It’s All a Preference

[Ronm1]

A lot of aluminum wiring is being used too. Not to worry though, these are much harder alloys than the ones they used for house wiring in the 1970s that burned down so many homes.

My understanding is the connections were the issue not the wiring itself. Same with chip leads when aluminum was/is used.

 

[Gregadd]

Hello Gregadd

Roger as in Green Mountain Audio?? Roger Sanders of Sandersounds sytems.

OK they can be manipulated by changing the scales, but what does this have to do with the posted response graphs?? These have the same scales so the comparison is a valid one.

Different than the the one I copipd form RUR.

 

[Soundminded]

Let me explain. The "live" tests were done with listener in the room or in situ. The "binaural" tests were done one using binaural recordings of the loudspeakers, later played back over earphones. The point of comparing live and binaural recordings of loudspeakers was to test the accuracy of the binaural system for doing controlled listening experiments where we could do instantaneous A/B/C/D comparisons of the same or different loudspeakers in different positions and different room acoustics. It turned out that the positional and room effects played a much greater factor in the results when you could instantaneously compare loudspeakers among different positions or rooms via the magic of the binaural system. In contrast, in the live/insitu tests, the room effects were very, very minimal: this is evidence that we quickly adapt to room acoustics within certain limitations.

The research showed that loudspeaker position and seat position can have a significant acoustical interaction with the room (i.e. from boundary effects, how the speaker couples into room modes). The positional effects can change the loudspeaker preference rating by 10-25%. When I moved from NRC to Harman, we decided to control these positional effects via an automated speaker shuffler. Before that we had to retest all the speakers in all the position to balance out any positional effects.

As long as the speakers are monopoles at low frequencies then the position will be the same for all speakers under test. The Martin Logan Prodigy is a hybrid electrostatic design with a monopole subwoofer like the other dynamic speakers we test. So, at low frequencies where the room dominates what we hear, it was not placed in any disadvantage relative to the other speakers under test.

It was placed far away from the side walls which the manufacturer recommends, and about 3-4 ft from the back wall. The main reason it was rarely poorly is because of its poor spectral balance and multiple resonances apparent in its anechoic measurements (speaker M in this graph). It is also very directional, and its spectral balance changes radically as you move off-axis unlike the other speaker tested. It actually sounds less harsh and bright as you move further off-axis, eventually changing from too bright to too dull. The resonances however appear in all on-axis and off-axis curves, so no matter where you position yourself or the speaker it will sound colored (unless you position yourself well outside the room.

We recently tested a less expensive ML model using over 200 high school/college students from Los Angeles and Japan and found that it received lower scores from listener sitting closer to the on-axis positions. In my view a speaker should have consistently good sound over a wide angle so that it is room friendly and listening seat friendly characteristic. So far, we've found no seat in our listening room where the ML is not considered colored, and the speaker measurements clearly tell us why

Hi Sean;

How did you test speakers where the manufacturer's recommendations were not in the position most favorable to the type you found "most pleasing" such as for example "Audio Note" where the manufacturer recommends placing the speaker in the corner of a room? What consideration did you give to speakers that had superior radiating patterns to the ones you found most preferable but were not flat without additional equalization and might have scored much higher if they were optimized as part of a system?

 

[Phelonious Ponk]

You are sooooooooo late it isn't even funny. The first time I saw people harvesting scrap copper was in 1980. It probably went on long before that. Abandoned houses are prime targets for scavangers stealing plumbing. Now China is buying up as much copper as they can lay their hands on. Copper has become so expensive that I now usually choose transformers with aluminum windings. Copper is just too expensive. A lot of aluminum wiring is being used too. Not to worry though, these are much harder alloys than the ones they used for house wiring in the 1970s that burned down so many homes.

I've got two houses full of the stuff. At the price I'm getting for these cables, I can retire on the pipes from a couple of 60s split levels, no problem.

Tim

 

[tonmeister2008]

Hi Sean;

How did you test speakers where the manufacturer's recommendations were not in the position most favorable to the type you found "most pleasing" such as for example "Audio Note" where the manufacturer recommends placing the speaker in the corner of a room? What consideration did you give to speakers that had superior radiating patterns to the ones you found most preferable but were not flat without additional equalization and might have scored much higher if they were optimized as part of a system?

