Originally Posted by Ratman
OTOH... can you explain how it does apply?
Maxwell's equations; more specifically Ampère's law and Faraday's law. The qualitative explanation is that an incident electromagnetic field (say, a RF plane wave) induces eddy currents in a solid, closed conducting surface (the shielding). The eddy currents in turn cause a temporary charge redistribution and an opposing magnetic field that cancel the incident electromagnetic field. The net result is that a solid, conductive, closed surface will mostly reflect plane waves up to a certain frequency (limited by properties of most "real" conductors like resistance). You know the Faraday cage to screen static electric fields and attenuate RF radiation, this is the same thing.
And... how that recommendation would be accomplished.
The basic RF shielding effect is accomplished by a coaxial cable. However, you rarely see coaxial cable used for speaker wiring for a few reasons. First, it's not frequently made with the necessary characteristic impedance (basically the square root of the inductance divided by the capacitance)--which must match the source and load impedances for maximum power transfer--and it would probably be pretty expensive to do so. If the characteristic impedance doesn't match the source and load impedance, the signal will reflect at the ends and the cable will dissipate power through a variety of mechanisms (so you can't just use your 75 ohm coax CATV cable to hook up a speaker!). Second, and more importantly, as I said before, the large amplitude signal used to drive a speaker isn't going to audibly degrade with the addition of some microvolt-amplitude RF noise. Third, if you really need EM wave shielding, some electrical conduit will be a lot cheaper than coax cable.
To go a step further and separate the signal and shield grounds, you would usually use something like a triaxial cable, which is pretty rare even in the scientific world where it is sometimes needed. This shields the signal reference conductor with another conductor that is usually connected to chassis ground. We use triax cables for low-noise measurements (femto-, pico-, nanoampere levels) where even those eddy currents induced by RF waves in the outer conductor (the signal reference in a normal coax) can cause measurable noise (we also built a Faraday cage room with aluminum foil
Sorry for any confusion, I didn't emphasize enough that separating signal and shield ground is something that is only necessary when you're dealing with very small signals. My comment was just a random musing to indicate that having a separate shield isn't totally snake oil. Speco2003 is correct that it isn't a concern for speakers, I was just wondering if he had an additional comment since he seemed to skip over my brief original statement that explained why.