Proof vs. Press: Images Match, Colour Panels Don’t Match

[Stephen Marsh]

Proof vs. Press: Images Match, Colour Panels Don’t Match.

I have seen this often enough over the years with many different proofing systems and presses etc.

Has anybody else seen this, or have any links to cited cases, or any other info?

The photographic image content of the press sheet may match a certified proof very closely to the “eyeball”, however large panels of colour may not match the proof.

I understand that large panels of solid and tint colour builds can affect how the press prints in these areas. No the press sheet does not have “solid ink take up bars”, there is no room, they only have some slur/doubling, solid primary, grey balance, and 80/40 tints etc. Colour bar measurements are fairly consistent, whether the patches are above the large solid blocks of colour or not.

The proof and press solid primary and 40/80 tint values appear to be within the F39/ISO Coated specification.

I can provide colorimetric results that would help with an answer. Any ideas?


Stephen Marsh

 

[gordo]

Proof vs. Press: Images Match, Colour Panels Don’t Match.

I have seen this often enough over the years with many different proofing systems and presses etc.

Has anybody else seen this, or have any links to cited cases, or any other info?

The photographic image content of the press sheet may match a certified proof very closely to the “eyeball”, however large panels of colour may not match the proof.

Because the image areas typically have a great many small discrete areas of color the eye/brain integrates the overall image appearance. You can't distinguish a small area of individual color to make a comparison. So you compare the overall appearance. A large flat area of color, on the other hand, allows you to make a direct one to one comparison. A large area of flat color can manifest in an image - like a large blue sky - at which point you'll have the same issue of "match" that you have with a large flat screen tint build. It's one of the reasons that press operators cut the press sheet through an area important imagery and place it over the proof to make a direct comparison. If they just look at the image on the press sheet an compare it to the proof it may appear the same even though a direct cut-through comparison shows they are different.

best, gordo

 

[Stephen Marsh]

Thanks Gordo .

In "by the numbers" proofing and printing, there is an expectation of a closer match with little work, colour just falls into line.

I know what LAB values the F39 spec calls for. I know the proof matches these values for the large solid area. I know that the press does not match the spec. or the proof (either visually or by spectro).

Press solid and tint values meausure OK, but the panel colour is off .

So if the plate curve is adjusted for the large panels, then the images may no longer match and the tint values may no longer be to spec.

The press guys say the proof is off, however the proof is proven to be to spec. In the press guy’s mind, it obviously cant be their printing that is off, as the image matches pretty good - so the proof must be wrong!


Stephen Marsh

 

[Chasfinch]

Just out of curiosity what is the tint build and do you have a 4 color tint book. I am finding out that photos are the easiest to make "look" correct - I have been having several issues with dot gain density as well as ink tack. And almost any picture looks fine- but get to large areas of 4 color builds and you can see that the press is not running correctly. The pressman at my shop are open and work with me.

 

[Stephen Marsh]

Just out of curiosity what is the tint build and do you have a 4 color tint book. I am finding out that photos are the easiest to make "look" correct - I have been having several issues with dot gain density as well as ink tack. And almost any picture looks fine- but get to large areas of 4 color builds and you can see that the press is not running correctly. The pressman at my shop are open and work with me.

Specific details follow:

CMYK Build: m25 y95

Fogra39 Target Value: L*78 a*11 b*76 (rounded off)

Press Measured Value: L*81.3 a*8.2 b*81.1

Proof Measured Value: L*78.6 a*11.2 b*76.0

Target to Press dE76: 6.6 (Tolerance < 6 dE76)

Target to Proof dE76: 0.6 (Tolerance < 6 dE76)

Summary: The proof is great (both visually and by the numbers), the press is just out of tolerance and is too light. I have also compared the proof and the press result to the Altona Test Suite print on type 1 paper for 30m100y (being the closest CMYK build).

_____

Press L*a*b* values for magenta solid & tints:

Press 100M = L*49.5 a*74.5 b*0.5
Target 100M = L*48.0 a*74.0 b*-3.0 (rounded off)

Target to Press dE76: 3.8 (Tolerance < 5 dE76)


Press 40M = L*77.5 a*26.6 b*-8.0
Target 40M = L*76.0 a*26.0 b*-7 (rounded off)

Target to Press dE76: 1.9

_____


Stephen Marsh

 

[Prepper]

Matching

Matching

Some thoughts, but first some comments on the process itself. I too get frustrated when things like this come up, but it is just part of the process that we can't control, there is variation and sometimes that variation is larger than we'd like it to be. Being in prepress for 30+ years, color management has improved our printing here and the matches we get a lot but it is still not perfect and is way too often for me, "pleasing" color, but still very close to or within industry accepted tolerances. A lot has to do with solids and which way they vary from the target and then also dot gains which vary from day to day, from press to press, from one paper to another, and even within any given run, so some runs will match really well, some may not, depends on a lot of things, all those variables, what color it is, what the gains are right then, etc. etc.

I wish it was as exact a science as it is in my mind, but, in the reality of offset printing, it isn't. Still marvel that it works as well as it does, that with a 20 micron FM screening, we can image dots on 4 different plates, mount them on four different cylinders of a press, print them within pretty tight register at 10,000 sheets an hour and control it enough that we can produce a fairly close and sometimes really close proof of what it's going to look like.

The color you are dealing with is 25% M, quarter to half tones are where I would expect dot gains to make the most difference in color. Dealing with a 2-color MY tint, any variation in M will probably change the color somewhat. Here it's lighter on press so M gains may be a little light today or at that point in the press run. And your Lab of 100% M shows the ink is a little light itself compared to the Target value to start with right? So the proof is really close to the target value but the M ink is lighter to start with.

