What Can We Hear?

[Gary Eickmeier]

There are lots and lots of discussions and descriptions of what we can hear
in the subjective press and in these halls of RAHE. Some of these
descriptions go into detail that is sometimes real, sometimes imagined,
leading to wild arguments about double blind listening tests vs subjectivism
and long experience. I think most of it misses the main points of genuine
"hard-nosed" listening and being honest with yourself about what you are
really hearing with your system. I mean, like, you want to think that your
speakers or whatever are better than they are, or that your ears and tastes
are so sophisticated that you can hear all of these marvelous aspects of
recorded sound - and you hope that the rest of us will believe you, that
either your hearing is so much better than ours, or your components are
"revealing" so much more due to their greater "accuracy."

So let's cut the bull**** and ask ourselves what CAN we really hear about
the original live sound and the reproduction, and thus what kind of
correlation can we draw between them to possibly see how far we have come,
or how far we can go. I struggle to come up with a catchy name for these
characteristics, but for now let's all them the Essential Elements of
Fidelity, or EEF.

THE ESSENTIAL ELEMENTS OF FIDELITY

OK, everybody stand up and shake your hands, wiggle you ears and noses,
whatever it takes to shake off all preconceptions of "inner detail," "phase
anomolies," "transient response," "togetherness," "toe tapping ability (my
personal favorite)," or any of the many other supposedly audible
characteristics of reproduced sound and let's start over again.

1. PHYSICAL SIZE - whether you're talking about the real thing or the
reproduction, we can hear the size of a room we are in. This is due to the
time between reflections, the characteristics of the reverberant tail, and
maybe some discrete echoes, which good spaces shouldn't have, but in any
case we can tell whether we are in a big or a small room. In the
reprocuction, one of the main problems is that the acoustics of the repro
room are superimposed on those of the recording, and we can sense that
fairly easily. This means, the larger you can design your room, the more
realistic it will sound, because it will be more like the real thing. It
also does away with some other pesky acoustical problems at the same time,
but right now I just want to point our that physical size is audible.

2. POWER - of course we can hear the enormous power of a symphony orchestra
or a big band, or even a smaller group. This means that the more power we
have in the reproduction, the closer to realism we will get. You can have
great fidelity in a boombox or a desktop computer speaker system, but it
will not have the POWER of the real thing unless and until you get some
speakers that can take any amount of power you can give them and get louder
without distorting, and amplifiers that have that power rather than the
audiophile fave raves of dainty 20W tube amps. Power is definitely audible.

3. WAVEFORM FIDELITY - I have been taken to task for calling it that,
because the actual shape of a waveform is not the point, but I can't think
of a term to describe what I mean by just simply the accuracy in the
electronic domain of the recorded signal transmission. This includes, of
course, frequency response, noise, and distortion. We struggled with these
for a long time in our audio history with LP records and magnetic tape. But
now with digital, we have essentially eliminated this characteristic from
being a problem in recording or reproduction. Still, it is one of the
factors that we can hear, so I list it for completeness.

4. SPATIAL CHARACTERISTICS - This is the biggie, the collective term for
the realistic reproduction of auditory perspective - the stereo effect and
all of the possible recording and reproduction systems and schemes. But for
now, all we need to point out is that we can hear the spatial
characteristics of live and reproduced sound, and those characteristics are
very important to both the enjoyment of live sound (and quality of a concert
hall) and the realism of the reproduction.

Now, not to impugn your intelligence or vast knowledge, if most of us think
we understand what this spatial stuff is all about, but to emphasize the
difference between the spatial and the temporal, the two aspects of a sound
field that get continually confused with each other, I would like to add an
illustration.

Your friend is a novice audiophile on his way back from Best Buy, where he
has purchased a new home theater system and learned all about hi fi from the
salesman. He sets up the speakers all up front, perhaps all in a row or all
on top of his "teevee" and he plays a movie or some music. It is very
accurate, plays all of the frequencies and the timings of the ambience that
were recorded, but it just doesn't sound realistic yet. So you go over and
show him how to correctly place all of the speakers to reconstruct a
semblance of the sound field that was recorded in his listening room. You
place the front speakers LCR for correct perspective of the frontal
soundstage, and you place the surround speakers back and to the sides, where
the ambience of the hall should come from. You have addressed the SPATIAL
characteristic, which has nothing to do with the temporal, but rather with
the directions form which the various sound fields arrive at the listener,
or exist in the listening room. I bring this up because of a frequent
question about my statements on getting the spatial more correct. They
always tell me that getting the spatial more correct can't work because you
can't make a small room sound like a concert hall. They have confused, or
"fused" the spatial and the temporal.

So I try to explain the difference but for now I only want to state that
these are the main characteristics of sound that we can hear and try to
reproduce. Do you have any other "biggies" that I have left out? I would be
fascinated.

So what is the state of the art of attempts to reproduce all of these and
how far can we go?

The physical size and waveform accuracy and power we can easily get a handle
on and improve, if for example we are in a small room and we understand that
limitation. The spatial stuff is the biggie and is where I say we need more
basic understanding of the process in order to get any further than we have
already come in 100 or so years.

The most basic and foolish mislead is thinking that good "stereo" comes from
the direct sound alone, and trying to kill the room reflections or design a
speaker that casts all of its sound toward your hapless ears. This
misconception, or mislead, is caused by the confusion between stereo and
binaural.

Stereo does not work like a "window into another acoustic." Rather, if you
think of it as a model of the original, in which your room is the performing
space and your speaker setup attempts to get the spatial closer to the
original, then you have a fighting chance for greater realism, but you also
inherit the understanding that it is not an "accuracy" process, and we can
never get all the way there. We cannot, in other words, totally get to the
goal of a "you are there" experience but rather more like a "they are here"
experience in which your room is the performing space and you design it for
good sound and arrange THE BIG THREE of speaker positioning, radiation
pattern, and room acoustics to get the model closer to the live situation.

So what can we hear? We can hear the spatial, spectral, and temporal
characteristics of our listening room and speaker situation, or layout,
superimposed on that of the recording, and we can hear the physical size,
power, and electronic accuracy of your system. When we play back any
recording, we CHANGE the spatial characteristics of the original to those of
our playback system and room.

That is slightly too bad, but once we understand the limitations of the
system and what can be achieved, we can stop worrying about false goals and
start concentrating on more fruitful paths that can lead to greater realism.

Gary Eickmeier

[Audio Empire]

On Mon, 14 May 2012 16:44:59 -0700, Gary Eickmeier wrote
(in article >):

> There are lots and lots of discussions and descriptions of what we can hear
> in the subjective press and in these halls of RAHE. Some of these
> descriptions go into detail that is sometimes real, sometimes imagined,
> leading to wild arguments about double blind listening tests vs subjectivism
> and long experience. I think most of it misses the main points of genuine
> "hard-nosed" listening and being honest with yourself about what you are
> really hearing with your system. I mean, like, you want to think that your
> speakers or whatever are better than they are, or that your ears and tastes
> are so sophisticated that you can hear all of these marvelous aspects of
> recorded sound - and you hope that the rest of us will believe you, that
> either your hearing is so much better than ours, or your components are
> "revealing" so much more due to their greater "accuracy."
>
> So let's cut the bull**** and ask ourselves what CAN we really hear about
> the original live sound and the reproduction, and thus what kind of
> correlation can we draw between them to possibly see how far we have come,
> or how far we can go. I struggle to come up with a catchy name for these
> characteristics, but for now let's all them the Essential Elements of
> Fidelity, or EEF.

The human ear is a sensitive tool, no doubt about it. Unfortunately, it is
also an interpretive tool. There is no sound preception without the brain and
the brain brings with it a lifetime of experience, preconceptions, and
personal preferences. It applies these factors to everything we hear and
they're difficult to overcome. In fact, I'd go so far as to say that we, as
individuals, can't overcome them. We are going to hear what we want to hear
or what we expect to hear. The only way around this is to take human bias out
of the equation with tests that are either totally objective (such as
measurements using instrumentation) or by taking part in listening tests
which remove as much of the human propensity for self-delusion as possible
and relying on the statistical results. To paraphrase an old adage: he who
trusts his own ear/brain interface to judge audio qualities is a fool.

> THE ESSENTIAL ELEMENTS OF FIDELITY
>
> OK, everybody stand up and shake your hands, wiggle you ears and noses,
> whatever it takes to shake off all preconceptions of "inner detail," "phase
> anomolies," "transient response," "togetherness," "toe tapping ability (my
> personal favorite)," or any of the many other supposedly audible
> characteristics of reproduced sound and let's start over again.

Phase anomalies are real. You can measure them and you can easily hear their
effects. All one has to do is walk past a stereo pair of speakers that are
wired out-of phase to instantly hear the results. You can also take a
recording made with a pair of spaced omnis and sum the two channels to mono
to immediately hear (and see, on an oscilloscope) some of the instruments go
away.

> 1. PHYSICAL SIZE - whether you're talking about the real thing or the
> reproduction, we can hear the size of a room we are in. This is due to the
> time between reflections, the characteristics of the reverberant tail, and
> maybe some discrete echoes, which good spaces shouldn't have, but in any
> case we can tell whether we are in a big or a small room. In the
> reprocuction, one of the main problems is that the acoustics of the repro
> room are superimposed on those of the recording, and we can sense that
> fairly easily. This means, the larger you can design your room, the more
> realistic it will sound, because it will be more like the real thing. It
> also does away with some other pesky acoustical problems at the same time,
> but right now I just want to point our that physical size is audible.

I have to disagree here. While you are correct about being able to tell, even
blindfolded, the approximate size of a room you're in, you are wrong about
the results. It is possible to make a small room sound good much more easily
than it is possible to make a large room sound good. Small rooms have all the
characteristics that you mention (and then some) but they are all fairly
controllable.

> 2. POWER - of course we can hear the enormous power of a symphony orchestra
> or a big band, or even a smaller group. This means that the more power we
> have in the reproduction, the closer to realism we will get. You can have
> great fidelity in a boombox or a desktop computer speaker system, but it
> will not have the POWER of the real thing unless and until you get some
> speakers that can take any amount of power you can give them and get louder
> without distorting, and amplifiers that have that power rather than the
> audiophile fave raves of dainty 20W tube amps. Power is definitely audible.

