setting up ABX tests at home

Re: setting up ABX tests at home

"bear" <bearlabs@netzero,net > wrote in message
news:fp582c01a2c@news1.newsguy,com

> Put it another way, IF the distortions present are of a
> particular type or class then you can put anything you
> like into the signal chain of lower or nil distortion and
> most likely not be able to identify any change. Since you
> have no way to determine the significance or type/class
> of distortions in your system, you don't know anything
> about the results of an ABX test other than the results
> being SPECIFIC to the test you performed.

This problem exists for any listening test. The only reason why it is more
apparent in DBTs is that DBTs clearly detect listeners who can't reliably
detect differences for whatever reason.

We've got yet another case where someone is specifically complaining about
ABX tests when in fact the problem exists for *any* listening test. The
reason why people don't notice this problem with sighted evalautions is the
fact that sighted evaluations have no means for addressing the problem of
false positives.

Re: setting up ABX tests at home

Arny Krueger wrote:
> "bear" <bearlabs@netzero,net > wrote in message
> news:fp582c01a2c@news1.newsguy,com
>
>> Put it another way, IF the distortions present are of a
>> particular type or class then you can put anything you
>> like into the signal chain of lower or nil distortion and
>> most likely not be able to identify any change. Since you
>> have no way to determine the significance or type/class
>> of distortions in your system, you don't know anything
>> about the results of an ABX test other than the results
>> being SPECIFIC to the test you performed.
>
> This problem exists for any listening test. The only reason why it is more
> apparent in DBTs is that DBTs clearly detect listeners who can't reliably
> detect differences for whatever reason.
>
> We've got yet another case where someone is specifically complaining about
> ABX tests when in fact the problem exists for *any* listening test. The
> reason why people don't notice this problem with sighted evalautions is the
> fact that sighted evaluations have no means for addressing the problem of
> false positives.
>

We can agree then that there are no tests that actually take into account this
issue, therefore all tests thus far are suspect in terms of the significance of
the test results?

So we can cease claiming that ABX tests have results that are generalizable?

Good.

I agree that ABX tests *can be* and *may be* a useful tool amongst many, nothing
more.

- -bear

Re: setting up ABX tests at home

"bear" <bearlabs@netzero,net > wrote in message
news:fp9jvb01hp7@news5.newsguy,com
> Arny Krueger wrote:
>> "bear" <bearlabs@netzero,net > wrote in message
>> news:fp582c01a2c@news1.newsguy,com
>>
>>> Put it another way, IF the distortions present are of a
>>> particular type or class then you can put anything you
>>> like into the signal chain of lower or nil distortion
>>> and most likely not be able to identify any change.
>>> Since you have no way to determine the significance or
>>> type/class of distortions in your system, you don't
>>> know anything about the results of an ABX test other
>>> than the results being SPECIFIC to the test you
>>> performed.
>>
>> This problem exists for any listening test. The only
>> reason why it is more apparent in DBTs is that DBTs
>> clearly detect listeners who can't reliably detect
>> differences for whatever reason. We've got yet another case where someone
>> is specifically
>> complaining about ABX tests when in fact the problem
>> exists for *any* listening test. The reason why people
>> don't notice this problem with sighted evalautions is
>> the fact that sighted evaluations have no means for
>> addressing the problem of false positives.

> We can agree then that there are no tests that actually
> take into account this issue, therefore all tests thus
> far are suspect in terms of the significance of the test
> results?

Not at all. In fact there is an international standard, ITU BS 1116, that
specifically addresses this situation.

http :// www ,it u.int/rec/R-REC-BS.1116/en

Re: setting up ABX tests at home - ITU controls?

Arny Krueger wrote:
> "bear" <bearlabs@netzero,net > wrote in message
> news:fp9jvb01hp7@news5.newsguy,com
>> Arny Krueger wrote:
>>> "bear" <bearlabs@netzero,net > wrote in message
>>> news:fp582c01a2c@news1.newsguy,com
>>>
>>>> Put it another way, IF the distortions present are of a
<snip>ressing the problem of false positives.
>
>> We can agree then that there are no tests that actually
>> take into account this issue, therefore all tests thus
>> far are suspect in terms of the significance of the test
>> results?
>
> Not at all. In fact there is an international standard, ITU BS 1116, that
> specifically addresses this situation.
>
> http :// www ,it u.int/rec/R-REC-BS.1116/en
>
Available only by subscription or purchase.

