Originally Posted by frankwp
To directly address your above statement; a coax cable will carry a square wave or a sine wave, or any other wave pattern. It doesn't know any better, just carries whatever it's given.
True, it will carry "whatever it's given" so long as "whatever" it is can be transmitted through a cable having its given characteristics. If it can't, the signal will be deformed, possibly enough such that it can't be received correctly. Obviously this becomes a much bigger issue when something on the rx side is trying to recover clock and figure out what's a digital 1 vs. 0 than something that just takes the input and amplifies it (e.g., a S-P/DIF link vs. an RCA connected to a power-amp).
That said, the whole thing is a "system" -- the transmitter, receiver, and cable characteristics all have to "align". The higher frequency signal that needs to be transmitted the more "rigorously" the system must be controlled (which is why, e.g., Ethernet cables have "rigid" specs that analog audio cables really don't need). For analog audio use, our ears govern how accurate this system needs to be (and, ours ears aren't the most accurate thing in the world). For digital, it's a bit more cut-and-dry because either it "works" or it "doesn't" (and "doesn't" is typically quite obvious, especially if there's robust error checking). However, getting it "to work" takes more effort and requires more "control" (e.g., if you don't use a 75ohm cable for S-P/DIF then it's more likely you won't be able to recover the signal and it "won't work" because you'll hear silence, lol).
Finally, I do wonder if this used to be a larger issue in the analog world. Is it not true that tube circuits are more impacted by impedance variance? Did this require "impedance matching" (including cable) beyond what we need today? Because, now-a-days, it's pretty common, as said in this thread, for the source device to have a low driving impedance and the receiving device to have a much higher load impedance; under these circumstances it's pretty hard to find a cable that can't transmit analog audio up to 20kHz with inaudible losses. This is especially so over the short distances typical of a home environment. Even more so when you figure in the potential for high frequency hearing loss among the folks most likely able to afford exotic cables, lol.