New guy to speaker building here. I'm thinking of trying an array of small full range speakers shaped like a sphere or a section of one to approximate a wavefront coming from a point source. Surely somebody's tried this before, other than the guys at UC Berkeley and Meyer Sound. The references I find are electronic musicians building speakers for their instruments, like at SLORK. Anybody tried something along these lines for music reproduction? Suggestions for inexpensive drivers for a first cut at it? Maybe 9 of these PE 2 inchers per side?
http://www.parts-express.com/pe/showdetl.cfm?Partnumber=289-124
Would a sphere or hemisphere eliminate the arrival time effects of a line array, especially if it's relatively small, while keeping the dynamics of spreading the work over several drivers? Am I nutz?
http://www.parts-express.com/pe/showdetl.cfm?Partnumber=289-124
Would a sphere or hemisphere eliminate the arrival time effects of a line array, especially if it's relatively small, while keeping the dynamics of spreading the work over several drivers? Am I nutz?
Last edited:
Welcome to the forums!
I like the idea of making a sphere, as it would make the sound very omni-directional. However, you could run into phase issues, which could result in cancelling various frequencies and highlighting others, which wouldn't sound brilliant.
You could try something more like a guitar stack (drivers arranged in a square) which a more conventional box. This would take away the phase issues.
Chris
I like the idea of making a sphere, as it would make the sound very omni-directional. However, you could run into phase issues, which could result in cancelling various frequencies and highlighting others, which wouldn't sound brilliant.
You could try something more like a guitar stack (drivers arranged in a square) which a more conventional box. This would take away the phase issues.
Chris
This gent has done something along the lines of your inquiry. Reading about his experience may be of some assistance.
See Here:www.dmitrynizh.com/labaffles.htm
I hope this is some useful information.
John
See Here:www.dmitrynizh.com/labaffles.htm
I hope this is some useful information.
John
Thanks for the pointer. It looks well worth reading in depth. What I'm envisioning is something like a soccer ball, maybe a bit bigger, with small drivers tightly clustered on it to minimize phase cancellations, or at least push them way up in frequency. I've seen concave arrays to combat phasing, but that puts listeners at a focal point. What I'm wondering about is bending the array the other direction, simulating a wave expanding from a point source. Haven't seen many designs like that, perhaps like the Quad ESL57 curves in an attempt to broaden the sweet spot. I'm wondering if it's been tried and found to be just awful for some reason...
A guitar stack in miniature is also an intriguing idea.
A guitar stack in miniature is also an intriguing idea.
You mean something like the Bozak B200XA tweeter array designed to mimic a large multi-cell theater horn in a relatively small/inexpensive package as shown in the Concert Grand (B310 series)?
http://www.hifilit.com/hifilit/Bozak/1960-7.jpg
http://www.hifilit.com/hifilit/Bozak/1968-5.jpg
http://www.hifilit.com/hifilit/Bozak/1968-4.jpg
If so, it works fairly well with the caveat that just like the big horns it has increasingly audible phasing issues as it collapses from a point source to a bunch of increasingly beaming 'cells' with increasing frequency. Mono ruled when it was introduced, so was actually an acceptable compromise to ameliorate off axis HF roll off (beaming), but stereo...........??? Yeck!
So, Rudy (Bozak) then adapted the early RCA theater line array concept for his stereo variants, the Symphony No.1 (B4000). Much better, but still somewhat lacking and not a commercial success AFAIK. http://www.hifilit.com/hifilit/Bozak/1968-5.jpg
Later, McIntosh would expand on the line array concept till it became a sonic tour de force with the XR290 'infinite' array. It takes a very high level of performance to impress me and these do it like few other systems have over the decades: http://www.roger-russell.com/columns/xr290prod.jpg
GM
http://www.hifilit.com/hifilit/Bozak/1960-7.jpg
http://www.hifilit.com/hifilit/Bozak/1968-5.jpg
http://www.hifilit.com/hifilit/Bozak/1968-4.jpg
If so, it works fairly well with the caveat that just like the big horns it has increasingly audible phasing issues as it collapses from a point source to a bunch of increasingly beaming 'cells' with increasing frequency. Mono ruled when it was introduced, so was actually an acceptable compromise to ameliorate off axis HF roll off (beaming), but stereo...........??? Yeck!
So, Rudy (Bozak) then adapted the early RCA theater line array concept for his stereo variants, the Symphony No.1 (B4000). Much better, but still somewhat lacking and not a commercial success AFAIK. http://www.hifilit.com/hifilit/Bozak/1968-5.jpg
Later, McIntosh would expand on the line array concept till it became a sonic tour de force with the XR290 'infinite' array. It takes a very high level of performance to impress me and these do it like few other systems have over the decades: http://www.roger-russell.com/columns/xr290prod.jpg
GM
I'd forgotten the Bozak tweeter array. Never heard one, and never was all that fond of Bozak tonal balance. Best Bozak system I ran across back in the day was a friend's tri-amped (with McIntosh) Concert Grands with a JBL Paragon sitting between them. Impressive sound, if a bit vague on imaging, if memory serves, and memory serves less & less after all these years. The big cabinet with all the MC225s, MC240s and MC275s was also impressive as heck.
