J-M, bonjour!
Let's consider the situation where the xover to a bass section is not an issue - for example, we want to only listen to the horn. Doubtless this has been done, if only for testing and evaluation.
In this case, is the group delay from running down to the mouth frequency, in your experience, objectionable or noticeable?
If so, how does it manifest itself in terms of a subjective description when listening to a stereo program?
I am asking this, since the xover that I use along with my present (exponential) horn(s) is "non-traditional" since it looks very much like a "brick wall" in the frequency domain at the roll-off - an impossibly steep roll-off. So, I am not counting at all on having a proper phase relationship between the horn and the bass section!
Let me explain: The way I see it is that if the bass section stops like a brick wall, then the horn is free to do whatever it does naturally without any frequencies interfering with its operation within its passband. Now at the frequencies below the horn's operational frequency (the mouth rolloff) the output of the horn drops very quickly - not quite a brick wall. So, in fact for less than 1 octave (in my system 125Hz - ~250hz) there is less than optimum cooperation between the two drivers. But in practice the horn is many dB down by the time you get to 125Hz. and there is actually a very narrow range of frequencies exactly at and below the actual xover freqs for the two drivers where things are clearly imperfect. But that's a very very narrow range that shows up mostly as a harmonic part of a musical tone and in reality isn't going to show up very often - and if it does will likely not be noticed in the flow of the music/sound.
I then adjust the physical position of the acoustic centers of the drivers using an impulse response for proper phase and best impulse.
Fyi, when I used to use the usual xover methods, which permitted the bass and horn to overlap through the xover frequency it always seemed to be less than satisfactory in total.
This is how I run it. Get the woofer out of the picture entirely so that the horn is pristine in its presentation.
It seems to work...
So, this is why I am curious about running the horn down into the mouth frequency alone - in essence that is what I am doing!
_-_-bear
Let's consider the situation where the xover to a bass section is not an issue - for example, we want to only listen to the horn. Doubtless this has been done, if only for testing and evaluation.
In this case, is the group delay from running down to the mouth frequency, in your experience, objectionable or noticeable?
If so, how does it manifest itself in terms of a subjective description when listening to a stereo program?
I am asking this, since the xover that I use along with my present (exponential) horn(s) is "non-traditional" since it looks very much like a "brick wall" in the frequency domain at the roll-off - an impossibly steep roll-off. So, I am not counting at all on having a proper phase relationship between the horn and the bass section!
Let me explain: The way I see it is that if the bass section stops like a brick wall, then the horn is free to do whatever it does naturally without any frequencies interfering with its operation within its passband. Now at the frequencies below the horn's operational frequency (the mouth rolloff) the output of the horn drops very quickly - not quite a brick wall. So, in fact for less than 1 octave (in my system 125Hz - ~250hz) there is less than optimum cooperation between the two drivers. But in practice the horn is many dB down by the time you get to 125Hz. and there is actually a very narrow range of frequencies exactly at and below the actual xover freqs for the two drivers where things are clearly imperfect. But that's a very very narrow range that shows up mostly as a harmonic part of a musical tone and in reality isn't going to show up very often - and if it does will likely not be noticed in the flow of the music/sound.
I then adjust the physical position of the acoustic centers of the drivers using an impulse response for proper phase and best impulse.
Fyi, when I used to use the usual xover methods, which permitted the bass and horn to overlap through the xover frequency it always seemed to be less than satisfactory in total.
This is how I run it. Get the woofer out of the picture entirely so that the horn is pristine in its presentation.
It seems to work...
So, this is why I am curious about running the horn down into the mouth frequency alone - in essence that is what I am doing!
_-_-bear
bear said:Now I be use horns mee too.
Ditto! But I was fortunate enough to hear good ones when I was young. That kinda stuck with me. So now I'll be trying low x-over with Radian drivers on big horns.
Bear, the modded DCX is no Ferrari, really- but maybe an Alpha Romeo 8C. You are more demanding than most, so it may not be your cup of tea (to mix metaphors) but for the rest of us, it will do nicely. =)
Hello,
There is nothing like "mouth frequency", this is a concept coming from the PA world (where everything has to me small) that I avoided myself since long time as I only use quasi infinite horns.
