Matching Directivity

[Trdat]

What's the simplest way to match directivity? For a DIY enthusiast who uses software DSP and knows enough about REW.

Firstly how do we do it on paper by interpreting driver specs?

Then once drivers are purchased how do we measure them to achieve controlled directivity?

I get we measure the off axis response. But what are we looking at after measurements have been taken? And do we have to gate the measurement or can we just interpret the measurement as is?

 

[alex-z]

You are looking to crossover the drivers at the point where their off-axis response is similar. This is the simplest form of directivity control, and is easily done by adding a waveguide to your tweeter. A wide baffle will assist in controlling the directivity further down.

A more advanced technique is using rear or side firing woofers to create cardoid mid-bass. Dutch & Dutch 8C is an example, the directivity is constant from 100-12000Hz AFAIK.

Edit: Oops, I meant the Kii Audio Three for the cardoid design.

 

[abdo123]

A more advanced technique is using rear or side firing woofers to create cardoid mid-bass. Dutch & Dutch 8C is an example, the directivity is constant from 100-12000Hz AFAIK.

The D&D8C does not do this, it releases out of phase sound from the cabinets via open slots on its side. The back woofers are merely subwoofers.

 

[Trdat]

You are looking to crossover the drivers at the point where their off-axis response is similar.

I have come across a wide baffle helping with directivity is there any hard and fast rules? Don't find too much info on that around...

When you say the off axis is similar what am I looking at? I am just tyring to get my head around this.

Below is the response for my tweeter and waveguide and the graph after is the response of my 4 inch. If you could direct me to what I am looking at and what the possible crossover point might be if directivity is the priority.

A more advanced technique is using rear or side firing woofers to create cardoid mid-bass. Dutch & Dutch 8C is an example, the directivity is constant from 100-12000Hz AFAIK.

Thats my next DIY project, cardoid mid bass with motion feedback for subs, absolutely massive horn. But first before I jump ahead need to get the basics downpat.

The above is a Scanspeak tweeter D2608 with a Scanspeak 4 inch 12W/8524G00.

The waveguide is

 

[alex-z]

The D&D8C does not do this, it releases out of phase sound from the cabinets via open slots on its side. The back woofers are merely subwoofers.

Oops, got it confused with the Kii Audio Three.

 

[Plcamp]

I have come across a wide baffle helping with directivity is there any hard and fast rules? Don't find too much info on that around...

When you say the off axis is similar what am I looking at? I am just tyring to get my head around this.

Below is the response for my tweeter and waveguide and the graph after is the response of my 4 inch. If you could direct me to what I am looking at and what the possible crossover point might be if directivity is the priority.

View attachment 177517


Thats my next DIY project, cardoid mid bass with motion feedback for subs, absolutely massive horn. But first before I jump ahead need to get the basics downpat.

The above is a Scanspeak tweeter D2608 with a Scanspeak 4 inch 12W/8524G00.

The waveguide is

If your waveguide is not approx the same width as your 4” driver, the off axis response will diverge from on axis response differently with frequency. What you are looking for is a frequency that is not too low (to evade tweeter distortion) at which the off axis response of the tweeter diverges by the same amount as the 4”. From the other direction you want that match to be not too high to cause the 4” breakup modes to become visible.

 

[alex-z]

I have come across a wide baffle helping with directivity is there any hard and fast rules? Don't find too much info on that around...

When you say the off axis is similar what am I looking at? I am just tyring to get my head around this.

Below is the response for my tweeter and waveguide and the graph after is the response of my 4 inch. If you could direct me to what I am looking at and what the possible crossover point might be if directivity is the priority.

The wider your baffle, the more low frequency energy is directed forward. A flat baffle controls the response 90 degrees to either side. The reason you don't see huge baffles is because the drivers themselves are not radiating that widely.

A practical example of baffle size assisting directivity is in-wall speakers. They have much smoother mid-bass response because there is no baffle step loss or associated reflections.

Start by looking at the woofer. You can see it is tightly grouped until 1000Hz, and then gradually drops off, before response anomalies start at 3000Hz. Your tweeter is also tightly grouped in the 1800-3000Hz range, so you want the crossover somewhere in that region. Lower is generally better because you get tighter vertical dispersion.

Your waveguide design doesn't look much like an optimized waveguide. The throat is relatively narrow so it will act more like a horn, an ideal waveguide generally has the same flare profile as the woofer diameter. If you want to see proper examples, look here:

Waveguides | somasonus

www.somasonus.net

 

[Trdat]

The wider your baffle, the more low frequency energy is directed forward. A flat baffle controls the response 90 degrees to either side. The reason you don't see huge baffles is because the drivers themselves are not radiating that widely.

Fascinating, I had come across this concept but not in its entirety. My understadning from what you have mentioned is that if the waves radiate 90 degrees such as below 200hz there is benefit to a wider baffle.

Start by looking at the woofer. You can see it is tightly grouped until 1000Hz, and then gradually drops off, before response anomalies start at 3000Hz. Your tweeter is also tightly grouped in the 1800-3000Hz range, so you want the crossover somewhere in that region. Lower is generally better because you get tighter vertical dispersion.

This I understand, I just don't get if its so simple how come so many companies get it so wrong? Or is it significantly more complicated than this when it comes to the crunch...?

Your waveguide design doesn't look much like an optimized waveguide. The throat is relatively narrow so it will act more like a horn, an ideal waveguide generally has the same flare profile as the woofer diameter. If you want to see proper examples, look here:

I know about Sonus but he didn't he have a waveguide for the D2608 tweeter so I went with one that was made specifically for the tweeter I had. I am not too fussed if it acts like more like a horn but as its not the same width as the woofer I am curious if its going to be a problem. And also curious what the tradeoffs and advantages a guide like this has compared to the Sonus waveguides?

His design is an MTM which simplifies the C-C distance issue but don't know how he mitigated the low end loading of the guide of which was narrower in width than the 4 inch he used.

 

[alex-z]

Fascinating, I had come across this concept but not in its entirety. My understadning from what you have mentioned is that if the waves radiate 90 degrees such as below 200hz there is benefit to a wider baffle.


This I understand, I just don't get if its so simple how come so many companies get it so wrong? Or is it significantly more complicated than this when it comes to the crunch...?



I know about Sonus but he didn't he have a waveguide for the D2608 tweeter so I went with one that was made specifically for the tweeter I had. I am not too fussed if it acts like more like a horn but as its not the same width as the woofer I am curious if its going to be a problem. And also curious what the tradeoffs and advantages a guide like this has compared to the Sonus waveguides?

His design is an MTM which simplifies the C-C distance issue but don't know how he mitigated the low end loading of the guide of which was narrower in width than the 4 inch he used.

Keep in mind a 200Hz wavelength is 1.7 metres. A baffle that wide isn't realistic unless you are building speakers into a wall. If you build something more realistic like a .5 metre baffle width, the directivity improvement would largely be in the 700-1500Hz region where the woofer is still radiating widely.

You can pretty easily adapt a waveguide to suit a different tweeter if the cutout diameter is the same. Just need to change the mounting.

A steeper waveguide profile causes narrower dispersion, and more overall output. This might be done for reduced distortion in the crossover region, rather than having an ideal directivity match.

 

[Trdat]

Keep in mind a 200Hz wavelength is 1.7 metres. A baffle that wide isn't realistic unless you are building speakers into a wall. If you build something more realistic like a .5 metre baffle width, the directivity improvement would largely be in the 700-1500Hz region where the woofer is still radiating widely.

So techinically you could do a .5 - 1 metre baffle those specific frequencies around that wavelenght, use a horn above the 700hz mark which is quite typical in a horn design and use cardoid or a large horn for the woofer frequencies below 600hz ish. Would like to see more DIY designs like this to emulate.

You can pretty easily adapt a waveguide to suit a different tweeter if the cutout diameter is the same. Just need to change the mounting.

I kind of new this but was a little hesitant to approach this myself. After a post in his Sonus's thread in which he mentioned that there isn't a specific waveguide for the D2608 I presumed that was due to the throat not matching to the tweeter inlet. But it seems I was wrong and it more about the mounting, if I have understod correctly.

A steeper waveguide profile causes narrower dispersion, and more overall output. This might be done for reduced distortion in the crossover region, rather than having an ideal directivity match.

Well, directivity would have been the preferance but as the design is for a centre speaker that is going to be somewhat similar to an already made left and right(Troel 3WC) but with a an added waveguide perhaps the narrower dispersion would work for the centre.

If it was a 3 way designed from scratch for left and right the Sonus waveguide options would and could have been better.

 

[Vineeth Kumar V]

For reference, here is an example of a wide baffle 3-way speaker designed by Kimmosto (author of VituixCAD software) that has directivity control down to lower midrange frequencies.

