Opinions please: Lightjet at 200 dpi?

Today I went to a local lab to have a Lightjet print made from one

of my better 4x5 stills. It was originally a 4000 dpi drum scan I

had reduced to 3000 dpi. I plan to print it at 20x24. When I got

there they told me they wanted the file the actual print size and

200 dpi. I could not believe 200 dpi. That works out to scanning a

4x5 at 1000 dpi at film for a 20x24. Also they were going to charge

me an extra $18 to resize it.

 

At any rate 200 dpi works out to a little less than 4lp/mm roughly

in print, but my neg is carrying a lot more detail at that level,

like 6-8 lp/mm or very very detailed.

 

I asked why only 200 dpi, and they said thats where it works best.

 

I dont buy it. Above 200 dpi probably slows their workflow too much.

 

I looked at one of their 20x24 4x5 prints on a wall and it was okay,

but it did not snap with detail. I looked at it with a loupe and you

could see what looked like scan lines too, but only under

magnification. Actually it looked like a digital print that had been

enlarged almost too much. It was just not crisp, but that could have

been the camera too.

 

Hmmmm

 

I then called a local photo nut friend and mentioned the 200 dpi

thing to him and he said it did not sound right, that all the

Lightjet printshops he knew of wanted 300 dpi final size files. I

also mentioned to him that 200 dpi at 20"x24" was equal to a 1000

dpi scan equiv.

 

I told him that if I was only scanning 4x5's at 1000 dpi I might as

well be shooting MF. Sounds to me like I am throwing away detail.

He laughed and said your right.

 

I think for a 4' x 5' print ould be fine but.......

 

What do you guys think

 

I understand to can print at R8 (200 dpi roughly), R12 (300) or R16.

At the risk of sounding like a Luddite, I think you should get a Beseler 45 or D2 with a color head! Michaelangelo wouldn't send a scale model of the David off to someone to make it into a full size statue for him, so why would you do that with your negatives?

I had some lightjet prints printed. They wanted from me a 305dpi file. I was told that 305dpi is maximum that their printer (Lighjet 5000) can print.<p>

My prints were drum scanned from a 6x9cm transparency; the size was roughly 20x24". If you do not want to print anything bigger then 6x9cm is just fine for the print size.

Troy:

 

I have experimented with this, and the difference between printing to the LJ at 200 dpi and 300 dpi is very subtle. In talking with folks at several major digital imaging firms on both coasts, they agree and add that the difference is also strongly dependent on the image, but more often than not virtually undetectable without a loupe. Looking at a 20x24 with a loupe is, well, loupy!

 

Printing at the equivalent of 1000 dpi will preserve detail up to about 20 lp/mm in the original chrome or neg. While most LF lenses resolve somewhat more, the reality is that most images don't have much content at those frequencies. These high spatial frequencies are mostly important for preserving edge sharpness, which can be dealt with by proper application of sharpening algorithms at the print resolution. The appearance of the print will depend much more heavily on the quality of that sharpening than on whether you feed the LF 200 or 300 dpi.

 

Another test I have run is to compare scans from MF and LF done at the same total pixels per image. I can assure you that 4000 pixels from LF (20x200dpi) aren't the same as 4000 pixels from MF. They are very close, but the numbers in each pixel matter, and the numbers from an LF scan are much more noise free from sampling a larger area of film. Again, the differences are subtle, but you don't even need a loupe to see this difference.

 

My local lab ususally requests 200 dpi for very large images (30x40+) but they let the customer decide. If you aren't comfortable with 200 dpi for a 20x24, use another lab!

That's the problem with digital printing with the equipment being used. You take a nice, sharp 4 x 5 and loose the detail. What a waste.

Well, this is a semi preliminary print that is going into a gallery. At some point depending on how this one turns out. I may print it really big, like 30x40 or 40x50 in the future, but I thought printing at 20x24 I would be printing at more lp/mm for this print.

 

Oh well

As Nancy said: "just say no".

 

I have run tests to see how much detail I can get to film with both air force test charts and real world photographs.

 

With Air Force test charts I can resolve 80 lp/mm and with real world photographs I can resolve the lettering on signs at 1.5 miles that reproduce to be 1/10 of a mm tall on the edge of a 4x5 transparancy.

 

If you assume that it takes about 10 rows of dots to resolve type, then it would require 2500 dpi to reproduce that level of detail at 4x5 or half that at 8x10.

