Why evaluating the sound of a single speaker is essential

[Jaxx1138]

That was the experimental condition used to establish the claim. Change the experimental setup and that's new territory.

I am not sure that is germane to discussing how you might conduct an experiment evaluating the NHT loudspeaker which requires some special condition for boundary reinforcement. I conducted quite a few of these tests. There were many tests conducted, including in-wall testing with a rotating wall surface, and tests where the speaker listening axis was adjusted to see if any differences in the subjective rankings were revealed. It did not change with any statistical significant difference.

 

[Snarfie]

Probably a silly question but is there a difference (regarding precieved sound character) evaluating 2 speakers in mono than putting 1 speaker in the middle. If so why.

Funny so i asked Chatgpt with the following answer even I can comprehend


That’s not a silly question at all—it's actually a very insightful one. Here's the clear and detailed answer:

---

Is there a difference between:​

1. Playing mono sound through one speaker centered in front of you, versus
2. Playing mono sound through two speakers (left and right) simultaneously

Yes, there is a difference in perceived sound character, and here’s why:

---

1. Sound Interference and Comb Filtering​

When you play the same mono signal through two speakers, the sound waves from each speaker interact with each other in space before reaching your ears. Depending on:

• Your listening position
• The distance between the speakers
• The acoustics of the room

You’ll get constructive and destructive interference—this leads to a phenomenon called comb filtering, which causes peaks and dips in the frequency response. This can color the sound, especially in the midrange and treble, and may make it seem less natural or more “phasey.”


By contrast, one single speaker in the middle produces a cleaner, interference-free sound field.

---

2. Phantom Center vs. Real Center​

When you use two speakers to play the same mono signal, your brain localizes the sound to a phantom center—an illusion created by equal sound from both sides. But:

• The phantom center image can be unstable depending on your exact position.
• If you move off-center, the illusion collapses and you hear the sound more from one side.

A single speaker placed physically in the center is more stable and consistent in localization, regardless of minor head movements.

---

3. Room Reflections​

Two speakers in a room excite more room modes and early reflections than one speaker. This can subtly affect the tonal character, perceived spaciousness, and clarity.

---

4. Measuring vs. Listening​

From a technical evaluation or measurement standpoint (like what Audio Science Review focuses on), listening with one speaker is preferred:

• It isolates the speaker’s own frequency response and distortion.
• It removes room and psychoacoustic variables introduced by dual-speaker interference.

---

Summary​

Setup	Pros	Cons
One mono speaker centered	Most accurate tonal and distortion evaluation; no interference	Not stereo, no spatial effects
Two speakers playing mono	Simulates typical stereo setup for mono content	Can introduce interference and coloration

 

[Jaxx1138]

That was the experimental condition used to establish the claim. Change the experimental setup and that's new territory.

I should add that it took years to get to the point where Dr Sean Olive and Dr Floyd Toole arrived where these tests got to level of efficacy and proficiency that was achieved at Harman. There were lots of tests and measurements as well as number crunching of large amounts data to determine statistical significant consistency in the results. The book that Dr Toole has put together is a lifetimes worth of work. This research is not a straight line. It looks easy when you come in to see a demonstration, but you are witnessing 20-30 years worth of work by multiple individuals. I was a small part of it. Harman does a lot of research and most of it you will never see or possibly even hear about (It is a corporation for profit after all). There is a lot more going on behind the scenes that the general public does not see. The headphone work, the subwoofer research, surround algorithms and many , many other things. I, Dr Olive and Dr Toole no longer work there. So I have no idea what is currently in the pipe in any of the research areas of work.

 

[SIY]

Harman does a lot of research and most of it you will never see or possibly even hear about

Fine, but again, there is no evidence that doing a different experimental setup will lead to the same results beyond your assertion that it doesn't matter. All we have to go on is the published data.

 

[Jaxx1138]

Fine, but again, there is no evidence that doing a different experimental setup will lead to the same results beyond your assertion that it doesn't matter. All we have to go on is the published data.