Why don't you paraphrase the key aspects for discussion?
Or excerpt key points (legal, you know...)

Thanks,

- -bear

Re: setting up ABX tests at home - ITU controls?

"bear" <bearlabs@netzero,net > wrote in message
news:fpnvc3066b@news3.newsguy,com

> Arny Krueger wrote:

>> "bear" <bearlabs@netzero,net > wrote in message
>> news:fp9jvb01hp7@news5.newsguy,com

>>> Arny Krueger wrote:

>>>> "bear" <bearlabs@netzero,net > wrote in message
>>>> news:fp582c01a2c@news1.newsguy,com
>
>>>>> Put it another way, IF the distortions present are of
>>>>> a suppressing the problem of false positives.

>>> We can agree then that there are no tests that actually
>>> take into account this issue, therefore all tests thus
>>> far are suspect in terms of the significance of the test
>>> results?

>> Not at all. In fact there is an international standard,
>> ITU BS 1116, that specifically addresses this situation.
>>
>> http :// www ,it u.int/rec/R-REC-BS.1116/en
>>
> Available only by subscription or purchase.

When I bought my copy, it cost me $11 and delivery was instantaneous. I
think the current price would be affected by the weak dollar, but not to an
unmanagable degree.

> Why don't you paraphrase the key aspects for discussion?

Here's a start:

Contents:

1. General
2. Experimental design
3. Selection of listening panels
4. Test method
5. Attributes
6. Programme material
7. Reproduction devices
8. Listening conditions
9. Statistical analysis
10. Presentation of the results of the statistical analyses
11. Contents of test reports.

Re: setting up ABX tests at home - ITU controls?

bear <bearlabs@netzero,net > wrote:
> Arny Krueger wrote:
> > "bear" <bearlabs@netzero,net > wrote in message
> > news:fp9jvb01hp7@news5.newsguy,com
> >> Arny Krueger wrote:
> >>> "bear" <bearlabs@netzero,net > wrote in message
> >>> news:fp582c01a2c@news1.newsguy,com
> >>>
> >>>> Put it another way, IF the distortions present are of a
> <snip>ressing the problem of false positives.
> >
> >> We can agree then that there are no tests that actually
> >> take into account this issue, therefore all tests thus
> >> far are suspect in terms of the significance of the test
> >> results?
> >
> > Not at all. In fact there is an international standard, ITU BS 1116, that
> > specifically addresses this situation.
> >
> > http :// www ,it u.int/rec/R-REC-BS.1116/en
> >
> Available only by subscription or purchase.

specifically, it costs a whopping $20.26 to download.

___
-S
"Hey pip squeak, who's L Ron, some new rapper?" -- Nic

Re: setting up ABX tests at home - ITU controls?

Steven Sullivan wrote:
> bear <bearlabs@netzero,net > wrote:
>> Arny Krueger wrote:
>>> "bear" <bearlabs@netzero,net > wrote in message
>>> news:fp9jvb01hp7@news5.newsguy,com
>>>> Arny Krueger wrote:
>>>>> "bear" <bearlabs@netzero,net > wrote in message
>>>>> news:fp582c01a2c@news1.newsguy,com
>>>>>
>>>>>> Put it another way, IF the distortions present are of a
>> <snip>ressing the problem of false positives.
>>>> We can agree then that there are no tests that actually
>>>> take into account this issue, therefore all tests thus
>>>> far are suspect in terms of the significance of the test
>>>> results?
>>> Not at all. In fact there is an international standard, ITU BS 1116, that
>>> specifically addresses this situation.
>>>
>>> http :// www ,it u.int/rec/R-REC-BS.1116/en
>>>
>> Available only by subscription or purchase.
>
> specifically, it costs a whopping $20.26 to download.
>
>
> -S
> "Hey pip squeak, who's L Ron, some new rapper?" -- Nic

Since no one wants to state what it actually says in the sections where it
applies to this discussion, I'll presume it says nothing that is dispositive.
Doubt that I will spend hard earned money to download a document that is of
questionable value, and certainly not to merely debate here on rahe.

