Calvin, come on, we are talking few hundred kilobits per second, probably at TTL levels, over a coax! If the cable is very long, then capacitance could become a problem, much sooner than attenuation or noise. As long as the signal looks decent enough that clocks can be kept in sync and noise is a couple of dB below the signal, you will get exactly the same result as with the highest quality optical fiber (or some hypotetical error-correcting protocol).

As for $8.000 cables: whoever has read a single book on experimental perceptive psychology (or has any experience in conducting experiments or observations with a subjective component in them) understands that any test similar to comparison of high-end Hi-Fi gear is absolutely useless unless it is double-blind. Mere A/B is not enough. Hi-Fi magazine writers who harangue against objective testing are either morons or frauds.

What is often neglected is quality of source. Analog vinyl record production techniques had reached very high level in their last years. CDs are (still) often slapped together by less than competent audio-engineers and post-producers (because of deceiving ease with which 'anybody' can tweak a recording using even a $500 PC program). Introduce, for example, a bit too much of phase distortion, and sound image or 'presence' is gone. Even an abrupt piece of 'digital silence' before and after 'concert hall silence' between two movements can be enough to ruin the experience. None of this is inherent problem of CDs.

(It is bad that it is very difficult to conduct double-blind tests between vinyl and CD, as vinyl is always easy to tell by lower dynamic range, clicks and hiss )

BTW, I have read here that empeg sounds 'dead', but also 'over-analytical' and 'harsh'. To me that sounds contradictory, but that could be due to terminology. Care to elaborate?