On Mon, Mar 4, 2013 at 4:40 PM, Michael Rogers wrote:

Last year I spent some time playing with audio encoding of data for transmission over handheld radios. The state of the art here is dialup modems - on a good day they can get 56,000 bits per second over a channel designed for voice, but that requires advanced modulation and error correction techniques. The radio hams have packet radio (AX.25 and APRS) running at 1,200-9,600 bps over long distances using simple modulation and no error correction. Some early home computers used audio cassettes for storage (300-1,200 bps, CUTS or Kansas City Standard).

Nice information, thanks. Would it be wrong to assume larger data rates to be attainable on an FM "link" than over the telephone line? For music etc. FM has far superior sound quality in any case.

If you want to support purely one-way communication (no acks), you'll need to forward error correct the data. Hamming codes and parity checks are simple to implement but they'll eat a lot of your bandwidth; Reed-Solomon codes are more bandwidth-efficient but also more complex.

Yes, I thought of that too in September. Luckily I've taken courses in abstract algebra and error-correcting codes at my university; I think I'd be able to write a working RS implementation from my theory books. Another thing I didn't tell in my first mail is that I've been wanting to design a protocol for metadata, too, since it doesn't really make sense to decode and save half files anyway. It would also make it possible to send the file names and file sizes beforehand so the receiver can know how much of the file s/he has already received. And yes, I want this to be truly one-way -- no acks. The idea is that I want the receiving end to need as little hardware as possible: one FM radio and one computer with a sound card (and this software). The sender obviously has access to an FM transmitter (or whatever becomes the sound carrier). This modulation algorithm should not provide authenticity of the sender, instead cryptographic signing of the data should happen at higher levels of the stack.

Some Java code for modulation, framing and error correction is here if you're interested:

http://briar.git.sf.net/git/gitweb.cgi?p=briar/sandpit;a=tree;f=src/net/sf/b... http://briar.git.sf.net/git/gitweb.cgi?p=briar/sandpit;a=tree;f=src/net/sf/b...

Thanks a lot! I'll have a look at it soon. JC -- Too many emails? Unsubscribe, change to digest, or change password by emailing moderator at companys@stanford.edu or changing your settings at https://mailman.stanford.edu/mailman/listinfo/liberationtech ----- End forwarded message ----- -- Eugen* Leitl <a href="http://leitl.org">leitl</a> http://leitl.org ______________________________________________________________ ICBM: 48.07100, 11.36820 http://www.ativel.com http://postbiota.org 8B29F6BE: 099D 78BA 2FD3 B014 B08A 7779 75B0 2443 8B29 F6BE