Feel free to excerpt key sections - it's legal.

- -bear

Re: setting up ABX tests at home - ITU controls?

bear <bearlabs@netzero,net > wrote:
> Steven Sullivan wrote:
> > bear <bearlabs@netzero,net > wrote:
> >> Arny Krueger wrote:
> >>> "bear" <bearlabs@netzero,net > wrote in message
> >>> news:fp9jvb01hp7@news5.newsguy,com
> >>>> Arny Krueger wrote:
> >>>>> "bear" <bearlabs@netzero,net > wrote in message
> >>>>> news:fp582c01a2c@news1.newsguy,com
> >>>>>
> >>>>>> Put it another way, IF the distortions present are of a
> >> <snip>ressing the problem of false positives.
> >>>> We can agree then that there are no tests that actually
> >>>> take into account this issue, therefore all tests thus
> >>>> far are suspect in terms of the significance of the test
> >>>> results?
> >>> Not at all. In fact there is an international standard, ITU BS 1116, that
> >>> specifically addresses this situation.
> >>>
> >>> http :// www ,it u.int/rec/R-REC-BS.1116/en
> >>>
> >> Available only by subscription or purchase.
> >
> > specifically, it costs a whopping $20.26 to download.
> >
> > ___
> > -S
> > "Hey pip squeak, who's L Ron, some new rapper?" -- Nic

> Since no one wants to state what it actually says in the sections where it
> applies to this discussion, I'll presume it says nothing that is dispositive.

I don't know what it says, as I don't own a copy. Then again, I'm not
the one insisting that what it says, needs to be posted here.

> Doubt that I will spend hard earned money to download a document that is of
> questionable value, and certainly not to merely debate here on rahe.

So, you've determined the value of an ITU recommendation, without knowing
what it is. That talent must come in handy.

> Feel free to excerpt key sections - it's legal.

Feel free to spend a whopping $20 to satisfy your persistent curiousity.
That's legal too. Btw, what do you spend each month for Usenet access?

___
-S
"Hey pip squeak, who's L Ron, some new rapper?" -- Nic

Re: setting up ABX tests at home

bear <bearlabs@netzero,net > wrote:
> Sonnova wrote:
> > On Mon, 11 Feb 2008 16:17:08 -0800, bear wrote
> > (in article <foqoi402tkd@news1.newsguy,com >):
> >
> >> Arny Krueger wrote:
> >>> "Greg Wormald" <greg.wormald@gmail,com > wrote in message
> >>> news:folrst02p34@news1.newsguy,com
> >>>
> >>>> The very set-up of much ABX testing (which is most often
> >>>> short segments of music quickly switched), cannot test
> >>>> for what most music lovers are seeking--long term
> >>>> enjoyment of music listening.
> >>> This is a false claim that I've heard on and off for about 30 years, now.
> >> Well Arny, where is the correlation between ABX tests and subjective
> >>

> <snip>

> > onnects are so bad that they cause non-trivial frequency response
> >>> variations, they are by definition, defective.
> >> This would appear to be a true statement.
> >> But it is not.
> >> The reason that it is not true is that the "ABX" tests that have been touted
> >> as
> >> providing this information are fraught with unknown variables.
> >> Let me list a few for the readers:
> >> - frequency response of the speakers
> >
> > if you are using the SAME speakers for each component under test, the
> > frequency response of the speakers is a constant, NOT a variable and
> > therefore is irrelevant. If you are using different speakers for the ABX
> > test, then you're doing it wrong by introducing more than one variable into
> > the test.

> You are simply not comprehending the significance of masking due to
> distortion(s) nor the effect of cascaded distortions.

> Put it another way, IF the distortions present are of a particular type or class
> then you can put anything you like into the signal chain of lower or nil
> distortion and most likely not be able to identify any change. Since you have no
> way to determine the significance or type/class of distortions in your system,
> you don't know anything about the results of an ABX test other than the results
> being SPECIFIC to the test you performed.

You're getting way too analytical. It's admirable from a scientific POV
to want to exhausively determine, a priori, whether a difference between
two treatments is audible. This is how mp3 codecs have evolved, for
example.

