Why Lower Crossover Frequency Does Not Always Improve Loudspeaker Directivity

[Arindal]

Mark Audio has some interesting drivers, eg Alpair 7ms . Still rolls of from around 5khz though would be less than many other drivers.

Never worked with that one, but measurements suggest directivity narrows down between 5k and 10k, maybe a bit smoother than with other rigid-diaphragm fullrange models.

Regarding treble directivity, something like Dayton RS75 is better than most fullrange cones, but it is rather a toy, and you would hear some of the phaseyness issues I have mentioned.

KEF developed the tangerine waveguide (aka phase plug) along with the MF cone shape to broaden the beam width at higher frequencies.

The beam width in the listening window is indeed broadened, everything looks excellent in a +-45deg horizontal window. But the more you move away from the 0deg axis, the more midrange cone and baffle are defining the directivity, and that one becomes pretty uneven.

The directive index figures published provide evidence that this goal has been achieved in practice. Not perfect, but very good, in my opinion

It depends on the application and window you are looking at. If we are talking solely about the listening window and typical early reflections from ceiling and floor, I agree. So in some kind of well-treated loft studio this might work. In your typical living environment, you might have reflective walls either behind the speakers or on the sides at a limited distance, so these define early reflections and reverb tonally.

How would one theoretically reach a constant directivity from a diaphragm driver, without crossing over to a super tweeter

For broad dispersion / low directivity index: Comparably small midrange (like 3" or 4") plus small tweeter (3/4" or dome-shaped 1"), both in a baffle-less or very slim baffle arrangement, with all drivers working above 300Hz in close proximity to each other (or in coaxial arrangement). In theory, B&W Nautilus (the original 4-way model) was such a concept.

If you want narrower dispersion or higher directivity index: Well, it is complicated. You have to define your ideal listening window pattern and directivity index and combine several measures of controlling directivity for different frequency bands (waveguide, bigger diaphragms, cardioids, line sources).

We should note that constant directivity is from psychoacoustical point most useful for frequency bands which our brain can distanguish between direct and reflected sound events. What happens below 300Hz and above 8,000Hz, is IMHO not as critical in terms of directivity, if the latter is not completely off.

 

[Ditonica]

It is significant enough in my understanding to create a disbalance between the highest octave the woofer is playing alone, the transitional band, and the lowest octave of the tweeter. With a non-waveguided tweeter, I would personally always opt for smaller midrange cones to have a smooth transition and balance between broader bands alike.

I agree. If designers have the opportunity to use either a smaller midrange driver or a wave-guided tweeter, they should do this. Both problems, diffraction and lobing, as well as the directivity of drivers themselves, exist, and both should be considered and treated.

The main problem with lobing is that it is not only affecting the overall d.i., but leading to significant dips in frequency response of the ceiling and floor/desk/console reflections, which is pretty important both for perceived tonality and imaging, while leaving the important side-wall reflections, responsible for localization issues, untouched. A constant d.i. is in general a reasonable goal in my understanding, but not should be achieved by balancing out dips in different windows.

I think this is quite an interesting topic as well, because the strict separation of horizontal and vertical directivity sometimes is just a theoretical simulation. Let's say a situation where the loudspeaker placement is parallel to the walls (a very common situation and even a manufacturer's recommendation for some loudspeakers). Technically, the floor reflection in the listening position between the speakers consists of both vertical and horizontal directivity. In addition, our listening rooms are not just walls; if a chair is closer to your listening position than the left/right wall, the earliest reflections can also be off the tweeter level, but lower or higher, so this horizontal reflection will have a vertical component.

 

[Arindal]

the strict separation of horizontal and vertical directivity sometimes is just a theoretical simulation. Let's say a situation where the loudspeaker placement is parallel to the walls (a very common situation and even a manufacturer's recommendation for some loudspeakers). Technically, the floor reflection in the listening position between the speakers consists of both vertical and horizontal directivity.

