Scanning B&W film via a Nikon LS-4000 vs. a Minolta Elite 5400

[didjiman]

Conventional wisdom says that a film scanner such as the Nikon

Coolscan does not scan B&W film well because the scanner light source

acts like the light source in a condenser enlarger ("collimated" light

source) and the B&W film's silver grains scatter the light and thus

the scans are more contrasty and the grains "clump" together. There is

a nice diagram illustrating the light source differences (condenser

vs. diffuser) on this web site:

http://www.normankoren.com/scanners.html. One important note is that

this site does not present any photos supporting this claim.

http://www.scanhancer.com/ sells a diffuser plate for the Minolta

Multipro scanner and they make a similar claim that the Multipro with

their Scan Enhancer (or the Minolta Elite 5400 with the Minolta's own

grain dissolver) makes less grainy scans, especially on color scans.

 

I have been using the Nikon LS-4000 scanners for over 3 years,

scanning well over 300 rolls of slides. I process my own slides on a

Jobo rotary processor and bulk scan the whole strip of 36-38 exposures

using the Nikon bulk film adapter. For the past six months, I have

been gradually doing more and more B&W, and playing with different

film types and developers. Again, the films are developed in the Jobo

and then scanned in. While searching for the web for scanning info, I

found the sites mentioned above. Since my local camera shop rents the

Elite 5400 per day basis, it is a cheap way to find out whether the

conventional wisdom is right or not. I use Vuescan to drive both

scanners, setting the film type to B&W, "Generic" vendor and selecting

"White Balance." Personally I do not find Vuescan's user interface

particularly friendly or intuitive, but it does produce good scans.

 

The short answer is that I see no noticeable contrast differences. The

so called "clumping" effect due to the collimated light sources is not

apparent at all. I have tested it on Delta 400, HP5+, Efke 100, and

Tri X, processed (mostly) in Xtol, and a few in Rodinal and D76. Some

of the rolls were pushed but most are used in the box speed. The light

scattering effect does show up in the Nikon scans as being more dusty.

Unfortunately, you cannot use the automatic dust removal system

(usually a technology called ICE) built into these scanners on B&W

film as ICE uses the IR channel to detect dust and scratches. So it is

worthwhile to keep your negatives clean. In fact, whether due to its

higher resolution or other factors, the Minolta scans seem to be more

grainy.

 

(All photos are not processed except with an USM of 120%/1/0 since

scanners are known to soften the scans) The following is the engine

compartment of a working 1911 Pierce Arrow.

 

This is an Elite 5400 scan. The film is not held flat by the scanner

mechanism so notice the left and right edges are out of focus:

http://www.dragonsgate.net/pub/richard/scanner_tests/elite5400.jpg

 

100% crop:

 

http://www.dragonsgate.net/pub/richard/scanner_tests/elite_enlarged.jpg

 

The Nikon 4000 scan.

 

http://www.dragonsgate.net/pub/richard/scanner_tests/Nikon4000.jpg

 

100% crop:

 

http://www.dragonsgate.net/pub/richard/scanner_tests/Nikon_enlarged.jpg

 

Since the Nikon scan is 4000 DPI, the 100% crop is smaller than the

Elite's. Notice also the hair or dust on the Nikon enlargement :-(

 

In summary: while there are real benefits to the diffuser light source

on the Elite 5400 with the grain dissolver option (e.g. less dusty

scans), there is no discernable differences otherwise. If anything,

the Nikon scans seems less grainy and are sharper since the film is

held flatter. The Nikon is also significantly faster. However Minolta

just releases Elite 5400 II so it may have even the score in the speed

front. The Minolta is also cheaper than the LS-4000 or the LS-5000

replacement.

[martin_ranger]

Richard,

 

excellent comparison. I guess I won't have to sell my Coolscan after all.

 

Martin

[erik scanhancer]

Richard,

 

Obviously we see different things. I would prefer the Minolta scan in your comparison,

although I'm not sure the Grain Dissolver was engaged. The Nikon scan shows typically

what would be called grain aliasing, as a result of the "underresolution" of the CCD for the

fine grain of the film. In this case the diffuseness of the lightsource makes only a minor

difference to the final result.

 

Could you explain why there are several people who use a Scanhancer (beta model) in their

Nikon 4000 and 5000 scanners if the differences won't matter that much? See for example

the pics in this thread:

 

http://www.photo.net/bboard/q-and-a-fetch-msg?msg_id=00BIiO

 

Or read what this Nikon 5000 user has to say:

 

http://groups.google.com/group/comp.periphs.scanners/msg/7fa2b98b1f02b4b9?

hl=nl&

 

Fairness demands to say that I am the maker of the Scanhancer, but believe me, I would

never have entered that "business" if I hadn't been a desperate photographer myself who

wanted to share his findings with others. Although to some it seems to appear different,

the selling of Scanhancers is definately not a "commercial" activity. Fortunately I make my

living through my photography and film work.

