Some things need need clearing up in my head...
The rear wave of a dipole speaker arrives some period after the forward radiating wave such that the two do not cancel each other out?
Seemingly there is still a cancellation between the now arrived rear wave and the current forward wave? Isn't this a fundamental drawback of dipoles?
Or is the out of phase relfected wave attentuated enough that this isn't a problem? (edit: and the same would apply to the rear-facing in-phase wave of bipoles).
The rear wave of a bipole is encapsulated? So it's just a sealed or ported enclosure with two symetrical baffles to overcome baffle step (perhaps without tweeters on both?). Conceivably, you could push-pull on both baffles...hmm...
A sealed enclosure through encapsulating the rear wave rids itself of such concerns. And since by adjusting enclosure size the driver parameters are adjusted, a wider variety of drivers are usable. Why opt for a dipole or bipole at all? can you say they have better transients, or better overall x or y?
What about using a large amount of stuffing material directly behind a dipole. It seems to me that this would help, but I haven't seen it?
The rear wave of a dipole speaker arrives some period after the forward radiating wave such that the two do not cancel each other out?
Seemingly there is still a cancellation between the now arrived rear wave and the current forward wave? Isn't this a fundamental drawback of dipoles?
Or is the out of phase relfected wave attentuated enough that this isn't a problem? (edit: and the same would apply to the rear-facing in-phase wave of bipoles).
The rear wave of a bipole is encapsulated? So it's just a sealed or ported enclosure with two symetrical baffles to overcome baffle step (perhaps without tweeters on both?). Conceivably, you could push-pull on both baffles...hmm...
A sealed enclosure through encapsulating the rear wave rids itself of such concerns. And since by adjusting enclosure size the driver parameters are adjusted, a wider variety of drivers are usable. Why opt for a dipole or bipole at all? can you say they have better transients, or better overall x or y?
What about using a large amount of stuffing material directly behind a dipole. It seems to me that this would help, but I haven't seen it?
You want to visit Siegfried Linkwitz's website for informtation on dipoles.
Subjectively, I find the Orions have the most natrual bass I've heard even when placed in a modest sized room. The midrange at least equals what you get out of an ESL (I need to listen to one again) although it integrates well with the bass and the sweet spot is much wider.
Obviously, a dipole is free of cabinet resonances and stored energy. Other interesting characteristics come from the dipole radiation pattern.
Once the wavelengths have become long enough that a driver illuminates the baffle edges you have some cancelation which results in a cosine alpha polar response attenuation (20 log cosine alpha: -3dB @ 45 degrees, -6dB @ 60 and at 90 degrees you're just picking up the reverberant field).
This is very different monopole which approximates an omnidirectional source once the frequency is low enough. Side wall reflelctions are less of a problem, and the ratio of on-axis to total power response is 4.8dB higher. Dipole bass is perceptually very different (more natural).
This is all very good.
The biggest problem for full range use is that once the wavelength exceeds Fequal =.17 v/D where v is the speed of sound and D the dipole path difference (half baffle width plus wing size on shallow baffles, wing depth on deep ones) you have dipole roll-off at 6dB/octave. Big silicon amps have enough power to work arround this, although it means excursion is increasing at 18dB/octave instead of 12dB/octave in a sealed box. Your last two octaves require oodles of displacement (a pair of 10" XLS subwoofers on a 15" deep are excursion limited to 100dB @ 40Hz and 82dB @ 20Hz). If you want to achieve realistic output levels, the high resulting woofer sensitivity before dipole roll-off forces the use of a line-level cross-over on the woofers and at least bi-amplification.
Dipoles also don't suffer or benefit from room gain.
If you stuff a dipole, you attenuate the back wave of high frequencies more and low frequencies less so you get a dipole behavior at low frequencies and cardoid at higher.
Subjectively, I find the Orions have the most natrual bass I've heard even when placed in a modest sized room. The midrange at least equals what you get out of an ESL (I need to listen to one again) although it integrates well with the bass and the sweet spot is much wider.
Obviously, a dipole is free of cabinet resonances and stored energy. Other interesting characteristics come from the dipole radiation pattern.
