A theoretical model for stereo imaging

[geickmei]

Note from moderator:

This thread was updated to include relevant content from another. Also updated the thread title to more accurately reflect the model's timeline...

The stereo presentation is a function of speaker positioning, radiation pattern, and acoustic qualties of the room. All recordings play against this structure, which is audible but not easily measured. For example, some speakers are very directional and image everything very forward and between the speakers. Some are more multidirectional and are able to recess the soundstage behind the plane of the speakers and wider than the speakers. The strengths of the direct and reflected sounds output by the speakers are what forms the soundstage. There is obviously much more than the direct sound that is responsible for imaging. Please see my article in the current (September) AudioXpress.

 

[RickS]

The stereo presentation is a function of speaker positioning, radiation pattern, and acoustic qualties of the room. All recordings play against this structure, which is audible but not easily measured. For example, some speakers are very directional and image everything very forward and between the speakers. Some are more multidirectional and are able to recess the soundstage behind the plane of the speakers and wider than the speakers. The strengths of the direct and reflected sounds output by the speakers are what forms the soundstage. There is obviously much more than the direct sound that is responsible for imaging. Please see my article in the current (September) AudioXpress.

Would be very interested in your article but can you share as AudioXpress requires a subscription?

 

[geickmei]

Would be very interested in your article but can you share as AudioXpress requires a subscription?

Rick I would not do that. I am calling attention to the article to make everyone aware of this most excellent magazine for anyone interested in audio research and engineering and general practical use. They pay the authors for the articles and hope to sell magazines and keep the process going! And I would be very interested in your take on the article!

 

[RickS]

Rick I would not do that. I am calling attention to the article to make everyone aware of this most excellent magazine for anyone interested in audio research and engineering and general practical use. They pay the authors for the articles and hope to sell magazines and keep the process going! And I would be very interested in your take on the article!

I understand but not much of a discussion if the information is not openly available. Maybe once the next issue is out? Our perhaps you just can share some of the key findings? I already get Voice Coil and their other pub so not paying for AudioXpress.

 

[geickmei]

You can get that issue only for just $9. In my humble opinion, the article is worth a study by every audio interested person - even Amir. I put it here in ASR to call attention to it in hopes of more discussion. It's a deep dive in more ways than one!

Archives - audioXpress Issues (PDFs)

Back Issues of audioXpress magazine in PDF format.Instant download upon purchase. Subscriptions available in Print and Digital formats.Every month, audioXpress combines the best educational articles on topic such audio electronics, speaker and headphone design, amplifiers (from tube to Class D)...

cc-webshop.com

Gary

 

[Duke]

You can get that issue only for just $9. In my humble opinion, the article is worth a study by every audio interested person - even Amir. I put it here in ASR to call attention to it in hopes of more discussion. It's a deep dive in more ways than one!

Archives - audioXpress Issues (PDFs)

Back Issues of audioXpress magazine in PDF format.Instant download upon purchase. Subscriptions available in Print and Digital formats.Every month, audioXpress combines the best educational articles on topic such audio electronics, speaker and headphone design, amplifiers (from tube to Class D)...

cc-webshop.com

Gary

I did not realize single issues could be purchased. Thank you for that information and for the link. I just bought the September 2025 issue so that I can read your article.

 

[amirm]

The stereo presentation is a function of speaker positioning, radiation pattern, and acoustic qualties of the room. All recordings play against this structure, which is audible but not easily measured. For example, some speakers are very directional and image everything very forward and between the speakers. Some are more multidirectional and are able to recess the soundstage behind the plane of the speakers and wider than the speakers.

"Multidirectional" or "very directional" speakers are special cases. If you wanted to know about them, you should have asked. Your comment was general criticism of my testing.

The strengths of the direct and reflected sounds output by the speakers are what forms the soundstage. There is obviously much more than the direct sound that is responsible for imaging. Please see my article in the current (September) AudioXpress.

Direct sound absolutely dominates sound stage. Are you telling that if there is a pan left or right, it is the room and speaker that determine that subjective impression???

Regardless, if you have a case, state it here. Don't send people to an article that cannot be viewed without payment. I do that all the time with research papers I post. You are here to make an argument. So make it.

 

[amirm]

You can get that issue only for just $9. In my humble opinion, the article is worth a study by every audio interested person - even Amir.

So you say. Without telling us what is in it, I have no interest in paying for it. For all I know, it could be what we already know (difference between wide and narrow dispersion) or some nonsense about uncontrolled listening test outcomes.

If your interest is in sharing knowledge, this is the place to do it. Take a lesson from Dr. Toole who routinely shares knowledge that is also in his book.

I am also not a fan of the setup of asking me a question and then answering it with your paid article. It is very uncool.

 

[geickmei]

I did not realize single issues could be purchased. Thank you for that information and for the link. I just bought the September 2025 issue so that I can read your article.

Big Duke! Thanks! Please let me know if I oversold it and what you think.

Gary

 

[geickmei]

"Multidirectional" or "very directional" speakers are special cases. If you wanted to know about them, you should have asked. Your comment was general criticism of my testing.

Direct sound absolutely dominates sound stage. Are you telling that if there is a pan left or right, it is the room and speaker that determine that subjective impression???

Regardless, if you have a case, state it here. Don't send people to an article that cannot be viewed without payment. I do that all the time with research papers I post. You are here to make an argument. So make it.

Huh what? I'm in trouble already? Faster than usual, but inevitable.

I am also not a fan of the setup of asking me a question and then answering it with your paid article. It is very uncool.

Sorry Amir. Just trying to use a probing question to create interest. The article is just like Floyd presents a long article to better explain some point, except that I don't feel it ethical to publish the article that still belongs to the magazine. I summarized it above, but the whole thing is 8 pages long.

I would be honored if you would read it and comment. You apparently still believe that stereo imaging is caused by just the direct sound from the speakers. I can't offer you the article without permission but I can offer you the short

and the longer

video on the same subject. The longer one actually has a little more info in it than the article, but is missing some of the deep dive and extra material that is in the article. It is a combined video from the short one and a BAS presentation I did a while ago. A lot of work went into the graphics and the script etc.

PS I enjoyed your takedown of phase response and hope you can - or have - done one on comb filtering, which I think is also bunk.

Gary

 

[Newman]

...I can offer you the short

and the longer

video on the same subject. ...

Gary

Gary, it is discouraged in ASR to post videos without a summary of its pertinent content. Can you edit please? cheers

 

[geickmei]

Hi Duke,

that article is called "An Image Model Theory for Stereophonic Sound, by Gary Eickmeier" (for the benefit of others reading this).

I'm quite interested in your short-form summary of its claims.

I'm also interested in your comments regarding its rigour. I think that Amir's statements in the second paragraph of post #122 are reflections of what researchers are saying, supported by experimental outcomes that have sufficient rigour to permit such claims.

(I am assuming it is only a 'theory' in the informal sense, ie "I have thought about something and I have a theory", and not a scientific theory. Let me know if you think it qualifies as the latter.)

cheers

Newman, if I had the resources of Harman or B&O behind me I could maybe satisfy your thirst for scientific testing of the theory, but I will have to be able to rely on my lifetime of observation of various speaker architectures and their spatial results instead. I am 82 now and running out of time to present all of these findings and I hope that others can continue on. It's fairly simple stuff - there are very obvious and audible imaging differences among speakers - differences that, as Mark Davis has said, depend mainly on radiation pattern. Maggies and open baffle sound different from box speakers. Bose 901s sounded different from everything else. Arnie Nudell put 12 tweeters on the back of the IRS. Linkwitz asked the AES what are the best radiation pattern, speaker positioning, and acoustics for establishing the AS, or Auditory Scene. My audio club in Michigan, SMWTMS, ran a double blind test for imaging among the Orion, a simple box speaker from Behringer, and one of my designs. I asked to enter because I knew the answers to all 3 questions and I won. That's the best I can offer.

Gary

Gary

 

[Duke]

Big Duke! Thanks! Please let me know if I oversold it and what you think.

Gary

It was definitely worth nine bucks to me.

If I understand correctly (and please feel free to correct or elaborate), your Image Model Theory is based on making deliberate use of the "virtual loudspeakers" create by the first- and second-order horizontal plane front and side wall reflections. The implication for speaker design would be that the speakers should be designed from the outset with these "virtual loudspeakers" in mind, so for this paradigm multi-directional speakers make more sense than conventional forward-firing-only design. And ideally the relative loudness (and possibly other attributes) of the driver arrays dedicated to creating specific "virtual loudspeaker" reflections would be optimized for their task.

The localization cues in the first-arrival sound are preserved largely by minimizing the first ipsilateral reflections, via either radiation pattern control or (if I recall correctly) broadband absorption at the those first sidewall reflection locations. Soundstage broadening and deepening comes from the other, longer-reflection-path "virtual loudspeaker" reflections.

One thing that stood out to me was your emphasis on the desirability of strong "corner secondary" reflections. I've also found these to be beneficial to spaciousness when combined with minimizing the first ipsilateral reflections.

Imo altogether a very innovative, deliberate, and "outside the box" approach to loudspeaker/room interaction, emphasizing the creation of an in-room sounfield with desirable attributes, rather than focusing primarily on the first-arrival sound.

 

[amirm]

I would be honored if you would read it and comment.

I bought and read it. I can't fathom what you are trying to say in there other than borrowing the concepts/research from Dr. Toole's team around ASW (apparent source width).

This is what you end with:

"In Summary
Due to the gradual development of audio history and a need
for continuity and compatibility, we have lost track of the real
problem—a need for an overarching theory for auditory perspective
systems, both head related and field type, carefully delineating the
principles and differences between them for directed development
of each. This has caused confusion between the two at times and
stunted the growth of research into the real problem: sound in
rooms and the relationships between the live and the reproduced
sound events. Field-type systems were largely neglected until
surround sound came along, but I think we still need a clean break
from two channels to reemphasize the difference from binaural."

I don't know what this rant has to do with anything. No one is confused about binaural recordings and what stereo tries to do.

"Stereo is a reconstruction of the acoustic event that was
recorded, within your listening room, using loudspeakers and
the reflecting surfaces around them."

That only holds for simplistic recordings of a live event. Much of the music we listen to is not an acoustic event but a composition created for enjoyment. The purpose of the playback system is to be true to this creative art, with little to nothing to do with any original live event.

"Due to the limitation of the number of channels available—two—
the stereo effect has been limited to a simple spatial broadening of
the size of the image field, instead of something more like the
field contained in the recording, which is larger, deeper, and more
spacious. With additional channels the spatial resolution can be
increased dramatically as well (enter surround sound!)."

That's right although you are just repeating what we already know. You have to remember that such effects in room are subtle. From my article that I wrote 10 to 15 years ago, quoting Dr. Toole's actual experimentation: https://www.audiosciencereview.com/...ds/perceptual-effects-of-room-reflections.13/

"“It was in this room [Dr. Toole’s Reference IEC room at National Research Council] that experience was gained in understanding the role of first reflections from the side walls. The drapes were on tracks, permitting them to easily be brought forward toward the listening area so listeners could compare impressions with natural and attenuated lateral reflections (see Figures 4.10a and 8.8). In stereo listening, the effect would be considered by most as being subtle, but to the extent that there was a preference in terms of sound and imaging quality, the votes favored having the side walls left in a reflective state. In mono listening, the voting definitely favored having the side walls reflective.""

I have no idea what you are doing by using a graph paper, and drawing circles around speakers to show the ASW effect. So what? How is this important in speaker measurements which you asked me about? You propose no objective metric whatsoever.

And you say all of this was motivated by the design of Bose 901???

 

[amirm]

You apparently still believe that stereo imaging is caused by just the direct sound from the speakers.

Not at all. Here is what I wrote you when you first asked:

Spatial effects by far are dominated by the content itself which is outside of the scope of audio gear review. Next in line is location and configuration of speakers in a room, not the speakers themselves. This aspect is not useful to comment on either as how I locate my speakers and the room I have is unique to my situation and unlikely to apply to anyone else.

Of course I can make up stuff about the soundstage as just about every reviewer does. But that is of zero value as it has no defensible foundation.

There is one aspect that is a direct function of the speaker and that is the off-axis response of the speaker. This is part and parcel of my reviews so there is coverage there.

As I have bolded above, I started with the content itself as the top contributor. I then went on to talk about the role that speaker placement has and interactions with the room. And finishing with off-axis response.

If you had told me you were a member of the industry, and had specific questions about all of this, i would have explained it all, including the obvious role ASW plays.

And no, I don't cover dipoles, Bose 901, etc. when someone is just asking me a generic question. Those speakers are in a world onto themselves.

 

[Newman]

Newman, if I had the resources of Harman or B&O behind me I could maybe satisfy your thirst for scientific testing of the theory, but I will have to be able to rely on my lifetime of observation of various speaker architectures and their spatial results instead. I am 82 now and running out of time to present all of these findings and I hope that others can continue on. It's fairly simple stuff - there are very obvious and audible imaging differences among speakers - differences that, as Mark Davis has said, depend mainly on radiation pattern. Maggies and open baffle sound different from box speakers. Bose 901s sounded different from everything else. Arnie Nudell put 12 tweeters on the back of the IRS. Linkwitz asked the AES what are the best radiation pattern, speaker positioning, and acoustics for establishing the AS, or Auditory Scene. My audio club in Michigan, SMWTMS, ran a double blind test for imaging among the Orion, a simple box speaker from Behringer, and one of my designs. I asked to enter because I knew the answers to all 3 questions and I won. That's the best I can offer.

