Poll - Do you linearize your plates?

[DCurry]

I'm glad you stated your own opinion regarding this, Gordo, as I know lots of folks (myself included) greatly respect your expertise.

I do think there would be some value in linearizing if a shop has more than one CTP device. By linearizing them each first, they would all be at the same baseline point at which the shop would then apply the press curves and they could output on any CTP for any press easily.

I guess they could make different press curves for each CTP device, but it seems more difficult to manage that way.

 

[Lukas Engqvist]

Can I ask if the AM screening curve is done for default (black angle) only or have you done curves for different angles. (I know vendors strive to get the dot shape right, wich is why there are variations on dot shape) Would be interesting to see how the curve would look comparing: round, rational tangent, Adobe Accurate. Agfa Balanced etc.

If the 2-curve method is used and say a 45° AM raster is used… and the 75° or 15° may have a slightly different curve this may even increase the deviations.
This kind of error would also be introduced using a one curve correction and/or K with CMY as common, though as I understand such a compromise would be only after a balanced decission.
I am asking because we did not find out plate meter to be accurate enough (due to plate contrast properties) for that kind of comparison.

 

[Hopkins Printing]

Yes, we linearize. We start with a linearization curve and then apply our calibration/compensation curve on top.

Cheers,
Jon Morgan
Hopkins Printing

 

[tmiller_iluvprinting]

Yes we linearize. We have two CTP devices made by different manufacturers. In our case linearization is a must to insure consistent dot percentages on plates between devices. If we only had one CTP device I don't believe linearization would be necessary because of the consistency of the devices.

Regards,
Todd

 

[John W]

Because we are folding carton and the client expectation is that all reprint cartons shall match the previous run standards, we only linearize. We have press curves capability but have not found a need to use them as of yet. But we do use different rip curves; eg: legacy, linear, G7, scooped for substrate etc to control the job on a job by job basis. We produce one bit plate image files and keep them live on a large RAID (about 7TB at the moment) to protect the integrity of the original plate image and simply merge onto the darkest color a current docket slug; eg: Dkt 12345 (K).tif + Dkt 54321 (Slug).tif = Dkt 54321 (KMrg).tif on glue flaps. This is also handy for our clients so their fulfillment inventory is as current as possible and assists them from using newer cartons before older cartons (ie: a selling feature).
John W

 

[Prepper]

No linearisation here

No linearisation here

We don't linearize since we went to CTP. Have seen no need to so far. Before with film we linearized and then applied G7 curves. Now we just output the plates as set up by manufacturer, print targets with no lin or curve applied, measure with Curve2 software and apply those curves to plates to print G7 at press. So far so good. One press, one CTP, I think if you had two or more you would need to make them match each other with lin first.

 

[gordo]

Because we are folding carton and the client expectation is that all reprint cartons shall match the previous run standards, we only linearize. We have press curves capability but have not found a need to use them as of yet. But we do use different rip curves; eg: legacy, linear, G7, scooped for substrate etc to control the job on a job by job basis.[SNIP]

This is somewhat confusing to me.
If you are only linearizing the plate. Then linearizing the plate is effectively a press curve since it gives you what you want on press.
Then you write that you use different "RIP" curves to achieve different press results. A RIP curve usually results in a curve that gets applied to the plate. So your RIP curves are plate curves.
So it appears that you always use one curve - to linearize the plate (plate curve) and then sometimes, on a job by job basis, you apply another one on top of that to achieve the results you want on press (a press curve).

Is that correct or a I misunderstanding you.

best, gordon

 

[ffstoney]

We don't have a second output device, but we have partnered with another shop that has the same device as ours. So if my device goes down hard, and we both have a linear curve applied, I can send him my 1bit tiffs which have a press curve applied and know with reasonable confidence that I will get the same results on press. Or at least close enough to get us through a couple of days until our device is repaired. So I guess I am saying that it is a safety issues for us.

 

[gordo]

Yes we linearize. We have two CTP devices made by different manufacturers. In our case linearization is a must to insure consistent dot percentages on plates between devices.

So you have at least two curves. One for each device to make them linear.

If you don't apply any other curves on top of those then your linearization curve is effectively a press curve (i.e. you need a linear plate in order to achieve the required tones on press). Another way of saying that is that you are not linearizing the plates from the two different systems. Instead you are applying a different press curve to each system. It's just coincidental that the result is a linear plate.

