Speaker Cables
Can You Hear the Difference?
In the early 1980s, esoteric high-end audio as we know it today was just taking off as an alternative to the massmarket equipment offered in neighborhood Tv/appliance stores. Fueled by an underground audio press that included magazines and newsletters such as Sound & Vision sister publication Stereophile, The Absolute Sound, International Audio Review, The Audio Critic, and others, a cottage industry emerged, one populated by small manufacturers of low-volume, high-priced exotica claiming greater faithfulness to the music than the gear reviewed and advertised in the pages of Stereo Review, High Fidelity, Audio, et al. Some of these claims were founded—true advances were indeed being made by start-ups run by technicians with first-class bonafides and good ears. But the High End also attracted its share of half-baked products and at least a few charlatans looking to cash in selling accessories that had little higher performance than a dime-store engagement ring.
In the midst of all this, the premium cable business emerged, driven in no small part by the success of the early Monster Cable products that followed the company’s founding by engineer/audiophile Noel Lee in 1979. The editors of our precursor Stereo Review were suspicious of the benefits of such speaker cables and interconnects, which were suddenly being proffered by an ever-widening mix of high-end specialists, often at prices far higher than Monster’s. The highly objective measurement-based testing approach employed by Julian Hirsch and his colleagues already ran counter to the high-end community’s subjective reviews, which focused solely on claimed sonic differences that SR’S instruments couldn’t detect. It wasn’t long before Stereo Review began positioning itself as the skeptical voice of reason in what its editors deemed an audio industry gone mad.
It was no surprise, then, that in 1983, the magazine jumped at the opportunity to conduct a double-blind listening test, which editorin-chief Bill Livingston and his colleagues hoped would reveal, scientifically, that high-end cables were indeed a hoax and provided no higher performance than the everyday lamp cord in common use at the time. The pitch was made by accomplished research psychiatrist Laurence Greenhill, an active member of the Westchester (county) Audiophile Society based in the New York City suburbs, and a reviewer for the audio journal High Performance Review. Greenhill’s science background and personal interest in acoustic perception and subjective listening had led him to write some articles on these subjects, and eventually to fascination with a new audio testing device called the ABX Double Blind Comparator System. This box (as explained in Greenhill’s article) used logic circuitry to randomly trigger relays that allowed listeners to compare two unknown sources instantaneously while eliminating the need for (and therefore, the bias of) a test administrator to select the sources; hence the term double-blind. The system had been developed by members of the
Southeastern Michigan Woofer and Tweeter Marching Society, a DIY hobbyist group out of the Detroit area (of which former long-time Stereo Review contributor Tom Nousaine was an active member).
By chance, Greenhill lived a few miles away from Hirsch, and their casual acquaintance provided the entrée. The SR editorial team worked up the test protocol with Greenhill and some Audiophile Society colleagues and Marching Society member/abx officer David Clark. To keep things from getting unruly, they settled on testing three speaker cables at 30-foot lengths, including thin, 24-gauge of the type found at electronics outlets; 16-gauge lamp cord bought at a hardware store; and 11.5gauge Monster Cable to represent the premium segment. As the most visible and successful purveyor in the category, Monster was an obvious choice, though, as Greenhill told me recently for this article, it’s noteworthy that Monster’s early speaker cable was essentially also twisted-filament copper like the other test subjects, with nothing particularly advanced in the cable design compared with later products that came from Monster and other manufacturers. An 11-member listening panel was recruited largely among experienced high-end audiophiles from the Westchester Audiophile Society, with Greenhill’s own young son and another high-schooler filling out the mix. The 50 hours of tests were conducted in Greenhill’s personal listening room with equipment borrowed from manufacturers.
The resulting article created a firestorm. As you’ll read, the panel identified, to a statistically significant degree, the 24-gauge from the other two contenders with pink noise as the source. More critically, they also identified, again with statistical significance, the Monster Cable from the 16-gauge with pink noise. But the latter results didn’t hold when choral music was used, and none of the Monster versus 16-gauge results passed the higher threshold of a 75 percent or greater detection rate said to be psychoacoustically significant.
Interestingly, it was not the results of the test (which were mixed, after all), but the article’s written conclusion that set off so much heated discussion. As originally submitted, Greenhill’s balanced conclusion acknowledged that the panel had indeed heard statistically valid differences, at least with pink noise, and that the test should be seen not as a blanket result for all audiophile cables, but simply as research on these three, largely similar test subjects that were distinguishable mostly by level changes easily associated with gauge.
