Push/Pull processing if Scanning Negative

[SergeyT]

... I know photographers who develop their black-and-white film for scanning to a much lower contrast gradient than would work well for traditional darkroom printing, claiming that the lower-contrast negatives scan better. That's fine unless you ever want to make darkroom prints...

All known to me film scanners feature the hardware (optical system and ADCs) capable of scanning color positive films. With the base density of negative and positive film being relatively equal, the DMax of modern color positive films well exceeds the DMax of any negative film (color or B&W) , hence the scanners hardware (ADC) is always underutilized when scanning negatives. Developing "their black-and-white film for scanning to a much lower contrast gradient than would work well for traditional darkroom printing" would make matters even worse as all the contrast expansion required for "digital" visualization (screen and prints) will be done past the hardware chain, by "stretching" scanned values mathematically in scanners software and/or editing programs like PS.

[SergeyT]

Part of a problem could be in poorly designed scanners software (or lack of understanding of its capabilities ) that outputs the data from scanned negatives in such a way that it creates the desire to develop film intended for scanning to a much lower contrast gradient

[koraks]

SergeyT, you might be right, assuming that the amplifiers and/or ADC's behind the CCD are fixed gain in scanners. I don't know if that assumption holds true. Do you?
S/N ratios might indeed be far worse when working with high densities due to rather obvious reasons. My personal experience with a few scanners suggests there's in any case certainly no quality gain when trying to scan through very dense areas as opposed to lighter ones. I'm quite sure I'm not alone in this observation.

[Bruce Watson]

All known to me film scanners feature the hardware (optical system and ADCs) capable of scanning color positive films.

Most consumer and prosumer flatbed scanners can't read the extreme density end of color positive films. It's difficult to do with flatbeds in general because of light scattering from adjacent pixels which you can't get away from in scan-a-line-at-a-time scanners. Most drum scanners, OTOH, can come sufficiently close, in large part because they are lighting just the one pixel at a time and thus controlling the lighting considerably better than a flatbed can.

With the basic density of negative and positive film being relatively equal, the DMax of modern color positive films well exceed the DMax of any negative film (color or B&W) , hence the scanners hardware (ADC) is always underutilized when scanning negatives.

While the Dmax of color positive films does normally (that is, almost always) exceed the Dmax of color negative and B&W negatives, it does not follow that the scanner hardware is under utilized when scanning negatives. This might be true for consumer and prosumer flatbeds. But it's certainly not true for most drum scanners, and I think for most professional flatbeds. Drum scanners in particular let the operator set the log amps right were the operator wants them, so the black and white points are exactly what the operator wants -- the scanner is neither underutilized nor over utilized. What that means is, you can make an exact match between the density range of the film and the digital range of the scanner. So if your scanner is good for, say, 14 bits of range, and your film density runs from 0.3 to 1.2, you can apply the 14 digital bits of range to cover exactly the 0.9 density range of the film. Hard to pull more tonality out of a film than that. Just sayin'.

That, and there's a lot more going on in scanning B&W negatives than just density. How you make that density has a lot to do with how the scanner can perform. Said another way, there's a large difference between scanning dye clouds and scanning metallic silver grain clumps.

Developing "their black-and-white film for scanning to a much lower contrast gradient than would work well for traditional darkroom printing" would make matters even worse as all the contrast expansion required for "digital" visualization (screen and prints) will be done past the hardware chain, by "stretching" scanned values mathematically in scanners software and/or editing programs like PS.

There was a time I thought this was true too. For a short while I reasoned that I should be developing B&W film to even higher densities for scanning, reasoning like you seem to be that "the hardware can handle it". But it turned out, the *film* can't handle that. And when scanned on a drum scanner, even though the scanner could handle it, the resulting scans were... awful? Because they faithfully reproduced how I had abused the film. My favorite film artifact was density waves at high contrast edges; looked sorta like ripples in a pond. Kodak had a name for that, but I've forgotten what they called it. Sigh... it was a long time ago.

So... I started lowering Dmax to see if there even was an optimum. It turned out there was. But it was different for different work flows. Just like finding your PEI, or your "N" development time. So you have to tune your workflow to get the most out of it. No big secret, this.

What I found was, with my camera, lenses, meter, film, developer, processing, scanner, software is that I got optimal results by keeping my Dmax about 1 to 1-1/2 stops lower down from where I would have wanted it to print in the darkroom. So... Dmax around 1.0.

Now the question is, why? And the answer isn't the density itself. It's how the density is created. Silver grain clumps are not translucent like dye clouds are. They reflect light. That is, light scatter. That is, Callier Effect. Just like you get in a darkroom enlarger. Scanners see this too.

