Speaker cable blind listening tests

[musicoverall]

My cable listening test has been completed. Tonight I find out how I scored and I'll have time this coming weekend to post the results which I will do if anyone is interested.

[theaudiohobby]

I am waiting for the results of these tests.

[musicoverall]

I am waiting for the results of these tests.

Sorry in advance for the length of this post.

First of all, I undertook this test not to prove anything to the world but only to arm myself with empirical evidence in the face of all the theory regarding cable sonics, theory that was not in accordance with my own anecdotal experience. With the exception of Markw, I'm not aware of anyone at A/R actually participating in blind cable tests. I'm all for theory and science because they can explain a lot about what we hear and how measurements correlate to sound but, either they don't totally dictate sound or we aren't measuring all the right things... so goeth the subjectivist mantra. I went into this with an open mind, of course believing I would score well but also open to the possibility of my own biases being the determinant for what I imagined was happening.

I tested my reference Cardas Neutral Reference speaker cable against Home Depot 12 awg "zipcord". I originally was going to use 16 awg zip from Radio Shack but was told that my results might be swayed since the Cardas is 8.5 guage. I couldn't find any zipcord in that exact guage so I got as close as I could. Both sets of cables were 1.5 meters in length, so I doubt the guage difference was a factor. A friend switched (or not) the cables each night over a period of 13 consecutive nights. He flipped a coin to determine which cable would be in place on which night. Interestingly, the zipcord was used 8 times and the Cardas only 5... but the coin ruled.

I used music with which I'm totally familiar... well recorded CD's of Bill Evans' "Sunday Night at the Village Vanguard" and Beethoven's 5th Symphony. The former is a jazz piano trio and the latter, obviously, a symphony. I played roughly 1/3 of each disc each night, the first 1/3 on the first night, the second 1/3 on the second, etc. Once I had played both discs in full once, I started over. I did not want to use the same 30 second snippet of music for each test because I felt that if I had to focus too diligently on a cymbal crash or an orchestra passage, there was no point in worrying about wire. The differences, if any, had to be apparent in the same manner as we listen to music... without any undue focus on a portion of the music or a portion of the instruments but by listening to it as a total piece.

Of the 13 trials, 3 of them ended up with me totally guessing. After listening, I found I had no idea whatsoever as to which cable was in use. I scored 1 correct of those 3. For two trials I was reasonably certain which cable was in use. I was only 50% correct. So far, I've scored just 2 corrects out of 5 trials.

However, during the other 8 trials, I felt I was absolutely positive which cable I was hearing. There was no doubt in my mind. I scored correctly 8 times out of 8. My total correct answers were 10, out of the 13 trials. about 77% correct.

77% correct is not the 90-95% that statistics state is required and it is not even the 80% correct that Markw allowed during his own listening tests. However, it's significant to me that when I was absolutely certain, I scored 100%. I can't say why I was incorrect on 3 trials or why I wasn't absolutely certain on 5 trials. Perhaps my state of mind was such that the differences were missed. But the sonic differences were very clear to me during 8 trials and I scored correctly each of those 8 times.

For the record, the zipcord very slightly muted detail which gave me the illusion of "less treble extension" or "more recessed midrange". But the major difference was in imaging/soundstaging. The zipcord sounded "confused" at times, and unable to sort out the positioning of the orchestra. It also seemed to slightly shift the musicians in the trio from their original positions (those positions known due to the liner notes). But remember, these are subtle differences - and not noticeable all the time. Getting 3 incorrect seems perfectly reasonable. I didn't focus on sound during these tests, I focused on music. If I had focused on sound, I may have proven something but my intent was mostly to find out how each cable played music. My test tended to bring ou the same differences I heard under sighted conditions but MUCH LESS NOTICEABLY. Which cable is "correct"? Good question. The Cardas sounded "better" and more accurate but I have no way of knowing if it actually was. I preferred it, plain and simple.

Once again, my tests don't prove anything, not even to me. But I find them to be powerful evidence that there is more work to be done before I can blindly (heh, heh) support theory without listening tests to support it. There's more work to be done when measuring cables, how they interact with other components, and how we humans hear things.

