Resolution for 175 line screen | Ledet Dan Margulis Photoshop Training

Dan Margulis Applied Color Theory

Resolution for 175 lpi Printing

pixel resolution for 175 lpi screen printing

Posted by: "lauphotomn"

Thu Feb�19,�2009 9:58�am (PST)

Good morning,

My wife and I photograph orchids for a calendar that is sold nationally by Borders Book Stores here in the USA. We just finished the photography for the 2010 calendar. The past three years we have always provided digital files at 300 ppi at final crop with bleed. The calendars print nicely, colors are great. This year we approached the project as we have in the past only to find out that they want a resolution of 350 ppi because the screen that will be used for printing on press will be 175 lpi or greater. In order to supply them with the requested resolution interpolation will be required to produce the final images.

I've come to believe through information obtained from associates, who have been closely involved in the separation and printing of their photographic work, that these requested resolutions are bloated and that a lot of information is discarded before ink hits paper. My fear is that because I need to create pixels that didn't exist and they and others will be thrown away that our quality will be less then it could have been. How do I find out if the press can even print with 175 lpi screen and truly how much ppi does it really need.

I hope that this isn't off topic or discussed before (I'm not sure how to use the archives), but this list has great minds on it and there is no better place to to ask this question.

Thanks in advance.

John Lauenstein

___________________________________________________________________________

Re: pixel resolution for 175 lpi screen printing

Posted by: "Jim Bean"

Thu Feb�19,�2009 10:37�am (PST)

I believe that your fears are unnecessary..... I would wager that the sharpening techniques they utilize will have far more impact... 300 or a resampled 350 image of a flower will likely print identically....

jim bean

___________________________________________________________________________

Re: pixel resolution for 175 lpi screen printing

Posted by: Henry Davis

Thu Feb�19,�2009 10:41�am (PST)

You may be left with trusting that the printing is done at 175lpi. You could ask for samples and measure for yourself if you wanted to verify.

350ppi is the absolute mathematical maximum resolution that can be utilized by 175lpi screening. It is a mathematical "law" so to speak. But, there is nothing implied in the math that a particular 350ppi image will *necessarily* look better than a 300ppi when printed. The printed results will depend more on the particular image and how it has been prepared.

If it were me, I would not upsample the images, but if I did, I would do the upsampling prior to sharpening. I would avoid upsampling the sharpened versions if those were the only versions available - I would just go with what I had.

Henry Davis

___________________________________________________________________________

Re: pixel resolution for 175 lpi screen printing

Posted by: J Walton

Thu Feb�19,�2009 11:44�am (PST)

On Thu, Feb 19, 2009 at 10:28 AM, Henry Davis wrote:

350ppi is the absolute mathematical maximum resolution that can be

utilized by 175lpi screening. It is a mathematical "law" so to speak.

That's another one of those urban legends that people repeat because they believe what they were told. Normally when this point gets challenged we are presented with a bunch of scientific data regarding radio waves. The reality is, there are advantages to having more than 2x the linescreen in some circumstances. I'm sure a search of our archives will reveal a few threads to that effect.

"I think we ought always to entertain our opinions with some measure of doubt. I shouldn't wish people dogmatically to believe any philosophy, not even mine." - Bertrand Russell

J Walton

___________________________________________________________________________

Re: pixel resolution for 175 lpi screen printing

Posted by: "lauphotomn" l

Thu Feb�19,�2009 11:44�am (PST)

--- "Jim Bean" wrote:

I believe that your fears are unnecessary..... I would wager that the sharpening techniques

they utilize will have far more impact... 300 or a resampled 350 image of a flower will likely

print identically....

Thanks for the reassurance, the up-sample is somewhere around 16 % if that, so my fears are probably unfounded.

John Lauenstein

___________________________________________________________________________

Re: pixel resolution for 175 lpi screen printing

Posted by: Henry Davis

Thu Feb�19,�2009 1:03�pm (PST)

On Feb 19, 2009, at 2:32 PM, J Walton wrote:

That's another one of those urban legends that people repeat because

they believe what they were told. Normally when this point gets

challenged we are presented with a bunch of scientific data regarding

radio waves. The reality is, there are advantages to having more than

2x the linescreen in some circumstances. I'm sure a search of our

archives will reveal a few threads to that effect.

Could you offer some concrete example of how the signal/noise problem is overcome by having greater than 2x the lpi?

Henry Davis

___________________________________________________________________________

Re: pixel resolution for 175 lpi screen printing

Posted by: "lauphotomn"

Thu Feb�19,�2009 1:03�pm (PST)

Henry Davis wrote:

You may be left with trusting that the printing is done at 175lpi.

