FM/Stochastic Screening

The Benefits
Below are listed some of the reasons why printers adopt FM/Stochastic screening for some, or all, of their presswork. In the next posting I'll be going into more detail on some of the benefits that have not been covered in past posts on this topic.

1 - No screen angle moiré
2 - No subject moiré
3 - No rosettes
4 - Photographic/contone look Some examples are mentioned HERE
5 - Greater tone and color stability as SIDs naturally vary during press run
6 - Larger color gamut The reason for the increased color gamut is explained in the post HERE
7 - Faster drying
8 - Reduced ink usage - About 10-15% depending on tone content of the presswork. The reason for reduced ink usage is explained HERE
9 - Tonal and color stability when misregistration occurs
10 - Halftone dot structure stability when misregistration occurs
11 - Competitive differentiator

It's important to understand that, when it comes to the press, how halftone dots are organized is not as important as the size of the actual dot. As a result, some of the benefits of FM/Stochastic screening are available to conventional AM/XM screening if the lpi (halftone frequency) is fine enough. Specifically benefits 6, 7, and 8 in the list above can be delivered by AM/XM screening if run at an equivalent lpi to the FM halftone. As a general guide, the lithographic AM/XM/FM equivalents for a "second order" FM screen at various dot sizes are shown below.
Color stability through the press run

Printing presses are stable - but not consistent. Solid ink densities will naturally vary through the press run, and when they do two things happen; dot gain varies, and ink density on the halftone dots varies. The result is variation in tone and color reproduction through the run.
When printers measure dot gain they measure it at the 50% tone value rather than at the 1% or 2% point because at around 50% dot gain variation is at its greatest.The tone variation happens because the large midtone dots in an AM/XM screen can accept more ink than the small dots in the highlights and deep shadows. Put another way, you can’t build a column of ink on a small dot, but the large dots around the 50% tone can take on the ink so they will grow the greatest. So, the midtone color gets darker and lighter as solid ink density naturally varies through the press run.

However, the halftone dots in an FM screen have a different ink profile.


FM screens effectively have small dots throughout the tone range. They act much like the highlight dots of an AM/XM screen and do not take on as much extra ink as SIDs naturally vary - so they remain more stable and hence the color remains more consistent through the run. For example, the below chart shows the dot gain profile (solid line) of a 133 lpi AM/XM screen and a 20 micron FM screen calibrated to match that dot gain profile. The dashed lines show the dot gain response when the solid ink density is raised by 30 points (i.e. 1.40 to 1.70) to magnify the difference. The curve with the small dashes shows the response of the FM screen, while the longer dashed line shows the response of the AM/XM screen.
The color and tone stability that FM screening provides is similar to that provided by heavy GCR color separations or ink reduction re-separation applications.

In some cases this on-press stability can be a two edged sword since, like heavy GCR separations, it does not allow the press operator the ability to move color quite as easily on press should the print buyer decide to be creative during a press approval. It also means that if the press operator makes a SID change, the region of color/tone shift will not be the same as that of an AM/XM screen.

That being said, consistency in color is usually the goal of presswork so the stability of FM screening is appreciated by printers and their customers as an excellent complement to their “print -by-the-numbers” goals.

Faster Ink Drying
One of the little known benefits of FM screening is that it helps the presswork dry faster. With FM screening, the tones are built with many small dots - effectively a “mist” rather than puddles of ink. The small FM dots carry a thinner ink film, and are distributed in a finer pattern that allows the ink to flash off its volatiles quicker causing the ink to dry faster than their AM/XM counterparts.

To illustrate how this benefit is achieved, looking at the same final tone value in the below image, on the left is a microphotograph of the large AM/XM dots while on the right is magnified the smaller FM dots.
In the below image I've taken those microphotographs of the dots and passed them through a 3D rendering program that translates density into height to see just how thick the ink film is. As you can see on the right, for the same final tone the FM screen has a thinner film of ink than the AM/XM dot.
The faster drying can also be demonstrated with a simple test. Below are three images printed on uncoated book stock, the lefthand image is 133 lpi, the center is 200 lpi, and the righhand image 20 micron FM. I then used a coin to scratch across the 3/C gradient as soon as the sheet came off the press:
It's very easy to see the marks made by the coin in the 133 and 200 lpi gradient - while the FM gradient is almost completely dry and hardly shows any marks from the coin at all.

