Hey, guys!
This is more of a philosophical question thank anything else, but it is one I’ve been asking myself since the beggining of my film journey. Is there such a thing as an “objective” film negative => digital conversion, and if not, is there a “most objective” one? (read on please)
DSLR or flatbed digital scanning of negatives is obviously something posterior to the analogue film process. It’s a method that allows to approximate the original intended results via digital emulation of physical properties (like shining through a negative with light and recording that light on paper). By this definition, it is clear why asking my question is inherently pointless. A digital representation of a physical image will always be that - a representation, however good it may be. No point trying to objectivise it, one should embrace the wide range of possibilities it offers to instead subjectively change and elaborate upon what has been provided through the development process.
However.
When I’ve been trying to get my colour DSLR scanning workflow underway, I was often plagued by this: people all the time say that certain stocks “look” a certain way, that over- and under- exposure does this or that… but with so many variables changing, exposure correcting being so flawless and stuff like that, these subtle differences are often smeared to the point where I think they’re a blemish, not a feature, and compensate for them. That in turn makes me unable to appreciate the medium itself and forces me to sort of create anew, to interpret the negative in an obtrusive and overedit-y way (yes, a “me-problem”, I’m aware). I’m looking for a way of making this process as unobtrusive to the original physical negative as possible.
For example, although I’ve been getting my images done with darktable, I’ve followed this helpful tutorial: Colour-Accurate Conversion of Negatives using GIMP 2.10 and RawTherapee 5.8 - PentaxForums.com . It’s a very automated process that leads to very uniform results, ones that are semi-free of any shifts and that don’t take any input from the human operator (if you tried, you could program the entire process and it would do the same thing as you would each time). I then tried get close to an image thusly generated with darktable in order to have a reference as to what the original tones of the negative actually are. You may think me naive (and I probably am naive), but in order to evaluate if, for example, I like one type of film more than another in a certain situation, I need to evaluate based on as unaltered of a base picture as possible. I’m unable to acquire this, because in my darktable workflow, I always find myself tweaking and correcting (usually to a point of over-correcting) everything. Not even after watching countless tutorials am I able to produce consistent results, and by that I don’t mean consistently nice or beautiful, but consistently representative of the film I’m shooting (to be frank, my photos just looks plain different to what other people post and exhibit. I don’t hate that fact, i hate that I don’t know why so I can’t emulate it if I ever want to).
Have you ever had this problem and found ways to mitigate it using darktable? Is there any semi-automatic process one is able to follow to get uniform results, provided that one has the well-exposed, well-developed and well-scanned negatives? Do you find the method highlighted in the link I sent “objective” in any way or is it just that it’s a sort of approximation that delivers uniform results (and it does) while also being unrepresentative of the original stock and negative image?
I’m interested in what you guys think about this!
My background is film and developing and printing film. I have printed millions of color images. Please take no offence when I say that you are mistaken to believe a film’s color can be objectively portrayed. Each manufactures film had characteristics (problems), each manufacturers processing chemicals and printing papers had characteristics that differed from their competitors. The closest you could get to objective color is color slide film presuming that you didn’t push or pull the processing of such said film. Photography is not objective in my view. It is a subjective medium. Just look at the PlayRaw category.
Well, colorimetry is a science…
You could get an objective measurement of the film colours (and the orange filter background in colour negatives, and derive an objective inversion function.
But that does not guarantee decent or even consistent results, between differences in film make, original exposition (light temperature plays a role), processing and storage of the negatives, and differences in digitising method.
Film records light in color-sensitive layers, usually more than three. Neighboring layers can inhibit or strengthen each other. It’s an astonishingly complex process.
Development of film is standardized, but still varies with the age of the developer, and various environmental factors. Local density and diffusion changes the development process across image areas depending on the image content.
Film used to be printed with another photochemical process, very much like the original negative, with all the same variations and vagaries.
Nowadays, film is scanned, and then printed on an inkjet printer. This is more repeatable than the photochemical process ever was. But it’s still subject to the color filters in the scanner and the built-in post-processing in the scanner and printer.
What even is “objective” in all of that? Frankly, it’s a miracle that this whole process manages to be close to reality at all.
More practically, I find that I can usually define a darktable style that can be applied to an entire roll of film. But there are stark differences between different film stocks, that go beyond what I’d consider acceptable in a realistic digital rendering. Professional film scanners also include white balance shifts and exposure/contrast adjustments that are set manually by the operators. I therefore conclude that there is no “objective” negative conversion. It’s always a “creative” rendition.
