Why the Harman mono speaker test was wrong for dipole planers

[DaveC]

Let me comment on another thing before more people want to hang me . And that is the notion of loudspeaker break in. Please don't fall for that.

Let me explain this with a personal story. We have two homes, one with city water and another from private well on our property. The water where we live is very high quality seeing how it rains all the time. We moved a few months ago to our vacation house. I come back to our main home with city water once in a while. I have tea in the morning. After staying in our vacation house for a month, I make my tea and I am shocked that it tastes like pool water! It is as if I am drinking Chlorine. We have lived in our primary house for 18 years and never ever did I detect this. Yet here I am, I can barely drink the water once I got used to having well water with no chlorine.

So what is going on here? Adaptation. Our brain has an amazing ability to filter out what is invariant. It took out the Chlorine since it was always there. I got used to it. Once I started to drink water without it, it built back up the sensitivity to it.

Back to loudspeakers, if you buy a loudspeaker, turn it on and it doesn't sound right, return it! That is the moment of truth. That is when your mind fresh and can detect variations that don't sound right. Do not keep the loudspeaker and let it "break in." Because what is happening instead is that your brain is starting to adapt and filter out what is bad about it sound. The speaker is not improving. It is your detection ability that is degrading!

If the manufacturer insists that such break in is necessary, ask them to break it in and then send it to you.

And yes, this has also been researched and break in makes almost no difference whatsoever to the tonal quality of the loudspeaker. The driver characteristics does change but is not reflected in loudspeaker performance.

Lol, Vas doubles after the woofer breaks in, but that's not audible, right?

 

[esldude]

Read post number 23.

A lacking description of measurements. He says for instance the ML is flat 700 hz to 20 khz. No tolerance given, no mention of smoothing of the response or not employed. We have other sources of measurements from several places. None are flat in the sense of amplifier flat or even audibly flat to a close tolerance. Below 700 hz is said to be 3 db up and stays that way until 55 hz. Again other measurements don't show an incredibly flat result.

Probably makes more sense if you could see the measurement results. But this description is lacking without the other information of the graphs etc of what is being described.

 

[Gregadd]

let me comment on another thing before more people want to hang me . And that is the notion of loudspeaker break in. Please don't fall for that.

Let me explain this with a personal story. We have two homes, one with city water and another from private well on our property. The water where we live is very high quality seeing how it rains all the time. We moved a few months ago to our vacation house. I come back to our main home with city water once in a while. I have tea in the morning. After staying in our vacation house for a month, i make my tea and i am shocked that it tastes like pool water! It is as if i am drinking chlorine. We have lived in our primary house for 18 years and never ever did i detect this. Yet here i am, i can barely drink the water once i got used to having well water with no chlorine.

So what is going on here? Adaptation. Our brain has an amazing ability to filter out what is invariant. It took out the chlorine since it was always there. I got used to it. Once i started to drink water without it, it built back up the sensitivity to it.

Back to loudspeakers, if you buy a loudspeaker, turn it on and it doesn't sound right, return it! That is the moment of truth. That is when your mind fresh and can detect variations that don't sound right. Do not keep the loudspeaker and let it "break in." because what is happening instead is that your brain is starting to adapt and filter out what is bad about it sound. The speaker is not improving. It is your detection ability that is degrading!

If the manufacturer insists that such break in is necessary, ask them to break it in and then send it to you.

And yes, this has also been researched and break in makes almost no difference whatsoever to the tonal quality of the loudspeaker. The driver characteristics does change but is not reflected in loudspeaker performance.

oy vey!

 

[microstrip]

(...)
So what do I think of name dropping of Linkwitz to invalidate the research being performed? Bad form on top of reluctance to learn.
(...)

A long post (thanks for your effort) just to divagate along your main idea - referring to public and accessible research and opinions of someone who does not agree with your beliefs is "bad form on top of reluctance to learn".

Probably it is the problem of all people who prefer speakers, electronics or rooms you do not endorse - "reluctance to learn".

The significant number of people, although in minority, who preferred the best selling designs in the listening tests - "reluctance to learn".

All those that post about tweaks, tubes, cables or sighted opinions - "reluctance to learn".

WBF - 99% of "reluctance to learn", the number is reducing as many frequent posters have left.

 

[amirm]

Have you done so amirm?

