This is a first glance at the negatives I did yesterday on this film:
Above is the negative as scanned. Note that my Epson scanner (like virtually all scanners) does a lot of auto-exposure magic that
cannot be disabled in the Epson scan software. So there's no absolute benchmark here; we'll have to interpret everything relative to how the engineers at Epson some 20 years ago decided we should be scanning film.
Shown above is the raw scan of the strips (Epson 4990 film holder). Scanned flat as positive and then used the Curves dialog to approximate a neutral color balance. Starting points in this curve adjustment are:
1: No lopping off of data on the left side of the histogram; these are the dense areas on the negative, so the highlights of the positive.
2: Truncate the shadows just after the bump of the violet hue of the film base itself, which is visible as a little bump/peak in the tail-end of the histogram (see illustration above, shown for the red channel).
This gives at first glance a very decent starting point and no glaring/major issues, apart from a very muted color palette. This is to be expected given the exceedingly drab conditions under which these were shot.
Taking one frame from the set and applying some entirely subjective adjustments to approximate the scene as I experienced it when I photographed it:
* Steepened the blue curve a tiny bit. The datasheet of this film suggests especially a significant blue/yellow crossover, but the adjustment I did here was really less than what I'd expect on that basis, so I guess most of it has already been 'dealt with' by the scanner and its software. It's a black box and there's nothing I can do about it.
* Took out a tiny bit of green from the midtones only. Just a matter of taste/subjectivity. Or maybe it does match with the stronger green response that the characteristic curves show - but this seems pretty much constant, so I would have expected it to be filtered out at this stage already.
* Applied an inverse S-curve on all channels. This does match up very well with the datasheet of the film, which shows a very distinct toe and shoulder. Countering this with an inverse S-curve does indeed bring the image much closer to natural in my view.
If I apply the adjustment curve above on all images, they seem to work out as pretty neutral.
What I take away from the film at this point:
* Scanning the negatives gives a reasonable starting point to take the results in whatever direction you'd want.
* There's nothing inherently outlandish so far. I've evidently not shot any saturated colors yet and that's where things will likely change, but for muted, neutral scenes, it's perfectly feasible to get a realistic looking result when scanning the film.
* It's a grainy film, but this flatbed scanner doesn't really bring it out, so it's not very conspicuous here. However, viewing the negatives with a loupe shows a remarkably gritty appearance. Prints and scans with a proper film scanner (I might make one or two in a minute) will bring this out, I expect.
* It's very contrasty. This is the only firm thing I'd say so far. But in the digital domain, that's all relative, since scanners easily deal with slides, and it's not like we're looking at that sort of density ranges here.
What this mostly shows, of course, is that you can take this whatever direction you want. For instance, I've seen some examples so far that lean in this direction, but I find I need to twist the curves pretty dramatically to get this kind of result, suggesting it's more of a choice than a given:
This is a very dramatic overall S-curve and some pulling and twisting of the separate color channels. But it's certainly not the case that this film gives these sorts of images "out of the box". If I take the same image and give it the adjustment curve I used to make the landscape pic sort of true to life, I get this:
Which is pretty much as it looked in real life.
I might do some wet prints tomorrow; as scarce as daylight is this season, I'd rather not block out what little there is.