offset moire

[shangai]

I have an old pic from a newspaper. It is printed using the offset method. What is the rigth procedure to remove this pattern from the picture using photoshop?.

[Elgreco]

There was another thread addressing this same problem.

Moire

Pete

[shangai]

OK, can you give some help about it?

[Phil_The_Rodent]

There was, ElGreco, but it was never resolved. I tried a couple of things, and the only thing you can really do is blur or shrink it until the moire disappears from what I could tell after a lot of playing...

[Cappy]

After you scan are you looking to eliminate or diminsh the moire on the screen or in Print or both.........?

[Phil_The_Rodent]

Oh yeah, thanks Cappy. Theory goes that you should be able to deal with it if you scan at the exact resolution the print was in. I was referring to fixing a moiré after the scan stage... of course, I've never tried this. Working in print, I would never be able to pull a good resolution off of a piece of print work. That is, without throwing some artistic filters at it. And if I'm throwing filters at it, then I wouldn't worry about moirés...

Do you have some suggestions, Cappy?

[shangai]

I have scanned it , and now the question is how the remoe the offset pattern, that is the dots from the scanned image
Thanks

[Cappy]

The premise lies within the screen angles of each color (CMYK) and the line frequency. If you could duplicate the exact dot size and shape, resolution, frequency and angle for each color when scanning the original and then match the configuratiion when printed you could lose the moire.
Typically it is a 45 degree angle for K / 75 degree angle for M / 90 degree angle for Y / and 105 Degree angle for C............... This, is of course based on a frequency of 60 lpi and a resolution of 300 dpi........... you could manipulate the screen frequency and try maybe 53 or 85, if my memory serves me right. It is a lot of trial and error but in my experience, the answer is in the scan or in the output configuration if going to film. If your taking it to the web, then it has to be done in the scan. If you could make each color a different layer and change your angles that may work. Nope, cause your always looking at RGB, that is what your monitor is giving you. It has to be done in the scan. Manipulate your resolutions and percentages......... I might try to simulate this if I get time and I will post the results....... It is always like shooting craps, the printed piece offers up so many variables even having the original configuration when out putting to film won't help. You have dot gain from the press, paper stretch, depending on the speed it was printed at you could have an elongated progression in your pattern and frequency, usually the press configuration is opaques first transparents last, so on press it is k (1st unit) c (2nd) m(3rd) y (last) so your frequency from K to Y could be 60 lpi on the k unit and 75 lpi on the y unit so ........
you would have to manilulate your line frequency from K to y trying different values for each color, resolution would have to be global, no option there and you would have to manipulate screen angles for each....... you will find your answer there....... then convert to RGB and hope............ Find the original, it is a lot easier and less challenging!

[Cappy]

You won't get rid of the dots, everything is dots................ I thought you were speaking of patterns........ that is what a moire is... your answer for the best your gonna get lies in trial and error based on my previous post............... good luck Just the fact that your scanned piece is on newsprint triples your degree of difficulty. Not to be depressing, just realistic.

[Edited on 0 by Cappy]

[Phil_The_Rodent]

Ask a silly question huh?

I'm thinking, Cappy, that if we're looking at a piece of newsprint, as long as it's completely mint condition (no stretching, folds, wrinkles, etc.) that if you capture the exact resolution, you won't really have to worry about screen angles (discuss amongst yourselves...). Take that little point of the scan which will be a pixel in the future. At any "supersaturated" point (containing some of each CMY&K), it should line up with a portion of each dot from the plates, correct? (I'm wondering if angling the dots will compress certain ratios along the X-Y axis). But if you capture the same portion of every dot, then the scan will average the colour and call it a pixel. Then you just have to adjust the levels (to compensate for whitespace) and curves (to compensate for ink inconsistency and paper colour) on the channels in PS. Again, I might be thinking to linearly. Actually I think I am. The screen angles will mess that up, won't it?