This thread is a continuation to the ABX debate a few of us have been having in another thread.

I say that the accepted method for the ABX tests is flawed & yeilds unprovable date,i also say that the method for these tests is unreliable & is set up with one goal in mind,that goal is to prove that all amps sound alike, not to show if any differences exist.

Im of the opinion that in order for any ABX test to be taken seriously that the sound in the room needs to be measured,recorded & compared against any listener results,the ABX group see's no need to gather measurable data from the room,they also see no need to measure what the test subjects are hearing.

It is also my opinion that the ABX tests are a fraud & that no effort has been put forth to qualify the test results, with real measured data ,yet the ABX group exclaims they have a fool proof test to show that all amps sound alike.

Please keep it civil guys :yes:

Whats your feelings on the ABX, is it real or a fraud.

...I say that the accepted method for the ABX tests is set up with one goal in mind,that goal is to prove that all amps sound alike, not to show if any differences exist.

...the ABX group see's no need to gather measurable data from the room,they also see no need to measure what the test subjects are hearing.

It is also my opinion that the ABX tests are a fraud & that no effort has been put forth to qualify the test results, with real measured data ,yet the ABX group exclaims they have a fool proof test to show that all amps sound alike.

Whats your feelings on the ABX, is it real or a fraud.



Please keep it civil guys :yes:


Too late.

ABX tests, like any method for conducting research, can be done well or poorly. The point of the ABX procedure is to ensure double-blind conditions, and by itself that doesn't have bias or inherent flaws. On the other hand, there are lots of things to control, and it is true that the easiest thing to do in research is find "support" for the "null hypothesis" of no difference. I didn't vote in the poll because none of the choices capture what the ABX procedure is about. You can believe or disbelieve a conclusion, but to believe or disbelieve a procedure is like saying that a hammer is true or false - it's just a tool, that may or may not be used well and appropriately.

I'm an experimental psychologist who does laboratory research, btw.

I'm an experimental psychologist who does laboratory research, btw.

How long before you become a proper psychologist? :D

I don't even know what ABX means........

...You can believe or disbelieve a conclusion, but to believe or disbelieve a procedure is like saying that a hammer is true or false - it's just a tool, that may or may not be used well and appropriately...

I'm an avid collector of eye-opening thoughts and phrases, and that one's a keeper.:thmbsp:

An ABX test...how to describe it succinctly?

Say you are comparing CD players. CD player #1 = A, CD player #2 = B. CD player A or CD player B is selected to play a passage, however, you don't know which player has been selected so it becomes the unknown = X. Then, after listening to X you can choose to listen to A or B, going back and forth between X and A or B as many times as you want. Ultimately, you have to choose whether X = A or X = B and a score is kept as to how often you were correct in your choices.

ABX is not about which sounds better, it's about being able to indentify a difference. Many people in audio don't like ABX tests because the results pretty much conclude things like, but not limited to, differences in cable "sound" can't be identified.

If a self-proclaimed "golden ears" roars like a lion that he can hear vast differences (or even subtle ones) between amps, preamps (tone out) and cdp, and then, if he has the nerve, takes a DB ABX test, he will be very unhappy to find that he cannot tell if any differences exist at all, let alone pick a specific device, according to the results.

The results will not be statistically viable. Unless your score is 99% or one better, it can be said that no differences exist.

Just saying you can hear a diff means nothing. You'd have to prove it.

That's why the golden ears won't take the test. That's why they make excuses about all the "flaws" in ABX testing. But... They still cannot prove what they assert.

The poll shows it all, although, at this point, we need more people to vote. But, I think the results will stay the same for the most part.

Here is what could be a better alternative to ABX:

http://theaudiocritic.com/blog/index.php?op=ViewArticle&articleId=35&blogId=1

Outshined, good link!:yes:

I learned a new term, "invincible ignorance".:D

Outshined, good link!:yes:

I learned a new term, "invincible ignorance".:D

Glad you liked the article. I'm hopeful others will also take a peek. (Not very long at all, guys) but very intriguing.

