Arjay
Time Traveller
I've been using a Nikon Coolscan V ED to scan my 35mm films. Generally, the scanner does what it is supposed to do, but it appears that its LED illumination unit produces a very harsh light which seems to accentuate silver halide film grain.
Scanning at the unit's native resolution (4000dpi) seems like the best way to minimize image quality losses, but at the same time, the scanner also reproduces the film's grain structure faithfully and obtrusively.
I feel that the 4000 dpi resolution is finer than the film's actual grain structure, so there might be two ways to solve the grain problem:
If you went the software route, which workflow did you use? Did you employ noise filtering software, and if so, using which settings?
If you used the frosted glass plate way, what were your findings? How did you circumvent scanner error messages for auto exposure measurement and autocalibration? Which scanner settings did you use in either the Nikon Scan or Vuescan scanning programs?
Scanning at the unit's native resolution (4000dpi) seems like the best way to minimize image quality losses, but at the same time, the scanner also reproduces the film's grain structure faithfully and obtrusively.
I feel that the 4000 dpi resolution is finer than the film's actual grain structure, so there might be two ways to solve the grain problem:
- Scan at native scanner resolution (~ 5900 x 3900 pixels for a 35mm negative, equivalent to ~ 23 Megapixels) and then downrez the image via software. I think this might be a viable option, because the original scanning resolution probably is higher than the film's actual (grain-based) resolution - at least for medium to high film sensitivities.
- Use a plate of frosted glass in front of the LED unit to slightly diffuse the light source. This measure has reportedly improved grain reduction on a Minolta DiMAGE ScanMulti PRO scanner that uses cold-cathode lighting (google 'Scanhancer'). It appears that it is possible to modify a Nikon MA-21 slide holder accessory to the Coolscan V ED with such a glass plate.
If you went the software route, which workflow did you use? Did you employ noise filtering software, and if so, using which settings?
If you used the frosted glass plate way, what were your findings? How did you circumvent scanner error messages for auto exposure measurement and autocalibration? Which scanner settings did you use in either the Nikon Scan or Vuescan scanning programs?
Bike Tourist
Well-known
I have the V ED. I have scanned my 50+ year old silver-based negs and accepted the grain and the fact that I can't use digital ICE on them. The grain helps the vintage look of my old German negatives. Of course, this does not help you.
One thing — I think when you make prints from your scan, the grain is not as apparent as it is on the computer screen.
You didn't say what your ultimate use is or why you wished to de-emphasize the grain. I find that in my microstock business I don't normally even try to get scans accepted because the agencies interpret any grain at all as "artifacts". I have had some success once-in-a-while by reducing the full size image to one with fewer pixels and that seems to smooth out things a little.
If you can bring yourself to use a non-silver c-41 process B+W film, such as XP2 or BW400CN, there is much less apparent grain with them.
Good luck!
One thing — I think when you make prints from your scan, the grain is not as apparent as it is on the computer screen.
You didn't say what your ultimate use is or why you wished to de-emphasize the grain. I find that in my microstock business I don't normally even try to get scans accepted because the agencies interpret any grain at all as "artifacts". I have had some success once-in-a-while by reducing the full size image to one with fewer pixels and that seems to smooth out things a little.
If you can bring yourself to use a non-silver c-41 process B+W film, such as XP2 or BW400CN, there is much less apparent grain with them.
Good luck!
Arjay
Time Traveller
You're asking an important question. Coming from a background of digital photography, I find that film grain - while apparently unavoidable - is a problem if I'm not aiming for a vintage look.
I am not producing pictures expressly to sell them as microstock, but I want to develop a workflow that is capable of producing fine art quality images.
Having restored scans of old slides (Ektachromes from the eighties) to publishable quality, I know there are ways to successfully reach that goal - but the question remains at what cost. In the past, this meant a lot of manual Photoshop work with lots of tweaks, masks, blur and sharpening filters, always dedicated to the specific properties of every single picture.. Of course, that's not feasible if I consider processing image files in two or three-digit numbers.
This is why I am looking for hints of how to speed up that process - either by modifying my scanning process or by procedures that can - at least in part - be implemented e.g. in a Photoshop action.
Pickett Wilson
Mentor
You have conflicting interests in your scanning. 
Seriously, scanners often emphasize grain in unattractive ways. The best solution I've seen is to use one of the noise reduction plugins. I use Noiseware Pro, but there are others. Is it sacrilege to use an NR plugin on scanned film? Well, you've already altered the negative by scanning it (scanners are not neutral devices). And if you are trying to produce fine art, then purity isn't your goal anyway.
Seriously, scanners often emphasize grain in unattractive ways. The best solution I've seen is to use one of the noise reduction plugins. I use Noiseware Pro, but there are others. Is it sacrilege to use an NR plugin on scanned film? Well, you've already altered the negative by scanning it (scanners are not neutral devices). And if you are trying to produce fine art, then purity isn't your goal anyway.
Bob Michaels
nobody special
I use the Minolta Multi Pro and that light source does seem to accentuate grain less than the Nikon light source. Very similar to a cold light head on an enlarger.
I have the Scanhancer for both 35mm and MF film. While I always use it on the rare color I do, I found that it seems to make no difference on b&w negs. Erik, the developer says it make a bit of difference but I cannot see it.
I have no information about how Scanhancer would work with the Nikon light source. I do believe the Scanhancer is more that any piece of frosted plastic. I also suspect if it solved the grain problems with the Nikon scanner, that Erik would be marketing it as such. He is accessible by e-mail so I suggest just asking him.
Mackinaw
Think Different
I read this someplace a few years back (maybe on RFF) and it works for me. When using Nikon Scan software, check the "enable post-processing" button and move the slider for "digital ROC" all the way to the left and the slider for "digital GEM" all the way to the right. This will greatly increase processing times but does a good job in minimizing the grain.
Jim B.
thirteen
Established
I scan my 35mm film with a coolscan 5000. I use this article as a guide when I'm concerned about grain
http://www.boeringa.demon.nl/menu_technic_optimalscanningresolution.htm
http://www.boeringa.demon.nl/menu_technic_optimalscanningresolution.htm
jarski
Mentor
hi Arjay, do you have samples of the grain problem from your scans ?
I use same scanner (with Vuescan), even at 2000dpi (due to its speed), and dont consider grain that big problem.
did spot one problem in my process though, that resulted really grainy and ugly files, but the problem was JAlbum, not the scanner.
I use same scanner (with Vuescan), even at 2000dpi (due to its speed), and dont consider grain that big problem.
did spot one problem in my process though, that resulted really grainy and ugly files, but the problem was JAlbum, not the scanner.
jke
Well-known
I do it on the film end. I changed my developing to decrease the grain or render it more suitable to scanning. That means in part looking for B&W films and developer combinations that scan well. The best (IMHO) are Fuji Acros, Fuji Neopan 400-1600 and Ilford 3200, all developed in HC-110. Constant level temperature at all stages is key - 20 degrees throughout, including wash water. Water not chemical stop. Minimal and gentle agitation - all the above get only 30 seconds of inversions in the first minute, and then only two inversions per minute after. The goal is a relatively thin negative. Grain then is the limiting factor, because everything else is adjustable post-scan. And when I scan (I have a Nikon Super Coolscan 9000) I do so with minimal adjustments in the scanner twain. I don't want the scanner software to make any big decisions - just record the raw information. All adjustments are Photoshop which is much better suited for doing those with appropriate subtlety.
gdi
Mentor
This is correct, you can reduce grain significantly by using GEM - though I would not use it at the maximum. I would suggest experimenting with incremental increases (I don't think I ever go over the second level with it.) And I turn ROC off - that is for faded color negs/trans.
Also, I have used a anti-glare glass and I can't see any difference in the scans.
craygc
Well-known
Don't know what you're doing but I don't ever see a "grain" problem. Only difference is Im using a Nikon 4000 ED scanner but basically the same. Everything switched off in Vuescan, never an issue...
not_in_good_order
Well-known
gdi
Mentor
Thanks,
I should have qualified my comments as being for the 9000 and not the V (or 5000). The illumination is different and I think one may need to use different GEM settings.
lawrence
Mentor
gdi
Mentor
It is listed in the specs on the Nikon site - newer LED type ("rod dispersion illumination") and 3-line CCD.
lawrence
Mentor
Arjay
Time Traveller
Here's a sample of Tri_X 400 developed in Prescysol EF (a tanning/staining developer), exposed @ EI 400 with a + 1 1/3 EV compensation for the snow background (scanned on a Cooloscan V using Nikon Scan @ 4000 dpi, shown at 100% magnification, no noise or grain reeduction):
Maybe I'm expecting too much, but I have no reference since I only processed digitally captured images in the past ...
Stuart John
Well-known
How do your prints look. If they are too grainy then you need to look into a way to reduce grain. You cold start my modifying your processing to produce finer grained negatives. I Always got nice 8x10 traditional darkroom prints from 35mm Trix developed in HC110 that did not show too much grain but I don't believe I would really go any bigger with Trix if I wanted fine grained prints. Trix is a traditional B&W film that is not exactly fine grained if you compare it to say Tmax100. Remember also that 100% crop from scans are rather like looking at huge over 40 inchs wide 100ppi prints. How big would you usually print your pictures it may be worth make a few test prints.
delft
Established
I too find my coolscan V sometimes too grainy for B+W. Now that I have access to a darkroom, I don't scan much 35mm B+W anymore. However, I've found the Neatimage software very effective for removing grain:
You can tweak the parameters (I think this was too agressive) and a jpg-only trial version is FREE!
Dirk
You can tweak the parameters (I think this was too agressive) and a jpg-only trial version is FREE!
Dirk
Steve M.
Mentor
Unfortunately, as you've seen, scanners will accentuate grain from silver films. jke has the most elegant solution, as it's much more effective to address the problem before hand rather than after the fact. But if you're like me, you have a lot of negs already developed.
I used your scanner for many years and what I finally did was try to fix things as best I could in PS w/ a little Gaussian blur, then slowly sharpen the image back up in several stages. Any noise ware software I used was a little too aggressive for my tastes and I lost too much detail. Tri-X is a tough one as it has a bit of grain in 35mm. It's beautiful grain, but gets noticeable at large print sizes. You'll just have to accept that as it's part of the film's nature, and beauty I might add. Try shooting it in 120. Totally different. Smoooooth, and the same great tonal structure.
My solution to the scanning issue was to get rid of the film scanner. For color negs they are fine, but I seldom shoot it. There's no competing w/ a good wet print from a darkroom in B&W. It's what the film was designed for, and it has a beauty and tonal range that cannot be duplicated any other way. Not to mention there are many special looks available using platinum, salt or many other fine print techniques.
I used your scanner for many years and what I finally did was try to fix things as best I could in PS w/ a little Gaussian blur, then slowly sharpen the image back up in several stages. Any noise ware software I used was a little too aggressive for my tastes and I lost too much detail. Tri-X is a tough one as it has a bit of grain in 35mm. It's beautiful grain, but gets noticeable at large print sizes. You'll just have to accept that as it's part of the film's nature, and beauty I might add. Try shooting it in 120. Totally different. Smoooooth, and the same great tonal structure.
My solution to the scanning issue was to get rid of the film scanner. For color negs they are fine, but I seldom shoot it. There's no competing w/ a good wet print from a darkroom in B&W. It's what the film was designed for, and it has a beauty and tonal range that cannot be duplicated any other way. Not to mention there are many special looks available using platinum, salt or many other fine print techniques.