I fully agree, and I recommend not to extensively stare at either vertical or horizontal groups of frequency response graphs and try to find even lines, as we do not need a precise calculation of a response at a precise angle, but rather an idea of what happens averaged with octave-broad frequency bands in a certain window. This is why in my understanding the spinorama calculations for ´side wall bounce´, ´ceiling/floor´ or rear hemisphere, are much easier to understand and reliably tell you what will happen in a room depending on the distance to the nearby boundaries. Taking your example, the lobing issue for the floor bounce would persists under different horizontal angles (aka toe-in), as the difference in distance between midrange and tweeter persists as well.

 

[Jonazkur]

Problem in higher xo frequensies is rising H3 in bass drivers. Its really hard to find for example 8” driver for 2 way that doesnt have distortion problems at +1k

 

[Waveform Fidelity]

Sure, we can also make a dedicated thread and share measuremens if you want, I'm game

Hi there

Havent got around to yanking out drivers for dimensions, but been toying with swapping SP1632 to LS50 drivers SP1753CA, until I noticed the VAS is about 6X greater , which makes sense because the LS50 driver is running in to the complete enclosure with a bass port, but the SP1632 in the R900 has its own small enclosure (hence acoustic roll-off of 12dB/Oct at around 400Hz) so 2 questions

Your R3 meta drivers are SP1753AA and LS50 are SP1753CA - my guess is that your R3 must be a stiffer driver (low compliance) in its own enclosure like the R900. In your design do you run it with its own enclosure and roughly what is that volume. Do you happen to know the VAS of the R3 driver and the Qts of your setup?…. Just wondered. I think for me the LS50 in a small enclosure is asking for trouble

 

[Ditonica]

Problem in higher xo frequensies is rising H3 in bass drivers. Its really hard to find for example 8” driver for 2 way that doesnt have distortion problems at +1k

Good point. An alternative solution is to choose a wider baffle to place the crossover point lower and avoid a common dip problem. Or a 3-way system with a smaller midrange.

 

[Jonazkur]

Good point. An alternative solution is to choose a wider baffle to place the crossover point lower and avoid a common dip problem. Or a 3-way system with a smaller midrange.

Imo best is a proper horn/waveguide and bass driver and cross low. Wider baffle could introduce other problems like diffraction and midrange narrowing.

 

[Arindal]

Imo best is a proper horn/waveguide and bass driver and cross low.

In theory a good solution, but practically you oftentimes increase lobing issues due to greater distance between bass and horn/waveguide center, as well as overly narrowing directivity in the upper treble bands. The latter comes as a side effect of many horns and waveguides being chunky in order to reach lower, so they narrow down treble dispersion too much. And oftentimes they just don´t sound right, from horn colorations to restricted midrange transparency.

Could you name an example which you find exemplary in this regards pls?

 

[Jonazkur]

In theory a good solution, but practically you oftentimes increase lobing issues due to greater distance between bass and horn/waveguide center, as well as overly narrowing directivity in the upper treble bands. The latter comes as a side effect of many horns and waveguides being chunky in order to reach lower, so they narrow down treble dispersion too much. And oftentimes they just don´t sound right, from horn colorations to restricted midrange transparency.

Could you name an example which you find exemplary in this regards pls?

There are good and bad waveguides/horns, lobing can be minimized when xo point is low enough and steep filter. And i rather take good horizontal and have little suckout in vertical what happens almost in every 2 way design to a degree. Here is polar from my 2 different horns made for dome tweeter without midrange or treble narrowing and pretty constant directivity. Both are 8” rectangular horns and pretty deep 6-7cm.

Horn coloration and bad midrange is often
associated with old poor horn designs with little knowledge about horns.

 

[Arindal]

lobing can be minimized when xo point is low enough and steep filter.

In theory yes. But how low and how steep would you filter if we are talking about 35cm C2C?

i rather take good horizontal and have little suckout in vertical what happens almost in every 2 way design to a degree.

Everyone is free to choose own compromises, but I would rather try to find a balance between horizontal and vertical reflections´ tonality. The latter is important in my understanding as ceiling and floor reflections are usually the first ones to arrive at the listener´s ear. And since coaxials and 3-way concepts have been invented, I don´t really see a reason for compromising here.

