Different dot rosette appears different color shade in one printed sheet

[DeltaE]

Hello!

I have an issue with a job (please check the attached picture for more detail)
- sheet size 750 x 540 mm, 157 gsm, coated paper.
- CMYK job with color sequence KCMY
- We print on a sheeted offset machine, just installed 4 months ago.
- I let machine run stable and take a sample sheet.
- On the picture you will see 50x zoom and 500x zoom by a digital microscope in ONE SHEET.
- We all find the different rosette of dots in the left-upward compared to the other position.
- Then I got a different grey shade in the left-upward position, it looks a little bit more yellowish than other.

I had remounted the plate in yellow unit, and tried to make diagonal value in print console nearly zero. But the rosette of dots is still the same.

I understand that if I can make the same rosette of dots in all 4 position, I will have same grey shade, am I right?
Anyone with experience in offset sheet machine please tell me how to get same rosette in all area of printed sheet.

Thanks in advance!
Regards,
DeltaE

 

[gordo]

If the rosette structure is altered - e.g. going from clear centered to dot centered - then you will get a color (and tonal) shift.

A possibility:

Back sheet flare This usually happens when a lightweight paper is printed with heavy ink coverage. As the sheet is pulled through the press the tail corners of the sheet fan out slightly causing misregister in the corners and results in a dot shift in the rosette with a subsequent shift in color.

Or, back stretch. This can occur if blanket and packing on one or more units may have been compressed during the pressrun, causing change in the print length.

Or waggle if there's a problem with the on press register system and/or side guides. But you'd likely see that effect in other corners.

Another possibility is rosette drift. That is the result of the halftone screen not being optimized for the plate size. Rosettes are high frequency moirés. If the frequency of the moiré is not optimized for the size of the plate then the rosette will drift and result in a color shift. If you can preview the screened bitmap before it's plated or if you can image all 4 colors on one plate then you can see if that's the problem.

Solutions:

For a neutral image like that you should be using a maximum GCR separation so as to minimize the formation of rosettes, or use FM screening which eliminates rosettes altogether and results in greater color stability on press.

 

[Alois Senefelder]

DeltaE,


Why involve expensive C.M.Y. print separations, when you are trying to reproduce a "Monochrome Halftone Image" ?????

Are you using "Printing Control Strips" if not - why not ?


In my opinion .... the leave edge L/Hand Image --- Black, a local slur in that area YOU do need to check the print control strips !!!!

1) Also check the Black Blanket tension is correct. NO slack area !


Regards, Alois

 

[DeltaE]

gordo
Please explain more about "...Another possibility is rosette drift. That is the result of the halftone screen not being optimized for the plate size. Rosettes are high frequency moirés. If the frequency of the moiré is not optimized for the size of the plate then the rosette will drift and result in a color shift..." I do not really understand the optimized halftone screen.

Your advice: imaging all four color in one plate show me something interesting. I found that the rosette in one plate in not the same in whole plate. Somewhere I can see clearly white in center and a very shape white circle, but there is some area not like that. So I think even the sheetfed machine has very good registration and very stable paper transfer system, I still can not get the same rosette in all area because it is different when combined CMYK in plate side.

Your advice about GCR, I have already received in another topic, I am very appreciated but to much black inside could make detail more grainy that print-buyer could reject.

BTW, we are using internal drum CtcP system. Do you think that could be the cause?

FM could be very good choice for me! Many thanks!

@alibryan,
We both know that middle area of sheet is stable area than on the side. But we can see the rosette of right side same as the lower-left position. And I also request printer to fine adjust registration four corners and of course mount plate then the diagonal value on press console is nearly zero. Then I can confirm that we have good registration when I check the registration marks visually by microscope.

@Alois,
Yes, of course we use some color strip. In the picture, it is not the printed sheet, it is only a artwork to help me showing my issue more clearly.
Blanket tension is a good point to check. Yes, I will.

Thank you all for your comment!
Regards,
DeltaE

 

[turbotom1052]

Some good suggestions so far. I like the one about paper quality as a first check. As a test try throwing about 100 sheets into the feeder of a different stock. It should be a stock with a bit more body, and most importantly, a stock that lays perfectly flat. No wavy edges for this test please. The reason I suggest 100 sheets is that you may have to make a couple of minor fit adjustments to the new paper. The other thing worth asking, is if your pressman is capable of making fit adjustments beyond the moves available on the press console? Sometimes with particularly fussy forms ,and IMO 4c black and white, along with any sort of textured images can be very fussy, there is the requirement to dig in a bit more than the moves available on the console. Most presses make available a bit of movement right on the plate clamps to eek out that little zonal move that sometimes makes the difference between a salable job and something for the recycle bin. Id also be looking at how smoothly the paper goes into the press infeed.

 

[gordo]

gordo
Please explain more about "...Another possibility is rosette drift. That is the result of the halftone screen not being optimized for the plate size. Rosettes are high frequency moirés. If the frequency of the moiré is not optimized for the size of the plate then the rosette will drift and result in a color shift..." I do not really understand the optimized halftone screen.

I'll try.

