Choosing B&W film to optimize scanning results

[Michael Mutmansky]

Folks,

I'm getting a film camera again and I need to understand what film choices are out there for my needs... 120 film (6x6 and 6x17 probably). I'll shoot, have it developed by a lab, and then scan frames for digital printing and also pt/pd, gum, and gravure work.

I want to bea able to scan the film and produce a high quality image that can go to about 24" on the side (and whatever length if it is a 6x17). That means I need to be thinking carefully about the scanning step to make sure I can get something that doesn't look like a smeared mess...

Obviously, as the final print gets larger, the need for a better scanner increases, so if I am going really large, I may have those scanned by a service, but until that final step up, I plan to do the scanning here (either with a Microtek M1 that I have, or through a digital copy camera setup, I'll do some testing to see which will work better, but I expect the camera will be the better approach).

However, one of the things I have been looking around the web for answers about is the choice of a good, fine grained B&W film that supports scanning... or I guess, some conparisons of what a scanner will do with the variety of films that are available. Everything I've seen indicates that Ilford XP2 Plus will scan a bit easier because it is built with dye clouds rather than silver, so it won't have the grain ailasing that other films might have. But how does that compare to a fine grained film like Delta 100?

Obviously, for low resolution scans (small prints), it won't be important, but I want to be able to go large whenever needed, so I will probably coose my film based on how well the scan results are of one vs. the other. Since I won't be messing with any ZS stuff and will be letting a lab do the wet step, I'm not going to fret characteristic curves nearly as much as I would if I were planning to go straight to a pt/pd or silver print.

I've been away from film for over a decade now, so I need to pick back up on some of this from where I left off...


Thanks, ---Michael

 

[Alan Edward Klein]

Tmax 100 and Tmax 400. Both scan well and were made for scanning. Use Xtol if the lab provides. Tmax uses T grain. Other developers are very good too Use the 100 for long shots on a tripod. If you absolutely need the speed, then shoot Tmax 400. I use an outside lab.
See my FLickr for samples of 4x5 and other formats.
Tmax 400 35mm and 4x5, sorry no MF. https://www.flickr.com/search/?sort=date-taken-desc&safe_search=1&tags=tmax400&user_id=55760757@N05&view_all=1
Tmax 100 6x7 and 4x5 https://www.flickr.com/search/?sort=date-taken-desc&safe_search=1&tags=tmax100&user_id=55760757@N05&view_all=1

 

[Light Capture]

10x enlargement should be critically sharp from 120 film regardless of scanning method.
My most used film stock is HP5, Delta 100, Tmax 100 and 400 in that order. I can easily achieve 10x enlargement from them either scanning with digital camera or Nikon 8000 (provided that original image is sharp enough).
Didn't see much difference when it comes to scanning.

Digital camera scanning will yield better results but to achieve that it needs very precise setup. Film and sensor have to be parallel to very high tolerance.
I'm using precision level and geared head to measure level in two directions on film side and than match it on either camera mount or lens filter mount if it's in good shape.
In addition to this AN glass holder will further improve results.

Any macro lens should do well for scanning 120 film with digital camera. Scanning setup gets important once magnification reaches 1:1 or more.

 

[runswithsizzers]

If you are mostly concerned about resolution/sharpness, then am guessing the limiting factor is more likely to be scanner than the film? I'm sure someone will correct me if I am wrong.

 

[Michael Mutmansky]

Yes, I'm not taxing a good fine grained B&W film with a 10-12x enlargement, but I am thining about grain ailasing and smoothness of grain as being an issue.

I used to own a drum scanner, so I've done enough scanning in the past to have a good understanding of film and scanner capabilities, but I have not done a comparision of XP2+ to silver grain film. In the past, I developed myself as well, so I had no issue doing what I wanted/needed to achieve the desired results for printing or scanning (for example, I used to use Ilford print developer PQ Universal for my ULF pt/pd negs to get the contrast I needed while avoiding B+F buildup).

My thought was that I should consider the XP2+ for B&W shooting because it (in theory, anyway) can scan easier/better than silver based film. I'm not sure it will meet my quality requirements, which is why I'd love to see some conparisions.

 

[MattKing]

I would suggest reaching out to Paul Ozello - fatso here on Photrio: http://www.paulozzello.com/m_home.htm

 

[Michael Mutmansky]

I would suggest reaching out to Paul Ozello - fatso here on Photrio: http://www.paulozzello.com/m_home.htm

Thanks, I'll do that.

 

[albireo]

My thought was that I should consider the XP2+ for B&W shooting because it (in theory, anyway) can scan easier/better than silver based film.

