Fixing negative underexposure in Negadoctor

[Neutral_Gray]

tl;dr: How would you fix a negative you scanned properly but which was already underexposed in the analog camera?

Imagine the following case:

• Fixed scan exposure for entire film roll
• Values mutually transferred from significant source images
• Adaptions only in print properties
• Brightest image has a D max of 1.50 dB (flash reflection/light from windows), but is otherwise underexposed

According to the manual, the paper grade (gamma) value should be D-max plus 2 to 3. So we land between 3.5 and 4.5. In “correctly” exposed images, this would be fine, but here it contributes to crushing the shadow details.
Aurélien doesn’t really demonstrate how to use Negadoctor in this case - apart from the option of raising print exposure. I’d be willing to accept a “washed-out” corrected image for the sake of retrieving the lost details in the shadows - but what’s the best way to process it?

Would you…

1. decrease paper grade (gamma) (below the recommendation from the manual - often rather between 1 and 2.50) until the histogram still touches the blackpoint - or, alternatively, with a color picker checking an intended midtones area, making the histogram lose contact with the blackpoint. You may combine the latter with decreasing paper black (density correction to spread the image across the histogram again
2. increase print exposure adjustment with a color picker on an intended midtones area - which potentially requires a heavy increase (~0.8EV) and may crush the highlights despite maximum dampening with paper gloss (specular highlights)
3. leave paper grade (gamma) at the recommended value and add other modules later in the pipeline. Here are some options:

• rgb curve (very difficult to set the points for pushing up the shadows without creating extreme colors)
• color balance rgb (selectively decreasing contrast and increasing brilliance & luminance - but individual channels quickly move out of the histogram)
• tone equalizer (must be moved behind negadoctor in the pipeline, works, but the allowed gain in the shadows is often not strong enough)
• filmic rgb (unconventional but possible with a custom middle gray value, though it tends to even increase contrast and brings more complications with different older color science options)

4. leave paper grade (gamma) around the recommended value and accept the underexposure?

[europlatus]

I rarely do much with the controls in Negadoctor these days. Instead, I prefer to use the tools I’m more proficient with and/or easier to use, such as Color Balance RGB, Tone Equalizer, etc.

So, after doing the basic conversion in Negadoctor (using the pickers or manually adjusting to taste), I would first try to adjust exposure with Tone Equalizer (ensuring it is placed AFTER Negadoctor in the pixelpipe) or with Color Balance RGB. I find the latter the most versatile for working with exposure and colour. One feature I particularly like is setting the contrast fulcrum and having more control over where to add contrast.

I also find the Local Contrast module to be very good at dealing with high dynamic range situations. It sometimes adds that extra separation you need in areas of quite even tonality.

But all of this I would approach on a case by case basis. It’s hard to come up with a workflow that will work for all scanned negatives. And, of course, there’s also the question of whether your intent is to improve upon the original capture or whether you want to replicate it for archival purposes.

Here’s an example of an image taken 25 years ago where I wanted to digitize the original print version, and also process it differently:

The version on the left is more faithful to the original print. It had high contrast with the person in the foreground and mountains silhouetted. After I scanned the negative, I realized how much more detail I could eke out of it, which resulted in the image on the right. I think both versions are interesting in their own right.

[Terry (Terry Pinfold)]

Is it possible to supply a sample image as a play raw?

[Neutral_Gray]

I could do that if I find one that’s both “underexposed” and not family-related. Beyond that, I don’t know if it is allowed here to post a 240MB dng raw.

[Neutral_Gray]

Thank you for the thorough explanation! Your sample images helped me come to terms with the idea that certain images need further treatment beyond Negadoctor. To give credit, where it is due, even Aurélien mentioned that.

With digital images, I was used to the idea that you can - and often should - still modify a lot of parameters, but with most analog scans I had the impression that they simply “worked” fine. Except for some very dark ones. First, I thought that something was wrong with the scanning setup. But your images reminded me that a) the metering modes of the old analog cameras could often focus on the “wrong” area (particularly with the cheap point and shoot ones my parents used), and b) most “cheap” lab prints back in the day simply had to be thrown away as “lost” or “messed up”, whereas we can now use corrections much more easily.

One thing about which I was most uncertain was whether it was better to do modifications within or after Negadoctor. I was afraid that after the “correct” (higher) paper grade (gamma) value was applied, the information was lost in the darker areas. At the same time, I wanted to leave as many parameters unchanged within the same film roll - in order to allow for at least a semi-automated workflow. Now, after testing some options with color balance rgb, I’ve managed to get better results than by fidgeting around with the parameters in Negadoctor’s print properties tab.

This step involves reducing contrast (25-50%), increasing global brilliance (or only brilliance in the shadows) and slightly reducing color in the shadows (yes, I say this vaguely because I still need to find out more about the different types of options here).

[europlatus]

@Neutral_Gray

Your sample images helped me come to terms with the idea that certain images need further treatment beyond Negadoctor. To give credit, where it is due, even Aurélien mentioned that.

I am really happy that Aurélien created Negadoctor, especially because the old Invert module was broken. I particularly like the way that Negadoctor fits into the RAW workflow, so you don’t need to do round trips into Photoshop or convert the file format, which is what I used to do before moving to Darktable. I also appreciate how the module has been designed to give a very accurate recreation of film and the print medium.

That said, I do find the module overly complex, and the function of several sliders overlap. So it gets confusing when you just want to tweak the contrast here or there, but it’s not clear which slider is best. (I fully appreciate that it might be clear to those who have fully grasped the function of each slider). Also, from what I know, the algorithms behind the colour corrections are very similar to those found in Color Balance RGB and Color Calibration, so I see no point in struggling with all the Red offset, green offset, etc. sliders. You can just do white balance adjustments in those more familiar modules.

So I have definitely made my peace with doing the bare minimum in Negadoctor and then doing everything else in other modules. For a while, I too was obsessed with doing everything in the best way from a technical standpoint and in the right order. But I found that trusting my eyes was just as effective.

The fact is that the success of scanning and developing negatives is very dependent on:

• the original film stock used
• the original camera/lens used
• the original settings / capture conditions / exposure used
• the scanning method (camera/scanner/software/light source…)
• the condition of the negative
• your processing workflow and expertise
• etc.
• • and, if you’re trying to emulate the original print, the lab’s original developing of the negatives, which has a huge impact on the print, as shown by my earlier example posted and the one below

There are so many variables to the outcome, which is why I now take everything on a case by case basis, or at least roll by roll. I often set up some presets for a particular roll only to find it’s useless for every other roll. So I’ve found limited use for specific settings.

The biggest difference for me came with improvements to my general knowledge and expertise with Darktable rather than learning the science behind analogue, digital and print used in photography (although that has helped). I have been using Darktable long enough now that I can do almost anything I want with it, and that has massively helped me get around problems encountered.

Here’s another example of where the lab went for very high contrast when they developed the neg (left image). I remember being very disappointed with this photo when I came back from my trip 25 years ago, thinking I’d horribly overexposed it. It turns out that the negative still had tons more detail in it, which I managed to reveal once I scanned and “developed” the neg in Darktable: