Can This Really Matter?

A bad (not to spec, defective) cable can cause trouble even in digital paths, but a solid cable will be no different from a really expensive one.  Bits are, in fact bits unless there is a serious defect (impedance, shielding, etc.)

Now, before people pop, there ARE issues int he digital path.  At the DAC chip itself (not the box) you must have timing. I explain this int he sonogyresearch blog -- but that only applies to SPDIF, not USB which is asynchronous.  A much bigger issue is ground and power noise, which is why sources, LPSs etc can affect sound.  I power all my sources with a custom low noise LPS, (linear power supply) and wherever possible I also use a bridge (e.g.: dedicated ethernet --> USB box) to further isolate the DAC from the noise of a computer. All streamers are computers, although they may be embedded in a fancy box.

You would not believe how much testing i have done on these various issues, both as me and through the company.  Many things surprise me and do matter.  Digital cables, past the point of engineering correctness, are not one of them.

Beware the placebo effect. One thing i can say, for sure, is every system sounds better with a glass of good wine.  really. And that tells you a a lot.  The wine (etc) variable must be accounted for in real listening if apples are to be compared to apples, which, when in this world, is my business.

to sgreg1 and mapman

1. Yes digital noise is the enemy - actually ANALOG noise on digital paths, created by high frequency digital circuits and switching power supplies. Cables will do little to fix this. Filters and isolation components like bridges do a lot.

2. The reconstruction of the 1s and 0s into PDM or PAM, and the reconstruction of that into a continuous waveform, is, of course, the entire issue.The question posed is quite different: do USB cables in particular, which carry an asynchronous stream, of 1s and 0s, have much effect? As noted above they can't really block much noise unless they are far more than cables. And unless they really suck, they cannot create bit errors.

oh, i should add to the above to finish the logic...and the digitally generated noise impacts a) the reconstruction process and b) the analog stage during and after reconstruction.  Some posit that it can affect the quantized PAM/PDM --> analog, by affecting zero crossings, but i think that is unlikely given how well proven digital transmission intergaity is (witness: gigabyte internet into your house...amazing)