Is it possible to cover the whole spectrum, high SPL, low distortion with a 2-way?

diyAudio Moderator
Joined 2008
Paid Member
whgeiger said:
At the point of inflection the radius of curvature of both curves should match.
You once suggested to me that LeCleach was a possible termination choice. I wonder whether your goals are focussed on diffraction minimisation, where another option would be to embrace and allow but control the inevitable diffraction in other ways?
 
diyAudio Moderator
Joined 2008
Paid Member
Interesting Mabat. However if the OS profile was convex, I'd expect that an axial ray rotated about the virtual apex of the waveguide would meet tangent to the profile at some middle point, rather than at its farther extreme. What do you think?
 
Last edited:
Oh yes! :)

The OS profile is convex. Second derivative would be zero at inflection point. It is positive everywhere, approaching zero in the limit. But it's quite obvious from a picture of a hyperbola. Joining with a clothoid (as shown), it remains convex everywhere. Convex means that any secant remains whole "above the contour".
 
Last edited:
Danley Pure Groove impression Part I

It's a little off topic, and completely subjective; the following short impression of 2 Danley Pure Groove setups:

Room 1:
2x J1, 16x TH118, 2x TH812, 2x SBH10 en 2x SHmini.

Room 2:
2x SH96HO, 2x TH812 which were switched off.

Amplifiers: MC2 Audio E90 & E45
DSP: XTA DP-448 & DP-444


In the big main room (1) both Jericho 1s are flown from the ceiling in stereo to cover the dancefloor, as well as the balconies on the second floor.
The SBH10s and SHminis are used for infill around the bars.
This install was taken care of by the Benelux distributor of Danley Sound Labs, MC2 Audio, XTA Electronics.

I've 'experienced' the J1s and the TH118s several times before.
Last saturday I decided to walk around and also listen upstairs, without earplugs (for short periods).
The party wasn't crowded compared to previous events, it made walking across the room much easier and also seemed to have a positive effect on overall soundquality.

First off, the J1 is one of the biggest Danley Synergy horns and according to Danley's website: "The loudest reference monitor ever created",
The J1-94 houses six 18 inch drivers, six 6 inch mid-range drivers and three 1.4 (BMS 4594) inch high frequency drivers, enough to produce >145 dB SPL.
1 Danley Jericho beats out the competition's 12 box line array.


It's quite special to experience these beasts during a live DJ performance, because in this club the J1s won't be used even close to the maximum output capability. The DJ played a wide variety of electronic music, including some tracks containing acoustic instruments and vocals. I heard a remix of Dire Straits' Brothers in Arms, as well as Kraftwerk's the Model.
While I prefer not to leave my ears unprotected for more than 5-10 minutes, the SPL in the room doesn't seem very high, due to the available headroom and subsequently low distortion levels.
Coverage is adequate for this (former) theatre. The sweetspot is large, obviously located around the center of the dancefloor.
Only if you're considerably off-axis you'll notice the loss of mid/highs.

Contrary to previous expectations, the Jericho's don't sound harsh at all. I guess the available headroom is a major contributing factor, as well as proper EQ-ing of the system.

The Jericho's sound like gigantic fullrange hifi loudspeakers. There's no typical horn signature to speak of. However, I don't think you could compare these beasts to large format studio monitors, because the SPL requirements are vastly different.
I am sure the Jericho's outperform many line arrays, because of the fullrange concept and point source nature. There's a surprisingly even distribution across the room and high levels of intelligibility/imaging/depth, which culiminates into a 'homogenic sonic picture'

Upstairs, at the front end of one of the balconies - only a few meters from the J1s - the SPL doesn't seem much higher compared to elsewhere in the room. In short, the Jericho's coverage is very consistent with distance. For those who haven't heard these yet: they do live up to the hype and expectations.

The lack of harshness is most surprising, considering those 3 BMS Coaxials spitting into the trapezium shaped, HOM generating combiner throat.
I have to admit, the combination of 3x BMS Coaxials, 6 compression loaded (sealed back) 6" drivers and 6x 18" doesn't make much sense. Apparently, it works and quite wonderfully so. I believe the J1 is among Tom Danley's personal favorites and I can see why.

The J1s may not be the most refined fullrange speakers on the market, their overall performance is very impressive. Used fullrange they are specified for 40-20k. The 18s would probably take some boosting of the low end in smaller venues.

The TH118s and TH812s are perfect companions to the Jerichos.
As mentioned by others, the tapped horns are louder than they appear.
They do a good job in pressurizing the room and completely lack the 'one note bass' syndrom. Instead there's a lot of tonality, definition and a sense of air in the bass/midbass.
 

Attachments

  • Pure Groove Kantine.png
    Pure Groove Kantine.png
    729.4 KB · Views: 279
  • 150826043-de-marktkantine-(1)-2 (2).jpg
    150826043-de-marktkantine-(1)-2 (2).jpg
    593.5 KB · Views: 271
  • De-Marktkantine-1 (1).jpg
    De-Marktkantine-1 (1).jpg
    185.7 KB · Views: 294
  • 60537827_147296873063482_2768695831570061022_n.jpg
    60537827_147296873063482_2768695831570061022_n.jpg
    173.7 KB · Views: 271
  • Fig1JerichoDanley.jpg
    Fig1JerichoDanley.jpg
    109 KB · Views: 267
Last edited:
diyAudio Moderator
Joined 2008
Paid Member
but once the mouth is big enough
I suppose, yes.
the only thing you care about is to ensure a smooth termination to mitigate further diffraction.
I think it should also be about how. If the wavefront doesn't "let go" of the walls then it can widen.

Since there is no natural termination for conical style horns, any attempt is already an independent mission.
 
diyAudio Moderator
Joined 2008
Paid Member
Not literally of course. I mean moving on with minimal disturbance. Maintaining its directivity moving out of the waveguide with minimal source creation. An example is the way that the LeCleach horn grants graceful independence to higher (gradually lower) frequencies, except that to do so it continues to increase its width.
 
To smoothly terminate a conical horn you need a curve with zero (or almost zero in a case of OSWG) curvature at the beginning. When I look at LeCleach profile I don't see this as suitable as the clothoid. I have no doubt it would still work well (it's only a very minor difference I guess), only that the resulting waveguide would have to be bigger in the end. Or not? Don't forget that I want to maintain the directivity constant as much as possible.
 
Last edited:
diyAudio Moderator
Joined 2008
Paid Member
you need a curve with zero
I'm not suggesting a simple radius termination, just that the wavefront not be allowed to "follow" the curve around.

There is going to be the same amount of diffraction on the whole. 'Minimising' could only mean 'spreading'. I think this is important.

Doing this, especially on a waveguide, may mean a smaller region within which to effect the diffraction. Then again, a radius in general does have a more diverse profile early on than an equivalent varying radius profile.

Is it better to have lower frequency axial resonances or room reflections?
 
Mitigating Reflectance

You once suggested to me that LeCleach was a possible termination choice. I wonder whether your goals are focussed on diffraction minimisation, where another option would be to embrace and allow but control the inevitable diffraction in other ways?

Beneficial diffraction occurs gradually in the throat of an OS horn.
Here we have a boundary condition of declining curvature which ideally transforms a plane outbound wave front into a oblate spheroidal surface that is approaching that of a spherical cap. While this spreads the wave energy uniformly within the horn body, this orderly propagation will become significantly disturbed if mouth reflectance is not mitigated. This requires reversing the curvature decline in the horn profile as the mouth is approached. The best way to do this is with a fairness transition curve.
Le Cleach's method will yield a satisfactory curve in this regard, but their others as well.

For the purposes of clarity I am not addressing the host of other design issues, including compression driver wave front matching and other propagation modes present in a finite horn. WHG
 
diyAudio Moderator
Joined 2008
Paid Member
this orderly propagation will become significantly disturbed if mouth reflectance is not mitigated.
I like to allow my lowest frequencies to pass thereabouts a mouth of a wavelength before attempting significant extra curvature (but after shedding the highs). I feel this is likely a significant mitigating factor in the need to do this.
 
Curvature

There is no point of inflection. If you mean at the join, yes, it should match (I guess). It will.

Thanks for the reading. I'm actually quite happy with how it's done now - this is the waveguide I will go for.

You are close enough. Design iteration might just be a little bit slower however.

1) I intended to recognize a point of curvature minimum where your "join" occurs. As the sign of curvature does not change, of course it is not a point of inflection.

2) A C-Bézier curve can follow the locus of a hyperbola and then depart from it to a fairness curve of increasing curvature proportional to its length. Euler's Spiral is just one member of the fairness curve family but requires a "join" to a hyperbolic throat segment. This tack to horn design is not new. [1]

3) Important to this approach (of terminating a hyperbolic throat in a fairness curve) is that it opens up an additional degree of freedom in throat design, otherwise fixed by specification of throat diameter (@ Slope = 0) and an asymptote (coverage) angle of a hyperbola.

WHG

[1] Post: 21st January 2013, 07:47 PM
Subject: A Contrary View
URL: Geddes on Waveguides

Excerpt: “This finding supports the conjecture that horns of infinite extent, that require mouth truncation to make them real, are a bad idea for loudspeaker applications. Alternatively, curves such as a segment from Euler’s spiral [1] might be a superior choice. Here boundary curvature gently increases with curve length to form a smooth transition to free space from the confines of throat aperture of a compression driver. One could argue that this approximation comes closer to meeting the mission of a loudspeaker horn where unlike that for a musical instrument, reflectance is to be avoided. We do not want to trap acoustical energy within the horn; we want all of it to leave!”
 
Last edited:
1) I intended to recognize a point of curvature minimum where your "join" occurs. As the sign of curvature does not change, of course it is not a point of inflection.
If you intended just that, then I don't understand why you called it an inflection point.

Curvatures of both segments will match at this junction, i.e. it should be smooth enough. There's no need to go for splines as I have it already working as it is. Is it possible to do it a bit differently? Of course it is. Will it be any noticeably better? Please show me that.

- It was year 2013. So where are all the waveguides terminated by Euler spiral segment? :) :wave:
 
Aren't you concerned that might help maintain the wavefront at the edges and allow it to widen, rather than dissipating it? (At the least a combination may serve)

What happens is a gradual reduction in the wavefronts amplitude at the edges. In my book I show how this has the net effect on the directivity of smoothing it, which is why it mitigates the midrange narrowing and the axial holes. They get "averaged out. The slower the mouth transition is the greater the averaging effect will be with the least amount of diffraction and reflections.

And let's all remember that whenever there is diffraction there must also be reflection. A new diffracted wave propagates out, but the counter to this wave must be reflected back down the device. So both are issues.

Widening of the wavefront is not the purpose of an OSWG. I guess minimizing/elimination the narrowing, the on-axis null and interferences at the mouth were areas of concern.
Dr. Geddes has opted for an asymetrical mouth, whereas mabat optimized the throat to mouth transition.
Please, correct me if I am wrong.

My commercial waveguides are all symmetric. I have built several non symmetric ones and they are usually not an improvement.