Re: [Discuss-gnuradio] FSK - newbie questions

[benjar]

Ed,

Your response is useful, but I too am curious about where to find some
documentation regarding the correlator block.

1) I can look at the source code for gr_simple_correlator, and see that it
is a "finite state machine", but I have no idea what the block is doing in
each of the states.  I look at the source code of gr_simple_correlator.cc
and see a few comments related to PN codes, but this is confusing because I
thought that has to do with de-spreading a DSSS signal, not necessarily FSK
demodulation.  A coherent explanation of what this thing is doing,
preferably put in the source code in a way that Doxygen can harvest it would
be very, very nice.  If you or anyone can point me to this, I would be happy
to put it in the source code myself.

2) True or False: The way this block gets used in the examples it looks like
it is making a hard decision about the state of a bit.  In other words,
there would be no way to generate a posterior probability from this block
and do some kind of soft decision decoding.  

3) What I am looking to do for my project is generate a set of posterior
probabilities from a 2-FSK modulated signal.  Do you know of any particular
blocks that do something like this already, or am I going to have to write
my own?

Thanks,

-Ben


Ed Criscuolo wrote:
> 
> Fergus Noble wrote:
>> 
>> Also I have noticed in some of the MSK examples use the quadrature 
>> demodulator block, is this applicable
>> to non-continuous phase FSK and if so is it more robust than the filter 
>> method?
> 
> I have used the quadrature demodulator block on GMSK (which is also
> non-continuous phase) with great success.
> 
>> 
>> Also, the next problem is bit recovery, I don't really understand the 
>> correlator block, could anyone
>> point me towards some documentation that would be grand. Am I supposed 
>> to be using this or the
>> clock recovery block?
> 
> The clock recovery block will resample the data stream at the data
> clock's frequency, yielding a stream of "soft bits". These can then
> be fed to a bit-slicer to produce a stream of binary '1's and '0's.
> 
> @(^.^)@  Ed
> 
> 
> _______________________________________________
> Discuss-gnuradio mailing list
> address@hidden
> http://lists.gnu.org/mailman/listinfo/discuss-gnuradio
> 
> 

-- 
View this message in context: 
http://www.nabble.com/FSK---newbie-questions-tp14117852p14505259.html
Sent from the GnuRadio mailing list archive at Nabble.com.