I use a de250. Its a great driver that I know well and have (mostly) had great success with. There are much more expensive drivers, but in all my testing I have not found any to be any better.
Back to beaming as not being a problem, it is if you toe-in your loudspeakers to widen the sweet spot. With a beaming device you have only one option and one seat in the sweet spot. OK for some, not OK with me. (There are other issues with beaming in reverberant rooms as well.)
It is simply not correct to say that constant directivity does not matter. Only a small few claim this, mostly because their designs are not CD. Kind of coincidence.
Back to beaming as not being a problem, it is if you toe-in your loudspeakers to widen the sweet spot. With a beaming device you have only one option and one seat in the sweet spot. OK for some, not OK with me. (There are other issues with beaming in reverberant rooms as well.)
It is simply not correct to say that constant directivity does not matter. Only a small few claim this, mostly because their designs are not CD. Kind of coincidence.
Yes, even when horn loaded, diaphragm excursion rises as frequency drops. And there's not a lot of room in a compression driver. So I don't like using them at low midrange frequencies (like under 400Hz). They're not made for that.
We were talking about this on another thread, but we got sidetracked I guess.
As I see it, we have a handful of competing priorities in horn design.
One is to optimize acoustic loading, which provides smoother response and lower distortion. A second priority is the ability to create a desired beamwidth, to be able to create a radiation pattern that covers the intended listening area. A third priority is to optimize uniformity of beamwidth, which provides equal coverage throughout the listening area and a more spectrally balanced reverberent field. And a fourth is to optimize wavefront propogation, which improves both response smoothness and beamwidth uniformity.
You can see evidence of fractured wavefronts both in the polars and often even as amplitude response ripple on any single axis. Discontinuities in a horn almost always present as ripples in acoustic impedance, which also show up in the amplitude response. Sometimes in a highly fractured wavefront, the interactions are so dense that the polars seem relatively uniform but this is usually limited to cases like lenses. Single diffraction slots usually manifest themselves in the polars and amplitude response. In addition to all these measurable features, subjectively, it has been suggested that fractured wavefront propogation also damages suble acoustic clues, degrading imaging and lessening the illusion of acoustic reality.
One thing I've seen though, is a predisposition of some designers to optimize one trait so much they lose sight of others. I've never understood, for example, the choise to employ a horn with physical characteristics that make it impossible to design a loudspeaker with widely spaced vertical nulls. If the vertical nulls form less than about 20° apart (10° above and below the forward centerline), then I personally think the speaker is pretty worthless because the nulls are huge and audible, and they're basically straight in front of the loudspeaker. No mater how good the speaker is in other aspects, this one flaw is a deal breaker, because it is huge.
On the other hand, I've seen other designs that seek to reduce or even eliminate lobing, which I think is an excellent goal, but the strategies employed usually either require a highly compromised horn design, or they have a lot of internal reflections, or both. These kinds of approaches are maybe good in prosound, where coverage and arrayability are extremely important, but for home hifi - where arrayability isn't needed - I think it is better to use other optimizations that don't fracture the wavefront so badly.
I've also wondered why some of the newer waveguide designers have almost completely ignored acoustic loading, bringing the polarization of the horn versus waveguide methods. I realize that conical horns and the closely related waveguides have characteristically poor acoustic loading, particularly at low frequencies. But to me, this makes it all the more important that the modern waveguide designer do some modeling and testing of their device to ensure it provides smooth response.
Horn designers are able to enjoy smooth response, nearly ruler flat. I'm not just talking about the collapsing pattern providing on-axis rising response that counters faling response from mass rolloff. I'm talking about horn loading throughout the passband, from lower cutoff through midband and up. A conical horn or waveguide will not be able to match the low end smoothness of an exponential horn, for example. But it can be made close, if done properly, and it still can provide constant directivity. To me, this is an important balance to strike.
As I see it, we have a handful of competing priorities in horn design.
One is to optimize acoustic loading, which provides smoother response and lower distortion. A second priority is the ability to create a desired beamwidth, to be able to create a radiation pattern that covers the intended listening area. A third priority is to optimize uniformity of beamwidth, which provides equal coverage throughout the listening area and a more spectrally balanced reverberent field. And a fourth is to optimize wavefront propogation, which improves both response smoothness and beamwidth uniformity.
