Low Reciprocity Failure Film

[Two23]

Kirk beat me to it. I shoot both Delta 100 and Fp4+ and always have a copy of the reciprocity curve with me. For a meter reading of 50 seconds I read the curve at 55 seconds.

Dang, you're making it harder to decide. Should I go with 55s, or 51s? Or maybe split the difference and go with 53s? Too complicated.


Kent in SD

[Leigh]

For a meter reading of 50 seconds I read the curve at 55 seconds.

For which film?

- Leigh

[Two23]

I'm kidding around. A few seconds won't matter either way.


Kent in SD

[Doremus Scudder]

I hate that phrase. It's not really a failure ...

Actually, it is... For photographic materials, "reciprocity" is the inverse relationship between the intensity (aperture) and duration (shutter speed) of light that gives the same density. As shutter speeds get longer, this relationship "fails" to predict the density; extra exposure is needed to achieve the target density. I'm not sure if the number of photons needed is thus increased or not, but it seems so

Best,

Doremus

[Bruce Watson]

I apologize for not being particularly articulate. But then again, I seldom am. If I could express myself well verbally, I wouldn't need photography so much. Sigh....

When I say that "reciprocity failure" isn't really a failure, what I'm talking about is the non-linearity of the film under low light conditions. That is, you're working in the toe of the film. And the film's response there isn't linear by definition. If you increase exposure to increase the number of photons (sufficient to make a latent image) hitting the shadow areas, you overexpose the areas of the film that aren't "in reciprocity" (that is, the parts that are on the linear part of the response curve). So... under develop for the highlights, yes?

Except that the old saw of "expose for the shadows and develop for the highlights" doesn't work so well here. Because the film response isn't linear here. So what you get is shadow detail, good highlight density, but what I call "mid range suck out" because the mids are under density. The resulting tonality isn't right, and it's also not in the right place.

I've never had good results printing film that's been treated this way, but it's not for lack of trying. I spent quite a lot of darkroom time trying to figure this out, and later when I was scanning found that it's still a problem -- thin mids are always going to look a little off even after correction. I don't know why, they just do. It's weird and sort of eerie.

This is the main reason I left Tri-X and went with TMY-2. I could finally work only in the linear part of the curve while still getting the texture and detail from the shadows under that rock in the river in full sunlight. That shadow was in the reciprocity region despite my 1/8 shutter speed, while the vast majority of the frame (98%?) was perfectly exposed. It wasn't until I made the trip back and exposed that same shot again, once with Tri-X, and again with TMY-2, both of which I'd dialed in (Zone System), and both of which I developed to their own individual N-1 development times, that I could understand exactly what was happening -- why I had to struggle so long and hard to get results I just wasn't happy with from Tri-X, while the TMY-2 print just leaped out of the paper with hardly any work at all.

It took me another couple of months of work to be sure of what I was seeing. But indeed, it was because of that dip in the low mids in this case because of working in the toe of the Tri-X curve. Those parts of the image, few as they might be, and scattered about in no reasonable pattern, just drove me crazy. And with TMY-2, they didn't, because they weren't in the toe anymore, and gave a linear response. The resulting tonality from the TMY-2 negative was just spot on.

Probably I'm one of the only people on this board to be bothered by this, or to have worked to figure out just what the heck was driving me bonkers. So it probably just doesn't matter to most of you. Fine, so be it. But just because it doesn't matter to you, doesn't mean it doesn't happen.

So my message is, if you're working in low light conditions, consider the implications this has on film response, even if you "correctly" compensate for both exposure (shadows) and development (highlights). And that is: you can correct for the two end points, but you can't correct for what happens in the mids. You can't alter the film's response curve; that's a decision made by the film designers and baked in during manufacturing. The only way out of this situation is to use a film that lets you work in the linear part of the response curve, at the light levels you have available.

[Kirk Gittings]

So my message is, if you're working in low light conditions, consider the implications this has on film response, even if you "correctly" compensate for both exposure (shadows) and development (highlights). And that is: you can correct for the two end points, but you can't correct for what happens in the mids. You can't alter the film's response curve; that's a decision made by the film designers and baked in during manufacturing. The only way out of this situation is to use a film that lets you work in the linear part of the response curve, at the light levels you have available.

All true and I get the point unless and until you scan the film.

[Drew Wiley]

Not really, Kirk. If something isn't on the film to begin with, adjusting the curve in PS won't change that. Besides, there were ways of doing that long before PS,
albeit rather tedious. It things are a little bit off, that's one matter; if the film choice is simply out of its league to begin with, that's another. Or by salvaging the curve, some other property might be compromised. Kinda like having a temporary spare on the car - yeah, you'll get down the road, but how far? What makes
reciprocity failure itself even more complicated is that spectral sensitivity and hence contrast filter factors can change dramatically at long exposures versus normal ones. This fact might not matter to some, but might to others. The only way to really know what is going on is to test, test, test, because there are darn
few if any printed references that tell you much about such sensitivity shifts.

[Michael R]

Bear in mind flare can be an important variable since it affects where the lower exposures fall on the film curve. Flare both raises the lower exposure values, and flattens local contrast - ie the very opposite effects of low intensity reciprocity failure.

You can't know what you're getting with any kind of precision.

[Bruce Watson]

All true and I get the point unless and until you scan the film.

I mostly scan my film. Do my own drum scanning. And I can't get it to work to the point I'm satisfied with the final print. So I'll have to disagree with you on this.

At least two things happen when you force the image into the nonlinear part of the response curve (and try to correct for it with exposure and development). First is the pushing of the mids into the lower mids and upper shadows. Building masks in a photoeditor to find all this is, well... problematic. And then, it's not a linear push. It's following the nonlinearity of the toe of the response curve. So you can break it up into lots of masks and move them each by somewhat different amounts depending on where they are in the curve. "That way madness lies" -- Shakespeare

Second, there's some compression of those tones that takes place. If sufficient compression takes place, it may be difficult to impossible to uncompress them. And that uncompression has to take place in conjunction with moving them back to where they should be (overlaps, feathering, etc.).

So while digital "recovery" in post might be theoretically possible, it's not easy (but maybe there's some nice algorithms out there by now that can do this, IDK), and for me at least it's not practical. It's way more work than just using a film that you can keep in the linear range of the response curve in the first place.

I guess I'm making an argument for knowing your craft, rather than fix it in post. I just wish 1) someone had told me, or 2) I had figured it out much sooner. I wasted way too much time chasing my tail trying to make my "favorite film" do what I needed it to do, when in fact that wasn't even possible.

[Andrew O'Neill]

For me, with FP4+, an indicated exposure of 15s should be increased to 32s from data that I generated several years ago. Manufacturers' data is usually a bit over the top. Acros is the king, followed by TMY-2. I'll be picking up a few boxes of Acros when I'm in Japan in March.
By the way, I'd love to see the photos of the old hotel, Kent!