I have. But better yet, went and found the original article with the measurements within: http://www.hi-fiworld.co.uk/loudspeakers/65-reviews/252-martin-logan-clx.html?showall=1

Let's review and quote what you extracted from it:

Frequency response of the CLX is flat from 700Hz all the way up to 20kHz, so it will sound even in its midband and upper midband/treble delivery.

You didn't show the measurements in the article so let's do that now:

(their site is getting errors -- see the green graph below on the right; it is the same thing)

How on earth does this graph read "flat from 700 Hz all the way up to 20 Khz?" There is a -7 db or so sharp dip around 4 Khz for example. We also have nice peaks at 2.5, and 3.5 Khz. Could those be caused by the room or the loudspeakers?

What smoothing was used? Not stated.

Can you tell how that loudspeaker sounds in your room from that measurement? No way. A single measurement by a single microphone has no prayer of telling you what you would hear in your room. Single line frequency measurements like this are exceptionally misleading. This is not an amplifier. You don't measure it this way.

Let's compare the measurements published in the Harman papers for the ML (which was a different model by the way than CLX) (left) and the one from your article (right):

The Harman measurements in raw format are actually 72 graphs at different points in space, not one, and all measured in anechoic chamber so there are no interactions between them. They distill that into different buckets such as listening window, early and late reflections. By comparing these we can see what happens when the reflections are added to the direct sound. Or put another way, how well behaved the loudspeaker is with respect to all the sound that it radiates into the room.

The two lines at the bottom of the graph show the deviation from direct sound. Focus there and notice how there is a deep drop starting around 2 Khz and continuing to about 10 Khz. That trough shows that the sum total of what you hear, i.e. direct and indirect sounds, has a big hollowing range in the mid to mid-high frequencies. There is no picture of a "flat response" as your review claimed. We care about what we hear and that is NOT flat. He talks about what his microphone picks up at close distance. Difference between professional research and DIY work done in that magazine.

It is these bottom lines in the left graph that guide us toward what sounds good to listeners in controlled testing. We want those graphs to be smooth and not have such large dips. Note that I did not say flat.

So am I advocating any old measurements? Absolutely not. This kind of old fashioned single measurement of loudspeakers presented in that review is totally wrong and practically useless. And the text that goes with it even more so.

Did I miss something you wanted me to read?

 

[amirm]

A long post (thanks for your effort) just to divagate along your main idea - referring to public and accessible research and opinions of someone who does not agree with your beliefs is "bad form on top of reluctance to learn".

Once again, I presented no beliefs. I present the research word for word. They are not my work or beliefs. I agree with them but you can't make them mine. You are disagreeing as such with research and go on to say they are crooks with financial motivations and have cooked up tests and research. That's a belief that you need to substantiate.

You then gave us the link to linkwitz web site with no mention of what in there supposed to back your point of view. So let me quote a link there where he went to AES in 2005 and wrote what he found notable there: http://www.linkwitzlab.com/Acoustics and Psychoacoustics of Loudspeakers in Small Rooms.doc

"The Acoustics and Psychoacoustics of Loudspeakers in Small Rooms: A Review.
For example, Floyd Toole gave this much more technical and extended version of his CEDIA talk. From the convention program tutorial synopsis: “The physical measures by which acousticians evaluate the performance of rooms have evolved in performance spaces.

[...]

Harman testing has shown that speakers with a fairly flat on-axis response and with a smooth and gentle off-axis level fade and roll-off are preferred by expert and amateur listeners. A speaker with an off-axis response that differs substantially from its on-axis response cannot be corrected with an equalizer. In-room measurements lower the accuracy of the data since they include a mix of direct and reflected sound. Since with most speakers the reflected sound has a different frequency response than the direct sound, an equalizer will cause more problems than it will cure."

The underlines are all his. Notice the emphasis on the point I just made in the discussion of ML measurements. How we want a smooth diversion from direct axis sound of the loudspeaker, and not one with large dips or variations in it. And how such preferences are formed by both trained and untrained listeners.

And as if to back what I just post word for word, he says this from Toole's presentation:

"He also noted that spatial-averaging of speaker measurements is critical —single-point measurements are erroneous and meaningless. Also, many types of measurements are required to characterize the sound, and that 20th-octave data is necessary to unmask narrow-bandwidth frequency response problems that would be averaged out and masked in third-octave measurements."

