Electrostatic speakers?

[RetroStereo]

I am new to speakers in general and would like to ask a pretty broad question on whether Electrostatic speakers are worth bothering with over conventional dynamic ones.

I have a lot of experience with headphones. There, electrostatic and planar drivers generally reign supreme over dynamics due to their rock bottom distortion that applies to the entirety of frequency response, including the low end. So even tho they very rarely have a natural bass shelf, you can easily add one with EQ and face no drawbacks because the drivers can take it no problem.

So I wonder if the same rules apply to electrostatic speakers. I looked through ASR and some other sources and couldn't find measurements that would give me a clear answer on this topic.
Looking at conventional speakers however, I have seen that they behave much more eratically than headphones. Many having distortion levels shooting through the roof in many places and pretty much always a sharp bass roll off. With only very few (usually also very expensive) speakers managing to avoid those problems.

Will electrostatic speakers help me combat those issues? Do they share the same technical advantages that electrostatic headphones have over dynamic headphones?

I'm biased. However, I'll try to give you an idea of what to expect. Most electrostatic speakers are dipoles whereas most dynamic speakers are not. They require a different mindset as to room placement and treatment. Is it worth the effort and expense? In my opinion, yes.
But there are classes of speakers that are also very similar to electrostatic, including ribbon, planar (similar to the headphones) and something called air motion transformers (AMT). All have their place and all can be part of exceptional speaker systems.
Two of the most famous names among electrostatic are Quad and MartinLogan. Both make exceptionally fine products, unfortunately, at exceptionally high prices, too. And because electrostatics are essentially giant capacitors, they can represent a brutal load on amplifiers. For example, my Monolith ESL panels are rated at a nominal 4 ohms; however, as frequencies reach ever higher that drops precipitously to 0.8 ohms at 20KHz and 0.6 ohms at 24KHz. Some amps simply implode, sometimes taking the panels with them. Luckily that is less true today.
Another problem is their sheer size. My panels are 26" wide and 48" tall on top of the enclosure that houses the dynamic woofer. The transition between the panel and it's dynamic partner depends on several factors, but the main one is simply square feet of surface. Mine crossover naturally at 6dB per octave at 125Hz so the woofer must take up the bass slack.
Therein is another problem. ESLs are really, really fast and large dynamic drivers just are not. When the Monolith was introduced, people loved its mids and highs but described its low end as muddy. I've resolved that issue with my monoliths by replacing the original woofers with what are probably the fastest model currently made. ML has also incorporated DSP in most of their top models.
A well designed dynamic speaker can have near vanishing levels of distortion until it is overdriven. It also can generally be driven at higher volume levels than most electrostatics. I haven't found that to be much of a problem. I can easily get neighbor complaint levels of sound from mine.
I really can't tell you that ESLs are for you. For me, I knew that I'd eventually have a pair after I first heard a pair of Dayton Wright XG8 models well over 50 years ago. Are they perfect? No. No speaker is. But I like the way they sound.

 

[Bjorn]

I am new to speakers in general and would like to ask a pretty broad question on whether Electrostatic speakers are worth bothering with over conventional dynamic ones.

I have a lot of experience with headphones. There, electrostatic and planar drivers generally reign supreme over dynamics due to their rock bottom distortion that applies to the entirety of frequency response, including the low end. So even tho they very rarely have a natural bass shelf, you can easily add one with EQ and face no drawbacks because the drivers can take it no problem.

So I wonder if the same rules apply to electrostatic speakers. I looked through ASR and some other sources and couldn't find measurements that would give me a clear answer on this topic.
Looking at conventional speakers however, I have seen that they behave much more eratically than headphones. Many having distortion levels shooting through the roof in many places and pretty much always a sharp bass roll off. With only very few (usually also very expensive) speakers managing to avoid those problems.

Will electrostatic speakers help me combat those issues? Do they share the same technical advantages that electrostatic headphones have over dynamic headphones?

We used a large planar driver in the Vera Coherence 12 for the reason you're mentioning and combined it with a dynamic woofer to compat the dynamic drawbacks of an electrostatic speaker.

