Hello Lynn, I just wanted to mention something regarding the DCX. It's a nifty tool for prototyping and whatnot, but the garbage becomes VERY noticable on a high efficiency system. I'm not one of those people who can hear the third chair violinist scratching his nose on recordings, but the hiss that the DCX adds to my HE system really takes the enjoyment out of listening to it. Now I don't have a modified one, that may take care of everything, but those modifications would have to bring that hiss down a LOT.
Lynn,
I have been following this thread with great interest and apreciate all the input and the way you have presented it, excellent work and I may even learn something... just not in this legue tho.
FWIW you may want to read what Thorsten Loech had to say about the Beringer op amps on post 6 on this thread
http://www.diyaudio.com/forums/showthread.php?postid=205180
I have been following this thread with great interest and apreciate all the input and the way you have presented it, excellent work and I may even learn something... just not in this legue tho.
FWIW you may want to read what Thorsten Loech had to say about the Beringer op amps on post 6 on this thread
http://www.diyaudio.com/forums/showthread.php?postid=205180
Dut
Hi
The good thing for test baffles is that you can run tests.
Sometimes the speaker is your device under test sometimes you are the DUT – me myself and I
I have to apologise for my csd – plot in post 797.
What is shows up as a stepped decay there is simply due to low frequency cone excursion out of its linear range.
To make something useful out of my fault for this thread, I run a couple of additional measurements focusing on the limitations in SPL with respect to XO frequency. XO I have chosen to be of the Linkwitz-Riley type at 24 dB.
All measurements were done at zero distance with no EQ applied. Speaker is the same modified Dynaudio 21W54 as before.
To give a taste of real SPL I adjusted the MLS signal to 100 dB A-weighted (300Hz LR-24dB cut).
The measurements compare the test baffle to a heavily damped closed box with 30 liter volume and almost the same front size.
First the 21W54 in the closed box no XO, no EQ smoothed and unsmoothed. The dip at 750Hz is considerably lower than in the OB like predictable from EDGE simulation:
Now the closed box with XO at 100 Hz, 150 Hz and 300 Hz:
It can be seen that the closed box decay of the resonance at 750 Hz smoothes from stepped to linear at around 150 Hz
Now the open baffle at 200 Hz 300 Hz and 400 Hz:
It can be seen that with the open baffle the decay of the resonance at 750 Hz smoothes from stepped to linear at around 300 Hz.
Sure, I could have simulated this with any speaker software as well but its always interesting to have something to visualise.
And living in the Alps I got sort of addicted to that kind of pictures.
Well, it's clouded today and we had fresh snow halfway down last week
Greetings
Michael
Hi
The good thing for test baffles is that you can run tests.
Sometimes the speaker is your device under test sometimes you are the DUT – me myself and I
I have to apologise for my csd – plot in post 797.
What is shows up as a stepped decay there is simply due to low frequency cone excursion out of its linear range.
To make something useful out of my fault for this thread, I run a couple of additional measurements focusing on the limitations in SPL with respect to XO frequency. XO I have chosen to be of the Linkwitz-Riley type at 24 dB.
All measurements were done at zero distance with no EQ applied. Speaker is the same modified Dynaudio 21W54 as before.
To give a taste of real SPL I adjusted the MLS signal to 100 dB A-weighted (300Hz LR-24dB cut).
The measurements compare the test baffle to a heavily damped closed box with 30 liter volume and almost the same front size.
First the 21W54 in the closed box no XO, no EQ smoothed and unsmoothed. The dip at 750Hz is considerably lower than in the OB like predictable from EDGE simulation:
An externally hosted image should be here but it was not working when we last tested it.
An externally hosted image should be here but it was not working when we last tested it.
Now the closed box with XO at 100 Hz, 150 Hz and 300 Hz:
An externally hosted image should be here but it was not working when we last tested it.
An externally hosted image should be here but it was not working when we last tested it.
