As for the topic, I would hesitate to think this is total crap from a physical standpoint. The idea is to put away with a low-level crossover-distortion like error. The electrical field has a very strong superimposed DC component which forces the charges in the isolator to "lock-in" on fixed positions/orientations rather than "moving around" and dissipate energy, slightly deforming a low level waveform when fed from a high impedance source. With low source impedance and high drive (volts) it's probably way down south of -120dB and irrelevant. In a phono cable though, error magnitudes may or may not be different and above hearing threshold. To get a grip on this one would need a more complex test setup, using a high-Z source at microvolt levels and a low-noise preamp, then some prost-processing on the ADC data with eg. time-tomain averaging, using the same cable with DC bias on and off in a subtractive test, maybe with giving it time to settle and/or "degauss" the cable with a high-level noise signal and all that. All in all a lot of effort with questionable outcome.
The good news is that microphones with +48V phantom power are using DC bias all the time, so if there is something to it, at least our recordings are not totally spoiled to begin with...