However, there's also another benefit with a speaker and the use of large planar driver that I find is often overlooked. And that is the minimization of vertical reflections due to a narrow dispersion. Contrary to Toole's research I have found in various AB tests that this matters a lot. Especially with various music material and recording of less quality. Besides a better imaging and insight into the recorded material the result is smoother sound, less harshness, and much less listening fatigue over time.

Working with acoustics over number years now, I have also seen that threating the ceiling reflections for wide vertical dispersion speakers are generally the most profound difference for many customers. More so then treatment at first reflection points on the side walls.
While we are more sensitive to what's coming laterally, I believe the reason for this is that vertical reflections almost in all rooms arrive very early and many speakers don't measure that well vertically. In rooms with beams or slanted ceiling, the vertical reflections also arrive from much more audible angles.
Anyway, it's one of the areas I disagree with Toole's findings. You will always stick with personal experience over a research. Perhaps listening over time with various music material is different vs short listening and especially with great recordings.

The narrow vertical dispersion is also what I personally like about electrostatic speakers. And the fact that one driver is covering either all or most of the frequency area leads to a very coherent sound stage with no break up in a senstive area. Clearly the speaker also needs to be accurate in tonality, without clear resonances, etc. in order to sound good, but that seems to be the case with the very best planar speakers. What I don't like about electostatics is the lack of dynamics and "punch". And the use of subwoofer doesn't close that gap 100% because there's a a need for better dynamics higher in frequency as well. The dipole design isn't my favourite either, sending just as much energy towards the rear as to the front. You need quality treatment on the front wall or place them far out. No speaker is perfect as we know.

As for distortion, I don't think traditional THD and IMD covers all what we hear. Dan Clarke said not too long ago that we can hear planar drivers sound cleaner, more detailed etc., but we don't really know today how to measure this. Maybe it's related to the fact that music containes more simultaneously tones compared to what our distortion measurements shows. I don't know. But in development of the Coherence 12, it was very clear the planar driver did something beneficial what we couldn't achieve with various pistonic drivers. Some is likely related to the dispersion (narrower and less room interaction), but most likely not all since we also heard it with a lot of acoustic treatment in the AB tests.

 

[RetroStereo]

I agree that the vertical dispersion pattern of large ESLs is an advantage (as it is with other large planar designs). However, I disagree that ESLs inherently lack dynamics and punch.
Back to the DW XG8 (and its successor, the XG10) - Mike Wright placed the stators inside an inert gas (sulfur hexafluoride) balloon in order to increase the stator gap and increase the driving charges on the panels without arcing. He had set them up in a home in Cary NC driven with a pair of Phase Linear 700 amps in mono mode. Those speakers had dynamics and punch equal to or better than any dynamic drivers I've ever heard.
I suspect the increased stator distance is one of the factors in dynamics and punch, but I think there is another: the amplifier driving the ESLs. Many amps simply cannot handle the capacitance load.
I'm using an amp design that doesn't get a lot of respect in some quarters but with ESLs, it excels. I'm driving each ESL panel with a Mcintosh MC-2255 in mono/parallel mode. This doubles the current and halves the impedance of the circuit. I have the Monolith panels connected to the 1 ohm tap. They drink the power and the amp behind Mac's big autoformers could care less about impedance drops at any frequency.
There may be other solutions or explanations, however this is working for me.

 

[NTK]

Back to the DW XG8 (and its successor, the XG10) - Mike Wright placed the stators inside an inert gas (sulfur hexafluoride) balloon in order to increase the stator gap and increase the driving charges on the panels without arcing. He had set them up in a home in Cary NC driven with a pair of Phase Linear 700 amps in mono mode. Those speakers had dynamics and punch equal to or better than any dynamic drivers I've ever heard.

Are you serious?

An SF6 "balloon" around the "stator" of the e-stat? How does the SF6 atmosphere transition to the normal breathable (i.e. life sustainable, not asphyxiating) atmosphere? You have a gas that is very different from normal atmospheric air (much denser, molecules with 7 atoms, as oppose to air which is ~99% nitrogen and oxygen, which are both diatomic, and the speed of sound in SF6 is about 40% of that of air). So the sound is propagating out from a highly physically and chemically non-uniform medium (from ??% SF6 to ~100% air), most probably with no control of how the gases are introduced/mixed/transitioned, and nobody raised any question?