An externally hosted image should be here but it was not working when we last tested it.
It can be seen that the closed box decay of the resonance at 750 Hz smoothes from stepped to linear at around 150 Hz
Now the open baffle at 200 Hz 300 Hz and 400 Hz:
An externally hosted image should be here but it was not working when we last tested it.
An externally hosted image should be here but it was not working when we last tested it.
An externally hosted image should be here but it was not working when we last tested it.
It can be seen that with the open baffle the decay of the resonance at 750 Hz smoothes from stepped to linear at around 300 Hz.
Sure, I could have simulated this with any speaker software as well but its always interesting to have something to visualise.
And living in the Alps I got sort of addicted to that kind of pictures.
An externally hosted image should be here but it was not working when we last tested it.
Well, it's clouded today and we had fresh snow halfway down last week
Greetings
Michael
Hi
If I understood right this is what you described:
Mic position at centre of WR and halfway the baffle height
Yep, that's really great news!
It should not be related to the EFFICINECY of your system unless the levels are not set properly??
Sad to hear that!
No, just plain and simple power supply biasing.
agree 100%
Greetings
Michael
If I understood right this is what you described:
Mic position at centre of WR and halfway the baffle height
An externally hosted image should be here but it was not working when we last tested it.
An externally hosted image should be here but it was not working when we last tested it.
Yep, that's really great news!
It should not be related to the EFFICINECY of your system unless the levels are not set properly??
Sad to hear that!
No, just plain and simple power supply biasing.
agree 100%
Greetings
Michael
Re: Studio Monitor RAAL Ribbons
No, I see the addition of a new driver as levelling the playing field between compression drivers and the new class of prosound ribbons. There's the Stage Accompany, and now the RAAL joins the field with what looks to me like a higher level of performance. Plus, we have contributors to this thread who have actually auditioned the SA vs the 1-high version of the RAAL, and preferred the RAAL.
I had quite an interesting conversation with Aleksandar of RAAL, and we compared our subjective impressions of horn, ribbon, and cone driver sound. We were comparing our impressions of the dynamic character of each technology - ordinary 87~92 dB direct-radiators start to noticeably compress around 90~95 dB, sound pretty flattened-out at 100 dB, and at levels below 60~70 dB, multiway systems become murky and confused-sounding.
Aleksandar mentioned that audiophile dome tweeters, as a group, have quite different impulse and CSD responses -60 dB below the usual test levels. The rise times slow down, and drivers parameters change. This failure of the drivers in a multiway system to track dynamically - over a wide range - is why complex audiophile systems sound disjointed and muffled at low levels (background music).
I mentioned that horns, in my experience, exaggerate dynamics, sometimes just a little bit, and sometimes very obviously, when they get into a high-order distortion regime with a compression driver approaches its limits. Aleksandar mentioned that the same thing happens at low levels - there's a dynamic-expansion effect that makes quiet sounds even quieter - and this is quite a different effect than the parameter change and "murk" of auidiophile-efficiency direct-radiators at low (microwatt) levels.
The big issue with ribbons, of course, is limited linear travel in the magnetic gap, along with the possibility of ribbon destruction - or stretching - due to over-excursion from excess power. Much of Aleksandar's work has been focussed on addressing high SPL's without the usual compromise of constraining the motion of the aluminum ribbon by gluing it to stretched plastic film, which substantially degrades impulse response and distortion. The issue with all of these technologies - horn, direct-radiator, ribbons, etc. - is retaining dynamically consistent performance at microwatt and 100+ dB levels.
FlorianO said:
No, I see the addition of a new driver as levelling the playing field between compression drivers and the new class of prosound ribbons. There's the Stage Accompany, and now the RAAL joins the field with what looks to me like a higher level of performance. Plus, we have contributors to this thread who have actually auditioned the SA vs the 1-high version of the RAAL, and preferred the RAAL.
