It’s All a Preference

[Soundminded]

Hello Soundminded

It's not for corner placement. One is a folded corner horn the other a bass reflex cabinet. They are apples and oranges and the K2 is leaps and bounds the better loudspeaker of the 2.

Rob

Building a loudspeaker system with a FR extended to 40 or 50 khz is not science, it's advertising. I've never thought much of horn speakers as high fidelity to begin with.

 

[Robh3606]

Building a loudspeaker system with a FR extended to 40 or 50 khz is not science, it's advertising. I've never thought much of horn speakers as high fidelity to begin with.

Well I can understand that on both points. I have 045Be's in my Array Horns. They go out to 50K but it's not about that they go out to 50K. It's more what they sound like from 8K up to 20K . They make damn nice sounding tweeters where it counts.

As far as horns are concerned too many people have formed opinions based on older designs from the 50's to give modern horns a serious listen. That's too bad for them. A modern horn simply doesn't sound like a horn. It doesn't have any of the obvious colorations that plagued the older horns.

Rob

 

[amirm]

Speaking about disagreement, your posting of the curve showing the coefficient of absorption as a function of frequency is extreme but typical for most materials, that is it increases with increasing frequency. But you also said that the first reflection of one speaker was very flat to 17 khz. I presume this was the reflection from the front wall, the wall behind the speaker and was the result of a rear firing tweeter as would be the case with the original Revel Salon Ultima. (am I right about that?)

No . The specs I gave are for a far more modest Revel speakers, namely the C52 center speaker:

I used that because I just bought 7 of them for my new theater! Yeh, it is strange to be using all center speakers but Harman made a bunch of these in maple and then realized this is not the "in" color so they decided to not further manufacture them. They provided a great promotional price to clear out what they had built. I am talking about all seven of them costing less than the seating for my theater!

As to your question regarding how it is spec'ed, here it is from the manual:

"In-Room Response • ±0.5dB from 70Hz to 18kHz
Indicates sound quality in context with other specifications.
A breakthrough measurement, this specification closely
correlates to sound quality in a single curve—a long-standing
goal of loudspeaker engineers.

In-room response is measured through the use of large
anechoic chambers. The loudspeaker’s response is measured
every 10 degrees, horizontally and vertically, for a total of
72 response measurements.

The in-room response curve is a prediction of how the
loudspeaker would measure in a typical room. Research
and observation reveals that ubiquitous on-axis response
curves cannot distinguish between two loudspeakers with
radically different sound qualities.

First Reflections Response
• ±0.5dB from 70Hz to 17kHz
Indicates the response listeners hear in relation to the first
reflections from walls, ceilings, and floors. This specification
indicates that Revel loudspeakers will remain accurate,
even in listening rooms that cast strong reflections.

Listening Window Response
• ±1.0dB from 65Hz to 18kHz
Indicates the on-axis response of the loudspeaker. An
improved on-axis measurement, this specification reduces
the visual confusion of inaudible interference. It retains full
accuracy without using “spectral smoothing,” which results
in significant data loss"

Do you think this was just an accident or do you think the speaker's output was carefully tailored to that room to give that result? What do you suppose would have happened to the FR of the reflection in a different room with a different wall material and covering (even the type of paint, water based latex most commonly used in US homes versus say oil base paint makes a difference in reflectivity.)

I see where the confusion comes from. This data is anechoic chamber measurements. They are not measured in rooms so are not subject to variations of the first reflection points. The naming here, first reflection, refers to off-axis response less than certain number of degrees.

And yes, you absolutely can screw up this flat response with improper room treatment. Put fabric on all wall surfaces and the response will take a nose dive at high frequencies.

I'm also curious to know what you think about why the rear firing tweeter was removed from the second version of Salon Ultima after Toole retired.

I asked Kevin Voecks about this. He said that the original design of the Salon 2 like the previous generation had a rear firing tweeter. But they found the new improved front firing tweeter to be so good that the rear one was no longer necessary to achieve good off-axis response. So they eliminated it as they went from prototype to production. Previous tweeters would get directional in as low as 8 Khz. Not so with the new Be tweeter.

 

[Phelonious Ponk]

Tim,

It could be any other respected brand and the reaction would be the same. However when the referred brand is well known for providing us with excellent sound quality it is natural that we should question the methodology used for these tests. The information available is scarce, diffuse and the parameters evaluated are not universally accepted. There were references to mono listening, using a DAC driving amplifiers directly, omission of details such as cables and correct setup of speakers. IMHO all this strongly influenced the results. May be Harman people considered that this is the typical use of their speakers and wanted to use these conditions to test preferences. Perfect. But in no way this a typical audiophile setup.

