale down" the larger source gamut, to make it fit within the smaller destination gamut. This will usually result in *all* colours of your image to change (more or less), but - when examined visually - the images will *appear* similar as the interrelation of colours is retained. If examined side-by-side, you will see - as you have! - a distinct difference in all colours. This is to be expected due to the "gamut scale-down" function.

Both Relative and Absolute intents are aimed for conversions, where the gamuts are virtually same (CMYK -> C'M'Y'K'). These intents will reproduce any colours within the destination gamut 1:1, while any colours outside the destination gamut will be clipped. (Exceptions do exist...)

Additionally, as others already explained, Absolute intent will keep the white point of the original *absolutely*, producing scum dot on your proof/print if the white point of the original were darker than the white point of your proofer/print paper. But when compared side-by-side, the images will be 1:1 in all such colours, that are common in source and destination gamuts.

Relative intent will remove the effect of the white point of the original (think of it as the underlaying paper of a printed image), and remap it to the white point of the destination.

When I explain this to my trainees, I ask them to imagine it as if you peeled off only the inks from a paper, and laid them down on another paper. If there is a significant colour/lightness difference between these papers, the same difference is reflected in all colours. This is naturally mostly visible in light tones.

Hope this helps!