(Of course you can profile and correct your process with a color chart, but that will negate the film characteristics, which does not seem to be what you’re trying to achieve.)
@Terry Hey! Thanks for the answer. I don’t think I should’ve used the word “objective”, thinking back. I should’ve used “faithful”. I understand what you’re telling me, but I also think there’s stuff to inquire about. I think just because a scanned photograph can be subjectively altered (and most often, it is), it doesn’t mean there isn’t an underlying set of characteristics that are inherent to a film stock or other. That set of characteristics, I’d think, translates into the scan in some way and it’s only a matter of editing skill and keen eyeing to make it out. That’s what I’m trying to get to.
@rvietor You’re right in supposing I won’t go as far as what you’re describing. The “points of divergence” where your final shot can differ from mine, even though we both shot kodak gold, are indeed overwhelming. But talking like that, it would seem there is no way to even know what kodak gold looks like because everyone would produce something different! Yet that is clearly not the case, therefore there must be some common ground. That’s what I’m trying to reach, or understand what it is in order to know I have reached it.
@bastibe Hey, Bastian. Your question “what even is ‘objective’ about converting the negative” is a good one. The negative is what is objective, and the processing of it isn’t, is what you’re saying. I have trouble subscribing to the idea that in it’s entirety, the quality of the final image is a product of creativity alone. And I still believe those qualities somehow translate to the scans.
I’ve gone through what you’re saying and you’re all probably right, I admit aiming at objectivity when working with film negatives is not something feasable and misses the point of the process. When a lab scans film for you and sends you the files, it can differ from lab to lab or coincide because they used the same scanner and process etc. Maybe my problem was that from the beginning, I assumed the way the labs do it is just different from what I do. That there isn’t any human input apart from “setting the machine up right”. I guess I’d like to approximate this method of conduct more, to eliminate me having to choose if I find the whites a tad bit red or not and if I should correct for it. I just don’t like correcting for much since I always get that creeping feeling that I’m correcting something that isn’t wrong. I think I’ll instead try to emulate results I like, for example: get lab scans AND negs, scan the negs myself and try getting them as close to the lab versions as possible. That could be a good exercise, not in achieving objective conversions, but at achieving relatively improved ones 
Different film stocks definitely have their characteristics, you’re on the right track there. The way the colors interact with each other, the way colors behave at different exposures, the tone curve, these all differ between film stocks.
I think I understand what you’re getting at. Film seems to effortlessly provide a pleasing render, and sort of absolves the photographer from the responsibility to “fix the colors”. Not unlike the various picture styles in Fuji cameras (or various others), but honed by countless engineers at Kodak and Fujifilm, back when color film was a billion-dollar industry.
The fact that things feel wrong means that you have an innate sensibility of what looks good to you. So there’s an implicit goal that you strive for. I found it useful to collect examples of what looks good to me. Try to analyze what it is. Is it the soft rolloff of highlights, that some negative film stocks provide? Is it the harsh blocked shadows of some positive films? The vivid dream-color of Velvia? The saturated shadow-browns of Kodachrome? The subtle pastels of overexposed Portra? The bold greens and cold hues of Superia? The golden warmth of Kodak Gold? Or even the alien colors of Aerochrome or cross processed film? Film is as varied as digital post processing, and you must match film stocks to situations in order to make best use of them.
Once you figure out what kind of rendering you like, it’s usually fairly straight-forward to replicate from a digital file. Or you could even take a shot of a color checker, and run darktable-chart to recreate that look in darktable. But at the end of the day it’s not about the characteristics of a particular film stock, but the characteristics that you enjoy personally.
I owned and run labs for many years and I can tell you that each lab produced different results. Some were just different and others were just crap. Many of my clientele were professionals and once a representative from one of the photographic film/paper suppliers looked in my waste bin and said what I am throwing away others labs were struggling to achieve.
You want to achieve faithful. Then to me that would mean a white dress looks white, a grey jumper looks grey and we trust the rest of the colors also reproduce faithfully. That is as close as I could hope for.
Thank you, immensely helpful response. I now understand it’s not the objectivity I strive for but the subjective objectivity sorry to make all of this difficult for all of you with how I think and speak, it was nice reading your responses. Cheers!