In the common run of 'audiophile' culture, however, all that's required is
to test its many, many claims of ALREADY HEARD difference. Which is to
say, to repeat the anecdotal 'sighted' experiments, where differences are
sometimes reported to be extravagantly large, this time using DBT
conditions. THis tells us whether the ORIGINAL claim was likely to be a
report of a real difference, or not.

Re: setting up ABX tests at home

On Fri, 15 Feb 2008 15:43:08 -0800, bear wrote
(in article <fp582c01a2c@news1.newsguy,com >):

> Sonnova wrote:
>> On Mon, 11 Feb 2008 16:17:08 -0800, bear wrote
>> (in article <foqoi402tkd@news1.newsguy,com >):
>>
>>> Arny Krueger wrote:
>>>> "Greg Wormald" <greg.wormald@gmail,com > wrote in message
>>>> news:folrst02p34@news1.newsguy,com
>>>>
>>>>> The very set-up of much ABX testing (which is most often
>>>>> short segments of music quickly switched), cannot test
>>>>> for what most music lovers are seeking--long term
>>>>> enjoyment of music listening.
>>>> This is a false claim that I've heard on and off for about 30 years, now.
>>> Well Arny, where is the correlation between ABX tests and subjective
>>>
>
> <snip>
>
>> onnects are so bad that they cause non-trivial frequency response
>>>> variations, they are by definition, defective.
>>> This would appear to be a true statement.
>>> But it is not.
>>> The reason that it is not true is that the "ABX" tests that have been
>>> touted
>>> as
>>> providing this information are fraught with unknown variables.
>>> Let me list a few for the readers:
>>> - frequency response of the speakers
>>
>> if you are using the SAME speakers for each component under test, the
>> frequency response of the speakers is a constant, NOT a variable and
>> therefore is irrelevant. If you are using different speakers for the ABX
>> test, then you're doing it wrong by introducing more than one variable into
>> the test.
>
> You are simply not comprehending the significance of masking due to
> distortion(s) nor the effect of cascaded distortions.
>
> Put it another way, IF the distortions present are of a particular type or
> class
> then you can put anything you like into the signal chain of lower or nil
> distortion and most likely not be able to identify any change. Since you have

> no
> way to determine the significance or type/class of distortions in your
> system,
> you don't know anything about the results of an ABX test other than the
> results
> being SPECIFIC to the test you performed.
>
>>
>>> - distortion vs. power vs. frequency for the speakers
>>
>> See above.
>>
>>> - polar response of the speakers vs. frequency
>>
>> Again irrelevant because the cables/interconnects are feeding the same set
>> of
>> speakers.
>>
>>> - reverberation time vs. frequency of the room
>>
>> Same room for both components under test, the room becomes irrelevant.
>
> That is simply wrong and incorrect IF you wish to generalize the results
> beyond
> the specific test.
>
> Let's do a simple mental excercise on this, ok?
> Put the speakers into a relatively small room made of ceramic tile, do the
> test
> there. Are those results meaningful? Yes? No?
>
> I assume the answer will be *no*.
> The point is that the room does matter - the question is to WHAT DEGREE. It
> is
> not irrelevant it is a non-zero factor. If it is not controlled and known,
> then
> it is a variable whose effect is UNKNOWN. Therefore the validity of the tests

> is
> compromised and can be called into question.
>
> Obviously, let's not push the thing into absurdity - at some point things can

> not be known - but the present crop of ABX tests that I know of don't take
> into
> account relatively simple, basic things that CAN be known.
>
>>
>>> etc. etc. etc....
>>>
>>> EVEN WITH this information, I have yet to read a published ABX test that
>>> would
>>> permit someone to precisely duplicate the test. Not enough information is
>>> provided.
>>
>> Not needed. All you need is a system and a way to switch between two
>> different components carrying the same signal at matched levels. This can
>> be
>> done by a ABX switch box or a third party who either switches or not on
>> command from the listener(s).
>
> If you wish to make this assertion, so be it.
> It is not valid science, nor does it yield particularly useful results except

