"Here is a compared wavefronts simulation between conical and Le Cléac'h horn of the same size at 1kHz."
That seems like a flawed comparison - the cone has no mouth radius.
The other is interesting in that it seems to suggest that there's benefit to continuing the mouth radius past 90 deg.
Would it be possible to show both profiles terminated with a large radius into the baffle plane?
That seems like a flawed comparison - the cone has no mouth radius.
The other is interesting in that it seems to suggest that there's benefit to continuing the mouth radius past 90 deg.
Would it be possible to show both profiles terminated with a large radius into the baffle plane?
Looking closer at the scale of the horns that Jean-Michel had presented, it seems the horns are quite large, about 1.2M+. What would happen when the wave length is close to the length of the horn, mouth curvature effects, etc.
Hornresp is a neat program, but it's really not implemented to explore these effects with the flexibility necessary.
Listening to the new small wave guide and the old wave guide with a modified throat, I must say both are very revealing, and absolute polarity of each instrument mixed into the music is so distinguishable that you can clearly identify mixed polarity in a piece of recording. Especially evident are vocals vs instruments; when I hear really good focus in the vocals, the instruments become a bit fuzzy, and vice versa.
Hornresp is a neat program, but it's really not implemented to explore these effects with the flexibility necessary.
Listening to the new small wave guide and the old wave guide with a modified throat, I must say both are very revealing, and absolute polarity of each instrument mixed into the music is so distinguishable that you can clearly identify mixed polarity in a piece of recording. Especially evident are vocals vs instruments; when I hear really good focus in the vocals, the instruments become a bit fuzzy, and vice versa.
Hello Timo,
For the Le Cléac'h horn, the results obtained with Hornresp fit quite good with the ones obtained with long duration BEM simulations.
I cannot see why it should be useful to simulate the wavefront at 10kHz of a Le Cléac'h horn having a 160Hz or 320Hz acoustic cut-off frequency...
But what is sure is that a conical horn is bad at every frequency when it comes to the point of the pressure field (and not only inside the conical horn).
Then, for a waveguide a large radius at the mouth solves partially (only partially) the effect of mouth diffraction but have no or very few effect on the HOMs created near the throat in the zone of rapide curvature. That was clear on the pressure field map from your simulation (and from other BEM simulations I could see).
Best regards from Paris,
Jean-Michel Le Cléac'h
For the Le Cléac'h horn, the results obtained with Hornresp fit quite good with the ones obtained with long duration BEM simulations.
I cannot see why it should be useful to simulate the wavefront at 10kHz of a Le Cléac'h horn having a 160Hz or 320Hz acoustic cut-off frequency...
But what is sure is that a conical horn is bad at every frequency when it comes to the point of the pressure field (and not only inside the conical horn).
Then, for a waveguide a large radius at the mouth solves partially (only partially) the effect of mouth diffraction but have no or very few effect on the HOMs created near the throat in the zone of rapide curvature. That was clear on the pressure field map from your simulation (and from other BEM simulations I could see).
Best regards from Paris,
Jean-Michel Le Cléac'h
tiki said:Hello Jean-Michel,
several horns look good up to 1kHz in Hornresponse. One should try higher frequencies to separate the wheat from the chaff.
Thanks and regards, Timo
Hello Soongsc,
The horn in the simulation you mentionned had a 160Hz cut-off (as the original Azura Horn built by Martin Seddon).
You asked what will happen for a frequency the wavelength will be similar to the horn length (1.2m), this means a frequency around 290Hz.
You'll find in attached file a graph obtained with Hornresp's wavefronts simulator of that horn at a frequency of 300Hz.
Best regards from Paris,
Jean-Michel Le Cléac'h
The horn in the simulation you mentionned had a 160Hz cut-off (as the original Azura Horn built by Martin Seddon).
You asked what will happen for a frequency the wavelength will be similar to the horn length (1.2m), this means a frequency around 290Hz.
You'll find in attached file a graph obtained with Hornresp's wavefronts simulator of that horn at a frequency of 300Hz.
Best regards from Paris,
Jean-Michel Le Cléac'h
soongsc said:
Attachments
Hello Ear
Sorry if my last message lead you to think that I mix conical horn and waveguide but no.
The waveguide part of my message was related to Tiki's simulation (an OS waveguide with large radius curved mouth).
In that simulation (as you confirmed) the lobes in the sound radiations could be due to Homs. As the mouth is curved, one could think that those Homs are generated in the rapid curvature change part near the throat of the OS waveguide (a feature you once described in one of your message).
In a Le Cléac'h horn, near the throat, such rapid curvature change doesn't exist and this could be one reason why less Homs are generated.
Best regards from Paris, France
Jean-Michel Le Cléac'h
Sorry if my last message lead you to think that I mix conical horn and waveguide but no.
The waveguide part of my message was related to Tiki's simulation (an OS waveguide with large radius curved mouth).
In that simulation (as you confirmed) the lobes in the sound radiations could be due to Homs. As the mouth is curved, one could think that those Homs are generated in the rapid curvature change part near the throat of the OS waveguide (a feature you once described in one of your message).
In a Le Cléac'h horn, near the throat, such rapid curvature change doesn't exist and this could be one reason why less Homs are generated.
Best regards from Paris, France
Jean-Michel Le Cléac'h
gedlee said:
Jmmlc said:
Is the "l" on your keyboard broke?
This is no proof that with your curvature "less Homs are generated" than in an OS, in fact I would expect the exact opposite. HOMs don't cause "lobes" and in fact they are usually small enough that they have little to no effect on the radiation pattern - this was proven by Gottfried Behler at Achen. His mistake was assuming that because they had no significant polar effect that they were not an issue. But a small polar effect does not always mean a small subjective effect.