Another example I've heard of but never heard was Dr. Amar Bose's first speaker system, which was a corner-resident 1/8 sphere made with something like 32 small 4" drivers. Not practical in many rooms, and he abandoned it for the original 901.
I see your point about becoming less a point radiator and more a "sonic disco ball" at high frequencies, so the drivers have to be small to avoid beaming and close together to avoid phasing. I figure if it turns out badly, I can build a line array with the drivers.
Another example I've heard of but never heard was Dr. Amar Bose's first speaker system, which was a corner-resident 1/8 sphere made with something like 32 small 4" drivers. Not practical in many rooms, and he abandoned it for the original 901.
I see your point about becoming less a point radiator and more a "sonic disco ball" at high frequencies, so the drivers have to be small to avoid beaming and close together to avoid phasing. I figure if it turns out badly, I can build a line array with the drivers.
If I understand your concept correctly this speaker does what you're trying to accomplish: http://www.raal-requisite.com/
Well, ideally you'll need at most, ~3/8" o.d. wide BW drivers which don't exist AFAIK, so unless all your HF hearing is gone, having a 'plan B' is a good plan.
Good luck with it, looking forward to your findings.
GM
The main object of my curiosity is the constant refrain "ideal point source". Is a point the ideal source? And what does it sound like with which recordings? Would it image well with recordings made with a spaced pair of mics? Seems quite possible. What about coincident pair mics? Multitrack recordings?
The best reproduction in my experience sounds like there is a pair of windows from my listening room into the concert hall. With good speakers, the windows can be fairly large and lots of music comes through. Sometimes the illusion is that my room is in the front row balcony, sometimes that it's onstage, especially with pop and jazz close-mic'd. Some speakers sound like there are drapes covering the windows, some sound like there are pipes between my room and the performance. Occasionally a system sounds like "the performance is in the room with me", and in fact it is - there is sufficient distortion & resonance that the source becomes the speaker, not the original performer. These systems are limited in the kind of material they play well, and become tiring to listen to.
I've been surprised the last few years by several "computer" speakers which sound much better than their very modest parts would suggest. They have common attributes: small full-range drivers (usually with a separate woofer), modest amplifiers, etc. Think H-K Soundsticks or Henry Kloss' original Cambridge Soundworks. This has me thinking about scaling up to a larger system with no crossover, small drivers which are less prone to cone flex, and so on.
The best reproduction in my experience sounds like there is a pair of windows from my listening room into the concert hall. With good speakers, the windows can be fairly large and lots of music comes through. Sometimes the illusion is that my room is in the front row balcony, sometimes that it's onstage, especially with pop and jazz close-mic'd. Some speakers sound like there are drapes covering the windows, some sound like there are pipes between my room and the performance. Occasionally a system sounds like "the performance is in the room with me", and in fact it is - there is sufficient distortion & resonance that the source becomes the speaker, not the original performer. These systems are limited in the kind of material they play well, and become tiring to listen to.
I've been surprised the last few years by several "computer" speakers which sound much better than their very modest parts would suggest. They have common attributes: small full-range drivers (usually with a separate woofer), modest amplifiers, etc. Think H-K Soundsticks or Henry Kloss' original Cambridge Soundworks. This has me thinking about scaling up to a larger system with no crossover, small drivers which are less prone to cone flex, and so on.
focused arrays.............
here are some links.
http://www.atssounds.com/page1.aspx
http://yu-ra.tripod.com/array.htm
http://speakers.sub-optimal.net/displayimage.php?album=16&pos=6
http://www.dakotaaudio.com/web/images/speaker-array-equal-distanc.gif
http://www.electronichouse.com/images/slideshow/Gryphon_Poseidon.jpg
http://www.dmitrynizh.com/line-arrays16.jpg
norman
here are some links.
http://www.atssounds.com/page1.aspx
http://yu-ra.tripod.com/array.htm
http://speakers.sub-optimal.net/displayimage.php?album=16&pos=6
http://www.dakotaaudio.com/web/images/speaker-array-equal-distanc.gif
http://www.electronichouse.com/images/slideshow/Gryphon_Poseidon.jpg
http://www.dmitrynizh.com/line-arrays16.jpg
norman
There can be problems inherent in attempting to simulate a point source by using multiple drivers in an array that produces multiple arrival times within the first millisecond or so (sound travels about 13.5 inches per millisecond, to give you a reference point). These problems ca be primarily perceptual or acoustic.
Multiple early arrivals can cause coloration and degrade clarity because the ear may not see them as a coherent whole, nor as an initial sound followed by reflections; rather, it may see them as one not-so-coherent initial sound. If you will inevitably have multiple early arrivals (within the first millisecond or so) from different sources reproducing the same frequency, it's best if they are simultaneous.
