Horn tweeters V.S. aluminum dome tweeters

Which tweeter provides the bigger sweet spot, the horn tweeter or an aluminum dome tweeter?

My local high-end theater equipment store just told me on the phone that horn tweeters have a very narrow listening area, while dome tweeters have a much wider sweet spot. In contrast, I found a few internet articles that say just the opposite . . .

I'm forced to have far from ideal room conditions for my theater set-up (stuck in a corner, but rotated about 30 degrees), so I was wondering which tweeter would provide me with the wider sweet spot, enabling me to accompany more guest seating positions?

Thanks for any help anyone can provide.

Bob

[salvasol]

Don't believe everything they say ... Do as you did, a research online to find out the differences

If you can, ask them if you can try them at home (What better way to decide which one are best for you ) ... if they can't loan you a pair, I'm sure they have a 30 day return/exchange policy ... you can buy a pair and excahnge/return if you don't like them

[avaserfi]

I believe I answered your question sufficiently here: http://forums.audioholics.com/forums...ad.php?t=40482

[Sonnie]

Yep... I think Andrew gets the bone on this one. Excellent answer, although I have no idea if he's telling the truth, but it sounds really good. Just picking at you Andrew.

[avaserfi]

Sonnie wrote:

Yep... I think Andrew gets the bone on this one. Excellent answer, although I have no idea if he's telling the truth, but it sounds really good. Just picking at you Andrew.

Why I'll teach you to pick at me!

DS-21 wrote:

A few things. First, all tweeters are "horn" tweeters, in the sense that all of them are loaded on a waveguide of some sort. (A flush baffle is just a 180deg waveguide.) Some of them are designed to improve power response (and thus imaging) by providing constant directivity over a given area, with a smooth rolloff of treble energy past that point. (Many coaxes, such as the KEF Uni-Q and Tannoy Dual Concentric, attempt constant directivity by using the cone profile as a CD waveguide.) Others (e.g. the Tractix flare used by Klipsch and others) does not attempt constant directivity.

While technically you are correct every tweeter is a "horn" this is not the common terminology most are familiar with and I was answering with respect to the OPs terminology.

DS-21 wrote:

A 180deg waveguide is basically hopeless, and leads to all kinds of problems in the midband power response, because it does not control the dispersion of the tweeter at the bottom of its passband at all.

I am not sure if I understand your thought process here. With proper integration of drivers (the midrange and the tweeter) this is not an issue at all. *Clarification* What I mean by this is: With the proper crossover settings along side a midrange and tweeter with similar on and off axis responses this issue can be overcome. I am fully aware of this issue due to my current testing in which I am doing exactly this. With a mock baffle set up I have encountered no such issues.

DS-21 wrote:

So generally, I think it's safe to say that the best imaging - especially the best imaging over a zone in a room, as opposed to a spot. - comes from constant directivity waveguides, with other waveguides being various levels of inferior to a good CD waveguide.

In terms of the loudspeaker itself it is off-axis response that creates imaging and the sense of a sweet spot. Simple physics dictates that for the largest sweet spot greater off-axis dispersion is required and for this to be pleasurable this response need to mimic the axial response.

Of course the room plays a hugely important role in this situation, but again the OP asked about which type of tweeter which is why my emphasis lay on the loudspeaker not the room.

As far as waveguide versus off-axis dispersion they are two different but interrelated things. A waveguide does exactly what the name implies where the horizontal and in this case less importantly vertical response are "guided." While this can limit a tweeter's off-axis dispersion characteristics it does not entirely control the linearity of these responses which is why this distinction is important. Waveguides can control actual dispersion, but the dispersion itself dictates imaging so I guess in a sense both the waveguide and the tweeters resulting dispersion characteristics will be indicative of imaging/sweet spot sensation. If one is looking for a way to quantify this data off-axis response will be the most important data to collect and analyze.

[Sonnie]

Clear this up for me... I don't want to misunderstand what I'm reading.

Does this make any speaker other than the Klipsch, KEF Uni-Q and Tannoy DC or similar type horn useless?

I guess all these speakers developers for about 95% of the speakers on the market are not so smart... and the hundreds of speaker reviewers giving good marks to all these speakers with flush baffle speakers just don't realize all the problems that exist in these speakers and really shouldn't be writing reviews?

I wonder how all these speakers companies with flush baffles stay in business with all the problems that exist in this design?

I wonder where my Martin Logan's fit into all this?