> by
> serendipitous result.
>
>>> Show me *one* that has the requisite information and controls.
>>
>> It's really very simple. after matching levels and using the same source
>> material, merely switch between the two units under test and determine if
>> you
>> can hear a difference. Listen as long as you like to either or both
>> contenders and decide whether or not you can tell ANY difference between
>> them. It's that simple.
>
> NOT that simple. Sorry. Wrong.
> For all the reasons stated.
> That test is ONLY valid in the system under test. Period.
>
>>> Dr. Earl Geddes' research - and he is a hardcore
>>> academician/scientist/researcher - has shown that absolute values of
>>> distortion do not translate at all to audibility of distortion.
>>
>> But that's a different discussion and has nothing whatsoever to do with
>> blind
>> or ABX testing.
>
> WRONG. It has *everything* to do with ABX and *all other listening tests*.
> Read it yourself.
> Report back.
> No reason to be ignorant of this new research!!
>
>>
>>> So, the sum total of all this is that ABX is a tool that may or may not
>>> yield
>>
>>> information that is useful. Period.
>>
>> Well, ABX/Blind testing can really only tell two things: Is there a
>> difference in the sound between two components (or interconnects or speaker
>> cables)? And if there are differences, what are they? It cannot tell which
>> is better (assuming that there are differences) as that is a totally
>> subjective call.
>
> That is not being discussed.
>
>>
>> Where this type of testing REALLY falls down is where the differences are
>> subtle (which they often are) and the listener doesn't have the trained
>> listening acumen to pick-up on those differences. Since there is no way to
>> eliminate or even pre-screen for this particular variable, the results are
>> often going to show that there is no difference, when in reality, the
>> listener(s) just missed them or really cannot hear them. The results never
>> work the other way where the unsophisticated listener reliably and
>> statistically hears differences that aren't there. I'm sure that you can
>> see
>> why that's so.
>
> Unfortunately, that is NOT what research has shown at all.
>
> You keep insisting on "differences that are not there."
> That is an assertion that may or may not correlate with tests that may or may

> not be effective at testing what they purport to be testing!!
>
> What you want to say is that *you believe* that the majority of differences
> reported by non-ABX test listeners are not real differences. That is your
> belief, which you are entitled to. There are no tests that adequately support

> such a point of view. Again, I urge you to actually read at least the
> abstract
> of Dr. Geddes work to understand WHY I keep saying this sort of thing!! :_)
>
> _-_-bear

I realize that this is usenet and people write stream of consciousness and
for brevity, and often people don't get the points across that they are
trying to convey due to poor wording, omitting certain facts by accident or,
occasionally out of purposeful obfuscation. Here, to paraphrase an old Amos
'n Andy radio routine "either I'm not hearing what you are saying, or you
aren't saying what I'm hearing." either way we don't seem to be
communicating. In any substitution test, you are listening for the
differences between two different components of the same type, whether that
be two preamps, two power amps. two speaker systems or two interconnects. If
the listeners cannot detect any difference between the sound of the two units
under test in a statistically meaningful manner after X number of tries, then
there are NO differences. The rest of the system OTHER than the units under
test, MUST be identical for both units. You can change out only one variable
at a time. If the rest of the ancillary equipment is the same for both, then
any differences perceived must be differences between the two components in
question, because they are the only two variables. All else being the same,
these red herrings (if I understand what you are saying correctly) about room
acoustics, speaker distortion, etc. are irrelevant because (unless you the
speakers are the devices under test) they are the SAME for both test samples.
If I switch between two preamps, for instance, and hear a difference, then it
is clearly the preamps which are different, because nothing else has changed
between the listening of one preamp or the other and the preamp cannot change
the characteristics of either the room or the speakers.

