Resolution & Contrast - Optical vs Digital printing from FILM

My ignorance is profound. Clue me in. I learned about everything I know about

photography by reading, and then Doing It until I'm competent to repeat the

practice successfully. I do not have the facility to do the following.

<p>

First, it is my experience that in optical enlarging I lose (rule of thumb)

roughly a third of the acutance of the negative, BUT it doesn't matter until I

get to about 10x. With 4x5, that's just perfectly fine.<p>

First: <u>Does one lose less detail when scanning LF and then digitally

printing LF on a conventional ink-jet printer at a nominal 360dpi than one

gets printing optically?</u><p>

Second: <u>Do large format (say 10x of nominal 4x5")printers really print at

360dpi (ppi,spi)?</u>

1) This is a religious question. You'll get arguments both ways because it depends on what detail you mean and where it ends up. For example, with a darkroom print you can use a loupe and see detail in the print that you can't see from anything like a normal viewing distance. When you put that loupe on a digital print, you don't see a lot of "extra" detail. But does this matter? Up to you.

 

What I can tell you is that in my experience I loose less detail when drum scanning 5x4 than when printing in the darkroom. There are some reasons for this. Chief amoung them are the specialized optics of drum scanning (it's more optically efficient because it's only looking at a very small part of the film at a time). Also, fluid mounting on a drum scanner puts the film in the exact plane of focus -- edge to edge sharpness is outstanding. This same level of edge to edge sharpness is difficult to obtain in the darkroom because of the precision alignment needed, especially difficult with 10x enlargments (125 x 100 cm, or 50 x 40 inch prints).

 

2) People often do this at first, but then they find that working on files that size is seriously painful. The curious then do some testing. They'll scan the same film at different output PPI and print a section from what would be a full size print for each scan and compare the sections side by side. Even very close I find it difficult to see the difference between 360 ppi and 300 ppi (highly detailed landscape work printed on an Epson printer). I can see more difference between 300 ppi and 240 ppi - I loose some sharpness and get some "coarsening" if you will. So most of my work at that enlargement level is done with an output of 300 ppi.

 

The only way to find out what output PPI you should use for your work is to test and see.

 

DISCLAIMER - the above discussion assumes drum scanning. If you are scanning on flat bed scanners (either consumer or pro) there will be some differences. The only way to know what works for you will be to make prints and compare.

Scanning how? There's a world of difference between a drum scan and a flatbed scan. And a world of price difference too.

Bruce, thanks for that. It's an "all depends" situation, I guess.

 

So, when you order a 40" wide inkjet print, do you think the printer really does 360dpi?

 

And regarding drum scans --- you all sure they are that much better than very good flatbed scans?

My experience is that scanning with a flatbed scanner and printing on a typical good quality inkjet printer such as an Epson 1280, one can't match resolution of fine detail obtainable by printing with an enlarger on photographic paper. But you have to make substantial enlargements---more than 4 X---and look at them closely in order to tell the difference. On the other hand, if you scan with a high quality scanner and print with a high quality printer such as a LightJet, then you may very well be able to do better. For color prints, you gain enough by proceding digitally through easier control of color balance that it might still be worth proceding digitally, even if you don't have the highest quality scanner or printer.

I know close to nothing about inkjets and their professional use, but I can't imagine any technical reason why a bigger inkjet should be worse than a smaller one besides that maybe nobody needs it because ordinary folks don't get enough pixels together.

 

Standing in your shoes I'd try to get hold of a small output sample of the LF inkjet before I'd buy it or order the huge print. If the final print is done sloppily, at lower res, I wouldn't pay for it.

 

Many people, even professionals confuse ppi, dpi and lpi. A standart offset print of 150lpi for example, demands a 300ppi file (to be on the safe side) and a 2400dpi plate writer to put a conventionally screened image out. A inkjet might demand / waste less pixels for safety than the presses I'm talking about because conventional offset screens have to be done at 15? angle of one color from each other to get results of similar look and the important colors like cyan are angeled roughly 45? to your monitor pixels, so there is a need for a 2x or at least 1.4x higher image resolution to avoid too much interpolation during definition of the output screen dots. Maybe a inkjet needs less, I don't understand the logic of modern screening methods (where not the dot size but the distance between dots varies) right now and the multitude of colors printed by photo inkjets adds to my confusion.

