i came across this at Will You Like My Speakers? page
"The nemesis of single-driver speakers is frequency modulation (FM) of the upper ranges by the bass. (Doepler distortion is a type of FM distortion, not the other way around.) A higher fequency becomes a warble tone, the frequency and depth of the warble depending on the frequency and driver displacement of the bass tone. The result is when there are many bass frequecies in the music, as the music becomes loud (more driver excusion) the upper ranges become harsh and tonally indistinct. The driver "falls apart""
does all fullrange drivers suffer this? or are there any exceptions?
any technique to tame this ?
"The nemesis of single-driver speakers is frequency modulation (FM) of the upper ranges by the bass. (Doepler distortion is a type of FM distortion, not the other way around.) A higher fequency becomes a warble tone, the frequency and depth of the warble depending on the frequency and driver displacement of the bass tone. The result is when there are many bass frequecies in the music, as the music becomes loud (more driver excusion) the upper ranges become harsh and tonally indistinct. The driver "falls apart""
does all fullrange drivers suffer this? or are there any exceptions?
any technique to tame this ?
As Bob says in the very first sentence: "Full range drivers are not for everyone", and he does a pretty good job of summarizing why.
I think the closing section answers your question:
(keep the change)
The mitigation of IM distortion effects is one of the primary reasons why many of us have come out of the closet about the use of "helper woofers" - i.e. what has not become categorized as FAST systems. When implemented with HP filters on the FR drivers, these are really nothing less than a 2-way system - the salient difference being moving the XO point out of the middle of the critical "telephone" band, allowing the FR driver to do all the things that Bob outlines, and that we longtime FR geeks appreciate.
I think the closing section answers your question:
emphasis (and typo corrections) mine, and this will be true of any brand and size of FR driver and enclosure design
(keep the change)The mitigation of IM distortion effects is one of the primary reasons why many of us have come out of the closet about the use of "helper woofers" - i.e. what has not become categorized as FAST systems. When implemented with HP filters on the FR drivers, these are really nothing less than a 2-way system - the salient difference being moving the XO point out of the middle of the critical "telephone" band, allowing the FR driver to do all the things that Bob outlines, and that we longtime FR geeks appreciate.
Member
Joined 2009
Paid Member
Doppler doesn't happen in a speaker driver, it's an oft repeated misunderstanding, like many in audio. There are of course various opportunities for non-linear behaviour in a speaker that can cause IM products and are mistakenly attributed to Doppler. However, I have generally found full range drivers to have only one issue - cone break up modes, otherwise they sound superb. And not all drivers have cone break up problems to be worried about. Parasitic cone break up oscillations could be subject to Doppler shift.
Last edited:
Well, as Rod Elliot (and no doubt others) would explain, the incorrect use of a technical term (as short-hand) does not eliminate the occurrence of a phenomena that could be attributed as "similar to .."
Doppler Distortion in loudspeakers
IINM, Full Range loudspeakers work over as wide a range as they do in no small part due to the designer's understanding, and careful use of what I think you're describing as "break-up" modes. An absolutely rigid "pistonic" cone driven by the conventional moving coil method would have rather limited frequency response; for example those massive car sub woofers with thick metal cones strong enough to support more than their own weight.
and doesn't your last sentence contradict your first?
Doppler Distortion in loudspeakers
IINM, Full Range loudspeakers work over as wide a range as they do in no small part due to the designer's understanding, and careful use of what I think you're describing as "break-up" modes. An absolutely rigid "pistonic" cone driven by the conventional moving coil method would have rather limited frequency response; for example those massive car sub woofers with thick metal cones strong enough to support more than their own weight.
and doesn't your last sentence contradict your first?
AMD/FMD exists. Doppler is essentially just a convenient term, a little like 'TL', accurate or otherwise. Problematic? They can (as in 'can') be, but depends on the listener, the material, the enclosure, the room, average & peak SPLs, listening distance... the list goes on. Same goes for cone breakup. As Chris points out, there's breakup & there's breakup. Most of the BW covered by fullrange drive units is created by resonant, not oscillatory, action (i.e. 'breakup'). Uncontrolled breakup is another matter entirely, and not exactly something that is desireable.
