IPB

Welcome Guest ( Log In | Register )

3 Pages V   1 2 3 >  
Reply to this topicStart new topic
Positive ABX test of loudspeaker cables
hybris
post Jul 18 2011, 21:05
Post #1





Group: Members
Posts: 97
Joined: 24-March 09
From: Oslo
Member No.: 68343



Hi,

Just wanted your thoughts about a recent ABX test done be a few members(users) of the norwegian audio forum www.avforum.no.

The test was an ABX-test with three participants, each participant did 10 trials.

The audio system was:
System Audio Explorer Master
2x Electrocompaniet AW180 monoblock
Electrocompaniet EC4.9
Electrocompaniet ECC 1 CD player
Hegel HD20 DAC


The cables tested was:
*Noname 4mm cable (2x3 meters) with banana plugs, price approx 100USD.
*Nordost Baldur (2x3 metres), price approx 1000USD

To get to the point, the result was as follows:

Participant 1 was able to correctly identify X 7 out of 10 times
Participant 2 was able to correctly identify X 8 out of 10 times
Participant 3 was able to correctly identify X 7 out of 10 times


I'm no expert in either statistics or mathematics (and I'm also a cable sceptic), and as far as I understand none of these results are statistically significant. But that all three tests should have a relatively high number of correct answers is interesting, and if we add the results together (I don't know if that is statistically correct), you would get 22 out of 30 correct, hvis IS statistically significant.

All in all it looks like a plausible indication that it is possible to discern loudspeaker cables from each other. There could of course be a number of weaknesses to the test itself and I'm awaiting some additional information on how the test was conducted, but preliminary information indicates that the test concept was reasonable.

Any immediate questions or comments?


--------------------
Thorbjorn
Go to the top of the page
+Quote Post
DVDdoug
post Jul 18 2011, 22:09
Post #2





Group: Members
Posts: 2565
Joined: 24-August 07
From: Silicon Valley
Member No.: 46454



I remain skeptical... I'm sure I wouldn't hear a difference. (But, I might guess 7 of 10 correctly!)

It's "interesting" that none of the participants did better than 8 of 10, or worse than 7 of ten.
Go to the top of the page
+Quote Post
hybris
post Jul 18 2011, 22:15
Post #3





Group: Members
Posts: 97
Joined: 24-March 09
From: Oslo
Member No.: 68343



QUOTE (DVDdoug @ Jul 18 2011, 23:09) *
I remain skeptical... I'm sure I wouldn't hear a difference. (But, I might guess 7 of 10 correctly!)

It's "interesting" that none of the participants did better than 8 of 10, or worse than 7 of ten.


I might guess 7/10 as well, but three times in a row? I'm sceptical as well, and it may well be a weakness in the test. That would be the first assumption.

But assuming the test methodology is correct, I think it's an interesting result. Hopefully the test methods will be confirmed, and the test repeated for verification.


--------------------
Thorbjorn
Go to the top of the page
+Quote Post
greynol
post Jul 18 2011, 22:40
Post #4





Group: Super Moderator
Posts: 10000
Joined: 1-April 04
From: San Francisco
Member No.: 13167



Without more information, we're left to speculation, which to me makes these these types of discussions are a complete waste of time.

I respectfully suggest people refrain starting such threads when there is such a blatant lack of pertinent information in the future.

Please take my comments from the point of view of a regular member and not a moderator. AFAIC, there is nothing in violation of the TOS regarding the content of this discussion, though it is questionable that it rises to a level deserving placement in the listening tests sub-forum; hearkening back to my original concern.

This post has been edited by greynol: Jul 18 2011, 22:42


--------------------
Concern trolls: not a myth.
Go to the top of the page
+Quote Post
hybris
post Jul 18 2011, 22:52
Post #5





Group: Members
Posts: 97
Joined: 24-March 09
From: Oslo
Member No.: 68343



QUOTE (greynol @ Jul 18 2011, 23:40) *
Without more information, we're left to speculation, which to me makes these these types of discussions are a complete waste of time.

I respectfully suggest people refrain starting such threads when there is such a blatant lack of pertinent information in the future.

