Hello,
For a part of my job I often work with orthogonal sets of curves, also I teach to my students the difference between an "equal area" transform which one keep area constant through the transform but not the shape and "equal angle" transform which one keep the shape constant but not the area.
Then I personnally live in an Euclidean world and I listen in that same Euclidean world that's why I would prefer if the wavefronts will be parallel in its "equidistant" meaning.
Just for the other people who want to undertand what it means, I put in attached file a graph everyone would recognize as the oblate spheroidal coordinate system (mu, theta). For what I understood the profile of the OS waveguide follows one curve in violet color. This means if I am right that the wavefronts are in green color.
I draw two segments for 2 different value of theta between 2 wavefronts. If the wavefronts were parallel in our Euclidean world the 2 blue segments should have the same length. Correct me if I am wrong but they have not.
So if the green curves are wavefronts and if speed of sound is constant (within the conditions in the waveguide) they are not reached at the same moment... But may be I am wrong in my interpretation and the green curves are not wavefronts...
Best regards from Paris, France
Jean-Michel Le Cléac'h
For a part of my job I often work with orthogonal sets of curves, also I teach to my students the difference between an "equal area" transform which one keep area constant through the transform but not the shape and "equal angle" transform which one keep the shape constant but not the area.
Then I personnally live in an Euclidean world and I listen in that same Euclidean world that's why I would prefer if the wavefronts will be parallel in its "equidistant" meaning.
Just for the other people who want to undertand what it means, I put in attached file a graph everyone would recognize as the oblate spheroidal coordinate system (mu, theta). For what I understood the profile of the OS waveguide follows one curve in violet color. This means if I am right that the wavefronts are in green color.
I draw two segments for 2 different value of theta between 2 wavefronts. If the wavefronts were parallel in our Euclidean world the 2 blue segments should have the same length. Correct me if I am wrong but they have not.
So if the green curves are wavefronts and if speed of sound is constant (within the conditions in the waveguide) they are not reached at the same moment... But may be I am wrong in my interpretation and the green curves are not wavefronts...
Best regards from Paris, France
Jean-Michel Le Cléac'h
gedlee said:
Attachments
Jmmlc said:
Your interpretation is not completely wrong, but I think that the point you try and make about "Then I personnally live in an Euclidean world and I listen in that same Euclidean world that's why I would prefer if the wavefronts will be parallel in its "equidistant" meaning." is completely without justification and given with only the weakest of support, which is basically "I don't like it." I don't like the idea that the waveguide contours for LeClerq or Avantgarde waveguides are drawn using equations that are completely incorrect as far as wave propagation goes and hence these contoure are drawn simply because they look good - no scientific justification is available. This kind of approach to engineering is simply not one that I could respect.
Because the OS solution is singular at the ends of the line it has to become complex (in a mathematical sense). But this is no different than the solution to a spherical wave near the origin where the phase angle between the pressure and the velocity are not in quadrature (as they are in the far field). So when talking about phase wavefronts, as you are, you must be careful about what it is you are talking about, the pressure wave or the velocity wave. For large off axis points very close to the origin disk ( the origin line in your drawing) the phase of the wavefronts for either the pressure or the velocity are not not coincident with the coordinate lines, but the group velocity of the magnitude of the wavefront is.
This effect becomes apparent in the OS because it changes with angle, while this situation does not change with angle in the spherical case and is not as apparent, but still occurs near the singularity. This gives the appearance of a changing wave speed, which it is for either the pressure or the velocity, but not for the wavefront when properly defined.
In any case, good waveguides are not usually used out to the angles that you have shown in your example and at points nearer the axis this effect is minimal - meaning that the pressure and velocity are very close to being in phase and as such the impedance is very close to be purely real. Because of the gradual slope of the impedance for an OS waveguide it can be shown that it will also have the lowest reactive component of any countour. I'm surprised that you missed this point.
Hello Earl,
Thank you to have (partly but largely) confirmed my analysis about the non equidistance of the "theorical wavefronts" in the OS waveguide.
About your direspect of the Le Cléac'h horn: this horn is designed in order to provide the most constant resistive acoustic loading to the driver and this can be maintained down to low frequency. People using Hornresp can verify it very easily and may compare to an OS waveguide (for which I could never obtain with Hornresp the low reactance you claim at least in the middle of its useful frequency range see attached graph).
Directivity of the Le Cleac'h will be controversial only at high frequency when using compression driver. The same Fc horn mounted on a 17centimeter loudspeaker is often judged as giving a very wide sweet spot. I cannot see anything in that an audiophile could disapprove.
