Novelty "Retro" Cell Phone Handset in an Acoustic Coupler?

On 17 April 2014 06:31, Mark J. Blair <nf6x at nf6x.net> wrote: I haven't tried it, but having worked in the field of codecs for cell phones (AMR-WB, VMR-WB and AMR-WB+ to be specific) I would caution you that the codecs on modern, digital phones are not at all designed for modem-type signals (eg QPSK etc) which - in their design - assume mostly linear channels with some bandwidth and noise floor. At low enough speeds (and 300 baud ought to qualify) the nonlinear distortion that the CELP codecs will introduce might not matter, but if it's RCELP (regularized or relaxed CELP) it is very much designed to modify the pitch imperceptibly to humans which would cause bit errors for any modem signal. (But I don't know of any RCELP in the GSM or CDMA standards, so this might not matter) In short, try it, but I'd be sceptical that you'd get good results. I suggest a different approach. Write an iOS or Android app that captures the mic and controls the speaker of the handset; do the modulating/demodulating on the phone, then the app can pass the data over ssh/telnet to a net-connected BBS. Or pass it to an emulated machine on the phone. (I've seen SIMH for Android - not sure about iOS). (I've been a TA for an undergraduate-level DSP course where students implemented 4-QPSK in C on a TI eval board (I designed that experiment with a nod to classic computing in mind) - I believe that could get you up to the 2400 baud standard, but with an acoustic coupler you couldn't go that high anyways. I can't remember if 300 Baud is actually simple frequency-shift keying (FSK) which is very simple to understand and implement) Cheers, Joachim. -- Joachim Thiemann :: http://jthiem.bitbucket.org :: http://signalsprocessed.blogspot.com