For a radiator of finite size, the dispersion pattern varies with frequency. At low frequencies with wavelengths large compared to the radiator dimension, the radiation is roughly omnidirectional. With increasing frequency (decreasing wavelength), the radiation pattern narrows and starts getting lobes. This is all due to classic interference effects and will happen even if the radiator is a perfect piston. If you look in a good acoustics text like Beranek, you'll see polar diagrams which illustrate this.
Now, is this a good thing? Like everything else pulled out of context, the answer is, "It depends." A very common situation where this is NOT a good thing comes when you try to mate a relatively large woofer with a relatively small tweeter. With practical choices of crossover frequency, the woofer's dispersion is narrowing, the tweeter still has very wide dispersion. So if you get the on-axis response flat, off axis response will have a dip in the upper part of the woofer's range and a peak in the lower part of the tweeter's range.
Because of that dip-and-peak, the reverberant energy will NOT have a flat response. So, OK, you adjust things to balance the reverberant energy. Oh-oh, the on-axis response isn't flat anymore. What you end up with is a speaker that either sounds colored everywhere or one that can sound fine as long as it is placed just so and the room is acoustically treated just so... The somewhat retarded chimp that's doing the review for a magazine notes that the speaker is sooooo good that it rewards every placement tweak that you can do.
There are many schools of thought regarding what an optimal polar pattern is, ranging from pencil-thin dispersion to flat on-axis but rolling off off-axis to full omni. What is never good is designing a speaker without first deciding on the target polar pattern and measuring the design against the goal.
Now, is this a good thing? Like everything else pulled out of context, the answer is, "It depends." A very common situation where this is NOT a good thing comes when you try to mate a relatively large woofer with a relatively small tweeter. With practical choices of crossover frequency, the woofer's dispersion is narrowing, the tweeter still has very wide dispersion. So if you get the on-axis response flat, off axis response will have a dip in the upper part of the woofer's range and a peak in the lower part of the tweeter's range.
Because of that dip-and-peak, the reverberant energy will NOT have a flat response. So, OK, you adjust things to balance the reverberant energy. Oh-oh, the on-axis response isn't flat anymore. What you end up with is a speaker that either sounds colored everywhere or one that can sound fine as long as it is placed just so and the room is acoustically treated just so... The somewhat retarded chimp that's doing the review for a magazine notes that the speaker is sooooo good that it rewards every placement tweak that you can do.
There are many schools of thought regarding what an optimal polar pattern is, ranging from pencil-thin dispersion to flat on-axis but rolling off off-axis to full omni. What is never good is designing a speaker without first deciding on the target polar pattern and measuring the design against the goal.
Konnichiwa,
It refers to a change in radiation pattern with frequency, where higher frequencies are radiated in a narrower angle than lower ones. Or in other words, if you have a flat on axis response and you measure the response significantly off-axis you will observe a progressive lessening of SPL above a certain frequency.
Because unless your speakers is sufficiently "beamy" even at lower frequencies it will generate a room reverbrant field which dominates the sound, in other words the spatial signature (and to a lesser degree the son ic one) replaces that of the recording venue.
Of course false High End doctrine demands a wide dispersion (in other words an absence of beaming) which in turn then forces system owners to significantly invest in acoustic treatment.
Sayonara
paulspencer said:
It refers to a change in radiation pattern with frequency, where higher frequencies are radiated in a narrower angle than lower ones. Or in other words, if you have a flat on axis response and you measure the response significantly off-axis you will observe a progressive lessening of SPL above a certain frequency.
paulspencer said:
Because unless your speakers is sufficiently "beamy" even at lower frequencies it will generate a room reverbrant field which dominates the sound, in other words the spatial signature (and to a lesser degree the son ic one) replaces that of the recording venue.
Of course false High End doctrine demands a wide dispersion (in other words an absence of beaming) which in turn then forces system owners to significantly invest in acoustic treatment.
Sayonara
My impression is that it's a bad thing because the off-axis response will tend to vary with frequency, starting off even and wide and narrowing as frequency rises. So if you attempt to design for a smooth power response, the on-axis response ends up shouting at you, and if you shoot for smooth on-axis response the in-room response is biased toward the lower end of the frequencies in question.
Then there's the issue of power response at the crossover point, where if the mid/bass driver is beaming there can be a jarring change in power response when the tweeter takes over.
edit: oops, beaten and with superior expanations to boot. Ah well, back to the coffee.
Then there's the issue of power response at the crossover point, where if the mid/bass driver is beaming there can be a jarring change in power response when the tweeter takes over.
edit: oops, beaten and with superior expanations to boot. Ah well, back to the coffee.
Konnichiwa,
Surely this is a good thing?
It makes sure of two things:
1) With rising frequency the stereo separation improves and any impulses or transients, which tend to dominate imaging, are less distorted by room generated reflections as they are lower in level.