SR’S editors, however, rewrote the ending to create something akin to a blanket condemnation of the category and pressured Greenhill to accept the changes, a decision he later regretted. “They felt my conclusion didn’t grab the reader as it should, but that was my point,” Greenhill recalled. “I wanted to conclude that this test wasn’t a fair judge of all cables, but revealed you could tell level differences based on the size of the cable. They wanted it to be a condemnation of [highend] audiophiles.” (You can read Stereophile’s description of the controversy on stereophile. com; search for “The Horse’s Mouth.”)
by Laurence Greenhill
The article as published was picked apart and attacked by the highend press, and the Monster Cable company, not unreasonably, felt unfairly maligned as expressed in a letter to the author from Noel Lee. Greenhill, for a time, became a pariah in the high-end community, though he continued reviewing and has for many years been a regular contributor to Stereophile. He now lives in California and, at 76, is still a practicing hospital physician and a teacher at the University of California, San Francisco. Today, 35 years later, the debate over audiophile cables remains as active as ever.— Rob Sabin
Some audiophiles believe that the seemingly innocuous wires used to connect stereo amplifiers to loudspeakers actually have a considerable effect on a system’s overall sound quality. Unsatisfied with the performance of 16-gauge heavy-duty lamp cord or “zip” cord (“gauge”
is a measure of thickness; the lower the number, the thicker the wire), these purists install exotic, expensive, and physically imposing cables instead. A 30-foot run of special audio cable may cost anywhere from $55 (for a pair of Monster Cables) to more than $300 (for Levinson wire). After purchase, these thick and massive wires are terminated with special lugs or pressure-fitting banana plugs ($25 per pair), coated with a contact cleaner (Cramolin), and installed with loving care.
Does it make any difference? Is all this trouble and expense necessary to achieve the best possible sound? The editors of Stereo Review have long maintained that for normal home cable runs (in the 20- to 50-foot range) to 8-ohm speakers, 16-gauge zip cord provides optimal power transmission between a system’s amplifier and speakers. The “official” view has been that heavier and/or more expensive audio cables represent—for ordinary home installations—electronic overkill.
Nevertheless, equipment-oriented audio hobbyists have continued to tinker and experiment with their systems’ connections. Quick to respond to a felt need, audio manufacturers have filled the marketplace with a variety of exotically named wires: Discwasher Smog-lifters I, AudioSource UHD, Live Wire 301X4, Mogami Neglex 2477, Oracle Leonische, Fulton Golds, Monster Power Line, Big Red Live Wire, and so on. Is this just another case of the mystique and glamour of high-end audio inducing otherwise sensible music lovers to waste their money? Or is there some real sonic advantage to using audiophile speaker cables?
Our view is that if there are audible differences among speaker cables, they probably derive from measurable differences in their respective electrical properties. Nelson Pass of Threshold Corp., for instance, claims that subtle sonic differences can be correlated with differences in electrical resistance, capacitance, and inductance, which vary according to overall cable length. And these factors can result in measurable differences in frequency response and signal level, depending on the particular speakers and amplifiers the cables connect. But the measurable differences in electrical characteristics and performance between audiophile cables and cheaper, 16-gauge zip cord seem too small to explain the apparently huge audible differences that are sometimes reported.
Just how different in electrical characteristics do cables have to be before there are audible effects on frequency response and signal level? In order to answer this question, among others, we embarked on a series of laboratory and listening tests. Because of logistical, time, and budget restrictions, only three cables were exhaustively tested: (1) 16-gauge heavy-duty lamp cord, purchased from a suburban hardware store for 30 cents a foot; (2) 30-foot lengths of New Monster Cable, costing $55 a pair; and (3) 24-gauge “loudspeaker cable,” available from many sources at about 3 cents a foot.
Although the high electrical resistance of 24-gauge wire makes it inadvisable for runs of more than a few feet, its low price and small diameter make it appealing to decorators, who use it unashamedly. It seemed the ideal “worst case” cable—if any cable would sound different, 24-gauge should be the one. The 16- and 24-gauge wires were cut to the same 30-foot lengths as the Monster Cable, and all the cables were terminated with Monster’s X-terminators for consistency in the connections.
Lab, Tests
For the first part of our tests, Julian Hirsch measured the three selected cables’ resistance and interconductor capacitance. The results are easily summarized: in 30-foot lengths, the 16-gauge zip cord had a resistance of 0.24 ohms, the Monster Cable 0.09 ohms, and the 24gauge “loudspeaker cable” 1.8 ohms; their interconductor capacitances were 420, 600, and 400 picofarads, respectively.
