Ultra simple spectrum analyzer question

Hello,
Recently I found this spectrum analyzer project http://www.astromag.co.uk/ssa/
It seems feasible and the resolution is quite high. The only great limitation I see is that the max span is 20KHz or up to 196KHz with very good sound blasters, so you have to advance the local oscillator frequency into 20KHz (or 196KHz) steps to cover larger bandwidths.
What is your oppinion about it?

Also I guess this setup suffers from the image signals, as these will appear at the mixer IF port, am I right?
Does it worth it, to try an I/Q mixer setup with a quadrature output DDS LO, combined with the I/Q receiver function of the spectrum lab software, in order to cancel out the image and lower the noise at the same time. I think this would make a much higher performance SA, what do you think?

I think the 'sound card o/p' in the diagrams should be inputs not outputs but yes, the system should work. You might find it difficult to get a clean enough output from a DDS in such a narrow bandwidth but it would be worth a try.

An alternative would be a dual-conversion system. Taking the RF into a wide band receiver with a low IF and using a tunable LF oscillator to bring the IF into range of the sound card input.

Brian.

Brian, if the direct conversion scheme is used instead of a dual conversion, then an I/Q downconversion scheme (with appropriate software or opamps) could help rejecting the unwanted sideband (KHz away from the carrier, but at the opposite direction), is that right?