[Gary Eickmeier]

All right, I am going to attempt to burst the dam of silence on RAHE by
telling about the new speakers that I have designed and had built by the
capable hands of Dan Neubecker in Indiana. He didn't debut it in the
InDIYana event because the venue was not suitable for this type of speaker.

Brieflly, this is the embodiment of all of my Image Model Theory tenets on
how a speaker should be designed. I said that what we hear in a speaker is
its image model, which is the radiation pattern as reflected by the surfaces
nearby in a pattern of actual and virtual speakers, or sources. I want my
image model to look like a lattice of evenly spaced sources like this:

https://s3.amazonaws.com/masters.gal...81994e4099169d

To achieve this image model, the speaker needed to have more output in the
reflected direction, and a certain "shaped" frontal radiation, so we put two
gain pots on the two front panels to be able to reduce the frontal output
and at the same time do the distance/intensity trading effect called for by
Mark Davis in the Soundfield One speaker. We have settled on a -9 dB for the
outer front panels and - 3 dB for the inner panels. The speakers thus had to
have drivers on four sides of a square cross section box with MTM drivers on
each face, something like this:

https://s3.amazonaws.com/masters.gal...8ab783642bb6ef

and the final design looks like this:

https://s3.amazonaws.com/masters.gal...98e22052d64852

which, in my room, looks like:

https://s3.amazonaws.com/masters.gal...1be93b6797f2cc

I hope all of these images come through OK.

The room is 21 x 31 ft, the speakers positioned 5 ft out from the front
wall, 5 ft in from the side walls, for an image model that has all real and
virtual sources spaced evenly 10 ft apart.

The sound is as spacious as it gets, has very sharp imaging, speakers
disappearing entirely and casting a sound field behind, between, and to the
sides beyond the separation of the speakers in a way that makes it seem like
the musicians are right there with you. This is a sat/sub system with the
IMPs doing the satellite part and a Velodyne F-1800 doing the subwoofer
chores. It is balanced, full range, spacious, and precise.

I have yet to test them against other high end speakers, but I don't know
what else a speaker can do better in any of the audible areas of speaker
performance. Frequency response, radiation pattern, and room positioning,
that's all there is, there is no more.

Questions? Comments?

Gary Eickmeier

[Andrew Haley]

Gary Eickmeier wrote:

Questions? Comments?

Sorry, those links don't work for me. I suggest you use tinyurl.com

Andrew.

[Gary Eickmeier]

Well, I must apologize for the links. They are not coming through for me on
this end, though they worked fine last night when I checked my SENT file. I
have used this site for posting pix before and it worked fine. Anyone
getting them OK?

Gary Eickmeier

[jason]

On 31 Aug 2014 02:24:08 GMT "Gary Eickmeier"
wrote in article

Moderators, fourth time is the last::
....

Questions? Comments?

Gary Eickmeier

I tried one of the links and it said permission from the poster was
necessary.

[Gary Eickmeier]

I have joined Flickr and posted my pix to that site, bringing them up one at
a time and copy and pasting the resulting URL. Let's see if it works this
time.

The image model drawing showing the two real and 6 virtual image speakers as
reflections from the other side of the walls nearby:

https://www.flickr.com/photos/127390...4/15089905192/

The basic design of the speakers, showing the drivers on all 4 faces in plan
and straight views:

https://www.flickr.com/photos/127390...4/15090259565/

The resulting speaker:

https://www.flickr.com/photos/127390...4/15089946922/

https://www.flickr.com/photos/127390...4/14903747037/

https://www.flickr.com/photos/127390...4/15067308146/

and how they are positioned in my room:

https://www.flickr.com/photos/127390...4/14903723228/

I certainly hope this works out this time!

Gary Eickmeier

[Andrew Haley]

Gary Eickmeier wrote:

The room is 21 x 31 ft, the speakers positioned 5 ft out from the front
wall, 5 ft in from the side walls, for an image model that has all real and
virtual sources spaced evenly 10 ft apart.

The sound is as spacious as it gets, has very sharp imaging, speakers
disappearing entirely and casting a sound field behind, between, and to the
sides beyond the separation of the speakers in a way that makes it seem like
the musicians are right there with you. This is a sat/sub system with the
IMPs doing the satellite part and a Velodyne F-1800 doing the subwoofer
chores. It is balanced, full range, spacious, and precise.

I have yet to test them against other high end speakers, but I don't know
what else a speaker can do better in any of the audible areas of speaker
performance. Frequency response, radiation pattern, and room positioning,
that's all there is, there is no more.

Questions? Comments?

Okay, I'll bite. I don't doubt that this arrangement would present a
huge soundstage. However, that first diagram, which shows laser-like
sound coming from the rear of the speakers is rather fanciful. In
reality, there will be a dispersal pattern which varies with frequency
and there won't be so much clean reflected virtual speakers as you
illustrate.

The other thing is that would worry me is that room looks extremely
lively, and I'm surprised it isn't plagued by flutter echo. Are there
some diffusers (books, etc.) on the back wall we can't see? I have a
somewhat smaller room which had similar properties, and the echo was
mostly cured by filling the back wall with books floor-to-ceiling and
some large pieces of furniture, neither of which your room has.

Andrew.

[Gary Eickmeier]

Andrew Haley wrote:
Gary Eickmeier wrote:

The room is 21 x 31 ft, the speakers positioned 5 ft out from the
front wall, 5 ft in from the side walls, for an image model that has
all real and virtual sources spaced evenly 10 ft apart.

The sound is as spacious as it gets, has very sharp imaging, speakers
disappearing entirely and casting a sound field behind, between, and
to the sides beyond the separation of the speakers in a way that
makes it seem like the musicians are right there with you. This is a
sat/sub system with the IMPs doing the satellite part and a Velodyne
F-1800 doing the subwoofer chores. It is balanced, full range,
spacious, and precise.

I have yet to test them against other high end speakers, but I don't
know what else a speaker can do better in any of the audible areas
of speaker performance. Frequency response, radiation pattern, and
room positioning, that's all there is, there is no more.

Questions? Comments?

Okay, I'll bite. I don't doubt that this arrangement would present a
huge soundstage. However, that first diagram, which shows laser-like
sound coming from the rear of the speakers is rather fanciful. In
reality, there will be a dispersal pattern which varies with frequency
and there won't be so much clean reflected virtual speakers as you
illustrate.

