Effect of Loudspeaker Directivity Compared with In-room Measurements

[Kvalsvoll]

And to me the status quo of sound (re)production is nothing to be happy about.

I would say, compared to when I was growing up, the situation today is far better - never before could you buy so much sound quality for so little money. Especially the more economic products have become far better in comparison.

But perhaps it is possible to improve sound reproduction, beyond what the typical box speaker can achieve. This is a technical challenge, and it involves both the speaker and the room, and this is why I am working on sound and hifi.

 

[Kvalsvoll]

You can read some thoughts from Sigfried Linkwitz here:

Frontiers

Conclusions

@JanRSmit I think S.L. answers those questions very well in those links.

I was mostly thinking about radiation pattern, and thus how room resonances affect the sound, but of course there are some benefits of not having to deal with box resonances in itself, and the reflected sound from inside the cabinet out through the cone.

 

[markus]

I would say, compared to when I was growing up, the situation today is far better - never before could you buy so much sound quality for so little money. Especially the more economic products have become far better in comparison.

Yes, and I said that already above

But perhaps it is possible to improve sound reproduction, beyond what the typical box speaker can achieve. This is a technical challenge, and it involves both the speaker and the room, and this is why I am working on sound and hifi.

I don't want to unnecessarily prolong this off-topic discussion but the problem is not that the speaker or room couldn't be made to perform in a certain way, the problem is a) audio's circle of confusion (as a lack of meaningful standards) and b) the lack of psychoacoustic research (which has to be at the core of any meaningful standard). Without that a cardioid or any other specific directivity pattern is just another solution to a problem that might not even exist.

So back on topic: Let's see the "Effect of Loudspeaker Directivity Compared with In-room Measurements", i.e. a comparison of multiple speakers with different directivities and how that changes the in-room response. At best in different rooms. Quite a task as you would need to present directivity data of different speaker designs and relate that to response changes in the room. I would expect a full set of measurements showing things like RT60, filtered IR (include Schroeder integral), clarity, spectrogram, etc.
If you do it in REW then please post the .mdat's so others can verify the results. Otherwise you open yourself up to the suspicion of having a hidden agenda.

 

[JanRSmit]

@JanRSmit I think S.L. answers those questions very well in those links.

I was mostly thinking about radiation pattern, and thus how room resonances affect the sound, but of course there are some benefits of not having to deal with box resonances in itself, and the reflected sound from inside the cabinet out through the cone.

Kvalsvoll, and Youngho, thanks for the response. Kvallvoll your info in this thread is about speaker-room interaction (and follow with great interest) , that is why i asked.

Yes i am aware of Linkwitz findings. I can also understand that the shape of the entire box has an influence in that interaction.

Regarding the box itself i am using short-fibre very well cleansed wool 'blobs' to fill the cavity, for 2 reasons: damping of the fundamental resonance of the driver, and damping of cavity resonances. This has proven to be a quite effective solution. When professionally active in loudspeaker production(many many years ago), we did a lot of testing and f.i. a test by placing a sheet of BAF (Bonded Acetate Fibre) in the cavity of a midrange (Thiel C2-77) together with wool and it sounded very "active" (very eufemistic statement ;-) ). The cavity resonances were indeed coming back via the driver and sounded awful. Removing the BAF layer completely removed this 'activeness'. Similar "active" effect happens if you card the wool by the way. We learned that the way the wool filling in the cavity was constructed is quite crucial.
The damping at low frequencies (this midrange resonance was ~ 100 Hz) is the result of the fibres reacting to pressure variations. The fibre structure resembles a pine-cone, and the 'petals' open or close under pressure variations, thus absorbing energy.

As Linkwitz states in one of his many publications, the box structure itself needs to be well damped as well. My current loudspeakers are constructed such that structural resonances and sound transmission is minimalized. The cross-section is an ellips truncated by the baffle, the wall material used is composed of thin (~1.5mm) layers of HPL (the brown core material of formica) glued with epoxy glue, and in the middle a damping layer (Constrained Layer Damping) specific for thick stiff structures (engine support structure and hull for GRP made yachts etc). The bulkheads separating the bass driver's cavities are also thick HPL plates. The part where the midrange and tweeter are is composed of layers of MDF with irregular shaped cavities for mid and tweeter, the total stack being some 30cm high. The baffle plate is also composed of two thick HPL layers with a CLD damping layer in between.
All in all the loudspeaker-box is quite dead both structurally and cavity wise. Also, given its high rigidity coupled with some serious mass (> 70kilo) it forms a excellent 'launching' platform for cone movements (action=reaction), which is also important.