 

[RetroStereo]

Are you serious?

An SF6 "balloon" around the "stator" of the e-stat? How does the SF6 atmosphere transition to the normal breathable (i.e. life sustainable, not asphyxiating) atmosphere? You have a gas that is very different from normal atmospheric air (much denser, molecules with 7 atoms, as oppose to air which is ~99% nitrogen and oxygen, which are both diatomic, and the speed of sound in SF6 is about 40% of that of air). So the sound is propagating out from a highly physically and chemically non-uniform medium (from ??% SF6 to ~100% air), most probably with no control of how the gases are introduced/mixed/transitioned, and nobody raised any question?

LOL! I didn't say it was practical; only that it worked. And I think they were manufactured for close to 20 years. I've spotted listings on eBay for used ones as recently as a few months ago.

Edited to add: 5 minutes with google will tell you probably more than you want to know. They were amazing and terrible at the same time. And they may have destroyed more amplifiers than any other speaker ever made. And, yes, they really contained sulfur hexafluoride. There was a "tire valve" at the bottom rear of the enclosure where the gas could be recharged from a cylinder available from welding supply sources.

 

[Floyd Toole]

I agree that the vertical dispersion pattern of large ESLs is an advantage (as it is with other large planar designs). However, I disagree that ESLs inherently lack dynamics and punch.
Back to the DW XG8 (and its successor, the XG10) - Mike Wright placed the stators inside an inert gas (sulfur hexafluoride) balloon in order to increase the stator gap and increase the driving charges on the panels without arcing. He had set them up in a home in Cary NC driven with a pair of Phase Linear 700 amps in mono mode. Those speakers had dynamics and punch equal to or better than any dynamic drivers I've ever heard.
I suspect the increased stator distance is one of the factors in dynamics and punch, but I think there is another: the amplifier driving the ESLs. Many amps simply cannot handle the capacitance load.
I'm using an amp design that doesn't get a lot of respect in some quarters but with ESLs, it excels. I'm driving each ESL panel with a Mcintosh MC-2255 in mono/parallel mode. This doubles the current and halves the impedance of the circuit. I have the Monolith panels connected to the 1 ohm tap. They drink the power and the amp behind Mac's big autoformers could care less about impedance drops at any frequency.
There may be other solutions or explanations, however this is working for me.

This has triggered a memory. Mike Wright brought a few prototypes to the NRCC to measure in our anechoic chamber. They made sound, played quite loud, but the sound they produced was measurably and audibly quite colored compared to the better cone and dome speakers of that period. So, his creative ideas needed refinement. He never returned, so I lost track of the product after that.

The attraction of electrostatic loudspeakers has been around for a very long time. The idea that the thin membrane is "faster" than relatively massive cones and domes has a superficial attraction - until it is realized that it has to be related to the strength of the motor driving the diaphragm. Electrostatic forces are small compared to those available from moving coil motors. I had a Tesla S that accelerated from 0 to 60 mph in about 3 seconds and it was a very heavy sedan- it's a "horsepower" thing. Transducers are minimum-phase devices, so the time domain response is predictable from the frequency response. There are no "speed" advantages to film diaphragms, including the Oscar Heil Air-Motion Transformers - he also visited me at the NRCC after I wrote an analysis of his driver showing that it was less perfect than promoted. The concept was good, but the materials of the time were not able to handle the heat.

To maximize the electrostatic forces the diaphragm displacement is limited by a small gap so diaphragms must be large to move the necessary volume of air at low frequencies. Enclosures create problems, so they are usually dipoles, radiating freely front and back. This adds "complexity" to the sound field that may or may not be appreciated. Large diaphragms are directional, so this becomes a distinguishing feature, and a problem, requiring subdivisions of the radiating surface, creating"woofer" and "tweeter" areas. The Quad ESL63 was a very elaborately subdivided panel, designed to radiate an approximate hemispherical wavefront. Unfortunately it had a large "tweeter", a 3 -inch circle, and limited bass output. It had felt pads in the back to damp the rear-radiated sound, so it was not a pure dipole. It was an ambitious design. I sat in the chair Peter Walker, the designer, used for voicing the speaker, and drank Scotch with him in his living room. A brilliant and very nice man.