Andrew -

No, it doesn't show laser like sound coming from the rear. That is just the
ray tracing technique for showing the path of reflected sound. If you take a
look at the desired radiation pattern at the lower left, you can see how I
wanted the mid to high frequencies to launch from the speaker in order to
achieve the image model above. Low freqs are always omni. But the IMP
satellites have nothing below 100 Hz. Not sure exactly what range it starts
to take on the desired radpat, but all you can do is aim the four panels as
shown and adjust their gains and see if the model takes shape or not.


The other thing is that would worry me is that room looks extremely
lively, and I'm surprised it isn't plagued by flutter echo. Are there
some diffusers (books, etc.) on the back wall we can't see? I have a
somewhat smaller room which had similar properties, and the echo was
mostly cured by filling the back wall with books floor-to-ceiling and
some large pieces of furniture, neither of which your room has.

Andrew.

Good question, and the picture doesn't zoom back far enough to show the wall
treatments. I made a half dozen half round absorber panels and mounted them
on the side walls staggered, not directly across from each other, so that
the laterals would be broken or absorbed before being reflected once again.
There is more and more absorption as you go back, so that most lateral and
diagonal waves will not return to be reflected again. Bottom line, there is
no flutter echo, as indicated by clapping my hands anywhere in the room.
Reflectivity is desired in the front of the room for creation of the model,
which is basically the big secret of most dipoles, bipoles, and omnis. Very
spacious sound with speakers disappearing and leaving nothing but the music.

The caution with reflecting type speakers is that the speakers MUST be
positioned very precisely or else they will lose focus. If those reflected
(virtual image) speakers are not evenly spaced you will get stretched
soloists and no center imaging. If any one of the elements of the model is
missing, radiation pattern, speaker positioning, and room acoustics, you
could miss these qualities and never discover all of the factors of stereo
imaging. I stumbled upon it by accident, then pursued the idea until I
finally figured it all out and had these built to do the concept on purpose
rather than by accident! The whole room was designed around the theory about
25 years ago, and now I am finally there!

Gary Eickmeier

[Andrew Haley]

Gary Eickmeier wrote:
Andrew Haley wrote:
Okay, I'll bite. I don't doubt that this arrangement would present a
huge soundstage. However, that first diagram, which shows laser-like
sound coming from the rear of the speakers is rather fanciful. In
reality, there will be a dispersal pattern which varies with frequency
and there won't be so much clean reflected virtual speakers as you
illustrate.

No, it doesn't show laser like sound coming from the rear. That is
just the ray tracing technique for showing the path of reflected
sound. If you take a look at the desired radiation pattern at the
lower left, you can see how I wanted the mid to high frequencies to
launch from the speaker in order to achieve the image model above.
Low freqs are always omni. But the IMP satellites have nothing below
100 Hz. Not sure exactly what range it starts to take on the desired
radpat, but all you can do is aim the four panels as shown and
adjust their gains and see if the model takes shape or not.

What frequency is that radiation pattern? It'd be interesting to know
how the pattern of those speakers varies with frequency.

The precedence effect predicts that these reflections will give the
impression of spaciousness but the second image won't be distinctly
separate as long as the walls are broadband reflectors: i.e. the
reflected sound has a similar spectrum to that of the direct sound.
So, nothing surprising there.

The caution with reflecting type speakers is that the speakers MUST
be positioned very precisely or else they will lose focus. If those
reflected (virtual image) speakers are not evenly spaced you will
get stretched soloists and no center imaging. If any one of the
elements of the model is missing, radiation pattern, speaker
positioning, and room acoustics, you could miss these qualities and
never discover all of the factors of stereo imaging. I stumbled upon
it by accident, then pursued the idea until I finally figured it all
out and had these built to do the concept on purpose rather than by
accident! The whole room was designed around the theory about 25
years ago, and now I am finally there!

What does this theory predict that conventional psychoacoustics does
not? In other words, is it possible to design an experiment which
would show whether the IMP is a superior theory to others?

Andrew.

[Gary Eickmeier]

"Andrew Haley" wrote in message
...


What does this theory predict that conventional psychoacoustics does
not? In other words, is it possible to design an experiment which
would show whether the IMP is a superior theory to others?

Andrew.

Sure. Just violate any of the three ingredients of the model. If a speaker
puts out all direct sound, as most do even today, then the majority of the
high freqs will be heard from the actual speakers only, limiting the depth
impression and spaciousness. If the speakers are mis-positioned, as in
placed too close to one or more reflecting surfaces, the model will collapse
to a clustering of sources near those reflections, leading to stretched
soloists and a hole in the middle. We have noticed these effects for a long
time now, but not realized what causes them. Finally, if the surfaces around
the speakers have been padded with a lot of sound absorbing material, the
back and side reflections will be greatly diminished in the high
frequencies, leading once again to a smaller soundstage with less depth.

I hope to demonstrate these effects with some AB testing if and when I can
get a few comparison speakers. Also note that these effects diminish with
increasing room size. Large room acoustics tend to swamp a lot of these
spatial effects of the playback system and also the acoustic signature of
the recording (the concert hall acoustic, if any was recorded).

Gary Eickmeier

[Andrew Haley]

Gary Eickmeier wrote:
"Andrew Haley" wrote in message
...

What does this theory predict that conventional psychoacoustics does
not? In other words, is it possible to design an experiment which
would show whether the IMP is a superior theory to others?

Sure. Just violate any of the three ingredients of the model. If a
speaker puts out all direct sound, as most do even today, then the
majority of the high freqs will be heard from the actual speakers
only, limiting the depth impression and spaciousness.

That's pretty conventional; in particular, Toole describes it at some
length.

If the speakers are mis-positioned, as in placed too close to one or
more reflecting surfaces, the model will collapse to a clustering of
sources near those reflections, leading to stretched soloists and a
hole in the middle. We have noticed these effects for a long time
now, but not realized what causes them.

That's not unconventional either.

Finally, if the surfaces around the speakers have been padded with a
lot of sound absorbing material, the back and side reflections will
be greatly diminished in the high frequencies, leading once again to
a smaller soundstage with less depth.