Again you are grossly oversimplifying. Power, in and of itself is merely a
means to an end. The end is moving air. The more air you can move the more
realistically an audio system will load the room with sound and the amount of
air any speaker can move with a Watt of electrical power from an amplifier is
determined by two things: the efficiency of the speaker and the volume of the
space one is trying to fill. The most realistic reproduction of the POWER of
a symphony orchestra I ever heard was from a pair of the big Klipschorn
corner horns in a smallish basement listening room in the early 1960's. The
system was owned by a high-school buddy of mine's dad. The effect was jaw
dropping. Here was a small pair of Heathkit Williamson power amps at about 25
Watts each that would drive these speakers to pump-out so much air, that,
like sitting front row center in a concert hall, while the NY Philharmonic
plays the climax or Ravel's "Bolero" , it literally made one's pants legs
flap with the music!. These speakers were so efficient, that you could play
them loud enough to have to scream to be heard over them with just the
earphone output of a garden-variety Japanese transistor radio, of the type
every teen carried around with him in those days. The Klipshorns were 50%
efficient. That meant that every two Watts of amplifier power produced one
acoustic Watt of sound (One acoustic Watt is defined as being equivalent to
107.5 dBSPL at around a meter from an omnidirectional source)! Of course,
most speakers are nowhere near that efficient and it can often take more than
100 electrical Watts to produce one Acoustic Watt at one meter. That would
make those speakers 1% efficient. Now, since sound pressure dissipates at a
rate inversely proportional to the distance one is from the speakers,
obviously, the larger the room, the more power it takes to maintain a
realistic sound power level in that room. After all, few people sit one meter
of less from their speakers. The point here is it's not the power itself
that's important, it's the amount of power the SPEAKERS need to reproduce the
SPL necessary to achieve the desired room loading. Some speakers can do it
with a 10 Watt SET amp and some require many hundreds of Watts for the same
effect. Now, understand, that we are talking about ONE and only one parameter
here. That's acoustical energy or volume. We aren't talking about frequency
response or imaging or any other sound characteristic. Fact is that while
Klipschorns were certainly realistically loud, they were not particularly
great sounding. They didn't have much bass below 50 Hz, and they certainly
weren't flat over the rest of the sdpectrum. I wouldn't want a pair in my
system, but at the time, when a 10 Watt amplifier was the norm, and a 25 Watt
amp was a behemoth, they produced the SPL that many audiophiles were looking
for.

> 3. WAVEFORM FIDELITY - I have been taken to task for calling it that,
> because the actual shape of a waveform is not the point, but I can't think
> of a term to describe what I mean by just simply the accuracy in the
> electronic domain of the recorded signal transmission. This includes, of
> course, frequency response, noise, and distortion. We struggled with these
> for a long time in our audio history with LP records and magnetic tape. But
> now with digital, we have essentially eliminated this characteristic from
> being a problem in recording or reproduction. Still, it is one of the
> factors that we can hear, so I list it for completeness.

Try "source signal accuracy." 8^)
>
> 4. SPATIAL CHARACTERISTICS - This is the biggie, the collective term for
> the realistic reproduction of auditory perspective - the stereo effect and
> all of the possible recording and reproduction systems and schemes. But for
> now, all we need to point out is that we can hear the spatial
> characteristics of live and reproduced sound, and those characteristics are
> very important to both the enjoyment of live sound (and quality of a concert
> hall) and the realism of the reproduction.
>
> Now, not to impugn your intelligence or vast knowledge, if most of us think
> we understand what this spatial stuff is all about, but to emphasize the
> difference between the spatial and the temporal, the two aspects of a sound
> field that get continually confused with each other, I would like to add an
> illustration.
>
> Your friend is a novice audiophile on his way back from Best Buy, where he
> has purchased a new home theater system and learned all about hi fi from the
> salesman. He sets up the speakers all up front, perhaps all in a row or all
> on top of his "teevee" and he plays a movie or some music. It is very
> accurate, plays all of the frequencies and the timings of the ambience that
> were recorded, but it just doesn't sound realistic yet. So you go over and
> show him how to correctly place all of the speakers to reconstruct a
> semblance of the sound field that was recorded in his listening room. You
> place the front speakers LCR for correct perspective of the frontal
> soundstage, and you place the surround speakers back and to the sides, where
> the ambience of the hall should come from. You have addressed the SPATIAL
> characteristic, which has nothing to do with the temporal, but rather with
> the directions form which the various sound fields arrive at the listener,
> or exist in the listening room. I bring this up because of a frequent
> question about my statements on getting the spatial more correct. They
> always tell me that getting the spatial more correct can't work because you
> can't make a small room sound like a concert hall. They have confused, or
> "fused" the spatial and the temporal.
>
> So I try to explain the difference but for now I only want to state that
> these are the main characteristics of sound that we can hear and try to
> reproduce. Do you have any other "biggies" that I have left out? I would be
> fascinated.
>
> So what is the state of the art of attempts to reproduce all of these and
> how far can we go?
>
> The physical size and waveform accuracy and power we can easily get a handle
> on and improve, if for example we are in a small room and we understand that
> limitation. The spatial stuff is the biggie and is where I say we need more
> basic understanding of the process in order to get any further than we have
> already come in 100 or so years.
>
> The most basic and foolish mislead is thinking that good "stereo" comes from
> the direct sound alone, and trying to kill the room reflections or design a
> speaker that casts all of its sound toward your hapless ears. This
> misconception, or mislead, is caused by the confusion between stereo and
> binaural.

Again, I disagree. Binaural sound only works with headphones. Stereo is
designed for speakers. Hall ambience can be captured by the recording
microphones, or it can be captured by an auxiliary pair of mikes placed at
some distance in the room away from the musicians. It can also be
artificially created and added either at the recording end of the chain, or
the playback end.

> Stereo does not work like a "window into another acoustic."

Actually, it is a pair of windows (in a traditional two-channel stereo
setup). They're called your speakers. Their job is to recreate, as accurately
as possible, the electrical signal fed to them. Ideally, the acoustic
wavefront they produce will be exactly like the audio signal they are fed.
The problem is that's an illusive goal. It's pretty much impossible, in fact.
We can get close, and we DO get closer all the time. but each step we take
toward that goal, gets smaller than the one before it. Taken metaphorically,
we'll never get there.


Rather, if you
> think of it as a model of the original, in which your room is the performing
> space and your speaker setup attempts to get the spatial closer to the
> original, then you have a fighting chance for greater realism, but you also
> inherit the understanding that it is not an "accuracy" process, and we can
> never get all the way there. We cannot, in other words, totally get to the
> goal of a "you are there" experience but rather more like a "they are here"
> experience in which your room is the performing space and you design it for
> good sound and arrange THE BIG THREE of speaker positioning, radiation
> pattern, and room acoustics to get the model closer to the live situation.

And this would be different from two open windows onto a space where a
musical ensemble is playing, how?

> So what can we hear? We can hear the spatial, spectral, and temporal
> characteristics of our listening room and speaker situation, or layout,
> superimposed on that of the recording, and we can hear the physical size,
> power, and electronic accuracy of your system. When we play back any
> recording, we CHANGE the spatial characteristics of the original to those of
> our playback system and room.
>
> That is slightly too bad, but once we understand the limitations of the
> system and what can be achieved, we can stop worrying about false goals and
> start concentrating on more fruitful paths that can lead to greater realism.

Which would be?

[Randy Yates[_2_]]

"Gary Eickmeier" > writes:
> [...]
> The most basic and foolish mislead is thinking that good "stereo" comes from
> the direct sound alone, and trying to kill the room reflections or design a
> speaker that casts all of its sound toward your hapless ears. This
> misconception, or mislead, is caused by the confusion between stereo and
> binaural.

There are serious problems in accurately recreating a sound field using
stereo in the manner you're describing. The most basic linear systems
theory and acoustics will tell us this logically.

In all cases the sound reaching the listener can be described in the
frequency domain as

L(\omega)= F(\omega) . H(\omega),

where F(\omega) is the signal, H(\omega) is the room response, and "."
represents multiplication. In the time domain this is

l(t) = f(t) * h(t),

where "*" denotes convolution and h(t) is the impulse response of the
room.[1]

Consider the case of a "dead" signal source, e.g., a single human
speaker in a small room, and a "live" room in which the impulse response
h(t) is significantly longer than that of the recording environment.
Then you will necessarily perceive the "wrong" sound field.

Was Dr. Duane Cooper, formerly of the University of Illinois/Urbana a
fool? I think not! He formulated a concept of "transaural processing"
which accounted for the room response, interaural crosstalk, and head
related transfer functions so that the original auditory event could be
faithfully recreated at the listener.

The problem with Cooper's method is that, with the technology at that
time, the listener was locked into one position - not very practical.
Also the formulated problem would only work for one listener.

However, there may be feasible solutions at this day and age involving
multiple emitters, tracking technology, and adaptive beamforming. All
TBD...
--
Randy Yates
Digital Signal Labs
http://www.digitalsignallabs.com

[Gary Eickmeier]

Oh boy. A treasure trove of binaural vs stereo confusion.

Listen fellers, I know you probably have engineering degrees to hold over my
head, and you know calculus, and I freely admit that Audio Empire has a lot
more experience with professional recording than I do, but... well, what I
am offering as an industrial designer (those who study whole systems for
understanding before putting pencil to paper to design a particular product)
is more conceptual than all of the objections that you have stated. We all
have preconceptions about how something is, or how something should be done,
and it is hard to shake off those preconceptions, especially if you have
based a lot of your life and your career and expertise on them. Galileo and
Copernicus didn't have it easy... they all laughed at Christopher Columbus
when he said the world was round... they all laughed when Edison recorded
sound...

To try and understand the magnitude of what I perceive I am up against,
imagine duking it out with a Chistian fundamentalist over atonement
theology. He will fight you tooth and nail using the bible as his reference
and "proof."

Allow me to just re-type one small section of my paper in which I propose an
analogy that uses the "window to another acoustic" that is one of the
primary conceptions you have both used to illustrate how stereo works.

(I discuss the difference between binaural and stereophonic, the lack of a
single stereo theory or explanation of how it works, what we are doing with
the process, the Bell labs experiments, the Blumlein patent.). I continue:



The trend to note with both of these versions is that stereo is thought to
operate as a sort of windowing or portaling process wherein the sound that
was recorded is simply being relayed to the listener by the reproduction
chain. Stereophonic sound is thought to be a "trick" that attempts to fool
the ears into hearing all audible spatial properties of live sound strictly
by means of lateralization - like looking through a portal into another
acoustic space. The degree of success of the illusion is thought to depend
on the "accuracy" of the system, and the status of stereo theory as we know
it today can be thought of as a search for greater and greater accuracy.