When you have any two, or more, repeating patterns (e.g. halftone screens) overlapping each other a moiré pattern, effectively an artifact, occurs. Because the overlapping screens have a frequency the resulting moiré will also have a frequency.

If you rotate the overlapping screens relative to one another the frequency of the moiré pattern changes:

As you can see in the above, when the angle difference between the screens is small the moiré pattern frequency is large.
When the angle difference between the screens is large the moiré pattern frequency is small.

The moiré pattern is least visible when its frequency repeats over a very small distance (e.g. 1/64 inches) or over a very large distance (e.g. 64 inches).

In 4/C printing the screen angles for K, C, and M are set at 30° apart which creates a small moiré pattern frequency. The moiré pattern that's formed when this happens is called a rosette.

The rosette can be either clear centered (typical) or dot centered. (Yours are clear-centered which is typical)

Unfortunately digital halftone screen angles other than 0° and 45° are not accurate. They're very close but not absolutely perfect.

The small angle error in the C and M screens causes the rosette to shift from clear-centered to dot-centered to clear-centered. You can see this happen if you create a halftone screen using PhotoShop which isn't able to create correct screen angles for C and M. When the rosette shifts from clear to dot center you get a tone/color shift as the overprint relationship of the ink changes.

Modern screening design will usually try to get the angles close enough so that the distance of the shift from clear-centered to dot-centered rosette is greater than the diagonal of the largest plate size that will be imaged in order to avoid the tone/color shift associated with the change from clear-centered to dot-centered rosette.

• Side bar. The Yellow printer is not part of the rosette because its frequency (lpi) is higher that the C, M, or K printers. The method used to increase screen angle accuracy is called "supercell" screening and some variation of it is used by most of today's RIPs.

 

[Alois Senefelder]

DeltaE,


Clarification: This 4 Colour Halftone Image, are you printing the same image 4 - UP on the Sheet ?

1) The L/H leave edge image is the problem ??

Regards, Alois

 

[DeltaE]

DeltaE,


Clarification: This 4 Colour Halftone Image, are you printing the same image 4 - UP on the Sheet ?

1) The L/H leave edge image is the problem ??

Regards, Alois

My problem is the grey of picture in lower left, lower right and upper right are same, but the grey in upper left is more yellowish than others.

regards,
DeltaE

 

[Alois Senefelder]

DeltaE,

Lost in Translation !!!


OK...... I'll start again, THE same image is printed down onto the Plate --- 4 times YES /NO

On a printed sheet 3 of 4 the images are printed correct - YES/NO

The 2 images at the GRIP EDGE of the sheet are OK YES?NO

The 2 images at the leave edge ... I (one) is OK and 1(one) not YES?NO


Regards, Alois

 

[DeltaE]

Alois,

You are right, all answers are YES.

rgds / deltaE

 

[Alois Senefelder]

Gentlemen and fellow Lithographers,

AT Last !!!

The podium is yours ... you need to get your "Thinking Hat" on


Regards, Alois

 

[gordo]

Gentlemen and fellow Lithographers,

AT Last !!!

The podium is yours ... you need to get your "Thinking Hat" on


Regards, Alois

Only two possibilities- sun spots or sorcery.
(Apologies, couldn’t resist. Now back to our regular programme.)

 

[DeltaE]

I'll try.

When you have any two, or more, repeating patterns (e.g. halftone screens) overlapping each other a moiré pattern, effectively an artifact, occurs. Because the overlapping screens have a frequency the resulting moiré will also have a frequency.

If you rotate the overlapping screens relative to one another the frequency of the moiré pattern changes:

As you can see in the above, when the angle difference between the screens is small the moiré pattern frequency is large.
When the angle difference between the screens is large the moiré pattern frequency is small.

The moiré pattern is least visible when its frequency repeats over a very small distance (e.g. 1/64 inches) or over a very large distance (e.g. 64 inches).

In 4/C printing the screen angles for K, C, and M are set at 30° apart which creates a small moiré pattern frequency. The moiré pattern that's formed when this happens is called a rosette.

The rosette can be either clear centered (typical) or dot centered. (Yours are clear-centered which is typical)

Unfortunately digital halftone screen angles other than 0° and 45° are not accurate. They're very close but not absolutely perfect.

The small angle error in the C and M screens causes the rosette to shift from clear-centered to dot-centered to clear-centered. You can see this happen if you create a halftone screen using PhotoShop which isn't able to create correct screen angles for C and M. When the rosette shifts from clear to dot center you get a tone/color shift as the overprint relationship of the ink changes.

Modern screening design will usually try to get the angles close enough so that the distance of the shift from clear-centered to dot-centered rosette is greater than the diagonal of the largest plate size that will be imaged in order to avoid the tone/color shift associated with the change from clear-centered to dot-centered rosette.

• Side bar. The Yellow printer is not part of the rosette because its frequency (lpi) is higher that the C, M, or K printers. The method used to increase screen angle accuracy is called "supercell" screening and some variation of it is used by most of today's RIPs.

