1996-10-08 - Re: Voice Stress Analysis of Debates?

Header Data

From: jim bell <jimbell@pacifier.com>
To: cypherpunks@count04.mry.scruznet.com
Message Hash: 517eab3d250ed10f595b1bc99f9c3b177c402841adfcafedf0962e1a458f2ed0
Message ID: <199610071830.LAA26780@mail.pacifier.com>
Reply To: N/A
UTC Datetime: 1996-10-08 00:07:54 UTC
Raw Date: Tue, 8 Oct 1996 08:07:54 +0800

Raw message

From: jim bell <jimbell@pacifier.com>
Date: Tue, 8 Oct 1996 08:07:54 +0800
To: cypherpunks@count04.mry.scruznet.com
Subject: Re: Voice Stress Analysis of Debates?
Message-ID: <199610071830.LAA26780@mail.pacifier.com>
MIME-Version: 1.0
Content-Type: text/plain


At 12:30 AM 10/7/96 -0700, cypherpunks@count04.mry.scruznet.com wrote:
>PSE/VSE etc relys  on a random 8-14 hz fm modulation appearing in the
> the 3rd formant band of the voice... the Dektor PSE retailed for 
approximately 5k in the early 1970's the hagoth hs-2 was the early 80;s for 
2k .. 

The technical details look vaguely familiar, but I probably last read them 
two decades ago.  It doesn't sound like  it would be too complicated to 
implement, with ordinary phase-lock loops, switched-capacitor filters, and 
other analog building blocks.

>I bought 2 VSE units based on a new chip for 89.95 apiece recently...
>this was on an article appearing in popular electronics in 1995

I'll  have to look that up.  Was it a DSP or some analog implementation?  
Most importantly, does it seem to WORK?  

>no software as of yet and it would be probably a FFT or DFT based algorithm 
>running on a sound blaster DSP(hardware DSP only) or as software
> on a fast 150 Mhz + pentium I can give the basic facts known... it is a truth
>detector only(not a lie detector), background music shows up a stress...
>and scrubbing the modulation shows up as constant stress or "possible 
decption"

Yes, it occurs to me that a vocorder-type compression/decompression of the 
voice would, effectively, remove FM modulation by averaging out short-term 
variations in voice frequencies.


>truth is indicated by the modualtion showing up randomly over small increments
>of time... the effect is caused by microscopic tremors of the larynx when the 
>speaker is relaxed and truthful the larnyx is relaxed and the random 
modulations
>show up,(or the speaker believes what he is saying to be the truth, when lying
>or under stress the larybx tightens and the random modulations
>disappear indicating stress... it would be  nice to have it in software...
>to run on a laptop soundboard and spread to the net as widely as pgp...

I think it's particularly revealing that the TV networks don't try to use 
it.  They're supposed to be looking for an edge, something to make the news 
seem more interesting.  The political establishment would see the 
publicizing of  this as going beyond an unwritten limitation on the media.

 
Jim Bell
jimbell@pacifier.com





Thread