Over Developing Delta 100 For More Contrast

[braxus]

Matt. For you.

 

[Andrew O'Neill]

What do you mean by sharper? In common developers, D100 has better edge acutance than TMX, but not quite as much detail. With the kind of special developing I use, TMX excels in both categories, and I'd place D100 about midway between TMX100 and TMY400.

Yes, Drew. Acutance.

 

[Huub]

In Xtol straight (only gives times for straight), it says 7.5 minutes.

My XTOL datasheet J-107 gives 10.5 minutes for 20 C at 1+1 for a contrast index of 0.58, 13 minutes for 1+2 and 15.5 minutes for 1+3 at the same temp and contrast index. For a contrast index of 0.65 the times are 12 minutes at 1+1, 15 minutes at 1+2 and 17.5 minutes at 1+3.
When using the 1+3 make sure you use at least 100ml stock solution per film, otherwise there won't be enough developing agent in your tank to get your film fully developed.

 

[koraks]

Chemical contrast gives you the separation without that trade-off.

Only at an optimum. Scanners struggle esp. at high densities; this is where s/n ratio drops. Overdevelopment of B&W negs to the extent @braxus has done makes s/n performance worse than it has to be, esp. in the highlights.

W.r.t. curve shape & acutance: in both regards, the digital domain is more flexible; the drawbacks you attribute to digital either don't exist or are easily worked around, and, again, with more flexibility than in the chemical domain.

Sorry, I don't see the advantage. Everyone should of course do as they please. There's that. If this is about the magic of the manual & chemical process, then sure, I get it. But let's separate that from the more rational approach. One isn't better than the other, but let's not confuse them and somehow pretend there's a very rational argument underlying what's happening here. There isn't.

 

[albireo]

@koraks, a few points:

On scanner S/N at high densities: yep, scanners struggle at high densities. But my experience and discussion above was specifically about black and white negatives (Delta 100 - remember?). You've said yourself, multiple times, on Photrio, that B&W negatives don't reach densities that are problematic for scanners, something you're conveniently forgetting just for the sake of argument, just above. You've cited logD ≈ 2.0 as typical Dmax for a normally-developed B&W negative many timesbefore. So let's do some middle school maths. Mild push of a low-contrast scene (which is at the core of what we're talking about here, I'm not talking about 'pushing Fomapan 100 to 1600 in full sun'):

- A flat scene: ~5 stops of brightness range → log exposure range = log₂(2⁵ = 32) × [log₁₀(2) = 0.301] = 1.505
- Delta 100 at normal development, gamma ≈ 0.62: density range = 0.62 × 1.505 = 0.93
- Add base + fog (~0.2): highlights land at ~1.13

Now push to N+1 (gamma ≈ 0.75):

- Density range = 0.75 × 1.505 = 1.13
- Highlights land at ~1.33

Even if we take a full-range scene (7 stops) at N+1:

- Log exposure range = log₂(128) × 0.301 = 2.11
- Density range = 0.75 × 2.11 = 1.58
- Highlights land at ~1.78

So mild push of a flat scene puts highlights at density 1.3. A full-range scene pushed one stop still sits under 1.8. So below the ~2.0 you yourself have cited as unproblematic. Dedicated scanners such as the Coolscans I use quote Dmax capability of 3.8–4.2. We're nowhere near that territory, by your own numbers.

are easily worked around, and, again, with more flexibility than in the chemical domain.

Noted, and I'm pleased you like putting your copy of Photoshop to good use, but without specifics this doesn't engage with the mechanism I described. Adjacency effects are spatially varying, exposure-dependent, and scene-adaptive. Just letting me know that digital is "more flexible" is not a counter-argument. I honestly see just a claim without support here.

On the "rational vs. romantic" framing. I gave three specific mechanisms with physical explanations. You responded to one (density/S/N) with a point that spectacularly contradicts your own prior statements, and to the other two with "easily worked around." I'm sure you can do it all digitally. That's not the point. The point is that you dismissed development control for hybrid photography as "moot", which it isn't, as the numbers above show.

That other people choose a different workflow to yours doesn't make their reasoning irrational. I'm not here to convince you of anything, just to correct the record for anyone reading who might otherwise take your unfair and unsupported "overdeveloping is a moot point if you're in a hybrid workflow" at face value.

I'll leave it there, I don't think further back-and-forth on this is productive, and we've drifted well off @braxus's original question.

 

[koraks]

The point is that you dismissed development control for hybrid photography as "moot"

Yup, it is in many cases, especially those where people don't have a clear concept of what they're trying to optimize in an objective sense. That's the case here.
There are instances where people do have a solid conception of what they're doing, and it's those instances that appear to answer to two characteristics:
1; they usually involve arguments that aren't really about technical optimization, but involve meta-level arguments involving non-rational (which to me is something different from irrational) concerns.
2; there's generally a subjective preference involved for solving the puzzle one way or another; e.g. "I know I could do it like X, but Y gets me where I want to be quicker within my personal preferences and limitations".