My experience with negative scanning is that exposure and development are more important variables to manipulate than film type or brand if optimal results are sought. I routinely scan XP2+ with my Coolscan 8000ED as well as Foma film and these are all capable of delivering excellent results if I was careful with how I exposed and how I developed. A poorly exposed XP2+ negative will scan poorly compared to a well exposed and developed Foma 400 negative. For my taste, no amount of post-processing can rescue a severely under-exposed, over-exposed, or over-developed negative.

Of course, optimal exposure & development for scanning is a different endeavour than trying to do the same for darkroom printing - but given your experience with drum scanning I'm sure you already know that!

 

[runswithsizzers]

My experience with negative scanning is that exposure and development are more important variables to manipulate than film type or brand if optimal results are sought. I routinely scan XP2+ with my Coolscan 8000ED as well as Foma film and these are all capable of delivering excellent results if I was careful with how I exposed and how I developed. A poorly exposed XP2+ negative will scan poorly compared to a well exposed and developed Foma 400 negative. For my taste, no amount of post-processing can rescue a severely under-exposed, over-exposed, or over-developed negative.

Of course, optimal exposure & development for scanning is a different endeavour than trying to do the same for darkroom printing - but given your experience with drum scanning I'm sure you already know that!

I don't want to sidetrack this thread, but can you please point to a reference that discusses optimal exposure and development when the negatives are specfically intended for scanning? Thank you.

 

[Alan Edward Klein]

I don't want to sidetrack this thread, but can you please point to a reference that discusses optimal exposure and development when the negatives are specfically intended for scanning? Thank you.

I;ve found that a properly exposed picture scans best and edits the easiest afterwards. Of course, there;s latitude in making adjustments after the scan in post processing. The more to the middle of the histogram you can scan it, the more adjustments you have at both ends. But the bottom line is to expose the film properly.

 

[albireo]

I don't want to sidetrack this thread, but can you please point to a reference that discusses optimal exposure and development when the negatives are specfically intended for scanning? Thank you.

That's a great question. The answer is that I am not aware of any serious references (and by serious here I mean 'peer reviewed').

The field of hybrid photography is relatively new and needs solid quantitative foundations that essentially do not exist currently. In my limited, but growing, personal experience, I'm finding that truisms such as 'just expose and develop optimising for a grade 2 condenser enlarger print and your scans will be FINE" just don't always hold true. A scanner CCD sensor is a fundamentally different device than the pair (condenser enlarger, grade 2 paper) and a direct translation doesn't even begin to address a question of optimality.

I think it would be interesting to put together a working group, team up and do methodical experimentation to this end.

 

[Alan Edward Klein]

That's a great question. The answer is that I am not aware of any serious references (and by serious here I mean 'peer reviewed').

The field of hybrid photography is relatively new and needs solid quantitative foundations that essentially do no exist currently. In my limited, but growing, personal experience, I'm finding that truisms such as 'just expose and develop optimising for a grade 2 condenser enlarger print and your scans will be FINE" just don't always hold true. A scanner CCD sensor is a fundamentally different device than the pair (condenser enlarger, grade 2 paper) and a direct translation doesn't even begin to address a question of optimality.

I think it would be interesting to put together a working group, team up and do methodical experimentation to this end.

For whatever it's worth, when I shoot medium format chromes like Velvia 50, I bracket +1 and -1 stop. I always found that the one of three that look the best on a light table was the best to scan. To me, that suggests one should shoot for proper exposure. I also found that when scanning BW bracketed negatives also +1 and -1 flat without adjustments in the scanner, the shot that had most of the data towards the middle of the histogram, was the easiest to adjust after the scan in post-processing and gave the best latitude and room to adjust. Again, that implies to me that you ought to shoot for the right exposure.

 

[gone]

I'd be more concerned about making negatives w/ a good range of tonal values, and tightening up my exposure and compositions. If the negatives look good on a light table, they'll scan well. Scanning will do weird things w/ the grain too. The negs that look grainy on my monitor don't look like that on the printing end.

 

[Oren Grad]

Yes, I'm not taxing a good fine grained B&W film with a 10-12x enlargement, but I am thining about grain ailasing and smoothness of grain as being an issue....

Yes, with the Nikon 5000/9000 scanners I found grain aliasing with silver-grain negatives from ISO 400 films very annoying, which was a problem because for 35 and medium format TX is my favorite film. If you're limited to that level of scanning capability, either XP2+ or TMX definitely helps with that aspect of image character. I scanned a bit of both when I had the 5000 and 9000 but never did a head-to-head comparison. My inclination would be to go with XP2+ as it will be more forgiving all around - in the field, in commercial processing and in scanning, the latter notwithstanding the freedom to manipulate the curve in post. However, 12x is larger than I print so I can't comment directly on that, and in any case your tastes in print image character may be different from mine.

