Scanning: A Story of a Quest to Scan B&W Film with a Happy Ending!

[monkeytumble]

Yes, there is a Scanning Nirvana. For me, it is Nikon CoolScan 9000 with FH-869G medium format film holder, a PC running Vista 64, Vuescan, and a eye-one display 2 screen calibration system. Not to mention B&W film, developing tanks, chemicals, thermometers, and assorted other film developing accessories.

It took a leap of faith (and few thousand dollars) that it could all work together and that I would be able to get it all to work together. But, it does and I did. And, I am now starting to generate results that I'm satisfied with.

There are a few things that I still have yet to incorporate into my evolving work flow, including wet mounting film and use of third party noise reduction and sharpening software. But, I am getting there. Was it worth it? In a word, yes!

To Sandy King, Nuno Souto, JD, and Christopher Crawford who provided helpful advise to find my path. Gentlemen, thank you.

To those on the quest, I say keep the faith, happy scanning endings to do happen!

Jay Decker

 

[nsouto]

Well done! Excellent results.
Enjoy that excellent scanner, it's a truly amazing tool.

 

[makanakijones]

Marvellous portrait!
I want to know (please) about the technical aspects of this photos (film and developing process).

 

[Jordan]

I love the tonality but they look a little over-sharpened to me. What kind of sharpening routine are you using?

 

[monkeytumble]

Marvellous portrait!
I want to know (please) about the technical aspects of this photos (film and developing process).

The photograph was taken in our music room with evening light entering the room from a window camera left. The background was a window and a redwood deck on the other side that was also lit by the evening light. The camera was a Bronica SQ-Ai with a 180mm f/4.5 lens. This lens focuses as close as 1-meter, which makes it ideal for medium format headshots. The camera was mounted on a tripod and scene was metered with an incident light meter. The exposure was 1/60 of a second at f/5.6. One other note regarding the camera, I use a 45 degree prism finder for portraits that helps me greatly during composition and focusing (and any other situation where I need to get lower for a more interesting perspective on the subject).

The image was captured on Tri-X 400, exposed at an EI of 1250, developed in Diafine, and fixed in TF-4. I use a standardized development process with Diafine that I picked up from the Figital Revolution web site (see the Diafine Developer UNIVERSAL TIME/ AGITATION PROCEDURE at http://figitalrevolution.com/2008/06/09/new-testing-results-for-inkaid-and-film-processing/). I have found Diafine to be ideal for my Hybrid process for the reasons Stephen Schaub describes at the Figital Revolution (if you are really interested go to http://figitalrevolution.com search on the word Diafine and go the last page of results and start at the beginning) and Sandy King presents in his July/Aug 2008 View Camera article (http://www.viewcamera.com/documents/pages48-55.pdf). The only thing that I wish that Diafine did was stained the negatives like Pryrocat HD. Maybe Sandy King can develop a staining and compensating developer that is as easy and robust to use as Diafine if he did, I for one would be all over it in a New York minute.

The negative was scanned with a Nikon CoolScan 9000 using the FH-869G film carrier. I quickly batch scan the roll of negatives in B&W at 1,333 dpi using Vuescan and import them into Lightroom where they receive some quick exposure and contrast adjustments and are evaluated. This portrait image is straight out of Lightroom with basic exposure and contrast adjustments and a little spotting. Images processed in Diafine tend to be rather flat with a long tonal scale, which is good for scanning and easily adjusted in software. The image was placed into an InDesign template that I use to create folio images using techniques that I learned/borrowed/stole from Brooks Jensen and Lenswork (Dead Link Removed. The post scan processing to get the image in this form including InDesign folio presentation template was quick, say 5 to 10 minutes. Note, this is my work print process. If this work print makes it through my project final print selection process, the negative will be wet mounted, rescanned as a positive at a higher resolution, and more elaborate processing will be applied, e.g., noise reduction, finer tonal adjustments, sharping, etc.. The 8.5x11 inkjet work print of this image is just short of stunning. The image its self is about 7-inches square in this folio presentation format.

The other image was generally created the same way. The differences were that the film was FP4+ exposed at an EI of 250, which incidentally was processed in the same tank with the roll of Tri-X that the portrait was on, the camera was Bronica RF645 with 65mm lens, and it received about 2 minutes of additional localized contrast adjustments in Lightroom. While the 8.5x11 work print of this image is good, this image would more definitely benefit from wet mount scanning at a higher resolution, more elaborate post scan processing, and probably most importantly, presentation as a larger print.