Notice also that the above descriptions are strictly two-dimensional
processes. The theories are based only on the direct sound radiatied from a
pair or a line of speakers. They are "blind" to the effects of loudspeaker
radiation pattern, positioning, and room acoustics. We started with the
system definition as a field type system, reproduced in a real acoustic
space by loudspeakers, but as far as the explanation of how it works goes,
the playback room might as well not even exist, and nowhere do we find
reflected sound incorporated as part of stereo theory.

AN ANALOGY

The best way to illustrate this highly conceptual problem is with an
analogy.

Many people have used the "brick wall" analogy - that stereo is something
like punching out two holes in a brick wall separating you from the
performance. Some writers widen the two holes and join them together, some
claim that their systems knock down the entire wall, but we are always
witnessing teh sound through a large portal, standing on the outside looking
in.

That's a good starting point, and a nice, simple analogy to make the desired
point, but let's take it one step further. Imagine your listening room
plunked down in the middle of Symphony Hall with you in it. We're going to
punch out first two holes (or a portal) in front of us, to "let the music
in." Then, the surround sound devotees will puch out some more holes in the
rear and perhaps side walls, to let all the ambience in. Under the
"accuracy" banner, we say that when the reproduction chain gets good enough,
the sound will be indistinguishable from this punched-out shell of a room,
with nothing between you and the music but air.

The caution at this point is that this would all be very fine thinking
except that, no matter how many channels we have, we will never quite make
it all the way because, in this analogy, we must remember that the sound can
get into the imaginary room but it can't get out, and so the sound still
bounces around the listening room with the time between reflections of the
smaller space.

The main point of this section, however, is that this is NOT a good analogy
at all.

Many people, especially audiophiles, have the impression that the recording
contains a perfect image of the performance as witnessed from the best seat
in the house. This may be true with binaural, but stereophonic is a very
much different process. The problem with the above analogy is that it
pictures the sound as having been "witnessed," or recorded, from the vantage
point of the listener in the room suspended in the middle of the concert
hall. This is not the case. What we have done is dispatched the microphones
up to the orchestra, recorded the musicians and the soundstage surrounding
them, and brought back the sound to be played again from entirely within our
room, not from outside with holes punched in the walls so we can hear it.
This is quite a different thing, and it forces us for the first time to
think of the listening room not as a nuisance variable but as the performing
space itself. For better or for worse, the room must be thought of as an
integral part of the sound, to be used to construct the same sort of spatial
patterns that existed in the real concert hall, rather than fought with
sound killing materials. I believe that this is for the better, because once
we reconstruct the sound fields in the playhback room, all of the
characteristics of live sound can be present, making the sound real and not
a trick. The stereophonic recording can be thought of as a concentrate, to
be mixed with the playback acoustic in a way that models the reproduction
after the real thing. Although we must inevitably hear some of the
listeining room along with the "flavor" of the recorded acoustic, the
realism can be stunning.

Gary Eickmeier

[Gary Eickmeier]

{Moderators - please allow edit}

Oh boy. This is what I was afraed of.

Listen fellers, I know you probably have engineering degrees to hold over my
head, and you know calculus, and I freely admit that Audio Empire has a lot
more experience with professional recording than I do, but... well, what I
am offering as an industrial designer (those who study whole systems for
understanding before putting pencil to paper to design a particular product)
is more conceptual than all of the objections that you have stated. We all
have preconceptions about how something is, or how something should be done,
and it is hard to shake off those preconceptions, especially if you have
based a lot of your life and your career and expertise on them. Galileo and
Copernicus didn't have it easy... they all laughed at Christopher Columbus
when he said the world was round... they all laughed when Edison recorded
sound...

Allow me to just re-type one small section of my paper in which I propose an
analogy that uses the "window to another acoustic" that is one of the
primary conceptions you have both used to illustrate how stereo works.

(I discuss the difference between binaural and stereophonic, the lack of a
single stereo theory or explanation of how it works, what we are doing with
the process, the Bell labs experiments, the Blumlein patent.). I continue:



The trend to note with both of these versions is that stereo is thought to
operate as a sort of windowing or portaling process wherein the sound that
was recorded is simply being relayed to the listener by the reproduction
chain. Stereophonic sound is thought to be a "trick" that attempts to fool
the ears into hearing all audible spatial properties of live sound strictly
by means of lateralization - like looking through a portal into another
acoustic space. The degree of success of the illusion is thought to depend
on the "accuracy" of the system, and the status of stereo theory as we know
it today can be thought of as a search for greater and greater accuracy.

Notice also that the above descriptions are strictly two-dimensional
processes. The theories are based only on the direct sound radiatied from a
pair or a line of speakers. They are "blind" to the effects of loudspeaker
radiation pattern, positioning, and room acoustics. We started with the
system definition as a field type system, reproduced in a real acoustic
space by loudspeakers, but as far as the explanation of how it works goes,
the playback room might as well not even exist, and nowhere do we find
reflected sound incorporated as part of stereo theory.

AN ANALOGY

The best way to illustrate this highly conceptual problem is with an
analogy.

Many people have used the "brick wall" analogy - that stereo is something
like punching out two holes in a brick wall separating you from the
performance. Some writers widen the two holes and join them together, some
claim that their systems knock down the entire wall, but we are always
witnessing teh sound through a large portal, standing on the outside looking
in.

That's a good starting point, and a nice, simple analogy to make the desired
point, but let's take it one step further. Imagine your listening room
plunked down in the middle of Symphony Hall with you in it. We're going to
punch out first two holes (or a portal) in front of us, to "let the music
in." Then, the surround sound devotees will puch out some more holes in the
rear and perhaps side walls, to let all the ambience in. Under the
"accuracy" banner, we say that when the reproduction chain gets good enough,
the sound will be indistinguishable from this punched-out shell of a room,
with nothing between you and the music but air.

The caution at this point is that this would all be very fine thinking
except that, no matter how many channels we have, we will never quite make
it all the way because, in this analogy, we must remember that the sound can
get into the imaginary room but it can't get out, and so the sound still
bounces around the listening room with the time between reflections of the
smaller space.

The main point of this section, however, is that this is NOT a good analogy
at all.

Many people, especially audiophiles, have the impression that the recording
contains a perfect image of the performance as witnessed from the best seat
in the house. This may be true with binaural, but stereophonic is a very
much different process. The problem with the above analogy is that it
pictures the sound as having been "witnessed," or recorded, from the vantage
point of the listener in the room suspended in the middle of the concert
hall. This is not the case. What we have done is dispatched the microphones
up to the orchestra, recorded the musicians and the soundstage surrounding
them, and brought back the sound to be played again from entirely within our
room, not from outside with holes punched in the walls so we can hear it.
This is quite a different thing, and it forces us for the first time to
think of the listening room not as a nuisance variable but as the performing
space itself. For better or for worse, the room must be thought of as an
integral part of the sound, to be used to construct the same sort of spatial
patterns that existed in the real concert hall, rather than fought with
sound killing materials. I believe that this is for the better, because once
we reconstruct the sound fields in the playhback room, all of the
characteristics of live sound can be present, making the sound real and not
a trick. The stereophonic recording can be thought of as a concentrate, to
be mixed with the playback acoustic in a way that models the reproduction
after the real thing. Although we must inevitably hear some of the
listeining room along with the "flavor" of the recorded acoustic, the
realism can be stunning.

Gary Eickmeier

[Andrew Haley]

Audio Empire > wrote:

> The Klipshorns were 50% efficient. That meant that every two Watts
> of amplifier power produced one acoustic Watt of sound (One acoustic
> Watt is defined as being equivalent to 107.5 dBSPL at around a meter
> from an omnidirectional source)!

Even if that famous 104 dB/W figure is true, AFAICR that's a half-
space measurement, and the horns are highly directional. It doesn't
equate to 50% efficiency, which would be miraculous. I reckon it's
about 15% efficiency, which is extremely good.

> Of course, most speakers are nowhere near that efficient and it can
> often take more than 100 electrical Watts to produce one Acoustic
> Watt at one meter. That would make those speakers 1% efficient. Now,
> since sound pressure dissipates at a rate inversely proportional to
> the distance one is from the speakers,

You're assuming omnidirectional raidators.

Andrew.

[Audio Empire]

On Wed, 16 May 2012 03:57:25 -0700, Gary Eickmeier wrote
(in article >):

> Oh boy. A treasure trove of binaural vs stereo confusion.
>
> Listen fellers, I know you probably have engineering degrees to hold over my
> head, and you know calculus, and I freely admit that Audio Empire has a lot
> more experience with professional recording than I do, but... well, what I
> am offering as an industrial designer (those who study whole systems for
> understanding before putting pencil to paper to design a particular product)
> is more conceptual than all of the objections that you have stated. We all
> have preconceptions about how something is, or how something should be done,
> and it is hard to shake off those preconceptions, especially if you have
> based a lot of your life and your career and expertise on them. Galileo and
> Copernicus didn't have it easy... they all laughed at Christopher Columbus
> when he said the world was round... they all laughed when Edison recorded
> sound...
>
> To try and understand the magnitude of what I perceive I am up against,
> imagine duking it out with a Chistian fundamentalist over atonement
> theology. He will fight you tooth and nail using the bible as his reference
> and "proof."

