[filmscanners] Re: Digital ICE

["Bob Frost" <bob@frost.name>]

Austin & Jack,

I was interested in the possible reasons for scans being apparently softer
than originals, but must confess that I don't understand 'light piping' at
all, and vaguely see how 'electronic crosstalk' might be involved.

I, naively perhaps, have often thought that simple things like 'sensor
motion blur' and 'scanner shake' might have something to do with it. When
looking and listening to my LS4000 scanning, I start to wonder how long the
sensors actually stays still, and how this relates to the length of time of
the exposure of the sensors. In addition, when I put my hand on the scanner
during scanning, the whole thing is vibrating, which surely can't be doing
anything to improve sharpness. Indeed, I have often thought that one good
reason for doing 16x sampling might be that the scanner is actually
stationary and vibration-free for perhaps 14 of the 16 samples.

I won't mind if you tell me that this is all rubbish, but some detail of
exposure times and sensor movement might be useful to know.

Bob Frost



----- Original Message -----
From: "Austin Franklin" <darkroom@ix.netcom.com>


Hi Jack,

> Okay Austin, I've got the Doctor looking over my shoulder right
> now. He says
> that with MOST scanners (especially ccd) because of the pixels sensor
> density, there is light piping between sensors causing a loss of
> sharpness.

I don't believe it's light piping, but electronic crosstalk between adjacent
sensors, which is very dependant on color (red is worst, blue second worst
and green the best).  If it was lightpiping, I would think it would be the
same across all three colors.  Please run that buy the Dr.

> He is
> telling me
> that a drum scanner is an exception because it masks the light going into
> and out of the film.

Now that is interesting.  I understand the principle, but my experience, is
that drum scanners actually require more sharpening than high end CCDs...at
least that's what I've seen in scans I've done on both a Howtek 4500 and a
Leaf 45.  Though I agree, what you say makes sense theoretically.

I do agree that the color smearing isn't an issue with PMTs, as it is with
CCDs, simply because PMTs only "read" one pixel at a time, and therefore
there is no possibility of this type of crosstalk.

> Does this help?

I wanted to see if you had some other answer than I had.  I don't know that
I believe the lightpipe issue...we've always attributed it to what I stated.
If you (or anyone for that matter) or the Dr. have actually done
experimentation that shows it is lightpiping...I would like to know.  Since
I have not done any actual sensor design, I just use them, I have not done
any of this, but it would be interesting to know.

----------------------------------------------------------------------------------------
Unsubscribe by mail to listserver@halftone.co.uk, with 'unsubscribe 
filmscanners'
or 'unsubscribe filmscanners_digest' (as appropriate) in the message title or 
body