The irony in all this is that the human ear is a remarkable sensory "device"


Maybe more so than any electronic device? :scratch2:

If a self-proclaimed "golden ears" roars like a lion that he can hear vast differences (or even subtle ones) between amps, preamps (tone out) and cdp, and then, if he has the nerve, takes a DB ABX test, he will be very unhappy to find that he cannot tell if any differences exist at all, let alone pick a specific device, according to the results.

Painting with a rather broad brush, aren't you? :)


http://theaudiocritic.com/blog/index.php?op=ViewArticle&articleId=35&blogId=1

I played with this a while ago. It's an interesting piece of software, and may be useful, but it has flaws too.

It assumes that the complete signal can be captured and stored. It then compares it to another captured and stored signal. The comparator outputs a difference signal, which you then listen to.

The storage and comparator are pretty easy. It's the A/D and D/A conversions that cause trouble. They may not be perfect.

Even if the conversions are perfect, there is no proof that an audible difference signal is equivalent to an audible difference between two components. (Two different meanings of difference there.)

If a self-proclaimed "golden ears" roars like a lion that he can hear vast differences (or even subtle ones) between amps, preamps (tone out) and cdp, and then, if he has the nerve, takes a DB ABX test, he will be very unhappy to find that he cannot tell if any differences exist at all, let alone pick a specific device, according to the results.

The results will not be statistically viable. Unless your score is 99% or one better, it can be said that no differences exist.

Just saying you can hear a diff means nothing. You'd have to prove it.

That's why the golden ears won't take the test. That's why they make excuses about all the "flaws" in ABX testing. But... They still cannot prove what they assert.

The poll shows it all, although, at this point, we need more people to vote. But, I think the results will stay the same for the most part.

Here is what could be a better alternative to ABX:

http://theaudiocritic.com/blog/index.php?op=ViewArticle&articleId=35&blogId=1

Proving what one asserts is easy,the steps were outlined in a previous thread & i can outline them again if anybody would like,so far ive seen nobody step up to the plate & do some of their own testing with measurements to back up what they assert,just more links to other peoples positions in order to reinforce their own standing.

Ive made my mind up on the subject from my own testing ,with measurements, to prove to myself if im really hearing anything or just imagining a difference.

Who picked the percentage that a test taker must get right in order to pass the test,even in medical testing the percentages are not 99%,could it be that the challenge is skewed to show a specific result:scratch2:

I'd also like to know why the ABX crowd of followers so adamantly against measuring what the ear is hearing? Why is it so evil to measure the speaker response from 2 like amplifiers with all levels matched.

The way the current ABX tests are set up the only thing proved is that human auditory memory is very weak & in most cases not able to remember detail.

Proving what one asserts is easy,the steps were outlined in a previous thread & i can outline them again if anybody would like,so far ive seen nobody step up to the plate & do some of their own testing with measurements to back up what they assert,just more links to other peoples positions in order to reinforce their own standing.

Ive made my mind up on the subject from my own testing ,with measurements, to prove to myself if im really hearing anything or just imagining a difference.

Who picked the percentage that a test taker must get right in order to pass the test,even in medical testing the percentages are not 99%,could it be that the challenge is skewed to show a specific result:scratch2:

I'd also like to know why the ABX crowd of followers so adamantly against measuring what the ear is hearing? Why is it so evil to measure the speaker response from 2 like amplifiers with all levels matched.

The way the current ABX tests are set up the only thing proved is that human auditory memory is very weak & in most cases not able to remember detail.

ABX testing results cannot be skewed; the results are from double blind testing. I don't know how much more un-biased any test can be.

OK, if you don't like 99%, I'll be OK with 95%, but not less. Statistics are not just ramdomly picked numbers. It is a science unto itself. Very diciplined.

Measuring the output of a speaker, or pair of speakers is something that I've never heard being done, or even eluded to. Too many interactive variables to SKEW the results.

One can listen to A or B and take all the time one wants to get a feel for what one hears. This lessens the short sonic memory affect.

This is nothing more than the "subjective" v "objective" school of thought.

