Re: filmscanners: Shadows and Scanwit 2720s

[Arthur Entlich <artistic@ampsc.com>]

Bigboy9955@aol.com wrote:
> 
> In a message dated 08/07/2001 12:40:48 PM Eastern Daylight Time,
> TonySleep@halftone.co.uk writes:
> 
> << No, not damaged. These sorts of horribleness are revealed when you try and
>  use a scanner beyond its capabilities. You are exposing behaviour which
>  would normally be hidden 'below' the black point, and then amplifying the
>  defects by boosting contrast. Basically, if a sensible black point doesn't
>  allow a decent scan you are stuffed.
> 
>  Regards
> 
>  Tony Sleep >>
> 
> Could this be what is buggering Art and his new SDII?  Asking for more than
> the scanner is capable?  I also have a "new" SDII and "seem" to get some nice
> prints from scans.  Of course, I don't know what I am doing yet.  But I would
> think blowing up a scan 2-300% would show the uglies from the scanner.  I
> understand pushing the limits to see how far you can go but....
> 
> Ed

I don't think this is an accurate portrayal of what I am encountering.

First off, the defects I noticed on the first Dimage Scan II were severe
enough to show up on 8 x 10 Epson prints, printed at 300 dpi.  But
beyond that, is a CCD sensor is designated to work at 2820 dpi, or about
2700 sensor elements, per color channel, I don't think it is a lot to
expect that all 2700 work.  If several random ones are either
miscalibrated or simply non-functional then the CCD chip is defective.

Would you accept a digital camera that had a bad sensor element in each
of the images?  In the case of film scanners the problem is magnified in
that each bad sensor element creates a line right across the whole
length of the image.

The second Dimage Scan has fewer bad elements, but it appears to suffer
from a wide area of miscalibrated elements.  I am still trying to
determine if this problem is due to the calibration routine failing
sometimes, which might, in part be a software problem.

Finally, I suspect that the run of Dimage Scan II up here in Canada
might have been a bad run, and it could be that your unit is perfectly
fine.  My first one had severe green shadow noise, while number two does
not, so there are some high inconsistencies in the production line.

The test I suggested for bad elements, by isolating the channels is just
a way to see some defects which otherwise might only show up with
certain subject matter or color combos.  Interestingly, I do not recall
seeing these problems with my HP S-20 which was an older and cheaper
model.  It had other problems, and was lower res, but it did not
manifest bad individual sensor elements.  I suspect if other scanners
manifested this, we'd be hearing more about it, as well.

Anyway, I'm hopeful the problem can be resolved via a properly
manufactured unit which used components which were more thoroughly
tested, or had higher reliability.

Art