Diluted C41 Tested

[koraks]

Peanuts will be provided to all pigeons involved.

 

[koraks]

To follow up on this; as mentioned, David has sent me his prints and negatives for me to mess around with - or, rather, to have a stab at digitizing them as well as I can. This is always a challenge, so the results shown here are not 100% accurate with the physical prints, but they're fairly close.

These are the scanned prints I received:

If you're wondering: yes, they are quite cyan/blue (except the 3rd). The actual prints are rather cool in tone, fairly low in overall density and contrast.

To get a feel for these negatives, I tried printing them as well:

As you can tell, I made some different decisions in filtration than David did. I think we optimized for different parts of the prints, and perhaps evaluated strips under different kinds of lighting. And then there's personal preference to account for, and I know I tend to print a bit warm usually.

Actually printing these also gave me some insight into the characteristics of the negatives:
* Contrast seemed fairly normal compared to regular C41.
* Grain seems a bit pronounced, but it may be the film itself (It's a Fuji 100 stock, not sure which, how old, etc.)
* Overall, the negatives have a significant yellow cast, requiring much lower than usual yellow filtration. Most of this filters out OK in the end.
* There does seem to be a quite significant crossover problem, affecting hue purity and color balance. Basically, I didn't manage to get an entirely neutral looking print from these negatives. I had to compromise for certain parts and allow other parts of the same image to shift to something I didn't intend. In particular, there seems to be going on something on the green/magenta axis that is not really linear.
The above matches quite well in all respects with the results I got from my own experiment with diluted C41 developer. I'd say that David's negatives were a little less shifted on the yellow axis and contrast was a little higher, suggesting that the Kodak developer does marginally better than the Fuji at this dilution. But both seem to run into crossover issues when used this way.

Finally, I did a very quick & dirty scan of the negatives themselves, but I didn't spend much time on color correcting them. They were smacked on the scanner glass just like that, so there are some sprocket hole reflections in the images that aren't really there in the negatives; ignore them. I did a quick attempt at some color balancing, but decided quite soon that I wouldn't be able to figure it out quite to my liking; normal C41 film tends to scan rather straightforwardly, but these negatives require more advanced filtering to correct the issues.
Edit: forgot to include the film strip; here it is:

 

[David Lyga]

And, I have this to say: If my method is deemed to be in error, I will accept that. To me, to MY eyes, under tungsten lighting, the prints I had sent seemed fine. But my eyes, my determinations, are not all that count in this world. I fully accept koraks' output even though I think that my physical prints were better. Sometimes, there is a difference between the segue from analog to digital but, here, I am not making excuses. Indeed, if you look at the prints I had 'scanned' with my cheap digital camera, maybe some will see that the colors are better.

Again, I fully accept this output by koraks and thank him profusely for his efforts. Maybe we have all learned something from this exercise. - David Lyga

 

[koraks]

David, I haven't examined your prints under tungsten light (in fact, we don't have any anymore around the house...), but I do think that they will likely look much better under that light.
It was quite valuable to actually have a go printing these negatives myself. It says so much more than going by what people post on Flickr etc.! I really think people should do this more, exchanging prints and negatives for real life comparisons.

 

[pentaxuser]

Thanks koraks for taking the time and trouble to do this. The prints look a lot less blue that the scans but there does seem to be a kind of a de-saturated look to them. It all depends on how closely you examine them. Looking at the prints by themselves a lot of people would browse through and not say that there was anything wrong with them. They do the job as prints. Against the prints from the same negatives but developed "normally" then I think most people would feel that there was something lacking in these prints but it would be along the lines of saying that prints from David's negs were OK but just not as good as prints from normally developed negs

pentaxuser

 

[David Lyga]

OK, whatever is decided upon I will abide by, without rancor. But, do take a look at what I had posted with my cheap digital camera before they were sent to koraks. I am not trying to defend myself, only to ask you to see if you see any difference. - David Lyga

 

[koraks]

Oh, there's a big difference between your digital photo and my scans of the same prints, for sure! I think the lighting conditions for the digital photo explain that difference. It's important to note that I corrected the scans against the real prints as viewed under overcast daylight conditions, so quite cool color temperature. My prints were also made for viewing under daylight/open shade.

 

[koraks]

The prints look a lot less blue that the scans but there does seem to be a kind of a de-saturated look to them.

Yes, muted colors. The hues aren't quite as pure with diluted developer compared to stock c41 it seems.