You could ask for samples and measure for yourself if you wanted to

verify.

I wish I could request samples. The actual production is handled by a publishing house and printed (unfortunately)in Korea.

350ppi is the absolute mathematical maximum resolution that can be

utilized by 175lpi screening. It is a mathematical "law" so to

speak. But, there is nothing implied in the math that a particular

350ppi image will *necessarily* look better than a 300ppi when

printed. The printed results will depend more on the particular

image and how it has been prepared.

Good info, but what is the minimum resolution that would result in a quality reproduction of the image? I know that this is a loaded question but for information sake lets say the image is near perfect, little or no post capture work is needed.

If it were me, I would not upsample the images, but if I did, I would

do the upsampling prior to sharpening. I would avoid upsampling the

sharpened versions if those were the only versions available - I

would just go with what I had.

My clients want to see that 350 ppi resolution in the file. Since the images are in Raw format I reassigned the resolution of 350 ppi and upsampled 108% to get to size in the Raw converter. No sharping. For comparison I did a version to upsample in photoshop using bicubic smoother and I couldn't see any difference when viewed at 100%.

I guess my question is do you really need twice the resolution in pixels for any given line screen to print a quality image at a given size?

Thanks for the reply Henery

John Lauenstein

___________________________________________________________________________ .

Re: pixel resolution for 175 lpi screen printing

Posted by: Henry Davis

Thu Feb�19,�2009 2:30�pm (PST)

If the images are 300ppi at print size then there's not much to sweat over about how low you could go - you have plenty enough resolution for good reproduction already. Since you understand that it's image dependent, then you understand that only experience will help gain judgement. For instance, what is the least amount of resolution you could get by with for a London Bridge scene on a very foggy day? One can only guess about this, but it's probably very low, like about .75 x lpi perhaps less. Whether printing grey or color could make a difference as well.

One doesn't necessarily need 2x the lpi to result in a quality image. If an image is highly contrast dependent and/or contains small, contrast dependent subject matter, then 2x the lpi is a solid rule to shoot for. For a product shot where the subject contains lines or type, or small hard edges etc. - 2x lpi would be a good rule of thumb.

You are very fortunate to have the Camera Raw version. There is probably some debate about sampling and sharpening as to which is best to use - the raw converter or USM Photoshop. I'll leave this part for others to argue, but I would be pleased to sharpen with either, provided it is done at the final sampling.

Henry Davis

___________________________________________________________________________

Re: pixel resolution for 175 lpi screen printing

Posted by: "RJay Hansen"

Thu Feb�19,�2009 2:30�pm (PST)

On Thu, Feb 19, 2009 at 1:46 PM, lauphotom wrote:

I guess my question is do you really need twice the resolution in pixels for

any given line screen to print a quality image at a given size?

No.

rjay

___________________________________________________________________________

Re: pixel resolution for 175 lpi screen printing

Posted by: "George Machen"

Thu Feb�19,�2009 2:30�pm (PST)

I've often heard that high line screen printing benefits from images whose dpi is more than twice the line screen, but never been given a clear reason why. One thing's for sure, that higher resolution will result in *softer* printed results, less detail - necessitating more sharpening to (try to) compensate. Unless the softer look is precisely the circumstance J Walton is talking about. (But there are other ways to get that softer look, no?) What else could the extra resolution contribute?

My understanding is that a one-to-one correspondence between image pixel resolution and halftone dot or rosette resolution (line screen) is sufficient to *fully* capture & reproduce an image's information, in principle, all else remaining equal.

The only reason I've ever heard (not involving hand-waving) where all else doesn't remain equal, to legitimately add more image resolution, is to compensate for halftone *screen angles* - the maximum of which would be to help suppress visibility of the most prominent K plate at 45 degrees, which Pythagorus works out at the square root of two, or an image resolution of about only 1.4 times the halftone screen. The industry standard of increasing that amount to double the line screen only was to allow a little headroom for upscaling the image after-the-fact.

And all that Nyquist sound theory analogy is inapplicable to printing because I don't think there's a correlate to "bandwidth" in an applicable sense. (And it's not Nyquist, anyway--it's Shannon.)

- George Machen

___________________________________________________________________________

Re: pixel resolution for 175 lpi screen printing

Posted by: J Walton

Thu Feb�19,�2009 2:30�pm (PST)

On Thu, Feb 19, 2009 at 11:45 AM, Henry Davis wrote:

Could you offer some concrete example of how the signal/noise problem

is overcome by having greater than 2x the lpi?

