This for the most part comes down to subjectivity.....a heavy burden every designer and developer in this industry must embrace at some point or another. We can take a well designed 12" woofer and play it to 1khz and get great dynamics and an equally good directivity match to the HF device. We can also take a well designed 5" midwoofer like a Revelator or Illuminator and bring it to 1khz, we only loose dynamics and some minor mismatch of directivity. We can get back some dynamics if we use a three way.....and if we cross it high enough we can eliminate the dreaded stand mounted 250khz suckout. I'm 100% confident the midrange resolution and detail surpasses that of the best 12" when the crossover is well designed. When we DIYers and commercial developers make these decisions on what they prefer sonically, de facto the design is based on subjective predications.
No doubt your design philosophy solves a lot of common room related problems and the Summas are a fantastic system to listen to. It's partially responsible for me getting into recreational speaker DIY ( I designed and build one off pro audio solutions for installed systems). But I just can't share the view that generally a large format two way with controlled directivity is the be all end all. This thread is intended to explore some other options. Thanks immensely for your view points and sharing your vast experiences.
So would I be correct in speculating that IF a three way system was designed around the preferred characteristics you build upon it should contain...?
A Large format dynamic sensitive driver of 15-18" placed close to the floor in a sealed enclosure and low passed at 400hz
A 6-6.5" dedicated high power handling midrange from 400hz to 1khz in waveguide of suitable size to provide a close directivity match to the HF device
A HF device of efficient and dynamic nature with response and near constant directivity from 1khz through 16khz employing a waveguide exhibiting OS design principles.
The 'middle' is where I would imagine you object.......I'd like to know why possibly?
No doubt your design philosophy solves a lot of common room related problems and the Summas are a fantastic system to listen to. It's partially responsible for me getting into recreational speaker DIY ( I designed and build one off pro audio solutions for installed systems). But I just can't share the view that generally a large format two way with controlled directivity is the be all end all. This thread is intended to explore some other options. Thanks immensely for your view points and sharing your vast experiences.
So would I be correct in speculating that IF a three way system was designed around the preferred characteristics you build upon it should contain...?
A Large format dynamic sensitive driver of 15-18" placed close to the floor in a sealed enclosure and low passed at 400hz
A 6-6.5" dedicated high power handling midrange from 400hz to 1khz in waveguide of suitable size to provide a close directivity match to the HF device
A HF device of efficient and dynamic nature with response and near constant directivity from 1khz through 16khz employing a waveguide exhibiting OS design principles.
The 'middle' is where I would imagine you object.......I'd like to know why possibly?
Last edited:
This is one thing that I had thought too. Use a smaller format mid driver with baffle moulding, to control its directivity, and to give it an off axis profile that would mate well with a much larger wave guide.
You'd then get the benefit of using a smaller driver for the mids, you'd not have to contend with any break up whatsoever as a result. Of course it would need to be a three way.
On the flip side though, for a good directivity match you need to cross 6.5" drivers at around 2.2-2.5khz. Many well designed soft cone 6.5" drivers can do this without a problem. I would imagine that you could find similar 12" drivers that would work exceptionally well up to around 1kHz. The AE TD12M springs to mind, or the TD12X if you wanted to sacrifice some sensitivity for more bass/linear travel.
Obviously, if you can make the 12" + wave guide work exceptionally well, it does beg the question, why would you want to go three way when you've got a 12" doing the mids and bass? Sure if you don't like it, use a TD12M along with a pair of floor mounted TD12Xs and cross them at 200-300Hz or so to eliminate floor bounce. But for the majority of users a high sensitivity 12" going down to 50Hz is going to deafen you before you can probably see it move more than a couple of mm.
I'd agree since there is no objective data to support your contention.
Being that everything is subjective, as you say, and we can't trust nor agree on "subjectivity', show me a speaker that measures better than a Summa and I will accept your contention. But otherwise it's just another unsubstantiated claim.
I don't do hypothetical designs, so I don't have an answer to your first question.
The answer to your second question is clearly a topic that I have covered ad-infinitum. Why I do what I do is no secret. What I don't understand is why others do what they do - short of the very real and very important aspect of budget. The way I do things is not cheap. So if my designs are out of your budget then the field is wide open for solutions, but I don't play in that field. I certainly appreciate many well designed speakers at the budget level - Behringer comes to mind (although the designer has left so its only the older stuff that I can attest to.)