 

I really don't see the point of shooting anything at above 35mm if you are going to throw all the detail away in the printing process.

 

I talked to someone at or local camera club the other night that had run side by side compairisons between 35mm and 6x6 and said that she couldn't see the difference in the prints.

 

After leaving the meeting I realized that she probably had digital prints made.

You might find the following website interesting.

 

clarkvision.com/imagedetail/printer-ppi/index.html

 

Roger Clarke claims to be able to see differences when printing at up to 600 ppi.

I must sound like a broken record but this question keeps comming up.

 

Ctien's book "Post Exposure" has some very good data on the limits of human vision. His data shows that there is a precievable difference up to 30 lp/mm but that above 6 lp/mm people will precive a single print with nothing to compair it to, as "sharp".

 

http://ctein.com/booksmpl.htm

 

(This is of course from negatives using the best techniques, I suspect with prints from hand held shots, it doesn't make much difference what you do, or what equipment you use.)

 

Williams book "Image Clarity" goes into much more detail and is a much drier read. It is also getting expensive.

 

Most people like the look of digital prints better than direct optical prints and given the cost differences I expect there to be fewer and fewer labs doing what I consider quality work.

 

I use the last lab in our town that makes direct optical color prints from negatives. No one in our town makes them from transparancies.

 

When they quit, I intend to transition to 90% B&W from my own dark room and mail order color.

Digital color printing is several orders of magnitude better than classical printing. You cannot compare it with classical enlarging. I do not believe that 35mm and 6x6cm had similar perceived quality. Photos from my 6x9cm transparencies are much better than from 35mm.Because of this reason I do not expose anymore almost any 35mm color film. I had my transparencies scanned on drum scanner. I am not sure whether you are accustomed with the lightjet technology. I had my photos printed on Kodak Endura paper with mean expected fading between 100 and 200 years; that is what Kodak says. It was printed at 305dpi and the detail that is in the file is also in the print, though, you might need a loupe to check it. <p>

Lightjet color printing is also mainstream technology for any large color fine prints. Almost all Gursky's photos were printed on a Lighjet printer by Grieger in Düsseldorf. The prints are often 4m (15ft) wide... I think most of his photos were scanned from 4x5".

"Digital color printing is several orders of magnitude better than classical printing."

 

Can you post some links to data to support that?

 

I am not looking for opinions no mater how good a photographer they come from.

 

The term "orders of magnitude" strongly implies scientifically reproducible tests that generate numbers. The only objective measurements that I am aware of for evaluating photographic prints are dynamic range, lp/mm, mtf, and granularity.

 

Objective testing shows that it is within most of our capabilities (large format photographers) to get detail in the 30 lp/mm range to an 8x10 print. It takes a minimum of 1525 dpi to reproduce that. I.E. 30 x 2 x 25.4 (It takes a minimum of two rows of dots to replicate a line pair a black line and a white space.)

 

How can it be possiable for a digital print to be several orders of magnitatued above that. If I remember correctly an "order of magnitude" is a decimal place or 10 times. Wouldn't several orders be 1,525,000 dpi?

"You cannot compare it with classical enlarging."

 

You just did.

Whoah! You don't have to rescan for each print size! Scan at the highest quality and resolution as the foundation for a master image, make the adjustments to produce a master image, then resample and resize to produce files for printing.

 

You're not really throwing "information" away! A print can only resolve so much. 300 ppi at the final size is the generally accepted standard. You won't get a better print at 1000 ppi than at 300-360 ppi, but it will take a lot longer because the print driver will resample anyway. 200 ppi seems a little thin. The pixelation might be virtually invisible at normal viewing distance, just a little soft. 200 ppi looks bad in a 4x6 inch print (which you view up close), but might be okay for a mural.

 

When sending images out for printing, it's best to do the sizing and sampling yourself. That way you maintain control over the results, including cropping, rather than relying on a technician with his own priorities. A greater problem is color balance, for which you need a calibrated system and a profile from the printing service.

To answer the original question, optimal Lightjet resolution is 304.8. Regardless of whether the diffs. between 200 and 300dpi are visible, that lab doesn't know what they're talking about. You should find one that does.

Photography is not mathematics. If you talk about print quality then you have to define what you understand under print quality. Assumptions and definitions come first. If you are only interested in resolution then theoretically it is possible to get the same quality for both methods up to a certain print size. The fact that someone is using an outdated method and not focusing on photography is his problem. If he can then compete with people that have much more time available for their philosophical and artistic growth then he is a genius.<p>

If I add to print quality the amount of reduced noise in dark areas then you cannot compete with a digital print anymore (this is just one example). That would count for me as one order of improvement. Why should I buy prints about which I know that they were not produced to current quality standards? Unless, you are someone famous or you photograph my family I will not do it.

