[lug] 8mm video vs. data tapes

[J. Wayde Allen]

On 29 Apr 2001, Tkil wrote:

> Is this still true with some of the fancier encoding methods that are
> being used on magnetic media?

I don't really know.  My comments were only related to the generic concept
of recording analog data versus digital data on a magnetic tape.  My first
reaction would be to think that "fancier encoding methods" is starting to
get you into the realm of digitally encoded analog signals or at the very
least more complicated analog modulation schemes.

>  From a discussion of PRML in:
> 
>    
> http://www.usenix.org/publications/library/proceedings/sec96/full_papers/gutmann/

I haven't had a chance to read the article you sent, but did skim it.  
Basically this paper is talking about how the recording of magnetic info
is not typically an all or nothing event.  In other words, not all of the
magnetic domains in the region of the read/write head will flip and with a
magnetic force microscope it is possible to extract multiple signals from
the tape.  

I think that some of the original work on this method of extracting data
from previously erased tapes was done by a couple of guys here at the
Boulder Labs.  I personally don't know much about it, but could dig up
some stuff if you are interested.

> it seems that (at least on hard drives) it's no longer an obvious
> string of ones and zeros (and it hasn't been for a while).  The author
> of the above paper compares it to the symbols used in modern modems
> (where you might have only a 4800 baud signal, but you're transferring
> 33600 bits per second -- implying 7 bits per symbol, or 128 distinct
> symbols.)

Well ... actually when you get right down to it computers, so far anyway,
have never really been completely digital.  All electronic circuitry has
an inherent noise level, and as such the logic signals have to exceed this
level in order to be determinant.  In other words you can say that a zero
is zero volts plus the intrinsic noise in the circuit.  A logical "1" has
to exceed the intrinsic noise thresh hold, and also has to include
variations in the power supply.  The point being that there is always some
range of voltages that represents a logic symbol.

When it comes to signal modulation such as you'd see for a modem, or a
wireless link you can get into lots of different schemes.  Are you
modulating amplitude, frequency, phase, or some combination of these
values?

> So "data tape" might not be so different from "analog tape" after all.

That is pretty much the point I think.  They are fundamentally the same
thing.  My understanding, meager as it may be, is that the differences lie
in the relative background noise levels.  

My Dad has a small recording studio, and I do know that if you erase and
reuse an analog tape, the noise level goes up.  This is due to the
residual signal from previously recording on the tape, and is what your
web article talks about.  If you are making a true analog recording this
is an issue.  If you are using a recording scheme where there is some
encoding going on then you could more than likely tolerate more background
tape noise.

- Wayde
  (wallen at lug.boulder.co.us)