It's possible to adequately suppress grain aliasing from ISO 400 or old-style medium speed silver-grain negatives if you throw enough scanning horsepower at the problem. You'll have your own conclusions about the pros and cons of drum scanning from your experience. Thinking about camera-based options, in my experience, a single capture of a 35mm frame using a carefully-tuned A7RIV setup can more or less tame grain aliasing from those films. However, even allowing for pixel shift, you'd need to stitch to accomplish the same with medium format negatives from those films, and the workflow for 6x17 would be really burdensome. A GFX 100(S) should be able to do a bit better, though with obviously higher cost and more hassle finding/adapting a suitable lens. Of course, for logistical and/or cost reasons you might also find yourself using a camera for copying XP2+ or TMX too.

If you think you might consider the GFX route, have a look at Jim Kasson's blog for his copy setup and his test results on various adapted lenses in that application.

Good luck - will be interested to hear what solution you end up with.

 

[Michael Mutmansky]

Oren,

Yes, I've seen Jim's setup. He can tend toward the extreme end on things, but I suspect that it will produce excellent results. I was a bit suprised that he is doing both a focus bracket and fussing with finding the flattest field lens for his needs. It seems that if he were focus bracketing, he wouldn't need to be that concerned with the lens, as long as it has very good sharpness (but not necessarily in a flat field).

I digitized several thousand 35mm chromes a few years ago for my father (mostly Kodachrome) and using a Sony a7r2 I was easily able to reproduce the grain in that film. I did some testing with the Fuji 50R as well and it improved on the Sony setup (I think because the 50R has only about a 70% well fill on the photosites, so it produces better pixel-to pixel seperation). THe Sony 90 macro lens has a little bit of field curvature at 1:1, so it required me to focus about 2/3 of the way to the edge and then stop down, but it worked very well. I actually shifted the shooting over to some Olympus macro bellows lenses because they are symmetrical in design and appeared to have better flatness of field... so I now have a more or less complete Olympus macro bellows setup here that I can use with the Fuji 50R and any other camera with an adapter.

My thought on the 6x17 is that I should be able to do that in 3 shots and stitch in LR before doing the negative conversion. 5 shots or maybe 6 is I wanted to use the longer side vertically to increase the sampling size, but I suspect that doing a full resolution 3x stitch will be resolving the grain on most films anyway, so there really won't be any reason to go higher. And yes, if I decide to get a 100MP body, I'll gain a good bit of pixel density right off the bat as well.

I used to have both a Nikon 5000 and the 9000 and also the old Colorgetter Falcon... They were remarkably good scanners for the money. I no longer have the comparision files, but I did sacns on the CS5000 and the CGF at max resolution (was it 4000spi for the CS5000?) and the Nikon scanner held it's own very, very well compared to the drum scanner. No, it couldn't handle the same dmax, and it had focus issues unless you had the slide glass mounted, but when optimally set up, it was really excellent for the money. The same for the SC8000/9000.

My thought is that I will probably want to try Ilford 100 delta and 400 delta in XTOL or similar and maybe even with a pull process to keep grain buildup from happening. Deal with the speed loss associated with that... I suspect it will produce better results that way, but before i go there, I will also just try straight processing, etc. and see what kind of results I get.

 

[albireo]

A correctly used Coolscan 9000 operating within factory specs most definitely focuses on the grain and does not produce any aliasing. Sounds like you used a faulty couple of units or you had very curly negatives.

Yes, with the Nikon 5000/9000 scanners I found grain aliasing with silver-grain negatives from ISO 400 films very annoying, which was a problem because for 35 and medium format TX is my favorite film. If you're limited to that level of scanning capability, either XP2+ or TMX definitely helps with that aspect of image character. I scanned a bit of both when I had the 5000 and 9000 but never did a head-to-head comparison. My inclination would be to go with XP2+ as it will be more forgiving all around - in the field, in commercial processing and in scanning, the latter notwithstanding the freedom to manipulate the curve in post. However, 12x is larger than I print so I can't comment directly on that, and in any case your tastes in print image character may be different from mine.

