Re: scanner density-value relation

Ryuji Suzuki wrote:
> I don't know why I can't find this info on web... but people on this
> list probably know a lot about this topic.
>
> When I scan a step tablet and plot the pixel value in the raw scan
> TIFF file, the relation is good for some range but then it will crap
> out. With my scanner (I still use Epson Expression 1600 with TPU
> because I don't print from scanned negatives) the value craps out way
> below the advertised Dmax value in relation to transmission density.
>
> Did anyone measure this relation and care to share the info here?

Ryuji,
The answer is noise, in particular the signal-to-noise ratio of the sensor.
As the light going through the film striking the CCD sensor decreases, the
signal generated by the sensor also lessens -- the "ratio" between sensor's
signal and scanner system's background operating "noise" is changing,
decreasing, becoming more obvious. When there's lots of light maybe the
ratio is 1000:1 when there's not much light the ratio might only be 5:1. You
should be able to see it "creeping" in on the histogram of a stouffer wedge
as the values darken. A better example would be to think in terms of Bob
Dylan's electric guitar pick-up. When Bob's playing loud all you hear is
music, but then the song quitens down, if the equipment isn't up to spec,
you can start to hear feedback, guitar string noises and ground loop buzz
problems in the amp. He never should have gone electric.

~m
Received on Tue Mar 8 22:12:39 2005