Fact is that what the entire G7 process began as an attempt to duplicate in offset lithography the linearization process that's already included somewhere in the routine of every RIP-driven digital printing device.
I disagree, Mike. The G7 process attempts to control a press in the same manner that you would a scanner. Maybe I'm being pedantic. If it seems I'm hair-splitting, I apologize. RIPs don't generally have a process to correct, or adjust, gray balance, and that's a fundamental part of scanning and G7 calibration.
Originally Posted by Correct Color
Once you're creating dots digitally it's to all that difficult of a process to establish a routine to do some sort of calibration/linearization (the terms loosely used somewhat overlap.)
I think it's very important to distinguish between the two terms. "Calibration" is the process of setting up output parameters. "Linearization" is the process of setting those output parameters to a straight-line response of some sort - for example, getting a 30% file value to measure a 30% on the plate. "Linearization" is a more restrictive term, and falls under the umbrella of "calibration".
Originally Posted by Correct Color
Of course it's much harder to do on a litho press because of all the variably factors involved, and up until digital plate making, it was pretty much impossible.
In the days of matchprints, the proof was directly tied to the film. 3M, then Imation, then Kodak, engineered the donor sheets to exhibit a TVI response. Divorcing proofing from the final output granted us great freedom to manipulate the appearance of each. Now we can optimize the proof and the plate separately.
Originally Posted by Correct Color
And understand that color management on a digital device is something that has to be done. It doesn't come in the box from the factory. If you want to attempt to hit a Gracol proofing standard with a digital device, first you have to calibrate and characterize the device,…
Not exactly. One might linearize the platesetter to ensure correct exposure, et cetera. I actually quit linearizing platesetters long ago. I get better results with the natural response of the plates. You must linearize an inkjet proofer before profiling. But, you don't 'linearize' a press.
It sounds like you're operating from the position that dot gain is a bad thing. It's neither good, nor bad, it just is. You can set up your platesetter so that a 30% in the file will yield 30% on press. I've done it, and it looks like crap. Trust me, we like dot gain. Printing without dot gain looks hollow and sharp. I've never ended up using linear plates. Some level of adjustment has always been applied to pull the plate out of linearity.
If you look at the ISO 12647-2 TVI curves and/or the G7/GRACoL NPDC, none of them are 'linear'. They all exhibit non-linear responses. What you're attempting to do is get your presses' dot gain responses to align with the dot gain of the standard you want to match.
In linearizing my goal isn't to directly make the output look a certain way, but rather to be able to control what that output should be. You get your input PDF already in terms of how how a certain press should respond, whether that's as "SWOP" or "Gracol" or whatever. But if I'm going to emulate any of those characteristics on my press, then step one is understanding the actual characteristics of my press, right?
Sure I could just leave it alone and say "Well, I'm running an offset press, that should be kinda like the theoretical offset press used for SWOP standards. I should be OK, then." And yeah that would kind of work, but if I want to do better than "kind of" I've got to actually understand the true response of my press so I can compensate for the differences between it and the theoretical press.
In linearizing my goal isn't to directly make the output look a certain way, but rather to be able to control what that output should be. [SNIP] But if I'm going to emulate any of those characteristics on my press, then step one is understanding the actual characteristics of my press, right?
Sure I could just leave it alone and say "Well, I'm running an offset press, that should be kinda like the theoretical offset press used for SWOP standards. I should be OK, then." And yeah that would kind of work, but if I want to do better than "kind of" I've got to actually understand the true response of my press so I can compensate for the differences between it and the theoretical press.
Rich has explained it well.
You do not need to linearize your plates in order to understand the response of your press.
If you linearize your press response you actually mask its response.
When a CtP device is calibrated the exposure and processing is set to provide a robust ink accepting dot on the plate and a clean non-printing image area. It will also be free of artifacts such as laser swath banding. Once calibrated the plate is not likely to be linear.
That non-linear plate will have a certain tone response on press when the press is run to the appropriate SIDs (or Lab values).
That tone response (among other things) will tell you how well the press is transferring ink.
It will also give you the information you need to build a tone reproduction compensation curve that will be applied to the plate in order for the plate to generate the desired tone response on press. The desired press tone response is probably going to be the ISO 12647 standard - which itself is not linear.