I don't think that's really an apt analogy. You have formulated, in a vacuum
of actual knowledge, and experience, it seems, a bunch of notions about how
you think this stuff works. People with knowledge (those engineering degrees
that you mention) and actual experience with the concepts that you blithely
throw around, try to explain to you why your notions are based on assumptions
not in evidence. It's not a "fight" between fundamentalism against atonement
theology, it's a fight between how you want things to be and how they
actually are. It's like someone with a deep background in physics and
electronics arguing with a layman who is convinced that he can hear the
difference between different expensive audio cables. The scientist KNOWS that
the cables can't sound any different, and can explain to the "true believer"
why this HAS to be so, but the true believer hasn't the background to follow
the argument and KNOWS what he thinks he hears.
>
> Allow me to just re-type one small section of my paper in which I propose an
> analogy that uses the "window to another acoustic" that is one of the
> primary conceptions you have both used to illustrate how stereo works.
>
> (I discuss the difference between binaural and stereophonic, the lack of a
> single stereo theory or explanation of how it works, what we are doing with
> the process, the Bell labs experiments, the Blumlein patent.). I continue:
>
>
>
> The trend to note with both of these versions is that stereo is thought to
> operate as a sort of windowing or portaling process wherein the sound that
> was recorded is simply being relayed to the listener by the reproduction
> chain. Stereophonic sound is thought to be a "trick" that attempts to fool
> the ears into hearing all audible spatial properties of live sound strictly
> by means of lateralization - like looking through a portal into another
> acoustic space. The degree of success of the illusion is thought to depend
> on the "accuracy" of the system, and the status of stereo theory as we know
> it today can be thought of as a search for greater and greater accuracy.
>
> Notice also that the above descriptions are strictly two-dimensional
> processes. The theories are based only on the direct sound radiatied from a
> pair or a line of speakers. They are "blind" to the effects of loudspeaker
> radiation pattern, positioning, and room acoustics. We started with the
> system definition as a field type system, reproduced in a real acoustic
> space by loudspeakers, but as far as the explanation of how it works goes,
> the playback room might as well not even exist, and nowhere do we find
> reflected sound incorporated as part of stereo theory.

Well the window analogy is useful to explain the listener's relationship to
the sound source, it is overly simplistic. Too simplistic to explain the
signal that the speakers are reproducing. For instance, windows will never
give the listener inside the room, any image specificity, any front-to-back
layering, or any image height. That;s because speakers are attempting to
reproduce an audio signal picked up by microphones with certain
characteristics that are very unlike "two open windows".
>
> AN ANALOGY
>
> The best way to illustrate this highly conceptual problem is with an
> analogy.
>
> Many people have used the "brick wall" analogy - that stereo is something
> like punching out two holes in a brick wall separating you from the
> performance.

Only in the sense that it explains the listener's RELATIONSHIP with the sound
field as produced by the speakers. It is not, by any stretch of the
imagination, a "model" for stereophonic sound and it would be a mistake to
see it that way.

[Gary Eickmeier]

"Audio Empire" > wrote in message
...


> The human ear is a sensitive tool, no doubt about it. Unfortunately, it is
> also an interpretive tool. There is no sound preception without the brain
> and
> the brain brings with it a lifetime of experience, preconceptions, and
> personal preferences. It applies these factors to everything we hear and
> they're difficult to overcome. In fact, I'd go so far as to say that we,
> as
> individuals, can't overcome them. We are going to hear what we want to
> hear
> or what we expect to hear. The only way around this is to take human bias
> out
> of the equation with tests that are either totally objective (such as
> measurements using instrumentation) or by taking part in listening tests
> which remove as much of the human propensity for self-delusion as possible
> and relying on the statistical results. To paraphrase an old adage: he who
> trusts his own ear/brain interface to judge audio qualities is a fool.

Careful. Our ears are the final arbiter in audibility of effects. You just
have to keep the brain at by with DBT. What I mean by hard-nosed listening
is trying to keep the overactive imagination out of the discussion and being
honest about what you are hearing. If you can tell that the sound is coming
from a pair of speakers, and you are not being fooled into thinking "you are
there," then admit it and try for some improvements.

> Phase anomalies are real. You can measure them and you can easily hear
> their
> effects. All one has to do is walk past a stereo pair of speakers that are
> wired out-of phase to instantly hear the results. You can also take a
> recording made with a pair of spaced omnis and sum the two channels to
> mono
> to immediately hear (and see, on an oscilloscope) some of the instruments
> go
> away.
>

Monophonic phase response is not audible, and wasting time time aligning
drivers is a dead end.


> I have to disagree here. While you are correct about being able to tell,
> even
> blindfolded, the approximate size of a room you're in, you are wrong about
> the results. It is possible to make a small room sound good much more
> easily
> than it is possible to make a large room sound good. Small rooms have all
> the
> characteristics that you mention (and then some) but they are all fairly
> controllable.

Would you agree that you cannot make a room sound bigger than it actually is
by playing a sound recorded in a larger room?


> Again you are grossly oversimplifying. Power, in and of itself is merely a
> means to an end. The end is moving air. The more air you can move the more
> realistically an audio system will load the room with sound and the amount
> of
> air any speaker can move with a Watt of electrical power from an amplifier
> is
> determined by two things: the efficiency of the speaker and the volume of
> the
> space one is trying to fill.

and so on. Obviously, all I mean is acoustic power. I don't care how you get
there, with normal speakers and a lot of power, or efficient horn speakers
and less powerful amps, just mean how loud can it go undistorted.
..

> Actually, it is a pair of windows (in a traditional two-channel stereo
> setup). They're called your speakers. Their job is to recreate, as
> accurately
> as possible, the electrical signal fed to them. Ideally, the acoustic
> wavefront they produce will be exactly like the audio signal they are fed.
> The problem is that's an illusive goal. It's pretty much impossible, in
> fact.
> We can get close, and we DO get closer all the time. but each step we take
> toward that goal, gets smaller than the one before it. Taken
> metaphorically,
> we'll never get there.

Good illustration of a false goal. Accuracy of what compared to what? Your
statement above would indicate that you think the ideal speaker is a point
source directional jobby that has no output to the rear or sides anc can
cast all of its output directly toward your ears. That would be perfect
"accuracy" of the wavefront compared to the electrical signal fed them

But stereo is not an "accuracy" process. That signal contains recorded
information from the venue that arrived at the microphones from widely
varying incident angles. I think you know what I mean. To shove all of that
recorded sound at you from just that one point in space would CHANGE the
spatial characteristics of the original to those of your speakers. So would
that be "accurate"?
>
>
> Rather, if you
>> think of it as a model of the original, in which your room is the
>> performing
>> space and your speaker setup attempts to get the spatial closer to the
>> original, then you have a fighting chance for greater realism, but you
>> also
>> inherit the understanding that it is not an "accuracy" process, and we
>> can
>> never get all the way there. We cannot, in other words, totally get to
>> the
>> goal of a "you are there" experience but rather more like a "they are
>> here"
>> experience in which your room is the performing space and you design it
>> for
>> good sound and arrange THE BIG THREE of speaker positioning, radiation
>> pattern, and room acoustics to get the model closer to the live
>> situation.
>
> And this would be different from two open windows onto a space where a
> musical ensemble is playing, how?

Please see the next thread, "What Can We Hear?".


>
>> So what can we hear? We can hear the spatial, spectral, and temporal
>> characteristics of our listening room and speaker situation, or layout,
>> superimposed on that of the recording, and we can hear the physical size,
>> power, and electronic accuracy of your system. When we play back any
>> recording, we CHANGE the spatial characteristics of the original to those
>> of
>> our playback system and room.
>>
>> That is slightly too bad, but once we understand the limitations of the
>> system and what can be achieved, we can stop worrying about false goals
>> and
>> start concentrating on more fruitful paths that can lead to greater
>> realism.
>
> Which would be?

Getting the spatial more correct by looking at the reproduction as a model
of the live situation, rather than a "window" to another space. This, in
turn, forces you to think about "the big three" of radiation pattern, room
positioning, and room acoustics.

Gary Eickmeier

[Audio Empire]

On Wed, 16 May 2012 16:41:19 -0700, Gary Eickmeier wrote
(in article >):

> "Audio Empire" > wrote in message
> ...
>
>
>> The human ear is a sensitive tool, no doubt about it. Unfortunately, it is
>> also an interpretive tool. There is no sound preception without the brain
>> and
>> the brain brings with it a lifetime of experience, preconceptions, and
>> personal preferences. It applies these factors to everything we hear and
>> they're difficult to overcome. In fact, I'd go so far as to say that we,
>> as
>> individuals, can't overcome them. We are going to hear what we want to
>> hear
>> or what we expect to hear. The only way around this is to take human bias
>> out
>> of the equation with tests that are either totally objective (such as
>> measurements using instrumentation) or by taking part in listening tests
>> which remove as much of the human propensity for self-delusion as possible
>> and relying on the statistical results. To paraphrase an old adage: he who
>> trusts his own ear/brain interface to judge audio qualities is a fool.
>
> Careful. Our ears are the final arbiter in audibility of effects. You just
> have to keep the brain at by with DBT. What I mean by hard-nosed listening
> is trying to keep the overactive imagination out of the discussion and being
> honest about what you are hearing. If you can tell that the sound is coming
> from a pair of speakers, and you are not being fooled into thinking "you are
> there," then admit it and try for some improvements.

I think I just said that...

>> Phase anomalies are real. You can measure them and you can easily hear
>> their
>> effects. All one has to do is walk past a stereo pair of speakers that are
>> wired out-of phase to instantly hear the results. You can also take a
>> recording made with a pair of spaced omnis and sum the two channels to
>> mono
>> to immediately hear (and see, on an oscilloscope) some of the instruments
>> go
>> away.
>>
>
> Monophonic phase response is not audible, and wasting time time aligning
> drivers is a dead end.

That's irrelevant to my comments. In stereo, phase differences are part of
what constitutes "channel separation" and therefore imaging. Also, there is
nothing of a "waste of time" about making sure all of your speakers are
in-phase and it's easy to hear when they are not. Yet, if I had a dollar for
every stereo store's demo room I've entered only to say to the salesperson
"Your speakers are out of phase", I could have a really good steak dinner at
Ruth's Chris!

Also, the fact that spaced omni's aren't phase coherent is one of the reasons
why Telarc's classical recordings never imaged as well as they should. I'd
say phase is very important. OTOH, I've never heard any REAL advantage to
"time aligned" drivers in speakers. It might be theoretically important, but
the difference certainly doesn't reach out and grab you.

>
>> I have to disagree here. While you are correct about being able to tell,
>> even
>> blindfolded, the approximate size of a room you're in, you are wrong about
>> the results. It is possible to make a small room sound good much more
>> easily
>> than it is possible to make a large room sound good. Small rooms have all
>> the
>> characteristics that you mention (and then some) but they are all fairly
>> controllable.
>
> Would you agree that you cannot make a room sound bigger than it actually is
> by playing a sound recorded in a larger room?

Of course you can't. However, you can use a DSP based playback reverb system
(such as those made by Lexicon) and speakers placed in the back of a smallish
room and by choosing the correct amounts of delay and reverb make the room
sound bigger. Like I said, a small room's anomalies are controllable by
various means.

