Apples to apples comparisons

[Chris Gittins]

The specific impetus for this post was the article in the July/August View Camera on PMK vs Pryocat HD. The bone I have to pick isn't specific to the materials discussed there. The prints which accompanied the article just raised an issue that's been on my mind for awhile: How does one make an appropriate comparison between negatives developed in A vs B, prints developed in C vs D, printed on paper E vs paper F, etc.? I don't mean that as a rhetorical question. What common _objective_ characteristics must A and B, C and D, etc. possess in order to meaningfully compare them? I'll limit my comments here to comparison of prints, but the same general considerations hold for negatives as well.

Picking on the prints in the View Camera article for a moment - with apologies to Steve Simmons, as I appreciate the objective of his article and the prints in question are only one example of a widespread phenomena - the prints made from negatives developed in A and B appeared to be printed differently. (For the purpose of this discussion its immaterial which was Developer A and which was Developer B.) The prints made from the negs developed in A are significantly darker than the negatives developed in B. You can tell its not simply an overall contrast effect because both the shadows and highlights are darker. With one print being darker than the other, I found it very hard to tell whether one negative was really giving better highlight or shadow separation or whether it was just an overall print density effect. If you print Zone III down to Zone II, then loss of shadow detail is no surprise. Neither is an apparent improvement in highlight separation if you print VIII down to VII, etc.

And that gets back to question of how you make a meaningful comparison of any two prints. It seems to me that if you want to do meaningful apples-to-apples comparison, then you have to pick at least two tie points (Zones) where the print densities match, e.g., Zone III and Zone VII. If you have one print darker than the other, it's going to be much harder to make that comparison. I'd imagine it can be done, but the difference in overall density is a big distraction. Do others concur that you want to match both a shadow Zone and a highlight Zone in the prints you're comparing? (Paper white and Dmax don't count.) I don't see another way to make a good comparison of highlight/shadow separation unless you have tie points. For example, if you don't, then how do you know the perceived effect isn't due simply to a mismatch in paper exposure scale to negative density range?

Other thoughts?

Chris

[Keith Fleming]

The reality is that all technical articles in magazines must be evaluated by the reader based on experience and judgment. Magazine reproductions, no matter how good, are not the same as the original prints. Those are not "prints" in the article--they are at least two generations away from the original prints. The originals were scanned, and that process may introduce slight differences. Another step is the printing onto paper. It's quite possible that the quality of printing may vary at times during the press run of an issue of a magazine, so your copy may not be the same as mine. Given the variables, the recourse is to read the article, study the illustrations, make a judgment call on the accuracy and validity of the article as a whole, and then study your own tests (which you may have done long ago). It's to be expected that you and I will not always agree with an author's conclusions; after all, we in the large format community can't even agree on which large format camera is best. The article is part of a dialogue that includes this forum, and that's a healthy thing.

[Alex Hawley]

I thnk you're right on the point Chris. I haven't read the article so I have no opiions on it. However, its a fundamental point of any scientific or near-scientific test that the number of variables must be minimized. Prefferably, there should be only one variable. Many times, that's not possible, so the other variables that are not being tested must be accounted for.

Here's how I did a test between Pyrocat HD and ABC pyro. First, I made 2 negatives of the same subject. The only difference in exposure was speed rating since the concensus of the time was Pyrocat allowed an extra stop of film speed. Of course, both negatives are from the same box of film.

Both negatives were developed by inspection in the different developers, noting the development times. Here's the major difference we are testing for and we know what and why it is.

Printing is done on the same paper. The ABC negative is printed first. Several trials are made until I feel I have a good print. The Pcat neg is next and exposed for the same time as the ABC neg. As to be expected, the exposure is not good for that respective negative but it does give a good measurement of the density difference between the negatives.

Then I proceed to get a good print out of the Pcat neg. Now we know the exposure time difference (remember, same paper and paper grade). After everything has dried, we can make a judgement on which negative printed best, and the differences between the two in processing are known and not subjective.