Please take my comments from the point of view of a regular member and not a moderator. AFAIC, there is nothing in violation of the TOS regarding the content of this discussion, though it is questionable that it rises to a level deserving placement in the listening tests sub-forum; hearkening back to my original concern.


I understand your concern, I just thought it would be interesting to get input from hydrogen audio with respect to possible weaknesses in the test. I also think it is an interesting topic, as documented ABX tests of loudspeaker cables (negative or positive) are quite few and far between, probably because they are quite time consuming. I also assume that there are a number of negative tests performed that we never hear about.

I also understand that information is somewhat lacking. I will attempt to get more information from the guys who did the test and supply it here (for those that are interested).

If a moderator thinks that the thread is better suited for another sub-forum, please move it. smile.gif


--------------------
Thorbjorn
Go to the top of the page
+Quote Post
greynol
post Jul 18 2011, 23:11
Post #6





Group: Super Moderator
Posts: 10000
Joined: 1-April 04
From: San Francisco
Member No.: 13167



I look forward to reading more information. Right now I can't take the results seriously; not when people can easily come up with multiple possible reasons that would invalidate the test.

This post has been edited by greynol: Jul 18 2011, 23:11


--------------------
Concern trolls: not a myth.
Go to the top of the page
+Quote Post
Notat
post Jul 18 2011, 23:30
Post #7





Group: Members
Posts: 581
Joined: 17-August 09
Member No.: 72373



We recently discussed a similar statistical result. Yes, as long as the participants made their decisions independently, you can combine the results and improve the case for significance. If so, there is something going on here. It could be a flaw in the test or it could be a real audible difference. The sound of cables is also something we've recently discussed. In the end, one should keep in mind that different doesn't necessarily indicate better. Determining which is better requires a different king of testing.

This post has been edited by Notat: Jul 18 2011, 23:31
Go to the top of the page
+Quote Post
pdq
post Jul 19 2011, 00:31
Post #8





Group: Members
Posts: 3391
Joined: 1-September 05
From: SE Pennsylvania
Member No.: 24233



There is one characteristic of speaker cables that can definitely make an audible difference, and that is resistance. How closely matched in resistance were these cables?
Go to the top of the page
+Quote Post
greynol
post Jul 19 2011, 00:51
Post #9





Group: Super Moderator
Posts: 10000
Joined: 1-April 04
From: San Francisco
Member No.: 13167



Series inductance and parallel capacitance will have a low-pass effect. Whether the resulting corner is low enough for this to be audible is the real question.

We really need to determine the nature of the test before diving into hypotheticals.

For the discussion to meander like this is rather pointless.


--------------------
Concern trolls: not a myth.
Go to the top of the page
+Quote Post
RobWansbeck
post Jul 19 2011, 01:32
Post #10





Group: Members
Posts: 23
Joined: 24-January 10
Member No.: 77468



Perhaps the way to look at this is that the listeners could not reliably distinguish between a no-name cable and an audiophile cable.
Go to the top of the page
+Quote Post
pdq
post Jul 19 2011, 02:44
Post #11





Group: Members
Posts: 3391
Joined: 1-September 05
From: SE Pennsylvania
Member No.: 24233



The term "reliably distinguish" is a bit vague. Do you mean with 100% accuracy?
Go to the top of the page
+Quote Post
andy o
post Jul 19 2011, 02:45
Post #12





Group: Members
Posts: 1311
Joined: 14-April 09
Member No.: 68950



Which "noname" cable costs 100US for 6 meters?
Go to the top of the page
+Quote Post
mixminus1
post Jul 19 2011, 04:04
Post #13





Group: Members
Posts: 688
Joined: 23-February 05
Member No.: 20097



My thoughts exactly - that's about $5.25/ft.

4 mm^2 is right in between 12 and 11 AWG, and I think you'd be hard-pressed to find 12 gauge zip cord that cost much more than $0.60/ft.

Decent banana plugs are a few dollars apiece - $10/ea. if you really want to get fancy - so yes, there's quite a discrepancy there.

In any event, as greynol has observed, we have no idea how the test was actually conducted - first and foremost, how was the switching accomplished?


