===================================================
= NB: Over 50% of this message is QUOTED, please =
= be more selective when quoting text =
===================================================
> From: Stainless Steel Rat [mailto:[EMAIL PROTECTED]]
> * Steve Hill <[EMAIL PROTECTED]> on Wed, 06 Jun 2001
> | On a CD the laser is reflected off the disc, onto a photodiode which
> | produces four analogue voltages. These are then used to
> perform tracking
> | and linear speed adjustments, and are also processed to produce the
> | digital data. The signals could just as easily be used to produce an
> | analogue audio signal.
>
> Could, but they are not. And though what is there could be
> interpreted as
> an analog signal, it really isn't. Consider this: I whisper "one". I
> shout "one". Has the value of "one" changed? My voice is
> analog, but the
> spoken signal is still digital.
If it wasn't really an analogue signal, there would be no need for error
correction coding. In practise, the analogue signal is converted into soft
decisions which vary from zero, through zero-ish, one-ish, to one. These are
then fed into the error corrector, which uses the redundancy introduced into
the data to estimate the most likely sequence of data bits. Continuing your
analogy, if you whisper 'one' but I can't hear what you say, I have to make
a guess on whether you said 'one' or 'zero'. The source data is digital, but
the signal after the effects of the channel are considered is analogue, and
must be converted back into a digital representation.
> [...]
> | No. The advantage of CD and MD against vinyl and tape is
> that they do not
> | wear out. The fact that the former are digital and the
> latter analogue is
> | co-incidental. DCC and DAT both wear out yet are digital.
>
> And are insignificant to consumers.
>
> | Laser discs were entirely analogue and do not.
>
> Actually, they do. And the format is (was) insignificant to
> consumers,
> too.
My point was not what was significant to consumers but rather that it is a
property medium, and not the format of the data stored on it, that
determines whether wear occurs.
> | Admittedly, due to the error correction, digital recordings
> will handle
> | wear better for a while before failing completely where as analogue
> | recordings deteriorate more gradually, but it is the medium that
> | determines whether wear occurs.
>
> Well, if you want to insist on picking nits, then consider
> this: microphone
> in to a solid state deck, real-time conversion to MPEG-1
> Layer III audio,
> and stored on compact flash cards. No analog storage
> involved anywhere.
The bits are stored by tunnelling electrons through the oxide layer,
generating a potential on the floating gate. That potential is analogue -
though, if you want to get pedantic, quantised. The real world is analogue
and, hence, all data stored in the real world is stored in an analogue form.
It's all a bit academic, though.
> [...]
> | I'm not sure that I understand the point that you are
> trying to make here.
>
> My point is that the original post making the claim that, paraphrased,
> "digital takes more space to store than analog because square
> waves take up
> more space," is wrong.
I can't argue with that being faulty logic. In the general case,
uncompressed digital signals take more space to store than the analogue
signal that they represent. However, if you just wanted to store square(ish)
waves, it would require much more bandwidth to store them in an analogue
manner than digitally. The key is that it is wrong to think of digital as
storing things as square waves.
S.
-----------------------------------------------------------------
To stop getting this list send a message containing just the word
"unsubscribe" to [EMAIL PROTECTED]
