Home Page | About Me | Home Entertainment | Home Entertainment Blog | Politics | Australian Libertarian Society Blog | Disclosures

Know It All: When Only the Best Will Do -- A Dire Digital Audio Cable

Published in Geare magazine, Issue #57, 2009

In most aspects of my thinking I endeavour to be a sceptic. That includes when I judge claims of science and faith, and also in my area of work: understanding and writing about home entertainment equipment.

My scepticism extends to my own judgements. I am not a computer. I do get things wrong. So I usually check my determinations multiple times.

Naturally my scepticism also extends to claims made for some of the cables used in home entertainment systems. Three things seem to distinguish these cables from regular low cost cables. The first is that the claims are entirely qualitative in their terms. They don't say that the cable impedance is lower, or its capacitance is of a different value. They don't mention anything that can be measured. Instead they invoke concepts of 'clarity' and 'musicality' and so on.

Second, these cables are expensive. They are typically one hundred dollars or more per metre.

The third is that they are bulky, look nice, have attractive plugs on their ends.

I would be inclined to think that they don't make any difference, and in my efforts to actually hear any differences I've been notably unsuccessful.

So why do some people hear differences between cables?

To me there seem to be two reasonable explanations. The first and most obvious is that there are differences, and these listeners have the sufficiently acute hearing and refined listening experience to enable them to detect it. This has some unpleasant implications, though.

To understand this, you need to accept that a cable cannot improve sound. It is a passive device. At its best, an interconnect cable will transport the sound without any degradation at all. But if it did alter the signal in some way, that alteration must be a degradation.

So, next time you are in a hifi shop that offers to sell you cable A for $100, or for even better results, cable B for $200, consider what they are saying. They are saying that they will charge you $100 for a cable which they know to actually damage the signal!

The second explanation is that those who hear marked differences between expensive cables A and B are imagining them. I write that not as a person who is attempting to demean these people, because I myself have imagined such differences. I have pressed a button on a piece of equipment and been surprised by the effect. And then noticed moments later that the circuit I had thought I was engaging had, in fact, not been engaged. In other words, I had imagined the change because I expected to hear a change.

When once, years ago, I reviewed a set of interconnect cables that retailed, I think, for about $400, I switched from cheap ones. Actually, cheap and nasty ones. The kind that are given away with cheap DVD players (except this was before DVD). To my surprise, there was significant additional detail revealed in the music from the CD player. I should have run with that impression.

Unfortunately, as a sceptic I then switched back to the original cheap and nasties, and lo and behold, there was all the additional detail I thought had been revealed by the expensive cables. The difference in the first switch of cables was not them, but my state of attention.

Nevertheless, that interconnects and speaker cables may affect analogue signals is not beyond the realms of possibility.

An analogue signal can be damaged by the smallest influence since at every point in the signal, within the amplitude constraints imposed by the equipment, it is infinitely variable. A point in the signal may be moved by some defect in the system by, say, a decibel. Or by a tenth of a decibel. Or by a thousandth of a decibel, or a millionth.

We can be confident that cables do not move points in the signal by even a hundredth of a decibel, for if they did we would measure the effect with standard instrumentation. Whether or not a move by a thousandth or a millionth of a decibel is audible is, perhaps, what this is all about.

The whole horrid mess
The whole horrid mess
Unfortunately, the subjective approach to this subject has been employed with digital cables. After all, with the advent of the DVD, interconnects have increasingly become digital. First there were optical and coaxial digital (S/PDIF). Now there is HDMI. If I had a nice little business selling $400 analogue interconnect cables, I may be receptive to the thought that a $400 TOSLINK cable will sound better than a $10 one. But that's just me.

But there are major problems with this on every level. One of the major advantages of digital audio is that the signal is robust. Remember, the signal consists of two voltages, one representing a '1' and the other representing a '0'. If a '1' is altered by a thousandth of a decibel, or a hundredth, or even a tenth, it doesn't matter. If the '1' is a one volt signal, then a 0.1 decibel of attenuation will make it a mere 0.989 volts, which the receiving piece of equipment will still take to be a '1'.

