My Takeaways From ASR Speaker Testing

[Blumlein 88]

Is it possible for any dipole (or bi-pole, for that matter) to get a great set of spin-o-rama measurements?

I have found that listening seat measurements do seem to loosely predict which speakers I might like. I heard the MLs before I read JA's review (the dealer handed me a copy), I can see why I probably preferred my own Salon2 configuration. I thought they made some instruments sound awesome (horns and strings, for example), but I didn't care so much for them with solo piano or a drum kit. My takeaway at the time was they might be euphonically colored. After seeing the JA review I did think that flat response curve is not something I've ever found realistic sounding since I started measuring perhaps ten or twelve years ago. I think the so-called Harman Curve sounds more realistic with my own recordings.

I've been tempted by electrostatics a couple of times, occasionally with MLs and especially with Sound Labs, and a long time ago with Quads. I just never pulled the trigger.

Which Sound Labs?

It would certainly be possible for an ESL to be made that had excellent directivity in the horizontal direction. Imagine some tall panels with delays in sections like how ESL-63's were done. Or a more segmented ESL-63 speaker. Those could have directivity which was designed to ace the target for a good rating by Harman criteria. I suppose you would have to dampen the rear as well.

I don't necessarily disagree that panels are colored in a sense that favors some music over others. Of course I'd like to do the Harman blind listening session with 3 good box speakers and say a Quad ESL-63 just to hear what is what without seeing the speaker. Thought they've no commercial reason for doing so, I'd like to see Harman or someone using their methods do some shoot outs with only panel speakers to see which panels are most preferred.

I've some Aura's. They are A3/M3 sized with a heavy steel frame Soundlab only made a short while. The frames were too costly.
This isn't mine, but mine look just like this.

 

[MattHooper]

It would certainly be possible for an ESL to be made that had excellent directivity in the horizontal direction. Imagine some tall panels with delays in sections like how ESL-63's were done. Or a more segmented ESL-63 speaker. Those could have directivity which was designed to ace the target for a good rating by Harman criteria. I suppose you would have to dampen the rear as well.

I don't necessarily disagree that panels are colored in a sense that favors some music over others. Of course I'd like to do the Harman blind listening session with 3 good box speakers and say a Quad ESL-63 just to hear what is what without seeing the speaker. Thought they've no commercial reason for doing so, I'd like to see Harman or someone using their methods do some shoot outs with only panel speakers to see which panels are most preferred.

I've some Aura's. They are A3/M3 sized with a heavy steel frame Soundlab only made a short while. The frames were too costly.
This isn't mine, but mine look just like this.

View attachment 58887

A soundlab owner! Very cool!

I enjoyed the various soundlab speakers I've heard. (And at one point I auditioned the really big full range ones...sorry, can't be arsed to look up the name). That was quite an experience.

I think it would be fascinating to do a blind shoot out with Quads at the HK labs as well.

I did a mini blind shoot out with the Quad ESL 63s and Spendor BC1 speakers at a pal's house. (He owned the '63s and later I owned them as well). It was just me blindfolded and him switching the signal between the two speakers. But it was fascinating how similar they actually sounded. It was ultimately the "boxier" character of the Spendors that seemed to give their identity away vs the completely "box free" sound of the Quads, which is of course a big part of the appeal of electrostatics.

 

[MattHooper]

Somewhat ironically, I actually think the reviews have given me some clarity on subjective reviews or the subjective portions of reviews (including my own).

What I've noticed from reading some subjective reviews after measurements.

• The measurements make me think a fair many subjective reviews get the primary sound descriptions right. Or, at least, reading two or three subjective reviews of a speaker can give you a pretty decent idea of sound characteristics that seems to match measurements. This speaker sounds bright, this speaker has a forward soundstage, that speaker has recessed mids. These aspects seem to roughly track the measurements from the samples I've read.
• There are details subjective reviews can be useful, as even the most extensive measurements out there are not completely comprehensive due to simple time constraints. SPL levels is a big one.
• Subjective reviews seem to fail in two or three places: First is value assessment. They might be able to characterize a speaker's sound, but not so much whether you can get the same sound quality for much less. Conversely, they might assume a cheap speaker can't possibly sound as good as a significantly more expensive one.
• It also shows the importance of room EQ below the transition frequency, as it can really affect the perception of the other frequencies.
• Even with measurements, people can disagree on the interpretation.