Re: setting up ABX tests at home

Sonnova wrote:
> On Fri, 15 Feb 2008 15:43:08 -0800, bear wrote
> (in article <fp582c01a2c@news1.newsguy,com >):
>
>> Sonnova wrote:
>>> On Mon, 11 Feb 2008 16:17:08 -0800, bear wrote
>>> (in article <foqoi402tkd@news1.newsguy,com >):
>>>
>>>> Arny Krueger wrote:
>>>>> "Greg Wormald" <greg.wormald@gmail,com > wrote in message
>>>>> news:folrst02p34@news1.newsguy,com
>>>>>
>>>>>> The very set-up of much ABX testing (which is most often
>>>>>> short segments of music quickly switched), cannot test
>>>>>> for what most music lovers are seeking--long term
>>>>>> enjoyment of music listening.
>>>>> Th

<massive snip>

>>
>> - -bear
>
> I realize that this is usenet and people write stream of consciousness and
> for brevity, and often people don't get the points across that they are
> trying to convey due to poor wording, omitting certain facts by accident or,
> occasionally out of purposeful obfuscation. Here, to paraphrase an old Amos
> 'n Andy radio routine "either I'm not hearing what you are saying, or you
> aren't saying what I'm hearing." either way we don't seem to be
> communicating. In any substitution test, you are listening for the
> differences between two different components of the same type, whether that
> be two preamps, two power amps. two speaker systems or two interconnects. If
> the listeners cannot detect any difference between the sound of the two units
> under test in a statistically meaningful manner after X number of tries, then
> there are NO differences.

Only if you say there are "no differences detected in this test."
It does NOT mean that there are no differences, or that said differences can
never be heard. This is where we diverge, in specific.

We also may diverge on the issue of the effect of doing the test itself on the
perception of listeners. We may not, depends...

I say you can not generalize as fact from any ABX test. Only can one say that
there is fact as to the result(s) of a given test. THEN we must examine closely
the test conditions - which seems to never take place, although the ABX'ers seem
to want to take the position of being "scientific." Here we diverge also.

The rest of the system OTHER than the units under
> test, MUST be identical for both units. You can change out only one variable
> at a time. If the rest of the ancillary equipment is the same for both, then
> any differences perceived must be differences between the two components in
> question, because they are the only two variables. All else being the same,
> these red herrings (if I understand what you are saying correctly) about room
> acoustics, speaker distortion, etc. are irrelevant because (unless you the
> speakers are the devices under test) they are the SAME for both test samples.
> If I switch between two preamps, for instance, and hear a difference, then it
> is clearly the preamps which are different, because nothing else has changed
> between the listening of one preamp or the other and the preamp cannot change
> the characteristics of either the room or the speakers.

Such a test is VALID ONLY FOR THAT TEST. And, ONLY if you ACCEPT the suitability
of the test method itself (the effect upon listeners... as mentioned).

In such a case I do not have any problem. The claims have to be denoted
appropriately.

Must I repeat this over and over in order for the simple issues to be understood?

I understand completely the ABX method(s)... we've been discussing it for at
least a decade here... time to get past this overly simplistic belief in its
infallibility and suitability for generalization of test results.

It is a useful tool that may or may not yield useful information - depending on
how it is used! Nothing more.

- -bear

Re: setting up ABX tests at home

On Sun, 17 Feb 2008 07:32:01 -0800, bear wrote
(in article <fp9k1h01hrv@news5.newsguy,com >):

> Sonnova wrote:
>> On Fri, 15 Feb 2008 15:43:08 -0800, bear wrote
>> (in article <fp582c01a2c@news1.newsguy,com >):
>>
>>> Sonnova wrote:
>>>> On Mon, 11 Feb 2008 16:17:08 -0800, bear wrote
>>>> (in article <foqoi402tkd@news1.newsguy,com >):
>>>>
>>>>> Arny Krueger wrote:
>>>>>> "Greg Wormald" <greg.wormald@gmail,com > wrote in message
>>>>>> news:folrst02p34@news1.newsguy,com
>>>>>>
>>>>>>> The very set-up of much ABX testing (which is most often
>>>>>>> short segments of music quickly switched), cannot test
>>>>>>> for what most music lovers are seeking--long term
>>>>>>> enjoyment of music listening.
>>>>>> Th
>
> <massive snip>
>
>>>
>>> _-_-bear
>>
>> I realize that this is usenet and people write stream of consciousness and
>> for brevity, and often people don't get the points across that they are
>> trying to convey due to poor wording, omitting certain facts by accident
>> or,
>> occasionally out of purposeful obfuscation. Here, to paraphrase an old Amos
>> 'n Andy radio routine "either I'm not hearing what you are saying, or you
>> aren't saying what I'm hearing." either way we don't seem to be
>> communicating. In any substitution test, you are listening for the
>> differences between two different components of the same type, whether that
>> be two preamps, two power amps. two speaker systems or two interconnects.
>> If
>> the listeners cannot detect any difference between the sound of the two
>> units
>> under test in a statistically meaningful manner after X number of tries,
>> then
>> there are NO differences.
>
> Only if you say there are "no differences detected in this test."
> It does NOT mean that there are no differences, or that said differences can
> never be heard. This is where we diverge, in specific.

