Curving Out "The Look" of an Emulsion

[mkochsch]

I preparing to print some cyanotype using Mike Ware's New Cyanotype formula (if you haven't tried it you really should, it's stunning, on the right paper at least). Anyway. I'm looking at my Stouffer 31 wedge and I notice that the highlights really falloff slowly -- unlike say gum or silver. I'm hitting around 2.2 logD for my range. But I'm wondering, and maybe someone with some Pt/Pd/Zia (or other) experience can chime in here, if I'm creating a curve to totally linearise the emulsion across the 101 Stepwedge am I also detroying some of the character of that emulsion? Is my New Cyano going to look more like a Silver print rather than having delicate highlights that would appear if I wasn't using a digital negative? Is any of this making sense?
~m

 

[dwross2]

You're making perfect sense. I have been having a similar niggle in the back of my brain, just not nearly as well thought out as your question. We're all in pretty new territory here. We're pulling together bits and pieces of the very old and the very new. It's all together too easy to fall back on comfortable and homogenous standards. Craftsmanship is (or should be) more complex than 'perfection'. Now, did that make any sense?

d

 

[Katharine Thayer]

Michael, you ask a very good question, one I've puzzled over for some time.

A year ago or so, I went through a very intense digital negative period where I didn't print anything but calibration prints and step tablets and color charts and the like (and drove Michael crazy, no doubt, by writing him and asking him dumb questions all the time). At one point during this period, I stopped at my local photo supply store and asked about ordering something I thought might help in the process, I forget what it was. Some sort of color target or checker or something; at any rate, whatever it was, cost over $300. Chuck, the proprietor of the store and a very competent photographer and teacher, asked me what I needed the thing for, and I explained to him about digital negatives, and how the idea is to make a curve that distributes the tonal scale across gum's short range. He looked at me as if I had lost my marbles, and asked, "What in the world would you want to do that for?" I said, rather lamely, because I couldn't think of a better reason for it, "well, it's what everyone is doing now, and so you kind of feel like you have to try it, anyway."

He guffawed and went off on a long rant about how stupid a thing that would be to do. I should have taken notes, because it was quite funny, but the basic idea was that my gum prints don't look like any other photographs in the world, and that is what is wonderful about them, and why would I want to try to match the tonal scale of other photographic processes; wouldn't that defeat the purpose of printing in gum? "If it ain't broke, don't fix it!" he said, and after going on like this for 15-20 minutes, he asked, "Have I said it enough different ways to make myself clear?" When I left the store, he called after me, "Don't you dare change a thing about how you make your prints."

I've thought about that off and on since, although I've also kept on playing around with the RNP and ChartThrob things.

I've noticed, in looking at online examples of gum prints, that the gum prints that are made using digital negatives to distribute the tones this way don't seem as interesting to me as the old style, more creative, less photographically tonal gum prints. I think that was what Chuck was trying to tell me: Why would we want to make gum prints look like silver prints? I'm not asking this question as if I thought I already knew the answer; I don't know the answer, at least I haven't answered the question to my own satisfaction yet. But I think it's a good question, and I'm glad Michael brought it up.

The nice thing about gum is that it will always resist being put into a different mold than its nature is comfortable with, anyway. So gum may not be a good example to bring to bear on your question, except that sometimes the most extreme examples illustrate a concept better than the subtler ones.
Katharine

 

[clay]

Here is what I think: make the digital negative behave like the 'real' negative that gives you the look you love in your prints.

What this means is that if you want to preserve the look is that you can either do a super-excellent photoshop editing job that mimics that final print look and then print it using a verbatim mapping of photoshop->negative->print, or just design your digineg process to generate negatives that give you the type of print look that you are trying to achieve, which may not necessarily correspond to a literal representation of the photoshop image.

That is why I was trying so hard to find a digineg process that allows me to print palladium with no contrast agent and preserve the wonderful toe inherent in the palladium process. Being forced to use a significant amount of contrast agent can end up giving you a print that is both literally and numerically correct, but esthetically unsatisfactory.

I think it sort of defeats the purpose to make palladium prints that look exactly like an inkjet print on matte paper when viewed from 18 inches. The inkjet print is a whole lot less hassle, and I have observed a noticeable lack of interest from anyone about the amount of pain involved in making a print.

 

[mkochsch]

The Emperor Has No Curves

That is why I was trying so hard to find a digineg process that allows me to print palladium with no contrast agent and preserve the wonderful toe inherent in the palladium process. Being forced to use a significant amount of contrast agent can end up giving you a print that is both literally and numerically correct, but esthetically unsatisfactory.

How much of a Palladium print's "look" was/is the influence of using a silver, or stained for that matter, negative? Surely that had an effect. And just to be clear, when you say "toe" of the print we are talking about the highlights of the print right, traditionally a product of the "shoulder" of the negative?

