Process Work

[meddington]

...when curves aren't printing right... the easiest place to start diagnosing the problem is to have an EASY to check baseline reference. Spend a quick 5 mins to burn a test strip on a plate, read the dots and prove that the CTP is burning linear, now you have shown the proof that the plates/curves are in line and you can move on to looking at the myriad of assorted press issues that could be causing the problem. to show that 100=100, 75=75, 50=50, 25=25 and all the rest is something anyone can understand.

With due respect, burning a test strip after the fact (if this is indeed what you mean) doesn't really confirm much about previously imaged plates other than they CAN be linearized. It doesn't reveal the the requested curve from the Rip was applied to the plate. Depending on when the plates were generated and how often the plates are linearized, it may not even confirm that the plates in question were imaged in conformance either (if plates were burned last week, but linearization updated 20 minutes ago). The best solution is to have a control element (as in Gordo's suggestion) on every plate that can be verified without going back to CTP. Moreover, as I stated in an earlier post, my experience has shown that CTP imaging/developing problems can have a negative visual effect without, or prior to, effecting the measured result. I wouldn't say this is a rule, but I strongly believe that linearization is not effective at correcting imaging/developing issues.

 

[Alith7]

Just to start, I original CTP was violet lasers which did fluctuate constantly, but I'll get back to that.

That's a lot of curves to manage. Maybe I'm misunderstanding you, but since you don't see the result of the linear curve because it gets clobbered by the press curve, I don't see how you can make sure that the plates are linear when you use a press curve. If you want linear patches to measure then you could simply impose a curved bitmap step wedge in the gripper margin that creates linear patches. That way you have linear patches if you want but still have the simplicity of a one curve workflow.

you do see the result. I'm shortening this for ease of typing, but the Xenith RIP sends over an uncalibrated, "uncurved" 39 patch test strip 0-100 1-BIT tif. The point of it is to verify that if the RIP sends the assorted screen values, that the CTP is burning those values. in the mid ranges, it's not as critical, however when you get into the lower 10% and upper 10% it becomes MUCH more critical that if you send a 97% dot, it's burning a 97% dot, and same for a 3% dot. what artificial curves you apply to increase dot gain, etc., where needed is done BEFORE your 1-BIT tifs are created, the CTP doesn't care about that. but if your RIP is sending 95%, 96%, 97%, 98%, 99% and the CTP is burning them at a solid 100%, that's not going to work so well. If you decide that you need to apply a calibration curve pushing your lower percentages, or opening your higher, that's up to you, and is applied separately.

I think the biggest issue is thinking of it as applying a "curve" like you would for a profile. to linearize the plate isn't really applying a "curve" like that, you're simply calibrating your device, same as you would a densitometer, a spectro or a plate reader.

If you do what you are saying you do then you do not need to apply a press curve on top of a linearizing curve in day to day production. Just use a press curve, and if there are any issues do what you say you do, burn the demonstration plate with a linearizing curve.
As an alternative, you could do what I suggested earlier - have a target in the gripper margin. The difference though is that if you include the linearized patch on every plate then there is no need to burn another plate to prove that that the CtP is burning linear.

A possible problem with showing the press people that the plate is burning linear is that it is not related to the dots they are using to print the job - i.e. the plate is actually not linear it has a curve applied to it. Telling them a fairy tale like that might result in them telling you some pressroom fairy tales if they take the attitude that your general prepress manager/operator doesn't understand or give a flying flip about the intricacies of the press curves.

A linearize patch on the plate would have the press curve applied to it, so would defeat the purpose.
It's not so much "fairytales" as the irritated owner with the stopped press and held up production, who doesn't understand the process just wants the quick proof that it's NOT the CTP that is the issue. They (owner & pressmen) ARE aware of the process in that the plates are linearized, and THEN a curve is applied to push an artificial dot gain, and adjust for press faults, so that's not a concern.

It sounds redundant because it is. It adds complication to the workflow without benefit. And in your case, as you've explained, you actually don't need the double curves workflow.
I've spent a lot of time in shops where everything is always prepress's fault, and the press operator's can do no wrong. It's not a healthy environment for either the employees or the company itself. IMHO, a more productive approach is education. It does work wonders and is a more sustainable strategy.

In my experience, using curves to fix a plate imaging problem does not work. So I don't see it as a safe guard.

I can only speak from what I've seen in hundreds of shops (and what a few vendors have claimed for their systems on this forum) is that the range of variation is within the "noise." I.e. The inherent variation doesn't require a change in curves. And if it's an imaging problem then curves will not fix it anyway.