If you want subjective, just say you can hear differences without scientific proof, and stick to reading Stereophile.

If you want objective, take a DB ABX (or, read the article in the link I provided and try that, although in all fairness, this new technology makes it difficult to implement right off. But it seems very valid to me, by credentialed scientists) and see what the science tells you.

ABX is not some monster under the bed. I've never heard of anyone passing a ABX test with statistically meaningful results. If the differences in sound between 2 like amps are that obvious, I would think someone would have, and should have, passed the test.

If anyone can point the way to refute this, please let us know.

The basic problem with ABX testing is that it usually doesn't produce desired results - that ridicuously overpriced "high end" equipment sounds better than "mediocre" equipment. I usually don't participate in these tests because I don't like the way they are conducted (too much pressure and not a relaxed atmosphere). I still think they are basically valid. If you are honest, you don't need the "X". Plain old AB testing works just as good.

Measuring the output of a speaker, or pair of speakers is something that I've never heard being done, or even eluded to. Too many interactive variables to SKEW the results.

We all know that speaker output is not measured ever,the variables you keep refering to are meaningless,as with any other test of anything on the planet, a variable ceases to be a variable if both test subjects are subjected to the same variable.At that point the variable becomes a constant.

Does it not?

If you want subjective, just say you can hear differences without scientific proof, and stick to reading Stereophile.


Is it possible for you to write one post on this subject without resorting to the use of catch words like "golden eared" or the now famous "stereophile" reader line,it stoped being cute about 100,000 posts ago on the forums.

You've yet to answer the question ive posed,have you ever taken any measurements from 2 like amplifiers with all levels matched.

BTW, you may be shocked to know, ive never owned a single audio mag nor do i own botique cables but i do own some very basic testing equipment which is something i recomend to all.

We all know that speaker output is not measured ever...

I don't follow what you mean or where you're going with this.

One of the prime points, afaik, of ABX is to mitigate effects that might steer people towards a foregone conclusion.

I've never seen anyone suggest ABX for a set of speakers or other things that are generally undisputed as having a clearly audible difference.

That's like suggesting a tube amp should be ABX'ed against SS amp. What's the point since most people can accept with little coaxing there is a pretty good chance they DO sound different?

ABX is not some monster under the bed. I've never heard of anyone passing a ABX test with statistically meaningful results. If the differences in sound between 2 like amps are that obvious, I would think someone would have, and should have, passed the test.

If anyone can point the way to refute this, please let us know.

The way has been pointed but you dont like it,a simple measurement would make quick work of backing up any test results.

True: the human ear is a very sensitive measuring instrument.
Also, the brain doesn't retain detail between comparisons.
It's all subjective. No sound is perceived exactly the same by different people.

It may be a viable test in the double-blind, but there's just too much hair-splitting in my opinion.

If two components are so close in their performance, then I'd just buy the one I liked.

Rick

I don't follow what you mean or where you're going with this.

One of the prime points, afaik, of ABX is to mitigate effects that might steer people towards a foregone conclusion.

I've never seen anyone suggest ABX for a set of speakers or other things that are generally undisputed as having a clearly audible difference.

That's like suggesting a tube amp should be ABX'ed against SS amp. What's the point since most people can accept with little coaxing there is a pretty good chance they DO sound different?

Ok,were going in circles here,were not comparing tubes to SS or anything else.

The whole point is this,the whole debate about amplifiers is do amplifiers sound different,the ABX test is used ad nauseum to say that differences in properly designed amplifiers do not exist to a point where they make a meaningfull difference in the overall system performance.

Ok now stay with me on this.

Anytime anybody mentions that the overall difference in their system was caused by a different amplifier their conclusion is instantly discarded & dismissed as unfounded & unscientific because they cant prove it,the only existing proof is a Blind A/B test or an ABX test,which both rely 100% on HUMAN auditory memory which is not proof but merely observation.