What I've been doing when creating curves is basically a one-to-one mapping of tones. "Fixing" the contrast so it follows in perfect mathematical order -- like an inkjet print. Seems though if I'd been using a densitometer to measure the contrast changes between steps the story would be different -- subtly different but one would see the natural characteristics of the emulsion revealed. Would it not stand to reason then that a secondary set of curves could be developed and applied before "linearisation" to reintroduce "the look" of the emulsion. This secondary set of curves would be more like a universal standard though. One generic curve for raw Palladium, one for Pt, Gum, Cyano et al. These curvae would be usable by all and, again to stress the point, applied to the original image prior to applying the "linearisation" curve. Thoughts?!

 

[Katharine Thayer]

Seems though if I'd been using a densitometer to measure the contrast changes between steps the story would be different -- subtly different but one would see the natural characteristics of the emulsion revealed. Would it not stand to reason then that a secondary set of curves could be developed and applied before "linearisation" to reintroduce "the look" of the emulsion. This secondary set of curves would be more like a universal standard though. One generic curve for raw Palladium, one for Pt, Gum, Cyano et al. These curvae would be usable by all and, again to stress the point, applied to the original image prior to applying the "linearisation" curve. Thoughts?!

Just to check how well I'm following this thread, it seems to me that this suggestion is similar to Clay's first suggestion, to adjust the monitor image to reflect the "look" you want in the emulsion, and then apply a curve to "linearise" those tones.

I'm just not able to picture what that generic curve would look like for gum, or more saliently perhaps, how one curve could possibly work for all gum emulsions and all ideas people might have about how they want their gum prints to look. So I guess I'd need to see some examples. My first thought is that this wouldn't be useful for gum, but I'm open to persuasion by a concrete example.
Katharine

 

[sanking]

The "look" of any process is not totally dependent on the process itself but is the result of a combination of factors, including the curve characteristics of the negative as well as the curve characteristics of the process. A further important determining factor is whether or not the process is self-masking.

Carbon and gum are relatively straight line processes and if you print from the same negative the final look will be quite unlike that of a pure palladium print that has a very long toe and shoulder. However, if you know in advance what look you want it is simple enough to just linearize your curve and, assuming your monitor is calibrated, adjust the tonal values on screen to what you want to see in the final print.

Another option, and this is what Mark Nelson does in Curve Calculator II, is provide hybrid type curves to warp the image on screen so that the final result mimics the look of a specific process.

I believe the first approach makes more sense for people who understand the sensitometry of their process and know what they want. The second approach seem more suited to those who work in a more automatic fashion and don't really understand how film and process curves inter-relate.

Sandy King

 

[Ben Altman]

Interesting thread... It seems to me that if one has good control of the digital side, the final print can have the traditional look of the wet process, or any other desired "look" that is within the physical capability of that wet process. To me this is what is exciting about the hybrid approach - all the possibilities that can be generated in digital combined with the hand-crafty making of alt. prints. But then I'm backing into the old-style darkroom techniques from a commercial/color/digital background, so I'm not attached to a particular process having a particular look.

I've been away from my darkroom for a few months, but here's the structure I have been working on:

- use a RIP (QTR) to make negatives that translate into prints (or step wedges) that accurately represent all the levels in a Photoshop test file.

- use a profile or a curve to set up the monitor to mimic those prints reasonably well.

- adjust the image file to produce the "look" I want for that print. In my case that is a multiple stage process - I have a fancy scanning capability that allows me to get a smooth scan that is close to what I want before I start stretching levels around in PS, and I'm shooting and processing my negatives to have a long scale that uses the best capability of the scanner, rather than matching them to a particular alt process.

So the idea is to use an integrated workflow that maximizes the potential of each step, giving the highest quality and greatest freedom of expression at the end-point, rather than one "expected" look for particular wet process.

 

[sanking]

Interesting thread... It seems to me that if one has good control of the digital side, the final print can have the traditional look of the wet process, or any other desired "look" that is within the physical capability of that wet process. To me this is what is exciting about the hybrid approach - all the possibilities that can be generated in digital combined with the hand-crafty making of alt. prints. But then I'm backing into the old-style darkroom techniques from a commercial/color/digital background, so I'm not attached to a particular process having a particular look.

I am sure you are right. With good control the final wet print can have any look you want, whether that be a desired look that mimics the use of in-camera negatives or something else. Or, from the other corner, you could make your digital prints from an inkjet printer look like wet prints from a specific process. Should not be difficult at all to create a curve and color that would replicate visually the look of long scale palladium.

