Loudspeaker Evaluation Protocol

[Roy W. Rising]

Many years ago I developed a Loudspeaker Evaluation Protocol for ABC-TV,
Hollywood. It easily and finally differentiates the good from the others.

The competition and hype (referred to as "Bose" by some, "Monster Cable" by
others) stands in the way of identifying the truly accurate systems. When
the reproducers are highly accurate and can be trusted, things like voicing
and room treatment fall to third place in importance.

It works like this: Imagine yourself blindfolded in a bedroom. Someone
you know speaks. Would you recognize the voice? Now imagine yourself in
the kitchen, still blindfolded. Someone else you know speaks. Would you
recognize the voice? You can see that the acoustics don't really matter.
The Ear-Brain processes away the aberations faster than light speed. As
long as the source is truly accurate, we hear what we need for judging the
sound, mix, etc.

I'm thinking of taking the Protocol to Consumer Reports, they have the
facility and do regular testing of loudspeakers. However, I'm in Los
Angeles and would like to find a local place, a sponsor and a small group
of "golden ears" to participate in the evaluations. I might approach
Harmon Industries (JBL, etc.) with the idea.

Any thoughts would be welcome. Thanks.

--
~ Roy
"If you notice the sound, it's wrong!"

[Kalman Rubinson]

On 24 Aug 2006 15:15:41 GMT, Roy W. Rising >
wrote:

>It works like this: Imagine yourself blindfolded in a bedroom. Someone
>you know speaks. Would you recognize the voice? Now imagine yourself in
>the kitchen, still blindfolded. Someone else you know speaks. Would you
>recognize the voice? You can see that the acoustics don't really matter.
>The Ear-Brain processes away the aberations faster than light speed. As
>long as the source is truly accurate, we hear what we need for judging the
>sound, mix, etc.

Here's the fly in that ointment. With a real source, you recognize
the voice for what it is but in the context of the current space and
its acoustics. That does not mean that it sounds the same, merely
that you associate it with what you already have in memory. It still
sounds like that voice in the bedroom or in the kitchen or.......

With an unknown source for which you do not have a memory model, you
cannot know what it should sound like, so the acoustics of the
listening room cannot be distinguished from the reproduced sounds.

Adaptation (which is not as fast as you say) has little to do with
this phenomenon.

Nonetheless, it would be nice to find a reliable and quick way to
evaluate loudspeakers.

Kal

[soundhaspriority]

"Kalman Rubinson" > wrote in message
...
> On 24 Aug 2006 15:15:41 GMT, Roy W. Rising >
> wrote:
>
>>It works like this: Imagine yourself blindfolded in a bedroom. Someone
>>you know speaks. Would you recognize the voice? Now imagine yourself in
>>the kitchen, still blindfolded. Someone else you know speaks. Would you
>>recognize the voice? You can see that the acoustics don't really matter.
>>The Ear-Brain processes away the aberations faster than light speed. As
>>long as the source is truly accurate, we hear what we need for judging the
>>sound, mix, etc.
>
> Here's the fly in that ointment. With a real source, you recognize
> the voice for what it is but in the context of the current space and
> its acoustics. That does not mean that it sounds the same, merely
> that you associate it with what you already have in memory. It still
> sounds like that voice in the bedroom or in the kitchen or.......
>
> With an unknown source for which you do not have a memory model, you
> cannot know what it should sound like, so the acoustics of the
> listening room cannot be distinguished from the reproduced sounds.
>
> Adaptation (which is not as fast as you say) has little to do with
> this phenomenon.
>
> Nonetheless, it would be nice to find a reliable and quick way to
> evaluate loudspeakers.
>
> Kal
>
His post lacks the details of a true protocol. However, it seems to be
reminiscent of the early AR demos, where a pair of speakers were set up in
an auditorium, behind some kind of a baffle, and the audience was challenged
to distinguish from a live orchestra.
Comments?

[Roy W. Rising]

Kalman Rubinson > wrote:

> With an unknown source for which you do not have a memory model, you
> cannot know what it should sound like, so the acoustics of the
> listening room cannot be distinguished from the reproduced sounds.
>
> Adaptation (which is not as fast as you say) has little to do with
> this phenomenon.
>
> Nonetheless, it would be nice to find a reliable and quick way to
> evaluate loudspeakers.
>
> Kal

I have to disagree. However, the Loudspeaker Evaluation Protocol does
what's needed ... objective differentiation of the good from the others.
~ Roy

--
~ Roy
"If you notice the sound, it's wrong!"

