CeramicMan said:
I agree that the difference tones is not the optimal way of showing that there is a doppler effect, since these can come about from non-linearities. I would like to see a measurement where the low frequency (100Hz) has been filtered out, and where the period time of the high frequency (2000Hz) is extracted for every period.
CeramicMan said:
It has nothing (well...) to do with the movement of the air, it is the movement, or rather the velocity, of the source that is important.
Seen in short term, at the time when the low frequency passes zero, the source moves towards (or away from) the listener. The velocity of the cone will cause a doppler effect, just as there is a doppler effect when the ambulance passes you on the street. The movement of the air particles isn't great in the ambulance case either, right?
The only way I can think of for the frequency of the 1kHz tone to actually change when a 100Hz tone is applied is when there is a design problem in the measurement setup. If a free-field microphone is rigidly fixed just a few centimetres from the offending cone, then yes, of course there will be Doppler distortion because air will be rushing past the microphone at 100Hz and the time taken for the 1kHz waves pulsate against it will vary. Such a setup is faulty because it relies on extremely high amplitudes that do not adequately reflect a normal listening environment. The microphone may even pick up wind noise.
Lets say that a certain measurement at 1 metre distance picks up a 100Hz tone with an air-displacement of +-10mm (maybe a machine-gun would be that loud?) Then the wavelength of a 344Hz tone (from the same perfect source) would change between 1010mm and 990mm, and the frequency would therefore change between 340.6Hz and 347.5Hz. At 10 metres distance, the air displacement of the 100Hz tone would be only +-0.1mm and the measurements of the "340.6Hz - 347.5Hz" warble tone would stabilize to "343.96Hz - 344.03Hz". This is assuming that all relevant frequencies travel through air at the same speed, or is someone is suggesting that some frequencies travel faster than others?
Besides, if you are listening to some soundz at such volumes, then the effects described above are not your biggest problem.
Notice that when an ambulance rushes past, the ambulance moves past the air, rather than compressing it. A more interesting and informative microphone setup than described above would be to place the microphone and speaker at opposite ends inside a sealed tube. If the theoretical speaker could then compress the air in the tube to half of its original volume, it would be interesting to see how long it takes for sound waves to travel half the distance. That would be a real measurement of distortion from the Doppler Effect. It would take into account changes in the speed of sound caused by changes like temperature, humidity, and of course pressure.
CM
Lets say that a certain measurement at 1 metre distance picks up a 100Hz tone with an air-displacement of +-10mm (maybe a machine-gun would be that loud?) Then the wavelength of a 344Hz tone (from the same perfect source) would change between 1010mm and 990mm, and the frequency would therefore change between 340.6Hz and 347.5Hz. At 10 metres distance, the air displacement of the 100Hz tone would be only +-0.1mm and the measurements of the "340.6Hz - 347.5Hz" warble tone would stabilize to "343.96Hz - 344.03Hz". This is assuming that all relevant frequencies travel through air at the same speed, or is someone is suggesting that some frequencies travel faster than others?
Besides, if you are listening to some soundz at such volumes, then the effects described above are not your biggest problem.Notice that when an ambulance rushes past, the ambulance moves past the air, rather than compressing it. A more interesting and informative microphone setup than described above would be to place the microphone and speaker at opposite ends inside a sealed tube. If the theoretical speaker could then compress the air in the tube to half of its original volume, it would be interesting to see how long it takes for sound waves to travel half the distance. That would be a real measurement of distortion from the Doppler Effect. It would take into account changes in the speed of sound caused by changes like temperature, humidity, and of course pressure.
CM
Furthermore, I suggest that the so-called Doppler distortion is locked to the measured amplitude of the test-tones. So 2 theoretical drivers, one with a much smaller diameter but higher excursion than the other will both have an identical effect. This is assuming that the drivers are both small enough to be effectively omni-directional.