What lead me to the question of the "theorical" non equidistant wavefronts in the OS waveguide was the difference between the published pulse responses for the OS waveguide at different angle (specifically 0° and 22.5°). I found that the difference in their shapes was quite noticeable... I have no other comments to do on that.
BTW: Is there a formula giving the evolution of the area of the wavefronts in the waveguide versus their distance to the throat?
Best regards from Paris, France
Jean-Michel Le Cléac'h
attached graph: comprison between acoustical resistane, acoustical reactance for an OS waveguide and a Le Cléac'h horn of same outer dimensions.
Thank you to have (partly but largely) confirmed my analysis about the non equidistance of the "theorical wavefronts" in the OS waveguide.
About your direspect of the Le Cléac'h horn: this horn is designed in order to provide the most constant resistive acoustic loading to the driver and this can be maintained down to low frequency. People using Hornresp can verify it very easily and may compare to an OS waveguide (for which I could never obtain with Hornresp the low reactance you claim at least in the middle of its useful frequency range see attached graph).
Directivity of the Le Cleac'h will be controversial only at high frequency when using compression driver. The same Fc horn mounted on a 17centimeter loudspeaker is often judged as giving a very wide sweet spot. I cannot see anything in that an audiophile could disapprove.
What lead me to the question of the "theorical" non equidistant wavefronts in the OS waveguide was the difference between the published pulse responses for the OS waveguide at different angle (specifically 0° and 22.5°). I found that the difference in their shapes was quite noticeable... I have no other comments to do on that.
BTW: Is there a formula giving the evolution of the area of the wavefronts in the waveguide versus their distance to the throat?
Best regards from Paris, France
Jean-Michel Le Cléac'h
attached graph: comprison between acoustical resistane, acoustical reactance for an OS waveguide and a Le Cléac'h horn of same outer dimensions.
gedlee said:
Attachments
Jmmlc said:
First, you must not have read what I wrote very carefully because the wavefronts are equidistant when defined correctly.
Second, thanks for proving my point that for equal cutoffs, the OS will have lower reactance and a smoother resistive load that the LeClearch horn. You, of course, have choosen to compare apples to oranges by NOT setting the cutoffs the same. When this is done what I claim will be clearly evident as true.
Of course there would be an equation giving the area as you desire. I would have no use for this equation so I have not developed it, but it probably wouldn't be too hard.
Hello --
I'm still a novice when it comes to horns, so please excuse the simple-minded questions. In past postings Earl Geddes has pointed out that the OS Waveguide is designed to be constant directivity whereas the Le Cléac'h has a directivity that narrows with frequency. So that is clearly one major difference between the designs.
In the two most recent posts there are some more data comparing the two approaches. On the one hand for a given physical size it would appear that the Le Cléac'h will have a usable frequency range nearly 10x lower than the OS Waveguide. However, Earl Geddes points out that the OS Waveguide has the lowest reactance and smoothest resistance for a given cutoff frequency.
I just want to make sure I'm reading these posts correctly. Earl, do you agree that the OS Waveguide will have to be 10x larger than the Le Cléac'h for a given cutoff frequency? This seems rather extreme to me and not at all the impression I get just from looking at pictures of the Summa (and its descendants).
I'm not trying to stir up a hornet's nest, just trying to get a clearer picture of what is happening with these two designs.
I suspect that there may be some assumptions used in the McBean HornResponse program that is skewing the results. For example, although I haven't used the program for many years, McBean once wrote that the acoustic output was for the total radiated power and not the on-axis frequency response. I'm not sure that this has any bearing on the current topic, but perhaps Mr. Geddes has some reservations about the results produced by Horn Response.
Thanks for your input.
I'm still a novice when it comes to horns, so please excuse the simple-minded questions. In past postings Earl Geddes has pointed out that the OS Waveguide is designed to be constant directivity whereas the Le Cléac'h has a directivity that narrows with frequency. So that is clearly one major difference between the designs.
In the two most recent posts there are some more data comparing the two approaches. On the one hand for a given physical size it would appear that the Le Cléac'h will have a usable frequency range nearly 10x lower than the OS Waveguide. However, Earl Geddes points out that the OS Waveguide has the lowest reactance and smoothest resistance for a given cutoff frequency.
I just want to make sure I'm reading these posts correctly. Earl, do you agree that the OS Waveguide will have to be 10x larger than the Le Cléac'h for a given cutoff frequency? This seems rather extreme to me and not at all the impression I get just from looking at pictures of the Summa (and its descendants).
I'm not trying to stir up a hornet's nest, just trying to get a clearer picture of what is happening with these two designs.
