230 Line Screen Banding

[tmason]

We are running a new 230 Line screen set on our True Flow 6 system with a Screen, Platerite 6600S Platesetter and Fuji LH-PJ plates. We are experiencing banding in gradations. The jobs were set up in Illustrator and In Design and look great at 175 and 200 line screens but when we run with the new screen set for 210 line and 230 line, the banding shows up. The 230 screen set is TF1-724SQ01-H12. This screen set has a seven degree offset so the angles are not the same as our 175 and 200 line screens.
We can re create the blends in Photoshop and make them look good but it would be impossible to do this for every job. Also, expecting the sales staff to "train" our customers isn't realistic.
Has anyone resolved this issue without switching to Hybrid screening or reworking every file with blends?

Thanks for your help,

Todd

 

[gordo]

We are running a new 230 Line screen set on our True Flow 6 system with a Screen, Platerite 6600S Platesetter and Fuji LH-PJ plates. We are experiencing banding in gradations. The jobs were set up in Illustrator and In Design and look great at 175 and 200 line screens but when we run with the new screen set for 210 line and 230 line, the banding shows up. The 230 screen set is TF1-724SQ01-H12. This screen set has a seven degree offset so the angles are not the same as our 175 and 200 line screens.

The 7 degree offset is a screening technique to prevent single channel moiré. It doesn't have anything to do with banding/shadestepping.
If your output device is 2400 dpi, or greater, you should not see shadestepping if you are running linear plates. However, you are likely not running linear and that could cause shadestepping. Where in the tone scale does shadestepping occur?
Hybrid screening only effects the highlight <5% and shadow dots >95%. If those tones are lost due to imaging resolution issues it can recover those tones by using fewer larger dots. I doubt that would have any impact on shadestepping though.

best, gordon p

 

[tmason]

Gordo,

We are running plates at 2400DPI and checking into running at 4000 if we can get a screen set for that resolution.
The Banding is occurring in the highlights 0-5%. We did run linear plates for our original G7 calibration and our colors match perfectly running 175, 200 or 230. We have not tried to run a linear 230 plate, with a gradation. I will do that today as a test. We do check our plate dots and we are not losing the 1% dot on our plates.
The strange thing is, we need to use a different screen set with 210 and higher screens. With our 200 line screen and the set that has no offset, we don't see the banding.
To sum it up, 200 line screens don't show the banding, 210 does.
Is it because of the screen set and is there a better screen set to use for 230 line screens?
How do shops run 600 line screens waterless?
Thanks

 

[tmason]

Gordo,

I read your piece on AM/XM screening and highlights. We are getting Fuji and Screen together to come up with the best screening technique for us.

thanks for your help.

 

[joz2k]

Hello tmason,
Which printer driver do you use? I had same problem a month ago. Green gradient with a difference 10-15% from end to end looked like 10 or more steps on the plate. I used Scitex and DotMate printer drivers. After swithching to Xitron I made new postscript file at perhaps 300 lines and about 4500 dots i cant remember exactly so there was no any gradient problem anymore... Gradients was made in Illustrator, i even tryed to rasterize them in Photoshop using dithering... same result. So by my side the problem was in Scitex printer driver i used in the beginning. Xitron ppd you can download if i'm not wrong from Xitron website and its free.
regards
Joz

 

[RickLee]

is the problem only in photoshop? you do know that photoshop if in 8 bit colour mode will only give you 256 shades for each colour channel. by comparison Illustrator EPS format in Postscript level 3 (PS3) will give you 4096 shades per colour.
The problem in photoshop comes with the length of area you are trying to cover. for example. if you want a gradient to cover 5" area and the blend goes from 1 percent to 5 percent. this means that each inch will show a change of 1 percent. each inch would show 3 distinct bands. formula (100 percent divided by 256 shades equals .39 so each shade represent .39 percent of colour. for each percent you will get 3 steps.
in photoshop you can get around this by deceiving the eye. Use the filters to add Noise to break up the pattern so you don't get straight lines or Diffuse or Displacement to move the adjacent pixel to blur the banding.

The other option would be to make smaller gradients or make the gradient in illustrator.

If you have a modern RIP you might be able to use the 16bit colour in photoshop to correct the problem but the whole image must have been made in 16 bit to start with.

All this also does not account for DPI in combination of LPI. That I believe you are already looking into. Don't believe the whole super cell structure crap to get more shade with less DPI.

 

[VladCanada]

Who knows a phisical LPI limitation for 64 chanels Platerite 6600S and LH-PJ plates?

 

[gordo]

All this also does not account for DPI in combination of LPI. That I believe you are already looking into. Don't believe the whole super cell structure crap to get more shade with less DPI.

Unfortunately Screen doesn't have any info on their screening options. I detect some irony there. All I could find on their website is:
"Product Overview (Conventional) Screen offers a number of options for conventional screening."

The halftone dot dithering in supercell screening that is used to overcome DPI/LPI constraints has been in service since the early 90s and is now used by virtually all prepress systems vendors in their conventional AM/XM screening. Far from being crap - it works very well. I would be surprised if Trueflow does not use it - but anything's improbable these days.

best, gordon p

 

[Bret Hesler]

I agree with Gordo, and I have had banding problems with non- supercell screening. It happened with the Paragon screening from Artwork Systems. 175LS was fine, a few problems at 200LS, major problems at 240, 283 and 340LS. There was no was no variation in the size of dots ( all one pixel dots, then all two pixel dots, then all three pixel dots,... ), so the number of levels was dropping as the LPI was increasing. We switched to the Concentric screening, which had supercell screening, and the problems went away.

Do you have a way to look at screened images files before they are plated. I ran a 256 level target from Don Hutcheson through the Nexus RIP to Agfa PrintDrive to look at the screening. It was plain as day when I zoomed in to the .4%, .8% and 1.2% values and saw only 1 pixel dots for all three levels.

Bret

 

[RickLee]

Perceptible banding will go away with super cell structures however you will not get 100% predictable dot placement therefore you will develop artifacting on your screen tints. if this acceptable then fine. most of the customers won't notice it.

The reason why multi/cuper cell creates the artifact is that the higher the LPI goes the smaller the spot gets. there comes a point when the 2400 DPI won't resolve the size needed especially in the 1,2,3 or 97, 98 or 99 percentages.

I have had Fuji trying to tell me that they have customers running 2400DPI and can output 10 micron FM. the problem is at 2400 DPI the minimum spot is 11 micons. you would need to use 2540 to get 10 micron spot size.

maybe I am being too critical but those are the numbers.