I do in fact know that it works as I have conducted test just like this. Not with the NHT specifically but other loudspeakers with non-standard radiation patterns. I conducted well over 500 double blind subjective tests at Harman during my time there. So there is evidence, but much of what was done was not published because it was not part of the primary goal that we were commissioned to do. I hate saying things like this as it appears pompous even to me. I do not recall the specifics on many of these tests as most of them were done over 20 years ago.

 

[Snarfie]

Funny so i asked Chatgpt with the following answer even I can comprehend


That’s not a silly question at all—it's actually a very insightful one. Here's the clear and detailed answer:

---

Is there a difference between:​

1. Playing mono sound through one speaker centered in front of you, versus
2. Playing mono sound through two speakers (left and right) simultaneously

Yes, there is a difference in perceived sound character, and here’s why:

---

1. Sound Interference and Comb Filtering​

When you play the same mono signal through two speakers, the sound waves from each speaker interact with each other in space before reaching your ears. Depending on:

• Your listening position
• The distance between the speakers
• The acoustics of the room

You’ll get constructive and destructive interference—this leads to a phenomenon called comb filtering, which causes peaks and dips in the frequency response. This can color the sound, especially in the midrange and treble, and may make it seem less natural or more “phasey.”


By contrast, one single speaker in the middle produces a cleaner, interference-free sound field.

---

2. Phantom Center vs. Real Center​

When you use two speakers to play the same mono signal, your brain localizes the sound to a phantom center—an illusion created by equal sound from both sides. But:

• The phantom center image can be unstable depending on your exact position.
• If you move off-center, the illusion collapses and you hear the sound more from one side.

A single speaker placed physically in the center is more stable and consistent in localization, regardless of minor head movements.

---

3. Room Reflections​

Two speakers in a room excite more room modes and early reflections than one speaker. This can subtly affect the tonal character, perceived spaciousness, and clarity.

---

4. Measuring vs. Listening​

From a technical evaluation or measurement standpoint (like what Audio Science Review focuses on), listening with one speaker is preferred:

• It isolates the speaker’s own frequency response and distortion.
• It removes room and psychoacoustic variables introduced by dual-speaker interference.

---

Summary​

Setup	Pros	Cons
One mono speaker centered	Most accurate tonal and distortion evaluation; no interference	Not stereo, no spatial effects
Two speakers playing mono	Simulates typical stereo setup for mono content	Can introduce interference and coloration

My follow up question for AI: It even came up with a recomended tracklist to use.

For most average listners making measurments is not possible if you have to bring it back to an average user who want to evelauate a speaker what do you advise.

You're right: most average users can’t make professional measurements or conduct blind tests with rotating platforms. But that doesn’t mean they can’t make trustworthy evaluations. There are simple, practical ways for a normal listener to get close to the truth about a speaker’s performance without needing lab gear.

---

Goal: Give the Average Listener a Reliable Way to Evaluate Speakers​

Here’s what I recommend—organized into steps anyone can follow, using only common sense and basic tools.

---

Step-by-Step Evaluation Guide (for the Average Listener)​

1. Start by Listening to Just One Speaker (Mono)​

• Why: This reveals tonal balance, clarity, and any obvious flaws like boxiness, harshness, or muddiness.
• How:
  • Place the speaker directly in front of you, tweeter at ear height.
  • Sit at a reasonable distance (~1.5 meters).
  • Play mono content (or stereo music with a strong centered vocal).
  • Use tracks with clear vocals, piano, and acoustic instruments.

Listen for: unnatural coloration, exaggerated highs/lows, vocals that sound too recessed or piercing, or general lack of clarity.

---

2. Switch to Stereo Listening in Your Room​

• Set them up in a near-equilateral triangle.
• Toe-in the speakers slightly (aiming them toward your ears).
• Sit centered between them.

Listen for: imaging (can you place instruments?), width, depth, and whether they remain tonally balanced in stereo. If a speaker is too bright or boomy in mono, stereo won’t fix it.