Or that. Again, Toole describes the benefits (in terms of perceived
spaciousness) of reflected sound, and in particular the benefits of
the reflected sound having the same spectrum as the direct sound. He
also talks about the precedence effect, and how it applies in real
rooms with reflective surfaces

I can't see anything at all in what you have descibed which is not in
accordance with well-known principles. Sure, it's not how speakers
(except some electrostatics and perhaps the Linkwitz designs) are
usually made, but that's a different thing altogether.

Andrew.

[Gary Eickmeier]

Andrew Haley wrote:
Gary Eickmeier wrote:
"Andrew Haley" wrote in message
...

What does this theory predict that conventional psychoacoustics does
not? In other words, is it possible to design an experiment which
would show whether the IMP is a superior theory to others?

Sure. Just violate any of the three ingredients of the model. If a
speaker puts out all direct sound, as most do even today, then the
majority of the high freqs will be heard from the actual speakers
only, limiting the depth impression and spaciousness.

That's pretty conventional; in particular, Toole describes it at some
length.

If the speakers are mis-positioned, as in placed too close to one or
more reflecting surfaces, the model will collapse to a clustering of
sources near those reflections, leading to stretched soloists and a
hole in the middle. We have noticed these effects for a long time
now, but not realized what causes them.

That's not unconventional either.

Finally, if the surfaces around the speakers have been padded with a
lot of sound absorbing material, the back and side reflections will
be greatly diminished in the high frequencies, leading once again to
a smaller soundstage with less depth.

Or that. Again, Toole describes the benefits (in terms of perceived
spaciousness) of reflected sound, and in particular the benefits of
the reflected sound having the same spectrum as the direct sound. He
also talks about the precedence effect, and how it applies in real
rooms with reflective surfaces

I can't see anything at all in what you have descibed which is not in
accordance with well-known principles. Sure, it's not how speakers
(except some electrostatics and perhaps the Linkwitz designs) are
usually made, but that's a different thing altogether.

Andrew.

OK, good, so it is all in accordance with "sound" principles. As I said, we
have observed a lot of these qualities from other speakers but not put it
all together in one theory of how it all fits together. We had the Bose 901
with its huge sound, but they did not specify the speaker positioning and
the importance of it. Also, it could have used a little more direct sound,
and not aimed straight ahead as they did it. We had the DBX Soundfield One,
with its distance/ intensity trading effect for off center listeners. But it
didn't have the negative directivity of more sound in the reflected
direction, or the speaker placement instructions. My design combines all of
these principles under the concept of Image Model Theory for stereophonic
sound, which says that the reproduction should be a reconstruction, or a 3
dimensional model, of the original sound fields rather than a "window" into
another acoustic space. Anything less will change the spatial
characteristics of the original to those of your speakers and room. In fact
that is always what we are doing in playback, which is why the playback
model should be as much like the live model as physically possible.

It seems to work, and work well.

Gary Eickmeier

[Gary Eickmeier]

"ScottW" wrote in message
...
"Gary Eickmeier" wrote in message
...
"Andrew Haley" wrote in message
...


What does this theory predict that conventional psychoacoustics does
not? In other words, is it possible to design an experiment which
would show whether the IMP is a superior theory to others?

Andrew.

Sure. Just violate any of the three ingredients of the model. If a
speaker
puts out all direct sound, as most do even today, then the majority of
the
high freqs will be heard from the actual speakers only, limiting the
depth
impression and spaciousness.

I find the most realistic depth and spatial impressions come from cues in
the
recording.
Not from cues imposed on playback by the room and room speaker
interactions.
When I was listening to dipoles (Quad 63s) I used rear wave diffusers
(idea I
took from Sound Labs Sallie) and while the impact was subtle I felt it
clarified
the image and increased spaciousness and depth. Multiple secondary
arrivals
just muddled and tended to flatten the image even when they were
substantially
delayed by distance to the walls.
So I cannot agree that high freqs direct from the speakers contributes to
a
limited depth and spaciousness impression. That is dependent upon the
recording. Some recordings will benefit from a direct approach with room
effects minimized while others may benefit from your approach.

ScottW

But don't you get the impression also that the soundstage is limited to
speaker to speaker in such a setup? There seems to be two schools of thought
in playback techniques. There is the let's hear the speakers only and
eliminate the room school, and the float a huge soundstage by means of
reflection from nearby surfaces school.

I say that in all cases what we hear is the presentation in front of us, and
that whatever you are doing with your speakers and room, that is what you
will change the recorded presentation to. In other words, simplest example,
play it back on a boombox with 10 inches separation and that is what you
will hear - it will sound like it is coming from a boombox, not from
Nashville. Play it on some Wilson WAMMs, and it will sound like big speakers
with a lot of bass and whatever the separation of the speakers is. Play it
on some Quads or Maggies or Martin Logans, and it will expand to fill the
front of the room in a larger soundstage with greater depth and out of
speaker imaging.

That is what I hear. On the more direct firing speakers in a dead room, to
me it sounds like the sound is coming from those two boxes rather than a
larger space. The left and right images collect at the speaker grills and ai
lose all suspension of disbelief.

I don't know of any way around all that except an entirely different system
known as binaural, in which each ear is fed its own signal with the
intention of fooling them into being in the original space. But try and do
that with normal stereo and legacy recordings and normal interaural
crosstalk, and your ears will not be fooled, they will hear whatever the
spatial qualities of your presentation happen to be. The sound will be heard
to be coming from where it IS coming, and little you can do about it except
change those spatial qualities to something more like the real thing.

And so my Image Model Theory. The "image model" of the sound presentation in
front of us is the umbrella explanation of all of the above spatial
qualities that we can hear. So I am saying that the object is to make the
image model of the playback as close to that of the live event as possible.
Take two channels, add a dose of direct sound, mix well with reflected in a
certain pattern, and you've got yerself a soundstage modeled after the live
fields.

It seems to work for me.

Gary Eickmeier

[Oregonian Haruspex]

On 2014-08-30 13:19:35 +0000, Gary Eickmeier said:

All right, I am going to attempt to burst the dam of silence on RAHE by
telling about the new speakers that I have designed and had built by the
capable hands of Dan Neubecker in Indiana. He didn't debut it in the
InDIYana event because the venue was not suitable for this type of speaker.

