A Bit On Star Quad Wire
An essay on wiring. Quite long, and so under the cut.
There're many wiring layouts, but the most interesting one for the purpose of headphone and interconnect wiring is star-quad.
First of all, for the unitiated, wiring material and scheme can affect sound greatly.
Contrary to what CD audio apologets try to push, humans do not listen to certain frequency ranges. In fact, what CD audio pushers imply is that humans listen to certain noises; but there's a great difference between hearing and listening, and what is perceived as noise.
Hearing is acknowledgement of a sound; a sound can be perceived as noise if it does not have a certain combination of overtones (even harmonics) that distinguishes music and pleasant, "warm", lush, full sounds. An even harmonic (an overtone) is a resonance at a certain frequency. So a first even harmonic to a fundamental tone of 60 Hz is 120 Hz, second at 240, etc. Progression's Fundamental*2, F*4, F*6, etc. An odd (dissonant) harmonic is a harmonic created by odd repetitions, e. g. 60 fundamental, 180 second harmonic, etc. (F*3, F*5, F*7...). A note played by a musical instrument will contain many harmonics, each perceived as part of a major 3D sphere of sound affecting the listener. In reality, the frequency range (and, more importantly, the whole bouquet of harmonics) of a musical instrument that's registrable by existing electronics is very high, in the hundreds of KHz.
What humans listen to though, is not a determined frequency range; but echoes. The whole combined, warm, lush, huge, intermodulating bouquet of echoes that arrives to one's ears is what sounds like music. Human mind is extremely sensitive to echo continuity and sound wave contour. A slightest distortion, stray odd harmonics caused by quantising during conversion to a lower-resolution digital format and ringer in the high-frequency harmonics are perceived as screeching and "hollowness" in music. This is how the mind separates music from noise: music is perceived as having a whole body if its reverberation impulse is untainted, especially in the 10-20 KHz range where spatial cues are listened for.
Human listening is echoic: humans listen to echoes and phase intensity. Harmonics and their shapes inside the whole "bouquet". Hence the ability to transmit harmonics is crucial to a recording format.
And there's the snag with CD audio... Low-resolution 44100 Hz sampling at 16-bit does two fundamental perversions: first, it deforms harmonics in the critical 10-20 KHz range (spatiality), second, it imposes a restriction on available frequency bands (only 4 samples/cycle at 11025 Hz, both negative and positive, but more importantly, audio is constrained to just a few bands of defined frequencies). This is also why lossy MP3 compression is not perceived as artificial by many listeners: it actually discards noise in the 11025+ Hz range, making CD-audio-sourced material more tolerable than original 44/16 audio (though of course killing a good part of whichever space definition there was left). Consumer audio electronics aren't very precise or sensitive in the 11-22 KHz range either. Most consumer system speakers (or headphones) can't even reproduce anything over 10-11 KHz reliably. Hence MP3-codec-lowpassed and -masking-rejected audio can introduce less distortion in consumer audio gear than CD audio.
In summary, what humans really listen to (as opposed to hearing) is the delicate bouquet of harmonics and reverberations; a slight deformation in the amplitude of a reverberation given away by a musical instrument can make a portion of the sound range appear as noise that will often be discarded by the mental processing. The mind is always on the listen, and will always pick music out of any noise; basically, the human mind is akin to a signal processor looking for the most harmonic in sound, and discarding everything that's perceived as irrelevant to music. A human's conscious perception is often composed of only the "good" harmonic material, not the harmonic distortion (but a quick switch from hi-fi speakers to consumer speakers can reveal all of the barminess that the low-end speakers really have before the mind focuses on discerning the music in the sound).
Which is why cable materials and wire arrangement matter a lot. In audio electronics, signal's transmitted as a fluctuation of electric current, with electricity modulated in analogue to the intensity of original sound. Digital signals, by the way, are also analogue in nature - a digital system of information coding is just an additional layer of logic on top of an analogue electronic system of data transmission. In short, digital signals are usually transmitted by square wave modulation. Hence the precision of analogue electronic circuits modulating the digital logical signal is what matters for any digital data protocol. Hence materials and analogue noise resistance are important.
Two commonly used conductive metals are copper and silver. Silver is superior (a truly noble metal) in every way - it has better conductivity properties, and it does not oxygenate easily. Copper is less prone to oxygenation (rust) than steel or iron, but it does lose conductivity with oxygenation, just like steel (the process of oxygenation is slower for copper). Silver, even when oxygenated, has a very slight loss of conductivity (and it oxygenates more slowly than copper).
A wire, no matter of which material, as long as is conductive, also acts as an antenna, picking up electromagnetic modulation from any nearby source. A power supply, a computer (notebooks are especially powerful electromagnetic radiators, more so than desktop PCs), even an AC extension or socket all have their EM fields which intermodulate with a cable. And well, it wouldn't be such a big problem if it were electromagnetic resonance at just one frequency, but the problem is, even slight noise affects not one frequency, but several - 60-Hz electric hum also rings at 120 Hz, 240, etc. (just listen to a bad ground loop).
The solution is complex: audio (and network) cables usually have more than one layer of isolation. A plastic coating, a section of wire designed to absorb interference, within which is buried the main signal wire enclosed in aluminium foil and its own plastic coat. But even that isn't enough, hence signal balancing and wire twisting.
A balanced signal is transmitted twice (positive and negative). What's picked up by the receiving side is actually not the sum of the original duplicated signal, but the difference. That way any interference that's picked up along the way is rejected (or, rather, ignored). But: this arrangement does require special circuitry and grounding.
Star quad wiring uses four conductors per channel. Two positive, two negative. Conventional twisted pair cable twists conductors to reduce electromagnetic interference, cancelling out noise (and more importantly, phase deformation and hence harmonic distortion). Star quad consists of two twisted pairs of wires; negative+negative and positive+positive. Each pair transmits a duplicate of the original signal with natural EMI phase cancelling (thanks to twisting). The benefit of this scheme is that it's akin to a "natural balanced" circuit when it comes to EMI.
In an unbalanced connection, the phase distortion will still be received by the cable, but it will be minimised and distributed evenly over both channels, and hence easier to remove with processing. Which is why star-quad is a natural choice for microphone connections (even unpowered unbalanced connections).
As for silver as a material for an audio cable, the choice is simple - silver has the best high-frequency latency for electric modulation, and hence the best perceived spatiality. There really is no other (common) material for high-fidelity audio connections.
Hence, silver star quad audio cable has the lowest phase distortion, lowest harmonic distortion, and the best dynamic range possible with a non-balanced connection such as the usual minijack in a battery-powered amplifier or player. This is more important for higher-impedance headphones. Speaker cable can do with less protection as conventional speaker signals are relatively low-impedance (4, 6, 8, 12, 16-ohm) and high-power (1W+). Headphone signals can be high-impedance (32-ohm and higher, all the way up to 1400 ohm and more for older models) and low-power (0.1W is a lot for many headphones, typical sensitivity is around 100 dB SPL/1 mW input power), and are more sensitive to distortion due to EMI and intermodulation.
All of this translates into a more realistic sound, with dynamics and an image of space/dimension/musicality unattainable with regular straight copper cables.