Dear Gordo,

Thank you for your explanation!
Now I understand clearly, and will try to find the way to deal with this.

best regards,
DeltaE

 

[DeltaE]

4-color printing is in effect, a controlled moiré. So if it only appears uneven in one area (assuming it's consistent from sheet to sheet), then it's either a problem with the imaged plate(s), or there's uneven pressure somewhere from the cylinders on the press. This probably leads back to an onpress fit problem and there are more than a few things that can cause this.

Whatever it is, there are ways to find it but I'm guessing the problem is something that goes on in most of the work coming off of this press. It's probably just more pronounced (easier to see) with this particular job.

We had tried to print this job on the other machine (older machine) then the result is worse.
The machine which we are using to print this job is a brand new famous German sheetfed machine. It is a good machine, and has been qualified by test jobs. I dont think any mechanical issue in it. The thing I am concerned that is 5 years old internal CtcP device.

Thank you for your attention!
Regards,
DeltaE

 

[DeltaE]

gordo ,

I have just read again your blog print-guide. And I understand your advice about rosette drift.
45 or 90 could not be exact 45 and 90 degree on the plates. Then I will get rosette drift in printout even trying to do as good registration as possible.
Remedy could be shift the screen angle 7,5 degree every screen. Do you think that could help to eliminate rosette drift?

Thank you!
Best regards,
DeltaE

 

[gordo]

gordo ,

I have just read again your blog print-guide. And I understand your advice about rosette drift.
45 or 90 could not be exact 45 and 90 degree on the plates. Then I will get rosette drift in printout even trying to do as good registration as possible.
Remedy could be shift the screen angle 7,5 degree every screen. Do you think that could help to eliminate rosette drift?

Thank you!
Best regards,
DeltaE

No. 45 or 90 are an exact 45 and 90 degree on the plates. The other angles cannot. (There is a micro issue even with a 90° and 45° screen angle because of how drum-based CtP lasers image plates but that's getting into the technical weeds)

The 7.5 degree screen rotation is to solve a completly different problem.

It is used to eliminate "single channel" moiré when the lpi of the screen is not an exact divisor of the dpi of the output device (e.g. 175 lp on a 2400 dpi device).

When the lpi of the screen is not an exact divisor of the dpi of the output device you can end up with an extra pixel on each halftone dot. If the extra pixels repeat halftone dot to halftone dot then they will have both angle and frequency. That angle and frequency can harmonically beat with the screen that they're on. So you now have two screens conflicting - the underlying screen e.g. the black halftone screen and the pattern formed by those extra pixels on each dot of the black halftone screen. Two screens like that create moiré. In this case within the screen itself - i.e. single channel moiré. The 7.5 degree rotation introduces "noise" in the screen and thereby eliminates the possibility of those extra pixels forming a repeating pattern and hence eliminates the possibility of single channel moiré.

 

[DeltaE]

No. 45 or 90 are an exact 45 and 90 degree on the plates. The other angles cannot. (There is a micro issue even with a 90° and 45° screen angle because of how drum-based CtP lasers image plates but that's getting into the technical weeds)

The 7.5 degree screen rotation is to solve a completly different problem.

It is used to eliminate "single channel" moiré when the lpi of the screen is not an exact divisor of the dpi of the output device (e.g. 175 lp on a 2400 dpi device).

When the lpi of the screen is not an exact divisor of the dpi of the output device you can end up with an extra pixel on each halftone dot. If the extra pixels repeat halftone dot to halftone dot then they will have both angle and frequency. That angle and frequency can harmonically beat with the screen that they're on. So you now have two screens conflicting - the underlying screen e.g. the black halftone screen and the pattern formed by those extra pixels on each dot of the black halftone screen. Two screens like that create moiré. In this case within the screen itself - i.e. single channel moiré. The 7.5 degree rotation introduces "noise" in the screen and thereby eliminates the possibility of those extra pixels forming a repeating pattern and hence eliminates the possibility of single channel moiré.

gordo
Yes, that is my mistake.
So if I want to eliminate the rosette drift due to not-exact angle (75 and 15), please tell me what I should do?

In your blog, you also mentioned that some well-designed screen, the rosette drift can be inhibited. Please tell how can I know which is well-designed screen?

Best regards,
DeltaE

 

[Alois Senefelder]

Gentlemen and fellow Lithographers,


Before we ramble of into the realms of Screen Rulings and Rosette Patterns, we have a problem to resolve.

Salient Points:

A print run of 4 > up x 4colour identical images printed KCMY, of these 3 were printed OK and 1 with tonal colour variations.

What are the ODDS of CtP imaging to reproduce 1 image out of 4 - incorrect >KCMY printer separations images ???

Next, we need to compare the suspect image on the plate, then on a printed sheet of each

single colour KCMY .


DeltaE, if these steps have been done- what are the conclusions ??

Regards, Alois

 

[gordo]

I didn't mean to suggest it was a press defect, only that it's either a plate problem, or a press problem. If it is from your platesetter, then most of your jobs probably have this problem but it just shows more in this particular job. Hopefully, you'll be able to find a solution. Good luck to you!

Correct. You can't just start throwing solutions at a problem without understanding what the problem is and what caused it.
I just suggested some possibilities - there are others.