This might be a difference in our terminology. In a CtP workflow the important thing is to set laser exposure and processing (or lack thereof) to the manufacturer's specifications so that the result is a robust halftone dot on the plate that maintains its integrity on press. That process is calibrating the CtP/plate imaging system. The result is a calibrated plate - it may not be linear, but it is calibrated.

The tones on plate are there to create a desired tone response in the presswork.
Curving a plate to make it linear will result in a certain tone reproduction on press.
Curving a plate to make it linear and then curving the curve to result in a certain tone reproduction on press makes the linearizing curve redundant. Why not simply curve the plate once to achieve the desired tone reproduction on press?

You do not need a linear plate in order to determine that the press is in spec or not. When you test a press to see if it's in spec you are primarily testing for things like ink transfer since that is the primary function of the press and what the press is designed to do. The tone reproduction in the presswork gives you an insight into how well that ink transfer is occurring irrespective of whether the plate is linear or not.

I'd have to go back and look at my CTP specs, as I don't remember if my AGFA Accento is violet or not, but our original CTP was, and the AGFA has similar problems sometimes, though I do have to admit, not as often.

the problem being this, a laser does not burn as hot/strong as it gets older. this does not always changes it's "sharpness" but it doesn't always burn the dot the same size. it doesn't always show up as banding, or chatter, either. This was especially a problem with the violet laser. Yes, in a perfect world, you call in the tech and have them fix the CTP back to factory spec. In a mid to small shop, that's not always an option, whether money or time restraints.

Maybe, the biggest difference in technique is the difference between a high-volume commercial printer vs. a quick-print shop or small sized commercial. The steps we have in place are for our small sized commercial shop. Most of the time, if it's broke, I have to fix it. I have limited tools, and an even more limited budget. I can see your point, I just don't agree with it. If I didn't have the ability to fix, or had a larger source of funds to call in a pro every time something wasn't quite right, I might be more inclined to agree with you. But as things stand, this is how I see them.

 

[Lukas Engqvist]

The Accento is a thermal, we used it with the azura plates, if you do have Agfa I would look into using their hybrid screens or sublima if your issues are in the 0-3% and 97-100%.
I'm following the conversation and it seems to be polarising into the two schools of linearise and the one curve only. As was said in another thread, when the issue was discussed, and I lift it because it is important to the overall results. The problem with a 2-curve (linear+tvi) as opposed to a 1 curve is the rounding off that can cause banding. Measuring too much data in these cases is the biggest issue, it's fine to do the 39 patch test, but once plotted you must use as few points as possible to define the curve. To make a good press curve does need some common sense. (This is true for any profiling work) a kink in the curve will me much more disturbing than a rounding off of some value. Especially use common sense when it comes to peripheral values (first and last 5%) since banding in soft shadows is very unbecoming.
There is the need for awareness of each tools limitations and precision. Keeping curves simple is key wether you pre-linearise or not.
In my experience creating a stable climate controlled dust free environment for your CTP (and possibly shared with inkjet proofers is worth it in saved down time , CTP life cycle and health of the prepress staff.

 

[gordo]

Just to start, I original CTP was violet lasers which did fluctuate constantly, but I'll get back to that. [SNIP] but if your RIP is sending 95%, 96%, 97%, 98%, 99% and the CTP is burning them at a solid 100%, that's not going to work so well. If you decide that you need to apply a calibration curve pushing your lower percentages, or opening your higher, that's up to you, and is applied separately.

I think the biggest issue is thinking of it as applying a "curve" like you would for a profile. to linearize the plate isn't really applying a "curve" like that, you're simply calibrating your device, same as you would a densitometer, a spectro or a plate reader.

A linearize patch on the plate would have the press curve applied to it, so would defeat the purpose.[SNIP}

I'd have to go back and look at my CTP specs, as I don't remember if my AGFA Accento is violet or not, but our original CTP was, and the AGFA has similar problems sometimes, though I do have to admit, not as often.

the problem being this, a laser does not burn as hot/strong as it gets older. this does not always changes it's "sharpness" but it doesn't always burn the dot the same size. [SNIP]
Maybe, the biggest difference in technique is the difference between a high-volume commercial printer vs. a quick-print shop or small sized commercial. The steps we have in place are for our small sized commercial shop. Most of the time, if it's broke, I have to fix it. I have limited tools, and an even more limited budget. I can see your point, I just don't agree with it. If I didn't have the ability to fix, or had a larger source of funds to call in a pro every time something wasn't quite right, I might be more inclined to agree with you. But as things stand, this is how I see them.

Thanks for your considered answer and providing more detail.

Just a few comments. I tend to agree with the printer's motto of "why fix it if it ain't broke" as appears to be so in your case. However, for the benefit of others who might be reading this thread...

