[Jerry]

Dick, I was told by a friend that when a speaker system's efficiency is
measured that the input signal (to the system) is always held at 2.83 volts
regardless of the speaker's nominal impedance. Then at one meter we
measured SPL.

Is this correct?

To be fair, I'd think we'd rather input 1 watt and take the speaker's
impedance into account. Now, I realize a speaker's impedance can vary
considerably over the audio spectrum, so deciding what is one watt is not
necessarily the easiest thing.

So, Dick, how are we supposed to compare speakers based upon their
efficiency specs?

Jerry

[[email protected]]

Jerry wrote:

Dick, I was told by a friend that when a speaker
system's efficiency is measured that the input signal
(to the system) is always held at 2.83 volts
regardless of the speaker's nominal impedance. Then at
one meter we measured SPL.

Is this correct?

To be fair, I'd think we'd rather input 1 watt and
take the speaker's impedance into account. Now, I
realize a speaker's impedance can vary considerably
over the audio spectrum, so deciding what is one watt
is not necessarily the easiest thing.

So, Dick, how are we supposed to compare speakers
based upon their efficiency specs?


You've identified one of the major problems in measuring
"efficiency." There are several more.

Efficiency in a loudspeaker is DEFINED as the ratio of
total real acoustical power output to total real
electrical power input, and is expressed as a percentage
figure, like 1%.. That's a precise, unambiguous
statement which contains all the gotchas and conditions
for precise measurement. Unfortunately, it hints at the
rather large practical difficulties of making such
measurements.

Recall the discussion earlier about the differences
between apparent power and real power? Well, the issues
underlying that discussion, as silly as it got, have
important implications. The "real" power, that is, the
actual power that is dissipated by the device is waht's
of importance here, because only the real power, i.e.,
that portion that is handled by the resistive component
of the impedance, is the ONLY part of the total power
that can actually make sound.

And, to complicate matters even more, those same
principles of real vs apparent (also called "imaginary"
as in sqrt(-1)) also affect the ACOUSTICAL power as
well. This is because the acoustical power is developed
across an acoustical "load" or impedance, which itself
has a real and apparent (imaginary) part, and the actual
dissipation of power (or first derivative of work wrt
time) is developed as actual sound.

There are other confounding factors. Among these are the
fact that the total acoustical power output can only be
sampled at discrete points, whether those points are
microphons or eardrums. They are sampling the total
power output of the system only over an extremely tiny
region of the total radiation space of the speaker.
Obviously, if the power over a spherical surface having
the speaker at its center is not constant, then any
assumptions about extrapolating the total power output
from a few tiny points are not going to be valid.

These are just a few of the confounding factors in
obtaining a true measure of "efficiency" as defined
above. You will seldom, if ever, see a true "efficiency"
measurement of a speaker, most especially consumer-
oriented loudspeakers (of which your AR3a is but one
example).

As it turns out, there are a couple of factors that aid
us in coming up with a usable figure. It's not
efficiency, but rather "sensitivity." Those factors
include:

1. The vast majority of amplifiers driving speakers are,
in the precise engineering sense of the term "voltage
sources." This means that the effective output
impedance of the source (amplifier) is insubstantial
when compared to the load (speaker). The consequence
of this is that most loudspeakers are designed to
achieve their frequency response when driven by a
constant voltage, not at constant power.

2. Over the range where the ear is most sensitive
(essentially, a loose definition of "midrange"), the
actual acoustic output of the speaker in most typical
listening positions is fairly independent of the
acoustical conditions in which the speaker finds
itself. That means as long as you are at or near what
is defined as the "principle listening axis"), the
response of the speaker over the critical listening
range is fairly independent of all but pathological
listening environmental conditions. (the room has the
greatest effect only in the bass and the lower
midrange: it's a big surprise to most people how
closely the in-room response of the loudpeaker
approximates the free-field response).