Were there noticeable differences between the two cables? Yes. Were they worth the cost? Well... they certainly aren't worth obsessing over. As you can read, they weren't always noticeable at all! This test served to reinforce strongly that cables should be the last thing someone tries to improve their system. But they CAN improve it. The worth of the investment is a personal decision.

I invite your comments/questions.

[shokhead]

Interesting. Good job.

[theaudiohobby]

Thanks very much for posting your results and conclusions.

[mystic]

Sorry in advance for the length of this post.

First of all, I undertook this test not to prove anything to the world but only to arm myself with empirical evidence in the face of all the theory regarding cable sonics, theory that was not in accordance with my own anecdotal experience. With the exception of Markw, I'm not aware of anyone at A/R actually participating in blind cable tests. I'm all for theory and science because they can explain a lot about what we hear and how measurements correlate to sound but, either they don't totally dictate sound or we aren't measuring all the right things... so goeth the subjectivist mantra. I went into this with an open mind, of course believing I would score well but also open to the possibility of my own biases being the determinant for what I imagined was happening.

I tested my reference Cardas Neutral Reference speaker cable against Home Depot 12 awg "zipcord". I originally was going to use 16 awg zip from Radio Shack but was told that my results might be swayed since the Cardas is 8.5 guage. I couldn't find any zipcord in that exact guage so I got as close as I could. Both sets of cables were 1.5 meters in length, so I doubt the guage difference was a factor. A friend switched (or not) the cables each night over a period of 13 consecutive nights. He flipped a coin to determine which cable would be in place on which night. Interestingly, the zipcord was used 8 times and the Cardas only 5... but the coin ruled.

I used music with which I'm totally familiar... well recorded CD's of Bill Evans' "Sunday Night at the Village Vanguard" and Beethoven's 5th Symphony. The former is a jazz piano trio and the latter, obviously, a symphony. I played roughly 1/3 of each disc each night, the first 1/3 on the first night, the second 1/3 on the second, etc. Once I had played both discs in full once, I started over. I did not want to use the same 30 second snippet of music for each test because I felt that if I had to focus too diligently on a cymbal crash or an orchestra passage, there was no point in worrying about wire. The differences, if any, had to be apparent in the same manner as we listen to music... without any undue focus on a portion of the music or a portion of the instruments but by listening to it as a total piece.

Of the 13 trials, 3 of them ended up with me totally guessing. After listening, I found I had no idea whatsoever as to which cable was in use. I scored 1 correct of those 3. For two trials I was reasonably certain which cable was in use. I was only 50% correct. So far, I've scored just 2 corrects out of 5 trials.

However, during the other 8 trials, I felt I was absolutely positive which cable I was hearing. There was no doubt in my mind. I scored correctly 8 times out of 8. My total correct answers were 10, out of the 13 trials. about 77% correct.

77% correct is not the 90-95% that statistics state is required and it is not even the 80% correct that Markw allowed during his own listening tests. However, it's significant to me that when I was absolutely certain, I scored 100%. I can't say why I was incorrect on 3 trials or why I wasn't absolutely certain on 5 trials. Perhaps my state of mind was such that the differences were missed. But the sonic differences were very clear to me during 8 trials and I scored correctly each of those 8 times.

For the record, the zipcord very slightly muted detail which gave me the illusion of "less treble extension" or "more recessed midrange". But the major difference was in imaging/soundstaging. The zipcord sounded "confused" at times, and unable to sort out the positioning of the orchestra. It also seemed to slightly shift the musicians in the trio from their original positions (those positions known due to the liner notes). But remember, these are subtle differences - and not noticeable all the time. Getting 3 incorrect seems perfectly reasonable. I didn't focus on sound during these tests, I focused on music. If I had focused on sound, I may have proven something but my intent was mostly to find out how each cable played music. My test tended to bring ou the same differences I heard under sighted conditions but MUCH LESS NOTICEABLY. Which cable is "correct"? Good question. The Cardas sounded "better" and more accurate but I have no way of knowing if it actually was. I preferred it, plain and simple.

