Poll - Do you linearize your plates?

[gordo]

Do you linearize your plates before applying a press curve (a two curve workflow - e.g. one to linearize followed by one to compensate for dot gain) or do you only apply a press curve to the uncalibrated plate (a one curve workflow - e.g. one to compensate for dot gain)?

 

[seejay]

Hi gordo,

we linearize our plates and check daily for any drift (albeit that since we started monitoring this 2-3 years ago, we have not seen any drift!). Part of the reason for this is that we also run 2 different plate output devices, from two different manufacturers (h-berg & screen) which have totally different imaging methods. the only way we can guarantee any level of consistency in what we supply to the pressroom is by working with a linearised plate. the main issue we have is that the quality and sharpness of the dot from both devices are quite a ways apart - one device producing a very sharp dot, the other device giving us a more 'furry' dot.

Why would you not linearize? Like any calibration, you have to have a known starting point! without linearization then the press curve will not be a stable measurement.

 

[Lukas Engqvist]

We measure the plates, just to be able to store the values, but do not linearise…*philosoppy being keeping the maths to a minimum, and don't want to introduce bad readings due to measuring device inaccuracy.

@seejay
I would compare linearising the plates shooting a dart at a target, and then introducing a hoope that you need to pass through. The hoope itself can make hitting the target trickier. However, if you do have a chemically unstable process and many variables in the process of making the plate. It would make good sense.

Please don't my opinion dissuade you from answering the poll.

Also I maybe should add we have prepress, platemaking and press under the same roof, and have a tight feedback loop. We see the CTP and press as two halves of the same flow.

 

[DCurry]

At my old shop we used to linearize first, then apply press curves. It did seem to overcomplicate things. When we switched plate vendors and had a team come in to help make the transition, one of the plate specialists thought we were crazy to be doing it that way. In his words, "you shouldn't curve a curve." That statement made a lot of sense to me, so we decided not to linearize and to simply make our press curves from the natural plate state.

If the natural plate state is not stable, then it would make sense to linearize it first so you are at a known starting point before applying press curves. However, we found we were very stable to begin with, so the linearization seemed unnecessary - the uncurved plate was the known value.

 

[gordo]

At my old shop we used to linearize first, then apply press curves.

Was there any discussion or reasoning done behind the decision to linearize first then apply press curves or was it done because it seemed like the right thing to do?

gordon p

 

[Green Printer]

We have not linearized in 12 years. Adjust exposure for holding plates smallest feature. Next make your curve. Simple and effective. When you run into a batch of plates that require an exposure change the curve will allows fall into place. All you have adjusted is the starting point by changing the exposure.

 

[DCurry]

Was there any discussion or reasoning done behind the decision to linearize first then apply press curves or was it done because it seemed like the right thing to do?

gordon p

Actually, I think it was the way the original techs set it up when we first went to CTP.

 

[vaithi47]

the linearising of the plate is to bring the exposure and development to standard levels.as the exposure ie the laser power and development ie the strength,temp,dwelltime of plate and the coating of the plate are variables.linearisation is done biweekly at our place to ensure the exp/dev is under control.

 

[meddington]

No linearization here and consistency is such that I have not seen a need.

 

[J]

the linearising of the plate is to bring the exposure and development to standard levels.as the exposure ie the laser power and development ie the strength,temp,dwelltime of plate and the coating of the plate are variables.linearisation is done biweekly at our place to ensure the exp/dev is under control.

Linearizing does not bring exposure and development to standard levels. Those are accomplished with laser power adjustments and chemistry strength and processing time/temperature etc. You can linearize a plate with a curve but that does not mean that the dots will hold ink or the background be clear or that exposure development is under control.

J

 

[DCurry]

The important thing is to always be starting from a known point before applying the press curves. It doesn't matter if you start from an unlinearized plate or one that has been linearized, as long as it is always the same starting point.

If there are issues with fluctuating chemistry, then it might make sense to linearize to compensate rather than dump chemistry all the time. I currently use AzuraTS plates, so there is no chemistry variable and so no linearization required.

 

[rich apollo]

Nein, nein, nein! No linearization here. I had issues with linearization introducing banding with non-periodic screening. As was said, I like to keep the math simple. I'm very conservative with the number of control points I'll put into the adjustment curve, too.

Less is more.

 

[gordo]

Since this is a very fundamental, technical, prepress workflow question I would have thought that the results would have been strongly skewed to favor one of the two options rather than evenly split (so far).