[Roy W. Rising]

"soundhaspriority" > wrote:
> >
> His post lacks the details of a true protocol. However, it seems to be
> reminiscent of the early AR demos, where a pair of speakers were set up
> in an auditorium, behind some kind of a baffle, and the audience was
> challenged to distinguish from a live orchestra.
> Comments?

Unlike the old AR blindfold demo, the Protocol uses A-X/Y comparison where
A is the source material and X & Y are the reproduced sounds from two
unknown loudspeakers. The key word is "comparison". ~ Roy

--
~ Roy
"If you notice the sound, it's wrong!"

[Paul Stamler]

"Roy W. Rising" > wrote in message
...

> It works like this: Imagine yourself blindfolded in a bedroom. Someone
> you know speaks. Would you recognize the voice? Now imagine yourself in
> the kitchen, still blindfolded. Someone else you know speaks. Would you
> recognize the voice? You can see that the acoustics don't really matter.
> The Ear-Brain processes away the aberations faster than light speed. As
> long as the source is truly accurate, we hear what we need for judging the
> sound, mix, etc.

Telephone your wife, partner, sig o or best friend. Do you recognize the
voice? Is your telephone a high-quality sound reproducer?

Peace,
Paul

[Scott Dorsey]

Paul Stamler > wrote:
>"Roy W. Rising" > wrote in message
>
>> It works like this: Imagine yourself blindfolded in a bedroom. Someone
>> you know speaks. Would you recognize the voice? Now imagine yourself in
>> the kitchen, still blindfolded. Someone else you know speaks. Would you
>> recognize the voice? You can see that the acoustics don't really matter.
>> The Ear-Brain processes away the aberations faster than light speed. As
>> long as the source is truly accurate, we hear what we need for judging the
>> sound, mix, etc.
>
>Telephone your wife, partner, sig o or best friend. Do you recognize the
>voice? Is your telephone a high-quality sound reproducer?

On the cellphone, I often don't.
--scott

--
"C'est un Nagra. C'est suisse, et tres, tres precis."

[Kalman Rubinson]

On 24 Aug 2006 16:40:14 GMT, Roy W. Rising >
wrote:

>Kalman Rubinson > wrote:
>
>> With an unknown source for which you do not have a memory model, you
>> cannot know what it should sound like, so the acoustics of the
>> listening room cannot be distinguished from the reproduced sounds.
>>
>> Adaptation (which is not as fast as you say) has little to do with
>> this phenomenon.
>>
>> Nonetheless, it would be nice to find a reliable and quick way to
>> evaluate loudspeakers.
>>
>> Kal
>
> I have to disagree. However, the Loudspeaker Evaluation Protocol does
> what's needed ... objective differentiation of the good from the others.

Well, until you divulge exactly what your test does, I can only judge
it on your faulty use of sensory biology.

Kal

[Roy W. Rising]

Kalman Rubinson > wrote:
> On 24 Aug 2006 16:40:14 GMT, Roy W. Rising >
> wrote:
>>
> Well, until you divulge exactly what your test does, I can only judge
> it on your faulty use of sensory biology.
>
> Kal

My 12:50 PM post gives a clue. Lipschitz and Vanerkooy have validated the
essence of the method. ~ Roy

--
~ Roy
"If you notice the sound, it's wrong!"

[Kalman Rubinson]

On 24 Aug 2006 18:21:44 GMT, Roy W. Rising >
wrote:

>Kalman Rubinson > wrote:
>> On 24 Aug 2006 16:40:14 GMT, Roy W. Rising >
>> wrote:
>>>
>> Well, until you divulge exactly what your test does, I can only judge
>> it on your faulty use of sensory biology.
>>
>> Kal
>
>My 12:50 PM post gives a clue. Lipschitz and Vanerkooy have validated the
>essence of the method. ~ Roy

If "A" is the original source, i.e., live performer, OK but that's
hardly practical in the vast majority of situations. If you limit the
"A" source to a live human voice, the results are similarly limited
(although valid) to the voice range. Either way, it seems to be
simply a "live-vs.-recorded" comparison. So, how can this be of
practical use in wide-range loudspeaker evaluation?

(BTW, there are still issues regarding the redundancy of the hall
ambience in the reproduced sound but that can be finessed, to a
degree, with careful miking.)

In neither case, again, is adaptation an issue.