So... what do I mean by "optimal results"? Same as the Zone System -- easy to print. When I optimized my negatives for scanning, I could actually, on occasion, make straight prints. I was never able to do this in darkroom printing. Never. But with this optimal-for-scanning workflow I came up with, I could make excellent prints (on occasion, I want to emphasis that again) that needed no dodging, burning, or other manipulation. It was eerie. It was cool. It saved a lot of time and work. And it looked better. Really worked well in the print.

After a lot more research I figured out why. Basically, I reduced Callier Effect sufficiently that my photographs were not experiencing much if any highlight compression. Highlight compression is one of the ways Callier Effect manifests itself. The light scattering from the highlight density of the negative tends to "contaminate" the dark areas by making them lighter, like a fog. When this is minimized, you don't have to fix the highlights in the print.

==================================================

But... you (anyone reading this, not just SergeyT) don't have to believe me. In fact, I'd be shocked if anyone did. It's easy enough to do your own testing find out what actually works for you and how you do things. Find the truth by doing the work. Evaluate. Come back here and tell this thread (or start a new thread) what you found. Just as I have now told you what I found. Information sharing is what this place is about -- we all want to learn from each other to improve our work.

[Alan Klein]

I wish there was a scientific explanation of what that "optimized for scanning" means . Otherwise consider that a marketing pitch.
Ok found some explanation and, as expected, nothing to worry about...
----
Press Release:

In a move that underscores their ongoing support of the professional photography market, Kodak have announced improved versions of their Kodak Professional Portra 400NC and 400VC films. The latest enhancements to the award winning Kodak Professional Portra colour negative film family build on the successful introduction of improved Portra films in 2006. The finer grain of the Portra 400NC and 400VC films, combined with the spectacular skin tone of the entire portfolio, enable professional photographers to create striking, breathtaking photographs.

“In the past 18 months, since we introduced the new generation of Portra film, we’ve sought ways to refine and improve the films to provide even better options for professional photographers,” said Mary Jane Hellyar, president, Film Products Group and executive vice president, Eastman Kodak Company. “These new 400NC and 400VC films, with their even finer grain, will enable photographers to create stunning images with spectacular skin tones and, in addition, provide improved scanning performance for greater enlargement capability.”



The new films offer:

Finer grain
Extended use of antenna dye technology enables Portra 400NC and 400VC films to deliver finer grain than ever before.

Outstanding scanning results
With finer grain and an emulsion overcoat specially designed for scanners, Portra 400 films reproduce beautifully, with either optical or digital output.
----

Those emulsions I believe use T grain which has different grain than the traditional films like Tri-x. Ektar negative color also uses them I believe.

As an aside I find Tmax 100 and 400 scan very nicely as well. The tones are greater and there's not as much contrast giving wide latitude to change the scan's appearance when editing digitally.

[Alan Klein]

Most consumer and prosumer flatbed scanners can't read the extreme density end of color positive films. It's difficult to do with flatbeds in general because of light scattering from adjacent pixels which you can't get away from in scan-a-line-at-a-time scanners. Most drum scanners, OTOH, can come sufficiently close, in large part because they are lighting just the one pixel at a time and thus controlling the lighting considerably better than a flatbed can.



While the Dmax of color positive films does normally (that is, almost always) exceed the Dmax of color negative and B&W negatives, it does not follow that the scanner hardware is under utilized when scanning negatives. This might be true for consumer and prosumer flatbeds. But it's certainly not true for most drum scanners, and I think for most professional flatbeds. Drum scanners in particular let the operator set the log amps right were the operator wants them, so the black and white points are exactly what the operator wants -- the scanner is neither underutilized nor over utilized. What that means is, you can make an exact match between the density range of the film and the digital range of the scanner. So if your scanner is good for, say, 14 bits of range, and your film density runs from 0.3 to 1.2, you can apply the 14 digital bits of range to cover exactly the 0.9 density range of the film. Hard to pull more tonality out of a film than that. Just sayin'.

That, and there's a lot more going on in scanning B&W negatives than just density. How you make that density has a lot to do with how the scanner can perform. Said another way, there's a large difference between scanning dye clouds and scanning metallic silver grain clumps.



There was a time I thought this was true too. For a short while I reasoned that I should be developing B&W film to even higher densities for scanning, reasoning like you seem to be that "the hardware can handle it". But it turned out, the *film* can't handle that. And when scanned on a drum scanner, even though the scanner could handle it, the resulting scans were... awful? Because they faithfully reproduced how I had abused the film. My favorite film artifact was density waves at high contrast edges; looked sorta like ripples in a pond. Kodak had a name for that, but I've forgotten what they called it. Sigh... it was a long time ago.

So... I started lowering Dmax to see if there even was an optimum. It turned out there was. But it was different for different work flows. Just like finding your PEI, or your "N" development time. So you have to tune your workflow to get the most out of it. No big secret, this.

