Why is there color crossover in divided C-41?

chuck94022 · Feb 7, 2013

After mixing the last batch of working solution from my Tetenal kit late last night, and facing having to buy another one when this stuff exhausts, I started thinking (never a good thing...). I had recently started working with Diafine in B&W, and have been impressed with its simplicity and flexibility. So of course, I thought, why not a divided developer? Wouldn't such a developer work for the first developer in C-41? And, as a divided developer, since it works to completion, doesn't that solve any crossover problems?

So some quick research on the web this morning showed me that in fact, this ground has been plowed before, by Patrick Dignan and by this guy: http://nelsonfoto.com/SMF/index.php?topic=9360.0

It seems that in fact, a divided development brew has been done, and it works at room temperature. This immediately sounds awesome to me, and seems a brew I can make myself, which is triple keen, assuming I can find the ingredients.

But then I read replies to Dave Lyga's post on diluted C-41. RPC reported that his sensitometry tests showed crossover in the negatives. Greg Davis showed results with images, but for a different alternative process.

I'm immediately concerned about crossover, because in my experience, it is not simple (or perhaps even possible, practically) to correct it sufficiently to get accurate colors across the spectrum. Tweak one color, you've lost another somewhere else in the image. Frustrating. I don't want it, I'm not a Lomographer.

So, my question is this, specifically for a divided development based process: why would there be crossover, if the developer works to completion in all layers? Wouldn't all the color layers be fully developed, and thus balanced? I am certain this question reveals my full ignorance of what happens in the C-41 process, which is the purpose of the post. Would those deeply knowledgeable about what is going on chemically respond? I would be curious to try a divided development process, but not if in the end I am faced with never ending crossover battles.

Athiril · Feb 8, 2013

I would assume colour crossover from the interlayer effects not being present. Otherwise if they're present.. then it'd just be mismatched curves, not under correction of colour crossover in the masking (ie: Not the green portion having exposure from blue and/or red in it).

If you're correcting digitally, then a Xrite ColorChecker Passport will cure what ails you, but it's best of course to start off with the best source.

Add a few mg (1-1.4mg per litre - in reference to this formula - (there was a url link here which no longer exists) ) of Potassium Iodide to at least part B (though you would likely want it in both parts) of Dignan's formula, skip the Benzotriazole. HAS is important in working solution developer as a preservative, and to keep replenishment rates low. So it may not be needed, though I would recommend some in part A to prevent excess fog. There are many suppliers of CD-4 in China. Check to see what acid salt they supply though, that'll change your mix ratios.

The Sodium Sulphite present activates at least some of the CD-4 in my experience (3.5g from calculation if using the H2SO4 salt of CD-4), iirc it's

H2SO4 + Na2SO3 -> Na2SO4 + H2O + SO2
98.079g + 126.043g

Some people skip the sulfite and find they got thin to no images. Having the developer activate only in bath B is not good enough, at least a thin image must form in Bath A for this to work reasonably.

You can replace the 4.5g of Sulfite with 9.5 grams of Sodium Bicarbonate to activate all CD-4 (H2SO4 salt), and then set the desired pH with sodium metabisulphite.

Mr Bill · Feb 8, 2013

chuck94022 said:
So, my question is this, specifically for a divided development based process: why would there be crossover, if the developer works to completion in all layers? Wouldn't all the color layers be fully developed, and thus balanced?

I think the simplest response is to say that no, the developer does NOT work to completion. If you doubt this, consider the situation of people who "push process" C-41 films. They use longer development times to increase the amount of development, which proves that the spec development is NOT to completion.

The development of C-41 films is an incredible balancing act between the processing chemicals and interactions within the emulsion, optimized for the spec conditions. So it should not be surprising that any variation might produce color crosses. I'm not sure how someone would measure these sensitometrically, except in the simple case of a neutral tone scale (photograph a gray scale, or the equivalent via a sensitometer). Even if no color crosses exist here, who can say if such would exist in a non-neutral case, such as human skin tones in portraiture?

I've done pretty extensive testing on pro color neg films for studio portraiture, doing evaluations for a large chain/finishing outfit. When new pro films were being introduced, we would start with sensitometric screening and simple shooting tests. These were followed up by exposure variation tests, where all would be optically printed onto the matching professional papers; the prints are hand-balanced to match one master control print, roughly to skin highlights. Professional color correctors, using special color booths, would evaluate the results, both critically and with respect to what an average person might notice. The results are too complicated to put into a written report, other than superficially, and certainly a densitometer is not adequate for this. If future questions should arise, we'd always pull the original test prints back out; maybe the test has to be repeated because of some other changes.