So I think if you can sift through some of the dramatic review language (which I admittedly enjoy sometimes) and read multiple reviews there's some value there. I'd try to read while ignoring the value assessments.

I agree with your points, and have argued for this numerous times here.

A perceptive subjective reviewer can do a pretty good job of describing the character of a speaker. Many, many times I have auditioned speakers and, whether I read a review before or after my audition, I find that reviewer had done quite a good job putting what I heard in to words.
(Certainly not all reviews/reviewers!). And though many here think subjective reviews get too flowery, reaching beyond mere workable terms like "bright" or "recessed," I find many of the characteristics subjective audiophile reviewers describe DO comport with the subjective effects of speakers, as I experience them. (Again...not all the time...but I get what they are referring to).

On the other hand, I understand that when someone becomes versed enough in understanding the measurements and correlating them exactly to what he hears, that person will have little need for someone else's subjective descriptions so long as the right measurements are supplied.

But not everyone is in that position, so subjective description is one way audiophiles can pass information among one another. It's not as reliable as objective measurements, but it's not rendered wholly useless, and sometimes I think it can go beyond, "documenting" a richer variety of the subjective effects of speakers than one tends to find in purely measurement-oriented discussions.

 

[PaulD]

Hey!

I resemble that remark!

Not certain how to interpret it, though.

Can't decide it it is one thought or two, since they are, if there are two, co-mingled in the same paragraph.

I know where you're coming from Ray, I resemble it too. I'm lucky enough to listen to JBL M2s at work daily in a nicely engineered studio (controlled reflections etc), and Quad electrostatics with subs at home (pretty uncontrolled domestic acoustics). To be honest the similarities far FAR outweigh the differences, they both sound superb, tonal balance is practically the same after a little EQ to both, etc etc.

 

[tuga]

Looking at all the measurements and what result you might get in a spin-o-rama graph those are not good looking measurements. They look pretty poor.

Do you mean these horizontal response measurements?
And are you taking into account that the vertical scale is only 30dB wide?

I see nowt to complain except perhaps a bit too much "air" (see green curve), at 45º the curve is flat.

I know Simon says, sorry Science says that we don't like a narrow dispersion, but they are wrong.

MartinLogan Renaissance ESL 15A, lateral response family at 50",
responses on mid-panel axis (black) and at: 15° (orange), 30° (green),
45° (blue), 60° (magenta), 75° (purple), 90° (turquoise) off axis.

source: https://www.stereophile.com/content...-renaissance-esl-15a-loudspeaker-measurements

 

[tuga]

1. Knowledge is power. We are much better off with these measurements than without.

I agree. I am grateful for @amirm 's efforts.

2. Expect the unexpected. There have been some surprises like the KEF R3 topping the field in passive speakers while the LS50 measured meh.

I feel that speakers should be divided into classes according to the performance potential of the topology (bass extension, max SPL).

3. Controversy is part of life. There was a lively discussion about why @amirm liked the M16 better than the R3. The mystery was solved with a little EQ.

Also it must be taken into account that these are sighted tests after seeing the measurements in not the best of conditions nor using the best methodology.

4. Don't frorget the preference index is only one number. It can be adjusted for adding a sub, but not for low output. Speakers with 4" mid/low drivers can't play loud.

The preference index is based on the false(?) premise that most people like a particular kind of speaker and thus that this is somehow relevant to you/me.

5. A little EQ can work miracles. My LS50's sound a lot better with some adjustments to deal with known frequency response deviations. Could preference scores be recomputed for EQ on some kind of a consistent basis? Conversely, bad dispersion can't be easily fixed. Do we need a separate dispersion index?

It might be possible to add a note to indicate whether a speaker's response is EQ'able.

Many aspects of performance can be antecipated just by looking at a photo of the speaker: EQ'ability is one, dispersion characteristics, max SPL, LF extension.

6. Some people must be really fortunate to have a perfect room. The Klippel shows simulated room response in what looks like a perfect room. My room goes nuts between 100 and 200Hz no matter what I put in there. Remember that uneven response lead to EQ-ing the R3's with a big improvement in sound.

What about how much your room impacts the sound above 500Hz, more so with wide-dispersion speakers?

This problem seems to be glossed-over for some reason...

7. I believe the science behind this is good. If you don't like the science, take the argument somewhere else.

The science is good, but not the use opinion polls to influence taste.