The faster drying of FM screening brings other benefits, namely:
1- It's easier to align presswork with proof - since they are both dry one doesn't have to account for dryback
2 - Reduced setoff
3 - Prints better on uncoated, newsprint, foils and plastics
4 - Less spray powder is needed
5 - Less distortion on perfecting presses
6 - Jobs are faster to bindery
7 - Heatset presses may be able to lower their drying ovens a few degrees and thus save on energy costs

LPI Integrity
Even though most printers have experienced the problem, few have understood it. LPI integrity is a benefit of FM/Stochastic screening that may be a bit difficult to understand at first.

AM/XM screens generate a rosette pattern when all the screens are at the correct angles. A rosette is a high frequency moiré and as such is very unstable. If there is a slight misregistration, what happens is that the rosette pattern of the AM/XM screen changes - typically from clear-centered to dot centered. The next few images show the results of a half row shift in mis-registration:As the rosette shifts from being clear-centered to being dot-centered - the visibility of the rosette changes. Effectively, the rosette shifts from being high frequency to low frequency. Put another way, the effective lpi of the halftone can shift from a fine 175 lpi to a coarse 85 lpi as a result of slight misregistration. So, a rosette pattern that occurs in important skin colors may not be visible when in register, suddenly becomes very visible with slight misregistration. However, FM screens do not produce rosettes and as a result they maintain their lpi image integrity when slight misregistration naturally occurs.

Color Integrity When Misregistration Occurs

You can have the best presses and the best plate imaging available, however, because you are covering a sheet of paper with a mix of oil and water, squeezing and pulling it through the press - some amount of misregistration is an inevitable fact of press life. The impact of that misregistration on color and tone reproduction will depend on whether you are using an AM/XM or FM screen.

Here is a full sized image:

And here, enlarged, is a section of the image using an AM/XM screen on the left and an FM screen on the right. Both are in register.

And here is the same section but with the Magenta for both screens out of register by a half row of dots (1/300th of an inch) - considered within normal printing tolerances.
Notice how the dot structure in the rosette visibly changes in the AM/XM screen (175 lpi in this case) when it's out of register but the FM screen (20 micron in this case) remains stable since it does not have a rosette.

More importantly note the change in AM/XM rosette structure from clear-centered to dot-centered and back to clear-centered as it drifts in and out out of register. It’s like turning a light on and off and back again - so the color goes darker and lighter and darker and lighter through the run.

Also, misregistration changes the overprint and ratios of wet and dry trap. A blue made from Cyan and Magenta dots beside each other is not the same blue as when Magenta dots overprint the Cyan. The result is a color shift. So while the AM/XM presswork is going darker and lighter it is also shifting color from bluish to redish and back again as slight misregistration occurs. This can be particularly troublesome in, for example, car brochures where there are large expansive areas of neutral grey made up of 4/c process. Again, since FM screens don’t have a rosette the color and tone are better preserved with natural slight misregistration in presswork.

The bottom line - FM screening can help deliver more presswork color consistency - which, after all, is one of the main goals of printing.

Potential Issues
So far it's all good - but what about the potential downsides of FM/Stochastic screening?

A printing press or plate does not "know" how a pattern of halftone dots is arranged - it does not know whether the screening is AM/XM or FM. So, the problems related to FM screening are not usually the result of the halftone pattern itself, but by the size of the dots that make up the halftone screen. In that respect, FM screening shares the same problems associated with any high frequency halftone - FM or AM/XM.

This old Planeta press prints breakfast cereal boxes using FM screening for press runs that can last a week.
This new KBA press prints daily newspapers, inserts/flyers and commercial work using FM screening exclusively.
Here are some of the issues that can arise using FM and AM/XM screening.

1 - Grainy appearance in flat tint areas. This usually doesn't occur with AM/XM. There are several possible causes:
1) Poor screen design and/or using a first order FM screen pattern (see HERE for explanation). This problem has been overcome by modern second order FM screens.
2) Graininess can also result from a lack of plate imaging integrity e.g. plate doesn't have a high enough resolution for the dot size being used, and inconsistent plate imaging. Again it generally occurs with older first order FM screen designs. Graininess can also occur when screening back a spot color. This is because spot colors are normally not formulated to be halftone screened and hence their pigment grind may be too coarse for the dot size being used.