If, after you have set your baseline by linearizing the plates from the two systems, you then apply another curve on top of the initial linearizing curves to achieve the presswork you desire. Then you have a third curve (two plate curves to linearize, one press curve to bend that condition to achieve the presswork you want). If that is the case, why not skip the linearizing process/curves and simply have two (press) curves that get applied to the plates so that the presswork from the two systems appears the same? No need for a third curve.

I don't see the need in a multi-CtP shop, even where the devices come from different manufacturers, to have more than one curve (a press curve) per device. Linearizing the plate first seems redundant because I don' see it achieving anything useful. It just adds complexity and the possibility of error.

best, gordon p

PS, My intention is not to promote one method over another, which is why I was reluctant to say anything before. My intent is to better understand why and how printers choose one method over another. Unlike many choices that printers face in production this should be a pretty objective, fundamental, technical issue (it's a "no brainer") and so it shouldn't get clouded with the religious fervor that other questions my prompt (e.g. what's the best workflow?).

 

[seejay]

Hi Gordo,

we have two different ctp devices which give totally different results on the same plate when outputting 'non-linearised'. e.g. one device prints slightly sharp (3% in the 50% dot), the other prints more-or less on the money.

The reason why we linearise is that when we do have to re-profile presses, we would have to run the test sheet on plates output by both devices.
By working from linearised plates, when we run the test form, we can then apply the same press curve to both devices - at least this is how it should work in theory.

 

[vansrv8er]

I would always run with a linearization curve. The simple thought behind is it is easy to track and if a portion of the process fails or weakens, laser, chemistry, plate, and a change in exposure or hardware on your platesetter must be made, it is easy to get back on track of where you were. Just re linearize! Otherwise your entire plate curve must be redone from tip to tail.

 

[gordo]

Hi Gordo, we have two different ctp devices which give totally different results on the same plate when outputting 'non-linearised'. e.g. one device prints slightly sharp (3% in the 50% dot), the other prints more-or less on the money.
The reason why we linearise is that when we do have to re-profile presses, we would have to run the test sheet on plates output by both devices.
By working from linearised plates, when we run the test form, we can then apply the same press curve to both devices - at least this is how it should work in theory.

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]

I would always run with a linearization curve. The simple thought behind is it is easy to track and if a portion of the process fails or weakens, laser, chemistry, plate, and a change in exposure or hardware on your platesetter must be made, it is easy to get back on track of where you were. Just re linearize! Otherwise your entire plate curve must be redone from tip to tail.

If the natural state of your CtP means that when you ask for 50% you get 45% (and so on through the tone scale) then that is easy to track - no need for a linearizing curve. If I ask for 50% and don't get 45% then I know something has changed.

I doubt very much that relinearizing a plate with a curve is the correct way to deal laser power, exposure, chemistry, or processing issues. Maybe the vendors or field service engineers could chime in on that idea.

best, Gordon p

 

[Lukas Engqvist]

I will attempt to throw in a question and an analogy.

When we say "press curve" we actually have two (or more) per press one coated and one uncoated, since the dotgain or TVI is the target?
I know that some would even here have a "press curve" and a "calibration curve" with enough decimals I suppose this could work, but that would then be a 3 curve system.

Let me brave an analogy.
However good a translator may be it there is information lost in each stage. If I would need to communicate in russian and must first translate my phrases to latin (latin being an old standard interchange anguage for schollars) I am sure things would be lost (actually did that "game" with google translate taking a text and translating it in multiple stages till it was undistinguishable).
If I go back to why I decided on the one curve method it was my experience with photoshop and seeing how destructive adjusting a picture by applying multiple adjustment curves was.

 

[Kristian]

We linearise the plate and then apply a compensation curve.

Our CTP system is very stable and has never varied from 50% on file coming out at 45% on a plate with no curve applied. We chose to linearise first so that if we change our CTP system or plate type we just have to generate new linearisation curves and the press curves remain as they are.

Seemed a simpler approach to me?

Kristian

 

[tmiller_iluvprinting]

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

Those of us using the G7 methodology would beg to differ with you Gordo. If you were to create a plate curve for each non-linear CTP device you would double your G7 run presstime(Our VP of Manufacturing's nose is beginning to bleed!) You would need to do calibration runs for each non-linear plate, and then verification runs for each. NOT very economically sound, especially in this day and age. It is IMHO, a much better procedure to linearize, and do one calibration run, create one curve and verify once. BTW- we have been using CTP for 9 years and I have yet to see the CTP devices drift.