--------------------
"Not sure what the question is, but the answer is probably no."
Go to the top of the page
+Quote Post
hybris
post Jul 19 2011, 08:45
Post #14





Group: Members
Posts: 97
Joined: 24-March 09
From: Oslo
Member No.: 68343



The "noname" cable was a 4mm^2 cable from a Norwegian company "Biltema" (www.biltema.no) that sells more or less everything. The cable can be bought in a roll of 10 metre for 219 NOK = 39USD with the current exchange rate. So 3.9 USD per metre. (link to the cable: http://www.biltema.no/no/Bil---Mc/Bilstere...oyttalerkabel/). It had identical nordost bananaplugs as the Nordost cable soldiered on.

So the actual cable 3.9USD X 6 metres (not counting the price of the banana plugs) was about 23.4USD. My mistake.


During the test they first listened to A, then B (without being informed wich was nordost and which was biltema), and then X, and were asked to determine if X was A or B. Each participant was allowed to choose what music to listen to, and for how long.

They were also allowed to listen to the cables sighted for a while before conducting the actual ABX test.

The participants did not do 10 listening tests in a row, the listeners alternated between them so that the listeners got some rest in between the tests.

PRELIMINARY INFORMATION:
Before the test it was claimed that it would be conducted as follows (this has not yet been confirmed after the test):

  • The listener is alone in the listening room while doing the ABX (There are a separate room for waiting to do the test)
  • The song that is played is selected by the listener, and played back at the same volume during each test
  • The listener is forced to put on ear phones before cable switching - making him unable to hear the switch
  • The listener is either blindfolded, or the change of cables is done behind a curtain, making him unable to see the switch
  • The listener is (by his own discretion) allowed to listen to parts of or an entire song
  • There is no talk between the listeners in between tests.


When it comes to the question about resistance, I would assume that both cables are thick enough to make this a small (non-existant?) issue at 3 metres. If not we should have a lot more positive tests out there. smile.gif


More information will hopefully follow.


--------------------
Thorbjorn
Go to the top of the page
+Quote Post
DonP
post Jul 19 2011, 12:23
Post #15





Group: Members (Donating)
Posts: 1471
Joined: 11-February 03
From: Vermont
Member No.: 4955



QUOTE (hybris @ Jul 18 2011, 16:05) *
To get to the point, the result was as follows:

Participant 1 was able to correctly identify X 7 out of 10 times
Participant 2 was able to correctly identify X 8 out of 10 times
Participant 3 was able to correctly identify X 7 out of 10 times


I'm no expert in either statistics or mathematics (and I'm also a cable sceptic), and as far as I understand none of these results are statistically significant. But that all three tests should have a relatively high number of correct answers is interesting, and if we add the results together (I don't know if that is statistically correct), you would get 22 out of 30 correct, hvis IS statistically significant.


I find it interesting that all 3 listeners had such close results. Did they tend to get it right on the same trial numbers? If so, then look for a difference between trials instead of a difference between cables.

Was the choice of X random (ie coin flip) or a test administrator selecting? Did all listeners get the same order?

In format, did a trial consist of one listen to A, B, and X, or could the listener ask for repeats? If, for example, he asks to hear B and then X you have to take care that he doesn't get clues from a shorter transition time if X is B, so no cable change would be required (or a disconnect/reconnect takes less time).

As was discussed in a recent topic of analog vs digital, double-blind is important. That is, there is no exposure to any one who knows which is which.

The phones to prevent hearing the switch, are they sufficient? These are usually meant to protect hearing, not to keep you from hearing at all, typically between 20 and 30 dB of isolation.

One way to skew results is picking the best ones. So was there a total of 3 participants, or is this the best 3 results out of more?


Go to the top of the page
+Quote Post
andy o
post Jul 19 2011, 15:21
Post #16





Group: Members
Posts: 1311
Joined: 14-April 09
Member No.: 68950



Link to the original thread, maybe, even if it's in another language? Also the fact that it wasn't immediate switching makes this even more suspicious, IMO. Also, were the switches completely random (as in throwing a coin or similar, instead of being at discretion of the testers)?
Go to the top of the page
+Quote Post
krabapple
post Jul 19 2011, 16:41
Post #17





Group: Members
Posts: 2221
Joined: 18-December 03
Member No.: 10538



Other considerations:

Looks like a 'homemade' cable ('banana plugs soldered on') versus a 'phile jewelry cable. Could use some measurement data on both cables themselves. Maybe the 'audiophile' cable is intentionally 'colored'. Maybe the soldering on the homemeade cable isn't good
.