Brieflly, this is the embodiment of all of my Image Model Theory tenets on
how a speaker should be designed. I said that what we hear in a speaker is
its image model, which is the radiation pattern as reflected by the surfaces
nearby in a pattern of actual and virtual speakers, or sources. I want my
image model to look like a lattice of evenly spaced sources like this:

https://s3.amazonaws.com/masters.gal...81994e4099169d


To achieve this image model, the speaker needed to have more output in the
reflected direction, and a certain "shaped" frontal radiation, so we put two
gain pots on the two front panels to be able to reduce the frontal output
and at the same time do the distance/intensity trading effect called for by
Mark Davis in the Soundfield One speaker. We have settled on a -9 dB for the
outer front panels and - 3 dB for the inner panels. The speakers thus had to
have drivers on four sides of a square cross section box with MTM drivers on
each face, something like this:

https://s3.amazonaws.com/masters.gal...8ab783642bb6ef


and the final design looks like this:

https://s3.amazonaws.com/masters.gal...98e22052d64852


which, in my room, looks like:

https://s3.amazonaws.com/masters.gal...1be93b6797f2cc


I hope all of these images come through OK.

The room is 21 x 31 ft, the speakers positioned 5 ft out from the front
wall, 5 ft in from the side walls, for an image model that has all real and
virtual sources spaced evenly 10 ft apart.

The sound is as spacious as it gets, has very sharp imaging, speakers
disappearing entirely and casting a sound field behind, between, and to the
sides beyond the separation of the speakers in a way that makes it seem like
the musicians are right there with you. This is a sat/sub system with the
IMPs doing the satellite part and a Velodyne F-1800 doing the subwoofer
chores. It is balanced, full range, spacious, and precise.

I have yet to test them against other high end speakers, but I don't know
what else a speaker can do better in any of the audible areas of speaker
performance. Frequency response, radiation pattern, and room positioning,
that's all there is, there is no more.

Questions? Comments?

Gary Eickmeier

What sort of test setup did you use to measure the results of your
design? Do you have a favorite set of test mics? What sort of
measurement instruments do you use to confirm that your speakers match
the mathematical model of how they should perform?

[Gary Eickmeier]

"Oregonian Haruspex" wrote in message
...

What sort of test setup did you use to measure the results of your design?
Do you have a favorite set of test mics? What sort of measurement
instruments do you use to confirm that your speakers match the
mathematical model of how they should perform?

Dear Oregon,

As I believe I have mentioned, I did not build them, a very talented DIY
speaker builder in Indiana built them to my desires. He has not written up a
description of the design factors, just said some minimum facts as follows:

"I haven't written anything on the IMP. I certainly have my crossover
schematics, but that is really not something I would expect to see posted
for something like this. It would make sense to write about how the back
panels share one passive crossover and each front panel has its own passive
XO, with adjustable lpad on the amp side, before the crossover, to allow for
control of the radiation pattern. Perhaps something how the cabinet is
divided vertically into 4 triangular shaped enclosure spaces, one for each
MTM to reduce standing waves. I could include language about the octagonal
shape with large 45º adjacent to all driver panels to reduce diffraction.
The MTM configuration to reduce the effect of ceiling and floor bounce
cancellations. Also, that each panel is designed such that the frequency
response is essentially flat on axis, as is the FR when measured in the
center between any two sets of drivers, when only those two panels are
turned on. It was assumed that power response would be relatively flat in
room with this approach and the long gated frequency response measurements
appear to support this. Anyway, I'm not sure there would be that much to
include in technical details."

I know he uses the windowed MLS technique for measurment, but not which
mikes. This one may have been easier to measure because it is just a
satellite speaker and you don't need to worry about the deep bass response.

But the design goal had not so much to do with frequency response as
radiation pattern. We start with a basic box radiating in four directions
evenly, then put some L-pads on the front two panels to be able to attenuate
the frontal pattern in a way that also gives the distance/intensity trading
that Mark Davis called for in the Soundfield One design. We chose to
attenuate the inner front by 3 dB and the outer front by 9 dB, so that as
you go from center to either side, you get more in the loudness from the
opposite channel and less from the one you are near. This compensates for
how close you are to the one speaker and makes the center imaging remain
stable.

So we took a guess as to how to make this pattern work and we haven't done a
careful radpat measurement yet to see if we are getting it very precisely,
but audibly it works quite well, because we are getting that "out of body"
projection of the sound outside the speakers themselves by means of
reflection of the rear wave, and also even imaging as you go around in the
room due to my required speaker placement and the frontal radpat of the
speakers. This is a combination of the design features of the Bose 901 and
the DBX Soundfield One, and an improvement on both. I also listen in
surround sound for even better spatial results.

Gary Eickmeier

[Peter Wieck]

Gary:

You mention the Bose and DBX speakers as your springpoint - I will not insu=
lt you by suggesting they were your inspiration. It has been my experience=
with a number of speakers of (mostly) conventional design that what arrive=
s at one's ears is perhaps, and at best, a third-cousin to the live experie=
nce, and for any number of obvious reasons. fixed-point microphones for the=
most part do not sit where one's ears sit, nor do any given pair (with a f=
ew deliberate exceptions) sit so either.=20

Nor does the typical conventional speaker duplicate the original sound prod=
ucer. A piano string may be from a few inches to several feet long, be made=
of drawn, hardened steel and often be wrapped with hardened brass wire. Th=
is is hardly duplicated by a speaker coil which, at best, might be in the 1=
2" - 14" (diameter) range. Further to this, each producer has its unique ch=
aracteristic, all of which is co-mingled by the reproducer.=20

Nor does the typical speaker duplicated the subtle, for lack of a better te=
rm, "elliptical" nature of live sound. As an analogy, consider even a small=
orchestra of perhaps 10 instruments. One sits in an audience dead-center (=
such as the 'ideal' (term used very loosely) listening room) to the orchest=
ra and one is just as far from the base viol as from the brass. Put a virtu=
al thumbtack on each of those instruments, and extend a string to each of y=
our years. Draw it across the back of each tack so it is a complete loop fr=
om each ear to the instruments, then across the orchestra. If one now moves=
off dead-center, the string will slide and your ears will describe an elli=
pse. Hence the term. And, the basis of the much-misunderstood term "soundst=
age".=20