>> Again you are grossly oversimplifying. Power, in and of itself is merely a
>> means to an end. The end is moving air. The more air you can move the more
>> realistically an audio system will load the room with sound and the amount
>> of
>> air any speaker can move with a Watt of electrical power from an amplifier
>> is
>> determined by two things: the efficiency of the speaker and the volume of
>> the
>> space one is trying to fill.
>
> and so on. Obviously, all I mean is acoustic power. I don't care how you get
> there, with normal speakers and a lot of power, or efficient horn speakers
> and less powerful amps, just mean how loud can it go undistorted.

Then you should have said acoustic power. As you wrote it, it pertained to
amplifier power. That's all I had to go by in my response.


>> Actually, it is a pair of windows (in a traditional two-channel stereo
>> setup). They're called your speakers. Their job is to recreate, as
>> accurately
>> as possible, the electrical signal fed to them. Ideally, the acoustic
>> wavefront they produce will be exactly like the audio signal they are fed.
>> The problem is that's an illusive goal. It's pretty much impossible, in
>> fact.
>> We can get close, and we DO get closer all the time. but each step we take
>> toward that goal, gets smaller than the one before it. Taken
>> metaphorically,
>> we'll never get there.
>
> Good illustration of a false goal. Accuracy of what compared to what?

That should be apparent. The speakers can only be accurate to what's fed them
and that's all I meant. The acoustic waveform emanating from the speakers
should look EXACTLY like the electrical signal fed to it. I.E. feed the
speakers a perfect square wave and it should produce a perfect square wave in
space - at any frequency in the audio spectrum. No overshoots, no rounding of
the flat tops of waves, no ringing. Obviously, that's impossible.

Your
> statement above would indicate that you think the ideal speaker is a point
> source directional jobby that has no output to the rear or sides anc can
> cast all of its output directly toward your ears. That would be perfect
> "accuracy" of the wavefront compared to the electrical signal fed them

I don't think that's obvious at all, and in fact, I said nothing of the sort!
The perfect speaker is, by definition, an infinitely small, omnidirectional
point source in the form of a pulsating sphere that can respond instantly to
every nuance of the signal fed to it (see above).
>
> But stereo is not an "accuracy" process. That signal contains recorded
> information from the venue that arrived at the microphones from widely
> varying incident angles. I think you know what I mean. To shove all of that
> recorded sound at you from just that one point in space would CHANGE the
> spatial characteristics of the original to those of your speakers. So would
> that be "accurate"?

Accuracy on that level is a goal. It is not an achievable goal, because the
physics of reality mitigate against it. But trying to get there (such as what
MBL is trying to do with the MB-101 MkII and their pulsating spheres) is a
worthy goal.

ther, if you
>>> think of it as a model of the original, in which your room is the
>>> performing
>>> space and your speaker setup attempts to get the spatial closer to the
>>> original, then you have a fighting chance for greater realism, but you
>>> also
>>> inherit the understanding that it is not an "accuracy" process, and we
>>> can
>>> never get all the way there. We cannot, in other words, totally get to
>>> the
>>> goal of a "you are there" experience but rather more like a "they are
>>> here"
>>> experience in which your room is the performing space and you design it
>>> for
>>> good sound and arrange THE BIG THREE of speaker positioning, radiation
>>> pattern, and room acoustics to get the model closer to the live
>>> situation.
>>
>> And this would be different from two open windows onto a space where a
>> musical ensemble is playing, how?
>
> Please see the next thread, "What Can We Hear?".
>
>
>>
>>> So what can we hear? We can hear the spatial, spectral, and temporal
>>> characteristics of our listening room and speaker situation, or layout,
>>> superimposed on that of the recording, and we can hear the physical size,
>>> power, and electronic accuracy of your system. When we play back any
>>> recording, we CHANGE the spatial characteristics of the original to those
>>> of
>>> our playback system and room.
>>>
>>> That is slightly too bad, but once we understand the limitations of the
>>> system and what can be achieved, we can stop worrying about false goals
>>> and
>>> start concentrating on more fruitful paths that can lead to greater
>>> realism.
>>
>> Which would be?
>
> Getting the spatial more correct by looking at the reproduction as a model
> of the live situation, rather than a "window" to another space. This, in
> turn, forces you to think about "the big three" of radiation pattern, room
> positioning, and room acoustics.

OK first of all, are you aware that at least 90% of all of the commercially
available classical recordings don't image anywhere near as well as the state
of the art has allowed since stereo recording began in the mid-Fifties? How
can you "get the spatial more correct" on the listening end when most record
companies can't get it right on the recording end, the end over which you and
I, as listeners have NO control?

And getting our system to be more like a model of the live situation can't be
done with two channels, but I'll let you in on something, you can get damn
close. I have symphony orchestra recordings that I have made with just two
cardioid condenser microphones mounted on a stereo "T"-bar with their axis's
90-degrees apart that image so well, that you can close your eyes and point
to every instrument in the orchestra. You can even tell that the back row of
brass is sitting up on risers and are higher than the woodwinds in front of
them! How's that for realistic soundstage. It can be done, it has been done
in commercial recordings, but it's SELDOM done. I can count on the fingers of
two hands the commercial classical recordings that I have (and I have
literally thousands both in Vinyl and CD) that image like that. You want to
go on a crusade for better sound? Start there.

[Gary Eickmeier]

"Audio Empire" > wrote in message
...
> On Wed, 16 May 2012 03:57:25 -0700, Gary Eickmeier wrote
> (in article >):
>
>> Oh boy. A treasure trove of binaural vs stereo confusion.
>>
>> Listen fellers, I know you probably have engineering degrees to hold over
>> my
>> head, and you know calculus, and I freely admit that Audio Empire has a
>> lot
>> more experience with professional recording than I do, but... well, what
>> I
>> am offering as an industrial designer (those who study whole systems for
>> understanding before putting pencil to paper to design a particular
>> product)
>> is more conceptual than all of the objections that you have stated. We
>> all
>> have preconceptions about how something is, or how something should be
>> done,
>> and it is hard to shake off those preconceptions, especially if you have
>> based a lot of your life and your career and expertise on them. Galileo
>> and
>> Copernicus didn't have it easy... they all laughed at Christopher
>> Columbus
>> when he said the world was round... they all laughed when Edison recorded
>> sound...
>>
>> To try and understand the magnitude of what I perceive I am up against,
>> imagine duking it out with a Chistian fundamentalist over atonement
>> theology. He will fight you tooth and nail using the bible as his
>> reference
>> and "proof."
>
> I don't think that's really an apt analogy. You have formulated, in a
> vacuum
> of actual knowledge, and experience, it seems, a bunch of notions about
> how
> you think this stuff works. People with knowledge (those engineering
> degrees
> that you mention) and actual experience with the concepts that you
> blithely
> throw around, try to explain to you why your notions are based on
> assumptions
> not in evidence. It's not a "fight" between fundamentalism against
> atonement
> theology, it's a fight between how you want things to be and how they
> actually are. It's like someone with a deep background in physics and
> electronics arguing with a layman who is convinced that he can hear the
> difference between different expensive audio cables. The scientist KNOWS
> that
> the cables can't sound any different, and can explain to the "true
> believer"
> why this HAS to be so, but the true believer hasn't the background to
> follow
> the argument and KNOWS what he thinks he hears.

OK, I tried to edit that paragraph, but I really did think it an apt
analogy, and you didn't follow it. It is not a fight between fundamentalists
and atonement theology; it is the fundamentalists who believe in atonement
theology. They support it as "experts" because of all they have learned from
the bible and bible school. But if you come along and try to tell them that
these notions are wrong, they will give all of the arguments that you have
just given me, that what I am saying is not supported by the "bible" of the
Handbook of Audio Engineering.

I have tried to point out that even among the "experts" there is no single
stereo theory, all laid out and accepted by all engineers. There is just
about as much variability in the audio engineering community as there is in
religion. Look at loudspeaker design man. No one knows what the hell he is
doing. There are dipoles, bipoles, omnis, and megaphones, line sources,
point sources, wallspeakers and free standing speakers. No one has a clue
how or why to do any of this, nor is there a guideline for any sort of
"correct" design or theory of stereo.

Nor am I just entering the room. I have been studying this stuff for almost
30 years now.

>>
>> Allow me to just re-type one small section of my paper in which I propose
>> an
>> analogy that uses the "window to another acoustic" that is one of the
>> primary conceptions you have both used to illustrate how stereo works.
>>
>> (I discuss the difference between binaural and stereophonic, the lack of
>> a
>> single stereo theory or explanation of how it works, what we are doing
>> with
>> the process, the Bell labs experiments, the Blumlein patent.). I
>> continue:
>>
>>
>>
>> The trend to note with both of these versions is that stereo is thought
>> to
>> operate as a sort of windowing or portaling process wherein the sound
>> that
>> was recorded is simply being relayed to the listener by the reproduction
>> chain. Stereophonic sound is thought to be a "trick" that attempts to
>> fool
>> the ears into hearing all audible spatial properties of live sound
>> strictly
>> by means of lateralization - like looking through a portal into another
>> acoustic space. The degree of success of the illusion is thought to
>> depend
>> on the "accuracy" of the system, and the status of stereo theory as we
>> know
>> it today can be thought of as a search for greater and greater accuracy.
>>
>> Notice also that the above descriptions are strictly two-dimensional
>> processes. The theories are based only on the direct sound radiatied from
>> a
>> pair or a line of speakers. They are "blind" to the effects of
>> loudspeaker
>> radiation pattern, positioning, and room acoustics. We started with the
>> system definition as a field type system, reproduced in a real acoustic
>> space by loudspeakers, but as far as the explanation of how it works
>> goes,
>> the playback room might as well not even exist, and nowhere do we find
>> reflected sound incorporated as part of stereo theory.
>
> Well the window analogy is useful to explain the listener's relationship
> to
> the sound source, it is overly simplistic. Too simplistic to explain the
> signal that the speakers are reproducing. For instance, windows will never
> give the listener inside the room, any image specificity, any
> front-to-back
> layering, or any image height. That;s because speakers are attempting to
> reproduce an audio signal picked up by microphones with certain
> characteristics that are very unlike "two open windows".

So you are agreeing with me up to this point - or what?