8. Never forget the result is about preference of A over B. Not everyone has the same tastes. That's why electrostatics and Maggies remain popular.

In other words, ignore the preference ratings or risk ignoring your own taste. Use the measurements to shortlist speakers worth listening to, then get off your bum and go listen.

 

[tuga]

It would certainly be possible for an ESL to be made that had excellent directivity in the horizontal direction. Imagine some tall panels with delays in sections like how ESL-63's were done. Or a more segmented ESL-63 speaker. Those could have directivity which was designed to ace the target for a good rating by Harman criteria. I suppose you would have to dampen the rear as well.

Is this not "excellent directivity in the horizontal direction"?

MartinLogan Montis, lateral response family
source: https://www.stereophile.com/content/martinlogan-montis-loudspeaker-measurements

 

[Juhazi]

For dipoles, multiway with dsp can give nice results https://www.diyaudio.com/forums/mul...aborative-speaker-project-56.html#post5003836

 

[Frank Dernie]

I am thrilled these measurements are being done.
I don't take any notice of any single factor of goodness, personally.

 

[Marakk]

Ignoring the ~2kHz dip, it’s sensitivity is ~89dB, which makes it excellent for home theater purposes.

Looking at it‘s normalized horizontal directivity, it is really well controlled, thus making it easy to EQ (it has low distortion ~2kHz), and since the majority of surround sound setups (ignoring HTiBs) employ the included room correction, it will sound pretty good.

Sure. Those are the saving graces, but the EQ recommendation doesn't seem to be a thing before we had these measurements, and there are a lot of builds without receivers that have room correction. I don't know percentages, but I can find you at least 10 instances of these speakers being used without the same via various threads on forums and reddit, if not more. Like for example those who saw rave reviews and upgraded just their speakers and used the same receiver they had maybe because they didn't know that room correction is something they really needed with these speakers.

Accurate measurements are now there for all of us to take calculated decisions. That includes work being done Rtings to measure soundbars. Some of those might actually sound good. Q80/Q90 for example measure really well, and as far as I can judge from the SFR measurements, better than Elac Debut 2.0 that Rtings had tested and posted on reddit.

So even if people want to guide people away from soundbars, using the work being done here, better recommendations can be given for each, that'll ensure that that speaker is actually good. Micca RB42 for example aren't that good. Pairing those up with a receiver might cost 200-250$, but is that an obvious choice over a soundbar? Vizio budget model and even TCL 7+ measured decently neutral. Of course, we can't correlate measurements here with Rtings directly, but my point was that even if people have to guide people away from soundbars, minimum recommendation should be for the Pioneers instead of the Micca.

 

[Hayabusa]

6. Some people must be really fortunate to have a perfect room. The Klippel shows simulated room response in what looks like a perfect room.

It would be great if after the measurement the Klippel sw could show a simulated room response with a room that actually matches your/my room.

 

[tuga]

It would be great if after the measurement the Klippel sw could show a simulated room response with a room that actually matches your/my room.

Maybe @amirm could pair up with Hunecke: https://www.hunecke.de/en/calculators/loudspeakers.html

Some years ago if you went to the Stereoplay (if I remember correctly) website the drop-down menus in that simulator would alow you to choose specific loudspeakers that I presume had been measured by the magazine, so it can be done.

 

[Blumlein 88]

Is this not "excellent directivity in the horizontal direction"?

MartinLogan Montis, lateral response family
source: https://www.stereophile.com/content/martinlogan-montis-loudspeaker-measurements

It is pretty good. But off axis is more choppy than these F228BE speakers as a comparison. And that off axis choppiness in one of the audible differences from the work Harman has done. That choppiness erodes how well speakers score in their blind auditioning. I find that surprising, but not unbelievable. And having worked with a number of ESL's that Monti is probably far better than most of them in that regard.

 

[Blumlein 88]

Do you mean these horizontal response measurements?
And are you taking into account that the vertical scale is only 30dB wide?

I see nowt to complain except perhaps a bit too much "air" (see green curve), at 45º the curve is flat.

I know Simon says, sorry Science says that we don't like a narrow dispersion, but they are wrong.

MartinLogan Renaissance ESL 15A, lateral response family at 50",
responses on mid-panel axis (black) and at: 15° (orange), 30° (green),
45° (blue), 60° (magenta), 75° (purple), 90° (turquoise) off axis.

source: https://www.stereophile.com/content...-renaissance-esl-15a-loudspeaker-measurements

Yes, the issue is the off axis is up and down too much. Not smooth enough, and on axis is a bit too up and down as well.