Given those two conditions are met, and they are met in
almost ever case in actual usage, then we can define an
alternative to a true "efficiency" called
"characteristic sensitivity," which is defined as
(quoted from IEC 168-5, section 14.3 "Characteristic
Sensitivity":

"The sound pressure produced at 1 m from the
reference point on the reference axis when the
loudspeaker is supplied with a pink-noise signal the
voltage of which corresponds TO A POWER OF 1 W in the
rated impedance. The bandwidth of the pink-noise
signal shall be limited to the rated frequency range
of the loudspeaker."

Notice the phrase "the voltage of which corresponds TO A
POWER OF 1 W in the rated impedance." (The emphasis is
as found in the standard.) Here we take care of the
variable impedance issue: it's the "rated impedance" of
the speaker (which is addressed in a different part of
the standard).

Using this definition of sensitivity, we have
encompassed three of the major issues outlined above:

1. The issue of defining power into a variable,
partially reactive impedance by using a voltage
reference and a nominal impedance value,

2. The issue of radiation pattern, potential room
loading issues at low frequencies, et al, by
specifying the measurement is done on the reference
axis at 1 meter from the reference point.

3. Using pink noise essentially eliminates the freqency-
dependent impedance variations as a factor. Integrate
the impedance of the loudspeaker over it's entire
rated range, and you'll find almost universally that
the result is VERY close to purely resistive.

Now, the 2.83 volt figure comes from the fact that if
the speaker has a rated impedance of 8 ohms, the voltage
needed to produce 1 watt into that 8 ohms is 2.83 volts.
For a 4 ohms speaker, it would be 2 volts.

However, the sensitivity figure can be stretched a wee
bit to incorporate what the "apparent" efficiency of
speakers with different rated impedances might be. Say
you have two loudspeakers of similar frequency range,
and of identical electroacoustic efficiency (say, 1%),
but one has a rated impedance of 8 ohms, the other is 4
ohms. With the volume control set to the same position,
guess what? The 4 ohm loudspeaker will sound louder than
the 8 ohm speaker, because it is putting out more
acoustics poweer. Why? because volume control set to the
same position, for the same musical poassages, the
amplifier produces the same voltage,, and that same
voltage produces more power in a lower impedance than in
a higher impedance. In fact, the 4 ohm loudspeaker wil
be producing 3 dB more sound pressure level for the same
input that the 8 ohm speaker will do.

To get back to your question more directly: your AR3a
has a "rated impedance" (aka "nominal impedance") of 4
ohms. That's what the manufacturer states in there
specifications, and that's what IEC 268-5 allows as the
figure. Into such a load, 2 volts RMS of pink noise
produces nominal 1 watt of power, thus for a 4 ohm
loudspeaker, the characteristic sensitivity would be
be measured under that driving voltage on the reference
axis of the speaker.

For a speaker with a rated impedance of 8 ohms, that
voltage would be 2.83 volts to produce the same nominal
1 watt.

If you're interested in calculating power budgets in a
multi-speaker system, efficiency is the more important
figure. But if you're calculating, for example, the
attenuation neede to match a midrange to a woofer, then
sensitvity from a fixed voltage is far more important,
since the two drivers are driven by the same amplifier
(assuming passive networks) or the relative figures are
needed for adjusting the gains of multiple amplifiers
(assuming multi-amp systems).

So, this is very long-winded, but it covers the problems
of measuring efficiency and why there exist "sensivity"
measurements: They're simply easier to perform and more
directly comparable, as long as the domain is well
defined. Sensitivity is much more a useful measurement
in your situation than raw efficiency. Efficiency is a
more significant figure when a concert engineer or an
acoustical architect is trying to put together a massive
sound system and is trying to decide whether they need
10,000 watts or 20,000 watts of amplifier power. That
difference has substantial economic and engineering
consequences that simply are of relatively minor
importance in a home listening situation by comparison.

[Jerry]

Dick wrote on 9/19/2006:

Jerry wrote:

So, Dick, how are we supposed to compare speakers
based upon their efficiency specs?

You've identified one of the major problems in measuring
"efficiency." There are several more.

Efficiency in a loudspeaker is DEFINED as the ratio of
total real acoustical power output to total real
electrical power input, and is expressed as a percentage
figure, like 1%.. That's a precise, unambiguous
statement which contains all the gotchas and conditions
for precise measurement. Unfortunately, it hints at the
rather large practical difficulties of making such
measurements.

Recall the discussion earlier about the differences
between apparent power and real power? Well, the issues
underlying that discussion, as silly as it got, have
important implications. The "real" power, that is, the
actual power that is dissipated by the device is waht's
of importance here, because only the real power, i.e.,
that portion that is handled by the resistive component
of the impedance, is the ONLY part of the total power
that can actually make sound.

And, to complicate matters even more, those same
principles of real vs apparent (also called "imaginary"
as in sqrt(-1)) also affect the ACOUSTICAL power as
well. This is because the acoustical power is developed
across an acoustical "load" or impedance, which itself
has a real and apparent (imaginary) part, and the actual
dissipation of power (or first derivative of work wrt
time) is developed as actual sound.

There are other confounding factors. Among these are the
fact that the total acoustical power output can only be
sampled at discrete points, whether those points are
microphons or eardrums. They are sampling the total
power output of the system only over an extremely tiny
region of the total radiation space of the speaker.
Obviously, if the power over a spherical surface having
the speaker at its center is not constant, then any
assumptions about extrapolating the total power output
from a few tiny points are not going to be valid.

These are just a few of the confounding factors in
obtaining a true measure of "efficiency" as defined
above. You will seldom, if ever, see a true "efficiency"
measurement of a speaker, most especially consumer-
oriented loudspeakers (of which your AR3a is but one
example).

As it turns out, there are a couple of factors that aid
us in coming up with a usable figure. It's not
efficiency, but rather "sensitivity." Those factors
include:

1. The vast majority of amplifiers driving speakers are,
in the precise engineering sense of the term "voltage
sources." This means that the effective output
impedance of the source (amplifier) is insubstantial
when compared to the load (speaker). The consequence
of this is that most loudspeakers are designed to
achieve their frequency response when driven by a
constant voltage, not at constant power.

2. Over the range where the ear is most sensitive
(essentially, a loose definition of "midrange"), the
actual acoustic output of the speaker in most typical
listening positions is fairly independent of the
acoustical conditions in which the speaker finds
itself. That means as long as you are at or near what
is defined as the "principle listening axis"), the
response of the speaker over the critical listening
range is fairly independent of all but pathological
listening environmental conditions. (the room has the
greatest effect only in the bass and the lower
midrange: it's a big surprise to most people how
closely the in-room response of the loudpeaker
approximates the free-field response).

Given those two conditions are met, and they are met in
almost ever case in actual usage, then we can define an
alternative to a true "efficiency" called
"characteristic sensitivity," which is defined as
(quoted from IEC 168-5, section 14.3 "Characteristic
Sensitivity":

"The sound pressure produced at 1 m from the
reference point on the reference axis when the
loudspeaker is supplied with a pink-noise signal the
voltage of which corresponds TO A POWER OF 1 W in the
rated impedance. The bandwidth of the pink-noise
signal shall be limited to the rated frequency range
of the loudspeaker."

Notice the phrase "the voltage of which corresponds TO A
POWER OF 1 W in the rated impedance." (The emphasis is
as found in the standard.) Here we take care of the
variable impedance issue: it's the "rated impedance" of
the speaker (which is addressed in a different part of
the standard).

Using this definition of sensitivity, we have
encompassed three of the major issues outlined above:

1. The issue of defining power into a variable,
partially reactive impedance by using a voltage
reference and a nominal impedance value,

2. The issue of radiation pattern, potential room
loading issues at low frequencies, et al, by
specifying the measurement is done on the reference
axis at 1 meter from the reference point.

3. Using pink noise essentially eliminates the freqency-
dependent impedance variations as a factor. Integrate
the impedance of the loudspeaker over it's entire
rated range, and you'll find almost universally that
the result is VERY close to purely resistive.

Now, the 2.83 volt figure comes from the fact that if
the speaker has a rated impedance of 8 ohms, the voltage
needed to produce 1 watt into that 8 ohms is 2.83 volts.
For a 4 ohms speaker, it would be 2 volts.