CM
CM
That is not the case. The modulation effect on the effective instantaneous change in position of the virtual source of the higher frequency is independent of distance listened to but reduces as the angle of measurement is increased away from the axis of sound generation up to 90 degrees.
More than a few mm of excursion will quite noticeably degrade midrange and high frequency sound reproduction handled by the same driver. The effects of this modulation have been likened to vocalists gargling or of musical instrument sound being listened to through a fan.
More than a few mm of excursion will quite noticeably degrade midrange and high frequency sound reproduction handled by the same driver. The effects of this modulation have been likened to vocalists gargling or of musical instrument sound being listened to through a fan.
Frequency modulation doesn't sound like that anyway. I've experimented with FM synthesis on old SoundBlaster clones to know that for sure. FM has to be pretty drastic before any deviation from a sine wave is audible, much more drastic than plus-minus a few dozen Hertz that might get measured close-up at extremely high volume. What you described is exactly how I would describe amplitude modulation caused by harmonic distortion.thoriated said:
The amplitude modulation is only called "intermodulation" distortion because at a loud low frequency the voice-coil moves in and out of the magnetic gap enough to modulate its effective strength. Then if the same speaker tries to reproduce high frequency sounds, their amplitude will vary at the rate of the low frequency. Hence the low frequency "modulates" the amplitude of relatively high frequencies. This is caused directly by harmonic distortion, and yes: I've heard it too, and it does sound like gargling.
CM
My speakers are the worse culprits here
I find this question of modulation distortion to be an interesting one. My own design uses 2" diameter drive units running full-range, so any ill-effects should be readily apparent.
Kelticwizard was kind enough to send me a copy of Paul Klipsch's paper for the Journal of the Audio Engineering Society: "Modulation Distortion in Loudspeakers". I have enormous respect for the work of Mr. Klipsch and I'm certainly not capable of debating the theoretical existence of Doppler distortion or FM distortion as he outlines it in this paper. The important question to me is how does it sound in practice?
Paul Klipsch's paper begins with an initial observation where two loudspeakers, one with direct radiator bass and the other with horn-loaded bass were compared in listening tests by 20 people. The drivers were identical as were the horn-loaded midrange and treble drivers. Despite the direct radiator offering a more extended bass with a cutoff of 30Hz compared with 45Hz for the horn-loaded bass, 19 of his 20 listeners thought that the one with horn-loaded bass was 'cleaner'. They used terms like "transparent", "better resolution" and "clarity of the inner voices". Both speakers were shown to be generating negligible harmonic distortion and had accurate power linearity through the frequency range. Paul Klipsch concluded that the cause of the difference in listening quality must be modulation distortion.
I'd like to contribute to this debate by offering a few observations about my own speakers (with the 2" full-range drivers), which have been described by a number of people as transparent, detailed and fast. Paul Messenger, in his recent review for Hi-Fi+ magazine, after commenting on the "very superior image definition, focus and precision compared to conventional loudspeakers", wrote: "The second very obvious thing here was the superior clarity and intelligibility of voices". He described the speakers as having "exceptional vocal clarity". This is despite the fact that the same 2" drivers dishing out the vocals are also on full bass duty down to 20Hz.
This is not my crude attempt to sell Seventh Veil speakers on this forum. I'm sure that there are no potential customers amongst our members here. I'm just trying to make the point that they cannot be readily classified as suffering in clarity or transparency compared to similarly sized conventional speakers with crossovers, despite the fact that they should be ideal candidates for higher doppler and modulation distortion.
When I initially tested the prototype speakers I used an A.R. active tube crossover for filtering the signal to the matching subwoofers. I tested using the main speakers full-range and filtered at 12db/octave below 100Hz. To my ears (and to my wife and friends) the sound was clearer and better without a filter on the full-range speakers.
What does all this mean with regard to doppler and modulation distortion? Perhaps the audible effects of these distortions are less than the degradation that a crossover unit or filter would cause. To be honest I really don't know.
