The ABX test in the commentary

[jj]

That's a much better test than many amateur or informal tests I've read about.

There are other people (Clarke, Krueger, Muller, etc) who can run a decent ABX test as well, using the actual "ABX Box", but of course that's a bit tough when we're swapping out power cables.

I do have a couple of small criticisms of the test.

First, an ABX test should have A and B as anchors. The test subjects should know what A and B are, i.e. nothing blind UNTIL 'X' is presented.

Second, switching in under 200 milliseconds is a good thing, if one can achieve it. The lowest (most peripheral) level of auditory memory starts to decay quickly.

Third, there needed to be some kind of a control, perhaps a set level difference, or something of the sort, included as one of the test conditions, in order to verify the test.

However, given the various differences in sound stated in reviews, the test certainly should have enough sensitivity to detect the previously claimed differences, and as we can tell from the numbers, nobody heard a thing.

People who wonder what this bit about power cables is all on about can head to http://www.cardas.com/content.php?area=products&content_id=4&pagestring=Power for a tour through some aspects of High-End Powercords.

As usual, I don't comment on actual products.

[Just thinking]

Sorry ... correction made.

[jj]

Originally posted by Just thinking
Sorry ... correction made.

wha? Who? Where?

[Just thinking]

Originally posted by jj
wha? Who? Where?

No, my mistake. I overlooked something in the commentary that when I re-read it answered my question.

[Ladewig]

Originally posted by jj


Second, switching in under 200 milliseconds is a good thing, if one can achieve it. The lowest (most peripheral) level of auditory memory starts to decay quickly.


Why? I imagine that the vast majority of listeners could distinguish between $4 speakers and $4000 speakers in an ABX test even if there were 60 seconds between the listening segments (assuming no noise during that switching time). Why is it unfair to hold a $0.25 cable / $2500 cable test to a higher standard?

[Jeff Corey]

Originally posted by Ladewig
Why? I imagine that the vast majority of listeners could distinguish between $4 speakers and $4000 speakers in an ABX test even if there were 60 seconds between the listening segments (assuming no noise during that switching time). Why is it unfair to hold a $0.25 cable / $2500 cable test to a higher standard?
Sure, just tell me where to buy $4 speakers. I can recyle zip cord from old Vegamatics for a free speakers cord that can't be told from a Monster in a fair test. Your dollar comparison folds.
Speakers are a lot more complex than cord.

[jimlintott]

Two things that I would love to see in audio tests are:

Hearing test results of the participants.

A non-human participant, ie a high end microphone placed at the optimal listening point. Surely if the human participants claim audible differences then these differences should be apparent in the mic's recording?

Recently I was part of a group listening to three pairs of speakers. One pair were cheap run of the mill, another were decent and the last were a very good pair. Everyone agreed that the best pair did indeed sound best. Mixed reaction to the other two. Some felt that the best pair would definately worth the extra money (about$1000 cdn more) but others said they didn't think the difference was that great.

It is sooo subjective.

[drkitten]

Originally posted by jimlintott


Recently I was part of a group listening to three pairs of speakers. One pair were cheap run of the mill, another were decent and the last were a very good pair. Everyone agreed that the best pair did indeed sound best. Mixed reaction to the other two. Some felt that the best pair would definately worth the extra money (about$1000 cdn more) but others said they didn't think the difference was that great.


Well, part of this is becaues "worth the extra money" is such a subjective and personal judgement. That's a judgment not only of how much better the music is to you, but of how important music in general is -- and of how much money you have to spend. If that $1000 is next month's rent, then no amount of difference would be worth not having a roof over the speakers.

[CurtC]

Originally posted by jimlintott
A non-human participant, ie a high end microphone placed at the optimal listening point. Surely if the human participants claim audible differences then these differences should be apparent in the mic's recording?Don't have much experience with audiophiles, do you?

Their claims are that the music sounds different to humans, so that's what you have to test. One aspect is that tiny, hardly measurable differences could be perceived as important to the listener (according to them). The other aspect is that major differences in what test equipment could pick up (such as nonlinear phase, or harmonics above 15 kHz) would not be perceived at all by listeners. No, you have to test what they claim.

[jj]

Originally posted by Ladewig
Why? I imagine that the vast majority of listeners could distinguish between $4 speakers and $4000 speakers in an ABX test even if there were 60 seconds between the listening segments (assuming no noise during that switching time). Why is it unfair to hold a $0.25 cable / $2500 cable test to a higher standard?

Exactly where am I doing that?