I recently hosted the local audio club to bring their speakers over to be measured and I will post the data on my site in the PolarMap database. Some speakers were pretty good, older KEF's for example, but some were borderline terrible - broke I would call them, and maybe they were, I can't judge that from the data. Stay tuned to my site and you can see a list of some 20 different designs in all their glory and ...
Did any of these speakers meet my design objectives? No, nothing really.
I'd have to agree.
I just finished measuring a B&C 12NDX100 and there is no sign of any "breakup" well past 1 kHz. It is the finest woofer that I have ever seen. As I said, going this high with a 12" or 15" takes a real effort of speaker design, but they do exist. Of course at nearly $400 retail this is not a budget solution.
I'm hoping the mfgrs measurements are jacked up then as they clearly display a 6-7db narrow peak who's resonance is clearly visible in the impedance sweep.
I see this as a similar question to this thread but I hope this is not too off topic.
But how would one build a speaker with good polars for a small living space that may not have the room for some Summas or SEOS12 based speakers?
I was trying to figure that out earlier today and I thought that a dipole might work but then you have to have the speakers for into the room....then I thought that you could build a box that is open around the drivers to try and have some sort of cardoid like response with cancelation trying to control the polars...
Then I also thought that you could try and use a speakers own beaming to try and control it by using a larger tweeter and matching the directivity with maybe a bit larger of midbass and allow the tweeter to continue to beam as you go up in FR....
purely speculation here and I am not highly experienced but from what I have read and kinda understand it would be a possible option....maybe?
But how would one build a speaker with good polars for a small living space that may not have the room for some Summas or SEOS12 based speakers?
I was trying to figure that out earlier today and I thought that a dipole might work but then you have to have the speakers for into the room....then I thought that you could build a box that is open around the drivers to try and have some sort of cardoid like response with cancelation trying to control the polars...
Then I also thought that you could try and use a speakers own beaming to try and control it by using a larger tweeter and matching the directivity with maybe a bit larger of midbass and allow the tweeter to continue to beam as you go up in FR....
purely speculation here and I am not highly experienced but from what I have read and kinda understand it would be a possible option....maybe?
Smooth, controlled directivity is always beneficial. Sudden changes in DI are bad, but gradually increasing towards highs is generally best, as per Geddes.
We have some lines to go (none of whic are "easy" to make wel!
Minimal directivity
- 3-4 -ways with smart selection of drivers and xo points/slopes (Reve, Avalon)
- omnipoles (Devialet)
Medium directivity
- 2-3-way with large waveguide for tweeter, low xo-point 1,5-2kHz (many, also coaxials)
High directivity
- Gedlee Summa/Econowave type 2-way with horn midtweeter
- dipoles (NaO Note, LX521)
- synergy horn (Danley)
- 2-3-way horns
We have some lines to go (none of whic are "easy" to make wel!
Minimal directivity
- 3-4 -ways with smart selection of drivers and xo points/slopes (Reve, Avalon)
- omnipoles (Devialet)
Medium directivity
- 2-3-way with large waveguide for tweeter, low xo-point 1,5-2kHz (many, also coaxials)
High directivity
- Gedlee Summa/Econowave type 2-way with horn midtweeter
- dipoles (NaO Note, LX521)
- synergy horn (Danley)
- 2-3-way horns
That is the prevailing view because the difference is enormous, but you don't hold it because you can't measure deffinition in a loudspeaker.
All you need to do is listen. If 15" woofer and 6" midrange sounds the same to you from 400 to 1KHz, then i guess it's ok - for you. When i did it the difference was heaven and earth for me.
I would post some measurements but i don't know of any that can show me the amount of deffinion lost in a 15" woofer in regard to 6" at that particular frequency range.
And there is no subjectivity envolved - you could play it to anyone and the advantage of smaller midrange would be evident.
Don't get me wrong, i love the way OSWG sounds and i love the sound of 15" woofer but when i played one next to the other, 15" two way sounded like it was broken in regard to three way.
Last edited:
I suppose you mean 12NBX100 ?
Interesting driver!
That breakup just above 1khz needs to be dealt with in the crossover though, so I guess you plan to use it with an active crossover.
It also implies a 3rd distortion bump in the 300-400Hz range, and probably IMD products at various places.
Why not go with the cheaper 12NW100 as the only difference looks to be the absence of demodulation rings, and nonlinear distortion is not a factor in your design?
Don't quite understand that because my measurements did not show the peak that large. There was the typical rise in response above 700 Hz, but not such a large peak. At any rate this is still well controlled and I don't use the driver above about 800 Hz, so its not a big issue.
Of course there are certain benefits of having more drivers sharing the frequencies in smaller bands. Lower IM-distortion more even horizontal dispersion etc. But it all depends on the design. There are other ways of handling these issues with fewer drivers. One of them is using waveguides and a bigger bass/midrange unit crossed over lower that minimizes the distortions from excursion. Another is having an especially well behaved smaller single unit handling a larger part of the frequency range just using other drivers at the extremes >5kHz and below 300Hz that requires moving more air or even a smaller diameter closely positioned membrane to reach the top end.
The point is it all depends on what drivers you plan to use in the design and how to best utilize their benefits and minimize their shortcomings.
There is no real reason to be afraid of cross-overs per say in a particular range thinking they do bad. You have to work with them to an advantage in the design.
The point is it all depends on what drivers you plan to use in the design and how to best utilize their benefits and minimize their shortcomings.
There is no real reason to be afraid of cross-overs per say in a particular range thinking they do bad. You have to work with them to an advantage in the design.
Yes on both accounts.
When using a large driver up as high as I do you have to be very careful with the LP filter to match it to the drivers rising acoustic response. But this can be done actively or passively. Active, of coarse, is much easier.
I do not worry about any form of nonlinearity that is excursion related because using such a big driver excursion is never an issue. But nonlinear inductance is a different issue. It can be high order, unlike excursion nonlinearities which are always very low order, it is broadband, unlike excursion which is a LF issue and it is current dependent. I can limit the excursion but I cannot limit the current. So I DO worry about flux modulation and would never use a woofer as high as I do that did not have a demodulation ring.
I do disagree here as there is no "good" crossover, only lessor bad ones. Lobbing will always occur in non-coincident drivers and coincident ones have there own set of problems. The higher the frequency the bigger the problems. So if one can be avoided above 1 kHz it is best to do so.
Dr. Geddes,
Which crossover topology do you think is least riddled with problems?
1) First-order crossovers with minimal separation between drivers (if possible less than 1/4 wavelength apart);
2) Higher-order crossovers with time delay between drivers, either physical or electronic;
3) Substractive delay crossovers;
4) FIR filters.
I have yet to see a first order LP filter work for what I do.
Other than that you are talking about topologies and I don't look at the problem in those terms. I use whatever works, but this is almost always a high order LP because of cone breakup problems in woofers. Waveguides have there own acoustic HP function so these often end up being first order electrically, but with a CD waveguide the electrical HP filter frequency is at the upper edge of the passband because the response is falling in the passband. There just isn't a generic topology that works with a waveguide.
Digital filters can be FIR or IIR, either way that works.
Other than that you are talking about topologies and I don't look at the problem in those terms. I use whatever works, but this is almost always a high order LP because of cone breakup problems in woofers. Waveguides have there own acoustic HP function so these often end up being first order electrically, but with a CD waveguide the electrical HP filter frequency is at the upper edge of the passband because the response is falling in the passband. There just isn't a generic topology that works with a waveguide.
Digital filters can be FIR or IIR, either way that works.
Earl have you ever considered trying out some of the drivers from Acoustic Elegance? They would seem right up your street in terms of specifications, price and performance.
I don't see any high directivity here. Dipoles are quite close to omni in practice, average DI ~5 dB which is typical value for small 2-way.
You can roll your eyes as much as you want but it's a fact. If we can't measure something now, that doesn't mean that we will never be able to measure it. The fact that you think that 15" woofer and 6" dedicated midrange sounds the same from 400Hz to 1KHz speaks a lot.
I can't measure deffinition but it is subjective term that we all very well know what it means as well as imaging, coloration and spaciousness.
Are you trying to say that just because you can't measure coloration it stops to be of any importance to you ?
We can debate on why smaller format midrange drivers sounds better than 15" in 400-1KHz area but do they sound better is not a matter of oppinion. It's a fact.
And anyone can check if i'm right simply by trying it on their own. People are usually prone to believe in things before they check it out by themselves so i don't expect to see too many guys trying to make 2way and 3way with the same woofer/waveguide combo just for testing. But hey, i did 🙂
I'm making a 3way (still work in progress) with a small format midrange that can play lower midrange as it should. Now why would i do that ? Hearing problem ? Desire to complicate my life with 3way xover ? I don't think so.
Last edited:
- Status
- Not open for further replies.