Re: [Discuss-gnuradio] SSB demodulation

[cswiger]

On Wed, 8 Sep 2004, Johnathan Corgan wrote:

>
> My point was that your algorithm for SSB demodulation is incorrect--it's
> not simply a matter of adding (or subtracting) the real and imaginary
> portions of the complex symbol stream.  The math is a bit complicated,
> but in the end you have apply a 90 deg. phase delay to the imaginary
> portion first, via what's called a Hilbert transform.
>
> It sounds like you think the frequency translation already provides the
> 90 deg. phase shift--there is one involved there but it is something
> completely unrelated.
>

Oh I see now - I was going by this block diagram:

http://thierry.leconte.chez.tiscali.fr/demod.html

and missed the part at the bottom that says, "The Q filter is the same
bandpass filter + a 90 phase shifter (Hilbert filter)"

Ok, that makes sense now - so everything on one side of the carrier
adds and the other side cancels. Gotcha.

--Chuck