[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [HLL] VP1 Comparison



Status: RO
Content-Length: 2540
Lines: 57

First an intro. I am Lucky, I have done crypto since 1989 and I have
extensive experience with digital telephony and isochronous protocols. I
am a long time subscriber to the Cypherpunks mailing list. [Hence the
domain cypherpunks.to :-]

 On Tue, 6 Jan 1998, Volker Ernst wrote:

> -----BEGIN PGP SIGNED MESSAGE-----
> 
> Hi,
> 
> 
> i had a quick look at the protokoll-specs for the "VP1"-Device.

What is the VP1 device and where can I find the specs?

> Here is a short comparison:
> 
> * They operate their 14400-Modem is *synchronous* operation.

Sure. 14.4 sync mode is the only way you can fit a  GSM 06.10 full rate
speech transcoder (13,000 bits/sec) plus protocol overhead into a V.32bis
connection. If you go async, you have to use a lower voice quality codec.
The next fallback is something around 7200bps. Which makes a huge
difference in voice quality.

 >   This means, that they do not have the trouble of baud-rate
>   variations between cpu/modem and no inserted/deleted bytes.

But you will have the problem that hardly any PC can talk sync to the
modem. Sync is best used in stand  alone devices such as the the GSP8191.
http://www.comsec.com/

[Meta:  the GSP8191 is by far the best phone encryption device available
today. The voice quality is excellent  and the protocol is as solid as it
gets. If you have not read the protocol specs, you would likely greatly
benefit by doing so now].

>   This makes a big difference in low-level frame-layout.
>   Because we use a asynchronous channel in between (to our modem),
>   we do have the problem of baud-rate-variations.
>   Because of this, we cannot make use of 100% bandwitdh.
>   We will also have to handle the situation of missing/inserted data.
>   Their synchronous communication gives them 14400 Bits all_the_time,
>   while our modem might even deliver no data at all (in error-condition).
>   This means that we might loose frame-sync due to com-errors,
>   they might get wrong bits - but the bit-count (sync-pos) is always ok !

OK. For "business quality" encrypted voice, assuming GSM codecs, you need
14.4 sync. You can get "sort of ok", in 9600bps async with a 7200bps
codec. But to be honest, I don't like the sound quality.

Next question: I am told you guys want this to work over IP. That's a
whole different ballgame. What are your plans for this?

-- Lucky Green <shamrock@cypherpunks.to> PGP v5 encrypted email preferred.
   "Tonga? Where the hell is Tonga? They have Cypherpunks there?"