[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Google
  Web www.spinics.net

Re: Sprintscan 120 issues/findings (comments welcomed)



On 15 Apr 2002 at 8:38, Kennedy McEwen wrote:

> In article <3CB9F8E2.23275.1E72AEA@localhost>, rafe.bustin@verizon.net 
> writes
> >
> >This specific behavior has also been noted and griped
> >about on the LS-8000 -- ie., histograms that fall off
> >abruptly, on the low end, at a value well above zero.
> >
> >On my Nikon, using NikonScan, this cutoff appears
> >at a count of around 20-30 (out of 255.)  The "fix"
> >for this problem (in NikonScan) is screwy, and probably
> >not applicable or transferable to the SS 120, is my guess.
> >
> >Frankly, this (mis)behavior doesn't concern me nearly as
> >much as the opposite problem (clipping.)
> >
> But this would appear to be clipping just the same Rafe - the analogue 
> signal is being clipped before the optical or digital range limits have 
> been reached.  I would be less concerned if the fall off wasn't abrupt.


The bins have to fall off somewhere, don't they,
Kennedy?

Having the CCD's analog values fall "comfortably" 
inside the A/D limiting values is no huge sin -- 
not compared to the opposite error, which really is 
clipping.  The analog values have to be manipulated 
into some range, no?  Better that the range is 
narrower than the A/D's range, than wider.

I mis-spoke a bit using the word "abrupt."  What 
I meant is that there are no pixels -- none at all -- 
represented in the histogram, below, say, 15 o 20 
or so.  (The exact value varies from image to image.)
But I didn't mean to imply a steep wall in the 
histogram at that point.

I feel good about a scan when I see the population 
fall off "nicely" to zero at both ends of the range.  
And in general, I see that on the Nikon.  It's just 
that the bottom end of the range isn't zero; it's 
more like 15 or 20.

Mathematically, I'd describe it this way:  the 
function that describes the envelope of the 
histogram has a derivative that approaches 
zero (more or less), as the function itself 
approaches zero.

It's ridiculously easy to fix;  just nudge up the 
black point slider in NikonScan's histogram 
(Levels) tool.  If that darkens the image too 
much, then nudge up the gamma a bit by 
pulling up on the curve.

Not sure what all the fuss is about.  Almost regret 
bringing it up.


rafe b.

-
Turn off HTML mail features. Keep quoted material short. Use accurate
subject lines. http://www.leben.com/lists for list instructions.

[Books]     [Home]     [Photos]     [Yosemite]     [Scanners]     [Steve's Art]     [The Gimp]     [100% Free Online Dating]     [PhotoForum]     [Epson Inkjet]

Powered by Linux