Once again, my tests don't prove anything, not even to me. But I find them to be powerful evidence that there is more work to be done before I can blindly (heh, heh) support theory without listening tests to support it. There's more work to be done when measuring cables, how they interact with other components, and how we humans hear things.

Were there noticeable differences between the two cables? Yes. Were they worth the cost? Well... they certainly aren't worth obsessing over. As you can read, they weren't always noticeable at all! This test served to reinforce strongly that cables should be the last thing someone tries to improve their system. But they CAN improve it. The worth of the investment is a personal decision.

I invite your comments/questions.

Perhaps you performed better statistically than you think you did. Other members can correct me if I'm wrong, but I believe 10 correct identifications in 13 trials has a p value of of 0.046, meaning there is less than a 5 percent chance your scores were random. I believe results with p values under 0.050 usually are considered statistically significant.

Thank you for doing the test and telling the members about it.

[musicoverall]

Perhaps you performed better statistically than you think you did. Other members can correct me if I'm wrong, but I believe 10 correct identifications in 13 trials has a p value of of 0.046, meaning there is less than a 5 percent chance your scores were random. I believe results with p values under 0.050 usually are considered statistically significant.

Thank you for doing the test and telling the members about it.

Interesting. But I'm afraid I can't correct you!

Certainly the guesses were random but I would have bet the farm on the 8 I was sure of.

[Swerd]

Thanks for taking the time and effort to do these listening tests. Its not so simple as people might think. Overall, your test seems fair and your conclusions are not out of line with your observations.

Were there noticeable differences between the two cables? Yes. Were they worth the cost? Well... they certainly aren't worth obsessing over. As you can read, they weren't always noticeable at all! This test served to reinforce strongly that cables should be the last thing someone tries to improve their system. But they CAN improve it. The worth of the investment is a personal decision.

Some people might quibble with your statement that cables CAN improve a system. I think your results of 77% correct suggests that cables MIGHT improve a system, but it would require a higher percentage of correct answers before you could claim cables CAN improve a system.

I tested my reference Cardas Neutral Reference speaker cable against Home Depot 12 awg zipcord. I originally was going to use 16 awg zip from Radio Shack but was told that my results might be swayed since the Cardas is 8.5 guage. I couldn't find any zipcord in that exact guage so I got as close as I could. Both sets of cables were 1.5 meters in length, so I doubt the guage difference was a factor.

I think the biggest variable in your test was the difference in gauges of the speaker wires you tested. You assumed that any difference in 12 and 8.5 gauge would not be a factor

I think that going from 12 to 8.5 gauge might explain the differences you heard. 8.5 gauge wire has more than twice as much copper as the same length of 12 gauge wire. This link http://www.powerstream.com/Wire_Size.htm explains the relationship between gauge and cross-sectional area of wire in AWG system. Every 3 gauge decrease doubles the wire cross sectional area. Two 12 gauge wires equal one 9 gauge wire. It would have been a better comparison if you bundled two 12 gauge wires together to compare to your Cardas 8.5 gauge wire.
So if anyone out reading this wants to try and repeat your test, they should try to eliminate or minimize differences in the gauge of the two wires.

[musicoverall]

Thanks for your comments - much appreciated and food for thought.

On the other hand, I've heard on this very site that unless we're talking about 24 awg wire or extremely long runs, differences in guage aren't noticeable in audio applications. Is that information wrong? We're talking less than 5 feet of wire here, much shorter than the average audio application.

Thanks for the "might" as opposed to "can" and you're correct... my error. Might is the word I intended.

As I mentioned, I simplified this test as much as possible by not putting myself on stage with short passages of music played over and over. If the differences didn't come out in normal listening, I determined that they weren't worth it. I rarely listen critically unless I'm auditioning a component... or writing a music review, but that's a different kind of critique.

[E-Stat]

Sorry in advance for the length of this post...Were there noticeable differences between the two cables? Yes. Were they worth the cost? Well... they certainly aren't worth obsessing over. As you can read, they weren't always noticeable at all! This test served to reinforce strongly that cables should be the last thing someone tries to improve their system. But they CAN improve it. The worth of the investment is a personal decision.