A properly designed performance stage will not be subject to standing/inter=
ference waves as might a poorly designed stage - and back in the day (datin=
g myself here) I was exposed as a grade-school student to the teething prob=
lems of Avery Fisher Hall at Lincoln Center - all kinds of interference and=
standing waves - and where I first learned of the term and began to have =
the inklings of appreciating good sound. This cannot be said of the typical=
rectangular listening venue. Which _WILL_ have such issues even if identif=
ied and minimized by good placement, design and other means.=20

This is a long way of stating that I am fascinated by your design and how y=
our longer-term impression of their performance has evolved. A few years ag=
o, when we moved into our "new" (built in 1890) house, my wife graciously p=
ermitted me to 'set up' in the largest room the library, 16 x 28 x 10. Wall=
s are plaster, floor is hardwood, four french doors, lots of glass. The two=
short walls are mostly shelves, there is a fireplace. The furniture is sof=
t and there are Afghan carpets on the floor. In any case, she also permitte=
d me to install Maggies - the MG-IIIa as it happens, rebuilt to as-new spec=
ifications and with new ribbon tweeters. Pretty amazing beasts. And the clo=
sest thing to a live experience I have been able to produce outside of a pa=
ir of first-issue AR9s that left me when we moved into our previous house w=
ith no possible venue for them. Most of the appeal with the Maggies is that=
I may sit in any number of places (imagine that ellipse) with about the sa=
me sense-of-place one gets in a moderately sized concert venue moving about=
the 'house'. The volume equals the virtual 'string length' if you will, so=
that sense of place/movement about the stage is much greater at lower volu=
mes than higher. I see that you have set up a 'sweet spot', something that =
Bose (at least) advertised early on as being unnecessary with the 901 desig=
n. How does your set-up respond to more than one listener, or perhaps even =
a few of them not sitting on the ideal arc? That, to me, is the single larg=
est problem with so-called 'high-end' installations. They are designed arou=
nd a single head at a single position at a given distance and height. Not v=
ery user-friendly and in my *opinion* far too high a price to pay for a per=
haps dubious result. Although you do seem to have such an installation in =
place, perhaps your speaker design is to increase your options? Might that =
single chair evolve into a couch?=20

Thanks in advance!

Peter Wieck
Melrose Park, PA=20

p.s.: Other speaker systems in place and operating a=20

AR3a
Revox Picollo
AR M5
AR Athena

[Gary Eickmeier]

Peter Wieck wrote:
Gary:

You mention the Bose and DBX speakers as your springpoint - I will
not insult you by suggesting they were your inspiration. It has been
my experience with a number of speakers of (mostly) conventional
design that what arrives at one's ears is perhaps, and at best, a
third-cousin to the live experience, and for any number of obvious
reasons. fixed-point microphones for the most part do not sit where
one's ears sit, nor do any given pair (with a few deliberate
exceptions) sit so either.

Nor does the typical conventional speaker duplicate the original
sound producer. A piano string may be from a few inches to several
feet long, be made of drawn, hardened steel and often be wrapped with
hardened brass wire. This is hardly duplicated by a speaker coil
which, at best, might be in the 12" - 14" (diameter) range. Further
to this, each producer has its unique characteristic, all of which is
co-mingled by the reproducer.

Nor does the typical speaker duplicated the subtle, for lack of a
better term, "elliptical" nature of live sound. As an analogy,
consider even a small orchestra of perhaps 10 instruments. One sits
in an audience dead-center (such as the 'ideal' (term used very
loosely) listening room) to the orchestra and one is just as far from
the base viol as from the brass. Put a virtual thumbtack on each of
those instruments, and extend a string to each of your years. Draw it
across the back of each tack so it is a complete loop from each ear
to the instruments, then across the orchestra. If one now moves off
dead-center, the string will slide and your ears will describe an
ellipse. Hence the term. And, the basis of the much-misunderstood
term "soundstage".

A properly designed performance stage will not be subject to
standing/interference waves as might a poorly designed stage - and
back in the day (dating myself here) I was exposed as a grade-school
student to the teething problems of Avery Fisher Hall at Lincoln
Center - all kinds of interference and standing waves - and where I
first learned of the term and began to have the inklings of
appreciating good sound. This cannot be said of the typical
rectangular listening venue. Which _WILL_ have such issues even if
identified and minimized by good placement, design and other means.

This is a long way of stating that I am fascinated by your design and
how your longer-term impression of their performance has evolved. A
few years ago, when we moved into our "new" (built in 1890) house, my
wife graciously permitted me to 'set up' in the largest room the
library, 16 x 28 x 10. Walls are plaster, floor is hardwood, four
french doors, lots of glass. The two short walls are mostly shelves,
there is a fireplace. The furniture is soft and there are Afghan
carpets on the floor. In any case, she also permitted me to install
Maggies - the MG-IIIa as it happens, rebuilt to as-new specifications
and with new ribbon tweeters. Pretty amazing beasts. And the closest
thing to a live experience I have been able to produce outside of a
pair of first-issue AR9s that left me when we moved into our previous
house with no possible venue for them. Most of the appeal with the
Maggies is that I may sit in any number of places (imagine that
ellipse) with about the same sense-of-place one gets in a moderately
sized concert venue moving about the 'house'. The volume equals the
virtual 'string length' if you will, so that sense of place/movement
about the stage is much greater at lower volumes than higher. I see
that you have set up a 'sweet spot', something that Bose (at least)
advertised early on as being unnecessary with the 901 design. How
does your set-up respond to more than one listener, or perhaps even a
few of them not sitting on the ideal arc? That, to me, is the single
largest problem with so-called 'high-end' installations. They are
designed around a single head at a single position at a given
distance and height. Not very user-friendly and in my *opinion* far
too high a price to pay for a perhaps dubious result. Although you
do seem to have such an installation in place, perhaps your speaker
design is to increase your options? Might that single chair evolve
into a couch?

Thanks in advance!

Hi Peter -

A most interesting post. I can't answer it all at once, here, but permit me
to give the basic idea or ideas.