 

[pentaxuser]

OK, whatever is decided upon I will abide by, without rancor. But, do take a look at what I had posted with my cheap digital camera before they were sent to koraks. I am not trying to defend myself, only to ask you to see if you see any difference. - David Lyga

I had a look at your pics of your prints, David. I can't see any crossover in either set of prints and the blue cast in the scans is just that, a blue cast in the scans. Not some cast that is there in the negs. I prefer some of your prints to koraks in terms of overall look except that his prints seem to have more depth in the colours. In the case of the flower tub perhaps too much depth. I think I prefer your print to his in that one. It may be that there is yet another exposure and filtration setting possible for the prints that would be better for my taste and that is what it may be i.e.taste as opposed to any process issue that is part of your process. I do not see a fundamental flaw being present here. What I have noticed when I was doing RA4 enlarger exposed prints is that a slight change of exposure and filtration can make a large difference to the look. As an example I made a print of the centre of Malahide, a town in the Republic of Ireland and it looked fine to me. Then out of curiosity I changed the filtration setting by as little as a couple of units on the dichroic head. Then I made 3 more prints of the same neg in the form of a ring around and with all four front of me I could see that a change of 2 units on a head that runs from 0-130 so not very much change created by 2 units, took out a very, very slight magenta cast that was visible because the central square in Malahide was grey block paving. It just not visible without the comparison present

So I am less sure now of what I said in my earlier post about whether the normal C41 process would have made any difference. It may well be down to exposure and filtration under the enlarger.

pentaxuser

 

[MattKing]

Thanks for doing all this koraks - and David too!
It is the crossover observations that strike to my heart. And non-linear ones at that!
I have one question - are there any negatives in that batch that include pictures of people?
I ask because in my experience, photos of people are the ones where crossover can be the most disturbing.
And by the way, I would never describe what David does as inherently wrong.
I would describe it as imposing compromises that may be acceptable to some, and unacceptable to others.

 

[MattKing]

I played a bit with the picture of the flowering plants. Giving it a bit of an adjustment in the colours and gamma improves the foilage, but replaces the blue cast with a green/cyan one.

 

[pentaxuser]

Good point about a person/people shot, Matt. I am reminded now of an amusing story that the late Roger Hicks told about the problem of crossover and casts He had taken a picture of a cat called Yeti in I think his parents-in-law's house in the U.S. The cat was pure white or so he and his wife remembered it and yet in the print it had a very, very subtle pink look to it. If he altered the filtration enough he got rid of the pink look in the cat but the rest of the colours in the picture were clearly wrong. Had he got the dreaded colour crossover due to a mistake in the processing? He was certain he had not and yet there was no way to eliminate the cast without making the rest of the print worse. He gave up in the end, rationalising that it was one of those things that you just have to put down to experience. Inexplicable things do happen etc

The next time in the U.S. the cat was still very much alive and well and guess what, it had this very subtle pink look to its fur

pentaxuser

 

[David Lyga]

Well, work was done here and we all had our say. Koraks worked hard at this and this is the answer.

Yes, with color, even a tiny change can be momentous. I am thankful to all for the output and comments. No, I had not taken pictures of people, but I always had thought that gray areas, like sidewalks, were the most revealing as to color crossover. We have all learned something today. - David Lyga

Where I had posted my prints using my cheap digital camera was in post # 73.

 

[MattKing]

I always had thought that gray areas, like sidewalks, were the most revealing as to color crossover. We have all learned something today

It is very helpful to include grey in a photo that you are using to analyze colour with.
But it is even more useful to include grey with a bunch of other references.
This one gives me a lot of material to reference - things like flesh tones that move from highlight to shadow. Let's see if the uploader will mangle it:

Not too bad as a usable example.

 

[Truzi]

This is great. After reading all the back-and-forth over the years we finally have a replication! Perhaps we need more replication (with the same set of negatives, whatever they are decided on).

Anyway, both David's and Koraks' look better than my last few experiences with a minilab, which is why I started doing my own c-41. Not that I'm as good as a proper minilab (let alone a pro setup), but why spend money having someone else do the same as I do? LOL.

To me, the fact that Koraks even attempted this on his own (let alone with David's negatives) is what I imagine a collaborative site to be. We need more things like this, regardless of final conclusions.

 

[Rafael Saffirio]

Finally, I did a very quick & dirty scan of the negatives themselves, but I didn't spend much time on color correcting them

Hi Koraks.
If is not too much trouble, could you upload scans of the negatives in Tiff format without converting them to positive?

 

[koraks]

Sure, let me see if I can help. The other scans are all made on an Epson 4990 flatbed, but for your request I resuscitated my old Minolta Scan Dual IV and did some scans. Here are three sets:
(1) a set of uncorrected 8-bit jpegs scanned as 'color negative', so a positive image with no further corrections besides what magic the Minolta software does on them (and that turns out to be quite a bit of magic in my experience).
(2) a set of resized 8-bit jpegs derived from the 1st set with some additional color corrections done manually by me to get something as close to a decent result as I personally see fit - so kind of subjective. Particularly pics #1 and #3 required separate Y/M adjustments for shadows, midtones and highlights and even then, it's still not exactly as it should be.
(3) the set of 16-bit 'scanned as positive' TIFFs. I downsized these to keep the files relatively small (~20Mb each). If you require full res scans, let me know, I can share them via Wetransfer.