It would be great if some of the vendors (Steve M, Kevin C, etc.) or their technical specialists (I know you watch this forum) would also cast a vote for which approach they suggest.

I find that the comments in the posts so far have been excellent and provide much food for thought.

best, gordon p

 

[Prepress_Parasite]

Great poll Gordo. We have since the film days carried over to CTP used a curve to be linear then another on top of it for adjustment. I have always wondered what problems this might cause in a CTP workflow.

 

[Deleted member 16349]

Since this is a very fundamental, technical, prepress workflow question I would have thought that the results would have been strongly skewed to favor one of the two options rather than evenly split (so far).

It would be great if some of the vendors (Steve M, Kevin C, etc.) or their technical specialists (I know you watch this forum) would also cast a vote for which approach they suggest.

I find that the comments in the posts so far have been excellent and provide much food for thought.

best, gordon p

I have no experience in this area but in general it seems to me that starting from a linear condition would be better. Practically speaking for this situation it probably does not matter since many seem to do just fine without linearization.

My view is that there is only one linear curve but an infinite number of nonlinear curves. I hope I am understanding this discussion. I am assuming that by linear one means that if the plate imaging system is told to make a 50% dot for example, the result is a 50% dot on the plate.

 

[gordo]

I have no experience in this area but in general it seems to me that starting from a linear condition would be better. Practically speaking for this situation it probably does not matter since many seem to do just fine without linearization.

My view is that there is only one linear curve but an infinite number of nonlinear curves. I hope I am understanding this discussion. I am assuming that by linear one means that if the plate imaging system is told to make a 50% dot for example, the result is a 50% dot on the plate.

I'll try and clarify.

Back in the day, in a film workflow, the process was to linearize the film output. I.e. Ask for 50% in the file and then measure 50% in the film (and supposedly the same correspondence would occur through the tone scale. Nobody measured what the dot size on the plate was. Linear film was the accepted file interchange format. Very few prepress shops would put a curve in the film to compensate for differences in dot gain on press.

Along comes CtP.

You now have the option to apply curves to the plates to compensate for differences in dot gain on press due to a variety of issues - paper choice, screening choice, press condition, etc. I.e. By not sending a linear plate to the press room you can "normalize" or "optimize" the plate for the given press condition.

So, you can have a "plate curve" and a "press curve."

The plate curve can be its natural (uncalibrated state). For example. With "X" CtP device and plate combination 50% in the file becomes 45% on the plate and 25% in the file becomes 32% on the plate. The plate has an inherent curve that is not linear.

A press curve could be used in order, for example, to make my 20 micron FM (which has a high dot gain) tonally match my 133 lpi AM which has a low dot gain. In that case I might want the requested 50% in the file to create a 35% dot on the plate in order to achieve a final tone of 68% in the presswork. I.e. both my FM and AM produce the same final tone on press from a given requested tone value.

So, the options are to:

1) First apply a linear "plate curve" so that requested tones in the file generate the same tone value on the plate. Then apply a "press curve" to create the needed tones on the plate to achieve the tone response I want on the press sheet. Note that if I need a linear plate in order to achieve my on press tone goals then the linearizing plate curve is also the press curve.

or

2) Accept the natural non-linear plate response and make that the base line. I would then only apply a press curve to the nonlinear (uncalibrated via plate curves) plate to achieve my on press tone objectives. The requirement is that the non-linear plate response is consistent. I.e. whenever I ask for a 50% tone I get a 45% on the plate.

best, gordon p

 

[Deleted member 16349]

I'll try and clarify.

Back in the day, in a film workflow, the process was to linearize the film output. I.e. Ask for 50% in the file and then measure 50% in the film (and supposedly the same correspondence would occur through the tone scale. Nobody measured what the dot size on the plate was. Linear film was the accepted file interchange format. Very few prepress shops would put a curve in the film to compensate for differences in dot gain on press.

Along comes CtP.

You now have the option to apply curves to the plates to compensate for differences in dot gain on press due to a variety of issues - paper choice, screening choice, press condition, etc. I.e. By not sending a linear plate to the press room you can "normalize" or "optimize" the plate for the given press condition.

So, you can have a "plate curve" and a "press curve."

The plate curve can be its natural (uncalibrated state). For example. With "X" CtP device and plate combination 50% in the file becomes 45% on the plate and 25% in the file becomes 32% on the plate. The plate has an inherent curve that is not linear.

