I would volunteer the following principles
(a) all engineering is a compromise ( I can do this for $10K but with $20K more I could reduce distortion another 0.001%)
(b) specs only measure the state of our knowledge, not what we are capable of hearing
This discussion is interesting to me because its the direction I need to go next. Being physics trained, when dealing with unknown equations, one tries to start with the most important variables first (1st order), establish the parameters, then move on to the next subtler variables (2nd order) establish those and so on.
All electrical equipment depends on power which makes it a first order variable... any imperfection in power will feed instability in all attached equipment. And you can easily demonstrate the interaction between devices attached to power in your own home.... flip on a heater or vacumm cleaner and you will see lights on the same circuit react during the initial power surge. And thats just what you can see, not the complexity induced by flourescent lights, refrigerators, air conditioners, and hair dryers.
So I believe it sound judgement to start first with a high quality power filter/voltage clamp. Interconnects on the other hand have differences but those differences become a complex hard to predict interplay connecting two dissimiliar electonic devices.. they can change the sound but how is as hard to predict as finding a comfortable shoe.
I find little merit in the discussion about power cords. I suspect the actual issue here is the end interface... loose connectivity or poor choice of metal contact at plug and equipment end. You could probably do better taking a stock copper cable at a home supply store and soldering a high quality temination at both ends. The power cord has the simplest task of all... feed 60hz at 120 volts with no capactive or inductive interference to a transformer. A power cord that addresses this at around $50 is a convenience factor but spending more on a passive device I find hard to justify.