I had quite an interesting conversation with Aleksandar of RAAL, and we compared our subjective impressions of horn, ribbon, and cone driver sound. We were comparing our impressions of the dynamic character of each technology - ordinary 87~92 dB direct-radiators start to noticeably compress around 90~95 dB, sound pretty flattened-out at 100 dB, and at levels below 60~70 dB, multiway systems become murky and confused-sounding.
Aleksandar mentioned that audiophile dome tweeters, as a group, have quite different impulse and CSD responses -60 dB below the usual test levels. The rise times slow down, and drivers parameters change. This failure of the drivers in a multiway system to track dynamically - over a wide range - is why complex audiophile systems sound disjointed and muffled at low levels (background music).
I mentioned that horns, in my experience, exaggerate dynamics, sometimes just a little bit, and sometimes very obviously, when they get into a high-order distortion regime with a compression driver approaches its limits. Aleksandar mentioned that the same thing happens at low levels - there's a dynamic-expansion effect that makes quiet sounds even quieter - and this is quite a different effect than the parameter change and "murk" of auidiophile-efficiency direct-radiators at low (microwatt) levels.
The big issue with ribbons, of course, is limited linear travel in the magnetic gap, along with the possibility of ribbon destruction - or stretching - due to over-excursion from excess power. Much of Aleksandar's work has been focussed on addressing high SPL's without the usual compromise of constraining the motion of the aluminum ribbon by gluing it to stretched plastic film, which substantially degrades impulse response and distortion. The issue with all of these technologies - horn, direct-radiator, ribbons, etc. - is retaining dynamically consistent performance at microwatt and 100+ dB levels.
Very interesting observations re: dynamics. I have heard the argument before wrt high efficiency drivers (exaggerated dynamics) but can't quite imagine a mechanism that would lead to that.
Driver parameters most certainly change with drive levels but that should be happening to any kind of mechanical construction.
Driver parameters most certainly change with drive levels but that should be happening to any kind of mechanical construction.
MBK said:
Beats me too, but that's what I hear, and I've heard plenty of live acoustic music. I've noticed that phase-plug geometry has an effect on this, so I'm surmising that turbulence effects are involved - no other form of audio reproducer has such extremely small internal acoustic surfaces and transmission paths.
Another interesting topic that arose from the discussion is that horn theory relies on a true plane wave entering the horn, but in practice, the wavefront from a cone driver, or the exit from a CD phase plug, isn't very close to a plane wave. The horn doesn't magically "straighten out" the wavefront - if anything, it gets more disordered, not less, as it makes its way through the horn expansion.
I have heard a horn that was driven with a nearly perfect plane wave. The horn was a large circular Tractrix with a 3" throat, driven by a single JansZen electrostatic element. It basically sounded like you'd expect, a REALLY BIG electrostat with lots of headroom and midbass presence. In tonal terms, no horn coloration at all, but it also had some of the physically forward, in-the-room spatial perspective of large-format horns - quite different than the faraway perspective of a typical large electrostatic panel.
Aleksandar at RAAL had similar experiences with his experimental horn-loaded ribbons - the horns sounded really good, but not the same as the open ribbons, with the most of the difference being a different sense of perspective on the music. That was my take on the electrostat-horn as well - the difference between transported to the musicians, versus the musicians being in the room. Much of this is due to polar-pattern differences illuminating the local environment, but I also suspect the ear is directly sensitive to the shape of the wavefront itself. I suspect a highly curved spherical wavefront is interpreted as coming from a location nearby the listener, and as the listener moves further off-axis, the smooth curvature becomes more disrupted.