I can't see how there is any misuse of "Doppler" in connection with IMD from a physically moving speaker diaphragm - calling it FM or PM, the zero mean cyclic motion of the diaphragm doesn't change the physics
the linear model of the driver's motion in response to signal works just fine - it is the step where you convert the extended physical motion into sound radiation that most people make the mistake
the cheap assumption is that the driver motion causes an ideal air velocity modulation in its nominal, immobile position
but the diaphragm really does move, does modify the radiation equation from the idealized "linear" assumption
I have seen people who occasionally try clubbing you with “argument from authority” citing their physics backgrounds double think themselves into bad corners on Doppler from speaker cones though
despite the ability to find papers measuring the effect, correlating with the equations...
the linear model of the driver's motion in response to signal works just fine - it is the step where you convert the extended physical motion into sound radiation that most people make the mistake
the cheap assumption is that the driver motion causes an ideal air velocity modulation in its nominal, immobile position
but the diaphragm really does move, does modify the radiation equation from the idealized "linear" assumption
I have seen people who occasionally try clubbing you with “argument from authority” citing their physics backgrounds double think themselves into bad corners on Doppler from speaker cones though
despite the ability to find papers measuring the effect, correlating with the equations...
Member
Joined 2009
Paid Member
I see it like this.
An idealized speaker driver has a perfectly rigid cone that moves in response to the music signal. The signal drives a current through voice-coil which interacts with the fixed magnetic field of the permanent magnet. So the cone moves relative to the fixed magnet. Both the high and low frequency signals are translated into motion of the cone relative to the fixed magnet. The time-average position of the cone is fixed. The cone + magnet form the radiating source and it's position is stationary. There is no Doppler shift in the traditional sense because the sound source is not moving. The classic example is a Fire truck siren that changes pitch as it comes towards you and then away from you. If this siren were a speaker driver, the whole thing magnet + cone is in motion toward and then away from you. For a stationary speaker there is no motion of the source.
If there were a sound source on the cone surface, say a small tweeter, then a low frequency vibration of the main cone would cause the tweeter to move back and forth and it would exhibit Doppler distortion. Cone break up modes are parasitic vibrations of the cone surface, they are stimulated by the music but they exist in the fabric of the cone and are not movement referenced to the fixed magnet but referenced to the cone surface. In this way, sound from cone break up is like the small tweeter glued to the cone, the source will move with the cone and Doppler distortion is possible.
It's not a case of fancy physics, its about separating out the movement of the source from the movement of the parts that constitute the source.
There are several sources of non-linear behaviour in a real speaker system and so it is conceivable that measurements of those sources might have similarities to the signature expected from Doppler distortion.
Last edited:
Idealized to keep the math used in analysis in the 5% of the problems that can be done with "regular" prblem solving techniques. This does not make it an ideal loudspeaker.
dave
Gee, I can remember when one of the models for an "idealized" speaker was a massless pulsating sphere - which to my thinking at the time (yes, that was late mid last century, so some "perception enhancement modalities" may have been involved, and perhaps are still flashing back ) suggested elasticity of the surface, if not imaginary materials and energy sources, and certainly not perfect rigidity
I can see your point about parasitic vibrations/resonances in cone (or other diaphragm type) being unrelated to the signal induced movement of the voice coil (or alternative motive method), but does not the type of wide frequency bandwidth that FR drivers such as under discussion here are targeted to achieve rely on a certain amount of controlled decoupling of emitting surfaces? as well as their profiles, variations in material density, perimeter termination / damping , and no doubt other factors?
Perhaps not (see first parenthesis above)
) suggested elasticity of the surface, if not imaginary materials and energy sources, and certainly not perfect rigidity I can see your point about parasitic vibrations/resonances in cone (or other diaphragm type) being unrelated to the signal induced movement of the voice coil (or alternative motive method), but does not the type of wide frequency bandwidth that FR drivers such as under discussion here are targeted to achieve rely on a certain amount of controlled decoupling of emitting surfaces? as well as their profiles, variations in material density, perimeter termination / damping , and no doubt other factors?