I can only speak from my experience. That is, I used to print both b&w and color neg in

the darkroom, but have been making LightJet prints from drum-scanned 4x5

transparencies since 1999. When I look at the old 16x20 b&w prints I used to make from

4x5 TMX in the darkroom, I still love the look of the prints, while the 16x20 prints from

Fuji NSP (the old version of NPS) are nice but fall short of the quality of modern Epsons or

LightJet prints. This goes back to the fundamental difference between the b&w and color

darkrooms, in that color darkroom work never reached the level of control possible (with

reasonable ease) in the b&w darkroom. Christopher Burkett will disagree, but he spends

far more time in the darkroom than I could ever tolerate. I prefer to be outside!

 

If you drumscan your 4x5 negatives or transparencies at 2400ppi+, and make prints at

20x24 or larger, I am sure you will be blown away by modern digital printing. Newer

flatbed scanners can do pretty well, but good drum scans still beat them hands down.

Imacon scans are better but still not as good as drum scans.

 

I recently created a master file from a 500MB drum scan of a 4x5 Velvia transparency I

made in Zion N.P. this fall. I printed a 40"x50" at 304.8 pixels-per-inch (120p-p-cm) with

a LightJet printer on Fuji Crystal Archive photographic paper. The detail and print quality

would blow you away!

 

I strongly recommend Color Folio of Sebastopol, CA (www.colorfolio.com) for drum scans

and color or b&w digital printing. The owner, Bob Cornelis, is very accessible and

extremely helpful.

One problem with conventional optical printing onto photographic paper has always been the quality of the lens used for the enlargment. Enlarger alignment gets a lot of attention, and is another profound problem for many people. But I think that the optical quality of the lens has generally been overlooked, and many peoples' general printing quality has really suffered as a result.

 

In his excellent book "Edge of Darkness: The Art, Craft, and Power of the High-Definition Monochrome Photograph", Barry Thornton goes over enlarging lens tests he's done which illuminate his point that the quality of enlarging lenses is often distressingly poor, even hellishly-expensive samples from the "best" manufacturers.

 

Over time, my own experience with this has led me to generally trust graphic arts "process" lenses over enlarging lenses; and when deciding on whether to keep a lens, to trust only results rather than reputation and reviews.

 

I just thought I'd throw this point out there. Before relegating "optical processing" onto the scrap heap, please evaluate your working tools and techniques very carefully.

 

That said, in fact, I'm intrigued by the possibilities for color printing in particular that the digital workflow now presents.

 

I think I'll probably stick with optical, "conventional" or "traditional" techniques for black and white, though, simply for the degree of control it gives me over the final result. I just don't see any reason to spend all the extra time doing it digitally -- which also begs the question "how many options for this print have I missed because the workflow itself just wore me out?"

 

"Traditional" color printing, on the other hand, is something I've never tried for various reasons, mostly having to do with the potential lack of control; the expense; and last but not least, the toxicity of the chemicals. I'm much more willing to put up with black and white chemistry, especially for the working speed and price range of materials.

 

The digital workflow really comes into its own when working with color, in my humble opinion.

 

Thanks for the above posts on that subject, very much!

That is really something lots of people overlook, that the enlarging lens has to be perfect and there are enlargers that realy hold your negative straight.

<b>Fred Gols</b><i> That is really something lots of people overlook, that the enlarging lens has to be perfect </i>[... snip good stuff ...] <p>

Indeed, Fred. Good of you to speak up. In fact, a negative that yields, say, 100 lp/mm that goes through a lens that also yields 100 /lpmm actually yields something like 65% (maximum) of the negative, so all optical prints are diminished before they hit the paper where they are diminishted just a tiny bit again.<p>

But in real experience, in real life, stunning images are made with a conventional enlarger, regardless. But you know that, Sir. I am speaking to the rest.

