al@greenspace.freeserve.co.uk schrieb:
> I've just got an Elite II and have been checking the deep shadows for
> issues, particularly the red channel banding. I have noticed something
> very odd. A straight scan (with or without ICE) shows red channel
> banding identical to what Ralf got along with some "telephone wires".
> However, when GEM is enabled, the banding pretty well disappears.
> And it also looks like some of the the other CCD anomalies like the
> "telephone wires" also improve or disappear.
Hi again,
with my defective Elite II I used GEM once at the default value of 50. I
hadn't noticed the banding then, so I can't tell whether it was present
or not, but at the value of 50 I noticed *huge* "averaging" of the
colors, just a bit like Gaussian blur would cause it. That softening
effect may well have evened out the banding.
> This seems to suggest that the banding (or lack of it with GEM) is not
> caused by any change in scanning speed (or anything else mechanical)
> but by a glitch in the calibration, which doesn't occur with GEM. Or
> alternatively, the algorithms in GEM are just good at cleaning up these
> sort of deep shadow problems......
You may find out by using GEM at a very low value (1 instead of the
default 50). If it is the cleaning effect of GEM, a value of 1 should
leave more of the banding than a value of 50. If it is just a matter of
the calibration data being re-written, the banding should be gone even
at a value of 1 (which I doubt will remove any grain at all).
> I suspect the issue is faulty calibration by the software (maybe made
> worse by poor QC) rather than it being purely hardware. Could well be
> good hardware let down by buggy software or firmware.
I may be mistaken but as far as I know *every* CCD unit would be
"faulty" in the sense of showing off "telephone wires" without prior
calibration. What I am not sure of is, for example, if a single CCD
pixel lost "efficiency" in the sense of electrical response to light
exposition (as a result of aging, dirt, you name it), whether or not the
Minolta Software would be able to correct this properly.
> I haven't reached any conclusions yet whether what I've seen will
> constitute real problems in normal scans. Certainly, using GEM and x4
> multiscanning on one of my (underexposed) slides I use as a test and
> pulling up the shadows, the result was far better than my old Elite.
If GEM, at a low setting that does not affect the overall appearance of
a picture, would be able to completely remedy the banding/"telephone
wire" issue, this might be a way out. I'll hopefully be able to try for
myself next week. If the price for getting rid of the banding is having
the image dramatically softened due to a high GEM setting, there will be
another Elite II on its way back to Minolta, and a Nikon LS-40 will take
its place :-))
> Incidentally, it doesn't look like the current support for the scanner in
> the latest version of Vuescan incorporates calibration routines. Pity
> really, as this would help clarify whether the hardware or the Minolta
> software is the issue.
But it would be interesting if the banding issue exists at all with
Vuescan. If Vuescan has the functionality of getting at least a raw scan
out of the Elite II, I'd be most interested in the results regarding the
problems discussed.
So long -
Ralf
--
My animal photo page on the WWW:
Find my PGP keys (RSA and DSS/DH) on PGP key servers
(use "TrustCenter" certified keys only)
----------------------------------------------------------------------------------------
Unsubscribe by mail to listserver@halftone.co.uk, with 'unsubscribe
filmscanners'
or 'unsubscribe filmscanners_digest' (as appropriate) in the message title or
body
