I read an interview today between John Atkinson from Stereophile and the late John Dunlavy. It's a very interesting read that I recommend checking out. One snippet from the interview that really stood out to me was this:
"Dunlavy: That's certainly true. We pay an awful lot of attention to the power response of the speaker into the room. Because that's one of the things that permits us to determine whether we're listening to a live instrument, let's say with our eyes closed, in a typical room. We hear two things. We hear the direct sound of the instrument, but we also hear all of the reflected sound, the reflections off of all of the boundaries of the room. And the ratio between that direct sound as a function of frequency and the reflected sound determines to our ears whether we perceive it as being realistic or not.
We spent a lot of time and money, over 20 years ago, doing measurements in an anechoic chamber of the three-dimensional response patterns of 17 different musical instruments, including drums, string bass, cello—we measured a bassoon, a clarinet, a violin. If a loudspeaker's directivity pattern is incapable of emulating the aggregate, the average of the patterns of all of these musical instruments, it will never sound "accurate."
Most musical instruments are almost omnidirectional at low frequencies, as are most loudspeakers, so it doesn't pose a problem. But as you go higher in frequency, to between 100Hz and 300Hz, if you don't get the beam-width of the speaker correct in this range—and by "correct" I mean that it simulates most live instruments—it will add warmth, unnatural warmth, to the sound of voices and musical instruments. It'll make the average male voice sound too chesty, very unnatural. As you go up higher in frequency, if you have a tweeter that radiates too broad a pattern...it's going to produce shrieky sounds, it's going to sound too zippy. I think everyone's experienced that, especially from inexpensive speakers that have a rising high end.
So a good designer certainly knows that he has to pay a lot of attention to the polar response of a loudspeaker."
I have seen many discussions about speaker directivity, but the approach Mr. Dunlavy mentions here I have not come across until now. I would be interested, and I'm hoping so will you all, in a discussion on this approach to speaker directivity. Thoughts?
Full interview here: https://www.stereophile.com/interviews/163/index.html
"Dunlavy: That's certainly true. We pay an awful lot of attention to the power response of the speaker into the room. Because that's one of the things that permits us to determine whether we're listening to a live instrument, let's say with our eyes closed, in a typical room. We hear two things. We hear the direct sound of the instrument, but we also hear all of the reflected sound, the reflections off of all of the boundaries of the room. And the ratio between that direct sound as a function of frequency and the reflected sound determines to our ears whether we perceive it as being realistic or not.
We spent a lot of time and money, over 20 years ago, doing measurements in an anechoic chamber of the three-dimensional response patterns of 17 different musical instruments, including drums, string bass, cello—we measured a bassoon, a clarinet, a violin. If a loudspeaker's directivity pattern is incapable of emulating the aggregate, the average of the patterns of all of these musical instruments, it will never sound "accurate."
Most musical instruments are almost omnidirectional at low frequencies, as are most loudspeakers, so it doesn't pose a problem. But as you go higher in frequency, to between 100Hz and 300Hz, if you don't get the beam-width of the speaker correct in this range—and by "correct" I mean that it simulates most live instruments—it will add warmth, unnatural warmth, to the sound of voices and musical instruments. It'll make the average male voice sound too chesty, very unnatural. As you go up higher in frequency, if you have a tweeter that radiates too broad a pattern...it's going to produce shrieky sounds, it's going to sound too zippy. I think everyone's experienced that, especially from inexpensive speakers that have a rising high end.
So a good designer certainly knows that he has to pay a lot of attention to the polar response of a loudspeaker."
I have seen many discussions about speaker directivity, but the approach Mr. Dunlavy mentions here I have not come across until now. I would be interested, and I'm hoping so will you all, in a discussion on this approach to speaker directivity. Thoughts?
Full interview here: https://www.stereophile.com/interviews/163/index.html
I'm not one who tries to make a speaker emulate an instrument and I don't believe it needs to, or that it necessarily should.
In cases where the ambience is on the recording, you need to hear it through the speakers. Designing for best imaging is the way to do this. One way is to avoid early room reflections. Keep power balanced for the later room reverb which doesn't impact imaging due to it's timing, but can be a problem if it is not balanced.
In cases where the recorded parts are closer to anechoic, designing the speakers this way is still of benefit. If you try to add listening room spaciousness you risk increasing listener fatigue if you don't do it right. This may in some way explain why some builders pursue dipole designs, diving in in the other direction as a usable alternative.
I don't believe that a near anechoic recording on a narrow directivity playback system with low listener fatigue is necessarily a bad thing despite people mentioning the 'headphone effect'. I hear stories of 'sterile' and 'analytical' but I feel these speakers must have other concerns.
In cases where the ambience is on the recording, you need to hear it through the speakers. Designing for best imaging is the way to do this. One way is to avoid early room reflections. Keep power balanced for the later room reverb which doesn't impact imaging due to it's timing, but can be a problem if it is not balanced.
In cases where the recorded parts are closer to anechoic, designing the speakers this way is still of benefit. If you try to add listening room spaciousness you risk increasing listener fatigue if you don't do it right. This may in some way explain why some builders pursue dipole designs, diving in in the other direction as a usable alternative.
I don't believe that a near anechoic recording on a narrow directivity playback system with low listener fatigue is necessarily a bad thing despite people mentioning the 'headphone effect'. I hear stories of 'sterile' and 'analytical' but I feel these speakers must have other concerns.
That is simply not correct. A lot of instruments behave very differently in terms of dispersion. And how much 'power response' is there not only depends on the dispersion but very much also on the listening room, the furniture, carpet, windows and speaker placement, listening distance etc. Of course a uniform/controlled dispersion provides a good average basis but in that range the location isn't very good yet and in that range you can fix that relative easily by an EQ and to control the dispersion in that range leads to speaker sizes which are not acceptable for the most people and/or room dimensions.
I think rather it should not.
I agree. But for the location it is also important to avoid asymmetrical reflections. Ie, one speaker close to the side wall, the other at the middle of a (back) wall. If the best location is a goal, the asymmetry of the reflections have to go first.
A dipole can help in these regards but many ppl don't realize a lot of the room impression comes from the positioning of the OB and they haven't tried a conventional speaker with that much distance to the adjacent walls. Another misconception is the aleged 'boxy' sound of conventional speakers. Many dipole speakers have a room impression that extremely wide and 'airy' but that is often quite exaggerated. I mean, it's completely fine if you like that but that doesn't make smaller room impressions wrong but often rather more realistic. Don't misunderstand me, there are amazing OB speakers but a dipole isn't better (or worse) by definition. All speakers are a compromise - It's simply a different compromise.
That depends also strongly on the listening distance and towards which characteristics the speaker development went.
While I don’t agree with much of this in regards to why, the power response argument still holds water…..thanks for sharing.
If every recording complied with the same recording standard (i.e., technique), then there would be a single correct way to design a speaker. But that is not the case. A "dry" recording might benefit from added reverberation in the listening environment, but the same environment could ruin the sound of a recording that has more of the ambience (i.e., not dry/wet) of the recording environment captured. A balance is called for.
One can theorise about it, but one of the most useful things done was to use empirical observation, as with the Harman studies. It still does not get you a singular correct speaker design, but applying this knowledge significantly improves your chances of making a generally very good loudspeaker.
No disrespect to John Dunlavy intended but consider that every manufacturer will (must) give the impression that they secretly know just the right amount of spice to add to their "unique" formula.
One can theorise about it, but one of the most useful things done was to use empirical observation, as with the Harman studies. It still does not get you a singular correct speaker design, but applying this knowledge significantly improves your chances of making a generally very good loudspeaker.
No disrespect to John Dunlavy intended but consider that every manufacturer will (must) give the impression that they secretly know just the right amount of spice to add to their "unique" formula.
That's a completely wrong assumption of the reproduction of sound. Since there will still be non-standard listening rooms (dimenstions, furniture etc), it's still not possible to grant the same listening result, even close. And that's even before considering the personal taste, listening abilities and media limitations or speaker placement.
