From: Phil Karn <karn@qualcomm.com>
To: Eric Blossom <MIKEINGLE@delphi.com
Message Hash: afa45706217152a03e4b2cba6497d904162ea7787ccd5bb0bcf0a9209f6814a0
Message ID: <9311180612.AA03568@servo>
Reply To: N/A
UTC Datetime: 1993-11-18 06:14:28 UTC
Raw Date: Wed, 17 Nov 93 22:14:28 PST
From: Phil Karn <karn@qualcomm.com>
Date: Wed, 17 Nov 93 22:14:28 PST
To: Eric Blossom <MIKEINGLE@delphi.com
Subject: Re: Secure phones - STU3
Message-ID: <9311180612.AA03568@servo>
MIME-Version: 1.0
Content-Type: text/plain
As far as I can tell, basic compatibility among STU-IIIs calls for the
2400 bps LPC-10 coder as a minimum. So it's universal. It works, but it
sounds pretty bad, so there's been a lot of work on newer and better
vocoders. Correct me if I'm wrong, but I suspect the Motorola coder you
mention is probably one of several manufacturer-specific algorithms,
sort of like the manufacturer-specific high speed modes you find on some
fax machines.
FED-STD-1016 CELP (Codebook Excited Linear Prediction) at 4800 bps seems
to be the up-and-coming standard for newer phones, and it indeed does
provide *much* better voice quality than LPC-10. Unfortunately, it also
requires many more DSP cycles.
Phil
Return to November 1993
Return to “Phil Karn <karn@qualcomm.com>”
1993-11-18 (Wed, 17 Nov 93 22:14:28 PST) - Re: Secure phones - STU3 - Phil Karn <karn@qualcomm.com>