What role should listening play in speaker design

[a4eaudio]

Are studio monitors also designed without listening?
I mean what In their design process make them more accurate than most hi fi speakers?

I have not seen a credible post that there exists a single speaker that the designer doesn't listen to before releasing it.

 

[Sal1950]

Tell everyone that you have turned over a new leaf. Say that in your younger days, you believed in oscilloscopes, measurements, low THD, etc. but that was not the path to sound quality. Now that you are older, you have realized that superior sound is not just measurements, it must involve the listening experience".

Haven't I heard much the same in some PS Audio ad's and interviews? Ole Paul knows exactly how to spin that story. LOL

"I mean what In their design process make them more accurate than most hi fi speakers?"

The marketing BS attached to them.

 

[levimax]

It is also appropriate to criticise him for doing so, if as you say, 'doing his job' includes deception.

After all, 'snake oil salesmen' are just doing their job. Doesn't make it okay.

cheers

In the US the laws allow for "puffery".

Puffery is an exaggerated, subjective, or boastful claim made in marketing ("best coffee in the world") that is not intended to be taken as a factual, verifiable statement. Legally permissible, it differs from false advertising because it does not make specific, measurable promises, allowing brands to build imagery without needing evidence.

 

[RexrothPigeon]

you’re just not interested in that type of communication.

Again you mischaracterize my position. This is becoming a pattern.

What I said is that I'm not interested in the massive amount of legwork I would need to exert in order to even begin to evaluate its use-value. And considering I have another source of information (measurements) whose use-value I have already evaluated to my satisifaction (and didn't have to read ad-supported mags or watch youtube), I'm making the easy and rational choice.

You, meanwhile, enjoy the magazines and youtube influencers and enjoyed that sort of thing long before ASR, so naturally you defend it to whatever extent you can without donning the mantle of measurements-denial. I don't begrudge you your fun, and if you find that stuff fun, that's reason enough. It's not necessary to try to convince people that it's also a valid source of information

 

[RexrothPigeon]

As I say: a tool is only useful if you choose to use it.

True. I do not choose to use my shoe as a frying pan, so my shoe is not a useful frying pan for me, a state of affairs almost perfectly analogous to my choice not to use enthusiast marketing publications and influencers to help me assess audio equipment

 

[MattHooper]

Again you mischaracterize my position. This is becoming a pattern.

I have no interest in mischaracterizing your view. In fact I’ve been trying to draw out exactly what it is so I can make sense of it.

We’re talking about communicating information via informal, listening, impressions, and subjective language…about which you wrote:

“I just really don't want to learn that stuff or be influenced by it, since it's the stock-in-trade of the grifters and snake oil salesmen and I don't have any need of it anyway.”

How in the world does “ you’re just not interested in that type of communication” misscharacterized that??? I can’t imagine anybody thinks it’s a mischaracterization.

And you just followed up with more reasons why you’re not interested in putting the effort into that form of communication.

I guess I’m beginning to believe you that communicating with you about sound would be quite difficult.

You, meanwhile, enjoy the magazines and youtube influencers and enjoyed that sort of thing long before ASR, so naturally you defend it to whatever extent you can without donning the mantle of measurements-denial.

That’s substituting attempted psychoanalysis rather than addressing the points I’ve been making. I don’t defend what I defend “ just because I enjoyed it before ASR.”
I defend what I take to be defendable.
I attempted to put together a coherent world view.

I’ve been in the trenches defending ASR ideas long before ASR was around, and I came in for plenty of pushback from the golden ear crowd given how often I pushed back against pseudoscience and pure subjectivism.

. I don't begrudge you your fun, and if you find that stuff fun, that's reason enough. It's not necessary to try to convince people that it's also a valid source of information

I could say it’s not necessary for you to convince anybody of anything as well.

But what we do here as enthusiasts is exchange opinions and viewpoints, and defend the viewpoints we think to be reasonable.

Just like you’re going to be doing that I’m going to be doing that. If I see you making claims about evaluating audio gear that seem dubious or unreasonable I’m likely to point out why that is, just as you do and I expect others here if I make dubious claims.

With that, I don’t think we can go further in these conversations so I will just move on.

Cheers

 

[Audionaut]

Does this suggest speakers can or should be designed without measuring?

Until Neville Thiele and Richard Small, the trial-and-error method was the norm in loudspeaker design. But this wasn’t because engineers didn’t want to measure; rather, the relationships between the parameters to be measured had not been fully described.
Floyd Toole then set many standards, and today we have Klippel Nearfield Scanning and Spinorama with a complete and, very importantly, comparable and reproducible description of
on-axis
off-axis
distortion and
dynamic response.
Anyone who doesn’t measure today has slept through the last 50 years of loudspeaker development, I would say.
Even hobbyists, like myself, now use turntables, REW, and calibrated measurement microphones to apply the whole process to the loudspeaker using DSP.
Why? Because it’s much easier to do today than it was not too long ago.
In that respect, one can reasonably expect every reputable loudspeaker manufacturer to base their development on extensive and reproducible measurements.