Some people can't "hear". They will likely not detect any differences between
components. Many others won't hear. They are so tied to their preconceived
notions that even if such tests proved them wrong, they wouldn't admit it,
even to themselves (like those who insist that all amplifiers, preamps, and
CD players sound the same, or that there is no difference between SACD and
Redbook CDs).

>
> We also may diverge on the issue of the effect of doing the test itself on
> the
> perception of listeners. We may not, depends...
>
> I say you can not generalize as fact from any ABX test. Only can one say that

> there is fact as to the result(s) of a given test. THEN we must examine
> closely
> the test conditions - which seems to never take place, although the ABX'ers
> seem
> to want to take the position of being "scientific." Here we diverge also.

This is where I do not understand you. As long as nothing else changes except
the device under test the rest of the system simply isn't that important to
the test. Sure the rest of the system must be good enough in the first place
that a difference, if it exists, can be detected. I wouldn't, for instance,
try to evaluate the effect of interconnects using a $50 amplified pair of
computer speakers. But assuming a decent set of speakers reasonably good
amplifiers and signal sources, if you switch cables and the sound of the
system doesn't change to the point where no one on the listening panel can
statistically, reliably tell which set of cables is which, then I would have
to deduce that there is a high probability that no difference between the
sound of the two interconnects exist.

>
> The rest of the system OTHER than the units under
>> test, MUST be identical for both units. You can change out only one
>> variable
>> at a time. If the rest of the ancillary equipment is the same for both,
>> then
>> any differences perceived must be differences between the two components in
>> question, because they are the only two variables. All else being the same,
>> these red herrings (if I understand what you are saying correctly) about
>> room
>> acoustics, speaker distortion, etc. are irrelevant because (unless you the
>> speakers are the devices under test) they are the SAME for both test
>> samples.
>> If I switch between two preamps, for instance, and hear a difference, then
>> it
>> is clearly the preamps which are different, because nothing else has
>> changed
>> between the listening of one preamp or the other and the preamp cannot
>> change
>> the characteristics of either the room or the speakers.
>
> Such a test is VALID ONLY FOR THAT TEST. And, ONLY if you ACCEPT the
> suitability
> of the test method itself (the effect upon listeners... as mentioned).
>
> In such a case I do not have any problem. The claims have to be denoted
> appropriately.
>
> Must I repeat this over and over in order for the simple issues to be
> understood?

To me you aren't doing a very good job of explaining it at all, but I think I
finally understand what you are saying. Basically, I think you are saying
that just because no difference between two components was detected using
set-up "A" (due to such factors as room acoustics, speaker distortion,
frequency response, etc) doesn't necessarily mean that differences don't
exist and won't be detected in set-up "B" with different room acoustics,
different speakers etc. All this time I thought you were saying that speaker
distortion, frequency response and room acoustics would affect different
amps, preamps, interconnects, or signal sources differently in the SAME
system during the same test. Is that what you are getting at? If so I am in
complete agreement and we would have saved a lot of time and bandwidth if you
had just said that at the beginning. Put it down to miscommunication.