I suspect that there may be some assumptions used in the McBean HornResponse program that is skewing the results. For example, although I haven't used the program for many years, McBean once wrote that the acoustic output was for the total radiated power and not the on-axis frequency response. I'm not sure that this has any bearing on the current topic, but perhaps Mr. Geddes has some reservations about the results produced by Horn Response.
Thanks for your input.
Hello,
One of the goal of the Le Cléac'h horn is that for outer dimensions smaller than an OS waveguide it can be used in the frequency range inside which its acoustic impedance is purely resistive.
To obtain an equivalent cut-off frequency (if you defined it by example at the point for which the resistive part of the impedance halves) as the example I shown in my last message, then the Le Cleac'h horn will be very small and only usable for a supertweeter...
Best regards from Paris, France
Jean-Michel Le Cléac'h
One of the goal of the Le Cléac'h horn is that for outer dimensions smaller than an OS waveguide it can be used in the frequency range inside which its acoustic impedance is purely resistive.
To obtain an equivalent cut-off frequency (if you defined it by example at the point for which the resistive part of the impedance halves) as the example I shown in my last message, then the Le Cleac'h horn will be very small and only usable for a supertweeter...
Best regards from Paris, France
Jean-Michel Le Cléac'h
gedlee said:
Rybaudio said:
exactly
people - normal listeners with normal hearing or even with impaired hearing!
but not "audiophiles"
that's the problem with the hifi/hiend industry - customers that form its core target market are very seldom normal listeners
best,
graaf
gedlee said:
Dr. Geddes, are You sure that "We" (=diy audio users) are always talking about "high fidelity reproduction"?
well, "audiophiles" are not interested in "high fidelity reproduction" at all
they are interested in having fun playing with sound reproduction equipment
they reach their nirvana not when they hear music properly reproduced but rather when they hear difference , especially after connecting a different power cord 😉 (audiophiles-shopaholics type) or after inserting a different capacitor in the power supply (audiophiles-DIY hobbyist type)
that is their hobby, not "high fidelity reproduction"
best,
graaf
Charles Hansen said:
No I don't agree with this conclusion at all. Theoretically the sizes would not be any different. I don't know what mistakes are being made in the comparisons that Mr. Le Cléac'h is making, but I don't agree with them.
The LF portion of all horns or waveguides work virtually identically and depend, as Webster assumed, only on the rate of change of area with length. If this average rate of change is the same, no matter what the details are, then the LF function must be the same. Where things get dramatically different is at the higher frequencies where the wall contour matters for diffraction.
It is easy to see that Mr. Le Cléac'h's argument, that it is the LFs that are different for his device and the OS, is incorrect because it is exactly this region where they HAVE TO BE the most similar. He keeps comparing designs that are not comparable and drawing conclusions from that invalid comparison.
Hello,
I waited for Earl's reply to your question and am not surprised by his reply.
The good question should be:
- we know how looks of a 160Hz Le Cléac'h horn. Even many people use the Azura AH160 with great pleasure and several of them even use it with no high pass crossover ... so what is the shape and dimension of a 160Hz waveguide (or whatever frequency below 600Hz you can want...).
I asked Earl to describe a OS waveguide having a 160Hz cut-off but until now he refused to do it.
But you may use hornresp and see if you can design a 160Hz OS waveguide ... you'll see it is a most difficult task.
Best regards
Jean-Michel Le Cléac'h
I waited for Earl's reply to your question and am not surprised by his reply.
The good question should be:
- we know how looks of a 160Hz Le Cléac'h horn. Even many people use the Azura AH160 with great pleasure and several of them even use it with no high pass crossover ... so what is the shape and dimension of a 160Hz waveguide (or whatever frequency below 600Hz you can want...).
I asked Earl to describe a OS waveguide having a 160Hz cut-off but until now he refused to do it.
But you may use hornresp and see if you can design a 160Hz OS waveguide ... you'll see it is a most difficult task.
Best regards
Jean-Michel Le Cléac'h
Charles Hansen said:
Jmmlc said:
Umm, to sum it up the Gedleel approach doesn't require the horn to load that low. It has different qualities for different purposes. Why do you insist on looking at it from such an obscure angle?
Jmmlc said:
Nor am I surprised by Jean-Michel response. I have already stated that I don't agree with what he has posted from Hornresp and that it is not my job to debug someone elses program. My work has been published in per-reviewed journals and has been place under substantial scrutiny and has held up admirably and changed the marketplace for good. Thus it is a standard which must be compared to when one is "validating" a piece of software. I contend that this has not been done and I won't go any further with this line of discussion.
