Resolution loss from transparency to print

[Nathan Potter]

Jac, by a strictly academic answer and with no enlargement at the paper, the chain of lp/mm you cite would result in about 22 lp/mm at the enlarging paper at 1:1. Each doubling of the enlargement halves the lp/mm at the paper.

Of course per the above arguments, all of which I think have a degree of validity, there is a lot more to the comparison than just academics.

For me, how the textural and fine detail of the original scene is preserved through the entire workflow is important in some cases. If the film original contains the desired level of detail then from my experience the best replication is obtained from a high quality optical system. I tend to agree with Drew on this.

OTOH I find a scanning digital workflow utilizing a 2000 nominal spi scan, for example, with judicial sharpening in PS can actually appear to be a sharper print. For some images I prefer this. My problem is that the digitally enhanced sharpness is to some degree artificial in that it does not replicate the fine detail or microcontrast in the original scene (or on film) like a high quality enlarging lens can. This is particularly true using a fairly high contrast enlarger.

I would certainly cite the Ilfochrome work of Chris Burkett for state of the art in optical reproduction but I also work with Ilfochrome and 2000 spi scanning and would generally confirm what I said above.

I'm not sure the difference is at all significant when we consider very high quality drum scanning in comparison to an all optical workflow.

Nate Potter, Austin TX.

[Jac@stafford.net]

Thank you all for the good discussion.

Visual intelligence, or how humans perceive is an academic interest of mine. I had the privilege of working with some excellent psychologists (the perceptual kind, not shrinks) and others in the sciences. One of the things that came up in conversation makes me wonder if having been a photographer and darkroom printer for forty years has changed the way I perceive some things, in particular I do not perceive the Cornsweet illusion. Nor is my vision confused by certain black-and-white tonal differences caused by surrounding and adjacencies. Has anyone else here a similar experience?

Note that I am still entirely susceptible to confusions of colors caused by surrounds and adjacencies. I have normal to excellent color perception, but not as good as the 1% who have the best. (I do very little color photography.)

Aside: A friend of mine is an accomplished, published documentary photographer. He does all 'straight photography' in MF and 6x17cm. At a show I mentioned to him that his prints were soft on the left edge and he responded, "That's the kind of thing only a photographer would notice." I think our discussions are very much along the same line. It is a good thing.

[Drew Wiley]

You can simply bypass a lot of the MTF calculations with a simple analogy. You have a lens on a view
camera (some better than others), then a piece of polyester or acetate film on the back. You expose it.
IF the film was flat, properly focussed etc. you view that on a lightbox under a high-power magnifier
and can often see a LOT of detail. Same principle at the enlarging station. Now look thru a loupe at
an inkjet print or lightjet print or a typical substrate. Not much there. This isn't about how good it looks
to an ordinary viewer (which might be fine) but about the academic ability to hold resolution. All the
number crunching in the world won't change that. You are only as good as your weakest link. A proper
optical system, given a highly detailed oringal neg or transparency, will transfer virtually all that information onto an equivalent polyester photosensitive surface. Paper and ink cannot. Again, none of
this is about esthetics, but simply why splitting hairs over published MTF data is a very minor issue
compared to the other problems in reproduction.

[january112013]

Hey Drew,
That last analysis cleared up my question. Thanks
Now, is there anywhere on the planet where darkroom printing is still done?
Jim

[paulr]

The proof may be in the pudding, Drew, but you haven't offered any. Just personal attlacks. I've invited anyone in the NYC area to come to my studio to look for themselves.

You say my enlarging settup was primitive; I'd love to hear about a more advanced one. But for the sake of argument, I'll agree, and just let people compare my inkjet prints to contact prints from the same negatives. I'll even put 3X enlarged ink prints against contact prints from the same negative.

Ascertaining sharpness by looking through a loupe is just blowing smoke; sharpness is highly size dependent. This wasn't obvious with analog tools, although people sensitive to the process who used staining developers and other techniques that produce edge effects figured this out. Digital tools make it obvious to anyone.

You're incorrect in saying you can't add sharpness. I believe you mean that you can't add detail, which is not at all the same thing. Sharpness is a subjective quality, and it's influenced primarily by MTF at a particular range of spatial frequencies. This can increased or reduced with great control.

I'm offering pudding for proof to anyone curious.

[Drew Wiley]

Paul - edge enhancement can be done all kinds of ways and is not the same thing as preserving detail. This is a bit academic
obviously, because if no one really sees micro-detail in the final print, is a mum subject. I do like it when one can work at the
visible limits however. Place a microfilm target in the glass neg carrier and then a piece of high resolution film on a vac easel
and print that and look at it through a high-power loupe or low power microscope. That starts telling you what optics can really do. Way back in the 1930 they were having contests seeing who could put the most number of complete texts of the
Bible on a microdot, and even back then there were film and lenses that could do about thirty! Of course, full scale tonality
photography is a somewhat different subject. I really don't care how someone chooses to print their own work, and there are
many viable specific media. The only issue I have is folks who go around waving their arms and screaming that newer technology trumps everything simply via its novelty. Then they jump thru a lot of software hoops attempting to imitate what
was done better a century ago using primitive techniques. In your case, you've been implying that digital printing actually
has some kind of edge relative to sharpness - and that would have bordered on nonsense a hundred years ago. Maybe it's
easier for things like squirrely 120 films, but there are all kinds of means of analog printing, which don't have the same
output bottenecks that color digital printing does (per resolution that is). Black and white work can be done a little more
precisely, esp by folks with serious prepress and press gear, but millions of bucks put even into that game won't match a
basic 8x10 contact print done correctly. It's technologically impossible. And don't go thinking we Calif folks who are a bunch
of hicks who don't understand what that means. If I want to see world class inkjet or laser printed work I can do so almost
in walking distance. Some of it gets installed in your city for huge sums. I work with some of these folks supplying them specialty gear and chatting esthetic strategy, and you'd be pretty amazed what they can come up with, as well as with their equipment and facilities.