Again, all underlines are his. Note how he is emphasizing the fact that single measurements like there was in Kenwood review of ML CLX are "erroneous and meaningless."

So your own expert you put forward to counter Toole, relies on and agrees with Dr. Toole on these key points related to the core of this topic. He, despite his expertise sat through the presentation and walked away with these points.

As I said, many experts in the industry follow this research. It is not that I am that lucky that the one name you put out there would support the research. This work is highly trusted and utilized in understanding of sound reproduction in rooms. And fortunately the concepts within it are not too complicated. If only you would try to learn than complain....

 

[RayDunzl]

I'm convinced.

My speakers are crap.

I still like them, though.

 

[microstrip]

Once again, I presented no beliefs. I present the research word for word. They are not my work or beliefs. I agree with them but you can't make them mine. You are disagreeing as such with research and go on to say they are crooks with financial motivations and have cooked up tests and research. That's a belief that you need to substantiate.(...)

I never said what you are stating, unfortunately you are using your usual style of free abusive interpretation of others words and then asking for proof of distorted claims once you have no arguments. It is ridiculous.

I have been very careful to separate the aggressive marketing from people as you from the excellent research of F. Toole. You insist on mixing it. I present views of other experts on audio, people can take their conclusions. I have taken mine, right or wrongly, it is my own choice. I hope our readers do the same by themselves, I am not in a crusade for "audio science". See http://www.linkwitzlab.com/Links/issue29-SL-letter.jpg from the S. Linkwitz site. See the fundamental need of using unamplified sound as a reference, something that completely separates their paradigms.

There are several other brief comments of deep disagreement between them - I am sure that an expert Google user as you has found them in his site. It is very common in every science.

There are plenty of people who have different opinions from F. Toole and debate them in the net. I do not want to show as someone against his research just because marketing people uses the Harman research against the competition, creating a conspirator theory about the high-end.

 

[amirm]

I'm convinced.

My speakers are crap.

I still like them, though.

And I still like you Ray .

 

[rbbert]

Perhaps worthy of note is that once Peter Aczel heard and tested one of Linkwitz' speakers, they became his reference, and have remained so through multiple redesigns and upgrades. Please don't take this as an endorsement for The Audio Critic, though

 

[amirm]

I have been very careful to separate the aggressive marketing from people as you from the excellent research of F. Toole. You insist on mixing it. I present views of other experts on audio, people can take their conclusions. I have taken mine, right or wrongly, it is my own choice. I hope our readers do the same by themselves, I am not in a crusade for "audio science". See http://www.linkwitzlab.com/Links/issue29-SL-letter.jpg from the S. Linkwitz site. See the fundamental need of using unamplified sound as a reference, something that completely separates their paradigms.

Reviewing what the research says is not 'marketing.' Marketing has sales intent. None is here. So let's not keep spinning the discussion. As to you having respect for Toole's research, this is what I responded to earlier:

Fantastic - in order to understand the competition success they recreated conditions in which they sounded poor to carry their tests.

Doesn't show an an ounce of respect.

That aside, let's review what the letter says:

He confirm yet again the simple concept behind Toole's research we have been discussing. So with respect to what is important in this regard, there is no disagreement.

Where he goes on is to advocate, dare I say market, the idea of open baffle dynamic drivers. To which I say, show me the results of double blind testing tests. I want to know what the listener ears heard, not an opinion of what should sound good. That's all. Do you have such data? Does he? If no, why do you say you respect Dr. Toole's research where the very core of it is not followed, i.e. confirmation through listening tests?

There are plenty of people who have different opinions from F. Toole and debate them in the net. I do not want to show as someone against his research just because marketing people uses the Harman research against the competition, creating a conspirator theory about the high-end.

Again, for the hundredth time, this is not "Harman" research. The worked started under NRC in Canada.

And I have never, ever seen Harman use any of this research data against their competitors in any marketing material. I have pushed them to publish data that they have on countless loudspeakers in the market. But out of professional courtesy, their have chosen to not do that. So please, park this marketing argument on the side. It doesn't not hold water in any form or fashion.

As to different opinions, there are millions of them. The whole reason to read this topic is to learn how we can verify those claims. And that verification comes from controlled testing, research that is published and accepted by others, etc. Not standing up and saying this and that loudspeaker design is better.