2 - Inability to align presswork color to an industry standard like SWOP, GRACoL7, etc. The higher the halftone screen frequency, AM/XM or FM, the larger the color gamut becomes (this is explained HERE and HERE). The extra gamut appears in one and two color screen tint builds making them appear more vibrant than the same screen tint build using a lower lpi AM/XM screen (e.g. 175 lpi). It usually only effects vector artwork and typically does not show up in raster images. If this is a problem it can be dealt with using color management and re-separating images to contaminate color builds to reduce the gamut. Unexpected color shifts can also occur when screening back a spot color. This is because spot colors are normally not formulated to be halftone screened and hence the impact of using a high frequency screen, with its increased gamut can lead to an unexpected final color appearance.

3 - Piling. Piling may occur on heatset web presses and is usually caused by not following proper blanket washup schedules or by poor ink rheology - i.e. using inks formulated for the large dots of a coarse AM/XM screen rather than formulated for the smaller FM dots. Piling is rare on sheetfed presses - typically an ik that works with a 200 lpi AM/XM screen will work with a 20-25 micron FM screen.

4 - Mottle. Mottle (splotchy appearance of flat tone areas) can appear with any high frequency screen. It is usually the result poor ink transfer (incorrect ink formulation), fountain solution condition, or paper surface/coating characteristics. Mottle can also occur when screening back a spot color. This is because spot colors are normally not formulated to be halftone screened and hence their pigment grind may be too coarse for the dot size being used.

5 - Poor image reproduction. The higher the lpi of the screen, AM/XM or FM, the higher the detail rendering and hence the better the detail, flaws, and artifacts of an original image are reproduced. Another problem can be a "softness" to the image reproduction. This is usually caused when high dpi images are resampled to a lower resolution when the PDF is created or resampled when the image goes through the RIP. The solution is to not resample images.

6 - General failure. The most common reasons that cause printers to fail with FM and high frequency AM/FM screening that I've encountered are:
1) Poor process control in the shop. Halftone screening in the range of 65-175 lpi is very tolerant of process variations however as the dots get smaller, process controls need to be tighter. That being said, the process controls that need to be in place for FM screening to succeed are no different than those needed for an efficient 175 lpi print manufacturing process. If there are any issues - high frequency screening will make those issues more visible.
2) Lack of buy-in. If the frontline people (e.g. prepress and/or press operators) have not bought into the idea of using FM screening - it will fail. Without buy-in the change merely represents a disruption to what they see as a smooth running process. "Why fix it if it ain't broke?"
3) Lack of use. If FM screening is only used for "special jobs" then press operators never get used to how it performs on press and as a result they will struggle with it. As with most activities, the more you do them the better you get at doing them. It is best if the shop makes the commitment to change to FM and make traditional AM/XM screening the special case exception.

In future posts I will be covering success strategies for implementing FM screening as well as general process control proceedures.

About 80% of all the directory printing in North America is done using FM screening.

Mass customization for print production

Print production is most cost-effective when it is standardized. However a standardized production process doesn’t allow the flexibility to meet unique customer needs since tuning production for each customer quickly increases costs. The solution to this dilemma lies in the manufacturing philosophy of "mass customization."

Mass customization combines the efficiencies of mass production with the flexibility of custom manufacturing and is based on the notion of customizing what’s visible and standardizing what’s invisible to the customer. For example, in the automobile manufacturing industry – Volkswagen, with only four standard platforms, manufactures over 30 different vehicles. Amazingly, the Volkswagen Beetle and Audi TT sit on the same chassis as the Volkswagen Jetta, Golf, Audi A3, Skoda, and SEAT offerings in Europe. The secret is that the parts the customer doesn’t see are standard across the different car models but the parts that the customer sees are unique to each model. Translated into printing terms, this means the printer standardizes on the few, key, presswork characteristics that provide maximum visible customization from the print-buyer perspective. These characteristics are then pre-tested and integrated into the production workflow. When the sales representative discusses the project with the client, one of the standardized (to the printer) packages can then be offered as an alternative “custom-tailored” solution for the buyer.

The key to success is in developing, pre-testing, and standardizing options rather than waiting for a specific customer request or experimenting on live jobs.