There is also what Gordo said, large flat tints, you will notice small differences in color a lot easier than images, and some colors more than others. Some colors a 3.8 dE76 isn't that noticeable, others it looks huge. Just nature of the process. And Murphy's law takes over a lot more than you'd think, if it can, it will, everything can match perfectly except for this one color and it will invariably be on the front cover of an important piece. For us anyway, it happens more than we'd like to admit.

All that being said, yes, pressmen always blame it on the plates, prepress always blames it on the press, but in reality it all really is just part of the process. As long as everyone is on the same page and doing their part the right way then you're getting the best results possible. If that tint is then off to far, someone has to decide whether it's ok or whether to spend time and money to tweak it in prepress if it can't be done with tweaking densities at the press without affecting other elements too much. There will almost always be a certain amount of compromise in the mix.

 

[iSeeker]

What aim do you use to determine your adjustment curves?
Linear perceived neutral scales (G7) or FOGRA A/B desired TVI curve for each ink (ISO)?
I think Images tend to better match when using curves that also adjust gray balance (G7), while linework tend to better match when keeping single-inked tones to target (ISO).
You mention that 40/80 is within ISO target, but what about the other tones (25/90)?

 

[Stephen Marsh]

Thank you for the reply Prepper!

Your thoughts on the process itself help to reset the expectation. Things have mostly become so good these days that when there is a major discrepancy in a single area it really stands out and is made into a big issue. With a ¾ page light orange panel, the difference was noted.

You are indeed correct that the solid and 40% tint of magenta are a little bit lighter than the target, even if they are within acceptable tolerance. You are also correct that this is a snapshot, it is yet to be determined if this is a trend or a single case.

My direct concern could easily stop at the proofing, there is nothing wrong with the proofing, so this should be considered the end of things from my official proofing support viewpoint, the rest is in-house for the customer to work out (they are not paying us to work out their process control and proof to press matching etc).


Stephen Marsh

 

[Stephen Marsh]

Thank you for the reply iSeeker. I am not sure what method that are using for their TVI curves. There are no 25/90 patches, the colour bar is very small (almost too small to measure with an i1 pro). There are only 40/80 tint patches of the primaries, solids, secondary overprints, gray balance patches, S/D targets etc. So I only have limited info to go by. I will inspect the M+Y overprint patches to see if there are any clues in ink trap.


Stephen Marsh

 

[Stephen Marsh]

Apparently the pressroom does not have a densitometer. They eyeball the CMY grey balance bars so that they look close to the 50%K value. I have checked the results and this midtone grey balance bar varies over the length of the sheet, it is not consistent.

Faced with such conditions, I think that it is next to pointless attempting to do anything in prepress that affects their plating paper curves (this is what the client wishes). I think that they are trying to hit a moving target, if the curves work this time, who is to say what will happen the next time? I would recommend that they only make these decisions after looking at many jobs to see if there is a trend.


Stephen Marsh

 

[rich apollo]

Stephen, if my thinking is correct, you're wrestling with two issues. 1) The magenta is printing a bit light in the quarter-tone. 2) The magenta ink is a bit on the warm side, to begin with. How is the M+Y overprint?

I'm guessing a little boost to the M curve would remedy this particular discrepency. Don't know what kind of issues it might open up.

 

[gordo]

Apparently the pressroom does not have a densitometer. They eyeball the CMY grey balance bars so that they look close to the 50%K value.

I was going to post something pithy about grey balance and G7 but couldn't beat what you wrote. ROTFL

Best, gordo

 

[Alois Senefelder]

Saga

Saga

Hello Stephen and gentlemen of the Pre- Press,


From reading this Saga, in my opinion this is On Press Problem

Some PDFs that I hope will be of "Interest and Value"


Regards, Alois

 

[tmiller_iluvprinting]

Apparently the pressroom does not have a densitometer. They eyeball the CMY grey balance bars so that they look close to the 50%K value.

Whew! I had to check the calendar to be sure it wasn't 1990!
How are the trap numbers?

Todd

 

[Prepper]

They eyeball the CMY grey balance bars so that they look close to the 50%K value.
Stephen Marsh

I wish our pressmen would even look at the gray balance bars! They have a scanning densitometer but don't check the gray and don't even notice that it's red, or green, or dark, or light. Still stuck on the "craftsman" way of adjusting everything to their eye.

 

[Stephen Marsh]

Whew! I had to check the calendar to be sure it wasn't 1990!
How are the trap numbers?

Todd

Target 100m100y trap/overprint (rounded):
L*47 a*68 b*48

Measured 100m100y trap/overprint:
L*48.9 a*71.4 b*47.2

dE76: 3.9 (< 6 dE76 tolerance)

So yes, the red overprint is lighter than the ideal target, however it is again within tolerance for error.


Stephen Marsh

 

[gordo]

What are the trap values as opposed to the Lab values?

Gordo

 

[Stephen Marsh]

What are the trap values as opposed to the Lab values?

Gordo

What values would you be interested in seeing Gordo? I have been using an i1 Pro and ColorPort for these spot measurements (I can provide Status T or E density, LAB, spectral etc).


Stephen Marsh

 

[gordo]

What values would you be interested in seeing Gordo? I have been using an i1 Pro and ColorPort for these spot measurements (I can provide Status T or E density, LAB, spectral etc).


Stephen Marsh

Overprint trap...using a densitometer...
Measure first color down.
Measure second color down.
Measure overprint.

http://www.aleqs.nl/documentatie/AN4TRAP_E.PDF

Oh,and the densities your running at would be good.

Best gordo