Re: setting up ABX tests at home

Sonnova wrote:
> On Sun, 17 Feb 2008 07:32:01 -0800, bear wrote
> (in article <fp9k1h01hrv@news5.newsguy,com >):
>
>> Sonnova wrote:
>>> On Fri, 15 Feb 2008 15:43:08 -0800, bear wrote
>>> (in article <fp582c01a2c@news1.newsguy,com >):
>>>
>>>> Sonnova wrote:
>>>>> On Mon, 11 Feb 2008 16:17:08 -0800, bear wrote
>>>>> (in article <foqoi402tkd@news1.newsguy,com >):
>>>>>
>>>>>> Arny Krueger wrote:
>>>>>>> "Greg Wormald" <greg.wormald@gmail,com > wrote in message
>>>>>>> news:folrst02p34@news1.newsguy,com
>>>>>>>
>>>>>>>> The very set-up of much ABX testing (which is most often
>>>>>>>> short segments of music quickly switched), cannot test
>>>>>>>> for what most music lovers are seeking--long term
>>>>>>>> enjoyment of music listening.
>>>>>>> Th
>> <massive snip>
<snipped again>
> Some people can't "hear". They will likely not detect any differences between
> components. Many others won't hear. They are so tied to their preconceived
> notions that even if such tests proved them wrong, they wouldn't admit it,
> even to themselves (like those who insist that all amplifiers, preamps, and
> CD players sound the same, or that there is no difference between SACD and
> Redbook CDs).

I'm in general agreement with this...
>
<snippo>
>
> This is where I do not understand you. As long as nothing else changes except
> the device under test the rest of the system simply isn't that important to
> the test. Sure the rest of the system must be good enough in the first place
> that a difference, if it exists, can be detected. I wouldn't, for instance,
> try to evaluate the effect of interconnects using a $50 amplified pair of
> computer speakers. But assuming a decent set of speakers reasonably good
> amplifiers and signal sources, if you switch cables and the sound of the
> system doesn't change to the point where no one on the listening panel can
> statistically, reliably tell which set of cables is which, then I would have
> to deduce that there is a high probability that no difference between the
> sound of the two interconnects exist.

A "high probability" perhaps. But only perhaps.
A deduction of this sort simply does not rise to the level of certainty required
in order to generalize results. If you do not see why this is so, then further
discussion is pointless?

>
>> The rest of the system OTHER than the units under
<snip>
>
> To me you aren't doing a very good job of explaining it at all, but I think I
> finally understand what you are saying. Basically, I think you are saying
> that just because no difference between two components was detected using
> set-up "A" (due to such factors as room acoustics, speaker distortion,
> frequency response, etc) doesn't necessarily mean that differences don't
> exist and won't be detected in set-up "B" with different room acoustics,
> different speakers etc.

This is true. I am saying this.

I am going beyond this to suggest that in some set-ups (as you call it) one will
be absolutely unable to detect differences (reliably) that are actually present.

> All this time I thought you were saying that speaker
> distortion, frequency response and room acoustics would affect different
> amps, preamps, interconnects, or signal sources differently in the SAME
> system during the same test. Is that what you are getting at? If so I am in
> complete agreement and we would have saved a lot of time and bandwidth if you
> had just said that at the beginning. Put it down to miscommunication.

Well, your sentence construction here is slightly unclear.
But, if by "differently" you mean that the system used in the test might "mask"
- meaning obscure the true sonic character of a given DUT - then yes, it may. My
point is that if this is not in anyway determined and controlled by some means,
then the results are only valid for the specific test conducted.

- -bear

PS. to the extent that a given speaker load (for example) causes a significant
enough difference in freq response and distortion in two amps under test, then
there would be a "difference" between them (if it was audible) - but this is not
really in question, imho. The question comes in terms of what rises to an
audible difference that can be reliably detected - compared to what so very many
people appear to hear when asked for subjective or casual observation.

Re: setting up ABX tests at home

On Feb 11, 7:17 pm, bear <bearl...@netzero,net > wrote:

> Dr. Earl Geddes' research - and he is a hardcore
> academician/scientist/researcher - has shown that absolute values of distortion
> do not translate at all to audibility of distortion.

Please provide a citation for this research. Thanks.

bob

Re: setting up ABX tests at home

bob wrote:
> On Feb 11, 7:17 pm, bear <bearl...@netzero,net > wrote:
>
>> Dr. Earl Geddes' research - and he is a hardcore
>> academician/scientist/researcher - has shown that absolute values of distortion
>> do not translate at all to audibility of distortion.
>
> Please provide a citation for this research. Thanks.
>
> bob

Try Google? He has his own website.
Dr. Earl Geddes.
JAES articles, if you wish to buy them too.

- -bear