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

[Dan Gisselquist]

Chuck,

Now I follow what you are doing ... thanks for the web link--it really  
helped explain things for me.

You're trying to pull off a single sideband from a complex sampled  
input.  (At least, that's what the phasing method of image rejection in  
the web-link discusses ...)  May I suggest an alternate approach?  Just  
use a complex filter:

Complex samples -> Complex filter -> Single sideband

Start with a lowpass filter that passes the bandwidth of your signal.   
Shift this filter up in frequency by multiplying it by a complex  
exponential at half your bandwidth ...

e^(j 2pi (BW/2) n)

If your original filter was real, like any good lowpass filter, the  
output of this operation will be a complex filter.  Apply this complex  
filter to your input and you'll have the output you are looking for.

How's this sound?

Dan

On 11/24/2004 03:53:14 PM, Matt Ettus wrote:
> Quoting cswiger <address@hidden>:
>
> > On Wed, 24 Nov 2004, Dan Gisselquist wrote:
> >
> > > Chuck,
> > >
> > > I think you are comparing apples to oranges.  Using the block
> diagram
> > > you've presented below, a more appropriate way to get complex
> sampling
> > > would be ...
> > >
> >
> > Hi Dan - Well untill I get something working any and all input is
> > welcome ;)   I'm working from figure 13 here:
> >
> > http://members.tripod.com/michaelgellis/mixerscom.html
> >
> >
> > Phasing method of image rejection.
>
>
> Basically, the hilbert transform must be on the real side of the
> mixers.  The
> input to both the delay and the hilbert filter must be the same for  
> it
> to work.
>
> Matt
>
>
> _______________________________________________
> Discuss-gnuradio mailing list
> address@hidden
> http://lists.gnu.org/mailman/listinfo/discuss-gnuradio