Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!uunet!cbmvax!grr
From: grr@cbmvax.commodore.com (George Robbins)
Newsgroups: comp.dcom.modems
Subject: Re: Inexpensive 9600 baud modems
Keywords: 9600, V.32, V.42, bis
Message-ID: <16309@cbmvax.commodore.com>
Date: 5 Dec 90 14:49:16 GMT
References: <136548@pyramid.pyramid.com>
Reply-To: grr@cbmvax.commodore.com (George Robbins)
Organization: Commodore, West Chester, PA
Lines: 69

In article <136548@pyramid.pyramid.com> lstowell@pyrnova.pyramid.com (Lon Stowell) writes:
> In article <5435@navy19.UUCP> benyukhi@motcid.UUCP (Ed Benyukhis) writes:
> >57 Kbps on the voice grade line that is band limited to 3.4 Khz is contrary
> >to both Shannon and Nyquist rules.
 
> Maybe so, but the POTS line, with its 3.4 Khz limit can easily
> handle 160 Kbit/second or so rates.  ISDN easily places 144 Kbit
> USER rate on this 2-wire line, using 2B1Q signalling...

Are you sure you're not confusing the wires with the bandwidth?

The standard local loop cables can handle a much wider bandwith
than the 3.4 KHz nominal channel width.  The nominal bandwidth is
imposed by various devices used to optimize the longer loops for
voice transmission, analog components in the central office and
the channelized multiplexing and bandwidth limited digitizing used
in the long distance network and digital switches.

There is also a secondary regulatory issue - when the telco supplies
a pair of wires for a connection to the direct dial network, one
set of bandwidth and signal level rules apply, when a similar pair
of wires from the same cable is offered for other purposes, these
particular rules no longer apply.

> Don't confuse the Nyquist limit on "baud" or carrier signalling
> interval over 3.4 KHz with the ability to encode and decode user
> information as multiple bits/baud over the bandwidth limited
> channel...   38.4 USER rate works quite nicely over V.32 modems
> by using Trellis (5:4 redundancy) and QAM (4:1) signalling at a
> baud rate of only 1200 on the analog channel...although the
> "raw" bit rate is only 9600...and V.32bis will push this to 14.4
> raw rate,,,should be able to use data compression to get 57.6
> out of a V.32bis modem easily...

All that is nice, but I think you'll also find that if the data
can be compressed then it's statistical properties are less than
optimal, i.e. the information is poorly encoded.  It seems that
Nyquist and Shannon still rule, but as with's Einstein's theory
of relativity, a quick grasp of some of the features of the theory
do not convey the complete understanding of the implications of
the theory or the conditions.

I might tend to agree with the orignal poster in the sense that
any modem will have real trouble giving 56KBPS with random data
over arbitrary direct dial connections - first of all the notion
that 4:1 compression is normal is bogus, 2:1 is typical of normal
text, while random data doesn't compress at all - second there
is are real theoretical limits on data throuput versus channel
quality that sooner or later intersect with the properties of
real-world direct dial connections.

Right now there seems to be pessimism as far as getting past the
14K-19.2KBPS range of real thruput.  I don't claim to know what the
theroretical limits are, but I'm happy to have modems that compress
or take advantage of unusually good connections as long as they
offer good thruput under impaired conditions and reliable operation
under any conditions.  V.32 seems a bit marginal in these terms, but
V.32bis may consitute a real improvement, not just because it's
"faster", but also because it's more agile.

In the meantime, my Trailblazers work reasonably well, not because
they do anything magical or really sophisticated, but because they
adapt well to (most) real world line conditions and implement
(reasonably) rugged protocols and error control techniques.

-- 
George Robbins - now working for,     uucp:   {uunet|pyramid|rutgers}!cbmvax!grr
but no way officially representing:   domain: grr@cbmvax.commodore.com
Commodore, Engineering Department     phone:  215-431-9349 (only by moonlite)