Nikon scanners, Vista 64 & Vuescan - Thank you Ed!

[scott_turner]

I just wanted to offer up a big THANK YOU! to Ed Hamrick and Vuescan for doing something which Nikon apparently

finds too difficult to do.

 

I just finished putting together a new 64 bit system for CS4 work. I'd originally planned to keep my old system

as a dedicated scanning system, as I knew that Nikon (and consequently Silverfast) didn't support 64-bit Vista.

I wasn't happy about it, as it would break my workflow, but I wanted the memory of a 64 bit system. Then I found

a posting here (a big thanks to Andre Noble as well!) which pointed out that if you install the latest Vuescan

and have it install drivers first - then install NikonScan but DON'T install Nikon's drivers - that NikonScan

will work.

 

And it does. Works just fine in fact, and more importantly to me, so does Silverfast AI Studio - something which

had not been addressed in the previous posting. Haven't found a single glitch yet.

 

So again - thank you Ed Hamrick! When digging up all of my old licenses for installing on the new machine, I

noted that I've been a Vuescan customer since April of 2001. One very affordable purchase has supported a

multitude of scanners with truckloads full of updates for me. Over the years, Vuescan has rescued me on numerous

occasions when other apps just couldn't do the job. Oh, I've cussed it at times, as it has its own way of doing

things, and some of those updates have caused problems. But it still remains in my toolbox as a valued piece of

software. And its now once again at the heart of every scan I make, even though I'm not usually using it directly.

 

The internet is full of gripes - its something we're all good at when things don't go the way we'd like - but too

often we forget to praise as well. This is a product worthy of some praise.

 

Scott

[mendel_leisk]

I've yet to get my head around NikonScan on *any* system: To date Vuescan is *indespensible* to me for running a Coolscan V.

[scott_turner]

Mendel,

 

You'll get no argument from me on that! NikonScan's "color management" is broken and incomplete. But while I've

been pretty much dedicated to Vuescan for color negatives, I've been using Silverfast AI Studio for transparency

scans, and Silverfast requires Nikon's drivers to be present in order to work. Thankfully, I discovered that the

current Vuescan drivers fill in that gap on Vista 64.

 

Lately, I've been revisiting Silverfast's NegaFix for scanning color negatives, and I'm slowly getting used to

it, if not actually learning to love it. I'm still using Vuescan and Erik Krause's workflow, however, if the

going gets tough with NegaFix on some images.

 

Scott

[joe_oliva]

...according to a well known Nikon USA Tech Rep representative, with whom I discussed the current state of NikonScan software earlier in the year. While he admitted NikonScan was due for an update, he was not optimistic that it would happen. I now use Ed Hamrick's excellent Vuescan as my default software for my Nikon 9000. Personally, I don't expect it to be long before Nikon discontinues the 9000, and I am looking to purchase a second unit as I have an extensive archive of 120 negatives.

[stwrtertbsratbs5]

I much prefer NikonScan to Vuescan. DIgital ICE 4 is far superior to the 'Dust off' feature in Vuescan. And, with a bit of practice, NikonScan can produces excellent results.

[scott_turner]

Robert,

 

It's not a matter of practice - I've had plenty of that over the years. :-)

 

Yes, NikonScan can produce reasonable results and it's a very nice app in many ways - I understand, for example,

why some would prefer it over Vuescan's somewhat quirky user interface. But as referenced above, it's color

management is both broken and incomplete, so a true calibrated workflow is simply not possible, and results vary

as output profiles are changed. And for me that's a dead-solid showstopper.

 

Vuescan does not have a "dust off" feature as you call it - it uses a well-implemented IR channel based cleaning

routine, just like ICE. While I agree that the ICE implementation is generally better, Vuescan's implementation

is still very good, and Vuescan can do something (actually several somethings...) that NikonScan can't, and

that's support calibration. I also use Silverfast which supports both true ICE 4, AND IT8 calibrated scanning.

 

But all of this is very much beside the point of my posting. The bottom line is that neither Silverfast nor

NikonScan would even be possible for me on a 64 bit OS were it not for Vuescan. And for that, I stand and applaud.

 

Scott

[stwrtertbsratbs5]

I don't make color corrections in NikonScan. So I suppose I've adapted my work flow to the tools strengths and weaknesses. I create linear scans that are not color managed and I apply a scanner profile in Photoshop. I wish it were better in some areas, but I've found that I can use it effectively.