Gary

Thanks Gary.

In your shorter video (the first one) you seem to be verbally concluding that flat sound power response is the key target for excellence. One of your key takeaways on your concluding slide is that we need to focus on power response not axial response, point #2 below.

However, Olive actually tested this theory and found it wanting. Toole relates the story of how Consumer Reports Magazine had the longstanding practice of rating the sound quality of stereo speakers on their sound power response, similar to your theory. Toole and Olive had been conducting controlled listening tests of loudspeakers and were identifying other factors as more important to listener preference. So they subjected numerous speakers that Consumer Reports had recently reviewed to their controlled double blind tests, to see how well the sound power correlated with preference.

© Toole, Sound Reproduction, 1st ed.

They found a slight negative correlation, and no statistical significance, linking sound power response to listener preference.

Since seeing that, I have tended to think of sound power as a useless indicator of subjective loudspeaker sound quality.

cheers

 

[geickmei]

It was definitely worth nine bucks to me.

If I understand correctly (and please feel free to correct or elaborate), your Image Model Theory is based on making deliberate use of the "virtual loudspeakers" create by the first- and second-order horizontal plane front and side wall reflections. The implication for speaker design would be that the speakers should be designed from the outset with these "virtual loudspeakers" in mind, so for this paradigm multi-directional speakers make more sense than conventional forward-firing-only design. And ideally the relative loudness (and possibly other attributes) of the driver arrays dedicated to creating specific "virtual loudspeaker" reflections would be optimized for their task.

The localization cues in the first-arrival sound are preserved largely by minimizing the first ipsilateral reflections, via either radiation pattern control or (if I recall correctly) broadband absorption at the those first sidewall reflection locations. Soundstage broadening and deepening comes from the other, longer-reflection-path "virtual loudspeaker" reflections.

One thing that stood out to me was your emphasis on the desirability of strong "corner secondary" reflections. I've also found these to be beneficial to spaciousness when combined with minimizing the first ipsilateral reflections.

Imo altogether a very innovative, deliberate, and "outside the box" approach to loudspeaker/room interaction, emphasizing the creation of an in-room sounfield with desirable attributes, rather than focusing primarily on the first-arrival sound.

Duke -

I think I love you! You said it better than I did. The main concept is that the "accuracy" model of thinking that stereo is made by the first arrival axial output of the speakers is wrong and is confusing the system with binaural. This has led us to making these directional speakers for the entire stereo era because of the thinking that it is two channels for two ears and that is that. Bust out of that kind of thinking and you are free to design speakers for radiation pattern and systems for multichannel and surround sound. It illustrates why wall speakers would be wrong and corner horns would have no spatial qualities.

The bit about the first sidewall reflections I discovered when building some prototypes, when I thought that those sidewall reflections would be great for broadening the soundstage. I had the speakers with greater output to the front/sides and soon discovered that the imaging was a great disappointment. As you move to one side the whole soundstage collapses to that side. So I interchanged the left and right to get a greater output to the front/inside, and the imaging was so good I found myself getting emotional about it. I could walk anywhere in the room, even up to and between the speakers, and it remained stable everywhere. The rear output was already greater by about 6 dB. My current thinking is that this problem is mainly because we have but two channels to work with, and each speaker is pulling fully half of the orchestra to one side if you get it wrong. Hence, Davis's time/intensity trading for the Soundfield One. He could have had IMT projectors if he just put more gain toward the front corners. I heard his Soundfield Ones at his home in Pacifica when I lived in California and we discussed it, but it was too late by then.

THANKS for the comments!

Gary

 

[geickmei]

Thanks Gary.

In your shorter video (the first one) you seem to be verbally concluding that flat sound power response is the key target for excellence. One of your key takeaways on your concluding slide is that we need to focus on power response not axial response, point #2 below.

View attachment 475454

However, Olive actually tested this theory and found it wanting. Toole relates the story of how Consumer Reports Magazine had the longstanding practice of rating the sound quality of stereo speakers on their sound power response, similar to your theory. Toole and Olive had been conducting controlled listening tests of loudspeakers and were identifying other factors as more important to listener preference. So they subjected numerous speakers that Consumer Reports had recently reviewed to their controlled double blind tests, to see how well the sound power correlated with preference.