Those days are long gone, and modern digital printing can probably, in principle, be adjusted to correct nearly any deficiencies.

Athiril · Feb 8, 2013

Mr Bill said:
I think the simplest response is to say that no, the developer does NOT work to completion. If you doubt this, consider the situation of people who "push process" C-41 films. They use longer development times to increase the amount of development, which proves that the spec development is NOT to completion.

The development of C-41 films is an incredible balancing act between the processing chemicals and interactions within the emulsion, optimized for the spec conditions. So it should not be surprising that any variation might produce color crosses. I'm not sure how someone would measure these sensitometrically, except in the simple case of a neutral tone scale (photograph a gray scale, or the equivalent via a sensitometer). Even if no color crosses exist here, who can say if such would exist in a non-neutral case, such as human skin tones in portraiture?

I've done pretty extensive testing on pro color neg films for studio portraiture, doing evaluations for a large chain/finishing outfit. When new pro films were being introduced, we would start with sensitometric screening and simple shooting tests. These were followed up by exposure variation tests, where all would be optically printed onto the matching professional papers; the prints are hand-balanced to match one master control print, roughly to skin highlights. Professional color correctors, using special color booths, would evaluate the results, both critically and with respect to what an average person might notice. The results are too complicated to put into a written report, other than superficially, and certainly a densitometer is not adequate for this. If future questions should arise, we'd always pull the original test prints back out; maybe the test has to be repeated because of some other changes.

Those days are long gone, and modern digital printing can probably, in principle, be adjusted to correct nearly any deficiencies.

I would imagine you would shoot a grey scale along with various colour strips, and then you could measure the grey scale and compare to colour strips and see how much 'wrong colour' is in the colour strips (ie: green and blue in a mostly pure red target).

Mr Bill · Feb 8, 2013

Athiril said:
I would imagine you would shoot a grey scale along with various colour strips, and then you could measure the grey scale and compare to colour strips and see how much 'wrong colour' is in the colour strips (ie: green and blue in a mostly pure red target).

Well, assuming that nothing else is wrong with the method, you would still only see a certain relative strength of a certain color. For example, say that you used a strongly red-colored step wedge as a target. So perhaps you might say that an 80% saturated red color does not show a color cross. But isn't it possible that a 30% saturated red MIGHT have a crossover? Or a mix of 30% red and 50% blue, etc? The interactions between film layers are too complicated, I think, to generalize from a small handful of specific color tests.

chuck94022 · Feb 8, 2013

So if I understand what you are saying, Mr. Bill, the layers are not just individually developing, but they are also interacting with each other. Is that the case? Also, I think you are implying that since the layers aren't developing to completion, with the two bath development they may also not be developing at the same rate (and worse, perhaps development slows at different rates in different layers, causing crossover in some tonal values but not others). Am I understanding this correctly?

chuck94022 · Feb 8, 2013

Athiril, the Colorchecker Passport is on my wishlist! Thanks.

Mr Bill · Feb 8, 2013

chuck94022 said:
So if I understand what you are saying, Mr. Bill, the layers are not just individually developing, but they are also interacting with each other. Is that the case?

Yes, exactly.

These effects seem to be very tricky. If you remember the introduction of Fuji Reala film, with the "fourth layer," they were able to counteract unwanted greenish effects under certain fluorescent lights. Later films seemed to handle this situation just about as well without a fourth layer. So some very tricky interimage effects are going on. These are commonly attributed to things called "development-inhibitor reactors," known as DIRs or Super DIRs. (I only know of these things from technical papers or the very occasional discussion with film engineers.)

I don't know much about the split developers so I shouldn't talk about them, but it is my opinion that modern C-41 film developing is such a finely balanced process that any tampering is more likely to unbalance things than it is to keep them the same. Just guessing, though.

Prof_Pixel · Feb 8, 2013

Mr Bill said:
it is my opinion that modern C-41 film developing is such a finely balanced process that any tampering is more likely to unbalance things than it is to keep them the same. Just guessing, though.

To modify that: the emulsion design and developer design are, together, a finely balanced system.

Photo Engineer · Feb 8, 2013

Amen Fred!

Look guys, the C41 film family consists of over 10 layers with 3 sets of 3 emulsions making for a total of 9 emulsion types in each film. They all have to reach the right point at the right time or 3' 15" at 100F. Any change to that will cause problems (within limits of about 15" for push or pull). In fact, it is so critical that it is better to over and under expose than it is to push or pull.

And, correct processing is also a diffusion controlled process involving the color correcting chemicals which must diffuse properly to give the correct colors.