Sure. It happens with small images that contain fine detail. Any time something is being described by only one pixel, it will inevitably raster on output. It will also look bad on-screen. We ran into that a lot before we required a MINIMUM scan size (in megabytes) from our scanning department. They were going off of the 2x lpi rule and we were having a lot of problems with rastering. Once we specified that all scans needed to be at least 10 or 15mb, that problem went away. It was then that I realized that the 2x lpi thing was just a rule of thumb, not a rule of law.

Think about the hard-and-fast rule you described earlier. Doesn't it seems a bit over-simplified? There is so much that goes into screening an image to plate...

1. The type of screening used - diamond, elliptical, square, round-square, dot-centered rosette, hole-centered rosette. Does NONE of that have an effect on fidelity? It's always 2x?

2. The resolution of the image/plate setter - an image screened at 1200 dpi will not look as good as an image at 2400 dpi. But THAT's not factored into this hard-and-fast rule?

3. The various algorithms used to convert 8-bit data to 1-bit. Some companies have better technology than others. Obviously you'd rather see a 1-bit tiff made by Screen or Heidelberg rather than a 1-bit made by Adobe or Xinet.

4. The quality of the image - is there grain, compression?

5. The content of the image - this is huge. An image of clouds does not need a lot of resolution. Something with fine detail (think a close-up shot of wind-blown hair, or a feather, or stripes).

6. The resolution of the image - here's the factor people ignore because they are stuck on some rule they NEVER tested themselves, they just believed what someone else said. Of course, that someone else has never tested it either...

I remember someone making an argument about the 2x frequency/detail issue. The conclusion they came to was that this would apply to the resolution of the image setter, not the lines per inch of the halftone. So you could never resolve anything more than 2400dpi if your imagesetter resolution was 1200dpi. Now, that's not my theory. I couldn't care less about that stuff. All I know is 2x lpi is a pretty good principle to apply, as long as you're smart enough to know when it doesn't fit.

J Walton

___________________________________________________________________________

Re: pixel resolution for 175 lpi screen printing

Posted by: "mac townsend"

Thu Feb�19,�2009 2:30�pm (PST)

Good info, but what is the minimum resolution that would result in a

quality reproduction of the image? I know that this is a loaded

question but for information sake lets say the image is near perfect,

little or no post capture work is needed.

there is no one answer that really fits all situations. "It depends" is the only accurate general purpose answer.

It depends first on the nature of the photograph.

to generalize a bit here, I've found images with sharp edges, especially ones at an angle, will want more resolution to avoid the "jaggies". Such as sailboats on the bay, or automotive product photography (the edges here being reflections on the paint), etc. In this case I usually use about 400 ppi@100%. In a few cases I have found that even more helps (especially with real sharp product photography).

Images with softer edges can, I've found, get by with less...most portraits, for instance. 200-300 ppi@100% is usually fine for as much as 175-200 lpi in these cases.

In the original query, orchids, I'd look at the sharpness of the edges and the background...they might well benefit from >300 ppi.

Then there is the stock. hard finish cast coat stocks have such ink hold out that more resolution and rulings can provide a better result, while softer stocks, like uncoated or text, can't carry these higher values and the image will usually look better/cleaner at lower lpi (133 or even 120 instead of 150, and lower resolution images). Unfortunately most shops these days run as high a ruling as they can on most ever stock.

Then the press/pressman. A 30 year old press is not going to hold the same dot that a new new one can hold. And the skill of the pressman also comes into play.

The rip tries to use all the available image, the the point of even charting different zones in a single dot (how depends on the rip itself and it's screening algorithms).

It is my understanding that significantly more image resolution than "needed" tends to "smooth" the image. Dan has referred to this in his books...using one illustration a section of grass. The detail in the grass consists of brownish dirt and shadow, while the overall dominant color is greens. with very high resolution the greens tend to overpower the browns and the image becomes smoother...in effect, loosing the detail that had been provided by the browns.

and what works for a litho press is totally different from what works for a toner or inkjet based process.

Mac Townsend

Adcom Graphics Digital Imaging

Fairfield, California

www.adcomgraphics.com

___________________________________________________________________________

Re: pixel resolution for 175 lpi screen printing

Posted by: Henry Davis

Thu Feb�19,�2009 4:42�pm (PST)

I won't argue that your method is better for you. I just wondered and was asking if you had come across a better explanation.

There is more math involved in this than I can grasp. But I do understand a little about spots and dots and halftone cells, enough to ask myself "is there an upper limit?" - and I believe there is a mathematical upper limit. I don't believe this just because of what I have heard or been told.

Regardless of the shape of the dot, the percentage area of the cell is finite. While there are qualitative differences in the kind of dot used, the area of the cell remains finite. Some imagery will benefit from a particular dot structure, but there is an upper limit nonetheless. For me, Nyquist's theorem is a reasonable explanation even if it's purpose was not for describing halftone printing.