And that monkey is just too funny!

[avaserfi]

Sonnie wrote:

Does this make any speaker other than the Klipsch, KEF Uni-Q and Tannoy DC or similar type horn useless?

Not at all. As I am sure you know no loudspeaker is perfect. This is why subjectivity comes to play so greatly when one goes out speaker shopping and listening you are narrowing down choices you are actually removing imperfections that you are more sensitive too.

Also, in some rooms it is preferable to have poor off-axis response. While in a technical sense this is not ideal there are certain situations such as an extremely reverberant room where off axis response would likely need to be tamed for maximum sound quality.

Sonnie wrote:

I guess all these speakers developers for about 95% of the speakers on the market are not so smart... and the hundreds of speaker reviewers giving good marks to all these speakers with flush baffle speakers just don't realize all the problems that exist in these speakers and really shouldn't be writing reviews.

One extremely important thing to realize with speaker reviews is that the room they are auditioned in plays a paramount role in sound quality. So the reviewer is reviewing the specific loudspeakers interaction with the given room.

Sonnie wrote:

I wonder how all these speakers companies with flush baffles stay in business with all the problems that exist in this design?

I don't know where you got this from, but flush baffles are used very often without problem. Internal to the tweeter (and thus in an area we cannot see or sometimes it just isn't noticed) there can be a waveguide which allows for proper dispersion characteristics without need for baffle compensation allowing for a flush mount with no problems.

If DS-21's reference to a 180 degree waveguide is where this came from I believe he was referring to a situation where a non-waveguided tweeter is used in conjunction with a flat baffle which is exactly what I plan on doing and as I have previously said is an extremely good idea if proper methodology is taken .

Baffle compensation is only really needed in the case where a non-waveguided tweeter is used and the tweeter in question has extremely linear off-axis response. If the compensation is not taken off-axis response will be ruined due to wave diffraction issues.

Sonnie wrote:

I wonder where my Martin Logan's fit into all this.

I am not aware of any credible 3rd party measurements on the Ascent I's. If you know of any I could give you my thoughts on the objective performance of the loudspeaker. If you wish not to jack this thread feel free to PM me or start another thread.

[Sonnie]

avaserfi wrote:

I don't know where you got this from, but flush baffles are used very often without problem. Internal to the tweeter (and thus in an area we cannot see or sometimes it just isn't noticed) there can be a waveguide which allows for proper dispersion characteristics without need for baffle compensation allowing for a flush mount with no problems.

If DS-21's reference to a 180 degree waveguide is where this came from I believe he was referring to a situation where a non-waveguided tweeter is used in conjunction with a flat baffle which is exactly what I plan on doing and as I have previously said is an extremely good idea if proper methodology is taken .

Baffle compensation is only really needed in the case where a non-waveguided tweeter is used and the tweeter in question has extremely linear off-axis response. If the compensation is not taken off-axis response will be ruined due to wave diffraction issues.

Yes... I was referring to his comments... (A flush baffle is just a 180deg waveguide.) and A 180deg waveguide is basically hopeless, and leads to all kinds of problems in the midband power response, because it does not control the dispersion of the tweeter at the bottom of its passband at all.

avaserfi wrote:

I am not aware of any credible 3rd party measurements on the Ascent I's. If you know of any I could give you my thoughts on the objective performance of the loudspeaker. If you wish not to jack this thread feel free to PM me or start another thread.

I am not aware of any either. I really haven't even studied the ML philosophy for that matter. I just can't get over how much I like the sound... even if someone discovers they are design flawed. In that case I would say I am glad they erred.

[avaserfi]

Sonnie wrote:

Yes... I was referring to his comments... (A flush baffle is just a 180deg waveguide.) and A 180deg waveguide is basically hopeless, and leads to all kinds of problems in the midband power response, because it does not control the dispersion of the tweeter at the bottom of its passband at all.

Yeah, I am pretty sure he was referring to a non-waveguided tweeter with a flat baffle. But I have already discussed why I disagree with this idea and why I even plan on using exactly what he said as hopeless in my current speaker design a couple posts back. Proper integration is key. Ambient field effects will change, but these can all be dealt with using proper planning and design.

Sonnie wrote:

I am not aware of any either. I really haven't even studied the ML philosophy for that matter. I just can't get over how much I like the sound... even if someone discovers they are design flawed. In that case I would say I am glad they erred.

Then enjoy and don't worry about it .

[Sonnie]

DS-21 wrote:

Note that when Toole started employing his blind testing methodology at Harman, good designs (waveguide-loaded tweeters, multiway centers) proliferated and bad designs (flush-mounted tweeters, toppled MTM centers) disappeared from all but the cheapest models, with the end result being that Harman has what I consider the deepest and broadest bench of good speakers right now bar none. Why? One obvious conclusion is that good designs implemented well sound better than bad designs implemented well.

So the upper end of JBL, Infinity, Revel are all some of the better designs? Actually all of Revels centers appear to be the toppled MTM design. Seems to be the same with JBL and Infinity. I'm assuming Harman owns more speakers companies I'm not aware of and maybe I'm not looking at the right models. Can you be more specific? Thanks!

[Sonnie]

I plan to get a pair of bookshelf speakers with a center for our great room, so this thread may help me decide what to get. I had looked at the Klipsch and the JBLs... particularly the RB-51 and L830.




Bob... if this discussion is not helping you and you feel I'm hi-jacking your thread, let me know and we'll move it. I'm hoping that it's helpful to you though... and others as well.

[avaserfi]

DS-21 wrote:

I've never heard a system with a flush-mounted tweeter that had a consistent sound on- and off- axis. All of them seem to get in trouble at the crossover region. Sure, there are some band-aids (e.g. the BBC dip) but to my ears they don't work as well as just doing it right to begin with.

This statement actually surprises me coming from you due to the extreme objectivity you have shown in other threads. Nonetheless subjectivity is important so...

This lack of consistent sound is likely due to the poor integration. Using a digital crossover along with matching axial and off-axis response you can achieve this consistent sound.

Using something like the Infinity 10PR80BZQ-FW02 4" mid range coupled with a tweeter such as the Neo PDR3 this response can be achieved. The Infinities off axis response at 60 degrees is nearly identical to the axial response up to 3000kHz. The PDR3 has an extremely linear response from around 2kHz to 20kHz with its off-axis response 60 degrees at 15kHz being only -5dB if flush mounted with proper baffle compensation. A combination such as this will allow for complete continuity with proper crossover use as the axial responses are nearly identical.

The real issue with such mounting is the typical DIYer and speaker manufacturer do not invest the time in testing/designing drivers until they are ideal for a given application which is why inconsistency occurs. I will say it can be much harder to achieve proper response using a flush mounted non-waveguide tweeter, but once implemented properly the results are far superior in both an objective and subjective* sense as off-axis dispersion is allowed to flourish in a way not conducive to the waveguided design.

*Both Ian Paisely of Mirage and Floyd Toole of Harman through various double blind tests of thousands of participants found wide off-axis dispersion with similar magnitude to on-axis is preferable.

One thing that I have noticed about these typical horn loaded designs is that their off-axis tweeter response is lacking in said linearity which has been shown to be preferable by the credible perceptual research. This is why I recommend them for highly reverberant rooms as the imaging would be better in such a case or for a near field application, but when compared with a loudspeaker that has a wider, more linear, dispersion pattern (all else being equal) in a acoustically treated environment subjective reviews will fall with the latter discussed loudspeaker.

As far as the discussion that there are only a handful brands putting out speakers worth buying that is simply ludicrous.For example B&W and Ascend Acoustics both quite a few speakers that perform well in the objective sense as laid out by perceptual research as well as the subjective sense. Furthermore, another Harman subsidiary that was never mentioned, Infinity has developed some speakers that are simply superb especially considering their price point. One such example is the Primus 360 (do not confuse this with the Primus 362). The perceptual research conducted does show issues with some designs, but more importantly, if fully understood, gives on an objective method of rating loudspeakers via measurements in the end design methodology should be ignored as it is not necessarily indicative of quality.

For those interested this link has all of Harman's child companies.

Sonnie wrote:

I plan to get a pair of bookshelf speakers with a center for our great room, so this thread may help me decide what to get. I had looked at the Klipsch and the JBLs... particularly the RB-51 and L830.




Bob... if this discussion is not helping you and you feel I'm hi-jacking your thread, let me know and we'll move it. I'm hoping that it's helpful to you though... and others as well.

Sonnie, if you tell me your budget for the two speakers I will make a suggestion as the best speaker in the price range tonight. There are far more issues in speaker design than just tweeter directivity that have been shown to increase or decrease listener ratings.

[Sonnie]

I'll start a new thread for those bookshelves and toss my budget in there.

I did mention Infinity as a Harman company though. I installed some 360's in a home theater system for my wife's boss and we were all impressed.

[no. 5]

avaserfi wrote:

One such example is the Primus 360 (do not confuse this with the Primus 362).

What changes were made between the 360 and 362?

[WmAx]

no. 5 wrote:

What changes were made between the 360 and 362?

A reliable source told me that the the new **2 series is an attempt to reduce driver production costs, at the expense of some quality. Based on the measurements at stereophile.com of a 162 vs. a 160 I had on hand, this certainly appears to be the case.

However, I personally do not think the slight quality reduction matters greatly based on the use of these speakers in stock form. The cabinets are resonant structurally, and do not even use sufficient acoustic absorption materials internally. However, some of those Primus **0 raw drivers measure fantastically - essentially good enough to use in unlimited class sound quality speaker systems. The mid-range used on the 360 is a case in point - I bought a replacement driver from Harman and analyzed this unit. The characteristic response properties are unbelievable considering the speaker system in which it was used. But as you know, you can put the best drivers in the world in a shoddy cabinet system with a less than ideal crossover and ruin any potential said drivers may have.

-Chris

[WmAx]

DS-21 wrote:

The Neo3 is basically at 180deg at 3kHz, and there's no way a 4" midrange can match that.

For one, the mid Andrew mentions as 4", actually has a 3" effective radiating area. Second, as soon as you put a Neo3PDR on a flat baffle, it can have no more effective than a +/- 90 degree radiation pattern, due to the physical size of the baffle in relation to the wavelengths. It is trivial to match these two drivers to have near identical response out to a * +/- 60-70 degree horizontal axis when integrated around 2.5khz. Which is frankly, perfect in practical terms, for perceptual purposes. This accounts for the window of the 1st reflection angle vector typical in most situations, so a near mirror reflection occurs as opposed to one with substantial deviations relative to the on axis response.

That's assuming one can stand that awful little tweeter. Forget the smooth FR, because it has lots of distortion and highly constricted dynamics. I've tried three variants of it, and I've learned my lesson every time.

What distortion would that be? I have measured this tweeter, and it has a notably low measured distortion. Certainly nothing near suspected audibility. Distortion measurements on that Zaph site also evidence it to have very low distortion.

-Chris

* In this thread, I intend to mean non-wave guide tweeters installed on baffles with very large radius edges, in the 3" or larger range, in order to maintain linearity on and off axis, thus minimizing diffracting characteristics of the baffle.

[avaserfi]

DS-21 wrote:

No, the real problem is one of physics: a big piston will in general have more limited directivity at high frequencies than a small piston, unless the directivity of the small piston at the bottom of its passband is controlled.

Very true. Hence the need for proper driver integration. If the drivers are not properly integrated they won't sound good together. I have repeated this throughout the thread and stand by it. The area of driver integration is where it seems not enough research is done. Various measurements must be taken of each driver individually as well as in conjunction to see how they interact as the loudspeaker is a system and the sum of its parts.

DS-21 wrote:

That is exactly what I'm talking about. The subjective aspect is where it's more important. For my tastes, I'm willing to put up with some top-octave irregularities to get the mid-tweeter transition region really right. Others may have opposite preferences.

I think you have missed my point - If proper time and care is taken neither of these situations need be an issue. Through the previously stated process of proper driver choice, baffle design and crossover implementation one can address both issues in a way to remove them from the system.

WmAx wrote:

For one, the mid Andrew mentions as 4", actually has a 3" effective radiating area. Second, as soon as you put a Neo3PDR on a flat baffle, it can have no more effective than a +/- 90 degree radiation pattern, due to the physical size of the baffle in relation to the wavelengths. It is trivial to match these two drivers to have near identical response out to a * +/- 60-70 degree horizontal axis when integrated around 2.5khz. Which is frankly, perfect in practical terms, for perceptual purposes. This accounts for the window of the 1st reflection angle vector typical in most situations, so a near mirror reflection occurs as opposed to one with substantial deviations relative to the on axis response.



What distortion would that be? I have measured this tweeter, and it has a notably low measured distortion. Certainly nothing near suspected audibility. Distortion measurements on that Zaph site also evidence it to have very low distortion.

-Chris

* In this thread, I intend to mean non-wave guide tweeters installed on baffles with very large radius edges, in the 3" or larger range, in order to maintain linearity on and off axis, thus minimizing diffracting characteristics of the baffle.

Couldn't have said it better myself.

[MatrixDweller]

It really all boils down to what sound you like the best.

[avaserfi]

MatrixDweller wrote:

It really all boils down to what sound you like the best.

While that is a noble thought there is a tremendous amount of research in the field that shows there are certain characteristics of tweeters that are desirable while there are others that are not. Numerous blind tests were conducted with thousands of subjects in search of this data.

Look up Ian Paisley on the AES website for example of such research.

[WmAx]

DS-21 wrote:

My point is, proper driver integration, absent some sort of directivity control of the bottom of the tweeter's passband, is impossible with most standard driver configurations, e.g. the 7" 2-way or MTM.

Whom is suggesting such an illogical combo as a 7" 'midrange' coupled to a standard direct radiation tweeter? Not Andrew. Not Chris(me).

And frankly, the more separate drivers in a system, the more cost required in crossover (be it digital/active or passive) and the smaller the listening window will often become. That suggests a 2-way configuration with a reasonably-sized midbass (10" or bigger, say, or twin 7's or 8's) and a coaxial or waveguide-loaded tweeter as the optimum configuration for home use, IMO.

Such an execution would ultimately mean poor off axis dispersion(undesirable from the perspective of general rules outlined by loudspekaer perceptual research) if you use for example, an 8" for most midrange duty, in a band over about 1000Hz. Generally, such a large mid-range would also have have poor cone behaviour so far as internal break up modes, if used over 1000Hz, resulting in a poor waterfall response. The unit may still have a relatively flat on-axis response, but have various other technical issues when used in such a high band.

Multiple drivers(in a 3 way system as an example) work sufficiently when properly integrated. I do not see the issue so far as relevant performance is concerned in a properly integrated system. A 2 way system is very difficult to pull off while having desirable measured characteristics are dictated as desirable by perceptual research.

-Chris

[MatrixDweller]

I personally like my Klipsch Reference speakers avaserfi. Klipsch has always advocated horn loading. I do like the sound of B&W and Paradigm speakers though and they're not (or at least most of them aren't) horn loaded.

Am I wrong in thinking that a basic dome tweeter is pretty much the same as a basic horn loaded design minus the horn? You can horn load any driver for that matter (ie: the Klipschorn or La Scala). Is the main goal of a horn to increase the sensitivity of the speaker without distorting the output too much?

As for sweet spot, would that not depend on the construction of either horn or tweeter. If you look at the aluminum dome tweeters on some Paradigm speakers (ie: the Signature series) they have a unique wave guide. I'm sure this sort of design would affect the direction of the frequencies emitted by the tweeter to enlarge the sweet spot and control any erratically dispersed waves.

Correct me if I'm way off in my assumptions.

[avaserfi]

MatrixDweller wrote:

Am I wrong in thinking that a basic dome tweeter is pretty much the same as a basic horn loaded design minus the horn? You can horn load any driver for that matter (ie: the Klipschorn or La Scala). Is the main goal of a horn to increase the sensitivity of the speaker without distorting the output too much?

Theoretically any driver could be horn loaded, but ideally you would want to use one that was designed for the specific application just as with any driver. Also, there are certain trade offs for this increased sensitivity, i.e., poor off-axis frequency response a common trait in most horn loaded designs.

Personally, I would never make the efficiency trade off in regards to high fidelity mid/far field listening. The cost for increased sensitivity just isn't worth the worsened off-axis response unless one needs a PA system in which fidelity isn't the primary concern. After all, proper amplification coupled with a well designed driver will allow for required SPLs in with inaudible compression and distortion assuming typical listening habits and even if the music that has extreme transient peaks of 110-115dB.

MatrixDweller wrote:

As for sweet spot, would that not depend on the construction of either horn or tweeter. If you look at the aluminum dome tweeters on some Paradigm speakers (ie: the Signature series) they have a unique wave guide. I'm sure this sort of design would affect the direction of the frequencies emitted by the tweeter to enlarge the sweet spot and control any erratically dispersed waves.

I refer you to this post as it should cover the question posed fully. Focus on the last section.

[WmAx]

DS-21 wrote:

Nonsense. Cone behavior depends fundamentally on cone design. There's no reason one shouldn't expect an 8" woofer to be just fine with a ~1.6kHz Fc, assuming it's coupled to a tweeter that matches its directivity at that frequency. Without a CD tweeter, of course, such a system would be hopeless.