I invite your comments/questions.

Thanks for taking the time to run your trials.

rw

[E-Stat]

I think that going from 12 to 8.5 gauge might explain the differences you heard. 8.5 gauge wire has more than twice as much copper as the same length of 12 gauge wire.

Well the often quoted "expert" Roger Russell would certainly disagree with your assertion governing such a short cable. In fact he has published a chart that "proves" otherwise.

rw

[musicoverall]

Well the often quoted "expert" Roger Russell would certainly disagree with your assertion governing such a short cable. In fact he has published a chart that "proves" otherwise.

rw

Well, apparently most if not all of the naysayer posters around here don't think much of Mr Russell's "proof", despite how much virtual ink they afford him. I've stopped counting how many times on this site I've read the advice that 12 or 16 guage lampcord is perfectly fine for all audio applications unless the wire length is x feet or more- x being something like 25 or 50 feet, I can't recall, but it was much, much longer than 5!

[E-Stat]

Well, apparently most if not all of the naysayer posters around here don't think much of Mr Russell's "proof", despite how much virtual ink they afford him. I've stopped counting how many times on this site I've read the advice that 12 or 16 guage lampcord is perfectly fine for all audio applications unless the wire length is x feet or more- x being something like 25 or 50 feet, I can't recall, but it was much, much longer than 5!

That's why I smiled. Russell's chart is based upon a simplistic formula he asserts is the basis for audibility. Something like "resistance doesn't exceed 5% of the total impedance of the system". Whatever.

While my JPS Labs cables are 9 gauge, I believe the real answer is based upon more than simple resistance.

rw

[musicoverall]

That's why I smiled. Russell's chart is based upon a simplistic formula he asserts is the basis for audibility. Something like "resistance doesn't exceed 5% of the total impedance of the system". Whatever. rw

Theories are pleasant things when you don't have to test them.

[shokhead]

I'll go with a hearing test over any instrument most test anytime. Misicoverall,thanks for the time you spent but i knew you would get ragged on. Some would rather relay on a test that only a dog could hear rather then a real life test.

[musicoverall]

I'll go with a hearing test over any instrument most test anytime. Misicoverall,thanks for the time you spent but i knew you would get ragged on. Some would rather relay on a test that only a dog could hear rather then a real life test.

I haven't been ragged on at all so far. I'm amazed that Robot Czar, Mash and several others haven't been all over this. Maybe they're just busy at the moment. But I'm expecting a much louder hue and cry then I've seen so far.

[RobotCzar]

I haven't been ragged on at all so far. I'm amazed that Robot Czar, Mash and several others haven't been all over this. Maybe they're just busy at the moment. But I'm expecting a much louder hue and cry then I've seen so far.

You shouldn't be ragged on for trying to test your beliefs.

As you say, your test really doesn't prove anything. The statisical issue is this: how often can a person be expected to get 10 of 13 trials correct when there are two alternatives for each trial. I don't have the time or inclination to see if 10 of 13 meets statistical significance or not, and I know it won't matter to you.

You fail to say (or I missed it) if the same cable was being used on all 8 trials you were certain of. I find that significant. I also find it significant that you did not feel certain on every trial. I don't buy that you were having an off day at all. To me that is a indication that something was cueing you on the days you were "certain". I assume you could not see the cables, how did you accomplish that? Did you also blind yourself from volume control positions? How did you deal with volume?

The bottom lline is that I think you probably could hear something different for the 8 trials you felt certain about. What you heard was very likely not due to 1.5 meters of different cable.

There is a lot of silliness in the world. Believing that you can hear better than instuments is one of them, it is simply ridiculuous in my opinion--like saying you can see further than a telescope. I really have no hope of changing the true believers in this forum, they have a religious-like belief in what they can hear and science and logic be damned. Fine, ignorance is bliss. But there are a fiew people with open minds who are new to the field and they deserve to know that there are facts beyond ignorant opinion. I salute you for at least attempting to run a test, which is better than most "audiophiilies" will do as they already "know" their answer.