In this context smooth directivity is just as important to me as a flat on-axis frequency response. A speaker that measures well only on-axis can still sound poor if its off-axis behavior is inconsistent.

However, I often find that standard measurements—especially for speakers—are too static and brief, as they fail to adequately capture the dynamic behavior.
For example, the sound characteristics of a speaker driver change as the voice coil heats up, which happens with many smaller speakers and at higher volumes, but is almost never reflected in measurements.

In summary:
I don't trust speaker designers who don't use modern measurement techniques, and neither should you.

 

[Sal1950]

You, meanwhile, enjoy the magazines and youtube influencers and enjoyed that sort of thing long before ASR, so naturally you defend it to whatever extent you can without donning the mantle of measurements-denial. I don't begrudge you your fun, and if you find that stuff fun, that's reason enough. It's not necessary to try to convince people that it's also a valid source of information

That's just Matt being Matt and riding the fence here, glad you figured him out.
You would think his butt would get sore after a while ?

 

[Audionaut]

I mean the whole discussion was about how listening fits in the designing process

Let’s start at the end: What is a speaker for?
To clean the kitchen? To drill a hole in a wall? To flip a steak on the grill?

I suspect that a speaker is designed, developed, and built ultimately for the purpose of listening, and in the course of this process, it might be a good idea for the developer(s) to listen as well.

Let’s look at cars and ask the question: Would you willingly and with a clear conscience buy a car that you knew was based entirely on measurements, without anyone ever having tested its actual driving characteristics?
In Formula 1, you can see it even more clearly: Millions are spent creating the optimal cars based on measurements, yet sometimes they still aren’t truly competitive on the track.
They work, but the final few percentage points can only be determined and improved through practical testing.
I see it exactly the same way with loudspeakers.
Today, you can probably design up to 95% of a speaker without ever hearing it; the rest is fine-tuning, which must include listening, because ultimately that is what the speaker is for and what it must be able to do: provide a listening experience.

 

[SigbergAudio]

I would like to summarize what the various POVs of speaker design as it pertains to listening test:

Camp 1) Yes, use measurements, but using uncontrolled, possibly unrepeatable listening tests conducted by the designer (and possibly others) is a key ingredient and it is OK and encouraged for the designer to freely, arbitrarily and subjectively adjusts the on-axis and/or off-axis frequency response as they deemed "accurate" and/or "artistic." Based on many posts by @sigbergaudio, that is not his MO, he does NOT do this. But if I am understanding this right, members such as @Arindal is on this camp. And we suspect AJ to be on this camp as well.

Camp 2) Yes, use measurements, but use listening tests (and not every listening test needs to be a formal controlled, repeatable with a statistically significant amount of test subjects), along with more measurements to decide on engineering trade off's between a flat anechoic response vs. in-room response, with the goal to achieve a smoother in-room response for either on-axis or a broader listening window and depending on the use cases that the speaker is designed for. This sounds like @sigbergaudio. And I can accept this because designing a speaker with excellent directivity control is hard and not all speakers have the same use cases, i.e. on-wall/in-wall speakers, solo listening, home theater, etc. Although, I personally am in the position where I don't have to make cost compromises in buying a speaker with excellent directivity, such that I am expecting a ruler flat on-axis. So while I accept this, the engineering compromises aren't applicable to my situation.

Camp 3) No compromises should be made to the on-axis response because that is the most important factor. You should only aim for ruler flat response. Of course, this assumes the overall speaker is half way decent designed. I am on this camp, but it's easy for me to say this since I am able to seek out and willing to spend the money on speakers with excellent directivity; and the systems in my house has one use case: me, the singular, stationary listener without too much room constraints.

Camp 4) Able to make a speaker with superb directivity, ruler flat on-axis response, absolutely believe in measurements but have controlled and repeatable listening tests demonstrating that in real life use cases, the audience actually prefer something different than what Dr. Toole's research demonstrates. That is the KEF camp, where they have a slightly more aggressive downward tilt.

I will obviously only partly agree with this, especially if I am to understand it correctly that the main difference between camp 2 and 4 is that you are implying Kef are making better speakers.

Also, I will of course suggest that if you have basically no limit to your budget, you absolutely should not go into camp 3 and purchase speakers that has been designed without listening. As I have suggested before, a ruler flat / non-sloped anechoic on-axis response doesn't necessarily translate to the correct sound in-room (for instance a lack of energy in the 100-300 / 100-500hz area is a typical problem).