You can see evidence of fractured wavefronts both in the polars and often even as amplitude response ripple on any single axis. Discontinuities in a horn almost always present as ripples in acoustic impedance, which also show up in the amplitude response. Sometimes in a highly fractured wavefront, the interactions are so dense that the polars seem relatively uniform but this is usually limited to cases like lenses. Single diffraction slots usually manifest themselves in the polars and amplitude response. In addition to all these measurable features, subjectively, it has been suggested that fractured wavefront propogation also damages suble acoustic clues, degrading imaging and lessening the illusion of acoustic reality.
One thing I've seen though, is a predisposition of some designers to optimize one trait so much they lose sight of others. I've never understood, for example, the choise to employ a horn with physical characteristics that make it impossible to design a loudspeaker with widely spaced vertical nulls. If the vertical nulls form less than about 20° apart (10° above and below the forward centerline), then I personally think the speaker is pretty worthless because the nulls are huge and audible, and they're basically straight in front of the loudspeaker. No mater how good the speaker is in other aspects, this one flaw is a deal breaker, because it is huge.
On the other hand, I've seen other designs that seek to reduce or even eliminate lobing, which I think is an excellent goal, but the strategies employed usually either require a highly compromised horn design, or they have a lot of internal reflections, or both. These kinds of approaches are maybe good in prosound, where coverage and arrayability are extremely important, but for home hifi - where arrayability isn't needed - I think it is better to use other optimizations that don't fracture the wavefront so badly.
I've also wondered why some of the newer waveguide designers have almost completely ignored acoustic loading, bringing the polarization of the horn versus waveguide methods. I realize that conical horns and the closely related waveguides have characteristically poor acoustic loading, particularly at low frequencies. But to me, this makes it all the more important that the modern waveguide designer do some modeling and testing of their device to ensure it provides smooth response.
Horn designers are able to enjoy smooth response, nearly ruler flat. I'm not just talking about the collapsing pattern providing on-axis rising response that counters faling response from mass rolloff. I'm talking about horn loading throughout the passband, from lower cutoff through midband and up. A conical horn or waveguide will not be able to match the low end smoothness of an exponential horn, for example. But it can be made close, if done properly, and it still can provide constant directivity. To me, this is an important balance to strike.
I am looking for a replacement for a Goto replica that plays down to 200Hz. Seems I will only be needing 350Hz. Was the JMLC-200T good enough to keep?
Wayne,
Your take on uniform directivity is strictly from horn perspective. Your designs illuminate one side wall and rear wall. Virtual mono source located between speakers would illuminate both side walls if it were real. True uniform directivity is 360 degrees.
Exploring true 360 degree coverage reveals misnomers about early reflections. With 360 degree coverage, room response uniformly equals speaker direct response. Imaging is detailed and remains so in much wider and deeper listening area. Such speakers can be placed in room or against boundaries and retain spectral balance. No room treatments or changes in equalization are needed.
Sensitivity of 90 degree waveguides/horns to room placement and toe in and the highly limited sweet spot show failure to uniformly cover listening area.
Your take on uniform directivity is strictly from horn perspective. Your designs illuminate one side wall and rear wall. Virtual mono source located between speakers would illuminate both side walls if it were real. True uniform directivity is 360 degrees.
Exploring true 360 degree coverage reveals misnomers about early reflections. With 360 degree coverage, room response uniformly equals speaker direct response. Imaging is detailed and remains so in much wider and deeper listening area. Such speakers can be placed in room or against boundaries and retain spectral balance. No room treatments or changes in equalization are needed.
Sensitivity of 90 degree waveguides/horns to room placement and toe in and the highly limited sweet spot show failure to uniformly cover listening area.
I would agree that DI-matched two-way speakers definitely have the problems you're describing. I recommend flanking subs or some kind of helper woofer blended with the mains to mitigate those problems. But that approach is a compromise done to allow a sort of "generic loudspeaker" to be built, one that works in most rooms.
A constant directivity cornerhorn, on the other hand, is a no-compromise solution that has none of those problems. In a sense, it is an omnidirectional source placed at the apex of the room's corner. So there is no self-interference from the "back wall" or the ipsilateral "side wall" because the sound source lies upon those boundaries.
As for the waveguide/horns used to build these loudspeakers, to me, the holy grail is a horn that provides constant directivity with the smoothness of a an exponential or LeCleach horn.
What I've found works very well is an OS/PS/EC device, any of which have the same flare profile as an OS waveguide/horn. But it is very important that the area expansion be chosen that provides smooth response. This can be done with mouth dimensions, e.g. aspect ratio and/or beamwidth. It should also have some mouth roundover, using something like an iterative expansion technique just like LeCleach uses. Not so much that the device has to sacrifice too much of its OS/PS/EC flare though.
