Plate Calibration - Harlequin RIP - Edit From Calibrated/Uncalibrated

[andy_baz]

Been trying to fine tune my plate calibration for our press. But when i used "Edit from Calibrated" and enter values taken from the final printed job...it completely throws out the calibration.

So far...

1.) I have calibrated the new batch of plates i need to use to get my plate setting. (this works as expected).

2.) Used FOGRA values for intended press (no problems here)

3.) Then Created the the first Uncalibrated plate (from RIP Calibration Manager) and run it on the press to get the values for and entered into 'Edit From Uncalibrated' setting.

4.) Created the Calibrated Plate (from RIP Calibration Manager) run plates on press and got close to what i expected.

5.) Heres where the issue is after running a few jobs i noticed that the calibration was still a bit too far out. So i then measured our colour bar tint values (20% 50% 70% etc we have these tints on every production job) i used these values to enter values into the Edit From Calibrated area in the rip.

6.) My next set of plates went from being very close to miles out. For example 50% was clocking 74% previously but now after entering 74% into edit fro calibrated my 50% tint now measures around 54%.

Should i have not used the colour bar tints we use on production jobs for values? but instead sent out another Calibrated plate from 'Calibration Manager' and used those values to enter into edit from calibrated?

I was under the impression i could use the colour bar values to keep fine tuning each job until i was happy.

I hope this makes sense.

 

[gordo]

IMHO you should never use the tone patches in the color bar on live jobs to build tone reproduction curves. There are several reasons for this including the fact that the color bar is affected by the live image area, it crosses several ink zones which may be set differently, it is at one location of the press sheet - the tail end, which may not represent inking of the rest of the sheet, and typically the press operator is setting ink according to the results on the live image area rather than the color bar which gets thrown away and so the values in the color bar may deviate from the standard target.
instead, do dedicated documented tone reproduction plates and press runs.

 

[andy_baz]

I understand where you're coming from and completely agree.

However when we do have the need to test plates i put additional tone patches in the centre of the sheet, at the tail at the head, at the head and at the sides (where possible) on the live job, so i'm aware of how the patches deviate on different jobs and different condition etc....and to be honest we arent that colour critical on most jobs.

My real concern was that we are currently as much as 15-20% (at the 50% patch) since using "edit from calibrated" in the rip so i guess im doing something wrong.

My question was more focused on how to use the "edit from calibrated" correctly, and whether the plates i send to my workflow (Xitron) direct from indesign as PDF are classed as Calibrated plates or whether calibrated plates has to be generated from "Calibration Manager" in the RIP.

I can get close enough for most jobs using steps 1-4 on original post, but i want to understand how to do finer adjustments properly rather than just starting from scratch each time.

 

[Ulrich]

Hi Andy

i am not sure that i got it right what you wrote:

Did you earn different results in print with the same curves?
(Than you have to check for equal density first and eventually relocating issues, but you know that...?)

you know that too?:

AFAIR there are four “curves” in the Harlekin RIP:

1. Calibration curve (for linerasation)

2. Tone curve (for adjusting the fine work in the real print to the different aims e.g. coated or unconted Substrates and/or FM or AM screens, I guess that is what you are looking for.)

3. Intended press (the “aim”, e.g FOGRA39, either to edite by hand the values from the specifications about the ideal values e.g. 53% in print for the 40% patch)

4. Actual press curve: this works together with the third (intended press),
here you have to write the real measured values taken from the first real print with a linerized plate.

I myself had only used the first and the second, it is possible to work only with the second (tone curve) too to reach the aim:

Again the example 40% patch: the real print in Cyan is maybe 59%, but the aim is 53%, than you have to edit in the tone curve via the “edit from calibrated curve”-button the 40-patch with 34 or 35 to come down with the print, if the next (test-) print is not satisfying yet in accuracy because you got maybe 55% instead of 53%, edit it again with 32...

No matter with which curves you are
working, to use all four is possible too, but important is that you “tell” them to the actual page setup by choosing them in the page setup dialog!


I do not know the xitron-workflow you are talking about, but sure is that the calibration will happens in the RIP and not before

Ulrich

 

[andy_baz]

Hi Ulrich thanks for your reply,

So let me explain further...

