I got into this with an audiophule type a few years ago. He, with a completely straight face, asserted that double-blind testing was an inherently flawed methodology for evaluating the objective marvelosity of some silly audiophule crap he was touting.
This has been a consistent argument from audiophiles for several years:
In short, for the important stuff, like "Do amplifiers or cables or differing storage media sound different", "blind testing" of any kind, single or double isn't likely to work because there are simply too many characteristics present and changing, and (if only because of the way human perception works) it's virtually impossible to isolate them and make sure that all of the testees are hearing the same test of the same thing in exactly the same way.
In short, "we believe in high fidelity but only in a purely non-falsifiable experiential sense." You can talk all you want about your error rates and THD+Ns but all they want to hear about is the "clarity," the "smoothness of the tone," and the "space around the instruments."
And these aren't crazy things to talk about, but insisting that a physical thing, a $300 ethernet cable, can actually create these things in a way that a cheaper one cannot is a kind of fetishism.