---

3. Use Familiar, Well-Recorded Tracks​

Stick to music you know well and that’s well-produced. Suggestions:

• “Don’t Know Why” – Norah Jones (vocal clarity, natural tone)
• “Hotel California” – Eagles (imaging, detail)
• “Keith Don’t Go” – Nils Lofgren (live feel, transients)
• “Take Five” – Dave Brubeck (instrument placement, cymbals)

Tip: Avoid judging with overly compressed pop or electronic music—it masks speaker flaws.

---

️ 4. Listen at Different Volumes​

• Some speakers sound fine at low volume but fall apart when pushed.
• Others need volume to “wake up.”

Listen for: distortion, harshness, or bass getting muddy at louder levels.

---

5. Compare to a Reference​

If possible:

• Bring in a speaker you already know (even a cheap or old one).
• A/B switching—even manually—is very revealing.

This resets your ear and helps you judge whether what you're hearing is genuinely better or just different.

---

6. Minimize the Room’s Influence​

If your room is very echoey:

• Put a rug between you and the speaker.
• Drape blankets over reflective surfaces temporarily.
• Move speakers away from walls (especially corners).
• Sit at least a meter from the rear wall.

You don’t need a treated studio—but reducing early reflections helps you hear the speaker, not the room.

---

What to Avoid​

• Judging speakers solely on bass: Bigger boom doesn’t mean better.
• Being fooled by volume: Louder often sounds better—use matched volume when comparing.
• Relying on marketing claims or watt ratings: Specs rarely tell you anything about sound quality.

---

Final Recommendation:​

If you're evaluating speakers in a store or during a home trial, ask yourself:

1. Can I listen to just one speaker in mono first?
2. Does it sound natural with vocals and acoustic instruments?
3. Can I pinpoint instrument positions when switching to stereo?
4. Do I feel like turning the music up—or off?
5. Can I listen for an hour without fatigue?

If a speaker passes these tests, it’s very likely a good one, even if you don't have lab measurements to back it up.

---

Would you like me to create a printable version of this checklist for in-store or home use?

 

[SIY]

So there is evidence, but much of what was done was not published because it was not part of the primary goal that we were commissioned to do.

When it goes beyond "trust me," we can have something to talk about. This is not at all meant as a slight to you, but in my profession, if results aren't published with sufficient detail for analysis and replication, they don't exist.

 

[Jaxx1138]

When it goes beyond "trust me," we can have something to talk about. This is not at all meant as a slight to you, but in my profession, if results aren't published with sufficient detail for analysis and replication, they don't exist.

Fair enough but I believe there is enough published data to establish the reliability of tests like this. You could just ask many of the PhD people in the forum at least for their opinion. Or ask Dr Toole himself if he believes that room effects due to boundaries near the loudspeaker under test would negate the subjective test results. But you definitely do not have to trust me.

 

[Heinrich]

Would you like me to create a printable version of this checklist for in-store or home use?

Not for me, now that ChatGpt fulfills its profession to be the parrot for every other chit chat on the world.

Question number two (you may have missed it, I stated number one above on references and memory items):

Given two speakers, one is common bread and butter style, the other is a bit off, but maybe commercially available for pros/:
an 18" bass/midrange and a 60x60 long-throw horn, but a best one, flat, controlled, good bass!

What would we - scientifically - expect from a listening session in mono in a smaller reflective room (given a reference or not) while the test panel is allowed to turn their heads?

Harman didn't address the range of their methodology. Fair enough, we wanted the standard, appreciated.

(Reminder: in science you want "it" to fail, look out for limits of a theory, there's where the beef is.)

 

[Arindal]

Two speakers in a room excite more room modes and early reflections than one speaker.

That is definitely wrong and I would recommend to stop reading A.I. comments on a complicated topic.

Harman does a lot of research and most of it you will never see or possibly even hear about (It is a corporation for profit after all).