First and foremost, I wrote a paper in 1989 (An Image Model Theory for
Stereophonic Sound) in which I tried to relate my discovery that stereo
doesn't work like most of us seem to think, by the lateral localization
alone setting up an "image" by fooling your hearing mechanism into hearing
"through" a large portal into another acoustic space. In my paper I proposed
that stereo is rather a "model" rather than a picture, a model that is
reconstructed within your listening room. This model is composed of the same
parts as the live model, namely a number of real (direct) sound sources
placed upon a model soundstage, surrounded by several virtual images of
those sources as reflections from the other side of the nearby walls. This
is nothing unknown to architectural acoustics, but describing stereo as
working like that is rather different and revolutionary in hi fi circles.
Bose proposed the same effect, but explained it as an "enhancement" more
than a whole new theory of sound, which called for a new approach to our
thinking about stereo theory. The DBX Soundfield One was a refinement on the
direct field portion of the model, which proposed a certain radiation
pattern that could stabilize the central imaging with what is now called
distance/ intensity trading which makes the farther speaker louder to
balance out the nearer speaker's closeness and move the image back to
center.

The whole theory requires a certain speaker positioning scheme (1/4 of the
room width in from the side walls and out from the front wall) in order to
keep the reflected images from clustering together in an uneven manner to
destroy the lateral or depth imaging. If you draw an image model of that
positioning scheme you can see what I mean. My room for example - it is 20
ft wide, so I place my speakers 5 ft in from side walls and out from front
wall and that makes all real and virtual image speakers exactly the same 10
ft apart from each other.

My room surfaces near the front are specularly reflective (another
difference from standard thinking) but with more absorption and diffusion as
you go back into the room. The direct to reflected ratio is adjusted to more
in the reflected domain, until the imaging "pops" outside the speakers
themselves and they disappear, leaving only the soundscape which extends
from wall to wall, across the entire front of my room, rather than just
speaker to speaker. NO sounds are EVER heard to be coming from the speaker
boxes themselves, an artifact which destroys the suspension of disbelief for
me. The worst speakers suffer the imaging of the extreme L or R instruments
collapsing to the speaker grills.

If you think about it, your Maggies have most of these same properties if
you position them correctly and aim them to cross in front of you, so that
if you move left of center the right speaker gets louder, and so that the
backwave is sent deep and wide, returning to you from front and side walls
as a massive secondary reflection from the corners, delayed more than 10 mS.
When the delayed sounds are done by reflection like this, the perspective on
the instruments changes as you walk around the room, just like a mirror
image would change visually.

This is pretty much my "ultimate" design, taking in all of the factors that
are audible about speaker reproduction. These are radiation pattern and
frequency response, positioning in the room, and power, or dynamics
achievable. There is nothing more you can manipulate about speaker sound in
rooms*. Not all speakers up to this point have done that, and no other
speaker in the world has this particular radiation pattern, nor does anyone
else say much about speaker positioning except as relates to frequency
response, which is not the basis of speaker positioning but rather imaging
should be the only criterion. I also use surround sound in my playback,
again under the principle of reconstructing all of the sound fields within
my room to mimic the live fields as much as feasible.

So yes, my longer term impression is still holding, but I still need to get
some people in here to give me their impressions and to get some other
speakers to compare them with in an AB fashion.

* The only other factor is room size, a larger room sounding more like the
original because tghe model is more like the size of the original.

Gary Eickmeier

[Peter Wieck]

The whole theory requires a certain speaker positioning scheme (1/4 of the=
=20
room width in from the side walls and out from the front wall) in order to=
=20
keep the reflected images from clustering together in an uneven manner to=
=20
destroy the lateral or depth imaging. If you draw an image model of that=20
positioning scheme you can see what I mean. My room for example - it is 20=
=20
ft wide, so I place my speakers 5 ft in from side walls and out from front=
=20
wall and that makes all real and virtual image speakers exactly the same 10=
=20
ft apart from each other.=20

__________________________________________________ __________________

Interesting. Back in the day when AR had their listening room at Grand Cent=
ral Station, they handed out literature by the ton, and spec. sheets by the=
freighter-load. Remember one could not purchase anything there. One thing=
they handed out was a white-paper on speaker placement with all sorts of t=
ips and tricks based on many often-not-considered stuff. And it should also=
be understood that AR put a center-channel output on their receivers as ea=
rly stereo recordings often greatly exaggerated separation. This is somethi=
ng of a nested issue - center channel/stereo, as even in the mid-late 1960s=
, the concept of 'stereo vs. mono' was as much a hot issue as the concept o=
f CD vs. Vinyl is today.=20

In any case, and highly distilled, AR posited that speakers should be place=
on the long wall of any given room, with one speaker being somewhere betwe=
en the 1/4 and 1/3 points, the other placed asymmetrically to it about half=
-way between the first speaker and the far corner. Then, the speakers were =
to be moved closer together until the best sound-stage was achieved - but s=
till kept asymmetrical to the corners. After which, at that point, they wer=
e to be moved as a fixed pair closer to or further from the nearer corner u=
ntil the best (clearest) bass was achieved. Oh, and the woofer centers were=
to be at least two woofer diameters from the floor. The backs were to be a=
gainst the wall. At no time were the speakers to be symmetrical about the l=
ong wall.=20

I have done that with my 3as (also driven by an AR receiver), and I will st=
ate for the record that even moving the pair (or either one) by a few inche=
s is easily perceived. We have a three-person couch facing the speakers abo=
ut 10 feet away, and the speakers are about 8 feet apart on a 17 foot wall.=
(room is 17 x 14 x 10). This is the "wife-friendly" stereo. All self-evide=
nt controls and very few of them. But the sound at either end of the couch =
or either flanking wing-back is quite remarkable. The two wing-backs and th=
e couch describe the same virtual ellipse I noted previously.=20

So, I am quite sympathetic to how sensitive any system might be to speaker =
placement, and most especially to reflections and refractions as created by=
room acoustics, materials and even canvas vs. glass-front paintings on the=
wall. Controlling all the variables is the key, and you have apparently in=
troduced many more potential variables with, I am guessing the potential fo=
r much better results if those variables are controlled adequately. What ha=
ppens if those speakers are in a more difficult location? Do they have basi=
c requirements such as minimum distances from a wall? Must they be symmetri=
cally placed or is there sufficient adjustment that they need not be?=20

