Oh for heaven's sake, it's a USB cable. Don't stress too much on this, a decently built $20 cable will give you everything you need. Spending big bux on a USB cable makes no sense at all...
You are right, Dalig, I have no experience with snake-oil cables. But I *do* have almost 30 years as a Ntwk. Admin/Programmer and I know that either the cable passes the bits or it doesn't. Bits is bits, folks, and NO amount of money will cause 2 sets of identical bits to sound different from each other. Any mfgr. who tells you so is, flat-out, a liar.
If you want a great cable without paying stoopid money, check out this AudioQuest for $35 delivered:
Folks who spend uber-bux on USB cables have far too much disposable income...
My opinion on the matter was expressed in Tony's other related recent thread.
A point I would add relates to the fact that both of the cables being asked about in this thread emphasize accurate control of impedance (or more precisely, characteristic impedance) as being one of their key features. The standard for the characteristic impedance of USB interconnections is, I believe, 90 ohms +/- 15%. If in fact impedance accuracy in a USB connection has anything to do with the sonics that are ultimately heard in a given system, best results will occur NOT when the cable has an impedance of 90.000 ohms, but when the impedance of the cable is matched as closely as possible to whatever the input and output impedances of the components it is connecting happen to be, within that +/- 15% tolerance. Signal reflections and waveform distortion (that may contribute to jitter or other sonically relevant effects to some degree with some DACs) occur not due to impedance inaccuracy relative to a standard value, but due to mismatches relative to the actual impedances of the components that are being connected.
I see no reason to expect that the unspecified impedances of the USB interfaces on the Squeezebox Touch and the PW DAC will be perfectly accurate, particularly in the case of the Squeezebox, and therefore no reason (everything else being equal) to expect that a 90.000 ohm cable will perform better than (or even as well as) a cable whose impedance is inaccurate, but happens to more closely match the impedances of those components.
Which is one more reason why, as I said in the other thread, sonic differences between good quality USB cables are conceivable to me, but consistency and predictability of those differences among different systems, different physical setups, and even different lengths of the same cable type, are not.