I would make one further observation though. My speakers were designed down to a specific size. Compared with some much larger speakers I have heard they sound as if they have less space and are more compressed . However, I believe this is also true of other speakers of a similar size. Is this compression or 'lack of spaciness' due to modulation distortion?
Perhaps.
I find this question of modulation distortion to be an interesting one. My own design uses 2" diameter drive units running full-range, so any ill-effects should be readily apparent.
Kelticwizard was kind enough to send me a copy of Paul Klipsch's paper for the Journal of the Audio Engineering Society: "Modulation Distortion in Loudspeakers". I have enormous respect for the work of Mr. Klipsch and I'm certainly not capable of debating the theoretical existence of Doppler distortion or FM distortion as he outlines it in this paper. The important question to me is how does it sound in practice?
Paul Klipsch's paper begins with an initial observation where two loudspeakers, one with direct radiator bass and the other with horn-loaded bass were compared in listening tests by 20 people. The drivers were identical as were the horn-loaded midrange and treble drivers. Despite the direct radiator offering a more extended bass with a cutoff of 30Hz compared with 45Hz for the horn-loaded bass, 19 of his 20 listeners thought that the one with horn-loaded bass was 'cleaner'. They used terms like "transparent", "better resolution" and "clarity of the inner voices". Both speakers were shown to be generating negligible harmonic distortion and had accurate power linearity through the frequency range. Paul Klipsch concluded that the cause of the difference in listening quality must be modulation distortion.
I'd like to contribute to this debate by offering a few observations about my own speakers (with the 2" full-range drivers), which have been described by a number of people as transparent, detailed and fast. Paul Messenger, in his recent review for Hi-Fi+ magazine, after commenting on the "very superior image definition, focus and precision compared to conventional loudspeakers", wrote: "The second very obvious thing here was the superior clarity and intelligibility of voices". He described the speakers as having "exceptional vocal clarity". This is despite the fact that the same 2" drivers dishing out the vocals are also on full bass duty down to 20Hz.
This is not my crude attempt to sell Seventh Veil speakers on this forum. I'm sure that there are no potential customers amongst our members here. I'm just trying to make the point that they cannot be readily classified as suffering in clarity or transparency compared to similarly sized conventional speakers with crossovers, despite the fact that they should be ideal candidates for higher doppler and modulation distortion.
When I initially tested the prototype speakers I used an A.R. active tube crossover for filtering the signal to the matching subwoofers. I tested using the main speakers full-range and filtered at 12db/octave below 100Hz. To my ears (and to my wife and friends) the sound was clearer and better without a filter on the full-range speakers.
What does all this mean with regard to doppler and modulation distortion? Perhaps the audible effects of these distortions are less than the degradation that a crossover unit or filter would cause. To be honest I really don't know.
I would make one further observation though. My speakers were designed down to a specific size. Compared with some much larger speakers I have heard they sound as if they have less space and are more compressed . However, I believe this is also true of other speakers of a similar size. Is this compression or 'lack of spaciness' due to modulation distortion?
Perhaps.
ucla88 said:
Unscientific ? Do me a favour. The whole analysis starts off
by presuming it occurs, rather than showing somehow that
it is inevitable. I can't find one reputable published paper
illustrating the effect with test results.
http://www.geocities.com/kreskovs/Doppler2.html
Just shows standard intermodulation products
due to non-linearity, f2+f1 and f2-f1.
And no evidence whatsoever of the higher frequency shifting
related to the excursion level of the lower frequency.
You'd think a test that would show frequencies not related
to IM distortion would be chosen, but it can't be done.
quoting rod elliot :
My thinking on this issue was recently challenged (thanks Chuckand we did a bit of e-mail exchange, and some demonstartion calculations, and I ran some tests on a speaker in my workshop. The simple fact of the matter is that doppler distortion is a furphy, not quite a lie, but a calculated twisting of the truth ....