Once the wavelengths have become long enough that a driver illuminates the baffle edges you have some cancelation which results in a cosine alpha polar response attenuation (20 log cosine alpha: -3dB @ 45 degrees, -6dB @ 60 and at 90 degrees you're just picking up the reverberant field).
This is very different monopole which approximates an omnidirectional source once the frequency is low enough. Side wall reflelctions are less of a problem, and the ratio of on-axis to total power response is 4.8dB higher. Dipole bass is perceptually very different (more natural).
This is all very good.
The biggest problem for full range use is that once the wavelength exceeds Fequal =.17 v/D where v is the speed of sound and D the dipole path difference (half baffle width plus wing size on shallow baffles, wing depth on deep ones) you have dipole roll-off at 6dB/octave. Big silicon amps have enough power to work arround this, although it means excursion is increasing at 18dB/octave instead of 12dB/octave in a sealed box. Your last two octaves require oodles of displacement (a pair of 10" XLS subwoofers on a 15" deep are excursion limited to 100dB @ 40Hz and 82dB @ 20Hz). If you want to achieve realistic output levels, the high resulting woofer sensitivity before dipole roll-off forces the use of a line-level cross-over on the woofers and at least bi-amplification.
Dipoles also don't suffer or benefit from room gain.
If you stuff a dipole, you attenuate the back wave of high frequencies more and low frequencies less so you get a dipole behavior at low frequencies and cardoid at higher.
Drew Eckhardt said:
Mostly true, but... The dipole baffle can still resonate and add its own coloration.
No argument that dipoles suffer this problem rather less than boxed speakers.
I apologize for complicating the discussion, but using dipoles complicates loudspeaker design and room interaction. Here are just two of many points to consider when designing dipoles. First, while musical instruments (acoustic sort) are not truly omni directional, they do produce sounds that are widely dispersed. A lot of time, effort, and money have been spent to create performance spaces where different musical instruments all sound great, yet most performance spaces are considered inadequate. The number of really great sounding performance spaces is a small number. All sorts of corrections have been tried, ranging from “clouds” suspended from the ceiling to computer based adaptive sound reinforcement systems. Few of the corrections have ever been judged unqualified successes. In addition, simply adding more energy by either increasing the number of instruments playing or changing the direction some of the instruments face does not correct or decrease the number of room modes and anomalies excited. This is not to say that the concept will not work, it is just much more a complex problem with a greater number of variables.
Two, just as the response off the front diaphragm of drivers varies from loudspeaker driver model to loudspeaker driver model, so does the sound coming off the front and back of the diaphragm. The difference in sound generated by the front of the diaphragm and the back of the diaphragm is greater with some drivers than it is with others. This just adds another layer of complexity in driver selection. Rear frequency response is a driver characteristic that is rarely if ever published. When you start mixing the front and a rear sound that are of themselves extremely different at differing delays it becomes an extremely complex design task. I have included the front and rear diaphragm response of a driver that is used by a few members of diyaudio. I have seen threads where this driver is pictured and named. The identity of the driver is not important, but I doubt this is the correct choice for a dipole design.
Best of luck with your designing,
Mark
Two, just as the response off the front diaphragm of drivers varies from loudspeaker driver model to loudspeaker driver model, so does the sound coming off the front and back of the diaphragm. The difference in sound generated by the front of the diaphragm and the back of the diaphragm is greater with some drivers than it is with others. This just adds another layer of complexity in driver selection. Rear frequency response is a driver characteristic that is rarely if ever published. When you start mixing the front and a rear sound that are of themselves extremely different at differing delays it becomes an extremely complex design task. I have included the front and rear diaphragm response of a driver that is used by a few members of diyaudio. I have seen threads where this driver is pictured and named. The identity of the driver is not important, but I doubt this is the correct choice for a dipole design.
Best of luck with your designing,
Mark
Attachments
Konnichiwa,
This is actually patently untrue. Dipoles realised with common (Magnet & Cone) drivers offer a much greater independence from room effects. Here is why:
1) Throughout the rooms modal range (up to around 300Hz a dipole will interact very different with the room modes (Kempe/Ferekidis). While a normal speaker will excite room modes most in a corner (pressure maxima) and least in the middle of the room raised to 1/2 ceiling hight (velocity maxima), the dipole behaves opposite (general speaker theory, many sources), so placing a dipole near corners (where speakers invariably end up) is much less of an issue.