View attachment 475457 © Toole, Sound Reproduction, 1st ed.

They found a slight negative correlation, and no statistical significance, linking sound power response to listener preference.

Since seeing that, I have tended to think of sound power as a useless indicator of subjective loudspeaker sound quality.

cheers

Newman,

I will have to take a closer look at that one, but I had been arguing with Floyd for a few years in the pages of the Journal about his test methods and especially the single speaker idea. We remain friends and correspond periodically, but all of his famous work was done before IMT. Speakers that are designed for axial response would probably just naturally have poor power response. The basic theory reason is that what you want is the response in the concert hall at about the critical distance which is where the main mikes should be placed and where we are listening in most systems. Actually, both should be flat or near the hall curve, not just the axial or the power. But notice that all of the room EQ systems place the mikes at the listening positions and EQ toward this hall curve that you would hear at that position live. I think Floyd's 4th edition is going to emphasize surround sound and declare stereo a dead format.

Gary

 

[Duke]

However, Olive actually tested this theory [that sound power is what matters most] and found it wanting. Toole relates the story of how Consumer Reports Magazine had the longstanding practice of rating the sound quality of stereo speakers on their sound power response, similar to your theory. Toole and Olive had been conducting controlled listening tests of loudspeakers and were identifying other factors as more important to listener preference. So they subjected numerous speakers that Consumer Reports had recently reviewed to their controlled double blind tests, to see how well the sound power correlated with preference.

View attachment 475457 © Toole, Sound Reproduction, 1st ed.

They found a slight negative correlation, and no statistical significance, linking sound power response to listener preference.

Since seeing that, I have tended to think of sound power as a useless indicator of subjective loudspeaker sound quality.

Imo it depends on the speaker's radiation pattern.

I agree that the power response is the wrong metric for evaluating a speaker which radiates forward only. But imo the power response is a relevant metric for a deliberately multi-directional speaker because it's giving you information about the spectral balance of those deliberate reflections.

(Imo for a multidirectional speaker one should look at the spectral balance of the direct sound and the reflected sound separately, and imo they should be very similar to one another.)

I'm also interested in your comments regarding its rigour. I think that Amir's statements in the second paragraph of post #122 are reflections of what researchers are saying, supported by experimental outcomes that have sufficient rigour to permit such claims.

(I am assuming it is only a 'theory' in the informal sense, ie "I have thought about something and I have a theory", and not a scientific theory. Let me know if you think it qualifies as the latter.)

Yes, I would call it an "informal theory" rather than a "scientific theory". So, not unlike the informal theory I have. And there are similarities; for example, both informal theories de-emphasize the first sidewall reflections and emphasize the later-onset reflections. I'd say that Gary goes further in that general direction than I do.

 

[geickmei]

With such questions about the reproduction I always look to the live sound fields for the answer. What are all of the acoustic qualities of the live sound that we can refer to for the answer. CERTAINLY live sound does not come from two points in space spouting out a narrowly dispersed stream of direct sound toward your ears. That would be absurd, but here we are in 2025 thinking just that.

No, what the situation is with the live event is two or more sources with a certain mix of direct and reflected sounds, some early and most later reverberant field sounds. These direct sounds that were recorded should be presented from the two or three direct field speakers but should we expect the early and late reflected sounds to come from the same points in the reproduction? Of course not. They should come from a multiplicity of incident angles, just like the original sounds did. We don't have enough channels to present all sounds from their correct directions but the best we can do is force direct and reflected to come from a decent mix of directions in general for most recordings. So we make the radiation pattern some direct and some reflected from realistic directions and this pattern, along with sufficient delay of the reflected, forms a more realistic sound field within the playback room. This is NOT additional reverberation or "echoes" because it all happens within the fusion time. It is strictly a spatial broadening effect caused by a summing localization between the speakers and the walls around them. We are literally shaping the sound fields in front of us with the speakers and their early reflections and first arrival sounds remain so and later recorded sounds are presented from more correct directions and the recorded timings, as desired.

This sort of analysis is what Bose went through for some 11 years before the introduction of the 901 speaker and he wrote several papers on how they arrived at those conclusions but we don't seem to be able to learn from each other and we all have to go through it on our own over and over again.

And again, the biggest obstacle has not been "but we hated the sound of the 901s" but rather the mistaken stereo theory that it is about direct sensory input to the ears from the two stereo channels that we can't seem to shake. That has caused the strong axial response of most speakers with poor power response and too strong direct to reflected ratio and reviewers and designers unaware of the spatial response of speakers in rooms.

Gary