... so, while it can be "proven", further analysis shows the real truth of the matter - the "bad guy" is intermodulation, and the good Prof. Doppler can rest easy, since the effects we hear are not related to his discovery of the Doppler effect at all. Further reading on this matter will be forthcoming, but for now, assume that Doppler "distortion" does not exist (despite the snake oil liberally spread over the topic by various individuals who shall remain nameless, but have also annointed cables with their alleged "magic" - charlatanism at its very finest
[Update added 30 Jul 2002]
sreten.The modulation is simply a result of the reproducing surface's instantaneous change in distance from the listener's ears because of modulation from a generally lower frequency signal reproduced by the same driver that requires significant cone excursion.
Consider this also: I have heard a somewhat related effect due to the entire speaker rocking back and forth (visibly by as much as a quarter inch p-p at the top in the case of a floor standing speaker on carpet) when playing higher energy bass transients. This rocking produced significant image instability and smearing of inner musical detail. The difference here is that even drivers reproducing no lf energy at all are affected (eg, the tweeters and midranges).
Think also that image lateral stability can be affected by out of phase lf energy moving cones in opposite directions while both are reproducing higher frequency info. The effect can be, depending on setup and listening geometry as if the acoustic center at the listening position is moving laterally by an inch or two which is far from ideal.
I've distinctly heard the 'fan' modulation effects with low distortion JBL 2226's working well within their linear Xmax (7.6mm one way) and crossed over at a relatively low 700hz.
This effect is also quite measurable and explains the popularity of horn loading in professional applications. Two tone measurements have shown doppler sidebands amounting to 10-20% intermodulation distortion as measured on spectral analysis gear when measuring bass reflex pro sound equipment running above its br tuning frequency & well within its power capability and excursion limits whereas the exact same drivers, but horn loaded at the same reproduced levels produce only a fifth or less as much such distortion.
sreten said:
Just shows standard intermodulation products
due to non-linearity, f2+f1 and f2-f1.
And no evidence whatsoever of frequency shifting related
to the excursion of the lower frequency.
You'd think a test that would show frequencies not related
to IM distortion would be chosen, but it can't be done.
Sreten,
I agree that there should be a better way to show this, and maybe I'll try if I find the time. Meanwhile, could you tell me what is wrong with this reasoning:
If a sound source moves towards me (like in the classic ambulance example) the fundamental frequency (F0) of the sound is shifted upwards by c/(c-v), where c is the speed of sound and v is the speed of the ambulance in the direction towards me. The factor is >1 if it moves against me, and <1 if it moves away from me.
Now, replace the ambulance with a very-long-throw loudspeaker. Let's say for the sake of the argument that it can move a meter or so. Let's also say that we feed it with a signal such that it moves with a constant velocity towards me plus a 2 kHz sinusoid.
Will there be a doppler shift now? If not, what is the difference between this and the ambulance?
OK, so finally, look at the normal loudspeaker, driven with 100 Hz and 2 kHz simultaneously. Take a short time window (~2 ms) when the cone has its maximum velocity towards me. Why should not the 2 kHz tone suffer from a doppler shift in this case?
Again, note that I don't say anything about its audibility and/or what magnitude this "problem" has, just that the effect should exist.
thoriated said:
The Xmax of a speaker doesn't define any linear range, it just gives an indication that above the Xmax the THD will be above some threshold like 10% or even 30%. THD directly causes one form of IMD, therefore horn design => lower cone excursion => lower THD => lower IMD. IMD from amplifiers is usually measured electrically, without relying on any acoustical effects from transducers.
I'm still not convinced that FM from the Doppler Effect is as simple as measuring the distance travelled by the transducer or from its velocity. The difference between the ambulance example and a transducer is that the air moves around the ambulance so the sound has progressively less and less air to travel through, whereas the transducer compresses the same number of particles into a smaller volume of space.
