I recently changed my speaker cable setup and was very surprised at
the difference in the sound. Basically I moved from having
entry-level Van-Den-Hull skyline cables single-wired to my speakers to
bi-wiring with the Skylines on the woofers and Kimber-Kables 4TC for
the tweeters. I was rather astounded at the difference in the sound.
To be honest I was not expecting much change, if any and only did the
upgrade because of a very good deal on the 4TC. Instead I am now very
convinced that cabling considerations may be, although not the most
important aspect, essential to any system and not something to be
ignored.

Here is my pseudo-scientific explaniation.

A few years ago I had a summer job of network support for a small
office. At the time There were about 40 computer workstations with
about 4 printers. One of my jobs was to set it up so that all of the
workstations would have access to a printer. This required the
purchasing of a lot of serial and parallel cables to attatch computers
to printer-sharing boxes.

The use of high quality cabling was ABSOLUTLY essential for this
purpose. Quite simply, using generic cables more than about 2m was
impossible for parallel printer cables. There was to much noise and
signal loss for the succesful and reliable transmission of data.
However, by using much more expensive shielded low capacitance cables
runs could be as long as 10m or even 15m.

Although serial cables (much lower speed signals) were not as finicky,
there were still issues with long runs with cheap cabling. Often I
had to reduce the data speed from 128kb/s to 64kb/s

What does all of this have to do with audio cables you may ask.
Everything. Cables essentially serve as carriers of data.

Assuming as a baseline the bandwidth of a cd (44000*16bits per channel
= 704 kb/s) the amount of data being transmitted is very considerable.
if one sample has a value of 5024 then that exact value must be what
is found at the other end of the interconnect/cable. If the value
recorded at the other end is 5024 +/- 1 then we are losing almost 2
bits of resolution on the sound or a drop in signal to noise of 12db.
Since the CD specification came from studies of human hearing this
loss of two bits is definetly noticable.

Transmitting 704kb/s over a serial connection would be quite a trick.
It can be done but only if care is taken into the way that the signal
is transmitted. Obviously this represents a much lower data rate than
ethernet connection but we must also consider that the type of signal
is very different. Speaker cables and interconnects can't use
modulation of high frequency signels(measured in Mhz and gHz), Rather
the method of signal tranmission is more like amplitude modulation
with relatively low (measured in kHz) to extremely low frequencies
(measured in Hz), where microvolts may represent the difference
between 16 bits and 14 bits of data being transmitted.

With speaker cables the difference is even more profound. Not only
must the source generate this signals with considerable voltage and
current, but the speaker's themselves the recipients are also acting
as sources of noise being transmitted back down the line. I would
suggest that there is a significant amout of data distortion present
even when using the finest amplifiers with excellent short cables.

As I said I only begain thinking about this after noticing a huge
difference to the sound of my system after changing the speaker
cables. Now I am just tring to come up with the theory to explain my
observations. Since I am somewhat familiar with the limitations of
cabling in other applications (computer data transmission) I am trying
to look at the mystery of the audable differences of cables from a
data point of view.

My System

Source:Cambridge Audio D300SE
PAdcom GFP-555 MKII
Power:Bryston 3BST
Speakers:Energy Exl-26
Interconnects:Van-Den-Hull D300SE MKIII
Cables Top : Kimber 4TC
Cables Bottom : VanDenHull Skyline