I'm not so sure that the theorem can't address the imagesetter as well. The upper limit lpi is determined by the imagesetter resolution. By what mathematical factors?

I have seen many images ruined by too much resolution. Many of them are small in dimensional size - increasingly so with the use of DLSR's. Full captures sized down to an inch or two visually demonstrate at print time that there is an upper limit.

The 2x lpi rule of thumb may not express the exactitude that we would like, but since it works so well with such a huge population of images, I'll be happy to use it as my upper limit. And, the Nyquist theorem is pretty convincing. It may be simplistic, and I'll be glad to take some grief on that point. But I take exception to the notion that I am merely aping - that would give apes a bad name. Do you always make use of Bertrand Russell's ramblings or did you pull this one out especially for me - or was this quote some sort of Freudian slip?

Henry Davis

___________________________________________________________________________

Re: pixel resolution for 175 lpi screen printing

Posted by: J Walton

Thu Feb�19,�2009 8:47�pm (PST)

On Thu, Feb 19, 2009 at 4:07 PM, Henry wrote:

I won't argue that your method is better for you. I just wondered

and was asking if you had come across a better explanation.

I think I explained that. I came across an exception to the 2x rule, even though I have been using it myself for the last 16 years. 2x works great until you get rastering in fine detail. If you see it, you know that it may need a little more resolution than that. Obviously there are cases where 1.2 or 1.5 look better. You end up getting more "action" in your shoot with less resolution, and sometimes that's good.

There is more math involved in this than I can grasp. But I do

understand a little about spots and dots and halftone cells, enough

to ask myself "is there an upper limit?" - and I believe there is a

mathematical upper limit. I don't believe this just because of what

I have heard or been told.

I believe there is an upper limit as well. I just happen to know that it is not 2x the lpi.

Regardless of the shape of the dot, the percentage area of the cell

is finite. While there are qualitative differences in the kind of

dot used, the area of the cell remains finite. Some imagery will

benefit from a particular dot structure, but there is an upper limit

nonetheless. For me, Nyquist's theorem is a reasonable explanation

even if it's purpose was not for describing halftone printing.

I am not a mathematician, and I have never claimed to be. But in my unprofessional opinion, Nyquist's (or whoever else's) theorem works by coincidence. That doesn't change the fact that 2x lpi is a nice place to start.

I'm not so sure that the theorem can't address the imagesetter as

well. The upper limit lpi is determined by the imagesetter

resolution. By what mathematical factors?

See above. (IOW, I have no idea)

I have seen many images ruined by too much resolution. Many of them

are small in dimensional size - increasingly so with the use of

DLSR's. Full captures sized down to an inch or two visually

demonstrate at print time that there is an upper limit.

I think it demonstrates that you gotta think, and that's all I'm saying.

The 2x lpi rule of thumb may not express the exactitude that we would

like, but since it works so well with such a huge population of

images, I'll be happy to use it as my upper limit.

Change "upper limit" to "upper limit 99% of the time" and we are in complete agreement.

But I take exception to the notion

that I am merely aping - that would give apes a bad name.

I think that's a harsh way to put it. I don't know how it affects apes or their reputations. ;-)

I think it is such a common opinion/feeling/theory/conviction/belief that it is accepted as surely as you accept it when someone tells you not to put your hand in the press while it's printing. And when there's math behind it even the scientists among us may tend to believe it without testing. It's not the easiest thing to test, but you know it when you see it.

Do you always make use of Bertrand Russell's ramblings

I never have before, but this one seemed to fit.

or did you pull this one out especially for me - or was this quote some sort of Freudian slip?

You shouldn't take any of this personally. I've read even less of Freud than I have of Russell. This is simply an issue where the consensus opinion, in my experience, is incorrect. But it is stated with such force that others accept it without argument. If you re-read the part of your original text which I quoted, it was stated in that same forceful manner.

Why not throw in a bonus quote...

"Where there is much desire to learn, there of necessity will be much arguing, much writing, many opinions; for opinions in good men is but knowledge in the making." - John Milton (1608 - 1674)

J Walton

___________________________________________________________________________

Re: pixel resolution for 175 lpi screen printing

Posted by: "David Riecks"

Thu Feb�19,�2009 8:47�pm (PST)

At 01:46 PM 2/19/2009, lauphotomn wrote:

Good info, but what is the minimum resolution that would result in a

quality reproduction of the image? I know that this is a loaded

question but for information sake lets say the image is near

perfect, little or no post capture work is needed.