Personally I don't try to replicate any specific look in my carbon prints. I simply adjust the image on screen in the manner that best suits the image and apply a curve that linearizes output.

Sandy King

 

[clay]

Nice summary in your two posts. In Hemingway speak: You get what you want.


In regard to your process statement, I was talking to Mike Kravit the other day, who used to do a whole lot of pt/pd printing a few years ago. He has gone totally digital, and used a photospectrometer to measure some of his old prints to get a very precise profile built for a matte inkjet paper that he claims is indistinguishable from a palladium print. I have not seen one, but he always seemed like a straight shooter to me. So it appears that you can make your process mimic another with a little work.

I am sure you are right. With good control the final wet print can have any look you want, whether that be a desired look that mimics the use of in-camera negatives or something else. Or, from the other corner, you could make your digital prints from an inkjet printer look like wet prints from a specific process. Should not be difficult at all to create a curve and color that would replicate visually the look of long scale palladium.

Personally I don't try to replicate any specific look in my carbon prints. I simply adjust the image on screen in the manner that best suits the image and apply a curve that linearizes output.

Sandy King

 

[sanking]

So it appears that you can make your process mimic another with a little work.

BTW, I took the QTR profile that you posted for the Epson 2200, designed for palladium and no restrainer, and designed a carbon tissue for it. Since my old tissues were based on ES of about 1.8, I had to bump up the amount of pigment in the tissue considerably so that I could print for the longer ES of about 2.8. And that of course meant using a stronger dichromate solution.

What I reasoned was that a curve that printed in a linear fashion a long toe and long shoulder process like palladium would print a straight line process like carbon with more separation in the shadows and more contrast in the highlights. Worked out pretty much as I anticipated and I have made a few very nice carbon prints using that profile, but there was an unexpected finding.

The finding is that when I plotted a curve from a step wedge print I found that the very long scale (log 2.8) carbon print had a much longer toe and shoulder than short scale (log 1.8) carbon. In looking back at some of my previous testing I can see that this is not an anomaly, but since I had always made my tissue optimized for ES of about 1.8 the difference was never so obvious.

So in some sense one gets pretty much the same thing with carbon that you get with pt/pd. When the ES is low, say 1.4-1.8, the curve is fairly straight line, but when you extend it out there is a lot of toe and shoulder. I suspect this may be due to some form self-masking with both processes, even though the mechanisms may be different in the two processes.

Bottom line is that your QTR profile, linearized for long scale palladium without restrainer, turned out to be also very closely linearized for long scale carbon.

Live another day, and continue to learn.


Sandy

 

[Katharine Thayer]

Interesting thread... It seems to me that if one has good control of the digital side, the final print can have the traditional look of the wet process, or any other desired "look" that is within the physical capability of that wet process.

I think this last phrase is key. I've been thinking, while mulling over this thread, that if gum were exempt from this discussion I'd say sure, that makes sense; by either method of generating a digital negative, one could probably achieve whatever "look" they wanted from any process. But as I said earlier in the thread, I can't make it make sense for gum.

The thing is, gum has an infinite number of "looks," one for every possible emulsion, and every emulsion has its own physical limitations that make its look inevitable. On reflection, I'm sure I don't know what my friend Chuck meant by the "look" of my gum prints, since I print with a great variety of different emulsions and thus a great range of "looks." I wouldn't know how to adjust the monitor to match the "look" of a gum print, as I have no idea what that "look" would look like. (And as I said, I'd be very curious to see what that generic curve would look like, that would replicate a "gum look.")

In general, a lightly pigmented mix will have a high-key, low-contrast look that's good for filling in subtle gradations in the highlights on a multiple print, or for printing a high-key one-coat print. But the darkest DMax of such a mix might be as low as .3 or .4, depending on the pigment and the degree of dilution; in other words, it's not going to give you any very dark values. A very heavily pigmented mix will have a high-contrast look, good for filling in the deepest shadows on a multiple print (if the pigment chosen is a dark enough pigment) or for a high-contrast one-coat print. If one chooses a fairly heavily pigmented mix of lamp black or Mars black, one can achieve a DMax of as high as 1.7 or even 2.00, but you won't get but a few dark steps from such a mix. (One can increase tonal range with such a mix by special measures such as gross overexposure followed by aggressive forced development, but those measures introduce their own tradeoffs and make the print necessarily less photographic in nature, so this exception rather proves the rule than disproves it.) A medium-pigmented mix will express the midtones nicely and include some shadow and highlight, so is the best choice for a one-coat where one wants to express something approximating a "normal" photographic tonal scale, like a silver print, but won't give either as dark a DMax or as subtle gradations in the highest highlights as can be obtained with either a more concentrated or a less concentrated mix. (Judging print densities from several test prints I've done of this type, I'd guess that the best DMax one can hope for from a mix that will give a fairly normal-appearing tonal scale, is around 1.3 or 1.4.)

