You give your account of how the sound is generated. (We agree that we are talking about the generation of sound).
But is not the Doppler effect THE EFFECT?
It doesn't matter that the sound is generated by this that or the other, or according to some model of sound generation.
If the sound is being generated it is being generated.
In our case, it's a vibrating piece of paper.
We seem to have regressed in the discussion.
I'm perfectly happy that a small disc in free air playing a combined high frequency and near DC frequency on a thought-experiment style linear motor produces Doppler shift - I was using this as incontrovertible proof of Doppler myself a few pages back.
However, I'm also happy that a tight fitting truck advancing down a tunnel, horn blaring, pushing all the air in its path ahead of it does not produce Doppler, for a stationary listener. There is no 'bunching' of the horn sound, because the medium itself is being pushed at the same speed as the horn's displacement.
Real speakers, I think, are somewhere between these two extremes, and are nearer to the no-Doppler situation.
Why do speakers have to be big? Why do they have to be in a box? I suspect that a speaker that suffered from Doppler would be a useless speaker for many other more important reasons than Doppler, and would simply not be built.
N.B. I plead guilty for having started this discussion, and having completely changed my mind about the issue over the course of it!
I'm perfectly happy that a small disc in free air playing a combined high frequency and near DC frequency on a thought-experiment style linear motor produces Doppler shift - I was using this as incontrovertible proof of Doppler myself a few pages back.
However, I'm also happy that a tight fitting truck advancing down a tunnel, horn blaring, pushing all the air in its path ahead of it does not produce Doppler, for a stationary listener. There is no 'bunching' of the horn sound, because the medium itself is being pushed at the same speed as the horn's displacement.
Real speakers, I think, are somewhere between these two extremes, and are nearer to the no-Doppler situation.
Why do speakers have to be big? Why do they have to be in a box? I suspect that a speaker that suffered from Doppler would be a useless speaker for many other more important reasons than Doppler, and would simply not be built.
N.B. I plead guilty for having started this discussion, and having completely changed my mind about the issue over the course of it!
Member
Joined 2009
Paid Member
Except that it does not produce a Doppler shift because the movement of the disc in an idealized case is always moving in a sinusoid with respect to the stationary magnet such that the wavelength is constant when observed by the listener also stationary with respect to the magnet. I'm at a loss to understand why so many people are 'stuck' on this
Last edited:
Except that it does not produce a Doppler shift because the movement of the disc in an idealized case is always moving in a sinusoid with respect to the stationary magnet such that the wavelength is constant when observed by the listener also stationary with respect to the magnet. I'm at a loss to understand why so many people are 'stuck' on this
The HF rides the LF signal. This is observable - simply 'scope the output of an amplifier playing music. Notice how the complete high frequency waves move up and down when the bass comes in.
The cone (being the source of the sound - the magnet does not excite the air) follows this movement.
As the high frequencies are pushed forward and back by the low frequencies, the source of the sound (still the cone) must move towards and away from the listener.
Following this, there has to be some modulation of the high frequency signal because of the low frequency signal.
Chris
Member
Joined 2009
Paid Member
Chris, yes, the HF rides on the LF signal, the actual signal applied to the voice coil is a wiggly shape that contains both frequencies, a direct super-position of them. Yes, you would observe this on a scope and yes the HF will bob up and down with the LF. What we both mean by this is that if we were to take an average of the HF waveform, a smooth line that follows the average HF signal level, it would be a flat line when no LF was present, and it would follow the LF up and down when the LF is applied. Agreed.
And yes, the cone will follow this movement. And yes the magnet does not excite the air, it is the cone. Agreed.
As the LF + HF drive the cone back and forth the cone does move toward and away from the listener. The cone will move slowly back and forth with the LF with a HF vibration super-posed on top of it. Agreed.
In an idealized linear speaker, following this it is tempting to conclude there must be some modulation of the HF. But I disagree with this because of the details involved.
* long post ahead *
What happens when there is only HF applied ?
With an HF signal applied to the voicecoil it drives the cone back and forth relative to the permanent magnet. Each time the cone moves toward the listener (and away from the magnet) it pushes on the air and each time it moves away from the listener (and towards the magnet) it rarefies the air, creating sound waves for the listener. The time that elapses between each 'push' of the cone towards the listener is the period of the HF oscillation. It does not vary between successive oscillations and so the listener perceives a pure tone. If we now grab the speaker driver by it's magnet and move the whole speaker assembly rapidly toward the listener (don't trip over the wires now !) then things are different. After an initial 'push' of air by the cone towards the listener the whole speaker moves closer to the listener (in the hands of the mad experimenter) and the next 'push' starts closer to him/her. As a result it arrives at the listener a little earlier than it would have done had the speaker been stationary. The time period between these two 'pushes' of air is now reduced and the listener experiences the HF as a higher frequency than before when the speaker was sat still. This is Doppler.
