[filmscanners] RE: Terms used in Dynamic Range calculations...

["Austin Franklin" <darkroom@ix.netcom.com>]

Hi Paul,

> Okay, I missed the "largest = maximum signal level - minimum signal level"
> thing. So it sounds like you're using the word "signal" to refer to signal
> plus noise, in which case the minimal "signal" is pure noise. In
> this case,
> I suppose one could arguably say that the ability to handle a
> maximum of 1mW
> of information while adding 1mW of noise constitutes a 0db
> dynamic range.

Agreed...and the equation bears that out:

DR (dB) = 10log10((1-0)/1) = 0

> In
> audio, they call that (S+N)/N, instead of S/N, and it's rarely an
important
> distinction because usually S>>N.

> But the reason I wouldn't define dynamic range that way

Which way?

> is that
> if the noise
> doesn't have the same spectrum as the signal, you can often extract the
> signal even when its power is much less than the noise.

There are noise reduction techniques for imaging, of course, but they
typically reduce/average the noise, not extract the signal from the noise.
I believe this application (scanner data) would be considered static in this
sense, as in the measurements are done at one point in time to static data,
not over time to a changing signal, and what you say may not be applicable?
I'm not sure what technique(s) you may be referring to.

> I gave the example
> of an ink jet print, which technically has a "dynamic range" of about 0db
> (since the noise amplitude is full scale), but sure doesn't look
> that way to
> the eye.

I'm not sure why you say that the noise amplitude is full scale...would you
please explain?

Regards,

Austin

----------------------------------------------------------------------------------------
Unsubscribe by mail to listserver@halftone.co.uk, with 'unsubscribe 
filmscanners'
or 'unsubscribe filmscanners_digest' (as appropriate) in the message title or 
body