>> AN ANALOGY
>>
>> The best way to illustrate this highly conceptual problem is with an
>> analogy.
>>
>> Many people have used the "brick wall" analogy - that stereo is something
>> like punching out two holes in a brick wall separating you from the
>> performance.
>
> Only in the sense that it explains the listener's RELATIONSHIP with the
> sound
> field as produced by the speakers. It is not, by any stretch of the
> imagination, a "model" for stereophonic sound and it would be a mistake to
> see it that way.

Where is the rest of it? Did you press the SEND too soon?

Gary Eickmeier

[Audio Empire]

On Thu, 17 May 2012 03:48:17 -0700, Gary Eickmeier wrote
(in article >):

> "Audio Empire" > wrote in message
> ...
>> On Wed, 16 May 2012 03:57:25 -0700, Gary Eickmeier wrote
>> (in article >):
>>
>>> Oh boy. A treasure trove of binaural vs stereo confusion.
>>>
>>> Listen fellers, I know you probably have engineering degrees to hold over
>>> my
>>> head, and you know calculus, and I freely admit that Audio Empire has a
>>> lot
>>> more experience with professional recording than I do, but... well, what
>>> I
>>> am offering as an industrial designer (those who study whole systems for
>>> understanding before putting pencil to paper to design a particular
>>> product)
>>> is more conceptual than all of the objections that you have stated. We
>>> all
>>> have preconceptions about how something is, or how something should be
>>> done,
>>> and it is hard to shake off those preconceptions, especially if you have
>>> based a lot of your life and your career and expertise on them. Galileo
>>> and
>>> Copernicus didn't have it easy... they all laughed at Christopher
>>> Columbus
>>> when he said the world was round... they all laughed when Edison recorded
>>> sound...
>>>
>>> To try and understand the magnitude of what I perceive I am up against,
>>> imagine duking it out with a Chistian fundamentalist over atonement
>>> theology. He will fight you tooth and nail using the bible as his
>>> reference
>>> and "proof."
>>
>> I don't think that's really an apt analogy. You have formulated, in a
>> vacuum
>> of actual knowledge, and experience, it seems, a bunch of notions about
>> how
>> you think this stuff works. People with knowledge (those engineering
>> degrees
>> that you mention) and actual experience with the concepts that you
>> blithely
>> throw around, try to explain to you why your notions are based on
>> assumptions
>> not in evidence. It's not a "fight" between fundamentalism against
>> atonement
>> theology, it's a fight between how you want things to be and how they
>> actually are. It's like someone with a deep background in physics and
>> electronics arguing with a layman who is convinced that he can hear the
>> difference between different expensive audio cables. The scientist KNOWS
>> that
>> the cables can't sound any different, and can explain to the "true
>> believer"
>> why this HAS to be so, but the true believer hasn't the background to
>> follow
>> the argument and KNOWS what he thinks he hears.
>
> OK, I tried to edit that paragraph, but I really did think it an apt
> analogy, and you didn't follow it. It is not a fight between fundamentalists
> and atonement theology; it is the fundamentalists who believe in atonement
> theology. They support it as "experts" because of all they have learned from
> the bible and bible school. But if you come along and try to tell them that
> these notions are wrong, they will give all of the arguments that you have
> just given me, that what I am saying is not supported by the "bible" of the
> Handbook of Audio Engineering.

You give yourself too much credit. This stuff isn't "theory" (as in an
untried hypothesis) this stuff is fact backed by mathematics and sound
acoustical physics. Your "notion" is backed by...what? Your own conviction
that you are right and the world of engineering is wrong? I'm afraid it takes
more than that to get your ideas accepted as fact. To even be considered, you
have to have bona fides. You haven't stated yours. A few non-controlled
experiments with some cheap microphones and some untutored listening is
simply not compelling. I'm not trying to put you down here, nor am I trying
to be unkind, but when one pontificates, one needs to have some credibility.
You certainly are within your rights to postulate any new theory you wish,
but without the proper foundation, you shouldn't get all defensive if nobody
follows you down that road.

> I have tried to point out that even among the "experts" there is no single
> stereo theory, all laid out and accepted by all engineers. There is just
> about as much variability in the audio engineering community as there is in
> religion. Look at loudspeaker design man. No one knows what the hell he is
> doing. There are dipoles, bipoles, omnis, and megaphones, line sources,
> point sources, wallspeakers and free standing speakers. No one has a clue
> how or why to do any of this, nor is there a guideline for any sort of
> "correct" design or theory of stereo.

That's a gross oversimplification. You have confused the fact that there are
many different ways to apply known electrical and acoustic principles with a
lack of knowledge and understanding on the part of those in the business.
This is not the case. Everybody knows what's needed , it's just that because
what's needed is ultimately impossible and so many faceted, that different
manufactures tend to concentrate on different parts of the overall puzzle.
For instance, manufacturer A might concentrate on getting a flat frequency
response across a wide room dispersion. Manufacturer B might concentrate on
image specificity while manufacturer C tries to get a realistic dynamic
range. Add to that the fact that some of these goals are, in reality,
mutually exclusive (for a speaker, not for real instruments playing in real
space).
>
> Nor am I just entering the room. I have been studying this stuff for almost
> 30 years now.

But what are your credentials that make your "study" worthy of attention?
IOW, why should anyone listen to you? You have made a number of errors in
your assumptions already. It's fine that you have ideas, but without a
grounding in physics and the experience to go along with it, you are bound to
go off the track at some point.
>
>>>
>>> Allow me to just re-type one small section of my paper in which I propose
>>> an
>>> analogy that uses the "window to another acoustic" that is one of the
>>> primary conceptions you have both used to illustrate how stereo works.
>>>
>>> (I discuss the difference between binaural and stereophonic, the lack of
>>> a
>>> single stereo theory or explanation of how it works, what we are doing
>>> with
>>> the process, the Bell labs experiments, the Blumlein patent.). I
>>> continue:
>>>
>>>
>>>
>>> The trend to note with both of these versions is that stereo is thought
>>> to
>>> operate as a sort of windowing or portaling process wherein the sound
>>> that
>>> was recorded is simply being relayed to the listener by the reproduction
>>> chain. Stereophonic sound is thought to be a "trick" that attempts to
>>> fool
>>> the ears into hearing all audible spatial properties of live sound
>>> strictly
>>> by means of lateralization - like looking through a portal into another
>>> acoustic space. The degree of success of the illusion is thought to
>>> depend
>>> on the "accuracy" of the system, and the status of stereo theory as we
>>> know
>>> it today can be thought of as a search for greater and greater accuracy.
>>>
>>> Notice also that the above descriptions are strictly two-dimensional
>>> processes. The theories are based only on the direct sound radiatied from
>>> a
>>> pair or a line of speakers. They are "blind" to the effects of
>>> loudspeaker
>>> radiation pattern, positioning, and room acoustics. We started with the
>>> system definition as a field type system, reproduced in a real acoustic
>>> space by loudspeakers, but as far as the explanation of how it works
>>> goes,
>>> the playback room might as well not even exist, and nowhere do we find
>>> reflected sound incorporated as part of stereo theory.
>>
>> Well the window analogy is useful to explain the listener's relationship
>> to
>> the sound source, it is overly simplistic. Too simplistic to explain the
>> signal that the speakers are reproducing. For instance, windows will never
>> give the listener inside the room, any image specificity, any
>> front-to-back
>> layering, or any image height. That;s because speakers are attempting to
>> reproduce an audio signal picked up by microphones with certain
>> characteristics that are very unlike "two open windows".
>
> So you are agreeing with me up to this point - or what?
>
>>> AN ANALOGY
>>>
>>> The best way to illustrate this highly conceptual problem is with an
>>> analogy.
>>>
>>> Many people have used the "brick wall" analogy - that stereo is something
>>> like punching out two holes in a brick wall separating you from the
>>> performance.
>>
>> Only in the sense that it explains the listener's RELATIONSHIP with the
>> sound
>> field as produced by the speakers. It is not, by any stretch of the
>> imagination, a "model" for stereophonic sound and it would be a mistake to
>> see it that way.
>
> Where is the rest of it? Did you press the SEND too soon?

I had no comments on the rest of it.

[Sebastian Kaliszewski]

Gary Eickmeier wrote:
[... large snip ...]
> Many people, especially audiophiles, have the impression that the recording
> contains a perfect image of the performance as witnessed from the best seat
> in the house. This may be true with binaural, but stereophonic is a very
> much different process. The problem with the above analogy is that it
> pictures the sound as having been "witnessed," or recorded, from the vantage
> point of the listener in the room suspended in the middle of the concert
> hall. This is not the case.

What is the case is that all we got is just a little part of a soundfield
sampled at some predetermined points in space, and those points and in fact
acoustical transformation assigned with all of them do differ significantly
among different recordings. And anyway, such a transformation loses vast
majority of the information -- for a start it transforms (projects) 3
dimensional data into just single dimension.

> What we have done is dispatched the microphones
> up to the orchestra, recorded the musicians and the soundstage surrounding
> them, and brought back the sound to be played again from entirely within our
> room, not from outside with holes punched in the walls so we can hear it.
> This is quite a different thing, and it forces us for the first time to
> think of the listening room not as a nuisance variable but as the performing
> space itself. For better or for worse, the room must be thought of as an
> integral part of the sound, to be used to construct the same sort of spatial
> patterns that existed in the real concert hall, rather than fought with
> sound killing materials.

So far so good.

> I believe that this is for the better, because once
> we reconstruct the sound fields in the playhback room, all of the
> characteristics of live sound can be present, making the sound real and not
> a trick.

One litlle problem -- it might well be physically impossible.

> The stereophonic recording can be thought of as a concentrate, to
> be mixed with the playback acoustic in a way that models the reproduction
> after the real thing.

Recrating various sound fields from arbitrary events at different venues by just
two (or even 5 or 7 or 8) speakers is physically impossible. Period.

But our ears-brain system perceives only part of full physical reality of
acousting waves propagating in the space. Its fairly sensitive to some parts of
that reality, while being pretty insensitive to others. Thus here enters
psychoacoustics and all that masking, various sensitivituies for distinct
phenomenons and so on.

Certainly some aspects of soundfield characteristics are unimportant while
others are crucial for realistic presentation. So maybe, just maybe, after
prunning ale the unimportant stuff we could get some model which for our human
senses resempbles reality. But the reality might be such, that it's not possible
at all with less than 10 channels. In that sesond case we need to drop some
important characteristics to reduce degreees of freedom enough -- then all we
have is a trick, an illusion, and somewhat faulty one. My gut feeling is that
that second case is in fact true, but that's just my intuition and I might be
proven wrong.