Here are just a few thought-starters for the kinds of standards that can be established within a multiple-standards print shop to customize presswork to better meet individual print buyer's needs:

1) 175 lpi GRACoL7. Standard commodity presswork.
2) 175 lpi at Dmaxx. This is presswork run at higher-than-standards solid ink densities (about 30+ points) in order to provide more saturated color.
3) FM/stochastic screening to provide near-contone fidelity and photographic detail.
4) C(2M)YK. This is CMYK with an alternate magenta ink hue. Changing the Magenta ink between two different hues allows favoring intense reds and oranges or blues and purples.
5) Hi-Fi via a ink set that uses more than the four process colors in order to expand the color gamut for better image fidelity or to replace spot colors through more effective screen tint builds.

Jim Rimmer - April 1, 1934 - January 8, 2010

Jim Rimmer, born April 1, 1934 died of throat cancer January 8, 2010.

He was a type designer, typefounder, printer, teacher, publisher, illustrator, mentor and friend to many. He was born and raised in Vancouver, British Columbia, Canada and attended Vancouver Technical School. During his freelance years he worked on projects for the major agencies and design studios in Vancouver, for corporations, airlines, mining and forestry companies. A large part of his work entailed letter design and lettering projects.
Rimmer was noted as the proprietor of the Pie Tree Press, located in New Westminster, a letterpress shop for which he designed many typefaces in metal, including Albertan, Kaatskill and Stern.
He had several presses, including a very large Colt’s Armory. He also had a complete Monotype setup, which let him cast individual letters for handsetting and complete pages of text when driven by punched paper tapes. But the most unusual thing he had was a pair of pantograph machines, which allowed him to engrave matrices for making new type faces.
He was an inspiration to so very many.
Some of his fonts are available HERE

RE:Print - "Job Security"

Imaging technology oddities

Just some miscellaneous imaging technology oddities lying around my desktop that I thought were interesting and didn't fit anywhere else.

Click on images to enlarge
Eugene Omar Goldbeck was a photographer who specialized in large scale group portrait, events, and scenes.
Shield made up of 30,000 troops. 1918
Portrait of Woodrow Wilson made up of 21,000 troops. 1918
Insignia made up of 5,100 troops. 1918

Christian Faur is an artist based in Granville, Ohio, who creates pictures out of crayons by packing thousands of them together so they become like the colored pixels on a TV screen.
He starts each work by scanning a photo into a computer and breaking the image down into colored blocks He then draws a grid that shows him exactly where to place each crayon The finished artworks are packed tightly into wooden frames. He actually makes the crayons himself, hand-casting each one in a mould.


Daniel Rozin is an artist who uses interactive art that allows the viewer to create the image. The image is made using a variety of moveable elements controlled by software and servo-motors.
This example is "Wooden Mirror" from 1999 and uses 830 square pieces of wood acting like digital pixels to deflect light and create different tones.

Press play arrow to view video.

Shapeways creates customizable 3D mini branding irons that click onto standard disposable lighters. You simply turn the lighter on for 30 seconds until the branding iron is hot and then brand away. They can create text or logo graphic branding irons.


Oleksiy Pikalo's latte printer is quite sweet.

ESCP Europe is a top ranked international business school. It has campuses in five major European cities: Paris, London, Madrid, Berlin, and Turin.
This video was made by the students using bodies and clothing to create images.

iPhone Prepress - Australia style

Dwane Hollands of Hollands Print Solutions Australia (http://www.hollands.com.au/) has what is likely one of the most unique workflows in the business – he can manage his prepress via iPhone.
At his current imaging speed he can do 12 plates an hour with two plate cassettes dedicated to his four colour Sakurai A2 machine. That equates to about 120 plates (60 plates per cassette) and, all things going well, he could drive a 10 hour, lights-out shift until his CtP ran out of plates and required human intervention.

However, his concern was about what would happen if something goes wrong and plate output stops during the night shift after the last person leaves the building. So he came up with a way to monitor the CtP Controller while he is off site. Since Kodak installs a VNC server on the Controller so they can dial in and operate the machine for remote servicing, Dwane took the same route and downloaded a free VNC client for his iPhone. At first he connected via Hollands' internal WiFi network and was able to log onto the VNC Server for the Magnus CtP controller. He then took it up a notch by implementing a VPN (Virtual Private Network) connection. With that setup he can click on windows and use menus. Zoom in using the pinch out and pan with his finger. He has also connected via the 3G network using his VPN and was again able to view and interact with the Magnus controller. The beauty of the VPN method is that he can connect to any VNC machine on the network - including his Matchprint and Prinergy Servers. On Leopard Macs, such as his iMac, he's able to simply switch on "Screen Sharing" to enable remote control, so it's cross platform as well.