John:

I recall Brian Lawler talking about this during an early digital seminar (mid 90's) and he'd done a test with the same digital file printed at varying "q" (quality) factors from 1.0 up to 2.5 or so. A version of this was in a book that I have from him if I can locate it. For most 150 line screen reproductions, there was little to no visible difference between the reproductions once you were 1.3~1.5 x the line screen.

This is also corroborated on the resolution page on the Universal Photographic Digital Imaging Guidelines website... "Although many publishers have standardized on 300 ppi for 150-line screens, the actual requirement is 1.3 to 2.0 times the line-screen resolution, so as a practical matter, smaller files can successfully work for a given final size."

http://www.updig.org/guidelines/ph_resolution.html

Obviously there are other factors such as the paper quality, finish and type of image that affect this issue as well.

My clients want to see that 350 ppi resolution in the file. Since

the images are in Raw format I reassigned the resolution of 350 ppi

and upsampled 108% to get to size in the Raw converter. No sharping.

For comparison I did a version to upsample in photoshop using

bicubic smoother and I couldn't see any difference when viewed at 100%.

2 times the line screen being the absolute minimum is an old "conventional wisdom" (and face it, many things that are considered conventional wisdom are neither conventional nor purport to be wisdom.

I guess my question is do you really need twice the resolution in

pixels for any given line screen to print a quality image at a given size?

Short answer, "no."

David

David Riecks (that's "i" before "e", but the "e" is silent)

___________________________________________________________________________

Re[2]: [colortheory] Re: pixel resolution for 175 lpi screen printin

Posted by: Mac Townsend

Fri Feb�20,�2009 3:57�am (PST)

Hello David,

Brian's conclusions were based on his observations at that level of screening technology.

it is very different now.

Brian was dealing with almost nothing in CPU power compared to now. 1:1 maybe worked well then when the Big Kahuna was a 68030 and rational tangent screening was the norm.

It no longer is..

The world is different and Brian has retired.

Your best bet is to create your own test samples and evaluate them. This is the classic way to go, BTW.

Can't do that? Sorry. Once upon a time there was a Service Bureau that would work with such questions. No longer when it is a company that won't permit such "since it is all software".

(I'm a service bureau shutting down in the next few months because "i did it myself" has become the standard)

One reason the Service Bureau was of value.

Best regards,

Mac Townsend

___________________________________________________________________________

Re: pixel resolution for 175 lpi screen printing

Posted by: Stephen Marsh

Fri Feb�20,�2009 3:57�am (PST)

Back in the day, a colour house that I used to work for drum scanned all cosmetic advertising images at Res 16 (160 ppcm or around 406 ppi), instead of the more common Res 12 that was used for editorial images appearing in similar conditions (120 ppcm, approx. 305 ppi). The scans contained female portrait and product shots. Even assuming a Q-factor of x2, this is much higher than is required for a 175 lpi halftone screen.

Sincerely,

Stephen Marsh

___________________________________________________________________________

Re: pixel resolution for 175 lpi screen printing

Posted by: Jacob Rus

Fri Feb�20,�2009 1:32�pm (PST)

As far as I can tell, the main thing that demonstrates is that the interpolation used to down-size the image (perhaps in the printer, or similar) did a poor job. If the image looks good at large size, there's no reason that its extra resolution would *harm* it when shrunk (i.e. as compared to a smaller native image), assuming it's processed correctly.

Cheers,

Jacob Rus

___________________________________________________________________________

Re: pixel resolution for 175 lpi screen printing

Posted by: Henry Davis

Fri Feb�20,�2009 3:52�pm (PST)

On Feb 20, 2009, at 3:06 PM, Jacob Rus wrote:

As far as I can tell, the main thing that demonstrates is that the

interpolation used to down-size the image (perhaps in the printer, or

similar) did a poor job. If the image looks good at large size,

there's no reason that its extra resolution would *harm* it when

shrunk (i.e. as compared to a smaller native image), assuming it's

processed correctly.

I agree to a point. When downsampling from 48MB to 525K, this would call for some special handling. But no matter how good the decimation method, an image undergoing this change is going to take some punishment - somewhere. Do you believe that this can be accomplished without any image quality loss?

For me, this demonstrates that decimating an image to this extreme is not a good practice. Sure, you can work them over to make them work better, but the practice of laying out lots of these monsters on a sheet is becoming more popular. Do you know of a rip that can better a human using Photoshop for this task?

Henry Davis

___________________________________________________________________________

Re: pixel resolution for 175 lpi screen printing

Posted by: J Walton

Fri Feb�20,�2009 6:07�pm (PST)

On Fri, Feb 20, 2009 at 2:16 PM, Henry wrote:

I agree to a point. When downsampling from 48MB to 525K, this would

call for some special handling. But no matter how good the

decimation method, an image undergoing this change is going to take

some punishment - somewhere.