Again, no contest that people have preferences and that's OK. What I do want to inject in discussions like these, at least occasionally, is that the tendency of some to try and 'sell' one workflow or another based on quasi-objective grounds is virtually always problematic. It's just not a linear optimization problem. It's one of personal preference, involving a rather fuzzy set of criteria that in part depend on highly personal considerations, including know-how and competence. This leaves the possibility that somebody is trying to do something in a very roundabout and possibly inefficient or even ineffective manner, while a more suitable approach would be feasible. Whether that approach is considered attractive, is for anyone to decide.

Now, your 'objective' arguments in my view just do not convincingly make a case for chemical optimization, or against doing it digitally.

Take the s/n ratio argument, for instance. If you think that one through, which goes beyond the simple calculation you show, you realize that you have to find an optimum between several factors that degrade signal quality; in particular sensor (thermal) noise in the scanner vs. film grain. It turns out that the optimal is dependent on the scanner, probably on scanning settings, film used, scene brightness range and to a lesser extent the developer used. Then there's the question of what the penalty (qualitative and quantitative) is of deviating from the mathematical optimal. Is it a steep decline on either side, or a gentle slope? I'm willing to bet a good bottle of wine that you haven't done the testing, and taking a (sensible!) and pragmatic approach to it and take a ballpark guess at it. Fine, but if you're taking an objective approach, you'd have to test a number of combinations, quantify (and/or qualitatively operationalize) the results and systematically compare those in order to find a decent optimum. Which, of course, turns out to be a compromise if you're working with roll film - after all, SBR was part of the mix...

As to curve shape - I'm not sure how there could be a discussion at all. Curve shape is so ridiculously easy to adjust digitally, whereas it's challenging to control (with precision) in chemical space. That doesn't mean the latter can't be done, or that over- or underdevelopment (there are divided camps on this, it appears) cannot produce very pleasing results for someone. I just don't see how it would be a superior approach, in an objective sense, in comparison with an approach that is so flexible, that can be repeated with variations without altering the original, that can be applied with local variation, etc. It's literally like split-grade burn/dodge printing, just quicker. What's not to like about including that option? How is the alternative more attractive if there's no compelling need to go there (e.g. contact printing in a process that doesn't allow for contrast manipulation etc.)?

The adjacency effect claim is problematic for two reasons.
Firstly, as I said before, it's contestable (IMO plain wrong) that it's somehow better to optimize this in film development. This is only true if you have no other (feasible, attractive) means downstream in the imaging chain. Especially when scanning, you do, since all manner of manipulations of contrasty edges (USM and related techniques) are readily available. Moreover, they can indeed be applied in various modes, degrees and with local differentiation. As to this allegedly being an unsupported claim - please refer to the zillions of online guides, videos etc. on sharpening digital images, on how to work the USM, local contrast etc. adjustments and what the various parameters do, how to work with masked layers, etc. The toolbox is virtually unlimited. Since you're adamant that 'your' way is best, I trust you're aware of the alternative? Secondly, since we're in this context discussing roll film, there's the issue of having multiple scenes on the same roll for which it's arguably very often (usually) not the case that 'one size fits all' presentation choices would work. In conjunction with edge effects being something that cannot be removed from a negative (they're analogous to a 'destructive' edit), there's a very valid argument to postpone such decisions to a later stage in the process. Furthermore, I found the inclusion of edge effects in your initial response a bit opportunistic since the story started with simple lengthening of development. I trust you're aware that edge effects are typically controlled/achieved with the combination of reduced/limited agitation and a highly dilute developer that's working on the edge of depletion. Braxus is using straight XTOL...the opposite of what you'd use if you wanted to create pronounced edge effects. Simply developing longer in XTOL (etc.) 1+0 won't produce such effects. Drifted off of the original question, indeed!

I don't doubt that you're getting somewhere you like with your images and your workflow. That's great, honestly. In this particular case, I see someone using Delta 100, they're presenting us with a number of images in which there's apparently (according to and paraphrasing OP) an issue with suboptimal quality of midtones, and they're using a hybrid workflow. In that context I think it's fair game to remark that overdevelopment just isn't going to fix the issue, and insofar it is, the same effect can be had by postponing such adjustment to a later stage, when it can be experimented with in order to optimize on an image-wise basis. To put it very bluntly, if someone shows the images in #11 and states there's a problem with mid-tones, and they allege that they're trying to solve it by overdeveloping the film, I think there's a very good set of reasons to point out they're looking in a direction that holds very few solutions for their specific problem.

 

[albireo]

@koraks , I'd find the "just do it digitally" argument more persuasive if your posted results demonstrated the superiority you're describing. You have an unlimited digital toolbox, decades of experience, and very firm opinions about the right way to do things - and yet, what you routinely post here doesn't reflect any of that advantage. Maybe the chemical domain isn't quite as "moot" as you think. Something in your chain isn't working, and I'd gently suggest the issue might not be on other people's end.

Aside from that, the density contradiction stands. I'll leave it to readers to draw their own conclusions.

Goodbye, this time for real, and enjoy your photoshopping!

 

[koraks]

Wow. That's just lame.

 

[pentaxuser]

Are you sure you didn’t intend to post this in the thread of (mostly incorrect) hot takes?

Must have been a thread I missed but based on Drew's post and Andrew's reply it looks as if no-one died and Ben Miller and Will Kane called it an honourable draw.

pentaxuser