In a later conversation with John Atwood, I was mentioning how the large-format Tractrix and Le Cleac'h horns have 90-degree dispersion in the technical sense - they sound good, FR response seems all there, etc. - but the real 3D quality and hyper-vivid tonality extends to about one seat-width. So in effect, the horns have 90 degrees of high-quality sound, but only 10 degrees of super-quality sound. If you're sitting where you can see the entire surface of the driver without obstruction, that's the super-quality zone. (In my experience, at least.) In a really dialled-in top-quality horn system, the best sound is either directly on-axis - looking right smack into the drivers - or just ever so slightly off-axis, depending on how the system was set up. When you sit far enough off-axis so the drivers are no longer visible and disappear behind the curve of the horn, there's a loss of focus and precision, compared to the on-axis (or close to on-axis) listening experience.
JA and I were wondering whether this is an artifact of the rough wavefront entering the horn - it emerges reasonably intact in a narrow beam (where the horn isn't doing much), but gets more disordered further off-axis, resulting in time dispersion and small-scale irregularities in the polar pattern. Cone drivers seem to be the reverse of this - they commonly have highly directional colorations that fall in a narrow 5~10-degree window, and sound smoother, more natural, and less "megaphone-like" slightly off-axis. It's part of the reason I'm considering a 5 to 10-degree slope on the open baffle, with a separate aiming assembly for the tweeter.
It would take an interesting measurement protocol showing impulse response vs off-axis, a sort of 3D waterfall. I think JVC once did this in an early 1980's AES paper, making some interesting graphs that looked like waves breaking on a beachfront. The ribbon/horn had much nicer looking off-axis impulse response than the usual CD/horn, with the mundane dome tweeter looking best, as expected.
P.S. About the Edge simulations - you're getting close, just use a greater rake angle, maybe 200 to 250 mm of overhang, larger drivers (315 mm), even spacing between the 6 drivers, and measure at 3 and 12 meters.
The low level resolution thing is interesting. Have not heard it talked about in years. It used to be a test for the resolving power of a system. If it held together at low levels, that was a good sign. Funny that many systems don't - but I never knew why.
Maybe this is the real reason that Capt. Kirk was alway bugging Scotty to push the Starship Enterprise to her limits. Kirk wasn't in any real hurry, he just thought the ship sounded better running flat out. Kind of like a Marshall amp stack.
Maybe this is the real reason that Capt. Kirk was alway bugging Scotty to push the Starship Enterprise to her limits. Kirk wasn't in any real hurry, he just thought the ship sounded better running flat out. Kind of like a Marshall amp stack.
Re: Dissing the DCX
Hello Lynn,
I have been reading this tread with great enjoyment. Very interesting and informative.
Maybe I am jumping little bit ahead but I think it might help some. Are you familiar ( I am sure you are) with Mr. LeCleach's approach to selecting the crossover points for drivers? Here is the link:
http://www.diyaudio.com/forums/showthread.php?postid=875687#post875687
I would like to know what is your opinion for this method. The document included is based on using the Behringer DCX digital crossover, but I believe it could be done in passive manner as well.
Shortly, by his method of selecting the crossover point for drivers through the Butterworth 3rd order for tweeter is shifted up by aprox. 1.4 x crossover point Fq. and crossover point for mid driver is lowered by aprox. 0.87 x crossover point Fq.
This could be very beneficial for ribbon drivers by bringing their cut off frequency up higher from the region where they are the most problematic.
I am very happy with this arangement in my set up with Aurum Cantus 3" ribbons and 101dB 10"PHL mids.
I still have to hear - see RAAL ribbons, hopefully this summer, which seems to be the real winner by many accounts and it feels like to be your tweeter choice. They seem to be much improved ribbons in handling the low fq. region with low distortion.
Hello Lynn,
I have been reading this tread with great enjoyment. Very interesting and informative.
Maybe I am jumping little bit ahead but I think it might help some. Are you familiar ( I am sure you are) with Mr. LeCleach's approach to selecting the crossover points for drivers? Here is the link:
http://www.diyaudio.com/forums/showthread.php?postid=875687#post875687
I would like to know what is your opinion for this method. The document included is based on using the Behringer DCX digital crossover, but I believe it could be done in passive manner as well.