However, the sensitivity figure can be stretched a wee
bit to incorporate what the "apparent" efficiency of
speakers with different rated impedances might be. Say
you have two loudspeakers of similar frequency range,
and of identical electroacoustic efficiency (say, 1%),
but one has a rated impedance of 8 ohms, the other is 4
ohms. With the volume control set to the same position,
guess what? The 4 ohm loudspeaker will sound louder than
the 8 ohm speaker, because it is putting out more
acoustics poweer. Why? because volume control set to the
same position, for the same musical poassages, the
amplifier produces the same voltage,, and that same
voltage produces more power in a lower impedance than in
a higher impedance. In fact, the 4 ohm loudspeaker wil
be producing 3 dB more sound pressure level for the same
input that the 8 ohm speaker will do.

To get back to your question more directly: your AR3a
has a "rated impedance" (aka "nominal impedance") of 4
ohms. That's what the manufacturer states in there
specifications, and that's what IEC 268-5 allows as the
figure. Into such a load, 2 volts RMS of pink noise
produces nominal 1 watt of power, thus for a 4 ohm
loudspeaker, the characteristic sensitivity would be
be measured under that driving voltage on the reference
axis of the speaker.

For a speaker with a rated impedance of 8 ohms, that
voltage would be 2.83 volts to produce the same nominal
1 watt.

If you're interested in calculating power budgets in a
multi-speaker system, efficiency is the more important
figure. But if you're calculating, for example, the
attenuation neede to match a midrange to a woofer, then
sensitvity from a fixed voltage is far more important,
since the two drivers are driven by the same amplifier
(assuming passive networks) or the relative figures are
needed for adjusting the gains of multiple amplifiers
(assuming multi-amp systems).

So, this is very long-winded, but it covers the problems
of measuring efficiency and why there exist "sensivity"
measurements: They're simply easier to perform and more
directly comparable, as long as the domain is well
defined. Sensitivity is much more a useful measurement
in your situation than raw efficiency. Efficiency is a
more significant figure when a concert engineer or an
acoustical architect is trying to put together a massive
sound system and is trying to decide whether they need
10,000 watts or 20,000 watts of amplifier power. That
difference has substantial economic and engineering
consequences that simply are of relatively minor
importance in a home listening situation by comparison.

Dick, thank you!!!

Your explanation wasn't long-winded, but very thorough and well thought out!

I just knew it wouldn't make any sense to measure the sensitivity at a fixed
voltage when speakers can have widely different impedance ratings. Under a
fixed voltage measurement system, the low impedance speakers would have a
tremendous advantage. Adjusting the rms input voltage accordingly really
"levels the playing field".

Next question, recently I saw some HT speakers advertised at 6 ohms and I've
seen HT amps rated at 6 ohms.

Dick, do you think 6 ohms is likely to become a new popular standard for
amps and speakers?

Regards,
Jerry

[[email protected]]

Jerry wrote:
Your explanation wasn't long-winded, but very
thorough and well thought out!

I just knew it wouldn't make any sense to measure
the sensitivity at a fixed voltage when speakers can
have widely different impedance ratings.

Actually, I cited one very common instance where
it DOES make sense: in matching woofers, midranges
and tweeters when designing passively crossed over
system

Under a fixed voltage measurement system, the low
impedance speakers would have a tremendous
advantage.

No, they WON'T have a "tremendous" advantage. If
you make the assumption that everything else is equal,
a 4 ohm speaker will ONLY have a 3 dB APPARENT
advantage, and 3 dB ain't much at all, certainly not
tremendous. Further, all things are NEVER equal. I
know of almost NO 8 ohm and 4 ohm loudspeakers
that are similar enough in overall response AND that
have the same efficiency. Take your AR3a's as but one
example: they're nominally rated at 4 ohms, but they are
VERY low efficiency speakers. Now, compare them to
a contemporary 8 ohm speaker, say, a JBL L100. They
are EASILY 10-12 dB more efficient, and even using the
same voltage, the JBL's will most assuredly ALWAYS
play louder.

Adjusting the rms input voltage accordingly really
"levels the playing field".

