Poll - Do you linearize your plates?

[seejay]

I have to say that I am of the same opinion as Todd on this one.
The driver for us is cost and time. When we do recalibrate, we can be talking about 1/2 a days work per press by the time we output plates, optimise the press, run the test sheet, measure and calculate TVI - reoutput with adjusted curves then verify - for a minimum of 3 stock types: matt, gloss & uncoated. Realistically, if we were to carry out this process for non-linearised plates from both ctp devices my gaffer would be asking some serious questions.
further, whilst neither of our CTP devices has drifted over the last 10 years, when we did invest in a replacement CTP 2 years ago the transition was virtually seamless. Install the device, linearise, overlay the press curve then we were good to go.

seejay

 

[gordo]

@seejay: You lost me.

If you were using a press curve instead of a press curve on top of a plate curve then your process would be (paraphrasing your words) - when we did invest in a replacement CTP 2 years ago the transition was virtually seamless. Install the device, adjust the press curve to bring the plate dots to the same as the previous plate condition then we are were good to go.

I'm not quite sure I understand what you mean by "recalibrate" however, it seems to me that the only difference might be a bit of extra press time and paper. All the other actions ("output plates, optimise the press, measure and calculate TVI - reoutput with adjusted curves then verify") would have to happen whether you linearize first or not. Or maybe I'm misunderstanding.

best, gordon p

 

[seejay]

@seejay: You lost me.

If you were using a press curve instead of a press curve on top of a plate curve then your process would be (paraphrasing your words) - when we did invest in a replacement CTP 2 years ago the transition was virtually seamless. Install the device, adjust the press curve to bring the plate dots to the same as the previous plate condition then we are were good to go.

I'm not quite sure I understand what you mean by "recalibrate" however, it seems to me that the only difference might be a bit of extra press time and paper. All the other actions ("output plates, optimise the press, measure and calculate TVI - reoutput with adjusted curves then verify") would have to happen whether you linearize first or not. Or maybe I'm misunderstanding.

best, gordon p

Yesterday was a long day - it was late in the day when I posted. The second paragraph should have read 'install the device, adjust the plate curve....' I'll try and clarify our reasoning:

Consistent colour for us is critical. We print a lot of repeat work where content changes, but corporate/house colours are consistent from issue to issue. We have to have a means of ensuring that colour doesn't drift regardless of what CTP device it is output from.

We monitor press results on a regular basis - daily spot checks for tolerance to TVI standard, quarterly checks using our calibration file. If the measurements have drifted (for whatever reason), we make an iterative adjustment to bring the curve back into line if the drift is small, or if the drift is more significant, we do a full calibration from scratch.

Using your methodology - because the 'non-linearised' characteristics of the two different devices are so different we would have to run each calibration using plates from both CTP devices as follows.


Step 1
device A - one set of linear plates (i.e. no press curve applied to the plate) for coated stocks (gloss & matt), then a second set of linear plates for uncoated stock.
device B (which prints sharp in highlight and mid-tone) - set of linear plates for coated stocks (gloss & matt), then a second set of linear plates for uncoated stock.

Step 2
measure results from device A for gloss, matt & uncoated; adjust curves; output 3 sets of verification plates for gloss, matt & uncoated curves - print & check
measure results from device B for gloss, matt & uncoated; adjust curves; output 3 sets of verification plates for gloss, matt & uncoated curves - print & check

By linearising the plates prior to starting the tests - we then know that both devices are outputting the same results to press.
In doing this, we half the work in verifying the results 1 set of plates for each of the 3 stocks.

My opinion is that if we had two devices that have the same/similar characteristics then there would be no need to linearise. However, because our devices are so different, if we didn't linearise we risk less consistent results from month to month/issue to issue.

Logisitics are also a consideration - linearising the plates means only one set of press curves to manage. This means less choice for operators to choose = less chance of selecting the wrong curve set.

hope this makes a little more sense.

S

 

[gordo]

@ seejay

I understand.

Let's for a moment put a wall up between prepress and pressroom.

You've already said that neither of your CTP devices has drifted over the last 10 years, but you do have two CtP that have quite different natural uncalibrated plate response curves.