By comparison, B&W is a snap. Oh, and even in B&W a divided developer must be fined tuned for time and formulation for your particular film!

Anyhow, the guy in the reference from the OP posted the most difficult pictures to judge for crossover. For example, a portrait might have had greenish highlights and beefy higher density areas. BTDT. Anyhow, Pat Dignan and I exchanged letters on that in the '90s and we even talked on the phone. By the end, I think that I had him convinced that the divided developer idea might be a bad one. IDK for sure.

So, in a word, don't do it. Do it the right way.

PE

Paul Verizzo · Feb 9, 2013

Divided Development, B&W or color

I have spent years experimenting with various B&W formulas, published and my own creations. My gosh, who wouldn't want a developer that was film, time, or otherwise so compensating for every situation? But truth be told, I NEVER got one that worked well. Either thin negatives or large grain. Many say that is because modern films aren't like those of 80 years ago. Whatever the reason, forget it.

Thats B&W! Imagine the complexity times ten or twenty for color. Patrick Dignan published a formula for divided development way back when. I've never tried it, but I've never seen anyone on this or other forums claiming "Eureka!" I do not recall one instance of his formula working.

Photo Engineer · Feb 9, 2013

Paul;

You are almost 100% correct. Where there is a slight difference in my observations and yours is that a divided developer for B&W is possible, but it must be individually tailored for each and every film due to Silver Halide type, emulsion thickness and etc. This is a painful and tedious task and therefore has never really been made popular. But yeah, you got it. And, color is worse!

PE

Paul Verizzo · Feb 9, 2013

Dear PE: Since we "Almost" agree, there is surely hope for peace in the Middle East! Glad we are on the same side ("almost") for once.

The lure of divided development is a sweet song, but as we both now know, it's just not practical. Wish it were other wise.

Photo Engineer · Feb 9, 2013

Now,we agree 100%.

Peace in the middle east is sadly not as easy to achieve.

PE

chuck94022 · Feb 10, 2013

I'm confused about your B&W comments, since Diafine seems to have a wide following and appears to work for many films. What am I missing?

Paul Verizzo · Feb 10, 2013

chuck94022 said:
I'm confused about your B&W comments, since Diafine seems to have a wide following and appears to work for many films. What am I missing?

Nothing to be confused by. They just don't work very well with modern films. Negatives tend to be thin. Although it's been a couple of decades or more since I used Diafine, I've mixed my own two baths for the same period of time. I've used other people's formulas and made up my own.

Some people have had decent results by overriding the whole intention of divided development. I.e., keep the film in the first bath longer. The pH is not neutral or low, the sodium sulfite does cause development. If you use a longer A bath, you are right back to time and temperature controls. I've acidified the A bath so no development would take place, used sodium carbonate in the B bath, and IIRC, got very thin negatives. Diafine and all others do develop in the A bath.

Also, I never found the grain to be fine, and it theoretically wouldn't be given the chemistry.

Photo Engineer · Feb 10, 2013

Chuck;

Let me give you an example. Film "A" has 300 mg of gelatin per square foot and 300 mg of Silver Halide. Film "B" has 600 mg of gelatin and 300 mg of Silver Halide in the same area. Now, film "B" will absorb 2x more solution 1 than film "A" and therefore will have that much more to carry into solution 2. You therefore get more development with film "B".

In this case, film "B" might be an old film and "A" might be a modern film, as per Paul's example above. And so, you have to either accept thinner negatives from film "A" or adjust solution 1 or 2 for these new conditions.

The fact that you do not observe it may only mean that you have not done side by side comparisons and / or, you are getting less than optimum result.

For this very reason, divided B&W developers never were really commercialized by EK, Fuji or Ilford.

BTW, the Wikipedia article is incorrect. It implies that commercial processing is done with a 2 bath developer. This is not so.

PE

nworth · Feb 10, 2013

Dignan did his work with early C-41 films. He did it correctly, using proper sensitometric and color control tools, and got a developer that worked decently and got many excellent reviews at the time. Although that developer tended to have the advantages of a black and white divided developer, it was still quite fussy about composition, times, and temperature. The reference the OP gave showed pictures made with a modified formula which really looked pretty bad, with magenta highlights (not shadows). Scanned postings are always difficult to judge, in any case. The deviations from the prescribed recipe may be part of the reason they look so bad. the formulas always were fussy. We need to note here that color films have changed a lot since Dignan did his work. The changes were calibrated for standard C-41 processing, not for Dignan's soup. The two bath solution may not work as well with these modern films as it did with the early films he used.