I agree with much of that. Many respected audiophile brands could have created the same reaction (though B&W didn't). And you're right, the set up is not the typical audiophile set up. But let's get the test conditions, even the measurements part of the test out of the discussion for a moment by ignoring the on-axis response figures. The rapid drop-off of off-axis response from Martin Logan panels should come as no surprise to anyone who has heard them, or even read a brochure for them. They've made it a copy point. You can't miss it it a room with them; any room. Now you can buy into the concept of "controlled directivity." ML is not the only brand doing it. But the premise of this study was that smooth, even, off axis response would be preferred by listeners. It was. It's really that simple. Treat the room well enough, stop those first reflections from mixing with the direct sound and I'll bet the Martin Logan's scores would come up. I'd also bet that all the speakers would sound worse, but that's another study.

Tim

 

[Phelonious Ponk]

IMHO their scientists were much more interested in extracting one parameter from the 70 measurements that could be correlated with the preferences shown in the listening tests than in comparisons with the competition. This was the real challenge. Unhappily the marketing guys spoiled it all.

What marketing guys? Show me one JBL, Infinity or Revel ad that quotes this research and this rises to the level of speculation.

Tim

 

[MylesBAstor]

I agree with much of that. Many respected audiophile brands could have created the same reaction (though B&W didn't). And you're right, the set up is not the typical audiophile set up. But let's get the test conditions, even the measurements part of the test out of the discussion for a moment by ignoring the on-axis response figures. The rapid drop-off of off-axis response from Martin Logan panels should come as no surprise to anyone who has heard them, or even read a brochure for them. They've made it a copy point. You can't miss it it a room with them; any room. Now you can buy into the concept of "controlled directivity." ML is not the only brand doing it. But the premise of this study was that smooth, even, off axis response would be preferred by listeners. It was. It's really that simple. Treat the room well enough, stop those first reflections from mixing with the direct sound and I'll bet the Martin Logan's scores would come up. I'd also bet that all the speakers would sound worse, but that's another study.

Tim

And I'll bet you that ML sound just quite wonderful the way they are. You really don't know what you're talking about. Until you live with a product, you know nothing. You've obviously never heard Summit-Xs set up properly with the right electronics or you wouldn't be saying what you're saying. Til then, you can speculate til the cows come home but it's laughable.

 

[Phelonious Ponk]

And I'll bet you that ML sound just quite wonderful the way they are. You really don't know what you're talking about. Until you live with a product, you know nothing. You've obviously never heard Summit-Xs set up properly with the right electronics or you wouldn't be saying what you're saying. Til then, you can speculate til the cows come home but it's laughable.

So Martin Logan's pitch for the benefit of narrow dispersion is a lie? How about Klipsch and Geddes? Are they lying about this feature as well?

I didn't say they sound bad, Myles. I said their response falls off rapidly off-axis. Martin Logan says the same thing. This is not speculation. This is their speaker design, as represented in their own communications. I've read their materials and their web site (these are things you do when charged with selling their speakers) and not once do they say that if properly set up, with the right electronics, one of their primary selling propositions, narrow, controlled dispersion, goes away.

The purpose of this test was to find out if the opposite - broad, even off-axis dispersion, was preferred by listeners. It was. That doesn't mean that some people, like you, won't prefer ML's approach and be happy. I had some customers that loved them. But that doesn't change ML panels' off-axis response or the listeners response to them in these tests.

Tim

 

[microstrip]

You are missing the point on this. This premise of the FR dominating the preference in blind testing was there before Toole was employed by Harmon. You asked if anyone has both read his book and his prior papers I have. Have you??

You know maybe you are right maybe it does. But for sure it certainly removes any sighted bias and bias attained through reading what's in Stereophile simply because it's a blind test. Surely you can see how what you read by a respected reviewer could influence you??

Rob

Rob,

Yes, I own and read the book and several of F. Toole papers. Otherwise I would immediately jump in the wagon that claims all this is Harman marketing!

The book is excellent, but as Amir already said, F. Toole and S. Olive realized their tests considering that differences between electronics, cables and digital sources are too small to influence the results. F. Toole spends a few lines on it in his book, referring to exactly the same papers almost all audiophiles do not consider acceptable - such as the old Douglas Self tests and the famous Moran and Meyer.

