[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Optimizing Negatives for Scanning



Ender100@aol.com wrote:
> ... I looked at the scan comparisons you showed online... the scanner noise that
> shows in your scans seems awfully regular for random noise... also it almost
> appears that some of the ccd's in the scanner are not functioning at the same
> level as others, giving you lines in the scan.

Noise does not have to be random to be destructive to a signal.

The thing about the ccd's is interesting, but aren't their information
combined to give 36-bits (in this scanner) of information for each pixel
at a physical location.  In other words, even though the ccd's are lined
up, each set of three is (for practical purposes) looking and recording
a single physical location.  Additionally, this would be expected to
only account for lines in the direction of the movement of the ccd's
(lines occur in both directions).

> I'm confused by what you said about bit depth and noise... I think many
> scanner and software combinations will do 16 bit scans... however, the better
> scanners will have reduced noise...
> ... I would rather
> start with a good 8 bit scan from a good scanner than a bad 16 bit from a low
> quality scanner...

The HP 6300C is supposedly one of the better desktop scanners (at least
HP thinks so).  It has a pixel depth of 12-bits (36-bits total for
color).  I do not recall of any desktop scanners having more than 36-bit
(bought mine in October 1999).

If you have an 8-bit scanner, scan a Stouffer 21-step and compare with
mine.  The uncalibrated Stouffers should be close enough (in the same
ball park), and if not perhaps we could mail one around.  I would also
be interested if someone has a 16-bit scanner to scan the Stouffer
21-step as well.  Comparison of high quality 8-bit, 12-bit, and 16-bit
scans might clear some of the confusion.

-- 
Jeffrey D. Mathias
http://home.att.net/~jeffrey.d.mathias/