From what I have read all the test showed was that in a specific spot listeners preferred non-esl speakers to a esl speaker, with certain music.
Nothing more.
And that is the biggest problem with ELS (and dipoles) - positioning is key. And not just the speakers position, but also your ears. To some this is a huge deterrent, to others it is a non-issue.
In full disclosure, I have MartinLogan Vistas.
I recently spent 3months with a fantastic pair of Thiel MCS1 speakers in the exact same location. (same price as the Vistas) While they did some things great, and they are a fantastic speaker...I prefered the sound coming from the ML's.
Sure the Thiels had a wider sweet-spot, took considerably less time to setup and were not nearly as directional in the HF. In the end I choose the MLs.
Nothing more.
And that is the biggest problem with ELS (and dipoles) - positioning is key. And not just the speakers position, but also your ears. To some this is a huge deterrent, to others it is a non-issue.
In full disclosure, I have MartinLogan Vistas.
I recently spent 3months with a fantastic pair of Thiel MCS1 speakers in the exact same location. (same price as the Vistas) While they did some things great, and they are a fantastic speaker...I prefered the sound coming from the ML's.
Sure the Thiels had a wider sweet-spot, took considerably less time to setup and were not nearly as directional in the HF. In the end I choose the MLs.
Don't think you can dismiss Toole and Olive that breezily. Their work is pretty clean and comprehensive.
But there should be evidence in their data if seat location had much effect, at least for any ESL speakers. BTW, they give the anechoic frequency plot and off-axis for "M"... it will make you vomit.
Hard to tell from little pictures, but their seating area wasn't like a movie theater but with a small number of seats mostly towards the center. In the ESL beam? Might be possible to tell from information in the book and articles.
Martin Logans and most any other ESL (and ionic) - starting with Janzens in 1960s - have always sounded great to me. I guess they tickle what I think is most important sub-set of sound quality.
Last edited:
I do not doubt that, and I was not dismissing them, only the tests conditions as being perfect. What we hear is influenced by the room and speakers interaction. I saw nothing in the report that indicated that the found the best spot for every speaker under test. Instead each speaker was tested in the same spot.
The problem with ESL/positioning is not just about horizontal dispersion (accounted for by sitting inside the beam) but also with the vertical positioning. Moving your head up and down can change the HF response noticeably. This will affect both 'heard' and 'measured' responses. Which may explain why people either love or hate ESL's.
Me too! When they get it right it is beautiful. Problem being, it is easier to get them wrong than right!
Easy to see that JRace knows what he is talking about (and posts with style). Therefore, the issue might be one of degree of "trust" in their research.
I like ESLs but I have sympathy for the problems of doing this kind of human measurement beyond criticism. My most charitable take on the situation is that maybe the cumulative effect of small biases in their method add up to low ratings for "M".
I like ESLs but I have sympathy for the problems of doing this kind of human measurement beyond criticism. My most charitable take on the situation is that maybe the cumulative effect of small biases in their method add up to low ratings for "M".
a speaker with flat on axis anechoic HF (last two octaves) response often sounds like dog poop in the real world. Ymmv.
I'm not trying to say that the ML speakers are any good, in case anyone thought that was the case.
But if you make a test with X number of standard speakers and one (not that great) ESL, that doesn't seem quite right either... what does this tell us that is of use??
_-_-bear
I'm not trying to say that the ML speakers are any good, in case anyone thought that was the case.
But if you make a test with X number of standard speakers and one (not that great) ESL, that doesn't seem quite right either... what does this tell us that is of use??
_-_-bear
Hello Bear
That cost has no bearing on what sounds good. It's the design and total system integration that determines the outcome. It was an obvious scapegoat in that respect.
They don't use just one type of recording. I can understand some of the points Sy was making about high SPL as an example but these tests are done at a relatively benign 70 or so SPL levels. They also all get the same program material so all have the same chance to either shine or burn.
I tend to agree that the tests should not be taken out of context. However under the test conditions, providing it is under the same test conditions a generalization can be made. I don't see any mass condemnation of all ESL's just a this speaker was "prefered" under these test conditions.
I don't agree that on axis is not important or that you won't be able to hear it. It's your first arrival and if it's CD your first reflections.
Rob
Last edited:
It does appear Toole has discovered what most people with ears have known for years ..
1. M speakers are horrible ( well not all... some are OK ) ....
2. It is impossible for a dipole to work in the same acoustic space as a monopole without proper acoustic treatment , so the test is flawed based on what has been discussed.