I said:

However, given the various differences in sound stated in reviews, the test certainly should have enough sensitivity to detect the previously claimed differences, and as we can tell from the numbers, nobody heard a thing.


It would seem to me that I'm not doing what you think I'm doing.

I do auditory testing as part of my living. I do it in various ways that are scientifically validated, etc, with controls, etc. I am telling people what is necessary for a "best case" test. I am ALSO saying that the effects alledged in the reviews should easily be captured in the DBT that was executed if they weren't the results of attentional focusing or simple self-influence of some other sort, and you know what, that's what the numbers suggest. The subjects heard nada. Zip. Nothing.

[jj]

Originally posted by jimlintott
Recently I was part of a group listening to three pairs of speakers. One pair were cheap run of the mill, another were decent and the last were a very good pair. Everyone agreed that the best pair did indeed sound best. Mixed reaction to the other two. Some felt that the best pair would definately worth the extra money (about$1000 cdn more) but others said they didn't think the difference was that great.

It is sooo subjective.

This isn't quite on-thread, but I feel obliged to point out that I have often tested $100 loudspeakers that were better, from both the listener's perspective and the analytic measurement perspective, than a $500 loudspeaker. I've also experienced the opposite.

Somewhere in that price range, one often goes from 2-way to 3-way speakers, and very often the crossover for the 3-way sits in the middle of the frequency range where people are very sensitive to problems, AND the crossover/driver combination are not very good.

That's not the only technical problem I've experienced in loudspeakers, of course. :)

[Ladewig]

Originally posted by Jeff Corey
Sure, just tell me where to buy $4 speakers. I can recyle zip cord from old Vegamatics for a free speakers cord that can't be told from a Monster in a fair test. Your dollar comparison folds.
Speakers are a lot more complex than cord.

http://www.sciplus.com/category.cfm?subsection=17&category=158

Two 2" SpeakersÂ_
Somewhere a bunch of transister radios are waiting for speakers that aren't going to show up, because we have them here. Build a couple of your own radios, or an interesting set of earphones, with (2) of these 4-ohm, 1.5-watt speakers. Each 2" cone is already on a 2-1/8" square mounting bracket. It's 1-1/4" deep OA. Frequency response is approx 1.5 KHz to 4KHz.
31607 2" SPEAKER $3.95 / PKG(2)

I am not sure why my dollar comparison folds. I still maintain that if I hooked up these (less than) $4 speakers next to some $4000 speakers, I and virtually everyone else could with 100% accuracy distinguish between the two during an ABX test. If a price difference of three orders of magnitude doesn't produce obvious results, then something is wrong. The prices in the cable test are closer to four or five orders of magnitude different.

[Ladewig]

jj
Second, switching in under 200 milliseconds is a good thing, if one can achieve it. The lowest (most peripheral) level of auditory memory starts to decay quickly.

Ladewig
Why? I imagine that the vast majority of listeners could distinguish between $4 speakers and $4000 speakers in an ABX test even if there were 60 seconds between the listening segments (assuming no noise during that switching time). Why is it unfair to hold a $0.25 cable / $2500 cable test to a higher standard?


jj
Exactly where am I doing that?

I am trying to figure out why switching should be done in a matter of milliseconds. If we are dealing with a product that claims to be worth 10- to 100- thousand times as much as another product, then isn't any switching less than 60 seconds sufficient?

[jj]

Originally posted by Ladewig
I am trying to figure out why switching should be done in a matter of milliseconds.


Because 200 milliseconds, give or take, is how long primary loudness memory lasts. You don't have to like the fact, but there is rather a lot of experimental results showing it. It really doesn't make much sense to argue this, the experiments are done, from Jont Allen's level-ranging to experiments done just about a half-century earlier, and that's it.


If we are dealing with a product that claims to be worth 10- to 100- thousand times as much as another product, then isn't any switching less than 60 seconds sufficient?

When did I say it wasn't? I said that the test was not of ultimate sensitivity. I also said, just like I previously quoted:


However, given the various differences in sound stated in reviews, the test certainly should have enough sensitivity to detect the previously claimed differences, and as we can tell from the numbers, nobody heard a thing.


The test should address the claimed difference in sound quality, be it large or small. In my experience, audiophiles have time and again locked into small test issues to explain their failure to hear something they previously discussed as "huge", ergo it pays, as much as one can, to run a test that offers them no wiggle room.

The price of the objects has NOTHING to say about the percieved differences.

[Bodhi Dharma Zen]

I know several audiophiles who will swear for a brand of cables. It is simply NONSENSE.

When they know which cable is connected they say things like: "it opened the soundstage", "the highs are not harsh anymore", "the bass become more solid", and such subjective terminology.