I"m not familiar with Audio Note. But clearly if a speaker is designed to go in a corner or in/on a wall, it would have to be tested that way if that is the intended use. It was for this reason we built another listening room with an in-wall/on-wall/near-wall speaker mover.

I'm not sure what you mean by "superior" radiating patterns. We recently tested a speaker with a LF cardioid pattern placing it close to the wall, and then retesting it away from the wall. Interestingly, it's sound quality ratings were relatively independent of its position in the room due to the cardioid pattern exciting fewer room modes. Not a bad way to go from a design point, if you don't have the luxury of one or more subwoofers to cancel modes over a seating area, or a well-designed room correction device for single seat correction.

 

[Gregadd]

While the small sample size of listeners doesn't [sic] ** I am loath to correct anyones'grammar, but a double negative changes the meaning of a sentence]not allow us to make generalizations to larger populations, nonetheless it is reassuring to find that both the American and Japanese students, regardless of their critical listening experience, recognized good sound when they heard it, and preferred it to the lower quality options.

We agree on something.

We can agree on something. Others do cite your work as proof of what the general population prefers.
It would appear that the reason kids don't own better sounding audio solutions has nothing to do with their "radical" tastes in sound quality, but more do with other factors (e.g. price, convenience, portability, marketing, fashion) that have nothing to do with sound quality. Music and audio companies should take notice that kids can indeed discriminate between good and bad sound, and prefer accurate sound, despite what the media has been falsely reporting for the last few years. With that out of the way, we should focus on figuring out how to sell sound quality to kids at affordable prices and form factors they desire to own.

Indeed they can. Of course the difficulty is building a consesnsus on what good sound is. Then we find those who are motivated to develop it. Unforunately the profit motive has driven the indusrty as a whole. Thankfully there are some exceptions.

 

[Soundminded]

I"m not familiar with Audio Note. But clearly if a speaker is designed to go in a corner or in/on a wall, it would have to be tested that way if that is the intended use. It was for this reason we built another listening room with an in-wall/on-wall/near-wall speaker mover.

I'm not sure what you mean by "superior" radiating patterns. We recently tested a speaker with a LF cardioid pattern placing it close to the wall, and then retesting it away from the wall. Interestingly, it's sound quality ratings were relatively independent of its position in the room due to the cardioid pattern exciting fewer room modes. Not a bad way to go from a design point, if you don't have the luxury of one or more subwoofers to cancel modes over a seating area, or a well-designed room correction device for single seat correction.

"I'm not sure what you mean by "superior" radiating patterns."

My understanding of the conclusions of Floyd Toole's research is anong other things that early lateral reflections are desirable. This would result from superior lateral propagation or as it's more commonly called dispersion. If one speaker has greater lateral dispersion than another it will create stronger early lateral reflections in a given room and if the FR as a function of lateral radiation is relatively uniform then we'd expect that lateral reflections to have a spectrum closer to the driect arriving field than another which isn't. But suppose that its overall FR is not so flat and could be improved through equalization, the result if it were might be superior. As an example, looking at the off axis output of the Revel Ultima Salon 2 tweeter published in figure 6 of Stereophile Magazine's review I linked to earlier it showed considerable falloff starting above 8.5 khz although the on axis response was relatively flat. By comparison, the 3/4" dome tweeter used in many AR designs has a far more uniform off axis response down only 5db at 15 khz 60 degrees off axis from its on axis response but it is known to have a high end rolloff even on axis. If this rolloff is compensated for wouldn't that result in a superior speaker by virtue of its better lateral reflections?

 

[Phelonious Ponk]

Let me explain. The "live" tests were done with the listener in the room "in situ" with the speakers. The "binaural" tests were done using binaural recordings of the loudspeakers, played back over earphones. The point of comparing live and binaural recordings of loudspeakers was to test the accuracy of the binaural system for doing controlled listening experiments where we could do instantaneous A/B/C/D comparisons of the same or different loudspeakers in different positions and different room acoustics. It turned out that the positional and room effects played a much greater factor in the results when you could instantaneously compare loudspeakers among different positions or rooms via the magic of the binaural system. In contrast, in the live/insitu tests, the room effects were very, very minimal: this is evidence that we quickly adapt to room acoustics within certain limitations.