But doing so simply ignore all of the other very large
mountains and canyons remaining on the field. The
point of my long article is that it is NEVER so simple.

Next question, recently I saw some HT speakers
advertised at 6 ohms and I've seen HT amps rated
at 6 ohms.

Dick, do you think 6 ohms is likely to become a new
popular standard for amps and speakers?

It has little if not nothing to do with "popularity," it's
a matter of manufacturing practicality.

Speakers are made out of magnets and voice coils
and such (obviously). What's not so obvious to most
people is that the components that make up a speaker
are simply not available in arbitrary sizes. For example,
the front plate thicknesses for the magnet assemblies
are available in only about 3 standard thicknesses. That
sets the depth of the magnet gap. The wire used in the
voice coil is only available in fixed gauges, and the basic
limitations of winding geometry mean that you can have
only an even number of layers in the voice coil.

As far as the magnet material itself (e.g. barium ferrite
ceramics), even though they are available in a wide
variety of sizes, it makes no sense to use anything but
the sized magnet that's sufficient to run the magnet gap
right at saturation. Anything more is simply a waste of
material, anything less is unstable.

So, you take these three properties, and you find very
quickly that your DC resistance and your Bl product
cannot have any arbitrary value, but rather are almost
"quantized" in a series of discrete values. For woofers,
DC resistances seem to want to fall in the 6.2-6.8 realm
(for a nominal 8 ohm impedance), around 5.1-5.5 (for
a nominal 6 ohm impedance), about 2.9-3.5 (for a nominal
4 ohm impedance) and so on.

The same sort of "quantization" is true for many cone
materials, surrounds, and so on. There is not a
continuum of arbitrayr values for each of the parameters
that go into making a loudspeaker.

Now, another problem you run into is that many modern
HT amplifiers do not behave well into low impedance
speakers, yet there are advantages to low impedance
loads for a variety of reasons. One example is that if
you're building passice networks, the lower the impedance,
the less copper you need in the inductors to cross them
over, and copper is not cheap.

So 6 ohms is a compromise constrained in many
dimensions: It has to be right about there because
of the discrete nature of the components that make
up the speaker, it's low enough to take advantage
of low impedance, it's high enough not to be too
taxing on lower-end HT receivers (where the bulk of
the market is).

As to "popularity," rest assured that no manufacturer
ever bothered to ask the consumer what impedance
they wanted. The answers would be pure noise anyway.

[Jerry]

Dick wrote in message on 9/20/2006:

Jerry wrote:
Your explanation wasn't long-winded, but very
thorough and well thought out!

I just knew it wouldn't make any sense to measure
the sensitivity at a fixed voltage when speakers can
have widely different impedance ratings.

Actually, I cited one very common instance where
it DOES make sense: in matching woofers, midranges
and tweeters when designing passively crossed over
system

Dick, I should have been clearer. My question was really about speaker
systems and comparisons between systems to determine minimum amp
requirements.

I would think that when attempting to match drivers within a speaker, you
have a fairly complex problem. Not only do the speakers differ, but then
the cabinets and frequency separating circuits will impact the SPL produced
by an individual driver. It gives me a headache just thinking about it.

Under a fixed voltage measurement system, the low
impedance speakers would have a tremendous
advantage.

No, they WON'T have a "tremendous" advantage. If
you make the assumption that everything else is equal,
a 4 ohm speaker will ONLY have a 3 dB APPARENT
advantage, and 3 dB ain't much at all, certainly not
tremendous. Further, all things are NEVER equal. I
know of almost NO 8 ohm and 4 ohm loudspeakers
that are similar enough in overall response AND that
have the same efficiency. Take your AR3a's as but one
example: they're nominally rated at 4 ohms, but they are
VERY low efficiency speakers. Now, compare them to
a contemporary 8 ohm speaker, say, a JBL L100. They
are EASILY 10-12 dB more efficient, and even using the
same voltage, the JBL's will most assuredly ALWAYS
play louder.

Well, we were talking about Allisons vs AR's. I don't think anyone would
claim either is particularily efficient.

Adjusting the rms input voltage accordingly really
"levels the playing field".