2) The off axis (room) sound will show the typhical "house curve" due to greater rolloff at higher frequencies, failure to do so produce invariably an overly bright sound.
If your speaker does not beam enough you in fact require acoustic room treatment to correct for the problems introduced.
This BTW illustrates the complete and utter idiocy of using wide dispersion high frequency systems (eg. dome tweeters).
Sayonara
HeatMiser said:
Surely this is a good thing?
It makes sure of two things:
1) With rising frequency the stereo separation improves and any impulses or transients, which tend to dominate imaging, are less distorted by room generated reflections as they are lower in level.
2) The off axis (room) sound will show the typhical "house curve" due to greater rolloff at higher frequencies, failure to do so produce invariably an overly bright sound.
If your speaker does not beam enough you in fact require acoustic room treatment to correct for the problems introduced.
This BTW illustrates the complete and utter idiocy of using wide dispersion high frequency systems (eg. dome tweeters).
Sayonara
Kuei Yang Wang said:
I'll defer to your greater expertise in the area of speaker/room interactions as to whether the varying reflective nature of listening rooms tends to naturally correct a change in dispersion with frequency. The point it though that it does vary, and particularly in multi-driver systems that fact seems to suggest some design considerations.
Konnichiwa,
Typhically early reflection points are not covered for diffusion or absorbtion, ceilings are always non-sound absorbing, the modern tendency towards laminate floors does also not help. So, at least for modern designs of room arrangements, nope, the envoironment does not help.
Absolutely. Well controlled directivity with no abrupt swings in directivity (as are charateristic for common 2/2.5 Way speakers with cone midranges and dome tweeters) is a fundamental requirement for high performance speakers (the regular absence of said feature in proclaimed "state of the art" speakers nonwithstanding).
Sayonara
HeatMiser said:
Typhically early reflection points are not covered for diffusion or absorbtion, ceilings are always non-sound absorbing, the modern tendency towards laminate floors does also not help. So, at least for modern designs of room arrangements, nope, the envoironment does not help.
HeatMiser said:
Absolutely. Well controlled directivity with no abrupt swings in directivity (as are charateristic for common 2/2.5 Way speakers with cone midranges and dome tweeters) is a fundamental requirement for high performance speakers (the regular absence of said feature in proclaimed "state of the art" speakers nonwithstanding).
Sayonara
Kuei Yang Wang said:
Idiocy? Bit harsh but nevertheless...
IMO Room treatment is essential for any repectable hifi system regardless of dispertion. Some like me go steps further and use DRC and EQ. Purists will say that's wrong but you've got to consider which offers less distortion and I can easily answer that.
I've used ribbons, dome and current ring radiators for HF, all of the gained from room treatment.
Konnichiwa,
Actually, I was being charitable.
A competently designed speaker (WRT Dispersion) is easily found to provide more accurate imaging and tonality in an untreated room than most conventional ones in moderatly treated rooms (eg. anything short of an anechonic chamber). This is not to say that sensible room treatment will not provide further improvement, merely that it becomes more of an option.
Room treatment and EQ/DRC (DRC is really just a specialised EQ) work very different and on different issues. EQ cannot correct for badly designed dispersion or adjust RT60 etc, room treatment (and speaker design) can.
Mixing the two up is quite inapropriate.
In other words you used devices that invariably have wide dispersion and are thusly singulary inaproppriate for reproducing music in acoustically relatively small spaces (ordinary living rooms for example).
Treating the poor dispersion control at the source much reduces the benefit of room treatment. Plus, not only tweeters are subject to overly wide dispersion.
It would seem ideal that a directivity index of 6db up to around 2...3KHz and then smoothly rising to 10db @20KHz seems the most desirable charateristic, even though it is difficult to achieve. Using Dipoles for LF we can reach 4.8db DI for low frequencies, suitable waveguides can produce 6db DI above around 300Hz without becoming unpleasantly large and having the DI smoothly increase above a few KHz is not particulary difficult either.
The speakers I currently use have 4.8db DI to around 800Hz and then increases smoothly to around 10..12db at the upper end of the audio range. This is achieved by the simple expedient of using a fairly large cone driver in a fairly large dipole and the addition of a narrow dispersion horn tweeter for very high frequencies. Room treatment past normal furnishings and room EQ have become largely unneccesary and are hence no longer in use.
Sayonara
ShinOBIWAN said:
Actually, I was being charitable.
ShinOBIWAN said:
A competently designed speaker (WRT Dispersion) is easily found to provide more accurate imaging and tonality in an untreated room than most conventional ones in moderatly treated rooms (eg. anything short of an anechonic chamber). This is not to say that sensible room treatment will not provide further improvement, merely that it becomes more of an option.
