[cmicali]

Poor quality or really long connections can cause jitter, especially since in most consumer stuff the clock is sent with the signal. That jitter can cause differences in the signal out from the d/a. That said, the difference is less than minuscule, and time and time again people post double blind tests that prove even people who consider themselves trained audiophiles can't tell the difference. Same goes for analog - in practice the cable doesn't make an audible difference, no matter what you use (see below)

Short answer: cables don't matter

The famous audioholics coat hanger test: http://forums.audioholics.com/forums/threads/speakers-when-i...

Also a great watch: http://xiph.org/video/vid1.shtml

[anentropic]

the whole jitter thing is nonsense in the context of something like an Ethernet cable since only the data is being sent, the packets only have to arrive 'fast enough', the timing doesn't matter.

[cmicali]

Yeah sorry - should have said that above applies to digital audio streamed over a wire and decoded directly (spdif, aes, etc)

You can always use a $6k rubidium clock http://www.antelopeaudio.com/en/products/10m-atomic-clock if you really want to make sure your bits are synchronized. It's good enough for wu-tang!

For buffered data being sent and decoded, like via ethernet, the bits are the same on both ends, so I'm not sure what effect you would even make up to say it could change the output.