Geddes on Waveguides

To repeat, getting CD with the lowest amount of diffraction is the goal - OS does that, period.

IF lowest amount of diffraction is the goal, buying one of your speakers is the most reasonable decision.
BUT is there any data that objectively shows the benefits of low diffraction in sound reproduction or is it just obsessing over a single factor of accurate reproduction?

Best, Markus
 
The profile has a fairly sharp discontinuity to it so I would expect decent CD up to the point at which this sharp transition is 1/2 wavelength. At that point it will likely go "all to heii". CD is not difficult, its been done for years, that point seems to be missing here. Its Narrow CD without diffraction that is difficult. I would expect more diffraction from that device than one of mine that much is certain. The device looks like an axi-symmetric diffraction horn, with a larger radius at the diffraction point than is usually done. To repeat, getting CD with the lowest amount of diffraction is the goal - OS does that, period.

And yet didn't they refer to this as an OS waveguide?
 
IF lowest amount of diffraction is the goal, buying one of your speakers is the most reasonable decision.
BUT is there any data that objectively shows the benefits of low diffraction in sound reproduction or is it just obsessing over a single factor of accurate reproduction?

Best, Markus

Markus

There isn't, nor is there ever likely to be, a conclusive study of the nature that you suggest. Its far too difficult an experiment to do. I did do, and you know this, a study of "diffraction like" effects and showed their audibility, so I believe that there is objective support for the claim that they can be a factor.

The only other data that there is are the subjective perceptions of a large group of people who listened to my designs. Are these perception due, in part, to a reduction of diffraction? - I have no doubt of that. Is diffraction reduction the major aspect of the design that results in these perceptions? Hard to tell. Is it simply just good engineering that "makes the difference"? There is no doubt that this is also a factor and I am sure a fairly substantial one. But simply put, if I knew the answer to "what makes my speakers sound so good", I'm not sure that I would disclose all the details. I give away a lot of my design criteria, take it as reliable or not, its up to you. And is it comprehensive? You can pretty much bet that its not.
 
Last edited:
Hello Earl,

Where have you seen the profile of the min-phase horn/waveguide with this discontinuity?

Did it was published on DiyAudio?

Best regards from Paris, France

Jean-Michel Le Cléac'h

Wasn't it posted in 3578? Looking back I believe that you may be correct and that I misundertood. It was the profile in 3578 that I was refering to and this may not be the same one as in the other speakers being talked about.

I'll admit to not paying very close attention. Until some real data is shown, as promised, its just another speaker with no data and a few casual and uncontrolled listening tests.
 
I think the horn in 3578 is supposed to be an OS flair with the taper at the throat left on the horn so it can be cut off at exactly the correct size to mate up to the exit flair of a compression driver. So if that point is at the tangent of the conical section and OS section, then it would be an OS flair from the entry of the horn out to the mouth (assuming it's an OS profile, but it looks like it). I don't think it's meant to be a diffraction horn with a radius at the break in the flair.

I can't tell what the other horn being discussed is (in 3595, for example).
 
I think the horn in 3578 is supposed to be an OS flair with the taper at the throat left on the horn so it can be cut off at exactly the correct size to mate up to the exit flair of a compression driver.

Hi John

That may be the case, but its a very bad idea to do things that way if it is. A "one size fits all" is basically bad for everything.
 
A little explanation to profile 3578 - basically it's an OS waveguide for B&C DE250. It has OS contour down to 1", where it has 14.6 deg full wall angle and then it continues with conical section (14.6 deg). This way it is possible to attach the driver preciselly accounting the production tolerances, etc.
 

Attachments

  • de250.jpg
    de250.jpg
    46 KB · Views: 385
I'm doing a lot of assuming, but I don't think one size fits all was the intent. I think the manufacturing process doesn't all them to mold an exact size throat, so they are leaving that conical section on there and then trimming it to be an exact size, but there is only one size intended.

Thats a lot of extension, and a pretty big tollerance, if that was the intent. I do something like that on my waveguides, but I only end up needing maybe 1/16".
 
On a different topic, when Toole talks about a spatial average to calculate an early reflections response curve in his book (around p378 iirc), do you know if he means an unweighted average of curves from various angles? For some of the other curves he calculates, he specifically mentions weighting to account for the various areas that various angles would cover (ie, for measurements every 10 degrees, the 0 degree measurement would have an area covering 0 to 5 and be 0.2% of the area of a sphere, the 10 degree curve would cover 5 to 15 degrees and be 1.5% of the area, etc.)
 
p. 378-379: "The early refl ections curve is an estimate of all single-bounce, fi rst
refl ections in a typical listening room. Measurements were made of
early refl ection “rays” in 15 domestic listening rooms. From these data,
a formula was developed for combining selected data from the 70
measurements to develop an estimate of the fi rst refl ections arriving at
the listening location in an “average” room (Devantier, 2002). It is the
average of the following:
— Floor bounce: average of 20°, 30°, 40° down
— Ceiling bounce: average of 40°, 50°, 60° up
— Front wall bounce: average of 0°, ±10°, ±20°, ±30° horizontal
— Side wall bounces: average of ±40°, ±50°, ±60°, ±70°, ±80° horizontal
— Rear wall bounces: average of 180°, ±90° horizontal
The number of “averages” mentioned in that description may make it
seem as though anything useful would be lost in statistics. However,
this turns out to be a very useful metric. Being a substantial spatial
average, a bump that appears in this curve, and in other curves is clear
evidence of a resonance. It is also, as will be seen, the basis for a good
prediction of what is measured in rooms."
 
On a different topic, when Toole talks about a spatial average to calculate an early reflections response curve in his book (around p378 iirc), do you know if he means an unweighted average of curves from various angles? For some of the other curves he calculates, he specifically mentions weighting to account for the various areas that various angles would cover (ie, for measurements every 10 degrees, the 0 degree measurement would have an area covering 0 to 5 and be 0.2% of the area of a sphere, the 10 degree curve would cover 5 to 15 degrees and be 1.5% of the area, etc.)

John

I argued precisely this point with Floyd and Sean Olive. Basically there is weighting, but they have never published what it is and won't tell you what it is. According to Sean, "you should be able to guess and you'd probably be right", but that's simply not acceptable. If Harman wants the world to accept the use of their technique then they have to disclose it fully so that others can attempt to duplicate their work. But they aren't going to do that. This is why I refuse to promote their technique, saying only that some weighted form of polar measurements has been found to be very useful.

Markus post makes another strong point that I argue with. If Harman used "typical" loudspeakers when they measured the rooms - and we don't actually know - then they would get an entirely different set of first reflections than if they had used a loudspeaker with a narrower directivity, such as mine. Hence they "built" the measurements arround what they make, and of course, the measurements of their products always come out the best.

I completely support the "essence" of what Floyd and Sean have done, but I also disagree with some of the details. Their approach is a huge step forward, but it is not without its potential pitfalls. And not supplying full disclosure is certainly NOT the way to garner widespread acceptance.
 
A little explanation to profile 3578 - basically it's an OS waveguide for B&C DE250. It has OS contour down to 1", where it has 14.6 deg full wall angle and then it continues with conical section (14.6 deg). This way it is possible to attach the driver preciselly accounting the production tolerances, etc.
I think the true throat is in the driver, this is the most critical part of the design. Any change in flare after than acts like any other diffraction horn. The long 14deg flare howers the horn extension frequency which would be evident in any AxiDriver sim if modelled to the true physical diaphragm location. So it seems the horn loading may not be optimum, thus the image focus can be lost in the mid frequencies probably below 2KHz based on an educated guess. I would expect this to be evident when the data with proper CSD resolution is shown.

If it takes so long to just get measured data, there is certainly something wrong. I could get useable data down to at least 400Hz in any everage home around here just taking equipment that fits into a large notebook bag.
 
Last edited:
While I don't have any problem believing a discontinuity of flare can set off
standing waves in the primary mode of the horn, I'm not so sure you can
accurately call it a "diffraction" so far back in the narrow end of the throat.

How wide is a 1/4wave at 20KHz? If that part of the flare is less wide, can
a true diffraction or high order mode be considered to exist?
 
While I don't have any problem believing a discontinuity of flare can set off
standing waves in the primary mode of the horn, I'm not so sure you can
accurately call it a "diffraction" so far back in the narrow end of the throat.

How wide is a 1/4wave at 20KHz? If that part of the flare is less wide, can
a true diffraction or high order mode be considered to exist?

This is one of the very difficult questions to answer. To be exact, HOMs always exist whenever the slope of the contour changes, but these can be very small. At what point do they become important? - thats a very grey area. And is it "diffraction", that's semantic as well. All modes exist all the time, they just have vanishingly small values. At some point these values jump way up, and for higher modes they can jump up fast and to very high values. An HOM can be many times more efficient than the primary mode, but only over a very narrow range of values. It's the kind of thing that is all but impossible to talk about in posts like this because of the intracasies involved. A detailed analysis would take a book to describe, and so a post like this is going to be very hazy at best.

But more to your specific point - if "a discontinuity of flare can set off
standing waves in the primary mode of the horn" then HOMs can also be present to balance out the "boundary conditions" at the reflection point. For example when the wave is reflected from the mouth, the sum of the reflected wave, the propagated wave and the HOMs must all equal the incident wave. It is possible for the reflected wave to be exactly the same shape as the primary wave, but this is unlikely due to non-uniform impedance across the boundary, this means that there have to be HOMs to account for the difference. Is this "diffraction"? To me it is, but that's a matter of interpretation. If we could write the mathematics we would of course agree on all the terms as equations and what they did, even if we called them different things.

" I'm not so sure you can accurately call it a "diffraction" so far back in the narrow end of the throat"

I guess that I don't see how location would have anything to do with it. There can be diffraction (by my way of thinking) right at the end of the driver where it connects with the waveguide.