RE: filmscanners: Link to Nikon 8000 banding example...

[Raphael Bustin <rafeb@channel1.com>]

On Wed, 18 Jul 2001, Lawrence Smith wrote:

> 
> > Seems pretty slight to me -
> 
> Perhaps but it shows up in prints.  As my prints sell for hundreds of $ they
> need to be perfect.
> >
> > What magnification are we looking at in the zoom?
> 
> About 66%
> 
> 
> > Somebody was talking
> > about making the 8000 slow down to use only one row of CCD sensors at a
> > time, instead of three (have I got the right scanner?) so it
> > would seem that
> > if you could turn on that option, you might get a preferable result.
> 
> It does help but it also takes about an 1.5 hours to do a 16x, 14bit scan of
> a 645 neg.  Clearly not an option.
> 
> The bottom line is that Nikon needs to decide that it has real issue on it's
> hands.  One second level tech essentially admitted that they did during one
> phone call but during a different conversation with the same individual he
> was adamant that the problem was not widespread and was limited to my first
> scanner.  Clearly that is not the case.  Anyway, for #k, they need to figure


Lawrence, why 16x ? Are you convinced that the 
results are that much better?

As for 14-bit vs. 8-bit color, you've already 
heard my rap on that.

In my case, using "Super Fine Scan" is only 
a 3 or 4 minute penalty.

Banding victims:  I assume banding is in 
the direction parallel to the CCD ?? (ie., 
perpendicular to the motion of the film 
holder?)

I have another theory about this that I'll 
try to check out tonight.  Surely you've 
noticed the "teeth" on the edges of the 
film holders, right?  I presume they're 
meant to engage with a gear of some sort 
inside the scanner, in order to move the 
film holder.  (Hopefully just for coarse 
pre-scan moves, and not for actual scanning.)

So I'm wondering.. is the pitch of the 
bands identical to the pitch of these 
teeth?

In all cases that I've seen, the bands 
are remarkably wide, and very uniform 
in width.

If this theory holds, it seems to me 
to indicate a very fundamental design 
issue.  These "teeth" have a very large 
pitch considering it's a 4000 dpi machine.


rafe b.