[didjiman]

Hi Erik, sorry, I don't mean to devalue the Scan Enhancer. My post was more in response to some people (not yours, I believe) statement that Nikon Coolscan is "useless" for B&W scan and that the Minolta Elite 5400 with the grain dissolver is *MUCH* better. For my cup of tea, my tests demonstrate there aren't a whole lot of difference. If I do not have a scanner currently, I may go out and get the Elite5400, or if I have a LS-5000, I may purchase the Beta Scan Enhancer, but for now, I do not see benefits or not continue to use my LS-4000.

 

Thanks.

[larry_kincaid1]

Erik,

As you know, the Minolta 5400II does not have the physical scanhancer type device like the previous 5400. The 5400II added ICE and GEM software. Have you developed a scanhancer that fits and works with the 5400II yet?

[didjiman]

re: Les

Well, I don't know about other people's workflow, but I stick the film in, open Vuescan, and click :-) Very much less than an hour for sure :-)

[jtk]

I think Nikon's issue with B&W is software-related.

 

Erik's comparison with and without Scanhancer look like they were done with Nikonscan. I don't know if Silverfast has the same problem, but Vuescan eliminates the problem.

 

A diffuser might reduce dust, but dust shouldn't be an issue if one is scanning in a reasonable environment and possesses a can of compressed air!

 

I do have a Scanhancer (Erik did you ever get the choral CD I ordered for you?) and have tried to use it with film strips...unfortunately it doesn't work with the SA21 filmstrip adapter because the scanner is designed to position film using optical clues, and Scanhancer evidently confuses it. Wouldn't be a problem with mounted slides.

 

I don't want to mount negatives in slide mounts, but that might allow Scanhancer to be effective.

[erik scanhancer]

Larry, I have no access to a 5400II and thus I am not developing a Scanhancer for that

model. Also, so far I have not seen any direct comparisons between the 5400 and the

5400II on the issue of grain. Maybe the 5400II doesn't really need a Scanhancer? (Given the

amount of requests I have received so far one would guess otherwise, but guessing is not

a very scientific way of jumping to conclusions. So I'll await the hard evidence.)

 

John, the comparison of the Nikon 5000 with and without Scanhancer was made with

Vuescan. The problems you see have nothing to do with the software used. The

appearance of dust and peppergrain is fully related to the optical properties of a scanner.

Of course software might accentuate or mess up the optical results in any way, but the

basic information should still be visible through the treated results.

 

The fact that you seem to have problems scanning with your Nikon 4000 and beta

Scanhancer sounds a bit strange to me as other users report troublefree scanning with

film strips. You did mount the Scanhancer as shown on the picture on my website, I hope?

There should be some open space left on one side in order to let the scanner calibrate

properly. (It's a bit OT, but since you brought it up: no, I never received anything from

you.)

 

 

I would like to make some common comments on the testing of different scanners side by

side. IMO there is no use in making scans for comparisons by means of software that does

anything "automatic". When you want to see how well a scanner picks up the emulsion of

the film you should scan in linear positive mode and not use any sharpening or other

image changer. Otherwise half of the test is going to show how well your scanning

software works with the specific scanner used. (This might also be an interesting test, but

it confuses the objective information about the hardware quality of a scanner.) Now a

linear scan of a B&W negative isn't very informative when it comes to final image quality,

so we should allow ourselves to make fully manual and registered corrections to see the

final quality. In fact, this way it won't matter whether we'll use VueScan, Silverfast or our

OEM drivers. We'll only try to capture the most basic scan which can be done with any of

the known drivers.

 

In order to judge the final result I would like to propose the following workflow:

 

1. Make a 16 bits linear positive scan of the B&W film (I suggest 4x sampling), without

anything else but exposure control in order to get the full histogram within limits. We'll

end up with a very weak negative representation of the final picture;

 

2. Now it is time to take the image to Photoshop and invert it. Assuming that the image

should contain peak whites and deep blacks without clipping, all empty information

should be removed from the file by cutting it off in Levels. Then go to Curves and create a

nice smooth curve that gives you the exact overall contrast that you want to see in your

final picture.

 

3. store the final result without any further changes, like USM or burning.

 

Repeat the above three steps with the other scanner that you want to make a comparison

to. Make very sure that the final overall contrast from step 2 looks exactly the same as in

your first scan. (You might need a different curve to get to that same result, so simply

storing the first one and applying it might not work.) Now both files can be compared at

100% viewing level. In case one scanner has a higher max resolution than the other, the

lower res image should be upsampled to the same size for a fair comparison.

 

I am looking forward to finally see some relevant comparisons on the net!