Set 1 (automatically adjusted full-size Minolta scans, 8b jpeg):
Image #1
Image #2
Image #3
Image #4

Set 2 (resized and with additional corrections based on [1], 8b jpeg):

Set 3 (scanned as positive to 16-bit tiff, resampled to 50% linear):
Image #1
Image #2
Image #3
Image #4

 

[pentaxuser]

koraks, these are very good. I know nothing of scanning nor do I intend to use that route but it raises two questions which are relevant to non scanning i.e. traditional enlarger with RA4 processing.
1.Did you have to do anything with the scans to correct crossover in the negatives

2. Could this look be achieved with traditional darkroom methods?

Thanks

pentaxuser

 

[koraks]

koraks, these are very good. I know nothing of scanning nor do I intend to use that route but it raises two questions which are relevant to non scanning i.e. traditional enlarger with RA4 processing.
1.Did you have to do anything with the scans to correct crossover in the negatives

Partly you're looking at the 'under-the-bonnet' digital processing of the Minolta scanning software, partly I made some manual adjustments to the images in set #2. Particularly in images #1 and #3 I did try to fix crossover issues, which were quite severe in both images, and much less so in the other two, where the issues were solved by the Minolta algorithms.

2. Could this look be achieved with traditional darkroom methods?

Not easily. That's what I tried in first instance, as shown in the initial post with the scans of the prints. Those are about as good as I could manage. However:
Image 1 looks better to me in my print than it does on screen in the adjusted film-scanned version. The digital version could easily be modified further.
Image 2 just works better in the film scan. There's no way I see (without color masks) to make the red come out like this in a real print without the rest of the scene turning into pink or orange.
Image 3 is like #1; I find it looks better in the actual print than it does in the film-scanned version. Just like #1, the digital file could fairly easily be further optimized.
Image 4 is very saturated in the film scan, and much more muted in the actual print. I prefer the latter version.

The main difference between the film scans and the prints are:
1: Saturation is lower in the actual prints. Printing them onto Kodak Endura paper would make a difference (both David and myself used Fuji paper, which is less saturated), but still not give as much flexibility as digital.
2: The extant crossover can to an extent be fixed or at least reduced in digital post processing. This is virtually not feasible in RA4 prints even with advanced techniques.

So short answer: (1) yes, and (2), no.

 

[pentaxuser]

Thanks koraks. For my education in terms of colour crossover what is it that I need to look for in the RA4 prints compared to your latest scans that indicates colour crossover in the negatives?

pentaxuser

 

[koraks]

Look at e.g. my print of the yellow car. The tarmac is green, while the yellow of the car bonnet leans towards magenta.
In my print from the Porsche, the right side of the tarmac (behind the car) is fairly neutral, on the left side it's also magenta/yellow.
Other than that, it's mostly the hues themselves that aren't quite right. The church building is impossible to filter for a neutral sandstone in the sunlit area while keeping the shadows cool without them tilting towards green.

It's often difficult to accurately put into words, but you really notice it when printing such negatives. You make a strip, decide that some part is a bit too magenta or green (e.g. the skies), so you adjust it and then a new problem pops up in a different area. This is also why some crossovers are in practice a little less obnoxious than others. If you have for example a blue/yellow crossover with dark areas going blue and highlights crossing over to yellow, it's sometimes not very noticeable if you print a sunlit scene, where the shadows will always be cooler than the sunlit areas. The fact that there's crossover in such a case is masked by our perception of the scene. However, if you then try to print e.g. a portrait from the same roll, problems do occur...

 

[MattKing]

+1 to what koraks has posted.
One thing to avoid when you are trying to analyze things like this - at least initially - is subjects like brightly coloured cars. They are nice to look at, but for almost everyone, as a colour reference they are misleading at best.

 

[pentaxuser]

Thanks koraks. My difficulty in in trying to scroll back and forth between two pages and David's digital photos of his prints and you prints and corrected scans but in David's picture of the yellow and red cars there seems to be a much smaller change if any between the front and back part of the road in the red car scene which looks more grey and the road in the yellow car scene seems fairly neutral. In both car scenes the fault appears to be over exposure rather than anything else photos,

I note that in the far pavement side of the road in the red car scene the colour doesn't change in any of the prints or scans but David's prints look overexposed
in much the same way as in a scene of a car submitted by Bormental. In his case crossover was suggested as well due to what is alleged to be overexposure I say alleged as Bormental is yet to submit his exposures. As far as I am aware Bormental does not use David's processing method and while his picture does look overexposed it doesn't look so overexposed to create crossover. Unless Ektar 100 ( the film in question in Bormental's case) is different from most C41 film then I had thought that its latitude was +4 stops in the overexposure direction? The water, certainly in the case of Bormental's alleged overexposure, is muddied by, in his case, deliberate scanning manipulation to make the scan look like he wants it to rather than a reflection of the actual negative.

I remain confused over the whole business of the causes of crossover. Is there any way of detecting crossover from the negatives alone?

Thanks

pentaxuser

 

[koraks]

Is there any way of detecting crossover from the negatives alone?

Yeah, color densitometry and grey patches.

I like to keep things simple. If it prints well easily, it's probably good. If printing gives me a hard time to get something realistic, I probably done goofed somewhere.

 

[MattKing]

Is there any way of detecting crossover from the negatives alone?

If the mask is a weird colour, and the negatives look unreasonably dense, the likelihood is much greater!
When one actually works with the scans, it reveals itself quickly whenever you have a subject that is both in the light and in shadow.