"Been reading this thread and while I don't want to rain on anyone's parade, I thought I'd correct a factual error I've seen mentioned a few times. With audio interconnects we are talking about unbalanced circuits, meaning a single signal traveling one direction and a ground. Twisted pairs will not reduce noise in unbalanced circuits." No cable will reduce noise unless it is faulty.
"On the other hand, when noise hits balanced circuits, which have a positive and negative signal, the twisted wires will enable the noise to cancel itself out." the twisted wires self shield partially and theoretically pick up the SAME amount of common mode noise which the balanced front end 'cancels' out.
I have found that for my line level connections (not turntables), unshielded cable works fine.
"But that just means I don't have EMI invading the cable--the twisting accomplishes nothing noisewise" This is plain wrong. You may have heaps of EMI up in the meg range overloading wide bandwidth preamps, upsetting amps and upsetting feedback loops. Twisting an unbalanced cable helps reduce massively the RF suseptibilty of a cable and associated equipment.
The ONLY reason you get away with untwisted unshielded cables is because the output impedance of your source is low enough to prevent truckloads of noise getting in. It is a BAD, BAD practice and has no basis whatsoever in good enginnering practice and certainly does not lead to a silent system.