Re: filmscanners: Was New Nikon performance, now dust

[Arthur Entlich <artistic@ampsc.com>]

Dave King wrote:


> 
> Perhaps not from a design perspective, but from a users perspective it
> seems perfectly reasonable to evaluate scan data in the context of end
> results.  After working on both scans, the Agfa, to my eye, has
> recorded more real image data.  Rafe brought up the idea of noise, and
> perhaps that explains the difference between these scans.  The LS-30
> scan appears sharper initially, but after working on both files I
> would have to say first impressions are misleading, the "sharpness"
> seems to be an artifact.  No matter how I sharpen the LS-30 scan, I
> can't get results that match the sharpened T-2500 scan for image
> detail and clarity, and tonal smoothness and sharpness of grain.
> 

First off, please excuse the many typos and grammatical errors in posts 
of the other day.  I had no internet service for 3 days and then had to 
deal with a barrage of mail, and I was up until 4 AM answering them, as 
well, some personally quite stressful situations came up as well, just 
to keep the pot stirred.  I'm a bit more rested now, and might even be 
able to type a coherent posting or two...

I have a few ideas in regard to the issues of sharpness.  Has Nikon 
added "unsharp masking to their scan?  Before everyone jumps on me 
stating that the Nikon scan is pure and unadulterated and raw, let me 
qualify my statement.  If Nikon found a way to do unsharp masking that 
was not in "firmware" code or their software interface/driver could it 
still me called sharpening?

Isn't in fact true that the LED light source itself does a type of edge 
sharpening?  Could the use of certain cut off, trim, or bias filters in 
the electronics actually "sharpen" an image?

Is it possible the reason the Nikon can't be further sharpened without 
more artifacts be because it is being "optimized" in some other manner 
than through software?

> 
> 
> I don't question the clarity of the dust spots is related to the focus
> of the scanner.  The darkening (exaggeration) of the dust appears to
> be a function of the infrared channel however, as Rob points out.  I
> have no problem with this either, as long as a dust removal algorithm
> takes care of it (it does), and I can use the scanner with all
> Kodachromes and B&W film and get results as good or better as with a
> conventional design (I can't). 

David, would you be kind enough to post the same two images that you did 
previously, but this time using the unsharp masking you feel best 
"glorifies" the Agfa scan.

What I'm wanting to see, is how the dust and dirt responds to that 
processing in software.  Do we end up with very similar dust and dirt in 
both after sharpening?  Also, I want to see the overall tonal ranges you 
refer to.

 I have the feeling that Nikon has
> addressed these problems in the new designs, but I would like to know
> how effectively before deciding on a next scanner purchase.  Both the
> Polaroid 120 and Nikonscan 8000 appear to be excellent with a slight
> edge going to the Nikon perhaps.  But is the Polaroid better for B&W
> and Kodachrome work?

The issue of Kodachrome and B&W with Nikon probably cannot be addressed 
easily.  Unless Nikon has developed a way to lessen the "native 
sharpening" that normally occurs, or ASF has developed a new way to deal 
with the IR scan (or the IR scan methods have been altered), I don't 
know of any way to make B&W or certain Kodachrome dye sets to become IR 
transparent.


Art



> 
> Dave