[Discuss-gnuradio] PMR data decode

[Matthew Burnham]

> > Is anyone working on deocding of data found in the PMR
> bands (e.g. MDT
> > terminals, pagers, etc.).
> I'm not familiar with those specs, but if you send more info, and 
> possibly some audio samples (assuming this signal is narrower than 20 
> kHz BW), it might be pretty easy to decode.

I can send some samples, will have to work out how to do it in linux as
I've not 'fixed' the input gain under Windows yet... But the protocol of
the things I'm interested in is unknown. I could easily first look into
POCSAG or something else known. Some of the signals might well be
MPT1327 trunking, which I don't know much about yet, but believe are low
enough bandwidth to get on my scanner so should be do-able (and the
standard is open so would be easier).

> > I'm quite impressed with GNU Radio so far
> > having got a nice display of the waveforms using the audio scope 
> > example. I'm not sure how to proceed in decoding in terms of the 
> > modules to link together. I envision something along these lines:
> 
> It depends largely on what type of modulation it uses. Assuming FSK, 
> the chain would most likely be
> 
> AudioSource
> FreqXlatingFIRfilter
> Quadraure Demod
> Filtering/Decimation
> Data Slicing
> Output

Okay, so perhaps my next question is which of these modules are ready to
plug in, and which do I need to do some work on.

> > This example is for data where the bits are manchester encoded in 
> > amplitude when viewed on the scope. (I originally believed
> this to be
> > FSK, but now I'm not sure as I've seen some other data signals that 
> > seem to be a sinewave at two discrete frequencies, which would 
> > necessitate further FSK decoding first. Anyway, that's a
> side issue on
> > my lack of understanding of FSK modulation.)
> > 
> > I've got my head around a lot of the architecture having
> read through
> > the entire mailing list archive and the (limited)
> documentation, but
> > I'm not sure how I would go about getting from an audio input to a 
> > data output. I connected a VrAudioSource to a VrFileSink, but I'm 
> > not sure of the format of the data in the file
> 
> If I understand your question correctly, the data file output is 
> pretty much raw binary.  It could be bytes, words, floats, complex 
> floats, etc.

I didn't explain well, but now you've suggested a suitable path (above),
I think its clearer in my mind. After data slicing I would have binary
data (in effect), how would I go about getting this out to a program to
decode the datastream? Naively, this would be the output section, and it
may well just 'work' once I get the parts above it together. Say I used
VrFileSink, would I end up with a binary stream of the data, or am I
missing something?

I can understand that down the chain to the data slicing point, the
values being passed are all floating point, but after this point they
become binary. Perhaps (in trying to answer my own question) Data
slicing and output are one unit? Data slicing being a Sink?

Cheers for such a quick reply!

-- 
Matthew Burnham