3. Aggregate measured response is more important than just measuring on axis and playing into the last 2 "octaves"..
4. The dead last jazz trio had to have included John Pizzarelli ...
1. M speakers are horrible ( well not all... some are OK ) ....
2. It is impossible for a dipole to work in the same acoustic space as a monopole without proper acoustic treatment , so the test is flawed based on what has been discussed.
3. Aggregate measured response is more important than just measuring on axis and playing into the last 2 "octaves"..
4. The dead last jazz trio had to have included John Pizzarelli ...
I've begun to notice a theme in speaker technology evaluations. It's widely acknowledged that dynamic driver-based speakers differ wildly in the quality of their design and in the quality of their implementation, but it seems to be assumed that all electrostatic speakers are fundamentally similar. This really is a baseless assertion/implication. I say this as a diy builder of ESLs, and one who thinks ESLs have a lot to offer. There are good ESLs and bad ESLs just as there are good three-way dynamic speakers and bad ones. If one ESL performs well or badly in a test, that does not by itself imply that all ESLs share the same characteristics. Some ESLs have carefully controlled resonances while others buzz and rattle. Some ESLs are designed for broad directivity, others are designed to beam as much as possible. Some use transformers designed for a wide dynamic range and others saturate at rather low levels. Why lump all ESLs into the same basket?
It's been my experience that great dynamic speakers are better than poor ESLs and great ESLs are better than poor dynamic speakers. When comparing the best examples of each type, it's less clear that there's a consistent winner. Each technology is well suited to a particular application, but these blanket statements about speaker technologies seem to me not to move us in a useful direction. Unfortunately, the hifi rags are the worst purveyors of this kind of oversimplification: ESLs are anemic; dynamic drivers lack resolution; ported bass systems are boomy; horns are honky; metal tweeters are harsh. Perhaps generalizations like these sell magazines but they don't move us all toward better reproduction of music.
I think Toole's observations are very useful, we just need to be careful when extrapolating his results.
Few
It's been my experience that great dynamic speakers are better than poor ESLs and great ESLs are better than poor dynamic speakers. When comparing the best examples of each type, it's less clear that there's a consistent winner. Each technology is well suited to a particular application, but these blanket statements about speaker technologies seem to me not to move us in a useful direction. Unfortunately, the hifi rags are the worst purveyors of this kind of oversimplification: ESLs are anemic; dynamic drivers lack resolution; ported bass systems are boomy; horns are honky; metal tweeters are harsh. Perhaps generalizations like these sell magazines but they don't move us all toward better reproduction of music.
I think Toole's observations are very useful, we just need to be careful when extrapolating his results.
Few
Last edited:
Strikingly true comments from Few.
I would differ in that I hear (of think I hear) an ESL common thread of cleanliness (or something) even if ESLs have various faults. These faults may or may not be in the same families as cone drivers and may be easy for some of us to live with, if we choose to do so (like some speakers beaming or like occasional buzzing or popping on some speakers which can be ignored as easily as a car horn honk outside).
Pushing air using Saran Wrap is fundamentally different than pushing it using a relatively massive block of cardboard. I don't think it is much of a stretch to think the sound propagated is detectably different. But as I said earlier, maybe I should listen to modern trick cone-and-dome systems.
ESL builders have it harder in examining the (high voltage) signal to the driver or to relate that to acoustic output. Also the matching transformers may do tricks to the signal more ornery than any low-level crossover in a multi-driver cone speaker. BTW, I am a true-believer in direct-drive HV amps for ESLs.... the final frontier.
I would differ in that I hear (of think I hear) an ESL common thread of cleanliness (or something) even if ESLs have various faults. These faults may or may not be in the same families as cone drivers and may be easy for some of us to live with, if we choose to do so (like some speakers beaming or like occasional buzzing or popping on some speakers which can be ignored as easily as a car horn honk outside).
Pushing air using Saran Wrap is fundamentally different than pushing it using a relatively massive block of cardboard. I don't think it is much of a stretch to think the sound propagated is detectably different. But as I said earlier, maybe I should listen to modern trick cone-and-dome systems.
ESL builders have it harder in examining the (high voltage) signal to the driver or to relate that to acoustic output. Also the matching transformers may do tricks to the signal more ornery than any low-level crossover in a multi-driver cone speaker. BTW, I am a true-believer in direct-drive HV amps for ESLs.... the final frontier.