From my experience audibility of the rise in group delay near the acoustical cut-off frequency depends:
1) on the sensitivity of the listener to phase distorsion
2) for people sensible to phase distortion, the sensitivity to phase distortion is reduced below 250z
3) on the loudspeaker (Fs, Qts...) and its rear load
2) means that if the horn possess a very low cut-off, then the use of a 6dB crossover or no crossover is possible. But one possible problem then, is that the sonic results depends a lot of the resonance frequency of the driver due to the fact that the loudspeaker is not well acoustically loaded below the cut-off frequency of the horn.
Many people are using Martin Seddon's Azura horn AH160 without any crososover.
For use of a horn the cut-off of which is above 250Hz, a crossover has to be used. If not, as an example the voice seems to come from very deep inside the horn and a bit like when you are in a cave (depends of the loudspeaker though).
Have you ever used my crossover / alignment spreadsheet.
With 2 French guys, until now by step by step method I began in 1982, we could only find 4 types of quasioptimal crossovers using conventional l crossovers low-pass nad high pass (Butterworth, Bessel, Linkwitz-Riley...)
2 versions of the spreadsheet exist, one with response curves + group delay curve the other one with response curves + square wave response.
They are downloadable on my friend Nicolas Davidenko's website at:
http://ndaviden.club.fr/outils/filtre_simul.zip
http://ndaviden.club.fr/outils/filtre_carre.zip
Best regards from Paris, France
Jean-Michel Le Cléac'h
There is nothing like "mouth frequency", this is a concept coming from the PA world (where everything has to me small) that I avoided myself since long time as I only use quasi infinite horns.
From my experience audibility of the rise in group delay near the acoustical cut-off frequency depends:
1) on the sensitivity of the listener to phase distorsion
2) for people sensible to phase distortion, the sensitivity to phase distortion is reduced below 250z
3) on the loudspeaker (Fs, Qts...) and its rear load
2) means that if the horn possess a very low cut-off, then the use of a 6dB crossover or no crossover is possible. But one possible problem then, is that the sonic results depends a lot of the resonance frequency of the driver due to the fact that the loudspeaker is not well acoustically loaded below the cut-off frequency of the horn.
Many people are using Martin Seddon's Azura horn AH160 without any crososover.
For use of a horn the cut-off of which is above 250Hz, a crossover has to be used. If not, as an example the voice seems to come from very deep inside the horn and a bit like when you are in a cave (depends of the loudspeaker though).
Have you ever used my crossover / alignment spreadsheet.
With 2 French guys, until now by step by step method I began in 1982, we could only find 4 types of quasioptimal crossovers using conventional l crossovers low-pass nad high pass (Butterworth, Bessel, Linkwitz-Riley...)
2 versions of the spreadsheet exist, one with response curves + group delay curve the other one with response curves + square wave response.
They are downloadable on my friend Nicolas Davidenko's website at:
http://ndaviden.club.fr/outils/filtre_simul.zip
http://ndaviden.club.fr/outils/filtre_carre.zip
Best regards from Paris, France
Jean-Michel Le Cléac'h
bear said:
we want to only listen to the horn. In this case, is the group delay from running down to the mouth frequency, in your experience, objectionable or noticeable?
If so, how does it manifest itself in terms of a subjective description when listening to a stereo program?
Fyi, when I used to use the usual xover methods, which permitted the bass and horn to overlap through the xover frequency it always seemed to be less than satisfactory in total.
So, this is why I am curious about running the horn down into the mouth frequency alone - in essence that is what I am doing!
_-_-bear
Bonjour J-M,
By "mouth frequency" I am refering the the lower frequency cutoff, which I presume is still related to the width/diameter of the horn??
I don't think that there are any compression drivers that can be operated full range except at micropower levels... I presume they are loading the horn with a "cone driver"??
When you say that the voice seems to come very deep within the horn, are you referring only to the case where a cone driver is used to load the horn? If so, this is not my case. I will only use a compression driver.
And by "voice" do you refer to the lower frequency part of the voice or the middle part, or the higher freqs, or the entire presentation of the horn seems to be moved back inside the horn? If so this is a very odd effect, one that is almost worth building a horn just to experience!
Have you noted the same effect with a compression driver, or has no one done it with a compression driver?
My goal would be to reach useable acoustic output at 250Hz. or lower if possible.
I'll download the spread sheets and see what they show...
I'm not sure why the use of a higher order xover effects the group delay in a positive way... perhaps that will become apparent after looking at the downloads...
_-_-bear
By "mouth frequency" I am refering the the lower frequency cutoff, which I presume is still related to the width/diameter of the horn??