Here is polar from my 2 different horns made for dome tweeter without midrange or treble narrowing and pretty constant directivity.

Interesting. Would it be possible to see the rear hemisphere polar plot pls, i.e. 90-180deg horizontally?

Horn coloration and bad midrange is often
associated with old poor horn designs with little knowledge about horns.

Don´t mean to disagree with you, as I tend to believe that in theory this is possible. But obviously it is difficult for a consumer to identify what are commercially available ´newer horns´, engineered by people who understand about sound quality. My personal skepticism is pretty much based on very disappointing expensive products, engineered in recent decades by the biggest names in horn design, among them brands with a reputation for having the only research institute for sound quality, so to say the scientific gold standard of sound quality ;-)

 

[markstein]

In theory yes. But how low and how steep would you filter if we are talking about 35cm C2C?



Everyone is free to choose own compromises, but I would rather try to find a balance between horizontal and vertical reflections´ tonality. The latter is important in my understanding as ceiling and floor reflections are usually the first ones to arrive at the listener´s ear. And since coaxials and 3-way concepts have been invented, I don´t really see a reason for compromising here.



Interesting. Would it be possible to see the rear hemisphere polar plot pls, i.e. 90-180deg horizontally?



Don´t mean to disagree with you, as I tend to believe that in theory this is possible. But obviously it is difficult for a consumer to identify what are commercially available ´newer horns´, engineered by people who understand about sound quality. My personal skepticism is pretty much based on very disappointing expensive products, engineered in recent decades by the biggest names in horn design, among them brands with a reputation for having the only research institute for sound quality, so to say the scientific gold standard of sound quality ;-)

Which commercially available loudspeakers you would recommend?

All I am aware of are a compromise in the one or other direction.

 

[Jonazkur]

In theory yes. But how low and how steep would you filter if we are talking about 35cm C2C?



Everyone is free to choose own compromises, but I would rather try to find a balance between horizontal and vertical reflections´ tonality. The latter is important in my understanding as ceiling and floor reflections are usually the first ones to arrive at the listener´s ear. And since coaxials and 3-way concepts have been invented, I don´t really see a reason for compromising here.



Interesting. Would it be possible to see the rear hemisphere polar plot pls, i.e. 90-180deg horizontally?



Don´t mean to disagree with you, as I tend to believe that in theory this is possible. But obviously it is difficult for a consumer to identify what are commercially available ´newer horns´, engineered by people who understand about sound quality. My personal skepticism is pretty much based on very disappointing expensive products, engineered in recent decades by the biggest names in horn design, among them brands with a reputation for having the only research institute for sound quality, so to say the scientific gold standard of sound quality ;-)

For example i have these diy speakers with 20cm C2C and 1100hz xo with 24db/oct slope.
With 35cm you want under 1k xo with atleast 24db/oct.

I havent personally heard any advantages in coaxial speakers vs good 2 ways when listening. Ive heard every genelec model and also blade 2 metas. Maybe very close nearfield is good with coaxials.

I can take better measurements in summer when i can get outside.

Many commercial horn speakers have some issues and arent always that well executed.

 

[Arindal]

Which commercially available loudspeakers you would recommend?

Very much depends on the application in question. But if you are referring to an optimal compromise in terms of all directivity-related issues, i.e. even response throughout the listening window, while avoiding vertical lobing, continuously narrowing dispersion and unwanted widening at the same time, there are several promising coaxials on the market. Some TAD models, MEGs, Genelecs with smaller coaxial driver (8341A for example), KSD, even some budget Technics and Elac come to mind.

have these diy speakers with 20cm C2C and 1100hz xo with 24db/oct slope.

That would exactly meet the maximum x-over frequency ceiling as calculated by D´Apollito, so I agree that a smooth transition with minimum of vertical lobing appears possible. The other question is how the tweeter dome actually sounds between 1 and 2K, and how it is possible to keep the dispersion to the rear hemisphere wide above 6K with what appears to be a linearly flared horn shape. Thanks for investigating and all the best for your concept!