Shortly, by his method of selecting the crossover point for drivers through the Butterworth 3rd order for tweeter is shifted up by aprox. 1.4 x crossover point Fq. and crossover point for mid driver is lowered by aprox. 0.87 x crossover point Fq.
This could be very beneficial for ribbon drivers by bringing their cut off frequency up higher from the region where they are the most problematic.
I am very happy with this arangement in my set up with Aurum Cantus 3" ribbons and 101dB 10"PHL mids.
I still have to hear - see RAAL ribbons, hopefully this summer, which seems to be the real winner by many accounts and it feels like to be your tweeter choice. They seem to be much improved ribbons in handling the low fq. region with low distortion.
Lynn,
even with a somehow irregular wavefront, I'd think we are dealing with wavefront shape aberrations at most in the cm range, far smaller than the most relevant wavelengths of audio. I would rather blame Geddes' "higher order modes" for the troubles of narrower horns off-axis, in other words, multiple internal reflections. Interestingly LeCleac'h in his worksheet gives the option between planar or spherical wave, but with a presumably irregular source this may then not even matter. If wavefront shape at the mouth does matter indeed, then a smaller exit CD would be better than a larger one (more point source like).
RE: turbulence I had the idea of using an irregular horn inside surface - either the golf ball pattern, used for turbulence control in the golf ball as well, or concentric rings, or in the case of a wooden horn, irregular ripples using sandblasting. Obviously that would throw off the whole precise horn contour calculations, but hey maybe this matters more than the contour? In any case trying this out would involve a lot of trial and error.
Re: spatial properties, forward or deep placement - FWIW in my dipoles at least, the delay setting for the tweeter can significantly alter the deep or forward pattern. If set correctly, the image is in a 3D position behind the speakers. If the delay is off, the very same speaker and x-o point can sound "forward". So this may have something to do with the arrival time of the HF wrt the LF.
BTW - an electrostatic horn, now that's exotic!
even with a somehow irregular wavefront, I'd think we are dealing with wavefront shape aberrations at most in the cm range, far smaller than the most relevant wavelengths of audio. I would rather blame Geddes' "higher order modes" for the troubles of narrower horns off-axis, in other words, multiple internal reflections. Interestingly LeCleac'h in his worksheet gives the option between planar or spherical wave, but with a presumably irregular source this may then not even matter. If wavefront shape at the mouth does matter indeed, then a smaller exit CD would be better than a larger one (more point source like).
RE: turbulence I had the idea of using an irregular horn inside surface - either the golf ball pattern, used for turbulence control in the golf ball as well, or concentric rings, or in the case of a wooden horn, irregular ripples using sandblasting. Obviously that would throw off the whole precise horn contour calculations, but hey maybe this matters more than the contour? In any case trying this out would involve a lot of trial and error.
Re: spatial properties, forward or deep placement - FWIW in my dipoles at least, the delay setting for the tweeter can significantly alter the deep or forward pattern. If set correctly, the image is in a 3D position behind the speakers. If the delay is off, the very same speaker and x-o point can sound "forward". So this may have something to do with the arrival time of the HF wrt the LF.
BTW - an electrostatic horn, now that's exotic!
Lynn,
Somewhere in the last couple pages I lost track of the working plan. What do you intend to cross the RAAL ribbon over to? The problem I forsee with crossing a larger format mid to a ribbon in the 2.5K region is the widely different polar response in the xo region, or do you think this doesn't matter?
I am just following along curiously to see where this goes...
Somewhere in the last couple pages I lost track of the working plan.
What do you intend to cross the RAAL ribbon over to? The problem I forsee with crossing a larger format mid to a ribbon in the 2.5K region is the widely different polar response in the xo region, or do you think this doesn't matter?I am just following along curiously to see where this goes...