Logic runs like this: If the bass is more "solid" then, in a blind test, they should be able to pick the "solid" bass from the "loose" one. Right?

But of course.... they cant. :D

[jj]

Originally posted by Bodhi Dharma Zen
I know several audiophiles who will swear for a brand of cables. It is simply NONSENSE.

Well, I know of at least 100's of people who believe that. Most of them are also implacably hostile to DBT's, and insist, with not a shred of evidence in the world, that DBT's don't work.

Need I say more?

[Ladewig]

Originally posted by jj
Because 200 milliseconds, give or take, is how long primary loudness memory lasts.

Can you provide me with a definition of "primary loudness memory"? When I look for it on Google and Altavista nothing shows up.

[Bodhi Dharma Zen]

Originally posted by jj
Most of them are also implacably hostile to DBT's, and insist, with not a shred of evidence in the world, that DBT's don't work.


I know, and I find it really interesting. What mechanism does this people have to avoid a bit of critical thinking about their hobby? It is not true that with the money spent on cables (for example) they can buy LOTS of music? ;)

I simply cant understand how some minds work :(

[Kopji]

There always seems to be an element of 'authority' in these wild claims. Double blind testing is deemed invalid because that would diminish their claim to be some kind of authority:a mysterious quality about them that sets them apart from we ordinary folk.

Is anyone surprised when we note the emperor has no clothes, it is met with opposition?

[jj]

Originally posted by Ladewig
Can you provide me with a definition of "primary loudness memory"? When I look for it on Google and Altavista nothing shows up.

Go to Yost or Moore, and read about the time response of the ear, and then go to a good psych book at read about stages of memory. I don't know where else to send you.

ETA: I think it was Elizebeth Cohen gave a talk about it a few years ago at an AES meeting, too, but memory is dim in that respect. You might look for a talk on testing, and ABX testing (watch out, there are two kinds, one much better than the other, due to the time lag issue and some other issues), from a New York AES meeting.

This isn't something that really gets much press, it's "old news". You may be having the problem that every author has invented his own name, too.

If you look in the thread on education that's in, um, I think community, I did discuss the memory issues in the auditory system a few days ago. The discusison is very schematic, but should convey more of what I'm talking about.

Now, perhaps I should explain a few things.

Intensity is the measured "sound pressure level" (not exactly, but it's a linear measurement in the air.

Loudness is the internal representation inside the listener. As
Fletcher, et al showed many years ago, this is something that is quite constant across unimpaired listeners.

Loudness in this sense is not a single number, it is a function of both bark frequency (See "Bark Scale") and time, and represents the neural firings coming down the audityr nerves almost exactly (needless to say that's much more modern work). The rate of firing is shown to be proportional (not quite linearly, of course, this is a biological thing we're talking about) to the loudness that Fletcher demonstrated nearly a century ago.

You can remember this "loudness" for a very short time, it's a peripheral position, barely into the CNS (depending on where you set the "line" it might not be), and this is where the most detail in the auditory system resides. The memory of this is what starts to fade substantially in 200 milliseconds. You can see this in a variety of work by Fletcher, Zwicker, Greenwood, etc, etc, etc... It's not exactly new work I'm refering to .

This memory is where the most fine distinctions can be cut, as demonstrated by Allen's level roving experiments (where he used level shifting instead of time to disrupt the memory), fades quickly, and is the most fragile level of auditory memory.

Jont Allen published a paper about "level roving" that is germane, you might try googling on that.

[jj]

Originally posted by Kopji
There always seems to be an element of 'authority' in these wild claims. Double blind testing is deemed invalid because that would diminish their claim to be some kind of authority:a mysterious quality about them that sets them apart from we ordinary folk.

Is anyone surprised when we note the emperor has no clothes, it is met with opposition?

After 25 years of it, not a bit of surprise. :(

There is also a desire to "putter", and digital systems, as well as modern analog pre-amps and amplifiers, destroy this, because unlike 1950's and 1960's equipment, they are to the point where the equipment that is most accurate in the technical sense is simply like all the rest except for features. N.B. This does not mean that all equipment is good, there's some real (*&(* out there, too. Anyhow, this quasi-perfection of equipment has done several things including REMOVING EUPHONIC DISTORTIONS.
(euphonic distortion: a distortion that makes things sound better)

So, we see modern puttering go in a number of directions, certainly not all of them meaningful in a technical or scientific sense.

This has been going on since the Edison phonograph, you know, there were people back then who demanded one use 'thorn' needles because they were 'natural'. It's been around as long as recorded sound.

