I asked earlier if Wireworld has an explanation. I looked and they don't have any that are convincing to me. They do have a page showing what they say is waveform error for generic USB vs theirs.
http://wireworldcable.com/Image_Library.html Look at the January 2013 white paper under "product sheets".
Here they say they played a music waveform through one cable and then differenced it against their USB cabled. I am assuming they recorded the result and subtracted one waveform from another. Just happens I spent some time doing just that with USB, and analog cables a bit over a year ago. I still have the files. Here is a screenshot of my result.
One file was a 96 khz recording of a Clark Terry song using a friend's Wireworld USB (was either Ultraviolet or Starlight didn't record which). Then again using a 33 foot monoprice USB extender the end of which was using a generic USB cable that came in an HP printer. I even changed the scale to match what Wireworld has best I could tell. They appeared to have it at 25%. Even if I zoomed in like their second picture it looks like a flat line. Zoomed in 5 times closer it looks a bit fuzzy. This isn't a zero. Doing the FFT on it, there is a peak of -107 db at 60 hz. Everywhere else is around -115db to 118 db. That is the basic noise level of the equipment on hand. Later with quieter equipment got the same result between USB cables just the noise floor was lower by 8-10 decibels. I didn't have the Wireworld on hand at that later time.
As there appears to be no difference greater than -100 db I dont' know how you could hear that. I also don't know the details of why they show a different result. I had only good consumer level equipment yet get far better performance than they are showing.
Edited by esldude - 7/11/13 at 9:32pm