I think what Jacob is getting at is that the image is getting downsized *somehow*. If it looks bad then it was downsized badly.

Some RIPs don't exactly work that way, though. If you have a 300 dpi image and you size it in InDesign to be 25%, you have a 1200 dpi image going to the RIP. And some RIPs just send that data on as-is. Back in the day Scitex RIPs would not do that, they would size things down to res 8, 12, 16... whatever you set for that queue.

Let's imagine you have a 48MB capture from a high-quality digital back. You might be using that for a few different purposes: outdoor, business card, web site. You obviously don't need the outdoor file size for use as a business card. So what do you do? You res it down! Shooting the image twice, and reducing the file size in-camera makes no sense. You res it down, and now the question is do you res it down to 2x lpi or 1.4 or 1.2 or whatever. Photoshop does a nice job of downsizing images, and I'd say the lion's share of retouchers will use bicubic or bicubic sharper and just let it fly.

If you don't res it down in Photoshop, you are leaving it up to the RIP. If the RIP (or an OPI solution in the middle) is set to down-size images, and that is *not* uncommon, then the question is really... How well does it do compared to Photoshop? You don't have any manual control over it, but in my experience most retouchers don't manually adjust res-downs anyway.

Do you know of a rip that can better

a human using Photoshop for this task?

I guess it depends on the RIP and the human. :-)

If you are in a high-volume environment (think hundreds or thousands of images per day), then a RIP can better a human in that the company goes out of business if the dedicate a person (or team of persons) to fine-tuning down-res's. If you are in a low-volume environment (think a few images a day, or a few images a week) then putting some attention to each image will almost always result in better quality. The question then is... how much better?

J Walton

___________________________________________________________________________

Re: pixel resolution for 175 lpi screen printing

Posted by: Jacob Rus

Fri Feb�20,�2009 8:48�pm (PST)

Quality loss compared to what? The implication of the original email as I understood it was that having a lower-resolution original file might give better quality, all else equal, because the higher resolution file must be downsampled and that "causes punishment". But that's silly. If the high-resolution file looks good at high resolution, then it should just as good as the low-resolution file, assuming that a reasonable interpolation method is used.

In other words, there's no reason to worry about having an original image with "too much" resolution, unless that comes along with dramatically expanded noise, reduced dynamic range, or similar. At the worst, it just takes up extra disk space.

If your high-resolution files are being printed fuzzily, that implies that the interpolation used (in the printer, or by the retoucher, or whatever) is performing poorly, and should be swapped out for a different algorithm.

Cheers,

Jacob Rus

___________________________________________________________________________

Image resolution

Posted by: "Jim Rich"

Sat Feb�21,�2009 8:19�am (PST)

Here is my .02.

Back in the day, say mid 1990's, I (and a partner) ran a series of tests related to image resolution and screening. We looked at ratios of pixels to halftone screen rulings (AM screen) such as 1:1, 1.4:1, 1.5:1, 1.75:1 and 2:1. One thing we found was that image content was a factor of this input image resolution. For example a resolution ratio of 1.4:1 might work for a dark looking image (lowkey) but not for a light looking image (highkey). This was reported at a Seybold event.

The conclusion from those series of tests, illustrated to me and others that, by doing using lower res images can break a workflow because of many things that happen in publishing. When changes take place you have to stop to go back and fix one or many images to re-ress them. The conventional wisdom from that experience was to use an image resolution in the 2:1 neighbor hood. Since technology has improved by leaps-and-bounds since then, it is cheaper and faster, it is often wiser to use slightly higher res file than lower res files.

Jim Rich

___________________________________________________________________________

Re: pixel resolution for 175 lpi screen printing

Posted by: Dan Margulis

Sat Feb�21,�2009 12:43�pm (PST)

Stephen Marsh writes,

Back in the day, a colour house that I used to work for drum scanned

all cosmetic advertising images at Res 16 (160 ppcm or around 406

ppi), instead of the more common Res 12 that was used for editorial

images appearing in similar conditions (120 ppcm, approx. 305 ppi).

The scans contained female portrait and product shots. Even assuming a

Q-factor of x2, this is much higher than is required for a 175 lpi

halftone screen.

This was also standard practice in the New York market I was working in. I've showed several examples in early books of the impact of scanning resolution on quality, mostly to refute the argument that you can never have too much resolution. Wildly excessive resolutions (meaning 4x or more of screen ruling) did in fact harm quality by yielding overly soft images. In pictures of young models this softness was considered desirable even though in most other categories it would not be.

The old saw about 2.0x resolution being best is much too conservative, particularly today. The exception is cases where there is some possibility that we may have to resize the image at the last minute and wish to take out some insurance.

