Frequencies addresses by power cords?

Ralph's comment isn't referring to distortion in the incoming AC voltage. He's referring to the current drawn by the component, which can (especially in the case of power amplifiers) consist of spikes of high amplitude lasting for just a small fraction of each 60 cycle period, with little or no current being drawn during the rest of the cycle. The brief duration of the spikes means that their frequency content extends to frequencies that are much higher than 60 Hz, and the bandwidth of the power cord has to be high enough to avoid significantly attenuating of any of that content.

So there can be a tradeoff between maximizing noise reduction and satisfying very rapid changes in demands by the component for current. I have no particular quantitative feel for what kinds of bandwidth numbers tend to be optimal with respect to that tradeoff (which btw will also be affected by the house wiring).

Regards,
-- Al

Well Al, what does this mean for power regenerators?

I don't think that any generally applicable conclusions can be drawn about power regenerators. From both a technical and an anecdotal standpoint their efficacy figures to be dependent upon their own design, the designs of the components they are powering, and the characteristics of the incoming AC.

Obviously regenerated AC that is essentially free of both noise and distortion figures to either be helpful or to not make a perceptible difference, depending unpredictably on the designs of the particular components and on the characteristics of the incoming AC. But whatever limitations a particular regenerator may have in terms of its responsiveness to abrupt changes in demand for current may or may not outweigh that benefit, depending on the designs of the components and the regenerator. And adding to all of that uncertainty is the possibility that the potential downsides of either using a regenerator or not using a regenerator, if present to some degree, may be subjectively preferable to some listeners with some systems.

So it's all essentially unpredictable as I see it.

Regards,
-- Al

Safe comment - unpredictable :-).

MIT and Transparent both use "Interface Boxes" which are nothing but "glorified" low pass filters that filter most signals above 20KHz.