Kal

[Roy W. Rising]

Kalman Rubinson > wrote:
> On 24 Aug 2006 18:21:44 GMT, Roy W. Rising >
> wrote:
>
> If "A" is the original source, i.e., live performer, OK but that's
> hardly practical in the vast majority of situations. If you limit the
> "A" source to a live human voice, the results are similarly limited
> (although valid) to the voice range. Either way, it seems to be
> simply a "live-vs.-recorded" comparison. So, how can this be of
> practical use in wide-range loudspeaker evaluation?
>
> (BTW, there are still issues regarding the redundancy of the hall
> ambience in the reproduced sound but that can be finessed, to a
> degree, with careful miking.)
>
> In neither case, again, is adaptation an issue.
>
> Kal

'A' is recorded program material ... voice, music, etc. 'X' and 'Y' are
loudspeakers in an anechoic environment or, at least, one where reflections
are late enough to be ignored. A test microphone returns the sound of a
speaker to a monitoring place where it can be compared to the source.
Switching is set up to allow comparison of the three - a known and two
unknowns. I have executed this protocol many times and all listeners agree
about which unknown most closely matches the known. The closer the match,
the more accurate the reproducer. The choice of the test mic is important.
It, too, must be accurate. I use a hand picked EV RE55. I have done
similar X/Y comparisons with the mic to develop confidence in its accuracy.
The monitor speaker should be as good as possible, but it is common to A,
X, and Y, so its influence is neutral. I hope this answers some questions.
~ Roy

--
~ Roy
"If you notice the sound, it's wrong!"

[Mark]

> >
> >Telephone your wife, partner, sig o or best friend. Do you recognize the
> >voice? Is your telephone a high-quality sound reproducer?
>
> On the cellphone, I often don't.
> --scott
>
> --
> "C'est un Nagra. C'est suisse, et tres, tres precis."

It can be a problem if you get your sig other confused with your
wife... :-)
Mark

[Kalman Rubinson]

On 24 Aug 2006 19:33:58 GMT, Roy W. Rising >
wrote:

>'A' is recorded program material ... voice, music, etc. 'X' and 'Y' are
>loudspeakers in an anechoic environment or, at least, one where reflections
>are late enough to be ignored. A test microphone returns the sound of a
>speaker to a monitoring place where it can be compared to the source.
>Switching is set up to allow comparison of the three - a known and two
>unknowns. I have executed this protocol many times and all listeners agree
>about which unknown most closely matches the known. The closer the match,
>the more accurate the reproducer. The choice of the test mic is important.
>It, too, must be accurate. I use a hand picked EV RE55. I have done
>similar X/Y comparisons with the mic to develop confidence in its accuracy.
>The monitor speaker should be as good as possible, but it is common to A,
>X, and Y, so its influence is neutral. I hope this answers some questions.

Thanks for this information which, indeed, explains a lot more than
your earlier posts and seems logical. It should serve, within the
limits of the microphone and electronic(two matched sets, of course),
to compare the on-axis, near-field responses of the speakers for
listeners who sit on-axis to "A." As such, it might be very useful
where such facilities are available.

I still do not think it is a comprehensive test for the way speakers
are used these days: in small rooms, in pairs or more, where spatial
radiation characteristics are important. Perhaps combining your
protocol with mechanized blind, paired audition facilities, such as
those at Harmon, would be very informative.

Kal

[Roy W. Rising]

Kalman Rubinson > wrote:
> On 24 Aug 2006 19:33:58 GMT, Roy W. Rising >
> wrote:
>
> Thanks for this information which, indeed, explains a lot more than
> your earlier posts and seems logical. It should serve, within the
> limits of the microphone and electronic(two matched sets, of course),
> to compare the on-axis, near-field responses of the speakers for
> listeners who sit on-axis to "A." As such, it might be very useful
> where such facilities are available.
>
> I still do not think it is a comprehensive test for the way speakers
> are used these days: in small rooms, in pairs or more, where spatial
> radiation characteristics are important. Perhaps combining your
> protocol with mechanized blind, paired audition facilities, such as
> those at Harmon, would be very informative.
>
> Kal

Actually, the tests I've done to date were by no means near-field. The
test mic was placed 3 or so meters from the speakers. The environment was
a not-very-dead sound stage. There was no difficulty differentiating the
better from the lesser reproducer. ~ Roy

--
~ Roy
"If you notice the sound, it's wrong!"