Dignan NCF-41 Divided Color Negative Developer
For C-41 films

Prebath
Water 800 ml
Sodium bisulfite 1 g
CD-4 11 g
Sodium sulfite (anh) 9 g
Water to make 1 l
pH at 75F = 6.5

Activator
Water 350 ml
Potassium carbonate 53 g
Potassium bromide 500 mg
Benzotriazole 2 mg
Water to make 1 l
pH at 75F = 11.8

Treat film in prebath at 75F for 3 minutes. Drain for 15 seconds. Treat in activator at 75F for 6 minutes in tank, with continuous agitation for the first 15 seconds and one inversion every 15 seconds thereafter. Treat in activator at 75F for 5-1/2 minutes in drum processor. Provide a 15 seconds drain after the activator and before the stop bath. An 87 g /l sodium bisulfite stop bath at pH 4.8 was recommended.
Ref: Dignan Photographic Newsletter, September 1979

Photo Engineer · Feb 10, 2013

And IIRC, Pat did that with just one C41 film, not several. Also, today's C41 films are very very different in composition including emulsion types and etc.

Yes, I talked to Pat about this and many other projects. At that time, he had a working color paper developer using CD4 and I told him what my dye stability tests looked like. He backed off.

Point is, unless you do a lot more tests than usual for APUGgers, then you run the risk of error. In that formula, maybe Portra 160 might work but Portra 400 would fail, or vice versa. IDK. Maybe neither would work. And Ektar would be a pain with its thin structure and 2 emulsion types.

Oh well, have fun.

PE

RPC · Feb 11, 2013

I experimented with Dignan's divided developer several years ago and got terrible results with Konica and Kodak films (Portra 160NC was the worst) while Fuji films fared much better with the best results with Pro 160S. The 160NC images were extremely thin with a very yellow mask while the Pro160S was almost acceptable, but overall contrast was a bit low and cyan contrast low, and this was quite visible in prints with some subject matter.

So, yes, different films give different results. One could probably reformulate it to work better with the different films but it is a sure bet you will always have some crossover, so applications would have to be non-critical. One advantage is cost. Whereas the second bath should be used one-shot, the first bath can be re-used multiple times.

Athiril · Feb 13, 2013

I've tested various split bathes on C-41 and ECN-2 films.

What I've found is that there -has- to be development (low contrast) in bath A, if only a little bit, otherwise, very thin negatives from very low EI's every time (and bad colour).

The carry-over amount in the emulsion is simply not enough to have developing only occur in B. Even with vastly increased developer levels in bath A.

And with this.. you can alter effects and colour by varying time in Bath A.

Some people have left out the sulphite in Dignan's formula and found they got essentially nothing on their negs. When they finally tried it with the sulphite, it was greatly improved.

Paul Verizzo · Feb 13, 2013

Athiril said:
I've tested various split bathes on C-41 and ECN-2 films.

What I've found is that there -has- to be development (low contrast) in bath A, if only a little bit, otherwise, very thin negatives from very low EI's every time (and bad colour).

The carry-over amount in the emulsion is simply not enough to have developing only occur in B. Even with vastly increased developer levels in bath A.

And with this.. you can alter effects and colour by varying time in Bath A.

Some people have left out the sulphite in Dignan's formula and found they got essentially nothing on their negs. When they finally tried it with the sulphite, it was greatly improved.

Yup, pretty much my experiences dealing with B&W divided developers. I loaded up Bath A to the maximum, I think it was hydroquinone solubility, whatever, and it wasn't enought to kick ass in Bath B.

Take that and multiply times three emulsions times all the layer interactions..............

Photo Engineer · Feb 13, 2013

Guys;

I wish you two were here. I have an awesome tutorial on this subject I am aching to give you. It is too long to write out as it would take several hours. There is so much to learn on this subject. But, you see the basics of it.

PE

chuck94022 · Feb 19, 2013

Photo Engineer said:
Oh well, have fun.

PE

in addition to all the very useful information in this thread, the above is probably the best admonition of all. I might -might- try divided c-41 at some point, just for fun. Obviously it won't be a go to solution for critical work!

Once I am back in the USA this summer, with more ready access to raw chems, I'll probably give it a go. In the mean time, I'm satisfied with spending my spare time improving my home brew water bath to deliver precision temperature control on the normal process.

Thanks for all the insight, as usual.

Photo Engineer · Feb 19, 2013

Just make sure that your database is proof against Chinese hacking!!!!!!

PE

Why is there color crossover in divided C-41?

Privacy & Transparency

Privacy & Transparency