[Drew Wiley]

Oh but Paul - please keep this in perspective. It's a tech debate, and I in no manner whatsoever am implying I don't like your
work in an esthetic sense, or wouldn't admire your own prints in person. And yes, it is kind to invite me to view it in person,
which I would do if I ever had opportunity.

[Bernice Loui]

Edge enhancement, up sampling or etc in post process does not increase the actual amount of information. If this held true then perpetual motion would be real and current information theory has been proven false. These processes can never equal real amounts of information acquired.

It seems the current norm for computer based printing is something like 300 dpi which translated to a dot size of 0.0033". In order to meet the Nyquist criteria for conveying stored information two points are required which means the actual smallest area possible to convey information would be 2x 0.0033" or 0.0066".

The average thickness of a human hair is 0.004".. which is quite visible to most individuals when places on a sheet of white paper at a distance of 12".

Point being, there is far more to image quality than simple lines per mm, dots per inch or any other highly simplified metric of visual acuity. How and the way we see is far more complex than that.

High resolution or information content alone does not make an expressive image.. it takes far more than just these metrics.


Bernice

Paul - edge enhancement can be done all kinds of ways and is not the same thing as preserving detail. This is a bit academic
obviously, because if no one really sees micro-detail in the final print, is a mum subject. I do like it when one can work at the
visible limits however. Place a microfilm target in the glass neg carrier and then a piece of high resolution film on a vac easel
and print that and look at it through a high-power loupe or low power microscope. That starts telling you what optics can really do. Way back in the 1930 they were having contests seeing who could put the most number of complete texts of the
Bible on a microdot, and even back then there were film and lenses that could do about thirty! Of course, full scale tonality
photography is a somewhat different subject. I really don't care how someone chooses to print their own work, and there are
many viable specific media. The only issue I have is folks who go around waving their arms and screaming that newer technology trumps everything simply via its novelty. Then they jump thru a lot of software hoops attempting to imitate what
was done better a century ago using primitive techniques. In your case, you've been implying that digital printing actually
has some kind of edge relative to sharpness - and that would have bordered on nonsense a hundred years ago. Maybe it's
easier for things like squirrely 120 films, but there are all kinds of means of analog printing, which don't have the same
output bottenecks that color digital printing does (per resolution that is). Black and white work can be done a little more
precisely, esp by folks with serious prepress and press gear, but millions of bucks put even into that game won't match a
basic 8x10 contact print done correctly. It's technologically impossible. And don't go thinking we Calif folks who are a bunch
of hicks who don't understand what that means. If I want to see world class inkjet or laser printed work I can do so almost
in walking distance. Some of it gets installed in your city for huge sums. I work with some of these folks supplying them specialty gear and chatting esthetic strategy, and you'd be pretty amazed what they can come up with, as well as with their equipment and facilities.

[Brian C. Miller]

Any recommendations on neg film?
Jim

I just realized that nobody answered this. Ektar 100 is currently "king" for finest grain. However, you should really use this film with a color meter, and balance the film with appropriate filters when you are exposing it. However, do try Portra. It's an artistic call, and a certain subject or scene might be suited to it better.

As for current darkroom work, there are a large number of B&W commercial printers producing up to 20x24. (I know there was a lab in NYC that printed larger, but they don't appear to be doing that any longer.) For larger work, Bob Carnie on the forum prints larger B&W. I don't know about color, so perhaps someone can chime in about a commercial lab which still prints color with an optical-chemical path.

[paulr]

We all agree that there's more to a print than detail, sharpness, and resolution ... which I am not listing as synonyms, because they are not.

This thread brought up issues of what the eye can see, which is what I'm addressing. I have divided this into two separate questions, because "what the eye can see" and "what makes a print look sharp, detailed and clear" are only distantly related, even though we often conflate them.

When I'm talking about sharpness, which is a major component of perceived image quality, I'm talking about a perceptual quality. Imaging scientists and perceptual psychologists start with perception. Like, if 19 out of 20 test subjects said that print A looked sharper than print B. The technicians then find which measureable qualities correlate with that impression.

Old fashioned metrics like total detail and maximum extinction resolution do not correlate. A print that only resolves 6 lp/mm can appear radically sharper and more detailed than one which resolves 30 lp/mm. The eye can't see all that detail, and the detail that it uses to form its impression of image quality is much lower frequencies: 1 to 5 lp/mm. This is why MTF curves have replaced resolution numbers when discussing optical quality.

The higher resolutions matter if you're looking through a loupe or a microfiche machine. This can be significant for some uses, but not for normal viewing of photographs.

My contact prints contain much, much more recoverable detail than my similar sized ink prints from the same negative. But the ink prints appear more sharp and detailed at normal viewing distances. The fact that they are storing less detail is irrelevant, because that added detail is not visible.

Drew keeps saying that my claims are "impossible," but the results are easily visible to anyone who wants to look. The optical science explaining this is also unequivocal, and is easy to lay your hands on today.