From that Dunlavy interview i read for nearly ten years regularly, i get what John said is not really different than what Earl Geddes or Dave Smith ( RIP) said too:
To have a smooth directivity behaviour from omni to some more narrow directivity happening* as a design goal.
The range given by Dunlavy 100/300hz is were there is full transition to 4 pi in many 'large width' ( 50cm width and more) loudspeakers ( when baffle step is then fully transitioned) and i interpret what he is saying as there must be a bsc compensation included in the design goal too.
It's the same about the tweeter 'step'. In fact i think he gave hint about how he defined crossover point and reasons to have doubled membrane area for low end on his flagship models...
* Earl Geddes insist on constant directivity from 1khz and up. Dave Smith wasn't bothered by a bit of narrowing in the upper octave, but otherwise he agreed with the fact smooth transition to this should occur. In essence this is very close to what Earl defend but with some different details.
It can be read here ( truth to be told both Dunlavy's and Smith interview i read regularly and often one after the other...):
https://www.tnt-audio.com/intervis/david_smith_e.html
To have a smooth directivity behaviour from omni to some more narrow directivity happening* as a design goal.
The range given by Dunlavy 100/300hz is were there is full transition to 4 pi in many 'large width' ( 50cm width and more) loudspeakers ( when baffle step is then fully transitioned) and i interpret what he is saying as there must be a bsc compensation included in the design goal too.
It's the same about the tweeter 'step'. In fact i think he gave hint about how he defined crossover point and reasons to have doubled membrane area for low end on his flagship models...
* Earl Geddes insist on constant directivity from 1khz and up. Dave Smith wasn't bothered by a bit of narrowing in the upper octave, but otherwise he agreed with the fact smooth transition to this should occur. In essence this is very close to what Earl defend but with some different details.
It can be read here ( truth to be told both Dunlavy's and Smith interview i read regularly and often one after the other...):
https://www.tnt-audio.com/intervis/david_smith_e.html
It might be that he meant that, his quote said said something different.
That's right because the narrowing in the upper octave isn't a design goal but rather the opposite, a compromise. Same confirmed Earl Geddes.
About Dunlavy quote: he was a genius for sure but he couldn't break law of physics.
To have directivity management at such low freq there is not many ways: either you go with waveguide or some kind of cardioid box are the obvious one ( to me at least). Neither of which he used in his own design so it only let one possible thing: baffle step from 'large width' box as it's a way to achieve some kind of 'smooth' transitioning from omni to hemispheric pattern.
Patrick Bateman's some years ago rambling on Von Hoecks box design to manage directivity got me into this, as until i read his posts i understood bsc only as a frequency related issue ( because it's displayed as this, high freq going up 6db over 4 octaves, i find there should be a directivity pattern transition accompanying this freq graph as this is not evident to me displayed that way even if i know it come from radiation pattern constraining...)... study of Dutch & Dutch 8c and how Martijn incorporated baffle step to achieve a wider freq area of directivity management ended convinced me this was what Dunlavy was about giving hint of this freq range and directivity behavior.
I might be wrong but there is enough evidence from other source to convince me it's what it's all about.
So how i interpret this cryptic sentence ( to me) is if there is omni radiation in the 300/100hz range then human voice sound chesty, and other instruments 'warmer' than they are in real life. Which could also explain why people tends to like small width design as with higher freq bsc they tend to 'embelish' sources.
That said that don't tell how much directivity he targeted at 300hz? I assumed from my large mains ( 54cm width) bsc center freq is circa 225hz. The SCIV were something like 40cm wide iirc, so center bsc is circa 285hz. Which means a 270* beamwidth as target ( +/- 130* each side of axis). Not far from a cardioid pattern ( but without back attenuation of course).
To have directivity management at such low freq there is not many ways: either you go with waveguide or some kind of cardioid box are the obvious one ( to me at least). Neither of which he used in his own design so it only let one possible thing: baffle step from 'large width' box as it's a way to achieve some kind of 'smooth' transitioning from omni to hemispheric pattern.