 

[SigbergAudio]

I will also still suggest that the measurements doesn't tell us everything there is to know. This is not really an argument for or against listening, but rather just pointing out that we still can't know everything by sitting on a forum and reading CEA2034 graphs.

Let's look at for instance the often used example of the Neuman KH-120 II (Preference score 6.58), the praised (rightly so) Revel Ultima Salon 2 (Preference score 6.27), and my own Sigberg Audio SBS G2 (Preference score 6.69).

They are three quite different systems both in design and price, but have a relatively similar preference score, and also relatively similar measurements.

If we use Neuman as the reference,

here is Neuman vs Revel

And here is KH 120II vs SBS G2:

These should sound very similar, and based on price it would be stupid to buy anything other than the KH-120II. I suggest there still is slightly more to the story.

I have no idea if Amir would say that any speaker he has tested that measures better than his Salons also sounds better than the Salons, perhaps he would. But I'm reasonably confident that at least I would not find it to be so.

 

[PristineSound]

I will obviously only partly agree with this, especially if I am to understand it correctly that the main difference between camp 2 and 4 is that you are implying Kef are making better speakers.

Also, I will of course suggest that if you have basically no limit to your budget, you absolutely should not go into camp 3 and purchase speakers that has been designed without listening. As I have suggested before, a ruler flat / non-sloped anechoic on-axis response doesn't necessarily translate to the correct sound in-room (for instance a lack of energy in the 100-300 / 100-500hz area is a typical problem).

I've amended this to camp 3: listening tests is also conducted, but it is used as a sanity check and only do so in a controlled and repeatable manner. If the results aren't as expected, a scientific investigation is conducted to make sense of it and any design flaws are fixed without compromising the ruler flat frequency response and without making the directivity any worse.

I am not implying KEF's approach is better because while I have heard both the blade 1 and 2 and they sound amazing, I have not heard the perfectly flat KEF vs. the slightly tilted KEF. I wasn't part of their listening tests. Who knows if I will say the perfectly flat KEF may or may not sound better than the tilted KEF or maybe they make no difference to me since that tilt is so minor.

But I can also confirm that my Neumann 120, Ascend ELX and Perlisten that's ruler flat, also sounds amazing to me. I would say my Neumann is up there.

 

[SigbergAudio]

I've amended this to camp 3: listening tests is also conducted, but it is used as a sanity check and only do so in a controlled and repeatable manner. If the results aren't as expected, a scientific investigation is conducted to make sense of it and any design flaws are fixed without compromising the ruler flat frequency response and without making the directivity any worse.

I am not implying KEF's approach is better because while I have heard both the blade 1 and 2 and they sound amazing, I have not heard the perfectly flat KEF vs. the slightly tilted KEF. I wasn't part of their listening tests. Who knows if I will say the perfectly flat KEF may or may not sound better than the tilted KEF or maybe they make no difference to me since that tilt is so minor.

But I can also confirm that my Neumann 120, Ascend ELX and Perlisten that's ruler flat, also sounds amazing to me. I would say my Neumann is up there.

Are we aware of any speaker manufacturers in camp 3, or is it a theoretical ideal?

Which Ascend and Perlisten do you own?

 

[PristineSound]

I have no idea if Amir would say that any speaker he has tested that measures better than his Salons also sounds better than the Salons, perhaps he would. But I'm reasonably confident that at least I would not find it to be so.

I believe Amir measured a Wilson and it wasn't so good, but he ultimately recommended it because it sounded pretty good in the listening test.

But I am not sure if he did any in-room measurements, perhaps compromises were made for the sake of better in-room response?

But I want to make a statement, I have the Neumann 120, they sound as good as they measure. They are easily the best speakers I have heard given the size.

 

[SigbergAudio]

But I want to make a statement, I have the Neumann 120, they sound as good as they measure. They are easily the best speakers I have heard given the size.

Here it is of course tempting to suggest that you are biased because you know how they measure.

 

[PristineSound]

I should reword it. When I say ruler flat, I meant comparatively to all the speakers out in the market, with the Neumann leading the way with the definition of ruler flat. But with Perlisten leading the way with directivity, but all three still have Excellent directivity relative to what is on the market.

 

[PristineSound]

Here it is of course tempting to suggest that you are biased because you know how they measure.

Yes there is the risk of that bias. And I don't have the equipments to do a blind listening test.

But I can also argue you may have suffered from the same bias in your uncontrolled listening test.

However, my Ascend costs a lot more and I suspect my bias wants to tell me the Ascend is better, and they are in many ways, but if you take the bass extension and punch away and the ultra wide soundstage away, man those Neumann are so much better.

 

[thewas]

Exactly, it's likely a branding/marketing exercise.

