timothy_nelson

I have them both also. The 50 Years book is fascinating, with photos of prototypes along

the development of the M features, brief profiles of some of the key individuals in that

development, and loads of insight into the conservative yet innovative philosophy that

kept the M so stable in concept and form over so many years. It gave me more

appreciation for the engineering cleverness that was involved in cramming the electronics

for the M6 and M7 into the fixed-form M body. The book is beautifully produced.

When I run Smart Sharpen on my iMac G4 (1 Gb RAM), I can view the preview of the

sharpening effect at any magnification, but when I run the actual sharpening (with progress

bar, etc.), I can only view the effect on screen at 100% or 66%. At lower magnification, the

view reverts to the unsharpened version. The file is indeed sharpened as seen in the 100%

view. It doesn't seem to be a memory limitation, since the same thing happens with files of

various sizes. I haven't seen any documentation of this issue--has anyone else experienced

this or does anyone know how to fix it?

To B Macy,

You didn't mention which film profile you used in Vuescan. This will influence the

distribution of tones across the histogram. From my experience, the difference in scan

contrast between NS and Vuescan is likely because NS clipped shadows and highlights to

achieve a pleasing contrast, while Vuescan fits the full tonal range into the file. This

appears flat in the raw scan of many negatives, but it means you got all the info from the

negative. You can then decide in Photoshop whether you want to adjust contrast by

clipping the ends or using curves or other tools. If you do a web search for negative scans

and NikonScan, you'll get quite a few hits in which people have tried all possible

adjustments of NS, and conclude that it simply clips most negatives, even when you scan

as RGB positive and invert. The LS-4000 really benefits from Vuescan--give it another

chance.

<Others may help you more in your quest, but I think using the LS-4000 for traditional

black and white films is definitely a ticket for disasterously poor results.>

I think the problem with the LS-4000 for B&W silver negs is mostly with the NikonScan

software. With Vuescan, I think it produces scans that are comparable to the tonal range

and grain structure I get with conventional enlarging with my Focomat enlarger--in fact,

probably better. The grain is in the negative, and it's accurately captured by the scanner.

Other light sources may suppress it more, just as with various enlarger light sources.

Disasterously poor results? I disagree. It does require dust spotting, but it can produce

beautiful B&W scans.

I've used the 4000 ED scanner for the last several years to scan my Tri-X and Delta 100

B&W negatives. Here are my recommendations:

1. Use Vuescan to operate the scanner. After much testing and adjustment, I found that all

of the available film profiles and methods within NikonScan clipped the highlights and

shadows of both B&W and color negatives, although NS works fine for slides. Vuescan

always produces full scale, unclipped scans.

2. Scan full resolution (4000 ppi), grayscale, 16-bit depth. For Tri-X I use the Vuescan

TMax professional 400 profile. For Delta 100, I use the TMax 100 profile. However, there

are many other Vuescan profiles you can test to see what works best for you.

3. Don't bother with multiscanning unless you see significant shadow noise in the scans. I

find it much preferable to use noise reduction software such as Noise Ninja to tame

shadow noise, and even to smooth grain. The latest versions of these NR softwares

operate in a "smart" fashion that enables you to dial in settings that eliminate the noise,

but protect edge information, so sharpening still works and you don't get the plastic,

blurred look that was the problem with earlier versions.

4. Make sure to check your scan histogram, rather than just the look of the scan. You want

a file that captures the entire range of the negative, even if the raw scan looks quite flat.

Do your tonal adjustments in Photoshop, not in the scanning software.

5. Scanning your B&W negatives (and subsequent removal of dust spots) may give you the

push to try shooting color negative film instead, since the scanner IR channel plus

software gives you automatic dust spotting. I switched to color neg about a year after I

started scanning, although I have a 30-year archive of B&W negs. I've found that scanned

color neg has very close to the tonal range of B&W, despite the "common knowledge"

reiterated by web savants that it has less range. Try it. Converting to B&W is trivial in

Photoshop, with many options for doing this.

If you are printing from scans, you can also consider color negative as source material, not

just C41 B&W negatives. Many ways to convert to B&W, including the option of virtual

filtering. I have 30 years-worth of TriX negs that I also scan, but it's a pleasure by

comparison to scan the C41 material, because the IR scratch and dust elimination works

so well. Not so with the conventional silver negs. Fujicolor NPH looks a lot like TriX when

converted, at least to me.