> Although we must inevitably hear some of the
> listeining room along with the "flavor" of the recorded acoustic, the
> realism can be stunning.

But what's the basis of your claim. I read your white paper and what you wrote
here and probably some time ago stumbled upon something from you on some audio
forum(s)). What I miss is that actual theory you're talking about. You describe
how thing should be without showing how to reach tha goal nor even presenting
any argument that this goal is acheivable at all.
Your ideas about speaker directivity and placement such reflected images extend
the audio scene in a room which is typically (much) smaller than recorded
performance venue are all interesting but they do not for a theory. They are yet
another trick withing a big bag of tricks already known.


rgds
\SK
--
"Never underestimate the power of human stupidity" -- L. Lang
--
http://www.tajga.org -- (some photos from my travels)

[Audio Empire]

On Tue, 22 May 2012 16:18:05 -0700, Sebastian Kaliszewski wrote
(in article >):

<snip>

>> Although we must inevitably hear some of the
> > listeining room along with the "flavor" of the recorded acoustic, the
> > realism can be stunning.
>
> But what's the basis of your claim. I read your white paper and what
> you wrote here and probably some time ago stumbled upon something
> from you on some audio forum(s)). What I miss is that actual
> theory you're talking about. You describe how thing should be
> without showing how to reach tha goal nor even presenting
> any argument that this goal is acheivable at all. Your ideas about
> speaker directivity and placement such reflected images extend the
> audio scene in a room which is typically (much) smaller than
> recorded performance venue are all interesting but they do not for
> a theory. They are yet another trick withing a big bag of tricks
> already known.

Well, that is certainly one of the problems with Mr. Eickmeier's
missives. The other, and the one I have been taking him to task over,
is that his "observations"(?) seem to totally ignore the other side of
the process; the recording.

There is no single type of recording that can be modeled by any single
speaker setup. There are as many different recording methods as there
are recording engineers/producers or, for that matter, recording
venues. Which is used for what circumstance depends upon many
different things, not the least of which is the engineer/producer's
"taste" (or more likely, lack of it) in sound. Since capturing
everything, just like reproducing everything, is impossible, people
tend to concentrate on those aspects of the sound that they feel are
important. Often what the recording people are trying to capture and
what the listener is trying to reproduce in his home are two different
things. For instance, if I'm a listener who really gets-off on
pin-point imaging and I have spent a small fortune putting together a
system that gets the imaging right at the expense of, perhaps, some
other characteristics that don't interest me as much (like, perhaps,
low bass) and I play a recording that's been made with 48 or more
tracks and a forest of microphones, then I'm not going to enjoy that
recording because it has no image. OTOH, if I like well reproduced,
clean treble and have designed my system to highlight that, and I play
a recording that has been made with microphones that are too bright
or a recording where the engineer has toned down the high frequencies
because he doesn't like "sparkling highs" , then that listener is not
going to get optimum sound from either one of those recordings on his
playback system. A speaker system that can give decent reproduction
irrespective of the recording or the recording venue, simply doesn't
exist and it can't exist. Even if one could buy the ultimate,
theoretical ideal of a speaker system, unless that ideal is carried
all the way back to the recording so that all the music one plays on
one's ideal playback system was captured with a theoretically ideal
microphone system and then recorded with totally transparent
electronics to a theoretically perfect recording device, all bets are
off and the system is still going to be fundamentally flawed. The kind
of control that I believe that Mr. Eickmeier is advocating here is
simply impossible because the variables are simply too many.

[KH]

On 5/22/2012 4:18 PM, Sebastian Kaliszewski wrote:
> Gary Eickmeier wrote:
> [... large snip ...]
> > Many people, especially audiophiles, have the impression that the
> recording
> > contains a perfect image of the performance as witnessed from the
> best seat
> > in the house.

Really? I don't know anyone that believes that. It's physically
impossible with current recording practices.


> > This may be true with binaural,

I fail to see how it could.

> > but stereophonic is a very
> > much different process. The problem with the above analogy is that it
> > pictures the sound as having been "witnessed," or recorded, from the
> vantage
> > point of the listener in the room suspended in the middle of the concert
> > hall. This is not the case.

Which would be ideal, but again, not practical.


> What is the case is that all we got is just a little part of a
> soundfield sampled at some predetermined points in space, and those
> points and in fact acoustical transformation assigned with all of them
> do differ significantly among different recordings. And anyway, such a
> transformation loses vast majority of the information -- for a start it
> transforms (projects) 3 dimensional data into just single dimension.

Exactly. With the exception of temporal and magnitude clues, all
directional information is lost. There is no quantitative or
qualitative difference - in the recorded signal - between say a 500hz
75db signal *recorded* from 120⁰ (from any given reference point) and
the same signal recorded from 0⁰ or 180⁰. Note qualification - recorded
signal - obviously there could be level cues resulting from non-linear
microphone responses, but that's irrelevant.

> > What we have done is dispatched the microphones
> > up to the orchestra, recorded the musicians and the soundstage
> surrounding
> > them, and brought back the sound to be played again from entirely
> within our
> > room, not from outside with holes punched in the walls so we can hear
> it.

This "holes punched in the walls" theory you seem to want to ascribe to
the world at large is something I've never heard anyone believe in or
allude to. It's a faulty image, one fraught with wave interference
problems if nothing else. My "image" of stereo is creating a realistic,
continuous sound image exactly analogous to a live venue. Never totally
achieved, but pretty close at times.


<snip>
>
> > I believe that this is for the better, because once
> > we reconstruct the sound fields in the playhback room, all of the
> > characteristics of live sound can be present, making the sound real
> and not
> > a trick.
>
> One litlle problem -- it might well be physically impossible.

I'd say it's certainly impossible as a practical matter.

>
> > The stereophonic recording can be thought of as a concentrate, to
> > be mixed with the playback acoustic in a way that models the
> reproduction
> > after the real thing.

The problem is that the stereophonic recording doesn't have the
information required to create a 3-D sound field. It's gone. You can
certainly use reflected sound to try and capture some sense of
spaciousness, but it cannot be accurate since the you cannot parse the
part of the recording resulting from reflected sound from that resulting
from direct sound. If you could, you still would be screwing up the
directional clues because you cannot recreate the same incident angles
and delay times experienced in the venue.

<snip>
> > Although we must inevitably hear some of the
> > listeining room along with the "flavor" of the recorded acoustic, the
> > realism can be stunning.

You have to hear primarily the listening room, by definition, since
you're using that room to create the reflections. That's one of the
rubs. You're reflecting both the originally direct sound, and the
originally reflected sound. That can't be accurate. Speaker placement
isn't going to be able to create information that just isn't in the
recording. The "sound field" created by your approach is no less a
contrived illusion than are the many other approaches taken by more
conventional designers and engineers.

Can that seem more realistic despite being totally inaccurate?
Probably, in some instances, with some recordings, in some rooms, to
some listeners. I've heard a number of dipole speakers, in numerous
configurations, and a number of setups over the years with 901's, and
while you can certainly achieve a degree of spaciousness with reflected
sound, I've yet to hear any such systems that sounded more "real" to me
than more conventional speakers, properly set up. In fact, the best
system with 901's I have ever heard was way back in the day, using Mac
gear with 901's mounted on stands. Those stands, however, were JBL
L-100's. Not exactly in keeping with your theory, but capable of
surprising realism at times.

> But what's the basis of your claim. I read your white paper and what you
> wrote here and probably some time ago stumbled upon something from you
> on some audio forum(s)). What I miss is that actual theory you're
> talking about. You describe how thing should be without showing how to
> reach tha goal nor even presenting any argument that this goal is
> acheivable at all.
> Your ideas about speaker directivity and placement such reflected images
> extend the audio scene in a room which is typically (much) smaller than
> recorded performance venue are all interesting but they do not for a
> theory. They are yet another trick withing a big bag of tricks already
> known.

My sentiments exactly. I don't see anything particularly "new" in this
approach either with the exception of a false dichotomy - i.e. there's
the *right* way (Gary's way) and myriad *wrong* ways to create realism.
From the test paper referenced, I see a set of criteria that represent
the testers' opinion of what aspects are important for realism. One
thing I see missing is a simple "which is more realistic" question as a
control for the evaluation criteria chosen. It may well be that any
given listener might rate, for e.g. the Orion as better than the
reference in 4 categories, and equal in two, and still feel the
reference was subjectively more realistic (i.e. any given evaluator may
have significantly different key parameters that signal "realism" to
them than those posed by the test questions).

Having heard neither the Behringers nor the Orions (and obviously not
Gary's brew), I can't comment on the appropriateness of the test
speakers other than to say I have serious doubts about comparison of
sub/satellite systems - relative to bass reproduction - to the Orions.
Using the subs with the Orions, with level matching, would seem more
appropriate.

Keith

[Audio Empire]

On Wed, 23 May 2012 04:01:02 -0700, KH wrote
(in article >):

> On 5/22/2012 4:18 PM, Sebastian Kaliszewski wrote:
>> Gary Eickmeier wrote:
>> [... large snip ...]
>>> Many people, especially audiophiles, have the impression that the recording
>>> contains a perfect image of the performance as witnessed from the best seat
>>> in the house.

> Really? I don't know anyone that believes that. It's physically
> impossible with current recording practices.

Amen. I think most audiophiles realize that there are so many different ways
to capture a musical performance, that they can't all be optimum. In fact, if
a perfect image of a performance is the goal, then most microphone setups are
just plain wrong.

>>> This may be true with binaural,
>
> I fail to see how it could.

Since a good binaural setup has a pair of high-quality omnidirectional mikes
mounted in a dummy head which has been designed to mimic the dimensions and
the acoustical properties of a skin-covered human head, it probably comes
closer to the ideal capture medium than any other microphone technique.
However, binaural playback only works through headphones, and the quality of
the playback depends, in large measure, on the quality of the headphones.

Even so, binaural recordings cannot differentiate between sounds coming from
directly in front of the dummy head or directly behind. A simple experiment
using a binaural setup will prove this point. While recording, walk around
the dummy head jingling a set of keys. Play the recording back through
headphones. images from the sides appear as they should (often with shocking
realism, especially if one is not used to the effect), but when the keys move
to the front or back of the head, they sound as if they are IN the listener's
head rather than in front or in back. Our ears don't have this problem and
humans can easily tell the direction from which a sound emanates,
irrespective of its direction.