 

[Duke]

the problem is a) audio's circle of confusion (as a lack of meaningful standards) and b) the lack of psychoacoustic research (which has to be at the core of any meaningful standard).

a) Audio's circle of confusion is not a problem that one small speaker manufacturer can address, as it arises from different speakers (and different room acoustic conditions) being used for the creation and playback of recordings. Nor is it a crippling problem. Note that the man who came up with this paradigm, Floyd Toole, uses home audio speakers and "tasteful upmixing" of stereo recordings in a fairly normal living room, so even he is not preoccupied with replicating the conditions in a recording studio. And his system probably sounds better than most recording studios.

b) We already have sufficient psychoacoustic research in this area. Research by Harman has shown the benefits of good radiation pattern control, and cardioid loading is clearly an effective technique for extending good radiation pattern control down lower in frequency than is otherwise possible in a comparable conventional enclosure. We can connect those dots ourselves without waiting for someone to do it for us in another peer-reviewed paper.

Without that a cardioid or any other specific directivity pattern is just another solution to a problem that might not even exist.

Poor radiation pattern control is known to be a problem. Therefore improving radiation pattern control would be a logical approach to addressing that problem. And cardioid loading is a technique which improves radiation pattern control.

Let's see the "Effect of Loudspeaker Directivity Compared with In-room Measurements", i.e. a comparison of multiple speakers with different directivities and how that changes the in-room response. At best in different rooms. Quite a task as you would need to present directivity data of different speaker designs and relate that to response changes in the room. I would expect a full set of measurements showing things like RT60, filtered IR (include Schroeder integral), clarity, spectrogram, etc.
If you do it in REW then please post the .mdat's so others can verify the results. Otherwise you open yourself up to the suspicion of having a hidden agenda.

In my opinion it is unreasonable to ask this of Kvalsvoll. Demanding an unreasonable level of proof is a rhetorical technique called "Sea Lioning":

And saying that Kvalsvoll "opens himself up to the suspicion of having a hidden agenda" if he fails to meet your demands is manipulative and insulting, in my opinion.

 

[markus]

a) Audio's circle of confusion is not a problem that one small speaker manufacturer can address, as it arises from different speakers (and different room acoustic conditions) being used for the creation and playback of recordings. Nor is it a crippling problem. Note that the man who came up with this paradigm, Floyd Toole, uses home audio speakers and "tasteful upmixing" of stereo recordings in a fairly normal living room, so even he is not preoccupied with replicating the conditions in a recording studio. And his system probably sounds better than most recording studios.

Argumentum ab auctoritate? The reason why "he is not preoccupied with replicating the conditions in a recording studio" is because he knows that the circle of confusion is in full force and obsessing about certain performance characteristics is likely just an exercise in futility.

b) We already have sufficient psychoacoustic research in this area. Research by Harman has shown the benefits of good radiation pattern control, and cardioid loading is clearly an effective technique for extending good radiation pattern control down lower in frequency than is otherwise possible in a comparable conventional enclosure. We can connect those dots ourselves without waiting for someone to do it for us in another peer-reviewed paper.

Sorry, I don't agree. Harman just showed that the smoothness of certain radiation patterns (which are in fact NOT cardioid) are preferred (in specific rooms with specific source material). Please do realize the limitations of the Harman research. A lot more research is necessary. Read through Toole's papers and you'll find a lot of research questions that no one has answered yet.
This thread also doesn't want to show that cardioid is desirable, does it? It wants to show the effect of directivity on the in-room response. Whether the effect is good or bad won't show up in those measurements. If one piece of wood is one foot long and the other is two feet then which one is "better"? Depends on the application. This a totally separate issue.

Poor radiation pattern control is known to be a problem. Therefore improving radiation pattern control would be a logical approach to addressing that problem. And cardioid loading is a technique which improves radiation pattern control.

Not sure what you're arguing against. Why do you assume I was talking about poor radiation pattern control? I said "specific directivity pattern" which means controlled directivity.