This is really interesting if I understand it properly. Obviously, the binaural recording, the headphone playback system, the phones themselves will have an effect on the reproduction, but I'm assuming you're using very accurate mics and headphones and comparing both measured and listener responses in room vs through the headphone system to understand exactly what impacts you're dealing with. I don't imagine it's easy, but the ability to compare not only speakers but placement, blind with an instantaneous A/B switch will yield very powerful information.

Please make a consumer active speaker. Please .

Tim

 

[tonmeister2008]

We agree on something.



Indeed they can. Of course the difficulty is building a consesnsus on what good sound is. Then we find those who are motivated to develop it. Unforunately the profit motive has driven the indusrty as a whole. Thankfully there are some exceptions.

Thanks for catching the double negative. I clearly couldn't decide whether to use "doesn't" or "does not" and ended up with a combination of the two I fixed it.

I don't believe reaching a consensus on what constitutes good sound is an issue. These 200+ student from US and Japan generally agreed on which speaker was the best one, and which one was the worst. The objective measurements clearly tell us why. It confirms results from previous studies.

There are bigger problems with the audio chain as far as sound quality goes, and I think most of them have to do with the quality of the content itself. The problem may get worse before it gets better as we move from physical media to streaming content from the cloud and using our cellphones to stream content (via Bluetooth) to our stereo systems at home and in our automobile. And these aren't just kids streaming music but anyone who owns a smart phone.

Loudness wars have eradicated the quality and dynamics of music. There is huge variability in quality of recordings, in part due to the lack of standardization of the recording/playback chain where the art is created. The music industry is generally resistant to science, technological advancements, and standards. And then we have music engineers/producers second guessing where we consume our content mix it to sound "good" over an iPhone speaker in a noisy bar. It kind of ruins it for anyone who has a decent sound system, so it would be better to optimize the content for maximum sound quality a let the user optimize the sound for the specific playback environment.

It turns out that the film industry has a 30-year old standard (SMPTE X-curve) that is not a standard at all, and many people in the industry ignore it because it's a lousy standard. Once we get the quality of the content under control perhaps we can work on standards for the consumer playback chain including the calibration of the loudspeaker/room interface.

So, we have some work to do.

 

[Gregadd]

I don't believe reaching a consensus on what constitutes good sound is an issue. These 200+ listeners on average agreed on which speaker was the best one, and the objective measurements clearly tell us why.

I don't doubt the results of your test. There may be a consensus at Harman, but the market does not reflect it. Or if you read this thread ,nor do I. You do agree the choices were limited.

 

[amirm]

I don't doubt the results of your test. There may be a consensus at Harman, but the market does not reflect it. Or if you read this thread ,nor do I. You do agree the choices were limited.

I don't get this argument that keeps getting repeated. Are you a worse lawyer Greg that one of your competitors who makes more money? Should I pick a law firm based on most number of clients and revenue?

 

[tonmeister2008]

This is really interesting if I understand it properly. Obviously, the binaural recording, the headphone playback system, the phones themselves will have an effect on the reproduction, but I'm assuming you're using very accurate mics and headphones and comparing both measured and listener responses in room vs through the headphone system to understand exactly what impacts you're dealing with. I don't imagine it's easy, but the ability to compare not only speakers but placement, blind with an instantaneous A/B switch will yield very powerful information.

Please make a consumer active speaker. Please .

Tim

When these studies were done at NRC (circa 1988-93) we used a KEMAR with pinnae, B&K microphones and prototype Etymotic earphones. We measured and calibrated the system to minimize any errors. It worked pretty well, but we've taken it one step further today and are doing something called Binaural Room Scanning. We take binaural impulse responses and/or make recordings at multiple angles. On playback we use a head-tracker that updates the auditory display every 30 msec so the signals at the ears correspond to the exact angle of head of the listener/mannequin immersed in the sound field. In this way, we get accurate localization and externalization of auditory images, which is necessary for realistic spatialization of sound sources. We can also do individualized listener calibrations to minimize errors related to the mannequin, headphones and how they couple to the individual listener.