 

[amirm]

Other than Harman, which other speaker manufacturers would you say design their products with the same performance characteristics the research demonstrates is most desirable?

I don't have a comprehensive list to offer. Only examples. One of them is KEF. Here is their write-up on their coaxial drivers that they are famous for: http://kef.com/html/us/innovation/uni-q/index.html

"The second advantage of Uni-Q is what we call ‘matched directivity’. With the treble unit mounted at the centre of the bass driver’s cone, its directivity (the spread of sound away from the main axis) is governed by the angle of the cone, which also largely determines the directivity of the bass driver. So with the coincident mounting of the two units, the directivity of the treble unit is adjusted to be virtually the same as that of the bass driver."

And this graph:

Notice the series of graphs and how they only deviate from each other in slope. No sudden dips.

There are a also a number of companies that spun off from NRC such as PSB, and Paradigm that somewhat follow this research although their work is dominated by marketing/business aspects. See this from their web site for example: http://www.psbspeakers.com/articles/Paul-Barton-at-the-NRC. And this from their marketing bit on one of their loudspeakers:

"The perfect proportion of Sound Power, combined with extremely even frequency response and uniform off-axis response, creates a massive, effortless sounding and precisely detailed sound field."

This is Paradigm: http://www.paradigm.com/support/about-paradigm.php

"Unlike almost every other audio manufacturer, we have our own sound research and development facilities—a 36,000-cubic-foot anechoic chamber to take accurate sound measurements and our own double blind listening rooms to test every product we produce."

 

[Don Hills]

... Let's not get too carriesd away. The prefer a standard that makes their product sell. ...

You have put the cart before the horse. They have worked out what sounds good to people, and they produce speakers to that standard.

 

[Argonaut]

Of course you have missed something, any creditable real world experience of that particular transducer, well positioned, and in a room optimised for its use. You state your position that the 'measured' should support the primary that is the audible, and yet you are unwilling to accept the later from myself, from professional reviewers, not to mention the several hundreds of paying customers whose acuity and perspicacity you quite obviously demean.

NK used the word '"essentially" flat, and other than the obvious room interactions, that you yourself offer in mitigation, it most certainly does not, as has been yapped about on your forum "measure poorly" nor invalidate any subjective considerations within that review as to how audio reproduction via that transducer manifest itself, Were you Present ?

The somewhat tardy realisation, on my part, that you are also in fact a 'Pedlar of Boxes', the portfolio of which would seem to place certain of your wares in competition with Martin Logan, does afford me an altogether 'Other' prism through which I might view your considerations on this subject matter.

 

[DaveC]

Amir, on the speaker break in thing I'm currently breaking in woofers and their initial Vas is half of the advertised figure. How is it at all possible that a doubling of the Vas, which indicates the woofer's suspension is becoming twice as compliant as before break-in, could possibly have no audible effect? Other speakers I have had have changed dramatically as well, but this is the first one I've measured T/S parameters to see the changes and I'm surprised by how large some of them are.

 

[TBone]

There are a also a number of companies that spun off from NRC such as PSB, and Paradigm that somewhat follow this research although their work is dominated by marketing/business aspects.

In the early years of API, they marketed/advertised differently. Purchased by Klipsch Group, this quote on their site ...

Founded in 1973 and headquartered in Toronto, Canada, Audio Products International Corp. (API) is one of the top manufacturers of high-performance loudspeakers in the world. API manufactures over 200 models of speakers and exports to over 50 countries under the company's three world-renowned brands: Energy Speaker Systems, Mirage and Athena Technologies. In 2006, API is entering the professional market with the launch of its Energy Pro division. The API headquarters includes a state-of-the-art research and design facility. Lead by Ian Paisley, API's VP of Engineering and world-renowned acoustic researcher, API was a pioneering member of the Canadian National Research Center's studies to determine the design characteristics of a great sounding loudspeaker. API's continued psycho-acoustic research along with their three anechoic test chambers, maintain the highest levels of engineering excellence and product performance.

 

[microstrip]

Reviewing what the research says is not 'marketing.' Marketing has sales intent. None is here. So let's not keep spinning the discussion.

There is a lot of marketing in the reviewing of that research. The style in the book and in the Harman site is completely different - anyone reading both will immediately see it. Selecting through texts is always marketing in these cases. Weren't you the one who said marketing is 90%?