What happens when a signal with both LF and HF is applied ?
As before, the signal is applied to the voicecoil and it drives the cone back and forth relative to the permanent magnet. Each time the cone moves toward the listener (and away from the magnet) it pushes on the air and each time it moves away from the listener (and towards the magnet) it rarefies the air, creating sound waves for the listener. With the mixed signal the cone is moving in a complex pattern, but it nevertheless pushes air toward the listener at each peak of the HF and each peak of the LF. Let's pay attention only the HF and look at the behaviour of the cone. Each peak of the HF in the applied signal from the amplifier is separated from subsequent peaks by the period of oscillation and this period is constant. Each of these signal peaks will generate a push of the cone towards the listener (remember the voicecoil is driven to move relative to the permanent magnet). The cone may not be idle during this process, it may be busy following any number of other peaks and valleys in the complex signal, but in an idealized speaker it will faithfully generate a push of air toward the listener for each peak of HF applied to the voicecoil. It will generate each push coicident with each applied signal peak. And so each push of air associated with the HF signal will be separate from successive pushes of air by the period of oscillation, which is constant. The resulting sound wave experienced by the listener will be complex, but the HF component will be made up of pressure waves where each peak was generated by a push of the cone. And because the time period between successive pushes of the cone is constant, the time period of the HF experienced by the listener will also be constant. The frequency will not vary. There is no Doppler effect. Unless we move the whole speaker assembly including the magnet.
Why do some people report measurements indicating a Doppler effect ?
An idealized linear speaker can be described by a constant set of parameters. For example the cone suspension is an ideal linear spring of 'spring constant k'. In a real world non-ideal speaker the movement of the cone causes the 'spring constant' to change once it moves away from it's resting position because the suspension is not a perfect linear spring. There are many other sources of non-linearity in the system. With a source of non-linearity the application of a LF signal means that the parameters of the speaker are changing with time (remember parametric oscillators ?). When you add a HF signal into this situation it no longer has an idealized linear system to translate the signal into sound and it produces distortion. Distortion in speakers is generally much higher than the other components in the audio chain. Such distortion can produce inter-modulation products. They are not Doppler shifts.
And yes, the cone will follow this movement. And yes the magnet does not excite the air, it is the cone. Agreed.
As the LF + HF drive the cone back and forth the cone does move toward and away from the listener. The cone will move slowly back and forth with the LF with a HF vibration super-posed on top of it. Agreed.
In an idealized linear speaker, following this it is tempting to conclude there must be some modulation of the HF. But I disagree with this because of the details involved.
* long post ahead *
What happens when there is only HF applied ?
With an HF signal applied to the voicecoil it drives the cone back and forth relative to the permanent magnet. Each time the cone moves toward the listener (and away from the magnet) it pushes on the air and each time it moves away from the listener (and towards the magnet) it rarefies the air, creating sound waves for the listener. The time that elapses between each 'push' of the cone towards the listener is the period of the HF oscillation. It does not vary between successive oscillations and so the listener perceives a pure tone. If we now grab the speaker driver by it's magnet and move the whole speaker assembly rapidly toward the listener (don't trip over the wires now !) then things are different. After an initial 'push' of air by the cone towards the listener the whole speaker moves closer to the listener (in the hands of the mad experimenter) and the next 'push' starts closer to him/her. As a result it arrives at the listener a little earlier than it would have done had the speaker been stationary. The time period between these two 'pushes' of air is now reduced and the listener experiences the HF as a higher frequency than before when the speaker was sat still. This is Doppler.
What happens when a signal with both LF and HF is applied ?