Not only Harman. Other manufacturers might do similar things, but they do not talk about their methods of getting to the point which defines aspects setting their products apart from competitors. That is completely legitimate and the standard procedure in this business. Independent institutions do controlled tests as well for different reasons (I can speak from the perspective of procurement issues and standardization between different institutions) and might never publish the results.

The interesting thing is: According to Dr. Toole no-one has replicated the mono-vs.-stereo experiments under controlled conditions with modern equipment and focussing on preference questions. While mono listening tests are widely accepted as the most sensitive method for discrimination tests and used during early stages of speaker design (such as testing drivers), no-one except from Harman and its followers have adopted this method for evaluating products in a stage they go to market and have to succeed in a stereo comparison over their rivals.

That would not be anyhow surprisingly, as different speaker companies use different evaluation methods. What is kind of making me rethink the whole thing, is that Harman was seemingly not very successful with expensive stereo speakers being a product of these R&D methods over the course of the last 25 years. It is even difficult to find a distributor or a dealer to arrange a demo session outside the US, even in markets which are full of hi-fi dealerships filled with products looking like being made by amateurs in comparison. That is surprising given the claim that superior performance in preference tests was the base for the underlying development and tuning methods.

On the other hand, Harman in the same period has been tremendously successful with mono or quasi-monaural speakers. It is safe to say that they have a dominant position with portable bluetooth speakers, and it is not a far stretch to attribute this success to solid R&D and research on consumer preferences. So I would not call it a surprise they seemingly have slowly lost interest in these R&D methods as a base for expensive stereo speaker development focussing on other products they are good in designing, like Dr. Olive was hinting in his statement. I would rather say they have been pretty patient.

Find the error.

 

[SIY]

You could just ask many of the PhD people in the forum at least for their opinion.

I'm one of them. And my opinion is that there's no data presented to validate the claim that the actual published experiment has validity outside of the basic test setup.

 

[Jaxx1138]

That would not be anyhow surprisingly, as different speaker companies use different evaluation methods. What is kind of making me rethink the whole thing, is that Harman was seemingly not very successful with expensive stereo speakers being a product of these R&D methods over the course of the last 25 years. It is even difficult to find a distributor or a dealer to arrange a demo session outside the US, even in markets which are full of hi-fi dealerships filled with products looking like being made by amateurs in comparison. That is surprising given the claim that superior performance in preference tests was the base for the underlying development and tuning methods.

It was a required metric for all consumer products while I was there. All Revel products went through the test and the large horn based JBL units. We sold far more JBL Everest units than the company had predicted. As far as seeing a demo, in these times you can rarely find a hi end dealer showing products that can be demonstrated in person here in North America. Harman is looking to sell as many units as possible. Hi-End never sells the quantity than the much cheaper units do. The meat and potatoes for Harman was the OEM car audio contracts. The Hi-End and pro gave use the brand recognition and prestige that marketing so desired. I think they under-sold the actual research and number of patents that Harman engineers and PhD researchers had acquired over the years.

 

[Jaxx1138]

I'm one of them. And my opinion is that there's no data presented to validate the claim that the actual published experiment has validity outside of the basic test setup.

There is a book based on a number of peer reviewed research papers. But if you need that particular speaker with that specific test conducted just for you, I doubt that most research will satisfy you. But you would not be the first.

 

[SIY]

But if you need that particular speaker with that specific test conducted just for you

That's not at all what my objection was, but you do you.

 

[Newman]

I'm one of them. And my opinion is that there's no data presented to validate the claim that the actual published experiment has validity outside of the basic test setup.

Thanks for your opinion.

Let's move on. There is zero data to support your opinion that speakers of uncommon configuration are not best reviewed in mono, or give relative spatial scores in stereo dramatically different to in mono. Insisting on data to disprove it is argumentative at best, like insisting that science has to disprove every ambit claim by people who just made up an opinion, or who hold some belief. If you have no data to back you up, you have nothing, and you have wasted an awful lot of this thread.

 

[Floyd Toole]

Do you know whether the MGC-1 used cross-correlation in its delayed side-firing arrays? I have tried to find information about this online but have been unable to.

