[cryptome] Re: Noise As a Unique Identifier. and voice print identification

  • From: coderman <coderman@xxxxxxxxx>
  • To: cryptome@xxxxxxxxxxxxx
  • Date: Sat, 6 Dec 2014 16:27:38 -0800

On 12/6/14, doug <douglasrankine2001@xxxxxxxxxxx> wrote:
> ...
> Back to noise and the Tempest, I did notice the other day, that when I
> used my new tablet to access my online trading account for the first
> time, that I had to confirm its use via my phone.  The phone call to me
> was automated, and obviously, my voice had been recorded by some kind of
> voice pattern recognition software, by the company concerned.  My wife
> and I were most surprised, because the quality of the phone call was
> such that it was almost...I say almost, imperceptible from a real human
> being.  The reason it was so surprising was because I speak with a very
> strong Scottish accent, and me missus says I mumble rather than speak
> coherent English, yet the software understood me perfectly. Voice
> pattern recognition has come a long way, since those early days...

it used to be military, then commercial, now widespread (to the individual).

in early 2000's we used voice print authentication for a commercial
directory assistance service.  "hello, infone" all it needed to give
confidence sufficient to authorize for charges.

(the margin of error, of course, tune-able to desired tolerances, and
indeed, this is where the bulk of complexity lies - too tight, and you
can't auth from a pay phone or noisy cell call. too loose, and the
number of false positives skyrockets...)

> P.S.Did you notice...In the last "Homeland" series, the "secret"
> services, whoever they were, used the control of a heart monitor to bump
> off one of the "opposition"?  Such a "fictional" series must be having
> some kind of effect on the masses, don't you think...

Cheney's pacemaker had the wireless capabilities removed, specifically
because of this risk, which at the time was under-appreciated by the
public. (recent medical device hacking research at security
conferences has raised the public awareness :)


best regards,

Other related posts:

  • » [cryptome] Re: Noise As a Unique Identifier. and voice print identification - coderman