<i>The only objective measurements that I am aware of for evaluating photographic prints are dynamic range, lp/mm, mtf, and granularity.</i><p>

How about this "objective" measurement? You locate a shadow area in your print that under normal cicumstances should be pure black. You produce pure black as one sample, and you use that area as other sample. You do a t-test on these two pixel populations. If it is not significant then you cannot reduce the noise too much. If it is significant you apply noise reduction and do the test again. When testing you compute the critical value for the test with a given level of significance. How about noise clustering, noise distribution? Many statistical tests assume that noise is Gaussian. You could really come up with many other "objective" measures that could be very relevant. Widen your mathematical horizons!

<i>You're not really throwing "information" away! A print can only resolve so much</i>

<p>

The second statement is certainly true. The first statement does not follow from the

second (just because the print doesn't show all the information in the negative or

transparency does not mean you're not throwing it away when you scan at low resolution),

and additionally it's simply false on the face of it. When you scan at a dot

resolution lower than the resolution of the film, you're throwing information away. Period.

<p>

If it's information you don't want or need, fine! Throw it away! But don't live in denial

about it.

<I>If I remember correctly an "order of magnitude" is a decimal place or 10 times.</i><P>An 'order of magnitude' is the degree of improvement my 8yr old niece can take better pictures than Neil. <P>

 

 

 

<I>Can you post some links to data to support that?</i><P>Sure, my own experience having LightJet prints made from the same MF and LF chromes and negs and then comparing them to optical prints I made myself. This is also "Photo.Net" where people engage in the photographic process, not "hyperlink.net" where second rate photographers like to argue against digital processes because they don't have any decent analog work to show for it.<P>In the past month I've noted two threads where a LightJet was blamed for bad prints, and in looking at the posted samples and carefully reading the anti-digital propoganda sites that were "hyperlinked" concluded the author wasn't using a LighJet at all, but in one case a Fuji Frontier (which is severely limited on the scanning stage) and in the other case it was an older Durst Lamda which *is* limited to 200 dpi, which sucks below making anything 10ft by 20ft. This being the case, we have some trolls here that will lie simply to push their anti-digital cause, and we shouldn't believe a damn word out of their mouth until they can post their own work that looks decent. Simply look at their work (Neil's) vs the galleries at West Coast Imaging, and conclude who knows what they are doing and has the higher standards. You'll also have hard time finding any acclaimed LF color photographer who doesn't rave about LightJets, yet these amatuers are trying to trying to get you to set up an optical darkroom and do Ilfochromes which is the lowest 'order of magnitude' color process that's ever been invented. <P>I've been making LightJet prints for years, and have yet to encounter one that shows 'scan lines' or pixelated artifacts, and my frame of reference is a minimum my optical 6x7 work printed on my old in house 1 1/2 story Durst enlarger, not some loser's 20x30 locker print from Fuji. If you give a LightJet clean information, it will deliver pristine prints that the best photographers won't be able to discern from conventional color prints. Everybody also blames the LightJet when the lab does a crappy scan or plays with the image too much in Photoshop. <P>If you were shooting arial work from 8x10 for the CIA and wanted to show as much detail in a 20x24 as possible, by all means an optical print will show more cracks in the sidewalk because of the limited spatial resolution of the LightJet. For a 20x24 from 4x5? No way. That's 1500dpi actual rez from the original film, which is plenty.<P>Troy needs to first go to a decent lab not run by amatuers trying to downscale his files because they get confused about the LightJet's interpolation ability with scans from small format film vs 1:1 resolution from large format film scans. That's a 'no-brainer'. Troy should also check what kind of printer the lab is using because it sure sounds like a 200dpi Durst Lambda and not a LightJet. I have no arguement with him that the lab is displaying what he thinks are soft prints. The problem is either the scan or the operator, and *not* the LightJet.<I><P>When you scan at a dot resolution lower than the resolution of the film</i><P>1500-2000dpi is the maximum you want to screw with scanning LF or MF film if you can get away with it because beyond that all you are doing is gaining better resolution of fuzzy dye clouds. Which brings rebuttals from 35mm photogs who are the ones 'living in denial' by trying to prove the smaller the format, the greater the resolution it has. My 6x7 work at 2000-2500 dpi looks MUCH better than 35mm at 4000dpi, and my 4x5 work scanned at 1000-1500 dpi looks cleaner than my 6x7 work scanned at 2000. Absolute film resolution has little to do with other than keep arm chair photographers occupied with trivia to argue about.<P>1500dpi works out to about a 20x24 from 4x5 film and 300dpi off a LightJet, all at 1:1, which is the sweet zone. You'll be hard pressed to get a sharper optical print than that, and sure as hell won't get a better Ilfochrome, which will be fuzzed up with the cranky contrast mask anyways.