Patrick Bateman's some years ago rambling on Von Hoecks box design to manage directivity got me into this, as until i read his posts i understood bsc only as a frequency related issue ( because it's displayed as this, high freq going up 6db over 4 octaves, i find there should be a directivity pattern transition accompanying this freq graph as this is not evident to me displayed that way even if i know it come from radiation pattern constraining...)... study of Dutch & Dutch 8c and how Martijn incorporated baffle step to achieve a wider freq area of directivity management ended convinced me this was what Dunlavy was about giving hint of this freq range and directivity behavior.
I might be wrong but there is enough evidence from other source to convince me it's what it's all about.
So how i interpret this cryptic sentence ( to me) is if there is omni radiation in the 300/100hz range then human voice sound chesty, and other instruments 'warmer' than they are in real life. Which could also explain why people tends to like small width design as with higher freq bsc they tend to 'embelish' sources.
That said that don't tell how much directivity he targeted at 300hz? I assumed from my large mains ( 54cm width) bsc center freq is circa 225hz. The SCIV were something like 40cm wide iirc, so center bsc is circa 285hz. Which means a 270* beamwidth as target ( +/- 130* each side of axis). Not far from a cardioid pattern ( but without back attenuation of course).
Last edited:
Yes Allen but then directivity is omni, and John insisted on this point so from his pov there is an issue there ( design goal) . At least this is how i interpret it.
It depends on how you couple it to the corner. You may be thinking of pointing a source into the corner but you could also install a source out of the corner. You could do point or cylindrical.
It's not easy but it is worth it.
It's not easy but it is worth it.
Hmm. Allen are you talking using corners as waveguide ( a la Klipsh or W.Payrham ( pi speakers) way) or direct radiating driver corner coupling (Allison effect)?
And i'm not saying it's bad or wrong, only i try to get what John Dunlavy design goal criteria were regarding this weird sentence.
And i'm not saying it's bad or wrong, only i try to get what John Dunlavy design goal criteria were regarding this weird sentence.
Then regardless of how, he is right to say that eliminating the baffle step and doing something more consistent over the whole spectrum is worthwhile.
Using the corner as a waveguide is a way to assist in that goal.. even though there are several exceptions in practice, for example the treble.
Using the corner as a waveguide is a way to assist in that goal.. even though there are several exceptions in practice, for example the treble.
I was not implying that you would get the same result in all listening environments. Just that you would be replaying the recording in a consistent way. Of course, the room would affect how it sounds, but no more than it would affect a live instrument. The problem with recordings is that they vary widely, which adds a level of uncertainty with regards to what kind of loudspeaker system suits the recording best.
We agree about SBIR in highs. And consistency/smooth behavior over as wide a range as possible.
But i don't understand he saying he try to eliminate baffle step, rather use it to achieve a goal over something else/side effect.
Some 'Aikido' approach to borrow a Martijn concept regarding how he play with the room rather than against it.
Clever, clever way to approach things imho.
@AllenB Is right, in a corner you already disperse omnidirectonal - in every direction, since the walls mirror the sound (-> Floyd Toole). But at the same time you eliminate the short reflections, which is the reason on why to avoid omnidirectional dispersion.
E: That's the same reason why you can measure low frequency so well in a groundplane measurement setup.
No need to try convince me about Allison effect ICG, it's been a year i'm well into Dutch & Dutch approach and i keep on telling how smart i find Martijn approach... 🙂
I see the point to use corner as a way to extend waveguide too ( Wayne Payrham approach i studyed too), here again you are preaching to the choir. 😉
Or Earl's pov about the fact that sub 500hz early reflections/ direct sound and room behavior tends to merge together ( i use wide box loudspeakers and never got the 'hate' about this kind of design...), i've studyed what Earl said too.
It's really how to make the link between what John Dunlavy's stated and how he implemented into his design which i have interest into.
I see the point to use corner as a way to extend waveguide too ( Wayne Payrham approach i studyed too), here again you are preaching to the choir. 😉
Or Earl's pov about the fact that sub 500hz early reflections/ direct sound and room behavior tends to merge together ( i use wide box loudspeakers and never got the 'hate' about this kind of design...), i've studyed what Earl said too.
It's really how to make the link between what John Dunlavy's stated and how he implemented into his design which i have interest into.