Now using the Harman results this is a bad thing when people listen to it vs a smoother responding speaker. You obviously don't like the Harman work, but who else has done such things? I haven't. I have lived very happily with ESL most of my adult life. I think in light of these preferences found which seem to be near universal, that more effort for a panel speaker to keep its good attributes and be smoother off axis would be a way forward. I think some of the choppiness is because the whole panel moves. So segmented by frequency panels might let you have your cake and eat it too.

 

[tuga]

It is pretty good. But off axis is more choppy than these F228BE speakers as a comparison. And that off axis choppiness in one of the audible differences from the work Harman has done. That choppiness erodes how well speakers score in their blind auditioning. I find that surprising, but not unbelievable. And having worked with a number of ESL's that Monti is probably far better than most of them in that regard.

View attachment 59000

If I'm not mistaken @John Atkinson believes that the choppiness is a measurement artifact which results from the closeness of the mic to the membrane. All measurements of ESLs show it.

PSB has done a much better work than Harman:

https://www.stereophile.com/content/psb-imagine-t3-loudspeaker-measurements

 

[Wes]

AFAIK< it is very difficult to measure dipoles. Most of us lack anechoic chambers too - tho you can take a speaker outside...

 

[Blumlein 88]

If I'm not mistaken @John Atkinson believes that the choppiness is a measurement artifact which results from the closeness of the mic to the membrane. All measurements of ESLs show it.

PSB has done a much better work than Harman:

https://www.stereophile.com/content/psb-imagine-t3-loudspeaker-measurements

Yes, I've heard that for years, and it makes sense. You'd still get some additional choppiness at the LP vs a smoother speaker like the PSB you show.

So what is your explanation for the panel doing poorly in a Harman style comparison? My most damming critique is they haven't done (to my knowledge) tests with other panels like Maggies, and other ESL's than the one ML speaker (which I thought was not one of their best as it didn't integrate with the wooofer very well when I've heard them). One reason I've liked panels is they don't have a box so you don't hear a box. Your better modern speakers don't have much of a boxy sound vs 20 or 30 years ago.

 

[MattHooper]

One reason I've liked panels is they don't have a box so you don't hear a box. Your better modern speakers don't have much of a boxy sound vs 20 or 30 years ago.

Exactly. This is one aspect that seems to go missing when people are arguing about measurements saying "this speaker is clearly poorer than the other." As individuals we can latch on to certain aspects of a presentation we value over others. If you find you notice the box coloration of box speakers, it it becomes something that bugs or distracts you, that very lack of artificial box sound can be a haven for you with panel speakers, and override other considerations to sound "better." (And it can go the opposite way too...which is why I moved away from panels, myself).

On the other hand, the research by Harmon et all has to be taken seriously at least in terms of blind test results, but I'm with you in that I'd love to have seen a wide variety of panel speakers tested.

 

[blueone]

One reason I've liked panels is they don't have a box so you don't hear a box. Your better modern speakers don't have much of a boxy sound vs 20 or 30 years ago.

Strongly agree, but I think there's more to the ESL enigma than lack of a box. I've always thought it had something to do with the comparatively massive wave launch surface versus point sources. Whatever the reasons are, on some music (especially acoustic jazz), the best ESLs have always attracted me.

The Quad ESL-63s did not endear themselves to me. I'm a solo piano fan, and the Quads just didn't have enough bass to provide a realistic illusion.

 

[tuga]

Yes, I've heard that for years, and it makes sense. You'd still get some additional choppiness at the LP vs a smoother speaker like the PSB you show.

So what is your explanation for the panel doing poorly in a Harman style comparison? My most damming critique is they haven't done (to my knowledge) tests with other panels like Maggies, and other ESL's than the one ML speaker (which I thought was not one of their best as it didn't integrate with the wooofer very well when I've heard them). One reason I've liked panels is they don't have a box so you don't hear a box. Your better modern speakers don't have much of a boxy sound vs 20 or 30 years ago.

That could be a possibility, people not being used to listening to panel speakers. People who enjoy and are used to listening to small standmounts tend to find large speakers "slow" and not as "spacious-sounding".

The Quad didn't do very well in mono but did a lot better in stereo.