[Kalman Rubinson]

On 24 Aug 2006 20:45:31 GMT, Roy W. Rising >
wrote:

>Kalman Rubinson > wrote:
>> On 24 Aug 2006 19:33:58 GMT, Roy W. Rising >
>> wrote:
>>
>> Thanks for this information which, indeed, explains a lot more than
>> your earlier posts and seems logical. It should serve, within the
>> limits of the microphone and electronic(two matched sets, of course),
>> to compare the on-axis, near-field responses of the speakers for
>> listeners who sit on-axis to "A." As such, it might be very useful
>> where such facilities are available.
>>
>> I still do not think it is a comprehensive test for the way speakers
>> are used these days: in small rooms, in pairs or more, where spatial
>> radiation characteristics are important. Perhaps combining your
>> protocol with mechanized blind, paired audition facilities, such as
>> those at Harmon, would be very informative.
>>
>> Kal
>
>Actually, the tests I've done to date were by no means near-field. The
>test mic was placed 3 or so meters from the speakers. The environment was
>a not-very-dead sound stage. There was no difficulty differentiating the
>better from the lesser reproducer. ~ Roy

Of course, that may depend on how much better or lesser they were!

Kal

[Roy W. Rising]

> On 24 Aug 2006 20:45:31 GMT, Roy W. Rising >
> wrote:
>
> >Actually, the tests I've done to date were by no means near-field. The
> >test mic was placed 3 or so meters from the speakers. The environment
> >was a not-very-dead sound stage. There was no difficulty
> >differentiating the better from the lesser reproducer. ~ Roy
>
Kalman Rubinson > wrote:
> Of course, that may depend on how much better or lesser they were!
>
> Kal

Surprisingly, even very subtle differences are discernable. That's the
nice thing about comparative evaluation. ~ Roy

--
~ Roy
"If you notice the sound, it's wrong!"

[Robert Morein]

"Roy W. Rising" > wrote in message
...
> Kalman Rubinson > wrote:
>> On 24 Aug 2006 18:21:44 GMT, Roy W. Rising >
>> wrote:
>>
>> If "A" is the original source, i.e., live performer, OK but that's
>> hardly practical in the vast majority of situations. If you limit the
>> "A" source to a live human voice, the results are similarly limited
>> (although valid) to the voice range. Either way, it seems to be
>> simply a "live-vs.-recorded" comparison. So, how can this be of
>> practical use in wide-range loudspeaker evaluation?
>>
>> (BTW, there are still issues regarding the redundancy of the hall
>> ambience in the reproduced sound but that can be finessed, to a
>> degree, with careful miking.)
>>
>> In neither case, again, is adaptation an issue.
>>
>> Kal
>
> 'A' is recorded program material ... voice, music, etc. 'X' and 'Y' are
> loudspeakers in an anechoic environment or, at least, one where
> reflections
> are late enough to be ignored. A test microphone returns the sound of a
> speaker to a monitoring place where it can be compared to the source.
> Switching is set up to allow comparison of the three - a known and two
> unknowns. I have executed this protocol many times and all listeners
> agree
> about which unknown most closely matches the known.

An obvious problem is that this measures the subjective performance of
speakers under anechoic conditions. Rooms affect different speakers
differently. Hence, I must reject your claim that your test is useful. It is
not.

[Roy W. Rising]

"Robert Morein" > wrote:
>
> An obvious problem is that this measures the subjective performance of
> speakers under anechoic conditions. Rooms affect different speakers
> differently. Hence, I must reject your claim that your test is useful. It
> is not.

If you have read all of the other exchanges on this topic, you'll see that
the conclusion to which you have jumped is invalid. The measurement is
neither subjective nor is the test environment necessarily anechoic.
Beyond that, empirical results of the protocol have proved the validity of
the mehtod. ~ Roy

--
~ Roy
"If you notice the sound, it's wrong!"

[soundhaspriority]

"Roy W. Rising" > wrote in message
...
> "Robert Morein" > wrote:
>>
>> An obvious problem is that this measures the subjective performance of
>> speakers under anechoic conditions. Rooms affect different speakers
>> differently. Hence, I must reject your claim that your test is useful. It
>> is not.
>
> If you have read all of the other exchanges on this topic, you'll see that
> the conclusion to which you have jumped is invalid. The measurement is
> neither subjective nor is the test environment necessarily anechoic.
> Beyond that, empirical results of the protocol have proved the validity of
> the mehtod. ~ Roy
>
I have read the entire thread. Your response is invalid. Your method is not
proved.