If you do not linearize your plates before applying a press curve - then once you've established the appropriate press curves (one for each device) - the plates will have the same dots on them. Different press curves resulting in the same final dots on plate. If you check your plates on a daily basis and there is any drift (which you say you don't have) it might show up in a change in the dots perhaps on one of the CtP devices. Then you would first try and determine why that's happened, and after that investigation you might decide to alter the press curve on that device to bring the plate dots back to your standard. In this case the press curve and the plate curve are in fact the same thing. There is no need to do a calibration run.

Now we jump over the wall.

If the presswork is drifting and the dots on the plates are correct then you need to determine what in the pressroom has changed that might have caused a shift in presswork since the plates have probably been eliminated as a cause since the dots on them are the same as they always have been. There is no need to do a calibration run. There may be a need to do a test run to understand why the press work is different when the plates have not changed. I would use my "Golden Reference" as the tool to investigate what's going on. I would be very concerned about doing iterative changes to plate curves in order to correct press problems.

The process that I would follow to set up the initial press calibration curve is not quite how you've described it.

Assuming that the two CtP devices have a different natural uncalibrated starting points then:

Step 1
device A - one set of plates with no curve applied to the plate for coated stocks (gloss & matt) and uncoated stock. Total of 4 plates.
device B - one set of plates with no curve applied to the plate for coated stocks (gloss & matt) and uncoated stock. Total of 4 plates.

However, you could combine device A and B into one press run (i.e. 4 plates only, one press run) in one of two ways:

1) Since you are building dot gain compensation curves, depending on what CtP devices you have you may be able to combine the two CtP outputs into just one test run by double burning. I.e. expose your test elements onto one half of one plate exposed in device A then expose the same elements on the other half by exposing it in device B.

2) Alternatively, it is also possible to create curves for CtP device B based on the results of CtP device A. In that case you would not have to go on press with plates from device B. You would use the response of device A to build a curve for device B. It's basically the same method as building any other tone reproduction curve.


Step 2
measure results from device A build curves then output 3 sets of verification plates for gloss, matt & uncoated curves - print & check
measure results from device B build curves then output 3 sets of verification plates for gloss, matt & uncoated curves - print & check

As in step 1, again you may be able to combine the two CtP outputs into just one test run by double burning. I.e. expose your curved test elements onto one half of one plate exposed in device A then expose the same elements with the different curve on the other half by exposing it in device B.

If you are successful with the results from device A then, because of your confidence in what you've done with device A and because the initial difference between device A and B are not that great, then you may decide that verifying device B with a press run is not needed.

best, gordon p

 

[kyle]

One curve. In our calibration process, we ran a test form that was ripped with no curve, and the response of the plate/press combination was measured. If (number of different platesetters or plates used) X (number of different press & paper combinations) is greater than the sum of those values, you would need fewer curves with a two-curve system. Using two curves does make the process of switching plates/platesetters easier, but with one curve you can always just do the math on readings of linear plate tests before and after the switch.

Our RIP (Apogee Prepress) quantizes to 256 values at every step, so having more than one curve applied would cause more levels to be skipped over, and would therefore increase banding.

 

[vaithi47]

seejay
i understand linearisation as 50%onfile to produce 50% on plate.this is to make sure that your exposing unit and your processor are in unison.suppose the plate shows 46% you have to do curve correction and do the test again to get 50 as 50.so this has nothing to do with the no of presses or no of kind of paper.if you have n no of ctp /processors you have to expose /develop n no of plates.plate sizes you expose need not be the press size,they can be smaller like a gto size plates,only they have to be the same make as your press plates.so linearisation of one ctp/processor takes may be 15 mins.the cost is the cost of a gto plate.if your ctp/processor have not much of fluctuations this can be done either once or twice a week.but this clears the ambiguity that if anything going wrong with your tvi is only because of your press and not your ctp/processor

 

[meddington]

Wow. I gotta say I'm totally surprised by the results of this poll. I would think that where consistency has been proven, there would be a move toward reducing the potential errors by eliminating the unnecessary step of curving a curve. In the case where two or more CTP units are in use, assuming they are proven consistent, press curves can be built off one press calibration run using one CTP workflow (CTP-a), and compensating the press curves for the other CTP devices (CTP-b, CTP-c, etc) by the value they differ from CTP-a's plates that were used on the calibration run. So if CTP-b is 5% off from CTP-a, compensate by 5% in the press curve and verify that your plates off both devices measure the same after imaging. Heck I'd at least try this before throwing on another press calibration run.