[mendel_leisk]

Scott, I've been going along swimmingly, using:

 

Coolscan V / Vuescan / Erik Kraus Super Advanced Workflow

 

until, I hit one roll that doesn't seem to work, delivering Vuescan Raws that are really underexposed, no pixels above maybe 100. I think it's due to the roll being overexposed. My guess is that setting and locking exposure based on the leader (Erik's workflow) works as long as you have decently exposed film. Actually it might *still* be the best bet. I've just left it for a while. I've found when confronted with seemingly insurmountable problems, just forgetting about them for a while can help ;)

 

Robert, I agree: Vuescan's cleaning is second fiddle to ICE. That said, everything needs to be weighed. With my current scanning, the color negatvie film I'm scanning is fairly clean, so cleaning isnot such a factor. If Ed Hamrick could improve his cleaning it would really pull up the weak link in his product. But I kind of doubt that's going to happen, it's been as is for several years now.

 

I've also found the Coolscans, at least the V I'm using, is merciless at revealing film scratches, comparted to say the Scan Elite 5400 with Grain Dissolver active. I've noticed this in particular on silver emulsion film scans, obviously done without ICE or Vuescan's cleaning. The highly directional light source on the Nikons makes for a quantum increase in cleaning chores.

 

Your workflow of outputting a "linear" scan through Nikonscan sounds similar to my Scan Elite 5400 color slide workflow. There I *needed* ICE, Vuescan cleaning was just not good enough. I would output 16 bit linears through Minolta Scan Utility, with ICE and Grain Dissolver, then use them with Vuescan scan-from-disk, with a special ICC profile.

[scott_turner]

Mendel,

 

I remain convinced that Erik's workflow is absolutely the best path to getting to the colors that were actually

captured on the film - and as a consequence for maintaining each film type's individual character. By taking

color corrections for the black and white points out of the equation (which is essentially what this workflow

allows you to do) you're left with B and W luminosity to set on your own, and, of course, neutral midtones. My

guess is that your lack of success with this one roll is, as you surmise an issue with that roll. That is, the

colors and tonal scale aren't right because, well, they're not right, leaving you trying to fix what was really

recorded.

 

If I have time over the next few weeks, I have an example image that I'm going to try to do a writeup on for

P.net - comparing what NikonScan does with its neutral (hah!) settings, with what can be accomplished with

Vuescan and Erik's workflow. Basically, this is a classic image which leaves you in an almost "can't get there

from here" situation relative to color corrections due to the extremely non-linear color errors induced by NS's

auto-routines. This same image was marginally better, BTW, using Silverfast's NegaFix, and the correction tools

are certainly better there, but the same non-linear color casts were evident, with the same problems involved in

removing them.

 

As for making linear scans and dealing with things later - the problem in doing so with a color negative remains

that you're dealing with too many unknowns and, short of a having a color card shot to refer to, nothing to set

as a standard. If the image is well behaved, i.e. the color channels of the actual image minus the mask color

reach limits together as all white and all black (that is all channels go from 0 to 255 when the mask is removed

from the equation), life is fairly simple - but then, these are the images that NikonScan will handle fine as

well. But the farther you deviate from that ideal, the more difficult an image is to deal with using other

standard techniques. These are the images where this workflow shines.

 

This test image also demonstrates nicely the differences between ICE and Vuescan's IR cleaning. Both do fairly

well with dust, but neither shine with scratches, with Vuescan lagging well behind in this area.

 

I still want "it all", and I still haven't found the holy grail for color negs, but at least with Vuescan I have

a reasonable answer for difficult images. I'll gladly deal with a bit more scratch cleanup knowing I can really

get the colors right.

 

Scott

[peter_gofton]

<p>Scott,<br>

I've been trying to run Nikon Scan on a Dell Studio running 64-bit Vista, and I installed VueScan as you suggest. I uninstalled NikonScan and reinstalled it without selecting either of the driver check-boxes. However, although VueScan works, NikonScan says it cannot find any active devices.<br>

Vista Device Manager says that I have a FireWire Scanner, using a driver from Hamrick Software, which looks right.<br>

Do you have any suggestions as to what to try?<br>

Peter</p>