It is unreasonable to ask this of Kvalsvoll. Demanding an unreasonable level of proof is a rhetorical technique called "Sea Lioning":

View attachment 173696

And saying that Kvalsvoll "opens himself up to the suspicion of having a hidden agenda" if he fails to meet your demands is manipulative and insulting.

Not sure why you feel the need to speak for him, and please stop the ad hominem.

Kvalsvoll opened this public thread. He wanted to show the "Effect of Loudspeaker Directivity Compared with In-room Measurements". He refuses to post his .mdat's. I'm just laying out what I think needs to be done to show the effect in a meaningful way. I'm also asking him to make his .mdat's available. How is that the same as "demanding an unreasonable level of proof"? And in my mind posting on audiosciencereview.com needs to mean more than just posting your subjective opinion.

 

[Duke]

Sorry, I don't agree. Harman just showed that the smoothness of certain radiation patterns (which are in fact NOT cardioid) are preferred (in specific rooms with specific source material). Please do realize the limitations of the Harman research. A lot more research is necessary.

Harman's research shows which radiation pattern characteristics are desirable. Cardioids generally have these characteristics moreso than comparable direct-radiator monopoles.

We will never have research which answers every question and objection we might raise. We will accomplish much more by applying the research we already have instead of dismissing it because it does not include this or that specific data point, especially if we can make our own reasonable evaluation just by thinking it through.

Not sure why you feel the need to speak for [Kvalsvoll], and please stop the ad hominem.

He is far more of a gentleman than I am. And I would have just let it go had you not said this:

[If you do not meet my demands] you open yourself up to the suspicion of having a hidden agenda.

Would you like to have your integrity called into question like that? If not, then don't do it to others.

 

[markus]

Harman's research shows which radiation pattern characteristics are desirable. Cardioids have these characteristics moreso than comparable direct-radiator monopoles.

You need to reread the papers. Seriously. If anything I would say that Harman research suggests that cardioid could be less preferred as the narrow directivity designs they tested that hold their directivity pattern down to lower frequencies had low preference ratings. That conclusion could be wrong just like the opposite conclusion might be wrong. The research just doesn't allow for any such conclusion.

We will never have research which answers every question and objection we might raise.

Strawman. We need answers to specific questions. That's not the same as demanding "research which answers every question and objection we might raise".

We will accomplish much more by applying the research we already have instead of dismissing it because it does not include this or that specific data point, especially if we can make our own reasonable evaluation just by thinking it through.

Pardon me, what research am I dismissing?

Would you like for your integrity to be questioned like that? If not, then don't do it to others.

I got nothing to sell. I didn't open this thread. He on the other hand is a manufacturer that happens to sell cardioid speakers and offers acoustic consulting services, comes to ASR and opens a thread about loudspeaker directivity comparing his speaker to a "traditional" design. Even linking to his site in the first post.

Again, science is based on certain principles like rigorous skepticism, careful observation or verifiability. Don't confuse that with sealioning.

 

[Duke]

You need to reread the papers. Seriously. If anything I would say that Harman research suggests that cardioid could be less preferred as the narrow directivity designs they tested that hold their pattern down to lower frequencies had low preference ratings. That conclusion could be wrong just like the opposite conclusion might be wrong. The research just doesn't allow for any such conclusion.

A cardioid has a radiation pattern width of ballpark 130 degrees (-6 dB at about 65 degrees off-axis to either side). Whether this is a narrow or wide pattern depends on what we're comparing it to. The advantage of the cardioid enclosure is that it can maintain its pattern across approximately two and a half octaves. Conventional monopole direct radiators won't do that.

Nor, to the best of my knowledge, has Harman published any tests wherein they identify any of the speakers as having cardioid behavior. If you know of any such tests, please let me know because I'd like to read about them and learn from them. It is possible that I am mistaken.

Again, science is based on certain principles like rigorous skepticism, careful observation or verifiability. Don't confuse that with sealioning.

I see nothing "scientific" about questioning someone's integrity. Can you leave that sort of thing out?

 

[markus]

A cardioid has a radiation pattern width of ballpark 130 degrees (-6 dB at about 65 degrees off-axis to either side). Whether this is a narrow or wide pattern depends on what we're comparing it to. The advantage of the cardioid enclosure is that it can maintain its pattern across approximately two and a half octaves. Conventional monopole direct radiators don't do that.