The 4 Curves:

Curve 1: Linearisation (Plate) Curve is ok.

Curve 2: Tone Curve, i have never used this and is set to 'None' on all my page setups.

Curve 3: Intended is FOGRA 39. no problems here.

Curve 4: Actual Press Curve.... Here is where im having issues.

- I use the print values taken from the linearised plate when printed on the press, and enter into Edit from Uncalibrated (actual curve).

- This gets my printed jobs close to the target values ( + 5-8%).

- But i want them closer/more accurate.

- What i then did was enter the values that were +5-8% out of tolerance into the 'Edit from Calibrated' (actual Curve) and expected this to adjust accordingly in my mind this seemed logical.

- My next printed job tint values were way out... -15-20% out of tolerance leaving my job very pale and washed out (almost linear when measured).



Im wondering now if perhaps these finer adjustments should have been made using the tone curve, but i have never used this, so do not know what to expect.


Your thoughts please

 

[Ulrich]

Hi Andy

Im wondering now if perhaps these finer adjustments should have been made using the tone curve, but i have never used this, so do not know what to expect.

my advice: just try it! ;-)

(AFAIR the Harlekin-RIP has no pipette to control the value of the roomed patches on screen before making the plate, but might be you can check your changes with a tool from ESKO, i remember a trial-running version for nothing from Flexo-Perfection, but it´s ten years ago and i can not find it now for sure, but may be it is this in an actual wardrobe: https://www.esko.com/de/downloads/trials/flexo-tools. To work with, you have to create a tiff-device-pagesetup with all the curves you want to check...)


okay, now i can follow you and think i am understanding your issue, i argued something like that

let me say again, that i myself never worked with the third (the ideal aim, intended press, e.g. values from FOGRA39) and the fourth (actual - just linearized printed - curve) in production. I preferred always the use of curve 1 + 2 ...

I agree with your logically thoughts about editing the fourth - actually - curve with "Edit from calibrated target", but if it does not work as expected, what is happened, either i would try a not logical step in using the right button (Edit from uncalibrated target) and writing here the actual print, which is almost 5-8% to high (no matter the first or second after starting the calibration procedure...) or i would go back one step, the plates from the linearized and first measured values which lead to the 5-8% to high result and than additional using the second - the tone curve - for correction in your intended direction: e.g. edit via the button on the left "Calibrated target" with values 5-8 less than you have mesuared. Save the curve with a easy to identifying name and tell it to your page setup, save that too.

Think about: Now you are working with four curves! Not easy to estimate the affect of a editing only one of them, for you and the RIP too ... ;-)


because that, I recommend strongly the way i was gone till two years ago (now i am working with Fuji XMF):


Use just the first (linerazation) and second curve or only the second (tone curve) - (see discussions about "linearization or not" in this forum or elsewhere...):

All you need is a documentation where you can find the increase-values for the output intent, e.g. the increase in FOGRA39 you will find on the right side on page 44 (curve A for CMY and curve B for K) here https://www.bvdm-online.de/fileadmin...Druck_2018.pdf (might be there is an english version to find...)

Back to your situation: You measure 59% at the 40%-patch in the linearized print, the plate which was printed is made only with curve 1 in the page-setup. The increase is 6% to high compared with the ideal aim in the documentation. Next step is, you build a tone-curve via left button "Edit from calibrated target". When you start that, 40% is 40 yet. You want 6 % less, so write 34% in the 40-patch. Do the same for all patches for all colors if neccesary (it will be for accurate work..). Do not forget to tell the saved curve to the page-setup (now, only working with curve 1 and curve 2).

(Yes, it is easy to load a file for the intended press, but it is to handle by mind also with having a look to a documantation about the aimed increase in the special intentions... ;-) )

You won´t hit it all exactly at once by doing only one editing, but after editing the tone curve again (via left button "Edit from calibrated target") once more a second run, a print from the linearized and now also via tone curve calibrated plate should be much more nearer to the aim as you are going actually with curve 1, 3 + 4 what leads you to about 15% less...

Trust in me, it works fine with only first and second curve, we had won a competition by the FOGRA on this way, but that is ten years ago... ;-)


Good look and succsess! (let us know...)

best

Ulrich

 

[gordo]

I’ve got to say, if I’m reading this thread correctly, what an overly complex and obtuse method to deal with such a simple issue. I find it hard to believe that any vendor would have actually developed and foisted it on their customers.

 

[Ulrich]

Gordo,

I know that one of my personally faults is to need always too many word’s for make an easy thing too complex... ;-)

But in this case, it won‘t help Andy to say: Just enter the right numbers into the right patches. Of course he can order a service guy by his vendor, too, but then he has to pay him every time, when he changes one or more parameters for something he can do by himself.

The Harlekin-RIP that i remember offers in fact four different curves for the calibration process to the user. Each with two buttons (Edit from calibrated target or uncalibrated), sometimes you have to enter just measured values and sometimes you have to edit them again. Here are two directions for manipulate a calibrated curve. A lot of opportunities at all, anything else than trivial stuff ;-)

The manual is not really a help and it is a pity that you have to expose a plate and may be even print it when you working with processless plates in absence of an expensive measurement-tool for them to control your steps, because there is no Pipette in the software as a tool for screen.

Your invention disappoints me: Not because you made one, but because I thought that I tried to simplify it with the recommendation for working only with one curve (the tone curve).
That was my intention.

Something went obviously wrong when he gets now 15-20% less by using the curves 1, 3 + 4. But what exactly?
May be he has not to enter the values from the just linerazed print via the right button (uncalibrated target), but via the left button (calibrated target)? (In the meaning of “already calibrated by curve 1, the linearization”). I mean 5-8%
is not really less as result for a linerized and a first automatically generated curve with the already entered values from the first print...

No one else here working with curves 1, 3 and 4 in the Harlekin who will enlighten us with his shared experience?

If not, I repeat my recommendation for working with curve 2!

Best

Ulrich

 

[gordo]

Gordo,

I know that one of my personally faults is to need always too many word’s for make an easy thing too complex... ;-)

That was not what I meant.

I don't have experience with the Harlequin RIP - however, in general, from a strictly technical point of view, there is only a need for a Target curve - i.e. the tone reproduction curve that you which to achieve on press - and a current Press curve - i.e. the tone reproduction curve that you are getting on press. The workflow can use those two curves to generate a tone reproduction compensation curve to be applied to the files so that you get the tone reproduction on press that you desire.

There is no need to linearize the plates.
There is no need to measure the plates.

It doesn't appear that the Harlequin RIP works that way. Instead it appears to be a very complicated process and one where (as you say) the manual is not really a help.

 

[andy_baz]

Here is a screenshot of the RIP interface...thought it might help to visualise the process!

 

[andy_baz]

The process does seem more complex that it needs to be i agree. It may be that i have made sound more complex than it needs to be!.

We have been using this workflow for around 12 years and it has always produced reasonable results.

For those 12 years when needing to re-calibrate, usually for a new batch of plates or a new press etc, i have always created Curve 1 (Linearize), then created Curve 4 (Actual Press), which has usually achieved pretty good results.

Also in those 12 years i have dabbled with how to make the finer adjustment, i.e. edit from calibrated etc and never had great success and just reverted to remaking the Curve 1 and Curve 4, a longer process but a process that worked.

I always read online about tone curves for adjustments or how people have used edit from calibrated to achieve the finer adjustments....im willing to give alternative methods a try such as using the tone curves, but need to understand more before i start editing the workflow.

 

[Ulrich]

Here is a screenshot of the RIP interface...thought it might help to visualise the process!

Good idea to post a screenshot: I notice by viewing it that my memory fools me!:
I have mistaken left and right concerning the Buttons "Edit from calibrated target" and "Edit from uncalibrated target" in my postings above...

I understand very good that you are not willing at once to throw overboard your usually handling with the Harlequin-RIP, just because someone in a forum told you to do it...

...but on the other hand: You stated that you haven´t achieved a satisfying fine tuning till today.




So again 1.:

it might be really helpfull to have a look to the ESKO hompage (the flexo perfection tool), it is an investion about 0,5 - 1 hour of your time:
If that is the tool i remember - in doubt ask their support by phone or mail - you are able with to check the effects of a modified Page Setup whithout exposuring a plate.
Just create a new Input with Device Tiff (if possible with also 2540 dpi - the same resolution as your platesetter...) and configure that with the Page Setup you have created for your platesetter you want to check. Notice or define it where the Folder the new started (Tiff-) Input-Channel is saving or should save the ripped tiffs (I mean you need "configure Device" for doing this?). Open the One-Bit ripped Tiff with the tool and you can measure the with all defined curves ripped 5- or 10- stepping patches from your testchart with a pipette (a tool the Harlequin doesn´t offer!), so you will know that your manipulation effects go in the right direction and dimension before exposuring or printing a plate.



So again 2.:

Even though of working myself with linearized plates till today, meanwhile i totally agree with Gordo that this way in the very most (if not all) cases is not necesary anymore. Shortly: This method is a relict from CtF (Computer to Film, a gone time in which the User Interface of the Harlequin RIP was developed and nearby said not changed till today-looking...) and if a changed parameter like plate-fabrication or a new press forces you to recalibrate your process you can adapt the changed prozess by modifying only one curve: CtP (Computer to Plate) is much more stable - above all with processless plates - than film which works with more sensible chemistry...

On the other hand it does not really hurt either to work with a linearized plate (except some time and power the RIP need more for respecting that curve in the process of generating the One-Bit-Tiffs...). So if you are used to work with and feel save with and your mainboards have fun and energy is not in focus: keep it, but it is not necessary...



Rethinking and reading your issue again

1. print: with linearized plate (only curve 1) Here i want to ask: The 50-patch shows in print which value?

2. print: with curve 1 and curve 3 and curve 4(#1), last curve edited by typing the measured values from the 1. print via left button "Edit from uncalibrated target = +5-8% (little bit to high for a standardized production)

3. print: with curve 1 and curve 3 and curve 4(#2), last curve edited by typing the measured values from the 2. print via right button "Edit from calibrated target = -15-20 (much to less to work with!)



I see this options for you to go on, in this sequence:

1. option:

Give a try to the tone-curve (curve 2) method. Because you have printed the linearized plate in this case you must keep the curve 1 in the page Setup, but deactivate curve 3+4. Next step: Edit a new tone curve via left button "Edit from uncalibrated target", by opening this dialogue you will read 50 in the 50-patch.

Your 1. print with only the linearized plate is measured there with in example 68% in Cyan (increase 18). You know the target 64% (increase 14) from a table in a documentation about FOGRA39. You are +4 (to much), so edit the 50-patch with 46 for coming down about 4. As i said before, this will probably not hit exactly in the next print at all patches, but you should earn not even a worse 2. print as with your method which leads you +5-8 away from the respective patch. Save this curve 2#1

May be with measured 61% as a possible example you are still "out" of a standard in the next, a "new" 2. print (with curve 1 and curve 2#1, edited with 46 in the 50 patch). So open again the existing tone curve 2#1 as a copy with the same button (left Button "Edit from uncalibrated target"), when you find again the former edited 46 there in the 50 patch everything is going right: Now start the fine tuning by sensible correcting the affected patch with maybe 47,3 instead the former edited 46 there for getting again a little bit more than 61, because the target is 64...

You will probably need a (now totally counted) third print (but the linearized plate you have already is included in this counting) i guess, but not a fourth, now you have expirienced the dimension in the effect of your manipulation and should there be left after measuring the 3. print a very few single patches to optimize, you can do it "blind" without another final print.

With this method i myself had worked about 14 years succsesfully and satisfying for AM coated and uncoated and FM, a spread very close or less 2% between cmy and only 1 or 2% away from the respective target is possible with!

Remember the opportunity with the Esko tool for checking a ripped file before exposuring to go for sure that all involved curves, page-setups and input channels partizipate from your actual modifications ...



2. option:

Try the above (in 1. option) described method via left Button "Edit from uncalibrated target" with curve 4 in that way, that you did not enter there - after the 2. print and before the 3. print - the measured values, but a "corrected" version of them.
E.g.: your 2. print leads you to printed +8 (too much) in the 50-patch, for example you edit it with 8 less than you read there by opening the dialogue (there is to read the value from the just linearized plate, right?).

Why?

The 2. print (curve 1, 3 + 4#1) leads you still to +5-8 (too much), if you now manipulate the values taken from this second print a little bit (around 5-8 less as measured) and make the RIP thinking these (5-8 less than mesuared) are the values you have got from the first print (only with curve 1), it should not generate a curve which leads you to a print with 5-8 more as targeted and hoped...

By the way, now i myself working with Fuji XMF it is the same thing at that point, that in the respective dialogue i have to type less than mesuared, when i will come down with the increase...



3.other options:

The pity with the 2. option is, that it is not documented or hardly to find in which way the curve 3 + 4 are working together. Forces higher values as targeted in curve 4 (actual) a weaker or a stronger compensation-curve together with curve 3 (Intended)

Of course it is possible to do more experiments, for example to manipulate the taken values from print 2 as suggested in the 2. option for editing via the right button "Edit from calibrated target" in curve 4.

I suspect the User Interface of the Harlequin RIP in that way, that the two buttons (Calibrated and Uncalibrated) only to find as an opportunity for editing each of all four curves, because the programming Creator was to lazy to kill one of them by entering the dialogue for a chosen curve, where a "Editing from calibrated target" does not make sense, for example the tone-curve. But may be I´m wrong and just too stupid for this stuff...



At least:

I am wondering about less resonance here, because i thought, the Harlequin is very much in use in the world.

So it is up to you Andy, when you go on and find the right editing and use of curve 4 for fine tuning, you can earn a lot of glory and honour and can help all the poor pigs out in the world they have to work with the Harlequin and will come to this planet looking for help...

;-)

Best

Ulrich

 

[Ulrich]

Here is a screenshot of the RIP interface...thought it might help to visualise the process!

one more thing:

in your screenshot is to read Device: DI-file and Color: Monochrome

Don´t you edit different curves for CMYK?

 

[gordo]

On the other hand it does not really hurt either to work with a linearized plate (except some time and power the RIP need more for respecting that curve in the process of generating the One-Bit-Tiffs...)

That is not actually true. If you use a curve to linearize the plate and then apply a second curve to achieve your desired tone reproduction on press then you run the risk of introducing shadestepping (banding). Also, you've introduced another process variable - i.e. potential point of failure - with no benefit. That being said there may be specific cases where you might want/need to linearize the plates first - but that would be rare.

Does the "manual" include definitions (a glossary) for the terms it uses (e.g. "calibrated target", monochrome, etc) when building tone reproduction curves? If so I'd like to see it (couldn't find anything on the web).

 

[andy_baz]

one more thing:

in your screenshot is to read Device: DI-file and Color: Monochrome

Don´t you edit different curves for CMYK?

The set of calibrations shown in the screenshot are for the plate calibration only...hence monochrome.

 

[andy_baz]

So i am thinking the best (and simplest) approach would be to...

1.) Run a set of plates with no settings in any of the curves.

2.) Measure the values of those plates.

3.) Adjust accordingly using a single tone curve.

4.) Repeat step 3 if required further.

It completely makes sense to just adjust at the end of process...so much more efficient.

Also thank you Ulrich for recommending ESKO i shall take a look...but in the mean time i have managed to configure a page setup to output TIFF's that i can view in Photoshop, this alone is very useful in seeing how any adjustments affect the output from the rip before it is plated.

 

[gordo]

So i am thinking the best (and simplest) approach would be to...

1.) Run a set of plates with no settings in any of the curves.

2.) Measure the values of those plates.

3.) Adjust accordingly using a single tone curve.

4.) Repeat step 3 if required further.

It completely makes sense to just adjust at the end of process...so much more efficient.

Also thank you Ulrich for recommending ESKO i shall take a look...but in the mean time i have managed to configure a page setup to output TIFF's that i can view in Photoshop, this alone is very useful in seeing how any adjustments affect the output from the rip before it is plated.