I havent personally heard any advantages in coaxial speakers vs good 2 ways when listening

I agree it is depending on the circumstances and expectations in terms of recordings, if this is really audible and decisive. With classical recordings containing meaningful reverb enveloping the phantom sources, and monitoring conditions involving console/desk reflections as well as a ´moving target´, I think it is pretty much audible. Maybe I am biased here, as I have been working with typical 2-way monitors early in my pro audio career (Genelec, Mackie, K+H and others), and that was the main shortcoming in my understanding.

 

[markstein]

Very much depends on the application in question. But if you are referring to an optimal compromise in terms of all directivity-related issues, i.e. even response throughout the listening window, while avoiding vertical lobing, continuously narrowing dispersion and unwanted widening at the same time, there are several promising coaxials on the market. Some TAD models, MEGs, Genelecs with smaller coaxial driver (8341A for example), KSD, even some budget Technics and Elac come to mind.

Interesting than you mention Geithain as a positive example in this context. Measurements of these are very rare and the few I found online do not look good...

 

[Arindal]

Measurements of these are very rare and the few I found online do not look good...

If I recall it correctly, they used to provide pro audio customers with detailed measurements including diffuse response graphs from a reverberant chamber. The one you linked, seems to look weird in a narrow band around 1.5K, that might be either linked to the specific type of open baffle cardioid used, or even be intentionally to apply some HRTF/diffuse-field correction to the reverb. Response within the listening window and broad-banded off-axis behavior look not too far off, particularly the midrange directivity control seems to work.

 

[Jonazkur]

Interesting than you mention Geithain as a positive example in this context. Measurements of these are very rare and the few I found online do not look good...

Looks like heavy diffraction and its no suprise if you look at the baffle.

 

[Arindal]

Looks like heavy diffraction and its no suprise if you look at the baffle.

Could you mark where you see evidence of diffraction in this graph, except from very narrow-banded ripples around 4K?

 

[Jonazkur]

Could you mark where you see evidence of diffraction in this graph, except from very narrow-banded ripples around 4K?

1-5k area there is almost 5db dip on axis. also 30 and 60 degrees huge peaks and dips its either diffraction or directivity is a mess on that area.

 

[Digital_Thor]

Hi there

Havent got around to yanking out drivers for dimensions, but been toying with swapping SP1632 to LS50 drivers SP1753CA, until I noticed the VAS is about 6X greater , which makes sense because the LS50 driver is running in to the complete enclosure with a bass port, but the SP1632 in the R900 has its own small enclosure (hence acoustic roll-off of 12dB/Oct at around 400Hz) so 2 questions

Your R3 meta drivers are SP1753AA and LS50 are SP1753CA - my guess is that your R3 must be a stiffer driver (low compliance) in its own enclosure like the R900. In your design do you run it with its own enclosure and roughly what is that volume. Do you happen to know the VAS of the R3 driver and the Qts of your setup?…. Just wondered. I think for me the LS50 in a small enclosure is asking for trouble

Well, the LS50 is modified to play a little bass, whereas the R3 is a dedicated midrange.. soo
I definitely use a small dedicated chamber for my R3 coax.... Almost copy paste from the R900... Just better coax, lower cross over and bigger woofers.

 

[Arindal]

1-5k area there is almost 5db dip on axis.

Strange, I don´t see that. There is a -2.5dB dip around 2.7K and a -2dB @4K on axis, but both very narrow-banded.

30 and 60 degrees huge peaks and dips its either diffraction or directivity is a mess on that area.

I see no evidence of diffraction here except for the little 4K ripple which might correspond with the intentional edge diffraction at the sub baffle´s edges.

Trying to judge directivity from just three different angles is a bit difficult, but all the peaks and dips you are mentioned, are pretty narrow-banded and alternating, which hints to the midrange open baffle cardiod as a cause. Not sure if this 1.5K behavior is intentional, but I wouldn't really worry about it. After all, perceived tonality of reverb is not a matter of straight lines on a graph, but rather of the relative SPL of neighboring octave-broad bands (or broader). Particularly around 1K and below it looks pretty good, and that is where usually problems originate from.