• Disclaimer: This Vendor's Forum is a paid-for commercial area. Unlike the rest of diyAudio, the Vendor has complete control of what may or may not be posted in this forum. If you wish to discuss technical matters outside the bounds of what is permitted by the Vendor, please use the non-commercial areas of diyAudio to do so.

DIY Waveguide loudspeaker kit

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Esperado,

Crikey! Earl would have to be the one least open to the charge of snake oil, or not providing measurements. I can't think of anyone further from being guilty of either. Even Genelec with their studio monitors only show off axis response at a couple of different angles. There is far more data contained in Earl's measurements than in those commonly seen.

I have compared various horn and waveguide profiles. In my opinion, the oblate spheroid is the best sounding of all of them so far and it also measures very well. I recall one that measured very well in the polar plots, but has a noticeable "horn honk." The oblate spheroid with the foam does not sound like a horn at all. It sounds like a good dome tweeter in terms of smoothness, but adds about 18 dB of output (at least) as well as a certain effortless quality and greater dynamics.
 
Earl would have to be the one least open to the charge of snake oil, or not providing measurements.
Please, don't misunderstand my postings. I don't said that it is Snake oil, but it was presented like, in general, snake oil is presented .
I do not knew who was Earl, and that was my first feeling, looking at his site.
Smoothed curves like designed by hand, no responses curves at 30°, and this method for presenting the polar is nice to have a good global idea in one sight, not to get a precise scale, on my point of view.
I know, for now what kind of work earl did those past years, but i still feel the style of his web site deserve-it. And i've read i'm not the only one in the audio business.
.
About "horn sound", it is the same kind of legend that the sound of cables. Just the bad designed horns or not well realized ones (resonant) presents that kind of color., reason why i build mine in plain wood. In fact, on my point of view, and when the size makes-it possible, compression drivers and horns are the best way to transform an electric signal in a faithful acoustic one. Thanks to the rigidity and little size of the diaphragm, and the gain of the horn. Efficiency is a very important part of a hifi enclosure. Only way to reduce dynamic distortions, to produce a realistic acoustic level etc....
.
Too, i do not agree with the choice of the motor in those loudspeakers. The motor makes a big difference. We tried TAD's one on the Aeria system with good results. But not so good than the JBL they (and i) use, despite the fact we where obliged to sort them, from some bad manufactured copies.
I, too, did not like Genelec at all, and avoided to mix with them. It was the old ones (coaxial, with those horrible and unbalanced trebles, i did not knows the last ones.

About smoothed curves, i know that, when working on an enclosure, it can help to gives a global idea of the energy. I know too it gives a glorious wrong image of linearity, so it looks like fake to me in a commercial data sheet. I'm very interested, looking in a response curve to can have an idea where the cone begin to fractionate, if a horn have huge resonances etc.
 
Last edited:
Please, don't misunderstand my postings. I don't said that it is Snake oil, but it was presented like, in general, snake oil is presented .
I do not knew who was Earl, and that was my first feeling, looking at his site.
Smoothed curves like designed by hand, no responses curves at 30°, and this method for presenting the polar is nice to have a good global idea in one sight, not to get a precise scale, on my point of view.

I understand but don't agree. Try running the application that is included and you can see the response at 30 degrees or any angle you choose. Who else provides that much data? Snake oil makers tend to provide no data at all and put in just enough pseudo science to convince those not really interested in it.

About "horn sound", it is the same kind of legend that the sound of cables. Just the bad designed horns or not well realized ones (resonant) presents that kind of color., reason why i build mine in plain wood.

Again, I can't agree at all. Take the same driver and set it up in different waveguides and horns, you will hear a difference. Some of them can be close enough that it's not so easy to pick, but even the good ones will measure and sound different. You can take two of them with the same construction and they will sound different for reasons other than the material itself. In my opinion, the design of the horn itself is the dominant aspect, but it's easy enough to deal with any issues related to the unit itself.

Regarding compression drivers sounding different I would tend to agree, but the most obvious ones are between units like the B&C that Earl uses and the Selenium titanium unit often used in the Ewave. Huge difference. If you compare the B&C to more expensive TAD units, I can't comment as I haven't tried it.
 
Again, I can't agree at all. Take the same driver and set it up in different waveguides and horns, you will hear a difference.
That's what i said: "Bad designed". My horn, by example, as other i knows, does not add accidents to the response curve of the driver, apart the low frequency roll off. And you can compare different diameters or depths, they all measure very close and sound the same. One thing is interesant is to compare the impedance curve of the motor in free air, and charged by the horn. If the horn is well designed, the impedance curve won't change. If your horn create resonances, you will see them in the impedance curve as well.
I would like to add that some other circular spherical waves horns with different designs give quite similar results, and no one present any "horn sound" at all. In fact, their response curves are very similar, some soo flat, like this one:
http://www.azurahorn.com/JA6681B on 425 Horn (2).pdf
I can believe that earl have a directivity controlled nice horn, but he is not alone in the world. Which can ensure to have the absolute truth, if even this truth exists ?

.
 
Esperado

You make some good points, the typical ones.

The horizontal response of the system is by far the more important than the vertical one because this is the same as the plane of our ears. Our hearing is much more accute to horizontal aberations than vertical ones. Hence, I will sacrifice some vertical problems to alleviate horizontal ones. My horizontal response is as good as it gets.

The polar response data that I post are 1/24 th octave results, they are not "smoothed" as you suggest.

As to sharper slopes, sharper slopes have greater phase change per Hz than shallow ones. This means that while the overlap bandwidth is narrower, the phase change within this bandwidth and hence the depth of the problems at crossover are greater. There is no magic bullet when it comes to crossovers, its what works. There is nothing standard about my crossovers because they are designed to get the best response no matter how steep or shallow they are.

As to Zobel's, a flat impedance curve is irrelavent when the amp is a voltage source ( and I hope the ones that get used are). When the amp has a "highish" output impedance then the Zobel can be a benifit, but there are other, and less costly, ways to achieve the same result. All in all the use of a Zobel is a lot like using "magic caps" or foil inductors. A lot of hype for little to no real effect.

Regarding the drivers, we built a set of Summas with TADs and another set with B&C - both crossovers were optimized for the driver sets. In a blind test of some 16 listeners the results showed that there was no clear preference for the TADs over the B&C (the B7C were slightly prefered) but there was a clear preference for the design over other designs. The obvious conclusion here is that the drivers do not matter nearly as much as the system design. When you do sighted tests of the TADs everyone is overwhelmed by how gorgious they are - so they MUST sound great! Fact is that they don't sound or measure any better than any other drivers.

Perhaps you should read my book if you want to know the theory behind my waveguides. There is no more complete analysis of the theory than you will find in that text. But I warn you, it is not easy reading in any language.
 
Last edited:
As to Zobel's, a flat impedance curve is irrelavent when the amp is a voltage source ( and I hope the ones that get used are). When the amp has a "highish" output impedance then the Zobel can be a benifit, but there are other, and less costly, ways to achieve the same result. All in all the use of a Zobel is a lot like using "magic caps" or foil inductors. A lot of hype for little to no real effect.
I am sorry, but Zobel and motional impedance compensation makes a noticeable difference you can measure and listen with evidence.

A pure voltage amp does not exist, that is the first reason.

The second one is that, delivering a constant tension/current, the amplifier will have less to ask to feedback's corrections, and phase swifts between A and V. Adding the fact that the response curve will be less altered with any serial wire Z.

The third one is it helps to calculate accurate filters.

One thing i like to insist on is the motional compensation. The change can be heard even with no amp, kicking the cone of the bass driver with the finger. The resonance is greatly reduced.
Here, too, and more than the Zobel, it helps to design accurate filters on the low pass side. Calculations applies with no tuning.
The result is visible a lot in the time domain, with some waterfall exploration. And audible with more controlled basses. Motional compensation has to be done for the loudspeaker in free air.. Accord tuning of the bass reflex has to be done with no compensation. When you add the two, it is amazing to look how it turns on the impedance curve. You just have to tune the serial resistance to get rid of too low z in the bottom.

The last thing i know fore sure id the way to position absorbent material in the enclosures:
instead of putting-it on the sides, to put in at all the centers of the volumes. Making a cross in all the dimensions. The difference is amazing, and you can see it again with waterfall, a reduced level in low medium on the response curve, no more bass reflex typical sound.

Just make the try, if not done yet.. On the bad side, I agree it add some cost in filter's components.

Btw: I don't believe in magic of any kind, like the sound of materials in the electric domain. I believe in their characteristics. I don't listen to cables's sound, never use audiophile caps or whatever, but appropriate less expensive caps as possible. IM distortions of some caps are measurable, like the ceramic ones, but i cannot understand the bad reputation of good electrolytic caps, as an example. I know the influence of too much R with coils, and the distortion they can add if their wires are not tightly glued. The only thing that is in the magic domain, on my side is, as far we are talking about passive filters, i like to design perfectly symetric ones.
At the end, my enclosures, at home, have a impedance curve less than +-1 Ohm around 6 Ohms on all the range. 5 only at the resonance of the Bass reflex. 6.5 at the cross over of the filter.
 
Last edited:
No one would agree with you that smoothing to 1/24th octave is a problem. I suppose you are one of those guys who believes that every bin of an FFT is "perfect". Well that is simply not the case. Each FFT bin has an error associated with it and only when a few bins are averaged together does one get an accurate result. The more bins the more accurate the average, but of course the less resolution in frequency. So one has to pick an "optimum". The almost universally accepted optimum is 1/20th to 1/10th octave.

No matter what other technique you use (other than FFT) there is SOME averaging. SO don't kid yourself that there is some "magic" measurement box or technique that is error and averaging free, because it does not exist.
 
I have had very good experience placing stuffing away from walls simply because that is here the high velocity is and thus more effective damping.
Impedance flattening proved to also audibly improve sound quality.
I quite agree with cap selection. There are many reasonably priced alternatives.
Interconnect impedance flattening is also important in my experience, which is also what has lead me to develop interconnects.
But let's face it, every engineer has his own thing about what is best combination for a design.
 
Last edited:
I am sorry, but Zobel and motional impedance compensation makes a noticeable difference you can measure and listen with evidence.

We will have to differ on this.

There is no published evidence to support your claims and they contradict my knowledge and experince. Amps with output impedances as low as .1 ohm are quite readily available and this level of impedance could only make less than a .1dB change in response. And as to the load affecting the amp, this to is nonsense in any good amplifier.

If you believe all this audio "snake oil" dogma then there is no wonder why you think that my approach is the falicious one.
 
No one would agree with you that smoothing to 1/24th octave is a problem. ...SO don't kid yourself that there is some "magic" measurement box or technique that is error and averaging free, because it does not exist.
If you are talking about tonal balance, i agree. But slow sweeping frequencies measurements in free air or big anechoic chamber give a better feeling aboutl the HQ cone"s material resonances due to fractioning. Helps to chose the best (far from perfect) hight cutoff frequency of a given loudspeaker, or to chose between several loudspeakers.. Nothing more.
About motors and horns, we all live in a poor real world. No one can knows what exactly happens in a specific motor and his phase piece. And it will differ from a model to an other. Of course it will be far from a spheroidal wave. In fact, far from any simple geometric figure at all the frequencies. Agree ?
Well, we all try to get some mathematical model in order to design horns. Some approaches gives quite good and close listening results. One can be better with a specific horn or room, worse with an other. No one, including you can pretend he own the perfect formula.
A specific attention to the first centimeters of the curve figure, witch is the one of the driver, slightly conical, and differ of any theoretical calculed curve. Did you manage this point and modify your drivers ?
About diffractions in horns and what you call HOM, i had too this kind of thinking, several years ago, and tried absorbent material too, thinking that it will damp the multi reflected lateral waves more than the direct ones, because they will have to across the damping material several times. But i had forgotten this idea as listening the result produced a feeling of some kind of "Dead" sound and medicine worse than the disease.

I would be very interested to listen to your solution, will try some dau to find someboly in Paris ,or better, Toulouse.

I wonder if some kind of long fibers oriented very light material ? Will try again with some very light fiberglass wool, fibers oriented in the depth direction of the horn. or, like for damping enclosures, some cross of thin absorbent material all along the center axe of the horn ?
Best regards.
 
If you believe all this audio "snake oil" dogma then there is no wonder why you think that my approach is the falicious one.
Where did i said your approach was "falicious"in any way ? Some paranoiac reaction ? Impossible to explore different technical approach things in a open minded way ?

Please, be gentle, i'm not that kind of audiophile you think, and i have a little experience in loudspeakers and enclosures, (i was head of the electro acoustic department of the first French hifi manufacturer during the seventeens). And sound engineer during more than 30 years, not very sensible to snake oils because my scientific formation. Jut to gives-you an idea about what i'm able to hear, compare, or not.
You knows 30years of contact with real musical instruments and musicians 10 hours a day, and more or less bad 'state of the art' professional equipments.
See what i mean ?
 
You need to understand at the very begining that I give no credence to subjectiuve listening tests that have not been done using rigorous scientific methods. Your personal experiences on any of this subject matter will make no points with me. I learned a long time ago that converations that have a "well it sounds good to me!" always become pointless. This is a claim that can neither be proven or disproven so the rest of the conversation can only be meaningless.

You did claim that my work sounded like "snake oil" - lots of people here noted that. So by falicious I meant this reference. I suspect that maybe you are seeing things a little differently now.

Compression drivers are, to me, a given, I can't make them so I can only use what is available. But you will find that I hold several patents on the proper way to make a phase plug for a compression driver, so this is not something for which I am unaware. There is still an optimum waveguide for any given driver. It just may not be the best possible design of the combination of the two. That drivers and horns are designed independently of one another is certainly a big problem. That needs to change and my patents show how that can be done.

There are none of my speakers in France - as yet - I have an order for some smaller baffles to France, but thats all. Shipping to the EU is prohibitive so there has not been much interest. One smart guy, years ago when they didn't charge for luggage, flew in from Germany and took back some speakers as baggage. It cost him less that way!!

I am very open minded, but with 45+ years in audio, I have heard it all. I am not going to argue about subjective things. My speakers are designed using science and science alone and I'll stand behind every inch of them from that standpoint, but I won't argue about "this speaker sounds better than that one" etc. Its all gets rather pointless.

You probably believe that to each person there is a "best" speaker and that this is different for everyone. Well, you see, I don't believe that, and I don't think that any of my customers believe that anymore either. There is a "best" speaker and it actually is the one that measures "best". Coincidence? I think not.
 
Last edited:
You did claim that my work sounded like "snake oil" -
I had said that the way you present-it on your web site... etc...
Not your work itself. it is not at all the same thing.

And a lot because it seems you consider all the other people, searching with the same passion than you, scientific methodologies too, and different approach as wrong or even stupid.
On my side, what can i say about the quality of your results or the validity of your theories without any hearing and measurements ? Did you provide somewhere formulas on witch i can experiment your horns designs ?

By example, you tell me that impedance compensation is "snake oil" ?
I believe then you have never even tried-it.

Ok: just one testing methodology.
Take you beloved amplifier.

Compensate the loudspeakers of one of your enclosure (i believe you know how to calculate those parallels circuits.)

Build a comparator with an Op amp gain 1 or more, feed the + with the input signal of your amplifier. The loudspeaker output have to be connected in // to your loudspeaker and a potentiometer followed by a serial resistance in the - of the operational amplifier.
Tune the potentiometer to reduce at the minimum the output signal of the Amp op in order to see only the difference between the input and the output of your amplifier.

Now, record on some multi-track a dynamic musical source, with drums and so...
Play the musical track at a normal listening level and, in the same time, record the out of the comparator on a new track.
Remove the impedances compensations, and make the same recording on an other track.

Compare the two tracks from the comparator. (and you can listen to).
Come back and argue against the utility of such a impedance compensation, now.
I will eat my hat.

And this test is just on the amplifier side.

On my side i do not believe in measurements, as i do not believe to my ears . I believe in what i hear that i can correlate to measurements, and the contrary.
and for a very good reason, we do not know how to measure everything, and the way we listen is a complex mix of mechanic-acoustical material( our ears) and complex neuronal treatment by our mind,including our desires and memory.
 
Last edited:
you consider all the other people, ... different approach as wrong or even stupid.

I do consider most of what is out there wrong, that much is correct, but not stupid, just behind the times. I used to believe that distortion was critically important in audio, I wasn't stupid, but I was wrong. I learned by doing just what you suggest - attempting to correlate what we hear to what we measure.

I have heard all your arguments before and checked them out. I'm not going to do it all again.
 
I have heard all your arguments before and checked them out. I'm not going to do it all again.
Well, i see any objective (scientific or technical) discussion is impossible with you. And seems to afraid you. You do not consider any evidence opposed to your "certainty", do not share any so calling "knowledge", and anything is wrong if it does not come from YOU.
Just think about the image you give of yourself, your intellectual honesty and your credibility when you says here :
Trust satisfied customers
and here:
I give no credence to subjectiuve listening tests
EOT, for me, i am sure i have nothing interesting to learn here, and that my fist feeling about reptilian lubricant was right.

Best regards and good sells.
 
Last edited:
I do consider most of what is out there wrong, that much is correct, but not stupid, just behind the times. I used to believe that distortion was critically important in audio, I wasn't stupid, but I was wrong. I learned by doing just what you suggest - attempting to correlate what we hear to what we measure.

...
It's not whether distortion is critically important or not, but at what level they become more critical than other factors. Due to the method of distortion measurement, it covers a very limited aspect of distortion that can be revealed.
 
Ex-Moderator R.I.P.
Joined 2005
well, there's no doubt in my mind that the worst distortion comes from crossover malfunction
it is very easy to hear when the right adjustment makes it less
and it can be so very small adjustment that you hardly believe it would make much difference
whether its phase related, or whatever, I don't know
but very audible, and makes a huge difference
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.