Many if not most of the common 90° OS and EOS horns are too shallow for their mouth area, and response is lumpy as a result. Common wisdom is the larger the mouth, the better the response, but this overlooks the flare profile. In the case of the OS/PS/EC flare used to create these devices - especially those with secondary flares or mouth radiusing - the overall dimensions have a great influence on response smoothness. This is demonstrated in the response chart comparison on the first post of this thread.
Again, many OS and EOS horns have 5dB or greater ripple, which the loudspeaker designer seeks to mitigate with notch filters. To me, that approach is not much better than the 1970s CD horns. Diffraction is better, but the internal standing waves are horrendous. What they have is a highly resonant horn.
So again, to me, the holy grail is smoothness of an exponential or LeCleach and the directivity of a OS, PS or EC flare. I have found this can be done with an OS/EC waveguide/horn with the right choices of area expansion and aspect ratio.
A constant directivity cornerhorn, on the other hand, is a no-compromise solution that has none of those problems. In a sense, it is an omnidirectional source placed at the apex of the room's corner. So there is no self-interference from the "back wall" or the ipsilateral "side wall" because the sound source lies upon those boundaries.
As for the waveguide/horns used to build these loudspeakers, to me, the holy grail is a horn that provides constant directivity with the smoothness of a an exponential or LeCleach horn.
What I've found works very well is an OS/PS/EC device, any of which have the same flare profile as an OS waveguide/horn. But it is very important that the area expansion be chosen that provides smooth response. This can be done with mouth dimensions, e.g. aspect ratio and/or beamwidth. It should also have some mouth roundover, using something like an iterative expansion technique just like LeCleach uses. Not so much that the device has to sacrifice too much of its OS/PS/EC flare though.
Many if not most of the common 90° OS and EOS horns are too shallow for their mouth area, and response is lumpy as a result. Common wisdom is the larger the mouth, the better the response, but this overlooks the flare profile. In the case of the OS/PS/EC flare used to create these devices - especially those with secondary flares or mouth radiusing - the overall dimensions have a great influence on response smoothness. This is demonstrated in the response chart comparison on the first post of this thread.
Again, many OS and EOS horns have 5dB or greater ripple, which the loudspeaker designer seeks to mitigate with notch filters. To me, that approach is not much better than the 1970s CD horns. Diffraction is better, but the internal standing waves are horrendous. What they have is a highly resonant horn.
So again, to me, the holy grail is smoothness of an exponential or LeCleach and the directivity of a OS, PS or EC flare. I have found this can be done with an OS/EC waveguide/horn with the right choices of area expansion and aspect ratio.
With "virtual" you mean a real source instead of a phantom source? Such a source also wouldn't illuminate the room the way a stereo configuration with two omni speakers would. How does that fit in your (Linkwitz's?) hypothesis about reflection perception?
Last edited:
Omni is the converse of the corner horn. One involves the room as much as possible the same as the other avoids it. Both can do their respective jobs well.
A reflected image from a freestanding speaker can be distinguished by its direction and delay, but this shouldn't happen with a corner horn. A point source in a corner is intended to produce idealised reflected images which occur in the same location, time and direction as the original..almost as if the wall weren't there.
I agree with your characterization. A constant directivity cornerhorn avoids any interaction from the nearest boundaries. It neatly sidesteps the issue of a reflected image because there isn't one. A constant directivity cornerhorn provides a textbook case of eighth-space radiation, one where the sound source is located at the apex of the boundaries.
As if two of six walls weren't there, or three of six with a point source at the junction of three walls, or four of six with a line source.
Well that's true, absolutely. One still benefits from multisubs for room modes. And the really late reflections from the most distant walls are still there too.
But the problematic early reflections - the ones that really mess things up - are nonexistant. They're not attenuated, not mitigated - they're simply not there.
When using a constant directivity cornerhorn configuration, the nearest boundaries are all essentially part of a trihedral ground plane.
But the problematic early reflections - the ones that really mess things up - are nonexistant. They're not attenuated, not mitigated - they're simply not there.
When using a constant directivity cornerhorn configuration, the nearest boundaries are all essentially part of a trihedral ground plane.
Sorry Wayne, but in practical situations there will be early reflections (and diffractions) in mid-high frequencies because the mid and tweeter are not at the apex . You have said that earlier yourself!