I wind up with only one variable besides the developer which is the point of the test; that was developing by inspection. Knowing the relative inaccuracy of the method lets me make a better judgement in comparing the prints resulting from the two developers.

[Alex Hawley]

Keith's point on trying to compare the prints in magazine is a valid one too. Its an exercise in futility trying to show subtle differences when the images are at least 2 or more generations from the original. Better to describe the diffrences in words.

[Jorge Gasteazoro]

IMO if you develop the films to the same contrast index, print with a fixed grade paper (not variable contrast) and expose the paper to the same time for all negatives, differences in the way the developer acted are seen on the prints.

I agree with Keith that a magazine reproduction might not be the best way to judge differences in prints. OTOH, the prints should look somewhat similar, maybe some change in contrast, but they should not look far darker or lighter.

Without having read the article let me make some guesses:

PMK was the better developer because it had better contrast, better edge effect and had the same speed as Pyrocat HD (Which is totally ridiculous).

If there was no sensitometric data to back up these claims, what you are reading is an opinion based on the authors eye balls, not an article comparing two developers.

[Alex Hawley]

Exactly, Jorge. I don't have any sensitometry equipment so I know I can't get an exacting match in CI. I try to compensate by making the best print I can from each negative. But the outcome, as you say, is still an opinion.

[Jay DeFehr]

While View Camera magazine would probably be the last place I looked for an objective comparison of anything, I think it's important to keep in mind what we are trying to compare. When we compare film developers, what is it about the developers that we are comparing? Film speed? Activity? tendency to fog? Grain formation? Stain color and proportionality? These attributes combine to characterise a staining developer, and each is measurable. I contend that a table listing these attributes for each developer and common films would be far more useful than the "proof-in-the-pudding" methodology of making prints for comparison, which introduces so many variables that nothing of value can be gleaned from the exercise. I think it's possible that some readers of View Camera don't really want to be confronted with raw data, regardless of its validity, and prefer to take the advice of an "expert", with nothing more significant than a few magazine reproductions and a testimonial to make the case for them. Mr Simmons finds himself in a quandary; how to qualify himself as an expert in the face of criticism from legitimate experts without doing the necessary work, that even if he were capable of, runs the risk of alienating too many of his readers. I don't envy him his position, and wish him the best of luck with his publication and reputation. In the meantime I will continue not to look to VC or its editor for objective, meaningful information.

[Jorge Gasteazoro]

Well yeah Alex, but if you are going to publish a comparison I would think you would want to provide hard data, not just, well this is how it looked to me.

Now my take when I did this was that if you take the time to produce a best print out of each negative form each of the developers, then there would not be any difference, all 3 were equally capable of producing outstanding prints. That is not proof that a developer is "better". The "proof" if we can call it that is which one produces the best first work print without any manipulations.

My observation was that if I wanted, I could have made excellent prints from each of the negatives that would have been virtually indistinguishable.

My conclusion was that WD2D (old formulation) produced the best first work print with the best film speed, but really with a little bit of extra work, all would have been equally good.

[Chris Gittins]

Just to reiterate, my point doesn't have anything to do with specific developers - PMK, Pyrocat, or otherwise. Keith makes a good point: magazine reproductions are several steps removed from the original prints. It should be no great surprise if there are significant discrepancies between the reproduction and the real thing.

That said, when you're doing your own testing, do you pick a couple of 'tie points' as I suggested above - which might necessitate changing paper grades if you're comparing negs where different development yielded different CI's? Do you go with Jorge's approach: develop to identical contrast index, print on fixed paper grade for identical times? Do you use a different approach? Are you as concerned about overall print brightness as with overall contrast? I'm interested in hearing what other people feel rates as an apples-to-apples comparison when comparing prints.

Chris

[Nick_3536]

Expose two negatives of a scene you know well. Process and make the best two prints you can. Look at the prints. Pick the one you like best. While I'm sure getting hard numbers will reduce bias but if in a blind test one print looks better to you isn't that the only thing that matters? I can understand if you're doing the test for others the need to produce hard numbers but if you're testing for yourself then shouldn't the test match your wants?