2) Throughout the lower midrange where the rear radiation still remains about the same in magnitude and opposite polarity to the front radiation the Dipole developes deep "nulls" in the sideways radiation, namely that which "bounces" off the sidewalls to cause early reflections that many find to "blurr" the stereo image (general speaker theory, many sources).
3) At frequencies high enough the combined actions of the driver basket's and enclosed air helmholz resonator leads to a 2nd order rolloff of the rear radiation (Backman) with the magnets "shadow" further rolling off the rear radiation at high frequencies. The speaker now transtions from a dipole to a cardiod/hypercardiode dispersion, further aided by the "beaming" due to the sound wavelength becomming small with respect to the diaphragm. Again, sideways and now also rear radiation are strongly suppressed, meaning the DIRECT sound (recording) dominates the sound at the listening position.
4) I have not seen this mentioned before, empirical testing of my own with friends has revealed a much reduced rise of the RT60 towards lower frequencies in normal rooms with dipoles, vs. Monopoles. This was so starteling when measuring that we at first believed we had made a mistake during measurement.
5) The LF transient in an open baffle is much more precise and symmetric than in ANY enclosure, including Horns (Briggs), simply due to the fact that driver sees the identical acoustic pressure envoironment on the In-Stroke and the Out-Stroke, whereas in enclosures thge in-stroke sees a compression of the air in the enclosure while on the out-stroke we have a rarefaction, the pressure curves for rarefaction and compression look different.
Most of the above Effetcs (minimised room mode interaction, excellent directivity [Index] down to very low frequencies and good DI to high frequencies, fairly even RT60 with frequency) can of course be attained with other methodes, however few of these methoes are compatible with the decor and size of normal living spaces, as other methodes of LF directivity control involve usually very large waveguides (Horns) and the suppression of room modes requires large helmholz bass traps.
Yet a speaker is NOT a musical instrument, it is in fact (ideally) an "inverse microphone". It's goal is to reproduce the recording, not to create the sound of a given instrument in the room (you need to use the right recording techniques and a good speaker system in a good room will do exactly that).
Yes, however this is a not so much a problem as it seems, in practice. In fact, it is DESIRABLE to have the rear radiation shall roll off, in fact as early as possible above the frequency where the room "transtits" into the "diffuse sound" region out of the "modal region".
Funny. While quite uneven at higher frequencies (where a little absorbent material on the wall behind the speaker such a wall-rug or a large potted plant with small leaves may be on order to absorb the rear wave), the fact that the rear output drops quite evenly and quickly above around 1 - 1.5KHz (hard to see on the graphs frequency scale) makes the driver eminently suitable for a Dipole Speaker that is a dipole at low frequencies and a monopole at high frequencies, just as room acoustics would demand anyway!
So, one would be rather interested in the drivers identity, so one might use it!
Sayonara
MarkMcK said:
This is actually patently untrue. Dipoles realised with common (Magnet & Cone) drivers offer a much greater independence from room effects. Here is why:
1) Throughout the rooms modal range (up to around 300Hz a dipole will interact very different with the room modes (Kempe/Ferekidis). While a normal speaker will excite room modes most in a corner (pressure maxima) and least in the middle of the room raised to 1/2 ceiling hight (velocity maxima), the dipole behaves opposite (general speaker theory, many sources), so placing a dipole near corners (where speakers invariably end up) is much less of an issue.
2) Throughout the lower midrange where the rear radiation still remains about the same in magnitude and opposite polarity to the front radiation the Dipole developes deep "nulls" in the sideways radiation, namely that which "bounces" off the sidewalls to cause early reflections that many find to "blurr" the stereo image (general speaker theory, many sources).
3) At frequencies high enough the combined actions of the driver basket's and enclosed air helmholz resonator leads to a 2nd order rolloff of the rear radiation (Backman) with the magnets "shadow" further rolling off the rear radiation at high frequencies. The speaker now transtions from a dipole to a cardiod/hypercardiode dispersion, further aided by the "beaming" due to the sound wavelength becomming small with respect to the diaphragm. Again, sideways and now also rear radiation are strongly suppressed, meaning the DIRECT sound (recording) dominates the sound at the listening position.
