These are 0 deg and 30 measurements of an equalized 3" driver. The main reason they are so much off above 10KHz is that enclosure edge diffraction cancels them out. Not too bad in comparison with horns? I would like to accomplish this with a driver twice the diameter.
An externally hosted image should be here but it was not working when we last tested it.
Last edited:
I understand that it changes crossover frequency location, but I can't figure out what you mean when you say "big effects on the crossover". Isn't it just a frequency shift, and a change of the LF driver to match the directivity of the waveguide at crossover frequency ? Other than that, is there something else to consider ?
One thing that helped me comprehend the Unity horns was the realization that it's basically a horn feeding a horn feeding a horn.
Once you get that, you realize a few things:
1) If designed properly, the high frequencies in a Unity horn do not "see" the midrange horn, and the midrange frequencies do not "see" the low frequency horn.
2) You can basically use an optimal driver for each frequency band. For instance, a Dayton ND91 isn't much of a tweeter, but it's a good midrange. Instead of trying to squeeze eight octaves out of it, I just use it in the range that it's best suited for.
The wavefront simulator in Hornresp is a great tool for visualizing how this works. It can show you how a wave basically glides down the center of a conical horn or waveguide, and once the wave is formed, it basically doesn't "see" the rest of the horn. Unless, of course, there's something in the way, which is why diffraction treatment is so important.
The key, of course, is to let the wave form properly. If it doesn't form properly in the correct location, no amount of EQ or DSP will bring it back. And this is why the transition at the waveguide throat is possibly the most important inch in the entire waveguide.
Hornresp can visualize wave front of a unity horn? I thought it does only single driver horns. I guess crossovers are used to control the range of each driver? Or does it just rely on acoustic filtering? Certainly would be nice to get it to work from the room corners so that they blend with the walls.
there are the later, lower frequency coupling holes in the walls too for the higher frequency wave to diffract - what are the practical opening area trade offs?
eaw's coaxial?:
http://eaw.com/docs/6_Technical_Information/White_Papers/AX_whtppr_L.pdf
seems like the Unity/Synergy lower frequency coupling holes can at best poorly approximate a ring radiator
eaw's coaxial?:
http://eaw.com/docs/6_Technical_Information/White_Papers/AX_whtppr_L.pdf
seems like the Unity/Synergy lower frequency coupling holes can at best poorly approximate a ring radiator
Last edited:
Basically Soongsc was trying to figure out how to get wide bandwidth out of a horn. And it can be done, definitely! But you run into two problems:
1) Bandwidth and efficiency are tied together. Widen the bandwidth and the efficiency drops. This isn't the end of the world though; we have a lot of efficiency in a compression driver. But you *really* lose a lot of efficiency every octave you go lower! For instance, I once measured my Summas and was surprised to see that the sensitivity was in the low 90s, even though the compression driver has a 'raw' sensitivity of 109dB. So four octaves of bandwidth comes at a price: a drop in efficiency of over 12 decibels.
2) Another problem is the phase and the CSD. If you look at the CSD plot of a compression driver, you'll notice that the CSD is clean in the lowest octaves, but gets grungier at higher and higher frequencies.
This problem is trickier; there's nothing you can do to clean up the CSD.
Lately I've been fascinated by the design approach of Richard Vandersteen and John Dunlavy. Those guys basically used a driver for two or three octaves, and no more.
Here's an example of what I mean.
The first measurements is the cumulative spectral delay of a B&C DE250 on a 60x90 horn, with no crossover.
The second measurement is the CSD of a Dunlavy Audio SC-IV/A. Even though the Dunlavy has five drivers and a gazillion crossover components, the Dunlavy has a cleaner CSD than the B&C. And keep in mind that both the drivers in the Dunlavy *and* the crossover components will insert a delay at certain frequencies. So the ability of the Dunlavy to cleanly decay over such a wide bandwidth is indicative of very careful driver selection and a crossover that's been finely tuned.
At home I have the Summas in one room and my Vandersteens in another. The Summas offer great dynamics and are 'easy on the ears.' The Vandersteens are delicate and revealing, but no match for the Summas dynamics.
The Unity approach seems to have the potential to combine both.
The one thing that vexes me is whether the HOMs will be too much to bear. Believe it or not, I find the Summas to be easier to listen to for long durations than just about any speaker I've ever heard, *including* the Vandersteens. And my hunch is that this is due to low diffraction and low HOMs.