Audio is a fascinating mixture of fact and folklore, with "opinions" supporting all possibilities. I have spent my life sorting out facts from fiction - and it has been rewarding and fun. The quest to be "different" is omnipresent, and it will be more difficult but won't diminish now that we have scientific guidance and competent engineers able to design loudspeakers approaching perfection much closer than ever before. That price is not a primary factor is a huge benefit for consumers. The "trickle-down" effect of knowledge is considerable. But marketing rubbish and folklore still abound making life more complicated than necessary for consumers who just want accurate sound.

I will say again - the bass response in the room cannot be solved by a stereo pair of loudspeakers and it is about 30% of our opinions of sound quality.

 

[pma]

The attraction of electrostatic loudspeakers has been around for a very long time. The idea that the thin membrane is "faster" than relatively massive cones and domes has a superficial attraction - until it is realized that it has to be related to the strength of the motor driving the diaphragm

I think the only good way to use an electrostatic driver is the horn loaded pressure speaker. As a student of professor Josef Merhaut in the seventies, I would dare to post this link:

US3590169A - Electrostatic horn loudspeaker - Google Patents

An electrostatic pressure loudspeaker with a diaphragm vibrating at the rhythm of a signal frequency actuated by electrostatic forces operating between the diaphragm and at least one back electrode. At least one side of the diaphragm is connected with a horn through the input cross section of...

patents.google.com

 

[DavidEdwinAston]

This has triggered a memory. Mike Wright brought a few prototypes to the NRCC to measure in our anechoic chamber. They made sound, played quite loud, but the sound they produced was measurably and audibly quite colored compared to the better cone and dome speakers of that period. So, his creative ideas needed refinement. He never returned, so I lost track of the product after that.

The attraction of electrostatic loudspeakers has been around for a very long time. The idea that the thin membrane is "faster" than relatively massive cones and domes has a superficial attraction - until it is realized that it has to be related to the strength of the motor driving the diaphragm. Electrostatic forces are small compared to those available from moving coil motors. I had a Tesla S that accelerated from 0 to 60 mph in about 3 seconds and it was a very heavy sedan- it's a "horsepower" thing. Transducers are minimum-phase devices, so the time domain response is predictable from the frequency response. There are no "speed" advantages to film diaphragms, including the Oscar Heil Air-Motion Transformers - he also visited me at the NRCC after I wrote an analysis of his driver showing that it was less perfect than promoted. The concept was good, but the materials of the time were not able to handle the heat.

To maximize the electrostatic forces the diaphragm displacement is limited by a small gap so diaphragms must be large to move the necessary volume of air at low frequencies. Enclosures create problems, so they are usually dipoles, radiating freely front and back. This adds "complexity" to the sound field that may or may not be appreciated. Large diaphragms are directional, so this becomes a distinguishing feature, and a problem, requiring subdivisions of the radiating surface, creating"woofer" and "tweeter" areas. The Quad ESL63 was a very elaborately subdivided panel, designed to radiate an approximate hemispherical wavefront. Unfortunately it had a large "tweeter", a 3 -inch circle, and limited bass output. It had felt pads in the back to damp the rear-radiated sound, so it was not a pure dipole. It was an ambitious design. I sat in the chair Peter Walker, the designer, used for voicing the speaker, and drank Scotch with him in his living room. A brilliant and very nice man.

Audio is a fascinating mixture of fact and folklore, with "opinions" supporting all possibilities. I have spent my life sorting out facts from fiction - and it has been rewarding and fun. The quest to be "different" is omnipresent, and it will be more difficult but won't diminish now that we have scientific guidance and competent engineers able to design loudspeakers approaching perfection much closer than ever before. That price is not a primary factor is a huge benefit for consumers. The "trickle-down" effect of knowledge is considerable. But marketing rubbish and folklore still abound making life more complicated than necessary for consumers who just want accurate sound.