I just use Reala 100 (Japan) profile to get partway there, then complete the color

adjustment in Photoshop. I think you'll need some Photoshop work, regardless of the

profile chosen, including the Generic profile. For me, the Photoshop Levels>Options dialog

box lets me correct most images almost instantly, with Snap Neutral Midtones, plus either

Find Dark & Light Colors or Enhance Per Channel Contrast. Most just need a little tweaking

after that, if anything.

If you scan B&W negatives, the pen & tablet are a must for dust spotting. It is a wonderful

improvement over spotting with the mouse.

I don't understand why the dynamic range of the scanner should be such an issue with

color negative film, which is a low contrast material. Yes, color negative can capture many

stops of scene range, but it compresses that into a relatively small density range on film.

Maybe a bigger issue with B&W negatives, but as far as I am aware, the need for great

dynamic range in film scanners is more for slides than for negatives. The challenge for

color negative seems more for post-scanning software to expand the relatively small

density range of the negative material back into a "full" pictorial range for viewing, and to

do it accurately.

Andrew,

I understand the advantage of doing selective white balances for regions of the image, as

you show. However, one can do this in Vuescan by saving a single raw scan and applying

different white balances after the fact. I understand you to say also that two scanner

exposures will benefit highlights and shadows. Are you saying that this improves the reach

of the scanner into those areas (greater range), or that the quality of those areas is

improved (same range as single scan)? I'm still having difficulty understanding how this

would improve a scan if the histogram does appear to capture the entire image, without

clipping. I often did multiple exposure scans with my Nikon scanner when I was struggling

with NikonScan, which always clipped my B&W and color negs, but Vuescan eliminated this

problem for me. For me the problem wasn't that the scanner range was unable to capture

the density range of the negative, it was inadequate software, with poor film profiles. I'm

not trying to criticize what obviously works very well for you, I just want to understand it.

Andrew,

Please explain what is gained in a high/low scan blend if one can do a single scan that

captures the negative's film base to Dmax range without clipping? If the histogram

touches down at both ends, isn't this all there is?

Perhaps someone with knowledge of sensitometry could chime in? I think we are confusing

the dynamic range of the scene with the density range of the negative. Once the exposure

is made and the film is developed, I believe the scene has been translated into the density

range of the negative or slide. How that density is interpreted after the scan comes from

the film profile your scanner software applies to the raw scan, but I think most film

scanners today do capture the entire range of pictorially useful film density. The high end

scanners may squeeze a little bit more out of the ends, but it's the film that is the real

limitation, not your scanner. The aim of the HDR multiple exposure approach is to

compress a greater scene range into the more limited printer or screen range.

I agree with the earlier poster that film scanners plus software (e.g., Vuescan) already

capture however many stops have been compressed from the scene into the negative

film's dynamic range. Look at the histogram of the scan. If the histogram includes all of

the density on the negative, I don't see how multiple scans will pull out more. The film has

already done the compression of the additional stops that you would be asking HDR to do

for you, unless you are combining multiple film exposures, instead of scanner exposures.

If you ask the scanner to penetrate blocked up highlight areas beyond the effective range

of the film exposure. by increasing the scanner exposure setting, you will likely get

something pretty ugly.

If you like taking low-light photos during your travel, be cautious about getting the

smallest lightest teles. I found that the thin TE was just too small and light to hold steady

at slow shutter speeds. I prefer the current Elmarit 90, not only for its optical performance,

but because it has extra mass that makes it easier for me to hold steady. In practice the

difference of a few hundred grams doesn't matter much when the lens is in your bag. The

physical dimensions are not much different between the TE and the Elmarit, especially

when you count the deep metal shade that is really required for the TE. The extra grams of

the Elmarit are worth it when it's in my hands.

A couple more observations & suggestions:

NikonScan tends to clip both B&W and color negatives, regardless of exposure settings or

profile chosen. Vuescan is much better for negatives.

Test all of the available Vuescan profiles for your B&W negs. The preview is fast enough

that it's easy to find the best for a particular negative/lighting condition, using the same

raw scan. The most useful to me have been TMax Professional 100 and 400 profiles, but I

use others depending on the particular negative.

If you want some grain smoothing, try one of the dedicated noise reduction programs

such as Noise Ninja or Neat Image. You can get the look of a slower film or larger format

using one of these "smart" noise reduction treatments that preserve "sharpness."