I have to admit I'm not entirely sure exactly what happens, and I think I'll spend some time trying to conjure up a simulation of air-particle motion to see what really happens. The way I see it, a transducer + free-air system acts approximately like a differentiator, ie: the highest cone velocity results in the highest sound pressure. Ever noticed that free-field step response measurements always ends up as impulse responses when viewed on a long enough time scale? Obviously the increased air pressure from a transducer's "step" decays with time, which suggests that there is longitudinal movement of air after the initial high-pressure step. If that is the case then it can be proven that FM modulation decays with distance, (we could use some experts here!)
I don't care that a high frequency sound becomes geometrically "squashed" by high-velocity cone movement, what I do want to know for sure, is: will that high frequency sound gradually become "un-squashed" as it moves further from the cone? Or will it keep the distorted shape despite the fact that the low frequency sound (causing the distortion in the first place) will lose pressure at increased distances?
CM
IMO the model of a "sond source moving at a velocity determined by th elow frequency content" is valid.
BTW: The air around the ambulance will also be compressed in front of it and rarefied behind it. The only difference is that the pressure differences are larger by some orders of magnitude than the pressure changes in front of a speaker cone !!!!!
I do however not know how much influence FMD has but I also assume that IMD is higher. Although both should be avoided.
Regards
Charles
BTW: The air around the ambulance will also be compressed in front of it and rarefied behind it. The only difference is that the pressure differences are larger by some orders of magnitude than the pressure changes in front of a speaker cone !!!!!
I do however not know how much influence FMD has but I also assume that IMD is higher. Although both should be avoided.
Regards
Charles
My simple problem is this : demonstration of the effect.
The mathematical relationships of IM are well known, the
distortion products being f2+f1 and f2-f1 for two tones.
The whole premise of doppler distortion as far as I can tell
is that the upper frequency is spread in range by the
excursion level of the lower frequency, at a rate determined
by this frequency, yet I have never seen the effect.
All demonstrations seem to confirm that IM is the only issue.
My understanding is in a perfect driver f2 and f1 add to form
a waveform and that waveform is produced by the driver.
As far as I understand it this is just accurately reflecting
the properties of the two superimposed waveforms.
I really can't see why this is presumed not to be the case.
Horn loading a driver will simply reduce its IM products.
Mounting one driver on another will cause the effect, as in this
case one driver is changing the relative time axis of the other,
ther waveform produced will not be the linearly superimposed
result of adding the two waveforms.
Descriptions of the effect seem to think the above analogy is
applicable to a single driver producing two tones, and its here
that I lose all confidence in the whole premise.
sreten.
The mathematical relationships of IM are well known, the
distortion products being f2+f1 and f2-f1 for two tones.
The whole premise of doppler distortion as far as I can tell
is that the upper frequency is spread in range by the
excursion level of the lower frequency, at a rate determined
by this frequency, yet I have never seen the effect.
All demonstrations seem to confirm that IM is the only issue.
My understanding is in a perfect driver f2 and f1 add to form
a waveform and that waveform is produced by the driver.
As far as I understand it this is just accurately reflecting
the properties of the two superimposed waveforms.
I really can't see why this is presumed not to be the case.
Horn loading a driver will simply reduce its IM products.
Mounting one driver on another will cause the effect, as in this
case one driver is changing the relative time axis of the other,
ther waveform produced will not be the linearly superimposed
result of adding the two waveforms.
Descriptions of the effect seem to think the above analogy is
applicable to a single driver producing two tones, and its here
that I lose all confidence in the whole premise.
sreten.
sreten said:
Agree. This would be the best.
sreten said:As far as I understand it this is just accurately reflecting
the properties of the two superimposed waveforms.
I really can't see why this is presumed not to be the case.
So what about the ambulance, will there be a doppler shift in that case?
And for the extra-long-throw speaker?
The problem is that you generate two different waveforms by two superimposed motions of a cone. I.e. you definitely don't seperately generate two waveforms and superimpose them linearily afterwards. Horn drivers generate less IMD within reasonable SPL limits thats agreed. But they also reduce FMD due to significantley reduced "cone" displacement and therefore also reduced cone velocity.
Regards
Charles
Regards
Charles
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.