The reasoning that linearization with multiple CTPs allows for a more simplified workflow as only one press curve (pre paper type, per press) needs to be maintained, or only one calibration run needs to be made seems specious to me. Either each CTP has a custom linearization curve (+ press curve), or each CTP has a custom press curve (built/compensated off of the one calibration run on CTP-a). As long as you know which device your throwing your data at when creating 1 bits, I don't see an advantage.

 

[Ozkan Hangisi]

Hello friends,

In these very hot days i faced with a hot subject here and wanted to write a few words. I am working in hundreds of printing houses for standardization and voted for linerazition of plate. As some colleagues here said it provides a good start point. Especially when printinghouse has more than one CTP device, one plate set from one CTP is enough for dotgain calibration. ( I assume CTPs r well adjusted and exposing same brand plate). Especially if the customer is using Photopolymer plates (negative), linerization is a must. Because without any plate curve, 50% area becomes 60%-65% with ideal expose power and it may harm smoothness of degrade on printed sheet.

But if the plate values is consist for weeks and uncalibrated plate dot values are not so far from actual values then removing lineraziton from system is considerable.


Regards

 

[disbellj]

Now, I just voted and said I use an uncalibrated plate with a curve, but I need to explain one thing:

I would linearize the plate first and then apply a curve to it, but my plates are already naturally linear within tolerance of 2 (tolerance given by platesetter manufacturer), so no need to linearize. Very cool when it works out like this. Means I set the press up once for coated house stock using G7, once for uncoated house stock using G7, and just have to make sure we stay within tolerance of 2 at platesetter for plate with no curves, and are able to use official (beta for uncoated) ICC profiles from IDEAlliance. Oh yeah, I do have to spend very little time once a week to make sure my proofer calibration check shows it's in tolerance (always around or under 1dE unless needs calibration).

I guess since I would linearize if I needed to, maybe Gordo needs to change my answer to the first option in the poll instead of the second one which I chose when asked.

Regards,

Don

 

[disbellj]

gordo,

This is where I disagree with you.

The person has 2 CTPs, and 3 plate curves. This sounds good/right.

Let's say that with linear plates, that when linear plates are put on press and G7 applied, the result for prepress is that we have our (press) plate curve. There is now NO NEED to set up the press again. What is left to do? Make sure the plates are linearized without that "press" curve, and then apply the "press" curve to those linear plates, and you have the final values on plate that you need at press.

So we have two platesetters, which might need two different "calibration" (linearization) curves. So if sending to platesetter #1, you send data that has had "platesetter #1 calibration" curve and "press" curve applied. To platesetter #2, you send data that has had "platesetter #2 calibration" curve and "press" curve applied.

If you didn't have the linearization step, then you would have to keep up with "two sets of numbers", the set of numbers that got you to target on press - for each press. Instead, you don't even have to keep up with those numbers at all. You just have to make sure the plate is linear when the "press" curve is not applied, and when the press curve (which never changes - or at least we ran here for many years without changing our "press" curves until I recently set up using G7, and now look forward to not having to set/change "press" curves again. Note here: For calibration of plates to linearize them, plates are output and measured. No need to press run on a continuing basis, just once to get press using G7. After that, no press runs needed).

Now here where I work, I have 2-4 curves, for one platesetter. One is "press" curve for 150 linescreen uncoated paper jobs. One is "press" curve for 175 linescreen coated paper jobs. Now these two curves aren't changing, as long as the press can print within tolerance and my plates are linear without these "press" curves. Also, if my 150 or 175 linescreen wasn't within tolerance of being linear, I could make a "calibration" curve for each linescreen, so could possibly have 4 curves being used in my system (up to 2 at a time for a job, depending on paper type), but can have only 2 - one per paper type (like we do, since linearization is not needed).

So really, it comes down to what do you want to "keep track of":
a. At least one set of numbers with the "press" curve applied, to be the master numbers you'd always want to make sure your uncalibrated plate with "press" curve applied measured within tolerances to.
-OR-
b. No numbers to keep up with, just make sure to output a plate with "calibration" curve applied, but not "press" curve applied, and make sure dot on plate is within tolerance to dot in file.

b is easier, and separates prepress and press, in case a problem arises.

If a problem arises, when doing b, prepress can just do b procedure and tell if the problem is in the platesetter or not, or if the problem is at press. Easily.

