Newsgroups: sci.engr.advanced-tv,sci.electronics,rec.video.satellite,comp.dcom.telecom.tech,comp.speech
Path: cantaloupe.srv.cs.cmu.edu!nntp.club.cc.cmu.edu!godot.cc.duq.edu!hudson.lm.com!netline-fddi.jpl.nasa.gov!elroy.jpl.nasa.gov!decwrl!netcomsv!netcom.com!jgk
From: Joe Keane <jgk@netcom.com>
Subject: Re: Spectral Inversion
Message-ID: <jgkD1I6x5.E7v@netcom.com>
Summary: It's lame.
Keywords: speech
Sender: jgk@netcom.com (Joe Keane)
Organization: none
References: <34or00$kms@papaioea.manawatu.planet.co.nz> <3cut35$p0j@eis.calstate.edu> <D0z7DL.Kqx@actrix.gen.nz> <3d13vm$hva@CSOS.ORST.EDU>
Date: Wed, 28 Dec 1994 04:24:40 GMT
Lines: 40
Xref: glinda.oz.cs.cmu.edu sci.engr.advanced-tv:2119 sci.electronics:113153 comp.dcom.telecom.tech:10686 comp.speech:4180

In article <3cut35$p0j@eis.calstate.edu>
John E. Lundgren <jlundgr@eis.calstate.edu> writes:
>The speech scramblers I've seen do not break the audio band up into 5 
>or more segments.  They merely choose a single frequency, and invert the 
>audio around that single frequency.  They use a 1496 chip which is a 
>multiplier or balanced modulator.
>
>Some of the newer cordless phones are using speech scrambling, and they 
>are using this inferior method to invert the speech.  They don't tell the 
>buyer that the system is easily defeated.

Right, this is lame.

In article <D0z7DL.Kqx@actrix.gen.nz>
Alex Ivopol <cyborg@atlantis.actrix.gen.nz> writes:
>I've seen a new 8 pin DIL chip which does spectrum inversion and also cuts
>the range 300-3000 Hz at a user defined point and swaps the two segments.
>There are about 12 or so cutpoints that can be selected by the user via a
>I2C bus and the cut point can be changed about 10 times per second. This is
>similar in concept to the "line cut & rotate" principle used to scramble
>pictures in the VideoCrypt system. If the cut point is selected via a pseudo
>random number generator this system can be made quite secure. Especially
>if the pseudo random number generator is reseeded every so often.

Sorry, this scheme is not at all secure, especially with only one cut-point.
Changing the cut-point may make it so you can't understand just by listening,
but it doesn't give real security.  Speech has a lot of redundancy, so you can
figure out what's being said, without cracking the random-number generator.
This sort of cut-and-paste is hardly a substitute for real encryption.

In article <3d13vm$hva@CSOS.ORST.EDU>
Joel Kolstad <kolstadj@CSOS.ORST.EDU> writes:
>I still think that, if someone really wants to listen to _your_ phone calls,
>it's much easier for them to go after your landlines long before they
>bother with your cordless phone.

Why go outside when you can work from the privacy of your home? :-)

--
Joe Keane, amateur mathematician
