Can you confirm one thing ... you appear to be saying that, because Hugh has done such a great job with the power supply, using a Felicia actually doesn't do much for a GK-1/Swift at all? Or have I got this wrong? 
Not exactly. My point was that the benefits of the Felicia were not as great with the Swift/GK-1 (as best as my aural memory serves me) as with the Modwright 9.0 SE, a very highly praised tube pre. I think this is due to the care and expertise Hugh brought to bear in the implementation of the powersupply design.
Specifically, Hugh chose custom, purpose designed, shielded EI core, dual mono transformers, as opposed to the Modwright's toroid. The benefits are the limited capacitive coupling of high frequencies (in technical jargon, crap) offered by his shielded EI transformers, as well as the far narrower bandwith of the transformers, in comparison to the typically interwinding unshielded, overwound toroid. Such toroids actually make resonably good audio transformers. So Hugh's design choices, although calling for far more expensive components, address much of the noise coupling concerns (at least in mains polluted NYC) that are also addressed by Felicia. The fact that he did this, and also managed a component with no discernable hum, humbles me. The major advantage of toroids (and almost a necessity in poweramps) is that they have limited emitted magnetic fields, which minimize inducement of hum.
But this excellent powersupply design does not obviate the benefits of balanced power. Symmetrical power yields symmetrical cancelation of capacitively coupled noise -
http://www.equitech.com/articles/articles.htmland while I do realize that Equitech has a vested interest in promoting their own products, the benefits have been applied in high precision instrumentation since the 60's, originally at Bell Labs.
To be honest, I don't percieve any overt noise or hum through the Swift when fed via a single rail AC supply. Nor do I have the requisite instrumentation needed to measure it. Regardless, those benefits of balanced power are immediately obvious when you hear it. I've not done this with the Swift yet, but I have configured Felicia with other components with a switch that allowed me to lift the center tap ground, and compare true balanced power to isolated (sometimes incorrectly called floating balanced) power, and those differences are quite apparent, with true balanced power providing substantial improvements. And at some point (when I tire of the joy of listening to the Swift/N+ synergy) I'll get around to it.
I'm tired of the anecdotal comments on the downside of balanced power. Certainly, for amplifiers sourcing the transformer(s) that will not saturate on peaks for low cost is problematic. But this is not true of low power source components.
Please realize that the Felicia was based on the availability of very low cost, surplus transformers here in the States. This is not as applicable outside the States (and actually, those specific transformers have become unobtainium). It was also designed to feed components with noisey switching suppies or toroidal fed linear suppies. Given that the Swift already has very excellent dual mono transformers, with minimal noise coupling, an alternative would be feeding it with a single split bobbin transformer having (outside North America) a 220-240v center tapped secondary and empirically determining the best betwixt and between capacitance of proper voltage ratings.
FWIW,
Paul