I have a very product oriented work flow. If you are seriously interested, Id be glad to share an overview of my product oriented work flow process.

 

[monkeytumble]

I love the tonality but they look a little over-sharpened to me. What kind of sharpening routine are you using?

The images were not sharpened. Think images were converted to a pdf in InDesign and then converted to a jpg in Acrobat. Through these conversions the image integrity started breaking down.

 

[Ray Heath]

g'day monkey
can you briefly explain what problems you initially encountered in scanning and how these images show a rectification of these issues?

 

[monkeytumble]

g'day monkey
can you briefly explain what problems you initially encountered in scanning and how these images show a rectification of these issues?

Superb question Ray!

I initially tried to use a flat bed scanner with a transparency adapter and a film holder for scanning 6x6 film. The problems where 1) the lack of dynamic range, 2) poor scanner focus, 3) curling of the film in film holder, which created additional focus problems, 4) the software's inability to interpret the grain in B&W negatives, and 5) the scanner manufacture's overstatement of the scanning resolution.

Moving to a quality dedicated film scanner, in my case the Nikon 9000, solved the hardware related problems, e.g., resolution, dynamic range, scanner focus.

For reasons unknown to me, Vuescan interprets the grain in B&W negatives better than any other software the I've used, including SilverFast. The smooth tones and grain in the image files and prints is world better than I anything I ever achieved with SilverFast or scanner manufacture provided software. Vuescan also works stabily with the 64-bit version of Windows Vista.

Using a film holder that sandwiches the film between glass plates solved the film curling problems. The glass holder keeps the film flat during scanning so that image is "grain sharp" across the entire frame. I'm working on fluid mounting now and it is an incremental improvement over the glass holder. The big step change improvement was moving to a glass holder.

That help?

 

[Ray Heath]

thnx monkey

so these images "look" grainy because they are (ISO 400 pushed to 1250)?

 

[monkeytumble]

thnx monkey

so these images "look" grainy because they are (ISO 400 pushed to 1250)?

They look grainy because the image brokedown a bit in the conversions from to pdf and then to jpg. If you look at the text you can clearly see signs that the text image is breaking down, the same thing is happening in the image.

Here's the pdf of the image: Portrait

 

[sanking]

Jay,

Both images of very nice. The outdoor landscape has a very dramatic look as if something really big is about to happen, and the portrait of the little girl, your daughter I assume, is very warm and intimate.

Regarding Pyrocat-HD, it is possible to use it as a two-bath developer. You dilute Stock A 1:10 with water to make working bath A, and Stock B 1:10 with water to make working bath B. Develop for five minutes in each at 75F. You can develop several more rolls in the working solution if in the same developing session, but at the end of the session discard the working solution.

Diafine with TRI-X gives quite a bit of grain. In the case of both images you posted the grain needs to be subdued, IMO. You should consider investing in a noise reduction software like Noise Ninja or Neat Image. With this software you can profile your film to reduce grain a lot without loss of sharpness.

Sandy King

 

[makanakijones]

In case you don't want to obtain grain I think is better to buy a digital camera.
Grain is film, grain is beautiful.

 

[sanking]

LF film that is contact printed or enlarged no more than about 3X-4X is virtually grain free, is capable of very fine detail, and has a kind of "creamy" tonal look. That is the kind of look I like in my B&W prints, and one should be able to get it with MF negatives with slow or medium speed films and appropriate exposure and development/scanniing technique.

Grain is only beautiful to me when it is so "fine" that it does not interfere with my concept of the way a photograph should look.

Sandy King

In case you don't want to obtain grain I think is better to buy a digital camera.
Grain is film, grain is beautiful.

 

[monkeytumble]

Jay,

Both images of very nice. The outdoor landscape has a very dramatic look as if something really big is about to happen, and the portrait of the little girl, your daughter I assume, is very warm and intimate.

Regarding Pyrocat-HD, it is possible to use it as a two-bath developer. You dilute Stock A 1:10 with water to make working bath A, and Stock B 1:10 with water to make working bath B. Develop for five minutes in each at 75F. You can develop several more rolls in the working solution if in the same developing session, but at the end of the session discard the working solution.

Diafine with TRI-X gives quite a bit of grain. In the case of both images you posted the grain needs to be subdued, IMO. You should consider investing in a noise reduction software like Noise Ninja or Neat Image. With this software you can profile your film to reduce grain a lot without loss of sharpness.

Sandy King

Thanks Sandy!