Active speakers would be nice. They exist as professional products and as computer/Ipod docking/mobile devices. Hopefully, it will happen soon for higher-end consumer speakers.

 

[tonmeister2008]

I don't doubt the results of your test. There may be a consensus at Harman, but the market does not reflect it. Or if you read this thread ,nor do I. You do agree the choices were limited.

I don't disagree with you that the consumer loudspeaker market is fragmented in term of what they consider quality. We all seem to agree that amplifiers, and BLU-Ray players should be +- 0.25 dB but with loudspeakers, all sense of rational thinking seems to go out the door. I think this has to do with poorly controlled listening evaluations and poor objective measurements because when you have good subjective and objective measurements , consensus is usually reached.

That said, I think there is convergence over the last few years among the larger, more successful audio manufacturers in terms of what they believe the optimal target should be. I know this to be true because I do a lot of competitive benchmarking of loudspeakers, and the differences among competitors have shrunk. They are reading the available scientific literature and the measurement equipment, design tools,etc have come down in cost. Some even have anechoic chambers for doing accurate measurements but they don't always calibrate them or use them to their advantage.

There are still cases, where marketing/sales are clearly voicing the speaker to sound hyped in the bass and treble (Speaker B) and not natural, or they are going for high midrange sensitivity at the expense of bass in order to be the loudest speaker (Speaker C) on the floor at Best Buy. Or in the case of Loudspeaker D they clearly don't have any serious subjective/objective measurement capability or truly believe that coloring/distorting the music differentiates them from the pack. I've argued with misguided marketing people who think accurate sound is boring and not sellable, but fortunately they are long gone, and I am still here.

But on average, the design target for the majority of companies is flat, extended frequency response. The better ones get it right off-axis as well. The differences are how well they achieve their target and the devil is in the details.

 

[MylesBAstor]

Thanks for catching the double negative. I clearly couldn't decide whether to use "doesn't" or "does not" and ended up with a combination of the two I fixed it.

I don't believe reaching a consensus on what constitutes good sound is an issue. These 200+ student from US and Japan generally agreed on which speaker was the best one, and which one was the worst. The objective measurements clearly tell us why. It confirms results from previous studies.

There are bigger problems with the audio chain as far as sound quality goes, and I think most of them have to do with the quality of the content itself. The problem may get worse before it gets better as we move from physical media to streaming content from the cloud and using our cellphones to stream content (via Bluetooth) to our stereo systems at home and in our automobile. And these aren't just kids streaming music but anyone who owns a smart phone.

Loudness wars have eradicated the quality and dynamics of music. There is huge variability in quality of recordings, in part due to the lack of standardization of the recording/playback chain where the art is created. The music industry is generally resistant to science, technological advancements, and standards. And then we have music engineers/producers second guessing where we consume our content mix it to sound "good" over an iPhone speaker in a noisy bar. It kind of ruins it for anyone who has a decent sound system, so it would be better to optimize the content for maximum sound quality a let the user optimize the sound for the specific playback environment.

It turns out that the film industry has a 30-year old standard (SMPTE X-curve) that is not a standard at all, and many people in the industry ignore it because it's a lousy standard. Once we get the quality of the content under control perhaps we can work on standards for the consumer playback chain including the calibration of the loudspeaker/room interface.

So, we have some work to do.

Geez, we do agree on something: the software Interesting concept since those who suggested the very same thing eg the importance of the front-enf here have been subjected to endless ridicule.

'Cept how do you "standardize" the front end/software when we all listen to different sources?

 

[RUR]

...The main reason it was rarely poorly is because of its poor spectral balance and multiple resonances apparent in its anechoic measurements (speaker M in this graph). It is also very directional, and its spectral balance changes radically as you move off-axis unlike the other speaker tested. It actually sounds less harsh and bright as you move further off-axis, eventually changing from sounding too bright to too dull. The resonances, however, appear in both the on-axis and off-axis curves, so no matter where you position yourself or the speaker it will sound colored (unless you position yourself well outside the room)....