4) I have not seen this mentioned before, empirical testing of my own with friends has revealed a much reduced rise of the RT60 towards lower frequencies in normal rooms with dipoles, vs. Monopoles. This was so starteling when measuring that we at first believed we had made a mistake during measurement.
5) The LF transient in an open baffle is much more precise and symmetric than in ANY enclosure, including Horns (Briggs), simply due to the fact that driver sees the identical acoustic pressure envoironment on the In-Stroke and the Out-Stroke, whereas in enclosures thge in-stroke sees a compression of the air in the enclosure while on the out-stroke we have a rarefaction, the pressure curves for rarefaction and compression look different.
Most of the above Effetcs (minimised room mode interaction, excellent directivity [Index] down to very low frequencies and good DI to high frequencies, fairly even RT60 with frequency) can of course be attained with other methodes, however few of these methoes are compatible with the decor and size of normal living spaces, as other methodes of LF directivity control involve usually very large waveguides (Horns) and the suppression of room modes requires large helmholz bass traps.
MarkMcK said:
Yet a speaker is NOT a musical instrument, it is in fact (ideally) an "inverse microphone". It's goal is to reproduce the recording, not to create the sound of a given instrument in the room (you need to use the right recording techniques and a good speaker system in a good room will do exactly that).
MarkMcK said:
Yes, however this is a not so much a problem as it seems, in practice. In fact, it is DESIRABLE to have the rear radiation shall roll off, in fact as early as possible above the frequency where the room "transtits" into the "diffuse sound" region out of the "modal region".
MarkMcK said:
Funny. While quite uneven at higher frequencies (where a little absorbent material on the wall behind the speaker such a wall-rug or a large potted plant with small leaves may be on order to absorb the rear wave), the fact that the rear output drops quite evenly and quickly above around 1 - 1.5KHz (hard to see on the graphs frequency scale) makes the driver eminently suitable for a Dipole Speaker that is a dipole at low frequencies and a monopole at high frequencies, just as room acoustics would demand anyway!
So, one would be rather interested in the drivers identity, so one might use it!
Sayonara
Konnichiwa,
Yes, BUT the point where the waves in effect "mix" ist at the listening position. By the time the rear sound has arrived there it will not have been delayed, but also have been attenuated, especially if you "toe in" the dipole speaker towards the listening position (usually around 30 degrees).
The directivity of the rear radiotion of a dipole is quite well known and understood, so we can plot the say three main path along which such sound will bounce around the room.
The strongest rear radiation is on axis, which is deflected by the rear-wall at an angle of around 60 degrees towards the sidewall, where the sound has an incoming angle of 30 degrees and is reflected on towards the rear wall, where the actual point where this (imaginary) line hit's the rear wall may very well be still notably to the side(s) of the listening position, making sure the sound takes another trip towards the rear wall (this time between the speakers) and is then reflected back towards the listener, most likely strongly diffused.
Any radiation that will be on shorter "routes" to the listener will be reduced in level due to the (rear) dispersion charateristic of the dipole and thus make much less of an impact.
The situation is still not exactly ideal, but much better than that for normal speakers, where sidewall and floor early reflections can louse up the response by far more.
Sayonara
Vikash said:
Yes, BUT the point where the waves in effect "mix" ist at the listening position. By the time the rear sound has arrived there it will not have been delayed, but also have been attenuated, especially if you "toe in" the dipole speaker towards the listening position (usually around 30 degrees).
The directivity of the rear radiotion of a dipole is quite well known and understood, so we can plot the say three main path along which such sound will bounce around the room.
The strongest rear radiation is on axis, which is deflected by the rear-wall at an angle of around 60 degrees towards the sidewall, where the sound has an incoming angle of 30 degrees and is reflected on towards the rear wall, where the actual point where this (imaginary) line hit's the rear wall may very well be still notably to the side(s) of the listening position, making sure the sound takes another trip towards the rear wall (this time between the speakers) and is then reflected back towards the listener, most likely strongly diffused.
Any radiation that will be on shorter "routes" to the listener will be reduced in level due to the (rear) dispersion charateristic of the dipole and thus make much less of an impact.
The situation is still not exactly ideal, but much better than that for normal speakers, where sidewall and floor early reflections can louse up the response by far more.
Sayonara
- Status
- Not open for further replies.