What I found was, with my camera, lenses, meter, film, developer, processing, scanner, software is that I got optimal results by keeping my Dmax about 1 to 1-1/2 stops lower down from where I would have wanted it to print in the darkroom. So... Dmax around 1.0.

Now the question is, why? And the answer isn't the density itself. It's how the density is created. Silver grain clumps are not translucent like dye clouds are. They reflect light. That is, light scatter. That is, Callier Effect. Just like you get in a darkroom enlarger. Scanners see this too.

So... what do I mean by "optimal results"? Same as the Zone System -- easy to print. When I optimized my negatives for scanning, I could actually, on occasion, make straight prints. I was never able to do this in darkroom printing. Never. But with this optimal-for-scanning workflow I came up with, I could make excellent prints (on occasion, I want to emphasis that again) that needed no dodging, burning, or other manipulation. It was eerie. It was cool. It saved a lot of time and work. And it looked better. Really worked well in the print.

After a lot more research I figured out why. Basically, I reduced Callier Effect sufficiently that my photographs were not experiencing much if any highlight compression. Highlight compression is one of the ways Callier Effect manifests itself. The light scattering from the highlight density of the negative tends to "contaminate" the dark areas by making them lighter, like a fog. When this is minimized, you don't have to fix the highlights in the print.

==================================================

But... you (anyone reading this, not just SergeyT) don't have to believe me. In fact, I'd be shocked if anyone did. It's easy enough to do your own testing find out what actually works for you and how you do things. Find the truth by doing the work. Evaluate. Come back here and tell this thread (or start a new thread) what you found. Just as I have now told you what I found. Information sharing is what this place is about -- we all want to learn from each other to improve our work.

Bruce, I scan with a V850 (V600 before). I only set the levels. I shoot film at box speed and the lab develops my film, normal. I don't print much but when I do, from the scan, will I be missing something or doing something wrong?

[SergeyT]

SergeyT, you might be right, assuming that the amplifiers and/or ADC's behind the CCD are fixed gain in scanners. I don't know if that assumption holds true. Do you?
S/N ratios might indeed be far worse when working with high densities due to rather obvious reasons. My personal experience with a few scanners suggests there's in any case certainly no quality gain when trying to scan through very dense areas as opposed to lighter ones. I'm quite sure I'm not alone in this observation.

I read some logs from high end flatbeds a while ago and looked at maintenance utilities. From what I remember they adjust light intensity and exposure times. Leads me to believe that they operate just like digital cameras (set ISO, set exposure, etc) no more sophisticated than that. Prosumer CCD scanners may even have fewer controls at hardware level.
In theory S/N comes to play at very extermes (ex: Velvia 50). Densities of a typical negative film are way way less than that, hence the effect of negative film densities on S/N can be ignored.

[koraks]

SergeyT, it probably works similar to a digital camera alright. Since a CCD is essentially a capacitor that charges as photons hit it, the obvious sensitivity choice you have as an engineer is deciding how long you have the integration period be before you read out the signal. This process of reading out likely involves at least one stage of analog amplification, so there too is an opportunity to set gain. A third place might be present if ADC's are used with variable gain.

It follows from this that the longer you make the integration period and the higher you set the analog gain (or the ADC gain), the poorer the S/N ratio will become. Hence, density of the negative can be expected to influence the S/N of the digital signal. My experience as a user suggests that scanning through high-density areas results in increased noise. Just like a digital camera will create higher noise when set to a higher ISO.

You might quibble at what point this becomes objectionable; in my experience, any negative that's within a normal printing range, will scan more or less fine, perhaps with the exception of the densest highlights on negatives that would require a very low grade to print well. Hence, you might also argue about the validity of developing negatives for a particularly low gamma (that will become problematic when printing in the darkroom) to optimize the scanning process. I personally don't think that's necessary or helpful in any way, so I personally would never 'optimize' negatives for scanning. And also, personally, this is where the otherwise interesting discussion about how scanners work under the hood ceases to be very relevant for me

[SergeyT]

Most consumer and prosumer flatbed scanners can't read the extreme density end of color positive films.

Yes , they can, just not as effectively as drum scanners do. And nobody will argue with the fact that the main advantage of drum scanners over CCD scanners in tone reproduction fidelity, which is due to the technology used in one vs the other. Nothing can compensate for the difference. But that is irrelevant to the topic of advantages / disadvantages of making less contrasty negatives for scanning.

While the Dmax of color positive films does normally (that is, almost always) exceed the Dmax of color negative and B&W negatives, it does not follow that the scanner hardware is under utilized when scanning negatives. This might be true for consumer and prosumer flatbeds. But it's certainly not true for most drum scanners, and I think for most professional flatbeds.

I would put all CCD scanners into the same category when it comes to the ability to adjust scanning parameters . They all have a CCD sensor with its physical properties that translate into the limited ability to handle DR of the scanned source. And that CCD is likely to have a linear response , just like digital cameras have. But none of them is likely to exceed their capabilities while dealing with negative film.

