[filmscanners] Re: Density vs Dynamic range>AUSTIN (2a)

["Anthony Atkielski" <anthony@atkielski.com>]

Paul writes:

> This is nonsense. A steady DC voltage contains
> an amount of information equal to log2 n, where
> n is the number of different DC voltages that
> can be distinguished by the receiver.

No.  In fact, you contradict yourself in this statement.  A steady DC
voltage does not change, and yet you speak of "the number of different DC
voltages that can be distinguished."  How will these different voltages be
distinguished, given that DC does not change?  The information carried by
the medium resides in the transitions embodied in that medium: changes from
one voltage to another, or from one place to another, and so on.  The only
way to use different DC voltages is by providing for a transition between
voltages; but a transition is a variation in the voltage, and is thus
effectively AC.  If there are no transitions, the voltage never changes, and
it really is DC; but in that case, it carries no information, either, and so
it is not a signal.

> It conveys this information even if it remains
> unchanged for a million years.

No.  It conveys information only if it changes.  The whole purpose of a
signal is to communicate information, and the only way to communicate
information is through change.  The only utility in a signal that does not
change for a million years arises when the signal _does_ change; if it never
changes, it may as well not exist, since it communicates no information.

> This information is exactly what is
> displayed by a DC voltmeter.

A DC voltmeter has a dial, allowing for changes in voltage.  If the incoming
voltage never changed, it wouldn't need a dial; it wouldn't even need an
on/off light, since the voltage would always be there.

These are practical applications of information theory, and they are
unavoidable.  Understanding them is important to understanding many aspects
of digital imaging.



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