The EMI filters are just a reason to want 30 KHz rather than 300 KHz. I've been doing a lot of probing with the SA function of the Instek MDO-2000E line. My 8560A just isn't sensitive enough below about 300 KHz which is where the majority of the noise is. Sadly I discovered that GW botched the calculation of the modulus so the series resonant notch of a crystal shows up as a peak.
So how would you evaluate the performance of an EMI filter? It surely looks like a transmission/reflection problem to me. I'm a retired reflection seismologist, so I'm very used to the time domain and impulse responses and to a large degree can look at either a time or a frequency domain plot and cartoon the other domain (thanks to Ronald Bracewell).
For that matter, how would you evaluate a crossover network? That's not 50 ohms. I would expect one would need a balun at each end, but I can't see why it matters that it's not an RF line. The conversion between balanced and unbalanced and the impedance matched would be required and the latter probably gets "interesting" if it's a 4 ohm speaker.