The cross correlation at issue is interaural (between the ears) cross correlation. It is a measure of how different the sounds are at the two ears. Low interaural cross correlation means that the sounds at the ears have a strong random component, as happens in reverberant spaces like concert halls. The result is a perception of spaciousness, envelopment, etc. Pinpoint imaging is not happening. The MGC-1 added laterally reflected sound in amounts determined by the user, with the result that the listener received additional delayed sound and interaural cross correlation would be reduced. It can be considered as an attempt at stereo-to-multichannel upmixing without additional loudspeaker locations. Nowadays, we would add side channel loudspeakers and use a surround processor - hopefully with an appealing up mixer (this is a challenge at the moment). I believe the side loudspeakers would be best located for music at around +/- 60 degrees, but that is likely to be program dependent, and how many movies are being watched

 

[thewas]

For poor recordings EQ or tone controls are always helpful to have, to quote Dr. Toole from another thread:

Neutrality wins, and simple tone controls or equalization can compensate for most problems in recordings or add coloration for personal preferences.

The rest I have the feeling is a bit of obsession of a perfection that doesn't and cannot even exist in first place in stereo reproduction.

 

[Duke]

The MGC-1 added laterally reflected sound in amounts determined by the user, with the result that the listener received additional delayed sound and interaural cross correlation would be reduced. It can be considered as an attempt at stereo-to-multichannel upmixing without additional loudspeaker locations. Nowadays, we would add side channel loudspeakers and use a surround processor - hopefully with an appealing up mixer (this is a challenge at the moment). I believe the side loudspeakers would be best located for music at around +/- 60 degrees, but that is likely to be program dependent, and how many movies are being watched

Thank you! I was wondering whether there was any processing of the sound delivered by the delayed side-firing arrays in the MGC-1, and apparently not (other than bandwidth and delay and level, if those can be called "processing"). My recollection is that the arrays were aimed such that their reflections would arrive from the approximately +/- 60 degree angles (which I recall from your book), but delayed by 20 milliseconds or so.

My guess is that 20 milliseconds delay would reduce the apparent source width (ASW) broadening effect of the first ipsilateral reflections, UNLESS the side-firing array was turned up pretty loud... ?

 

[Floyd Toole]

Probably a silly question but is there a difference (regarding precieved sound character) evaluating 2 speakers in mono than putting 1 speaker in the middle. If so why.

Because double mono centre images involve two signals arriving at the ears, one sound from each loudspeaker, called crosstalk, so all phantom images are comb filtered. A single loudspeaker in mono must sound different, and better, than the corrupted stereo phantom image. This is a dilemma for multichannel recording engineers because we (including me) have adapted to the distorted sound of stereo phantom images.

 

[Floyd Toole]

For about the fifteenth time, this puts both speakers on the far left or far right. This is very different than the setup used to establish the claimed superiority of mono evaluation (speaker at the center). There is no evidence that this new setup will be valid- it might be, but there's no data taking "might be" to "is."

Actually, when we first set up the Harman speaker shuffler we did some comparison tests with the speakers at stereo left and stereo right locations - the marvellous device permitted it. The results were never published, but we were satisfied that the centre location was giving us the same answers and is sounded less "awkward" to the listeners. There would of course be difference in HRTFs if the listener faced forward in both kinds of tests, but since we were looking for resonances - a comparison, not an absolute evaluation - the result was logical. There is some evidence that the brain normalizes differences in HRTFs for sounds arriving from different locations. The challenge for the brain in stereo is that the sound image is in front, but the sounds actually arrive from +/- 30 deg.

EDIT: I will add that the very first experiments done at the NRCC did not involve a shuffler, and as a result were much more labor and time intensive. In those tests the four loudspeakers being compared were arranged across the front of the room and their positions randomly varied during the repeated tests. No loudspeaker was in the centre location - the subjective preferences for timbrally neutral loudspeakers was evident - enough to justify another 50 years of work.