I will say again - the bass response in the room cannot be solved by a stereo pair of loudspeakers and it is about 30% of our opinions of sound quality.

Can I simply thank you, for this brilliant post. I saw Peter Walker at a Harrogate hifi show. Would never have dared to impose. To this day, would like to try electrostatics, as the Jecklin Float Electrostatic headphones I owned were the best sounding transducers in my long hifi search.
Best wishes to you.

 

[ElJaimito]

I have a pair of Quad ESL, the original electrostatic loudspeaker. These are late editions from the end of their production run. To me, I have found nothing that has compared within reason. They are still in original condition and work fine. A couple of years ago I gave them a thorough cleaning and was advised that if they continue to work, no other things need to be done. I dread the day when they stop working. I have put a lot of hours on them this past year with the lock down and worry that I will miss them when it comes time to move back to the city.

They truly are remarkable and hard to believe they were developed in the 50’s.

I had a pair and they were just as you say. However they are very sensitive to room positioning and size (IMO) especially with regard to bass, and I ended up selling mine when I moved to somewhere they just didn't fit. Not a criticism of the speakers. I regret that as they were a wonderful reference.

 

[Floyd Toole]

I had a pair and they were just as you say. However they are very sensitive to room positioning and size (IMO) especially with regard to bass, and I ended up selling mine when I moved to somewhere they just didn't fit. Not a criticism of the speakers. I regret that as they were a wonderful reference.

The original Quad ESL had good on-axis performance, but suffered from large off-axis variations, making them sound very different depending on the shape, size and reflectivity of the listening room. Perfect satisfaction was probably not possible, but during the "adjustment" period, the human mind adapts to what it is. A very dead room or close listening would be beneficial. The BBC would have been pleased in its very dead, reflection-controlled studios.

The ESL63 was a very different product, and Peter Walker addressed the inconsistent directivity with a complex design that preserved the desirably good direct sound while improving the timbre of the reflected sounds. It is an example of creative thinking and good engineering. It still was not a "perfect" loudspeaker, but in 1963 it was an achievement of merit.

Here is a comparison taken from Figures 1.3 and 4.7 in the 4th edition of my book.

 

[Keith_W]

Slide from my upcoming talk on bass for my local audio club. Hope you don't mind Dr. Toole, and keep those quotable quotes coming!

 

[kimmosto]

I will say again - the bass response in the room cannot be solved by a stereo pair of loudspeakers and it is about 30% of our opinions of sound quality.

Define "room", "loudspeakers" and "solved" please. What kind of living spaces, listening rooms; constructions and acoustics of the houses/buildings are included, what speaker and directivity concepts are included, what size categories are included, and how small errors and where is "solved"?

 

[Floyd Toole]

Define "room", "loudspeakers" and "solved" please. What kind of living spaces, listening rooms; constructions and acoustics of the houses/buildings are included, what speaker and directivity concepts are included, what size categories are included, and how small errors and where is "solved"?

Are you serious? Have you not seen the slide show, followed my postings on this forum, or read my book. All your questions have answers but right now I don't have the hours necessary to repeat what I have written. I suggest a forum search on my name and have a look. Dealing with bass problems alone takes two chapters in my book.

 

[kimmosto]

Are you serious? Have you not seen the slide show, followed my postings on this forum, or read my book. All your questions have answers but right now I don't have the hours necessary to repeat what I have written. I suggest a forum search on my name and have a look. Dealing with bass problems alone takes two chapters in my book.

I have your book, but I'm quite sure we have been living in different reality for more than two decades. I don't accept limited view to reality such as assumption about very sealed cubical rooms and conventional speaker concepts such as quite omnidirectional vented or sealed direct radiators at LF. That could be (unfortunate) reality for most of hifi consumers and investigators requiring some tightly specified room for simplified theories and tests, but it does not represent all possibilities and everyday practice for many of us.

