I too would love to understand the science behind the sonic impact. However, that's not to say I dis-believe it's possible without a proven explanation. That's because I have an asynchronous-USB DAC (Hegel HD30) and use USB to connect my Aurender N100 (the only digital output that music server provides actually). Still points under the Aurender made the DAC sound different. Changing USB cables made the DAC sound different... in ways that made me want, or not want, to listen to the music. I auditioned about 5 USB cables... everything from the Curious cable to Wireworld Platinum to Ansuz Diamond (@ $2K) to the Inakustik. All cables resulted in an appreciably different sound. It wasn't placebo or wanting to hear differences... rather the opposite... I was dumb-stuck and frustrated that an "asynchronous DAC that has it's own master clock and pulls data packets" could, in any way, sound different based on data transmission (or by whatever could account for stillpoints under the server). Lucky for me, the most affordable cable... the Inakustik, in my system, was the clear winner, and also the most affordable of every audiophile-cable I tried.
The bottom line is there appears to be more to "the sound" of USB protocol... even when the DAC owns the master clock. Whether it's noise leakage from the source to DAC via the cable, or whether it's something else that somehow affects the DAC's d/a process or output stage... or leaks along cables down to the preamp or amp... who knows. But it's there.
I'd love for good science to find out what's happening... not so audiophiles can stop arguing whether or not it's possible for them to hear what they are hearing, but rather so audio engineers can properly design their gear to be immune or less impacted by understood variables.