I'm not sure if it's even *possible* to reduce HOMs in a Unity horn to such a low level, but I aim to find out
Last edited:
John
I would agree with both your observation and your explanation about why. I have found that even reducing the diffraction in the room near the speakers helps. I am not sure if you have gone to that extreme or not.
Sometimes that extreme is as simple as a quilt over the TV between the speakers when doing "serious" listening. Wall mounting of modern flatscreens or using a front projector help with this.
Unfortunately for me, my broke-as-heck lifestyle precludes a dedicated listening space- the left speaker and stacked subs are all of about 15 feet from my front wall. Not a lot I can do besides use directional systems, aim them, and do the best I can with other "stuff".
My rearfiring section has a nightmare of a lip on it right now, which I'm managing with felt, but I don't hear the diffraction when in front of them. They appear to be DDS waveguides but with a square flange, and I'm treating that with felt.
Rear section (not mounted)
An externally hosted image should be here but it was not working when we last tested it.
Fullsize http://imageshack.us/a/img35/9905/d1kn.jpg
Front Section (excuse my mess, it's always a bit less than tidy, but the table behind the speaker is NOT representative of the cleanliness)
An externally hosted image should be here but it was not working when we last tested it.
Fullsize http://imageshack.us/a/img831/3810/img0391oe.jpg
Last edited:
There's an interesting/maddening aspect to this:
The main reason that people buy loudspeakers is to listen to music and to watch movies. People spend a lot of time doing both; I don't think it's unusual to spend 30, 40, even 50 hours a week doing this. When I am working I have music or podcasts on basically all the time.
The 'audiophile' speakers, like the Vandersteens, grab your attention right away. You can't miss the soundstage; the difference in soundstaging between a Vandersteen and your run of the mill loudspeaker is noticeable in literally two or three seconds.
So I can see how this 'audiophile sound' can really sell loudspeakers. The typical audiophile will generally buy a speaker based on an audition that lasts for minutes, maybe an hour at the most.
But here's the thing -
The sound of diffraction is something that takes quite a while to perceive.
For instance, I can listen to the Summas for hours, and I never feel the urge to turn them down, or turn them off.
With the Vandersteens, I definitely enjoy them a great deal, but I can't listen to them for more than a couple hours.
I think that difference is diffraction.
And I can *perceive* the sound of diffraction because the Summas have enlightened me to the absence of diffraction. But if I hadn't trained myself to perceive the sound of diffraction, I don't think that I would be aware of the reason that the Vandersteens are ever so slightly fatiguing, while the Summas are not.
Of course the Vandersteens are very good, and well worth the money. But there IS a slightly fatiguing quality to them.
I invested some time on the audio boards, and read about the more expensive Vandersteens, and how they addressed the shortcomings of the less expensive models. And the reviewers generally agreed that the Big Buck Models had better drivers.
And yes, they DO. But there is ONE other thing that the big buck models feature:
The more expensive models add wool around the drivers.
A small tweak, a small expense, but it likely reduces diffraction.
Oh I know! That's what's so baffling about Vandersteen. (No pun intended lol)
For instance, Vandersteen uses these driver baskets that are designed to combat diffraction.
But I personally think that this is semi-misguided; diffraction from *outside* the loudspeaker is a lot more important than diffraction *inside* the loudspeaker.
A simple roundover would probably make a big difference here.
It raises an interesting question:
1) The dipole radiation of something like John K's NAO Note will reduce diffraction. It does this destructively; the sound from the front will null out at the edges.
Is the reduction in diffraction one of the reasons that dipoles sound so good?
IE, I wonder if one could achieve some or all of the benefits of dipoles by simply reducing diffraction. To do this would be simple; just use spherical enclosures.
It's hard to say if this would work. First, it's hard to quantify a reduction in diffraction. The only practical way to quantify the reduction is to build two enclosures, one with sharp edges, and one without.
I'd love it if I could just go out and listen to some round speakers, but I can only think of a handful. Anthony Gallo Acoustics and B&W come to mind. The Gallos sound good, but they have limited bandwidth and dynamics. The B&Ws sound great, but at that price, they better.
For instance, Vandersteen uses these driver baskets that are designed to combat diffraction.
But I personally think that this is semi-misguided; diffraction from *outside* the loudspeaker is a lot more important than diffraction *inside* the loudspeaker.
A simple roundover would probably make a big difference here.
It raises an interesting question:
1) The dipole radiation of something like John K's NAO Note will reduce diffraction. It does this destructively; the sound from the front will null out at the edges.
Is the reduction in diffraction one of the reasons that dipoles sound so good?
IE, I wonder if one could achieve some or all of the benefits of dipoles by simply reducing diffraction. To do this would be simple; just use spherical enclosures.