KS-483

Here is a thread where Kimmosto has given a whole lot of design tips about how to take care of different directivity-related aspects in every step of speaker design (The whole thread is a golden reference about directivity control aspects):

VituixCAD

Normalized off, Still doesn't look right. The cyan (light blue) response is 30º and it does not look like the merged version.

www.diyaudio.com

Here is another thread where most of the above information and then some more can be found:

VituixCAD v2 - HTGuide Forum

For information, VituixCAD speaker simulator has been updated from version 1.1 to 2.0 with the following additions and changes (so far): Crossover schematic changed from formal block ladder to free form supporting unconventional topologies. Predefined passive blocks replaced with individual...

www.htguide.com

To me personally, reading these threads have been sort of enlightening regarding learning some speaker design concepts..
To very briefly summarize the key ideas, directivity control requires a few things taken care of:

1) Planning and selection of transducers/drivers and directivity control elements such as waveguides, horns, etc based on the overall speaker concept requirements, which may consider approximate intended/required driver crossover frequencies, driver mechanical limits etc

2) Planning and construction of a physical structure/acoustic construct that holds the drivers and let them do their work while minimally interfering/aiding directivity control

3) Planning and implementation of a crossover (active/passive) for the drivers, again designed from the point of view that it is another element that controls directivity.

Thanks
Vineeth

 

[YSDR]

What about the exact directivity matching of the drivers/waveguides? There is some magical happening when they are exactly match around the crossover region?
I ask, because my speaker have a slight (few dB) directivity mismatch between the mid and the horn loaded tweeter. If I want an exact match, then I need a larger mid or a smaller horn.
What do you guys think? Is it worth the effort, or should I just use it the way it is now?

 

[howard416]

The waveguide is

What waveguide is that?

 

[Plcamp]

There is some magical happening when they are exactly match around the crossover region?

I expect it’s better when they match not only at the crossover freq, but on both sides as far as possible…enhanced if they are the same, but probably more important with lower order crossover slopes?

 

[fluid]

What waveguide is that?

It's the one for the latest Elsinore design

https://www.customanalogue.com/elsinore/elsinore_index.htm

 

[TurtlePaul]

What about the exact directivity matching of the drivers/waveguides? There is some magical happening when they are exactly match around the crossover region?
I ask, because my speaker have a slight (few dB) directivity mismatch between the mid and the horn loaded tweeter. If I want an exact match, then I need a larger mid or a smaller horn.
What do you guys think? Is it worth the effort, or should I just use it the way it is now?

The problem when there isn’t a match at the crossover frequency is that you cant have both flat ox-axis response and smooth in-room response. You can have smooth on axis but the speaker will sound bright because of the wide directivity or you can have a pleasant sounding speaker in-room with a BBC dip on-axis. 2 dB strikes me as not crazy though (is this 45 degrees?) and even Revel 2-ways with waveguides can have this much error. I would think you could aim to voice it with a 1 dB on axis dip to minimize the sound power response gain.

 

[airborne]

directivity is mainly a function of dimensions - namely height and width - and in some cases ( such as horns and cardioids ) depth too ...

due to size limitations and cost cutting practical designs compromise directivity ...

if you want exemplary directivity be prepared to have gigantic speakers ...

on the most basic level a tweeter's directivity will match that of the woofer when the tweeter waveguide / horn has comparable height and width to that of the woofer - so for example a 8 X 8 inch tweeter horn will match a 8" woofer well when crossed at around the horn cutoff frequency ...

if possible i would use a larger horn than woofer, like a 10 X 10 horn crossed to a 6" midrange would be super ...

but that's obviously a generalization because horn cutoff frequency is a function of depth and flare rate and not simply mouth area ... so you're assuming here that a horn was designed optimally ... but most horns are designed smaller than they need to be to work properly so at that point you sort of have to think ...

i'm sorry i know it didn't answer your question but yours is a hard question ...

also consider that off-axis and power response are not the same thing ...

off-axis response is a consideration for when you slightly move around but are in the near-field or non-reverberant environment like outdoors ...

power response is a consideration for when you are in the far field in a reverberant environment, such as for example a gymnasium where you're hearing 90% reflected sound and at best 10% direct sound ...

you would ideally want to optimize all three - on axis, off-axis and power response - in practice it's so hard i don't think anybody even attempts this ... most people just choose to focus on optimizing what they know they actually can optimize and just pretend that other things do not matter or even exist ( but they do ! )