I don't think that there are any compression drivers that can be operated full range except at micropower levels... I presume they are loading the horn with a "cone driver"??
When you say that the voice seems to come very deep within the horn, are you referring only to the case where a cone driver is used to load the horn? If so, this is not my case. I will only use a compression driver.
And by "voice" do you refer to the lower frequency part of the voice or the middle part, or the higher freqs, or the entire presentation of the horn seems to be moved back inside the horn? If so this is a very odd effect, one that is almost worth building a horn just to experience!
Have you noted the same effect with a compression driver, or has no one done it with a compression driver?
My goal would be to reach useable acoustic output at 250Hz. or lower if possible.
I'll download the spread sheets and see what they show...
I'm not sure why the use of a higher order xover effects the group delay in a positive way... perhaps that will become apparent after looking at the downloads...
_-_-bear
Hello Bear,
Keele in a famous paper: "Optimum mouth size" in 1973
http://www.xlrtechs.com/dbkeele.com/PDF/Keele (1973-09 AES Preprint) - Optimum Horn Mouth Size.pdf
used a very simplistic theorical model of the horn (and specifically the modelized horns profiles were calculated on a expansion of planar wavefronts) to study the effect of mouth size on rippling in the frequency response curve. He concluded that for a given cutoff of the horn it exist an optimal circumference and added it is useless to use a larger mouth size.
In fact more recently it has been shown that such an optimum doesn't exist (if we calculate the profile with curved wavefronts) and using larger mouth lead to be smoother response curve and less reflections from the mouth.
This can be verified easily using Hornresp.
It is still convenient in PA in order to reduce the size of the horns while accepting some ripple in the frequency response to use Keele's formula but IMHO it should not be used in HIFI.
I know few persons who are often using large Le Cléac'h horns (Azura Horns or Musique Concrete) on compression drivers without any high-pass crossover (but not at too high SPL levels)
For the same cut-off of the horns -the first one using a cone loudspeaker the other one a compression driver- there is only a small difference between the effect of the first formants (low frequency components) of the voice seeming outcoming from deep in the throat of the horn (this difference is probably due to the difference in the diameters of the throat.)
My spreadsheet in order to model the behavior of crossover/alignment is limited to classical polynomials filters (Butterworth, Bessel...) and to the 4th order.
For more information in English by a user of my method:
http://freerider.dyndns.org/anlage/LeCleach.htm
Best regards from Paris, France
Jean-Michel Le Cléac'h
Keele in a famous paper: "Optimum mouth size" in 1973
http://www.xlrtechs.com/dbkeele.com/PDF/Keele (1973-09 AES Preprint) - Optimum Horn Mouth Size.pdf
used a very simplistic theorical model of the horn (and specifically the modelized horns profiles were calculated on a expansion of planar wavefronts) to study the effect of mouth size on rippling in the frequency response curve. He concluded that for a given cutoff of the horn it exist an optimal circumference and added it is useless to use a larger mouth size.
In fact more recently it has been shown that such an optimum doesn't exist (if we calculate the profile with curved wavefronts) and using larger mouth lead to be smoother response curve and less reflections from the mouth.
This can be verified easily using Hornresp.
It is still convenient in PA in order to reduce the size of the horns while accepting some ripple in the frequency response to use Keele's formula but IMHO it should not be used in HIFI.
I know few persons who are often using large Le Cléac'h horns (Azura Horns or Musique Concrete) on compression drivers without any high-pass crossover (but not at too high SPL levels)
For the same cut-off of the horns -the first one using a cone loudspeaker the other one a compression driver- there is only a small difference between the effect of the first formants (low frequency components) of the voice seeming outcoming from deep in the throat of the horn (this difference is probably due to the difference in the diameters of the throat.)
My spreadsheet in order to model the behavior of crossover/alignment is limited to classical polynomials filters (Butterworth, Bessel...) and to the 4th order.
For more information in English by a user of my method:
http://freerider.dyndns.org/anlage/LeCleach.htm
Best regards from Paris, France
Jean-Michel Le Cléac'h
bear said:Bonjour J-M,
By "mouth frequency" I am refering the the lower frequency cutoff, which I presume is still related to the width/diameter of the horn??
I don't think that there are any compression drivers that can be operated full range except at micropower levels... I presume they are loading the horn with a "cone driver"??
When you say that the voice seems to come very deep within the horn, are you referring only to the case where a cone driver is used to load the horn? If so, this is not my case. I will only use a compression driver.