 

[goryu]

I think the work done over decades by Dr. Toole is commendable. The theory is well reasoned and the results are supportive. In practice, we know from the marketplace that many successful products do not align well with the criteria said to be "most preferred" from the experimental data. We also know that less than technically accurate amps have a large and loyal following. Would the owners of these products choose them in Dr. Toole's lab tests? From the data, likely not. In the real world, many (most?) people don't choose products based on test results, measurements, science, or logic. There are often factors that are important that have nothing to do with audio performance- form factor, eye appeal, cost, warranty, etc., all play a differing role in purchasing decisions. Let's not forget the important role that marketing plays- a good story will sell more than a good test result in many cases. The marketplace is full of debunked quackery based products that just keep on selling.

With the importance of one's room, the 30% figure quoted above about low frequency performance, dsp and room treatment options, surely there must be a fairly wide variation in the factors said to be critical in a loud speaker for "preferrable preformance" that would allow for some remediation in one's room.

I will continue to use measurements as a tool to winnow the candidates but the final decision will always be my own subjective preference when heard in my room with my other components.

 

[Floyd Toole]

I have your book, but I'm quite sure we have been living in different reality for more than two decades. I don't accept limited view to reality such as assumption about very sealed cubical rooms and conventional speaker concepts such as quite omnidirectional vented or sealed direct radiators at LF. That could be (unfortunate) reality for most of hifi consumers and investigators requiring some tightly specified room for simplified theories and tests, but it does not represent all possibilities and everyday practice for many of us.

The entire purpose of our research has been to understand the relationships between what is measured in an anechoic chamber or in the listening room at the listener's location, and what is heard. No special assumptions were made about the construction or shape of the room, nor about the physical configuration of the loudspeaker design. In reality what you call "everyday practice" has been the goal from the outset, otherwise it is all just an academic exercise. The scientific literature is full of such studies, but I am a practical engineer at heart and solving real-world problems is the only sensible goal.

You said: "tightly specified room for simplified theories and tests" make my results unrepresentative of what "many of us" are interested in. The scientific method requires that experimental variables be controlled so that when an experiment is done the changes in subjective opinions can be attributed to one variable at a time. That is how one learns what is and is not important, and what needs to be done to improve a problematic situation.

As pointed out in the slide show, in normal listening situations there are several variables simultaneously influencing listener opinions, not even considering the numerous non-auditory factors involved in sighted tests. Over the years we have investigated and conducted costly and time-consuming experiments on the separable influences of loudspeakers, listeners, listening room shape and size, and more. There is nothing in these tests that make the results inappropriate for any normal consumer, including the "reality" that I suspect you live in. How is it different from the average audio enthusiast?

The entire purpose has been to provide guidance in selecting timbrally neutral loudspeakers, and providing methods by which the bass problem in rooms can be alleviated for one or more listeners. Doing so well requires measurements. Anechoic data on the loudspeakers. In-room measurements at low frequencies in the listening room. The 4th edition of the book provides the latest insights of me and my co-authors.

 

[Floyd Toole]

Let's not forget the important role that marketing plays- a good story will sell more than a good test result in many cases. The marketplace is full of debunked quackery based products that just keep on selling.

Yes indeed! I have known several audio salespeople over the years and they take pride in being able to sell what they need or want to move to a customer. The strategies are quite clever. In the early years, a few salespeople participated in my listening tests, and brought products for evaluation, which was wonderful for the growing database. Some did not last, though, commenting that what they had learned interfered with their ability to sell the products that the store stocked..

Indeed price and appearance matter to a buyer. I see the knowledge we now have as a huge advantage because competent engineers can design predictably good sound from loudspeakers of many different visual appearances and prices. Spinorama data shown by Amir and Erin, and Soundstage's NRC anechoic data confirm it.

 

[mps]

The original Quad ESL had good on-axis performance, but suffered from large off-axis variations, making them sound very different depending on the shape, size and reflectivity of the listening room. Perfect satisfaction was probably not possible, but during the "adjustment" period, the human mind adapts to what it is. A very dead room or close listening would be beneficial. The BBC would have been pleased in its very dead, reflection-controlled studios.

Thanks, Dr. Toole. Does this explain my experience that the best way to listen to my ESL57s is in the near-field? (Not that easy with such big speakers )

 

[mrehorst]

It's not that hard to set up ESLs for near field listening...

 

[norman bates]

Lol