I suppose a simpler way of saying all that is that by multiple printing in gum, one can easily replicate any look imaginable, but there is no gum emulsion that will replicate every possible look in one printing. Only if the tonal range of the desired look is fairly short (say 1.4 or less) can it be done, even with a very carefully crafted perfectly linearised negative. If one wants to make a print with a longer tonal scale than any emulsion can print in one printing, then multiple printings are required.

These fundamental truths of gum emulsions hold, regardless of the negative that's used. So the idea that one could simply linearise the tones on the monitor, or apply a generic "gum curve," and automatically achieve any "look" by printing the result on some undefined gum emulsion, isn't consistent with my knowledge of gum emulsions and how they work. Ben's caveat that the general idea holds only within the physical capabilities of the given wet process is well taken, and at the same time points out how poorly the idea applies to gum.
Katharine

 

[Katharine Thayer]

Only if the tonal range of the desired look is fairly short (say 1.4 or less) can it be done, even with a very carefully crafted perfectly linearised negative.

On reflection, that number may be too high even for the most "ideal" emulsions (ideal in the sense of being capable of the longest possible tonal range). I'd need to measure the DMin and DMax of a number of step tablets to get a solid estimate; it's interesting to consider that, to my knowledge at least, (and nothing comes up with a google search either) no one has ever done this for gum print tonalities.

 

[Ben Altman]

Well that taught me about five times more about gum printing than I knew before - great explanation!

It sounds as if the way I work would be way too limiting for gums. With POP Pd/Pt, which is what I've been working on, my approach requires picking a particular mix and characterizing it. So I make adjustments to the final image using the way the negative is printed, not with the emulsion on the paper. If I want to use different chemistry, I have to figure out a new ink curve in QTR to match it. So for me the "physical limitations" of the wet process are mainly the color of the print from a particular mix and the way the particular tones render in that material on that paper. The long-term challenge is to make the characterization process easy enough that I have ready access to a variety of mixes.

I suppose you could pick, say, three different gum mixes and characterize them, but it sounds as if that might lose the flexibility that you like. Although it might be interesting to try. So I guess for gum the use of digital is to get a nice neg to work with at the size you want, leaving most of the control in the wet side.

The nice thing about this stuff is doing whatever works - and learning from whatever works for everyone else

Ben

 

[Katharine Thayer]

The nice thing about this stuff is doing whatever works - and learning from whatever works for everyone else

Ben

I couldn't agree more. Thanks for comments.
kt

 

[sanking]

I suppose you could pick, say, three different gum mixes and characterize them, but it sounds as if that might lose the flexibility that you like. Although it might be interesting to try. So I guess for gum the use of digital is to get a nice neg to work with at the size you want, leaving most of the control in the wet side.


Ben

I would do it that way, and quite a number of people do. This way of working may not offer maximum flexibility but it does allow one to make duplicate type prints and, I believe, get more consistent results.

I am fairly certain that a master gum printer like Keith Taylor has special mixes that are developed in a fairly mechanical way. If he did not I don't believe there is any way he could get the kind of consistent results, especially in control of color, that he does.

I should also point out that the ability to create emulsions of many different colors and contrasts is also possible with the carbon process, and quite a bit of control is also possible in carbon with development as in gum. However, for practical reasons I adjust all of the emulsions so that they can be used with a common curve. If I did not most of my prints would be surprises and it would be virtually impossible to make duplicate prints or editions.

Sandy King

 

[mkochsch]

Curve Test

View attachment 145 View attachment 146
I went ahead and built my curve for Ware's cyanotype last night. My negative's colour was 255,80,80 incidentally. It's long highlight scale is characterised by the quick rise at the beginning of the curve. The area highlighted (in red), above the median line (yellow) demonstrates that in order to straighten (curve out "the look") the highlight scale of the emulsion the curve is subtracting density from the image at this point in the curve. At about point 24,24 you can see I've laid down a point, this is where the input and output of the curve match (i.e. 24 per cent in my 101-step original matches the output in the print, also 24 per cent) and the curve begins to add density to the negative to fill in the mid-tones and shadows of the print correctly, typical for most curves using CYMK inks using "colour" neg methods. I think one solution to saving "the look" of that long highlight scale would be to lower or somewhat flatten the curved area in red so it is close(r) to the yellow median line. Straightening it too much might cause an unnatural contrast shift in the print, but I'm thinking just nudging it down a little would be a reasonable tweak. Thoughts? BTW the area from 0-24 per cent measured about two-and-one-third stops (step 16-23 on my T3110 wedge) Q: I forget, where do highlights end and mid-tones begin?

~m