Curve section - here you go completely wrong. Typical color neg films might have wide exposure latitude overall, but just look at the published dye curves and see just how fast that color spikes
intersect and cross-talk. You can't clean up things at that point in the curve, Photoshop or otherwise. And to get a crisp scan of what's left requires either a large sample size (i.e., large film)or a very high quality drum scan because the curve shape of the geometry changes rapidly. But what do I know? I only spent a week with a spectophotometer making a true gray card that is actually 18% gray all the way from IR to UV. Commercial gray cards aren't even 18% at the middle.
I probably have a different concept of a lightbox than most folks too.
The general problem with negative film is that if there is no colour reference in the picture (if there is no test frame with a known colour patch target) then the 1-hour printer is left in the cold about filtration, and takes all the brown stuff when the result is disappointing the client.
Negative films sadly need a colour reference. Tourists don't put colour references in their shots and they happen to be disappointed by the prints they receive from the shop. It's not the shop fault, it's the medium. Professionals working in a colour "managed" environment use colour charts as a matter of fact and might prefer negative film.
I didn't scan negatives a lot, but I did discover that's a lot of head scratching involved when filtering without a reference, at least a neutral grey.
I don't know about how exact can slide film be in their "best" part of the characteristic curve. I think that when one works in a studio under controlled light (catalogue work or things like that) it's not uncommon that film is used on its linear part of the characteristic curve, where slide film can be on par with negative film.
Negative film has better characteristic curves on the film sheet, but they do extend much more than with slide film (wider dynamic range), and all that dynamic range is not necessarily used in a controlled-light situation, in the studio.
So I am not surprised to hear that slide film, when used in its limited "best" part, is on par with negative film for exactitude of colour rendition. But negative film is chromatically "correct" on a much broader dynamic range than slides which makes it an easier choice.
I have always preferred "neutral" films like Astia and I do think that their appeal rests on their fidelity, not on the surprise-surprise effect (which can be said for "saturated" film like Velvia).
Color neg film has more range - way more than one actually needs in a critical studio environment.
This stuff is engineered for pleasing skintones first, everything else second. If you want accurate
colors in a high contrast environment there's only one way to get it - a tricolor camera using straight
line black and white film! But if you want to get into copying museum painting nowdays go get a
Beterlight scanning back before they themselves go extinct and put a good Apo El Nikkor on it for
only another three grand. Or buy some of the real-deal chrome 8X10 dupe film from my freezer, though I suspect some it has already gone bad, at least for critical work. End of an era. So time to
apply a crowbar to Portra 160 and leverage some new tricks (I've already started).
I think that HRST has it right and Drew has it wrong. I have run literally thousands of comparisons between E6 type films and C41 type films when designing color negative film and being responsible for color reproduction. The masks and DIR couplers give C41 films a color fidelity that cannot even be approached by any reversal film.
Yes, slide films look beautiful, but they are not accurate which is the essence of the question in the OP. I vote for the Portra family. And, you get 2 shots at correction, the original auto corrects and you can make corrections in the print.
Before you ask, I'll answer. The comparisons were all Kodak C41 films ve Fuji, Konica and Agfa color C41 films, and all Kodak E6 films, Kodachrome, and again Fuji, Konica and Agfa E6 films. They were printed and viewed. Prints were made on type R, C and Cibachrome print materials, and also using Ektaflex R and C materials. As I said, it was a huge comparison.
I might add that part of this test also compared two sensitizing dyes and two magenta couplers in the paper and C41 film, and two cyan couplers in the paper. It also spanned 2 color developers for the paper. Did I mention that this was a huge test? Drew, you have not done one tenth of one percent of the tests I have. And, you have not had to have your results passed by a panel of observers.
Is that enough for you guys? Geez, no wonder Kodak film is so expensive. We did (and do) it right!
This. Unmasked films cannot match masked films. The dyes are not perfect. Therefore unmasked films cannot produce accurate results. You don't need any tests to tell you this, other than very simple logic and a little bit of math.
Originally Posted by hrst
"The multiple generations of film involved in the color neg-
pos process (as many as four in the motion picture chain)
served, however, to emphasize the colorimetric imperfec-
tions of the subtractive primary dyes. Ideally, a cyan dye, for
example, should control only red light by absorbing between
600 and 700 nm. But virtually all cyan dyes also had significant
unwanted absorptions in the blue-green. This led to desatu-
rated or “muddy” colors, especially in successive generations of
film. The ingenious solution to this problem was a technology
called “integral color masking,” invented by W. T. “Bunny”
Hanson of Kodak and introduced in Ektacolor films in 1949.
