I would like to point out a few very real and very measurable factors that under certain circumstances influence perceived sonic performance. The discussion is limited to signal cables, not speaker cables.
1. Shield performance part 1 - resistance.
Voltage imposed across the resistive portion of the impedance of the shield is superimposed on the signal carried by the cable.
In coaxial cables, shield and conductors are magnetically coupled. This means that a coaxial cable will act as a common mode filter with an inductance equal to the inductance of a piece of wire with the same external geometry as the coax. When a cable is used to connect two pieces of equipment, current will flow through the shield in order to equalise the ground potentials. As a result, a voltage is imposed on the shield.
The coupling between the shield and the conductor then becomes crucial, because it insures the same voltage is also imposed on the conductor such that the voltage differential between shield and conductor at either end is the same.
There is a limit to this, however, and that is the resistance of the shield. The impedance of the shield consists of its inductance (not to be confused with the inductance of the transmission line which is much smaller) and its resistance. Thus part of the voltage across the cable is in the inductance and another part is in the resistance. The voltage across the inductive portion is coupled to the signal conductor and thus eliminated from the output voltage. The voltage dropped across the resistance, however, is not coupled. This voltage is found at the output of the cable in addition to the wanted audio signal.
Many audio cables are built with two wires and a shield, with only the wires connecting to the receiving end. The same argument holds for these cables, where the ground conductor is substituted for the shield.
At low frequencies, the resistive component dominates. Over the entire audio band and under realistic conditions (leakage current is sourced by a high impedance), the coax (or twisted-pair in an unbalanced setting) cables are quite unable to carry an audio signal unscathed. For RF applications, coax is fine because the shield inductance produces an impedance which is very high compared to that of the earth current source.
2. Shield performance part 2 - optical coverage.
Coax cables should have a very closely braided shield. Any visible holes will allow capacitive coupling between something outside the shield and the conductor inside. In the case of balanced connections the signal may be presumed to affect both wires equally, thus being readily removed by the differential input circuit on the receiving end. Still, also balanced cables stand to benefit from solid shielding when exposed to large field gradients (ie. passing very close to a source of electromagnetical interference)
3. Microphonics - Triboelectric effect.
All cables are microphonic. Some more so than others. An unterminated cable is liable to produce tens of millivolts into a high-impedance input when tapped. Several techniques exist to reduce triboelectric effects, ranging from adding cotton padding in the cable over making part of the insulation conductive to simply adding exterior padding. High-end audio cable manufacturers prefer the latter because it makes the cables look bulky.
Providing a low impedance output is the most effective, but tests show that we're talking less than 10 ohms before the effect becomes to small to measure.
4. Microphonics - Condenser mic effect.
The capacitance of the cable varies under mechanical stress. If there is a voltage on the cable and the source impedance is high, varying the cable capacitance produces an error proportional to the signal voltage and the mechanical excitation. Again, competently low impedance is required before these effects can be declared negligible.
Both microphony mechanisms are more serious than you'd think. Some cables are so microphonic that when they are connected to a microphone input and the volume is turned up the whole thing starts howling. No joke. I've demonstrated this several times.
5. Frequency response of digital cables
All digital audio cables pass the bits correctly. What we hear is unfortunately not just the data, but also the timing. The same signal encodes both the audio data and the clock. The clock is recovered from the edges of the data. If the cable has an insufficient frequency response, the impulse response becomes longer than the duration of the symbols (bits). The recovered clock becomes dependent on the data (intersymbol interference). These effects are exacerbated by low-frequency limitations on the digital input and output circuits (baseline shift).
Furthermore, if the cable is not impedance-matched, reflections will constitute an even longer impulse response so you might get intersymbol interference from bits that are separated by several positions if the cable is long enough.
By far the easiest way of tackling this problem is to distribute clock and data separately, as is standard practice in recording studios where all equipment is synchronised to a house clock.
If that is not possible, use real 75 ohm coax and design the SPDIF interface circuits to have the widest bandwidth possible on either end of the spectrum.
A more complicated but technically more satisfying way is to make the converter impervious to clock jitter. It takes a PLL with a corner frequency well below 1Hz (and therefore a low-jitter VCO!) to make the effects of incoming jitter truly inaudible. Few designers feel up to the task and turn to asynchronous SRC's which go quite far along the way as well.
I hope this short outlook shows that in proper science, the notion that cables don't make a difference is a huge oversimplification.
What can be said is that cables shouldn't make a difference but that is an entirely different ball game. Very, very little equipment is that well designed.
Want to enjoy fewer advertisements and more features? Click here to become a Hardware Analysis registered user.
It is exceedingly easy to find out who I work for. Google is an efficient tool for such purposes. That you have not taken out the few minutes necessary to do so before posting a sneer is as telling as the fact that you have not taken out time to verify for yourself the very real and astoundingly easily measurable cable properties I point out.
The political alliance you are suggesting is one which I would rather expect from people as shortsighted as you show yourself to be.
What I find so interesting here is that out of the blue, two people make a post within 5 minutes of eachother, to attack [the author of] a technical expose of electrical behaviour of cable which was posted months ago, while completely failing to make any discernable point themselves. Please state what it is you find to be in error, along with a proper rebuttal based on physics. Pending that, I consider the discussion closed.
Bruno, go easy on these guys, they don't have the education to fully grasp the contents of your post, let alone be able to provide you with a bulletproof rebuttal. Much appreciated you took the time to slap these boys around a bit, I'm sure one day they'll realize their comments were ill addressed and lacked any real substance.