ShinOBIWAN said:
Room treatment and EQ/DRC (DRC is really just a specialised EQ) work very different and on different issues. EQ cannot correct for badly designed dispersion or adjust RT60 etc, room treatment (and speaker design) can.
Mixing the two up is quite inapropriate.
ShinOBIWAN said:
In other words you used devices that invariably have wide dispersion and are thusly singulary inaproppriate for reproducing music in acoustically relatively small spaces (ordinary living rooms for example).
Treating the poor dispersion control at the source much reduces the benefit of room treatment. Plus, not only tweeters are subject to overly wide dispersion.
It would seem ideal that a directivity index of 6db up to around 2...3KHz and then smoothly rising to 10db @20KHz seems the most desirable charateristic, even though it is difficult to achieve. Using Dipoles for LF we can reach 4.8db DI for low frequencies, suitable waveguides can produce 6db DI above around 300Hz without becoming unpleasantly large and having the DI smoothly increase above a few KHz is not particulary difficult either.
The speakers I currently use have 4.8db DI to around 800Hz and then increases smoothly to around 10..12db at the upper end of the audio range. This is achieved by the simple expedient of using a fairly large cone driver in a fairly large dipole and the addition of a narrow dispersion horn tweeter for very high frequencies. Room treatment past normal furnishings and room EQ have become largely unneccesary and are hence no longer in use.
Sayonara
Without offering my own opinion, it should be said that there are various approaches to "proper" dispersion, all with their own adherents. The one thing all approaches have in common is people who claim that their paradigm is the one-and-only.
Where issues of competence and idiocy come in is whether dispersion was adequately considered and whether that consideration was followed up by appropriate and effective design.
Where issues of competence and idiocy come in is whether dispersion was adequately considered and whether that consideration was followed up by appropriate and effective design.
It's as if no one here has heard of dipoles or waveguides. Especially suprising given, the very knowledgable posters Shinobiwan & Kuei Yang Wang.
The idiocy of using a dome? Siegrfried Linkwitz is an idiot for using a SEAS Millenium and Kuei Yang Wang is smart for not using domes huh? Thats interesting. Very interesting. Perhaps Mr. Wang can kindly point out some reviews by his peers of his design/s. I would certainly be interested in hearing them. Any magazine reviews to start with? So that I can get an idea of what I might get to hear? Graphs, distortion measurements, controlled listening tests, etc. ? I have this strange affinity to see some factual data, rather than Audiophile ramblings. That's just me.
I'll just throw in a pic. to illustrate my attempt at controlled directivity throughout the speakers entire bandwith. Dipole, waveguide, and of course, a idiot Ring radiator, ugly cousin of the dome.
Take from it what you may.
BTW, Paul, to address your question directly, beaming is generally a bad thing, but like all things, represents part of your design that has to be addressed one way or another.
I'm from the school of blame the designer, not the parts (domes,etc.) or enclosure type (sealed box,ported,TL,OB,etc).
Look within first.
Cheers,
AJ
** EDIT - And off course SY posts something as I slooowwwlly typed!
The idiocy of using a dome? Siegrfried Linkwitz is an idiot for using a SEAS Millenium and Kuei Yang Wang is smart for not using domes huh? Thats interesting. Very interesting. Perhaps Mr. Wang can kindly point out some reviews by his peers of his design/s. I would certainly be interested in hearing them. Any magazine reviews to start with? So that I can get an idea of what I might get to hear? Graphs, distortion measurements, controlled listening tests, etc. ? I have this strange affinity to see some factual data, rather than Audiophile ramblings. That's just me.
I'll just throw in a pic. to illustrate my attempt at controlled directivity throughout the speakers entire bandwith. Dipole, waveguide, and of course, a idiot Ring radiator, ugly cousin of the dome.
Take from it what you may.
BTW, Paul, to address your question directly, beaming is generally a bad thing, but like all things, represents part of your design that has to be addressed one way or another.
I'm from the school of blame the designer, not the parts (domes,etc.) or enclosure type (sealed box,ported,TL,OB,etc).
Look within first.
Cheers,
AJ
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.
** EDIT - And off course SY posts something as I slooowwwlly typed!
Konnichiwa,
It should also be noted that, if accurate reproduction of the recording is the goal and if basic acoustics are considered there is only one specific set of set of parameters that achieve this in acoustically "live" and acoustically "small" rooms, namely a set of parameters that minimises room interactions.
Of course, accurate reproduction of the recording may not be desired. Then other sets of parameters are required.
Competence is in selecting the right parameters for the desired result. In that sense the Bose 901 is a by far more competent design than the common 6.5" woofer & 1" dome tweeter "High Fidelity" speaker.