A press curve could be used in order, for example, to make my 20 micron FM (which has a high dot gain) tonally match my 133 lpi AM which has a low dot gain. In that case I might want the requested 50% in the file to create a 35% dot on the plate in order to achieve a final tone of 68% in the presswork. I.e. both my FM and AM produce the same final tone on press from a given requested tone value.

So, the options are to:

1) First apply a linear "plate curve" so that requested tones in the file generate the same tone value on the plate. Then apply a "press curve" to create the needed tones on the plate to achieve the tone response I want on the press sheet. Note that if I need a linear plate in order to achieve my on press tone goals then the linearizing plate curve is also the press curve.

or

2) Accept the natural non-linear plate response and make that the base line. I would then only apply a press curve to the nonlinear (uncalibrated via plate curves) plate to achieve my on press tone objectives. The requirement is that the non-linear plate response is consistent. I.e. whenever I ask for a 50% tone I get a 45% on the plate.

best, gordon p

Thanks Gordon, I think I follow the steps.

In option 2, if one put a control strip on the plate to be used to determine that the imaging of the plate is correct and this control strip was made of gray scale screens such as 5%, 10%, 20%.... 90%, 100%, would one expect that they measure as the non linear values of the plate curve?

 

[gordo]

Thanks Gordon, I think I follow the steps.

In option 2, if one put a control strip on the plate to be used to determine that the imaging of the plate is correct and this control strip was made of gray scale screens such as 5%, 10%, 20%.... 90%, 100%, would one expect that they measure as the non linear values of the plate curve?

A picture is worth...

This is the plate curve of one vendor's uncalibrated thermal plate (other plate CtP combos will have their own unique curve).

The bottom numbers are the requested tone values in the file - 5%, 10%, 20%.... 90%, 100%. The "0" line represents linearity. I.e. if the plate was linear then that 0 line would straight and be the "plate curve".

Once the engineer has set laser exposure and processing (or lack thereof) to the manufacturer's specs the result is a plate that will carry ink properly but typically does not have a linear response (i.e. a straight line). The thick line that dips below the 0 line is the natural uncalibrated plate curve after the engineer has done their work. In this case it is a positive thermal plate and the uncalibrated plate curve results in a negative value through the tones. So, a 50% request has resulted in about a 47% on plate. Also note that it is typically not a classic Bell curve - there is no symmetry. As I said, other CtP/plate combinations would have their own characteristic natural curves.

So, the question is: do you apply a plate curve to linearize that natural curve and zero it - then apply any needed press curve to achieve the press tonality you want? Or do you only apply a curve to the existing natural plate curve that results in the dots needed on the plate to achieve the final press tonality?

Hope that's clearer. Gordo

 

[Lukas Engqvist]

A question about the curve. Is the curve above true for all angles and all rasters? Or would you have a series of plate curves, for each angle and each raster ay each resolution?

If the deviation is due to the problem of dots being digital as opposed to analogue then a "curve" would be a missrepresentation, a scatter graph would be more accurate, but ofcourse harder to apply. I guess what I am saying is that even if I can see that a perfect linearisation would perhapps be an advantage. The potential errors and the complexity of keeping all the variables without error and atune to the processs seem impractical.

So even if it is not a perfect bell curve, at least with the combination of plates and raster we use it is reasonable to consider the deviation from the bell curve negligible (ie less than the deviation by imperfect readings).

 

[gordo]

A question about the curve. Is the curve above true for all angles and all rasters? Or would you have a series of plate curves, for each angle and each raster ay each resolution?

If the deviation is due to the problem of dots being digital as opposed to analogue then a "curve" would be a missrepresentation, a scatter graph would be more accurate, but ofcourse harder to apply. I guess what I am saying is that even if I can see that a perfect linearisation would perhapps be an advantage. The potential errors and the complexity of keeping all the variables without error and atune to the processs seem impractical.

So even if it is not a perfect bell curve, at least with the combination of plates and raster we use it is reasonable to consider the deviation from the bell curve negligible (ie less than the deviation by imperfect readings).

In my experience the characteristic uncalibrated curve is related to the raster. For example, here are the curves for the original 175 lpi AM, compared with 1st order and 2nd order FM screens on the same plate:

IMHO it doesn't matter whether the result is a bell curve or not. What matters is that, whatever the curve is, that it's consistent. Basically, I agree with your approach - I see no value in linearization before applying a press curve. But I'm not voting in this poll because I want to know, and understand, what people are doing and why, rather than what I think they should be doing.

best, gordon p