Silverfast and the "new" Portra 400

[olifaunt]

Even the lastest version of Silverfast (as sold in 2019) doesn't have a profile for the new Portra 400 (by "new" I mean since 2010). The Portra profiles for Portra 400 in Silverfast are for the older VC/NC versions of Portra that have been off the market for almost a decade. None look right when I scan Portra with them. Support told me to use NC but it doesn't look all that great. Any advice?

 

[Larry Cloetta]

I love Silverfast, but there are many Negafix profiles they don’t have, likely will never have. Still don’t even have an ACROS profile, which drives many people up the wall. It’s always been the case that it’s best to just find a profile you like best, doesn’t even have to be a Kodak profile in your case, start with that and use it to make a custom profile following their directions for doing so, and use that going forward. Those profiles are just their “opinion” anyway, Silverfast has never claimed they were perfect.
Even if there is a Negafix profile for a given film stock I might be using, I rarely use it right out of the bottle.
Their “failure” to hand me perfect presets for everything initially frustrated me, but I eventually figured out I was better off modifying existing ones anyway.

 

[olifaunt]

Yes, it's frustrating because I like the look of Portra I get from pro labs (it is more or less the same across labs; I just can't afford to use them anymore) and do film in part because I dislike this kind of digital PP decisions since I am not good at making them - what looks like a good color decision to me one day will look way wrong the next day. Even their B&W profiles have color casts that take an extra step to turn to gray scale in another editor. I mean, how hard would it have been to make the B&W profiles be gray scale? I feel like it's almost as if they have a 1990s code base with hard-coded profiles that they are unable to modify anymore.

 

[Mackinaw]

You don't have to use Kodak C-41 profiles, try Fuji, Agfa, etc. And play round with the individual RGB curves in the histogram. You can make a film look anyway you want once you master those.

Jim B.

 

[Jamie123]

The version of Silverfast Ai (8) that I have for my Epson V850 does actually have Portra 160 and Portra 400 profiles (in addition to NC and VC profiles).
That being said, I usually just pick whichever one looks best. Film profiles are never going to be quite accurate anyways since the orange mask of negative film can vary quite a bit depending on how long ago it's been shot, i.e. how it was stored. I've noticed in the past that films that were shot as little as 4 months apart and were developed at the same time have a noticeably different orange mask.

What you're seeing with the scans you get from pro labs is mostly the result of the scanners they use (in addition to some tweaking by the operator ofc). Minilab scanners such as the Noritsu are well known for the nice colors they produce from negatives. In my opinion they are the closest to what I can get in the color darkroom.

If I were you, I'd see if there's a color darkroom in your area where you can rent some time and do your own prints. Believe me, it's much more rewarding than sitting in front of the computer and labouring over a scan.


PS: a lot of people are having good results with the Negative Lab Pro plug-in and it seems to work with RAW files from Vuescan and Silverfast so you might want to give that a try.

 

[olifaunt]

You don't have to use Kodak C-41 profiles, try Fuji, Agfa, etc. And play round with the individual RGB curves in the histogram. You can make a film look anyway you want once you master those.

Ah, but I could never master color adjustments in several years of trying to learn digital color post-processing, so that's a lost cause. Some people have the visual version of perfect pitch and can get consistent looks from one picture to the other. I just don't. For digital I finally settled on sticking with camera jpeg output to at least have consistency, but eventually stopped using digital because camera jpegs from modern Japanese cameras never quite look entirely right to me.

I have a knack for recognizing instantly what I like in others' work. I just can't get there myself unfortunately. When I use different negafix profiles, I might get something that looks okay for one picture. For the next one it doesn't look right so again I get an inconsistent body of work.

Is Vuescan any better for Portra 400?

 

[olifaunt]

The version of Silverfast Ai (8) that I have for my Epson V850 does actually have Portra 160 and Portra 400 profiles (in addition to NC and VC profiles).

Okay, that makes zero sense though. Why would supported profiles be different depending on scanner? Have they not heard of modularity in software design? The scanner doesn't care which film you load in it. So Silverfast acts, for absolutely no logical reason, analogously to having a camera that could be used to take pictures when loaded with only old Portras, not the new Portra. Nobody would ever design a camera that way.

Another strange thing is that scanning the same negative using the VC profile in Silverfast will give a more vivid scan than NC. If you think about it, that's the opposite of what it should be, almost as if NC/VC were not different films stocks but simply different printing instructions for the same film stock. Maybe that was Kodak's business model, slapping different labels on the same stock, who knows. Alternatively, Silverfast programmers got confused in their labeling of the two profiles.

Again, does Vuescan does any of this better?

 

[Kai-san]

If you have the SE or SE plus version it will not have a full set of profiles. For that you need the AI Studio version 8.8. There should not be any difference for different scanners. I have the 8.8 version for Epson V750 and Reflecta RPS 10M, both have the Portra profiles.

 

[Jamie123]

Again, does Vuescan does any of this better?

I never quite got the hang of Vuescan but you could just try it out yoursef. You can test it for free but your scans will have a watermark across the image.

If you indeed have the Silverfast SE version and not the Ai version you could, of course, upgrade but I'm somewhat skeptical that you'll be happy with it. The Negafix profiles aren't really what one would expect, i.e. making the film look "right". It's more like they give the scan a certain look that they decided is what fits a certain film stock. Sometimes it looks good, a lot of times it takes a considerable amount of tweaking.
As far as I know the SE version of Silverfast lets you scan in Silverfast RAW format. Since NegativeLabPro now supports Silverfast RAW files I would give that plugin a try. (Again, I think you can try it out for free.)

The whole idea behind film profiles is somewhat questionable the more you think about it. As you've pointed out, there's no reason why setting the profile to VC should increase saturation. The "vivid color" should be present in the negative, not in the scanner profile. One would expect that all a profile does is compensate for the orange base of the specified film. However, as I've mentioned, the tint of the orange base can vary quite a lot depending on the age and/or storage of the film so it's not possible to have a fixed software setting that always works.

I share your aggravation about scanning and I also have a hard time getting the colors right. What's interesting, though, is that I don't really have those problems in the color darkroom. I can see quite quickly what has to be tweaked and can get the colors right more easily than when I'm working with a scan. With a darkroom print either it looks good or it doesn't. With a scan, though, often I like some things in the image and others not. Like, the midtones are good, but the highlights are off, or the greens are nice but the blues are off etc. etc. And every time I tweak what's wrong it throws off something else.

 

[Ted Striker]

You don't have to use Kodak C-41 profiles, try Fuji, Agfa, etc. And play round with the individual RGB curves in the histogram. You can make a film look anyway you want once you master those.

Jim B.

Can't you just turn profiles "off"?

 

[wpb]

Back in the film days the lab would have a “subscription” to test negatives from all the major film brands/emulations. Each strip would have 5 frames; -2 and -1, normal exposure, and +1 and +2. With these negatives the printer would make prints and then the machine would scan them and create a “slope” so it would color correct for improper exposure. Films between (say -1.5) would be adjusted according to the slope of the curve created during calibration. Without the proper slope your (even slightly) over or underexposed negatives will have color variances compared to the properly exposed negative.

Just removing the orange mask was only part of the equation, and only works if all of the frames are properly exposed.