Try using a beanbag. There will almost always be a stable object (chair, wall, etc.) you

can hold the camera & beanbag against. Do you really want to be fiddling with a

tripod when you're on the move? You can make or buy zippered beanbags that you

can fill on site with a suitable material from the local market. I prefer dry linseed or

lentils. You can dump the filler whenever you want to lighten your load.

Or if you use a Mac, there are nice PanoTools Photoshop plug-ins and support files at

www.kekus.com, including lens databases, tutorials, etc.

I've used all 4 of the 90 Elmarits and TE's over the years, all clean glass, no cement

problems, used with hoods, etc. Here's my personal experience:

1st Elmarit--beautiful crisp images, no flare if used with hood, just too big and heavy

fat TE--too soft, always disappointed me when used in kit along with 50/2 and 35/2 (1st

or 2nd generation at that time)

thin TE--can make nice images, but never could control/predict flare, even with the metal

hood. Too light and small for me to hold steady at slow speeds.

current Elmarit---ahhh! Just right. Produces images that have the look and sharpness of

those from my other Leica lenses (from the generation just before ASPH).

Interesting comparison, but I believe it would be a better test if you turned off

sharpening for both programs. If they use different sharpening algorithms, each may

have a different effect on grain appearance.

The deciding issue for me on Vuescan vs Nikon Scan, aside from the greater control

offered by Vuescan, was that NS consistently produced files with clipped shadows

from both color and BW negatives, regardless of any settings made. I have the

impression that NS was designed for slides only, with functionality for negatives as an

afterthought.

You don't need anything but Photoshop. I do this using File>Automate>Contact Sheet

II and an initial page size of 10.5 x 13.5 inches. This creates a sheet with too much

spacing between thumbnails, for which I recorded a Photoshop Action to rearrange to

an 8 x 10 sheet that looks exactly like an analog contact sheet.

I scan Widelux negatives (24 x 58) on my Coolscan 4000, using the normal SA-21

strip film holder, and using Vuescan to run the scanner. I shoot color negative film, so

I first prescan a frame that includes some clear (orange) film base (black image such

as film leader), then lock exposure, prescan again and lock film base. You can also

lock the white balance, although I usually don't find it necessary unless one side of

the image is very different from the other. Now adjust frame offset to frame the first

scan to left edge of the panorama, save that scan, then use frame offset another

20mm or so to shift the frame to right edge (with generous overlap to first exposure).

Once you find the right amount of frame offset by trial and error for your camera, it

will always be the same number. You can blend the two halves in Photoshop with a

variety of methods, or use any of many available panorama stitching programs (I use

PTMac from www.kirkus.com). You can also use XBlend to blend the overlap

automatically. It's all easier than it sounds. The essential part is locking the exposure,

film base, and white balance of the two half-scans so they match. I found Nikonscan

to be worthless for scanning negatives, either black and white or color. For a digital

contact sheet, you can scan an entire strip of panoramic negatives using Vuescan, by

entering half the frame to frame distance and making stepwise half-frame scans

across the strip without overlaps, although contacts are easier to make with a flat bed

scanner.

Roger and Bill,

Thanks for your comments. I do routinely prescan and lock in the base color, according to

Ed Hamrick's advanced workflow suggestions. I also do a manual white balance as Roger

describes. However, I have the same problem that Bill describes, which is that none of the

profiles available produce a good color rendition for some scenes. I believe that even using

Hamrick's IT8 target profiling method is only an approximation, because as I understand

it, it makes a linear color interpretation, while the more sophisticated and accurate profiles

are complex curves. So far, my best starting points have been with the Fuji Super G

profiles rather than Generic or Reala, but they still need adjustments in Photoshop.

It is surprising that profiles for the newer Fuji films have not been available. There are

dozens of threads on this at the comp.periph.scanners usegroup. I like NPH enough that I

bite the bullet and make the post-scan adjustments in Photoshop, but it's so much easier

with Reala and Superia 400, which usually require no more than a slight boost in color

saturation, coming straight out of Vuescan.

Has anyone shot an IT8 target and prepared a profile for Fujicolor NPH for Vuescan,

as described in Ed Hamrick's User Guide? The built-in Vuescan profile for Reala 100

(Japan) works fine for Reala and for Press 400 (Fujicolor 400), but I switched a while

ago to NPH, and none of the built-in profiles work too well. If anyone has made an

NPH profile (for Nikon LS4000 scanner), would you be willing to share it? Thanks!