Resolution needs do not increase proportionally to an increase in screen ruling. A 175-line screen does *not* require twice as much resolution as the minimum needed for quality work at 85 lines.

Digital captures do not require as much resolution as scanned film did. There isn't as much capture noise and there's no film grain to exaggerate. So, today, I am completely comfortable with 250 pixels per inch for 175-line work.

With the advent in recent years of digital cameras with ridiculously high claimed resolutions, I brought up the known problem of excessive resolution in scanning, and suggested that perhaps photographers should shoot certain subjects from further away, to reduce the effective resolution of the capture. During the preparation of PP5E, three professional photographers shot a series of exposures to help figure out whether there was anything to such a theory.

What we learned is that any such effect is dwarfed by the many variables in setting up cameras at different distances. It looked like the effect was present to some degree, but not serious enough to warrant recommending any change in procedure.

Dan Margulis

___________________________________________________________________________

Re: pixel resolution for 175 lpi screen printing

Posted by: Stephen Marsh

Sat Feb�21,�2009 5:27�pm (PST)

Dan Margulis wrote:

This was also standard practice in the New York market I was working

in. I've showed several examples in early books of the impact of

scanning resolution on quality, mostly to refute the argument that you

can never have too much resolution. Wildly excessive resolutions

(meaning 4x or more of screen ruling) did in fact harm quality by

yielding overly soft images. In pictures of young models this softness

was considered desirable even though in most other categories it would

not be.

Dan, I am wondering how modern software, hardware and workflows changed the "rule of thumb" developed in the previous 25 years. You go into some of this below, as well as in your Professional Photoshop books.

As many here know, when drum scans ruled, larger scanning operations often split the scanning process into two sections - preplanning and scanning. The preplanner would mount/unmount the originals to the drums and work out the file naming and print size/resolution for each scan, making the scanner operator more productive (keep that drum spinning!).

The approximate output size of the image was generally known before scanning, otherwise it was often assumed to be +/- 20% in size from the original scan size/resolution. In OPI situations, layout artists were supplied with a low resolution FPO duplicate of the high resolution scan. Artists were instructed not to crop the file, however it could be enlarged or reduced 20% in size in the layout (otherwise a rescan would be required). Scans were sharpened for output at the intended scan/print size, therefore greatly enlarging or reducing the image and sharpening halos would often lead to less than optimal results, as well as the image having too little or too much resolution for the intended halftone screen.

Anyway, that was scanning, where one could control the the resolution and size of the scan. With digital capture, the modern workflow has changed. The dog used to wag the tail, now the tail appears to wag the dog!

Today, many designers and layout artists seem to place native (sensor) resolution files into the layout software and to then reduce the scale size in the layout, rather than resizing a dupe in Photoshop and applying appropriate output sharpening. The result is often a very high resolution origianl that is unsharpened, being scaled in layout to say 20% size or lower (or at the other extreme, enlarged 300%). Additionally a rotation or other transform may also be applied in layout. Then repeat the process another twenty times for the rest of the page in the brochure - for each and every page!

Although many disliked the harsh sharpening often applied to many drum scans, at least the scan was at the correct print size/resolution for output, with output sharpening applied. Images had a crisp look. Today, I commonly see blurry mush in print, original unsharpened images greatly reduced in scale in layout software. Perhaps some images were sharpened, however they have now been reduced in size in layout (greatly increasing resolution) and the sharpening halos have now lost all of their subtle impact.

Above and beyond these factors, I think that the RIP is going to play a critical role in the quality of the final output, as discussed in this thread there are many variables going on in both the front and back end of the process. Even though RIP X and RIP Y may both use the "same" officially licensed Adobe PostScript Level 3, sending the same file to each RIP can lead to different output. This can be the case with two slightly different RIPs from the same vendor, not to mention different vendors with different screening methods.

Resolution needs do not increase proportionally to an increase in

screen ruling. A 175-line screen does *not* require twice as much

resolution as the minimum needed for quality work at 85 lines.

Although I take many vendor claims with a grain of salt, back in the 1990's AGFA produced an excellent range of educational and reference material (even if they did use AGFA technology in all their examples). Anyway, their "conventional wisdom" was that a Q-Factor of x2 was optimal for scans below 133 lpi, while a QF of 1.5 was appropriate for images over 133 lpi (due to diminishing returns with higher screen rulings, coarser screens being more demanding in this respect). The print samples for a 175 lpi screen used 263 ppi and did not appear to suffer for being below x2 QF.

Many images that are "ideal" for 150 lpi at a x2 QF are output at a 175 lpi screen. All other things being equal, the variables that appear to matter the most are image content, sharpening and the imagesetter RIP.