Definitely OT, but I thought some of you would like to see some inspired DIY "enclosure" designs for OB speakers (see "dipole subwoofer"):
http://www.audiocircle.com/gallery/index.php?cat=11046
Oh, and while I'm off topic, you might like to see this Lyngdorf/Steinway dipole just announced in Stereophile:
http://www.stereophile.com/news/052107steinway/
http://www.audiocircle.com/gallery/index.php?cat=11046
Oh, and while I'm off topic, you might like to see this Lyngdorf/Steinway dipole just announced in Stereophile:
http://www.stereophile.com/news/052107steinway/
I hope this is not considered too off topic, but the configuration of this speaker is so similar to what we are discussing that I thought it appropriate to inform everyone.
I see Emerald Physics has changed the design of their show model significantly with their first production model.
It is instructive to note which design elements of the CS1 prototype they chose to retain.
Now it is a dual 15" two way, with the 15's crossed at 900 Hz to a 1" compression driver in a 12" waveguide. Sensitivity is 100 dB/W/M.
http://www.emeraldphysics.com/products.html
I see Emerald Physics has changed the design of their show model significantly with their first production model.
It is instructive to note which design elements of the CS1 prototype they chose to retain.
Now it is a dual 15" two way, with the 15's crossed at 900 Hz to a 1" compression driver in a 12" waveguide. Sensitivity is 100 dB/W/M.
http://www.emeraldphysics.com/products.html
JoshK said:Lynn,
Somewhere in the last couple pages I lost track of the working plan.
I am just following along curiously to see where this goes...
As mentioned before, polar pattern is about #4 or #5 on my list of technical design requirements - desirable, nice to have, but sound quality and smoothness of crossover come far ahead. I feel that the modern emphasis on polar-pattern control is borrowed from the PA and theater-sound world, which has a completely different set of requirements than music playback in the home. The most important characteristics by far, as I see it, are natural timbre, natural dynamics, and a vivid, true-to-life quality. These parameters are controlled by the selection of drivers, and the success of system integration (good crossover design).
After natural-sounding timbre is achieved, or at least in sight as a reasonable goal (I feel most high-end speakers fail in this area), then good dynamic tracking at very quiet, medium, and high levels is desirable. This is even more rare in high-end speakers, which usually are unlistenable at background-music levels. Although horn systems can have serious problems with timbre, they excel in dynamic tracking, certainly compared to almost all audiophile systems, with the expensive, complex ones being the worst by far.
So my primary goal is getting these two main things right - timbre and dynamics. Polar pattern is nice, but way less important than the first two things, which are about primary musical qualities, and represent a serious and difficult technical challenge. Just getting timbre right is no small feat, considering how rare this is - just compare almost any loudspeaker to a live symphony. We all have a ways to go here.
For all of the patient readers who have persisted thus far, you can probably see the new speaker (still thinking of a name, probably out of the Greek pantheon) is going to be not one, but a family of designs. For the HF portion, I can see two quite reasonable and potentially good-sounding variations: a 1.4" or 2" BMS or Radian compression driver with a 300~500 Hz Le Cleac'h profile horn and 1.2~2 kHz crossover, or a twin (or triple!) RAAL tweeter version with a crossover around 1.8~2.2 kHz, depending on where the widerange driver wants to be crossed.
I plan to build the system both ways, with the final choice pretty much a subjective preference. Both versions will have the tweeter assembly on a central mounting post that offers fine-tuning adjustments in the front-to-back plane, as well as up-and-down and left-and-right tilt adjustments. As mentioned elsewhere, this is critical for getting the subjective impression of depth and perspective right, and will offer a fair comparison between CD and ribbon tweeters. The interchangeable platform - which I haven't figured out yet - seems like the only good way to fine-tune the tweeter to the room, and provide precise arrival-time adjustments for the crossover.