As to you having respect for Toole's research, this is what I responded to earlier:

Doesn't show an an ounce of respect.

No it does not show disrespect, as you pretend. It comments a real fact and is my opinion - if you really want to study the competition you should recreate the conditions in which it sounds better - not those every Martin Logan owner (a big number as you said) will say are not representative.

That aside, let's review what the letter says:

He confirm yet again the simple concept behind Toole's research we have been discussing. So with respect to what is important in this regard, there is no disagreement.

Where he goes on is to advocate, dare I say market, the idea of open baffle dynamic drivers. To which I say, show me the results of double blind testing tests. I want to know what the listener ears heard, not an opinion of what should sound good. That's all. Do you have such data? Does he? If no, why do you say you respect Dr. Toole's research where the very core of it is not followed, i.e. confirmation through listening tests?

Again, for the hundredth time, this is not "Harman" research. The worked started under NRC in Canada.

And I have never, ever seen Harman use any of this research data against their competitors in any marketing material. I have pushed them to publish data that they have on countless loudspeakers in the market. But out of professional courtesy, their have chosen to not do that. So please, park this marketing argument on the side. It doesn't not hold water in any form or fashion.

As to different opinions, there are millions of them. The whole reason to read this topic is to learn how we can verify those claims. And that verification comes from controlled testing, research that is published and accepted by others, etc. Not standing up and saying this and that loudspeaker design is better.

Five non representative lines of the whole page. What is more relevant is not even the concept of the open baffle dipole, it is much more than that. It is that another speaker concept can performs better in stereo than what is being proposed by the research of NRC and Harman. That there are alternatives. That there is place for the high-end.

And sorry, I am not interested in proving anything in the dying science section.

 

[amirm]

Amir, on the speaker break in thing I'm currently breaking in woofers and their initial Vas is half of the advertised figure. How is it at all possible that a doubling of the Vas, which indicates the woofer's suspension is becoming twice as compliant as before break-in, could possibly have no audible effect? Other speakers I have had have changed dramatically as well, but this is the first one I've measured T/S parameters to see the changes and I'm surprised by how large some of them are.

It will take me a while to locate references to give you. For now, the first piece of data was as you guessed it, at Harman. They pre-tested drivers as you did and found noticeable difference. But once in an enclosure, the effect almost vanished. It was noticeable as a tiny variation and that was likely inaudible. Since then, I have run into other people analyzing the same but in a quick search I could not find it. If I come across it, I will post.

Edit: just did a quick search online and ran into one of the articles I had read: http://www.audioholics.com/loudspeaker-design/speaker-break-in-fact-or-fiction

"The compliance of the air enclosed within the enclosure is significantly less than the driver's suspension compliance, therefore the enclosure has the dominant influence where it comes to determining things such as system resonance. Thus, owing to the physics of the enclosed box loudspeaker system, any pre/post differences in driver suspension mechanical compliance are constrained from having as large an effect on parameters such as resonance frequency, etc as intuition might lead one to believe it should.

Taken together, it's clear the volume of air confined within the sealed cabinet of the enclosed box loudspeaker system moderates any measurable and/or audible changes that might arise as a consequence of driver compliance changes."

 

[amirm]

What is more relevant is not even the concept of the open baffle dipole, it is much more than that. It is that another speaker concept can performs better in stereo than what is being proposed by the research of NRC and Harman. That there are alternatives. That there is place for the high-end.

Neither NRC nor Harman talk about approaches to speaker design. Harman itself makes different types from arrays to horns and traditional loudspeakers. What the research entails is proving that the design works through controlled listening tests, and a set of measurements that seem to correlate well with them. How you get there as a designer, is up to you.

That is what ultimately matters: how the loudspeaker sounds. Everything else is talk in dire need of verification beyond ad-hoc listening tests in different rooms.

And there most definitely is room for the high-end. No one has said otherwise.

 

[microstrip]

(...)
This is Paradigm: http://www.paradigm.com/support/about-paradigm.php

"Unlike almost every other audio manufacturer, we have our own sound research and development facilities—a 36,000-cubic-foot anechoic chamber to take accurate sound measurements and our own double blind listening rooms to test every product we produce."

Do they test burn-in using these facilities and double blind listening tests? Quoted from their site OM104 owner manual.