The sound that any test subject is hearing in an ABX or A/B test is obviously from a pair of speakers,any variables like room conditions are no longer variables because the speaker system is affected by the same variables & so are both amplifiers,these variables have been taken out of the equasition in any A/B or ABX test & are now called a constant.If the sound being heard is now reliable enough to be a constant then it can & should also be measured for any differences in the spectrum.

If the differences in amplifiers are so small then why is taking any measurements of speaker output fought so tiresly,remember now that any variables that might apply were turned into a constant by the test control protocol ,so why would these so called variables that you refer to only apply to taking a measurement & not to the human hearing.

Please explain how the variables differ from listener experience in an ABX to measured data.

True: the human ear is a very sensitive measuring instrument.
Also, the brain doesn't retain detail between comparisons.

Good point,the detail is exactly what im refering to,when a simple measurement is taken,at the speaker,differences in the audible spectrum can & will show up.

Seems like that testing would be as affective as trying to identify the different stars in the sky by the naked eye along over and over again with some reliability. Since most audio gear I listen too sounds relatively the same, I make my gear choices by sound, looks and by price. :scratch2:

...is like saying that a hammer is true or false - it's just a tool...

I'm in complete agreement with Fisherdude on this one. RichPA, any restrictions on using this phrase :D?

When you can't enjoy listening to a good system simply because of an ABX 'failure', then what's the point? I'd rather not know, and they can send me to hell as I bask in my ignorance.

I'm in complete agreement with Fisherdude on this one. RichPA, any restrictions on using this phrase :D?

Help yourself :)

If the differences in amplifiers are so small then why is taking any measurements of speaker output fought so tiresly,remember now that any variables that might apply were turned into a constant by the test control protocol ,so why would these so called variables that you refer to only apply to taking a measurement & not to the human hearing.



Can't say as I recall being part of an ABX debate where the pro-ABX people said measurements were off limits. From my recollection of past discussions I've been in, the ABX side and pro-measurement side are one in the same.

Now, it could be debated what the relevance of the measurement is but that's a whole different ballgame. For example, if a measurement show there is a difference of 0.1dB at 50kHz, is it at all relevant to what a person can hear?

In the end, the arguments always end up the same - the pro-ABX side insisting you can't prove you hear a difference, and the anti-ABX side insisting you can't prove I don't. Of course, negative proof is essentially impossible, therein lies the logical fallacy of the latter (negative proof) argument.

A good amp should sound like nothing, and it seems they do. I keep an open mind about things, generally, but a double-blind test done well is pretty much irrefutable. You can either hear a difference, or you can't. If you can't pick a winner a reasonable amount of the time, then you just can't hear the difference, you only think that you can. It really does seem pretty basic - nobody, anywhere has ever passed a double-blind test done properly.

In the end it doesn't matter -- buy what you like, enjoy it for what it is, or what you think it is. I could never be bothered to take a double blind test because I just don't care enough; I buy what I like based on it's appearance and whether or not I like the sound. Enjoy your stuff guys, and don't sweat over whether it sounds better/worse than an Adcom, Marantz, Bryston or Sansui...

A good amp should sound like nothing, and it seems they do. I keep an open mind about things, generally, but a double-blind test done well is pretty much irrefutable. You can either hear a difference, or you can't. If you can't pick a winner a reasonable amount of the time, then you just can't hear the difference, you only think that you can. It really does seem pretty basic - nobody, anywhere has ever passed a double-blind test done properly.

In the end it doesn't matter -- buy what you like, enjoy it for what it is, or what you think it is. I could never be bothered to take a double blind test because I just don't care enough; I buy what I like based on it's appearance and whether or not I like the sound. Enjoy your stuff guys, and don't sweat over whether it sounds better/worse than an Adcom, Marantz, Bryston or Sansui...

Very nicely said :thmbsp:

Any amp worth its' salt should neither add, nor subtract from the original source. It merely amplifies the signal to speaker level, and, therefore, can be considered a "black box."

Some consideration should be given when buying a new or older amp should be only its' ability to drive difficult speaker loads; you may need one that can drive 2 Ohms or even less. But, this is the exception, not the rule. Most decent amps can drive almost any speaker.