Regards,
Todd

 

[Lukas Engqvist]

@todd I am not understanding why your nose is bleeding? Are you saying that a press is so stable that within the G7 methodology you consider that running curves is a one time thing? Am I hearing that running linear plates is all that is needed? As I read the G7 manual a TVI curve is important. And then the complexity is when you have X amount of output devices on Y amouts of presses. In many to many flows it makes sense to use two curves. In a smaller shop where there is 1-2 CTPs and 1-2 presses then the extra linear plates is not necessary. Would have been interesting to see if this is the real case. And see if there is a correlation between size of plant and methodology.

I only work at a smaller plant, but for us the variations are at the press. Normalising the dot gain is a continuously monitored process with extra stripps added at press sheets on normal production runs. If there is some continued drifts in dotgain then the curves are adjusted (after making sure it is not a one time occurence due to any of the factors that could affect the dotgain) and the adjusted state is confirmed in the monitoring of the dotgain patches.

 

[gordo]

Those of us using the G7 methodology would beg to differ with you Gordo. [SNIP] It is IMHO, a much better procedure to linearize, and do one calibration run, create one curve and verify once. BTW- we have been using CTP for 9 years and I have yet to see the CTP devices drift.

Regards, Todd

Some quotes on the topic from the G7 guides: (now we are getting into religion after all 'cause I'm quoting from the "scriptures" LOL)

6.2 Origin of NPDC curves
To determine the 'natural' NPDC curves of commercial CtP-based printing, G7 research analyzed
numerous press runs made with ISO-standard ink and paper, and a variety of plate types imaged on
“un-calibrated” CtP systems (no RIP curves applied, not even to “linearize” the plate).

5.4 Set up the RIP
Set up the plate making RIP exactly as you would for a normal job, but clear out any values in the
current calibration table, or begin with a new, empty table. The first press run is best made with ‘un-
calibrated’ plates – i.e. no calibration values in the RIP.
IMPORTANT: Do NOT linearize the plate-setter so that measured dot values on plate exactly match
original file percentages. Contrary to common belief, this may reduce accuracy of subsequent steps.

a. PRINTING IDEALIZED TARGETS VLAUES - Achieving calibration condition with raw or linear plates, not requiring a curve, is an ideal situation.

Hmmmmmm.

Also because you have been using CTP for 9 years and are have yet to see the CTP devices drift is another reason why you don't have to linearize to bring it to a standard condition.

best, gordon p

 

[tmiller_iluvprinting]

My point is that because of the 5% difference between CTP devices our decision to linearize is based on economic reasoning. We eat the bill for calibration and verification pressruns, to double the bill is a bad
business decision. Linearization in this case serves its purpose and has little or no affect on the final outcome,
except to make 2 different CTP devices image the same size dot. Much easier on the 'bottomline' than 2 different pressruns, again IMHO.

Regards,
Todd

 

[gordo]

My point is that because of the 5% difference between CTP devices our decision to linearize is based on economic reasoning. We eat the bill for calibration and verification pressruns, to double the bill is a bad business decision. Linearization in this case serves its purpose and has little or no affect on the final outcome, except to make 2 different CTP devices image the same size dot. Much easier on the 'bottomline' than 2 different pressruns, again IMHO.

Regards, Todd

I think that your economic justification is a good approach to the question.

In my experience, for most sheetfed shops, the internal actual cost of doing test press runs is very, very low - a few hundred dollars in ink, paper, and plates. In the shop that I worked at it was low enough that we could cost justify doing press proofs for certain projects that involved tricky special color effects with pantone colors. So I don't believe that the savings would be significant. Maybe in a web environment the savings would be worthwhile. There is also the possibility of double burning the plate - i.e. image half the plate on one device and the other half on the other CtP so that one pressrun, instead of two, would provide the data needed to build curves for both CtP devices.

So, put another way, for the few shops that meet some specific criteria, they need to evaluate the implementation cost differences vs risks (e.g. prepress errors and shadestepping) for a linearize first then apply a press curve workflow vs a press curve on uncalibrated plates.

Thanks Todd!

best, Gordon p