Yep but whether that's an advantage or not has yet to be seen and it certainly isn't what Harman research suggests.

I see nothing "scientific" about questioning someone's integrity. Leave that out.

I was referring to your insulting "sealioning" comment.

 

[sigbergaudio]

Yep but whether that's an advantage or not has yet to be seen and it certainly isn't what Harman research suggests.

Somewhat anecdotal, but those who've heard speakers like DutchDutch and Kii seem to have a rather uniform and positive experience, including Erin who recently reviewed both. It's not extremely far fetched to assume a uniform off-axis response within a wide listening window combined with reduced activation of reflections from side and rear walls would be a good thing. Sounds pretty much in line with the Harman research as well.

 

[Duke]

Yep but whether that's an advantage or not has yet to be seen and it certainly isn't what Harman research suggests.

I disagree, loudspeakers whose off-axis energy tracks their on-axis energy well (most Revels, JBL M2) have fared well in Harman's controlled blind listening tests, which imo implies that that's the right direction to pursue. The Revel Salon 2's horizontal radiation pattern width is approximately 150 degrees across most of the spectrum, while the M2's horizontal pattern width is approximately 120 degrees across most of the spectrum, so a cardioid's 130 degree pattern is arguably "in the ballpark".

I was referring to your insulting "sealioning" comment.

And I was referring to your insulting "[If you do not meet my demands] you open yourself up to the suspicion of having a hidden agenda" comment.

I'll knock it off it you will.

Truce?

 

[markus]

Somewhat anecdotal, but those who've heard speakers like DutchDutch and Kii seem to have a rather uniform and positive experience, including Erin who recently reviewed both. It's not extremely far fetched to assume a uniform off-axis response within a wide listening window combined with reduced activation of reflections from side and rear walls would be a good thing. Sounds pretty much in line with the Harman research as well.

Not really. Harman research showed that spaciousness (=lateral reflections) were preferred by "recreational listeners". It also showed that "the worst offenders" are floor and ceiling reflections. A cardioid doesn't help here. Those two speakers you've mentioned are still excellent.

 

[markus]

I disagree, loudspeakers whose off-axis energy tracks their on-axis energy well (most Revels, JBL M2) have fared well in Harman's controlled blind listening tests, which imo implies that that's the right direction to pursue. The Revel Salon 2's horizontal radiation pattern width is approximately 150 degrees across most of the spectrum, while the M2's horizontal pattern width is approximately 120 degrees across most of the spectrum, so a cardioid's 130 degree pattern is arguably "in the ballpark".

And they all become omni at the low end, except the cardioid – which becomes omni at the lower end

 

[Duke]

And they all become omni at the low end,

As do the vast majority of other speakers Harman has tested, some of which scored terribly. The Salon 2 and JBL M2 scored exceptionally well in large part because they had good pattern control across most of the spectrum.

The Harman data shows that speakers which maintain good pattern control down to a lower frequency than others tend to out-score those others. So imo it is reasonable to extrapolate that maintaining good pattern control down further still would be a further improvement.

Do you have reason to believe that extending good pattern control down lower than the Salon and M2 would not be beneficial?

... except the cardioid.

Which, to the best of my knowledge, Harman has not tested, but the Kii and Dutch & Dutch have been thoroughly measured by others, and the measurements predict high preference scores.

(Not sure about the Kii, but the Dutch & Dutch is a monopole in the bass region.)

 

[markus]

As do the vast majority of other speakers Harman has tested, some of which scored terribly. So omni down low is not clearly an advantage.

I've never said that bad speaker design is an advantage I'm sure you can f*ck up any design, including cardioid.

The Salon 2 and JBL M2 scored exceptionally well in large part because they had good pattern control across most of the spectrum.

What the Harman data does show is that speakers which maintain good pattern control down to a lower frequency than others tend to out-score those others. So imo it is reasonable to extrapolate that maintaining good pattern control down further still would be a further improvement.

Do you have reason to believe that extending good pattern control down lower than the Salon and M2 would not be beneficial?



Which, to the best of my knowledge, Harman has not tested, but the Kii and Dutch & Dutch have been thoroughly measured by others, and the measurements predict high preference scores.