Acoustic problems can arise from the interaction of drivers in an array at frequencies where the horizontal or vertical distance between them is equal to a half-wavelength or more. At those frequencies, we begin to get radiation pattern narrowing and destructive intereference (cancellation of certain frequencies in certain planes). So we don't get the smooth distribution of sound that our intuition led us to expect; instead we get a frequency response that may well change significantly according to our position, as well as a net loss of high frequency energy from destructive interference.
Power tapering of vertical arrays is one way of addressing these issues, and there are others, but arrays are not as easy to do well as would appear at first glance.
Multiple early arrivals can cause coloration and degrade clarity because the ear may not see them as a coherent whole, nor as an initial sound followed by reflections; rather, it may see them as one not-so-coherent initial sound. If you will inevitably have multiple early arrivals (within the first millisecond or so) from different sources reproducing the same frequency, it's best if they are simultaneous.
Acoustic problems can arise from the interaction of drivers in an array at frequencies where the horizontal or vertical distance between them is equal to a half-wavelength or more. At those frequencies, we begin to get radiation pattern narrowing and destructive intereference (cancellation of certain frequencies in certain planes). So we don't get the smooth distribution of sound that our intuition led us to expect; instead we get a frequency response that may well change significantly according to our position, as well as a net loss of high frequency energy from destructive interference.
Power tapering of vertical arrays is one way of addressing these issues, and there are others, but arrays are not as easy to do well as would appear at first glance.
Last edited:
to me, a besel array helps due to more info in the center due to louder drivers (precedence effect), but then you lose the increase in dynamics / clarity due to even power to all drivers, and yes you still have smear, less but it is still there.
I don't buy into the 1/2 wavelength center to center spacing.
I heard a 3' tall .75" tweeter array at 20', it sounded wrong, similar to listening to a panel speaker and perceiving a 6' tall guitar. Or sounded like a giant was singing.
Drives me nuts............
I think a strong tweeter with a bunch of full ranges in an array above/below it could work.
Norman
I don't buy into the 1/2 wavelength center to center spacing.
I heard a 3' tall .75" tweeter array at 20', it sounded wrong, similar to listening to a panel speaker and perceiving a 6' tall guitar. Or sounded like a giant was singing.
Drives me nuts............
I think a strong tweeter with a bunch of full ranges in an array above/below it could work.
Norman
I wasn't familiar with Bessel Arrays. They are certainly interesting, particularly the sum-difference stereo variants. It's not intuitively obvious why they work, so I've got another topic to research. Thanks.
audiokinesis, I understand your points, and I'm beginning to think my original idea of a ball-shaped array isn't likely to pan out well. However, nothing ventured... Sometimes unlikely designs work in practice, and I've never heard a ball array.
According to US patent #2006/0159289A1, which contains a lot of Bessel designs, it looks like the out-of phase driver(s) cancel most of the array's gain in sensitivity and power handling. Sort of what intuition would expect, no? Looks to me like you're wiring drivers out of phase so as to cancel the contribution from drivers at one side of the line. In a 5-driver line, 2 of the full-power drivers cancel, and the 2 end drivers are half-power. Naturally this will make the array act smaller, like a pair, in fact. Looks like a lot of hardware for little gain.
Me confused.
Me confused.
mcfabo,
in a focused array there is no vert dispersion.........
I can't show or link any measurements, they are not feasable nor marketable to 99% of people.
I remember someone commenting on TerryO's focused arrays that if you stood up, it was like someone switched off a tweeter.
The shape is similar to a radar dish or satelite dish, your ears are the same distance to all drivers..............
http://speakers.sub-optimal.net/displayimage.php?album=16&pos=6
Norman
in a focused array there is no vert dispersion.........
I can't show or link any measurements, they are not feasable nor marketable to 99% of people.
I remember someone commenting on TerryO's focused arrays that if you stood up, it was like someone switched off a tweeter.
The shape is similar to a radar dish or satelite dish, your ears are the same distance to all drivers..............
http://speakers.sub-optimal.net/displayimage.php?album=16&pos=6
Norman
thanx norman bates,
non-uniform dispersion forces to listen in one place.
also that the dispersion is large at low frequency undergoes a strong variation when changing the array. the reflectet environment will be certanly tonal unbalanced.
is reasonable a variant in which the speakers are not focused toward a single point, but are only slightly oriented in order to manipulate the curves of vertical dispersion (to avoid the normal dip between the drive)?
bye,
fabio
non-uniform dispersion forces to listen in one place.
also that the dispersion is large at low frequency undergoes a strong variation when changing the array. the reflectet environment will be certanly tonal unbalanced.
is reasonable a variant in which the speakers are not focused toward a single point, but are only slightly oriented in order to manipulate the curves of vertical dispersion (to avoid the normal dip between the drive)?
bye,
fabio
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.