Anectdotes, Theories and Facts
One of the things that has struck me while perusing (and occasionally contributing to) online conversations about musical gear is that many participants are fundamentally, and in certain cases willfully ignorant of the fundamental difference between an anectdote and a fact. Ignorant? It's not a term I use lightly. It certainly does apply, though.
- 1. A lack of knowledge. A state of being unaware or uniformed.
- 2. (pejorative) The choice to not act or behave in accordance with regard to certain information in order to suit one's needs or beliefs.
This behavior recently came to light in a discussion regarding the influence of instrument cables on the sound of an amplified guitar. As you may be aware if you've read my articles about cables, my experience has shown me that total cable capacitance is the single most influential factor in determining how a cable will change the sound of a guitar.
To briefly summarize, the cable is part of a resonant circuit which consists of the guitar pickups and controls, the cable and the amplifier's input stage. Given the same guitar and amp, choosing a cable having a different total capacitance will change the resonant frequency of the circuit which will in turn change the sound of the guitar. This mechanism is well understood, measurable, audible and experimentally verifiable.
Even better, there is predictive value in this knowledge: I know that using a longer cable of the same construction (or a cable of the same length but with a higher capacitance per unit length) will shift the resonant frequency to a lower frequency, thus emphasizing different harmonics and also rolling off more of the high frequency content. Conversely, a lower-capacitance cable will shift the resonant peak and high-frequency response upwards. I can take advantage of that knowledge to fine-tune the upper-midrange response of my guitar.
I can also use this knowledge to guide my choice of cables based upon the premise that a cable's sound depends upon its capacitance. I know that if I have two cables which exhibit the same total capacitance, then they'll sound the same in my rig. That leaves me free to base my choice on other discernible factors such as cost, reliability, flexibility, sheilding or color.
There's a growing interest (among both professional and amateur musicians) in the use of high-end "audiophile-style" cables for musical instruments. Manufacturers and marketers alike make extravagant claims for the benefits and subtle sonic "signatures" of their cables using fuzzy, non-specific descriptions of unmeasurable and unverifiable effects which have absolutely zero predictive value for the purchaser. Sellers advance plausible-sounding theories using obscurantist language to support their "proprietary" claims.
- 1. deliberate obscurity or evasion of clarity
- 2. opposition to the increase and spread of knowledge
The advertising copy used to support these claims often borrows from known physical principles in order to lend credence to the hyperbole. With the exception of capacitance and sheild coverage, which although measurable (or perhaps because they are measurable) are often omitted from the specifications of audiophile cables, virtually all of the claimed benefits of audiophile cable construction do not apply under the conditions in which the cables are used. To put it another way, the benefits of audiophile cables are grossly misrepresented and overstated based upon misapplication and misinterpretation of the physical science upon which the vendor constructs his claims.
Here's a short list of some of the more common claims made by audiophile cable vendors, together with my rebuttals:
- skin effect
- This is a very popular claim based upon the observation that certain electromagnetic effects cause a wire to conduct higher-frequency alternating currents closer to its surface. This observation has led to various arguments by audiophile cable manufacturers. A common argument is that the skin effect, in combination with other plausible physical mechanisms, somehow affects the propagation of different frequencies in different ways, such that the cable subtly alters the phase relationship between high- and low-frequency content. However, skin effect is completely ignorable at the frequencies and currents involved in moving a signal from guitar to amp.
- Some vendors, perhaps realizing that skin effect has been rather vigorously debunked by skeptics of audiophilia, like to point out that a cable has an inherent amount of inductance and that inductance resists a change in current. In fact, inductance opposes the flow of higher frequency signals and delays high frequency information relative to low frequency information. All of that is strictly true. And all of it is totally ignorable at the frequencies and currents involved in moving a signal from guitar to amp.
- Like inductance, this is another red herring. The resistance of an instrument cable is many orders of magnitude less than other circuit elements. Again, completely ignorable.
- polished conductors
- This is a good one, because it's related to skin effect. The argument goes something like this: "Since we know that skin effect pushes high frequency information closer to the surface of a conductor, it must be beneficial if the the surface of the conductor is polished in order to facilitate the movement of electrons in the so-called surface layer. This one really takes the cake with respect to a complete lack of understanding of the relative scale of the size of an electron with respect to the terrain afforded by a drawn conductor, whether polished or not.
