Re: [Discuss-gnuradio] Weaver SSB (was: gr.firdes_hilbert () usage)

[Matt Ettus]

Quoting Eric Blossom <address@hidden>:

> On Fri, Nov 26, 2004 at 06:38:50PM -0500, cswiger wrote:
> > On Fri, 26 Nov 2004, Matt Ettus wrote:
> >
> > >
> > > That should work.  But you don't need to have 2 filters on each side.
> You could
> > > combine lpf1i with lpf2i.  Same for the q side.
> > >
> >
> > Well I tried that at first, but at the rf sampling rate (8e6),
> > 1.8Khz cutoff, and 300hz width the system would  lock up
> > after taking just under 65536 bytes. No errors, just stops.
> > So I decided to break the filters into two easy chunks, one
> > at rf sampling decimates down to audio level with easy taps
> > and the other does the sharp 1.8Khz slicing at af sampling rate.
>
> Chuck,
>
> Can you send me the python code that caused it to lock up?
> I'll see what I can figure out.


My guess is that a 300 Hz transition BW at an 8MHz sample rate caused there to
be a filter at least several thousand taps long.  That would make it appear to
lock up.  It makes sense to do the decimation in multiple steps so that there
don't need to be multiple stages.

matt