Again, I see transducers in general and speakers specifically as about the =
last frontier of audio. Amplification is pretty much 'done' since the 60s f=
or tubes and the 80s (a stretch, probably the 70s) for solid-state. Vinyl, =
Tape, CD, whatever are also pretty much done. Speakers and cartridges. Spea=
kers everyone needs. So, if you do have a cutting edge design that also suc=
ceeds in its intent, that would be truly remarkable.=20

Peter Wieck
Melrose Park, PA

[Gary Eickmeier]

Peter Wieck wrote:

Interesting. Back in the day when AR had their listening room at
Grand Central Station, they handed out literature by the ton, and
spec. sheets by the freighter-load. Remember one could not purchase
anything there. One thing they handed out was a white-paper on
speaker placement with all sorts of tips and tricks based on many
often-not-considered stuff. And it should also be understood that AR
put a center-channel output on their receivers as early stereo
recordings often greatly exaggerated separation. This is something of
a nested issue - center channel/stereo, as even in the mid-late
1960s, the concept of 'stereo vs. mono' was as much a hot issue as
the concept of CD vs. Vinyl is today.

In any case, and highly distilled, AR posited that speakers should be
place on the long wall of any given room, with one speaker being
somewhere between the 1/4 and 1/3 points, the other placed
asymmetrically to it about half-way between the first speaker and the
far corner. Then, the speakers were to be moved closer together until
the best sound-stage was achieved - but still kept asymmetrical to
the corners. After which, at that point, they were to be moved as a
fixed pair closer to or further from the nearer corner until the best
(clearest) bass was achieved. Oh, and the woofer centers were to be
at least two woofer diameters from the floor. The backs were to be
against the wall. At no time were the speakers to be symmetrical
about the long wall.

I have done that with my 3as (also driven by an AR receiver), and I
will state for the record that even moving the pair (or either one)
by a few inches is easily perceived. We have a three-person couch
facing the speakers about 10 feet away, and the speakers are about 8
feet apart on a 17 foot wall. (room is 17 x 14 x 10). This is the
"wife-friendly" stereo. All self-evident controls and very few of
them. But the sound at either end of the couch or either flanking
wing-back is quite remarkable. The two wing-backs and the couch
describe the same virtual ellipse I noted previously.

So, I am quite sympathetic to how sensitive any system might be to
speaker placement, and most especially to reflections and refractions
as created by room acoustics, materials and even canvas vs.
glass-front paintings on the wall. Controlling all the variables is
the key, and you have apparently introduced many more potential
variables with, I am guessing the potential for much better results
if those variables are controlled adequately. What happens if those
speakers are in a more difficult location? Do they have basic
requirements such as minimum distances from a wall? Must they be
symmetrically placed or is there sufficient adjustment that they need
not be?

Again, I see transducers in general and speakers specifically as
about the last frontier of audio. Amplification is pretty much 'done'
since the 60s for tubes and the 80s (a stretch, probably the 70s) for
solid-state. Vinyl, Tape, CD, whatever are also pretty much done.
Speakers and cartridges. Speakers everyone needs. So, if you do have
a cutting edge design that also succeeds in its intent, that would be
truly remarkable.

Interesting post Peter. On the AR speaker positioning recommendations, as
usual they were doing it for frequency response reasons, and not imaging. No
authority today would recommend an assymetrical positioning scheme.

I discovered my whole stereo theory by radically repositioning my Bose 901s
while stationed in England. There, they have plaster walls that are highly
reflective, and of course the 901 is a reflecting speaker, so you notice
changes in position more easily in this situation. Their instruction book
just says at least a foot and a half from all walls, so that is where I had
them, for disastrous results. But being in England, I noticed they were
pulling their speakers well out from the walls in an arrangement something
like I have described, so I said to myself why should their speakers call
for different placement than mine or any other design? It's only sound and
acoustics, so let's try it, and the result was a minor miracle. So I
dedicated my life from that point on to explaining what had just happened
and how the reflected sound should be a part of stereo theory.

I don't think my design would work as well in a non-dedicated listening
room, but I am not all that interested in less than ideal conditions. My
goal is to establish how an ultimate system functions, then people can
deviate from that as they please, but will know what causes what when they
do that.

Stereo is a fascinating problem. Practitioners have gravitated to the best
recording techniques over the years in spite of such a varied landscape of
playback techniques that all sound different from each other. If you read
the textbooks on stereo theory or speakers and rooms you will see the
infamous 60 degree triangle with two speakers and a listener, with no room
or walls included in the diagram. My Image Model Theory, or IMT,
incorporates the walls and room as part of the explanation for the first
time, and describes how to use them. I believe that the image model drawing
of any acoustical situation explains far more of what we are hearing in a
given situation, and gives direction for improvement in modeling the
reproduction after the live model.

Gary Eickmeier

[Peter Wieck]

On Friday, October 3, 2014 6:53:20 AM UTC-4, Gary Eickmeier wrote:

Interesting post Peter. On the AR speaker positioning recommendations, as=
=20
=20
usual they were doing it for frequency response reasons, and not imaging.=
No=20
=20
authority today would recommend an asymmetrical positioning scheme.
=20
Snipped=20

Gary Eickmeier

I think you missed the part that stated the speakers were to be moved relat=
ive to each other (mostly closer together) until the best sound-stage was r=
ealized - and only then moved as a fixed pair until the best bass was achie=
ved. I also think that AR was the Vilchur and Allison were the first engin=
eers to address bass head-on. Then, as now, bass was expensive, and their d=
esigns were quite expensive relative to many of their competitors. Large ma=
gnets, long-throw voice-coils, specific cabinet design requiring careful as=
sembly and more. And, as an unhappy offshoot, heavy power requirements.=20

And, for the first time, a legitimate 25-30Hz speaker was on the market wit=
hin the reach of a moderately heeled consumer.=20

Venturing into opinion he I am not much enamored with Authorities -=20
especially as it applies to speakers. Most have drunk the cool-aid of small=
=20
drivers (which, bluntly, don't move much air), sweet-spots and other popula=
r=20
persuasions that have some systems set up in such a way as there *can* be o=
nly one listener whose ears must be within a few cm/inches of a specific lo=
cation in all three axes. I have no argument with the fact that some such s=
ystems can sound ethereally beautiful, if a little thin. But they are quite=
selfish.=20