As before, the signal is applied to the voicecoil and it drives the cone back and forth relative to the permanent magnet. Each time the cone moves toward the listener (and away from the magnet) it pushes on the air and each time it moves away from the listener (and towards the magnet) it rarefies the air, creating sound waves for the listener. With the mixed signal the cone is moving in a complex pattern, but it nevertheless pushes air toward the listener at each peak of the HF and each peak of the LF. Let's pay attention only the HF and look at the behaviour of the cone. Each peak of the HF in the applied signal from the amplifier is separated from subsequent peaks by the period of oscillation and this period is constant. Each of these signal peaks will generate a push of the cone towards the listener (remember the voicecoil is driven to move relative to the permanent magnet). The cone may not be idle during this process, it may be busy following any number of other peaks and valleys in the complex signal, but in an idealized speaker it will faithfully generate a push of air toward the listener for each peak of HF applied to the voicecoil. It will generate each push coicident with each applied signal peak. And so each push of air associated with the HF signal will be separate from successive pushes of air by the period of oscillation, which is constant. The resulting sound wave experienced by the listener will be complex, but the HF component will be made up of pressure waves where each peak was generated by a push of the cone. And because the time period between successive pushes of the cone is constant, the time period of the HF experienced by the listener will also be constant. The frequency will not vary. There is no Doppler effect. Unless we move the whole speaker assembly including the magnet.
Why do some people report measurements indicating a Doppler effect ?
An idealized linear speaker can be described by a constant set of parameters. For example the cone suspension is an ideal linear spring of 'spring constant k'. In a real world non-ideal speaker the movement of the cone causes the 'spring constant' to change once it moves away from it's resting position because the suspension is not a perfect linear spring. There are many other sources of non-linearity in the system. With a source of non-linearity the application of a LF signal means that the parameters of the speaker are changing with time (remember parametric oscillators ?). When you add a HF signal into this situation it no longer has an idealized linear system to translate the signal into sound and it produces distortion. Distortion in speakers is generally much higher than the other components in the audio chain. Such distortion can produce inter-modulation products. They are not Doppler shifts.
Last edited:
Except that it does not produce a Doppler shift because the movement of the disc in an idealized case is always moving in a sinusoid with respect to the stationary magnet such that the wavelength is constant when observed by the listener also stationary with respect to the magnet. I'm at a loss to understand why so many people are 'stuck' on this
Not sure I follow you. I was talking about a disc vibrating at, say 10 kHz, being whizzed along a linear motor in proportion to a much lower frequency waveform (almost DC), in the same direction for a considerable time & distance. The 10 kHz vibration is clearly not symmetrical around a stationary point...
But I'm a non-Doppler convert, and I suggest that there is no Doppler when the speaker has a flat frequency response over all the frequencies it is being driven at. Near enough. It doesn't matter how big the speaker is, or how big its cone displacement, as long as it can shift all the air necessary to match its displacement. If it's a rubbish speaker, and air can get sucked round the back (like the vibrating disc) at low frequencies etc. then it will suffer from Doppler if driven at frequencies it is not suited for. This is because it is not effectively 'moving the medium' upon which the higher frequencies are superimposed, thus leading to 'bunching' or 'stretching' of them.
While the suspension, motor system, etc, remain linear, I can't see why there's a problem with the above.
Member
Joined 2009
Paid Member
I was going to post, but decided that many a sensible answer has already issued forth from DBmandrake.
Doppler shift may not be the correct term, semantic trolls abound here I see.
Frequency/doppler modulation perhaps.
Audible? YES. Although less than could be assumed, by virtue of the fact that few 'instruments' have dissonant harmonics, (or groups of instruments playing dissonant melodies), which would probably highlight the phenomena. Also, Since we are 'accustomed' to accepting a slight modulation of harmonic frequencies at the frequency of the fundamental. In reality, it would take a modulation of several 'cents' to be noticeable.
Phase shift due to VC inductance:
Audible?...Debatable, but more likely than not IMHO. After all no such shift occurs in nature.
Objectionable?
In FRs, alot less so than a multiway with a poor phase transition, but it IS still there.
until we have fullrange drivers that resonate the air alone, or resonate in the same way, it will always exist.
The biggest point factor is the single largest drawback of ANY speaker. transmission of the mechanical force, from the VC(exciter) to the air(media). The theory may be understood, but im pretty sure that no 2 speakers will be the same, even of the same driver model.
What we really need is a driver that is large enough that it can pressurise air at VLFs as efficiently as it can at HFs. radiation resistance....? Im not a Dr but Im also fairly sure this is actually impossible, with current technology, and maybe just impossible full stop, due to the fluid nature of air at lower velocities.
pointless discussion, food for trolls everywhere.
Maybe try listening to your speakers at altitude, or in a bathesphere, or in a HeliOx environment....pressure is the key, is it not?
Doppler shift may not be the correct term, semantic trolls abound here I see.
