[Discuss-gnuradio] Re: ssb phasing demod revisited

[Alberto di Bene]

Talking of SSB demod, I have used with success an alternative technique
that doesn't make use of the Hilbert transform, and which works best if
your signal is already in the frequency domain in complex form (I/Q),
at baseband.

Remembering that the first half of the complex spectrum describes the
positive frequencies (i.e.the USB) and the second half the negative
frequencies (i.e. the LSB), then you simply discard the unwanted half,
if what you are interested in is the LSB then you mirror it, then you do
an inverse *real* FFT, and what you get are nice time domain samples,
in real format, ready to be sent to the DAC.

There are also performance advantages if your Hilbert filter is longer
than, let's say, abt. 40 - 50 taps.

73  Alberto  I2PHD