BTW, SWOP is not a press standard. It is a specification for the inputs (e.g. proofs) destined for heatset web presses.
A PDF - I hope you will find of interest and value - BUT
It might muddy the water - !
Regards, Alois
Interesting but very confusing if you follow the instructions. It starts by saying that you pick a point on the curve in quadrant 1 and then draw a line horizontally to the curve in quadrant 4. Well, if you draw a horizontal line from quadrant 1 - it can never touch the curve in quadrant 4 so you're stumped from the beginning.
That being said, I think the principle might be akin to the explanation of how to build press curves explained here:
Hmm... interesting instructions. I have to agree with Gordo though that something appears funny. Either the description is a bit off or someone has gone and relabeled the Quadrants.
It does show the idea, though, except that this is a bit old-school. It looks like the instructions are intended to help someone deal with photo-negative reproduction from an original hard-copy. Largely the same sort of problems I think, but at the same time not quite the same thing; extra steps one way, missing considerations another.
I'm kind of thinking that everybody here is basically saying the same, thing, and that most of the smoke is a matter of terminology. I'm suspecting that the term "linearization" maybe has some meanings I'm not aware of.
I do prefer to think of things that way, though. It seems easier to me to think of a the press as a linear device, and consider the delta between linear and the theoretical dot-gain curves for a profile, rather than to try to think of two separate curves and the deltas between them. If there's a problem with that, I don't understand it.
I'm not actually having such big issues with getting a decent color match; it's more keeping a good color match when my back is turned that's been challenging.
Are there any really excellent color management books out there? Theory, and best practices? What little I do know basically comes from the PDF spec, RIP manuals and the frequently contradictory online tutorials.
I'm kind of thinking that everybody here is basically saying the same, thing, and that most of the smoke is a matter of terminology. I'm suspecting that the term "linearization" maybe has some meanings I'm not aware of.
I do prefer to think of things that way, though. It seems easier to me to think of a the press as a linear device, and consider the delta between linear and the theoretical dot-gain curves for a profile, rather than to try to think of two separate curves and the deltas between them. If there's a problem with that, I don't understand it.
Not to put too fine a point on it, however it is important that terminology is used correctly (or defined when used) otherwise comments are easily misunderstood - especially for those whose first language is not English.
The terms calibration and linearization do not mean the same thing - especially in a CtP workflow - so they shouldn't be used interchangeably.
A press is not a linear imaging device.
What is important is that a specific tone request in the original file results in a specific tone measured on the presswork. E.g. Where I ask for 50% in my file it should measure 64% in the presswork. If one is printing to a published standard like ISO 12647-2 then the required final presswork tone values are specified for the tone range for each color. So, I know what tones are requested in the original file and I know what tones the calibrated but non-linearized plate delivers on press and with those two data points I can build a tone reproduction curve to be applied to the plate that will map the requested tone values in the file to tone values on the plate that will create the required tone values in the press work. It's very simple and straightforward. Linearization and ICC profiles are not involved.
I'm kind of thinking that everybody here is basically saying the same, thing, and that most of the smoke is a matter of terminology. I'm suspecting that the term "linearization" maybe has some meanings I'm not aware of.
Yeah, that's pretty much it. Actually of course and as all these guys have said, linearization has one literal meaning...and calibration has another literal meaning. However, in digital printing, 'calibration' has come to usually mean doing a linearization routine.
That's what I meant when I said the two terms have come to be somewhat overlapping.
Since I stay almost exclusively in large and grand format these days I'll certainly defer to Rich and Gordo in their expertise in the nuts and bolts of the litho area, but the point is that by whatever name you call it, the point of all these routines is to get neutrals under control. The issue was always that if you could define the start point as zero, the chroma value of the ink as a constant, your max density as constant, and the paper white as a constant, then if you could define the neutrals consistently, you'd have printing as a true commodity, one printer absolutely indistinguishable from another.
How exciting.
Mainly that's why I stay in large and grand format, where that mentality has yet to take hold.
LinkBack URL
About LinkBacks
Originally Posted by Correct Color
Reply With Quote