It's possible to adequately suppress grain aliasing from ISO 400 or old-style medium speed silver-grain negatives if you throw enough scanning horsepower at the problem. You'll have your own conclusions about the pros and cons of drum scanning from your experience. Thinking about camera-based options, in my experience, a single capture of a 35mm frame using a carefully-tuned A7RIV setup can more or less tame grain aliasing from those films. However, even allowing for pixel shift, you'd need to stitch to accomplish the same with medium format negatives from those films, and the workflow for 6x17 would be really burdensome. A GFX 100(S) should be able to do a bit better, though with obviously higher cost and more hassle finding/adapting a suitable lens. Of course, for logistical and/or cost reasons you might also find yourself using a camera for copying XP2+ or TMX too.

If you think you might consider the GFX route, have a look at Jim Kasson's blog for his copy setup and his test results on various adapted lenses in that application.

Good luck - will be interested to hear what solution you end up with.

 

[Oren Grad]

Yes, I've seen Jim's setup. He can tend toward the extreme end on things, but I suspect that it will produce excellent results. I was a bit suprised that he is doing both a focus bracket and fussing with finding the flattest field lens for his needs. It seems that if he were focus bracketing, he wouldn't need to be that concerned with the lens, as long as it has very good sharpness (but not necessarily in a flat field).

I think the focus bracketing is also addressing a different problem - alignment. If you're willing to make a camera-plus-copy-stand into a dedicated scanner at a fixed reproduction ratio, you can contrive ways to lock everything in place so it's not necessary to redo the painstaking adjustment every time. But if you work at variable ratios and/or need to use the camera and stand for other purposes, it's a problem. The focus bracketing should be able to compensate for some alignment slop, at the cost of way more processing overhead. Pick your poison!

I'm currently exploring higher-precision mounting/leveling solutions to make the alignment problem less burdensome for my own copying.

ADDED: Was just looking at some of Jim's posts about his copy setup - he states that the focus stacking is to address negative curvature and alignment issues.

 

[Michael Mutmansky]

A correctly used Coolscan 9000 operating within factory specs most definitely focuses on the grain and does not produce any aliasing. Sounds like you used a faulty couple of units or you had very curly negatives.

Grain aliasing is a complex problem and focus is only one part of it. If fact, focus (rather, defocusing) is one method to help alleviate it.

The classical definition of grain aliasing is the accentuation of the film grain in a scan because the scanning sampling pattern overlays on the grain in a manner that can introduce a higher visual impact of the grain (thus, the grain looks, uh, "grainier") than it is in the actual film.

Factors that impact this are the size of the grain, the sampling frequency of the scanner, the size of the scanner sampling "dot", the spacing of these "dots", whether there is a gap between the dots (or if they overlap), and a bunch of other factors related to the optical performance of the scanner and also the nature of the film grains. Plus, throw in the moon cycle and whether Mercury is rising for good measure...

I garauntee a CS9000 can and will produce grain aliasing at times and there may not be much you can do to avoid it when the conditions are just right, other than change the film you shoot. Since that scanner doesn't allow focus adjustments (to defocus the sensor slightly) (at least I don't rmember it being able to do that), and you can't change the sampling "dot" size, you have very few options to avoid the issue. But it can certainly produce grain aliasing.

 

[Michael Mutmansky]

I think the focus bracketing is also addressing a different problem - alignment. If you're willing to make a camera-plus-copy-stand into a dedicated scanner at a fixed reproduction ratio, you can contrive ways to lock everything in place so it's not necessary to redo the painstaking adjustment every time. But if you work at variable ratios and/or need to use the camera and stand for other purposes, it's a problem. The focus bracketing should be able to compensate for some alignment slop, at the cost of way more processing overhead. Pick your poison!

I'm currently exploring higher-precision mounting/leveling solutions to make the alignment problem less burdensome for my own copying.

ADDED: Was just looking at some of Jim's posts about his copy setup - he states that the focus stacking is to address negative curvature and alignment issues.

Yes, good point. That will certainly help eliminte any issues related to having the camera or film move on the rail and cause alignment issues.

 

[Oren Grad]

Yes, good point. That will certainly help eliminte any issues related to having the camera or film move on the rail and cause alignment issues.

I use a glass sandwich carrier so negative flatness isn't an issue here (at the cost of more effort cleaning and spotting!). But FWIW, in my setup I've found tilt to be as much or more of a problem as fore-aft drift when I'm working with the copy stand and continually having to re-mount the camera and the negative carrier. When you're at the point of paying attention to the grain character it doesn't take much deviation to cause obvious problems in the capture.

The latest high-MP systems are wonderfully powerful tools, but just as in field work, taking full advantage of the extra information-recording capacity for copying requires more rigorous technique all around.

