Austin wrote:
>Rob Wrote:
>> Unfortunately this isn't what it actually does. The "analog
>> gain" actually changes the integration time ie. how long the
>> CCD is exposed for. The longer the exposure, the greater the
>> "analog gain".
> Where did you get the idea (as in a source of information) that
> exposure time is "analog gain"?
Ed Hamrick did testing which demonstrated that the "analog gain"
feature adjusted the integration time. I believe he traced the
SCSI command calls used by Nikonscan to control the scanner.
Adjusting the exposure level in Vuescan effectively does the
same thing as the "analog gain". Ed hasn't put a scale on a
slider comparing the numbers to EV.
> Gain is something entirely different than exposure time.
I'm aware of that. Nikon introduced the confusion by calling
the exposure compensation in Nikonscan "analog gain". It could
be thought of as "gain", but not in the correct meaning of the
term to an engineer.
> Time is, well, time. Gain is amplification of a signal...
> though that does increase over time, I don't believe that
> qualifies as "analog gain".
You'd need to argue that one with Nikon.
> In every digital imaging device I've ever designed or
> worked on, gain has specifically been the control of
> the output amplifier between the CCD and the A/D, and
> is typically determined by calibration. It is used
> to offset the variance in illumination that can occur
> over time.
Agreed. You're thinking from the perspective of the engineer
not the sales department. Personally I think they should
have called it "exposure compensation" then they wouldn't
have used "analog gain" incorrectly. My understanding is
that the analog circuitry in the Nikons is set up in the
factory at calibration time. Making a scan brighter or
darker is done by varying the exposure time.
I haven't heard of any "consumer" CCD film scanner which
does actually alter the analogue signal. As I think you
mentioned, such a thing could be used to increase the
dynamic range of the system, but the only scanners I've
heard mentioned actually doing this were certain drum
scanners.
Rob
Rob Geraghty harper@wordweb.com