But doing so simply ignore all of the other very large
mountains and canyons remaining on the field. The
point of my long article is that it is NEVER so simple.

No question. The Allisons among other things are designed to solve some
frequency "suck out" problems that exist in AR's.

Next question, recently I saw some HT speakers
advertised at 6 ohms and I've seen HT amps rated
at 6 ohms.

Dick, do you think 6 ohms is likely to become a new
popular standard for amps and speakers?

It has little if not nothing to do with "popularity," it's
a matter of manufacturing practicality.

Speakers are made out of magnets and voice coils
and such (obviously). What's not so obvious to most
people is that the components that make up a speaker
are simply not available in arbitrary sizes. For example,
the front plate thicknesses for the magnet assemblies
are available in only about 3 standard thicknesses. That
sets the depth of the magnet gap. The wire used in the
voice coil is only available in fixed gauges, and the basic
limitations of winding geometry mean that you can have
only an even number of layers in the voice coil.

As far as the magnet material itself (e.g. barium ferrite
ceramics), even though they are available in a wide
variety of sizes, it makes no sense to use anything but
the sized magnet that's sufficient to run the magnet gap
right at saturation. Anything more is simply a waste of
material, anything less is unstable.

So, you take these three properties, and you find very
quickly that your DC resistance and your Bl product
cannot have any arbitrary value, but rather are almost
"quantized" in a series of discrete values. For woofers,
DC resistances seem to want to fall in the 6.2-6.8 realm
(for a nominal 8 ohm impedance), around 5.1-5.5 (for
a nominal 6 ohm impedance), about 2.9-3.5 (for a nominal
4 ohm impedance) and so on.

The same sort of "quantization" is true for many cone
materials, surrounds, and so on. There is not a
continuum of arbitrayr values for each of the parameters
that go into making a loudspeaker.

Now, another problem you run into is that many modern
HT amplifiers do not behave well into low impedance
speakers, yet there are advantages to low impedance
loads for a variety of reasons. One example is that if
you're building passice networks, the lower the impedance,
the less copper you need in the inductors to cross them
over, and copper is not cheap.

So 6 ohms is a compromise constrained in many
dimensions: It has to be right about there because
of the discrete nature of the components that make
up the speaker, it's low enough to take advantage
of low impedance, it's high enough not to be too
taxing on lower-end HT receivers (where the bulk of
the market is).

As to "popularity," rest assured that no manufacturer
ever bothered to ask the consumer what impedance
they wanted. The answers would be pure noise anyway.

Dick, thanks for taking the time to answer my questions. Your description
of the "quantization" of speaker impedances is truly enlightening. My guess
is while "popular" might not be the correct word, it seems that a number of
HT amp manufacturers are following the lead of the speaker manufacturers and
specifying their amp's power at 6 ohms.

It's probably also a marketing "gimmick". By that I mean those crumby
little amps probably cannot handle a 4 ohm load. Since they can handle a 6
ohm load and since the output wattage for 6 ohms is HIGHER than for 8 ohms,
they quote the 6 ohm wattage. This allow them to claim total power of
1000's of watts. Of course it's all BS, as most ratings are NOT full range,
but at 1000 Hz.

Regards,
Jerry

[[email protected]]

Jerry wrote:
Dick wrote in message on 9/20/2006:
Actually, I cited one very common instance where
it DOES make sense: in matching woofers, midranges
and tweeters when designing passively crossed over
system

I would think that when attempting to match drivers
within a speaker, you have a fairly complex problem.
Not only do the speakers differ, but then the cabinets
and frequency separating circuits will impact the SPL
produced by an individual driver. It gives me a
headache just thinking about it.

While that's all true, it still does not chnage the fact that
you need t match the drivers on power efficiency, but
instead on voltage sensitivity. All these other factors do
not change that requirement at all, indeed, the make it
all the more important.

Under a fixed voltage measurement system, the low
impedance speakers would have a tremendous
advantage.