Also, was there any plausible way the cable switching person could have cued the subjects (consciously or not)?

This post has been edited by krabapple: Jul 19 2011, 16:41
Go to the top of the page
+Quote Post
hybris
post Jul 19 2011, 19:05
Post #18





Group: Members
Posts: 97
Joined: 24-March 09
From: Oslo
Member No.: 68343



A little more information:


*A trial consisted of one listen to A, B and X, no repeats.
*There was a total of three participants, no results have been removed or hidden
*I'm not sure what the point was with soldering on banana plugs on the cheap cables, I have asked about that
*The order of X was not the same for each listener.
*It has not been fully explained how the change was made, but it does not appear to be full double blind.

The first tester guessed correct on the cheap cable 3 times and the expensive cable 4 times.

The second tester guessed correct on the cheap cable 5 times and the expensive cable 3 times

The third tester guessed correct on the cheap cable 4 times and the expensive cable 3 times

Link to the original thread (long):
http://avforum.no/forum/kabler/113328-skik...-avsluttet.html

Link to post with an excel sheet with the results. Probably understandable even though it is in norwegian, let me know if not:
http://avforum.no/forum/kabler/113328-skik...tml#post1702429


A complete explanation of the test methodology has not yet been posted, I am suspecting that the importance of all these factors is not entirely evident of the guys who did the test, but I'm doing my best to get the relevant information.

My assumption so far is that there is some kind of weakness present here. The most significant so far being that it appears that the guy doing the switch knew which cables were being played, and did the switch while the listener was in the room. In what sense the listener was deprived of sight/hearing while the switch was made has not yet been fully explained.

This post has been edited by hybris: Jul 19 2011, 19:05


--------------------
Thorbjorn
Go to the top of the page
+Quote Post
hybris
post Jul 19 2011, 21:25
Post #19





Group: Members
Posts: 97
Joined: 24-March 09
From: Oslo
Member No.: 68343



Some more information:

It was not a double blind, the guy who switched cables knew what was connected.

The cable switch took about 15 seconds each time, and the cables were "switched" even though it was the same cable that was about to be listened to again. As the cables had identical banana plugs the switch was identical both in timing and "sound" when plugged into the sockets.

The switch was done behind a curtain, and the listeners never saw any of the cables before or during the test (which were referred to as cable A or cable B at all times), so they had no practical way of identifying the cables.

The reason why the cheap cables was supplied with identical banana plugs as the expensive ones was to make the test (everything was the same except the actual cable) and switch as identical as possible.

The sequence of X was made up during the test by the guy who switched cables, it was not predetermined (but not randomized by flipping a coin either).



As far as I can see it seems to be a pretty well carried out test, but it was not double blind, so you do have the issue of someone who know what was connected being in the room with the listener during the cable switch (the guy who switched left the room during the actual listening, he just went in to switch cable between A, B, X and then left again between each one).


--------------------
Thorbjorn
Go to the top of the page
+Quote Post
andy o
post Jul 19 2011, 21:34
Post #20





Group: Members
Posts: 1311
Joined: 14-April 09
Member No.: 68950



That doesn't seem very rigorous at all, nor very well thought out.
Go to the top of the page
+Quote Post
hybris
post Jul 19 2011, 21:43
Post #21





Group: Members
Posts: 97
Joined: 24-March 09
From: Oslo
Member No.: 68343



QUOTE (andy o @ Jul 19 2011, 22:34) *
That doesn't seem very rigorous at all, nor very well thought out.


What are the obvious shortcomings, besides the fact that it wasn't double blind? Any reasonable suggestions for improvements are welcome, as another test is being planned for a later date to confirm the result.