Hi Sean,

Long ago and far away in 1997, the ex-owner of ML responded to similar measurements in a magazine review by asserting that ideal measurement distance for a 48" electrostatic element is necessarily different than for a point source. That response may be found here. Does this make any sense to you or does it fall into the "exculpatory marketing babble" category?

 

[terryj]

good to hear from you Sean.

Perhaps not so good nthat you get flooded with questions!

You have been addressing a few of the 'problems' with your results, one has been the tracks used to base the auditions on. The charge is (I think) limited in number and hence possibly deliberately chosen to slant results. The limited in number does make 'sense' if you follow when wondering about universality.

I'd imagine these were not chosen on an ad hoc basis, is there some insight or reasoning behind them you could share? How valid could the claims of skewing the results be?

 

[Phelonious Ponk]

Geez, we do agree on something: the software Interesting concept since those who suggested the very same thing eg the importance of the front-enf here have been subjected to endless ridicule.

'Cept how do you "standardize" the front end/software when we all listen to different sources?

When has anyone here been ridiculed for suggesting that recording quality is inconsistent? I mean by anyone but Frank? I think almost all the objectivists on this board would agree that there is a serious quality/consistency problem in recording. Perhaps you meant something else by software?

Tim

 

[microstrip]

Let me explain. The "live" tests were done with the listener in the room "in situ" with the speakers. The "binaural" tests were done using binaural recordings of the loudspeakers, played back over earphones. The point of comparing live and binaural recordings of loudspeakers was to test the accuracy of the binaural system for doing controlled listening experiments where we could do instantaneous A/B/C/D comparisons of the same or different loudspeakers in different positions and different room acoustics. It turned out that the positional and room effects played a much greater factor in the results when you could instantaneously compare loudspeakers among different positions or rooms via the magic of the binaural system. In contrast, in the live/insitu tests, the room effects were very, very minimal: this is evidence that we quickly adapt to room acoustics within certain limitations.

The research showed that loudspeaker position and seat position can have a significant acoustical interaction with the room (i.e. from boundary effects, how the speaker couples into room modes). The positional effects can change the loudspeaker preference rating by 10-25%. When I moved from NRC to Harman, we decided to control these positional effects via an automated speaker shuffler. Before that, we had to retest all the speakers in all the positions to balance out any positional effects -- very time consuming.

As long as the speakers are monopoles at low frequencies then the position will be the same for all speakers under test. The Martin Logan Prodigy is a hybrid electrostatic design with a monopole subwoofer like the other dynamic speakers we test. So, at low frequencies where the room dominates what we hear, it was not placed in any disadvantage relative to the other speakers under test.

It was placed far away from the side walls which the manufacturer recommends, and about 3-4 ft from the back wall. The main reason it was rarely poorly is because of its poor spectral balance and multiple resonances apparent in its anechoic measurements (speaker M in this graph). It is also very directional, and its spectral balance changes radically as you move off-axis unlike the other speaker tested. It actually sounds less harsh and bright as you move further off-axis, eventually changing from sounding too bright to too dull. The resonances, however, appear in both the on-axis and off-axis curves, so no matter where you position yourself or the speaker it will sound colored (unless you position yourself well outside the room).

We recently tested a less expensive ML model using over 200 high school/college students from Los Angeles and Japan and it was rated as the least preferred choice. It also received lower scores from listeners sitting closer to its on-axis positions. In my view, a well-designed speaker should have consistently good sound over a wide angle so that it is both room friendly and seat friendly. This way, more of direct and reflected the sound reaching the listener(s) will be higher quality, and the sweet spot for ideal listening is also increased. So far, we've found no seat in our listening room where the ML does not sound colored, and the speaker measurements clearly tell us why.

We generally do all of our tests using single speakers because we've shown evidence that listeners are more sensitive to distortions present in the loudspeakers when auditioning a single source versus multiple sources. If loudspeakers score well in mono they generally score higher in stereo and surround, although the relative rank ordering of the loudspeakers doesn't generally change as you move from mono to stereo. This reference documents some of the evidence we've gathered. Floyd Toole has also written about it in his book and here:

Sean,

Thanks for answering. I am not questioning your work or what constitutes a good sound. You are using a statistical preference, and for this type of studies it is the correct approach.