>>> but stereophonic is a very
>>> much different process. The problem with the above analogy is that it
>>> pictures the sound as having been "witnessed," or recorded, from the vantage
>>> point of the listener in the room suspended in the middle of the concert
>>> hall. This is not the case.
>
> Which would be ideal, but again, not practical.
>
>
>> What is the case is that all we got is just a little part of a
>> soundfield sampled at some predetermined points in space, and those
>> points and in fact acoustical transformation assigned with all of them
>> do differ significantly among different recordings. And anyway, such a
>> transformation loses vast majority of the information -- for a start it
>> transforms (projects) 3 dimensional data into just single dimension.
>
> Exactly. With the exception of temporal and magnitude clues, all
> directional information is lost. There is no quantitative or
> qualitative difference - in the recorded signal - between say a 500hz
> 75db signal *recorded* from 120 (from any given reference point) and
> the same signal recorded from 0 or 180. Note qualification - recorded
> signal - obviously there could be level cues resulting from non-linear
> microphone responses, but that's irrelevant.
>
>>> What we have done is dispatched the microphones
>>> up to the orchestra, recorded the musicians and the soundstage surrounding
>>> them, and brought back the sound to be played again from entirely within our
>>> room, not from outside with holes punched in the walls so we can hear it.
>
> This "holes punched in the walls" theory you seem to want to ascribe to
> the world at large is something I've never heard anyone believe in or
> allude to. It's a faulty image, one fraught with wave interference
> problems if nothing else. My "image" of stereo is creating a realistic,
> continuous sound image exactly analogous to a live venue. Never totally
> achieved, but pretty close at times.

The "windows on a performance" analogy works only to explain the listener='s
relationship to the sound source - and then only in the most fundamental way
(it assumes very directional speakers with no back-wave). It cannot be used
to describe the recorded performance's relationship with the listening room
at all.

> <snip>
>>
>>> I believe that this is for the better, because once
>>> we reconstruct the sound fields in the playhback room, all of the
>>> characteristics of live sound can be present, making the sound real and not
>>> a trick.
>>
>> One litlle problem -- it might well be physically impossible.
>
> I'd say it's certainly impossible as a practical matter.
>
>>
>>> The stereophonic recording can be thought of as a concentrate, to
>>> be mixed with the playback acoustic in a way that models the reproduction
>>> after the real thing.
>
> The problem is that the stereophonic recording doesn't have the
> information required to create a 3-D sound field. It's gone. You can
> certainly use reflected sound to try and capture some sense of
> spaciousness, but it cannot be accurate since the you cannot parse the
> part of the recording resulting from reflected sound from that resulting
> from direct sound. If you could, you still would be screwing up the
> directional clues because you cannot recreate the same incident angles
> and delay times experienced in the venue.

Yep.

> <snip>
>>> Although we must inevitably hear some of the
>>> listeining room along with the "flavor" of the recorded acoustic, the
>>> realism can be stunning.
>
> You have to hear primarily the listening room, by definition, since
> you're using that room to create the reflections. That's one of the
> rubs. You're reflecting both the originally direct sound, and the
> originally reflected sound. That can't be accurate. Speaker placement
> isn't going to be able to create information that just isn't in the
> recording. The "sound field" created by your approach is no less a
> contrived illusion than are the many other approaches taken by more
> conventional designers and engineers.

Very true.

[KH]

On 5/23/2012 5:15 PM, Audio Empire wrote:
> On Wed, 23 May 2012 04:01:02 -0700, KH wrote
> (in >):
><snip>
>>>> This may be true with binaural,
>>
>> I fail to see how it could.
>
> Since a good binaural setup has a pair of high-quality omnidirectional mikes
> mounted in a dummy head which has been designed to mimic the dimensions and
> the acoustical properties of a skin-covered human head, it probably comes
> closer to the ideal capture medium than any other microphone technique.
> However, binaural playback only works through headphones, and the quality of
> the playback depends, in large measure, on the quality of the headphones.
>
> Even so, binaural recordings cannot differentiate between sounds coming from
> directly in front of the dummy head or directly behind.

That's pretty much my point. You can come closer, but still no cigar.

<snip>
>> This "holes punched in the walls" theory you seem to want to ascribe to
>> the world at large is something I've never heard anyone believe in or
>> allude to. It's a faulty image, one fraught with wave interference
>> problems if nothing else. My "image" of stereo is creating a realistic,
>> continuous sound image exactly analogous to a live venue. Never totally
>> achieved, but pretty close at times.
>
> The "windows on a performance" analogy works only to explain the listener='s
> relationship to the sound source - and then only in the most fundamental way
> (it assumes very directional speakers with no back-wave). It cannot be used
> to describe the recorded performance's relationship with the listening room
> at all.

And pitiful speakers those would be indeed. I think one misconception
that Gary exhibits is a belief that there *is* some point or
presentation that would be universally agreed upon as "most realistic".
To a large extent that, as you've alluded to in this thread already,
is largely a matter of preference (whether image size, pinpoint imaging,
etc.) depending upon, in large part, the factors each individual
listener finds most central to the illusion. The fact that there are
various audiophile groups that, respectively, find box, horn, dipole,
and omnidirectional (e.g. MBL's - for the rich) to be most realistic
would tend to support that conclusion rather well IMO.

I think if we were to generalize anything to "all" audiophiles, it would
be that *they* recognize that the recording is the first, and most
fundamentally challenging part of the whole process. It is clear, to
most IMO, that the effectiveness of anything done on the playback side
of the process, no matter how innovative, or clever, will always be
limited by the information contained in the recorded signal.

Keith

[Randy Yates[_2_]]

Audio Empire > writes:
> [...]
> Since a good binaural setup has a pair of high-quality omnidirectional mikes
> mounted in a dummy head which has been designed to mimic the dimensions and
> the acoustical properties of a skin-covered human head, it probably comes
> closer to the ideal capture medium than any other microphone technique.
> However, binaural playback only works through headphones, and the quality of
> the playback depends, in large measure, on the quality of the headphones.
>
> Even so, binaural recordings cannot differentiate between sounds coming from
> directly in front of the dummy head or directly behind. A simple experiment
> using a binaural setup will prove this point. While recording, walk around
> the dummy head jingling a set of keys. Play the recording back through
> headphones. images from the sides appear as they should (often with shocking
> realism, especially if one is not used to the effect), but when the keys move
> to the front or back of the head, they sound as if they are IN the listener's
> head rather than in front or in back. Our ears don't have this problem and
> humans can easily tell the direction from which a sound emanates,
> irrespective of its direction.

This may be due to the concha resonance. As I understand it, one cannot
simply play back a binaural recording using earphones but must equalize
for this resonance as well.
--
Randy Yates
Digital Signal Labs
http://www.digitalsignallabs.com

[Audio Empire]

On Thu, 24 May 2012 10:46:32 -0700, Randy Yates wrote
(in article >):

> Audio Empire > writes:
>> [...]
>> Since a good binaural setup has a pair of high-quality omnidirectional mikes
>> mounted in a dummy head which has been designed to mimic the dimensions and
>> the acoustical properties of a skin-covered human head, it probably comes
>> closer to the ideal capture medium than any other microphone technique.
>> However, binaural playback only works through headphones, and the quality of
>> the playback depends, in large measure, on the quality of the headphones.
>>
>> Even so, binaural recordings cannot differentiate between sounds coming from
>> directly in front of the dummy head or directly behind. A simple experiment
>> using a binaural setup will prove this point. While recording, walk around
>> the dummy head jingling a set of keys. Play the recording back through
>> headphones. images from the sides appear as they should (often with shocking
>> realism, especially if one is not used to the effect), but when the keys
>> move
>> to the front or back of the head, they sound as if they are IN the
>> listener's
>> head rather than in front or in back. Our ears don't have this problem and
>> humans can easily tell the direction from which a sound emanates,
>> irrespective of its direction.
>
> This may be due to the concha resonance. As I understand it, one cannot
> simply play back a binaural recording using earphones but must equalize
> for this resonance as well.
>

You may have a point there. If so, that's a big flaw in binaural recording. I
can't imagine that equalizing for this resonance would be easy to do without
extensive auditory measurements being made.

[Arny Krueger[_4_]]

"Audio Empire" > wrote in message
...
> On Thu, 24 May 2012 10:46:32 -0700, Randy Yates wrote
> (in article >):
>
>> Audio Empire > writes:
>>> [...]

>>> Since a good binaural setup has a pair of high-quality omnidirectional
>>> mikes
>>> mounted in a dummy head which has been designed to mimic the dimensions
>>> and
>>> the acoustical properties of a skin-covered human head, it probably
>>> comes
>>> closer to the ideal capture medium than any other microphone technique.
>>> However, binaural playback only works through headphones, and the
>>> quality of
>>> the playback depends, in large measure, on the quality of the
>>> headphones.

>>> Even so, binaural recordings cannot differentiate between sounds coming
>>> from
>>> directly in front of the dummy head or directly behind. A simple
>>> experiment
>>> using a binaural setup will prove this point. While recording, walk
>>> around
>>> the dummy head jingling a set of keys. Play the recording back through
>>> headphones. images from the sides appear as they should (often with
>>> shocking
>>> realism, especially if one is not used to the effect), but when the keys
>>> move
>>> to the front or back of the head, they sound as if they are IN the
>>> listener's
>>> head rather than in front or in back. Our ears don't have this problem
>>> and
>>> humans can easily tell the direction from which a sound emanates,
>>> irrespective of its direction.

I have to admit I've never tried that, even though I've done some binaural
listening. The most impressive demo I've participated in involved a
portable binaural recording system disguised as eyeglasses with real time
monitoring.

>> This may be due to the concha resonance. As I understand it, one cannot
>> simply play back a binaural recording using earphones but must equalize
>> for this resonance as well.

> You may have a point there. If so, that's a big flaw in binaural
> recording. I
> can't imagine that equalizing for this resonance would be easy to do
> without
> extensive auditory measurements being made.

When you properly equalize a set of earphones for flat response
subjectively, any such resonances are dealt with automagically. Of course
you need a competent equalizer - a 4 to 6 band full parametric would be my
tool of choice. Unfortunately I know of no portable digital players with
this feature - most top out with 5 band graphic equalizers which are
amazingly blunt sticks for actually doing things well.