Perhaps not (see first parenthesis above)
this is the key - as I said the diaphragm motion is a linear function of the drive signal to a very high degree of accuracy
so the high frequency motion is exactly like "a sound source on the cone surface, say a small tweeter" - by superposition
and the position and velocity of this high frequency "source" is moving with the motion of the diaphragm from the bass - so it is Doppler shifted
the "nonlinearity" causing the FM/PM/Doppler distortion is in the conversion of the linear cone motion into sound - over the varying position/displacement of the cone in the air - the high frequencies have no way to "know" the average position of the cone - they radiate from it wherever it is in its bass driven spatial position/velocity curve
there is no way to separately claim "breakup modes" are shifted and some other cone motion isn't
Rod Elliot gets it right eventually even if his measurements aren't terribly clear - they are measurement and evidence that "naysayers" have to explain
Last edited:
well, obviously, some (myself certainly encluded) enjoy the babble more than anything
its nice to get the basic physics of loudspeaker sound production right
builds some confidence when a speaker manufacturer seems to be getting it right
but the question of when Doppler distortion from a single driver becomes audible/objectionable is as the 1st post stated both music level and frequency content related
builds some confidence when a speaker manufacturer seems to be getting it right
but the question of when Doppler distortion from a single driver becomes audible/objectionable is as the 1st post stated both music level and frequency content related
Member
Joined 2009
Paid Member
Hi Fellas,
There's a good number of very interesting comments being produced but errors are also being made.
1 - The perfect cone is Not rigid. Most driver designers working on bass-wide and wide band classes will have their own flex ratio and mass/transmission formulas.
2 - Signal output transmission modulation variation is Not the sole "nemesis" for/of wide band drivers. Having tested several sub woofs, woofs and bass-mids over the years, the LF signal reception variance is significant in many of these drivers, certainly audible in the many anechoic and music tests we've made. Most of this testing was done prior to the design and production of Markaudio's Woofer No. 6 (EL166) and were instrumental in the formulation of its design.
Historically, the design of most wide-band and full-range driver were based on short stroke designs. I strongly suspect that the issue of modulation has become confused with the LF output limitations of these drivers.
In the case of Markaudio units, many of the Japanese testers (Ozawa San for MJ Magazine, Matsumoto san of Lean Audio, Clarke san etc etc) have cottoned on the benefits of Markaudio's long linear throw, low mass power train design; Observing little of no audible differentiation in musical output on a typical mixed frequency spectrum that includes LF loadings.
Bob Brines himself has made several recent references to the practical application of applying LF gain loads to Markaudio drivers. I believe the reference to Bob's page posted the thread starter should be read in the context of older wide-band and full-range driver operational designs.
To help out, I'd suggest the following guide to getting a good bass response without any "fluffiness' becoming audible:
For single point source projects, work on the basis on only deploying around 50% of the driver's quoted X-max for extended LF loads. Make use of the more efficient box concepts (Pensils being a prime example - Thanks to Dr. Scott) that minimise LF generation losses. Size your project according to your room requirements. Don't expect small drivers to cope well in large room situations.
These sorts of issues are interesting and get guys thinking and debating, but I can see that in some comments, the lack of knowledge on the operational properties of drivers is particularly evident. By all means guys, keep posting while applying more thinking and research time before committing to the "bold" statement.
Any comments that are severely technically in-accurate to the point they mis-lead others, I'll have to delete them, as many folks rely on the Markaudio section for reasonable technical accuracy, to help them learn better and get more from their projects.
Cheers
Mark.
There's a good number of very interesting comments being produced but errors are also being made.
1 - The perfect cone is Not rigid. Most driver designers working on bass-wide and wide band classes will have their own flex ratio and mass/transmission formulas.