The cables were then connected between a high-quality power amplifier (a Perreaux PMF 21508) and two pairs of high-quality speakers (Spendor BC-1S and KEF 105.2s). A 1,000-Hz square wave was fed to the speakers through the cables, and the cables’ effects on the signal were monitored by subtracting the waveform at the speaker terminals from the waveform at the amplifier output. Photographs were made of the oscilloscope displays. There were no observable differences between any of the cables driving either speaker system. All the measurements were then set aside and not shared with the listening panel until after the listening tests.
Open Listening
In the initial open (non-blind) phase of the listening tests, the listeners individually evaluated the sound of seven different music selections played with each of the three different cables while knowing which cable was in use at each moment. They were asked to fill out an elaborate, eight-page questionnaire. In addition to ten questions about the sound of the different cables, the form asked about the panelist’s age, occupation, degree and kind of involvement with audio, and any prior attitudes toward the specific cables being tested. The procedure was designed to collect information on the panelists’ various biases and to suggest how those biases might change in a group-listening situation. For example, a listener’s preference might be influenced by knowledge of a friend’s preference or by the reputation that a particular brand or type of cable has acquired through reviews in the “underground” hi-fi magazines.
Controlled Listening
The heart of our whole project was the set of controlled listening tests. These differed from the initial open tests in that the listening procedure was designed to maximize psychoacoustical fairness. That is, in order to keep the listeners’ biases about the weight, appearance, cost, or brand of the cables from influencing their sonic preferences, the tests in this part were double blind—neither the listeners nor the test administrator knew which cable was being listened to. Switching between the different cables was practically instantaneous (less than
50 milliseconds) in order to make subtle sonic differences as apparent as possible (human beings have notoriously poor long-term memories for sounds).
To make possible such a double-blind, instantaneous-switching test, we used a laboratory-grade audio comparator developed and sold by the ABX Company. It consists of control circuits and relays that can rapidly switch between different inputs. The listener compares sounds and decides whether a particular source, designated X, is sonically the same or different from each of two other sources, designated A or B (hence ABX). The trick is that source X is actually either A or B; each connection is made at random by the ABX comparator and is not identified for either the listener or the tester. The listener writes his response on a test form. Digital memory circuits store the sequence of connections for retrieval and analysis at the end of the test.
Level matching between cables to compensate for their differing electrical-loss characteristics was not performed for every comparison in this part. The basic premise of these tests was that measurable differences among cables would become audible if they were large enough. Matching all measurable parameters—including signal loss—would make the hypothesis untestable. For one comparison, however, of Monster Cable with the 24-gauge speaker wire, we did compensate for level differences by switching in a precision attenuator between the preamplifier and the power amplifier whenever the lower-resistance Monster Cable was selected by the ABX system.
The purpose of this was to determine if there were any audible differences between the two cables other than those resulting from a volume-level change.
Some audiophiles refuse to accept the validity of this kind of controlled listening test. They argue that either the test procedure or the switching systems will mask important sonic differences. The whole procedure and test setups are designed to increase listeners’ sensitivity to small differences and to bypass some of the limitations of the human hearing system so that the results will be applicable to all listeners, not just those participating in the tests. Let’s review three of the main features of this phase of our tests: (1) The ability to switch almost instantaneously from one cable to another with a precision set of low-resistance relays makes perceived sonic differences as vivid as they can be given our limited sonic memories. (2) The double-blind listening setup relieves the tester and the listeners from having to compensate for preconceptions about the different cables. (3) The use of numerous listeners making a large number of comparisons produces sufficient data for meaningful statistical analysis and generalization.
The procedure for the controlled listening tests can be quickly summarized: six cable comparisons were carried out using the ABX comparator and either pink noise or a choral music selection. Monster Cable was compared with 24-gauge wire, gain matched and unmatched; next, Monster Cable was compared with 16-gauge zip cord; then the 16-gauge and 24-gauge cables were compared with each other. Each comparison consisted of a series of fifteen tests; in each test, as explained previously, the listener was asked to identify which of two constant sources, A and B, was the same as the switched source X. The choice and order of cables being compared and program sources being used were determined by a table of random numbers, with switching randomly controlled by the ABX comparator.
Results were examined only after each listener completed the series of comparisons, and results were held confidential so that each panel member knew only his own score—after all, the egos of these “golden ears” were on the line. Moreover, in our account of the results we will
consider mainly the performance of the group as a whole; individual panelists will be identified by code letter—to protect the guilty! Further details of the test setup can be found in the box labeled “Test Methods and Equipment.”
