From: Derek Atkins <warlord@MIT.EDU>
Date: Sun, 31 Oct 93 19:53:47 PST
To: Matthew J Ghio <mg5n+@andrew.cmu.edu>
Subject: Re: Secure Phone Progress (fwd)
In-Reply-To: <0gp7NE600awUQ5lkYP@andrew.cmu.edu>
Message-ID: <9311010351.AA03794@hodge.MIT.EDU>
MIME-Version: 1.0
Content-Type: text/plain


Matt, 

There are a few flaws in your statements.  You are correct that human
hearing is up to 20 kHz, and again with the 40 kHz sampling (this is
why CD is 44.1 kHz, to be able to fully obtain a 40 kHz sample with
non-ideal filters).

However, we are talking about Secure Phones.  Phones have a dynamic
range of 3.6 kHz, with a sampling of 8000 samples/sec!  Also, the
phone uses 8-bit mu-law data, for the 64000 bits/sec that you commonly
hear.

So, without compression, you need a 64 Kbps link to transmit
phone-quality audio.  The problem with a Secure Phone is getting the
data compression to reduce the data to below 14.4 Kbps, in order to
use a 14.4 Kbps modem.  You need a real-time compression to do this,
which is where the DSP chips come in!  The encryption isn't a problem,
as most encryption algorithms can easily do 64 Kbps!

-derek

         Derek Atkins, SB '93 MIT EE, G MIT Media Laboratory
     Secretary, MIT Student Information Processing Board (SIPB)
         PGP key available from pgp-public-keys@pgp.mit.edu
            warlord@MIT.EDU       PP-ASEL        N1NWH