Have a couple questions for you...

1) Love FP4+ developed in Pyrocat-HD! What EI do you shoot FP4+ when you use it in a two bath developer? And, what is the agitation do you use with tank development? What about LF drum development?

2) You are right about the grain of Tri-X in Diafine. I use Noise Ninja and it can eliminate the grain. However, there is an aethetic to grain that I do like in some photographs. I've just developed my first roll of Acros 100 in Diafine, haven't scanned it yet, and it looks like it is going to be beautiful... (wonder how Acros looks when developed in Pyrocat-HD?)

Thank you for your help getting to this point... now it is up to me to see if I can produce any interesting work.

Jay

 

[sanking]

Jay,

I would expose FP4 at an EI of 100 for development in a two bath developer. For tank development I agitate for the first thirty seconds, then for ten seconds every minute thereafter. For drum development with constant agitation dilute the working solutions 1:1 with water and develop for the same time, but discard the diluted solution after use.

Acros100 is my favorite film for MF work. It gives beautiful results in both Pyrocat-HD and in divided D23. Both Pyrocat and divided D23 give slightly finer grain then Diafine, IMO. Course, my observations are based on scanning since I don't optically enlarge film anymore.

The advantages of scanning, correcting the file in Photoshop and then printing digitally are very great as most everyone who has worked this way would agree. It is so obvious that "even a cave man would get it," to use a modified line from a Geico commercial.

Sandy

Thanks Sandy!

Have a couple questions for you...

1) Love FP4+ developed in Pyrocat-HD! What EI do you shoot FP4+ when you use it in a two bath developer? And, what is the agitation do you use with tank development? What about LF drum development?

2) You are right about the grain of Tri-X in Diafine. I use Noise Ninja and it can eliminate the grain. However, there is an aethetic to grain that I do like in some photographs. I've just developed my first roll of Acros 100 in Diafine, haven't scanned it yet, and it looks like it is going to be beautiful... (wonder how Acros looks when developed in Pyrocat-HD?)

Thank you for your help getting to this point... now it is up to me to see if I can produce any interesting work.

Jay

 

[monkeytumble]

I would expose FP4 at an EI of 100 for development in a two bath developer. For tank development I agitate for the first thirty seconds, then for ten seconds every minute thereafter. For drum development with constant agitation dilute the working solutions 1:1 with water and develop for the same time, but discard the diluted solution after use.

Acros100 is my favorite film for MF work. It gives beautiful results in both Pyrocat-HD and in divided D23. Both Pyrocat and divided D23 give slightly finer grain then Diafine, IMO. Course, my observations are based on scanning since I don't optically enlarge film anymore.

Sandy, thank you for the information on divided Pyrocat-HD.

Since I returned to film capture, I have been working with negatives that have developed in four developers: 1) Pyrocat-HD, 2) HC-110 (using the 1+49 dilutions by Jason Brunner at http://www.jasonbrunner.com/hc110.html), 3) Diafine, and 4) D-23 (from a single shot solution recipe by Mike Brenner in the July 1975 issue of Petersen's Photographic).

Here are my general observations these four developers with FP4+ film:

1) From the perspective of film grain, the D-23 produces the finest grain when optically magnified.

2) From the perspective of producing the longest tonal scale, Diafine produces the longest tonal rendering (qualative assessment based comparison the histograms of similarly lit images).

3) Scanned highlights on film developed in Pyrocat-HD have really nice aethetic to me.

4) All four developers have produced negatives that have scanned well enough to produced good prints. However, the compensating nature of Diafine seems to produce consistently good scanning results.

Right now I really like Diafine, because it is easy to use with non-critical developing temperature and time, it's long life, an apparent effective film speed increase, and it consistently produces negatives that scan well. In you judgment and opinion do divided D-23 and divided Pyrocat-HD produce sensibly better results in the print, e.g., can they make the difference between a technically marginal print and a good print? If so, why, e.g., fine grain, staining effect, etc.? How have you answered this question for yourself?

 

[sanking]

I have not tested the limits of any of these three developers in terms of print size that would seriously challenge any of them with the fine grain B&W films that I use. If you look with a microscope at an Acros negative developed in Diafine, divided D23 and divided Pyrocat-HD you will see a bit more grain in the one developed in Diafine, with the other two about equal. However, the scanner, method of scan, and type of post-processing introduce variables that are greater than the observed differences in the negative. I personally stick with divided D23 and divided Pyrocat-HD for Acros because I like to mix my own developers and both of these developers are very inexpensive to mix and use.

At a final print size of about 18X21" from a 6X7cm Acros negative I really don't think you will see much different in image quality. If you were to push this out to 36X41" I suspect that the divided D23 and divided Pyrocat-HD negatives will give better prints, but that is an untested opinion.

Sandy King

Right now I really like Diafine, because it is easy to use with non-critical developing temperature and time, it's long life, an apparent effective film speed increase, and it consistently produces negatives that scan well. In you judgment and opinion do divided D-23 and divided Pyrocat-HD produce sensibly better results in the print, e.g., can they make the difference between a technically marginal print and a good print? If so, why, e.g., fine grain, staining effect, etc.? How have you answered this question for yourself?

 

[Ray Heath]

I have not tested the limits of any of these three developers in terms of print size that would seriously challenge any of them with the fine grain B&W films that I use. If you look with a microscope at an Acros negative developed in Diafine, divided D23 and divided Pyrocat-HD you will see a bit more grain in the one developed in Diafine, with the other two about equal. However, the scanner, method of scan, and type of post-processing introduce variables that are greater than the observed differences in the negative. I personally stick with divided D23 and divided Pyrocat-HD for Acros because I like to mix my own developers and both of these developers are very inexpensive to mix and use.

At a final print size of about 18X21" from a 6X7cm Acros negative I really don't think you will see much different in image quality. If you were to push this out to 36X41" I suspect that the divided D23 and divided Pyrocat-HD negatives will give better prints, but that is an untested opinion.

Sandy King

Jay and Sandy please forgive my past indiscretions and allow me a (long) question;

Sandy, your proposing quite large print sizes, would you have the prints made externally, what current digital printers will do justice to your carefully prepared files, what is the availabilty of such devices, and what would one expect to pay for such a service?

 

[sanking]

Ray,

I don't normally print any larger than seventeen inches wide because that is the limit of my printer, an Epson 3800. The detail possible from the 3800, and in most current generation Canon, Epson and HP photo printers, is well beyond the limit of human resolution so if there is sufficient detail in the image file digital prints can be as sharp as optical prints. Studies on the the human eye show that 5 lp/mm is the threshold of human resolution, whereas the current photo printers are capable of producing 7-10 lp/mm. The Epson 3800 cost me around $900. Photo printers with 13" wide printing areas re available for around $500 USD new.

Whether or not the image file used to make the digital print actually has the equivalent of 5 lp/mm depends on many factors, primarily the amount of detail in the original negative and in the quality of the scan. However, a Mamiya 7II negative, exposed with the camera on a tripod at an optimum aperture, should be able to give resolution of around 80 lp/mm. If you can capture all of that detail with your scanner, which requires *real* resolution of around 4000 spi, you should be able to print up to 36X44" and still have resolution equivalent to 5 lp/mm. I know that is possible because I saw last year an exhibition of color digital prints of this size made from drum scans of Mamiy 7II transparencies. At normal viewing distance the prints looked as good as one would expect from 4X5 or even larger color film.

The key of course to digital printing from scanned film is the scanner. Pulling all of the detail out of a good piece of film requires at minimum a dedicated film scanner like the Nikon LS-9000, and a drum scan would be even better.

So what is the minimum cost to do high quality digital printing of scanned film? For MF, it would be around $3000, assuming about $2K for the scanner and $1K for a photo printer. I don't shoot 35mm so that would be another comparison.

Sandy King

Jay and Sandy please forgive my past indiscretions and allow me a (long) question;

Sandy, your proposing quite large print sizes, would you have the prints made externally, what current digital printers will do justice to your carefully prepared files, what is the availabilty of such devices, and what would one expect to pay for such a service?

 

[jd callow]

Add to the minimum:
image editing software priced from 100.00 to 500.00
computer with enough power, RAM and hdd space priced form 0 (already sitting on my desk) to the sky's the limit

Highly recommended:
Monitor Calibration: 100 - 300.00
RIP (rastor image processor) 200.00 to 10k

Really nice to have:
Reflective sensitometer with printer calibration/profile software (300.00 w/o rip to 10k with rip)


Without a monitor calibration device you are not getting a large portion of the benefit of the ADD/ADA workflow.

Figure for 35mm:
2k for scanner & printer, 150.00 for calibration, and 700 for hw/sw upgrades to the avg home machine: 2850.00

For MF:
3k for scanner & printer, 150.00 for calibration, and 1000 for hw/sw upgrades: 4150.00

For 4x5:
6k for (used imacon or low end drum) scanner plus 24" printer, 150 for calibration, and maybe an additional 1250.00 in upgrades for a total of 7400.00

Cost of ownership can be very high as well. Imacons can be finicky, drums use consumables and be expensive to fix. Ink jet printers can easily consume ink costing more than the printer and paper can get very expensive as well and we must not forget the cost of upgrading software every couple years.

There is a real cost to going digital in a substantial way and it is a personal decision as to the value of doing so.


I don't like inkjets prints enough to purchase a really nice one for personal use. For my work I much prefer digital RA4 which can run anywhere from $3 to about $20 per square foot. By not buying an ink jet and sending the work out to be printed you can knock 1k off of the cost and possibly break even on the cost of ink v. the cost of digital RA4 output.

 

[Kerik]

JD, what would your recommend for a monitor calibration device? Whatever the current eye one thingy is?

 

[jd callow]

That or the spyder. After or above that they are really just intended so that you can properly match PMS colours and corporate ID/product standards -- in other words The spyder and eye 1 will get you 80%+ of the way.

 

[sanking]

Your figure for MF seems about where I am.

For 35mm, about which I know little or nothing, I thought that dedicated film scanners of 4000 spi were available for less than $1k? Have these all been discontinued? If so, there must be many of them on the used market?

For 4X5, an Epson V750 at less the $1k would give acceptable results at up to 3X-4X.

BTW, note that many of these expenses can be calcuated on a multi-year basis. A professional quality scanner, assuming it does not need repair, should give results as good in five years as today. By that time there will probably be digital cameras that exceed MF in quality at a reasonable cost. I have no doubt but that for color work my 15 mp Canon 50D is at least as good as the best 35mm color, and it is not far from 6X4.5 quality.

Sandy King

Figure for 35mm:
2k for scanner & printer, 150.00 for calibration, and 700 for hw/sw upgrades to the avg home machine: 2850.00

For MF:
3k for scanner & printer, 150.00 for calibration, and 1000 for hw/sw upgrades: 4150.00

For 4x5:
6k for (used imacon or low end drum) scanner plus 24" printer, 150 for calibration, and maybe an additional 1250.00 in upgrades for a total of 7400.00

 

[pschwart]

Add to the minimum:
image editing software priced from 100.00 to 500.00
computer with enough power, RAM and hdd space priced form 0 (already sitting on my desk) to the sky's the limit

Highly recommended:
Monitor Calibration: 100 - 300.00
RIP (rastor image processor) 200.00 to 10k

Really nice to have:
Reflective sensitometer with printer calibration/profile software (300.00 w/o rip to 10k with rip)


Without a monitor calibration device you are not getting a large portion of the benefit of the ADD/ADA workflow.

Figure for 35mm:
2k for scanner & printer, 150.00 for calibration, and 700 for hw/sw upgrades to the avg home machine: 2850.00

For MF:
3k for scanner & printer, 150.00 for calibration, and 1000 for hw/sw upgrades: 4150.00

For 4x5:
6k for (used imacon or low end drum) scanner plus 24" printer, 150 for calibration, and maybe an additional 1250.00 in upgrades for a total of 7400.00

Cost of ownership can be very high as well. Imacons can be finicky, drums use consumables and be expensive to fix. Ink jet printers can easily consume ink costing more than the printer and paper can get very expensive as well and we must not forget the cost of upgrading software every couple years.

There is a real cost to going digital in a substantial way and it is a personal decision as to the value of doing so.


I don't like inkjets prints enough to purchase a really nice one for personal use. For my work I much prefer digital RA4 which can run anywhere from $3 to about $20 per square foot. By not buying an ink jet and sending the work out to be printed you can knock 1k off of the cost and possibly break even on the cost of ink v. the cost of digital RA4 output.

There is a real cost to going digital in a substantial way and it is a personal decision as to the value of doing so.
No argument there. For a meaningful $$$ comparison we need to consider the cost of building a darkroom from the ground up, capable of the same output. I do think the cost is a red herring -- all approaches are costly, and for those of us using processes that require a negative for contact printing (palladium, carbon, ...) scanning offers practical solutions, especially if one shoots 35mm or medium format. As Sandy has shown, there is objective evidence that the digital workflow does not sacrifice quality, and there is even more headroom (read: less expensive hardware required ) if one makes prints 11x14 or smaller.

 

[jd callow]

I've used the V700 and the M1 and I would not consider either as being good for my purposes. YMMV