Re: looking for research partner [block cipher design]

From: Rob Warnock (rpw3@rpw3.org)
Date: 02/25/03 

From: rpw3@rpw3.org (Rob Warnock)
Date: Tue, 25 Feb 2003 06:31:21 -0600

Nicol So <see.signature> wrote:
+---------------
| Rob Warnock wrote:
| > p.s. Note that "baud" is used above with the meaning of symbols per second.
| > Technically, it's defined as "the reciprocal of the smallest signalling
| > interval", but only inefficient encoding schemes send less than one symbol
| > per signalling interval, so for conversational purposes: baud == symbols/s.
|
| I don't think that's correct. In the definition I'm familiar with, one
| baud equals one signaling event per second. It's not defined in terms of
| the shortest inter-event interval. (There's a difference between the two
| definitions, because signaling events need not happen at regular
| intervals.)
+---------------

That's correct, there *is* a difference between the two definitions,
but the one I gave is also the correct one. Some encodings do not have
uniform clock transitions (e.g., MFM comes to mind), but the baud rate
for any encoding is the reciprocal of the *smallest* inter-symbol interval
within that encoding.

See "Technical Aspects of Data Communication", by John E. McNamara,
which says basically the same thing: "A baud is the reciprocal of the
unit interval." Or as given in FOLDOC, "...the reciprocal of the duration
of the shortest signalling element."

Or at <URL:http://www.its.bldrdoc.gov/fs-1037/dir-004/_0559.htm>:

        One baud corresponds to a rate of one unit interval per second,
        where the modulation rate is expressed as the reciprocal of the
        duration in seconds of the shortest unit interval.

Etc., etc., etc...

The distinction is particularly important in bandwidth-constrained
environments, where the *smallest* inter-symbol interval (1/baud)
sets the minimum bandwidth required to carry the signal faithfully
without corruption.

Note that the Nyquist limit sets a hard floor on bandwidth consumed
at 0.5 Hz/Baud, but in practice that's not attainable, since there are
no such things as perfect filters and perfect inter-symbol interference
compensation (equalization). Practical limits (depending on cost, mainly!)
range from 0.75-2.0 Hz/Baud, with the lower values being harder to attain.

-Rob

-----
Rob Warnock, PP-ASEL-IA <rpw3@rpw3.org>
627 26th Avenue <URL:http://rpw3.org/>
San Mateo, CA 94403 (650)572-2607

Relevant Pages

  • Re: Speed
    ... while the signalling speed refers to the ... consider unencoded NRZ data transmission. ... many encoding schemes that are used to transmit digital ... Serial fiber optic transmission is commonly done at 10Gb/s. ...
    (sci.electronics.basics)
  • Re: SETI is pointless
    ... I have not left out bandwidth. ... I have fully understood the impact long distance signalling HAS on ... AT HIGH CARRIER WAVE FREQUENCIES OVER VERY LONG DISTANCES YOU ... That's why OSETI uses signalling rates of less than 1GHZ despite the ...
    (sci.space.policy)
  • Re: Powerline modem
    ... >turned off by the power company, during periods of high demand, using a ... Note however that these systems use signalling frequencies of 200 Hz ... Bandwidth and other constraints limited ... at zone substations (signal passes through one transformer) or at grid ...
    (sci.electronics.design)
  • Re: Does more bits-per-character but less characters-per-second make better use of b
    ... Gigabit ethernet uses such a multi-level signal ... Multilevel signalling depends on high SNR, but you win logarithmically with transmitted power, rather than linearly with bandwidth or number of parallel channels. ...
    (sci.electronics.design)