No, they WON'T have a "tremendous" advantage. If
you make the assumption that everything else is equal,
a 4 ohm speaker will ONLY have a 3 dB APPARENT
advantage, and 3 dB ain't much at all, certainly not
tremendous. Further, all things are NEVER equal. I
know of almost NO 8 ohm and 4 ohm loudspeakers
that are similar enough in overall response AND that
have the same efficiency. Take your AR3a's as but one
example: they're nominally rated at 4 ohms, but they are
VERY low efficiency speakers. Now, compare them to
a contemporary 8 ohm speaker, say, a JBL L100. They
are EASILY 10-12 dB more efficient, and even using the
same voltage, the JBL's will most assuredly ALWAYS
play louder.

Well, we were talking about Allisons vs AR's. I don't think
anyone would claim either is particularily efficient.

I bleieve you original post on this particular thread mentioned
neither. I picked two speakers available during the same
period of time that were commondly available and relatively
popular. The differences in the efficiency between these two
and, indeed, MOST speakers is NOT because of the differences
in impedance, but more fundamental differences in the
design and implementation. The resulting impedance can
often be merely a consequence of the design.

Adjusting the rms input voltage accordingly really
"levels the playing field".

But doing so simply ignore all of the other very large
mountains and canyons remaining on the field. The
point of my long article is that it is NEVER so simple.

No question. The Allisons among other things are
designed to solve some frequency "suck out" problems
that exist in AR's.

No, that's not what I am tlaking about. The "suckout"
problem (which is not an AR3a problem, but a problem
with ANY source placed some distance from a reflective
boundary) is, at best, a tertiary problem that has little
impact on overall efficiency.

WHat I was talking about, and I'll restate in the hopes
that it will be clearer, the difference in power conversion
efficiency resulting from differences in impedance is
NOT an issue of the magnitdue you seem to think it
is. It's almost unimportant. And the primary reason is
that the difference in efficiency between speakers is
NOT primarily due to differences in impedance: there are
far more fundamental agents to these differences than
that. The AR3a vs JBL L100 proves that point. The L100
is mush more efficient NOT because the impedance is
lower (in fact, it's HIGHER), but primarily because the
moving mass of the cone is substantially lower AND
because the Bl product is higher.

As to "popularity," rest assured that no manufacturer
ever bothered to ask the consumer what impedance
they wanted. The answers would be pure noise anyway.

Dick, thanks for taking the time to answer my questions.
Your description of the "quantization" of speaker impedances
is truly enlightening. My guess is while "popular" might
not be the correct word, it seems that a number of
HT amp manufacturers are following the lead of the speaker
manufacturers and specifying their amp's power at 6 ohms.

No, it's quite the other way around. The design-manufacturer-
market cycle for a receiver is FAR more expensive and MUCH
longer than that for a speaker. That which is easier to change
(the speaker design) is FAR more likely to follow the lead of
which has the longer and more expensive lead times (receiver).

It's probably also a marketing "gimmick". By that I mean
those crumby little amps probably cannot handle a 4 ohm
load. Since they can handle a 6 ohm load and since the
output wattage for 6 ohms is HIGHER than for 8 ohms,
they quote the 6 ohm wattage. This allow them to claim
total power of 1000's of watts. Of course it's all BS, as
most ratings are NOT full range, but at 1000 Hz.

Come on, do the math. All other things being equal, the
MOST you can claim for an amplifier driving 6 ohms vs
8 ohms is about 33% more power. 1.25 dB. Big freakin'
deal. And if the amplifier is already in serious current limiting
at 4 ohms, you can bet it's nitin' at the margins at 6 and likely
at 8 as well. So, from a purely technical viewpoint, it's
a wash.

And, if you assertion is correct that it's all gimmikery
(which I will not take the time to argue), then why would
a manufacturer who's going to resort to such even
bother with matters like "impedance?" Why not just make
up any ol' number as, in fact, some do?

There is still such a thing as the FTC amplifier rule,
which goes some way to providing a common means of
comparison.