2 - Signal output transmission modulation variation is Not the sole "nemesis" for/of wide band drivers. Having tested several sub woofs, woofs and bass-mids over the years, the LF signal reception variance is significant in many of these drivers, certainly audible in the many anechoic and music tests we've made. Most of this testing was done prior to the design and production of Markaudio's Woofer No. 6 (EL166) and were instrumental in the formulation of its design.
Historically, the design of most wide-band and full-range driver were based on short stroke designs. I strongly suspect that the issue of modulation has become confused with the LF output limitations of these drivers.
In the case of Markaudio units, many of the Japanese testers (Ozawa San for MJ Magazine, Matsumoto san of Lean Audio, Clarke san etc etc) have cottoned on the benefits of Markaudio's long linear throw, low mass power train design; Observing little of no audible differentiation in musical output on a typical mixed frequency spectrum that includes LF loadings.
Bob Brines himself has made several recent references to the practical application of applying LF gain loads to Markaudio drivers. I believe the reference to Bob's page posted the thread starter should be read in the context of older wide-band and full-range driver operational designs.
To help out, I'd suggest the following guide to getting a good bass response without any "fluffiness' becoming audible:
For single point source projects, work on the basis on only deploying around 50% of the driver's quoted X-max for extended LF loads. Make use of the more efficient box concepts (Pensils being a prime example - Thanks to Dr. Scott) that minimise LF generation losses. Size your project according to your room requirements. Don't expect small drivers to cope well in large room situations.
These sorts of issues are interesting and get guys thinking and debating, but I can see that in some comments, the lack of knowledge on the operational properties of drivers is particularly evident. By all means guys, keep posting while applying more thinking and research time before committing to the "bold" statement.
Any comments that are severely technically in-accurate to the point they mis-lead others, I'll have to delete them, as many folks rely on the Markaudio section for reasonable technical accuracy, to help them learn better and get more from their projects.
Cheers
Mark.
Last edited:
I am just addressing "Doppler" FM, PM distortion from an ideal speaker - assuming for argument that motor/suspension/diaphragm are all "ideal", giving linear, uniform motion of the diaphragm with respect to electrical input- which may not be "ideal" for a full range driver which has to control sensitivity and directivity by clever use/balancing of "non ideal" material parameters
Klippel measures loudspeakers - includes Doppler distortion in their zoo of driver distortion mechanisms - show how to differentiate, recognize "Doppler" in measurement
Klippel measures loudspeakers - includes Doppler distortion in their zoo of driver distortion mechanisms - show how to differentiate, recognize "Doppler" in measurement
http://www.klippel.de/fileadmin/kli...linearities?Causes,Parameters,Symptoms_06.pdf
http://www.klippel.de/fileadmin/kli...es/AN_10_Loudspeaker_FM_and_AM_Distortion.pdf
Last edited:
Member
Joined 2009
Paid Member
I'm glad that Mark is here for us to make great sounding drivers without me needing to wade through the math
Describing it as phase modulation turned on some light bulbs for me (as opposed to Doppler) - I think I can see where phase modulation can arise - the creation of successive peaks of the sound waves can occur when the cone surface is closer or further from the listener and this difference in distance translates to a difference in phase of the peaks. The distance the cone moves however is typically tiny so the phase modulation is also tiny and there are other distortion contributors in any speaker that are more important to pay attention to than this.
Therefore, it seems to me that there is nothing to be concerned about - on the contrary, using a full range driver, means there is no cross-over which means it will have orders of magnitude better phase integrity over multi-way speakers.
Describing it as phase modulation turned on some light bulbs for me (as opposed to Doppler) - I think I can see where phase modulation can arise - the creation of successive peaks of the sound waves can occur when the cone surface is closer or further from the listener and this difference in distance translates to a difference in phase of the peaks. The distance the cone moves however is typically tiny so the phase modulation is also tiny and there are other distortion contributors in any speaker that are more important to pay attention to than this.
Therefore, it seems to me that there is nothing to be concerned about - on the contrary, using a full range driver, means there is no cross-over which means it will have orders of magnitude better phase integrity over multi-way speakers.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.