No one has ever passed a DB ABX test because the differences in sound, if they exist at all, are so infinitesimal that it becomes a non-issue.

This is my position, and my opinion. Just like no one can ever change my mind when it comes to politics, no one can change my mind about this, either.

I would say the same is true for the believers in perceptable sound diffs in amps. They will always believe it. And they have every right to, if they wish. Just as I am entitled to my opinion.

That's like suggesting a tube amp should be ABX'ed against SS amp. What's the point since most people can accept with little coaxing there is a pretty good chance they DO sound different?

Except that ABX'ers do consider tube amps to be indistiguishable from SS amps.

IMO it's unreasonable to set a 95% level of accuracy as the statistically significant level for choosing between amps.

It doesn't have to be a huge percentage, just has to be a statistically significant number, ie; a percentage that is very unlikely to have been produced by sheer chance. If you could pick a winner 60% of the time, then you can more than likely hear a difference -- 60% is not a coin toss or a wild guess (as long as the number of samples is adequate, of course).

Except that ABX'ers do consider tube amps to be indistiguishable from SS amps.

IMO it's unreasonable to set a 95% level of accuracy as the statistically significant level for choosing between amps.

Perhaps some do, but generally not in the ABX debates I've read though. Most of those have been on a different forum.

I'm not a statistician, but I do sometimes work with them and levels such as you mention are necessary to establish a high degree of confidence in the results. Anything much less results in essentially maybe, maybe not. And, all that does is to foster the negative proof argument - "You can't prove I don't hear it, therefore it is true that I can hear it.".

EDIT: The main reason, IMO, ABXer's insist on ABX is to get rid of as much "maybe" factor as possible. 60% leaves lots of "maybe" wiggle room.

How long before you become a proper psychologist? :D

You koalas crack me up!

:thmbsp:

Any amp worth its' salt should neither add, nor subtract from the original source. It merely amplifies the signal to speaker level, and, therefore, can be considered a "black box."

This is nearly a direct quote from the product trainings I do. The only job of the equipment in the playback chain is to faithfully reproduce the source material, period.

A good amp should sound like nothing, and it seems they do. I keep an open mind about things, generally, but a double-blind test done well is pretty much irrefutable. You can either hear a difference, or you can't. If you can't pick a winner a reasonable amount of the time, then you just can't hear the difference, you only think that you can. It really does seem pretty basic - nobody, anywhere has ever passed a double-blind test done properly.

In the end it doesn't matter -- buy what you like, enjoy it for what it is, or what you think it is. I could never be bothered to take a double blind test because I just don't care enough; I buy what I like based on it's appearance and whether or not I like the sound. Enjoy your stuff guys, and don't sweat over whether it sounds better/worse than an Adcom, Marantz, Bryston or Sansui...

....what I stated in post 23. It's all subjective in the long run.

Now, it could be debated what the relevance of the measurement is but that's a whole different ballgame. For example, if a measurement show there is a difference of 0.1dB at 50kHz, is it at all relevant to what a person can hear?

I guess for me the relevance of measuring speaker response is along the same lines as your example,for me the differences are relevant to what each person can hear, with a test that includes measurements the measurements just might show that there is some truth to what people say about needing some time with an amp in their system to hear any differences.

I wasnt a firm believer that amps can sound entirely different until i started measuring my own amps response from the speaker,then i became a believer.

I'm one of those cursed with good musical ears, have to say what? alot if someone is talking to me and can't hear the phone in the other room, but I can here the differance in amps, and this is with nice fat rat-shack cables,can't even think about getting crazed on the fancy wires, I wouldn't get any sleep.
Craig

RichPA is right on here. For instance, as part of ABX, there should be a preliminary phase of "testing the test" just to make sure that "A" and "B" sides have no audible differences due to wiring, switch contact differences, etc. You simply switch sources a number of times and see if peoples' choices tend to follow one side. You can also use the same source for both A and B and see if one side is chosen most of the time, or if people just tell you they can't hear a difference. But this is why I never ABX. It's just too much time and effort to do it justice.

It has flaws but is necessary.