Last edited:
I'll try to hit all the notes - with a hammer?
I didn't mention cost in my comments... not sure why it is relevant?
Toole's chart shows that one of my fave's "pink noise" scored highest on discernability. BUT I know that pink noise's primary usefulness is in discerning frequency response variations, nothing more.
IF as stated the tests were run at ~70db SPL, it seems like a "cheat" since differences in speakers become larger as the SPL goes up, since the distortion levels go up (usually in a non-linear fashion).
IF we do not really know the precise test condtions, so that WE could replicate the tests, then as is my oft repeated criticism of all of these tests, they are not scientifically correct as published.
Few, I agree, the results from these test are not generalizable in the main - one can draw basic conclusions that might apply to "mid level" (the high-end of Harmon?) commercial designs - which is Toole's main thrust it seems to me.
markusG, it seems to me that this is not a comparision of ESLs vs. dynamic speakers at all... what is in question is Toole's test methods, protocols and most importantly results - results since he and others like to draw conclusions from said results. I always question the conclusions because the results are based on a test condition that is so unclear as to make WHAT is really being tested completely and absolutely uncertain. (yeah, they're testing "speakers"... )
_-_-bear
I didn't mention cost in my comments... not sure why it is relevant?
Toole's chart shows that one of my fave's "pink noise" scored highest on discernability. BUT I know that pink noise's primary usefulness is in discerning frequency response variations, nothing more.
IF as stated the tests were run at ~70db SPL, it seems like a "cheat" since differences in speakers become larger as the SPL goes up, since the distortion levels go up (usually in a non-linear fashion).
IF we do not really know the precise test condtions, so that WE could replicate the tests, then as is my oft repeated criticism of all of these tests, they are not scientifically correct as published.
Few, I agree, the results from these test are not generalizable in the main - one can draw basic conclusions that might apply to "mid level" (the high-end of Harmon?) commercial designs - which is Toole's main thrust it seems to me.
markusG, it seems to me that this is not a comparision of ESLs vs. dynamic speakers at all... what is in question is Toole's test methods, protocols and most importantly results - results since he and others like to draw conclusions from said results. I always question the conclusions because the results are based on a test condition that is so unclear as to make WHAT is really being tested completely and absolutely uncertain. (yeah, they're testing "speakers"...
)_-_-bear
Last edited:
"Taste" - now that is a very slippery slope.
Whatever significance it could have in whatever framework of discussion, it certainly isn't helpful when discussing which car motor gives the best torque. Likewise, while "taste" in car applications (say, 4-wheel off-road versus sport driving) obviously dictates torque requirements, it tends to be a discussion which is pretty self-evident once the terms of the application are set.
A major successful thrust of Toole's work is to push non-debatable concepts like "taste" further to the background.
Last edited:
Baloney
. It's all well documented what the room conditions are like. You should be able to build the room although I can't imagine why you would think that it is necessary to do so.https://docs.google.com/present/view?id=dgkrf7p2_85fxbrc4d7
Audio Musings by Sean Olive: The Harman International Reference Listening Room
http://www.harman.com/EN-US/OurComp...te Papers/HarmanWhitePaperMLLListeningLab.pdf
Rob
Last edited:
ben,
The HV signal isn't particularly hard to measure or otherwise discern... and the transformer's behavior is measureable, probably not more of a contributing factor than any transformer, or perhaps inductor in an xover... and the distributed force over the surface of a very light membrane (not "saran wrap" but ok...) is not an unknown...
The way that the ESL radiates sound - generally a larger flat surface - compared to a dynamic driver is important since the polar response and amount of energy per unit of surface area is typically different. The breakup modes are also somewhat different.
The problem (if there is one) with direct drive is matching the impedance to the panels properly... ymmv.
_-_-bear
The HV signal isn't particularly hard to measure or otherwise discern... and the transformer's behavior is measureable, probably not more of a contributing factor than any transformer, or perhaps inductor in an xover... and the distributed force over the surface of a very light membrane (not "saran wrap" but ok...) is not an unknown...
The way that the ESL radiates sound - generally a larger flat surface - compared to a dynamic driver is important since the polar response and amount of energy per unit of surface area is typically different. The breakup modes are also somewhat different.
The problem (if there is one) with direct drive is matching the impedance to the panels properly... ymmv.
_-_-bear
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.