Apple is going to beat all competitors

Federico Lorenzi florenzi at gmail.com
Fri Sep 7 23:40:16 CEST 2007


Wow, this thread has gone really off topic :) But I figure I'd add in my 2c.
A GSM Data call can carry up to 9.6kb/s. Using a good voice codec
like Speex, one can get voice still sounding decent at 2kb/s.
I don't really know much about the inner workings of GSM, but is there
anything stopping one Neo from establishing a data link with another
via CSD, then using netcat or something to simply pipe audio from the
microphone into speex and then into a command line blowfish / twofish
or whatever encryption you want, then on the other side do the same,
but in reverse.

I have actually done what I mentioned above
(Mic -> Speex -> Encrypt -> Comm -> Decrypt -> Speex -> Speaker)
on my computer using some simple python scripts, and it works fine.
The major issue IMO would be getting it from one point to another.

Cheers
Federico

PS) I probably don't know what I'm talking about, eat with a pinch of salt.

On 9/7/07, Mikko Rauhala <mjrauhal at cc.helsinki.fi> wrote:
> pe, 2007-09-07 kello 13:35 -0700, Shawn Rutledge kirjoitti:
> > It's easy to think first of doing it over a data connection.  But I
> > can imagine more of a signal-processing approach.
>
> This has been hashed to death. Short of it: the GSM codec actively tries
> to throw away anything that doesn't sound like speech. This is not very
> analog encryption friendly, and we can't bypass the codec for voice
> calls since that's in the closed TI firmware.
>
> This is also a repeat, but for new people, I'll mention that
> cryptophone.de does a proprietary encryption-capable GSM phone (through
> GSM data); though the protocol is their own, they apparently have it
> documented, and welcome others to do compatible implementations. (Their
> source is also available, but for purely review and verification
> purposes; it's not free and anyone doing a reimplementation would be
> wise to avoid looking at it.) Thus for anyone interested in doing
> encrypted calls with the Neo, it would perhaps be worth considering
> using their ready-made protocol to leverage their installed base.
>
> I haven't reviewed their protocol though, nor am I highly qualified to
> do that. Also, there's the usual potential future issues with the
> protocol being controlled by a single firm; basically that's just
> something to consider for those who eventually get hacking on this.
> Sadly, it's too big a project for me.
>
> --
> Mikko Rauhala   - mjr at iki.fi     - <URL:http://www.iki.fi/mjr/>
> Transhumanist   - WTA member     - <URL:http://www.transhumanism.org/>
> Singularitarian - SIAI supporter - <URL:http://www.singinst.org/>
>
>
> _______________________________________________
> OpenMoko community mailing list
> community at lists.openmoko.org
> http://lists.openmoko.org/mailman/listinfo/community
>




More information about the community mailing list