 

[albireo]

I garauntee a CS9000 can and will produce grain aliasing at times and there may not be much you can do to avoid it when the conditions are just right, other than change the film you shoot. Since that scanner doesn't allow focus adjustments (to defocus the sensor slightly) (at least I don't rmember it being able to do that), and you can't change the sampling "dot" size, you have very few options to avoid the issue. But it can certainly produce grain aliasing.

This is a controversial topic and one that doesn't belong to this thread, so I will refrain from commenting in depth with my experience on this issue. The cause of grain aliasing is clear to anyone with some level of understanding of sampling theory. However, identifying it - distinguishing it from plain unaliased grain - is not easy.

I will just add that all examples of claimed "grain aliasing" I've seen (as related to high quality consumer dedicated film scanners endowed with CCD line sensors and excellent optics such as the Coolscans) were traceable to emphasized grain effects in strongly overdeveloped, and/or underexposed negatives which had been produced in a (failed) attempt to optimise for a traditional wet printing workflow. These were then strongly altered digitally to restore or recover some of the desired content in spite of sub-optimal exposure and processing. Again, with my scanner, my exposure/processing workflow and the film/camera combinations I use, grain aliasing has not been an issue so far.

 

[PhilBurton]

Of course, optimal exposure & development for scanning is a different endeavour than trying to do the same for darkroom printing - but given your experience with drum scanning I'm sure you already know that!

As someone who doesn't know the difference, what are the key points? I'm hoping that "soon" I can get my Nikon F2 bodies out of the closet and start doing B&W film with home processing.

 

[destroya]

I shoot 6x17 scan and print (cause i dont have a 5x7 or larger enlarger). you are talking 24 inches on the height if I read you post right, so a 10x enlargement. for B&W most 100/125 speed films and slower will work as well as tmax 400. if you are having a lab develop, then I would suggest doing it yourself if you are as worried about results as you seem to be. for me, the best developer FOR SCANNING has been pyro-M. it really helps with the grain. even tri-x 120 shots enlarged 10x dont show grain or mushiness which seems to worry you. also, the scanner is gonna make or break your results. I used to scan my 6x17 film on a nikon 9000 and stitch in photoshop, but after getting a flextight scanner i can now do it on 1 pass. slide film will scan your size no problem. I have scanned and printed at the sizes you are talking about and they look great. but be careful no to over sharpen you scans!

all things being equal, the scanner will be the deciding factor for you. a flatbed will not work for what you are looking to do. if its really important, get a drum scan, be done and stop worrying about it. otherwise, its a compromise.

i have shot a total of 1 roll of xp2 and it didnt work for me. I would look at delta 100, tmax 100, acros or panF

john

 

[tokam]

If you are looking to get the best out of your BW negs using a hybrid workflow then checkout Youtube channel of Martin Henson and associated forum digitalmonochromeforum. His technique of linear raw scanning and contrast grading during post processing produce results that would take many years experience as a master darkroom printer to achieve. Using both a Coolcan 5000ED and Epson V700 I am already seeing a huge improvement although the learning curve is a bit steep. Even seeing a reduction in grain. His technique revolves around avoiding the image manipulation caused by the scanner manufacturers software and Vuescan and providing a 'raw' image for subsequent manipulation in post.

 

[alanrockwood]

Grain aliasing is being discussed in this thread. I am pretty familiar with Fourier transforms and aliasing from a mathematical point of view. However, I have yet to see a good discussion on the origin of grain aliasing that goes beyond vague generalities but is reasonably rigorous and also reasonably accessible to me. Can anyone point to a good discussion of the origin of grain aliasing.

For the record, aliasing occurs when there are frequencies (spatial frequencies in this case) that are above the Nyquist limit. The signal acquisition system (the scanner) must scan at a rate that is more than twice that of the highest spatial frequency in the image. Any signal frequencies above that limit will be shifted to lower frequency, resulting in artifacts in the image.

Grain can be considered as somewhat of a random signal, noise if you will. If grain aliasing is truly the result of aliasing in the usual meaning of the term, then the high spatial frequency components of the image can be shifted down in frequency. Possibly this is at the heart of why some scans seem to emphasize grain, but I'm not understanding this in any more detail than this generality.

I started to say that grain is noise. Noise can be characterized by a frequency spectrum. So I guess one way to look at this problem is to ask what is the noise spectrum of grain. If it is true white noise then the frequency spectrum would extend to infinity, with equal amounts of noise at each frequency. This could produce a lot of aliasing. In real life the frequency spectrum of grain won't extend to infinity, and the frequency spectrum is probably not truly flat, so grain won't be true white noise, especially in the high frequency range, but what does the noise spectrum of grain look like?

Also Michael pointed out that grain aliasing can be affected by a lot of things, including the structure of the sampling device...