This post has been edited by hybris: Jul 19 2011, 21:44


--------------------
Thorbjorn
Go to the top of the page
+Quote Post
DigitalDictator
post Jul 19 2011, 22:49
Post #22





Group: Members
Posts: 313
Joined: 9-August 02
From: SoFo
Member No.: 3002



QUOTE (hybris @ Jul 19 2011, 09:45) *
The "noname" cable was a 4mm^2 cable from a Norwegian company "Biltema" (www.biltema.no) that sells more or less everything.

Actually, Biltema is a Swedish company tongue.gif

This post has been edited by greynol: Jul 19 2011, 22:56
Reason for edit: Trimmed unnecessarily long quote, please consider being more considerate in the future.
Go to the top of the page
+Quote Post
RobWansbeck
post Jul 20 2011, 00:50
Post #23





Group: Members
Posts: 23
Joined: 24-January 10
Member No.: 77468



QUOTE (pdq @ Jul 19 2011, 01:44) *
The term "reliably distinguish" is a bit vague. Do you mean with 100% accuracy?


Although the overall result was statistically significant none of the individual listeners could achieve a statistically significant result. This is at odds with the advertising blurb, e.g.

“ The high tolerance conductor spacing used in this cable means that musical performance reaches new heights. For detail, resolution and transparency Baldur will surprise and delight you. “

One would have hoped that a cable that surprises and delights would have been more detectable.

Suggestions for improvements would be for more than ten tests to see if individual listeners could achieve a statistically significant score and for listeners to attempt to describe the difference. For example, was cable X more delightful or was it a bit louder.
Go to the top of the page
+Quote Post
hellokeith
post Jul 20 2011, 09:13
Post #24





Group: Members
Posts: 288
Joined: 14-August 06
Member No.: 34027



A bit simplified, but: p=.5, n=30, x=22 --> P(x) = .0055

I would say at the least that it merits positive correlation.
Go to the top of the page
+Quote Post
hybris
post Jul 20 2011, 09:19
Post #25





Group: Members
Posts: 97
Joined: 24-March 09
From: Oslo
Member No.: 68343



QUOTE (RobWansbeck @ Jul 20 2011, 01:50) *
QUOTE (pdq @ Jul 19 2011, 01:44) *
The term "reliably distinguish" is a bit vague. Do you mean with 100% accuracy?


Although the overall result was statistically significant none of the individual listeners could achieve a statistically significant result. This is at odds with the advertising blurb, e.g.

“ The high tolerance conductor spacing used in this cable means that musical performance reaches new heights. For detail, resolution and transparency Baldur will surprise and delight you. “

One would have hoped that a cable that surprises and delights would have been more detectable.

Suggestions for improvements would be for more than ten tests to see if individual listeners could achieve a statistically significant score and for listeners to attempt to describe the difference. For example, was cable X more delightful or was it a bit louder.


They should of course have more listeners and more tests, but they spent all day to do this many tests, so doing larger tests effectively is a challenge.

One of the listeners have provided their experience, in Norwegian - but a short summary of his impressions:

Before the test he did not think it was possible to tell the difference. If it was any difference, he assumed it would be that one of the cable sounded more "muddy" or less clear than the other.

Listening material: Hanne Boel, CD ”Black Wolf” – Medley Records. Track: ”Black Wolf”, the first two minutes.

During the sighted listening before the test, he did actually think there was a small difference, but it was small and it was hard to put a finger on exactly what was different. He had to really concentrate to tell the difference.

If he was pressed to explain what the difference was, he'd say that one of the cables had a more spacious sound, and somewhat crisper in some parts of the track. None of the cables sounded muddier than the other, and all instruments and the frequency response sounded the same.

During the test:
The first few times he felt fairly confident that he was able to guess correctly, but it became progressively harder to keep his concentration up. A break with some food helped. Even though he thought he could tell the difference during the test, he was surprised that he had been able to guess correct that many times.

His conclusion was that based on the test it appears that there may be a small difference, but it is so small that he does not find that it justfies the price of expensive cables compared to other investments as room treatment and other components.


--------------------
Thorbjorn
Go to the top of the page
+Quote Post

3 Pages V   1 2 3 >
Reply to this topicStart new topic
1 User(s) are reading this topic (1 Guests and 0 Anonymous Users)
0 Members:

 



RSS Lo-Fi Version Time is now: 23rd August 2014 - 14:46