If doing a, can prepress tell if problem is in prepress or at press as easily? Not in my opinion. Why? Because they are tied together.

Linearization provides separation from press, and therefore quality control for prepress dept better (made easier) IMHO.

In fact, Kristian said it the best of all.

Regards,

Don

Yikes! I'm becoming an apologist for natural state plates.

You're saying you have two curves to linearize the plates (one for each CtP). Then the same press curve is applied to both.

Instead, by starting with an unlinearized plate you would have two press curves - one for one CtP and another for the other CtP so that when you run your test form you get the same press result from both sets of plates. Two curves instead of three to achieve the same result. Much less chance of error or compounding errors. The linearizing curves are not providing any value to you.

Make sense?

best, Gordon p

 

[gordo]

@ disbellj

GP: I'll go through your points and try and see where the differences are.
First to clarify terms so we're talking about the same thing. I use the term "plate curve" to mean a curve that changes the tones on plate to some desired tone values without consideration for what that curve achieves on press. I use the term "press curve" to mean a curve that is applied to the plate in order to achieve a certain tone response in the press work.
So, someone who first linearizes the plate and then applies another curve to achieve G7 is applying two curves: a plate curve followed by a press curve.
Alternatively, someone who only applies a curve to linearize a plate and then prints from that plate without applying another curve is applying one curve - what I would call a press curve since the linearizing curve is intended to achieve a certain tone response on press. The fact that it happens to linearize the plate is irrelevant.

I don't know if these are the right terms for distinguishing these two types of curves - but they seem to make sense.

disbellj: gordo, This is where I disagree with you. The person has 2 CTPs, and 3 plate curves. This sounds good/right.

GP: I believe he has two plate curves in order to linearize his 2 CtPs (one each). He then has three press curves one each for the different substrates. A total of 5 curves.

disbellj: Let's say that with linear plates, that when linear plates are put on press and G7 applied, the result for prepress is that we have our (press) plate curve. There is now NO NEED to set up the press again. What is left to do? Make sure the plates are linearized without that "press" curve, and then apply the "press" curve to those linear plates, and you have the final values on plate that you need at press.

GP: I think that's correct if both plates have had plate curves applied to linearize them and if you are setting up for one substrate.

disbellj: So we have two platesetters, which might need two different "calibration" (linearization) curves. So if sending to platesetter #1, you send data that has had "platesetter #1 calibration" curve and "press" curve applied. To platesetter #2, you send data that has had "platesetter #2 calibration" curve and "press" curve applied.[/]

GP: Yes.

disbellj: If you didn't have the linearization step, then you would have to keep up with "two sets of numbers", the set of numbers that got you to target on press - for each press.

GP: No. If you did not have the linearization step you would only need to keep up with one set of numbers - the ones that result on plate from the application of the press curve.

disbellj: Instead, you don't even have to keep up with those numbers at all. You just have to make sure the plate is linear when the "press" curve is not applied[SNIP]

GP: With your method you apply a linearizing plate curve which you do not have to measure. You the apply a press curve which you do have to measure. If everything is correct then the numbers after applying the press curve to the linear plate will have certain values. If the values are correct then the plate can go to press.

With my method you apply a press curve to the uncalibrated plate. If everything is correct then the numbers after applying the press curve to the uncalibrated plate will have certain values. If the values are correct then the plate can go to press.

Now, if we didn't get the final numbers on the plate that we expected;

You would need to check that the press curve had been applied, that the linearizing plate curve had been applied and what the uncalibrated status of the plate had been (since your linearizing curve depends on it)

I would only have to check that the press curve had been applied and what the uncalibrated status of the plate had been.

disbellj: Now here where I work, I have 2-4 curves, for one platesetter. One is "press" curve for 150 linescreen uncoated paper jobs. One is "press" curve for 175 linescreen coated paper jobs. Now these two curves aren't changing, as long as the press can print within tolerance and my plates are linear without these "press" curves.

GP: Not unusual number of curves. And if you skipped the linearizing plate curve you would do just as well. (see the previous point)

disbellj: Also, if my 150 or 175 linescreen wasn't within tolerance of being linear, I could make a "calibration" curve for each linescreen, so could possibly have 4 curves being used in my system (up to 2 at a time for a job, depending on paper type), but can have only 2 - one per paper type (like we do, since linearization is not needed).