While there are some teething issues (it disconnects after 6 or 7 minutes) he thinks the likely cause is his lack of VPN expertise or the free software he is using. That being said, as it currently operates, it's completely effective for the company's needs as it only takes about 30 seconds to see if the machine is still outputting plates, or if it requires a shop visit to fix the problem. Ironically enough, at the moment the only thing stopping Hollands from running a lights out night shift is the fact that someone lives next to their factory and the compressor for the CtP is quite loud and runs frequently. Even so, now there's nothing stopping Dwane from working 24/7 from anywhere he finds himself.

Heidelberg's Color Atlas

Heidelberg has recently published their own color atlas titled: "Process Colour Look Up Guide" - and it fills a much needed void.
Now, color atlases have been around for quite some time (see my post HERE) however their useability has suffered in many ways due to a lack of adherence to any effective print specification. Heidelberg has solved that problem by producing their atlas according to ISO 12647-2 Amd. 1:2007 printed, in the UK, on both coated and uncoated paper. That should mean that the color atlas would be valid to specify screen tint builds at any printshop that prints according to the specification.
I strongly suggest you contact your Heidelberg representative and try to secure a copy for yourself - they will likely go quickly.

Moiré

A moiré pattern is an artifact that occurs in the print reproduction process when any two, or more, repeating patterns overlap each other.
Moiré in the print reproduction process is similar to the distortion effect on television when a presenter's clothing includes a striped or crisscross pattern as in the gentleman's shirt in the short video below from the WhatTheyThink.com website:
Click image to play
It that case, the presenter's stripe patterned shirt is "harmonically beating" i.e. has a similar frequency and angle to the video camera's sensor and/or the pixels on the computer's display. This results in the appearance of a secondary pattern or "moiré."

The most common types of moiré encountered in the print production process.

Scanning/sampling moiré
These artifacts are caused by the frequency/angle of the scanner sensor (flat bed and drum scanners, or digital camera sensor) harmonically beating with a pattern in the object being scanned. In this case the artwork ends up having the moiré embedded in it and is part of the image. For example, the original pattern in the pinstriped shirt below (left) acquires a moiré pattern when scanned (right).
If you encounter this type of moiré, there is a Photoshop technique developed by John Wheeler that may help you eliminate it. The tutorial is here: http://tinyurl.com/3mmzv4h

Moiré can also be introduced when a halftone printed image is scanned. In the picture below, the top image is how the photograph in a magazine appears to the eye and below it the result when the image is scanned.
In this case, the moiré is caused by the halftone dots in the magazine reproduction harmonically beating with the scanner's CCD array.

Subject moiré
These artifacts are caused when the halftone screen that is being used to reproduce the image on press harmonically beats with a pattern in the image being reproduced as in the example of the striped shirt below:This artifact is sometimes referred to as "screening moiré" since it is the halftone screen that is causing the problem. However, I use the term "screening moiré" to refer to a different problem - see below.

Screening moiré
Screening moiré, which is a term that is sometimes confused with subject moiré, is actually an artifact caused by either an inappropriate or incorrect halftone screen angle within a CMYK image. With modern screening systems this is rarely a problem. What is most likely to happen is that a screening moiré that is already present is somehow made more visible. For example, the image at left below is a blow-up of a screen tint made of Yellow and Cyan. Because the Y and C screen angles are less than 30° apart they create a moiré. However, because Yellow is so much lighter than Cyan the moiré is not normally visible.However, if the Yellow printer becomes contaminated, as at right above, the existing moiré can become very visible.

Another cause of screening moiré can occur if a prescreened (bitmap) graphic is imaged on a device that has a different resolution than the original art. In the below example a prescreened image that was created at 2400 dpi (standard for North American imagesetters) has been imaged on a 2540 dpi device (standard for the rest of the world):The result is a severe moiré in what should be a flat background screen tone.