And by "voice" do you refer to the lower frequency part of the voice or the middle part, or the higher freqs, or the entire presentation of the horn seems to be moved back inside the horn? If so this is a very odd effect, one that is almost worth building a horn just to experience!
Have you noted the same effect with a compression driver, or has no one done it with a compression driver?
My goal would be to reach useable acoustic output at 250Hz. or lower if possible.
I'll download the spread sheets and see what they show...
I'm not sure why the use of a higher order xover effects the group delay in a positive way... perhaps that will become apparent after looking at the downloads...
_-_-bear
Jmmlc said:
Hi Everyone,
I noticed the same thing myself when developing and testing the original Hornresp in 1970 - three years before the Keele paper was published
.Unlike Keele though, I came to the conclusion that the "optimum mouth size" phenomenon was due to a limitation in the plane wavefront model itself.
As Jean-Michel points out, the limitation disappears when an isophase wavefront model is used.
Kind regards,
David
Hello,
There have been Le Cléac'h horns built in concrete, in wood, in "papier maché", in fiberglass + resin, in polystrene foam (yes) and now in plain aluminum.
Give a look to those magnificient horns built by a French guy:
http://nsm01.casimages.com/img/2009/04/25//090425111818545253540869.jpg
Best regards from Paris, France
Jean-Michel Le Cléac'h
There have been Le Cléac'h horns built in concrete, in wood, in "papier maché", in fiberglass + resin, in polystrene foam (yes) and now in plain aluminum.
Give a look to those magnificient horns built by a French guy:
http://nsm01.casimages.com/img/2009/04/25//090425111818545253540869.jpg
Best regards from Paris, France
Jean-Michel Le Cléac'h
Hello,
Please note that the small Le Cléac'h horn on left hand replace the original diffraction horn of a TAD ET703 tweeter.
The larger one on right hand was designed for a TAD TD2001 compresssion driver.
Both possess a 360° complete rolled back mouth.
http://nsm01.casimages.com/img/2009/04/25//090425111818545253540869.jpg
Best regards from Paris, France
Jean-Michel Le Cléac'h
Please note that the small Le Cléac'h horn on left hand replace the original diffraction horn of a TAD ET703 tweeter.
The larger one on right hand was designed for a TAD TD2001 compresssion driver.
Both possess a 360° complete rolled back mouth.
http://nsm01.casimages.com/img/2009/04/25//090425111818545253540869.jpg
Best regards from Paris, France
Jean-Michel Le Cléac'h
hello Jean Michel
i have seen these last year at ebay :
http://www.audiovoice-acoustics.com/forum/showthread.php?t=468&highlight=lecleach
these horns are relatively easy to make, ( after the drawing with the dimensions is ready ). Just programming of a CNC lathe takes some time, but ones made that, on the lathe, they are turned fastly.
The chrome process is also not difficult , and fastly done. The only limitation i see is size. Beyond 12" might be difficult to make. Its also not everyones taste, to have such a " eyecatcher " in the livingroom. I personally like it more discrete.
Angelo
i have seen these last year at ebay :
http://www.audiovoice-acoustics.com/forum/showthread.php?t=468&highlight=lecleach
these horns are relatively easy to make, ( after the drawing with the dimensions is ready ). Just programming of a CNC lathe takes some time, but ones made that, on the lathe, they are turned fastly.
The chrome process is also not difficult , and fastly done. The only limitation i see is size. Beyond 12" might be difficult to make. Its also not everyones taste, to have such a " eyecatcher " in the livingroom. I personally like it more discrete.
Angelo
Hello Angelo,
For what it seems there is no chrome plating, those horns are manufactured in aluminum and there is no surface treatment butonly a mirror polishing final step.
(I don't know who made them but could see 3 pairs of them at an audiophile meeting organized by a French forum 2 weeks ago.
I could measure one. Here attached its waterfall
Best regards from Paris, France,
Jean-Michel Le Cléac'h
For what it seems there is no chrome plating, those horns are manufactured in aluminum and there is no surface treatment butonly a mirror polishing final step.
(I don't know who made them but could see 3 pairs of them at an audiophile meeting organized by a French forum 2 weeks ago.
I could measure one. Here attached its waterfall
Best regards from Paris, France,
Jean-Michel Le Cléac'h
Attachments
Wow, smooth waterfall! (down to 2kHz... even smooth below, but not as fast)
What driver is on the horn?
