AVReview is planning a blind cable test in London, November 2007. At the moment we’re thinking of including speaker cable, interconnects and mains leads, from very cheap to scarily expensive. If you’d like to get involved or want to make suggestions as to how the test should be conducted, tell us here.
I'd love to! It would be a great opportunity to contrast and compare the relative merits of different equipment. eg Would changing the mains cable be more rewarding than say an upgraded interconnect. It would be most interesting if this were a 'blind' test and if we were able to rank performance without knowledge of comparative costs, and even guess how much the item would retail at.
A good old Pepsi challenge is a great idea. However, I would like to see some attempt at removing experimenter's bias by not having advertising in the test area and by the experimenter not having primary interests.
Each test sequence itself should have a suitable number of samples, i.e. enough to allow the results to suggest things like 'expensive cables are better than cheap cables' but not so many that you can't really recall the relative merit of sample 53 against sample 4.
Results can be skewed by trying to test for more than one thing at a time. An example would be testing 'the relative merit of high quality speaker cables' and 'how do cheap and expensive speaker cables compare'. A poor sounding 'cheap' sample may push all other samples upwards, reducing the ability to discern differences in the 'high quality' samples.
The subject fills in a form and hands it in at the end of test. The design of the form is critical. For example, bias is introduced if the subject is asked to order the test items from good to bad. A more open question allowing samples to occupy the same category is recommended. An additional space for subect's comments is fine however.
I could go on, but you get the idea. If you can get your methodology sorted then your results will carry weight and will stand up to peer review. This is a great thing for the consumer and I'm sure the manufacturers will agree
It's a shame Mike Reid is no longer with us, or I would have suggested the test be based on the classic TV show Runaround.
This could be a splendid enterprise in which I would love to participate, work permitting.
I propose an ABX methodology, for which I am sure that JF would be willing to knock up the hardware. Here's how I see it working. The format of an ABX test is simple. The listener has two test cables (or amplifiers or whatever) called A & B. The test equipment will have randomly assigned X to be either A or B - no person need know which in any one instance. Once the umpire has started a new test (which effectively tosses the coin and determines whether X=A or X=B), it's all down to the test paticipant. He has a box with five buttons marked A, B, X, X is A, X is B. The participant may switch between A, B and X as much as he wants by pressing buttons A, B and X! When he has decided that X is actually A or actually B, he presses "X is A" or "X is B" to indicate his choice. This in turn causes an indicator to illuminate on the umpire's box, saying RIGHT or WRONG. This must be private to the umpire who records the result without reaction and initiates the next test. It is desirable that each participant should be allowed about a dozen shots at a particular test. Perhaps each participant can be allowed to supply his own CD of choice with which to take the test.
The actual switching is carried out by a third box with big, high quality relays. Of course there will need to be short lengths of cable connecting the relay box with the amplifier and speakers, but by using several metres of cable as the test samples, the two cables A & B will be dominant.
In the interests of fairness to individuals, it is possibly desirable that the score kept by the umpire should not attach names to the score sheet, merely "Participant 1" and so on. This will ensure that nobody feels pressured to achieve a particular result on account of their espoused beliefs. To prevent any kind of peer pressure, the actual test room should be occupied only by the umpire and participant, the latter having chosen his participant number at random.
The down side of this approach is that if several folks participate, it would only be possible to test relatively few cables in a day. Simplicity is key. Perhaps confine the speaker test to a good general purpose cable (a quid a metre from Maplin perhaps) and something truly expensive, similarly for the interconnect. Don't attempt too much in one day. If the exercise proves fruitful, it will be a great talking point and there will be many opportunities to develop the tests and the way they are conducted.
So we are up for a true ABX testing of cables, a revolutionary test that has never been carried out before, no I think not, at best we could repeat many tests made before that concluded, no repeatable conclusions were evident, and no conclusion was gained.
Many have done these tests in the past and they all come to the same conclusion, and that there is nothing to prove, no one in reality can hear one cable from another, the cable business is no more than a complete scam, there is no scientific reason for it and it is only a marketing ploy.
I have to admit a most remarkable marketing ploy, to sell a product that has basically no value at all, let alone for many hundreds or even thousands of pounds, but that is successful marketing.
IMO the whole cable issue is a complete con, one of the best kept secrets of the Hi Fi industry, and one that should be made public, for all to see what a scam it is. But like so many things in life that is only MY OPINION.
I have similar opinions on the energy saving of CFL lights, if your interested then e-mail me and I will send you the whole article.
Thanks for the input folks. How would a Friday in November suit - 9th, 16th, or 23rd?
Pluto - good point about keeping the participants' responses anonymous, though as to the switching of cables I had been thinking of using the much lower tech solution of a curtain to hide which cables were being used. I presume the main advantage of the box would be that changes between cables could be effected much more quickly?
"as to the switching of cables I had been thinking of using the much lower tech solution of a curtain to hide which cables were being used. I presume the main advantage of the box would be that changes between cables could be effected much more quickly?"
That is absolutely the point. While there are methods of preventing the participants from knowing what equipment is under test, the curtain approach demands that several people listen as a group, because you can't keep performing a major swapping operation for every individual taking part. Once you have a group listening session, objectivity is lost. It only takes one individual in the group to go "mmm" or "eek" to influence the others. The other practical problem with groups is that few get the hot seat - not that this makes that much difference in my view but it's an unhelpful factor.
But the REAL benefit of proper switching is, as you rightly state, instant comparison. I have taken part in ABX experiments where the switching was so clean it was hard to believe a switch had actually occurred. Under conditions like this, you have a real chance of detecting whether a cable swap, and nothing more, really does have an effect on the listening experience.
Personally, I would like to concentrate on Maplin £1/m versus an expensive cable of similar style - i.e. a cable of good weight (to ensure low R and high current capability) and a construction that an electrical engineer would consider "not too wacky". If the day proves a success (or even partial success), the scope for further tests is immense.
I'd love to be involved, and can make any of those dates... BTW will any of those cables be LCR tested, and do you plan to check this to see if there's any possible correlation?
I wasn't planning on having cables LCR tested beforehand, since we'd be trying to establish if there is any audible difference, rather than trying to establish possible causes. As you say Pluto, we could always do that afterwards if needs be.
BTW Dana, where do you stand on the cables debate?
Pretty neutral really, while I don't belive that differing metal conductors will change the sound in any way whatsoever (providing there's adequate CSA), I think that differing construction methods could change the sound wrt capacitance etc.
The main key to the test will be to ensure that all volume levels are closely matched, attenuating leads will sound duller imo, so that needs to be accounted for... BTW I have a small hand neld LCR meter if that's any help?
I've no objection to an ABX box if no-one else has, but I was thinking of having the equipment behind a screen in any case, so no-one would be able to see when, or indeed if, cables were being exchanged. The only advantage I can see for the box would be speed, but if we use terminated cables, changeover shouldn't take too long in any case.
The main key to the test will be to ensure that all volume levels are closely matched
Any apparent difference of loudness can only be due to the cable itself - nothing else should change - and that is only likely if one of the cables has significantly higher resistance than the other. As I've said before, I feel this first test could best be used to investigate two cables of broadly similar physical characteristics, the difference being a hundredfold in the price!
Re small hand held LCR meter, can it measure milliOhms, microHenrys, picoFarads as over say 3 metres we are looking at some very small values.
It gives what the overriding electrical profile of the cable looks like, as if it was a component, so therefore a lot of ic's that I've measured tend to be in the range 50-90 picoFarads for example - hope that helps?