- strand jumping
- This one's rather creative, too. Supposedly in a stranded cable the higher frequencies (because of our old friend the skin effect) will randomly "jump" from strand to strand making the total path length unpredictably time-varying and therefore "blurring" the high frequency signal content. Considering that the velocity of propagation is about 80% of the speed of light and that the distance is a couple tens of feet or less, we're talking about nanoseconds of difference. Even assuming that the argument is true, a few nanoseconds is nothing considering that the period of the highest frequency information coming from a guitar is more than a hundred microseconds. Again, the scale of the effect is off by many orders of magnitude. Interestingly, the strand-jumping argument is most often advanced as a reason for using solid-core conductors rather than stranded conductors; it seems like the cable manufacturers can't even agree on which physical principles they're going to pervert for marketing purposes.
- The argument here is that boundaries between copper grains or crystals (formed during the process of cooling the copper as it's drawn through a die) affect the flow of electrons in an unpredictable manner. This conveniently ignores that a current introduced at one end of a conductor has to flow through the other end in an exactly equal amount. Check your electrical engineering textbooks for Thevenin's Theorem, guys... It's the aggregate flow of current that matters in an electrical circuit, not the path taken by an individual electron. This is yet another example of cable manufacturers' misapplication of scale.
- At least one manufacturer has introduced a carbon-core cable which claims to mitigate the effects of things like metallic grain boundaries. Of course, carbon is a poor conductor and is an excellent generator of thermal shot noise. Perhaps the laws of physics (and the other vendors' claims that lower resistance is better) don't apply because of the manufacturer's proprietary formulation...
- dielectric absorption
- The insulating material between the center conductor and the shield of an instrument cable is the dielectric material between the two sides of a capacitor. A capacitor stores energy as a charge in the space occupied by the dielectric. The dielectric itself can hold a charge which will oppose a change in voltage somewhat beyond the opposition you'd expect based upon a purely capacitive cable. In effect, the dielectric appears as a tiny voltage source which varies with the signal. Yet again, this effect can be safely ignored at the voltages and currents present in an instrument lead.
Those are the more plausible sounding claims advanced by audiophile cable manufacturers. If you weren't paying attention in your high-school or college physics classes, you'll probably know enough to recognize some of the terminology, but not enough to understand how it's being misused. This is unintentional ignorance. You don't know what you don't know. We all have gaps in what we know with certainty. No one is an expert on everything.
Vendors count upon this unintentional ignorance to sell products where the incremental benefit is greatly exceeded by the incremental cost. In the case of high-end cables, the incremental benefit is infinitesimal while the incremental cost can be up to three orders of magnitude above the cost of a pedestrian yet perfectly servicable cable.
As I mentioned earlier, high-end cable vendors don't use terminology that a customer could use to predict the effect of a cable in the intended application. This means that a consumer must rely on subjective opinions formed by reading the manufacturer's literature or by collecting and evaluating subjective opinions formed by other purchasers. It's interesting to observer how often these independent subjective opinions conflict; one could infer that these opinions have little, if any, predictive benefit to the prospective purchaser.
All of this is, of course, very good for the vendor. It's darned near impossible to make a really bad cable at any price point given that purchasers will always base their evaluation on subjective factors. Pick any cable and see whether it does what you expect. If it doesn't, try another one. This has to be beneficial for the vendors... With nothing to guide customers except florid prose, how else can they make a choice except to buy the "best" (i.e.: most expensive) cable that they can afford?
This is an extremely common tactic in marketing. The tactic probably has a fancy name in business school classes, but I call it FUD: "fear, uncertainty and doubt". Customers are motivated to make choices that they won't "regret" when they compare their decisions to those made by their peers and acquaintances. The internet, in particular lifestyle-focused consumer discussion groups (e.g.: guitarist discussion boards), has done wonders to enhance the efficacy of FUD marketing.
Of course, some people -- vendors and customers alike -- will say that not every audible difference can be explained in measurable terms. And this is, to a certain extent, true. However, it's foolish to attribute changes to esoteric causes when a simpler explanation will suffice.
One problem with the claims of high-end cable vendors is that there's no apparent attempt to separate out the contributions of the claimed "improvements" in a way that would allow observation of the supposed improvement in isolation. For example, it's necessary to observe two cables differing only in the method of drawing the wire in order to determine the validity of the hypothesis that the wire-drawing method has an effect on the sound. The unchanged aspects of the two cables serve as a control for the experiment. Notice that I'm not insisting that the differences be measurable. It's enough that the differences be observable whenever the experiments are conducted under identical controlled conditions. Such knowledge would go a long way toward advancing the state of the art, but I can find no evidence that anyone has ever attempted to conduct a properly-controlled experiment. Draw your own conclusions...