Using an 8" unit up to that point(1.6khz example) will result in poor off axis response. It is NOT an option to use a waveguide that limits the tweeter to such an extent. It is non-ideal compared to a system that can keep a more even dispersion characteristic off axis, according to the perceptual research by Paisley, and by Toole.

All of them have some drawbacks. Three-ways more often than not sound incoherent, t

If you want to classify 'most' speakers, then 'most' speakers, regardless of design, are rubbish IMO, for ideal sound reproduction quality. I have extremely high standards and requirements in many different measured areas, and most speakers regardless of price, do not come close to meeting these standards I have set(which are based strictly on the standing perceptual research). It is not easy by any stretch of the imagination to design a system that meets all of these standards(one of which includes ideal mirrored off axis response up to a minimum of +/- 60 degrees, across the entire audible bandwidth relevant for music reproduction). I can make compromises for special purpose speakers. For example, a music production studio generally has early reflections that are either highly dampened or simply not reflecting back to the listening position, and generally in mid or near-field placement. Off axis response is not important in such an environment - there for I will waive my requirements for the off axis response standards in this case. For example: B&W 802D makes a superb monopole studio monitor in these conditions. It is very linear, has superbly low resonance drivers and a nearly inert cabinet system for minimum timbre distortion. But that speaker does not have an off axis response that meets my standards for a room that (ideally) uses the horizontal reflection points to increase perceived sound quality. But please note that my preference is for omnipolar, or near-omnipolar dispersion. As far as monopoles, the 802D is a superb speaker for all intents and purposes.

-Chris

[Sonnie]

DS-21 wrote:

You seem to keep mentioning Toole. Note what his speakers, which I consider some of the best currently available at their given price points, are doing to control tweeter directivity at the bottom of their respective passbands!

What speakers does he own?

[WmAx]

DS-21 wrote:

Philosophical difference, I guess. I see no need for a 120deg pattern. Quite the contrary, in fact. I side more with Geddes and Danley in having found that in the typical domestic living room a much narrower coverage area is desirable, because you get direct radiation at the listening position and less reflected sound. As I see it, "reflections" should come from the side and/or rear channels, not be spewed haphazardly about the room by the mains. The smaller the room, the narrower I would try to go, down to an arbitrary minimum of maybe 45deg or so.

But as Toole and other credible researchers have found, when the 1st horizontal reflections are very similar to the on axis sound, this reflected sound enhances perceived sound quality, when it is within a suitable time delay window. Read Toole's 'Loudspeakers and Rooms for Sound Reproduction: A Scientific Review', published in the JAES in 2006. It is a summary work going over the various modern research and conclusions based on such.

The other approach with significant theoretical and listening test support, which I take is your approach even though you've not talked much about the second prong of it, is to go wider with lots more room treatment.

It is true that to optimize the potential for a very wide response, certain additional treatments are needed. For example, cross channel talk should be minimized, also specified in such research. Due to the wide radiation pattern of an omnipolar, ideally, a middle room front dividing absorber would be used. Not by any chance desirable for cosmetics. I use such a device and it works as expected. Other than this, and lack of treatments at the 1st reflection points, the acoustical treatment is similar to any other application.

Alas, most room treatments that work are ugly, so attacking the problem from the speakers rather than the room makes more intuitive sense to me.

I am a function over form person. I will deal with the specific 'ugly' treatment method in order to achieve higher potential sound quality.

That said, look at the [url=http://www.danleysoundlabs.com/SYNERGY_horn.asp?MODEL=SH%20100]Danley Sound Labs SH-100 as an example of a 8" coax with very good polar measurements and a 100deg pattern.

I seriously doubt that unit would measure up to my requirements, in regards to several parameters.

True, but it becomes a lot easier once people get over their silly fear of waveguides. You seem to keep mentioning Toole. Note what his speakers, which I consider some of the best currently available at their given price points, are doing to control tweeter directivity at the bottom of their respective passbands!

Well, Toole does not actually design speakers. His research is used towards improving speakers for his former employer, Harman. There is also more than one reason to use shallow waveguide tweeters. One major benefit, especially for a speaker manufacturer, is to make the tweeter immune to the baffle edge diffraction. While normally, a baffle would require large radiuses(in the 2" range minimum for ideal behaviour) to prevent baffle diffraction transfer function errors, which can be costly to use in production, a shallow waveguide makes the cabinet edge irrelevant for the tweeter and is a low cost solution. My computer speaker monitor system uses tweeters with shallow waveguides for this very reason.

-Chris