But as the study conclusions about the ML Prodigy are quite different from mine (I have owned them for a few years long ago) and after some fine tuning of the system they sounded really good, I am not happy with your conclusions. As I have also listened to them in other systems playing even better than in my system, I must question either my preference or the conditions your study. As I am not alone in preference (by chance I found now that the Stereophile review of the Ultima Salon 2 and Prodigy were done by the same reviewer – an interesting reading freely available at the magazine site that I recommend to everyone following this thread - and thanks to the internet at that time I made a few friends around the world owning Prodigy, I have looked at the listening conditions of the published tests with some care.

IMHO it seems that our divergence is mostly due to what you expect from a sound reproduction system and the expectations of a typical audiophile. You want a speaker that does not need fine tuning and performs independently of the system where it is inserted for a large audience using mainstream music. Audiophiles usually accept to compromise in one or more of these conditions to get a little more of the emotion existing in the recording. But it seems that for the Harman tests all of them are fundamental and mandatory. Also, it seems to me that the conditions used for the tests of the ML Prodigy were not fair to the manufacturer and to the users of this speaker who enjoyed or still enjoy it, as you did not optimize them, and specially, you were not looking for the extra that this type of speaker can offer over typical box speakers.

I must tell I am not a typical music consumer – I listen mainly to acoustical instrumental music, classical and some jazz and ethnical music, using the type of recordings Harman chooses for their tests mainly for entertaining friends in hifi auditions. Orchestral chamber music, ancient music and jazz, Mahler, Shostakovitch, Jordi Savall, etc. populate my recordings shelves. I am not the type of listener who considers of extreme importance if the piano was a Steinway or Bösendorfer, or if the recording was mixed in a JBL or B&W monitor, but I want to recreate as much as possible the feelings I have when I listen to real music in good auditoriums - not the exact perceptions. Chamber music is particularly challenging in this aspect – as it does not have a conductor you must feel the connection and complicity between the players in the way they play their instruments. Finales in symphonies should transmit power, not loudness, and sections of the orchestra should sound with correct weight and spatial presence. Sometime ago I bought the F. Toole book and really enjoyed reading it - specially part one : Understanding the principles. It helped me to understand what I was looking for in sound reproduction.

An important question can be asked after reading this study. What were the speakers used as monitors during the mastering of the recordings used in these tests?

Thanks for participating in this debate.

 

[Soundminded]

I don't get this argument that keeps getting repeated. Are you a worse lawyer Greg that one of your competitors who makes more money? Should I pick a law firm based on most number of clients and revenue?

If the issue is to determine market preference, wouldn't the best way be to find out what model sold the most units? It would probably be something like AR3a or AR2ax. Maybe KLH Model 6 or 17. It would not be Revel Salon Ultima. I still think it's strange how even ardent audiophiles don't take Toole's advice about installing 4 subwoofers in the corners or midwall in a room to get uniform bass response. That strikes me as a real contribution and a genuine problem solved. There's a lot of conventional thinking in this industry. The unwillingness to experiment with radical new ideas usually spells the decline of an industry, technology that has matured past its sell by date and no prospect of newer better technology on the horizon.

 

[amirm]

If the issue is to determine market preference, wouldn't the best way be to find out what model sold the most units?

No, why should it? Let's say I make the tastiest ice cream tomorrow and only sell from porch of my house. Why would my sales indicate my ice cream is better or worse that Dryers?

Now, if everyone bought their speakers from the same place and they were shown side by side as I heard them behind a curtain, I am telling you that they would own 90% of the market. But such is not the case.

I still think it's strange how even ardent audiophiles don't take Toole's advice about installing 4 subwoofers in the corners or midwall in a room to get uniform bass response. That strikes me as a real contribution and a genuine problem solved. There's a lot of conventional thinking in this industry. The unwillingness to experiment with radical new ideas usually spells the decline of an industry, technology that has matured past its sell by date and no prospect of newer better technology on the horizon.

Indeed getting the bass right makes a huge difference. Boomy and improper bass has a big impact on fidelity of a speaker system.