It's hard to say if this would work. First, it's hard to quantify a reduction in diffraction. The only practical way to quantify the reduction is to build two enclosures, one with sharp edges, and one without.
I'd love it if I could just go out and listen to some round speakers, but I can only think of a handful. Anthony Gallo Acoustics and B&W come to mind. The Gallos sound good, but they have limited bandwidth and dynamics. The B&Ws sound great, but at that price, they better.
Last edited:
It is true that we generally have to do some trade offs. The louder you have to go, the more you have to trade off other aspects. This is why i like to first know about how loud I need to make a system play.
My interest in horns came due to it's potential to generate a uniform wave front, which is impossible when you go multiple drivers in a horn no matter what you say. All experiments show that waves can seem to merge at a certain point, but eventually they pass through each other. On the other hand, compression drivers incorporate an addition energy storage mechanism in attempt to reform the wave front, this creates additional delayed release of stored energy like a spring. You could add damping to dissipate that energy faster, but you also might suffer in transient onset which effects bandwidth. If you consider what the foam plug does, that is a very good example if introducing some damping and minimal distributed diffraction to get a good tuning point.
Since I have established what I consider a good reference using a 3" wide range driver, the challenge is to both increase the bandwidth and maximum SPL without noticeable sound degradation. I am not the only person that evaluates the sound quality either. Since the maximum SPL of the current 3" design is the SPL of the low frequency, so naturally adding a larger driver to use two drivers was the natural way to go. But then, if I could create a more unified wave front using a horn, that seemed like a perfect solution, however, I could not get the horn to do that considering the size, and necessary crossover point, so I am leaving horns for when I need higher SPLs. Rather, it seems there is potential for a larger full range driver possibility. So that is the next level for me, but I have not given up on horns, just sort of limited how I would use them.
We are going to have diffraction one way or another, it just amounts to how they occur and how they effect our perception of sound. The unity horn design seems to me like a great way to turn a room into sort of a large horn, but I am not decided on that until I get to larger systems. Generally I think if you can keep the diffraction energy low and dominant frequency out of the audio range as much as possible, that would be best.
My interest in horns came due to it's potential to generate a uniform wave front, which is impossible when you go multiple drivers in a horn no matter what you say. All experiments show that waves can seem to merge at a certain point, but eventually they pass through each other. On the other hand, compression drivers incorporate an addition energy storage mechanism in attempt to reform the wave front, this creates additional delayed release of stored energy like a spring. You could add damping to dissipate that energy faster, but you also might suffer in transient onset which effects bandwidth. If you consider what the foam plug does, that is a very good example if introducing some damping and minimal distributed diffraction to get a good tuning point.
Since I have established what I consider a good reference using a 3" wide range driver, the challenge is to both increase the bandwidth and maximum SPL without noticeable sound degradation. I am not the only person that evaluates the sound quality either. Since the maximum SPL of the current 3" design is the SPL of the low frequency, so naturally adding a larger driver to use two drivers was the natural way to go. But then, if I could create a more unified wave front using a horn, that seemed like a perfect solution, however, I could not get the horn to do that considering the size, and necessary crossover point, so I am leaving horns for when I need higher SPLs. Rather, it seems there is potential for a larger full range driver possibility. So that is the next level for me, but I have not given up on horns, just sort of limited how I would use them.
We are going to have diffraction one way or another, it just amounts to how they occur and how they effect our perception of sound. The unity horn design seems to me like a great way to turn a room into sort of a large horn, but I am not decided on that until I get to larger systems. Generally I think if you can keep the diffraction energy low and dominant frequency out of the audio range as much as possible, that would be best.
There is a simple way to determine if a horn is producing a uniform wave front:
1) look at the polar response
2) look at the phase response
To understand how this is possible, you have to understand that there's a delay on the mids and the woofers. If it wasn't for that delay, you wouldn't get the uniform wavefront that you want.
Now you might look at the horn and say "there's a reflection in the throat, and a reflection will certain screw up the wavefront."
And that would be true, except for one thing:
The frequency of the wavelength that's reflected into the throat of a Synergy horn is so low, it basically merges with the original wavefront.
For instance:
- In a Synergy horn the midrange is playing two octaves from 350hz to 1400hz.
- There's a gap about six centimeters long, between the throat and the midranges.
- 5600hz is six centimeters long.
So there IS a reflection, but that reflection happens so close, it all but blends into the original wavefront.