These are the kind of precision adjustments that should be part of any serious high-end speaker, and rarely are. I've heard enough high-resolution systems in Colorado and Switzerland to know that when you're working with the highest-performance drivers, fine-tuning becomes very important, just like it is on a quality tone-arm.
panomaniac said:
A test I've used for many years is really simple: playing favorite pieces of music at background-music levels (50 dB or so), chatting with friends and listening casually. With real musicians nearby, this is a very pleasant diversion that sweetens the whole environment, and a charming historical tradition going back to the Renaissance.
With most high-end speakers I've heard, you end up being annoyed at the background noise, and have to turn off the "big" system, because of its intrusive qualities. With the simplest and cheapest single-driver speakers, though, background music sounds pleasant - not as sweet as live music, but at least not intrusive and annoying.
This simple and casual test is actually quite revealing - it tells you a great deal about system coherence and dynamic tracking. If a system only "sounds good" or "sounds hi-fi" at SPLs of 70 dB and above - well, you have a problem. This is part of the reason I've steered clear of 3 and 4-way systems - getting dynamic tracking right with systems this complex is very, very difficult.
MBK said:
I think a smaller exit throat could easily be worse, not better. A rough wavefront is going to be scrambled more going through the greater expansion in a high-ratio horn - I think of optical analogies here, with laser light going through lumpy, irregular-surface optics (the phase plug), and then magnified further. If the throat is half the size, well, the expansion from throat to mouth is twice as big, magnifying the initial abberations twice as much.
Considering that ALL horn equations are based on underlying assumptions of perfectly spherical or planar wavefronts entering the horn - which is no more than an acoustic lens, or waveguide - the real-world departure from perfect spherical or planar wavefronts is quite significant. Optical designers don't dismiss departures from plane behaviour, so why do horn designers? I think this is one of those "simplifying assumptions" that make horns easier to design but have real-world consequences.
Once you remove the typical 1/3 octave smoothing - and even Geddes does this, based on the published curves - there are many very narrow notches in the response curves of horns, which indicate small-scale irregularities in the wavefront coming out of the horn. This is swept under the rug, but I think it is significant in terms of what we hear. I also suspect this is the real reason that we hear a small region of exceptional performance in the central 10 degrees of arc - well, the horn isn't doing very much there! The sound in that region is flying right through the horn, straight from the diaphragm or phase-plug, right to the listener. Note how much subjective resolution is lost when you sit 45 degrees off-axis - it doesn't sound the same as on-axis at all, even though it may measure fairly similarly.
Actually, I really like what horns do well - the fantastic dynamics, vivid tone colors, and what I like most, the way they sound when played quietly, where all the details are kept, instead of disappearing into the murk, as with ordinary audiophile drivers. But this business of irregular wavefronts, as well as the much larger-scale HOMs at the bottom of the working range, need to be addressed, analyzed, and solved, not swept under the rug.
I heard the JansZen/Tractrix horn system when I visited Hawaii in the mid-Nineties as a guest of Hiroshi Ito. Sadly, both Hiroshi and the creator of the quite remarkable horn system have passed on, so I have no further connections except for the memory. Perhaps somebody in the Honolulu Audio Society might remember the people involved.
The system itself was quite idiosyncratic. Two really big red fiberglass Tractrix horns, probably with at least 3-foot mouth diameters, with a single JansZen tweeter element clearly visible thorugh the throat. The amplification was equally bizarre, a 1-watt MOSFET run as a single-ended amplifier. The system didn't play super-loud, but it was the amp, not the electrostat, that limited things. Very big and spacious sounding, NOTHING like a Lowther or AER in a similar-sized Oris horn, and essentially no "horn" coloration in the usual sense. I'm pretty sure the electrostat got credit for that. I'd expect HOM's would be lower as well, since the electrostat, unlike a compression driver or Lowther, is almost completely acoustically transparent, thus no way for standing waves to be set up between the horn-mouth and hard reflective phase-plugs or Lowther magnet assembly.