Having, in fact, been involved with a number of manufacturers,
most of whom were trying to some extent to make a decent
effort and reasonably representing their products, one thing
is very apparent: "power" is a feature, not a property. A high
"power" commands as much attention with the average
consumer and is as important as the number of buttons
and displays and surround modes. The REAL market for
true voltage-source amplifiers with capabilities of driving
wide ranges of impedance is EXTREMELY small when
compared to the total consumer audio market. Indeed, if
that portion of the market where to suddenly vanish altogether
or, alternatively, double overnight, it would have essentially no
economic impact on the total market.

You provide the technical evidence to support this assertion.
In another thread, you stated on a number of occasions that
you spent most of your time listening at levels that drew very
small amouints of power from the amplifier, like a watt. This
is just as true for the vast majority of HT users, even when
listening to movies. The total amount of power used the vast
majority of times is simply not very great. And many of these
receivers that might have a hard time driving full power into
a 4 ohm load are perfectly happy doing 1/3 of their rated power
and less.

[Jerry]

Dick wrote on 9/24/2006:
Jerry wrote:

It's probably also a marketing "gimmick". By that I mean
those crumby little amps probably cannot handle a 4 ohm
load. Since they can handle a 6 ohm load and since the
output wattage for 6 ohms is HIGHER than for 8 ohms,
they quote the 6 ohm wattage. This allow them to claim
total power of 1000's of watts. Of course it's all BS, as
most ratings are NOT full range, but at 1000 Hz.

Come on, do the math. All other things being equal, the
MOST you can claim for an amplifier driving 6 ohms vs
8 ohms is about 33% more power. 1.25 dB. Big freakin'
deal. And if the amplifier is already in serious current limiting
at 4 ohms, you can bet it's nitin' at the margins at 6 and likely
at 8 as well. So, from a purely technical viewpoint, it's
a wash.

Dick, 33% to a marketing executive is HUGE!! Some of the HT systems have 7
amps and often the advertising is showing the combined power output of all 7
channels.

Then, when you read the "fine print" you find that the power ratings are NOT
full range, but just at 1000 Hz. JVC is just one example, although I read
in a forum where JVC is no longer the corporation it once was.

And, if you assertion is correct that it's all gimmikery
(which I will not take the time to argue), then why would
a manufacturer who's going to resort to such even
bother with matters like "impedance?" Why not just make
up any ol' number as, in fact, some do?

There is still such a thing as the FTC amplifier rule,
which goes some way to providing a common means of
comparison.

Having, in fact, been involved with a number of manufacturers,
most of whom were trying to some extent to make a decent
effort and reasonably representing their products, one thing
is very apparent: "power" is a feature, not a property. A high
"power" commands as much attention with the average
consumer and is as important as the number of buttons
and displays and surround modes. The REAL market for
true voltage-source amplifiers with capabilities of driving
wide ranges of impedance is EXTREMELY small when
compared to the total consumer audio market. Indeed, if
that portion of the market where to suddenly vanish altogether
or, alternatively, double overnight, it would have essentially no
economic impact on the total market.

Dick, I suspect you are correct. In the box stores like Circus City and
Worst Buy, what you see advertised are complete HT sound systems. Looks
like the vast majority of consumers purchase packages that include amps and
speakers. So, neither the manufacturers nor the consumers worry much about
matching the impedance of speakers to amps.

And this isn't all. The other "hot" audio market is docking stations for
ipods.

My son can't go anywhere without his ipod and I just can't get into it.
Alas, technology is passing me by.

You provide the technical evidence to support this assertion.
In another thread, you stated on a number of occasions that
you spent most of your time listening at levels that drew very
small amouints of power from the amplifier, like a watt. This
is just as true for the vast majority of HT users, even when
listening to movies. The total amount of power used the vast
majority of times is simply not very great. And many of these
receivers that might have a hard time driving full power into
a 4 ohm load are perfectly happy doing 1/3 of their rated power
and less.

No doubt! I would NOT be surprised to find that half of the total power
used is in the powered sub-woofer.

I think I am firmly stuck in the past, because I just can't listen to music
with those synthesized channels. I still really enjoy my AR-3a's and two
channel amps for music.

Now for movies (and I watch very, very few) I think digital dolby in
multi-channel is pretty entertaining. I don't like that system very much at
all when it comes to listening to cd's.

Regards,
Jerry