Stephen Marsh

___________________________________________________________________________

Re: pixel resolution for 175 lpi screen printing

Posted by: Henry Davis

Mon Feb�23,�2009 11:38�am (PST)

On Feb 20, 2009, at 10:19 PM, Jacob Rus wrote:

In other words, there's no reason to worry about having an original

image with "too much" resolution, unless that comes along with

dramatically expanded noise, reduced dynamic range, or similar. At

the worst, it just takes up extra disk space.

Forgetting problems of noise, dynamic range and other exposure issues, let's stick with resolution alone and consider this statement:

there's no reason to worry about having an original

image with "too much" resolution"

I know that there is such a thing as too much resolution at print size and it impacts quality, disc space and time. Both mathematical and experiential evidence lead me to conclude that there is an upper limit. Our opinions differ on this. C'est la vie.

Henry Davis

___________________________________________________________________________ .

Re: pixel resolution for 175 lpi screen printing

Posted by: Michael Jahn

Mon Feb�23,�2009 1:10�pm (PST)

yes - too much resolution is easy to understand.

imagine that I have a 500 words in an image at 600 ppi that is on a 4x4 inch square.

now imagine that I place that image into an InDesign page, and reduce it so that it now fits into a .25 x.25 inch square

even if I printed this at 300 lines per inch, there is not enough resolution when making the plate that you could even read the words on the plate.

you will not be able to print it and read the words.

so, yes, this would be a waste of time - as much of the detail required would be downsampled in the RIP

hope this helps.

Michael Jahn

___________________________________________________________________________

Re: pixel resolution for 175 lpi screen printing

Posted by: Mihai Necsulescu

Tue Feb�24,�2009 4:39�am (PST)

Hi Michael,

imagine that I have a 500 words in an image at 600 ppi that is on a 4x4 inch square.

now imagine that I place that image into an InDesign page, and reduce

it so that it now fits into a .25 x.25 inch square

even if I printed this at 300 lines per inch, there is not enough

resolution when making the plate that you could even read the words on

the plate. you will not be able to print it and read the words.

I printed whole Holy Bible, Old and New Testament in English language on a single sheet A4 1 side paper with my CTP in 2000 on a 10 years old printing machine.

At 1500 dpi only.

And yes, you could read the words of wisdom with a 10X magnifier glass on both paper and plate.

Don't mix B&W 1 bit images with color images!

And according to my experience as a prepress service bureau provider (still exist in Romania but ...) we have done films at 150 and 175 lpi with 1200 dpi and we make plates up to 420 lpi with 1500 dpi only. The secret stays in RIP and accuracy of exposure engine. Then come the calibration and keeping it constant.

Regarding image resolution for 175 lpi we use 225 dpi for all work except images where very fine details or text appear.

There we go to 300 or even 400 dpi.

Image is at scale 100% in page layout program.

Best results is after converting to CMYK the image you apply an Unshap filter with 150, 0.5,0.5 or 150, 1,0 or more when examining the image at 66% on the screen in Photoshop if the image respect the resolution stated above. Too much resolution will slow the printing, file transfer, ripping and storage of the files.

But most important will make your printed image soft. If you are working with a human face then you don't want to be visible the skin pores but otherwise you need images that stand out from the printed matter.

And we have worked during the time with more then 100 printing houses and it

works great.

Best regards,

Mihai Necsulescu

Zoom-Soft srl

Romania

___________________________________________________________________________

Re: pixel resolution for 175 lpi screen printing

Posted by: "George Machen" gmachen@triad.rr.com � i_love_gwyneth_paltrow

Tue Feb�24,�2009 7:24�pm (PST)

This is interesting:

http: //www.xs4all.nl/~bvdwolf/main/foto/down_sample/down_sample.htm

- George Machen

___________________________________________________________________________

Re: pixel resolution for 175 lpi screen printing

Posted by: Stephen Marsh

Tue Feb�24,�2009 11:45�pm (PST)

Interesting, however non CG images can lead to different results, so it will often be a case of comparing linear resampling to gamma encoded resampling to see which is preferred for the image at hand (there can often be little difference in results). Similar to sharpening in some respects (linear or gamma).

More here:

www.4p8.com/eric.brasseur/gamma.html

www.openphotographyforums.com/forums/showthread.php?t= 4624

www.glennchan.info/articles/technical/resampling/linearLight.htm

www.teamten.com/lawrence/graphics/gamma/

Regards,

Stephen Marsh

___________________________________________________________________________

Brian Lawler retired?

Posted by: "pt210w"

Tue Feb�24,�2009 12:25�pm (PST)

Brian was dealing with almost nothing in CPU power compared to now.