The test I linked shows the frequency axis.
Test in Sound & Recording
The diagrams of Sound & Recording are always normalized.
Btw, directivity of the LSR 305 & 308 looks very similar regarding the waveguide.
Sound & Recording LSR 305 & 308
There is always strong beaming in the highs.
Sound & Recording LSR 305 & 308
There is always strong beaming in the highs.
The strong beaming is fairly common. I'm more concerned about the lack of pattern control even for the M2 (as Earl has said, nothing better in the LSR305's price-range, but that is inexcusable for the M2). Seems like there is little advantage compared to a good direct radiator design except for the high SPL the system is capable of.
Common maybe, but not up-to-date.
A well designed waveguide maintains a nearly constant directivity up to the highs. Please take a look at the recent models from Neumann and Genelec. They don't start to beam above 10 kHz. They beam at much higher frequencies and the KH 420 doesn't beam at all up to 20 kHz considering the first two isobares.
Neumann KH 420:
An externally hosted image should be here but it was not working when we last tested it.
I designed waveguides myself. Yes, it is possible to avoid beaming.
The LSR 305 looks best in terms of directivity of the three models. The small woofer and the enclosure fits best to the waveguide. The bigger models are worse in this regard.
So Earl is right when he said that the 305 is a very good speaker for that price.
Common maybe, but not up-to-date.
A well designed waveguide maintains a nearly constant directivity up to the highs. Please take a look at the recent models from Neumann and Genelec. They don't start to beam above 10 kHz. They beam at much higher frequencies and the KH 420 doesn't beam at all up to 20 kHz considering the first two isobares.
Neumann KH 420:
An externally hosted image should be here but it was not working when we last tested it.
I designed waveguides myself. Yes, it is possible to avoid beaming.
The LSR 305 looks best in terms of directivity of the three models. The small woofer and the enclosure fits best to the waveguide. The bigger models are worse in this regard.
So Earl is right when he said that the 305 is a very good speaker for that price.
Agreed on the 305. The only blemish would the inevitably higher baffle-step frequency at which the 4-pi transition is achieved.
When I said common, I meant in the general loudspeaker world. At the cutting-edge ultra-high-end, the M2's performance is really quite average. No advancement on OS waveguides that I can see.
There are even cheap models with waveguides or horns doesn't beam that strong. It is not a matter of price.
But you are right. Typical hifi speakers beam much, too.
I partly agree. But the oblate spheroid waveguide is not a reference in terms of directivity control. All normalized diagrams I've seen (and simulated myself) looked even worse than the M2.
Gedlee Nathan:
The diagrams on Earl's homepage are not normalized and don't show these flaws. But if you take a closer look at his diagrams you'll see a hot spot on 0° in the range of 1 - 3 kHz. This corresponds to the strong narrowing in the normalized diagram.
And this is the reason why I don't like non-normalized directivity patterns. They simply don't show directivity, but frequency responses under different angles. They are misleading.
Last edited:
Even cheap models with waveguides or horns doesn't beam that strong.
I partly agree. But the oblate spheroid waveguide is not a reference in terms of directivity control. All normalized diagrams I've seen (and simulated myself) looked even worse than the M2.
Gedlee Nathan:
The diagrams on Earl's homepage are not normalized and don't show these flaws. But if you take a closer look at his diagrams you'll see a hot spot on 0° in the range of 1 - 3 kHz. This corresponds to the strong narrowing in the normalized diagram.
And this is the reason why I don't like non-normalized directivity patterns. They simply don't show directivity, but frequency responses under different angles. They are misleading.
Interesting. The 1-3kHz narrowing from the OS is much more than those I've seen from SEOS measurements, but I suppose that is to be expected from a more elliptical shape. Nevertheless, of particular interest is the seemingly lesser beaming from the Nathan compared to the M2. I suspect the 'reflectors' JBL put at the horn aperture are much of the reason. They said as much in their promotional material: that the 10-20kHz performance was deliberate and something they designed into the waveguide/driver combo. Probably a result of their listening tests.
Elliptical or not is not important. The contour forms the directivity (together with driver's size etc.). And both dimensions (x,y) can be considered independently.
Regarding SEOS: from what I've seen it seems to be a pretty good horn. But there are other good horns with equal directivity. The only special thing about it is the pretty low price.
I don't see any advantage in JBL's "weird" contour. In my eyes it is nothing more than an optical feature. It makes their speakers look more unique and "technical advanced".
Last edited:
Elliptical or not is not important. The contour forms the directivity (together with driver's size etc.). And both dimensions (x,y) can be considered independently.
Regarding SEOS: from what I've seen it seems to be a good horn. But there are other good horns with equal directivity. The only special thing about it is the pretty low price.
I don't see any advantage in JBL's "weird" contour. In my eyes it is nothing more than an optical feature. It makes their speakers look more unique and "technical advanced".
Yes, SEOS is the cheapest 'good' horn I've seen.
As for JBL's contour, I'm surprised you'd say that. It seems to be a compromise between high-output and wide coverage. Why wouldn't JBL's complex contour offer anything more than OS/SEOS or any common DIY design?
Waveguides/horns are always a compromise between efficency and directivity (especially in the highs). But JBL is not any better in this regard than others. Please compare maximum SPL from JBL M2 and Neumann KH 420. M2 has worse directivity and worse maximum SPL in the mids/highs. So where is the advantage?
You have picked the worst example of an OS to compare with speakers that cost many many times what a Nathan cost. Why don't you look at an NS-15, which would blow away any example that you have shown (except that the plotting is done differently which makes comparison almost impossible.) I abhor normalized plots because they completely distort the picture if there is anything non-flat on-axis. Since I don't design my speakers for listening on-axis I don't worry so much about it. To normalize and be even remotely fair you would have to do this along the 20 degree axis.