Actually you are using the bass in the same manner as a downfiring bass driver in a many subwoofers, but in 1/4 instead of 1/2 space!
Can we even say that "semilate" reflections are not there?
A Pi Speakers Corner Horn
Actually you are using the bass in the same manner as a downfiring bass driver in a many subwoofers, but in 1/4 instead of 1/2 space!
Can we even say that "semilate" reflections are not there?
A Pi Speakers Corner Horn
Last edited:
your not suppose to say that!
Well, sure, there are varying degrees of quality, as with anything else. And then there is the matter of how much is audible, and all that jazz.
The best implementation would be where the sound sources are truly flush with the boundaries. The worst implementation is one where the sound sources are several wavelengths away from the boundaries, all the way through the passband. I think it's safe to assume that there is a range where the acoustic distance is close enough to provide benefit.
But one thing is for sure - If the sound source is further than a couple wavelengths from a boundary, then it acts as a reflector, and self-interference will result. If radiation is omnidirectional, the the sound directed at the boundary is as loud as the sound directed at the listener, so self-interference is severe.
In cases where the sound source is acoustically close at the low end of its passband, but becomes acoustically distant as frequency rises, then it will have no self-interference down low, but it may at higher frequencies. However, if the sound source becomes directional at higher frequencies, that helps attenuate the self-interference by the nature of directivity. I would suggest that is an improvement over a non-directional source.
What I see is typical speakers with direct radiating midwoofers tend to be placed a few feet away from walls. This is just the nature of the typical home installation. What this does is to create a worst-case scenario with respect to self-interference. The interference notches form in the mid hundred Hertz range, where the midwoofers are radiating omnidirectionally. So the response aberration is very strong.
If you can move the woofer and/or midrange drivers close enough to the apex of the corner where this doesn't happen, it's a huge benefit, in my opinion. It sure sounds a lot different, and measures cleaner too. Then if the horn(s) provide directivity at higher frequencies, that helps even more.
But yeah, it would be even better to flush mount the drivers into the walls. Of course, that kind of permanent installation precludes the possibility of movement, but it is a really nice way to go, in my opinion. I think there are other less permanent approaches that work pretty well too, though, and my implementation of a constant directivity cornerhorn is the best soluton for non-permanent installation I've come up with thus far.
The best implementation would be where the sound sources are truly flush with the boundaries. The worst implementation is one where the sound sources are several wavelengths away from the boundaries, all the way through the passband. I think it's safe to assume that there is a range where the acoustic distance is close enough to provide benefit.
But one thing is for sure - If the sound source is further than a couple wavelengths from a boundary, then it acts as a reflector, and self-interference will result. If radiation is omnidirectional, the the sound directed at the boundary is as loud as the sound directed at the listener, so self-interference is severe.
In cases where the sound source is acoustically close at the low end of its passband, but becomes acoustically distant as frequency rises, then it will have no self-interference down low, but it may at higher frequencies. However, if the sound source becomes directional at higher frequencies, that helps attenuate the self-interference by the nature of directivity. I would suggest that is an improvement over a non-directional source.
What I see is typical speakers with direct radiating midwoofers tend to be placed a few feet away from walls. This is just the nature of the typical home installation. What this does is to create a worst-case scenario with respect to self-interference. The interference notches form in the mid hundred Hertz range, where the midwoofers are radiating omnidirectionally. So the response aberration is very strong.
If you can move the woofer and/or midrange drivers close enough to the apex of the corner where this doesn't happen, it's a huge benefit, in my opinion. It sure sounds a lot different, and measures cleaner too. Then if the horn(s) provide directivity at higher frequencies, that helps even more.
But yeah, it would be even better to flush mount the drivers into the walls. Of course, that kind of permanent installation precludes the possibility of movement, but it is a really nice way to go, in my opinion. I think there are other less permanent approaches that work pretty well too, though, and my implementation of a constant directivity cornerhorn is the best soluton for non-permanent installation I've come up with thus far.
They very much do if you want uniform dispersion...
Reflections mean that you have unwanted sound added to your speakers output and that means your sound is then distorted. I think that falls under the category of "it matters". Having speakers which are directive goes a long way to help minimize such reflections. The less directive your loudspeakers the more reflections you will have. Best regards Moray James.
Reflections mean that you have unwanted sound added to your speakers output and that means your sound is then distorted. I think that falls under the category of "it matters". Having speakers which are directive goes a long way to help minimize such reflections. The less directive your loudspeakers the more reflections you will have. Best regards Moray James.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.