 

"How about this "objective" measurement?"

 

It sounds like you are either describing MTF or Dynamic Range. Both have been used for quite a while.

 

MTF is the ability of a photographic system to transition from pure black to pure white with increaseing resolution.

 

Dynamic range is the range of tones and the ability to resolve the difference between tones.

 

What a lot of people like about digital prints is that the process throws out detail that is too small for the eye to resolve and that is precieved as noise. However, that issue predates digital by tens of years and photographers have been using darkroom techniques and equipment to do the very same thing for almost as long as photography.

 

I did my own survey with two prints one digital and one optical. The digital print had blocked up shadows, blown highlights and reduced detail. The cream colored walls of my house were ice box white in the digital print. In a nearly 2 to 1 majority, people prefered the digital print. The word used most often was "crisp".

 

However, in terms of faithfulness to the origional scene using any objective measurement, the digital print was garbage.

<i>blown highlights...</i><p>

I think that my lab does a pretty decent job. I had to produce a 305dpi file for them if I wanted optimum quality. They recommended Adobe RGB98 as color space. The prints are really good. They are exactly what the file looks like on my calibrated monitor. I had to spent some time with them talking about profiles and how I can improve my printing. They argued that whatever gets done to make the process transparent I will need to evaluate test prints anyway. I took the print to a gallery and the guy was really surprised, He called it the best COLOR print that he has ever seen. Then we argued about longevity of the print. He insisted that nothing can beat his platinum print. I mentioned that Kodak says that Endura paper should be good up to 100 years under normal display conditions, and up to 200 years under archival conditions. He was surprised. He had no idea about how much the technology has progressed.<p>

You cannot compare maturity in digital printing technology and digital cameras. Digital printing is maybe 15 years ahead. The best quality available technology currently available is Large format photography, high end drum scanning, Photoshop with experienced user, and Lighjet printers.<p>

In fact once the image is in digital form I could produce a plenty of useful tests. Just take a look as some statistical analysis or data mining books...<p>

The fact that some people believe that they can beat a digital print with their enlarger just makes me happy. It means less competition for me. The moment that there is a 50-100MPix LF digital back available on the market for a reasonable price I will switch to a pure digital work flow. I want to concentrate on my ideas, philosophy, and artistic expression rather than on technical details.

Digital printing is way ahead of digital cameras. Yes.

 

"Just take a look as some statistical analysis or data mining books..."

 

I have some small experience with statistics, I am in quality management. ALL the statistics that we deal with start with accurate measurements and design of experiments. Are there statistics that can be done without quantitative data?

 

However, bigger, more, less, smaller can be either better or worse depending on the desire of the consumer.

 

I have a 2" X 4 1/2" transparency. I paid for two different labs to make prints from it. One was scanned and printed digitally, and the other was a direct optical print (Illfochrome). I like the optical print better. Others like the digital better. It is a matter of opinion. However, there are details that are resolvable on the optical print that are not resolvable on the digital print. Further they are logically too small to be resolvable on the digital print. I think that they are important, others don't.

 

What I have never seen, either from viewing prints side by side, or understood from analyzing data: is that a digital 8x10 from a 4x5 scan is better from a digital 8x10 from 35mm.

 

It seems to me that a 35mm negative carefully made and scanned at 4000 dpi will have more than enough information for an 8x10 digital print. I don't see the point of having 16 times as much before you start the process.

 

 

It just seems to me that if you are going to shoot large format and print to smaller than 11x14 you are condemned to a wet darkroom for it to make sense.

 

I have, to satisfy my on curiosity, had 12000 dpi drum scans made from sections of a large format transparency. From that, I conclude that I can get more information off a transparency or negative with a drum scan than I can with an enlarger, I just can't figure out what to do with it other than make prints measured in feet instead of inches.