Drum scanners in particular let the operator set the log amps right were the operator wants them, so the black and white points are exactly what the operator wants -- the scanner is neither underutilized nor over utilized. What that means is, you can make an exact match between the density range of the film and the digital range of the scanner. So if your scanner is good for, say, 14 bits of range, and your film density runs from 0.3 to 1.2, you can apply the 14 digital bits of range to cover exactly the 0.9 density range of the film. Hard to pull more tonality out of a film than that. Just sayin'.

Some do, some don't. Tango driven by NewColor hides a lot of hardware controls, if not most of them, from the operator and operator has to rely on whatever automation was embedded into the software by NewColor developers. Nevertheless, the combo delivers outstanding scans from any sort of film with ease.

That, and there's a lot more going on in scanning B&W negatives than just density. How you make that density has a lot to do with how the scanner can perform. Said another way, there's a large difference between scanning dye clouds and scanning metallic silver grain clumps.


There was a time I thought this was true too. For a short while I reasoned that I should be developing B&W film to even higher densities for scanning, reasoning like you seem to be that "the hardware can handle it". But it turned out, the *film* can't handle that. And when scanned on a drum scanner, even though the scanner could handle it, the resulting scans were... awful? Because they faithfully reproduced how I had abused the film. My favorite film artifact was density waves at high contrast edges; looked sorta like ripples in a pond. Kodak had a name for that, but I've forgotten what they called it. Sigh... it was a long time ago.

So... I started lowering Dmax to see if there even was an optimum. It turned out there was. But it was different for different work flows. Just like finding your PEI, or your "N" development time. So you have to tune your workflow to get the most out of it. No big secret, this.

What I found was, with my camera, lenses, meter, film, developer, processing, scanner, software is that I got optimal results by keeping my Dmax about 1 to 1-1/2 stops lower down from where I would have wanted it to print in the darkroom. So... Dmax around 1.0.

Now the question is, why? And the answer isn't the density itself. It's how the density is created. Silver grain clumps are not translucent like dye clouds are. They reflect light. That is, light scatter. That is, Callier Effect. Just like you get in a darkroom enlarger. Scanners see this too.

So... what do I mean by "optimal results"? Same as the Zone System -- easy to print. When I optimized my negatives for scanning, I could actually, on occasion, make straight prints. I was never able to do this in darkroom printing. Never. But with this optimal-for-scanning workflow I came up with, I could make excellent prints (on occasion, I want to emphasis that again) that needed no dodging, burning, or other manipulation. It was eerie. It was cool. It saved a lot of time and work. And it looked better. Really worked well in the print.

After a lot more research I figured out why. Basically, I reduced Callier Effect sufficiently that my photographs were not experiencing much if any highlight compression. Highlight compression is one of the ways Callier Effect manifests itself. The light scattering from the highlight density of the negative tends to "contaminate" the dark areas by making them lighter, like a fog. When this is minimized, you don't have to fix the highlights in the print.

That's an interesting dive-in into physics behind certain B&W film properties. And I'm glad that your discoveries help you to achieve desired results.

And BTW, in some way they also imply that it is impossible to make good darkroom prints as the required for darkroom printing densities lead to "abuse" of film.

I do not recall to ever advocating for abusing anything. And I'm not practicing photography for scientific research purposes. From my personal experience, the major effect on aesthetics of my prints have the composition that was chosen to photograph, the quality of light on the photographed scene and my ability to choose the correct camera settings for a given film. If these were great than I was extremely likely to get an excellent print. The rest is secondary and would fall together nicely if I simply followed my "standard" times and "formulas" for development and scanning. In other words, a properly exposed and normally developed negative film will cause no issues with scanning on any prosumer+ film scanner.

And since we are on this topic, I have a feeling that just because there is a complete freedom in how B&W film can be exposed and developed , the said exposure and development becomes a universe on their own and some photographers tend to make a bigger deal out of it than it really is. Let's compare that to color negative film. There is a single process that is called C-41 with particular chemistry, temps and timings for every step of the process . And there are few variations to it for + and - development, which are also particular and documented. In other words, it is a standard that guarantees and actually delivers consistent and predictable results. Does it make color photography inferior compared to B&W? What about color positive and E-6 process - practically no room to wiggle and nevertheless capable of delivering outstanding results. Yes, I would be concerned with what scanner to choose to get the best out of E-6 film but both normally exposed and developed C-41 or B&W films should be a no brainer for any film scanner's hardware (no need to bother with optimizing the densities). Tone reproduction fidelity of drum vs CCD is clear but not relevant to this topic.

[Tin Can]

I find it amazing how much a Photo EPSON Flatbed can pull out of a very thin negative

VERY VERY THIN