The listeners were all males with an average age of about thirtynine years (ranging from thirteen to sixty). Although two were highschool students, most of the remaining nine were middle-aged professionals with a serious involvement with audio. Six panelists owned expensive esoteric twin-lead speaker cables, two owned interwoven audiophile cables, and one even used 14-gauge zip cord. Seven thought that controlled, double-blind tests like those we used were valid, but the others thought such tests missed the boat.
Half the panelists thought that sonic differences between speaker cables could not be explained by measurable differences in the cables’ electrical characteristics. Before the listening tests even began, ten out of the eleven listeners expected the “better” cables to show improved bass, “punchiness,” and frequency response.
The left-hand portion of Table I summarizes the pre-test attitudes of the listeners. In terms of attitudes, Monster Cable scored significantly higher on appearance and reputation than either 16- or 24-gauge cable. Although 16-gauge scored highest on frequency of use and cost-effectiveness, the preference was not statistically significant in these categories. Generally, the panel seemed to prefer Monster Cable from the start, suggesting that it would come out on top in the non-blind open listening tests, which is just what happened.
The Open Tests
When the listeners knew which cable they were listening to, Monster Cable scored significantly higher, as the data in the right side of
Table I show. It was preferred to 24-gauge in seven out of ten categories. Monster Cable also scored significantly higher than 16-gauge in three categories. In the pink-noise portions of the test, 16-gauge scored higher than 24-gauge.
Subjective comments from the panelists were consistent with the numerical scoring. They reported that Monster Cable conveyed deeper bass, more impact, more ambience, and a fuller, lusher sound with greater transparency and that the program could be played louder than with the other cables. In comparison, the sound with 16-gauge was said to be “congealed and homogenized,” revealing less “space and expanse” in the music. Listeners also noted that 24-gauge produced a drop in level and rolloffs in frequency response at both ends of the audio spectrum. The “clipped and compressed” qualities of 24-gauge were even said to give it a “New England” sound!
It is important to emphasize again that these ratings and comments came from listeners who knew which cables they were listening to. The panelists’ strong pre-test biases in favor of Monster Cable could have influenced their reactions during the open listening tests no matter how conscientiously they tried to ignore them. Eliminating the listeners’ awareness of which cable was being auditioned was, therefore, a major consideration in our design of the controlled listening tests.
The Controlled Tests
The most striking result of the controlled listening tests (see Table
II) was the selecting out of the 24-gauge wire. All the comparisons in which it was not matched in gain level with the other cables show a high degree of statistical significance. With pink noise, every panelist was able to distinguish non-level-matched 24-gauge wire from either of the two heavier cables in at least twelve out of fifteen tries, or 75 percent, which is an accepted threshold of significance in psychoacoustical testing. Even when the levels were matched between 24-gauge wire and Monster Cable (Comparison No. 2), four listeners still heard a difference. While one listener might have just been making lucky guesses, it is very unlikely that four out of eleven were doing so.
It is reasonable to assume that because Monster Cable is thicker and has lower electrical resistance than 24-gauge wire, it is a more electrically “accurate” signal conductor. The audible differences between them in our controlled listening tests strongly suggest, therefore, that 24-gauge is too thin for optimal audio performance—in 30-foot lengths, at least.
When 16-gauge wire was pitted against Monster Cable using pink noise as the program material (Comparison No. 3), three of the panelists correctly identified it in twelve out of fifteen tries. Again, it is very unlikely that this could have occurred by chance (the odds are less than 1 in 999). But when choral music was used instead of pink noise (Comparison No. 5), none of the panelists correctly distinguished 16-gauge from Monster Cable to a psychoacoustically significant degree (75 percent of the time). This suggests that both 16-gauge and Monster Cable are fine conductors for music signals of typical complexity, but a very sensitive listener under the best conditions might find an audible difference—not necessarily a preference— between them on wide-band (pink-noise) test signals.
In sum, five out of the six sets of controlled listening comparisons gave results that are statistically significant using the binomial distribution (see caption for Table II). In only two sets of comparisons, however—both testing 24-gauge wire against a heavier cable using pink noise—were the results above the stricter 75 percent threshold usually considered necessary to establish an audibly significant difference.
Comments
While it may not be news that the thin 24-gauge speaker cable sounded different from the other cables tested, what does this result mean in terms of cable measurements and their audible effects? A 1- to 2-db decrease in sound level was measured for the 24-gauge wire during the pink-noise listening tests, and a number of panel members reported a loss of treble and bass response with 24-gauge during open listening. Was the wire also changing the perceived frequency balance?
