Dealing with under/over-exposed & -developed negatives; scanning or enlarging

[koraks]

Even if a digital camera can capture better images, how can we be sure for this?

There's plenty of knowledge about this generated through the past 3 decades of research. The long & short of it is that camera capture can be perfectly adequate at a very high level of image reproduction. Only for very particular purposes like conservation there can arguably be a need to move beyond regular RGB camera capture.

Generally, we must calibrate the camera and workflow based on the widely recognized results from darkroom prints and traditional scanning.

Calibration assumes a golden standard/baseline, which does not exist for color or B&W negatives.

 

[250swb]

WRONG. We're discussing negatives here. Dedicated film scanners are made so as to be able to scan a wide range of film, very specifically including color transparencies. The density range of those far surpasses that of B&W (let alone color) negatives. Hence, the dynamic range of a film scanner is very high comparative to the density range of a typical (even overexposed and overdeveloped) B&W negative.


The camera has no user-adjustable hardware gain and thus by default is set to a relatively wide dynamic range (it needs to capture a typical daylight scene after all). Since the gamma of a negative is well below 1, this means that even for a wide-range scene, a negative will only occupy part of the dynamic range of the camera sensor. In practice, this is not really a problem given the relatively clean (noiseless) signal from the camera especially if it's used in RAW format (yielding a bit depth of 12 on 15+ year old cameras or more on modern ones). Moreover, it's doubtful whether a scanner performs much better - and in practice it can be argued to do much worse, since hardware gain may not be available either (instead, software 'gain' is used and simply truncates the available signal depth) and sensor noise performance on typically 20+ year old scanners is likely to be far inferior to that of a modern camera CMOS sensor.

There's a lot of hearsay, misunderstandings and incorrect assumptions underlying amateur scanning practices (and some pro conservation scanning as well). What's particularly frustrating is that these falsehoods keep being rehashed on forums, YouTube channels etc. without being challenged, and thus, people keep making choices that go counter to their aims, and spend countless hours trying to reinvent wheels that have been around in perfectly usable form since about the 1990s.

You appear to have changed the topic from 'difficult negatives' to a generalised discussion? Have you ever owned a film scanner? Serious question because making a low contrast scan to get the maximum tonal range is a tried and tested technique. The reason is because film scanners left to themselves fall off the cliff at either end of their dynamic range and respond less well to deep shadows or bright highlights in the negative than recording mid tones. With 'difficult' negatives (or slides) this a exaggerated, and this is hard to believe I know, but the scanner software has no way of knowing what the negative is like. Amazing isn't it, the scanner is actually a dumb instrument, so the photographer has to do something, and it seems that is what you are objecting to. It is not re-inventing the wheel, it is acknowledging that wheels now have tyres on them, things have moved on.

 

[250swb]

Professional film scanners have a much larger dynamic range than needed for negatives, as they were originally designed to extract maximum detail from slides. Some so-called “professional” scanners are actually modified flatbeds, I suspect you might be referring them. Many features in new digital cameras, like pixel shift or multi-exposure or multiple native ISOs or 16/48 bit images, were standard in older film scanners, which didn’t have to account for shutter speed or subject movement.

It's an interesting idea that you think the OP has a professional film scanner, otherwise I think we have to live in the real world. A slide has a typical DR of 4-5 stops, a negative can be 15 stops, so it doesn't actually need a professional scanner to extract 4-5 stops of information.

 

[Yezishu]

It's an interesting idea that you think the OP has a professional film scanner, otherwise I think we have to live in the real world. A slide has a typical DR of 4-5 stops, a negative can be 15 stops, so it doesn't actually need a professional scanner to extract 4-5 stops of information.

You may want to double-check whether you’re confusing information recording capability with display capability.
Professional scanners often handle slide films with Dmax above 4.0, and even desktop scanners can measure up to 4.2–5.0.
As for negatives, their transmittance light does not reach 15 stops. Their 15-stop information is compressed can only be realized after printing to a positive. As PE noted about Kodak’s process, positives for cinema can reach densities of 0.2–4.0 or higher, but their oringinal negatives typically have a much smaller density range, with Dmax around 3.0.

WHY does slide film have a narrower exposure latitude?

If some of the "experts" could explain it it would be nice, i have been wondering for a while, why is it?

www.photrio.com

 

[koraks]

Have you ever owned a film scanner?

Yes. I also understand the difference between essential concepts like film dynamic range and transmission silver/dye density. I think this discussion only makes sense if you understand those basics first.

Also, you might want to define 'professional film scanner' so we're all on the same page. I for now assume that most of us work with anything in-between consumer flatbeds like Canon 9000f and Epson v-series scanners up to old, salvaged drum scanners, with the majority using Epson flatbeds or dedicated 'pro-sumer' film scanners like the Minolta Scan Dual series, Nikon LS series and more recent Plustek units. In terms of dynamic range, they all cover a similar bandwidth and for all of these devices it's true that we do not strictly speaking know how gain control is done, i.e. through which combination of hard- and software measures, although some hints can be derived from e.g. noise profiles or scan speeds.

You appear to have changed the topic from 'difficult negatives' to a generalised discussion?

I'm OK with discussing general scanning technique in relation to subjects like dynamic range and noise performance in the Scanning subforum. I was responding to an erroneous assumption on your part that resulted in a factually wrong recommendation within the context of this thread, so I feel it was relevant.