E-6 to C-41: A Transition

[Ben Syverson]

Now that I've handed you the tools to correct for contrast disparities, let's consider spectral response (color).


First, here is the spectral sensitivity for E100G overlaid on top of Portra 400. I adjusted the curves for E100G vertically, so the Y axis becomes a relative rather than objective measure. But even so, we can see that the curves align quite well. All you need to do is check out the Spectral Response of digital sensors to see how much these curves can vary yet still produce a similar image. The fact that Kodak's E6 and C41 materials have such a similar spectral response is an absolute feat of engineering. They are both incredibly, incredibly neutral.

The E6 film's very slightly "sharper" individual spectral curves mean it has slightly more natural color separation (saturation), but this is obviously easily replicated. For absolute precision, you could subtract the E100G response from Portra's, and apply those curves to the ab channels of a Lab image, but honestly, it will be pretty indistinguishable from a few targeted HSV adjustments.

Next, consider the spectral dye density curve for Portra 400:

I've taken the liberty of subtracting Dmin from the Neutral curve, to create a combined relative density curve of a neutral subject (the red line). The big dip in the orange part of the spectrum is actually intentional (that's the orange mask). Normalizing the curves the way I instructed in my previous post would bring that section of the spectrum back up

Finally, let's overlay the combined E100G curve with Portra 400's:

Hey, what do you know. They are extraordinarily similar. The biggest difference is that dip in orange response, which is again because of the orange mask, and is corrected out either in scanning or printing.

So what have we learned? Kodak's C41 emulsions should be trivially easy to match to their E6. The most important adjustment will be adding a nice S-curve to the extremely linear response of the negative film. In terms of color reproduction, they should already be quite close, but targeted HSV adjustments should help individual areas of the spectrum match closely.

[Drew Wiley]

Well, I like what you're illustating, Ben. But what you regard as very minor differences in the spectral sensitivity curves I perceive as pretty significant. One has
to get used to interpreting these things and what they mean in practical terms.
Basically, once a neutral is in effect, you can't remove it. Makes no difference whether we're talking film or industrial colorants - same principles. Portrait-style color neg films are particularly adept at creating neturals which mimic skin tones, but at the same time will "dump" analogous hues in nature the same direction. The question is, what other colors can you salvage-engineer in terms of priority? Significant progress has been made, first with Portra VC, and possibly now in Ektar, to invade territory formerly held only by chromes. The curves are getting more
similar. And making a fuss about it has its benefits, because this problem isn't
going to go away. Maybe a good percent of the time, things will be easy to correct;
but if LF chrome film is outright unobtainable, then that other 10% of applications
becomes very important to de-bug too.

[Drew Wiley]

Your comparison graphs themselves, Ben ... It's logarithmic, so the difference between the E100G vs Portra 400 show a curve overlap fully two stop of density further into each spike - that's a gigantic difference in terms of how much the inherent neturals or cross-talk between the respective dye layers begin. (With Ektar it's obviously less). To an important degree, your own graph overlays disprove the point you are trying to emphasize. Those neutrals are there. You can change your color balance, contrast, and overall curve, but there is no way to remove those neutrals or any related hue contamination from an adjacent dye layer (that is, unless you want to "paint" or dither or whatever, applying information which never existed in the film itself). But in terms of making these particular neg films look "more" like E100G, what you have done is certainly a useful illustration, because many indivduals will want to stick to basic PS controls. Yet what I'm implying doesn't
outright negate what you are attempting to do, because we are using different tools
for an analogous problem. I like putting a bit of fire under this debate, because folks
like you do come up with very useful tidbits of image mgt.

[Brian Ellis]

Guys, having tried a few times to shine some light into the cracks in Drew's granite like belief systems...............

[Roger Cole]

Hence the "overkill" of the subject, Bob. I'm not doing this just to be academically correct. I want options if tranny film &/or the relevant printing materials are lost;

Print materials for them are pretty much lost already with type R paper gone and Ilfochrome now an exorbitantly expensive special order item.

Back to the regularly scheduled fracas.

[Ben Syverson]

Well, I like what you're illustating, Ben. But what you regard as very minor differences in the spectral sensitivity curves I perceive as pretty significant. One has
to get used to interpreting these things and what they mean in practical terms.
Basically, once a neutral is in effect, you can't remove it. Makes no difference whether we're talking film or industrial colorants - same principles. Portrait-style color neg films are particularly adept at creating neturals which mimic skin tones, but at the same time will "dump" analogous hues in nature the same direction.