You keep trying to elevate the improtance of your contour, but the fact remains that it is not CD. Until you have a contour that is CD it is meaningless to compare it to one that is.
Hello Earl,
(sorry for the missing "l" last time , few letters are worn on my keyboard...)
May be the word "lobe" was badly chosen, I should have used your own words: "angular dependence".
quote from
http://www.diyaudio.com/forums/showthread.php?postid=1745665#post1745665
"Timo: Does this small simulation, trying to express the wavguide, show the stated HOMs?"
http://ibtk.de/project/speaker/linearray/simulation/20081222_LA-WG_Hornresp.gif
"Earl: it may be that those are HOM, they are clearly not smooth wavefront in an angular sense so they have to have some angular dependance and as such would have to be HOM."
Best regards from Paris, France
Jean-Michel Le Cléac'h
(sorry for the missing "l" last time , few letters are worn on my keyboard...)
May be the word "lobe" was badly chosen, I should have used your own words: "angular dependence".
quote from
http://www.diyaudio.com/forums/showthread.php?postid=1745665#post1745665
"Timo: Does this small simulation, trying to express the wavguide, show the stated HOMs?"
http://ibtk.de/project/speaker/linearray/simulation/20081222_LA-WG_Hornresp.gif
"Earl: it may be that those are HOM, they are clearly not smooth wavefront in an angular sense so they have to have some angular dependance and as such would have to be HOM."
Best regards from Paris, France
Jean-Michel Le Cléac'h
gedlee said:
Jmmlc said:
"it MAY BE that those are HOM" seems to me to state pretty clearly that I am not convinced that these sims are accurate or representative of anything. They are very coarse and the HOM are, in general, fairly subtle - they are indeed hard to measure. While I love that wave simulation tool, I would never count on it to be accurate enough to draw any conclusions from at frequencies where the wavelengths are quite short. Numerical sims are notoriuosly unstable for short wavelengths.
Hi Jean-Michel,Jmmlc said:
It seems that even before the cut-off frequency, the directivity widens. Also, it seems with the conical horn simulation, the HOMs are generated by the mouth if the throat conditions are ideal and the wave length is significantly smaller than the length of the horn/guide. With the two wave guides I have, the larger one has more honk to it even though it has an elliptical cross section at the mouth. The smaller one seems to have less honk.
With the Le Cléac'h horn, less HOM's are generated from the throat, and the expansion is such a way that HOMs do not have a chance to form until probably at very high frequencies due to the short fast flare. I somehow get a feeling that throat flare can be widened to a certain point before HOMs start becoming significant. So how to have a good directivity control without introducing lots of HOMs will require two different flare equations for the throat and mouth. This is currently what I hope to discover. I certainly wish Hornresp had the capability to do this.
I am curious as to what it would be like if the horn/guide cut-off frequency, driver cut-off frequency, and directivity controlled frequency were all pretty much the same.
It will be a few more days till I actually do any measurements. Hope to train my ears to see if my ears hear certain things that I might see in the data.
soongsc said:
Soongsc
This is completely false. The HOM that are generated at the throat depend on how the driver connects to the device. In those sims the source was a point source which will generate lots of HOM in any device. But the longer and straighter the throat (in any sim where the source is not a plane wave) the less diffraction there will be at lower freqs. A straight tube will not have any. After looking at the sims they are completely inaccurate at higher freqs (the only place where waveguides do much of anything)because only a point source at the throat is possible. Don't judge anything based on those simple examples.
Earl,
You certainly have much better skills handling that Java applet than I. The sim looks really nice.
Now, if we look at the throat size, guide size, and wave length. The wave guide is quite short compared to the wave length. I think if we look at a larger guide with the same throat at the same wave length, HOMs will start to show. However, if then we start teaking the flare on the outer half of the wave guide, we should see things becoming better in terms of HOMs.
When the wave guide shape is optimum for it's size, the remaining thing we have not simed is the actual diaphragm to the throat.
You certainly have much better skills handling that Java applet than I. The sim looks really nice.
Now, if we look at the throat size, guide size, and wave length. The wave guide is quite short compared to the wave length. I think if we look at a larger guide with the same throat at the same wave length, HOMs will start to show. However, if then we start teaking the flare on the outer half of the wave guide, we should see things becoming better in terms of HOMs.
When the wave guide shape is optimum for it's size, the remaining thing we have not simed is the actual diaphragm to the throat.
soongsc said:
That depends on what you mean by "larger". HOMs do not depend on the length of the device in any way so there will not be more if it is just longer with a larger mouth. But if by larger you mean wider angle, then yes, the HOMs do get generated at a lower frequency in that case.
Well, there are bits and pieces scattered around which I cannot remember the sources of, but one place that I started out at is here.
http://www.quarter-wave.com/Horns/Horn_Theory.html
Basically, when the wave velocity is not in parallel with the wall it's travelling along, reflections will occur with is the beginning of HOMs. Friction will cause the velocity to gradually change direction. However, if the wall is not long enough, then the reflection does not build up strong.
Having said this, I have opened a can of worms. But I'm not going to try to convince anyone in forums as usual.
http://www.quarter-wave.com/Horns/Horn_Theory.html
Basically, when the wave velocity is not in parallel with the wall it's travelling along, reflections will occur with is the beginning of HOMs. Friction will cause the velocity to gradually change direction. However, if the wall is not long enough, then the reflection does not build up strong.
Having said this, I have opened a can of worms. But I'm not going to try to convince anyone in forums as usual.