The technique added “colored couplers,” which bore an
attached pre-formed azo dye, to the normal colorless coupler
in the layer. For example, the added cyan dye-forming colored
coupler carried a blue-green dye that would be released and
washed out of the film to the extent that cyan dye with its
unwanted blue-green absorption was formed. The result
was equivalent to a “perfect” cyan image dye overlaid with a
uniform density to blue green. While this gave the negative an
orange cast, it required only a longer cyan exposure in making
the positive print. So revolutionary was this improvement that
virtually all negative films would adopt this technology once it
was free of patent restrictions."
Last edited by Athiril; 06-07-2012 at 09:07 PM. Click to view previous post history.
Sponsored Ad. (Subscribers to APUG have the option to remove this ad.)
Exactly this. Astia looks nice and is very accurate for a chrome, but it is not nearly as accurate as Portra.
Originally Posted by Athiril
Next thing to keep in mind is that films have only three (sometimes four) output channels (dyes) and reproduce colour "accurately" only to human eyes. A "perfectly accurate" tri-colour film would have three photosensitive dyes, each with the same spectral sensitivity as one of the cone-types in a human eye, and with an absorption spectrum to match and therefore expose the paper appropriately. So colour accuracy is heavily dependent on the papers used too, as well as how the film exposes the paper (correspondence between absorption spectra in the film and sensitivity spectra in the paper). Masking is necessary to deal with differences in both the sensitivity and absorption spectra of the film with respect to human vision.
Obviously the RA4 cannot be masked for the same reason a chrome cannot be masked, however my understanding is that the masking in the film is designed to make up for dye deficiencies in both the film and the paper. If you have no mask at all (i.e. a chrome), a whole bunch of colour errors will creep in.
We differentiate slightly different hues (wavelengths) by comparison of intensity of response from different cone types, which means we cannot differentiate between mixtures of light (a little blue plus a little green) and a monochromatic source somewhere between the two. Films have the same issues. When you look at a print or chrome, it's showing a few narrow spikes of fairly monochromatic light, with magnitudes chosen so as to excite cones in the eyes at the right ratios to believe that there is a full spectrum present when there is not. If you had non-human eyes, a colour print designed for humans would look totally wrong.
Edit: I see Athiril has posted in a section about the reason for the mask being mostly-orange. The primary correction is to the cyan dye but there are other, more-subtle spectral corrections used in modern films.
Last edited by polyglot; 06-07-2012 at 09:36 PM. Click to view previous post history.
I never thought my seemingly simple question would spark such an interesting discussion
Nerds. No one expects nerds on the internet
Drew, it seems to me that while you have real hands-on experience on the topic, it also seems you have some problems in the very basics of color theory. Or then, I might be misreading you.
And they NEED to intersect and "cross-talk". It is the only way to give good color accuracy in a tricolor RGB system! It simulates the eye. We come to the very basics of the color theory here; the optimum RGB system would use wide band "input" filters when capturing the images, to simulate how the eye works, and when presenting the image, narrow band "output" filters to eliminate any crosstalk happening in the eye at that stage.
Originally Posted by DREW WILEY
It is actually a common misconception to think that the input stage would use narrow filters too.
If the coupler dye absorption spectra would not intersect and "cross-talk" at all, some yellow objects would be rendered black and some yellow! We cannot tell the difference of these objects by eye when looking at them, but a narrow-band RGB imaging system would show it. This system would not be called "accurate".
Wavelengths between the primaries have to expose both of the records in order to be reproduced in the RGB system.
Not a single word here makes any sense to me, sorry!
You can't clean up things at that point in the curve, Photoshop or otherwise. And to get a crisp scan of what's left requires either a large sample size (i.e., large film)or a very high quality drum scan because the curve shape of the geometry changes rapidly.
It is a bit arrogant to assume that others here have no hands-on experience on these topics. For example, I have built a film scanner from scratch (for my motion picture laboratory) and am currently building version 2, as well as motion picture contact printer etc, not to mention the lab. I had no problems getting good color fidelity and accuracy out of 16mm motion picture color neg from the first try. It was as simple as adjusting the RGB levels of the light to get a neutral near-white gray from Dmin. In this sense, color negative has as much "reference" as color slide does.