Basically I don't see your discussion as fair. You admitted that the designs that you praise are yours - a little convenient? And then you show the worst examples of the competition.
I'm sorry, but this just isn't true. The axial hole that you see on the Nathan is a result of the coherent adding of the edge diffraction (which is out of phase) in the far field. If the edge where elliptical (or square) then this coherence would be smeared and the hole would go away (or at least be minimized). If you are a waveguide designer I would have thought that you would understand this.
Hence, what happens is that when you normalize the data, the axial hole, which is NOT part of the directivity control, completely distorts the control provided by the device. A larger mouth, with a larger radius would almost completely negate the hole (as it does in the NS-15). Making the mouth elliptical - even if the waveguide stayed round - would also minimize the axial hole.
Last edited:
I understand your point by not listening on-axis and showing not normalized plots. But don't claim that your waveguides are "constant directivity". They are not. Post a normalized plot and averyone can see it.
I never stated that they are my designs. What makes you think that?
I don't work at Neumann or Genelec. Designing speakers is just a hobby of mine.
Show me your best design in a normalized plot and I will discuss it as objectively as I discuss Neumann's or Genelec's speakers (they have worse designs as well). I honor every good design, not whole companies.
The on-axis hole is a flaw of your contour. Of course it will be smeared with a different contour in y direction. But it is easily possible to design a circular waveguide without this on-axis hole. And this is easy to prove.
This is wrong - clearly - since the hole is absent in the NS-15 and has exactly the same contour as the Nathan.
Show me a 90 degree waveguide that is round, only 10" across that doesn't have a hole and I will agree with you. Otherwise you are just stating false opinions that you have.
It is far easier to make a wide waveguide without the axial hole than a narrow one. None of the examples that you show are as narrow as mine.
Of course the whole is part of the directivity. Even if you call it different.
Who made you "king" of definitions?
Then please post it normalized. The diagram on your website shows the same peak under 0° around 2 kHz as your other speakers. The hole may be absent, but the directivity is still not constant.
Sica Q07030A (80°):
Limmer 860 (60°):
Both radiate with even less than 90°. They don't show an on-axis hole (above 10 kHz the driver doesn't fit good enough on the Limmer). The contour is just different.
This is nonsense. You just have to optimize the contour manually and don't apply a simple mathematical model.
No need to be offending. The definition is simple and dictated by logic. It can be derived by the energy response which is well defined. When you don't normalize it to 0° then you show only frequency responses under different angles. The raw directivity can only be seen when normalized.
Normalized plots are standard in the industry.
And btw. a hole at 0° means the same for directivity as a widening under higher angles. These artifacts are equivalent. In directivity and energy response you can't differentiate them. Of course you can optimize for energy response and "fill" it with reflections. But what if the room doesn't reflect enough in this frequency region? In my eyes this is a bad tradeoff.
Attachments
Dr.Keele about CD horn midband beaminig
It seems to me that mid-range beaming is something that can be expected, on almost conical horn, as Dr. Keele has shown years before, and has suggested to enhance horn wall flare in the last 1/3 section of the horn towards horn mouth.
"Keele_(1975-05_AES_Preprint)_-_Whats_So_Sacred_Exp_Horns.pdf"
regards
ivica
Hi,There are even cheap models with waveguides or horns doesn't beam that strong. It is not a matter of price.
But you are right. Typical hifi speakers beam much, too.
I partly agree. But the oblate spheroid waveguide is not a reference in terms of directivity control. All normalized diagrams I've seen (and simulated myself) looked even worse than the M2.
Gedlee Nathan:
The diagrams on Earl's homepage are not normalized and don't show these flaws. But if you take a closer look at his diagrams you'll see a hot spot on 0° in the range of 1 - 3 kHz. This corresponds to the strong narrowing in the normalized diagram.
And this is the reason why I don't like non-normalized directivity patterns. They simply don't show directivity, but frequency responses under different angles. They are misleading.
It seems to me that mid-range beaming is something that can be expected, on almost conical horn, as Dr. Keele has shown years before, and has suggested to enhance horn wall flare in the last 1/3 section of the horn towards horn mouth.
"Keele_(1975-05_AES_Preprint)_-_Whats_So_Sacred_Exp_Horns.pdf"
regards
ivica
By your strict definition nothing that has been shown is "constant directivity". I am not arguing this issue with you anymore because we simply have different definitions.
But your statement above "And btw. a hole at 0° means the same for directivity as a widening under higher angles. These artifacts are equivalent. In directivity and energy response you can't differentiate them."
This is entirely false simply because different angles represent different areas and as such different energies (since energy is a per unit area quantity. Thus a hole on axis, has a very small if not negligible effect on the power response, while a peak off axis has a huge effect. This is precisely why I do not normalize the plots because it completely distorts the energy by increasing the off axis - in terms of energy - more than the on-axis. Visually it is completely wrong.
What is correct is to show the data unchanged and plot both the power and the DI. This is a perceptually correct interpretation (which Toole and Olive also use). I will never accept normalized plots as meaningful.
Last edited:
Hi ivica - it is simply misunderstood that this "beaming" is a result of the treatment of the mouth more than anything to do with the contour. Small and Fincham also found this at KEF when they put a tweeter at the apex of the woofer. There was an axial hole, which initially they did not understand. I explained it to them because I had seen this before in simulations and in real devices. It is not evident in anything but a perfectly round device. Its location and depth depend on the wall angles and the mouth radius. The hole tends downward in frequency and lower in magnitude with wider mouths and wider angles. So worst case is narrow angle with a small mouth - i.e. a Nathan.
Last edited: