FrankWW said:
I will try from memory
Horn length is NOT important
Throath and smooth transistion to "hornpath" is important
Mouth rounding is important
Mouth size determines lowest FR
Other than that I dont recall any other information
Does it look like this ? ... well, throath need reworking, but its hard to draw
Attachments
macka said:
I'm not in complete agreement here, but mostly.
I think that the total power response SHAPE is important as this sets the reverberant field color and the direct anhd reverberant fields need to match. Being smooth is necessary but not sufficient to good sound quality. The direct and reverberant fields need to track each other in total and not just over a listening window.
People do nat pay enough attention to polar response period. The vertical as well as the horizontal. But the horizontal is the more important just as horizontal reflections in a room are the more important.
Correct - no single horn or waveguide can do much more than about a decade - I've said that before.
As to the equivalence of Jean-Michel's horns and my waveguides, as I said they are different animals and will certainly sound different. Which one any particular listener may like is a matter of preference, but I've done horns like his and I know what they sound like, and well, I use waveguides.
Hi Earl
“Correct - no single horn or waveguide can do much more than about a decade - I've said that before.”
Perhaps that’s true for a conventional horn, but the Synergy horns fall into a different category.
The SH-50 for example is one horn but can re-produce a waveshape like a square wave over a decade and of course has much wider response than that.
While these are not “hifi” speakers per say, they seem to sound ok to hifi / horn people.
http://www.audioasylum.com/forums/hug/messages/13/134216.html
Best,
Tom Danley
“Correct - no single horn or waveguide can do much more than about a decade - I've said that before.”
Perhaps that’s true for a conventional horn, but the Synergy horns fall into a different category.
The SH-50 for example is one horn but can re-produce a waveshape like a square wave over a decade and of course has much wider response than that.
While these are not “hifi” speakers per say, they seem to sound ok to hifi / horn people.
http://www.audioasylum.com/forums/hug/messages/13/134216.html
Best,
Tom Danley
Tom
I don't question that your approach does extend the usual useful range of a typical horn, especially with the caveat that they are not Hi-Fi. My comments were all aimed at Hi-Fi class sound quality not pro applications where power and output reign.
If we are comparing quotes here are a few from some DIY listeners.
I don't question that your approach does extend the usual useful range of a typical horn, especially with the caveat that they are not Hi-Fi. My comments were all aimed at Hi-Fi class sound quality not pro applications where power and output reign.
If we are comparing quotes here are a few from some DIY listeners.
Hi Earl
I am not contesting that your speakers sound good, my point was that there is a way to get much more bandwidth from one horn. Also, if done properly, it can be indistinguishable in energy vs time or radiated pattern from a single tiny but very strong wide band driver in a large CD horn
Though the SH-50’s do not have a “home” appearance or shape, size, boutique construction or drivers, hifi marketing or pricing structure, it is certainly not the sound that makes them “not hifi”.
The link i added was a comment from a fellow who holds a speaker tasting every year, "Oswald's Mill" and has been listening to the SH-50's, nearly all of the feedback we get is from pro- and commercial sound and wouldn't apply as much.
What I see is that in many ways, accurate reproduction becomes vastly more difficult as the room size, distance to the audience and number of people in the listening area increases.
That is the job they were designed for “large scale hifi”, they just happen to work well loafing along in a living room too and the approach may allow an exception to the decade rule.
Best,
Tom
I am not contesting that your speakers sound good, my point was that there is a way to get much more bandwidth from one horn. Also, if done properly, it can be indistinguishable in energy vs time or radiated pattern from a single tiny but very strong wide band driver in a large CD horn
Though the SH-50’s do not have a “home” appearance or shape, size, boutique construction or drivers, hifi marketing or pricing structure, it is certainly not the sound that makes them “not hifi”.
The link i added was a comment from a fellow who holds a speaker tasting every year, "Oswald's Mill" and has been listening to the SH-50's, nearly all of the feedback we get is from pro- and commercial sound and wouldn't apply as much.
What I see is that in many ways, accurate reproduction becomes vastly more difficult as the room size, distance to the audience and number of people in the listening area increases.
That is the job they were designed for “large scale hifi”, they just happen to work well loafing along in a living room too and the approach may allow an exception to the decade rule.
Best,
Tom
Hey Tom -
I do agree that with room size and SPL comes a whole array of difficulties that require different compromises to solve. And as I have said on numerous occasions your approach is optimized for its target and mine for mine.
Where I continue to be a little less convinced is with words like "Also, if done properly, it can be indistinguishable in energy vs time or radiated pattern from a single tiny but very strong wide band driver in a large CD horn". Perhaps there are CD horns for which your approach would be indistinguishable, but I would challenge that it would be indistinguishable from my approach. Admitedly there is no data either way, except for some subjective data by a few people who have heard both.
I continue to want to see good polars and frequency response plots of both approaches which are comparable. None of the data that I have seen is comparable as the conditions are vastly different. And I'm not saying that one is right and the other wrong its just that they are different. For example I have seen your data for a 15TBX100 and it doesn't look like mine - I don't know who's is the more correct, only that they are different. Yours does look better than mine, thats for sure.
From a purely theoretical point of view your designs should have problems with HOM and mouth reflections.
I'd love to measure both systems under identical circumstances, but I don't see that happening soon. So we'll have to agree to disagree.
Thanks for the posts.
I do agree that with room size and SPL comes a whole array of difficulties that require different compromises to solve. And as I have said on numerous occasions your approach is optimized for its target and mine for mine.
Where I continue to be a little less convinced is with words like "Also, if done properly, it can be indistinguishable in energy vs time or radiated pattern from a single tiny but very strong wide band driver in a large CD horn". Perhaps there are CD horns for which your approach would be indistinguishable, but I would challenge that it would be indistinguishable from my approach. Admitedly there is no data either way, except for some subjective data by a few people who have heard both.
I continue to want to see good polars and frequency response plots of both approaches which are comparable. None of the data that I have seen is comparable as the conditions are vastly different. And I'm not saying that one is right and the other wrong its just that they are different. For example I have seen your data for a 15TBX100 and it doesn't look like mine - I don't know who's is the more correct, only that they are different. Yours does look better than mine, thats for sure.
From a purely theoretical point of view your designs should have problems with HOM and mouth reflections.
I'd love to measure both systems under identical circumstances, but I don't see that happening soon. So we'll have to agree to disagree.
Thanks for the posts.
Looking at the data for the SH50, I would think higher frequency extension would be desireable for home use. Wonder what the CSD and impulse would look like.
Hi Earl
I continue to want to see good polar’s and frequency response plots of both approaches which are comparable. None of the data that I have seen is comparable as the conditions are vastly different. And I'm not saying that one is right and the other wrong its just that they are different. For example I have seen your data for a 15TBX100 and it doesn't look like mine - I don't know who's is the more correct, only that they are different. Yours does look better than mine, that’s for sure.
Keep in mind we are sort of stuck where the customers demand EASE data and CLF files of the measurements and the requirements in gathering that spherical data are beyond what I can do.
Mag /Phase and Polar’s etc sure but not 5 degree points on a sphere at a large enough distance.
Fwiw, the 15TBX100 driver curve I had posted once was for that driver in a Tapped horn, not operating as a direct radiator and as I recall, it was four close coupled cabinets where the response is uncannily flat and were measured outdoors away from buildings etc..
From a purely theoretical point of view your designs should have problems with HOM and mouth reflections.
Perhaps so but I suspect it is also partly a “shades of gray” issue.
For instance, I have found that the driver I use in that box is particularly well behaved on that kind of horn (conical / pyramidal) as its origin (a conical horn within the driver) is only about 3 / 8 inch dia and does not use “rings of summation” like most drivers which can interfere or cause HOM’s..
While it may not present the horn entrance with a classical plane wave, the diverging nature is I think better for this use as it begins life as an acoustically small point source at the apex of the horn..
I'd love to measure both systems under identical circumstances, but I don't see that happening soon. So we'll have to agree to disagree.
Agreed;
Tom
I’m thinking about another BBQ later this summer / fall, would you be up for a drive?
Hi soonqsc
Apparently they are adding stuff on the web site at the moment but if you go later and download the white paper of the Synergy horn, there is a measured impulse response there.
The hf response for this driver does extend past 20KHz however the TEF is (in these measurements) was set as 20Khz for the upper limit and the absence of data past that makes either end of the response droop a bit at each band edge. This can be eliminated if one measures past the display limits but that was not done in this case.
Best,
Tom Danley
I continue to want to see good polar’s and frequency response plots of both approaches which are comparable. None of the data that I have seen is comparable as the conditions are vastly different. And I'm not saying that one is right and the other wrong its just that they are different. For example I have seen your data for a 15TBX100 and it doesn't look like mine - I don't know who's is the more correct, only that they are different. Yours does look better than mine, that’s for sure.
Keep in mind we are sort of stuck where the customers demand EASE data and CLF files of the measurements and the requirements in gathering that spherical data are beyond what I can do.
Mag /Phase and Polar’s etc sure but not 5 degree points on a sphere at a large enough distance.
Fwiw, the 15TBX100 driver curve I had posted once was for that driver in a Tapped horn, not operating as a direct radiator and as I recall, it was four close coupled cabinets where the response is uncannily flat and were measured outdoors away from buildings etc..
From a purely theoretical point of view your designs should have problems with HOM and mouth reflections.
Perhaps so but I suspect it is also partly a “shades of gray” issue.
For instance, I have found that the driver I use in that box is particularly well behaved on that kind of horn (conical / pyramidal) as its origin (a conical horn within the driver) is only about 3 / 8 inch dia and does not use “rings of summation” like most drivers which can interfere or cause HOM’s..
While it may not present the horn entrance with a classical plane wave, the diverging nature is I think better for this use as it begins life as an acoustically small point source at the apex of the horn..
I'd love to measure both systems under identical circumstances, but I don't see that happening soon. So we'll have to agree to disagree.
Agreed;
Tom
I’m thinking about another BBQ later this summer / fall, would you be up for a drive?
Hi soonqsc
Apparently they are adding stuff on the web site at the moment but if you go later and download the white paper of the Synergy horn, there is a measured impulse response there.
The hf response for this driver does extend past 20KHz however the TEF is (in these measurements) was set as 20Khz for the upper limit and the absence of data past that makes either end of the response droop a bit at each band edge. This can be eliminated if one measures past the display limits but that was not done in this case.
Best,
Tom Danley
Tom Danley said:
Tom
I was thinking more about the drivers on the sides of the horn and the effect that they would have on the wavefront - HOM's can be created along the horn as well as at the throat. The Unity's that I have seen have very sharp mouth terminations. Everything is "shades of grey". Some grays work better in some situations than other, thats all.
Let me know about the Bar-B-Q, I'd certainly come if I could.
And if you are ever near here you should come and hear my theater.
Tom Danley said:
Tom,
I have a pair of Unity's from Nick at Lambda (who is back in the woofer game again). I have TAD 2001's on them now, but was curious if drivers like you use (I recall you mentioned BMS at one time) would work better, for the reasons you cite. I made the entrance holes for the mids smaller, as John Hancock suggested at one time, so reflections there should be reduced. I also got some foam from Dr. Geddes. To my ears, it does remove any remaining slight trace of horn character.
I have a DEQX so I can do a wide variety of crossovers. I have played a little, crossing the 15" mid directly to the horn at about 1K. It sounds good, but adding the mids and crossing sounds a little better to me. The difference may just be a matter of taste, as I don't hear anything wrong either way, and I could live with either.
Sheldon
tinitus said:
My point was: there is no big secret.
Pages @24 onward posts by joshk, AJ, Earl Geddes, and KSTR.
There were posts earlier in the thread about importance of getting the match between the driver throat angle and wave guide mouth angle correct.
There were earlier posts with other spreadsheets.
audiokinesis said:
Last night I had the Summas playing, and walked over by the amp to change tracks. I was about one foot away from the right speaker, and noticed that there was still a soundstage - even though the left speaker was clear on the other side of the room.
I've never seen another speaker that does that.
My hypothesis is that this is due to the exceptionally large waveguide on the compression driver, and all the attention paid to termination and diffraction reduction.
Tom Danley said:
Tom, did you ever see the project I did a few years back, putting a Unity clone in my car? I took the best of both worlds - the OS waveguide and foam plug, combined with the unity design.
Unity Horn in a Honda Accord FTW!
I'm taking another crack at it, since I made a fatal mistake when I did it two years ago (the midranges are tuned wrong - didn't realize this for a year.)
Unity Horn in a Honda Accord - 2nd Try
Hello Jzagaja,
The minphase curve and the group delay curve you obtained on the pulse response I measured are similar to the ones I obtained with my own Matlab routine.
I presented such group delay curve I mesured on an older Jerzual 320 Le Cléac'h horn in the conference about "phase distortion" I gave at ETF in 2004 (too bad the document is no more downoadable on the website http://www.triodefestival.net )
But the same group delay curve can be seen in page 46 of the document of the conference in French I gave in March 2005
at the French association Melaudia: http://www.melaudia.net/zdoc/distorsion_de_phase.pdf
About your question :
let me allow some (French) humour.
After you published here few CSD graphs better for the Le Cléac'h horns than for other horns and waveguide, someone wrote that CSD is useless, then now you published a very smooth min phase curve for the Le CLéac'h horn with a perfectly flat group delay curve above 1kHz, so be sure that someone will say that minphase plots are useless too...
Thanks again for the analysis.
Best regards from Paris, France
Jean-Michel Le Cléac'h
The minphase curve and the group delay curve you obtained on the pulse response I measured are similar to the ones I obtained with my own Matlab routine.
I presented such group delay curve I mesured on an older Jerzual 320 Le Cléac'h horn in the conference about "phase distortion" I gave at ETF in 2004 (too bad the document is no more downoadable on the website http://www.triodefestival.net )
But the same group delay curve can be seen in page 46 of the document of the conference in French I gave in March 2005
at the French association Melaudia: http://www.melaudia.net/zdoc/distorsion_de_phase.pdf
About your question :
jzagaja said:
let me allow some (French) humour.
After you published here few CSD graphs better for the Le Cléac'h horns than for other horns and waveguide, someone wrote that CSD is useless, then now you published a very smooth min phase curve for the Le CLéac'h horn with a perfectly flat group delay curve above 1kHz, so be sure that someone will say that minphase plots are useless too...
Thanks again for the analysis.
Best regards from Paris, France
Jean-Michel Le Cléac'h
Dear Jean,
I was told by the experts that horns aren't minphase devices and cannot be corrected with minphase filters. Well I've tried them (digital) with good results. Does it mean that minphase component cannot be used for an ebaluation - I don't know.
I was told by the experts that horns aren't minphase devices and cannot be corrected with minphase filters. Well I've tried them (digital) with good results. Does it mean that minphase component cannot be used for an ebaluation - I don't know.
Hello,
Most horns create reflected waves from mouth to throat. If you see the pulse response of such horns (+ loudspeaker) you see small pulses delayed from the main pulse by the time for the wave to go from mouth to throat and then come back from throat to mouth (= more or less the double of the time for the sound to move along a distance equal to the double of the length of the horn.)
How can you imagine that some equalization can make disapppear those delayed pulses? It is impossible and this explain than most horns cannot be considered as minimum phase device.
There is no such reflected waves in a "complete" Le Cléac'h horn which one behaves more accurately as a minimum phase device (at least above 1kHz if Fc = 320Hz)!
Reflected waves may appear as tails parallel to the time axis at different frequencies on CSD graphs. You can see those tails on the CSD graph of the JBL horn (and of the waveguide eventually), you can see only some noise but no such tails, on the CSD graph of the Le Cléac'h horn. Also some elements on the group delay curves may be interpreted in the case of the JBL and the waveguide as due to reflected waves from mouth to throat...
When the Le Cléac'h horn is mounted on a compression driver the impedance of which obeys to a simple equivalent model ( this means: without parasitic phenomenons leading to delayed energy release), then it is benefitial to feed the driver by a current source. You cannot do that with bad drivers. The result is that the response curve of the compression driver mounted on the Le Cléac'h horn becomes independant on the impedance curve .
Give a look to the attached graph:
On left (my measurements), the blue curve shows the very smooth impedance curve of the TD2001 when mounted on a J321 Le Cléac'h horn. In red you see the response curve of the TD2001 monted on the J321 horn when source impedance is very very low (voltage control).
Please note:
Every shape element on the response curve (for F >450Hz where the horn behaves like a nearly pure resistive load) possess its "inverted" equivalent on the impedance curve (this is not a coincidence, this has a meaning!).
On right: according to Ken Kinoshita (in JAES, 1978), the hole in the response curve of the TD2001 when voltage driven is due to the cavity beneath the coil. The resonance in that cavity is translated on the impedance curve as a peak in the frequency interval 1000Hz - 2000Hz. If you use a voltage source to feed the driver the current through the coil will then be reduced in that frequency interval and there will be a dip in the response curve (the famous 1600Hz or 1900Hz hole in the curve of the TD2001). A current drive will avoid this problem (this is not a simple equalization as one could eventually say...) .
Best regards from Paris, France
Jean-Michel Le Cléac'h
Most horns create reflected waves from mouth to throat. If you see the pulse response of such horns (+ loudspeaker) you see small pulses delayed from the main pulse by the time for the wave to go from mouth to throat and then come back from throat to mouth (= more or less the double of the time for the sound to move along a distance equal to the double of the length of the horn.)
How can you imagine that some equalization can make disapppear those delayed pulses? It is impossible and this explain than most horns cannot be considered as minimum phase device.
There is no such reflected waves in a "complete" Le Cléac'h horn which one behaves more accurately as a minimum phase device (at least above 1kHz if Fc = 320Hz)!
Reflected waves may appear as tails parallel to the time axis at different frequencies on CSD graphs. You can see those tails on the CSD graph of the JBL horn (and of the waveguide eventually), you can see only some noise but no such tails, on the CSD graph of the Le Cléac'h horn. Also some elements on the group delay curves may be interpreted in the case of the JBL and the waveguide as due to reflected waves from mouth to throat...
When the Le Cléac'h horn is mounted on a compression driver the impedance of which obeys to a simple equivalent model ( this means: without parasitic phenomenons leading to delayed energy release), then it is benefitial to feed the driver by a current source. You cannot do that with bad drivers. The result is that the response curve of the compression driver mounted on the Le Cléac'h horn becomes independant on the impedance curve .
Give a look to the attached graph:
On left (my measurements), the blue curve shows the very smooth impedance curve of the TD2001 when mounted on a J321 Le Cléac'h horn. In red you see the response curve of the TD2001 monted on the J321 horn when source impedance is very very low (voltage control).
Please note:
Every shape element on the response curve (for F >450Hz where the horn behaves like a nearly pure resistive load) possess its "inverted" equivalent on the impedance curve (this is not a coincidence, this has a meaning!).
On right: according to Ken Kinoshita (in JAES, 1978), the hole in the response curve of the TD2001 when voltage driven is due to the cavity beneath the coil. The resonance in that cavity is translated on the impedance curve as a peak in the frequency interval 1000Hz - 2000Hz. If you use a voltage source to feed the driver the current through the coil will then be reduced in that frequency interval and there will be a dip in the response curve (the famous 1600Hz or 1900Hz hole in the curve of the TD2001). A current drive will avoid this problem (this is not a simple equalization as one could eventually say...) .
Best regards from Paris, France
Jean-Michel Le Cléac'h
jzagaja said:
Attachments
Jmmlc said:
And some American humor:
Sounds like your response to constant directivity control in your horns.
🙂
- Status
- Not open for further replies.