I think this illustrates something important about diffraction and reflections though, which is that they're a much bigger problem at high frequency. At 1khz the wavelengths are so long, it takes a relatively large object to generate diffraction. But at four kilohertz it's a whole different story; the wavelengths are much shorter and even something two or three centimeters in diameter can generate it.
If so it's mostly a vertical effect you'd hear, which depending on the wavelength could be a reflection. (..owing to the lower mid and bass cabinet tops.)
Horizontal effects probably don't even begin until about 3 kHz, and are likely well off-axis.
My guess on fatigue effects would be an over-driven tweeter with relatively high non-linear effects (the result of use/abuse with a 1st order electrical filter).
The radiation pattern of each driver also effects the combined wave front. So the more drivers you have, the more complicated the design considerations, and the more cost involved. The polar pattern cannot correctly identify uniform wave front, but of you take measurements from outside of the horn to the inside, if the CSD look pretty much the same and decays as fast as a direct radiating driver, then I would say it is a good horn.
Last edited:
Subjectively, I find that the sound of diffraction is an 'edge' to high frequencies.
I hear this 'edge' at all volume levels in the Vandersteens.
If the effect I was hearing was due to a tweeter being overdriven, I think it would be less noticeable at lower volume.
Possibly the best data we have on diffraction reduction is an innocuous thread I started on diymobileaudio.com. It was a total goof - I basically suggested that people stick their tweeters in spheres.
But the thread took off like wildfire, and years down the road, dozens or even hundreds of people have taken their tweeters and placed them in spheres.
It's a good source of diffraction data because we have a set of listeners who heard their tweeters *without* diffraction treatment, and *with* treatment. It's a long thread, but the general response seems to be threefold:
1) better imaging
2) the tweeter sounds "smoother"
3) the improvement is audible and unmistakable
It's the 'smoother' part which is particularly interesting.
Here's one of many reviews from the thread:
"i enjoy the spheres I did in the Scion tC with Hertz ML28 tweeters so much, i'm doing a set for my own car using my Oz-25 tweeters which are currently off-axis on the a-pillars. i have a separate set of a-pillars so i can switch back and forth if necessary (or while the spheres are curing/drying).
the sound stage width is greatly widened and it's just amazing how you can't locate a tweeter that is staring you in the face. diffraction is very real, as is the resulting phase-interference/comb-filtering that ensues."
Car Audio | DiyMobileAudio.com | Car Stereo Forum - View Single Post - Improve Your Soundstage for $2
I've trained myself to notice the sound of drivers being overdriven, and subjectively it sounds 'congested' to me. One interesting thing about that type of distortion is that I don't find it particularly offensive or fatiguing. For instance, I have some small speakers that exhibit that effect, and they don't make me want to turn them off. YES, the congestion can obscure the details of a good recording, but it's not fatiguing. In fact, I think a case could be made that overdriving a speaker may make it sound *better* to some people. Some people *prefer* the sound of harmonic distortion.
But diffraction, IMHO, is different. It's fatiguing.
Here's an example of this:
There's a bar near me with very bad sound. Just that white-hot horn honk that gives prosound gear a bad name. And if you look at the horns, you see the culprit. They use a horn like this:
An externally hosted image should be here but it was not working when we last tested it.
Now a big sharp edge like that, in front of the high frequency driver, is basically going to generate tons of diffraction. But I'm 'tuned in' to that sound now, and I can hear it in small amounts in a lot of conventional boxes too.
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.
Another example would be my Control Now speakers and my Kef speakers. My Kefs don't even have a hint of diffraction. And the enclosure is egg shaped. My Control Now speakers definitely have it; and they also have an abrupt edge on the enclosure. Basically the tweeter waveguide needs a proper roundover on the edge, a bigger waveguide or both. In the pic you can see that there's a *small* roundover, but a *small* roundover won't cut it. 2000hz is 17 centimeters long; so a roundover will need to be at least 4-8cm long do to some good here.
Interestingly, the Kefs also 'disappear' in a way that the JBLs cannot.
The Kefs measure worse than the JBLs, so this isn't a frequency response thing. The JBLs have superior power and frequency response. (I've measured both.)
Last edited:
Maybe it is.
Make short work of it and test it with some damping material.
Target each area individually (vertical vs. horizontal), and then in total.
In a proper multi driver horn, the radiators are within one quarter wavelength so they function virtually as a single driver.
It takes a really really long time to understand all the complexities of a Synergy horn.
One of the odder revelations, once you understand all of the puzzle pieces, is that adding drivers actually makes a Synergy horn perform better, not worse!
This is rather unique in audio.