Even if there is a misinterpretation, then this will cause an error to be noticed by the equipment receiving the signal since it is all packaged with checksums and the like to ensure data integrity. In this case the error correction circuitry will fix it, generally with the redundacy in the signal with no loss of quality.

Third, if the error is too big even for redundancy-based correction, then the error correction circuitry will have a go at interpolating missing information. Instead of a nicely curved line in the signal, you're likely to get a straight ramp in the signal from one voltage level to another. At its worst, this is nasty and obvious, sounding rather like a click.

Finally, if the error is so huge that there is a gap in valid data for a noticable period, will you get a clearly audible drop-out.

What you won't get -- unless you count clicks, drop outs and total loss of signal to be such -- are differences in a sense of musicality, or the authority of the bass, or the detail in the high frequencies, or any of those other fuzzy subjective terms used to describe high fidelity sound.

But that's just me saying that. So how about an experiment?

Check out the first photo. That is an interconnect. Not the usual kind, to be sure, but then I didn't want it to be.

I reasoned that if a really high quality one metre long digital audio cable costing hundreds of dollars can make a noticable improvement in sound, a really crappy one ought to noticably degrade it.

So I checked my cable cupboard and selected eight of them. I ignored the section where I keep video cables and went straight to the audio cables. Remember, composite video cables are intended to carry a 7MHz bandwidth, which nicely encompasses any S/PDIF-style digital audio. Audio cables are only intended to carry a 20kHz bandwidth, or possibly as much as 100kHz. CD-style digital audio requires 1.411 megabits per second for the signal alone, plus overheads for error correction and signal management, so the cable needs to comfortably handle the megahertz range of signals.

The red and the white plugs are the ends of a quarter century old cable
The red and the white plugs are the ends
of a quarter century old cable
I ended up with an interesting collection. I had a couple of home-made subwoofer cables (for which I used microphone cable and Jaycar-purchased RCA plugs), a cheap subwoofer cable that came with some system or other, one half of a five metre stereo Jaycar audio cable, one half of an inexpensive ten metre stereo Dick Smith cable. Oh, and both sides of the interconnect that was packaged my first CD player, the original Sony model, the CDP-101, which I purchased new over a quarter of a century ago. To join these cables I used a collection of double ended female adaptors I had.

The net result is that my coaxial digital audio cable, connecting my DVD player and home theatre receiver, is a 44 metre long hodge-podge of cables of indifferent qualities, joined by cheap RCA connectors and piled roughly on my floor.

Then I selected some listening material. I did a quick check with some CDs -- which worked fine -- but then I moved to DTS CDs. These package a DTS surround sound signal in a CD-like bitstream to fool equipment into playing them. The home theatre receiver quickly realises that if played back as PCM, the speakers will produce nothing but extremely nasty high level noise, and so engages the DTS decoder. The nice thing about this is that while the demands on the system are the same as CD with regard to bitrate, if there were any significant signal problems, the DTS would fail in an obvious way.

I played all the way through the DTS CDs of the movie Titanic soundtrack, a 'Best Of' The Police, Seventh Sojourn by The Moody Blues and, to finish off, the DVD Audio disc of Queen's A Night At The Opera, with the DTS 96/24 (bitrate 1.5Mbps) audio selected.

No problems. There were no dropouts, no unseemly distortion.

So I raised the stakes by increasing the bitrate to 4.608Mbps. I played the Queen again, this time selecting the two channel track with 96kHz, 24 bit sound. Then I played the Livingston Taylor DVD Ink -- also 24 bit, 96kHz stereo at 4.608Mbps. This is over three times the bandwidth of standard CD audio.

There was one tiny dropout during the playing of Ink. I rewound immediately, but couldn't reproduce it, so I couldn't nail down the cause.

But what I can say is that if the ability of a cobbled-together 44 metre cable to deliver digital audio at 3.5 times the rate of standard CD sound is so very nearly perfectly reliable, I have grave doubts about the ability of a $400 one metre long digital audio cable to improve the sound quality over a $20 one.

© 2002-2010, Stephen Dawson