Improving flat panel speakers in blind test results

[Blumlein 88]

https://secure.aes.org/forum/pubs/conventions/?elib=20603

Another for those with AES access to possible help discuss.

Blind testing of preference for conventional vs flat panel speakers. Two non-commercial flat panels with panel damping were able to score close to conventional speakers in blind preference testing while one commercial panel speaker scored much lower.

 

[MattHooper]

Not a subscriber so can't read it, but thanks for bringing it to attention.

Very interesting.

I wonder which commercial flat panel was used. I'm guessing initials "ML." (Though maybe the experimenters didn't want to use hybrid panel designs, so maybe something like Quads?).

I had Quad 63s for a few years, and I enjoy the ESL 57s. Never loved the Martin Logan speakers myself, but I get their appeal.

 

[amirm]

These are not flat panel speakers you are thinking of in hifi. They are instead panels that drive flat surfaces, often used in "architectural audio" to hide speakers behind drywall, etc.

Here is a shot of them:

They are saying that using an array of drivers with damped material can sound as good as typical bookshelf speakers (e.g. KEF LS50) and that standard prediction models (i.e. Olive et. al.) don't work for them. That is, they sound better than the model predicts. The hypothesize that their large size may have something to do with this.

 

[cjfrbw]

You can buy the transducers that attach to any planar surface for reasonable prices. One guy put them on guitars and used them as speakers. It's a subset of the 'bending wave' concept.

 

[solderdude]

discussed here as well


https://www.amazon.co.uk/DAEX25-Sound-Exciter-Speaker-Pair/dp/B001EYEM8C

 

[Blumlein 88]

These are not flat panel speakers you are thinking of in hifi. They are instead panels that drive flat surfaces, often used in "architectural audio" to hide speakers behind drywall, etc.

Here is a shot of them:
View attachment 42935

They are saying that using an array of drivers with damped material can sound as good as typical bookshelf speakers (e.g. KEF LS50) and that standard prediction models (i.e. Olive et. al.) don't work for them. That is, they sound better than the model predicts. The hypothesize that their large size may have something to do with this.

Okay, I was wondering if that were the case since I couldn't see the paper.

 

[tmtomh]

discussed here as well


https://www.amazon.co.uk/DAEX25-Sound-Exciter-Speaker-Pair/dp/B001EYEM8C

This video was the first thing that came to my mind too. Thanks for sharing.

 

[Sancus]

It's sort of staggering to me that somebody would spend the time and effort to make a video showing how to build those things, and show measurements, but their measurements make no sense(single point measurement, 1/24 smoothing and max out at 7.44khz? what?), and they don't actually show any measurements of the final setup, despite claiming that it's much flatter than the earlier measurements. I mean, what a mess, unless of course that's intentional or the guy is just semi-crazy.

 

[tmtomh]

https://secure.aes.org/forum/pubs/conventions/?elib=20603

Another for those with AES access to possible help discuss.

Blind testing of preference for conventional vs flat panel speakers. Two non-commercial flat panels with panel damping were able to score close to conventional speakers in blind preference testing while one commercial panel speaker scored much lower.

One think comes to mind: If they can get these close to conventional speakers in blind preference testing, then they should be more than good enough to use in background-music applications and could possibly be both more visually appealing (since they're invisible) and less expensive than decent in-wall or in-ceiling speakers.

 

[sfdoddsy]

There’s a very long thread at Audiocircle singing the praises of surface driven speakers.

https://www.audiocircle.com/index.php?topic=70541.

Personally, whilst I’m a fan of out of the box thinking, I’m also a fan of physics.

Such a speaker cannot be accurate.

Some (or even many) people like the sound of inaccurate speakers. Look at the success of B&W since they moved from reference to preference.

The old 801 Matrix measured like a textbook. Flat on axis, flat off axis.

Every recent B&W has a deliberate midrange dip, then a deliberate peak, then occasional flatness if you listen at just the right spot off axis.

Depressingly, all the new KEF R series have a lesser, but still deliberate, dip in the lower midrange.

Because people prefer that sound in listening tests.

At the extreme, many people (including some supposedly experienced reviewers at Stereophile) love the sound of the Zu Essence, recipient of the single worst set of independent measurements I’ve ever seen.

https://www.stereophile.com/content/zu-essence-loudspeaker-measurements

That’s not a speaker, it is a tone control.

I remember stories of when recorded music first came out 100 years ago and listeners couldn’t tell the difference between live and shellac.

It is also what I keep banging on about with the obsession here and elsewhere that the Harman spinorama and house curves are the holy grail.

They are accurate predictors of what people like in blind tests. But that’s not the same as predicting accurate speakers.

 

[Putter]

There’s a very long thread at Audiocircle singing the praises of surface driven speakers.

https://www.audiocircle.com/index.php?topic=70541.

Personally, whilst I’m a fan of out of the box thinking, I’m also a fan of physics.

Such a speaker cannot be accurate.

There's a little too much hand waving in that response. While I'll concede that NXT type transducers are not likely to find use as full range transducers, they've found applications in commercial speakers such as the various Cambridge Aero speakers as a midrange/tweeter and and as a midrange in the Philharmonic Audio BMR Philharmonitor which got rave reviews and great measurements from Audioholics.

 

[solderdude]

since they moved from reference to preference

I like those words ...

 

[dguillor]

I am using a single panel, Distributed Mode Loudspeaker, in my main system now. The reason I like it, is that it has unique acoustic characteristics. It is described as diffuse omnidirectional. It’s really pretty amazing. If you walk around the panel, it sounds almost the same, all you way around. In addition, the volume drops away much slower as you move away from the panel than for conventional speakers. I really like having sound that’s equal around the room.

 

[dguillor]

There’s a very long thread at Audiocircle singing the praises of surface driven speakers.

https://www.audiocircle.com/index.php?topic=70541.

Personally, whilst I’m a fan of out of the box thinking, I’m also a fan of physics.

Such a speaker cannot be accurate.

Some (or even many) people like the sound of inaccurate speakers. Look at the success of B&W since they moved from reference to preference.

The old 801 Matrix measured like a textbook. Flat on axis, flat off axis.

Every recent B&W has a deliberate midrange dip, then a deliberate peak, then occasional flatness if you listen at just the right spot off axis.

Depressingly, all the new KEF R series have a lesser, but still deliberate, dip in the lower midrange.

Because people prefer that sound in listening tests.

At the extreme, many people (including some supposedly experienced reviewers at Stereophile) love the sound of the Zu Essence, recipient of the single worst set of independent measurements I’ve ever seen.

https://www.stereophile.com/content/zu-essence-loudspeaker-measurements

That’s not a speaker, it is a tone control.

I remember stories of when recorded music first came out 100 years ago and listeners couldn’t tell the difference between live and shellac.

It is also what I keep banging on about with the obsession here and elsewhere that the Harman spinorama and house curves are the holy grail.

They are accurate predictors of what people like in blind tests. But that’s not the same as predicting accurate speakers.

They do have a rough response because their sound comes from exciting the vibration modes of the panel, but the response of a traditional speaker system is only smooth with a single speaker anechoically. As soon as you put two speakers in a room you also have a very rough response, especially away from the sweet spot.