How does the driver measure in another "good horn" by way of comparison? (if such data is available)
_-_-bear
PS. Sweet looking horn... but the speed of sound in aluminum is faster than in air...so I wonder about that aspect, and if the rear side of the horn is damped by anything??
What driver is on the horn?
How does the driver measure in another "good horn" by way of comparison? (if such data is available)
_-_-bear
PS. Sweet looking horn... but the speed of sound in aluminum is faster than in air...so I wonder about that aspect, and if the rear side of the horn is damped by anything??
Limono,
I don't know if you saw already Lecleac'h horn for ET703;
here is a link to my pictures, in desorder
http://picasaweb.google.fr/fredsonqc/Bezon19Avril2009?authkey=Gv1sRgCNfqjo2K_8-5RQ
maybe you need a google/gmail acount...
I did ask a complete horn 1/2" feat specially for ET703:
Size is only 12cm and not much longer than the original one.
Cut off is 1700Hz with T=0,7 (expansion factor);
I did that acoustical cut off because then the complete horne goes back in a nice plan and also to feat well the screws! (not much place,I provide the special screws feat with);
There is a cone coming in the tweeter that is made specially for ET703,I guess we can do different for different super tweeter.
The guy who made that is audio passionate and very well worker:It's not so easy as somebody said because of the complete profile,but he is really busy professionaly;
Also, the purpose is not to get a 2x1700Hz electrical cut off, since as you know ET703 is usually need for a 5kHz at 12dB/octave minimum, but perhaps in private room use(low level) you can trust in a 4kHz 18 or 24dB/oct without breaking your diaphragm!
My advise,If I can, is to use the super tweeter at the higher cut off and 18dB/oct minimum; of course,depend of application you espect to use It.
But, in any case, the improove of the complete horn Lecleac'h profile is great since the acoustical impedance adaptation is much more better! For theory ask Jean-Michel about explanation.
I find this super tweeter just a little smoother and very precise with this horn against a "hard" sound before.
And you need a very silent system to appreciate the improove of to add a super tweeter
FRED fredsonqc @ free . fr
I don't know if you saw already Lecleac'h horn for ET703;
here is a link to my pictures, in desorder
http://picasaweb.google.fr/fredsonqc/Bezon19Avril2009?authkey=Gv1sRgCNfqjo2K_8-5RQ
maybe you need a google/gmail acount...
I did ask a complete horn 1/2" feat specially for ET703:
Size is only 12cm and not much longer than the original one.
Cut off is 1700Hz with T=0,7 (expansion factor);
I did that acoustical cut off because then the complete horne goes back in a nice plan and also to feat well the screws! (not much place,I provide the special screws feat with);
There is a cone coming in the tweeter that is made specially for ET703,I guess we can do different for different super tweeter.
The guy who made that is audio passionate and very well worker:It's not so easy as somebody said because of the complete profile,but he is really busy professionaly;
Also, the purpose is not to get a 2x1700Hz electrical cut off, since as you know ET703 is usually need for a 5kHz at 12dB/octave minimum, but perhaps in private room use(low level) you can trust in a 4kHz 18 or 24dB/oct without breaking your diaphragm!
My advise,If I can, is to use the super tweeter at the higher cut off and 18dB/oct minimum; of course,depend of application you espect to use It.
But, in any case, the improove of the complete horn Lecleac'h profile is great since the acoustical impedance adaptation is much more better! For theory ask Jean-Michel about explanation.
I find this super tweeter just a little smoother and very precise with this horn against a "hard" sound before.
And you need a very silent system to appreciate the improove of to add a super tweeter
FRED fredsonqc @ free . fr
Fred
Thanks so much for pictures and details on the ET-703 horn. Now I can see how it is done . I wondered where the screws are to mount the horn. My guess is that with round horn the afficiency of the tweeter went down from original 107db right? As to the crossover I think in home enviroment we can forgo what manufacturer suggest for professional use. 5kHz at 6dB should be still OK for fraction of a Watt usually needed.
Regards, L
Thanks so much for pictures and details on the ET-703 horn. Now I can see how it is done . I wondered where the screws are to mount the horn. My guess is that with round horn the afficiency of the tweeter went down from original 107db right? As to the crossover I think in home enviroment we can forgo what manufacturer suggest for professional use. 5kHz at 6dB should be still OK for fraction of a Watt usually needed.
Regards, L