Resampling moiré
Moiré artifacts can be introduced when images are resampled (have their resolution changed) somewhere in the production process.
Original resolution car grill at left. Resampled car grill at right.
Moiré, caused by resampling, usually occurs if the image is resized in a page layout program, or when the document is exported as a PDF, or as a result of the RIP settings when the document is processed by prepress.

Obscure moirés
These moirés are fairly rare, but do happen. When other explanations fail, these causes may be worth investigating.

Demosaicing moiré
On rare occasions you may encounter a "demosaicing" moiré. These occur when images with small-scale detail near the resolution limit of the digital sensor in a digital camera sometimes cause the demosaicing algorithm to produce repeating patterns, color artifacts or pixels arranged in an unrealistic maze-like pattern. On the left a properly demosaiced image, and on the right one in which the demosaicing algorithm has caused colored moiré artifacts in the fence and side of the building:Click image to enlarge
Single channel moiré
The standard photomechanical screen angles do not work best with digital screens. As a result some output device, halftone dot shape, screen angle and frequency combinations can result in moiré within one screen resulting in "single channel moiré."
One solution to avoid this problem was the development of shifted angles. The angular distance between screen angles remains more or less the same however all the angles are shifted by 7.5°. This has the effect of adding "noise" to the halftone screen and hence eliminating the moiré. For that reason, some individual screen sets may vary the requested screen angles slightly in order to overcome the potential for single channel moiré.

Paper related moiré
During the paper manufacturing process the side of a sheet paper that is in contact with the wire or forming fabric of the paper machine is the wire side (also called the reverse or bottom side). The wire side is usually not quite as smooth as the top or felt side and may carry a subtle impression of the wire pattern. If that pattern harmonically beats with the halftone screen pattern a subtle moiré will appear in the presswork. It often appears, and is confused, as a mottle. The difference is that mottling appears as random splotches while wire side paper related moiré appears as splotches that form a periodic pattern.

Avoiding moiré
One of the unfortunate effects of the use of inkjet proofing is that moiré artifacts are no longer detected during proofing cycles where there is an opportunity to deal with them. Instead, they are usually first seen on press at which point the job may need to be stopped resulting in increased costs and schedule disruption.

While the geometries of moiré formation are well understood, I'm not aware of any prepress system that incorporates moiré detection/prediction during document processing. So, the key thing is that print specifiers and prepress technicians have to take responsibility to reduce the likelihood of moirés occurring in the first place as well as being aware of image types that have the potential to form moiré artifacts.

Tips for avoiding scanning/sampling moiré
1- Use descreening software if your scanner application has this option.
2- Try scanning at a resolution equal to the halftone lpi used for the printed image.
Left image scanned at 600 dpi and rescreened. Right image scanned at 150 dpi (same as printed) ready for rescreening.
3- Try scanning the image by placing and scanning the original at different angles.
Left image scanned at 90° shows moiré on face. Right image scanned at 30° - no moiré on face.
Tips for avoiding subject moiré
1- Identify images with the potential for moiré. If the prepress workflow allows it, export the processed halftoned bitmap files of the suspect images. Proof them by viewing at 100% on a monitor or outputting to a laser printer at a magnification equal to the resolution of the printer. (e.g. if the bitmap is 2400 dpi, and the laser printer is 600 dpi, then output the bitmap at 400% (2400/600 = 4)). View the proofs on screen or on paper from a distance to see if a moiré is present.
2- Use FM/stochastic screening. Because this type of screening has no frequency or angle it avoids subject moiré completely.
3- Use FM/stochastic screening for the screen that is causing the subject moiré - typically it's the black printer.
4- Swap screen angles usually the black for magenta.
5- Change separation method to UCR instead of GCR.

Tips for avoiding screening moiré
1- Use FM/stochastic screening for the yellow printer. If you're using a 150-200 lpi AM/XM screen then use a 35 micron FM/stochastic since it will have a similar dot gain curve.
2- Make sure that incoming halftone screened bitmap files have a resolution that is equal to or an even divisor of the resolution of the output device. Make sure that those bitmapped images have not been resized in a page layout application.

Tips for avoiding resampling moiré
1- Import images into page layout applications at 100% - do not resize in the application.
2- Images should have a resolution that is an equal divisor of the output device. E.g. 300/400/600 dpi are even divisors of 2,400 dpi.
3- Make sure that PDF creation applications are set to not resample images.
PDF creation settings
4- Make sure that prepress RIP settings are set to not resample images.