[Steven Sullivan]

Roy W. Rising > wrote:
> Kalman Rubinson > wrote:
> > On 24 Aug 2006 16:40:14 GMT, Roy W. Rising >
> > wrote:
> >>
> > Well, until you divulge exactly what your test does, I can only judge
> > it on your faulty use of sensory biology.
> >
> > Kal

> My 12:50 PM post gives a clue. Lipschitz and Vanerkooy have validated the
> essence of the method. ~ Roy


I expect Floyd Toole and Sean Olive have read Lip****z and Vanderkooy.
You might want to look into T & O's speaker evaluation protocols, as
developed at the CRC and Harman, and see how and why they vary from yours.





___
-S
"As human beings, we understand the world through simile, analogy,
metaphor, narrative and, sometimes, claymation." - B. Mason

[Roy W. Rising]

Steven Sullivan > wrote:
> Roy W. Rising > wrote:
> > Kalman Rubinson > wrote:
> > > On 24 Aug 2006 16:40:14 GMT, Roy W. Rising >
> > > wrote:
> > >>
> > > Well, until you divulge exactly what your test does, I can only judge
> > > it on your faulty use of sensory biology.
> > >
> > > Kal
>
> > My 12:50 PM post gives a clue. Lipschitz and Vanerkooy have validated
> > the essence of the method. ~ Roy
>
> I expect Floyd Toole and Sean Olive have read Lip****z and Vanderkooy.
> You might want to look into T & O's speaker evaluation protocols, as
> developed at the CRC and Harman, and see how and why they vary from
> yours.
>
> ___
> -S
> "As human beings, we understand the world through simile, analogy,
> metaphor, narrative and, sometimes, claymation." - B. Mason

Steven ~ Please point me to the protocols you mentioned, I'd like to read
them. Mine grew out of TV's "Golden Age" practice of micing singers with
boom mics while playing backing tracks to them through Altec A7
loudspeakers. The voices sounded fine on the mics, but the track leakage
was terrible! Things were much better when the speaker's output sounded
almost exactly like the tracks. (The protocol was described in an earlier
follow-up.)

--
~ Roy
"If you notice the sound, it's wrong!"

[Steven Sullivan]

Roy W. Rising > wrote:
> Steven Sullivan > wrote:
> > Roy W. Rising > wrote:
> > > Kalman Rubinson > wrote:
> > > > On 24 Aug 2006 16:40:14 GMT, Roy W. Rising >
> > > > wrote:
> > > >>
> > > > Well, until you divulge exactly what your test does, I can only judge
> > > > it on your faulty use of sensory biology.
> > > >
> > > > Kal
> >
> > > My 12:50 PM post gives a clue. Lipschitz and Vanerkooy have validated
> > > the essence of the method. ~ Roy
> >
> > I expect Floyd Toole and Sean Olive have read Lip****z and Vanderkooy.
> > You might want to look into T & O's speaker evaluation protocols, as
> > developed at the CRC and Harman, and see how and why they vary from
> > yours.
> >
> > ___
> > -S
> > "As human beings, we understand the world through simile, analogy,
> > metaphor, narrative and, sometimes, claymation." - B. Mason

> Steven ~ Please point me to the protocols you mentioned, I'd like to read
> them.


You can look up and purchase the relevant JAES articles here:

http://www.aes.org/journal/search.cfm



___
-S
"As human beings, we understand the world through simile, analogy,
metaphor, narrative and, sometimes, claymation." - B. Mason

[Roy W. Rising]

Steven Sullivan > wrote:
>
> You can look up and purchase the relevant JAES articles here:
>
> http://www.aes.org/journal/search.cfm
>
> ___
> -S
> "As human beings, we understand the world through simile, analogy,
> metaphor, narrative and, sometimes, claymation." - B. Mason

Thanks. Maybe when I win the lotto. ;-p

--
~ Roy
"If you notice the sound, it's wrong!"

[Scott Dorsey]

Roy W. Rising > wrote:
>Steven Sullivan > wrote:
>>
>> You can look up and purchase the relevant JAES articles here:
>>
>> http://www.aes.org/journal/search.cfm
>
>Thanks. Maybe when I win the lotto. ;-p

Try your local library then. It'll take a little longer, but won't cost
anything.
--scott

--
"C'est un Nagra. C'est suisse, et tres, tres precis."