Returning for a moment to the perception versus measurement argument, let's consider the claims that human perception somehow exceeds the sensitivity of available measurement technologies. That may well be true. Let's take a moment, though, to consider the scale of the purported effects versus the environment in which the human perception is taking place.
Instrument cables operate in the voltage domain where a factor of ten change in voltage corresponds to a 20 dB change in perceived volume. A cranked guitar amplifier can push about 115 dB to 120 dB SPL. In a very quiet room a relatively noise-free guitar amplifier might generate hum and hiss at a level of about 40 dB to 50 SPL. The difference is the dynamic range of the amplifier: typically 65 dB to 80 dB SPL. To bring this back into the realm of orders of magnitude, 60 dB corresponds to three orders of magnitude (a factor of 1,000) while 80 dB corresponds to four orders of magnitude (a factor of 10,000). Even the smaller 60 dB figure yields a huge difference in loudness.
Now imagine for a moment that you have two players side by side with identical guitars, cables and amps. One plays flat out with everything cranked. The other plays as quietly as the amp can go and still be perceived as putting out a guitar sound through the speakers, just barely louder than the hum and noise. What will you hear if both players play the same thing at the same time? Will be be able to hear any difference in the sound of the louder rig based upon what the player with quieter rig does? You can certainly do this experiment if you wish. You can make it more interesting by having the quieter rig and player be and play something completely different than the louder player and rig; I'll bet that you won't be able to hear the slightest contribution from the quieter rig. This little experiment tells us that a difference of three to four orders of magnitude masks the smaller signal very effectively. Many of the effects touted as "significant" by high-end cable vendors are five or more orders of magnitude down from the signal. Do you think you'll be able to hear that if you can't hear a three- or four-order difference?
The problem is compounded when you consider that playing your rig anywhere but in the very quiet room cuts the dynamic range to more like 20 dB, which further masks any contributions that are 100 dB or more below the guitar signal.
One interesting conjecture is that the minutae of cable construction and material make a much bigger difference (or perhaps any difference at all) to guitarists who play through high-gain amps. I can imagine that the extreme amounts of gain (60 dB, typically) added to create the high gain sound might bring some of the "microscopic" effects up to near-audible levels. I don't know whether those low-level details would be masked by the distortion or not. It's entirely possible that the differences can be explained by macroscopic effects such as cable capacitance; there's just not enough reliable, verifiable data available to form a coherent hypothesis. I'm glad that I play a low-gain rig and don't have to lose sleep over details like that...
On the other hand, some cable consumers do claim to hear significant differences among cables even if playing guitar through a low-gain rig with the capacitances closely matched. I've done that experiment myself and don't notice anything interesting. Changing total capacitance, of course, alters the EQ. But two cables of varying manufacture and quality having roughly equal total capacitance end up sounding the same to me. Maybe my perceptive skills aren't sharp enough to hear a difference if one exists. Or perhaps my expectations color my perceptions. All I know is that I can hear the differences postulated by the macroscopic theory and can not hear any differences postulated by the various microscopic theories advanced by high-end cable vendors.
The psychological term cognitive dissonance suggests an alternative explanation for why purchasers of expensive audiophile cables "hear" clear differences where theory and measurement suggest no significant effect upon the sound, and why purchasers' perceptions often conflict with one another. Advertising copy and peer reviews set an expectation for the performance of the high-end cable, which is invariably priced at a significant multiple of non-audiophile cables. Should the actual performance of the cable fail to meet a priori expectations, the consumer is left with the choice of either admitting that they made a mistake and seeking recompense from the retailer or inventing reasons to enjoy their purchase. The latter approach is most peoples' path of least resistance.
The problem with adjusting one's expectations due to cognitive dissonance is that it is an act which has consequences beyond the person who has subjected his expectations to the mechanisms of denial. Such persons tend to seek validation among like-minded peers, turning a simple mistake (caveat emptor) into a boon for purveyors of snake oil. The groupthink mentality captures more victims, and the cycle repeats. These people are normally not susceptible to either rational arguments or conflicting observations; they believe what they need to believe. This is willful ignorance.