Julian Hirsch’s measurements of the 24-gauge wire showed it to be much higher in resistance than Monster Cable or 16-gauge wire. Its 1.8-ohm resistance resulted in a 1.76-db insertion loss with an 8-ohm resistive load, which means that fully one-third of the amplifier’s output was being dissipated by the cable over the 30-foot runs we used. This accounts for the unanimously audible decrease in level. Frequency response was also affected because the 24-gauge cable’s relatively high series resistance interacted with the system impedance of the KEF 105.2 speakers used in the tests to produce 1.25-db frequencyresponse depressions at 100, 1,000, and 4,000 Hz. This perhaps explains the subjective impression that with the 24-gauge wire the sound was “duller.” Note that even in the double-blind, matched-level test of 24-gauge vs. Monster Cable (Comparison No. 2), four panel members heard differences. The level-matching procedure corrected only for an overall level mismatch, not for frequency-response changes caused by the 24-gauge wire.
The panelists’ initial preference for Monster Cable was only partly vindicated by their ability to distinguish it from 16-gauge cable in the controlled pink-noise test (Comparison No. 3). What the panelists noticed during this comparison was either the result of the measured 0.16-db insertion-loss difference between the two cables or the corresponding 0.04-db frequency-response variation when they were connected to the KEF 105.2 speakers. The former is far more likely.
It seems that tests with real music signals decrease the ability of listeners to distinguish small sonic differences between cables. Although three panelists heard a difference between 16-gauge and Monster Cable with pink noise, the panel as a whole was unable to hear any significant difference between them with choral music
(Comparison No. 5). Even when Monster Cable was compared with 24-gauge using choral music (Comparison No. 6), only three panel members had psychoacoustically significant scores (twelve out of fifteen or better), though the group’s score in this non-level-matched comparison did reach the less strict level of “statistical significance.” It seems likely that the differences between 24-gauge and Monster Cable would be even harder to hear (if audible at all) if the levels were matched and typical music, not pink noise, were used.
Because of our short auditory memories, noticing very subtle sonic differences requires a more or less constant signal as well as instantaneous switching between the components being compared. One must wonder how real are the vast sonic differences reported in the underground audio press among such components as cables, preamplifiers, and power amplifiers, since the results derive from uncontrolled, non-levelmatched listening tests using only wide-dynamicrange musical material.
The final significant conclusion one can draw from our data is that at least one genuine “golden ear” does exist. Listener J on our panel had 80 percent “hits” (psychoacoustically significant scores) with an average score of 12.7 (out of 15). Obviously, certain listeners—whether through talent, training, or experience—can hear small differences between components. But the majority of the panelists, although by no means poor or unskilled listeners, heard only differences resulting from the relatively large, easily measured variations in signal level and frequency response in the 24-gauge comparisons. Note, however, that J was the only panel member to score a hit for both music and level-matched pink-noise comparisons between Monster Cable and 24-gauge (Nos. 2 and 6). In these two tests only a third of the listeners achieved psychoacoustically significant scores, showing that the audible differences in these tests were not very obvious despite the overall statistical significance of the group’s scores. Our tests were certainly not exhaustive, since only three different cables were compared. Yet the results demonstrate that while Monster Cable and 16-gauge lamp cord are both audibly different from and probably superior to 24-gauge wire, 16-gauge is good enough to be indistinguishable from Monster Cable when playing music. An esoteric cable would have to be substantially better than Monster Cable in order to be demonstrably superior to 16-gauge wire. One of the listeners on the panel ran a quick but controlled listening test of Monster Cable against high-capacitance Mogami Cable (with its “damper” removed) and 8-gauge Levinson HF-LOC twin-lead, both products more costly than Monster Cable. He did no better in distinguishing Monster Cable from the other two than chance would allow.
So what do our fifty hours of testing, scoring, comparing, and listening to speaker cables amount to? Only that 16-gauge lamp cord and Monster Cable are indistinguishable from each other with music and seem to be superior to the 24-gauge wire commonly sold or given away as “speaker cable.” Remember, however, that it was a measurable characteristic—higher resistance per foot—that made 24-gauge sound different from the other cables. If the cable runs were only 6 instead of 30 feet, the overall cable resistances would have been lower and our tests would probably have found no audible differences between the three cables. This project was unable to validate the sonic benefits claimed for exotic speaker cables over common 16-gauge zip cord. We can only conclude, therefore, that there is little advantage besides pride of ownership in using these thick, expensive wires.
Laurence Greenhill is a research psychiatrist who is currently studying the effects of drugs on perception. He has been an audio enthusiast since the late Sixties and has written a number of articles on perception, audio, and subjective listening tests.