I was talking to two amplifier designers at our last hifi show. One of them sells a Class AB amp that he designs and manufactures himself for a fairly reasonable price. He said "I can't believe that people are willing to pay >$100k for an amplifier!". I told him: "this is what you do. Take your amplifier and stuff it into a really nice and expensive case. No need to change the design, just make sure the build quality and aesthetics are elevated. Tell everyone that you have turned over a new leaf. Say that in your younger days, you believed in oscilloscopes, measurements, low THD, etc. but that was not the path to sound quality. Now that you are older, you have realized that superior sound is not just measurements, it must involve the listening experience".

He said: "you cynical bastard, you're telling me to sell the same amplifier for 10x the price but with different marketing spin?". The other amplifier designer, who was chuckling away, said "he's right, that's what you should do". I pointed out it will cost him next to nothing, and such a halo product might improve the sales of his more reasonably priced amplifier. If it also had a suitable upgrade in aesthetics. He muttered something about "it will cost me my conscience".

Talking about amplifiers, a fun story about audio electronic engineers doing their "voicing" by listening.
One of the world best best studio electronics engineer gave an interview from which I electronically translated a part below:

Friedemann Kootz: Your developments are really planned
planned, measured and tested down to the last detail.
and tested. Do you listen to the devices
do you listen to the devices you develop?

Thomas Funk: No. For example, our MTX monitoring controller
for example, I heard it for the first time about two
for the first time about two years ago, even though
have had it in our product range for nine years.
for nine years. Our customers from the high-end
always have astonished faces when I
the single loudspeaker box in the back under the
the desk on the metal shelf.
Some of the 'high-enders' sometimes react
downright mischievous when I tell them
that we don't do listening tests. But
it's easy to explain. We develop
technology, the most important property of which
is absolute linearity. If I have
a compressor, an equalizer, or anything else that would shape
sound shaping technology,
I would also do listening tests to see if it sounds
to check if it sounds 'good'. If I remember
remember, it happened once that I developed a headphone
developed a headphone amplifier
and listened to it under full load with about 10
dB overload. At that time I wanted to
to find out whether the stability
of the circuit in the absolute limit range
changes so much that it would have to be
would have to be modified. But since this was not the
case, this 'listening test' was exhausted after ten minutes.
exhausted after ten minutes.

Full original interview in German language can be found here http://www.funk-tonstudiotechnik.de/1 Studiomagazin 1-2009 Interview.pdf

Of course loudspeakers are a different thing due to their inherent multidimensional character and limitations, nevertheless a refreshing story.

 

[SigbergAudio]

I should reword it. When I say ruler flat, I meant comparatively to all the speakers out in the market, with the Neumann leading the way with the definition of ruler flat. But with Perlisten leading the way with directivity, but all three still have Excellent directivity relative to what is on the market.

Yes, the problem is that ruler flat on-axis doesn't tell us everything we need to know.

An even flatter and more highly rated speaker (albeit at a higher price) is the Kii Three. While an impressive speaker in many ways, I must say (with great respect to the designer) it is not the best speaker I have ever heard. Can the reason for this be found in the measurements? Perhaps.

 

[PristineSound]

Yes, the problem is that ruler flat on-axis doesn't tell us everything we need to know.

An even flatter and more highly rated speaker (albeit at a higher price) is the Kii Three. While an impressive speaker in many ways, I must say (with great respect to the designer) it is not the best speaker I have ever heard. Can the reason for this be found in the measurements? Perhaps.

I would like to think I am well read with regards to speaker design, I also would like to think I am well experienced listening to speakers compare to the general population. I am not a DIY speaker designer, and I do not have the test equipment, so I am no where as experience of yourself and others.

But I do believe from my experience listening to so many speakers in my home and other people's home, that the CEA2034 is super important and I have never heard any speakers that didn't sound good but measured well on the CEA2034. I may have heard few speakers that sounded good and didn't measure well on the CEA2034. As such, I believe a ruler flat on axis response with excellent directivity should be king.

BUT. . .there are other measurements outside of the CEA2034 that I believe don't get enough attention, distortion is one. I also kind of agree with you that there are things that I hear that is not explained by the CEA2034, but perhaps can be explained by other measurements. For example, my Perlisten has this impeccable mid-range that is not explainable by the frequency response, I would like to think it has something to do with the narrow vertical dispersion, but then again, research tells us that we humans aren't sensitive to vertical sound reflections, I would also like to think that the tweeter material has something to do with it, but others has pointed out to me that Dr. Toole have once stated that his research shows tweeter material made no difference in a blind listening test.

In short, I agree that on-axis response alone or even with directivity, don't tell you everything about how a speaker would sound, which is my experience (without a proper blind ABX). but I still believe that on-axis along with great directivity would guarantee you a pair of speakers sounding good, how good is a different matter.