 

I suspect that laser digital printing technology is capiable of way better than the 2, 3 or 400 dpi choices that we are being offered but that the write times and costs limit the market to the point that no one is going to offer the service. One of our local labs has a machine that will print at either 200 or 400 dpi but even if you specifically request 400 dpi they print at 200. They think 200 is good enough and if you don't like it you can go elsewhere...I did.

I'd like to be clear that I agree with Scott, and don't want to make any comment on the

quality or lack thereof of digital prints, since I don't make them and therefore don't know

about them. My only point was about the information in an analog original, which IS

thrown away if it's scanned at too low a resolution. Scott's point that not all the

information in a negative or slide is image information is a good one - dye clouds, grain,

etc... are information, but they're not image information.

<I>I've been making LightJet prints for years, and have yet to encounter one that shows 'scan lines' or pixelated artifacts [...]</I>

 

<P>

One time, I got a "bad" Lightjet print (from Calypso). It had some nasty artifacts that looked like the "halftones" you see in old comic books. Calypso reprinted it, and it was perfect.

 

<P>

Occasionally, something can screw up in the Lightjet process, but that is an error, and it can be corrected.

The short answer is find another lab. You should be sending files that are 305 ppi to the LightJet for best prints.

 

As for scanning, I always scan at the highest resolution I can and then resize in Photoshop to the ppi required for the print size.

 

For a 20x24 inch print from 4x5, at 305 dpi for the LightJet, you need an image that is 2288 ppi - this number even accounts for resampling by the printer.

 

As for digital versus wet darkroom...

 

35mm to 8x10 print: wet darkroom and digital prints - wet darkroom (from a negative) looks as good. Transparency - digital looks better because an internegative degrades the image, and Ilfochrome color is not close to accurate - especially in the blues. It has it's own look that can be quite nice, but let's not pretend it reproduces color accurately - it doesn't.

 

Larger prints from larger formats? That's a bit trickier. Depends upon the enlarger, enlarger lenses and, most importantly, the person doing the printing. At 20x24 - my vote would be the digital print.

 

If you can resolve the freaking dye clouds in the film with a scan - just how much image information can you be missing - OR not printing? When the LightJet is working at it's highest resolution, it is making 12 lines per millimeter. That's really, really, really....small. Like 3/10,000 of an inch??

 

You do understand that the printer is making continuous lines the width of the paper? The print is made up of teeny-tiny continuous-tone parallel lines? It's not making dots or pixels?

 

If you see pixellation, then the image was not scanned correctly, or, if the file was from a digital camera, then the pixel matrix was not large enough to support the print size - scale the freaking image down to the correct size.

 

This is NOT a digital problem - it's a stupid human problem.

 

If you get lousy digital prints - don't blame the machine, blame the person doing the digital work and their workflow.

 

If you can't get prints that color match the film image exactly - you suck at digital work and should probably use a wet darkroom - or spend some time and learn how to work digital images properly.

 

There is absolutely no reason that you cannot get color prints from a digital system that are equal to, or in most cases, superior to wet darkroom prints - especially at larger print sizes.

 

Funny, you never hear people talk about diffraction, illumination falloff, internal enlarger reflections, edge sharpness, dirty lenses, dust, - or any of the other myriad of optical problems that kill detail,contrast, and resolution in a print made with an enlarger. Believe me - optical systems degrade the image too!

 

The anti-digital print expert always starts with the premise that a wet darkroom print retains all of the image information - it's 100% perfect - and digital somehow degrades it.

 

I have to wonder if they've really ever MADE big prints with an enlarger or are even faintly acquainted with digital imaging systems- or if they are, if they're even half-way competent.

 

Next, I'll be hearing about how by scanning the image it's lost it's "soul," which is so readily rendered by light passing through the film onto paper - barf, barf...

So I am looking at this print hanging on the labs wall, my thought is it looks a bit soft. Now supposedly I would assume they are going to hang their best. I ask for a loupe, mostly because i have never seen a lightjet print under magnification and do understand that they are perfectly smooth, with no artifacts, and continuous toned then I see bands. These did not appear to be in the scan, but at the print level. Dont ask me why but that was my gut. The super tiny bands were very even and identical. I would assume scan lines would fade in and out and would not be so even and regular, but I am not sure about that ??

 

So whats the deal. A bad operator or do they really have a lightjet. They do advertise lightjet prints, but now I am having my doubts. also they do have a 2 day turn around.

 

Maybe I should have them run one to check the quality ?