Pico,

 

Lot's of sound advice. My own experience largely concurs with your own initial point - My

experience is that for prints to 16 x 20 or thereabouts full frame optical enlargements of

fine grain 4x5 film have a more pleasing rendering of fine detail than drum scan / Lambda

prints from the same neg. This is of course contingent on the premise that you want an

emphasis on detail in the image. The optical images may not be as consistently sharp

edge to edge, but they are sharper in the middle. The problem I found with digital seems

to lie in the murky world of default unsharp masking at various stages in the process.

These defaults work well with bigger enlargements but mess up the finest detail of a

modest enlargement from a 4x5 neg. With such modest enlargements acutance might be

higher with the digital prints, giving them more 'snap' but resolution is not as high. This is

my empirical experience of side by side comparisons. I can only guess that this is because

the output resolution of Lambda, Lightjet etc. is largely fixed independent of image output

size, while the actual resolving power of a given enlarger / lens combination varies on a

sliding scale relative to the actual size of the enlargement. For smaller enlargements the

optical print has potentially significantly higher resolution, whilst with bigger

enlargements things change - I think the limitations of the enlarger and lens become more

of an issue relative to the fixed output resolution of digital. Digital probably trumps for

bigger enlargements. In saying all of this, there are so many mitigating factors in either

workflow that any hypothesis is largely useless. The only way is testing - direct side by

side assessment.

In using 36 and 54 inch roll fed printers for the publics inputs, there is a trend for the customer to get lost in the vortex of ppi/dpi/resolution. Upsizing average inputs doesnt usually make a print any better, it just ticks off a printer because it bogs the RIP with useless info. We printers get alot of average or poor inputs, and alot of big egos. We often downsize retarded inputs if there is no change in the final print appearance. Few folks really use high end drum scans. Some folks do these weird steps, they scan their 4x5 on a flatbed, then print it 8x10 on inkjet ; then scan the print at aq zillion dpi to make a monster bloaded 1 gig file. This fills the ego of the customer, huge megapixel file with no thought of the missing shadows and highlights. Printers often need the original 4x5 to do decent print, since theses half baked upizing events often butcher the image. Property done either a digital or an optical print can be great. There is a huge goober factor with the publics focus on ppi/dpi /resolution versus actually pulling the real detail from a 4x5 original. There is no reason for a printer to print a giant print at 40x50 inchs at 360 ppi if it really is just a flatbeds input that chops of the higher end detail. Sadly one czan often downsize the publics inputs 10x and never see any difference, sometimes even 1000 times when folks are experts and do theses goofy upsizing events. Some RIPS just automaticalyy discard the useless info.

<i>So, when you order a 40" wide inkjet print, do you think the printer really does 360dpi?</i>

<p>

Depends on the printer. Epson wide format printers have a native resolution of 360 ppi. If you send them something else, the Epson driver will resize automatically. It does this because the mathmatics of building a dither pattern are much easier from a "fixed" base of 360 ppi. Canon and HP use 300 ppi IIRC. Lightjets use 304ppi (120 ppcm). And on and on...

<p>

When you use a RIP, all bets are off. Many (most?) RIPs will do the complex math to get the chosen printer to use all the data that your file contains regardless of your output ppi. That's part of why people use RIPs.

<p>

All that said, this doesn't answer the unasked question: how much resolution should you send to the printer? This is more dependent on the image and its intended use. In many cases you'll get an excellent print sending 240ppi (or less) to the printer. This is expecially true when the prints are big -- because people tend to view prints from a distance more or less equal to the print's diagonal measurement.

<p>

<i>And regarding drum scans --- you all sure they are that much better than very good flatbed scans?</i>

<p>

I am, yes. I put my money where my mouth is too. I bought my own drum scanner.

<p>

How long this stays the case is anyone's guess. Eventually a professional CCD scanner will equal a PMT scanner in sharpness, and even Dmax. But we haven't seen it yet...