GP: OK.

disbellj: So really, it comes down to what do you want to "keep track of":
a. At least one set of numbers with the "press" curve applied, to be the master numbers you'd always want to make sure your uncalibrated plate with "press" curve applied measured within tolerances to.
-OR-
b. No numbers to keep up with, just make sure to output a plate with "calibration" curve applied, but not "press" curve applied, and make sure dot on plate is within tolerance to dot in file.

GP: I disagree with that breakdown. In day to day work, both your method and mine result in keeping track of one set of numbers - the final values of the dots on the plate. The difference is what you have to do if you do not get what you expect on plate. You've got more variables and points of failure to check than I do (as described earlier).
[SNIP]

disbellj: If doing a, can prepress tell if problem is in prepress or at press as easily? Not in my opinion. Why? Because they are tied together.

GP: As long as the plates have the correct, expected, dots on them then if there is a problem at press then it is likely a pressroom problem. If the dots on plate are not what is expected then the problem is likely prepress. I've just go fewer point to check to determine where the problem is compared to you.

disbellj: Linearization provides separation from press, and therefore quality control for prepress dept better (made easier) IMHO.

GP: I don't see that at all because it doesn't matter whether you linearize or not - nobody will ever know since you destroy the linearization when you apply a press curve. What matters is that the dots on the plates that you throw over the wall into the press room have the tone values that are required to do the job.

best, gordon p

 

[Lukas Engqvist]

Am I a geek if I'm enjoying this thread?

 

[gordo]

Am I a geek if I'm enjoying this thread?

Me too. I haven't had as much fun since the old days of arguing with color experts who insisted that the proof should always follow the press. I find it very interesting that there can be so much disagreement - even confusion - on such a basic question as linearizing plates. The logic that is used to support the particular chosen workflow is also enlightening.

best gordon p

 

[rich apollo]

I'm very surprised at seejay's responses. I'm surprised that the two devices image so differently. I really didn't realize there could be such a difference introduced by the imaging device.

 

[spectrecom]

We always linearize the plate on every install of CTP we do. We then ask the customer to run the dot gain scales but so many pressmen don't use a densitometer and the owners say they can afford to buy one so what then really is the point to go further??? Sometimes I am amazed, the cost of a good densitometer has got to be less than the wasted time and materials in one month of a pressman guessing...

 

[gordo]

We always linearize the plate on every install of CTP we do.

It might be different for you as a vendor doing an install vs how a printer should operate in a production environment - especially given the densitometerless and clueless customers you mentioned. Linearizing the plate brings it to a specific status and enables you to get install sign-off without having to provide professional services for free.

That being said, do you advise your customers that they should first linearize their devices and then apply any required press curve to deal with dot gain issues? Or do you advise them to run an uncalibrated/unlinearized plate and just build an appropriate press curve? Assuming, of course, that they have a densitometer they can lay their hands on

best, gordon p

 

[sandpiper]

We have 2 CTP devices for newspaper production . We average the measurements and create a single linear curve, followed by a press curve.

 

[J]

We have 2 CTP devices for newspaper production . We average the measurements and create a single linear curve, followed by a press curve.

Why not just simply use a press curve instead? What is the point of linearizing first? Seems completely redundant to me.

J

 

[RickS]

We linearize first then curve. Twice over my CTP time frame we have had the platesetter break down and not be the same calibration with new parts. Naturally the pressroom and scheduling were freaking by the time the platesetter was fixed. When it was up again the calibration was off. Not a lot, but 3-4%. No problem, relinearize and plates were coming out. If for no other reason I stick with this method.

 

[fiatlux]

We linearize first then curve. Twice over my CTP time frame we have had the platesetter break down and not be the same calibration with new parts. Naturally the pressroom and scheduling were freaking by the time the platesetter was fixed. When it was up again the calibration was off. Not a lot, but 3-4%. No problem, relinearize and plates were coming out. If for no other reason I stick with this method.

Right, given the same situation this is what we would do (paraphrasing your post).
The platesetter breaks down and after repairs it's not the same calibration with new parts. Naturally the pressroom and scheduling are freaking by the time the platesetter was fixed. When it is up again the calibration is off. Not a lot, but 3-4%. No problem, apply a curve so that it delivers the same plate tones as it did before. One step solution.

Linearizing makes it a two step solution that you do that does nothing other than increase complexity.

FL