[musicoverall]

You fail to say (or I missed it) if the same cable was being used on all 8 trials you were certain of. I find that significant. I also find it significant that you did not feel certain on every trial. I don't buy that you were having an off day at all. To me that is a indication that something was cueing you on the days you were "certain". I assume you could not see the cables, how did you accomplish that? Did you also blind yourself from volume control positions? How did you deal with volume?.

No, it wasn't the same cable on all 8 trials. Unfortunately, my notes are at home but I guessed the Cardas as zip twice and the zip as Cardas once - I recall that.

I could not see the cables as they were wrapped in a cheesy homemade paper wrap and taped shut. I used the remote for volume and it doesn't click so I couldn't count them. Believe me on one thing - I had no cues of any kind. There would be no real point in being dishonest since my test won't sway anyone. It was done for my purposes only.

As for "knowing" the answer, well... I still don't. To me, the cables had significant if subtle differences that I picked up on 77% of the time. To you, there's another explanation. The world appears to be spinning on the usual axis, no? I will continue to encourage those with open minds to perform their own tests rather than blindly (heh, heh) follow either the subjectivist or the objectivist path. I think that's reasonable given my test results - reasonable, if not necessarily valid from a scientific POV.

[E-Stat]

Believing that you can hear better than instuments is one of them, it is simply ridiculuous in my opinion--like saying you can see further than a telescope.

Who said anything about "hearing" better than instruments? The presumption that a few metrics fully characterize all the audible aspects of reproducing musical content using real world speakers in real world environments is yours.

rw

[musicoverall]

You fail to say (or I missed it) if the same cable was being used on all 8 trials you were certain of. .

Trial # - Cable - (P)ositive/(R)easonably certain/(G)uess - (C)orrect/(I)ncorrect

1 - Zip - P - C
2 - Zip - P - C
3 - Zip - R - I
4 - Cardas - P - C
5 - Zip - P - C
6 - Cardas - G - I
7 - Zip - P - C
8 - Zip - R - C
9 - Card - P - C
10 -Zip - P - C
11 -Cardas - G - I
12 - Cardas - P - C
13 - Zip - G - C

Now if I were as much of a subjectivist as my set up man, I might claim:

1) I got #3 wrong because I might have assumed the same cable wouldn't have been in place 3 days in a row.
2) I got the Cardas wrong in Trial 11 and ended up with guesses during 2 of the last 3 trials because I was tired.

But I'll leave that to him!

[mystic]

As you say, your test really doesn't prove anything. The statisical issue is this: how often can a person be expected to get 10 of 13 trials correct when there are two alternatives for each trial. I don't have the time or inclination to see if 10 of 13 meets statistical significance or not, and I know it won't matter to you.

I think the results of the test were statistically significant. I said the following in post #6 of this thread:

"Perhaps you performed better statistically than you think you did. Other members can correct me if I'm wrong, but I believe 10 correct identifications in 13 trials has a p value of of 0.046, meaning there is less than a 5 percent chance your scores were random. I believe results with p values under 0.050 usually are considered statistically significant."

[shokhead]

Does anyone belive if they walked into somebodys home,sat down to listen to a cd and talk to there friend and one main was HD zipcord and the other was cardas neutral ref,they would look at there friend and say,wow,somethings wrong,it doesnt sound right? 100% NO! There is no way in hell. Well,maybe my dog would hear something,maybe not.

[musicoverall]

Does anyone belive if they walked into somebodys home,sat down to listen to a cd and talk to there friend and one main was HD zipcord and the other was cardas neutral ref,they would look at there friend and say,wow,somethings wrong,it doesnt sound right? 100% NO! There is no way in hell. Well,maybe my dog would hear something,maybe not.

I'd no sooner use my friends audio system as a reference for sonic differences than I would his eyeglasses for a vision test. My problem with the now-famous power cord DBT was that the equipment used was all borrowed. No participant in the test had that system as their frame of reference. That's why I used my own during my listening tests.

[RobotCzar]

I think the results of the test were statistically significant. I said the following in post #6 of this thread:

"Perhaps you performed better statistically than you think you did. Other members can correct me if I'm wrong, but I believe 10 correct identifications in 13 trials has a p value of of 0.046, meaning there is less than a 5 percent chance your scores were random. I believe results with p values under 0.050 usually are considered statistically significant."