The reference can be a neutral print with neutral filtration, or a positive copy on a print film if it is to be viewed on a light table. It only necessitates SOME bookkeeping and laboratory discipline skills, and some calibrated equipment to do this well.
The difference is just that color negative film ALLOWS corrections and some people get confused with this possibility!
In a real-world photography anyway, corrections are almost ALWAYS needed. The only real reference was the scene when it was photographed. The slide may look completely different, so it makes no sense to call it reference. Actually, shooting in studio is the only exception and even that requires a lot of skill to make the slide have the desired color and contrast. Maybe some people, again, are lost when they have learned that skill with a lot of hard work and forget it is only for one specific corner case and not very descriptive on the theory of color imaging.
If the aim is to have something to look at light table, then you have to compare slide film with a print of a color negative made on print film (in the correct meaning of the words) stock.
Do you know how motion picture films are/were produced? Do you know that everything is shot on negative, and that they can copy original negative to a negative-working interpositive, that being copied on an internegative, copied on a release print film? There can be four color "negative" film copies in a row, and the end result can still have astonishing life-like color accuracy. Try copying a slide three times.
This all comes back to linear curves and color masking. Please take a look at the curves of any today's color negative film. You will see they are unbelievably linear. More linear than most B&W stock available. Also, the spectral sensitivity set is so good that you wouldn't gain that much benefit from using B&W separations. After all, you would probably end up choosing too narrow bands and get an interesting effect of increased saturation and false color, like slide film. This is, however, not called "accuracy".
I agree that for a studio lighting conditions, a certain slide film designed for just this special purpose can do remarkably well. And it probably does better than those color negatives NOT designed for this purpose, especially if you shoot those without 80A or 80B filter. So we have a general product having 10x to 100x more accurate colors, and then we have a very special corner case with a specifically engineered product just for this purpose, and it can do better in this case than the general product. And there's nothing wrong using it, quite the opposite.
Also, we all are leaving out the fact that if we need a specific output format, the game changes. If we NEED a projectable slide, then slide film is a very natural choice. If we NEED a print, then color negative film is a very natural choice. If we need a digital file and if we have a usable scanner with a usable software, color negative film gives better color accuracy for most purposes, but slide film might still look better if that look is desired.
We are lucky to still have so many choices left.
One more question to digest; have you looked at the same slide
(1) on a small light table
(2) on a large light table
(3) on a light table, with everything except the slide masked black
(4) on a light table with 2800K tubes
(5) on a light table with 4000K tubes
(6) on a light table with 5600K tubes
(7) on a light table with 9000K tubes
(8) on a light table with poor color rendition index
(9) on a light table with good color rendition index
(10) with a projector equipped with a 110V or 220V mains halogen bulb
(11) with a projector equipped with a 24V halogen bulb
(12) with a projector equipped with a xenon arc bulb
I have a slide which has absolutely stunning delicate light-blue hues. It looks very good scanned or on a light table with 5600K tubes, but horrible with a halogen projector. Completely different image. It was supposed to be a REFERENCE as it is a SLIDE! Go figure. I ended up cutting a piece of light-blue wratten filter with the slide for projection.
Last edited by hrst; 06-08-2012 at 02:06 AM. Click to view previous post history.
Originally Posted by Athiril
Here is how I understand the masking of unwanted absorptions in a color negative. Someone correct me if I am wrong:
The cyan dye absorbs some green and blue light which it shouldn't, making it have a red component. The magenta dye absorbs some blue, which it shouldn't, making it have a yellow component. These red and yellow components combine to give the dyes an overall unwanted orange component. The yellow dye does not have any significant absorption problems.
The cyan dye coupler is colored red in manufacture, and the magenta dye coupler is colored yellow in manufacture. If a colored dye coupler does not convert to dye during processing, it retains its original color. If it converts to dye, it has its new dye color, plus the color component of the dye's unwanted absorption, which is the same as its original color. Therefore, after processing, the orange color is uniform all over the film. It consists of two things: the residual dye coupler, forming the "mask", which is orange, and, what it is masking: the unwanted absorptions, also orange. The mask forms a positive orange image, and the unwanted absorptions form a negative orange image, which cancel each other. This uniform orange color is simply filtered out during printing, and along with it the effects of the mask and the unwanted absorptions.