This, I think, is not supported by the data. Saying that Portra "dumps" colors to neutral is nonsense, when you can see for yourself that the discrete spectral curves for Portra 400 are well separated. E100G's color response is marginally steeper, which means marginally less cross-color contamination, but the differences we're talking about would be measured in fractions of a camera stop.

What it means in English is that a particular green on Portra may have a quarter stop too much blue when compared to E100G. Talk about splitting hairs. If you can even see it, it would be eminently correctible with one targeted HSV adjustment.

I'm honestly starting to get insulted on behalf of the Kodak engineers who slaved to make these films as phenomenally accurate as they are.

[Drew Wiley]

Ben - the difference in green is very frustrating. I spent about two years trying to
tame it and decided either it's a no-go and back to chromes, or I'm going to have to
try some advanced masking controls or even printing from sep negs. The problem
isn't Crystal Archive paper - which is the same thing used in many digital printers.
It's pretty amenable. And getting nice bright green-greens is pretty easy with Portra.
It's when you get neutral-inflected greens; they get rather cyan contaminated, just
like in the old Vericolor, just not quite as bad. As Roger noted, Ciba is getting dreadfully expensive, and it has serious gamut problems of its own, while Type C
color paper is quite affordable, especially if direct optically printed. Supergloss is
a lot more money, but still way below Ciba. Chromes are still far superior in gamut
(that is, within their intended scale - don't tell me about blues in the blacks which
are clear off the deep end). And don't tell me you can take the contamination out
with a scan, not unless you can overturn certain laws of color which are behing
everything out there, including your scanner and the way your printer inks have
been programmed and balanced. If you want to arugue fine, but you're going to
have to start your fight even before VanGogh arrived. Cross-talk across the color
wheel results in neturals (brown, beiges, greiges, etc - if that is what a film is
engineered to do, you can't stop it!)

[Drew Wiley]

Another minor note - don't tell my problem is my darkroom. I don't use white light,
and my system is very analagous to the narrow-band lasers used in high-end digital
printers. The "mud" in a neg image comes mainly from the film itself. Scanning chromes and outputting onto Crystal Archive is one method of working cleaner; inkjet from chrome scans is another, but with a lot more potential for metamerism.
Ciba is downright goofy, and requires massive mask bludgeoning. The silver bullet ... there ain't none, but dye transfer is about as close as it gets.

[Drew Wiley]

Ad infinitum ... but I could plop down a Fuji optical print made directly from a standardized Portra neg of the Macbeath chart, and you would say the color reproduction is damn near perfect ... well, at least until I set down a significantly
masked Ciba from a chrome of the same chart... the Fuji print would win the skintone and netural patches, the Ciba everything else.... But then if I set down a
dye transfer version made from color separations from the same chrome, well then,
at that point the color neg Crystal Archive version starts looking pretty muddy. If it
had been made with a conventional additive enlarger, it would be even dirtier, but
nowhere near as bad as an old Vericolor S print on Ektacolor paper.

[Ben Syverson]

Chromes are still far superior in gamut
(that is, within their intended scale - don't tell me about blues in the blacks which
are clear off the deep end). And don't tell me you can take the contamination out
with a scan, not unless you can overturn certain laws of color which are behing
everything out there, including your scanner and the way your printer inks have
been programmed and balanced.

I don't know why you insist that C41 films are designed to have cross-color contamination. They're not. And it's really not an issue. Believe it or not, Kodak film is actually designed to reproduce colors accurately. The reason why Portra films are labeled "portrait" films is precisely because they have such long, linear response curves, NOT because of some funky desaturated color response.

And to claim that Chromes have better gamut "within their intended scale" is sort of like saying cellphone cameras have better resolution than DSLRs "within their intended print size." It is literally just complete nonsense.

You can think of Gamut as a 3D map of what colors can be represented by a format. C41 film has much greater latitude than E6 film, yet has similar spectral sensitivity, so it's objectively true that C41 has a larger gamut. If you mapped Portra 400's gamut in 3D, E100G's would fit neatly within it.

Honestly, you can't just write things and will them into truth. If you want hard data, I've given you the hard data. If you want real-world empirical evidence, several of us have chimed in with our personal experiences. I'm sure there is no hope of convincing you of anything, but please stop posting disinformation to this forum, because that type of folksy anti-wisdom spreads far too easily. We've all seen it happen.