At 10:42 AM 12/02/1999 -0500, Arnold G. Reinhold wrote:
>http://www.dragonsystems.com/products/audiomining/
>
>"New AudioMining Technology Uses Award-Winning Speech Recognition
>Engine to Quickly Capture and Index Information Contained in Recorded
>Video Footage, Radio Broadcasts, Telephone Conversations, Call Center
>Dialogues, Help Desk Recordings, and More
It's easier to do that with single calls than with large numbers at once;
scaling can be hard or at least expensive. A T3 carries 672 calls,
unless there's voice compression or silence suppression increasing the
capacity (which there probably is, for international circuits),
and an OC48 carries 48 T3s. So if you need a P133 to track one call,
you need about 30000 of them to track a wavelength. I suppose that's
only a few million bucks if designed efficiently, but that's more than
the fiber optic network equipment on each end costs the telcos :-)
(plus you need the euivalent fiber optic equipment as well.)
This may still be affordable for international circuits, but it's
unrealistic for land-based communications. For instance, AT&T announced
having installed a thousand or two OC48-segments a year,
though many long-distance calls go across multiple segments
because they're going a long ways. Adds up to a lot of wiretapware.
On the other hand, if you're only looking for a small number of
Usual Suspects, or can use dialing information or other call setup data
to figure out which calls you want to eavesdrop on,
the scale becomes much more manageable.
Thanks!
Bill
Bill Stewart, [EMAIL PROTECTED]
PGP Fingerprint D454 E202 CBC8 40BF 3C85 B884 0ABE 4639