I expect that what is being done is that equal mouth and throat areas are being entered with a 45° waveguide, which will yield a very wide very short device which has poor LF capabilities, but CD performance. How does one compare a non-CD horn with a CD waveguide? There really isn't any way to do this. One is CD and one is not. Thats the absurdity of this discussion. If one wants CD (which I contend is a necessity) then horns won't do this and you must use a waveguide. Even Jean-Michel admits to this. Hence I will discuss whether CD is desirable, necessary or whatever, but comparing a CD waveguide to a non-CD horn is pointless.
Hello,
In the simulations I described, both the Le Cléac'h horn and the Waveguide have the same length, throat area and mouth area
throat area: S1 = 2,26 cm²
mouth area: S2 = 11354,06 cm²
length = 146,02 cm
Both use my TAD TD2001 modelisation as driver.
That's a very funny thing that Hornresp while using the same method for every kind of horns or waveguides gives results similar to theory for Salmon horns, conical, etc and not for the OS waveguide... ;-)
Best regards from Paris,
Jean-Michel Le CLéac'h
In the simulations I described, both the Le Cléac'h horn and the Waveguide have the same length, throat area and mouth area
throat area: S1 = 2,26 cm²
mouth area: S2 = 11354,06 cm²
length = 146,02 cm
Both use my TAD TD2001 modelisation as driver.
That's a very funny thing that Hornresp while using the same method for every kind of horns or waveguides gives results similar to theory for Salmon horns, conical, etc and not for the OS waveguide... ;-)
Best regards from Paris,
Jean-Michel Le CLéac'h
Hence I will discuss whether CD is desirable, necessary or whatever.
Hi Earl,
i have not read your Summa paper, and don't understand the mathematical equations , and have not followed all your writings and your posts about CD horns. So my knowledge is limited. However, it interests me in a practical way. I want to implement for my system the best possible contour, and since i build my horns, i have freedom to do so. Beside constant directivity of wave front of your OS waveguide, why is OS superior compared to LeCleac'h, Tractrix , and other contours ? how about tone, naturality, neutrality. Have you made direct comparisons ? What are the effective advantages ?
Angelo
Hi Earl,
i have not read your Summa paper, and don't understand the mathematical equations , and have not followed all your writings and your posts about CD horns. So my knowledge is limited. However, it interests me in a practical way. I want to implement for my system the best possible contour, and since i build my horns, i have freedom to do so. Beside constant directivity of wave front of your OS waveguide, why is OS superior compared to LeCleac'h, Tractrix , and other contours ? how about tone, naturality, neutrality. Have you made direct comparisons ? What are the effective advantages ?
Angelo
angeloitacare said:
Dr Geddes time isn't free - wouldn't it be respectful to read the paper before asking questions?
As I see it, you're basically saying "I'm too lazy to even read the paper you spent years refining, could you summarize it for me in a paragraph.?"
If you want a subjective analysis of the OS waveguide performance, take a look a the thread here:
http://www.diyaudio.com/forums/showthread.php?threadid=122318&goto=newpost
I know I'm being a bit rude myself, but it's frustrating to see his time abused like this.
angeloitacare said:
You really should read the white paper, and the reviews. They have all the answers that you are looking for and it saves me a whole lot of typing.
Bottom line is that CD is an absolute requirement for good sound quality in a small room. Everyone agrees with this (well almsot everyone) including Floyd Toole and Sigfried Linkwitz. Hence you can't just say "Beside constant directivity of wave front of your OS waveguide" because CD is everything. But achieving CD without colorations from diffraction, etc. is no small task. And its not just the OS contour but the foam plug that makes all the difference, we have just tended to argue here amount the contour. First you have to get the contour right, then the foam plug goes to the next step. But read the reviews and then the white paper and if you have more questions, I'd be glad to address them.
Member
Joined 2003
Since the OS flare needs to be coupled to a low diffraction mouth, and the LeCleach may be one of the most highly evolved mouths (in terms of diffraction performance vs diameter), what would be the result of mating an OS waveguide with a large LeCleach mouth?
This is a sincere question in search of sincere response. (No sparks please.)
Thanks,
Paul
This is a sincere question in search of sincere response. (No sparks please.)
Thanks,
Paul
Paul W said:
All horns or waveguides need low diffraction mouths not just OS. Based on my research and the results from the waveguides that I have tested it all comes down to the radius. The larger the better. I don't see how a varying radius would be of any benefit. If I had, I would have used one.
Jean-Michel,
A little off-topic here, but could you post the input parameters you use to model the TAD driver in Hornresp? Thank you.
Rgs, JLH
A little off-topic here, but could you post the input parameters you use to model the TAD driver in Hornresp? Thank you.
Rgs, JLH