1.) Run a set of plates with no settings in any of the curves. No curves.

2.) Don't measure the values of those plates. The only need for measuring plates is to ensure that the plates image consistently - it is not for creating press tone curves. If your plates are not imaged consistently then you cannot build curves since you will be trying to hit a moving target,

3.) Don't adjust accordingly using a single tone curve. Don't bother with the tones on the plates - you are selling presswork not plates. What happens tone-wise on the press is what's important not what happens tone-wise on the plate.

4.) Don't repeat step 3 if required further. If you do it the right way it is unlikely that you will need to make any adjustment to your tone curve.

What you do - if your plating is consistent - is to:

1 have a target - a desired tone reproduction in your presswork
2 output your tone patches to your plate. Do not apply any curves. It is preferable not to include images as that may confuse the press operator.
3 run on press at your target densities/Lab values or whatever
4 measure the tones in your presswork.

You should be able to enter the tone values you got in your presswork into your RIP. The RIP then compares what you got to what you wanted and using that data builds a compensating tone curve that remaps the tones in your source files so that they will deliver your target tones in your press.

5 do a confirmation press run with the compensation curve calculated by the RIP
6 Celebrate success

NB you can see a more detailed version of the process here: https://the-print-guide.blogspot.com/2010/02/principle-of-dot-gain-compensation.html

 

[Ulrich]

That is not actually true. If you use a curve to linearize the plate and then apply a second curve to achieve your desired tone reproduction on press then you run the risk of introducing shadestepping (banding). Also, you've introduced another process variable - i.e. potential point of failure - with no benefit.

Thank you again Gordo for this lesson. I really appreciate it as such. Banding is a phenomenal issue I never noticed in practice since the end of the 90ies with gradients out of quark xpress-files and I don’t think that Andy has problems with that although he works with linearized plates till now...

i remember the risk of banding as a strong argument for the FM-profiles when they had been published about ten years ago: “when the files already are converted into the FM-profiled separations in which the remarkable higher increase that come with FM is already respected, there is no longer a need for such strong compensation curves in the RIP you will need as by working with FOGRA39-separated files you want to print with FM...”

So, does it really matter a difference when you take two curves or only one for such a strong compensation curve at least? I do not understand the difference, but that’s not the point here... ;-)

Does the "manual" include definitions (a glossary) for the terms it uses (e.g. "calibrated target", monochrome, etc) when building tone reproduction curves? If so I'd like to see it (couldn't find anything on the web).

i find only this in the www (now at home typing in a smartphone...), may be the search in Acrobat will do the job of a glossary, hope the link works:

http://xitron.com/downloads/support...tor_documentation/Navigator 8 User Manual.pdf


Ulrich

 

[Ulrich]

Now we have really fun together:

You should be able to enter the tone values you got in your presswork into your RIP. The RIP then compares what you got to what you wanted and using that data builds a compensating tone curve that remaps the tones in your source files so that they will deliver your target tones in your press.

5 do a confirmation press run with the compensation curve calculated by the RIP

that is exactly what Andy did already with curve 3 (Target/Intended press) and curve 4 (press curve via left button “Edit from uncalibrated target”) and what leads him to a result 5-8% above the target.
The only difference to your recommendation is that he uses a linearization (curve 1), too, which really does not matter in this case. It is for sure not the effect of the linearization curve what leads him to a higher result as aimed in the harlequin dialogue ...

6 Celebrate success

That is what he want!

Ulrich

 

[andy_baz]

So i am thinking the best (and simplest) approach would be to...

1.) Run a set of plates with no settings in any of the curves.

2.) Measure the values of those plates.

3.) Adjust accordingly using a single tone curve.

4.) Repeat step 3 if required further.

It completely makes sense to just adjust at the end of process...so much more efficient.

Also thank you Ulrich for recommending ESKO i shall take a look...but in the mean time i have managed to configure a page setup to output TIFF's that i can view in Photoshop, this alone is very useful in seeing how any adjustments affect the output from the rip before it is plated.

Apologies I misled you in this message...

when i said measure those plates I meant measure those values of that plate when printed.

i beleive we are now talking the same method