Hi Tom, good to see your contribution here!
Hmm - regarding conical horns, I've heard the Cogent system twice now (at the RMAF each time), and the field-coil had intriguing tonal properties - surprisingly different than Alnico, and different again from ceramic. Very subtle and delicate tone colors, quite interesting.
I assume Bill Woods designed the conical mid-horn on the Cogent, and I regret to report that both times, I heard a fair amount - more than I can personally accept - of what I consider horn coloration. Other people sitting right next to me heard none, and were really turned-on by the sound. I really like the Cogent team, and am quite curious what their technically intriguing drivers would sound like with a different horn profile.
I can only guess perception of this particular sonic characteristic is highly individualized, much like audibility of DHT/triode vs EL34/pentode vs MOSFET/transistor amplifiers. Long story short, I've never heard a conical I've liked, which I admit makes me somewhat prejudiced against them. I'm sure such an animal could exist - I just haven't heard one yet, that's all.
Among the horns that I've heard that seemed to have reasonably low coloration, I'd list Dr. Geddes, LeCleac'h, Tractrix, and large-format multicells. No Altec sectorals, JBL things with wavy lenses on them, or JBL/Altec theatre/PA constant-directivity horns with kinks in the profile. I've yet to hear any horn or waveguide with no horn coloration, unfortunately, but then again, all loudspeakers have pretty obvious colorations of one kind or another. Ultimately, it all comes down to what set of colorations we can personally accept - and not being intimidated by the opinions and preferences of reviewers or industry "experts".
I don't want the high-efficiency dipole speaker to get endlessly sidetracked by the search for a perfect horn, so I'm leaving the options open while the horn design community continues their research. Nor do I want to spend the next 5 years becoming a horn expert. I'm relying on others to move the art forward - it's beyond the scope of what I want to attempt right now.
I'm very much looking forward to seeing more time-domain data, especially the 0~2 mSec decay characteristics compared to direct-radiator equivalents - I'm guessing that's what I'm responding to when I grumble about "horn coloration".
Looking forward to carrying out experiments just like what Tom suggested, and many more, in the January/February/March timeframe as my personal mobility improves and I start measuring and listening downstairs. Breaking the system into 3 stacked modules, one for each part of the spectrum, should greatly simplify decision-making and make assembly and portability more straightforward.
Hmm - regarding conical horns, I've heard the Cogent system twice now (at the RMAF each time), and the field-coil had intriguing tonal properties - surprisingly different than Alnico, and different again from ceramic. Very subtle and delicate tone colors, quite interesting.
I assume Bill Woods designed the conical mid-horn on the Cogent, and I regret to report that both times, I heard a fair amount - more than I can personally accept - of what I consider horn coloration. Other people sitting right next to me heard none, and were really turned-on by the sound. I really like the Cogent team, and am quite curious what their technically intriguing drivers would sound like with a different horn profile.
I can only guess perception of this particular sonic characteristic is highly individualized, much like audibility of DHT/triode vs EL34/pentode vs MOSFET/transistor amplifiers. Long story short, I've never heard a conical I've liked, which I admit makes me somewhat prejudiced against them. I'm sure such an animal could exist - I just haven't heard one yet, that's all.
Among the horns that I've heard that seemed to have reasonably low coloration, I'd list Dr. Geddes, LeCleac'h, Tractrix, and large-format multicells. No Altec sectorals, JBL things with wavy lenses on them, or JBL/Altec theatre/PA constant-directivity horns with kinks in the profile. I've yet to hear any horn or waveguide with no horn coloration, unfortunately, but then again, all loudspeakers have pretty obvious colorations of one kind or another. Ultimately, it all comes down to what set of colorations we can personally accept - and not being intimidated by the opinions and preferences of reviewers or industry "experts".
I don't want the high-efficiency dipole speaker to get endlessly sidetracked by the search for a perfect horn, so I'm leaving the options open while the horn design community continues their research. Nor do I want to spend the next 5 years becoming a horn expert. I'm relying on others to move the art forward - it's beyond the scope of what I want to attempt right now.
I'm very much looking forward to seeing more time-domain data, especially the 0~2 mSec decay characteristics compared to direct-radiator equivalents - I'm guessing that's what I'm responding to when I grumble about "horn coloration".
Looking forward to carrying out experiments just like what Tom suggested, and many more, in the January/February/March timeframe as my personal mobility improves and I start measuring and listening downstairs. Breaking the system into 3 stacked modules, one for each part of the spectrum, should greatly simplify decision-making and make assembly and portability more straightforward.
Hi Lynn,
If you are sensitive to horn coloration you might try a modified LeCleach contour.
I have been in contact with John Hasquin, who is one of the most experienced people when it comes to round horns. He has build and evaluated many tractrix and LeCleach horns.
Here is his reply to me about his modified contour ( I did not ask him for permission to post his reply to me, but since he already posted this information on the triode mafia forum , I take the liberty to post it here):
The T factor (for the LeCleach horn) that I ended up with was 1.3415
It comes from my equation of ((Sqrt((Sqrt(2))^(1/pi))-1) + 1)
Sounds much more open than the normal tractrix. However, does not do the lower frequencies as well as the tractrix. There are always trade offs.
Rgs, John
If you are sensitive to horn coloration you might try a modified LeCleach contour.
I have been in contact with John Hasquin, who is one of the most experienced people when it comes to round horns. He has build and evaluated many tractrix and LeCleach horns.
Here is his reply to me about his modified contour ( I did not ask him for permission to post his reply to me, but since he already posted this information on the triode mafia forum , I take the liberty to post it here):
The T factor (for the LeCleach horn) that I ended up with was 1.3415
It comes from my equation of ((Sqrt((Sqrt(2))^(1/pi))-1) + 1)
Sounds much more open than the normal tractrix. However, does not do the lower frequencies as well as the tractrix. There are always trade offs.
Rgs, John
Radian said:
That's what I've heard too. In subjective terms, the LeCleac'h seems to have substantially lower diffraction and time-domain "clutter" than other profiles. I think that was probably one of the design goals, since these things don't happen by accident - you have to make them a design priority. Having heard the same drivers on a Tractrix and LeCleac'h, I'd use exactly the same words as John Hasquin - the LeCleac'h is much more open and spacious-sounding than the Tractrix, but also just a little thinner-sounding as well - but not by much.
This is partly why I want a 12 or 15-inch driver that can confidently meet a LeCleac'h horn at 600~800 Hz - with good-to-superb tonality, generous dynamics, and very low distortion (to complement the horn upstairs). The array of drivers that come in below the baffle peak are more about linear volume displacement than anything, since the drivers are well within the piston range from 160 Hz on down.
By contrast, the single driver that goes up to meet the horn is going to be working with the range where Le nonlinearities, pole-piece construction, and choice of magnet material make a big difference to the overall sound of the system. Even though the 12 or 15-inch driver is still within the piston-band, tone-color variations are pretty noticeable and obvious. This is why the Altec/GPA 416 Alnico and the 12-inch Alnico Tone Tubby are in the running - both of these drivers are known for "good tone", and I've heard them for myself and agree with the general consensus.
It is not an accident the two drivers most responsible for the overall system "tone color" are the single 12~15" driver and the mid horn, while the ribbon working above 5~7 kHz and the drivers working below 160 Hz have less color.
Hello,
This makes a lot of sense, however I don't understand how is it possible to match the output of the horn with the wide range 12'' driver. Reading Magnetar's posts it seems that the "magic" number would be around 103dB/1W/1M. The horn will do it, the bass array can do it but I wonder if just one 12" in the 160Hz-600Hz region can deliver this kind of dynamics.
Why not just a 3 way with a very high quality bass array crossed to the horn at around 400-500Hz where the problems asociated with these kind of arrays are not problematic. This way the dynamics capabilities of the horn and direct radiators are more uniform...
I hope this makes some sens...
This makes a lot of sense, however I don't understand how is it possible to match the output of the horn with the wide range 12'' driver. Reading Magnetar's posts it seems that the "magic" number would be around 103dB/1W/1M. The horn will do it, the bass array can do it but I wonder if just one 12" in the 160Hz-600Hz region can deliver this kind of dynamics.
Why not just a 3 way with a very high quality bass array crossed to the horn at around 400-500Hz where the problems asociated with these kind of arrays are not problematic. This way the dynamics capabilities of the horn and direct radiators are more uniform...
I hope this makes some sens...
Tom Danley said:
I have two questions if you are still watching this thread.
1) Why the specific 60º recommendation? Is that 60º from side wall to side wall (as opposed to on axis bisector to side wall)?
2) I don't quite grasp all the diffraction implication, can you explain a bit more about what the mouth roundover does? My naive thinking would make it seem that no roundover would produce "diffraction" that would add horn colorations. Why also the >=1.4 x factor?
When I looked at the Cogent/AH conical, I was wishing for a big roundover like Geddes has. I heard Geddes' Summas and didn't pick up on any horn colorations.
Dear Lynn,
Nobel prize for someone who can show me diffraction part from total speaker/horn response and model it. At which level it occurs?
Does my tweeter in Tractrix horn exhibit diffraction or/and time clutter?
http://www.diyaudio.com/forums/showthread.php?postid=1370902#post1370902
I've decided to use shelving LR4 active crossovers for my 400Hz Tractrix and 105dB 38cm woofer.
Merry Christmas
--
Jack
Nobel prize for someone who can show me diffraction part from total speaker/horn response and model it. At which level it occurs?
Does my tweeter in Tractrix horn exhibit diffraction or/and time clutter?
http://www.diyaudio.com/forums/showthread.php?postid=1370902#post1370902
I've decided to use shelving LR4 active crossovers for my 400Hz Tractrix and 105dB 38cm woofer.
Merry Christmas
--
Jack
Attachments
Re: XBaffle sim
Zene
Plug parameters in WinISD, UniBox or whatever using sealed enclosure volume of 10 x Vd or greater to simulate an IB and look at slope. Edge's best speaker placement shows -3dB at 100hz. Even common sense indicates this speaker will not get to -3dB of 40hz in an OB without considerable boost. As long as it's used as UB I have no problem and it does look good. Thx for the find. I'm a P.audio fan. Just did not want anyone to think this speaker will do bass.mikey_audiogeek said:
Zene
Hi Lynn
Bill Woods was a horn designer who worked at Yorkville for many years and is a great guy. His love was always hifi and when he left the company he set up his own business making horns.
http://www.acoustichorn.com/
He had worked at Yorkville back when the Unity horn products were designed and built there under license, the Unity horn was the precursor to the Synergy Horns we sell at work in commercial sound. Anyway, the Unity horns in the U-15 and U215 also used a basically conical horn and the two are very similar.
Thinking back, I would have to say the most difficult part dealing with conical horns is that even a perfect CD horn, perfectly driven still reveals the drivers acoustic power response which is not flat.
Making the response flat into the complex electrical load the driver represents is doable but requires a spice like approach. Hence, text book filters for horn crossovers normally have some big irregularities loaded with real drivers.
Often what was thought to be an acoustic resonance turns out to be the drivers impedance curve interacting, producing peaks and dips etc.
To see if it will do what you want, an active approach is the fastest I think, then you just deal with the response and not the load.
The down side of horns and all large sources (all point sources) is that unless two drivers are about 1 / 3 wl or less apart, they don’t add coherently and then have cancellations and lobes.
This points to a type of coloration that I have not heard with a name but it is something that gives your ears que’s as to where in space, how far away the speaker is.
I mean when the speaker is producing a voice or music a short distance away and you close your eyes it is usually easy to guess how far away it is
On the speaker I do at work, as I got the time between ranges down and made the dimensions correct so it acted like one driver, the ability to tell how far away the speaker was largely went away.
In stereo, this effect on imaging was hugely in the right direction, now all there was were Que’s from the recording and a mono signal hangs in midspace.
Anyway I am sure there is some type of “self interference” lets say that you can hear when you have multiple sources which are too far apart, that identifies the position of the loudspeaker as a source in space instead of reproducing the signal as the source.
Another informal place I think this occurs is in floor monitors (on the floor in front of singers). I was able to make a particular coax driver on a short horn preserve time / not interfere the same way.
The weird thing is without that source interference, singers have the monitors way quieter than before, its like they hear more of what they need to sing, I don’t know if that makes sense?
Pdan, JoshK
In the early days of CD horns which had straight walls, it was found that these maintained a fixed radiation angle down to some frequency, near pattern loss, where the coverage suddenly got narrower before widening out. Don Keele figured out a rule of thumb for fixing this, involving making the last third of the horn a larger angle to off set the narrowing as the portion of the horn governing pattern reaches the mouth end.
In practice, this end detail is not that critical and it would seem that increasing the total angle from 1.3 to 1.5 times the original angle, works well when you measure the radiation at every angle (spherical measure).
The round over between the two angles isn’t specified and horns can work fine with a sharp break in angle
However, like proper looking speaker wire, intellectually it should be a radius, in reality (measuring where the sound goes), as the mouth size increases, one finds less and less energy up high is radiated out of pattern angle. While a nice radius looks good and philosophically is “right”, on larger conical horns, the sound waves seem immune to that logic and largely ignores those mouth details.. I would round it over because it looks right, might help and couldn't hurt.
I suggested 60 degrees as a safe horn angle as the larger horn wall angles and larger exit diameters tend to have increasing issues with sound "bouncing around" inside up high.
Picture this, that ideally, the horn apex extends down to where its diameter is about ¼ across at the highest frequency. In this case the sound will propagate outward as if this were a part of a spherical section. Sounds easy but how big is ¼ at say 20KHz?, Way less than the one inch exit so what goes on inside the driver can effect directivity significantly in the upper octaves.
The BMS driver for example has a conical horn that begins at about 3/8 inch dia and does not involve multiple path lengths which could add improperly.
Anyway, I am hopelessly biased towards the horns I do for a living but that is what I use in my home too and I could never go back .
Have fun, Happy Holidays
Best,
Tom Danley
It’s a White Christmas for sure around here, 14 degrees F. now, blowing snow, a fresh coat of ice on the sled ramp.
Bill Woods was a horn designer who worked at Yorkville for many years and is a great guy. His love was always hifi and when he left the company he set up his own business making horns.
http://www.acoustichorn.com/
He had worked at Yorkville back when the Unity horn products were designed and built there under license, the Unity horn was the precursor to the Synergy Horns we sell at work in commercial sound. Anyway, the Unity horns in the U-15 and U215 also used a basically conical horn and the two are very similar.
Thinking back, I would have to say the most difficult part dealing with conical horns is that even a perfect CD horn, perfectly driven still reveals the drivers acoustic power response which is not flat.
Making the response flat into the complex electrical load the driver represents is doable but requires a spice like approach. Hence, text book filters for horn crossovers normally have some big irregularities loaded with real drivers.
Often what was thought to be an acoustic resonance turns out to be the drivers impedance curve interacting, producing peaks and dips etc.
To see if it will do what you want, an active approach is the fastest I think, then you just deal with the response and not the load.
The down side of horns and all large sources (all point sources) is that unless two drivers are about 1 / 3 wl or less apart, they don’t add coherently and then have cancellations and lobes.
This points to a type of coloration that I have not heard with a name but it is something that gives your ears que’s as to where in space, how far away the speaker is.
I mean when the speaker is producing a voice or music a short distance away and you close your eyes it is usually easy to guess how far away it is
On the speaker I do at work, as I got the time between ranges down and made the dimensions correct so it acted like one driver, the ability to tell how far away the speaker was largely went away.
In stereo, this effect on imaging was hugely in the right direction, now all there was were Que’s from the recording and a mono signal hangs in midspace.
Anyway I am sure there is some type of “self interference” lets say that you can hear when you have multiple sources which are too far apart, that identifies the position of the loudspeaker as a source in space instead of reproducing the signal as the source.
Another informal place I think this occurs is in floor monitors (on the floor in front of singers). I was able to make a particular coax driver on a short horn preserve time / not interfere the same way.
The weird thing is without that source interference, singers have the monitors way quieter than before, its like they hear more of what they need to sing, I don’t know if that makes sense?
Pdan, JoshK
In the early days of CD horns which had straight walls, it was found that these maintained a fixed radiation angle down to some frequency, near pattern loss, where the coverage suddenly got narrower before widening out. Don Keele figured out a rule of thumb for fixing this, involving making the last third of the horn a larger angle to off set the narrowing as the portion of the horn governing pattern reaches the mouth end.
In practice, this end detail is not that critical and it would seem that increasing the total angle from 1.3 to 1.5 times the original angle, works well when you measure the radiation at every angle (spherical measure).
The round over between the two angles isn’t specified and horns can work fine with a sharp break in angle
However, like proper looking speaker wire, intellectually it should be a radius, in reality (measuring where the sound goes), as the mouth size increases, one finds less and less energy up high is radiated out of pattern angle. While a nice radius looks good and philosophically is “right”, on larger conical horns, the sound waves seem immune to that logic and largely ignores those mouth details.. I would round it over because it looks right, might help and couldn't hurt.
I suggested 60 degrees as a safe horn angle as the larger horn wall angles and larger exit diameters tend to have increasing issues with sound "bouncing around" inside up high.
Picture this, that ideally, the horn apex extends down to where its diameter is about ¼ across at the highest frequency. In this case the sound will propagate outward as if this were a part of a spherical section. Sounds easy but how big is ¼ at say 20KHz?, Way less than the one inch exit so what goes on inside the driver can effect directivity significantly in the upper octaves.
The BMS driver for example has a conical horn that begins at about 3/8 inch dia and does not involve multiple path lengths which could add improperly.
Anyway, I am hopelessly biased towards the horns I do for a living but that is what I use in my home too and I could never go back .
Have fun, Happy Holidays
Best,
Tom Danley
It’s a White Christmas for sure around here, 14 degrees F. now, blowing snow, a fresh coat of ice on the sled ramp.
Merry Christmas to all !
http://members.aon.at/kinotechnik/diyaudio/dipol/IM_speaker-disto.xls
PS:
Comments and corrections welcome.
Greetings
Michael
http://members.aon.at/kinotechnik/diyaudio/dipol/IM_speaker-disto.xls
PS:
Comments and corrections welcome.
Greetings
Michael
Thank you very much Tom!
How about something like this:www.google.com/patents?vid=USPAT6334505 ?
Cilla
How about something like this:www.google.com/patents?vid=USPAT6334505 ?
Cilla
Merry Christmas season greetings to all on this thread (my favorite) from cool and wet Victoria, BC, on the southern tip of Vancouver Island.
The dry snow of Colorado does appeal!
Although it is 40º F (4ºC) here and raining lightly, about a hundred miles up the island we do have Mt Washington which currently has a snow base of 7 1/2 feet.
The dry snow of Colorado does appeal!
Although it is 40º F (4ºC) here and raining lightly, about a hundred miles up the island we do have Mt Washington which currently has a snow base of 7 1/2 feet.
Radian said:
Well, shovelling 4 to 6 inches of snow off the driveway gets old after a while - especially when Karna, who is all of 5' 2" high, is doing all the work. It took her about 2 hours to get it done today, while I watched from inside. I'm still not fit enough to do heavy work outdoors, and walking on snow, ice, or a complicated mix of the two takes my full attention combined with a cane. I think we'll be buying a Honda snow-blower to deal with this stuff.
Returning to the audio theme, there's a good link over in the Dr. Geddes thread describing the GuangZhou (Canton) Hi-Fi Show. Most interesting.
Having grown up in Hong Kong from 1961 to 1967, it's interesting seeing high-end audio spreading into mainland China. GuangZhou (Canton), Hong Kong, and Macau are about 60 miles from each other - Canton is at the head of the Pearl River delta, and Macau and Hong Kong at the ends, on either side. When I was there, they had fast hydrofoil service between Macau and HK Harbour, with Canton and China off-limits to all but highly placed politicians on state visits - in fact, the mid-Sixties was the height of the Cultural Revolution, and China was in turmoil. The situation is very different now, with both Macau and HK in Chinese hands, and China itself much more open to the world.
I'm not surprised the Pearl River Delta has become an industrial powerhouse. When I was in Hong Kong, and on a followup visit in 1971, I was surprised by the mix of low-tech plastic toys (GI Joe dolls) and high-tech manufacturing (magnetic core memory used by NASA for the Apollo program, and the first desktop calculator in the world to use the Intel 4004 processor). It was obvious they were going to give the Japanese, Germans, and the USA a very serious run for the money on being the world's most advanced industrial-manufacturing area.
Since I grew up in Kobe/Osaka (1956~61) and Hong Kong (1961~67) and saw all kinds of things being made in both places, I've always been interested in manufacturing (working at Audionics and Tektronix). Small details in manufacturing - not usually disclosed in product sales literature or obvious in measurements - is what makes the difference between indifferent, good, and really superb products.
I changed plans. and just hiked where I could see the road.Multiple tweeters
Gentlemen,
Would someone either point me to a thread about multiple tweeters or provide an answer as to why I don't see more speaker systems with mutiple tweeters.
With the beaming of tweeters as a problem, I wonder why Bozak's solution hasn't been used more. The least he used was a pair of tweeters on about a 15 degree spread to either side of center. I always considered his bank of tweeters assembled in a hemispherical array as a grand, if not grandiose, solution.
I find that a second tweeter facing away from the listener
(bipolar?) helps to fill in the upper register.
With the availability of low cost piezo tweeters, would an array be worth trying or would the sound quality suffer?
dobias
Gentlemen,
Would someone either point me to a thread about multiple tweeters or provide an answer as to why I don't see more speaker systems with mutiple tweeters.
With the beaming of tweeters as a problem, I wonder why Bozak's solution hasn't been used more. The least he used was a pair of tweeters on about a 15 degree spread to either side of center. I always considered his bank of tweeters assembled in a hemispherical array as a grand, if not grandiose, solution.
I find that a second tweeter facing away from the listener
(bipolar?) helps to fill in the upper register.
With the availability of low cost piezo tweeters, would an array be worth trying or would the sound quality suffer?
dobias
