Gee.......how did I ever miss this one? I gotta pay more attention to what goes on around here.
OK.....briefly.........the impedance of the cable, transport and receiver have to be matched. In the case of SPDIF, that means 75 ohms. 75 ohms means BNC, and not RCAs. (75 ohm RCAs can not exist, the laws of physics say so.)
Anyway......if there is ANY impedance mismatch, part of the signal will bounce back and forth. (Reflections......in nerd-speak. ) The reflected signal can cause disturbances in the clock recovery. (Don't ask me to 'splain why......it is just the nature of a poorly designed protocol.) When you have a system where all 3 components are matched, impedance-wise, then you should not be able to hear any differences when you futz around trying different combinations.
Sadly, most stuff is not even close to being 75 ohms. Not cables, not receivers, and transports are usually far off the mark. Transports are ususally designed to produce the least amount of EMI, and that spells disaster for SPDIF.
Receivers usually suffer from input stages that use RS422 inputs, that have lots of hysterisis. Which is a form of regenerative feedback......which means lots of reflections from reactive elements being coupled back into the input, blah, blah....
(A very famous maker of digital gear once designed a DAC that had a 5-pole filter on the input. Any wonder why it sounded like crud? And to think that they actually offered me a job once, to help fix all their problems. How dumb did they think I am?)
As for cables.......well......RCA jacks and mystery coax...or worse.........twisted pair.......make getting an impedance match darn near impossible.
Which gives rise to an entire industry of after-market kludges, designed to move money from your wallet and into someone else's.
(Actually....a buddy of mine used to make probably the only one that ever worked as claimed. Only to see it skewered on some DIY website. My, my.......)
Did that answer the question?