[DevilsAdvocate]

Originally posted by Ladewig
Why? I imagine that the vast majority of listeners could distinguish between $4 speakers and $4000 speakers in an ABX test even if there were 60 seconds between the listening segments (assuming no noise during that switching time). Why is it unfair to hold a $0.25 cable / $2500 cable test to a higher standard? The quality of low end products to high end products is not a straight line. I can (probably) tell the difference between a $400 computer and a $1000 computer by just working with it for a few minutes. But telling the difference betwen a $1000 computer and a $8000 computer would take some work.

The issue is that high-end top-of-the-line stuff gives much smaller subtler changes, but the price just dramatically for those small subtle changes. This is true for cars, clothes, computers, macaroni and cheese--everything. When you go high-end, just a "smidgen better" can cost big bucks. And sometimes it is worth it if you need that. So you are right that a $.25 cable compared to a $2500 cable should be significant. But if an audiophile wants the absolute best and believes that the $2500 cable is a "smidgen better", then it may be worth it to them. It is not a question of degree of imporovement, but a question of any imporovement at all--because a true audiophile will pay anything for even a "smidgen better".

[Bodhi Dharma Zen]

Originally posted by DevilsAdvocate
So you are right that a $.25 cable compared to a $2500 cable should be significant. But if an audiophile wants the absolute best and believes that the $2500 cable is a "smidgen better", then it may be worth it to them.

This is the point. He is being robbed. There is NO evidence, whatsoever, about any kind of "micro small improvement".

Nothing, Nada. Zero.

[marting]

I really don't have any problem with this. In a sense it is a sort of voluntary redistribution system from the stupid rich to a wide range of others including retailers, publishers, workers, etc. Unlike medical quackery, or psychic readers, it does little harm.

marty

[jj]

Originally posted by DevilsAdvocate
The issue is that high-end top-of-the-line stuff gives much smaller subtler changes, but the price just dramatically for those small subtle changes.

While I'm tired of fighting about this issue, why don't you go read my notes about memory vs. the auditory system in one of those threads about writing a teaching unit on critical thinking?

You can, just by refocusing (by thought) your attention in an auditory stimulus, change what you hear substantially, in fact you do remember different things (because you listened to a different set of information that was available at the loudness level, something everyone I know of does).

Given that, to show "subtle differences" one needs DBT's, or some kind of blind, falsiable system, in order to detect differences.

[Ladewig]

Originally posted by DevilsAdvocate
The quality of low end products to high end products is not a straight line. I can (probably) tell the difference between a $400 computer and a $1000 computer by just working with it for a few minutes. But telling the difference betwen a $1000 computer and a $8000 computer would take some work.

The issue is that high-end top-of-the-line stuff gives much smaller subtler changes, but the price just dramatically for those small subtle changes. This is true for cars, clothes, computers, macaroni and cheese--everything. When you go high-end, just a "smidgen better" can cost big bucks.

I disagree with the analogy. Everyone, not just experts, can tell the difference between a $20,000 car and a $200,000,000 car. between a $400 computer and a $4,000,000 computer, between a $2 plate of macaroni and cheese and a $20,000 plate of macaroni and cheese. If you are charging one-thousand to ten-thousand times as much as the next best thing, then the difference shouldn't be so subtle that only people who can see the emperor's new clothes can tell the difference.

[DevilsAdvocate]

Originally posted by Ladewig
I disagree with the analogy. Everyone, not just experts, can tell the difference between a $20,000 car and a $200,000,000 car. between a $400 computer and a $4,000,000 computer, between a $2 plate of macaroni and cheese and a $20,000 plate of macaroni and cheese. If you are charging one-thousand to ten-thousand times as much as the next best thing, then the difference shouldn't be so subtle that only people who can see the emperor's new clothes can tell the difference. I'm not supporting any of this nonsense in any way, I'm just saying why it might be reasonable to believe in the nonsense. I don't think I am wrong in saying that there is usually a much smaller difference in quality when prices get high. Argggh. I can't think of a good analogy. It's like paying $100,000 for a car or paying $1,000,000 for car. What do you get for the extra $900,000? Not much. But people do it. Or paying $20,000 for a computer instead of $5,000, or even $2,000. Yes, you get top of the line, but it isn't really that much better. If you don't really use it to the capacity where it will make a difference, you are spending loads of money for "top of the line". But people do it anyway. For that little extra edge. Which they probably won't even use. And don't even realize. And which maybe isn't actually there. And is a complete fraud. But they will pay "top dollar" for the "top of the line".