Also, in their care to reproduce typical listening conditions and fast switching IMHO they compromised one audiophile key point - optimization of speaker placement. When I read that all speakers were placed at the same place, to be sure they all are evaluated in equality, or evaluated in mono I do not think that this did not bias the results significantly. Surely some others will disagree.

 

[microstrip]

What marketing guys? Show me one JBL, Infinity or Revel ad that quotes this research and this rises to the level of speculation.

Tim

Tim,
Sorry, I will not enter this battle - it will only introduce more noise in this debate. Any one interested can use google to find it. Or look for Harman press releases.

 

[terryj]

IMHO their scientists were much more interested in extracting one parameter from the 70 measurements that could be correlated with the preferences shown in the listening tests than in comparisons with the competition. This was the real challenge. Unhappily the marketing guys spoiled it all.

I resisted this initially....

I think there is a fundamental misunderstanding here. It is ONE parameter (FR), however it was not a single on axis measurement. It was seventy or so frequency measurements in many different polar positions.....

In other words, NOT seventy different parameters from which a single parameter (FR) was 'chosen'.

Whilst I am here, it might be worthwhile following up on a post from Rob. I alluded to this, that FR can also be an excellent indicator of time.

Ok, not sure how to bold or copy and past from this! but in any case, read about 'resonances'. They do show up clearly on a FR response, and we can see that this response is hardly free from these 'time based' errors.

It is simple to create ad hoc answers or rebuttals, only too often to turn out to not be based in reality.

I keep trying to hone in on the 'essential' reasons for the outright denial of these results and conclusions (a ninety percent success rate in prediction is not to be sneezed at...try that at the track or the stock market. But this is audio remember)

Is it as simple as some (many?) needing audio to be indefinable? The second we begin to isolate any sort of predictive (scientific in other words) power the magic is lost?? Is it as simple as that?

IF audio is ultimately all preference, that there is NO transference from one person to another then what value the audio press? What earthly use is any sort of audio review?

The measurements do not indicate anything useful, our ears are 'ours' and no one elses, we can not determine anything until we experience it ourselves, what use then for audio magazines?

And, who is more likely to read them. Not me for one...I'd hazard the people most likely to read them are the ones who argue that measurements mean little, have no predictive power and all taste is preference.

Precisely the arguments against subjective press.

 

[Soundminded]

No . The specs I gave are for a far more modest Revel speakers, namely the C52 center speaker:

I used that because I just bought 7 of them for my new theater! Yeh, it is strange to be using all center speakers but Harman made a bunch of these in maple and then realized this is not the "in" color so they decided to not further manufacture them. They provided a great promotional price to clear out what they had built. I am talking about all seven of them costing less than the seating for my theater!

As to your question regarding how it is spec'ed, here it is from the manual:

"In-Room Response • ±0.5dB from 70Hz to 18kHz
Indicates sound quality in context with other specifications.
A breakthrough measurement, this specification closely
correlates to sound quality in a single curve—a long-standing
goal of loudspeaker engineers.

In-room response is measured through the use of large
anechoic chambers. The loudspeaker’s response is measured
every 10 degrees, horizontally and vertically, for a total of
72 response measurements.

The in-room response curve is a prediction of how the
loudspeaker would measure in a typical room. Research
and observation reveals that ubiquitous on-axis response
curves cannot distinguish between two loudspeakers with
radically different sound qualities.

First Reflections Response
• ±0.5dB from 70Hz to 17kHz
Indicates the response listeners hear in relation to the first
reflections from walls, ceilings, and floors. This specification
indicates that Revel loudspeakers will remain accurate,
even in listening rooms that cast strong reflections.

Listening Window Response
• ±1.0dB from 65Hz to 18kHz
Indicates the on-axis response of the loudspeaker. An
improved on-axis measurement, this specification reduces
the visual confusion of inaudible interference. It retains full
accuracy without using “spectral smoothing,” which results
in significant data loss"


I see where the confusion comes from. This data is anechoic chamber measurements. They are not measured in rooms so are not subject to variations of the first reflection points. The naming here, first reflection, refers to off-axis response less than certain number of degrees.

And yes, you absolutely can screw up this flat response with improper room treatment. Put fabric on all wall surfaces and the response will take a nose dive at high frequencies.


I asked Kevin Voecks about this. He said that the original design of the Salon 2 like the previous generation had a rear firing tweeter. But they found the new improved front firing tweeter to be so good that the rear one was no longer necessary to achieve good off-axis response. So they eliminated it as they went from prototype to production. Previous tweeters would get directional in as low as 8 Khz. Not so with the new Be tweeter.