Get out your copy of "Sound reproduction". The metrics for the Harman preference calculation are listed there. Flatness of the directivity curve is not one of them.

Is there any chance this thread gets back on topic?

 

[Kvalsvoll]

He is far more of a gentleman than I am.

Thank you, but I am not so sure this is true, you are the gentleman here.

Communicating on-line and on the forums can at times be challenging, and we often write in a way that comes out a bit different from what we intended. Especially in these times, where the world has become increasingly hostile, and this hostile and rude behavior has come to be accepted.

For me, I have found that using my name and not hiding under some obscure alias, makes me think one more time before attacking and writing insulting comments. But - it does not always work out.

Giving someone who write a comment that I find negative some slack, can be a good idea. Perhaps I did not understand. But there is a limit, to what one can and should find acceptable.

Now, back to what this is - a technical thread about something sound-related.

 

[Kvalsvoll]

Audio's circle of confusion is not a problem that one small speaker manufacturer can address, as it arises from different speakers (and different room acoustic conditions) being used for the creation and playback of recordings. Nor is it a crippling problem.

Yes, this is not really a problem. It works just fine to have a very different system to play back something that was created in a studio, where the sound was different.

And the faults are mostly realted to frequency response and tonality, so they can often be corrected. I am lazy, so I just listen on and accept that "yes, this recording form the 80ies has less bass", and then "They sure did turn up the bass on this Billie Eilish".

 

[markus]

Yes, this is not really a problem. It works just fine to have a very different system to play back something that was created in a studio, where the sound was different.

And the faults are mostly realted to frequency response and tonality, so they can often be corrected. I am lazy, so I just listen on and accept that "yes, this recording form the 80ies has less bass", and then "They sure did turn up the bass on this Billie Eilish".

I agree there are general trends that can be seen (heard) in recordings over the decades but audio's circle of confusion has a much deeper impact.

The speaker-room interface is the foundation on which a recording, mixing and mastering engineer makes his creative decisions. Not only in terms of overall spectral balance. It influences his choice of miking techniques, application of processing like reverberation or compression, equalization, levels. In short it influences every important aspect of a recording.

The consumer has no way of knowing what speaker-room interface was used in the making of a recording. He can't even know at what level a recording was mixed and mastered (recently even the movie industry gave up reference level when they started doing home remixes). Furthermore the speaker-room interface can vary a lot from studio to studio. Hence the consumer has no way of knowing how he could "counteract" these effects or how his speaker-room interface ideally should look like.

You could say, "adjust to taste" on the reproduction side. That is a valid approach – like drawing on a Picasso because you don't like how it looks – but make no mistake, that is not faithful reproduction. If at the same time now someone is obsessing about specific speaker directivity and room treatments – which is the same as drawing the very same thing on any picture, not just that particular Picasso – then the whole approach becomes more or less an exercise in futility. Nothing more but cargo cult really. And like any belief system it provokes "interesting" reactions when challenged.

My personal goal is faithful reproduction. I do want to hear the art as the artist intended. So no, I don't think that audio's circle of confusion is just a minor thing. I think it prevents the art from being "seen" in its entirety. And I think it takes away opportunities for the artist to be even more creative – not that every recording would be a piece of art or even creative but some are.

 

[Duke]

My personal goal is faithful reproduction. I do want to hear the art as the artist intended.

That goal seems ideal on paper, but is it realistic?

The consumer has no way of knowing what speaker-room interface was used in the making of a recording... Hence the consumer has no way of knowing how he could "counteract" these effects or how his speaker-room interface ideally should look like.

If these unknowns preclude the possibility of "hearing what the artist intended", then doing so is not really a realistic goal, is it?

So IS there a different paradigm which offers a more pragmatic starting point? How about this:

A speaker system should do two things: It should get the direct sound right, and it should get the reflected sound right.

The advantage of this paradigm is, we no longer need to know and replicate the acoustics of the studio control room (which was optimized to be maximally revealing rather than maximally enjoyable anyway). All we need do is optimize our own system and room for presenting the direct and reflected sound to the listeners. "How to" is beyond the scope of this post, and delves into room acoustics as well as loudspeaker directivity. My point here is only to suggest a more pragmatic paradigm as a starting point:

A speaker system should do two things: It should get the direct sound right, and it should get the reflected sound right.