Frequency/doppler modulation perhaps.
Audible? YES. Although less than could be assumed, by virtue of the fact that few 'instruments' have dissonant harmonics, (or groups of instruments playing dissonant melodies), which would probably highlight the phenomena. Also, Since we are 'accustomed' to accepting a slight modulation of harmonic frequencies at the frequency of the fundamental. In reality, it would take a modulation of several 'cents' to be noticeable.
Phase shift due to VC inductance:
Audible?...Debatable, but more likely than not IMHO. After all no such shift occurs in nature.
Objectionable?
In FRs, alot less so than a multiway with a poor phase transition, but it IS still there.
until we have fullrange drivers that resonate the air alone, or resonate in the same way, it will always exist.
The biggest point factor is the single largest drawback of ANY speaker. transmission of the mechanical force, from the VC(exciter) to the air(media). The theory may be understood, but im pretty sure that no 2 speakers will be the same, even of the same driver model.
What we really need is a driver that is large enough that it can pressurise air at VLFs as efficiently as it can at HFs. radiation resistance....? Im not a Dr but Im also fairly sure this is actually impossible, with current technology, and maybe just impossible full stop, due to the fluid nature of air at lower velocities.
pointless discussion, food for trolls everywhere.
Maybe try listening to your speakers at altitude, or in a bathesphere, or in a HeliOx environment....pressure is the key, is it not?
Last edited:
Oh man, now you've done it. That's completely wrong, and shows no understanding of doppler at all. Don't feel alone though, there have even been some school books with similar serious errors.
The 'bunching' of the air, the change in air density, has nothing to do with the pitch shift due to doppler; it's a second subject. If the listener were also moving with the train and air then there would still be the same number of wavefronts being generated per sexond and the same nuber of wavefronts passing the listener and no pitch shift to his ears, right? The wavefronts don't just disappear os accumulate more and more as long as it's moving LOL. So we agree that a moving listener in the truck ahead with both trucks and the air between moving, hears no pitch shift.
But if the listener is stationary, more wavefronts will pass his ears per second, right? That's pitch shift due to doppler.
Last edited:
Electrical signal translates into force.
The force translates into acceleration.
An externally hosted image should be here but it was not working when we last tested it.
Here is how position changes with time and acceleration (upper graph):
I easily found this after searching:
Doppler Distortion: Piston Vibrating in a Tube Including the Effect of Excursion
Should provide lots of insight for the mathies.
I don't know how well any of it settles the debate, since it doesn't seem to for even the author.
Seems like it should be easily settled with FFT. If it's intermodulation you'll get a bunch of clean spikes. If it's doppler you'll get broadening of the HF center by the amount of LF FM. Clear enough to me. Tubes, no tubes. No matter. Who wants to do a lab session and report back?
The problem is all about Velocity though, so I'd suggest several mm of LF excursion for the test.
Doppler Distortion: Piston Vibrating in a Tube Including the Effect of Excursion
Should provide lots of insight for the mathies.
I don't know how well any of it settles the debate, since it doesn't seem to for even the author.
Seems like it should be easily settled with FFT. If it's intermodulation you'll get a bunch of clean spikes. If it's doppler you'll get broadening of the HF center by the amount of LF FM. Clear enough to me. Tubes, no tubes. No matter. Who wants to do a lab session and report back?
The problem is all about Velocity though, so I'd suggest several mm of LF excursion for the test.
Last edited:
Now we stuck in translation of voltage and current on the coil to acceleration of membrane and air while people who call us semantic trolls still way behind us thinking of direct translation of signals to distance.
Your velocity is somewhere in the middle.
Ok, let's think of one more experiment.
We have a box with round hole covered by membrane. Inside of the box we have 2 independent sources of waves, one with 10 Hz frequency, another with 2 KHz frequency. Obviously, since sources of sound are separated there is no Doppler effect. Right?
But membrane that cover the hole in the box vibrates with both frequencies. Does it produce Doppler effect?
Last edited:
Yes, I think you're right. Damn! I was thinking that the situation of the air moving with the truck must be different from the situation in free air, but it's really no different from the observer moving towards a stationary source and slicing through the ripples faster. I think.
But... that's DC. I still have an idea that the situation of a tweeter glued to a woofer cone must be different from a tweeter glued to an identical, but wire mesh, woofer cone. Accompanying the movement of the tweeter is a pulse of lower frequency air pressure. Does this do anything to neutralise the Doppler-type shift? Or does it have no net effect?
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.