The FR of the first reflection looks excellent....too excellent....suspiciously excellent. This begs the question how was it done. There's more to this than meets the eye so a lot more must be asked and answered about it to take it at face value. I don't care how good the new tweeter is, it's not omnidirectional, its output very far off axis even frontally will fall significantly with frequency compared to it on axis response and its output to the rear at 15 khz will be next to nil. This means that the first reflection could not have come from the front or side walls. Clearly there is no way to prevent radiation to the rear and sides above the cutoff frequency of 200 hz. This can easily be confirmed by taking the speaker out of doors and walking around it noticing the change in tonal balance with program material having significant high frequency content. The next logical conclusion is that the first reflection comes from the wall behind the listener. The next question is what is its material. Since reflectance of all materials is a function of frequency and each one is different, the material must have been selected to have as uniform a spectral reflectance as possible. This data for different materials can probably be found from ASTM and ASA. It's important for among other things design of noise control such as in open landscape offices. Was it something very hard like concrete, brick, or thick glass? What is the raw data of the 70 output measurements of FR as a function of solid angle. That would tell us a lot more about how the speaker really performs and we can draw our own conclusions about what the consequences would be in a particular room.

This is consistent with Harman Kardon's advertising style in the 1960s. For example, the A1000 solid state integrated amplifier was advertised as flat from DC to Light. Another ad for it had FR flat from DC to over 1 mhz. At what, one watt? Even my A500 tube amp was advertised as flat to 70 khz. Pardon me if I don't draw the conclusions Harman wants me to, this raised a lot more questions about the test methods than it answers about the speaker performace for me. I don't see anything about the design of this speaker that can control the FR of room reflections any better than other speakers and the geometry of the tweeter which in part determines its geometric propagational pattern doesn't look to me like it's even close to AR's 3/4" model or Roy Allison's. So what else can you tell us about how the measurements were conducted and what ALL of the raw data for these speakers show?

 

[Phelonious Ponk]

Is it as simple as some (many?) needing audio to be indefinable? The second we begin to isolate any sort of predictive (scientific in other words) power the magic is lost?? Is it as simple as that?

I think that's a big part of it. The rest, I think, is strenuous denial of the poor performance of a respected audiophile brand. The thing that makes it downright surreal is what these measurements say is the problem with Martin Logans -- a steep drop-off of response off axis -- is integral to Martin Logan's design and a featured benefit of their marketing pitch. They call it "Controlled Dispersion." Go to their site and search the term. Klipsch calls it "Controlled Directivity," I think. Earl Geddes just calls a better way to build a speaker, as far as I know, but they all tout low off-axis ouput as a benefit. And if you look at the measurements in this study, there it is. The idea, according to ML, is to limit the output off axis so much that it virtually eliminates sidewall reflections. But the participants in the listening test that followed the measurements evidently heard those reflections, and didn't like them.

Or perhaps they heard something else altogether, something that's not in those measurements at all. Maybe the direct line the Canadian research group and Harman have drawn, over three decades of research and product development, between smooth, even, off-axis response and preference is all wrong and it's something else that people are preferring. Anything is possible.

Tim

 

[Robh3606]

Also, in their care to reproduce typical listening conditions and fast switching IMHO they compromised one audiophile key point - optimization of speaker placement. When I read that all speakers were placed at the same place, to be sure they all are evaluated in equality, or evaluated in mono I do not think that this did not bias the results significantly. Surely some others will disagree.

Hello microstrip

I can understand where you are coming from but I don't agree and here is why. If you were in a typical showroom the speakers obviuosly could not occupy the same physical space. We all know how critical placement can be for smooth LF response where the speakers is in the modal region below the transition zone.

So with a pair of speakers in that showroom you have different placement of the pairs at the mercy of the room. One pair could be in the best spot the other won't be. If one is pair is placed where they are driving modal peaks it will be immediately obvious when you switch between the pairs. How it goes as far as preference would be determined by how the listener "hears" the peak. One pair could be heard as bass shy compared to the peaked pair or the peaked pair as boomy.

Obviously under these conditions one pair is not getting a fair evaluation.

This is where the shuffler comes in. It places the speakers in the same physical location which removes one more variable for the evaluation process. Now you could argue that it's not optimum placement and frankly I don't see how anyone could not see that and I see your point. That said both speakers will be affected in a similar fashion by their placement in the room. At least under these conditions both speakers get the the same peaks and suckout's from any modal effects due to placement.

The flip side is they are doing mono evaluations so there is a very good chance they are using the room sweet spot to do this. That's just common sense and not something they would be likely to overlook. Let face it we don't deliberately put our speakers in areas where we get the worst response so why would they??

Nothing is ever going to be perfect. The best you can do is remove as many variables as you can based on the current and most up to date understanding of acoustics and all the rest.

Rob

 

[Phelonious Ponk]

And the whole point of speaker placement is almost moot given the circumstances. Not to beat the same point over and over again, but I think it is the critical point:

The purpose of the measurements was to understand the linearity on and off-axis FR of the speakers in the test.

The purpose of the listening tests was to discover any correlation between off-axis FR and preference.

Martin Logans have extremely uneven off-axis response by design. This doesn't change with speaker placement.

Is there something else, something exceedingly wonderful that Martin Logans do that may have radically altered the results of the preference testing had they been placed differently? Maybe. That's another hypothesis, another test.

Tim

 

[Soundminded]

Hello microstrip

I can understand where you are coming from but I don't agree and here is why. If you were in a typical showroom the speakers obviuosly could not occupy the same physical space. We all know how critical placement can be for smooth LF response where the speakers is in the modal region below the transition zone.

So with a pair of speakers in that showroom you have different placement of the pairs at the mercy of the room. One pair could be in the best spot the other won't be. If one is pair is placed where they are driving modal peaks it will be immediately obvious when you switch between the pairs. How it goes as far as preference would be determined by how the listener "hears" the peak. One pair could be heard as bass shy compared to the peaked pair or the peaked pair as boomy.

Obviously under these conditions one pair is not getting a fair evaluation.

This is where the shuffler comes in. It places the speakers in the same physical location which removes one more variable for the evaluation process. Now you could argue that it's not optimum placement and frankly I don't see how anyone could not see that and I see your point. That said both speakers will be affected in a similar fashion by their placement in the room. At least under these conditions both speakers get the the same peaks and suckout's from any modal effects due to placement.

The flip side is they are doing mono evaluations so there is a very good chance they are using the room sweet spot to do this. That's just common sense and not something they would be likely to overlook. Let face it we don't deliberately put our speakers in areas where we get the worst response so why would they??

Nothing is ever going to be perfect. The best you can do is remove as many variables as you can based on the current and most up to date understanding of acoustics and all the rest.

Rob

"This is where the shuffler comes in."

In Toole's room with Toole's recordings. What about in a different room with different recordings. Entirely different and opposite conclusions might be drawn. Toole no more addresses the variables of recordings or room acoustics than anyone else does as far as I can tell.

BTW, constant directivity speakers do have a very useful application...in designing sound reinforcement systems, their original and legitimate purpose. In fact it is so useful that computer programs can reliably design systems for you with given data about the speakers, the space, and the SPL and frequency range requirements to maximize gain before feedback. It will give you the number of speakers, their precise placement, and the required amplifier power. As for use in home sound reproducing systems, I don't see it.

 

[microstrip]

Hello microstrip

I can understand where you are coming from but I don't agree and here is why. If you were in a typical showroom the speakers obviuosly could not occupy the same physical space. We all know how critical placement can be for smooth LF response where the speakers is in the modal region below the transition zone.
(...)

Nothing is ever going to be perfect. The best you can do is remove as many variables as you can based on the current and most up to date understanding of acoustics and all the rest.

Rob

Rob,

I am not alone. See the abstract of this Sean Olive article. Would love to read it, but it is not accessible for free.


The Effects of Loudspeaker Placement on Listener Preference Ratings

Through the use of an acoustically adjustable listening room and a binaural recording and reproduction system, live and binaural subjective evaluations were made of different loudspeakers placed in different room locations. The experimental results from both tests show that listener preference ratings for different loudspeakers are significantly influenced by the loudspeaker location within the room. In fact, the positional effects can be larger than the subjective differences between the loudspeakers themselves. The binaural evaluations indicate that listener preferences are influenced significantly by interactions between the loudspeaker, its location, and the type of program material auditioned. These secondary effects were less significant in the live tests, suggesting that traditional real-time listening tests may be inadequate for measuring or controlling these effects.

Authors: Olive, Sean E.; Schuck, Peter L.; Sally, Sharon L.; Bonneville, Marc E.
Affiliations: Institute for Microstructural Sciences, National Research Council, Ottawa, Ont., Canada ; Canadian Audio Research Consortium, Scarborough, Ont., Canada(See document for exact affiliation information.)
JAES Volume 42 Issue 9 pp. 651-669; September 1994

I am not saying that the Harman people did not think about it , but I would like evidence that each speaker position was optimized.

 

[Soundminded]

So far all I've heard about is that there are 70 anechoic measurements of each speaker I haven't seen, Harman's speakers sound best in its own room with its own program material. I have no idea how the "golden ears" panel was selected or what qualifications other than training Harman provided they have for making critical judgments. Their source material so far seems to be limited to James Taylor and test tones deciding which frequency band(s) has been equalized from one instance to the next. I'm really not sure about the reliability of the science here in drawing any valid conclusions.

 

[KeithR]

I am sorry it comes across that way. As I noted to Mark and repeatedly so, this work started well before Harman and its value is manifested in car audio products today, not in masses of audiophiles buying the product. It is unfortunate that the thread keeps becoming a shootout between brands. It is not about that. It is about some rare science in a confusing world of sound reproduction and acoustics. I cannot separate the bias from the data. It is there. If we are suggesting that we should throw the baby out with the bath water, let's have a show of hands on that.

For now, you should know that I drive no income from my company and have had next to no sales generated from this forum. That is not by accident. I work hard to make sure I don't mix my company and the forum. But ultimately I can't zero out the connection, lest I also zero out sharing of the information.

Amir-

Firstly, I respect your contributions to the forum very much as a poster and as a moderator. This isn't a personal jibe.

However, in this thread you have specified that every other speaker you have had in store has not unseated the Revels and that you have participated in Harman studies that lead to listener preferences that make Revels sound better than others. Your other point is Harman has the money backing it to do superior analysis than other smaller speaker companies. You are unabashedly a fan of their techniques, studies, and methodology to the point of being "giddy" about it all.

The logical conclusion is that Revels are best and you are a dealer for them because of it. That comes across as a sales pitch to me.

I just think there is a conflict of interest for you in this thread, despite your efforts to be objective. If there are NRC studies that you can talk about, it would appear different. In some respect, you are an extension of Harman in this thread---and no speaker manufacturer would participate in such a thread other than to state or correct misrepresented facts.

 

[microstrip]

IMHO their scientists were much more interested in extracting one parameter from the 70 measurements that could be correlated with the preferences shown in the listening tests than in comparisons with the competition. This was the real challenge. Unhappily the marketing guys spoiled it all.

I resisted this initially....

I think there is a fundamental misunderstanding here. It is ONE parameter (FR), however it was not a single on axis measurement. It was seventy or so frequency measurements in many different polar positions.....

In other words, NOT seventy different parameters from which a single parameter (FR) was 'chosen'. (...)

Yes, Terryj there is a fundamental misunderstanding. Between what are exactly parameters, measurements and correlations. IMHO, the main point of the Harman work is scientifically establishing a correlation between measurements - yes the preference was measured, as well as many FR's.

In order to do it scientifically the problem must be formulated mathematically. For this the parameters must be expressed by simple numbers that can be analyzed using statistical tools, that give results and errors that must be appreciated. For me, the fantastic part of the work is taking all the data from 70 comprehensive measurements (having a resolution of 48 points per octave each, if I remember well) and reducing it to a reduced set of data (one parameter) that can be meaningful analyzed mathematically versus the preference parameter and validated. It is an huge work and it is why their papers are tens of pages long. Just reading the graphs and figure captions is not enough to appreciate their work.

BTW, I am referring to A" Multiple Regression Model for Predicting Loudspeaker Preference Using Objective Measurements: Part I and Part II" . This is the base of their scientific work. Sentences such "There are clear visual correlations between listeners' loudspeaker preferences and the set of frequency graphs" showing graphs made with four known speakers for me are just futile single cases. But forums love debating them.

 

[JackD201]

And the whole point of speaker placement is almost moot given the circumstances. Not to beat the same point over and over again, but I think it is the critical point:

The purpose of the measurements was to understand the linearity on and off-axis FR of the speakers in the test.

The purpose of the listening tests was to discover any correlation between off-axis FR and preference.

Martin Logans have extremely uneven off-axis response by design. This doesn't change with speaker placement.

Is there something else, something exceedingly wonderful that Martin Logans do that may have radically altered the results of the preference testing had they been placed differently? Maybe. That's another hypothesis, another test.

Tim

If I'm not mistaken the purpose was to find out what off axis FR behavior was preferred by the panel. The charts represent what the listeners were made to listen to.

Question: Did the panelist all do their evaluations from a set sweet spot?