You are correct that a p value of .046 is statistically significant at the (generally accepted) aphla of 0.05. You do not indicate how you calculated the value so I don't know if your 0.046 value is correct (as I didn't calculate it). It very well might be correct, but I am reluctant to just take people's word on these things (given the nature of some people who respond in this forum). If you want me to accept your value, please say how you got it. To some extent the statistical result is irrelavent as formal experimental procedures were not followed and the number of trails is very small (13). I don't think statistical significance is extremely important in tihs disucssion as the experimenter was not preprared to accept an alphal level prior to the experiement and we can all agree than 10 of 13 is a positive result in this case.

[RobotCzar]

Who said anything about "hearing" better than instruments? The presumption that a few metrics fully characterize all the audible aspects of reproducing musical content using real world speakers in real world environments is yours.
rw

I generally don't respond to e-stat as he seems to be operating with a different logic than I use. His comment here is, however, such a clear example of this difference in logic that I cannot resist.

It is very common for people of his persuasion to make up arguments that they can "win" and attribute them to his opponent (this is often called creating a "straw man"). At no time did I suggest that "a few metrics" would charaterize all aspects of reproducing musical content. I never said how many metrics are involved, and I am not referring to musical content because that is entirely in the head of the listener. Music is a subjective human concept the appreciation of which cannot be totally quatitatively characterized. Music is like "truth" or "beauty". It is astounding how many high end hobbiest seem unable to distinguish the concepts of "sound" and "music". Thus, people will describe the "musicality" of equipment, which is ridiculous.

However, a surprisingly few metrics would suffice to characterize sound reproduced by speakers. Audio equipment has the job of accurately reproducing sound (at least that is my idea of the goal). Equipment knows nothing of music and cannot tell sound from music. Music is not a phsyical entitiy and therefore is not subject to physics. Sound is a pretty simple physical phenomina (a 3-dimentional sound presure wave) that can COMPLETELY be described in terms of physical "metrics" (i.e., measurements). Air pressure at a particular location can be characterized by one number (pressure), sound is a dynamic phenomina (the change in air pressure over time). This change requires only one more metric that can be expressed as frequency or wavelength. These two numbers are why a digital sound recording format has only two parameters sampling rate (e.g., 44 KHz) and sample size (.e.g., 16 bits).

Judging sound reproduction requires more data. In this case we are comparing the accuracy of a sound that is recorded (a "live" sound) to one that is reproduced.

Now, as all rationalists, and apparently e-stat, agree that the physical environment of a sound will effect the sound (have a functional relationship to the air pressure measured), so it is very difficult to compare the accuracy of a sound produced in one environment and reproduced in another. (Thus rationalists point to the listening room and speaker location as significant factors of reproduced audio.) Measurements could tell us EXACTLY how accurate a reproduced sound is in a given environment (so e-stat is wrong), but these measurements are complicated and never provided because the speaker maker doesn't know what your environment is. Therefore, some subjectivity (using human ears and opinion) is required to judge the accuracy of speakers in real environments (i.e., we don't have the measurements for most real environments).

Because we must depend on opinion because of a lack of hard data for the wide range of home listening environments, speaker locations, and room contents DOES NOT mean that ears are superior for judging speakers than measurements, they simply are all we have in most cases. This situation also does not mean the measurements we do have are worthless, they still somewhat describe how a speaker performs, just not exactly in every environment.

Ears are very subjective in that human perception is relative and subject to ABEs because those types of ears are more useful for survival. They absolutely cannot be used for the utlimate judgement of accuracy. For one thing, people tend to like "pretty" or "dramatic" or even overly "detailed" sound rather than simply accurate sound. Also, being subjective, what one person "likes" has zero relevance to what another person likes. Accuracy, on the other hand, is objective--it is possible to measure the accuracy of a reproduced sound. I am after accuracy, not what other people like, so I do give preference to objective measurements by instruments more accurate than the human ear and not subject to the mental effects of perception.