Sayonara
SY said:
It should also be noted that, if accurate reproduction of the recording is the goal and if basic acoustics are considered there is only one specific set of set of parameters that achieve this in acoustically "live" and acoustically "small" rooms, namely a set of parameters that minimises room interactions.
Of course, accurate reproduction of the recording may not be desired. Then other sets of parameters are required.
Competence is in selecting the right parameters for the desired result. In that sense the Bose 901 is a by far more competent design than the common 6.5" woofer & 1" dome tweeter "High Fidelity" speaker.
Sayonara
Konnichiwa,
And it is as if some people cannot read. I have used the same terms earlier in the thread.
Perhaps I should have qualified this as "Dome tweeter without waveguide", except such a device is generally known as "Horn". ;-)
Note that I do not direct my comments at anyone specifically, but the industry in general.
In my view SL would have been well advised to use a Ribbon with a waveguide to limit horizontal dispersion, in order to attain a more uniform dispersion. What the fact that he does not do so suggests I leave to others.
Next time you are in London, drop by. Commercially our (realhi-fi.com) descition was to promote Hyperion Speakers, as they invariably stand a much better chance to sell to average punters. Thus my own designs are sidelined (on my own suggestion).
Funny.
My approach is very similar. Just like you I use a hornloaded ring-radiator (in my case a smaller voicecoil design with a tighter radiation pattern, but in principle the same) and a cone driver (arguably a larger one) on an open baffle.
My crossover to the Sub-Woofers is lower and I use a sealed sub, as in my room the crossover between resonant and pressure mode of the room falls near that crossover anyway.
I also note that your approach differs in critical areas from that of SL. So you agree in your own designs that he is in fact wrong in his applications.
I am too. Hence my comments on the use of the dome. It is not the dome that is to blame, it is the designers use of said device.
Sayonara
AJinFLA said:
And it is as if some people cannot read. I have used the same terms earlier in the thread.
AJinFLA said:
Perhaps I should have qualified this as "Dome tweeter without waveguide", except such a device is generally known as "Horn". ;-)
Note that I do not direct my comments at anyone specifically, but the industry in general.
In my view SL would have been well advised to use a Ribbon with a waveguide to limit horizontal dispersion, in order to attain a more uniform dispersion. What the fact that he does not do so suggests I leave to others.
AJinFLA said:
Next time you are in London, drop by. Commercially our (realhi-fi.com) descition was to promote Hyperion Speakers, as they invariably stand a much better chance to sell to average punters. Thus my own designs are sidelined (on my own suggestion).
AJinFLA said:
Funny.
My approach is very similar. Just like you I use a hornloaded ring-radiator (in my case a smaller voicecoil design with a tighter radiation pattern, but in principle the same) and a cone driver (arguably a larger one) on an open baffle.
My crossover to the Sub-Woofers is lower and I use a sealed sub, as in my room the crossover between resonant and pressure mode of the room falls near that crossover anyway.
I also note that your approach differs in critical areas from that of SL. So you agree in your own designs that he is in fact wrong in his applications.
AJinFLA said:
I am too. Hence my comments on the use of the dome. It is not the dome that is to blame, it is the designers use of said device.
Sayonara
A question for AJinFLA on midrange baffle width....
Xover compensation and dispersion patterns
AJ, you used a very narrow dipole baffle for your midrange and I would like to understand your reasons.
With a Xover ~110 Hz to the dipole bass, my first design test would be to use a midrange baffle width that was 6db down at this 110 Hz Xover point, something like 18"-20" wide and similar height. The design goal would be to avoid Xover baffle compensation to the midrange, and possibly simplify the combined acoustic + electrical Xover between midrange and woofer.
Xover compensation and dispersion patterns
AJ, you used a very narrow dipole baffle for your midrange and I would like to understand your reasons.
With a Xover ~110 Hz to the dipole bass, my first design test would be to use a midrange baffle width that was 6db down at this 110 Hz Xover point, something like 18"-20" wide and similar height. The design goal would be to avoid Xover baffle compensation to the midrange, and possibly simplify the combined acoustic + electrical Xover between midrange and woofer.
Member
Joined 2003
Here is an example of an intentionally "beamy" speaker. Vertical dispersion is limited to reduce room reflections. At least for now, dipole for further reduction. Mid-tweet waveguides can also be changed.
As Kuei says, removing the room allows me to hear more detail in the performance. But...everyone is entitled to their own preferences.
Paul
As Kuei says, removing the room allows me to hear more detail in the performance. But...everyone is entitled to their own preferences.
Paul
An externally hosted image should be here but it was not working when we last tested it.
Actually I agree with that. The One Way™ is to dispense with speakers entirely and go with binaural recordings and earphones. The fidelity of such a system far surpasses anything that speakers will ever be capable of since all room effects are removed. Indeed, it can be indistinguishable from reality. Of course it's impractical, since recordings must be personalized.SY said:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.