It is dangerous to use the strategies that may have worked with one imaging technology and apply them to a very different technology.

Here's why.

The violet CtP you started with exposes plates photographically. As an analogy, it's like exposing a piece of black and white film with a flashlight (torch). If the exposure is very brief you'll get a small dot of exposure at the center where the light is brightest. If you increase the exposure brightness (laser intensity) the area that is exposed will grow. So exposure (and processing) can affect dot size.

Your Agfa Accento, on the other hand, is a thermal CtP device. And while people say that they "expose" plates in a thermal CtP device - that is actually a misnomer.
Instead of exposing the plate coating, a thermal device uses the laser to "cook" the coating on the plate.
The analogy would be if you cracked open an egg in a pan and then blasted it with a blow torch aimed at the yolk. If the intensity of the blow torch flame was too great then the surface of the yolk would be cooked but the interior would remain raw. If the intensity of the flame was too low then the yolk may never reach the point of cooking. In either case, the white surrounding the yolk would not be cooked.

So, what the technician does during the thermal CtP installation is to set the laser intensity so that it cooks the plate coating to provide a robust image on the plate. Because if the laser intensity is not correct, then the coating is not cooked properly and the result is a coating that will not perform on press as per its specification. Because thermal plate coatings are fairly binary, variations in laser intensity will have little effect on the size of the dot. So, if the dot is not cooked properly then the highlight dots, being small, may not withstand processing and/or the press environment and hence be lost. Shadow dots may not clean out completely and hence would also be lost.

That cooking of the plate setup is called "calibrating" the plate imaging - (akin to your example of calibrating a densitometer). It has very little to do with the size of the halftone dots that will be produced. In that sense, applying a curve to the plate is not calibrating it (since it is already calibrated), instead it is modifying its tone response which, since another curve will be applied to achieve the desired tone response on press, makes that first curve redundant.

The result of calibrating the CtP will typically be a non-linear tone response, which will result in a certain tone response on press which can be changed with the use of a press curve applied to the calibrated but non-linear plate.

If there is a problem with imaging small highlight and shadow dots (as you mentioned) the solution is not to change the "cooking" of the plate coating by changing laser intensity because that might alter the performance of the plate on press. Now, you might try and use a curve to make larger dots i.e. make 1% dots into 2% or 3% dots, but that could cause muddy highlights and bad fade-offs on vignettes.

So, instead, the strategy to recover those lost highlight and shadow tones is to use a hybrid AM (a.k.a. XM) halftone screen like Agfa :Sublima, Kodak Maxtone, or other. With that type of screening AM dots are constrained to the smallest reproducible dot size. To make a lighter tone, i.e. recover the part of the tone scale that's being lost, the screening starts removing those reproducible-sized dots. Because there are fewer dots the area appears lighter and the tones are recovered while preserving the performance capabilities of the plate.

best, gordo

 

[meddington]

I think the biggest issue is thinking of it as applying a "curve" like you would for a profile. *to linearize the plate isn't really applying a "curve" like that, you're simply calibrating your device, same as you would a densitometer, a spectro or a plate reader.

If you're referring to using rip software to "linearize" or "calibrate" the tonal output of a plate, not sure how this wouldn't be applying a curve.

 

[David Milisock]

With due respect, burning a test strip after the fact (if this is indeed what you mean) doesn't really confirm much about previously imaged plates other than they CAN be linearized. It doesn't reveal the the requested curve from the Rip was applied to the plate. Depending on when the plates were generated and how often the plates are linearized, it may not even confirm that the plates in question were imaged in conformance either (if plates were burned last week, but linearization updated 20 minutes ago). The best solution is to have a control element (as in Gordo's suggestion) on every plate that can be verified without going back to CTP. Moreover, as I stated in an earlier post, my experience has shown that CTP imaging/developing problems can have a negative visual effect without, or prior to, effecting the measured result. I wouldn't say this is a rule, but I strongly believe that linearization is not effective at correcting imaging/developing issues.

>burning a test strip after the fact (if this is indeed what you mean) doesn't really confirm much about previously imaged plates
Correct but it can confirm if the plate setter hardware is functioning properly on it's own. Applying the curve is another fact.

 

[Alith7]

ok! I concede! LOL
thank you everyone for being more determined than I to prove you're right!
looks like I have some more digging/research to do and need to take a fresh look at my calibration process.

On a side note, I'm WELL aware that a clean room would be a HUGE help, unfortunately, that's just not possible. Maybe "someday" when we move to the new shop, I'll try to push harder in the building stage for a clean room set-up.