To me, a system should be able to function at several levels, the first bei=
ng able to produce excellent sound at live or near-live volumes. As in not =
necessarily Rock Concert ear-bleed levels (although that is surprisingly ea=
sy to produce), but at mid-row seating full orchestra levels (which is=20
surprisingly hard to produce).=20

The second being able to produce pleasant background noise that will permit=
a friendly conversation without shouting, yet still be specifically pleasi=
ng to the ear. Many speakers are incapable of producing clear noise at low =
volume.=20
=20
The last, being able to do this with some pretty articulate music - and my=
=20
test pieces to this are the Saint-Saens Organ Symphony (Phila. Orchestra=20
recorded DD at St. Francis DeSales church in West Philadelphia), Kiri Te Ka=
nawa doing Exultate Jubilate (Mozart), and any of several Purcell trumpet p=
ieces. For the 'other' music, I use a few recent recordings of Emmylou Harr=
is - somehow the female voice tests speakers like nothing else.=20

The maggies are the closest thing I have to an ideal speaker. Not quite an =
AR9 at the low end, but extremely clean and precise at the high end. I do n=
ot feel the need for a sub-woofer, they do the Bombard Pipes quite well. Th=
e 3as are very nearly as good but in a much smaller room. They would be los=
t in the place of the maggies. Filling ~4,600 cubic feet is not an easy tas=
k. Doing it without slinging a lot of mud is much harder.=20

So, I remain fascinated by the potential in _any_ new speaker design. You m=
ay be selling your design short by dismissing 'less than ideal' conditions=
=20
out-of-hand, and I think you should also question authority and do some gen=
uine experimentation under those less than ideal situations. I have found i=
n over 40 years around this hobby that well-designed speakers are remarkabl=
y forgiving of bad situations, whereas marginal designs fall flat if outsid=
e their comfort zone. This remains a hobby of discovery for me - I switch s=
tuff around and ring the changes on a more-or-less weekly basis - there are=
five (5) active systems shared between this house and the summer house wit=
h the makings of several others going in and out on a regular basis.=20

Peter Wieck
Melrose Park, PA

[Gary Eickmeier]

Peter Wieck wrote:

Hello again Peter -

Without re-quoting your entire post -


So, I remain fascinated by the potential in _any_ new speaker design.
You may be selling your design short by dismissing 'less than ideal'
conditions
out-of-hand, and I think you should also question authority and do
some genuine experimentation under those less than ideal situations.
I have found in over 40 years around this hobby that well-designed
speakers are remarkably forgiving of bad situations, whereas marginal
designs fall flat if outside their comfort zone. This remains a hobby
of discovery for me - I switch stuff around and ring the changes on a
more-or-less weekly basis - there are five (5) active systems shared
between this house and the summer house with the makings of several
others going in and out on a regular basis.

Peter Wieck
Melrose Park, PA

Yes, wow - question authority - the story of my audio life. There are some
who agree with most of what I have said about stereo, but more of them
disagree and want the direct sound only theory. Your arguments against the
sweet spot type of speaker are spot on - so to speak. We have this problem
simply because of a side issue of using just two speakers for stereo in the
beginning, but that is not an "essential" part of stereo theory.

I have used the analogy of using a single microphone and speaker for each
instrument in an ideal system. You mike them relatively closely, then play
each channel on speakers with a radiation pattern that is similar to the
instrument it is reproducing, and all placed in positions that are
geometrically similar to where the instrumnts were, in a performing space
that has good acoustics, not a Sonex plastered goofball audio room. You
would have ultimate realism because the system IS REAL and is playing the
performance with the same sound patterns in the room as there were
originally, there is no sweet spot, you can walk around in the room and get
different perspectives on the players, everything with ultimate realism.
Something like a player piano playing back the exact same notes and strokes
as the original on the same instrument in the same room.

That is the root basic idea of a field-type system such as stereophonic
(including surround sound). Two speaker stereo is just a simplification of
such a system, pared down to as few channels as you can get away with and
still have auditory perspective. BUT THEN, the fact that we have only two
ears causes this great confusion about the fundamental nature of the system,
that it is a two ears - two speakers system, with the sound from each
channel shooting into your ears and fooling you into hearing stereo. NO -
the summing localization that is at work is a wonderful effect, enabling us
to use just the two speakers, but it is not a fundamental reason, or
explanation, of how stereo should be done.

With that realization, you jump back to reality and legacy two channel
recordings and ask yourself, OK so how should we be doing this speaker and
room stuff? The answer is that we need to consider the spatial nature of
live sound and try to mimic that in the playback. The speakers, then, rather
than seen as simple direct radiators become Image Model Projectors, using a
combination of direct and reflected sound to set up sound patterns that are
as much like the live patterns as feasible with just a few channels.

Finally, there is no acoustical free lunch. If and when you realize that you
shouldn't do realistic stereo with two small monitors in a small dead room,
hopefully you can see that the size of the playback room is just as much a
part of the equation as any other factor, a larger room sounding more real
because the size of the model is more like the real thing.

So to summarize, the "sweet spot" problem is a red herring and not a
necessary part of stereo theory, just a side effect of doing it with so few
speakers. You can play music in a non-ideal room or in a car or on a
boombox, but then you don't expect ultimate realism, just pleasant sound and
some aspects of fidelity such as frequency response or loudness.

The world of audio "authority" thinks that the process is one of "accuracy"
which is a great sounding word to an engineer, but is not what we are doing
with a field-type system. They think that if the microphones capture the
original with perfect frequency and phase response etc and the speakers
shoot that same "signal" right back into your ears, you will be done, and
the history of audio research has been a search for greater and greater
accuracy. But as you have said, we are already there with every component
but the speakers. The answer to this seeming quandary is that we are doing
realism, not accuracy. We must study sound fields and the spatial nature of
sound before we can realize the true goal we are all after.

When I see some new miracle speaker from some high end maker that has all of
the drivers on one side of the box, I pull my hair out, which is why I am so
bald. If you ever get down to Florida on vacation or whatever, write an
Email back for directions or look me up in Lakeland.

If you would like to continure this, I love to talk audio and we could take
it up in Email or on the phone. Nobody else has lurked on in here because
most of them have heard my ideas before.

Gary Eickmeier

863-670-0850